Solutions Manual (JJC)

1 FUNCTIONS AND MODELS
1.1 Four Ways to Represent a Function

In exercises requiring estimations or approximations, your answers may vary slightly from the answers given here.
1. (a) The point (31. 32) is on the graph of 1, so 1(31) = 32.
(b) When r = 2, j is about 2.8, so 1(2) E 2.8.
(c) 1(r) = 2 is equivalent to j = 2. When j = 2, we have r = 33 and r = 1.
(d) Reasonable estimates for r when j = 0 are r = 32.5 and r = 0.3.
(e) The domain of 1 consists of all r-values on the graph of 1. For this function, the domain is 33 $ r $ 3, or [33. 3].
The range of 1 consists of all j-values on the graph of 1. For this function, the range is 32 $ j $ 3, or [32. 3].
(f ) As r increases from 31 to 3, j increases from 32 to 3. Thus, 1 is increasing on the interval [31. 3].
3. From Figure 1 in the text, the lowest point occurs at about (t. o) = (12. 385). The highest point occurs at about (17. 115).
Thus, the range of the vertical ground acceleration is 385 $ o $ 115. Written in interval notation, we get [385. 115].
5. No, the curve is not the graph of a function because a vertical line intersects the curve more than once. Hence, the curve fails
the Vertical Line Test.
7. Yes, the curve is the graph of a function because it passes the Vertical Line Test. The domain is [33. 2] and the range
is [33. 32) [31. 3].
9. The persons weight increased to about 160 pounds at age 20 and stayed fairly steady for 10 years. The persons weight
dropped to about 120 pounds for the next 5 years, then increased rapidly to about 170 pounds. The next 30 years saw a gradual
increase to 190 pounds. Possible reasons for the drop in weight at 30 years of age: diet, exercise, health problems.
11. The water will cool down almost to freezing as the ice
melts. Then, when the ice has melted, the water will
slowly warm up to room temperature.
13. Of course, this graph depends strongly on the
geographical location!
15. As the price increases, the amount sold decreases.

17.
9
10 CHAPTER 1 FUNCTIONS AND MODELS
19. (a) (b) From the graph, we estimate the number of
cell-phone subscribers worldwide to be about
92 million in 1995 and 485 million in 1999.
21. 1(r) = 3r
2
3r + 2.
1(2) = 3(2)
2
32 + 2 = 12 32 + 2 = 12.
1(32) = 3(32)
2
3(32) + 2 = 12 + 2 + 2 = 16.
1(o) = 3o
2
3o + 2.
1(3o) = 3(3o)
2
3(3o) + 2 = 3o
2
+o + 2.
1(o + 1) = 3(o + 1)
2
3(o + 1) + 2 = 3(o
2
+ 2o + 1) 3o 31 + 2 = 3o
2
+ 6o + 3 3o + 1 = 3o
2
+ 5o + 4.
21(o) = 2 1(o) = 2(3o
2
3o + 2) = 6o
2
32o + 4.
1(2o) = 3(2o)
2
3(2o) + 2 = 3(4o
2
) 32o + 2 = 12o
2
32o + 2.
1(o
2
) = 3(o
2
)
2
3(o
2
) + 2 = 3(o
4
) 3o
2
+ 2 = 3o
4
3o
2
+ 2.
[1(o)]
2
=

3o
2
3o + 2
2
=

3o
2
3o + 2
3o
2
3o + 2
= 9o
4
33o
3
+ 6o
2
33o
3
+o
2
32o + 6o
2
32o + 4 = 9o
4
36o
3
+ 13o
2
34o + 4.
1(o +/) = 3(o +/)
2
3(o +/) + 2 = 3(o
2
+ 2o/ +/
2
) 3o 3/ + 2 = 3o
2
+ 6o/ + 3/
2
3o 3/ + 2.
23. 1(r) = 4 + 3r 3r
2
, so 1(3 +/) = 4 + 3(3 +/) 3(3 +/)
2
= 4 + 9 + 3/ 3(9 + 6/ +/
2
) = 4 33/ 3/
2
,
and
1(3 +/) 31(3)
/
=
(4 33/ 3/
2
) 34
/
=
/(33 3/)
/
= 33 3/.
25.
1(r) 31(o)
r 3o
=
1
r
3
1
o
r 3o
=
o 3r
ro
r 3o
=
o 3r
ro(r 3o)
=
31(r 3o)
ro(r 3o)
= 3
1
or
27. 1(r) = r(3r 31) is dened for all r except when 0 = 3r 31 C r =
1
3
, so the domain
is

r M R | r 6=
1
3
=

3".
1
3
1
3
. "
.
29. 1(t) =
I
t +
3
I
t is dened when t D 0. These values of t give real number results for
I
t, whereas any value of t gives a real
number result for
3
I
t. The domain is [0. ").
31. /(r) = 1
4
I
r
2
35r is dened when r
2
35r 0 C r(r 35) 0. Note that r
2
35r 6= 0 since that would result in
division by zero. The expression r(r 35) is positive if r < 0 or r 5. (See Appendix A for methods for solving
inequalities.) Thus, the domain is (3". 0) (5. ").
SECTION 1.1 FOUR WAYS TO REPRESENT A FUNCTION 11
33. 1(r) = 5 is dened for all real numbers, so the domain is R, or (3". ").
The graph of 1 is a horizontal line with j-intercept 5.
35. 1(t) = t
2
36t is dened for all real numbers, so the domain is R, or
(3". "). The graph of 1 is a parabola opening upward since the coefcient
of t
2
is positive. To nd the t-intercepts, let j = 0 and solve for t.
0 = t
2
36t = t(t 36) i t = 0 and t = 6. The t-coordinate of the
vertex is halfway between the t-intercepts, that is, at t = 3. Since
1(3) = 3
2
36 3 = 39, the vertex is (3. 39).
37. o(r) =
I
r 35 is dened when r 35 D 0 or r D 5, so the domain is [5. ").
Since j =
I
r 35 i j
2
= r 35 i r = j
2
+5, we see that o is the
top half of a parabola.
39. G(r) =
3r + |r|
r
. Since |r| =
r if r D 0
3r if r < 0
, we have
G(r) =
3r +r
r
if r 0
3r 3r
r
if r < 0
=
4r
r
if r 0
2r
r
if r < 0
=
4 if r 0
2 if r < 0
Note that G is not dened for r = 0. The domain is (3". 0) (0. ").
41. 1(r) =
r + 2 if r < 0
1 3r if r D 0
The domain is R.
43. 1(r) =
r + 2 if r $ 31
r
2
if r 31
Note that for r = 31, both r + 2 and r
2
are equal to 1.
The domain is R.
45. Recall that the slope i of a line between the two points (r
1
. j
1
) and (r
2
. j
2
) is i =
j
2
3j
1
r
2
3r
1
and an equation of the line
connecting those two points is j 3j
1
= i(r 3r
1
). The slope of this line segment is
7 3(33)
5 31
=
5
2
, so an equation is
j 3(33) =
5
2
(r 31). The function is 1(r) =
5
2
r 3
11
2
, 1 $ r $ 5.
47. We need to solve the given equation for j. r + (j 31)
2
= 0 C (j 31)
2
= 3r C j 31 =
I
3r C
j = 1
I
3r. The expression with the positive radical represents the top half of the parabola, and the one with the negative
radical represents the bottom half. Hence, we want 1(r) = 1 3
I
3r. Note that the domain is r $ 0.
49. For 0 $ r $ 3, the graph is the line with slope 31 and j-intercept 3, that is, j = 3r + 3. For 3 < r $ 5, the graph is the line
with slope 2 passing through (3. 0); that is, j 30 = 2(r 33), or j = 2r 36. So the function is
1(r) =
3r + 3 if 0 $ r $ 3
2r 36 if 3 < r $ 5
51. Let the length and width of the rectangle be 1 and W. Then the perimeter is 21 + 2W = 20 and the area is = 1W.
Solving the rst equation for W in terms of 1 gives W =
20 321
2
= 10 31. Thus, (1) = 1(10 31) = 10131
2
. Since
lengths are positive, the domain of is 0 < 1 < 10. If we further restrict 1 to be larger than W, then 5 < 1 < 10 would be
the domain.
53. Let the length of a side of the equilateral triangle be r. Then by the Pythagorean Theorem, the height j of the triangle satises
j
2
+
1
2
r
2
= r
2
, so that j
2
= r
2
3
1
4
r
2
=
3
4
r
2
and j =
I
3
2
r. Using the formula for the area of a triangle,
=
1
2
(base)(height), we obtain (r) =
1
2
(r)
I
3
2
r
=
I
3
4
r
2
, with domain r 0.
55. Let each side of the base of the box have length r, and let the height of the box be /. Since the volume is 2, we know that
2 = /r
2
, so that / = 2r
2
, and the surface area is o = r
2
+ 4r/. Thus, o(r) = r
2
+ 4r(2r
2
) = r
2
+ (8r), with
domain r 0.
57. The height of the box is r and the length and width are 1 = 20 32r, W = 12 32r. Then \ = 1Wr and so
\ (r) = (20 32r)(12 32r)(r) = 4(10 3r)(6 3r)(r) = 4r(60 316r +r
2
) = 4r
3
364r
2
+ 240r.
The sides 1, W, and r must be positive. Thus, 1 0 C 20 32r 0 C r < 10;
W 0 C 12 32r 0 C r < 6; and r 0. Combining these restrictions gives us the domain 0 < r < 6.
59. (a) (b) On $14,000, tax is assessed on $4000, and 10%($4000) = $400.
On $26,000, tax is assessed on $16,000, and
10%($10,000) + 15%($6000) = $1000 + $900 = $1900.
(c) As in part (b), there is $1000 tax assessed on $20,000 of income, so
the graph of T is a line segment from (10,000. 0) to (20,000. 1000).
The tax on $30,000 is $2500, so the graph of T for r 20,000 is
the ray with initial point (20,000. 1000) that passes through
(30,000. 2500).
SECTION 1.2 MATHEMATICAL MODELS: A CATALOG OF ESSENTIAL FUNCTIONS

13
61. 1 is an odd function because its graph is symmetric about the origin. o is an even function because its graph is symmetric with
respect to the j-axis.
63. (a) Because an even function is symmetric with respect to the j-axis, and the point (5. 3) is on the graph of this even function,
the point (35. 3) must also be on its graph.
(b) Because an odd function is symmetric with respect to the origin, and the point (5. 3) is on the graph of this odd function,
the point (35. 33) must also be on its graph.
65. 1(r) =
r
r
2
+ 1
.
1(3r) =
3r
(3r)
2
+ 1
=
3r
r
2
+ 1
= 3
r
r
2
+ 1
= 31(r).
So 1 is an odd function.
67. 1(r) =
r
r + 1
, so 1(3r) =
3r
3r + 1
=
r
r 31
.
Since this is neither 1(r) nor 31(r), the function 1 is
neither even nor odd.
69. 1(r) = 1 + 3r
2
3r
4
.
1(3r) = 1 + 3(3r)
2
3(3r)
4
= 1 + 3r
2
3r
4
= 1(r).
So 1 is an even function.
1.2 Mathematical Models: A Catalog of Essential Functions
1. (a) 1(r) =
5
I
r is a root function with n = 5.
(b) o(r) =
I
1 3r
2
is an algebraic function because it is a root of a polynomial.
(c) /(r) = r
9
+r
4
is a polynomial of degree 9.
(d) v(r) =
r
2
+ 1
r
3
+r
is a rational function because it is a ratio of polynomials.
(e) c(r) = tan2r is a trigonometric function.
(f ) t(r) = log
10
r is a logarithmic function.
3. We notice from the gure that o and / are even functions (symmetric with respect to the j-axis) and that 1 is an odd function
(symmetric with respect to the origin). So (b)
j = r
5
must be 1. Since o is atter than / near the origin, we must have

(c)
j = r
8
matched with o and (a)
j = r
2
matched with /.
5. (a) An equation for the family of linear functions with slope 2
is j = 1(r) = 2r +/, where / is the j-intercept.
(b) 1(2) = 1 means that the point (2. 1) is on the graph of 1. We can use the
point-slope form of a line to obtain an equation for the family of linear
functions through the point (2. 1). j 31 = i(r 32), which is equivalent
to j = ir + (1 32i) in slope-intercept form.
(c) To belong to both families, an equation must have slope i = 2, so the equation in part (b), j = ir + (1 32i),
becomes j = 2r 33. It is the only function that belongs to both families.
7. All members of the family of linear functions 1(r) = c 3r have graphs
that are lines with slope 31. The j-intercept is c.
9. Since 1(31) = 1(0) = 1(2) = 0, 1 has zeros of 31, 0, and 2, so an equation for 1 is 1(r) = o[r 3(31)](r 30)(r 32),
or 1(r) = or(r + 1)(r 32). Because 1(1) = 6, well substitute 1 for rand 6 for 1(r).
6 = o(1)(2)(31) i 32o = 6 i o = 33, so an equation for 1 is 1(r) = 33r(r + 1)(r 32).
11. (a) 1 = 200, so c = 0.04171(o + 1) = 0.0417(200)(o + 1) = 8.34o + 8.34. The slope is 8.34, which represents the
change in mg of the dosage for a child for each change of 1 year in age.
(b) For a newborn, o = 0, so c = 8.34 mg.
13. (a) (b) The slope of
9
5
means that 1 increases
9
5
degrees for each increase
of 1
C. (Equivalently, 1 increases by 9 when C increases by 5

and 1 decreases by 9 when C decreases by 5.) The 1-intercept of
32 is the Fahrenheit temperature corresponding to a Celsius
temperature of 0.
SECTION 1.2 MATHEMATICAL MODELS: A CATALOG OF ESSENTIAL FUNCTIONS 15
15. (a) Using ` in place of r and T in place of j, we nd the slope to be
T
2
3T
1
`
2
3`
1
=
80 370
173 3113
=
10
60
=
1
6
. So a linear
equation is T 380 =
1
6
(` 3173) C T 380 =
1
6
` 3
173
6
C T =
1
6
` +
307
6
307
6
= 51.16
.
(b) The slope of
1
6
means that the temperature in Fahrenheit degrees increases one-sixth as rapidly as the number of cricket
chirps per minute. Said differently, each increase of 6 cricket chirps per minute corresponds to an increase of 1
F.
(c) When ` = 150, the temperature is given approximately by T =
1
6
(150) +
307
6
= 76.16
F E 76
F.
17. (a) We are given
change in pressure
10 feet change in depth
=
4.34
10
= 0.434. Using 1 for pressure and d for depth with the point
(d. 1) = (0. 15), we have the slope-intercept form of the line, 1 = 0.434d + 15.
(b) When 1 = 100, then 100 = 0.434d + 15 C 0.434d = 85 C d =
85
0.434
E 195.85 feet. Thus, the pressure is
100 lbin
2
at a depth of approximately 196 feet.
19. (a) The data appear to be periodic and a sine or cosine function would make the best model. A model of the form
1(r) = o cos(/r) +c seems appropriate.
(b) The data appear to be decreasing in a linear fashion. A model of the form1(r) = ir +/ seems appropriate.
Some values are given to many decimal places. These are the results given by several computer algebra systems rounding is left to the reader.
21. (a)
A linear model does seem appropriate.
(b) Using the points (4000. 14.1) and (60,000. 8.2), we obtain
j 314.1 =
8.2 314.1
60,000 34000
(r 34000) or, equivalently,
j E 30.000105357r + 14.521429.
(c) Using a computing device, we obtain the least squares regression line j = 30.0000997855r + 13.950764.
The following commands and screens illustrate how to nd the least squares regression line on a TI-83 Plus.
Enter the data into list one (L1) and list two (L2). Press to enter the editor.
Find the regession line and store it in Y1. Press .
Note from the last gure that the regression line has been stored in Y
1
and that Plot1 has been turned on (Plot1 is
highlighted). You can turn on Plot1 from the Y= menu by placing the cursor on Plot1 and pressing or by
pressing .
Now press to produce a graph of the data and the regression
line. Note that choice 9 of the ZOOM menu automatically selects a window
that displays all of the data.
(d) When r = 25,000, j E 11.456; or about 11.5 per 100 population.
(e) When r = 80,000, j E 5.968; or about a 6% chance.
(f ) When r = 200,000, j is negative, so the model does not apply.
23. (a)
A linear model does seem appropriate.
(b)
Using a computing device, we obtain the least squares
regression line j = 0.089119747r 3158.2403249,
where r is the year and j is the height in feet.
(c) When r = 2000, the model gives j E 20.00 ft. Note that the actual winning height for the 2000 Olympics is less than the
winning height for 1996so much for that prediction.
(d) When r = 2100, j E 28.91 ft. This would be an increase of 9.49 ft from 1996 to 2100. Even though there was an increase
of 8.59 ft from 1900 to 1996, it is unlikely that a similar increase will occur over the next 100 years.
25. Using a computing device, we obtain the cubic
function j = or
3
+/r
2
+cr +d with
o = 0.0012937, / = 37.06142, c = 12,823,
and d = 37,743,770. When r = 1925,
j E 1914 (million).
SECTION 1.3 NEW FUNCTIONS FROM OLD FUNCTIONS

17
1.3 New Functions from Old Functions
1. (a) If the graph of 1 is shifted 3 units upward, its equation becomes j = 1(r) + 3.
(b) If the graph of 1 is shifted 3 units downward, its equation becomes j = 1(r) 33.
(c) If the graph of 1 is shifted 3 units to the right, its equation becomes j = 1(r 33).
(d) If the graph of 1 is shifted 3 units to the left, its equation becomes j = 1(r + 3).
(e) If the graph of 1 is reected about the r-axis, its equation becomes j = 31(r).
(f ) If the graph of 1 is reected about the j-axis, its equation becomes j = 1(3r).
(g) If the graph of 1 is stretched vertically by a factor of 3, its equation becomes j = 31(r).
(h) If the graph of 1 is shrunk vertically by a factor of 3, its equation becomes j =
1
3
1(r).
3. (a) (graph 3) The graph of 1 is shifted 4 units to the right and has equation j = 1(r 34).
(b) (graph 1) The graph of 1 is shifted 3 units upward and has equation j = 1(r) + 3.
(c) (graph 4) The graph of 1 is shrunk vertically by a factor of 3 and has equation j =
1
3
1(r).
(d) (graph 5) The graph of 1 is shifted 4 units to the left and reected about the r-axis. Its equation is j = 31(r + 4).
(e) (graph 2) The graph of 1 is shifted 6 units to the left and stretched vertically by a factor of 2. Its equation is
j = 21(r + 6).
5. (a) To graph j = 1(2r) we shrink the graph of 1
horizontally by a factor of 2.
The point (4. 31) on the graph of 1 corresponds to the
point
1
2
4. 31
= (2. 31).
(b) To graph j = 1
1
2
r
we stretch the graph of 1

horizontally by a factor of 2.
point (2 4. 31) = (8. 31).
(c) To graph j = 1(3r) we reect the graph of 1 about
the j-axis.
point (31 4. 31) = (34. 31).
(d) To graph j = 31(3r) we reect the graph of 1 about
the j-axis, then about the r-axis.
point (31 4. 31 31) = (34. 1).
18

CHAPTER 1 FUNCTIONS AND MODELS
7. The graph of j = 1(r) =
I
3r 3r
2
has been shifted 4 units to the left, reected about the r-axis, and shifted downward
1 unit. Thus, a function describing the graph is
j = 31
. .. .
reect
about r-axis
1 (r + 4)
. .. .
shift
4 units left
3 1
. .. .
shift
1 unit left
This function can be written as
j = 31(r + 4) 31 = 3
3(r + 4) 3(r + 4)
2
31 = 3
3r + 12 3(r
2
+ 8r + 16) 31 = 3
I
3r
2
35r 34 31
9. j = 3r
3
: Start with the graph of j = r
3
and reect
about the r-axis. Note: Reecting about the j-axis
gives the same result since substituting 3r for r gives
us j = (3r)
3
= 3r
3
.
11. j = (r + 1)
2
: Start with the graph of j = r
2
and shift 1 unit to the left.
13. j = 1 + 2 cos r: Start with the graph of j = cos r, stretch vertically by a factor of 2, and then shift 1 unit upward.
15. j = sin(r2): Start with the graph of j = sinr and stretch horizontally by a factor of 2.
SECTION 1.3 NEW FUNCTIONS FROM OLD FUNCTIONS 19
17. j =
I
r + 3 : Start with the graph of
j =
I
r and shift 3 units to the left.
19. j =
1
2
(r
2
+ 8r) =
1
2
(r
2
+ 8r + 16) 38 =
1
2
(r + 4)
2
38: Start with the graph of j = r
2
, compress vertically by a
factor of 2, shift 4 units to the left, and then shift 8 units downward.
0 0 0 0
21. j = 2(r + 1): Start with the graph of j = 1r, shift 1 unit to the left, and then stretch vertically by a factor of 2.
23. j = |sinr|: Start with the graph of j = sinr and reect all the parts of the graph below the r-axis about the r-axis.
25. This is just like the solution to Example 4 except the amplitude of the curve (the 30
N curve in Figure 9 on June 21) is

14 312 = 2. So the function is 1(t) = 12 + 2 sin
2r
365
(t 380)
. March 31 is the 90th day of the year, so the model gives

1(90) E 12.34 h. The daylight time (5:51 AM to 6:18 PM) is 12 hours and 27 minutes, or 12.45 h. The model value differs
from the actual value by
12.45312.34
12.45
E 0.009, less than 1%.
27. (a) To obtain j = 1(|r|), the portion of the graph of j = 1(r) to the right of the j-axis is reected about the j-axis.
(b) j = sin|r| (c) j =

|r|
29. 1(r) = r
3
+ 2r
2
; o(r) = 3r
2
31. 1 = R for both 1 and o.
(1 +o)(r) = (r
3
+ 2r
2
) + (3r
2
31) = r
3
+ 5r
2
31, 1 = R.
(1 3o)(r) = (r
3
+ 2r
2
) 3(3r
2
31) = r
3
3r
2
+ 1, 1 = R.
(1o)(r) = (r
3
+ 2r
2
)(3r
2
31) = 3r
5
+ 6r
4
3r
3
32r
2
, 1 = R.
1
o
(r) =
r
3
+ 2r
2
3r
2
31
, 1 =

r | r 6=
1
I
3
since 3r
2
31 6= 0.
31. 1(r) = r
2
31, 1 = R; o(r) = 2r + 1, 1 = R.
(a) (1 o)(r) = 1(o(r)) = 1(2r + 1) = (2r + 1)
2
31 = (4r
2
+ 4r + 1) 31 = 4r
2
+ 4r, 1 = R.
(b) (o 1)(r) = o(1(r)) = o(r
2
31) = 2(r
2
31) + 1 = (2r
2
32) + 1 = 2r
2
31, 1 = R.
(c) (1 1)(r) = 1(1(r)) = 1(r
2
31) = (r
2
31)
2
31 = (r
4
32r
2
+ 1) 31 = r
4
32r
2
, 1 = R.
(d) (o o)(r) = o(o(r)) = o(2r + 1) = 2(2r + 1) + 1 = (4r + 2) + 1 = 4r + 3, 1 = R.
33. 1(r) = 1 33r; o(r) = cos r. 1 = R for both 1 and o, and hence for their composites.
(a) (1 o)(r) = 1(o(r)) = 1(cos r) = 1 33 cos r.
(b) (o 1)(r) = o(1(r)) = o(1 33r) = cos(1 33r).
(c) (1 1)(r) = 1(1(r)) = 1(1 33r) = 1 33(1 33r) = 1 33 + 9r = 9r 32.
(d) (o o)(r) = o(o(r)) = o(cos r) = cos(cos r) [Note that this is not cos r cos r.]
35. 1(r) = r +
1
r
, 1 = {r | r 6= 0}; o(r) =
r + 1
r + 2
, 1 = {r | r 6= 32}
(a) (1 o)(r) = 1(o(r)) = 1
r + 1
r + 2
=
r + 1
r + 2
+
1
r + 1
r + 2
=
r + 1
r + 2
+
r + 2
r + 1
=
(r + 1)(r + 1) + (r + 2)(r + 2)
(r + 2)(r + 1)
=
r
2
+ 2r + 1
r
2
+ 4r + 4
(r + 2)(r + 1)
=
2r
2
+ 6r + 5
(r + 2)(r + 1)
Since o(r) is not dened for r = 32 and 1(o(r)) is not dened for r = 32 and r = 31,
the domain of (1 o)(r) is 1 = {r | r 6= 32. 31}.
(b) (o 1)(r) = o(1(r)) = o
r +
1
r
r +
1
r
+ 1
r +
1
r
+ 2
=
r
2
+ 1 +r
r
r
2
+ 1 + 2r
r
=
r
2
+r + 1
r
2
+ 2r + 1
=
r
2
+r + 1
(r + 1)
2
Since 1(r) is not dened for r = 0 and o(1(r)) is not dened for r = 31,
the domain of (o 1)(r) is 1 = {r | r 6= 31. 0}.
(c) (1 1)(r) = 1(1(r)) = 1
r +
1
r
=

r +
1
r
+
1
r +
1
i
= r +
1
r
+
1
i
2
+1
i
= r +
1
r
+
r
r
2
+ 1
=
r(r)
r
2
+ 1
+ 1
r
2
+ 1
+r(r)
r(r
2
+ 1)
=
r
4
+r
2
+r
2
+ 1 +r
2
r(r
2
+ 1)
=
r
4
+ 3r
2
+ 1
r(r
2
+ 1)
. 1 = {r | r 6= 0}
SECTION 1.3 NEW FUNCTIONS FROM OLD FUNCTIONS 21
(d) (o o)(r) = o(o(r)) = o
r + 1
r + 2
=
r + 1
r + 2
+ 1
r + 1
r + 2
+ 2
=
r + 1 + 1(r + 2)
r + 2
r + 1 + 2(r + 2)
r + 2
=
r + 1 +r + 2
r + 1 + 2r + 4
=
2r + 3
3r + 5
Since o(r) is not dened for r = 32 and o(o(r)) is not dened for r = 3
5
3
,
the domain of (o o)(r) is 1 =

r | r 6= 32. 3
5
3
.
37. (1 o /)(r) = 1(o(/(r))) = 1(o(r 31)) = 1(2(r 31)) = 2(r 31) + 1 = 2r 31
39. (1 o /)(r) = 1(o(/(r))) = 1(o(r
3
+ 2)) = 1[(r
3
+ 2)
2
]
= 1(r
6
+ 4r
3
+ 4) =

(r
6
+ 4r
3
+ 4) 33 =
I
r
6
+ 4r
3
+ 1
41. Let o(r) = r
2
+ 1 and 1(r) = r
10
. Then (1 o)(r) = 1(o(r)) = 1(r
2
+ 1) = (r
2
+ 1)
10
= 1(r).
43. Let o(r) =
3
I
r and 1(r) =
r
1 +r
. Then (1 o)(r) = 1(o(r)) = 1(
3
I
r) =
3
I
r
1 +
3
I
r
= 1(r).
45. Let o(t) = cos t and 1(t) =
I
t. Then (1 o)(t) = 1(o(t)) = 1(cos t) =
I
cos t = &(t).
47. Let /(r) = r
2
, o(r) = 3
i
, and 1(r) = 1 3 r. Then
(1 o /)(r) = 1(o(/(r))) = 1(o(r
2
)) = 1
3
i
2
= 1 33
i
2
= 1(r).
49. Let /(r) =
I
r, o(r) = sec r, and 1(r) = r
4
. Then
(1 o /)(r) = 1(o(/(r))) = 1(o(
I
r)) = 1(sec
I
r) = (sec
I
r)
4
= sec
4
(
I
r) = 1(r).
51. (a) o(2) = 5, because the point (2. 5) is on the graph of o. Thus, 1(o(2)) = 1(5) = 4, because the point (5. 4) is on the
graph of 1.
(b) o(1(0)) = o(0) = 3
(c) (1 o)(0) = 1(o(0)) = 1(3) = 0
(d) (o 1)(6) = o(1(6)) = o(6). This value is not dened, because there is no point on the graph of o that has
r-coordinate 6.
(e) (o o)(32) = o(o(32)) = o(1) = 4
(f ) (1 1)(4) = 1(1(4)) = 1(2) = 32
53. (a) Using the relationship distance = rate time with the radius v as the distance, we have v(t) = 60t.
(b) = v
2
i ( v)(t) = (v(t)) = (60t)
2
= 3600t
2
. This formula gives us the extent of the rippled area
(in cm
2
) at any time t.
55. (a) From the gure, we have a right triangle with legs 6 and d, and hypotenuse c.
By the Pythagorean Theorem, d
2
+ 6
2
= c
2
i c = 1(d) =
I
d
2
+ 36.
(b) Using d = vt, we get d = (30 kmhr)(t hr) = 30t (in km). Thus,
d = o(t) = 30t.
(c) (1 o)(t) = 1(o(t)) = 1(30t) =

(30t)
2
+ 36 =
I
900t
2
+ 36. This function represents the distance between the
lighthouse and the ship as a function of the time elapsed since noon.
57. (a)
1(t) =
0 if t < 0
1 if t D 0
(b)
\ (t) =
0 if t < 0
120 if t D 0
so \ (t) = 1201(t).
(c) Starting with the formula in part (b), we replace 120 with 240 to reect the
different voltage. Also, because we are starting 5 units to the right of t = 0,
we replace t with t 35. Thus, the formula is \ (t) = 2401(t 35).
59. If 1(r) = i
1
r + /
1
and o(r) = i
2
r + /
2
, then
(1 o)(r) = 1(o(r)) = 1(i2r +/2) = i1(i2r +/2) +/1 = i1i2r +i1/2 +/1.
So 1 o is a linear function with slope i
1
i
2
.
61. (a) By examining the variable terms in o and /, we deduce that we must square o to get the terms 4r
2
and 4r in /. If we let
1(r) = r
2
+c, then (1 o)(r) = 1(o(r)) = 1(2r + 1) = (2r + 1)
2
+c = 4r
2
+ 4r + (1 +c). Since
/(r) = 4r
2
+ 4r + 7, we must have 1 +c = 7. So c = 6 and 1(r) = r
2
+ 6.
(b) We need a function o so that 1(o(r)) = 3(o(r)) + 5 = /(r). But
/(r) = 3r
2
+ 3r + 2 = 3(r
2
+r) + 2 = 3(r
2
+r 31) + 5, so we see that o(r) = r
2
+r 31.
63. (a) If 1 and o are even functions, then 1(3r) = 1(r) and o(3r) = o(r).
(i) (1 +o)(3r) = 1(3r) +o(3r) = 1(r) +o(r) = (1 +o)(r), so 1 +o is an even function.
(ii) (1o)(3r) = 1(3r) o(3r) = 1(r) o(r) = (1o)(r), so 1o is an even function.
(b) If 1 and o are odd functions, then 1(3r) = 31(r) and o(3r) = 3o(r).
(i) (1 +o)(3r) = 1(3r) +o(3r) = 31(r) + [3o(r)] = 3[1(r) +o(r)] = 3(1 +o)(r),
so 1 +o is an odd function.
(ii) (1o)(3r) = 1(3r) o(3r) = 31(r) [3o(r)] = 1(r) o(r) = (1o)(r), so 1o is an even function.
65. We need to examine /(3r).
/(3r) = (1 o)(3r) = 1(o(3r)) = 1(o(r)) [because o is even] = /(r)
Because /(3r) = /(r), / is an even function.
SECTION 1.4 GRAPHING CALCULATORS AND COMPUTERS

23
1.4 Graphing Calculators and Computers
1. 1(r) =
I
r
3
35r
2
(a) [35. 5] by [35. 5]
(There is no graph shown.)
(b) [0. 10] by [0. 2] (c) [0. 10] by [0. 10]
The most appropriate graph is produced in viewing rectangle (c).
3. Since the graph of 1(r) = 5 + 20r 3r
2
is a
parabola opening downward, an appropriate viewing
rectangle should include the maximum point.
5. 1(r) =
4
I
81 3r
4
is dened when 81 3r
4
D 0 C
r
4
$ 81 C |r| $ 3, so the domain of 1 is [33. 3]. Also
0 $
4
I
81 3r
4
$
4
I
81 = 3, so the range is [0. 3].
7. The graph of 1(r) = r
3
3225r is symmetric with respect to the origin.
Since 1(r) = r
3
3225r = r(r
2
3225) = r(r + 15)(r 315), there
are r-intercepts at 0, 315, and 15. 1(20) = 3500.
9. The period of o(r) = sin(1000r) is
2r
1000
E 0.0063 and its range is
[31. 1]. Since 1(r) = sin
2
(1000r) is the square of o, its range is
[0. 1] and a viewing rectangle of [30.01. 0.01] by [0. 1.1] seems
appropriate.
11. The domain of j =
I
r is r D 0, so the domain of 1(r) = sin
I
r is [0. ")
and the range is [31. 1]. With a little trial-and-error experimentation, we nd
that an Xmax of 100 illustrates the general shape of 1, so an appropriate
viewing rectangle is [0. 100] by [31.5. 1.5].
13. The rst term, 10 sinr, has period 2 and range [310. 10]. It will be the dominant term in any large graph of
j = 10 sinr + sin100r, as shown in the rst gure. The second term, sin100r, has period
2r
100
=
r
50
and range [31. 1].
It causes the bumps in the rst gure and will be the dominant term in any small graph, as shown in the view near the
origin in the second gure.
15. We must solve the given equation for j to obtain equations for the upper and
lower halves of the ellipse.
4r
2
+ 2j
2
= 1 C 2j
2
= 1 34r
2
C j
2
=
1 34r
2
2
C
j =
1 34r
2
2
17. From the graph of j = 3r
2
36r +1
and j = 0.23r 32.25 in the viewing
rectangle [31. 3] by [32.5. 1.5], it is
difcult to see if the graphs intersect.
If we zoom in on the fourth quadrant,
we see the graphs do not intersect.
19. From the graph of 1(r) = r
3
39r
2
34, we see that there is one solution
of the equation 1(r) = 0 and it is slightly larger than 9. By zooming in or
using a root or zero feature, we obtain r E 9.05.
21. We see that the graphs of 1(r) = r
2
and o(r) = sinr intersect twice. One
solution is r = 0. The other solution of 1 = o is the r-coordinate of the
point of intersection in the rst quadrant. Using an intersect feature or
zooming in, we nd this value to be approximately 0.88. Alternatively, we
could nd that value by nding the positive zero of /(r) = r
2
3sinr.
Note: After producing the graph on a TI-83 Plus, we can nd the approximate value 0.88 by using the following keystrokes:
. The 1 is just a guess for 0.88.
SECTION 1.4 GRAPHING CALCULATORS AND COMPUTERS 25
23. o(r) = r
3
10 is larger than 1(r) = 10r
2
whenever r 100.
25. We see from the graphs of j = |sinr 3r| and j = 0.1 that there are
two solutions to the equation |sinr 3r| = 0.1: r E 30.85 and
r E 0.85. The condition |sinr 3r| < 0.1 holds for any r lying
between these two values, that is, 30.85 < r < 0.85.
27. (a) The root functions j =
I
r,
j =
4
I
r and j =
6
I
r
(b) The root functions j = r,
j =
3
I
r and j =
5
I
r
(c) The root functions j =
I
r, j =
3
I
r,
j =
4
I
r and j =
5
I
r
(d) For any n, the nth root of 0 is 0 and the nth root of 1 is 1; that is, all nth root functions pass through the points (0. 0)
and (1. 1).
For odd n, the domain of the nth root function is R, while for even n, it is {r M R | r D 0}.
Graphs of even root functions look similar to that of
I
r, while those of odd root functions resemble that of
3
I
r.
As n increases, the graph of
q
I
r becomes steeper near 0 and atter for r 1.
29. 1(r) = r
4
+cr
2
+r. If c < 31.5, there are three humps: two minimum points
and a maximum point. These humps get atter as c increases, until at c = 31.5
two of the humps disappear and there is only one minimum point. This single
hump then moves to the right and approaches the origin as c increases.
31. j = r
n
2
3i
. As n increases, the maximum of the
function moves further from the origin, and gets
larger. Note, however, that regardless of n, the
function approaches 0 as r <".
33. j
2
= cr
3
+r
2
. If c < 0, the loop is to the right of the origin, and if c is positive,
it is to the left. In both cases, the closer c is to 0, the larger the loop is. (In the
limiting case, c = 0, the loop is innite, that is, it doesnt close.) Also, the
larger |c| is, the steeper the slope is on the loopless side of the origin.
35. The graphing window is 95 pixels wide and we want to start with r = 0 and end with r = 2. Since there are 94 gaps
between pixels, the distance between pixels is
2r30
94
. Thus, the r-values that the calculator actually plots are r = 0 +
2r
94
n,
where n = 0, 1, 2, . . . , 93, 94. For j = sin2r, the actual points plotted by the calculator are
2r
94
n. sin
2
2r
94
n
for
n = 0, 1, . . . , 94. For j = sin96r, the points plotted are
2r
94
n. sin
96
2r
94
n
for n = 0, 1, . . . , 94. But

sin
96
2r
94
n
= sin
94
2r
94
n + 2
2r
94
n
= sin
2n + 2
2r
94
n
= sin
2
2r
94
n
[by periodicity of sine]. n = 0, 1, . . . , 94

So the j-values, and hence the points, plotted for j = sin96r are identical to those plotted for j = sin2r.
Note: Try graphing j = sin94r. Can you see why all the j-values are zero?
1.5 Exponential Functions
1. (a) 1(r) = o
i
, o 0 (b) R (c) (0. ") (d) See Figures 4(c), 4(b), and 4(a), respectively.
3. All of these graphs approach 0 as r <3", all of them pass through the point
(0. 1), and all of them are increasing and approach "as r <". The larger the
base, the faster the function increases for r 0, and the faster it approaches 0 as
r <3".
Note: The notation r <" can be thought of as r becomes large at this point.
More details on this notation are given in Chapter 2.
5. The functions with bases greater than 1 (3
i
and 10
i
) are increasing, while those
with bases less than 1
1
3
i
and

1
10
are decreasing. The graph of
1
3
i
is the
reection of that of 3
i
about the j-axis, and the graph of
1
10
i
is the reection of
that of 10
i
about the j-axis. The graph of 10
i
increases more quickly than that of
3
i
for r 0, and approaches 0 faster as r <3".
7. We start with the graph of j = 4
i
(Figure 3) and then
shift 3 units downward. This shift doesnt affect the
domain, but the range of j = 4
i
33 is (33. ") .
There is a horizontal asymptote of j = 33.
j = 4
i
j = 4
i
33
SECTION 1.5 EXPONENTIAL FUNCTIONS 27
9. We start with the graph of j = 2
i
(Figure 3),
reect it about the j-axis, and then about the
r-axis (or just rotate 180
to handle both
reections) to obtain the graph of j = 32
3i
.
In each graph, j = 0 is the horizontal
asymptote.
j = 2
i
j = 2
3i
j = 32
3i
11. We start with the graph of j = c
i
(Figure 13) and reect about the j-axis to get the graph of j = c
3i
. Then we compress the
graph vertically by a factor of 2 to obtain the graph of j =
1
2
c
3i
and then reect about the r-axis to get the graph of
j = 3
1
2
c
3i
. Finally, we shift the graph upward one unit to get the graph of j = 1 3
1
2
c
3i
.
13. (a) To nd the equation of the graph that results from shifting the graph of j = c
i
2 units downward, we subtract 2 from the
original function to get j = c
i
32.
(b) To nd the equation of the graph that results from shifting the graph of j = c
i
2 units to the right, we replace r with r 32
in the original function to get j = c
(i32)
.
(c) To nd the equation of the graph that results from reecting the graph of j = c
i
about the r-axis, we multiply the original
function by 31 to get j = 3c
i
.
(d) To nd the equation of the graph that results from reecting the graph of j = c
i
about the j-axis, we replace r with 3r in
the original function to get j = c
3i
.
(e) To nd the equation of the graph that results from reecting the graph of j = c
i
about the r-axis and then about the
j-axis, we rst multiply the original function by 31 (to get j = 3c
i
) and then replace r with 3r in this equation to
get j = 3c
3i
.
15. (a) The denominator 1 +c
i
is never equal to zero because c
i
0, so the domain of 1(r) = 1(1 +c
i
) is R.
(b) 1 3c
i
= 0 C c
i
= 1 C r = 0, so the domain of 1(r) = 1(1 3c
i
) is (3". 0) (0. ").
17. Use j = Co
i
with the points (1. 6) and (3. 24). 6 = Co
1

C =
6
a
and 24 = Co
3
i 24 =

6
o
o
3
i
4 = o
2
i o = 2 [since o 0] and C =
6
2
= 3. The function is 1(r) = 3 2
i
.
19. If 1(r) = 5
i
, then
1(r +/) 31(r)
/
=
5
i+I
35
i
/
=
5
i
5
I
35
i
/
=
5
i
5
I
31
/
= 5
i
5
I
31
/
.
21. 2 ft = 24 in, 1(24) = 24
2
in = 576 in = 48 ft. o(24) = 2
24
in = 2
24
(12 5280) mi E 265 mi
23. The graph of o nally surpasses that of 1 at r E 35.8.
25. (a) Fifteen hours represents 5 doubling periods (one doubling period is three hours). 100 2
5
= 3200
(b) In t hours, there will be t3 doubling periods. The initial population is 100,
so the population j at time t is j = 100 2
I'3
.
(c) t = 20 i j = 100 2
20'3
E 10,159
(d) We graph j
1
= 100 2
i'3
and j
2
= 50,000. The two curves intersect at
r E 26.9, so the population reaches 50,000 in about 26.9 hours.
27. An exponential model is j = o/
I
, where o = 3.154832569 10
312
and / = 1.017764706. This model gives j(1993) E 5498 million and
j(2010) E 7417 million.
29. From the graph, it appears that 1 is an odd function (1 is undened for r = 0).
To prove this, we must show that 1(3r) = 31(r).
1(3r) =
1 3c
1'(3i)
1 +c
1'(3i)
=
1 3c
(31'i)
1 +c
(31'i)
=
1 3
1
c
1'i
1 +
1
c
1'i
c
1'i
c
1'i
=
c
1'i
31
c
1'i
+ 1
= 3
1 3c
1'i
1 +c
1'i
= 31(r)
so 1 is an odd function.
1.6 Inverse Functions and Logarithms
1. (a) See Denition 1.
(b) It must pass the Horizontal Line Test.
3. 1 is not one-to-one because 2 6= 6, but 1(2) = 2.0 = 1(6).
5. No horizontal line intersects the graph of 1 more than once. Thus, by the Horizontal Line Test, 1 is one-to-one.
SECTION 1.6 INVERSE FUNCTIONS AND LOGARITHMS 29
7. The horizontal line j = 0 (the r-axis) intersects the graph of 1 in more than one point. Thus, by the Horizontal Line Test,
1 is not one-to-one.
9. The graph of 1(r) = r
2
32r is a parabola with axis of symmetry r = 3
/
2o
= 3
32
2(1)
= 1. Pick any r-values equidistant
from 1 to nd two equal function values. For example, 1(0) = 0 and 1(2) = 0, so 1 is not one-to-one.
11. o(r) = 1r. r
1
6= r
2
i 1r
1
6= 1r
2
i o (r
1
) 6= o (r
2
), so o is one-to-one.
Geometric solution: The graph of o is the hyperbola shown in Figure 14 in Section 1.2. It passes the Horizontal Line Test,
so o is one-to-one.
13. A football will attain every height / up to its maximum height twice: once on the way up, and again on the way down. Thus,
even if t
1
does not equal t
2
, 1(t
1
) may equal 1(t
2
), so 1 is not 1-1.
15. Since 1(2) = 9 and 1 is 1-1, we know that 1
31
(9) = 2. Remember, if the point (2. 9) is on the graph of 1, then the point
(9. 2) is on the graph of 1
31
.
17. First, we must determine r such that o(r) = 4. By inspection, we see that if r = 0, then o(r) = 4. Since o is 1-1 (o is an
increasing function), it has an inverse, and o
31
(4) = 0.
19. We solve C =
5
9
(1 332) for 1:
9
5
C = 1 332 i 1 =
9
5
C + 32. This gives us a formula for the inverse function, that
is, the Fahrenheit temperature 1 as a function of the Celsius temperature C. 1 D 3459.67 i
9
5
C + 32 D 3459.67 i
9
5
C D 3491.67 i C D 3273.15, the domain of the inverse function.
21. 1(r) =
I
10 33r i j =
I
10 33r (j D 0) i j
2
= 10 33r i 3r = 10 3j
2
i r = 3
1
3
j
2
+
10
3
.
Interchange r and j: j = 3
1
3
r
2
+
10
3
. So 1
31
(r) = 3
1
3
r
2
+
10
3
. Note that the domain of 1
31
is r D 0.
23. j = 1(r) = c
i
3
i lnj = r
3
i r =
3
I
lnj. Interchange r and j: j =
3
I
lnr. So 1
31
(r) =
3
I
lnr.
25. j = 1(r) = ln(r + 3) i r + 3 = c
u
i r = c
u
33. Interchange r and j: j = c
i
33. So 1
31
(r) = c
i
33.
27. j = 1(r) = r
4
+ 1 i j 31 = r
4
i r =
4
I
j 31 (not since
r D 0). Interchange x and y: j =
4
I
r 31. So 1
31
(r) =
4
I
r 31. The
graph of j =
4
I
r 31 is just the graph of j =
4
I
r shifted right one unit.
From the graph, we see that 1 and 1
31
are reections about the line j = r.
29. Reect the graph of 1 about the line j = r. The points (31. 32), (1. 31),
(2. 2), and (3. 3) on 1 are reected to (32. 31), (31. 1), (2. 2), and (3. 3)
on 1
31
.
31. (a) It is dened as the inverse of the exponential function with base o, that is, log
a
r = j C o
u
= r.
(b) (0. ") (c) R (d) See Figure 11.
33. (a) log
5
125 = 3 since 5
3
= 125. (b) log
3
1
27
= 33 since 3
33
=
1
3
3
=
1
27
.
35. (a) log
2
6 3log
2
15 + log
2
20 = log
2
(
6
15
) + log
2
20 [by Law 2]
= log
2
(
6
15
20) [by Law 1]
= log
2
8, and log
2
8 = 3 since 2
3
= 8.
(b) log
3
100 3log
3
18 3log
3
50 = log
3
100
18
3log
3
50 = log
3
100
1850
= log
3
(
1
9
), and log
3
1
9
= 32 since 3
32
=
1
9
.
37. ln5 + 5 ln3 = ln5 + ln3
5
[by Law 3]
= ln(5 3
5
) [by Law 1]
= ln1215
39. ln(1 +r
2
) +
1
2
lnr 3lnsinr = ln(1 +r
2
) + lnr
1'2
3lnsinr = ln[(1 +r
2
)
I
r] 3lnsinr = ln
(1 +r
2
)
I
r
sinr
41. To graph these functions, we use log
1.5
r =
lnr
ln1.5
and log
50
r =
lnr
ln50
.
These graphs all approach 3"as r <0
+
, and they all pass through the
point (1. 0). Also, they are all increasing, and all approach "as r <".
The functions with larger bases increase extremely slowly, and the ones with
smaller bases do so somewhat more quickly. The functions with large bases
approach the j-axis more closely as r <0
+
.
43. 3 ft = 36 in, so we need r such that log
2
r = 36 C r = 2
36
= 68,719,476,736. In miles, this is
68,719,476,736 in
1 ft
12 in

1 mi
5280 ft
E 1,084,587.7 mi.
45. (a) Shift the graph of j = log
10
r ve units to the left to
obtain the graph of j = log
10
(r +5). Note the vertical
asymptote of r = 35.
j = log
10
r j = log
10
(r + 5)
(b) Reect the graph of j = lnr about the r-axis to obtain
the graph of j = 3lnr.
j = lnr j = 3lnr
47. (a) 2 lnr = 1 i lnr =
1
2
i r = c
1'2
=
I
c
(b) c
3i
= 5 i 3r = ln5 i r = 3ln5
49. (a) 2
i35
= 3 C log
2
3 = r 35 C r = 5 + log
2
3.
Or: 2
i35
= 3 C ln
2
i35
= ln3 C (r 35) ln2 = ln3 C r 35 =

ln3
ln2
C r = 5 +
ln3
ln2
SECTION 1.6 INVERSE FUNCTIONS AND LOGARITHMS 31
(b) lnr + ln(r 31) = ln(r(r 31)) = 1 C r(r 31) = c
1
C r
2
3r 3c = 0. The quadratic formula (with o = 1,
/ = 31, and c = 3c) gives r =
1
2
1
I
1 + 4c
, but we reject the negative root since the natural logarithm is not
dened for r < 0. So r =
1
2
1 +
I
1 + 4c
.
51. (a) c
i
< 10 i lnc
i
< ln10 i r < ln10 i r M (3". ln10)
(b) lnr 31 i c
ln i
c
31
i r c
31
i r M (1c. ")
53. (a) For 1(r) =
I
3 3c
2i
, we must have 3 3c
2i
D 0 i c
2i
$ 3 i 2r $ ln3 i r $
1
2
ln3. Thus, the domain
of 1 is (3".
1
2
ln3].
(b) j = 1(r) =
I
3 3c
2i
[note that j D 0] i j
2
= 3 3c
2i
i c
2i
= 3 3j
2
i 2r = ln(3 3j
2
) i
r =
1
2
ln(3 3j
2
). Interchange r and j: j =
1
2
ln(3 3r
2
). So 1
31
(r) =
1
2
ln(3 3r
2
). For the domain of 1
31
, we must
have 3 3r
2
0 i r
2
< 3 i |r| <
I
3 i 3
I
3 < r <
I
3 i 0 $ r <
I
3 since r D 0. Note that the
domain of 1
31
, [0.
I
3 ), equals the range of 1.
55. We see that the graph of j = 1(r) =
I
r
3
+r
2
+r + 1 is increasing, so 1 is 1-1.
Enter r =

j
3
+j
2
+j + 1 and use your CAS to solve the equation for j.
Using Derive, we get two (irrelevant) solutions involving imaginary expressions,
as well as one which can be simplied to the following:
j = 1
31
(r) = 3
3
I
4
6
3
I
1327r
2
+ 20 3
3
I
1 + 27r
2
320 +
3
I
2
where 1 = 3
I
3
I
27r
4
340r
2
+ 16.
Maple and Mathematica each give two complex expressions and one real expression, and the real expression is equivalent
to that given by Derive. For example, Maples expression simplies to
1
6
A
2'3
38 32A
1'3
2A
1'3
, where
A = 108r
2
+ 12
I
48 3120r
2
+ 81r
4
380.
57. (a) n = 100 2
I'3
i
n
100
= 2
I'3
i log
2
n
100
=
t
3
i t = 3 log
2
n
100
. Using formula (10), we can write

this as t = 1
31
(n) = 3
ln(n100)
ln2
. This function tells us how long it will take to obtain n bacteria (given the number n).
(b) n = 50,000 i t = 1
31
(50,000) = 3
ln
50,000
100
ln2
= 3
ln500
ln2
E 26.9 hours
59. (a) sin
31
I
3
2
=
r
3
since sin
r
3
=
I
3
2
and
r
3
is in
3
r
2
.
r
2
.
(b) cos
31
(31) = since cos = 31 and is in [0. ].
61. (a) arctan1 =
r
4
since tan
r
4
= 1 and
r
4
is in
3
r
2
.
r
2
.
(b) sin
31 1
I
2
=
r
4
since sin
r
4
=
1
I
2
and
r
4
is in
3
r
2
.
r
2
.
63. (a) In general, tan(arctanr) = r for any real number r. Thus, tan(arctan10) = 10.
(b) sin
31
sin
7r
3
= sin
31
sin
r
3
= sin
31
I
3
2
=
r
3
since sin
r
3
=
I
3
2
and
r
3
is in
3
r
2
.
r
2
.
[Recall that
7r
3
=
r
3
+ 2 and the sine function is periodic with period 2.]
65. Let j = sin
31
r. Then 3
r
2
$ j $
r
2
i cos j D 0, so cos(sin
31
r) = cos j =

1 3sin
2
j =
I
1 3r
2
.
67. Let j = tan
31
r. Then tanj = r, so from the triangle we see that
sin(tan
31
r) = sinj =
r
I
1 +r
2
.
69. The graph of sin
31
r is the reection of the graph of
sinr about the line j = r.
71. o(r) = sin
31
(3r + 1).
Domain (o) = {r | 31 $ 3r + 1 $ 1} = {r | 32 $ 3r $ 0} =

r | 3
2
3
$ r $ 0
=

3
2
3
. 0
.
Range (o) =
j | 3
r
2
$ j $
r
2
=

3
r
2
.
r
2
.
73. (a) If the point (r. j) is on the graph of j = 1(r), then the point (r 3c. j) is that point shifted c units to the left. Since 1 is
1-1, the point (j. r) is on the graph of j = 1
31
(r) and the point corresponding to (r 3c. j) on the graph of 1 is
(j. r 3c) on the graph of 1
31
. Thus, the curves reection is shifted down the same number of units as the curve itself is
shifted to the left. So an expression for the inverse function is o
31
(r) = 1
31
(r) 3c.
(b) If we compress (or stretch) a curve horizontally, the curves reection in the line j = r is compressed (or stretched)
vertically by the same factor. Using this geometric principle, we see that the inverse of /(r) = 1(cr) can be expressed as
/
31
(r) = (1c) 1
31
(r).
CHAPTER 1 REVIEW 33
1 Review
1. (a) A function 1 is a rule that assigns to each element r in a set exactly one element, called 1(r), in a set 1. The set is
called the domain of the function. The range of 1 is the set of all possible values of 1(r) as r varies throughout the
domain.
(b) If 1 is a function with domain , then its graph is the set of ordered pairs {(r. 1(r)) | r M }.
(c) Use the Vertical Line Test on page 16.
2. The four ways to represent a function are: verbally, numerically, visually, and algebraically. An example of each is given
below.
Verbally: An assignment of students to chairs in a classroom (a description in words)
Numerically: A tax table that assigns an amount of tax to an income (a table of values)
Visually: A graphical history of the Dow Jones average (a graph)
Algebraically: A relationship between distance, rate, and time: d = vt (an explicit formula)
3. (a) An even function 1 satises 1(3r) = 1(r) for every number r in its domain. It is symmetric with respect to the j-axis.
(b) An odd function o satises o(3r) = 3o(r) for every number r in its domain. It is symmetric with respect to the origin.
4. A function 1 is called increasing on an interval 1 if 1(r1) < 1(r2) whenever r1 < r2 in 1.
5. A mathematical model is a mathematical description (often by means of a function or an equation) of a real-world
phenomenon.
6. (a) Linear function: 1(r) = 2r + 1, 1(r) = or +/ 7.
(b) Power function: 1(r) = r
2
, 1(r) = r
a
(c) Exponential function: 1(r) = 2
i
, 1(r) = o
i
(d) Quadratic function: 1(r) = r
2
+r + 1, 1(r) = or
2
+/r +c
(e) Polynomial of degree 5: 1(r) = r
5
+ 2
(f ) Rational function: 1(r) =
r
r + 2
, 1(r) =
1(r)
Q(r)
where 1(r) and
Q(r) are polynomials
8. (a) (b)
(c) (d) (e)
(f ) (g) (h)
9. (a) The domain of 1 +o is the intersection of the domain of 1 and the domain of o; that is, K 1.
(b) The domain of 1o is also K 1.
(c) The domain of 1o must exclude values of r that make o equal to 0; that is, {r M K 1 | o(r) 6= 0}.
10. Given two functions 1 and o, the composite function 1 o is dened by (1 o) (r) = 1(o (r)). The domain of 1 o is the
set of all r in the domain of o such that o(r) is in the domain of 1.
11. (a) If the graph of 1 is shifted 2 units upward, its equation becomes j = 1(r) + 2.
(b) If the graph of 1 is shifted 2 units downward, its equation becomes j = 1(r) 32.
(c) If the graph of 1 is shifted 2 units to the right, its equation becomes j = 1(r 32).
(d) If the graph of 1 is shifted 2 units to the left, its equation becomes j = 1(r + 2).
(e) If the graph of 1 is reected about the r-axis, its equation becomes j = 31(r).
(f ) If the graph of 1 is reected about the j-axis, its equation becomes j = 1(3r).
(g) If the graph of 1 is stretched vertically by a factor of 2, its equation becomes j = 21(r).
(h) If the graph of 1 is shrunk vertically by a factor of 2, its equation becomes j =
1
2
1(r).
(i) If the graph of 1 is stretched horizontally by a factor of 2, its equation becomes j = 1
1
2
r
.
(j) If the graph of 1 is shrunk horizontally by a factor of 2, its equation becomes j = 1(2r).
12. (a) A function 1 is called a one-to-one function if it never takes on the same value twice; that is, if 1(r
1
) 6= 1(r
2
) whenever
r
1
6= r
2
. (Or, 1 is 1-1 if each output corresponds to only one input.)
Use the Horizontal Line Test: A function is one-to-one if and only if no horizontal line intersects its graph more than
once.
(b) If 1 is a one-to-one function with domain and range 1, then its inverse function 1
31
has domain 1 and range and is
dened by
1
31
(j) = r C 1(r) = j
for any j in 1. The graph of 1
31
is obtained by reecting the graph of 1 about the line j = r.
CHAPTER 1 REVIEW 35
13. (a) The inverse sine function 1(r) = sin
31
r is dened as follows:
sin
31
r = j C sinj = r and 3
2
$ j $

2
Its domain is 31 $ r $ 1 and its range is 3
2
$ j $

2
.
(b) The inverse cosine function 1(r) = cos
31
cos
31
r = j C cos j = r and 0 $ j $
Its domain is 31 $ r $ 1 and its range is 0 $ j $ .
(c) The inverse tangent function 1(r) = tan
31
tan
31
r = j C tanj = r and 3
2
< j <

2
Its domain is R and its range is 3
2
< j <

2
.
1. False. Let 1(r) = r
2
, c = 31, and t = 1. Then 1(c +t) = (31 + 1)
2
= 0
2
= 0, but
1(c) +1(t) = (31)
2
+ 1
2
= 2 6= 0 = 1(c +t).
2
. Then 1(3r) = (3r)
2
= 9r
2
and 31(r) = 3r
2
. So 1(3r) 6= 31(r).
5. True. See the Vertical Line Test.
3
. Then 1 is one-to-one and 1
31
(r) =
3
I
r. But 11(r) = 1r
3
, which is not equal to 1
31
(r).
9. True. The function lnr is an increasing function on (0. ").
11. False. Let r = c
2
and o = c. Then
lnr
lno
=
lnc
2
lnc
=
2 lnc
lnc
= 2 and ln
r
o
= ln
c
2
c
= lnc = 1, so in general the statement
is false. What is true, however, is that ln
r
o
= lnr 3lno.
13. False. For example, tan
31
20 is dened; sin
31
20 and cos
31
20 are not.
1. (a) When r = 2, j E 2.7. Thus, 1(2) E 2.7.
(b) 1(r) = 3 i r E 2.3, 5.6
(c) The domain of 1 is 36 $ r $ 6, or [36. 6].
(d) The range of 1 is 34 $ j $ 4, or [34. 4].
(e) 1 is increasing on [34. 4], that is, on 34 $ r $ 4.
(f ) 1 is not one-to-one since it fails the Horizontal Line Test.
(g) 1 is odd since its graph is symmetric about the origin.
3. 1(r) = r
2
32r + 3, so 1(o +/) = (o +/)
2
32(o +/) + 3 = o
2
+ 2o/ +/
2
32o 32/ + 3, and
1(o +/) 31(o)
/
=
(o
2
+ 2o/ +/
2
32o 32/ + 3) 3(o
2
32o + 3)
/
=
/(2o +/ 32)
/
= 2o +/ 32.
5. 1(r) = 2(3r 31). Domain: 3r 31 6= 0 i 3r 6= 1 i r 6=
1
3
. 1 =

3".
1
3
1
3
. "
Range: all reals except 0 (j = 0 is the horizontal asymptote for 1.) 1 = (3". 0) (0. ")
7. /(r) = ln(r + 6). Domain: r + 6 0 i r 36. 1 = (36. ")
Range: r + 6 0, so ln(r + 6) takes on all real numbers and, hence, the range is R.
1 = (3". ")
9. (a) To obtain the graph of j = 1(r) + 8, we shift the graph of j = 1(r) up 8 units.
(b) To obtain the graph of j = 1(r + 8), we shift the graph of j = 1(r) left 8 units.
(c) To obtain the graph of j = 1 + 21(r), we stretch the graph of j = 1(r) vertically by a factor of 2, and then shift the
resulting graph 1 unit upward.
(d) To obtain the graph of j = 1(r 32) 32, we shift the graph of j = 1(r) right 2 units (for the 32 inside the
parentheses), and then shift the resulting graph 2 units downward.
(e) To obtain the graph of j = 31(r), we reect the graph of j = 1(r) about the r-axis.
(f ) To obtain the graph of j = 1
31
(r), we reect the graph of j = 1(r) about the line j = r (assuming 1 is oneto-one).
11. j = 3sin2r: Start with the graph of j = sinr, compress horizontally by a factor of 2, and reect about the r-axis.
13. j =
1
2
(1 +c
i
):
Start with the graph of j = c
i
,
shift 1 unit upward, and compress
vertically by a factor of 2.
15. 1(r) =
1
r + 2
:
Start with the graph of 1(r) = 1r
and shift 2 units to the left.
CHAPTER 1 REVIEW 37
17. (a) The terms of 1 are a mixture of odd and even powers of r, so 1 is neither even nor odd.
(b) The terms of 1 are all odd powers of r, so 1 is odd.
(c) 1(3r) = c
3(3i)
2
= c
3i
2
= 1(r), so 1 is even.
(d) 1(3r) = 1 + sin(3r) = 1 3sinr. Now 1(3r) 6= 1(r) and 1(3r) 6= 31(r), so 1 is neither even nor odd.
19. 1(r) = lnr, 1 = (0. "); o(r) = r
2
39, 1 = R.
(a) (1 o)(r) = 1(o(r)) = 1(r
2
39) = ln(r
2
39).
Domain: r
2
39 0 i r
2
9 i |r| 3 i r M (3". 33) (3. ")
(b) (o 1)(r) = o(1(r)) = o(lnr) = (lnr)
2
39. Domain: r 0, or (0. ")
(c) (1 1)(r) = 1(1(r)) = 1(lnr) = ln(lnr). Domain: lnr 0 i r c
0
= 1, or (1. ")
(d) (o o)(r) = o(o(r)) = o(r
2
39) = (r
2
39)
2
39. Domain: r M R, or (3". ")
21. Many models appear to be plausible. Your choice depends on whether you
think medical advances will keep increasing life expectancy, or if there is
bound to be a natural leveling-off of life expectancy. A linear model,
j = 0.2493r 3423.4818, gives us an estimate of 77.6 years for the
year 2010.
23. We need to know the value of r such that 1(r) = 2r + lnr = 2. Since r = 1 gives us j = 2, 1
31
(2) = 1.
25. (a) c
2 ln 3
=

c
ln 3
2
= 3
2
= 9
(b) log
10
25 + log
10
4 = log
10
(25 4) = log
10
100 = log
10
10
2
= 2
(c) tan
arcsin
1
2
= tan
r
6
=
1
I
3
(d) Let 0 = cos
31 4
5
, so cos 0 =
4
5
. Then sin
cos
31 4
5
= sin0 =
I
1 3cos
2
0 =

1 3
4
5
2
=

9
25
=
3
5
.
27. (a) The population would reach 900 in about 4.4 years.
(b) 1 =
100,000
100 + 900c
3I
i 1001 + 9001c
3I
= 100,000 i 9001c
3I
= 100,000 31001 i
c
3I
=
100,000 31001
9001
i 3t = ln
1000 31
91
i t = 3ln
1000 31
91
, or ln
91
1000 31
; this is the time

required for the population to reach a given number 1.
(c) 1 = 900 i t = ln
9 900
1000 3900
= ln81 E 4.4 years, as in part (a).

PRINCIPLES OF PROBLEM SOLVING
1. By using the area formula for a triangle,
1
2
(base) (height), in two ways, we see that
1
2
(4) (j) =
1
2
(/) (o), so o =
4j
/
. Since 4
2
+j
2
= /
2
, j =
I
/
2
316, and
o =
4
I
/
2
316
/
.
3. |2r 31| =
2r 31 if r D
1
2
1 32r if r <
1
2
and |r + 5| =
r + 5 if r D 35
3r 35 if r < 35
Therefore, we consider the three cases r < 35, 35 $ r <
1
2
, and r D
1
2
.
If r < 35, we must have 1 32r 3(3r 35) = 3 C r = 3, which is false, since we are considering r < 35.
If 35 $ r <
1
2
, we must have 1 32r 3(r + 5) = 3 C r = 3
7
3
.
If r D
1
2
, we must have 2r 31 3(r + 5) = 3 C r = 9.
So the two solutions of the equation are r = 3
7
3
and r = 9.
5. 1(r) =
r
2
34 |r| + 3
. If r D 0, then 1(r) =

r
2
34r + 3
= |(r 31)(r 33)|.

Case (i): If 0 < r $ 1, then 1(r) = r
2
34r + 3.
Case (ii): If 1 < r $ 3, then 1(r) = 3(r
2
34r + 3) = 3r
2
+ 4r 33.
Case (iii): If r 3, then 1(r) = r
2
34r + 3.
This enables us to sketch the graph for r D 0. Then we use the fact that 1 is an even
function to reect this part of the graph about the j-axis to obtain the entire graph. Or, we
could consider also the cases r < 33, 33 $ r < 31, and 31 $ r < 0.
7. Remember that |o| = o if o D 0 and that |o| = 3o if o < 0. Thus,
r + |r| =
2r if r D 0
0 if r < 0
and j + |j| =
2j if j D 0
0 if j < 0
We will consider the equation r + |r| = j + |j| in four cases.
(1) r D 0. j D 0
2r = 2j
r = j
(2) r D 0, j < 0
2r = 0
r = 0
(3) r < 0, j D 0
0 = 2j
0 = j
(4) r < 0. j < 0
0 = 0
Case 1 gives us the line j = r with nonnegative r and j.
Case 2 gives us the portion of the j-axis with j negative.
Case 3 gives us the portion of the r-axis with r negative.
Case 4 gives us the entire third quadrant.
39
40 CHAPTER 1 PRINCIPLES OF PROBLEM SOLVING
9. |r| + |j| $ 1. The boundary of the region has equation |r| + |j| = 1. In quadrants
I, II, III, and IV, this becomes the lines r +j = 1, 3r +j = 1, 3r 3j = 1, and
r 3j = 1 respectively.
11. (log
2
3)(log
3
4)(log
4
5) (log
31
32) =

ln3
ln2
ln4
ln3
ln5
ln4
ln32
ln31
=
ln32
ln2
=
ln2
5
ln2
=
5 ln2
ln2
= 5
13. ln
r
2
32r 32
$ 0 i r
2
32r 32 $ c
0
= 1 i r
2
32r 33 $ 0 i (r 33)(r + 1) $ 0 i r M [31. 3].
Since the argument must be positive, r
2
32r 32 0 i

r 3
1 3
I
3
r 3
1 +
I
3
0 i
r M
3". 1 3
I
3
1 +
I
3. "
. The intersection of these intervals is
31. 1 3
I
3
1 +
I
3. 3
.
15. Let d be the distance traveled on each half of the trip. Let t
1
and t
2
be the times taken for the rst and second halves of the trip.
For the rst half of the trip we have t
1
= d30 and for the second half we have t
2
= d60. Thus, the average speed for the
entire trip is
total distance
total time
=
2d
t1 +t2
=
2d
d
30
+
d
60
60
60
=
120d
2d +d
=
120d
3d
= 40. The average speed for the entire trip
is 40 mih.
17. Let o
n
be the statement that 7
n
31 is divisible by 6.
o1 is true because 7
1
31 = 6 is divisible by 6.
Assume o
I
is true, that is, 7
I
31 is divisible by 6. In other words, 7
I
31 = 6i for some positive integer i. Then
7
I+1
31 = 7
I
7 31 = (6i+ 1) 7 31 = 42i+ 6 = 6(7i+ 1), which is divisible by 6, so o
I+1
is true.
Therefore, by mathematical induction, 7
n
31 is divisible by 6 for every positive integer n.
19. 1
0
(r) = r
2
and 1
n+1
(r) = 1
0
(1
n
(r)) for n = 0. 1. 2. . . ..
1
1
(r) = 1
0
(1
0
(r)) = 1
0
r
2
=

r
2
2
= r
4
, 1
2
(r) = 1
0
(1
1
(r)) = 1
0
(r
4
) = (r
4
)
2
= r
8
,
13(r) = 10(12(r)) = 10(r
8
) = (r
8
)
2
= r
16
, . . .. Thus, a general formula is 1n(r) = r
2
q+1
.
2 LIMITS AND DERIVATIVES
2.1 The Tangent and Velocity Problems
1. (a) Using 1(15. 250), we construct the following table:
t Q slope = i
TC
5 (5. 694)
6943250
5315
= 3
444
10
= 344.4
10 (10. 444)
4443250
10315
= 3
194
5
= 338.8
20 (20. 111)
1113250
20315
= 3
139
5
= 327.8
25 (25. 28)
283250
25315
= 3
222
10
= 322.2
30 (30. 0)
03250
30315
= 3
250
15
= 316.6
(b) Using the values of t that correspond to the points
closest to 1 (t = 10 and t = 20), we have
338.8 + (327.8)
2
= 333.3
(c) From the graph, we can estimate the slope of the
tangent line at 1 to be
3300
9
= 333.3.
3. (a)
r Q i
TC
(i) 0.5 (0.5. 0.333333) 0.333333
(ii) 0.9 (0.9. 0.473684) 0.263158
(iii) 0.99 (0.99. 0.497487) 0.251256
(iv) 0.999 (0.999. 0.499750) 0.250125
(v) 1.5 (1.5. 0.6) 0.2
(vi) 1.1 (1.1. 0.523810) 0.238095
(vii) 1.01 (1.01. 0.502488) 0.248756
(viii) 1.001 (1.001. 0.500250) 0.249875
(b) The slope appears to be
1
4
.
(c) j 3
1
2
=
1
4
(r 31) or j =
1
4
r +
1
4
.
5. (a) j = j(t) = 40t 316t
2
. At t = 2, j = 40(2) 316(2)
2
= 16. The average velocity between times 2 and 2 +/ is
ave
=
j(2 +/) 3j(2)
(2 +/) 32
=
40(2 +/) 316(2 +/)

2
316
/
=
324/ 316/
2
/
= 324 316/, if / 6= 0.
(i) [2. 2.5]: / = 0.5,
ave
= 332 fts (ii) [2. 2.1]: / = 0.1,
ave
= 325.6 fts
(iii) [2. 2.05]: / = 0.05,
ave
= 324.8 fts (iv) [2. 2.01]: / = 0.01,
ave
= 324.16 fts
(b) The instantaneous velocity when t = 2 (/ approaches 0) is 324 fts.
41
42

CHAPTER 2 LIMITS AND DERIVATIVES
7. (a) (i) On the interval [1. 3],
ave
=
c(3) 3c(1)
3 31
=
10.7 31.4
2
=
9.3
2
= 4.65 ms.
(ii) On the interval [2. 3],
ave
=
c(3) 3c(2)
3 32
=
10.7 35.1
1
= 5.6 ms.
(iii) On the interval [3. 5],
ave
=
c(5) 3c(3)
5 33
=
25.8 310.7
2
=
15.1
2
= 7.55 ms.
(iv) On the interval [3. 4],
ave
=
c(4) 3c(3)
4 33
=
17.7 310.7
1
= 7 ms.
(b) Using the points (2. 4) and (5. 23) from the approximate tangent
line, the instantaneous velocity at t = 3 is about
23 34
5 32
E 6.3 ms.
9. (a) For the curve j = sin(10r) and the point 1(1. 0):
r Q i
TC
2 (2. 0) 0
1.5 (1.5. 0.8660) 1.7321
1.4 (1.4. 30.4339) 31.0847
1.3 (1.3. 30.8230) 32.7433
1.2 (1.2. 0.8660) 4.3301
1.1 (1.1. 30.2817) 32.8173
r Q i
TC
0.5 (0.5. 0) 0
0.6 (0.6. 0.8660) 32.1651
0.7 (0.7. 0.7818) 32.6061
0.8 (0.8. 1) 35
0.9 (0.9. 30.3420) 3.4202
As r approaches 1, the slopes do not appear to be approaching any particular value.
(b) We see that problems with estimation are caused by the frequent
oscillations of the graph. The tangent is so steep at 1 that we need to
take r-values much closer to 1 in order to get accurate estimates of
its slope.
(c) If we choose r = 1.001, then the point Q is (1.001. 30.0314) and i
TC
E 331.3794. If r = 0.999, then Q is
(0.999. 0.0314) and i
TC
= 331.4422. The average of these slopes is 331.4108. So we estimate that the slope of the
tangent line at 1 is about 331.4.
SECTION 2.2 THE LIMIT OF A FUNCTION

43
2.2 The Limit of a Function
1. As r approaches 2, 1(r) approaches 5. [Or, the values of 1(r) can be made as close to 5 as we like by taking r sufciently
close to 2 (but r 6= 2).] Yes, the graph could have a hole at (2. 5) and be dened such that 1(2) = 3.
3. (a) lim
i<33
1(r) = "means that the values of 1(r) can be made arbitrarily large (as large as we please) by taking r
sufciently close to 33 (but not equal to 33).
(b) lim
i<4
+
1(r) = 3"means that the values of 1(r) can be made arbitrarily large negative by taking r sufciently close to 4
through values larger than 4.
5. (a) 1(r) approaches 2 as r approaches 1 from the left, so lim
i<1
1(r) = 2.
(b) 1(r) approaches 3 as r approaches 1 from the right, so lim
i<1
+
1(r) = 3.
(c) lim
i<1
1(r) does not exist because the limits in part (a) and part (b) are not equal.
(d) 1(r) approaches 4 as r approaches 5 from the left and from the right, so lim
i<5
1(r) = 4.
(e) 1(5) is not dened, so it doesnt exist.
7. (a) lim
I<0
o(t) = 31 (b) lim

I<0
+
o(t) = 32
(c) lim
I<0
o(t) does not exist because the limits in part (a) and part (b) are not equal.
(d) lim
I<2
o(t) = 2 (e) lim

I<2
+
o(t) = 0
(f ) lim
I<2
o(t) does not exist because the limits in part (d) and part (e) are not equal.
(g) o(2) = 1 (h) lim
I<4
o(t) = 3
9. (a) lim
i<37
1(r) = 3" (b) lim
i<33
1(r) = " (c) lim
i<0
1(r) = "
(d) lim
i<6
1(r) = 3" (e) lim

i<6
+
1(r) = "
(f ) The equations of the vertical asymptotes are r = 37, r = 33, r = 0, and r = 6.
11. (a) lim
i<0
1(r) = 1
(b) lim
i<0
+
1(r) = 0
(c) lim
i<0
1(r) does not exist because the limits
in part (a) and part (b) are not equal.
13. lim
i<1
1(r) = 2, lim
i<1
+
1(r) = 32, 1(1) = 2 15. lim
i<3
+
1(r) = 4, lim
i<3
1(r) = 2, lim
i<32
1(r) = 2,
1(3) = 3, 1(32) = 1
44

17. For 1(r) =
r
2
32r
r
2
3r 32
:
r 1(r)
2.5 0.714286
2.1 0.677419
2.05 0.672131
2.01 0.667774
2.005 0.667221
2.001 0.666778
r 1(r)
1.9 0.655172
1.95 0.661017
1.99 0.665552
1.995 0.666110
1.999 0.666556
It appears that lim
i<2
r
2
32r
r
2
3r 32
= 0.
6 =
2
3
.
19. For 1(r) =
c
i
31 3r
r
2
:
r 1(r)
1 0.718282
0.5 0.594885
0.1 0.517092
0.05 0.508439
0.01 0.501671
r 1(r)
31 0.367879
30.5 0.426123
30.1 0.483742
30.05 0.491770
30.01 0.498337
It appears that lim
i<0
c
i
31 3r
r
2
= 0.5 =
1
2
.
21. For 1(r) =
I
r + 4 32
r
:
r 1(r)
1 0.236068
0.5 0.242641
0.1 0.248457
0.05 0.249224
0.01 0.249844
r 1(r)
31 0.267949
30.5 0.258343
30.1 0.251582
30.05 0.250786
30.01 0.250156
It appears that lim
i<0
I
r + 4 32
r
= 0.25 =
1
4
.
23. For 1(r) =
r
6
31
r
10
31
:
r 1(r)
0.5 0.985337
0.9 0.719397
0.95 0.660186
0.99 0.612018
0.999 0.601200
r 1(r)
1.5 0.183369
1.1 0.484119
1.05 0.540783
1.01 0.588022
1.001 0.598800
It appears that lim
i<1
r
6
31
r
10
31
= 0.6 =
3
5
.
25. lim
i<33
+
r + 2
r + 3
= 3"since the numerator is negative and the denominator approaches 0 from the positive side as r <33
+
.
27. lim
i<1
2 3r
(r 31)
2
= "since the numerator is positive and the denominator approaches 0 through positive values as r <1.
29. Let t = r
2
39. Then as r <3
+
, t <0
+
, and lim
i<3
+
ln(r
2
39) = lim
I<0
+
lnt = 3"by (3).
31. lim
i<2r
rcsc r = lim
i<2r
r
sinr
= 3"since the numerator is positive and the denominator approaches 0 through negative
values as r <2
3
.
33. (a) 1(r) =
1
r
3
31
.
From these calculations, it seems that
lim
i<1
1(r) = 3"and lim

i<1
+
1(r) = ".
r 1(r)
0.5 31.14
0.9 33.69
0.99 333.7
0.999 3333.7
0.9999 33333.7
0.99999 333,333.7
r 1(r)
1.5 0.42
1.1 3.02
1.01 33.0
1.001 333.0
1.0001 3333.0
1.00001 33,333.3
SECTION 2.2 THE LIMIT OF A FUNCTION

45
(b) If r is slightly smaller than 1, then r
3
31 will be a negative number close to 0, and the reciprocal of r
3
31, that is, 1(r),
will be a negative number with large absolute value. So lim
i<1
1(r) = 3".
If r is slightly larger than 1, then r
3
31 will be a small positive number, and its reciprocal, 1(r), will be a large positive
number. So lim
i<1
+
1(r) = ".
(c) It appears from the graph of 1 that
lim
i<1
1(r) = 3"and lim

i<1
+
1(r) = ".
35. (a) Let /(r) = (1 +r)
1'i
.
r /(r)
30.001 2.71964
30.0001 2.71842
30.00001 2.71830
30.000001 2.71828
0.000001 2.71828
0.00001 2.71827
0.0001 2.71815
0.001 2.71692
It appears that lim
i<0
(1 +r)
1'i
E 2.71828, which is approximately c.
In Section 3.6 we will see that the value of the limit is exactly c.
(b)
37. For 1(r) = r
2
3(2
i
1000):
(a)
r 1(r)
1 0.998000
0.8 0.638259
0.6 0.358484
0.4 0.158680
0.2 0.038851
0.1 0.008928
0.05 0.001465
It appears that lim
i<0
1(r) = 0.
(b)
r 1(r)
0.04 0.000572
0.02 30.000614
0.01 30.000907
0.005 30.000978
0.003 30.000993
0.001 30.001000
It appears that lim
i<0
1(r) = 30.001.
46

39. No matter how many times we zoom in toward the origin, the graphs of 1(r) = sin(r) appear to consist of almost-vertical
lines. This indicates more and more frequent oscillations as r <0.
41. There appear to be vertical asymptotes of the curve j = tan(2 sinr) at r E 0.90
and r E 2.24. To nd the exact equations of these asymptotes, we note that the
graph of the tangent function has vertical asymptotes at r =
r
2
+n. Thus, we
must have 2 sinr =
r
2
+n, or equivalently, sinr =
r
4
+
r
2
n. Since
31 $ sinr $ 1, we must have sinr =
r
4
and so r = sin
31 r
4
(corresponding
to r E 0.90). Just as 150
is the reference angle for 30
, 3sin
31 r
4
is the
reference angle for sin
31 r
4
. So r =
3sin
31 r
4
are also equations of

vertical asymptotes (corresponding to r E 2.24).
2.3 Calculating Limits Using the Limit Laws
1. (a) lim
i<2
[1(r) + 5o(r)] = lim
i<2
1(r) + lim
i<2
[5o(r)] [Limit Law 1]
= lim
i<2
1(r) + 5 lim
i<2
o(r) [Limit Law 3]
= 4 + 5(32) = 36
(b) lim
i<2
[o(r)]
3
=

lim
i<2
o(r)
3
[Limit Law 6]
= ( 32)
3
= 38
(c) lim
i<2
1(r) =

lim
i<2
1(r) [Limit Law 11]
=
I
4 = 2
SECTION 2.3 CALCULATING LIMITS USING THE LIMIT LAWS

47
(d) lim
i<2
31(r)
o(r)
=
lim
i<2
[31(r)]
lim
i<2
o(r)
[Limit Law 5]
=
3 lim
i<2
1(r)
lim
i<2
o(r)
[Limit Law 3]
=
3(4)
32
= 36
(e) Because the limit of the denominator is 0, we cant use Limit Law 5. The given limit, lim
i<2
o(r)
/(r)
, does not exist because the
denominator approaches 0 while the numerator approaches a nonzero number.
(f ) lim
i<2
o(r) /(r)
1(r)
=
lim
i<2
[o(r) /(r)]
lim
i<2
1(r)
[Limit Law 5]
=
lim
i<2
o(r) lim
i<2
/(r)
lim
i<2
1(r)
[Limit Law 4]
=
32 0
4
= 0
3. lim
i<32
(3r
4
+ 2r
2
3r + 1) = lim
i<32
3r
4
+ lim
i<32
2r
2
3 lim
i<32
r + lim
i<32
1 [Limit Laws 1 and 2]
= 3 lim
i<32
r
4
+ 2 lim
i<32
r
2
3 lim
i<32
r + lim
i<32
1 [3]
= 3(32)
4
+ 2(32)
2
3(32) + (1) [9, 8, and 7]
= 48 + 8 + 2 + 1 = 59
5. lim
i<8
(1 +
3
I
r) (2 36r
2
+r
3
) = lim
i<8
(1 +
3
I
r) lim
i<8
(2 36r
2
+r
3
) [Limit Law 4]
=

lim
i<8
1 + lim
i<8
3
I
r
lim
i<8
2 36 lim
i<8
r
2
+ lim
i<8
r
3
[1, 2, and 3]
=

1 +
3
I
8
2 36 8
2
+ 8
3
[7, 10, 9]
= (3)(130) = 390
7. lim
i<1
1 + 3r
1 + 4r
2
+ 3r
4
3
=

lim
i<1
1 + 3r
1 + 4r
2
+ 3r
4
3
[6]
=
lim
i<1
(1 + 3r)
lim
i<1
(1 + 4r
2
+ 3r
4
)
3
[5]
=
lim
i<1
1 + 3 lim
i<1
r
lim
i<1
1 + 4 lim
i<1
r
2
+ 3 lim
i<1
r
4
3
[2, 1, and 3]
=

1 + 3(1)
1 + 4(1)
2
+ 3(1)
4
3
=

4
8
3
=

1
2
3
=
1
8
[7, 8, and 9]
9. lim
i<4
I
16 3r
2
=

lim
i<4
(16 3r
2
) [11]
=

lim
i<4
16 3 lim
i<4
r
2
[2]
=

16 3(4)
2
= 0 [7 and 9]
48

11. lim
i<2
r
2
+r 36
r 32
= lim
i<2
(r + 3)(r 32)
r 32
= lim
i<2
(r + 3) = 2 + 3 = 5
13. lim
i<2
r
2
3r + 6
r 32
does not exist since r 32 <0 but r
2
3r + 6 <8 as r <2.
15. lim
I<33
t
2
39
2t
2
+ 7t + 3
= lim
I<33
(t + 3)(t 33)
(2t + 1)(t + 3)
= lim
I<33
t 33
2t + 1
=
33 33
2(33) + 1
=
36
35
=
6
5
17. lim
I<0
(4 +/)
2
316
/
= lim
I<0
(16 + 8/ +/
2
) 316
/
= lim
I<0
8/ +/
2
/
= lim
I<0
/(8 +/)
/
= lim
I<0
(8 +/) = 8 + 0 = 8
19. By the formula for the sum of cubes, we have
lim
i<32
r + 2
r
3
+ 8
= lim
i<32
r + 2
(r + 2)(r
2
32r + 4)
= lim
i<32
1
r
2
32r + 4
=
1
4 + 4 + 4
=
1
12
.
21. lim
I<9
9 3t
3 3
I
t
= lim
I<9
3 +
I
t

3 3
I
t
3 3
I
t
= lim
I<9
3 +
I
t
= 3 +
I
9 = 6
23. lim
i<7
I
r + 2 33
r 37
= lim
i<7
I
r + 2 33
r 37

I
r + 2 + 3
I
r + 2 + 3
= lim
i<7
(r + 2) 39
(r 37)
I
r + 2 + 3
= lim
i<7
r 37
(r 37)
I
r + 2 + 3
= lim
i<7
1
I
r + 2 + 3
=
1
I
9 + 3
=
1
6
25. lim
i<34
1
4
+
1
r
4 +r
= lim
i<34
r + 4
4r
4 +r
= lim
i<34
r + 4
4r(4 +r)
= lim
i<34
1
4r
=
1
4(34)
= 3
1
16
27. lim
i<16
4 3
I
r
16r 3r
2
= lim
i<16
(4 3
I
r)(4 +
I
r)
(16r 3r
2
)(4 +
I
r)
= lim
i<16
16 3r
r(16 3r)(4 +
I
r)
= lim
i<16
1
r(4 +
I
r)
=
1
16
4 +
I
16
=
1
16(8)
=
1
128
29. lim
I<0
1
t
I
1 +t
3
1
t
= lim
I<0
1 3
I
1 +t
t
I
1 +t
= lim
I<0
1 3
I
1 +t
1 +
I
1 +t
t
I
t + 1
1 +
I
1 +t
= lim
I<0
3t
t
I
1 +t
1 +
I
1 +t
= lim
I<0
31
I
1 +t
1 +
I
1 +t
=
31
I
1 + 0
1 +
I
1 + 0
= 3
1
2
31. (a)
lim
i<0
r
I
1 + 3r 31
E
2
3
(b)
r 1(r)
30.001 0.6661663
30.0001 0.6666167
30.00001 0.6666617
30.000001 0.6666662
0.000001 0.6666672
0.00001 0.6666717
0.0001 0.6667167
0.001 0.6671663
The limit appears to be
2
3
.
SECTION 2.3 CALCULATING LIMITS USING THE LIMIT LAWS

49
(c) lim
i<0
r
I
1 + 3r 31

I
1 + 3r + 1
I
1 + 3r + 1
= lim
i<0
r
I
1 + 3r + 1
(1 + 3r) 31
= lim
i<0
r
I
1 + 3r + 1
3r
=
1
3
lim
i<0
I
1 + 3r + 1
[Limit Law 3]
=
1
3
lim
i<0
(1 + 3r) + lim
i<0
1
[1 and 11]
=
1
3
lim
i<0
1 + 3 lim
i<0
r + 1
[1, 3, and 7]
=
1
3
I
1 + 3 0 + 1
[7 and 8]
=
1
3
(1 + 1) =
2
3
33. Let 1(r) = 3r
2
, o(r) = r
2
cos 20r and /(r) = r
2
. Then
31 $ cos 20r $ 1 i 3r
2
$ r
2
cos 20r $ r
2
i 1(r) $ o(r) $ /(r).
So since lim
i<0
1(r) = lim
i<0
/(r) = 0, by the Squeeze Theorem we have
lim
i<0
o(r) = 0.
35. We have lim
i<4
(4r 39) = 4(4) 39 = 7 and lim
i<4
r
2
34r + 7
= 4
2
34(4) + 7 = 7. Since 4r 39 $ 1(r) $ r
2
34r + 7
for r D 0, lim
i<4
1(r) = 7 by the Squeeze Theorem.
37. 31 $ cos(2r) $ 1 i 3r
4
$ r
4
cos(2r) $ r
4
. Since lim
i<0
3r
4
= 0 and lim
i<0
r
4
= 0, we have
lim
i<0
r
4
cos(2r)
= 0 by the Squeeze Theorem.

39. |r 33| =
r 33 if r 33 D 0
3(r 33) if r 33 < 0
=
r 33 if r D 3
3 3r if r < 3
Thus, lim
i<3
+
(2r + |r 33|) = lim
i<3
+
(2r +r 33) = lim
i<3
+
(3r 3 3) = 3(3) 3 3 = 6 and
lim
i<3
(2r + |r 33|) = lim

i<3
(2r + 3 3r) = lim

i<3
(r + 3) = 3 + 3 = 6. Since the left and right limits are equal,

lim
i<3
(2r + |r 33|) = 6.
41.

2r
3
3r
2
r
2
(2r 31)
r
2
|2r 31| = r
2
|2r 31|
|2r 31| =
2r 31 if 2r 31 D 0
3(2r 31) if 2r 31 < 0
=
2r 31 if r D 0.5
3(2r 31) if r < 0.5
So
2r
3
3r
2
= r
2
[3(2r 31)] for r < 0.5.
Thus, lim
i<0.5
2r 31
|2r
3
3r
2
|
= lim
i<0.5
2r 31
r
2
[3(2r 31)]
= lim
i<0.5
31
r
2
=
31
(0.5)
2
=
31
0.25
= 34.
50

43. Since |r| = 3r for r < 0, we have lim
i<0
1
r
3
1
|r|
= lim
i<0
1
r
3
1
3r
= lim
i<0
2
r
, which does not exist since the
denominator approaches 0 and the numerator does not.
45. (a) (b) (i) Since sgnr = 1 for r 0, lim
i<0
+
sgnr = lim
i<0
+
1 = 1.
(ii) Since sgnr = 31 for r < 0, lim
i<0
sgn r = lim
i<0
31 = 31.
(iii) Since lim
i<0
sgnr 6= lim
i<0
+
sgnr, lim
i<0
sgnr does not exist.
(iv) Since |sgnr| = 1 for r 6= 0, lim
i<0
|sgnr| = lim
i<0
1 = 1.
47. (a) (i) lim
i<1
+
1(r) = lim
i<1
+
r
2
31
|r 31|
= lim
i<1
+
r
2
31
r 31
= lim
i<1
+
(r + 1) = 2 (c)
(ii) lim
i<1
1(r) = lim
i<1
r
2
31
|r 31|
= lim
i<1
r
2
31
3(r 31)
= lim
i<1
3(r + 1) = 32
(b) No, lim
i<1
1(r) does not exist since lim
i<1
+
1(r) 6= lim
i<1
1(r).
49. (a) (i) [[r]] = 32 for 32 $ r < 31, so lim
i<32
+
[[r]] = lim
i<32
+
(32) = 32
(ii) [[r]] = 33 for 33 $ r < 32, so lim
i<32
[[r]] = lim
i<32
(33) = 33.
The right and left limits are different, so lim
i<32
[[r]] does not exist.
(iii) [[r]] = 33 for 33 $ r < 32, so lim
i<32.4
[[r]] = lim
i<32.4
(33) = 33.
(b) (i) [[r]] = n 31 for n 31 $ r < n, so lim
i<n
[[r]] = lim
i<n
(n 31) = n 31.
(ii) [[r]] = n for n $ r < n + 1, so lim
i<n
+
[[r]] = lim
i<n
+
n = n.
(c) lim
i<a
[[r]] exists C o is not an integer.
51. The graph of 1(r) = [[r]] + [[3r]] is the same as the graph of o(r) = 31 with holes at each integer, since 1(o) = 0 for any
integer o. Thus, lim
i<2
1(r) = 31 and lim

i<2
+
1 (r) = 31, so lim
i<2
1(r) = 31. However,
1(2) = [[2]] + [[32]] = 2 + (32) = 0, so lim
i<2
1(r) 6= 1(2).
53. Since j(r) is a polynomial, j(r) = o
0
+o
1
r +o
2
r
2
+ +o
n
r
n
. Thus, by the Limit Laws,
lim
i<a
j(r) = lim
i<a
o
0
+o
1
r +o
2
r
2
+ +o
n
r
n
= o
0
+o
1
lim
i<a
r +o
2
lim
i<a
r
2
+ +o
n
lim
i<a
r
n
= o
0
+o
1
o +o
2
o
2
+ +o
n
o
n
= j(o)
Thus, for any polynomial j, lim
i<a
j(r) = j(o).
SECTION 2.4 THE PRECISE DEFINITION OF A LIMIT

51
55. lim
i<1
[1(r) 38] = lim
i<1
1(r) 38
r 31
(r 31)
= lim
i<1
1(r) 38
r 31
lim
i<1
(r 31) = 10 0 = 0.
Thus, lim
i<1
1(r) = lim
i<1
{[1(r) 38] + 8} = lim
i<1
[1(r) 38] + lim
i<1
8 = 0 + 8 = 8.
Note: The value of lim
i<1
1(r) 38
r 31
does not affect the answer since its multiplied by 0. Whats important is that lim
i<1
1(r) 38
r 31
exists.
57. Observe that 0 $ 1(r) $ r
2
for all r, and lim
i<0
0 = 0 = lim
i<0
r
2
. So, by the Squeeze Theorem, lim
i<0
1(r) = 0.
59. Let 1(r) = 1(r) and o(r) = 1 31(r), where 1 is the Heaviside function dened in Exercise 1.3.57.
Thus, either 1 or o is 0 for any value of r. Then lim
i<0
1(r) and lim
i<0
o(r) do not exist, but lim
i<0
[1(r)o(r)] = lim
i<0
0 = 0.
61. Since the denominator approaches 0 as r <32, the limit will exist only if the numerator also approaches
0 as r < 32. In order for this to happen, we need lim
i<32
3r
2
+or +o + 3
= 0 C
3(32)
2
+o(32) +o + 3 = 0 C 12 32o +o + 3 = 0 C o = 15. With o = 15, the limit becomes
lim
i<32
3r
2
+ 15r + 18
r
2
+r 32
= lim
i<32
3(r + 2)(r + 3)
(r 31)(r + 2)
= lim
i<32
3(r + 3)
r 31
=
3(32 + 3)
32 31
=
3
33
= 31.
2.4 The Precise Definition of a Limit
1. On the left side of r = 2, we need |r 32| <

10
7
32
=
4
7
. On the right side, we need |r 32| <

10
3
32
=
4
3
. For both of
these conditions to be satised at once, we need the more restrictive of the two to hold, that is, |r 32| <
4
7
. So we can choose
c =
4
7
, or any smaller positive number.
3. The leftmost question mark is the solution of
I
r = 1.6 and the rightmost,
I
r = 2.4. So the values are 1.6
2
= 2.56 and
2.4
2
= 5.76. On the left side, we need |r 34| < |2.56 34| = 1.44. On the right side, we need |r 34| < |5.76 34| = 1.76.
To satisfy both conditions, we need the more restrictive condition to holdnamely, |r 34| < 1.44. Thus, we can choose
c = 1.44, or any smaller positive number.
5. From the graph, we nd that tanr = 0.8 when r E 0.675, so
r
4
3c
1
E 0.675 i c
1
E
r
4
30.675 E 0.1106. Also, tanr = 1.2
when r E 0.876, so
r
4
+c
2
E 0.876 i c
2
= 0.876 3
r
4
E 0.0906.
Thus, we choose c = 0.0906 (or any smaller positive number) since this is
the smaller of c
1
and c
2
.
52

7. For - = 1, the denition of a limit requires that we nd c such that
4 +r 33r
3
32
< 1 C 1 < 4 +r 33r

3
< 3
whenever 0 < |r 31| < c. If we plot the graphs of j = 1, j = 4 +r 33r
3
and j = 3 on the same screen, we see that we
need 0.86 $ r $ 1.11. So since |1 30.86| = 0.14 and |1 31.11| = 0.11, we choose c = 0.11 (or any smaller positive
number). For - = 0.1, we must nd c such that
4 +r 33r
3
32
< 0.1 C 1.9 < 4 +r 33r

3
< 2.1 whenever
0 < |r 31| < c. From the graph, we see that we need 0.988 $ r $ 1.012. So since |1 30.988| = 0.012 and
|1 31.012| = 0.012, we choose c = 0.012 (or any smaller positive number) for the inequality to hold.
9. (a) From the graph, we nd that j = tan
2
r = 1000 when r E 1.539 and
r E 1.602 for r near
r
2
. Thus, we get c E 1.602 3
r
2
E 0.031 for
A = 1000.
(b) From the graph, we nd that j = tan
2
r = 10,000 when r E 1.561 and
r E 1.581 for r near
r
2
. Thus, we get c E 1.581 3
r
2
E 0.010 for
A = 10,000.
11. (a) = v
2
and = 1000 cm
2
i v
2
= 1000 i v
2
=
1000
r
i v =

1000
r
(v 0) E 17.8412 cm.
(b) |31000| $ 5 i 35 $ v
2
31000 $ 5 i 1000 35 $ v
2
$ 1000 + 5 i
995
r
$ v $
1005
r
i 17.7966 $ v $ 17.8858.

1000
r
3
995
r
E 0.04466 and
1005
r
3
1000
r
E 0.04455. So
if the machinist gets the radius within 0.0445 cm of 17.8412, the area will be within 5 cm
2
of 1000.
(c) r is the radius, 1(r) is the area, o is the target radius given in part (a), 1 is the target area (1000), - is the tolerance in the
area (5), and c is the tolerance in the radius given in part (b).
13. (a) |4r 38| = 4 |r 32| < 0.1 C |r 32| <
0.1
4
, so c =
0.1
4
= 0.025.
(b) |4r 38| = 4 |r 32| < 0.01 C |r 32| <
0.01
4
, so c =
0.01
4
= 0.0025.
SECTION 2.4 THE PRECISE DEFINITION OF A LIMIT

53
15. Given - 0, we need c 0 such that if 0 < |r 31| < c, then
|(2r + 3) 35| < -. But |(2r + 3) 35| < - C
|2r 32| < - C 2 |r 31| < - C |r 31| < -2.
So if we choose c = -2, then 0 < |r 31| < c i
|(2r + 3) 35| < -. Thus, lim
i<1
(2r + 3) = 5 by the denition of a limit.
17. Given - 0, we need c 0 such that if 0 < |r 3(33)| < c, then
|(1 34r) 313| < -. But |(1 34r) 313| < - C
|34r 312| < - C |34| |r + 3| < - C |r 3(33)| < -4.
So if we choose c = -4, then 0 < |r 3(33)| < c i
|(1 34r) 313| < -. Thus, lim
i<33
(1 34r) = 13 by the denition of
a limit.
x
r
5
3
3
5
< - C
1
5
|r 33| < - C |r 33| < 5-.
So choose c = 5-. Then 0 < |r 33| < c i |r 33| < 5- i
|r 33|
5
< - i
r
5
3
3
5
< -. By the denition

of a limit, lim
i<3
r
5
=
3
5
.
r
2
+r 36
r 32
35
< - C
(r + 3)(r 32)
r 32
35
< - C |r + 3 35| < - [r 6= 2] C |r 32| < -. So choose c = -.

Then 0 < |r 32| < c i |r 32| < - i |r + 3 35| < - i
(r + 3)(r 32)
r 32
35
< - [r 6= 2] i
r
2
+r 36
r 32
35
< -. By the denition of a limit, lim

i<2
r
2
+r 36
r 32
= 5.
23. Given - 0, we need c 0 such that if 0 < |r 3o| < c, then |r 3o| < -. So c = - will work.
r
2
30
< - C r
2
< - C |r| <
I
-. Take c =
I
-.
Then 0 < |r 30| < c i

r
2
30
< -. Thus, lim

i<0
r
2
= 0 by the denition of a limit.
|r| 30
< -. But
|r|
= |r|. So this is true if we pick c = -.

Thus, lim
i<0
|r| = 0 by the denition of a limit.
54

r
2
34r + 5
31
< - C

r
2
34r + 4
< - C
(r 32)
2
< -. So take c =
I
-. Then 0 < |r 32| < c C |r 32| <
I
- C

(r 32)
2
< -. Thus,
lim
i<2
r
2
34r + 5

31. Given - 0, we need c 0 such that if 0 < |r 3(32)| < c, then
r
2
31
33
< - or upon simplifying we need
r
2
34
< - whenever 0 < |r + 2| < c. Notice that if |r + 2| < 1, then 31 < r + 2 < 1 i 35 < r 32 < 33 i
|r 32| < 5. So take c = min{-5. 1}. Then 0 < |r + 2| < c i |r 32| < 5 and |r + 2| < -5, so
r
2
31
33
= |(r + 2)(r 32)| = |r + 2| |r 32| < (-5)(5) = -. Thus, by the denition of a limit, lim
i<32
(r
2
31) = 3.
33. Given - 0, we let c = min
2.
s
8
. If 0 < |r 33| < c, then |r 33| < 2 i 32 < r 33 < 2 i

4 < r + 3 < 8 i |r + 3| < 8. Also |r 33| <
s
8
, so
r
2
39
= |r + 3| |r 33| < 8
s
8
= -. Thus, lim
i<3
r
2
= 9.
35. (a) The points of intersection in the graph are (r
1
. 2.6) and (r
2
. 3.4)
with r
1
E 0.891 and r
2
E 1.093. Thus, we can take c to be the
smaller of 1 3r
1
and r
2
31. So c = r
2
31 E 0.093.
(b) Solving r
3
+r + 1 = 3 +- gives us two nonreal complex roots and one real root, which is
r(-) =
216 + 108- + 12
I
336 + 324- + 81-
2
2'3
312
6
216 + 108- + 12
I
336 + 324- + 81-
2
1'3
. Thus, c = r(-) 31.
(c) If - = 0.4, then r(-) E 1.093 272 342 and c = r(-) 31 E 0.093, which agrees with our answer in part (a).
37. 1. Guessing a value for c Given - 0, we must nd c 0 such that |
I
r 3
I
o| < - whenever 0 < |r 3o| < c. But
|
I
r 3
I
o| =
|r 3o|
I
r +
I
o
< - (from the hint). Now if we can nd a positive constant C such that
I
r +
I
o C then
|r 3o|
I
r +
I
o
<
|r 3o|
C
< -, and we take |r 3o| < C-. We can nd this number by restricting r to lie in some interval
centered at o. If |r 3o| <
1
2
o, then 3
1
2
o < r 3o <
1
2
o i
1
2
o < r <
3
2
o i
I
r +
I
o

1
2
o +
I
o, and so
C =

1
2
o +
I
o is a suitable choice for the constant. So |r 3o| <

1
2
o +
I
o
-. This suggests that we let

c = min
1
2
o.
1
2
o +
I
o
.
2. Showing that c works Given - 0, we let c = min
1
2
o.
1
2
o +
I
o
. If 0 < |r 3o| < c, then

|r 3o| <
1
2
o i
I
r +
I
o

1
2
o +
I
o (as in part 1). Also |r 3o| <

1
2
o +
I
o
-, so
|
I
r 3
I
o | =
|r 3o|
I
r +
I
o
<
o2 +
I
o
o2 +
I
o
= -. Therefore, lim
i<a
I
r =
I
o by the denition of a limit.
SECTION 2.5 CONTINUITY

55
39. Suppose that lim
i<0
1(r) = 1. Given - =
1
2
, there exists c 0 such that 0 < |r| < c i |1(r) 31| <
1
2
. Take any rational
number v with 0 < |v| < c. Then 1(v) = 0, so |0 31| <
1
2
, so 1 $ |1| <
1
2
. Now take any irrational number c with
0 < |c| < c. Then 1(c) = 1, so |1 31| <
1
2
. Hence, 1 31 <
1
2
, so 1
1
2
. This contradicts 1 <
1
2
, so lim
i<0
1(r) does not
exist.
41.
1
(r + 3)
4
10,000 C (r + 3)
4
<
1
10,000
C |r + 3| <
1
4
I
10,000
C |r 3(33)| <
1
10
43. Given A < 0 we need c 0 so that lnr < A whenever 0 < r < c; that is, r = c
ln i
< c
L
whenever 0 < r < c. This
suggests that we take c = c
L
. If 0 < r < c
L
, then lnr < lnc
L
= A. By the denition of a limit, lim
i<0
+
lnr = 3".
2.5 Continuity
1. From Denition 1, lim
i<4
1(r) = 1(4).
3. (a) The following are the numbers at which 1 is discontinuous and the type of discontinuity at that number: 34 (removable),
32 ( jump), 2 ( jump), 4 (innite).
(b) 1 is continuous from the left at 32 since lim
i<32
1(r) = 1(32). 1 is continuous from the right at 2 and 4 since

lim
i<2
+
1(r) = 1(2) and lim
i<4
+
1(r) = 1(4). It is continuous from neither side at 34 since 1(34) is undened.
5. The graph of j = 1(r) must have a discontinuity at r = 3 and must show that lim
i<3
1(r) = 1(3).
7. (a) (b) There are discontinuities at times t = 1, 2, 3, and 4. A person
parking in the lot would want to keep in mind that the charge will
jump at the beginning of each hour.
9. Since 1 and o are continuous functions,
lim
i<3
[21(r) 3o(r)] = 2 lim
i<3
1(r) 3 lim
i<3
o(r) [by Limit Laws 2 and 3]
= 21(3) 3o(3) [by continuity of 1 and o at r = 3]
= 2 5 3o(3) = 10 3o(3)
Since it is given that lim
i<3
[21(r) 3o(r)] = 4, we have 10 3o(3) = 4, so o(3) = 6.
56

11. lim
i<31
1(r) = lim
i<31
r + 2r
3
4
=

lim
i<31
r + 2 lim
i<31
r
3
4
=

31 + 2(31)
3
4
= (33)
4
= 81 = 1(31).
By the denition of continuity, 1 is continuous at o = 31.
13. For o 2, we have
lim
i<a
1(r) = lim
i<a
2r + 3
r 32
=
lim
i<a
(2r + 3)
lim
i<a
(r 32)
[Limit Law 5]
=
2 lim
i<a
r + lim
i<a
3
lim
i<a
r 3 lim
i<a
2
[1, 2, and 3]
=
2o + 3
o 32
[7 and 8]
= 1(o)
Thus, 1 is continuous at r = o for every o in (2. "); that is, 1 is continuous on (2. ").
15. 1(r) = ln|r 32| is discontinuous at 2 since 1(2) = ln0 is not dened.
17. 1(r) =
c
i
if r < 0
r
2
if r D 0
The left-hand limit of 1 at o = 0 is lim
i<0
1(r) = lim
i<0
c
i
= 1. The
right-hand limit of 1 at o = 0 is lim
i<0
+
1(r) = lim
i<0
+
r
2
= 0. Since these
limits are not equal, lim
i<0
1(r) does not exist and 1 is discontinuous at 0.
19. 1(r) =
cos r if r < 0
0 if r = 0
1 3r
2
if r 0
lim
i<0
1(r) = 1, but 1(0) = 0 6= 1, so 1 is discontinuous at 0.
21. 1(r) =
r
r
2
+ 5r + 6
is a rational function. So by Theorem 5 (or Theorem 7), 1 is continuous at every number in its domain,
r | r
2
+ 5r + 6 6= 0
= {r | (r + 3)(r + 2) 6= 0} = {r | r 6= 33, 32} or (3". 33) (33. 32) (32. ").

23. By Theorem 5, the polynomials r
2
and 2r 31 are continuous on (3". "). By Theorem 7, the root function
I
r is
continuous on [0. "). By Theorem 9, the composite function
I
2r 31 is continuous on its domain, [
1
2
. ").
By part 1 of Theorem 4, the sum1(r) = r
2
+
I
2r 31 is continuous on [
1
2
. ").

57
25. By Theorem 7, the exponential function c
35I
and the trigonometric function cos 2t are continuous on (3". ").
By part 4 of Theorem 4, 1(t) = c
35I
cos 2t is continuous on (3". ").
27. By Theorem 5, the polynomial t
4
31 is continuous on (3". "). By Theorem 7, lnr is continuous on its domain, (0. ").
By Theorem 9, ln
t
4
31
is continuous on its domain, which is
t | t
4
31 0
=

t | t
4
1
= {t | |t| 1} = (3". 31) (1. ")

29. The function j =
1
1 +c
1'i
is discontinuous at r = 0 because the
left- and right-hand limits at r = 0 are different.
31. Because we are dealing with root functions, 5 +
I
r is continuous on [0. "),
I
r + 5 is continuous on [35. "), so the
quotient 1(r) =
5 +
I
r
I
5 +r
is continuous on [0. "). Since 1 is continuous at r = 4, lim
i<4
1(r) = 1(4) =
7
3
.
33. Because r
2
3r is continuous on R, the composite function 1(r) = c
i
2
3i
is continuous on R, so
lim
i<1
1(r) = 1(1) = c
1 31
= c
0
= 1.
35. 1(r) =
r
2
if r < 1
I
r if r D 1
By Theorem 5, since 1(r) equals the polynomial r
2
on (3". 1), 1 is continuous on (3". 1). By Theorem 7, since 1(r)
equals the root function
I
r on (1. "). 1 is continuous on (1. "). At r = 1, lim
i<1
1(r) = lim
i<1
r
2
= 1 and
lim
i<1
+
1(r) = lim
i<1
+
I
r = 1. Thus, lim
i<1
1(r) exists and equals 1. Also, 1(1) =
I
1 = 1. Thus, 1 is continuous at r = 1.
We conclude that 1 is continuous on (3". ").
37. 1(r) =
1 +r
2
if r $ 0
2 3r if 0 < r $ 2
(r 32)
2
if r 2
1 is continuous on (3". 0), (0. 2), and (2. ") since it is a polynomial on
each of these intervals. Now lim
i<0
1(r) = lim
i<0
(1 +r
2
) = 1 and lim
i<0
+
1(r) = lim
i<0
+
(2 3r) = 2. so 1 is
discontinuous at 0. Since 1(0) = 1, 1 is continuous from the left at 0. Also, lim
i<2
1(r) = lim
i<2
(2 3r) = 0.
lim
i<2
+
1(r) = lim
i<2
+
(r 32)
2
= 0, and 1(2) = 0, so 1 is continuous at 2. The only number at which 1 is discontinuous is 0.
58

39. 1(r) =
r + 2 if r < 0
c
i
if 0 $ r $ 1
2 3r if r 1
1 is continuous on (3". 0) and (1. ") since on each of these intervals
it is a polynomial; it is continuous on (0. 1) since it is an exponential.
Now lim
i<0
1(r) = lim
i<0
(r + 2) = 2 and lim
i<0
+
1(r) = lim
i<0
+
c
i
= 1, so 1 is discontinuous at 0. Since 1(0) = 1, 1 is
continuous from the right at 0. Also lim
i<1
1(r) = lim
i<1
c
i
= c and lim
i<1
+
1(r) = lim
i<1
+
(2 3r) = 1, so 1 is discontinuous
at 1. Since 1(1) = c, 1 is continuous from the left at 1.
41. 1(r) =
cr
2
+ 2r if r < 2
r
3
3cr if r D 2
1 is continuous on (3". 2) and (2. "). Now lim
i<2
1(r) = lim
i<2
cr
2
+ 2r
= 4c + 4 and
lim
i<2
+
1(r) = lim
i<2
+
r
3
3cr
= 8 32c. So 1 is continuous C 4c +4 = 8 32c C 6c = 4 C c =

2
3
. Thus, for 1
to be continuous on (3". "), c =
2
3
.
43. (a) 1(r) =
r
4
31
r 31
=
(r
2
+ 1)(r
2
31)
r 31
=
(r
2
+ 1)(r + 1)(r 31)
r 31
= (r
2
+ 1)(r + 1) [or r
3
+r
2
+r + 1]
for r 6= 1. The discontinuity is removable and o(r) = r
3
+r
2
+r + 1 agrees with 1 for r 6= 1 and is continuous on R.
(b) 1(r) =
r
3
3r
2
32r
r 32
=
r(r
2
3r 32)
r 32
=
r(r 32)(r + 1)
r 32
= r(r + 1) [or r
2
+r] for r 6= 2. The discontinuity
is removable and o(r) = r
2
+r agrees with 1 for r 6= 2 and is continuous on R.
(c) lim
i<r
1(r) = lim
i<r
[[sinr]] = lim
i<r
0 = 0 and lim
i<r
+
1(r) = lim
i<r
+
[[sinr]] = lim
i<r
+
(31) = 31, so lim
i<r
1(r) does not
exist. The discontinuity at r = is a jump discontinuity.
45. 1(r) = r
2
+ 10 sinr is continuous on the interval [31. 32], 1(31) E 957, and 1(32) E 1030. Since 957 < 1000 < 1030,
there is a number c in (31. 32) such that 1(c) = 1000 by the Intermediate Value Theorem. Note: There is also a number c in
(332. 331) such that 1(c) = 1000.
47. 1(r) = r
4
+r 33 is continuous on the interval [1. 2]. 1(1) = 31, and 1(2) = 15. Since 31 < 0 < 15, there is a number c
in (1. 2) such that 1(c) = 0 by the Intermediate Value Theorem. Thus, there is a root of the equation r
4
+r 33 = 0 in the
interval (1. 2).
49. 1(r) = cos r 3r is continuous on the interval [0. 1], 1(0) = 1, and 1(1) = cos 1 31 E 30.46. Since 30.46 < 0 < 1, there
is a number c in (0. 1) such that 1(c) = 0 by the Intermediate Value Theorem. Thus, there is a root of the equation
cos r 3r = 0, or cos r = r, in the interval (0. 1).

59
51. (a) 1(r) = cos r 3r
3
is continuous on the interval [0. 1], 1(0) = 1 0, and 1(1) = cos 1 31 E 30.46 < 0. Since
1 0 30.46, there is a number c in (0. 1) such that 1(c) = 0 by the Intermediate Value Theorem. Thus, there is a root
of the equation cos r 3r
3
= 0, or cos r = r
3
, in the interval (0. 1).
(b) 1(0.86) E 0.016 0 and 1(0.87) E 30.014 < 0, so there is a root between 0.86 and 0.87, that is, in the interval
(0.86. 0.87).
53. (a) Let 1(r) = 100c
3i'100
30.01r
2
. Then 1(0) = 100 0 and
1(100) = 100c
31
3100 E 363.2 < 0. So by the Intermediate
Value Theorem, there is a number c in (0. 100) such that 1(c) = 0.
This implies that 100c
3c'100
= 0.01c
2
.
(b) Using the intersect feature of the graphing device, we nd that the
root of the equation is r = 70.347, correct to three decimal places.
55. (i) If 1 is continuous at o, then by Theorem 8 with o(/) = o +/, we have
lim
I<0
1(o +/) = 1
lim
I<0
(o +/)
= 1(o).
(U) Let - 0. Since lim
I<0
1(o +/) = 1(o), there exists c 0 such that 0 < |/| < c i
|1(o +/) 31(o)| < -. So if 0 < |r 3o| < c, then |1(r) 31(o)| = |1(o + (r 3o)) 31(o)| < -.
Thus, lim
i<a
1(r) = 1(o) and so 1 is continuous at o.
57. As in the previous exercise, we must show that lim
I<0
cos(o +/) = cos o to prove that the cosine function is continuous.
lim
I<0
cos(o +/) = lim
I<0
(cos o cos / 3sino sin/) = lim
I<0
(cos o cos /) 3 lim
I<0
(sino sin/)
=

lim
I<0
cos o
lim
I<0
cos /
lim
I<0
sino
lim
I<0
sin/
= (cos o)(1) 3(sino)(0) = cos o

59. 1(r) =
0 if r is rational
1 if r is irrational
is continuous nowhere. For, given any number o and any c 0, the interval (o 3c. o +c)
contains both innitely many rational and innitely many irrational numbers. Since 1(o) = 0 or 1, there are innitely many
numbers r with 0 < |r 3o| < c and |1(r) 31(o)| = 1. Thus, lim
i<a
1(r) 6= 1(o). [In fact, lim
i<a
1(r) does not even exist.]
61. If there is such a number, it satises the equation r
3
+1 = r C r
3
3r +1 = 0. Let the left-hand side of this equation be
called 1(r). Now 1(32) = 35 < 0, and 1(31) = 1 0. Note also that 1(r) is a polynomial, and thus continuous. So by the
Intermediate Value Theorem, there is a number c between 32 and 31 such that 1(c) = 0, so that c = c
3
+ 1.
63. 1(r) = r
4
sin(1r) is continuous on (3". 0) (0. ") since it is the product of a polynomial and a composite of a
trigonometric function and a rational function. Now since 31 $ sin(1r) $ 1, we have 3r
4
$ r
4
sin(1r) $ r
4
. Because
60

lim
i<0
(3r
4
) = 0 and lim
i<0
r
4
= 0, the Squeeze Theorem gives us lim
i<0
(r
4
sin(1r)) = 0, which equals 1(0). Thus, 1 is
continuous at 0 and, hence, on (3". ").
65. Dene &(t) to be the monks distance from the monastery, as a function of time, on the rst day, and dene d(t) to be his
distance from the monastery, as a function of time, on the second day. Let 1 be the distance from the monastery to the top of
the mountain. From the given information we know that &(0) = 0, &(12) = 1, d(0) = 1 and d(12) = 0. Now consider the
function & 3d, which is clearly continuous. We calculate that (& 3d)(0) = 31 and (& 3d)(12) = 1. So by the
Intermediate Value Theorem, there must be some time t
0
between 0 and 12 such that (& 3d)(t
0
) = 0 C &(t
0
) = d(t
0
).
So at time t
0
after 7:00 AM, the monk will be at the same place on both days.
2.6 Limits at Infinity; Horizontal Asymptotes
1. (a) As r becomes large, the values of 1(r) approach 5.
(b) As r becomes large negative, the values of 1(r) approach 3.
3. (a) lim
i<2
1(r) = " (b) lim
i<31
1(r) = " (c) lim

i<31
+
1(r) = 3"
(d) lim
i<"
1(r) = 1 (e) lim
i<3"
1(r) = 2 (f ) Vertical: r = 31, r = 2; Horizontal: j = 1, j = 2
5. 1(0) = 0, 1(1) = 1,
lim
i<"
1(r) = 0,
1 is odd
7. lim
i<2
1(r) = 3", lim
i<"
1(r) = ",
lim
i<3"
1(r) = 0, lim
i<0
+
1(r) = ",
lim
i<0
1(r) = 3"
9. 1(0) = 3, lim
i<0
1(r) = 4,
lim
i<0
+
1(r) = 2,
lim
i<3"
1(r) = 3", lim
i<4
1(r) = 3",
lim
i<4
+
1(r) = ", lim
i<"
1(r) = 3
11. If 1(r) = r
2
2
i
, then a calculator gives 1(0) = 0, 1(1) = 0.5, 1(2) = 1, 1(3) = 1.125, 1(4) = 1, 1(5) = 0.78125,
1(6) = 0.5625, 1(7) = 0.3828125, 1(8) = 0.25, 1(9) = 0.158203125, 1(10) = 0.09765625, 1(20) E 0.00038147,
1(50) E 2.2204 10
312
, 1(100) E 7.8886 10
327
.
It appears that lim
i<"
r
2
2
i
= 0.
SECTION 2.6 LIMITS AT INFINITY; HORIZONTAL ASYMPTOTES

61
13. lim
i<"
3r
2
3r + 4
2r
2
+ 5r 38
= lim
i<"
(3r
2
3r + 4)r
2
(2r
2
+ 5r 38)r
2
[divide both the numerator and denominator by r
2
(the highest power of r thatappears in the denominator)]
=
lim
i<"
(3 31r + 4r
2
)
lim
i<"
(2 + 5r 38r
2
)
[Limit Law 5]
=
lim
i<"
3 3 lim
i<"
(1r) + lim
i<"
(4r
2
)
lim
i<"
2 + lim
i<"
(5r) 3 lim
i<"
(8r
2
)
[Limit Laws 1 and 2]
=
3 3 lim
i<"
(1r) + 4 lim
i<"
(1r
2
)
2 + 5 lim
i<"
(1r) 38 lim
i<"
(1r
2
)
[Limit Laws 7 and 3]
=
3 30 + 4(0)
2 + 5(0) 38(0)
[Theorem 5 of Section 2.5]
=
3
2
15. lim
i<"
1
2r + 3
= lim
i<"
1r
(2r + 3)r
=
lim
i<"
(1r)
lim
i<"
(2 + 3r)
=
lim
i<"
(1r)
lim
i<"
2 + 3 lim
i<"
(1r)
=
0
2 + 3(0)
=
0
2
= 0
17. lim
i<3"
1 3r 3r
2
2r
2
37
= lim
i<3"
(1 3r 3r
2
)r
2
(2r
2
37)r
2
=
lim
i<3"
(1r
2
31r 31)
lim
i<3"
(2 37r
2
)
=
lim
i<3"
(1r
2
) 3 lim
i<3"
(1r) 3 lim
i<3"
1
lim
i<3"
2 37 lim
i<3"
(1r
2
)
=
0 30 31
2 37(0)
= 3
1
2
19. Divide both the numerator and denominator by r
3
(the highest power of r that occurs in the denominator).
lim
i<"
r
3
+ 5r
2r
3
3r
2
+ 4
= lim
i<"
r
3
+ 5r
r
3
2r
3
3r
2
+ 4
r
3
= lim
i<"
1 +
5
r
2
2 3
1
r
+
4
r
3
=
lim
i<"
1 +
5
r
2
lim
i<"
2 3
1
r
+
4
r
3
=
lim
i<"
1 + 5 lim
i<"
1
r
2
lim
i<"
2 3 lim
i<"
1
r
+ 4 lim
i<"
1
r
3
=
1 + 5(0)
2 30 + 4(0)
=
1
2
21. First, multiply the factors in the denominator. Then divide both the numerator and denominator by &
4
.
lim
u<"
4&
4
+ 5
(&
2
32)(2&
2
31)
= lim
u<"
4&
4
+ 5
2&
4
35&
2
+ 2
= lim
u<"
4&
4
+ 5
&
4
2&
4
35&
2
+ 2
&
4
= lim
u<"
4 +
5
&
4
2 3
5
&
2
+
2
&
4
=
lim
u<"
4 +
5
&
4
lim
u<"
2 3
5
&
2
+
2
&
4
=
lim
u<"
4 + 5 lim
u<"
1
&
4
lim
u<"
2 35 lim
u<"
1
&
2
+ 2 lim
u<"
1
&
4
=
4 + 5(0)
2 35(0) + 2(0)
=
4
2
= 2
62

23. lim
i<"
I
9r
6
3r
r
3
+ 1
= lim
i<"
I
9r
6
3rr
3
(r
3
+ 1)r
3
=
lim
i<"
(9r
6
3r)r
6
lim
i<"
(1 + 1r
3
)
[since r
3
=
I
r
6
for r 0]
=
lim
i<"
9 31r
5
lim
i<"
1 + lim
i<"
(1r
3
)
=
lim
i<"
9 3 lim
i<"
(1r
5
)
1 + 0
=
I
9 30 = 3
25. lim
i<"
I
9r
2
+r 33r
= lim
i<"
I
9r
2
+r 33r
I
9r
2
+r + 3r
I
9r
2
+r + 3r
= lim
i<"
I
9r
2
+r
2
3(3r)
2
I
9r
2
+r + 3r
= lim
i<"
9r
2
+r
39r
2
I
9r
2
+r + 3r
= lim
i<"
r
I
9r
2
+r + 3r

1r
1r
= lim
i<"
rr
9r
2
r
2
+rr
2
+ 3rr
= lim
i<"
1
9 + 1r + 3
=
1
I
9 + 3
=
1
3 + 3
=
1
6
27. lim
i<"
I
r
2
+or 3
I
r
2
+/r
= lim
i<"
I
r
2
+or 3
I
r
2
+/r
I
r
2
+or +
I
r
2
+/r
I
r
2
+or +
I
r
2
+/r
= lim
i<"
(r
2
+or) 3(r
2
+/r)
I
r
2
+or +
I
r
2
+/r
= lim
i<"
[(o 3/)r]r
I
r
2
+or +
I
r
2
+/r
I
r
2
= lim
i<"
o 3/
1 +or +
1 +/r
=
o 3/
I
1 + 0 +
I
1 + 0
=
o 3/
2
29. lim
i<"
r +r
3
+r
5
1 3r
2
+r
4
= lim
i<"
(r +r
3
+r
5
)r
4
(1 3r
2
+r
4
)r
4
[divide by the highest power of r in the denominator]
= lim
i<"
1r
3
+ 1r +r
1r
4
31r
2
+ 1
= "
because (1r
3
+ 1r +r) <"and (1r
4
31r
2
+ 1) <1 as r <".
31. lim
i<3"
(r
4
+r
5
) = lim
i<3"
r
5
(
1
i
+ 1) [factor out the largest power of r] = 3" because r
5
<3"and 1r + 1 <1
as r <3".
Or: lim
i<3"
r
4
+r
5
= lim
i<3"
r
4
(1 +r) = 3".
33. lim
i<"
1 3c
i
1 + 2c
i
= lim
i<"
(1 3c
i
)c
i
(1 + 2c
i
)c
i
= lim
i<"
1c
i
31
1c
i
+ 2
=
0 31
0 + 2
= 3
1
2
35. Since 31 $ cos r $ 1 and c
32i
0, we have 3c
32i
$ c
32i
cos r $ c
32i
. We know that lim
i<"
(3c
32i
) = 0 and
lim
i<"
c
32i
= 0, so by the Squeeze Theorem, lim

i<"
(c
32i
cos r) = 0.
37. (a)
From the graph of 1(r) =
I
r
2
+r + 1 +r, we estimate
the value of lim
i<3"
1(r) to be 30.5.
(b)
r 1(r)
310,000 30.4999625
3100,000 30.4999962
31,000,000 30.4999996
From the table, we estimate the limit
to be 30.5.

63
(c) lim
i<3"
I
r
2
+r + 1 +r
= lim
i<3"
I
r
2
+r + 1 +r
I
r
2
+r + 1 3r
I
r
2
+r + 1 3r
= lim
i<3"
r
2
+r + 1
3r
2
I
r
2
+r + 1 3r
= lim
i<3"
(r + 1)(1r)
I
r
2
+r + 1 3r
(1r)
= lim
i<3"
1 + (1r)
3
1 + (1r) + (1r
2
) 31
=
1 + 0
3
I
1 + 0 + 0 31
= 3
1
2
Note that for r < 0, we have
I
r
2
= |r| = 3r, so when we divide the radical by r, with r < 0, we get
1
r
I
r
2
+r + 1 = 3
1
I
r
2
I
r
2
+r + 1 = 3
1 + (1r) + (1r
2
).
39. lim
i<"
2r + 1
r 32
= lim
i<"
2r + 1
r
r 32
r
= lim
i<"
2 +
1
r
1 3
2
r
=
lim
i<"
2 +
1
r
lim
i<"
1 3
2
r
=
lim
i<"
2 + lim
i<"
1
r
lim
i<"
1 3 lim
i<"
2
r
=
2 + 0
1 30
= 2, so j = 2 is a horizontal asymptote.
The denominator r 32 is zero when r = 2 and the numerator is not zero, so we
investigate j = 1(r) =
2r + 1
r 32
as r approaches 2. lim
i<2
1(r) = 3"because as
r <2
3
the numerator is positive and the denominator approaches 0 through
negative values. Similarly, lim
i<2
+
1(r) = ". Thus, r = 2 is a vertical asymptote.
The graph conrms our work.
41. lim
i<"
2r
2
+r 31
r
2
+r 32
= lim
i<"
2r
2
+r 31
r
2
r
2
+r 32
r
2
= lim
i<"
2 +
1
r
3
1
r
2
1 +
1
r
3
2
r
2
=
lim
i<"
2 +
1
r
3
1
r
2
lim
i<"
1 +
1
r
3
2
r
2
=
lim
i<"
2 + lim
i<"
1
r
3 lim
i<"
1
r
2
lim
i<"
1 + lim
i<"
1
r
32 lim
i<"
1
r
2
=
2 + 0 30
1 + 0 32(0)
j = 1(r) =
2r
2
+r 31
r
2
+r 32
=
(2r 31)(r + 1)
(r + 2)(r 31)
, so lim
i<32
1(r) = ",
lim
i<32
+
1(r) = 3", lim
i<1
1(r) = 3", and lim

i<1
+
1(r) = ". Thus, r = 32
and r = 1 are vertical asymptotes. The graph conrms our work.
43. j = 1(r) =
r
3
3r
r
2
36r + 5
=
r(r
2
31)
(r 31)(r 35)
=
r(r + 1)(r 31)
(r 31)(r 35)
=
r(r + 1)
r 35
= o(r) for r 6= 1.
The graph of o is the same as the graph of 1 with the exception of a hole in the
graph of 1 at r = 1. By long division, o(r) =
r
2
+r
r 35
= r + 6 +
30
r 35
.
As r <", o(r) <", so there is no horizontal asymptote. The denominator
of o is zero when r = 5. lim
i<5
o(r) = 3"and lim

i<5
+
o(r) = ", so r = 5 is a
vertical asymptote. The graph conrms our work.
64

45. From the graph, it appears j = 1 is a horizontal asymptote.
lim
i<"
3r
3
+ 500r
2
r
3
+ 500r
2
+ 100r + 2000
= lim
i<"
3r
3
+ 500r
2
r
3
r
3
+ 500r
2
+ 100r + 2000
r
3
= lim
i<"
3 + (500r)
1 + (500r) + (100r
2
) + (2000r
3
)
=
3 + 0
1 + 0 + 0 + 0
The discrepancy can be explained by the choice of the viewing window. Try
[3100,000. 100,000] by [31. 4] to get a graph that lends credibility to our
calculation that j = 3 is a horizontal asymptote.
47. Lets look for a rational function.
(1) lim
i<"
1(r) = 0 i degree of numerator < degree of denominator
(2) lim
i<0
1(r) = 3" i there is a factor of r
2
in the denominator (not just r, since that would produce a sign
change at r = 0), and the function is negative near r = 0.
(3) lim
i<3
1(r) = "and lim

i<3
+
1(r) = 3" i vertical asymptote at r = 3; there is a factor of (r 33) in the
denominator.
(4) 1(2) = 0 i 2 is an r-intercept; there is at least one factor of (r 32) in the numerator.
Combining all of this information and putting in a negative sign to give us the desired left- and right-hand limits gives us
1(r) =
2 3r
r
2
(r 33)
as one possibility.
49. j = 1(r) = r
4
3r
6
= r
4
(1 3r
2
) = r
4
(1 +r)(1 3r). The j-intercept is
1(0) = 0. The r-intercepts are 0, 31, and 1 [found by solving 1(r) = 0 for r].
Since r
4
0 for r 6= 0, 1 doesnt change sign at r = 0. The function does change
sign at r = 31 and r = 1. As r <", 1(r) = r
4
(1 3r
2
) approaches 3"
because r
4
<"and (1 3r
2
) <3".
51. j = 1(r) = (3 3r)(1 +r)
2
(1 3r)
4
. The j-intercept is 1(0) = 3(1)
2
(1)
4
= 3.
The r-intercepts are 3, 31, and 1. There is a sign change at 3, but not at 31 and 1.
When r is large positive, 3 3r is negative and the other factors are positive, so
lim
i<"
1(r) = 3". When r is large negative, 3 3r is positive, so
lim
i<3"
1(r) = ".
53. (a) Since 31 $ sinr $ 1 for all r. 3
1
r
$
sinr
r
$
1
r
for r 0. As r <", 31r <0 and 1r <0, so by the Squeeze
Theorem, (sinr)r <0. Thus, lim
i<"
sinr
r
= 0.

65
(b) From part (a), the horizontal asymptote is j = 0. The function
j = (sinr)r crosses the horizontal asymptote whenever sinr = 0;
that is, at r = n for every integer n. Thus, the graph crosses the
asymptote an innite number of times.
55. Divide the numerator and the denominator by the highest power of r in Q(r).
(a) If deg 1 < deg Q, then the numerator <0 but the denominator doesnt. So lim
i<"
[1(r)Q(r)] = 0.
(b) If deg 1 deg Q, then the numerator <"but the denominator doesnt, so lim
i<"
[1(r)Q(r)] = "
(depending on the ratio of the leading coefcients of 1 and Q).
57. lim
i<"
5
I
r
I
r 31

1
I
r
1
I
r
= lim
i<"
5
1 3(1r)
=
5
I
1 30
= 5 and
lim
i<"
10c
i
321
2c
i

1c
i
1c
i
= lim
i<"
10 3(21c
i
)
2
=
10 30
2
= 5. Since
10c
i
321
2c
i
< 1(r) <
5
I
r
I
r 31
,
we have lim
i<"
59. (a) lim
I<"
(t) = lim
I<"
1 3c
3gI'r
=
W
(1 30) =
W
(b) We graph (t) = 1 3c
39.8I
and (t) = 0.99
W
, or in this case,
(t) = 0.99. Using an intersect feature or zooming in on the point of
intersection, we nd that t E 0.47 s.
61. Let o(r) =
3r
2
+ 1
2r
2
+r + 1
and 1(r) = |o(r) 31.5|. Note that
lim
i<"
o(r) =
3
2
and lim
i<"
1(r) = 0. We are interested in nding the
r-value at which 1(r) < 0.05. From the graph, we nd that r E 14.804,
so we choose ` = 15 (or any larger number).
63. For - = 0.5, we need to nd ` such that
I
4r
2
+ 1
r + 1
3(32)
< 0.5 C
32.5 <
I
4r
2
+ 1
r + 1
< 31.5 whenever r $ `. We graph the three parts of this
inequality on the same screen, and see that the inequality holds for r $ 36. So we
choose ` = 36 (or any smaller number).
For - = 0.1, we need 32.1 <
I
4r
2
+ 1
r + 1
< 31.9 whenever r $ `. From the
graph, it seems that this inequality holds for r $ 322. So we choose ` = 322
(or any smaller number).
66

65. (a) 1r
2
< 0.0001 C r
2
10.0001 = 10 000 C r 100 (r 0)
(b) If - 0 is given, then 1r
2
< - C r
2
1- C r 1
I
-. Let ` = 1
I
-.
Then r ` i r
1
I
-
i
1
r
2
30
=
1
r
2
< -, so lim
i<"
1
r
2
= 0.
67. For r < 0, |1r 30| = 31r. If - 0 is given, then 31r < - C r < 31-.
Take ` = 31-. Then r < ` i r < 31- i |(1r) 30| = 31r < -, so lim
i<3"
(1r) = 0.
69. Given A 0, we need ` 0 such that r ` i c
i
A. Now c
i
A C r lnA, so take
` = max(1. lnA). (This ensures that ` 0.) Then r ` = max(1. lnA) i c
i
max(c. A) D A,
so lim
i<"
c
i
= ".
71. Suppose that lim
i<"
1(r) = 1. Then for every - 0 there is a corresponding positive number ` such that |1(r) 31| < -
whenever r `. If t = 1r, then r ` C 0 < 1r < 1` C 0 < t < 1`. Thus, for every - 0 there is a
corresponding c 0 (namely 1`) such that |1(1t) 31| < - whenever 0 < t < c. This proves that
lim
I<0
+
1(1t) = 1 = lim
i<"
1(r).
Now suppose that lim
i<3"
1(r) = 1. Then for every - 0 there is a corresponding negative number ` such that
|1(r) 31| < - whenever r < `. If t = 1r, then r < ` C 1` < 1r < 0 C 1` < t < 0. Thus, for every
- 0 there is a corresponding c 0 (namely 31`) such that |1(1t) 31| < - whenever 3c < t < 0. This proves that
lim
I<0
1(1t) = 1 = lim
i<3"
1(r).
2.7 Derivatives and Rates of Change
1. (a) This is just the slope of the line through two points: i
TC
=
{j
{r
=
1(r) 31(3)
r 33
.
(b) This is the limit of the slope of the secant line 1Q as Qapproaches 1: i = lim
i<3
1(r) 31(3)
r 33
.
3. (a) (i) Using Denition 1 with 1(r) = 4r 3r
2
and 1(1. 3),
i= lim
i<a
1(r) 31(o)
r 3o
= lim
i<1
(4r 3r
2
) 33
r 31
= lim
i<1
3(r
2
34r + 3)
r 31
= lim
i<1
3(r 31)(r 33)
r 31
= lim
i<1
(3 3r) = 3 31 = 2
(ii) Using Equation 2 with 1(r) = 4r 3r
2
and 1(1. 3),
i= lim
I<0
1(o +/) 31(o)
/
= lim
I<0
1(1 +/) 31(1)
/
= lim
I<0
4(1 +/) 3(1 +/)

2
33
/
= lim
I<0
4 + 4/ 31 32/ 3/
2
33
/
= lim
I<0
3/
2
+ 2/
/
= lim
I<0
/(3/ + 2)
/
= lim
I<0
(3/ + 2) = 2
(b) An equation of the tangent line is j 31(o) = 1
0
(o)(r 3o) i j 31(1) = 1
0
(1)(r 31) i j 33 = 2(r 31),
or j = 2r + 1.
SECTION 2.7 DERIVATIVES AND RATES OF CHANGE

67
(c) The graph of j = 2r + 1 is tangent to the graph of j = 4r 3r
2
at the
point (1. 3). Now zoom in toward the point (1. 3) until the parabola and
the tangent line are indistiguishable.
5. Using (1) with 1(r) =
r 31
r 32
and 1(3. 2),
i= lim
i<a
1(r) 31(o)
r 3o
= lim
i<3
r 31
r 32
32
r 33
= lim
i<3
r 31 32(r 32)
r 32
r 33
= lim
i<3
3 3r
(r 32)(r 33)
= lim
i<3
31
r 32
=
31
1
= 31
Tangent line: j 32 = 31(r 33) C j 32 = 3r + 3 C j = 3r + 5
7. Using (1), i = lim
i<1
I
r 3
I
1
r 31
= lim
i<1
(
I
r 31)(
I
r + 1)
(r 31)(
I
r + 1)
= lim
i<1
r 31
(r 31)(
I
r + 1)
= lim
i<1
1
I
r + 1
=
1
2
.
Tangent line: j 31 =
1
2
(r 31) C j =
1
2
r +
1
2
9. (a) Using (2) with j = 1(r) = 3 + 4r
2
32r
3
,
i= lim
I<0
1(o +/) 31(o)
/
= lim
I<0
3 + 4(o +/)
2
32(o +/)
3
3(3 + 4o
2
32o
3
)
/
= lim
I<0
3 + 4(o
2
+ 2o/ +/
2
) 32(o
3
+ 3o
2
/ + 3o/
2
+/
3
) 33 34o
2
+ 2o
3
/
= lim
I<0
3 + 4o
2
+ 8o/ + 4/
2
32o
3
36o
2
/ 36o/
2
32/
3
33 34o
2
+ 2o
3
/
= lim
I<0
8o/ + 4/
2
36o
2
/ 36o/
2
32/
3
/
= lim
I<0
/(8o + 4/ 36o
2
36o/ 32/
2
)
/
= lim
I<0
(8o + 4/ 36o
2
36o/ 32/
2
) = 8o 36o
2
(b) At (1. 5): i = 8(1) 36(1)
2
= 2, so an equation of the tangent line
is j 35 = 2(r 31) C j = 2r + 3.
At (2. 3): i = 8(2) 36(2)
2
= 38, so an equation of the tangent
line is j 33 = 38(r 32) C j = 38r + 19.
(c)
11. (a) The particle is moving to the right when c is increasing; that is, on the intervals (0. 1) and (4. 6). The particle is moving to
the left when c is decreasing; that is, on the interval (2. 3). The particle is standing still when c is constant; that is, on the
intervals (1. 2) and (3. 4).
68

(b) The velocity of the particle is equal to the slope of the tangent line of the
graph. Note that there is no slope at the corner points on the graph. On the
interval (0. 1). the slope is
3 30
1 30
= 3. On the interval (2. 3), the slope is
1 33
3 32
= 32. On the interval (4. 6), the slope is
3 31
6 34
= 1.
13. Let c(t) = 40t 316t
2
.
(2) = lim
I<2
c(t) 3c(2)
t 32
= lim
I<2
40t 316t
2
316
t 32
= lim
I<2
316t
2
+ 40t 316
t 32
= lim
I<2
38
2t
2
35t + 2
t 32
= lim
I<2
38(t 32)(2t 31)
t 32
= 38 lim
I<2
(2t 31) = 38(3) = 324
Thus, the instantaneous velocity when t = 2 is 324 fts.
15. (o) = lim
I<0
c(o +/) 3c(o)
/
= lim
I<0
1
(o +/)
2
3
1
o
2
/
= lim
I<0
o
2
3(o +/)
2
o
2
(o +/)
2
/
= lim
I<0
o
2
3(o
2
+ 2o/ +/
2
)
/o
2
(o +/)
2
= lim
I<0
3(2o/ +/
2
)
/o
2
(o +/)
2
= lim
I<0
3/(2o +/)
/o
2
(o +/)
2
= lim
I<0
3(2o +/)
o
2
(o +/)
2
=
32o
o
2
o
2
=
32
o
3
ms
So (1) =
32
1
3
= 32 ms, (2) =
32
2
3
= 3
1
4
ms, and (3) =
32
3
3
= 3
2
27
ms.
17. o
0
(0) is the only negative value. The slope at r = 4 is smaller than the slope at r = 2 and both are smaller than the slope at
r = 32. Thus, o
0
(0) < 0 < o
0
(4) < o
0
(2) < o
0
(32).
19. We begin by drawing a curve through the origin with a
slope of 3 to satisfy 1(0) = 0 and 1
0
(0) = 3. Since
1
0
(1) = 0, we will round off our gure so that there is
a horizontal tangent directly over r = 1. Last, we
make sure that the curve has a slope of 31 as we pass
over r = 2. Two of the many possibilities are shown.
21. Using Denition 2 with 1(r) = 3r
2
35r and the point (2. 2), we have
1
0
(2) = lim
I<0
1(2 +/) 31(2)
/
= lim
I<0
3(2 +/)
2
35(2 +/)
32
/
= lim
I<0
(12 + 12/ + 3/
2
310 35/) 32
/
= lim
I<0
3/
2
+ 7/
/
= lim
I<0
(3/ + 7) = 7
So an equation of the tangent line at (2. 2) is j 32 = 7(r 32) or j = 7r 312.
SECTION 2.7 DERIVATIVES AND RATES OF CHANGE

69
23. (a) Using Denition 2 with 1(r) = 5r(1 +r
2
) and the point (2. 2), we have
1
0
(2) = lim
I<0
1(2 +/) 31(2)
/
= lim
I<0
5(2 +/)
1 + (2 +/)
2
32
/
= lim
I<0
5/ + 10
/
2
+ 4/ + 5
32
/
= lim
I<0
5/ + 10 32(/
2
+ 4/ + 5)
/
2
+ 4/ + 5
/
= lim
I<0
32/
2
33/
/(/
2
+ 4/ + 5)
= lim
I<0
/(32/ 33)
/(/
2
+ 4/ + 5)
= lim
I<0
32/ 33
/
2
+ 4/ + 5
=
33
5
So an equation of the tangent line at (2. 2) is j 32 = 3
3
5
(r 32) or j = 3
3
5
r +
16
5
.
(b)
25. Use Denition 2 with 1(r) = 3 32r + 4r
2
.
1
0
(o) = lim
I<0
1(o +/) 31(o)
/
= lim
I<0
[3 32(o +/) + 4(o +/)
2
] 3(3 32o + 4o
2
)
/
= lim
I<0
(3 32o 32/ + 4o
2
+ 8o/ + 4/
2
) 3(3 32o + 4o
2
)
/
= lim
I<0
32/ + 8o/ + 4/
2
/
= lim
I<0
/(32 + 8o + 4/)
/
= lim
I<0
(32 + 8o + 4/) = 32 + 8o
27. Use Denition 2 with 1(t) = (2t + 1)(t + 3).
1
0
(o) = lim
I<0
1(o +/) 31(o)
/
= lim
I<0
2(o +/) + 1
(o +/) + 3
3
2o + 1
o + 3
/
= lim
I<0
(2o + 2/ + 1)(o + 3) 3(2o + 1)(o +/ + 3)
/(o +/ + 3)(o + 3)
= lim
I<0
(2o
2
+ 6o + 2o/ + 6/ +o + 3) 3(2o
2
+ 2o/ + 6o +o +/ + 3)
/(o +/ + 3)(o + 3)
= lim
I<0
5/
/(o +/ + 3)(o + 3)
= lim
I<0
5
(o +/ + 3)(o + 3)
=
5
(o + 3)
2
29. Use Denition 2 with 1(r) = 1
I
r + 2.
1
0
(o) = lim
I<0
1(o +/) 31(o)
/
= lim
I<0
1
(o +/) + 2
3
1
I
o + 2
/
= lim
I<0
I
o + 2 3
I
o +/ + 2
I
o +/ + 2
I
o + 2
/
= lim
I<0
I
o + 2 3
I
o +/ + 2
/
I
o +/ + 2
I
o + 2

I
o + 2 +
I
o +/ + 2
I
o + 2 +
I
o +/ + 2
= lim
I<0
(o + 2) 3(o +/ + 2)
/
I
o +/ + 2
I
o + 2
I
o + 2 +
I
o +/ + 2
= lim
I<0
3/
/
I
o +/ + 2
I
o + 2
I
o + 2 +
I
o +/ + 2
= lim
I<0
31
I
o +/ + 2
I
o + 2
I
o + 2 +
I
o +/ + 2
=
31
I
o + 2
2
I
o + 2
= 3
1
2(o + 2)
3'2
Note that the answers to Exercises 31 36 are not unique.
31. By Denition 2, lim
I<0
(1 +/)
10
31
/
= 1
0
(1), where 1(r) = r
10
and o = 1.
Or: By Denition 2, lim
I<0
(1 +/)
10
31
/
= 1
0
(0), where 1(r) = (1 +r)
10
and o = 0.
70

33. By Equation 3, lim
i<5
2
i
332
r 35
= 1
0
(5), where 1(r) = 2
i
and o = 5.
35. By Denition 2, lim
I<0
cos( +/) + 1
/
= 1
0
(), where 1(r) = cos r and o = .
Or: By Denition 2, lim
I<0
cos( +/) + 1
/
= 1
0
(0), where 1(r) = cos( +r) and o = 0.
37. (5) = 1
0
(5) = lim
I<0
1(5 +/) 31(5)
/
= lim
I<0
[100 + 50(5 +/) 34.9(5 +/)
2
] 3[100 + 50(5) 34.9(5)
2
]
/
= lim
I<0
(100 + 250 + 50/ 34.9/
2
349/ 3122.5) 3(100 + 250 3122.5)
/
= lim
I<0
34.9/
2
+/
/
= lim
I<0
/(34.9/ + 1)
/
= lim
I<0
(34.9/ + 1) = 1 ms
The speed when t = 5 is |1| = 1 ms.
39. The sketch shows the graph for a room temperature of 72
and a refrigerator
temperature of 38
. The initial rate of change is greater in magnitude than the

rate of change after an hour.
41. (a) (i) [2000. 2002]:
1(2002) 31(2000)
2002 32000
=
77 355
2
=
22
2
= 11 percentyear
(ii) [2000. 2001]:
1(2001) 31(2000)
2001 32000
=
68 355
1
= 13 percentyear
(iii) [1999. 2000]:
1(2000) 31(1999)
2000 31999
=
55 339
1
= 16 percentyear
(b) Using the values from (ii) and (iii), we have
13 + 16
2
= 14.5 percentyear.
(c) Estimating as (1999. 40) and 1 as (2001. 70), the slope at 2000 is
70 340
2001 31999
=
30
2
= 15 percentyear.
43. (a) (i)
{C
{r
=
C(105) 3C(100)
105 3100
=
6601.25 36500
5
= $20.25unit.
(ii)
{C
{r
=
C(101) 3C(100)
101 3100
=
6520.05 36500
1
= $20.05unit.
(b)
C(100 +/) 3C(100)
/
=
5000 + 10(100 +/) + 0.05(100 +/)

2
36500
/
=
20/ + 0.05/
2
/
= 20 + 0.05/, / 6= 0
So the instantaneous rate of change is lim
I<0
C(100 +/) 3C(100)
/
= lim
I<0
(20 + 0.05/) = $20unit.
45. (a) 1
0
(r) is the rate of change of the production cost with respect to the number of ounces of gold produced. Its units are
dollars per ounce.
SECTION 2.8 THE DERIVATIVE AS A FUNCTION

71
(b) After 800 ounces of gold have been produced, the rate at which the production cost is increasing is $17ounce. So the cost
of producing the 800th (or 801st) ounce is about $17.
(c) In the short term, the values of 1
0
(r) will decrease because more efcient use is made of start-up costs as r increases. But
eventually 1
0
(r) might increase due to large-scale operations.
47. T
0
(10) is the rate at which the temperature is changing at 10:00 AM. To estimate the value of T
0
(10), we will average the
difference quotients obtained using the times t = 8 and t = 12. Let =
T(8) 3T(10)
8 310
=
72 381
32
= 4.5 and
1 =
T(12) 3T(10)
12 310
=
88 381
2
= 3.5. Then T
0
(10) = lim
I<10
T(t) 3T(10)
t 310
E
+1
2
=
4.5 + 3.5
2
= 4
Fh.
49. (a) o
0
(T) is the rate at which the oxygen solubility changes with respect to the water temperature. Its units are (mgL)
C.
(b) For T = 16
C, it appears that the tangent line to the curve goes through the points (0. 14) and (32. 6). So
o
0
(16) E
6 314
32 30
= 3
8
32
= 30.25 (mgL)
C. This means that as the temperature increases past 16
C, the oxygen
solubility is decreasing at a rate of 0.25 (mgL)
C.
51. Since 1(r) = rsin(1r) when r 6= 0 and 1(0) = 0, we have
1
0
(0) = lim
I<0
1(0 +/) 31(0)
/
= lim
I<0
/sin(1/) 30
/
= lim
I<0
sin(1/). This limit does not exist since sin(1/) takes the
values 31 and 1 on any interval containing 0. (Compare with Example 4 in Section 2.2.)
2.8 The Derivative as a Function
1. It appears that 1 is an odd function, so 1
0
will be an even
functionthat is, 1
0
(3o) = 1
0
(o).
(a) 1
0
(33) E 1.5
(c) 1
0
(31) E 0
(e) 1
0
(1) E 0
(g) 1
0
(3) E 1.5
(b) 1
0
(32) E 1
(d) 1
0
(0) E 34
(f ) 1
0
(2) E 1
3. (a)
0
= II, since from left to right, the slopes of the tangents to graph (a) start out negative, become 0, then positive, then 0, then
negative again. The actual function values in graph II follow the same pattern.
(b)
0
= IV, since from left to right, the slopes of the tangents to graph (b) start out at a xed positive quantity, then suddenly
become negative, then positive again. The discontinuities in graph IV indicate sudden changes in the slopes of the tangents.
(c)
0
= I, since the slopes of the tangents to graph (c) are negative for r < 0 and positive for r 0, as are the function values of
graph I.
(d)
0
= III, since from left to right, the slopes of the tangents to graph (d) are positive, then 0, then negative, then 0, then
positive, then 0, then negative again, and the function values in graph III follow the same pattern.
72

Hints for Exercises 4 11: First plot r-intercepts on the graph of 1
0
for any horizontal tangents on the graph of 1. Look for any corners on the graph
of 1there will be a discontinuity on the graph of 1
0
. On any interval where 1 has a tangent with positive (or negative) slope, the graph of 1
0
will be
positive (or negative). If the graph of the function is linear, the graph of 1
0
will be a horizontal line.
5. 7.
9. 11.
13. It appears that there are horizontal tangents on the graph of A for t = 1963
and t = 1971. Thus, there are zeros for those values of t on the graph of
A
0
. The derivative is negative for the years 1963 to 1971.
15.
The slope at 0 appears to be 1 and the slope at 1
appears to be 2.7. As r decreases, the slope gets
closer to 0. Since the graphs are so similar, we might
guess that 1
0
(r) = c
i
.

73
17. (a) By zooming in, we estimate that 1
0
(0) = 0, 1
0
1
2
= 1, 1
0
(1) = 2,
and 1
0
(2) = 4.
(b) By symmetry, 1
0
(3r) = 31
0
(r). So 1
0
3
1
2
= 31, 1
0
(31) = 32,
and 1
0
(32) = 34.
(c) It appears that 1
0
(r) is twice the value of r, so we guess that 1
0
(r) = 2r.
(d) 1
0
(r) = lim
I<0
1(r +/) 31(r)
/
= lim
I<0
(r +/)
2
3r
2
/
= lim
I<0
r
2
+ 2/r +/
2
3r
2
/
= lim
I<0
2/r +/
2
/
= lim
I<0
/(2r +/)
/
= lim
I<0
(2r +/) = 2r
19. 1
0
(r) = lim
I<0
1(r +/) 31(r)
/
= lim
I<0
1
2
(r +/) 3
1
3
1
2
r 3
1
3
/
= lim
I<0
1
2
r +
1
2
/ 3
1
3
3
1
2
r +
1
3
/
= lim
I<0
1
2
/
/
= lim
I<0
1
2
=
1
2
Domain of 1 = domain of 1
0
= R.
21. 1
0
(t) = lim
I<0
1(t +/) 31(t)
/
= lim
I<0
5(t +/) 39(t +/)

2
3(5t 39t
2
)
/
= lim
I<0
5t + 5/ 39(t
2
+ 2t/ +/
2
) 35t + 9t
2
/
= lim
I<0
5t + 5/ 39t
2
318t/ 39/
2
35t + 9t
2
/
= lim
I<0
5/ 318t/ 39/
2
/
= lim
I<0
/(5 318t 39/)
/
= lim
I<0
(5 318t 39/) = 5 318t
0
= R.
23. 1
0
(r) = lim
I<0
1(r +/) 31(r)
/
= lim
I<0
(r +/)
3
33(r +/) + 5
3(r
3
33r + 5)
/
= lim
I<0
r
3
+ 3r
2
/ + 3r/
2
+/
3
33r 33/ + 5
r
3
33r + 5
/
= lim
I<0
3r
2
/ + 3r/
2
+/
3
33/
/
= lim
I<0
/
3r
2
+ 3r/ +/
2
33
/
= lim
I<0
3r
2
+ 3r/ +/
2
33
= 3r
2
33
0
= R.
25. o
0
(r) = lim
I<0
o(r +/) 3o(r)
/
= lim
I<0
1 + 2(r +/) 3
I
1 + 2r
/
1 + 2(r +/) +
I
1 + 2r
1 + 2(r +/) +
I
1 + 2r
= lim
I<0
(1 + 2r + 2/) 3(1 + 2r)
/
1 + 2(r +/) +
I
1 + 2r
= lim
I<0
2
I
1 + 2r + 2/ +
I
1 + 2r
=
2
2
I
1 + 2r
=
1
I
1 + 2r
Domain of o =

3
1
2
. "
, domain of o
0
=

3
1
2
. "
.
74

27. G
0
(t) = lim
I<0
G(t +/) 3G(t)
/
= lim
I<0
4(t +/)
(t +/) + 1
3
4t
t + 1
/
= lim
I<0
4(t +/)(t + 1) 34t(t +/ + 1)
(t +/ + 1)(t + 1)
/
= lim
I<0
4t
2
+ 4/t + 4t + 4/
4t
2
+ 4/t + 4t
/(t +/ + 1)(t + 1)
= lim
I<0
4/
/(t +/ + 1)(t + 1)
= lim
I<0
4
(t +/ + 1)(t + 1)
=
4
(t + 1)
2
Domain of G = domain of G
0
= (3". 31) (31. ").
29. 1
0
(r) = lim
I<0
1(r +/) 31(r)
/
= lim
I<0
(r +/)
4
3r
4
/
= lim
I<0
r
4
+ 4r
3
/ + 6r
2
/
2
+ 4r/
3
+/
4
3r
4
/
= lim
I<0
4r
3
/ + 6r
2
/
2
+ 4r/
3
+/
4
/
= lim
I<0
4r
3
+ 6r
2
/ + 4r/
2
+/
3
= 4r
3
0
= R.
31. (a) 1
0
(r) = lim
I<0
1(r +/) 31(r)
/
= lim
I<0
[(r +/)
4
+ 2(r +/)] 3(r
4
+ 2r)
/
= lim
I<0
r
4
+ 4r
3
/ + 6r
2
/
2
+ 4r/
3
+/
4
+ 2r + 2/ 3r
4
32r
/
= lim
I<0
4r
3
/ + 6r
2
/
2
+ 4r/
3
+/
4
+ 2/
/
= lim
I<0
/(4r
3
+ 6r
2
/ + 4r/
2
+/
3
+ 2)
/
= lim
I<0
(4r
3
+ 6r
2
/ + 4r/
2
+/
3
+ 2) = 4r
3
+ 2
(b) Notice that 1
0
(r) = 0 when 1 has a horizontal tangent, 1
0
(r) is
positive when the tangents have positive slope, and 1
0
(r) is
negative when the tangents have negative slope.
33. (a) l
0
(t) is the rate at which the unemployment rate is changing with respect to time. Its units are percent per year.
(b) To nd l
0
(t), we use lim
I<0
l(t +/) 3l(t)
/
E
l(t +/) 3l(t)
/
for small values of /.
For 1993: l
0
(1993) E
l(1994) 3l(1993)
1994 31993
=
6.1 36.9
1
= 30.80
For 1994: We estimate l
0
(1994) by using / = 31 and / = 1, and then average the two results to obtain a nal estimate.
/ = 31 i l
0
(1994) E
l(1993) 3l(1994)
1993 31994
=
6.9 36.1
31
= 30.80;
/ = 1 i l
0
(1994) E
l(1995) 3l(1994)
1995 31994
=
5.6 36.1
1
= 30.50.
So we estimate that l
0
(1994) E
1
2
[(30.80) + (30.50)] = 30.65.
t 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002
l
0
(t) 30.80 30.65 30.35 30.35 30.45 30.35 30.25 0.25 0.90 1.10

75
35. 1 is not differentiable at r = 34, because the graph has a corner there, and at r = 0, because there is a discontinuity there.
37. 1 is not differentiable at r = 31, because the graph has a vertical tangent there, and at r = 4, because the graph has a corner
there.
39. As we zoom in toward (31. 0), the curve appears more and more like a
straight line, so 1(r) = r +
|r| is differentiable at r = 31. But no

matter how much we zoom in toward the origin, the curve doesnt straighten
outwe cant eliminate the sharp point (a cusp). So 1 is not differentiable
at r = 0.
41. o = 1, / = 1
0
, c = 1
00
. We can see this because where o has a horizontal tangent, / = 0, and where / has a horizontal tangent,
c = 0. We can immediately see that c can be neither 1 nor 1
0
, since at the points where c has a horizontal tangent, neither o
nor / is equal to 0.
43. We can immediately see that o is the graph of the acceleration function, since at the points where o has a horizontal tangent,
neither c nor / is equal to 0. Next, we note that o = 0 at the point where / has a horizontal tangent, so / must be the graph of
the velocity function, and hence, /
0
= o. We conclude that c is the graph of the position function.
45. 1
0
(r) = lim
I<0
1(r +/) 31(r)
/
= lim
I<0
1 + 4(r +/) 3(r +/)

2
3(1 + 4r 3r
2
)
/
= lim
I<0
(1 + 4r + 4/ 3r
2
32r/ 3/
2
) 3(1 + 4r 3r
2
)
/
= lim
I<0
4/ 32r/ 3/
2
/
= lim
I<0
(4 32r 3/) = 4 32r
1
00
(r) = lim
I<0
1
0
(r +/) 31
0
(r)
/
= lim
I<0
[4 32(r +/)] 3(4 32r)
/
= lim
I<0
32/
/
= lim
I<0
(32) = 32
We see from the graph that our answers are reasonable because the graph of
1
0
is that of a linear function and the graph of 1
00
is that of a constant
function.
76

47. 1
0
(r) = lim
I<0
1(r +/) 31(r)
/
= lim
I<0
2(r +/)
2
3(r +/)
3
3(2r
2
3r
3
)
/
= lim
I<0
/(4r + 2/ 33r
2
33r/ 3/
2
)
/
= lim
I<0
(4r + 2/ 33r
2
33r/ 3/
2
) = 4r 33r
2
1
00
(r) = lim
I<0
1
0
(r +/) 31
0
(r)
/
= lim
I<0
4(r +/) 33(r +/)

2
3(4r 33r
2
)
/
= lim
I<0
/(4 36r 33/)
/
= lim
I<0
(4 36r 33/) = 4 36r
1
000
(r) = lim
I<0
1
00
(r +/) 31
00
(r)
/
= lim
I<0
[4 36(r +/)] 3(4 36r)
/
= lim
I<0
36/
/
= lim
I<0
(36) = 36
1
(4)
(r) = lim
I<0
1
000
(r +/) 31
000
(r)
/
= lim
I<0
36 3(36)
/
= lim
I<0
0
/
= lim
I<0
(0) = 0
The graphs are consistent with the geometric interpretations of the
derivatives because 1
0
has zeros where 1 has a local minimum and a local
maximum, 1
00
has a zero where 1
0
has a local maximum, and 1
000
is a
constant function equal to the slope of 1
00
.
49. (a) Note that we have factored r 3o as the difference of two cubes in the third step.
1
0
(o) = lim
i<a
1(r) 31(o)
r 3o
= lim
i<a
r
1'3
3o
1'3
r 3o
= lim
i<a
r
1'3
3o
1'3
(r
1'3
3o
1'3
)(r
2'3
+r
1'3
o
1'3
+o
2'3
)
= lim
i<a
1
r
2'3
+r
1'3
o
1'3
+o
2'3
=
1
3o
2'3
or
1
3
o
32'3
(b) 1
0
(0) = lim
I<0
1(0 +/) 31(0)
/
= lim
I<0
3
I
/ 30
/
= lim
I<0
1
/
2'3
. This function increases without bound, so the limit does not
exist, and therefore 1
0
(0) does not exist.
(c) lim
i<0
|1
0
(r)| = lim
i<0
1
3r
2'3
= "and 1 is continuous at r = 0 (root function), so 1 has a vertical tangent at r = 0.
51. 1(r) = |r 36| =
r 36 if r 36 D 6
3(r 36) if r 36 < 0
=
r 36 if r D 6
6 3r if r < 6
So the right-hand limit is lim
i<6
+
1(r) 31(6)
r 36
= lim
i<6
+
|r 36| 30
r 36
= lim
i<6
+
r 36
r 36
= lim
i<6
+
1 = 1, and the left-hand limit
is lim
i<6
1(r) 31(6)
r 36
= lim
i<6
|r 36| 30
r 36
= lim
i<6
6 3r
r 36
= lim
i<6
(31) = 31. Since these limits are not equal,

1
0
(6) = lim
i<6
1(r) 31(6)
r 36
does not exist and 1 is not differentiable at 6.
However, a formula for 1
0
is 1
0
(r) =
1 if r 6
31 if r < 6
Another way of writing the formula is 1
0
(r) =
r 36
|r 36|
.

77
53. (a) 1(r) = r|r| =
r
2
if r D 0
3r
2
if r < 0
(b) Since 1(r) = r
2
for r D 0, we have 1
0
(r) = 2r for r 0.
[See Exercise 2.8.17(d).] Similarly, since 1(r) = 3r
2
for
r < 0, we have 1
0
(r) = 32r for r < 0. At r = 0, we have
1
0
(0) = lim
i<0
1(r) 31(0)
r 30
= lim
i<0
r|r|
r
= lim
i<0
|r| = 0.
So 1 is differentiable at 0. Thus, 1 is differentiable for all r.
(c) From part (b), we have 1
0
(r) =
2r if r D 0
32r if r < 0
= 2 |r|.
55. (a) If 1 is even, then
1
0
(3r) = lim
I<0
1(3r +/) 31(3r)
/
= lim
I<0
1[3(r 3/)] 31(3r)
/
= lim
I<0
1(r 3/) 31(r)
/
= 3 lim
I<0
1(r 3/) 31(r)
3/
[let {r = 3/]
= 3 lim
{i<0
1(r +{r) 31(r)
{r
= 31
0
(r)
Therefore, 1
0
is odd.
(b) If 1 is odd, then
1
0
(3r) = lim
I<0
1(3r +/) 31(3r)
/
= lim
I<0
1[3(r 3/)] 31(3r)
/
= lim
I<0
31(r 3/) +1(r)
/
= lim
I<0
1(r 3/) 31(r)
3/
[let {r = 3/]
= lim
{i<0
1(r +{r) 31(r)
{r
= 1
0
(r)
Therefore, 1
0
is even.
57. In the right triangle in the diagram, let {j be the side opposite angle c and {r
the side adjacent angle c. Then the slope of the tangent line /
is i = {j{r = tanc. Note that 0 < c <
r
2
. We know (see Exercise 17)
that the derivative of 1(r) = r
2
is 1
0
(r) = 2r. So the slope of the tangent to
the curve at the point (1. 1) is 2. Thus, c is the angle between 0 and
r
2
whose
tangent is 2; that is, c = tan
31
2 E 63
.
78

2 Review
1. (a) lim
i<a
1(r) = 1: See Denition 2.2.1 and Figures 1 and 2 in Section 2.2.
(b) lim
i<a
+
1(r) = 1: See the paragraph after Denition 2.2.2 and Figure 9(b) in Section 2.2.
(c) lim
i<a
1(r) = 1: See Denition 2.2.2 and Figure 9(a) in Section 2.2.

(d) lim
i<a
1(r) = ": See Denition 2.2.4 and Figure 12 in Section 2.2.
(e) lim
i<"
1(r) = 1: See Denition 2.6.1 and Figure 2 in Section 2.6.
2. In general, the limit of a function fails to exist when the function does not approach a xed number. For each of the following
functions, the limit fails to exist at r = 2.
The left- and right-hand
limits are not equal.
There is an
innite discontinuity.
There are an innite
number of oscillations.
3. (a) (g) See the statements of Limit Laws 1 6 and 11 in Section 2.3.
4. See Theorem 3 in Section 2.3.
5. (a) See Denition 2.2.6 and Figures 1214 in Section 2.2.
(b) See Denition 2.6.3 and Figures 3 and 4 in Section 2.6.
6. (a) j = r
4
: No asymptote (b) j = sinr: No asymptote
(c) j = tanr: Vertical asymptotes r =
r
2
+n, n an integer (d) j = tan
31
r: Horizontal asymptotes j =
r
2
(e) j = c
i
: Horizontal asymptote j = 0
lim
i<3"
c
i
= 0
(f ) j = lnr: Vertical asymptote r = 0
lim
i<0
+
lnr = 3"
(g) j = 1r: Vertical asymptote r = 0,

horizontal asymptote j = 0
(h) j =
I
r: No asymptote
7. (a) A function 1 is continuous at a number o if 1(r) approaches 1(o) as r approaches o; that is, lim
i<a
1(r) = 1(o).
(b) A function 1 is continuous on the interval (3". ") if 1 is continuous at every real number o. The graph of such a
function has no breaks and every vertical line crosses it.
CHAPTER 2 REVIEW

79
8. See Theorem 2.5.10.
9. See Denition 2.7.1.
10. See the paragraph containing Formula 3 in Section 2.7.
11. (a) The average rate of change of j with respect to r over the interval [r
1
. r
2
] is
1(r
2
) 31(r
1
)
r
2
3r
1
.
(b) The instantaneous rate of change of j with respect to r at r = r
1
is lim
i
2
<i
1
1(r
2
) 31(r
1
)
r
2
3r
1
.
12. See Denition 2.7.2. The pages following the denition discuss interpretations of 1
0
(o) as the slope of a tangent line to the
graph of 1 at r = o and as an instantaneous rate of change of 1(r) with respect to r when r = o.
13. See the paragraphs before and after Example 6 in Section 2.8.
14. (a) A function 1 is differentiable at a number o if its derivative 1
0
exists
at r = o; that is, if 1
0
(o) exists.
(c)
(b) See Theorem 2.8.4. This theorem also tells us that if 1 is not
continuous at o, then 1 is not differentiable at o.
15. See the discussion and Figure 7 on page 159.
1. False. Limit Law 2 applies only if the individual limits exist (these dont).
3. True. Limit Law 5 applies.
5. False. Consider lim
i<5
r(r 35)
r 35
or lim
i<5
sin(r 35)
r 35
. The rst limit exists and is equal to 5. By Example 3 in Section 2.2,
we know that the latter limit exists (and it is equal to 1).
7. True. A polynomial is continuous everywhere, so lim
i<l
j(r) exists and is equal to j(/).
9. True. See Figure 8 in Section 2.6.
11. False. Consider 1(r) =
1(r 31) if r 6= 1
2 if r = 1
13. True. Use Theorem 2.5.8 with o = 2, / = 5, and o(r) = 4r
2
311. Note that 1(4) = 3 is not needed.
15. True, by the denition of a limit with - = 1.
17. False. See the note after Theorem 4 in Section 2.8.
19. False.
d
2
j
dr
2
is the second derivative while
dj
dr
2
is the rst derivative squared. For example, if j = r,
then
d
2
j
dr
2
= 0, but
dj
dr
2
= 1.
80

1. (a) (i) lim
i<2
+
1(r) = 3 (ii) lim
i<33
+
1(r) = 0
(iii) lim
i<33
1(r) does not exist since the left and right limits are not equal. (The left limit is 32.)
(iv) lim
i<4
1(r) = 2
(v) lim
i<0
1(r) = " (vi) lim
i<2
1(r) = 3"
(vii) lim
i<"
1(r) = 4 (viii) lim
i<3"
1(r) = 31
(b) The equations of the horizontal asymptotes are j = 31 and j = 4.
(c) The equations of the vertical asymptotes are r = 0 and r = 2.
(d) 1 is discontinuous at r = 33, 0, 2, and 4. The discontinuities are jump, innite, innite, and removable, respectively.
3. Since the exponential function is continuous, lim
i<1
c
i
3
3i
= c
131
= c
0
= 1.
5. lim
i<33
r
2
39
r
2
+ 2r 33
= lim
i<33
(r + 3)(r 33)
(r + 3)(r 31)
= lim
i<33
r 33
r 31
=
33 33
33 31
=
36
34
=
3
2
7. lim
I<0
(/ 31)
3
+ 1
/
= lim
I<0
/
3
33/
2
+ 3/ 31
+ 1
/
= lim
I<0
/
3
33/
2
+ 3/
/
= lim
I<0
/
2
33/ + 3
= 3
Another solution: Factor the numerator as a sum of two cubes and then simplify.
lim
I<0
(/ 31)
3
+ 1
/
= lim
I<0
(/ 31)
3
+ 1
3
/
= lim
I<0
[(/ 31) + 1]
(/ 31)
2
31(/ 31) + 1
2
/
= lim
I<0
(/ 31)
2
3/ + 2
= 1 30 + 2 = 3
9. lim
r<9
I
v
(v 39)
4
= "since (v 39)
4
<0 as v <9 and
I
v
(v 39)
4
0 for v 6= 9.
11. lim
u<1
&
4
31
&
3
+ 5&
2
36&
= lim
u<1
(&
2
+ 1)(&
2
31)
&(&
2
+ 5& 36)
= lim
u<1
(&
2
+ 1)(& + 1)(& 31)
&(& + 6)(& 31)
= lim
u<1
(&
2
+ 1)(& + 1)
&(& + 6)
=
2(2)
1(7)
=
4
7
13. Since r is positive,
I
r
2
= |r| = r. Thus,
lim
i<"
I
r
2
39
2r 36
= lim
i<"
I
r
2
39
I
r
2
(2r 36)r
= lim
i<"
1 39r
2
2 36r
=
I
1 30
2 30
=
1
2
15. Let t = sinr. Then as r <
3
, sinr <0
+
, so t <0
+
. Thus, lim
i<r
ln(sinr) = lim
I<0
+
lnt = 3".
17. lim
i<"
I
r
2
+ 4r + 1 3r
= lim
i<"
I
r
2
+ 4r + 1 3r
1

I
r
2
+ 4r + 1 +r
I
r
2
+ 4r + 1 +r
= lim
i<"
(r
2
+ 4r + 1) 3r
2
I
r
2
+ 4r + 1 +r
= lim
i<"
(4r + 1)r
(
I
r
2
+ 4r + 1 +r)r
divide by r =
I
r
2
for r 0
= lim
i<"
4 + 1r
1 + 4r + 1r
2
+ 1
=
4 + 0
I
1 + 0 + 0 + 1
=
4
2
= 2
CHAPTER 2 REVIEW

81
19. Let t = 1r. Then as r <0
+
, t <", and lim
i<0
+
tan
31
(1r) = lim
I<"
tan
31
t =

2
.
21. From the graph of j =

cos
2
r
r
2
, it appears that j = 0 is the horizontal
asymptote and r = 0 is the vertical asymptote. Now 0 $ (cos r)
2
$ 1 i
0
r
2
$
cos
2
r
r
2
$
1
r
2
i 0 $
cos
2
r
r
2
$
1
r
2
. But lim
i<"
0 = 0 and
lim
i<"
1
r
2
= 0, so by the Squeeze Theorem, lim
i<"
cos
2
r
r
2
= 0.
Thus, j = 0 is the horizontal asymptote. lim
i<0
cos
2
r
r
2
= "because cos
2
r <1 and r
2
<0 as r <0, so r = 0 is the
vertical asymptote.
23. Since 2r 31 $ 1(r) $ r
2
for 0 < r < 3 and lim
i<1
(2r 31) = 1 = lim
i<1
r
2
, we have lim
i<1
25. Given - 0, we need c 0 such that if 0 < |r 32| < c, then |(14 35r) 34| < -. But |(14 35r) 34| < - C
|35r + 10| < - C |35| |r 32| < - C |r 32| < -5. So if we choose c = -5, then 0 < |r 32| < c i
|(14 35r) 34| < -. Thus, lim
i<2
(14 35r) = 4 by the denition of a limit.
27. Given - 0, we need c 0 so that if 0 < |r 32| < c, then
r
2
33r 3(32)
< -. First, note that if |r 32| < 1, then

31 < r 32 < 1, so 0 < r 31 < 2 i |r 31| < 2. Now let c = min{-2. 1}. Then 0 < |r 32| < c i
r
2
33r 3(32)
= |(r 32)(r 31)| = |r 32| |r 31| < (-2)(2) = -.

Thus, lim
i<2
r
2
33r

29. (a) 1(r) =
I
3r if r < 0, 1(r) = 3 3r if 0 $ r < 3, 1(r) = (r 33)
2
if r 3.
(i) lim
i<0
+
1(r) = lim
i<0
+
(3 3r) = 3 (ii) lim
i<0
1(r) = lim
i<0
I
3r = 0
(iii) Because of (i) and (ii), lim
i<0
1(r) does not exist. (iv) lim
i<3
1(r) = lim
i<3
(3 3r) = 0
(v) lim
i<3
+
1(r) = lim
i<3
+
(r 33)
2
= 0 (vi) Because of (iv) and (v), lim
i<3
1(r) = 0.
(b) 1 is discontinuous at 0 since lim
i<0
1(r) does not exist.
1 is discontinuous at 3 since 1(3) does not exist.
(c)
31. sinr is continuous on R by Theorem 7 in Section 2.5. Since c
i
is continuous on R, c
sin i
is continuous on R by Theorem 9 in
Section 2.5. Lastly, r is continuous on R since its a polynomial and the product rc
sin i
is continuous on its domain R by
Theorem 4 in Section 2.5.
33. 1(r) = 2r
3
+r
2
+ 2 is a polynomial, so it is continuous on [32. 31] and 1(32) = 310 < 0 < 1 = 1(31). So by the
Intermediate Value Theorem there is a number c in (32. 31) such that 1(c) = 0, that is, the equation 2r
3
+r
2
+ 2 = 0 has a
root in (32. 31).
82

35. (a) The slope of the tangent line at (2. 1) is
lim
i<2
1(r) 31(2)
r 32
= lim
i<2
9 32r
2
31
r 32
= lim
i<2
8 32r
2
r 32
= lim
i<2
32(r
2
34)
r 32
= lim
i<2
32(r 32)(r + 2)
r 32
= lim
i<2
[32(r + 2)] = 32 4 = 38
(b) An equation of this tangent line is j 31 = 38(r 32) or j = 38r + 17.
37. (a) c = c(t) = 1 + 2t +t
2
4. The average velocity over the time interval [1. 1 +/] is
ave
=
c(1 +/) 3c(1)
(1 +/) 31
=
1 + 2(1 +/) + (1 +/)
2
4 3134
/
=
10/ +/
2
4/
=
10 +/
4
So for the following intervals the average velocities are:
(i) [1. 3]: / = 2,
ave
= (10 + 2)4 = 3 ms (ii) [1. 2]: / = 1,
ave
= (10 + 1)4 = 2.75 ms
(iii) [1. 1.5]: / = 0.5,
ave
= (10 + 0.5)4 = 2.625 ms (iv) [1. 1.1]: / = 0.1,
ave
= (10 + 0.1)4 = 2.525 ms
(b) When t = 1, the instantaneous velocity is lim
I<0
c(1 +/) 3c(1)
/
= lim
I<0
10 +/
4
=
10
4
= 2.5 ms.
39. (a) 1
0
(2) = lim
i<2
1(r) 31(2)
r 32
= lim
i<2
r
3
32r 34
r 32
= lim
i<2
(r 32)
r
2
+ 2r + 2
r 32
= lim
i<2
r
2
+ 2r + 2
= 10
(c)
(b) j 34 = 10(r 32) or j = 10r 316
41. (a) 1
0
(v) is the rate at which the total cost changes with respect to the interest rate. Its units are dollars(percent per year).
(b) The total cost of paying off the loan is increasing by $1200(percent per year) as the interest rate reaches 10%. So if the
interest rate goes up from 10% to 11%, the cost goes up approximately $1200.
(c) As v increases, C increases. So 1
0
(v) will always be positive.
43.
CHAPTER 2 REVIEW

83
45. (a) 1
0
(r) = lim
I<0
1(r +/) 31(r)
/
= lim
I<0
3 35(r +/) 3
I
3 35r
/
3 35(r +/) +
I
3 35r
3 35(r +/) +
I
3 35r
= lim
I<0
[3 35(r +/)] 3(3 35r)
/
3 35(r +/) +
I
3 35r
= lim
I<0
35
3 35(r +/) +
I
3 35r
=
35
2
I
3 35r
(b) Domain of 1: (the radicand must be nonnegative) 3 35r D 0 i
5r $ 3 i r M
3".
3
5
Domain of 1
0
: exclude
3
5
because it makes the denominator zero;
r M
3".
3
5
(c) Our answer to part (a) is reasonable because 1

0
(r) is always negative and 1
is always decreasing.
47. 1 is not differentiable: at r = 34 because 1 is not continuous, at r = 31 because 1 has a corner, at r = 2 because 1 is not
continuous, and at r = 5 because 1 has a vertical tangent.
49. C
0
(1990) is the rate at which the total value of US currency in circulation is changing in billions of dollars per year. To
estimate the value of C
0
(1990), we will average the difference quotients obtained using the times t = 1985 and t = 1995.
Let =
C(1985) 3C(1990)
1985 31990
=
187.3 3271.9
35
=
384.6
35
= 16.92 and
1 =
C(1995) 3C(1990)
1995 31990
=
409.3 3271.9
5
=
137.4
5
= 27.48. Then
C
0
(1990) = lim
I<1990
C(t) 3C(1990)
t 31990
E
+1
2
=
16.92 + 27.48
2
=
44.4
2
= 22.2 billion dollarsyear.
51. |1(r)| $ o(r) C 3o(r) $ 1(r) $ o(r) and lim
i<a
o(r) = 0 = lim
i<a
3o(r).
Thus, by the Squeeze Theorem, lim
i<a
1(r) = 0.
PROBLEMS PLUS
1. Let t =
6
I
r, so r = t
6
. Then t <1 as r <1, so
lim
i<1
3
I
r 31
I
r 31
= lim
I<1
t
2
31
t
3
31
= lim
I<1
(t 31)(t + 1)
(t 31) (t
2
+t + 1)
= lim
I<1
t + 1
t
2
+t + 1
=
1 + 1
1
2
+ 1 + 1
=
2
3
.
Another method: Multiply both the numerator and the denominator by (
I
r + 1)
3
I
r
2
+
3
I
r + 1
.
3. For 3
1
2
< r <
1
2
, we have 2r 31 < 0 and 2r + 1 0, so |2r 31| = 3(2r 31) and |2r + 1| = 2r + 1.
Therefore, lim
i<0
|2r 31| 3|2r + 1|
r
= lim
i<0
3(2r 31) 3(2r + 1)
r
= lim
i<0
34r
r
= lim
i<0
(34) = 34.
5. Since [[r]] $ r < [[r]] + 1, we have
[[r]]
[[r]]
$
r
[[r]]
<
[[r]] + 1
[[r]]
i 1 $
r
[[r]]
< 1 +
1
[[r]]
for r D 1. As r <", [[r]] <",
so
1
[[r]]
<0 and 1 +
1
[[r]]
<1. Thus, lim
i<"
r
[[r]]
= 1 by the Squeeze Theorem.
7. 1 is continuous on (3". o) and (o. "). To make 1 continuous on R, we must have continuity at o. Thus,
lim
i<a
+
1(r) = lim
i<a
1(r) i lim
i<a
+
r
2
= lim
i<a
(r + 1) i o
2
= o + 1 i o
2
3o 31 = 0 i
[by the quadratic formula] o =

1
I
5
2 E 1.618 or 30.618.
9. lim
i<a
1(r) = lim
i<a
1
2
[1(r) +o(r)] +
1
2
[1(r) 3o(r)]
=
1
2
lim
i<a
[1(r) +o(r)] +
1
2
lim
i<a
[1(r) 3o(r)]
=
1
2
2 +
1
2
1 =
3
2
,
and lim
i<a
o(r) = lim
i<a
[1(r) +o(r)] 31(r)
= lim
i<a
[1(r) +o(r)] 3 lim
i<a
1(r) = 2 3
3
2
=
1
2
.
So lim
i<a
[1(r)o(r)] =
lim
i<a
1(r)

lim
i<a
o(r)
=
3
2

1
2
=
3
4
.
Another solution: Since lim
i<a
[1(r) +o(r)] and lim
i<a
[1(r) 3o(r)] exist, we must have
lim
i<a
[1(r) +o(r)]
2
=
lim
i<a
[1(r) +o(r)]
2
and lim
i<a
[1(r) 3o(r)]
2
=
lim
i<a
[1(r) 3o(r)]
2
, so
lim
i<a
[1(r) o(r)] = lim
i<a
1
4
[1(r) +o(r)]
2
3[1 (r) 3o(r)]
2
[because all of the 1

2
and o
2
cancel]
=
1
4
lim
i<a
[1(r) +o(r)]
2
3 lim
i<a
[1(r) 3o(r)]
2
=
1
4
2
2
31
2
=
3
4
.
11. (a) Consider G(r) = T(r + 180
) 3T(r). Fix any number o. If G(o) = 0, we are done: Temperature at o =Temperature

at o + 180
. If G(o) 0, then G(o + 180
) = T(o + 360
) 3T(o + 180
) = T(o) 3T(o + 180
) = 3G(o) < 0.
Also, G is continuous since temperature varies continuously. So, by the Intermediate Value Theorem, Ghas a zero on the
interval [o. o + 180
]. If G(o) < 0, then a similar argument applies.

85
86

CHAPTER 2 PROBLEMS PLUS
(b) Yes. The same argument applies.
(c) The same argument applies for quantities that vary continuously, such as barometric pressure. But one could argue that
altitude above sea level is sometimes discontinuous, so the result might not always hold for that quantity.
13. (a) Put r = 0 and j = 0 in the equation: 1(0 + 0) = 1(0) +1(0) + 0
2
0 + 0 0
2
i 1(0) = 21(0).
Subtracting 1(0) from each side of this equation gives 1(0) = 0.
(b) 1
0
(0) = lim
I<0
1(0 +/) 31(0)
/
= lim
I<0
1(0) +1(/) + 0
2
/ + 0/
2
31(0)
/
= lim
I<0
1(/)
/
= lim
i<0
1(r)
r
= 1
(c) 1
0
(r) = lim
I<0
1(r +/) 31(r)
/
= lim
I<0
1(r) +1(/) +r
2
/ +r/
2
31(r)
/
= lim
I<0
1(/) +r
2
/ +r/
2
/
= lim
I<0
1(/)
/
+r
2
+r/
= 1 +r
2
3 DIFFERENTIATION RULES
3.1 Derivatives of Polynomials and Exponential Functions
1. (a) c is the number such that lim
I<0
c
I
31
/
= 1.
(b)
r
2.7
i
31
r
30.001 0.9928
30.0001 0.9932
0.001 0.9937
0.0001 0.9933
r
2.8
i
31
r
30.001 1.0291
30.0001 1.0296
0.001 1.0301
0.0001 1.0297
From the tables (to two decimal places),
lim
I<0
2.7
I
31
/
= 0.99 and lim
I<0
2.8
I
31
/
= 1.03.
Since 0.99 < 1 < 1.03, 2.7 < c < 2.8.
3. 1(r) = 186.5 is a constant function, so its derivative is 0, that is, 1
0
(r) = 0.
5. 1(t) = 2 3
2
3
t i 1
0
(t) = 0 3
2
3
= 3
2
3
7. 1(r) = r
3
34r + 6 i 1
0
(r) = 3r
2
34(1) + 0 = 3r
2
34
9. 1(t) =
1
4
(t
4
+ 8) i 1
0
(t) =
1
4
(t
4
+ 8)
0
=
1
4
(4t
431
+ 0) = t
3
11. j = r
32'5
i j
0
= 3
2
5
r
(32'5)31
= 3
2
5
r
37'5
= 3
2
5r
7'5
13. \ (v) =
4
3
v
3
i \
0
(v) =
4
3
3v
2
= 4v
2
15. (c) = 3
12
c
5
= 312c
35
i
0
(c) = 312(35c
36
) = 60c
36
or 60c
6
17. G(r) =
I
r 32c
i
= r
1'2
32c
i
i G
0
(r) =
1
2
r
31'2
32c
i
=
1
2
I
r
32c
i
19. 1(r) = (
1
2
r)
5
=

1
2
5
r
5
=
1
32
r
5
i 1
0
(r) =
1
32
(5r
4
) =
5
32
r
4
21. j = or
2
+/r +c i j
0
= 2or +/
23. j =
r
2
+ 4r + 3
I
r
= r
3'2
+ 4r
1'2
+ 3r
31'2
i
j
0
=
3
2
r
1'2
+ 4
1
2
r
31'2
+ 3
3
1
2
r
33'2
=
3
2
I
r +
2
I
r
3
3
2r
I
r
note that r
3'2
= r
2'2
r
1'2
= r
I
r
The last expression can be written as

3r
2
2r
I
r
+
4r
2r
I
r
3
3
2r
I
r
=
3r
2
+ 4r 33
2r
I
r
.
25. j = 4
2
i j
0
= 0 since 4
2
is a constant.
87
88

CHAPTER 3 DIFFERENTIATION RULES
27. We rst expand using the Binomial Theorem (see Reference Page 1).
1(r) = (r +r
31
)
3
= r
3
+ 3r
2
r
31
+ 3r(r
31
)
2
+ (r
31
)
3
= r
3
+ 3r + 3r
31
+r
33
i
1
0
(r) = 3r
2
+ 3 + 3(31r
32
) + (33r
34
) = 3r
2
+ 3 33r
32
33r
34
29. & =
5
I
t + 4
I
t
5
= t
1'5
+ 4t
5'2
i &
0
=
1
5
t
34'5
+ 4
5
2
t
3'2
=
1
5
t
34'5
+ 10t
3'2
or 1
5
5
I
t
4
+ 10
I
t
3
31. : =

j
10
+1c
u
= j
310
+1c
u
i :
0
= 310j
311
+1c
u
= 3
10
j
11
+1c
u
33. j =
4
I
r = r
1'4
i j
0
=
1
4
r
33'4
=
1
4
4
I
r
3
. At (1. 1), j
0
=
1
4
and an equation of the tangent line is
j 31 =
1
4
(r 31) or j =
1
4
r +
3
4
.
35. j = r
4
+ 2c
i
i j
0
= 4r
3
+ 2c
i
. At (0. 2), j
0
= 2 and an equation of the tangent line is j 32 = 2(r 30)
or j = 2r + 2. The slope of the normal line is 3
1
2
(the negative reciprocal of 2) and an equation of the normal line is
j 32 = 3
1
2
(r 30) or j = 3
1
2
r + 2.
37. j = 3r
2
3r
3
i j
0
= 6r 33r
2
.
At (1. 2), j
0
= 6 33 = 3, so an equation of the tangent line is
j 32 = 3(r 31) or j = 3r 31.
39. 1(r) = c
i
35r i 1
0
(r) = c
i
35.
Notice that 1
0
0
is positive
when 1 is increasing, and 1
0
is negative when 1 is decreasing.
41. 1(r) = 3r
15
35r
3
+ 3 i 1
0
(r) = 45r
14
315r
2
.
Notice that 1
0
0
is positive
when 1 is increasing, and 1
0
is negative when 1 is decreasing.
SECTION 3.1 DERIVATIVES OF POLYNOMIALS AND EXPONENTIAL FUNCTIONS

89
43. (a) (b) From the graph in part (a), it appears that 1
0
is zero at r
1
E 31.25, r
2
E 0.5,
and r
3
E 3. The slopes are negative (so 1
0
is negative) on (3". r
1
) and
(r
2
. r
3
). The slopes are positive (so 1
0
is positive) on (r
1
. r
2
) and (r
3
. ").
(c) 1(r) = r
4
33r
3
36r
2
+ 7r + 30 i
1
0
(r) = 4r
3
39r
2
312r + 7
45. 1(r) = r
4
33r
3
+ 16r i 1
0
(r) = 4r
3
39r
2
+ 16 i 1
00
(r) = 12r
2
318r
47. 1(r) = 2r 35r
3'4
i 1
0
(r) = 2 3
15
4
r
31'4
i 1
00
(r) =
15
16
r
35'4
Note that 1
0
is negative when 1 is decreasing and positive when 1 is
increasing. 1
00
is always positive since 1
0
is always increasing.
49. (a) c = t
3
33t i (t) = c
0
(t) = 3t
2
33 i o(t) =
0
(t) = 6t
(b) o(2) = 6(2) = 12 ms
2
(c) (t) = 3t
2
33 = 0 when t
2
= 1, that is, t = 1 and o(1) = 6 ms
2
.
51. The curve j = 2r
3
+ 3r
2
312r + 1 has a horizontal tangent when j
0
= 6r
2
+ 6r 312 = 0 C 6(r
2
+r 32) = 0 C
6(r + 2)(r 31) = 0 C r = 32 or r = 1. The points on the curve are (32. 21) and (1. 36).
53. j = 6r
3
+ 5r 33 i i = j
0
= 18r
2
+ 5, but r
2
D 0 for all r, so i D 5 for all r.
55. The slope of the line 12r 3j = 1 (or j = 12r 31) is 12, so the slope of both lines tangent to the curve is 12.
j = 1 +r
3
i j
0
= 3r
2
. Thus, 3r
2
= 12 i r
2
= 4 i r = 2, which are the r-coordinates at which the tangent
lines have slope 12. The points on the curve are (2. 9) and (32. 37), so the tangent line equations are j 39 = 12(r 32)
or j = 12r 315 and j + 7 = 12(r + 2) or j = 12r + 17.
90

57. The slope of j = r
2
35r + 4 is given by i = j
0
= 2r 35. The slope of r 33j = 5 C j =
1
3
r 3
5
3
is
1
3
,
so the desired normal line must have slope
1
3
, and hence, the tangent line to the parabola must have slope 33. This occurs if
2r 35 = 33 i 2r = 2 i r = 1. When r = 1, j = 1
2
35(1) + 4 = 0, and an equation of the normal line is
j 30 =
1
3
(r 31) or j =
1
3
r 3
1
3
.
59. Let
o. o
2
be a point on the parabola at which the tangent line passes through the
point (0. 34). The tangent line has slope 2o and equation j 3(34) = 2o(r 30) C
j = 2or 34. Since
o. o
2
also lies on the line, o

2
= 2o(o) 34, or o
2
= 4. So
o = 2 and the points are (2. 4) and (32. 4).
61. 1
0
(r) = lim
I<0
1(r +/) 31(r)
/
= lim
I<0
1
r +/
3
1
r
/
= lim
I<0
r 3(r +/)
/r(r +/)
= lim
I<0
3/
/r(r +/)
= lim
I<0
31
r(r +/)
= 3
1
r
2
63. Let 1(r) = or
2
+/r +c. Then 1
0
(r) = 2or +/ and 1
00
(r) = 2o. 1
00
(2) = 2 i 2o = 2 i o = 1.
1
0
(2) = 3 i 2(1)(2) +/ = 3 i 4 +/ = 3 i / = 31.
1(2) = 5 i 1(2)
2
+ (31)(2) +c = 5 i 2 +c = 5 i c = 3. So 1(r) = r
2
3r + 3.
65. j = 1(r) = or
3
+/r
2
+cr +d i 1
0
(r) = 3or
2
+ 2/r +c. The point (32. 6) is on 1, so 1(32) = 6 i
38o + 4/ 32c +d = 6 (1). The point (2. 0) is on 1, so 1(2) = 0 i 8o + 4/ + 2c +d = 0 (2). Since there are
horizontal tangents at (32. 6) and (2. 0), 1
0
(2) = 0. 1
0
(32) = 0 i 12o 34/ +c = 0 (3) and 1
0
(2) = 0 i
12o + 4/ +c = 0 (4). Subtracting equation (3) from (4) gives 8/ = 0 i / = 0. Adding (1) and (2) gives 8/ + 2d = 6,
so d = 3 since / = 0. From (3) we have c = 312o, so (2) becomes 8o + 4(0) + 2(312o) + 3 = 0 i 3 = 16o i
o =
3
16
. Now c = 312o = 312
3
16
= 3
9
4
and the desired cubic function is j =
3
16
r
3
3
9
4
r + 3.
67. 1(r) = 2 3r if r $ 1 and 1(r) = r
2
32r + 2 if r 1. Now we compute the right- and left-hand derivatives dened in
Exercise 2.8.54:
1
0
3
(1) = lim
I<0
1(1 +/) 31(1)

/
= lim
I<0
2 3(1 +/) 31
/
= lim
I<0
3/
/
= lim
I<0
31 = 31 and
1
0
+
(1) = lim
I<0
+
1(1 +/) 31(1)
/
= lim
I<0
+
(1 +/)
2
32(1 +/) + 2 31
/
= lim
I<0
+
/
2
/
= lim
I<0
+
/ = 0.
Thus, 1
0
(1) does not exist since 1
0
3
(1) 6= 1
0
+
(1), so 1
is not differentiable at 1. But 1
0
(r) = 31 for r < 1
and 1
0
(r) = 2r 32 if r 1.
SECTION 3.1 DERIVATIVES OF POLYNOMIALS AND EXPONENTIAL FUNCTIONS

91
69. (a) Note that r
2
39 < 0 for r
2
< 9 C |r| < 3 C 33 < r < 3. So
1(r) =
r
2
39 if r $ 33
3r
2
+ 9 if 33 < r < 3
r
2
39 if r D 3
i 1
0
(r) =
2r if r < 33
32r if 33 < r < 3
2r if r 3
=
2r if |r| 3
32r if |r| < 3
To show that 1
0
(3) does not exist we investigate lim
I<0
1(3 +/) 31(3)
/
by computing the left- and right-hand derivatives
dened in Exercise 2.8.54.
1
0
3
(3) = lim
I<0
1(3 +/) 31(3)

/
= lim
I<0
[3(3 +/)
2
+ 9] 30
/
= lim
I<0
(36 3/) = 36 and

1
0
+
(3) = lim
I<0
+
1(3 +/) 31(3)
/
= lim
I<0
+
(3 +/)
2
39
30
/
= lim
I<0
+
6/ +/
2
/
= lim
I<0
+
(6 +/) = 6.
Since the left and right limits are different,
lim
I<0
1(3 +/) 31(3)
/
does not exist, that is, 1
0
(3)
does not exist. Similarly, 1
0
(33) does not exist.
Therefore, 1 is not differentiable at 3 or at 33.
(b)
71. Substituting r = 1 and j = 1 into j = or
2
+/r gives us o +/ = 1 (1). The slope of the tangent line j = 3r 32 is 3 and the
slope of the tangent to the parabola at (r. j) is j
0
= 2or +/. At r = 1, j
0
= 3 i 3 = 2o +/ (2). Subtracting (1) from
(2) gives us 2 = o and it follows that / = 31. The parabola has equation j = 2r
2
3r.
73. j = 1(r) = or
2
i 1
0
(r) = 2or. So the slope of the tangent to the parabola at r = 2 is i = 2o(2) = 4o. The slope
of the given line, 2r +j = / C j = 32r +/, is seen to be 32, so we must have 4o = 32 C o = 3
1
2
. So when
r = 2, the point in question has j-coordinate 3
1
2
2
2
= 32. Now we simply require that the given line, whose equation is
2r +j = /, pass through the point (2. 32): 2(2) + (32) = / C / = 2. So we must have o = 3
1
2
and / = 2.
75. 1 is clearly differentiable for r < 2 and for r 2. For r < 2, 1
0
(r) = 2r, so 1
0
3
(2) = 4. For r 2, 1
0
(r) = i, so
1
0
+
(2) = i. For 1 to be differentiable at r = 2, we need 4 = 1
0
3
(2) = 1
0
+
(2) = i. So 1(r) = 4r +/. We must also have
continuity at r = 2, so 4 = 1(2) = lim
i<2
+
1(r) = lim
i<2
+
(4r +/) = 8 +/. Hence, / = 34.
77. Solution 1: Let 1(r) = r
1000
. Then, by the denition of a derivative, 1
0
(1) = lim
i<1
1(r) 31(1)
r 31
= lim
i<1
r
1000
31
r 31
.
But this is just the limit we want to nd, and we know (from the Power Rule) that 1
0
(r) = 1000r
999
, so
1
0
(1) = 1000(1)
999
= 1000. So lim
i<1
r
1000
31
r 31
= 1000.
Solution 2: Note that (r
1000
31) = (r 31)(r
999
+r
998
+r
997
+ +r
2
+r + 1). So
lim
i<1
r
1000
31
r 31
= lim
i<1
(r 31)(r
999
+r
998
+r
997
+ +r
2
+r + 1)
r 31
= lim
i<1
(r
999
+r
998
+r
997
+ +r
2
+r + 1)
= 1 + 1 + 1 + + 1 + 1 + 1
. .. .
= 1000, as above.
1000 ones
92

79. j = r
2
i j
0
= 2r, so the slope of a tangent line at the point (o. o
2
) is j
0
= 2o and the slope of a normal line is 31(2o),
for o 6= 0. The slope of the normal line through the points (o. o
2
) and (0. c) is
o
2
3c
o 30
, so
o
2
3c
o
= 3
1
2o
i
o
2
3c = 3
1
2
i o
2
= c 3
1
2
. The last equation has two solutions if c
1
2
, one solution if c =
1
2
, and no solution if
c <
1
2
. Since the j-axis is normal to j = r
2
regardless of the value of c (this is the case for o = 0), we have three normal lines
if c
1
2
and one normal line if c $
1
2
.
3.2 The Product and Quotient Rules
1. Product Rule: j = (r
2
+ 1)(r
3
+ 1) i
j
0
= (r
2
+ 1)(3r
2
) + (r
3
+ 1)(2r) = 3r
4
+ 3r
2
+ 2r
4
+ 2r = 5r
4
+ 3r
2
+ 2r.
Multiplying rst: j = (r
2
+ 1)(r
3
+ 1) = r
5
+r
3
+r
2
+ 1 i j
0
= 5r
4
+ 3r
2
+ 2r (equivalent).
3. By the Product Rule, 1(r) = (r
3
+ 2r)c
i
i
1
0
(r) = (r
3
+ 2r)(c
i
)
0
+c
i
(r
3
+ 2r)
0
= (r
3
+ 2r)c
i
+c
i
(3r
2
+ 2)
= c
i
[(r
3
+ 2r) + (3r
2
+ 2)] = c
i
(r
3
+ 3r
2
+ 2r + 2)
5. By the Quotient Rule, j =
c
i
r
2
i j
0
=
r
2
d
dr
(c
i
) 3c
i
d
dr
(r
2
)
(r
2
)
2
=
r
2
(c
i
) 3c
i
(2r)
r
4
=
rc
i
(r 32)
r
4
=
c
i
(r 32)
r
3
.
The notations
PR
i and
QR
i indicate the use of the Product and Quotient Rules, respectively.
7. o(r) =
3r 31
2r + 1
QR
i o
0
(r) =
(2r + 1)(3) 3(3r 31)(2)
(2r + 1)
2
=
6r + 3 36r + 2
(2r + 1)
2
=
5
(2r + 1)
2
9. \ (r) = (2r
3
+ 3)(r
4
32r)
PR
i
\
0
(r) = (2r
3
+ 3)(4r
3
32) + (r
4
32r)(6r
2
) = (8r
6
+ 8r
3
36) + (6r
6
312r
3
) = 14r
6
34r
3
36
11. 1(j) =
1
j
2
3
3
j
4
(j + 5j
3
) =

j
32
33j
34
j + 5j
3
PR
i
1
0
(j) = (j
32
33j
34
)(1 + 15j
2
) + (j + 5j
3
)(32j
33
+ 12j
35
)
= (j
32
+ 15 33j
34
345j
32
) + (32j
32
+ 12j
34
310 + 60j
32
)
= 5 + 14j
32
+ 9j
34
or 5 + 14j
2
+ 9j
4
13. j =
r
3
1 3r
2
QR
i j
0
=
(1 3r
2
) (3r
2
) 3r
3
(32r)
(1 3r
2
)
2
=
r
2
(3 33r
2
+ 2r
2
)
(1 3r
2
)
2
=
r
2
(3 3r
2
)
(1 3r
2
)
2
SECTION 3.2 THE PRODUCT AND QUOTIENT RULES

93
15. j =
t
2
+ 2
t
4
33t
2
+ 1
QR
i
j
0
=
(t
4
33t
2
+ 1)(2t) 3(t
2
+ 2)(4t
3
36t)
(t
4
33t
2
+ 1)
2
=
2t[(t
4
33t
2
+ 1) 3(t
2
+ 2)(2t
2
33)]
(t
4
33t
2
+ 1)
2
=
2t(t
4
33t
2
+ 1 32t
4
34t
2
+ 3t
2
+ 6)
(t
4
33t
2
+ 1)
2
=
2t(3t
4
34t
2
+ 7)
(t
4
33t
2
+ 1)
2
17. j = (v
2
32v)c
r
PR
i j
0
= (v
2
32v)(c
r
) +c
r
(2v 32) = c
r
(v
2
32v + 2v 32) = c
r
(v
2
32)
19. j =

3
32
I
=
2
32
I
=
2
32
1'2
i j
0
= 2 32
1
2
31'2
= 2 3
31'2
.
We can change the form of the answer as follows: 2 3
31'2
= 2 3
1
I
=
2
I
31
I
=
2
3'2
31
I
21. 1(t) =
2t
2 +
I
t
QR
i 1
0
(t) =
(2 +t
1'2
)(2) 32t
1
2
t
31'2
(2 +
I
t )
2
=
4 + 2t
1'2
3t
1'2
(2 +
I
t )
2
=
4 +t
1'2
(2 +
I
t )
2
or
4 +
I
t
(2 +
I
t )
2
23. 1(r) =

1 +Cc
i
QR
i 1
0
(r) =
(1 +Cc
i
) 0 3(Cc
i
)
(1 +Cc
i
)
2
= 3
Cc
i
(1 +Cc
i
)
2
25. 1(r) =
r
r +cr
i 1
0
(r) =
(r +cr)(1) 3r(1 3cr
2
)
r +
c
r
2
=
r +cr 3r +cr
r
2
+c
r
2
=
2cr
(r
2
+c)
2
r
2
r
2
r
2
=
2cr
(r
2
+c)
2
27. 1(r) = r
4
c
i
i 1
0
(r) = r
4
c
i
+c
i
4r
3
=

r
4
+ 4r
3
c
i

or r
3
c
i
(r + 4)
i
1
00
(r) = (r
4
+ 4r
3
)c
i
+c
i
(4r
3
+ 12r
2
) = (r
4
+ 4r
3
+ 4r
3
+ 12r
2
)c
i
= (r
4
+ 8r
3
+ 12r
2
)c
i

or r
2
c
i
(r + 2)(r + 6)
29. 1(r) =
r
2
1 + 2r
i 1
0
(r) =
(1 + 2r)(2r) 3r
2
(2)
(1 + 2r)
2
=
2r + 4r
2
32r
2
(1 + 2r)
2
=
2r
2
+ 2r
(1 + 2r)
2
i
1
00
(r) =
(1 + 2r)
2
(4r + 2) 3(2r
2
+ 2r)(1 + 4r + 4r
2
)
0
[(1 + 2r)
2
]
2
=
2(1 + 2r)
2
(2r + 1) 32r(r + 1)(4 + 8r)
(1 + 2r)
4
=
2(1 + 2r)[(1 + 2r)
2
34r(r + 1)]
(1 + 2r)
4
=
2(1 + 4r + 4r
2
34r
2
34r)
(1 + 2r)
3
=
2
(1 + 2r)
3
31. j =
2r
r + 1
i j
0
=
(r + 1)(2) 3(2r)(1)
(r + 1)
2
=
2
(r + 1)
2
.
At (1. 1), j
0
=
1
2
, and an equation of the tangent line is j 31 =
1
2
(r 31), or j =
1
2
r +
1
2
.
33. j = 2rc
i
i j
0
= 2(r c
i
+c
i
1) = 2c
i
(r + 1).
At (0. 0), j
0
= 2c
0
(0 + 1) = 2 1 1 = 2, and an equation of the tangent line is j 30 = 2(r 30), or j = 2r. The slope of
the normal line is 3
1
2
, so an equation of the normal line is j 30 = 3
1
2
(r 30), or j = 3
1
2
r.
94

35. (a) j = 1(r) =
1
1 +r
2
i
1
0
(r) =
(1 +r
2
)(0) 31(2r)
(1 +r
2
)
2
=
32r
(1 +r
2
)
2
. So the slope of the
tangent line at the point
31.
1
2
is 1
0
(31) =
2
2
2
=
1
2
and its
equation is j 3
1
2
=
1
2
(r + 1) or j =
1
2
r + 1.
(b)
37. (a) 1(r) =
c
i
r
3
i 1
0
(r) =
r
3
(c
i
) 3c
i
(3r
2
)
(r
3
)
2
=
r
2
c
i
(r 33)
r
6
=
c
i
(r 33)
r
4
(b) 1
0
= 0 when 1 has a horizontal tangent line, 1
0
is negative when
1 is decreasing, and 1
0
is positive when 1 is increasing.
39. (a) 1(r) = (r 31)c
i
i 1
0
(r) = (r 31)c
i
+c
i
(1) = c
i
(r 31 + 1) = rc
i
.
1
00
(r) = r(c
i
) +c
i
(1) = c
i
(r + 1)
(b) 1
0
= 0 when 1 has a horizontal tangent and 1
00
= 0 when 1
0
has a
horizontal tangent. 1
0
is negative when 1 is decreasing and positive when 1
is increasing. 1
00
is negative when 1
0
is decreasing and positive when 1
0
is
increasing. 1
00
is negative when 1 is concave down and positive when 1 is
concave up.
41. 1(r) =
r
2
1 +r
i 1
0
(r) =
(1 +r)(2r) 3r
2
(1)
(1 +r)
2
=
2r + 2r
2
3r
2
(1 +r)
2
=
r
2
+ 2r
r
2
+ 2r + 1
i
1
00
(r) =
(r
2
+ 2r + 1)(2r + 2) 3(r
2
+ 2r)(2r + 2)
(r
2
+ 2r + 1)
2
=
(2r + 2)(r
2
+ 2r + 1 3r
2
32r)
[(r + 1)
2
]
2
=
2(r + 1)(1)
(r + 1)
4
=
2
(r + 1)
3
,
so 1
00
(1) =
2
(1 + 1)
3
=
2
8
=
1
4
.
43. We are given that 1(5) = 1, 1
0
(5) = 6, o(5) = 33, and o
0
(5) = 2.
(a) (1o)
0
(5) = 1(5)o
0
(5) +o(5)1
0
(5) = (1)(2) + (33)(6) = 2 318 = 316
(b)

1
o
0
(5) =
o(5)1
0
(5) 31(5)o
0
(5)
[o(5)]
2
=
(33)(6) 3(1)(2)
(33)
2
= 3
20
9
(c)

o
1
0
(5) =
1(5)o
0
(5) 3o(5)1
0
(5)
[1(5)]
2
=
(1)(2) 3(33)(6)
(1)
2
= 20
45. 1(r) = c
i
o(r) i 1
0
(r) = c
i
o
0
(r) +o(r)c
i
= c
i
[o
0
(r) +o(r)]. 1
0
(0) = c
0
[o
0
(0) +o(0)] = 1(5 + 2) = 7
SECTION 3.2 THE PRODUCT AND QUOTIENT RULES

95
47. (a) From the graphs of 1 and o, we obtain the following values: 1(1) = 2 since the point (1. 2) is on the graph of 1;
o(1) = 1 since the point (1. 1) is on the graph of o; 1
0
(1) = 2 since the slope of the line segment between (0. 0) and (2. 4)
is
4 30
2 30
= 2; o
0
(1) = 31 since the slope of the line segment between (32. 4) and (2. 0) is
0 34
2 3(32)
= 31.
Now &(r) = 1(r)o(r), so &
0
(1) = 1(1)o
0
(1) +o(1) 1
0
(1) = 2 (31) + 1 2 = 0.
(b) (r) = 1(r)o(r), so
0
(5) =
o(5)1
0
(5) 31(5)o
0
(5)
[o(5)]
2
=
2
3
1
3
33
2
3
2
2
=
3
8
3
4
= 3
2
3
49. (a) j = ro(r) i j
0
= ro
0
(r) +o(r) 1 = ro
0
(r) +o(r)
(b) j =
r
o(r)
i j
0
=
o(r) 1 3ro
0
(r)
[o(r)]
2
=
o(r) 3ro
0
(r)
[o(r)]
2
(c) j =
o(r)
r
i j
0
=
ro
0
(r) 3o(r) 1
(r)
2
=
ro
0
(r) 3o(r)
r
2
51. If j = 1(r) =
r
r + 1
, then 1
0
(r) =
(r + 1)(1) 3r(1)
(r + 1)
2
=
1
(r + 1)
2
. When r = o, the equation of the tangent line is
j 3
o
o + 1
=
1
(o + 1)
2
(r 3o). This line passes through (1. 2) when 2 3
o
o + 1
=
1
(o + 1)
2
(1 3o) C
2(o + 1)
2
3o(o + 1) = 1 3o C 2o
2
+ 4o + 2 3o
2
3o 31 +o = 0 C o
2
+ 4o + 1 = 0.
The quadratic formula gives the roots of this equation as o =
34
4
2
34(1)(1)
2(1)
=
34
I
12
2
= 32
I
3,
so there are two such tangent lines. Since
1
32
I
3
=
32
I
3
32
I
3 + 1
=
32
I
3
31
I
3

31 ~
I
3
31 ~
I
3
=
2 2
I
3 ~
I
3 33
1 33
=
31
I
3
32
=
1 ~
I
3
2
,
the lines touch the curve at
32 +
I
3.
1 3
I
3
2
E (30.27. 30.37)
and 1
32 3
I
3.
1 +
I
3
2
E (33.73. 1.37).
53. If 1(t) denotes the population at time t and (t) the average annual income, then T(t) = 1(t)(t) is the total personal
income. The rate at which T(t) is rising is given by T
0
(t) = 1(t)
0
(t) +(t)1
0
(t) i
T
0
(1999) = 1(1999)
0
(1999) +(1999)1
0
(1999) = (961,400)($1400yr) + ($30,593)(9200yr)
= $1,345,960,000yr + $281,455,600yr = $1,627,415,600yr
So the total personal income was rising by about $1.627 billion per year in 1999.
The term 1(t)
0
(t) E $1.346 billion represents the portion of the rate of change of total income due to the existing
populations increasing income. The term (t)1
0
(t) E $281 million represents the portion of the rate of change of total
income due to increasing population.
96

We will sometimes use the form 1
0
o +1o
0
rather than the form 1o
0
+o1
0
for the Product Rule.
55. (a) (1o/)
0
= [(1o)/]
0
= (1o)
0
/ + (1o)/
0
= (1
0
o +1o
0
)/ + (1o)/
0
= 1
0
o/ +1o
0
/ +1o/
0
(b) Putting 1 = o = / in part (a), we have
d
dr
[1(r)]
3
= (111)
0
= 1
0
11 +11
0
1 +111
0
= 3111
0
= 3[1(r)]
2
1
0
(r).
(c)
d
dr
(c
3i
) =
d
dr
(c
i
)
3
= 3(c
i
)
2
c
i
= 3c
2i
c
i
= 3c
3i
57. For 1(r) = r
2
c
i
, 1
0
(r) = r
2
c
i
+c
i
(2r) = c
i
(r
2
+ 2r). Similarly, we have
1
00
(r) = c
i
(r
2
+ 4r + 2)
1
000
(r) = c
i
(r
2
+ 6r + 6)
1
(4)
(r) = c
i
(r
2
+ 8r + 12)
1
(5)
(r) = c
i
(r
2
+ 10r + 20)
It appears that the coefcient of r in the quadratic term increases by 2 with each differentiation. The pattern for the
constant terms seems to be 0 = 1 0, 2 = 2 1, 6 = 3 2, 12 = 4 3, 20 = 5 4. So a reasonable guess is that
1
(n)
(r) = c
i
[r
2
+ 2nr +n(n 31)].
Proof: Let o
n
be the statement that 1
(n)
(r) = c
i
[r
2
+ 2nr +n(n 31)].
1. o
1
is true because 1
0
(r) = c
i
(r
2
+ 2r).
2. Assume that o
I
is true; that is, 1
(I)
(r) = c
i
[r
2
+ 2Ir +I(I 31)]. Then
1
(I+1)
(r) =
d
dr
1
(I)
(r)
= c
i
(2r + 2I) + [r
2
+ 2Ir +I(I 31)]c
i
= c
i
[r
2
+ (2I + 2)r + (I
2
+I)] = c
i
[r
2
+ 2(I + 1)r + (I + 1)I]
This shows that o
I+1
is true.
3. Therefore, by mathematical induction, o
n
is true for all n; that is, 1
(n)
(r) = c
i
[r
2
+ 2nr +n(n 31)] for every
positive integer n.
3.3 Derivatives of Trigonometric Functions
1. 1(r) = 3r
2
32 cos r i 1
0
(r) = 6r 32(3sinr) = 6r + 2 sinr
3. 1(r) = sinr +
1
2
cot r i 1
0
(r) = cos r 3
1
2
csc
2
r
5. o(t) = t
3
cos t i o
0
(t) = t
3
(3sint) + (cos t) 3t
2
= 3t
2
cos t 3t
3
sint or t
2
(3 cos t 3t sint)
7. /(0) = csc 0 +c
0
cot 0 i /
0
(0) = 3csc 0 cot 0 +c
0
(3csc
2
0) + (cot 0)c
0
= 3csc 0 cot 0 +c
0
(cot 0 3csc
2
0)
9. j =
r
2 3tanr
i j
0
=
(2 3tanr)(1) 3r(3sec
2
r)
(2 3tanr)
2
=
2 3tanr +rsec
2
r
(2 3tanr)
2
11. 1(0) =
sec 0
1 + sec 0
i
1
0
(0) =
(1 + sec 0)(sec 0 tan0) 3(sec 0)(sec 0 tan0)
(1 + sec 0)
2
=
(sec 0 tan0) [(1 + sec 0) 3sec 0]
(1 + sec 0)
2
=
sec 0 tan0
(1 + sec 0)
2
SECTION 3.3 DERIVATIVES OF TRIGONOMETRIC FUNCTIONS

97
13. j =
sinr
r
2
i j
0
=
r
2
cos r 3(sinr)(2r)
(r
2
)
2
=
r(rcos r 32 sinr)
r
4
=
rcos r 32 sinr
r
3
15. Using Exercise 3.2.55(a), 1(r) = rc
i
csc r i
1
0
(r) = (r)
0
c
i
csc r +r(c
i
)
0
csc r +rc
i
(csc r)
0
= 1c
i
csc r +rc
i
csc r +rc
i
(3cot r csc r)
= c
i
csc r(1 +r 3rcot r)
17.
d
dr
(csc r) =
d
dr
1
sinr
=
(sinr)(0) 31(cos r)
sin
2
r
=
3cos r
sin
2
r
= 3
1
sinr

cos r
sinr
= 3csc r cot r
19.
d
dr
(cot r) =
d
dr
cos r
sinr
=
(sinr)(3sinr) 3(cos r)(cos r)
sin
2
r
= 3
sin
2
r + cos
2
r
sin
2
r
= 3
1
sin
2
r
= 3csc
2
r
21. j = sec r i j
0
= sec r tanr, so j
0
(
r
3
) = sec
r
3
tan
r
3
= 2
I
3. An equation of the tangent line to the curve j = sec r
at the point
r
3
. 2
is j 32 = 2
I
3
r 3
r
3
or j = 2
I
3 r + 2 3
2
3
I
3 .
23. j = r +cos r i j
0
= 1 3sinr. At (0. 1), j
0
= 1, and an equation of the tangent line is j 31 = 1(r 30), or j = r +1.
25. (a) j = 2rsinr i j
0
= 2(rcos r + sinr 1). At
r
2
.
,
j
0
= 2
r
2
cos
r
2
+ sin
r
2
= 2(0 + 1) = 2, and an equation of the

tangent line is j 3 = 2
r 3
r
2
, or j = 2r.
(b)
27. (a) 1(r) = sec r 3r i 1
0
(r) = sec r tanr 31
(b) Note that 1
0
= 0 where 1 has a minimum. Also note that 1
0
is negative
when 1 is decreasing and 1
0
29. 1(0) = 0 sin0 i 1
0
(0) = 0 (cos 0) + (sin0) 1 = 0 cos 0 + sin0 i
1
00
(0) = 0 (3sin0) + (cos 0) 1 + cos 0 = 30 sin0 + 2 cos 0
31. (a) 1(r) =
tanr 31
sec r
i
1
0
(r) =
sec r(sec
2
r) 3(tanr 31)(sec rtanr)
(sec r)
2
=
sec r(sec
2
r 3tan
2
r + tanr)
sec
2
r
=
1 + tanr
sec r
(b) 1(r) =
tanr 31
sec r
=
sinr
cos r
31
1
cos r
=
sinr 3cos r
cos r
1
cos r
= sinr 3cos r i 1
0
(r) = cos r 3(3sinr) = cos r + sinr
(c) From part (a), 1
0
(r) =
1 + tanr
sec r
=
1
sec r
+
tanr
sec r
= cos r + sinr, which is the expression for 1
0
(r) in part (b).
98

33. 1(r) = r + 2 sinr has a horizontal tangent when 1
0
(r) = 0 C 1 + 2 cos r = 0 C cos r = 3
1
2
C
r =
2r
3
+ 2n or
4r
3
+ 2n, where n is an integer. Note that
4r
3
and
2r
3
are
r
3
units from . This allows us to write the
solutions in the more compact equivalent form (2n + 1)
r
3
, n an integer.
35. (a) r(t) = 8 sint i (t) = r
0
(t) = 8 cos t i o(t) = r
00
(t) = 38 sint
(b) The mass at time t =
2r
3
has position r
2r
3
= 8 sin
2r
3
= 8
I
3
2
= 4
I
3, velocity
2r
3
= 8 cos
2r
3
= 8
3
1
2
= 34,
and acceleration o
2r
3
= 38 sin
2r
3
= 38
I
3
2
= 34
I
3. Since
2r
3
< 0, the particle is moving to the left.

37. From the diagram we can see that sin0 = r10 C r = 10 sin0. We want to nd the rate
of change of r with respect to 0, that is, drd0. Taking the derivative of r = 10 sin0, we get
drd0 = 10(cos 0). So when 0 =
r
3
,
ui
u0
= 10 cos
r
3
= 10
1
2
= 5 ftrad.
39. lim
i<0
sin3r
r
= lim
i<0
3 sin3r
3r
[multiply numerator and denominator by 3]
= 3 lim
3i<0
sin3r
3r
[as r <0, 3r <0]
= 3 lim
0<0
sin0
0
[let 0 = 3r]
= 3(1) [Equation 2]
= 3
41. lim
I<0
tan6t
sin2t
= lim
I<0
sin6t
t

1
cos 6t

t
sin2t
= lim
I<0
6 sin6t
6t
lim
I<0
1
cos 6t
lim
I<0
2t
2 sin2t
= 6 lim
I<0
sin6t
6t
lim
I<0
1
cos 6t

1
2
lim
I<0
2t
sin2t
= 6(1)
1
1

1
2
(1) = 3
43. lim
0<0
sin(cos 0)
sec 0
=
sin
lim
0<0
cos 0
lim
0<0
sec 0
=
sin1
1
= sin1
45. Divide numerator and denominator by 0. (sin0 also works.)
lim
0<0
sin0
0 + tan0
= lim
0<0
sin0
0
1 +
sin0
0

1
cos 0
=
lim
0<0
sin0
0
1 + lim
0<0
sin0
0
lim
0<0
1
cos 0
=
1
1 + 1 1
=
1
2
47. lim
i<r'4
1 3tanr
sinr 3cos r
= lim
i<r'4
1 3
sinr
cos r
cos r
(sinr 3cos r) cos r
= lim
i<r'4
cos r 3sinr
(sinr 3cos r) cos r
= lim
i<r'4
31
cos r
=
31
1
I
2
= 3
I
2
49. (a)
d
dr
tanr =
d
dr
sinr
cos r
i sec
2
r =
cos r cos r 3sinr(3sinr)
cos
2
r
=
cos
2
r + sin
2
r
cos
2
r
. So sec
2
r =
1
cos
2
r
.
(b)
d
dr
sec r =
d
dr
1
cos r
i sec r tanr =
(cos r)(0) 31(3sinr)
cos
2
r
. So sec r tanr =
sinr
cos
2
r
.
SECTION 3.4 THE CHAIN RULE

99
(c)
d
dr
(sinr + cos r) =
d
dr
1 + cot r
csc r
i
cos r 3sinr =
csc r(3csc
2
r) 3(1 + cot r)(3csc r cot r)
csc
2
r
=
csc r[3csc
2
r + (1 + cot r) cot r]
csc
2
r
=
3csc
2
r + cot
2
r + cot r
csc r
=
31 + cot r
csc r
So cos r 3sinr =
cot r 31
csc r
.
51. By the denition of radian measure, c = v0, where v is the radius of the circle. By drawing the bisector of the angle 0, we can
see that sin
0
2
=
d2
v
i d = 2v sin
0
2
. So lim
0<0
+
c
d
= lim
0<0
+
v0
2v sin(02)
= lim
0<0
+
2 (02)
2 sin(02)
= lim
0<0
02
sin(02)
= 1.
[This is just the reciprocal of the limit lim
i<0
sin i
i
= 1 combined with the fact that as 0 <0,
0
2
<0 also.]
3.4 The Chain Rule
1. Let & = o(r) = 4r and j = 1(&) = sin&. Then
dj
dr
=
dj
d&
d&
dr
= (cos &)(4) = 4 cos 4r.
3. Let & = o(r) = 1 3r
2
and j = 1(&) = &
10
. Then
dj
dr
=
dj
d&
d&
dr
= (10&
9
)(32r) = 320r(1 3r
2
)
9
.
5. Let & = o(r) =
I
r and j = 1(&) = c
u
. Then
dj
dr
=
dj
d&
d&
dr
= (c
u
)
1
2
r
31'2
= c
I
i
1
2
I
r
=
c
I
i
2
I
r
.
7. 1(r) = (r
4
+ 3r
2
32)
5
i 1
0
(r) = 5(r
4
+ 3r
2
32)
4
d
dr
r
4
+ 3r
2
32
= 5(r
4
+ 3r
2
32)
4
(4r
3
+ 6r)
or 10r(r
4
+ 3r
2
32)
4
(2r
2
+ 3)
9. 1(r) =
4
I
1 + 2r +r
3
= (1 + 2r +r
3
)
1'4
i
1
0
(r) =
1
4
(1 + 2r +r
3
)
33'4
d
dr
(1 + 2r +r
3
) =
1
4(1 + 2r +r
3
)
3'4
(2 + 3r
2
) =
2 + 3r
2
4(1 + 2r +r
3
)
3'4
=
2 + 3r
2
4
4
(1 + 2r +r
3
)
3
11. o(t) =
1
(t
4
+ 1)
3
= (t
4
+ 1)
33
i o
0
(t) = 33(t
4
+ 1)
34
(4t
3
) = 312t
3
(t
4
+ 1)
34
=
312t
3
(t
4
+ 1)
4
13. j = cos(o
3
+r
3
) i j
0
= 3sin(o
3
+r
3
) 3r
2
[o
3
is just a constant] = 33r
2
sin(o
3
+r
3
)
15. j = rc
3Ii
i j
0
= r
c
3Ii
(3I)
+c
3Ii
1 = c
3Ii
(3Ir + 1)

or (1 3Ir)c
3Ii
17. o(r) = (1 + 4r)

5
(3 +r 3r
2
)
8
i
o
0
(r) = (1 + 4r)
5
8(3 +r 3r
2
)
7
(1 32r) + (3 +r 3r
2
)
8
5(1 + 4r)
4
4
= 4(1 + 4r)
4
(3 +r 3r
2
)
7
2(1 + 4r)(1 32r) + 5(3 +r 3r

2
)
= 4(1 + 4r)
4
(3 +r 3r
2
)
7
(2 + 4r 316r
2
) + (15 + 5r 35r
2
)
= 4(1 + 4r)
4
(3 +r 3r
2
)
7
(17 + 9r 321r
2
)
100

19. j = (2r 35)
4
(8r
2
35)
33
i
j
0
= 4(2r 35)
3
(2)(8r
2
35)
33
+ (2r 35)
4
(33)(8r
2
35)
34
(16r)
= 8(2r 35)
3
(8r
2
35)
33
348r(2r 35)
4
(8r
2
35)
34
[This simplies to 8(2r 35)
3
(8r
2
35)
34
(34r
2
+ 30r 35).]
21. j =

r
2
+ 1
r
2
31
3
i
j
0
= 3
r
2
+ 1
r
2
31
d
dr
r
2
+ 1
r
2
31
= 3
r
2
+ 1
r
2
31
(r
2
31)(2r) 3(r
2
+ 1)(2r)
(r
2
31)
2
= 3
r
2
+ 1
r
2
31
2r[r
2
31 3(r
2
+ 1)]
(r
2
31)
2
= 3
r
2
+ 1
r
2
31
2r(32)
(r
2
31)
2
=
312r(r
2
+ 1)
2
(r
2
31)
4
23. j = c
i cos i
i j
0
= c
i cos i
d
dr
(rcos r) = c
icos i
[r(3sinr) + (cos r) 1] = c
icos i
(cos r 3rsinr)
25. 1(:) =
: 31
: + 1
=

: 31
: + 1
1'2
i
1
0
(:) =
1
2
: 31
: + 1
31'2
d
d:
: 31
: + 1
=
1
2
: + 1
: 31
1'2
(: + 1)(1) 3(: 31)(1)

(: + 1)
2
=
1
2
(: + 1)
1'2
(: 31)
1'2

: + 1 3: + 1
(: + 1)
2
=
1
2
(: + 1)
1'2
(: 31)
1'2

2
(: + 1)
2
=
1
(: 31)
1'2
(: + 1)
3'2
27. j =
v
I
v
2
+ 1
i
j
0
=
I
v
2
+ 1 (1) 3v
1
2
(v
2
+ 1)
31'2
(2v)
I
v
2
+ 1
2
=
I
v
2
+ 1 3
v
2
I
v
2
+ 1
I
v
2
+ 1
2
=
I
v
2
+ 1
I
v
2
+ 1 3v
2
I
v
2
+ 1
I
v
2
+ 1
2
=
v
2
+ 1
3v
2
I
v
2
+ 1
3
=
1
(v
2
+ 1)
3'2
or (v
2
+ 1)
33'2
Another solution: Write j as a product and make use of the Product Rule. j = v(v
2
+ 1)
31'2
i
j
0
= v 3
1
2
(v
2
+ 1)
33'2
(2v) + (v
2
+ 1)
31'2
1 = (v
2
+ 1)
33'2
[3v
2
+ (v
2
+ 1)
1
] = (v
2
+ 1)
33'2
(1) = (v
2
+ 1)
33'2
.
The step that students usually have trouble with is factoring out (v
2
+ 1)
33'2
. But this is no different than factoring out r
2
from r
2
+r
5
; that is, we are just factoring out a factor with the smallest exponent that appears on it. In this case, 3
3
2
is
smaller than 3
1
2
.
29. j = sin(tan2r) i j
0
= cos(tan2r)
d
dr
(tan2r) = cos(tan2r) sec
2
(2r)
d
dr
(2r) = 2 cos(tan2r) sec
2
(2r)
31. Using Formula 5 and the Chain Rule, j = 2
sin ri
i
j
0
= 2
sin ri
(ln2)
d
dr
(sinr) = 2
sin ri
(ln2) cos r = 2
sin ri
( ln2) cos r

101
33. j = sec
2
r + tan
2
r = (sec r)
2
+ (tanr)
2
i
j
0
= 2(sec r)(sec r tanr) + 2(tanr)(sec
2
r) = 2 sec
2
r tanr + 2 sec
2
r tanr = 4 sec
2
r tanr
35. j = cos
1 3c
2i
1 +c
2i
i
j
0
= 3sin
1 3c
2i
1 +c
2i
d
dr
1 3c
2i
1 +c
2i
= 3sin
1 3c
2i
1 +c
2i
(1 +c
2i
)(32c
2i
) 3(1 3c
2i
)(2c
2i
)
(1 +c
2i
)
2
= 3sin
1 3c
2i
1 +c
2i
32c
2i
(1 +c
2i
) + (1 3c
2i
)
(1 +c
2i
)
2
= 3sin
1 3c
2i
1 +c
2i
32c
2i
(2)
(1 +c
2i
)
2
=
4c
2i
(1 +c
2i
)
2
sin
1 3c
2i
1 +c
2i
37. j = cot
2
(sin0) = [cot(sin0)]
2
i
j
0
= 2[cot(sin0)]
d
d0
[cot(sin0)] = 2 cot(sin0) [3csc
2
(sin0) cos 0] = 32 cos 0 cot(sin0) csc
2
(sin0)
39. 1(t) = tan(c
I
) +c
tan I
i 1
0
(t) = sec
2
(c
I
)
d
dt
(c
I
) +c
tan I
d
dt
(tant) = sec
2
(c
I
) c
I
+c
tan I
sec
2
t
41. 1(t) = sin
2
c
sin
2
I
=

sin
c
sin
2
I
2
i
1
0
(t) = 2
sin
c
sin
2
I
d
dt
sin
c
sin
2
I
= 2 sin
c
sin
2
I
cos
c
sin
2
I
d
dt
c
sin
2
I
= 2 sin
c
sin
2
I
cos
c
sin
2
I
c
sin
2
I
d
dt
sin
2
t = 2 sin
c
sin
2
I
cos
c
sin
2
I
c
sin
2
I
2 sint cos t
= 4 sin
c
sin
2
I
cos
c
sin
2
I
c
sin
2
I
sint cos t
43. o(r) = (2vo
ri
+ n)
i
o
0
(r) = j(2vo
ri
+n)
31
d
dr
(2vo
ri
+n) = j(2vo
ri
+n)
31
2vo
ri
(lno) v = 2v
2
j(lno)(2vo
ri
+n)
31
o
ri
45. j = cos
sin(tanr) = cos(sin(tanr))
1'2
i
j
0
= 3sin(sin(tanr))
1'2
d
dr
(sin(tanr))
1'2
= 3sin(sin(tanr))
1'2
1
2
(sin(tanr))
31'2
d
dr
(sin(tanr))
=
3sin
sin(tanr)
2
sin(tanr)
cos(tanr)
d
dr
tanr =
3sin
sin(tanr)
2
sin(tanr)
cos(tanr) sec
2
(r)
=
3 cos(tanr) sec
2
(r) sin
sin(tanr)
2
sin(tanr)
47. /(r) =
I
r
2
+ 1 i /
0
(r) =
1
2
(r
2
+ 1)
31'2
(2r) =
r
I
r
2
+ 1
i
/
00
(r) =
I
r
2
+ 1 1 3r
1
2
(r
2
+ 1)
31'2
(2r)
I
r
2
+ 1
2
=
r
2
+ 1
31'2
(r
2
+ 1) 3r
2
(r
2
+ 1)
1
=
1
(r
2
+ 1)
3'2
102

49. j = c
oi
sinor i j
0
= c
oi
o cos or + sinor cc
oi
= c
oi
(o cos or +csinor) i
j
00
= c
oi
(3o
2
sinor +co cos or) + (o cos or +csinor) cc
oi
= c
oi
(3o
2
sinor +co cos or +co cos or +c
2
sinor) = c
oi
(c
2
sinor 3o
2
sinor + 2co cos or)
= c
oi
(c
2
3o
2
) sinor + 2co cos or
51. j = (1 + 2r)
10
i j
0
= 10(1 + 2r)
9
2 = 20(1 + 2r)
9
.
At (0. 1), j
0
= 20(1 + 0)
9
= 20, and an equation of the tangent line is j 31 = 20(r 30), or j = 20r + 1.
53. j = sin(sinr) i j
0
= cos(sinr) cos r. At (. 0), j
0
= cos(sin) cos = cos(0) (31) = 1(31) = 31, and an
equation of the tangent line is j 30 = 31(r 3), or j = 3r +.
55. (a) j =
2
1 +c
3i
i j
0
=
(1 +c
3i
)(0) 32(3c
3i
)
(1 +c
3i
)
2
=
2c
3i
(1 +c
3i
)
2
.
At (0. 1), j
0
=
2c
0
(1 +c
0
)
2
=
2(1)
(1 + 1)
2
=
2
2
2
=
1
2
. So an equation of the
tangent line is j 31 =
1
2
(r 30) or j =
1
2
r + 1.
(b)
57. (a) 1(r) = r
I
2 3r
2
= r(2 3 r
2
)
1'2
i
1
0
(r) = r
1
2
(2 3r
2
)
31'2
(32r) + (2 3r
2
)
1'2
1 = (2 3r
2
)
31'2
3r
2
+ (2 3r
2
)
=
2 32r
2
I
2 3r
2
(b) 1
0
= 0 when 1 has a horizontal tangent line, 1
0
is negative when 1 is
decreasing, and 1
0
59. For the tangent line to be horizontal, 1
0
(r) = 0. 1(r) = 2 sinr + sin
2
r i 1
0
(r) = 2 cos r + 2 sinr cos r = 0 C
2 cos r(1 + sinr) = 0 C cos r = 0 or sinr = 31, so r =
r
2
+ 2n or
3r
2
+ 2n, where n is any integer. Now
1
r
2
= 3 and 1
3r
2
= 31, so the points on the curve with a horizontal tangent are
r
2
+ 2n. 3
and
3r
2
+ 2n. 31
,
where n is any integer.
61. 1(r) = 1(o(r)) i 1
0
(r) = 1
0
(o(r)) o
0
(r), so 1
0
(5) = 1
0
(o(5)) o
0
(5) = 1
0
(32) 6 = 4 6 = 24
63. (a) /(r) = 1(o(r)) i /
0
(r) = 1
0
(o(r)) o
0
(r), so /
0
(1) = 1
0
(o(1)) o
0
(1) = 1
0
(2) 6 = 5 6 = 30.
(b) 1(r) = o(1(r)) i 1
0
(r) = o
0
(1(r)) 1
0
(r), so 1
0
(1) = o
0
(1(1)) 1
0
(1) = o
0
(3) 4 = 9 4 = 36.
65. (a) &(r) = 1(o(r)) i &
0
(r) = 1
0
(o(r))o
0
(r). So &
0
(1) = 1
0
(o(1))o
0
(1) = 1
0
(3)o
0
(1). To nd 1
0
(3), note that 1 is
linear from (2. 4) to (6. 3), so its slope is
3 34
6 32
= 3
1
4
. To nd o
0
(1), note that o is linear from (0. 6) to (2. 0), so its slope
is
0 36
2 30
= 33. Thus, 1
0
(3)o
0
(1) =
3
1
4
(33) =
3
4
.
(b) (r) = o(1(r)) i
0
(r) = o
0
(1(r))1
0
(r). So
0
(1) = o
0
(1(1))1
0
(1) = o
0
(2)1
0
(1), which does not exist since
o
0
(2) does not exist.

103
(c) n(r) = o(o(r)) i n
0
(r) = o
0
(o(r))o
0
(r). So n
0
(1) = o
0
(o(1))o
0
(1) = o
0
(3)o
0
(1). To nd o
0
(3), note that o is
linear from (2. 0) to (5. 2), so its slope is
2 30
5 32
=
2
3
. Thus, o
0
(3)o
0
(1) =

2
3
(33) = 32.
67. (a) 1(r) = 1(c
i
) i 1
0
(r) = 1
0
(c
i
)
d
dr
(c
i
) = 1
0
(c
i
)c
i
(b) G(r) = c
](i)
i G
0
(r) = c
](i)
d
dr
1(r) = c
](i)
1
0
(r)
69. v(r) = 1(o(/(r))) i v
0
(r) = 1
0
(o(/(r))) o
0
(/(r)) /
0
(r), so
v
0
(1) = 1
0
(o(/(1))) o
0
(/(1)) /
0
(1) = 1
0
(o(2)) o
0
(2) 4 = 1
0
(3) 5 4 = 6 5 4 = 120
71. 1(r) = 1(31(41(r))) i
1
0
(r) = 1
0
(31(41(r)))
d
dr
(31(41(r))) = 1
0
(31(41(r))) 31
0
(41(r))
d
dr
(41(r))
= 1
0
(31(41(r))) 31
0
(41(r)) 41
0
(r), so
1
0
(0) = 1
0
(31(41(0))) 31
0
(41(0)) 41
0
(0) = 1
0
(31(4 0)) 31
0
(4 0) 4 2 = 1
0
(3 0) 3 2 4 2 = 2 3 2 4 2 = 96.
73. j = c
3i
+1rc
3i
i
j
0
= (3c
3i
) +1[r(3c
3i
) +c
3i
1] = 3c
3i
+1c
3i
31rc
3i
= (1 3)c
3i
31rc
3i
i
j
00
= (1 3)(3c
3i
) 31[r(3c
3i
) +c
3i
1] = (31)c
3i
31c
3i
+1rc
3i
= (321)c
3i
+1rc
3i
,
so j
00
+ 2j
0
+j = (321)c
3i
+1rc
3i
+ 2[(1 3)c
3i
31rc
3i
] +c
3i
+1rc
3i
= [(321) + 2(1 3) +]c
3i
+ [1 321 +1]rc
3i
= 0.
75. The use of 1, 1
2
, . . ., 1
n
is just a derivative notation (see text page 157). In general, 11(2r) = 21
0
(2r),
1
2
1(2r) = 41
00
(2r), . . ., 1
n
1(2r) = 2
n
1
(n)
(2r). Since 1(r) = cos r and 50 = 4(12) + 2, we have
1
(50)
(r) = 1
(2)
(r) = 3cos r, so 1
50
cos 2r = 32
50
cos 2r.
77. c(t) = 10 +
1
4
sin(10t) i the velocity after t seconds is (t) = c
0
(t) =
1
4
cos(10t)(10) =
5r
2
cos(10t) cms.
79. (a) 1(t) = 4.0 + 0.35 sin
2t
5.4
i
d1
dt
=

0.35 cos
2t
5.4
2
5.4
=
0.7
5.4
cos
2t
5.4
=
7
54
cos
2t
5.4
(b) At t = 1,
d1
dt
=
7
54
cos
2
5.4
E 0.16.
81. c(t) = 2c
31.5I
sin2t i
(t) = c
0
(t) = 2[c
31.5I
(cos 2t)(2) + (sin2t)c
31.5I
(31.5)] = 2c
31.5I
(2 cos 2t 31.5 sin2t)
104

83. By the Chain Rule, o(t) =
d
dt
=
d
dc
dc
dt
=
d
dc
(t) = (t)
d
dc
. The derivative ddt is the rate of change of the velocity
with respect to time (in other words, the acceleration) whereas the derivative ddc is the rate of change of the velocity with
respect to the displacement.
85. (a) Using a calculator or CAS, we obtain the model Q = o/
I
with o E 100.0124369 and / E 0.000045145933.
(b) Use Q
0
(t) = o/
I
ln/ (from Formula 5) with the values of o and / from part (a) to get Q
0
(0.04) E 3670.63 jA.
The result of Example 2 in Section 2.1 was 3670 jA.
87. (a) Derive gives o
0
(t) =
45(t 32)
8
(2t + 1)
10
without simplifying. With either Maple or Mathematica, we rst get
o
0
(t) = 9
(t 32)
8
(2t + 1)
9
318
(t 32)
9
(2t + 1)
10
, and the simplication command results in the expression given by Derive.
(b) Derive gives j
0
= 2(r
3
3r + 1)
3
(2r + 1)
4
(17r
3
+ 6r
2
39r + 3) without simplifying. With either Maple or
Mathematica, we rst get j
0
= 10(2r + 1)
4
(r
3
3r + 1)
4
+ 4(2r + 1)
5
(r
3
3r + 1)
3
(3r
2
31). If we use
Mathematicas Factor or Simplify, or Maples factor, we get the above expression, but Maples simplify gives
the polynomial expansion instead. For locating horizontal tangents, the factored form is the most helpful.
89. (a) If 1 is even, then 1(r) = 1(3r). Using the Chain Rule to differentiate this equation, we get
1
0
(r) = 1
0
(3r)
d
dr
(3r) = 31
0
(3r). Thus, 1
0
(3r) = 31
0
(r), so 1
0
is odd.
(b) If 1 is odd, then 1(r) = 31(3r). Differentiating this equation, we get 1
0
(r) = 31
0
(3r)(31) = 1
0
(3r), so 1
0
is
even.
91. (a)
d
dr
(sin
n
r cos nr) = nsin
n31
r cos r cos nr + sin
n
r(3nsinnr) [Product Rule]
= nsin
n31
r(cos nr cos r 3sinnr sinr) [factor out nsin
n31
r]
= nsin
n31
rcos(nr +r) [Addition Formula for cosine]
= nsin
n31
rcos[(n + 1)r] [factor out r]
(b)
d
dr
(cos
n
r cos nr) = ncos
n31
r(3sinr) cos nr + cos
n
r(3nsinnr) [Product Rule]
= 3ncos
n31
r(cos nr sinr + sinnr cos r) [factor out 3ncos
n31
r]
= 3ncos
n31
rsin(nr +r) [Addition Formula for sine]
= 3ncos
n31
rsin[(n + 1)r] [factor out r]
93. Since 0
=

r
180
0 rad, we have
d
d0
(sin0
) =
d
d0
sin
r
180
0
=
r
180
cos
r
180
0 =
r
180
cos 0
.
95. The Chain Rule says that
dj
dr
=
dj
d&
d&
dr
, so
d
2
j
dr
2
=
d
dr
dj
dr
=
d
dr
dj
d&
d&
dr
=

d
dr
dj
d&
d&
dr
+
dj
d&
d
dr
d&
dr
[Product Rule]
=

d
d&
dj
d&
d&
dr
d&
dr
+
dj
d&
d
2
&
dr
2
=
d
2
j
d&
2
d&
dr
2
+
dj
d&
d
2
&
dr
2
SECTION 3.5 IMPLICIT DIFFERENTIATION

105
3.5 Implicit Differentiation
1. (a)
d
dr
(rj + 2r + 3r
2
) =
d
dr
(4) i (r j
0
+j 1) + 2 + 6r = 0 i rj
0
= 3j 32 36r i
j
0
=
3j 32 36r
r
or j
0
= 36 3
j + 2
r
.
(b) rj + 2r + 3r
2
= 4 i rj = 4 32r 33r
2
i j =
4 32r 33r
2
r
=
4
r
32 33r, so j
0
= 3
4
r
2
33.
(c) From part (a), j
0
=
3j 32 36r
r
=
3(4r 32 33r) 32 36r
r
=
34r 33r
r
= 3
4
r
2
33.
3. (a)
d
dr
1
r
+
1
j
=
d
dr
(1) i 3
1
r
2
3
1
j
2
j
0
= 0 i 3
1
j
2
j
0
=
1
r
2
i j
0
= 3
j
2
r
2
(b)
1
r
+
1
j
= 1 i
1
j
= 1 3
1
r
=
r 31
r
i j =
r
r 31
, so j
0
=
(r 31)(1) 3(r)(1)
(r 31)
2
=
31
(r 31)
2
.
(c) j
0
= 3
j
2
r
2
= 3
[r(r 31)]
2
r
2
= 3
r
2
r
2
(r 31)
2
= 3
1
(r 31)
2
5.
d
dr
r
3
+j
3
=
d
dr
(1) i 3r
2
+ 3j
2
j
0
= 0 i 3j
2
j
0
= 33r
2
i j
0
= 3
r
2
j
2
7.
d
dr
(r
2
+rj 3j
2
) =
d
dr
(4) i 2r +r j
0
+j 1 32j j
0
= 0 i
rj
0
32j j
0
= 32r 3j i (r 32j) j
0
= 32r 3j i j
0
=
32r 3j
r 32j
=
2r +j
2j 3r
9.
d
dr
r
4
(r +j)
=
d
dr
j
2
(3r 3j)
i r
4
(1 +j
0
) + (r +j) 4r
3
= j
2
(3 3j
0
) + (3r 3j) 2j j
0
i
r
4
+r
4
j
0
+ 4r
4
+ 4r
3
j = 3j
2
3j
2
j
0
+ 6rj j
0
32j
2
j
0
i r
4
j
0
+ 3j
2
j
0
36rj j
0
= 3j
2
35r
4
34r
3
j i
(r
4
+ 3j
2
36rj) j
0
= 3j
2
35r
4
34r
3
j i j
0
=
3j
2
35r
4
34r
3
j
r
4
+ 3j
2
36rj
11.
d
dr
(r
2
j
2
+rsinj) =
d
dr
(4) i r
2
2j j
0
+j
2
2r +rcos j j
0
+ sinj 1 = 0 i
2r
2
j j
0
+rcos j j
0
= 32rj
2
3sinj i (2r
2
j +rcos j)j
0
= 32rj
2
3sinj i j
0
=
32rj
2
3sinj
2r
2
j +rcos j
13.
d
dr
(4 cos r sinj) =
d
dr
(1) i 4 [cos r cos j j
0
+ sinj (3sinr)] = 0 i
j
0
(4 cos r cos j) = 4 sinr sinj i j
0
=
4 sinr sinj
4 cos r cos j
= tanr tanj
15.
d
dr
(c
i'u
) =
d
dr
(r 3 j) i c
i'u
d
dr
r
j
= 1 3 j
0
i
c
i'u
j 1 3r j
0
j
2
= 1 3j
0
i c
i'u
1
j
3
rc
i'u
j
2
j
0
= 1 3j
0
i j
0
3
rc
i'u
j
2
j
0
= 1 3
c
i'u
j
i
j
0
1 3
rc
i'u
j
2
=
j 3c
i'u
j
i j
0
=
j 3c
i'u
j
j
2
3rc
i'u
j
2
=
j(j 3c
i'u
)
j
2
3rc
i'u
106

17.

rj = 1 +r
2
j i
1
2
(rj)
31'2
(rj
0
+j 1) = 0 +r
2
j
0
+j 2r i
r
2
rj
j
0
+
j
2
rj
= r
2
j
0
+ 2rj i
j
0
r
2
rj
3r
2
= 2rj 3
j
2
rj
i j
0
r 32r
2
rj
2
rj
=
4rj
rj 3j
2
rj
i j
0
=
4rj
rj 3j
r 32r
2
rj
19.
d
dr
(c
u
cos r) =
d
dr
[1 + sin(rj)] i c
u
(3sinr) + cos r c
u
j
0
= cos(rj) (rj
0
+j 1) i
3c
u
sinr +c
u
cos r j
0
= rcos(rj) j
0
+j cos(rj) i c
u
cos r j
0
3rcos(rj) j
0
= c
u
sinr +j cos(rj) i
[c
u
cos r 3rcos(rj)] j
0
= c
u
sinr +j cos(rj) i j
0
=
c
u
sinr +j cos(rj)
c
u
cos r 3rcos(rj)
21.
d
dr
1(r) +r
2
[1(r)]
3
=
d
dr
(10) i 1
0
(r) +r
2
3[1(r)]
2
1
0
(r) + [1(r)]
3
2r = 0. If r = 1, we have
1
0
(1) + 1
2
3[1(1)]
2
1
0
(1) + [1(1)]
3
2(1) = 0 i 1
0
(1) + 1 3 2
2
1
0
(1) + 2
3
2 = 0 i
1
0
(1) + 121
0
(1) = 316 i 131
0
(1) = 316 i 1
0
(1) = 3
16
13
.
23.
d
dj
(r
4
j
2
3r
3
j + 2rj
3
) =
d
dj
(0) i r
4
2j +j
2
4r
3
r
0
3(r
3
1 +j 3r
2
r
0
) + 2(r 3j
2
+j
3
r
0
) = 0 i
4r
3
j
2
r
0
33r
2
j r
0
+ 2j
3
r
0
= 32r
4
j +r
3
36rj
2
i (4r
3
j
2
33r
2
j + 2j
3
) r
0
= 32r
4
j +r
3
36rj
2
i
r
0
=
dr
dj
=
32r
4
j +r
3
36rj
2
4r
3
j
2
33r
2
j + 2j
3
25. r
2
+rj +j
2
= 3 i 2r +rj
0
+j 1 + 2jj
0
= 0 i rj
0
+ 2j j
0
= 32r 3j i j
0
(r + 2j) = 32r 3j i
j
0
=
32r 3j
r + 2j
. When r = 1 and j = 1, we have j
0
=
32 31
1 + 2 1
=
33
3
= 31, so an equation of the tangent line is
j 31 = 31(r 31) or j = 3r + 2.
27. r
2
+j
2
= (2r
2
+ 2j
2
3r)
2
i 2r + 2j j
0
= 2(2r
2
+ 2j
2
3r)(4r + 4j j
0
31). When r = 0 and j =
1
2
, we have
0 +j
0
= 2(
1
2
)(2j
0
31) i j
0
= 2j
0
31 i j
0
= 1, so an equation of the tangent line is j 3
1
2
= 1(r 30)
or j = r +
1
2
.
29. 2(r
2
+ j
2
)
2
= 25(r
2
3 j
2
) i 4(r
2
+ j
2
)(2r + 2j j
0
) = 25(2r 3 2j j
0
) i
4(r +j j
0
)(r
2
+j
2
) = 25(r 3j j
0
) i 4j j
0
(r
2
+j
2
) + 25jj
0
= 25r 34r(r
2
+j
2
) i
j
0
=
25r 34r(r
2
+j
2
)
25j + 4j(r
2
+j
2
)
. When r = 3 and j = 1, we have j
0
=
75 3120
25 +40
= 3
45
65
= 3
9
13
,
so an equation of the tangent line is j 31 = 3
9
13
(r 33) or j = 3
9
13
r +
40
13
.
31. (a) j
2
= 5r
4
3r
2
i 2j j
0
= 5(4r
3
) 32r i j
0
=
10r
3
3r
j
.
So at the point (1. 2) we have j
0
=
10(1)
3
31
2
=
9
2
, and an equation
of the tangent line is j 32 =
9
2
(r 31) or j =
9
2
r 3
5
2
.
(b)

107
33. 9r
2
+ j
2
= 9 i 18r + 2j j
0
= 0 i 2j j
0
= 318r i j
0
= 39rj i
j
00
= 39
j 1 3r j
0
j
2
= 39
j 3r(39rj)
j
2
= 39
j
2
+ 9r
2
j
3
= 39
9
j
3
[since x and y must satisfy the original
equation, 9r
2
+j
2
= 9]. Thus, j
00
= 381j
3
.
35. r
3
+ j
3
= 1 i 3r
2
+ 3j
2
j
0
= 0 i j
0
= 3
r
2
j
2
i
j
00
= 3
j
2
(2r) 3r
2
2j j
0
(j
2
)
2
= 3
2rj
2
32r
2
j(3r
2
j
2
)
j
4
= 3
2rj
4
+ 2r
4
j
j
6
= 3
2rj(j
3
+r
3
)
j
6
= 3
2r
j
5
,
since r and j must satisfy the original equation, r
3
+j
3
= 1.
37. (a) There are eight points with horizontal tangents: four at r E 1.57735 and
four at r E 0.42265.
(b) j
0
=
3r
2
36r + 2
2(2j
3
33j
2
3j + 1)
i j
0
= 31 at (0. 1) and j
0
=
1
3
at (0. 2).
Equations of the tangent lines are j = 3r + 1 and j =
1
3
r + 2.
(c) j
0
= 0 i 3r
2
36r + 2 = 0 i r = 1
1
3
I
3
(d) By multiplying the right side of the equation by r 33, we obtain the rst
graph. By modifying the equation in other ways, we can generate the other
graphs.
j(j
2
31)(j 32)
= r(r 31)(r 32)(r 33)
j(j
2
34)(j 32)
= r(r 31)(r 32)
j(j + 1)(j
2
31)(j 32)
= r(r 31)(r 32)
(j + 1)(j
2
31)(j 32)
= (r 31)(r 32)
r(j + 1)(j
2
31)(j 32)
= j(r 31)(r 32)
j(j
2
+ 1)(j 32)
= r(r
2
31)(r 32)
j(j + 1)(j
2
32)
= r(r 31)(r
2
32)
108

39. From Exercise 29, a tangent to the lemniscate will be horizontal if j
0
= 0 i 25r 34r(r
2
+j
2
) = 0 i
r[25 34(r
2
+j
2
)] = 0 i r
2
+j
2
=
25
4
(1). (Note that when r is 0, j is also 0, and there is no horizontal tangent
at the origin.) Substituting
25
4
for r
2
+j
2
in the equation of the lemniscate, 2(r
2
+j
2
)
2
= 25(r
2
3j
2
), we get
r
2
3j
2
=
25
8
(2). Solving (1) and (2), we have r
2
=
75
16
and j
2
=
25
16
, so the four points are
5
I
3
4
.
5
4
.
41.
r
2
o
2
3
j
2
/
2
= 1 i
2r
o
2
3
2jj
0
/
2
= 0 i j
0
=
/
2
r
o
2
j
i an equation of the tangent line at (r
0
. j
0
) is
j 3j
0
=
/
2
r
0
o
2
j
0
(r 3r
0
). Multiplying both sides by
j
0
/
2
gives
j
0
j
/
2
3
j
2
0
/
2
=
r
0
r
o
2
3
r
2
0
o
2
. Since (r
0
. j
0
) lies on the hyperbola,
we have
r
0
r
o
2
3
j
0
j
/
2
=
r
2
0
o
2
3
j
2
0
/
2
= 1.
43. If the circle has radius v, its equation is r
2
+j
2
= v
2
i 2r + 2jj
0
= 0 i j
0
= 3
r
j
, so the slope of the tangent line
at 1(r
0
. j
0
) is 3
r
0
j
0
. The negative reciprocal of that slope is
31
3r
0
j
0
=
j
0
r
0
, which is the slope of O1, so the tangent line at
1 is perpendicular to the radius O1.
45. j = tan
31
I
r i j
0
=
1
1 +
I
r
2

d
dr
I
r
=
1
1 +r
1
2
r
31'2
=
1
2
I
r(1 +r)
47. j = sin
31
(2r + 1) i
j
0
=
1
1 3(2r + 1)
2

d
dr
(2r + 1) =
1
1 3(4r
2
+ 4r + 1)
2 =
2
I
34r
2
34r
=
1
I
3r
2
3r
49. G(r) =
I
1 3r
2
arccos r i G
0
(r) =
I
1 3r
2

31
I
1 3r
2
+ arccos r
1
2
(1 3r
2
)
31'2
(32r) = 31 3
rarccos r
I
1 3r
2
51. /(t) = cot
31
(t) + cot
31
(1t) i
/
0
(t) = 3
1
1 +t
2
3
1
1 + (1t)
2

d
dt
1
t
= 3
1
1 +t
2
3
t
2
t
2
+ 1

3
1
t
2
= 3
1
1 +t
2
+
1
t
2
+ 1
= 0.
Note that this makes sense because /(t) =

2
for t 0 and /(t) =
3
2
for t < 0.
53. j = cos
31
(c
2i
) i j
0
= 3
1
1 3(c
2i
)
2

d
dr
(c
2i
) = 3
2c
2i
I
1 3c
4i
55. 1(r) =
I
1 3r
2
arcsinr i 1
0
(r) =
I
1 3r
2

1
I
1 3r
2
+ arcsinr
1
2
1 3r
2
31'2
(32r) = 1 3
rarcsinr
I
1 3r
2
Note that 1
0
= 0 where the graph of 1 has a horizontal tangent. Also note
that 1
0
is negative when 1 is decreasing and 1
0
is positive when 1 is
increasing.

109
57. Let j = cos
31
r. Then cos j = r and 0 $ j $ i 3sinj
dj
dr
= 1 i
dj
dr
= 3
1
sinj
= 3
1
1 3cos
2
j
= 3
1
I
1 3r
2
. [Note that sinj D 0 for 0 $ j $ .]
59. r
2
+j
2
= v
2
is a circle with center O and or +/j = 0 is a line through O [assume o
and / are not both zero]. r
2
+j
2
= v
2
i 2r + 2jj
0
= 0 i j
0
= 3rj, so the
slope of the tangent line at 1
0
(r
0
. j
0
) is 3r
0
j
0
. The slope of the line O1
0
is j
0
r
0
,
which is the negative reciprocal of 3r
0
j
0
. Hence, the curves are orthogonal, and the
families of curves are orthogonal trajectories of each other.
61. j = cr
2
i j
0
= 2cr and r
2
+ 2j
2
= I [assume I 0] i 2r + 4jj
0
= 0 i
2jj
0
= 3r i j
0
= 3
r
2(j)
= 3
r
2(cr
2
)
= 3
1
2cr
, so the curves are orthogonal if
c 6= 0. If c = 0, then the horizontal line j = cr
2
= 0 intersects r
2
+ 2j
2
= I orthogonally
at
I
I. 0
, since the ellipse r

2
+ 2j
2
= I has vertical tangents at those two points.
63. To nd the points at which the ellipse r
2
3rj +j
2
= 3 crosses the r-axis, let j = 0 and solve for r.
j = 0 i r
2
3r(0) + 0
2
= 3 C r =
I
3. So the graph of the ellipse crosses the r-axis at the points
I
3. 0
.
Using implicit differentiation to nd j
0
, we get 2r 3rj
0
3j + 2jj
0
= 0 i j
0
(2j 3r) = j 32r C j
0
=
j 32r
2j 3r
.
So j
0
at
I
3. 0
is
0 32
I
3
2(0) 3
I
3
= 2 and j
0
at
3
I
3. 0
is
0 + 2
I
3
2(0) +
I
3
= 2. Thus, the tangent lines at these points are parallel.
65. r
2
j
2
+rj = 2 i r
2
2jj
0
+j
2
2r +r j
0
+j 1 = 0 C j
0
(2r
2
j +r) = 32rj
2
3j C
j
0
= 3
2rj
2
+j
2r
2
j +r
. So 3
2rj
2
+j
2r
2
j +r
= 31 C 2rj
2
+j = 2r
2
j +r C j(2rj + 1) = r(2rj + 1) C
j(2rj + 1) 3r(2rj + 1) = 0 C (2rj + 1)(j 3r) = 0 C rj = 3
1
2
or j = r. But rj = 3
1
2
i
r
2
j
2
+rj =
1
4
3
1
2
6= 2, so we must have r = j. Then r
2
j
2
+rj = 2 i r
4
+r
2
= 2 C r
4
+r
2
32 = 0 C
(r
2
+ 2)(r
2
31) = 0. So r
2
= 32, which is impossible, or r
2
= 1 C r = 1. Since r = j, the points on the curve
where the tangent line has a slope of 31 are (31. 31) and (1. 1).
67. (a) If j = 1
31
(r), then 1(j) = r. Differentiating implicitly with respect to r and remembering that j is a function of r,
we get 1
0
(j)
dj
dr
= 1, so
dj
dr
=
1
1
0
(j)
i

1
31
0
(r) =
1
1
0
(1
31
(r))
.
(b) 1(4) = 5 i 1
31
(5) = 4. By part (a),
1
31
0
(5) =
1
1
0
(1
31
(5))
=
1
1
0
(4)
= 1
2
3
=
3
2
.
110

69. r
2
+ 4j
2
= 5 i 2r + 4(2jj
0
) = 0 i j
0
= 3
r
4j
. Now let / be the height of the lamp, and let (o. /) be the point of
tangency of the line passing through the points (3. /) and (35. 0). This line has slope (/ 30)[3 3(35)] =
1
8
/. But the
slope of the tangent line through the point (o. /) can be expressed as j
0
= 3
o
4/
, or as
/ 30
o 3(35)
=
/
o + 5
[since the line
passes through (35. 0) and (o. /)], so 3
o
4/
=
/
o + 5
C 4/
2
= 3o
2
35o C o
2
+ 4/
2
= 35o. But o
2
+ 4/
2
= 5
[since (o. /) is on the ellipse], so 5 = 35o C o = 31. Then 4/
2
= 3o
2
35o = 31 35(31) = 4 i / = 1, since the
point is on the top half of the ellipse. So
/
8
=
/
o + 5
=
1
31 + 5
=
1
4
i / = 2. So the lamp is located 2 units above the
r-axis.
3.6 Derivatives of Logarithmic Functions
1. The differentiation formula for logarithmic functions,
d
dr
(log
a
r) =
1
rlno
, is simplest when o = c because lnc = 1.
3. 1(r) = sin(lnr) i 1
0
(r) = cos(lnr)
d
dr
lnr = cos(lnr)
1
r
=
cos(lnr)
r
5. 1(r) = log
2
(1 33r) i 1
0
(r) =
1
(1 33r) ln2
d
dr
(1 33r) =
33
(1 33r) ln2
or
3
(3r 31) ln2
7. 1(r) =
5
I
lnr = (lnr)
1'5
i 1
0
(r) =
1
5
(lnr)
34'5
d
dr
(lnr) =
1
5(lnr)
4'5

1
r
=
1
5r
5
(lnr)
4
9. 1(r) = sinrln(5r) i 1
0
(r) = sinr
1
5r

d
dr
(5r) +ln(5r) cos r =
sinr 5
5r
+cos rln(5r) =
sinr
r
+cos rln(5r)
11. 1(t) = ln
(2t + 1)
3
(3t 31)
4
= ln(2t + 1)
3
3 ln(3t 31)
4
= 3 ln(2t + 1) 3 4 ln(3t 3 1) i
1
0
(t) = 3
1
2t + 1
2 34
1
3t 31
3 =
6
2t + 1
3
12
3t 31
, or combined,
36(t + 3)
(2t + 1)(3t 31)
.
13. o(r) = ln
r
I
r
2
31
= lnr + ln(r
2
3 1)
1'2
= lnr +
1
2
ln(r
2
3 1) i
o
0
(r) =
1
r
+
1
2

1
r
2
31
2r =
1
r
+
r
r
2
31
=
r
2
31 +r r
r(r
2
31)
=
2r
2
31
r(r
2
31)
15. 1(&) =
ln&
1 + ln(2&)
i
1
0
(&) =
[1 + ln(2&)]
1
u
3ln&
1
2u
2
[1 + ln(2&)]
2
=
1
u
[1 + ln(2&) 3ln&]
[1 + ln(2&)]
2
=
1 + (ln2 + ln&) 3ln&
&[1 + ln(2&)]
2
=
1 + ln2
&[1 + ln(2&)]
2
17. j = ln
2 3r 35r
2
i j
0
=
1
2 3r 35r
2
(31 310r) =
310r 31
2 3r 35r
2
or
10r + 1
5r
2
+r 32
19. j = ln(c
3i
+rc
3i
) = ln(c
3i
(1 +r)) = ln(c
3i
) + ln(1 + r) = 3r + ln(1 +r) i
j
0
= 31 +
1
1 +r
=
31 3r + 1
1 +r
= 3
r
1 +r
SECTION 3.6 DERIVATIVES OF LOGARITHMIC FUNCTIONS

111
21. j = 2rlog
10
I
r = 2rlog
10
r
1'2
= 2r
1
2
log
10
r = rlog
10
r i j
0
= r
1
rln10
+ log
10
r 1 =
1
ln10
+ log
10
r
Note:
1
ln10
=
lnc
ln10
= log
10
c, so the answer could be written as
1
ln10
+ log
10
r = log
10
c + log
10
r = log
10
cr.
23. j = r
2
ln(2r) i j
0
= r
2
1
2r
2 + ln(2r) (2r) = r + 2rln(2r) i
j
00
= 1 + 2r
1
2r
2 + ln(2r) 2 = 1 + 2 + 2 ln(2r) = 3 + 2 ln(2r)
25. j = ln
r +
I
1 +r
2
i
j
0
=
1
r +
I
1 +r
2
d
dr
r +
I
1 +r
2
=
1
r +
I
1 +r
2
1 +
1
2
(1 +r
2
)
31'2
(2r)
=
1
r +
I
1 +r
2
1 +
r
I
1 +r
2
=
1
r +
I
1 +r
2

I
1 +r
2
+r
I
1 +r
2
=
1
I
1 +r
2
i
j
00
= 3
1
2
(1 +r
2
)
33'2
(2r) =
3r
(1 +r
2
)
3'2
27. 1(r) =
r
1 3ln(r 31)
i
1
0
(r) =
[1 3ln(r 31)] 1 3r
31
r 31
[1 3ln(r 31)]
2
=
(r 31)[1 3ln(r 31)] +r
r 31
[1 3ln(r 31)]
2
=
r 31 3(r 31) ln(r 31) +r
(r 31)[1 3ln(r 31)]
2
=
2r 31 3(r 31) ln(r 31)
(r 31)[1 3ln(r 31)]
2
Dom(1) = {r | r 31 0 and 1 3ln(r 31) 6= 0} = {r | r 1 and ln(r 31) 6= 1}
=

r | r 1 and r 31 6= c
1
= {r | r 1 and r 6= 1 +c} = (1. 1 +c) (1 +c. ")

29. 1(r) = ln(r
2
32r) i 1
0
(r) =
1
r
2
32r
(2r 32) =
2(r 31)
r(r 32)
.
Dom(1) = {r | r(r 32) 0} = (3". 0) (2. ").
31. 1(r) =
lnr
r
2
i 1
0
(r) =
r
2
(1r) 3(lnr)(2r)
(r
2
)
2
=
r 32rlnr
r
4
=
r(1 32 lnr)
r
4
=
1 32 lnr
r
3
,
so 1
0
(1) =
1 32 ln1
1
3
=
1 32 0
1
= 1.
33. j = ln
rc
i
2
= lnr + lnc
i
2
= lnr +r
2
i j
0
=
1
r
+ 2r. At (1. 1), the slope of the tangent line is
j
0
(1) = 1 + 2 = 3, and an equation of the tangent line is j 31 = 3(r 31), or j = 3r 32.
35. 1(r) = sinr + lnr i 1
0
(r) = cos r + 1r.
This is reasonable, because the graph shows that 1 increases when 1
0
is
positive, and 1
0
(r) = 0 when 1 has a horizontal tangent.
112

37. j = (2r + 1)
5
(r
4
33)
6
i lnj = ln
(2r + 1)
5
(r
4
33)
6
i lnj = 5 ln(2r + 1) + 6 ln(r

4
33) i
1
j
j
0
= 5
1
2r + 1
2 + 6
1
r
4
33
4r
3
i
j
0
= j
10
2r + 1
+
24r
3
r
4
33
= (2r + 1)
5
(r
4
33)
6
10
2r + 1
+
24r
3
r
4
33
.
[The answer could be simplied to j
0
= 2(2r + 1)
4
(r
4
33)
5
(29r
4
+ 12r
3
315), but this is unnecessary.]
39. j =
sin
2
r tan
4
r
(r
2
+ 1)
2
i lnj = ln(sin
2
r tan
4
r) 3 ln(r
2
+ 1)
2
i
lnj = ln(sinr)
2
+ ln(tanr)
4
3ln(r
2
+ 1)
2
i lnj = 2 ln|sinr| + 4 ln|tanr| 32 ln(r
2
+ 1) i
1
j
j
0
= 2
1
sinr
cos r + 4
1
tanr
sec
2
r 32
1
r
2
+ 1
2r i j
0
=
sin
2
r tan
4
r
(r
2
+ 1)
2
2 cot r +
4 sec
2
r
tanr
3
4r
r
2
+ 1
41. j = r
i
i lnj = lnr
i
i lnj = r lnr i j
0
j = r(1r) + (lnr) 1 i j
0
= j(1 + lnr) i
j
0
= r
i
(1 + lnr)
43. j = r
sin i
i lnj = lnr
sin i
i lnj = sinrlnr i
j
0
j
= (sinr)
1
r
+ (lnr)(cos r) i
j
0
= j
sinr
r
+ lnrcos r
i j
0
= r
sin i
sinr
r
+ lnr cos r
45. j = (cos r)
i
i lnj = ln(cos r)
i
i lnj = rlncos r i
1
j
j
0
= r
1
cos r
(3sinr) + lncos r 1 i
j
0
= j
lncos r 3
rsinr
cos r
i j
0
= (cos r)
i
(lncos r 3rtanr)
47. j = (tanr)
1'i
i lnj = ln(tanr)
1'i
i lnj =
1
r
lntanr i
1
j
j
0
=
1
r

1
tanr
sec
2
r + lntanr
3
1
r
2
i j
0
= j
sec
2
r
rtanr
3
lntanr
r
2
i
j
0
= (tanr)
1'i
sec
2
r
rtanr
3
lntanr
r
2
or j
0
= (tanr)
1'i
1
r
csc rsec r 3
lntanr
r
49. j = ln(r
2
+j
2
) i j
0
=
1
r
2
+j
2
d
dr
(r
2
+j
2
) i j
0
=
2r + 2jj
0
r
2
+j
2
i r
2
j
0
+j
2
j
0
= 2r + 2jj
0
i
r
2
j
0
+j
2
j
0
32jj
0
= 2r i (r
2
+j
2
32j)j
0
= 2r i j
0
=
2r
r
2
+j
2
32j
51. 1(r) = ln(r 31) i 1
0
(r) =
1
(r 31)
= (r 31)
31
i 1
00
(r) = 3(r 31)
32
i 1
000
(r) = 2(r 31)
33
i
1
(4)
(r) = 32 3(r 31)
34
i i 1
(n)
(r) = (31)
n31
2 3 4 (n 31)(r 31)
3n
= (31)
n31
(n 31)!
(r 31)
n
53. If 1(r) = ln(1 +r), then 1
0
(r) =
1
1 +r
, so 1
0
(0) = 1.
Thus, lim
i<0
ln(1 +r)
r
= lim
i<0
1(r)
r
= lim
i<0
1(r) 31(0)
r 30
= 1
0
(0) = 1.
SECTION 3.7 RATES OF CHANGE IN THE NATURAL AND SOCIAL SCIENCES

113
3.7 Rates of Change in the Natural and Social Sciences
1. (a) c = 1(t) = t
3
312t
2
+ 36t i (t) = 1
0
(t) = 3t
2
324t + 36
(b) (3) = 27 372 + 36 = 39 fts
(c) The particle is at rest when (t) = 0. 3t
2
324t + 36 = 0 C 3(t 32)(t 36) = 0 C t = 2 s or 6 s.
(d) The particle is moving in the positive direction when (t) 0. 3(t 32)(t 36) 0 C 0 $ t < 2 or t 6.
(e) Since the particle is moving in the positive direction and in the
negative direction, we need to calculate the distance traveled in the
intervals [0. 2], [2. 6], and [6. 8] separately.
|1(2) 31(0)| = |32 30| = 32.
|1(6) 31(2)| = |0 332| = 32.
|1(8) 31(6)| = |32 30| = 32.
The total distance is 32 + 32 + 32 = 96 ft.
(f )
(g) (t) = 3t
2
324t + 36 i
o(t) =
0
(t) = 6t 324.
o(3) = 6(3) 324 = 36 (fts)s or fts
2
.
(h )
(i) The particle is speeding up when and o have the same sign. This occurs when 2 < t < 4 [ and o are both negative] and
when t 6 [ and o are both positive]. It is slowing down when and o have opposite signs; that is, when 0 $ t < 2 and
when 4 < t < 6.
3. (a) c = 1(t) = cos(t4) i (t) = 1
0
(t) = 3sin(t4) (4)
(b) (3) = 3
r
4
sin
3r
4
= 3
r
4

I
2
2
= 3
r
I
2
8
fts [E 30.56]
(c) The particle is at rest when (t) = 0. 3
r
4
sin
rI
4
= 0 i sin
rI
4
= 0 i
rI
4
= n i t = 0, 4, 8 s.
(d) The particle is moving in the positive direction when (t) 0. 3
r
4
sin
rI
4
0 i sin
rI
4
< 0 i 4 < t < 8.
(e) From part (c), (t) = 0 for t = 0. 4. 8. As in Exercise 1, well
nd the distance traveled in the intervals [0. 4] and [4. 8].
|1(4) 31(0)| = |31 31| = 2
|1(8) 31(4)| = |1 3(31)| = 2.
The total distance is 2 + 2 = 4 ft.
(f )

(g) (t) = 3
4
sin
t
4
i
o(t) =
0
(t) = 3
4
cos
t
4

4
= 3
2
16
cos
t
4
.
o(3) = 3
2
16
cos
3
4
= 3
2
16
3
I
2
2
=

2
I
2
32
(fts)s or fts
2
.
(h )
114

(i) The particle is speeding up when and o have the same sign. This occurs when 0 < t < 2 or 8 < t < 10 [ and o are
both negative] and when 4 < t < 6 [ and o are both positive]. It is slowing down when and o have opposite signs;
that is, when 2 < t < 4 and when 6 < t < 8.
5. (a) From the gure, the velocity is positive on the interval (0. 2) and negative on the interval (2. 3). The acceleration o is
positive (negative) when the slope of the tangent line is positive (negative), so the acceleration is positive on the interval
(0. 1), and negative on the interval (1. 3). The particle is speeding up when and o have the same sign, that is, on the
interval (0. 1) when 0 and o 0, and on the interval (2. 3) when < 0 and o < 0. The particle is slowing down
when and o have opposite signs, that is, on the interval (1. 2) when 0 and o < 0.
(b) 0 on (0. 3) and < 0 on (3. 4). o 0 on (1. 2) and o < 0 on (0. 1) and (2. 4). The particle is speeding up on (1. 2)
[ 0, o 0] and on (3. 4) [ < 0, o < 0]. The particle is slowing down on (0. 1) and (2. 3) [ 0, o < 0].
7. (a) c(t) = t
3
34.5t
2
37t i (t) = c
0
(t) = 3t
2
39t 37 = 5 C 3t
2
39t 312 = 0 C
3(t 34)(t + 1) = 0 C t = 4 or 31. Since t D 0, the particle reaches a velocity of 5 ms at t = 4 s.
(b) o(t) =
0
(t) = 6t 39 = 0 C t = 1.5. The acceleration changes from negative to positive, so the velocity changes
from decreasing to increasing. Thus, at t = 1.5 s, the velocity has its minimum value.
9. (a) / = 10t 30.83t
2
i (t) =
d/
dt
= 10 31.66t, so (3) = 10 31.66(3) = 5.02 ms.
(b) / = 25 i 10t 30.83t
2
= 25 i 0.83t
2
310t + 25 = 0 i t =
10
I
17
1.66
E 3.54 or 8.51.
The value t
1
=
10 3
I
17
1.66
corresponds to the time it takes for the stone to rise 25 m and t
2
=
10 +
I
17
1.66
corresponds to the
time when the stone is 25 m high on the way down. Thus, (t
1
) = 10 31.66
10 3
I
17
1.66
=
I
17 E 4.12 ms.
11. (a) (r) = r
2
i
0
(r) = 2r.
0
(15) = 30 mm
2
mm is the rate at which
the area is increasing with respect to the side length as r reaches 15 mm.
(b) The perimeter is 1(r) = 4r, so
0
(r) = 2r =
1
2
(4r) =
1
2
1(r). The
gure suggests that if {r is small, then the change in the area of the square
is approximately half of its perimeter (2 of the 4 sides) times {r. From the
gure, { = 2r({r) + ({r)
2
. If {r is small, then { E 2r({r) and
so {{r E 2r.
13. (a) Using (v) = v
2
, we nd that the average rate of change is:
(i)
(3) 3(2)
3 32
=
9 34
1
= 5 (ii)
(2.5) 3(2)
2.5 32
=
6.25 34
0.5
= 4.5
(iii)
(2.1) 3(2)
2.1 32
=
4.41 34
0.1
= 4.1
(b) (v) = v
2
i
0
(v) = 2v, so
0
(2) = 4.
SECTION 3.7 RATES OF CHANGE IN THE NATURAL AND SOCIAL SCIENCES

115
(c) The circumference is C(v) = 2v =
0
(v). The gure suggests that if {v is small,
then the change in the area of the circle (a ring around the outside) is approximately equal
to its circumference times {v. Straightening out this ring gives us a shape that is approx-
imately rectangular with length 2v and width {v, so { E 2v({v). Algebraically,
{ = (v +{v) 3(v) = (v +{v)
2
3v
2
= 2v({v) +({v)
2
.
So we see that if {v is small, then { E 2v({v) and therefore, {{v E 2v.
15. o(v) = 4v
2
i o
0
(v) = 8v i
(a) o
0
(1) = 8 ft
2
ft (b) o
0
(2) = 16 ft
2
ft (c) o
0
(3) = 24 ft
2
ft
As the radius increases, the surface area grows at an increasing rate. In fact, the rate of change is linear with respect to the
radius.
17. The mass is 1(r) = 3r
2
, so the linear density at r is a(r) = 1
0
(r) = 6r.
(a) a(1) = 6 kgm (b) a(2) = 12 kgm (c) a(3) = 18 kgm
Since a is an increasing function, the density will be the highest at the right end of the rod and lowest at the left end.
19. The quantity of charge is Q(t) = t
3
32t
2
+ 6t + 2, so the current is Q
0
(t) = 3t
2
34t + 6.
(a) Q
0
(0.5) = 3(0.5)
2
34(0.5) + 6 = 4.75 A (b) Q
0
(1) = 3(1)
2
34(1) + 6 = 5 A
The current is lowest when Q
0
has a minimum. Q
00
(t) = 6t 34 < 0 when t <
2
3
. So the current decreases when t <
2
3
and
increases when t
2
3
. Thus, the current is lowest at t =
2
3
s.
21. (a) To nd the rate of change of volume with respect to pressure, we rst solve for \ in terms of 1.
1\ = C i \ =
C
1
i
d\
d1
= 3
C
1
2
.
(b) From the formula for d\d1 in part (a), we see that as 1 increases, the absolute value of d\d1 decreases.
Thus, the volume is decreasing more rapidly at the beginning.
(c) o = 3
1
\
d\
d1
= 3
1
\
3
C
1
2
=
C
(1\ )1
=
C
C1
=
1
1
23. In Example 6, the population function was n = 2
I
n
0
. Since we are tripling instead of doubling and the initial population is
400, the population function is n(t) = 400 3
I
. The rate of growth is n
0
(t) = 400 3
I
ln3, so the rate of growth after
2.5 hours is n
0
(2.5) = 400 3
2.5
ln3 E 6850 bacteriahour.
25. (a) 1920: i
1
=
1860 3 1750
1920 3 1910
=
110
10
= 11, i
2
=
2070 3 1860
1930 3 1920
=
210
10
= 21,
(i
1
+i
2
)/ 2 = (11 + 21)2 = 16 millionyear
1980: i
1
=
4450 3 3710
1980 3 1970
=
740
10
= 74, i
2
=
5280 3 4450
1990 3 1980
=
830
10
= 83,
(i
1
+i
2
)/ 2 = (74 + 83)2 = 78.5 millionyear
(b) 1(t) = ot
3
+/t
2
+ct +d (in millions of people), where o E 0.0012937063, / E 37.061421911, c E 12,822.97902,
and d E 37,743,770.396.
(c) 1(t) = ot
3
+/t
2
+ct +d i 1
0
(t) = 3ot
2
+ 2/t +c (in millions of people per year)
116

(d) 1
0
(1920) = 3(0.0012937063)(1920)
2
+ 2(37.061421911)(1920) + 12,822.97902
E 14.48 millionyear [smaller than the answer in part (a), but close to it]
1
0
(1980) E 75.29 millionyear (smaller, but close)
(e) 1
0
(1985) E 81.62 millionyear, so the rate of growth in 1985 was about 81.62 millionyear.
27. (a) Using =
1
4|
(1
2
3v
2
) with 1 = 0.01, | = 3, 1 = 3000, and = 0.027, we have as a function of v:
(v) =
3000
4(0.027)3
(0.01
2
3v
2
). (0) = 0.925 cms, (0.005) = 0.694 cms, (0.01) = 0.
(b) (v) =
1
4|
(1
2
3v
2
) i
0
(v) =
1
4|
(32v) = 3
1v
2|
. When | = 3, 1 = 3000, and = 0.027, we have
0
(v) = 3
3000v
2(0.027)3
.
0
(0) = 0,
0
(0.005) = 392.592 (cms)cm, and
0
(0.01) = 3185.185 (cms)cm.
(c) The velocity is greatest where v = 0 (at the center) and the velocity is changing most where v = 1 = 0.01 cm
(at the edge).
29. (a) C(r) = 1200 + 12r 30.1r
2
+ 0.0005r
3
i C
0
(r) = 12 30.2r + 0.0015r
2
$yard, which is the marginal cost
function.
(b) C
0
(200) = 12 30.2(200) + 0.0015(200)
2
= $32yard, and this is the rate at which costs are increasing with respect to
the production level when r = 200. C
0
(200) predicts the cost of producing the 201st yard.
(c) The cost of manufacturing the 201st yard of fabric is C(201) 3C(200) = 3632.2005 33600 E $32.20, which is
approximately C
0
(200).
31. (a) (r) =
j(r)
r
i
0
(r) =
rj
0
(r) 3j(r) 1
r
2
=
rj
0
(r) 3j(r)
r
2
.
0
(r) 0 i (r) is increasing; that is, the average productivity increases as the size of the workforce increases.
(b) j
0
(r) is greater than the average productivity i j
0
(r) (r) i j
0
(r)
j(r)
r
i rj
0
(r) j(r) i
rj
0
(r) 3j(r) 0 i
rj
0
(r) 3j(r)
r
2
0 i
0
(r) 0.
33. 1\ = n1T i T =
1\
n1
=
1\
(10)(0.0821)
=
1
0.821
(1\ ). Using the Product Rule, we have
dT
dt
=
1
0.821
[1(t)\
0
(t) +\ (t)1
0
(t)] =
1
0.821
[(8)(30.15) + (10)(0.10)] E 30.2436 Kmin.
35. (a) If the populations are stable, then the growth rates are neither positive nor negative; that is,
dC
dt
= 0 and
dW
dt
= 0.
(b) The caribou go extinct means that the population is zero, or mathematically, C = 0.
(c) We have the equations
dC
dt
= oC 3/CW and
dW
dt
= 3cW +dCW. Let dCdt = dWdt = 0, o = 0.05, / = 0.001,
c = 0.05, and d = 0.0001 to obtain 0.05C 30.001CW = 0 (1) and 30.05W + 0.0001CW = 0 (2). Adding 10 times
(2) to (1) eliminates the CW-terms and gives us 0.05C 30.5W = 0 i C = 10W. Substituting C = 10W into (1)
SECTION 3.8 EXPONENTIAL GROWTH AND DECAY

117
results in 0.05(10W) 30.001(10W)W = 0 C 0.5W 30.01W
2
= 0 C 50W 3W
2
= 0 C
W(50 3W) = 0 C W = 0 or 50. Since C = 10W, C = 0 or 500. Thus, the population pairs (C. W) that lead to
stable populations are (0. 0) and (500. 50). So it is possible for the two species to live in harmony.
3.8 Exponential Growth and Decay
1. The relative growth rate is
1
1
d1
dt
= 0.7944, so
d1
dt
= 0.79441 and, by Theorem 2, 1(t) = 1(0)c
0.7944I
= 2c
0.7944I
.
Thus, 1(6) = 2c
0.7944(6)
E 234.99 or about 235 members.
3. (a) By Theorem 2, 1(t) = 1(0)c
II
= 100c
II
. Now 1(1) = 100c
I(1)
= 420 i c
I
=
420
100
i I = ln4.2.
So 1(t) = 100c
(ln 4.2)I
= 100(4.2)
I
.
(b) 1(3) = 100(4.2)
3
= 7408.8 E 7409 bacteria
(c) d1dt = I1 i 1
0
(3) = I 1(3) = (ln4.2)
100(4.2)
3
[from part (a)] E 10,632 bacteriahour

(d) 1(t) = 100(4.2)
I
= 10,000 i (4.2)
I
= 100 i t = (ln100)(ln4.2) E 3.2 hours
5. (a) Let the population (in millions) in the year t be 1(t). Since the initial time is the year 1750, we substitute t 31750 for t in
Theorem 2, so the exponential model gives 1(t) = 1(1750)c
I(I31750)
. Then 1(1800) = 980 = 790c
I(180031750)
i
980
790
= c
I(50)
i ln
980
790
= 50I i I =
1
50
ln
980
790
E 0.0043104. So with this model, we have
1(1900) = 790c
I(190031750)
E 1508 million, and 1(1950) = 790c
I(195031750)
E 1871 million. Both of these
estimates are much too low.
(b) In this case, the exponential model gives 1(t) = 1(1850)c
I(I31850)
i 1(1900) = 1650 = 1260c
I(190031850)
i
ln
1650
1260
= I(50) i I =
1
50
ln
1650
1260
E 0.005393. So with this model, we estimate
1(1950) = 1260c
I(195031850)
E 2161 million. This is still too low, but closer than the estimate of 1(1950) in part (a).
(c) The exponential model gives 1(t) = 1(1900)c
I(I31900)
i 1(1950) = 2560 = 1650c
I(195031900)
i
ln
2560
1650
= I(50) i I =
1
50
ln
2560
1650
E 0.008785. With this model, we estimate
1(2000) = 1650c
I(200031900)
E 3972 million. This is much too low. The discrepancy is explained by the fact that the
world birth rate (average yearly number of births per person) is about the same as always, whereas the mortality rate
(especially the infant mortality rate) is much lower, owing mostly to advances in medical science and to the wars in the rst
part of the twentieth century. The exponential model assumes, among other things, that the birth and mortality rates will
remain constant.
7. (a) If j = [N
2
O
5
] then by Theorem 2,
dj
dt
= 30.0005j i j(t) = j(0)c
30.0005I
= Cc
30.0005I
.
(b) j(t) = Cc
30.0005I
= 0.9C i c
30.0005I
= 0.9 i 30.0005t = ln0.9 i t = 32000 ln0.9 E 211 s
118

9. (a) If j(t) is the mass (in mg) remaining after t years, then j(t) = j(0)c
II
= 100c
II
.
j(30) = 100c
30I
=
1
2
(100) i c
30I
=
1
2
i I = 3(ln2)30 i j(t) = 100c
3(ln 2)I'30
= 100 2
3I'30
(b) j(100) = 100 2
3100'30
E 9.92 mg
(c) 100c
3(ln 2)I'30
= 1 i 3(ln2)t30 = ln
1
100
i t = 330
ln 0.01
ln 2
E 199.3 years
11. Let j(t) be the level of radioactivity. Thus, j(t) = j(0)c
3II
and I is determined by using the half-life:
j(5730) =
1
2
j(0) i j(0)c
3I(5730)
=
1
2
j(0) i c
35730I
=
1
2
i 35730I = ln
1
2
i I = 3
ln
1
2
5730
=
ln2
5730
.
If 74% of the
14
C remains, then we know that j(t) = 0.74j(0) i 0.74 = c
3I(ln 2)'5730
i ln0.74 = 3
t ln2
5730
i
t = 3
5730(ln0.74)
ln2
E 2489 E 2500 years.
13. (a) Using Newtons Law of Cooling,
dT
dt
= I(T 3T
s
), we have
dT
dt
= I(T 375). Now let j = T 375, so
j(0) = T(0) 375 = 185 375 = 110, so j is a solution of the initial-value problemdjdt = Ij with j(0) = 110 and by
Theorem 2 we have j(t) = j(0)c
II
= 110c
II
.
j(30) = 110c
30I
= 150 375 i c
30I
=
75
110
=
15
22
i I =
1
30
ln
15
22
, so j(t) = 110c
1
30
I ln(
15
22
)
and
j(45) = 110c
45
30
ln(
15
22
)
E 62
F. Thus, T(45) E 62 + 75 = 137
F.
(b) T(t) = 100 i j(t) = 25. j(t) = 110c
1
30
I ln(
15
22
)
= 25 i c
1
30
I ln(
15
22
)
=
25
110
i
1
30
t ln
15
22
= ln
25
110
i
t =
30 ln
25
110
ln
15
22
E 116 min.
15.
dT
dt
= I(T 320). Letting j = T 320, we get
dj
dt
= Ij, so j(t) = j(0)c
II
. j(0) = T(0) 320 = 5 320 = 315, so
j(25) = j(0)c
25I
= 315c
25I
, and j(25) = T(25) 320 = 10 320 = 310, so 315c
25I
= 310 i c
25I
=
2
3
. Thus,
25I = ln
2
3
and I =
1
25
ln
2
3
, so j(t) = j(0)c
II
= 315c
(1'25) ln(2'3)I
. More simply, c
25I
=
2
3
i c
I
=

2
3
1'25
i
c
II
=

2
3
I'25
i j(t) = 315
2
3
I'25
.
(a) T(50) = 20 +j(50) = 20 315
2
3
50'25
= 20 315
2
3
2
= 20 3
20
3
= 13.
C
(b) 15 = T(t) = 20 +j(t) = 20 315
2
3
I'25
i 15
2
3
I'25
= 5 i

2
3
I'25
=
1
3
i
(t25) ln
2
3
= ln
1
3
i t = 25 ln
1
3
ln
2
3
E 67.74 min.
17. (a) Let 1(/) be the pressure at altitude /. Then d1d/ = I1 i 1(/) = 1(0)c
II
= 101.3c
II
.
1(1000) = 101.3c
1000I
= 87.14 i 1000I = ln
87.14
101.3
i I =
1
1000
ln
87.14
101.3
i
1(/) = 101.3 c
1
1000
Iln(
87=14
101=3
)
, so 1(3000) = 101.3c
3 ln(
87=14
101=3
)
E 64.5 kPa.
(b) 1(6187) = 101.3 c
6187
1000
ln(
87=14
101=3
)
E 39.9 kPa
SECTION 3.9 RELATED RATES

119
19. (a) Using =
0
1 +
v
n
nI
with
0
= 3000, v = 0.05, and t = 5, we have:
(i) Annually: n = 1; = 3000
1 +
0.05
1
15
= $3828.84
(ii) Semiannually: n = 2; = 3000
1 +
0.05
2
25
= $3840.25
(iii) Monthly: n = 12; = 3000
1 +
0.05
12
125
= $3850.08
(iv) Weekly: n = 52; = 3000
1 +
0.05
52
525
= $3851.61
(v) Daily: n = 365; = 3000
1 +
0.05
365
3655
= $3852.01
(vi) Continuously: = 3000c
(0.05)5
= $3852.08
(b) ddt = 0.05 and (0) = 3000.
3.9 Related Rates
1. \ = r
3
i
d\
dt
=
d\
dr
dr
dt
= 3r
2
dr
dt
3. Let c denote the side of a square. The squares area is given by = c
2
. Differentiating with respect to t gives us
d
dt
= 2c
dc
dt
. When = 16, c = 4. Substitution 4 for c and 6 for
dc
dt
gives us
d
dt
= 2(4)(6) = 48 cm
2
s.
5. \ = v
2
/ = (5)
2
/ = 25/ i
d\
dt
= 25
d/
dt
i 3 = 25
d/
dt
i
d/
dt
=
3
25
mmin.
7. j = r
3
+ 2r i
dj
dt
=
dj
dr
dr
dt
= (3r
2
+ 2)(5) = 5(3r
2
+ 2). When r = 2,
dj
dt
= 5(14) = 70.
9. :
2
= r
2
+j
2
i 2:
d:
dt
= 2r
dr
dt
+ 2j
dj
dt
i
d:
dt
=
1
:
r
dr
dt
+j
dj
dt
. When r = 5 and j = 12,

:
2
= 5
2
+ 12
2
i :
2
= 169 i : = 13. For
dr
dt
= 2 and
dj
dt
= 3.
d:
dt
=
1
13
(5 2 + 12 3) =
46
13
.
11. (a) Given: a plane ying horizontally at an altitude of 1 mi and a speed of 500 mih passes directly over a radar station.
If we let t be time (in hours) and r be the horizontal distance traveled by the plane (in mi), then we are given
that drdt = 500 mih.
(b) Unknown: the rate at which the distance from the plane to the station is increasing
when it is 2 mi from the station. If we let j be the distance from the plane to the station,
then we want to nd djdt when j = 2 mi.
(c)
(d) By the Pythagorean Theorem, j
2
= r
2
+ 1 i 2j (djdt) = 2r(drdt).
(e)
dj
dt
=
r
j
dr
dt
=
r
j
(500). Since j
2
= r
2
+ 1, when j = 2, r =
I
3, so
dj
dt
=
I
3
2
(500) = 250
I
3 E 433 mih.
13. (a) Given: a man 6 ft tall walks away from a street light mounted on a 15-ft-tall pole at a rate of 5 fts. If we let t be time (in s)
and r be the distance from the pole to the man (in ft), then we are given that drdt = 5 fts.
120

(b) Unknown: the rate at which the tip of his shadow is moving when he is 40 ft
from the pole. If we let j be the distance from the man to the tip of his
shadow (in ft), then we want to nd
d
dt
(r +j) when r = 40 ft.
(c)
(d) By similar triangles,
15
6
=
r +j
j
i 15j = 6r + 6j i 9j = 6r i j =
2
3
r.
(e) The tip of the shadow moves at a rate of
d
dt
(r +j) =
d
dt
r +
2
3
r
=
5
3
dr
dt
=
5
3
(5) =
25
3
fts.
15. We are given that
dr
dt
= 60 mih and
dj
dt
= 25 mih. :
2
= r
2
+j
2
i
2:
d:
dt
= 2r
dr
dt
+ 2j
dj
dt
i :
d:
dt
= r
dr
dt
+j
dj
dt
i
d:
dt
=
1
:
r
dr
dt
+j
dj
dt
.
After 2 hours, r = 2 (60) = 120 and j = 2 (25) = 50 i : =
I
120
2
+ 50
2
= 130,
so
d:
dt
=
1
:
r
dr
dt
+j
dj
dt
=
120(60) + 50(25)
130
= 65 mih.
dr
dt
= 4 fts and
dj
dt
= 5 fts. :
2
= (r +j)
2
+ 500
2
i
2:
d:
dt
= 2(r +j)
dr
dt
+
dj
dt
. 15 minutes after the woman starts, we have

r = (4 fts)(20 min)(60 smin) = 4800 ft and j = 5 15 60 = 4500 i
: =

(4800 + 4500)
2
+ 500
2
=
I
86,740,000, so
d:
dt
=
r +j
:
dr
dt
+
dj
dt
=
4800 + 4500
I
86,740,000
(4 + 5) =
837
I
8674
E 8.99 fts.
19. =
1
2
//, where / is the base and / is the altitude. We are given that
d/
dt
= 1 cmmin and
d
dt
= 2 cm
2
min. Using the
Product Rule, we have
d
dt
=
1
2
/
d/
dt
+/
d/
dt
. When / = 10 and = 100, we have 100 =

1
2
/(10) i
1
2
/ = 10 i
/ = 20, so 2 =
1
2
20 1 + 10
d/
dt
i 4 = 20 + 10
d/
dt
i
d/
dt
=
4 320
10
= 31.6 cmmin.
dr
dt
= 35 kmh and
dj
dt
= 25 kmh. :
2
= (r +j)
2
+ 100
2
i
2:
d:
dt
= 2(r +j)
dr
dt
+
dj
dt
. At 4:00 PM, r = 4(35) = 140 and j = 4(25) = 100 i

: =

(140 + 100)
2
+ 100
2
=
I
67,600 = 260, so
d:
dt
=
r +j
:
dr
dt
+
dj
dt
=
140 + 100
260
(35 + 25) =
720
13
E 55.4 kmh.
SECTION 3.9 RELATED RATES

121
23. If C = the rate at which water is pumped in, then
d\
dt
= C 310,000, where
\ =
1
3
v
2
/ is the volume at time t. By similar triangles,
v
2
=
/
6
i v =
1
3
/ i
\ =
1
3
1
3
/
2
/ =
r
27
/
3
i
d\
dt
=

9
/
2
d/
dt
. When / = 200 cm,
d/
dt
= 20 cmmin, so C 310,000 =

9
(200)
2
(20) i C = 10,000 +
800,000
9
E 289,253 cm
3
min.
25. The gure is labeled in meters. The area of a trapezoid is
1
2
(base
1
+base
2
)(height), and the volume \ of the 10-meter-long trough is 10.
Thus, the volume of the trapezoid with height / is \ = (10)
1
2
[0.3 + (0.3 + 2o)]/.
By similar triangles,
o
/
=
0.25
0.5
=
1
2
, so 2o = / i \ = 5(0.6 +/)/ = 3/ + 5/
2
.
Now
d\
dt
=
d\
d/
d/
dt
i 0.2 = (3 + 10/)
d/
dt
i
d/
dt
=
0.2
3 + 10/
. When / = 0.3,
d/
dt
=
0.2
3 + 10(0.3)
=
0.2
6
mmin =
1
30
mmin or
10
3
cmmin.
d\
dt
= 30 ft
3
min. \ =
1
3
v
2
/ =
1
3
/
2
2
/ =
/
3
12
i
d\
dt
=
d\
d/
d/
dt
i 30 =
/
2
4
d/
dt
i
d/
dt
=
120
/
2
.
When / = 10 ft,
d/
dt
=
120
10
2
=
6
5
E 0.38 ftmin.
29. =
1
2
//, but / = 5 m and sin0 =
/
4
i / = 4 sin0, so =
1
2
(5)(4 sin0) = 10 sin0.
We are given
d0
dt
= 0.06 rads, so
d
dt
=
d
d0
d0
dt
= (10 cos 0)(0.06) = 0.6 cos 0.
When 0 =

3
,
d
dt
= 0.6
cos

3
= (0.6)
1
2
= 0.3 m
2
s.
31. Differentiating both sides of 1\ = C with respect to t and using the Product Rule gives us 1
d\
dt
+\
d1
dt
= 0 i
d\
dt
= 3
\
1
d1
dt
. When \ = 600, 1 = 150 and
d1
dt
= 20, so we have
d\
dt
= 3
600
150
(20) = 380. Thus, the volume is
decreasing at a rate of 80 cm
3
min.
33. With 1
1
= 80 and 1
2
= 100,
1
1
=
1
1
1
+
1
1
2
=
1
80
+
1
100
=
180
8000
=
9
400
, so 1 =
400
9
. Differentiating
1
1
=
1
1
1
+
1
1
2
with respect to t, we have 3
1
1
2
d1
dt
= 3
1
1
2
1
d1
1
dt
3
1
1
2
2
d1
2
dt
i
d1
dt
= 1
2
1
1
2
1
d1
1
dt
+
1
1
2
2
d1
2
dt
. When 1
1
= 80 and
1
2
= 100,
d1
dt
=
400
2
9
2
1
80
2
(0.3) +
1
100
2
(0.2)
=
107
810
E 0.132 ls.
122

35. We are given d0dt = 2
min =
r
90
radmin. By the Law of Cosines,
r
2
= 12
2
+ 15
2
32(12)(15) cos 0 = 369 3360 cos 0 i
2r
dr
dt
= 360 sin0
d0
dt
i
dr
dt
=
180 sin0
r
d0
dt
. When 0 = 60
,
r =
I
369 3360 cos 60
=
I
189 = 3
I
21, so
dr
dt
=
180 sin60
3
I
21
90
=

I
3
3
I
21
=
I
7
21
E 0.396 mmin.
37. (a) By the Pythagorean Theorem, 4000
2
+j
2
= /
2
. Differentiating with respect to t,
we obtain 2j
dj
dt
= 2/
d/
dt
. We know that
dj
dt
= 600 fts, so when j = 3000 ft,
/ =
I
4000
2
+ 3000
2
=
I
25,000,000 = 5000 ft
and
d/
dt
=
j
/
dj
dt
=
3000
5000
(600) =
1800
5
= 360 fts.
(b) Here tan0 =
j
4000
i
d
dt
(tan0) =
d
dt
j
4000
i sec
2
0
d0
dt
=
1
4000
dj
dt
i
d0
dt
=
cos
2
0
4000
dj
dt
. When
j = 3000 ft,
dj
dt
= 600 fts, / = 5000 and cos 0 =
4000
/
=
4000
5000
=
4
5
, so
d0
dt
=
(45)
2
4000
(600) = 0.096 rads.
39. cot 0 =
r
5
i 3csc
2
0
d0
dt
=
1
5
dr
dt
i 3
csc

3
=
1
5
dr
dt
i
dr
dt
=
5
6
2
I
3
2
=
10
9
kmmin [E 130 mih]
dr
dt
= 300 kmh. By the Law of Cosines,
j
2
= r
2
+ 1
2
32(1)(r) cos 120
= r
2
+ 1 32r
3
1
2
= r
2
+r + 1, so
2j
dj
dt
= 2r
dr
dt
+
dr
dt
i
dj
dt
=
2r + 1
2j
dr
dt
. After 1 minute, r =
300
60
= 5 km i
j =
I
5
2
+ 5 + 1 =
I
31 km i
dj
dt
=
2(5) + 1
2
I
31
(300) =
1650
I
31
E 296 kmh.
43. Let the distance between the runner and the friend be /. Then by the Law of Cosines,
/
2
= 200
2
+100
2
32 200 100 cos 0 = 50,000 340,000 cos 0 (-). Differentiating
implicitly with respect to t, we obtain 2/
d/
dt
= 340,000(3sin0)
d0
dt
. Now if 1 is the
distance run when the angle is 0 radians, then by the formula for the length of an arc
on a circle, c = v0, we have 1 = 1000, so 0 =
1
100
1 i
d0
dt
=
1
100
d1
dt
=
7
100
. To substitute into the expression for
d/
dt
, we must know sin0 at the time when / = 200, which we nd from (-): 200
2
= 50,000 340,000 cos 0 C
cos 0 =
1
4
i sin0 =

1 3
1
4
2
=
I
15
4
. Substituting, we get 2(200)
d/
dt
= 40,000
I
15
4
7
100
i
d/dt =
7
I
15
4
E 6.78 ms. Whether the distance between them is increasing or decreasing depends on the direction in which
the runner is running.
SECTION 3.10 LINEAR APPROXIMATIONS AND DIFFERENTIALS

123
3.10 Linear Approximations and Differentials
1. 1(r) = r
4
+ 3r
2
i 1
0
(r) = 4r
3
+ 6r, so 1(31) = 4 and 1
0
(31) = 310.
Thus, 1(r) = 1(31) +1
0
(31)(r 3(31)) = 4 + (310)(r + 1) = 310r 36.
3. 1(r) = cos r i 1
0
(r) = 3sinr, so 1
r
2
= 0 and 1
0
r
2
= 31.
Thus, 1(r) = 1
r
2
+1
0
r
2
r 3
r
2
= 0 31
r 3
r
2
= 3r +
r
2
.
5. 1(r) =
I
1 3r i 1
0
(r) =
31
2
I
1 3r
, so 1(0) = 1 and 1
0
(0) = 3
1
2
.
Therefore,
I
1 3r = 1(r) E 1(0) +1
0
(0)(r 30) = 1 +
3
1
2
(r 30) = 1 3
1
2
r.
So
I
0.9 =
I
1 30.1 E 1 3
1
2
(0.1) = 0.95
and
I
0.99 =
I
1 30.01 E 1 3
1
2
(0.01) = 0.995.
7. 1(r) =
3
I
1 3r = (1 3r)
1'3
i 1
0
(r) = 3
1
3
(1 3r)
32'3
, so 1(0) = 1
and 1
0
(0) = 3
1
3
. Thus, 1(r) E 1(0) +1
0
(0)(r 30) = 1 3
1
3
r. We need
3
I
1 3r 30.1 < 1 3
1
3
r <
3
I
1 3r + 0.1, which is true when
31.204 < r < 0.706.
9. 1(r) =
1
(1 + 2r)
4
= (1 + 2r)
34
i
1
0
(r) = 34(1 + 2r)
35
(2) =
38
(1 + 2r)
5
, so 1(0) = 1 and 1
0
(0) = 38.
Thus, 1(r) E 1(0) +1
0
(0)(r 30) = 1 + (38)(r 30) = 1 38r.
We need
1
(1 + 2r)
4
30.1 < 1 38r <
1
(1 + 2r)
4
+ 0.1, which is true
when 30.045 < r < 0.055.
11. (a) The differential dj is dened in terms of dr by the equation dj = 1
0
(r) dr. For j = 1(r) = r
2
sin2r,
1
0
(r) = r
2
cos 2r 2 + sin2r 2r = 2r(rcos 2r + sin2r), so dj = 2r(rcos 2r + sin2r) dr.
(b) For j = 1(t) = ln
I
1 +t
2
=
1
2
ln(1 +t
2
), 1
0
(t) =
1
2

1
1 +t
2
2t =
t
1 +t
2
, so dj =
t
1 +t
2
dt.
13. (a) For j = 1(&) =
& + 1
& 31
, 1
0
(&) =
(& 31)(1) 3(& + 1)(1)
(& 31)
2
=
32
(& 31)
2
, so dj =
32
(& 31)
2
d&.
(b) For j = 1(v) = (1 +v
3
)
32
, 1
0
(v) = 32(1 +v
3
)
33
(3v
2
) =
36v
2
(1 +v
3
)
3
, so dj =
36v
2
(1 +v
3
)
3
dv.
15. (a) j = c
i'10
i dj = c
i'10
1
10
dr =
1
10
c
i'10
dr
(b) r = 0 and dr = 0.1 i dj =
1
10
c
0'10
(0.1) = 0.01.
124

17. (a) j = tanr i dj = sec
2
rdr
(b) When r = 4 and dr = 30.1, dj = [sec(4)]
2
(30.1) =
I
2
2
(30.1) = 30.2.
19. j = 1(r) = 2r 3r
2
, r = 2, {r = 30.4 i
{j = 1(1.6) 31(2) = 0.64 30 = 0.64
dj = (2 32r) dr = (2 34)(30.4) = 0.8
21. j = 1(r) = 2r, r = 4, {r = 1 i
{j = 1(5) 31(4) =
2
5
3
2
4
= 30.1
dj = 3
2
r
2
dr = 3
2
4
2
(1) = 30.125
23. To estimate (2.001)
5
, well nd the linearization of 1(r) = r
5
at o = 2. Since 1
0
(r) = 5r
4
, 1(2) = 32, and 1
0
(2) = 80,
we have 1(r) = 32 + 80(r 32) = 80r 3128. Thus, r
5
E 80r 3128 when r is near 2 , so
(2.001)
5
E 80(2.001) 3128 = 160.08 3128 = 32.08.
25. To estimate (8.06)
2'3
, well nd the linearization of 1(r) = r
2'3
at o = 8. Since 1
0
(r) =
2
3
r
31'3
= 2
3
3
I
r
,
1(8) = 4, and 1
0
(8) =
1
3
, we have 1(r) = 4 +
1
3
(r 38) =
1
3
r +
4
3
. Thus, r
2'3
E
1
3
r +
4
3
when r is near 8, so
(8.06)
2'3
E
1
3
(8.06) +
4
3
=
12.06
3
= 4.02.
27. j = 1(r) = tanr i dj = sec
2
rdr. When r = 45
and dr = 31
,
dj = sec
2
45
(3180) =
I
2
2
(3180) = 390, so tan44
= 1(44
) E 1(45
) +dj = 1 390 E 0.965.

29. j = 1(r) = sec r i 1
0
(r) = sec r tanr, so 1(0) = 1 and 1
0
(0) = 1 0 = 0. The linear approximation of 1 at 0 is
1(0) +1
0
(0)(r 30) = 1 + 0(r) = 1. Since 0.08 is close to 0, approximating sec 0.08 with 1 is reasonable.
31. j = 1(r) = lnr i 1
0
(r) = 1r, so 1(1) = 0 and 1
0
(1) = 1. The linear approximation of 1 at 1 is
1(1) +1
0
(1)(r 31) = 0 + 1(r 31) = r 31. Now 1(1.05) = ln1.05 E 1.05 31 = 0.05, so the approximation
is reasonable.
33. (a) If r is the edge length, then \ = r
3
i d\ = 3r
2
dr. When r = 30 and dr = 0.1, d\ = 3(30)
2
(0.1) = 270, so the
maximum possible error in computing the volume of the cube is about 270 cm
3
. The relative error is calculated by dividing
the change in \ , {\ , by \ . We approximate {\ with d\ .
Relative error =
{\
\
E
d\
\
=
3r
2
dr
r
3
= 3
dr
r
= 3
0.1
30
= 0.01.
Percentage error = relative error 100% = 0.01 100% = 1%.
SECTION 3.10 LINEAR APPROXIMATIONS AND DIFFERENTIALS

125
(b) o = 6r
2
i do = 12rdr. When r = 30 and dr = 0.1, do = 12(30)(0.1) = 36, so the maximum possible error in
computing the surface area of the cube is about 36 cm
2
.
Relative error =
{o
o
E
do
o
=
12rdr
6r
2
= 2
dr
r
= 2
0.1
30
= 0.006.
Percentage error = relative error 100% = 0.006 100% = 0.6%.
35. (a) For a sphere of radius v, the circumference is C = 2v and the surface area is o = 4v
2
, so
v =
C
2
i o = 4
C
2
2
=
C
2
i do =
2
C dC. When C = 84 and dC = 0.5, do =

2
(84)(0.5) =
84
,
so the maximum error is about
84
E 27 cm
2
. Relative error E
do
o
=
84
84
2
=
1
84
E 0.012
(b) \ =
4
3
v
3
=
4
3
C
2
3
=
C
3
6
2
i d\ =
1
2
2
C
2
dC. When C = 84 and dC = 0.5,
d\ =
1
2
2
(84)
2
(0.5) =
1764
2
, so the maximum error is about
1764
2
E 179 cm
3
.
The relative error is approximately
d\
\
=
1764
2
(84)
3
(6
2
)
=
1
56
E 0.018.
37. (a) \ = v
2
/ i {\ E d\ = 2v/dv = 2v/{v
(b) The error is
{\ 3d\ = [(v +{v)
2
/ 3v
2
/] 32v/{v = v
2
/ + 2v/{v +({v)
2
/ 3v
2
/ 32v/{v = ({v)
2
/.
39. \ = 11 i 1 =
\
1
i d1 = 3
\
1
2
d1. The relative error in calculating 1 is
{1
1
E
d1
1
=
3(\1
2
) d1
\1
= 3
d1
1
.
Hence, the relative error in calculating 1 is approximately the same (in magnitude) as the relative error in 1.
41. (a) dc =
dc
dr
dr = 0 dr = 0 (b) d(c&) =
d
dr
(c&) dr = c
d&
dr
dr = c d&
(c) d(& +) =
d
dr
(& +) dr =

d&
dr
+
d
dr
dr =
d&
dr
dr +
d
dr
dr = d& +d
(d) d(&) =
d
dr
(&) dr =

&
d
dr
+
d&
dr
dr = &
d
dr
dr +
d&
dr
dr = &d + d&
(e) d
&
=
d
dr
&
dr =

d&
dr
3&
d
dr
2
dr =

d&
dr
dr 3&
d
dr
dr
2
=
d& 3&d
2
(f ) d (r
n
) =
d
dr
(r
n
) dr = nr
n31
dr
43. (a) The graph shows that 1
0
(1) = 2, so 1(r) = 1(1) +1
0
(1)(r 31) = 5 + 2(r 31) = 2r + 3.
1(0.9) E 1(0.9) = 4.8 and 1(1.1) E 1(1.1) = 5.2.
(b) From the graph, we see that 1
0
(r) is positive and decreasing. This means that the slopes of the tangent lines are positive,
but the tangents are becoming less steep. So the tangent lines lie above the curve. Thus, the estimates in part (a) are too
large.
126

3.11 Hyperbolic Functions
1. (a) sinh0 =
1
2
(c
0
3c
0
) = 0 (b) cosh0 =
1
2
(c
0
+c
0
) =
1
2
(1 + 1) = 1
3. (a) sinh(ln2) =
c
ln 2
3c
3ln 2
2
=
c
ln 2
3(c
ln 2
)
31
2
=
2 32
31
2
=
2 3
1
2
2
=
3
4
(b) sinh2 =
1
2
(c
2
3c
32
) E 3.62686
5. (a) sech0 =
1
cosh0
=
1
1
= 1 (b) cosh
31
1 = 0 because cosh0 = 1.
7. sinh(3r) =
1
2
[c
3i
3c
3(3i)
] =
1
2
(c
3i
3c
i
) = 3
1
2
(c
3i
3c
i
) = 3sinhr
9. coshr + sinhr =
1
2
(c
i
+c
3i
) +
1
2
(c
i
3c
3i
) =
1
2
(2c
i
) = c
i
11. sinhr coshj + coshr sinhj =

1
2
(c
i
3c
3i
)
1
2
(c
u
+c
3u
)
1
2
(c
i
+c
3i
)
1
2
(c
u
3c
3u
)
=
1
4
[(c
i+u
+c
i3u
3c
3i+u
3c
3i3u
) + (c
i+u
3c
i3u
+c
3i+u
3c
3i3u
)]
=
1
4
(2c
i+u
32c
3i3u
) =
1
2
[c
i+u
3c
3(i+u)
] = sinh(r +j)
13. Divide both sides of the identity cosh
2
r 3sinh
2
r = 1 by sinh
2
r:
cosh
2
r
sinh
2
r
3
sinh
2
r
sinh
2
r
=
1
sinh
2
r
C coth
2
r 31 = csch
2
r.
15. Putting j = r in the result from Exercise 11, we have
sinh2r = sinh(r +r) = sinhr coshr + coshr sinhr = 2 sinhr coshr.
17. tanh(lnr) =
sinh(lnr)
cosh(lnr)
=
(c
ln i
3c
3ln i
)2
(c
ln i
+c
3ln i
)2
=
r 3(c
ln i
)
31
r + (c
ln i
)
31
=
r 3r
31
r +r
31
=
r 31r
r + 1r
=
(r
2
31)r
(r
2
+ 1)r
=
r
2
31
r
2
+ 1
19. By Exercise 9, (coshr + sinhr)
n
= (c
i
)
n
= c
ni
= coshnr + sinhnr.
21. sechr =
1
coshr
i sechr =
1
53
=
3
5
.
cosh
2
r 3sinh
2
r = 1 i sinh
2
r = cosh
2
r 31 =
5
3
2
31 =
16
9
i sinhr =
4
3
[because r 0].
cschr =
1
sinhr
i cschr =
1
43
=
3
4
.
tanhr =
sinhr
coshr
i tanhr =
43
53
=
4
5
.
cothr =
1
tanhr
i cothr =
1
45
=
5
4
.
23. (a) lim
i<"
tanhr = lim
i<"
c
i
3c
3i
c
i
+c
3i

c
3i
c
3i
= lim
i<"
1 3c
32i
1 +c
32i
=
1 30
1 + 0
= 1
(b) lim
i<3"
tanhr = lim
i<3"
c
i
3c
3i
c
i
+c
3i

c
i
c
i
= lim
i<3"
c
2i
31
c
2i
+ 1
=
0 31
0 + 1
= 31
SECTION 3.11 HYPERBOLIC FUNCTIONS

127
(c) lim
i<"
sinhr = lim
i<"
c
i
3c
3i
2
= "
(d) lim
i<3"
sinhr = lim
i<3"
c
i
3c
3i
2
= 3"
(e) lim
i<"
sechr = lim
i<"
2
c
i
+c
3i
= 0
(f ) lim
i<"
cothr = lim
i<"
c
i
+c
3i
c
i
3c
3i

c
3i
c
3i
= lim
i<"
1 +c
32i
1 3c
32i
=
1 + 0
1 30
= 1 [Or: Use part (a)]
(g) lim
i<0
+
cothr = lim
i<0
+
coshr
sinhr
= ", since sinhr <0 through positive values and coshr <1.
(h) lim
i<0
cothr = lim
i<0
coshr
sinhr
= 3", since sinhr <0 through negative values and coshr <1.
(i) lim
i<3"
cschr = lim
i<3"
2
c
i
3c
3i
= 0
25. Let j = sinh
31
r. Then sinhj = r and, by Example 1(a), cosh
2
j 3sinh
2
j = 1 i [with coshj 0]
coshj =
1 + sinh
2
j =
I
1 +r
2
. So by Exercise 9, c
u
= sinhj + coshj = r +
I
1 +r
2
i j = ln
r +
I
1 +r
2
.
27. (a) Let j = tanh
31
r. Then r = tanhj =
sinhj
coshj
=
(c
u
3c
3u
)2
(c
u
+c
3u
)2

c
u
c
u
=
c
2u
31
c
2u
+ 1
i rc
2u
+r = c
2u
31 i
1 +r = c
2u
3rc
2u
i 1 +r = c
2u
(1 3r) i c
2u
=
1 +r
1 3r
i 2j = ln
1 +r
1 3r
i j =
1
2
ln
1 +r
1 3r
.
(b) Let j = tanh
31
r. Then r = tanhj, so from Exercise 18 we have
c
2u
=
1 + tanhj
1 3tanhj
=
1 +r
1 3r
i 2j = ln
1 +r
1 3r
i j =
1
2
ln
1 +r
1 3r
.
29. (a) Let j = cosh
31
r. Then coshj = r and j D 0 i sinhj
dj
dr
= 1 i
dj
dr
=
1
sinhj
=
1
cosh
2
j 31
=
1
I
r
2
31
[since sinhj D 0 for j D 0]. Or: Use Formula 4.
(b) Let j = tanh
31
r. Then tanhj = r i sech
2
j
dj
dr
= 1 i
dj
dr
=
1
sech
2
j
=
1
1 3tanh
2
j
=
1
1 3r
2
.
Or: Use Formula 5.
(c) Let j = csch
31
r. Then cschj = r i 3cschj cothj
dj
dr
= 1 i
dj
dr
= 3
1
cschj cothj
. By Exercise 13,
cothj =
csch
2
j + 1 =
I
r
2
+ 1. If r 0, then cothj 0, so cothj =
I
r
2
+ 1. If r < 0, then cothj < 0,
so cothj = 3
I
r
2
+ 1. In either case we have
dj
dr
= 3
1
cschj cothj
= 3
1
|r|
I
r
2
+ 1
.
(d) Let j = sech
31
r. Then sechj = r i 3sechj tanhj
dj
dr
= 1 i
dj
dr
= 3
1
sechj tanhj
= 3
1
sechj
1 3sech
2
j
= 3
1
r
I
1 3r
2
. [Note that j 0 and so tanhj 0.]
128

(e) Let j = coth
31
r. Then cothj = r i 3csch
2
j
dj
dr
= 1 i
dj
dr
= 3
1
csch
2
j
=
1
1 3coth
2
j
=
1
1 3r
2
by Exercise 13.
31. 1(r) = rsinhr 3coshr i 1
0
(r) = r(sinhr)
0
+ sinhr 1 3sinhr = rcoshr
33. /(r) = ln(coshr) i /
0
(r) =
1
coshr
(coshr)
0
=
sinhr
coshr
= tanhr
35. j = c
cosh 3i
i j
0
= c
cosh 3i
sinh3r 3 = 3c
cosh 3i
sinh3r
37. 1(t) = sech
2
(c
I
) = [sech(c
I
)]
2
i
1
0
(t) = 2[sech(c
I
)] [sech(c
I
)]
0
= 2 sech(c
I
)
3sech(c
I
) tanh(c
I
) c
I
= 32c
I
sech
2
(c
I
) tanh(c
I
)
39. j = arctan(tanhr) i j
0
=
1
1 + (tanhr)
2
(tanhr)
0
=
sech
2
r
1 + tanh
2
r
41. G(r) =
1 3coshr
1 + coshr
i
G
0
(r) =
(1 + coshr)(3sinhr) 3(1 3coshr)(sinhr)
(1 + coshr)
2
=
3sinhr 3sinhr coshr 3sinhr + sinhr coshr
(1 + coshr)
2
=
32 sinhr
(1 + coshr)
2
43. j = tanh
31
I
r i j
0
=
1
1 3
I
r
2

1
2
r
31'2
=
1
2
I
r(1 3r)
45. j = rsinh
31
(r3) 3
I
9 +r
2
i
j
0
= sinh
31
r
3
+r
13
1 + (r3)
2
3
2r
2
I
9 +r
2
= sinh
31
r
3
+
r
I
9 +r
2
3
r
I
9 +r
2
= sinh
31
r
3
47. j = coth
31
I
r
2
+ 1 i j
0
=
1
1 3(r
2
+ 1)
2r
2
I
r
2
+ 1
= 3
1
r
I
r
2
+ 1
49. As the depth d of the water gets large, the fraction
2d
1
gets large, and from Figure 3 or Exercise 23(a), tanh
2d
1
approaches 1. Thus, =
o1
2
tanh
2d
1
o1
2
(1) =
o1
2
.
51. (a) j = 20 cosh(r20) 315 i j
0
= 20 sinh(r20)
1
20
= sinh(r20). Since the right pole is positioned at r = 7,
we have j
0
(7) = sinh
7
20
E 0.3572.
(b) If c is the angle between the tangent line and the r-axis, then tanc = slope of the line = sinh
7
20
, so
c = tan
31
sinh
7
20
E 0.343 rad E 19.66
. Thus, the angle between the line and the pole is 0 = 90
3c E 70.34
.
53. (a) j = sinhir + 1coshir i j
0
= icoshir + i1sinhir i
j
00
= i
2
sinhir +i
2
1coshir = i
2
(sinhir +1coshir) = i
2
j
CHAPTER 3 REVIEW

129
(b) From part (a), a solution of j
00
= 9j is j(r) = sinh3r +1cosh3r. So 34 = j(0) = sinh0 +1cosh0 = 1, so
1 = 34. Now j
0
(r) = 3cosh3r 312 sinh3r i 6 = j
0
(0) = 3 i = 2, so j = 2 sinh3r 34 cosh3r.
55. The tangent to j = coshr has slope 1 when j
0
= sinhr = 1 i r = sinh
31
1 = ln
1 +
I
2
, by Equation 3.
Since sinhr = 1 and j = coshr =

1 + sinh
2
r, we have coshr =
I
2. The point is
ln
1 +
I
2
.
I
2
.
57. If oc
i
+/c
3i
= ccosh(r +o) [or csinh(r +o)], then
oc
i
+/c
3i
=
o
2
c
i+c
c
3i3c
=
o
2
c
i
c
c
c
3i
c
3c
=

o
2
c
c
c
i
o
2
c
3c
c
3i
. Comparing coefcients of c
i
and c
3i
, we have o =
o
2
c
c
(1) and / =
o
2
c
3c
(2). We need to nd c and o. Dividing equation (1) by equation (2)
gives us
a
l
= c
2c
i (-) 2o = ln
a
l
i o =
1
2
ln
a
l
. Solving equations (1) and (2) for c

c
gives us
c
c
=
2o
c
and c
c
=
c
2/
, so
2o
c
=
c
2/
i c
2
= 4o/ i c = 2
I
o/.
(-) If
a
l
0, we use the + sign and obtain a cosh function, whereas if
a
l
< 0, we use the 3 sign and obtain a sinh
function.
In summary, if o and / have the same sign, we have oc
i
+/c
3i
= 2
I
o/ cosh
r +
1
2
ln
a
l
, whereas, if o and / have the

opposite sign, then oc
i
+/c
3i
= 2
I
3o/ sinh
r +
1
2
ln
3
a
l
.
3 Review
1. (a) The Power Rule: If n is any real number, then
d
dr
(r
n
) = nr
n31
. The derivative of a variable base raised to a constant
power is the power times the base raised to the power minus one.
(b) The Constant Multiple Rule: If c is a constant and 1 is a differentiable function, then
d
dr
[c1(r)] = c
d
dr
1(r).
The derivative of a constant times a function is the constant times the derivative of the function.
(c) The Sum Rule: If 1 and o are both differentiable, then
d
dr
[1(r) +o(r)] =
d
dr
1(r) +
d
dr
o(r). The derivative of a sum
of functions is the sum of the derivatives.
(d) The Difference Rule: If 1 and o are both differentiable, then
d
dr
[1(r) 3o(r)] =
d
dr
1(r) 3
d
dr
o(r). The derivative of a
difference of functions is the difference of the derivatives.
(e) The Product Rule: If 1 and o are both differentiable, then
d
dr
[1(r) o(r)] = 1(r)
d
dr
o(r) +o(r)
d
dr
1(r). The
derivative of a product of two functions is the rst function times the derivative of the second function plus the second
function times the derivative of the rst function.
130

(f ) The Quotient Rule: If 1 and o are both differentiable, then
d
dr
1(r)
o(r)
=
o(r)
d
dr
1(r) 31(r)
d
dr
o(r)
[ o(r)]
2
.
The derivative of a quotient of functions is the denominator times the derivative of the numerator minus the numerator
times the derivative of the denominator, all divided by the square of the denominator.
(g) The Chain Rule: If 1 and o are both differentiable and 1 = 1 o is the composite function dened by 1(r) = 1(o(r)),
then 1 is differentiable and 1
0
is given by the product 1
0
(r) = 1
0
(o(r)) o
0
(r). The derivative of a composite function is
the derivative of the outer function evaluated at the inner function times the derivative of the inner function.
2. (a) j = r
n
i j
0
= nr
n31
(b) j = c
i
i j
0
= c
i
(c) j = o
i
i j
0
= o
i
lno (d) j = lnr i j
0
= 1r
(e) j = log
a
r i j
0
= 1(rlno) (f ) j = sinr i j
0
= cos r
(g) j = cos r i j
0
= 3sinr (h) j = tanr i j
0
= sec
2
r
(i) j = csc r i j
0
= 3csc r cot r ( j) j = sec r i j
0
= sec r tanr
(k) j = cot r i j
0
= 3csc
2
r (l) j = sin
31
r i j
0
= 1
I
1 3r
2
(m) j = cos
31
r i j
0
= 31
I
1 3r
2
(n) j = tan
31
r i j
0
= 1(1 +r
2
)
(o) j = sinhr i j
0
= coshr (p) j = coshr i j
0
= sinhr
(q) j = tanhr i j
0
= sech
2
r (r) j = sinh
31
r i j
0
= 1
I
1 +r
2
(s) j = cosh
31
r i j
0
= 1
I
r
2
31 (t) j = tanh
31
r i j
0
= 1(1 3r
2
)
3. (a) c is the number such that lim
I<0
c
I
31
/
= 1.
(b) c = lim
i<0
(1 +r)
1'i
(c) The differentiation formula for j = o
i
[j
0
= o
i
lno] is simplest when o = c because lnc = 1.
(d) The differentiation formula for j = log
a
r [j
0
= 1(rlno)] is simplest when o = c because lnc = 1.
4. (a) Implicit differentiation consists of differentiating both sides of an equation involving r and j with respect to r, and then
solving the resulting equation for j
0
.
(b) Logarithmic differentiation consists of taking natural logarithms of both sides of an equation j = 1(r), simplifying,
differentiating implicitly with respect to r, and then solving the resulting equation for j
0
.
5. (a) The linearization 1 of 1 at r = o is 1(r) = 1(o) +1
0
(o)(r 3o).
(b) If j = 1(r), then the differential dj is given by dj = 1
0
(r) dr.
(c) See Figure 5 in Section 3.10.
CHAPTER 3 REVIEW

131
1. True. This is the Sum Rule.
3. True. This is the Chain Rule.
5. False.
d
dr
1
I
r
=
1
0
I
r
2
I
r
by the Chain Rule.
7. False.
d
dr
10
i
= 10
i
ln10
9. True.
d
dr
(tan
2
r) = 2 tanr sec
2
r, and
d
dr
(sec
2
r) = 2 sec r(sec r tanr) = 2 tanr sec
2
r.
Or:
d
dr
(sec
2
r) =
d
dr
(1 + tan
2
r) =
d
dr
(tan
2
r).
11. True. o(r) = r
5
i o
0
(r) = 5r
4
i o
0
(2) = 5(2)
4
= 80, and by the denition of the derivative,
lim
i<2
o(r) 3o(2)
r 32
= o
0
(2) = 80.
1. j = (r
4
3 3r
2
+ 5)
3
i
j
0
= 3(r
4
33r
2
+ 5)
2
d
dr
(r
4
33r
2
+ 5) = 3(r
4
33r
2
+ 5)
2
(4r
3
36r) = 6r(r
4
33r
2
+ 5)
2
(2r
2
33)
3. j =
I
r +
1
3
I
r
4
= r
1'2
+r
34'3
i j
0
=
1
2
r
31'2
3
4
3
r
37'3
=
1
2
I
r
3
4
3
3
I
r
7
5. j = 2r
I
r
2
+ 1 i
j
0
= 2r
1
2
(r
2
+ 1)
31'2
(2r) +
I
r
2
+ 1 (2) =
2r
2
I
r
2
+ 1
+ 2
I
r
2
+ 1 =
2r
2
+ 2(r
2
+ 1)
I
r
2
+ 1
=
2(2r
2
+ 1)
I
r
2
+ 1
7. j = c
sin 20
i j
0
= c
sin 20
d
d0
(sin20) = c
sin 20
(cos 20)(2) = 2 cos 20 c
sin 20
9. j =
t
1 3t
2
i j
0
=
(1 3t
2
)(1) 3t(32t)
(1 3t
2
)
2
=
1 3t
2
+ 2t
2
(1 3t
2
)
2
=
t
2
+ 1
(1 3t
2
)
2
11. j =
I
rcos
I
r i
j
0
=
I
r

cos
I
r
0
+ cos
I
r

I
r
0
=
I
r
3sin
I
r
1
2
r
31'2
+ cos
I
r
1
2
r
31'2
=
1
2
r
31'2
3
I
r sin
I
r + cos
I
r
=
cos
I
r 3
I
r sin
I
r
2
I
r
13. j =
c
1'i
r
2
i j
0
=
r
2
(c
1'i
)
0
3c
1'i
r
2
0
(r
2
)
2
=
r
2
(c
1'i
)(31r
2
) 3c
1'i
(2r)
r
4
=
3c
1'i
(1 + 2r)
r
4
132

15.
d
dr
(rj
4
+r
2
j) =
d
dr
(r + 3j) i r 4j
3
j
0
+j
4
1 +r
2
j
0
+j 2r = 1 + 3j
0
i
j
0
(4rj
3
+r
2
33) = 1 3j
4
32rj i j
0
=
1 3j
4
32rj
4rj
3
+r
2
33
17. j =
sec 20
1 + tan20
i
j
0
=
(1 + tan20)(sec 20 tan20 2) 3(sec 20)(sec
2
20 2)
(1 + tan20)
2
=
2 sec 20 [(1 + tan20) tan20 3sec
2
20]
(1 + tan20)
2
=
2 sec 20 (tan20 + tan
2
20 3sec
2
20)
(1 + tan20)
2
=
2 sec 20 (tan20 31)
(1 + tan20)
2
1 + tan
2
r = sec
2
r
19. j = c
ci
(c sinr 3cos r) i
j
0
= c
ci
(c cos r + sinr) +cc
ci
(c sinr 3cos r) = c
ci
(c
2
sinr 3c cos r +c cos r + sinr)
= c
ci
(c
2
sinr + sinr) = c
ci
sinr(c
2
+ 1)
21. j = 3
iln i
i j
0
= 3
iln i
ln3
d
dr
(rlnr) = 3
i ln i
ln3
r
1
r
+ lnr 1
= 3
iln i
ln3(1 + lnr)
23. j = (1 3r
31
)
31
i
j
0
= 31(1 3r
31
)
32
[3(31r
32
)] = 3(1 31r)
32
r
32
= 3((r 31)r)
32
r
32
= 3(r 31)
32
25. sin(rj) = r
2
3j i cos(rj)(rj
0
+j 1) = 2r 3j
0
i rcos(rj)j
0
+j
0
= 2r 3j cos(rj) i
j
0
[rcos(rj) + 1] = 2r 3j cos(rj) i j
0
=
2r 3j cos(rj)
rcos(rj) + 1
27. j = log
5
(1 + 2r) i j
0
=
1
(1 + 2r) ln5
d
dr
(1 + 2r) =
2
(1 + 2r) ln5
29. j = lnsinr 3
1
2
sin
2
r i j
0
=
1
sinr
cos r 3
1
2
2 sinr cos r = cot r 3sinr cos r
31. j = rtan
31
(4r) i j
0
= r
1
1 + (4r)
2
4 + tan
31
(4r) 1 =
4r
1 + 16r
2
+ tan
31
(4r)
33. j = ln|sec 5r + tan5r| i
j
0
=
1
sec 5r + tan5r
(sec 5r tan5r 5 + sec
2
5r 5) =
5 sec 5r(tan5r + sec 5r)
sec 5r + tan5r
= 5 sec 5r
35. j = cot(3r
2
+ 5) i j
0
= 3csc
2
(3r
2
+ 5)(6r) = 36rcsc
2
(3r
2
+ 5)
37. j = sin
tan
I
1 +r
3
i j
0
= cos
tan
I
1 +r
3
sec
2
I
1 +r
3
3r
2
2
I
1 +r
3
39. j = tan
2
(sin0) = [tan(sin0)]
2
i j
0
= 2[tan(sin0)] sec
2
(sin0) cos 0
CHAPTER 3 REVIEW

133
41. j =
I
r + 1 (2 3r)
5
(r + 3)
7
i lnj =
1
2
ln(r +1) +5 ln(2 3r) 37 ln(r +3) i
j
0
j
=
1
2(r + 1)
+
35
2 3r
3
7
r + 3
i
j
0
=
I
r + 1 (2 3r)
5
(r + 3)
7
1
2(r + 1)
3
5
2 3r
3
7
r + 3
or j
0
=
(2 3r)
4
(3r
2
355r 352)
2
I
r + 1 (r + 3)
8
.
43. j = rsinh(r
2
) i j
0
= rcosh(r
2
) 2r + sinh(r
2
) 1 = 2r
2
cosh(r
2
) + sinh(r
2
)
45. j = ln(cosh3r) i j
0
= (1 cosh3r)(sinh3r)(3) = 3 tanh3r
47. j = cosh
31
(sinhr) i j
0
=
1
(sinhr)
2
31
coshr =
coshr
sinh
2
r 31
49. j = cos
c
I
tan 3i
i
j
0
= 3sin
c
I
tan 3i
c
I
tan 3i
0
= 3sin
c
I
tan 3i
c
I
tan 3i
1
2
(tan3r)
31'2
sec
2
(3r) 3
=
33 sin
c
I
tan 3i
c
I
tan 3i
sec
2
(3r)
2
I
tan3r
51. 1(t) =
I
4t + 1 i 1
0
(t) =
1
2
(4t + 1)
31'2
4 = 2(4t + 1)
31'2
i
1
00
(t) = 2(3
1
2
)(4t + 1)
33'2
4 = 34(4t + 1)
3'2
, so 1
00
(2) = 349
3'2
= 3
4
27
.
53. r
6
+j
6
= 1 i 6r
5
+ 6j
5
j
0
= 0 i j
0
= 3r
5
j
5
i
j
00
= 3
j
5
(5r
4
) 3r
5
(5j
4
j
0
)
(j
5
)
2
= 3
5r
4
j
4
j 3r(3r
5
j
5
)
j
10
= 3
5r
4
(j
6
+r
6
)j
5
j
6
= 3
5r
4
j
11
55. We rst show it is true for n = 1: 1(r) = rc
i
i 1
0
(r) = rc
i
+c
i
= (r + 1)c
i
. We now assume it is true
for n = I: 1
(I)
(r) = (r +I)c
i
. With this assumption, we must show it is true for n = I + 1:
1
(I+1)
(r) =
d
dr
1
(I)
(r)
=
d
dr
[(r +I)c
i
] = (r +I)c
i
+c
i
= [(r +I) + 1]c
i
= [r + (I + 1)]c
i
.
Therefore, 1
(n)
(r) = (r +n)c
i
by mathematical induction.
57. j = 4 sin
2
r i j
0
= 4 2 sinrcos r. At
r
6
. 1
, j
0
= 8
1
2

I
3
2
= 2
I
3, so an equation of the tangent line
is j 31 = 2
I
3
r 3
r
6
, or j = 2
I
3 r + 1 3
I
33.
59. j =
I
1 + 4 sinr i j
0
=
1
2
(1 + 4 sinr)
31'2
4 cos r =
2 cos r
I
1 + 4 sinr
.
At (0. 1), j
0
=
2
I
1
= 2, so an equation of the tangent line is j 31 = 2(r 30), or j = 2r + 1.
61. j = (2 +r)c
3i
i j
0
= (2 +r)(3c
3i
) +c
3i
1 = c
3i
[3(2 +r) + 1] = c
3i
(3r 31).
At (0. 2), j
0
= 1(31) = 31, so an equation of the tangent line is j 32 = 31(r 30), or j = 3r + 2.
The slope of the normal line is 1, so an equation of the normal line is j 32 = 1(r 30), or j = r + 2.
134

63. (a) 1(r) = r
I
5 3r i
1
0
(r) = r
1
2
(5 3r)
31'2
(31)
+
I
5 3r =
3r
2
I
5 3r
+
I
5 3r
2
I
5 3r
2
I
5 3r
=
3r
2
I
5 3r
+
2(5 3r)
2
I
5 3r
=
3r + 10 32r
2
I
5 3r
=
10 33r
2
I
5 3r
(b) At (1. 2): 1
0
(1) =
7
4
.
So an equation of the tangent line is j 32 =
7
4
(r 31) or j =
7
4
r +
1
4
.
At (4. 4): 1
0
(4) = 3
2
2
= 31.
So an equation of the tangent line is j 34 = 31(r 34) or j = 3r + 8.
(c)
(d) The graphs look reasonable, since 1
0
is positive where 1 has tangents with
positive slope, and 1
0
is negative where 1 has tangents with negative slope.
65. j = sinr + cos r i j
0
= cos r 3sinr = 0 C cos r = sinr and 0 $ r $ 2 C r =
r
4
or
5r
4
, so the points
are
r
4
.
I
2
and
5r
4
. 3
I
2
.
67. 1(r) = (r 3o)(r 3/)(r 3c) i 1
0
(r) = (r 3/)(r 3c) + (r 3o)(r 3c) + (r 3o)(r 3/).
So
1
0
(r)
1(r)
=
(r 3/)(r 3c) + (r 3o)(r 3c) + (r 3o)(r 3/)
(r 3o)(r 3/)(r 3c)
=
1
r 3o
+
1
r 3/
+
1
r 3c
.
Or: 1(r) = (r 3o)(r 3/)(r 3c) i ln|1(r)| = ln|r 3o| + ln|r 3/| + ln|r 3c| i
1
0
(r)
1(r)
=
1
r 3o
+
1
r 3/
+
1
r 3c
69. (a) /(r) = 1(r) o(r) i /
0
(r) = 1(r) o
0
(r) +o(r) 1
0
(r) i
/
0
(2) = 1(2) o
0
(2) +o(2) 1
0
(2) = (3)(4) + (5)(32) = 12 310 = 2
(b) 1(r) = 1(o(r)) i 1
0
(r) = 1
0
(o(r)) o
0
(r) i 1
0
(2) = 1
0
(o(2)) o
0
(2) = 1
0
(5)(4) = 11 4 = 44
71. 1(r) = r
2
o(r) i 1
0
(r) = r
2
o
0
(r) +o(r)(2r) = r[ro
0
(r) + 2o(r)]
73. 1(r) = [ o(r)]
2
i 1
0
(r) = 2[ o(r)] o
0
(r) = 2o(r) o
0
(r)
75. 1(r) = o(c
i
) i 1
0
(r) = o
0
(c
i
) c
i
77. 1(r) = ln|o(r)| i 1
0
(r) =
1
o(r)
o
0
(r) =
o
0
(r)
o(r)
CHAPTER 3 REVIEW

135
79. /(r) =
1(r) o(r)
1(r) +o(r)
i
/
0
(r) =
[1(r) +o(r)] [1(r) o
0
(r) +o(r) 1
0
(r)] 31(r) o(r) [1
0
(r) +o
0
(r)]
[1(r) +o(r)]
2
=
[1(r)]
2
o
0
(r) +1(r) o(r) 1
0
(r) +1(r) o(r) o
0
(r) + [ o(r)]
2
1
0
(r) 31(r) o(r) 1
0
(r) 31(r) o(r) o
0
(r)
[1(r) +o(r)]
2
=
1
0
(r) [ o(r)]
2
+o
0
(r) [1(r)]
2
[1(r) +o(r)]
2
81. Using the Chain Rule repeatedly, /(r) = 1(o(sin4r)) i
/
0
(r) = 1
0
(o(sin4r))
d
dr
(o(sin4r)) = 1
0
(o(sin4r)) o
0
(sin4r)
d
dr
(sin4r) = 1
0
(o(sin4r))o
0
(sin4r)(cos 4r)(4).
83. j = [ln(r + 4)]
2
i j
0
= 2[ln(r + 4)]
1
1
r + 4
1 = 2
ln(r + 4)
r + 4
and j
0
= 0 C ln(r + 4) = 0 C
r + 4 = c
0
i r + 4 = 1 C r = 33, so the tangent is horizontal at the point (33. 0).
85. j = 1(r) = or
2
+/r +c i 1
0
(r) = 2or +/. We know that 1
0
(31) = 6 and 1
0
(5) = 32, so 32o +/ = 6 and
10o +/ = 32. Subtracting the rst equation from the second gives 12o = 38 i o = 3
2
3
. Substituting 3
2
3
for o in the
rst equation gives / =
14
3
. Now 1(1) = 4 i 4 = o +/ +c, so c = 4 +
2
3
3
14
3
= 0 and hence, 1(r) = 3
2
3
r
2
+
14
3
r.
87. c(t) = c
3cI
cos(.t +c) i
(t) = c
0
(t) = {c
3cI
[3. sin(.t +c)] + cos(.t +c)(3cc
3cI
)} = 3c
3cI
[. sin(.t +c) +c cos(.t +c)] i
o(t) =
0
(t) = 3{c
3cI
[.
2
cos(.t +c) 3c. sin(.t +c)] + [. sin(.t +c) +c cos(.t +c)](3cc
3cI
)}
= 3c
3cI
[.
2
cos(.t +c) 3c. sin(.t +c) 3c. sin(.t +c) 3c
2
cos(.t +c)]
= 3c
3cI
[(.
2
3c
2
) cos(.t +c) 32c. sin(.t +c)] = c
3cI
[(c
2
3.
2
) cos(.t +c) + 2c. sin(.t +c)]
89. (a) j = t
3
312t + 3 i (t) = j
0
= 3t
2
312 i o(t) =
0
(t) = 6t
(b) (t) = 3(t
2
34) 0 when t 2, so it moves upward when t 2 and downward when 0 $ t < 2.
(c) Distance upward = j(3) 3j(2) = 36 3(313) = 7,
Distance downward = j(0) 3j(2) = 3 3(313) = 16. Total distance = 7 + 16 = 23.
(d) (e) The particle is speeding up when and o have the same sign, that is,
when t 2. The particle is slowing down when and o have opposite
signs; that is, when 0 < t < 2.
91. The linear density a is the rate of change of mass i with respect to length r.
i = r
1 +
I
r
= r +r
3'2
i a = didr = 1 +
3
2
I
r, so the linear density when r = 4 is 1 +
3
2
I
4 = 4 kgm.
136

93. (a) j(t) = j(0)c
II
= 200c
II
i j(0.5) = 200c
0.5I
= 360 i c
0.5I
= 1.8 i 0.5I = ln1.8 i
I = 2 ln1.8 = ln(1.8)
2
= ln3.24 i j(t) = 200c
(ln 3.24)I
= 200(3.24)
I
(b) j(4) = 200(3.24)
4
E 22,040 bacteria
(c) j
0
(t) = 200(3.24)
I
ln3.24, so j
0
(4) = 200(3.24)
4
ln3.24 E 25,910 bacteria per hour
(d) 200(3.24)
I
= 10,000 i (3.24)
I
= 50 i t ln3.24 = ln50 i t = ln50 ln3.24 E 3.33 hours
95. (a) C
0
(t) = 3IC(t) i C(t) = C(0)c
3II
by Theorem 9.4.2. But C(0) = C
0
, so C(t) = C
0
c
3II
.
(b) C(30) =
1
2
C
0
since the concentration is reduced by half. Thus,
1
2
C
0
= C
0
c
330I
iln
1
2
= 330I i
I = 3
1
30
ln
1
2
=
1
30
ln2. Since 10% of the original concentration remains if 90% is eliminated, we want the value of t
such that C(t) =
1
10
C
0
. Therefore,
1
10
C
0
= C
0
c
3I(ln 2)'30
i ln0.1 = 3t(ln2)30 i t = 3
30
ln 2
ln0.1 E 100 h.
97. If r = edge length, then \ = r
3
i d\dt = 3r
2
drdt = 10 i drdt = 10(3r
2
) and o = 6r
2
i
dodt = (12r) drdt = 12r[10(3r
2
)] = 40r. When r = 30, dodt =
40
30
=
4
3
cm
2
min.
99. Given d/dt = 5 and drdt = 15, nd d:dt. :
2
= r
2
+/
2
i
2:
d:
dt
= 2r
dr
dt
+ 2/
d/
dt
i
d:
dt
=
1
:
(15r + 5/). When t = 3,
/ = 45 + 3(5) = 60 and r = 15(3) = 45 i : =
I
45
2
+ 60
2
= 75,
so
d:
dt
=
1
75
[15(45) + 5(60)] = 13 fts.
101. We are given d0dt = 30.25 radh. tan0 = 400r i
r = 400 cot 0 i
dr
dt
= 3400 csc
2
0
d0
dt
. When 0 =
r
6
,
dr
dt
= 3400(2)
2
(30.25) = 400 fth.
103. (a) 1(r) =
3
I
1 + 3r = (1 + 3r)
1'3
i 1
0
(r) = (1 + 3r)
32'3
, so the linearization of 1 at o = 0 is
1(r) = 1(0) +1
0
(0)(r 30) = 1
1'3
+ 1
32'3
r = 1 + r. Thus,
3
I
1 + 3r E 1 +r i
3
I
1.03 =
3
1 + 3(0.01) E 1 + (0.01) = 1.01.

(b) The linear approximation is
3
I
1 + 3r E 1 +r, so for the required accuracy
we want
3
I
1 + 3r 30.1 < 1 +r <
3
I
1 + 3r + 0.1. From the graph,
it appears that this is true when 30.23 < r < 0.40.
CHAPTER 3 REVIEW

137
105. = r
2
+
1
2
1
2
r
2
=

1 +
r
8
r
2
i d =

2 +
r
4
rdr. When r = 60
and dr = 0.1, d =

2 +
r
4
60(0.1) = 12 +
3r
2
, so the maximum error is
approximately 12 +
3r
2
E 16.7 cm
2
.
107. lim
I<0
4
I
16 +/ 32
/
=

d
dr
4
I
r
i =16
=
1
4
r
33'4
i =16
=
1
4
4
I
16
3
=
1
32
109. lim
i<0
I
1 + tanr 3
I
1 + sinr
r
3
= lim
i<0
I
1 + tanr 3
I
1 + sinr
I
1 + tanr +
I
1 + sinr
r
3
I
1 + tanr +
I
1 + sinr
= lim
i<0
(1 + tanr) 3(1 + sinr)
r
3
I
1 + tanr +
I
1 + sinr
= lim
i<0
sinr(1 cos r 31)
r
3
I
1 + tanr +
I
1 + sinr

cos r
cos r
= lim
i<0
sinr(1 3cos r)
r
3
I
1 + tanr +
I
1 + sinr
cos r

1 + cos r
1 + cos r
= lim
i<0
sinr sin
2
r
r
3
I
1 + tanr +
I
1 + sinr
cos r(1 + cos r)

=

lim
i<0
sinr
r
3
lim
i<0
1
I
1 + tanr +
I
1 + sinr
cos r(1 + cos r)

= 1
3
1
I
1 +
I
1
1 (1 + 1)
=
1
4
111.
d
dr
[1(2r)] = r
2
i 1
0
(2r) 2 = r
2
i 1
0
(2r) =
1
2
r
2
. Let t = 2r. Then 1
0
(t) =
1
2
1
2
t
2
=
1
8
t
2
, so 1
0
(r) =
1
8
r
2
.
PROBLEMS PLUS
1. Let o be the r-coordinate of Q. Since the derivative of j = 1 3r
2
is j
0
= 32r, the slope at Qis 32o. But since the triangle
is equilateral, OOC =
I
31, so the slope at Q is 3
I
3. Therefore, we must have that 32o = 3
I
3 i o =
I
3
2
.
Thus, the point Q has coordinates
I
3
2
. 1 3
I
3
2
=

I
3
2
.
1
4
and by symmetry, 1 has coordinates
3
I
3
2
.
1
4
.
3.

We must show that v (in the gure) is halfway between j and q, that is,
v = (j +q)2. For the parabola j = or
2
+/r +c, the slope of the tangent line is
given by j
0
= 2or +/. An equation of the tangent line at r = j is
j 3(oj
2
+/j +c) = (2oj +/)(r 3j). Solving for j gives us
j = (2oj +/)r 32oj
2
3/j + (oj
2
+/j +c)
or j = (2oj +/)r +c 3oj
2
(1)
Similarly, an equation of the tangent line at r = q is
j = (2oq +/)r +c 3oq
2
(2)
We can eliminate j and solve for r by subtracting equation (1) from equation (2).
[(2oq +/) 3(2oj +/)]r 3oq
2
+oj
2
= 0
(2oq 32oj)r = oq
2
3oj
2
2o(q 3j)r = o(q
2
3j
2
)
r =
o(q +j)(q 3j)
2o(q 3j)
=
j +q
2
Thus, the r-coordinate of the point of intersection of the two tangent lines, namely v, is (j +q)2.
5. Let j = tan
31
r. Then tanj = r, so from the triangle we see that
sin(tan
31
r) = sinj =
r
I
1 +r
2
. Using this fact we have that
sin(tan
31
(sinhr)) =
sinhr
1 + sinh
2
r
=
sinhr
coshr
= tanhr.
Hence, sin
31
(tanhr) = sin
31
(sin(tan
31
(sinhr))) = tan
31
(sinhr).
7. We use mathematical induction. Let o
n
be the statement that
d
n
dr
n
(sin
4
r + cos
4
r) = 4
n31
cos(4r +n2).
o
1
is true because
d
dr
(sin
4
r + cos
4
r) = 4 sin
3
r cos r 34 cos
3
r sinr = 4 sinr cos r
sin
2
r 3cos
2
r
r
= 34 sinr cos r cos 2r = 32 sin2r cos 2 = 3sin4r = sin(34r)
= cos
r
2
3(34r)
= cos
r
2
+ 4r
= 4
n31
cos
4r +n
r
2
when n = 1
[continued]
139
140

Now assume o
I
is true, that is,
d
I
dr
I
sin
4
r + cos
4
r
= 4
I31
cos
4r +I
r
2
. Then
d
I+1
dr
I+1
(sin
4
r + cos
4
r) =
d
dr
d
I
dr
I
(sin
4
r + cos
4
r)
=
d
dr
4
I31
cos
4r +I
r
2
= 34
I31
sin
4r +I
r
2
d
dr
4r +I
r
2
= 34
I
sin
4r +I
r
2
= 4
I
sin
34r 3I
r
2
= 4
I
cos
r
2
3
34r 3I
r
2
= 4
I
cos
4r + (I + 1)
r
2
which shows that o

I+1
is true.
Therefore,
d
n
dr
n
(sin
4
r + cos
4
r) = 4
n31
cos
4r +n
r
2
for every positive integer n, by mathematical induction.

Another proof: First write
sin
4
r + cos
4
r = (sin
2
r + cos
2
r)
2
32 sin
2
r cos
2
r = 1 3
1
2
sin
2
2r = 1 3
1
4
(1 3cos 4r) =
3
4
+
1
4
cos 4r
Then we have
d
n
dr
n
(sin
4
r + cos
4
r) =
d
n
dr
n
3
4
+
1
4
cos 4r
=
1
4
4
n
cos
4r +n
= 4
n31
cos
4r +n
.
9. We must nd a value r
0
such that the normal lines to the parabola j = r
2
at r = r
0
intersect at a point one unit from the
points
r
0
. r
2
0
. The normals to j = r
2
at r = r
0
have slopes 3
1
2r
0
and pass through
r
0
. r
2
0
respectively, so the
normals have the equations j 3r
2
0
= 3
1
2r
0
(r 3r
0
) and j 3r
2
0
=
1
2r
0
(r +r
0
). The common j-intercept is r
2
0
+
1
2
.
We want to nd the value of r
0
for which the distance from
0. r
2
0
+
1
2
to
r
0
. r
2
0
equals 1. The square of the distance is

(r
0
30)
2
+
r
2
0
3
r
2
0
+
1
2
2
= r
2
0
+
1
4
= 1 C r
0
=
I
3
2
. For these values of r
0
, the j-intercept is r
2
0
+
1
2
=
5
4
, so
the center of the circle is at
0.
5
4
.
Another solution: Let the center of the circle be (0. o). Then the equation of the circle is r
2
+ (j 3o)
2
= 1.
Solving with the equation of the parabola, j = r
2
, we get r
2
+ (r
2
3o)
2
= 1 C r
2
+r
4
32or
2
+o
2
= 1 C
r
4
+ (1 32o)r
2
+o
2
31 = 0. The parabola and the circle will be tangent to each other when this quadratic equation in r
2
has equal roots; that is, when the discriminant is 0. Thus, (1 32o)
2
34(o
2
31) = 0 C
1 34o + 4o
2
34o
2
+ 4 = 0 C 4o = 5, so o =
5
4
. The center of the circle is
0.
5
4
.
11. We can assume without loss of generality that 0 = 0 at time t = 0, so that 0 = 12t rad. [The angular velocity of the wheel
is 360 rpm = 360 (2 rad)(60 s) = 12 rads.] Then the position of as a function of time is
= (40 cos 0. 40 sin0) = (40 cos 12t. 40 sin12t), so sinc =
j
1.2 m
=
40 sin0
120
=
sin0
3
=
1
3
sin12t.
(a) Differentiating the expression for sinc, we get cos c
dc
dt
=
1
3
12 cos 12t = 4 cos 0. When 0 =

3
, we have
sinc =
1
3
sin0 =
I
3
6
, so cos c =
1 3
I
3
6
2
=
11
12
and
dc
dt
=
4 cos
r
3
cos c
=
2
1112
=
4
I
3
I
11
E 6.56 rads.

141
(b) By the Law of Cosines, |1|
2
= |O|
2
+ |O1|
2
32 |O| |O1| cos 0 i
120
2
= 40
2
+ |O1|
2
32 40 |O1| cos 0 i |O1|
2
3(80 cos 0) |O1| 312,800 = 0 i
|O1| =
1
2
80 cos 0
I
6400 cos
2
0 + 51,200
= 40 cos 0 40
I
cos
2
0 + 8 = 40
cos 0 +
I
8 + cos
2
0
cm
[since |O1| 0]. As a check, note that |O1| = 160 cm when 0 = 0 and |O1| = 80
I
2 cm when 0 =
r
2
.
(c) By part (b), the r-coordinate of 1 is given by r = 40
cos 0 +
I
8 + cos
2
0
, so
dr
dt
=
dr
d0
d0
dt
= 40
3sin0 3
2 cos 0 sin0
2
I
8 + cos
2
0
12 = 3480 sin0
1 +
cos 0
I
8 + cos
2
0
cms.
In particular, drdt = 0 cms when 0 = 0 and drdt = 3480 cms when 0 =
r
2
.
13. Consider the statement that
d
n
dr
n
(c
ai
sin/r) = v
n
c
ai
sin(/r + n0). For n = 1,
d
dr
(c
ai
sin/r) = oc
ai
sin/r +/c
ai
cos /r, and
vc
ai
sin(/r +0) = vc
ai
[sin/rcos 0 + cos /rsin0] = vc
ai
o
v
sin/r +
/
v
cos /r
= oc
ai
sin/r +/c
ai
cos /r
since tan0 =
/
o
i sin0 =
/
v
and cos 0 =
o
v
. So the statement is true for n = 1.
Assume it is true for n = I. Then
d
I+1
dr
I+1
(c
ai
sin/r) =
d
dr
v
I
c
ai
sin(/r +I0)
= v
I
oc
ai
sin(/r +I0) +v
I
c
ai
/ cos(/r +I0)
= v
I
c
ai
[o sin(/r +I0) +/ cos(/r +I0)]
But
sin[/r + (I + 1)0] = sin[(/r +I0) +0] = sin(/r +I0) cos 0 + sin0 cos(/r +I0) =
a
r
sin(/r +I0) +
l
r
cos(/r +I0).
Hence, o sin(/r +I0) +/ cos(/r +I0) = v sin[/r + (I + 1)0]. So
d
I+1
dr
I+1
(c
ai
sin/r) = v
I
c
ai
[o sin(/r+I0)+/ cos(/r+I0)] = v
I
c
ai
[v sin(/r+(I+1)0)] = v
I+1
c
ai
[sin(/r+(I+1)0)].
Therefore, the statement is true for all n by mathematical induction.
15. It seems from the gure that as 1 approaches the point (0. 2) from the right, r
T
<"and j
T
<2
+
. As 1 approaches the
point (3. 0) from the left, it appears that r
T
<3
+
and j
T
<". So we guess that r
T
M (3. ") and j
T
M (2. "). It is
more difcult to estimate the range of values for r
^
and j
^
. We might perhaps guess that r
^
M (0. 3),
and j
^
M (3". 0) or (32. 0).
In order to actually solve the problem, we implicitly differentiate the equation of the ellipse to nd the equation of the
tangent line:
r
2
9
+
j
2
4
= 1 i
2r
9
+
2j
4
j
0
= 0, so j
0
= 3
4
9
r
j
. So at the point (r
0
. j
0
) on the ellipse, an equation of the
142

tangent line is j 3j
0
= 3
4
9
r
0
j
0
(r 3r
0
) or 4r
0
r + 9j
0
j = 4r
2
0
+ 9j
2
0
. This can be written as
r
0
r
9
+
j
0
j
4
=
r
2
0
9
+
j
2
0
4
= 1,
because (r
0
. j
0
) lies on the ellipse. So an equation of the tangent line is
r
0
r
9
+
j
0
j
4
= 1.
Therefore, the r-intercept r
T
for the tangent line is given by
r
0
r
T
9
= 1 C r
T
=
9
r
0
, and the j-intercept j
T
is given
by
j
0
j
T
4
= 1 C j
T
=
4
j
0
.
So as r
0
takes on all values in (0. 3), r
T
takes on all values in (3. "), and as j
0
takes on all values in (0. 2), j
T
takes on
all values in (2. "). At the point (r
0
. j
0
) on the ellipse, the slope of the normal line is 3
1
j
0
(r
0
. j
0
)
=
9
4
j
0
r
0
, and its
equation is j 3j
0
=
9
4
j
0
r
0
(r 3r
0
). So the r-intercept r
^
for the normal line is given by 0 3j
0
=
9
4
j
0
r
0
(r
^
3r
0
) i
r
^
= 3
4r
0
9
+r
0
=
5r
0
9
, and the j-intercept j
^
is given by j
^
3j
0
=
9
4
j
0
r
0
(0 3r
0
) i j
^
= 3
9j
0
4
+j
0
= 3
5j
0
4
.
So as r
0
takes on all values in (0. 3), r
^
takes on all values in
0.
5
3
, and as j
0
takes on all values in (0. 2), j
^
takes on
all values in
3
5
2
. 0
.
17. (a) If the two lines 1
1
and 1
2
have slopes i
1
and i
2
and angles of
inclination c
1
and c
2
, then i
1
= tan c
1
and i
2
= tan c
2
. The triangle
in the gure shows that c
1
+c + (180
3c
2
) = 180
and so
c = c
2
3c
1
. Therefore, using the identity for tan(r 3j), we have
tanc = tan(c
2
3c
1
) =
tanc
2
3tanc
1
1 + tanc
2
tanc
1
and so tan c =
i
2
3i
1
1 +i
1
i
2
.
(b) (i) The parabolas intersect when r
2
= (r 32)
2
i r = 1. If j = r
2
, then j
0
= 2r, so the slope of the tangent
to j = r
2
at (1. 1) is i
1
= 2(1) = 2. If j = (r 32)
2
, then j
0
= 2(r 32), so the slope of the tangent to
j = (r 32)
2
at (1. 1) is i
2
= 2(1 32) = 32. Therefore, tanc =
i
2
3i
1
1 +i
1
i
2
=
32 32
1 + 2(32)
=
4
3
and
so c = tan
31
4
3
E 53
[or 127
].
(ii) r
2
3j
2
= 3 and r
2
34r +j
2
+ 3 = 0 intersect when r
2
34r + (r
2
33) + 3 = 0 C 2r(r 32) = 0 i
r = 0 or 2, but 0 is extraneous. If r = 2, then j = 1. If r
2
3j
2
= 3 then 2r 32jj
0
= 0 i j
0
= rj and
r
2
34r +j
2
+ 3 = 0 i 2r 34 + 2jj
0
= 0 i j
0
=
2 3r
j
. At (2. 1) the slopes are i
1
= 2 and
i
2
= 0, so tanc =
0 32
1 +2 0
= 32 i c E 117
. At (2. 31) the slopes are i

1
= 32 and i
2
= 0.
so tanc =
0 3(32)
1 + (32) (0)
= 2 i c E 63
[or 117
].

143
19. Since _1OQ = _OQ1 = 0, the triangle QO1 is isosceles, so
|Q1| = |1O| = r. By the Law of Cosines, r
2
= r
2
+v
2
32vrcos 0. Hence,
2vrcos 0 = v
2
, so r =
v
2
2v cos 0
=
v
2 cos 0
. Note that as j <0
+
, 0 <0
+
(since
sin0 = jv), and hence r <
v
2 cos 0
=
v
2
. Thus, as 1 is taken closer and closer
to the r-axis, the point 1 approaches the midpoint of the radius O.
21. lim
i<0
sin(o + 2r) 32 sin(o +r) + sino
r
2
= lim
i<0
sino cos 2r + cos o sin2r 32 sino cos r 32 cos o sinr + sino
r
2
= lim
i<0
sino (cos 2r 32 cos r + 1) + cos o (sin2r 32 sinr)
r
2
= lim
i<0
sino (2 cos
2
r 31 32 cos r + 1) + cos o (2 sinr cos r 32 sinr)
r
2
= lim
i<0
sino (2 cos r)(cos r 31) + cos o (2 sinr)(cos r 31)
r
2
= lim
i<0
2(cos r 31)[sino cos r + cos o sinr](cos r + 1)
r
2
(cos r + 1)
= lim
i<0
32 sin
2
r[sin(o +r)]
r
2
(cos r + 1)
= 32 lim
i<0
sinr
r
sin(o +r)
cos r + 1
= 32(1)
2
sin(o + 0)
cos 0 + 1
= 3sino
23. Let 1(r) = c
2i
and o(r) = I
I
r [I 0]. From the graphs of 1 and o,
we see that 1 will intersect o exactly once when 1 and o share a tangent
line. Thus, we must have 1 = o and 1
0
= o
0
at r = o.
1(o) = o(o) i c
2a
= I
I
o (-)
and 1
0
(o) = o
0
(o) i 2c
2a
=
I
2
I
o
i c
2a
=
I
4
I
o
.
So we must have I
I
o =
I
4
I
o
i

I
o
2
=
I
4I
i o =
1
4
. From (-), c
2(1'4)
= I
14 i
I = 2c
1'2
= 2
I
c E 3.297.
25. j =
r
I
o
2
31
3
2
I
o
2
31
arctan
sinr
o +
I
o
2
31 + cos r
. Let I = o +
I
o
2
31. Then
j
0
=
1
I
o
2
31
3
2
I
o
2
31

1
1 + sin
2
r(I + cos r)
2

cos r(I + cos r) + sin
2
r
(I + cos r)
2
=
1
I
o
2
31
3
2
I
o
2
31

I cos r + cos
2
r + sin
2
r
(I + cos r)
2
+ sin
2
r
=
1
I
o
2
31
3
2
I
o
2
31

I cos r + 1
I
2
+ 2I cos r + 1
=
I
2
+ 2I cos r + 1 32I cos r 32
I
o
2
31 (I
2
+ 2I cos r + 1)
=
I
2
31
I
o
2
31 (I
2
+ 2I cos r + 1)
But I
2
= 2o
2
+ 2o
I
o
2
31 31 = 2o
o +
I
o
2
31
31 = 2oI 31, so I
2
+ 1 = 2oI, and I
2
31 = 2(oI 31).
[continued]
144

So j
0
=
2(oI 31)
I
o
2
31 (2oI + 2I cos r)
=
oI 31
I
o
2
31I (o + cos r)
. But oI 31 = o
2
+o
I
o
2
31 31 = I
I
o
2
31,
so j
0
= 1(o + cos r).
27. j = r
4
32r
2
3r i j
0
= 4r
3
34r 31. The equation of the tangent line at r = o is
j 3(o
4
32o
2
3o) = (4o
3
34o 31)(r 3o) or j = (4o
3
34o 31)r + (33o
4
+ 2o
2
) and similarly for r = /. So if at
r = o and r = / we have the same tangent line, then 4o
3
34o 31 = 4/
3
34/ 31 and 33o
4
+2o
2
= 33/
4
+2/
2
. The rst
equation gives o
3
3/
3
= o 3/ i (o 3/)(o
2
+o/ +/
2
) = (o 3/). Assuming o 6= /, we have 1 = o
2
+o/ +/
2
.
The second equation gives 3(o
4
3/
4
) = 2(o
2
3/
2
) i 3(o
2
3/
2
)(o
2
+/
2
) = 2(o
2
3/
2
) which is true if o = 3/.
Substituting into 1 = o
2
+o/ +/
2
gives 1 = o
2
3o
2
+o
2
i o = 1 so that o = 1 and / = 31 or vice versa. Thus,
the points (1. 32) and (31. 0) have a common tangent line.
As long as there are only two such points, we are done. So we show that these are in fact the only two such points.
Suppose that o
2
3/
2
6= 0. Then 3(o
2
3/
2
)(o
2
+/
2
) = 2(o
2
3/
2
) gives 3(o
2
+/
2
) = 2 or o
2
+/
2
=
2
3
.
Thus, o/ = (o
2
+o/ +/
2
) 3(o
2
+/
2
) = 1 3
2
3
=
1
3
, so / =
1
3o
. Hence, o
2
+
1
9o
2
=
2
3
, so 9o
4
+ 1 = 6o
2
i
0 = 9o
4
36o
2
+1 = (3o
2
31)
2
. So 3o
2
31 = 0 i o
2
=
1
3
i /
2
=
1
9o
2
=
1
3
= o
2
, contradicting our assumption
that o
2
6= /
2
.
29. Because of the periodic nature of the lattice points, it sufces to consider the points in the 5 2 grid shown. We can see that
the minimum value of v occurs when there is a line with slope
2
5
which touches the circle centered at (3. 1) and the circles
centered at (0. 0) and (5. 2).
To nd 1, the point at which the line is tangent to the circle at (0. 0), we simultaneously solve r
2
+j
2
= v
2
and
j = 3
5
2
r i r
2
+
25
4
r
2
= v
2
i r
2
=
4
29
v
2
i r =
2
I
29
v, j = 3
5
I
29
v. To nd Q, we either use symmetry or
solve (r 33)
2
+(j 31)
2
= v
2
and j 31 = 3
5
2
(r 33). As above, we get r = 3 3
2
I
29
v, j = 1 +
5
I
29
v. Now the slope of
the line 1Q is
2
5
, so i
TC
=
1 +
5
I
29
v 3
3
5
I
29
v
3 3
2
I
29
v 3
2
I
29
v
=
1 +
10
I
29
v
3 3
4
I
29
v
=
I
29 + 10v
3
I
29 34v
=
2
5
i
5
I
29 + 50v = 6
I
29 38v C 58v =
I
29 C v =
I
29
58
. So the minimum value of v for which any line with slope
2
5
intersects circles with radius v centered at the lattice points on the plane is v =
I
29
58
E 0.093.

145
31. By similar triangles,
v
5
=
/
16
i v =
5/
16
. The volume of the cone is
\ =
1
3
v
2
/ =
1
3
5/
16
2
/ =
25
768
/
3
, so
d\
dt
=
25
256
/
2
d/
dt
. Now the rate of
change of the volume is also equal to the difference of what is being added
(2 cm
3
min) and what is oozing out (Iv|, where v| is the area of the cone and I
is a proportionality constant). Thus,
d\
dt
= 2 3Iv|.
Equating the two expressions for
d\
dt
and substituting / = 10,
d/
dt
= 30.3, v =
5(10)
16
=
25
8
, and
|
I
281
=
10
16
C
| =
5
8
I
281, we get
25
256
(10)
2
(30.3) = 2 3I
25
8

5
8
I
281 C
125I
I
281
64
= 2 +
750
256
. Solving for I gives us
I =
256 + 375
250
I
281
. To maintain a certain height, the rate of oozing, Iv|, must equal the rate of the liquid being poured in;
that is,
d\
dt
= 0. Thus, the rate at which we should pour the liquid into the container is
Iv| =
256 + 375
250
I
281

25
8

5
I
281
8
=
256 + 375
128
E 11.204 cm
3
min
4 APPLICATIONS OF DIFFERENTIATION
4.1 Maximum and Minimum Values
1. A function 1 has an absolute minimum at r = c if 1(c) is the smallest function value on the entire domain of 1, whereas
1 has a local minimum at c if 1(c) is the smallest function value when r is near c.
3. Absolute maximum at c, absolute minimum at v, local maximum at c, local minima at / and v, neither a maximum nor a
minimum at o and d.
5. Absolute maximum value is 1(4) = 5; there is no absolute minimum value; local maximum values are 1(4) = 5 and
1(6) = 4; local minimum values are 1(2) = 2 and 1(1) = 1(5) = 3.
7. Absolute minimum at 2, absolute maximum at 3,
local minimum at 4
9. Absolute maximum at 5, absolute minimum at 2,
local maximum at 3, local minima at 2 and 4
11. (a) (b) (c)
13. (a) Note: By the Extreme Value Theorem,
1 must not be continuous; because if it
were, it would attain an absolute
minimum.
(b)
147
148

CHAPTER 4 APPLICATIONS OF DIFFERENTIATION
15. 1(r) = 8 33r, r D 1. Absolute maximum 1(1) = 5; no
local maximum. No absolute or local minimum.
17. 1(r) = r
2
, 0 < r < 2. No absolute or local maximum or
minimum value.
19. 1(r) = r
2
, 0 $ r < 2. Absolute minimum 1(0) = 0; no
local minimum. No absolute or local maximum.
21. 1(r) = r
2
, 33 $ r $ 2. Absolute maximum
1(33) = 9. No local maximum. Absolute and local
minimum 1(0) = 0.
23. 1(r) = lnr, 0 < r $ 2. Absolute maximum
1(2) = ln2 r 0.69; no local maximum. No absolute or
local minimum.
25. 1(r) = 1 3
I
r. Absolute maximum 1(0) = 1; no local
maximum. No absolute or local minimum.
27. 1(r) =
1 3r if 0 $ r < 2
2r 34 if 2 $ r $ 3
Absolute maximum 1(3) = 2; no local maximum. No absolute or local minimum.
SECTION 4.1 MAXIMUM AND MINIMUM VALUES

149
29. 1(r) = 5r
2
+ 4r i 1
0
(r) = 10r + 4. 1
0
(r) = 0 i r = 3
2
5
, so 3
2
5
is the only critical number.
31. 1(r) = r
3
+ 3r
2
324r i 1
0
(r) = 3r
2
+ 6r 324 = 3(r
2
+ 2r 38).
1
0
(r) = 0 i 3(r + 4)(r 32) = 0 i r = 34, 2. These are the only critical numbers.
33. c(t) = 3t
4
+ 4t
3
36t
2
i c
0
(t) = 12t
3
+ 12t
2
312t. c
0
(t) = 0 i 12t(t
2
+t 31) i
t = 0 or t
2
+t 3 1 = 0. Using the quadratic formula to solve the latter equation gives us
t =
31
1
2
34(1)(31)
2(1)
=
31
I
5
2
E 0.618, 31.618. The three critical numbers are 0,
31
I
5
2
.
35. o(j) =
j 31
j
2
3j + 1
i
o
0
(j) =
(j
2
3j + 1)(1) 3(j 31)(2j 31)
(j
2
3j + 1)
2
=
j
2
3j + 1 3(2j
2
33j + 1)
(j
2
3j + 1)
2
=
3j
2
+ 2j
(j
2
3j + 1)
2
=
j(2 3j)
(j
2
3j + 1)
2
.
o
0
(j) = 0 i j = 0, 2. The expression j
2
3j + 1 is never equal to 0, so o
0
(j) exists for all real numbers.
The critical numbers are 0 and 2.
37. /(t) = t
3'4
32t
1'4
i /
0
(t) =
3
4
t
31'4
3
2
4
t
33'4
=
1
4
t
33'4
(3t
1'2
32) =
3
I
t 32
4
4
I
t
3
.
/
0
(t) = 0 i 3
I
t = 2 i
I
t =
2
3
i t =
4
9
. /
0
(t) does not exist at t = 0, so the critical numbers are 0 and
4
9
.
39. 1(r) = r
4'5
(r 34)
2
i
1
0
(r) = r
4'5
2(r 34) + (r 34)
2
4
5
r
31'5
=
1
5
r
31'5
(r 34)[5 r 2 + (r 34) 4]
=
(r 34)(14r 316)
5r
1'5
=
2(r 34)(7r 38)
5r
1'5
1
0
(r) = 0 i r = 4,
8
7
. 1
0
(0) does not exist. Thus, the three critical numbers are 0,
8
7
, and 4.
41. 1(0) = 2 cos 0 + sin
2
0 i 1
0
(0) = 32 sin0 + 2 sin0 cos 0. 1
0
(0) = 0 i 2 sin0 (cos 0 31) = 0 i sin0 = 0
or cos 0 = 1 i 0 = n [n an integer] or 0 = 2n. The solutions 0 = n include the solutions 0 = 2n, so the critical
numbers are 0 = n.
43. 1(r) = r
2
c
33i
i 1
0
(r) = r
2
(33c
33i
) +c
33i
(2r) = rc
33i
(33r + 2). 1
0
(r) = 0 i r = 0,
2
3
[c
33i
is never equal to 0]. 1
0
(r) always exists, so the critical numbers are 0 and
2
3
.
45. The graph of 1
0
(r) = 5c
30.1|i|
sinr 31 has 10 zeros and exists
everywhere, so 1 has 10 critical numbers.
150

47. 1(r) = 3r
2
312r + 5, [0. 3]. 1
0
(r) = 6r 312 = 0 C r = 2. Applying the Closed Interval Method, we nd that
1(0) = 5, 1(2) = 37, and 1(3) = 34. So 1(0) = 5 is the absolute maximum value and 1(2) = 37 is the absolute minimum
value.
49. 1(r) = 2r
3
33r
2
312r + 1, [32. 3]. 1
0
(r) = 6r
2
36r 312 = 6(r
2
3r 32) = 6(r 32)(r + 1) = 0 C
r = 2. 31. 1(32) = 33, 1(31) = 8, 1(2) = 319, and 1(3) = 38. So 1(31) = 8 is the absolute maximum value and
1(2) = 319 is the absolute minimum value.
51. 1(r) = r
4
32r
2
+ 3, [32. 3]. 1
0
(r) = 4r
3
34r = 4r(r
2
31) = 4r(r + 1)(r 31) = 0 C r = 31, 0, 1.
1(32) = 11, 1(31) = 2, 1(0) = 3, 1(1) = 2, 1(3) = 66. So 1(3) = 66 is the absolute maximum value and 1(1) = 2 is
the absolute minimum value.
53. 1(r) =
r
r
2
+ 1
, [0. 2]. 1
0
(r) =
(r
2
+ 1) 3r(2r)
(r
2
+ 1)
2
=
1 3r
2
(r
2
+ 1)
2
= 0 C r = 1, but 31 is not in [0. 2]. 1(0) = 0,
1(1) =
1
2
, 1(2) =
2
5
. So 1(1) =
1
2
is the absolute maximum value and 1(0) = 0 is the absolute minimum value.
55. 1(t) = t
I
4 3t
2
, [31. 2].
1
0
(t) = t
1
2
(4 3t
2
)
31'2
(32t) + (4 3t
2
)
1'2
1 =
3t
2
I
4 3t
2
+
I
4 3t
2
=
3t
2
+ (4 3t
2
)
I
4 3t
2
=
4 32t
2
I
4 3t
2
.
1
0
(t) = 0 i 4 32t
2
= 0 i t
2
= 2 i t =
I
2, but t = 3
I
2 is not in the given interval, [31. 2].
1
0
(t) does not exist if 4 3t
2
= 0 i t = 2, but 32 is not in the given interval. 1(31) = 3
I
3, 1
I
2
= 2, and
1(2) = 0. So 1
I
2
= 2 is the absolute maximum value and 1(31) = 3

I
3 is the absolute minimum value.
57. 1(t) = 2 cos t + sin2t, [0, 2].
1
0
(t) = 32 sint + cos 2t 2 = 32 sint + 2(1 32 sin
2
t) = 32(2 sin
2
t + sint 31) = 32(2 sint 31)(sint + 1).
1
0
(t) = 0 i sint =
1
2
or sint = 31 i t =
r
6
. 1(0) = 2, 1(
r
6
) =
I
3 +
1
2
I
3 =
3
2
I
3 r 2.60, and 1(
r
2
) = 0.
So 1(
r
6
) =
3
2
I
3 is the absolute maximum value and 1(
r
2
) = 0 is the absolute minimum value.
59. 1(r) = rc
3i
2
'8
, [31. 4]. 1
0
(r) = r c
3i
2
'8
(3
i
4
) +c
3i
2
'8
1 = c
3i
2
'8
(3
i
2
4
+ 1). Since c
3i
2
'8
is never 0,
1
0
(r) = 0 i 3r
2
4 + 1 = 0 i 1 = r
2
4 i r
2
= 4 i r = 2, but 32 is not in the given interval, [31. 4].
1(31) = 3c
31'8
E 30.88, 1(2) = 2c
31'2
E 1.21, and 1(4) = 4c
32
E 0.54. So 1(2) = 2c
31'2
is the absolute maximum
value and 1(31) = 3c
31'8
is the absolute minimum value.
61. 1(r) = ln(r
2
+r + 1), [31. 1]. 1
0
(r) =
1
r
2
+r + 1
(2r + 1) = 0 C r = 3
1
2
. Since r
2
+r + 1 0 for all r, the
domain of 1 and 1
0
is R. 1(31) = ln1 = 0, 1
3
1
2
= ln
3
4
r 30.29, and 1(1) = ln3 r 1.10. So 1(1) = ln3 r 1.10 is
the absolute maximum value and 1
3
1
2
= ln
3
4
r 30.29 is the absolute minimum value.
SECTION 4.1 MAXIMUM AND MINIMUM VALUES

151
63. 1(r) = r
a
(1 3r)
l
, 0 $ r $ 1, o 0, / 0.
1
0
(r) = r
a
/(1 3r)
l31
(31) + (1 3r)
l
or
a31
= r
a31
(1 3r)
l31
[r /(31) + (1 3r) o]
= r
a31
(1 3r)
l31
(o 3or 3/r)
At the endpoints, we have 1(0) = 1(1) = 0 [the minimum value of 1 ]. In the interval (0. 1), 1
0
(r) = 0 C r =
o
o +/
.
1
o
o +/
=

o
o +/
1 3
o
o +/
l
=
o
a
(o +/)
a
o +/ 3o
o +/
l
=
o
a
(o +/)
a

/
l
(o +/)
l
=
o
a
/
l
(o +/)
a+l
.
So 1
o
o +/
=
o
a
/
l
(o +/)
a+l
is the absolute maximum value.
65. (a) From the graph, it appears that the absolute maximum value is about
1(30.77) = 2.19, and the absolute minimum value is about
1(0.77) = 1.81.
(b) 1(r) = r
5
3r
3
+ 2 i 1
0
(r) = 5r
4
33r
2
= r
2
(5r
2
33). So 1
0
(r) = 0 i r = 0,
3
5
.
1
3
5
=

3
3
5
5
3
3
5
3
+ 2 = 3
3
5
3
5
+
3
5
3
5
+ 2 =

3
5
3
9
25
3
5
+ 2 =
6
25
3
5
+ 2 (maximum)
and similarly, 1
3
5
= 3
6
25
3
5
+ 2 (minimum).
67. (a) From the graph, it appears that the absolute maximum value is about
1(0.75) = 0.32, and the absolute minimum value is 1(0) = 1(1) = 0;
that is, at both endpoints.
(b) 1(r) = r
I
r 3r
2
i 1
0
(r) = r
1 32r
2
I
r 3r
2
+
I
r 3r
2
=
(r 32r
2
) + (2r 32r
2
)
2
I
r 3r
2
=
3r 34r
2
2
I
r 3r
2
.
So 1
0
(r) = 0 i 3r 34r
2
= 0 i r(3 34r) = 0 i r = 0 or
3
4
.
1(0) = 1(1) = 0 (minimum), and 1
3
4
=
3
4
3
4
3
3
4
2
=
3
4
3
16
=
3
I
3
16
(maximum).
69. The density is dened as a =
mass
volume
=
1000
\ (T)
(in gcm
3
). But a critical point of a will also be a critical point of \
[since
da
dT
= 31000\
32
d\
dT
and \ is never 0], and \ is easier to differentiate than a.
\ (T) = 999.87 30.06426T + 0.0085043T
2
30.0000679T
3
i \
0
(T) = 30.06426 + 0.0170086T 30.0002037T
2
.
Setting this equal to 0 and using the quadratic formula to nd T, we get
T =
30.0170086
I
0.0170086
2
34 0.0002037 0.06426
2(30.0002037)
E 3.9665
C or 79.5318
C. Since we are only interested

152

in the region 0
C $ T $ 30
C, we check the density a at the endpoints and at 3.9665
C: a(0) E
1000
999.87
E 1.00013;
a(30) E
1000
1003.7628
E 0.99625; a(3.9665) E
1000
999.7447
E 1.000255. So water has its maximum density at
about 3.9665
C.
71. Let o = 30.000 032 37, / = 0.000 903 7, c = 30.008 956, d = 0.03629, c = 30.04458, and 1 = 0.4074.
Then o(t) = ot
5
+/t
4
+ct
3
+dt
2
+ct +1 and o
0
(t) = 5ot
4
+ 4/t
3
+ 3ct
2
+ 2dt +c.
We now apply the Closed Interval Method to the continuous function o on the interval 0 $ t $ 10. Since o
0
exists for all t,
the only critical numbers of o occur when o
0
(t) = 0. We use a rootnder on a CAS (or a graphing device) to nd that
o
0
(t) = 0 when t
1
r 0.855, t
2
r 4.618, t
3
r 7.292, and t
4
r 9.570. The values of o at these critical numbers are
o(t
1
) r 0.39, o(t
2
) r 0.43645, o(t
3
) r 0.427, and o(t
4
) r 0.43641. The values of o at the endpoints of the interval are
o(0) r 0.41 and o(10) r 0.435. Comparing the six numbers, we see that sugar was most expensive at t
2
r 4.618
(corresponding roughly to March 1998) and cheapest at t
1
r 0.855 (June 1994).
73. (a) (v) = I(v
0
3v)v
2
= Iv
0
v
2
3Iv
3
i
0
(v) = 2Iv
0
v 33Iv
2
.
0
(v) = 0 i Iv(2v
0
33v) = 0 i
v = 0 or
2
3
v
0
(but 0 is not in the interval). Evaluating at
1
2
v
0
,
2
3
v
0
, and v
0
, we get
1
2
v
0
=
1
8
Iv
3
0
,
2
3
v
0
=
4
27
Iv
3
0
,
and (v
0
) = 0. Since
4
27

1
8
, attains its maximum value at v =
2
3
v
0
. This supports the statement in the text.
(b) From part (a), the maximum value of is
4
27
Iv
3
0
.
(c)
75. 1(r) = r
101
+r
51
+r + 1 i 1
0
(r) = 101r
100
+ 51r
50
+ 1 D 1 for all r, so 1
0
(r) = 0 has no solution. Thus, 1(r)
has no critical number, so 1(r) can have no local maximum or minimum.
77. If 1 has a local minimum at c, then o(r) = 31(r) has a local maximum at c, so o
0
(c) = 0 by the case of Fermats Theorem
proved in the text. Thus, 1
0
(c) = 3o
0
(c) = 0.
4.2 The Mean Value Theorem
1. 1(r) = 5 312r + 3r
2
, [1, 3]. Since 1 is a polynomial, it is continuous and differentiable on R, so it is continuous on [1. 3]
and differentiable on (1. 3). Also 1(1) = 34 = 1(3). 1
0
(c) = 0 C 312 + 6c = 0 C c = 2, which is in the open
interval (1. 3), so c = 2 satises the conclusion of Rolles Theorem.
SECTION 4.2 THE MEAN VALUE THEOREM

153
3. 1(r) =
I
r 3
1
3
r, [0. 9]. 1, being the difference of a root function and a polynomial, is continuous and differentiable
on [0. "), so it is continuous on [0. 9] and differentiable on (0. 9). Also, 1(0) = 0 = 1(9). 1
0
(c) = 0 C
1
2
I
c
3
1
3
= 0 C 2
I
c = 3 C
I
c =
3
2
i c =
9
4
, which is in the open interval (0. 9), so c =
9
4
satises the
conclusion of Rolles Theorem.
5. 1(r) = 1 3r
2'3
. 1(31) = 1 3(31)
2'3
= 1 31 = 0 = 1(1). 1
0
(r) = 3
2
3
r
31'3
, so 1
0
(c) = 0 has no solution. This
does not contradict Rolles Theorem, since 1
0
(0) does not exist, and so 1 is not differentiable on (31. 1).
7.
1(8) 31(0)
8 30
=
6 34
8
=
1
4
. The values of c which satisfy 1
0
(c) =
1
4
seem to be about c = 0.8, 3.2, 4.4, and 6.1.
9. (a), (b) The equation of the secant line is
j 35 =
8.5 35
8 31
(r 31) C j =
1
2
r +
9
2
.
(c) 1(r) = r + 4r i 1
0
(r) = 1 34r
2
.
So 1
0
(c) =
1
2
i c
2
= 8 i c = 2
I
2, and
1(c) = 2
I
2 +
4
2
I
2
= 3
I
2. Thus, an equation of the
tangent line is j 33
I
2 =
1
2
r 32
I
2
C
j =
1
2
r + 2
I
2.
11. 1(r) = 3r
2
+ 2r + 5, [31. 1]. 1 is continuous on [31. 1] and differentiable on (31. 1) since polynomials are continuous
and differentiable on R. 1
0
(c) =
1(/) 31(o)
/ 3o
C 6c + 2 =
1(1) 31(31)
1 3(31)
=
10 36
2
= 2 C 6c = 0 C
c = 0, which is in (31. 1).
13. 1(r) = c
32i
, [0. 3]. 1 is continuous and differentiable on R, so it is continuous on [0. 3] and differentiable on (0. 3).
1
0
(c) =
1(/) 31(o)
/ 3o
C 32c
32c
=
c
36
3c
0
3 30
C c
32c
=
1 3c
36
6
C 32c = ln
1 3c
36
6
C
c = 3
1
2
ln
1 3c
36
6
E 0.897, which is in (0. 3).

154

15. 1(r) = (r 33)
32
i 1
0
(r) = 32 (r 33)
33
. 1(4) 31(1) = 1
0
(c)(4 31) i
1
1
2
3
1
(32)
2
=
32
(c 33)
3
3 i
3
4
=
36
(c 33)
3
i (c 33)
3
= 38 i c 33 = 32 i c = 1, which is not in the open interval (1. 4). This does not
contradict the Mean Value Theorem since 1 is not continuous at r = 3.
17. Let 1(r) = 1 + 2r +r
3
+ 4r
5
. Then 1(31) = 36 < 0 and 1(0) = 1 0. Since 1 is a polynomial, it is continuous, so the
Intermediate Value Theorem says that there is a number c between 31 and 0 such that 1(c) = 0. Thus, the given equation has
a real root. Suppose the equation has distinct real roots o and / with o < /. Then 1(o) = 1(/) = 0. Since 1 is a polynomial, it
is differentiable on (o. /) and continuous on [o. /]. By Rolles Theorem, there is a number v in (o. /) such that 1
0
(v) = 0. But
1
0
(r) = 2 + 3r
2
+ 20r
4
D 2 for all r, so 1
0
(r) can never be 0. This contradiction shows that the equation cant have two
distinct real roots. Hence, it has exactly one real root.
19. Let 1(r) = r
3
315r +c for r in [32. 2]. If 1 has two real roots o and / in [32. 2], with o < /, then 1(o) = 1(/) = 0. Since
the polynomial 1 is continuous on [o. /] and differentiable on (o. /), Rolles Theorem implies that there is a number v in (o. /)
such that 1
0
(v) = 0. Now 1
0
(v) = 3v
2
315. Since v is in (o. /), which is contained in [32. 2], we have |v| < 2, so v
2
< 4.
It follows that 3v
2
315 < 3 4 315 = 33 < 0. This contradicts 1
0
(v) = 0, so the given equation cant have two real roots
in [32. 2]. Hence, it has at most one real root in [32. 2].
21. (a) Suppose that a cubic polynomial 1(r) has roots o
1
< o
2
< o
3
< o
4
, so 1(o
1
) = 1(o
2
) = 1(o
3
) = 1(o
4
).
By Rolles Theorem there are numbers c
1
, c
2
, c
3
with o
1
< c
1
< o
2
, o
2
< c
2
< o
3
and o
3
< c
3
< o
4
and
1
0
(c
1
) = 1
0
(c
2
) = 1
0
(c
3
) = 0. Thus, the second-degree polynomial 1
0
(r) has three distinct real roots, which is
impossible.
(b) We prove by induction that a polynomial of degree n has at most n real roots. This is certainly true for n = 1. Suppose
that the result is true for all polynomials of degree n and let 1(r) be a polynomial of degree n + 1. Suppose that 1(r) has
more than n + 1 real roots, say o
1
< o
2
< o
3
< < o
n+1
< o
n+2
. Then 1(o
1
) = 1(o
2
) = = 1(o
n+2
) = 0.
By Rolles Theorem there are real numbers c
1
. . . . . c
n+1
with o
1
< c
1
< o
2
. . . . . o
n+1
< c
n+1
< o
n+2
and
1
0
(c
1
) = = 1
0
(c
n+1
) = 0. Thus, the nth degree polynomial 1
0
(r) has at least n + 1 roots. This contradiction shows
that 1(r) has at most n + 1 real roots.
23. By the Mean Value Theorem, 1(4) 31(1) = 1
0
(c)(4 31) for some c M (1. 4). But for every c M (1. 4) we have
1
0
(c) D 2. Putting 1
0
(c) D 2 into the above equation and substituting 1(1) = 10, we get
1(4) = 1(1) +1
0
(c)(4 31) = 10 + 31
0
(c) D 10 + 3 2 = 16. So the smallest possible value of 1(4) is 16.
25. Suppose that such a function 1 exists. By the Mean Value Theorem there is a number 0 < c < 2 with
1
0
(c) =
1(2) 31(0)
2 30
=
5
2
. But this is impossible since 1
0
(r) $ 2 <
5
2
for all r, so no such function can exist.
SECTION 4.3 HOW DERIVATIVES AFFECT THE SHAPE OF A GRAPH

155
27. We use Exercise 26 with 1(r) =
I
1 +r, o(r) = 1 +
1
2
r, and o = 0. Notice that 1(0) = 1 = o(0) and
1
0
(r) =
1
2
I
1 +r
<
1
2
= o
0
(r) for r 0. So by Exercise 26, 1(/) < o(/) i
I
1 +/ < 1 +
1
2
/ for / 0.
Another method: Apply the Mean Value Theorem directly to either 1(r) = 1 +
1
2
r 3
I
1 +r or o(r) =
I
1 +r on [0. /].
29. Let 1(r) = sinr and let / < o. Then 1(r) is continuous on [/. o] and differentiable on (/. o). By the Mean Value Theorem,
there is a number c M (/. o) with sino 3sin/ = 1(o) 31(/) = 1
0
(c)(o 3/) = (cos c)(o 3/). Thus,
|sino 3sin/| $ |cos c| |/ 3o| $ |o 3/|. If o < /, then |sino 3sin/| = |sin/ 3sino| $ |/ 3o| = |o 3/|. If o = /, both
sides of the inequality are 0.
31. For r 0, 1(r) = o(r), so 1
0
(r) = o
0
(r). For r < 0, 1
0
(r) = (1r)
0
= 31r
2
and o
0
(r) = (1 + 1r)
0
= 31r
2
, so
again 1
0
(r) = o
0
(r). However, the domain of o(r) is not an interval [it is (3". 0) (0. ")] so we cannot conclude that
1 3o is constant (in fact it is not).
33. Let 1(r) = arcsin
r 31
r + 1
32 arctan
I
r +
r
2
. Note that the domain of 1 is [0. "). Thus,
1
0
(r) =
1
1 3
r 31
r + 1
2
(r + 1) 3(r 31)
(r + 1)
2
3
2
1 +r

1
2
I
r
=
1
I
r(r + 1)
3
1
I
r(r + 1)
= 0.
Then 1(r) = C on (0. ") by Theorem 5. By continuity of 1, 1(r) = C on [0. "). To nd C, we let r = 0 i
arcsin(31) 32 arctan(0) +
r
2
= C i 3
r
2
30 +
r
2
= 0 = C. Thus, 1(r) = 0 i
arcsin
r 31
r + 1
= 2 arctan
I
r 3
r
2
.
35. Let o(t) and /(t) be the position functions of the two runners and let 1(t) = o(t) 3/(t). By hypothesis,
1(0) = o(0) 3/(0) = 0 and 1(/) = o(/) 3/(/) = 0, where / is the nishing time. Then by the Mean Value Theorem,
there is a time c, with 0 < c < /, such that 1
0
(c) =
1(/) 31(0)
/ 30
. But 1(/) = 1(0) = 0, so 1
0
(c) = 0. Since
1
0
(c) = o
0
(c) 3/
0
(c) = 0, we have o
0
(c) = /
0
(c). So at time c, both runners have the same speed o
0
(c) = /
0
(c).
4.3 How Derivatives Affect the Shape of a Graph
1. (a) 1 is increasing on (1. 3) and (4. 6). (b) 1 is decreasing on (0. 1) and (3. 4).
(c) 1 is concave upward on (0. 2). (d) 1 is concave downward on (2. 4) and (4. 6).
(e) The point of inection is (2. 3).
3. (a) Use the Increasing/Decreasing (I/D) Test. (b) Use the Concavity Test.
(c) At any value of r where the concavity changes, we have an inection point at (r. 1(r)).
156

5. (a) Since 1
0
(r) 0 on (1. 5), 1 is increasing on this interval. Since 1
0
(r) < 0 on (0. 1) and (5. 6), 1 is decreasing on these
intervals.
(b) Since 1
0
(r) = 0 at r = 1 and 1
0
changes from negative to positive there, 1 changes from decreasing to increasing and has
a local minimum at r = 1. Since 1
0
(r) = 0 at r = 5 and 1
0
changes from positive to negative there, 1 changes from
increasing to decreasing and has a local maximum at r = 5.
7. There is an inection point at r = 1 because 1
00
(r) changes from negative to positive there, and so the graph of 1 changes
from concave downward to concave upward. There is an inection point at r = 7 because 1
00
(r) changes from positive to
negative there, and so the graph of 1 changes from concave upward to concave downward.
9. (a) 1(r) = 2r
3
+ 3r
2
336r i 1
0
(r) = 6r
2
+ 6r 336 = 6(r
2
+r 36) = 6(r + 3)(r 32).
We dont need to include the 6 in the chart to determine the sign of 1
0
(r).
Interval r + 3 r 32 1
0
(r) 1
r < 33 3 3 + increasing on (3". 33)
33 < r < 2 + 3 3 decreasing on (33. 2)
r 2 + + + increasing on (2. ")
(b) 1 changes from increasing to decreasing at r = 33 and from decreasing to increasing at r = 2. Thus, 1(33) = 81 is a
local maximum value and 1(2) = 344 is a local minimum value.
(c) 1
0
(r) = 6r
2
+ 6r 336 i 1
00
(r) = 12r + 6. 1
00
(r) = 0 at r = 3
1
2
, 1
00
(r) 0 C r 3
1
2
, and
1
00
(r) < 0 C r < 3
1
2
. Thus, 1 is concave upward on
3
1
2
. "
and concave downward on
3". 3
1
2
. There is an
inection point at
3
1
2
. 1
3
1
2
=

3
1
2
.
37
2
.
11. (a) 1(r) = r
4
32r
2
+ 3 i 1
0
(r) = 4r
3
34r = 4r
r
2
31
= 4r(r + 1)(r 31).

Interval r + 1 r r 31 1
0
(r) 1
r < 31 3 3 3 3 decreasing on (3". 31)
31 < r < 0 + 3 3 + increasing on (31. 0)
0 < r < 1 + + 3 3 decreasing on (0. 1)
r 1 + + + + increasing on (1. ")
(b) 1 changes from increasing to decreasing at r = 0 and from decreasing to increasing at r = 31 and r = 1. Thus,
1(0) = 3 is a local maximum value and 1(1) = 2 are local minimum values.
(c) 1
00
(r) = 12r
2
34 = 12
r
2
3
1
3
= 12
r + 1
I
3
r 31
I
3
. 1
00
(r) 0 C r < 31
I
3 or r 1
I
3 and
1
00
(r) < 0 C 31
I
3 < r < 1
I
3. Thus, 1 is concave upward on
3". 3
I
33
and
I
33. "
and concave
downward on
3
I
33.
I
33
. There are inection points at
I
33.
22
9
.

157
13. (a) 1(r) = sinr + cos r, 0 $ r $ 2. 1
0
(r) = cos r 3sinr = 0 i cos r = sinr i 1 =
sinr
cos r
i
tanr = 1 i r =
r
4
or
5r
4
. Thus, 1
0
(r) 0 C cos r 3sinr 0 C cos r sinr C 0 < r <
r
4
or
5r
4
< r < 2 and 1
0
(r) < 0 C cos r < sinr C
r
4
< r <
5r
4
. So 1 is increasing on
0.
r
4
and
5r
4
. 2
and 1
is decreasing on
r
4
.
5r
4
.
(b) 1 changes from increasing to decreasing at r =
r
4
and from decreasing to increasing at r =
5r
4
. Thus, 1
r
4
=
I
2 is a
local maximum value and 1
5r
4
= 3
I
2 is a local minimum value.
(c) 1
00
(r) = 3sinr 3cos r = 0 i 3sinr = cos r i tanr = 31 i r =
3r
4
or
7r
4
. Divide the interval
(0. 2) into subintervals with these numbers as endpoints and complete a second derivative chart.
Interval 1
00
(r) = 3sinr 3cos r Concavity
0.
3r
4
1
00
r
2
= 31 < 0 downward
3r
4
.
7r
4
1
00
() = 1 0 upward
7r
4
. 2
1
00
11r
6
=
1
2
3
1
2
I
3 < 0 downward
There are inection points at
3r
4
. 0
and
7r
4
. 0
.
15. (a) 1(r) = c
2i
+c
3i
i 1
0
(r) = 2c
2i
3c
3i
. 1
0
(r) 0 C 2c
2i
c
3i
C c
3i
1
2
C 3r ln
1
2
C
r
1
3
(ln1 3ln2) C r 3
1
3
ln2 [r 30.23] and 1
0
(r) < 0 if r < 3
1
3
ln2. So 1 is increasing on
3
1
3
ln2. "
and 1 is decreasing on
3". 3
1
3
ln2
.
(b) 1 changes from decreasing to increasing at r = 3
1
3
ln2. Thus,
1
3
1
3
ln2
= 1
ln
3
12
= c
2 ln
3
I
1'2
+c
3ln
3
I
1'2
= c
ln
3
I
1'4
+c
ln
3
I
2
=
3
14 +
3
I
2 = 2
32'3
+ 2
1'3
[r 1.89]
is a local minimum value.
(c) 1
00
(r) = 4c
2i
+c
3i
0 [the sum of two positive terms]. Thus, 1 is concave upward on (3". ") and there is no
point of inection.
17. (a) j = 1(r) =
lnr
I
r
. (Note that 1 is only dened for r 0.)
1
0
(r) =
I
r(1r) 3lnr
1
2
r
31'2
r
=
1
I
r
3
lnr
2
I
r
r

2
I
r
2
I
r
=
2 3lnr
2r
3'2
0 C 2 3lnr 0 C
lnr < 2 C r < c
2
. Therefore 1 is increasing on
0. c
2
and decreasing on
c
2
. "
.
(b) 1 changes from increasing to decreasing at r = c
2
, so 1(c
2
) =
lnc
2
I
c
2
=
2
c
is a local maximum value.
(c) 1
00
(r) =
2r
3'2
(31r) 3(2 3lnr)(3r
1'2
)
(2r
3'2
)
2
=
32r
1'2
+ 3r
1'2
(lnr 32)
4r
3
=
r
1'2
(32 + 3 lnr 36)
4r
3
=
3 lnr 38
4r
5'2
1
00
(r) = 0 C lnr =
8
3
C r = c
8'3
. 1
00
(r) 0 C r c
8'3
, so 1 is concave upward on (c
8'3
. ") and
concave downward on (0. c
8'3
). There is an inection point at
c
8'3
.
8
3
c
34'3
E (14.39. 0.70).
158

19. 1(r) = r
5
35r + 3 i 1
0
(r) = 5r
4
35 = 5(r
2
+ 1)(r + 1)(r 31).
First Derivative Test: 1
0
(r) < 0 i 31 < r < 1 and 1
0
(r) 0 i r 1 or r < 31. Since 1
0
changes from
positive to negative at r = 31, 1(31) = 7 is a local maximum value; and since 1
0
changes from negative to positive at r = 1,
1(1) = 31 is a local minimum value.
Second Derivative Test: 1
00
(r) = 20r
3
. 1
0
(r) = 0 C r = 1. 1
00
(31) = 320 < 0 i 1(31) = 7 is a local
maximum value. 1
00
(1) = 20 0 i 1(1) = 31 is a local minimum value.
Preference: For this function, the two tests are equally easy.
21. 1(r) = r +
I
1 3r i 1
0
(r) = 1 +
1
2
(1 3r)
31'2
(31) = 1 3
1
2
I
1 3r
. Note that 1 is dened for 1 3r D 0; that is,
for r $ 1. 1
0
(r) = 0 i 2
I
1 3r = 1 i
I
1 3r =
1
2
i 1 3r =
1
4
i r =
3
4
. 1
0
does not exist at r = 1,
but we cant have a local maximum or minimum at an endpoint.
First Derivative Test: 1
0
(r) 0 i r <
3
4
and 1
0
(r) < 0 i
3
4
< r < 1. Since 1
0
changes from positive to
negative at r =
3
4
, 1
3
4
=
5
4
Second Derivative Test: 1
00
(r) = 3
1
2
3
1
2
(1 3r)
33'2
(31) = 3
1
4
I
1 3r
3
.
1
00
3
4
= 32 < 0 i 1
3
4
=
5
4
Preference: The First Derivative Test may be slightly easier to apply in this case.
23. (a) By the Second Derivative Test, if 1
0
(2) = 0 and 1
00
(2) = 35 < 0, 1 has a local maximum at r = 2.
(b) If 1
0
(6) = 0, we know that 1 has a horizontal tangent at r = 6. Knowing that 1
00
(6) = 0 does not provide any additional
information since the Second Derivative Test fails. For example, the rst and second derivatives of j = (r 36)
4
,
j = 3(r 36)
4
, and j = (r 36)
3
all equal zero for r = 6, but the rst has a local minimum at r = 6, the second has a
local maximum at r = 6, and the third has an inection point at r = 6.
25. 1
0
(0) = 1
0
(2) = 1
0
(4) = 0 i horizontal tangents at r = 0, 2, 4.
1
0
(r) 0 if r < 0 or 2 < r < 4 i 1 is increasing on (3". 0) and (2. 4).
1
0
(r) < 0 if 0 < r < 2 or r 4 i 1 is decreasing on (0. 2) and (4. ").
1
00
(r) 0 if 1 < r < 3 i 1 is concave upward on (1. 3).
1
00
(r) < 0 if r < 1 or r 3 i 1 is concave downward on (3". 1) and (3. ").
There are inection points when r = 1 and 3.
27. 1
0
(r) 0 if |r| < 2 i 1 is increasing on (32. 2).
1
0
(r) < 0 if |r| 2 i 1 is decreasing on (3". 32)
and (2. "). 1
0
(32) = 0 i horizontal tangent at r = 32.
lim
i<2
|1
0
(r)| = " i there is a vertical asymptote or
vertical tangent (cusp) at r = 2. 1
00
(r) 0 if r 6= 2 i
1 is concave upward on (3". 2) and (2. ").

159
29. The function must be always decreasing (since the rst derivative is always negative)
and concave downward (since the second derivative is always negative).
31. (a) 1 is increasing where 1
0
is positive, that is, on (0. 2), (4. 6), and (8. "); and decreasing where 1
0
is negative, that is,
on (2. 4) and (6. 8).
(b) 1 has local maxima where 1
0
changes from positive to negative, at r = 2 and at r = 6, and local minima where 1
0
changes
from negative to positive, at r = 4 and at r = 8.
(c) 1 is concave upward (CU) where 1
0
is increasing, that is, on (3. 6) and (6. "),
and concave downward (CD) where 1
0
is decreasing, that is, on (0. 3).
(e)
(d) There is a point of inection where 1 changes from being CD to being CU, that
is, at r = 3.
33. (a) 1(r) = 2r
3
33r
2
312r i 1
0
(r) = 6r
2
36r 312 = 6(r
2
3r 32) = 6(r 32)(r + 1).
1
0
(r) 0 C r < 31 or r 2 and 1
0
(r) < 0 C 31 < r < 2. So 1 is increasing on (3". 31) and (2. "),
and 1 is decreasing on (31. 2).
(b) Since 1 changes from increasing to decreasing at r = 31, 1(31) = 7 is a local
maximum value. Since 1 changes from decreasing to increasing at r = 2,
1(2) = 320 is a local minimum value.
(d)
(c) 1
00
(r) = 6(2r 31) i 1
00
(r) 0 on
1
2
. "
and 1
00
(r) < 0 on
3".
1
2
.
So 1 is concave upward on
1
2
. "
and concave downward on
3".
1
2
. There
is a change in concavity at r =
1
2
, and we have an inection point at
1
2
. 3
13
2
.
35. (a) 1(r) = 2 + 2r
2
3r
4
i 1
0
(r) = 4r 34r
3
= 4r(1 3r
2
) = 4r(1 +r)(1 3r). 1
0
(r) 0 C r < 31 or
0 < r < 1 and 1
0
(r) < 0 C 31 < r < 0 or r 1. So 1 is increasing on (3". 31) and (0. 1) and 1 is decreasing
on (31. 0) and (1. ").
(b) 1 changes from increasing to decreasing at r = 31 and r = 1, so 1(31) = 3 and 1(1) = 3 are local maximum values.
1 changes from decreasing to increasing at r = 0, so 1(0) = 2 is a local minimum value.
160

(c) 1
00
(r) = 4 312r
2
= 4(1 33r
2
). 1
00
(r) = 0 C 1 33r
2
= 0 C
r
2
=
1
3
C r = 1
I
3. 1
00
(r) 0 on
31
I
3. 1
I
3
and 1
00
(r) < 0
on
3". 31
I
3
and
1
I
3. "
. So 1 is concave upward on
31
I
3. 1
I
3
and 1 is concave downward on
3". 31
I
3
and
1
I
3. "
. 1
1
I
3
= 2 +
2
3
3
1
9
=
23
9
. There are points of inection
at
1
I
3.
23
9
.
(d)
37. (a) /(r) = (r + 1)
5
35r 32 i /
0
(r) = 5(r + 1)
4
35. /
0
(r) = 0 C 5(r + 1)
4
= 5 C (r + 1)
4
= 1 i
(r + 1)
2
= 1 i r + 1 = 1 or r + 1 = 31 i r = 0 or r = 32. /
0
(r) 0 C r < 32 or r 0 and
/
0
(r) < 0 C 32 < r < 0. So / is increasing on (3". 32) and (0. ") and / is decreasing on (32. 0).
(b) /(32) = 7 is a local maximum value and /(0) = 31 is a local minimum value. (d)
(c) /
00
(r) = 20(r + 1)
3
= 0 C r = 31. /
00
(r) 0 C r 31 and
/
00
(r) < 0 C r < 31, so / is CU on (31. ") and / is CD on (3". 31).
There is a point of inection at (31. /(31)) = (31. 3).
39. (a) (r) = r
I
r + 3 i
0
(r) = r
1
2
(r+3)
31'2
+
I
r + 3 1 =
r
2
I
r + 3
+
I
r + 3 =
r + 2(r + 3)
2
I
r + 3
=
3r + 6
2
I
r + 3
.
The domain of is [33. ").
0
(r) 0 for r 32 and
0
(r) < 0 for 33 < r < 32, so is increasing on (32. ")
and decreasing on (33. 32).
(b) (32) = 32 is a local minimum value. (d)
(c)
00
(r) =
2
I
r + 3 3 3(3r + 6)
1
I
r + 3
2
I
r + 3
2
=
6(r + 3) 3(3r + 6)
4(r + 3)
3'2
=
3r + 12
4(r + 3)
3'2
=
3(r + 4)
4(r + 3)
3'2
00
(r) 0 for all r 33, so is concave upward on (33. "). There is no inection point.
41. (a) C(r) = r
1'3
(r + 4) = r
4'3
+ 4r
1'3
i C
0
(r) =
4
3
r
1'3
+
4
3
r
32'3
=
4
3
r
32'3
(r + 1) =
4(r + 1)
3
3
I
r
2
. C
0
(r) 0 if
31 < r < 0 or r 0 and C
0
(r) < 0 for r < 31, so C is increasing on (31. ") and C is decreasing on (3". 31).

161
(b) C(31) = 33 is a local minimum value. (d)
(c) C
00
(r) =
4
9
r
32'3
3
8
9
r
35'3
=
4
9
r
35'3
(r 32) =
4(r 32)
9
3
I
r
5
.
C
00
(r) < 0 for 0 < r < 2 and C
00
(r) 0 for r < 0 and r 2, so C is
concave downward on (0. 2) and concave upward on (3". 0) and (2. ").
There are inection points at (0. 0) and
2. 6
3
I
2
E (2. 7.56).
43. (a) 1(0) = 2 cos 0 + cos
2
0, 0 $ 0 $ 2 i 1
0
(0) = 32 sin0 + 2 cos 0 (3sin0) = 32 sin0 (1 + cos 0).
1
0
(0) = 0 C 0 = 0. . and 2. 1
0
(0) 0 C < 0 < 2 and 1
0
(0) < 0 C 0 < 0 < . So 1 is increasing on
(. 2) and 1 is decreasing on (0. ).
(b) 1() = 31 is a local minimum value.
(c) 1
0
(0) = 32 sin0 (1 + cos 0) i
1
00
(0) = 32 sin0 (3sin0) + (1 + cos 0)(32 cos 0) = 2 sin
2
0 32 cos 0 32 cos
2
0
= 2(1 3cos
2
0) 32 cos 0 32 cos
2
0 = 34 cos
2
0 32 cos 0 + 2
= 32(2 cos
2
0 + cos 0 31) = 32(2 cos 0 31)(cos 0 + 1)
Since 32(cos 0 + 1) < 0 [for 0 6= ], 1
00
(0) 0 i 2 cos 0 31 < 0 i cos 0 <
1
2
i
r
3
< 0 <
5r
3
and
1
00
(0) < 0 i cos 0
1
2
i 0 < 0 <
r
3
or
5r
3
< 0 < 2. So 1 is CU on
r
3
.
5r
3
and 1 is CD on
0.
r
3
and
5r
3
. 2
. There are points of inection at
r
3
. 1
r
3
=

r
3
.
5
4
and
5r
3
. 1
5r
3
=

5r
3
.
5
4
.
(d)
45. 1(r) =
r
2
r
2
31
=
r
2
(r + 1)(r 31)
has domain (3". 31) (31. 1) (1. ").
(a) lim
i<"
1(r) = lim
i<"
r
2
r
2
(r
2
31)r
2
= lim
i<"
1
1 31r
2
=
1
1 30
= 1, so j = 1 is a HA.
lim
i<31
r
2
r
2
31
= "since r
2
<1 and (r
2
31) <0
+
as r <31
3
, so r = 31 is a VA.
lim
i<1
+
r
2
r
2
31
= "since r
2
<1 and (r
2
31) <0
+
as r <1
+
, so r = 1 is a VA.
(b) 1(r) =
r
2
r
2
31
i 1
0
(r) =
(r
2
31)(2r) 3r
2
(2r)
(r
2
31)
2
=
2r[(r
2
31) 3r
2
]
(r
2
31)
2
=
32r
(r
2
31)
2
. Since (r
2
31)
2
is
positive for all r in the domain of 1, the sign of the derivative is determined by the sign of 32r. Thus, 1
0
(r) 0 if r < 0
(r 6= 31) and 1
0
(r) < 0 if r 0 (r 6= 1). So 1 is increasing on (3". 31) and (31. 0), and 1 is decreasing on (0. 1)
and (1. ").
(c) 1
0
(r) = 0 i r = 0 and 1(0) = 0 is a local maximum value.
162

(d) 1
00
(r) =
(r
2
31)
2
(32) 3(32r) 2(r
2
31)(2r)
[(r
2
31)
2
]
2
=
2(r
2
31)[3(r
2
31) + 4r
2
]
(r
2
31)
4
=
2(3r
2
+ 1)
(r
2
31)
3
.
The sign of 1
00
(r) is determined by the denominator; that is, 1
00
(r) 0 if
|r| 1 and 1
00
(r) < 0 if |r| < 1. Thus, 1 is CU on (3". 31) and (1. "),
and 1 is CD on (31. 1). There are no inection points.
(e)
47. (a) lim
i<3"
I
r
2
+ 1 3r
= "and
lim
i<"
I
r
2
+ 1 3r
= lim
i<"
I
r
2
+ 1 3r
I
r
2
+ 1 +r
I
r
2
+ 1 +r
= lim
i<"
1
I
r
2
+ 1 +r
= 0, so j = 0 is a HA.
(b) 1(r) =
I
r
2
+ 1 3r i 1
0
(r) =
r
I
r
2
+ 1
31. Since
r
I
r
2
+ 1
< 1 for all r, 1
0
(r) < 0, so 1 is decreasing on R.
(c) No minimum or maximum
(d) 1
00
(r) =
(r
2
+ 1)
1'2
(1) 3r
1
2
(r
2
+ 1)
31'2
(2r)
I
r
2
+ 1
2
=
(r
2
+ 1)
1'2
3
r
2
(r
2
+ 1)
1'2
r
2
+ 1
=
(r
2
+ 1) 3r
2
(r
2
+ 1)
3'2
=
1
(r
2
+ 1)
3'2
0,
so 1 is CU on R. No IP
(e)
49. 1(r) = ln(1 3lnr) is dened when r 0 (so that lnr is dened) and 1 3lnr 0 [so that ln(1 3lnr) is dened].
The second condition is equivalent to 1 lnr C r < c. so 1 has domain (0. c).
(a) As r <0
+
. lnr <3". so 1 3lnr <"and 1(r) <". As r <c
3
. lnr <1
3
. so 1 3lnr <0
+
and
1(r) <3". Thus, r = 0 and r = c are vertical asymptotes. There is no horizontal asymptote.
(b) 1
0
(r) =
1
1 3lnr
3
1
r
= 3
1
r(1 3lnr)
< 0 on (0. c) . Thus, 1 is decreasing on its domain, (0. c) .
(c) 1
0
(r) 6= 0 on (0. c) . so 1 has no local maximum or minimum value. (e)
(d) 1
00
(r) = 3
3[r(1 3lnr)]
0
[r(1 3lnr)]
2
=
r(31r) + (1 3lnr)
r
2
(1 3lnr)
2
= 3
lnr
r
2
(1 3lnr)
2
so 1
00
(r) 0 C lnr < 0 C 0 < r < 1. Thus, 1 is CU on (0. 1)
and CD on (1. c) . There is an inection point at (1. 0) .
51. (a) lim
i<"
c
31'(i+1)
= 1 since 31(r + 1) <0, so j = 1 is a HA. lim
i<31
+
c
31'(i+1)
= 0 since 31(r + 1) <3",
lim
i<31
c
31'(i+1)
= "since 31(r + 1) <", so r = 31 is a VA.
(b) 1(r) = c
31'(i+1)
i 1
0
(r) = c
31'(i+1)
3(31)
1
(r + 1)
2
[Reciprocal Rule] = c
31'(i+1)
(r + 1)
2
i
1
0
(r) 0 for all r except 31, so 1 is increasing on (3". 31) and (31. ").
(c) There is no local maximum or minimum.

163
(d) 1
00
(r) =
(r + 1)
2
c
31'(i+1)
1(r + 1)
2
3c
31'(i+1)
[2(r + 1)]
[(r + 1)
2
]
2
=
c
31'(i+1)
[1 3(2r + 2)]
(r + 1)
4
= 3
c
31'(i+1)
(2r + 1)
(r + 1)
4
i
1
00
(r) 0 C 2r + 1 < 0 C r < 3
1
2
, so 1 is CU on (3". 31)
and
31. 3
1
2
, and CD on
3
1
2
, "
. 1 has an IP at
3
1
2
. c
32
.
(e)
53. The nonnegative factors (r + 1)
2
and (r 36)
4
do not affect the sign of 1
0
(r) = (r + 1)
2
(r 33)
5
(r 36)
4
.
So 1
0
(r) 0 i (r 33)
5
0 i r 33 0 i r 3. Thus, 1 is increasing on the interval (3. ").
55. (a) From the graph, we get an estimate of 1(1) E 1.41 as a local maximum
value, and no local minimum value.
1(r) =
r + 1
I
r
2
+ 1
i 1
0
(r) =
1 3r
(r
2
+ 1)
3'2
.
1
0
(r) = 0 C r = 1. 1(1) =
2
I
2
=
I
2 is the exact value.
(b) From the graph in part (a), 1 increases most rapidly somewhere between r = 3
1
2
and r = 3
1
4
. To nd the exact value,
we need to nd the maximum value of 1
0
, which we can do by nding the critical numbers of 1
0
.
1
00
(r) =
2r
2
33r 31
(r
2
+ 1)
5'2
= 0 C r =
3
I
17
4
. r =
3 +
I
17
4
corresponds to the minimum value of 1
0
.
The maximum value of 1
0
is at
3 3
I
17
4
.
7
6
3
I
17
6
E (30.28. 0.69).
57. 1(r) = cos r +
1
2
cos 2r i 1
0
(r) = 3sinr 3sin2r i 1
00
(r) = 3cos r 32 cos 2r
(a) From the graph of 1, it seems that 1 is CD on (0. 1), CU on (1. 2.5), CD on
(2.5. 3.7), CU on (3.7. 5.3), and CD on (5.3. 2). The points of inection
appear to be at (1. 0.4), (2.5. 30.6), (3.7. 30.6), and (5.3. 0.4).
(b) From the graph of 1
00
(and zooming in near the zeros), it seems that 1 is CD
on (0. 0.94), CU on (0.94. 2.57), CD on (2.57. 3.71), CU on (3.71. 5.35),
and CD on (5.35. 2). Rened estimates of the inection points are
(0.94. 0.44), (2.57. 30.63), (3.71. 30.63), and (5.35. 0.44).
59. In Maple, we dene 1 and then use the command
plot(diff(diff(f,x),x),x=-2..2);. In Mathematica, we dene 1
and then use Plot[Dt[Dt[f,x],x],{x,-2,2}]. We see that 1
00
0 for
r < 30.6 and r 0.0 [E 0.03] and 1
00
< 0 for 30.6 < r < 0.0. So 1 is CU
on (3". 30.6) and (0.0. ") and CD on (30.6. 0.0).
164

61. (a) The rate of increase of the population is initially very small, then gets larger until it reaches a maximum at about
t = 8 hours, and decreases toward 0 as the population begins to level off.
(b) The rate of increase has its maximum value at t = 8 hours.
(c) The population function is concave upward on (0. 8) and concave downward on (8. 18).
(d) At t = 8, the population is about 350, so the inection point is about (8. 350).
63. Most students learn more in the third hour of studying than in the eighth hour, so 1(3) 31(2) is larger than 1(8) 31(7).
In other words, as you begin studying for a test, the rate of knowledge gain is large and then starts to taper off, so 1
0
(t)
decreases and the graph of 1 is concave downward.
65. o(t) = t
c
3II
with = 0.01, j = 4, and I = 0.07. We will nd the
zeros of 1
00
for 1(t) = t
c
3II
.
1
0
(t) = t
(3Ic
3II
) +c
3II
(jt
31
) = c
3II
(3It
+jt
31
)
1
00
(t) = c
3II
(3Ijt
31
+j(j 31)t
32
) + (3It
+jt
31
)(3Ic
3II
)
= t
32
c
3II
[3Ijt +j(j 31) +I
2
t
2
3Ijt]
= t
32
c
3II
(I
2
t
2
32Ijt +j
2
3j)
Using the given values of j and I gives us 1
00
(t) = t
2
c
30.07I
(0.0049t
2
30.56t + 12). So o
00
(t) = 0.011
00
(t) and its zeros
are t = 0 and the solutions of 0.0049t
2
30.56t + 12 = 0, which are t
1
=
200
7
r 28.57 and t
2
=
600
7
r 85.71.
At t
1
minutes, the rate of increase of the level of medication in the bloodstream is at its greatest and at t
2
minutes, the rate of
decrease is the greatest.
67. 1(r) = or
3
+/r
2
+cr +d i 1
0
(r) = 3or
2
+ 2/r +c.
We are given that 1(1) = 0 and 1(32) = 3, so 1(1) = o +/ +c +d = 0 and
1(32) = 38o + 4/ 32c +d = 3. Also 1
0
(1) = 3o + 2/ +c = 0 and
1
0
(32) = 12o 34/ +c = 0 by Fermats Theorem. Solving these four equations, we get
o =
2
9
, / =
1
3
, c = 3
4
3
, d =
7
9
, so the function is 1(r) =
1
9
2r
3
+ 3r
2
312r + 7
.
69. j =
1 +r
1 +r
2
i j
0
=
(1 +r
2
)(1) 3(1 +r)(2r)
(1 +r
2
)
2
=
1 32r 3r
2
(1 +r
2
)
2
i
j
00
=
(1 +r
2
)
2
(32 32r) 3(1 32r 3r
2
) 2(1 +r
2
)(2r)
[(1 +r
2
)
2
]
2
=
2(1 +r
2
)[(1 +r
2
)(31 3r) 3(1 32r 3r
2
)(2r)]
(1 +r
2
)
4
=
2(31 3r 3r
2
3r
3
32r + 4r
2
+ 2r
3
)
(1 +r
2
)
3
=
2(r
3
+ 3r
2
33r 31)
(1 +r
2
)
3
=
2(r 31)(r
2
+ 4r + 1)
(1 +r
2
)
3
So j
00
= 0 i r = 1, 32
I
3. Let o = 32 3
I
3, / = 32 +
I
3, and c = 1. We can show that 1(o) =
1
4
1 3
I
3
,

165
1(/) =
1
4
1 +
I
3
, and 1(c) = 1. To show that these three points of inection lie on one straight line, well show that the
slopes i
ac
and i
lc
are equal.
i
ac
=
1(c) 31(o)
c 3o
=
1 3
1
4
1 3
I
3
1 3
32 3
I
3
=
3
4
+
1
4
I
3
3 +
I
3
=
1
4
i
lc
=
1(c) 31(/)
c 3/
=
1 3
1
4
1 +
I
3
1 3
32 +
I
3
=
3
4
3
1
4
I
3
3 3
I
3
=
1
4
71. Suppose that 1 is differentiable on an interval 1 and 1
0
(r) 0 for all r in 1 except r = c. To show that 1 is increasing on 1,
let r
1
, r
2
be two numbers in 1 with r
1
< r
2
.
Case 1 r
1
< r
2
< c. Let J be the interval {r M 1 | r < c}. By applying the Increasing/Decreasing Test to 1
on J, we see that 1 is increasing on J, so 1(r
1
) < 1(r
2
).
Case 2 c < r
1
< r
2
. Apply the Increasing/Decreasing Test to 1 on 1 = {r M 1 | r c}.
Case 3 r
1
< r
2
= c. Apply the proof of the Increasing/Decreasing Test, using the Mean Value Theorem (MVT)
on the interval [r
1
. r
2
] and noting that the MVT does not require 1 to be differentiable at the endpoints
of [r
1
. r
2
].
Case 4 c = r
1
< r
2
. Same proof as in Case 3.
Case 5 r
1
< c < r
2
. By Cases 3 and 4, 1 is increasing on [r
1
. c] and on [c. r
2
], so 1(r
1
) < 1(c) < 1(r
2
).
In all cases, we have shown that 1(r
1
) < 1(r
2
). Since r
1
, r
2
were any numbers in 1 with r
1
< r
2
, we have shown that 1 is
increasing on 1.
73. (a) Since 1 and o are positive, increasing, and CU on 1 with 1
00
and o
00
never equal to 0, we have 1 0, 1
0
D 0, 1
00
0,
o 0, o
0
D 0, o
00
0 on 1. Then (1o)
0
= 1
0
o +1o
0
i (1o)
00
= 1
00
o + 21
0
o
0
+1o
00
D 1
00
o +1o
00
0 on 1 i
1o is CU on 1.
(b) In part (a), if 1 and o are both decreasing instead of increasing, then 1
0
$ 0 and o
0
$ 0 on 1, so we still have 21
0
o
0
D 0
on 1. Thus, (1o)
00
= 1
00
o + 21
0
o
0
+1o
00
D 1
00
o +1o
00
0 on 1 i 1o is CU on 1 as in part (a).
(c) Suppose 1 is increasing and o is decreasing [with 1 and o positive and CU]. Then 1
0
D 0 and o
0
$ 0 on 1, so 21
0
o
0
$ 0
on 1 and the argument in parts (a) and (b) fails.
Example 1. 1 = (0. "), 1(r) = r
3
, o(r) = 1r. Then (1o)(r) = r
2
, so (1o)
0
(r) = 2r and
(1o)
00
(r) = 2 0 on 1. Thus, 1o is CU on 1.
Example 2. 1 = (0. "), 1(r) = 4r
I
r, o(r) = 1r. Then (1o)(r) = 4
I
r, so (1o)
0
(r) = 2
I
r and
(1o)
00
(r) = 31
I
r
3
< 0 on 1. Thus, 1o is CD on 1.
Example 3. 1 = (0. "), 1(r) = r
2
, o(r) = 1r. Thus, (1o)(r) = r, so 1o is linear on 1.
75. 1(r) = tanr 3r i 1
0
(r) = sec
2
r 31 0 for 0 < r <
r
2
since sec
2
r 1 for 0 < r <
r
2
. So 1 is increasing
on
0.
r
2
. Thus, 1(r) 1(0) = 0 for 0 < r <

r
2
i tanr 3r 0 i tanr r for 0 < r <
r
2
.
166

77. Let the cubic function be 1(r) = or
3
+/r
2
+cr +d i 1
0
(r) = 3or
2
+ 2/r +c i 1
00
(r) = 6or + 2/.
So 1 is CU when 6or + 2/ 0 C r 3/(3o), CD when r < 3/(3o), and so the only point of inection occurs when
r = 3/(3o). If the graph has three r-intercepts r
1
, r
2
and r
3
, then the expression for 1(r) must factor as
1(r) = o(r 3r
1
)(r 3r
2
)(r 3r
3
). Multiplying these factors together gives us
1(r) = o[r
3
3(r
1
+r
2
+r
3
)r
2
+ (r
1
r
2
+r
1
r
3
+r
2
r
3
)r 3r
1
r
2
r
3
]
Equating the coefcients of the r
2
-terms for the two forms of 1 gives us / = 3o(r
1
+r
2
+r
3
). Hence, the r-coordinate of
the point of inection is 3
/
3o
= 3
3o(r
1
+r
2
+r
3
)
3o
=
r
1
+r
2
+r
3
3
.
79. By hypothesis o = 1
0
is differentiable on an open interval containing c. Since (c. 1(c)) is a point of inection, the concavity
changes at r = c, so 1
00
(r) changes signs at r = c. Hence, by the First Derivative Test, 1
0
has a local extremum at r = c.
Thus, by Fermats Theorem 1
00
(c) = 0.
81. Using the fact that |r| =
I
r
2
, we have that o(r) = r
I
r
2
i o
0
(r) =
I
r
2
+
I
r
2
= 2
I
r
2
= 2 |r| i
o
00
(r) = 2r
r
2
31'2
=
2r
|r|
< 0 for r < 0 and o
00
(r) 0 for r 0, so (0. 0) is an inection point. But o
00
(0) does not
exist.
83. (a) 1(r) = r
4
sin
1
r
i 1
0
(r) = r
4
cos
1
r
3
1
r
2
+ sin
1
r
(4r
3
) = 4r
3
sin
1
r
3r
2
cos
1
r
.
o(r) = r
4
2 + sin
1
r
= 2r
4
+1(r) i o
0
(r) = 8r
3
+1
0
(r).
/(r) = r
4
32 + sin
1
r
= 32r
4
+1(r) i /
0
(r) = 38r
3
+1
0
(r).
It is given that 1(0) = 0, so 1
0
(0) = lim
i<0
1(r) 31(0)
r 30
= lim
i<0
r
4
sin
1
r
30
r
= lim
i<0
r
3
sin
1
r
. Since
3
r
3
$ r
3
sin
1
r
$
r
3
and lim
i<0
r
3
= 0, we see that 1
0
(0) = 0 by the Squeeze Theorem. Also,
o
0
(0) = 8(0)
3
+1
0
(0) = 0 and /
0
(0) = 38(0)
3
+1
0
(0) = 0, so 0 is a critical number of 1, o, and /.
For r
2n
=
1
2n
[n a nonzero integer], sin
1
r
2n
= sin2n = 0 and cos
1
r
2n
= cos 2n = 1, so 1
0
(r
2n
) = 3r
2
2n
< 0.
For r
2n+1
=
1
(2n + 1)
, sin
1
r
2n+1
= sin(2n + 1) = 0 and cos
1
r
2n+1
= cos(2n + 1) = 31, so
1
0
(r
2n+1
) = r
2
2n+1
0. Thus, 1
0
changes sign innitely often on both sides of 0.
Next, o
0
(r
2n
) = 8r
3
2n
+1
0
(r
2n
) = 8r
3
2n
3r
2
2n
= r
2
2n
(8r
2n
31) < 0 for r
2n
<
1
8
, but
o
0
(r
2n+1
) = 8r
3
2n+1
+r
2
2n+1
= r
2
2n+1
(8r
2n+1
+ 1) 0 for r
2n+1
3
1
8
, so o
0
changes sign innitely often on both
sides of 0.
Last, /
0
(r
2n
) = 38r
3
2n
+1
0
(r
2n
) = 38r
3
2n
3r
2
2n
= 3r
2
2n
(8r
2n
+ 1) < 0 for r
2n
3
1
8
and
/
0
(r
2n+1
) = 38r
3
2n+1
+r
2
2n+1
= r
2
2n+1
(38r
2n+1
+1) 0 for r
2n+1
<
1
8
, so /
0
changes sign innitely often on both
sides of 0.
SECTION 4.4 INDETERMINATE FORMS AND LHOSPITALS RULE

167
(b) 1(0) = 0 and since sin
1
r
and hence r
4
sin
1
r
is both positive and negative ininitely often on both sides of 0, and
arbitrarily close to 0, 1 has neither a local maximum nor a local minimum at 0.
Since 2 + sin
1
r
D 1, o(r) = r
4
2 + sin
1
r
0 for r 6= 0, so o(0) = 0 is a local minimum.

Since 32 + sin
1
r
$ 31, /(r) = r
4
32 + sin
1
r
< 0 for r 6= 0, so /(0) = 0 is a local maximum.

4.4 Indeterminate Forms and L'Hospital's Rule
Note: The use of lHospitals Rule is indicated by an H above the equal sign:
H
=
1. (a) lim
i<a
1(r)
o(r)
is an indeterminate form of type
0
0
.
(b) lim
i<a
1(r)
j(r)
= 0 because the numerator approaches 0 while the denominator becomes large.
(c) lim
i<a
/(r)
j(r)
= 0 because the numerator approaches a nite number while the denominator becomes large.
(d) If lim
i<a
j(r) = "and 1(r) <0 through positive values, then lim
i<a
j(r)
1(r)
= ". [For example, take o = 0, j(r) = 1r
2
,
and 1(r) = r
2
.] If 1(r) <0 through negative values, then lim
i<a
j(r)
1(r)
= 3". [For example, take o = 0, j(r) = 1r
2
,
and 1(r) = 3r
2
.] If 1(r) <0 through both positive and negative values, then the limit might not exist. [For example,
take o = 0, j(r) = 1r
2
, and 1(r) = r.]
(e) lim
i<a
j(r)
q(r)
is an indeterminate form of type
"
"
.
3. (a) When r is near o, 1(r) is near 0 and j(r) is large, so 1(r) 3j(r) is large negative. Thus, lim
i<a
[1(r) 3j(r)] = 3".
(b) lim
i<a
[ j(r) 3q(r)] is an indeterminate form of type "3".
(c) When r is near o, j(r) and q(r) are both large, so j(r) +q(r) is large. Thus, lim
i<a
[ j(r) +q(r)] = ".
5. This limit has the form
0
0
. We can simply factor and simplify to evaluate the limit.
lim
i<1
r
2
31
r
2
3r
= lim
i<1
(r + 1)(r 31)
r(r 31)
= lim
i<1
r + 1
r
=
1 + 1
1
= 2
0
0
. lim
i<1
r
9
31
r
5
31
H
= lim
i<1
9r
8
5r
4
=
9
5
lim
i<1
r
4
=
9
5
(1) =
9
5
0
0
. lim
i<(r'2)
+
cos r
1 3sinr
H
= lim
i<(r'2)
+
3sinr
3cos r
= lim
i<(r'2)
+
tanr = 3".
0
0
. lim
I<0
c
I
31
t
3
H
= lim
I<0
c
I
3t
2
= "since c
I
<1 and 3t
2
<0
+
as t <0.
168

0
0
. lim
i<0
tanjr
tanqr
H
= lim
i<0
j sec
2
jr
q sec
2
qr
=
j(1)
2
q(1)
2
=
j
q
"
"
. lim
i<"
lnr
I
r
H
= lim
i<"
1r
1
2
r
31'2
= lim
i<"
2
I
r
= 0
17. lim
i<0
+
[(lnr)r] = 3"since lnr <3"as r <0
+
and dividing by small values of r just increases the magnitude of the
quotient (lnr)r. LHospitals Rule does not apply.
"
"
. lim
i<"
c
i
r
3
H
= lim
i<"
c
i
3r
2
H
= lim
i<"
c
i
6r
H
= lim
i<"
c
i
6
= "
0
0
. lim
i<0
c
i
31 3r
r
2
H
= lim
i<0
c
i
31
2r
H
= lim
i<0
c
i
2
=
1
2
0
0
. lim
i<0
tanhr
tanr
H
= lim
i<0
sech
2
r
sec
2
r
=
sech
2
0
sec
2
0
=
1
1
= 1
0
0
. lim
I<0
5
I
33
I
t
H
= lim
I<0
5
I
ln5 33
I
ln3
1
= ln5 3ln3 = ln
5
3
0
0
. lim
i<0
sin
31
r
r
H
= lim
i<0
1
I
1 3r
2
1
= lim
i<0
1
I
1 3r
2
=
1
1
= 1
0
0
. lim
i<0
1 3cos r
r
2
H
= lim
i<0
sinr
2r
H
= lim
i<0
cos r
2
=
1
2
31. lim
i<0
r + sinr
r + cos r
=
0 + 0
0 + 1
=
0
1
= 0. LHospitals Rule does not apply.
0
0
. lim
i<1
1 3r + lnr
1 + cos r
H
= lim
i<1
31 + 1r
3 sinr
H
= lim
i<1
31r
2
3
2
cos r
=
31
3
2
(31)
= 3
1
2
0
0
. lim
i<1
r
a
3or +o 31
(r 31)
2
H
= lim
i<1
or
a31
3o
2(r 31)
H
= lim
i<1
o(o 31)r
a32
2
=
o(o 31)
2
0
0
. lim
i<0
cos r 31 +
1
2
r
2
r
4
H
= lim
i<0
3sinr +r
4r
3
H
= lim
i<0
3cos r + 1
12r
2
H
= lim
i<0
sinr
24r
H
= lim
i<0
cos r
24
=
1
24
39. This limit has the form " 0.
lim
i<"
rsin(r) = lim
i<"
sin(r)
1r
H
= lim
i<"
cos(r)(3r
2
)
31r
2
= lim
i<"
cos(r) = (1) =
41. This limit has the form " 0. Well change it to the form
0
0
.
lim
i<0
cot 2rsin6r = lim
i<0
sin6r
tan2r
H
= lim
i<0
6 cos 6r
2 sec
2
2r
=
6(1)
2(1)
2
= 3
43. This limit has the form" 0. lim
i<"
r
3
c
3i
2
= lim
i<"
r
3
c
i
2
H
= lim
i<"
3r
2
2rc
i
2
= lim
i<"
3r
2c
i
2
H
= lim
i<"
3
4rc
i
2
= 0

169
45. This limit has the form 0 (3").
lim
i<1
+
lnr tan(r2) = lim
i<1
+
lnr
cot(r2)
H
= lim
i<1
+
1r
(32) csc
2
(r2)
=
1
(32)(1)
2
= 3
2
47. This limit has the form"3".

lim
i<1
r
r 31
3
1
lnr
= lim
i<1
rlnr 3(r 31)
(r 31) lnr
H
= lim
i<1
r(1r) + lnr 31
(r 31)(1r) + lnr
= lim
i<1
lnr
1 3(1r) + lnr
H
= lim
i<1
1r
1r
2
+ 1r

r
2
r
2
= lim
i<1
r
1 +r
=
1
1 + 1
=
1
2
49. We will multiply and divide by the conjugate of the expression to change the form of the expression.
lim
i<"
I
r
2
+r 3r
= lim
i<"
I
r
2
+r 3r
1

I
r
2
+r +r
I
r
2
+r +r
= lim
i<"
r
2
+r
3r
2
I
r
2
+r +r
= lim
i<"
r
I
r
2
+r +r
= lim
i<"
1
1 + 1r + 1
=
1
I
1 + 1
=
1
2
As an alternate solution, write
I
r
2
+r 3r as
I
r
2
+r 3
I
r
2
, factor out
I
r
2
, rewrite as (
1 + 1r 31)(1r), and
apply lHospitals Rule.
51. The limit has the form "3"and we will change the form to a product by factoring out r.
lim
i<"
(r 3lnr) = lim
i<"
r
1 3
lnr
r
= "since lim
i<"
lnr
r
H
= lim
i<"
1r
1
= 0.
53. j = r
i
2
i lnj = r
2
lnr, so lim
i<0
+
lnj = lim
i<0
+
r
2
lnr = lim
i<0
+
lnr
1r
2
H
= lim
i<0
+
1r
32r
3
= lim
i<0
+
3
1
2
r
2
= 0 i
lim
i<0
+
r
i
2
= lim
i<0
+
c
ln u
= c
0
= 1.
55. j = (1 32r)
1'i
i lnj =
1
r
ln(1 32r), so lim
i<0
lnj = lim
i<0
ln(1 32r)
r
H
= lim
i<0
32(1 32r)
1
= 32 i
lim
i<0
(1 32r)
1'i
= lim
i<0
c
ln u
= c
32
.
57. j =

1 +
3
r
+
5
r
2
i
i lnj = rln
1 +
3
r
+
5
r
2
i
lim
i<"
lnj = lim
i<"
ln
1 +
3
r
+
5
r
2
1r
H
= lim
i<"
3
3
r
2
3
10
r
3
1 +
3
r
+
5
r
2
31r
2
= lim
i<"
3 +
10
r
1 +
3
r
+
5
r
2
= 3,
so lim
i<"
1 +
3
r
+
5
r
2
i
= lim
i<"
c
ln u
= c
3
.
59. j = r
1'i
i lnj = (1r) lnr i lim
i<"
lnj = lim
i<"
lnr
r
H
= lim
i<"
1r
1
= 0 i
lim
i<"
r
1'i
= lim
i<"
c
ln u
= c
0
= 1
170

61. j = (4r + 1)
cot i
i lnj = cot r ln(4r + 1), so lim
i<0
+
lnj = lim
i<0
+
ln(4r + 1)
tanr
H
= lim
i<0
+
4
4r + 1
sec
2
r
= 4 i
lim
i<0
+
(4r + 1)
cot i
= lim
i<0
+
c
ln u
= c
4
.
63. j = (cos r)
1'i
2
i lnj =
1
r
2
lncos r i lim
i<0
+
lnj = lim
i<0
+
lncos r
r
2
H
= lim
i<0
+
3tanr
2r
H
= lim
i<0
+
3sec
2
r
2
= 3
1
2
i lim
i<0
+
(cos r)
1'i
2
= lim
i<0
+
c
ln u
= c
31'2
= 1
I
c
65. From the graph, if r = 500, j E 7.36. The limit has the form 1
"
.
Now j =

1 +
2
r
i
i lnj = rln
1 +
2
r
i
lim
i<"
lnj = lim
i<"
ln(1 + 2r)
1r
H
= lim
i<"
1
1 + 2r
3
2
r
2
31r
2
= 2 lim
i<"
1
1 + 2r
= 2(1) = 2 i
lim
i<"
1 +
2
r
i
= lim
i<"
c
ln u
= c
2
[E 7.39]
67. From the graph, it appears that lim
i<0
1(r)
o(r)
= lim
i<0
1
0
(r)
o
0
(r)
= 0.25.
We calculate lim
i<0
1(r)
o(r)
= lim
i<0
c
i
31
r
3
+ 4r
H
= lim
i<0
c
i
3r
2
+ 4
=
1
4
.
69. lim
i<"
c
i
r
n
H
= lim
i<"
c
i
nr
n31
H
= lim
i<"
c
i
n(n 31)r
n32
H
=
H
= lim
i<"
c
i
n!
= "
71. lim
i<"
r
I
r
2
+ 1
H
= lim
i<"
1
1
2
(r
2
+ 1)
31'2
(2r)
= lim
i<"
I
r
2
+ 1
r
. Repeated applications of lHospitals Rule result in the
original limit or the limit of the reciprocal of the function. Another method is to try dividing the numerator and denominator
by r: lim
i<"
r
I
r
2
+ 1
= lim
i<"
rr
r
2
r
2
+ 1r
2
= lim
i<"
1
1 + 1r
2
=
1
1
= 1
73. First we will nd lim
n<"
1 +
v
n
nI
, which is of the form 1
"
. j =

1 +
v
n
nI
i lnj = nt ln
1 +
v
n
, so
lim
n<"
lnj = lim
n<"
nt ln
1 +
v
n
= t lim
n<"
ln(1 +vn)
1n
H
= t lim
n<"
3vn
2
(1 +vn)(31n
2
)
= t lim
n<"
v
1 +jn
= tv i
lim
n<"
j = c
rI
. Thus, as n <", =
0
1 +
v
n
nI
<
0
c
rI
.

171
75. lim
T<0
+
1(1) = lim
T<0
+
c
T
+c
3T
c
T
3c
3T
3
1
1
= lim
T<0
+
1
c
T
+c
3T
31
c
T
3c
3T
(c
T
3c
3T
) 1
= lim
T<0
+
1c
T
+1c
3T
3c
T
+c
3T
1c
T
31c
3T
form is
0
0
H
= lim
T<0
+
1c
T
+c
T
1 +1
3c
3T
+c
3T
1 3c
T
+
3c
3T
1c
T
+c
T
1 3[1(3c
3T
) +c
3T
1]
= lim
T<0
+
1c
T
31c
3T
1c
T
+c
T
+1c
3T
3c
3T
= lim
T<0
+
c
T
3c
3T
c
T
+
c
T
1
+c
3T
3
c
3T
1
[divide by 1]
=
0
2 +1
, where 1 = lim
T<0
+
c
T
3c
3T
1
form is
0
0
H
= lim
T<0
+
c
T
+c
3T
1
=
1 + 1
1
= 2
Thus, lim
T<0
+
1(1) =
0
2 + 2
= 0.
77. We see that both numerator and denominator approach 0, so we can use lHospitals Rule:
lim
i<a
I
2o
3
r 3r
4
3o
3
I
oor
o 3
4
I
or
3
H
= lim
i<a
1
2
(2o
3
r 3r
4
)
31'2
(2o
3
34r
3
) 3o
1
3
(oor)
32'3
o
2
3
1
4
(or
3
)
33'4
(3or
2
)
=
1
2
(2o
3
o 3o
4
)
31'2
(2o
3
34o
3
) 3
1
3
o
3
(o
2
o)
32'3
3
1
4
(oo
3
)
33'4
(3oo
2
)
=
(o
4
)
31'2
(3o
3
) 3
1
3
o
3
(o
3
)
32'3
3
3
4
o
3(a
4
)33'4
=
3o 3
1
3
o
3
3
4
=
4
3
4
3
o
=
16
9
o
79. Since 1(2) = 0, the given limit has the form
0
0
.
lim
i<0
1(2 + 3r) +1(2 + 5r)
r
H
= lim
i<0
1
0
(2 + 3r) 3 +1
0
(2 + 5r) 5
1
= 1
0
(2) 3 +1
0
(2) 5 = 81
0
(2) = 8 7 = 56
81. Since lim
I<0
[1(r +/) 31(r 3/)] = 1(r) 31(r) = 0 (1 is differentiable and hence continuous) and lim
I<0
2/ = 0, we use
lHospitals Rule:
lim
I<0
1(r +/) 31(r 3/)
2/
H
= lim
I<0
1
0
(r +/)(1) 31
0
(r 3/)(31)
2
=
1
0
(r) +1
0
(r)
2
=
21
0
(r)
2
= 1
0
(r)
1(r +/) 31(r 3/)
2/
is the slope of the secant line between
(r 3/. 1(r 3/)) and (r +/. 1(r +/)). As / <0, this line gets closer
to the tangent line and its slope approaches 1
0
(r).
83. (a) We show that lim
i<0
1(r)
r
n
= 0 for every integer n D 0. Let j =
1
r
2
. Then
lim
i<0
1(r)
r
2n
= lim
i<0
c
31'i
2
(r
2
)
n
= lim
u<"
j
n
c
u
H
= lim
u<"
nj
n31
c
u
H
=
H
= lim
u<"
n!
c
u
= 0 i
lim
i<0
1(r)
r
n
= lim
i<0
r
n
1(r)
r
2n
= lim
i<0
r
n
lim
i<0
1(r)
r
2n
= 0. Thus, 1
0
(0) = lim
i<0
1(r) 31(0)
r 30
= lim
i<0
1(r)
r
= 0.
172

(b) Using the Chain Rule and the Quotient Rule we see that 1
(n)
(r) exists for r 6= 0. In fact, we prove by induction that for
each n D 0, there is a polynomial j
n
and a non-negative integer I
n
with 1
(n)
(r) = j
n
(r)1(r)r
I
q
for r 6= 0. This is
true for n = 0; suppose it is true for the nth derivative. Then 1
0
(r) = 1(r)(2r
3
), so
1
(n+1)
(r) =

r
I
q
[j
0
n
(r) 1(r) +j
n
(r)1
0
(r)] 3I
n
r
I
q
31
j
n
(r) 1(r)
r
32I
q
=

r
I
q
j
0
n
(r) +j
n
(r)
2r
3
3I
n
r
I
q
31
j
n
(r)
1(r)r
32I
q
=

r
I
q
+3
j
0
n
(r) + 2j
n
(r) 3I
n
r
I
q
+2
j
n
(r)
1(r)r
3(2I
q
+3)
which has the desired form.
Now we show by induction that 1
(n)
(0) = 0 for all n. By part (a), 1
0
(0) = 0. Suppose that 1
(n)
(0) = 0. Then
1
(n+1)
(0) = lim
i<0
1
(n)
(r) 31
(n)
(0)
r 30
= lim
i<0
1
(n)
(r)
r
= lim
i<0
j
n
(r) 1(r)r
I
q
r
= lim
i<0
j
n
(r) 1(r)
r
I
q
+1
= lim
i<0
j
n
(r) lim
i<0
1(r)
r
I
q
+1
= j
n
(0) 0 = 0
4.5 Summary of Curve Sketching
1. j = 1(r) = r
3
+r = r(r
2
+ 1) A. 1 is a polynomial, so 1 = R.
B. r-intercept = 0, j-intercept = 1(0) = 0 C. 1(3r) = 31(r), so 1 is
odd; the curve is symmetric about the origin. D. 1 is a polynomial, so there is
no asymptote. E. 1
0
(r) = 3r
2
+ 1 0, so 1 is increasing on (3". ").
F. There is no critical number and hence, no local maximum or minimum value.
G. 1
00
(r) = 6r 0 on (0. ") and 1
00
(r) < 0 on (3". 0), so 1 is CU on
(0. ") and CD on (3". 0). Since the concavity changes at r = 0, there is an
inection point at (0. 0).
H.
3. j = 1(r) = 2 315r + 9r
2
3r
3
= 3(r 32)
r
2
37r + 1
A. 1 = R B. j-intercept: 1(0) = 2; r-intercepts:

1(r) = 0 i r = 2 or (by the quadratic formula) r =
7
I
45
2
E 0.15, 6.85 C. No symmetry D. No asymptote
E. 1
0
(r) = 315 + 18r 33r
2
= 33(r
2
36r + 5)
= 33(r 31)(r 35) 0 C 1 < r < 5
so 1 is increasing on (1. 5) and decreasing on (3". 1) and (5. ").
F. Local maximum value 1(5) = 27, local minimum value 1(1) = 35
G. 1
00
(r) = 18 36r = 36(r 33) 0 C r < 3, so 1 is CU on (3". 3)
and CD on (3. "). IP at (3. 11)
H.
SECTION 4.5 SUMMARY OF CURVE SKETCHING

173
5. j = 1(r) = r
4
+ 4r
3
= r
3
(r + 4) A. 1 = R B. j-intercept: 1(0) = 0;
r-intercepts: 1(r) = 0 C r = 34. 0 C. No symmetry
D. No asymptote E. 1
0
(r) = 4r
3
+ 12r
2
= 4r
2
(r + 3) 0 C
r 33, so 1 is increasing on (33. ") and decreasing on (3". 33).
F. Local minimum value 1(33) = 327, no local maximum
G. 1
00
(r) = 12r
2
+ 24r = 12r(r + 2) < 0 C 32 < r < 0, so 1 is CD
on (32. 0) and CU on (3". 32) and (0. "). IP at (0. 0) and (32. 316)
H.
7. j = 1(r) = 2r
5
35r
2
+ 1 A. 1 = R B. j-intercept: 1(0) = 1 C. No symmetry D. No asymptote
E. 1
0
(r) = 10r
4
310r = 10r(r
3
31) = 10r(r 31)(r
2
+r + 1), so 1
0
(r) < 0 C 0 < r < 1 and 1
0
(r) 0 C
r < 0 or r 1. Thus, 1 is increasing on (3". 0) and (1. ") and decreasing on (0. 1). F. Local maximum value 1(0) = 1,
local minimum value 1(1) = 32 G. 1
00
(r) = 40r
3
310 = 10(4r
3
31)
so 1
00
(r) = 0 C r = 1
3
I
4. 1
00
(r) 0 C r 1
3
I
4 and
1
00
(r) < 0 C r < 1
3
I
4, so 1 is CD on
3". 1
3
I
4
and CU
on
1
3
I
4. "
. IP at
1
3
I
4
. 1 3
9
2
3
I
4
E (0.630. 30.786)
H.
9. j = 1(r) = r(r 31) A. 1 = {r | r 6= 1} = (3". 1) (1. ") B. r-intercept = 0, j-intercept = 1(0) = 0
C. No symmetry D. lim
i<"
r
r 31
= 1, so j = 1 is a HA. lim
i<1
r
r 31
= 3", lim
i<1
+
r
r 31
= ", so r = 1 is a VA.
E. 1
0
(r) =
(r 31) 3r
(r 31)
2
=
31
(r 31)
2
< 0 for r 6= 1, so 1 is
decreasing on (3". 1) and (1. ") . F. No extreme values
G. 1
00
(r) =
2
(r 31)
3
0 C r 1, so 1 is CU on (1. ") and
CD on (3". 1). No IP
H.
11. j = 1(r) = 1(r
2
39) A. 1 = {r | r 6= 3} = (3". 33) (33. 3) (3. ") B. j-intercept = 1(0) = 3
1
9
, no
r-intercept C. 1(3r) = 1(r) i 1 is even; the curve is symmetric about the j-axis. D. lim
i<"
1
r
2
39
= 0, so j = 0
is a HA. lim
i<3
1
r
2
39
= 3", lim
i<3
+
1
r
2
39
= ", lim
i<33
1
r
2
39
= ", lim
i<33
+
1
r
2
39
= 3", so r = 3 and r = 33
are VA. E. 1
0
(r) = 3
2r
(r
2
39)
2
0 C r < 0 (r 6= 33) so 1 is increasing on (3". 33) and (33. 0) and
174

decreasing on (0. 3) and (3. "). F. Local maximum value 1(0) = 3
1
9
.
G. j
00
=
32(r
2
39)
2
+ (2r)2(r
2
39)(2r)
(r
2
39)
4
=
6(r
2
+ 3)
(r
2
39)
3
0 C
r
2
9 C r 3 or r < 33, so 1 is CU on (3". 33) and (3. ") and
CD on (33. 3). No IP
H.
13. j = 1(r) = r(r
2
+ 9) A. 1 = R B. j-intercept: 1(0) = 0; r-intercept: 1(r) = 0 C r = 0
C. 1(3r) = 31(r), so 1 is odd and the curve is symmetric about the origin. D. lim
i<"
[r(r
2
+ 9)] = 0, so j = 0 is a
HA; no VA E. 1
0
(r) =
(r
2
+ 9)(1) 3r(2r)
(r
2
+ 9)
2
=
9 3r
2
(r
2
+ 9)
2
=
(3 +r)(3 3r)
(r
2
+ 9)
2
0 C 33 < r < 3, so 1 is increasing
on (33. 3) and decreasing on (3". 33) and (3. "). F. Local minimum value 1(33) = 3
1
6
, local maximum
value 1(3) =
1
6
1
00
(r) =
(r
2
+ 9)
2
(32r) 3(9 3r
2
) 2(r
2
+ 9)(2r)
[(r
2
+ 9)
2
]
2
=
(2r)(r
2
+ 9)[3(r
2
+ 9) 32(9 3r
2
)]
(r
2
+ 9)
4
=
2r(r
2
327)
(r
2
+ 9)
3
= 0 C r = 0.
I
27 = 3
I
3
G. 1
00
(r) =
(r
2
+ 9)
2
(32r) 3(9 3r
2
) 2(r
2
+ 9)(2r)
[(r
2
+ 9)
2
]
2
=
(2r)(r
2
+ 9)
3(r
2
+ 9) 32(9 3r
2
)
(r
2
+ 9)
4
=
2r(r
2
327)
(r
2
+ 9)
3
= 0 C r = 0,
I
27 = 3
I
3 H.
1
00
(r) 0 C 33
I
3 < r < 0 or r 3
I
3, so 1 is CU on
33
I
3. 0
and
3
I
3. "
, and CD on
3". 33
I
3
and
0. 3
I
3
. There are three

inection points: (0. 0) and
3
I
3.
1
12
I
3
.
15. j = 1(r) =
r 31
r
2
A. 1 = {r | r 6= 0} = (3". 0) (0. ") B. No j-intercept; r-intercept: 1(r) = 0 C r = 1
i<"
r 31
r
2
i<0
r 31
r
2
= 3", so r = 0 is a VA.
E. 1
0
(r) =
r
2
1 3(r 31) 2r
(r
2
)
2
=
3r
2
+ 2r
r
4
=
3(r 32)
r
3
, so 1
0
(r) 0 C 0 < r < 2 and 1
0
(r) < 0 C
r < 0 or r 2. Thus, 1 is increasing on (0. 2) and decreasing on (3". 0)
and (2. "). F. No local minimum, local maximum value 1(2) =
1
4
.
G. 1
00
(r) =
r
3
(31) 3[3(r 32)] 3r
2
(r
3
)
2
=
2r
3
36r
2
r
6
=
2(r 33)
r
4
.
1
00
(r) is negative on (3". 0) and (0. 3) and positive on (3. "), so 1 is CD
on (3". 0) and (0. 3) and CU on (3. "). IP at
3.
2
9
H.

175
17. j = 1(r) =
r
2
r
2
+ 3
=
(r
2
+ 3) 33
r
2
+ 3
= 1 3
3
r
2
+ 3
A. 1 = R B. j-intercept: 1(0) = 0;
r-intercepts: 1(r) = 0 C r = 0 C. 1(3r) = 1(r), so 1 is even; the graph is symmetric about the j-axis.
D. lim
i<"
r
2
r
2
+ 3
= 1, so j = 1 is a HA. No VA. E. Using the Reciprocal Rule, 1
0
(r) = 33
32r
(r
2
+ 3)
2
=
6r
(r
2
+ 3)
2
.
1
0
(r) 0 C r 0 and 1
0
(r) < 0 C r < 0, so 1 is decreasing on (3". 0) and increasing on (0. ").
F. Local minimum value 1(0) = 0, no local maximum.
G. 1
00
(r) =
(r
2
+ 3)
2
6 36r 2(r
2
+ 3) 2r
[(r
2
+ 3)
2
]
2
=
6(r
2
+ 3)[(r
2
+ 3) 34r
2
]
(r
2
+ 3)
4
=
6(3 33r
2
)
(r
2
+ 3)
3
=
318(r + 1)(r 31)
(r
2
+ 3)
3
1
00
(r) is negative on (3". 31) and (1. ") and positive on (31. 1),
so 1 is CD on (3". 31) and (1. ") and CU on (31. 1). IP at
1.
1
4
H.
19. j = 1(r) = r
I
5 3r A. The domain is {r | 5 3r D 0} = (3". 5] B. j-intercept: 1(0) = 0;
r-intercepts: 1(r) = 0 C r = 0, 5 C. No symmetry D. No asymptote
E. 1
0
(r) = r
1
2
(5 3r)
31'2
(31) + (5 3r)
1'2
1 =
1
2
(5 3r)
31'2
[3r + 2(5 3r)] =
10 33r
2
I
5 3r
0 C
r <
10
3
, so 1 is increasing on
3".
10
3
and decreasing on
10
3
. 5
.
F. Local maximum value 1
10
3
=
10
9
I
15 E 4.3; no local minimum
G. 1
00
(r) =
2(5 3r)
1'2
(33) 3(10 33r) 2
1
2
(5 3r)
31'2
(31)
2
I
5 3r
2
=
(5 3r)
31'2
[36(5 3r) + (10 33r)]
4(5 3r)
=
3r 320
4(5 3r)
3'2
1
00
(r) < 0 for r < 5, so 1 is CD on (3". 5). No IP
H.
21. j = 1(r) =
I
r
2
+r 32 =

(r + 2)(r 31) A. 1 = {r | (r + 2)(r 31) D 0} = (3". 32] [1. ")
B. j-intercept: none; r-intercepts: 32 and 1 C. No symmetry D. No asymptote
E. 1
0
(r) =
1
2
(r
2
+r 32)
31'2
(2r + 1) =
2r + 1
2
I
r
2
+r 32
, 1
0
(r) = 0 if r = 3
1
2
, but 3
1
2
is not in the domain.
1
0
(r) 0 i r 3
1
2
and 1
0
(r) < 0 i r < 3
1
2
, so (considering the domain) 1 is increasing on (1. ") and
1 is decreasing on (3". 32). F. No local extrema
G. 1
00
(r) =
2(r
2
+r 32)
1'2
(2) 3(2r + 1) 2
1
2
(r
2
+r 32)
31'2
(2r + 1)
2
I
r
2
+r 32
2
=
(r
2
+r 32)
31'2
4(r
2
+r 32) 3(4r
2
+ 4r + 1)
4(r
2
+r 32)
=
39
4(r
2
+r 32)
3'2
< 0
so 1 is CD on (3". 32) and (1. "). No IP
H.
176

23. j = 1(r) = r
I
r
2
+ 1 A. 1 = R B. j-intercept: 1(0) = 0; r-intercepts: 1(r) = 0 i r = 0
C. 1(3r) = 31(r), so 1 is odd; the graph is symmetric about the origin.
D. lim
i<"
1(r) = lim
i<"
r
I
r
2
+ 1
= lim
i<"
rr
I
r
2
+ 1r
= lim
i<"
rr
I
r
2
+ 1
I
r
2
= lim
i<"
1
1 + 1r
2
=
1
I
1 + 0
= 1
and
lim
i<3"
1(r) = lim
i<3"
r
I
r
2
+ 1
= lim
i<3"
rr
I
r
2
+ 1r
= lim
i<3"
rr
I
r
2
+ 1
3
I
r
2
= lim
i<3"
1
3
1 + 1r
2
=
1
3
I
1 + 0
= 31 so j = 1 are HA.
No VA.
E. 1
0
(r) =
I
r
2
+ 1 3r
2r
2
I
r
2
+ 1
[(r
2
+ 1)
1'2
]
2
=
r
2
+ 1 3r
2
(r
2
+ 1)
3'2
=
1
(r
2
+ 1)
3'2
0 for all r, so 1 is increasing on R.
F. No extreme values
G. 1
00
(r) = 3
3
2
(r
2
+ 1)
35'2
2r =
33r
(r
2
+ 1)
5'2
, so 1
00
(r) 0 for r < 0
and 1
00
(r) < 0 for r 0. Thus, 1 is CU on (3". 0) and CD on (0. ").
IP at (0. 0)
H.
25. j = 1(r) =
I
1 3r
2
r A. 1 = {r | |r| $ 1, r 6= 0} = [31. 0) (0. 1] B. r-intercepts 1, no j-intercept
C. 1(3r) = 31(r), so the curve is symmetric about (0. 0) . D. lim
i<0
+
I
1 3r
2
r
= ", lim
i<0
I
1 3r
2
r
= 3",
so r = 0 is a VA. E. 1
0
(r) =
3r
2
I
1 3r
2
3
I
1 3r
2
r
2
= 3
1
r
2
I
1 3r
2
< 0, so 1 is decreasing
on (31. 0) and (0. 1). F. No extreme values
G. 1
00
(r) =
2 33r
2
r
3
(1 3r
2
)
3'2
0 C 31 < r < 3
2
3
or 0 < r <

2
3
, so
1 is CU on
31. 3
2
3
and
0.
2
3
and CD on
2
3
. 0
and
2
3
. 1
.
IP at
2
3
.
1
I
2
H.
27. j = 1(r) = r 33r
1'3
A. 1 = R B. j-intercept: 1(0) = 0; r-intercepts: 1(r) = 0 i r = 3r
1'3
i
r
3
= 27r i r
3
327r = 0 i r(r
2
327) = 0 i r = 0, 3
I
3 C. 1(3r) = 31(r), so 1 is odd;
the graph is symmetric about the origin. D. No asymptote E. 1
0
(r) = 1 3r
32'3
= 1 3
1
r
2'3
=
r
2'3
31
r
2'3
.
1
0
(r) 0 when |r| 1 and 1
0
(r) < 0 when 0 < |r| < 1, so 1 is increasing on (3". 31) and (1. "), and
decreasing on (31. 0) and (0. 1) [hence decreasing on (31. 1) since 1 is
continuous on (31. 1)]. F. Local maximum value 1(31) = 2, local minimum
value 1(1) = 32 G. 1
00
(r) =
2
3
r
35'3
< 0 when r < 0 and 1
00
(r) 0
when r 0, so 1 is CD on (3". 0) and CU on (0. "). IP at (0. 0)
H.

177
29. j = 1(r) =
3
I
r
2
31 A. 1 = R B. j-intercept: 1(0) = 31; r-intercepts: 1(r) = 0 C r
2
31 = 0 C
r = 1 C. 1(3r) = 1(r), so the curve is symmetric about the j-axis. D. No asymptote
E. 1
0
(r) =
1
3
(r
2
31)
32'3
(2r) =
2r
3
3
(r
2
31)
2
. 1
0
(r) 0 C r 0 and 1
0
(r) < 0 C r < 0, so 1 is
increasing on (0. ") and decreasing on (3". 0). F. Local minimum value 1(0) = 31
G. 1
00
(r) =
2
3

(r
2
31)
2'3
(1) 3r
2
3
(r
2
31)
31'3
(2r)
[(r
2
31)
2'3
]
2
=
2
9

(r
2
31)
31'3
[3(r
2
31) 34r
2
]
(r
2
31)
4'3
= 3
2(r
2
+ 3)
9(r
2
31)
5'3
1
00
(r) 0 C 31 < r < 1 and 1
00
(r) < 0 C r < 31 or r 1, so
1 is CU on (31. 1) and 1 is CD on (3". 31) and (1. "). IP at (1. 0)
H.
31. j = 1(r) = 3 sinr 3sin
3
r A. 1 = R B. j-intercept: 1(0) = 0; r-intercepts: 1(r) = 0 i
sinr(3 3sin
2
r) = 0 i sinr = 0 [since sin
2
r $ 1 < 3] i r = n, n an integer.
C. 1(3r) = 31(r), so 1 is odd; the graph (shown for 32 $ r $ 2) is symmetric about the origin and periodic
with period 2. D. No asymptote E. 1
0
(r) = 3 cos r 33 sin
2
r cos r = 3 cos r(1 3sin
2
r) = 3 cos
3
r.
1
0
(r) 0 C cos r 0 C r M
2n 3
r
2
. 2n +
r
2
for each integer n, and 1

0
(r) < 0 C cos r < 0 C
r M
2n +
r
2
. 2n +
3r
2
for each integer n. Thus, 1 is increasing on
2n 3
r
2
. 2n +
r
2
for each integer n,

and 1 is decreasing on
2n +
r
2
. 2n +
3r
2
for each integer n.

F. 1 has local maximum values 1(2n +
r
2
) = 2 and local minimum values 1(2n +
3r
2
) = 32.
G. 1
00
(r) = 39 sinr cos
2
r = 39 sinr(1 3sin
2
r) = 39 sinr(1 3sinr)(1 + sinr).
1
00
(r) < 0 C sinr 0 and sinr 6= 1 C r M
2n. 2n +
r
2
2n +
r
2
. 2n +
for some integer n.

1
00
(r) 0 C sinr < 0 and sinr 6= 1 C r M
(2n 31). (2n 31) +

r
2
(2n 31) +
r
2
. 2n
for some integer n. Thus, 1 is CD on the intervals
2n.
2n +
1
2
and
2n +
1
2
. (2n + 1)
[hence CD on the intervals (2n. (2n + 1) )]

for each integer n, and 1 is CU on the intervals
(2n 31).
2n 3
1
2
and
2n 3
1
2
. 2n
[hence CU on the intervals ((2n 31). 2n)]

for each integer n. 1 has inection points at (n. 0) for each integer n.
H.
33. j = 1(r) = rtanr, 3
r
2
< r <
r
2
A. 1 =

3
r
2
.
r
2
B. Intercepts are 0 C. 1(3r) = 1(r), so the curve is

symmetric about the j-axis. D. lim
i<(r'2)
rtanr = "and lim

i<3(r'2)
+
rtanr = ", so r =
r
2
and r = 3
r
2
are VA.
E. 1
0
(r) = tanr +rsec
2
r 0 C 0 < r <
r
2
, so 1 increases on
0.
r
2
and decreases on
3
r
2
. 0
. F. Absolute and local minimum value 1(0) = 0.

G. j
00
= 2 sec
2
r + 2rtanr sec
2
r 0 for 3
r
2
< r <
r
2
, so 1 is
CU on
3
r
2
.
r
2
. No IP
H.
178

35. j = 1(r) =
1
2
r 3sinr, 0 < r < 3 A. 1 = (0. 3) B. No j-intercept. The r-intercept, approximately 1.9, can be
found using Newtons Method. C. No symmetry D. No asymptote E. 1
0
(r) =
1
2
3cos r 0 C cos r <
1
2
C
r
3
< r <
5r
3
or
7r
3
< r < 3, so 1 is increasing on
r
3
.
5r
3
and
7r
3
. 3
and decreasing on
0.
r
3
and
5r
3
.
7r
3
.
F. Local minimum value 1
r
3
=
r
6
3
I
3
2
, local maximum value
1
5r
3
=
5r
6
+
I
3
2
, local minimum value 1
7r
3
=
7r
6
3
I
3
2
G. 1
00
(r) = sinr 0 C 0 < r < or 2 < r < 3, so 1 is CU on
(0. ) and (2. 3) and CD on (. 2). IPs at
.
r
2
and (2. )
H.
37. j = 1(r) =
sinr
1 + cos r
when
cos { 6= 1
=
sinr
1 + cos r

1 3cos r
1 3cos r
=
sinr(1 3cos r)
sin
2
r
=
1 3cos r
sinr
= csc r 3cot r
A. The domain of 1 is the set of all real numbers except odd integer multiples of . B. j-intercept: 1(0) = 0; r-intercepts:
r = n, n an even integer. C. 1(3r) = 31(r), so 1 is an odd function; the graph is symmetric about the origin and has
period 2. D. When n is an odd integer, lim
i<(nr)
1(r) = "and lim

i<(nr)
+
1(r) = 3", so r = n is a VA for each odd
integer n. No HA. E. 1
0
(r) =
(1 + cos r) cos r 3sinr(3sinr)
(1 + cos r)
2
=
1 + cos r
(1 + cos r)
2
=
1
1 + cos r
. 1
0
(r) 0 for all r
except odd multiples of , so 1 is increasing on ((2I 31). (2I + 1)) for each integer I. F. No extreme values
G. 1
00
(r) =
sinr
(1 + cos r)
2
0 i sinr 0 i
r M (2I. (2I + 1)) and 1
00
(r) < 0 on ((2I 31). 2I) for each
integer I. 1 is CU on (2I. (2I + 1)) and CD on ((2I 31). 2I)
for each integer I. 1 has IPs at (2I. 0) for each integer I.
H.
39. j = 1(r) = c
sin i
A. 1 = R B. j-intercept: 1(0) = c
0
= 1; r-intercepts: none, since c
sin i
0 C. 1 is periodic
with period 2, so we determine EG for 0 $ r $ 2. D. No asymptote E. 1
0
(r) = c
sin i
cos r. 1
0
(r) 0 C
cos r 0 i r is in
0.
r
2
or
3r
2
. 2
[ 1 is increasing] and 1
0
(r) < 0 i r is in
r
2
.
3r
2
[ 1 is decreasing].
F. Local maximum value 1
r
2
= c and local minimum value 1
3r
2
= c
31
G. 1
00
(r) = c
sin i
(3sinr) + cos r(c
sin i
cos r) = c
sin i
(cos
2
r 3sinr). 1
00
(r) = 0 C cos
2
r 3sinr = 0 C
1 3sin
2
r 3sinr = 0 C sin
2
r + sinr 31 = 0 i sinr =
31
I
5
2
i c = sin
31
31 +
I
5
2
E 0.67 and
o = 3c E 2.48. 1
00
(r) < 0 on (c. o) [ 1 is CD] and 1
00
(r) 0 on (0. c) and (o. 2) [ 1 is CU].
The inection points occur when r = c, o.
H.

179
41. j = 1(1 +c
3i
) A. 1 = R B. No r-intercept; j-intercept = 1(0) =
1
2
. C. No symmetry
D. lim
i<"
1(1 +c
3i
) =
1
1 +0
= 1 and lim
i<3"
1(1 +c
3i
) = 0 since lim
i<3"
c
3i
= "], so 1 has horizontal asymptotes
j = 0 and j = 1. E. 1
0
(r) = 3(1 +c
3i
)
32
(3c
3i
) = c
3i
(1 +c
3i
)
2
. This is positive for all r, so 1 is increasing on R.
F. No extreme values G. 1
00
(r) =
(1 +c
3i
)
2
(3c
3i
) 3c
3i
(2)(1 +c
3i
)(3c
3i
)
(1 +c
3i
)
4
=
c
3i
(c
3i
31)
(1 +c
3i
)
3
The second factor in the numerator is negative for r 0 and positive for r < 0,
and the other factors are always positive, so 1 is CU on (3", 0) and CD
on (0. "). IP at
0,
1
2
H.
43. j = 1(r) = r 3lnr A. 1 = (0. ") B. j-intercept: none (0 is not in the domain); r-intercept: 1(r) = 0 C
r = lnr, which has no solution, so there is no r-intercept. C. No symmetry D. lim
i<0
+
(r 3lnr) = ". so r = 0
is a VA. E. 1
0
(r) = 1 31r 0 i 1 1r i r 1 and
1
0
(r) < 0 i 0 < r < 1, so 1 is increasing on (1. ") and 1 is decreasing
on (0. 1). F. Local minimum value 1(1) = 1; no local maximum value
G. 1
00
(r) =
1
r
2
0 for all r, so 1 is CU on (0. "). No IP
H.
45. j = 1(r) = (1 +c
i
)
32
=
1
(1 +c
i
)
2
A. 1 = R B. j-intercept: 1(0) =
1
4
. r-intercepts: none [since 1(r) 0]
i<"
1(r) = 0 and lim
i<3"
1(r) = 1, so j = 0 and j = 1 are HA; no VA
E. 1
0
(r) = 32(1 +c
i
)
33
c
i
=
32c
i
(1 +c
i
)
3
< 0, so 1 is decreasing on R F. No local extrema
G. 1
00
(r) = (1 +c
i
)
33
(32c
i
) + (32c
i
)(33)(1 +c
i
)
34
c
i
= 32c
i
(1 +c
i
)
34
[(1 +c
i
) 33c
i
] =
32c
i
(1 32c
i
)
(1 +c
i
)
4
.
1
00
(r) 0 C 1 32c
i
< 0 C c
i
1
2
C r ln
1
2
and
1
00
(r) < 0 C r < ln
1
2
, so 1 is CU on
ln
1
2
. "
and CD on
3". ln
1
2
.
IP at
ln
1
2
.
4
9
H.
47. j = 1(r) = ln(sinr)
A. 1 = {r in R | sinr 0} =
"
n=3"
(2n. (2n + 1) ) = (34. 33) (32. 3) (0. ) (2. 3)
B. No j-intercept; r-intercepts: 1(r) = 0 C ln(sinr) = 0 C sinr = c
0
= 1 C r = 2n +
r
2
for each
integer n. C. 1 is periodic with period 2. D. lim
i<(2nr)
+
1(r) = 3"and lim
i<[(2n+1)r]
1(r) = 3", so the lines

r = n are VAs for all integers n. E. 1
0
(r) =
cos r
sinr
= cot r, so 1
0
(r) 0 when 2n < r < 2n +
r
2
for each
integer n, and 1
0
(r) < 0 when 2n +
r
2
< r < (2n + 1). Thus, 1 is increasing on
2n. 2n +
r
2
and
180

decreasing on
2n +
r
2
. (2n + 1)
for each integer n.

F. Local maximum values 1
2n +
r
2
= 0, no local minimum.
G. 1
00
(r) = 3csc
2
r < 0, so 1 is CD on (2n. (2n + 1)) for
each integer n. No IP
H.
49. j = 1(r) = rc
3i
2
A. 1 = R B. Intercepts are 0 C. 1(3r) = 31(r), so the curve is symmetric
about the origin. D. lim
i<"
rc
3i
2
= lim
i<"
r
c
i
2
H
= lim
i<"
1
2rc
i
2
= 0, so j = 0 is a HA.
E. 1
0
(r) = c
3i
2
32r
2
c
3i
2
= c
3i
2
(1 32r
2
) 0 C r
2
<
1
2
C |r| <
1
I
2
3
1
I
2
.
1
I
2
and decreasing on
3". 3
1
I
2
and
1
I
2
. "
. F. Local maximum value 1
1
I
2
= 1
I
2c, local minimum
value 1
3
1
I
2
= 31
I
2c G. 1
00
(r) = 32rc
3i
2
(1 32r
2
) 34rc
3i
2
= 2rc
3i
2
(2r
2
33) 0 C
r

3
2
or 3
3
2
< r < 0, so 1 is CU on
3
2
. "
and
3
2
. 0
and CD on
3". 3
3
2
and
0.
3
2
.
IP are (0, 0) and
3
2
.
3
2
c
33'2
.
H.
51. j = 1(r) = c
3i
+c
32i
A. 1 = R B. j-intercept = 1(0) = 2; no r-intercept C. No symmetry D. No asymptote
E. 1
0
(r) = 3c
3i
32c
32i
, so 1
0
(r) 0 C 3c
3i
2c
32i
[multiply by c
2i
] C
c
5i
2
3
C 5r ln
2
3
C r
1
5
ln
2
3
E 30.081. Similarly, 1
0
(r) < 0 C
r <
1
5
ln
2
3
. 1 is decreasing on
3".
1
5
ln
2
3
and increasing on
1
5
ln
2
3
. "
.
F. Local minimum value 1
1
5
ln
2
3
=

2
3
3'5
+
2
3
32'5
E 1.96; no local maximum.
G. 1
00
(r) = 9c
3i
+ 4c
32i
, so 1
00
(r) 0 for all r, and 1 is CU on (3". "). No IP
H.
53. i = 1() =
i
0
1 3
2
c
2
. The m-intercept is 1(0) = i
0
. There are no -intercepts. lim
r<c
1() = ", so = c is a VA.

1
0
() = 3
1
2
i
0
(1 3
2
c
2
)
33'2
(32c
2
) =
i
0
c
2
(1 3
2
c
2
)
3'2
=
i
0
c
2
(c
2
3
2
)
3'2
c
3
=
i
0
c
(c
2
3
2
)
3'2
0, so 1 is
increasing on (0. c). There are no local extreme values.
1
00
() =
(c
2
3
2
)
3'2
(i
0
c) 3i
0
c
3
2
(c
2
3
2
)
1'2
(32)
[(c
2
3
2
)
3'2
]
2
=
i
0
c(c
2
3
2
)
1'2
[(c
2
3
2
) + 3
2
]
(c
2
3
2
)
3
=
i
0
c(c
2
+ 2
2
)
(c
2
3
2
)
5'2
0,
so 1 is CU on (0. c). There are no inection points.

181
55. j = 3
W
2411
r
4
+
W1
1211
r
3
3
W1
2
2411
r
2
= 3
W
2411
r
2
r
2
321r +1
2
=
3W
2411
r
2
(r 31)
2
= cr
2
(r 31)
2
where c = 3
W
2411
is a negative constant and 0 $ r $ 1. We sketch
1(r) = cr
2
(r 31)
2
for c = 31. 1(0) = 1(1) = 0.
1
0
(r) = cr
2
[2(r 31)] + (r 31)
2
(2cr) = 2cr(r 31)[r + (r 31)] = 2cr(r 31)(2r 31). So for 0 < r < 1,
1
0
(r) 0 C r(r 31)(2r 31) < 0 [since c < 0] C 12 < r < 1 and 1
0
(r) < 0 C 0 < r < 12.
Thus, 1 is increasing on (12. 1) and decreasing on (0. 12), and there is a local and absolute
minimum at the point (12. 1(12)) =

12. c1
4
16
. 1
0
(r) = 2c[r(r 31)(2r 31)] i
1
00
(r) = 2c[1(r 31)(2r 31) +r(1)(2r 31) +r(r 31)(2)] = 2c(6r
2
361r +1
2
) = 0 C
r =
61
I
121
2
12
=
1
2
1
I
3
6
1, and these are the r-coordinates of the two inection points.
57. j =
r
2
+ 1
r + 1
. Long division gives us: r 31
r + 1 r
2
+ 1
r
2
+r
3r + 1
3r 31
2
Thus, j = 1(r) =
r
2
+ 1
r + 1
= r 31 +
2
r + 1
and 1(r) 3(r 31) =
2
r + 1
=
2
r
1 +
1
r
[for r 6= 0] <0 as r <".
So the line j = r 31 is a slant asymptote (SA).
59. j =
4r
3
32r
2
+ 5
2r
2
+r 33
. Long division gives us: 2r 32
2r
2
+r 33 4r
3
32r
2
+ 5
4r
3
+ 2r
2
36r
3 4r
2
+ 6r + 5
3 4r
2
32r + 6
8r 31
Thus, j = 1(r) =
4r
3
32r
2
+ 5
2r
2
+r 33
= 2r 32 +
8r 31
2r
2
+r 33
and 1(r) 3(2r 32) =
8r 31
2r
2
+r 33
=
8
r
3
1
r
2
2 +
1
r
3
3
r
2
[for r 6= 0] <0 as r <". So the line j = 2r 32 is a SA.
61. j = 1(r) =
32r
2
+ 5r 31
2r 31
= 3r + 2 +
1
2r 31
A. 1 =

r M R | r 6=
1
2
=

3".
1
2
1
2
. "
B. j-intercept: 1(0) = 1; r-intercepts: 1(r) = 0 i 32r

2
+ 5r 31 = 0 i r =
35
I
17
34
i r E 0.22, 2.28.
[continued]
182

i<(1'2)
1(r) = 3"and lim

i<(1'2)
+
1(r) = ", so r =
1
2
is a VA.
lim
i<"
[1(r) 3(3r + 2)] = lim
i<"
1
2r 31
= 0, so the line j = 3r + 2 is a SA.
E. 1
0
(r) = 31 3
2
(2r 31)
2
< 0 for r 6=
1
2
, so 1 is decreasing on
3".
1
2
and
1
2
. "
. F. No extreme values G. 1
0
(r) = 31 32(2r 31)
32
i
1
00
(r) = 32(32)(2r 31)
33
(2) =
8
(2r 31)
3
, so 1
00
(r) 0 when r
1
2
and
1
00
(r) < 0 when r <
1
2
. Thus, 1 is CU on
1
2
. "
and CD on
3".
1
2
. No IP
H.
63. j = 1(r) = (r
2
+ 4)r = r + 4r A. 1 = {r | r 6= 0} = (3". 0) (0. ") B. No intercept
C. 1(3r) = 31(r) i symmetry about the origin D. lim
i<"
(r + 4r) = "but 1(r) 3r = 4r <0 as r <",
so j = r is a slant asymptote. lim
i<0
+
(r + 4r) = "and
lim
i<0
(r + 4r) = 3", so r = 0 is a VA. E. 1

0
(r) = 1 34r
2
0 C
r
2
4 C r 2 or r < 32, so 1 is increasing on (3". 32) and (2. ") and
decreasing on (32. 0) and (0. 2). F. Local maximum value 1(32) = 34, local
minimum value 1(2) = 4 G. 1
00
(r) = 8r
3
0 C r 0 so 1 is CU on
(0. ") and CD on (3". 0). No IP
H.
65. j = 1(r) =
2r
3
+r
2
+ 1
r
2
+ 1
= 2r + 1 +
32r
r
2
+ 1
A. 1 = R B. j-intercept: 1(0) = 1; r-intercept: 1(r) = 0 i
0 = 2r
3
+r
2
+ 1 = (r + 1)(2r
2
3r + 1) i r = 31 C. No symmetry D. No VA
lim
i<"
[1(r) 3(2r + 1)] = lim
i<"
32r
r
2
+ 1
= lim
i<"
32r
1 + 1r
2
= 0, so the line j = 2r + 1 is a slant asymptote.
E. 1
0
(r) = 2 +
(r
2
+ 1)(32) 3(32r)(2r)
(r
2
+ 1)
2
=
2(r
4
+ 2r
2
+ 1) 32r
2
32 + 4r
2
(r
2
+ 1)
2
=
2r
4
+ 6r
2
(r
2
+ 1)
2
=
2r
2
(r
2
+ 3)
(r
2
+ 1)
2
so 1
0
(r) 0 if r 6= 0. Thus, 1 is increasing on (3". 0) and (0. "). Since 1 is continuous at 0, 1 is increasing on R.
F. No extreme values
G. 1
00
(r) =
(r
2
+ 1)
2
(8r
3
+ 12r) 3(2r
4
+ 6r
2
) 2(r
2
+ 1)(2r)
[(r
2
+ 1)
2
]
2
=
4r(r
2
+ 1)[(r
2
+ 1)(2r
2
+ 3) 32r
4
36r
2
]
(r
2
+ 1)
4
=
4r(3r
2
+ 3)
(r
2
+ 1)
3
so 1
00
(r) 0 for r < 3
I
3 and 0 < r <
I
3, and 1
00
(r) < 0 for
3
I
3 < r < 0 and r
I
3. 1 is CU on
3". 3
I
3
and
0.
I
3
,
and CD on
3
I
3. 0
and
I
3. "
. There are three IPs: (0. 1),
3
I
3. 3
3
2
I
3 + 1
E (31.73. 31.60), and

I
3.
3
2
I
3 + 1
E (1.73. 3.60).
H.

183
67. j = 1(r) = r 3tan
31
r, 1
0
(r) = 1 3
1
1 +r
2
=
1 +r
2
31
1 +r
2
=
r
2
1 +r
2
,
1
00
(r) =
(1 +r
2
)(2r) 3r
2
(2r)
(1 +r
2
)
2
=
2r(1 +r
2
3r
2
)
(1 +r
2
)
2
=
2r
(1 +r
2
)
2
.
lim
i<"
1(r) 3
r 3
r
2
= lim
i<"
r
2
3tan
31
r
=
r
2
3
r
2
= 0, so j = r 3
r
2
is a SA.
Also, lim
i<3"
1(r) 3
r +
r
2
= lim
i<3"
3
r
2
3tan
31
r
= 3
r
2
3
3
r
2
= 0,
so j = r +
r
2
is also a SA. 1
0
(r) D 0 for all r, with equality C r = 0, so 1 is
increasing on R. 1
00
(r) has the same sign as r, so 1 is CD on (3". 0) and CU on
(0. "). 1(3r) = 31(r), so 1 is an odd function; its graph is symmetric about the
origin. 1 has no local extreme values. Its only IP is at (0. 0).
69.
r
2
o
2
3
j
2
/
2
= 1 i j =
/
o
I
r
2
3o
2
. Now
lim
i<"
/
o
I
r
2
3o
2
3
/
o
r
=
/
o
lim
i<"
I
r
2
3o
2
3r
I
r
2
3o
2
+r
I
r
2
3o
2
+r
=
/
o
lim
i<"
3o
2
I
r
2
3o
2
+r
= 0,
which shows that j =
/
o
r is a slant asymptote. Similarly,
lim
i<"
3
/
o
I
r
2
3o
2
3
3
/
o
r
= 3
/
o
lim
i<"
3o
2
I
r
2
3o
2
+r
= 0, so j = 3
/
o
r is a slant asymptote.
71. lim
i<"
1(r) 3r
3
= lim
i<"
r
4
+ 1
r
3
r
4
r
= lim
i<"
1
r
= 0, so the graph of 1 is asymptotic to that of j = r
3
.
A. 1 = {r | r 6= 0} B. No intercept C. 1 is symmetric about the origin. D. lim
i<0
r
3
+
1
r
= 3"and
lim
i<0
+
r
3
+
1
r
= ", so r = 0 is a vertical asymptote, and as shown above, the graph of 1 is asymptotic to that of j = r
3
.
E. 1
0
(r) = 3r
2
31r
2
0 C r
4
1
3
C |r|
1
4
I
3
3". 3
1
4
I
3
and
1
4
I
3
. "
and
decreasing on
3
1
4
I
3
. 0
and
0.
1
4
I
3
. F. Local maximum value

1
3
1
4
I
3
= 34 3
35'4
, local minimum value 1
1
4
I
3
= 4 3
35'4
G. 1
00
(r) = 6r + 2r
3
0 C r 0, so 1 is CU on (0. ") and CD
on (3". 0). No IP
H.
184

4.6 Graphing with Calculus and Calculators
1. 1(r) = 4r
4
332r
3
+ 89r
2
395r + 29 i 1
0
(r) = 16r
3
396r
2
+ 178r 395 i 1
00
(r) = 48r
2
3192r + 178.
1(r) = 0 C r E 0.5, 1.60; 1
0
(r) = 0 C r E 0.92, 2.5, 2.58 and 1
00
(r) = 0 C r E 1.46, 2.54.
From the graphs of 1
0
, we estimate that 1
0
< 0 and that 1 is decreasing on (3". 0.92) and (2.5. 2.58), and that 1
0
0 and 1
is increasing on (0.92. 2.5) and (2.58. ") with local minimum values 1(0.92) E 35.12 and 1(2.58) E 3.998 and local
maximum value 1(2.5) = 4. The graphs of 1
0
make it clear that 1 has a maximum and a minimum near r = 2.5, shown more
clearly in the fourth graph.
From the graph of 1
00
, we estimate that 1
00
0 and that 1 is CU on
(3". 1.46) and (2.54. "), and that 1
00
< 0 and 1 is CD on (1.46. 2.54).
There are inection points at about (1.46. 31.40) and (2.54. 3.999).
3. 1(r) = r
6
310r
5
3400r
4
+ 2500r
3
i 1
0
(r) = 6r
5
350r
4
31600r
3
+ 7500r
2
i
1
00
(r) = 30r
4
3200r
3
34800r
2
+ 1500r
SECTION 4.6 GRAPHING WITH CALCULUS AND CALCULATORS

185
From the graph of 1
0
, we estimate that 1 is decreasing on (3". 315), increasing on (315. 4.40), decreasing
on (4.40. 18.93), and increasing on (18.93. "), with local minimum values of 1(315) E 39,700,000 and
1(18.93) E 312,700,000 and local maximum value 1(4.40) E 53,800. From the graph of 1
00
, we estimate that 1 is CU on
(3". 311.34), CD on (311.34. 0), CU on (0. 2.92), CD on (2.92. 15.08), and CU on (15.08. "). There is an inection
point at (0. 0) and at about (311.34. 36,250,000), (2.92. 31,800), and (15.08. 38,150,000).
5. 1(r) =
r
r
3
3r
2
34r + 1
i 1
0
(r) =
32r
3
+r
2
+ 1
(r
3
3r
2
34r + 1)
2
i 1
00
(r) =
2(3r
5
33r
4
+ 5r
3
36r
2
+ 3r + 4)
(r
3
3r
2
34r + 1)
3
We estimate from the graph of 1 that j = 0 is a horizontal asymptote, and that there are vertical asymptotes at r = 31.7,
r = 0.24, and r = 2.46. From the graph of 1
0
, we estimate that 1 is increasing on (3". 31.7), (31.7. 0.24), and (0.24. 1),
and that 1 is decreasing on (1. 2.46) and (2.46. "). There is a local maximum value at 1(1) = 3
1
3
. From the graph of 1
00
, we
estimate that 1 is CU on (3". 31.7), (30.506. 0.24), and (2.46. "), and that 1 is CD on (31.7. 30.506) and (0.24. 2.46).
There is an inection point at (30.506. 30.192).
7. 1(r) = r
2
34r + 7 cos r, 34 $ r $ 4. 1
0
(r) = 2r 34 37 sinr i 1
00
(r) = 2 37 cos r.
1(r) = 0 C r E 1.10; 1
0
(r) = 0 C r E 31.49, 31.07, or 2.89; 1
00
(r) = 0 C r = cos
31
2
7
E 1.28.
0
, we estimate that 1 is decreasing (1
0
< 0) on (34. 31.49), increasing on (31.49. 31.07), decreasing
on (31.07. 2.89), and increasing on (2.89. 4), with local minimum values 1(31.49) E 8.75 and 1(2.89) E 39.99 and local
maximum value 1(31.07) E 8.79 (notice the second graph of 1). From the graph of 1
00
, we estimate that 1 is CU (1
00
0)
on (34. 31.28), CD on (31.28. 1.28), and CU on (1.28. 4). There are inection points at about (31.28. 8.77)
and (1.28. 31.48).
186

9. 1(r) = 1 +
1
r
+
8
r
2
+
1
r
3
i 1
0
(r) = 3
1
r
2
3
16
r
3
3
3
r
4
= 3
1
r
4
(r
2
+ 16r + 3) i
1
00
(r) =
2
r
3
+
48
r
4
+
12
r
5
=
2
r
5
(r
2
+ 24r + 6).
From the graphs, it appears that 1 increases on (315.8. 30.2) and decreases on (3". 315.8), (30.2. 0), and (0. "); that 1
has a local minimum value of 1(315.8) E 0.97 and a local maximum value of 1(30.2) E 72; that 1 is CD on (3". 324)
and (30.25. 0) and is CU on (324. 30.25) and (0. "); and that 1 has IPs at (324. 0.97) and (30.25. 60).
To nd the exact values, note that 1
0
= 0 i r =
316
I
256 312
2
= 38
I
61 [E 30.19 and 315.81].
1
0
is positive (1 is increasing) on
38 3
I
61. 38 +
I
61
and 1
0
is negative (1 is decreasing) on
3". 38 3
I
61
38 +
I
61. 0
, and (0. "). 1

00
= 0 i r =
324
I
576 324
2
= 312
I
138 [E 30.25 and 323.75]. 1
00
is
positive (1 is CU) on
312 3
I
138. 312 +
I
138
and (0. ") and 1

00
is negative (1 is CD) on
3". 312 3
I
138
and
312 +
I
138. 0
.
11. (a) 1(r) = r
2
lnr. The domain of 1 is (0. ").
(b) lim
i<0
+
r
2
lnr = lim
i<0
+
lnr
1r
2
H
= lim
i<0
+
1r
32r
3
= lim
i<0
+
3
r
2
2
= 0.
There is a hole at (0. 0).
(c) It appears that there is an IP at about (0.2. 30.06) and a local minimum at (0.6. 30.18). 1(r) = r
2
lnr i
1
0
(r) = r
2
(1r) + (lnr)(2r) = r(2 lnr + 1) 0 C lnr 3
1
2
C r c
31'2
1
I
c. "
, decreasing on
0. 1
I
c
. By the FDT, 1
1
I
c
= 31(2c) is a local minimum value. This point is

approximately (0.6065. 30.1839), which agrees with our estimate.
1
00
(r) = r(2r) + (2 lnr + 1) = 2 lnr + 3 0 C lnr 3
3
2
C r c
33'2
, so 1 is CU on (c
33'2
. ")
and CD on (0. c
33'2
). IP is (c
33'2
. 33(2c
3
)) E (0.2231. 30.0747).

187
13. 1(r) =
(r + 4)(r 33)
2
r
4
(r 31)
has VA at r = 0 and at r = 1 since lim
i<0
1(r) = 3",
lim
i<1
1(r) = 3"and lim

i<1
+
1(r) = ".
1(r) =
r + 4
r

(r 33)
2
r
2
r
4
r
3
(r 31)
dividing numerator
and denominator by r
3
=
(1 + 4r)(1 33r)
2
r(r 31)
<0
as r <", so 1 is asymptotic to the r-axis.
Since 1 is undened at r = 0, it has no j-intercept. 1(r) = 0 i (r +4)(r 33)
2
= 0 i r = 34 or r = 3, so 1 has
r-intercepts 34 and 3. Note, however, that the graph of 1 is only tangent to the r-axis and does not cross it at r = 3, since 1 is
positive as r <3
3
and as r <3
+
.
From these graphs, it appears that 1 has three maximum values and one minimum value. The maximum values are
approximately 1(35.6) = 0.0182, 1(0.82) = 3281.5 and 1(5.2) = 0.0145 and we know (since the graph is tangent to the
r-axis at r = 3) that the minimum value is 1(3) = 0.
15. 1(r) =
r
2
(r + 1)
3
(r 32)
2
(r 34)
4
i 1
0
(r) = 3
r(r + 1)
2
(r
3
+ 18r
2
344r 316)
(r 32)
3
(r 34)
5
[from CAS].
0
, it seems that the critical points which indicate extrema occur at r E 320, 30.3, and 2.5, as estimated
in Example 3. (There is another critical point at r = 31, but the sign of 1
0
does not change there.) We differentiate again,
obtaining 1
00
(r) = 2
(r + 1)(r
6
+ 36r
5
+ 6r
4
3628r
3
+ 684r
2
+ 672r + 64)
(r 32)
4
(r 34)
6
.
00
, it appears that 1 is CU on (335.3. 35.0), (31. 30.5), (30.1. 2), (2. 4) and (4. ") and CD
188

on (3". 335.3), (35.0. 31) and (30.5. 30.1). We check back on the graphs of 1 to nd the j-coordinates of the
inection points, and nd that these points are approximately (335.3. 30.015), (35.0. 30.005), (31. 0), (30.5. 0.00001),
and (30.1. 0.0000066).
17. j = 1(r) =
I
r
r
2
+r + 1
. From a CAS, j
0
= 3
3r
2
+r 31
2
I
r(r
2
+r + 1)
2
and j
00
=
15r
4
+ 10r
3
315r
2
36r 31
4r
3'2
(r
2
+r + 1)
3
.
1
0
(r) = 0 C r E 0.43, so 1 is increasing on (0. 0.43) and decreasing on (0.43. "). There is a local maximum value of
1(0.43) E 0.41. 1
00
(r) = 0 C r E 0.94, so 1 is CD on (0. 0.94) and CU on (0.94. "). There is an inection point at
(0.94. 0.34).
19. j = 1(r) =
I
r + 5 sinr, r $ 20.
From a CAS, j
0
=
5 cos r + 1
2
I
r + 5 sinr
and j
00
= 3
10 cos r + 25 sin
2
r + 10rsinr + 26
4(r + 5 sinr)
3'2
.
Well start with a graph of o(r) = r + 5 sinr. Note that 1(r) =
o(r) is only dened if o(r) D 0. o(r) = 0 C r = 0

or r E 34.91, 34.10, 4.10, and 4.91. Thus, the domain of 1 is [34.91. 34.10] [0. 4.10] [4.91. 20].
From the expression for j
0
, we see that j
0
= 0 C 5 cos r + 1 = 0 i r
1
= cos
31
3
1
5
E 1.77 and
r
2
= 2 3r
1
E 34.51 (not in the domain of 1 ). The leftmost zero of 1
0
is r
1
32 E 34.51. Moving to the right, the zeros
of 1
0
are r
1
, r
1
+ 2, r
2
+ 2, r
1
+ 4, and r
2
+ 4. Thus, 1 is increasing on (34.91. 34.51), decreasing on
(34.51. 34.10), increasing on (0. 1.77), decreasing on (1.77. 4.10), increasing on (4.91. 8.06), decreasing on (8.06. 10.79),
increasing on (10.79. 14.34), decreasing on (14.34. 17.08), and increasing on (17.08. 20). The local maximum values are
1(34.51) E 0.62, 1(1.77) E 2.58, 1(8.06) E 3.60, and 1(14.34) E 4.39. The local minimum values are 1(10.79) E 2.43
and 1(17.08) E 3.49.
1 is CD on (34.91. 34.10), (0. 4.10), (4.91. 9.60), CU on (9.60. 12.25), CD
on (12.25. 15.81), CU on (15.81. 18.65), and CD on (18.65. 20). There are
inection points at (9.60. 2.95), (12.25. 3.27), (15.81. 3.91), and (18.65. 4.20).

189
21. j = 1(r) =
1 3c
1'i
1 +c
1'i
. From a CAS, j
0
=
2c
1'i
r
2
(1 +c
1'i
)
2
and j
00
=
32c
1'i
(1 3c
1'i
+ 2r + 2rc
1'i
)
r
4
(1 +c
1'i
)
3
.
1 is an odd function dened on (3". 0) (0. "). Its graph has no x- or y-intercepts. Since lim
i<"
1(r) = 0, the x-axis
is a HA. 1
0
(r) 0 for r 6= 0, so 1 is increasing on (3". 0) and (0. "). It has no local extreme values.
1
00
(r) = 0 for r E 0.417, so 1 is CU on (3". 30.417), CD on (30.417. 0), CU on (0. 0.417), and CD on (0.417. ").
1 has IPs at (30.417. 0.834) and (0.417. 30.834).
23. (a) 1(r) = r
1'i
(b) Recall that o
l
= c
l ln a
. lim
i<0
+
r
1'i
= lim
i<0
+
c
(1'i) ln i
. As r <0
+
,
lnr
r
<3", so r
1'i
= c
(1'i) ln i
<0. This
indicates that there is a hole at (0. 0). As r <", we have the indeterminate form "
0
. lim
i<"
r
1'i
= lim
i<"
c
(1'i) ln i
,
but lim
i<"
lnr
r
H
= lim
i<"
1r
1
= 0, so lim
i<"
r
1'i
= c
0
= 1. This indicates that j = 1 is a HA.
(c) Estimated maximum: (2.72. 1.45). No estimated minimum. We use logarithmic differentiation to nd any critical
numbers. j = r
1'i
i lnj =
1
r
lnx i
j
0
j
=
1
r

1
r
+ (lnr)
3
1
r
2
i j
0
= r
1'i
1 3lnr
r
2
= 0 i
lnr = 1 i r = c. For 0 < r < c, j
0
0 and for r c, j
0
< 0, so 1(c) = c
1'c
is a local maximum value. This
point is approximately (2.7183. 1.4447), which agrees with our estimate.
(d) From the graph, we see that 1
00
(r) = 0 at r E 0.58 and r E 4.37. Since 1
00
changes sign at these values, they are r-coordinates of inection points.
190

25.
From the graph of 1(r) = sin(r + sin3r) in the viewing rectangle [0. ] by [31.2. 1.2], it looks like 1 has two maxima
and two minima. If we calculate and graph 1
0
(r) = [cos(r + sin3r)] (1 + 3 cos 3r) on [0. 2], we see that
the graph of 1
0
appears to be almost tangent to the r-axis at about r = 0.7. The graph of
1
00
= 3[sin(r + sin3r)] (1 + 3 cos 3r)
2
+ cos(r + sin3r)(39 sin3r) is even more interesting near this r-value:
it seems to just touch the r-axis.
If we zoom in on this place on the graph of 1
00
, we see that 1
00
actually does cross the axis twice near r = 0.65,
indicating a change in concavity for a very short interval. If we look at the graph of 1
0
on the same interval, we see that it
changes sign three times near r = 0.65, indicating that what we had thought was a broad extremum at about r = 0.7 actually
consists of three extrema (two maxima and a minimum). These maximum values are roughly 1(0.59) = 1 and 1(0.68) = 1,
and the minimum value is roughly 1(0.64) = 0.99996. There are also a maximum value of about 1(1.96) = 1 and minimum
values of about 1(1.46) = 0.49 and 1(2.73) = 30.51. The points of inection on (0. ) are about (0.61. 0.99998),
(0.66. 0.99998), (1.17. 0.72), (1.75. 0.77), and (2.28. 0.34). On (. 2), they are about (4.01. 30.34), (4.54. 30.77),
(5.11. 30.72), (5.62. 30.99998), and (5.67. 30.99998). There are also IP at (0. 0) and (. 0). Note that the function is odd
and periodic with period 2, and it is also rotationally symmetric about all points of the form ((2n + 1). 0), n an integer.
27. 1(r) = r
4
+cr
2
= r
2
r
2
+c
. Note that 1 is an even function. For c D 0, the only r-intercept is the point (0. 0). We
calculate 1
0
(r) = 4r
3
+ 2cr = 4r
r
2
+
1
2
c
i 1
00
(r) = 12r
2
+ 2c. If c D 0, r = 0 is the only critical point and there
is no inection point. As we can see from the examples, there is no change in the basic shape of the graph for c D 0; it merely

191
becomes steeper as c increases. For c = 0, the graph is the simple curve j = r
4
. For c < 0, there are r-intercepts at 0
and at
I
3c. Also, there is a maximum at (0. 0), and there are
minima at
3
1
2
c. 3
1
4
c
2
. As c <3", the r-coordinates of

these minima get larger in absolute value, and the minimum points
move downward. There are inection points at
3
1
6
c. 3
5
36
c
2
,
which also move away from the origin as c <3".
29.
c = 0 is a transitional value we get the graph of j = 1. For c 0, we see that there is a HA at j = 1, and that the graph
spreads out as c increases. At rst glance there appears to be a minimum at (0. 0), but 1(0) is undened, so there is no
minimum or maximum. For c < 0, we still have the HA at j = 1, but the range is (1. ") rather than (0. 1). We also have
a VA at r = 0. 1(r) = c
3c'i
2
i 1
0
(r) = c
3c'i
2
2c
r
3
i 1
00
(r) =
2c(2c 33r
2
)
r
6
c
c'i
2
.
1
0
(r) 6= 0 and 1
0
(r) exists for all r 6= 0 (and 0 is not in the domain of 1), so there are no maxima or minima.
1
00
(r) = 0 i r =
2c3, so if c 0, the inection points spread out as c increases, and if c < 0, there are no IP.
For c 0. there are IP at
2c3. c
33'2
. Note that the j-coordinate of the IP is constant.

31. Note that c = 0 is a transitional value at which the graph consists of the r-axis. Also, we can see that if we substitute 3c for c,
the function 1(r) =
cr
1 +c
2
r
2
will be reected in the r-axis, so we investigate only positive values of c (except c = 31, as a
demonstration of this reective property). Also, 1 is an odd function. lim
i<"
1(r) = 0, so j = 0 is a horizontal asymptote for
all c. We calculate 1
0
(r) =
(1 +c
2
r
2
)c 3cr(2c
2
r)
(1 +c
2
r
2
)
2
= 3
c(c
2
r
2
31)
(1 +c
2
r
2
)
2
. 1
0
(r) = 0 C c
2
r
2
31 = 0 C r = 1c.
So there is an absolute maximum value of 1(1c) =
1
2
and an absolute minimum value of 1(31c) = 3
1
2
. These extrema
have the same value regardless of c, but the maximum points move closer to the j-axis as c increases.
1
00
(r) =
(32c
3
r)(1 +c
2
r
2
)
2
3(3c
3
r
2
+c)[2(1 +c
2
r
2
)(2c
2
r)]
(1 +c
2
r
2
)
4
=
(32c
3
r)(1 +c
2
r
2
) + (c
3
r
2
3c)(4c
2
r)
(1 +c
2
r
2
)
3
=
2c
3
r(c
2
r
2
33)
(1 +c
2
r
2
)
3
1
00
(r) = 0 C r = 0 or
I
3c, so there are inection points at (0. 0) and
at
I
3c.
I
34
. Again, the j-coordinate of the inection points does not

depend on c, but as c increases, both inection points approach the j-axis.
192

33. 1(r) = cr + sinr i 1
0
(r) = c + cos r i 1
00
(r) = 3sinr
1(3r) = 31(r), so 1 is an odd function and its graph is symmetric with respect to the origin.
1(r) = 0 C sinr = 3cr, so 0 is always an r-intercept.
1
0
(r) = 0 C cos r = 3c, so there is no critical number when |c| 1. If |c| $ 1, then there are innitely
many critical numbers. If r
1
is the unique solution of cos r = 3c in the interval [0. ], then the critical numbers are 2n r
1
,
where n ranges over the integers. (Special cases: When c = 1, r
1
= 0; when c = 0, r =
r
2
; and when c = 31, r
1
= .)
1
00
(r) < 0 C sinr 0, so 1 is CD on intervals of the form (2n. (2n + 1)). 1 is CU on intervals of the form
((2n 31). 2n). The inection points of 1 are the points (2n. 2nc), where n is an integer.
If c D 1, then 1
0
(r) D 0 for all r, so 1 is increasing and has no extremum. If c $ 31, then 1
0
(r) $ 0 for all r, so 1 is
decreasing and has no extremum. If |c| < 1, then 1
0
(r) 0 C cos r 3c C r is in an interval of the form
(2n 3r
1
. 2n +r
1
) for some integer n. These are the intervals on which 1 is increasing. Similarly, we
nd that 1 is decreasing on the intervals of the form (2n +r
1
. 2(n + 1) 3r
1
). Thus, 1 has local maxima at the points
2n +r
1
, where 1 has the values c(2n +r
1
) + sinr
1
= c(2n +r
1
) +
I
1 3c
2
, and 1 has local minima at the points
2n 3r
1
, where we have 1(2n 3r
1
) = c(2n 3r
1
) 3sinr
1
= c(2n 3r
1
) 3
I
1 3c
2
.
The transitional values of c are 31 and 1. The inection points move vertically, but not horizontally, when c changes.
When |c| D 1, there is no extremum. For |c| < 1, the maxima are spaced
2 apart horizontally, as are the minima. The horizontal spacing between
maxima and adjacent minima is regular (and equals ) when c = 0, but
the horizontal space between a local maximum and the nearest local
minimum shrinks as |c| approaches 1.
35. If c < 0, then lim
i<3"
1(r) = lim
i<3"
rc
3ci
= lim
i<3"
r
c
ci
H
= lim
i<3"
1
cc
ci
= 0, and lim
i<"
1(r) = ".
If c 0, then lim
i<3"
1(r) = 3", and lim
i<"
1(r)
H
= lim
i<"
1
cc
ci
= 0.
If c = 0, then 1(r) = r, so lim
i<"
1(r) = ", respectively.
So we see that c = 0 is a transitional value. We now exclude the case c = 0, since we know how the function behaves
in that case. To nd the maxima and minima of 1, we differentiate: 1(r) = rc
3ci
i
1
0
(r) = r(3cc
3ci
) +c
3ci
= (1 3cr)c
3ci
. This is 0 when 1 3cr = 0 C r = 1c. If c < 0 then this
represents a minimum value of 1(1c) = 1(cc), since 1
0
(r) changes from negative to positive at r = 1c;
SECTION 4.7 OPTIMIZATION PROBLEMS

193
and if c 0, it represents a maximum value. As |c| increases, the maximum or
minimum point gets closer to the origin. To nd the inection points, we
differentiate again: 1
0
(r) = c
3ci
(1 3cr) i
1
00
(r) = c
3ci
(3c) + (1 3cr)(3cc
3ci
) = (cr 32)cc
3ci
. This changes sign
when cr 32 = 0 C r = 2c. So as |c| increases, the points of inection get
closer to the origin.
37. (a) 1(r) = cr
4
32r
2
+ 1. For c = 0, 1(r) = 32r
2
+ 1, a parabola whose vertex, (0. 1), is the absolute maximum. For
c 0, 1(r) = cr
4
32r
2
+ 1 opens upward with two minimum points. As c <0, the minimum points spread apart and
move downward; they are below the r-axis for 0 < c < 1 and above for c 1. For c < 0, the graph opens downward, and
has an absolute maximum at r = 0 and no local minimum.
(b) 1
0
(r) = 4cr
3
34r = 4cr(r
2
31c) [c 6= 0]. If c $ 0, 0 is the only critical number.
1
00
(r) = 12cr
2
34, so 1
00
(0) = 34 and there is a local maximum at
(0. 1(0)) = (0. 1), which lies on j = 1 3r
2
. If c 0, the critical
numbers are 0 and 1
I
c. As before, there is a local maximum at
(0. 1(0)) = (0. 1), which lies on j = 1 3r
2
.
1
00
1
I
c
= 12 34 = 8 0, so there is a local minimum at

r = 1
I
c. Here 1
1
I
c
= c(1c
2
) 32c + 1 = 31c + 1.
But
1
I
c. 31c + 1
lies on j = 1 3r
2
since 1 3
1
I
c
2
= 1 31c.
4.7 Optimization Problems
1. (a)
First Number Second Number Product
1 22 22
2 21 42
3 20 60
4 19 76
5 18 90
6 17 102
7 16 112
8 15 120
9 14 126
10 13 130
11 12 132
We neednt consider pairs where the rst number
is larger than the second, since we can just
interchange the numbers in such cases. The
answer appears to be 11 and 12, but we have
considered only integers in the table.
194

(b) Call the two numbers r and j. Then r +j = 23, so j = 23 3r. Call the product 1. Then
1 = rj = r(23 3r) = 23r 3r
2
, so we wish to maximize the function 1(r) = 23r 3r
2
. Since 1
0
(r) = 23 32r,
we see that 1
0
(r) = 0 C r =
23
2
= 11.5. Thus, the maximum value of 1 is 1(11.5) = (11.5)
2
= 132.25 and it
occurs when r = j = 11.5.
Or: Note that 1
00
(r) = 32 < 0 for all r, so 1 is everywhere concave downward and the local maximum at r = 11.5
must be an absolute maximum.
3. The two numbers are r and
100
r
, where r 0. Minimize 1(r) = r +
100
r
. 1
0
(r) = 1 3
100
r
2
=
r
2
3100
r
2
. The critical
number is r = 10. Since 1
0
(r) < 0 for 0 < r < 10 and 1
0
(r) 0 for r 10, there is an absolute minimum at r = 10.
The numbers are 10 and 10.
5. If the rectangle has dimensions r and j, then its perimeter is 2r + 2j = 100 m, so j = 50 3r. Thus, the area is
= rj = r(50 3r). We wish to maximize the function (r) = r(50 3r) = 50r 3r
2
, where 0 < r < 50. Since
0
(r) = 50 32r = 32(r 325),
0
(r) 0 for 0 < r < 25 and
0
(r) < 0 for 25 < r < 50. Thus, has an absolute
maximum at r = 25, and (25) = 25
2
= 625 m
2
. The dimensions of the rectangle that maximize its area are r = j = 25 m.
(The rectangle is a square.)
7. We need to maximize Y for ` D 0. Y (`) =
I`
1 +`
2
i
Y
0
(`) =
(1 +`
2
)I 3I`(2`)
(1 +`
2
)
2
=
I(1 3`
2
)
(1 +`
2
)
2
=
I(1 +`)(1 3`)
(1 +`
2
)
2
. Y
0
(`) 0 for 0 < ` < 1 and Y
0
(`) < 0
for ` 1. Thus, Y has an absolute maximum of Y (1) =
1
2
I at ` = 1.
9. (a)
The areas of the three gures are 12,500, 12,500, and 9000 ft
2
. There appears to be a maximum area of at least 12,500 ft
2
.
(b) Let r denote the length of each of two sides and three dividers.
Let j denote the length of the other two sides.
(c) Area = length width = j r
(d) Length of fencing = 750 i 5r + 2j = 750
(e) 5r + 2j = 750 i j = 375 3
5
2
r i (r) =
375 3
5
2
r
r = 375r 3
5
2
r
2
(f )
0
(r) = 375 35r = 0 i r = 75. Since
00
(r) = 35 < 0 there is an absolute maximum when r = 75. Then
j =
375
2
= 187.5. The largest area is 75
375
2
= 14,062.5 ft
2
. These values of r and j are between the values in the rst
and second gures in part (a). Our original estimate was low.

195
11. rj = 1.5 10
6
, so j = 1.5 10
6
r. Minimize the amount of fencing, which is
3r + 2j = 3r + 2(1.5 10
6
r) = 3r + 3 10
6
r = 1(r).
1
0
(r) = 3 33 10
6
r
2
= 3(r
2
310
6
)r
2
. The critical number is r = 10
3
and
1
0
(r) < 0 for 0 < r < 10
3
and 1
0
(r) 0 if r 10
3
, so the absolute minimum
occurs when r = 10
3
and j = 1.5 10
3
.
The eld should be 1000 feet by 1500 feet with the middle fence parallel to the short side of the eld.
13. Let / be the length of the base of the box and / the height. The surface area is 1200 = /
2
+4// i / = (1200 3/
2
)(4/).
The volume is \ = /
2
/ = /
2
(1200 3/
2
)4/ = 300/ 3/
3
4 i \
0
(/) = 300 3
3
4
/
2
.
\
0
(/) = 0 i 300 =
3
4
/
2
i /
2
= 400 i / =
I
400 = 20. Since \
0
(/) 0 for 0 < / < 20 and \
0
(/) < 0 for
/ 20, there is an absolute maximum when / = 20 by the First Derivative Test for Absolute Extreme Values (see page 324).
If / = 20, then / = (1200 320
2
)(4 20) = 10, so the largest possible volume is /
2
/ = (20)
2
(10) = 4000 cm
3
.
15. 10 = (2n)(n)/ = 2n
2
/, so / = 5n
2
. The cost is
C(n) = 10(2n
2
) + 6[2(2n/) + 2/n] + 6(2n
2
)
= 32n
2
+ 36n/ = 32n
2
+ 180n
C
0
(n) = 64n 3180n
2
= 4(16n
3
345)n
2
i n =
3
45
16
is the critical number. C
0
(n) < 0 for 0 < n <
3
45
16
and
C
0
(n) 0 for n
3
45
16
. The minimum cost is C
45
16
= 32(2.8125)
2'3
+ 180
I
2.8125 E $191.28.
17. The distance from a point (r. j) on the line j = 4r + 7 to the origin is
(r 30)
2
+ (j 30)
2
=

r
2
+j
2
. However, it is
easier to work with the square of the distance; that is, 1(r) =

r
2
+j
2
2
= r
2
+j
2
= r
2
+ (4r + 7)
2
. Because the
distance is positive, its minimum value will occur at the same point as the minimum value of 1.
1
0
(r) = 2r + 2(4r + 7)(4) = 34r + 56, so 1
0
(r) = 0 C r = 3
28
17
.
1
00
(r) = 34 0, so 1 is concave upward for all r. Thus, 1 has an absolute minimum at r = 3
28
17
. The point closest to the
origin is (r. j) =
3
28
17
. 4
3
28
17
+ 7
=

3
28
17
.
7
17
.
19. From the gure, we see that there are two points that are farthest away from
(1. 0). The distance d from to an arbitrary point 1(r. j) on the ellipse is
d =

(r 31)
2
+ (j 30)
2
and the square of the distance is
o = d
2
= r
2
32r + 1 +j
2
= r
2
32r + 1 + (4 34r
2
) = 33r
2
32r + 5.
o
0
= 36r 32 and o
0
= 0 i r = 3
1
3
. Now o
00
= 36 < 0, so we know
that o has a maximum at r = 3
1
3
. Since 31 $ r $ 1, o(31) = 4,
o
3
1
3
=
16
3
, and o(1) = 0, we see that the maximum distance is
16
3
. The corresponding j-values are
j =
4 34
3
1
3
2
=
32
9
=
4
3
I
2 E 1.89. The points are
3
1
3
.
4
3
I
2
.
196

21. The area of the rectangle is (2r)(2j) = 4rj. Also v
2
= r
2
+j
2
so
j =
I
v
2
3r
2
, so the area is (r) = 4r
I
v
2
3r
2
. Now
0
(r) = 4
I
v
2
3r
2
3
r
2
I
v
2
3r
2
= 4
v
2
32r
2
I
v
2
3r
2
. The critical number is
r =
1
I
2
v. Clearly this gives a maximum.
j =
v
2
3
1
I
2
v
2
=

1
2
v
2
=
1
I
2
v = r, which tells us that the rectangle is a square. The dimensions are 2r =
I
2v
and 2j =
I
2v.
23. The height / of the equilateral triangle with sides of length 1 is
I
3
2
1,
since /
2
+ (12)
2
= 1
2
i /
2
= 1
2
3
1
4
1
2
=
3
4
1
2
i
/ =
I
3
2
1. Using similar triangles,
I
3
2
1 3j
r
=
I
3
2
1
12
=
I
3 i
I
3 r =
I
3
2
1 3j i j =
I
3
2
1 3
I
3 r i j =
I
3
2
(1 32r).
The area of the inscribed rectangle is (r) = (2r)j =
I
3 r(1 32r) =
I
3 1r 32
I
3 r
2
, where 0 $ r $ 12. Now
0 =
0
(r) =
I
3 1 34
I
3 r i r =
I
3 1
4
I
3
= 14. Since (0) = (12) = 0, the maximum occurs when

r = 14, and j =
I
3
2
1 3
I
3
4
1 =
I
3
4
1, so the dimensions are 12 and
I
3
4
1.
25. The area of the triangle is
(r) =
1
2
(2t)(v +r) = t(v +r) =
I
v
2
3r
2
(v +r). Then
0 =
0
(r) = v
32r
2
I
v
2
3r
2
+
I
v
2
3r
2
+r
32r
2
I
v
2
3r
2
= 3
r
2
+vr
I
v
2
3r
2
+
I
v
2
3r
2
i
r
2
+vr
I
v
2
3r
2
=
I
v
2
3r
2
i r
2
+vr = v
2
3r
2
i 0 = 2r
2
+vr 3v
2
= (2r 3v)(r +v) i
r =
1
2
v or r = 3v. Now (v) = 0 = (3v) i the maximum occurs where r =
1
2
v, so the triangle has height
v +
1
2
v =
3
2
v and base 2
v
2
3
1
2
v
2
= 2
3
4
v
2
=
I
3v.
27. The cylinder has volume \ = j
2
(2r). Also r
2
+j
2
= v
2
i j
2
= v
2
3r
2
, so
\ (r) = (v
2
3r
2
)(2r) = 2(v
2
r 3r
3
), where 0 $ r $ v.
\
0
(r) = 2
v
2
33r
2
= 0 i r = v
I
3. Now \ (0) = \ (v) = 0, so there is a
maximum when r = v
I
3 and \
v
I
3
= (v
2
3v
2
3)
2v
I
3
= 4v
3
3
I
3
.

197
29. The cylinder has surface area
2(area of the base) + (lateral surface area) = 2(radius)
2
+ 2(radius)(height)
= 2j
2
+ 2j(2r)
Now r
2
+j
2
= v
2
i j
2
= v
2
3r
2
i j =
I
v
2
3r
2
, so the surface area is
o(r) = 2(v
2
3r
2
) + 4r
I
v
2
3r
2
. 0 $ r $ v
= 2v
2
32r
2
+ 4
r
I
v
2
3r
2
Thus, o
0
(r) = 0 34r + 4
r
1
2
(v
2
3r
2
)
31'2
(32r) + (v
2
3r
2
)
1'2
1
= 4
3r 3
r
2
I
v
2
3r
2
+
I
v
2
3r
2
= 4
3r
I
v
2
3r
2
3r
2
+v
2
3r
2
I
v
2
3r
2
o
0
(r) = 0 i r
I
v
2
3r
2
= v
2
32r
2
(-) i

r
I
v
2
3r
2
2
= (v
2
32r
2
)
2
i
r
2
(v
2
3r
2
) = v
4
34v
2
r
2
+ 4r
4
i v
2
r
2
3r
4
= v
4
34v
2
r
2
+ 4r
4
i 5r
4
35v
2
r
2
+v
4
= 0.
This is a quadratic equation in r
2
. By the quadratic formula, r
2
=
5
I
5
10
v
2
, but we reject the root with the + sign since it
doesnt satisfy (-). [The right side is negative and the left side is positive.] So r =

5 3
I
5
10
v. Since o(0) = o(v) = 0, the
maximum surface area occurs at the critical number and r
2
=
5 3
I
5
10
v
2
i j
2
= v
2
3
5 3
I
5
10
v
2
=
5 +
I
5
10
v
2
i
the surface area is
2
5 +
I
5
10
v
2
+ 4
5 3
I
5
10
5 +
I
5
10
v
2
= v
2
2
5+
I
5
10
+ 4
t
(53
I
5)(5+
I
5)
10
= v
2
5+
I
5
5
+
2
I
20
5
= v
2
5+
I
5 +22
I
5
5
= v
2
5 +5
I
5
5
= v
2
1 +
I
5
.
31. rj = 384 i j = 384r. Total area is
(r) = (8 +r)(12 + 384r) = 12(40 +r + 256r), so
0
(r) = 12(1 3256r
2
) = 0 i r = 16. There is an absolute minimum
when r = 16 since
0
(r) < 0 for 0 < r < 16 and
0
(r) 0 for r 16.
When r = 16, j = 38416 = 24, so the dimensions are 24 cm and 36 cm.
33. Let r be the length of the wire used for the square. The total area is
(r) =

r
4
2
+
1
2
10 3r
3
I
3
2
10 3r
3
=
1
16
r
2
+
I
3
36
(10 3r)
2
, 0 $ r $ 10
0
(r) =
1
8
r 3
I
3
18
(10 3r) = 0 C
9
72
r +
4
I
3
72
r 3
40
I
3
72
= 0 C r =
40
I
3
9 +4
I
3
. Now (0) =

I
3
36
100 E 4.81,
(10) =
100
16
= 6.25 and
40
I
3
9 +4
I
3
E 2.72, so
(a ) The maximum area occurs when r = 10 m, and all the wire is used for the square.
(b) The minimum area occurs when r =
40
I
3
9 +4
I
3
E 4.35 m.
198

35. The volume is \ = v
2
/ and the surface area is
o(v) = v
2
+ 2v/ = v
2
+ 2v
\
v
2
= v
2
+
2\
v
.
o
0
(v) = 2v 3
2\
v
2
= 0 i 2v
3
= 2\ i v =
3
cm.
This gives an absolute minimum since o
0
(v) < 0 for 0 < v <
3
and o
0
(v) 0 for v
3
.
When v =
3
, / =
\
v
2
=
\
(\)
2'3
=
3
cm.
37. /
2
+v
2
= 1
2
i \ =
r
3
v
2
/ =
r
3
(1
2
3/
2
)/ =
r
3
(1
2
/ 3/
3
).
\
0
(/) =
r
3
(1
2
33/
2
) = 0 when / =
1
I
3
1. This gives an absolute maximum, since
\
0
(/) 0 for 0 < / <
1
I
3
1 and \
0
(/) < 0 for /
1
I
3
1. The maximum volume is
\
1
I
3
1
=
r
3
1
I
3
1
3
3
1
3
I
3
1
3
=
2
9
I
3
1
3
.
1
1
=
1 3/
v
(1). The volume of the inner cone is \ =
1
3
v
2
/,
so well solve (1) for /.
1v
1
= 1 3/ i
/ = 1 3
1v
1
=
1131v
1
=
1
1
(13v) (2).
Thus, \(v) =

3
v
2
1
1
(13v) =
1
31
(1v
2
3v
3
) i
\
0
(v) =
1
31
(21v 33v
2
) =
1
31
v(2133v).
\
0
(v) = 0 i v = 0 or 21 = 3v i v =
2
3
1 and from (2), / =
1
1
13
2
3
1
=
1
1
1
3
1
=
1
3
1.
\
0
(v) changes from positive to negative at v =
2
3
1, so the inner cone has a maximum volume of
\ =
1
3
v
2
/ =
1
3
2
3
1
1
3
1
=
4
27

1
3
1
2
1, which is approximately 15% of the volume of the larger cone.
41. 1(1) =
1
2
1
(1+v)
2
i
1
0
(1) =
(1+v)
2
1
2
31
2
1 2(1+v)
[(1+v)
2
]
2
=
(1
2
+ 21v +v
2
)1
2
321
2
1
2
321
2
1v
(1+v)
4
=
1
2
v
2
31
2
1
2
(1+v)
4
=
1
2
(v
2
31
2
)
(1+v)
4
=
1
2
(v +1)(v 31)
(1+v)
4
=
1
2
(v 31)
(1+v)
3
1
0
(1) = 0 i 1 = v i 1(v) =
1
2
v
(v +v)
2
=
1
2
v
4v
2
=
1
2
4v
.
The expression for 1
0
(1) shows that 1
0
(1) 0 for 1 < v and 1
0
(1) < 0 for 1 v. Thus, the maximum value of the
power is 1
2
(4v), and this occurs when 1 = v.

199
43. o = 6c/ 3
3
2
c
2
cot 0 + 3c
2
I
3
2
csc 0
(a)
do
d0
=
3
2
c
2
csc
2
0 33c
2
I
3
2
csc 0 cot 0 or
3
2
c
2
csc 0
csc 0 3
I
3 cot 0
.
(b)
do
d0
= 0 when csc 0 3
I
3 cot 0 = 0 i
1
sin0
3
I
3
cos 0
sin0
= 0 i cos 0 =
1
I
3
. The First Derivative Test shows
that the minimum surface area occurs when 0 = cos
31
1
I
3
E 55
.
(c) If cos 0 =
1
I
3
, then cot 0 =
1
I
2
and csc 0 =
I
3
I
2
, so the surface area is
o = 6c/ 3
3
2
c
2 1
I
2
+ 3c
2
I
3
2
I
3
I
2
= 6c/ 3
3
2
I
2
c
2
+
9
2
I
2
c
2
= 6c/ +
6
2
I
2
c
2
= 6c
/ +
1
2
I
2
c
45. Here T(r) =

I
r
2
+ 25
6
+
5 3r
8
, 0 $ r $ 5 i T
0
(r) =
r
6
I
r
2
+ 25
3
1
8
= 0 C 8r = 6
I
r
2
+ 25 C
16r
2
= 9(r
2
+ 25) C r =
15
I
7
. But
15
I
7
5, so T has no critical number. Since T(0) E 1.46 and T(5) E 1.18, he
should row directly to 1.
47. There are (6 3r) km over land and
I
r
2
+ 4 km under the river.
We need to minimize the cost C (measured in $100,000) of the pipeline.
C(r) = (6 3r)(4) +
I
r
2
+ 4
(8) i
C
0
(r) = 34 + 8
1
2
(r
2
+ 4)
31'2
(2r) = 34 +
8r
I
r
2
+ 4
.
C
0
(r) = 0 i 4 =
8r
I
r
2
+ 4
i
I
r
2
+ 4 = 2r i r
2
+ 4 = 4r
2
i 4 = 3r
2
i r
2
=
4
3
i
r = 2
I
3 [0 $ r $ 6]. Compare the costs for r = 0, 2
I
3, and 6. C(0) = 24 + 16 = 40,
C
2
I
3
= 24 38
I
3 + 32
I
3 = 24 + 24
I
3 E 37.9, and C(6) = 0 + 8
I
40 E 50.6. So the minimum cost is about
$3.79 million when 1 is 6 32
I
3 E 4.85 km east of the renery.
49. The total illumination is 1(r) =
3I
r
2
+
I
(10 3r)
2
, 0 < r < 10. Then
1
0
(r) =
36I
r
3
+
2I
(10 3r)
3
= 0 i 6I(10 3r)
3
= 2Ir
3
i
3(10 3r)
3
= r
3
i
3
I
3 (10 3r) = r i 10
3
I
3 3
3
I
3 r = r i 10
3
I
3 = r +
3
I
3 r i
10
3
I
3 =
1 +
3
I
3
r i r =
10
3
I
3
1 +
3
I
3
E 5.9 ft. This gives a minimum since 1
00
(r) 0 for 0 < r < 10.
200

51. Every line segment in the rst quadrant passing through (o. /) with endpoints on the r-
and j-axes satises an equation of the form j 3/ = i(r 3o), where i < 0. By setting
r = 0 and then j = 0, we nd its endpoints, (0. / 3oi) and 1
o 3
l
r
. 0
. The
distance d from to 1 is given by d =

[
o 3
l
r
30]
2
+ [0 3(/ 3oi)]
2
.
It follows that the square of the length of the line segment, as a function of i, is given by
o(i) =
o 3
/
i
2
+ (oi3/)
2
= o
2
3
2o/
i
+
/
2
i
2
+o
2
i
2
32o/i+/
2
. Thus,
o
0
(i) =
2o/
i
2
3
2/
2
i
3
+ 2o
2
i32o/ =
2
i
3
(o/i3/
2
+o
2
i
4
3o/i
3
)
=
2
i
3
[/(oi3/) +oi
3
(oi3/)] =
2
i
3
(oi3/)(/ +oi
3
)
Thus, o
0
(i) = 0 C i = /o or i = 3
3
l
a
. Since /o 0 and i < 0, imust equal 3
3
l
a
. Since
2
i
3
< 0, we see
that o
0
(i) < 0 for i < 3
3
l
a
and o
0
(i) 0 for i 3
3
l
a
. Thus, o has its absolute minimum value when i = 3
3
l
a
.
That value is
o
3
3
l
a
=

o +/
3
a
l
2
+
3o
3
l
a
3/
2
=

o +
3
I
o/
2
2
+
3
I
o
2
/ +/
2
= o
2
+ 2o
4'3
/
2'3
+o
2'3
/
4'3
+o
4'3
/
2'3
+ 2o
2'3
/
4'3
+/
2
= o
2
+ 3o
4'3
/
2'3
+ 3o
2'3
/
4'3
+/
2
The last expression is of the form r
3
+ 3r
2
j + 3rj
2
+j
3
[= (r +j)
3
] with r = o
2'3
and j = /
2'3
,
so we can write it as (o
2'3
+/
2'3
)
3
and the shortest such line segment has length
I
o = (o
2'3
+/
2'3
)
3'2
.
53. (a) If c(r) =
C(r)
r
, then, by Quotient Rule, we have c
0
(r) =
rC
0
(r) 3C(r)
r
2
. Now c
0
(r) = 0 when rC
0
(r) 3C(r) = 0
and this gives C
0
(r) =
C(r)
r
= c(r). Therefore, the marginal cost equals the average cost.
(b) (i) C(r) = 16,000 + 200r + 4r
3'2
, C(1000) = 16,000 + 200,000 + 40,000
I
10 E 216,000 + 126,491, so
C(1000) E $342,491. c(r) = C(r)r =
16,000
r
+ 200 + 4r
1'2
, c(1000) E $342.49unit. C
0
(r) = 200 + 6r
1'2
,
C
0
(1000) = 200 + 60
I
10 E $389.74unit.
(ii) We must have C
0
(r) = c(r) C 200 + 6r
1'2
=
16,000
r
+ 200 + 4r
1'2
C 2r
3'2
= 16,000 C
r = (8,000)
2'3
= 400 units. To check that this is a minimum, we calculate
c
0
(r) =
316,000
r
2
+
2
I
r
=
2
r
2
(r
3'2
38000). This is negative for r < (8000)
2'3
= 400, zero at r = 400,
and positive for r 400, so c is decreasing on (0. 400) and increasing on (400. "). Thus, c has an absolute minimum
at r = 400. [Note: c
00
(r) is not positive for all r 0.]
(iii) The minimum average cost is c(400) = 40 + 200 + 80 = $320unit.

201
55. (a) We are given that the demand function j is linear and j(27,000) = 10, j(33,000) = 8, so the slope is
10 38
27,000 333,000
= 3
1
3000
and an equation of the line is j 310 =

3
1
3000
(r 327,000) i
j = j(r) = 3
1
3000
r + 19 = 19 3(r3000).
(b) The revenue is 1(r) = rj(r) = 19r 3(r
2
3000) i 1
0
(r) = 19 3(r1500) = 0 when r = 28,500. Since
1
00
(r) = 311500 < 0, the maximum revenue occurs when r = 28,500 i the price is j(28,500) = $9.50.
57. (a) As in Example 6, we see that the demand function j is linear. We are given that j(1000) = 450 and deduce that
j(1100) = 440, since a $10 reduction in price increases sales by 100 per week. The slope for j is
440 3450
1100 31000
= 3
1
10
,
so an equation is j 3450 = 3
1
10
(r 31000) or j(r) = 3
1
10
r + 550.
(b) 1(r) = rj(r) = 3
1
10
r
2
+ 550r. 1
0
(r) = 3
1
5
r + 550 = 0 when r = 5(550) = 2750.
j(2750) = 275, so the rebate should be 450 3275 = $175.
(c) C(r) = 68,000 + 150r i 1(r) = 1(r) 3C(r) = 3
1
10
r
2
+ 550r 368,000 3150r = 3
1
10
r
2
+ 400r 368,000,
1
0
(r) = 3
1
5
r + 400 = 0 when r = 2000. j(2000) = 350. Therefore, the rebate to maximize prots should be
450 3350 = $100.
59. Here c
2
= /
2
+/
2
4, so /
2
= c
2
3/
2
4. The area is =
1
2
/
c
2
3/
2
4.
Let the perimeter be j, so 2c +/ = j or c = (j 3/)2 i
(/) =
1
2
/
(j 3/)
2
4 3/
2
4 = /
j
2
32j/4. Now
0
(/) =
j
2
32j/
4
3
/j4
j
2
32j/
=
33j/ +j
2
4
j
2
32j/
.
Therefore,
0
(/) = 0 i 33j/ +j
2
= 0 i / = j3. Since
0
(/) 0 for / < j3 and
0
(/) < 0 for / j3, there
is an absolute maximum when / = j3. But then 2c +j3 = j, so c = j3 i c = / i the triangle is equilateral.
61. Note that |1| = |1| + |11| i 5 = r + |11| i |11| = 5 3r.
Using the Pythagorean Theorem for {111 and {11C gives us
1(r) = |1| + |11| + |C1| = r +
(5 3r)
2
+ 2
2
+
(5 3r)
2
+ 3
2
= r +
I
r
2
310r + 29 +
I
r
2
310r + 34 i
1
0
(r) = 1 +
r 35
I
r
2
310r + 29
+
r 35
I
r
2
310r + 34
. From the graphs of 1
and 1
0
, it seems that the minimum value of 1 is about 1(3.59) = 9.35 m.
202

63. The total time is
T(r) = (time from to C) + (time from C to 1)
=
I
o
2
+r
2
1
+
/
2
+ (d 3r)
2
2
, 0 < r < d
T
0
(r) =
r
1
I
o
2
+r
2
3
d 3r
/
2
+ (d 3r)
2
=
sin0
1
1
3
sin0
2
2
The minimum occurs when T
0
(r) = 0 i
sin0
1
1
=
sin0
2
2
.
[Note: T
00
(r) 0]
65. j
2
= r
2
+:
2
, but triangles C11 and 1C are similar, so
:8 = r
4
I
r 34
i : = 2r
I
r 34. Thus, we minimize
1(r) = j
2
= r
2
+ 4r
2
(r 34) = r
3
(r 34), 4 < r $ 8.
1
0
(r) =
(r 34)(3r
2
) 3r
3
(r 34)
2
=
r
2
[3(r 34) 3r]
(r 34)
2
=
2r
2
(r 36)
(r 34)
2
= 0
when r = 6. 1
0
(r) < 0 when r < 6, 1
0
(r) 0 when r 6, so the minimum
occurs when r = 6 in.
67. It sufces to maximize tan0. Now
3t
1
= tan(w +0) =
tanw + tan0
1 3tanw tan0
=
t + tan0
1 3t tan0
. So
3t(1 3t tan0) = t + tan0 i 2t = (1 + 3t
2
) tan0 i tan0 =
2t
1 + 3t
2
.
Let 1(t) = tan0 =
2t
1 + 3t
2
i 1
0
(t) =
2
1 + 3t
2
32t(6t)
(1 + 3t
2
)
2
=
2
1 33t
2
(1 + 3t
2
)
2
= 0 C 1 33t
2
= 0 C
t =
1
I
3
since t D 0. Now 1
0
(t) 0 for 0 $ t <
1
I
3
and 1
0
(t) < 0 for t
1
I
3
, so 1 has an absolute maximum when t =
1
I
3
and tan0 =
2
1
I
3
1 + 3
1
I
3
2
=
1
I
3
i 0 =
r
6
. Substituting for t and 0 in 3t = tan(w +0) gives us
I
3 = tan
w +
r
6
i w =
r
6
.

203
69. From the gure, tanc =
5
r
and tano =
2
3 3r
. Since
c +o +0 = 180
= , 0 = 3tan
31
5
r
3tan
31
2
3 3r
i
d0
dr
= 3
1
1 +
5
r
3
5
r
2
3
1
1 +
2
3 3r
2
(3 3r)
2
=
r
2
r
2
+ 25

5
r
2
3
(3 3r)
2
(3 3r)
2
+ 4

2
(3 3r)
2
.
Now
d0
dr
= 0 i
5
r
2
+ 25
=
2
r
2
36r + 13
i 2r
2
+ 50 = 5r
2
330r + 65 i
3r
2
330r + 15 = 0 i r
2
310r + 5 = 0 i r = 5 2
I
5. We reject the root with the + sign, since it is larger
than 3. d0dr 0 for r < 5 32
I
5 and d0dr < 0 for r 5 32
I
5, so 0 is maximized when
|1| = r = 5 32
I
5 E 0.53.
71. In the small triangle with sides o and c and hypotenuse W, sin0 =
o
W
and
cos 0 =
c
W
. In the triangle with sides / and d and hypotenuse 1, sin0 =
d
1
and
cos 0 =
/
1
. Thus, o = W sin0, c = W cos 0, d = 1sin0, and / = 1cos 0, so the
area of the circumscribed rectangle is
(0) = (o +/)(c +d) = (W sin0 +1cos 0)(W cos 0 +1sin0)
= W
2
sin0 cos 0 +W1sin
2
0 +1W cos
2
0 +1
2
sin0 cos 0
= 1W sin
2
0 +1W cos
2
0 + (1
2
+W
2
) sin0 cos 0
= 1W(sin
2
0 + cos
2
0) + (1
2
+W
2
)
1
2
2 sin0 cos 0 = 1W +
1
2
(1
2
+W
2
) sin20, 0 $ 0 $
r
2
This expression shows, without calculus, that the maximum value of (0) occurs when sin20 = 1 C 20 =
r
2
i
0 =
r
4
. So the maximum area is
r
4
= 1W +
1
2
(1
2
+W
2
) =
1
2
(1
2
+ 21W +W
2
) =
1
2
(1 +W)
2
.
73. (a) If I = energykm over land, then energykm over water = 1.4I.
So the total energy is 1 = 1.4I
I
25 +r
2
+I(13 3r), 0 $ r $ 13,
and so
d1
dr
=
1.4Ir
(25 +r
2
)
1'2
3I.
Set
d1
dr
= 0: 1.4Ir = I(25 +r
2
)
1'2
i 1.96r
2
= r
2
+ 25 i 0.96r
2
= 25 i r =
5
I
0.96
E 5.1.
Testing against the value of 1 at the endpoints: 1(0) = 1.4I(5) + 13I = 20I, 1(5.1) E 17.9I, 1(13) E 19.5I.
Thus, to minimize energy, the bird should y to a point about 5.1 km from 1.
204

(b) If W1 is large, the bird would y to a point C that is closer to 1 than to 1 to minimize the energy used ying over water.
If W1 is small, the bird would y to a point C that is closer to 1 than to 1 to minimize the distance of the ight.
1 = W
I
25 +r
2
+1(13 3r) i
d1
dr
=
Wr
I
25 +r
2
31 = 0 when
W
1
=
I
25 +r
2
r
. By the same sort of
argument as in part (a), this ratio will give the minimal expenditure of energy if the bird heads for the point r km from 1.
(c) For ight direct to 1, r = 13, so from part (b), W1 =
I
25 +13
2
13
E 1.07. There is no value of W1 for which the bird
should y directly to 1. But note that lim
i<0
+
(W1) = ", so if the point at which 1 is a minimum is close to 1, then
W1 is large.
(d) Assuming that the birds instinctively choose the path that minimizes the energy expenditure, we can use the equation for
d1dr = 0 from part (a) with 1.4I = c, r = 4, and I = 1: c(4) = 1 (25 + 4
2
)
1'2
i c =
I
414 E 1.6.
4.8 Newton's Method
1. (a) The tangent line at r = 1 intersects the r-axis at r E 2.3, so
r
2
E 2.3. The tangent line at r = 2.3 intersects the r-axis at
r E 3, so r
3
E 3.0.
(b) r
1
= 5 would not be a better rst approximation than r
1
= 1 since the tangent line is nearly horizontal. In fact, the second
approximation for r
1
= 5 appears to be to the left of r = 1.
3. Since r
1
= 3 and j = 5r 34 is tangent to j = 1(r) at r = 3, we simply need to nd where the tangent line intersects the
r-axis. j = 0 i 5r
2
34 = 0 i r
2
=
4
5
.
5. 1(r) = r
3
+ 2r 34 i 1
0
(r) = 3r
2
+ 2, so r
n+1
= r
n
3
r
3
n
+ 2r
n
34
3r
2
n
+ 2
. Now r
1
= 1 i
r
2
= 1 3
1 + 2 34
3 1
2
+ 2
= 1 3
31
5
= 1.2 i r
3
= 1.2 3
(1.2)
3
+ 2(1.2) 34
3(1.2)
2
+ 2
E 1.1797.
7. 1(r) = r
5
3r 31 i 1
0
(r) = 5r
4
31, so r
n+1
= r
n
3
r
5
n
3r
n
31
5r
4
n
31
. Now r
1
= 1 i
r
2
= 1 3
1 31 31
5 31
= 1 3
3
1
4
= 1.25 i r
3
= 1.25 3
(1.25)
5
31.25 31
5(1.25)
4
31
E 1.1785.
9. 1(r) = r
3
+r + 3 i 1
0
(r) = 3r
2
+ 1, so r
n+1
= r
n
3
r
3
n
+r
n
+ 3
3r
2
n
+ 1
.
Now r
1
= 31 i
r
2
= 31 3
(31)
3
+ (31) + 3
3(31)
2
+ 1
= 31 3
31 31 + 3
3 + 1
= 31 3
1
4
= 31.25.
Newtons method follows the tangent line at (31. 1) down to its intersection with
the r-axis at (31.25. 0), giving the second approximation r
2
= 31.25.
SECTION 4.8 NEWTONS METHOD

205
11. To approximate r =
5
I
20 (so that r
5
= 20), we can take 1(r) = r
5
320. So 1
0
(r) = 5r
4
, and thus,
r
n+1
= r
n
3
r
5
n
320
5r
4
n
. Since
5
I
32 = 2 and 32 is reasonably close to 20. well use r
1
= 2. We need to nd approximations
until they agree to eight decimal places. r
1
= 2 i r
2
= 1.85, r
3
E 1.82148614, r
4
E 1.82056514,
r
5
E 1.82056420 Er
6
. So
5
I
20 E 1.82056420, to eight decimal places.
Here is a quick and easy method for nding the iterations for Newtons method on a programmable calculator.
(The screens shown are from the TI-84 Plus, but the method is similar on other calculators.) Assign 1(r) = r
5
320
to Y
1
, and 1
0
(r) = 5r
4
to Y
2
. Now store r
1
= 2 in X and then enter X3Y
1
Y
2
<X to get r
2
= 1.85. By successively
pressing the ENTER key, you get the approximations r
3
, r
4
, . . . .
In Derive, load the utility le SOLVE. Enter NEWTON(x5-20,x,2) and then APPROXIMATE to get
[2, 1.85, 1.82148614, 1.82056514, 1.82056420]. You can request a specic iteration by adding a fourth argument. For
example, NEWTON(x5-20,x,2,2) gives [2. 1.85. 1.82148614].
In Maple, make the assignments 1 := r <r5 320;, o := r <r 31(r)1(1)(r);, and r:= 2.;. Repeatedly execute
the command r := o(r); to generate successive approximations.
In Mathematica, make the assignments 1[r_ ] := r5 320, o[r_ ] := r 31[r]1
0
[r], and r = 2. Repeatedly execute the
command r = o[r] to generate successive approximations.
13. 1(r) = r
4
32r
3
+ 5r
2
36 i 1
0
(r) = 4r
3
36r
2
+ 10r i r
n+1
= r
n
3
r
4
n
32r
3
n
+ 5r
2
n
36
4r
3
n
36r
2
n
+ 10r
n
. We need to nd
approximations until they agree to six decimal places. Well let r
1
equal the midpoint of the given interval, [1. 2].
r
1
= 1.5 i r
2
= 1.2625, r
3
E 1.218808, r
4
E 1.217563, r
5
E 1.217562 E r
6
. So the root is 1.217562 to six decimal
places.
15. sinr = r
2
, so 1(r) = sinr 3r
2
i 1
0
(r) = cos r 32r i
r
n+1
= r
n
3
sinr
n
3r
2
n
cos r
n
32r
n
. From the gure, the positive root of sin r = r
2
is
near 1. r
1
= 1 i r
2
E 0.891396, r
3
E 0.876985, r
4
E 0.876726 E r
5
. So
the positive root is 0.876726, to six decimal places.
206

17. From the graph, we see that there appear to be points of intersection near
r = 30.7 and r = 1.2. Solving r
4
= 1 +r is the same as solving
1(r) = r
4
3r 31 = 0. 1(r) = r
4
3r 31 i 1
0
(r) = 4r
3
31,
so r
n+1
= r
n
3
r
4
n
3r
n
31
4r
3
n
31
.
r
1
= 30.7 r
1
= 1.2
r
2
E 30.725253 r
2
E 1.221380
r
3
E 30.724493 r
3
E 1.220745
r
4
E 30.724492 E r
5
r
4
E 1.220744 E r
5
To six decimal places, the roots of the equation are 30.724492 and 1.220744.
19. From the graph, we see that there appear to be points of intersection near
r = 1.5 and r = 3. Solving (r 32)
2
= lnr is the same as solving
1(r) = (r 32)
2
3lnr = 0. 1(r) = (r 32)
2
3lnr i
1
0
(r) = 2(r 32) 31r, so r
n+1
= r
n
3
(r
n
32)
2
3lnr
n
2(r
n
32) 31r
n
.
r
1
= 1.5 r
1
= 3
r
2
E 1.406721 r
2
E 3.059167
r
3
E 1.412370 r
3
E 3.057106
r
4
E 1.412391 E r
5
r
4
E 3.057104 E r
5
To six decimal places, the roots of the equation are 1.412391 and 3.057104.
21. From the graph, there appears to be a point of intersection near r = 0.6.
Solving cos r =
I
r is the same as solving 1(r) = cos r 3
I
r = 0.
1(r) = cos r 3
I
r i 1
0
(r) = 3sinr 31
2
I
r
, so
r
n+1
= r
n
3
cos r
n
3
I
r
n
3sinr
n
31
2
I
r
. Now r
1
= 0.6 i r
2
E 0.641928,
r
3
E 0.641714 E r
4
. To six decimal places, the root of the equation is 0.641714.
23. 1(r) = r
6
3r
5
36r
4
3r
2
+r + 10 i
1
0
(r) = 6r
5
35r
4
324r
3
32r + 1 i
r
n+1
= r
n
3
r
6
n
3r
5
n
36r
4
n
3r
2
n
+r
n
+ 10
6r
5
n
35r
4
n
324r
3
n
32r
n
+ 1
.
From the graph of 1, there appear to be roots near 31.9, 31.2, 1.1, and 3.
SECTION 4.8 NEWTONS METHOD

207
r
1
= 31.9
r
2
E 31.94278290
r
3
E 31.93828380
r
4
E 31.93822884
r
5
E 31.93822883 E r
6
r
1
= 31.2
r
2
E 31.22006245
r
3
E 31.21997997 E r
4
r
1
= 1.1
r
2
E 1.14111662
r
3
E 1.13929741
r
4
E 1.13929375 E r
5
r
1
= 3
r
2
E 2.99
r
3
E 2.98984106
r
4
E 2.98984102 E r
5
To eight decimal places, the roots of the equation are 31.93822883. 31.21997997. 1.13929375, and 2.98984102.
25. From the graph, j = r
2
I
2 3r 3r
2
and j = 1 intersect twice, at r E 32 and
at r E 31. 1(r) = r
2
I
2 3r 3r
2
31 i
1
0
(r) = r
2
1
2
(2 3r 3r
2
)
31'2
(31 32r) + (2 3r 3r
2
)
1'2
2r
=
1
2
r(2 3r 3r
2
)
31'2
[r(31 32r) + 4(2 3r 3r
2
)]
=
r(8 35r 36r
2
)
2
(2 +r)(1 3r)
,
so r
n+1
= r
n
3
r
2
n
I
2 3r
n
3r
2
n
31
r
n
(8 35r
n
36r
2
n
)
2
(2 +r
n
)(1 3r
n
)
. Trying r
1
= 32 wont work because 1
0
(32) is undened, so well
try r
1
= 31.95.
r
1
= 31.95
r
2
E 31.98580357
r
3
E 31.97899778
r
4
E 31.97807848
r
5
E 31.97806682
r
6
E 31.97806681 E r
7
r
1
= 30.8
r
2
E 30.82674444
r
3
E 30.82646236
r
4
E 30.82646233 E r
5
To eight decimal places, the roots of the equation are 31.97806681 and 30.82646233.
27. Solving 4c
3i
2
sinr = r
2
3r + 1 is the same as solving
1(r) = 4c
3i
2
sinr 3r
2
+r 31 = 0.
1
0
(r) = 4c
3i
2
(cos r 32rsinr) 32r + 1 i
r
n+1
= r
n
3
4c
3i
2
q
sinr
n
3r
2
n
+r
n
31
4c
3i
2
q
(cos r
n
32r
n
sinr
n
) 32r
n
+ 1
.
From the gure, we see that the graphs intersect at approximately r = 0.2 and r = 1.1.
r
1
= 0.2
r
2
E 0.21883273
r
3
E 0.21916357
r
4
E 0.21916368 E r
5
r
1
= 1.1
r
2
E 1.08432830
r
3
E 1.08422462 E r
4
To eight decimal places, the roots of the equation are 0.21916368 and 1.08422462.
208

29. (a) 1(r) = r
2
3 o i 1
0
(r) = 2r, so Newtons method gives
r
n+1
= r
n
3
r
2
n
3o
2r
n
= r
n
3
1
2
r
n
+
o
2r
n
=
1
2
r
n
+
o
2r
n
=
1
2
r
n
+
o
r
n
.
(b) Using (a) with o = 1000 and r
1
=
I
900 = 30, we get r
2
E 31.666667, r
3
E 31.622807, and r
4
E 31.622777 E r
5
.
So
I
1000 E 31.622777.
31. 1(r) = r
3
33r + 6 i 1
0
(r) = 3r
2
33. If r
1
= 1, then 1
0
(r
1
) = 0 and the tangent line used for approximating r
2
is
horizontal. Attempting to nd r
2
results in trying to divide by zero.
33. For 1(r) = r
1'3
, 1
0
(r) =
1
3
r
32'3
and
r
n+1
= r
n
3
1(r
n
)
1
0
(r
n
)
= r
n
3
r
1'3
n
1
3
r
32'3
n
= r
n
33r
n
= 32r
n
.
Therefore, each successive approximation becomes twice as large as the
previous one in absolute value, so the sequence of approximations fails to
converge to the root, which is 0. In the gure, we have r
1
= 0.5,
r
2
= 32(0.5) = 31, and r
3
= 32(31) = 2.
35. (a) 1(r) = r
6
3r
4
+ 3r
3
32r i 1
0
(r) = 6r
5
34r
3
+ 9r
2
32 i
1
00
(r) = 30r
4
312r
2
+ 18r. To nd the critical numbers of 1, well nd the
zeros of 1
0
. From the graph of 1
0
, it appears there are zeros at approximately
r = 31.3, 30.4, and 0.5. Try r
1
= 31.3 i
r
2
= r
1
3
1
0
(r
1
)
1
00
(r
1
)
E 31.293344 i r
3
E 31.293227 E r
4
.
Now try r
1
= 30.4 i r
2
E 30.443755 i r
3
E 30.441735 i r
4
E 30.441731 E r
5
. Finally try
r
1
= 0.5 i r
2
E 0.507937 i r
3
E 0.507854 E r
4
. Therefore, r = 31.293227, 30.441731, and 0.507854 are
all the critical numbers correct to six decimal places.
(b) There are two critical numbers where 1
0
changes from negative to positive, so 1 changes from decreasing to increasing.
1(31.293227) E 32.0212 and 1(0.507854) E 30.6721, so 32.0212 is the absolute minimum value of 1 correct to four
decimal places.
37. From the gure, we see that j = 1(r) = c
cos i
is periodic with period 2. To
nd the r-coordinates of the IP, we only need to approximate the zeros of j
00
on [0. ]. 1
0
(r) = 3c
cos i
sinr i 1
00
(r) = c
cos i
sin
2
r 3cos r
. Since
c
cos i
6= 0, we will use Newtons method with o(r) = sin
2
r 3cos r,
o
0
(r) = 2 sinr cos r + sinr, and r
1
= 1. r
2
E 0.904173,
r
3
E 0.904557 E r
4
. Thus, (0.904557. 1.855277) is the IP.
SECTION 4.9 ANTIDERIVATIVES

209
39. We need to minimize the distance from (0. 0) to an arbitrary point (r. j) on the
curve j = (r 31)
2
. d =

r
2
+j
2
i
d(r) =
r
2
+ [(r 31)
2
]
2
=

r
2
+ (r 31)
4
. When d
0
= 0, d will be
minimized and equivalently, c = d
2
will be minimized, so we will use Newtons
method with 1 = c
0
and 1
0
= c
00
.
1(r) = 2r + 4(r 31)
3
i 1
0
(r) = 2 + 12(r 31)
2
, so r
n+1
= r
n
3
2r
n
+ 4(r
n
31)
3
2 + 12(r
n
31)
2
. Try r
1
= 0.5 i
r
2
= 0.4, r
3
E 0.410127, r
4
E 0.410245 E r
5
. Now d(0.410245) E 0.537841 is the minimum distance and the point on
the parabola is (0.410245. 0.347810), correct to six decimal places.
41. In this case, = 18,000, 1 = 375, and n = 5(12) = 60. So the formula =
1
j
[1 3(1 +j)
3n
] becomes
18,000 =
375
r
[1 3(1 +r)
360
] C 48r = 1 3(1 +r)
360
[multiply each term by (1 +r)
60
] C
48r(1 +r)
60
3(1 +r)
60
+ 1 = 0. Let the LHS be called 1(r), so that
1
0
(r) = 48r(60)(1 +r)
59
+ 48(1 +r)
60
360(1 +r)
59
= 12(1 +r)
59
[4r(60) + 4(1 +r) 35] = 12(1 +r)
59
(244r 31)
r
n+1
= r
n
3
48r
n
(1 +r
n
)
60
3(1 +r
n
)
60
+ 1
12(1 +r
n
)
59
(244r
n
31)
. An interest rate of 1% per month seems like a reasonable estimate for
r = j. So let r
1
= 1% = 0.01, and we get r
2
E 0.0082202, r
3
E 0.0076802, r
4
E 0.0076291, r
5
E 0.0076286 E r
6
.
Thus, the dealer is charging a monthly interest rate of 0.76286% (or 9.55% per year, compounded monthly).
4.9 Antiderivatives
1. 1(r) = r 33 = r
1
33 i 1(r) =
r
1+1
1 + 1
33r +C =
1
2
r
2
33r +C
Check: 1
0
(r) =
1
2
(2r) 33 + 0 = r 33 = 1(r)
3. 1(r) =
1
2
+
3
4
r
2
3
4
5
r
3
i 1(r) =
1
2
r +
3
4
r
2+1
2 + 1
3
4
5
r
3+1
3 + 1
+C =
1
2
r +
1
4
r
3
3
1
5
r
4
+C
Check: 1
0
(r) =
1
2
+
1
4
(3r
2
) 3
1
5
(4r
3
) + 0 =
1
2
+
3
4
r
2
3
4
5
r
3
= 1(r)
5. 1(r) = (r + 1)(2r 31) = 2r
2
+r 31 i 1(r) = 2
1
3
r
3
+
1
2
r
2
3r +C =
2
3
r
3
+
1
2
r
2
3r +C
7. 1(r) = 5r
1'4
37r
3'4
i 1(r) = 5
r
1'4 +1
1
4
+ 1
37
r
3'4 +1
3
4
+ 1
+C = 5
r
5'4
54
37
r
7'4
74
+C = 4r
5'4
34r
7'4
+C
9. 1(r) = 6
I
r 3
6
I
r = 6r
1'2
3r
1'6
i
1(r) = 6
r
1'2+1
1
2
+ 1
3
r
1'6+1
1
6
+ 1
+C = 6
r
3'2
32
3
r
7'6
76
+C = 4r
3'2
3
6
7
r
7'6
+C
210

11. 1(r) =
10
r
9
= 10r
39
has domain (3". 0) (0. "), so 1(r) =
10r
38
38
+C
1
= 3
5
4r
8
+C
1
if r < 0
3
5
4r
8
+C
2
if r 0
See Example 1(b) for a similar problem.
13. 1(&) =
&
4
+ 3
I
&
&
2
=
&
4
&
2
+
3&
1'2
&
2
= &
2
+ 3&
33'2
i
1(&) =
&
3
3
+ 3
&
33'2+1
332 + 1
+C =
1
3
&
3
+ 3
&
31'2
312
+C =
1
3
&
3
3
6
I
&
+C
15. o(0) = cos 0 35 sin0 i G(0) = sin0 35(3cos 0) +C = sin0 + 5 cos 0 +C
17. 1(r) = 5c
i
33 coshr i 1(r) = 5c
i
33 sinhr +C
19. 1(r) =
r
5
3r
3
+ 2r
r
4
= r 3
1
r
+
2
r
3
= r 3
1
r
+ 2r
33
i
1(r) =
r
2
2
3ln|r| + 2
r
33+1
33 + 1
+C =
1
2
r
2
3ln|r| 3
1
r
2
+C
21. 1(r) = 5r
4
32r
5
i 1(r) = 5
r
5
5
32
r
6
6
+C = r
5
3
1
3
r
6
+C.
1(0) = 4 i 0
5
3
1
3
0
6
+C = 4 i C = 4, so 1(r) = r
5
3
1
3
r
6
+ 4.
The graph conrms our answer since 1(r) = 0 when 1 has a local maximum, 1 is
positive when 1 is increasing, and 1 is negative when 1 is decreasing.
23. 1
00
(r) = 6r + 12r
2
i 1
0
(r) = 6
r
2
2
+ 12
r
3
3
+ C = 3r
2
+ 4r
3
+C i
1(r) = 3
r
3
3
+ 4
r
4
4
+Cr +1 = r
3
+r
4
+Cr +1 [C and 1 are just arbitrary constants]
25. 1
00
(r) =
2
3
r
2'3
i 1
0
(r) =
2
3
r
5'3
53
+C =
2
5
r
5'3
+C i 1(r) =
2
5
r
8'3
83
+Cr +1 =
3
20
r
8'3
+Cr +1
27. 1
000
(t) = c
I
i 1
00
(t) = c
I
+C i 1
0
(t) = c
I
+Ct +1 i 1(t) = c
I
+
1
2
Ct
2
+1t +1
29. 1
0
(r) = 1 36r i 1(r) = r 33r
2
+C. 1(0) = C and 1(0) = 8 i C = 8, so 1(r) = r 33r
2
+ 8.
31. 1
0
(r) =
I
r(6 + 5r) = 6r
1'2
+ 5r
3'2
i 1(r) = 4r
3'2
+ 2r
5'2
+C.
1(1) = 6 +C and 1(1) = 10 i C = 4, so 1(r) = 4r
3'2
+ 2r
5'2
+ 4.
33. 1
0
(t) = 2 cos t + sec
2
t i 1(t) = 2 sint + tant +C because 32 < t < 2.
1
r
3
= 2
I
32
+
I
3 +C = 2
I
3 +C and 1
r
3
= 4 i C = 4 32
I
3, so 1(t) = 2 sint + tant + 4 32
I
3.

211
35. 1
0
(r) = r
31'3
has domain (3". 0) (0. ") i 1(r) =
3
2
r
2'3
+C
1
if r 0
3
2
r
2'3
+C
2
if r < 0
1(1) =
3
2
+C
1
and 1(1) = 1 i C
1
= 3
1
2
. 1(31) =
3
2
+C
2
and 1(31) = 31 i C
2
= 3
5
2
.
Thus, 1(r) =
3
2
r
2'3
3
1
2
if r 0
3
2
r
2'3
3
5
2
if r < 0
37. 1
00
(r) = 24r
2
+ 2r + 10 i 1
0
(r) = 8r
3
+r
2
+ 10r +C. 1
0
(1) = 8 + 1 + 10 +C and 1
0
(1) = 33 i
19 +C = 33 i C = 322, so 1
0
(r) = 8r
3
+r
2
+ 10r 322 and hence, 1(r) = 2r
4
+
1
3
r
3
+ 5r
2
322r +1.
1(1) = 2 +
1
3
+ 5 322 +1 and 1(1) = 5 i 1 = 22 3
7
3
=
59
3
, so 1(r) = 2r
4
+
1
3
r
3
+ 5r
2
322r +
59
3
.
39. 1
00
(0) = sin0 + cos 0 i 1
0
(0) = 3cos 0 + sin0 +C. 1
0
(0) = 31 +C and 1
0
(0) = 4 i C = 5, so
1
0
(0) = 3cos 0 + sin0 + 5 and hence, 1(0) = 3sin0 3cos 0 + 50 +1. 1(0) = 31 +1 and 1(0) = 3 i 1 = 4,
so 1(0) = 3sin0 3cos 0 + 50 + 4.
41. 1
00
(r) = 2 312r i 1
0
(r) = 2r 36r
2
+C i 1(r) = r
2
32r
3
+Cr +1.
1(0) = 1 and 1(0) = 9 i 1 = 9. 1(2) = 4 316 + 2C + 9 = 2C 33 and 1(2) = 15 i 2C = 18 i
C = 9, so 1(r) = r
2
32r
3
+ 9r + 9.
43. 1
00
(r) = 2 + cos r i 1
0
(r) = 2r + sinr +C i 1(r) = r
2
3cos r +Cr +1.
1(0) = 31 +1 and 1(0) = 31 i 1 = 0. 1
r
2
=
2
4 +
r
2
C and 1
r
2
= 0 i

r
2
C = 3
2
4 i
C = 3
r
2
, so 1(r) = r
2
3cos r 3
r
2
r.
45. 1
00
(r) = r
32
, r 0 i 1
0
(r) = 31r +C i 1(r) = 3ln|r| +Cr +1 = 3lnr +Cr +1 [since r 0].
1(1) = 0 i C +1 = 0 and 1(2) = 0 i 3ln2 +2C +1 = 0 i 3ln2 +2C 3C = 0 [since 1 = 3C] i
3ln2 +C = 0 i C = ln2 and 1 = 3ln2. So 1(r) = 3lnr + (ln2)r 3ln2.
47. Given 1
0
(r) = 2r +1, we have 1(r) = r
2
+r +C. Since 1 passes through (1. 6), 1(1) = 6 i 1
2
+1 +C = 6 i
C = 4. Therefore, 1(r) = r
2
+r + 4 and 1(2) = 2
2
+ 2 + 4 = 10.
49. / is the antiderivative of 1. For small r, 1 is negative, so the graph of its antiderivative must be decreasing. But both o and c
are increasing for small r, so only / can be 1s antiderivative. Also, 1 is positive where / is increasing, which supports our
conclusion.
51. The graph of 1 must start at (0. 1). Where the given graph, j = 1(r), has a
local minimum or maximum, the graph of 1 will have an inection point.
Where 1 is negative (positive), 1 is decreasing (increasing).
Where 1 changes from negative to positive, 1 will have a minimum.
Where 1 changes from positive to negative, 1 will have a maximum.
Where 1 is decreasing (increasing), 1 is concave downward (upward).
212

53.
1
0
(r) =
2 if 0 $ r < 1
1 if 1 < r < 2
31 if 2 < r $ 3
i 1(r) =
2r +C if 0 $ r < 1
r +1 if 1 < r < 2
3r +1 if 2 < r $ 3
1(0) = 31 i 2(0) +C = 31 i C = 31. Starting at the point
(0. 31) and moving to the right on a line with slope 2 gets us to the point (1. 1).
The slope for 1 < r < 2 is 1, so we get to the point (2. 2). Here we have used the fact that 1 is continuous. We can include the
point r = 1 on either the rst or the second part of 1. The line connecting (1. 1) to (2. 2) is j = r, so 1 = 0. The slope for
2 < r $ 3 is 31, so we get to (3. 1). 1(3) = 1 i 33 +1 = 1 i 1 = 4. Thus
1(r) =
2r 31 if 0 $ r $ 1
r if 1 < r < 2
3r + 4 if 2 $ r $ 3
Note that 1
0
(r) does not exist at r = 1 or at r = 2.
55. 1(r) =
sinr
1 +r
2
, 32 $ r $ 2
Note that the graph of 1 is one of an odd function, so the graph of 1 will
be one of an even function.
57. (t) = c
0
(t) = sint 3cos t i c(t) = 3cos t 3sint +C. c(0) = 31 +C and c(0) = 0 i C = 1, so
c(t) = 3cos t 3sint + 1.
59. o(t) =
0
(t) = t 32 i (t) =
1
2
t
2
32t +C. (0) = C and (0) = 3 i C = 3, so (t) =
1
2
t
2
32t + 3 and
c(t) =
1
6
t
3
3t
2
+ 3t +1. c(0) = 1 and c(0) = 1 i 1 = 1, and c(t) =
1
6
t
3
3t
2
+ 3t + 1.
61. o(t) =
0
(t) = 10 sint + 3 cos t i (t) = 310 cos t + 3 sint +C i c(t) = 310 sint 33 cos t +Ct +1.
c(0) = 33 +1 = 0 and c(2) = 33 + 2C +1 = 12 i 1 = 3 and C =
6
r
. Thus,
c(t) = 310 sint 33 cos t +
6
r
t + 3.
63. (a) We rst observe that since the stone is dropped 450 m above the ground, (0) = 0 and c(0) = 450.
0
(t) = o(t) = 39.8 i (t) = 39.8t +C. Now (0) = 0 i C = 0, so (t) = 39.8t i
c(t) = 34.9t
2
+1. Last, c(0) = 450 i 1 = 450 i c(t) = 450 34.9t
2
.
(b) The stone reaches the ground when c(t) = 0. 450 34.9t
2
= 0 i t
2
= 4504.9 i t
1
=

4504.9 E 9.58 s.

213
(c) The velocity with which the stone strikes the ground is (t
1
) = 39.8
4504.9 E 393.9 ms.

(d) This is just reworking parts (a) and (b) with (0) = 35. Using (t) = 39.8t +C, (0) = 35 i 0 +C = 35 i
(t) = 39.8t 35. So c(t) = 34.9t
2
35t +1 and c(0) = 450 i 1 = 450 i c(t) = 34.9t
2
35t + 450.
Solving c(t) = 0 by using the quadratic formula gives us t =

5
I
8845
(39.8) i t
1
E 9.09 s.
65. By Exercise 64 with o = 39.8, c(t) = 34.9t
2
+
0
t +c
0
and (t) = c
0
(t) = 39.8t +
0
. So
[(t)]
2
= (39.8t +
0
)
2
= (9.8)
2
t
2
319.6
0
t +
2
0
=
2
0
+ 96.04t
2
319.6
0
t =
2
0
319.6
34.9t
2
+
0
t
.
But 34.9t
2
+
0
t is just c(t) without the c
0
term; that is, c(t) 3c
0
. Thus, [(t)]
2
=
2
0
319.6 [c(t) 3c
0
].
67. Using Exercise 64 with o = 332,
0
= 0, and c
0
= / (the height of the cliff ), we know that the height at time t is
c(t) = 316t
2
+/. (t) = c
0
(t) = 332t and (t) = 3120 i 332t = 3120 i t = 3.75, so
0 = c(3.75) = 316(3.75)
2
+/ i / = 16(3.75)
2
= 225 ft.
69. Marginal cost = 1.92 30.002r = C
0
(r) i C(r) = 1.92r 30.001r
2
+1. But C(1) = 1.92 30.001 +1 = 562 i
1 = 560.081. Therefore, C(r) = 1.92r 30.001r
2
+ 560.081 i C(100) = 742.081, so the cost of producing
100 items is $742.08.
71. Taking the upward direction to be positive we have that for 0 $ t $ 10 (using the subscript 1 to refer to 0 $ t $ 10),
o
1
(t) = 3(9 30.9t) =
0
1
(t) i
1
(t) = 39t + 0.45t
2
+
0
, but
1
(0) =
0
= 310 i
1
(t) = 39t + 0.45t
2
310 = c
0
1
(t) i c
1
(t) = 3
9
2
t
2
+ 0.15t
3
310t +c
0
. But c
1
(0) = 500 = c
0
i
c
1
(t) = 3
9
2
t
2
+ 0.15t
3
310t + 500. c
1
(10) = 3450 + 150 3100 + 500 = 100, so it takes
more than 10 seconds for the raindrop to fall. Now for t 10, o(t) = 0 =
0
(t) i
(t) = constant =
1
(10) = 39(10) + 0.45(10)
2
310 = 355 i (t) = 355.
At 55 ms, it will take 10055 E 1.8 s to fall the last 100 m. Hence, the total time is 10 +
100
55
=
130
11
E 11.8 s.
73. o(t) = I, the initial velocity is 30 mih = 30
5280
3600
= 44 fts, and the nal velocity (after 5 seconds) is
50 mih = 50
5280
3600
=
220
3
fts. So (t) = It +C and (0) = 44 i C = 44. Thus, (t) = It + 44 i
(5) = 5I + 44. But (5) =
220
3
, so 5I + 44 =
220
3
i 5I =
88
3
i I =
88
15
E 5.87 fts
2
.
75. Let the acceleration be o(t) = I kmh
2
. We have (0) = 100 kmh and we can take the initial position c(0) to be 0.
We want the time t
]
for which (t) = 0 to satisfy c(t) < 0.08 km. In general,
0
(t) = o(t) = I, so (t) = It +C, where
C = (0) = 100. Now c
0
(t) = (t) = It + 100, so c(t) =
1
2
It
2
+ 100t +1, where 1 = c(0) = 0.
Thus, c(t) =
1
2
It
2
+ 100t. Since (t
]
) = 0, we have It
]
+ 100 = 0 or t
]
= 3100I, so
c(t
]
) =
1
2
I
3
100
I
2
+ 100
3
100
I
= 10,000
1
2I
3
1
I
= 3
5,000
I
. The condition c(t
]
) must satisfy is
3
5,000
I
< 0.08 i 3
5,000
0.08
I [I is negative] i I < 362,500 kmh
2
, or equivalently,
I < 3
3125
648
E 34.82 ms
2
.
214

77. (a) First note that 90 mih = 90
5280
3600
fts = 132 fts. Then o(t) = 4 fts
2
i (t) = 4t +C, but (0) = 0 i
C = 0. Now 4t = 132 when t =
132
4
= 33 s, so it takes 33 s to reach 132 fts. Therefore, taking c(0) = 0, we have
c(t) = 2t
2
, 0 $ t $ 33. So c(33) = 2178 ft. 15 minutes = 15(60) = 900 s, so for 33 < t $ 933 we have
(t) = 132 fts i c(933) = 132(900) + 2178 = 120,978 ft = 22.9125 mi.
(b) As in part (a), the train accelerates for 33 s and travels 2178 ft while doing so. Similarly, it decelerates for 33 s and travels
2178 ft at the end of its trip. During the remaining 900 366 = 834 s it travels at 132 fts, so the distance traveled is
132 834 = 110,088 ft. Thus, the total distance is 2178 + 110,088 + 2178 = 114,444 ft = 21.675 mi.
(c) 45 mi = 45(5280) = 237,600 ft. Subtract 2(2178) to take care of the speeding up and slowing down, and we have
233,244 ft at 132 fts for a trip of 233,244132 = 1767 s at 90 mih. The total time is
1767 + 2(33) = 1833 s = 30 min 33 s = 30.55 min.
(d) 37.5(60) = 2250 s. 2250 32(33) = 2184 s at maximum speed. 2184(132) + 2(2178) = 292,644 total feet or
292,6445280 = 55.425 mi.
4 Review
1. A function 1 has an absolute maximum at r = c if 1(c) is the largest function value on the entire domain of 1, whereas 1 has
a local maximum at c if 1(c) is the largest function value when r is near c. See Figure 4 in Section 4.1.
2. (a) See Theorem 4.1.3.
(b) See the Closed Interval Method before Example 8 in Section 4.1.
3. (a) See Theorem 4.1.4.
(b) See Denition 4.1.6.
4. (a) See Rolles Theorem at the beginning of Section 4.2.
(b) See the Mean Value Theorem in Section 4.2. Geometric interpretationthere is some point 1 on the graph of a function 1
[on the interval (o. /)] where the tangent line is parallel to the secant line that connects (o. 1(o)) and (/. 1(/)).
5. (a) See the I/D Test before Example 1 in Section 4.3.
(b) If the graph of 1 lies above all of its tangents on an interval 1, then it is called concave upward on 1.
(c) See the Concavity Test before Example 4 in Section 4.3.
(d) An inection point is a point where a curve changes its direction of concavity. They can be found by determining the points
at which the second derivative changes sign.
6. (a) See the First Derivative Test after Example 1 in Section 4.3.
(b) See the Second Derivative Test before Example 6 in Section 4.3.
(c) See the note before Example 7 in Section 4.3.
CHAPTER 4 REVIEW

215
7. (a) See lHospitals Rule and the three notes that follow it in Section 4.4.
(b) Write 1o as
1
1o
or
o
11
.
(c) Convert the difference into a quotient using a common denominator, rationalizing, factoring, or some other method.
(d) Convert the power to a product by taking the natural logarithm of both sides of j = 1
g
or by writing 1
g
as c
g ln ]
.
8. Without calculus you could get misleading graphs that fail to show the most interesting features of a function.
See the discussion at the beginning of Section 4.5 and the rst paragraph in Section 4.6.
9. (a) See Figure 3 in Section 4.8.
(b) r
2
= r
1
3
1(r
1
)
1
0
(r
1
)
(c) r
n+1
= r
n
3
1(r
n
)
1
0
(r
n
)
(d) Newtons method is likely to fail or to work very slowly when 1
0
(r
1
) is close to 0. It also fails when 1
0
(r
.
) is undened,
such as with 1(r) = 1r 32 and r
1
= 1.
10. (a) See the denition at the beginning of Section 4.9.
(b) If 1
1
and 1
2
are both antiderivatives of 1 on an interval 1, then they differ by a constant.
1. False. For example, take 1(r) = r
3
, then 1
0
(r) = 3r
2
and 1
0
(0) = 0, but 1(0) = 0 is not a maximum or minimum;
(0. 0) is an inection point.
3. False. For example, 1(r) = r is continuous on (0. 1) but attains neither a maximum nor a minimum value on (0. 1).
Dont confuse this with 1 being continuous on the closed interval [o. /], which would make the statement true.
5. True. This is an example of part (b) of the I/D Test.
7. False. 1
0
(r) = o
0
(r) i 1(r) = o(r) +C. For example, if 1(r) = r +2 and o(r) = r +1, then 1
0
(r) = o
0
(r) = 1,
but 1(r) 6= o(r).
9. True. The graph of one such function is sketched.
11. True. Let r
1
< r
2
where r
1
. r
2
M 1. Then 1(r
1
) < 1(r
2
) and o(r
1
) < o(r
2
) [since 1 and o are increasing on 1 ],
so (1 +o)(r
1
) = 1(r
1
) +o(r
1
) < 1(r
2
) +o(r
2
) = (1 +o)(r
2
).
13. False. Take 1(r) = r and o(r) = r 31. Then both 1 and o are increasing on (0. 1). But 1(r) o(r) = r(r 31) is not
increasing on (0. 1).
216

15. True. Let r
1
. r
2
M 1 and r
1
< r
2
. Then 1(r
1
) < 1(r
2
) [1 is increasing] i
1
1(r
1
)

1
1(r
2
)
[1 is positive] i
o(r
1
) o(r
2
) i o(r) = 11(r) is decreasing on 1.
17. True. If 1 is periodic, then there is a number j such that 1(r +j) = 1(j) for all r. Differentiating gives
1
0
(r) = 1
0
(r +j) (r +j)
0
= 1
0
(r +j) 1 = 1
0
(r +j), so 1
0
is periodic.
19. True. By the Mean Value Theorem, there exists a number c in (0. 1) such that 1(1) 31(0) = 1
0
(c)(1 30) = 1
0
(c).
Since 1
0
(c) is nonzero, 1(1) 31(0) 6= 0, so 1(1) 6= 1(0).
1. 1(r) = r
3
36r
2
+ 9r + 1, [2. 4]. 1
0
(r) = 3r
2
312r + 9 = 3(r
2
34r + 3) = 3(r 31)(r 33). 1
0
(r) = 0 i
r = 1 or r = 3, but 1 is not in the interval. 1
0
(r) 0 for 3 < r < 4 and 1
0
(r) < 0 for 2 < r < 3, so 1(3) = 1 is a local
minimum value. Checking the endpoints, we nd 1(2) = 3 and 1(4) = 5. Thus, 1(3) = 1 is the absolute minimum value and
1(4) = 5 is the absolute maximum value.
3. 1(r) =
3r 34
r
2
+ 1
, [32. 2]. 1
0
(r) =
(r
2
+ 1)(3) 3(3r 34)(2r)
(r
2
+ 1)
2
=
3(3r
2
38r 33)
(r
2
+ 1)
2
=
3(3r + 1)(r 33)
(r
2
+ 1)
2
.
1
0
(r) = 0 i r = 3
1
3
or r = 3, but 3 is not in the interval. 1
0
(r) 0 for 3
1
3
< r < 2 and 1
0
(r) < 0 for
32 < r < 3
1
3
, so 1
3
1
3
=
35
10'9
= 3
9
2
is a local minimum value. Checking the endpoints, we nd 1(32) = 32 and
1(2) =
2
5
. Thus, 1
3
1
3
= 3
9
2
is the absolute minimum value and 1(2) =
2
5
is the absolute maximum value.
5. 1(r) = r + sin2r, [0. ]. 1
0
(r) = 1 + 2 cos 2r = 0 C cos 2r = 3
1
2
C 2r =
2r
3
or
4r
3
C r =
r
3
or
2r
3
.
1
00
(r) = 34 sin2r, so 1
00
r
3
= 34 sin
2r
3
= 32
I
3 < 0 and 1
00
2r
3
= 34 sin
4r
3
= 2
I
3 0, so
1
r
3
=
r
3
+
I
3
2
E 1.91 is a local maximum value and 1
2r
3
=
2r
3
3
I
3
2
E 1.23 is a local minimum value. Also 1(0) = 0
and 1() = , so 1(0) = 0 is the absolute minimum value and 1() = is the absolute maximum value.
0
0
. lim
i<0
tanr
ln(1 +r)
H
= lim
i<0
sec
2
r
1(1 +r)
=
1
2
11
=
0
0
. lim
i<0
c
4i
31 34r
r
2
H
= lim
i<0
4c
4i
34
2r
H
= lim
i<0
16c
4i
2
= lim
i<0
8c
4i
= 8 1 = 8
11. This limit has the form" 0. lim
i<"
r
3
c
3i
= lim
i<"
r
3
c
i
H
= lim
i<"
3r
2
c
i
H
= lim
i<"
6r
c
i
H
= lim
i<"
6
c
i
= 0
CHAPTER 4 REVIEW

217
13. This limit has the form"3".
lim
i<1
+
r
r 31
3
1
lnr
= lim
i<1
+
rlnr 3r + 1
(r 31) lnr
H
= lim
i<1
+
r (1r) + lnr 31
(r 31) (1r) + lnr
= lim
i<1
+
lnr
1 31r + lnr
H
= lim
i<1
+
1r
1r
2
+ 1r
=
1
1 + 1
=
1
2
15. 1(0) = 0, 1
0
(32) = 1
0
(1) = 1
0
(9) = 0, lim
i<"
1(r) = 0, lim
i<6
1(r) = 3",
1
0
(r) < 0 on (3". 32), (1. 6), and (9. "), 1
0
(r) 0 on (32. 1) and (6. 9),
1
00
(r) 0 on (3". 0) and (12. "), 1
00
(r) < 0 on (0. 6) and (6. 12)
17. 1 is odd, 1
0
(r) < 0 for 0 < r < 2, 1
0
(r) 0 for r 2,
1
00
(r) 0 for 0 < r < 3, 1
00
(r) < 0 for r 3, lim
i<"
1(r) = 32
19. j = 1(r) = 2 32r 3r
3
A. 1 = R B. j-intercept: 1(0) = 2.
The r-intercept (approximately 0.770917) can be found using Newtons
Method. C. No symmetry D. No asymptote
E. 1
0
(r) = 32 33r
2
= 3(3r
2
+ 2) < 0, so 1 is decreasing on R.
F. No extreme value G. 1
00
(r) = 36r < 0 on (0. ") and 1
00
(r) 0 on
(3". 0), so 1 is CD on (0. ") and CU on (3". 0). There is an IP at (0. 2).
H.
21. j = 1(r) = r
4
33r
3
+ 3r
2
3r = r(r 31)
3
A. 1 = R B. j-intercept: 1(0) = 0; r-intercepts: 1(r) = 0 C
r = 0 or r = 1 C. No symmetry D. 1 is a polynomial function and hence, it has no asymptote.
E. 1
0
(r) = 4r
3
39r
2
+ 6r 31. Since the sum of the coefcients is 0, 1 is a root of 1
0
, so
1
0
(r) = (r 31)
4r
2
35r + 1
= (r 31)
2
(4r 31). 1
0
(r) < 0 i r <
1
4
, so 1 is decreasing on
3".
1
4
and 1 is increasing on
1
4
. "
. F. 1
0
(r) does not change sign at r = 1, so
there is not a local extremum there. 1
1
4
= 3
27
256
is a local minimum value.
G. 1
00
(r) = 12r
2
318r + 6 = 6(2r 31)(r 31). 1
00
(r) = 0 C r =
1
2
or 1. 1
00
(r) < 0 C
1
2
< r < 1 i 1 is CD on
1
2
. 1
and CU on
3".
1
2
and (1. "). There are inection points at
1
2
. 3
1
16
and (1. 0).

H.
218

23. j = 1(r) =
1
r(r 33)
2
A. 1 = {r | r 6= 0. 3} = (3". 0) (0. 3) (3. ") B. No intercepts. C. No symmetry.
D. lim
i<"
1
r(r 33)
2
i<0
+
1
r(r 33)
2
= ", lim
i<0
1
r(r 33)
2
= 3", lim
i<3
1
r(r 33)
2
= ",
so r = 0 and r = 3 are VA. E. 1
0
(r) = 3
(r 33)
2
+ 2r(r 33)
r
2
(r 33)
4
=
3(1 3r)
r
2
(r 33)
3
i 1
0
(r) 0 C 1 < r < 3,
so 1 is increasing on (1. 3) and decreasing on (3". 0), (0. 1), and (3. ").
F. Local minimum value 1(1) =
1
4
G. 1
00
(r) =
6(2r
2
34r + 3)
r
3
(r 33)
4
.
Note that 2r
2
34r + 3 0 for all r since it has negative discriminant.
So 1
00
(r) 0 C r 0 i 1 is CU on (0. 3) and (3. ") and
CD on (3". 0). No IP
H.
25. j = 1(r) =
r
2
r + 8
= r 38 +
64
r + 8
A. 1 = {r | r 6= 38} B. Intercepts are 0 C. No symmetry
D. lim
i<"
r
2
r + 8
= ", but 1(r) 3(r 38) =
64
r + 8
<0 as r <", so j = r 38 is a slant asymptote.
lim
i<38
+
r
2
r + 8
= "and lim
i<38
r
2
r + 8
= 3", so r = 38 is a VA. E. 1
0
(r) = 1 3
64
(r + 8)
2
=
r(r + 16)
(r + 8)
2
0 C
r 0 or r < 316, so 1 is increasing on (3". 316) and (0. ") and
decreasing on (316. 38) and (38. 0) .
F. Local maximum value 1(316) = 332, local minimum value 1(0) = 0
G. 1
00
(r) = 128(r + 8)
3
0 C r 38, so 1 is CU on (38. ") and
CD on (3". 38). No IP
H.
27. j = 1(r) = r
I
2 +r A. 1 = [32. ") B. j-intercept: 1(0) = 0; r-intercepts: 32 and 0 C. No symmetry
D. No asymptote E. 1
0
(r) =
r
2
I
2 +r
+
I
2 +r =
1
2
I
2 +r
[r + 2(2 +r)] =
3r + 4
2
I
2 +r
= 0 when r = 3
4
3
, so 1 is
decreasing on
32. 3
4
3
and increasing on
3
4
3
. "
. F. Local minimum value 1
3
4
3
= 3
4
3
2
3
= 3
4
I
6
9
E 31.09,
no local maximum
G. 1
00
(r) =
2
I
2 +r 3 3(3r + 4)
1
I
2 +r
4(2 +r)
=
6(2 +r) 3(3r + 4)
4(2 +r)
3'2
=
3r + 8
4(2 +r)
3'2
1
00
(r) 0 for r 32, so 1 is CU on (32. "). No IP
H.
CHAPTER 4 REVIEW

219
29. j = 1(r) = sin
2
r 32 cos r A. 1 = R B. j-intercept: 1(0) = 32 C. 1(3r) = 1(r), so 1 is symmetric with respect
to the j-axis. 1 has period 2. D. No asymptote E. j
0
= 2 sinrcos r + 2 sinr = 2 sinr(cos r + 1). j
0
= 0 C
sinr = 0 or cos r = 31 C r = n or r = (2n + 1). j
0
0 when sinr 0, since cos r + 1 D 0 for all r.
Therefore, j
0
0 [and so 1 is increasing] on (2n. (2n + 1)); j
0
< 0 [and so 1 is decreasing] on ((2n 31). 2n).
F. Local maximum values are 1((2n + 1)) = 2; local minimum values are 1(2n) = 32.
G. j
0
= sin2r + 2 sinr i j
00
= 2 cos 2r + 2 cos r = 2(2 cos
2
r 31) + 2 cos r = 4 cos
2
r + 2 cos r 32
= 2(2 cos
2
r + cos r 31) = 2(2 cos r 31)(cos r + 1)
j
00
= 0 C cos r =
1
2
or 31 C r = 2n
r
3
or r = (2n + 1).
j
00
0 [and so 1 is CU] on
2n 3
r
3
. 2n +
r
3
; j
00
$ 0 [and so 1 is CD]
on
2n +
r
3
. 2n +
5r
3
. There are inection points at
2n
r
3
. 3
1
4
.
H.
31. j = 1(r) = sin
31
(1r) A. 1 = {r | 31 $ 1r $ 1} = (3". 31] [1. ") . B. No intercept
C. 1(3r) = 31(r), symmetric about the origin D. lim
i<"
sin
31
(1r) = sin
31
(0) = 0, so j = 0 is a HA.
E. 1
0
(r) =
1
1 3(1r)
2
3
1
r
2
=
31
I
r
4
3r
2
< 0, so 1 is decreasing on (3". 31) and (1. ") .
F. No local extreme value, but 1(1) =
r
2
is the absolute maximum value
and 1(31) = 3
r
2
is the absolute minimum value.
G. 1
00
(r) =
4r
3
32r
2(r
4
3r
2
)
3'2
=
r
2r
2
31
(r
4
3r
2
)
3'2
0 for r 1 and 1
00
(r) < 0
for r < 31, so 1 is CU on (1. ") and CD on (3". 31). No IP
H.
33. j = 1(r) = rc
32i
A. 1 = R B. j-intercept: 1(0) = 0; r-intercept: 1(r) = 0 C r = 0
i<"
rc
32i
= lim
i<"
r
c
2i
H
= lim
i<"
1
2c
2i
= 0, so j = 0 is a HA.
E. 1
0
(r) = r(32c
32i
) +c
32i
(1) = c
32i
(32r + 1) 0 C 32r + 1 0 C r <
1
2
and 1
0
(r) < 0 C r
1
2
,
so 1 is increasing on
3".
1
2
and decreasing on
1
2
. "
. F. Local maximum value 1
1
2
=
1
2
c
31
= 1(2c);
no local minimum value
G. 1
00
(r) = c
32i
(32) + (32r + 1)(32c
32i
)
= 2c
32i
[31 3(32r + 1)] = 4 (r 31) c
32i
.
1
00
(r) 0 C r 1 and 1
00
(r) < 0 C r < 1, so 1 is
CU on (1. ") and CD on (3". 1). IP at (1. 1(1)) = (1. c
32
)
H.
220

35. 1(r) =
r
2
31
r
3
i 1
0
(r) =
r
3
(2r) 3
r
2
31
3r
2
r
6
=
3 3r
2
r
4
i
1
00
(r) =
r
4
(32r) 3
3 3r
2
4r
3
r
8
=
2r
2
312
r
5
Estimates: From the graphs of 1
0
and 1
00
, it appears that 1 is increasing on
(31.73. 0) and (0. 1.73) and decreasing on (3". 31.73) and (1.73. ");
1 has a local maximum of about 1(1.73) = 0.38 and a local minimum of about
1(31.7) = 30.38; 1 is CU on (32.45. 0) and (2.45. "), and CD on
(3". 32.45) and (0. 2.45); and 1 has inection points at about
(32.45. 30.34) and (2.45. 0.34).
Exact: Now 1
0
(r) =
3 3r
2
r
4
is positive for 0 < r
2
< 3, that is, 1 is increasing
on
3
I
3. 0
and
0.
I
3
; and 1
0
(r) is negative (and so 1 is decreasing) on
3". 3
I
3
and
I
3. "
. 1
0
(r) = 0 when r =
I
3.
1
0
goes from positive to negative at r =
I
3, so 1 has a local maximum of
1
I
3
=
(
I
3 )
2
31
(
I
3 )
3
=
2
I
3
9
; and since 1 is odd, we know that maxima on the
interval (0. ") correspond to minima on (3". 0), so 1 has a local minimum of
1
3
I
3
= 3
2
I
3
9
. Also, 1
00
(r) =
2r
2
312
r
5
is positive (so 1 is CU) on
3
I
6. 0
and
I
6. "
, and negative (so 1 is CD) on
3". 3
I
6
and
0.
I
6
. There are IP at
I
6.
5
I
6
36
and
3
I
6. 3
5
I
6
36
.
37. 1(r) = 3r
6
35r
5
+r
4
35r
3
32r
2
+ 2 i 1
0
(r) = 18r
5
325r
4
+ 4r
3
315r
2
34r i
1
00
(r) = 90r
4
3100r
3
+ 12r
2
330r 34
0
and 1
00
, it appears that 1 is increasing on (30.23. 0) and (1.62. ") and decreasing on (3". 30.23)
and (0. 1.62); 1 has a local maximum of about 1(0) = 2 and local minima of about 1(30.23) = 1.96 and 1(1.62) = 319.2;
CHAPTER 4 REVIEW

221
1 is CU on (3". 30.12) and (1.24. ") and CD on (30.12. 1.24); and 1 has inection points at about (30.12. 1.98) and
(1.24. 312.1).
39. From the graph, we estimate the points of inection to be about (0.82. 0.22).
1(r) = c
31'i
2
i 1
0
(r) = 2r
33
c
31'i
2
i
1
00
(r) = 2[r
33
(2r
33
)c
31'i
2
+c
31'i
2
(33r
34
)] = 2r
36
c
31'i
2
2 33r
2
.
This is 0 when 2 33r
2
= 0 C r =
2
3
, so the inection points
are
2
3
. c
33'2
.
41. 1(r) =
cos
2
r
I
r
2
+r + 1
, 3 $ r $ i 1
0
(r) = 3
cos r[(2r + 1) cos r + 4(r
2
+r + 1) sinr]
2(r
2
+r + 1)
3'2
i
1
00
(r) = 3
(8r
4
+ 16r
3
+ 16r
2
+ 8r + 9) cos
2
r 38(r
2
+r + 1)(2r + 1) sinr cos r 38(r
2
+r + 1)
2
sin
2
r
4(r
2
+r + 1)
5'2
1(r) = 0 C r =
r
2
; 1
0
(r) = 0 C r E 32.96, 31.57, 30.18, 1.57, 3.01;
1
00
(r) = 0 C r E 32.16, 30.75, 0.46, and 2.21.
The r-coordinates of the maximum points are the values at which 1
0
changes from positive to negative, that is, 32.96, 30.18,
and 3.01. The r-coordinates of the minimum points are the values at which 1
0
changes from negative to positive, that is,
31.57 and 1.57. The r-coordinates of the inection points are the values at which 1
00
changes sign, that is, 32.16, 30.75,
0.46, and 2.21.
43. The family of functions 1(r) = ln(sinr +C) all have the same period and all
have maximum values at r =
r
2
+ 2n. Since the domain of ln is (0. "), 1 has
a graph only if sinr +C 0 somewhere. Since 31 $ sinr $ 1, this happens
if C 31, that is, 1 has no graph if C $ 31. Similarly, if C 1, then
sinr +C 0 and 1 is continuous on (3". "). As C increases, the graph of
1 is shifted vertically upward and attens out. If 31 < C $ 1, 1 is dened
222

where sinr +C 0 C sinr 3C C sin
31
(3C) < r < 3sin
31
(3C). Since the period is 2, the domain of 1
is
2n + sin
31
(3C). (2n + 1) 3sin
31
(3C)
, n an integer.
45. Let 1(r) = 3r +2 cos r + 5. Then 1(0) = 7 0 and 1(3) = 33 32 +5 = 33 +3 = 33( 31) < 0. and since 1 is
continuous on R (hence on [3. 0]), the Intermediate Value Theorem assures us that there is at least one zero of 1 in [3. 0].
Now 1
0
(r) = 3 32 sinr 0 implies that 1 is increasing on R, so there is exactly one zero of 1, and hence, exactly one real
root of the equation 3r + 2 cos r + 5 = 0.
47. Since 1 is continuous on [32. 33] and differentiable on (32. 33), then by the Mean Value Theorem there exists a number c in
(32. 33) such that 1
0
(c) =
1
5
c
34'5
=
5
I
33 3
5
I
32
33 332
=
5
I
33 32, but
1
5
c
34'5
0 i
5
I
33 32 0 i
5
I
33 2. Also
1
0
is decreasing, so that 1
0
(c) < 1
0
(32) =
1
5
(32)
34'5
= 0.0125 i 0.0125 1
0
(c) =
5
I
33 32 i
5
I
33 < 2.0125.
Therefore, 2 <
5
I
33 < 2.0125.
49. (a) o(r) = 1(r
2
) i o
0
(r) = 2r1
0
(r
2
) by the Chain Rule. Since 1
0
(r) 0 for all r 6= 0, we must have 1
0
(r
2
) 0 for
r 6= 0, so o
0
(r) = 0 C r = 0. Now o
0
(r) changes sign (from negative to positive) at r = 0, since one of its factors,
1
0
(r
2
), is positive for all r, and its other factor, 2r, changes from negative to positive at this point, so by the First
Derivative Test, 1 has a local and absolute minimum at r = 0.
(b) o
0
(r) = 2r1
0
(r
2
) i o
00
(r) = 2[r1
00
(r
2
)(2r) +1
0
(r
2
)] = 4r
2
1
00
(r
2
) +21
0
(r
2
) by the Product Rule and the Chain
Rule. But r
2
0 for all r 6= 0, 1
00
(r
2
) 0 [since 1 is CU for r 0], and 1
0
(r
2
) 0 for all r 6= 0, so since all of its
factors are positive, o
00
(r) 0 for r 6= 0. Whether o
00
(0) is positive or 0 doesnt matter [since the sign of o
00
does not
change there]; o is concave upward on R.
51. If 1 = 0, the line is vertical and the distance from r = 3
C
to (r
1
. j
1
) is
r
1
+
C
=
|r
1
+1j
1
+C|
I
2
+1
2
, so assume
1 6= 0. The square of the distance from (r
1
. j
1
) to the line is 1(r) = (r 3r
1
)
2
+ (j 3j
1
)
2
where r +1j +C = 0, so
we minimize 1(r) = (r 3r
1
)
2
+
1
r 3
C
1
3j
1
2
i 1
0
(r) = 2 (r 3r
1
) + 2
1
r 3
C
1
3j
1
.
1
0
(r) = 0 i r =
1
2
r
1
31j
1
3C
2
+1
2
and this gives a minimum since 1
00
(r) = 2
1 +

2
1
2
0. Substituting
this value of r into 1(r) and simplifying gives 1(r) =
(r
1
+1j
1
+C)
2
2
+1
2
, so the minimum distance is
1(r) =
|r
1
+1j
1
+C|
I
2
+1
2
.
CHAPTER 4 REVIEW

223
j
r
=
v
I
r
2
32vr
, so the area of the triangle is
(r) =
1
2
(2j)r = rj =
vr
2
I
r
2
32vr
i
0
(r) =
2vr
I
r
2
32vr 3vr
2
(r 3v)
I
r
2
32vr
r
2
32vr
=
vr
2
(r 33v)
(r
2
32vr)
3'2
= 0
when r = 3v.
0
(r) < 0 when 2v < r < 3v,
0
(r) 0 when r 3v. So r = 3v gives a minimum and (3v) =
v(9v
2
)
I
3 v
= 3
I
3 v
2
.
55. We minimize 1(r) = |1| + |11| + |1C| = 2
I
r
2
+ 16 + (5 3r),
0 $ r $ 5. 1
0
(r) = 2r
I
r
2
+ 16 31 = 0 C 2r =
I
r
2
+ 16 C
4r
2
= r
2
+ 16 C r =
4
I
3
. 1(0) = 13, 1
4
I
3
E 11.9, 1(5) E 12.8, so the

minimum occurs when r =
4
I
3
E 2.3.
57. = 1
1
C
+
C
1
i
d
d1
=
1
2
(1C) + (C1)
1
C
3
C
1
2
= 0 C
1
C
=
C
1
2
C 1
2
= C
2
C 1 = C.
This gives the minimum velocity since
0
< 0 for 0 < 1 < C and
0
0 for 1 C.
59. Let r denote the number of $1 decreases in ticket price. Then the ticket price is $12 3$1(r), and the average attendance is
11,000 + 1000(r). Now the revenue per game is
1(r) = (price per person) (number of people per game)
= (12 3r)(11,000 + 1000r) = 31000r
2
+ 1000r + 132,000
for 0 $ r $ 4 [since the seating capacity is 15,000] i 1
0
(r) = 32000r + 1000 = 0 C r = 0.5. This is a
maximum since 1
00
(r) = 32000 < 0 for all r. Now we must check the value of 1(r) = (12 3r)(11,000 + 1000r) at
r = 0.5 and at the endpoints of the domain to see which value of r gives the maximum value of 1.
1(0) = (12)(11,000) = 132,000, 1(0.5) = (11.5)(11,500) = 132,250, and 1(4) = (8)(15,000) = 120,000. Thus, the
maximum revenue of $132,250 per game occurs when the average attendance is 11,500 and the ticket price is $11.50.
61. 1(r) = r
5
3r
4
+ 3r
2
33r 32 i 1
0
(r) = 5r
4
34r
3
+ 6r 33, so r
n+1
= r
n
3
r
5
n
3r
4
n
+ 3r
2
n
33r
n
32
5r
4
n
34r
3
n
+ 6r
n
33
.
Now r
1
= 1 i r
2
= 1.5 i r
3
E 1.343860 i r
4
E 1.300320 i r
5
E 1.297396 i
r
6
E 1.297383 E r
7
, so the root in [1. 2] is 1.297383, to six decimal places.
224

63. 1(t) = cos t +t 3t
2
i 1
0
(t) = 3sint + 1 32t. 1
0
(t) exists for all
t, so to nd the maximum of 1, we can examine the zeros of 1
0
.
From the graph of 1
0
, we see that a good choice for t
1
is t
1
= 0.3.
Use o(t) = 3sint + 1 32t and o
0
(t) = 3cos t 32 to obtain
t
2
E 0.33535293, t
3
E 0.33541803 E t
4
. Since 1
00
(t) = 3cos t 32 < 0
for all t, 1(0.33541803) E 1.16718557 is the absolute maximum.
65. 1
0
(r) = cos r 3(1 3r
2
)
31'2
= cos r 3
1
I
1 3r
2
i 1(r) = sinr 3sin
31
r +C
67. 1
0
(r) =
I
r
3
+
3
I
r
2
= r
3'2
+r
2'3
i 1(r) =
r
5'2
52
+
r
5'3
53
+C =
2
5
r
5'2
+
3
5
r
5'3
+C
69. 1
0
(t) = 2t 33 sint i 1(t) = t
2
+ 3 cos t +C.
1(0) = 3 +C and 1(0) = 5 i C = 2, so 1(t) = t
2
+ 3 cos t + 2.
71. 1
00
(r) = 1 36r + 48r
2
i 1
0
(r) = r 33r
2
+ 16r
3
+C. 1
0
(0) = C and 1
0
(0) = 2 i C = 2, so
1
0
(r) = r 33r
2
+ 16r
3
+ 2 and hence, 1(r) =
1
2
r
2
3r
3
+ 4r
4
+ 2r +1.
1(0) = 1 and 1(0) = 1 i 1 = 1, so 1(r) =
1
2
r
2
3r
3
+ 4r
4
+ 2r + 1.
73. (t) = c
0
(t) = 2t 3
1
1 +t
2
i c(t) = t
2
3tan
31
t +C.
c(0) = 0 30 +C = C and c(0) = 1 i C = 1, so c(t) = t
2
3tan
31
t + 1.
75. (a) Since 1 is 0 just to the left of the j-axis, we must have a minimum of 1 at the same place since we are increasing through
(0. 0) on 1. There must be a local maximum to the left of r = 33, since 1 changes from positive to negative there.
(b) 1(r) = 0.1c
i
+ sinr i
1(r) = 0.1c
i
3cos r +C. 1(0) = 0 i
0.1 31 +C = 0 i C = 0.9, so
1(r) = 0.1c
i
3cos r + 0.9.
(c)
77. Choosing the positive direction to be upward, we have o(t) = 39.8 i (t) = 39.8t +
0
, but (0) = 0 =
0
i
(t) = 39.8t = c
0
(t) i c(t) = 34.9t
2
+c
0
, but c(0) = c
0
= 500 i c(t) = 34.9t
2
+ 500. When c = 0,
34.9t
2
+ 500 = 0 i t
1
=

500
4.9
E 10.1 i (t
1
) = 39.8
500
4.9
E 398.995 ms. Since the canister has been
designed to withstand an impact velocity of 100 ms, the canister will not burst.
CHAPTER 4 REVIEW

225
79. (a) The cross-sectional area of the rectangular beam is
= 2r 2j = 4rj = 4r
I
100 3r
2
, 0 $ r $ 10, so
d
dr
= 4r
1
2
(100 3r
2
)
31'2
(32r) + (100 3r
2
)
1'2
4
=
34r
2
(100 3r
2
)
1'2
+ 4(100 3r
2
)
1'2
=
4[3r
2
+
100 3r
2
]
(100 3r
2
)
1'2
.
d
dr
= 0 when 3r
2
+
100 3r
2
= 0 i r
2
= 50 i r =
I
50 E 7.07 i j =

100 3
I
50
2
=
I
50.
Since (0) = (10) = 0, the rectangle of maximum area is a square.
(b) The cross-sectional area of each rectangular plank (shaded in the gure) is
= 2r
j 3
I
50
= 2r
I
100 3r
2
3
I
50
, 0 $ r $
I
50, so
d
dr
= 2
I
100 3r
2
3
I
50
+ 2r
1
2
(100 3r
2
)
31'2
(32r)
= 2(100 3r
2
)
1'2
32
I
50 3
2r
2
(100 3r
2
)
1'2
Set
d
dr
= 0: (100 3r
2
) 3
I
50 (100 3r
2
)
1'2
3r
2
= 0 i 100 32r
2
=
I
50 (100 3r
2
)
1'2
i
10,000 3400r
2
+ 4r
4
= 50(100 3r
2
) i 4r
4
3350r
2
+ 5000 = 0 i 2r
4
3175r
2
+ 2500 = 0 i
r
2
=
175
I
10,625
4
E 69.52 or 17.98 i r E 8.34 or 4.24. But 8.34
I
50, so r
1
E 4.24 i
j 3
I
50 =

100 3r
2
1
3
I
50 E 1.99. Each plank should have dimensions about 8
1
2
inches by 2 inches.
(c) From the gure in part (a), the width is 2r and the depth is 2j, so the strength is
o = I(2r)(2j)
2
= 8Irj
2
= 8Ir(100 3r
2
) = 800Ir 38Ir
3
, 0 $ r $ 10. dodr = 800I 324Ir
2
= 0 when
24Ir
2
= 800I i r
2
=
100
3
i r =
10
I
3
i j =

200
3
=
10
I
2
I
3
=
I
2 r. Since o(0) = o(10) = 0, the
maximum strength occurs when r =
10
I
3
. The dimensions should be
20
I
3
E 11.55 inches by
20
I
2
I
3
E 16.33 inches.
81. We rst show that
r
1 +r
2
< tan
31
r for r 0. Let 1(r) = tan
31
r 3
r
1 +r
2
. Then
1
0
(r) =
1
1 +r
2
3
1(1 +r
2
) 3r(2r)
(1 +r
2
)
2
=
(1 +r
2
) 3(1 3r
2
)
(1 +r
2
)
2
=
2r
2
(1 +r
2
)
2
0 for r 0. So 1(r) is increasing
on (0. "). Hence, 0 < r i 0 = 1(0) < 1(r) = tan
31
r 3
r
1 +r
2
. So
r
1 +r
2
< tan
31
r for 0 < r. We next show
that tan
31
r < r for r 0. Let /(r) = r 3tan
31
r. Then /
0
(r) = 1 3
1
1 +r
2
=
r
2
1 +r
2
0. Hence, /(r) is increasing
on (0. "). So for 0 < r, 0 = /(0) < /(r) = r 3tan
31
r. Hence, tan
31
r < r for r 0, and we conclude that
r
1 +r
2
< tan
31
r < r for r 0.
PROBLEMS PLUS
1. Let j = 1(r) = c
3i
2
. The area of the rectangle under the curve from3r to r is (r) = 2rc
3i
2
where r D 0. We maximize
(r):
0
(r) = 2c
3i
2
34r
2
c
3i
2
= 2c
3i
2
1 32r
2
= 0 i r =
1
I
2
. This gives a maximum since
0
(r) 0
for 0 $ r <
1
I
2
and
0
(r) < 0 for r
1
I
2
. We next determine the points of inection of 1(r). Notice that
1
0
(r) = 32rc
3i
2
= 3(r). So 1
00
(r) = 3
0
(r) and hence, 1
00
(r) < 0 for 3
1
I
2
< r <
1
I
2
and 1
00
(r) 0 for r < 3
1
I
2
and r
1
I
2
. So 1(r) changes concavity at r =
1
I
2
, and the two vertices of the rectangle of largest area are at the inection
points.
3. First, we recognize some symmetry in the inequality:
c
i+u
rj
D c
2
C
c
i
r

c
u
j
D c c. This suggests that we need to show
that
c
i
r
D c for r 0. If we can do this, then the inequality
c
u
j
D c is true, and the given inequality follows. 1(r) =
c
i
r
i
1
0
(r) =
rc
i
3c
i
r
2
=
c
i
(r 31)
r
2
= 0 i r = 1. By the First Derivative Test, we have a minimum of 1(1) = c, so
c
i
r D c for all r.
5. Let 1 = lim
i<0
or
2
+ sin/r + sincr + sindr
3r
2
+ 5r
4
+ 7r
6
. Now 1 has the indeterminate form of type
0
0
, so we can apply lHospitals
Rule. 1 = lim
i<0
2or +/ cos /r +c cos cr +d cos dr
6r + 20r
3
+ 42r
5
. The denominator approaches 0 as r <0, so the numerator must also
approach 0 (because the limit exists). But the numerator approaches 0 +/ +c +d, so / +c +d = 0. Apply lHospitals Rule
again. 1 = lim
i<0
2o 3/
2
sin/r 3c
2
sincr 3d
2
sindr
6 + 60r
2
+ 210r
4
=
2o 30
6 + 0
=
2o
6
, which must equal 8.
2o
6
= 8 i o = 24.
Thus, o +/ +c +d = o + (/ +c +d) = 24 + 0 = 24.
7. Differentiating r
2
+rj +j
2
= 12 implicitly with respect to r gives 2r +j +r
dj
dr
+ 2j
dj
dr
= 0, so
dj
dr
= 3
2r +j
r + 2j
.
At a highest or lowest point,
dj
dr
= 0 C j = 32r. Substituting 32r for j in the original equation gives
r
2
+r(32r) + (32r)
2
= 12, so 3r
2
= 12 and r = 2. If r = 2, then j = 32r = 34, and if r = 32 then j = 4.
Thus, the highest and lowest points are (32. 4) and (2. 34).
9. j = r
2
i j
0
= 2r, so the slope of the tangent line at 1(o. o
2
) is 2o and the slope of the normal line is 3
1
2o
for o 6= 0.
An equation of the normal line is j 3o
2
= 3
1
2o
(r 3o). Substitute r
2
for j to nd the r-coordinates of the two points of
intersection of the parabola and the normal line. r
2
3o
2
= 3
r
2o
+
1
2
i 2or
2
+r 32o
3
3o = 0 i
227
228

r =
31
1 34(2o)(32o
3
3o)
2(2o)
=
31
I
1 + 16o
4
+ 8o
2
4o
=
31
(4o
2
+ 1)
2
4o
=
31 (4o
2
+ 1)
4o
=
4o
2
4o
or
34o
2
32
4o
, or equivalently, o or 3o 3
1
2o
So the point Q has coordinates
3o 3
1
2o
.
3o 3
1
2o
. The square o of the distance from 1 to Qis given by

o =

3o 3
1
2o
3o
2
+
3o 3
1
2o
2
3o
2
2
=

32o 3
1
2o
2
+
o
2
+ 1 +
1
4o
2
3o
2
2
=

4o
2
+ 2 +
1
4o
2
1 +
1
4o
2
2
=

4o
2
+ 2 +
1
4o
2
+ 1 +
2
4o
2
+
1
16o
4
= 4o
2
+ 3 +
3
4o
2
+
1
16o
4
o
0
= 8o 3
6
4o
3
3
4
16o
5
= 8o 3
3
2o
3
3
1
4o
5
=
32o
6
36o
2
31
4o
5
. The only real positive zero of the equation o
0
= 0 is
o =
1
I
2
. Since o
00
= 8 +
9
2o
4
+
5
4o
6
0, o =
1
I
2
corresponds to the shortest possible length of the line segment 1Q.
11. 1(r) =
o
2
+o 36
cos 2r + (o 32)r + cos 1 i 1

0
(r) = 3
o
2
+o 36
sin2r(2) + (o 32). The derivative exists

for all r, so the only possible critical points will occur where 1
0
(r) = 0 C 2(o 32)(o + 3) sin2r = o 32 C
either o = 2 or 2(o + 3) sin2r = 1, with the latter implying that sin2r =
1
2(o + 3)
. Since the range of sin2r is [31. 1],
this equation has no solution whenever either
1
2(o + 3)
< 31 or
1
2(o + 3)
1. Solving these inequalities, we get
3
7
2
< o < 3
5
2
.
13. =

r
1
. r
2
1
and 1 =

r
2
. r
2
2
, where r
1
and r
2
are the solutions of the quadratic equation r
2
= ir +/. Let 1 =

r. r
2
and set
1
= (r
1
. 0), 1
1
= (r
2
. 0), and 1
1
= (r. 0). Let 1(r) denote the area of triangle 11. Then 1(r) can be expressed
in terms of the areas of three trapezoids as follows:
1(r) = area (
1
11
1
) 3area (
1
11
1
) 3area (1
1
111
1
)
=
1
2
r
2
1
+r
2
2
(r
2
3r
1
) 3
1
2
r
2
1
+r
2
(r 3r
1
) 3
1
2
r
2
+r
2
2
(r
2
3r)
After expanding and canceling terms, we get
1(r) =
1
2
r
2
r
2
1
3r
1
r
2
2
3rr
2
1
+r
1
r
2
3r
2
r
2
+rr
2
2
=
1
2
r
2
1
(r
2
3r) +r
2
2
(r 3r
1
) +r
2
(r
1
3r
2
)
1
0
(r) =
1
2
3r
2
1
+r
2
2
+ 2r(r
1
3r
2
)
. 1
00
(r) =
1
2
[2(r
1
3r
2
)] = r
1
3r
2
< 0 since r
2
r
1
.
1
0
(r) = 0 i 2r(r
1
3r
2
) = r
2
1
3r
2
2
i r
T
=
1
2
(r
1
+r
2
).
1(r
T
) =
1
2
r
2
1
1
2
(r
2
3r
1
)
+r
2
2
1
2
(r
2
3r
1
)
+
1
4
(r
1
+r
2
)
2
(r
1
3r
2
)
=
1
2
1
2
(r
2
3r
1
)
r
2
1
+r
2
2
3
1
4
(r
2
3r
1
)(r
1
+r
2
)
2
=
1
8
(r
2
3r
1
)
r
2
1
+r
2
2
r
2
1
+ 2r
1
r
2
+r
2
2
=
1
8
(r
2
3r
1
)
r
2
1
32r
1
r
2
+r
2
2
=
1
8
(r
2
3r
1
)(r
1
3r
2
)
2
=
1
8
(r
2
3r
1
)(r
2
3r
1
)
2
=
1
8
(r
2
3r
1
)
3

229
To put this in terms of i and /, we solve the system j = r
2
1
and j = ir
1
+/, giving us r
2
1
3ir
1
3/ = 0 i
r
1
=
1
2
i3
I
i
2
+ 4/
. Similarly, r
2
=
1
2
i+
I
i
2
+ 4/
. The area is then

1
8
(r
2
3r
1
)
3
=
1
8
I
i
2
+ 4/
3
,
and is attained at the point 1
r
T
. r
2
T
= 1
1
2
i.
1
4
i
2
.
Note: Another way to get an expression for 1(r) is to use the formula for an area of a triangle in terms of the coordinates of
the vertices: 1(r) =
1
2
r
2
r
2
1
3r
1
r
2
2
r
1
r
2
3rr
2
1
rr
2
2
3r
2
r
2
.
15. Suppose that the curve j = o
i
intersects the line j = r. Then o
i
0
= r
0
for some r
0
0, and hence o = r
1'i
0
0
. We nd the
maximum value of o(r) = r
1'i
, 0, because if o is larger than the maximum value of this function, then the curve j = o
i
does not intersect the line j = r. o
0
(r) = c
(1'i) ln i
3
1
r
2
lnr +
1
r

1
r
= r
1'i
1
r
2
(1 3lnr). This is 0 only where

r = c, and for 0 < r < c, 1
0
(r) 0, while for r c, 1
0
(r) < 0, so o has an absolute maximum of o(c) = c
1'c
. So if
j = o
i
intersects j = r, we must have 0 < o $ c
1'c
. Conversely, suppose that 0 < o $ c
1'c
. Then o
c
$ c, so the graph of
j = o
i
lies below or touches the graph of j = r at r = c. Also o
0
= 1 0, so the graph of j = o
i
lies above that of j = r
at r = 0. Therefore, by the Intermediate Value Theorem, the graphs of j = o
i
and j = r must intersect somewhere between
r = 0 and r = c.
17. Note that 1(0) = 0, so for r 6= 0,
1(r) 31(0)
r 30
1(r)
r
=
|1(r)|
|r|
$
| sinr|
|r|
=
sinr
r
.
Therefore, |1
0
(0)| =
lim
i<0
1(r) 31(0)
r 30
= lim
i<0
1(r) 31(0)
r 30
$ lim
i<0
sinr
r
= 1.
But 1(r) = o
1
sinr +o
2
sin2r + +o
n
sinnr i 1
0
(r) = o
1
cos r + 2o
2
cos 2r + +no
n
cos nr, so
|1
0
(0)| = |o
1
+ 2o
2
+ +no
n
| $ 1.
Another solution: We are given that
n
I=1
o
I
sinIr
$ | sinr|. So for r close to 0, and r 6= 0, we have
I=1
o
I
sinIr
sinr
$ 1 i lim
i<0
I=1
o
I
sinIr
sinr
$ 1 i
I=1
o
I
lim
i<0
sinIr
sinr
$ 1. But by lHospitals Rule,

lim
i<0
sinIr
sinr
= lim
i<0
I cos Ir
cos r
= I, so
I=1
Io
I
$ 1.
19. (a) Distance = rate time, so time = distancerate. T
1
=
1
c
1
, T
2
=
2 |11|
c
1
+
|1o|
c
2
=
2/sec 0
c
1
+
132/tan0
c
2
,
T
3
=
2
/
2
+ 1
2
/4
c
1
=
I
4/
2
+1
2
c
1
.
(b)
dT
2
d0
=
2/
c
1
sec 0 tan0 3
2/
c
2
sec
2
0 = 0 when 2/sec 0
1
c
1
tan0 3
1
c
2
sec 0
= 0 i
1
c
1
sin0
cos 0
3
1
c
2
1
cos 0
= 0 i
sin0
c
1
cos 0
=
1
c
2
cos 0
i sin0 =
c
1
c
2
. The First Derivative Test shows that this gives
a minimum.
230

(c) Using part (a) with 1 = 1 and T
1
= 0.26, we have T
1
=
1
c
1
i c
1
=
1
0.26
E 3.85 kms. T
3
=
I
4/
2
+1
2
c
1
i
4/
2
+1
2
= T
2
3
c
2
1
i / =
1
2
T
2
3
c
2
1
31
2
=
1
2
(0.34)
2
(10.26)
2
31
2
E 0.42 km. To nd c
2
, we use sin0 =
c
1
c
2
from part (b) and T
2
=
2/sec 0
c
1
+
132/tan0
c
2
from part (a). From the gure,
sin0 =
c
1
c
2
i sec 0 =
c
2
c
2
2
3c
2
1
and tan0 =
c
1
c
2
2
3c
2
1
, so
T
2
=
2/c
2
c
1
c
2
2
3c
2
1
+
1
c
2
2
3c
2
1
32/c
1
c
2
c
2
2
3c
2
1
. Using the values for T
2
[given as 0.32],
/, c
1
, and 1. we can graph Y
1
= T
2
and Y
2
=
2/c
2
c
1
c
2
2
3c
2
1
+
1
c
2
2
3c
2
1
32/c
1
c
2
c
2
2
3c
2
1
and nd their intersection points.
Doing so gives us c
2
E 4.10 and 7.66, but if c
2
= 4.10, then 0 = arcsin(c
1
c
2
) E 69.6
, which implies that point o is to

the left of point 1 in the diagram. So c
2
= 7.66 kms.
21. Let o = |11| and / = |11| as shown in the gure. Since / = |11| +|11|,
|11| = / 3/. Now
|11| = |11| +|11| = o +/ 3/ =
I
v
2
3r
2
+/ 3
(d 3r)
2
+o
2
=
I
v
2
3r
2
+/ 3
(d 3r)
2
+
I
v
2
3r
2
2
=
I
v
2
3r
2
+/ 3
I
d
2
32dr +r
2
+v
2
3r
2
Let 1(r) =
I
v
2
3r
2
+/ 3
I
d
2
+v
2
32dr.
1
0
(r) =
1
2
(v
2
3r
2
)
31'2
(32r) 3
1
2
(d
2
+v
2
32dr)
31'2
(32d) =
3r
I
v
2
3r
2
+
d
I
d
2
+v
2
32dr
.
1
0
(r) = 0 i
r
I
v
2
3r
2
=
d
I
d
2
+v
2
32dr
i
r
2
v
2
3r
2
=
d
2
d
2
+v
2
32dr
i
d
2
r
2
+v
2
r
2
32dr
3
= d
2
v
2
3d
2
r
2
i 0 = 2dr
3
32d
2
r
2
3v
2
r
2
+d
2
v
2
i
0 = 2dr
2
(r 3d) 3v
2
(r
2
3d
2
) i 0 = 2dr
2
(r 3d) 3v
2
(r +d)(r 3d) i 0 = (r 3d)[2dr
2
3v
2
(r +d)]
But d v r, so r 6= d. Thus, we solve 2dr
2
3 v
2
r 3 dv
2
= 0 for r:
r =
3(3v
2
)
(3v
2
)
2
34(2d)(3dv
2
)
2(2d)
=
v
2
I
v
4
+ 8d
2
v
2
4d
. Because
I
v
4
+ 8d
2
v
2
v
2
, the negative can be
discarded. Thus, r =
v
2
+
I
v
2
I
v
2
+ 8d
2
4d
=
v
2
+v
I
v
2
+ 8d
2
4d
[v 0] =
v
4d
v +
I
v
2
+ 8d
2
. The maximum value

of |11| occurs at this value of r.

231
23. \ =
4
3
v
3
i
d\
dt
= 4v
2
dv
dt
. But
d\
dt
is proportional to the surface area, so
d\
dt
= I 4v
2
for some constant I.
Therefore, 4v
2
dv
dt
= I 4v
2
C
dv
dt
= I = constant. An antiderivative of I with respect to t is It, so v = It +C.
When t = 0, the radius v must equal the original radius v
0
, so C = v
0
, and v = It +v
0
. To nd I we use the fact that when
t = 3, v = 3I +v
0
and \ =
1
2
\
0
i
4
3
(3I +v
0
)
3
=
1
2

4
3
v
3
0
i (3I +v
0
)
3
=
1
2
v
3
0
i 3I +v
0
=
1
3
I
2
v
0
i
I =
1
3
v
0
1
3
I
2
31
. Since v = It +v
0
, v =
1
3
v
0
1
3
I
2
31
t +v
0
. When the snowball has melted completely we have
v = 0 i
1
3
v
0
1
3
I
2
31
t +v
0
= 0 which gives t =
3
3
I
2
3
I
2 31
. Hence, it takes
3
3
I
2
3
I
2 31
33 =
3
3
I
2 31
E 11 h 33 min
longer.
5 INTEGRALS
5.1 Areas and Distances
1. (a) Since 1 is increasing, we can obtain a lower estimate by using
left endpoints. We are instructed to use ve rectangles, so n = 5.
1
5
=
5
.=1
1(r
.31
) {r [{r =
l3a
n
=
10 30
5
= 2]
= 1(r
0
) 2 +1(r
1
) 2 +1(r
2
) 2 +1(r
3
) 2 +1(r
4
) 2
= 2 [1(0) +1(2) +1(4) +1(6) +1(8)]
E 2(1 + 3 + 4.3 + 5.4 + 6.3) = 2(20) = 40
Since 1 is increasing, we can obtain an upper estimate by using
right endpoints.
1
5
=
5
.=1
1(r
.
) {r
= 2 [1(r
1
) +1(r
2
) +1(r
3
) +1(r
4
) +1(r
5
)]
= 2 [1(2) +1(4) +1(6) +1(8) +1(10)]
E 2(3 + 4.3 + 5.4 + 6.3 + 7) = 2(26) = 52
Comparing 1
5
to 1
5
, we see that we have added the area of the rightmost upper rectangle, 1(10) 2, to the sum and
subtracted the area of the leftmost lower rectangle, 1(0) 2, from the sum.
(b) 1
10
=
10
.=1
1(r
.31
) {r [{r =
10 30
10
= 1]
= 1 [1(r
0
) +1(r
1
) + +1(r
9
)]
= 1(0) +1(1) + +1 (9)
E 1 + 2.1 + 3 + 3.7 + 4.3 + 4.9 + 5.4 + 5.8 + 6.3 + 6.7
= 43.2
1
10
=
10
.=1
1(r
.
) {r = 1(1) +1(2) + +1(10)
= 1
10
+ 1 1(10) 31 1(0)

add rightmost upper rectangle,
subtract leftmost lower rectangle
= 43.2 + 7 31 = 49.2
233
234

CHAPTER 5 INTEGRALS
3. (a) 1
4
=
4
.=1
1(r
.
) {r

{r =
2 3 0
4
=

8
=

4
.=1
1(r
.
)
{r
= [1(r
1
) +1(r
2
) +1(r
3
) +1(r
4
)] {r
=

cos
r
8
+ cos
2r
8
+ cos
3r
8
+ cos
4r
8
r
8
E (0.9239 + 0.7071 + 0.3827 + 0)
r
8
E 0.7908
Since 1 is decreasing on [0. 2], an underestimate is obtained by using the right endpoint approximation, 1
4
.
(b) 1
4
=
4
.=1
1(r
.31
) {r =

4
.=1
1 (r
.31
)
{r
= [1(r
0
) +1(r
1
) +1(r
2
) +1(r
3
)] {r
=

cos 0 + cos
r
8
+ cos
2r
8
+ cos
3r
8
r
8
E (1 + 0.9239 + 0.7071 + 0.3827)
r
8
E 1.1835
1
4
is an overestimate. Alternatively, we could just add the area of the leftmost upper rectangle and subtract the area of the
rightmost lower rectangle; that is, 1
4
= 1
4
+1(0)
r
8
31
r
2
r
8
.
5. (a) 1(r) = 1 +r
2
and {r =
2 3(31)
3
= 1 i
1
3
= 1 1(0) + 1 1(1) + 1 1(2) = 1 1 + 1 2 + 1 5 = 8.
{r =
2 3(31)
6
= 0.5 i
1
6
= 0.5[1(30.5) +1(0) +1(0.5) +1(1) +1(1.5) +1(2)]
= 0.5(1.25 + 1 + 1.25 + 2 + 3.25 + 5)
= 0.5(13.75) = 6.875
(b) 1
3
= 1 1(31) + 1 1(0) + 1 1(1) = 1 2 + 1 1 + 1 2 = 5
1
6
= 0.5[1(31) +1(30.5) +1(0) +1(0.5) +1(1) +1(1.5)]
= 0.5(2 + 1.25 + 1 + 1.25 + 2 + 3.25)
= 0.5(10.75) = 5.375
(c) A
3
= 1 1(30.5) + 1 1(0.5) + 1 1(1.5)
= 1 1.25 + 1 1.25 + 1 3.25 = 5.75
A
6
= 0.5[1(30.75) +1(30.25) +1(0.25)
+ 1(0.75) +1(1.25) +1(1.75)]
= 0.5(1.5625 + 1.0625 + 1.0625 + 1.5625 + 2.5625 + 4.0625)
= 0.5(11.875) = 5.9375
(d) A
6
appears to be the best estimate.
SECTION 5.1 AREAS AND DISTANCES

235
7. Here is one possible algorithm (ordered sequence of operations) for calculating the sums:
1 Let SUM = 0, X_MIN = 0, X_MAX = 1, N = 10 (depending on which sum we are calculating),
DELTA_X = (X_MAX - X_MIN)/N, and RIGHT_ENDPOINT = X_MIN + DELTA_X.
2 Repeat steps 2a, 2b in sequence until RIGHT_ENDPOINT X_MAX.
2a Add (RIGHT_ENDPOINT)^4 to SUM.
2b Add DELTA_X to RIGHT_ENDPOINT.
At the end of this procedure, (DELTA_X)(SUM) is equal to the answer we are looking for. We nd that
1
10
=
1
10
10
.=1
j
10
4
E 0.2533, 1
30
=
1
30
30
.=1
j
30
4
E 0.2170, 1
50
=
1
50
50
.=1
j
50
4
E 0.2101. and
1
100
=
1
100
100
.=1
j
100
4
E 0.2050. It appears that the exact area is 0.2.
The following display shows the program SUMRIGHT and its output from a TI-83 Plus calculator. To generalize the
program, we have input (rather than assign) values for Xmin, Xmax, and N. Also, the function, r
4
, is assigned to Y
1
, enabling
us to evaluate any right sum merely by changing Y
1
and running the program.
9. In Maple, we have to perform a number of steps before getting a numerical answer. After loading the student package
[command: with(student);] we use the command
left_sum:=leftsum(1/(x2+1),x=0..1,10 [or 30, or 50]); which gives us the expression in summation
notation. To get a numerical approximation to the sum, we use evalf(left_sum);. Mathematica does not have a special
command for these sums, so we must type them in manually. For example, the rst left sum is given by
(1/10)*Sum[1/(((i-1)/10)2+1)],{i,1,10}], and we use the N command on the resulting output to get a
numerical approximation.
In Derive, we use the LEFT_RIEMANN command to get the left sums, but must dene the right sums ourselves.
(We can dene a new function using LEFT_RIEMANN with I ranging from 1 to n instead of from 0 to n 31.)
(a) With 1(r) =
1
r
2
+ 1
, 0 $ r $ 1, the left sums are of the form 1
n
=
1
n
n
.=1
1
.31
n
2
+ 1
. Specically, 1
10
E 0.8100,
1
30
E 0.7937, and 1
50
E 0.7904. The right sums are of the form 1
n
=
1
n
n
.=1
1
.
n
2
+ 1
. Specically, 1
10
E 0.7600,
1
30
E 0.7770, and 1
50
E 0.7804.
236

CHAPTER 5 INTEGRALS
(b) In Maple, we use the leftbox (with the same arguments as left_sum) and rightbox commands to generate the
graphs.
left endpoints, n = 10 left endpoints, n = 30 left endpoints, n = 50
right endpoints, n = 10 right endpoints, n = 30 right endpoints, n = 50
(c) We know that since j = 1(r
2
+ 1) is a decreasing function on (0. 1), all of the left sums are larger than the actual area,
and all of the right sums are smaller than the actual area. Since the left sum with n = 50 is about 0.7904 < 0.791 and the
right sum with n = 50 is about 0.7804 0.780, we conclude that 0.780 < 1
50
< exact area < 1
50
< 0.791, so the
exact area is between 0.780 and 0.791.
11. Since is an increasing function, 1
6
will give us a lower estimate and 1
6
will give us an upper estimate.
1
6
= (0 fts)(0.5 s) + (6.2)(0.5) + (10.8)(0.5) + (14.9)(0.5) + (18.1)(0.5) + (19.4)(0.5) = 0.5(69.4) = 34.7 ft
1
6
= 0.5(6.2 + 10.8 + 14.9 + 18.1 + 19.4 + 20.2) = 0.5(89.6) = 44.8 ft
13. Lower estimate for oil leakage: 1
5
= (7.6 + 6.8 + 6.2 + 5.7 + 5.3)(2) = (31.6)(2) = 63.2 L.
Upper estimate for oil leakage: 1
5
= (8.7 + 7.6 + 6.8 + 6.2 + 5.7)(2) = (35)(2) = 70 L.
15. For a decreasing function, using left endpoints gives us an overestimate and using right endpoints results in an underestimate.
We will use A
6
to get an estimate. {t = 1, so
A
6
= 1[(0.5) +(1.5) +(2.5) +(3.5) +(4.5) +(5.5)] E 55 + 40 + 28 + 18 + 10 + 4 = 155 ft
For a very rough check on the above calculation, we can draw a line from (0. 70) to (6. 0) and calculate the area of the
triangle:
1
2
(70)(6) = 210. This is clearly an overestimate, so our midpoint estimate of 155 is reasonable.
17. 1(r) =
4
I
r, 1 $ r $ 16. {r = (16 31)n = 15n and r
.
= 1 +j {r = 1 + 15jn.
= lim
n<"
1
n
= lim
n<"
n
.=1
1(r
.
) {r = lim
n<"
n
.=1
4
1 +
15j
n

15
n
.
19. 1(r) = rcos r, 0 $ r $
r
2
. {r = (
r
2
30)n =
r
2
n and r
.
= 0 +j {r =
r
2
jn.
= lim
n<"
1
n
= lim
n<"
n
.=1
1(r
.
) {r = lim
n<"
n
.=1
j
2n
cos
j
2n

2n
.
SECTION 5.2 THE DEFINITE INTEGRAL

237
21. lim
n<"
n
.=1
4n
tan
j
4n
can be interpreted as the area of the region lying under the graph of j = tanr on the interval
0.
r
4
,
since for j = tanr on
0.
r
4
with {r =
4 30
n
=

4n
, r
.
= 0 +j {r =
j
4n
, and r
W
.
= r
.
, the expression for the area is
= lim
n<"
n
.=1
1 (r
W
.
) {r = lim
n<"
n
.=1
tan
j
4n

4n
. Note that this answer is not unique, since the expression for the area is
the same for the function j = tan(r 3I) on the interval
I. I +
r
4
, where I is any integer.

23. (a) j = 1(r) = r
5
. {r =
2 30
n
=
2
n
and r
.
= 0 + j {r =
2j
n
.
= lim
n<"
1
n
= lim
n<"
n
.=1
1(r
.
) {r = lim
n<"
n
.=1
2j
n
2
n
= lim
n<"
n
.=1
32j
5
n
5

2
n
= lim
n<"
64
n
6
n
.=1
j
5
.
(b)
n
.=1
j
5
CAS
=
n
2
(n + 1)
2
2n
2
+ 2n 31
12
(c) lim
n<"
64
n
6

n
2
(n + 1)
2
2n
2
+ 2n 31
12
=
64
12
lim
n<"
n
2
+ 2n + 1
2n
2
+ 2n 31
n
2
n
2
=
16
3
lim
n<"
1 +
2
n
+
1
n
2
2 +
2
n
3
1
n
2
=
16
3
1 2 =
32
3
25. j = 1(r) = cos r. {r =
/ 30
n
=
/
n
and r
.
= 0 +j {r =
/j
n
.
= lim
n<"
1
n
= lim
n<"
n
.=1
1(r
.
) {r = lim
n<"
n
.=1
cos
/j
n
/
n
CAS
= lim
n<"
/ sin
1
2n
+ 1
2nsin
/
2n
3
/
2n
CAS
= sin/
If / =
r
2
, then = sin
r
2
= 1.
5.2 The Definite Integral
1. 1(r) = 3 3
1
2
r, 2 $ r $ 14. {r =
/ 3o
n
=
14 32
6
= 2.
Since we are using left endpoints, r
W
.
= r
.31
.
1
6
=
6
.=1
1(r
.31
) {r
= ({r) [1(r
0
) +1(r
1
) +1(r
2
) +1(r
3
) +1(r
4
) +1(r
5
)]
= 2[1(2) +1(4) +1(6) +1(8) +1(10) +1(12)]
= 2[2 + 1 + 0 + (31) + (32) + (33)] = 2(33) = 36
The Riemann sum represents the sum of the areas of the two rectangles above the r-axis minus the sum of the areas of the
three rectangles below the r-axis; that is, the net area of the rectangles with respect to the r-axis.
238

CHAPTER 5 INTEGRALS
3. 1(r) = c
i
32, 0 $ r $ 2. {r =
/ 3o
n
=
2 30
4
=
1
2
.
Since we are using midpoints, r
W
.
= r
.
=
1
2
(r
.31
+r
.
).
A
4
=
4
.=1
1(r
.
) {r = ({r) [1(r
1
) +1(r
2
) +1(r
3
) +1(r
4
)]
=
1
2
1
4
+1
3
4
+1
5
4
+1
7
4
=
1
2
(c
1'4
32) + (c
3'4
32) + (c
5'4
32) + (c
7'4
32)
r 2.322986
The Riemann sum represents the sum of the areas of the three rectangles above the r-axis minus the area of the rectangle
below the r-axis; that is, the net area of the rectangles with respect to the r-axis.
5. {r = (/ 3o)n = (8 30)4 = 84 = 2.
(a) Using the right endpoints to approximate

8
0
1(r) dr, we have
4
.=1
1(r
.
) {r = 2[1(2) +1(4) +1(6) +1(8)] E 2[1 + 2 + (32) + 1] = 4.
(b) Using the left endpoints to approximate

8
0
1(r) dr, we have
4
.=1
1(r
.31
) {r = 2[1(0) +1(2) +1(4) +1(6)] E 2[2 + 1 + 2 + (32)] = 6.
(c) Using the midpoint of each subinterval to approximate
8
0
1(r) dr, we have
4
.=1
1(r
.
) {r = 2[1(1) +1(3) +1(5) +1(7)] E 2[3 + 2 + 1 + (31)] = 10.
7. Since 1 is increasing, 1
5
$
25
0
1(r) dr $ 1
5
.
Lower estimate = 1
5
=
5
.=1
1(r
.31
) {r = 5[1(0) +1(5) +1(10) +1(15) +1(20)]
= 5(342 337 325 36 + 15) = 5(395) = 3475
Upper estimate = 1
5
=
5
.=1
1(r
.
) {r = 5[1(5) +1(10) +1(15) +1(20) +1(25)]
= 5(337 325 36 + 15 + 36) = 5(317) = 385
9. {r = (10 32)4 = 2, so the endpoints are 2, 4, 6, 8, and 10, and the midpoints are 3, 5, 7, and 9. The Midpoint Rule
gives
10
2
I
r
3
+ 1 dr E
4
.=1
1(r
.
) {r = 2
I
3
3
+ 1 +
I
5
3
+ 1 +
I
7
3
+ 1 +
I
9
3
+ 1
E 124.1644.
11. {r = (1 30)5 = 0.2, so the endpoints are 0, 0.2, 0.4, 0.6, 0.8, and 1, and the midpoints are 0.1, 0.3, 0.5, 0.7, and 0.9.
The Midpoint Rule gives
1
0
sin(r
2
) dr E
5
.=1
1(r
.
) {r = 0.2
sin(0.1)
2
+ sin(0.3)
2
+ sin(0.5)
2
+ sin(0.7)
2
+ sin(0.9)
2
E 0.3084.

239
13. In Maple, we use the command with(student); to load the sum and box commands, then
m:=middlesum(sin(x2),x=0..1,5); which gives us the sum in summation notation, then M:=evalf(m); which
gives A
5
E 0.30843908, conrming the result of Exercise 11. The command middlebox(sin(x2),x=0..1,5)
generates the graph. Repeating for n = 10 and n = 20 gives A
10
E 0.30981629 and A
20
E 0.31015563.
15. Well create the table of values to approximate
r
0
sinrdr by using the
program in the solution to Exercise 5.1.7 with Y
1
= sinr, Xmin = 0,
Xmax = , and n = 5, 10, 50, and 100.
The values of 1
n
appear to be approaching 2.
n 1
n
5 1.933766
10 1.983524
50 1.999342
100 1.999836
17. On [2. 6], lim
n<"
n
.=1
r
.
ln(1 +r
2
.
) {r =

6
2
rln(1 +r
2
) dr.
19. On [1. 8], lim
n<"
n
.=1
2r
W
.
+ (r
W
.
)
2
{r =

8
1
I
2r +r
2
dr.
21. Note that {r =
5 3(31)
n
=
6
n
and r
.
= 31 +j {r = 31 +
6j
n
.
5
31
(1 + 3r) dr = lim
n<"
n
.=1
1(r
.
) {r = lim
n<"
n
.=1
1 + 3
31 +
6j
n
6
n
= lim
n<"
6
n
n
.=1
32 +
18j
n
= lim
n<"
6
n
.=1
(32) +
n
.=1
18j
n
= lim
n<"
6
n
32n +
18
n
n
.=1
j
= lim
n<"
6
n
32n +
18
n

n(n + 1)
2
= lim
n<"
312 +
108
n
2

n(n + 1)
2
= lim
n<"
312 + 54
n + 1
n
= lim
n<"
312 + 54
1 +
1
n
= 312 + 54 1 = 42
23. Note that {r =
2 30
n
=
2
n
and r
.
= 0 +j {r =
2j
n
.
2
0
2 3r
2
dr = lim
n<"
n
.=1
1(r
.
) {r = lim
n<"
n
.=1
2 3
4j
2
n
2
2
n
= lim
n<"
2
n
.=1
2 3
4
n
2
n
.=1
j
2
= lim
n<"
2
n
2n 3
4
n
2
n
.=1
j
2
= lim
n<"
4 3
8
n
3

n(n + 1)(2n + 1)
6
= lim
n<"
4 3
4
3

n + 1
n

2n + 1
n
= lim
n<"
4 3
4
3
1 +
1
n
2 +
1
n
= 4 3
4
3
1 2 =
4
3
240

CHAPTER 5 INTEGRALS
25. Note that {r =
2 31
n
=
1
n
and r
.
= 1 +j {r = 1 +j(1n) = 1 +jn.
2
1
r
3
dr = lim
n<"
n
.=1
1(r
.
) {r = lim
n<"
n
.=1
1 +
j
n
1
n
= lim
n<"
1
n
n
.=1
n +j
n
3
= lim
n<"
1
n
4
n
.=1
n
3
+ 3n
2
j + 3nj
2
+j
3
= lim
n<"
1
n
4
.=1
n
3
+
n
.=1
3n
2
j +
n
.=1
3nj
2
+
n
.=1
j
3
= lim
n<"
1
n
4
n n
3
+ 3n
2
n
.=1
j + 3n
n
.=1
j
2
+
n
.=1
j
3
= lim
n<"
1 +
3
n
2

n(n + 1)
2
+
3
n
3

n(n + 1)(2n + 1)
6
+
1
n
4

n
2
(n + 1)
2
4
= lim
n<"
1 +
3
2

n + 1
n
+
1
2

n + 1
n

2n + 1
n
+
1
4

(n + 1)
2
n
2
= lim
n<"
1 +
3
2
1 +
1
n
+
1
2
1 +
1
n
2 +
1
n
+
1
4
1 +
1
n
= 1 +
3
2
+
1
2
2 +
1
4
= 3.75
27.

l
a
rdr = lim
n<"
/ 3o
n
n
.=1
o +
/ 3o
n
j
= lim
n<"
o(/ 3o)
n
n
.=1
1 +
(/ 3o)
2
n
2
n
.=1
j
= lim
n<"
o(/ 3o)
n
n +
(/ 3o)
2
n
2

n(n + 1)
2
= o (/ 3o) + lim
n<"
(/ 3o)
2
2
1 +
1
n
= o(/ 3o) +
1
2
(/ 3o)
2
= (/ 3o)
o +
1
2
/ 3
1
2
o
= (/ 3o)
1
2
(/ +o) =
1
2
/
2
3o
2
29. 1(r) =
r
1 +r
5
, o = 2, / = 6, and {r =
6 32
n
=
4
n
. Using Theorem 4, we get r
W
.
= r
.
= 2 +j {r = 2 +
4j
n
,
so
6
2
r
1 +r
5
dr = lim
n<"
1
n
= lim
n<"
n
.=1
2 +
4j
n
1 +
2 +
4j
n
5

4
n
.
31. {r = ( 3 0)n = n and r
W
.
= r
.
= jn.
r
0
sin5rdr = lim
n<"
n
.=1
(sin5r
.
)
= lim
n<"
n
.=1
sin
5j
n
n
CAS
= lim
n<"
1
n
cot
5
2n
CAS
=
2
5
=
2
5
33. (a) Think of
2
0
1(r) dr as the area of a trapezoid with bases 1 and 3 and height 2. The area of a trapezoid is =
1
2
(/ +1)/,
so
2
0
1(r) dr =
1
2
(1 + 3)2 = 4.
(b)

5
0
1(r) dr =

2
0
1(r) dr
trapezoid
+

3
2
1(r) dr
rectangle
+

5
3
1(r) dr
triangle
=
1
2
(1 + 3)2 + 3 1 +
1
2
2 3 = 4 + 3 + 3 = 10
(c)

7
5
1(r) dr is the negative of the area of the triangle with base 2 and height 3.

7
5
1(r) dr = 3
1
2
2 3 = 33.
(d)

9
7
1(r) dr is the negative of the area of a trapezoid with bases 3 and 2 and height 2, so it equals
3
1
2
(1 + /)/ = 3
1
2
(3 + 2)2 = 35. Thus,
9
0
1(r) dr =

5
0
1(r) dr +
7
5
1(r) dr +
9
7
1(r) dr = 10 + (33) + (35) = 2.

241
35.

3
0
1
2
r 31
dr can be interpreted as the area of the triangle above the x-axis

minus the area of the triangle below the x-axis; that is,
1
2
(1)
1
2
3
1
2
(2)(1) =
1
4
31 = 3
3
4
.
0
37.

0
33
1 +
I
9 3r
2
dr can be interpreted as the area under the graph of

1(r) = 1 +
I
9 3r
2
between r = 33 and r = 0. This is equal to one-quarter
the area of the circle with radius 3, plus the area of the rectangle, so
0
33
1 +
I
9 3r
2
dr =
1
4
3
2
+ 1 3 = 3 +
9
4
.
39.

2
31
|r| dr can be interpreted as the sum of the areas of the two shaded
triangles; that is,
1
2
(1)(1) +
1
2
(2)(2) =
1
2
+
4
2
=
5
2
.
0
41.

r
r
sin
2
rcos
4
rdr = 0 since the limits of intergration are equal.
43.

1
0
(5 36r
2
) dr =

1
0
5 dr 36
1
0
r
2
dr = 5(1 30) 36
1
3
= 5 32 = 3
45.

3
1
c
i+2
dr =

3
1
c
i
c
2
dr = c
2
3
1
c
i
dr = c
2
(c
3
3c) = c
5
3c
3
47.

2
32
1(r) dr +
5
2
1(r) dr 3
31
32
1(r) dr =

5
32
1(r) dr +
32
31
1(r) dr [by Property 5 and reversing limits]
=

5
31
1(r) dr [Property 5]
49.

9
0
[21(r) + 3o(r)] dr = 2
9
0
1(r) dr + 3
9
0
o(r) dr = 2(37) + 3(16) = 122
51. Using Integral Comparison Property 8, i $ 1(r) $ A i i(2 30) $
2
0
1(r) dr $ A(2 30) i
2i $
2
0
1(r) dr $ 2A.
53. If 31 $ r $ 1, then 0 $ r
2
$ 1 and 1 $ 1 + r
2
$ 2, so 1 $
I
1 +r
2
$
I
2 and
1[1 3(31)] $
1
31
I
1 +r
2
dr $
I
2 [1 3(31)] [Property 8]; that is, 2 $
1
31
I
1 +r
2
dr $ 2
I
2.
55. If 1 $ r $ 4, then 1 $
I
r $ 2. so 1(4 31) $
4
1
I
rdr $ 2(4 31); that is, 3 $
4
1
I
rdr $ 6.
57. If
r
4
$ r $
r
3
, then 1 $ tanr $
I
3, so 1
r
3
3
r
4
r'3
r'4
tanrdr $
I
3
r
3
3
r
4
or
r
12
$
r'3
r'4
tanrdr $
r
12
I
3.
59. The only critical number of 1(r) = rc
3i
on [0. 2] is r = 1. Since 1(0) = 0, 1(1) = c
31
E 0.368, and
1(2) = 2c
32
E 0.271, we know that the absolute minimum value of 1 on [0. 2] is 0, and the absolute maximum is c
31
. By
Property 8. 0 $ rc
3i
$ c
31
for 0 $ r $ 2 i 0(2 30) $
2
0
rc
3i
dr $ c
31
(2 30) i 0 $
2
0
rc
3i
dr $ 2c.
61.
I
r
4
+ 1 D
I
r
4
= r
2
, so
3
1
I
r
4
+ 1 dr D
3
1
r
2
dr =
1
3
3
3
31
3
=
26
3
.
242

CHAPTER 5 INTEGRALS
63. Using right endpoints as in the proof of Property 2, we calculate
l
a
c1(r) dr = lim
n<"
n
.=1
c1(r
.
) {r = lim
n<"
c
n
.=1
1(r
.
) {r = c lim
n<"
n
.=1
1(r
.
) {r = c
l
a
1(r) dr.
65. Since 3|1(r)| $ 1(r) $ |1(r)|, it follows from Property 7 that
3
l
a
|1(r)| dr $
l
a
1(r) dr $
l
a
|1(r)| dr i

l
a
1(r) dr
l
a
|1(r)| dr
Note that the denite integral is a real number, and so the following property applies: 3o $ / $ o i |/| $ o for all real
numbers / and nonnegative numbers o.
67. To show that 1 is integrable on [0. 1] . we must show that lim
n<"
n
.=1
1(r
W
.
) {r exists. Let n denote a positive integer and divide
the interval [0. 1] into n equal subintervals
0.
1
n
1
n
.
2
n
, ... ,
n 31
n
. 1
. If we choose r
W
.
to be a rational number in the ith
subinterval, then we obtain the Riemann sum
n
.=1
1(r
W
.
)
1
n
= 0, so lim
n<"
n
.=1
1(r
W
.
)
1
n
= lim
n<"
0 = 0. Now suppose we
choose r
W
.
to be an irrational number. Then we get
n
.=1
1(r
W
.
)
1
n
=
n
.=1
1
1
n
= n
1
n
= 1 for each n, so
lim
n<"
n
.=1
1(r
W
.
)
1
n
= lim
n<"
1 = 1. Since the value of lim
n<"
n
.=1
1(r
W
.
) {r depends on the choice of the sample points r
W
.
, the
limit does not exist, and 1 is not integrable on [0. 1].
69. lim
n<"
n
.=1
j
4
n
5
= lim
n<"
n
.=1
j
4
n
4

1
n
= lim
n<"
n
.=1
j
n
4
1
n
. At this point, we need to recognize the limit as being of the form
lim
n<"
n
.=1
1(r
.
) {r, where {r = (1 30)n = 1n, r
.
= 0 +j {r = jn, and 1(r) = r
4
. Thus, the denite integral
is
1
0
r
4
dr.
71. Choose r
.
= 1 +
j
n
and r
W
.
=
I
r
.31
r
.
=
1 +
j 31
n
1 +
j
n
. Then
2
1
r
32
dr = lim
n<"
1
n
n
.=1
1
1 +
. 31
n
1 +
.
n
= lim
n<"
n
n
.=1
1
(n +j 31)(n +j)
= lim
n<"
n
n
.=1
1
n +j 31
3
1
n +j
[by the hint] = lim

n<"
n
n31
.=0
1
n +j
3
n
.=1
1
n +j
= lim
n<"
n
1
n
+
1
n + 1
+ +
1
2n 31
1
n + 1
+ +
1
2n 31
+
1
2n
= lim
n<"
n
1
n
3
1
2n
= lim
n<"
1 3
1
2
=
1
2
5.3 The Fundamental Theorem of Calculus
1. One process undoes what the other one does. The precise version of this statement is given by the Fundamental Theorem of
Calculus. See the statement of this theorem and the paragraph that follows it on page 387.
SECTION 5.3 THE FUNDAMENTAL THEOREM OF CALCULUS

243
3. (a) o(r) =

i
0
1(t) dt.
o(0) =
0
0
1(t) dt = 0
o(1) =
1
0
1(t) dt = 1 2 = 2 [rectangle],
o(2) =
2
0
1(t) dt =

1
0
1(t) dt +
2
1
1(t) dt = o(1) +
2
1
1(t) dt
= 2 + 1 2 +
1
2
1 2 = 5 [rectangle plus triangle],
o(3) =
3
0
1(t) dt = o(2) +
3
2
1(t) dt = 5 +
1
2
1 4 = 7,
o(6) = o(3) +
6
3
1(t) dt [the integral is negative since 1 lies under the r-axis]
= 7 +
1
2
2 2 + 1 2
= 7 34 = 3
(d)
(b) o is increasing on (0. 3) because as r increases from 0 to 3, we keep adding more area.
(c) o has a maximum value when we start subtracting area; that is, at r = 3.
5. (a) By FTC1 with 1(t) = t
2
and o = 1, o(r) =
i
1
t
2
dt i
o
0
(r) = 1(r) = r
2
.
(b) Using FTC2, o(r) =

i
1
t
2
dt =

1
3
t
3
i
1
=
1
3
r
3
3
1
3
i o
0
(r) = r
2
.
7. 1(t) =
1
t
3
+ 1
and o(r) =

i
1
1
t
3
+ 1
dt, so by FTC1, o
0
(r) = 1(r) =
1
r
3
+ 1
. Note that the lower limit, 1, could be any
real number greater than 31 and not affect this answer.
9. 1(t) = t
2
sint and o(j) =
u
2
t
2
sint dt, so by FTC1, o
0
(j) = 1(j) = j
2
sinj.
11. 1(r) =
r
i
I
1 + sec t dt = 3
i
r
I
1 + sec t dt i 1
0
(r) = 3
d
dr
i
r
I
1 + sec t dt = 3
I
1 + sec r
13. Let & =
1
r
. Then
d&
dr
= 3
1
r
2
. Also,
d/
dr
=
d/
d&
d&
dr
, so
/
0
(r) =
d
dr
1'i
2
arctant dt =
d
d&
u
2
arctant dt
d&
dr
= arctan&
d&
dr
= 3
arctan(1r)
r
2
.
15. Let & = tanr. Then
d&
dr
= sec
2
r. Also,
dj
dr
=
dj
d&
d&
dr
, so
j
0
=
d
dr
tan i
0
t +
I
t dt =
d
d&
u
0
t +
I
t dt
d&
dr
=

& +
I
&
d&
dr
=

tanr +
I
tanrsec
2
r.
17. Let n = 1 33r. Then
dn
dr
= 33. Also,
dj
dr
=
dj
dn
dn
dr
, so
j
0
=
d
dr
1
133i
&
3
1 +&
2
d& =
d
dn
1
u
&
3
1 +&
2
d&
dn
dr
= 3
d
dn
u
1
&
3
1 +&
2
d&
dn
dr
= 3
n
3
1 +n
2
(33) =
3(1 33r)
3
1 + (1 33r)
2
19.

2
31
r
3
32r
dr =

r
4
4
3r
2
2
31
=

2
4
4
32
2
(31)
4
4
3(31)
2
= (4 34) 3
1
4
31
= 0 3
3
3
4
=
3
4
244

CHAPTER 5 INTEGRALS
21.

4
1
(5 32t + 3t
2
) dt =

5t 3t
2
+t
3
4
1
= (20 316 + 64) 3(5 31 + 1) = 68 35 = 63
23.

1
0
r
4'5
dr =

5
9
r
9'5
1
0
=
5
9
30 =
5
9
25.

2
1
3
t
4
dt = 3
2
1
t
34
dt = 3
t
33
33
2
1
=
3
33
1
t
3
2
1
= 31
1
8
31
=
7
8
27.

2
0
r(2 +r
5
) dr =

2
0
(2r +r
6
) dr =

r
2
+
1
7
r
7
2
0
=

4 +
128
7
3(0 + 0) =
156
7
29.

9
1
r 31
I
r
dr =

9
1
r
I
r
3
1
I
r
dr =

9
1
(r
1'2
3r
31'2
) dr =

2
3
r
3'2
32r
1'2
9
1
=

2
3
27 32 3
2
3
32
= 12 3
3
4
3
=
40
3
31.

r'4
0
sec
2
t dt =

tant
r'4
0
= tan
r
4
3tan0 = 1 30 = 1
33.

2
1
(1 + 2j)
2
dj =

2
1
(1 + 4j + 4j
2
) dj =

j + 2j
2
+
4
3
j
3
2
1
=

2 + 8 +
32
3
1 + 2 +
4
3
=
62
3
3
13
3
=
49
3
35.

9
1
1
2r
dr =
1
2
9
1
1
r
dr =
1
2
ln|r|
9
1
=
1
2
(ln9 3ln1) =
1
2
ln9 30 = ln9
1'2
= ln3
37.

I
3'2
1'2
6
I
1 3t
2
dt = 6
I
3'2
1'2
1
I
1 3t
2
dt = 6
sin
31
t
I
3'2
1'2
= 6
sin
31
I
3
2
3sin
31
1
2
= 6
r
3
3
r
6
= 6
r
6
=
39.

1
31
c
u+1
d& =

c
u+1
1
31
= c
2
3c
0
= c
2
31 [or start with c
u+1
= c
u
c
1
]
41. If 1(r) =
sinr if 0 $ r < 2
cos r if 2 $ r $
then
r
0
1(r) dr =

r'2
0
sinrdr +
r
r'2
cos rdr =

3cos r
r'2
0
+ [sinr]
r
r'2
= 3cos
r
2
+ cos 0 + sin 3sin
r
2
= 30 + 1 + 0 31 = 0
Note that 1 is integrable by Theorem 3 in Section 5.2.
43. 1(r) = r
34
is not continuous on the interval [32. 1], so FTC2 cannot be applied. In fact, 1 has an innite discontinuity at
r = 0, so
1
32
r
34
dr does not exist.
45. 1(0) = sec 0 tan0 is not continuous on the interval [3. ], so FTC2 cannot be applied. In fact, 1 has an innite
discontinuity at r = 2, so
r
r'3
sec 0 tan0 d0 does not exist.
47. From the graph, it appears that the area is about 60. The actual area is
27
0
r
1'3
dr =

3
4
r
4'3
27
0
=
3
4
81 30 =
243
4
= 60.75. This is
3
4
of the
area of the viewing rectangle.
SECTION 5.3 THE FUNDAMENTAL THEOREM OF CALCULUS

245
49. It appears that the area under the graph is about
2
3
of the area of the viewing
rectangle, or about
2
3
E 2.1. The actual area is
r
0
sinrdr = [3cos r]
r
0
= (3cos ) 3(3cos 0) = 3(31) + 1 = 2.
51.

2
31
r
3
dr =

1
4
r
4
2
31
= 4 3
1
4
=
15
4
= 3.75
53. o(r) =

3i
2i
&
2
31
&
2
+ 1
d& =

0
2i
&
2
31
&
2
+ 1
d& +
3i
0
&
2
31
&
2
+ 1
d& = 3
2i
0
&
2
31
&
2
+ 1
d& +
3i
0
&
2
31
&
2
+ 1
d& i
o
0
(r) = 3
(2r)
2
31
(2r)
2
+ 1

d
dr
(2r) +
(3r)
2
31
(3r)
2
+ 1

d
dr
(3r) = 32
4r
2
31
4r
2
+ 1
+ 3
9r
2
31
9r
2
+ 1
55. j =

i
3
I
i
I
t sint dt =

1
I
i
I
t sint dt +
i
3
1
I
t sint dt = 3
I
i
1
I
t sint dt +
i
3
1
I
t sint dt i
j
0
= 3
4
I
r(sin
I
r)
d
dr
(
I
r) +r
3'2
sin(r
3
)
d
dr
r
3
= 3
4
I
rsin
I
r
2
I
r
+r
3'2
sin(r
3
)(3r
2
)
= 3r
7'2
sin(r
3
) 3
sin
I
r
2
4
I
r
57. 1(r) =

i
1
1(t) dt i 1
0
(r) = 1(r) =

i
2
1
I
1 +&
4
&
d&
since 1(t) =
I
2
1
I
1 +&
4
&
d&
i
1
00
(r) = 1
0
(r) =
1 + (r
2
)
4
r
2

d
dr
r
2
=
I
1 +r
8
r
2
2r =
2
I
1 +r
8
r
. So 1
00
(2) =
I
1 + 2
8
=
I
257.
59. By FTC2,
4
1
1
0
(r) dr = 1(4) 31(1), so 17 = 1(4) 312 i 1(4) = 17 + 12 = 29.
61. (a) The Fresnel function o(r) =
i
0
sin
r
2
t
2
dt has local maximum values where 0 = o

0
(r) = sin
r
2
t
2
and
o
0
changes from positive to negative. For r 0, this happens when
r
2
r
2
= (2n 31) [odd multiples of ] C
r
2
= 2(2n 31) C r =
I
4n 32, n any positive integer. For r < 0, o
0
changes from positive to negative where
r
2
r
2
= 2n [even multiples of ] C r
2
= 4n C r = 32
I
n. o
0
does not change sign at r = 0.
(b) o is concave upward on those intervals where o
00
(r) 0. Differentiating our expression for o
0
(r), we get
o
00
(r) = cos
r
2
r
2
2
r
2
r
= rcos
r
2
r
2
. For r 0, o
00
(r) 0 where cos(
r
2
r
2
) 0 C 0 <
r
2
r
2
<
r
2
or
2n 3
1
2
<
r
2
r
2
<

2n +
1
2
, n any integer C 0 < r < 1 or

I
4n 31 < r <
I
4n + 1, n any positive integer.
For r < 0, o
00
(r) 0 where cos(
r
2
r
2
) < 0 C

2n 3
3
2
<
r
2
r
2
<

2n 3
1
2
, n any integer C
4n 33 < r
2
< 4n 31 C
I
4n 33 < |r| <
I
4n 31 i
I
4n 33 < 3r <
I
4n 31 i
246

CHAPTER 5 INTEGRALS
3
I
4n 33 r 3
I
4n 31, so the intervals of upward concavity for r < 0 are
3
I
4n 31. 3
I
4n 33
, n any
positive integer. To summarize: o is concave upward on the intervals (0. 1),
3
I
3. 31
,
I
3.
I
5
3
I
7. 3
I
5
,
I
7. 3
, . . . .
(c) In Maple, we use plot({int(sin(Pi*t2/2),t=0..x),0.2},x=0..2);. Note that
Maple recognizes the Fresnel function, calling it FresnelS(x). In Mathematica, we use
Plot[{Integrate[Sin[Pi*t2/2],{t,0,x}],0.2},{x,0,2}]. In Derive, we load the utility le
FRESNEL and plot FRESNEL_SIN(x). From the graphs, we see that
i
0
sin
r
2
t
2
dt = 0.2 at r E 0.74.
63. (a) By FTC1, o
0
(r) = 1(r). So o
0
(r) = 1(r) = 0 at r = 1. 3. 5. 7, and 9. o has local maxima at r = 1 and 5 (since 1 = o
0
changes from positive to negative there) and local minima at r = 3 and 7. There is no local maximum or minimum at
r = 9, since 1 is not dened for r 9.
(b) We can see from the graph that
1
0
1 dt
<

3
1
1 dt
<

5
3
1 dt
<

7
5
1 dt
<

9
7
1 dt
. So o(1) =

1
0
1 dt
,
o(5) =

5
0
1 dt = o(1) 3
3
1
1 dt
5
3
1 dt
, and o(9) =

9
0
1 dt = o(5) 3
7
5
1 dt
9
7
1 dt
. Thus,
o(1) < o(5) < o(9), and so the absolute maximum of o(r) occurs at r = 9.
(c) o is concave downward on those intervals where o
00
< 0. But o
0
(r) = 1(r),
so o
00
(r) = 1
0
(r), which is negative on (approximately)
1
2
. 2
, (4. 6) and
(8. 9). So o is concave downward on these intervals.
(d)
65. lim
n<"
n
.=1
j
3
n
4
= lim
n<"
1 30
n
n
.=1
j
n
3
=

1
0
r
3
dr =

r
4
4
1
0
=
1
4
67. Suppose / < 0. Since 1 is continuous on [r +/. r], the Extreme Value Theorem says that there are numbers & and in
[r +/. r] such that 1(&) = i and 1() = A, where i and A are the absolute minimum and maximum values of 1 on
[r +/. r]. By Property 8 of integrals, i(3/) $
i
i+I
1(t) dt $ A(3/); that is, 1(&)(3/) $ 3
i+I
i
1(t) dt $ 1()(3/).
Since 3/ 0, we can divide this inequality by 3/: 1(&) $
1
/
i+I
i
1 (t) dt $ 1(). By Equation 2,
o(r +/) 3o(r)
/
=
1
/
i+I
i
1(t) dt for / 6= 0, and hence 1(&) $
o(r +/) 3o(r)
/
$ 1(), which is Equation 3 in the
case where / < 0.
SECTION 5.4 INDEFINITE INTEGRALS AND THE NET CHANGE THEOREM

247
69. (a) Let 1(r) =
I
r i 1
0
(r) = 1(2
I
r) 0 for r 0 i 1 is increasing on (0. "). If r D 0, then r
3
D 0, so
1 +r
3
D 1 and since 1 is increasing, this means that 1
1 +r
3
D 1(1) i
I
1 +r
3
D 1 for r D 0. Next let
o(t) = t
2
3t i o
0
(t) = 2t 31 i o
0
(t) 0 when t D 1. Thus, o is increasing on (1. "). And since o(1) = 0,
o(t) D 0 when t D 1. Now let t =
I
1 +r
3
, where r D 0.
I
1 +r
3
D 1 (from above) i t D 1 i o(t) D 0 i
1 +r
3
3
I
1 +r
3
D 0 for r D 0. Therefore, 1 $
I
1 +r
3
$ 1 +r
3
for r D 0.
(b) From part (a) and Property 7:

1
0
1 dr $
1
0
I
1 +r
3
dr $
1
0
(1 +r
3
) dr C
1
0
$
1
0
I
1 +r
3
dr $
r +
1
4
r
4
1
0
C 1 $
1
0
I
1 +r
3
dr $ 1 +
1
4
= 1.25.
71. 0 <
r
2
r
4
+r
2
+ 1
<
r
2
r
4
=
1
r
2
on [5. 10], so
0 $
10
5
r
2
r
4
+r
2
+ 1
dr <
10
5
1
r
2
dr =

3
1
r
10
5
= 3
1
10
3
3
1
5
=
1
10
= 0.1.
73. Using FTC1, we differentiate both sides of 6 +
i
a
1(t)
t
2
dt = 2
I
r to get
1(r)
r
2
= 2
1
2
I
r
i 1(r) = r
3'2
.
To nd o, we substitute r = o in the original equation to obtain 6 +
a
a
1(t)
t
2
dt = 2
I
o i 6 + 0 = 2
I
o i
3 =
I
o i o = 9.
75. (a) Let 1(t) =
I
0
1(c) dc. Then, by FTC1, 1
0
(t) = 1(t) = rate of depreciation, so 1(t) represents the loss in value over the
interval [0. t].
(b) C(t) =
1
t
I
0
1(c) dc
=
+1(t)
t
represents the average expenditure per unit of t during the interval [0. t],
assuming that there has been only one overhaul during that time period. The company wants to minimize average
expenditure.
(c) C(t) =
1
t
I
0
1(c) dc
. Using FTC1, we have C

0
(t) = 3
1
t
2
I
0
1(c) dc
+
1
t
1(t).
C
0
(t) = 0 i t 1(t) = +
I
0
1(c) dc i 1(t) =
1
t
I
0
1(c) dc
= C(t).
5.4 Indefinite Integrals and the Net Change Theorem
1.
d
dr
I
r
2
+ 1 +C
=
d
dr
r
2
+ 1
1'2
+C
=
1
2
r
2
+ 1
31'2
2r + 0 =
r
I
r
2
+ 1
3.
d
dr
sinr 3
1
3
sin
3
r +C
=
d
dr
sinr 3
1
3
(sinr)
3
+C
= cos r 3
1
3
3(sinr)
2
(cos r) + 0
= cos r(1 3sin
2
r) = cos r(cos
2
r) = cos
3
r
5.

(r
2
+r
32
) dr =
r
3
3
+
r
31
31
+C =
1
3
r
3
3
1
r
+C
248

CHAPTER 5 INTEGRALS
7.

r
4
3
1
2
r
3
+
1
4
r 32
dr =
r
5
5
3
1
2
r
4
4
+
1
4
r
2
2
32r +C =
1
5
r
5
3
1
8
r
4
+
1
8
r
2
32r +C
9.

(1 3t)(2 +t
2
) dt =

(2 32t +t
2
3t
3
) dt = 2t 32
t
2
2
+
t
3
3
3
t
4
4
+C = 2t 3t
2
+
1
3
t
3
3
1
4
t
4
+C
11.

r
3
32
I
r
r
dr =

r
3
r
3
2r
1'2
r
dr =

(r
2
32r
31'2
) dr =
r
3
3
32
r
1'2
12
+C =
1
3
r
3
34
I
r +C
13.

(sinr + sinhr) dr = 3cos r + coshr +C
15.

(0 3csc 0 cot 0) d0 =
1
2
0
2
+ csc 0 +C
17.

(1 + tan
2
c) dc =

sec
2
cdc = tanc +C
19.

cos r +
1
2
r
dr = sinr +
1
4
r
2
+C. The members of the family
in the gure correspond to C = 35, 0, 5, and 10.
21.

2
0
(6r
2
34r + 5) dr =

6
1
3
r
3
34
1
2
r
2
+ 5r
2
0
=

2r
3
32r
2
+ 5r
2
0
= (16 38 + 10) 30 = 18
23.

0
31
(2r 3c
i
) dr =

r
2
3c
i
0
31
= (0 31) 3
1 3c
31
= 32 + 1c
25.

2
32
(3& + 1)
2
d& =

2
32
9&
2
+ 6& + 1
d& =

9
1
3
&
3
+ 6
1
2
&
2
+&
2
32
=

3&
3
+ 3&
2
+&
2
32
= (24 + 12 + 2) 3(324 + 12 32) = 38 3(314) = 52
27.

4
1
I
t (1 +t) dt =

4
1
(t
1'2
+t
3'2
) dt =

2
3
t
3'2
+
2
5
t
5'2
4
1
=

16
3
+
64
5
2
3
+
2
5
=
14
3
+
62
5
=
256
15
29.

31
32
4j
3
+
2
j
3
dj =

4
1
4
j
4
+ 2
1
32
j
32
31
32
=

j
4
3
1
j
2
31
32
= (1 31) 3
16 3
1
4
= 3
63
4
31.

1
0
r
3
I
r +
4
I
r
dr =

1
0
(r
4'3
+r
5'4
) dr =

3
7
r
7'3
+
4
9
r
9'4
1
0
=

3
7
+
4
9
30 =
55
63
33.

4
1
5rdr =
I
5
4
1
r
31'2
dr =
I
5
2
I
r
4
1
=
I
5 (2 2 32 1) = 2
I
5
35.

r
0
(4 sin0 33 cos 0) d0 =

34 cos 0 33 sin0
r
0
= (4 30) 3(34 30) = 8
37.

r'4
0
1 + cos
2
0
cos
2
0
d0 =

r'4
0
1
cos
2
0
+
cos
2
0
cos
2
0
d0 =

r'4
0
(sec
2
0 + 1) d0
=

tan0 +0
r'4
0
=

tan
r
4
+
r
4
3(0 + 0) = 1 +
r
4
SECTION 5.4 INDEFINITE INTEGRALS AND THE NET CHANGE THEOREM

249
39.

64
1
1 +
3
I
r
I
r
dr =

64
1
1
r
1'2
+
r
1'3
r
1'2
dr =

64
1
r
31'2
+r
(1'3) 3(1'2)
dr =

64
1
(r
31'2
+r
31'6
) dr
=

2r
1'2
+
6
5
r
5'6
64
1
=

16 +
192
5
2 +
6
5
= 14 +
186
5
=
256
5
41.

1'
I
3
0
t
2
31
t
4
31
dt =

1'
I
3
0
t
2
31
(t
2
+ 1)(t
2
31)
dt =

1'
I
3
0
1
t
2
+ 1
dt =

arctant
1'
I
3
0
= arctan
1
I
3
3arctan0
=
r
6
30 =
r
6
43.

2
31
(r 32 |r|) dr =

0
31
[r 32(3r)] dr +
2
0
[r 32(r)] dr =

0
31
3rdr +
2
0
(3r) dr = 3
1
2
r
2
0
31
3
1
2
r
2
2
0
= 3
0 3
1
2
3(2 30) = 3
7
2
= 33.5
45. The graph shows that j = r +r
2
3r
4
has r-intercepts at r = 0 and at
r = o E 1.32. So the area of the region that lies under the curve and above the
r-axis is
a
0
(r +r
2
3r
4
) dr =

1
2
r
2
+
1
3
r
3
3
1
5
r
5
a
0
=

1
2
o
2
+
1
3
o
3
3
1
5
o
5
30 E 0.84
47. =

2
0
2j 3j
2
dj =

j
2
3
1
3
j
3
2
0
=

4 3
8
3
30 =
4
3
49. If n
0
(t) is the rate of change of weight in pounds per year, then n(t) represents the weight in pounds of the child at age t. We
know from the Net Change Theorem that
10
5
n
0
(t) dt = n(10) 3n(5), so the integral represents the increase in the childs
weight (in pounds) between the ages of 5 and 10.
51. Since v(t) is the rate at which oil leaks, we can write v(t) = 3\
0
(t), where \ (t) is the volume of oil at time t. [Note that the
minus sign is needed because \ is decreasing, so \
0
(t) is negative, but v(t) is positive.] Thus, by the Net Change Theorem,
120
0
v(t) dt = 3
120
0
\
0
(t) dt = 3[\ (120) 3\ (0)] = \ (0) 3\ (120), which is the number of gallons of oil that leaked
from the tank in the rst two hours (120 minutes).
53. By the Net Change Theorem,
5000
1000
1
0
(r) dr = 1(5000) 31(1000), so it represents the increase in revenue when
production is increased from 1000 units to 5000 units.
55. In general, the unit of measurement for
l
a
1(r) dr is the product of the unit for 1(r) and the unit for r. Since 1(r) is
measured in newtons and r is measured in meters, the units for
100
0
1(r) dr are newton-meters. (A newton-meter is
abbreviated Nm and is called a joule.)
57. (a) Displacement =

3
0
(3t 35) dt =

3
2
t
2
35t
3
0
=
27
2
315 = 3
3
2
m
(b) Distance traveled =

3
0
|3t 35| dt =

5'3
0
(5 33t) dt +
3
5'3
(3t 35) dt
=

5t 3
3
2
t
2
5'3
0
+
3
2
t
2
35t
3
5'3
=
25
3
3
3
2

25
9
+
27
2
315 3
3
2

25
9
3
25
3
=
41
6
m
59. (a)
0
(t) = o(t) = t + 4 i (t) =
1
2
t
2
+ 4t +C i (0) = C = 5 i (t) =
1
2
t
2
+ 4t + 5 ms
(b) Distance traveled =

10
0
|(t)| dt =

10
0
1
2
t
2
+ 4t + 5
dt =

10
0
1
2
t
2
+ 4t + 5
dt =

1
6
t
3
+ 2t
2
+ 5t
10
0
=
500
3
+ 200 + 50 = 416
2
3
m
250

CHAPTER 5 INTEGRALS
61. Since i
0
(r) = a(r), i =

4
0
a(r) dr =

4
0
9 + 2
I
r
dr =

9r +
4
3
r
3'2
4
0
= 36 +
32
3
30 =
140
3
= 46
2
3
kg.
63. Let c be the position of the car. We know from Equation 2 that c(100) 3c(0) =

100
0
(t) dt. We use the Midpoint Rule for
0 $ t $ 100 with n = 5. Note that the length of each of the ve time intervals is 20 seconds =
20
3600
hour =
1
180
hour.
So the distance traveled is
100
0
(t) dt E
1
180
[(10) +(30) +(50) +(70) +(90)] =
1
180
(38 + 58 + 51 + 53 + 47) =
247
180
E 1.4 miles.
65. From the Net Change Theorem, the increase in cost if the production level is raised
from 2000 yards to 4000 yards is C(4000) 3 C(2000) =

4000
2000
C
0
(r) dr.
4000
2000
C
0
(r) dr =

4000
2000
3 30.01r + 0.000006r
2
dr =

3r 30.005r
2
+ 0.000002r
3
4000
2000
= 60,00032,000 = $58,000
67. (a) We can nd the area between the Lorenz curve and the line j = r by subtracting the area under j = 1(r) from the area
under j = r. Thus,
coefcient of inequality =
area between Lorenz curve and line j = r
area under line j = r
=
1
0
[r 31(r)] dr
1
0
rdr
=
1
0
[r 31(r)] dr
[r
2
2]
1
0
=
1
0
[r 31(r)] dr
12
= 2
1
0
[r 31(r)] dr
(b) 1(r) =
5
12
r
2
+
7
12
r i 1(50%) = 1
1
2
=
5
48
+
7
24
=
19
48
= 0.39583, so the bottom 50% of the households receive
at most about 40% of the income. Using the result in part (a),
coefcient of inequality = 2
1
0
[r 31(r)] dr = 2
1
0
r 3
5
12
r
2
3
7
12
r
dr = 2
1
0
5
12
r 3
5
12
r
2
dr
= 2
1
0
5
12
(r 3r
2
) dr =
5
6
1
2
r
2
3
1
3
r
3
1
0
=
5
6
1
2
3
1
3
=
5
6
1
6
=
5
36
5.5 The Substitution Rule
1. Let & = 3r. Then d& = 3dr. so dr = 3d&. Thus,
c
3i
dr =

c
u
(3d&) = 3c
u
+C = 3c
3i
+C. Dont forget that it
is often very easy to check an indenite integration by differentiating your answer. In this case,
d
dr
(3c
3i
+C) = 3[c
3i
(31)] = c
3i
, the desired result.
3. Let & = r
3
+ 1. Then d& = 3r
2
dr and r
2
dr =
1
3
d&, so
r
2
r
3
+ 1 dr =

I
&
1
3
d&
=
1
3
&
3'2
32
+C =
1
3

2
3
&
3'2
+C =
2
9
(r
3
+ 1)
3'2
+C.
5. Let & = cos 0. Then d& = 3sin0 d0 and sin0 d0 = 3d&, so
cos
3
0 sin0 d0 =

&
3
(3d&) = 3
&
4
4
+C = 3
1
4
cos
4
0 +C.
7. Let & = r
2
. Then d& = 2rdr and rdr =
1
2
d&, so
rsin(r
2
) dr =

sin&
1
2
d&
= 3
1
2
cos & +C = 3
1
2
cos(r
2
) +C.
SECTION 5.5 THE SUBSTITUTION RULE

251
9. Let & = 3r32. Then d& = 3 dr and dr =
1
3
d&, so
(3r32)
20
dr =

&
20
1
3
d&
=
1
3

1
21
&
21
+C =
1
63
(3r32)
21
+C.
11. Let & = 2r + r
2
. Then d& = (2 + 2r) dr = 2(1 + r) dr and (r + 1) dr =
1
2
d&, so
(r + 1)
2r +r
2
dr =

I
&

1
2
d&
=
1
2
&
3'2
32
+C =
1
3
2r +r
2
3'2
+C.
Or: Let & =
I
2r +r
2
. Then &
2
= 2r +r
2
i 2&d& = (2 + 2r) dr i &d& = (1 +r) dr, so
(r + 1)
I
2r +r
2
dr =

& &d& =

&
2
d& =
1
3
&
3
+C =
1
3
(2r +r
2
)
3'2
+C.
13. Let & = 5 3 3r. Then d& = 33 dr and dr = 3
1
3
d&, so
dr
5 33r
=

1
&
3
1
3
d&
= 3
1
3
ln|&| +C = 3
1
3
ln|5 33r| +C.
15. Let & = t. Then d& = dt and dt =
1
r
d&, so
sint dt =

sin&
1
r
d&
=
1
r
(3cos &) +C = 3
1
r
cos t +C.
17. Let & = 3or + /r
3
. Then d& = (3o + 3/r
2
) dr = 3(o + /r
2
) dr, so
o +/r
2
I
3or +/r
3
dr =

1
3
d&
&
1'2
=
1
3
&
31'2
d& =
1
3
2&
2
+C =
2
3
3or +/r
3
+C.
19. Let & = lnr. Then d& =
dr
r
, so
(lnr)
2
r
dr =

&
2
d& =
1
3
&
3
+C =
1
3
(lnr)
3
+C.
21. Let & =
I
t. Then d& =
dt
2
I
t
and
1
I
t
dt = 2 d&, so
cos
I
t
I
t
dt =

cos &(2 d&) = 2 sin& +C = 2 sin
I
t +

C.
23. Let & = sin0. Then d& = cos 0 d0, so
cos 0 sin
6
0 d0 =

&
6
d& =
1
7
&
7
+C =
1
7
sin
7
0 +C.
25. Let & = 1 +c
i
. Then d& = c
i
dr, so
c
i
I
1 +c
i
dr =
I
&d& =
2
3
&
3'2
+C =
2
3
(1 +c
i
)
3'2
+C.
Or: Let & =
I
1 +c
i
. Then &
2
= 1 + c
i
and 2&d& = c
i
dr, so
c
i
I
1 +c
i
dr =

& 2&d& =
2
3
&
3
+C =
2
3
(1 +c
i
)
3'2
+C.
27. Let & = 1 +:
3
. Then d& = 3:
2
d: and :
2
d: =
1
3
d&, so
:
2
3
I
1 +:
3
d: =

&
31'3
1
3
d&
=
1
3

3
2
&
2'3
+C =
1
2
(1 +:
3
)
2'3
+C.
29. Let & = tanr. Then d& = sec
2
rdr, so
c
tan i
sec
2
rdr =

c
u
d& = c
u
+C = c
tan i
+C.
31. Let & = sinr. Then d& = cos rdr, so
cos r
sin
2
r
dr =

1
&
2
d& =

&
32
d& =
&
31
31
+C = 3
1
&
+C = 3
1
sinr
+C
[or 3csc r +C ].
33. Let & = cot r. Then d& = 3csc
2
rdr and csc
2
rdr = 3d&, so
I
cot rcsc
2
rdr =

I
&(3d&) = 3
&
3'2
32
+C = 3
2
3
(cot r)
3'2
+C.
252

CHAPTER 5 INTEGRALS
35.

sin2r
1 + cos
2
r
dr = 2
sinrcos r
1 + cos
2
r
dr = 21. Let & = cos r. Then d& = 3sinrdr, so
21 = 32
&d&
1 +&
2
= 32
1
2
ln(1 +&
2
) +C = 3ln(1 +&
2
) +C = 3ln(1 + cos
2
r) +C.
Or: Let & = 1 + cos
2
r.
37.

cot rdr =

cos r
sinr
dr. Let & = sinr. Then d& = cos rdr, so
cot rdr =

1
&
d& = ln|&| +C = ln|sinr| +C.
39. Let & = sec r. Then d& = sec r tanrdr, so
sec
3
r tanrdr =

sec
2
r(sec r tanr) dr =

&
2
d& =
1
3
&
3
+C =
1
3
sec
3
r +C.
41. Let & = sin
31
r. Then d& =
1
I
1 3r
2
dr, so
dr
I
1 3r
2
sin
31
r
=

1
&
d& = ln|&| +C = ln
sin
31
r
+C.
43. Let & = 1 +r
2
. Then d& = 2rdr, so
1 +r
1 +r
2
dr =

1
1 +r
2
dr +
r
1 +r
2
dr = tan
31
r +
1
2
d&
&
= tan
31
r +
1
2
ln|&| +C
= tan
31
r +
1
2
ln
1 +r
2
+C = tan
31
r +
1
2
ln
1 +r
2
+C [since 1 +r
2
0].
45. Let & = r + 2. Then d& = dr, so
r
4
I
r + 2
dr =

& 32
4
I
&
d& =

(&
3'4
32&
31'4
) d& =
4
7
&
7'4
32
4
3
&
3'4
+C
=
4
7
(r + 2)
7'4
3
8
3
(r + 2)
3'4
+C
In Exercises 4750, let 1(r) denote the integrand and 1(r) its antiderivative (with C = 0).
47. 1(r) = r(r
2
31)
3
. & = r
2
31 i d& = 2rdr. so
r(r
2
31)
3
dr =

&
3
1
2
d&
=
1
8
&
4
+C =
1
8
(r
2
31)
4
+C
Where 1 is positive (negative), 1 is increasing (decreasing). Where 1
changes from negative to positive (positive to negative), 1 has a local
minimum (maximum).
49. 1(r) = sin
3
r cos r. & = sinr i d& = cos rdr, so
sin
3
r cos rdr =

&
3
d& =
1
4
&
4
+C =
1
4
sin
4
r +C
Note that at r =
r
2
, 1 changes from positive to negative and 1 has a local
maximum. Also, both 1 and 1 are periodic with period , so at r = 0 and
at r = , 1 changes from negative to positive and 1 has local minima.
51. Let & = r 31, so d& = dr. When r = 0, & = 31; when r = 2, & = 1. Thus,
2
0
(r 31)
25
dr =

1
31
&
25
d& = 0 by
Theorem 7(b), since 1(&) = &
25
is an odd function.
SECTION 5.5 THE SUBSTITUTION RULE

253
53. Let & = 1 + 2r
3
, so d& = 6r
2
dr. When r = 0, & = 1; when r = 1, & = 3. Thus,
1
0
r
2
1 + 2r
3
5
dr =

3
1
&
5
1
6
d&
=
1
6
1
6
&
6
3
1
=
1
36
(3
6
31
6
) =
1
36
(729 31) =
728
36
=
182
9
.
55. Let & = t4, so d& =
1
4
dt. When t = 0, & = 0; when t = , & = 4. Thus,
r
0
sec
2
(t4) dt =

r'4
0
sec
2
&(4 d&) = 4
tan&
r'4
0
= 4
tan
r
4
3tan0
= 4(1 30) = 4.
57.

r'6
3r'6
tan
3
0 d0 = 0 by Theorem 7(b), since 1(0) = tan
3
0 is an odd function.
59. Let & = 1r, so d& = 31r
2
dr. When r = 1, & = 1; when r = 2, & =
1
2
. Thus,
2
1
c
1'i
r
2
dr =

1'2
1
c
u
(3d&) = 3
c
u
1'2
1
= 3(c
1'2
3c) = c 3
I
c.
61. Let & = 1 + 2r, so d& = 2 dr. When r = 0, & = 1; when r = 13, & = 27. Thus,
13
0
dr
3
(1 + 2r)
2
=

27
1
&
32'3
1
2
d&
=

1
2
3&
1'3
27
1
=
3
2
(3 31) = 3.
63. Let & = r
2
+o
2
, so d& = 2rdr and rdr =
1
2
d&. When r = 0, & = o
2
; when r = o, & = 2o
2
. Thus,
a
0
r
r
2
+o
2
dr =

2a
2
a
2
&
1'2
1
2
d&
=
1
2
2
3
&
3'2
2a
2
a
2
=

1
3
&
3'2
2a
2
a
2
=
1
3
(2o
2
)
3'2
3(o
2
)
3'2
=
1
3
2
I
2 31
o
3
65. Let & = r 31, so & + 1 = r and d& = dr. When r = 1, & = 0; when r = 2, & = 1. Thus,
2
1
r
I
r 31 dr =

1
0
(& + 1)
I
&d& =

1
0
(&
3'2
+&
1'2
) d& =

2
5
&
5'2
+
2
3
&
3'2
1
0
=
2
5
+
2
3
=
16
15
.
67. Let & = lnr, so d& =
dr
r
. When r = c, & = 1; when r = c
4
; & = 4. Thus,
c
4
c
dr
r
I
lnr
=

4
1
&
31'2
d& = 2
&
1'2
4
1
= 2(2 31) = 2.
69. Let & = c
:
+:, so d& = (c
:
+ 1) d:. When : = 0, & = 1; when : = 1, & = c + 1. Thus,
1
0
c
:
+ 1
c
:
+:
d: =

c+1
1
1
&
d& =

ln|&|
c+1
1
= ln|c + 1| 3ln|1| = ln(c + 1).
71. From the graph, it appears that the area under the curve is about
1 +
a little more than

1
2
1 0.7
, or about 1.4. The exact area is given by

=

1
0
I
2r + 1 dr. Let & = 2r +1, so d& = 2 dr. The limits change to
2 0 + 1 = 1 and 2 1 + 1 = 3, and
=

3
1
I
&
1
2
d&
=
1
2
2
3
&
3'2
3
1
=
1
3
3
I
3 31
=
I
3 3
1
3
E 1.399.
73. First write the integral as a sum of two integrals:
1 =

2
32
(r + 3)
I
4 3r
2
dr = 1
1
+1
2
=

2
32
r
I
4 3r
2
dr +
2
32
3
I
4 3r
2
dr. 1
1
= 0 by Theorem 7(b), since
1(r) = r
I
4 3r
2
is an odd function and we are integrating from r = 32 to r = 2. We interpret 1
2
as three times the area of
a semicircle with radius 2, so 1 = 0 + 3
1
2
2
2
= 6.
254

CHAPTER 5 INTEGRALS
75. First Figure Let & =
I
r, so r = &
2
and dr = 2&d&. When r = 0, & = 0; when r = 1, & = 1. Thus,
1
=

1
0
c
I
i
dr =

1
0
c
u
(2&d&) = 2
1
0
&c
u
d&.
Second Figure
2
=

1
0
2rc
i
dr = 2
1
0
&c
u
d&.
Third Figure Let & = sinr, so d& = cos rdr. When r = 0, & = 0; when r =
r
2
, & = 1. Thus,
3
=

r'2
0
c
sin i
sin2rdr =

r'2
0
c
sin i
(2 sinr cos r) dr =

1
0
c
u
(2&d&) = 2
1
0
&c
u
d&.
Since
1
=
2
=
3
, all three areas are equal.
77. The rate is measured in liters per minute. Integrating fromt = 0 minutes to t = 60 minutes will give us the total amount of oil
that leaks out (in liters) during the rst hour.
60
0
v(t) dt =

60
0
100c
30.01I
dt [& = 30.01t, d& = 30.01dt]
= 100
30.6
0
c
u
(3100 d&) = 310,000
c
u
30.6
0
= 310,000(c
30.6
31) r 4511.9 r 4512 liters
79. The volume of inhaled air in the lungs at time t is
\ (t) =

I
0
1(&) d& =

I
0
1
2
sin
2r
5
&
d& =

2rI'5
0
1
2
sin
5
2r
d

substitute =
2r
5
&, d =
2r
5
d&
=
5
4r
3cos
2rI'5
0
=
5
4r
3cos
2r
5
t
+ 1
=
5
4r
1 3cos
2r
5
t
liters
81. Let & = 2r. Then d& = 2 dr, so
2
0
1(2r) dr =

4
0
1(&)
1
2
d&
=
1
2
4
0
1(&) d& =
1
2
(10) = 5.
83. Let & = 3r. Then d& = 3dr, so
l
a
1(3r) dr =

3l
3a
1(&)(3d&) =
3a
3l
1(&) d& =

3a
3l
1(r) dr
From the diagram, we see that the equality follows from the fact that we are
reecting the graph of 1, and the limits of integration, about the j-axis.
85. Let & = 1 3r. Then r = 1 3& and dr = 3d&, so
1
0
r
a
(1 3r)
l
dr =

0
1
(1 3&)
a
&
l
(3d&) =
1
0
&
l
(1 3&)
a
d& =

1
0
r
l
(1 3r)
a
dr.
87.
rsinr
1 + cos
2
r
= r
sinr
2 3sin
2
r
= r1(sinr), where 1(t) =
t
2 3t
2
. By Exercise 86,
r
0
rsinr
1 + cos
2
r
dr =

r
0
r1(sinr) dr =

2
r
0
1(sinr) dr =

2
r
0
sinr
1 + cos
2
r
dr
Let & = cos r. Then d& = 3sinrdr. When r = , & = 31 and when r = 0, & = 1. So
r
0
sinr
1 + cos
2
r
dr = 3
31
1
d&
1 +&
2
=

2
1
31
d&
1 +&
2
=

2
tan
31
&
1
31
=

2
[tan
31
1 3tan
31
(31)] =

2
4
3
=

2
4
CHAPTER 5 REVIEW

255
5 Review
1. (a)

n
.=1
1(r
W
.
) {r is an expression for a Riemann sum of a function 1.
r
W
.
is a point in the jth subinterval [r
.31
. r
.
] and {r is the length of the subintervals.
(b) See Figure 1 in Section 5.2.
(c) In Section 5.2, see Figure 3 and the paragraph beside it.
2. (a) See Denition 5.2.2.
(b) See Figure 2 in Section 5.2.
(c) In Section 5.2, see Figure 4 and the paragraph by it (contains net area).
3. See the Fundamental Theorem of Calculus after Example 9 in Section 5.3.
4. (a) See the Net Change Theorem after Example 5 in Section 5.4.
(b)

I
2
I
1
v(t) dt represents the change in the amount of water in the reservoir between time t
1
and time t
2
.
5. (a)

120
60
(t) dt represents the change in position of the particle from t = 60 to t = 120 seconds.
(b)

120
60
|(t)| dt represents the total distance traveled by the particle from t = 60 to 120 seconds.
(c)

120
60
o(t) dt represents the change in the velocity of the particle from t = 60 to t = 120 seconds.
6. (a)

1(r) dr is the family of functions {1 | 1
0
= 1}. Any two such functions differ by a constant.
(b) The connection is given by the Net Change Theorem:

l
a
1(r) dr =

1(r) dr
l
a
if 1 is continuous.
7. The precise version of this statement is given by the Fundamental Theorem of Calculus. See the statement of this theorem and
the paragraph that follows it at the end of Section 5.3.
8. See the Substitution Rule (5.5.4). This says that it is permissible to operate with the dr after an integral sign as if it were a
differential.
1. True by Property 2 of the Integral in Section 5.2.
3. True by Property 3 of the Integral in Section 5.2.
5. False. For example, let 1(r) = r
2
. Then
1
0
I
r
2
dr =

1
0
rdr =
1
2
, but
1
0
r
2
dr =

1
3
=
1
I
3
.
7. True by Comparison Property 7 of the Integral in Section 5.2.
9. True. The integrand is an odd function that is continuous on [31. 1], so the result follows from Theorem 5.5.7(b).
256

CHAPTER 5 INTEGRALS
11. False. The function 1(r) = 1r
4
is not bounded on the interval [32. 1]. It has an innite discontinuity at r = 0, so it is
not integrable on the interval. (If the integral were to exist, a positive value would be expected, by Comparison
Property 6 of Integrals.)
13. False. For example, the function j = |r| is continuous on R, but has no derivative at r = 0.
15. False.

l
a
1(r) dr is a constant, so
d
dr
l
a
1(r) dr
= 0, not 1(r) [unless 1(r) = 0]. Compare the given statement

carefully with FTC1, in which the upper limit in the integral is r.
1. (a) 1
6
=
6
.=1
1(r
.31
) {r [{r =
6 30
6
= 1]
= 1(r
0
) 1 +1(r
1
) 1 +1(r
2
) 1 +1(r
3
) 1 +1(r
4
) 1 +1(r
5
) 1
E 2 + 3.5 + 4 + 2 + (31) + (32.5) = 8
The Riemann sum represents the sum of the areas of the four rectangles
above the r-axis minus the sum of the areas of the two rectangles below the
r-axis.
(b) A
6
=
6
.=1
1(r
.
) {r [{r =
6 30
6
= 1]
= 1(r
1
) 1 +1(r
2
) 1 +1(r
3
) 1 +1(r
4
) 1 +1(r
5
) 1 +1(r
6
) 1
= 1(0.5) +1(1.5) +1(2.5) +1(3.5) +1(4.5) +1(5.5)
E 3 + 3.9 + 3.4 + 0.3 + (32) + (32.9) = 5.7
3.

1
0
r +
I
1 3r
2
dr =

1
0
rdr +
1
0
I
1 3r
2
dr = 1
1
+1
2
.
1
1
can be interpreted as the area of the triangle shown in the gure
and 1
2
can be interpreted as the area of the quarter-circle.
Area =
1
2
(1)(1) +
1
4
()(1)
2
=
1
2
+
r
4
.
5.

6
0
1(r) dr =

4
0
1(r) dr +
6
4
1(r) dr i 10 = 7 +
6
4
1(r) dr i

6
4
1(r) dr = 10 37 = 3
7. First note that either o or / must be the graph of
i
0
1(t) dt, since
0
0
1(t) dt = 0, and c(0) 6= 0. Now notice that / 0 when c
is increasing, and that c 0 when o is increasing. It follows that c is the graph of 1(r), / is the graph of 1
0
(r), and o is the
graph of
i
0
1(t) dt.
9.

2
1
8r
3
+ 3r
2
dr =

8
1
4
r
4
+ 3
1
3
r
3
2
1
=

2r
4
+r
3
2
1
=

2 2
4
+ 2
3
3(2 + 1) = 40 33 = 37
11.

1
0
1 3r
9
dr =

r 3
1
10
r
10
1
0
=

1 3
1
10
30 =
9
10
CHAPTER 5 REVIEW

257
13.

9
1
I
& 32&
2
&
d& =

9
1
(&
31'2
32&) d& =

2&
1'2
3&
2
9
1
= (6 381) 3(2 31) = 376
15. Let & = j
2
+ 1, so d& = 2j dj and j dj =
1
2
d&. When j = 0, & = 1; when j = 1, & = 2. Thus,
1
0
j(j
2
+ 1)
5
dj =

2
1
&
5
1
2
d&
=
1
2
1
6
&
6
2
1
=
1
12
(64 31) =
63
12
=
21
4
.
17.

5
1
dt
(t 34)
2
does not exist because the function 1(t) =
1
(t 34)
2
has an innite discontinuity at t = 4;
that is, 1 is discontinuous on the interval [1. 5].
19. Let & =
3
, so d& = 3
2
d. When = 0, & = 0; when = 1, & = 1. Thus,
1
0

2
cos(
3
) d =

1
0
cos &
1
3
d&
=
1
3
sin&
1
0
=
1
3
(sin1 30) =
1
3
sin1.
21.

r'4
3r'4
t
4
tant
2 + cos t
dt = 0 by Theorem 5.5.7(b), since 1(t) =
t
4
tant
2 + cos t
is an odd function.
23.

1 3r
r
2
dr =

1
r
31
2
dr =

1
r
2
3
2
r
+ 1
dr = 3
1
r
32 ln|r| +r +C
25. Let & = r
2
+ 4r. Then d& = (2r + 4) dr = 2(r + 2) dr, so
r + 2
I
r
2
+ 4r
dr =

&
31'2
1
2
d&
=
1
2
2&
1'2
+C =
I
& +C =

r
2
+ 4r +C.
27. Let & = sint. Then d& = cos t dt, so
sint cos t dt =

&
1
r
d&
=
1
r

1
2
&
2
+C =
1
2r
(sint)
2
+C.
29. Let & =
I
r. Then d& =
dr
2
I
r
, so
c
I
i
I
r
dr = 2
c
u
d& = 2c
u
+C = 2c
I
i
+C.
31. Let & = ln(cos r). Then d& =
3sinr
cos r
dr = 3tanrdr, so
tanr ln(cos r) dr = 3
&d& = 3
1
2
&
2
+C = 3
1
2
[ln(cos r)]
2
+C.
33. Let & = 1 +r
4
. Then d& = 4r
3
dr, so
r
3
1 +r
4
dr =
1
4
1
&
d& =
1
4
ln|&| +C =
1
4
ln
1 +r
4
+C.
35. Let & = 1 + sec 0. Then d& = sec 0 tan0 d0, so
sec 0 tan0
1 + sec 0
d0 =

1
1 + sec 0
(sec 0 tan0 d0) =

1
&
d& = ln|&| +C = ln|1 + sec 0| +C.
37. Since r
2
34 < 0 for 0 $ r < 2 and r
2
34 0 for 2 < r $ 3, we have
r
2
34
= 3(r
2
34) = 4 3r
2
for 0 $ r < 2 and
r
2
34
= r
2
34 for 2 < r $ 3. Thus,
3
0
r
2
34
dr =

2
0
(4 3r
2
) dr +
3
2
(r
2
34) dr =

4r 3
r
3
3
2
0
+
r
3
3
34r
3
2
=

8 3
8
3
30 + (9 312) 3
8
3
38
=
16
3
33 +
16
3
=
32
3
3
9
3
=
23
3
258

CHAPTER 5 INTEGRALS
In Exercises 39 and 40, let 1(r) denote the integrand and 1(r) its antiderivative (with C = 0).
39. Let & = 1 + sinr. Then d& = cos rdr, so
cos rdr
I
1 + sinr
=

&
31'2
d& = 2&
1'2
+C = 2
I
1 + sinr +C.
41. From the graph, it appears that the area under the curve j = r
I
r between r = 0
and r = 4 is somewhat less than half the area of an 8 4 rectangle, so perhaps
about 13 or 14. To nd the exact value, we evaluate
4
0
r
I
rdr =

4
0
r
3'2
dr =

2
5
r
5'2
4
0
=
2
5
(4)
5'2
=
64
5
= 12.8.
43. 1(r) =
i
0
t
2
1 +t
3
dt i 1
0
(r) =
d
dr
i
0
t
2
1 +t
3
dt =
r
2
1 +r
3
45. Let & = r
4
. Then
d&
dr
= 4r
3
. Also,
do
dr
=
do
d&
d&
dr
, so
o
0
(r) =
d
dr
i
4
0
cos(t
2
) dt =
d
d&
u
0
cos(t
2
) dt
d&
dr
= cos(&
2
)
d&
dr
= 4r
3
cos(r
8
).
47. j =

i
I
i
c
I
t
dt =

1
I
i
c
I
t
dt +
i
1
c
I
t
dt = 3
I
i
1
c
I
t
dt +
i
1
c
I
t
dt i
dj
dr
= 3
d
dr
I
i
1
c
I
t
dt
+
d
dr
i
1
c
I
t
dt
. Let & =
I
r. Then
d
dr
I
i
1
c
I
t
dt =
d
dr
u
1
c
I
t
dt =
d
d&
u
1
c
I
t
dt
d&
dr
=
c
u
&

1
2
I
r
=
c
I
i
I
r

1
2
I
r
=
c
I
i
2r
,
so
dj
dr
= 3
c
I
i
2r
+
c
i
r
.
49. If 1 $ r $ 3, then
I
1
2
+ 3 $
I
r
2
+ 3 $
I
3
2
+ 3 i 2 $
I
r
2
+ 3 $ 2
I
3, so
2(3 31) $
3
1
I
r
2
+ 3 dr $ 2
I
3(3 31); that is, 4 $
3
1
I
r
2
+ 3 dr $ 4
I
3.
51. 0 $ r $ 1 i 0 $ cos r $ 1 i r
2
cos r $ r
2
i

1
0
r
2
cos rdr $
1
0
r
2
dr =
1
3
r
3
1
0
=
1
3
[Property 7].
53. cos r $ 1 i c
i
cos r $ c
i
i

1
0
c
i
cos rdr $
1
0
c
i
dr = [c
i
]
1
0
= c 31
55. {r = (3 30)6 =
1
2
, so the endpoints are 0,
1
2
, 1,
3
2
, 2,
5
2
, and 3, and the midpoints are
1
4
,
3
4
,
5
4
,
7
4
,
9
4
, and
11
4
.
The Midpoint Rule gives
3
0
sin(r
3
) dr r
6
.=1
1(r
.
) {r =
1
2
sin
1
4
3
+ sin
3
4
3
+ sin
5
4
3
+ sin
7
4
3
+ sin
9
4
3
+ sin
11
4
r 0.280981.
CHAPTER 5 REVIEW

259
57. Note that v(t) = /
0
(t), where /(t) = the number of barrels of oil consumed up to time t. So, by the Net Change Theorem,
8
0
v(t) dt = /(8) 3/(0) represents the number of barrels of oil consumed from Jan. 1, 2000, through Jan. 1, 2008.
59. We use the Midpoint Rule with n = 6 and {t =
24 30
6
= 4. The increase in the bee population was
24
0
v(t) dt E A
6
= 4[v(2) +v(6) +v(10) +v(14) +v(18) +v(22)]
E 4[50 + 1000 + 7000 + 8550 + 1350 + 150] = 4(18,100) = 72,400
61. Let & = 2 sin0. Then d& = 2 cos 0 d0 and when 0 = 0, & = 0; when 0 =
r
2
, & = 2. Thus,
r'2
0
1(2 sin0) cos 0 d0 =

2
0
1(&)
1
2
d&
=
1
2
2
0
1(&) d& =
1
2
2
0
1(r) dr =
1
2
(6) = 3.
63. Area under the curve j = sinhcr between r = 0 and r = 1 is equal to 1 i
1
0
sinhcrdr = 1 i
1
c
coshcr
1
0
= 1 i
1
c
(coshc 31) = 1 i
coshc 31 = c i coshc = c + 1. From the graph, we get c = 0 and
c E 1.6161, but c = 0 isnt a solution for this problem since the curve
j = sinhcr becomes j = 0 and the area under it is 0. Thus, c E 1.6161.
65. Using FTC1, we differentiate both sides of the given equation,
i
0
1(t) dt = rc
2i
+
i
0
c
3I
1(t) dt, and get
1(r) = c
2i
+ 2rc
2i
+c
3i
1(r) i 1(r)
1 3c
3i
= c
2i
+ 2rc
2i
i 1(r) =
c
2i
(1 + 2r)
1 3c
3i
.
67. Let & = 1(r) and d& = 1
0
(r) dr. So 2
l
a
1(r)1
0
(r) dr = 2
](l)
](a)
&d& =

&
2
](l)
](a)
= [1(/)]
2
3[1(o)]
2
.
69. Let & = 1 3r. Then d& = 3dr, so
1
0
1(1 3r) dr =

0
1
1(&)(3d&) =
1
0
1(&) d& =

1
0
1(r) dr.
71. The shaded region has area
1
0
1(r) dr =
1
3
. The integral
1
0
1
31
(j) dj
gives the area of the unshaded region, which we know to be 1 3
1
3
=
2
3
.
So
1
0
1
31
(j) dj =
2
3
.
PROBLEMS PLUS
1. Differentiating both sides of the equation rsinr =

i
2
0
1(t) dt (using FTC1 and the Chain Rule for the right side) gives
sinr +rcos r = 2r1(r
2
). Letting r = 2 so that 1(r
2
) = 1(4), we obtain sin2 + 2 cos 2 = 41(4), so
1(4) =
1
4
(0 + 2 1) =
r
2
.
3. Differentiating the given equation,
i
0
1(t) dt = [1(r)]
2
, using FTC1 gives 1(r) = 21(r) 1
0
(r) i
1(r)[21
0
(r) 31] = 0, so 1(r) = 0 or 1
0
(r) =
1
2
. Since 1(r) is never 0, we must have 1
0
(r) =
1
2
and 1
0
(r) =
1
2
i
1(r) =
1
2
r +C. To nd C, we substitute into the given equation to get
i
0
1
2
t +C
dt =

1
2
r +C
2
C
1
4
r
2
+Cr =
1
4
r
2
+Cr +C
2
. It follows that C
2
= 0, so C = 0, and 1(r) =
1
2
r.
5. 1(r) =

g(i)
0
1
I
1 +t
3
dt, where o(r) =

cos i
0
[1 + sin(t
2
)] dt. Using FTC1 and the Chain Rule (twice) we have
1
0
(r) =
1
1 + [o(r)]
3
o
0
(r) =
1
1 + [o(r)]
3
[1 + sin(cos
2
r)](3sinr). Now o
r
2
=

0
0
[1 + sin(t
2
)] dt = 0, so
1
0
r
2
=
1
I
1 + 0
(1 + sin0)(31) = 1 1 (31) = 31.
7. By lHospitals Rule and the Fundamental Theorem, using the notation exp(j) = c
u
,
lim
i<0
i
0
(1 3tan2t)
1'I
dt
r
H
= lim
i<0
(1 3tan2r)
1'i
1
= exp
lim
i<0
ln(1 3tan2r)
r
H
= exp
lim
i<0
32 sec
2
2r
1 3tan2r
= exp
32 1
2
1 30
= c
32
9. 1(r) = 2 +r 3r
2
= (3r + 2)(r + 1) = 0 C r = 2 or r = 31. 1(r) D 0 for r M [31. 2] and 1(r) < 0 everywhere
else. The integral
l
a
(2 +r 3r
2
) dr has a maximum on the interval where the integrand is positive, which is [31. 2]. So
o = 31, / = 2. (Any larger interval gives a smaller integral since 1(r) < 0 outside [31. 2]. Any smaller interval also gives a
smaller integral since 1(r) D 0 in [31. 2].)
11. (a) We can split the integral
n
0
[[r]] dr into the sum
n
.=1
.
.31
[[r]] dr
. But on each of the intervals [j 31. j) of integration,

[[r]] is a constant function, namely j 31. So the ith integral in the sum is equal to (j 31)[j 3(j 31)] = (j 31). So the
original integral is equal to
n
.=1
(j 31) =
n31
.=1
j =
(n 31)n
2
.
261
262

PROBLEMS PLUS
(b) We can write
l
a
[[r]] dr =

l
0
[[r]] dr 3
a
0
[[r]] dr.
Now
l
0
[[r]] dr =

[[l]]
0
[[r]] dr +
l
[[l]]
[[r]] dr. The rst of these integrals is equal to
1
2
([[/]] 31) [[/]],
by part (a), and since [[r]] = [[/]] on [[[/]] . /], the second integral is just [[/]] (/ 3[[/]]). So
l
0
[[r]] dr =
1
2
([[/]] 31) [[/]] + [[/]] (/ 3[[/]]) =
1
2
[[/]] (2/ 3[[/]] 31) and similarly
a
0
[[r]] dr =
1
2
[[o]] (2o 3[[o]] 31).
Therefore,
l
a
[[r]] dr =
1
2
[[/]] (2/ 3[[/]] 31) 3
1
2
[[o]] (2o 3[[o]] 31).
13. Let Q(r) =

i
0
1(t) dt =

ot +
/
2
t
2
+
c
3
t
3
+
d
4
t
4
i
0
= or +
/
2
r
2
+
c
3
r
3
+
d
4
r
4
. Then Q(0) = 0, and Q(1) = 0 by the
given condition, o +
/
2
+
c
3
+
d
4
= 0. Also, Q
0
(r) = 1(r) = o +/r +cr
2
+dr
3
by FTC1. By Rolles Theorem, applied to
Q on [0. 1], there is a number v in (0. 1) such that Q
0
(v) = 0, that is, such that 1(v) = 0. Thus, the equation 1(r) = 0 has a
root between 0 and 1.
More generally, if 1 (r) = o
0
+o
1
r +o
2
r
2
+ +o
n
r
n
and if o
0
+
o
1
2
+
o
2
3
+ +
o
n
n + 1
= 0. then the equation
1(r) = 0 has a root between 0 and 1. The proof is the same as before:
Let Q(r) =

i
0
1(t) dt = o
0
r +
o
1
2
r
2
+
o
2
3
r
3
+ +
o
n
n + 1
r
n
. Then Q(0) = Q(1) = 0 and Q
0
(r) = 1(r). By
Rolles Theorem applied to Qon [0. 1], there is a number v in (0. 1) such that Q
0
(v) = 0, that is, such that 1(v) = 0.
15. Note that
d
dr
i
0
u
0
1(t) dt
d&
=

i
0
1(t) dt by FTC1, while
d
dr
i
0
1(&)(r 3&) d&
=
d
dr
i
0
1(&) d&
3
d
dr
i
0
1(&)&d&
=

i
0
1(&) d& +r1(r) 31(r)r =

i
0
1(&) d&
Hence,
i
0
1(&)(r 3&) d& =

i
0
u
0
1(t) dt
d& +C. Setting r = 0 gives C = 0.

17. lim
n<"
1
I
n
I
n + 1
+
1
I
n
I
n + 2
+ +
1
I
n
I
n +n
= lim
n<"
1
n
n
n + 1
+
n
n + 2
+ +
n
n +n
= lim
n<"
1
n
1 + 1n
+
1
1 + 2n
+ +
1
I
1 + 1
= lim
n<"
1
n
n
.=1
1
j
n

where 1(r) =
1
I
1 +r
=

1
0
1
I
1 +r
dr =

2
I
1 +r
1
0
= 2
I
2 31
6 APPLICATIONS OF INTEGRATION
6.1 Areas Between Curves
1. =

i=4
i=0
(j
T
3j
T
) dr =

4
0
(5r 3r
2
) 3r
dr =

4
0
(4r 3r
2
) dr =

2r
2
3
1
3
r
3
4
0
=

32 3
64
3
3(0) =
32
3
3. =

u=1
u=31
(r
T
3r
1
) dj =

1
31
c
u
3(j
2
32)
dj =

1
31
c
u
3j
2
+ 2
dj
=

c
u
3
1
3
j
3
+ 2j
1
31
=

c
1
3
1
3
+ 2
c
31
+
1
3
32
= c 3
1
c
+
10
3
5. =

2
31
(9 3r
2
) 3(r + 1)
dr
=

2
31
(8 3r 3r
2
) dr
=

8r 3
r
2
2
3
r
3
3
2
31
=

16 32 3
8
3
38 3
1
2
+
1
3
= 22 33 +
1
2
=
39
2
7. The curves intersect when r = r
2
C r
2
3r = 0 C
r(r 31) = 0 C r = 0 or 1.
=

1
0
(r 3r
2
) dr =

1
2
r
2
3
1
3
r
3
1
0
=
1
2
3
1
3
=
1
6
9. =

2
1
1
r
3
1
r
2
dr =

lnr +
1
r
2
1
=

ln2 +
1
2
3(ln1 + 1)
= ln2 3
1
2
E 0.19
11. =

1
0
I
r 3r
2
dr
=

2
3
r
3'2
3
1
3
r
3
1
0
=
2
3
3
1
3
=
1
3
263
264

CHAPTER 6 APPLICATIONS OF INTEGRATION
13. 12 3r
2
= r
2
36 C 2r
2
= 18 C
r
2
= 9 C r = 3, so
=

3
33
(12 3r
2
) 3(r
2
36)
dr
= 2
3
0
18 32r
2
dr [by symmetry]
= 2
18r 3
2
3
r
3
3
0
= 2 [(54 318) 30]
= 2(36) = 72
15. The curves intersect when tanr = 2 sinr (on [33. 3]) C sinr = 2 sinr cos r C
2 sinr cos r 3sinr = 0 C sinr(2 cos r 31) = 0 C sinr = 0 or cos r =
1
2
C r = 0 or r =
r
3
.
=

r'3
3r'3
(2 sinr 3tanr) dr
= 2
r'3
0
(2 sinr 3tanr) dr [by symmetry]
= 2
32 cos r 3ln|sec r|
r'3
0
= 2 [(31 3ln2) 3(32 30)]
= 2(1 3ln2) = 2 32 ln2
17.
1
2
r =
I
r i
1
4
r
2
= r i r
2
34r = 0 i r(r 34) = 0 i r = 0 or 4, so
=

4
0
I
r 3
1
2
r
dr +
9
4
1
2
r 3
I
r
dr =

2
3
r
3'2
3
1
4
r
2
4
0
+
1
4
r
2
3
2
3
r
3'2
9
4
=

16
3
34
30
81
4
318
4 3
16
3
=
81
4
+
32
3
326 =
59
12
SECTION 6.1 AREAS BETWEEN CURVES

265
19. 2j
2
= 4 +j
2
C j
2
= 4 C j = 2, so
=

2
32
(4 +j
2
) 32j
2
dj
= 2
2
0
(4 3j
2
) dj [by symmetry]
= 2
4j 3
1
3
j
3
2
0
= 2
8 3
8
3
=
32
3
21. The curves intersect when 1 3j
2
= j
2
31 C 2 = 2j
2
C j
2
= 1 C j = 1.
=

1
31
(1 3j
2
) 3(j
2
31)
dj
=

1
31
2(1 3j
2
) dj
= 2 2
1
0
(1 3j
2
) dj
= 4
j 3
1
3
j
3
1
0
= 4
1 3
1
3
=
8
3
23. Notice that cos r = sin2r = 2 sinrcos r C
2 sinrcos r 3cos r = 0 C cos r(2 sinr 31) = 0 C
2 sinr = 1 or cos r = 0 C r =
r
6
or
r
2
.
=

r'6
0
(cos r 3sin2r) dr +
r'2
r'6
(sin2r 3cos r) dr
=

sinr +
1
2
cos 2r
r'6
0
+
3
1
2
cos 2r 3sinr
r'2
r'6
=
1
2
+
1
2

1
2
3
0 +
1
2
1
1
2
31
3
1
2

1
2
3
1
2
=
1
2
25. The curves intersect when r
2
=
2
r
2
+ 1
C
r
4
+r
2
= 2 C r
4
+r
2
32 = 0 C
(r
2
+ 2)(r
2
31) = 0 C r
2
= 1 C r = 1.
=

1
31
2
r
2
+ 1
3r
2
dr = 2
1
0
2
r
2
+ 1
3r
2
dr
= 2
2 tan
31
r 3
1
3
r
3
1
0
= 2
2
r
4
3
1
3
= 3
2
3
E 2.47
266

27. 1r = r C 1 = r
2
C r = 1 and 1r =
1
4
r C
4 = r
2
C r = 2, so for r 0,
=

1
0
r 3
1
4
r
dr +
2
1
1
r
3
1
4
r
dr
=

1
0
3
4
r
dr +
2
1
1
r
3
1
4
r
dr
=

3
8
r
2
1
0
+
ln|r| 3
1
8
r
2
2
1
=
3
8
+
ln2 3
1
2
0 3
1
8
= ln2
29. An equation of the line through (0. 0) and (2. 1) is j =
1
2
r; through (0. 0)
and (31. 6) is j = 36r; through (2. 1) and (31. 6) is j = 3
5
3
r +
13
3
.
=

0
31
3
5
3
r +
13
3
3(36r)
dr +
2
0
3
5
3
r +
13
3
3
1
2
r
dr
=

0
31
13
3
r +
13
3
dr +
2
0
3
13
6
r +
13
3
dr
=
13
3
0
31
(r + 1) dr +
13
3
2
0
3
1
2
r + 1
dr
=
13
3
1
2
r
2
+r
0
31
+
13
3
3
1
4
r
2
+r
2
0
=
13
3
0 3
1
2
31
+
13
3
[(31 + 2) 30] =
13
3

1
2
+
13
3
1 =
13
2
31. The curves intersect when sinr = cos 2r (on [0. 2]) C sinr = 1 32 sin
2
r C 2 sin
2
r + sinr 31 = 0 C
(2 sinr 31)(sinr + 1) = 0 i sinr =
1
2
i r =
r
6
.
=

r'2
0
|sinr 3cos 2r| dr
=

r'6
0
(cos 2r 3sinr) dr +
r'2
r'6
(sinr 3cos 2r) dr
=

1
2
sin2r + cos r
r'6
0
+
3cos r 3
1
2
sin2r
r'2
r'6
=

1
4
I
3 +
1
2
I
3
3(0 + 1) + (0 30) 3
3
1
2
I
3 3
1
4
I
3
=
3
2
I
3 31
33. Let 1(r) = cos
2
r
4
3sin
2
r
4
and {r =
1 30
4
.
The shaded area is given by
=

1
0
1(r) dr E A
4
=
1
4
1
8
+1
3
8
+1
5
8
+1
7
8
E 0.6407

267
35. From the graph, we see that the curves intersect at r = 0 and r = o E 0.896, with
rsin(r
2
) r
4
on (0. o). So the area of the region bounded by the curves is
=

a
0
rsin(r
2
) 3r
4
dr =

3
1
2
cos(r
2
) 3
1
5
r
5
a
0
= 3
1
2
cos(o
2
) 3
1
5
o
5
+
1
2
E 0.037
37. From the graph, we see that the curves intersect at
r = o r 31.11. r = / r 1.25. and r = c r 2.86. with
r
3
33r + 4 3r
2
32r on (o. /) and 3r
2
32r r
3
33r + 4
on (/. c). So the area of the region bounded by the curves is
=

l
a
(r
3
33r + 4) 3(3r
2
32r)
dr +
c
l
(3r
2
32r) 3(r
3
33r + 4)
dr
=

l
a
(r
3
33r
2
3r + 4) dr +
c
l
(3r
3
+ 3r
2
+r 34) dr
=

1
4
r
4
3r
3
3
1
2
r
2
+ 4r
l
a
+
3
1
4
r
4
+r
3
+
1
2
r
2
34r
c
l
r 8.38
39. As the gure illustrates, the curves j = r and j = r
5
36r
3
+ 4r
enclose a four-part region symmetric about the origin (since
r
5
36r
3
+ 4r and r are odd functions of r). The curves intersect
at values of r where r
5
36r
3
+ 4r = r; that is, where
r(r
4
36r
2
+ 3) = 0. That happens at r = 0 and where
r
2
=
6
I
36 312
2
= 3
I
6; that is, at r = 3
3 +
I
6, 3
3 3
I
6, 0,
3 3
I
6, and
3 +
I
6.
The exact area is
2
I
3+
I
6
0
(r
5
36r
3
+ 4r) 3r
dr = 2
I
3+
I
6
0
r
5
36r
3
+ 3r
dr
= 2
I
33
I
6
0
(r
5
36r
3
+ 3r) dr + 2
I
3+
I
6
I
33
I
6
(3r
5
+ 6r
3
33r) dr
CAS
= 12
I
6 39
268

41. 1 second =
1
3600
hour, so 10 s =
1
360
h. With the given data, we can take n = 5 to use the Midpoint Rule.
{t =
1'36030
5
=
1
1800
, so
distance
Kelly
3distance
Chris
=

1'360
0

1
dt 3
1'360
0

C
dt =

1'360
0
(
1
3
C
) dt
E A
5
=
1
1800
[(
1
3
C
)(1) + (
1
3
C
)(3) + (
1
3
C
)(5)
+(
1
3
C
)(7) + (
1
3
C
)(9)]
=
1
1800
[(22 320) + (52 346) + (71 362) + (86 375) + (98 386)]
=
1
1800
(2 + 6 + 9 + 11 + 12) =
1
1800
(40) =
1
45
mile, or 117
1
3
feet
43. Let /(r) denote the height of the wing at r cm from the left end.
r A
5
=
200 30
5
[/(20) +/(60) +/(100) +/(140) +/(180)]
= 40(20.3 + 29.0 + 27.3 + 20.5 + 8.7) = 40(105.8) = 4232 cm
2
45. We know that the area under curve between t = 0 and t = r is
i
0

/
(t) dt = c
/
(r), where
/
(t) is the velocity of car A
and c
A
is its displacement. Similarly, the area under curve 1 between t = 0 and t = r is
i
0

B
(t) dt = c
B
(r).
(a) After one minute, the area under curve is greater than the area under curve 1. So car A is ahead after one minute.
(b) The area of the shaded region has numerical value c
A
(1) 3c
B
(1), which is the distance by which A is ahead of B after
1 minute.
(c) After two minutes, car B is traveling faster than car A and has gained some ground, but the area under curve from t = 0
to t = 2 is still greater than the corresponding area for curve 1, so car A is still ahead.
(d) From the graph, it appears that the area between curves and 1 for 0 $ t $ 1 (when car A is going faster), which
corresponds to the distance by which car A is ahead, seems to be about 3 squares. Therefore, the cars will be side by side
at the time r where the area between the curves for 1 $ t $ r (when car B is going faster) is the same as the area for
0 $ t $ 1. From the graph, it appears that this time is r E 2.2. So the cars are side by side when t E 2.2 minutes.
47. To graph this function, we must rst express it as a combination of explicit
functions of j; namely, j = r
I
r + 3. We can see from the graph that the loop
extends from r = 33 to r = 0, and that by symmetry, the area we seek is just
twice the area under the top half of the curve on this interval, the equation of the
top half being j = 3r
I
r + 3. So the area is = 2
0
33
3r
I
r + 3
dr. We
substitute & = r + 3, so d& = dr and the limits change to 0 and 3, and we get
= 32
3
0
[(& 33)
I
&] d& = 32
3
0
(&
3'2
33&
1'2
) d&
= 32
2
5
&
5'2
32&
3'2
3
0
= 32
2
5
3
2
I
3
32
3
I
3
=
24
5
I
3

269
49. By the symmetry of the problem, we consider only the rst quadrant, where
j = r
2
i r =

j. We are looking for a number / such that
l
0
j dj =

4
l
j dj i
2
3
j
3'2
l
0
=
2
3
j
3'2
4
l
i
/
3'2
= 4
3'2
3/
3'2
i 2/
3'2
= 8 i /
3'2
= 4 i / = 4
2'3
E 2.52.
51. We rst assume that c 0, since c can be replaced by 3c in both equations without changing the graphs, and if c = 0 the
curves do not enclose a region. We see from the graph that the enclosed area lies between r = 3c and r = c, and by
symmetry, it is equal to four times the area in the rst quadrant. The enclosed area is
= 4
c
0
(c
2
3r
2
) dr = 4
c
2
r 3
1
3
r
3
c
0
= 4
c
3
3
1
3
c
3
= 4
2
3
c
3
=
8
3
c
3
So = 576 C
8
3
c
3
= 576 C c
3
= 216 C c =
3
I
216 = 6.
Note that c = 36 is another solution, since the graphs are the same.
53. The curve and the line will determine a region when they intersect at two or
more points. So we solve the equation r(r
2
+ 1) = ir i
r = r(ir
2
+i) i r(ir
2
+i) 3r = 0 i
r(ir
2
+i31) = 0 i r = 0 or ir
2
+i31 = 0 i
r = 0 or r
2
=
1 3i
i
i r = 0 or r =
1
i
31. Note that if i = 1, this has only the solution r = 0, and no region
is determined. But if 1i31 0 C 1i 1 C 0 < i < 1, then there are two solutions. [Another way of seeing
this is to observe that the slope of the tangent to j = r(r
2
+ 1) at the origin is j
0
(0) = 1 and therefore we must have
0 < i < 1.] Note that we cannot just integrate between the positive and negative roots, since the curve and the line cross at
the origin. Since ir and r(r
2
+ 1) are both odd functions, the total area is twice the area between the curves on the interval
0.
1i31
. So the total area enclosed is

2
I
1'r31
0
r
r
2
+ 1
3ir
dr = 2
1
2
ln(r
2
+ 1) 3
1
2
ir
2
I
1'r31
0
= [ln(1i31 + 1) 3i(1i31)] 3(ln1 30)
= ln(1i) 31 +i = i3lni31
270

6.2 Volumes
1. A cross-section is a disk with radius 2 3
1
2
r. so its area is (r) =
2 3
1
2
r
2
.
\ =

2
1
(r) dr =

2
1

2 3
1
2
r
2
dr
=
2
1
4 32r +
1
4
r
2
dr
=
4r 3r
2
+
1
12
r
3
2
1
=
8 34 +
8
12
4 31 +
1
12
1 +
7
12
=
19
12
3. A cross-section is a disk with radius 1r, so its area is

(r) = (1r)
2
.
\ =

2
1
(r) dr =

2
1

1
r
2
dr
=
2
1
1
r
2
dr =
3
1
r
2
1
=
3
1
2
3(31)
=
r
2
5. A cross-section is a disk with radius 2
j, so its area is
(j) =
2
.
\ =

9
0
(j) dj =

9
0

2
dj = 4
9
0
j dj
= 4
1
2
j
2
9
0
= 2(81) = 162
7. A cross-section is a washer (annulus) with inner
radius r
3
and outer radius r, so its area is
(r) = (r)
2
3(r
3
)
2
= (r
2
3r
6
).
\ =

1
0
(r) dr =

1
0
(r
2
3r
6
) dr
=
1
3
r
3
3
1
7
r
7
1
0
=
1
3
3
1
7
=
4
21
SECTION 6.2 VOLUMES

271
9. A cross-section is a washer with inner radius j
2
and outer radius 2j, so its area is
(j) = (2j)
2
3(j
2
)
2
= (4j
2
3j
4
).
\ =

2
0
(j) dj =
2
0
(4j
2
3j
4
) dj
=
4
3
j
3
3
1
5
j
5
2
0
=
32
3
3
32
5
=
64
15
11. A cross-section is a washer with inner radius 1 3

I
r and outer radius 1 3r, so its area is
(r) = (1 3r)
2
3
1 3
I
r
2
=
(1 32r +r
2
) 3
1 32
I
r +r
33r +r
2
+ 2
I
r
.
\ =

1
0
(r) dr =
1
0
33r +r
2
+ 2
I
r
dr
=
3
3
2
r
2
+
1
3
r
3
+
4
3
r
3'2
1
0
=
3
3
2
+
5
3
=
r
6
13. A cross-section is a washer with inner radius (1 + sec r) 31 = sec r and outer radius 3 31 = 2, so its area is
(r) =
2
2
3(sec r)
2
= (4 3sec
2
r).
\ =

r'3
3r'3
(r) dr =

r'3
3r'3
(4 3sec
2
r) dr
= 2
r'3
0
(4 3sec
2
r) dr [by symmetry]
= 2
4r 3tanr
r'3
0
= 2
4r
3
3
I
3
30
= 2
4r
3
3
I
3
15. \ =

1
31
(1 3j
2
)
2
dj = 2
1
0
(1 3j
2
)
2
dj
= 2
1
0
(1 32j
2
+j
4
) dj
= 2
j 3
2
3
j
3
+
1
5
j
5
1
0
= 2
8
15
=
16
15
272

17. j = r
2
i r =
I
j for r D 0. The outer radius is the distance fromr = 31 to r =

j and the inner radius is the
distance from r = 31 to r = j
2
.
\ =

1
0

j 3(31)
2
3
j
2
3(31)
dj =
1
0
j + 1
2
3(j
2
+ 1)
2
dj
=
1
0
j + 2
j + 1 3j
4
32j
2
31
dj =
1
0
j + 2
j 3j
4
32j
2
dj
=
1
2
j
2
+
4
3
j
3'2
3
1
5
j
5
3
2
3
j
3
1
0
=
1
2
+
4
3
3
1
5
3
2
3
=
29
30
19. R
1
about O (the line j = 0): \ =

1
0
(r) dr =

1
0
(r
3
)
2
dr =
1
0
r
6
dr =
1
7
r
7
1
0
=

7
21. R
1
about 1 (the line r = 1):
\ =

1
0
(j) dj =

1
0

1 3
3
2
dj =
1
0
(1 32j
1'3
+j
2'3
) dj =
j 3
3
2
j
4'3
+
3
5
j
5'3
1
0
=
1 3
3
2
+
3
5
=
r
10
23. R
2
about O (the line j = 0):
\ =

1
0
(r) dr =

1
0
(1)
2
3
I
r
dr =
1
0
(1 3r) dr =
r 3
1
2
r
2
1
0
=
1 3
1
2
=
r
2
25. R
2
\ =

1
0
(j) dj =

1
0
(1)
2
3(1 3j
2
)
2
dj =
1
0
1 3(1 32j
2
+j
4
)
dj =
1
0
(2j
2
3j
4
) dj
=
2
3
j
3
3
1
5
j
5
1
0
=
2
3
3
1
5
=
7
15
27. R
3
about O (the line j = 0):
\ =

1
0
(r) dr =

1
0
I
r
2
3(r
3
)
2
dr =
1
0
(r 3r
6
) dr =
1
2
r
2
3
1
7
r
7
1
0
=
1
2
3
1
7
=
5
14
.
Note: Let R= R
1
+ R
2
+ R
3
. If we rotate Rabout any of the segments O, OC, 1, or 1C, we obtain a right circular
cylinder of height 1 and radius 1. Its volume is v
2
/ = (1)
2
1 = . As a check for Exercises 19, 23, and 27, we can add the
answers, and that sum must equal . Thus,
r
7
+
r
2
+
5r
14
=

2 +7 +5
14
= .
SECTION 6.2 VOLUMES

273
29. R
3
\ =

1
0
(j) dj =

1
0
(1 3j
2
)
2
3
1 3
3
dj =
1
0
(1 32j
2
+j
4
) 3(1 32j
1'3
+j
2'3
)
dj
=
1
0
(32j
2
+j
4
+ 2j
1'3
3j
2'3
) dj =
3
2
3
j
3
+
1
5
j
5
+
3
2
j
4'3
3
3
5
j
5'3
1
0
=
3
2
3
+
1
5
+
3
2
3
3
5
=
13
30
Note: See the note in Exercise 27. For Exercises 21, 25, and 29, we have
r
10
+
7r
15
+
13r
30
=

3 +14 +13
30
= .
31. \ =
r'4
0
(1 3tan
3
r)
2
dr
33. \ =
r
0
(1 30)
2
3(1 3sinr)
2
dr
=
r
0
1
2
3(1 3sinr)
2
dr
35. \ =
I
8
3
I
8
[3 3(32)]
2
3
j
2
+ 1 3(32)
dj
=
2
I
2
32
I
2
5
2
3
1 +j
2
+ 2
dj
37. j = 2 +r
2
cos r and j = r
4
+r + 1 intersect at
r = o r 31.288 and r = / r 0.884.
\ =
l
a
[(2 +r
2
cos r)
2
3(r
4
+r + 1)
2
]dr r 23.780
274

39. \ =
r
0
sin
2
r 3(31)
2
3[0 3(31)]
2
dr
CAS
=
11
8

2
41.
r'2
0
cos
2
rdr describes the volume of the solid obtained by rotating the region R=

(r. j) | 0 $ r $
r
2
. 0 $ j $ cos r
of the rj-plane about the r-axis.

43.
1
0
(j
4
3j
8
) dj =
1
0
(j
2
)
2
3(j
4
)
2
dj describes the volume of the solid obtained by rotating the region

R=

(r. j) | 0 $ j $ 1. j
4
$ r $ j
2
of the rj-plane about the j-axis.

45. There are 10 subintervals over the 15-cm length, so well use n = 102 = 5 for the Midpoint Rule.
\ =

15
0
(r) dr E A
5
=
1530
5
[(1.5) +(4.5) +(7.5) +(10.5) +(13.5)]
= 3(18 + 79 + 106 + 128 + 39) = 3 370 = 1110 cm
3
47. (a) \ =

10
2
[1(r)]
2
dr r
10 32
4
[1(3)]
2
+ [1(5)]
2
+ [1(7)]
2
+ [1(9)]
2
r 2
(1.5)
2
+ (2.2)
2
+ (3.8)
2
+ (3.1)
2
r 196 units
3
(b) \ =

4
0

(outer radius)
2
3(inner radius)
2
dj
r
4 30
4
(9.9)
2
3(2.2)
2
(9.7)
2
3(3.0)
2
(9.3)
2
3(5.6)
2
(8.7)
2
3(6.5)
2
r 838 units
3
49. Well form a right circular cone with height / and base radius v by
revolving the line j =
r
I
r about the r-axis.
\ =
I
0
v
/
r
2
dr =
I
0
v
2
/
2
r
2
dr =
v
2
/
2
1
3
r
3
I
0
=
v
2
/
2
1
3
/
3
=
1
3
v
2
/
Another solution: Revolve r = 3
v
/
j +v about the j-axis.
\ =
I
0
3
v
/
j +v
2
dj
W
=
I
0
v
2
/
2
j
2
3
2v
2
/
j +v
2
dj
=
v
2
3/
2
j
3
3
v
2
/
j
2
+v
2
j
I
0
=
1
3
v
2
/ 3v
2
/ +v
2
/
=
1
3
v
2
/
W
Or use substitution with & = v 3
v
/
j and d& = 3
v
/
dj to get
0
r
&
2
3
/
v
d&
= 3
/
v
1
3
&
3
0
r
= 3
/
v
3
1
3
v
3
=
1
3
v
2
/.
SECTION 6.2 VOLUMES

275
51. r
2
+j
2
= v
2
C r
2
= v
2
3j
2
\ =
r
r3I
v
2
3j
2
dj =
v
2
j 3
j
3
3
r
r3I
=
v
3
3
v
3
3
v
2
(v 3/) 3
(v 3/)
3
3
2
3
v
3
3
1
3
(v 3/)
3v
2
3(v 3/)
2
=
1
3
2v
3
3(v 3/)
3v
2
3
v
2
32v/ +/
2
=
1
3
2v
3
3(v 3/)
2v
2
+ 2v/ 3/
2
=
1
3
2v
3
32v
3
32v
2
/ +v/
2
+ 2v
2
/ + 2v/
2
3/
3
=
1
3
3v/
2
3/
3
=
1
3
/
2
(3v 3/), or, equivalently, /
2
v 3
/
3
53. For a cross-section at height j, we see from similar triangles that

c2
/2
=
/ 3j
/
, so c = /
1 3
j
/
.
Similarly, for cross-sections having 2/ as their base and o replacing c, o = 2/
1 3
j
/
. So
\ =

I
0
(j) dj =

I
0
1 3
j
/
2/
1 3
j
/
dj
=

I
0
2/
2
1 3
j
/
2
dj = 2/
2
I
0
1 3
2j
/
+
j
2
/
2
dj
= 2/
2
j 3
j
2
/
+
j
3
3/
2
I
0
= 2/
2
/ 3/ +
1
3
/
=
2
3
/
2
/ [ =
1
3
1/ where 1 is the area of the base, as with any pyramid.]
55. A cross-section at height : is a triangle similar to the base, so well multiply the legs of the base triangle, 3 and 4, by a
proportionality factor of (5 3:)5. Thus, the triangle at height : has area
(:) =
1
2
3
5 3:
5
5 3:
5
= 6
1 3
:
5
2
, so
\ =

5
0
(:) d: = 6
5
0
1 3
:
5
2
d: = 6
0
1
&
2
(35 d&)

& = 1 3:5,
d& = 3
1
5
d:
= 330
1
3
&
3
0
1
= 330
3
1
3
= 10 cm
3
57. If | is a leg of the isosceles right triangle and 2j is the hypotenuse,
then |
2
+|
2
= (2j)
2
i 2|
2
= 4j
2
i |
2
= 2j
2
.
\ =

2
32
(r) dr = 2
2
0
(r) dr = 2
2
0
1
2
(|)(|) dr = 2
2
0
j
2
dr
= 2
2
0
1
4
(36 39r
2
) dr =
9
2
2
0
(4 3r
2
) dr
=
9
2
4r 3
1
3
r
3
2
0
=
9
2
8 3
8
3
= 24
276

59. The cross-section of the base corresponding to the coordinate r has length
j = 1 3r. The corresponding square with side c has area
(r) = c
2
= (1 3r)
2
= 1 32r +r
2
. Therefore,
\ =

1
0
(r) dr =

1
0
(1 32r +r
2
) dr
=

r 3r
2
+
1
3
r
3
1
0
=

1 31 +
1
3
30 =
1
3
Or:

1
0
(1 3r)
2
dr =

0
1
&
2
(3d&) [& = 1 3r] =

1
3
&
3
1
0
=
1
3
61. The cross-section of the base / corresponding to the coordinate r has length 1 3r
2
. The height / also has length 1 3r
2
,
so the corresponding isosceles triangle has area (r) =
1
2
// =
1
2
(1 3r
2
)
2
. Therefore,
\ =

1
31
1
2
(1 3r
2
)
2
dr
= 2
1
2
1
0
(1 32r
2
+r
4
) dr [by symmetry]
=

r 3
2
3
r
3
+
1
5
r
5
1
0
=

1 3
2
3
+
1
5
30 =
8
15
63. (a) The torus is obtained by rotating the circle (r 31)
2
+j
2
= v
2
about
the j-axis. Solving for r, we see that the right half of the circle is given by
r = 1+
v
2
3j
2
= 1(j) and the left half by r = 13
v
2
3j
2
= o(j). So
\ =
r
3r
[1(j)]
2
3[o(j)]
2
dj
= 2
r
0
1
2
+ 21
v
2
3j
2
+v
2
3j
2
1
2
321
v
2
3j
2
+v
2
3j
2
dj
= 2
r
0
41
v
2
3j
2
dj = 81
r
0
v
2
3j
2
dj
(b) Observe that the integral represents a quarter of the area of a circle with radius v, so
81
r
0
v
2
3j
2
dj = 81
1
4
v
2
= 2
2
v
2
1.
65. (a) Volume(o
1
) =
I
0
(:) d: = Volume(o
2
) since the cross-sectional area (:) at height : is the same for both solids.
(b) By Cavalieris Principle, the volume of the cylinder in the gure is the same as that of a right circular cylinder with radius v
and height /, that is, v
2
/.
SECTION 6.2 VOLUMES

277
67. The volume is obtained by rotating the area common to two circles of radius v, as
shown. The volume of the right half is
\
right
=
r'2
0
j
2
dr =
r'2
0
v
2
3
1
2
v +r
dr
=
v
2
r 3
1
3
1
2
v +r
r'2
0
=
1
2
v
3
3
1
3
v
3
0 3
1
24
v
3
=
5
24
v
3
So by symmetry, the total volume is twice this, or
5
12
v
3
.
Another solution: We observe that the volume is the twice the volume of a cap of a sphere, so we can use the formula from
Exercise 51 with / =
1
2
v: \ = 2
1
3
/
2
(3v 3/) =
2
3
1
2
v
3v 3
1
2
v
=
5
12
v
3
.
69. Take the r-axis to be the axis of the cylindrical hole of radius v.
A quarter of the cross-section through j. perpendicular to the
j-axis, is the rectangle shown. Using the Pythagorean Theorem
twice, we see that the dimensions of this rectangle are
r =

1
2
3j
2
and : =

v
2
3j
2
, so
1
4
(j) = r: =

v
2
3j
2
1
2
3j
2
, and
\ =

r
3r
(j) dj =

r
3r
4
v
2
3j
2
1
2
3j
2
dj = 8
r
0
v
2
3j
2
1
2
3j
2
dj
71. (a) The radius of the barrel is the same at each end by symmetry, since the
function j = 13cr
2
is even. Since the barrel is obtained by rotating
the graph of the function j about the r-axis, this radius is equal to the
value of j at r =
1
2
/, which is 13c
1
2
/
2
= 13d = v.
(b) The barrel is symmetric about the j-axis, so its volume is twice the volume of that part of the barrel for r 0. Also, the
barrel is a volume of rotation, so
\ = 2
I'2
0
j
2
dr = 2
I'2
0
13cr
2
2
dr = 2
1
2
r 3
2
3
1cr
3
+
1
5
c
2
r
5
I'2
0
= 2
1
2
1
2
/ 3
1
12
1c/
3
+
1
160
c
2
/
5
Trying to make this look more like the expression we want, we rewrite it as \ =
1
3
/
21
2
+
1
2
3
1
2
1c/
2
+
3
80
c
2
/
4
.
But 1
2
3
1
2
1c/
2
+
3
80
c
2
/
4
=

13
1
4
c/
2
2
3
1
40
c
2
/
4
= (13d)
2
3
2
5
1
4
c/
2
2
= v
2
3
2
5
d
2
.
Substituting this back into \ , we see that \ =
1
3
/
21
2
+v
2
3
2
5
d
2
, as required.
278

6.3 Volumes by Cylindrical Shells
1. If we were to use the washer method, we would rst have to locate the
local maximum point (o. /) of j = r(r 31)
2
using the methods of
Chapter 4. Then we would have to solve the equation j = r(r 31)
2
for r in terms of j to obtain the functions r = o
1
(j) and r = o
2
(j)
shown in the rst gure. This step would be difcult because it involves
the cubic formula. Finally we would nd the volume using
\ =
l
0
[o
1
(j)]
2
3[o
2
(j)]
2
dj.
Using shells, we nd that a typical approximating shell has radius r, so its circumference is 2r. Its height is j, that is,
r(r 31)
2
. So the total volume is
\ =

1
0
2r
r(r 31)
2
dr = 2
1
0
r
4
32r
3
+r
2
dr = 2
r
5
5
32
r
4
4
+
r
3
3
1
0
=

15
3. \ =

2
1
2r
1
r
dr = 2
2
1
1 dr
= 2
2
1
= 2(2 31) = 2
5. \ =

1
0
2rc
3i
2
dr. Let & = r
2
.
Thus, d& = 2rdr, so
\ =
1
0
c
3u
d& =
3c
3u
1
0
= (1 31c).
7. The curves intersect when 4(r 32)
2
= r
2
34r + 7 C 4r
2
316r + 16 = r
2
34r + 7 C
3r
2
312r + 9 = 0 C 3(r
2
34r + 3) = 0 C 3(r 31)(r 33) = 0, so r = 1 or 3.
\ = 2
3
1
r
2
34r + 7
34(r 32)
2
dr = 2
3
1
r(r
2
34r + 7 34r
2
+ 16r 316)
dr
= 2
3
1
r(33r
2
+ 12r 39)
dr = 2(33)
3
1
(r
3
34r
2
+ 3r) dr = 36
1
4
r
4
3
4
3
r
3
+
3
2
r
2
3
1
= 36
81
4
336 +
27
2
1
4
3
4
3
+
3
2
= 36
20 336 + 12 +
4
3
= 36
3
8
3
= 16
SECTION 6.3 VOLUMES BY CYLINDRICAL SHELLS

279
9. \ =

2
1
2j(1 +j
2
) dj = 2
2
1
(j +j
3
) dj = 2
1
2
j
2
+
1
4
j
4
2
1
= 2
(2 + 4) 3
1
2
+
1
4
= 2
21
4
=
21
2

11. \ = 2
8
0
j(
3
j 30)
dj
= 2
8
0
j
4'3
dj = 2
3
7
j
7'3
8
0
=
6
7
(8
7'3
) =
6
7
(2
7
) =
768
7

13. The height of the shell is 2 3
1 + (j 32)
2
= 1 3(j 32)
2
= 1 3
j
2
34j + 4
= 3j
2
+ 4j 33.
\ = 2
3
1
j(3j
2
+ 4j 33) dj
= 2
3
1
(3j
3
+ 4j
2
33j) dj
= 2
3
1
4
j
4
+
4
3
j
3
3
3
2
j
2
3
1
= 2
3
81
4
+ 36 3
27
2
3
1
4
+
4
3
3
3
2
= 2
8
3
=
16
3

15. The shell has radius 2 3r, circumference 2(2 3r), and height r
4
.
\ =

1
0
2(2 3r)r
4
dr
= 2
1
0
(2r
4
3r
5
) dr
= 2
2
5
r
5
3
1
6
r
6
1
0
= 2
2
5
3
1
6
30
= 2
7
30
=
7
15
280

17. The shell has radius r 31, circumference 2(r 31), and height (4r 3r
2
) 33 = 3r
2
+ 4r 33.
\ =

3
1
2(r 31)(3r
2
+ 4r 33) dr
= 2
3
1
(3r
3
+ 5r
2
37r + 3) dr
= 2
3
1
4
r
4
+
5
3
r
3
3
7
2
r
2
+ 3r
3
1
= 2
3
81
4
+ 45 3
63
2
+ 9
3
1
4
+
5
3
3
7
2
+ 3
= 2
4
3
=
8
3
19. The shell has radius 1 3j, circumference 2(1 3j), and height 1 3
3
j

j = r
3
C r =
3
.
\ =

1
0
2(1 3j)(1 3j
1'3
) dj
= 2
1
0
(1 3j 3j
1'3
+j
4'3
) dj
= 2
j 3
1
2
j
2
3
3
4
j
4'3
+
3
7
j
7'3
1
0
= 2
1 3
1
2
3
3
4
+
3
7
30
= 2
5
28
=
5
14
21. \ =

2
1
2rlnrdr 23. \ =

1
0
2[r 3(31)]
sin
r
2
r 3r
4
dr
25. \ =

r
0
2(4 3j)
I
sinj dj
SECTION 6.3 VOLUMES BY CYLINDRICAL SHELLS

281
27. \ =

1
0
2r
I
1 +r
3
dr. Let 1(r) = r
I
1 +r
3
.
Then the Midpoint Rule with n = 5 gives
1
0
1(r) dr r
130
5
[1(0.1) +1(0.3) +1(0.5) +1(0.7) +1(0.9)]
r 0.2(2.9290)
Multiplying by 2 gives \ r 3.68.
29.

3
0
2r
5
dr = 2
3
0
r(r
4
) dr. The solid is obtained by rotating the region 0 $ j $ r
4
, 0 $ r $ 3 about the y-axis using
cylindrical shells.
31.

1
0
2(3 3j)(1 3j
2
) dj. The solid is obtained by rotating the region bounded by (i) r = 1 3j
2
, r = 0, and j = 0 or
(ii) r = j
2
, r = 1, and j = 0 about the line j = 3 using cylindrical shells.
33. From the graph, the curves intersect at r = 0 and r = o r 0.56,
with
I
r + 1 c
i
on the interval (0. o). So the volume of the solid
obtained by rotating the region about the j-axis is
\ = 2
a
0
r
I
r + 1
3c
i
dr r 0.13.
35. \ = 2
r'2
0
r
2
3r
sin
2
r 3sin
4
r
dr
CAS
=
1
32
3
37. Use shells:
\ =

4
2
2r(3r
2
+ 6r 38) dr = 2
4
2
(3r
3
+ 6r
2
38r) dr
= 2
3
1
4
r
4
+ 2r
3
34r
2
4
2
= 2[(364 + 128 364) 3(34 + 16 316)]
= 2(4) = 8
39. Use shells:
\ =

4
1
2[r 3(31)][5 3(r + 4r)] dr
= 2
4
1
(r + 1)(5 3r 34r) dr
= 2
4
1
(5r 3r
2
34 + 5 3r 34r) dr
= 2
4
1
(3r
2
+ 4r + 1 34r) dr = 2
3
1
3
r
3
+ 2r
2
+r 34 lnr
4
1
= 2
3
64
3
+ 32 + 4 34 ln4
3
1
3
+ 2 + 1 30
= 2(12 34 ln4) = 8(3 3ln4)

282

41. Use disks: r
2
+ (j 31)
2
= 1 C r =
1 3(j 31)
2
\ =
2
0
1 3(j 31)
2
2
dj =
2
0
(2j 3j
2
) dj
=
j
2
3
1
3
j
3
2
0
=
4 3
8
3
=
4
3
43. Use shells:

\ = 2
r
0
2r
I
v
2
3r
2
dr
= 32
r
0
(v
2
3r
2
)
1'2
(32r) dr
=

32
2
3
(v
2
3r
2
)
3'2
r
0
= 3
4
3
(0 3v
3
) =
4
3
v
3

45. \ = 2
r
0
r
3
/
v
r +/
dr = 2/
r
0
3
r
2
v
+r
dr
= 2/
3
r
3
3v
+
r
2
2
r
0
= 2/
v
2
6
=
v
2
/
3

6.4 Work
1. W = 1d = iod = (40)(9.8)(1.5) = 588 J
3. W =

l
a
1(r) dr =

9
0
10
(1 +r)
2
dr = 10
10
1
1
&
2
d& [& = 1 +r. d& = dr] = 10
3
1
&
10
1
= 10
3
1
10
+ 1
= 9 ft-lb
5. The force function is given by 1(r) (in newtons) and the work (in joules) is the area under the curve, given by
8
0
1(r) dr =

4
0
1(r) dr +
8
4
1(r) dr =
1
2
(4)(30) + (4)(30) = 180 J.
7. 10 = 1(r) = Ir =
1
3
I [4 inches =
1
3
foot], so I = 30 lbft and 1(r) = 30r. Now 6 inches =
1
2
foot, so
W =

1'2
0
30rdr =

15r
2
1'2
0
=
15
4
ft-lb.
9. (a) If
0.12
0
Irdr = 2 J, then 2 =

1
2
Ir
2
0.12
0
=
1
2
I(0.0144) = 0.0072I and I =
2
0.0072
=
2500
9
E 277.78 Nm.
Thus, the work needed to stretch the spring from 35 cm to 40 cm is
0.10
0.05
2500
9
rdr =

1250
9
r
2
1'10
1'20
=
1250
9
1
100
3
1
400
=
25
24
E 1.04 J.
(b) 1(r) = Ir, so 30 =
2500
9
r and r =
270
2500
m = 10.8 cm
SECTION 6.4 WORK

283
11. The distance from 20 cm to 30 cm is 0.1 m, so with 1(r) = Ir, we get W
1
=

0.1
0
Irdr = I
1
2
r
2
0.1
0
=
1
200
I.
Now W
2
=

0.2
0.1
Irdr = I
1
2
r
2
0.2
0.1
= I
4
200
3
1
200
=
3
200
I. Thus, W
2
= 3W
1
.
In Exercises 13 20, n is the number of subintervals of length {r, and r
W
.
is a sample point in the jth subinterval [r
.31
. r
.
].
13. (a) The portion of the rope from r ft to (r +{r) ft below the top of the building weighs
1
2
{r lb and must be lifted r
W
.
ft,
so its contribution to the total work is
1
2
r
W
.
{r ft-lb. The total work is
W = lim
n<"
n
.=1
1
2
r
W
.
{r =

50
0
1
2
rdr =

1
4
r
2
50
0
=
2500
4
= 625 ft-lb
Notice that the exact height of the building does not matter (as long as it is more than 50 ft).
(b) When half the rope is pulled to the top of the building, the work to lift the top half of the rope is
W
1
=

25
0
1
2
rdr =

1
4
r
2
25
0
=
625
4
ft-lb. The bottom half of the rope is lifted 25 ft and the work needed to accomplish
that is W
2
=

50
25
1
2
25 dr =
25
2
50
25
=
625
2
ft-lb. The total work done in pulling half the rope to the top of the building
is W = W
1
+W
2
=
625
2
+
625
4
=
3
4
625 =
1875
4
ft-lb.
15. The work needed to lift the cable is lim
n<"
n
.=1
2r
W
.
{r =

500
0
2rdr =

r
2
500
0
= 250,000 ft-lb. The work needed to lift
the coal is 800 lb 500 ft = 400,000 ft-lb. Thus, the total work required is 250,000 + 400,000 = 650,000 ft-lb.
17. At a height of r meters (0 $ r $ 12), the mass of the rope is (0.8 kgm)(12 3r m) = (9.6 30.8r) kg and the mass of the
water is
36
12
kgm
(12 3r m) = (36 33r) kg. The mass of the bucket is 10 kg, so the total mass is
(9.6 30.8r) +(36 33r) +10 = (55.6 33.8r) kg, and hence, the total force is 9.8(55.6 33.8r) N. The work needed to lift
the bucket {r m through the jth subinterval of [0. 12] is 9.8(55.6 33.8r
W
.
){r, so the total work is
W = lim
n<"
n
.=1
9.8(55.6 33.8r
W
.
) {r =

12
0
(9.8)(55.6 33.8r) dr = 9.8
55.6r 31.9r
2
12
0
= 9.8(393.6) E 3857 J
19. A slice of water {r m thick and lying at a depth of r
W
.
m (where 0 $ r
W
.
$
1
2
) has volume (2 1 {r) m
3
, a mass of
2000 {r kg, weighs about (9.8)(2000 {r) = 19,600 {r N, and thus requires about 19,600r
W
.
{r J of work for its removal.
So W = lim
n<"
n
.=1
19,600r
W
.
{r =

1'2
0
19,600rdr =

9800r
2
1'2
0
= 2450 J.
21. A rectangular slice of water {r m thick and lying r m above the bottom has width r m and volume 8r{r m
3
. It weighs
about (9.8 1000)(8r{r) N, and must be lifted (5 3r) m by the pump, so the work needed is about
(9.8 10
3
)(5 3r)(8r{r) J. The total work required is
W E
3
0
(9.8 10
3
)(5 3r)8rdr = (9.8 10
3
)
3
0
(40r 38r
2
) dr = (9.8 10
3
)
20r
2
3
8
3
r
3
3
0
= (9.8 10
3
)(180 372) = (9.8 10
3
)(108) = 1058.4 10
3
E 1.06 10
6
J
284

23. Let r measure depth (in feet) below the spout at the top of the tank. A horizontal
disk-shaped slice of water {r ft thick and lying at coordinate r has radius
3
8
(16 3r) ft (-) and volume v
2
{r =
9
64
(16 3r)
2
{r ft
3
. It weighs
about (62.5)
9r
64
(16 3r)
2
{r lb and must be lifted r ft by the pump, so the
work needed to pump it out is about (62.5)r
9r
64
(16 3r)
2
{r ft-lb. The total
work required is
W E
8
0
(62.5)r
9r
64
(16 3r)
2
dr = (62.5)
9r
64
8
0
r(256 332r +r
2
) dr
= (62.5)
9r
64
8
0
(256r 332r
2
+r
3
) dr = (62.5)
9r
64
128r
2
3
32
3
r
3
+
1
4
r
4
8
0
= (62.5)
9
64
11. 264
3
= 33,000 E 1.04 10
5
ft-lb
(-) From similar triangles,
J
8 r
=
3
8
.
So : = 3 + J = 3 +
3
8
(8 r)
=
3(8)
8
+
3
8
(8 r)
=
3
8
(16 r)
25. If only 4.7 10
5
J of work is done, then only the water above a certain level (call
it /) will be pumped out. So we use the same formula as in Exercise 21, except that
the work is xed, and we are trying to nd the lower limit of integration:
4.7 10
5
E
3
I
(9.8 10
3
)(5 3r)8rdr =

9.8 10
3
20r
2
3
8
3
r
3
3
I
C
4.7
9.8
10
2
E 48 =

20 3
2
3
8
3
3
3
20/
2
3
8
3
/
3
C
2/
3
315/
2
+ 45 = 0. To nd the solution of this equation, we plot 2/
3
315/
2
+ 45 between / = 0 and / = 3.
We see that the equation is satised for / E 2.0. So the depth of water remaining in the tank is about 2.0 m.
27. \ = v
2
r, so \ is a function of r and 1 can also be regarded as a function of r. If \
1
= v
2
r
1
and \
2
= v
2
r
2
, then
W =

i
2
i
1
1(r) dr =

i
2
i
1
v
2
1(\ (r)) dr =

i
2
i
1
1(\ (r)) d\ (r) [Let \ (r) = v
2
r, so d\ (r) = v
2
dr.]
=

1
2
1
1
1(\ ) d\ by the Substitution Rule.
29. W =

l
a
1(v) dv =

l
a
G
i
1
i
2
v
2
dv = Gi
1
i
2
31
v
l
a
= Gi
1
i
2
1
o
3
1
/
6.5 Average Value of a Function

1. 1
ave
=
1
l 3a
l
a
1(r) dr =
1
4 30
4
0
(4r 3r
2
) dr =
1
4
2r
2
3
1
3
r
3
4
0
=
1
4
32 3
64
3
30
=
1
4
32
3
=
8
3
3. o
ave
=
1
l 3a
l
a
o(r) dr =
1
8 31
8
1
3
I
rdr =
1
7
3
4
r
4'3
8
1
=
3
28
(16 31) =
45
28
5. 1
ave
=
1
5 30
5
0
tc
3I
2
dt =
1
5
325
0
c
u
3
1
2
d&

& = 3t
2
, d& = 32t dt, t dt = 3
1
2
d&
= 3
1
10
c
u
325
0
= 3
1
10
(c
325
31) =
1
10
(1 3c
325
)
SECTION 6.5 AVERAGE VALUE OF A FUNCTION

285
7. /
ave
=
1
r 30
r
0
cos
4
r sinrdr =
1
r
31
1
&
4
(3d&) [& = cos r, d& = 3sinrdr]
=
1
r
1
31
&
4
d& =
1
r
2
1
0
&
4
d& [by Theorem 5.5.7] =
2
r
1
5
&
5
1
0
=
2
5r
9. (a) 1
ave
=
1
5 3 2
5
2
(r 33)
2
dr =
1
3
1
3
(r 33)
3
5
2
=
1
9
2
3
3(31)
3
=
1
9
(8 + 1) = 1
(c)
(b) 1(c) = 1
ave
C (c 33)
2
= 1 C
c 33 = 1 C c = 2 or 4
11. (a) 1
ave
=
1
3 0
r
0
(2 sinr 3sin2r) dr
=
1
r
32 cos r +
1
2
cos 2r
r
0
=
1
r
2 +
1
2
32 +
1
2
=
4
r
(c)
(b) 1(c) = 1
ave
C 2 sinc 3sin2c =
4
r
C
c
1
E 1.238 or c
2
E 2.808
13. 1 is continuous on [1. 3], so by the Mean Value Theorem for Integrals there exists a number c in [1. 3] such that
3
1
1(r) dr = 1(c)(3 31) i 8 = 21(c); that is, there is a number c such that 1(c) =
8
2
= 4.
15. 1
ave
=
1
50 320
50
20
1(r) dr E
1
30
A
3
=
1
30

50 320
3
[1(25) +1(35) +1(45)] =
1
3
(38 + 29 + 48) =
115
3
= 38
1
3
17. Let t = 0 and t = 12 correspond to 9 AM and 9 PM, respectively.
T
ave
=
1
12 30
12
0
50 + 14 sin
1
12
t
dt =
1
12
50t 314
12
r
cos
1
12
t
12
0
=
1
12
50 12 + 14
12
r
+ 14
12
r
=

50 +
28
r
F E 59
F
19. a
ave
=
1
8
8
0
12
I
r + 1
dr =
3
2
8
0
(r + 1)
31'2
dr =

3
I
r + 1
8
0
= 9 33 = 6 kgm
21. \
ave
=
1
5
5
0
\ (t) dt =
1
5
5
0
5
4r
1 3cos
2
5
t
dt =
1
4r
5
0
1 3cos
2
5
t
dt
=
1
4r
t 3
5
2r
sin
2
5
t
5
0
=
1
4r
[(5 30) 30] =
5
4r
E 0.4 L
23. Let 1(r) =

i
a
1(t) dt for r in [o. /]. Then 1 is continuous on [o. /] and differentiable on (o. /), so by the Mean Value
Theorem there is a number c in (o. /) such that 1(/) 31(o) = 1
0
(c)(/ 3o). But 1
0
(r) = 1(r) by the Fundamental
Theorem of Calculus. Therefore,
l
a
1(t) dt 30 = 1(c)(/ 3o).
286

6 Review
1. (a) See Section 6.1, Figure 2 and Equations 6.1.1 and 6.1.2.
(b) Instead of using top minus bottom and integrating from left to right, we use right minus left and integrate from bottom
to top. See Figures 11 and 12 in Section 6.1.
2. The numerical value of the area represents the number of meters by which Sue is ahead of Kathy after 1 minute.
3. (a) See the discussion in Section 6.2, near Figures 2 and 3, ending in the Denition of Volume.
(b) See the discussion between Examples 5 and 6 in Section 6.2. If the cross-section is a disk, nd the radius in terms of r or j
and use = (radius)
2
. If the cross-section is a washer, nd the inner radius v
in
and outer radius v
out
and use
=
v
2
out
v
2
in
.
4. (a) \ = 2v/{v = (circumference)(height)(thickness)
(b) For a typical shell, nd the circumference and height in terms of r or j and calculate
\ =

l
a
(circumference)(height)(dr or dj), where o and / are the limits on r or j.
(c) Sometimes slicing produces washers or disks whose radii are difcult (or impossible) to nd explicitly. On other
occasions, the cylindrical shell method leads to an easier integral than slicing does.
5.

6
0
1(r) dr represents the amount of work done. Its units are newton-meters, or joules.
6. (a) The average value of a function 1 on an interval [o. /] is 1
ave
=
1
/ 3o
l
a
1(r) dr.
(b) The Mean Value Theorem for Integrals says that there is a number c at which the value of 1 is exactly equal to the average
value of the function, that is, 1(c) = 1
ave
. For a geometric interpretation of the Mean Value Theorem for Integrals, see
Figure 2 in Section 6.5 and the discussion that accompanies it.
1. The curves intersect when r
2
= 4r 3r
2
C 2r
2
34r = 0 C
2r(r 32) = 0 C r = 0 or 2.
=

2
0
(4r 3r
2
) 3r
2
dr =

2
0
(4r 32r
2
) dr
=

2r
2
3
2
3
r
3
2
0
=

8 3
16
3
30
=
8
3
3. If r D 0, then | r| = r, and the graphs intersect when r = 1 32r
2
C 2r
2
+r 31 = 0 C (2r 31)(r +1) = 0 C
r =
1
2
or 31, but 31 < 0. By symmetry, we can double the area from r = 0 to r =
1
2
.
= 2
1'2
0
(1 32r
2
) 3r
dr = 2
1'2
0
(32r
2
3r + 1) dr
= 2
3
2
3
r
3
3
1
2
r
2
+r
1'2
0
= 2
3
1
12
3
1
8
+
1
2
30
= 2
7
24
=
7
12
CHAPTER 6 REVIEW

287
5. =

2
0
sin
r
2
3(r
2
32r)
dr
=

3
2
cos
r
2
3
1
3
r
3
+r
2
2
0
=

2
r
3
8
3
+ 4
3
2
r
30 + 0
=
4
3
+
4
r
7. Using washers with inner radius r
2
and outer radius 2r, we have
\ =
2
0
(2r)
2
3(r
2
)
2
dr =
2
0
(4r
2
3r
4
) dr
=
4
3
r
3
3
1
5
r
5
2
0
=
32
3
3
32
5
= 32
2
15
=
64
15
9. \ =
3
33
(9 3j
2
) 3(31)
2
3[0 3(31)]
2
dj
= 2
3
0
(10 3j
2
)
2
31
dj = 2
3
0
(100 320j
2
+j
4
31) dj
= 2
3
0
(99 320j
2
+j
4
) dj = 2
99j 3
20
3
j
3
+
1
5
j
5
3
0
= 2
297 3180 +
243
5
=
1656
5

11. The graph of r
2
3j
2
= o
2
is a hyperbola with right and left branches.
Solving for j gives us j
2
= r
2
3o
2
i j =
I
r
2
3o
2
.
Well use shells and the height of each shell is
I
r
2
3o
2
3
3
I
r
2
3o
2
= 2
I
r
2
3o
2
.
The volume is \ =

a+I
a
2r 2
I
r
2
3o
2
dr. To evaluate, let & = r
2
3o
2
,
so d& = 2rdr and rdr =
1
2
d&. When r = o, & = 0, and when r = o +/,
& = (o +/)
2
3o
2
= o
2
+ 2o/ +/
2
3o
2
= 2o/ +/
2
.
Thus, \ = 4
2aI+I
2
0
I
&
1
2
d&
= 2
2
3
&
3'2
2aI+I
2
0
=
4
3
2o/ +/
2
3'2
.
13. A shell has radius
r
2
3r, circumference 2
r
2
3r
, and height cos

2
r 3
1
4
.
j = cos
2
r intersects j =
1
4
when cos
2
r =
1
4
C
cos r =
1
2
[ |r| $ 2] C r =
r
3
.
\ =

r'3
3r'3
2
2
3r
cos
2
r 3
1
4
dr
288

15. (a) A cross-section is a washer with inner radius r
2
and outer radius r.
\ =

1
0

(r)
2
3(r
2
)
2
dr =

1
0
(r
2
3r
4
) dr =
1
3
r
3
3
1
5
r
5
1
0
=
1
3
3
1
5
=
2
15
(b) A cross-section is a washer with inner radius j and outer radius
j.
\ =

1
0

2
3j
2
dj =

1
0
(j 3j
2
) dj =
1
2
j
2
3
1
3
j
3
1
0
=
1
2
3
1
3
=
r
6
(c) A cross-section is a washer with inner radius 2 3r and outer radius 2 3r
2
.
\ =

1
0

(2 3r
2
)
2
3(2 3r)
2
dr =

1
0
(r
4
35r
2
+ 4r) dr =
1
5
r
5
3
5
3
r
3
+ 2r
2
1
0
=
1
5
3
5
3
+ 2
=
8
15
17. (a) Using the Midpoint Rule on [0. 1] with 1(r) = tan(r
2
) and n = 4, we estimate
=

1
0
tan(r
2
) dr E
1
4
tan
1
8
+ tan
3
8
+ tan
5
8
+ tan
7
8
E
1
4
(1.53) E 0.38
(b) Using the Midpoint Rule on [0. 1] with 1(r) = tan
2
(r
2
) (for disks) and n = 4, we estimate
\ =

1
0
1(r) dr E
1
4
tan
2
1
8
+ tan
2
3
8
+ tan
2
5
8
+ tan
2
7
8
E
r
4
(1.114) E 0.87
19.

r'2
0
2rcos rdr =

r'2
0
(2r) cos rdr
The solid is obtained by rotating the region R=

(r. j) | 0 $ r $
r
2
. 0 $ j $ cos r
about the j-axis.

21.

r
0
(2 3sinr)
2
dr
The solid is obtained by rotating the region R= {(r. j) | 0 $ r $ . 0 $ j $ 2 3sinr} about the r-axis.
23. Take the base to be the disk r
2
+j
2
$ 9. Then \ =

3
33
(r) dr, where (r
0
) is the area of the isosceles right triangle
whose hypotenuse lies along the line r = r
0
in the rj-plane. The length of the hypotenuse is 2
I
9 3r
2
and the length of
each leg is
I
2
I
9 3r
2
. (r) =
1
2
I
2
I
9 3r
2
2
= 9 3r
2
, so
\ = 2
3
0
(r) dr = 2
3
0
(9 3r
2
) dr = 2
9r 3
1
3
r
3
3
0
= 2(27 39) = 36
25. Equilateral triangles with sides measuring
1
4
r meters have height
1
4
rsin60
=
I
3
8
r. Therefore,
(r) =
1
2

1
4
r
I
3
8
r =
I
3
64
r
2
. \ =

20
0
(r) dr =
I
3
64
20
0
r
2
dr =
I
3
64
1
3
r
3
20
0
=
8000
I
3
64 3
=
125
I
3
3
m
3
.
27. 1(r) = Ir i 30 N = I(15 312) cm i I = 10 Ncm = 1000 Nm. 20 cm312 cm = 0.08 m i
W =

0.08
0
Irdr = 1000
0.08
0
rdr = 500
r
2
0.08
0
= 500(0.08)
2
= 3.2 Nm = 3.2 J.
CHAPTER 6 REVIEW

289
29. (a) The parabola has equation j = or
2
with vertex at the origin and passing through
(4. 4). 4 = o 4
2
i o =
1
4
i j =
1
4
r
2
i r
2
= 4j i
r = 2
j. Each circular disk has radius 2
j and is moved 4 3j ft.

W =
4
0

2
62.5(4 3j) dj = 250
4
0
j(4 3j) dj
= 250
2j
2
3
1
3
j
3
4
0
= 250
32 3
64
3
=
8000r
3
E 8378 ft-lb
(b) In part (a) we knew the nal water level (0) but not the amount of work done. Here
we use the same equation, except with the work xed, and the lower limit of
integration (that is, the nal water level call it /) unknown: W = 4000 C
250
2j
2
3
1
3
j
3
4
I
= 4000 C
16
r
=

32 3
64
3
2/
2
3
1
3
/
3
C
/
3
36/
2
+ 32 3
48
r
= 0. We graph the function 1(/) = /
3
36/
2
+ 32 3
48
r
on the interval [0. 4] to see where it is 0. From the graph, 1(/) = 0 for / E 2.1.
So the depth of water remaining is about 2.1 ft.
31. lim
I<0
1
ave
= lim
I<0
1
(r +/) 3r
i+I
i
1(t) dt = lim
I<0
1(r +/) 31(r)
/
, where 1(r) =

i
a
1(t) dt. But we recognize this
limit as being 1
0
(r) by the denition of a derivative. Therefore, lim
I<0
1
ave
= 1
0
(r) = 1(r) by FTC1.
PROBLEMS PLUS
1. (a) The area under the graph of 1 from 0 to t is equal to
I
0
1(r) dr, so the requirement is that
I
0
1(r) dr = t
3
for all t. We
differentiate both sides of this equation with respect to t (with the help of FTC1) to get 1(t) = 3t
2
. This function is
positive and continuous, as required.
(b) The volume generated from r = 0 to r = / is
l
0
[1(r)]
2
dr. Hence, we are given that /
2
=

l
0
[1(r)]
2
dr for all
/ 0. Differentiating both sides of this equation with respect to / using the Fundamental Theorem of Calculus gives
2/ = [1(/)]
2
i 1(/) =
2/, since 1 is positive. Therefore, 1(r) =
2r.
3. Let o and / be the r-coordinates of the points where the line intersects the
curve. From the gure, 1
1
= 1
2
i
a
0
c 3
8r 327r
3
dr =
l
a
8r 327r
3
3c
dr
cr 34r
2
+
27
4
r
4
a
0
=
4r
2
3
27
4
r
4
3cr
l
a
oc 34o
2
+
27
4
o
4
=
4/
2
3
27
4
/
4
3/c
4o
2
3
27
4
o
4
3oc
0 = 4/
2
3
27
4
/
4
3/c = 4/
2
3
27
4
/
4
3/
8/ 327/
3
= 4/
2
3
27
4
/
4
38/
2
+ 27/
4
=
81
4
/
4
34/
2
=/
2
81
4
/
2
34
So for / 0, /
2
=
16
81
i / =
4
9
. Thus, c = 8/ 327/
3
= 8
4
9
327
64
729
=
32
9
3
64
27
=
32
27
.
5. (a) \ = /
2
(v 3/3) =
1
3
/
2
(3v 3/). See the solution to Exercise 6.2.51.
(b) The smaller segment has height / = 1 3 r and so by part (a) its volume is
\ =
1
3
(1 3r)
2
[3(1) 3(1 3r)] =
1
3
(r 31)
2
(r + 2). This volume must be
1
3
of the total volume of the sphere,
which is
4
3
(1)
3
. So
1
3
(r 31)
2
(r + 2) =
1
3
4
3
i (r
2
32r + 1)(r + 2) =
4
3
i r
3
33r + 2 =
4
3
i
3r
3
39r + 2 = 0. Using Newtons method with 1(r) = 3r
3
39r + 2, 1
0
(r) = 9r
2
39, we get
r
n+1
= r
n
3
3r
3
n
39r
n
+ 2
9r
2
n
39
. Taking r
1
= 0, we get r
2
E 0.2222, and r
3
E 0.2261 E r
4
, so, correct to four decimal
places, r E 0.2261.
(c) With v = 0.5 and c = 0.75, the equation r
3
33vr
2
+ 4v
3
c = 0 becomes r
3
33(0.5)r
2
+ 4(0.5)
3
(0.75) = 0 i
r
3
3
3
2
r
2
+ 4
1
8
3
4
= 0 i 8r
3
312r
2
+ 3 = 0. We use Newtons method with 1(r) = 8r
3
312r
2
+ 3,
1
0
(r) = 24r
2
324r, so r
n+1
= r
n
3
8r
3
n
312r
2
n
+ 3
24r
2
n
324r
n
. Take r
1
= 0.5. Then r
2
E 0.6667, and r
3
E 0.6736 E r
4
.
So to four decimal places the depth is 0.6736 m.
291
292

(d) (i) From part (a) with v = 5 in., the volume of water in the bowl is
\ =
1
3
/
2
(3v 3/) =
1
3
/
2
(15 3/) = 5/
2
3
1
3
/
3
. We are given that
d\
dt
= 0.2 in
3
s and we want to nd
d/
dt
when / = 3. Now
d\
dt
= 10/
d/
dt
3/
2
d/
dt
, so
d/
dt
=
0.2
(10/ 3/
2
)
. When / = 3, we have
d/
dt
=
0.2
(10 3 33
2
)
=
1
105
E 0.003 ins.
(ii) From part (a), the volume of water required to ll the bowl from the instant that the water is 4 in. deep is
\ =
1
2

4
3
(5)
3
3
1
3
(4)
2
(15 34) =
2
3
125 3
16
3
11 =
74
3
. To nd the time required to ll the bowl we divide
this volume by the rate: Time =
74r'3
0.2
=
370r
3
E 387 s E 6.5 min.
7. We are given that the rate of change of the volume of water is
d\
dt
= 3I(r), where I is some positive constant and (r) is
the area of the surface when the water has depth r. Now we are concerned with the rate of change of the depth of the water
with respect to time, that is,
dr
dt
. But by the Chain Rule,
d\
dt
=
d\
dr
dr
dt
, so the rst equation can be written
d\
dr
dr
dt
= 3I(r) (-). Also, we know that the total volume of water up to a depth r is \ (r) =
i
0
(c) dc, where (c) is
the area of a cross-section of the water at a depth c. Differentiating this equation with respect to r, we get d\dr = (r).
Substituting this into equation -, we get (r)(drdt) = 3I(r) i drdt = 3I, a constant.
9. We must nd expressions for the areas and 1, and then set them equal and see what this says about the curve C. If
1 =

o. 2o
2
, then area is just
a
0
(2r
2
3r
2
) dr =

a
0
r
2
dr =
1
3
o
3
. To nd area 1, we use j as the variable of
integration. So we nd the equation of the middle curve as a function of j: j = 2r
2
C r =

j2, since we are
concerned with the rst quadrant only. We can express area 1 as
2a
2
0
j2 3C(j)
dj =

4
3
(j2)
3'2
2a
2
0
3
2a
2
0
C(j) dj =
4
3
o
3
3
2a
2
0
C(j) dj
where C(j) is the function with graph C. Setting = 1, we get
1
3
o
3
=
4
3
o
3
3
2a
2
0
C(j) dj C

2a
2
0
C(j) dj = o
3
.
Now we differentiate this equation with respect to o using the Chain Rule and the Fundamental Theorem:
C(2o
2
)(4o) = 3o
2
i C(j) =
3
4
j2, where j = 2o
2
. Now we can solve for j: r =
3
4
j2 i
r
2
=
9
16
(j2) i j =
32
9
r
2
.
11. (a) Stacking disks along the j-axis gives us \ =

I
0
[1(j)]
2
dj.
(b) Using the Chain Rule,
d\
dt
=
d\
d/

d/
dt
= [1(/)]
2
d/
dt
.

293
(c) I
I
/ = [1(/)]
2
d/
dt
. Set
d/
dt
= C: [1(/)]
2
C = I
I
/ i [1(/)]
2
=
I
C
I
/ i 1(/) =
I
C
/
1'4
; that
is, 1(j) =
I
C
j
1'4
. The advantage of having
d/
dt
= C is that the markings on the container are equally spaced.
13. The cubic polynomial passes through the origin, so let its equation be
j = jr
3
+qr
2
+vr. The curves intersect when jr
3
+qr
2
+vr = r
2
C
jr
3
+ (q 31)r
2
+vr = 0. Call the left side 1(r). Since 1(o) = 1(/) = 0.
another form of 1 is
1(r) = jr(r 3o)(r 3/) = jr[r
2
3(o +/)r +o/]
= j[r
3
3(o +/)r
2
+o/r]
Since the two areas are equal, we must have
a
0
1(r) dr = 3
l
a
1(r) dr i
[1(r)]
a
0
= [1 (r)]
a
l
i 1(o) 31(0) = 1(o) 31(/) i 1(0) = 1(/), where 1 is an antiderivative of 1.
Now 1(r) =
1(r) dr =

j[r
3
3(o +/)r
2
+o/r] dr = j
1
4
r
4
3
1
3
(o +/)r
3
+
1
2
o/r
2
+C, so
1(0) = 1(/) i C = j
1
4
/
4
3
1
3
(o +/)/
3
+
1
2
o/
3
+C i 0 = j
1
4
/
4
3
1
3
(o +/)/
3
+
1
2
o/
3
i
0 = 3/ 34(o +/) + 6o [multiply by 12(j/
3
), / 6= 0] i 0 = 3/ 34o 34/ + 6o i / = 2o.
Hence, / is twice the value of o.
15. We assume that 1 lies in the region of positive r. Since j = r
3
is an odd
function, this assumption will not affect the result of the calculation. Let
1 =

o. o
3
. The slope of the tangent to the curve j = r

3
at 1 is 3o
2
, and so
the equation of the tangent is j 3o
3
= 3o
2
(r 3o) C j = 3o
2
r 32o
3
.
We solve this simultaneously with j = r
3
to nd the other point of intersection:
r
3
= 3o
2
r 32o
3
C (r 3o)
2
(r + 2o) = 0. So Q =

32o. 38o
3
is
the other point of intersection. The equation of the tangent at Q is
j 3(38o
3
) = 12o
2
[r 3(32o)] C j = 12o
2
r + 16o
3
. By symmetry,
this tangent will intersect the curve again at r = 32(32o) = 4o. The curve lies above the rst tangent, and
below the second, so we are looking for a relationship between =

a
32a
r
3
3(3o
2
r 32o
3
)
dr and
1 =

4a
32a
(12o
2
r + 16o
3
) 3r
3
dr. We calculate =

1
4
r
4
3
3
2
o
2
r
2
+ 2o
3
r
a
32a
=
3
4
o
4
3(36o
4
) =
27
4
o
4
, and
1 =

6o
2
r
2
+ 16o
3
r 3
1
4
r
4
4a
32a
= 96o
4
3(312o
4
) = 108o
4
. We see that 1 = 16 = 2
4
. This is because our
calculation of area 1 was essentially the same as that of area , with o replaced by 32o, so if we replace o with 32o in our
expression for , we get
27
4
(32o)
4
= 108o
4
= 1.
7 TECHNIQUES OF INTEGRATION
7.1 Integration by Parts
1. Let & = lnr, d = r
2
dr i d& =
1
i
dr, =
1
3
r
3
. Then by Equation 2,
r
2
lnrdr = (lnr)
1
3
r
3
3

1
3
r
3
1
i
dr =
1
3
r
3
lnr 3
1
3
r
2
dr =
1
3
r
3
lnr 3
1
3
1
3
r
3
+C
=
1
3
r
3
lnr 3
1
9
r
3
+C

or
1
3
r
3
lnr 3
1
3
+C
Note: A mnemonic device which is helpful for selecting & when using integration by parts is the LIATE principle of precedence for &:
Logarithmic
Inverse trigonometric
Algebraic
Trigonometric
Exponential
If the integrand has several factors, then we try to choose among them a & which appears as high as possible on the list. For example, in
rc
2i
dr
the integrand is rc
2i
, which is the product of an algebraic function (r) and an exponential function (c
2i
). Since Algebraic appears before Exponential,
we choose & = r. Sometimes the integration turns out to be similar regardless of the selection of & and d, but it is advisable to refer to LIATE when in
doubt.
3. Let & = r, d = cos 5rdr i d& = dr, =
1
5
sin5r. Then by Equation 2,
rcos 5rdr =
1
5
rsin5r 3
1
5
sin5rdr =
1
5
rsin5r +
1
25
cos 5r +C.
5. Let & = v, d = c
r'2
dv i d& = dv, = 2c
r'2
. Then
vc
r'2
dv = 2vc
r'2
3
2c
r'2
dv = 2vc
r'2
34c
r'2
+C.
7. Let & = r
2
, d = sinrdr i d& = 2rdr and = 3
1
r
cos r. Then
1 =

r
2
sinrdr = 3
1
r
r
2
cos r +
2
r
rcos rdr (-). Next let l = r, d\ = cos rdr i dl = dr,

\ =
1
r
sinr, so
rcos rdr =
1
r
rsinr 3
1
r
sinrdr =
1
r
rsinr +
1
r
2
cos r +C
1
.
Substituting for

rcos rdr in (-), we get
1 = 3
1
r
r
2
cos r +
2
r
1
r
rsinr +
1
r
2
cos r +C
1
= 3
1
r
r
2
cos r +
2
r
2
rsinr +
2
r
3
cos r +C, where C =
2
r
C
1
.
9. Let & = ln(2r + 1), d = dr i d& =
2
2r + 1
dr, = r. Then
ln(2r + 1) dr = rln(2r + 1) 3
2r
2r + 1
dr = rln(2r + 1) 3
(2r + 1) 31
2r + 1
dr
= rln(2r + 1) 3

1 3
1
2r + 1
dr = rln(2r + 1) 3r +
1
2
ln(2r + 1) +C
=
1
2
(2r + 1) ln(2r + 1) 3r +C
11. Let & = arctan4t, d = dt i d& =
4
1 + (4t)
2
dt =
4
1 + 16t
2
dt, = t. Then
arctan4t dt = t arctan4t 3
4t
1 + 16t
2
dt = t arctan4t 3
1
8
32t
1 + 16t
2
dt = t arctan4t 3
1
8
ln(1 + 16t
2
) +C.
295
296

CHAPTER 7 TECHNIQUES OF INTEGRATION
13. Let & = t, d = sec
2
2t dt i d& = dt, =
1
2
tan2t. Then
t sec
2
2t dt =
1
2
t tan2t 3
1
2
tan2t dt =
1
2
t tan2t 3
1
4
ln|sec 2t| +C.
15. First let & = (lnr)
2
, d = dr i d& = 2 lnr
1
i
dr, = r. Then by Equation 2,
1 =

(lnr)
2
dr = r(lnr)
2
32
rlnr
1
i
dr = r(lnr)
2
32
lnrdr. Next let l = lnr, d\ = dr i

dl = 1rdr, \ = r to get
lnrdr = rlnr 3
r (1r) dr = rlnr 3
dr = rlnr 3r +C
1
. Thus,
1 = r(lnr)
2
32(rlnr 3r +C
1
) = r(lnr)
2
32rlnr + 2r +C, where C = 32C
1
.
17. First let & = sin30, d = c
20
d0 i d& = 3 cos 30 d0, =
1
2
c
20
. Then
1 =

c
20
sin30 d0 =
1
2
c
20
sin30 3
3
2
c
20
cos 30 d0. Next let l = cos 30, d\ = c
20
d0 i dl = 33 sin30 d0,
\ =
1
2
c
20
to get
c
20
cos 30 d0 =
1
2
c
20
cos 30 +
3
2
c
20
sin30 d0. Substituting in the previous formula gives
1 =
1
2
c
20
sin30 3
3
4
c
20
cos 30 3
9
4
c
20
sin30 d0 =
1
2
c
20
sin30 3
3
4
c
20
cos 30 3
9
4
1 i
13
4
1 =
1
2
c
20
sin30 3
3
4
c
20
cos 30 +C
1
. Hence, 1=
1
13
c
20
(2 sin30 33 cos 30) +C, where C =
4
13
C
1
.
19. Let & = t, d = sin3t dt i d& = dt, = 3
1
3
cos 3t. Then
r
0
t sin3t dt =

3
1
3
t cos 3t
r
0
+
1
3
r
0
cos 3t dt =

1
3
30
+
1
9
sin3t
r
0
=
r
3
.
21. Let & = t, d = cosht dt i d& = dt. = sinht. Then
1
0
t cosht dt =

t sinht
1
0
3
1
0
sinht dt = (sinh1 3sinh0) 3
cosht
1
0
= sinh1 3(cosh1 3cosh0)
= sinh1 3cosh1 + 1.
We can use the denitions of sinh and cosh to write the answer in terms of c:
sinh1 3cosh1 + 1 =
1
2
(c
1
3c
31
) 3
1
2
(c
1
+c
31
) + 1 = 3c
31
+ 1 = 1 31c.
23. Let & = lnr, d = r
32
dr i d& =
1
r
dr, = 3r
31
. By (6),
2
1
lnr
r
2
dr =

3
lnr
r
2
1
+
2
1
r
32
dr = 3
1
2
ln2 + ln1 +
3
1
r
2
1
= 3
1
2
ln2 + 0 3
1
2
+ 1 =
1
2
3
1
2
ln2.
25. Let & = j, d =
dj
c
2u
= c
32u
dj i d& = dj, = 3
1
2
c
32u
. Then
1
0
j
c
2u
dj =

3
1
2
jc
32u
1
0
+
1
2
1
0
c
32u
dj =

3
1
2
c
32
+ 0
3
1
4
c
32u
1
0
= 3
1
2
c
32
3
1
4
c
32
+
1
4
=
1
4
3
3
4
c
32
.
27. Let & = cos
31
r, d = dr i d& = 3
dr
I
1 3r
2
, = r. Then
1 =

1'2
0
cos
31
rdr =

rcos
31
r
1'2
0
+
1'2
0
rdr
I
1 3r
2
=
1
2

r
3
+
3'4
1
t
31'2
3
1
2
dt
, where t = 1 3r
2
i
dt = 32rdr. Thus, 1 =
r
6
+
1
2
1
3'4
t
31'2
dt =
r
6
+
I
t
1
3'4
=
r
6
+ 1 3
I
3
2
=
1
6
+ 6 33
I
3
.
SECTION 7.1 INTEGRATION BY PARTS

297
29. Let & = ln(sinr), d = cos rdr i d& =
cos r
sinr
dr, = sinr. Then
1 =

cos rln(sinr) dr = sinrln(sinr) 3
cos rdr = sinrln(sinr) 3sinr +C.

Another method: Substitute t = sinr, so dt = cos rdr. Then 1 =

lnt dt = t lnt 3t +C (see Example 2) and so
1 = sinr(lnsinr 31) +C.
31. Let & = (lnr)
2
, d = r
4
dr i d& = 2
lnr
r
dr, =
r
5
5
. By (6),
2
1
r
4
(lnr)
2
dr =

r
5
5
(lnr)
2
2
1
32
2
1
r
4
5
lnrdr =
32
5
(ln2)
2
30 32
2
1
r
4
5
lnrdr.
Let l = lnr, d\ =
r
4
5
dr i dl =
1
r
dr, \ =
r
5
25
.
Then
2
1
r
4
5
lnrdr =

r
5
25
lnr
2
1
3
2
1
r
4
25
dr =
32
25
ln2 30 3
r
5
125
2
1
=
32
25
ln2 3
32
125
3
1
125
.
So
2
1
r
4
(lnr)
2
dr =
32
5
(ln2)
2
32
32
25
ln2 3
31
125
=
32
5
(ln2)
2
3
64
25
ln2 +
62
125
.
33. Let j =
I
r, so that dj =
1
2
r
31'2
dr =
1
2
I
r
dr =
1
2j
dr. Thus,
cos
I
rdr =

cos j (2j dj) = 2
j cos j dj. Now

use parts with & = j, d = cos j dj, d& = dj, = sinj to get
j cos j dj = j sinj 3
sinj dj = j sinj + cos j +C

1
,
so
cos
I
rdr = 2j sinj + 2 cos j +C = 2
I
rsin
I
r + 2 cos
I
r +C.
35. Let r = 0
2
, so that dr = 20 d0. Thus,
I
r
I
r/2
0
3
cos
0
2
d0 =

I
r
I
r/2
0
2
cos
0
2
1
2
(20 d0) =
1
2
r
r/2
rcos rdr. Now use
parts with & = r, d = cos rdr, d& = dr, = sinr to get
1
2
r
r/2
rcos rdr =
1
2
rsinr
r
r/2
3
r
r/2
sinrdr
=
1
2
rsinr + cos r
r
r/2
=
1
2
( sin + cos ) 3
1
2
r
2
sin
r
2
+ cos
r
2
=
1
2
( 0 31) 3
1
2
r
2
1 + 0
= 3
1
2
3
r
4
37. Let j = 1 +r. so that dj = dr. Thus,
rln(1 +r) dr =

(j 31) lnj dj. Now use parts with & = lnj. d = (j 31) dj,
d& =
1
u
dj, =
1
2
j
2
3j to get
(j 31) lnj dj =

1
2
j
2
3j
lnj 3

1
2
j 31
dj =
1
2
j(j 32) lnj 3
1
4
j
2
+j +C
=
1
2
(1 +r)(r 31) ln(1 +r) 3
1
4
(1 +r)
2
+ 1 +r +C,
which can be written as
1
2
(r
2
31) ln(1 +r) 3
1
4
r
2
+
1
2
r +
3
4
+C.
In Exercises 39 42, let 1(r) denote the integrand and 1(r) its antiderivative (with C = 0).
39. Let & = 2r + 3, d = c
i
dr i d& = 2 dr, = c
i
. Then
(2r + 3)c
i
dr = (2r + 3)c
i
32
c
i
dr = (2r + 3)c
i
32c
i
+C
= (2r + 1) c
i
+C
We see from the graph that this is reasonable, since 1 has a minimum where
1 changes from negative to positive.
298

41. Let & =
1
2
r
2
, d = 2r
I
1 +r
2
dr i d& = rdr, =
2
3
(1 +r
2
)
3'2
.
Then
r
3
I
1 +r
2
dr =
1
2
r
2
2
3
(1 +r
2
)
3'2
3
2
3
r(1 +r
2
)
3'2
dr
=
1
3
r
2
(1 +r
2
)
3'2
3
2
3

2
5

1
2
(1 +r
2
)
5'2
+C
=
1
3
r
2
(1 +r
2
)
3'2
3
2
15
(1 +r
2
)
5'2
+C
Another method: Use substitution with & = 1 +r
2
to get
1
5
(1 +r
2
)
5'2
3
1
3
(1 +r
2
)
3'2
+C.
43. (a) Take n = 2 in Example 6 to get
sin
2
rdr = 3
1
2
cos rsinr +
1
2
1 dr =
r
2
3
sin2r
4
+C.
(b)

sin
4
rdr = 3
1
4
cos rsin
3
r +
3
4
sin
2
rdr = 3
1
4
cos rsin
3
r +
3
8
r 3
3
16
sin2r +C.
45. (a) From Example 6,
sin
n
rdr = 3
1
n
cos r sin
n31
r +
n 31
n
sin
n32
rdr. Using (6),
r'2
0
sin
n
rdr =

3
cos r sin
n31
r
n
r'2
0
+
n 31
n
r'2
0
sin
n32
rdr
= (0 30) +
n 31
n
r'2
0
sin
n32
rdr =
n 31
n
r'2
0
sin
n32
rdr
(b) Using n = 3 in part (a), we have
r'2
0
sin
3
rdr =
2
3
r'2
0
sinrdr =

3
2
3
cos r
r'2
0
=
2
3
.
Using n = 5 in part (a), we have
r'2
0
sin
5
rdr =
4
5
r'2
0
sin
3
rdr =
4
5

2
3
=
8
15
.
(c) The formula holds for n = 1 (that is, 2n + 1 = 3) by (b). Assume it holds for some I D 1. Then
r'2
0
sin
2I+1
rdr =
2 4 6 (2I)
3 5 7 (2I + 1)
. By Example 6,
r'2
0
sin
2I+3
rdr =
2I + 2
2I + 3
r'2
0
sin
2I+1
rdr =
2I + 2
2I + 3

2 4 6 (2I)
3 5 7 (2I + 1)
=
2 4 6 (2I)[2 (I + 1)]
3 5 7 (2I + 1)[2 (I + 1) + 1]
,
so the formula holds for n = I + 1. By induction, the formula holds for all n D 1.
47. Let & = (lnr)
n
, d = dr i d& = n(lnr)
n31
(drr), = r. By Equation 2,
(lnr)
n
dr = r(lnr)
n
3
nr(lnr)
n31
(drr) = r(lnr)
n
3n
(lnr)
n31
dr.
49.

tan
n
rdr =

tan
n32
r tan
2
rdr =

tan
n32
r(sec
2
r 31) dr =

tan
n32
r sec
2
rdr 3
tan
n32
rdr
= 1 3
tan
n32
rdr.
Let & = tan
n32
r, d = sec
2
rdr i d& = (n 32) tan
n33
r sec
2
rdr, = tanr. Then, by Equation 2,
1 = tan
n31
r 3(n 32)
tan
n32
r sec
2
rdr
11 = tan
n31
r 3(n 32)1
(n 31)1 = tan
n31
r
1 =
tan
n31
r
n 31
Returning to the original integral,
tan
n
rdr =
tan
n31
r
n 31
3
tan
n32
rdr.
SECTION 7.1 INTEGRATION BY PARTS

299
51. By repeated applications of the reduction formula in Exercise 47,
(lnr)
3
dr = r(lnr)
3
33
(lnr)
2
dr = r(lnr)
3
33
r(lnr)
2
32
(lnr)
1
dr
= r(lnr)
3
33r(lnr)
2
+ 6
r(lnr)
1
31
(lnr)
0
dr
= r(lnr)
3
33r(lnr)
2
+ 6rlnr 36
1 dr = r(lnr)
3
33r(lnr)
2
+ 6rlnr 36r +C
53. Area =

5
0
rc
30.4i
dr. Let & = r, d = c
30.4i
dr i
d& = dr, = 32.5c
30.4i
. Then
area =

32.5rc
30.4i
5
0
+ 2.5
5
0
c
30.4i
dr
= 312.5c
32
+ 0 + 2.5
32.5c
30.4i
5
0
= 312.5c
32
36.25(c
32
31) = 6.25 318.75c
32
or
25
4
3
75
4
c
32
55. The curves j = rsinr and j = (r 32)
2
intersect at o E 1.04748 and
/ E 2.87307, so
area =

l
a
[rsinr 3(r 32)
2
] dr
=

3rcos r + sinr 3
1
3
(r 32)
3
l
a
[by Example 1]
E 2.81358 30.63075 = 2.18283
57. \ =

1
0
2rcos(r2) dr. Let & = r, d = cos(r2) dr i d& = dr, =
2
r
sin(r2).
\ = 2
rsin
r
2
1
0
32
2
1
0
sin
r
2
dr = 2
30
34
3
2
cos
r
2
1
0
= 4 +
8
(0 31) = 4 3
8
.
59. Volume =

0
31
2(1 3r)c
3i
dr. Let & = 1 3r, d = c
3i
dr i d& = 3dr, = 3c
3i
.
\ = 2
(1 3r)(3c
3i
)
0
31
32
0
31
c
3i
dr = 2
(r 31)(c
3i
) +c
3i
0
31
= 2
rc
3i
0
31
= 2(0 +c) = 2c
61. The average value of 1(r) = r
2
lnr on the interval [1. 3] is 1
ave
=
1
3 31
3
1
r
2
lnrdr =
1
2
1.
Let & = lnr, d = r
2
dr i d& = (1r) dr, =
1
3
r
3
.
So 1 =

1
3
r
3
lnr
3
1
3
3
1
1
3
r
2
dr = (9 ln3 30) 3
1
9
r
3
3
1
= 9 ln3 3
3 3
1
9
= 9 ln3 3
26
9
.
Thus, 1
ave
=
1
2
1 =
1
2
9 ln3 3
26
9
=
9
2
ln3 3
13
9
.
63. Since (t) 0 for all t, the desired distance is c(t) =
I
0
(n) dn =

I
0
n
2
c
3u
dn.
First let & = n
2
, d = c
3u
dn i d& = 2ndn, = 3c
3u
. Then c(t) =
3n
2
c
3u
I
0
+ 2
I
0
nc
3u
dn.
Next let l = n, d\ = c
3u
dn i dl = dn, \ = 3c
3u
. Then
c(t) = 3t
2
c
3I
+ 2
3nc
3u
I
0
+
I
0
c
3u
dn
= 3t
2
c
3I
+ 2
3tc
3I
+ 0 +
3c
3u
I
0
= 3t
2
c
3I
+ 2(3tc
3I
3c
3I
+ 1) = 3t
2
c
3I
32tc
3I
32c
3I
+ 2 = 2 3c
3I
(t
2
+ 2t + 2) meters
65. For 1 =

4
1
r1
00
(r) dr, let & = r, d = 1
00
(r) dr i d& = dr, = 1
0
(r). Then
1 =

r1
0
(r)
4
1
3
4
1
1
0
(r) dr = 41
0
(4) 31 1
0
(1) 3[1(4) 31(1)] = 4 3 31 5 3(7 32) = 12 35 35 = 2.
We used the fact that 1
00
is continuous to guarantee that 1 exists.
300

67. Using the formula for volumes of rotation and the gure, we see that
Volume =

u
0
/
2
dj 3
c
0
o
2
dj 3
u
c
[o(j)]
2
dj = /
2
d 3o
2
c 3
u
c
[o(j)]
2
dj. Let j = 1(r),
which gives dj = 1
0
(r) dr and o(j) = r, so that \ = /
2
d 3 o
2
c 3
l
a
r
2
1
0
(r) dr.
Now integrate by parts with & = r
2
, and d = 1
0
(r) dr i d& = 2rdr, = 1(r), and
l
a
r
2
1
0
(r) dr =

r
2
1(r)
l
a
3
l
a
2r1(r) dr = /
2
1(/) 3o
2
1(o) 3
l
a
2r1(r) dr, but 1(o) = c and 1(/) = d i
\ = /
2
d 3o
2
c 3
/
2
d 3o
2
c 3
l
a
2r1(r) dr
=

l
a
2r1(r) dr.
7.2 Trigonometric Integrals
The symbols
s
= and
c
= indicate the use of the substitutions {& = sinr. d& = cos rdr} and {& = cos r. d& = 3sinrdr}, respectively.
1.

sin
3
r cos
2
rdr =

sin
2
r cos
2
r sinrdr =

(1 3cos
2
r) cos
2
r sinrdr
c
=

(1 3&
2
)&
2
(3d&)
=

(&
2
31)&
2
d& =

(&
4
3&
2
) d& =
1
5
&
5
3
1
3
&
3
+C =
1
5
cos
5
r 3
1
3
cos
3
r +C
3.

3r'4
r'2
sin
5
r cos
3
rdr =

3r'4
r'2
sin
5
r cos
2
r cos rdr =

3r'4
r'2
sin
5
r(1 3sin
2
r) cos rdr
s
=

I
2'2
1
&
5
(1 3&
2
) d&
=

I
2'2
1
(&
5
3&
7
) d& =

1
6
&
6
3
1
8
&
8
I
2'2
1
=

1'8
6
3
1'16
8
1
6
3
1
8
= 3
11
384
5. Let j = r, so dj = dr and
sin
2
(r) cos
5
(r) dr =
1
r
sin
2
j cos
5
j dj =
1
r
sin
2
j cos
4
j cos j dj
=
1
r
sin
2
j (1 3sin
2
j)
2
cos j dj
s
=
1
r
&
2
(1 3&
2
)
2
d& =
1
r
(&
2
32&
4
+&
6
) d&
=
1
r
1
3
&
3
3
2
5
&
5
+
1
7
&
7
+C =
1
3r
sin
3
j 3
2
5r
sin
5
j +
1
7r
sin
7
j +C
=
1
3r
sin
3
(r) 3
2
5r
sin
5
(r) +
1
7r
sin
7
(r) +C
7.

r'2
0
cos
2
0 d0 =

r'2
0
1
2
(1 + cos 20) d0 [half-angle identity]
=
1
2
0 +
1
2
sin20
r'2
0
=
1
2
r
2
+ 0
3(0 + 0)
=
r
4
9.

r
0
sin
4
(3t) dt =

r
0
sin
2
(3t)
2
dt =

r
0
1
2
(1 3cos 6t)
2
dt =
1
4
r
0
(1 32 cos 6t + cos
2
6t) dt
=
1
4
r
0
1 32 cos 6t +
1
2
(1 + cos 12t)
dt =
1
4
r
0
3
2
32 cos 6t +
1
2
cos 12t
dt
=
1
4
3
2
t 3
1
3
sin6t +
1
24
sin12t
r
0
=
1
4
3r
2
30 + 0
3(0 30 + 0)
=
3r
8
11.

(1 + cos 0)
2
d0 =

(1 + 2 cos 0 + cos
2
0) d0 = 0 + 2 sin0 +
1
2
(1 + cos 20) d0
= 0 + 2 sin0 +
1
2
0 +
1
4
sin20 +C =
3
2
0 + 2 sin0 +
1
4
sin20 +C
13.

r'2
0
sin
2
r cos
2
rdr =

r'2
0
1
4
(4 sin
2
r cos
2
r) dr =

r'2
0
1
4
(2 sinr cos r)
2
dr =
1
4
r'2
0
sin
2
2rdr
=
1
4
r'2
0
1
2
(1 3cos 4r) dr =
1
8
r'2
0
(1 3cos 4r) dr =
1
8
r 3
1
4
sin4r
r'2
0
=
1
8
r
2
=
r
16
SECTION 7.2 TRIGONOMETRIC INTEGRALS

301
15.

cos
5
c
I
sinc
dc =

cos
4
c
I
sinc
cos cdc =

1 3sin
2
c
2
I
sinc
cos cdc
s
=

(1 3&
2
)
2
I
&
d&
=

1 32&
2
+&
4
&
1'2
d& =

(&
31'2
32&
3'2
+&
7'2
) d& = 2&
1'2
3
4
5
&
5'2
+
2
9
&
9'2
+C
=
2
45
&
1'2
(45 318&
2
+ 5&
4
) +C =
2
45
I
sinc(45 318 sin
2
c + 5 sin
4
c) +C
17.

cos
2
r tan
3
rdr =

sin
3
r
cos r
dr
c
=

(1 3&
2
)(3d&)
&
=

31
&
+&
d&
= 3ln|&| +
1
2
&
2
+C =
1
2
cos
2
r 3ln|cos r| +C
19.

cos r + sin2r
sinr
dr =

cos r + 2 sinr cos r
sinr
dr =

cos r
sinr
dr +
2 cos rdr
s
=

1
&
d& + 2 sinr
= ln|&| + 2 sinr +C = ln|sinr| + 2 sinr +C
Or: Use the formula
cot rdr = ln|sinr| +C.

21. Let & = tanr, d& = sec
2
rdr. Then
sec
2
r tanrdr =

&d& =
1
2
&
2
+C =
1
2
tan
2
r +C.
Or: Let = sec r, d = sec r tanrdr. Then
sec
2
r tanrdr =

d =
1
2
2
+C =
1
2
sec
2
r +C.
23.

tan
2
rdr =

(sec
2
r 31) dr = tanr 3r +C
25.

sec
6
t dt =

sec
4
t sec
2
t dt =

(tan
2
t + 1)
2
sec
2
t dt =

(&
2
+ 1)
2
d& [& = tant, d& = sec
2
t dt]
=

(&
4
+ 2&
2
+ 1) d& =
1
5
&
5
+
2
3
&
3
+& +C =
1
5
tan
5
t +
2
3
tan
3
t + tant +C
27.

r'3
0
tan
5
r sec
4
rdr =

r'3
0
tan
5
r(tan
2
r + 1) sec
2
rdr =

I
3
0
&
5
(&
2
+ 1) d& [& = tanr, d& = sec
2
rdr]
=

I
3
0
(&
7
+&
5
) d& =

1
8
&
8
+
1
6
&
6
I
3
0
=
81
8
+
27
6
=
81
8
+
9
2
=
81
8
+
36
8
=
117
8
Alternate solution:
r'3
0
tan
5
r sec
4
rdr =

r'3
0
tan
4
r sec
3
r sec r tanrdr =

r'3
0
(sec
2
r 31)
2
sec
3
r sec r tanrdr
=

2
1
(&
2
31)
2
&
3
d& [& = sec r, d& = sec r tanrdr] =

2
1
(&
4
32&
2
+ 1)&
3
d&
=

2
1
(&
7
32&
5
+&
3
) d& =

1
8
&
8
3
1
3
&
6
+
1
4
&
4
2
1
=

32 3
64
3
+ 4
1
8
3
1
3
+
1
4
=
117
8
29.

tan
3
r sec rdr =

tan
2
r sec r tanrdr =

(sec
2
r 31) sec r tanrdr
=

(&
2
31) d& [& = sec r, d& = sec r tanrdr] =
1
3
&
3
3& +C =
1
3
sec
3
r 3sec r +C
31.

tan
5
rdr =

(sec
2
r 31)
2
tanrdr =

sec
4
r tanrdr 32
sec
2
r tanrdr +
tanrdr
=

sec
3
r sec r tanrdr 32
tanr sec
2
rdr +
tanrdr
=
1
4
sec
4
r 3tan
2
r + ln|sec r| +C [or
1
4
sec
4
r 3sec
2
r + ln|sec r| +C]
33.

tan
3
0
cos
4
0
d0 =

tan
3
0 sec
4
0 d0 =

tan
3
0 (tan
2
0 + 1) sec
2
0 d0
=

&
3
(&
2
+ 1) d& [& = tan0, d& = sec
2
0 d0]
=

(&
5
+&
3
) d& =
1
6
&
6
+
1
4
&
4
+C =
1
6
tan
6
0 +
1
4
tan
4
0 +C
302

35. Let & = r. d = sec r tanrdr i d& = dr. = sec r. Then
r sec r tanrdr = rsec r 3
sec rdr = rsec r 3ln|sec r + tanr| +C.

37.

r'2
r'6
cot
2
rdr =

r'2
r'6
(csc
2
r 31) dr =

3cot r 3r
r'2
r'6
=

0 3
r
2
3
I
3 3
r
6
=
I
3 3
r
3
39.

cot
3
c csc
3
cdc =

cot
2
c csc
2
c csc c cot cdc =

(csc
2
c 31) csc
2
c csc c cot cdc
=

(&
2
31)&
2
(3d&) [& = csc c, d& = 3csc ccot cdc]
=

(&
2
3&
4
) d& =
1
3
&
3
3
1
5
&
5
+C =
1
3
csc
3
c 3
1
5
csc
5
c +C
41. 1 =

csc rdr =

csc r(csc r 3cot r)
csc r 3cot r
dr =

3csc r cot r + csc
2
r
csc r 3cot r
dr. Let & = csc r 3cot r i
d& = (3csc r cot r + csc
2
r) dr. Then 1 =

d&& = ln|&| = ln|csc r 3cot r| +C.
43.

sin8r cos 5rdr
2a
=

1
2
[sin(8r 35r) + sin(8r + 5r)] dr =
1
2
sin3rdr +
1
2
sin13rdr
= 3
1
6
cos 3r 3
1
26
cos 13r +C
45.

sin50 sin0 d0
2b
=

1
2
[cos(50 30) 3cos(50 +0)] d0 =
1
2
cos 40 d0 3
1
2
cos 60 d0 =
1
8
sin40 3
1
12
sin60 +C
47.

1 3tan
2
r
sec
2
r
dr =

cos
2
r 3sin
2
r
dr =

cos 2rdr =
1
2
sin2r +C
49. Let & = tan(t
2
) i d& = 2t sec
2
(t
2
) dt. Then
t sec
2
(t
2
) tan
4
(t
2
) dt =

&
4
1
2
d&
=
1
10
&
5
+C =
1
10
tan
5
(t
2
) +C.
In Exercises 5154, let 1(r) denote the integrand and 1(r) its antiderivative (with C = 0).
51. Let & = r
2
, so that d& = 2rdr. Then
rsin
2
(r
2
) dr =

sin
2
&
1
2
d&
=
1
2
1
2
(1 3cos 2&) d&
=
1
4
& 3
1
2
sin2&
+C =
1
4
& 3
1
4
1
2
2 sin& cos &
+C
=
1
4
r
2
3
1
4
sin(r
2
) cos(r
2
) +C
We see from the graph that this is reasonable, since 1 increases where 1 is positive and 1 decreases where 1 is negative.
Note also that 1 is an odd function and 1 is an even function.
53.

sin3r sin6rdr =

1
2
[cos(3r 36r) 3cos(3r + 6r)] dr
=
1
2
(cos 3r 3cos 9r) dr

=
1
6
sin3r 3
1
18
sin9r +C
Notice that 1(r) = 0 whenever 1 has a horizontal tangent.
55. 1
ave
=
1
2r
r
3r
sin
2
r cos
3
rdr =
1
2r
r
3r
sin
2
r(1 3sin
2
r) cos rdr
=
1
2r
0
0
&
2
(1 3&
2
) d& [where & = sinr]
= 0
SECTION 7.2 TRIGONOMETRIC INTEGRALS

303
57. =

r'4
3r'4
(cos
2
r 3sin
2
r) dr =

r'4
3r'4
cos 2rdr
= 2
r'4
0
cos 2rdr = 2
1
2
sin2r
r'4
0
=

sin2r
r'4
0
= 1 30 = 1
59. It seems from the graph that
2r
0
cos
3
rdr = 0, since the area below the
r-axis and above the graph looks about equal to the area above the axis and
below the graph. By Example 1, the integral is
sinr 3
1
3
sin
3
r
2r
0
= 0.
Note that due to symmetry, the integral of any odd power of sinr or cos r
between limits which differ by 2n (n any integer) is 0.
61. Using disks, \ =

r
r'2
sin
2
rdr =
r
r'2
1
2
(1 3cos 2r) dr =
1
2
r 3
1
4
sin2r
r
r'2
=
r
2
30 3
r
4
+ 0
=
r
2
4
63. Using washers,
\ =

r'4
0

(1 3sinr)
2
3(1 3cos r)
2
dr
=
r'4
0
(1 32 sinr + sin
2
r) 3(1 32 cos r + cos
2
r)
dr
=
r'4
0
(2 cos r 32 sinr + sin
2
r 3cos
2
r) dr
=
r'4
0
(2 cos r 32 sinr 3cos 2r) dr =
2 sinr + 2 cos r 3
1
2
sin2r
r'4
0
=
I
2 +
I
2 3
1
2
3(0 + 2 30)
2
I
2 3
5
2
65. c = 1(t) =

I
0
sin.& cos
2
.&d&. Let j = cos .& i dj = 3. sin.&d&. Then
c = 3
1
.
cos .I
1
j
2
dj = 3
1
.
1
3
j
3
cos .I
1
=
1
3.
(1 3cos
3
.t).
67. Just note that the integrand is odd [1(3r) = 31(r)].
Or: If i 6= n, calculate
r
3r
sinir cos nrdr =

r
3r
1
2
[sin(i3n)r + sin(i+n)r] dr =
1
2
3
cos(i3n)r
i3n
3
cos(i+n)r
i+n
r
3r
= 0
If i = n, then the rst term in each set of brackets is zero.
69.

r
3r
cos ir cos nrdr =

r
3r
1
2
[cos(i3n)r + cos(i+n)r] dr.
If i 6= n, this is equal to
1
2
sin(i3n)r
i3n
+
sin(i+n)r
i+n
r
3r
= 0.
If i = n, we get
r
3r
1
2
[1 + cos(i+n)r] dr =

1
2
r
r
3r
+
sin(i+n)r
2(i+n)
r
3r
= + 0 = .
304

7.3 Trigonometric Substitution
1. Let r = 3 sec 0, where 0 $ 0 <
r
2
or $ 0 <
3r
2
. Then
dr = 3 sec 0 tan0 d0 and
I
r
2
39 =
I
9 sec
2
0 39 =

9(sec
2
0 31) =
I
9 tan
2
0
= 3 |tan0| = 3 tan0 for the relevant values of 0.
1
r
2
I
r
2
39
dr =

1
9 sec
2
0 3 tan0
3 sec 0 tan0 d0 =
1
9
cos 0 d0 =
1
9
sin0 +C =
1
9
I
r
2
39
r
+C
Note that 3sec(0 +) = sec 0, so the gure is sufcient for the case $ 0 <
3r
2
.
3. Let r = 3 tan0, where 3
r
2
< 0 <
r
2
. Then dr = 3 sec
2
0 d0 and
I
r
2
+ 9 =
I
9 tan
2
0 + 9 =

9(tan
2
0 + 1) =
I
9 sec
2
0
= 3 |sec 0| = 3 sec 0 for the relevant values of 0.
r
3
I
r
2
+ 9
dr =

3
3
tan
3
0
3 sec 0
3 sec
2
0 d0 = 3
3
tan
3
0 sec 0d0 = 3
3
tan
2
0 tan0 sec 0 d0
= 3
3
(sec
2
0 31) tan0 sec 0 d0 = 3
3
(&
2
31) d& [& = sec 0, d& = sec 0 tan0 d0]
= 3
3
1
3
&
3
3&
+C = 3
3
1
3
sec
3
0 3sec 0
+C = 3
3
1
3
r
2
+ 9
3'2
3
3
3
I
r
2
+ 9
3
+C
=
1
3
(r
2
+ 9)
3'2
39
I
r
2
+ 9 +C or
1
3
(r
2
318)
I
r
2
+ 9 +C
5. Let t = sec 0, so dt = sec 0 tan0 d0, t =
I
2 i 0 =
r
4
, and t = 2 i 0 =
r
3
. Then
2
I
2
1
t
3
I
t
2
31
dt =

r'3
r'4
1
sec
3
0 tan0
sec 0 tan0 d0 =

r'3
r'4
1
sec
2
0
d0 =

r'3
r'4
cos
2
0 d0
=

r'3
r'4
1
2
(1 + cos 20) d0 =
1
2
0 +
1
2
sin20
r'3
r'4
=
1
2
r
3
+
1
2
I
3
2
r
4
+
1
2
1
=
1
2
r
12
+
I
3
4
3
1
2
=
r
24
+
I
3
8
3
1
4
7. Let r = 5 sin0, so dr = 5 cos 0 d0. Then
1
r
2
I
25 3r
2
dr =

1
5
2
sin
2
0 5 cos 0
5 cos 0 d0 =
1
25
csc
2
0 d0
= 3
1
25
cot 0 +C = 3
1
25
I
25 3r
2
r
+C
SECTION 7.3 TRIGONOMETRIC SUBSTITUTION

305
9. Let r = 4 tan0, where 3
r
2
< 0 <
r
2
. Then dr = 4 sec
2
0 d0 and
I
r
2
+ 16 =
I
16 tan
2
0 + 16 =

16(tan
2
0 + 1)
=
I
16 sec
2
0 = 4 |sec 0|
= 4 sec 0 for the relevant values of 0.
dr
I
r
2
+ 16
=

4 sec
2
0 d0
4 sec 0
=

sec 0 d0 = ln|sec 0 + tan0| +C
1
= ln
I
r
2
+ 16
4
+
r
4
+C
1
= ln
I
r
2
+ 16 +r
3ln|4| +C
1
= ln
I
r
2
+ 16 +r
+C, where C = C
1
3ln4.
(Since
I
r
2
+ 16 +r 0, we dont need the absolute value.)
11. Let 2r = sin0, where 3
r
2
$ 0 $
r
2
. Then r =
1
2
sin0, dr =
1
2
cos 0 d0,
and
I
1 34r
2
=

1 3(2r)
2
= cos 0.
I
1 34r
2
dr =

cos 0
1
2
cos 0
d0 =
1
4
(1 + cos 20) d0
=
1
4
0 +
1
2
sin20
+C =
1
4
(0 + sin0 cos 0) +C
=
1
4
sin
31
(2r) + 2r
I
1 34r
2
+C
13. Let r = 3 sec 0, where 0 $ 0 <
r
2
or $ 0 <
3r
2
. Then
dr = 3 sec 0 tan0 d0 and
I
r
2
39 = 3 tan0, so
I
r
2
39
r
3
dr =

3 tan0
27 sec
3
0
3 sec 0 tan0 d0 =
1
3
tan
2
0
sec
2
0
d0
=
1
3
sin
2
0 d0 =
1
3
1
2
(1 3cos 20) d0 =
1
6
0 3
1
12
sin20 +C =
1
6
0 3
1
6
sin0 cos 0 +C
=
1
6
sec
31
r
3
3
1
6
I
r
2
39
r
3
r
+C =
1
6
sec
31
r
3
3
I
r
2
39
2r
2
+C
15. Let r = o sin0, dr = o cos 0 d0, r = 0 i 0 = 0 and r = o i 0 =
r
2
. Then
a
0
r
2
I
o
2
3r
2
dr =

r'2
0
o
2
sin
2
0 (o cos 0) o cos 0 d0 = o
4
r'2
0
sin
2
0 cos
2
0 d0 = o
4
r'2
0
1
2
(2 sin0 cos 0)
2
d0
=
o
4
4
r'2
0
sin
2
20 d0 =
o
4
4
r'2
0
1
2
(1 3cos 40) d0 =
o
4
8
0 3
1
4
sin40
r'2
0
=
o
4
8
2
30
30
=

16
o
4
17. Let & = r
2
37, so d& = 2rdr. Then
r
I
r
2
37
dr =
1
2
1
I
&
d& =
1
2
2
I
& +C =

r
2
37 +C.
306

19. Let r = tan0, where 3
r
2
< 0 <
r
2
. Then dr = sec
2
0 d0
and
I
1 +r
2
= sec 0, so
I
1 +r
2
r
dr =

sec 0
tan0
sec
2
0 d0 =

sec 0
tan0
(1 + tan
2
0) d0
=

(csc 0 + sec 0 tan0) d0
= ln|csc 0 3cot 0| + sec 0 +C [by Exercise 7.2.41]
= ln
I
1 +r
2
r
3
1
r
+
I
1 +r
2
1
+C = ln
I
1 +r
2
31
r
+
I
1 +r
2
+C
21. Let r =
3
5
sin0, so dr =
3
5
cos 0 d0, r = 0 i 0 = 0, and r = 0.6 i 0 =
r
2
. Then
0.6
0
r
2
I
9 325r
2
dr =

r'2
0
3
5
2
sin
2
0
3 cos 0
3
5
cos 0 d0
=
9
125
r'2
0
sin
2
0 d0
=
9
125
r'2
0
1
2
(1 3cos 20) d0 =
9
250
0 3
1
2
sin20
r'2
0
=
9
250
r
2
30
30
=
9
500
23. 5 + 4r 3r
2
= 3(r
2
34r + 4) + 9 = 3(r 32)
2
+ 9. Let
r 32 = 3 sin0, 3
r
2
$ 0 $
r
2
, so dr = 3 cos 0 d0. Then
I
5 + 4r 3r
2
dr =

9 3(r 32)
2
dr =

9 39 sin
2
0 3 cos 0 d0
=
I
9 cos
2
0 3 cos 0 d0 =

9 cos
2
0 d0
=
9
2
(1 + cos 20) d0 =
9
2
0 +
1
2
sin20
+C
=
9
2
0 +
9
4
sin20 +C =
9
2
0 +
9
4
(2 sin0 cos 0) +C
=
9
2
sin
31
r 32
3
+
9
2

r 32
3

I
5 + 4r 3r
2
3
+C
=
9
2
sin
31
r 32
3
+
1
2
(r 32)
I
5 + 4r 3r
2
+C
25. r
2
+r + 1 =

r
2
+r +
1
4
+
3
4
=

r +
1
2
2
+
I
3
2
2
. Let
r +
1
2
=
I
3
2
tan0, so dr =
I
3
2
sec
2
0 d0 and
I
r
2
+r + 1 =
I
3
2
sec 0.
Then
[continued]

307
r
I
r
2
+r + 1
dr =

I
3
2
tan0 3
1
2
I
3
2
sec 0
I
3
2
sec
2
0 d0
=

I
3
2
tan0 3
1
2
sec 0 d0 =

I
3
2
tan0 sec 0 d0 3
1
2
sec 0 d0
=
I
3
2
sec 0 3
1
2
ln|sec 0 + tan0| +C
1
=
I
r
2
+r + 1 3
1
2
ln
2
I
3
I
r
2
+r + 1 +
2
I
3
r +
1
2
+C
1
=
I
r
2
+r + 1 3
1
2
ln
2
I
3
I
r
2
+r + 1 +
r +
1
2
+C
1
=
I
r
2
+r + 1 3
1
2
ln
2
I
3
3
1
2
ln
I
r
2
+r + 1 +r +
1
2
+C
1
=
I
r
2
+r + 1 3
1
2
ln
I
r
2
+r + 1 +r +
1
2
+C, where C = C
1
3
1
2
ln
2
I
3
27. r
2
+ 2r = (r
2
+ 2r + 1) 31 = (r + 1)
2
31. Let r + 1 = 1 sec 0,
so dr = sec 0 tan0 d0 and
I
r
2
+ 2r = tan0. Then
I
r
2
+ 2rdr =

tan0 (sec 0 tan0 d0) =

tan
2
0 sec 0 d0
=

(sec
2
0 31) sec 0 d0 =

sec
3
0 d0 3
sec 0 d0
=
1
2
sec 0 tan0 +
1
2
ln|sec 0 + tan0| 3ln|sec 0 + tan0| +C
=
1
2
sec 0 tan0 3
1
2
ln|sec 0 + tan0| +C =
1
2
(r + 1)
I
r
2
+ 2r 3
1
2
ln
r + 1 +
I
r
2
+ 2r
+C
29. Let & = r
2
, d& = 2rdr. Then
r
I
1 3r
4
dr =
I
1 3&
2
1
2
d&
=
1
2
cos 0 cos 0 d0

where u = sin 0, Ju = cos 0 J0,
and
s
1 u
2
= cos 0
=
1
2
1
2
(1 + cos 20) d0 =
1
4
0 +
1
8
sin20 +C =
1
4
0 +
1
4
sin0 cos 0 +C
=
1
4
sin
31
& +
1
4
&
I
1 3&
2
+C =
1
4
sin
31
(r
2
) +
1
4
r
2
I
1 3r
4
+C
31. (a) Let r = o tan0, where 3
r
2
< 0 <
r
2
. Then
I
r
2
+o
2
= o sec 0 and
dr
I
r
2
+o
2
=

o sec
2
0 d0
o sec 0
=

1
= ln
I
r
2
+o
2
o
+
r
o
+C
1
= ln
r +
I
r
2
+o
2
+C where C = C
1
3ln|o|
(b) Let r = o sinht, so that dr = o cosht dt and
I
r
2
+o
2
= o cosht. Then
dr
I
r
2
+o
2
=

o cosht dt
o cosht
= t +C = sinh
31
r
o
+C.
33. The average value of 1(r) =
I
r
2
31r on the interval [1. 7] is
1
7 31
7
1
I
r
2
31
r
dr =
1
6
o
0
tan0
sec 0
sec 0 tan0 d0

where r = sec 0, Jr = sec 0 tan 0 J0,
s
r
2
1 = tan 0, and o = sec
1
7
=
1
6
o
0
tan
2
0 d0 =
1
6
o
0
(sec
2
0 31) d0 =
1
6
tan0 30
o
0
=
1
6
(tanc 3c) =
1
6
I
48 3sec
31
7
308

35. Area of 41OQ =
1
2
(v cos 0)(v sin0) =
1
2
v
2
sin0 cos 0. Area of region 1Q1 =

r
r cos 0
I
v
2
3r
2
dr.
Let r = v cos & i dr = 3v sin&d& for 0 $ & $
r
2
. Then we obtain
I
v
2
3r
2
dr =

v sin&(3v sin&) d& = 3v
2
sin
2
&d& = 3
1
2
v
2
(& 3sin& cos &) +C
= 3
1
2
v
2
cos
31
(rv) +
1
2
r
I
v
2
3r
2
+C
so area of region 1Q1=
1
2
3v
2
cos
31
(rv) +r
I
v
2
3r
2
r
r cos 0
=
1
2
0 3(3v
2
0 +v cos 0 v sin0)
=
1
2
v
2
0 3
1
2
v
2
sin0 cos 0
and thus, (area of sector 1O1) = (area of 41OQ) + (area of region 1Q1) =
1
2
v
2
0.
37. From the graph, it appears that the curve j = r
2
I
4 3r
2
and the
line j = 2 3r intersect at about r = o E 0.81 and r = 2, with
r
2
I
4 3r
2
2 3r on (o. 2). So the area bounded by the curve and the line is
E
2
a
r
2
I
4 3r
2
3(2 3r)
dr =

2
a
r
2
I
4 3r
2
dr 3
2r 3
1
2
r
2
2
a
.
To evaluate the integral, we put r = 2 sin0, where 3
r
2
$ 0 $
r
2
. Then
dr = 2 cos 0 d0, r = 2 i 0 = sin
31
1 =
r
2
, and r = o i 0 = c = sin
31
(o2) E 0.416. So
2
a
r
2
I
4 3r
2
dr E
r'2
o
4 sin
2
0 (2 cos 0)(2 cos 0 d0) = 4
r'2
o
sin
2
20 d0 = 4
r'2
o
1
2
(1 3cos 40) d0
= 2
0 3
1
4
sin40
r'2
o
= 2
r
2
30
c 3
1
4
(0.996)
E 2.81
Thus, E 2.81 3
2 2 3
1
2
2
2
2o 3
1
2
o
2
E 2.10.
39. (a) Let t = o sin0, dt = o cos 0 d0, t = 0 i 0 = 0 and t = r i 0 = sin
31
(ro). Then
i
0
o
2
3t
2
dt =

sin
1
(i'a)
0
o cos 0 (o cos 0 d0)
= o
2
sin
1
(i'a)
0
cos
2
0 d0 =
o
2
2
sin
1
(i'a)
0
(1 + cos 20) d0
=
o
2
2
0 +
1
2
sin20
sin
1
(i'a)
0
=
o
2
2
0 + sin0 cos 0
sin
1
(i'a)
0
=
o
2
2
sin
31
r
o
+
r
o

I
o
2
3r
2
o
30
=
1
2
o
2
sin
31
(ro) +
1
2
r
I
o
2
3r
2
(b) The integral
i
0
I
o
2
3t
2
dt represents the area under the curve j =
I
o
2
3t
2
between the vertical lines t = 0 and t = r.
The gure shows that this area consists of a triangular region and a sector of the circle t
2
+j
2
= o
2
. The triangular region
has base r and height
I
o
2
3r
2
, so its area is
1
2
r
I
o
2
3r
2
. The sector has area
1
2
o
2
0 =
1
2
o
2
sin
31
(ro).

309
41. Let the equation of the large circle be r
2
+j
2
= 1
2
. Then the equation of
the small circle is r
2
+(j 3/)
2
= v
2
, where / =
I
1
2
3v
2
is the distance
between the centers of the circles. The desired area is
=

r
3r
/ +
I
v
2
3r
2
3
I
1
2
3r
2
dr
= 2
r
0
/ +
I
v
2
3r
2
3
I
1
2
3r
2
dr
= 2
r
0
/ dr + 2
r
0
I
v
2
3r
2
dr 32
r
0
I
1
2
3r
2
dr
The rst integral is just 2/v = 2v
I
1
2
3v
2
. The second integral represents the area of a quarter-circle of radius v, so its value
is
1
4
v
2
. To evaluate the other integral, note that
I
o
2
3r
2
dr =

o
2
cos
2
0 d0 [r = o sin0, dr = o cos 0 d0] =

1
2
o
2
(1 + cos 20) d0
=
1
2
o
2
0 +
1
2
sin20
+C =
1
2
o
2
(0 + sin0 cos 0) +C
=
o
2
2
arcsin
r
o
+
o
2
2
r
o
I
o
2
3r
2
o
+C =
o
2
2
arcsin
r
o
+
r
2
I
o
2
3r
2
+C
Thus, the desired area is
= 2v
I
1
2
3v
2
+ 2
1
4
v
2
1
2
arcsin(r1) +r
I
1
2
3r
2
r
0
= 2v
I
1
2
3v
2
+
1
2
v
2
3
1
2
arcsin(v1) +v
I
1
2
3v
2
= v
I
1
2
3v
2
+
r
2
v
2
31
2
arcsin(v1)
43. We use cylindrical shells and assume that 1 v. r
2
= v
2
3(j 31)
2
i r =
v
2
3(j 31)
2
,
so o(j) = 2
v
2
3(j 31)
2
and
\ =

T+r
T3r
2j 2
v
2
3(j 31)
2
dj =

r
3r
4(& +1)
I
v
2
3&
2
d& [where u = u 1]
= 4
r
3r
&
I
v
2
3&
2
d& + 41
r
3r
I
v
2
3&
2
d&

where u = : sin 0 , Ju = : cos 0 J0
in the second integral
= 4
3
1
3
(v
2
3&
2
)
3'2
r
3r
+ 41
r'2
3r'2
v
2
cos
2
0 d0 = 3
4r
3
(0 30) + 41v
2
r'2
3r'2
cos
2
0 d0
= 21v
2
r'2
3r'2
(1 + cos 20) d0 = 21v
2
0 +
1
2
sin20
r'2
3r'2
= 2
2
1v
2
Another method: Use washers instead of shells, so \ = 81
r
0
v
2
3j
2
dj as in Exercise 6.2.63(a), but evaluate the
integral using j = v sin0.
310

7.4 Integration of Rational Functions by Partial Fractions
1. (a)
2r
(r + 3)(3r + 1)
=

r + 3
+
1
3r + 1
(b)
1
r
3
+ 2r
2
+r
=
1
r(r
2
+ 2r + 1)
=
1
r(r + 1)
2
=

r
+
1
r + 1
+
C
(r + 1)
2
3. (a)
r
4
+ 1
r
5
+ 4r
3
=
r
4
+ 1
r
3
(r
2
+ 4)
=

r
+
1
r
2
+
C
r
3
+
1r +1
r
2
+ 4
(b)
1
(r
2
39)
2
=
1
[(r + 3)(r 33)]
2
=
1
(r + 3)
2
(r 33)
2
=

r + 3
+
1
(r + 3)
2
+
C
r 33
+
1
(r 33)
2
5. (a)
r
4
r
4
31
=
(r
4
31) + 1
r
4
31
= 1 +
1
r
4
31
[or use long division] = 1 +
1
(r
2
31)(r
2
+ 1)
= 1 +
1
(r 31)(r + 1)(r
2
+ 1)
= 1 +

r 31
+
1
r + 1
+
Cr +1
r
2
+ 1
(b)
t
4
+t
2
+ 1
(t
2
+ 1)(t
2
+ 4)
2
=
t +1
t
2
+ 1
+
Ct +1
t
2
+ 4
+
1t +1
(t
2
+ 4)
2
7.

r
r 36
dr =

(r 36) + 6
r 36
dr =

1 +
6
r 36
dr = r + 6 ln|r 36| +C
9.
r 39
(r + 5)(r 32)
=

r + 5
+
1
r 32
. Multiply both sides by (r + 5)(r 32) to get r 39 = (r 32) +1(r + 5)(W), or
equivalently, r 39 = (+1)r 32+ 51. Equating coefcients of r on each side of the equation gives us 1 = +1 (1)
and equating constants gives us 39 = 32+ 51 (2). Adding two times (1) to (2) gives us 37 = 71 C 1 = 31 and
hence, = 2. [Alternatively, to nd the coefcients and 1, we may use substitution as follows: substitute 2 for r in (W) to
get 37 = 71 C 1 = 31, then substitute 35 for r in (W) to get 314 = 37 C = 2.] Thus,
r 39
(r + 5)(r 32)
dr =

2
r + 5
+
31
r 32
dr = 2 ln|r + 5| 3ln|r 32| +C.

11.
1
r
2
31
=
1
(r + 1)(r 31)
=

r + 1
+
1
r 31
. Multiply both sides by (r + 1)(r 31) to get 1 = (r 31) +1(r + 1).
Substituting 1 for r gives 1 = 21 C 1 =
1
2
. Substituting 31 for r gives 1 = 32 C = 3
1
2
. Thus,
3
2
1
r
2
31
dr =

3
2
312
r + 1
+
12
r 31
dr =

3
1
2
ln|r + 1| +
1
2
ln|r 31|
3
2
=

3
1
2
ln4 +
1
2
ln2
3
1
2
ln3 +
1
2
ln1
=
1
2
(ln2 + ln3 3ln4)

or
1
2
ln
3
2
13.

or
r
2
3/r
dr =

or
r(r 3/)
dr =

o
r 3/
dr = o ln|r 3/| +C
SECTION 7.4 INTEGRATION OF RATIONAL FUNCTIONS BY PARTIAL FRACTIONS

311
15.
r
3
32r
2
34
r
3
32r
2
= 1 +
34
r
2
(r 32)
. Write
34
r
2
(r 32)
=

r
+
1
r
2
+
C
r 32
. Multiplying both sides by r
2
(r 32) gives
34 = r(r 32) +1(r 32) +Cr
2
. Substituting 0 for r gives 34 = 321 C 1 = 2. Substituting 2 for r gives
34 = 4C C C = 31. Equating coefcients of r
2
, we get 0 = +C, so = 1. Thus,
4
3
r
3
32r
2
34
r
3
32r
2
dr =

4
3
1 +
1
r
+
2
r
2
3
1
r 32
dr =

r + ln|r| 3
2
r
3ln|r 32|
4
3
=

4 + ln4 3
1
2
3ln2
3 + ln3 3
2
3
30
=
7
6
+ ln
2
3
17.
4j
2
37j 312
j(j + 2)(j 33)
=

j
+
1
j + 2
+
C
j 33
i 4j
2
37j 312 = (j + 2)(j 33) +1j(j 33) +Cj(j + 2). Setting
j = 0 gives 312 = 36, so = 2. Setting j = 32 gives 18 = 101, so 1 =
9
5
. Setting j = 3 gives 3 = 15C, so C =
1
5
.
Now
2
1
4j
2
37j 312
j(j + 2)(j 33)
dj =

2
1
2
j
+
95
j + 2
+
15
j 33
dj =

2 ln|j| +
9
5
ln|j + 2| +
1
5
ln|j 33|
2
1
= 2 ln2 +
9
5
ln4 +
1
5
ln1 32 ln1 3
9
5
ln3 3
1
5
ln2
= 2 ln2 +
18
5
ln2 3
1
5
ln2 3
9
5
ln3 =
27
5
ln2 3
9
5
ln3 =
9
5
(3 ln2 3ln3) =
9
5
ln
8
3
19.
1
(r + 5)
2
(r 31)
=

r + 5
+
1
(r + 5)
2
+
C
r 31
i 1 = (r + 5)(r 31) +1(r 31) +C(r + 5)
2
.
Setting r = 35 gives 1 = 361, so 1 = 3
1
6
. Setting r = 1 gives 1 = 36C, so C =
1
36
. Setting r = 32 gives
1 = (3)(33) +1(33) +C
3
2
= 39331 + 9C = 39+
1
2
+
1
4
= 39+
3
4
, so 9 = 3
1
4
and = 3
1
36
. Now
1
(r + 5)
2
(r 31)
dr =

3136
r + 5
3
16
(r + 5)
2
+
136
r 31
dr = 3
1
36
ln|r + 5| +
1
6(r + 5)
+
1
36
ln|r 31| +C.
21. r
r
2
+ 4
r
3
+ 0r
2
+ 0r + 4
r
3
+ 4r
34r + 4
By long division,
r
3
+ 4
r
2
+ 4
= r +
34r + 4
r
2
+ 4
. Thus,
r
3
+ 4
r
2
+ 4
dr =

r +
34r + 4
r
2
+ 4
dr =

r 3
4r
r
2
+ 4
+
4
r
2
+ 2
2
dr
=
1
2
r
2
34
1
2
ln
r
2
+ 4
+ 4
1
2
tan
31
r
2
+C =
1
2
r
2
32 ln(r
2
+ 4) + 2 tan
31
r
2
+C
23.
5r
2
+ 3r 32
r
3
+ 2r
2
=
5r
2
+ 3r 32
r
2
(r + 2)
=

r
+
1
r
2
+
C
r + 2
. Multiply by r
2
(r + 2) to
get 5r
2
+ 3r 3 2 = r(r + 2) + 1(r + 2) + Cr
2
. Set r = 32 to get C = 3, and take
r = 0 to get 1 = 31. Equating the coefcients of r
2
gives 5 = +C i = 2. So
5r
2
+ 3r 32
r
3
+ 2r
2
dr =

2
r
3
1
r
2
+
3
r + 2
dr = 2 ln|r| +
1
r
+ 3 ln|r + 2| +C.
312

25.
10
(r 31)(r
2
+ 9)
=

r 31
+
1r +C
r
2
+ 9
. Multiply both sides by (r 31)
r
2
+ 9
to get
10 =
r
2
+ 9
+ (1r +C)(r 31) (-). Substituting 1 for r gives 10 = 10 C = 1. Substituting 0 for r gives
10 = 93C i C = 9(1) 310 = 31. The coefcients of the r
2
-terms in (-) must be equal, so 0 = +1 i
1 = 31. Thus,
10
(r 31)(r
2
+ 9)
dr =

1
r 31
+
3r 31
r
2
+ 9
dr =

1
r 31
3
r
r
2
+ 9
3
1
r
2
+ 9
dr
= ln|r 31| 3
1
2
ln(r
2
+ 9) 3
1
3
tan
31
i
3
+C
In the second term we used the substitution & = r
2
+ 9 and in the last term we used Formula 10.
27.
r
3
+r
2
+ 2r + 1
(r
2
+ 1)(r
2
+ 2)
=
r +1
r
2
+ 1
+
Cr +1
r
2
+ 2
. Multiply both sides by
r
2
+ 1
r
2
+ 2
to get
r
3
+r
2
+ 2r + 1 = (r +1)
r
2
+ 2
+ (Cr +1)
r
2
+ 1
C
r
3
+r
2
+ 2r + 1 =

r
3
+1r
2
+ 2r + 21
Cr
3
+1r
2
+Cr +1
C
r
3
+r
2
+ 2r + 1 = (+C)r
3
+ (1 +1)r
2
+ (2+C)r + (21 +1). Comparing coefcients gives us the following
system of equations:
+C = 1 (1) 1 +1 = 1 (2)
2+C = 2 (3) 21 +1 = 1 (4)
Subtracting equation (1) from equation (3) gives us = 1, so C = 0. Subtracting equation (2) from equation (4) gives us
1 = 0, so 1 = 1. Thus, 1 =

r
3
+r
2
+ 2r + 1
(r
2
+ 1)(r
2
+ 2)
dr =

r
r
2
+ 1
+
1
r
2
+ 2
dr. For
r
r
2
+ 1
dr, let & = r
2
+ 1
so d& = 2rdr and then
r
r
2
+ 1
dr =
1
2
1
&
d& =
1
2
ln|&| +C =
1
2
ln
r
2
+ 1
+C. For
1
r
2
+ 2
dr, use
Formula 10 with o =
I
2. So
1
r
2
+ 2
dr =

1
r
2
+
I
2
2
dr =
1
I
2
tan
31
r
I
2
+C.
Thus, 1 =
1
2
ln
r
2
+ 1
+
1
I
2
tan
31
r
I
2
+C.
29.

r + 4
r
2
+ 2r + 5
dr =

r + 1
r
2
+ 2r + 5
dr +
3
r
2
+ 2r + 5
dr =
1
2
(2r + 2) dr
r
2
+ 2r + 5
+
3 dr
(r + 1)
2
+ 4
=
1
2
ln
r
2
+ 2r + 5
+ 3
2 d&
4(&
2
+ 1)
where r + 1 = 2u,
and Jr = 2 Ju
=
1
2
ln(r
2
+ 2r + 5) +
3
2
tan
31
& +C =
1
2
ln(r
2
+ 2r + 5) +
3
2
tan
31
r + 1
2
+C
31.
1
r
3
31
=
1
(r 31)(r
2
+r + 1)
=

r 31
+
1r +C
r
2
+r + 1
i 1 =
r
2
+r + 1
+ (1r +C)(r 31).

Take r = 1 to get =
1
3
. Equating coefcients of r
2
and then comparing the constant terms, we get 0 =
1
3
+1, 1 =
1
3
3C,

313
so 1 = 3
1
3
, C = 3
2
3
i
1
r
3
31
dr =

1
3
r 31
dr +
3
1
3
r 3
2
3
r
2
+r + 1
dr =
1
3
ln|r 31| 3
1
3
r + 2
r
2
+r + 1
dr
=
1
3
ln|r 31| 3
1
3
r + 12
r
2
+r + 1
dr 3
1
3
(32) dr
(r + 12)
2
+ 34
=
1
3
ln|r 31| 3
1
6
ln
r
2
+r + 1
3
1
2
2
I
3
tan
31
r +
1
2
I
3
+1
=
1
3
ln|r 31| 3
1
6
ln(r
2
+r + 1) 3
1
I
3
tan
31
1
I
3
(2r + 1)
+1
33. Let & = r
4
+ 4r
2
+ 3. so that d& = (4r
3
+ 8r) dr = 4(r
3
+ 2r) dr, r = 0 i & = 3, and r = 1 i & = 8.
Then
1
0
r
3
+ 2r
r
4
+ 4r
2
+ 3
dr =

8
3
1
&
1
4
d&
=
1
4
ln|&|
8
3
=
1
4
(ln8 3ln3) =
1
4
ln
8
3
.
35.
1
r(r
2
+ 4)
2
=

r
+
1r +C
r
2
+ 4
+
1r +1
(r
2
+ 4)
2
i 1 = (r
2
+ 4)
2
+ (1r +C)r(r
2
+ 4) + (1r +1)r. Setting r = 0
gives 1 = 16, so =
1
16
. Now compare coefcients.
1 =
1
16
(r
4
+ 8r
2
+ 16) + (1r
2
+Cr)(r
2
+ 4) +1r
2
+1r
1 =
1
16
r
4
+
1
2
r
2
+ 1 +1r
4
+Cr
3
+ 41r
2
+ 4Cr +1r
2
+1r
1 =
1
16
+1
r
4
+Cr
3
+
1
2
+ 41 +1
r
2
+ (4C +1)r + 1
So 1 +
1
16
= 0 i 1 = 3
1
16
, C = 0,
1
2
+ 41 +1 = 0 i 1 = 3
1
4
, and 4C +1 = 0 i 1 = 0. Thus,
dr
r(r
2
+ 4)
2
=

1
16
r
+
3
1
16
r
r
2
+ 4
+
3
1
4
r
(r
2
+ 4)
2
dr =
1
16
ln|r| 3
1
16

1
2
ln
r
2
+ 4
3
1
4
3
1
2
1
r
2
+ 4
+C
=
1
16
ln|r| 3
1
32
ln(r
2
+ 4) +
1
8(r
2
+ 4)
+C
37.
r
2
33r + 7
(r
2
34r + 6)
2
=
r +1
r
2
34r + 6
+
Cr +1
(r
2
34r + 6)
2
i r
2
33r + 7 = (r +1)(r
2
34r + 6) +Cr +1 i
r
2
33r + 7 = r
3
+ (34+1)r
2
+ (6341 +C)r + (61 +1). So = 0, 34+1 = 1 i 1 = 1,
6341 +C = 33 i C = 1, 61 +1 = 7 i 1 = 1. Thus,
1 =

r
2
33r + 7
(r
2
34r + 6)
2
dr =

1
r
2
34r + 6
+
r + 1
(r
2
34r + 6)
2
dr
=

1
(r 32)
2
+ 2
dr +
r 32
(r
2
34r + 6)
2
dr +
3
(r
2
34r + 6)
2
dr
= 1
1
+1
2
+1
3
.
[continued]
314

1
1
=

1
(r 32)
2
+
I
2
2
dr =
1
I
2
tan
31
r 32
I
2
+C
1
1
2
=
1
2
2r 34
(r
2
34r + 6)
2
dr =
1
2
1
&
2
d& =
1
2
3
1
&
+C
2
= 3
1
2(r
2
34r + 6)
+C
2
1
3
= 3
(r 32)
2
+
I
2
2
dr = 3
1
[2(tan
2
0 + 1)]
2
I
2 sec
2
0 d0

r 2 =
s
2 tan 0,
Jr =
s
2 sec
2
0 J0
=
3
I
2
4
sec
2
0
sec
4
0
d0 =
3
I
2
4
cos
2
0 d0 =
3
I
2
4
1
2
(1 + cos 20) d0
=
3
I
2
8
0 +
1
2
sin20
+C
3
=
3
I
2
8
tan
31
r 32
I
2
+
3
I
2
8
1
2
2 sin0 cos 0
+C
3
=
3
I
2
8
tan
31
r 32
I
2
+
3
I
2
8

r 32
I
r
2
34r + 6

I
2
I
r
2
34r + 6
+C
3
=
3
I
2
8
tan
31
r 32
I
2
+
3(r 32)
4(r
2
34r + 6)
+C
3
So 1 = 1
1
+1
2
+1
3
[C = C
1
+C
2
+C
3
]
=
1
I
2
tan
31
r 32
I
2
+
31
2(r
2
34r + 6)
+
3
I
2
8
tan
31
r 32
I
2
+
3(r 32)
4(r
2
34r + 6)
+C
=

4
I
2
8
+
3
I
2
8
tan
31
r 32
I
2
+
3(r 32) 32
4(r
2
34r + 6)
+C =
7
I
2
8
tan
31
r 32
I
2
+
3r 38
4(r
2
34r + 6)
+C
39. Let & =
I
r + 1. Then r = &
2
3 1, dr = 2&d& i
dr
r
I
r + 1
=

2&d&
(&
2
31) &
= 2
d&
&
2
31
= ln
& 31
& + 1
+C = ln
I
r + 1 31
I
r + 1 + 1
+C.
41. Let & =
I
r, so &
2
= r and dr = 2&d&. Thus,
16
9
I
r
r 34
dr =

4
3
&
&
2
34
2&d& = 2
4
3
&
2
&
2
34
d& = 2
4
3
1 +
4
&
2
34
d& [by long division]

= 2 + 8
4
3
d&
(& + 2)(& 32)
(-)
Multiply
1
(& + 2)(& 32)
=

& + 2
+
1
& 32
by (& + 2)(& 32) to get 1 = (& 32) +1(& + 2). Equating coefcients we
get +1 = 0 and 32+ 21 = 1. Solving gives us 1 =
1
4
and = 3
1
4
, so
1
(& + 2)(& 32)
=
314
& + 2
+
14
& 32
and (-) is
2 + 8
4
3
314
& + 2
+
14
& 32
d& = 2 + 8
3
1
4
ln|& + 2| +
1
4
ln|& 32|
4
3
= 2 +
2 ln|& 32| 32 ln|& + 2|
4
3
= 2 + 2
ln
& 32
& + 2
4
3
= 2 + 2
ln
2
6
3ln
1
5
= 2 + 2 ln
2'6
1'5
= 2 + 2 ln
5
3
or 2 + ln
5
3
2
= 2 + ln
25
9

315
43. Let & =
3
I
r
2
+ 1. Then r
2
= &
3
31, 2rdr = 3&
2
d& i
r
3
dr
3
I
r
2
+ 1
=

(&
3
31)
3
2
&
2
d&
&
=
3
2
(&
4
3&) d&
=
3
10
&
5
3
3
4
&
2
+C =
3
10
(r
2
+ 1)
5'3
3
3
4
(r
2
+ 1)
2'3
+C
45. If we were to substitute & =
I
r, then the square root would disappear but a cube root would remain. On the other hand, the
substitution & =
3
I
r would eliminate the cube root but leave a square root. We can eliminate both roots by means of the
substitution & =
6
I
r. (Note that 6 is the least common multiple of 2 and 3.)
Let & =
6
I
r. Then r = &
6
, so dr = 6&
5
d& and
I
r = &
3
,
3
I
r = &
2
. Thus,
dr
I
r 3
3
I
r
=

6&
5
d&
&
3
3&
2
= 6
&
5
&
2
(& 31)
d& = 6
&
3
& 31
d&
= 6

&
2
+& + 1 +
1
& 31
d& [by long division]

= 6
1
3
&
3
+
1
2
&
2
+& + ln|& 31|
+C = 2
I
r + 3
3
I
r + 6
6
I
r + 6 ln
6
I
r 31
+C
47. Let & = c
i
. Then r = ln&, dr =
d&
&
i
c
2i
dr
c
2i
+ 3c
i
+ 2
=

&
2
(d&&)
&
2
+ 3& + 2
=

&d&
(& + 1)(& + 2)
=

31
& + 1
+
2
& + 2
d&
= 2 ln|& + 2| 3ln|& + 1| +C = ln
(c
i
+ 2)
2
c
i
+ 1
+C
49. Let & = tant, so that d& = sec
2
t dt. Then
sec
2
t
tan
2
t + 3 tant + 2
dt =

1
&
2
+ 3& + 2
d& =

1
(& + 1)(& + 2)
d&.
Now
1
(& + 1)(& + 2)
=

& + 1
+
1
& + 2
i 1 = (& + 2) + 1(& + 1).
Setting & = 32 gives 1 = 31, so 1 = 31. Setting & = 31 gives 1 = .
Thus,
1
(& + 1)(& + 2)
d& =

1
& + 1
3
1
& + 2
d& = ln|& + 1| 3ln|& + 2| +C = ln|tant + 1| 3ln|tant + 2| +C.

51. Let & = ln(r
2
3r + 2), d = dr. Then d& =
2r 31
r
2
3r + 2
dr, = r, and (by integration by parts)
ln(r
2
3r + 2) dr = rln(r
2
3r + 2) 3
2r
2
3r
r
2
3r + 2
dr = rln(r
2
3r + 2) 3

2 +
r 34
r
2
3r + 2
dr
= rln(r
2
3r + 2) 32r 3
1
2
(2r 31)
r
2
3r + 2
dr +
7
2
dr
(r 3
1
2
)
2
+
7
4
= rln(r
2
3r + 2) 32r 3
1
2
ln(r
2
3r + 2) +
7
2
I
7
2
d&
7
4
(&
2
+ 1)
where r
1
2
=
s
7
2
u,
Jr =
s
7
2
Ju,
(r
1
2
)
2
+
7
4
=
7
4
(u
2
+ 1)
= (r 3
1
2
) ln(r
2
3r + 2) 32r +
I
7 tan
31
& +C
= (r 3
1
2
) ln(r
2
3r + 2) 32r +
I
7 tan
31
2r 31
I
7
+C
316

53. From the graph, we see that the integral will be negative, and we guess
that the area is about the same as that of a rectangle with width 2 and
height 0.3, so we estimate the integral to be 3(2 0.3) = 30.6. Now
1
r
2
32r 33
=
1
(r 33)(r + 1)
=

r 33
+
1
r + 1
C
1 = (+1)r +331, so = 31 and 331 = 1 C =
1
4
and 1 = 3
1
4
, so the integral becomes
2
0
dr
r
2
32r 33
=
1
4
2
0
dr
r 33
3
1
4
2
0
dr
r + 1
=
1
4
ln|r 33| 3ln|r + 1|
2
0
=
1
4
ln
r 33
r + 1
2
0
=
1
4
ln
1
3
3ln3
= 3
1
2
ln3 E 30.55
55.

dr
r
2
32r
=

dr
(r 31)
2
31
=

d&
&
2
31
[put & = r 31]
=
1
2
ln
& 31
& + 1
+C [by Equation 6] =
1
2
ln
r 32
r
+C
57. (a) If t = tan
r
2
, then
r
2
= tan
31
t. The gure gives
cos
r
2
=
1
I
1 +t
2
and sin
r
2
=
t
I
1 +t
2
.
(b) cos r = cos
2
r
2
= 2 cos
2
r
2
31
= 2
1
I
1 +t
2
2
31 =
2
1 +t
2
31 =
1 3t
2
1 +t
2
(c)
r
2
= arctant i r = 2 arctant i dr =
2
1 +t
2
dt
59. Let t = tan(r2). Then, using the expressions in Exercise 57, we have
1
3 sinr 34 cos r
dr =

1
3
2t
1 +t
2
34
1 3t
2
1 +t
2
2 dt
1 +t
2
= 2
dt
3(2t) 34(1 3t
2
)
=

dt
2t
2
+ 3t 32
=

dt
(2t 31)(t + 2)
=

2
5
1
2t 31
3
1
5
1
t + 2
dt [using partial fractions]

=
1
5
ln|2t 31| 3ln|t + 2|
+C =
1
5
ln
2t 31
t + 2
+C =
1
5
ln
2 tan(r2) 31
tan(r2) + 2
+C
61. Let t = tan(r2). Then, by Exercise 57,
r'2
0
sin2r
2 + cos r
dr =

r'2
0
2 sinr cos r
2 + cos r
dr =

1
0
2
2t
1 +t
2

1 3t
2
1 +t
2
2 +
1 3t
2
1 +t
2
2
1 +t
2
dt =

1
0
8t(1 3t
2
)
(1 +t
2
)
2
2(1 +t
2
) + (1 3t
2
)
dt
=

1
0
8t
1 3t
2
(t
2
+ 3)(t
2
+ 1)
2
dt = 1

317
If we now let & = t
2
, then
1 3t
2
(t
2
+ 3)(t
2
+ 1)
2
=
1 3&
(& + 3)(& + 1)
2
=

& + 3
+
1
& + 1
+
C
(& + 1)
2
i
1 3& = (& + 1)
2
+1(& + 3)(& + 1) +C(& + 3). Set & = 31 to get 2 = 2C, so C = 1. Set & = 33 to get 4 = 4, so
= 1. Set & = 0 to get 1 = 1 + 31 + 3, so 1 = 31. So
1 =

1
0
8t
t
2
+ 3
3
8t
t
2
+ 1
+
8t
(t
2
+ 1)
2
dt =

4 ln(t
2
+ 3) 34 ln(t
2
+ 1) 3
4
t
2
+ 1
1
0
= (4 ln4 34 ln2 32) 3(4 ln3 30 34) = 8 ln2 34 ln2 34 ln3 + 2 = 4 ln
2
3
+ 2
63. By long division,
r
2
+ 1
3r 3r
2
= 31 +
3r + 1
3r 3r
2
. Now
3r + 1
3r 3r
2
=
3r + 1
r(3 3r)
=

r
+
1
3 3r
i 3r + 1 = (3 3r) +1r. Set r = 3 to get 10 = 31, so 1 =
10
3
. Set r = 0 to
get 1 = 3, so =
1
3
. Thus, the area is
2
1
r
2
+ 1
3r 3r
2
dr =

2
1
31 +
1
3
r
+
10
3
3 3r
dr =

3r +
1
3
ln|r| 3
10
3
ln|3 3r|
2
1
=

32 +
1
3
ln2 30
31 + 0 3
10
3
ln2
= 31 +
11
3
ln2
65.
1 +o
1 [(v 31)1 3o]
=

1
+
1
(v 31)1 3o
i 1 +o = [(v 31)1 3o] +11 = [(v 31)+1] 1 3o i
(v 31)+1 = 1, 3 = 1 i = 31, 1 = v. Now
t =

1 +o
1 [(v 31)1 3o]
d1 =

31
1
+
v
(v 31)1 3o
d1 = 3
d1
1
+
v
v 31
v 31
(v 31)1 3o
d1
so t = 3ln1 +
v
v 31
ln|(v 31)1 3o| +C. Here v = 0.10 and o = 900, so
t = 3ln1 +
0.1
30.9
ln|30.91 3900| +C = 3ln1 3
1
9
ln(|31| |0.91 + 900|) = 3ln1 3
1
9
ln(0.91 + 900) +C.
When t = 0, 1 = 10,000, so 0 = 3ln10,000 3
1
9
ln(9900) +C. Thus, C = ln10,000 +
1
9
ln9900 [E 10.2326], so our
equation becomes
t = ln10,000 3ln1 +
1
9
ln9900 3
1
9
ln(0.91 + 900) = ln
10,000
1
+
1
9
ln
9900
0.91 + 900
= ln
10,000
1
+
1
9
ln
1100
0.11 + 100
= ln
10,000
1
+
1
9
ln
11,000
1 + 1000
67. (a) In Maple, we dene 1(r), and then use convert(f,parfrac,x); to obtain
1(r) =
24,1104879
5r + 2
3
668323
2r + 1
3
943880,155
3r 37
+
(22,098r + 48,935)260,015
r
2
+r + 5
In Mathematica, we use the command Apart, and in Derive, we use Expand.
318

(b)

1(r) dr =
24,110
4879

1
5
ln|5r + 2| 3
668
323

1
2
ln|2r + 1| 3
9438
80,155

1
3
ln|3r 37|
+
1
260,015
22,098
r +
1
2
+ 37,886
r +
1
2
2
+
19
4
dr +C
=
24,110
4879

1
5
ln|5r + 2| 3
668
323

1
2
ln|2r + 1| 3
9438
80,155

1
3
ln|3r 37|
+
1
260,015
22,098
1
2
ln
r
2
+r + 5
+ 37,886
4
19
tan
31
1
I
19'4
r +
1
2
+C
=
4822
4879
ln|5r + 2| 3
334
323
ln|2r + 1| 3
3146
80,155
ln|3r 37| +
11,049
260,015
ln
r
2
+r + 5
+
75,772
260,015
I
19
tan
31
1
I
19
(2r + 1)
+C
Using a CAS, we get
4822 ln(5r + 2)
4879
3
334 ln(2r + 1)
323
3
3146 ln(3r 37)
80,155
+
11,049 ln(r
2
+r + 5)
260,015
+
3988
I
19
260,015
tan
31
I
19
19
(2r + 1)
The main difference in this answer is that the absolute value signs and the constant of integration have been omitted. Also,
the fractions have been reduced and the denominators rationalized.
69. There are only nitely many values of r where Q(r) = 0 (assuming that Qis not the zero polynomial). At all other values of
r, 1(r)Q(r) = G(r)Q(r), so 1(r) = G(r). In other words, the values of 1 and G agree at all except perhaps nitely
many values of r. By continuity of 1 and G, the polynomials 1 and G must agree at those values of r too.
More explicitly: if o is a value of r such that Q(o) = 0, then Q(r) 6= 0 for all r sufciently close to o. Thus,
1(o) = lim
i<a
1(r) [by continuity of 1]
= lim
i<a
G(r) [whenever Q(r) 6= 0]
= G(o) [by continuity of G]
7.5 Strategy for Integration
1. Let & = sinr, so that d& = cos rdr. Then
cos r(1 +sin

2
r) dr =

(1 +&
2
) d& = &+
1
3
&
3
+C = sinr +
1
3
sin
3
r +C.
3.

sinr + sec r
tanr
dr =

sinr
tanr
+
sec r
tanr
dr =

(cos r + csc r) dr = sinr + ln|csc r 3cot r| +C
5.

2
0
2t
(t 33)
2
dt =

31
33
2(& + 3)
&
2
d&

u = | 3,
Ju = J|
=

31
33
2
&
+
6
&
2
d& =

2 ln|&| 3
6
&
31
33
= (2 ln1 + 6) 3(2 ln3 + 2) = 4 32 ln3 or 4 3ln9
7. Let & = arctanj. Then d& =
dj
1 +j
2
i

1
31
c
arctan u
1 +j
2
dj =

r'4
3r'4
c
u
d& =

c
u
r'4
3r'4
= c
r'4
3c
3r'4
.
SECTION 7.5 STRATEGY FOR INTEGRATION

319
9.

3
1
v
4
lnv dv
u = ln :,
Ju =
J:
:
Ju = :
4
J:,
u =
1
5
:
5
=

1
5
v
5
lnv
3
1
3
3
1
1
5
v
4
dv =
243
5
ln3 30 3
1
25
v
5
3
1
=
243
5
ln3 3
243
25
3
1
25
=
243
5
ln3 3
242
25
11.

r 31
r
2
34r + 5
dr =

(r 32) + 1
(r 32)
2
+ 1
dr =

&
&
2
+ 1
+
1
&
2
+ 1
d& [& = r 32, d& = dr]

=
1
2
ln(&
2
+ 1) + tan
31
& +C =
1
2
ln(r
2
34r + 5) + tan
31
(r 32) +C
13.

sin
3
0 cos
5
0 d0 =

cos
5
0 sin
2
0 sin0 d0 = 3
cos
5
0 (1 3cos
2
0)(3sin0) d0
= 3
&
5
(1 3&
2
) d&

u = cos 0,
Ju = sin 0 J0
=

(&
7
3&
5
) d& =
1
8
&
8
3
1
6
&
6
+C =
1
8
cos
8
0 3
1
6
cos
6
0 +C
Another solution:
sin
3
0 cos
5
0 d0 =

sin
3
0 (cos
2
0)
2
cos 0 d0 =

sin
3
0 (1 3sin
2
0)
2
cos 0 d0
=

&
3
(1 3&
2
)
2
d&

u = sin 0,
Ju = cos 0 J0
=

&
3
(1 32&
2
+&
4
) d&
=

(&
3
32&
5
+&
7
) d& =
1
4
&
4
3
1
3
&
6
+
1
8
&
8
+C =
1
4
sin
4
0 3
1
3
sin
6
0 +
1
8
sin
8
0 +C
15. Let r = sin0, where 3
r
2
$ 0 $
r
2
. Then dr = cos 0 d0 and (1 3r
2
)
1'2
= cos 0,
so
dr
(1 3r
2
)
3'2
=

cos 0 d0
(cos 0)
3
=

sec
2
0 d0 = tan0 +C =
r
I
1 3r
2
+C.
17.

rsin
2
rdr

u = r,
Ju = Jr
Ju = sin
2
rJr,
u =
U
sin
2
rJr =
U
1
2
(1 cos 2r) Jr =
1
2
r
1
2
sin rcos r
=
1
2
r
2
3
1
2
rsinrcos r 3

1
2
r 3
1
2
sinrcos r
dr
=
1
2
r
2
3
1
2
rsinrcos r 3
1
4
r
2
+
1
4
sin
2
r +C =
1
4
r
2
3
1
2
rsinrcos r +
1
4
sin
2
r +C
Note:

sinrcos rdr =

c dc =
1
2
c
2
+C [where c = sinr, dc = cos rdr].
A slightly different method is to write
rsin
2
rdr =

r
1
2
(1 3cos 2r) dr =
1
2
rdr 3
1
2
rcos 2rdr. If we evaluate

the second integral by parts, we arrive at the equivalent answer
1
4
r
2
3
1
4
rsin2r 3
1
8
cos 2r +C.
19. Let & = c
i
. Then
c
i+c
{
dr =

c
c
{
c
i
dr =

c
u
d& = c
u
+C = c
c
{
+C.
21. Let t =
I
r, so that t
2
= r and 2t dt = dr. Then
arctan
I
rdr =

arctant (2t dt) = 1. Now use parts with
& = arctant, d = 2t dt i d& =
1
1 +t
2
dt, = t
2
. Thus,
1 = t
2
arctant 3
t
2
1 +t
2
dt = t
2
arctant 3

1 3
1
1 +t
2
dt = t
2
arctant 3t + arctant +C
= rarctan
I
r 3
I
r + arctan
I
r +C

or (r + 1) arctan
I
r 3
I
r +C
320

23. Let & = 1 +
I
r. Then r = (& 3 1)
2
, dr = 2(& 3 1) d& i
1
0
1 +
I
r
8
dr =

2
1
&
8
2(& 31) d& = 2
2
1
(&
9
3&
8
) d& =

1
5
&
10
32
1
9
&
9
2
1
=
1024
5
3
1024
9
3
1
5
+
2
9
=
4097
45
.
25.
3r
2
32
r
2
32r 38
= 3 +
6r + 22
(r 34)(r + 2)
= 3 +

r 34
+
1
r + 2
i 6r + 22 = (r + 2) +1(r 34). Setting
r = 4 gives 46 = 6, so =
23
3
. Setting r = 32 gives 10 = 361, so 1 = 3
5
3
. Now
3r
2
32
r
2
32r 38
dr =

3 +
233
r 34
3
53
r + 2
dr = 3r +
23
3
ln|r 34| 3
5
3
ln|r + 2| +C.
27. Let & = 1 +c
i
, so that d& = c
i
dr = (& 31) dr. Then
1
1 +c
i
dr =

1
&

d&
& 31
=

1
&(& 31)
d& = 1. Now
1
&(& 31)
=

&
+
1
& 31
i 1 = (& 31) +1&. Set & = 1 to get 1 = 1. Set & = 0 to get 1 = 3, so = 31.
Thus, 1 =

31
&
+
1
& 31
d& = 3ln|&| + ln|& 31| +C = 3ln(1 +c

i
) + lnc
i
+C = r 3ln(1 +c
i
) +C.
Another method: Multiply numerator and denominator by c
3i
and let & = c
3i
+ 1. This gives the answer in the
form 3ln(c
3i
+ 1) +C.
29.

5
0
3n 31
n + 2
dn =

5
0
3 3
7
n + 2
dn =

3n 37 ln|n + 2|
5
0
= 15 37 ln7 + 7 ln2
= 15 + 7(ln2 3ln7) = 15 + 7 ln
2
7
31. As in Example 5,

1 +r
1 3r
dr =
I
1 +r
I
1 3r

I
1 +r
I
1 +r
dr =

1 +r
I
1 3r
2
dr =

dr
I
1 3r
2
+
rdr
I
1 3r
2
= sin
31
r 3
1 3r
2
+C.
Another method: Substitute & =

(1 +r)(1 3r).
33. 3 32r 3r
2
= 3(r
2
+ 2r + 1) + 4 = 4 3(r + 1)
2
. Let r + 1 = 2 sin0,
where 3
r
2
$ 0 $
r
2
. Then dr = 2 cos 0 d0 and
I
3 32r 3r
2
dr =

4 3(r + 1)
2
dr =

4 34 sin
2
0 2 cos 0 d0
= 4
cos
2
0 d0 = 2
(1 + cos 20) d0
= 20 + sin20 +C = 20 + 2 sin0 cos 0 +C
= 2 sin
31
r + 1
2
+ 2
r + 1
2

I
3 32r 3r
2
2
+C
= 2 sin
31
r + 1
2
+
r + 1
2
I
3 32r 3r
2
+C
35. Because 1(r) = r
8
sinr is the product of an even function and an odd function, it is odd.
Therefore,
1
31
r
8
sinrdr = 0 [by (5.5.7)(b)].
37.

r'4
0
cos
2
0 tan
2
0 d0 =

r'4
0
sin
2
0 d0 =

r'4
0
1
2
(1 3cos 20) d0 =

1
2
0 3
1
4
sin20
r'4
0
=

r
8
3
1
4
3(0 30) =
r
8
3
1
4

321
39. Let & = sec 0, so that d& = sec 0 tan0 d0. Then
sec 0 tan0
sec
2
0 3sec 0
d0 =

1
&
2
3&
d& =

1
&(& 31)
d& = 1. Now
1
&(& 31)
=

&
+
1
& 31
i 1 = (& 31) +1&. Set & = 1 to get 1 = 1. Set & = 0 to get 1 = 3, so = 31.
Thus, 1 =

31
&
+
1
& 31
d& = 3ln|&| + ln|& 31| +C = ln|sec 0 31| 3ln|sec 0| +C [or ln|1 3cos 0| +C].
41. Let & = 0, d = tan
2
0 d0 =

sec
2
0 31
d0 i d& = d0 and = tan0 30. So
0 tan
2
0 d0 = 0(tan0 30) 3
(tan0 30) d0 = 0 tan0 30

2
3ln|sec 0| +
1
2
0
2
+C
= 0 tan0 3
1
2
0
2
3ln|sec 0| +C
43. Let & = 1 +c
i
, so that d& = c
i
dr. Then
c
i
I
1 +c
i
dr =

&
1'2
d& =
2
3
&
3'2
+C =
2
3
(1 +c
i
)
3'2
+C.
Or: Let & =
I
1 +c
i
, so that &
2
= 1 + c
i
and 2&d& = c
i
dr. Then
c
i
I
1 +c
i
dr =

& 2&d& =

2&
2
d& =
2
3
&
3
+C =
2
3
(1 +c
i
)
3'2
+C.
45. Let t = r
3
. Then dt = 3r
2
dr i 1 =

r
5
c
3i
3
dr =
1
3
tc
3I
dt. Now integrate by parts with & = t, d = c
3I
dt:
1 = 3
1
3
tc
3I
+
1
3
c
3I
dt = 3
1
3
tc
3I
3
1
3
c
3I
+C = 3
1
3
c
3i
3
(r
3
+ 1) +C.
47. Let & = r 31, so that d& = dr. Then
r
3
(r 31)
34
dr =

(& + 1)
3
&
34
d& =

(&
3
+ 3&
2
+ 3& + 1)&
34
d& =

(&
31
+ 3&
32
+ 3&
33
+&
34
) d&
= ln|&| 33&
31
3
3
2
&
32
3
1
3
&
33
+C = ln|r 31| 33(r 31)
31
3
3
2
(r 31)
32
3
1
3
(r 31)
33
+C
49. Let & =
I
4r + 1 i &
2
= 4r + 1 i 2&d& = 4 dr i dr =
1
2
&d&. So
1
r
I
4r + 1
dr =

1
2
&d&
1
4
(&
2
31) &
= 2
d&
&
2
31
= 2
1
2
ln
& 31
& + 1
+C [by Formula 19]

= ln
I
4r + 1 31
I
4r + 1 + 1
+C
51. Let 2r = tan0 i r =
1
2
tan0, dr =
1
2
sec
2
0 d0,
I
4r
2
+ 1 = sec 0, so
dr
r
I
4r
2
+ 1
=

1
2
sec
2
0 d0
1
2
tan0 sec 0
=

sec 0
tan0
d0 =

csc 0 d0
= 3ln|csc 0 + cot 0| +C [or ln|csc 0 3cot 0| +C]
= 3ln
I
4r
2
+ 1
2r
+
1
2r
+C

or ln
I
4r
2
+ 1
2r
3
1
2r
+C
53.

r
2
sinh(ir)dr =
1
i
r
2
cosh(ir) 3
2
i
rcosh(ir) dr

u = r
2
,
Ju = 2rJr
Ju = sinh(nr) Jr,
u =
1
p
cosh(nr)
=
1
i
r
2
cosh(ir) 3
2
i
1
r
rsinh(ir) 3
1
i
sinh(ir) dr

I = r,
JI = Jr
J\ = cosh(nr) Jr,
\ =
1
p
sinh(nr)
=
1
i
r
2
cosh(ir) 3
2
i
2
rsinh(ir) +
2
i
3
cosh(ir) +C
322

55. Let & =
I
r, so that r = &
2
and dr = 2&d&. Then
dr
r +r
I
r
=

2&d&
&
2
+&
2
&
=

2
&(1 +&)
d& = 1.
Now
2
&(1 +&)
=

&
+
1
1 +&
i 2 = (1 +&) +1&. Set & = 31 to get 2 = 31, so 1 = 32. Set & = 0 to get 2 = .
Thus, 1 =

2
&
3
2
1 +&
d& = 2 ln|&| 32 ln|1 +&| +C = 2 ln

I
r 32 ln
1 +
I
r
+C.
57. Let & =
3
I
r +c. Then r = &
3
3c i
r
3
I
r +c dr =

(&
3
3c)& 3&
2
d& = 3
(&
6
3c&
3
) d& =
3
7
&
7
3
3
4
c&
4
+C =
3
7
(r +c)
7'3
3
3
4
c(r +c)
4'3
+C
cos r cos
3
(sinr) dr =

cos
3
&d& =

cos
2
& cos &d& =

(1 3sin
2
&) cos &d&
=

(cos & 3sin
2
& cos &) d& = sin& 3
1
3
sin
3
& +C = sin(sinr) 3
1
3
sin
3
(sinr) +C
61. Let j =
I
r so that dj =
1
2
I
r
dr i dr = 2
I
rdj = 2j dj. Then
I
rc
I
i
dr =

jc
u
(2j dj) =
2j
2
c
u
dj

u = 2u
2
,
Ju = 4u Ju
Ju = t
|
Ju,
u = t
|
= 2j
2
c
u
3
4jc
u
dj

I = 4u,
JI = 4 Ju
J\ = t
|
Ju,
\ = t
|
= 2j
2
c
u
3
4jc
u
3
4c
u
dj
= 2j
2
c
u
34jc
u
+ 4c
u
+C
= 2(j
2
32j + 2)c
u
+C = 2
r 32
I
r + 2
c
I
i
+C
63. Let & = cos
2
r, so that d& = 2 cos r(3sinr) dr. Then
sin2r
1 + cos
4
r
dr =

2 sinr cos r
1 + (cos
2
r)
2
dr =

1
1 +&
2
(3d&) = 3tan
31
& +C = 3tan
31
(cos
2
r) +C.
65.

dr
I
r + 1 +
I
r
=

1
I
r + 1 +
I
r

I
r + 1 3
r
I
r
I
r + 1 3
I
r
dr =

r + 1 3
I
r
dr
=
2
3
(r + 1)
3'2
3r
3'2
+C
67. Let r = tan0, so that dr = sec
2
0 d0, r =
I
3 i 0 =
r
3
, and r = 1 i 0 =
r
4
. Then
I
3
1
I
1 +r
2
r
2
dr =

r'3
r'4
sec 0
tan
2
0
sec
2
0 d0 =

r'3
r'4
sec 0 (tan
2
0 + 1)
tan
2
0
d0 =

r'3
r'4
sec 0 tan
2
0
tan
2
0
+
sec 0
tan
2
0
d0
=

r'3
r'4
(sec 0 + csc 0 cot 0) d0 =

ln|sec 0 + tan0| 3csc 0
r'3
r'4
=

ln
2 +
I
3
3
2
I
3
ln
I
2 + 1
3
I
2
=
I
2 3
2
I
3
+ ln
2 +
I
3
3ln
1 +
I
2

323
69. Let & = c
i
. Then r = ln&, dr = d&& i
c
2i
1 +c
i
dr =

&
2
1 +&
d&
&
=

&
1 +&
d& =

1 3
1
1 +&
d& = & 3ln|1 +&| +C = c

i
3ln(1 +c
i
) +C.
71. Let 0 = arcsinr, so that d0 =
1
I
1 3r
2
dr and r = sin0. Then
r + arcsinr
I
1 3r
2
dr =

(sin0 +0) d0 = 3cos 0 +
1
2
0
2
+C
= 3
I
1 3r
2
+
1
2
(arcsinr)
2
+C
73.
1
(r 32)(r
2
+ 4)
=

r 32
+
1r +C
r
2
+ 4
i 1 = (r
2
+4) +(1r+C)(r32) = (+1)r
2
+(C321)r+(432C).
So 0 = +1 = C 321, 1 = 432C. Setting r = 2 gives =
1
8
i 1 = 3
1
8
and C = 3
1
4
. So
1
(r 32)(r
2
+ 4)
dr =

1
8
r 32
+
3
1
8
r 3
1
4
r
2
+ 4
dr =
1
8
dr
r 32
3
1
16
2rdr
r
2
+ 4
3
1
4
dr
r
2
+ 4
=
1
8
ln|r 32| 3
1
16
ln(r
2
+ 4) 3
1
8
tan
31
(r2) +C
75. Let j =
I
1 +c
i
, so that j
2
= 1 +c
i
, 2j dj = c
i
dr, c
i
= j
2
31, and r = ln(j
2
31). Then
rc
i
I
1 +c
i
dr =

ln(j
2
31)
j
(2j dj) = 2
[ln(j + 1) + ln(j 31)] dj

= 2[(j + 1) ln(j + 1) 3(j + 1) + (j 31) ln(j 31) 3(j 31)] +C [by Example 7.1.2]
= 2[j ln(j + 1) + ln(j + 1) 3j 31 +j ln(j 31) 3ln(j 31) 3j + 1] +C
= 2[j(ln(j + 1) + ln(j 31)) + ln(j + 1) 3ln(j 31) 32j] +C
= 2
j ln(j
2
31) + ln
j + 1
j 31
32j
+C = 2
I
1 +c
i
ln(c
i
) + ln
I
1 +c
i
+ 1
I
1 +c
i
31
32
I
1 +c
i
+C
= 2r
I
1 +c
i
+ 2 ln
I
1 +c
i
+ 1
I
1 +c
i
31
34
I
1 +c
i
+C = 2(r 32)
I
1 +c
i
+ 2 ln
I
1 +c
i
+ 1
I
1 +c
i
31
+C
77. Let & = r
3'2
so that &
2
= r
3
and d& =
3
2
r
1'2
dr i
I
rdr =
2
3
d&. Then
I
r
1 +r
3
dr =

2
3
1 +&
2
d& =
2
3
tan
31
& +C =
2
3
tan
31
(r
3'2
) +C.
79. Let & = r, d = sin
2
rcos rdr i d& = dr, =
1
3
sin
3
r. Then
rsin
2
r cos rdr =
1
3
rsin
3
r 3
1
3
sin
3
rdr =
1
3
rsin
3
r 3
1
3
(1 3cos
2
r) sinrdr
=
1
3
rsin
3
r +
1
3
(1 3j
2
) dj

u = cos r,
Ju = sin rJr
=
1
3
rsin
3
r +
1
3
j 3
1
9
j
3
+C =
1
3
rsin
3
r +
1
3
cos r 3
1
9
cos
3
r +C
324

81. The function j = 2rc
i
2
does have an elementary antiderivative, so well use this fact to help evaluate the integral.
(2r
2
+ 1)c
i
2
dr =

2r
2
c
i
2
dr +
c
i
2
dr =

r
2rc
i
2
dr +
c
i
2
dr
= rc
i
2
3
c
i
2
dr +
c
i
2
dr

u = r,
Ju = Jr
Ju= 2rt
{
2
Jr,
u= t
{
2
= rc
i
2
+C
7.6 Integration Using Tables and Computer Algebra Systems
Keep in mind that there are several ways to approach many of these exercises, and different methods can lead to different forms of the answer.
1. We could make the substitution & =
I
2 r to obtain the radical
I
7 3&
2
and then use Formula 33 with o =
I
7.
Alternatively, we will factor
I
2 out of the radical and use o =

7
2
.
I
7 32r
2
r
2
dr =
I
2

7
2
3r
2
r
2
dr
33
=
I
2
3
1
r
7
2
3r
2
3sin
31
r
7
2
+C
= 3
1
r
I
7 32r
2
3
I
2 sin
31
2
7
r
+C
3. Let & = r i d& = dr, so
sec
3
(r) dr =
1
r
sec
3
&d&
71
=
1
r
1
2
sec &tan& +
1
2
ln|sec & + tan&|
+C
=
1
2r
sec rtanr +
1
2r
ln|sec r + tanr| +C
5.

1
0
2rcos
31
rdr
91
= 2
2r
2
31
4
cos
31
r 3
r
I
1 3r
2
4
1
0
= 2
1
4
0 30
3
1
4

r
2
30
= 2
r
8
=
r
4
7. Let & = r, so that d& = dr. Then
tan
3
(r) dr =

tan
3
&
1
r
d&
=
1
r
tan
3
&d&
69
=
1
r
1
2
tan
2
& + ln|cos &|
+C
=
1
2r
tan
2
(r) +
1
r
ln|cos (r)| +C
9. Let & = 2r and o = 3. Then d& = 2 dr and
dr
r
2
I
4r
2
+ 9
=

1
2
d&
&
2
4
I
&
2
+o
2
= 2
d&
&
2
I
o
2
+&
2
28
= 32
I
o
2
+&
2
o
2
&
+C
= 32
I
4r
2
+ 9
9 2r
+C = 3
I
4r
2
+ 9
9r
+C
11.

0
31
t
2
c
3I
dt
97
=

1
31
t
2
c
3I
0
31
3
2
31
0
31
tc
3I
dt = c + 2
0
31
tc
3I
dt
96
= c + 2
1
(31)
2
(3t 31) c
3I
0
31
= c + 2
3c
0
+ 0
= c 32
13.

tan
3
(1:)
:
2
d:

u = 1/:,
Ju = J:/:
2
= 3
tan
3
&d&
69
= 3
1
2
tan
2
& 3ln|cos &| +C
= 3
1
2
tan
2
1
:
3ln
cos
1
:
+C
SECTION 7.6 INTEGRATION USING TABLES AND COMPUTER ALGEBRA SYSTEMS

325
15. Let & = c
i
, so that d& = c
i
dr and c
2i
= &
2
. Then
c
2i
arctan(c
i
) dr =

&
2
arctan&
d&
&
=

&arctan&d&
92
=
&
2
+ 1
2
arctan& 3
&
2
+C =
1
2
(c
2i
+ 1) arctan(c
i
) 3
1
2
c
i
+C
17. Let : = 6 +4j 34j
2
= 6 3(4j
2
34j +1) +1 = 7 3(2j 31)
2
, & = 2j 31, and o =
I
7. Then : = o
2
3&
2
, d& = 2 dj,
and
6 + 4j 34j
2
dj =

j
I
: dj =

1
2
(& + 1)
I
o
2
3&
2
1
2
d& =
1
4
&
I
o
2
3&
2
d& +
1
4
I
o
2
3&
2
d&
=
1
4
I
o
2
3&
2
d& 3
1
8
(32&)
I
o
2
3&
2
d&
30
=
&
8
I
o
2
3&
2
+
o
2
8
sin
31
&
o
3
1
8
I
ndn

u = o
2
u
2
,
Ju = 2uJu
=
2j 31
8
6 + 4j 34j
2
+
7
8
sin
31
2j 31
I
7
3
1
8

2
3
n
3'2
+C
=
2j 31
8
6 + 4j 34j
2
+
7
8
sin
31
2j 31
I
7
3
1
12
(6 + 4j 34j
2
)
3'2
+C
This can be rewritten as
6 + 4j 34j
2
1
8
(2j 31) 3
1
12
(6 + 4j 34j
2
)
+
7
8
sin
31
2j 31
I
7
+C
=

1
3
j
2
3
1
12
j 3
5
8
6 + 4j 34j
2
+
7
8
sin
31
2j 31
I
7
+C
=
1
24
(8j
2
32j 315)
6 + 4j 34j
2
+
7
8
sin
31
2j 31
I
7
+C
19. Let & = sinr. Then d& = cos rdr, so
sin
2
r cos r ln(sinr) dr =

&
2
ln&d&
101
=
&
2+1
(2 + 1)
2
[(2 + 1) ln& 31] +C =
1
9
&
3
(3 ln& 31) +C
=
1
9
sin
3
r[3 ln(sinr) 31] +C
21. Let & = c
i
and o =
I
3. Then d& = c
i
dr and
c
i
3 3c
2i
dr =

d&
o
2
3&
2
19
=
1
2o
ln
& +o
& 3o
+C =
1
2
I
3
ln
c
i
+
I
3
c
i
3
I
3
+C.
23.

sec
5
rdr
77
=
1
4
tanr sec
3
r +
3
4
sec
3
rdr
77
=
1
4
tanr sec
3
r +
3
4
1
2
tanr sec r +
1
2
sec rdr
14
=
1
4
tanr sec
3
r +
3
8
tanr sec r +
3
8
ln|sec r + tanr| +C
25. Let & = lnr and o = 2. Then d& = drr and

4 + (lnr)
2
r
dr =

o
2
+&
2
d&
21
=
&
2
o
2
+&
2
+
o
2
2
ln
& +
o
2
+&
2
+C
=
1
2
(lnr)
4 + (lnr)
2
+ 2 ln
lnr +
4 + (lnr)
2
+C
326

27. Let & = c
i
. Then r = ln&, dr = d&&, so

c
2i
31 dr =
I
&
2
31
&
d&
41
=

&
2
31 3cos
31
(1&) +C =

c
2i
31 3cos
31
(c
3i
) +C.
29.

r
4
dr
I
r
10
32
=

r
4
dr
(r
5
)
2
32
=
1
5
d&
I
&
2
32
u= r
5
,
Ju= 5r
4
Jr
43
=
1
5
ln
& +
I
&
2
32
+C =
1
5
ln
r
5
+
I
r
10
32
+C
31. Using cylindrical shells, we get
\ = 2
2
0
r r
4 3r
2
dr = 2
2
0
r
2
4 3r
2
dr
31
= 2
r
8
(2r
2
34)
4 3r
2
+
16
8
sin
31
r
2
2
0
= 2[(0 + 2 sin
31
1) 3(0 + 2 sin
31
0] = 2
2

2
= 2
2
33. (a)
d
d&
1
/
3
o +/& 3
o
2
o +/&
32o ln|o +/&|
+C
=
1
/
3
/ +
/o
2
(o +/&)
2
3
2o/
(o +/&)
=
1
/
3
/(o +/&)
2
+/o
2
3(o +/&)2o/
(o +/&)
2
=
1
/
3
/
3
&
2
(o +/&)
2
=
&
2
(o +/&)
2
(b) Let t = o +/& i dt = / d&. Note that & =
t 3o
/
and d& =
1
/
dt.
&
2
d&
(o +/&)
2
=
1
/
3
(t 3o)
2
t
2
dt =
1
/
3
t
2
32ot +o
2
t
2
dt =
1
/
3

1 3
2o
t
+
o
2
t
2
dt
=
1
/
3
t 32o ln|t| 3
o
2
t
+C =
1
/
3
o +/& 3
o
2
o +/&
32o ln|o +/&|
+C
35. Maple and Mathematica both give
sec
4
rdr =
2
3
tanr +
1
3
tanr sec
2
r, while Derive gives the second
term as
sinr
3 cos
3
r
=
1
3
sinr
cos r
1
cos
2
r
=
1
3
tanr sec
2
r. Using Formula 77, we get
sec
4
rdr =
1
3
tanr sec
2
r +
2
3
sec
2
rdr =
1
3
tanr sec
2
r +
2
3
tanr +C.
37. Derive gives
r
2
I
r
2
+ 4 dr =
1
4
r(r
2
+ 2)
I
r
2
+ 4 32 ln
I
r
2
+ 4 +r
. Maple gives
1
4
r(r
2
+ 4)
3'2
3
1
2
r
I
r
2
+ 4 32 arcsinh
1
2
r
. Applying the command convert(%,ln); yields

1
4
r(r
2
+ 4)
3'2
3
1
2
r
I
r
2
+ 4 32 ln
1
2
r +
1
2
I
r
2
+ 4
=
1
4
r(r
2
+ 4)
1'2
(r
2
+ 4) 32
32 ln
r +
I
r
2
+ 4
=
1
4
r(r
2
+ 2)
I
r
2
+ 4 32 ln
I
r
2
+ 4 +r
+ 2 ln2
Mathematica gives
1
4
r(2 +r
2
)
I
3 +r
2
32 arcsinh(r2). Applying the TrigToExp and Simplify commands gives
1
4
r(2 +r
2
)
I
4 +r
2
38 log
1
2
r +
I
4 +r
2
=
1
4
r(r
2
+ 2)
I
r
2
+ 4 32 ln
r +
I
4 +r
2
+ 2 ln2, so all are

equivalent (without constant).
SECTION 7.6 INTEGRATION USING TABLES AND COMPUTER ALGEBRA SYSTEMS

327
Now use Formula 22 to get
r
2
2
2
+r
2
dr =
r
8
(2
2
+ 2r
2
)
I
2
2
+r
2
3
2
4
8
ln
r +
I
2
2
+r
2
+C
=
r
8
(2)(2 +r
2
)
I
4 +r
2
32 ln
r +
I
4 +r
2
+C
=
1
4
r(r
2
+ 2)
I
r
2
+ 4 32 ln
I
r
2
+ 4 +r
+C
39. Maple gives
r
I
1 + 2rdr =
1
10
(1 + 2r)
5'2
3
1
6
(1 +2r)
3'2
, Mathematica gives
I
1 + 2r
2
5
r
2
+
1
15
r 3
1
15
, and Derive
gives
1
15
(1 + 2r)
3'2
(3r 31). The rst two expressions can be simplied to Derives result. If we use Formula 54, we get
r
I
1 + 2rdr =
2
15(2)
2
(3 2r 32 1)(1 + 2r)
3'2
+C =
1
30
(6r 32)(1 + 2r)
3'2
+C =
1
15
(3r 31)(1 + 2r)
3'2
.
41. Maple gives
tan
5
rdr =
1
4
tan
4
r 3
1
2
tan
2
r +
1
2
ln(1 + tan
2
r), Mathematica gives
tan
5
rdr =
1
4
[31 32 cos(2r)] sec
4
r 3ln(cos r), and Derive gives
tan
5
rdr =
1
4
tan
4
r 3
1
2
tan
2
r 3ln(cos r).
These expressions are equivalent, and none includes absolute value bars or a constant of integration. Note that Mathematicas
and Derives expressions suggest that the integral is undened where cos r < 0, which is not the case. Using Formula 75,
tan
5
rdr =
1
5 31
tan
531
r 3
tan
532
rdr =
1
4
tan
4
r 3
tan
3
rdr. Using Formula 69,
tan
3
rdr =
1
2
tan
2
r + ln|cos r| +C, so
tan
5
rdr =
1
4
tan
4
r 3
1
2
tan
2
r 3ln|cos r| +C.
43. (a) 1(r) =
1(r) dr =

1
r
I
1 3r
2
dr
35
= 3
1
1
ln
1 +
I
1 3r
2
r
+C = 3ln
1 +
I
1 3r
2
r
+C.
1 has domain
r | r 6= 0, 1 3r
2
0
= {r | r 6= 0, |r| < 1} = (31. 0) (0. 1). 1 has the same domain.

(b) Derive gives 1(r) = ln
I
1 3r
2
31
3lnr and Mathematica gives 1(r) = lnr 3ln
1 +
I
1 3r
2
.
Both are correct if you take absolute values of the logarithm arguments, and both would then have the
same domain. Maple gives 1(r) = 3arctanh
1
I
1 3r
2
. This function has domain
|r| < 1. 31 < 1

I
1 3r
2
< 1
=

r

|r| < 1, 1
I
1 3r
2
< 1
=

r

|r| < 1.
I
1 3r
2
1
= ,
the empty set! If we apply the command convert(%,ln); to Maples answer, we get
3
1
2
ln
1
I
1 3r
2
+ 1
+
1
2
ln
1 3
1
I
1 3r
2
, which has the same domain, .

45. Maple gives the antiderivative
1(r) =
r
2
31
r
4
+r
2
+ 1
dr = 3
1
2
ln(r
2
+r + 1) +
1
2
ln(r
2
3r + 1).
We can see that at 0, this antiderivative is 0. From the graphs, it appears that 1 has
a maximum at r = 31 and a minimum at r = 1 [since 1
0
(r) = 1(r) changes
sign at these r-values], and that 1 has inection points at r E 31.7, r = 0, and
r E 1.7 [since 1(r) has extrema at these r-values].
328

47. Since 1(r) = sin
4
r cos
6
r is everywhere positive, we know that its antiderivative 1 is increasing. Maple gives
1(r) dr = 3
1
10
sin
3
r cos
7
r 3
3
80
sinr cos
7
r +
1
160
cos
5
r sinr +
1
128
cos
3
r sinr +
3
256
cos r sinr +
3
256
r
and this expression is 0 at r = 0.
1 has a minimum at r = 0 and a maximum at r = .
1 has inection points where 1
0
changes sign, that is, at r E 0.7, r = 2,
and r E 2.5.
7.7 Approximate Integration
1. (a) {r = (/ 3o)n = (4 30)2 = 2
1
2
=
2
.=1
1(r
.31
) {r = 1(r
0
) 2 +1(r
1
) 2 = 2 [1(0) +1(2)] = 2(0.5 + 2.5) = 6
1
2
=
2
.=1
1(r
.
) {r = 1(r
1
) 2 +1(r
2
) 2 = 2 [1(2) +1(4)] = 2(2.5 + 3.5) = 12
A
2
=
2
.=1
1(r
.
){r = 1(r
1
) 2 +1(r
2
) 2 = 2 [1(1) +1(3)] E 2(1.6 + 3.2) = 9.6
(b) 1
2
is an underestimate, since the area under the small rectangles is less than
the area under the curve, and 1
2
is an overestimate, since the area under the
large rectangles is greater than the area under the curve. It appears that A
2
is an overestimate, though it is fairly close to 1. See the solution to
Exercise 45 for a proof of the fact that if 1 is concave down on [o. /], then
the Midpoint Rule is an overestimate of

l
a
1(r) dr.
(c) T
2
=

1
2
{r
[1(r
0
) + 21(r
1
) +1(r
2
)] =
2
2
[1(0) + 21(2) +1(4)] = 0.5 + 2(2.5) + 3.5 = 9.
This approximation is an underestimate, since the graph is concave down. Thus, T
2
= 9 < 1. See the solution to
Exercise 45 for a general proof of this conclusion.
(d) For any n, we will have 1
n
< T
n
< 1 < A
n
< 1
n
.
3. 1(r) = cos
r
2
, {r =
1 30
4
=
1
4
(a) T
4
=
1
4 2
1(0) + 21
1
4
+ 21
2
4
+ 21
3
4
+1(1)
E 0.895759
(b) A
4
=
1
4
1
8
+1
3
8
+1
5
8
+1
7
8
E 0.908907
The graph shows that 1 is concave down on [0. 1]. So T
4
is an
underestimate and A
4
is an overestimate. We can conclude that
0.895759 <

1
0
cos
r
2
dr < 0.908907.
SECTION 7.7 APPROXIMATE INTEGRATION

329
5. 1(r) = r
2
sinr, {r =
/ 3o
n
=
30
8
=

8
(a) A
8
=
r
8
r
16
+1
3r
16
+1
5r
16
+ +1
15r
16
E 5.932957
(b) o
8
=
r
8 3
1(0) + 41
r
8
+ 21
2r
8
+ 41
3r
8
+ 21
4r
8
+ 41
5r
8
+ 21
6r
8
+ 41
7r
8
+1()
E 5.869247
Actual:

r
0
r
2
sinrdr
84
=

3r
2
cos r
r
0
+ 2
r
0
rcos rdr
83
=

3
2
(31) 30
+ 2
cos r +rsinr
r
0
=
2
+ 2[(31 + 0) 3(1 + 0)] =
2
34 E 5.869604
Errors: 1
L
= actual 3A
8
=

r
0
r
2
sinrdr 3A
8
E 30.063353
1
S
= actual 3o
8
=

r
0
r
2
sinrdr 3o
8
E 0.000357
7. 1(r) =
4
I
1 +r
2
, {r =
2 30
8
=
1
4
(a) T
8
=
1
4 2
1(0) + 21
1
4
+ 21
1
2
+ + 21
3
2
+ 21
7
4
+1(2)
E 2.413790
(b) A
8
=
1
4
1
8
+1
3
8
+ +1
13
8
+1
15
8
E 2.411453
(c) o
8
=
1
4 3
1(0) + 41
1
4
+ 21
1
2
+ 41
3
4
+ 21(1) + 41
5
4
+ 21
3
2
+ 41
7
4
+1(2)
E 2.412232
9. 1(r) =
lnr
1 +r
, {r =
2 31
10
=
1
10
(a) T
10
=
1
10 2
[1(1) + 21(1.1) + 21(1.2) + + 21(1.8) + 21(1.9) +1(2)] E 0.146879
(b) A
10
=
1
10
[1(1.05) +1(1.15) + +1(1.85) +1(1.95)] E 0.147391
(c) o
10
=
1
10 3
[1(1) + 41(1.1) + 21(1.2) + 41(1.3) + 21(1.4) + 41(1.5) + 21(1.6) + 41(1.7)
+ 21(1.8) + 41(1.9) +1(2)]
E 0.147219
11. 1(t) = sin(c
I'2
), {t =
1
2
30
8
=
1
16
(a) T
8
=
1
16 2
1(0) + 21
1
16
+ 21
2
16
+ + 21
7
16
+1
1
2
E 0.451948
(b) A
8
=
1
16
1
32
+1
3
32
+1
5
32
+ +1
13
32
+1
15
32
E 0.451991
(c) o
8
=
1
16 3
1(0) + 41
1
16
+ 21
2
16
+ + 41
7
16
+1
1
2
E 0.451976
13. 1(t) = c
I
I
sint, {t =
4 30
8
=
1
2
(a) T
8
=
1
2 2
1(0) + 21
1
2
+ 21(1) + 21
3
2
+ 21(2) + 21
5
2
+ 21(3) + 21
7
2
+1(4)
E 4.513618
(b) A
8
=
1
2
1
4
+1
3
4
+1
5
4
+1
7
4
+1
9
4
+1
11
4
+1
13
4
+1
15
4
E 4.748256
(c) o
8
=
1
2 3
1(0) + 41
1
2
+ 21(1) + 41
3
2
+ 21(2) + 41
5
2
+ 21(3) + 41
7
2
+1(4)
E 4.675111
15. 1(r) =
cos r
r
, {r =
5 31
8
=
1
2
(a) T
8
=
1
2 2
1(1) + 21
3
2
+ 21(2) + + 21(4) + 21
9
2
+1(5)
E 30.495333
(b) A
8
=
1
2
5
4
+1
7
4
+1
9
4
+1
11
4
+1
13
4
+1
15
4
+1
17
4
+1
19
4
E 30.543321
(c) o
8
=
1
2 3
1(1) + 41
3
2
+ 21(2) + 41
5
2
+ 21(3) + 41
7
2
+ 21(4) + 41
9
2
+1(5)
E 30.526123
330

17. 1(j) =
1
1 +j
5
, {j =
3 30
6
=
1
2
(a) T
6
=
1
2 2
1(0) + 21
1
2
+ 21
2
2
+ 21
3
2
+ 21
4
2
+ 21
5
2
+1(3)
E 1.064275
(b) A
6
=
1
2
1
4
+1
3
4
+1
5
4
+1
7
4
+1
9
4
+1
11
4
E 1.067416
(c) o
6
=
1
2 3
1(0) + 41
1
2
+ 21
2
2
+ 41
3
2
+ 21
4
2
+ 41
5
2
+1(3)
E 1.074915
19. 1(r) = cos(r
2
), {r =
1 30
8
=
1
8
(a) T
8
=
1
8 2
1(0) + 2
1
8
+1
2
8
+ +1
7
8
+1(1)
E 0.902333
A
8
=
1
8
1
16
+1
3
16
+1
5
16
+ +1
15
16
= 0.905620
(b) 1(r) = cos(r
2
), 1
0
(r) = 32rsin(r
2
), 1
00
(r) = 32 sin(r
2
) 34r
2
cos(r
2
). For 0 $ r $ 1, sin and cos are positive,
so |1
00
(r)| = 2 sin(r
2
) + 4r
2
cos(r
2
) $ 2 1 + 4 1 1 = 6 since sin(r
2
) $ 1 and cos
r
2
$ 1 for all r,
and r
2
$ 1 for 0 $ r $ 1. So for n = 8, we take 1 = 6, o = 0, and / = 1 in Theorem 3, to get
|1
T
| $ 6 1
3
(12 8
2
) =
1
128
= 0.0078125 and |1
L
| $
1
256
= 0.00390625. [A better estimate is obtained by noting
from a graph of 1
00
that |1
00
(r)| $ 4 for 0 $ r $ 1.]
(c) Take 1 = 6 [as in part (b)] in Theorem 3. |1
T
| $
1(/ 3o)
3
12n
2
$ 0.0001 C
6(1 30)
3
12n
2
$ 10
34
C
1
2n
2
$
1
10
4
C 2n
2
D 10
4
C n
2
D 5000 C n D 71. Take n = 71 for T
n
. For 1
L
, again take 1 = 6 in
Theorem 3 to get |1
L
| $ 10
34
C 4n
2
D 10
4
C n
2
D 2500 C n D 50. Take n = 50 for A
n
.
21. 1(r) = sinr, {r =
r 30
10
=
r
10
(a) T
10
=
r
10 2
1(0) + 21
r
10
+ 21
2r
10
+ + 21
9r
10
+1()
E 1.983524
A
10
=
r
10
r
20
+1
3r
20
+1
5r
20
+ +1
19r
20
E 2.008248
o
10
=
r
10 3
1(0) + 41
r
10
+ 21
2r
10
+ 41
3r
10
+ + 41
9r
10
+1()
E 2.000110
Since 1 =

r
0
sinrdr =

3cos r
r
0
= 1 3(31) = 2, 1
T
= 1 3T
10
E 0.016476, 1
L
= 1 3A
10
E 30.008248,
and 1
S
= 1 3o
10
E 30.000110.
(b) 1(r) = sinr i

1
(n)
(r)
$ 1, so take 1 = 1 for all error estimates.

|1
T
| $
1(/ 3o)
3
12n
2
=
1( 30)
3
12(10)
2
=

3
1200
E 0.025839. |1
L
| $
|1
T
|
2
=

3
2400
E 0.012919.
|1
S
| $
1(/ 3o)
5
180n
4
=
1( 30)
5
180(10)
4
=

5
1,800,000
E 0.000170.
The actual error is about 64% of the error estimate in all three cases.
(c) |1
T
| $ 0.00001 C

3
12n
2
$
1
10
5
C n
2
D
10
5
3
12
i n D 508.3. Take n = 509 for T
n.
|1
L
| $ 0.00001 C

3
24n
2
$
1
10
5
C n
2
D
10
5
3
24
i n D 359.4. Take n = 360 for A
n.
|1
S
| $ 0.00001 C

5
180n
4
$
1
10
5
C n
4
D
10
5
5
180
i n D 20.3.
Take n = 22 for o
n
(since n must be even).

331
23. (a) Using a CAS, we differentiate 1(r) = c
cos i
twice, and nd that
1
00
(r) = c
cos i
(sin
2
r 3cos r). From the graph, we see that the maximum
value of |1
00
(r)| occurs at the endpoints of the interval [0. 2].
Since 1
00
(0) = 3c, we can use 1 = c or 1 = 2.8.
(b) A CAS gives A
10
E 7.954926518. (In Maple, use student[middlesum].)
(c) Using Theorem 3 for the Midpoint Rule, with 1 = c, we get |1
L
| $
c(2 30)
3
24 10
2
E 0.280945995.
With 1 = 2.8, we get |1
L
| $
2.8(2 30)
3
24 10
2
= 0. 289391916.
(d) A CAS gives 1 E 7.954926521.
(e) The actual error is only about 3 10
39
, much less than the estimate in part (c).
(f) We use the CAS to differentiate twice more, and then graph
1
(4)
(r) = c
cos i
(sin
4
r 36 sin
2
r cos r + 3 37 sin
2
r + cos r).
From the graph, we see that the maximum value of
1
(4)
(r)
occurs at the
endpoints of the interval [0. 2]. Since 1
(4)
(0) = 4c, we can use 1 = 4c
or 1 = 10.9.
(g) A CAS gives o
10
E 7.953789422. (In Maple, use student[simpson].)
(h) Using Theorem 4 with 1 = 4c, we get |1
S
| $
4c(2 30)
5
180 10
4
E 0.059153618.
With 1 = 10.9, we get |1
S
| $
10.9(2 30)
5
180 10
4
E 0.059299814.
(i) The actual error is about 7.954926521 37.953789422 E 0.00114. This is quite a bit smaller than the estimate in part (h),
though the difference is not nearly as great as it was in the case of the Midpoint Rule.
( j) To ensure that |1
S
| $ 0.0001, we use Theorem 4: |1
S
| $
4c(2)
5
180 n
4
$ 0.0001 i
4c(2)
5
180 0.0001
$ n
4
i
n
4
D 5,915,362 C n D 49.3. So we must take n D 50 to ensure that |1 3o
n
| $ 0.0001.
(1 = 10.9 leads to the same value of n.)
25. 1 =

1
0
rc
i
dr = [(r 31)c
i
]
1
0
= 0 3(31) = 1, 1(r) = rc
i
, {r = 1n
n = 5: 1
5
=
1
5
[1(0) +1(0.2) +1(0.4) +1(0.6) +1(0.8)] E 0.742943
1
5
=
1
5
[1(0.2) +1(0.4) +1(0.6) +1(0.8) +1(1)] E 1.286599
T
5
=
1
5 2
[1(0) + 21(0.2) + 21(0.4) + 21(0.6) + 21(0.8) +1(1)] E 1.014771
A
5
=
1
5
[1(0.1) +1(0.3) +1(0.5) +1(0.7) +1(0.9)] E 0.992621
1
1
= 1 31
5
E 1 30.742943 = 0.257057
1
T
E 1 31.286599 = 30.286599
1
T
E 1 31.014771 = 30.014771
1
L
E 1 30.992621 = 0.007379
[parts or Formula 96]
332

n = 10: 1
10
=
1
10
[1(0) +1(0.1) +1(0.2) + +1(0.9)] E 0.867782
1
10
=
1
10
[1(0.1) +1(0.2) + +1(0.9) +1(1)] E 1.139610
T
10
=
1
10 2
{1(0) + 2[1(0.1) +1(0.2) + +1(0.9)] +1(1)} E 1.003696
A
10
=
1
10
[1(0.05) +1(0.15) + +1(0.85) +1(0.95)] E 0.998152
1
1
= 1 31
10
E 1 30.867782 = 0.132218
1
T
E 1 31.139610 = 30.139610
1
T
E 1 31.003696 = 30.003696
1
L
E 1 30.998152 = 0.001848
n = 20: 1
20
=
1
20
[1(0) +1(0.05) +1(0.10) + +1(0.95)] E 0.932967
1
20
=
1
20
[1(0.05) +1(0.10) + +1(0.95) +1(1)] E 1.068881
T
20
=
1
20 2
{1(0) + 2[1(0.05) +1(0.10) + +1(0.95)] +1(1)} E 1.000924
A
20
=
1
20
[1(0.025) +1(0.075) +1(0.125) + +1(0.975)] E 0.999538
1
1
= 1 31
20
E 1 30.932967 = 0.067033
1
T
E 1 31.068881 = 30.068881
1
T
E 1 31.000924 = 30.000924
1
L
E 1 30.999538 = 0.000462
n 1
n
1
n
T
n
A
n
5 0.742943 1.286599 1.014771 0.992621
10 0.867782 1.139610 1.003696 0.998152
20 0.932967 1.068881 1.000924 0.999538
n 1
1
1
T
1
T
1
L
5 0.257057 30.286599 30.014771 0.007379
10 0.132218 30.139610 30.003696 0.001848
20 0.067033 30.068881 30.000924 0.000462
Observations:
1. 1
1
and 1
T
are always opposite in sign, as are 1
T
and 1
L
.
2. As n is doubled, 1
1
and 1
T
are decreased by about a factor of 2, and 1
T
and 1
L
are decreased by a factor of about 4.
3. The Midpoint approximation is about twice as accurate as the Trapezoidal approximation.
4. All the approximations become more accurate as the value of n increases.
5. The Midpoint and Trapezoidal approximations are much more accurate than the endpoint approximations.
27. 1 =

2
0
r
4
dr =

1
5
r
5
2
0
=
32
5
30 = 6.4, 1(r) = r
4
, {r =
2 30
n
=
2
n
n = 6: T
6
=
2
6 2
1(0) + 2
1
3
+1
2
3
+1
3
3
+1
4
3
+1
5
3
+1(2)
E 6.695473
A
6
=
2
6
1
6
+1
3
6
+1
5
6
+1
7
6
+1
9
6
+1
11
6
E 6.252572
o
6
=
2
6 3
1(0) + 41
1
3
+ 21
2
3
+ 41
3
3
+ 21
4
3
+ 41
5
3
+1(2)
E 6.403292
1
T
= 1 3T
6
E 6.4 36.695473 = 30.295473
1
L
E 6.4 36.252572 = 0.147428
1
S
E 6.4 36.403292 = 30.003292

333
n = 12: T
12
=
2
12 2
1(0) + 2
1
6
+1
2
6
+1
3
6
+ +1
11
6
+1(2)
E 6.474023
A
6
=
2
12
1
12
+1
3
12
+1
5
12
+ +1
23
12
E 6.363008
o
6
=
2
12 3
1(0) + 41
1
6
+ 21
2
6
+ 41
3
6
+ 21
4
6
+ + 41
11
6
+1(2)
E 6.400206
1
T
= 1 3T
12
E 6.4 36.474023 = 30.074023
1
L
E 6.4 36.363008 = 0.036992
1
S
E 6.4 36.400206 = 30.000206
n T
n
A
n
o
n
6 6.695473 6.252572 6.403292
12 6.474023 6.363008 6.400206
n 1
T
1
L
1
S
6 30.295473 0.147428 30.003292
12 30.074023 0.036992 30.000206
Observations:
1. 1
T
and 1
L
are opposite in sign and decrease by a factor of about 4 as n is doubled.
2. The Simpsons approximation is much more accurate than the Midpoint and Trapezoidal approximations, and 1
S
seems to
decrease by a factor of about 16 as n is doubled.
29. {r = (/ 3o)n = (6 30)6 = 1
(a) T
6
=
{i
2
[1(0) + 21(1) + 21(2) + 21(3) + 21(4) + 21(5) +1(6)]
E
1
2
[3 + 2(5) + 2(4) + 2(2) + 2(2.8) + 2(4) + 1]
=
1
2
(39.6) = 19.8
(b) A
6
= {r[1(0.5) +1(1.5) +1(2.5) +1(3.5) +1(4.5) +1(5.5)]
E 1[4.5 + 4.7 + 2.6 + 2.2 + 3.4 + 3.2]
= 20.6
(c) o
6
=
{i
3
[1(0) + 41(1) + 21(2) + 41(3) + 21(4) + 41(5) +1(6)]
E
1
3
[3 + 4(5) + 2(4) + 4(2) + 2(2.8) + 4(4) + 1]
=
1
3
(61.6) = 20.53
31. (a) We are given the function values at the endpoints of 8 intervals of length 0.4, so well use the Midpoint Rule with
n = 82 = 4 and {r = (3.2 30)4 = 0.8.
3.2
0
1(r) dr E A
4
= 0.8[1(0.4) +1(1.2) +1(2.0) +1(2.8)] = 0.8[6.5 + 6.4 + 7.6 + 8.8]
= 0.8(29.3) = 23.44
(b) 34 $ 1
00
(r) $ 1 i |1
00
(r)| $ 4, so use 1 = 4, o = 0, / = 3.2, and n = 4 in Theorem 3.
So |1
L
| $
4(3.2 30)
3
24(4)
2
=
128
375
= 0.3413.
334

33. By the Net Change Theorem, the increase in velocity is equal to
6
0
o(t) dt. We use Simpsons Rule with n = 6 and
{t = (6 30)6 = 1 to estimate this integral:
6
0
o(t) dt E o
6
=
1
3
[o(0) + 4o(1) + 2o(2) + 4o(3) + 2o(4) + 4o(5) +o(6)]
E
1
3
[0 + 4(0.5) + 2(4.1) + 4(9.8) + 2(12.9) + 4(9.5) + 0] =
1
3
(113.2) = 37.73 fts
35. By the Net Change Theorem, the energy used is equal to
6
0
1(t) dt. We use Simpsons Rule with n = 12 and
{t =
6 30
12
=
1
2
to estimate this integral:
6
0
1(t) dt E o
12
=
1'2
3
[1(0) + 41(0.5) + 21(1) + 41(1.5) + 21(2) + 41(2.5) + 21(3)
+ 41(3.5) + 21(4) + 41(4.5) + 21(5) + 41(5.5) +1(6)]
=
1
6
[1814 + 4(1735) + 2(1686) + 4(1646) + 2(1637) + 4(1609) + 2(1604)
+ 4(1611) + 2(1621) + 4(1666) + 2(1745) + 4(1886) + 2052]
=
1
6
(61,064) = 10,177.3 megawatt-hours
37. Let j = 1(r) denote the curve. Using cylindrical shells, \ =

10
2
2r1(r) dr = 2
10
2
r1(r) dr = 21
1
.
Now use Simpsons Rule to approximate 1
1
:
1
1
E o
8
=
10 32
3(8)
[21(2) + 4 31(3) + 2 41(4) + 4 51(5) + 2 61(6) + 4 71(7) + 2 81(8) + 4 91(9) + 101(10)]
E
1
3
[2(0) + 12(1.5) + 8(1.9) + 20(2.2) + 12(3.0) + 28(3.8) + 16(4.0) + 36(3.1) + 10(0)]
=
1
3
(395.2)
Thus, \ E 2
1
3
(395.2) E 827.7 or 828 cubic units.
39. Using disks, \ =

5
1
(c
31'i
)
2
dr =
5
1
c
32'i
dr = 1
1
. Now use Simpsons Rule with 1(r) = c
32'i
to approximate
1
1
. 1
1
E o
8
=
5 31
3(8)
[1(1) + 41(1.5) + 21(2) + 41(2.5) + 21(3) + 41(3.5) + 21(4) + 41(4.5) +1(5)] E
1
6
(11.4566)
Thus, \ E
1
6
(11.4566) E 6.0 cubic units.
41. 1(0) =
`
2
sin
2
I
I
2
, where I =
`d sin0
`
, ` = 10,000, d = 10
34
, and ` = 632.8 10
39
. So 1(0) =
(10
4
)
2
sin
2
I
I
2
,
where I =
(10
4
)(10
34
) sin0
632.8 10
39
. Now n = 10 and {0 =
10
36
3(310
36
)
10
= 2 10
37
, so
A
10
= 2 10
37
[1(30.0000009) +1(30.0000007) + +1(0.0000009)] E 59.4.
43. Consider the function 1 whose graph is shown. The area
2
0
1(r) dr
is close to 2. The Trapezoidal Rule gives
T
2
=
2 30
2 2
[1(0) + 21(1) +1(2)] =
1
2
[1 + 2 1 + 1] = 2.
The Midpoint Rule gives A
2
=
2 30
2
[1(0.5) +1(1.5)] = 1[0 + 0] = 0,
so the Trapezoidal Rule is more accurate.
SECTION 7.8 IMPROPER INTEGRALS

335
45. Since the Trapezoidal and Midpoint approximations on the interval [o. /] are the sums of the Trapezoidal and Midpoint
approximations on the subintervals [r
.31
. r
.
], j = 1. 2. . . . . n, we can focus our attention on one such interval. The condition
1
00
(r) < 0 for o $ r $ / means that the graph of 1 is concave down as in Figure 5. In that gure, T
n
is the area of the
trapezoid Q11,
l
a
1(r) dr is the area of the region Q111, and A
n
is the area of the trapezoid 1C1, so
T
n
<

l
a
1(r) dr < A
n
. In general, the condition 1
00
< 0 implies that the graph of 1 on [o. /] lies above the chord joining the
points (o. 1(o)) and (/. 1(/)). Thus,
l
a
1(r) dr T
n
. Since A
n
is the area under a tangent to the graph, and since 1
00
< 0
implies that the tangent lies above the graph, we also have A
n

l
a
1(r) dr. Thus, T
n
<

l
a
1(r) dr < A
n
.
47. T
n
=
1
2
{r[1(r
0
) + 21(r
1
) + + 21(r
n31
) +1(r
n
)] and
A
n
= {r[1(r
1
) +1(r
2
) + +1(r
n31
) +1(r
n
)], where r
.
=
1
2
(r
.31
+r
.
). Now
T
2n
=
1
2
1
2
{r
[1(r
0
) + 21(r
1
) + 21(r
1
) + 21(r
2
) + 21(r
2
) + + 21(r
n31
) + 21(r
n31
) + 21(r
n
) +1(r
n
)]
so
1
2
(T
n
+A
n
) =
1
2
T
n
+
1
2
A
n
=
1
4
{r[1(r
0
) + 21(r
1
) + + 21(r
n31
) +1(r
n
)] +
1
4
{r[21(r
1
) + 21(r
2
) + + 21(r
n31
) + 21(r
n
)]
= T
2n
7.8 Improper Integrals
1. (a) Since
"
1
r
4
c
3i
4
dr has an innite interval of integration, it is an improper integral of Type I.
(b) Since j = sec r has an innite discontinuity at r =
r
2
,
r'2
0
sec rdr is a Type II improper integral.
(c) Since j =
r
(r 32)(r 33)
has an innite discontinuity at r = 2,
2
0
r
r
2
35r + 6
dr is a Type II improper integral.
(d) Since
0
3"
1
r
2
+ 5
dr has an innite interval of integration, it is an improper integral of Type I.
3. The area under the graph of j = 1r
3
= r
33
between r = 1 and r = t is
(t) =

I
1
r
33
dr =

3
1
2
r
32
I
1
= 3
1
2
t
32
3
3
1
2
=
1
2
31
2t
2
. So the area for 1 $ r $ 10 is

(10) = 0.5 30.005 = 0.495, the area for 1 $ r $ 100 is (100) = 0.5 30.00005 = 0.49995, and the area for
1 $ r $ 1000 is (1000) = 0.5 30.0000005 = 0.4999995. The total area under the curve for r D 1 is
lim
I<"
(t) = lim
I<"
1
2
31(2t
2
)
=
1
2
.
336

5. 1 =

"
1
1
(3r + 1)
2
dr = lim
I<"
I
1
1
(3r + 1)
2
dr. Now
1
(3r + 1)
2
dr =
1
3
1
&
2
d& [u = 3r + 1, Ju = 3 Jr] = 3
1
3&
+C = 3
1
3(3r + 1)
+C,
so 1 = lim
I<"
3
1
3(3r + 1)
I
1
= lim
I<"
3
1
3(3t + 1)
+
1
12
= 0 +
1
12
=
1
12
. Convergent
7.

31
3"
1
I
2 3n
dn = lim
I<3"
31
I
1
I
2 3n
dn = lim
I<3"
32
I
2 3n
31
I
[u = 2 u, Ju = Ju]
= lim
I<3"
32
I
3 + 2
I
2 3t
= ". Divergent
9.

"
4
c
3u'2
dj = lim
I<"
I
4
c
3u'2
dj = lim
I<"
32c
3u'2
I
4
= lim
I<"
(32c
3I'2
+ 2c
32
) = 0 + 2c
32
= 2c
32
.
Convergent
11.

"
3"
rdr
1 +r
2
=

0
3"
rdr
1 +r
2
+
"
0
rdr
1 +r
2
and
0
3"
rdr
1 +r
2
= lim
I<3"
1
2
ln
1 +r
2
0
t
= lim
I<3"
0 3
1
2
ln
1 +t
2
= 3". Divergent
13.

"
3"
rc
3i
2
dr =

0
3"
rc
3i
2
dr +
"
0
rc
3i
2
dr.
0
3"
rc
3i
2
dr = lim
I<3"
3
1
2
c
3i
2
0
I
= lim
I<3"
3
1
2
1 3c
3I
2
= 3
1
2
1 = 3
1
2
, and
"
0
rc
3i
2
dr = lim
I<"
3
1
2
c
3i
2
I
0
= lim
I<"
3
1
2
c
3I
2
31
= 3
1
2
(31) =
1
2
.
Therefore,
"
3"
rc
3i
2
dr = 3
1
2
+
1
2
= 0. Convergent
15.

"
2r
sin0 d0 = lim
I<"
I
2r
sin0 d0 = lim
I<"
3cos 0
I
2r
= lim
I<"
(3cos t + 1). This limit does not exist, so the integral is
divergent. Divergent
17.

"
1
r + 1
r
2
+ 2r
dr = lim
I<"
I
1
1
2
(2r + 2)
r
2
+ 2r
dr =
1
2
lim
I<"
ln(r
2
+ 2r)
I
1
=
1
2
lim
I<"
ln(t
2
+ 2t) 3ln3
= ".
Divergent
19.

"
0
cc
35s
dc = lim
I<"
I
0
cc
35s
dc = lim
I<"
3
1
5
cc
35s
3
1
25
c
35s

by integration by
parts with u = s
= lim
I<"
3
1
5
tc
35I
3
1
25
c
35I
+
1
25
= 0 30 +
1
25
[by lHospitals Rule]
=
1
25
. Convergent

337
21.

"
1
lnr
r
dr = lim
I<"
(lnr)
2
2
I
1
by substitution with
u = ln r, Ju = Jr/r
= lim
I<"
(lnt)
2
2
= ". Divergent
23.

"
3"
r
2
9 +r
6
dr =

0
3"
r
2
9 +r
6
dr +
"
0
r
2
9 +r
6
dr = 2
"
0
r
2
9 +r
6
dr [since the integrand is even].
Now

r
2
dr
9 +r
6
u = r
3
Ju = 3r
2
Jr
=

1
3
d&
9 +&
2
u = 3u
Ju = 3 Ju
=

1
3
(3 d)
9 + 9
2
=
1
9
d
1 +
2
=
1
9
tan
31
+C =
1
9
tan
31
&
3
+C =
1
9
tan
31
r
3
3
+C,
so 2
"
0
r
2
9 +r
6
dr = 2 lim
I<"
I
0
r
2
9 +r
6
dr = 2 lim
I<"
1
9
tan
31
r
3
3
I
0
= 2 lim
I<"
1
9
tan
31
t
3
3
=
2
9

2
=

9
.
Convergent
25.

"
c
1
r(lnr)
3
dr = lim
I<"
I
c
1
r(lnr)
3
dr = lim
I<"
ln I
1
&
33
d&

u = ln r,
Ju = Jr/r
= lim
I<"
3
1
2&
2
ln I
1
= lim
I<"
3
1
2 (lnt)
2
+
1
2
= 0 +
1
2
=
1
2
. Convergent
27.

1
0
3
r
5
dr = lim
I<0
+
1
I
3r
35
dr = lim
I<0
+
3
3
4r
4
1
I
= 3
3
4
lim
I<0
+
1 3
1
t
4
= ". Divergent
29.

14
32
dr
4
I
r + 2
= lim
I<32
+
14
I
(r + 2)
31'4
dr = lim
I<32
+
4
3
(r + 2)
3'4
14
I
=
4
3
lim
I<32
+
16
3'4
3(t + 2)
3'4
=
4
3
(8 30) =
32
3
. Convergent
31.

3
32
dr
r
4
=

0
32
dr
r
4
+
3
0
dr
r
4
, but
0
32
dr
r
4
= lim
I<0
3
r
33
3
I
32
= lim
I<0
3
1
3t
3
3
1
24
= ". Divergent
33. There is an innite discontinuity at r = 1.

33
0
(r 31)
31'5
dr =

1
0
(r 31)
31'5
dr +
33
1
(r 31)
31'5
dr. Here
1
0
(r 31)
31'5
dr = lim
I<1
I
0
(r 31)
31'5
dr = lim
I<1
5
4
(r 31)
4'5
I
0
= lim
I<1
5
4
(t 31)
4'5
3
5
4
= 3
5
4
and
33
1
(r 31)
31'5
dr = lim
I<1
+
33
I
(r 31)
31'5
dr = lim
I<1
+
5
4
(r 31)
4'5
33
I
= lim
I<1
+
5
4
16 3
5
4
(t 31)
4'5
= 20.
Thus,
33
0
(r 31)
31'5
dr = 3
5
4
+ 20 =
75
4
. Convergent
35. 1 =

3
0
dr
r
2
36r + 5
=

3
0
dr
(r 31)(r 35)
= 1
1
+1
2
=

1
0
dr
(r 31)(r 35)
+
3
1
dr
(r 31)(r 35)
.
Now
1
(r 31)(r 35)
=

r 31
+
1
r 35
i 1 = (r 35) +1(r 31).
Set r = 5 to get 1 = 41, so 1 =
1
4
. Set r = 1 to get 1 = 34, so = 3
1
4
. Thus
338

1
1
= lim
I<1
I
0
3
1
4
r 31
+
1
4
r 35
dr = lim
I<1
3
1
4
ln|r 31| +
1
4
ln|r 35|
I
0
= lim
I<1
3
1
4
ln|t 31| +
1
4
ln|t 35|
3
1
4
ln|31| +
1
4
ln|35|
= ", since lim

I<1
3
1
4
ln|t 31|
= ".
Since 1
1
is divergent, 1 is divergent.
37.

0
31
c
1'i
r
3
dr = lim
I<0
I
31
1
r
c
1'i
1
r
2
dr = lim
I<0
1'I
31
&c
u
(3d&)

u = 1/r,
Ju = Jr/r
2
= lim
I<0
(& 31)c
u
31
1'I
use parts
or Formula 96
= lim
I<0
32c
31
3
1
t
31
c
1'I
= 3
2
c
3 lim
s<3"
(c 31)c
s
[s = 1/|] = 3
2
c
3 lim
s<3"
c 31
c
3s
H
= 3
2
c
3 lim
s<3"
1
3c
3s
= 3
2
c
30 = 3
2
c
. Convergent
39. 1 =

2
0
:
2
ln: d: = lim
I<0
+
2
I
:
2
ln: d: = lim
I<0
+
:
3
3
2
(3 ln: 31)
2
I
integrate by parts
or use Formula 101
= lim
I<0
+
8
9
(3 ln2 31) 3
1
9
t
3
(3 lnt 31)
=
8
3
ln2 3
8
9
3
1
9
lim
I<0
+
t
3
(3 lnt 31)
=
8
3
ln2 3
8
9
3
1
9
1.
Now 1 = lim
I<0
+
t
3
(3 lnt 31)
= lim
I<0
+
3 lnt 31
t
33
H
= lim
I<0
+
3t
33t
4
= lim
I<0
+
3t
3
= 0.
Thus, 1 = 0 and 1 =
8
3
ln2 3
8
9
. Convergent
41. Area =

1
3"
c
i
dr = lim
I<3"
c
i
1
I
= c 3 lim
I<3"
c
I
= c
43. Area =

"
3"
2
r
2
+ 9
dr = 2 2
"
0
1
r
2
+ 9
dr = 4 lim
I<"
I
0
1
r
2
+ 9
dr
= 4 lim
I<"
1
3
tan
31
r
3
I
0
=
4
3
lim
I<"
tan
31
t
3
30
=
4
3

2
=
2
3
45. Area =

r'2
0
sec
2
rdr = lim
I<(r'2)
I
0
sec
2
rdr = lim
I<(r'2)
[tanr]
I
0
= lim
I<(r'2)
(tant 30) = "

Innite area

339
47. (a)
t

I
1
o(r) dr
2 0.447453
5 0.577101
10 0.621306
100 0.668479
1000 0.672957
10,000 0.673407
o(r) =
sin
2
r
r
2
.
It appears that the integral is convergent.
(b) 31 $ sinr $ 1 i 0 $ sin
2
r $ 1 i 0 $
sin
2
r
r
2
$
1
r
2
. Since
"
1
1
r
2
dr is convergent
[Equation 2 with j = 2 1],

"
1
sin
2
r
r
2
dr is convergent by the Comparison Theorem.
(c) Since
"
1
1(r) dr is nite and the area under o(r) is less than the area under 1(r)
on any interval [1. t],
"
1
o(r) dr must be nite; that is, the integral is convergent.
49. For r 0,
r
r
3
+ 1
<
r
r
3
=
1
r
2
.

"
1
1
r
2
dr is convergent by Equation 2 with j = 2 1, so
"
1
r
r
3
+ 1
dr is convergent
by the Comparison Theorem.

1
0
r
r
3
+ 1
dr is a constant, so
"
0
r
r
3
+ 1
dr =

1
0
r
r
3
+ 1
dr +
"
1
r
r
3
+ 1
dr is also
convergent.
51. For r 1, 1(r) =
r + 1
I
r
4
3r

r + 1
I
r
4

r
r
2
=
1
r
, so
"
2
1(r) dr diverges by comparison with
"
2
1
r
dr, which diverges
by Equation 2 with j = 1 $ 1. Thus,
"
1
1(r) dr =

2
1
1(r) dr +
"
2
1(r) dr also diverges.
53. For 0 < r $ 1,
sec
2
r
r
I
r

1
r
3'2
. Now
1 =

1
0
r
33'2
dr = lim
I<0
+
1
I
r
33'2
dr = lim
I<0
+
32r
31'2
1
I
= lim
I<0
+
32 +
2
I
t
= ", so 1 is divergent, and by

comparison,
1
0
sec
2
r
r
I
r
is divergent.
340

55.

"
0
dr
I
r(1 +r)
=

1
0
dr
I
r(1 +r)
+
"
1
dr
I
r(1 +r)
= lim
I<0
+
1
I
dr
I
r(1 +r)
+ lim
I<"
I
1
dr
I
r(1 +r)
. Now
dr
I
r(1 +r)
=

2&d&
&(1 +&
2
)
u =
s
r, r = u
2
,
Jr = 2uJu
= 2
d&
1 +&
2
= 2 tan
31
& +C = 2 tan
31
I
r +C, so
"
0
dr
I
r(1 +r)
= lim
I<0
+
2 tan
31
I
r
1
I
+ lim
I<"
2 tan
31
I
r
I
1
= lim
I<0
+
r
4
32 tan
31
I
t
+ lim
I<"
2 tan
31
I
t 32
r
4
=
r
2
30 + 2
r
2
3
r
2
= .
57. If j = 1, then
1
0
dr
r
= lim
I<0
+
1
I
dr
r
= lim
I<0
+
[lnr]
1
I
= ". Divergent.
If j 6= 1, then
1
0
dr
r
= lim
I<0
+
1
I
dr
r
[note that the integral is not improper if j < 0]

= lim
I<0
+
r
3+1
3j + 1
1
I
= lim
I<0
+
1
1 3j
1 3
1
t
31
If j 1, then j 31 0, so
1
t
31
<"as t <0
+
, and the integral diverges.
If j < 1, then j 31 < 0, so
1
t
31
<0 as t <0
+
and
1
0
dr
r
=
1
1 3j
lim
I<0
+
1 3t
13
=
1
1 3j
.
Thus, the integral converges if and only if j < 1, and in that case its value is
1
1 3j
.
59. First suppose j = 31. Then
1
0
r
lnrdr =

1
0
lnr
r
dr = lim
I<0
+
1
I
lnr
r
dr = lim
I<0
+
1
2
(lnr)
2
1
I
= 3
1
2
lim
I<0
+
(lnt)
2
= 3", so the
integral diverges. Now suppose j 6= 31. Then integration by parts gives
lnrdr =
r
+1
j + 1
lnr 3
j + 1
dr =
r
+1
j + 1
lnr 3
r
+1
(j + 1)
2
+C. If j < 31, then j + 1 < 0, so
1
0
r
lnrdr = lim
I<0
+
r
+1
j + 1
lnr 3
r
+1
(j + 1)
2
1
I
=
31
(j + 1)
2
3
1
j + 1
lim
I<0
+
t
+1
lnt 3
1
j + 1
= ".
If j 31, then j + 1 0 and
1
0
r
lnrdr =
31
(j + 1)
2
3
1
j + 1
lim
I<0
+
lnt 31(j + 1)
t
3(+1)
H
=
31
(j + 1)
2
3
1
j + 1
lim
I<0
+
1t
3(j + 1)t
3(+2)
=
31
(j + 1)
2
+
1
(j + 1)
2
lim
I<0
+
t
+1
=
31
(j + 1)
2
Thus, the integral converges to 3
1
(j + 1)
2
if j 31 and diverges otherwise.

341
61. (a) 1 =

"
3"
rdr =

0
3"
rdr +
"
0
rdr, and
"
0
rdr = lim
I<"
I
0
rdr = lim
I<"
1
2
r
2
I
0
= lim
I<"
1
2
t
2
30
= ",
so 1 is divergent.
(b)

I
3I
rdr =

1
2
r
2
I
3I
=
1
2
t
2
3
1
2
t
2
= 0, so lim
I<"
I
3I
rdr = 0. Therefore,
"
3"
rdr 6= lim
I<"
I
3I
rdr.
63. Volume =

"
1

1
r
2
dr = lim
I<"
I
1
dr
r
2
= lim
I<"
3
1
r
I
1
= lim
I<"
1 3
1
t
= < ".
65. Work =

"
T
1 dv = lim
I<"
I
T
GiA
v
2
dv = lim
I<"
GiA
1
1
3
1
t
=
GiA
1
. The initial kinetic energy provides the work,
so
1
2
i
2
0
=
GiA
1
i
0
=
2GA
1
.
67. We would expect a small percentage of bulbs to burn out in the rst few hundred hours, most of the bulbs to burn out after
close to 700 hours, and a few overachievers to burn on and on.
(a)
(b) v(t) = 1
0
(t) is the rate at which the fraction 1(t) of burnt-out bulbs increases as t increases. This could be interpreted as
a fractional burnout rate.
(c)

"
0
v(t) dt = lim
i<"
1(r) = 1, since all of the bulbs will eventually burn out.
69. 1 =

"
a
1
r
2
+ 1
dr = lim
I<"
I
a
1
r
2
+ 1
dr = lim
I<"
tan
31
r
I
a
= lim
I<"
tan
31
t 3tan
31
o
=
r
2
3tan
31
o.
1 < 0.001 i
r
2
3tan
31
o < 0.001 i tan
31
o
r
2
30.001 i o tan
r
2
30.001
E 1000.
71. (a) 1(c) =

"
0
1(t)c
3sI
dt =

"
0
c
3sI
dt = lim
n<"
3
c
3sI
c
n
0
= lim
n<"
c
3sn
3c
+
1
c
. This converges to
1
c
only if c 0.
Therefore 1(c) =
1
c
with domain {c | c 0}.
(b) 1(c) =

"
0
1(t)c
3sI
dt =

"
0
c
I
c
3sI
dt = lim
n<"
n
0
c
I(13s)
dt = lim
n<"
1
1 3c
c
I(13s)
n
0
= lim
n<"
c
(13s)n
1 3c
3
1
1 3c
This converges only if 1 3c < 0 i c 1, in which case 1(c) =

1
c 31
with domain {c | c 1}.
342

(c) 1(c) =

"
0
1(t)c
3sI
dt = lim
n<"
n
0
tc
3sI
dt. Use integration by parts: let & = t, d = c
3sI
dt i d& = dt,
= 3
c
3sI
c
. Then 1(c) = lim
n<"
3
t
c
c
3sI
3
1
c
2
c
3sI
n
0
= lim
n<"
3n
cc
sn
3
1
c
2
c
sn
+ 0 +
1
c
2
=
1
c
2
only if c 0.
Therefore, 1(c) =
1
c
2
and the domain of 1 is {c | c 0}.
73. G(c) =
"
0
1
0
(t)c
3sI
dt. Integrate by parts with & = c
3sI
, d = 1
0
(t) dt i d& = 3cc
3sI
, = 1(t):
G(c) = lim
n<"
1(t)c
3sI
n
0
+c
"
0
1(t)c
3sI
dt = lim
n<"
1(n)c
3sn
31(0) +c1(c)
But 0 $ 1(t) $ Ac
aI
i 0 $ 1(t)c
3sI
$ Ac
aI
c
3sI
and lim
I<"
Ac
I(a3s)
= 0 for c o. So by the Squeeze Theorem,
lim
I<"
1(t)c
3sI
= 0 for c o i G(c) = 0 31(0) +c1(c) = c1(c) 31(0) for c o.
75. We use integration by parts: let & = r, d = rc
3i
2
dr i d& = dr, = 3
1
2
c
3i
2
. So
"
0
r
2
c
3i
2
dr = lim
I<"
3
1
2
rc
3i
2
I
0
+
1
2
"
0
c
3i
2
dr = lim
I<"
3
t
2c
I
2
+
1
2
"
0
c
3i
2
dr =
1
2
"
0
c
3i
2
dr
(The limit is 0 by lHospitals Rule.)
77. For the rst part of the integral, let r = 2 tan0 i dr = 2 sec
2
0 d0.
1
I
r
2
+ 4
dr =

2 sec
2
0
2 sec 0
d0 =

sec 0 d0 = ln|sec 0 + tan0|.
From the gure, tan0 =
r
2
, and sec 0 =
I
r
2
+ 4
2
. So
1 =

"
0
1
I
r
2
+ 4
3
C
r + 2
dr = lim
I<"
ln
I
r
2
+ 4
2
+
r
2
3C ln|r + 2|
I
0
= lim
I<"
ln
I
t
2
+ 4 +t
2
3C ln(t + 2) 3(ln1 3C ln2)
= lim
I<"
ln
I
t
2
+ 4 +t
2 (t + 2)
C
+ ln2
C
= ln
lim
I<"
t +
I
t
2
+ 4
(t + 2)
C
+ ln2
C31
Now 1 = lim
I<"
t +
I
t
2
+ 4
(t + 2)
C
H
= lim
I<"
1 +t
I
t
2
+ 4
C (t + 2)
C31
=
2
C lim
I<"
(t + 2)
C31
.
If C < 1, 1 = "and 1 diverges.
If C = 1, 1 = 2 and 1 converges to ln2 + ln2
0
= ln2.
If C 1, 1 = 0 and 1 diverges to 3".
79. No, 1 =

"
0
1(r) dr must be divergent. Since lim
i<"
1(r) = 1, there must exist an ` such that if r D `, then 1(r) D
1
2
.
Thus, 1 = 1
1
+1
2
=

^
0
1(r) dr +
"
^
1(r) dr, where 1
1
is an ordinary denite integral that has a nite value, and 1
2
is
improper and diverges by comparison with the divergent integral
"
^
1
2
dr.
CHAPTER 7 REVIEW

343
7 Review
1. See Formula 7.1.1 or 7.1.2. We try to choose & = 1(r) to be a function that becomes simpler when differentiated (or at least
not more complicated) as long as d = o
0
(r) dr can be readily integrated to give .
2. See the Strategy for Evaluating
sin
r
rcos
n
rdr on page 462.
3. If
I
o
2
3r
2
occurs, try r = o sin0; if
I
o
2
+r
2
occurs, try r = o tan0, and if
I
r
2
3o
2
occurs, try r = o sec 0. See the
Table of Trigonometric Substitutions on page 467.
4. See Equation 2 and Expressions 7, 9, and 11 in Section 7.4.
5. See the Midpoint Rule, the Trapezoidal Rule, and Simpsons Rule, as well as their associated error bounds, all in Section 7.7.
We would expect the best estimate to be given by Simpsons Rule.
6. See Denitions 1(a), (b), and (c) in Section 7.8.
7. See Denitions 3(b), (a), and (c) in Section 7.8.
8. See the Comparison Theorem after Example 8 in Section 7.8.
1. False. Since the numerator has a higher degree than the denominator,
r
r
2
+ 4
r
2
34
= r +
8r
r
2
34
= r +

r + 2
+
1
r 32
.
3. False. It can be put in the form

r
+
1
r
2
+
C
r 34
.
5. False. This is an improper integral, since the denominator vanishes at r = 1.
4
0
r
r
2
31
dr =

1
0
r
r
2
31
dr +
4
1
r
r
2
31
dr and
1
0
r
r
2
31
dr = lim
I<1
I
0
r
r
2
31
dr = lim
I<1
1
2
ln
r
2
31
I
0
= lim
I<1
1
2
ln
t
2
31
= "
So the integral diverges.
7. False. See Exercise 61 in Section 7.8.
9. (a) True. See the end of Section 7.5.
(b) False. Examples include the functions 1(r) = c
i
2
, o(r) = sin(r
2
), and /(r) =
sinr
r
.
11. False. If 1(r) = 1r, then 1 is continuous and decreasing on [1. ") with lim
i<"
1(r) = 0, but
"
1
1(r) dr is divergent.
13. False. Take 1(r) = 1 for all r and o(r) = 31 for all r. Then
"
a
1(r) dr = " [divergent]
and
"
a
o(r) dr = 3" [divergent], but
"
a
[1(r) +o(r)] dr = 0 [convergent].
344

1.

5
0
r
r + 10
dr =

5
0
1 3
10
r + 10
dr =

r 310 ln(r + 10)
5
0
= 5 310 ln15 + 10 ln10
= 5 + 10 ln
10
15
= 5 + 10 ln
2
3
3.

r'2
0
cos 0
1 + sin0
d0 =

ln(1 + sin0)
r'2
0
= ln2 3ln1 = ln2
5.

r'2
0
sin
3
0 cos
2
0 d0 =

r'2
0
(1 3cos
2
0) cos
2
0 sin0 d0 =

0
1
(1 3&
2
)&
2
(3d&)

u = cos 0,
Ju = sin 0 J0
=

1
0
(&
2
3&
4
) d& =

1
3
&
3
3
1
5
&
5
1
0
=

1
3
3
1
5
30 =
2
15
7. Let & = lnt, d& = dtt. Then
sin(lnt)
t
dt =

sin&d& = 3cos & +C = 3cos(lnt) +C.
9.

4
1
r
3'2
lnrdr

u = ln r,
Ju = Jr/r
Ju = r
3@2
Jr,
u =
2
5
r
5'2
=
2
5
r
5'2
lnr
4
1
3
2
5
4
1
r
3'2
dr =
2
5
(32 ln4 3ln1) 3
2
5
2
5
r
5'2
4
1
=
2
5
(64 ln2) 3
4
25
(32 31) =
128
5
ln2 3
124
25
or
64
5
ln4 3
124
25
11. Let r = sec 0. Then
2
1
I
r
2
31
r
dr =

r'3
0
tan0
sec 0
sec 0 tan0 d0 =

r'3
0
tan
2
0 d0 =

r'3
0
(sec
2
0 31) d0 =

tan0 30
r'3
0
=
I
3 3
r
3
.
13. Let t =
3
I
r. Then t
3
= r and 3t
2
dt = dr, so
c
3
I
i
dr =

c
I
3t
2
dt = 31. To evaluate 1, let & = t
2
,
d = c
I
dt i d& = 2t dt, = c
I
, so 1 =

t
2
c
I
dt = t
2
c
I
3
2tc
I
dt. Now let l = t, d\ = c
I
dt i
dl = dt, \ = c
I
. Thus, 1 = t
2
c
I
32
tc
I
3
c
I
dt
= t
2
c
I
32tc
I
+ 2c
I
+C
1
, and hence
31 = 3c
I
(t
2
32t + 2) +C = 3c
3
I
i
(r
2'3
32r
1'3
+ 2) +C.
15.
r 31
r
2
+ 2r
=
r 31
r(r + 2)
=

r
+
1
r + 2
i r 31 = (r + 2) +1r. Set r = 32 to get 33 = 321, so 1 =
3
2
. Set r = 0
to get 31 = 2, so = 3
1
2
. Thus,
r 31
r
2
+ 2r
dr =

3
1
2
r
+
3
2
r + 2
dr = 3
1
2
ln|r| +
3
2
ln|r + 2| +C.
17. Integrate by parts with & = r, d = sec r tanrdr i d& = dr, = sec r:
rsec r tanrdr = rsec r 3
sec rdr
14
= rsec r 3ln|sec r + tanr| +C.
19.

r + 1
9r
2
+ 6r + 5
dr =

r + 1
(9r
2
+ 6r + 1) + 4
dr =

r + 1
(3r + 1)
2
+ 4
dr

u= 3r + 1,
Ju= 3 Jr
=

1
3
(& 31)
+ 1
&
2
+ 4
1
3
d&
=
1
3

1
3
(& 31) + 3
&
2
+ 4
d&
=
1
9
&
&
2
+ 4
d& +
1
9
2
&
2
+ 2
2
d& =
1
9

1
2
ln(&
2
+ 4) +
2
9

1
2
tan
31
1
2
&
+C
=
1
18
ln(9r
2
+ 6r + 5) +
1
9
tan
31
1
2
(3r + 1)
+C
CHAPTER 7 REVIEW

345
21.

dr
I
r
2
34r
=

dr
(r
2
34r + 4) 34
=

dr
(r 32)
2
32
2
=

2 sec 0 tan0 d0
2 tan0
r 2 = 2 sec 0,
Jr = 2 sec 0 tan 0 J0
=

1
= ln
r 32
2
+
I
r
2
34r
2
+C
1
= ln
r 32 +
I
r
2
34r
+C, where C = C
1
3ln2
23. Let r = tan0, so that dr = sec
2
0 d0. Then
dr
r
I
r
2
+ 1
=

sec
2
0 d0
tan0 sec 0
=

sec 0
tan0
d0
=

csc 0 d0 = ln|csc 0 3cot 0| +C
= ln
I
r
2
+ 1
r
3
1
r
+C = ln
I
r
2
+ 1 31
r
+C
25.
3r
3
3r
2
+ 6r 34
(r
2
+ 1)(r
2
+ 2)
=
r +1
r
2
+ 1
+
Cr +1
r
2
+ 2
i 3r
3
3r
2
+ 6r 34 = (r +1)
r
2
+ 2
+ (Cr +1)
r
2
+ 1
.
Equating the coefcients gives +C = 3, 1 +1 = 31, 2+C = 6, and 21 +1 = 34 i
= 3, C = 0, 1 = 33, and 1 = 2. Now
3r
3
3r
2
+ 6r 34
(r
2
+ 1)(r
2
+ 2)
dr = 3
r 31
r
2
+ 1
dr + 2
dr
r
2
+ 2
=
3
2
ln
r
2
+ 1
33 tan
31
r +
I
2 tan
31
r
I
2
+C.
27.

r'2
0
cos
3
r sin2rdr =

r'2
0
cos
3
r(2 sinr cos r) dr =

r'2
0
2 cos
4
r sinrdr =

3
2
5
cos
5
r
r'2
0
=
2
5
29. The product of an odd function and an even function is an odd function, so 1(r) = r
5
sec r is an odd function.
By Theorem 5.5.7(b),
1
31
r
5
sec rdr = 0.
31. Let & =
I
c
i
31. Then &
2
= c
i
31 and 2&d& = c
i
dr. Also, c
i
+ 8 = &
2
+ 9. Thus,
ln 10
0
c
i
I
c
i
31
c
i
+ 8
dr =

3
0
& 2&d&
&
2
+ 9
= 2
3
0
&
2
&
2
+ 9
d& = 2
3
0
1 3
9
&
2
+ 9
d&
= 2
& 3
9
3
tan
31
&
3
3
0
= 2
(3 33 tan
31
1) 30
= 2
3 33

4
= 6 3
3
2
33. Let r = 2 sin0 i

4 3r
2
3'2
= (2 cos 0)
3
, dr = 2 cos 0 d0, so
r
2
(4 3r
2
)
3'2
dr =

4 sin
2
0
8 cos
3
0
2 cos 0 d0 =

tan
2
0 d0 =

sec
2
0 31
d0
= tan0 30 +C =
r
I
4 3r
2
3sin
31
r
2
+C
346

35.

1
I
r +r
3'2
dr =

dr
r(1 +
I
r)
=

dr
I
r
1 +
I
r
u = 1 +
s
r,
Ju =
Jr
2
s
r
=

2 d&
I
&
=

2&
31'2
d&
= 4
I
& +C = 4
1 +
I
r +C
37.

(cos r + sinr)
2
cos 2rdr =

cos
2
r + 2 sinrcos r + sin
2
r
cos 2rdr =

(1 + sin2r) cos 2rdr
=

cos 2rdr +
1
2
sin4rdr =
1
2
sin2r 3
1
8
cos 4r +C
Or:

(cos r + sinr)
2
cos 2rdr =

(cos r + sinr)
2
(cos
2
r 3sin
2
r) dr
=

(cos r + sinr)
3
(cos r 3sinr) dr =
1
4
(cos r + sinr)
4
+C
1
39. Well integrate 1 =

rc
2i
(1 + 2r)
2
dr by parts with & = rc
2i
and d =
dr
(1 + 2r)
2
. Then d& = (r 2c
2i
+c
2i
1) dr
and = 3
1
2

1
1 + 2r
, so
1 = 3
1
2

rc
2i
1 + 2r
3

3
1
2

c
2i
(2r + 1)
1 + 2r
dr = 3
rc
2i
4r + 2
+
1
2

1
2
c
2i
+C = c
2i
1
4
3
r
4r + 2
+C
Thus,
1'2
0
rc
2i
(1 + 2r)
2
dr =

c
2i
1
4
3
r
4r + 2
1'2
0
= c
1
4
3
1
8
31
1
4
30
=
1
8
c 3
1
4
.
41.

"
1
1
(2r + 1)
3
dr = lim
I<"
I
1
1
(2r + 1)
3
dr = lim
I<"
I
1
1
2
(2r + 1)
33
2 dr = lim
I<"
3
1
4(2r + 1)
2
I
1
= 3
1
4
lim
I<"
1
(2t + 1)
2
3
1
9
= 3
1
4
0 3
1
9
=
1
36
43.

dr
rlnr
u = ln r,
Ju = Jr/r
=

d&
&
= ln|&| +C = ln|lnr| +C. so
"
2
dr
rlnr
= lim
I<"
I
2
dr
rlnr
= lim
I<"
ln|lnr|
I
2
= lim
I<"
[ln(lnt) 3ln(ln2)] = ", so the integral is divergent.
45.

4
0
lnr
I
r
dr = lim
I<0
+
4
I
lnr
I
r
dr
W
= lim
I<0
+
2
I
r lnr 34
I
r
4
I
= lim
I<0
+
(2 2 ln4 34 2) 3
2
I
t lnt 34
I
t
WW
= (4 ln4 38) 3(0 30) = 4 ln4 38
(W) Let & = lnr, d =
1
I
r
dr i d& =
1
r
dr, = 2
I
r. Then
lnr
I
r
dr = 2
I
r lnr 32
dr
I
r
= 2
I
r lnr 34
I
r +C
(WW) lim
I<0
+
2
I
t lnt
= lim
I<0
+
2 lnt
t
31'2
H
= lim
I<0
+
2t
3
1
2
t
33'2
= lim
I<0
+
34
I
t
= 0
47.

1
0
r 31
I
r
dr = lim
I<0
+
1
I
r
I
r
3
1
I
r
dr = lim
I<0
+
1
I
(r
1'2
3r
31'2
) dr = lim
I<0
+
2
3
r
3'2
32r
1'2
1
I
= lim
I<0
+
2
3
32
2
3
t
3'2
32t
1'2
= 3
4
3
30 = 3
4
3
CHAPTER 7 REVIEW

347
49. Let & = 2r + 1. Then
"
3"
dr
4r
2
+ 4r + 5
=

"
3"
1
2
d&
&
2
+ 4
=
1
2
0
3"
d&
&
2
+ 4
+
1
2
"
0
d&
&
2
+ 4
=
1
2
lim
I<3"
1
2
tan
31
1
2
&
0
I
+
1
2
lim
I<"
1
2
tan
31
1
2
&
I
0
=
1
4
0 3
3
r
2
+
1
4
r
2
30
=
r
4
.
51. We rst make the substitution t = r + 1, so ln(r
2
+ 2r + 2) = ln
(r + 1)
2
+ 1
= ln(t
2
+ 1). Then we use parts with
& = ln(t
2
+ 1), d = dt:
ln(t
2
+ 1) dt = t ln(t
2
+ 1) 3
t(2t) dt
t
2
+ 1
= t ln(t
2
+ 1) 32
t
2
dt
t
2
+ 1
= t ln(t
2
+ 1) 32

1 3
1
t
2
+ 1
dt
= t ln(t
2
+ 1) 32t + 2 arctant +C
= (r + 1) ln(r
2
+ 2r + 2) 32r + 2 arctan(r + 1) +1, where 1 = C 32
[Alternatively, we could have integrated by parts immediately with
& = ln(r
2
+ 2r + 2).] Notice from the graph that 1 = 0 where 1 has a
horizontal tangent. Also, 1 is always increasing, and 1 D 0.
53. From the graph, it seems as though
2r
0
cos
2
r sin
3
rdr is equal to 0.
To evaluate the integral, we write the integral as
1 =

2r
0
cos
2
r(1 3cos
2
r) sinrdr and let & = cos r i
d& = 3sinrdr. Thus, 1 =

1
1
&
2
(1 3&
2
)(3d&) = 0.
55.
I
4r
2
34r 33 dr =

(2r 31)
2
34 dr

u = 2r 1,
Ju = 2 Jr
=
I
&
2
32
2
1
2
d&
39
=
1
2
&
2
I
&
2
32
2
3
2
2
2
ln
& +
I
&
2
32
2
+C =
1
4
&
I
&
2
34 3ln
& +
I
&
2
34
+C
=
1
4
(2r 31)
I
4r
2
34r 33 3ln
2r 31 +
I
4r
2
34r 33
+C
cos r
4 + sin
2
rdr =
I
2
2
+&
2
d&
21
=
&
2
I
2
2
+&
2
+
2
2
2
ln
& +
I
2
2
+&
2
+C
=
1
2
sinr
4 + sin
2
r + 2 ln
sinr +
4 + sin
2
r
+C
59. (a)
d
d&
3
1
&
I
o
2
3&
2
3sin
31
&
o
+C
=
1
&
2
I
o
2
3&
2
+
1
I
o
2
3&
2
3
1
1 3&
2
o
2

1
o
=

o
2
3&
2
31'2
1
&
2
o
2
3&
2
+ 1 31
=
I
o
2
3&
2
&
2
348

(b) Let & = o sin0 i d& = o cos 0 d0, o
2
3&
2
= o
2
1 3sin
2
0
= o
2
cos
2
0.
I
o
2
3&
2
&
2
d& =

o
2
cos
2
0
o
2
sin
2
0
d0 =

1 3sin
2
0
sin
2
0
d0 =

(csc
2
0 31) d0 = 3cot 0 30 +C
= 3
I
o
2
3&
2
&
3sin
31
&
o
+C
61. For n D 0,
"
0
r
n
dr = lim
I<"
r
n+1
(n + 1)
I
0
= ". For n < 0,
"
0
r
n
dr =

1
0
r
n
dr +
"
1
r
n
dr. Both integrals are
improper. By (7.8.2), the second integral diverges if 31 $ n < 0. By Exercise 7.8.57, the rst integral diverges if n $ 31.
Thus,
"
0
r
n
dr is divergent for all values of n.
63. 1(r) =
1
lnr
, {r =
/ 3o
n
=
4 32
10
=
1
5
(a) T
10
=
1
5 2
{1(2) + 2[1(2.2) +1(2.4) + +1(3.8)] +1(4)} E 1.925444
(b) A
10
=
1
5
[1(2.1) +1(2.3) +1(2.5) + +1(3.9)] E 1.920915
(c) o
10
=
1
5 3
[1(2) + 41(2.2) + 21(2.4) + + 21(3.6) + 41(3.8) +1(4)] E 1.922470
65. 1(r) =
1
lnr
i 1
0
(r) = 3
1
r(lnr)
2
i 1
00
(r) =
2 + lnr
r
2
(lnr)
3
=
2
r
2
(lnr)
3
+
1
r
2
(lnr)
2
. Note that each term of
1
00
(r) decreases on [2. 4], so well take 1 = 1
00
(2) E 2.022. |1
T
| $
1(/ 3o)
3
12n
2
E
2.022(4 32)
3
12(10)
2
= 0.01348 and
|1
L
| $
1(/ 3o)
3
24n
2
= 0.00674. |1
T
| $ 0.00001 C
2.022(8)
12n
2
$
1
10
5
C n
2
D
10
5
(2.022)(8)
12
i n D 367.2.
Take n = 368 for T
n
. |1
L
| $ 0.00001 C n
2
D
10
5
(2.022)(8)
24
i n D 259.6. Take n = 260 for A
n
.
67. {t =

10
60
30
10 =
1
60
.
Distance traveled =

10
0
dt E o
10
=
1
60 3
[40 + 4(42) + 2(45) + 4(49) + 2(52) + 4(54) + 2(56) + 4(57) + 2(57) + 4(55) + 56]
=
1
180
(1544) = 8.57 mi
69. (a) 1(r) = sin(sinr). A CAS gives
1
(4)
(r) = sin(sinr)[cos
4
r + 7 cos
2
r 33]
+ cos(sinr)
6 cos
2
rsinr + sinr
From the graph, we see that
1
(4)
(r)
< 3.8 for r M [0. ].

(b) We use Simpsons Rule with 1(r) = sin(sinr) and {r =
r
10
:
r
0
1(r) dr E
r
10 3
1(0) + 41
r
10
+ 21
2r
10
+ + 41
9r
10
+1()
E 1.786721
From part (a), we know that
1
(4)
(r)
< 3.8 on [0. ], so we use Theorem 7.7.4 with 1 = 3.8, and estimate the error
as |1
S
| $
3.8( 3 0)
5
180(10)
4
E 0.000646.
CHAPTER 7 REVIEW

349
(c) If we want the error to be less than 0.00001, we must have |1
S
| $
3.8
5
180n
4
$ 0.00001,
so n
4
D
3.8
5
180(0.00001)
E 646,041.6 i n D 28.35. Since n must be even for Simpsons Rule, we must have n D 30
to ensure the desired accuracy.
71.
r
3
r
5
+ 2
$
r
3
r
5
=
1
r
2
for r in [1. ").

"
1
1
r
2
dr is convergent by (7.8.2) with j = 2 1. Therefore,
"
1
r
3
r
5
+ 2
dr is
convergent by the Comparison Theorem.
73. For r in
0.
r
2
, 0 $ cos
2
r $ cos r. For r in
r
2
.
, cos r $ 0 $ cos
2
r. Thus,
area =

r'2
0
(cos r 3cos
2
r) dr +
r
r'2
(cos
2
r 3cos r) dr
=

sinr 3
1
2
r 3
1
4
sin2r
r'2
0
+
1
2
r +
1
4
sin2r 3sinr
r
r'2
=

1 3
r
4
30
r
2
3
r
4
31
= 2
75. Using the formula for disks, the volume is
\ =

r'2
0
[1(r)]
2
dr =
r'2
0
(cos
2
r)
2
dr =
r'2
0
1
2
(1 + cos 2r)
2
dr
=
r
4
r'2
0
(1 + cos
2
2r + 2 cos 2r) dr =
r
4
r'2
0
1 +
1
2
(1 + cos 4r) + 2 cos 2r
dr
=
r
4
3
2
r +
1
2
1
4
sin4r
+ 2
1
2
sin2r
r'2
0
=
r
4
3r
4
+
1
8
0 + 0
30
=
3r
2
16
77. By the Fundamental Theorem of Calculus,
"
0
1
0
(r) dr = lim
I<"
I
0
1
0
(r) dr = lim
I<"
[1(t) 31(0)] = lim
I<"
1(t) 31(0) = 0 31(0) = 31(0).
79. Let & = 1r i r = 1& i dr = 3(1&
2
) d&.
"
0
lnr
1 +r
2
dr =

0
"
ln(1&)
1 + 1&
2
3
d&
&
2
=

0
"
3ln&
&
2
+ 1
(3d&) =
0
"
ln&
1 +&
2
d& = 3
"
0
ln&
1 +&
2
d&
Therefore,
"
0
lnr
1 +r
2
dr = 3
"
0
lnr
1 +r
2
dr = 0.
PROBLEMS PLUS
1.
By symmetry, the problem can be reduced to nding the line r = c such that the shaded area is one-third of the area of the
quarter-circle.An equation of the semicircle is j =
I
49 3r
2
, so we require that
c
0
I
49 3r
2
dr =
1
3

1
4
(7)
2
C
1
2
r
I
49 3r
2
+
49
2
sin
31
(r7)
c
0
=
49
12
[by Formula 30] C
1
2
c
I
49 3c
2
+
49
2
sin
31
(c7) =
49
12
.
This equation would be difcult to solve exactly, so we plot the left-hand side as a function of c, and nd that the equation
holds for c E 1.85. So the cuts should be made at distances of about 1.85 inches from the center of the pizza.
3. The given integral represents the difference of the shaded areas, which appears to
be 0. It can be calculated by integrating with respect to either r or j, so we nd r
in terms of j for each curve: j =
3
I
1 3r
7
i r =
7
1 3j
3
and
j =
7
I
1 3r
3
i r =
3
1 3j
7
, so
1
0
1 3j
7
3
7
1 3j
3
dj =

1
0
7
I
1 3r
3
3
3
I
1 3r
7
dr. But this

equation is of the form : = 3:. So
1
0
3
I
1 3r
7
3
7
I
1 3r
3
dr = 0.
5. The area of the remaining part of the circle is given by
= 41 = 4
a
0
o
2
3r
2
3
/
o
o
2
3r
2
dr = 4
1 3
/
o
a
0
o
2
3r
2
dr
30
=
4
o
(o 3/)
r
2
I
o
2
3r
2
+
o
2
2
sin
31
r
o
a
0
=
4
o
(o 3/)
0 +
o
2
2

2
30
=
4
o
(o 3/)
o
2
= o(o 3/),
which is the area of an ellipse with semiaxes o and o 3/.
Alternate solution: Subtracting the area of the ellipse from the area of the circle gives us o
2
3o/ = o (o 3/),
as calculated above. (The formula for the area of an ellipse was derived in Example 2 in Section 7.3.)
351
352

7. Recall that cos cos 1 =
1
2
[cos(+1) + cos(31)]. So
1(r) =

r
0
cos t cos(r 3t) dt =
1
2
r
0
[cos(t +r 3t) + cos(t 3r +t)] dt =
1
2
r
0
[cos r + cos(2t 3r)] dt
=
1
2
t cos r +
1
2
sin(2t 3r)
r
0
=
r
2
cos r +
1
4
sin(2 3r) 3
1
4
sin(3r)
=
r
2
cos r +
1
4
sin(3r) 3
1
4
sin(3r) =
r
2
cos r
The minimum of cos r on this domain is 31, so the minimum value of 1(r) is 1() = 3
r
2
.
9. In accordance with the hint, we let 1
I
=

1(13i
2
)I
0
dr, and we nd an expression for 1
I+1
in terms of 1
I
. We integrate 1
I+1
by parts with & = (1 3r
2
)
I+1
i d& = (I + 1)(1 3r
2
)
I
(32r), d = dr i = r, and then split the remaining
integral into identiable quantities:
1
I+1
= r(1 3r
2
)
I+1
1
0
+ 2(I + 1)
1
0
r
2
(1 3r
2
)
I
dr = (2I + 2)
1
0
(1 3r
2
)
I
[1 3(1 3r
2
)] dr
= (2I + 2)(1
I
31
I+1
)
So 1
I+1
[1 + (2I + 2)] = (2I + 2)1
I
i 1
I+1
=
2I + 2
2I + 3
1
I
. Now to complete the proof, we use induction:
1
0
= 1 =
2
0
(0!)
2
1!
, so the formula holds for n = 0. Now suppose it holds for n = I. Then
1
I+1
=
2I + 2
2I + 3
1
I
=
2I + 2
2I + 3
2
2I
(I!)
2
(2I + 1)!
=
2(I + 1)2
2I
(I!)
2
(2I + 3)(2I + 1)!
=
2(I + 1)
2I + 2

2(I + 1)2
2I
(I!)
2
(2I + 3)(2I + 1)!
=
[2(I + 1)]
2
2
2I
(I!)
2
(2I + 3)(2I + 2)(2I + 1)!
=
2
2(I+1)
[(I + 1)!]
2
[2(I + 1) + 1]!
So by induction, the formula holds for all integers n D 0.
11. 0 < o < /. Now
1
0
[/r +o(1 3r)]
I
dr =

l
a
&
I
(/ 3o)
d& [u = br + o(1 r)] =

&
I+1
(t + 1)(/ 3o)
l
a
=
/
I+1
3o
I+1
(t + 1)(/ 3o)
.
Now let j = lim
I<0
/
I+1
3o
I+1
(t + 1)(/ 3o)
1'I
. Then lnj = lim
I<0
1
t
ln
/
I+1
3o
I+1
(t + 1)(/ 3o)
. This limit is of the form 00,

so we can apply lHospitals Rule to get
lnj = lim
I<0
/
I+1
ln/ 3o
I+1
lno
/
I+1
3o
I+1
3
1
t + 1
=
/ ln/ 3o lno
/ 3o
31 =
/ ln/
/ 3o
3
olno
/ 3o
3lnc = ln
/
l'(l3a)
co
a'(l3a)
.
Therefore, j = c
31
/
l
o
a
1'(l3a)
.

353
13. An equation of the circle with center (0. c) and radius 1 is r
2
+ (j 3c)
2
= 1
2
, so
an equation of the lower semicircle is j = c 3
I
1 3r
2
. At the points of tangency,
the slopes of the line and semicircle must be equal. For r D 0, we must have
j
0
= 2 i
r
I
1 3r
2
= 2 i r = 2
I
1 3r
2
i r
2
= 4(1 3r
2
) i
5r
2
= 4 i r
2
=
4
5
i r =
2
5
I
5 and so j = 2
2
5
I
5
=
4
5
I
5.
The slope of the perpendicular line segment is 3
1
2
, so an equation of the line segment is j 3
4
5
I
5 = 3
1
2
r 3
2
5
I
5
C
j = 3
1
2
r +
1
5
I
5 +
4
5
I
5 C j = 3
1
2
r +
I
5, so c =
I
5 and an equation of the lower semicircle is j =
I
5 3
I
1 3r
2
.
Thus, the shaded area is
2
(2'5)
I
5
0
I
5 3
1 3r
2
32r
dr
30
= 2
I
5 r 3
r
2
I
1 3r
2
3
1
2
sin
31
r 3r
2
(2'5)
I
5
0
= 2
2 3
I
5
5

1
I
5
3
1
2
sin
31
2
I
5
3
4
5
32(0)
= 2
1 3
1
2
sin
31
2
I
5
= 2 3sin
31
2
I
5
15. We integrate by parts with & =

1
ln(1 +r +t)
, d = sint dt, so d& =
31
(1 +r +t)[ln(1 +r +t)]
2
and = 3cos t. The
integral becomes
1 =

"
0
sint dt
ln(1 +r +t)
= lim
l<"
3cos t
ln(1 +r +t)
l
0
3
l
0
cos t dt
(1 +r +t)[ln(1 +r +t)]
2
= lim
l<"
3cos /
ln(1 +r +/)
+
1
ln(1 +r)
+
"
0
3cos t dt
(1 +r +t)[ln(1 +r +t)]
2
=
1
ln(1 +r)
+J
where J =

"
0
3cos t dt
(1 +r +t)[ln(1 +r +t)]
2
. Now 31 $ 3cos t $ 1 for all t; in fact, the inequality is strict except
at isolated points. So 3
"
0
dt
(1 +r +t)[ln(1 +r +t)]
2
< J <

"
0
dt
(1 +r +t)[ln(1 +r +t)]
2
C
3
1
ln(1 +r)
< J <
1
ln(1 +r)
C 0 < 1 <
2
ln(1 +r)
.
8 FURTHER APPLICATIONS OF INTEGRATION
8.1 Arc Length
1. j = 2r 35 i 1 =

3
31
1 + (djdr)
2
dr =

3
31
1 + (2)
2
dr =
I
5 [3 3(31)] = 4
I
5.
The arc length can be calculated using the distance formula, since the curve is a line segment, so
1 = [distance from (31. 37) to (3. 1)] =

[3 3(31)]
2
+ [1 3(37)]
2
=
I
80 = 4
I
5
3. j = cos r i djdr = 3sinr i 1 + (djdr)
2
= 1 + sin
2
r. So 1 =

2r
0
1 + sin
2
rdr.
5. r = j +j
3
i drdj = 1 + 3j
2
i 1 + (drdj)
2
= 1 + (1 + 3j
2
)
2
= 9j
4
+ 6j
2
+ 2.
So 1 =

4
1
9j
4
+ 6j
2
+ 2 dj.
7. j = 1 + 6r
3'2
i djdr = 9r
1'2
i 1 + (djdr)
2
= 1 + 81r. So
1 =

1
0
I
1 + 81rdr =

82
1
&
1'2
1
81
d&

u = 1 + 81r,
Ju = 81 Jr
=
1
81

2
3
&
3'2
82
1
=
2
243
82
I
82 31
9. j =
r
5
6
+
1
10r
3
i
dj
dr
=
5
6
r
4
3
3
10
r
34
i
1 + (djdr)
2
= 1 +
25
36
r
8
3
1
2
+
9
100
r
38
=
25
36
r
8
+
1
2
+
9
100
r
38
=

5
6
r
4
+
3
10
r
34
2
. So
1 =

2
1
5
6
r
4
+
3
10
r
34
2
dr =

2
1
5
6
r
4
+
3
10
r
34
dr =

1
6
r
5
3
1
10
r
33
2
1
=

32
6
3
1
80
1
6
3
1
10
=
31
6
+
7
80
=
1261
240
11. r =
1
3
I
j (j 33) =
1
3
j
3'2
3 j
1'2
i drdj =
1
2
j
1'2
3
1
2
j
31'2
i
1 + (drdj)
2
= 1 +
1
4
j 3
1
2
+
1
4
j
31
=
1
4
j +
1
2
+
1
4
j
31
=

1
2
j
1'2
+
1
2
j
31'2
2
. So
1 =

9
1
1
2
j
1'2
+
1
2
j
31'2
dj =
1
2
2
3
j
3'2
+ 2j
1'2
9
1
=
1
2
2
3
27 + 2 3
2
3
1 + 2 1
=
1
2
24 3
8
3
=
1
2
64
3
=
32
3
.
13. j = ln(sec r) i
dj
dr
=
sec r tanr
sec r
= tanr i 1 +
dj
dr
2
= 1 + tan
2
r = sec
2
r, so
1 =

r'4
0
I
sec
2
rdr =

r'4
0
|sec r| dr =

r'4
0
sec rdr =

ln(sec r + tanr)
r'4
0
= ln
I
2 + 1
3ln(1 + 0) = ln
I
2 + 1
355
356

CHAPTER 8 FURTHER APPLICATIONS OF INTEGRATION
15. j = ln(1 3 r
2
) i j
0
=
1
1 3r
2
(32r) i
1 +
dj
dr
2
= 1 +
4r
2
(1 3r
2
)
2
=
1 32r
2
+r
4
+ 4r
2
(1 3r
2
)
2
=
1 + 2r
2
+r
4
(1 3r
2
)
2
=
(1 +r
2
)
2
(1 3r
2
)
2
i
1 +
dj
dr
2
=
1 +r
2
1 3r
2
2
=
1 +r
2
1 3r
2
= 31 +
2
1 3r
2
[by division] = 31 +
1
1 +r
+
1
1 3r
[partial fractions].
So 1 =

1'2
0
31 +
1
1 +r
+
1
1 3r
dr =

3r + ln|1 +r| 3ln|1 3r|
1'2
0
=

3
1
2
+ ln
3
2
3ln
1
2
30 = ln3 3
1
2
.
17. j = c
i
i j
0
= c
i
i 1 + (j
0
)
2
= 1 +c
2i
. So
1 =

1
0
1 +c
2i
dr =

c
1
1 +&
2
d&
&
u = t
{
, so
r = ln u, Jr = Ju/u
=

c
1
I
1 +&
2
&
2
&d&
=

I
1+c
2
I
2
2
31
d

u =
s
1 + u
2
, so
u
2
= 1 + u
2
, u Ju = uJu
=

I
1+c
2
I
2
1 +
12
31
3
12
+ 1
d
=

+
1
2
ln
31
+ 1
I
1+c
2
I
2
=
I
1 +c
2
+
1
2
ln
I
1 +c
2
31
I
1 +c
2
+ 1
3
I
2 3
1
2
ln
I
2 31
I
2 + 1
=
I
1 +c
2
3
I
2 + ln
I
1 +c
2
31
31 3ln
I
2 31
Or: Use Formula 23 for

I
1 +&
2
&
d&, or substitute & = tan0.

19. j =
1
2
r
2
i djdr = r i 1 + (djdr)
2
= 1 +r
2
. So
1=

1
31
I
1 +r
2
dr = 2
1
0
I
1 +r
2
dr [by symmetry]
21
= 2
i
2
I
1 +r
2
+
1
2
ln
r +
I
1 +r
2
1
0
or substitute
r = tan 0
= 2
1
2
I
2 +
1
2
ln
1 +
I
2
0 +
1
2
ln1
=
I
2 + ln
1 +
I
2
21. From the gure, the length of the curve is slightly larger than the hypotenuse
of the triangle formed by the points (1. 0), (3. 0), and (3. 1(3)) E (3. 15),
where j = 1(r) =
2
3
(r
2
31)
3'2
. This length is about
I
15
2
+ 2
2
E 15, so
we might estimate the length to be 15.5.
j =
2
3
(r
2
31)
3'2
i j
0
= (r
2
31)
1'2
(2r) i 1 + (j
0
)
2
= 1 + 4r
2
(r
2
31) = 4r
4
34r
2
+ 1 = (2r
2
31)
2
,
so, using the fact that 2r
2
3 1 0 for 1 $ r $ 3,
1 =

3
1
(2r
2
31)
2
dr =

3
1
2r
2
31
dr =

3
1
(2r
2
31) dr =

2
3
r
3
3r
3
1
= (18 33) 3
2
3
31
=
46
3
= 15.3.
23. j = rc
3i
i djdr = c
3i
3rc
3i
= c
3i
(1 3r) i 1 + (djdr)
2
= 1 +c
32i
(1 3r)
2
. Let
1(r) =
1 + (djdr)
2
=

1 +c
32i
(1 3r)
2
. Then 1 =

5
0
1(r) dr. Since n = 10, {r =
5 30
10
=
1
2
. Now
1 E o
10
=
1'2
3
[1(0) + 41
1
2
+ 21(1) + 41
3
2
+ 21(2) + 41
5
2
+ 21(3) + 41
7
2
+ 21(4) + 41
9
2
+1(5)]
E 5.115840
The value of the integral produced by a calculator is 5.113568 (to six decimal places).
SECTION 8.1 ARC LENGTH

357
25. j = sec r i djdr = sec r tanr i 1 =

r'3
0
1(r) dr, where 1(r) =
I
1 + sec
2
r tan
2
r.
Since n = 10, {r =
3 30
10
=

30
. Now
1 E o
10
=
30
3
1(0) + 41

30
+ 21
2
30
+ 41
3
30
+ 21
4
30
+ 41
5
30
+ 21
6
30
+ 41
7
30
+ 21
8
30
+ 41
9
30
+1
E 1.569619.
27. (a)
(b) Let 1(r) = j = r
3
I
4 3r. The polygon with one side is just
the line segment joining the points (0. 1(0)) = (0. 0) and
(4. 1(4)) = (4. 0), and its length 1
1
= 4.
The polygon with two sides joins the points (0. 0),
(2. 1(2)) =
2. 2
3
I
2
and (4. 0). Its length

1
2
=

(2 30)
2
+
2
3
I
2 30
2
+
(4 32)
2
+
0 32
3
I
2
2
= 2
I
4 + 2
8'3
E 6.43
Similarly, the inscribed polygon with four sides joins the points (0. 0),
1.
3
I
3
2. 2
3
I
2
, (3. 3), and (4. 0),

so its length
1
3
=

1 +
3
I
3
2
+
1 +
2
3
I
2 3
3
I
3
2
+
1 +
3 32
3
I
2
2
+
I
1 + 9 E 7.50
(c) Using the arc length formula with
dj
dr
= r
1
3
(4 3r)
32'3
(31)
+
3
I
4 3r =
12 34r
3(4 3r)
2'3
, the length of the curve is
1 =

4
0
1 +
dj
dr
2
dr =

4
0
1 +
12 34r
3(4 3r)
2'3
2
dr.
(d) According to a CAS, the length of the curve is 1 E 7.7988. The actual value is larger than any of the approximations in
part (b). This is always true, since any approximating straight line between two points on the curve is shorter than the
length of the curve between the two points.
358

29. j = lnr i djdr = 1r i 1 + (djdr)
2
= 1 + 1r
2
= (r
2
+ 1)r
2
i
1 =

2
1
r
2
+ 1
r
2
dr =

2
1
I
1 +r
2
r
dr
23
=

1 +r
2
3ln
1 +
I
1 +r
2
r
2
1
=
I
5 3ln
1 +
I
5
2
3
I
2 + ln
1 +
I
2
31. j
2'3
= 1 3r
2'3
i j = (1 3r
2'3
)
3'2
i
dj
dr
=
3
2
(1 3r
2'3
)
1'2
3
2
3
r
31'3
= 3r
31'3
(1 3r
2'3
)
1'2
i
dj
dr
2
= r
32'3
(1 3r
2'3
) = r
32'3
31. Thus
1 = 4
1
0
1 + (r
32'3
31) dr = 4
1
0
r
31'3
dr = 4 lim
I<0
+
3
2
r
2'3
1
I
= 6.
33. j = 2r
3'2
i j
0
= 3r
1'2
i 1 + (j
0
)
2
= 1 + 9r. The arc length function with starting point 1
0
(1. 2) is
c(r) =
i
1
I
1 + 9t dt =

2
27
(1 + 9t)
3'2
i
1
=
2
27
(1 + 9r)
3'2
310
I
10
.
35. j = sin
31
r +
I
1 3r
2
i j
0
=
1
I
1 3r
2
3
r
I
1 3r
2
=
1 3r
I
1 3r
2
i
1 + (j
0
)
2
= 1 +
(1 3r)
2
1 3r
2
=
1 3r
2
+ 1 32r +r
2
1 3r
2
=
2 32r
1 3r
2
=
2(1 3r)
(1 +r)(1 3r)
=
2
1 +r
i
1 + (j
0
)
2
=
2
1 +r
. Thus, the arc length function with starting point (0. 1) is given by
c(r) =
i
0
1 + [1
0
(t)]
2
dt =

i
0
2
1 +t
dt =
I
2
2
I
1 +t
i
0
= 2
I
2
I
1 +r 31
.
37. The prey hits the ground when j = 0 C 180 3
1
45
r
2
= 0 C r
2
= 45 180 i r =
I
8100 = 90,
since r must be positive. j
0
= 3
2
45
r i 1 + (j
0
)
2
= 1 +
4
45
2
r
2
, so the distance traveled by the prey is
1 =

90
0
1 +
4
45
2
r
2
dr =

4
0
1 +&
2
45
2
d&

u =
2
45
r,
Ju =
2
45
Jr
21
=
45
2
1
2
&
I
1 +&
2
+
1
2
ln
& +
I
1 +&
2
4
0
=
45
2
2
I
17 +
1
2
ln
4 +
I
17
= 45
I
17 +
45
4
ln
4 +
I
17
E 209.1 m
39. The sine wave has amplitude 1 and period 14, since it goes through two periods in a distance of 28 in., so its equation is
j = 1 sin
2r
14
r
= sin
r
7
r
. The width n of the at metal sheet needed to make the panel is the arc length of the sine curve
from r = 0 to r = 28. We set up the integral to evaluate n using the arc length formula with
uu
ui
=
r
7
cos
r
7
r
:
1 =

28
0
1 +
r
7
cos
r
7
r
2
dr = 2
14
0
1 +
r
7
cos
r
7
r
2
dr. This integral would be very difcult to evaluate exactly,
so we use a CAS, and nd that 1 E 29.36 inches.
41. j =

i
1
I
t
3
31 dt i djdr =
I
r
3
31 [by FTC1] i 1 + (djdr)
2
= 1 +
I
r
3
31
2
= r
3
i
1 =

4
1
I
r
3
dr =

4
1
r
3'2
dr =
2
5
r
5'2
4
1
=
2
5
(32 31) =
62
5
= 12.4
SECTION 8.2 AREA OF A SURFACE OF REVOLUTION

359
8.2 Area of a Surface of Revolution
1. j = r
4
i djdr = 4r
3
i dc =

1 + (djdr)
2
dr =
I
1 + 16r
6
dr
(a) By (7), an integral for the area of the surface obtained by rotating the curve about the r-axis is
o =

2j dc =

1
0
2r
4
I
1 + 16r
6
dr.
(b) By (8), an integral for the area of the surface obtained by rotating the curve about the j-axis is
o =

2rdc =

1
0
2r
I
1 + 16r
6
dr.
3. j = tan
31
r i
dj
dr
=
1
1 +r
2
i dc =
1 +
dj
dr
2
dr =

1 +
1
(1 +r
2
)
2
dr.
(a) By (7), o =

2j dc =

1
0
2 tan
31
r
1 +
1
(1 +r
2
)
2
dr.
(b) By (8), o =

2rdc =

1
0
2r
1 +
1
(1 +r
2
)
2
dr.
5. j = r
3
i j
0
= 3r
2
. So
o =

2
0
2j
1 + (j
0
)
2
dr = 2
2
0
r
3
I
1 + 9r
4
dr [u = 1 + 9r
4
, Ju = 36r
3
Jr]
=
2r
36
145
1
I
&d& =
r
18
2
3
&
3'2
145
1
=
r
27
145
I
145 31
7. j =
I
1 + 4r i j
0
=
1
2
(1 + 4r)
31'2
(4) =
2
I
1 + 4r
i

1 + (j
0
)
2
=
1 +
4
1 + 4r
=
5 + 4r
1 + 4r
. So
o =

5
1
2j
1 + (j
0
)
2
dr = 2
5
1
I
1 + 4r
5 + 4r
1 + 4r
dr = 2
5
1
I
4r + 5 dr
= 2
25
9
I
&
1
4
d&

u = 4r + 5,
Ju = 4 Jr
=
2r
4
2
3
&
3'2
25
9
=
r
3
(25
3'2
39
3'2
) =
r
3
(125 327) =
98
3

9. j = sinr i j
0
= cos r i 1 + (j
0
)
2
= 1 +
2
cos
2
(r). So
o =

1
0
2j
1 + (j
0
)
2
dr = 2
1
0
sinr
1 +
2
cos
2
(r) dr

u = t cos tr,
Ju = t
2
sin trJr
= 2
3r
r
1 +&
2
3
1
2
d&
=
2
r
3r
1 +&
2
d&
=
4
r
0
1 +&
2
d&
21
=
4
&
2
1 +&
2
+
1
2
ln
& +
1 +&
2
r
0
=
4
2
I
1 +
2
+
1
2
ln
+
I
1 +
2
30
= 2
I
1 +
2
+
2
ln
+
I
1 +
2
11. r =
1
3
(j
2
+ 2)
3'2
i drdj =
1
2
(j
2
+ 2)
1'2
(2j) = j
j
2
+ 2 i 1 + (drdj)
2
= 1 +j
2
(j
2
+ 2) = (j
2
+ 1)
2
.
So o = 2
2
1
j(j
2
+ 1) dj = 2
1
4
j
4
+
1
2
j
2
2
1
= 2
4 + 2 3
1
4
3
1
2
=
21r
2
.
13. j =
3
I
r i r = j
3
i 1 + (drdj)
2
= 1 + 9j
4
. So
o = 2
2
1
r
1 + (drdj)
2
dj = 2
2
1
j
3
1 + 9j
4
dj =
2r
36
2
1
1 + 9j
4
36j
3
dj =
r
18
2
3
1 + 9j
4
3'2
2
1
=
r
27
145
I
145 310
I
10
360

15. r =

o
2
3j
2
i drdj =
1
2
(o
2
3 j
2
)
31'2
(32j) = 3j
o
2
3j
2
i
1 +
dr
dj
2
= 1 +
j
2
o
2
3j
2
=
o
2
3j
2
o
2
3j
2
+
j
2
o
2
3j
2
=
o
2
o
2
3j
2
i
o =

a'2
0
2
o
2
3j
2
o
o
2
3j
2
dj = 2
a'2
0
odj = 2o
a'2
0
= 2o
o
2
30
= o
2
.
Note that this is
1
4
the surface area of a sphere of radius o, and the length of the interval j = 0 to j = o2 is
1
4
the length of the
interval j = 3o to j = o.
17. j = lnr i djdr = 1r i 1 + (djdr)
2
= 1 + 1r
2
i o =

3
1
2 lnr
1 + 1r
2
dr.
Let 1(r) = lnr
1 + 1r
2
. Since n = 10, {r =
3 31
10
=
1
5
. Then
o E o
10
= 2
1'5
3
[1(1) + 41(1.2) + 21(1.4) + + 21(2.6) + 41(2.8) +1(3)] E 9.023754.
19. j = sec r i djdr = sec r tanr i 1 + (djdr)
2
= 1 + sec
2
r tan
2
r i
o =

r'3
0
2 sec r
I
1 + sec
2
r tan
2
rdr. Let 1(r) = sec r
I
1 + sec
2
r tan
2
r. Since n = 10, {r =
3 30
10
=

30
.
Then o E o
10
= 2
30
3
1(0) + 41

30
+ 21
2
30
+ + 21
8
30
+ 41
9
30
+1
E 13.527296.
21. j = 1r i dc =

1 + (djdr)
2
dr =

1 + (31r
2
)
2
dr =

1 + 1r
4
dr i
o =

2
1
2
1
r
1 +
1
r
4
dr = 2
2
1
I
r
4
+ 1
r
3
dr = 2
4
1
I
&
2
+ 1
&
2
1
2
d&
[u = r
2
, Ju = 2rJr]
=
4
1
I
1 +&
2
&
2
d&
24
=
3
I
1 +&
2
&
+ ln
& +
1 +&
2
4
1
=
3
I
17
4
+ ln
4 +
I
17
+
I
2
1
3ln
1 +
I
2
=

4
4 ln
I
17 + 4
34 ln
I
2 + 1
3
I
17 + 4
I
2
23. j = r
3
and 0 $ j $ 1 i j
0
= 3r
2
and 0 $ r $ 1.
o =

1
0
2r
1 + (3r
2
)
2
dr = 2
3
0
I
1 +&
2
1
6
d&

u = 3r
2
,
Ju = 6rJr
=
r
3
3
0
I
1 +&
2
d&
21
= [or use CAS]
r
3
1
2
&
I
1 +&
2
+
1
2
ln
& +
I
1 +&
2
3
0
=
r
3
3
2
I
10 +
1
2
ln
3 +
I
10
=
r
6
3
I
10 + ln
3 +
I
10
25. o = 2
"
1
j
1 +
dj
dr
2
dr = 2
"
1
1
r
1 +
1
r
4
dr = 2
"
1
I
r
4
+ 1
r
3
dr. Rather than trying to
evaluate this integral, note that
I
r
4
+ 1
I
r
4
= r
2
for r 0. Thus, if the area is nite,
o = 2
"
1
I
r
4
+ 1
r
3
dr 2
"
1
r
2
r
3
dr = 2
"
1
1
r
dr. But we know that this integral diverges, so the area o is
innite.
SECTION 8.2 AREA OF A SURFACE OF REVOLUTION

361
27. Since o 0, the curve 3oj
2
= r(o 3r)
2
only has points with r D 0.
[3oj
2
D 0 i r(o 3r)
2
D 0 i r D 0.]
The curve is symmetric about the x-axis (since the equation is unchanged
when j is replaced by 3j). j = 0 when r = 0 or o, so the curves loop
extends from r = 0 to r = o.
d
dr
(3oj
2
) =
d
dr
[r(o 3r)
2
] i 6oj
dj
dr
= r 2(o 3r)(31) + (o 3r)
2
i
dj
dr
=
(o 3r)[32r +o 3r]
6oj
i
dj
dr
2
=
(o 3r)
2
(o 33r)
2
36o
2
j
2
=
(o 3r)
2
(o 33r)
2
36o
2

3o
r(o 3r)
2
the last fraction

is 1/u
2
=
(o 33r)
2
12or
i
1 +
dj
dr
2
= 1 +
o
2
36or + 9r
2
12or
=
12or
12or
+
o
2
36or + 9r
2
12or
=
o
2
+ 6or + 9r
2
12or
=
(o + 3r)
2
12or
for r 6= 0.
(a) o =

a
i=0
2j dc = 2
a
0
I
r(o 3r)
I
3o

o + 3r
I
12or
dr = 2
a
0
(o 3r)(o + 3r)
6o
dr
=

3o
a
0
(o
2
+ 2or 33r
2
) dr =

3o
o
2
r +or
2
3r
3
a
0
=

3o
(o
3
+o
3
3o
3
) =

3o
o
3
=
o
2
3
.
Note that we have rotated the top half of the loop about the x-axis. This generates the full surface.
(b) We must rotate the full loop about the y-axis, so we get double the area obtained by rotating the top half of the loop:
o = 2 2
a
i=0
rdc = 4
a
0
r
o + 3r
I
12or
dr =
4
2
I
3o
a
0
r
1'2
(o + 3r) dr =
2
I
3o
a
0
(or
1'2
+ 3r
3'2
) dr
=
2
I
3o
2
3
or
3'2
+
6
5
r
5'2
a
0
=
2
I
3
3
I
o
2
3
o
5'2
+
6
5
o
5'2
=
2
I
3
3
2
3
+
6
5
o
2
=
2
I
3
3
28
15
o
2
=
56
I
3 o
2
45
29. (a)
r
2
o
2
+
j
2
/
2
= 1 i
j (djdr)
/
2
= 3
r
o
2
i
dj
dr
= 3
/
2
r
o
2
j
i
1 +
dj
dr
2
= 1 +
/
4
r
2
o
4
j
2
=
/
4
r
2
+o
4
j
2
o
4
j
2
=
/
4
r
2
+o
4
/
2
1 3r
2
o
2
o
4
/
2
(1 3r
2
o
2
)
=
o
4
/
2
+/
4
r
2
3o
2
/
2
r
2
o
4
/
2
3o
2
/
2
r
2
=
o
4
+/
2
r
2
3o
2
r
2
o
4
3o
2
r
2
=
o
4
3
o
2
3/
2
r
2
o
2
(o
2
3r
2
)
The ellipsoids surface area is twice the area generated by rotating the rst-quadrant portion of the ellipse about the r-axis.
Thus,
o = 2
a
0
2j
1 +
dj
dr
2
dr = 4
a
0
/
o
o
2
3r
2
o
4
3(o
2
3/
2
)r
2
o
I
o
2
3r
2
dr =
4/
o
2
a
0
o
4
3(o
2
3/
2
)r
2
dr
=
4/
o
2
a
I
a
2
3l
2
0
o
4
3&
2
d&
I
o
2
3/
2
u =
s
o
2
b
2
r
30
=
4/
o
2
I
o
2
3/
2
&
2
o
4
3&
2
+
o
4
2
sin
31
&
o
2
a
I
a
2
3l
2
0
=
4/
o
2
I
o
2
3/
2
o
I
o
2
3/
2
2
o
4
3o
2
(o
2
3/
2
) +
o
4
2
sin
31
I
o
2
3/
2
o
= 2
/
2
+
o
2
/ sin
31
I
o
2
3/
2
o
I
o
2
3/
2
362

(b)
r
2
o
2
+
j
2
/
2
= 1 i
r(drdj)
o
2
= 3
j
/
2
i
dr
dj
= 3
o
2
j
/
2
r
i
1 +
dr
dj
2
= 1 +
o
4
j
2
/
4
r
2
=
/
4
r
2
+o
4
j
2
/
4
r
2
=
/
4
o
2
(1 3j
2
/
2
) +o
4
j
2
/
4
o
2
(1 3j
2
/
2
)
=
o
2
/
4
3o
2
/
2
j
2
+o
4
j
2
o
2
/
4
3o
2
/
2
j
2
=
/
4
3/
2
j
2
+o
2
j
2
/
4
3/
2
j
2
=
/
4
3(/
2
3o
2
)j
2
/
2
(/
2
3j
2
)
The oblate spheroids surface area is twice the area generated by rotating the rst-quadrant portion of the ellipse about the
j-axis. Thus,
o = 2
l
0
2 r
1 +
dr
dj
2
dj = 4
l
0
o
/
/
2
3j
2
/
4
3(/
2
3o
2
)j
2
/
/
2
3j
2
dj
=
4o
/
2
l
0
/
4
3(/
2
3o
2
)j
2
dj =
4o
/
2
l
I
l
2
3a
2
0
/
4
3&
2
d&
I
/
2
3o
2
u =
s
b
2
o
2
u
30
=
4o
/
2
I
/
2
3o
2
&
2
I
/
4
3&
2
+
/
4
2
sin
31
&
/
2
l
I
l
2
3a
2
0
=
4o
/
2
I
/
2
3o
2
/
I
/
2
3o
2
2
/
4
3/
2
(/
2
3o
2
) +
/
4
2
sin
31
I
/
2
3o
2
/
= 2
o
2
+
o/
2
sin
31
I
/
2
3o
2
/
I
/
2
3o
2
Notice that this result can be obtained from the answer in part (a) by interchanging o and /.
31. The analogue of 1(r
W
.
) in the derivation of (4) is now c 3 1(r
W
.
), so
o = lim
n<"
n
.=1
2[c 31(r
W
.
)]
1 + [1
0
(r
W
.
)]
2
{r =

l
a
2[c 31(r)]
1 + [1
0
(r)]
2
dr.
33. For the upper semicircle, 1(r) =
I
v
2
3r
2
, 1
0
(r) = 3r
I
v
2
3r
2
. The surface area generated is
o
1
=

r
3r
2
v 3
v
2
3r
2
1 +
r
2
v
2
3r
2
dr = 4
r
0
v 3
v
2
3r
2
v
I
v
2
3r
2
dr
= 4
r
0
v
2
I
v
2
3r
2
3v
dr
For the lower semicircle, 1(r) = 3
I
v
2
3r
2
and 1
0
(r) =
r
I
v
2
3r
2
, so o
2
= 4
r
0
v
2
I
v
2
3r
2
+v
dr.
Thus, the total area is o = o
1
+o
2
= 8
r
0
v
2
I
v
2
3r
2
dr = 8
v
2
sin
31
r
v
r
0
= 8v
2
= 4
2
v
2
.
35. In the derivation of (4), we computed a typical contribution to the surface area to be 2
j
.31
+j
.
2
|1
.31
1
.
|,
the area of a frustum of a cone. When 1(r) is not necessarily positive, the approximations j
.
= 1(r
.
) E 1(r
W
.
) and
j
.31
= 1(r
.31
) E 1(r
W
.
) must be replaced by j
.
= |1(r
.
)| E |1(r
W
.
)| and j
.31
= |1(r
.31
)| E |1(r
W
.
)|. Thus,
2
j
.31
+j
.
2
|1
.31
1
.
| E 2 |1(r
W
.
)|
1 + [1
0
(r
W
.
)]
2
{r. Continuing with the rest of the derivation as before,
we obtain o =

l
a
2 |1(r)|
1 + [1
0
(r)]
2
dr.
SECTION 8.3 APPLICATIONS TO PHYSICS AND ENGINEERING

363
8.3 Applications to Physics and Engineering
1. The weight density of water is c = 62.5 lbft
3
.
(a) 1 = cd E (62.5 lbft
3
)(3 ft) = 187.5 lbft
2
(b) 1 = 1 E (187.5 lbft
2
)(5 ft)(2 ft) = 1875 lb. ( is the area of the bottom of the tank.)
(c) As in Example 1, the area of the jth strip is 2 ({r) and the pressure is cd = cr
.
. Thus,
1 =

3
0
cr 2 dr E (62.5)(2)
3
0
rdr = 125
1
2
r
2
3
0
= 125
9
2
= 562.5 lb.
In Exercises 39, n is the number of subintervals of length {r and r
W
.
is a sample point in the jth subinterval [r
.31
. r
.
].
3. Set up a vertical x-axis as shown, with r = 0 at the waters surface and r increasing in the
downward direction. Then the area of the jth rectangular strip is 6 {r and the pressure on
the strip is cr
W
.
(where c E 62.5 lbft
3
). Thus, the hydrostatic force on the strip is
cr
W
.
6 {r and the total hydrostatic force E
n
.=1
cr
W
.
6 {r. The total force
1 = lim
n<"
n
.=1
cr
W
.
6 {r =

6
2
cr 6 dr = 6c
6
2
rdr = 6c
1
2
r
2
6
2
= 6c(18 32) = 96c E 6000 lb
5. Set up a vertical x-axis as shown. The base of the triangle shown in the gure
has length
3
2
3(r
W
.
)
2
, so n
.
= 2
9 3(r
W
.
)
2
, and the area of the jth
rectangular strip is 2
9 3(r
W
.
)
2
{r. The jth rectangular strip is (r
W
.
31) m
below the surface level of the water, so the pressure on the strip is ao(r
W
.
31).
The hydrostatic force on the strip is ao(r
W
.
31) 2
9 3(r
W
.
)
2
{r and the total
force on the plate E
n
.=1
ao(r
W
.
31) 2
9 3(r
W
.
)
2
{r. The total force
1 = lim
n
.=1
ao(r
W
.
31) 2
9 3(r
W
.
)
2
{r = 2ao
3
1
(r 31)
I
9 3r
2
dr
= 2ao
3
1
r
I
9 3r
2
dr 32ao
3
1
I
9 3r
2
dr
30
= 2ao
3
1
3
(9 3r
2
)
3'2
3
1
32ao
r
2
I
9 3r
2
+
9
2
sin
31
r
3
3
1
= 2ao
0 +
1
3
8
I
8
32ao
0 +
9
2

r
2
1
2
I
8 +
9
2
sin
31
1
3
=
32
3
I
2 ao 3
9r
2
ao + 2
I
2 ao + 9
sin
31
1
3
ao =

38
3
I
2 3
9r
2
+ 9 sin
31
1
3
ao
E 6.835 1000 9.8 E 6.7 10
4
N
Note: If you set up a typical coordinate system with the water level at j = 31, then 1 =

31
33
ao(31 3j)2
9 3j
2
dj.
364

7. Set up a vertical x-axis as shown. Then the area of the jth rectangular strip is
2 3
2
I
3
r
W
.
{r.

n
.
2
=
I
3 3r
W
.
I
3
, so n
.
= 2 3
2
I
3
r
W
.
.
The pressure on the strip is aor

W
.
, so the hydrostatic force on the strip is
aor
W
.
2 3
2
I
3
r
W
.
{r and the hydrostatic force on the plate E

n
.=1
aor
W
.
2 3
2
I
3
r
W
.
{r.
The total force
1 = lim
n<"
n
.=1
aor
W
.
2 3
2
I
3
r
W
.
{r =

I
3
0
aor
2 3
2
I
3
r
dr = ao
I
3
0
2r 3
2
I
3
r
2
dr
= ao
r
2
3
2
3
I
3
r
3
I
3
0
= ao [(3 32) 30] = ao E 1000 9.8 = 9.8 10
3
N
9. Set up coordinate axes as shown in the gure. The length of the jth strip is
2
25 3(j
W
.
)
2
and its area is 2
25 3(j
W
.
)
2
{j. The pressure on this strip is
approximately cd
.
= 62.5(7 3j
W
.
) and so the force on the strip is approximately
62.5(7 3j
W
.
)2
25 3(j
W
.
)
2
{j. The total force
1 = lim
n<"
n
.=1
62.5(7 3j
W
.
)2
25 3(j
W
.
)
2
{j = 125
5
0
(7 3j)
25 3j
2
dj
= 125
5
0
7
25 3j
2
dj 3
5
0
j
25 3j
2
dj
= 125
5
0
25 3j
2
dj 3
3
1
3
(25 3j
2
)
3'2
5
0
= 125
1
4
5
2
+
1
3
(0 3125)
= 125
175r
4
3
125
3
E 11,972 E 1.2 10
4
lb
11. Set up a vertical x-axis as shown. Then the area of the jth rectangular strip is
o
/
(2/ 3r
W
.
) {r.

n
.
2/ 3r
W
.
=
2o
2/
, so n
.
=
o
/
(2/ 3r
W
.
).
The pressure on the strip is cr

W
.
, so the hydrostatic force on the plate
E
n
.=1
cr
W
.
o
/
(2/ 3r
W
.
) {r. The total force
1 = lim
n<"
n
.=1
cr
W
.
o
/
(2/ 3r
W
.
) {r = c
o
/
I
0
r(2/ 3r) dr =
oc
/
I
0
2/r 3r
2
dr
=
oc
/
/r
2
3
1
3
r
3
I
0
=
oc
/
/
3
3
1
3
/
3
=
oc
/
2/
3
3
=
2
3
co/
2

365
8
4
I
3
=
n
.
r
W
.
i n
.
=
2r
W
.
I
3
. The area of the jth
rectangular strip is
2r
W
.
I
3
{r and the pressure on it is ao
4
I
3 3r
W
.
.
1 =

4
I
3
0
ao
4
I
3 3r
2r
I
3
dr = 8ao
4
I
3
0
rdr 3
2ao
I
3
4
I
3
0
r
2
dr
= 4ao
r
2
4
I
3
0
3
2ao
3
I
3
r
3
4
I
3
0
= 192ao 3
2ao
3
I
3
64 3
I
3 = 192ao 3128ao = 64ao
E 64(840)(9.8) E 5.27 10
5
N
15. (a) The top of the cube has depth d = 1 m320 cm = 80 cm = 0.8 m.
1 = aod E (1000)(9.8)(0.8)(0.2)
2
= 313.6 E 314 N
(b) The area of a strip is 0.2 {r and the pressure on it is aor
W
.
.
1 =

1
0.8
aor(0.2) dr = 0.2ao
1
2
r
2
1
0.8
= (0.2ao)(0.18) = 0.036ao = 0.036(1000)(9.8) = 352.8 E 353 N
17. (a) The area of a strip is 20 {r and the pressure on it is cr
.
.
1 =

3
0
cr20 dr = 20c
1
2
r
2
3
0
= 20c
9
2
= 90c
= 90(62.5) = 5625 lb E 5.63 10
3
lb
(b) 1 =

9
0
cr20 dr = 20c
1
2
r
2
9
0
= 20c
81
2
= 810c = 810(62.5) = 50,625 lb E 5.06 10
4
lb.
(c) For the rst 3 ft, the length of the side is constant at 40 ft. For 3 < r $ 9, we can use similar triangles to nd the length o:
o
40
=
9 3r
6
i o = 40
9 3r
6
.
1 =

3
0
cr40 dr +
9
3
cr(40)
9 3r
6
dr = 40c
1
2
r
2
3
0
+
20
3
c
9
3
(9r 3r
2
) dr = 180c +
20
3
c
9
2
r
2
3
1
3
r
3
9
3
= 180c +
20
3
c
729
2
3243
81
2
39
= 180c + 600c = 780c = 780(62.5) = 48,750 lb E 4.88 10

4
lb
(d) For any right triangle with hypotenuse on the bottom,
sin0 =
{r
hypotenuse
i
hypotenuse = {rcsc 0 = {r
I
40
2
+ 6
2
6
=
I
409
3
{r.
1 =

9
3
cr20
I
409
3
dr =
1
3
20
I
409
1
2
r
2
9
3
=
1
3
10
I
409 c(81 39) E 303,356 lb E 3.03 10
5
lb
19. 1 =

5
2
aor n(r) dr, where n(r) is the width of the plate at depth r. Since n = 6, {r =
5 32
6
=
1
2
, and
1 E o
6
= ao
1'2
3
[2 n(2) + 4 2.5 n(2.5) + 2 3 n(3) + 4 3.5 n(3.5) + 2 4 n(4) + 4 4.5 n(4.5) + 5 n(5)]
=
1
6
ao(2 0 + 10 0.8 + 6 1.7 + 14 2.4 + 8 2.9 + 18 3.3 + 5 3.6)
=
1
6
(1000)(9.8)(152.4) E 2.5 10
5
N
366

21. The moment A of the system about the origin is A =
2
.=1
i
.
r
.
= i
1
r
1
+i
2
r
2
= 40 2 + 30 5 = 230.
The mass i of the system is i =
2
.=1
i
.
= i
1
+i
2
= 40 + 30 = 70.
The center of mass of the system is Ai =
230
70
=
23
7
.
23. i =
3
.=1
i
.
= 6 + 5 + 10 = 21.
A
i
=
3
.=1
i
.
j
.
= 6(5) + 5(32) + 10(31) = 10; A
u
=
3
.=1
i
.
r
.
= 6(1) + 5(3) + 10(32) = 1.
r =
A
u
i
=
1
21
and j =
A
i
i
=
10
21
, so the center of mass of the system is

1
21
.
10
21
.
25. Since the region in the gure is symmetric about the y-axis, we know
that r = 0. The region is bottom-heavy, so we know that j < 2,
and we might guess that j = 1.5.
=

2
32
(4 3r
2
) dr = 2
2
0
(4 3r
2
) dr = 2
4r 3
1
3
r
3
2
0
= 2
8 3
8
3
=
32
3
.
r =
1
/
2
32
r(4 3r
2
) dr = 0 since 1(r) = r(4 3r
2
) is an odd
function (or since the region is symmetric about the y-axis).
j =
1
/
2
32
1
2
(4 3r
2
)
2
dr =
3
32

1
2
2
2
0
(16 38r
2
+r
4
) dr =
3
32
16r 3
8
3
r
3
+
1
5
r
5
2
0
=
3
32
32 3
64
3
+
32
5
= 3
1 3
2
3
+
1
5
= 3
8
15
=
8
5
Thus, the centroid is (r. j) =
0.
8
5
.
27. The region in the gure is right-heavy and bottom-heavy, so we know
r 0.5 and j < 1, and we might guess that r = 0.6 and j = 0.9.
=

1
0
c
i
dr = [c
i
]
1
0
= c 31.
r =
1
/
1
0
rc
i
dr =
1
c 31
[rc
i
3c
i
]
1
0
[by parts]
=
1
c 31
[0 3(31)] =
1
c 31
.
j =
1
/
1
0
1
2
(c
i
)
2
dr =
1
c 31

1
4
c
2i
1
0
=
1
4(c 31)
c
2
31
=
c +1
4
.
1
c 31
.
c +1
4
E (0.58. 0.93).
29. =

1
0
(r
1'2
3r
2
) dr =

2
3
r
3'2
3
1
3
r
3
1
0
=

2
3
3
1
3
30 =
1
3
.
r =
1
/
1
0
r(r
1'2
3r
2
) dr = 3
1
0
(r
3'2
3r
3
) dr
= 3
2
5
r
5'2
3
1
4
r
4
1
0
= 3
2
5
3
1
4
= 3
3
20
=
9
20
.
j =
1
/
1
0
1
2
(r
1'2
)
2
3(r
2
)
2
dr = 3
1
2

1
0
(r 3r
4
) dr
=
3
2
1
2
r
2
3
1
5
r
5
1
0
=
3
2
1
2
3
1
5
=
3
2
3
10
=
9
20
.
9
20
.
9
20
.

367
31. =

r'4
0
(cos r 3sinr) dr =

sinr + cos r
r'4
0
=
I
2 31.
r =
31
r'4
0
r(cos r 3sinr) dr
=
31
r(sinr + cos r) + cos r 3sinr
r'4
0
[integration by parts]
=
31
r
4
I
2 31
=
1
4
I
2 31
I
2 31
.
j =
31
r'4
0
1
2
(cos
2
r 3sin
2
r) dr =
1
2/
r'4
0
cos 2rdr =
1
4/
sin2r
r'4
0
=
1
4
=
1
4
I
2 31
.

I
2 34
4
I
2 31
.
1
4
I
2 31
E (0.27. 0.60).
33. From the gure we see that j = 0. Now
=

5
0
2
I
5 3rdr = 2
3
2
3
(5 3r)
3'2
5
0
= 2
0 +
2
3
5
3'2
=
20
3
I
5,
so
r =
1
/
5
0
r
I
5 3r 3
3
I
5 3r
dr =
1
/
5
0
2r
I
5 3rdr
=
1
/
0
I
5
2
5 3&
2
&(32&) d&

u =
s
5 r, r = 5 u
2
,
u
2
= 5 r, Jr = 2uJu
=
4
/
I
5
0
&
2
(5 3&
2
) d& =
4
/
5
3
&
3
3
1
5
&
5
I
5
0
=
3
5
I
5
25
3
I
5 35
I
5
= 5 33 = 2.
Thus, the centroid is (r. j) = (2. 0).
35. The line has equation j =
3
4
r. =
1
2
(4)(3) = 6, so i = a = 10(6) = 60.
A
i
= a
4
0
1
2
3
4
r
2
dr = 10
4
0
9
32
r
2
dr =
45
16
1
3
r
3
4
0
=
45
16
64
3
= 60
A
u
= a
4
0
r
3
4
r
dr =
15
2
4
0
r
2
dr =
15
2
1
3
r
3
4
0
=
15
2
64
3
= 160
r =
A
u
i
=
160
60
=
8
3
and j =
A
i
i
=
60
60
= 1. Thus, the centroid is (r. j) =
8
3
. 1
.
37. =

2
0
(2
i
3r
2
) dr =

2
i
ln2
3
r
3
3
2
0
=

4
ln2
3
8
3
3
1
ln2
=
3
ln2
3
8
3
E 1.661418.
r =
1
2
0
r(2
i
3r
2
) dr =
1
2
0
(r2
i
3r
3
) dr
=
1
r2
i
ln2
3
2
i
(ln2)
2
3
r
4
4
2
0
[use parts]
=
1
8
ln2
3
4
(ln2)
2
34 +
1
(ln2)
2
=
1
8
ln2
3
3
(ln2)
2
34
E
1
(1.297453) E 0.781
[continued]
368

j =
1
2
0
1
2
[(2
i
)
2
3(r
2
)
2
] dr =
1
2
0
1
2
(2
2i
3r
4
) dr =
1

1
2
2
2i
2 ln2
3
r
5
5
2
0
=
1

1
2
16
2 ln2
3
32
5
3
1
2 ln2
=
1
15
4 ln2
3
16
5
E
1
(2.210106) E 1.330
Thus, the centroid is (r. j) E (0.781. 1.330).
Since the position of a centroid is independent of density when the density is constant, we will assume for convenience that a = 1 in Exercises 38
and 39.
39. Choose r- and j-axes so that the base (one side of the triangle) lies along
the r-axis with the other vertex along the positive j-axis as shown. From
geometry, we know the medians intersect at a point
2
3
of the way from each
vertex (along the median) to the opposite side. The median from 1 goes to
the midpoint
1
2
(o +c). 0
of side C, so the point of intersection of the

medians is
2
3

1
2
(o +c).
1
3
/
=

1
3
(o +c).
1
3
/
.
This can also be veried by nding the equations of two medians, and solving them simultaneously to nd their point of
intersection. Now let us compute the location of the centroid of the triangle. The area is =
1
2
(c 3o)/.
r =
1
0
a
r
/
o
(o 3r) dr +
c
0
r
/
c
(c 3r) dr
=
1
/
o
0
a
(or 3r
2
) dr +
/
c
c
0
cr 3r
2
dr
=
/
o
1
2
or
2
3
1
3
r
3
0
a
+
/
c
1
2
cr
2
3
1
3
r
3
c
0
=
/
o
3
1
2
o
3
+
1
3
o
3
+
/
c
1
2
c
3
3
1
3
c
3
=
2
o (c 3 o)

3o
3
6
+
2
c (c 3 o)

c
3
6
=
1
3 (c 3 o)
(c
2
3o
2
) =
o + c
3
and j =
1
0
a
1
2
/
o
(o 3r)
2
dr +
c
0
1
2
/
c
(c 3r)
2
dr
=
1
/
2
2o
2
0
a
(o
2
32or +r
2
) dr +
/
2
2c
2
c
0
(c
2
32cr +r
2
) dr
=
1
/
2
2o
2
o
2
r 3or
2
+
1
3
r
3
0
a
+
/
2
2c
2
c
2
r 3cr
2
+
1
3
r
3
c
0
=
1
/
2
2o
2
3o
3
+o
3
3
1
3
o
3
+
/
2
2c
2
c
3
3c
3
+
1
3
c
3
=
1
/
2
6
(3o +c)
=
2
(c 3 o) /

(c 3o)/
2
6
=
/
3
o +c
3
.
/
3
, as claimed.
Remarks: Actually the computation of j is all that is needed. By considering each side of the triangle in turn to be the base,
we see that the centroid is
1
3
of the way from each side to the opposite vertex and must therefore be the intersection of the
medians.

369
The computation of j in this problem (and many others) can be
simplied by using horizontal rather than vertical approximating rectangles.
If the length of a thin rectangle at coordinate j is /(j), then its area is
/(j) {j, its mass is a/(j) {j, and its moment about the r-axis is
{A
i
= aj/(j) {j. Thus,
A
i
=

aj/(j) dj and j =
aj/(j) dj
a
=
1
j/(j) dj
In this problem, /(j) =
c 3o
/
(/ 3j) by similar triangles, so
j =
1
l
0
c 3o
/
j(/ 3j) dj =
2
/
2
l
0
(/j 3j
2
) dj =
2
/
2
1
2
/j
2
3
1
3
j
3
l
0
=
2
/
2

/
3
6
=
/
3
Notice that only one integral is needed when this method is used.
41. Divide the lamina into two triangles and one rectangle with respective masses of 2, 2 and 4, so that the total mass is 8. Using
the result of Exercise 39, the triangles have centroids
31.
2
3
and
1.
2
3
. The centroid of the rectangle (its center) is
0. 3
1
2
.
So, using Formulas 5 and 7, we have j =
A
i
i
=
1
i
3
.=1
i
.
j
.
=
1
8
2
3
+ 2
2
3
+ 4
3
1
2
=
1
8
2
3
=
1
12
, and r = 0,
since the lamina is symmetric about the line r = 0. Thus, the centroid is (r. j) =
0.
1
12
.
43.

l
a
(cr +d) 1(r) dr =

l
a
cr1(r) dr +
l
a
d1(r) dr = c
l
a
r1(r) dr +d
l
a
1(r) dr = cr+d
l
a
1(r) dr [by (8)]
= cr
l
a
1(r) dr +d
l
a
1(r) dr = (cr +d)
l
a
1(r) dr
45. A cone of height / and radius v can be generated by rotating a right triangle
about one of its legs as shown. By Exercise 39, r =
1
3
v, so by the Theorem of
Pappus, the volume of the cone is
\ = d =

1
2
base height
(2r) =
1
2
v/ 2
1
3
v
=
1
3
v
2
/.
47. Suppose the region lies between two curves j = 1(r) and j = o(r) where 1(r) D o(r), as illustrated in Figure 13.
Choose points r
.
with o = r
0
< r
1
< < r
n
= / and choose r
W
.
to be the midpoint of the jth subinterval; that is,
r
W
.
= r
.
=
1
2
(r
.31
+r
.
). Then the centroid of the jth approximating rectangle 1
.
is its center C
.
=

r
.
.
1
2
[1(r
.
) +o(r
.
)]
.
Its area is [1(r
.
) 3o(r
.
)] {r, so its mass is
a[1(r
.
) 3o(r
.
)] {r. Thus, A
u
(1
.
) = a[1(r
.
) 3o(r
.
)] {r r
.
= ar
.
[1(r
.
) 3o(r
.
)] {r and
A
i
(1
.
) = a[1(r
.
) 3o(r
.
)] {r
1
2
[1(r
.
) +o(r
.
)] = a
1
2
1(r
.
)
2
3o(r
.
)
2
{r. Summing over j and taking the limit

as n < ", we get A
u
= lim
n<"
.
ar
.
[1(r
.
) 3o(r
.
)] {r = a
l
a
r[1(r) 3o(r)] dr and
A
i
= lim
n<"
.
a
1
2
1(r
.
)
2
3o(r
.
)
2
{r = a
l
a
1
2
1(r)
2
3o(r)
2
dr.
Thus, r =
A
u
i
=
A
u
a
=
1
l
a
r[1(r) 3o(r)] dr and j =
A
i
i
=
A
i
a
=
1
l
a
1
2
1(r)
2
3o(r)
2
dr.
370

8.4 Applications to Economics and Biology
1. By the Net Change Theorem, C(2000) 3C(0) =
2000
0
C
0
(r) dr i
C(2000) = 20,000 +
2000
0
(5 30.008r + 0.000009r
2
) dr = 20,000 +
5r 30.004r
2
+ 0.000003r
3
2000
0
= 20,000 + 10,000 30.004(4,000,000) + 0.000003(8,000,000,000) = 30,000 316,000 + 24,000
= $38,000
3. If the production level is raised from 1200 units to 1600 units, then the increase in cost is
C(1600) 3C(1200) =

1600
1200
C
0
(r) dr =

1600
1200
(74 + 1.1r 30.002r
2
+ 0.00004r
3
) dr
=

74r + 0.55r
2
3
0.002
3
r
3
+ 0.00001r
4
1600
1200
= 64,331,733.33 320,464,800 = $43,866,933.33
5. j(r) = 10 i
450
r + 8
= 10 i r + 8 = 45 i r = 37.
Consumer surplus =
37
0
[j(r) 310] dr =

37
0
450
r + 8
310
dr
=
450 ln(r + 8) 310r
37
0
= (450 ln45 3370) 3450 ln8
= 450 ln
45
8
3370 E $407.25
7. 1 = j
S
(r) i 400 = 200 + 0.2r
3'2
i 200 = 0.2r
3'2
i 1000 = r
3'2
i r = 1000
2'3
= 100.
Producer surplus =

100
0
[1 3j
S
(r)] dr =

100
0
[400 3(200 + 0.2r
3'2
)] dr =

100
0
200 3
1
5
r
3'2
dr
=

200r 3
2
25
r
5'2
100
0
= 20,000 38,000 = $12,000
9. j(r) =
800,000c
3i'5000
r + 20,000
= 16 i r = r
1
E 3727.04.
Consumer surplus =

i
1
0
[ j(r) 316] dr E $37,753
11. 1(8) 31(4) =

8
4
1
0
(t) dt =

8
4
I
t dt =

2
3
t
3'2
8
4
=
2
3
16
I
2 38
E $9.75 million
13. ` =

l
a
r
3I
dr =
r
3I+1
3I + 1
l
a
=

1 3I
(/
13I
3o
13I
).
Similarly,
l
a
r
13I
dr =
r
23I
2 3I
l
a
=

2 3I
(/
23I
3o
23I
).
Thus, r =
1
`
l
a
r
13I
dr =
[(2 3I)](/
23I
3o
23I
)
[(1 3I)](/
13I
3o
13I
)
=
(1 3I)(/
23I
3o
23I
)
(2 3I)(/
13I
3o
13I
)
.
15. 1 =
11
4
8|
=
(4000)(0.008)
4
8(0.027)(2)
E 1.19 10
34
cm
3
s
SECTION 8.5 PROBABILITY

371
17. From (3), 1 =

T
0
c (t) dt
=
6
201
, where
1 =

10
0
tc
30.6I
dt =

1
(30.6)
2
(30.6t 31) c
30.6I
10
0
integrating
by parts
=
1
0.36
(37c
36
+ 1)
Thus, 1 =
6(0.36)
20(1 37c
36
)
=
0.108
1 37c
36
E 0.1099 Ls or 6.594 Lmin.
19. As in Example 2, we will estimate the cardiac output using Simpsons Rule with {t = (16 30)8 = 2.
16
0
c(t) dt E
2
3
[c(0) + 4c(2) + 2c(4) + 4c(6) + 2c(8) + 4c(10) + 2c(12) + 4c(14) +c(16)]
E
2
3
[0 + 4(6.1) + 2(7.4) + 4(6.7) + 2(5.4) + 4(4.1) + 2(3.0) + 4(2.1) + 1.5]
=
2
3
(109.1) = 72.73 mg sL
Therefore, 1 E

72.73
=
7
72.73
E 0.0962 Ls or 5.77 Lmin.
8.5 Probability
1. (a)

40,000
30,000
1(r) dr is the probability that a randomly chosen tire will have a lifetime between 30,000 and 40,000 miles.
(b)

"
25,000
1(r) dr is the probability that a randomly chosen tire will have a lifetime of at least 25,000 miles.
3. (a) In general, we must satisfy the two conditions that are mentioned before Example 1namely, (1) 1(r) D 0 for all r,
and (2)
"
3"
1(r) dr = 1. For 0 $ r $ 4, we have 1(r) =
3
64
r
I
16 3r
2
D 0, so 1(r) D 0 for all r. Also,
"
3"
1(r) dr =

4
0
3
64
r
I
16 3r
2
dr = 3
3
128
4
0
(16 3r
2
)
1'2
(32r) dr = 3
3
128
2
3
(16 3r
2
)
3'2
4
0
= 3
1
64
16 3r
2
3'2
4
0
= 3
1
64
(0 364) = 1.
Therefore, 1 is a probability density function.
(b) 1(A < 2) =

2
3"
1(r) dr =

2
0
3
64
r
I
16 3r
2
dr = 3
3
128
2
0
(16 3r
2
)
1'2
(32r) dr
= 3
3
128
2
3
(16 3r
2
)
3'2
2
0
= 3
1
64
(16 3r
2
)
3'2
2
0
= 3
1
64
(12
3'2
316
3'2
)
=
1
64
64 312
I
12
=
1
64
64 324
I
3
= 1 3
3
8
I
3 E 0.350481
and (2)
"
3"
1(r) dr = 1. If c D 0, then 1(r) D 0, so condition (1) is satised. For condition (2), we see that
"
3"
1(r) dr =

"
3"
c
1 +r
2
dr and
Similarly,
"
0
c
1 +r
2
dr = lim
I<"
I
0
c
1 +r
2
dr = c lim
I<"
tan
31
r
I
0
= c lim
I<"
tan
31
t = c
0
3"
c
1 +r
2
dr = c
, so

"
3"
c
1 +r
2
dr = 2c
= c.
Since c must equal 1, we must have c = 1 so that 1 is a probability density function.
372

(b) 1 (31 < A < 1) =

1
31
1
1 +r
2
dr =
2
1
0
1
1 +r
2
dr =
2
tan
31
r
1
0
=
2
4
30
=
1
2
and (2)
"
3"
1(r) dr = 1. Since 1(r) = 0 or 1(r) = 0.1, condition (1) is satised. For condition (2), we see that
"
3"
1(r) dr =

10
0
0.1 dr =

1
10
r
10
0
= 1. Thus, 1(r) is a probability density function for the spinners values.
(b) Since all the numbers between 0 and 10 are equally likely to be selected, we expect the mean to be halfway between the
endpoints of the interval; that is, r = 5.
j =

"
3"
r1(r) dr =

10
0
r(0.1) dr =

1
20
r
2
10
0
=
100
20
= 5, as expected.
9. We need to nd i so that
"
r
1(t) dt =
1
2
i lim
i<"
i
r
1
5
c
3I'5
dt =
1
2
i lim
i<"
1
5
(35)c
3I'5
i
r
=
1
2
i
(31)(0 3c
3r'5
) =
1
2
i c
3r'5
=
1
2
i 3i5 = ln
1
2
i i = 35 ln
1
2
= 5 ln2 E 3.47 min.
11. We use an exponential density function with j = 2.5 min.
(a) 1(A 4) =
"
4
1(t) dt = lim
i<"
i
4
1
2.5
c
3I'2.5
dt = lim
i<"
3c
3I'2.5
i
4
= 0 +c
34'2.5
E 0.202
(b) 1(0 $ A $ 2) =

2
0
1(t) dt =

3c
3I'2.5
2
0
= 3c
32'2.5
+ 1 E 0.551
(c) We need to nd a value o so that 1(A D o) = 0.02, or, equivalently, 1(0 $ A $ o) = 0.98 C
a
0
1(t) dt = 0.98 C

3c
3I'2.5
a
0
= 0.98 C 3c
3a'2.5
+ 1 = 0.98 C c
3a'2.5
= 0.02 C
3o2.5 = ln0.02 C o = 32.5 ln
1
50
= 2.5 ln50 E 9.78 min E 10 min. The ad should say that if you arent served
within 10 minutes, you get a free hamburger.
13. 1(A D 10) =

"
10
1
4.2
I
2
exp
3
(r 39.4)
2
2 4.2
2
dr. To avoid the improper integral we approximate it by the integral from

10 to 100. Thus, 1(A D 10) E
100
10
1
4.2
I
2
exp
3
(r 39.4)
2
2 4.2
2
dr E 0.443 (using a calculator or computer to estimate

the integral), so about 44 percent of the households throw out at least 10 lb of paper a week.
Note: We cant evaluate 1 31(0 $ A $ 10) for this problem since a signicant amount of area lies to the left of A = 0.
15. (a) 1(0 $ A $ 100) =

100
0
1
8
I
2
exp
3
(r 3112)
2
2 8
2
dr E 0.0668 (using a calculator or computer to estimate the

integral), so there is about a 6.68% chance that a randomly chosen vehicle is traveling at a legal speed.
(b) 1(A D 125) =

"
125
1
8
I
2
exp
3
(r 3112)
2
2 8
2
dr =

"
125
1(r) dr. In this case, we could use a calculator or computer
to estimate either
300
125
1(r) dr or 1 3
125
0
1(r) dr. Both are approximately 0.0521, so about 5.21% of the motorists are
targeted.
SECTION 8.5 PROBABILITY

373
17. 1(j 32o $ A $ j + 2o) =

+2
32
1
o
I
2
exp
3
(r 3j)
2
2o
2
dr. Substituting t =
r 3j
o
and dt =
1
o
dr gives us
2
32
1
o
I
2
c
3I
2
/2
(o dt) =
1
I
2
2
32
c
3I
2
/2
dt E 0.9545.
19. (a) First j(v) =
4
o
3
0
v
2
c
32r'a
0
D 0 for v D 0. Next,
"
3"
j(v) dv =

"
0
4
o
3
0
v
2
c
32r'a
0
dv =
4
o
3
0
lim
I<"
I
0
v
2
c
32r'a
0
dv
By using parts, tables, or a CAS , we nd that
r
2
c
li
dr = (c
li
/
3
)(/
2
r
2
32/r + 2). (-)
Next, we use (-) (with / = 32o
0
) and lHospitals Rule to get
4
o
3
0
o
3
0
38
(32)
= 1. This satises the second condition for

a function to be a probability density function.
(b) Using lHospitals Rule,
4
o
3
0
lim
r<"
v
2
c
2r'a
0
=
4
o
3
0
lim
r<"
2v
(2o
0
)c
2r'a
0
=
2
o
2
0
lim
r<"
2
(2o
0
)c
2r'a
0
= 0.
To nd the maximum of j, we differentiate:
j
0
(v) =
4
o
3
0
v
2
c
32r'a
0
3
2
o
0
+c
32r'a
0
(2v)
=
4
o
3
0
c
32r'a
0
(2v)
3
v
o
0
+ 1
j
0
(v) = 0 C v = 0 or 1 =
v
o
0
C v = o
0
[o
0
E 5.59 10
311
m].
j
0
(v) changes from positive to negative at v = o
0
, so j(v) has its maximum value at v = o
0
.
(c) It is fairly difcult to nd a viewing rectangle, but knowing the maximum
value from part (b) helps.
j(o
0
) =
4
o
3
0
o
2
0
c
32a
0
'a
0
=
4
o
0
c
32
E 9,684,098,979
With a maximum of nearly 10 billion and a total area under the curve of 1,
we know that the hump in the graph must be extremely narrow.
(d) 1(v) =
r
0
4
o
3
0
c
2
c
32s'a
0
dc i 1(4o
0
) =
4a
0
0
4
o
3
0
c
2
c
32s'a
0
dc. Using (-) from part (a) [with / = 32o
0
],
1(4o
0
) =
4
o
3
0
c
32s'a
0
38o
3
0
4
o
2
0
c
2
+
4
o
0
c + 2
4a
0
0
=
4
o
3
0
o
3
0
38
[c
38
(64 + 16 + 2) 31(2)] = 3
1
2
(82c
38
32)
= 1 341c
38
E 0.986
(e) j =

"
3"
vj(v) dv =
4
o
3
0
lim
I<"
I
0
v
3
c
32r'a
0
dv. Integrating by parts three times or using a CAS, we nd that
r
3
c
li
dr =
c
li
/
4
/
3
r
3
33/
2
r
2
+ 6/r 36
. So with / = 3
2
o
0
, we use lHospitals Rule, and get
j =
4
o
3
0
3
o
4
0
16
(36)
=
3
2
o
0
.
374

1. (a) The length of a curve is dened to be the limit of the lengths of the inscribed polygons, as described near Figure 3 in
Section 8.1.
(b) See Equation 8.1.2.
(c) See Equation 8.1.4.
2. (a) o =

l
a
21(r)
1 + [1
0
(r)]
2
dr
(b) If r = o(j), c $ j $ d, then o =

u
c
2j
1 + [o
0
(j)]
2
dj.
(c) o =

l
a
2r
1 + [1
0
(r)]
2
dr or o =

u
c
2o(j)
1 + [o
0
(j)]
2
dj
3. Let c(r) be the cross-sectional length of the wall (measured parallel to the surface of the uid) at depth r. Then the hydrostatic
force against the wall is given by 1 =

l
a
crc(r) dr, where o and / are the lower and upper limits for r at points of the wall
and c is the weight density of the uid.
4. (a) The center of mass is the point at which the plate balances horizontally.
(b) See Equations 8.3.8.
5. If a plane region Rthat lies entirely on one side of a line / in its plane is rotated about /, then the volume of the resulting solid
is the product of the area of Rand the distance traveled by the centroid of R.
6. See Figure 3 in Section 8.4, and the discussion which precedes it.
7. (a) See the denition in the rst paragraph of the subsection Cardiac Output in Section 8.4.
(b) See the discussion in the second paragraph of the subsection Cardiac Output in Section 8.4.
8. A probability density function 1 is a function on the domain of a continuous random variable A such that
l
a
1(r) dr
measures the probability that A lies between o and /. Such a function 1 has nonnegative values and satises the relation
T
1(r) dr = 1, where 1 is the domain of the corresponding random variable A. If 1 = R, or if we dene 1(r) = 0 for real
numbers r M 1, then
"
3"
1(r) dr = 1. (Of course, to work with 1 in this way, we must assume that the integrals of 1 exist.)
9. (a)

130
0
1(r) dr represents the probability that the weight of a randomly chosen female college student is less than
130 pounds.
(b) j =

"
3"
r1(r) dr =

"
0
r1(r) dr
(c) The median of 1 is the number i such that
"
r
1(r) dr =
1
2
.
10. See the discussion near Equation 3 in Section 8.5.
8 Review
CHAPTER 8 REVIEW

375
1. j =
1
6
(r
2
+ 4)
3'2
i djdr =
1
4
(r
2
+ 4)
1'2
(2r) i
1 + (djdr)
2
= 1 +
1
2
r(r
2
+ 4)
1'2
2
= 1 +
1
4
r
2
(r
2
+ 4) =
1
4
r
4
+r
2
+ 1 =

1
2
r
2
+ 1
2
.
Thus, 1 =

3
0
1
2
r
2
+ 1
2
dr =

3
0
1
2
r
2
+ 1
dr =

1
6
r
3
+r
3
0
=
15
2
.
3. (a) j =
r
4
16
+
1
2r
2
=
1
16
r
4
+
1
2
r
32
i
dj
dr
=
1
4
r
3
3r
33
i
1 + (djdr)
2
= 1 +
1
4
r
3
3r
33
2
= 1 +
1
16
r
6
3
1
2
+r
36
=
1
16
r
6
+
1
2
+r
36
=

1
4
r
3
+r
33
2
.
Thus, 1 =

2
1
1
4
r
3
+r
33
dr =

1
16
r
4
3
1
2
r
32
2
1
=

1 3
1
8
1
16
3
1
2
=
21
16
.
(b) o =

2
1
2r
1
4
r
3
+r
33
dr = 2
2
1
1
4
r
4
+r
32
dr = 2
1
20
r
5
3
1
i
2
1
= 2
32
20
3
1
2
1
20
31
= 2
8
5
3
1
2
3
1
20
+ 1
= 2
41
20
=
41
10
5. j = c
3i
2
i djdr = 32rc
3i
2
i 1 + (djdr)
2
= 1 + 4r
2
c
32i
2
. Let 1(r) =
1 + 4r
2
c
32i
2
. Then
1 =

3
0
1(r) dr E o
6
=
(3 30)6
3
[1(0) + 41(0.5) + 21(1) + 41(1.5) + 21(2) + 41(2.5) +1(3)] E 3.292287
7. j =

i
1
I
t 31 dt i djdr =

I
r 31 i 1 + (djdr)
2
= 1 +
I
r 31
=
I
r.
Thus, 1 =

16
1
I
rdr =

16
1
r
1'4
dr =
4
5
r
5'4
16
1
=
4
5
(32 31) =
124
5
.
9. As in Example 1 of Section 8.3,
o
2 3r
=
1
2
i 2o = 2 3r and n = 2(1.5 +o) = 3 + 2o = 3 + 2 3r = 5 3r.
Thus, 1 =

2
0
aor(5 3r) dr = ao
5
2
r
2
3
1
3
r
3
2
0
= ao
10 3
8
3
=
22
3
c [ao = c] E
22
3
62.5 E 458 lb.
11. =

4
0
I
r 3
1
2
r
dr =

2
3
r
3'2
3
1
4
r
2
4
0
=
16
3
34 =
4
3
r =
1
/
4
0
r
I
r 3
1
2
r
dr =
3
4
4
0
r
3'2
3
1
2
r
2
dr
=
3
4
2
5
r
5'2
3
1
6
r
3
4
0
=
3
4
64
5
3
64
6
=
3
4
64
30
=
8
5
j =
1
/
4
0
1
2
I
r
2
3
1
2
r
dr =
3
4
4
0
1
2
r 3
1
4
r
2
dr =
3
8
1
2
r
2
3
1
12
r
3
4
0
=
3
8
8 3
16
3
=
3
8
8
3
= 1
8
5
. 1
.
13. An equation of the line passing through (0, 0) and (3, 2) is j =
2
3
r. =
1
2
3 2 = 3. Therefore, using Equations 8.3.8,
r =
1
3
3
0
r
2
3
r
dr =
2
27
r
3
3
0
= 2 and j =
1
3
3
0
1
2
2
3
r
2
dr =
2
81
r
3
3
0
=
2
3
. Thus, the centroid is (r. j) =
2.
2
3
.
376

15. The centroid of this circle, (1. 0), travels a distance 2(1) when the lamina is rotated about the j-axis. The area of the circle
is (1)
2
. So by the Theorem of Pappus, \ = (2r) = (1)
2
2(1) = 2
2
.
17. r = 100 i 1 = 2000 30.1(100) 30.01(100)
2
= 1890
Consumer surplus =

100
0
[ j(r) 31] dr =

100
0
2000 30.1r 30.01r

2
31890
dr
=

110r 30.05r
2
3
0.01
3
r
3
100
0
= 11,000 3500 3
10,000
3
E $7166.67
19. 1(r) =
r
20
sin
r
10
r
if 0 $ r $ 10
0 if r < 0 or r 10
(a) 1(r) D 0 for all real numbers r and
"
3"
1(r) dr =

10
0
r
20
sin
r
10
r
dr =
r
20

10
r
3cos
r
10
r
10
0
=
1
2
(3cos + cos 0) =
1
2
(1 + 1) = 1
Therefore, 1 is a probability density function.
(b) 1(A < 4) =

4
3"
1(r) dr =

4
0
r
20
sin
r
10
r
dr =
1
2
3cos
r
10
r
4
0
=
1
2
3cos
2r
5
+ cos 0
E
1
2
(30.309017 + 1) E 0.3455
(c) j =

"
3"
r1(r) dr =

10
0
r
20
rsin
r
10
r
dr
=

r
0
r
20

10
r
&(sin&)
10
r
d& [u =

10
r, Ju =

10
Jr]
=
5
r
r
0
&sin&d&
82
=
5
r
[sin& 3&cos &]
r
0
=
5
r
[0 3(31)] = 5
This answer is expected because the graph of 1 is symmetric about the
line r = 5.
21. (a) The probability density function is 1(t) =
0 if t < 0
1
8
c
3I'8
if t D 0
1(0 $ A $ 3) =

3
0
1
8
c
3I'8
dt =

3c
3I'8
3
0
= 3c
33'8
+ 1 E 0.3127
(b) 1(A 10) =

"
10
1
8
c
3I'8
dt = lim
i<"
3c
3I'8
i
10
= lim
i<"
(3c
3i'8
+c
310'8
) = 0 +c
35'4
E 0.2865
(c) We need to nd i such that 1(A D i) =
1
2
i

"
r
1
8
c
3I'8
dt =
1
2
i lim
i<"
3c
3I'8
i
r
=
1
2
i
lim
i<"
(3c
3i'8
+c
3r'8
) =
1
2
i c
3r'8
=
1
2
i 3i8 = ln
1
2
i i = 38 ln
1
2
= 8 ln2 E 5.55 minutes.
PROBLEMS PLUS
1. r
2
+j
2
$ 4j C r
2
+ (j 32)
2
$ 4, so o is part of a circle, as shown
in the diagram. The area of o is
1
0
4j 3j
2
dj
113
=

u32
2
4j 3j
2
+ 2 cos
31
23u
2
1
0
[o = 2]
= 3
1
2
I
3 + 2 cos
31
1
2
32 cos
31
1
= 3
I
3
2
+ 2
r
3
32(0) =
2r
3
3
I
3
2
Another method (without calculus): Note that 0 = _C1 =
r
3
, so the area is
(area of sector O1) 3(area of 41C) =
1
2
2
2
r
3
3
1
2
(1)
I
3 =
2r
3
3
I
3
2
3. (a) The two spherical zones, whose surface areas we will call o
1
and o
2
, are
generated by rotation about the j-axis of circular arcs, as indicated in the gure.
The arcs are the upper and lower portions of the circle r
2
+j
2
= v
2
that are
obtained when the circle is cut with the line j = d. The portion of the upper arc
in the rst quadrant is sufcient to generate the upper spherical zone. That
portion of the arc can be described by the relation r =
v
2
3j
2
for
d $ j $ v. Thus, drdj = 3j
v
2
3j
2
and
dc =
1 +
dr
dj
2
dj =
1 +
j
2
v
2
3j
2
dj =
v
2
v
2
3j
2
dj =
v dj
v
2
3j
2
From Formula 8.2.8 we have
o
1
=
r
u
2r
1 +
dr
dj
2
dj =
r
u
2
v
2
3j
2
v dj
v
2
3j
2
=
r
u
2v dj = 2v(v 3d)
Similarly, we can compute o
2
=

u
3r
2r
1 + (drdj)
2
dj =
u
3r
2v dj = 2v(v +d). Note that o
1
+o
2
= 4v
2
,
the surface area of the entire sphere.
(b) v = 3960 mi and d = v (sin75
) E 3825 mi,
so the surface area of the Arctic Ocean is about
2v(v3d) E 2(3960)(135) E 3.3610
6
mi
2
.
377
378

(c) The area on the sphere lies between planes j = j
1
and j = j
2
, where j
2
3j
1
= /. Thus, we compute the surface area on
the sphere to be o =
u
2
u
1
2r
1 +
dr
dj
2
dj =
u
2
u
1
2v dj = 2v(j
2
3j
1
) = 2v/.
This equals the lateral area of a cylinder of radius v and height /, since such
a cylinder is obtained by rotating the line r = v about the j-axis, so the
surface area of the cylinder between the planes j = j
1
and j = j
2
is
=
u
2
u
1
2r
1 +
dr
dj
2
dj =
u
2
u
1
2v
1 + 0
2
dj
= 2vj
u
2
u=u
1
= 2v(j
2
3j
1
) = 2v/
(d) / = 2v sin23.45
E 3152 mi, so the surface area of the

Torrid Zone is 2v/ E 2(3960)(3152) E 7.84 10
7
mi
2
.
5. (a) Choose a vertical r-axis pointing downward with its origin at the surface. In order to calculate the pressure at depth :,
consider n subintervals of the interval [0. :] by points r
.
and choose a point r
W
.
M [r
.31
. r
.
] for each j. The thin layer of
water lying between depth r
.31
and depth r
.
has a density of approximately a(r
W
.
), so the weight of a piece of that layer
with unit cross-sectional area is a(r
W
.
)o {r. The total weight of a column of water extending from the surface to depth :
(with unit cross-sectional area) would be approximately
n
.=1
a(r
W
.
)o {r. The estimate becomes exact if we take the limit
as n <"; weight (or force) per unit area at depth : is W = lim
n<"
n
.=1
a(r
W
.
)o {r. In other words, 1(:) =
:
0
a(r)o dr.
More generally, if we make no assumptions about the location of the origin, then 1(:) = 1
0
+
:
0
a(r)o dr, where 1
0
is
the pressure at r = 0. Differentiating, we get d1d: = a(:)o.
(b) 1 =
r
3r
1(1 +r) 2
I
v
2
3r
2
dr
=
r
3r
1
0
+
1+i
0
a
0
c
:'1
o d:
2
I
v
2
3r
2
dr
= 1
0
r
3r
2
I
v
2
3r
2
dr +a
0
o1
r
3r
c
(1+i)'1
31
2
I
v
2
3r
2
dr
= (1
0
3a
0
o1)
r
3r
2
I
v
2
3r
2
dr +a
0
o1
r
3r
c
(1+i)'1
2
I
v
2
3r
2
dr
= (1
0
3a
0
o1)
v
2
+a
0
o1c
1'1
r
3r
c
i'1
2
I
v
2
3r
2
dr

379
7. To nd the height of the pyramid, we use similar triangles. The rst gure shows a cross-section of the pyramid passing
through the top and through two opposite corners of the square base. Now |11| = /, since it is a radius of the sphere, which
has diameter 2/ since it is tangent to the opposite sides of the square base. Also, |1| = / since 411 is isosceles. So the
height is |1| =
I
/
2
+/
2
=
I
2 /.
We rst observe that the shared volume is equal to half the volume of the sphere, minus the sum of the four equal volumes
(caps of the sphere) cut off by the triangular faces of the pyramid. See Exercise 6.2.51 for a derivation of the formula for the
volume of a cap of a sphere. To use the formula, we need to nd the perpendicular distance / of each triangular face from the
surface of the sphere. We rst nd the distance d from the center of the sphere to one of the triangular faces. The third gure
shows a cross-section of the pyramid through the top and through the midpoints of opposite sides of the square base. From
similar triangles we nd that
d
/
=
|1|
|C|
=
I
2 /
/
2
+
I
2 /
2
i d =
I
2 /
2
I
3/
2
=
I
6
3
/
So / = / 3d = / 3
I
6
3
/ =
3 3
I
6
3
/. So, using the formula \ = /
2
(v 3/3) from Exercise 6.2.51 with v = /, we nd that
the volume of each of the caps is
3 3
I
6
3
/
/ 3
3 3
I
6
3 3
/
=
15 36
I
6
9

6 +
I
6
9
/
3
=
2
3
3
7
27
I
6
/
3
. So, using our rst
observation, the shared volume is \ =
1
2
4
3
/
3
34
2
3
3
7
27
I
6
/
3
=
28
27
I
6 32
/
3
.
9. We can assume that the cut is made along a vertical line r = / 0, that the
disks boundary is the circle r
2
+j
2
= 1, and that the center of mass of the
smaller piece (to the right of r = /) is
1
2
. 0
. We wish to nd / to two
decimal places. We have
1
2
= r =
1
l
r 2
I
1 3r
2
dr
1
l
2
I
1 3r
2
dr
. Evaluating the
numerator gives us 3
1
l
(1 3r
2
)
1'2
(32r) dr = 3
2
3
1 3r
2
3'2
1
l
= 3
2
3
0 3
1 3/
2
3'2
=
2
3
(1 3/
2
)
3'2
.
Using Formula 30 in the table of integrals, we nd that the denominator is
r
I
1 3r
2
+ sin
31
r
1
l
=

0 +
r
2
/
I
1 3/
2
+ sin
31
/
. Thus, we have
1
2
= r =
2
3
(1 3/
2
)
3'2
r
2
3/
I
1 3/
2
3sin
31
/
, or,
equivalently,
2
3
(1 3/
2
)
3'2
=
r
4
3
1
2
/
I
1 3/
2
3
1
2
sin
31
/. Solving this equation numerically with a calculator or CAS, we
obtain / E 0.138173, or / = 0.14 m to two decimal places.
380

11. If / = 1, then 1 =
area under j = 1sin0
area of rectangle
=
r
0
1sin0 d0
1
=
[3cos 0]
r
0
=
3(31) + 1
=
2
.
If / = 12, then 1 =
area under j =
1
2
1sin0
area of rectangle
=
r
0
1
2
1sin0 d0
1
=
[3cos 0]
r
0
2
=
2
2
=
1
.
9 DIFFERENTIAL EQUATIONS
9.1 Modeling with Differential Equations
1. j = r 3r
31
i j
0
= 1 +r
32
. To show that j is a solution of the differential equation, we will substitute the expressions
for j and j
0
in the left-hand side of the equation and show that the left-hand side is equal to the right-hand side.
LHS= rj
0
+j = r(1 +r
32
) + (r 3r
31
) = r +r
31
+r 3r
31
= 2r =RHS
3. (a) j = c
ri
i j
0
= vc
ri
i j
00
= v
2
c
ri
. Substituting these expressions into the differential equation
2j
00
+j
0
3j = 0, we get 2v
2
c
ri
+vc
ri
3c
ri
= 0 i (2v
2
+v 31)c
ri
= 0 i
(2v 31)(v + 1) = 0 [since c
ri
is never zero] i v =
1
2
or 31.
(b) Let v
1
=
1
2
and v
2
= 31, so we need to show that every member of the family of functions j = oc
i'2
+/c
3i
is a
solution of the differential equation 2j
00
+j
0
3j = 0.
j = oc
i'2
+/c
3i
i j
0
=
1
2
oc
i'2
3/c
3i
i j
00
=
1
4
oc
i'2
+/c
3i
.
LHS = 2j
00
+j
0
3j = 2
1
4
oc
i'2
+/c
3i
1
2
oc
i'2
3/c
3i
3(oc
i'2
+/c
3i
)
=
1
2
oc
i'2
+ 2/c
3i
+
1
2
oc
i'2
3/c
3i
3oc
i'2
3/c
3i
=

1
2
o +
1
2
o 3o
c
i'2
+ (2/ 3/ 3/)c
3i
= 0 = RHS
5. (a) j = sinr i j
0
= cos r i j
00
= 3sinr.
LHS = j
00
+j = 3sinr + sinr = 0 6= sinr, so j = sinr is not a solution of the differential equation.
(b) j = cos r i j
0
= 3sinr i j
00
= 3cos r.
LHS = j
00
+j = 3cos r + cos r = 0 6= sinr, so j = cos r is not a solution of the differential equation.
(c) j =
1
2
rsinr i j
0
=
1
2
(rcos r + sinr) i j
00
=
1
2
(3rsinr + cos r + cos r).
LHS = j
00
+j =
1
2
(3rsinr + 2 cos r) +
1
2
rsinr = cos r 6= sinr, so j =
1
2
rsinr is not a solution of the
differential equation.
(d) j = 3
1
2
rcos r i j
0
= 3
1
2
(3rsinr + cos r) i j
00
= 3
1
2
(3rcos r 3sinr 3sinr).
LHS = j
00
+j = 3
1
2
(3rcos r 32 sinr) +
3
1
2
rcos r
= sinr = RHS, so j = 3
1
2
rcos r is a solution of the
differential equation.
7. (a) Since the derivative j
0
= 3j
2
is always negative (or 0 if j = 0), the function j must be decreasing (or equal to 0) on any
interval on which it is dened.
(b) j =
1
r +C
i j
0
= 3
1
(r +C)
2
. LHS = j
0
= 3
1
(r +C)
2
= 3
1
r +C
2
= 3j
2
= RHS
381
382

CHAPTER 9 DIFFERENTIAL EQUATIONS
(c) j = 0 is a solution of j
0
= 3j
2
that is not a member of the family in part (b).
(d) If j(r) =
1
r +C
, then j(0) =
1
0 +C
=
1
C
. Since j(0) = 0.5,
1
C
=
1
2
i C = 2, so j =
1
r + 2
.
9. (a)
d1
dt
= 1.21
1 3
1
4200
. Now
d1
dt
0 i 1 3
1
4200
0 [assuming that 1 0] i
1
4200
< 1 i
1 < 4200 i the population is increasing for 0 < 1 < 4200.
(b)
d1
dt
< 0 i 1 4200
(c)
d1
dt
= 0 i 1 = 4200 or 1 = 0
11. (a) This function is increasing and also decreasing. But djdt = c
I
(j 31)
2
D 0 for all t, implying that the graph of the
solution of the differential equation cannot be decreasing on any interval.
(b) When j = 1, djdt = 0, but the graph does not have a horizontal tangent line.
13. (a) 1 increases most rapidly at the beginning, since there are usually many simple, easily-learned sub-skills associated with
learning a skill. As t increases, we would expect d1dt to remain positive, but decrease. This is because as time
progresses, the only points left to learn are the more difcult ones.
(b)
d1
dt
= I(A 31) is always positive, so the level of performance 1
is increasing. As 1 gets close to A, d1dt gets close to 0; that is,
the performance levels off, as explained in part (a).
(c)
9.2 Direction Fields and Euler's Method
1. (a) (b) It appears that the constant functions j = 0, j = 32, and j = 2 are
equilibrium solutions. Note that these three values of j satisfy the
given differential equation j
0
= j
1 3
1
4
j
2
.
3. j
0
= 2 3j. The slopes at each point are independent of r, so the slopes are the same along each line parallel to the r-axis.
Thus, III is the direction eld for this equation. Note that for j = 2, j
0
= 0.
5. j
0
= r +j 31 = 0 on the line j = 3r +1. Direction eld IV satises this condition. Notice also that on the line j = 3r we
have j
0
= 31, which is true in IV.
SECTION 9.2 DIRECTION FIELDS AND EULERS METHOD

383
7. (a) j(0) = 1
(b) j(0) = 2
(c) j(0) = 31
9.
r j j
0
= 1 +j
0 0 1
0 1 2
0 2 3
0 33 32
0 32 31
Note that for j = 31, j
0
= 0. The three solution curves sketched go
through (0. 0), (0. 31), and (0. 32).
11.
r j j
0
= j 32r
32 32 2
32 2 6
2 2 32
2 32 36
Note that j
0
= 0 for any point on the line j = 2r. The slopes are
positive to the left of the line and negative to the right of the line. The
solution curve in the graph passes through (1. 0).
13.
r j j
0
= j +rj
0 2 2
1 2 4
33 2 ~4
Note that j
0
= j(r + 1) = 0 for any point on j = 0 or on r = 31.
The slopes are positive when the factors j and r + 1 have the same
sign and negative when they have opposite signs. The solution curve
in the graph passes through (0. 1).
384

15. In Maple, we can use either directionfield (in Maples share library) or
DEtools[DEplot] to plot the direction eld. To plot the solution, we can
either use the initial-value option in directionfield, or actually solve the
equation.
In Mathematica, we use PlotVectorField for the direction eld, and the
Plot[Evaluate[. . .]] construction to plot the solution, which is
j = 2 arctan
c
i
3
'3
tan
1
2
.
In Derive, use Direction_Field (in utility le ODE_APPR) to plot the direction eld. Then use
DSOLVE1(-x2*SIN(y),1,x,y,0,1) (in utility le ODE1) to solve the equation. Simplify each result.
17. The direction eld is for the differential equation j
0
= j
3
34j.
1 = lim
I<"
j(t) exists for 32 $ c $ 2;
1 = 2 for c = 2 and 1 = 0 for 32 < c < 2.
For other values of c, 1 does not exist.
19. (a) j
0
= 1(r. j) = j and j(0) = 1 i r
0
= 0, j
0
= 1.
(i) / = 0.4 and j
1
= j
0
+/1(r
0
. j
0
) i j
1
= 1 + 0.4 1 = 1.4. r
1
= r
0
+/ = 0 + 0.4 = 0.4,
so j
1
= j (0.4) = 1.4.
(ii) / = 0.2 i r
1
= 0.2 and r
2
= 0.4, so we need to nd j
2
.
j
1
= j
0
+ /1(r
0
. j
0
) = 1 + 0.2j
0
= 1 + 0.2 1 = 1.2,
j
2
= j
1
+/1(r
1
. j
1
) = 1.2 + 0.2j
1
= 1.2 + 0.2 1.2 = 1.44.
(iii) / = 0.1 i r
4
= 0.4, so we need to nd j
4
. j
1
= j
0
+/1(r
0
. j
0
) = 1 + 0.1j
0
= 1 + 0.1 1 = 1.1,
j
2
= j
1
+/1(r
1
. j
1
) = 1.1 + 0.1j
1
= 1.1 + 0.1 1.1 = 1.21,
j
3
= j
2
+/1(r
2
. j
2
) = 1.21 + 0.1j
2
= 1.21 + 0.1 1.21 = 1.331,
j
4
= j
3
+/1(r
3
. j
3
) = 1.331 + 0.1j
3
= 1.331 + 0.1 1.331 = 1.4641.
(b) We see that the estimates are underestimates since
they are all below the graph of j = c
i
.
SECTION 9.2 DIRECTION FIELDS AND EULERS METHOD

385
(c) (i) For / = 0.4: (exact value) 3(approximate value) = c
0.4
31.4 E 0.0918
(ii) For / = 0.2: (exact value) 3(approximate value) = c
0.4
31.44 E 0.0518
(iii) For / = 0.1: (exact value) 3(approximate value) = c
0.4
31.4641 E 0.0277
Each time the step size is halved, the error estimate also appears to be halved (approximately).
21. / = 0.5, r
0
= 1, j
0
= 0, and 1(r. j) = j 32r.
Note that r
1
= r
0
+/ = 1 + 0.5 = 1.5, r
2
= 2, and r
3
= 2.5.
j
1
= j
0
+/1(r
0
. j
0
) = 0 + 0.51(1. 0) = 0.5[0 32(1)] = 31.
j
2
= j
1
+/1(r
1
. j
1
) = 31 + 0.51(1.5. 31) = 31 + 0.5[31 32(1.5)] = 33.
j
3
= j
2
+/1(r
2
. j
2
) = 33 + 0.51(2. 33) = 33 + 0.5[33 32(2)] = 36.5.
j
4
= j
3
+/1(r
3
. j
3
) = 36.5 + 0.51(2.5. 36.5) = 36.5 + 0.5[36.5 32(2.5)] = 312.25.
23. / = 0.1, r
0
= 0, j
0
= 1, and 1(r. j) = j +rj.
Note that r
1
= r
0
+/ = 0 + 0.1 = 0.1, r
2
= 0.2, r
3
= 0.3, and r
4
= 0.4.
j
1
= j
0
+/1(r
0
. j
0
) = 1 + 0.11(0. 1) = 1 + 0.1[1 + (0)(1)] = 1.1.
j
2
= j
1
+/1(r
1
. j
1
) = 1.1 + 0.11(0.1. 1.1) = 1.1 + 0.1[1.1 + (0.1)(1.1)] = 1.221.
j
3
= j
2
+/1(r
2
. j
2
) = 1.221 + 0.11(0.2. 1.221) = 1.221 + 0.1[1.221 + (0.2)(1.221)] = 1.36752.
j
4
= j
3
+/1(r
3
. j
3
) = 1.36752 + 0.11(0.3. 1.36752) = 1.36752 + 0.1[1.36752 + (0.3)(1.36752)]
= 1.5452976.
j
5
= j
4
+/1(r
4
. j
4
) = 1.5452976 + 0.11(0.4. 1.5452976)
= 1.5452976 + 0.1[1.5452976 + (0.4)(1.5452976)] = 1.761639264.
Thus, j(0.5) E 1.7616.
25. (a) djdr + 3r
2
j = 6r
2
i j
0
= 6r
2
33r
2
j. Store this expression in Y
1
and use the following simple program to
evaluate j(1) for each part, using H = / = 1 and N = 1 for part (i), H = 0.1 and N = 10 for part (ii), and so forth.
/ <H: 0 <X: 3 <Y:
For(I, 1, N): Y + H Y
1
<Y: X + H <X:
End(loop):
Display Y. [To see all iterations, include this statement in the loop.]
(i) H = 1, N = 1 i j(1) = 3
(ii) H = 0.1, N = 10 i j(1) E 2.3928
(iii) H = 0.01, N = 100 i j(1) E 2.3701
(iv) H = 0.001. N = 1000 i j(1) E 2.3681
(b) j = 2 +c
3i
3
i j
0
= 33r
2
c
3i
3
LHS = j
0
+ 3r
2
j = 33r
2
c
3i
3
+ 3r
2
2 +c
3i
3
= 33r
2
c
3i
3
+ 6r
2
+ 3r
2
c
3i
3
= 6r
2
= RHS
j(0) = 2 +c
30
= 2 + 1 = 3
386

(c) The exact value of j(1) is 2 +c
31
3
= 2 +c
31
.
(i) For / = 1: (exact value) 3(approximate value) = 2 +c
31
33 E 30.6321
(ii) For / = 0.1: (exact value) 3(approximate value) = 2 +c
31
32.3928 E 30.0249
(iii) For / = 0.01: (exact value) 3(approximate value) = 2 +c
31
32.3701 E 30.0022
(iv) For / = 0.001: (exact value) 3(approximate value) = 2 +c
31
32.3681 E 30.0002
In (ii)(iv), it seems that when the step size is divided by 10, the error estimate is also divided by 10 (approximately).
27. (a) 1
dQ
dt
+
1
C
Q = 1(t) becomes 5Q
0
+
1
0.05
Q = 60
or Q
0
+ 4Q = 12.
(b) From the graph, it appears that the limiting value of the
charge Q is about 3.
(c) If Q
0
= 0, then 4Q = 12 i Q = 3 is an
equilibrium solution.
(d)
(e) Q
0
+ 4Q = 12 i Q
0
= 12 34Q. Now Q(0) = 0, so t
0
= 0 and Q
0
= 0.
Q
1
= Q
0
+/1(t
0
. Q
0
) = 0 + 0.1(12 34 0) = 1.2
Q
2
= Q
1
+/1(t
1
. Q
1
) = 1.2 + 0.1(12 34 1.2) = 1.92
Q
3
= Q
2
+/1(t
2
. Q
2
) = 1.92 + 0.1(12 34 1.92) = 2.352
Q
4
= Q
3
+/1(t
3
. Q
3
) = 2.352 + 0.1(12 34 2.352) = 2.6112
Q
5
= Q
4
+/1(t
4
. Q
4
) = 2.6112 + 0.1(12 34 2.6112) = 2.76672
Thus, Q
5
= Q(0.5) E 2.77 C.
9.3 Separable Equations
1.
dj
dr
=
j
r
i
dj
j
=
dr
r
[j 6= 0] i

dj
j
=

dr
r
i ln|j| = ln|r| + C i
|j| = c
ln|i|+C
= c
ln|i|
c
C
= c
C
|r| i j = 1r, where 1 = c
C
is a constant. (In our derivation, 1 was nonzero,
but we can restore the excluded case j = 0 by allowing 1 to be zero.)
3. (r
2
+ 1)j
0
= rj i
dj
dr
=
rj
r
2
+ 1
i
dj
j
=
rdr
r
2
+ 1
[j 6= 0] i

dj
j
=

rdr
r
2
+ 1
i
ln|j| =
1
2
ln(r
2
+ 1) +C [& = r
2
+ 1, d& = 2rdr] = ln(r
2
+ 1)
1'2
+ lnc
C
= ln
c
C
I
r
2
+ 1
i
|j| = c
C
I
r
2
+ 1 i j = 1
I
r
2
+ 1, where 1 = c
C
is a constant. (In our derivation, 1 was nonzero, but we can
restore the excluded case j = 0 by allowing 1 to be zero.)
SECTION 9.3 SEPARABLE EQUATIONS

387
5. (1 + tanj) j
0
= r
2
+ 1 i (1 + tanj)
dj
dr
= r
2
+ 1 i

1 +
sinj
cos j
dj = (r
2
+ 1) dr i

1 3
3sinj
cos j
dj =

(r
2
+ 1) dr i j 3ln|cos j| =
1
3
r
3
+r +C.
Note: The left side is equivalent to j + ln|sec j|.
7.
dj
dt
=
tc
I
j
1 +j
2
i j
1 +j
2
dj = tc
I
dt i

j
1 +j
2
dj =

tc
I
dt i
1
3
1 +j
2
3'2
= tc
I
3c
I
+C
[where the rst integral is evaluated by substitution and the second by parts] i 1 +j
2
= [3(tc
I
3c
I
+C)]
2'3
i
j =
[3(tc
I
3c
I
+C)]
2'3
31
9.
d&
dt
= 2 + 2& +t +t& i
d&
dt
= (1 +&)(2 +t) i

d&
1 +&
=

(2 +t)dt [& 6= 31] i
ln|1 +&| =
1
2
t
2
+ 2t +C i |1 +&| = c
I
2
'2 +2I +C
= 1c
I
2
'2 +2I
, where 1 = c
C
i 1 +& = 1c
I
2
'2 +2I
i
& = 31 1c
I
2
'2 +2I
where 1 0. & = 31 is also a solution, so & = 31 +c
I
2
'2 +2I
, where is an arbitrary constant.
11.
dj
dr
=
r
j
i j dj = rdr i

j dj =

rdr i
1
2
j
2
=
1
2
r
2
+C. j(0) = 33 i
1
2
(33)
2
=
1
2
(0)
2
+C i C =
9
2
, so
1
2
j
2
=
1
2
r
2
+
9
2
i j
2
= r
2
+ 9 i j = 3
I
r
2
+ 9 since j(0) = 33 < 0.
13. rcos r = (2j +c
3u
) j
0
i rcos rdr = (2j +c
3u
) dj i

(2j +c
3u
) dj =

rcos rdr i
j
2
+
1
3
c
3u
= rsinr + cos r +C [where the second integral is evaluated using integration by parts].
Now j(0) = 0 i 0 +
1
3
= 0 + 1 +C i C = 3
2
3
. Thus, a solution is j
2
+
1
3
c
3u
= rsinr + cos r 3
2
3
.
We cannot solve explicitly for j.
15.
d&
dt
=
2t + sec
2
t
2&
, &(0) = 35.

2&d& =

2t + sec
2
t
dt i &
2
= t
2
+ tant +C,
where [&(0)]
2
= 0
2
+ tan0 +C i C = (35)
2
= 25. Therefore, &
2
= t
2
+ tant + 25, so & =
I
t
2
+ tant + 25.
Since &(0) = 35, we must have & = 3
I
t
2
+ tant + 25.
17. j
0
tanr = o +j, 0 < r < 2 i
dj
dr
=
o +j
tanr
i
dj
o +j
= cot rdr [o +j 6= 0] i
dj
o +j
=

cos r
sinr
dr i ln|o +j| = ln|sinr| +C i |o +j| = c
ln|sin i|+C
= c
ln|sin i|
c
C
= c
C
|sinr| i
o +j = 1sinr, where 1 = c
C
. (In our derivation, 1 was nonzero, but we can restore the excluded case
j = 3o by allowing 1 to be zero.) j(3) = o i o +o = 1sin
i 2o = 1
I
3
2
i 1 =
4o
I
3
.
Thus, o +j =
4o
I
3
sinr and so j =
4o
I
3
sinr 3o.
19. If the slope at the point (r. j) is rj, then we have
dj
dr
= rj i
dj
j
= rdr [j 6= 0] i

dj
j
=

rdr i
ln|j| =
1
2
r
2
+C. j(0) = 1 i ln1 = 0 +C i C = 0. Thus, |j| = c
i
2
'2
i j = c
i
2
'2
, so j = c
i
2
'2
since j(0) = 1 0. Note that j = 0 is not a solution because it doesnt satisfy the initial condition j(0) = 1.
388

21. & = r +j i
d
dr
(&) =
d
dr
(r +j) i
d&
dr
= 1 +
dj
dr
, but
dj
dr
= r +j = &, so
d&
dr
= 1 +& i
d&
1 +&
= dr [& 6= 31] i

d&
1 +&
=

dr i ln|1 +&| = r +C i |1 +&| = c
i+C
i
1 +& = c
C
c
i
i & = c
C
c
i
31 i r +j = c
C
c
i
31 i j = 1c
i
3r 31, where 1 = c
C
6= 0.
If & = 31, then 31 = r +j i j = 3r 31, which is just j = 1c
i
3r 31 with 1 = 0. Thus, the general solution
is j = 1c
i
3r 31, where 1 M R.
23. (a) j
0
= 2r
1 3j
2
i
dj
dr
= 2r
1 3j
2
i
dj
1 3j
2
= 2rdr i

dj
1 3j
2
=

2rdr i
sin
31
j = r
2
+C for 3
r
2
$ r
2
+C $
r
2
.
(b) j(0) = 0 i sin
31
0 = 0
2
+C i C = 0,
so sin
31
j = r
2
and j = sin
r
2
for
3
2 $ r $
2.
(c) For
1 3j
2
to be a real number, we must have 31 $ j $ 1; that is, 31 $ j(0) $ 1. Thus, the initial-value problem
j
0
= 2r
1 3j
2
, j(0) = 2 does not have a solution.
25.
dj
dr
=
sinr
sinj
, j(0) =

2
. So
sinj dj =

sinrdr C
3cos j = 3cos r +C C cos j = cos r 3C. From the initial condition,
we need cos
r
2
= cos 0 3C i 0 = 1 3C i C = 1, so the solution is
cos j = cos r 31. Note that we cannot take cos
31
of both sides, since that would
unnecessarily restrict the solution to the case where 31 $ cos r 31 C 0 $ cos r,
as cos
31
is dened only on [31. 1]. Instead we plot the graph using Maples
plots[implicitplot] or Mathematicas Plot[Evaluate[ ]].
27. (a)
r j j
0
= 1j
0 0.5 2
0 30.5 32
0 1 1
0 31 31
0 2 0.5
r j j
0
= 1j
0 32 30.5
0 4 0.25
0 3 0.3
0 0.25 4
0 0.3 3

389
(b) j
0
= 1j i djdr = 1j i
j dj = dr i

j dj =

dr i
1
2
j
2
= r +C i
j
2
= 2(r +C) or j =
2(r +C).
(c)
29. The curves r
2
+ 2j
2
= I
2
form a family of ellipses with major axis on the r-axis. Differentiating gives
d
dr
(r
2
+ 2j
2
) =
d
dr
(I
2
) i 2r + 4jj
0
= 0 i 4jj
0
= 32r i j
0
=
3r
2j
. Thus, the slope of the tangent line
at any point (r. j) on one of the ellipses is j
0
=
3r
2j
, so the orthogonal trajectories
must satisfy j
0
=
2j
r
C
dj
dr
=
2j
r
C
dj
j
= 2 =
dr
r
C
dj
j
= 2
dr
r
C ln|j| = 2 ln|r| +C
1
C ln|j| = ln|r|
2
+C
1
C
|j| = c
ln i
2
+C
1
C j = r
2
c
C
1
= Cr
2
. This is a family of parabolas.
31. The curves j = Ir form a family of hyperbolas with asymptotes r = 0 and j = 0. Differentiating gives
d
dr
(j) =
d
dr
I
r
i j
0
= 3
I
r
2
i j
0
= 3
rj
r
2
[since j = Ir i rj = I] i j
0
= 3
j
r
. Thus, the slope
of the tangent line at any point (r. j) on one of the hyperbolas is j
0
= 3jr,
so the orthogonal trajectories must satisfy j
0
= rj C
dj
dr
=
r
j
C
j dj = rdr C

j dj =

rdr C
1
2
j
2
=
1
2
r
2
+C
1
C
j
2
= r
2
+C
2
C r
2
3j
2
= C. This is a family of hyperbolas with
asymptotes j = r.
33. From Exercise 9.2.27,
dQ
dt
= 12 34Q C

dQ
12 34Q
=

dt C 3
1
4
ln|12 34Q| = t +C C
ln|12 34Q| = 34t 34C C |12 34Q| = c
34I34C
C 12 34Q = 1c
34I
[1 = c
34C
] C
4Q = 12 31c
34I
C Q = 3 3c
34I
[ = 14]. Q(0) = 0 C 0 = 3 3 C = 3 C
Q(t) = 3 33c
34I
. As t <", Q(t) <3 30 = 3 (the limiting value).
35.
d1
dt
= I(A 31) C

d1
1 3A
=

(3I) dt C ln|1 3A| = 3It +C C |1 3A| = c
3II+C
C
1 3A = c
3II
[ = c
C
] C 1 = A +c
3II
. If we assume that performance is at level 0 when t = 0, then
1(0) = 0 C 0 = A + C = 3A C 1(t) = A 3Ac
3II
. lim
I<"
1(t) = A 3A 0 = A.
390

37. (a) If o = /, then
dr
dt
= I(o 3r)(/ 3r)
1'2
becomes
dr
dt
= I(o 3r)
3'2
i (o 3r)
33'2
dr = I dt i
(o 3r)
33'2
dr =

I dt i 2(o 3r)
31'2
= It +C [by substitution] i
2
It +C
=
I
o 3r i
2
It +C
2
= o 3r i r(t) = o 3
4
(It +C)
2
. The initial concentration of HBr is 0, so r(0) = 0 i
0 = o 3
4
C
2
i
4
C
2
= o i C
2
=
4
o
i C = 2
I
o [C is positive since It +C = 2(o 3r)
31'2
0].
Thus, r(t) = o 3
4
(It + 2
I
o )
2
.
(b)
dr
dt
= I(o 3r)(/ 3r)
1'2
i
dr
(o 3r)
I
/ 3r
= I dt i

dr
(o 3r)
I
/ 3r
=

I dt (-).
From the hint, & =
I
/ 3r i &
2
= / 3r i 2&d& = 3dr, so
dr
(o 3r)
I
/ 3r
=

32&d&
[o 3(/ 3&
2
)]&
= 32
d&
o 3/ +&
2
= 32
d&
I
o 3/
2
+&
2
17
= 32
1
I
o 3/
tan
31
&
I
o 3/
So (-) becomes
32
I
o 3/
tan
31
I
/ 3r
I
o 3/
= It +C. Now r(0) = 0 i C =
32
I
o 3/
tan
31
I
/
I
o 3/
and we have
32
I
o 3/
tan
31
I
/ 3r
I
o 3/
= It 3
2
I
o 3/
tan
31
I
/
I
o 3/
i
2
I
o 3/
tan
31
/
o 3/
3tan
31
/ 3r
o 3/
= It i
t(r) =
2
I
I
o 3/
tan
31
/
o 3/
3tan
31
/ 3r
o 3/
.
39. (a)
dC
dt
= v 3IC i
dC
dt
= 3(IC 3v) i

dC
IC 3v
=

3dt i (1I) ln|IC 3v| = 3t +A
1
i
ln|IC 3v| = 3It +A
2
i |IC 3v| = c
3II+L
2
i IC 3v = A
3
c
3II
i IC = A
3
c
3II
+v i
C(t) = A
4
c
3II
+vI. C(0) = C
0
i C
0
= A
4
+vI i A
4
= C
0
3vI i
C(t) = (C
0
3vI)c
3II
+vI.
(b) If C
0
< vI, then C
0
3vI < 0 and the formula for C(t) shows that C(t) increases and lim
I<"
C(t) = vI.
As t increases, the formula for C(t) shows how the role of C
0
steadily diminishes as that of vI increases.
41. (a) Let j(t) be the amount of salt (in kg) after t minutes. Then j(0) = 15. The amount of liquid in the tank is 1000 L at all
times, so the concentration at time t (in minutes) is j(t)1000 kgL and
dj
dt
= 3
j(t)
1000
kg
L
10
L
min
= 3
j(t)
100
kg
min
.
dj
j
= 3
1
100
dt i lnj = 3
t
100
+C, and j(0) = 15 i ln15 = C, so lnj = ln15 3
t
100
.
It follows that ln
j
15
= 3
t
100
and
j
15
= c
3I'100
, so j = 15c
3I'100
kg.
(b) After 20 minutes, j = 15c
320'100
= 15c
30.2
E 12.3 kg.

391
43. Let j(t) be the amount of alcohol in the vat after t minutes. Then j(0) = 0.04(500) = 20 gal. The amount of beer in the vat
is 500 gallons at all times, so the percentage at time t (in minutes) is j(t)500 100, and the change in the amount of alcohol
with respect to time t is
dj
dt
= rate in 3 rate out = 0.06
5
gal
min
3
j(t)
500
5
gal
min
= 0.3 3
j
100
=
30 3j
100
gal
min
.
Hence,
dj
30 3j
=

dt
100
and 3ln|30 3j| =
1
100
t +C. Because j(0) = 20, we have 3ln10 = C, so
3ln|30 3j| =
1
100
t 3ln10 i ln|30 3j| = 3t100 + ln10 i ln|30 3j| = lnc
3I'100
+ ln10 i
ln|30 3j| = ln(10c
3I'100
) i |30 3j| = 10c
3I'100
. Since j is continuous, j(0) = 20, and the right-hand side is
never zero, we deduce that 30 3j is always positive. Thus, 30 3j = 10c
3I'100
i j = 30 310c
3I'100
. The
percentage of alcohol is j(t) = j(t)500 100 = j(t)5 = 6 32c
3I'100
. The percentage of alcohol after one hour is
j(60) = 6 32c
360'100
E 4.9.
45. Assume that the raindrop begins at rest, so that (0) = 0. didt = Ii and (i)
0
= oi i i
0
+i
0
= oi i
i
0
+(Ii) = oi i
0
+I = o i
d
dt
= o 3I i

d
o 3I
=

dt i
3(1I) ln|o 3I| = t +C i ln|o 3I| = 3It 3IC i o 3I = c
3II
. (0) = 0 i = o.
So I = o 3oc
3II
i = (oI)(1 3c
3II
). Since I 0, as t <", c
3II
<0 and therefore, lim
I<"
(t) = oI.
47. (a) The rate of growth of the area is jointly proportional to
(t) and A 3(t); that is, the rate is proportional to the

product of those two quantities. So for some constant I, ddt = I
I
(A 3). We are interested in the maximum of
the function ddt (when the tissue grows the fastest), so we differentiate, using the Chain Rule and then substituting for
ddt from the differential equation:
d
dt
d
dt
= I
I
(31)
d
dt
+ (A 3)
1
2
31'2
d
dt
=
1
2
I
31'2
d
dt
[32+ (A 3)]
=
1
2
I
31'2
I
I
(A 3)
[A 33] =
1
2
I
2
(A 3)(A 33)
This is 0 when A 3 = 0 [this situation never actually occurs, since the graph of (t) is asymptotic to the line j = A,
as in the logistic model] and when A 33 = 0 C (t) = A3. This represents a maximum by the First Derivative
Test, since
d
dt
d
dt
goes from positive to negative when (t) = A3.

(b) From the CAS, we get (t) = A
Cc
I
LII
31
Cc
I
LII
+ 1
2
. To get C in terms of the initial area
0
and the maximum area A,
we substitute t = 0 and =
0
= (0):
0
= A
C 31
C + 1
2
C (C + 1)
I
0
= (C 31)
I
A C
C
I
0
+
I
0
= C
I
A 3
I
A C
I
A +
I
0
= C
I
A 3 C
I
0
C
I
A +
I
0
= C
I
A 3
I
C C =
I
A +
I
0
I
A 3
I
0
. [Notice that if
0
= 0, then C = 1.]
392

9.4 Models for Population Growth
1. (a) d1dt = 0.051 30.00051
2
= 0.051(1 30.011) = 0.051(1 31100). Comparing to Equation 4,
d1dt = I1(1 311), we see that the carrying capacity is 1 = 100 and the value of I is 0.05.
(b) The slopes close to 0 occur where 1 is near 0 or 100. The largest slopes appear to be on the line 1 = 50. The solutions
are increasing for 0 < 1
0
< 100 and decreasing for 1
0
100.
(c) All of the solutions approach 1 = 100 as t increases. As in
part (b), the solutions differ since for 0 < 1
0
< 100 they are
increasing, and for 1
0
100 they are decreasing. Also, some
have an IP and some dont. It appears that the solutions which
have 1
0
= 20 and 1
0
= 40 have inection points at 1 = 50.
(d) The equilibrium solutions are 1 = 0 (trivial solution) and 1 = 100. The increasing solutions move away from 1 = 0 and
all nonzero solutions approach 1 = 100 as t <".
3. (a)
dj
dt
= Ij
1 3
j
1
i j(t) =
1
1 +c
3II
with =
1 3j(0)
j(0)
. With 1 = 8 10
7
, I = 0.71, and
j(0) = 2 10
7
, we get the model j(t) =
8 10
7
1 + 3c
30.71I
, so j(1) =
8 10
7
1 + 3c
30.71
E 3.23 10
7
kg.
(b) j(t) = 4 10
7
i
8 10
7
1 + 3c
30.71I
= 4 10
7
i 2 = 1 + 3c
30.71I
i c
30.71I
=
1
3
i
30.71t = ln
1
3
i t =
ln3
0.71
E 1.55 years
5. (a) We will assume that the difference in the birth and death rates is 20 millionyear. Let t = 0 correspond to the year 1990
and use a unit of 1 billion for all calculations. I E
1
1
d1
dt
=
1
5.3
(0.02) =
1
265
, so
d1
dt
= I1
1 3
1
1
=
1
265
1
1 3
1
100
, 1 in billions
(b) =
1 31
0
1
0
=
100 35.3
5.3
=
947
53
E 17.8679. 1(t) =
1
1 +c
3II
=
100
1 +
947
53
c
3(1'265)I
, so 1(10) E 5.49 billion.
(c) 1(110) E 7.81, and 1(510) E 27.72. The predictions are 7.81 billion in the year 2100 and 27.72 billion in 2500.
(d) If 1 = 50, then 1(t) =
50
1 +
447
53
c
3(1'265)I
. So 1(10) E 5.48, 1(110) E 7.61, and 1(510) E 22.41. The predictions
become 5.48 billion in the year 2000, 7.61 billion in 2100, and 22.41 billion in the year 2500.
7. (a) Our assumption is that
dj
dt
= Ij(1 3j), where j is the fraction of the population that has heard the rumor.
SECTION 9.4 MODELS FOR POPULATION GROWTH

393
(b) Using the logistic equation (4),
d1
dt
= I1
1 3
1
1
, we substitute j =
1
1
, 1 = 1j, and
d1
dt
= 1
dj
dt
,
to obtain 1
dj
dt
= I(1j)(1 3j) C
dj
dt
= Ij(1 3j), our equation in part (a).
Now the solution to (4) is 1(t) =
1
1 +c
3II
, where =
1 31
0
1
0
.
We use the same substitution to obtain 1j =
1
1 +
1 31j
0
1j
0
c
3II
i j =
j
0
j
0
+ (1 3j
0
)c
3II
.
Alternatively, we could use the same steps as outlined in the solution of Equation 4.
(c) Let t be the number of hours since 8 AM. Then j
0
= j(0) =
80
1000
= 0.08 and j(4) =
1
2
, so
1
2
= j(4) =
0.08
0.08 + 0.92c
34I
. Thus, 0.08 + 0.92c
34I
= 0.16, c
34I
=
0.08
0.92
=
2
23
, and c
3I
=

2
23
1'4
,
so j =
0.08
0.08 + 0.92(223)
I'4
=
2
2 + 23(223)
I'4
. Solving this equation for t, we get
2j + 23j
2
23
I'4
= 2 i

2
23
I'4
=
2 32j
23j
i

2
23
I'4
=
2
23

1 3j
j
i

2
23
I'431
=
1 3j
j
.
It follows that
t
4
31 =
ln[(1 3j)j]
ln
2
23
, so t = 4
1 +
ln((1 3j)j)
ln
2
23
.
When j = 0.9,
1 3j
j
=
1
9
, so t = 4
1 3
ln9
ln
2
23
E 7.6 h or 7 h 36 min. Thus, 90% of the population will have heard

the rumor by 3:36 PM.
9. (a)
d1
dt
= I1
1 3
1
1
i
d
2
1
dt
2
= I
3
1
1
d1
dt
1 3
1
1
d1
dt
= I
d1
dt
3
1
1
+ 1 3
1
1
= I
I1
1 3
1
1
1 3
21
1
= I
2
1
1 3
1
1
1 3
21
1
(b) 1 grows fastest when 1

0
has a maximum, that is, when 1
00
= 0. From part (a), 1
00
= 0 C 1 = 0, 1 = 1,
or 1 = 12. Since 0 < 1 < 1, we see that 1
00
= 0 C 1 = 12.
11. Following the hint, we choose t = 0 to correspond to 1960 and subtract
94,000 from each of the population gures. We then use a calculator to
obtain the models and add 94,000 to get the exponential function
1
T
(t) = 1578.3(1.0933)
I
+ 94,000 and the logistic function
1
1
(t) =
32,658.5
1 + 12.75c
30.1706I
+ 94,000. 1
1
is a reasonably accurate
model, while 1
T
is not, since an exponential model would only be used
for the rst few data points.
13. (a)
d1
dt
= I1 3i = I
1 3
i
I
. Let j = 1 3
i
I
, so
dj
dt
=
d1
dt
and the differential equation becomes
dj
dt
= Ij.
The solution is j = j
0
c
II
i 1 3
i
I
=

1
0
3
i
I
c
II
i 1(t) =
i
I
+
1
0
3
i
I
c
II
.
394

(b) Since I 0, there will be an exponential expansion C 1
0
3
i
I
0 C i < I1
0
.
(c) The population will be constant if 1
0
3
i
I
= 0 C i = I1
0
. It will decline if 1
0
3
i
I
< 0 C i I1
0
.
(d) 1
0
= 8,000,000, I = c 3o = 0.016, i = 210,000 i i I1
0
(= 128,000), so by part (c), the population was
declining.
15. (a) The term 315 represents a harvesting of sh at a constant ratein this case, 15 shweek. This is the rate at which sh
are caught.
(b) (c) From the graph in part (b), it appears that 1(t) = 250 and 1(t) = 750
are the equilibrium solutions. We conrm this analytically by solving the
equation d1dt = 0 as follows: 0.081(1 311000) 315 = 0 i
0.081 30.000081
2
315 = 0 i
30.00008(1
2
310001 + 187,500) = 0 i
(1 3250)(1 3750) = 0 i 1 = 250 or 750.
(d) For 0 < 1
0
< 250, 1(t) decreases to 0. For 1
0
= 250, 1(t) remains
constant. For 250 < 1
0
< 750, 1(t) increases and approaches 750.
For 1
0
= 750, 1(t) remains constant. For 1
0
750, 1(t) decreases
and approaches 750.
(e)
d1
dt
= 0.081
1 3
1
1000
315 C 3
100,000
8

d1
dt
= (0.081 30.000081
2
315)
3
100,000
8
C
312,500
d1
dt
= 1
2
310001 + 187,500 C
d1
(1 3250)(1 3750)
= 3
1
12,500
dt C

31500
1 3250
+
1500
1 3750
d1 = 3
1
12,500
dt C

1
1 3250
3
1
1 3750
d1 =
1
25
dt C
ln|1 3250| 3ln|1 3750| =
1
25
t +C C ln
1 3250
1 3750
=
1
25
t +C C
1 3250
1 3750
= c
I'25+C
= Ic
I'25
C
1 3250
1 3750
= Ic
I'25
C 1 3250 = 1Ic
I'25
3750Ic
I'25
C 1 31Ic
I'25
= 250 3750Ic
I'25
C
1(t) =
250 3750Ic
I'25
1 3Ic
I'25
. If t = 0 and 1 = 200, then 200 =
250 3750I
1 3I
C 200 3200I = 250 3750I C
550I = 50 C I =
1
11
. Similarly, if t = 0 and 1 = 300, then
I = 3
1
9
. Simplifying 1 with these two values of I gives us
1(t) =
250(3c
I'25
311)
c
I'25
311
and 1(t) =
750(c
I'25
+ 3)
c
I'25
+ 9
.
SECTION 9.4 MODELS FOR POPULATION GROWTH

395
17. (a)
d1
dt
= (I1)
1 3
1
1
1 3
i
1
. If i < 1 < 1, then d1dt = (+)(+)(+) = + i 1 is increasing.

If 0 < 1 < i, then d1dt = (+)(+)(3) = 3 i 1 is decreasing.
(b) I = 0.08, 1 = 1000, and i = 200 i
d1
dt
= 0.081
1 3
1
1000
1 3
200
1
For 0 < 1
0
< 200, the population dies out. For 1
0
= 200, the population
is steady. For 200 < 1
0
< 1000, the population increases and approaches
1000. For 1
0
1000, the population decreases and approaches 1000.
The equilibrium solutions are 1(t) = 200 and 1(t) = 1000.
(c)
d1
dt
= I1
1 3
1
1
1 3
i
1
= I1
1 31
1
1 3i
1
=
I
1
(1 31)(1 3i) C
d1
(1 31)(1 3i)
=

I
1
dt. By partial fractions,
1
(1 31)(1 3i)
=

1 31
+
1
1 3i
, so
(1 3i) +1(1 31) = 1.
If 1 = i, 1 =
1
1 3i
; if 1 = 1, =
1
1 3i
, so
1
1 3i

1
1 31
+
1
1 3i
d1 =

I
1
dt i
1
1 3i
(3ln|1 31| + ln|1 3i|) =
I
1
t +A i
1
1 3i
ln
1 3i
1 31
=
I
1
t +A i
ln
1 3i
1 31
= (1 3i)
I
1
t +A
1
C
1 3i
1 31
= 1c
(13r)(I'1)I
[1 = c
L
1
].
Let t = 0:
1
0
3i
1 31
0
= 1. So
1 3i
1 31
=
1
0
3i
1 31
0
c
(13r)(I'1)I
. Solving for 1, we get
1(t) =
i(1 31
0
) +1(1
0
3i)c
(13r)(I'1)I
1 31
0
+ (1
0
3i)c
(13r)(I'1)I
.
(d) If 1
0
< i, then 1
0
3i < 0. Let `(t) be the numerator of the expression for 1(t) in part (c). Then
`(0) = 1
0
(1 3i) 0, and 1
0
3i < 0 C lim
I<"
1(1
0
3i)c
(13r)(I'1)I
= 3" i lim
I<"
`(t) = 3".
Since ` is continuous, there is a number t such that `(t) = 0 and thus 1(t) = 0. So the species will become extinct.
19. (a) d1dt = I1 cos(vt 3c) i (d1)1 = I cos(vt 3c) dt i

(d1)1 = I
cos(vt 3c) dt i
ln1 = (Iv) sin(vt 3c) +C. (Since this is a growth model, 1 0 and we can write ln1 instead of ln|1|.) Since
1(0) = 1
0
, we obtain ln1
0
= (Iv) sin(3c) +C = 3(Iv) sinc +C i C = ln1
0
+ (Iv) sinc. Thus,
ln1 = (Iv) sin(vt 3c) + ln1
0
+ (Iv) sinc, which we can rewrite as ln(11
0
) = (Iv)[sin(vt 3c) + sinc] or,
after exponentiation, 1(t) = 1
0
c
(I'r)[sin(rI3c)+sin c]
.
396

(b) As I increases, the amplitude
increases, but the minimum value
stays the same.
As v increases, the amplitude and
the period decrease.
A change in c produces slight
adjustments in the phase shift and
amplitude.
1(t) oscillates between 1
0
c
(I'r)(1+sin c)
and 1
0
c
(I'r)(31+sin c)
(the extreme values are attained when vt 3c is an odd
multiple of
r
2
), so lim
I<"
1(t) does not exist.
21. By Equation (7), 1(t) =
1
1 +c
3II
. By comparison, if c = (ln)I and & =
1
2
I(t 3c), then
1 + tanh& = 1 +
c
u
3c
3u
c
u
+c
3u
=
c
u
+c
3u
c
u
+c
3u
+
c
u
3c
3u
c
u
+c
3u
=
2c
u
c
u
+c
3u

c
3u
c
3u
=
2
1 +c
32u
and c
32u
= c
3I(I3c)
= c
Ic
c
3II
= c
ln /
c
3II
= c
3II
, so
1
2
1
1 + tanh
1
2
I(t 3c)
=
1
2
[1 + tanh&] =
1
2

2
1 +c
32u
=
1
1 +c
32u
=
1
1 +c
3II
= 1(t).
9.5 Linear Equations
1. j
0
+ cos r = j i j
0
+ (31)j = 3cos r is linear since it can be put into the standard linear form (1),
j
0
+1(r) j = Q(r).
3. jj
0
+rj = r
2
i j
0
+r = r
2
j i j
0
3r
2
j = 3r is not linear since it cannot be put into the standard linear
form (1), j
0
+1(r) j = Q(r).
5. Comparing the given equation, j
0
+ 2j = 2c
i
, with the general form, j
0
+1(r)j = Q(r), we see that 1(r) = 2 and the
integrating factor is 1(r) = c
U
T(i)ui
= c
U
2 ui
= c
2i
. Multiplying the differential equation by 1(r) gives
c
2i
j
0
+ 2c
2i
j = 2c
3i
i (c
2i
j)
0
= 2c
3i
i c
2i
j =

2c
3i
dr i c
2i
j =
2
3
c
3i
+C i j =
2
3
c
i
+Cc
32i
.
7. rj
0
32j = r
2
[divide by r] i j
0
+
3
2
r
j = r (-).
1(r) = c
U
T(i) ui
= c
U
(32'i) ui
= c
32 ln|i|
= c
ln|i|
2
= c
ln(1'i
2
)
= 1r
2
. Multiplying the differential equation (-)
by 1(r) gives
1
r
2
j
0
3
2
r
3
j =
1
r
i

1
r
2
j
0
=
1
r
i
1
r
2
j = ln|r| +C i
j = r
2
(ln|r| +C ) = r
2
ln|r| +Cr
2
.
SECTION 9.5 LINEAR EQUATIONS

397
9. Since 1(r) is the derivative of the coefcient of j
0
[1(r) = 1 and the coefcient is r], we can write the differential equation
rj
0
+j =
I
r in the easily integrable form (rj)
0
=
I
r i rj =
2
3
r
3'2
+C i j =
2
3
I
r +Cr.
11. sinr
dj
dr
+(cos r) j = sin(r
2
) i [(sinr) j]
0
= sin(r
2
) i (sinr) j =

sin(r
2
) dr i j =
sin(r
2
) dr +C
sinr
.
13. (1 +t)
d&
dt
+& = 1 +t, t 0 [divide by 1 +t] i
d&
dt
+
1
1 +t
& = 1 (-), which has the
form &
0
+1(t) & = Q(t). The integrating factor is 1(t) = c
U
T(I) uI
= c
U
[1'(1+I)] uI
= c
ln(1+I)
= 1 +t.
Multiplying (-) by 1(t) gives us our original equation back. We rewrite it as [(1 +t)&]
0
= 1 +t. Thus,
(1 +t)& =

(1 +t) dt = t +
1
2
t
2
+C i & =
t +
1
2
t
2
+C
1 +t
or & =
t
2
+ 2t + 2C
2(t + 1)
.
15. j
0
= r +j i j
0
+ (31)j = r. 1(r) = c
U
(31) ui
= c
3i
. Multiplying by c
3i
gives c
3i
j
0
3c
3i
j = rc
3i
i
(c
3i
j)
0
= rc
3i
i c
3i
j =

rc
3i
dr = 3rc
3i
3c
3i
+C [integration by parts with & = r, d = c
3i
dr] i
j = 3r 31 +Cc
i
. j(0) = 2 i 31 +C = 2 i C = 3, so j = 3r 31 + 3c
i
.
17.
d
dt
32t = 3t
2
c
I
2
, (0) = 5. 1(t) = c
U
(32I) uI
= c
3I
2
. Multiply the differential equation by 1(t) to get
c
3I
2 d
dt
32tc
3I
2
= 3t
2
i

c
3I
2
0
= 3t
2
i c
3I
2
=

3t
2
dt = t
3
+C i = t
3
c
I
2
+Cc
I
2
.
5 = (0) = 0 1 +C 1 = C, so = t
3
c
I
2
+ 5c
I
2
.
19. rj
0
= j +r
2
sinr i j
0
3
1
r
j = rsinr. 1(r) = c
U
(31'i) ui
= c
3ln i
= c
ln i
1
=
1
r
.
Multiplying by
1
r
gives
1
r
j
0
3
1
r
2
j = sinr i

1
r
j
0
= sinr i
1
r
j = 3cos r +C i j = 3rcos r +Cr.
j() = 0 i 3 (31) +C = 0 i C = 31, so j = 3rcos r 3r.
21. rj
0
+ 2j = c
i
i j
0
+
2
r
j =
c
i
r
.
1(r) = c
U
(2'i) ui
= c
2 ln|i|
=

c
ln|i|
2
= |r|
2
= r
2
.
Multiplying by 1(r) gives r
2
j
0
+ 2rj = rc
i
i (r
2
j)
0
= rc
i
i
r
2
j =

rc
i
dr = (r 31)c
i
+C [by parts] i
j = [(r 31)c
i
+C]r
2
. The graphs for C = 35, 33, 31, 1, 3, 5, and 7 are
shown. C = 1 is a transitional value. For C < 1, there is an inection point and
for C 1, there is a local minimum. As |C| gets larger, the branches get
further from the origin.
398

23. Setting & = j
13n
,
d&
dr
= (1 3n) j
3n
dj
dr
or
dj
dr
=
j
n
1 3n
d&
dr
=
&
n'(13n)
1 3n
d&
dr
. Then the Bernoulli differential equation
becomes
&
n'(13n)
1 3n
d&
dr
+1(r)&
1'(13n)
= Q(r)&
n'(13n)
or
d&
dr
+ (1 3n)1(r)& = Q(r)(1 3n).
25. Here j
0
+
2
r
j =
j
3
r
2
, so n = 3, 1(r) =
2
r
and Q(r) =
1
r
2
. Setting & = j
32
, & satises &
0
3
4&
r
= 3
2
r
2
.
Then 1(r) = c
U
(34'i) ui
= r
34
and & = r
4
3
2
r
6
dr +C
= r
4
2
5r
5
+C
= Cr
4
+
2
5r
.
Thus, j =
Cr
4
+
2
5r
31'2
.
27. (a) 2
d1
dt
+ 101 = 40 or
d1
dt
+ 51 = 20. Then the integrating factor is c
U
5 uI
= c
5I
. Multiplying the differential equation
by the integrating factor gives c
5I
d1
dt
+ 51c
5I
= 20c
5I
i (c
5I
1)
0
= 20c
5I
i
1(t) = c
35I
20c
5I
dt +C
= 4 +Cc
35I
. But 0 = 1(0) = 4 +C, so 1(t) = 4 34c
35I
.
(b) 1(0.1) = 4 34c
30.5
E 1.57 A
29. 5
dQ
dt
+ 20Q = 60 with Q(0) = 0 C. Then the integrating factor is c
U
4 uI
= c
4I
, and multiplying the differential
equation by the integrating factor gives c
4I
dQ
dt
+ 4c
4I
Q = 12c
4I
i (c
4I
Q)
0
= 12c
4I
i
Q(t) = c
34I
12c
4I
dt +C
= 3 +Cc
34I
. But 0 = Q(0) = 3 +C so Q(t) = 3(1 3c
34I
) is the charge at time t
and 1 = dQdt = 12c
34I
is the current at time t.
31.
d1
dt
+I1 = IA, so 1(t) = c
U
I uI
= c
II
. Multiplying the differential equation
by 1(t) gives c
II
d1
dt
+I1c
II
= IAc
II
i (c
II
1)
0
= IAc
II
i
1(t) = c
3II
IAc
II
dt +C
= A +Cc
3II
, I 0. Furthermore, it is
reasonable to assume that 0 $ 1(0) $ A, so 3A $ C $ 0.
33. j(0) = 0 kg. Salt is added at a rate of
0.4
kg
L
5
L
min
= 2
kg
min
. Since solution is drained from the tank at a rate of
3 Lmin, but salt solution is added at a rate of 5 Lmin, the tank, which starts out with 100 L of water, contains (100 + 2t) L
of liquid after t min. Thus, the salt concentration at time t is
j(t)
100 + 2t
kg
L
. Salt therefore leaves the tank at a rate of
j(t)
100 + 2t
kg
L
3
L
min
=
3j
100 + 2t
kg
min
. Combining the rates at which salt enters and leaves the tank, we get
SECTION 9.6 PREDATOR-PREY SYSTEMS

399
dj
dt
= 2 3
3j
100 + 2t
. Rewriting this equation as
dj
dt
+
3
100 + 2t
j = 2, we see that it is linear.

1(t) = exp
3 dt
100 + 2t
= exp
3
2
ln(100 + 2t)
= (100 + 2t)
3'2
Multiplying the differential equation by 1(t) gives (100 + 2t)
3'2
dj
dt
+ 3(100 + 2t)
1'2
j = 2(100 + 2t)
3'2
i
[(100 + 2t)
3'2
j]
0
= 2(100 + 2t)
3'2
i (100 + 2t)
3'2
j =
2
5
(100 + 2t)
5'2
+C i
j =
2
5
(100 + 2t) +C(100 + 2t)
33'2
. Now 0 = j(0) =
2
5
(100) +C 100
33'2
= 40 +
1
1000
C i C = 340,000, so
j =

2
5
(100 + 2t) 340,000(100 + 2t)
33'2
kg. From this solution (no pun intended), we calculate the salt concentration
at time t to be C(t) =
j(t)
100 + 2t
=
340,000
(100 + 2t)
5'2
+
2
5
kg
L
. In particular, C(20) =
340,000
140
5'2
+
2
5
E 0.2275
kg
L
and j(20) =
2
5
(140) 340,000(140)
33'2
E 31.85 kg.
35. (a)
d
dt
+
c
i
= o and 1(t) = c
U
(c'r) uI
= c
(c'r)I
, and multiplying the differential equation by
1(t) gives c
(c'r)I
d
dt
+
cc
(c'r)I
i
= oc
(c'r)I
i

c
(c'r)I
0
= oc
(c'r)I
. Hence,
(t) = c
3(c'r)I
oc
(c'r)I
dt +1
= ioc +1c
3(c'r)I
. But the object is dropped from rest, so (0) = 0 and
1 = 3ioc. Thus, the velocity at time t is (t) = (ioc)[1 3c
3(c'r)I
].
(b) lim
I<"
(t) = ioc
(c) c(t) =
(t) dt = (ioc)[t + (ic)c

3(c'r)I
] +c
1
where c
1
= c(0) 3i
2
oc
2
.
c(0) is the initial position, so c(0) = 0 and c(t) = (ioc)[t + (ic)c
3(c'r)I
] 3i
2
oc
2
.
9.6 Predator-Prey Systems
1. (a) drdt = 30.05r + 0.0001rj. If j = 0, we have drdt = 30.05r, which indicates that in the absence of j, r declines at
a rate proportional to itself. So r represents the predator population and j represents the prey population. The growth of
the prey population, 0.1j (from djdt = 0.1j 30.005rj), is restricted only by encounters with predators (the term
30.005rj). The predator population increases only through the term 0.0001rj; that is, by encounters with the prey and
not through additional food sources.
(b) djdt = 30.015j + 0.00008rj. If r = 0, we have djdt = 30.015j, which indicates that in the absence of r, j would
decline at a rate proportional to itself. So j represents the predator population and r represents the prey population. The
growth of the prey population, 0.2r (from drdt = 0.2r 30.0002r
2
30.006rj = 0.2r(1 30.001r) 30.006rj), is
restricted by a carrying capacity of 1000 [from the term 1 30.001r = 1 3r1000] and by encounters with predators (the
term 30.006rj). The predator population increases only through the term 0.00008rj; that is, by encounters with the prey
and not through additional food sources.
400

3. (a) At t = 0, there are about 300 rabbits and 100 foxes. At t = t
1
, the number
of foxes reaches a minimum of about 20 while the number of rabbits is
about 1000. At t = t
2
, the number of rabbits reaches a maximum of about
2400, while the number of foxes rebounds to 100. At t = t
3
, the number of
rabbits decreases to about 1000 and the number of foxes reaches a
maximum of about 315. As t increases, the number of foxes decreases
greatly to 100, and the number of rabbits decreases to 300 (the initial
populations), and the cycle starts again.
(b)
5.
7.
dW
d1
=
30.02W + 0.000021W
0.08130.0011W
C (0.08 30.001W)1dW = (30.02 + 0.000021)W d1 C
0.08 30.001W
W
dW =
30.02 + 0.000021
1
d1 C

0.08
W
30.001
dW =

3
0.02
1
+ 0.00002
d1 C
0.08 ln|W| 30.001W = 30.02 ln|1| + 0.000021+1 C 0.08 lnW + 0.02 ln1 = 0.001W + 0.000021+1 C
ln
W
0.08
1
0.02
= 0.000021 + 0.001W +1 C W
0.08
1
0.02
= c
0.00002T+0.001V+1
C
1
0.02
W
0.08
= Cc
0.00002T
c
0.001V
C
1
0.02
W
0.08
c
0.00002T
c
0.001V
= C. In general, if
dj
dr
=
3vj +/rj
Ir 3orj
, then C =
r
r
j
I
c
li
c
au
.
9. (a) Letting W = 0 gives us d1dt = 0.081(1 30.00021). d1dt = 0 C 1 = 0 or 5000. Since d1dt 0 for
0 < 1 < 5000, we would expect the rabbit population to increase to 5000 for these values of 1. Since d1dt < 0 for
1 5000, we would expect the rabbit population to decrease to 5000 for these values of 1. Hence, in the absence of
wolves, we would expect the rabbit population to stabilize at 5000.
(b) 1 and W are constant i 1
0
= 0 and W
0
= 0 i
0 = 0.081(1 30.00021) 30.0011W

0 = 30.02W + 0.000021W
0 = 1[0.08(1 30.00021) 30.001W]

0 = W(30.02 + 0.000021)
The second equation is true if W = 0 or 1 =
0.02
0.00002
= 1000. If W = 0 in the rst equation, then either 1 = 0 or
CHAPTER 9 REVIEW

401
1 =
1
0.0002
= 5000 [as in part (a)]. If 1 = 1000, then 0 = 1000[0.08(1 30.0002 1000) 30.001W ] C
0 = 80(1 30.2) 3W C W = 64.
Case (i): W = 0, 1 = 0: both populations are zero
Case (ii): W = 0, 1 = 5000: see part (a)
Case (iii): 1 = 1000, W = 64: the predator/prey interaction balances and the populations are stable.
(c) The populations of wolves and rabbits uctuate around 64 and 1000, respectively, and eventually stabilize at those values.
(d)
9 Review
1. (a) A differential equation is an equation that contains an unknown function and one or more of its derivatives.
(b) The order of a differential equation is the order of the highest derivative that occurs in the equation.
(c) An initial condition is a condition of the form j(t
0
) = j
0
.
2. j
0
= r
2
+j
2
D 0 for all r and j. j
0
= 0 only at the origin, so there is a horizontal tangent at (0. 0), but nowhere else. The
graph of the solution is increasing on every interval.
3. See the paragraph preceding Example 1 in Section 9.2.
4. See the paragraph next to Figure 14 in Section 9.2.
5. A separable equation is a rst-order differential equation in which the expression for djdr can be factored as a function of r
times a function of j, that is, djdr = o(r)1(j). We can solve the equation by integrating both sides of the equation
dj1(j) = o(r)dr and solving for j.
6. A rst-order linear differential equation is a differential equation that can be put in the form
dj
dr
+1(r) j = Q(r), where 1
and Q are continuous functions on a given interval. To solve such an equation, multiply it by the integrating factor
1(r) = c
U
T(i)ui
to put it in the form [1(r) j]
0
= 1(r) Q(r) and then integrate both sides to get 1(r) j =

1(r) Q(r) dr,
that is, c
U
T(i) ui
j =

c
U
T(i) ui
Q(r) dr. Solving for j gives us j = c
3
U
T(i) ui
c
U
T(i) ui
Q(r) dr.
402

7. (a)
dj
dt
= Ij; the relative growth rate,
1
j
dj
dt
, is constant.
(b) The equation in part (a) is an appropriate model for population growth, assuming that there is enough room and nutrition to
support the growth.
(c) If j(0) = j
0
, then the solution is j(t) = j
0
c
II
.
8. (a) d1dt = I1(1 311), where 1 is the carrying capacity.
(b) The equation in part (a) is an appropriate model for population growth, assuming that the population grows at a rate
proportional to the size of the population in the beginning, but eventually levels off and approaches its carrying capacity
because of limited resources.
9. (a) d1dt = I1 3o1o and dodt = 3vo +/1o.
(b) In the absence of sharks, an ample food supply would support exponential growth of the sh population, that is,
d1dt = I1, where I is a positive constant. In the absence of sh, we assume that the shark population would decline at a
rate proportional to itself, that is, dodt = 3vo, where v is a positive constant.
1. True. Since j
4
D 0, j
0
= 31 3j
4
< 0 and the solutions are decreasing functions.
3. False. r +j cannot be written in the form o(r)1(j).
5. True. c
i
j
0
= j i j
0
= c
3i
j i j
0
+ (3c
3i
)j = 0, which is of the form j
0
+1(r) j = Q(r), so the
equation is linear.
7. True. By comparing
dj
dt
= 2j
1 3
j
5
with the logistic differential equation (9.4.4), we see that the carrying
capacity is 5; that is, lim
I<"
j = 5.
1. (a) (b) lim
I<"
j(t) appears to be nite for 0 $ c $ 4. In fact
lim
I<"
j(t) = 4 for c = 4, lim
I<"
j(t) = 2 for 0 < c < 4, and
lim
I<"
j(t) = 0 for c = 0. The equilibrium solutions are
j(t) = 0, j(t) = 2, and j(t) = 4.
CHAPTER 9 REVIEW

403
3. (a) We estimate that when r = 0.3, j = 0.8, so j(0.3) E 0.8.
(b) / = 0.1, r
0
= 0, j
0
= 1 and 1(r. j) = r
2
3j
2
. So j
n
= j
n31
+ 0.1
r
2
n31
3j
2
n31
. Thus,
j
1
= 1 + 0.1
0
2
31
2
= 0.9, j
2
= 0.9 + 0.1
0.1
2
30.9
2
= 0.82, j
3
= 0.82 + 0.1
0.2
2
30.82
2
= 0.75676.
This is close to our graphical estimate of j(0.3) E 0.8.
(c) The centers of the horizontal line segments of the direction eld are located on the lines j = r and j = 3r.
When a solution curve crosses one of these lines, it has a local maximum or minimum.
5. j
0
= rc
3sin i
3j cos r i j
0
+ (cos r) j = rc
3sin i
(-). This is a linear equation and the integrating factor is
1(r) = c
U
cos i ui
= c
sin i
. Multiplying (-) by c
sin i
gives c
sin i
j
0
+c
sin i
(cos r) j = r i (c
sin i
j)
0
= r i
c
sin i
j =
1
2
r
2
+C i j =

1
2
r
2
+C
c
3sin i
.
7. 2jc
u
2
j
0
= 2r + 3
I
r i 2jc
u
2 dj
dr
= 2r + 3
I
r i 2jc
u
2
dj =

2r + 3
I
r
dr i
2jc
u
2
dj =

2r + 3
I
r
dr i c
u
2
= r
2
+ 2r
3'2
+C i j
2
= ln(r
2
+ 2r
3'2
+C) i
j =
ln(r
2
+ 2r
3'2
+C)
9.
dv
dt
+ 2tv = v i
dv
dt
= v 32tv = v(1 32t) i

dv
v
=

(1 32t) dt i ln|v| = t 3t
2
+C i
|v| = c
I3I
2
+C
= Ic
I3I
2
. Since v(0) = 5, 5 = Ic
0
= I. Thus, v(t) = 5c
I3I
2
.
11. rj
0
3j = rlnr i j
0
3
1
r
j = lnr. 1(r) = c
U
(31'i) ui
= c
3ln|i|
=

c
ln|i|
31
= |r|
31
= 1r since the condition
j(1) = 2 implies that we want a solution with r 0. Multiplying the last differential equation by 1(r) gives
1
r
j
0
3
1
r
2
j =
1
r
lnr i

1
r
j
0
=
1
r
lnr i
1
r
j =

lnr
r
dr i
1
r
j =
1
2
(lnr)
2
+C i
j =
1
2
r(lnr)
2
+Cr. Now j(1) = 2 i 2 = 0 +C i C = 2, so j =
1
2
r(lnr)
2
+ 2r.
13.
d
dr
(j) =
d
dr
(Ic
i
) i j
0
= Ic
i
= j, so the orthogonal trajectories must have j
0
= 3
1
j
i
dj
dr
= 3
1
j
i
j dj = 3dr i

j dj = 3
dr i
1
2
j
2
= 3r +C i r = C 3
1
2
j
2
, which are parabolas with a horizontal axis.
404

15. (a) Using (4) and (7) in Section 9.4, we see that for
d1
dt
= 0.11
1 3
1
2000
with 1(0) = 100, we have I = 0.1,

1 = 2000, 1
0
= 100, and =
2000 3100
100
= 19. Thus, the solution of the initial-value problem is
1(t) =
2000
1 + 19c
30.1I
and 1(20) =
2000
1 + 19c
32
E 560.
(b) 1 = 1200 C 1200 =
2000
1 + 19c
30.1I
C 1 + 19c
30.1I
=
2000
1200
C 19c
30.1I
=
5
3
31 C
c
30.1I
=

2
3
19 C 30.1t = ln
2
57
C t = 310 ln
2
57
E 33.5.
17. (a)
d1
dt
x 1
"
31 i
d1
dt
= I(1
"
31) i

d1
1
"
31
=

I dt i 3ln|1
"
31| = It +C i
ln|1
"
31| = 3It 3C i |1
"
31| = c
3II3C
i 1
"
31 = c
3II
i 1 = 1
"
3c
3II
.
At t = 0, 1 = 1(0) = 1
"
3 i = 1
"
31(0) i 1(t) = 1
"
3[1
"
31(0)]c
3II
.
(b) 1
"
= 53 cm, 1(0) = 10 cm, and I = 0.2 i 1(t) = 53 3(53 310)c
30.2I
= 53 343c
30.2I
.
19. Let 1 represent the population and 1 the number of infected people. The rate of spread d1dt is jointly proportional to 1 and
to 1 31, so for some constant I,
d1
dt
= I1(1 31) = (I1)1
1 3
1
1
. From Equation 9.4.7 with 1 = 1 and I replaced by

I1, we have 1(t) =
1
1 +c
3ITI
=
1
0
1
1
0
+ (1 31
0
)c
3ITI
.
Now, measuring t in days, we substitute t = 7, 1 = 5000, 1
0
= 160 and 1(7) = 1200 to nd I:
1200 =
160 5000
160 + (5000 3160)c
350007I
C 3 =
2000
160 + 4840c
335,000I
C 480 + 14,520c
335,000I
= 2000 C
c
335,000I
=
2000 3480
14,520
C 335,000I = ln
38
363
C I =
31
35,000
ln
38
363
E 0.00006448. Next, let
1 = 5000 80% = 4000, and solve for t: 4000 =
160 5000
160 + (5000 3160)c
3I5000I
C 1 =
200
160 + 4840c
35000II
C
160 + 4840c
35000II
= 200 C c
35000II
=
200 3160
4840
C 35000It = ln
1
121
C
t =
31
5000I
ln
1
121
=
1
1
7
ln
38
363
ln
1
121
= 7
ln121
ln
363
38
E 14.875. So it takes about 15 days for 80% of the population
to be infected.
21.
d/
dt
= 3
1
\
/
I +/
i

I +/
/
d/ =

3
1
\
dt i

1 +
I
/
d/ = 3
1
\
1 dt i
/ +I ln/ = 3
1
\
t +C. This equation gives a relationship between / and t, but it is not possible to isolate / and express it in
terms of t.
23. (a) drdt = 0.4r(1 30.000005r) 30.002rj, djdt = 30.2j + 0.000008rj. If j = 0, then
drdt = 0.4r(1 30.000005r), so drdt = 0 C r = 0 or r = 200,000, which shows that the insect population
increases logistically with a carrying capacity of 200,000. Since drdt 0 for 0 < r < 200,000 and drdt < 0 for
r 200,000, we expect the insect population to stabilize at 200,000.
CHAPTER 9 REVIEW

405
(b) r and j are constant i r
0
= 0 and j
0
= 0 i
0 = 0.4r(1 30.000005r) 30.002rj

0 = 30.2j + 0.000008rj
0 = 0.4r[(1 30.000005r) 30.005j]

0 = j(30.2 + 0.000008r)
The second equation is true if j = 0 or r =
0.2
0.000008
= 25,000. If j = 0 in the rst equation, then either r = 0
or r =
1
0.000005
= 200,000. If r = 25,000, then 0 = 0.4(25,000)[(1 30.000005 25,000) 30.005j] i
0 = 10,000[(1 30.125) 30.005j] i 0 = 8750 350j i j = 175.
Case (i): j = 0, r = 0: Zero populations
Case (ii): j = 0, r = 200,000: In the absence of birds, the insect population is always 200,000.
Case (iii): r = 25,000, j = 175: The predator/prey interaction balances and the populations are stable.
(c) The populations of the birds and insects uctuate
around 175 and 25,000, respectively, and
eventually stabilize at those values.
(d)
25. (a)
d
2
j
dr
2
= I
1 +
dj
dr
2
. Setting : =
dj
dr
, we get
d:
dr
= I
I
1 +:
2
i
d:
I
1 +:
2
= I dr. Using Formula 25 gives
ln
: +
I
1 +:
2
= Ir +c i : +
I
1 +:
2
= Cc
Ii
[where C = c
c
] i
I
1 +:
2
= Cc
Ii
3: i
1 +:
2
= C
2
c
2Ii
32Cc
Ii
: +:
2
i 2Cc
Ii
: = C
2
c
2Ii
31 i : =
C
2
c
Ii
3
1
2C
c
3Ii
. Now
dj
dr
=
C
2
c
Ii
3
1
2C
c
3Ii
i j =
C
2I
c
Ii
+
1
2CI
c
3Ii
+C
0
. From the diagram in the text, we see that j(0) = o
and j(/) = /. o = j(0) =
C
2I
+
1
2CI
+C
0
i C
0
= o 3
C
2I
3
1
2CI
i
j =
C
2I
(c
Ii
31) +
1
2CI
(c
3Ii
31) +o. From / = j(/), we nd / =
C
2I
(c
Il
31) +
1
2CI
(c
3Il
31) +o
and / =
C
2I
(c
3Il
31) +
1
2CI
(c
Il
31) +o. Subtracting the second equation from the rst, we get
0 =
C
I
c
Il
3c
3Il
2
3
1
CI
c
Il
3c
3Il
2
=
1
I
C 3
1
C
sinhI/.
Now I 0 and / 0, so sinhI/ 0 and C = 1. If C = 1, then
j =
1
2I
(c
Ii
31) +
1
2I
(c
3Ii
31) +o =
1
I
c
Ii
+c
3Ii
2
3
1
I
+o = o +
1
I
(coshIr 31). If C = 31,
then j = 3
1
2I
(c
Ii
31) 3
1
2I
(c
3Ii
31) +o =
31
I
c
Ii
+c
3Ii
2
+
1
I
+o = o 3
1
I
(coshIr 31).
Since I 0, coshIr D 1, and j D o, we conclude that C = 1 and j = o +
1
I
(coshIr 31), where
406

/ = j(/) = o +
1
I
(coshI/ 31). Since cosh(I/) = cosh(3I/), there is no further information to extract from the
condition that j(/) = j(3/). However, we could replace o with the expression / 3
1
I
(coshI/ 31), obtaining
j = / +
1
I
(coshIr 3coshI/). It would be better still to keep o in the expression for j, and use the expression for / to
solve for I in terms of o, /, and /. That would enable us to express j in terms of r and the given parameters o, /, and /.
Sadly, it is not possible to solve for I in closed form. That would have to be done by numerical methods when specic
parameter values are given.
(b) The length of the cable is
1=

l
3l
1 + (djdr)
2
dr =

l
3l
1 + sinh
2
Irdr =

l
3l
coshIrdr = 2
l
0
coshIrdr
= 2
(1I) sinhIr
l
0
= (2I) sinhI/
PROBLEMS PLUS
1. We use the Fundamental Theorem of Calculus to differentiate the given equation:
[1(r)]
2
= 100 +
i
0
[1(t)]
2
+ [1
0
(t)]
2
dt i 21(r)1
0
(r) = [1(r)]
2
+ [1
0
(r)]
2
i
[1(r)]
2
+ [1
0
(r)]
2
321(r)1
0
(r) = 0 i [1(r) 31
0
(r)]
2
= 0 C 1(r) = 1
0
(r). We can solve this as a separable
equation, or else use Theorem 9.4.2 with I = 1, which says that the solutions are 1(r) = Cc
i
. Now [1(0)]
2
= 100, so
1(0) = C = 10, and hence 1(r) = 10c
i
are the only functions satisfying the given equation.
3. 1
0
(r) = lim
I<0
1(r +/) 31(r)
/
= lim
I<0
1(r) [1(/) 31]
/
[since 1(r +/) = 1(r)1(/)]
= 1(r) lim
I<0
1(/) 31
/
= 1(r) lim
I<0
1(/) 31(0)
/ 30
= 1(r)1
0
(0) = 1(r)
Therefore, 1
0
(r) = 1(r) for all r and from Theorem 9.4.2 we get 1(r) = c
i
. Now 1(0) = 1 i = 1 i
1(r) = c
i
.
5. The area under the graph of 1 from 0 to r is proportional to the (n + 1)st power of 1(r) translates to
i
0
1(t) dt = I[1(r)]
n+1
for some constant I. By FTC1,
d
dr
i
0
1(t) dt =
d
dr
I[1(r)]
n+1
i
1(r) = I(n + 1)[1(r)]
n
1
0
(r) i 1 = I(n + 1)[1(r)]
n31
1
0
(r) i 1 = I(n + 1)j
n31
dj
dr
i
I(n + 1)j
n31
dj = dr i

I(n + 1)j
n31
dj =

dr i I(n + 1)
1
n
j
n
= r +C.
Now 1(0) = 0 i 0 = 0 +C i C = 0 and then 1(1) = 1 i I(n + 1)
1
n
= 1 i I =
n
n + 1
,
so j
n
= r and j = 1(r) = r
1'n
.
7. Let j(t) denote the temperature of the peach pie t minutes after 5:00 PM and 1 the temperature of the room. Newtons Law of
Cooling gives us djdt = I(j 31). Solving for j we get
dj
j 31
= I dt i ln|j 31| = It +C i
|j 31| = c
II+C
i j 31 = c
II
c
C
i j = Ac
II
+1, where A is a nonzero constant. We are given
temperatures at three times.
j(0) = 100 i 100 = A +1 i 1 = 100 3A
j(10) = 80 i 80 = Ac
10I
+1 (1)
j(20) = 65 i 65 = Ac
20I
+1 (2)
Substituting 100 3A for 1in (1) and (2) gives us
320 = Ac
10I
3A (3) and 335 = Ac
20I
3A (4)
Dividing (3) by (4) gives us
320
335
=
A
c
10I
31
A(c
20I
31)
i
4
7
=
c
10I
31
c
20I
31
i 4c
20I
34 = 7c
10I
37 i
407
408

4c
20I
37c
10I
+ 3 = 0. This is a quadratic equation in c
10I
.

4c
10I
33
c
10I
31
= 0 i c
10I
=
3
4
or 1 i
10I = ln
3
4
or ln1 i I =
1
10
ln
3
4
since I is a nonzero constant of proportionality. Substituting
3
4
for c
10I
in (3) gives us
320 = A
3
4
3A i 320 = 3
1
4
A i A = 80. Now 1 = 100 3A so 1 = 20
C.
9. (a) While running from (1. 0) to (r. j), the dog travels a distance
c =

1
i
1 + (djdr)
2
dr = 3
i
1
1 + (djdr)
2
dr, so
dc
dr
= 3
1 + (djdr)
2
. The dog and rabbit run at the same speed, so the
rabbits position when the dog has traveled a distance c is (0. c). Since the
dog runs straight for the rabbit,
dj
dr
=
c 3j
0 3r
(see the gure).
Thus, c = j 3r
dj
dr
i
dc
dr
=
dj
dr
3
r
d
2
j
dr
2
+ 1
dj
dr
= 3r
d
2
j
dr
2
. Equating the two expressions for
dc
dr
gives us r
d
2
j
dr
2
=
1 +
dj
dr
2
, as claimed.
(b) Letting : =
dj
dr
, we obtain the differential equation r
d:
dr
=
I
1 +:
2
, or
d:
I
1 +:
2
=
dr
r
. Integrating:
lnr =

d:
I
1 +:
2
25
= ln
: +
1 +:
2
+C. When r = 1, : = djdr = 0, so ln1 = ln1 +C. Therefore,

C = ln1, so lnr = ln
I
1 +:
2
+:
+ ln1 = ln
1
I
1 +:
2
+:
i r = 1
I
1 +:
2
+:
i
I
1 +:
2
=
r
1
3: i 1 +:
2
=

r
1
2
3
2r:
1
+:
2
i

r
1
2
32:
r
1
31 = 0 i
: =
(r1)
2
31
2(r1)
=
r
2
31
2
21r
=
r
21
3
1
2
1
r
[for r 0]. Since : =
dj
dr
, j =
r
2
41
3
1
2
lnr +C
1
.
Since j = 0 when r = 1, 0 =
1
4
3
1
2
ln1 + C
1
i C
1
=
1
2
ln1 3
1
4
. Thus,
j =
r
2
41
3
1
2
lnr +
1
2
ln1 3
1
4
=
r
2
31
2
41
3
1
2
ln
r
1
.
(c) As r <0
+
, j <", so the dog never catches the rabbit.
11. (a) We are given that \ =
1
3
v
2
/, d\dt = 60,000 ft
3
h, and v = 1.5/ =
3
2
/. So \ =
1
3
3
2
/
2
/ =
3
4
/
3
i
d\
dt
=
3
4
3/
2
d/
dt
=
9
4
/
2
d/
dt
. Therefore,
d/
dt
=
4(d\dt)
9/
2
=
240,000
9/
2
=
80,000
3/
2
(-) i
3/
2
d/ =

80,000 dt i /
3
= 80,000t +C. When t = 0, / = 60. Thus, C = 60
3
= 216,000, so
/
3
= 80,000t + 216,000. Let / = 100. Then 100
3
= 1,000,000 = 80,000t + 216,000 i
80,000t = 784,000 i t = 9.8, so the time required is 9.8 hours.

409
(b) The oor area of the silo is 1 = 200
2
= 40,000 ft
2
, and the area of the base of the pile is
= v
2
=
3
2
/
2
=
9r
4
/
2
. So the area of the oor which is not covered when / = 60 is
1 3 = 40,000 38100 = 31,900 E 100,217 ft
2
. Now =
9r
4
/
2
i ddt =
9r
4
2/(d/dt),
and from (-) in part (a) we know that when / = 60, d/dt =
80,000
3(60)
2
=
200
27
fth. Therefore,
ddt =
9r
4
(2)(60)
200
27
= 2000 E 6283 ft
2
h.
(c) At / = 90 ft, d\dt = 60,000 320,000 = 40,000 ft
3
h. From (-) in part (a),
d/
dt
=
4(d\dt)
9/
2
=
4(40,000)
9/
2
=
160,000
9/
2
i

9/
2
d/ =

160,000 dt i 3/
3
= 160,000t +C. When t = 0,
/ = 90; therefore, C = 3 729,000 = 2,187,000. So 3/
3
= 160,000t + 2,187,000. At the top, / = 100 i
3(100)
3
= 160,000t + 2,187,000 i t =
813,000
160,000
E 5.1. The pile reaches the top after about 5.1 h.
13. Let 1(o. /) be any point on the curve. If i is the slope of the tangent line at 1, then i = j
0
(o), and an equation of the
normal line at 1 is j 3/ = 3
1
i
(r 3o), or equivalently, j = 3
1
i
r +/ +
o
i
. The j-intercept is always 6, so
/ +
o
i
= 6 i
o
i
= 6 3/ i i =
o
6 3/
. We will solve the equivalent differential equation
dj
dr
=
r
6 3j
i
(6 3j) dj = rdr i

(6 3j) dj =

rdr i 6j 3
1
2
j
2
=
1
2
r
2
+C i 12j 3j
2
= r
2
+1.
Since (3. 2) is on the curve, 12(2) 32
2
= 3
2
+1 i 1 = 11. So the curve is given by 12j 3j
2
= r
2
+ 11 i
r
2
+j
2
312j + 36 = 311 + 36 i r
2
+ (j 36)
2
= 25, a circle with center (0. 6) and radius 5.
10 PARAMETRIC EQUATIONS AND POLAR COORDINATES
10.1 Curves Defined by Parametric Equations
1. r = 1 +
I
t, j = t
2
34t, 0 $ t $ 5
t 0 1 2 3 4 5
r 1 2 1 +
I
2
2.41
1 +
I
3
2.73
3 1 +
I
5
3.24
j 0 33 34 33 0 5
3. r = 5 sint, j = t
2
, 3 $ t $
t 3 32 0 2
r 0 35 0 5 0
j
2
9.87
2
4
2.47
0
2
4
2.47
2
9.87
5. r = 3t 35, j = 2t + 1
(a)
t 32 31 0 1 2 3 4
r 311 38 35 32 1 4 7
j 33 31 1 3 5 7 9
(b) r = 3t 35 i 3t = r + 5 i t =
1
3
(r + 5) i
j = 2
1
3
(r + 5) + 1, so j =
2
3
r +
13
3
.
7. r = t
2
32, j = 5 32t, 33 $ t $ 4
(a)
t 33 32 31 0 1 2 3 4
r 7 2 31 32 31 2 7 14
j 11 9 7 5 3 1 31 33
(b) j = 5 32t i 2t = 5 3j i t =
1
2
(5 3j) i
r =

1
2
(5 3j)
2
32, so r =
1
4
(5 3j)
2
32, 33 $ j $ 11.
411
412

CHAPTER 10 PARAMETRIC EQUATIONS AND POLAR COORDINATES
9. r =
I
t, j = 1 3t
(a)
t 0 1 2 3 4
r 0 1 1.414 1.732 2
j 1 0 31 32 33
(b) r =
I
t i t = r
2
i j = 1 3t = 1 3r
2
. Since t D 0, r D 0.
So the curve is the right half of the parabola j = 1 3r
2
.
11. (a) r = sin0, j = cos 0, 0 $ 0 $ . r
2
+j
2
= sin
2
0 + cos
2
0 = 1. Since
0 $ 0 $ , we have sin0 D 0, so r D 0. Thus, the curve is the right half of
the circle r
2
+j
2
= 1.
(b)

13. (a) r = sint. j = csc t, 0 < t <
r
2
. j = csc t =
1
sint
=
1
r
. For 0 < t <
r
2
,
we have 0 < r < 1 and j 1. Thus, the curve is the portion
of the hyperbola j = 1r with j 1.
(b)
15. (a) r = c
2I
i 2t = lnr i t =
1
2
lnr.
j = t + 1 =
1
2
lnr + 1.
(b)
17. (a) r = sinht, j = cosht i j
2
3r
2
= cosh
2
t 3sinh
2
t = 1. Since
j = cosht D 1, we have the upper branch of the hyperbola j
2
3r
2
= 1.
(b)
19. r = 3 + 2 cos t, j = 1 + 2 sint, 2 $ t $ 32. By Example 4 with v = 2, / = 3, and I = 1, the motion of the particle
takes place on a circle centered at (3. 1) with a radius of 2. As t goes from
r
2
to
3r
2
, the particle starts at the point (3. 3) and
moves counterclockwise to (3. 31) [one-half of a circle].
21. r = 5 sint, j = 2 cos t i sint =
r
5
, cos t =
j
2
. sin
2
t + cos
2
t = 1 i

r
5
2
+
j
2
2
= 1. The motion of the
particle takes place on an ellipse centered at (0. 0). As t goes from 3 to 5, the particle starts at the point (0. 32) and moves
clockwise around the ellipse 3 times.
SECTION 10.1 CURVES DEFINED BY PARAMETRIC EQUATIONS

413
23. We must have 1 $ r $ 4 and 2 $ j $ 3. So the graph of the curve must be contained in the rectangle [1. 4] by [2. 3].
25. When t = 31, (r. j) = (0. 31). As t increases to 0, r decreases to 31 and j
increases to 0. As t increases from 0 to 1, r increases to 0 and j increases to 1.
As t increases beyond 1, both r and j increase. For t < 31, r is positive and
decreasing and j is negative and increasing. We could achieve greater accuracy
by estimating r- and j-values for selected values of t from the given graphs and
plotting the corresponding points.
27. When t = 0 we see that r = 0 and j = 0, so the curve starts at the origin. As t
increases from 0 to
1
2
, the graphs show that j increases from 0 to 1 while r
increases from 0 to 1, decreases to 0 and to 31, then increases back to 0, so we
arrive at the point (0. 1). Similarly, as t increases from
1
2
to 1, j decreases from 1
to 0 while r repeats its pattern, and we arrive back at the origin. We could achieve greater accuracy by estimating r- and
j-values for selected values of t from the given graphs and plotting the corresponding points.
29. As in Example 6, we let j = t and r = t 33t
3
+t
5
and use a t-interval of [33. 3].
31. (a) r = r
1
+ (r
2
3r
1
)t, j = j
1
+ (j
2
3j
1
)t, 0 $ t $ 1. Clearly the curve passes through 1
1
(r
1
. j
1
) when t = 0 and
through 1
2
(r
2
. j
2
) when t = 1. For 0 < t < 1, r is strictly between r
1
and r
2
and j is strictly between j
1
and j
2
. For
every value of t, r and j satisfy the relation j 3j
1
=
j
2
3j
1
r
2
3r
1
(r 3r
1
), which is the equation of the line through
1
1
(r
1
. j
1
) and 1
2
(r
2
. j
2
).
Finally, any point (r. j) on that line satises
j 3j
1
j
2
3j
1
=
r 3r
1
r
2
3r
1
; if we call that common value t, then the given
parametric equations yield the point (r. j); and any (r. j) on the line between 1
1
(r
1
. j
1
) and 1
2
(r
2
. j
2
) yields a value of
t in [0. 1]. So the given parametric equations exactly specify the line segment from 1
1
(r
1
. j
1
) to 1
2
(r
2
. j
2
).
(b) r = 32 + [3 3(32)]t = 32 + 5t and j = 7 + (31 37)t = 7 38t for 0 $ t $ 1.
33. The circle r
2
+ (j 31)
2
= 4 has center (0. 1) and radius 2, so by Example 4 it can be represented by r = 2 cos t,
j = 1 + 2 sint, 0 $ t $ 2. This representation gives us the circle with a counterclockwise orientation starting at (2. 1).
(a) To get a clockwise orientation, we could change the equations to r = 2 cos t, j = 1 32 sint, 0 $ t $ 2.
(b) To get three times around in the counterclockwise direction, we use the original equations r = 2 cos t, j = 1 +2 sint with
the domain expanded to 0 $ t $ 6.
414

(c) To start at (0. 3) using the original equations, we must have r
1
= 0; that is, 2 cos t = 0. Hence, t =
r
2
. So we use
r = 2 cos t, j = 1 + 2 sint,
r
2
$ t $
3r
2
.
Alternatively, if we want t to start at 0, we could change the equations of the curve. For example, we could use
r = 32 sint, j = 1 + 2 cos t, 0 $ t $ .
35. Big circle: Its centered at (2. 2) with a radius of 2, so by Example 4, parametric equations are
r = 2 + 2 cos t. j = 2 + 2 sint. 0 $ t $ 2
Small circles: They are centered at (1. 3) and (3. 3) with a radius of 0.1. By Example 4, parametric equations are
(left) r = 1 + 0.1 cos t. j = 3 + 0.1 sint. 0 $ t $ 2
and (right) r = 3 + 0.1 cos t. j = 3 + 0.1 sint. 0 $ t $ 2
Semicircle: Its the lower half of a circle centered at (2. 2) with radius 1. By Example 4, parametric equations are
r = 2 + 1 cos t. j = 2 + 1 sint. $ t $ 2
To get all four graphs on the same screen with a typical graphing calculator, we need to change the last t-interval to [0. 2] in
order to match the others. We can do this by changing t to 0.5t. This change gives us the upper half. There are several ways to
get the lower halfone is to change the + to a 3 in the j-assignment, giving us
r = 2 + 1 cos(0.5t). j = 2 31 sin(0.5t). 0 $ t $ 2
37. (a) r = t
3
i t = r
1'3
, so j = t
2
= r
2'3
.
We get the entire curve j = r
2'3
traversed in a left to
right direction.
(b) r = t
6
i t = r
1'6
, so j = t
4
= r
4'6
= r
2'3
.
Since r = t
6
D 0, we only get the right half of the
curve j = r
2'3
.
(c) r = c
33I
= (c
3I
)
3
[so c
3I
= r
1'3
],
j = c
32I
= (c
3I
)
2
= (r
1'3
)
2
= r
2'3
.
If t < 0, then r and j are both larger than 1. If t 0, then r and j are between 0
and 1. Since r 0 and j 0, the curve never quite reaches the origin.
39. The case
r
2
< 0 < is illustrated. C has coordinates (v0. v) as in Example 7,
and Q has coordinates (v0. v +v cos( 30)) = (v0. v(1 3cos 0))
[since cos( 3c) = cos cos c + sin sinc = 3cos c], so 1 has coordinates
(v0 3v sin( 30). v(1 3cos 0)) = (v(0 3sin0). v(1 3cos 0))
[since sin( 3c) = sin cos c 3cos sinc = sinc]. Again we have the
parametric equations r = v(0 3sin0), j = v(1 3cos 0).
SECTION 10.1 CURVES DEFINED BY PARAMETRIC EQUATIONS

415
41. It is apparent that r = |OQ| and j = |Q1| = |oT|. From the diagram,
r = |OQ| = o cos 0 and j = |oT| = / sin0. Thus, the parametric equations are
r = o cos 0 and j = / sin0. To eliminate 0 we rearrange: sin0 = j/ i
sin
2
0 = (j/)
2
and cos 0 = ro i cos
2
0 = (ro)
2
. Adding the two
equations: sin
2
0 + cos
2
0 = 1 = r
2
o
2
+j
2
/
2
. Thus, we have an ellipse.
43. C = (2o cot 0. 2o), so the r-coordinate of 1 is r = 2o cot 0. Let 1 = (0. 2o).
Then _O1 is a right angle and _O1 = 0, so |O| = 2o sin0 and
= ((2o sin0) cos 0. (2o sin0) sin0). Thus, the j-coordinate of 1
is j = 2o sin
2
0.
45. (a) There are 2 points of intersection:
(33. 0) and approximately (32.1. 1.4).
(b) A collision point occurs when r
1
= r
2
and j
1
= j
2
for the same t. So solve the equations:
3 sint = 33 + cos t (1)
2 cos t = 1 + sint (2)
From (2), sint = 2 cos t 31. Substituting into (1), we get 3(2 cos t 31) = 33 + cos t i 5 cos t = 0 (-) i
cos t = 0 i t =
r
2
or
3r
2
. We check that t =
3r
2
satises (1) and (2) but t =
r
2
does not. So the only collision point
occurs when t =
3r
2
, and this gives the point (33. 0). [We could check our work by graphing r
1
and r
2
together as
functions of t and, on another plot, j
1
and j
2
as functions of t. If we do so, we see that the only value of t for which both
pairs of graphs intersect is t =
3r
2
.]
(c) The circle is centered at (3. 1) instead of (33. 1). There are still 2 intersection points: (3. 0) and (2.1. 1.4), but there are
no collision points, since (-) in part (b) becomes 5 cos t = 6 i cos t =
6
5
1.
47. r = t
2
. j = t
3
3ct. We use a graphing device to produce the graphs for various values of c with 3 $ t $ . Note that all
the members of the family are symmetric about the r-axis. For c < 0, the graph does not cross itself, but for c = 0 it has a
cusp at (0. 0) and for c 0 the graph crosses itself at r = c, so the loop grows larger as c increases.
416

49. Note that all the Lissajous gures are symmetric about the r-axis. The parameters o and / simply stretch the graph in the
r- and j-directions respectively. For o = / = n = 1 the graph is simply a circle with radius 1. For n = 2 the graph crosses
itself at the origin and there are loops above and below the r-axis. In general, the gures have n 31 points of intersection,
all of which are on the j-axis, and a total of n closed loops.
o = / = 1 n = 2 n = 3
10.2 Calculus with Parametric Curves
1. r = t sint, j = t
2
+t i
dj
dt
= 2t + 1,
dr
dt
= t cos t + sint, and
dj
dr
=
djdt
drdt
=
2t + 1
t cos t + sint
.
3. r = t
4
+ 1, j = t
3
+t; t = 31.
dj
dt
= 3t
2
+ 1,
dr
dt
= 4t
3
, and
dj
dr
=
djdt
drdt
=
3t
2
+ 1
4t
3
. When t = 31,
(r. j) = (2. 32) and djdr =
4
34
= 31, so an equation of the tangent to the curve at the point corresponding to t = 31
is j 3(32) = (31)(r 32), or j = 3r.
5. r = c
I
I
, j = t 3lnt
2
; t = 1.
dj
dt
= 1 3
2t
t
2
= 1 3
2
t
,
dr
dt
=
c
I
I
2
I
t
, and
dj
dr
=
djdt
drdt
=
1 32t
c
I
I
2
I
t

2t
2t
=
2t 34
I
tc
I
I
.
When t = 1, (r. j) = (c. 1) and
dj
dr
= 3
2
c
, so an equation of the tangent line is j 31 = 3
2
c
(r 3c), or j = 3
2
c
r + 3.
7. (a) r = 1 + lnt, j = t
2
+ 2; (1. 3).
dj
dt
= 2t.
dr
dt
=
1
t
. and
dj
dr
=
djdt
drdt
=
2t
1t
= 2t
2
.
At (1. 3), r = 1 + lnt = 1 i lnt = 0 i t = 1 and
dj
dr
= 2, so an equation of the tangent is j 33 = 2(r 31),
or j = 2r + 1.
(b) r = 1 + lnt i r 31 = lnt i t = c
i31
, so j = (c
i31
)
2
+ 2 = c
2i32
+ 2 and
dj
dr
= 2c
2i32
.
When r = 1,
dj
dr
= 2c
0
= 2, so an equation of the tangent is j = 2r + 1, as in part (a).
9. r = 6 sint, j = t
2
+t; (0. 0).
dj
dr
=
djdt
drdt
=
2t + 1
6 cos t
. The point (0. 0) corresponds to t = 0, so the
slope of the tangent at that point is
1
6
. An equation of the tangent is therefore
j 30 =
1
6
(r 30), or j =
1
6
r.
SECTION 10.2 CALCULUS WITH PARAMETRIC CURVES

417
11. r = 4 + t
2
, j = t
2
+ t
3
i
dj
dr
=
djdt
drdt
=
2t + 3t
2
2t
= 1 +
3
2
t i
d
2
j
dr
2
=
d
dr
dj
dr
=
d(djdr)dt
drdt
=
(ddt)
1 +
3
2
t
2t
=
32
2t
=
3
4t
.
The curve is CU when
d
2
j
dr
2
0, that is, when t 0.
13. r = t 3 c
I
, j = t + c
3I
i
dj
dr
=
djdt
drdt
=
1 3c
3I
1 3c
I
=
1 3
1
c
I
1 3c
I
=
c
I
31
c
I
1 3c
I
= 3c
3I
i
d
2
j
dr
2
=
d
dt
dj
dr
drdt
=
d
dt
(3c
3I
)
drdt
=
c
3I
1 3c
I
.
The curve is CU when c
I
< 1 [since c
3I
0] i t < 0.
15. r = 2 sint, j = 3 cos t, 0 < t < 2.
dj
dr
=
djdt
drdt
=
33 sint
2 cos t
= 3
3
2
tant, so
d
2
j
dr
2
=
d
dt
dj
dr
drdt
=
3
3
2
sec
2
t
2 cos t
= 3
3
4
sec
3
t.
The curve is CU when sec
3
t < 0 i sec t < 0 i cos t < 0 i
r
2
< t <
3r
2
.
17. r = 10 3t
2
, j = t
3
312t.
dj
dt
= 3t
2
312 = 3(t + 2)(t 32), so
dj
dt
= 0 C
t = 2 C (r. j) = (6. ~16).
dr
dt
= 32t, so
dr
dt
= 0 C t = 0 C (r. j) = (10. 0).
The curve has horizontal tangents at (6. 16) and a vertical
tangent at (10. 0).
19. r = 2 cos 0, j = sin20.
dj
d0
= 2 cos 20, so
dj
d0
= 0 C 20 =
r
2
+n
[n an integer] C 0 =
r
4
+
r
2
n C
(r. j) =
I
2. 1
. Also,
dr
d0
= 32 sin0, so
dr
d0
= 0 C 0 = n C (r. j) = (2. 0).
The curve has horizontal tangents at
I
2. 1
(four points), and vertical tangents at (2. 0).

21. From the graph, it appears that the rightmost point on the curve r = t 3t
6
, j = c
I
is about (0.6. 2). To nd the exact coordinates, we nd the value of t for which the
graph has a vertical tangent, that is, 0 = drdt = 1 36t
5
C t = 1
5
I
6.
Hence, the rightmost point is
1
5
I
6 31
6
5
I
6
. c
1'
5
I
6
=

5 6
36'5
. c
6
1@5
E (0.58. 2.01).
418

23. We graph the curve r = t
4
32t
3
32t
2
, j = t
3
3t in the viewing rectangle [32. 1.1] by [30.5. 0.5]. This rectangle
corresponds approximately to t M [31. 0.8].
We estimate that the curve has horizontal tangents at about (31. 30.4) and (30.17. 0.39) and vertical tangents at
about (0. 0) and (30.19. 0.37). We calculate
dj
dr
=
djdt
drdt
=
3t
2
31
4t
3
36t
2
34t
. The horizontal tangents occur when
djdt = 3t
2
31 = 0 C t =
1
I
3
, so both horizontal tangents are shown in our graph. The vertical tangents occur when
drdt = 2t(2t
2
33t 32) = 0 C 2t(2t +1)(t 32) = 0 C t = 0, 3
1
2
or 2. It seems that we have missed one vertical
tangent, and indeed if we plot the curve on the t-interval [31.2. 2.2] we see that there is another vertical tangent at (38. 6).
25. r = cos t, j = sint cos t. drdt = 3sint, djdt = 3sin
2
t +cos
2
t = cos 2t.
(r. j) = (0. 0) C cos t = 0 C t is an odd multiple of
r
2
. When t =
r
2
,
drdt = 31 and djdt = 31, so djdr = 1. When t =
3r
2
, drdt = 1 and
djdt = 31. So djdr = 31. Thus, j = r and j = 3r are both tangent to the
curve at (0. 0).
27. r = v0 3d sin0, j = v 3d cos 0.
(a)
dr
d0
= v 3d cos 0,
dj
d0
= d sin0, so
dj
dr
=
d sin0
v 3d cos 0
.
(b) If 0 < d < v, then |d cos 0| $ d < v, so v 3d cos 0 D v 3d 0. This shows that drd0 never vanishes,
so the trochoid can have no vertical tangent if d < v.
29. r = 2t
3
, j = 1 + 4t 3t
2
i
dj
dr
=
djdt
drdt
=
4 32t
6t
2
. Now solve
dj
dr
= 1 C
4 32t
6t
2
= 1 C
6t
2
+ 2t 34 = 0 C 2(3t 32)(t + 1) = 0 C t =
2
3
or t = 31. If t =
2
3
, the point is
16
27
.
29
9
, and if t = 31,
the point is (32. 34).
31. By symmetry of the ellipse about the r- and j-axes,
= 4
a
0
j dr = 4
0
r'2
/ sin0 (3osin0) d0 = 4o/
r'2
0
sin
2
0 d0 = 4o/
r'2
0
1
2
(1 3cos 20) d0
= 2o/
0 3
1
2
sin20
r'2
0
= 2o/
r
2
= o/

419
33. The curve r = 1 +c
I
, j = t 3t
2
= t(1 3t) intersects the r-axis when j = 0,
that is, when t = 0 and t = 1. The corresponding values of r are 2 and 1 +c.
The shaded area is given by
i=1+c
i=2
(j
T
3j
T
) dr =

I=1
I=0
[j(t) 30] r
0
(t) dt =

1
0
(t 3t
2
)c
I
dt
=

1
0
tc
I
dt 3
1
0
t
2
c
I
dt =

1
0
tc
I
dt 3
t
2
c
I
1
0
+ 2
1
0
tc
I
dt [Formula 97 or parts]
= 3
1
0
tc
I
dt 3(c 30) = 3
(t 31)c
I
1
0
3c [Formula 96 or parts]
= 3[0 3(31)] 3c = 3 3c
35. r = v0 3d sin0, j = v 3d cos 0.
=

2rr
0
j dr =

2r
0
(v 3d cos 0)(v 3d cos 0) d0 =

2r
0
(v
2
32dv cos 0 +d
2
cos
2
0) d0
=

v
2
0 32dv sin0 +
1
2
d
2
0 +
1
2
sin20
2r
0
= 2v
2
+d
2
37. r = t 3 t
2
, j =
4
3
t
3'2
, 1 $ t $ 2. drdt = 1 3 2t and djdt = 2t
1'2
, so
(drdt)
2
+ (djdt)
2
= (1 32t)
2
+ (2t
1'2
)
2
= 1 34t + 4t
2
+ 4t = 1 + 4t
2
.
Thus, 1 =

l
a
(drdt)
2
+ (djdt)
2
dt =

2
1
I
1 + 4t
2
dt E 3.1678.
39. r = t + cos t, j = t 3sint, 0 $ t $ 2. drdt = 1 3sint and djdt = 1 3cos t, so
ui
uI
2
+
uu
uI
2
= (1 3sint)
2
+ (1 3cos t)
2
= (1 32 sint + sin
2
t) + (1 32 cos t + cos
2
t) = 3 32 sint 32 cos t.
Thus, 1 =

l
a
(drdt)
2
+ (djdt)
2
dt =

2r
0
I
3 32 sint 32 cos t dt E 10.0367.
41. r = 1 + 3t
2
, j = 4 + 2t
3
, 0 $ t $ 1.
drdt = 6t and djdt = 6t
2
, so (drdt)
2
+ (djdt)
2
= 36t
2
+ 36t
4
.
Thus, 1 =

1
0
I
36t
2
+ 36t
4
dt =

1
0
6t
I
1 +t
2
dt
= 6
2
1
I
&

1
2
d&
[u = 1 + |
2
, Ju = 2| J|]
= 3
2
3
&
3'2
2
1
= 2(2
3'2
31) = 2
2
I
2 31
43. r =
t
1 +t
, j = ln(1 +t), 0 $ t $ 2.
dr
dt
=
(1 +t) 1 3t 1
(1 +t)
2
=
1
(1 +t)
2
and
dj
dt
=
1
1 +t
,
so
dr
dt
2
+
dj
dt
2
=
1
(1 +t)
4
+
1
(1 +t)
2
=
1
(1 +t)
4
1 + (1 +t)
2
=
t
2
+ 2t + 2
(1 +t)
4
. Thus,
1 =

2
0
I
t
2
+ 2t + 2
(1 +t)
2
dt =

3
1
I
&
2
+ 1
&
2
d&

u = | + 1,
Ju = J|
24
=

3
I
&
2
+ 1
&
+ ln
& +
&
2
+ 1
3
1
= 3
I
10
3
+ ln
3 +
I
10
+
I
2 3ln
1 +
I
2
420

45. r = c
I
cos t, j = c
I
sint, 0 $ t $ .
ui
uI
2
+
uu
uI
2
= [c
I
(cos t 3sint)]
2
+ [c
I
(sint + cos t)]
2
= (c
I
)
2
(cos
2
t 32 cos t sint + sin
2
t)
+ (c
I
)
2
(sin
2
t + 2 sint cos t + cos
2
t
= c
2I
(2 cos
2
t + 2 sin
2
t) = 2c
2I
Thus, 1 =

r
0
I
2c
2I
dt =

r
0
I
2 c
I
dt =
I
2

c
I
r
0
=
I
2 (c
r
31).
47. r = c
I
3t, j = 4c
I'2
, 38 $ t $ 3
ui
uI
2
+
uu
uI
2
= (c
I
31)
2
+ (2c
I'2
)
2
= c
2I
32c
I
+ 1 + 4c
I
= c
2I
+ 2c
I
+ 1 = (c
I
+ 1)
2
1 =

3
38
(c
I
+ 1)
2
dt =

3
38
(c
I
+ 1) dt =

c
I
+t
3I
38
= (c
3
+ 3) 3(c
38
38) = c
3
3c
38
+ 11
49. r = t 3 c
I
, j = t + c
I
, 36 $ t $ 6.
ui
uI
2
+
uu
uI
2
= (1 3c
I
)
2
+ (1 +c
I
)
2
= (1 32c
I
+c
2I
) + (1 + 2c
I
+c
2I
) = 2 + 2c
2I
, so 1 =

6
36
I
2 + 2c
2I
dt.
Set 1(t) =
I
2 + 2c
2I
. Then by Simpsons Rule with n = 6 and {t =
63(36)
6
= 2, we get
1 E
2
3
[1(36) + 41(34) + 21(32) + 41(0) + 21(2) + 41(4) +1(6)] E 612.3053.
51. r = sin
2
t, j = cos
2
t, 0 $ t $ 3.
(drdt)
2
+ (djdt)
2
= (2 sint cos t)
2
+ (32 cos t sint)
2
= 8 sin
2
t cos
2
t = 2 sin
2
2t i
Distance =

3r
0
I
2 |sin2t| dt = 6
I
2
r'2
0
sin2t dt [by symmetry] = 33
I
2
cos 2t
r'2
0
= 33
I
2 (31 31) = 6
I
2.
The full curve is traversed as t goes from 0 to
r
2
, because the curve is the segment of r +j = 1 that lies in the rst quadrant
(since r, j D 0), and this segment is completely traversed as t goes from 0 to
r
2
. Thus, 1 =

r'2
0
sin2t dt =
I
2, as above.
53. r = o sin0, j = / cos 0, 0 $ 0 $ 2.
ui
uI
2
+
uu
uI
2
= (o cos 0)
2
+ (3/ sin0)
2
= o
2
cos
2
0 +/
2
sin
2
0 = o
2
(1 3sin
2
0) +/
2
sin
2
0
= o
2
3(o
2
3/
2
) sin
2
0 = o
2
3c
2
sin
2
0 = o
2
1 3
c
2
o
2
sin
2
0
= o
2
(1 3c
2
sin
2
0)
So 1 = 4
r'2
0
o
2
1 3c
2
sin
2
0
d0 [by symmetry] = 4o
r'2
0
1 3c
2
sin
2
0 d0.
55. (a) r = 11 cos t 34 cos(11t2), j = 11 sint 34 sin(11t2).
Notice that 0 $ t $ 2 does not give the complete curve because
r(0) 6= r(2). In fact, we must take t M [0. 4] in order to obtain the
complete curve, since the rst term in each of the parametric equations has
period 2 and the second has period
2r
11'2
=
4r
11
, and the least common
integer multiple of these two numbers is 4.

421
(b) We use the CAS to nd the derivatives drdt and djdt, and then use Formula 1 to nd the arc length. Recent versions
of Maple express the integral
4r
0
(drdt)
2
+ (djdt)
2
dt as 881
2
I
2 j
, where 1(r) is the elliptic integral
1
0
I
1 3r
2
t
2
I
1 3t
2
dt and j is the imaginary number
I
31.
Some earlier versions of Maple (as well as Mathematica) cannot do the integral exactly, so we use the command
evalf(Int(sqrt(diff(x,t)2+diff(y,t)2),t=0..4*Pi)); to estimate the length, and nd that the arc
length is approximately 294.03. Derives Para_arc_length function in the utility le Int_apps simplies the
integral to 11
4r
0
34 cos t cos
11I
2
34 sint sin
11I
2
+ 5 dt.
57. r = 1 +tc
I
, j = (t
2
+ 1)c
I
, 0 $ t $ 1.
ui
uI
2
+
uu
uI
2
= (tc
I
+c
I
)
2
+ [(t
2
+ 1)c
I
+c
I
(2t)]
2
= [c
I
(t + 1)]
2
+ [c
I
(t
2
+ 2t + 1)]
2
= c
2I
(t + 1)
2
+c
2I
(t + 1)
4
= c
2I
(t + 1)
2
[1 + (t + 1)
2
], so
o =

2j dc =

1
0
2(t
2
+ 1)c
I
c
2I
(t + 1)
2
(t
2
+ 2t + 2) dt =

1
0
2(t
2
+ 1)c
2I
(t + 1)
I
t
2
+ 2t + 2 dt E 103.5999
59. r = t
3
, j = t
2
, 0 $ t $ 1.

ui
uI
2
+
uu
uI
2
=

3t
2
2
+ (2t)
2
= 9t
4
+ 4t
2
.
o =

1
0
2j
ui
uI
2
+
uu
uI
2
dt =

1
0
2t
2
9t
4
+ 4t
2
dt = 2
1
0
t
2
t
2
(9t
2
+ 4) dt
= 2
13
4
& 34
9
I
&
1
18
d&

u = 9|
2
+ 4, |
2
= (u 4)/9,
Ju = 18| J|, so | J| =
1
18
Ju
=
2
9 18
13
4
(&
3'2
34&
1'2
) d&
=
r
81
2
5
&
5'2
3
8
3
&
3'2
13
4
=
r
81

2
15
3&
5'2
320&
3'2
13
4
=
2r
1215
3 13
2
I
13 320 13
I
13
3(3 32 320 8)
=
2r
1215
247
I
13 + 64
61. r = o cos
3
0, j = o sin
3
0, 0 $ 0 $
r
2
.

ui
u0
2
+
uu
u0
2
= (33o cos
2
0 sin0)
2
+ (3o sin
2
0 cos 0)
2
= 9o
2
sin
2
0 cos
2
0.
o =

r'2
0
2 o sin
3
0 3o sin0 cos 0 d0 = 6o
2
r'2
0
sin
4
0 cos 0 d0 =
6
5
o
2
sin
5
0
r'2
0
=
6
5
o
2
63. r = t +t
3
, j = t 3
1
t
2
, 1 $ t $ 2.
ui
uI
= 1 + 3t
2
and
uu
uI
= 1 +
2
t
3
, so
ui
uI
2
+
uu
uI
2
= (1 + 3t
2
)
2
+
1 +
2
t
3
2
and o =

2j dc =

2
1
2
t 3
1
t
2
(1 + 3t
2
)
2
+
1 +
2
t
3
2
dt E 59.101.
65. r = 3t
2
, j = 2t
3
, 0 $ t $ 5 i

ui
uI
2
+
uu
uI
2
= (6t)
2
+ (6t
2
)
2
= 36t
2
(1 +t
2
) i
o =

5
0
2r
(drdt)
2
+ (djdt)
2
dt =

5
0
2(3t
2
)6t
I
1 +t
2
dt = 18
5
0
t
2
I
1 +t
2
2t dt
= 18
26
1
(& 31)
I
&d&

u = 1 + |
2
,
Ju = 2| J|
= 18
26
1
(&
3'2
3&
1'2
) d& = 18
2
5
&
5'2
3
2
3
&
3'2
26
1
= 18
2
5
676
I
26 3
2
3
26
I
26
2
5
3
2
3
=
24
5

949
I
26 + 1
422

67. If 1
0
is continuous and 1
0
(t) 6= 0 for o $ t $ /, then either 1
0
(t) 0 for all t in [o. /] or 1
0
(t) < 0 for all t in [o. /]. Thus, 1
is monotonic (in fact, strictly increasing or strictly decreasing) on [o. /]. It follows that 1 has an inverse. Set 1 = o 1
31
,
that is, dene 1 by 1(r) = o(1
31
(r)). Then r = 1(t) i 1
31
(r) = t, so j = o(t) = o(1
31
(r)) = 1(r).
69. (a) c = tan
31
dj
dr
i
dc
dt
=
d
dt
tan
31
dj
dr
=
1
1 + (djdr)
2
d
dt
dj
dr
. But
dj
dr
=
djdt
drdt
=
j
r
i
d
dt
dj
dr
=
d
dt
j
r
=
j r 3 r j
r
2
i
dc
dt
=
1
1 + ( j r)
2
j r 3 r j
r
2
=
r j 3 r j
r
2
+ j
2
. Using the Chain Rule, and the
fact that c =

I
0
ui
uI
2
+
uu
uI
2
dt i
us
uI
=

ui
uI
2
+
uu
uI
2
=

r
2
+ j
2
1'2
, we have that
dc
dc
=
dcdt
dcdt
=

r j 3 r j
r
2
+ j
2
1
( r
2
+ j
2
)
1'2
=
r j 3 r j
( r
2
+ j
2
)
3'2
. So r =
dc
dc
r j 3 r j
( r
2
+ j
2
)
3'2
=
| r j 3 r j|
( r
2
+ j
2
)
3'2
.
(b) r = r and j = 1(r) i r = 1, r = 0 and j =
dj
dr
, j =
d
2
j
dr
2
.
So r =
1 (d
2
jdr
2
) 30 (djdr)
[1 + (djdr)
2
]
3'2
=
d
2
jdr
2
[1 + (djdr)
2
]
3'2
.
71. r = 0 3sin0 i r = 1 3cos 0 i r = sin0, and j = 1 3cos 0 i j = sin0 i j = cos 0. Therefore,
r =
cos 0 3cos
2
0 3sin
2
0
[(1 3cos 0)
2
+ sin
2
0]
3'2
=
cos 0 3(cos
2
0 + sin
2
0)
(1 32 cos 0 + cos
2
0 + sin
2
0)
3'2
=
|cos 0 31|
(2 32 cos 0)
3'2
. The top of the arch is
characterized by a horizontal tangent, and from Example 2(b) in Section 10.2, the tangent is horizontal when 0 = (2n 31),
so take n = 1 and substitute 0 = into the expression for r: r =
|cos 31|
(2 32 cos )
3'2
=
|31 31|
[2 32(31)]
3'2
=
1
4
.
73. The coordinates of T are (v cos 0. v sin0). Since T1 was unwound from
arc T, T1 has length v0. Also _1TQ = _1T13_QT1 =
1
2
30,
so 1 has coordinates r = v cos 0 +v0 cos
1
2
30
= v(cos 0 +0 sin0),
j = v sin0 3v0 sin
1
2
30
= v(sin0 30 cos 0).

10.3 Polar Coordinates
1. (a)

2.
r
3
By adding 2 to
r
3
, we obtain the point
2.
7r
3
. The direction
opposite
r
3
is
4r
3
, so
32.
4r
3
is a point that satises the v < 0

requirement.
SECTION 10.3 POLAR COORDINATES

423
(b)

1. 3
3r
4
u A 0:

1. 3
3r
4
+ 2
=

1.
5r
4
u ? 0:

31. 3
3r
4
+
=

31.
r
4
(c)

31.
r
2
u A 0:
3(31).
r
2
+
=

1.
3r
2
u ? 0:
31.
r
2
+ 2
=

31.
5r
2
3. (a) r = 1 cos = 1(31) = 31 and

j = 1 sin = 1(0) = 0 give us
the Cartesian coordinates (31. 0).
(b) r = 2 cos
3
2r
3
= 2
3
1
2
= 31 and
j = 2 sin
3
2r
3
= 2
3
I
3
2
= 3
I
3
give us
31. 3
I
3
.
(c) r = 32 cos
3r
4
= 32
3
I
2
2
=
I
2 and
j = 32 sin
3r
4
= 32
I
2
2
= 3
I
2
gives us
I
2. 3
I
2
.
5. (a) r = 2 and j = 32 i v =

2
2
+ (32)
2
= 2
I
2 and 0 = tan
31
32
2
= 3
r
4
. Since (2. 32) is in the fourth
quadrant, the polar coordinates are (i)
2
I
2.
7r
4
and (ii)
32
I
2.
3r
4
.
(b) r = 31 and j =
I
3 i v =

(31)
2
+
I
3
2
= 2 and 0 = tan
31
I
3
31
=
2r
3
. Since
31.
I
3
is in the second
quadrant, the polar coordinates are (i)
2.
2r
3
and (ii)
32.
5r
3
.
424

7. The curves v = 1 and v = 2 represent circles with center
O and radii 1 and 2. The region in the plane satisfying
1 $ v $ 2 consists of both circles and the shaded region
between them in the gure.
9. The region satisfying 0 $ v < 4 and 32 $ 0 < 6
does not include the circle v = 4 nor the line 0 =
r
6
.
11. 2 < v < 3,
5r
3
$ 0 $
7r
3
13. Converting the polar coordinates (2. 3) and (4. 23) to Cartesian coordinates gives us
2 cos
r
3
. 2 sin
r
3
=

1.
I
3
and
4 cos
2r
3
. 4 sin
2r
3
=

32. 2
I
3
. Now use the distance formula.

d =

(r
2
3r
1
)
2
+ (j
2
3j
1
)
2
=

(32 31)
2
+
2
I
3 3
I
3
2
=
I
9 + 3 =
I
12 = 2
I
3
15. v = 2 C

r
2
+j
2
= 2 C r
2
+j
2
= 4, a circle of radius 2 centered at the origin.
17. v = 3 sin0 i v
2
= 3v sin0 C r
2
+j
2
= 3j C r
2
+
j 3
3
2
2
=

3
2
2
, a circle of radius
3
2
centered at
0.
3
2
.
The rst two equations are actually equivalent since v
2
= 3v sin0 i v(v 33 sin0) = 0 i v = 0 or v = 3 sin0. But
v = 3 sin0 gives the point v = 0 (the pole) when 0 = 0. Thus, the single equation v = 3 sin0 is equivalent to the compound
condition (v = 0 or v = 3 sin0).
19. v = csc 0 C v =
1
sin0
C v sin0 = 1 C j = 1, a horizontal line 1 unit above the r-axis.
21. r = 3 C v cos 0 = 3 C v = 3 cos 0 C v = 3 sec 0.
23. r = 3j
2
C v cos 0 = 3v
2
sin
2
0 C cos 0 = 3v sin
2
0 C v = 3
cos 0
sin
2
0
= 3cot 0 csc 0.
25. r
2
+j
2
= 2cr C v
2
= 2cv cos 0 C v
2
32cv cos 0 = 0 C v(v 32c cos 0) = 0 C v = 0 or v = 2c cos 0.
v = 0 is included in v = 2c cos 0 when 0 =
r
2
+n, so the curve is represented by the single equation v = 2c cos 0.
27. (a) The description leads immediately to the polar equation 0 =
r
6
, and the Cartesian equation j = tan
r
6
r =
1
I
3
r is
slightly more difcult to derive.
(b) The easier description here is the Cartesian equation r = 3.

425
29. 0 = 36 31. v = sin0 C v
2
= v sin0 C r
2
+j
2
= j C
r
2
+
j 3
1
2
2
=

1
2
2
. The reasoning here is the same
as in Exercise 17. This is a circle of radius
1
2
centered at
0.
1
2
.
33. v = 2(1 3sin0). This curve is a cardioid. 35. v = 0, 0 D 0
37. v = 4 sin30
39. v = 2 cos 40
41. v = 1 32 sin0
426

43. v
2
= 9 sin20
45. v = 2 cos
3
2
0
47. v = 1 + 2 cos 20
49. For 0 = 0, , and 2, v has its minimum value of about 0.5. For 0 =
r
2
and
3r
2
, v attains its maximum value of 2.
We see that the graph has a similar shape for 0 $ 0 $ and $ 0 $ 2.
51. r = (v) cos 0 = (4 + 2 sec 0) cos 0 = 4 cos 0 + 2. Now, v <" i
(4 + 2 sec 0) <" i 0 <
r
2
3
or 0 <
3r
2
+
[since we need only
consider 0 $ 0 < 2], so lim
r<"
r = lim
0<r'2
(4 cos 0 + 2) = 2. Also,
v <3" i (4 + 2 sec 0) <3" i 0 <
r
2
+
or 0 <
3r
2
3
, so
lim
r<3"
r = lim
0<r'2
+
(4 cos 0 + 2) = 2. Therefore, lim
r<"
r = 2 i r = 2 is a vertical asymptote.

427
53. To show that r = 1 is an asymptote we must prove lim
r<"
r = 1.
r = (v) cos 0 = (sin0 tan0) cos 0 = sin
2
0. Now, v <" i sin0 tan0 <" i
0 <
r
2
3
, so lim
r<"
r = lim
0<r'2
sin
2
0 = 1. Also, v <3" i sin0 tan0 <3" i
0 <
r
2
+
, so lim
r<3"
r = lim
0<r'2
+
sin
2
0 = 1. Therefore, lim
r<"
r = 1 i r = 1 is
a vertical asymptote. Also notice that r = sin
2
0 D 0 for all 0, and r = sin
2
0 $ 1 for all 0. And r 6= 1, since the curve is not
dened at odd multiples of
r
2
. Therefore, the curve lies entirely within the vertical strip 0 $ r < 1.
55. (a) We see that the curve v = 1 +c sin0 crosses itself at the origin, where v = 0 (in fact the inner loop corresponds to
negative v-values,) so we solve the equation of the limaon for v = 0 C c sin0 = 31 C sin0 = 31c. Now if
|c| < 1, then this equation has no solution and hence there is no inner loop. But if c < 31, then on the interval (0. 2)
the equation has the two solutions 0 = sin
31
(31c) and 0 = 3sin
31
(31c), and if c 1, the solutions are
0 = + sin
31
(1c) and 0 = 2 3sin
31
(1c). In each case, v < 0 for 0 between the two solutions, indicating a loop.
(b) For 0 < c < 1, the dimple (if it exists) is characterized by the fact that j has a local maximum at 0 =
3r
2
. So we determine
for what c-values
d
2
j
d0
2
is negative at 0 =
3r
2
, since by the Second Derivative Test this indicates a maximum:
j = v sin0 = sin0 +c sin
2
0 i
dj
d0
= cos 0 + 2c sin0 cos 0 = cos 0 +c sin20 i
d
2
j
d0
2
= 3sin0 + 2c cos 20.
At 0 =
3r
2
, this is equal to 3(31) + 2c(31) = 1 32c, which is negative only for c
1
2
. A similar argument shows that
for 31 < c < 0, j only has a local minimum at 0 =
r
2
(indicating a dimple) for c < 3
1
2
.
57. v = 2 sin0 i r = v cos 0 = 2 sin0 cos 0 = sin20, j = v sin0 = 2 sin
2
0 i
dj
dr
=
djd0
drd0
=
2 2 sin0 cos 0
cos 20 2
=
sin20
cos 20
= tan20
When 0 =

6
,
dj
dr
= tan
2

6
= tan

3
=
I
3. [Another method: Use Equation 3.]
59. v = 10 i r = v cos 0 = (cos 0)0, j = v sin0 = (sin0)0 i
dj
dr
=
djd0
drd0
=
sin0(310
2
) + (10) cos 0
cos 0(310
2
) 3(10) sin0

0
2
0
2
=
3sin0 +0 cos 0
3cos 0 30 sin0
When 0 = ,
dj
dr
=
30 +(31)
3(31) 3(0)
=
3
1
= 3.
61. v = cos 20 i r = v cos 0 = cos 20 cos 0, j = v sin0 = cos 20 sin0 i
dj
dr
=
djd0
drd0
=
cos 20 cos 0 + sin0 (32 sin20)
cos 20 (3sin0) + cos 0 (32 sin20)
When 0 =

4
,
dj
dr
=
0
I
22
+
I
22
(32)
0
3
I
22
+
I
22
(32)
=
3
I
2
3
I
2
= 1.
428

63. v = 3 cos 0 i r = v cos 0 = 3 cos 0 cos 0, j = v sin0 = 3 cos 0 sin0 i
uu
u0
= 33 sin
2
0 + 3 cos
2
0 = 3 cos 20 = 0 i 20 =
r
2
or
3r
2
C 0 =
r
4
or
3r
4
.
So the tangent is horizontal at
3
I
2
.
r
4
and
3
3
I
2
.
3r
4

same as
3
I
2
. 3
r
4
.
ui
u0
= 36 sin0 cos 0 = 33 sin20 = 0 i 20 = 0 or C 0 = 0 or
r
2
. So the tangent is vertical at (3. 0) and
0.
r
2
.
65. v = 1 + cos 0 i r = v cos 0 = cos 0 (1 + cos 0), j = v sin0 = sin0 (1 + cos 0) i
uu
u0
= (1 + cos 0) cos 0 3sin
2
0 = 2 cos
2
0 + cos 0 31 = (2 cos 0 31)(cos 0 + 1) = 0 i cos 0 =
1
2
or 31 i
0 =
r
3
, , or
5r
3
i horizontal tangent at
3
2
.
r
3
, (0. ), and
3
2
.
5r
3
.
ui
u0
= 3(1 + cos 0) sin0 3cos 0 sin0 = 3sin0 (1 + 2 cos 0) = 0 i sin0 = 0 or cos 0 = 3
1
2
i
0 = 0, ,
2r
3
, or
4r
3
i vertical tangent at (2. 0),
1
2
.
2r
3
, and
1
2
.
4r
3
.
Note that the tangent is horizontal, not vertical when 0 = , since lim
0<r
djd0
drd0
= 0.
67. v = 2 + sin0 i r = v cos 0 = (2 + sin0) cos 0, j = v sin0 = (2 + sin0) sin0 i
uu
u0
= (2 + sin0) cos 0 + sin0 cos 0 = cos 0 2(1 + sin0) = 0 i cos 0 = 0 or sin0 = 31 i
0 =
r
2
or
3r
2
i horizontal tangent at
3.
r
2
and
1.
3r
2
.
ui
u0
= (2 + sin0)(3sin0) + cos 0 cos 0 = 32 sin0 3sin
2
0 + 1 3sin
2
0 = 32 sin
2
0 32 sin0 + 1 i
sin0 =
2
I
4 + 8
34
=
2 2
I
3
34
=
1 3
I
3
32
1 +
I
3
32
< 31
i
0
1
= sin
31
3
1
2
+
1
2
I
3
and 0
2
= 30
1
i vertical tangent at
3
2
+
1
2
I
3. 0
1
and
3
2
+
1
2
I
3. 0
2
.
Note that v(0
1
) = 2 + sin
sin
31
3
1
2
+
1
2
I
3
= 2 3
1
2
+
1
2
I
3 =
3
2
+
1
2
I
3.
69. v = o sin0 + / cos 0 i v
2
= ov sin0 +/v cos 0 i r
2
+j
2
= oj +/r i
r
2
3/r +
1
2
/
2
+j
2
3oj +
1
2
o
2
=

1
2
/
2
+
1
2
o
2
i

r 3
1
2
/
2
+
j 3
1
2
o
2
=
1
4
(o
2
+/
2
), and this is a circle
with center
1
2
/.
1
2
o
and radius
1
2
I
o
2
+/
2
.
Note for Exercises 7176: Maple is able to plot polar curves using the polarplot command, or using the coords=polar option in a regular
plot command. In Mathematica, use PolarPlot. In Derive, change to Polar under Options State. If your graphing device cannot
plot polar equations, you must convert to parametric equations. For example, in Exercise 71, r = v cos 0 = [1 + 2 sin(02)] cos 0,
j = v sin0 = [1 + 2 sin(02)] sin0.
71. v = 1 + 2 sin(02). The parameter interval is [0. 4]. 73. v = c
sin 0
32 cos(40). The parameter interval is [0. 2].

429
75. v = 2 35 sin(06). The parameter interval is [36. 6].
77. It appears that the graph of v = 1 + sin
0 3
r
6
is the same shape as

the graph of v = 1 + sin0, but rotated counterclockwise about the
origin by
r
6
. Similarly, the graph of v = 1 + sin
0 3
r
3
is rotated by
r
3
. In general, the graph of v = 1(0 3c) is the same shape as that of
v = 1(0), but rotated counterclockwise through c about the origin.
That is, for any point (v
0
. 0
0
) on the curve v = 1(0), the point
(v
0
. 0
0
+c) is on the curve v = 1(0 3c), since v
0
= 1(0
0
) = 1((0
0
+c) 3c).
79. (a) v = sinn0.
n = 2 n = 3 n = 4 n = 5
From the graphs, it seems that when n is even, the number of loops in the curve (called a rose) is 2n, and when n is odd,
the number of loops is simply n. This is because in the case of n odd, every point on the graph is traversed twice, due to
the fact that
v(0 +) = sin[n(0 +)] = sinn0 cos n + cos n0 sinn =
sinn0 if n is even
3sinn0 if n is odd
(b) The graph of v = |sinn0| has 2n loops whether n is odd or even, since v(0 +) = v(0).
n = 2 n = 3 n = 4 n = 5
430

81. v =
1 3o cos 0
1 +o cos 0
. We start with o = 0, since in this case the curve is simply the circle v = 1.
As o increases, the graph moves to the left, and its right side becomes attened. As o increases through about 0.4, the right
side seems to grow a dimple, which upon closer investigation (with narrower 0-ranges) seems to appear at o E 0.42 [the
actual value is
I
2 31]. As o <1, this dimple becomes more pronounced, and the curve begins to stretch out horizontally,
until at o = 1 the denominator vanishes at 0 = , and the dimple becomes an actual cusp. For o 1 we must choose our
parameter interval carefully, since v <"as 1 +o cos 0 <0 C 0 <cos
31
(31o). As o increases from 1, the curve
splits into two parts. The left part has a loop, which grows larger as o increases, and the right part grows broader vertically,
and its left tip develops a dimple when o E 2.42 [actually,
I
2 + 1]. As o increases, the dimple grows more and more
pronounced. If o < 0, we get the same graph as we do for the corresponding positive o-value, but with a rotation through
about the pole, as happened when c was replaced with 3c in Exercise 80.
o = 0 o = 0.3 o = 0.41. |0| $ 0.5
o = 0.42,|0| $ 0.5 o = 0.9. |0| $ 0.5 o = 1. |0| $ 0.1
o = 2
o = 2.41, |0 3| $ 0.2
o = 2.42, |0 3| $ 0.2
o = 4
SECTION 10.4 AREAS AND LENGTHS IN POLAR COORDINATES

431
83. tanw = tan(c 30) =
tanc 3tan0
1 + tanc tan0
=
dj
dr
3tan0
1 +
dj
dr
tan0
=
djd0
drd0
3tan0
1 +
djd0
drd0
tan0
=
dj
d0
3
dr
d0
tan0
dr
d0
+
dj
d0
tan0
=
dv
d0
sin0 +v cos 0
3tan0
dv
d0
cos 0 3v sin0
dv
d0
cos 0 3v sin0
+ tan0
dv
d0
sin0 +v cos 0
=
v cos 0 +v
sin
2
0
cos 0
dv
d0
cos 0 +
dv
d0

sin
2
0
cos 0
=
v cos
2
0 +v sin
2
0
dv
d0
cos
2
0 +
dv
d0
sin
2
0
=
v
dvd0
10.4 Areas and Lengths in Polar Coordinates
1. v = 0
2
, 0 $ 0 $
r
4
. =

r'4
0
1
2
v
2
d0 =

r'4
0
1
2
(0
2
)
2
d0 =

r'4
0
1
2
0
4
d0 =

1
10
0
5
r'4
0
=
1
10
r
4
5
=
1
10,240
5
3. v = sin0,
r
3
$ 0 $
2r
3
.
=

2r'3
r'3
1
2
sin
2
0 d0 =
1
4
2r'3
r'3
(1 3cos 20) d0 =
1
4
0 3
1
2
sin20
2r'3
r'3
=
1
4
2r
3
3
1
2
sin
4r
3
3
r
3
+
1
2
sin
2r
3
=
1
4
2r
3
3
1
2
3
I
3
2
3
r
3
+
1
2
I
3
2
=
1
4
r
3
+
I
3
2
=
r
12
+
I
3
8
5. v =
I
0, 0 $ 0 $ 2. =

2r
0
1
2
v
2
d0 =

2r
0
1
2
I
0
2
d0 =

2r
0
1
2
0 d0 =

1
4
0
2
2r
0
=
2
7. v = 4 + 3 sin0, 3
r
2
$ 0 $
r
2
.
=

r'2
3r'2
1
2
((4 + 3 sin0)
2
d0 =
1
2
r'2
3r'2
(16 + 24 sin0 + 9 sin
2
0) d0 =
1
2
r'2
3r'2
(16 + 9 sin
2
0) d0 [by Theorem 5.5.7(b)]
=
1
2
2
r'2
0
16 + 9
1
2
(1 3cos 20)
d0 [by Theorem 5.5.7(a)]

=

r'2
0
41
2
3
9
2
cos 20
d0 =

41
2
0 3
9
4
sin20
r'2
0
=

41r
4
30
3(0 30) =
41r
4
9. The area above the polar axis is bounded by v = 3 cos 0 for 0 = 0
to 0 = 2 [not ]. By symmetry,
= 2
r'2
0
1
2
v
2
d0 =

r'2
0
(3 cos 0)
2
d0 = 3
2
r'2
0
cos
2
0 d0
= 9
r'2
0
1
2
(1 + cos 20) d0 =
9
2
0 +
1
2
sin20
r'2
0
=
9
2
r
2
+ 0
3(0 + 0)
=
9r
4
Also, note that this is a circle with radius
3
2
, so its area is
3
2
2
=
9r
4
.
11. The curve goes through the pole when 0 = 4, so well nd the area for
0 $ 0 $ 4 and multiply it by 4.
= 4
r'4
0
1
2
v
2
d0 = 2
r'4
0
(4 cos 20) d0
= 8
r'4
0
cos 20 d0 = 4
sin20
r'4
0
= 4
432

13. One-sixth of the area lies above the polar axis and is bounded by the curve
v = 2 cos 30 for 0 = 0 to 0 = 6.
= 6
r'6
0
1
2
(2 cos 30)
2
d0 = 12
r'6
0
cos
2
30 d0
=
12
2
r'6
0
(1 + cos 60) d0
= 6
0 +
1
6
sin60
r'6
0
= 6
r
6
=
15. =

2r
0
1
2
(1 + 2 sin60)
2
d0 =
1
2
2r
0
(1 + 4 sin60 + 4 sin
2
60) d0
=
1
2
2r
0
1 + 4 sin60 + 4
1
2
(1 3cos 120)
d0
=
1
2
2r
0
(3 + 4 sin60 32 cos 120) d0
=
1
2
30 3
2
3
cos 60 3
1
6
sin120
2r
0
=
1
2
(6 3
2
3
30) 3
0 3
2
3
30
= 3
17. The shaded loop is traced out from 0 = 0 to 0 = 2.
=

r'2
0
1
2
v
2
d0 =
1
2
r'2
0
sin
2
20 d0
=
1
2
r'2
0
1
2
(1 3cos 40) d0 =
1
4
0 3
1
4
sin40
r'2
0
=
1
4
r
2
=
r
8
19. v = 0 i 3 cos 50 = 0 i 50 =
r
2
i 0 =
r
10
.
=

r'10
3r'10
1
2
(3 cos 50)
2
d0 =

r'10
0
9 cos
2
50 d0 =
9
2
r'10
0
(1 + cos 100) d0 =
9
2
0 +
1
10
sin100
r'10
0
=
9r
20
21. This is a limaon, with inner loop traced
out between 0 =
7r
6
and
11r
6
[found by
solving v = 0].
= 2
3r'2
7r'6
1
2
(1 + 2 sin0)
2
d0 =

3r'2
7r'6
1 + 4 sin0 + 4 sin
2
0
d0 =

3r'2
7r'6
1 + 4 sin0 + 4
1
2
(1 3cos 20)
d0
=

0 34 cos 0 + 20 3sin20
3r'2
7r'6
=

9r
2
7r
2
+ 2
I
3 3
I
3
2
= 3
3
I
3
2
23. 2 cos 0 = 1 i cos 0 =
1
2
i 0 =
r
3
or
5r
3
.
= 2
r'3
0
1
2
[(2 cos 0)
2
31
2
] d0 =

r'3
0
(4 cos
2
0 31) d0
=

r'3
0
1
2
(1 + cos 20)
31
d0 =

r'3
0
(1 + 2 cos 20) d0
=

0 + sin20
r'3
0
=
r
3
+
I
3
2

433
25. To nd the area inside the leminiscate v
2
= 8 cos 20 and outside the circle v = 2,
we rst note that the two curves intersect when v
2
= 8 cos 20 and v = 2,
that is, when cos 20 =
1
2
. For 3 < 0 $ , cos 20 =
1
2
C 20 = 3
or 53 C 0 = 6 or 56. The gure shows that the desired area is
4 times the area between the curves from 0 to 6. Thus,
= 4
r'6
0
1
2
(8 cos 20) 3
1
2
(2)
2
d0 = 8
r'6
0
(2 cos 20 31) d0
= 8
sin20 30
r'6
0
= 8
I
32 36
= 4
I
3 343
27. 3 cos 0 = 1 + cos 0 C cos 0 =
1
2
i 0 =
r
3
or 3
r
3
.
= 2
r'3
0
1
2
[(3 cos 0)
2
3(1 + cos 0)
2
] d0
=

r'3
0
(8 cos
2
0 32 cos 0 31) d0 =

r'3
0
[4(1 + cos 20) 32 cos 0 31] d0
=

r'3
0
(3 + 4 cos 20 32 cos 0) d0 =

30 + 2 sin20 32 sin0
r'3
0
= +
I
3 3
I
3 =
29.
I
3 cos 0 = sin0 i
I
3 =
sin0
cos 0
i tan0 =
I
3 i 0 =
r
3
.
=

r'3
0
1
2
(sin0)
2
d0 +
r'2
r'3
1
2
I
3 cos 0
2
d0
=

r'3
0
1
2

1
2
(1 3cos 20) d0 +
r'2
r'3
1
2
3
1
2
(1 + cos 20) d0
=
1
4
0 3
1
2
sin20
r'3
0
+
3
4
0 +
1
2
sin20
r'2
r'3
=
1
4
r
3
3
I
3
4
30
+
3
4
r
2
+ 0
r
3
+
I
3
4
=
r
12
3
I
3
16
+
r
8
3
3
I
3
16
=
5r
24
3
I
3
4
31. sin20 = cos 20 i
sin20
cos 20
= 1 i tan20 = 1 i 20 =
r
4
i
0 =
r
8
i
= 8 2
r'8
0
1
2
sin
2
20 d0 = 8
r'8
0
1
2
(1 3cos 40) d0
= 4
0 3
1
4
sin40
r'8
0
= 4
r
8
3
1
4
1
=
r
2
31
33. sin20 = cos 20 i tan20 = 1 i 20 =
r
4
i 0 =
r
8
= 4
r'8
0
1
2
sin20 d0 [since v
2
= sin20]
=

r'8
0
2 sin20 d0 =

3cos 20
r'8
0
= 3
1
2
I
2 3(31) = 1 3
1
2
I
2
434

35. The darker shaded region (from 0 = 0 to 0 = 23) represents
1
2
of the desired area plus
1
2
of the area of the inner loop.
From this area, well subtract
1
2
of the area of the inner loop (the lighter shaded region from0 = 23 to 0 = ), and then
double that difference to obtain the desired area.
= 2
2r'3
0
1
2
1
2
+ cos 0
2
d0 3
r
2r'3
1
2
1
2
+ cos 0
2
d0
=

2r'3
0
1
4
+ cos 0 + cos
2
0
d0 3
r
2r'3
1
4
+ cos 0 + cos
2
0
d0
=

2r'3
0
1
4
+ cos 0 +
1
2
(1 + cos 20)
d0
3

r
2r'3
1
4
+ cos 0 +
1
2
(1 + cos 20)
d0
=

0
4
+ sin0 +
0
2
+
sin20
4
2r'3
0
3
0
4
+ sin0 +
0
2
+
sin20
4
r
2r'3
=

r
6
+
I
3
2
+
r
3
3
I
3
8
r
4
+
r
2
r
6
+
I
3
2
+
r
3
3
I
3
8
=
r
4
+
3
4
I
3 =
1
4
+ 3
I
3
37. The pole is a point of intersection.

1 + sin0 = 3 sin0 i 1 = 2 sin0 i sin0 =
1
2
i
0 =
r
6
or
5r
6
.
The other two points of intersection are
3
2
.
r
6
and
3
2
.
5r
6
.
39. 2 sin20 = 1 i sin20 =
1
2
i 20 =
r
6
,
5r
6
,
13r
6
, or
17r
6
.
By symmetry, the eight points of intersection are given by
(1. 0), where 0 =
r
12
,
5r
12
,
13r
12
, and
17r
12
, and
(31. 0), where 0 =
7r
12
,
11r
12
,
19r
12
, and
23r
12
.
[There are many ways to describe these points.]
41. The pole is a point of intersection. sin0 = sin20 = 2 sin0 cos 0 C
sin0 (1 32 cos 0) = 0 C sin0 = 0 or cos 0 =
1
2
i
0 = 0, ,
r
3
, or 3
r
3
i the other intersection points are
I
3
2
.
r
3
and
I
3
2
.
2r
3
[by symmetry].

435
43.
From the rst graph, we see that the pole is one point of intersection. By zooming in or using the cursor, we nd the 0-values
of the intersection points to be c E 0.88786 E 0.89 and 3c E 2.25. (The rst of these values may be more easily
estimated by plotting j = 1 + sinr and j = 2r in rectangular coordinates; see the second graph.) By symmetry, the total
area contained is twice the area contained in the rst quadrant, that is,
= 2
o
0
1
2
(20)
2
d0 + 2
r'2
o
1
2
(1 + sin0)
2
d0 =

o
0
40
2
d0 +
r'2
o
1 + 2 sin0 +
1
2
(1 3cos 20)
d0
=

4
3
0
3
o
0
+
0 32 cos 0 +
1
2
0 3
1
4
sin20
r'2
o
=
4
3
c
3
+
r
2
+
r
4
c 32 cos c +
1
2
c 3
1
4
sin2c
E 3.4645
45. 1 =

l
a
v
2
+ (dvd0)
2
d0 =

r'3
0
(3 sin0)
2
+ (3 cos 0)
2
d0 =

r'3
0
9(sin
2
0 + cos
2
0) d0
= 3
r'3
0
d0 = 3
r'3
0
= 3
r
3
= .
As a check, note that the circumference of a circle with radius
3
2
is 2
3
2
= 3, and since 0 = 0 to =
r
3
traces out
1
3
of the
circle (from 0 = 0 to 0 = ),
1
3
(3) = .
47. 1 =

l
a
v
2
+ (dvd0)
2
d0 =

2r
0
(0
2
)
2
+ (20)
2
d0 =

2r
0
0
4
+ 40
2
d0
=

2r
0
0
2
(0
2
+ 4) d0 =

2r
0
0
0
2
+ 4 d0
Now let & = 0
2
+ 4, so that d& = 20 d0

0 d0 =
1
2
d&
and
2r
0
0
0
2
+ 4 d0 =

4r
2
+4
4
1
2
I
&d& =
1
2

2
3
&
3'2
4(r
2
+1)
4
=
1
3
[4
3'2
(
2
+ 1)
3'2
34
3'2
] =
8
3
[(
2
+ 1)
3'2
31]
49. The curve v = 3 sin20 is completely traced with 0 $ 0 $ 2. v
2
+
ur
u0
2
= (3 sin20)
2
+ (6 cos 20)
2
i
1 =

2r
0
9 sin
2
20 + 36 cos
2
20 d0 E 29.0653
436

51. The curve v = sin
0
2
is completely traced with 0 $ 0 $ 4. v

2
+
ur
u0
2
= sin
2
0
2
1
2
cos
0
2
2
i
1 =

4r
0
sin
2
0
2
+
1
4
cos
2
0
2
d0 E 9.6884
53. The curve v = cos
4
(04) is completely traced with 0 $ 0 $ 4.
v
2
+ (dvd0)
2
= [cos
4
(04)]
2
+
4 cos
3
(04) (3sin(04))
1
4
2
= cos
8
(04) + cos
6
(04) sin
2
(04)
= cos
6
(04)[cos
2
(04) + sin
2
(04)] = cos
6
(04)
1 =

4r
0
cos
6
(04) d0 =

4r
0
cos
3
(04)
d0
= 2
2r
0
cos
3
(04) d0 [since cos
3
(04) D 0 for 0 $ 0 $ 2] = 8
r'2
0
cos
3
&d&

& =
1
4
0
68
= 8
1
3
(2 + cos
2
&) sin&
r'2
0
=
8
3
[(2 1) 3(3 0)] =
16
3
55. (a) From (10.2.7),
o =

l
a
2j
(drd0)
2
+ (djd0)
2
d0
=

l
a
2j
v
2
+ (dvd0)
2
d0 [from the derivation of Equation 10.4.5]
=

l
a
2v sin0
v
2
+ (dvd0)
2
d0
(b) The curve v
2
= cos 20 goes through the pole when cos 20 = 0 i
20 =
r
2
i 0 =
r
4
. Well rotate the curve from 0 = 0 to 0 =
r
4
and double
this value to obtain the total surface area generated.
v
2
= cos 20 i 2v
dv
d0
= 32 sin20 i

dv
d0
2
=
sin
2
20
v
2
=
sin
2
20
cos 20
.
o = 2
r'4
0
2
I
cos 20 sin0
cos 20 +
sin
2
20
cos 20 d0 = 4
r'4
0
I
cos 20 sin0
cos
2
20 + sin
2
20
cos 20
d0
= 4
r'4
0
I
cos 20 sin0
1
I
cos 20
d0 = 4
r'4
0
sin0 d0 = 4
3cos 0
r'4
0
= 34
I
2
2
31
= 2
2 3
I
2
SECTION 10.5 CONIC SECTIONS

437
10.5 Conic Sections
1. r = 2j
2
i j
2
=
1
2
r. 4j =
1
2
, so j =
1
8
. The vertex
is (0. 0), the focus is
1
8
. 0
, and the directrix is r = 3

1
8
.
3. 4r
2
= 3j i r
2
= 3
1
4
j. 4j = 3
1
4
, so j = 3
1
16
.
The vertex is (0. 0), the focus is
0. 3
1
16
, and the
directrix is j =
1
16
.
5. (r + 2)
2
= 8 (j 33). 4j = 8, so j = 2. The vertex is
(32. 3), the focus is (32. 5), and the directrix is j = 1.
7. j
2
+ 2j + 12r + 25 = 0 i
j
2
+ 2j + 1 = 312r 324 i
(j + 1)
2
= 312(r + 2). 4j = 312, so j = 33.
The vertex is (32. 31), the focus is (35. 31), and the
directrix is r = 1.
9. The equation has the formj
2
= 4jr, where j < 0.
Since the parabola passes through (31. 1), we have
1
2
= 4j(31), so 4j = 31 and an equation is j
2
= 3r
or r = 3j
2
. 4j = 31, so j = 3
1
4
and the focus is
3
1
4
. 0
while the directrix is r =

1
4
.
11.
r
2
9
+
j
2
5
= 1 i o =
I
9 = 3, / =
I
5,
c =
I
o
2
3/
2
=
I
9 35 = 2. The ellipse is centered at
(0. 0), with vertices at (3. 0). The foci are (2. 0).
438

13. 4r
2
+j
2
= 16 i
r
2
4
+
j
2
16
= 1 i
o =
I
16 = 4, / =
I
4 = 2,
c =
I
o
2
3/
2
=
I
16 34 = 2
I
3. The ellipse is
centered at (0. 0), with vertices at (0. 4). The foci
are
0. 2
I
3
.
15. 9r
2
318r + 4j
2
= 27 C
9(r
2
32r + 1) + 4j
2
= 27 + 9 C
9(r 31)
2
+ 4j
2
= 36 C
(r 31)
2
4
+
j
2
9
= 1 i
o = 3, / = 2, c =
I
5 i center (1. 0),
vertices (1. 3), foci
1.
I
5
17. The center is (0. 0), o = 3, and / = 2, so an equation is

r
2
4
+
j
2
9
= 1. c =
I
o
2
3/
2
=
I
5, so the foci are
0.
I
5
.
19.
r
2
144
3
j
2
25
= 1 i o = 12, / = 5, c =
I
144 + 25 = 13 i
center (0. 0), vertices (12. 0), foci (13. 0), asymptotes j =
5
12
r.
Note: It is helpful to draw a 2o-by-2/ rectangle whose center is the center
of the hyperbola. The asymptotes are the extended diagonals of the
rectangle.
21. j
2
3r
2
= 4 C
j
2
4
3
r
2
4
= 1 i o =
I
4 = 2 = /,
c =
I
4 + 4 = 2
I
2 i center (0. 0), vertices (0. 2),
foci
0. 2
I
2
, asymptotes j = r
23. 4r
2
3j
2
324r 34j + 28 = 0 C
4(r
2
36r + 9) 3(j
2
+ 4j + 4) = 328 + 36 34 C
4(r 33)
2
3(j + 2)
2
= 4 C
(r 33)
2
1
3
(j + 2)
2
4
= 1 i
o =
I
1 = 1, / =
I
4 = 2, c =
I
1 + 4 =
I
5 i
center (3. 32), vertices (4. 32) and (2. 32), foci
3
I
5. 32
,
asymptotes j + 2 = 2(r 33).

439
25. r
2
= j + 1 C r
2
= 1(j + 1). This is an equation of a parabola with 4j = 1, so j =
1
4
. The vertex is (0. 31) and the
focus is
0. 3
3
4
.
27. r
2
= 4j 32j
2
C r
2
+ 2j
2
34j = 0 C r
2
+ 2(j
2
32j + 1) = 2 C r
2
+ 2(j 31)
2
= 2 C
r
2
2
+
(j 31)
2
1
= 1. This is an equation of an ellipse with vertices at
I
2. 1
. The foci are at
I
2 31. 1
= (1. 1).
29. j
2
+2j = 4r
2
+3 C j
2
+2j +1 = 4r
2
+4 C (j +1)
2
34r
2
= 4 C
(j + 1)
2
4
3r
2
= 1. This is an equation
of a hyperbola with vertices (0. 31 2) = (0. 1) and (0. 33). The foci are at
0. 31
I
4 + 1
=

0. 31
I
5
.
31. The parabola with vertex (0. 0) and focus (0. 32) opens downward and has j = 32, so its equation is r
2
= 4jj = 38j.
33. The distance from the focus (34. 0) to the directrix r = 2 is 2 3(34) = 6, so the distance from the focus to the vertex is
1
2
(6) = 3 and the vertex is (31. 0). Since the focus is to the left of the vertex, j = 33. An equation is j
2
= 4j(r + 1) i
j
2
= 312(r + 1).
35. A parabola with vertical axis and vertex (2. 3) has equation j 33 = o(r 32)
2
. Since it passes through (1. 5), we have
5 33 = o(1 32)
2
i o = 2, so an equation is j 33 = 2(r 32)
2
.
37. The ellipse with foci (2. 0) and vertices (5. 0) has center (0. 0) and a horizontal major axis, with o = 5 and c = 2,
so /
2
= o
2
3c
2
= 25 34 = 21. An equation is
r
2
25
+
j
2
21
= 1.
39. Since the vertices are (0. 0) and (0. 8), the ellipse has center (0. 4) with a vertical axis and o = 4. The foci at (0. 2) and (0. 6)
are 2 units from the center, so c = 2 and / =
I
o
2
3c
2
=
I
4
2
32
2
=
I
12. An equation is
(r 30)
2
/
2
+
(j 34)
2
o
2
= 1 i
r
2
12
+
(j 34)
2
16
= 1.
41. An equation of an ellipse with center (31. 4) and vertex (31. 0) is
(r + 1)
2
/
2
+
(j 34)
2
4
2
= 1. The focus (31. 6) is 2 units
from the center, so c = 2. Thus, /
2
+ 2
2
= 4
2
i /
2
= 12, and the equation is
(r + 1)
2
12
+
(j 34)
2
16
= 1.
43. An equation of a hyperbola with vertices (3. 0) is
r
2
3
2
3
j
2
/
2
= 1. Foci (5. 0) i c = 5 and 3
2
+/
2
= 5
2
i
/
2
= 25 39 = 16, so the equation is
r
2
9
3
j
2
16
= 1.
45. The center of a hyperbola with vertices (33. 34) and (33. 6) is (33. 1), so o = 5 and an equation is
(j 31)
2
5
2
3
(r + 3)
2
/
2
= 1. Foci (33. 37) and (33. 9) i c = 8, so 5
2
+/
2
= 8
2
i /
2
= 64 325 = 39 and the
equation is
(j 31)
2
25
3
(r + 3)
2
39
= 1.
440

47. The center of a hyperbola with vertices (3. 0) is (0. 0), so o = 3 and an equation is
r
2
3
2
3
j
2
/
2
= 1.
Asymptotes j = 2r i
/
o
= 2 i / = 2(3) = 6 and the equation is
r
2
9
3
j
2
36
= 1.
49. In Figure 8, we see that the point on the ellipse closest to a focus is the closer vertex (which is a distance
o 3c from it) while the farthest point is the other vertex (at a distance of o +c). So for this lunar orbit,
(o 3c) + (o +c) = 2o = (1728 + 110) + (1728 + 314), or o = 1940; and (o +c) 3(o 3c) = 2c = 314 3110,
or c = 102. Thus, /
2
= o
2
3c
2
= 3,753,196, and the equation is
r
2
3,763,600
+
j
2
3,753,196
= 1.
51. (a) Set up the coordinate system so that is (3200. 0) and 1 is (200. 0).
|1| 3|11| = (1200)(980) = 1,176,000 ft =
2450
11
mi = 2o i o =
1225
11
, and c = 200 so
/
2
= c
2
3o
2
=
3,339,375
121
i
121r
2
1,500,625
3
121j
2
3,339,375
= 1.
(b) Due north of 1 i r = 200 i
(121)(200)
2
1,500,625
3
121j
2
3,339,375
= 1 i j =
133,575
539
E 248 mi
53. The function whose graph is the upper branch of this hyperbola is concave upward. The function is
j = 1(r) = o
1 +
r
2
/
2
=
o
/
I
/
2
+r
2
, so j
0
=
o
/
r(/
2
+ r
2
)
31'2
and
j
00
=
o
/
(/
2
+r
2
)
31'2
3r
2
(/
2
+r
2
)
33'2
= o/(/
2
+r
2
)
33'2
0 for all r, and so 1 is concave upward.
55. (a) If I 16, then I 316 0, and
r
2
I
+
j
2
I 316
= 1 is an ellipse since it is the sum of two squares on the left side.
(b) If 0 < I < 16, then I 316 < 0, and
r
2
I
+
j
2
I 316
= 1 is a hyperbola since it is the difference of two squares on the
left side.
(c) If I < 0, then I 316 < 0, and there is no curve since the left side is the sum of two negative terms, which cannot equal 1.
(d) In case (a), o
2
= I, /
2
= I 316, and c
2
= o
2
3/
2
= 16, so the foci are at (4. 0). In case (b), I 316 < 0, so o
2
= I,
/
2
= 16 3I, and c
2
= o
2
+/
2
= 16, and so again the foci are at (4. 0).
57. r
2
= 4jj i 2r = 4jj
0
i j
0
=
r
2j
, so the tangent line at (r
0
. j
0
) is
j 3
r
2
0
4j
=
r
0
2j
(r 3r
0
). This line passes through the point (o. 3j) on the
directrix, so 3j 3
r
2
0
4j
=
r
0
2j
(o 3r
0
) i 34j
2
3r
2
0
= 2or
0
32r
2
0
C
r
2
0
32or
0
34j
2
= 0 C r
2
0
32or
0
+o
2
= o
2
+ 4j
2
C

441
(r
0
3o)
2
= o
2
+ 4j
2
C r
0
= o
o
2
+ 4j
2
. The slopes of the tangent lines at r = o
o
2
+ 4j
2
are
o
o
2
+ 4j
2
2j
, so the product of the two slopes is
o +
o
2
+ 4j
2
2j

o 3
o
2
+ 4j
2
2j
=
o
2
3(o
2
+ 4j
2
)
4j
2
=
34j
2
4j
2
= 31,
showing that the tangent lines are perpendicular.
59. For r
2
+ 4j
2
= 4, or r
2
4 +j
2
= 1, use the parametrization r = 2 cos t, j = sin t, 0 $ t $ 2 to get
1 = 4
r'2
0
(drdt)
2
+ (djdt)
2
dt = 4
r'2
0
4 sin
2
t + cos
2
t dt = 4
r'2
0
3 sin
2
t + 1 dt
Using Simpsons Rule with n = 10, {t =
r'2 30
10
=
r
20
, and 1(t) =
3 sin
2
t + 1, we get
1 E
4
3
r
20
1(0) + 41
r
20
+ 21
2r
20
+ + 21
8r
20
+ 41
9r
20
+1
r
2
E 9.69
61.
r
2
o
2
3
j
2
/
2
= 1 i
j
2
/
2
=
r
2
3o
2
o
2
i j =
/
o
I
r
2
3o
2
.
= 2
c
a
/
o
r
2
3o
2
dr
39
=
2/
o
r
2
r
2
3o
2
3
o
2
2
ln
r +
r
2
3o
2
c
a
=
/
o
c
I
c
2
3o
2
3o
2
ln
c +
I
c
2
3o
2
+o
2
ln|o|
Since o
2
+/
2
= c
2
. c
2
3o
2
= /
2
, and
I
c
2
3o
2
= /.
=
/
o
c/ 3o
2
ln(c +/) +o
2
lno
=
/
o
c/ +o
2
(lno 3ln(/ +c))
= /
2
co +o/ ln[o(/ +c)], where c
2
= o
2
+/
2
.
63. Differentiating implicitly,
r
2
o
2
+
j
2
/
2
= 1 i
2r
o
2
+
2jj
0
/
2
= 0 i j
0
= 3
/
2
r
o
2
j
[j 6= 0]. Thus, the slope of the tangent
line at 1 is 3
/
2
r
1
o
2
j
1
. The slope of 1
1
1 is
j
1
r
1
+c
and of 1
2
1 is
j
1
r
1
3c
. By the formula in Problem 17 on text page 268,
we have
tanc =
j
1
r
1
+c
+
/
2
r
1
o
2
j
1
1 3
/
2
r
1
j
1
o
2
j
1
(r
1
+c)
=
o
2
j
2
1
+/
2
r
1
(r
1
+c)
o
2
j
1
(r
1
+c) 3/
2
r
1
j
1
=
o
2
/
2
+/
2
cr
1
c
2
r
1
j
1
+o
2
cj
1
using b
2
r
2
1
+ o
2
u
2
1
= o
2
b
2
,
and o
2
b
2
= c
2
=
/
2
cr
1
+o
2
cj
1
(cr
1
+o
2
)
=
/
2
cj
1
and
tano =
3
/
2
r
1
o
2
j
1
3
j
1
r
1
3c
1 3
/
2
r
1
j
1
o
2
j
1
(r
1
3c)
=
3o
2
j
2
1
3/
2
r
1
(r
1
3c)
o
2
j
1
(r
1
3c) 3/
2
r
1
j
1
=
3o
2
/
2
+/
2
cr
1
c
2
r
1
j
1
3o
2
cj
1
=
/
2
cr
1
3o
2
cj
1
(cr
1
3o
2
)
=
/
2
cj
1
Thus, c = o.
442

10.6 Conic Sections in Polar Coordinates
1. The directrix j = 6 is above the focus at the origin, so we use the form with +c sin0 in the denominator. [See Theorem 6
and Figure 2(c).] v =
cd
1 +c sin0
=
7
4
6
1 +
7
4
sin0
=
42
4 + 7 sin0
3. The directrix r = 35 is to the left of the focus at the origin, so we use the form with 3c cos 0 in the denominator.
v =
cd
1 3c cos 0
=
3
4
5
1 3
3
4
cos 0
=
15
4 33 cos 0
5. The vertex (4. 32) is 4 units below the focus at the origin, so the directrix is 8 units below the focus (d = 8), and we use the
form with 3c sin0 in the denominator. c = 1 for a parabola, so an equation is v =
cd
1 3c sin0
=
1(8)
1 31 sin0
=
8
1 3sin0
.
7. The directrix v = 4 sec 0 (equivalent to v cos 0 = 4 or r = 4) is to the right of the focus at the origin, so we will use the form
with +c cos 0 in the denominator. The distance from the focus to the directrix is d = 4, so an equation is
v =
cd
1 +c cos 0
=
1
2
(4)
1 +
1
2
cos 0

2
2
=
4
2 + cos 0
.
9. v =
1
1 + sin0
=
cd
1 +c sin0
, where d = c = 1.
(a) Eccentricity = c = 1
(b) Since c = 1, the conic is a parabola.
(c) Since +c sin0 appears in the denominator, the directrix is above the
focus at the origin. d = |1|| = 1, so an equation of the directrix is j = 1.
(d) The vertex is at
1
2
.
r
2
, midway between the focus and the directrix.

11. v =
12
4 3sin0

14
14
=
3
1 3
1
4
sin0
, where c =
1
4
and cd = 3 i d = 12.
(a) Eccentricity = c =
1
4
(b) Since c =
1
4
< 1, the conic is an ellipse.
(c) Since 3c sin0 appears in the denominator, the directrix is below the focus
at the origin. d = |1|| = 12, so an equation of the directrix is j = 312.
(d) The vertices are
4.
r
2
and
12
5
.
3r
2
, so the center is midway between them,

that is,
4
5
.
r
2
.
SECTION 10.6 CONIC SECTIONS IN POLAR COORDINATES

443
13. v =
9
6 + 2 cos 0

16
16
=
32
1 +
1
3
cos 0
, where c =
1
3
and cd =
3
2
i d =
9
2
.
(a) Eccentricity = c =
1
3
(b) Since c =
1
3
< 1, the conic is an ellipse.
(c) Since +c cos 0 appears in the denominator, the directrix is to the right of
the focus at the origin. d = |1|| =
9
2
, so an equation of the directrix is
r =
9
2
.
9
8
. 0
and
9
4
.

that is,
9
16
.
.
15. v =
3
4 38 cos 0

14
14
=
34
1 32 cos 0
, where c = 2 and cd =
3
4
i d =
3
8
.
(a) Eccentricity = c = 2
(b) Since c = 2 1, the conic is a hyperbola.
(c) Since 3c cos 0 appears in the denominator, the directrix is to the left of
the focus at the origin. d = |1|| =
3
8
, so an equation of the directrix is
r = 3
3
8
.
3
3
4
. 0
and
1
4
.

that is,
1
2
.
.
17. (a) v =
1
1 32 sin0
, where c = 2 and cd = 1 i d =
1
2
. The eccentricity
c = 2 1, so the conic is a hyperbola. Since 3c sin0 appears in the
denominator, the directrix is below the focus at the origin. d = |1|| =
1
2
,
so an equation of the directrix is j = 3
1
2
. The vertices are
31.
r
2
and
1
3
.
3r
2
, so the center is midway between them, that is,
2
3
.
3r
2
.
(b) By the discussion that precedes Example 4, the equation
is v =
1
1 32 sin
0 3
3r
4
.
444

19. For c < 1 the curve is an ellipse. It is nearly circular when c is close to 0. As c
increases, the graph is stretched out to the right, and grows larger (that is, its
right-hand focus moves to the right while its left-hand focus remains at the
origin.) At c = 1, the curve becomes a parabola with focus at the origin.
21. |11| = c |1|| i v = c[d 3v cos( 30)] = c(d +v cos 0) i
v(1 3c cos 0) = cd i v =
cd
1 3c cos 0
23. |11| = c |1|| i v = c[d 3v sin(0 3)] = c(d +v sin0) i
v(1 3c sin0) = cd i v =
cd
1 3c sin0
25. We are given c = 0.093 and o = 2.28 10
8
. By (7), we have
v =
o(1 3c
2
)
1 +c cos 0
=
2.28 10
8
[1 3(0.093)
2
]
1 + 0.093 cos 0
E
2.26 10
8
1 + 0.093 cos 0
27. Here 2o = length of major axis = 36.18 AU i o = 18.09 AU and c = 0.97. By (7), the equation of the orbit is
v =
18.09
1 3(0.97)
2
1 30.97 cos 0
E
1.07
1 30.97 cos 0
. By (8), the maximum distance from the comet to the sun is
18.09(1 + 0.97) E 35.64 AU or about 3.314 billion miles.
29. The minimum distance is at perihelion, where 4.6 10
7
= v = o(1 3c) = o(1 30.206) = o(0.794) i
o = 4.6 10
7
0.794. So the maximum distance, which is at aphelion, is
v = o(1 +c) =
4.6 10
7
0.794
(1.206) E 7.0 10
7
km.
31. From Exercise 29, we have c = 0.206 and o(1 3c) = 4.6 10
7
km. Thus, o = 4.6 10
7
0.794. From (7), we can write the
equation of Mercurys orbit as v = o
1 3c
2
1 3c cos 0
. So since
dv
d0
=
3o(1 3c
2
)c sin0
(1 3c cos 0)
2
i
v
2
+
dv
d0
2
=
o
2
(1 3c
2
)
2
(1 3c cos 0)
2
+
o
2
(1 3c
2
)
2
c
2
sin
2
0
(1 3c cos 0)
4
=
o
2
(1 3c
2
)
2
(1 3c cos 0)
4
(1 32c cos 0 +c
2
)
CHAPTER 10 REVIEW

445
the length of the orbit is
1 =

2r
0
v
2
+ (dvd0)
2
d0 = o(1 3c
2
)
2r
0
I
1 +c
2
32c cos 0
(1 3c cos 0)
2
d0 E 3.6 10
8
km
This seems reasonable, since Mercurys orbit is nearly circular, and the circumference of a circle of radius o
is 2o E 3.6 10
8
km.
10 Review
1. (a) A parametric curve is a set of points of the form (r. j) = (1(t). o(t)), where 1 and o are continuous functions of a
variable t.
(b) Sketching a parametric curve, like sketching the graph of a function, is difcult to do in general. We can plot points on the
curve by nding 1(t) and o(t) for various values of t, either by hand or with a calculator or computer. Sometimes, when
1 and o are given by formulas, we can eliminate t from the equations r = 1(t) and j = o(t) to get a Cartesian equation
relating r and j. It may be easier to graph that equation than to work with the original formulas for r and j in terms of t.
2. (a) You can nd
dj
dr
as a function of t by calculating
dj
dr
=
djdt
drdt
[if drdt 6= 0].
(b) Calculate the area as
l
a
j dr =

c
o
o(t) 1
0
(t)dt [or
o
c
o(t) 1
0
(t)dt if the leftmost point is (1(o). o(o)) rather
than (1(c). o(c))].
3. (a) 1 =

c
o
(drdt)
2
+ (djdt)
2
dt =

c
o
[1
0
(t)]
2
+ [o
0
(t)]
2
dt
(b) o =

c
o
2j
(drdt)
2
+ (djdt)
2
dt =

c
o
2o(t)
[1
0
(t)]
2
+ [o
0
(t)]
2
dt
4. (a) See Figure 5 in Section 10.3.
(b) r = v cos 0, j = v sin0
(c) To nd a polar representation (v. 0) with v D 0 and 0 $ 0 < 2, rst calculate v =

r
2
+j
2
. Then 0 is specied by
cos 0 = rv and sin0 = jv.
5. (a) Calculate
dj
dr
=
dj
d0
dr
d0
=
d
d0
(j)
d
d0
(r)
=
d
d0
(v sin0)
d
d0
(v cos 0)
=
dv
d0
sin0 +v cos 0
dv
d0
cos 0 3v sin0
, where v = 1(0).
(b) Calculate =

l
a
1
2
v
2
d0 =

l
a
1
2
[1(0)]
2
d0
(c) 1 =

l
a
(drd0)
2
+ (djd0)
2
d0 =

l
a
v
2
+ (dvd0)
2
d0 =

l
a
[1(0)]
2
+ [1
0
(0)]
2
d0
6. (a) A parabola is a set of points in a plane whose distances from a xed point 1 (the focus) and a xed line | (the directrix)
are equal.
(b) r
2
= 4jj; j
2
= 4jr
446

7. (a) An ellipse is a set of points in a plane the sum of whose distances from two xed points (the foci) is a constant.
(b)
r
2
o
2
+
j
2
o
2
3c
2
= 1.
8. (a) A hyperbola is a set of points in a plane the difference of whose distances from two xed points (the foci) is a constant.
This difference should be interpreted as the larger distance minus the smaller distance.
(b)
r
2
o
2
3
j
2
c
2
3o
2
= 1
(c) j =
I
c
2
3o
2
o
r
9. (a) If a conic section has focus 1 and corresponding directrix |, then the eccentricity c is the xed ratio |11| |1|| for points
1 of the conic section.
(b) c < 1 for an ellipse; c 1 for a hyperbola; c = 1 for a parabola.
(c) r = d: v =
cd
1 +c cos 0
. r = 3d: v =
cd
1 3c cos 0
. j = d: v =
cd
1 +c sin0
. j = 3d: v =
cd
1 3c sin0
.
1. False. Consider the curve dened by r = 1(t) = (t 31)
3
and j = o(t) = (t 31)
2
. Then o
0
(t) = 2(t 31), so o
0
(1) = 0,
but its graph has a vertical tangent when t = 1. Note: The statement is true if 1
0
(1) 6= 0 when o
0
(1) = 0.
3. False. For example, if 1(t) = cos t and o(t) = sint for 0 $ t $ 4, then the curve is a circle of radius 1, hence its length
is 2, but
4r
0
[1
0
(t)]
2
+ [o
0
(t)]
2
dt =

4r
0
(3sint)
2
+ (cos t)
2
dt =

4r
0
1 dt = 4, since as t increases
from 0 to 4, the circle is traversed twice.
5. True. The curve v = 1 3sin20 is unchanged if we rotate it through 180
about O because
1 3sin2(0 +) = 1 3sin(20 + 2) = 1 3sin20. So its unchanged if we replace v by 3v. (See the discussion
after Example 8 in Section 10.3.) In other words, its the same curve as v = 3(1 3sin20) = sin20 31.
7. False. The rst pair of equations gives the portion of the parabola j = r
2
with r D 0, whereas the second pair of equations
traces out the whole parabola j = r
2
.
9. True. By rotating and translating the parabola, we can assume it has an equation of the form j = cr
2
, where c 0.
The tangent at the point
o. co
2
is the line j 3co

2
= 2co(r 3o); i.e., j = 2cor 3co
2
. This tangent meets
the parabola at the points
r. cr
2
where cr
2
= 2cor 3co
2
. This equation is equivalent to r
2
= 2or 3o
2
[since c 0]. But r
2
= 2or 3o
2
C r
2
32or +o
2
= 0 C (r 3o)
2
= 0 C r = o C
r. cr
2
=

o. co
2
. This shows that each tangent meets the parabola at exactly one point.
CHAPTER 10 REVIEW

447
1. r = t
2
+ 4t, j = 2 3t, 34 $ t $ 1. t = 2 3j, so
r = (2 3j)
2
+ 4(2 3j) = 4 34j +j
2
+ 8 34j = j
2
38j + 12 C
r + 4 = j
2
38j + 16 = (j 34)
2
. This is part of a parabola with vertex
(34. 4), opening to the right.
3. j = sec 0 =
1
cos 0
=
1
r
. Since 0 $ 0 $ 2, 0 < r $ 1 and j D 1.
This is part of the hyperbola j = 1r.
5. Three different sets of parametric equations for the curve j =
I
r are
(i) r = t, j =
I
t
(ii) r = t
4
, j = t
2
(iii) r = tan
2
t, j = tant, 0 $ t < 2
There are many other sets of equations that also give this curve.
7. (a) The Cartesian coordinates are r = 4 cos
2r
3
= 4
3
1
2
= 32 and
j = 4 sin
2r
3
= 4
I
3
2
= 2
I
3, that is, the point
32. 2
I
3
.
(b) Given r = 33 and j = 3, we have v =

(33)
2
+ 3
2
=
I
18 = 3
I
2. Also, tan0 =
j
r
i tan0 =
3
33
, and since
(33. 3) is in the second quadrant, 0 =
3r
4
. Thus, one set of polar coordinates for (33. 3) is
3
I
2.
3r
4
, and two others are
3
I
2.
11r
4
and
33
I
2.
7r
4
.
9. v = 1 3cos 0. This cardioid is
symmetric about the polar axis.
448

11. v = cos 30. This is a
three-leaved rose. The curve is
traced twice.
13. v = 1 + cos 20. The curve is
symmetric about the pole and
both the horizontal and vertical
axes.
15. v =
3
1 + 2 sin0
i c = 2 1, so the conic is a hyperbola. dc = 3 i
d =
3
2
and the form +2 sin0 imply that the directrix is above the focus at
the origin and has equation j =
3
2
. The vertices are
1.
r
2
and
33.
3r
2
.
17. r +j = 2 C v cos 0 +v sin0 = 2 C v(cos 0 + sin0) = 2 C v =
2
cos 0 + sin0
19. v = (sin0)0. As 0 <", v <0.
As 0 <0, v <1. In the rst gure,
there are an innite number of
r-intercepts at r = n, n a nonzero
integer. These correspond to pole
points in the second gure.
21. r = lnt, j = 1 +t
2
; t = 1.
dj
dt
= 2t and
dr
dt
=
1
t
, so
dj
dr
=
djdt
drdt
=
2t
1t
= 2t
2
.
When t = 1, (r. j) = (0. 2) and djdr = 2.
23. v = c
30
i j = v sin0 = c
30
sin0 and r = v cos 0 = c
30
cos 0 i
dj
dr
=
djd0
drd0
=
ur
u0
sin0 +v cos 0
ur
u0
cos 0 3v sin0
=
3c
30
sin0 +c
30
cos 0
3c
30
cos 0 3c
30
sin0

3c
0
3c
0
=
sin0 3cos 0
cos 0 + sin0
.
When 0 = ,
dj
dr
=
0 3(31)
31 + 0
=
1
31
= 31.
CHAPTER 10 REVIEW

449
25. r = t + sint, j = t 3cos t i
dj
dr
=
djdt
drdt
=
1 + sint
1 + cos t
i
d
2
j
dr
2
=
d
dt
dj
dr
drdt
=
(1 + cos t) cos t 3(1 + sint)(3sint)
(1 + cos t)
2
1 + cos t
=
cos t + cos
2
t + sint + sin
2
t
(1 + cos t)
3
=
1 + cos t + sint
(1 + cos t)
3
27. We graph the curve r = t
3
33t, j = t
2
+t + 1 for 32.2 $ t $ 1.2.
By zooming in or using a cursor, we nd that the lowest point is about
(1.4. 0.75). To nd the exact values, we nd the t-value at which
djdt = 2t + 1 = 0 C t = 3
1
2
C (r. j) =

11
8
.
3
4
.
29. r = 2o cos t 3o cos 2t i
dr
dt
= 32o sint + 2o sin2t = 2o sint(2 cos t 31) = 0 C
sint = 0 or cos t =
1
2
i t = 0,
r
3
, , or
5r
3
.
j = 2o sint 3o sin2t i
dj
dt
= 2o cos t 32o cos 2t = 2o
1 + cos t 32 cos
2
t
= 2o(1 3cos t)(1 + 2 cos t) = 0 i

t = 0,
2r
3
, or
4r
3
.
Thus the graph has vertical tangents where
t =
r
3
, and
5r
3
, and horizontal tangents where
t =
2r
3
and
4r
3
. To determine what the slope is
where t = 0, we use lHospitals Rule to evaluate
lim
I<0
djdt
drdt
= 0, so there is a horizontal tangent
there.
t r j
0 o 0
r
3
3
2
o
I
3
2
o
2r
3
3
1
2
o
3
I
3
2
o
33o 0
4r
3
3
1
2
o 3
3
I
3
2
o
5r
3
3
2
o 3
I
3
2
o
31. The curve v
2
= 9 cos 50 has 10 petals. For instance, for 3
r
10
$ 0 $
r
10
, there are two petals, one with v 0 and one
with v < 0.
= 10
r'10
3r'10
1
2
v
2
d0 = 5
r'10
3r'10
9 cos 50 d0 = 5 9 2
r'10
0
cos 50 d0 = 18
sin50
r'10
0
= 18
33. The curves intersect when 4 cos 0 = 2 i cos 0 =
1
2
i 0 =
r
3
for 3 $ 0 $ . The points of intersection are
2.
r
3
and
2. 3
r
3
.
450

35. The curves intersect where 2 sin0 = sin0 + cos 0 i
sin0 = cos 0 i0 =
r
4
, and also at the origin (at which 0 =
3r
4
on the second curve).
=

r'4
0
1
2
(2 sin0)
2
d0 +
3r'4
r'4
1
2
(sin0 + cos 0)
2
d0
=

r'4
0
(1 3cos 20) d0 +
1
2
3r'4
r'4
(1 + sin20) d0
=

0 3
1
2
sin20
r'4
0
+
1
2
0 3
1
4
cos 20
3r'4
r'4
=
1
2
( 31)
37. r = 3t
2
, j = 2t
3
.
1=

2
0
(drdt)
2
+ (djdt)
2
dt =

2
0
(6t)
2
+ (6t
2
)
2
dt =

2
0
I
36t
2
+ 36t
4
dt =

2
0
I
36t
2
I
1 +t
2
dt
=

2
0
6 |t|
I
1 +t
2
dt = 6
2
0
t
I
1 +t
2
dt = 6
5
1
&
1'2
1
2
d&

u = 1 + |
2
, Ju = 2| J|
= 6
1
2

2
3
&
3'2
5
1
= 2(5
3'2
31) = 2
5
I
5 31
39. 1 =

2r
r
v
2
+ (dvd0)
2
d0 =

2r
r
(10)
2
+ (310
2
)
2
d0 =

2r
r
0
2
+ 1
0
2
d0
24
=
0
2
+ 1
0
+ ln
0 +
0
2
+ 1
2r
r
=
I
2
+ 1
3
I
4
2
+ 1
2
+ ln
2 +
I
4
2
+ 1
+
I
2
+ 1
=
2
I
2
+ 1 3
I
4
2
+ 1
2
+ ln
2 +
I
4
2
+ 1
+
I
2
+ 1
41. r = 4
I
t, j =
t
3
3
+
1
2t
2
, 1 $ t $ 4 i
o =

4
1
2j
(drdt)
2
+ (djdt)
2
dt =

4
1
2
1
3
t
3
+
1
2
t
32
2
I
t
2
+ (t
2
3t
33
)
2
dt
= 2
4
1
1
3
t
3
+
1
2
t
32

(t
2
+t
33
)
2
dt = 2
4
1
1
3
t
5
+
5
6
+
1
2
t
35
dt = 2
1
18
t
6
+
5
6
t 3
1
8
t
34
4
1
=
471,295
1024

43. For all c except 31, the curve is asymptotic to the line r = 1. For
c < 31, the curve bulges to the right near j = 0. As c increases, the
bulge becomes smaller, until at c = 31 the curve is the straight line r = 1.
As c continues to increase, the curve bulges to the left, until at c = 0 there
is a cusp at the origin. For c 0, there is a loop to the left of the origin,
whose size and roundness increase as c increases. Note that the r-intercept
of the curve is always 3c.
CHAPTER 10 REVIEW

451
45.
r
2
9
+
j
2
8
= 1 is an ellipse with center (0. 0).
o = 3, / = 2
I
2, c = 1 i
foci (1. 0), vertices (3. 0).
47. 6j
2
+r 336j + 55 = 0 C
6(j
2
36j + 9) = 3(r + 1) C
(j 33)
2
= 3
1
6
(r + 1), a parabola with vertex (31. 3),
opening to the left, j = 3
1
24
i focus
3
25
24
. 3
and
directrix r = 3
23
24
.
49. The ellipse with foci (4. 0) and vertices (5. 0) has center (0. 0) and a horizontal major axis, with o = 5 and c = 4,
so /
2
= o
2
3c
2
= 5
2
34
2
= 9. An equation is
r
2
25
+
j
2
9
= 1.
51. The center of a hyperbola with foci (0. 4) is (0. 0), so c = 4 and an equation is
j
2
o
2
3
r
2
/
2
= 1.
The asymptote j = 3r has slope 3, so
o
/
=
3
1
i o = 3/ and o
2
+/
2
= c
2
i (3/)
2
+/
2
= 4
2
i
10/
2
= 16 i /
2
=
8
5
and so o
2
= 16 3
8
5
=
72
5
. Thus, an equation is
j
2
725
3
r
2
85
= 1, or
5j
2
72
3
5r
2
8
= 1.
53. r
2
= 3(j 3100) has its vertex at (0. 100), so one of the vertices of the ellipse is (0. 100). Another form of the equation of a
parabola is r
2
= 4j(j 3100) so 4j(j 3100) = 3(j 3100) i 4j = 31 i j = 3
1
4
. Therefore the shared focus is
found at
0.
399
4
so 2c =
399
4
30 i c =
399
8
and the center of the ellipse is
0.
399
8
. So o = 100 3
399
8
=
401
8
and
/
2
= o
2
3c
2
=
401
2
3399
2
8
2
= 25. So the equation of the ellipse is
r
2
/
2
+
j 3
399
8
2
o
2
= 1 i
r
2
25
+
j 3
399
8
401
8
2
= 1,
or
r
2
25
+
(8j 3399)
2
160,801
= 1.
55. Directrix r = 4 i d = 4, so c =
1
3
i v =
cd
1 +c cos 0
=
4
3 + cos 0
.
57. In polar coordinates, an equation for the circle is v = 2o sin0. Thus, the coordinates of Q are r = v cos 0 = 2o sin0 cos 0
and j = v sin0 = 2o sin
2
0. The coordinates of 1 are r = 2o cot 0 and j = 2o. Since 1 is the midpoint of Q1, we use the
midpoint formula to get r = o(sin0 cos 0 + cot 0) and j = o(1 + sin
2
0).
PROBLEMS PLUS
1. r =

I
1
cos &
&
d&, j =

I
1
sin&
&
d&, so by FTC1, we have
dr
dt
=
cos t
t
and
dj
dt
=
sint
t
. Vertical tangent lines occur when
dr
dt
= 0 C cos t = 0. The parameter value corresponding to (r. j) = (0, 0) is t = 1, so the nearest vertical tangent
occurs when t =
r
2
. Therefore, the arc length between these points is
1 =

r'2
1
dr
dt
2
+
dj
dt
2
dt =

r'2
1
cos
2
t
t
2
+
sin
2
t
t
2
dt =

r'2
1
dt
t
=

lnt
r'2
1
= ln
r
2
3. In terms of r and j, we have r = v cos 0 = (1 +c sin0) cos 0 = cos 0 +c sin0 cos 0 = cos 0 +
1
2
c sin20 and
j = v sin0 = (1 +c sin0) sin0 = sin0 +c sin
2
0. Now 31 $ sin0 $ 1 i 31 $ sin0 +c sin
2
0 $ 1 +c $ 2, so
31 $ j $ 2. Furthermore, j = 2 when c = 1 and 0 =
r
2
, while j = 31 for c = 0 and 0 =
3r
2
. Therefore, we need a viewing
rectangle with 31 $ j $ 2.
To nd the r-values, look at the equation r = cos 0 +
1
2
c sin20 and use the fact that sin20 D 0 for 0 $ 0 $
r
2
and
sin20 $ 0 for 3
r
2
$ 0 $ 0. [Because v = 1 +c sin0 is symmetric about the j-axis, we only need to consider
3
r
2
$ 0 $
r
2
.] So for 3
r
2
$ 0 $ 0, r has a maximum value when c = 0 and then r = cos 0 has a maximum value
of 1 at 0 = 0. Thus, the maximum value of r must occur on
0.
r
2
with c = 1. Then r = cos 0 +

1
2
sin20 i
ui
u0
= 3sin0 + cos 20 = 3sin0 + 1 32 sin
2
0 i
ui
u0
= 3(2 sin0 31)(sin0 + 1) = 0 when sin0 = 31 or
1
2
[but sin0 6= 31 for 0 $ 0 $
r
2
]. If sin0 =
1
2
, then 0 =
r
6
and
r = cos
r
6
+
1
2
sin
r
3
=
3
4
I
3. Thus, the maximum value of r is
3
4
I
3, and,
by symmetry, the minimum value is 3
3
4
I
3. Therefore, the smallest
viewing rectangle that contains every member of the family of polar curves
v = 1 +c sin0, where 0 $ c $ 1, is
3
3
4
I
3.
3
4
I
3
[31. 2].
5. (a) If (o. /) lies on the curve, then there is some parameter value t
1
such that
3t
1
1 +t
3
1
= o and
3t
2
1
1 +t
3
1
= /. If t
1
= 0,
the point is (0. 0), which lies on the line j = r. If t
1
6= 0, then the point corresponding to t =
1
t
1
is given by
r =
3(1t
1
)
1 + (1t
1
)
3
=
3t
2
1
t
3
1
+ 1
= /, j =
3(1t
1
)
2
1 + (1t
1
)
3
=
3t
1
t
3
1
+ 1
= o. So (/. o) also lies on the curve. [Another way to see
this is to do part (e) rst; the result is immediate.] The curve intersects the line j = r when
3t
1 +t
3
=
3t
2
1 +t
3
i
t = t
2
i t = 0 or 1, so the points are (0. 0) and
3
2
.
3
2
.
453
454

(b)
dj
dt
=
(1 +t
3
)(6t) 33t
2
(3t
2
)
(1 +t
3
)
2
=
6t 33t
4
(1 +t
3
)
2
= 0 when 6t 33t
4
= 3t(2 3t
3
) = 0 i t = 0 or t =
3
I
2, so there are
horizontal tangents at (0. 0) and
3
I
2.
3
I
4
. Using the symmetry from part (a), we see that there are vertical tangents at
(0. 0) and
3
I
4.
3
I
2
.
(c) Notice that as t <31
+
, we have r <3"and j <". As t <31
3
, we have r <"and j <3". Also
j 3(3r 31) = j +r + 1 =
3t + 3t
2
+ (1 +t
3
)
1 +t
3
=
(t + 1)
3
1 +t
3
=
(t + 1)
2
t
2
3t + 1
<0 as t <31. So j = 3r 31 is a
slant asymptote.
(d)
dr
dt
=
(1 +t
3
)(3) 33t(3t
2
)
(1 +t
3
)
2
=
3 36t
3
(1 +t
3
)
2
and from part (b) we have
dj
dt
=
6t 33t
4
(1 +t
3
)
2
. So
dj
dr
=
djdt
drdt
=
t(2 3t
3
)
1 32t
3
.
Also
d
2
j
dr
2
=
d
dt
dj
dr
drdt
=
2(1 +t
3
)
4
3(1 32t
3
)
3
0 C t <
1
3
I
2
.
So the curve is concave upward there and has a minimum point at (0. 0)
and a maximum point at
3
I
2.
3
I
4
. Using this together with the

information from parts (a), (b), and (c), we sketch the curve.
(e) r
3
+j
3
=

3t
1 +t
3
3
+
3t
2
1 +t
3
3
=
27t
3
+ 27t
6
(1 +t
3
)
3
=
27t
3
(1 +t
3
)
(1 +t
3
)
3
=
27t
3
(1 +t
3
)
2
and
3rj = 3
3t
1 +t
3
3t
2
1 +t
3
=
27t
3
(1 +t
3
)
2
, so r
3
+j
3
= 3rj.
(f ) We start with the equation from part (e) and substitute r = v cos 0, j = v sin0. Then r
3
+j
3
= 3rj i
v
3
cos
3
0 +v
3
sin
3
0 = 3v
2
cos 0 sin0. For v 6= 0, this gives v =
3 cos 0 sin0
cos
3
0 + sin
3
0
. Dividing numerator and denominator
by cos
3
0, we obtain v =
3
1
cos 0
sin0
cos 0
1 +
sin
3
0
cos
3
0
=
3 sec 0 tan0
1 + tan
3
0
.
(g) The loop corresponds to 0 M
0.
r
2
, so its area is
=

r'2
0
v
2
2
d0 =
1
2
r'2
0
3 sec 0 tan0
1 + tan
3
0
2
d0 =
9
2
r'2
0
sec
2
0 tan
2
0
(1 + tan
3
0)
2
d0 =
9
2
"
0
&
2
d&
(1 +&
3
)
2
[let u = tan 0]
= lim
l<"
9
2
3
1
3
(1 +&
3
)
31
l
0
=
3
2
(h) By symmetry, the area between the folium and the line j = 3r 31 is equal to the enclosed area in the third quadrant,
plus twice the enclosed area in the fourth quadrant. The area in the third quadrant is
1
2
, and since j = 3r 31 i
v sin0 = 3v cos 0 31 i v = 3
1
sin0 + cos 0
, the area in the fourth quadrant is
1
2
3r'4
3r'2
3
1
sin0 + cos 0
2
3
3 sec 0 tan0
1 + tan
3
0
d0
CAS
=
1
2
. Therefore, the total area is
1
2
+ 2
1
2
=
3
2
.
11 INFINITE SEQUENCES AND SERIES
11.1 Sequences
1. (a) A sequence is an ordered list of numbers. It can also be dened as a function whose domain is the set of positive integers.
(b) The terms o
n
approach 8 as n becomes large. In fact, we can make o
n
as close to 8 as we like by taking n sufciently
large.
(c) The terms o
n
become large as n becomes large. In fact, we can make o
n
as large as we like by taking n sufciently large.
3. o
n
= 1 3(0.2)
n
, so the sequence is {0.8. 0.96. 0.992. 0.9984. 0.99968. . . .}.
5. o
n
=
3 (31)
n
n!
, so the sequence is
33
1
.
3
2
.
33
6
.
3
24
.
33
120
. . . .
=

33.
3
2
. 3
1
2
.
1
8
. 3
1
40
. . . .
.
7. o
1
= 3, o
n+1
= 2o
n
31. Each term is dened in terms of the preceding term.
o
2
= 2o
1
31 = 2(3) 31 = 5. o
3
= 2o
2
31 = 2(5) 31 = 9. o
4
= 2o
3
31 = 2(9) 31 = 17.
o
5
= 2o
4
31 = 2(17) 31 = 33. The sequence is {3. 5. 9. 17. 33. . . .}.
9.

1.
1
3
.
1
5
.
1
7
.
1
9
. . . .
. The denominator of the nth term is the nth positive odd integer, so o
n
=
1
2n 31
.
11. {2. 7. 12. 17. . . .}. Each term is larger than the preceding one by 5, so o
n
= o
1
+d(n 31) = 2 + 5(n 31) = 5n 33.
13.

1. 3
2
3
.
4
9
. 3
8
27
. . . .
. Each term is 3
2
3
times the preceding one, so o
n
=

3
2
3
n31
.
15. The rst six terms of o
n
=
n
2n + 1
are
1
3
,
2
5
,
3
7
,
4
9
,
5
11
,
6
13
. It appears that the sequence is approaching
1
2
.
lim
n<"
n
2n + 1
= lim
n<"
1
2 + 1n
=
1
2
17. o
n
= 1 3(0.2)
n
, so lim
n<"
o
n
= 1 30 = 1 by (9). Converges
19. o
n
=
3 + 5n
2
n +n
2
=
(3 + 5n
2
)n
2
(n +n
2
)n
2
=
5 + 3n
2
1 + 1n
, so o
n
<
5 + 0
1 + 0
= 5 as n <". Converges
21. Because the natural exponential function is continuous at 0, Theorem 7 enables us to write
lim
n<"
o
n
= lim
n<"
c
1'n
= c
lim
q$4
(1'n)
= c
0
= 1. Converges
23. If /
n
=
2n
1 + 8n
, then lim
n<"
/
n
= lim
n<"
(2n)n
(1 + 8n)n
= lim
n<"
2
1n + 8
=
2
8
=

4
. Since tan is continuous at
r
4
, by
Theorem 7, lim
n<"
tan
2n
1 + 8n
= tan
lim
n<"
2n
1 + 8n
= tan

4
= 1. Converges
455
456

CHAPTER 11 INFINITE SEQUENCES AND SERIES
25. o
n
=
(31)
n31
n
n
2
+ 1
=
(31)
n31
n + 1n
, so 0 $ |o
n
| =
1
n + 1n
$
1
n
<0 as n <", so o
n
<0 by the Squeeze Theorem and
Theorem 6. Converges
27. o
n
= cos( n2). This sequence diverges since the terms dont approach any particular real number as n <".
The terms take on values between 31 and 1.
29. o
n
=
(2n 31)!
(2n + 1)!
=
(2n 31)!
(2n + 1)(2n)(2n 31)!
=
1
(2n + 1)(2n)
<0 as n <". Converges
31. o
n
=
c
n
+c
3n
c
2n
31

c
3n
c
3n
=
1 +c
32n
c
n
3c
3n
<0 as n <"because 1 +c
32n
<1 and c
n
3c
3n
<". Converges
33. o
n
= n
2
c
3n
=
n
2
c
n
. Since lim
i<"
r
2
c
i
H
= lim
i<"
2r
c
i
H
= lim
i<"
2
c
i
= 0, it follows from Theorem 3 that lim
n<"
o
n
= 0. Converges
35. 0 $
cos
2
n
2
n
$
1
2
n
[since 0 $ cos
2
n $ 1], so since lim
n<"
1
2
n
= 0,
cos
2
n
2
n
converges to 0 by the Squeeze Theorem.

37. o
n
= nsin(1n) =
sin(1n)
1n
. Since lim
i<"
sin(1r)
1r
= lim
I<0
+
sint
t
[where t = 1r] = 1, it follows from Theorem 3
that {o
n
} converges to 1.
39. j =

1 +
2
r
i
i lnj = rln
1 +
2
r
, so
lim
i<"
lnj = lim
i<"
ln(1 + 2r)
1r
H
= lim
i<"
1
1 + 2r
3
2
r
2
31r
2
= lim
i<"
2
1 + 2r
= 2 i
lim
i<"
1 +
2
r
i
= lim
i<"
c
ln u
= c
2
, so by Theorem 3, lim
n<"
1 +
2
n
n
= c
2
. Convergent
41. o
n
= ln(2n
2
+ 1) 3ln(n
2
+ 1) = ln
2n
2
+ 1
n
2
+ 1
= ln
2 + 1n
2
1 + 1n
2
<ln2 as n <". Convergent

43. {0. 1. 0. 0. 1. 0. 0. 0. 1. . . .} diverges since the sequence takes on only two values, 0 and 1, and never stays arbitrarily close to
either one (or any other value) for n sufciently large.
45. o
n
=
n!
2
n
=
1
2

2
2

3
2

(n 31)
2

n
2
D
1
2

n
2
[for n 1] =
n
4
<"as n <", so {o
n
} diverges.
47. From the graph, it appears that the sequence converges to 1.
{(32c)
n
} converges to 0 by (9), and hence {1 + (32c)
n
}
converges to 1 + 0 = 1.
SECTION 11.1 SEQUENCES

457
49. From the graph, it appears that the sequence converges to
1
2
.
As n <",
o
n
=
3 + 2n
2
8n
2
+n
=
3n
2
+ 2
8 + 1n
i
0 + 2
8 + 0
=
1
4
=
1
2
,
so lim
n<"
o
n
=
1
2
.
51. From the graph, it appears that the sequence {o
n
} =

n
2
cos n
1 +n
2
is
divergent, since it oscillates between 1 and 31 (approximately). To
prove this, suppose that {o
n
} converges to 1. If /
n
=
n
2
1 +n
2
, then
{/
n
} converges to 1, and lim
n<"
o
n
/
n
=
1
1
= 1. But
o
n
/
n
= cos n, so
lim
n<"
o
n
/
n
does not exist. This contradiction shows that {o
n
} diverges.
53. From the graph, it appears that the sequence approaches 0.
0 < o
n
=
1 3 5 (2n 31)
(2n)
n
=
1
2n

3
2n

5
2n

2n 31
2n
$
1
2n
(1) (1) (1) =
1
2n
<0 as n <"
So by the Squeeze Theorem,
1 3 5 (2n 31)
(2n)
n
converges to 0.
55. (a) o
n
= 1000(1.06)
n
i o
1
= 1060, o
2
= 1123.60, o
3
= 1191.02, o
4
= 1262.48, and o
5
= 1338.23.
(b) lim
n<"
o
n
= 1000 lim
n<"
(1.06)
n
, so the sequence diverges by (9) with v = 1.06 1.
57. If |v| D 1, then {v
n
} diverges by (9), so {nv
n
} diverges also, since |nv
n
| = n|v
n
| D |v
n
|. If |v| < 1 then
lim
i<"
rv
i
= lim
i<"
r
v
3i
H
= lim
i<"
1
(3lnv) v
3i
= lim
i<"
v
i
3lnv
= 0, so lim
n<"
nv
n
= 0, and hence {nv
n
} converges
whenever |v| < 1.
59. Since {o
n
} is a decreasing sequence, o
n
o
n+1
for all n D 1. Because all of its terms lie between 5 and 8, {o
n
} is a
bounded sequence. By the Monotonic Sequence Theorem, {o
n
} is convergent; that is, {o
n
} has a limit 1. 1 must be less than
8 since {o
n
} is decreasing, so 5 $ 1 < 8.
61. o
n
=
1
2n + 3
is decreasing since o
n+1
=
1
2(n + 1) + 3
=
1
2n + 5
<
1
2n + 3
= o
n
for each n D 1. The sequence is
bounded since 0 < o
n
$
1
5
for all n D 1. Note that o
1
=
1
5
.
458

63. The terms of o
n
= n(31)
n
alternate in sign, so the sequence is not monotonic. The rst ve terms are 31, 2, 33, 4, and 35.
Since lim
n<"
|o
n
| = lim
n<"
n = ", the sequence is not bounded.
65. o
n
=
n
n
2
+ 1
denes a decreasing sequence since for 1(r) =
r
r
2
+ 1
, 1
0
(r) =
r
2
+ 1
(1) 3r(2r)
(r
2
+ 1)
2
=
1 3r
2
(r
2
+ 1)
2
$ 0
for r D 1. The sequence is bounded since 0 < o
n
$
1
2
for all n D 1.
67. For
I
2,
2
I
2,
2
I
2, . . .
, o
1
= 2
1'2
, o
2
= 2
3'4
, o
3
= 2
7'8
, . . ., so o
n
= 2
(2
q
31)'2
q
= 2
13(1'2
q
)
.
lim
n<"
o
n
= lim
n<"
2
13(1'2
q
)
= 2
1
= 2.
Alternate solution: Let 1 = lim
n<"
o
n
. (We could show the limit exists by showing that {o
n
} is bounded and increasing.)
Then 1 must satisfy 1 =
I
2 1 i 1
2
= 21 i 1(1 32) = 0. 1 6= 0 since the sequence increases, so 1 = 2.
69. o
1
= 1, o
n+1
= 3 3
1
o
n
. We show by induction that {o
n
} is increasing and bounded above by 3. Let 1
n
be the proposition
that o
n+1
o
n
and 0 < o
n
< 3. Clearly 1
1
is true. Assume that 1
n
is true. Then o
n+1
o
n
i
1
o
n+1
<
1
o
n
i
3
1
o
n+1
3
1
o
n
. Now o
n+2
= 3 3
1
o
n+1
3 3
1
o
n
= o
n+1
C 1
n+1
. This proves that {o
n
} is increasing and bounded
above by 3, so 1 = o
1
< o
n
< 3, that is, {o
n
} is bounded, and hence convergent by the Monotonic Sequence Theorem.
If 1 = lim
n<"
o
n
, then lim
n<"
o
n+1
= 1 also, so 1 must satisfy 1 = 3 311 i 1
2
331 + 1 = 0 i 1 =
3
I
5
2
.
But 1 1, so 1 =
3 +
I
5
2
.
71. (a) Let o
n
be the number of rabbit pairs in the nth month. Clearly o
1
= 1 = o
2
. In the nth month, each pair that is
2 or more months old (that is, o
n32
pairs) will produce a new pair to add to the o
n31
pairs already present. Thus,
o
n
= o
n31
+o
n32
, so that {o
n
} = {1
n
}, the Fibonacci sequence.
(b) o
n
=
1
n+1
1
n
i o
n31
=
1
n
1
n31
=
1
n31
+1
n32
1
n31
= 1 +
1
n32
1
n31
= 1 +
1
1
n31
/1
n32
= 1 +
1
o
n32
. If 1 = lim
n<"
o
n
,
then 1 = lim
n<"
o
n31
and 1 = lim
n<"
o
n32
, so 1 must satisfy 1 = 1 +
1
1
i 1
2
31 31 = 0 i 1 =
1 +
I
5
2
[since 1 must be positive].
73. (a) From the graph, it appears that the sequence
n
5
n!
converges to 0, that is, lim

n<"
n
5
n!
= 0.
SECTION 11.1 SEQUENCES

459
(b)
From the rst graph, it seems that the smallest possible value of ` corresponding to - = 0.1 is 9, since n
5
n! < 0.1
whenever n D 10, but 9
5
9! 0.1. From the second graph, it seems that for - = 0.001, the smallest possible value for `
is 11 since n
5
n! < 0.001 whenever n D 12.
75. Theorem 6: If lim
n<"
|o
n
| = 0 then lim
n<"
3|o
n
| = 0, and since 3|o
n
| $ o
n
$ |o
n
|, we have that lim
n<"
o
n
= 0 by the
Squeeze Theorem.
77. To Prove: If lim
n<"
o
n
= 0 and {/
n
} is bounded, then lim
n<"
(o
n
/
n
) = 0.
Proof: Since {/
n
} is bounded, there is a positive number A such that |/
n
| $ A and hence, |o
n
| |/
n
| $ |o
n
| A for
all n D 1. Let - 0 be given. Since lim
n<"
o
n
= 0, there is an integer ` such that |o
n
30| <
-
A
if n `. Then
|o
n
/
n
30| = |o
n
/
n
| = |o
n
| |/
n
| $ |o
n
| A = |o
n
30| A <
-
A
A = - for all n `. Since - was arbitrary,
lim
n<"
(o
n
/
n
) = 0.
79. (a) First we show that o o
1
/
1
/.
o
1
3/
1
=
a +l
2
3
I
o/ =
1
2
o 32
I
o/ +/
=
1
2
I
o 3
I
/
2
0 [since o /] i o
1
/
1
. Also
o 3o
1
= o 3
1
2
(o +/) =
1
2
(o 3/) 0 and / 3/
1
= / 3
I
o/ =
I
/
I
/ 3
I
o
< 0, so o o
1
/
1
/. In the same
way we can show that o
1
o
2
/
2
/
1
and so the given assertion is true for n = 1. Suppose it is true for n = I, that is,
o
I
o
I+1
/
I+1
/
I
. Then
o
I+2
3/
I+2
=
1
2
(o
I+1
+/
I+1
) 3
o
I+1
/
I+1
=
1
2
o
I+1
32
o
I+1
/
I+1
+/
I+1
=
1
2
I
o
I+1
3
/
I+1
2
0,
o
I+1
3o
I+2
= o
I+1
3
1
2
(o
I+1
+/
I+1
) =
1
2
(o
I+1
3/
I+1
) 0, and
/
I+1
3/
I+2
= /
I+1
3
o
I+1
/
I+1
=

/
I+1
/
I+1
3
I
o
I+1
< 0 i o
I+1
o
I+2
/
I+2
/
I+1
,
so the assertion is true for n = I + 1. Thus, it is true for all n by mathematical induction.
(b) From part (a) we have o o
n
o
n+1
/
n+1
/
n
/, which shows that both sequences, {o
n
} and {/
n
}, are
monotonic and bounded. So they are both convergent by the Monotonic Sequence Theorem.
(c) Let lim
n<"
o
n
= c and lim
n<"
/
n
= o. Then lim
n<"
o
n+1
= lim
n<"
o
n
+/
n
2
i c =
c +o
2
i
2c = c +o i c = o.
460

81. (a) Suppose {j
n
} converges to j. Then j
n+1
=
/j
n
o +j
n
i lim
n<"
j
n+1
=
/ lim
n<"
j
n
o + lim
n<"
j
n
i j =
/j
o +j
i
j
2
+oj = /j i j(j +o 3/) = 0 i j = 0 or j = / 3o.
(b) j
n+1
=
/j
n
o +j
n
=
/
o
j
n
1 +
j
n
o
<

/
o
j
n
since 1 +
j
n
o
1.
(c) By part (b), j
1
<

/
o
j
0
, j
2
<

/
o
j
1
<

/
o
2
j
0
, j
3
<

/
o
j
2
<

/
o
3
j
0
, etc. In general, j
n
<

/
o
n
j
0
,
so lim
n<"
j
n
$ lim
n<"
/
o
n
j
0
= 0 since / < o.

By result 9. lim
n<"
v
n
= 0 if 31 < v < 1. Here v =
/
o
M (0. 1) .
(d) Let o < /. We rst show, by induction, that if j

0
< / 3o, then j
n
< / 3o and j
n+1
j
n
.
For n = 0, we have j
1
3j
0
=
/j
0
o +j
0
3j
0
=
j
0
(/ 3o 3j
0
)
o +j
0
0 since j
0
< / 3o. So j
1
j
0
.
Now we suppose the assertion is true for n = I, that is, j
I
< / 3o and j
I+1
j
I
. Then
/ 3o 3j
I+1
= / 3o 3
/j
I
o +j
I
=
o(/ 3o) +/j
I
3oj
I
3/j
I
o +j
I
=
o(/ 3o 3j
I
)
o +j
I
0 because j
I
< / 3o. So
j
I+1
< / 3o. And j
I+2
3j
I+1
=
/j
I+1
o +j
I+1
3j
I+1
=
j
I+1
(/ 3o 3j
I+1
)
o +j
I+1
0 since j
I+1
< / 3o. Therefore,
j
I+2
j
I+1
. Thus, the assertion is true for n = I + 1. It is therefore true for all n by mathematical induction.
A similar proof by induction shows that if j
0
/ 3o, then j
n
/ 3o and {j
n
} is decreasing.
In either case the sequence {j
n
} is bounded and monotonic, so it is convergent by the Monotonic Sequence Theorem.
It then follows from part (a) that lim
n<"
j
n
= / 3o.
11.2 Series
1. (a) A sequence is an ordered list of numbers whereas a series is the sum of a list of numbers.
(b) A series is convergent if the sequence of partial sums is a convergent sequence. A series is divergent if it is not convergent.
3.
n c
n
1 32.40000
2 31.92000
3 32.01600
4 31.99680
5 32.00064
6 31.99987
7 32.00003
8 31.99999
9 32.00000
10 32.00000
From the graph and the table, it seems that the series converges to 32. In fact, it is a geometric
series with o = 32.4 and v = 3
1
5
, so its sum is
"
n=1
12
(35)
n
=
32.4
1 3
3
1
5
=
32.4
1.2
= 32.
Note that the dot corresponding to n = 1 is part of both {o
n
} and {c
n
}.
TI-86 Note: To graph {o
n
} and {c
n
}, set your calculator to Param mode and DrawDot mode. (DrawDot is under
SECTION 11.2 SERIES

461
GRAPH, MORE, FORMT (F3).) Now under E(t)= make the assignments: xt1=t, yt1=12/(-5)t, xt2=t,
yt2=sum seq(yt1,t,1,t,1). (sum and seq are under LIST, OPS (F5), MORE.) Under WIND use
1,10,1,0,10,1,-3,1,1 to obtain a graph similar to the one above. Then use TRACE (F4) to see the values.
5.
n c
n
1 1.55741
2 30.62763
3 30.77018
4 0.38764
5 32.99287
6 33.28388
7 32.41243
8 39.21214
9 39.66446
10 39.01610
The series
"
n=1
tann diverges, since its terms do not approach 0.
7.
n c
n
1 0.29289
2 0.42265
3 0.50000
4 0.55279
5 0.59175
6 0.62204
7 0.64645
8 0.66667
9 0.68377
10 0.69849
From the graph and the table, it seems that the series converges.
I
n=1
1
I
n
3
1
I
n + 1
=

1
I
1
3
1
I
2
1
I
2
3
1
I
3
+ +
1
I
I
3
1
I
I + 1
= 1 3
1
I
I + 1
,
so
"
n=1
1
I
n
3
1
I
n + 1
= lim
I<"
1 3
1
I
I + 1
= 1.
9. (a) lim
n<"
o
n
= lim
n<"
2n
3n + 1
=
2
3
, so the sequence {o
n
} is convergent by (11.1.1).
(b) Since lim
n<"
o
n
=
2
3
6= 0, the series
"
n=1
o
n
is divergent by the Test for Divergence.
11. 3 + 2 +
4
3
+
8
9
+ is a geometric series with rst term o = 3 and common ratio v =
2
3
. Since |v| =
2
3
< 1, the series
converges to
a
1 3r
=
3
132'3
=
3
1'3
= 9.
13. 3 34 +
16
3
3
64
9
+ is a geometric series with ratio v = 3
4
3
. Since |v| =
4
3
1, the series diverges.
462

15.
"
n=1
6(0.9)
n31
is a geometric series with rst term o = 6 and ratio v = 0.9. Since |v| = 0.9 < 1, the series converges to
o
1 3v
=
6
1 30.9
=
6
0.1
= 60.
17.
"
n=1
(33)
n31
4
n
=
1
4
"
n=1
3
3
4
n31
. The latter series is geometric with o = 1 and ratio v = 3
3
4
. Since |v| =
3
4
< 1, it
converges to
1
1 3(334)
=
4
7
. Thus, the given series converges to
1
4
4
7
=
1
7
.
19.
"
n=0
n
3
n+1
=
1
3
"
n=0
n
is a geometric series with ratio v =

3
. Since |v| 1, the series diverges.
21.
"
n=1
1
2n
=
1
2
"
n=1
1
n
diverges since each of its partial sums is
1
2
times the corresponding partial sum of the harmonic series
"
n=1
1
n
, which diverges.

If
"
n=1
1
2n
were to converge, then
"
n=1
1
n
would also have to converge by Theorem 8(i).
In general, constant multiples of divergent series are divergent.

23.
"
I=2
I
2
I
2
31
diverges by the Test for Divergence since lim
I<"
o
I
= lim
I<"
I
2
I
2
31
= 1 6= 0.
25. Converges.
"
n=1
1 + 2
n
3
n
=
"
n=1
1
3
n
+
2
n
3
n
=
"
n=1
1
3
n
+
2
3
[sum of two convergent geometric series]

=
13
1 313
+
23
1 323
=
1
2
+ 2 =
5
2
27.
"
n=1
q
I
2 = 2 +
I
2 +
3
I
2 +
4
I
2 + diverges by the Test for Divergence since
lim
n<"
o
n
= lim
n<"
q
I
2 = lim
n<"
2
1'n
= 2
0
= 1 6= 0.
29.
"
n=1
ln
n
2
+ 1
2n
2
+ 1
diverges by the Test for Divergence since

lim
n<"
o
n
= lim
n<"
ln
n
2
+ 1
2n
2
+ 1
= ln
lim
n<"
n
2
+ 1
2n
2
+ 1
= ln
1
2
6= 0.
31.
"
n=1
arctan n diverges by the Test for Divergence since lim
n<"
o
n
= lim
n<"
arctann =
r
2
6= 0.
33.
"
n=1
1
c
n
=
"
n=1
1
c
n
is a geometric series with rst term o =
1
c
and ratio v =
1
c
. Since |v| =
1
c
< 1, the series converges
to
1c
1 31c
=
1c
1 31c

c
c
=
1
c 31
. By Example 6,
"
n=1
1
n(n + 1)
= 1. Thus, by Theorem 8(ii),
"
n=1
1
c
n
+
1
n(n + 1)
=
"
n=1
1
c
n
+
"
n=1
1
n(n + 1)
=
1
c 31
+ 1 =
1
c 31
+
c 31
c 31
=
c
c 31
.
SECTION 11.2 SERIES

463
35. Using partial fractions, the partial sums of the series
"
n=2
2
n
2
31
are
c
n
=
n
.=2
2
(j 31)(j + 1)
=
n
.=2
1
j 31
3
1
j + 1
=

1 3
1
3
1
2
3
1
4
1
3
3
1
5
+ +
1
n 33
3
1
n 31
1
n 32
3
1
n
This sum is a telescoping series and c

n
= 1 +
1
2
3
1
n 31
3
1
n
.
Thus,
"
n=2
2
n
2
31
= lim
n<"
c
n
= lim
n<"
1 +
1
2
3
1
n 31
3
1
n
=
3
2
.
37. For the series
"
n=1
3
n(n + 3)
, c
n
=
n
.=1
3
j(j + 3)
=
n
.=1
1
j
3
1
j + 3
[using partial fractions]. The latter sum is
1 3
1
4
1
2
3
1
5
1
3
3
1
6
1
4
3
1
7
+ +
1
n33
3
1
n
1
n32
3
1
n+1
1
n31
3
1
n+2
1
n
3
1
n+3
= 1 +
1
2
+
1
3
3
1
n+1
3
1
n+2
3
1
n+3
[telescoping series]
Thus,
"
n=1
3
n(n + 3)
= lim
n<"
c
n
= lim
n<"
1 +
1
2
+
1
3
3
1
n+1
3
1
n+2
3
1
n+3
= 1 +
1
2
+
1
3
=
11
6
. Converges
39. For the series
"
n=1
c
1'n
3c
1'(n+1)
,
c
n
=
n
.=1
c
1'.
3c
1'(.+1)
= (c
1
3c
1'2
) + (c
1'2
3c
1'3
) + +
c
1'n
3c
1'(n+1)
= c 3c
1'(n+1)
[telescoping series]
Thus,
"
n=1
c
1'n
3c
1'(n+1)
= lim
n<"
c
n
= lim
n<"
c 3c
1'(n+1)
= c 3c
0
= c 31. Converges
41. 0.2 =
2
10
+
2
10
2
+ is a geometric series with o =
2
10
and v =
1
10
. It converges to
o
1 3v
=
210
1 3110
=
2
9
.
43. 3.417 = 3 +
417
10
3
+
417
10
6
+ . Now
417
10
3
+
417
10
6
417
10
3
and v =
1
10
3
.
It converges to
o
1 3v
=
41710
3
1 3110
3
=
41710
3
99910
3
=
417
999
. Thus, 3.417 = 3 +
417
999
=
3414
999
=
1138
333
.
45. 1.5342 = 1.53 +
42
10
4
+
42
10
6
+ . Now
42
10
4
+
42
10
6
42
10
4
and v =
1
10
2
.
It converges to
o
1 3v
=
4210
4
1 3110
2
=
4210
4
9910
2
=
42
9900
.
Thus, 1.5342 = 1.53 +
42
9900
=
153
100
+
42
9900
=
15,147
9900
+
42
9900
=
15,189
9900
or
5063
3300
.
47.
"
n=1
r
n
3
n
=
"
n=1
r
3
n
is a geometric series with v =
r
3
, so the series converges C |v| < 1 C
|r|
3
< 1 C |r| < 3;
that is, 33 < r < 3. In that case, the sum of the series is
o
1 3v
=
r3
1 3r3
=
r3
1 3r3

3
3
=
r
3 3r
.
464

49.
"
n=0
4
n
r
n
=
"
n=0
(4r)
n
is a geometric series with v = 4r, so the series converges C |v| < 1 C 4 |r| < 1 C
|r| <
1
4
. In that case, the sum of the series is
1
1 34r
.
51.
"
n=0
cos
n
r
2
n
is a geometric series with rst term 1 and ratio v =
cos r
2
, so it converges C |v| < 1. But |v| =
|cos r|
2
$
1
2
for all r. Thus, the series converges for all real values of r and the sum of the series is
1
1 3(cos r)2
=
2
2 3cos r
.
53. After dening 1, We use convert(f,parfrac); in Maple, Apart in Mathematica, or Expand Rational and
Simplify in Derive to nd that the general term is
3n
2
+ 3n + 1
(n
2
+n)
3
=
1
n
3
3
1
(n + 1)
3
. So the nth partial sum is
c
n
=
n
I=1
1
I
3
3
1
(I + 1)
3
=

1 3
1
2
3
1
2
3
3
1
3
3
+ +
1
n
3
3
1
(n + 1)
3
= 1 3
1
(n + 1)
3
The series converges to lim
n<"
c
n
= 1. This can be conrmed by directly computing the sum using sum(f,1..infinity);
(in Maple), Sum[f,{n,1,Infinity}] (in Mathematica), or Calculus Sum (from 1 to ") and Simplify (in Derive).
55. For n = 1, o
1
= 0 since c
1
= 0. For n 1,
o
n
= c
n
3c
n31
=
n 31
n + 1
3
(n 31) 31
(n 31) + 1
=
(n 31)n 3(n + 1)(n 32)
(n + 1)n
=
2
n(n + 1)
Also,
"
n=1
o
n
= lim
n<"
c
n
= lim
n<"
1 31n
1 + 1n
= 1.
57. (a) The rst step in the chain occurs when the local government spends 1 dollars. The people who receive it spend a fraction c
of those 1 dollars, that is, 1c dollars. Those who receive the 1c dollars spend a fraction c of it, that is, 1c
2
dollars.
Continuing in this way, we see that the total spending after n transactions is
o
n
= 1 +1c +1c
2
+ +1c
n1
=
1(1 3c
n
)
1 3c
by (3).
(b) lim
n<"
o
n
= lim
n<"
1(1 3c
n
)
1 3c
=
1
1 3c
lim
n<"
(1 3c
n
) =
1
1 3c
since 0 < c < 1 i lim

n<"
c
n
= 0
=
1
c
[since c +c = 1] = I1 [since I = 1c]
If c = 0.8, then c = 1 3c = 0.2 and the multiplier is I = 1c = 5.
59.
"
n=2
(1 +c)
3n
is a geometric series with o = (1 +c)
32
and v = (1 +c)
31
, so the series converges when
(1 +c)
31
< 1 C |1 +c| 1 C 1 +c 1 or 1 +c < 31 C c 0 or c < 32. We calculate the sum of the

series and set it equal to 2:
(1 +c)
32
1 3(1 +c)
31
= 2 C

1
1 +c
2
= 2 32
1
1 +c
C 1 = 2(1 +c)
2
32(1 +c) C
2c
2
+ 2c 31 = 0 C c =
32
I
12
4
=

I
3 31
2
. However, the negative root is inadmissible because 32 <
3
I
3 31
2
< 0.
So c =
I
3 31
2
.
SECTION 11.2 SERIES

465
61. c
s
q
= c
1+
1
2
+
1
3
++
1
n
= c
1
c
1'2
c
1'3
c
1'n
(1 + 1)
1 +
1
2

1 +
1
3
1 +
1
n
[c
i
1 +r]
=
2
1
3
2
4
3

n + 1
n
= n + 1
Thus, c
s
q
n + 1 and lim
n<"
c
s
q
= ". Since {c
n
} is increasing, lim
n<"
c
n
= ", implying that the harmonic series is
divergent.
63. Let d
n
be the diameter of C
n
. We draw lines from the centers of the C
.
to
the center of 1 (or C), and using the Pythagorean Theorem, we can write
1
2
+
1 3
1
2
d
1
2
=

1 +
1
2
d
1
2
C
1 =
1 +
1
2
d
1
2
3
1 3
1
2
d
1
2
= 2d
1
[difference of squares] i d
1
=
1
2
.
Similarly,
1 =
1 +
1
2
d
2
2
3
1 3d
1
3
1
2
d
2
2
= 2d
2
+ 2d
1
3d
2
1
3d
1
d
2
= (2 3d
1
)(d
1
+d
2
) C
d
2
=
1
2 3d
1
3d
1
=
(1 3d
1
)
2
2 3d
1
, 1 =

1 +
1
2
d
3
2
3
1 3d
1
3d
2
3
1
2
d
3
2
C d
3
=
[1 3(d
1
+d
2
)]
2
2 3(d
1
+d
2
)
, and in general,
d
n+1
=
1 3
n
.=1
d
.
2
2 3
n
.=1
d
.
. If we actually calculate d
2
and d
3
from the formulas above, we nd that they are
1
6
=
1
2 3
and
1
12
=
1
3 4
respectively, so we suspect that in general, d
n
=
1
n(n + 1)
. To prove this, we use induction: Assume that for all
I $ n, d
I
=
1
I(I + 1)
=
1
I
3
1
I + 1
. Then
n
.=1
d
.
= 1 3
1
n + 1
=
n
n + 1
[telescoping sum]. Substituting this into our
formula for d
n+1
, we get d
n+1
=
1 3
n
n + 1
2
2 3
n
n + 1
=
1
(n + 1)
2
n + 2
n + 1
=
1
(n + 1)(n + 2)
, and the induction is complete.
Now, we observe that the partial sums
n
.=1
d
.
of the diameters of the circles approach 1 as n <"; that is,
"
n=1
o
n
=
"
n=1
1
n(n + 1)
= 1, which is what we wanted to prove.
65. The series 1 31 + 1 31 + 1 31 + diverges (geometric series with v = 31) so we cannot say that
0 = 1 31 + 1 31 + 1 31 + .
67.

"
n=1
co
n
= lim
n<"
n
.=1
co
.
= lim
n<"
c
n
.=1
o
.
= c lim
n<"
n
.=1
o
.
= c
"
n=1
o
n
, which exists by hypothesis.
69. Suppose on the contrary that
(o
n
+/
n
) converges. Then
(o
n
+/
n
) and
o
n
are convergent series. So by Theorem 8,
[(o
n
+/
n
) 3o
n
] would also be convergent. But
[(o
n
+/
n
) 3o
n
] =

/
n
, a contradiction, since
/
n
is given to be
divergent.
71. The partial sums {c
n
} form an increasing sequence, since c
n
3c
n31
= o
n
0 for all n. Also, the sequence {c
n
} is bounded
since c
n
$ 1000 for all n. So by the Monotonic Sequence Theorem, the sequence of partial sums converges, that is, the series
o
n
is convergent.
466

73. (a) At the rst step, only the interval
1
3
.
2
3
(length
1
3
) is removed. At the second step, we remove the intervals
1
9
.
2
9
and
7
9
.
8
9
, which have a total length of 2
1
3
2
. At the third step, we remove 2
2
intervals, each of length
1
3
3
. In general,
at the nth step we remove 2
n31
intervals, each of length
1
3
n
, for a length of 2
n31
1
3
n
=
1
3
2
3
n31
. Thus, the total
length of all removed intervals is
"
n=1
1
3
2
3
n31
=
1'3
1 32'3
= 1
geometric series with o =

1
3
and v =
2
3
. Notice that at
the nth step, the leftmost interval that is removed is
1
3
n
.
2
3
, so we never remove 0, and 0 is in the Cantor set. Also,

the rightmost interval removed is
1 3
2
3
n
. 1 3
1
3
, so 1 is never removed. Some other numbers in the Cantor set

are
1
3
,
2
3
,
1
9
,
2
9
,
7
9
, and
8
9
.
(b) The area removed at the rst step is
1
9
; at the second step, 8
1
9
2
; at the third step, (8)
2
1
9
3
. In general, the area
removed at the nth step is (8)
n31
1
9
n
=
1
9
8
9
n31
, so the total area of all removed squares is
"
n=1
1
9
8
9
n31
=
19
1 3 89
= 1.
75. (a) For
"
n=1
n
(n + 1)!
, c
1
=
1
1 2
=
1
2
, c
2
=
1
2
+
2
1 2 3
=
5
6
, c
3
=
5
6
+
3
1 2 3 4
=
23
24
,
c
4
=
23
24
+
4
1 2 3 4 5
=
119
120
. The denominators are (n + 1)!, so a guess would be c
n
=
(n + 1)! 31
(n + 1)!
.
(b) For n = 1, c
1
=
1
2
=
2! 31
2!
, so the formula holds for n = 1. Assume c
I
=
(I + 1)! 31
(I + 1)!
. Then
c
I+1
=
(I + 1)! 31
(I + 1)!
+
I + 1
(I + 2)!
=
(I + 1)! 31
(I + 1)!
+
I + 1
(I + 1)!(I + 2)
=
(I + 2)! 3(I + 2) +I + 1
(I + 2)!
=
(I + 2)! 31
(I + 2)!
Thus, the formula is true for n = I + 1. So by induction, the guess is correct.
(c) lim
n<"
c
n
= lim
n<"
(n + 1)! 31
(n + 1)!
= lim
n<"
1 3
1
(n + 1)!
= 1 and so
"
n=1
n
(n + 1)!
= 1.
11.3 The Integral Test and Estimates of Sums
1. The picture shows that o
2
=
1
2
1.3
<

2
1
1
r
1.3
dr,
o
3
=
1
3
1.3
<

3
2
1
r
1.3
dr, and so on, so
"
n=2
1
n
1.3
<

"
1
1
r
1.3
dr. The
integral converges by (7.8.2) with j = 1.3 1, so the series converges.
3. The function 1(r) = 1
5
I
r = r
31'5
is continuous, positive, and decreasing on [1. "), so the Integral Test applies.
"
1
r
31'5
dr = lim
I<"
I
1
r
31'5
dr = lim
I<"
5
4
r
4'5
I
1
= lim
I<"
5
4
t
4'5
3
5
4
= ", so
"
n=1
1
5
I
n diverges.
SECTION 11.3 THE INTEGRAL TEST AND ESTIMATES OF SUMS

467
5. The function 1(r) =
1
(2r + 1)
3
"
1
1
(2r + 1)
3
dr = lim
I<"
I
1
1
(2r + 1)
3
dr = lim
I<"
3
1
4
1
(2r + 1)
2
I
1
= lim
I<"
3
1
4(2t + 1)
2
+
1
36
=
1
36
.
Since this improper integral is convergent, the series
"
n=1
1
(2n + 1)
3
is also convergent by the Integral Test.
7. 1(r) = rc
3i
is continuous and positive on [1. "). 1
0
(r) = 3rc
3i
+c
3i
= c
3i
(1 3r) < 0 for r 1, so 1 is decreasing
on [1. "). Thus, the Integral Test applies.
"
1
rc
3i
dr = lim
l<"
l
1
rc
3i
dr = lim
l<"
3rc
3i
3c
3i
l
1
[by parts] = lim
l<"
[3/c
3l
3c
3l
+c
31
+c
31
] = 2c
since lim
l<"
/c
3l
= lim
l<"
(/c
l
)
H
= lim
l<"
(1c
l
) = 0 and lim
l<"
c
3l
= 0. Thus,
"
n=1
nc
3n
converges.
9. The series
"
n=1
1
n
0.85
is a j-series with j = 0.85 $ 1, so it diverges by (1). Therefore, the series
"
n=1
2
n
0.85
must also diverge,
for if it converged, then
"
n=1
1
n
0.85
would have to converge [by Theorem 8(i) in Section 11.2].
11. 1 +
1
8
+
1
27
+
1
64
+
1
125
+ =
"
n=1
1
n
3
. This is a j-series with j = 3 1, so it converges by (1).
13. 1 +
1
3
+
1
5
+
1
7
+
1
9
+ =
"
n=1
1
2n 31
. The function 1(r) =
1
2r 31
is
continuous, positive, and decreasing on [1. "), so the Integral Test applies.
"
1
1
2r 31
dr = lim
I<"
I
1
1
2r 31
dr = lim
I<"
1
2
ln|2r 31|
I
1
=
1
2
lim
I<"
(ln(2t 31) 30) = ", so the series
"
n=1
1
2n 31
diverges.
15.
"
n=1
5 32
I
n
n
3
= 5
"
n=1
1
n
3
32
"
n=1
1
n
5'2
by Theorem 11.2.8, since
"
n=1
1
n
3
and
"
n=1
1
n
5'2
both converge by (1)
with j = 3 1 and j =
5
2
1
. Thus,
"
n=1
5 32
I
n
n
3
converges.
1
r
2
+ 4
is continuous, positive, and decreasing on [1. "), so we can apply the Integral Test.
"
1
1
r
2
+ 4
dr = lim
I<"
I
1
1
r
2
+ 4
dr = lim
I<"
1
2
tan
31
r
2
I
1
=
1
2
lim
I<"
tan
31
t
2
3tan
31
1
2
=
1
2
2
3tan
31
1
2
Therefore, the series

"
n=1
1
n
2
+ 4
converges.
19.
"
n=1
lnn
n
3
=
"
n=2
lnn
n
3
since
ln1
1
= 0. The function 1(r) =
lnr
r
3
is continuous and positive on [2. ").
1
0
(r) =
r
3
(1r) 3(lnr)(3r
2
)
(r
3
)
2
=
r
2
33r
2
lnr
r
6
=
1 33 lnr
r
4
< 0 C 1 33 lnr < 0 C lnr
1
3
C
468

r c
1'3
E 1.4, so 1 is decreasing on [2. "), and the Integral Test applies.
"
2
lnr
r
3
dr = lim
I<"
I
2
lnr
r
3
dr
(+)
= lim
I<"
3
lnr
2r
2
3
1
4r
2
I
1
= lim
I<"
3
1
4t
2
(2 lnt + 1) +
1
4
(++)
=
1
4
, so the series
"
n=2
lnn
n
3
converges.
(-): & = lnr, d = r
33
dr i d& = (1r) dr, = 3
1
2
r
32
, so
lnr
r
3
dr = 3
1
2
r
32
lnr 3
3
1
2
r
32
(1r) dr = 3
1
2
r
32
lnr +
1
2
r
33
dr = 3
1
2
r
32
lnr 3
1
4
r
32
+C.
(--): lim
I<"
3
2 lnt + 1
4t
2
H
= 3 lim
I<"
2t
8t
= 3
1
4
lim
I<"
1
t
2
= 0.
21. 1(r) =
1
rlnr
is continuous and positive on [2. "), and also decreasing since 1
0
(r) = 3
1 + lnr
r
2
(lnr)
2
< 0 for r 2, so we can
use the Integral Test.
"
2
1
rlnr
dr = lim
I<"
[ln(lnr)]
I
2
= lim
I<"
[ln(lnt) 3ln(ln2)] = ", so the series
"
n=2
1
nlnn
diverges.
23. The function 1(r) = c
1'i
r
2
[o(r) = c
1'i
is decreasing and dividing by r
2
doesnt change that fact.]
"
1
1(r) dr = lim
I<"
I
1
c
1'i
r
2
dr = lim
I<"
3c
1'i
I
1
= 3 lim
I<"
(c
1'I
3c) = 3(1 3c) = c 31, so the series
"
n=1
c
1'n
n
2
converges.
1
r
3
+r
is continuous, positive, and decreasing on [1. "), so the Integral Test applies. We use partial
fractions to evaluate the integral:
"
1
1
r
3
+r
dr = lim
I<"
I
1
1
r
3
r
1 +r
2
dr = lim
I<"
lnr 3
1
2
ln(1 +r
2
)
I
1
= lim
I<"
ln
r
I
1 +r
2
I
1
= lim
I<"
ln
t
I
1 +t
2
3ln
1
I
2
= lim
I<"
ln
1
1 + 1t
2
+
1
2
ln2
=
1
2
ln2
so the series
"
n=1
1
n
3
+n
converges.
27. We have already shown (in Exercise 21) that when j = 1 the series
"
n=2
1
n(lnn)
diverges, so assume that j 6= 1.

1(r) =
1
r(lnr)
is continuous and positive on [2. "), and 1

0
(r) = 3
j + lnr
r
2
(lnr)
+1
< 0 if r c
3
, so that 1 is eventually
decreasing and we can use the Integral Test.
"
2
1
r(lnr)
dr = lim
I<"
(lnr)
13
1 3j
I
2
[for j 6= 1] = lim
I<"
(lnt)
13
1 3j
3
(ln2)
13
1 3j
This limit exists whenever 1 3j < 0 C j 1, so the series converges for j 1.

SECTION 11.3 THE INTEGRAL TEST AND ESTIMATES OF SUMS

469
29. Clearly the series cannot converge if j D 3
1
2
, because then lim
n<"
n(1 +n
2
)
6= 0. So assume j < 3
1
2
. Then
1(r) = r(1 +r
2
)
is continuous, positive, and eventually decreasing on [1. "), and we can use the Integral Test.
"
1
r(1 +r
2
)
dr = lim
I<"
1
2

(1 +r
2
)
+1
j + 1
I
1
=
1
2(j + 1)
lim
I<"
[(1 +t
2
)
+1
32
+1
].
This limit exists and is nite C j + 1 < 0 C j < 31, so the series converges whenever j < 31.
31. Since this is a j-series with j = r, c(r) is dened when r 1. Unless specied otherwise, the domain of a function 1 is the
set of real numbers r such that the expression for 1(r) makes sense and denes a real number. So, in the case of a series, its
the set of real numbers r such that the series is convergent.
33. (a) 1(r) =
1
r
2
is positive and continuous and 1
0
(r) = 3
2
r
3
is negative for r 0, and so the Integral Test applies.
"
n=1
1
n
2
E c
10
=
1
1
2
+
1
2
2
+
1
3
2
+ +
1
10
2
E 1.549768.
1
10
$
"
10
1
r
2
dr = lim
I<"
31
r
I
10
= lim
I<"
3
1
t
+
1
10
=
1
10
, so the error is at most 0.1.
(b) c
10
+
"
11
1
r
2
dr $ c $ c
10
+
"
10
1
r
2
dr i c
10
+
1
11
$ c $ c
10
+
1
10
i
1.549768 + 0.090909 = 1.640677 $ c $ 1.549768 + 0.1 = 1.649768, so we get c E 1.64522 (the average of 1.640677
and 1.649768) with error $ 0.005 (the maximum of 1.649768 31.64522 and 1.64522 31.640677, rounded up).
(c) 1
n
$
"
n
1
r
2
dr =
1
n
. So 1
n
< 0.001 if
1
n
<
1
1000
C n 1000.
35. 1(r) = 1(2r + 1)
6
is continuous, positive, and decreasing on [1. "), so the Integral Test applies. Using (2),
1
n
$
"
n
(2r + 1)
36
dr = lim
I<"
31
10(2r + 1)
5
I
n
=
1
10(2n + 1)
5
. To be correct to ve decimal places, we want
1
10(2n + 1)
5
$
5
10
6
C (2n + 1)
5
D 20,000 C n D
1
2
5
I
20,000 31
E 3.12, so use n = 4.
c
4
=
4
n=1
1
(2n + 1)
6
=
1
3
6
+
1
5
6
+
1
7
6
+
1
9
6
E 0.001 446 E 0.00145.
37.
"
n=1
n
31.001
=
"
n=1
1
n
1.001
is a convergent j-series with j = 1.001 1. Using (2), we get
1
n
$
"
n
r
31.001
dr = lim
I<"
r
30.001
30.001
I
n
= 31000 lim
I<"
1
r
0.001
I
n
= 31000
3
1
n
0.001
=
1000
n
0.001
.
We want 1
n
< 0.000 000 005 C
1000
n
0.001
< 5 10
39
C n
0.001
1000
5 10
39
C
n

2 10
11
1000
= 2
1000
10
11,000
E 1.07 10
301
10
11,000
= 1.07 10
11,301
.
470

39. (a) From the gure, o
2
+o
3
+ +o
n
$
n
1
1(r) dr, so with
1(r) =
1
r
,
1
2
+
1
3
+
1
4
+ +
1
n
$
n
1
1
r
dr = lnn.
Thus, c
n
= 1 +
1
2
+
1
3
+
1
4
+ +
1
n
$ 1 + lnn.
(b) By part (a), c
10
6 $ 1 + ln10
6
E 14.82 < 15 and
c
10
9 $ 1 + ln10
9
E 21.72 < 22.
41. /
ln n
=

c
ln l
ln n
=

c
ln n
ln l
= n
ln l
=
1
n
3ln l
. This is a j-series, which converges for all / such that 3ln/ 1 C
ln/ < 31 C / < c
31
C / < 1c [with / 0].
11.4 The Comparison Tests
1. (a) We cannot say anything about
o
n
. If o
n
/
n
for all n and
/
n
is convergent, then
o
n
could be convergent or
divergent. (See the note after Example 2.)
(b) If o
n
< /
n
for all n, then
o
n
is convergent. [This is part (i) of the Comparison Test.]
3.
n
2n
3
+ 1
<
n
2n
3
=
1
2n
2
<
1
n
2
for all n D 1, so
"
n=1
n
2n
3
+ 1
converges by comparison with
"
n=1
1
n
2
, which converges
because it is a p-series with j = 2 1.
5.
n + 1
n
I
n

n
n
I
n
=
1
I
n
for all n D 1, so
"
n=1
n + 1
n
I
n
diverges by comparison with
"
n=1
1
I
n
, which diverges because it is a
p-series with j =
1
2
$ 1.
7.
9
n
3 + 10
n
<
9
n
10
n
=

9
10
n
for all n D 1.
"
n=1
9
10
n
is a convergent geometric series
|v| =
9
10
< 1
, so
"
n=1
9
n
3 + 10
n
converges by the Comparison Test.
9.
cos
2
n
n
2
+ 1
$
1
n
2
+ 1
<
1
n
2
, so the series
"
n=1
cos
2
n
n
2
+ 1
converges by comparison with the j-series
"
n=1
1
n
2
[j = 2 1].
11.
n 31
n4
n
is positive for n 1 and
n 31
n4
n
<
n
n4
n
=
1
4
n
=

1
4
n
, so
"
n=1
n 31
n4
n
converges by comparison with the convergent
geometric series
"
n=1
1
4
n
.
13.
arctann
n
1.2
<
2
n
1.2
for all n D 1, so
"
n=1
arctann
n
1.2
converges by comparison with

2
"
n=1
1
n
1.2
, which converges because it is a
constant times a p-series with j = 1.2 1.
15.
2 + (31)
n
n
I
n
$
3
n
I
n
, and
"
n=1
3
n
I
n
converges because it is a constant multiple of the convergent j-series
"
n=1
1
n
I
n
j =
3
2
1
, so the given series converges by the Comparison Test.

SECTION 11.4 THE COMPARISON TESTS

471
17. Use the Limit Comparison Test with o
n
=
1
I
n
2
+ 1
and /
n
=
1
n
:
lim
n<"
o
n
/
n
= lim
n<"
n
I
n
2
+ 1
= lim
n<"
1
1 + (1n
2
)
= 1 0. Since the harmonic series
"
n=1
1
n
diverges, so does
"
n=1
1
I
n
2
+ 1
.
n
=
1 + 4
n
1 + 3
n
and /
n
=
4
n
3
n
:
lim
n<"
o
n
/
n
= lim
n<"
1 + 4
n
1 + 3
n
4
n
3
n
= lim
n<"
1 + 4
n
1 + 3
n

3
n
4
n
= lim
n<"
1 + 4
n
4
n

3
n
1 + 3
n
= lim
n<"
1
4
n
+ 1
1
1
3
n
+ 1
= 1 0
Since the geometric series
/
n
=

4
3
n
diverges, so does
"
n=1
1 + 4
n
1 + 3
n
. Alternatively, use the Comparison Test with
1 + 4
n
1 + 3
n

1 + 4
n
3
n
+ 3
n

4
n
2(3
n
)
=
1
2
4
3
n
or use the Test for Divergence.
n
=
I
n + 2
2n
2
+n + 1
and /
n
=
1
n
3'2
:
lim
n<"
o
n
/
n
= lim
n<"
n
3'2
I
n + 2
2n
2
+n + 1
= lim
n<"
(n
3'2
I
n + 2 )(n
3'2
I
n)
(2n
2
+n + 1)n
2
= lim
n<"
1 + 2n
2 + 1n + 1n
2
=
I
1
2
=
1
2
0.
Since
"
n=1
1
n
3'2
is a convergent p-series
j =
3
2
1
, the series
"
n=1
I
n + 2
2n
2
+n + 1
also converges.
n
=
5 + 2n
(1 +n
2
)
2
and /
n
=
1
n
3
:
lim
n<"
o
n
/
n
= lim
n<"
n
3
(5 + 2n)
(1 +n
2
)
2
= lim
n<"
5n
3
+ 2n
4
(1 +n
2
)
2

1n
4
1(n
2
)
2
= lim
n<"
5
n
+ 2
1
n
2
+ 1
2
= 2 0. Since
"
n=1
1
n
3
is a convergent
j-series [j = 3 1], the series
"
n=1
5 + 2n
(1 +n
2
)
2
also converges.
25. If o
n
=
1 +n +n
2
I
1 +n
2
+n
6
and /
n
=
1
n
, then lim
n<"
o
n
/
n
= lim
n<"
n +n
2
+n
3
I
1 +n
2
+n
6
= lim
n<"
1n
2
+ 1n + 1
1n
6
+ 1n
4
+ 1
= 1 0,
so
"
n=1
1 +n +n
2
I
1 +n
2
+n
6
diverges by the Limit Comparison Test with the divergent harmonic series
"
n=1
1
n
.
n
=

1 +
1
n
2
c
3n
and /
n
= c
3n
: lim
n<"
o
n
/
n
= lim
n<"
1 +
1
n
2
= 1 0. Since
"
n=1
c
3n
=
"
n=1
1
c
n
|v| =
1
c
< 1
, the series
"
n=1
1 +
1
n
2
c
3n
also converges.
29. Clearly n! = n(n 31)(n 32) (3)(2) D 2 2 2 2 2 = 2
n31
, so
1
n!
$
1
2
n31
.
"
n=1
1
2
n31
is a convergent geometric
series
|v| =
1
2
< 1
, so
"
n=1
1
n!
472

n
= sin
1
n
and /
n
=
1
n
. Then
o
n
and
/
n
are series with positive terms and
lim
n<"
o
n
/
n
= lim
n<"
sin(1n)
1n
= lim
0<0
sin0
0
= 1 0. Since
"
n=1
/
n
is the divergent harmonic series,
"
n=1
sin(1n) also diverges. [Note that we could also use lHospitals Rule to evaluate the limit:
lim
i<"
sin(1r)
1r
H
= lim
i<"
cos(1r)
31r
2
31r
2
= lim
i<"
cos
1
r
= cos 0 = 1.]
33.
10
n=1
1
I
n
4
+ 1
=
1
I
2
+
1
I
17
+
1
I
82
+ +
1
I
10,001
E 1.24856. Now
1
I
n
4
+ 1
<
1
I
n
4
=
1
n
2
, so the error is
1
10
$ T
10
$
"
10
1
r
2
dr = lim
I<"
3
1
r
I
10
= lim
I<"
3
1
t
+
1
10
=
1
10
= 0.1.
35.
10
n=1
1
1 + 2
n
=
1
3
+
1
5
+
1
9
+ +
1
1025
E 0.76352. Now
1
1 + 2
n
<
1
2
n
, so the error is
1
10
$ T
10
=
"
n=11
1
2
n
=
12
11
1 312
[geometric series] E 0.00098.
37. Since
d
n
10
n
$
9
10
n
for each n, and since
"
n=1
9
10
n
|v| =
1
10
< 1
, 0.d
1
d
2
d
3
. . . =
"
n=1
d
n
10
n
will always converge by the Comparison Test.
39. Since
o
n
converges, lim
n<"
o
n
= 0, so there exists ` such that |o
n
30| < 1 for all n ` i 0 $ o
n
< 1 for
all n ` i 0 $ o
2
n
$ o
n
. Since
o
n
converges, so does
o
2
n
by the Comparison Test.
41. (a) Since lim
n<"
o
n
/
n
= ", there is an integer ` such that
o
n
/
n
1 whenever n `. (Take A = 1 in Denition 11.1.5.)
Then o
n
/
n
whenever n ` and since
/
n
is divergent,
o
n
is also divergent by the Comparison Test.
(b) (i) If o
n
=
1
lnn
and /
n
=
1
n
for n D 2, then lim
n<"
o
n
/
n
= lim
n<"
n
lnn
= lim
i<"
r
lnr
H
= lim
i<"
1
1r
= lim
i<"
r = ",
so by part (a),
"
n=2
1
lnn
is divergent.
(ii) If o
n
=
lnn
n
and /
n
=
1
n
, then
"
n=1
/
n
is the divergent harmonic series and lim
n<"
o
n
/
n
= lim
n<"
lnn = lim
i<"
lnr = ",
so
"
n=1
o
n
diverges by part (a).
43. lim
n<"
no
n
= lim
n<"
o
n
1n
, so we apply the Limit Comparison Test with /
n
=
1
n
. Since lim
n<"
no
n
0 we know that either both
series converge or both series diverge, and we also know that
"
n=1
1
n
diverges [j-series with j = 1]. Therefore,
o
n
must be
divergent.
SECTION 11.5 ALTERNATING SERIES

473
45. Yes. Since
o
n
is a convergent series with positive terms, lim
n<"
o
n
= 0 by Theorem 11.2.6, and
/
n
=

sin(o
n
) is a
series with positive terms (for large enough n). We have lim
n<"
/
n
o
n
= lim
n<"
sin(o
n
)
o
n
= 1 0 by Theorem 3.3.2. Thus,
/
n
is also convergent by the Limit Comparison Test.
11.5 Alternating Series
1. (a) An alternating series is a series whose terms are alternately positive and negative.
(b) An alternating series
"
n=1
(31)
n31
/
n
converges if 0 < /
n+1
$ /
n
for all n and lim
n<"
/
n
= 0. (This is the Alternating
Series Test.)
(c) The error involved in using the partial sumc
n
as an approximation to the total sumc is the remainder 1
n
= c 3c
n
and the
size of the error is smaller than /
n+1
; that is, |1
n
| $ /
n+1
. (This is the Alternating Series Estimation Theorem.)
3.
4
7
3
4
8
+
4
9
3
4
10
+
4
11
3 =
"
n=1
(31)
n31
4
n + 6
. Now /
n
=
4
n + 6
0, {/
n
} is decreasing, and lim
n<"
/
n
= 0, so the
series converges by the Alternating Series Test.
5.
"
n=1
o
n
=
"
n=1
(31)
n31
1
2n + 1
=
"
n=1
(31)
n31
/
n
. Now /
n
=
1
2n + 1
0, {/
n
n<"
/
n
= 0, so the
series converges by the Alternating Series Test.
7.
"
n=1
o
n
=
"
n=1
(31)
n
3n 31
2n + 1
=
"
n=1
(31)
n
/
n
. Now lim
n<"
/
n
= lim
n<"
3 31n
2 + 1n
=
3
2
6= 0. Since lim
n<"
o
n
6= 0
(in fact the limit does not exist), the series diverges by the Test for Divergence.
9. /
n
=
n
10
n
0 for n D 1. {/
n
} is decreasing for n D 1 since
r
10
i
0
=
10
i
(1) 3r 10
i
ln10
(10
i
)
2
=
10
i
(1 3rln10)
(10
i
)
2
=
1 3rln10
10
i
< 0 for 1 3rln10 < 0 i rln10 1 i
r
1
ln10
E 0.4. Also, lim
n<"
/
n
= lim
n<"
n
10
n
= lim
i<"
r
10
i
H
= lim
i<"
r
10
i
ln10
= 0. Thus, the series
"
n=1
(31)
n
n
10
n
converges by the Alternating Series Test.
11. /
n
=
n
2
n
3
+ 4
0 for n D 1. {/
n
} is decreasing for n D 2 since
r
2
r
3
+ 4
0
=
(r
3
+ 4)(2r) 3r
2
(3r
2
)
(r
3
+ 4)
2
=
r(2r
3
+ 8 33r
3
)
(r
3
+ 4)
2
=
r(8 3r
3
)
(r
3
+ 4)
2
< 0 for r 2. Also,
lim
n<"
/
n
= lim
n<"
1n
1 + 4n
3
= 0. Thus, the series
"
n=1
(31)
n+1
n
2
n
3
+ 4
13.
"
n=2
(31)
n
n
lnn
. lim
n<"
n
lnn
= lim
i<"
r
lnr
H
= lim
i<"
1
1r
= ", so the series diverges by the Test for Divergence.
474

15.
"
n=1
cos n
n
3'4
=
"
n=1
(31)
n
n
3'4
. /
n
=
1
n
3'4
is decreasing and positive and lim
n<"
1
n
3'4
= 0, so the series converges by the
Alternating Series Test.
17.
"
n=1
(31)
n
sin
. /
n
= sin
0 for n D 2 and sin
D sin

n + 1
, and lim
n<"
sin
= sin0 = 0, so the series

19.
n
n
n!
=
n n n
1 2 n
D n i lim
n<"
n
n
n!
= " i lim
n<"
(31)
n
n
n
n!
does not exist. So the series diverges by the Test for
Divergence.
21.
n o
n
c
n
1 1 1
2 30.35355 0.64645
3 0.19245 0.83890
4 30.125 0.71390
5 0.08944 0.80334
6 30.06804 0.73530
7 0.05399 0.78929
8 30.04419 0.74510
9 0.03704 0.78214
10 30.03162 0.75051
By the Alternating Series Estimation Theorem, the error in the approximation
"
n=1
(31)
n31
n
3'2
E 0.75051 is |c 3c
10
| $ /
11
= 1(11)
3'2
E 0.0275 (to four
decimal places, rounded up).
23. The series
"
n=1
(31)
n+1
n
6
satises (i) of the Alternating Series Test because
1
(n + 1)
6
<
1
n
6
and (ii) lim
n<"
1
n
6
= 0, so the
series is convergent. Now /
5
=
1
5
6
= 0.000064 0.00005 and /
6
=
1
6
6
E 0.00002 < 0.00005, so by the Alternating Series
Estimation Theorem, n = 5. (That is, since the 6th term is less than the desired error, we need to add the rst 5 terms to get the
sum to the desired accuracy.)
25. The series
"
n=0
(31)
n
10
n
n!
satises (i) of the Alternating Series Test because
1
10
n+1
(n + 1)!
<
1
10
n
n!
and (ii) lim
n<"
1
10
n
n!
= 0,
so the series is convergent. Now /
3
=
1
10
3
3!
E 0.000 167 0.000 005 and /
4
=
1
10
4
4!
= 0.000 004 < 0.000 005, so by
the Alternating Series Estimation Theorem, n = 4 (since the series starts with n = 0, not n = 1). (That is, since the 5th term
is less than the desired error, we need to add the rst 4 terms to get the sum to the desired accuracy.)
27. /
7
=
1
7
5
=
1
16,807
E 0.000 059 5, so
"
n=1
(31)
n+1
n
5
E c
6
=
6
n=1
(31)
n+1
n
5
= 1 3
1
32
+
1
243
3
1
1024
+
1
3125
3
1
7776
E 0.972 080. Adding /
7
to c
6
does not change
the fourth decimal place of c
6
, so the sum of the series, correct to four decimal places, is 0.9721.
SECTION 11.6 ABSOLUTE CONVERGENCE AND THE RATIO AND ROOT TESTS

475
29. /
7
=
7
2
10
7
= 0.000 004 9, so
"
n=1
(31)
n31
n
2
10
n
E c
6
=
6
n=1
(31)
n31
n
2
10
n
=
1
10
3
4
100
+
9
1000
3
16
10,000
+
25
100,000
3
36
1,000,000
= 0.067 614. Adding /
7
to c
6
does not change the fourth decimal place of c
6
, so the sum of the series, correct to four decimal places, is 0.0676.
31.
"
n=1
(31)
n31
n
= 1 3
1
2
+
1
3
3
1
4
+ +
1
49
3
1
50
+
1
51
3
1
52
+ . The 50th partial sum of this series is an
underestimate, since
"
n=1
(31)
n31
n
= c
50
+
1
51
3
1
52
1
53
3
1
54
+ , and the terms in parentheses are all positive.

The result can be seen geometrically in Figure 1.
33. Clearly /
n
=
1
n +j
is decreasing and eventually positive and lim
n<"
/
n
= 0 for any j. So the series converges (by the
Alternating Series Test) for any j for which every /
n
is dened, that is, n +j 6= 0 for n D 1, or j is not a negative integer.
35.

/
2n
=

1(2n)
2
clearly converges (by comparison with the j-series for j = 2). So suppose that
(31)
n31
/
n
converges. Then by Theorem 11.2.8(ii), so does
(31)
n31
/
n
+/
n
= 2
1 +
1
3
+
1
5
+
= 2
1
2n 31
. But this
diverges by comparison with the harmonic series, a contradiction. Therefore,
(31)
n31
/
n
must diverge. The Alternating
Series Test does not apply since {/
n
} is not decreasing.
11.6 Absolute Convergence and the Ratio and Root Tests
1. (a) Since lim
n<"
o
n+1
o
n
= 8 1, part (b) of the Ratio Test tells us that the series
o
n
is divergent.
(b) Since lim
n<"
o
n+1
o
n
= 0.8 < 1, part (a) of the Ratio Test tells us that the series
o
n
is absolutely convergent (and
therefore convergent).
(c) Since lim
n<"
o
n+1
o
n
= 1, the Ratio Test fails and the series
o
n
might converge or it might diverge.
3.
"
n=0
(310)
n
n!
. Using the Ratio Test, lim
n<"
o
n+1
o
n
= lim
n<"
(310)
n+1
(n + 1)!

n!
(310)
n
= lim
n<"
310
n + 1
= 0 < 1, so the series is

absolutely convergent.
5.
"
n=1
(31)
n+1
4
I
n
converges by the Alternating Series Test, but
"
n=1
1
4
I
n
is a divergent j-series
j =
1
4
$ 1
, so the given series

is conditionally convergent.
476

7. lim
I<"
o
I+1
o
I
= lim
I<"
(I + 1)
2
3
I+1
I
2
3
= lim
I<"
I + 1
I
2
3
1
=
2
3
lim
I<"
1 +
1
I
=
2
3
(1) =
2
3
< 1, so the series
"
n=1
I
2
3
I
is absolutely convergent by the Ratio Test. Since the terms of this series are positive, absolute convergence is the
same as convergence.
9. lim
n<"
o
n+1
o
n
= lim
n<"
(1.1)
n+1
(n + 1)
4

n
4
(1.1)
n
= lim
n<"
(1.1)n
4
(n + 1)
4
= (1.1) lim
n<"
1
(n + 1)
4
n
4
= (1.1) lim
n<"
1
(1 + 1n)
4
= (1.1)(1) = 1.1 1,
so the series
"
n=1
(31)
n
(1.1)
n
n
4
diverges by the Ratio Test.
11. Since 0 $
c
1'n
n
3
$
c
n
3
= c
1
n
3
and
"
n=1
1
n
3
is a convergent j-series [j = 3 1],
"
n=1
c
1'n
n
3
converges, and so
"
n=1
(31)
n
c
1'n
n
3
is absolutely convergent.
13. lim
n<"
o
n+1
o
n
= lim
n<"
10
n+1
(n + 2) 4
2n+3

(n + 1) 4
2n+1
10
n
= lim
n<"
10
4
2

n + 1
n + 2
=
5
8
< 1, so the series
"
n=1
10
n
(n + 1)4
2n+1
is absolutely convergent by the Ratio Test. Since the terms of this series are positive, absolute convergence is the same as
convergence.
15.

(31)
n
arctann
n
2
<
2
n
2
, so since
"
n=1
2
n
2
=

2
"
n=1
1
n
2
converges (j = 2 1), the given series
"
n=1
(31)
n
arctann
n
2
converges absolutely by the Comparison Test.
17.
"
n=2
(31)
n
lnn
converges by the Alternating Series Test since lim
n<"
1
lnn
= 0 and
1
lnn
is decreasing. Now lnn < n, so

1
lnn

1
n
, and since
"
n=2
1
n
is the divergent (partial) harmonic series,
"
n=2
1
lnn
diverges by the Comparison Test. Thus,
"
n=2
(31)
n
lnn
19.
|cos (n3)|
n!
$
1
n!
and
"
n=1
1
n!
converges (use the Ratio Test), so the series
"
n=1
cos(n3)
n!
converges absolutely by the
Comparison Test.
21. lim
n<"
q
|o
n
| = lim
n<"
n
2
+ 1
2n
2
+ 1
= lim
n<"
1 + 1n
2
2 + 1n
2
=
1
2
< 1, so the series
"
n=1
n
2
+ 1
2n
2
+ 1
n
is absolutely convergent by the
Root Test.
23. lim
n<"
q
|o
n
| = lim
n<"
q
1 +
1
n
n
2
= lim
n<"
1 +
1
n
n
= c 1 (by Equation 3.6.6), so the series
"
n=1
1 +
1
n
n
2
diverges by the Root Test.
SECTION 11.6 ABSOLUTE CONVERGENCE AND THE RATIO AND ROOT TESTS

477
25. Use the Ratio Test with the series
1 3
1 3
3!
+
1 3 5
5!
3
1 3 5 7
7!
+ + (31)
n31
1 3 5 (2n 31)
(2n 31)!
+ =
"
n=1
(31)
n31
1 3 5 (2n 31)
(2n 31)!
.
lim
n<"
o
n+1
o
n
= lim
n<"
(31)
n
1 3 5 (2n 31)[2(n + 1) 31]
[2(n + 1) 31]!

(2n 31)!
(31)
n31
1 3 5 (2n 31)
= lim
n<"
(31)(2n + 1)(2n 31)!

(2n + 1)(2n)(2n 31)!
= lim
n<"
1
2n
= 0 < 1,
so the given series is absolutely convergent and therefore convergent.
27.
"
n=1
2 4 6 (2n)
n!
=
"
n=1
(2 1) (2 2) (2 3) (2 n)
n!
=
"
n=1
2
n
n!
n!
=
"
n=1
2
n
, which diverges by the Test for
Divergence since lim
n<"
2
n
= ".
29. By the recursive denition, lim
n<"
o
n+1
o
n
= lim
n<"
5n + 1
4n + 3
=
5
4
1, so the series diverges by the Ratio Test.
31. (a) lim
n<"
1(n + 1)
3
1n
3
= lim
n<"
n
3
(n + 1)
3
= lim
n<"
1
(1 + 1n)
3
= 1. Inconclusive
(b) lim
n<"
(n + 1)
2
n+1

2
n
n
= lim
n<"
n + 1
2n
= lim
n<"
1
2
+
1
2n
=
1
2
. Conclusive (convergent)
(c) lim
n<"
(33)
n
I
n + 1

I
n
(33)
n31
= 3 lim
n<"
n
n + 1
= 3 lim
n<"
1
1 + 1n
= 3. Conclusive (divergent)
(d) lim
n<"
I
n + 1
1 + (n + 1)
2

1 +n
2
I
n
= lim
n<"
1 +
1
n

1n
2
+ 1
1n
2
+ (1 + 1n)
2
= 1. Inconclusive
33. (a) lim
n<"
o
n+1
o
n
= lim
n<"
r
n+1
(n + 1)!

n!
r
n
= lim
n<"
r
n + 1
= |r| lim
n<"
1
n + 1
= |r| 0 = 0 < 1, so by the Ratio Test the
series
"
n=0
r
n
n!
converges for all r.
(b) Since the series of part (a) always converges, we must have lim
n<"
r
n
n!
= 0 by Theorem 11.2.6.
35. (a) c
5
=
5
n=1
1
n2
n
=
1
2
+
1
8
+
1
24
+
1
64
+
1
160
=
661
960
E 0.68854. Now the ratios
v
n
=
o
n+1
o
n
=
n2
n
(n + 1)2
n+1
=
n
2(n + 1)
form an increasing sequence, since
v
n+1
3v
n
=
n + 1
2(n + 2)
3
n
2(n + 1)
=
(n + 1)
2
3n(n + 2)
2(n + 1)(n + 2)
=
1
2(n + 1)(n + 2)
0. So by Exercise 34(b), the error
in using c
5
is 1
5
$
o
6
1 3 lim
n<"
v
n
=
1
6 2
6
1 312
=
1
192
E 0.00521.
(b) The error in using c
n
as an approximation to the sum is 1
n
=
o
n+1
1 3
1
2
=
2
(n + 1)2
n+1
. We want 1
n
< 0.00005 C
1
(n + 1)2
n
< 0.00005 C (n + 1)2
n
20,000. To nd such an n we can use trial and error or a graph. We calculate
(11 + 1)2
11
= 24,576, so c
11
=
11
n=1
1
n2
n
E 0.693109 is within 0.00005 of the actual sum.
478

37. (i) Following the hint, we get that |o
n
| < v
n
for n D `, and so since the geometric series
"
n=1
v
n
converges [0 < v < 1],
the series
"
n=^
|o
n
| converges as well by the Comparison Test, and hence so does
"
n=1
|o
n
|, so
"
n=1
o
n
is absolutely
convergent.
(ii) If lim
n<"
q
|o
n
| = 1 1, then there is an integer ` such that
q
|o
n
| 1 for all n D `, so |o
n
| 1 for n D `. Thus,
lim
n<"
o
n
6= 0, so
"
n=1
o
n
diverges by the Test for Divergence.
(iii) Consider
"
n=1
1
n
[diverges] and
"
n=1
1
n
2
[converges]. For each sum, lim
n<"
q
|o
n
| = 1, so the Root Test is inconclusive.
39. (a) Since
o
n
is absolutely convergent, and since
o
+
n
$ |o
n
| and
o
3
n
$ |o
n
| (because o
+
n
and o
3
n
each equal
either o
n
or 0), we conclude by the Comparison Test that both
o
+
n
and
o
3
n
must be absolutely convergent.
Or: Use Theorem 11.2.8.
(b) We will show by contradiction that both
o
+
n
and
o
3
n
must diverge. For suppose that
o
+
n
converged. Then so
would
o
+
n
3
1
2
o
n
by Theorem 11.2.8. But
o
+
n
3
1
2
o
n
=

1
2
(o
n
+ |o
n
|) 3
1
2
o
n
=
1
2
|o
n
|, which
diverges because
o
n
is only conditionally convergent. Hence,
o
+
n
cant converge. Similarly, neither can
o
3
n
.
11.7 Strategy for Testing Series
1.
1
n + 3
n
<
1
3
n
=

1
3
n
for all n D 1.
"
n=1
1
3
n
|v| =
1
3
< 1
, so
"
n=1
1
n + 3
n
3. lim
n<"
|o
n
| = lim
n<"
n
n + 2
= 1, so lim
n<"
o
n
= lim
n<"
(31)
n
n
n + 2
does not exist. Thus, the series
"
n=1
(31)
n
n
n + 2
diverges by
the Test for Divergence.
5. lim
n<"
o
n+1
o
n
= lim
n<"
(n + 1)
2
2
n
(35)
n+1

(35)
n
n
2
2
n31
= lim
n<"
2(n + 1)
2
5n
2
=
2
5
lim
n<"
1 +
1
n
2
=
2
5
(1) =
2
5
< 1, so the series
"
n=1
n
2
2
n31
(35)
n
converges by the Ratio Test.
7. Let 1(r) =
1
r
I
lnr
. Then 1 is positive, continuous, and decreasing on [2. "), so we can apply the Integral Test.
Since
1
r
I
lnr
dr
& = lnr,
d& = drr
=

&
31'2
d& = 2&
1'2
+ C = 2
I
lnr +C, we nd
"
2
dr
r
I
lnr
= lim
I<"
I
2
dr
r
I
lnr
= lim
I<"
2
I
lnr
I
2
= lim
I<"
2
I
lnt 32
I
ln2
= ". Since the integral diverges, the

given series
"
n=2
1
n
I
lnn
diverges.
SECTION 11.7 STRATEGY FOR TESTING SERIES

479
9.
"
I=1
I
2
c
3I
=
"
I=1
I
2
c
I
. Using the Ratio Test, we get
lim
I<"
o
I+1
o
I
= lim
I<"
(I + 1)
2
c
I+1

c
I
I
2
= lim
I<"
I + 1
I
1
c
= 1
2
1
c
=
1
c
< 1, so the series converges.
11. /
n
=
1
nlnn
0 for n D 2, {/
n
n<"
/
n
= 0, so the given series
"
n=2
(31)
n+1
nlnn
converges by the
13. lim
n<"
o
n+1
o
n
= lim
n<"
3
n+1
(n + 1)
2
(n + 1)!

n!
3
n
n
2
= lim
n<"
3(n + 1)
2
(n + 1)n
2
= 3 lim
n<"
n + 1
n
2
= 0 < 1, so the series
"
n=1
3
n
n
2
n!
15. lim
n<"
o
n+1
o
n
= lim
n<"
(n + 1)!
2 5 8 (3n + 2)[3(n + 1) + 2]

2 5 8 (3n + 2)
n!
= lim
n<"
n + 1
3n + 5
=
1
3
< 1,
so the series
"
n=0
n!
2 5 8 (3n + 2)
17. lim
n<"
2
1'n
= 2
0
= 1, so lim
n<"
(31)
n
2
1'n
does not exist and the series
"
n=1
(31)
n
2
1'n
19. Let 1(r) =
lnr
I
r
. Then 1
0
(r) =
2 3lnr
2r
3'2
< 0 when lnr 2 or r c
2
, so
lnn
I
n
is decreasing for n c
2
.
By lHospitals Rule, lim
n<"
lnn
I
n
= lim
n<"
1n
1
2
I
n
= lim
n<"
2
I
n
= 0, so the series
"
n=1
(31)
n
lnn
I
n
converges by the
21.
"
n=1
(32)
2n
n
n
=
"
n=1
4
n
n
. lim
n<"
q
|o
n
| = lim
n<"
4
n
= 0 < 1, so the given series is absolutely convergent by the Root Test.
23. Using the Limit Comparison Test with o
n
= tan
1
n
and /
n
=
1
n
, we have
lim
n<"
o
n
/
n
= lim
n<"
tan(1n)
1n
= lim
i<"
tan(1r)
1r
H
= lim
i<"
sec
2
(1r) (31r
2
)
31r
2
= lim
i<"
sec
2
(1r) = 1
2
= 1 0. Since
"
n=1
/
n
is the divergent harmonic series,
"
n=1
o
n
is also divergent.
25. Use the Ratio Test. lim
n<"
o
n+1
o
n
= lim
n<"
(n + 1)!
c
(n+1)
2

c
n
2
n!
= lim
n<"
(n + 1)n! c
n
2
c
n
2
+2n+1
n!
= lim
n<"
n + 1
c
2n+1
= 0 < 1, so
"
n=1
n!
c
n
2
converges.
27.

"
2
lnr
r
2
dr = lim
I<"
3
lnr
r
3
1
r
I
1
[using integration by parts]
H
= 1. So
"
n=1
lnn
n
2
converges by the Integral Test, and since
I lnI
(I + 1)
3
<
I lnI
I
3
=
lnI
I
2
, the given series
"
I=1
I lnI
(I + 1)
3
480

29.
"
n=1
o
n
=
"
n=1
(31)
n
1
coshn
=
"
n=1
(31)
n
/
n
. Now /
n
=
1
coshn
0, {/
n
n<"
/
n
= 0, so the series
Or: Write
1
coshn
=
2
c
n
+c
3n
<
2
c
n
and
"
n=1
1
c
n
is a convergent geometric series, so
"
n=1
1
coshn
is convergent by the
Comparison Test. So
"
n=1
(31)
n
1
coshn
is absolutely convergent and therefore convergent.
31. lim
I<"
o
I
= lim
I<"
5
I
3
I
+ 4
I
= [divide by 4
I
] lim
I<"
(54)
I
(34)
I
+ 1
= "since lim
I<"
3
4
I
= 0 and lim
I<"
5
4
I
= ".
Thus,
"
I=1
5
I
3
I
+ 4
I
33. Let o
n
=
sin(1n)
I
n
and /
n
=
1
n
I
n
. Then lim
n<"
o
n
/
n
= lim
n<"
sin(1n)
1n
= 1 0, so
"
n=1
sin(1n)
I
n
converges by limit
comparison with the convergent j-series
"
n=1
1
n
3'2
[j = 32 1].
35. lim
n<"
q
|o
n
| = lim
n<"
n
n + 1
n
2
'n
= lim
n<"
1
[(n + 1) n]
n
=
1
lim
n<"
(1 + 1n)
n
=
1
c
< 1, so the series
"
n=1
n
n + 1
n
2
converges by the Root Test.
37. lim
n<"
q
|o
n
| = lim
n<"
(2
1'n
31) = 1 31 = 0 < 1, so the series
"
n=1
q
I
2 31
n
converges by the Root Test.
11.8 Power Series
1. A power series is a series of the form
"
n=0
c
n
r
n
= c
0
+c
1
r +c
2
r
2
+c
3
r
3
+ , where r is a variable and the c
n
s are
constants called the coefcients of the series.
More generally, a series of the form
"
n=0
c
n
(r 3o)
n
= c
0
+c
1
(r 3o) +c
2
(r 3o)
2
+ is called a power series in
(r 3o) or a power series centered at o or a power series about o, where o is a constant.
3. If o
n
=
r
n
I
n
, then lim
n<"
o
n+1
o
n
= lim
n<"
r
n+1
I
n + 1

I
n
r
n
= lim
n<"
r
I
n + 1
I
n
= lim
n<"
|r|
1 + 1n
= |r|.
By the Ratio Test, the series
"
n=1
r
n
I
n
converges when |r| < 1, so the radius of convergence 1 = 1. Now well check the
endpoints, that is, r = 1. When r = 1, the series
"
n=1
1
I
n
diverges because it is a j-series with j =
1
2
$ 1. When r = 31,
the series
"
n=1
(31)
n
I
n
converges by the Alternating Series Test. Thus, the interval of convergence is 1 = [31. 1).
5. If o
n
=
(31)
n31
r
n
n
3
, then
lim
n<"
o
n+1
o
n
= lim
n<"
(31)
n
r
n+1
(n + 1)
3

n
3
(31)
n31
r
n
= lim
n<"
(31)rn
3
(n + 1)
3
= lim
n<"
n
n + 1
3
|r|
= 1
3
|r| = |r|. By the
SECTION 11.8 POWER SERIES

481
Ratio Test, the series
"
n=1
(31)
n31
r
n
n
3
converges when |r| < 1, so the radius of convergence 1 = 1. Now well check the
endpoints, that is, r = 1. When r = 1, the series
"
n=1
(31)
n31
n
3
converges by the Alternating Series Test. When r = 31,
the series
"
n=1
(31)
n31
(31)
n
n
3
= 3
"
n=1
1
n
3
converges because it is a constant multiple of a convergent j-series [j = 3 1].
Thus, the interval of convergence is 1 = [31. 1].
7. If o
n
=
r
n
n!
, then lim
n<"
o
n+1
o
n
= lim
n<"
r
n+1
(n + 1)!

n!
r
n
= lim
n<"
r
n + 1
= |r| lim
n<"
1
n + 1
= |r| 0 = 0 < 1 for all real r.
So, by the Ratio Test, 1 = "and 1 = (3". ").
9. If o
n
= (31)
n
n
2
r
n
2
n
, then
lim
n<"
o
n+1
o
n
= lim
n<"
(n + 1)
2
r
n+1
2
n+1

2
n
n
2
r
n
= lim
n<"
r(n + 1)
2
2n
2
= lim
n<"
|r|
2
1 +
1
n
=
|r|
2
(1)
2
=
1
2
|r|. By the
"
n=1
(31)
n
n
2
r
n
2
n
converges when
1
2
|r| < 1 C |r| < 2, so the radius of convergence is 1 = 2.
When r = 2, both series
"
n=1
(31)
n
n
2
(2)
n
2
n
=
"
n=1
(~1)
n
n
2
diverge by the Test for Divergence since
lim
n<"
(~1)
n
n
2
= ". Thus, the interval of convergence is 1 = (32. 2).

11. o
n
=
(32)
n
r
n
4
I
n
, so lim
n<"
o
n+1
o
n
= lim
n<"
2
n+1
|r|
n+1
4
I
n + 1

4
I
n
2
n
|r|
n
= lim
n<"
2 |r|
4
n
n + 1
= 2 |r|, so by the Ratio Test, the
series converges when 2 |r| < 1 C |r| <
1
2
, so 1 =
1
2
. When r = 3
1
2
, we get the divergent j-series
"
n=1
1
4
I
n
j =
1
4
$ 1
. When r =
1
2
, we get the series
"
n=1
(31)
n
4
I
n
, which converges by the Alternating Series Test.
Thus, 1 =

3
1
2
.
1
2
.
13. If o
n
= (31)
n
r
n
4
n
lnn
, then lim
n<"
o
n+1
o
n
= lim
n<"
r
n+1
4
n+1
ln(n + 1)

4
n
lnn
r
n
=
|r|
4
lim
n<"
lnn
ln(n + 1)
=
|r|
4
1
[by lHospitals Rule] =
|r|
4
. By the Ratio Test, the series converges when
|r|
4
< 1 C |r| < 4, so 1 = 4. When
r = 34,
"
n=2
(31)
n
r
n
4
n
lnn
=
"
n=2
[(31)(34)]
n
4
n
lnn
=
"
n=2
1
lnn
. Since lnn < n for n D 2,
1
lnn

1
n
and
"
n=2
1
n
is the
divergent harmonic series (without the n = 1 term),
"
n=2
1
lnn
is divergent by the Comparison Test. When r = 4,
"
n=2
(31)
n
r
n
4
n
lnn
=
"
n=2
(31)
n
1
lnn
, which converges by the Alternating Series Test. Thus, 1 = (34. 4].
482

15. If o
n
=
(r 32)
n
n
2
+ 1
, then lim
n<"
o
n+1
o
n
= lim
n<"
(r 32)
n+1
(n + 1)
2
+ 1

n
2
+ 1
(r 32)
n
= |r 32| lim
n<"
n
2
+ 1
(n + 1)
2
+ 1
= |r 32|. By the
"
n=0
(r 32)
n
n
2
+ 1
converges when |r 32| < 1 [1 = 1] C 31 < r 32 < 1 C 1 < r < 3. When
r = 1, the series
"
n=0
(31)
n
1
n
2
+ 1
converges by the Alternating Series Test; when r = 3, the series
"
n=0
1
n
2
+ 1
converges by
comparison with the p-series
"
n=1
1
n
2
[j = 2 1]. Thus, the interval of convergence is 1 = [1. 3].
17. If o
n
=
3
n
(r + 4)
n
I
n
, then lim
n<"
o
n+1
o
n
= lim
n<"
3
n+1
(r + 4)
n+1
I
n + 1

I
n
3
n
(r + 4)
n
= 3 |r + 4| lim
n<"
I
n
I
n + 1
= 3 |r + 4|.
By the Ratio Test, the series
"
n=1
3
n
(r + 4)
n
I
n
converges when 3 |r + 4| < 1 C |r + 4| <
1
3
1 =
1
3
C
3
1
3
< r + 4 <
1
3
C 3
13
3
< r < 3
11
3
. When r = 3
13
3
, the series
"
n=1
(31)
n
1
I
n
converges by the Alternating Series
Test; when r = 3
11
3
, the series
"
n=1
1
I
n
diverges
j =
1
2
$ 1
. Thus, the interval of convergence is 1 =

3
13
3
. 3
11
3
.
19. If o
n
=
(r 32)
n
n
n
, then lim
n<"
q
|o
n
| = lim
n<"
|r 32|
n
= 0, so the series converges for all r (by the Root Test).
1 = "and 1 = (3". ").
21. o
n
=
n
/
n
(r 3o)
n
, where / 0.
lim
n<"
o
n+1
o
n
= lim
n<"
(n + 1) |r 3o|
n+1
/
n+1

/
n
n|r 3o|
n
= lim
n<"
1 +
1
n
|r 3o|
/
=
|r 3o|
/
.
By the Ratio Test, the series converges when
|r 3o|
/
< 1 C |r 3o| < / [so 1 = /] C 3/ < r 3o < / C
o 3/ < r < o +/. When |r 3o| = /, lim
n<"
|o
n
| = lim
n<"
n = ", so the series diverges. Thus, 1 = (o 3/. o +/).
23. If o
n
= n! (2r 31)
n
, then lim
n<"
o
n+1
o
n
= lim
n<"
(n + 1)! (2r 31)

n+1
n!(2r 31)
n
= lim
n<"
(n + 1) |2r 31| <"as n <"
for all r 6=
1
2
. Since the series diverges for all r 6=
1
2
, 1 = 0 and 1 =

1
2
.
25. lim
n<"
o
n+1
o
n
= lim
n<"
|4r + 1|
n+1
(n + 1)
2

n
2
|4r + 1|
n
= lim
n<"
|4r + 1|
(1 + 1n)
2
= |4r + 1|, so by the Ratio Test, the series
converges when |4r + 1| < 1 C 31 < 4r +1 < 1 C 32 < 4r < 0 C 3
1
2
< r < 0, so 1 =
1
4
. When r = 3
1
2
,
the series becomes
"
n=1
(31)
n
n
2
, which converges by the Alternating Series Test. When r = 0, the series becomes
"
n=1
1
n
2
,
a convergent j-series [j = 2 1]. 1 =

3
1
2
. 0
.
SECTION 11.8 POWER SERIES

483
27. If o
n
=
r
n
1 3 5 (2n 31)
, then
lim
n<"
o
n+1
o
n
= lim
n<"
r
n+1
1 3 5 (2n 31)(2n + 1)

1 3 5 (2n 31)
r
n
= lim
n<"
|r|
2n + 1
= 0 < 1. Thus, by the
"
n=1
r
n
1 3 5 (2n 31)
converges for all real r and we have 1 = "and 1 = (3". ").
29. (a) We are given that the power series
"
n=0
c
n
r
n
is convergent for r = 4. So by Theorem 3, it must converge for at least
34 < r $ 4. In particular, it converges when r = 32; that is,
"
n=0
c
n
(32)
n
is convergent.
(b) It does not follow that
"
n=0
c
n
(34)
n
is necessarily convergent. [See the comments after Theorem 3 about convergence at
the endpoint of an interval. An example is c
n
= (31)
n
(n4
n
).]
31. If o
n
=
(n!)
I
(In)!
r
n
, then
lim
n<"
o
n+1
o
n
= lim
n<"
[(n + 1)!]
I
(In)!
(n!)
I
[I(n + 1)]!
|r| = lim
n<"
(n + 1)
I
(In +I)(In +I 31) (In + 2)(In + 1)
|r|
= lim
n<"
(n + 1)
(In + 1)
(n + 1)
(In + 2)

(n + 1)
(In +I)
|r|
= lim
n<"
n + 1
In + 1
lim
n<"
n + 1
In + 2
lim
n<"
n + 1
In +I
|r|
=

1
I
I
|r| < 1 C |r| < I
I
for convergence, and the radius of convergence is 1 = I
I
.
33. No. If a power series is centered at o, its interval of convergence is symmetric about o. If a power series has an innite radius
of convergence, then its interval of convergence must be (3". "), not [0. ").
35. (a) If o
n
=
(31)
n
r
2n+1
n!(n + 1)! 2
2n+1
, then
lim
n<"
o
n+1
o
n
= lim
n<"
r
2n+3
(n + 1)!(n + 2)! 2
2n+3

n!(n + 1)! 2
2n+1
r
2n+1
=

r
2
2
lim
n<"
1
(n + 1)(n + 2)
= 0 for all r.
So J
1
(r) converges for all r and its domain is (3". ").
(b), (c) The initial terms of J
1
(r) up to n = 5 are o
0
=
r
2
,
o
1
= 3
r
3
16
, o
2
=
r
5
384
, o
3
= 3
r
7
18,432
, o
4
=
r
9
1,474,560
,
and o
5
= 3
r
11
176,947,200
. The partial sums seem to
approximate J
1
(r) well near the origin, but as |r| increases,
we need to take a large number of terms to get a good
approximation.
484

37. c
2n31
= 1 + 2r +r
2
+ 2r
3
+r
4
+ 2r
5
+ +r
2n32
+ 2r
2n31
= 1(1 + 2r) +r
2
(1 + 2r) +r
4
(1 + 2r) + +r
2n32
(1 + 2r) = (1 + 2r)(1 +r
2
+r
4
+ +r
2n32
)
= (1 + 2r)
1 3r
2n
1 3r
2
[by (11.2.3)] with v = r
2
] <
1 + 2r
1 3r
2
as n <" [by (11.2.4)], when |r| < 1.
Also c
2n
= c
2n31
+r
2n
<
1 + 2r
1 3r
2
since r
2n
<0 for |r| < 1. Therefore, c
n
<
1 + 2r
1 3r
2
since c
2n
and c
2n31
both
approach
1 + 2r
1 3r
2
as n <". Thus, the interval of convergence is (31. 1) and 1(r) =
1 + 2r
1 3r
2
.
39. We use the Root Test on the series
c
n
r
n
. We need lim
n<"
q
|c
n
r
n
| = |r| lim
n<"
q
|c
n
| = c |r| < 1 for convergence, or
|r| < 1c, so 1 = 1c.
41. For 2 < r < 3,
c
n
r
n
diverges and
d
n
r
n
converges. By Exercise 11.2.69,
(c
n
+d
n
) r
n
diverges. Since both series
converge for |r| < 2, the radius of convergence of
(c
n
+d
n
) r
n
is 2.
11.9 Representations of Functions as Power Series
1. If 1(r) =
"
n=0
c
n
r
n
has radius of convergence 10, then 1
0
(r) =
"
n=1
nc
n
r
n31
also has radius of convergence 10 by
Theorem 2.
3. Our goal is to write the function in the form
1
1 3v
, and then use Equation (1) to represent the function as a sum of a power
series. 1(r) =
1
1 +r
=
1
1 3(3r)
=
"
n=0
(3r)
n
=
"
n=0
(31)
n
r
n
with |3r| < 1 C |r| < 1, so 1 = 1 and 1 = (31. 1).
5. 1(r) =
2
3 3r
=
2
3
1
1 3r3
=
2
3
"
n=0
r
3
n
or, equivalently, 2
"
n=0
1
3
n+1
r
n
. The series converges when
r
3
< 1,
that is, when |r| < 3, so 1 = 3 and 1 = (33. 3).
7. 1(r) =
r
9 +r
2
=
r
9
1
1 + (r3)
2
=
r
9
1
1 3{3(r3)
2
}
=
r
9
"
n=0
r
3
n
=
r
9
"
n=0
(31)
n
r
2n
9
n
=
"
n=0
(31)
n
r
2n+1
9
n+1
The geometric series
"
n=0
r
3
n
converges when
r
3
< 1 C
r
2
9
< 1 C |r|
2
< 9 C |r| < 3, so
1 = 3 and 1 = (33. 3).
9. 1(r) =
1 +r
1 3r
= (1 +r)
1
1 3r
= (1 +r)
"
n=0
r
n
=
"
n=0
r
n
+
"
n=0
r
n+1
= 1 +
"
n=1
r
n
+
"
n=1
r
n
= 1 + 2
"
n=1
r
n
.
The series converges when |r| < 1, so 1 = 1 and 1 = (31. 1).
A second approach: 1(r) =
1 +r
1 3r
=
3(1 3r) + 2
1 3r
= 31 + 2
1
1 3r
= 31 + 2
"
n=0
r
n
= 1 + 2
"
n=1
r
n
.
A third approach:
1(r) =
1 +r
1 3r
= (1 +r)
1
1 3r
= (1 +r)(1 +r +r
2
+r
3
+ )
= (1 +r +r
2
+r
3
+ ) + (r +r
2
+r
3
+r
4
+ ) = 1 + 2r + 2r
2
+ 2r
3
+ = 1 + 2
"
n=1
r
n
.
SECTION 11.9 REPRESENTATIONS OF FUNCTIONS AS POWER SERIES

485
11. 1(r) =
3
r
2
3r 32
=
3
(r 32)(r + 1)
=

r 32
+
1
r + 1
i 3 = (r + 1) +1(r 32). Let r = 2 to get = 1 and
r = 31 to get 1 = 31. Thus
3
r
2
3r 32
=
1
r 32
3
1
r + 1
=
1
32
1
1 3(r2)
3
1
1 3(3r)
= 3
1
2
"
n=0
r
2
n
3
"
n=0
(3r)
n
=
"
n=0
3
1
2
1
2
n
31(31)
n
r
n
=
"
n=0
(31)
n+1
3
1
2
n+1
r
n
We represented 1 as the sum of two geometric series; the rst converges for r M (32. 2) and the second converges for (31. 1).
Thus, the sum converges for r M (31. 1) = 1.
13. (a) 1(r) =
1
(1 +r)
2
=
d
dr
31
1 +r
= 3
d
dr
"
n=0
(31)
n
r
n
[from Exercise 3]
=
"
n=1
(31)
n+1
nr
n31
[from Theorem 2(i)] =
"
n=0
(31)
n
(n + 1)r
n
with 1 = 1.
In the last step, note that we decreased the initial value of the summation variable n by 1, and then increased each
occurrence of n in the term by 1 [also note that (31)
n+2
= (31)
n
].
(b) 1(r) =
1
(1 +r)
3
= 3
1
2
d
dr
1
(1 +r)
2
= 3
1
2
d
dr
"
n=0
(31)
n
(n + 1)r
n
[from part (a)]

= 3
1
2
"
n=1
(31)
n
(n + 1)nr
n31
=
1
2
"
n=0
(31)
n
(n + 2)(n + 1)r
n
with 1 = 1.
(c) 1(r) =
r
2
(1 +r)
3
= r
2
1
(1 +r)
3
= r
2
1
2
"
n=0
(31)
n
(n + 2)(n + 1)r
n
[from part (b)]
=
1
2
"
n=0
(31)
n
(n + 2)(n + 1)r
n+2
To write the power series with r
n
rather than r
n+2
, we will decrease each occurrence of n in the term by 2 and increase
the initial value of the summation variable by 2. This gives us
1
2
"
n=2
(31)
n
(n)(n 31)r
n
with 1 = 1.
15. 1(r) = ln(5 3r) = 3
dr
5 3r
= 3
1
5
dr
1 3r5
= 3
1
5

"
n=0
r
5
dr = C 3
1
5
"
n=0
r
n+1
5
n
(n + 1)
= C 3
"
n=1
r
n
n5
n
Putting r = 0, we get C = ln5. The series converges for |r5| < 1 C |r| < 5, so 1 = 5.
17.
1
2 3r
=
1
2(1 3r2)
=
1
2
"
n=0
r
2
n
=
"
n=0
1
2
n+1
r
n
for
r
2
< 1 C |r| < 2. Now

1
(r 32)
2
=
d
dr
1
2 3r
=
d
dr
"
n=0
1
2
n+1
r
n
=
"
n=1
n
2
n+1
r
n31
=
"
n=0
n + 1
2
n+2
r
n
. So
1(r) =
r
3
(r 32)
2
= r
3
"
n=0
n + 1
2
n+2
r
n
=
"
n=0
n + 1
2
n+2
r
n+3
or
"
n=3
n 32
2
n31
r
n
for |r| < 2. Thus, 1 = 2 and 1 = (32. 2).
486

19. 1(r) =
r
r
2
+ 16
=
r
16
1
1 3(3r
2
16)
=
r
16
"
n=0
3
r
2
16
n
=
r
16
"
n=0
(31)
n
1
16
n
r
2n
=
"
n=0
(31)
n
1
16
n+1
r
2n+1
.
The series converges when
3r
2
16
< 1 C r
2
< 16 C |r| < 4, so 1 = 4. The partial sums are c
1
=
r
16
,
c
2
= c
1
3
r
3
16
2
, c
3
= c
2
+
r
5
16
3
, c
4
= c
3
3
r
7
16
4
, c
5
= c
4
+
r
9
16
5
, . . . . Note that c
1
corresponds to the rst term of the innite
sum, regardless of the value of the summation variable and the value of the exponent.
As n increases, c
n
(r) approximates 1 better on the interval of convergence, which is (34. 4).
21. 1(r) = ln
1 +r
1 3r
= ln(1 +r) 3ln(1 3r) =
dr
1 +r
+
dr
1 3r
=

dr
1 3(3r)
+
dr
1 3r
=

"
n=0
(31)
n
r
n
+
"
n=0
r
n
dr =

[(1 3r +r
2
3r
3
+r
4
3 ) + (1 +r +r
2
+r
3
+r
4
+ )] dr
=

(2 + 2r
2
+ 2r
4
+ ) dr =

"
n=0
2r
2n
dr = C +
"
n=0
2r
2n+1
2n + 1
But 1(0) = ln
1
1
= 0, so C = 0 and we have 1(r) =
"
n=0
2r
2n+1
2n + 1
with 1 = 1. If r = 1, then 1(r) = 2
"
n=0
1
2n + 1
,
which both diverge by the Limit Comparison Test with /
n
=
1
n
. The partial sums are c
1
=
2r
1
, c
2
= c
1
+
2r
3
3
,
c
3
= c
2
+
2r
5
5
, . . . .
As n increases, c
n
(r) approximates 1 better on the interval of convergence, which is (31. 1).
SECTION 11.9 REPRESENTATIONS OF FUNCTIONS AS POWER SERIES

487
23.
t
1 3t
8
= t
1
1 3t
8
= t
"
n=0
(t
8
)
n
=
"
n=0
t
8n+1
i

t
1 3t
8
dt = C +
"
n=0
t
8n+2
8n + 2
. The series for
1
1 3t
8
converges
when
t
8
< 1 C |t| < 1, so 1 = 1 for that series and also the series for t(1 3t
8
). By Theorem 2, the series for
t
1 3t
8
dt also has 1 = 1.
25. By Example 7, tan
31
r =
"
n=0
(31)
n
r
2n+1
2n + 1
with 1 = 1, so
r 3tan
31
r = r 3
r 3
r
3
3
+
r
5
5
3
r
7
7
+
=
r
3
3
3
r
5
5
+
r
7
7
3 =
"
n=1
(31)
n+1
r
2n+1
2n + 1
and
r 3tan
31
r
r
3
=
"
n=1
(31)
n+1
r
2n32
2n + 1
, so
r 3tan
31
r
r
3
dr = C +
"
n=1
(31)
n+1
r
2n31
(2n + 1)(2n 31)
= C +
"
n=1
(31)
n+1
r
2n31
4n
2
31
. By Theorem 2, 1 = 1.
27.
1
1 +r
5
=
1
1 3(3r
5
)
=
"
n=0
3r
5
n
=
"
n=0
(31)
n
r
5n
i
1
1 +r
5
dr =

"
n=0
(31)
n
r
5n
dr = C +
"
n=0
(31)
n
r
5n+1
5n + 1
. Thus,
1 =

0.2
0
1
1 +r
5
dr =

r 3
r
6
6
+
r
11
11
3
0.2
0
= 0.2 3
(0.2)
6
6
+
(0.2)
11
11
3 . The series is alternating, so if we use
the rst two terms, the error is at most (0.2)
11
11 E 1.9 10
39
. So 1 E 0.2 3(0.2)
6
6 E 0.199989 to six decimal places.
29. We substitute 3r for r in Example 7, and nd that
rarctan(3r) dr =

r
"
n=0
(31)
n
(3r)
2n+1
2n + 1
dr =

"
n=0
(31)
n
3
2n+1
r
2n+2
2n + 1
dr = C +
"
n=0
(31)
n
3
2n+1
r
2n+3
(2n + 1)(2n + 3)
So

0.1
0
rarctan(3r) dr =

3r
3
1 3
3
3
3
r
5
3 5
+
3
5
r
7
5 7
3
3
7
r
9
7 9
+
0.1
0
=
1
10
3
3
9
5 10
5
+
243
35 10
7
3
2187
63 10
9
+ .
The series is alternating, so if we use three terms, the error is at most
2187
63 10
9
E 3.5 10
38
. So
0.1
0
rarctan(3r) dr E
1
10
3
3
9
5 10
5
+
243
35 10
7
E 0.000 983 to six decimal places.
31. Using the result of Example 6, ln(1 3 r) = 3
"
n=1
r
n
n
, with r = 30.1, we have
ln1.1 = ln[1 3(30.1)] = 0.1 3
0.01
2
+
0.001
3
3
0.0001
4
+
0.00001
5
3 . The series is alternating, so if we use only
the rst four terms, the error is at most
0.00001
5
= 0.000002. So ln1.1 E 0.1 3
0.01
2
+
0.001
3
3
0.0001
4
E 0.09531.
488

33. (a) J
0
(r) =
"
n=0
(31)
n
r
2n
2
2n
(n!)
2
. J
0
0
(r) =
"
n=1
(31)
n
2nr
2n31
2
2n
(n!)
2
, and J
00
0
(r) =
"
n=1
(31)
n
2n(2n 31)r
2n32
2
2n
(n!)
2
, so
r
2
J
00
0
(r) +rJ
0
0
(r) +r
2
J
0
(r) =
"
n=1
(31)
n
2n(2n 31)r
2n
2
2n
(n!)
2
+
"
n=1
(31)
n
2nr
2n
2
2n
(n!)
2
+
"
n=0
(31)
n
r
2n+2
2
2n
(n!)
2
=
"
n=1
(31)
n
2n(2n 31)r
2n
2
2n
(n!)
2
+
"
n=1
(31)
n
2nr
2n
2
2n
(n!)
2
+
"
n=1
(31)
n31
r
2n
2
2n32
[(n 31)!]
2
=
"
n=1
(31)
n
2n(2n 31)r
2n
2
2n
(n!)
2
+
"
n=1
(31)
n
2nr
2n
2
2n
(n!)
2
+
"
n=1
(31)
n
(31)
31
2
2
n
2
r
2n
2
2n
(n!)
2
=
"
n=1
(31)
n
2n(2n 31) + 2n 32
2
n
2
2
2n
(n!)
2
r
2n
=
"
n=1
(31)
n
4n
2
32n + 2n 34n
2
2
2n
(n!)
2
r
2n
= 0
(b)

1
0
J
0
(r) dr =

1
0
"
n=0
(31)
n
r
2n
2
2n
(n!)
2
dr =

1
0
1 3
r
2
4
+
r
4
64
3
r
6
2304
+
dr
=

r 3
r
3
3 4
+
r
5
5 64
3
r
7
7 2304
+
1
0
= 1 3
1
12
+
1
320
3
1
16,128
+
Since
1
16,128
E 0.000062, it follows from The Alternating Series Estimation Theorem that, correct to three decimal places,
1
0
J
0
(r) dr E 1 3
1
12
+
1
320
E 0.920.
35. (a) 1(r) =
"
n=0
r
n
n!
i 1
0
(r) =
"
n=1
nr
n31
n!
=
"
n=1
r
n31
(n 31)!
=
"
n=0
r
n
n!
= 1(r)
(b) By Theorem 9.4.2, the only solution to the differential equation d1(r)dr = 1(r) is 1(r) = 1c
i
, but 1(0) = 1,
so 1 = 1 and 1(r) = c
i
.
Or: We could solve the equation d1(r)dr = 1(r) as a separable differential equation.
37. If o
n
=
r
n
n
2
, then by the Ratio Test, lim
n<"
o
n+1
o
n
= lim
n<"
r
n+1
(n + 1)
2

n
2
r
n
= |r| lim
n<"
n
n + 1
2
= |r| < 1 for
convergence, so 1 = 1. When r = 1,
"
n=1
r
n
n
2
=
"
n=1
1
n
2
which is a convergent j-series (j = 2 1), so the interval of
convergence for 1 is [31. 1]. By Theorem 2, the radii of convergence of 1
0
and 1
00
are both 1, so we need only check the
endpoints. 1(r) =
"
n=1
r
n
n
2
i 1
0
(r) =
"
n=1
nr
n31
n
2
=
"
n=0
r
n
n + 1
, and this series diverges for r = 1 (harmonic series)
and converges for r = 31 (Alternating Series Test), so the interval of convergence is [31. 1). 1
00
(r) =
"
n=1
nr
n31
n + 1
diverges
at both 1 and 31 (Test for Divergence) since lim
n<"
n
n + 1
= 1 6= 0, so its interval of convergence is (31. 1).
SECTION 11.10 TAYLOR AND MACLAURIN SERIES

489
39. By Example 7, tan
31
r =
"
n=0
(31)
n
r
2n+1
2n + 1
for |r| < 1. In particular, for r =
1
I
3
, we
have

6
= tan
31
1
I
3
=
"
n=0
(31)
n
1
I
3
2n+1
2n + 1
=
"
n=0
(31)
n
1
3
n
1
I
3
1
2n + 1
, so
=
6
I
3
"
n=0
(31)
n
(2n + 1)3
n
= 2
I
3
"
n=0
(31)
n
(2n + 1)3
n
.
11.10 Taylor and Maclaurin Series
1. Using Theorem 5 with
"
n=0
/
n
(r 35)
n
, /
n
=
1
(n)
(o)
n!
, so /
8
=
1
(8)
(5)
8!
.
3. Since 1
(n)
(0) = (n + 1)!, Equation 7 gives the Maclaurin series
"
n=0
1
(n)
(0)
n!
r
n
=
"
n=0
(n + 1)!
n!
r
n
=
"
n=0
(n + 1)r
n
. Applying the Ratio Test with o
n
= (n + 1)r
n
gives us
lim
n<"
o
n+1
o
n
= lim
n<"
(n + 2)r
n+1
(n + 1)r
n
= |r| lim
n<"
n + 2
n + 1
= |r| 1 = |r|. For convergence, we must have |r| < 1, so the
radius of convergence 1 = 1.
5.
n 1
(n)
(r) 1
(n)
(0)
0 (1 3r)
32
1
1 2(1 3r)
33
2
2 6(1 3r)
34
6
3 24(1 3r)
35
24
4 120(1 3r)
36
120
.
.
.
.
.
.
.
.
.
(1 3r)
32
= 1(0) +1
0
(0)r +
1
00
(0)
2!
r
2
+
1
000
(0)
3!
r
3
+
1
(4)
(0)
4!
r
4
+
= 1 + 2r +
6
2
r
2
+
24
6
r
3
+
120
24
r
4
+
= 1 + 2r + 3r
2
+ 4r
3
+ 5r
4
+ =
"
n=0
(n + 1)r
n
lim
n<"
o
n+1
o
n
= lim
n<"
(n + 2)r
n+1
(n + 1)r
n
= |r| lim
n<"
n + 2
n + 1
= |r| (1) = |r| < 1
for convergence, so 1 = 1.
7.
n 1
(n)
(r) 1
(n)
(0)
0 sinr 0
1 cos r
2 3
2
sinr 0
3 3
3
cos r 3
3
4
4
sinr 0
5
5
cos r
5
.
.
.
.
.
.
.
.
.
sinr = 1(0) +1
0
(0)r +
1
00
(0)
2!
r
2
+
1
000
(0)
3!
r
3
+
1
(4)
(0)
4!
r
4
+
1
(5)
(0)
5!
r
5
+
= 0 +r + 0 3

3
3!
r
3
+ 0 +

5
5!
r
5
+
= r 3

3
3!
r
3
+

5
5!
r
5
3

7
7!
r
7
+
=
"
n=0
(31)
n

2n+1
(2n + 1)!
r
2n+1
lim
n<"
o
n+1
o
n
= lim
n<"
2n+3
r
2n+3
(2n + 3)!

(2n + 1)!
2n+1
r
2n+1
= lim
n<"
2
r
2
(2n + 3)(2n + 2)
= 0 < 1 for all r, so 1 = ".
490

9.
n 1
(n)
(r) 1
(n)
(0)
0 c
5i
1
1 5c
5i
5
2 5
2
c
5i
25
3 5
3
c
5i
125
4 5
4
c
5i
625
.
.
.
.
.
.
.
.
.
c
5i
=
"
n=0
1
(n)
(0)
n!
r
n
=
"
n=0
5
n
n!
r
n
.
lim
n<"
o
n+1
o
n
= lim
n<"
5
n+1
|r|
n+1
(n + 1)!

n!
5
n
|r|
n
= lim
n<"
5 |r|
n + 1
= 0 < 1
for all r, so 1 = ".
11.
n 1
(n)
(r) 1
(n)
(0)
0 sinhr 0
1 coshr 1
2 sinhr 0
3 coshr 1
4 sinhr 0
.
.
.
.
.
.
.
.
.
1
(n)
(0) =
0 if n is even
1 if n is odd
so sinhr =
"
n=0
r
2n+1
(2n + 1)!
.
Use the Ratio Test to nd 1. If o
n
=
r
2n+1
(2n + 1)!
, then
lim
n<"
o
n+1
o
n
= lim
n<"
r
2n+3
(2n + 3)!

(2n + 1)!
r
2n+1
= r
2
lim
n<"
1
(2n + 3)(2n + 2)
= 0 < 1 for all r, so 1 = ".
13.
n 1
(n)
(r) 1
(n)
(1)
0 r
4
33r
2
+ 1 31
1 4r
3
36r 32
2 12r
2
36 6
3 24r 24
4 24 24
5 0 0
6 0 0
.
.
.
.
.
.
.
.
.
1
(n)
(r) = 0 for n D 5, so 1 has a nite series expansion about o = 1.
1(r) = r
4
33r
2
+ 1 =
4
n=0
1
(n)
(1)
n!
(r 31)
n
=
31
0!
(r 31)
0
+
32
1!
(r 31)
1
+
6
2!
(r 31)
2
+
24
3!
(r 31)
3
+
24
4!
(r 31)
4
= 31 32(r 31) + 3(r 31)
2
+ 4(r 31)
3
+ (r 31)
4
A nite series converges for all r, so 1 = ".
15. 1(r) = c
i
i 1
(n)
(r) = c
i
, so 1
(n)
(3) = c
3
and c
i
=
"
n=0
c
3
n!
(r 33)
n
. If o
n
=
c
3
n!
(r 33)
n
, then
lim
n<"
o
n+1
o
n
= lim
n<"
c
3
(r 33)
n+1
(n + 1)!

n!
c
3
(r 33)
n
= lim
n<"
|r 33|
n + 1
= 0 < 1 for all r, so 1 = ".
17.
n 1
(n)
(r) 1
(n)
()
0 cos r 31
1 3sinr 0
2 3cos r 1
3 sinr 0
4 cos r 31
.
.
.
.
.
.
.
.
.
cos r =
"
I=0
1
(I)
()
I!
(r 3)
I
= 31 +
(r 3)
2
2!
3
(r 3)
4
4!
+
(r 3)
6
6!
3
=
"
n=0
(31)
n+1
(r 3)
2n
(2n)!
lim
n<"
o
n+1
o
n
= lim
n<"
|r 3|
2n+2
(2n + 2)!

(2n)!
|r 3|
2n
= lim
n<"
|r 3|
2
(2n + 2)(2n + 1)
= 0 < 1
for all r, so 1 = ".

491
19.
n 1
(n)
(r) 1
(n)
(9)
0 r
31'2 1
3
1 3
1
2
r
33'2
3
1
2

1
3
3
2
3
4
r
35'2
3
1
2

3
3
2
1
3
5
3 3
15
8
r
37'2
3
1
2

3
3
2
3
5
2
1
3
7
.
.
.
.
.
.
.
.
.
1
I
r
=
1
3
3
1
2 3
3
(r 39) +
3
2
2
3
5
(r 39)
2
2!
3
3 5
2
3
3
7
(r 39)
3
3!
+
=
1
3
+
"
n=1
(31)
n
1 3 5 (2n 31)
2
n
3
2n+1
n!
(r 39)
n
.
lim
n<"
o
n+1
o
n
= lim
n<"
1 3 5 (2n 31)[2(n + 1) 31] |r 39|

n+1
2
n+1
3
[2(n+1)+1]
(n + 1)!

2
n
3
2n+1
n!
1 3 5 (2n 31) |r 39|
n
= lim
n<"
(2n + 1) |r 39|
2 3
2
(n + 1)
=
1
9
|r 39| < 1
for convergence, so |r 39| < 9 and 1 = 9.
21. If 1(r) = sinr, then 1
(n+1)
(r) =
n+1
sinr or
n+1
cos r. In each case,
1
(n+1)
(r)
$
n+1
, so by Formula 9
with o = 0 and A =
n+1
, |1
n
(r)| $

n+1
(n + 1)!
|r|
n+1
=
|r|
n+1
(n + 1)!
. Thus, |1
n
(r)| <0 as n <"by Equation 10.
So lim
n<"
1
n
(r) = 0 and, by Theorem 8, the series in Exercise 7 represents sinr for all r.
23. If 1(r) = sinhr, then for all n, 1
(n+1)
(r) = coshr or sinhr. Since |sinhr| < |coshr| = coshr for all r, we have
1
(n+1)
(r)
$ coshr for all n. If d is any positive number and |r| $ d, then
1
(n+1)
(r)
$ coshr $ coshd, so by
Formula 9 with o = 0 and A = coshd, we have |1
n
(r)| $
coshd
(n + 1)!
|r|
n+1
. It follows that |1
n
(r)| <0 as n <"for
|r| $ d (by Equation 10). But d was an arbitrary positive number. So by Theorem 8, the series represents sinhr for all r.
25. The general binomial series in (17) is
(1 +r)
I
=
"
n=0
I
n
r
n
= 1 +Ir +
I(I 31)
2!
r
2
+
I(I 31)(I 32)
3!
r
3
+ .
(1 +r)
1'2
=
"
n=0
1
2
n
r
n
= 1 +
1
2
r +
1
2
3
1
2
2!
r
2
+
1
2
3
1
2
3
3
2
3!
r
3
+
= 1 +
r
2
3
r
2
2
2
2!
+
1 3 r
3
2
3
3!
3
1 3 5 r
4
2
4
4!
+
= 1 +
r
2
+
"
n=2
(31)
n31
1 3 5 (2n 33)r
n
2
n
n!
for |r| < 1, so 1 = 1.
492

27.
1
(2 +r)
3
=
1
[2(1 +r2)]
3
=
1
8
1 +
r
2
33
=
1
8
"
n=0
33
n
r
2
n
. The binomial coefcient is
33
n
=
(33)(34)(35) (33 3n + 1)
n!
=
(33)(34)(35) [3(n + 2)]
n!
=
(31)
n
2 3 4 5 (n + 1)(n + 2)
2 n!
=
(31)
n
(n + 1)(n + 2)
2
Thus,
1
(2 +r)
3
=
1
8
"
n=0
(31)
n
(n + 1)(n + 2)
2
r
n
2
n
=
"
n=0
(31)
n
(n + 1)(n + 2)r
n
2
n+4
for
r
2
< 1 C |r| < 2, so 1 = 2.

29. sinr =
"
n=0
(31)
n
r
2n+1
(2n + 1)!
i 1(r) = sin(r) =
"
n=0
(31)
n
(r)
2n+1
(2n + 1)!
=
"
n=0
(31)
n

2n+1
(2n + 1)!
r
2n+1
, 1 = ".
31. c
i
=
"
n=0
r
n
n!
i c
2i
=
"
n=0
(2r)
n
n!
=
"
n=0
2
n
r
n
n!
, so 1(r) = c
i
+c
2i
=
"
n=0
1
n!
r
n
+
"
n=0
2
n
n!
r
n
=
"
n=0
2
n
+ 1
n!
r
n
,
1 = ".
33. cos r =
"
n=0
(31)
n
r
2n
(2n)!
i cos
1
2
r
2
=
"
n=0
(31)
n
1
2
r
2
2n
(2n)!
=
"
n=0
(31)
n
r
4n
2
2n
(2n)!
, so
1(r) = rcos
1
2
r
2
=
"
n=0
(31)
n
1
2
2n
(2n)!
r
4n+1
, 1 = ".
35. We must write the binomial in the form (1+ expression), so well factor out a 4.
r
I
4 +r
2
=
r
4(1 +r
2
4)
=
r
2
1 +r
2
4
=
r
2
1 +
r
2
4
31'2
=
r
2
"
n=0
3
1
2
n
r
2
4
n
=
r
2
1 +
3
1
2
r
2
4
+
3
1
2
3
3
2
2!
r
2
4
2
+
3
1
2
3
3
2
3
5
2
3!
r
2
4
3
+
=
r
2
+
r
2
"
n=1
(31)
n
1 3 5 (2n 31)
2
n
4
n
n!
r
2n
=
r
2
+
"
n=1
(31)
n
1 3 5 (2n 31)
n! 2
3n+1
r
2n+1
and
r
2
4
< 1 C
|r|
2
< 1 C |r| < 2, so 1 = 2.
37. sin
2
r =
1
2
(1 3cos 2r) =
1
2
1 3
"
n=0
(31)
n
(2r)
2n
(2n)!
=
1
2
1 31 3
"
n=1
(31)
n
(2r)
2n
(2n)!
=
"
n=1
(31)
n+1
2
2n31
r
2n
(2n)!
,
1 = "
39. cos r =
"
n=0
(31)
n
r
2n
(2n)!
i 1(r) = cos
r
2
=
"
n=0
(31)
n
r
2
2n
(2n)!
=
"
n=0
(31)
n
r
4n
(2n)!
, 1 = "
Notice that, as n increases, T
n
(r)
becomes a better approximation to 1(r).

493
41. c
i
(11)
=
"
n=0
r
n
n!
, so c
3i
=
"
n=0
(3r)
n
n!
=
"
n=0
(31)
n
r
n
n!
, so
1(r) = rc
3i
=
"
n=0
(31)
n
1
n!
r
n+1
= r 3r
2
+
1
2
r
3
3
1
6
r
4
+
1
24
r
5
3
1
120
r
6
+
=
"
n=1
(31)
n31
r
n
(n 31)!
The series for c
i
converges for all r, so the same is true of the series
for 1(r); that is, 1 = ". From the graphs of 1 and the rst few Taylor
polynomials, we see that T
n
(r) provides a closer t to 1(r) near 0 as n increases.
43. c
i
=
"
n=0
r
n
n!
, so c
30.2
=
"
n=0
(30.2)
n
n!
= 1 30.2 +
1
2!
(0.2)
2
3
1
3!
(0.2)
3
+
1
4!
(0.2)
4
3
1
5!
(0.2)
5
+
1
6!
(0.2)
6
3 .
But
1
6!
(0.2)
6
= 8.8 10
38
, so by the Alternating Series Estimation Theorem, c
30.2
E
5
n=0
(30.2)
n
n!
E 0.81873, correct to
ve decimal places.
45. (a) 1
I
1 3r
2
=

1 +
3r
2
31'2
= 1 +
3
1
2
3r
2
3
1
2
3
3
2
2!
3r
2
2
+
3
1
2
3
3
2
3
5
2
3!
3r
2
3
+
= 1 +
"
n=1
1 3 5 (2n 31)
2
n
n!
r
2n
(b) sin
31
r =

1
I
1 3r
2
dr = C +r +
"
n=1
1 3 5 (2n 31)
(2n + 1)2
n
n!
r
2n+1
= r +
"
n=1
1 3 5 (2n 31)
(2n + 1)2
n
n!
r
2n+1
since 0 = sin
31
0 = C.
47. cos r
(16)
=
"
n=0
(31)
n
r
2n
(2n)!
i cos(r
3
) =
"
n=0
(31)
n
(r
3
)
2n
(2n)!
=
"
n=0
(31)
n
r
6n
(2n)!
i
rcos(r
3
) =
"
n=0
(31)
n
r
6n+1
(2n)!
i

rcos(r
3
) dr = C +
"
n=0
(31)
n
r
6n+2
(6n + 2)(2n)!
, with 1 = ".
49. cos r
(16)
=
"
n=0
(31)
n
r
2n
(2n)!
i cos r 31 =
"
n=1
(31)
n
r
2n
(2n)!
i
cos r 31
r
=
"
n=1
(31)
n
r
2n31
(2n)!
i
cos r 31
r
dr = C +
"
n=1
(31)
n
r
2n
2n (2n)!
, with 1 = ".
51. By Exercise 47,
rcos(r
3
) dr = C +
"
n=0
(31)
n
r
6n+2
(6n + 2)(2n)!
, so
1
0
rcos(r
3
) dr =

"
n=0
(31)
n
r
6n+2
(6n + 2)(2n)!
1
0
=
"
n=0
(31)
n
(6n + 2)(2n)!
=
1
2
3
1
8 2!
+
1
14 4!
3
1
20 6!
+ , but
1
20 6!
=
1
14,400
E 0.000 069, so
1
0
rcos(r
3
) dr E
1
2
3
1
16
+
1
336
E 0.440 (correct to three decimal places) by the
Alternating Series Estimation Theorem.
494

53.
I
1 +r
4
= (1 +r
4
)
1'2
=
"
n=0
12
n
(r
4
)
n
, so

1 +r
4
dr = C +
"
n=0
12
n
r
4n+1
4n + 1
and hence, since 0.4 < 1,
we have
1 =

0.4
0
1 +r
4
dr =
"
n=0
12
n
(0.4)
4n+1
4n + 1
= (1)
(0.4)
1
0!
+
1
2
1!
(0.4)
5
5
+
1
2
3
1
2
2!
(0.4)
9
9
+
1
2
3
1
2
3
3
2
3!
(0.4)
13
13
+
1
2
3
1
2
3
3
2
3
5
2
4!
(0.4)
17
17
+
= 0.4 +
(0.4)
5
10
3
(0.4)
9
72
+
(0.4)
13
208
3
5(0.4)
17
2176
+
Now
(0.4)
9
72
E 3.6 10
36
< 5 10
36
, so by the Alternating Series Estimation Theorem, 1 E 0.4 +
(0.4)
5
10
E 0.40102
(correct to ve decimal places).
55. lim
i<0
r 3tan
31
r
r
3
= lim
i<0
r 3
r 3
1
3
r
3
+
1
5
r
5
3
1
7
r
7
+
r
3
= lim
i<0
1
3
r
3
3
1
5
r
5
+
1
7
r
7
3
r
3
= lim
i<0
1
3
3
1
5
r
2
+
1
7
r
4
3
=
1
3
since power series are continuous functions.
57. lim
i<0
sinr 3r +
1
6
r
3
r
5
= lim
i<0
r 3
1
3!
r
3
+
1
5!
r
5
3
1
7!
r
7
+
3r +
1
6
r
3
r
5
= lim
i<0
1
5!
r
5
3
1
7!
r
7
+
r
5
= lim
i<0
1
5!
3
r
2
7!
+
r
4
9!
3
=
1
5!
=
1
120
since power series are continuous functions.
59. From Equation 11, we have c
3i
2
= 1 3
r
2
1!
+
r
4
2!
3
r
6
3!
+ and we know that cos r = 1 3
r
2
2!
+
r
4
4!
3 from
Equation 16. Therefore, c
3i
2
cos r =

1 3r
2
+
1
2
r
4
3
1 3
1
2
r
2
+
1
24
r
4
3
. Writing only the terms with

degree $ 4, we get c
3i
2
cos r = 1 3
1
2
r
2
+
1
24
r
4
3r
2
+
1
2
r
4
+
1
2
r
4
+ = 1 3
3
2
r
2
+
25
24
r
4
+ .
61.
r
sinr
(15)
=
r
r 3
1
6
r
3
+
1
120
r
5
3
.
1 +
1
6
r
2
+
7
360
r
4
+
r 3
1
6
r
3
+
1
120
r
5
3 r
r 3
1
6
r
3
+
1
120
r
5
3
1
6
r
3
3
1
120
r
5
+
1
6
r
3
3
1
36
r
5
+
7
360
r
5
+
7
360
r
5
+

From the long division above,
r
sinr
= 1 +
1
6
r
2
+
7
360
r
4
+ .
63.
"
n=0
(31)
n
r
4n
n!
=
"
n=0
3r
4
n
n!
= c
3i
4
, by (11).

495
65.
"
n=0
(31)
n
2n+1
4
2n+1
(2n + 1)!
=
"
n=0
(31)
n
r
4
2n+1
(2n + 1)!
= sin
r
4
=
1
I
2
, by (15).
67. 3 +
9
2!
+
27
3!
+
81
4!
+ =
3
1
1!
+
3
2
2!
+
3
3
3!
+
3
4
4!
+ =
"
n=1
3
n
n!
=
"
n=0
3
n
n!
31 = c
3
31, by (11).
69. Assume that |1
000
(r)| $ A, so 1
000
(r) $ A for o $ r $ o +d. Now
i
a
1
000
(t) dt $
i
a
A dt i
1
00
(r) 31
00
(o) $ A(r 3o) i 1
00
(r) $ 1
00
(o) +A(r 3o). Thus,
i
a
1
00
(t) dt $
i
a
[1
00
(o) +A(t 3o)] dt i
1
0
(r) 31
0
(o) $ 1
00
(o)(r 3o) +
1
2
A(r 3o)
2
i 1
0
(r) $ 1
0
(o) +1
00
(o)(r 3o) +
1
2
A(r 3o)
2
i
i
a
1
0
(t) dt $

i
a
1
0
(o) +1
00
(o)(t 3o) +
1
2
A(t 3o)
2
dt i
1(r) 3 1(o) $ 1
0
(o)(r 3 o) +
1
2
1
00
(o)(r 3 o)
2
+
1
6
A(r 3 o)
3
. So
1(r) 3 1(o) 3 1
0
(o)(r 3 o) 3
1
2
1
00
(o)(r 3 o)
2
$
1
6
A(r 3 o)
3
. But
1
2
(r) = 1(r) 3T
2
(r) = 1(r) 31(o) 31
0
(o)(r 3o) 3
1
2
1
00
(o)(r 3o)
2
, so 1
2
(r) $
1
6
A(r 3o)
3
.
A similar argument using 1
000
(r) D 3A shows that 1
2
(r) D 3
1
6
A(r 3o)
3
. So |1
2
(r
2
)| $
1
6
A |r 3o|
3
.
Although we have assumed that r o, a similar calculation shows that this inequality is also true if r < o.
71. (a) o(r) =
"
n=0
I
n
r
n
i o
0
(r) =
"
n=1
I
n
nr
n31
, so
(1 +r)o
0
(r) = (1 +r)
"
n=1
I
n
nr
n31
=
"
n=1
I
n
nr
n31
+
"
n=1
I
n
nr
n
=
"
n=0
I
n + 1
(n + 1)r
n
+
"
n=0
I
n
nr
n
Replace n with n + 1
in the rst series
=
"
n=0
(n + 1)
I(I 31)(I 32) (I 3n + 1)(I 3n)
(n + 1)!
r
n
+
"
n=0
(n)
I(I 31)(I 32) (I 3n + 1)
n!
r
n
=
"
n=0
(n + 1)I(I 31)(I 32) (I 3n + 1)
(n + 1)!
[(I 3n) +n] r
n
= I
"
n=0
I(I 31)(I 32) (I 3n + 1)
n!
r
n
= I
"
n=0
I
n
r
n
= Io(r)
Thus, o
0
(r) =
Io(r)
1 +r
.
(b) /(r) = (1 +r)
3I
o(r) i
/
0
(r) = 3I(1 +r)
3I31
o(r) + (1 +r)
3I
o
0
(r) [Product Rule]
= 3I(1 +r)
3I31
o(r) + (1 +r)
3I
Io(r)
1 +r
[from part (a)]
= 3I(1 +r)
3I31
o(r) +I(1 +r)
3I31
o(r) = 0
(c) From part (b) we see that /(r) must be constant for r M (31. 1), so /(r) = /(0) = 1 for r M (31. 1).
Thus, /(r) = 1 = (1 +r)
3I
o(r) C o(r) = (1 +r)
I
for r M (31. 1).
496

11.11 Applications of Taylor Polynomials
1. (a)
n 1
(n)
(r) 1
(n)
(0) T
n
(r)
0 cos r 1 1
1 3sinr 0 1
2 3cos r 31 1 3
1
2
r
2
3 sinr 0 1 3
1
2
r
2
4 cos r 1 1 3
1
2
r
2
+
1
24
r
4
5 3sinr 0 1 3
1
2
r
2
+
1
24
r
4
6 3cos r 31 1 3
1
2
r
2
+
1
24
r
4
3
1
720
r
6
(b)
r 1 T
0
= T
1
T
2
= T
3
T
4
= T
5
T
6
r
4
0.7071 1 0.6916 0.7074 0.7071
r
2
0 1 30.2337 0.0200 30.0009
31 1 33.9348 0.1239 31.2114
(c) As n increases, T
n
(r) is a good approximation to 1(r) on a larger and larger interval.
3.
n 1
(n)
(r) 1
(n)
(2)
0 1r
1
2
1 31r
2
3
1
4
2 2r
3 1
4
3 36r
4
3
3
8
T
3
(r) =
3
n=0
1
(n)
(2)
n!
(r 32)
n
=
1
2
0!
3
1
4
1!
(r 32) +
1
4
2!
(r 32)
2
3
3
8
3!
(r 32)
3
=
1
2
3
1
4
(r 32) +
1
8
(r 32)
2
3
1
16
(r 32)
3
5.
n 1
(n)
(r) 1
(n)
(2)
0 cos r 0
1 3sinr 31
2 3cos r 0
3 sinr 1
T
3
(r) =
3
n=0
1
(n)
(2)
n!
r 3
r
2
n
= 3
r 3
r
2
+
1
6
r 3
r
2
3
SECTION 11.11 APPLICATIONS OF TAYLOR POLYNOMIALS

497
7.
n 1
(n)
(r) 1
(n)
(0)
0 arcsinr 0
1
1
I
1 3r
2
1
2
r
(1 3r
2
)
3'2
0
3
2r
2
+ 1
(1 3r
2
)
5'2
1
T
3
(r) =
3
n=0
1
(n)
(0)
n!
r
n
= r +
r
3
6
9.
n 1
(n)
(r) 1
(n)
(0)
0 rc
32i
0
1 (1 32r)c
32i
1
2 4(r 31)c
32i
34
3 4(3 32r)c
32i
12
T
3
(r) =
3
n=0
1
(n)
(0)
n!
r
n
=
0
1
1 +
1
1
r
1
+
34
2
r
2
+
12
6
r
3
= r 32r
2
+ 2r
3
11. You may be able to simply nd the Taylor polynomials for
1(r) = cot r using your CAS. We will list the values of 1
(n)
(4)
for n = 0 to n = 5.
n 0 1 2 3 4 5
1
(n)
(4) 1 32 4 316 80 3512
T
5
(r) =
5
n=0
1
(n)
(4)
n!
r 3
r
4
n
= 1 32
r 3
r
4
+ 2
r 3
r
4
2
3
8
3
r 3
r
4
3
+
10
3
r 3
r
4
4
3
64
15
r 3
r
4
5
For n = 2 to n = 5, T
n
(r) is the polynomial consisting of all the terms up to and including the
r 3
r
4
n
term.
13.
n 1
(n)
(r) 1
(n)
(4)
0
I
r 2
1
1
2
r
31'2 1
4
2 3
1
4
r
33'2
3
1
32
3
3
8
r
35'2
(a) 1(r) =
I
r E T
2
(r) = 2 +
1
4
(r 34) 3
132
2!
(r 34)
2
= 2 +
1
4
(r 34) 3
1
64
(r 34)
2
(b) |1
2
(r)| $
A
3!
|r 34|
3
, where |1
000
(r)| $ A. Now 4 $ r $ 4.2 i
|r 34| $ 0.2 i |r 34|
3
$ 0.008. Since 1
000
(r) is decreasing
on [4. 4.2], we can take A = |1
000
(4)| =
3
8
4
35'2
=
3
256
, so
|1
2
(r)| $
3256
6
(0.008) =
0.008
512
= 0.000 015 625.
498

(c)
From the graph of |1
2
(r)| = |
I
r 3T
2
(r)|, it seems that the
error is less than 1.52 10
35
on [4. 4.2].
15.
n 1
(n)
(r) 1
(n)
(1)
0 r
2'3
1
1
2
3
r
31'3 2
3
2 3
2
9
r
34'3
3
2
9
3
8
27
r
37'3 8
27
4 3
56
81
r
310'3
(a) 1(r) = r
2'3
E T
3
(r) = 1 +
2
3
(r 31) 3
29
2!
(r 31)
2
+
827
3!
(r 31)
3
= 1 +
2
3
(r 31) 3
1
9
(r 31)
2
+
4
81
(r 31)
3
(b) |1
3
(r)| $
A
4!
|r 31|
4
, where
1
(4)
(r)
$ A. Now 0.8 $ r $ 1.2 i

|r 31| $ 0.2 i |r 31|
4
$ 0.0016. Since
1
(4)
(r)
is decreasing
on [0.8. 1.2], we can take A =

1
(4)
(0.8)
=
56
81
(0.8)
310'3
, so
|1
3
(r)| $
56
81
(0.8)
310'3
24
(0.0016) E 0.000 096 97.
(c)
3
(r)| =

r
2'3
3T
3
(r)
, it seems that the

error is less than 0.000 053 3 on [0.8. 1.2].
17.
n 1
(n)
(r) 1
(n)
(0)
0 sec r 1
1 sec r tanr 0
2 sec r(2 sec
2
r 31) 1
3 sec r tanr(6 sec
2
r 31)
(a) 1(r) = sec r E T
2
(r) = 1 +
1
2
r
2
(b) |1
2
(r)| $
A
3!
|r|
3
, where
1
(3)
(r)
$ A. Now 30.2 $ r $ 0.2 i |r| $ 0.2 i |r|

3
$ (0.2)
3
.
1
(3)
(r) is an odd function and it is increasing on [0. 0.2] since sec r and tanr are increasing on [0. 0.2],
so
1
(3)
(r)
$ 1
(3)
(0.2) E 1.085 158 892. Thus, |1
2
(r)| $
1
(3)
(0.2)
3!
(0.2)
3
E 0.001 447.
(c)
2
(r)| = |sec r 3T
2
error is less than 0.000 339 on [30.2. 0.2].

499
19.
n 1
(n)
(r) 1
(n)
(0)
0 c
i
2
1
1 c
i
2
(2r) 0
2 c
i
2
(2 + 4r
2
) 2
3 c
i
2
(12r + 8r
3
) 0
4 c
i
2
(12 + 48r
2
+ 16r
4
)
(a) 1(r) = c
i
2
E T
3
(r) = 1 +
2
2!
r
2
= 1 +r
2
(b) |1
3
(r)| $
A
4!
|r|
4
, where
1
(4)
(r)
$ A. Now 0 $ r $ 0.1 i
r
4
$ (0.1)
4
, and letting r = 0.1 gives
|1
3
(r)| $
c
0.01
(12 + 0.48 + 0.0016)
24
(0.1)
4
E 0.00006.
(c)
3
(r)| =

c
i
2
3T
3
(r)
, it appears that the

error is less than 0.000 051 on [0. 0.1].
21.
n 1
(n)
(r) 1
(n)
(0)
0 rsinr 0
1 sinr +rcos r 0
2 2 cos r 3rsinr 2
3 33 sinr 3rcos r 0
4 34 cos r +rsinr 34
5 5 sinr +rcos r
(a) 1(r) = rsinr E T
4
(r) =
2
2!
(r 30)
2
+
34
4!
(r 30)
4
= r
2
3
1
6
r
4
(b) |1
4
(r)| $
A
5!
|r|
5
, where
1
(5)
(r)
$ A. Now 31 $ r $ 1 i
|r| $ 1, and a graph of 1
(5)
(r) shows that
1
(5)
(r)
$ 5 for 31 $ r $ 1.
Thus, we can take A = 5 and get |1
4
(r)| $
5
5!
1
5
=
1
24
= 0.0416.
(c)
4
(r)| = |rsinr 3T
4
error is less than 0.0082 on [31. 1].
23. From Exercise 5, cos r = 3
r 3
r
2
+
1
6
r 3
r
2
3
+1
3
(r), where |1
3
(r)| $
A
4!
r 3
r
2
4
with
1
(4)
(r)
= |cos r| $ A = 1. Now r = 80
= (90
310
) =

r
2
3
r
18
=
4r
9
radians, so the error is
1
3
4r
9
$
1
24
r
18
4
E 0.000 039, which means our estimate would not be accurate to ve decimal places. However,
T
3
= T
4
, so we can use
1
4
4r
9
$
1
120
r
18
5
E 0.000 001. Therefore, to ve decimal places,
cos 80
E 3
3
r
18
+
1
6
3
r
18
3
E 0.17365.
25. All derivatives of c
i
are c
i
, so |1
n
(r)| $
c
i
(n + 1)!
|r|
n+1
, where 0 < r < 0.1. Letting r = 0.1,
1
n
(0.1) $
c
0.1
(n + 1)!
(0.1)
n+1
< 0.00001, and by trial and error we nd that n = 3 satises this inequality since
500

1
3
(0.1) < 0.0000046. Thus, by adding the four terms of the Maclaurin series for c
i
corresponding to n = 0, 1, 2, and 3,
we can estimate c
0.1
to within 0.00001. (In fact, this sum is 1.10516 and c
0.1
E 1.10517.)
27. sinr = r 3
1
3!
r
3
+
1
5!
r
5
3 . By the Alternating Series
Estimation Theorem, the error in the approximation
sinr = r 3
1
3!
r
3
is less than
1
5!
r
5
< 0.01 C
r
5
< 120(0.01) C |r| < (1.2)

1'5
E 1.037. The curves
j = r 3
1
6
r
3
and j = sinr 30.01 intersect at r E 1.043, so
the graph conrms our estimate. Since both the sine function
and the given approximation are odd functions, we need to check the estimate only for r 0. Thus, the desired range of
values for r is 31.037 < r < 1.037.
29. arctanr = r 3
r
3
3
+
r
5
5
3
r
7
7
+ . By the Alternating Series
Estimation Theorem, the error is less than
3
1
7
r
7
< 0.05 C
r
7
< 0.35 C |r| < (0.35)

1'7
E 0.8607. The curves
j = r 3
1
3
r
3
+
1
5
r
5
and j = arctanr + 0.05 intersect at
r E 0.9245, so the graph conrms our estimate. Since both the
arctangent function and the given approximation are odd functions,
we need to check the estimate only for r 0. Thus, the desired
range of values for r is 30.86 < r < 0.86.
31. Let c(t) be the position function of the car, and for convenience set c(0) = 0. The velocity of the car is (t) = c
0
(t) and the
acceleration is o(t) = c
00
(t), so the second degree Taylor polynomial is T
2
(t) = c(0) +(0)t +
o(0)
2
t
2
= 20t +t
2
. We
estimate the distance traveled during the next second to be c(1) E T
2
(1) = 20 + 1 = 21 m. The function T
2
(t) would not be
accurate over a full minute, since the car could not possibly maintain an acceleration of 2 ms
2
for that long (if it did, its nal
speed would be 140 ms E 313 mih!).
33. 1 =
q
1
2
3
q
(1 +d)
2
=
q
1
2
3
q
1
2
(1 +d1)
2
=
q
1
2
1 3
1 +
d
1
32
.
We use the Binomial Series to expand (1 +d1)
32
:
1 =
q
1
2
1 3
1 32
d
1
+
2 3
2!
d
1
2
3
2 3 4
3!
d
1
3
+
=
q
1
2
d
1
33
d
1
2
+ 4
d
1
3
3
E
q
1
2
2
d
1
= 2qd
1
1
3
when 1 is much larger than d; that is, when 1 is far away from the dipole.

501
35. (a) If the water is deep, then 2d1 is large, and we know that tanhr <1 as r <". So we can approximate
tanh(2d1) E 1, and so
2
E o1(2) C E
o1(2).
(b) From the table, the rst term in the Maclaurin series of
tanhr is r, so if the water is shallow, we can approximate
tanh
2d
1
E
2d
1
, and so
2
E
o1
2

2d
1
C E
I
od.
n 1
(n)
(r) 1
(n)
(0)
0 tanhr 0
1 sech
2
r 1
2 32 sech
2
rtanhr 0
3 2 sech
2
r(3 tanh
2
r 31) 32
(c) Since tanhr is an odd function, its Maclaurin series is alternating, so the error in the approximation
tanh
2d
1
E
2d
1
is less than the rst neglected term, which is
|1
000
(0)|
3!
2d
1
3
=
1
3
2d
1
3
.
If 1 10d, then
1
3
2d
1
3
<
1
3
2
1
10
3
=

3
375
, so the error in the approximation
2
= od is less
than
o1
2

3
375
E 0.0132o1.
37. (a) 1 is the length of the arc subtended by the angle 0, so 1 = 10 i
0 = 11. Now sec 0 = (1+C)1 i 1sec 0 = 1+C i
C = 1sec 0 31 = 1sec(11) 31.
(b) Extending the result in Exercise 17, we have 1
(4)
(r) = sec r(18 sec
2
r tan
2
r + 6 sec
4
r 3sec
2
r 3tan
2
r),
so 1
(4)
(0) = 5, and sec r E T
4
(r) = 1 +
1
2
r
2
+
5
24
r
4
. By part (a),
C E 1
1 +
1
2
1
1
2
+
5
24
1
1
31 = 1+
1
2
1
1
2
1
2
+
5
24
1
1
4
1
4
31 =
1
2
21
+
51
4
241
3
.
(c) Taking 1 = 100 km and 1 = 6370 km, the formula in part (a) says that
C = 1sec(11) 31 = 6370 sec(1006370) 36370 E 0.785 009 965 44 km.
The formula in part (b) says that C E
1
2
21
+
51
4
241
3
=
100
2
2 6370
+
5 100
4
24 6370
3
E 0.785 009 957 36 km.
The difference between these two results is only 0.000 000 008 08 km, or 0.000 008 08 m!
39. Using 1(r) = T
n
(r) +1
n
(r) with n = 1 and r = v, we have 1(v) = T
1
(v) +1
1
(v), where T
1
is the rst-degree Taylor
polynomial of 1 at o. Because o = r
n
, 1(v) = 1(r
n
) +1
0
(r
n
)(v 3r
n
) +1
1
(v). But v is a root of 1, so 1(v) = 0
and we have 0 = 1(r
n
) +1
0
(r
n
)(v 3r
n
) +1
1
(v). Taking the rst two terms to the left side gives us
1
0
(r
n
)(r
n
3v) 31(r
n
) = 1
1
(v). Dividing by 1
0
(r
n
), we get r
n
3v 3
1(r
n
)
1
0
(r
n
)
=
1
1
(v)
1
0
(r
n
)
. By the formula for Newtons
method, the left side of the preceding equation is r
n+1
3v, so |r
n+1
3v| =
1
1
(v)
1
0
(r
n
)
. Taylors Inequality gives us

|1
1
(v)| $
|1
00
(v)|
2!
|v 3r
n
|
2
. Combining this inequality with the facts |1
00
(r)| $ A and |1
0
(r)| D 1 gives us
|r
n+1
3v| $
A
21
|r
n
3v|
2
.
502

11 Review
(b) See Denition 11.2.2.
(c) The terms of the sequence {o
n
} approach 3 as n becomes large.
(d) By adding sufciently many terms of the series, we can make the partial sums as close to 3 as we like.
(b) A sequence is monotonic if it is either increasing or decreasing.
(c) By Theorem 11.1.12, every bounded, monotonic sequence is convergent.
3. (a) See (4) in Section 11.2.
(b) The j-series
"
n=1
1
n
is convergent if j 1.
4. If
o
n
= 3, then lim
n<"
o
n
= 0 and lim
n<"
c
n
= 3.
5. (a) Test for Divergence: If lim
n<"
o
n
does not exist or if lim
n<"
o
n
6= 0, then the series
"
n=1
o
n
is divergent.
(b) Integral Test: Suppose 1 is a continuous, positive, decreasing function on [1. ") and let o
n
= 1(n). Then the series
"
n=1
o
n
is convergent if and only if the improper integral
"
1
1(r) dr is convergent. In other words:
(i) If
"
1
1(r) dr is convergent, then
"
n=1
o
n
is convergent.
(ii) If
"
1
1(r) dr is divergent, then
"
n=1
o
n
is divergent.
(c) Comparison Test: Suppose that
o
n
and
/
n
are series with positive terms.
(i) If
/
n
is convergent and o
n
$ /
n
for all n, then
o
n
is also convergent.
(ii) If
/
n
is divergent and o
n
D /
n
for all n, then
o
n
is also divergent.
(d) Limit Comparison Test: Suppose that
o
n
and
/
n
are series with positive terms. If lim
n<"
(o
n
/
n
) = c, where c is a
nite number and c 0, then either both series converge or both diverge.
(e) Alternating Series Test: If the alternating series
"
n=1
(31)
n31
/
n
= /
1
3/
2
+/
3
3/
4
+/
5
3/
6
+ [/
n
0]
satises (i) /
n+1
$ /
n
for all n and (ii) lim
n<"
/
n
= 0, then the series is convergent.
(f ) Ratio Test:
(i) If lim
n<"
o
n+1
o
n
= 1 < 1, then the series

"
n=1
o
n
is absolutely convergent (and therefore convergent).
(ii) If lim
n<"
o
n+1
o
n
= 1 1 or lim
n<"
o
n+1
o
n
= ", then the series

"
n=1
o
n
is divergent.
(iii) If lim
n<"
o
n+1
o
n
= 1, the Ratio Test is inconclusive; that is, no conclusion can be drawn about the convergence or
divergence of
o
n.
CHAPTER 11 REVIEW

503
(g) Root Test:
(i) If lim
n<"
q
|o
n
| = 1 < 1, then the series
"
n=1
o
n
is absolutely convergent (and therefore convergent).
(ii) If lim
n<"
q
|o
n
| = 1 1 or lim
n<"
q
|o
n
| = ", then the series
"
n=1
o
n
is divergent.
(iii) If lim
n<"
q
|o
n
| = 1, the Root Test is inconclusive..
6. (a) A series
o
n
is called absolutely convergent if the series of absolute values
|o
n
| is convergent.
(b) If a series
o
n
is absolutely convergent, then it is convergent.
(c) A series
o
n
is called conditionally convergent if it is convergent but not absolutely convergent.
7. (a) Use (3) in Section 11.3.
(b) See Example 5 in Section 11.4.
(c) By adding terms until you reach the desired accuracy given by the Alternating Series Estimation Theorem on page 712.
8. (a)
"
n=0
c
n
(r 3o)
n
(b) Given the power series
"
n=0
c
n
(r 3o)
n
, the radius of convergence is:
(i) 0 if the series converges only when r = o
(ii) "if the series converges for all r, or
(iii) a positive number 1 such that the series converges if |r 3o| < 1 and diverges if |r 3o| 1.
(c) The interval of convergence of a power series is the interval that consists of all values of r for which the series converges.
Corresponding to the cases in part (b), the interval of convergence is: (i) the single point {o}, (ii) all real numbers, that is,
the real number line (3". "), or (iii) an interval with endpoints o 31 and o +1 which can contain neither, either, or
both of the endpoints. In this case, we must test the series for convergence at each endpoint to determine the interval of
convergence.
9. (a), (b) See Theorem 11.9.2.
10. (a) T
n
(r) =
n
.=0
1
(.)
(o)
j!
(r 3o)
.
(b)
"
n=0
1
(n)
(o)
n!
(r 3o)
n
(c)
"
n=0
1
(n)
(0)
n!
r
n
[o = 0 in part (b)]
(d) See Theorem 11.10.8.
(e) See Taylors Inequality (11.10.9).
11. (a)(e) See Table 1 on page 743.
12. See the binomial series (11.10.17) for the expansion. The radius of convergence for the binomial series is 1.
504

1. False. See Note 2 after Theorem 11.2.6.
3. True. If lim
n<"
o
n
= 1, then given any - 0, we can nd a positive integer ` such that |o
n
31| < - whenever n `.
If n `, then 2n + 1 ` and |o
2n+1
31| < -. Thus, lim
n<"
o
2n+1
= 1.
5. False. For example, take c
n
= (31)
n
(n6
n
).
7. False, since lim
n<"
o
n+1
o
n
= lim
n<"
1
(n + 1)
3

n
3
1
= lim
n<"
n
3
(n + 1)
3

1n
3
1n
3
= lim
n<"
1
(1 + 1n)
3
= 1.
9. False. See the note after Example 2 in Section 11.4.
11. True. See (9) in Section 11.1.
13. True. By Theorem 11.10.5 the coefcient of r
3
is
1
000
(0)
3!
=
1
3
i 1
000
(0) = 2.
Or: Use Theorem 11.9.2 to differentiate 1 three times.
15. False. For example, let o
n
= /
n
= (31)
n
. Then {o
n
} and {/
n
} are divergent, but o
n
/
n
= 1, so {o
n
/
n
} is convergent.
17. True by Theorem 11.6.3. [
(31)
n
o
n
is absolutely convergent and hence convergent.]
19. True. 0.99999 . . . = 0.9 +0.9(0.1)
1
+0.9(0.1)
2
+0.9(0.1)
3
+ =
"
n=1
(0.9)(0.1)
n31
=
0.9
1 30.1
= 1 by the formula
for the sum of a geometric series [o = o
1
(1 3v)] with ratio v satisfying |v| < 1.
1.

2 +n
3
1 + 2n
3
converges since lim

n<"
2 +n
3
1 + 2n
3
= lim
n<"
2n
3
+ 1
1n
3
+ 2
=
1
2
.
3. lim
n<"
o
n
= lim
n<"
n
3
1 +n
2
= lim
n<"
n
1n
2
+ 1
= ", so the sequence diverges.
5. |o
n
| =
nsinn
n
2
+ 1
$
n
n
2
+ 1
<
1
n
, so |o
n
| <0 as n <". Thus, lim
n<"
o
n
= 0. The sequence {o
n
} is convergent.
7.

1 +
3
n
4n
is convergent. Let j =

1 +
3
r
4i
. Then
lim
i<"
lnj = lim
i<"
4rln(1 + 3r) = lim
i<"
ln(1 + 3r)
1(4r)
H
= lim
i<"
1
1 + 3r
3
3
r
2
31(4r
2
)
= lim
i<"
12
1 + 3r
= 12, so
lim
i<"
j = lim
n<"
1 +
3
n
4n
= c
12
.
9. We use induction, hypothesizing that o
n31
< o
n
< 2. Note rst that 1 < o
2
=
1
3
(1 + 4) =
5
3
< 2, so the hypothesis holds
for n = 2. Now assume that o
I31
< o
I
< 2. Then o
I
=
1
3
(o
I31
+ 4) <
1
3
(o
I
+ 4) <
1
3
(2 + 4) = 2. So o
I
< o
I+1
< 2,
and the induction is complete. To nd the limit of the sequence, we note that 1 = lim
n<"
o
n
= lim
n<"
o
n+1
i
1 =
1
3
(1 + 4) i 1 = 2.
CHAPTER 11 REVIEW

505
11.
n
n
3
+ 1
<
n
n
3
=
1
n
2
, so
"
n=1
n
n
3
+ 1
converges by the Comparison Test with the convergent j-series
"
n=1
1
n
2
[ j = 2 1].
13. lim
n<"
o
n+1
o
n
= lim
n<"
(n + 1)
3
5
n+1

5
n
n
3
= lim
n<"
1 +
1
n
1
5
=
1
5
< 1, so
"
n=1
n
3
5
n
15. Let 1(r) =
1
r
I
lnr
. Then 1 is continuous, positive, and decreasing on [2. "), so the Integral Test applies.
"
2
1(r) dr = lim
I<"
I
2
1
r
I
lnr
dr

u = ln r, Ju =
1
r
Jr
= lim
I<"
ln I
ln 2
&
31'2
d& = lim
I<"
2
I
&
ln I
ln 2
= lim
I<"
2
I
lnt 32
I
ln2
= ".
so the series
"
n=2
1
n
I
lnn
diverges.
17. |o
n
| =
cos 3n
1 + (1.2)
n
$
1
1 + (1.2)
n
<
1
(1.2)
n
=

5
6
n
, so
"
n=1
|o
n
| converges by comparison with the convergent geometric
series
"
n=1
5
6
n

v =
5
6
< 1
. It follows that
"
n=1
o
n
converges (by Theorem 3 in Section 11.6).
19. lim
n<"
o
n+1
o
n
= lim
n<"
1 3 5 (2n 31)(2n + 1)
5
n+1
(n + 1)!

5
n
n!
1 3 5 (2n 31)
= lim
n<"
2n + 1
5(n + 1)
=
2
5
< 1, so the series
21. /
n
=
I
n
n + 1
0, {/
n
n<"
/
n
= 0, so the series
"
n=1
(31)
n31
I
n
n + 1
converges by the Alternating
Series Test.
23. Consider the series of absolute values:
"
n=1
n
31'3
is a p-series with j =
1
3
$ 1 and is therefore divergent. But if we apply the
Alternating Series Test, we see that /
n
=
1
3
I
n
0, {/
n
n<"
/
n
= 0, so the series
"
n=1
(31)
n31
n
31'3
converges. Thus,
"
n=1
(31)
n31
n
31'3
25.

o
n+1
o
n
(31)
n+1
(n + 2)3
n+1
2
2n+3

2
2n+1
(31)
n
(n + 1)3
n
=
n + 2
n + 1

3
4
=
1 + (2n)
1 + (1n)

3
4
<
3
4
< 1 as n <", so by the Ratio
Test,
"
n=1
(31)
n
(n + 1)3
n
2
2n+1
is absolutely convergent.
27.
"
n=1
(33)
n31
2
3n
=
"
n=1
(33)
n31
(2
3
)
n
=
"
n=1
(33)
n31
8
n
=
1
8
"
n=1
(33)
n31
8
n31
=
1
8
"
n=1
3
3
8
n31
=
1
8
1
1 3(338)
=
1
8

8
11
=
1
11
506

29.
"
n=1
[tan
31
(n + 1) 3tan
31
n] = lim
n<"
c
n
= lim
n<"
[(tan
31
2 3tan
31
1) + (tan
31
3 3tan
31
2) + + (tan
31
(n + 1) 3tan
31
n)]
= lim
n<"
[tan
31
(n + 1) 3tan
31
1] =
r
2
3
r
4
=
r
4
31. 1 3c +
c
2
2!
3
c
3
3!
+
c
4
4!
3 =
"
n=0
(31)
n
c
n
n!
=
"
n=0
(3c)
n
n!
= c
3c
since c
i
=
"
n=0
r
n
n!
for all r.
33. coshr =
1
2
(c
i
+c
3i
) =
1
2
"
n=0
r
n
n!
+
"
n=0
(3r)
n
n!
=
1
2
1 +r +
r
2
2!
+
r
3
3!
+
r
4
4!
+
1 3r +
r
2
2!
3
r
3
3!
+
r
4
4!
3
=
1
2
2 + 2
r
2
2!
+ 2
r
4
4!
+
= 1 +
1
2
r
2
+
"
n=2
r
2n
(2n)!
D 1 +
1
2
r
2
for all r
35.
"
n=1
(31)
n+1
n
5
= 1 3
1
32
+
1
243
3
1
1024
+
1
3125
3
1
7776
+
1
16,807
3
1
32,768
+ .
Since /
8
=
1
8
5
=
1
32,768
< 0.000031,
"
n=1
(31)
n+1
n
5
E
7
n=1
(31)
n+1
n
5
E 0.9721.
37.
"
n=1
1
2 + 5
n
E
8
n=1
1
2 + 5
n
E 0.18976224. To estimate the error, note that
1
2 + 5
n
<
1
5
n
, so the remainder term is
1
8
=
"
n=9
1
2 + 5
n
<
"
n=9
1
5
n
=
15
9
1 315
= 6.4 10
37

geometric series with o =
1
5
9
and v =
1
5
.
39. Use the Limit Comparison Test. lim
n<"
n+1
n
o
n
o
n
= lim
n<"
n + 1
n
= lim
n<"
1 +
1
n
= 1 0.
Since
|o
n
| is convergent, so is
n + 1
n
o
n
, by the Limit Comparison Test.

41. lim
n<"
o
n+1
o
n
= lim
n<"
|r + 2|
n+1
(n + 1) 4
n+1

n4
n
|r + 2|
n
= lim
n<"
n
n + 1
|r + 2|
4
=
|r + 2|
4
< 1 C |r + 2| < 4, so 1 = 4.
|r + 2| < 4 C 34 < r + 2 < 4 C 36 < r < 2. If r = 36, then the series
"
n=1
(r + 2)
n
n4
n
becomes
"
n=1
(34)
n
n4
n
=
"
n=1
(31)
n
n
, the alternating harmonic series, which converges by the Alternating Series Test. When r = 2, the
series becomes the harmonic series
"
n=1
1
n
, which diverges. Thus, 1 = [36. 2).
43. lim
n<"
o
n+1
o
n
= lim
n<"
2
n+1
(r 33)
n+1
I
n + 4

I
n + 3
2
n
(r 33)
n
= 2 |r 33| lim
n<"
n + 3
n + 4
= 2 |r 33| < 1 C |r 33| <
1
2
,
so 1 =
1
2
. |r 33| <
1
2
C 3
1
2
< r 33 <
1
2
C
5
2
< r <
7
2
. For r =
7
2
, the series
"
n=1
2
n
(r 33)
n
I
n + 3
becomes
"
n=0
1
I
n + 3
=
"
n=3
1
n
1'2
, which diverges

j =
1
2
$ 1
, but for r =
5
2
, we get
"
n=0
(31)
n
I
n + 3
, which is a convergent
alternating series, so 1 =

5
2
.
7
2
.
CHAPTER 11 REVIEW

507
45.
n 1
(n)
(r) 1
(n)
r
6
0 sinr
1
2
1 cos r
I
3
2
2 3sinr 3
1
2
3 3cos r 3
I
3
2
4 sinr
1
2
.
.
.
.
.
.
.
.
.
sinr = 1
+1
0
r 3

6
+
1
00
2!
r 3

6
2
+
1
(3)
3!
r 3

6
3
+
1
(4)
4!
r 3

6
4
+
=
1
2
1 3
1
2!
r 3

6
2
+
1
4!
r 3

6
4
3
+
I
3
2
r 3

6
3
1
3!
r 3

6
3
+
=
1
2
"
n=0
(31)
n
1
(2n)!
r 3

6
2n
+
I
3
2
"
n=0
(31)
n
1
(2n + 1)!
r 3

6
2n+1
47.
1
1 +r
=
1
1 3(3r)
=
"
n=0
(3r)
n
=
"
n=0
(31)
n
r
n
for |r| < 1 i
r
2
1 +r
=
"
n=0
(31)
n
r
n+2
with 1 = 1.
49.
1
1 3r
=
"
n=0
r
n
for |r| < 1 i ln(1 3r) = 3
dr
1 3r
= 3
"
n=0
r
n
dr = C 3
"
n=0
r
n+1
n + 1
.
ln(1 30) = C 30 i C = 0 i ln(1 3r) = 3
"
n=0
r
n+1
n + 1
=
"
n=1
3r
n
n
with 1 = 1.
51. sinr =
"
n=0
(31)
n
r
2n+1
(2n + 1)!
i sin(r
4
) =
"
n=0
(31)
n
(r
4
)
2n+1
(2n + 1)!
=
"
n=0
(31)
n
r
8n+4
(2n + 1)!
for all r, so the radius of
convergence is ".
53. 1(r) =
1
4
I
16 3r
=
1
4
16(1 3r16)
=
1
4
I
16
1 3
1
16
r
1'4
=
1
2
1 3
1
16
r
31'4
=
1
2
1 +
3
1
4
3
r
16
3
1
4
3
5
4
2!
3
r
16
2
+
3
1
4

3
5
4

3
9
4
3!
3
r
16
3
+
=
1
2
+
"
n=1
1 5 9 (4n 33)
2 4
n
n! 16
n
r
n
=
1
2
+
"
n=1
1 5 9 (4n 33)
2
6n+1
n!
r
n
for
3
r
16
< 1 C |r| < 16, so 1 = 16.

55. c
i
=
"
n=0
r
n
n!
, so
c
i
r
=
1
r
"
n=0
r
n
n!
=
"
n=0
r
n31
n!
= r
31
+
"
n=1
r
n31
n!
=
1
r
+
"
n=1
r
n31
n!
and
c
i
r
dr = C + ln|r| +
"
n=1
r
n
n n!
.
508

57. (a)
n 1
(n)
(r) 1
(n)
(1)
0 r
1'2
1
1
1
2
r
31'2 1
2
2 3
1
4
r
33'2
3
1
4
3
3
8
r
35'2 3
8
4 3
15
16
r
37'2
3
15
16
.
.
.
.
.
.
.
.
.
I
r E T
3
(r) = 1 +
12
1!
(r 31) 3
14
2!
(r 31)
2
+
38
3!
(r 31)
3
= 1 +
1
2
(r 31) 3
1
8
(r 31)
2
+
1
16
(r 31)
3
(b) (c) |1
3
(r)| $
A
4!
|r 31|
4
, where
1
(4)
(r)
$ A with
1
(4)
(r) = 3
15
16
r
37'2
. Now 0.9 $ r $ 1.1 i
30.1 $ r 31 $ 0.1 i (r 31)
4
$ (0.1)
4
,
and letting r = 0.9 gives A =
15
16(0.9)
7'2
, so
|1
3
(r)| $
15
16(0.9)
7'2
4!
(0.1)
4
E 0.000 005 648
E 0.000 006 = 6 10
36
(d)
3
(r)| = |
I
r 3T
3
(r)|, it appears
that the error is less than 5 10
36
on [0.9. 1.1].
59. sinr =
"
n=0
(31)
n
r
2n+1
(2n + 1)!
= r 3
r
3
3!
+
r
5
5!
3
r
7
7!
+ , so sinr 3r = 3
r
3
3!
+
r
5
5!
3
r
7
7!
+ and
sinr 3r
r
3
= 3
1
3!
+
r
2
5!
3
r
4
7!
+ . Thus, lim
i<0
sinr 3r
r
3
= lim
i<0
3
1
6
+
r
2
120
3
r
4
5040
+
= 3
1
6
.
61. 1(r) =
"
n=0
c
n
r
n
i 1(3r) =
"
n=0
c
n
(3r)
n
=
"
n=0
(31)
n
c
n
r
n
(a) If 1 is an odd function, then 1(3r) = 31(r) i
"
n=0
(31)
n
c
n
r
n
=
"
n=0
3c
n
r
n
. The coefcients of any power series
are uniquely determined (by Theorem 11.10.5), so (31)
n
c
n
= 3c
n
.
If n is even, then (31)
n
= 1, so c
n
= 3c
n
i 2c
n
= 0 i c
n
= 0. Thus, all even coefcients are 0, that is,
c
0
= c
2
= c
4
= = 0.
(b) If 1 is even, then 1(3r) = 1(r) i
"
n=0
(31)
n
c
n
r
n
=
"
n=0
c
n
r
n
i (31)
n
c
n
= c
n
.
If n is odd, then (31)
n
= 31, so 3c
n
= c
n
i 2c
n
= 0 i c
n
= 0. Thus, all odd coefcients are 0,
that is, c
1
= c
3
= c
5
= = 0.
PROBLEMS PLUS
1. It would be far too much work to compute 15 derivatives of 1. The key idea is to remember that 1
(n)
(0) occurs in the
coefcient of r
n
in the Maclaurin series of 1. We start with the Maclaurin series for sin: sinr = r 3
r
3
3!
+
r
5
5!
3 .
Then sin(r
3
) = r
3
3
r
9
3!
+
r
15
5!
3 , and so the coefcient of r
15
is
1
(15)
(0)
15!
=
1
5!
. Therefore,
1
(15)
(0) =
15!
5!
= 6 7 8 9 10 11 12 13 14 15 = 10,897,286,400.
3. (a) From Formula 14a in Appendix D, with r = j = 0, we get tan20 =
2 tan0
1 3tan
2
0
, so cot 20 =
1 3tan
2
0
2 tan0
i
2 cot 20 =
1 3tan
2
0
tan0
= cot 0 3tan0. Replacing 0 by
1
2
r, we get 2 cot r = cot
1
2
r 3tan
1
2
r, or
tan
1
2
r = cot
1
2
r 32 cot r.
(b) From part (a) with
r
2
n31
in place of r, tan
r
2
n
= cot
r
2
n
32 cot
r
2
n31
, so the nth partial sum of
"
n=1
1
2
n
tan
r
2
n
is
c
n
=
tan(r2)
2
+
tan(r4)
4
+
tan(r8)
8
+ +
tan(r2
n
)
2
n
=

cot(r2)
2
3cot r
cot(r4)
4
3
cot(r2)
2
cot(r8)
8
3
cot(r4)
4
+
+
cot(r2
n
)
2
n
3
cot(r2
n31
)
2
n31
= 3cot r +
cot(r2
n
)
2
n
[telescoping sum]
Now
cot(r2
n
)
2
n
=
cos(r2
n
)
2
n
sin(r2
n
)
=
cos(r2
n
)
r

r2
n
sin(r2
n
)
<
1
r
1 =
1
r
as n <"since r2
n
<0
for r 6= 0. Therefore, if r 6= 0 and r 6= I where I is any integer, then
"
n=1
1
2
n
tan
r
2
n
= lim
n<"
c
n
= lim
n<"
3cot r +
1
2
n
cot
r
2
n
= 3cot r +
1
r
If r = 0, then all terms in the series are 0, so the sum is 0.
5. (a) At each stage, each side is replaced by four shorter sides, each of length
1
3
of the side length at the preceding stage. Writing c
0
and /
0
for the
number of sides and the length of the side of the initial triangle, we
generate the table at right. In general, we have c
n
= 3 4
n
and
/
n
=

1
3
n
, so the length of the perimeter at the nth stage of construction
is j
n
= c
n
/
n
= 3 4
n
1
3
n
= 3
4
3
n
.
c
0
= 3 /
0
= 1
c
1
= 3 4 /
1
= 13
c
2
= 3 4
2
/
2
= 13
2
c
3
= 3 4
3
/
3
= 13
3
.
.
.
.
.
.
(b) j
n
=
4
n
3
n31
= 4
4
3
n31
. Since
4
3
1, j
n
<"as n <".
509
510

(c) The area of each of the small triangles added at a given stage is one-ninth of the area of the triangle added at the preceding
stage. Let o be the area of the original triangle. Then the area o
n
of each of the small triangles added at stage n is
o
n
= o
1
9
n
=
o
9
n
. Since a small triangle is added to each side at every stage, it follows that the total area
n
added to the
gure at the nth stage is
n
= c
n31
o
n
= 3 4
n31
o
9
n
= o
4
n31
3
2n31
. Then the total area enclosed by the snowake
curve is = o +
1
+
2
+
3
+ = o +o
1
3
+o
4
3
3
+o
4
2
3
5
+o
4
3
3
7
+ . After the rst term, this is a
geometric series with common ratio
4
9
, so = o +
o3
1 3
4
9
= o +
o
3

9
5
=
8o
5
. But the area of the original equilateral
triangle with side 1 is o =
1
2
1 sin

3
=
I
3
4
. So the area enclosed by the snowake curve is
8
5

I
3
4
=
2
I
3
5
.
7. (a) Let o = arctanr and / = arctanj. Then, from Formula 14b in Appendix D,
tan(o 3/) =
tano 3tan/
1 + tano tan/
=
tan(arctanr) 3tan(arctanj)
1 + tan(arctanr) tan(arctanj)
=
r 3j
1 +rj
Now arctanr 3arctanj = o 3/ = arctan(tan(o 3/)) = arctan
r 3j
1 +rj
since 3
r
2
< o 3/ <
r
2
.
(b) From part (a) we have
arctan
120
119
3arctan
1
239
= arctan
120
119
3
1
239
1 +
120
119

1
239
= arctan
28,561
28,441
28,561
28,441
= arctan1 =
r
4
(c) Replacing j by 3j in the formula of part (a), we get arctanr + arctanj = arctan
r +j
1 3rj
. So
4 arctan
1
5
= 2
arctan
1
5
+ arctan
1
5
= 2 arctan
1
5
+
1
5
1 3
1
5

1
5
= 2 arctan
5
12
= arctan
5
12
+ arctan
5
12
= arctan
5
12
+
5
12
1 3
5
12

5
12
= arctan
120
119
Thus, from part (b), we have 4 arctan
1
5
3arctan
1
239
= arctan
120
119
3arctan
1
239
=
r
4
.
(d) From Example 7 in Section 11.9 we have arctanr = r 3
r
3
3
+
r
5
5
3
r
7
7
+
r
9
9
3
r
11
11
+ , so
arctan
1
5
=
1
5
3
1
3 5
3
+
1
5 5
5
3
1
7 5
7
+
1
9 5
9
3
1
11 5
11
+
This is an alternating series and the size of the terms decreases to 0, so by the Alternating Series Estimation Theorem,
the sum lies between c
5
and c
6
, that is, 0.197395560 < arctan
1
5
< 0.197395562.
(e) From the series in part (d) we get arctan
1
239
=
1
239
3
1
3 239
3
+
1
5 239
5
3 . The third term is less than
2.6 10
313
, so by the Alternating Series Estimation Theorem, we have, to nine decimal places,
arctan
1
239
E c
2
E 0.004184076. Thus, 0.004184075 < arctan
1
239
< 0.004184077.

511
(f ) From part (c) we have = 16 arctan
1
5
34 arctan
1
239
, so from parts (d) and (e) we have
16(0.197395560) 34(0.004184077) < < 16(0.197395562) 34(0.004184075) i
3.141592652 < < 3.141592692. So, to 7 decimal places, E 3.1415927.
9. We start with the geometric series
"
n=0
r
n
=
1
1 3r
, |r| < 1, and differentiate:
"
n=1
nr
n31
=
d
dr
"
n=0
r
n
=
d
dr
1
1 3r
=
1
(1 3r)
2
for |r| < 1 i
"
n=1
nr
n
= r
"
n=1
nr
n31
=
r
(1 3r)
2
for |r| < 1. Differentiate again:
"
n=1
n
2
r
n31
=
d
dr
r
(1 3r)
2
=
(1 3r)
2
3r 2(1 3r)(31)
(1 3r)
4
=
r + 1
(1 3r)
3
i
"
n=1
n
2
r
n
=
r
2
+r
(1 3r)
3
i
"
n=1
n
3
r
n31
=
d
dr
r
2
+r
(1 3r)
3
=
(1 3r)
3
(2r + 1) 3(r
2
+r)3(1 3r)
2
(31)
(1 3r)
6
=
r
2
+ 4r + 1
(1 3r)
4
i
"
n=1
n
3
r
n
=
r
3
+ 4r
2
+r
(1 3r)
4
, |r| < 1. The radius of convergence is 1 because that is the radius of convergence for the
geometric series we started with. If r = 1, the series is
n
3
(1)
n
, which diverges by the Test For Divergence, so the
interval of convergence is (31. 1).
11. ln
1 3
1
n
2
= ln
n
2
31
n
2
= ln
(n + 1)(n 31)
n
2
= ln[(n + 1)(n 31)] 3lnn
2
= ln(n + 1) + ln(n 31) 32 lnn = ln(n 31) 3lnn 3lnn + ln(n + 1)
= ln
n 31
n
3[lnn 3ln(n + 1)] = ln
n 31
n
3ln
n
n + 1
.
Let c
I
=
I
n=2
ln
1 3
1
n
2
=
I
n=2
ln
n 31
n
3ln
n
n + 1
for I D 2. Then
c
I
=

ln
1
2
3ln
2
3
ln
2
3
3ln
3
4
+ +
ln
I 31
I
3ln
I
I + 1
= ln
1
2
3ln
I
I + 1
, so
"
n=2
ln
1 3
1
n
2
= lim
I<"
c
I
= lim
I<"
ln
1
2
3ln
I
I + 1
= ln
1
2
3ln1 = ln1 3ln2 3ln1 = 3ln2.
13. (a) The x-intercepts of the curve occur where sinr = 0 C r = n,
n an integer. So using the formula for disks (and either a CAS or
sin
2
r =
1
2
(1 3cos 2r) and Formula 99 to evaluate the integral),
the volume of the nth bead is
\
n
=
nr
(n31)
(c
3i'10
sinr)
2
dr =
nr
(n31)
c
3i'5
sin
2
rdr
=
250r
101
(c
3(n31)r'5
3c
3nr'5
)
512

(b) The total volume is
"
0
c
3i'5
sin
2
rdr =
"
n=1
\
n
=
250r
101
"
n=1
[c
3(n31)r'5
3c
3nr'5
] =
250r
101
[telescoping sum].
Another method: If the volume in part (a) has been written as \
n
=
250r
101
c
3nr'5
(c
r'5
31), then we recognize
"
n=1
\
n
as a geometric series with o =
250r
101
(1 3c
3r'5
) and v = c
3r'5
.
15. If 1 is the length of a side of the equilateral triangle, then the area is =
1
2
1
I
3
2
1 =
I
3
4
1
2
and so 1
2
=
4
I
3
.
Let v be the radius of one of the circles. When there are n rows of circles, the gure shows that
1 =
I
3 v +v + (n 32)(2v) +v +
I
3 v = v
2n 32 + 2
I
3
, so v =
1
2
n +
I
3 31
.
The number of circles is 1 + 2 + +n =
n(n + 1)
2
, and so the total area of the circles is
n
=
n(n + 1)
2
v
2
=
n(n + 1)
2

1
2
4
n +
I
3 31
2
=
n(n + 1)
2

4
I
3
4
n +
I
3 31
2
=
n(n + 1)
n +
I
3 31
2
I
3
i
=
n(n + 1)
n +
I
3 31
2
I
3
=
1 + 1n
1 +
I
3 31
2
I
3
<

2
I
3
as n <"
17. As in Section 11.9 we have to integrate the function r
i
by integrating series. Writing r
i
= (c
ln i
)
i
= c
iln i
and using the
Maclaurin series for c
i
, we have r
i
= (c
ln i
)
i
= c
iln i
=
"
n=0
(rlnr)
n
n!
=
"
n=0
r
n
(lnr)
n
n!
. As with power series, we can
integrate this series term-by-term:

1
0
r
i
dr =
"
n=0
1
0
r
n
(lnr)
n
n!
dr =
"
n=0
1
n!
1
0
r
n
(lnr)
n
dr. We integrate by parts
with & = (lnr)
n
, d = r
n
dr, so d& =
n(lnr)
n31
r
dr and =
r
n+1
n + 1
:
1
0
r
n
(lnr)
n
dr = lim
I<0
+
1
I
r
n
(lnr)
n
dr = lim
I<0
+
r
n+1
n + 1
(lnr)
n
1
I
3 lim
I<0
+
1
I
n
n + 1
r
n
(lnr)
n31
dr
= 0 3
n
n + 1
1
0
r
n
(lnr)
n31
dr
(where lHospitals Rule was used to help evaluate the rst limit). Further integration by parts gives
1
0
r
n
(lnr)
I
dr = 3
I
n + 1
1
0
r
n
(lnr)
I31
dr and, combining these steps, we get
1
0
r
n
(lnr)
n
dr =
(31)
n
n!
(n + 1)
n
1
0
r
n
dr =
(31)
n
n!
(n + 1)
n+1
i
1
0
r
i
dr =
"
n=0
1
n!
1
0
r
n
(lnr)
n
dr =
"
n=0
1
n!
(31)
n
n!
(n + 1)
n+1
=
"
n=0
(31)
n
(n + 1)
n+1
=
"
n=1
(31)
n31
n
n
.

513
19. Let 1(r) =
"
r=0
c
r
r
r
and o(r) = c
](i)
=
"
n=0
d
n
r
n
. Then o
0
(r) =
"
n=0
nd
n
r
n31
, so nd
n
occurs as the coefcient
of r
n31
. But also
o
0
(r) = c
](i)
1
0
(r) =

"
n=0
d
n
r
n
"
r=1
ic
r
r
r31
=

d
0
+d
1
r +d
2
r
2
+ +d
n31
r
n31
+
c
1
+ 2c
2
r + 3c
3
r
2
+ +nc
n
r
n31
+
so the coefcient of r
n31
is c
1
d
n31
+ 2c
2
d
n32
+ 3c
3
d
n33
+ +nc
n
d
0
=
n
.=1
jc
.
d
n3.
. Therefore, nd
n
=
n
.=1
jc
.
d
n3.
.
21. Call the series o. We group the terms according to the number of digits in their denominators:
o =

1
1
+
1
2
+ +
1
8
+
1
9
. .. .
o
1
+

1
11
+ +
1
99
. .. .
o
2
+

1
111
+ +
1
999
. .. .
o
3
+
Now in the group o
n
, since we have 9 choices for each of the n digits in the denominator, there are 9
n
terms.
Furthermore, each term in o
n
is less than
1
10
q1
[except for the rst term in o
1
]. So o
n
< 9
n
1
10
q1
= 9
9
10
n31
.
Now
"
n=1
9
9
10
n31
is a geometric series with o = 9 and v =
9
10
< 1. Therefore, by the Comparison Test,
o =
"
n=1
o
n
<
"
n=1
9
9
10
n31
=
9
1 39'10
= 90.
23. & = 1 +
r
3
3!
+
r
6
6!
+
r
9
9!
+ , = r +
r
4
4!
+
r
7
7!
+
r
10
10!
+ , n =
r
2
2!
+
r
5
5!
+
r
8
8!
+ .
Use the Ratio Test to show that the series for &, , and n have positive radii of convergence ("in each case), so
Theorem 11.9.2 applies, and hence, we may differentiate each of these series:
d&
dr
=
3r
2
3!
+
6r
5
6!
+
9r
8
9!
+ =
r
2
2!
+
r
5
5!
+
r
8
8!
+ = n
Similarly,
d
dr
= 1 +
r
3
3!
+
r
6
6!
+
r
9
9!
+ = &, and
dn
dr
= r +
r
4
4!
+
r
7
7!
+
r
10
10!
+ = .
So &
0
= n,
0
= &, and n
0
= . Now differentiate the left hand side of the desired equation:
d
dr
(&
3
+
3
+n
3
33&n) = 3&
2
&
0
+ 3
2
0
+ 3n
2
n
0
33(&
0
n +&
0
n +&n
0
)
= 3&
2
n + 3
2
& + 3n
2
33(n
2
+&
2
n +&
2
) = 0 i
&
3
+
3
+n
3
33&n = C. To nd the value of the constant C, we put r = 0 in the last equation and get
1
3
+ 0
3
+ 0
3
33(1 0 0) = C i C = 1, so &
3
+
3
+n
3
33&n = 1.
13 VECTORS AND THE GEOMETRY OF SPACE ET 12
13.1 Three-Dimensional Coordinate Systems ET 12.1
1. We start at the origin, which has coordinates (0. 0. 0). First we move 4 units along the positive r-axis, affecting only the
r-coordinate, bringing us to the point (4. 0. 0). We then move 3 units straight downward, in the negative :-direction. Thus
only the :-coordinate is affected, and we arrive at (4. 0. 33).
3. The distance from a point to the r:-plane is the absolute value of the j-coordinate of the point. Q(35. 31. 4) has the
j-coordinate with the smallest absolute value, so Q is the point closest to the r:-plane. 1(0. 3. 8) must lie in the j:-plane
since the distance from1 to the j:-plane, given by the r-coordinate of 1, is 0.
5. The equation r +j = 2 represents the set of all points in
R
3
whose r- and j-coordinates have a sum of 2, or
equivalently where j = 2 3r. This is the set
{(r. 2 3r. :) | r M R. : M R} which is a vertical plane
that intersects the rj-plane in the line j = 2 3r, : = 0.
7. We can nd the lengths of the sides of the triangle by using the distance formula between pairs of vertices:
|1Q| =

(7 33)
2
+ [0 3(32)]
2
+ [1 3(33)]
2
=
I
16 + 4 + 16 = 6
|Q1| =

(1 37)
2
+ (2 30)
2
+ (1 31)
2
=
I
36 + 4 + 0 =
I
40 = 2
I
10
|11| =

(3 31)
2
+ (32 32)
2
+ (33 31)
2
=
I
4 + 16 + 16 = 6
The longest side is Q1, but the Pythagorean Theorem is not satised: |1Q|
2
+ |11|
2
6= |Q1|
2
. Thus 1Q1 is not a right
triangle. 1Q1 is isosceles, as two sides have the same length.
9. (a) First we nd the distances between points:
|1| =

(3 32)
2
+ (7 34)
2
+ (32 32)
2
=
I
26
|1C| =

(1 33)
2
+ (3 37)
2
+ [3 3(32)]
2
=
I
45 = 3
I
5
|C| =

(1 32)
2
+ (3 34)
2
+ (3 32)
2
=
I
3
In order for the points to lie on a straight line, the sum of the two shortest distances must be equal to the longest distance.
Since
I
26 +
I
3 6= 3
I
5, the three points do not lie on a straight line.
105
106 CHAPTER 13 VECTORS AND THE GEOMETRY OF SPACE ET CHAPTER 12
(b) First we nd the distances between points:
|11| =

(1 30)
2
+ [32 3(35)]
2
+ (4 35)
2
=
I
11
|11| =

(3 31)
2
+ [4 3(32)]
2
+ (2 34)
2
=
I
44 = 2
I
11
|11| =

(3 30)
2
+ [4 3(35)]
2
+ (2 35)
2
=
I
99 = 3
I
11
Since |11| + |11| = |11|, the three points lie on a straight line.
11. An equation of the sphere with center (1. 34. 3) and radius 5 is (r 31)
2
+ [j 3(34)]
2
+ (: 33)
2
= 5
2
or
(r 31)
2
+ (j + 4)
2
+ (: 33)
2
= 25. The intersection of this sphere with the r:-plane is the set of points on the sphere
whose j-coordinate is 0. Putting j = 0 into the equation, we have (r 31)
2
+ 4
2
+ (: 33)
2
= 25. j = 0 or
(r 31)
2
+ (: 33)
2
= 9. j = 0, which represents a circle in the r:-plane with center (1. 0. 3) and radius 3.
13. The radius of the sphere is the distance between (4. 3. 31) and (3. 8. 1): v =

(3 34)
2
+ (8 33)
2
+ [1 3(31)]
2
=
I
30.
Thus, an equation of the sphere is (r 33)
2
+ (j 38)
2
+ (: 31)
2
= 30.
15. Completing squares in the equation r
2
+ j
2
+ :
2
3 6r + 4j 3 2: = 11 gives
(r
2
36r + 9) + (j
2
+ 4j + 4) + (:
2
32: + 1) = 11 + 9 + 4 + 1 i (r 33)
2
+ (j + 2)
2
+ (: 31)
2
= 25, which we
recognize as an equation of a sphere with center (3. 32. 1) and radius 5.
17. Completing squares in the equation 2r
2
3 8r + 2j
2
+ 2:
2
+ 24: = 1 gives
2(r
2
34r + 4) + 2j
2
+ 2(:
2
+ 12: + 36) = 1 + 8 + 72 i 2(r 32)
2
+ 2j
2
+ 2(: + 6)
2
= 81 i
(r 32)
2
+j
2
+ (: + 6)
2
=
81
2
, which we recognize as an equation of a sphere with center (2. 0. 36) and radius
81
2
= 9
I
2.
19. (a) If the midpoint of the line segment from 1
1
(r
1
. j
1
. :
1
) to 1
2
(r
2
. j
2
. :
2
) is Q =

r
1
+r
2
2
.
j
1
+j
2
2
.
:
1
+:
2
2
,
then the distances |1
1
Q| and |Q1
2
| are equal, and each is half of |1
1
1
2
|. We verify that this is the case:
|1
1
1
2
| =

(r
2
3r
1
)
2
+ (j
2
3j
1
)
2
+ (:
2
3:
1
)
2
|1
1
Q| =

1
2
(r
1
+r
2
) 3r
1
2
+
1
2
(j
1
+j
2
) 3j
1
2
+
1
2
(:
1
+:
2
) 3:
1
2
=

1
2
r
2
3
1
2
r
1
2
+
1
2
j
2
3
1
2
j
1
2
+
1
2
:
2
3
1
2
:
1
2
=

1
2
(r
2
3r
1
)
2
+ (j
2
3j
1
)
2
+ (:
2
3:
1
)
2
=
1
2
(r
2
3r
1
)
2
+ (j
2
3j
1
)
2
+ (:
2
3:
1
)
2
=
1
2
|1
1
1
2
|
|Q1
2
| =

r
2
3
1
2
(r
1
+r
2
)
2
+
j
2
3
1
2
(j
1
+j
2
)
2
+
:
2
3
1
2
(:
1
+:
2
)
2
=

1
2
r
2
3
1
2
r
1
2
+
1
2
j
2
3
1
2
j
1
2
+
1
2
:
2
3
1
2
:
1
2
=

1
2
(r
2
3r
1
)
2
+ (j
2
3j
1
)
2
+ (:
2
3:
1
)
2
=
1
2
(r
2
3r
1
)
2
+ (j
2
3j
1
)
2
+ (:
2
3:
1
)
2
=
1
2
|1
1
1
2
|
So Q is indeed the midpoint of 1
1
1
2
.
SECTION 13.1 THREE-DIMENSIONAL COORDINATE SYSTEMS ET SECTION 12.1

107
(b) By part (a), the midpoints of sides 1, 1C and Care 1
1
3
1
2
. 1. 4
, 1
2
1.
1
2
. 5
and 1
3
5
2
.
3
2
. 4
. (Recall that a median

of a triangle is a line segment from a vertex to the midpoint of the opposite side.) Then the lengths of the medians are:
|1
2
| =

0
2
+
1
2
32
2
+ (5 33)
2
=

9
4
+ 4 =

25
4
=
5
2
|11
3
| =

5
2
+ 2
2
+
3
2
2
+ (4 35)
2
=

81
4
+
9
4
+ 1 =

94
4
=
1
2
I
94
|C1
1
| =

3
1
2
34
2
+ (1 31)
2
+ (4 35)
2
=

81
4
+ 1 =
1
2
I
85
21. (a) Since the sphere touches the rj-plane, its radius is the distance from its center, (2. 33. 6), to the rj-plane, namely 6.
Therefore v = 6 and an equation of the sphere is (r 32)
2
+ (j + 3)
2
+ (: 36)
2
= 6
2
= 36.
(b) The radius of this sphere is the distance from its center (2. 33. 6) to the j:-plane, which is 2. Therefore, an equation is
(r 32)
2
+ (j + 3)
2
+ (: 36)
2
= 4.
(c) Here the radius is the distance from the center (2. 33. 6) to the r:-plane, which is 3. Therefore, an equation is
(r 32)
2
+ (j + 3)
2
+ (: 36)
2
= 9.
23. The equation j = 34 represents a plane parallel to the r:-plane and 4 units to the left of it.
25. The inequality r 3 represents a half-space consisting of all points in front of the plane r = 3.
27. The inequality 0 $ : $ 6 represents all points on or between the horizontal planes : = 0 (the rj-plane) and : = 6.
29. The inequality r
2
+j
2
+:
2
$ 3 is equivalent to
r
2
+j
2
+:
2
$
I
3, so the region consists of those points whose distance
from the origin is at most
I
3. This is the set of all points on or inside the sphere with radius
I
3 and center (0. 0. 0).
31. Here r
2
+:
2
$ 9 or equivalently
I
r
2
+:
2
$ 3 which describes the set of all points in R
3
whose distance from the j-axis is
at most 3. Thus, the inequality represents the region consisting of all points on or inside a circular cylinder of radius 3 with
axis the j-axis.
33. This describes all points whose r-coordinate is between 0 and 5, that is, 0 < r < 5.
35. This describes a region all of whose points have a distance to the origin which is greater than v, but smaller than 1. So
inequalities describing the region are v <

r
2
+j
2
+:
2
< 1, or v
2
< r
2
+j
2
+:
2
< 1
2
.
37. (a) To nd the r- and j-coordinates of the point 1, we project it onto 1
2
and
project the resulting point Qonto the r- and j-axes. To nd the
:-coordinate, we project 1 onto either the r:-plane or the j:-plane
(using our knowledge of its r- or j-coordinate) and then project the
resulting point onto the :-axis. (Or, we could draw a line parallel to QO
from 1 to the :-axis.) The coordinates of 1 are (2. 1. 4).
(b) is the intersection of 1
1
and 1
2
, 1 is directly below the
j-intercept of 1
2
, and C is directly above the r-intercept of 1
2
.
39. We need to nd a set of points
1(r. j. :)

|1| = |11|
(r + 1)
2
+ (j 35)
2
+ (: 33)
2
=

(r 36)
2
+ (j 32)
2
+ (: + 2)
2
i
(r + 1)
2
+ (j 35) + (: 33)
2
= (r 36)
2
+ (j 32)
2
+ (: + 2)
2
i
r
2
+ 2r + 1 +j
2
310j + 25 +:
2
36: + 9 = r
2
312r + 36 +j
2
34j + 4 +:
2
+ 4: + 4 i 14r 36j 310: = 9.
Thus the set of points is a plane perpendicular to the line segment joining and 1 (since this plane must contain the
perpendicular bisector of the line segment 1).
13.2 Vectors ET 12.2
1. (a) The cost of a theater ticket is a scalar, because it has only magnitude.
(b) The current in a river is a vector, because it has both magnitude (the speed of the current) and direction at any given
location.
(c) If we assume that the initial path is linear, the initial ight path from Houston to Dallas is a vector, because it has both
magnitude (distance) and direction.
(d) The population of the world is a scalar, because it has only magnitude.
3. Vectors are equal when they share the same length and direction (but not necessarily location). Using the symmetry of the
parallelogram as a guide, we see that
3<
1 =
33<
1C,
33<
1 =
33<
C1,
33<
11 =
33<
11, and
3<
1 =
33<
C1.
5. (a) (b) (c) (d)
7. a = h32 32. 1 33i = h34. 32i 9. a = h2 3(31). 2 33i = h3. 31i
11. a = h2 30. 3 33. 31 31i = h2. 0. 32i 13. h31. 4i + h6. 32i = h31 + 6. 4 + (32)i = h5. 2i
SECTION 13.2 VECTORS ET SECTION 12.2

109
15. h0. 1. 2i + h0. 0. 33i = h0 + 0. 1 + 0. 2 + (33)i
= h0. 1. 31i
17. a +b = h5 + (33) . 312 + (36)i = h2. 318i
2a + 3b = h10. 324i + h39. 318i = h1. 342i
|a| =

5
2
+ (312)
2
=
I
169 = 13
|a 3b| = |h5 3(33). 312 3(36)i| = |h8. 36i| =

8
2
+ (36)
2
=
I
100 = 10
19. a +b = (i + 2 j 33 k) + (32 i 3j + 5 k) = 3i +j + 2k
2a + 3b = 2 (i + 2 j 33 k) + 3 (32 i 3j + 5 k) = 2 i + 4 j 36 k 36 i 33 j + 15 k =34 i +j + 9k
|a| =

1
2
+ 2
2
+ (33)
2
=
I
14
|a 3b| = |(i + 2 j 33 k) 3(32 i 3j + 5 k)| = |3 i + 3 j 38 k| =

3
2
+ 3
2
+ (38)
2
=
I
82
21. |33 i + 7 j| =

(33)
2
+ 7
2
=
I
58, so u =
1
I
58
(33 i + 7 j) = 3
3
I
58
i +
7
I
58
j.
23. The vector 8 i 3j + 4 k has length |8 i 3j + 4 k| =

8
2
+ (31)
2
+ 4
2
=
I
81 = 9, so by Equation 4 the unit vector with
the same direction is
1
9
(8 i 3j + 4 k) =
8
9
i 3
1
9
j +
4
9
k.
25. From the gure, we see that the r-component of v is
1
= |v| cos(3) = 4
1
2
= 2 and the j-component is
2
= |v| sin(3) = 4
I
3
2
= 2
I
3. Thus
v = h
1
.
2
i =

2. 2
I
3
.
27. The velocity vector v makes an angle of 40
with the horizontal and

has magnitude equal to the speed at which the football was thrown.
From the gure, we see that the horizontal component of v is
|v| cos 40
= 60 cos 40
E 45.96 ft/s and the vertical component is

|v| sin40
= 60 sin40
E 38.57 ft/s.
29. The given force vectors can be expressed in terms of their horizontal and vertical components as 3300 i and
200 cos 60
i + 200 sin60
j = 200
1
2
i + 200
I
3
2
j = 100 i + 100
I
3 j. The resultant force F is the sum of these two
vectors: F = (3300 + 100) i +
0 + 100
I
3
j = 3200 i + 100
I
3 j. Then we have
|F| E
(3200)
2
+
100
I
3
2
=
I
70,000 = 100
I
7 E 264.6 N. Let 0 be the angle F makes with the positive r-axis.
Then tan0 =
100
I
3
3200
= 3
I
3
2
and the terminal point of F lies in the second quadrant, so
0 = tan
31
3
I
3
2
+ 180
E 340.9
+ 180
= 139.1
.
31. With respect to the waters surface, the womans velocity is the vector sum of the velocity of the ship with respect to the water,
and the womans velocity with respect to the ship. If we let north be the positive j-direction, then
v = h0. 22i + h33. 0i = h33. 22i. The womans speed is |v| =
I
9 + 484 E 22.2 mih. The vector v makes an angle 0
with the east, where 0 = tan
31
22
33
E 98
. Therefore, the womans direction is about N(98 390)
W = N8
W.
33. Let T
1
and T
2
represent the tension vectors in each side of the
clothesline as shown in the gure. T
1
and T
2
have equal vertical
components and opposite horizontal components, so we can write
T
1
= 3oi +/ j and T
2
= o i +/ j [o. / 0]. By similar triangles,
/
o
=
0.08
4
i o = 50/. The force due to gravity
acting on the shirt has magnitude 0.8o E (0.8)(9.8) = 7.84 N, hence we have w = 37.84 j. The resultant T
1
+T
2
of the tensile forces counterbalances w, so T
1
+T
2
= 3w i (3o i +/ j) + (o i +/ j) = 7.84 j i
(350/ i +/ j) + (50/ i +/ j) = 2/ j = 7.84 j i / =
7.84
2
= 3.92 and o = 50/ = 196. Thus the tensions are
T
1
= 3oi +/ j = 3196 i + 3.92 j and T
2
= o i +/ j = 196 i + 3.92 j.
Alternatively, we can nd the value of 0 and proceed as in Example 7.
35. The slope of the tangent line to the graph of j = r
2
at the point (2. 4) is
dj
dr
i=2
= 2r
i=2
= 4
and a parallel vector is i + 4 j which has length |i + 4 j| =
I
1
2
+ 4
2
=
I
17, so unit vectors parallel to the tangent line are
1
I
17
(i + 4 j).
37. By the Triangle Law,
3<
1 +
33<
1C =
3<
C. Then
3<
1 +
33<
1C +
3<
C =
3<
C +
3<
C, but
3<
C +
3<
C =
3<
C +
3
3<
C
= 0.
So
3<
1 +
33<
1C +
3<
C = 0.
39. (a), (b) (c) From the sketch, we estimate that c E 1.3 and t E 1.6.
(d) c = c a +t b C 7 = 3c + 2t and 1 = 2c 3t.
Solving these equations gives c =
9
7
and t =
11
7
.
41. |r 3r
0
| is the distance between the points (r. j. :) and (r
0
. j
0
. :
0
), so the set of points is a sphere with radius 1 and
center (r
0
. j
0
. :
0
).
Alternate method: |r 3r
0
| = 1 C

(r 3r
0
)
2
+ (j 3j
0
)
2
+ (: 3:
0
)
2
= 1 C
(r 3r
0
)
2
+ (j 3j
0
)
2
+ (: 3:
0
)
2
= 1, which is the equation of a sphere with radius 1 and center (r
0
. j
0
. :
0
).
SECTION 13.3 THE DOT PRODUCT ET SECTION 12.3

111
43. a + (b +c) = ho
1
. o
2
i + (h/
1
. /
2
i + hc
1
. c
2
i) = ho
1
. o
2
i + h/
1
+c
1
. /
2
+c
2
i
= ho
1
+/
1
+c
1
. o
2
+/
2
+c
2
i = h(o
1
+/
1
) +c
1
. (o
2
+/
2
) +c
2
i
= ho
1
+/
1
. o
2
+/
2
i + hc
1
. c
2
i = (ho
1
. o
2
i + h/
1
. /
2
i) + hc
1
. c
2
i
= (a +b) +c
45. Consider triangle 1C, where 1 and 1 are the midpoints of 1 and 1C. We know that
3<
1 +
33<
1C =
3<
C (1) and
33<
11 +
33<
11 =
33<
11 (2). However,
33<
11 =
1
2
3<
1, and
33<
11 =
1
2
33<
1C. Substituting these expressions for
33<
11 and
33<
11 into
(2) gives
1
2
3<
1 +
1
2
33<
1C =
33<
11. Comparing this with (1) gives
33<
11 =
1
2
3<
C. Therefore
3<
C and
33<
11 are parallel and
33<
11
=
1
2
3<
C
.
13.3 The Dot Product ET 12.3
1. (a) a b is a scalar, and the dot product is dened only for vectors, so (a b) c has no meaning.
(b) (a b) c is a scalar multiple of a vector, so it does have meaning.
(c) Both |a| and b c are scalars, so |a| (b c) is an ordinary product of real numbers, and has meaning.
(d) Both a and b +c are vectors, so the dot product a (b +c) has meaning.
(e) a b is a scalar, but c is a vector, and so the two quantities cannot be added and a b +c has no meaning.
(f ) |a| is a scalar, and the dot product is dened only for vectors, so |a| (b +c) has no meaning.
3. a b =

32.
1
3
h35. 12i = (32)(35) +
1
3
(12) = 10 + 4 = 14
5. a b =

4. 1.
1
4
h6. 33. 38i = (4)(6) + (1)(33) +
1
4
(38) = 19
7. a b = (i 32 j + 3 k) (5 i + 9 k) = (1)(5) + (32)(0) + (3)(9) = 32
9. a b = |a| |b| cos 0 = (6)(5) cos
2r
3
= 30
3
1
2
= 315
11. u. v. and w are all unit vectors, so the triangle is an equilateral triangle. Thus the angle between u and v is 60
and
u v = |u| |v| cos 60
= (1)(1)
1
2
=
1
2
. If w is moved so it has the same initial point as u, we can see that the angle
between them is 120
and we have u w = |u| |w| cos 120
= (1)(1)
3
1
2
= 3
1
2
.
13. (a) i j = h1. 0. 0i h0. 1. 0i = (1)(0) + (0)(1) + (0)(0) = 0. Similarly, j k = (0)(0) + (1)(0) + (0)(1) = 0 and
k i = (0)(1) + (0)(0) + (1)(0) = 0.
Another method: Because i, j, and k are mutually perpendicular, the cosine factor in each dot product (see Theorem 3)
is cos
r
2
= 0.
(b) By Property 1 of the dot product, i i = |i|
2
= 1
2
= 1 since i is a unit vector. Similarly, j j = |j|
2
= 1 and
k k = |k|
2
= 1.
15. |a| =

(38)
2
+ 6
2
= 10, |b| =

I
7
2
+ 3
2
= 4, and a b = (38)
I
7
+ (6)(3) = 18 38
I
7. From Corollary 6,
we have cos 0 =
a b
|a| |b|
=
18 38
I
7
10 4
=
9 34
I
7
20
. So the angle between a and b is 0 = cos
31
9 34
I
7
20
E 95
.
17. |a| =

3
2
+ (31)
2
+ 5
2
=
I
35, |b| =

(32)
2
+ 4
2
+ 3
2
=
I
29, and a b = (3)(32) + (31)(4) + (5)(3) = 5. Then
cos 0 =
a b
|a| |b|
=
5
I
35
I
29
=
5
I
1015
and the angle between a and b is 0 = cos
31
5
I
1015
E 81
.
19. |a| =
I
0
2
+ 1
2
+ 1
2
=
I
2, |b| =

1
2
+ 2
2
+ (33)
2
=
I
14, and a b = (0)(1) + (1)(2) + (1)(33) = 31.
Then cos 0 =
a b
|a| |b|
=
31
I
2
I
14
=
31
2
I
7
and 0 = cos
31
3
1
2
I
7
E 101
.
21. Let o, /, and c be the angles at vertices , 1, and C respectively.
Then o is the angle between vectors
3<
1 and
3<
C, / is the angle
between vectors
3<
1 and
33<
1C, and c is the angle between vectors
3<
C and
33<
C1.
Thus cos o =
3<
1
3<
C
3<
1
3<
C
=
h2. 6i h32. 4i
I
2
2
+ 6
2
(32)
2
+ 4
2
=
1
I
40
I
20
(34 + 24) =
20
I
800
=
I
2
2
and
o = cos
31
I
2
2
= 45
. Similarly, cos / =
3<
1
33<
1C
3<
1
33<
1C
=
h32. 36i h34. 32i
I
4 + 36
I
16 + 4
=
1
I
40
I
20
(8 + 12) =
20
I
800
=
I
2
2
so / = cos
31
I
2
2
= 45
and c = 180
3(45
+ 45
) = 90
.
Alternate solution: Apply the Law of Cosines three times as follows: cos o =
33<
1C
2
3
3<
1
2
3
3<
C
2
2
3<
1
3<
C
,
cos / =
3<
C
2
3
3<
1
2
3
33<
1C
2
2
3<
1
33<
1C
, and cos c =
3<
1
2
3
3<
C
2
3
33<
1C
2
2
3<
C
33<
1C
.
23. (a) a b = (35)(6) + (3)(38) + (7)(2) = 340 6= 0, so a and b are not orthogonal. Also, since a is not a scalar multiple
of b, a and b are not parallel.
(b) a b = (4)(33) + (6)(2) = 0, so a and b are orthogonal (and not parallel).
(c) a b = (31)(3) + (2)(4) + (5)(31) = 0, so a and b are orthogonal (and not parallel).
(d) Because a = 3
2
3
b, a and b are parallel.
25.
33<
Q1 = h31. 33. 2i,
3<
Q1 = h4. 32. 31i, and
33<
Q1
3<
Q1 = 34 + 6 32 = 0. Thus
33<
Q1 and
3<
Q1 are orthogonal, so the angle of
the triangle at vertex Qis a right angle.
27. Let a = o
1
i +o
2
j +o
3
k be a vector orthogonal to both i +j and i +k. Then a (i +j) = 0 C o
1
+o
2
= 0 and
a (i +k) = 0 C o
1
+o
3
= 0, so o
1
= 3o
2
= 3o
3
. Furthermore a is to be a unit vector, so 1 = o
2
1
+o
2
2
+o
2
3
= 3o
2
1
implies o
1
=
1
I
3
. Thus a =
1
I
3
i 3
1
I
3
j 3
1
I
3
k and a = 3
1
I
3
i +
1
I
3
j +
1
I
3
k are two such unit vectors.
SECTION 13.3 THE DOT PRODUCT ET SECTION 12.3

113
29. Since |h3. 4. 5i| =
I
9 + 16 + 25 =
I
50 = 5
I
2, using Equations 8 and 9 we have cos c =
3
5
I
2
, cos o =
4
5
I
2
, and
cos ~ =
5
5
I
2
=
1
I
2
. The direction angles are given by c = cos
31
3
5
I
2
E 65
, o = cos
31
4
5
I
2
E 56
, and
~ = cos
31
1
I
2
= 45
.
31. Since |2 i + 3 j 36 k| =
I
4 + 9 + 36 =
I
49 = 7, Equations 8 and 9 give cos c =
2
7
, cos o =
3
7
, and cos ~ =
36
7
, while
c = cos
31
2
7
E 73
, o = cos
31
3
7
E 65
, and ~ = cos
31
3
6
7
E 149
.
33. |hc. c. ci| =
I
c
2
+c
2
+c
2
=
I
3c [since c 0], so cos c = cos o = cos ~ =
c
I
3c
=
1
I
3
and
c = o = ~ = cos
31
1
I
3
E 55
.
35. |a| =

3
2
+ (34)
2
= 5. The scalar projection of b onto a is comp
a
b =
a b
|a|
=
3 5 + (34) 0
5
= 3 and the vector
projection of b onto a is proj
a
b =

a b
|a|
a
|a|
= 3
1
5
h3. 34i =

9
5
. 3
12
5
.
37. |a| =
I
9 + 36 + 4 = 7 so the scalar projection of b onto a is comp
a
b =
a b
|a|
=
1
7
(3 + 12 36) =
9
7
. The vector
a
b =
9
7
a
|a|
=
9
7

1
7
h3. 6. 32i =
9
49
h3. 6. 32i =

27
49
.
54
49
. 3
18
49
.
39. |a| =
I
4 + 1 + 16 =
I
21 so the scalar projection of b onto a is comp
a
b =
a b
|a|
=
0 31 + 2
I
21
=
1
I
21
while the vector
a
b =
1
I
21
a
|a|
=
1
I
21

2 i 3j + 4 k
I
21
=
1
21
(2 i 3j + 4 k) =
2
21
i 3
1
21
j +
4
21
k.
41. (orth
a
b) a = (b 3proj
a
b) a = b a 3(proj
a
b) a = b a 3
a b
|a|
2
a a = b a 3
a b
|a|
2
|a|
2
= b a 3a b = 0.
So they are orthogonal by (7).
43. comp
a
b =
a b
|a|
= 2 C a b = 2 |a| = 2
I
10. If b = h/
1
. /
2
. /
3
i, then we need 3/
1
+ 0/
2
31/
3
= 2
I
10.
One possible solution is obtained by taking /
1
= 0, /
2
= 0, /
3
= 32
I
10. In general, b =

c. t. 3c 32
I
10
, c, t M R.
45. The displacement vector is D = (6 30) i + (12 310) j + (20 38) k = 6 i + 2 j + 12 k so by Equation 12 the work done is
W = F D = (8 i 36 j + 9 k) (6 i + 2 j + 12 k) = 48 312 + 108 = 144 joules.
47. Here |D| = 80 ft, |F| = 30 lb, and 0 = 40
. Thus
W = F D = |F| |D| cos 0 = (30)(80) cos 40
= 2400 cos 40
E 1839 ft-lb.
49. First note that n = ho. /i is perpendicular to the line, because if Q
1
= (o
1
. /
1
) and Q
2
= (o
2
. /
2
) lie on the line, then
n
333<
Q
1
Q
2
= oo
2
3oo
1
+//
2
3//
1
= 0, since oo
2
+//
2
= 3c = oo
1
+//
1
from the equation of the line.
Let 1
2
= (r
2
. j
2
) lie on the line. Then the distance from 1
1
to the line is the absolute value of the scalar projection
of
333<
1
1
1
2
onto n. comp
n
333<
1
1
1
2
=
|n hr
2
3r
1
. j
2
3j
1
i|
|n|
=
|or
2
3or
1
+/j
2
3/j
1
|
I
o
2
+/
2
=
|or
1
+/j
1
+c|
I
o
2
+/
2
since or
2
+/j
2
= 3c. The required distance is
|3 32 +34 3 + 5|
I
3
2
+ 4
2
=
13
5
.
51. For convenience, consider the unit cube positioned so that its back left corner is at the origin, and its edges lie along the
coordinate axes. The diagonal of the cube that begins at the origin and ends at (1. 1. 1) has vector representation h1. 1. 1i.
The angle 0 between this vector and the vector of the edge which also begins at the origin and runs along the r-axis [that is,
h1. 0. 0i] is given by cos 0 =
h1. 1. 1i h1. 0. 0i
|h1. 1. 1i| |h1. 0. 0i|
=
1
I
3
i 0 = cos
31
1
I
3
E 55
.
53. Consider the HCH combination consisting of the sole carbon atom and the two hydrogen atoms that are at (1. 0. 0) and
(0. 1. 0) (or any HCH combination, for that matter). Vector representations of the line segments emanating from the
carbon atom and extending to these two hydrogen atoms are
1 3
1
2
. 0 3
1
2
. 0 3
1
2
=

1
2
. 3
1
2
. 3
1
2
and
0 3
1
2
. 1 3
1
2
. 0 3
1
2
=

3
1
2
.
1
2
. 3
1
2
. The bond angle, 0, is therefore given by

cos 0 =
1
2
. 3
1
2
. 3
1
2
3
1
2
.
1
2
. 3
1
2
1
2
. 3
1
2
. 3
1
2
3
1
2
.
1
2
. 3
1
2
=
3
1
4
3
1
4
+
1
4
3
4
3
4
= 3
1
3
i 0 = cos
31
3
1
3
E 109.5
.
55. Let a = ho
1
. o
2
. o
3
i and = h/
1
. /
2
. /
3
i.
Property 2: a b = ho
1
. o
2
. o
3
i h/
1
. /
2
. /
3
i = o
1
/
1
+o
2
/
2
+o
3
/
3
= /
1
o
1
+/
2
o
2
+/
3
o
3
= h/
1
. /
2
. /
3
i ho
1
. o
2
. o
3
i = b a
Property 4: (c a) b = hco
1
. co
2
. co
3
i h/
1
. /
2
. /
3
i = (co
1
)/
1
+ (co
2
)/
2
+ (co
3
)/
3
= c (o
1
/
1
+o
2
/
2
+o
3
/
3
) = c (a b) = o
1
(c/
1
) +o
2
(c/
2
) +o
3
(c/
3
)
= ho
1
. o
2
. o
3
i hc/
1
. c/
2
. c/
3
i = a (c b)
Property 5: 0 a = h0. 0. 0i ho
1
. o
2
. o
3
i = (0)(o
1
) + (0)(o
2
) + (0)(o
3
) = 0
57. |a b| =

|a| |b| cos 0
= |a| |b| |cos 0|. Since |cos 0| $ 1, |a b| = |a| |b| |cos 0| $ |a| |b|.
Note: We have equality in the case of cos 0 = 1, so 0 = 0 or 0 = , thus equality when a and b are parallel.
59. (a) The Parallelogram Law states that the sum of the squares of the
lengths of the diagonals of a parallelogram equals the sum of the
squares of its (four) sides.
(b) |a +b|
2
= (a +b) (a +b) = |a|
2
+ 2(a b) + |b|
2
and |a 3b|
2
= (a 3b) (a 3b) = |a|
2
32(a b) + |b|
2
.
Adding these two equations gives |a +b|
2
+ |a 3b|
2
= 2 |a|
2
+ 2 |b|
2
.
SECTION 13.4 THE CROSS PRODUCT ET SECTION 12.4

115
13.4 The Cross Product ET 12.4
1. a b =
i j k
6 0 32
0 8 0
0 32
8 0
i 3
6 32
0 0
j +
6 0
0 8
k
= [0 3(316)] i 3(0 30) j + (48 30) k = 16 i + 48 k
Now (a b) a = h16. 0. 48i h6. 0. 32i = 96 + 0 396 = 0 and (a b) b = h16. 0. 48i h0. 8. 0i = 0 + 0 + 0 = 0, so
a b is orthogonal to both a and b.
3. a b =
i j k
1 3 32
31 0 5
3 32
0 5
i 3
1 32
31 5
j +
1 3
31 0
k
= (15 30) i 3(5 32) j + [0 3(33)] k = 15 i 33 j + 3 k
Since (a b) a = (15 i 33 j + 3 k) (i + 3 j 32 k) = 15 39 36 = 0, a b is orthogonal to a.
Since (a b) b = (15 i 33 j + 3 k) (3i + 5 k) = 315 + 0 + 15 = 0, a b is orthogonal to b.
5. a b =
i j k
1 31 31
1
2
1
1
2
31 31
1
1
2
i 3
1 31
1
2
1
2
j +
1 31
1
2
1
k
=

3
1
2
3(31)
i 3
1
2
3(3
1
2
)
j +
1 3(3
1
2
)
k =
1
2
i 3j +
3
2
k
Now (a b) a =

1
2
i 3j +
3
2
k
(i 3 j 3 k) =
1
2
+ 1 3
3
2
= 0 and
(a b) b =

1
2
i 3j +
3
2
k
1
2
i +j +
1
2
k
=
1
4
31 +
3
4
= 0, so a b is orthogonal to both a and b.
7. a b =
i j k
t t
2
t
3
1 2t 3t
2
t
2
t
3
2t 3t
2
i 3
t t
3
1 3t
2
j +
t t
2
1 2t
k
= (3t
4
32t
4
) i 3(3t
3
3t
3
) j + (2t
2
3t
2
) k = t
4
i 32t
3
j +t
2
k
Since (a b) a =
t
4
. 32t
3
. t
2
t. t
2
. t
3
= t
5
32t
5
+t
5
= 0, a b is orthogonal to a.
Since (a b) b =
t
4
. 32t
3
. t
2
1. 2t. 3t
2
= t
4
34t
4
+ 3t
4
= 0, a b is orthogonal to b.
9. According to the discussion preceding Theorem 8, i j = k, so (i j) k = k k = 0 [by Example 2].
11. (j 3k) (k 3i) = (j 3k) k + (j 3k) (3i) by Property 3 of Theorem 8
= j k + (3k) k +j (3i) + (3k) (3i) by Property 4 of Theorem 8
= (j k) + (31)(k k) + (31)(j i) + (31)
2
(k i) by Property 2 of Theorem 8
= i + (31) 0 + (31)(3k) +j = i +j +k by Example 2 and the
discussion preceding Theorem 8
13. (a) Since b c is a vector, the dot product a (b c) is meaningful and is a scalar.
(b) b c is a scalar, so a (b c) is meaningless, as the cross product is dened only for two vectors.
(c) Since b c is a vector, the cross product a (b c) is meaningful and results in another vector.
(d) a b is a scalar, so the cross product (a b) c is meaningless.
(e) Since (a b) and (c d) are both scalars, the cross product (a b) (c d) is meaningless.
(f ) a b and c d are both vectors, so the dot product (a b) (c d) is meaningful and is a scalar.
15. If we sketch u and v starting from the same initial point, we see that
the angle between them is 30
. Using Theorem 6, we have

|u v| = |u| |v| sin30
= (6)(8)
1
2
= 24.
By the right-hand rule, u v is directed into the page.
17. a b =
i j k
1 2 1
0 1 3
2 1
1 3
i 3
1 1
0 3
j +
1 2
0 1
k = (6 31) i 3(3 30) j + (1 30) k = 5 i 33 j +k

b a =
i j k
0 1 3
1 2 1
1 3
2 1
i 3
0 3
1 1
j +
0 1
1 2
k = (1 36) i 3(0 33) j + (0 31) k = 35 i + 3 j 3k

Notice a b = 3b a here, as we know is always true by Theorem 8.
19. We know that the cross product of two vectors is orthogonal to both. So we calculate
h1. 31. 1i h0. 4. 4i =
i j k
1 31 1
0 4 4
31 1
4 4
i 3
1 1
0 4
j +
1 31
0 4
k = 38 i 34 j + 4 k.
So two unit vectors orthogonal to both are
h38. 34. 4i
I
64 + 16 + 16
=
h38. 34. 4i
4
I
6
, that is,
3
2
I
6
. 3
1
I
6
.
1
I
6
and
2
I
6
.
1
I
6
. 3
1
I
6
.
21. Let a = ho
1
. o
2
. o
3
i. Then
0 a =
i j k
0 0 0
o
1
o
2
o
3
0 0
o
2
o
3
i 3
0 0
o
1
o
3
j +
0 0
o
1
o
2
k = 0,
a 0 =
i j k
o
1
o
2
o
3
0 0 0
o
2
o
3
0 0
i 3
o
1
o
3
0 0
j +
o
1
o
2
0 0
k = 0.
23. a b = ho
2
/
3
3o
3
/
2
. o
3
/
1
3o
1
/
3
. o
1
/
2
3o
2
/
1
i
= h(31)(/
2
o
3
3/
3
o
2
) . (31)(/
3
o
1
3/
1
o
3
) . (31)(/
1
o
2
3/
2
o
1
)i
= 3h/
2
o
3
3/
3
o
2
. /
3
o
1
3/
1
o
3
. /
1
o
2
3/
2
o
1
i = 3b a
SECTION 13.4 THE CROSS PRODUCT ET SECTION 12.4

117
25. a (b +c) = a h/
1
+c
1
. /
2
+c
2
. /
3
+c
3
i
= ho
2
(/
3
+c
3
) 3o
3
(/
2
+c
2
) , o
3
(/
1
+c
1
) 3o
1
(/
3
+c
3
) , o
1
(/
2
+c
2
) 3o
2
(/
1
+c
1
)i
= ho
2
/
3
+o
2
c
3
3o
3
/
2
3o
3
c
2
, o
3
/
1
+o
3
c
1
3o
1
/
3
3o
1
c
3
, o
1
/
2
+o
1
c
2
3o
2
/
1
3o
2
c
1
i
= h(o
2
/
3
3o
3
/
2
) + (o
2
c
3
3o
3
c
2
) , (o
3
/
1
3o
1
/
3
) + (o
3
c
1
3o
1
c
3
) , (o
1
/
2
3o
2
/
1
) + (o
1
c
2
3o
2
c
1
)i
= ho
2
/
3
3o
3
/
2
. o
3
/
1
3o
1
/
3
. o
1
/
2
3o
2
/
1
i + ho
2
c
3
3o
3
c
2
. o
3
c
1
3o
1
c
3
. o
1
c
2
3o
2
c
1
i
= (a b) + (a c)
27. By plotting the vertices, we can see that the parallelogram is determined by the
vectors
3<
1 = h2. 3i and
33<
1 = h4. 32i. We know that the area of the parallelogram
determined by two vectors is equal to the length of the cross product of these vectors.
In order to compute the cross product, we consider the vector
3<
1 as the three-
dimensional vector h2. 3. 0i (and similarly for
33<
1), and then the area of
parallelogram1C1 is
3<
1
33<
1
i j k
2 3 0
4 32 0
= |(0) i 3(0) j + (34 312) k| = |316 k| = 16

29. (a) Because the plane through 1, Q, and 1 contains the vectors
33<
1Q and
3<
11, a vector orthogonal to both of these vectors
(such as their cross product) is also orthogonal to the plane. Here
33<
1Q = h31. 2. 0i and
3<
11 = h31. 0. 3i, so
33<
1Q
3<
11 = h(2)(3) 3(0)(0). (0)(31) 3(31)(3). (31)(0) 3(2)(31)i = h6. 3. 2i
Therefore, h6. 3. 2i (or any scalar multiple thereof) is orthogonal to the plane through 1, Q, and 1.
(b) Note that the area of the triangle determined by 1, Q, and 1 is equal to half of the area of the parallelogram determined by
the three points. From part (a), the area of the parallelogram is
33<
1Q
3<
11
= |h6. 3. 2i| =
I
36 + 9 + 4 = 7. so the area
of the triangle is
1
2
(7) =
7
2
.
31. (a)
33<
1Q = h4. 3. 32i and
3<
11 = h5. 5. 1i, so a vector orthogonal to the plane through 1, Q, and 1 is
33<
1Q
3<
11 = h(3)(1) 3(32)(5). (32)(5) 3(4)(1). (4)(5) 3(3)(5)i = h13. 314. 5i [or any scalar mutiple thereof ].
(b) The area of the parallelogram determined by
33<
1Q and
3<
11 is
33<
1Q
3<
11
= |h13. 314. 5i| =

13
2
+ (314)
2
+ 5
2
=
I
390, so the area of triangle 1Q1 is
1
2
I
390.
33. We know that the volume of the parallelepiped determined by a, b, and c is the magnitude of their scalar triple product, which
is a (b c) =
6 3 31
0 1 2
4 32 5
= 6
1 2
32 5
3 3
0 2
4 5
+ (31)
0 1
4 32
= 6(5 + 4) 33(0 38) 3(0 34) = 82.

Thus the volume of the parallelepiped is 82 cubic units.
35. a =
33<
1Q = h2. 1. 1i, b =
3<
11 = h1. 31. 2i, and c =
3<
1o = h0. 32. 3i.
a (b c) =
2 1 1
1 31 2
0 32 3
= 2
31 2
32 3
3 1
1 2
0 3
+ 1
1 31
0 32
= 2 33 32 = 33,
so the volume of the parallelepiped is 3 cubic units.
37. u (v w) =
1 5 32
3 31 0
5 9 34
= 1
31 0
9 34
35
3 0
5 34
+ (32)
3 31
5 9
= 4 + 60 364 = 0, which says that the volume

of the parallelepiped determined by u, v and w is 0, and thus these three vectors are coplanar.
39. The magnitude of the torque is || = |r F| = |r| |F| sin0 = (0.18 m)(60 N) sin(70 + 10)
= 10.8 sin80
E 10.6 Nm.
41. Using the notation of the text, r = h0. 0.3. 0i and F has direction h0. 3. 34i. The angle 0 between them can be determined by
cos 0 =
h0. 0.3. 0i h0. 3. 34i
|h0. 0.3. 0i| |h0. 3. 34i|
i cos 0 =
0.9
(0.3)(5)
i cos 0 = 0.6 i 0 E 53.1
. Then || = |r| |F| sin0 i

100 = 0.3 |F| sin53.1
i |F| E 417 N.
43. (a) The distance between a point and a line is the length of the perpendicular
from the point to the line, here
3<
1o
= d. But referring to triangle 1Qo,

d =

3<
1o
33<
Q1
sin0 = |b| sin0. But 0 is the angle between

33<
Q1 = b
and
3<
Q1 = a. Thus by Theorem 6, sin0 =
|a b|
|a| |b|
and so d = |b| sin0 =
|b| |a b|
|a| |b|
=
|a b|
|a|
.
(b) a =
3<
Q1 = h31. 32. 31i and b =
33<
Q1 = h1. 35. 37i. Then
a b = h(32)(37) 3(31)(35). (31)(1) 3(31)(37). (31)(35) 3(32)(1)i = h9. 38. 7i.
Thus the distance is d =
|a b|
|a|
=
1
I
6
I
81 + 64 + 49 =

194
6
=

97
3
.
45. (a 3b) (a +b) = (a 3b) a + (a 3b) b by Property 3 of Theorem 8
= a a + (3b) a +a b + (3b) b by Property 4 of Theorem 8
= (a a) 3(b a) + (a b) 3(b b) by Property 2 of Theorem 8 (with c = 31)
= 0 3(b a) + (a b) 30 by Example 2
= (a b) + (a b) by Property 1 of Theorem 8
= 2(a b)
47. a (b c) +b (c a) +c (a b)
= [(a c)b 3(a b)c] + [(b a)c 3(b c)a] + [(c b)a 3(c a)b] by Exercise 46
= (a c)b 3(a b)c + (a b)c 3(b c)a + (b c)a 3(a c)b = 0
SECTION 13.5 EQUATIONS OF LINES AND PLANES ET SECTION 12.5

119
49. (a) No. If a b = a c, then a (b 3c) = 0, so a is perpendicular to b 3c, which can happen if b 6= c. For example,
let a = h1. 1. 1i, b = h1. 0. 0i and c = h0. 1. 0i.
(b) No. If a b = a c then a (b 3c) = 0, which implies that a is parallel to b 3c, which of course can happen
if b 6= c.
(c) Yes. Since a c = a b, a is perpendicular to b 3c, by part (a). From part (b), a is also parallel to b 3c. Thus since
a 6= 0 but is both parallel and perpendicular to b 3c, we have b 3c = 0, so b = c.
13.5 Equations of Lines and Planes ET 12.5
1. (a) True; each of the rst two lines has a direction vector parallel to the direction vector of the third line, so these vectors are
each scalar multiples of the third direction vector. Then the rst two direction vectors are also scalar multiples of each
other, so these vectors, and hence the two lines, are parallel.
(b) False; for example, the r- and j-axes are both perpendicular to the :-axis, yet the r- and j-axes are not parallel.
(c) True; each of the rst two planes has a normal vector parallel to the normal vector of the third plane, so these two normal
vectors are parallel to each other and the planes are parallel.
(d) False; for example, the rj- and j:-planes are not parallel, yet they are both perpendicular to the r:-plane.
(e) False; the r- and j-axes are not parallel, yet they are both parallel to the plane : = 1.
(f ) True; if each line is perpendicular to a plane, then the lines direction vectors are both parallel to a normal vector for the
plane. Thus, the direction vectors are parallel to each other and the lines are parallel.
(g) False; the planes j = 1 and : = 1 are not parallel, yet they are both parallel to the r-axis.
(h) True; if each plane is perpendicular to a line, then any normal vector for each plane is parallel to a direction vector for the
line. Thus, the normal vectors are parallel to each other and the planes are parallel.
(i) True; see Figure 9 and the accompanying discussion.
( j) False; they can be skew, as in Example 3.
(k) True. Consider any normal vector for the plane and any direction vector for the line. If the normal vector is perpendicular
to the direction vector, the line and plane are parallel. Otherwise, the vectors meet at an angle 0, 0
$ 0 < 90
, and the
line will intersect the plane at an angle 90
30.
3. For this line, we have r
0
= 2 i + 2.4 j + 3.5 k and v = 3 i + 2 j 3 k, so a vector equation is
r = r
0
+t v = (2 i +2.4 j +3.5 k) +t(3 i +2 j 3k) = (2 +3t) i +(2.4 +2t) j +(3.5 3t) k and parametric equations are
r = 2 + 3t, j = 2.4 + 2t, : = 3.5 3t.
5. A line perpendicular to the given plane has the same direction as a normal vector to the plane, such as
n = h1. 3. 1i. So r
0
= i + 6 k, and we can take v = i + 3 j +k. Then a vector equation is
r = (i + 6 k) +t(i + 3 j +k) = (1 +t) i + 3t j + (6 +t) k, and parametric equations are r = 1 +t, j = 3t, : = 6 +t.
7. The vector v = h34 31. 3 33. 0 32i = h35. 0. 32i is parallel to the line. Letting 1
0
= (1. 3. 2), parametric equations are
r = 1 35t, j = 3 + 0t = 3, : = 2 32t, while symmetric equations are
r 31
35
=
: 32
32
. j = 3. Notice here that the
direction number / = 0, so rather than writing
j 33
0
in the symmetric equation we must write the equation j = 3 separately.
9. v =

2 30. 1 3
1
2
. 33 31
=

2.
1
2
. 34
, and letting 1
0
= (2. 1. 33), parametric equations are r = 2 + 2t, j = 1 +
1
2
t,
: = 33 34t, while symmetric equations are
r 32
2
=
j 31
12
=
: + 3
34
or
r 32
2
= 2j 32 =
: + 3
34
.
11. The line has direction v = h1. 2. 1i. Letting 1
0
= (1. 31. 1), parametric equations are r = 1 +t, j = 31 + 2t, : = 1 +t
and symmetric equations are r 31 =
j + 1
2
= : 31.
13. Direction vectors of the lines are v
1
= h32 3(34). 0 3(36). 33 31i = h2. 6. 34i and
v
2
= h5 310. 3 318. 14 34i = h35. 315. 10i, and since v
2
= 3
5
2
v
1
, the direction vectors and thus the lines are parallel.
15. (a) The line passes through the point (1. 35. 6) and a direction vector for the line is h31. 2. 33i, so symmetric equations for
the line are
r 31
31
=
j + 5
2
=
: 36
33
.
(b) The line intersects the rj-plane when : = 0, so we need
r 31
31
=
j + 5
2
=
0 36
33
or
r 31
31
= 2 i r = 31,
j + 5
2
= 2 i j = 31. Thus the point of intersection with the rj-plane is (31. 31. 0). Similarly for the j:-plane,
we need r = 0 i 1 =
j + 5
2
=
: 36
33
i j = 33, : = 3. Thus the line intersects the j:-plane at (0. 33. 3). For
the r:-plane, we need j = 0 i
r 31
31
=
5
2
=
: 36
33
i r = 3
3
2
, : = 3
3
2
. So the line intersects the r:-plane
at
3
3
2
. 0. 3
3
2
.
17. From Equation 4, the line segment from r
0
= 2 i 3 j + 4 k to r
1
= 4 i + 6 j + k is
r(t) = (1 3t) r
0
+t r
1
= (1 3t)(2 i 3j + 4 k) +t(4 i + 6 j +k) = (2 i 3j + 4 k) +t(2 i + 7 j 33 k), 0 $ t $ 1.
19. Since the direction vectors are v
1
= h36. 9. 33i and v
2
= h2. 33. 1i, we have v
1
= 33v
2
so the lines are parallel.
21. Since the direction vectors h1. 2. 3i and h34. 33. 2i are not scalar multiples of each other, the lines are not parallel, so we
check to see if the lines intersect. The parametric equations of the lines are 1
1
: r = t, j = 1 + 2t, : = 2 + 3t and 1
2
:
r = 3 34c, j = 2 33c, : = 1 + 2c. For the lines to intersect, we must be able to nd one value of t and one value of c that
produce the same point from the respective parametric equations. Thus we need to satisfy the following three equations:
t = 3 34c, 1 + 2t = 2 33c, 2 + 3t = 1 + 2c. Solving the rst two equations we get t = 31, c = 1 and checking, we see
that these values dont satisfy the third equation. Thus the lines arent parallel and dont intersect, so they must be skew lines.
23. Since the plane is perpendicular to the vector h32. 1. 5i, we can take h32. 1. 5i as a normal vector to the plane.
(6. 3. 2) is a point on the plane, so setting o = 32, / = 1, c = 5 and r
0
= 6, j
0
= 3, :
0
= 2 in Equation 7 gives
32(r 36) + 1(j 33) + 5(: 32) = 0 or 32r +j + 5: = 1 to be an equation of the plane.
25. i +j 3k = h1. 1. 31i is a normal vector to the plane and (1. 31. 1) is a point on the plane, so setting o = 1, / = 1, c = 31,
r
0
= 1, j
0
= 31, :
0
= 1 in Equation 7 gives 1 (r 31) + 1[j 3(31)] 31(: 31) = 0 or r +j 3: = 31 to be an equation
of the plane.
27. Since the two planes are parallel, they will have the same normal vectors. So we can take n = h2. 31. 3i, and an equation of
the plane is 2(r 30) 31(j 30) + 3(: 30) = 0 or 2r 3j + 3: = 0.

121
29. Since the two planes are parallel, they will have the same normal vectors. So we can take n = h3. 0. 37i, and an equation of
the plane is 3(r 34) + 0[j 3(32)] 37(: 33) = 0 or 3r 37: = 39.
31. Here the vectors a = h1 30. 0 31. 1 31i = h1. 31. 0i and b = h1 30. 1 31. 0 31i = h1. 0. 31i lie in the plane, so
a b is a normal vector to the plane. Thus, we can take n = a b = h1 30. 0 + 1. 0 + 1i = h1. 1. 1i. If 1
0
is the point
(0. 1. 1), an equation of the plane is 1(r 30) + 1(j 31) + 1(: 31) = 0 or r +j +: = 2.
33. Here the vectors a = h8 33. 2 3(31). 4 32i = h5. 3. 2i and b = h31 33. 32 3(31). 33 32i = h34. 31. 35i lie in
the plane, so a normal vector to the plane is n = a b = h315 + 2. 38 + 25. 35 + 12i = h313. 17. 7i and an equation of
the plane is 313(r 33) + 17[j 3(31)] + 7(: 32) = 0 or 313r + 17j + 7: = 342.
35. If we rst nd two nonparallel vectors in the plane, their cross product will be a normal vector to the plane. Since the given
line lies in the plane, its direction vector a = h32. 5. 4i is one vector in the plane. We can verify that the given point (6. 0. 32)
does not lie on this line, so to nd another nonparallel vector b which lies in the plane, we can pick any point on the line and
nd a vector connecting the points. If we put t = 0, we see that (4. 3. 7) is on the line, so
b = h6 34. 0 33. 32 37i = h2. 33. 39i and n = a b = h345 + 12. 8 318. 6 310i = h333. 310. 34i. Thus, an
equation of the plane is 333(r 36) 310(j 30) 34[: 3(32)] = 0 or 33r + 10j + 4: = 190.
37. A direction vector for the line of intersection is a = n
1
n
2
= h1. 1. 31i h2. 31. 3i = h2. 35. 33i, and a is parallel to the
desired plane. Another vector parallel to the plane is the vector connecting any point on the line of intersection to the given
point (31. 2. 1) in the plane. Setting r = 0, the equations of the planes reduce to j 3: = 2 and 3j + 3: = 1 with
simultaneous solution j =
7
2
and : =
3
2
. So a point on the line is
0.
7
2
.
3
2
and another vector parallel to the plane is
31. 3
3
2
. 3
1
2
. Then a normal vector to the plane is n = h2. 35. 33i
31. 3
3
2
. 3
1
2
= h32. 4. 38i and an equation of

the plane is 32(r + 1) + 4(j 32) 38(: 31) = 0 or r 32j + 4: = 31.
39. To nd the r-intercept we set j = : = 0 in the equation 2r + 5j +: = 10
and obtain 2r = 10 i r = 5 so the r-intercept is (5. 0. 0). When
r = : = 0 we get 5j = 10 i j = 2, so the j-intercept is (0. 2. 0).
Setting r = j = 0 gives : = 10, so the :-intercept is (0. 0. 10) and we
graph the portion of the plane that lies in the rst octant.
41. Setting j = : = 0 in the equation 6r 33j + 4: = 6 gives 6r = 6 i
r = 1, when r = : = 0 we have 33j = 6 i j = 32, and r = j = 0
implies 4: = 6 i : =
3
2
, so the intercepts are (1. 0. 0), (0. 32. 0), and
(0. 0.
3
2
). The gure shows the portion of the plane cut off by the coordinate
planes.
43. Substitute the parametric equations of the line into the equation of the plane: (3 3t) 3(2 +t) + 2(5t) = 9 i
8t = 8 i t = 1. Therefore, the point of intersection of the line and the plane is given by r = 3 31 = 2, j = 2 + 1 = 3,
and : = 5(1) = 5. that is, the point (2. 3. 5).
45. Parametric equations for the line are r = t, j = 1 +t, : =
1
2
t and substituting into the equation of the plane gives
4(t) 3(1 +t) + 3
1
2
t
= 8 i
9
2
t = 9 i t = 2. Thus r = 2, j = 1 + 2 = 3, : =
1
2
(2) = 1 and the point of
intersection is (2. 3. 1).
47. Setting r = 0, we see that (0. 1. 0) satises the equations of both planes, so that they do in fact have a line of intersection.
v = n
1
n
2
= h1. 1. 1i h1. 0. 1i = h1. 0. 31i is the direction of this line. Therefore, direction numbers of the intersecting
line are 1, 0, 31.
49. Normal vectors for the planes are n
1
= h1. 4. 33i and n
2
= h33. 6. 7i, so the normals (and thus the planes) arent parallel.
But n
1
n
2
= 33 + 24 321 = 0, so the normals (and thus the planes) are perpendicular.
51. Normal vectors for the planes are n
1
= h1. 1. 1i and n
2
= h1. 31. 1i. The normals are not parallel, so neither are the planes.
Furthermore, n
1
n
2
= 1 31 + 1 = 1 6= 0, so the planes arent perpendicular. The angle between them is given by
cos 0 =
n
1
n
2
|n
1
| |n
2
|
=
1
I
3
I
3
=
1
3
i 0 = cos
31
1
3
E 70.5
.
53. The normals are n
1
= h1. 34. 2i and n
2
= h2. 38. 4i. Since n
2
= 2n
1
, the normals (and thus the planes) are parallel.
55. (a) To nd a point on the line of intersection, set one of the variables equal to a constant, say : = 0. (This will fail if the line of
intersection does not cross the rj-plane; in that case, try setting r or j equal to 0.) The equations of the two planes reduce
to r +j = 1 and r + 2j = 1. Solving these two equations gives r = 1, j = 0. Thus a point on the line is (1. 0. 0).
A vector v in the direction of this intersecting line is perpendicular to the normal vectors of both planes, so we can take
v = n
1
n
2
= h1. 1. 1i h1. 2. 2i = h2 32. 1 32. 2 31i = h0. 31. 1i. By Equations 2, parametric equations for the
line are r = 1, j = 3t, : = t.
(b) The angle between the planes satises cos 0 =
n
1
n
2
|n
1
| |n
2
|
=
1 + 2 + 2
I
3
I
9
=
5
3
I
3
. Therefore 0 = cos
31
5
3
I
3
E 15.8
.
57. Setting : = 0, the equations of the two planes become 5r 32j = 1 and 4r +j = 6. Solving these two equations gives
r = 1, j = 2 so a point on the line of intersection is (1. 2. 0). A vector v in the direction of this intersecting line is
perpendicular to the normal vectors of both planes. So we can use v = n
1
n
2
= h5. 32. 32i h4. 1. 1i = h0. 313. 13i or
equivalently we can take v = h0. 31. 1i, and symmetric equations for the line are r = 1,
j 32
31
=
:
1
or r = 1, j 32 = 3:.
59. The distance from a point (r. j. :) to (1. 0. 32) is d
1
=

(r 31)
2
+j
2
+ (: + 2)
2
and the distance from (r. j. :) to
(3. 4. 0) is
(r 33)
2
+ (j 34)
2
+:
2
. The plane consists of all points (r. j. :) where d
1
= d
2
i d
2
1
= d
2
2
C
(r 31)
2
+j
2
+ (: + 2)
2
= (r 33)
2
+ (j 34)
2
+:
2
C
r
2
32r +j
2
+:
2
+ 4: + 5 = r
2
36r +j
2
38j +:
2
+ 25 C 4r + 8j + 4: = 20 so an equation for the plane is
4r + 8j + 4: = 20 or equivalently r + 2j +: = 5.
Alternatively, you can argue that the segment joining points (1. 0. 32) and (3. 4. 0) is perpendicular to the plane and the plane
includes the midpoint of the segment.

123
61. The plane contains the points (o. 0. 0), (0. /. 0) and (0. 0. c). Thus the vectors a = h3o. /. 0i and b = h3o. 0. ci lie in the
plane, and n = a b = h/c 30. 0 +oc. 0 +o/i = h/c. oc. o/i is a normal vector to the plane. The equation of the plane is
therefore /cr +ocj +o/: = o/c + 0 + 0 or /cr +ocj +o/: = o/c. Notice that if o 6= 0, / 6= 0 and c 6= 0 then we can
rewrite the equation as
r
o
+
j
/
+
:
c
= 1. This is a good equation to remember!
63. Two vectors which are perpendicular to the required line are the normal of the given plane, h1. 1. 1i, and a direction vector for
the given line, h1. 31. 2i. So a direction vector for the required line is h1. 1. 1i h1. 31. 2i = h3. 31. 32i. Thus 1 is given
by hr. j. :i = h0. 1. 2i +th3. 31. 32i, or in parametric form, r = 3t, j = 1 3t, : = 2 32t.
65. Let 1
.
have normal vector n
.
. Then n
1
= h4. 32. 6i, n
2
= h4. 32. 32i, n
3
= h36. 3. 39i, n
4
= h2. 31. 31i. Now
n
1
= 3
2
3
n
3
, so n
1
and n
3
are parallel, and hence 1
1
and 1
3
are parallel; similarly 1
2
and 1
4
are parallel because n
2
= 2n
4
.
However, n
1
and n
2
are not parallel.

0. 0.
1
2
lies on 1
1
, but not on 1
3
, so they are not the same plane, but both 1
2
and 1
4
contain the point (0. 0. 33), so these two planes are identical.
67. Let Q = (1. 3. 4) and 1 = (2. 1. 1), points on the line corresponding to t = 0 and t = 1. Let
1 = (4. 1. 32). Then a =
3<
Q1 = h1. 32. 33i, b =
33<
Q1 = h3. 32. 36i. The distance is
d =
|a b|
|a|
=
|h1. 32. 33i h3. 32. 36i|
|h1. 32. 33i|
=
|h6. 33. 4i|
|h1. 32. 33i|
=
6
2
+ (33)
2
+ 4
2
1
2
+ (32)
2
+ (33)
2
=
I
61
I
14
=
64
14
.
69. By Equation 9, the distance is 1 =
|or
1
+/j
1
+c:
1
+d|
I
o
2
+/
2
+c
2
=
|3(1) + 2(32) + 6(4) 35|
I
3
2
+ 2
2
+ 6
2
=
|18|
I
49
=
18
7
.
71. Put j = : = 0 in the equation of the rst plane to get the point (2. 0. 0) on the plane. Because the planes are parallel, the
distance 1 between them is the distance from (2. 0. 0) to the second plane. By Equation 9,
1 =
|4(2) 36(0) + 2(0) 33|
4
2
+ (36)
2
+ (2)
2
=
5
I
56
=
5
2
I
14
or
5
I
14
28
.
73. The distance between two parallel planes is the same as the distance between a point on one of the planes and the other plane.
Let 1
0
= (r
0
. j
0
. :
0
) be a point on the plane given by or +/j +c: +d
1
= 0. Then or
0
+/j
0
+c:
0
+d
1
= 0 and the
distance between 1
0
and the plane given by or +/j + c: +d
2
= 0 is, from Equation 9,
1 =
|or
0
+/j
0
+c:
0
+d
2
|
I
o
2
+/
2
+c
2
=
|3d
1
+d
2
|
I
o
2
+/
2
+c
2
=
|d
1
3d
2
|
I
o
2
+/
2
+c
2
.
75. 1
1
: r = j = : i r = j (1). 1
2
: r + 1 = j2 = :3 i r + 1 = j2 (2). The solution of (1) and (2) is
r = j = 32. However, when r = 32, r = : i : = 32, but r + 1 = :3 i : = 33, a contradiction. Hence the
lines do not intersect. For 1
1
, v
1
= h1. 1. 1i, and for 1
2
, v
2
= h1. 2. 3i, so the lines are not parallel. Thus the lines are skew
lines. If two lines are skew, they can be viewed as lying in two parallel planes and so the distance between the skew lines
would be the same as the distance between these parallel planes. The common normal vector to the planes must be
perpendicular to both h1. 1. 1i and h1. 2. 3i, the direction vectors of the two lines. So set
n = h1. 1. 1i h1. 2. 3i = h3 32. 33 + 1. 2 31i = h1. 32. 1i. From above, we know that (32. 32. 32) and (32. 32. 33)
are points of 1
1
and 1
2
respectively. So in the notation of Equation 8, 1(32) 32(32) +1(32) +d
1
= 0 i d
1
= 0 and
1(32) 32(32) + 1(33) +d
2
= 0 i d
2
= 1.
By Exercise 73, the distance between these two skew lines is 1 =
|0 31|
I
1 + 4 + 1
=
1
I
6
.
Alternate solution (without reference to planes): A vector which is perpendicular to both of the lines is
n = h1. 1. 1i h1. 2. 3i = h1. 32. 1i. Pick any point on each of the lines, say (32. 32. 32) and (32. 32. 33), and form the
vector b = h0. 0. 1i connecting the two points. The distance between the two skew lines is the absolute value of the scalar
projection of b along n, that is, 1 =
|n b|
|n|
=
|1 0 32 0 + 1 1|
I
1 + 4 + 1
=
1
I
6
.
77. If o 6= 0, then or +/j +c: +d = 0 i o(r +do) +/(j 30) +c(: 30) = 0 which by (7) is the scalar equation of the
plane through the point (3do. 0. 0) with normal vector ho. /. ci. Similarly, if / 6= 0 (or if c 6= 0) the equation of the plane can
be rewritten as o(r 30) +/(j +d/) +c(: 30) = 0 [or as o(r 30) +/(j 30) +c(: +dc) = 0] which by (7) is the
scalar equation of a plane through the point (0. 3d/. 0) [or the point (0. 0. 3dc)] with normal vector ho. /. ci.
13.6 Cylinders and Quadric Surfaces ET 12.6
1. (a) In R
2
, the equation j = r
2
represents a parabola.
(b) In R
3
, the equation j = r
2
doesnt involve :, so any
horizontal plane with equation : = I intersects the graph
in a curve with equation j = r
2
. Thus, the surface is a
parabolic cylinder, made up of innitely many shifted
copies of the same parabola. The rulings are parallel to
the :-axis.
(c) In R
3
, the equation : = j
2
also represents a parabolic
cylinder. Since r doesnt appear, the graph is formed by
moving the parabola : = j
2
in the direction of the r-axis.
Thus, the rulings of the cylinder are parallel to the r-axis.
SECTION 13.6 CYLINDERS AND QUADRIC SURFACES ET SECTION 12.6

125
3. Since r is missing from the equation, the vertical traces
j
2
+4:
2
= 4, r = I, are copies of the same ellipse in the
plane r = I. Thus, the surface j
2
+ 4:
2
= 4 is an elliptic
cylinder with rulings parallel to the r-axis.
5. Since : is missing, each horizontal trace r = j
2
, : = I,
is a copy of the same parabola in the plane : = I. Thus,
the surface r 3j
2
= 0 is a parabolic cylinder with rulings
parallel to the :-axis.
7. Since j is missing, each vertical trace : = cos r, j = I is a copy of a cosine curve in the plane j = I. Thus, the surface
: = cos r is a cylindrical surface with rulings parallel to the j-axis.
9. (a) The traces of r
2
+j
2
3:
2
= 1 in r = I are j
2
3:
2
= 1 3I
2
, a family of hyperbolas. (Note that the hyperbolas are
oriented differently for 31 < I < 1 than for I < 31 or I 1.) The traces in j = I are r
2
3:
2
= 1 3I
2
, a similar
family of hyperbolas. The traces in : = I are r
2
+j
2
= 1 +I
2
, a family of circles. For I = 0, the trace in the
rj-plane, the circle is of radius 1. As |I| increases, so does the radius of the circle. This behavior, combined with the
hyperbolic vertical traces, gives the graph of the hyperboloid of one sheet in Table 1.
(b) The shape of the surface is unchanged, but the hyperboloid is
rotated so that its axis is the j-axis. Traces in j = I are circles,
while traces in r = I and : = I are hyperbolas.
(c) Completing the square in j gives r
2
+ (j + 1)
2
3:
2
= 1. The
surface is a hyperboloid identical to the one in part (a) but shifted
one unit in the negative j-direction.
11. For r = j
2
+ 4:
2
, the traces in r = I are j
2
+ 4:
2
= I. When I 0 we
have a family of ellipses. When I = 0 we have just a point at the origin, and
the trace is empty for I < 0. The traces in j = I are r = 4:
2
+I
2
, a
family of parabolas opening in the positive r-direction. Similarly, the traces
in : = I are r = j
2
+ 4I
2
, a family of parabolas opening in the positive
r-direction. We recognize the graph as an elliptic paraboloid with axis the
r-axis and vertex the origin.
13. r
2
= j
2
+ 4:
2
. The traces in r = I are the ellipses j
2
+ 4:
2
= I
2
. The
traces in j = I are r
2
34:
2
= I
2
, hyperbolas for I 6= 0 and two
intersecting lines if I = 0. Similarly, the traces in : = I are
r
2
3j
2
= 4I
2
, hyperbolas for I 6= 0 and two intersecting lines if I = 0.
We recognize the graph as an elliptic cone with axis the r-axis and vertex
the origin.
15. 3r
2
+ 4j
2
3:
2
= 4. The traces in r = I are the hyperbolas
4j
2
3:
2
= 4 +I
2
. The traces in j = I are r
2
+:
2
= 4I
2
34, a family of
circles for |I| 1, and the traces in : = I are 4j
2
3r
2
= 4 +I
2
, a family
of hyperbolas. Thus the surface is a hyperboloid of two sheets with
axis the j-axis.
17. 36r
2
+j
2
+36:
2
= 36. The traces in r = I are j
2
+36:
2
= 36(1 3I
2
),
a family of ellipses for |I| < 1. (The traces are a single point for |I| = 1
and are empty for |I| 1.) The traces in j = I are the circles
36r
2
+ 36:
2
= 36 3I
2
C r
2
+:
2
= 1 3
1
36
I
2
, |I| < 6, and the
traces in : = I are the ellipses 36r
2
+j
2
= 36(1 3I
2
), |I| < 1. The
graph is an ellipsoid centered at the origin with intercepts r = 1, j = 6,
: = 1.
19. j = :
2
3r
2
. The traces in r = I are the parabolas j = :
2
3I
2
;
the traces in j = I are I = :
2
3r
2
, which are hyperbolas (note the hyperbolas
are oriented differently for I 0 than for I < 0); and the traces in : = I are
the parabolas j = I
2
3r
2
. Thus,
j
1
=
:
2
1
2
3
r
2
1
2
is a hyperbolic paraboloid.
21. This is the equation of an ellipsoid: r
2
+ 4j
2
+ 9:
2
= r
2
+
j
2
(12)
2
+
:
2
(13)
2
= 1, with r-intercepts 1, j-intercepts
1
2
and :-intercepts
1
3
. So the major axis is the r-axis and the only possible graph is VII.
SECTION 13.6 CYLINDERS AND QUADRIC SURFACES ET SECTION 12.6

127
23. This is the equation of a hyperboloid of one sheet, with o = / = c = 1. Since the coefcient of j
2
is negative, the axis of the
hyperboloid is the j-axis, hence the correct graph is II.
25. There are no real values of r and : that satisfy this equation for j < 0, so this surface does not extend to the left of the
r:-plane. The surface intersects the plane j = I 0 in an ellipse. Notice that j occurs to the rst power whereas r and :
occur to the second power. So the surface is an elliptic paraboloid with axis the j-axis. Its graph is VI.
27. This surface is a cylinder because the variable j is missing from the equation. The intersection of the surface and the r:-plane
is an ellipse. So the graph is VIII.
29. :
2
= 4r
2
+ 9j
2
+ 36 or 34r
2
39j
2
+:
2
= 36 or
3
r
2
9
3
j
2
4
+
:
2
36
= 1 represents a hyperboloid of two
sheets with axis the :-axis.
31. r = 2j
2
+ 3:
2
or r =
j
2
12
+
:
2
13
or
r
6
=
j
2
3
+
:
2
2
represents an elliptic paraboloid with vertex (0. 0. 0) and
axis the r-axis.
33. Completing squares in j and : gives
4r
2
+ (j 32)
2
+ 4(: 33)
2
= 4 or
r
2
+
(j 32)
2
4
+ (: 33)
2
= 1, an ellipsoid with
center (0. 2. 3).
35. Completing squares in all three variables gives
(r 32)
2
3(j + 1)
2
+ (: 31)
2
= 0 or
(j + 1)
2
= (r 32)
2
+ (: 31)
2
, a circular cone with
center (2. 31. 1) and axis the horizontal line r = 2,
: = 1.
37. Solving the equation for : we get : =
1 + 4r
2
+j
2
, so we plot separately : =

1 + 4r
2
+j
2
and
: = 3
1 + 4r
2
+j
2
.
To restrict the :-range as in the second graph, we can use the option view = -4..4 in Maples plot3d command, or
PlotRange - {-4,4} in Mathematicas Plot3D command.
39. Solving the equation for : we get : =
4r
2
+j
2
, so we plot separately : =

4r
2
+j
2
and : = 3
4r
2
+j
2
.
41.
43. The surface is a paraboloid of revolution (circular paraboloid) with vertex at the origin, axis the j-axis and opens to the right.
Thus the trace in the j:-plane is also a parabola: j = :
2
, r = 0. The equation is j = r
2
+:
2
.
45. Let 1 = (r, j, :) be an arbitrary point equidistant from (31, 0, 0) and the plane r = 1. Then the distance from 1 to
(31, 0, 0) is
(r + 1)
2
+j
2
+:
2
and the distance from 1 to the plane r = 1 is |r 31|
I
1
2
= |r 31|
(by Equation 13.5.9 [ ET 12.5.9]). So |r 31| =

(r + 1)
2
+j
2
+:
2
C (r 31)
2
= (r + 1)
2
+j
2
+:
2
C
r
2
32r + 1 = r
2
+ 2r + 1 +j
2
+:
2
C 34r = j
2
+:
2
. Thus the collection of all such points 1 is a circular
paraboloid with vertex at the origin, axis the r-axis, which opens in the negative direction.
47. (a) An equation for an ellipsoid centered at the origin with intercepts r = o, j = /, and : = c is
r
2
o
2
+
j
2
/
2
+
:
2
c
2
= 1.
Here the poles of the model intersect the :-axis at : = 6356.523 and the equator intersects the r- and j-axes at
r = 6378.137, j = 6378.137, so an equation is
r
2
(6378.137)
2
+
j
2
(6378.137)
2
+
:
2
(6356.523)
2
= 1
(b) Traces in : = I are the circles
r
2
(6378.137)
2
+
j
2
(6378.137)
2
= 1 3
I
2
(6356.523)
2
C
r
2
+j
2
= (6378.137)
2
3
6378.137
6356.523
2
I
2
.
CHAPTER 13 REVIEW ET CHAPTER 12 129
(c) To identify the traces in j = ir we substitute j = ir into the equation of the ellipsoid:
r
2
(6378.137)
2
+
(ir)
2
(6378.137)
2
+
:
2
(6356.523)
2
= 1
(1 +i
2
)r
2
(6378.137)
2
+
:
2
(6356.523)
2
= 1
r
2
(6378.137)
2
(1 +i
2
)
+
:
2
(6356.523)
2
= 1
As expected, this is a family of ellipses.
49. If (o. /. c) satises : = j
2
3r
2
, then c = /
2
3o
2
. 1
1
: r = o +t, j = / +t, : = c + 2(/ 3o)t,
1
2
: r = o +t, j = / 3t, : = c 32(/ +o)t. Substitute the parametric equations of 1
1
into the equation
of the hyperbolic paraboloid in order to nd the points of intersection: : = j
2
3r
2
i
c + 2(/ 3o)t = (/ +t)
2
3(o +t)
2
= /
2
3o
2
+ 2(/ 3o)t i c = /
2
3o
2
. As this is true for all values of t,
1
1
lies on : = j
2
3r
2
. Performing similar operations with 1
2
gives: : = j
2
3r
2
i
c 32(/ +o)t = (/ 3t)
2
3(o +t)
2
= /
2
3o
2
32(/ +o)t i c = /
2
3o
2
. This tells us that all of 1
2
also lies on
: = j
2
3r
2
.
51. The curve of intersection looks like a bent ellipse. The projection
of this curve onto the rj-plane is the set of points (r. j. 0) which
satisfy r
2
+j
2
= 1 3j
2
C r
2
+ 2j
2
= 1 C
r
2
+
j
2
1
I
2
2
= 1. This is an equation of an ellipse.
13 Review ET 12
1. A scalar is a real number, while a vector is a quantity that has both a real-valued magnitude and a direction.
2. To add two vectors geometrically, we can use either the Triangle Law or the Parallelogram Law, as illustrated in Figures 3
and 4 in Section 13.2 [ ET 12.2]. Algebraically, we add the corresponding components of the vectors.
3. For c 0, c a is a vector with the same direction as a and length c times the length of a. If c < 0, ca points in the opposite
direction as a and has length |c| times the length of a. (See Figures 7 and 15 in Section 13.2 [ ET 12.2].) Algebraically, to nd
c a we multiply each component of a by c.
4. See (1) in Section 13.2 [ ET 12.2].
5. See Theorem 13.3.3 [ ET 12.3.3] and Denition 13.3.1 [ ET 12.3.1].
6. The dot product can be used to nd the angle between two vectors and the scalar projection of one vector onto another. In
particular, the dot product can determine if two vectors are orthogonal. Also, the dot product can be used to determine the
work done moving an object given the force and displacement vectors.
7. See the boxed equations on page 819 [ ET 783] as well as Figures 4 and 5 and the accompanying discussion on pages 81819
[ ET 78283] .
8. See Theorem 13.4.6 [ ET 12.4.6] and the preceding discussion; use either (1) or (4) in Section 13.4 [ ET 12.4].
9. The cross product can be used to create a vector orthogonal to two given vectors as well as to determine if two vectors are
parallel. The cross product can also be used to nd the area of a parallelogram determined by two vectors. In addition, the
cross product can be used to determine torque if the force and position vectors are known.
10. (a) The area of the parallelogram determined by a and b is the length of the cross product: |a b|.
(b) The volume of the parallelepiped determined by a, b, and c is the magnitude of their scalar triple product: |a (b c)|.
11. If an equation of the plane is known, it can be written as or +/j +c: +d = 0. A normal vector, which is perpendicular to the
plane, is ho. /. ci (or any scalar multiple of ho. /. ci). If an equation is not known, we can use points on the plane to nd two
non-parallel vectors which lie in the plane. The cross product of these vectors is a vector perpendicular to the plane.
12. The angle between two intersecting planes is dened as the acute angle between their normal vectors. We can nd this angle
using Corollary 13.3.6 [ ET 12.3.6].
13. See (1), (2), and (3) in Section 13.5 [ ET 12.5].
14. See (5), (6), and (7) in Section 13.5 [ ET 12.5].
15. (a) Two (nonzero) vectors are parallel if and only if one is a scalar multiple of the other. In addition, two nonzero vectors are
parallel if and only if their cross product is 0.
(b) Two vectors are perpendicular if and only if their dot product is 0.
(c) Two planes are parallel if and only if their normal vectors are parallel.
16. (a) Determine the vectors
33<
1Q = ho
1
. o
2
. o
3
i and
3<
11 = h/
1
. /
2
. /
3
i. If there is a scalar t such that
ho
1
. o
2
. o
3
i = t h/
1
. /
2
. /
3
i, then the vectors are parallel and the points must all lie on the same line.
Alternatively, if
33<
1Q
3<
11 = 0, then
33<
1Q and
3<
11 are parallel, so 1, Q, and 1 are collinear.
Thirdly, an algebraic method is to determine an equation of the line joining two of the points, and then check whether or
not the third point satises this equation.
(b) Find the vectors
33<
1Q = a,
3<
11 = b,
3<
1o = c. a b is normal to the plane formed by 1, Q and 1, and so o lies on this
plane if a b and c are orthogonal, that is, if (a b) c = 0. (Or use the reasoning in Example 5 in Section 13.4
[ ET 12.4].)
Alternatively, nd an equation for the plane determined by three of the points and check whether or not the fourth point
satises this equation.
17. (a) See Exercise 13.4.43 [ ET 12.4.43].
(b) See Example 8 in Section 13.5 [ ET 12.5].
(c) See Example 10 in Section 13.5 [ ET 12.5].
18. The traces of a surface are the curves of intersection of the surface with planes parallel to the coordinate planes. We can nd
the trace in the plane r = I (parallel to the j:-plane) by setting r = I and determining the curve represented by the resulting
equation. Traces in the planes j = I (parallel to the r:-plane) and : = I (parallel to the rj-plane) are found similarly.
19. See Table 1 in Section 13.6 [ ET 12.6].
1. True, by Theorem 13.3.2 [ ET 12.3.2], property 2.
3. True. If 0 is the angle between u and v, then by Theorem 13.4.6 [ ET 12.4.6],
|u v| = |u| |v| sin0 = |v| |u| sin0 = |v u|.
(Or, by Theorem 13.4.8 [ ET 12.4.8], |u v| = |3v u| = |31| |v u| = |v u|.)
5. Theorem 13.4.8 [ ET 12.4.8], property 2 tells us that this is true.
7. This is true by Theorem 13.4.8 [ ET 12.4.8], property 5.
9. This is true because u v is orthogonal to u (see Theorem 13.4.5 [ ET 12.4.5]), and the dot product of two orthogonal vectors
is 0.
11. If |u| = 1, |v| = 1 and 0 is the angle between these two vectors (so 0 $ 0 $ ), then by Theorem 13.4.6 [ ET 12.4.6],
|u v| = |u| |v| sin0 = sin0, which is equal to 1 if and only if 0 =
r
2
(that is, if and only if the two vectors are orthogonal).
Therefore, the assertion that the cross product of two unit vectors is a unit vector is false.
13. This is false. In R
2
, r
2
+j
2
= 1 represents a circle, but
(r. j. :) | r
2
+j
2
= 1
represents a three-dimensional surface,

namely, a circular cylinder with axis the :-axis.
15. False. For example, i j = 0 but i 6= 0 and j 6= 0.
17. This is true. If u and v are both nonzero, then by (7) in Section 13.3 [ET 12.3], u v = 0 implies that u and v are orthogonal.
But uv = 0 implies that u and v are parallel (see Corollary 13.4.7 [ET 12.4.7]). Two nonzero vectors cant be both parallel
and orthogonal, so at least one of u, v must be 0.
1. (a) The radius of the sphere is the distance between the points (31. 2. 1) and (6. 32. 3), namely,
[6 3(31)]
2
+ (32 32)
2
+ (3 31)
2
=
I
69. By the formula for an equation of a sphere (see page 804 [ET 768]),
an equation of the sphere with center (31. 2. 1) and radius
I
69 is (r + 1)
2
+ (j 32)
2
+ (: 31)
2
= 69.
(b) The intersection of this sphere with the j:-plane is the set of points on the sphere whose r-coordinate is 0. Putting r = 0
into the equation, we have (j 32)
2
+ (: 31)
2
= 68. r = 0 which represents a circle in the j:-plane with center (0. 2. 1)
and radius
I
68.
(c) Completing squares gives (r 34)
2
+ (j + 1)
2
+ (: + 3)
2
= 31 + 16 + 1 + 9 = 25. Thus the sphere is centered at
(4. 31. 33) and has radius 5.
3. u v = |u| |v| cos 45
= (2)(3)
I
2
2
= 3
I
2. |u v| = |u| |v| sin45
= (2)(3)
I
2
2
= 3
I
2.
By the right-hand rule, u v is directed out of the page.
5. For the two vectors to be orthogonal, we need h3. 2. ri h2r. 4. ri = 0 C (3)(2r) + (2)(4) + (r)(r) = 0 C
r
2
+ 6r + 8 = 0 C (r + 2)(r + 4) = 0 C r = 32 or r = 34.
7. (a) (u v) w = u (v w) = 2
(b) u (wv) = u [3(v w)] = 3u (v w) = 32
(c) v (u w) = (v u) w = 3(u v) w = 32
(d) (u v) v = u (v v) = u 0 = 0
9. For simplicity, consider a unit cube positioned with its back left corner at the origin. Vector representations of the diagonals
joining the points (0. 0. 0) to (1. 1. 1) and (1. 0. 0) to (0. 1. 1) are h1. 1. 1i and h31. 1. 1i. Let 0 be the angle between these
two vectors. h1. 1. 1i h31. 1. 1i = 31 + 1 + 1 = 1 = |h1. 1. 1i| |h31. 1. 1i| cos 0 = 3 cos 0 i cos 0 =
1
3
i
0 = cos
31
1
3
E 71
.
11.
3<
1 = h1. 0. 31i,
3<
C = h0. 4. 3i, so
(a) a vector perpendicular to the plane is
3<
1
3<
C = h0 + 4. 3(3 + 0). 4 30i = h4. 33. 4i.
(b)
1
2
3<
1
3<
C
=
1
2
I
16 + 9 + 16 =
I
41
2
.
13. Let 1
1
be the magnitude of the force directed 20
away from the direction of shore, and let 1

2
be the magnitude of the other
force. Separating these forces into components parallel to the direction of the resultant force and perpendicular to it gives
1
1
cos 20
+ 1
2
cos 30
= 255 (1), and 1

1
sin20
3 1
2
sin30
= 0 i 1
1
= 1
2
sin30
sin20
(2). Substituting (2)

into (1) gives 1
2
(sin30
cot 20
+ cos 30
) = 255 i 1
2
E 114 N. Substituting this into (2) gives 1
1
E 166 N.
15. The line has direction v = h33. 2. 3i. Letting 1
0
= (4. 31. 2), parametric equations are
r = 4 33t, j = 31 + 2t, : = 2 + 3t.
17. A direction vector for the line is a normal vector for the plane, n = h2. 31. 5i, and parametric equations for the line are
r = 32 + 2t, j = 2 3t, : = 4 + 5t.
19. Here the vectors a = h4 33. 0 3(31) . 2 31i = h1. 1. 1i and b = h6 33. 3 3(31). 1 31i = h3. 4. 0i lie in the plane,
so n = a b = h34. 3. 1i is a normal vector to the plane and an equation of the plane is
34(r 33) + 3(j 3(31)) + 1(: 31) = 0 or 34r + 3j +: = 314.
21. Substitution of the parametric equations into the equation of the plane gives 2r 3j +: = 2(2 3t) 3(1 + 3t) + 4t = 2 i
3t + 3 = 2 i t = 1. When t = 1, the parametric equations give r = 2 31 = 1, j = 1 + 3 = 4 and : = 4. Therefore,
the point of intersection is (1. 4. 4).
23. Since the direction vectors h2. 3. 4i and h6. 31. 2i arent parallel, neither are the lines. For the lines to intersect, the three
equations 1 + 2t = 31 + 6c, 2 + 3t = 3 3c, 3 + 4t = 35 + 2c must be satised simultaneously. Solving the rst two
equations gives t =
1
5
, c =
2
5
and checking we see these values dont satisfy the third equation. Thus the lines arent parallel
and they dont intersect, so they must be skew.
25. n
1
= h1. 0. 31i and n
2
= h0. 1. 2i. Setting : = 0, it is easy to see that (1. 3. 0) is a point on the line of intersection of
r 3: = 1 and j + 2: = 3. The direction of this line is v
1
= n
1
n
2
= h1. 32. 1i. A second vector parallel to the desired
plane is v
2
= h1. 1. 32i, since it is perpendicular to r +j 32: = 1. Therefore, the normal of the plane in question is
n = v
1
v
2
= h4 31. 1 + 2. 1 + 2i = 3 h1. 1. 1i. Taking (r
0
. j
0
. :
0
) = (1. 3. 0), the equation we are looking for is
(r 31) + (j 33) +: = 0 C r +j +: = 4.
27. By Exercise 13.5.73 [ ET 12.5.73], 1 =
|2 324|
I
26
=
22
I
26
.
29. The equation r = : represents a plane perpendicular to
the r:-plane and intersecting the r:-plane in the line
r = :, j = 0.
31. The equation r
2
= j
2
+ 4:
2
represents a (right elliptical)
cone with vertex at the origin and axis the r-axis.
33. An equivalent equation is 3r
2
+
j
2
4
3:
2
= 1, a
hyperboloid of two sheets with axis the j-axis. For
|j| 2, traces parallel to the r:-plane are circles.
35. Completing the square in j gives
4r
2
+ 4(j 31)
2
+:
2
= 4 or r
2
+ (j 31)
2
+
:
2
4
= 1,
an ellipsoid centered at (0. 1. 0).
37. 4r
2
+j
2
= 16 C
r
2
4
+
j
2
16
= 1. The equation of the ellipsoid is
r
2
4
+
j
2
16
+
:
2
c
2
= 1, since the horizontal trace in the
plane : = 0 must be the original ellipse. The traces of the ellipsoid in the j:-plane must be circles since the surface is obtained
by rotation about the r-axis. Therefore, c
2
= 16 and the equation of the ellipsoid is
r
2
4
+
j
2
16
+
:
2
16
= 1 C
4r
2
+j
2
+:
2
= 16.
PROBLEMS PLUS
1. Since three-dimensional situations are often difcult to visualize and work with, let
us rst try to nd an analogous problem in two dimensions. The analogue of a cube
is a square and the analogue of a sphere is a circle. Thus a similar problem in two
dimensions is the following: if ve circles with the same radius v are contained in a
square of side 1 m so that the circles touch each other and four of the circles touch
two sides of the square, nd v.
The diagonal of the square is
I
2. The diagonal is also 4v + 2r. But r is the diagonal of a smaller square of side v. Therefore
r =
I
2 v i
I
2 = 4v + 2r = 4v + 2
I
2 v =

4 + 2
I
2
v i v =
I
2
4 +2
I
2
.
Lets use these ideas to solve the original three-dimensional problem. The diagonal of the cube is
I
1
2
+ 1
2
+ 1
2
=
I
3.
The diagonal of the cube is also 4v + 2r where r is the diagonal of a smaller cube with edge v. Therefore
r =
I
v
2
+v
2
+v
2
=
I
3 v i
I
3 = 4v + 2r = 4v + 2
I
3 v =

4 + 2
I
3
v. Thus v =
I
3
4 + 2
I
3
=
2
I
3 3 3
2
.
The radius of each ball is
I
3 3
3
2
m.
3. (a) We nd the line of intersection 1 as in Example 13.5.7(b) [ ET 12.5.7(b)]. Observe that the point (31. c. c) lies on both
planes. Now since 1 lies in both planes, it is perpendicular to both of the normal vectors n
1
and n
2
, and thus parallel to
their cross product n
1
n
2
=
i j k
c 1 1
1 3c c
=

2c. 3c
2
+ 1. 3c
2
31
. So symmetric equations of 1 can be written as

r + 1
32c
=
j 3c
c
2
31
=
: 3c
c
2
+ 1
, provided that c 6= 0, 1.
If c = 0, then the two planes are given by j +: = 0 and r = 31, so symmetric equations of 1 are r = 31, j = 3:. If
c = 31, then the two planes are given by 3r +j +: = 31 and r +j +: = 31, and they intersect in the line r = 0,
j = 3: 31. If c = 1, then the two planes are given by r +j +: = 1 and r 3j +: = 1, and they intersect in the line
j = 0, r = 1 3:.
(b) If we set : = t in the symmetric equations and solve for r and j separately, we get r + 1 =
(t 3c)(32c)
c
2
+ 1
,
j 3c =
(t 3c)(c
2
31)
c
2
+ 1
i r =
32ct + (c
2
31)
c
2
+ 1
, j =
(c
2
31)t + 2c
c
2
+ 1
. Eliminating c from these equations, we
have r
2
+j
2
= t
2
+ 1. So the curve traced out by 1 in the plane : = t is a circle with center at (0. 0. t) and
radius
I
t
2
+ 1.
(c) The area of a horizontal cross-section of the solid is (:) = (:
2
+ 1), so \ =

1
0
(:)d: =
1
3
:
3
+:
1
0
=
4r
3
.
135
136

PROBLEMS PLUS
5. (a) When 0 = 0
s
, the block is not moving, so the sum of the forces on the block
must be 0, thus N+F+W = 0. This relationship is illustrated
geometrically in the gure. Since the vectors form a right triangle, we have
tan(0
s
) =
|F|
|N|
=
j
s
n
n
= j
s
.
(b) We place the block at the origin and sketch the force vectors acting on the block, including the additional horizontal force
H, with initial points at the origin. We then rotate this system so that F lies along the positive r-axis and the inclined plane
is parallel to the r-axis.
|F| is maximal, so |F| = j
s
n for 0 0
s
. Then the vectors, in terms of components parallel and perpendicular to the
inclined plane, are
N = n j F = (j
s
n) i
W = (3io sin0) i + (3io cos 0) j H = (/
min
cos 0) i + (3/
min
sin0) j
Equating components, we have
j
s
n 3io sin0 +/
min
cos 0 = 0 i /
min
cos 0 +j
s
n = io sin0 (1)
n 3io cos 0 3/
min
sin0 = 0 i /
min
sin0 +io cos 0 = n (2)
(c) Since (2) is solved for n, we substitute into (1):
/
min
cos 0 +j
s
(/
min
sin0 +io cos 0) =io sin0 i
/
min
cos 0 +/
min
j
s
sin0 =io sin0 3ioj
s
cos 0 i
/
min
= io
sin0 3j
s
cos 0
cos 0 +j
s
sin0
= io
tan0 3j
s
1 +j
s
tan0
From part (a) we know j

s
= tan0
s
, so this becomes /
min
= io
tan0 3tan0
s
1 + tan0
s
tan0
and using a trigonometric identity,

this is io tan(0 30
s
) as desired.
Note for 0 = 0
s
, /
min
= io tan0 = 0, which makes sense since the block is at rest for 0
s
, thus no additional force H
is necessary to prevent it from moving. As 0 increases, the factor tan(0 30
s
), and hence the value of /
min
, increases
slowly for small values of 0 30
s
but much more rapidly as 0 30
s
becomes signicant. This seems reasonable, as the
PROBLEMS PLUS

137
steeper the inclined plane, the less the horizontal components of the various forces affect the movement of the block, so we
would need a much larger magnitude of horizontal force to keep the block motionless. If we allow 0 <90
, corresponding
to the inclined plane being placed vertically, the value of /
min
is quite large; this is to be expected, as it takes a great
amount of horizontal force to keep an object from moving vertically. In fact, without friction (so 0
s
= 0), we would have
0 <90
i /
min
<", and it would be impossible to keep the block from slipping.
(d) Since /
max
is the largest value of / that keeps the block from slipping, the force of friction is keeping the block from
moving up the inclined plane; thus, F is directed down the plane. Our system of forces is similar to that in part (b), then,
except that we have F = 3(j
s
n) i. (Note that |F| is again maximal.) Following our procedure in parts (b) and (c), we
equate components:
3j
s
n 3io sin0 +/
max
cos 0 = 0 i /
max
cos 0 3j
s
n = io sin0
n 3io cos 0 3/
max
sin0 = 0 i /
max
sin0 +io cos 0 = n
Then substituting,
/
max
cos 0 3j
s
(/
max
sin0 +io cos 0) = io sin0 i
/
max
cos 0 3/
max
j
s
sin0 = io sin0 +ioj
s
cos 0 i
/
max
= io
sin0 +j
s
cos 0
cos 0 3j
s
sin0
= io
tan0 +j
s
1 3j
s
tan0
= io
tan0 + tan0
s
1 3tan0
s
tan0
= io tan(0 +0
s
)
We would expect /
max
to increase as 0 increases, with similar behavior as we established for /
min
, but with /
max
values
always larger than /
min
. We can see that this is the case if we graph /
max
as a function of 0, as the curve is the graph of
/
min
translated 20
s
to the left, so the equation does seem reasonable. Notice that the equation predicts /
max
<"as
0 <(90
3 0
s
). In fact, as /
max
increases, the normal force increases as well. When (90
3 0
s
) $ 0 $ 90
, the
horizontal force is completely counteracted by the sum of the normal and frictional forces, so no part of the horizontal
force contributes to moving the block up the plane no matter how large its magnitude.
14 VECTOR FUNCTIONS ET 13
14.1 Vector Functions and Space Curves ET 13.1
1. The component functions
I
4 3t
2
, c
33I
, and ln(t + 1) are all dened when 4 3t
2
D 0 i 32 $ t $ 2 and
t + 1 0 i t 31, so the domain of r is (31. 2].
3. lim
I<0
+
cos t = cos 0 = 1, lim
I<0
+
sint = sin0 = 0, lim
I<0
+
t lnt = lim
I<0
+
lnt
1t
= lim
I<0
+
1t
31t
2
= lim
I<0
+
3t = 0
[by lHospitals Rule]. Thus lim
I<0
+
hcos t. sint. t lnti =

lim
I<0
+
cos t. lim
I<0
+
sint. lim
I<0
+
t lnt
= h1. 0. 0i.
5. lim
I<0
c
33I
= c
0
= 1, lim
I<0
t
2
sin
2
t
= lim
I<0
1
sin
2
t
t
2
=
1
lim
I<0
sin
2
t
t
2
=
1
lim
I<0
sint
t
2
=
1
1
2
= 1
and lim
I<0
cos 2t = cos 0 = 1. Thus the given limit equals i +j +k.
7. The corresponding parametric equations for this curve are r = sint, j = t.
We can make a table of values, or we can eliminate the parameter: t = j i
r = sinj, with j M R. By comparing different values of t, we nd the direction in
which t increases as indicated in the graph.
9. The corresponding parametric equations are r = t, j = cos 2t, : = sin2t.
Note that j
2
+:
2
= cos
2
2t + sin
2
2t = 1, so the curve lies on the circular
cylinder j
2
+:
2
= 1. Since r = t, the curve is a helix.
11. The corresponding parametric equations are r = 1, j = cos t, : = 2 sint.
Eliminating the parameter in j and : gives j
2
+ (:2)
2
= cos
2
t + sin
2
t = 1
or j
2
+:
2
4 = 1. Since r = 1, the curve is an ellipse centered at (1. 0. 0) in
the plane r = 1.
139
140 CHAPTER 14 VECTOR FUNCTIONS ET CHAPTER 13
13. The parametric equations are r = t
2
, j = t
4
, : = t
6
. These are positive
for t 6= 0 and 0 when t = 0. So the curve lies entirely in the rst quadrant.
The projection of the graph onto the rj-plane is j = r
2
, j 0, a half parabola.
On the r:-plane : = r
3
, : 0, a half cubic, and the j:-plane, j
3
= :
2
.
15. Taking r
0
= h0. 0. 0i and r
1
= h1. 2. 3i, we have from Equation 13.5.4 [ET 12.5.4]
r(t) = (1 3t) r
0
+t r
1
= (1 3t) h0. 0. 0i +t h1. 2. 3i, 0 $ t $ 1 or r(t) = ht. 2t. 3ti, 0 $ t $ 1.
Parametric equations are r = t, j = 2t, : = 3t, 0 $ t $ 1.
17. Taking r
0
= h1. 31. 2i and r
1
= h4. 1. 7i, we have
r(t) = (1 3t) r
0
+t r
1
= (1 3t) h1. 31. 2i +t h4. 1. 7i, 0 $ t $ 1 or r(t) = h1 + 3t. 31 + 2t. 2 + 5ti, 0 $ t $ 1.
Parametric equations are r = 1 + 3t, j = 31 + 2t, : = 2 + 5t, 0 $ t $ 1.
19. r = cos 4t, j = t, : = sin4t. At any point (r. j. :) on the curve, r
2
+:
2
= cos
2
4t + sin
2
4t = 1. So the curve lies on a
circular cylinder with axis the j-axis. Since j = t, this is a helix. So the graph is VI.
21. r = t, j = 1(1 +t
2
), : = t
2
. Note that j and : are positive for all t. The curve passes through (0. 1. 0) when t = 0.
As t <", (r. j. :) <(". 0. "), and as t <3", (r. j. :) <(3". 0. "). So the graph is IV.
23. r = cos t, j = sint, : = sin5t. r
2
+j
2
= cos
2
t + sin
2
t = 1, so the curve lies on a circular cylinder with axis the
:-axis. Each of r, j and : is periodic, and at t = 0 and t = 2 the curve passes through the same point, so the curve repeats
itself and the graph is V.
25. If r = t cos t, j = t sint, : = t, then r
2
+j
2
= t
2
cos
2
t +t
2
sin
2
t = t
2
= :
2
,
so the curve lies on the cone :
2
= r
2
+j
2
. Since : = t, the curve is a spiral on
this cone.
27. Parametric equations for the curve are r = t, j = 0, : = 2t 3t
2
. Substituting into the equation of the paraboloid gives
2t 3t
2
= t
2
i 2t = 2t
2
i t = 0, 1. Since r(0) = 0 and r(1) = i +k, the points of intersection
are (0. 0. 0) and (1. 0. 1).
29. r(t) = hcos t sin2t. sint sin2t. cos 2ti.
We include both a regular plot and a plot
showing a tube of radius 0.08 around the
curve.
SECTION 14.1 VECTOR FUNCTIONS AND SPACE CURVES ET SECTION 13.1

141
31. r(t) = ht. t sint. t cos ti
33. r = (1 + cos 16t) cos t, j = (1 + cos 16t) sint, : = 1 + cos 16t. At any
point on the graph,
r
2
+j
2
= (1 + cos 16t)
2
cos
2
t + (1 + cos 16t)
2
sin
2
t
= (1 + cos 16t)
2
= :
2
, so the graph lies on the cone r
2
+j
2
= :
2
.
From the graph at left, we see that this curve looks like the projection of a
leaved two-dimensional curve onto a cone.
35. If t = 31, then r = 1, j = 4, : = 0, so the curve passes through the point (1. 4. 0). If t = 3, then r = 9, j = 38, : = 28,
so the curve passes through the point (9. 38. 28). For the point (4. 7. 36) to be on the curve, we require j = 1 33t = 7 i
t = 32. But then : = 1 + (32)
3
= 37 6= 36, so (4. 7. 36) is not on the curve.
37. Both equations are solved for :, so we can substitute to eliminate ::

r
2
+j
2
= 1 +j i r
2
+j
2
= 1 + 2j +j
2
i
r
2
= 1 + 2j i j =
1
2
(r
2
31). We can form parametric equations for the curve C of intersection by choosing a
parameter r = t, then j =
1
2
(t
2
31) and : = 1 +j = 1 +
1
2
(t
2
31) =
1
2
(t
2
+ 1). Thus a vector function representing C
is r(t) = t i +
1
2
(t
2
31) j +
1
2
(t
2
+ 1) k.
39. The projection of the curve C of intersection onto the
rj-plane is the circle r
2
+j
2
= 4. : = 0. Then we can write
r = 2 cos t, j = 2 sint, 0 $ t $ 2. Since C also lies on
the surface : = r
2
, we have : = r
2
= (2 cos t)
2
= 4 cos
2
t.
Then parametric equations for C are r = 2 cos t, j = 2 sint,
: = 4 cos
2
t, 0 $ t $ 2.
41. For the particles to collide, we require r
1
(t) = r
2
(t) C

t
2
. 7t 312. t
2
=

4t 33. t
2
. 5t 36
. Equating components
gives t
2
= 4t 33, 7t 312 = t
2
, and t
2
= 5t 36. From the rst equation, t
2
34t +3 = 0 C (t 33)(t 31) = 0 so t = 1
or t = 3. t = 1 does not satisfy the other two equations, but t = 3 does. The particles collide when t = 3, at the
point (9. 9. 9).
43. (a) lim
I<a
u(t) + lim
I<a
v(t) =
lim
I<a
&
1
(t). lim
I<a
&
2
(t). lim
I<a
&
3
(t)
lim
I<a
1
(t). lim
I<a
2
(t). lim
I<a
3
(t)
and the limits of these

component functions must each exist since the vector functions both possess limits as t <o. Then adding the two vectors
and using the addition property of limits for real-valued functions, we have that
lim
I<a
u(t) + lim
I<a
v(t) =

lim
I<a
&
1
(t) + lim
I<a
1
(t). lim
I<a
&
2
(t) + lim
I<a
2
(t). lim
I<a
&
3
(t) + lim
I<a
3
(t)
=

lim
I<a
[&
1
(t) +
1
(t)] . lim
I<a
[&
2
(t) +
2
(t)] . lim
I<a
[&
3
(t) +
3
(t)]
= lim
I<a
h&
1
(t) +
1
(t). &
2
(t) +
2
(t). &
3
(t) +
3
(t)i [using (1) backward]
= lim
I<a
[u(t) +v(t)]
(b) lim
I<a
cu(t) = lim
I<a
hc&
1
(t). c&
2
(t). c&
3
(t)i =

lim
I<a
c&
1
(t). lim
I<a
c&
2
(t). lim
I<a
c&
3
(t)
=

c lim
I<a
&
1
(t). c lim
I<a
&
2
(t). c lim
I<a
&
3
(t)
= c
lim
I<a
&
1
(t). lim
I<a
&
2
(t). lim
I<a
&
3
(t)
= c lim
I<a
h&
1
(t). &
2
(t). &
3
(t)i = c lim
I<a
u(t)
(c) lim
I<a
u(t) lim
I<a
v(t) =

lim
I<a
&
1
(t). lim
I<a
&
2
(t). lim
I<a
&
3
(t)
lim
I<a
1
(t). lim
I<a
2
(t). lim
I<a
3
(t)
=

lim
I<a
&
1
(t)

lim
I<a
1
(t)
lim
I<a
&
2
(t)

lim
I<a
2
(t)
lim
I<a
&
3
(t)

lim
I<a
3
(t)
= lim
I<a
&
1
(t)
1
(t) + lim
I<a
&
2
(t)
2
(t) + lim
I<a
&
3
(t)
3
(t)
= lim
I<a
[&
1
(t)
1
(t) +&
2
(t)
2
(t) +&
3
(t)
3
(t)] = lim
I<a
[u(t) v(t)]
(d) lim
I<a
u(t) lim
I<a
v(t) =
lim
I<a
&
1
(t). lim
I<a
&
2
(t). lim
I<a
&
3
(t)
lim
I<a
1
(t). lim
I<a
2
(t). lim
I<a
3
(t)
=

lim
I<a
&
2
(t)

lim
I<a
3
(t)
lim
I<a
&
3
(t)

lim
I<a
2
(t)
lim
I<a
&
3
(t)

lim
I<a
1
(t)
lim
I<a
&
1
(t)

lim
I<a
3
(t)
lim
I<a
&
1
(t)

lim
I<a
2
(t)
lim
I<a
&
2
(t)

lim
I<a
1
(t)
=

lim
I<a
[&
2
(t)
3
(t) 3&
3
(t)
2
(t)] . lim
I<a
[&
3
(t)
1
(t) 3&
1
(t)
3
(t)] .
lim
I<a
[&
1
(t)
2
(t) 3&
2
(t)
1
(t)]
= lim
I<a
h&
2
(t)
3
(t) 3&
3
(t)
2
(t). &
3
(t)
1
(t) 3&
1
(t)
3
(t). &
1
(t)
2
(t) 3&
2
(t)
1
(t)i
= lim
I<a
[u(t) v(t)]
45. Let r(t) = h1 (t) . o (t) . /(t)i and b = h/
1
. /
2
. /
3
i. If lim
I<a
r(t) = b, then lim
I<a
r(t) exists, so by (1),
b = lim
I<a
r(t) =
lim
I<a
1(t). lim
I<a
o(t). lim
I<a
/(t)
. By the denition of equal vectors we have lim

I<a
1(t) = /
1
, lim
I<a
o(t) = /
2
and lim
I<a
/(t) = /
3
. But these are limits of real-valued functions, so by the denition of limits, for every - 0 there exists
c
1
0, c
2
0, c
3
0 so that if 0 < |t 3o| < c
1
then |1(t) 3/
1
| < -3, if 0 < |t 3o| < c
2
then |o(t) 3/
2
| < -3, and
if 0 < |t 3o| < c
3
then |/(t) 3/
3
| < -3. Letting c = minimum of {c
1
. c
2
. c
3
}, then if 0 < |t 3o| < c we have
|1(t) 3/
1
| + |o(t) 3/
2
| + |/(t) 3/
3
| < -3 +-3 +-3 = -. But
|r(t) 3b| = |h1(t) 3/
1
. o(t) 3/
2
. /(t) 3/
3
i| =

(1(t) 3/
1
)
2
+ (o(t) 3/
2
)
2
+ (/(t) 3/
3
)
2
$
[1(t) 3/
1
]
2
+
[o(t) 3/
2
]
2
+
[/(t) 3/
3
]
2
= |1(t) 3/
1
| + |o(t) 3/
2
| + |/(t) 3/
3
|
SECTION 14.2 DERIVATIVES AND INTEGRALS OF VECTOR FUNCTIONS ET SECTION 13.2

143
Thus for every - 0 there exists c 0 such that if 0 < |t 3o| < c then
|r(t) 3b| $ |1(t) 3/
1
| + |o(t) 3/
2
| + |/(t) 3/
3
| < -. Conversely, suppose for every - 0, there exists c 0 such
that if 0 < |t 3o| < c then |r(t) 3b| < - C |h1(t) 3/
1
. o(t) 3/
2
. /(t) 3/
3
i| < - C
[1(t) 3/
1
]
2
+ [o(t) 3/
2
]
2
+ [/(t) 3/
3
]
2
< - C [1(t) 3/
1
]
2
+ [o(t) 3/
2
]
2
+ [/(t) 3/
3
]
2
< -
2
. But each term
on the left side of the last inequality is positive, so if 0 < |t 3o| < c, then [1(t) 3/
1
]
2
< -
2
, [o(t) 3/
2
]
2
< -
2
and
[/(t) 3/
3
]
2
< -
2
or, taking the square root of both sides in each of the above, |1(t) 3/
1
| < -, |o(t) 3/
2
| < - and
|/(t) 3/
3
| < -. And by denition of limits of real-valued functions we have lim
I<a
1(t) = /
1
, lim
I<a
o(t) = /
2
and
lim
I<a
/(t) = /
3
. But by (1), lim
I<a
r(t) =
lim
I<a
1(t). lim
I<a
o(t). lim
I<a
/(t)
, so lim
I<a
r(t) = h/
1
. /
2
. /
3
i = b.
14.2 Derivatives and Integrals of Vector Functions ET 13.2
1. (a)
(b)
r(4.5) 3r(4)
0.5
= 2[r(4.5) 3r(4)], so we draw a vector in the same
direction but with twice the length of the vector r(4.5) 3r(4).
r(4.2) 3r(4)
0.2
= 5[r(4.2) 3r(4)], so we draw a vector in the same
direction but with 5 times the length of the vector r(4.2) 3r(4).
(c) By Denition 1, r
0
(4) = lim
I<0
r(4 +/) 3r(4)
/
. T(4) =
r
0
(4)
|r
0
(4)|
.
(d) T(4) is a unit vector in the same direction as r
0
(4), that is, parallel to the tangent line to the curve at r(4) with length 1.
3. Since (r + 2)
2
= t
2
= j 31 i
j = (r + 2)
2
31, the curve is a
parabola.
(a), (c) (b) r
0
(t) = h1. 2ti,
r
0
(31) = h1. 32i
5. r = sint, j = 2 cos t so
r
2
+ (j2)
2
= 1 and the curve is
an ellipse.
(a), (c) (b) r
0
(t) = cos t i 32 sint j,
r
0
=
I
2
2
i 3
I
2 j
7. Since j = c
3I
= (c
I
)
3
= r
3
, the
curve is part of a cubic cuve. Note
that here, r 0.
(a), (c) (b) r
0
(t) = c
I
i + 3c
3I
j,
r
0
(0) = i + 3 j
9. r
0
(t) =

d
dt
[t sint] .
d
dt
t
2
.
d
dt
[t cos 2t]
= ht cos t + sint. 2t. t(3sin2t) 2 + cos 2ti

= ht cos t + sint. 2t. cos 2t 32t sin2ti
11. r(t) = i 3j +c
4I
k i r
0
(t) = 0 i + 0 j + 4c
4I
k = 4c
4I
k
13. r(t) = c
I
2
i 3 j + ln(1 + 3t) k i r
0
(t) = 2tc
I
2
i +
3
1 + 3t
k
15. r
0
(t) = 0 +b + 2t c = b + 2t c by Formulas 1 and 3 of Theorem 3.
17. r
0
(t) =

3tc
3I
+c
3I
. 2(1 +t
2
). 2c
I
i r
0
(0) = h1. 2. 2i. So |r
0
(0)| =
I
1
2
+ 2
2
+ 2
2
=
I
9 = 3 and
T(0) =
r
0
(0)
|r
0
(0)|
=
1
3
h1. 2. 2i =

1
3
.
2
3
.
2
3
.
19. r
0
(t) = 3sint i + 3 j + 4 cos 2t k i r
0
(0) = 3 j + 4 k. Thus
T(0) =
r
0
(0)
|r
0
(0)|
=
1
I
0
2
+ 3
2
+ 4
2
(3 j + 4 k) =
1
5
(3 j + 4 k) =
3
5
j +
4
5
k.
21. r(t) =

t. t
2
. t
3
i r
0
(t) =

1. 2t. 3t
2
. Then r
0
(1) = h1. 2. 3i and |r
0
(1)| =
I
1
2
+ 2
2
+ 3
2
=
I
14, so
T(1) =
r
0
(1)
|r
0
(1)|
=
1
I
14
h1. 2. 3i =

1
I
14
.
2
I
14
.
3
I
14
. r
00
(t) = h0. 2. 6ti, so
r
0
(t) r
00
(t) =
i j k
1 2t 3t
2
0 2 6t
2t 3t
2
2 6t
i 3
1 3t
2
0 6t
j +
1 2t
0 2
k
= (12t
2
36t
2
) i 3(6t 30) j + (2 30) k =

6t
2
. 36t. 2
SECTION 14.2 DERIVATIVES AND INTEGRALS OF VECTOR FUNCTIONS ET SECTION 13.2

145
23. The vector equation for the curve is r(t) =

1 + 2
I
t. t
3
3t. t
3
+t
, so r
0
(t) =

1
I
t. 3t
2
31. 3t
2
+ 1
. The point
(3. 0. 2) corresponds to t = 1, so the tangent vector there is r
0
(1) = h1. 2. 4i. Thus, the tangent line goes through the point
(3. 0. 2) and is parallel to the vector h1. 2. 4i. Parametric equations are r = 3 +t, j = 2t, : = 2 + 4t.
25. The vector equation for the curve is r(t) =
c
3I
cos t. c
3I
sint. c
3I
, so
r
0
(t) =
c
3I
(3sint) + (cos t)(3c
3I
), c
3I
cos t + (sint)(3c
3I
), (3c
3I
)
=

3c
3I
(cos t + sint). c
3I
(cos t 3sint). 3c
3I
The point (1. 0. 1) corresponds to t = 0, so the tangent vector there is

r
0
(0) =

3c
0
(cos 0 + sin0). c
0
(cos 0 3sin0). 3c
0
= h31. 1. 31i. Thus, the tangent line is parallel to the vector

h31. 1. 31i and parametric equations are r = 1 + (31)t = 1 3t, j = 0 + 1 t = t, : = 1 + (31)t = 1 3t.
27. r(t) =
t. c
3I
. 2t 3t
2
i r
0
(t) =
1. 3c
3I
. 2 32t
. At (0. 1. 0),
t = 0 and r
0
(0) = h1. 31. 2i. Thus, parametric equations of the tangent
line are r = t, j = 1 3t, : = 2t.
29. r(t) = ht cos t. t. t sinti i r
0
(t) = hcos t 3t sint. 1. t cos t + sinti.
At (3. . 0), t = and r
0
() = h31. 1. 3i. Thus, parametric equations
of the tangent line are r = 3 3t, j = +t, : = 3t.
31. The angle of intersection of the two curves is the angle between the two tangent vectors to the curves at the point of
intersection. Since r
0
1
(t) =

1. 2t. 3t
2
and t = 0 at (0. 0. 0), r

0
1
(0) = h1. 0. 0i is a tangent vector to r
1
at (0. 0. 0). Similarly,
r
0
2
(t) = hcos t. 2 cos 2t. 1i and since r
2
(0) = h0. 0. 0i, r
0
2
(0) = h1. 2. 1i is a tangent vector to r
2
at (0. 0. 0). If 0 is the angle
between these two tangent vectors, then cos 0 =
1
I
1
I
6
h1. 0. 0i h1. 2. 1i =
1
I
6
and 0 = cos
31
1
I
6
E 66
.
33.

1
0
(16t
3
i 39t
2
j + 25t
4
k) dt =

1
0
16t
3
dt
i 3
1
0
9t
2
dt
j +
1
0
25t
4
dt
k
=

4t
4
1
0
i 3
3t
3
1
0
j +
5t
5
1
0
k = 4 i 33 j + 5 k
35.

r'2
0
(3 sin
2
t cos t i + 3 sint cos
2
t j + 2 sint cos t k) dt
=

r'2
0
3 sin
2
t cos t dt
i +
r'2
0
3 sint cos
2
t dt
j +
r'2
0
2 sint cos t dt
k
=

sin
3
t
r'2
0
i +
3cos
3
t
r'2
0
j+
sin
2
t
r'2
0
k = (1 30) i + (0 + 1) j + (1 30) k = i +j +k
37.

c
I
i + 2t j + lnt k
dt =

c
I
dt
i +
2t dt
j +
lnt dt
k
= c
I
i +t
2
j + (t lnt 3t) k +C, where Cis a vector constant of integration.
39. r
0
(t) = 2t i + 3t
2
j +
I
t k i r(t) = t
2
i +t
3
j +
2
3
t
3'2
k +C, where Cis a constant vector.
But i +j = r (1) = i +j +
2
3
k +C. Thus C = 3
2
3
k and r(t) = t
2
i +t
3
j +
2
3
t
3'2
3
2
3
k.
For Exercises 4144, let u(t) = h&
1
(t). &
2
(t). &
3
(t)i and v(t) = h
1
(t).
2
(t).
3
(t)i. In each of these exercises, the procedure is to apply
Theorem 2 so that the corresponding properties of derivatives of real-valued functions can be used.
41.
d
dt
[u(t) +v(t)] =
d
dt
h&
1
(t) +
1
(t). &
2
(t) +
2
(t). &
3
(t) +
3
(t)i
=

d
dt
[&
1
(t) +
1
(t)] .
d
dt
[&
2
(t) +
2
(t)] .
d
dt
[&
3
(t) +
3
(t)]
= h&
0
1
(t) +
0
1
(t). &
0
2
(t) +
0
2
(t). &
0
3
(t) +
0
3
(t)i
= h&
0
1
(t). &
0
2
(t) . &
0
3
(t)i + h
0
1
(t).
0
2
(t).
0
3
(t)i = u
0
(t) +v
0
(t)
43.
d
dt
[u(t) v(t)] =
d
dt
h&
2
(t)
3
(t) 3&
3
(t)
2
(t). &
3
(t)
1
(t) 3&
1
(t)
3
(t). &
1
(t)
2
(t) 3&
2
(t)
1
(t)i
= h&
0
2
3
(t) +&
2
(t)
0
3
(t) 3&
0
3
(t)
2
(t) 3&
3
(t)
0
2
(t).
&
0
3
(t)
1
(t) +&
3
(t)
0
1
(t) 3&
0
1
(t)
3
(t) 3&
1
(t)
0
3
(t).
&
0
1
(t)
2
(t) +&
1
(t)
0
2
(t) 3&
0
2
(t)
1
(t) 3&
2
(t)
0
1
(t)i
= h&
0
2
(t)
3
(t) 3&
0
3
(t)
2
(t) . &
0
3
(t)
1
(t) 3&
0
1
(t)
3
(t). &
0
1
(t)
2
(t) 3&
0
2
(t)
1
(t)i
+h&
2
(t)
0
3
(t) 3&
3
(t)
0
2
(t). &
3
(t)
0
1
(t) 3&
1
(t)
0
3
(t). &
1
(t)
0
2
(t) 3&
2
(t)
0
1
(t)i
= u
0
(t) v(t) +u(t) v
0
(t)
Alternate solution: Let r(t) = u(t) v(t). Then
r(t +/) 3r(t) = [u(t +/) v(t +/)] 3[u(t) v(t)]
= [u(t +/) v(t +/)] 3[u(t) v(t)] + [u(t +/) v(t)] 3[u(t +/) v(t)]
= u(t +/) [v(t +/) 3v(t)] + [u(t +/) 3u(t)] v(t)
(Be careful of the order of the cross product.) Dividing through by / and taking the limit as / <0 we have
r
0
(t) = lim
I<0
u(t +/) [v(t +/) 3v(t)]
/
+ lim
I<0
[u(t +/) 3u(t)] v(t)
/
= u(t) v
0
(t) +u
0
(t) v(t)
by Exercise 14.1.43(a) [ET 13.1.43(a)] and Denition 1.
45.
d
dt
[u(t) v(t)] = u
0
(t) v(t) +u(t) v
0
(t) [by Formula 4 of Theorem 3]
= hcos t. 3sint. 1i ht. cos t. sinti + hsint. cos t. ti h1. 3sint. cos ti
= t cos t 3cos t sint + sint + sint 3cos t sint +t cos t
= 2t cos t + 2 sint 32 cos t sint
SECTION 14.3 ARC LENGTH AND CURVATURE ET SECTION 13.3

147
47.
d
dt
[r(t) r
0
(t)] = r
0
(t) r
0
(t) +r(t) r
00
(t) by Formula 5 of Theorem 3. But r
0
(t) r
0
(t) = 0 (by Example 2 in
Section 13.4 [ET 12.4]). Thus,
d
dt
[r(t) r
0
(t)] = r(t) r
00
(t).
49.
d
dt
|r(t)| =
d
dt
[r(t) r(t)]
1'2
=
1
2
[r(t) r(t)]
31'2
[2r(t) r
0
(t)] =
1
|r(t)|
r(t) r
0
(t)
51. Since u(t) = r(t) [r
0
(t) r
00
(t)],
u
0
(t) = r
0
(t) [r
0
(t) r
00
(t)] +r(t)
d
dt
[r
0
(t) r
00
(t)]
= 0 +r(t) [r
00
(t) r
00
(t) +r
0
(t) r
000
(t)] [since r
0
(t) z r
0
(t) r
00
(t)]
= r(t) [r
0
(t) r
000
(t)] [since r
00
(t) r
00
(t) = 0]
14.3 Arc Length and Curvature ET 13.3
1. r(t) = h2 sint. 5t. 2 cos ti i r
0
(t) = h2 cos t. 5. 32 sinti i |r
0
(t)| =

(2 cos t)
2
+ 5
2
+ (32 sint)
2
=
I
29.
Then using Formula 3, we have 1 =

10
310
|r
0
(t)| dt =

10
310
I
29 dt =
I
29 t
10
310
= 20
I
29.
3. r(t) =
I
2 t i +c
I
j +c
3I
k i r
0
(t) =
I
2 i +c
I
j 3c
3I
k i
|r
0
(t)| =

I
2
2
+ (c
I
)
2
+ (3c
3I
)
2
=
I
2 +c
2I
+c
32I
=

(c
I
+c
3I
)
2
= c
I
+c
3I
[since c
I
+c
3I
0].
Then 1 =

1
0
|r
0
(t)| dt =

1
0
(c
I
+c
3I
) dt =

c
I
3c
3I
1
0
= c 3c
31
.
5. r(t) = i +t
2
j +t
3
k i r
0
(t) = 2t j + 3t
2
k i |r
0
(t)| =
I
4t
2
+ 9t
4
= t
I
4 + 9t
2
[since t D 0].
Then 1 =

1
0
|r
0
(t)| dt =

1
0
t
I
4 + 9t
2
dt =
1
18

2
3
(4 + 9t
2
)
3'2
1
0
=
1
27
(13
3'2
34
3'2
) =
1
27
(13
3'2
38).
7. r(t) =
I
t. t. t
2
i r
0
(t) =

1
2
I
t
. 1. 2t
i |r
0
(t)| =
1
2
I
I
2
+ 1
2
+ (2t)
2
=

1
4I
+ 1 + 4t
2
, so
1 =

4
1
|r
0
(t)| dt =

4
1
1
4I
+ 1 + 4t
2
dt E 15.3841.
9. r(t) = hsint. cos t. tanti i r
0
(t) =

cos t. 3sint. sec
2
t
i
|r
0
(t)| =

cos
2
t + (3sint)
2
+ (sec
2
t)
2
=
I
1 + sec
4
t and 1 =

r'4
0
|r
0
(t)| dt =

r'4
0
I
1 + sec
4
t dt E 1.2780.
11. The projection of the curve C onto the rj-plane is the curve r
2
= 2j or j =
1
2
r
2
, : = 0. Then we can choose the parameter
r = t i j =
1
2
t
2
. Since C also lies on the surface 3: = rj, we have : =
1
3
rj =
1
3
(t)(
1
2
t
2
) =
1
6
t
3
. Then parametric
equations for C are r = t, j =
1
2
t
2
, : =
1
6
t
3
and the corresponding vector equation is r(t) =

t.
1
2
t
2
.
1
6
t
3
. The origin
corresponds to t = 0 and the point (6. 18. 36) corresponds to t = 6, so
1 =

6
0
|r
0
(t)| dt =

6
0
1. t.
1
2
t
2
dt =

6
0
1
2
+t
2
+
1
2
t
2
2
dt =

6
0
1 +t
2
+
1
4
t
4
dt
=

6
0
(1 +
1
2
t
2
)
2
dt =

6
0
(1 +
1
2
t
2
) dt =

t +
1
6
t
3
6
0
= 6 + 36 = 42
13. r(t) = 2t i + (1 33t) j + (5 + 4t) k i r
0
(t) = 2 i 33 j + 4 k and
us
uI
= |r
0
(t)| =
I
4 + 9 + 16 =
I
29. Then
c = c(t) =

I
0
|r
0
(&)| d& =

I
0
I
29 d& =
I
29 t. Therefore, t =
1
I
29
c, and substituting for t in the original equation, we
have r(t(c)) =
2
I
29
c i +
1 3
3
I
29
c
j +
5 +
4
I
29
c
k.
15. Here r(t) = h3 sint. 4t. 3 cos ti, so r
0
(t) = h3 cos t. 4. 33 sinti and |r
0
(t)| =

9 cos
2
t + 16 + 9 sin
2
t =
I
25 = 5.
The point (0. 0. 3) corresponds to t = 0, so the arc length function beginning at (0. 0. 3) and measuring in the positive
direction is given by c(t) =

I
0
|r
0
(&)| d& =

I
0
5 d& = 5t. c(t) = 5 i 5t = 5 i t = 1, thus your location after
moving 5 units along the curve is (3 sin1. 4. 3 cos 1).
17. (a) r(t) = h2 sint. 5t. 2 cos ti i r
0
(t) = h2 cos t. 5. 32 sinti i |r
0
(t)| =

4 cos
2
t + 25 + 4 sin
2
t =
I
29.
Then T(t) =
r
0
(t)
|r
0
(t)|
=
1
I
29
h2 cos t. 5. 32 sinti or
2
I
29
cos t.
5
I
29
. 3
2
I
29
sint
.
T
0
(t) =
1
I
29
h32 sint. 0. 32 cos ti i |T
0
(t)| =
1
I
29
4 sin
2
t + 0 + 4 cos
2
t =
2
I
29
. Thus
N(t) =
T
0
(t)
|T
0
(t)|
=
1
I
29
2
I
29
h32 sint. 0. 32 cos ti = h3sint. 0. 3cos ti.
(b) r(t) =
|T
0
(t)|
|r
0
(t)|
=
2
I
29
I
29
=
2
29
19. (a) r(t) =
I
2 t. c
I
. c
3I
i r
0
(t) =
I
2. c
I
. 3c
3I
i |r
0
(t)| =
I
2 +c
2I
+c
32I
=

(c
I
+c
3I
)
2
= c
I
+c
3I
.
Then
T(t) =
r
0
(t)
|r
0
(t)|
=
1
c
I
+c
3I
I
2. c
I
. 3c
3I
=
1
c
2I
+ 1
I
2c
I
. c
2I
. 31

after multiplying by
c
I
c
I
and
T
0
(t) =
1
c
2I
+ 1
I
2c
I
. 2c
2I
. 0
3
2c
2I
(c
2I
+ 1)
2
I
2c
I
. c
2I
. 31
=
1
(c
2I
+ 1)
2
(c
2I
+ 1)
I
2c
I
. 2c
2I
. 0
32c
2I
I
2c
I
. c
2I
. 31
=
1
(c
2I
+ 1)
2
I
2c
I
1 3c
2I
. 2c
2I
. 2c
2I
Then
|T
0
(t)| =
1
(c
2I
+ 1)
2
2c
2I
(1 32c
2I
+c
4I
) + 4c
4I
+ 4c
4I
=
1
(c
2I
+ 1)
2
2c
2I
(1 + 2c
2I
+c
4I
)
=
1
(c
2I
+ 1)
2
2c
2I
(1 +c
2I
)
2
=
I
2c
I
(1 +c
2I
)
(c
2I
+ 1)
2
=
I
2 c
I
c
2I
+ 1
Therefore
N(t) =
T
0
(t)
|T
0
(t)|
=
c
2I
+ 1
I
2 c
I
1
(c
2I
+ 1)
2
I
2 c
I
(1 3c
2I
). 2c
2I
. 2c
2I
=
1
I
2 c
I
(c
2I
+ 1)
I
2 c
I
(1 3c
2I
). 2c
2I
. 2c
2I
=
1
c
2I
+ 1
1 3c
2I
.
I
2 c
I
.
I
2 c
I
(b) r(t) =
|T
0
(t)|
|r
0
(t)|
=
I
2 c
I
c
2I
+ 1

1
c
I
+c
3I
=
I
2 c
I
c
3I
+ 2c
I
+c
3I
=
I
2 c
2I
c
4I
+ 2c
2I
+ 1
=
I
2 c
2I
(c
2I
+ 1)
2

149
21. r(t) = t
2
i +t k i r
0
(t) = 2t i +k, r
00
(t) = 2 i, |r
0
(t)| =

(2t)
2
+ 0
2
+ 1
2
=
I
4t
2
+ 1, r
0
(t) r
00
(t) = 2 j,
|r
0
(t) r
00
(t)| = 2. Then r(t) =
|r
0
(t) r
00
(t)|
|r
0
(t)|
3
=
2
I
4t
2
+ 1
3
=
2
(4t
2
+ 1)
3'2
.
23. r(t) = 3t i + 4 sint j + 4 cos t k i r
0
(t) = 3 i + 4 cos t j 34 sint k, r
00
(t) = 34 sint j 34 cos t k,
|r
0
(t)| =

9 + 16 cos
2
t + 16 sin
2
t =
I
9 + 16 = 5, r
0
(t) r
00
(t) = 316 i + 12 cos t j 312 sint k,
|r
0
(t) r
00
(t)| =

256 + 144 cos
2
t + 144 sin
2
t =
I
400 = 20. Then r(t) =
|r
0
(t) r
00
(t)|
|r
0
(t)|
3
=
20
5
3
=
4
25
.
25. r(t) =

t. t
2
. t
3
i r
0
(t) =

1. 2t. 3t
2
. The point (1. 1. 1) corresponds to t = 1, and r

0
(1) = h1. 2. 3i i
|r
0
(1)| =
I
1 + 4 + 9 =
I
14. r
00
(t) = h0. 2. 6ti i r
00
(1) = h0. 2. 6i. r
0
(1) r
00
(1) = h6. 36. 2i, so
|r
0
(1) r
00
(1)| =
I
36 + 36 + 4 =
I
76. Then r(1) =
|r
0
(1) r
00
(1)|
|r
0
(1)|
3
=
I
76
I
14
3
=
1
7
19
14
.
27. 1(r) = 2r 3 r
2
, 1
0
(r) = 2 3 2r, 1
00
(r) = 32,
r(r) =
|1
00
(r)|
[1 + (1
0
(r))
2
]
3'2
=
|32|
[1 + (2 32r)
2
]
3'2
=
2
(4r
2
38r + 5)
3'2
29. 1(r) = 4r
5'2
, 1
0
(r) = 10r
3'2
, 1
00
(r) = 15r
1'2
,
r(r) =
|1
00
(r)|
[1 + (1
0
(r))
2
]
3'2
=
15r
1'2
[1 + (10r
3'2
)
2
]
3'2
=
15
I
r
(1 + 100r
3
)
3'2
31. Since j
0
= j
00
= c
i
, the curvature is r(r) =
|j
00
(r)|
[1 + (j
0
(r))
2
]
3'2
=
c
i
(1 +c
2i
)
3'2
= c
i
(1 +c
2i
)
33'2
.
To nd the maximum curvature, we rst nd the critical numbers of r(r):
r
0
(r) = c
i
(1 +c
2i
)
33'2
+c
i
3
3
2
(1 +c
2i
)
35'2
(2c
2i
) = c
i
1 +c
2i
33c
2i
(1 +c
2i
)
5'2
= c
i
1 32c
2i
(1 +c
2i
)
5'2
.
r
0
(r) = 0 when 1 32c
2i
= 0, so c
2i
=
1
2
or r = 3
1
2
ln2. And since 1 32c
2i
0 for r < 3
1
2
ln2 and
1 32c
2i
< 0 for r 3
1
2
ln2, the maximum curvature is attained at the point
3
1
2
ln2. c
(3ln 2)'2
=

3
1
2
ln2.
1
I
2
.
Since lim
i<"
c
i
(1 +c
2i
)
33'2
= 0. r(r) approaches 0 as r <".
33. (a) C appears to be changing direction more quickly at 1 than Q, so we would expect the curvature to be greater at 1.
(b) First we sketch approximate osculating circles at 1 and Q. Using the
axes scale as a guide, we measure the radius of the osculating circle
at 1 to be approximately 0.8 units, thus a =
1
r
i
r =
1
a
E
1
0.8
E 1.3. Similarly, we estimate the radius of the
osculating circle at Q to be 1.4 units, so r =
1
a
E
1
1.4
E 0.7.
35. j = r
32
i j
0
= 32r
33
, j
00
= 6r
34
, and
r(r) =
|j
00
|
[1 + (j
0
)
2
]
3'2
=
6r
34
[1 + (32r
33
)
2
]
3'2
=
6
r
4
(1 + 4r
36
)
3'2
.
The appearance of the two humps in this graph is perhaps a little surprising, but it is
explained by the fact that j = r
32
increases asymptotically at the origin from both
directions, and so its graph has very little bend there. (Note that r(0) is undened.)
37. Notice that the curve / has two inection points at which the graph appears almost straight. We would expect the curvature to
be 0 or nearly 0 at these values, but the curve o isnt near 0 there. Thus, o must be the graph of j = 1(r) rather than the graph
of curvature, and / is the graph of j = r(r).
39. Using a CAS, we nd (after simplifying)
r(t) =
6
I
4 cos
2
t 312 cos t + 13
(17 312 cos t)
3'2
. (To compute cross
products in Maple, use the VectorCalculus package and
the CrossProduct(a,b) command; in Mathematica, use
Cross[a,b].) Curvature is largest at integer multiples of 2.
41. r = c
I
cos t i r = c
I
(cos t 3sint) i r = c
I
(3sint 3cos t) +c
I
(cos t 3sint) = 32c
I
sint,
j = c
I
sint i j = c
I
(cos t + sint) i j = c
I
(3sint + cos t) +c
I
(cos t + sint) = 2c
I
cos t. Then
r(t) =
| r j 3 j r|
[ r
2
+ j
2
]
3'2
=
c
I
(cos t 3sint)(2c
I
cos t) 3c
I
(cos t + sint)(32c
I
sint)
([c
I
(cos t 3sint)]
2
+ [c
I
(cos t + sint)]
2
)
3'2
=
2c
2I
(cos
2
t 3sint cos t + sint cos t + sin
2
t)
c
2I
(cos
2
2
t + cos
2
t + 2 cos t sint + sin
2
t)
3'2
=
2c
2I
(1)
[c
2I
(1 + 1)]
3'2
=
2c
2I
c
3I
(2)
3'2
=
1
I
2 c
I
43.

1.
2
3
. 1
corresponds to t = 1. T(t) =
r
0
(t)
|r
0
(t)|
=
2t. 2t
2
. 1
I
4t
2
+ 4t
4
+ 1
=
2t. 2t
2
. 1
2t
2
+ 1
, so T(1) =
2
3
.
2
3
.
1
3
.
T
0
(t) = 34t(2t
2
+ 1)
32
2t. 2t
2
. 1
+ (2t
2
+ 1)
31
h2. 4t. 0i (by Formula 3 of Theorem 14.2 [ET 13.2])
= (2t
2
+ 1)
32
38t
2
+ 4t
2
+ 2. 38t
3
+ 8t
3
+ 4t. 34t
= 2(2t
2
+ 1)
32
1 32t
2
. 2t. 32t
N(t) =
T
0
(t)
|T
0
(t)|
=
2(2t
2
+ 1)
32
1 32t
2
. 2t. 32t
2(2t
2
+ 1)
32
(1 32t
2
)
2
+ (2t)
2
+ (32t)
2
=
1 32t
2
. 2t. 32t
I
1 34t
2
+ 4t
4
+ 8t
2
=
1 32t
2
. 2t. 32t
1 + 2t
2
N(1) =

3
1
3
.
2
3
. 3
2
3
and B(1) = T(1) N(1) =

3
4
9
3
2
9
. 3
3
4
9
+
1
9
.
4
9
+
2
9
=

3
2
3
.
1
3
.
2
3
.
45. (0. . 32) corresponds to t = . r(t) = h2 sin3t. t. 2 cos 3ti i
T(t) =
r
0
(t)
|r
0
(t)|
=
h6 cos 3t. 1. 36 sin3ti
36 cos
2
3t + 1 + 36 sin
2
3t
=
1
I
37
h6 cos 3t. 1. 36 sin3ti.
T() =
1
I
37
h36. 1. 0i is a normal vector for the normal plane, and so h36. 1. 0i is also normal. Thus an equation for the

151
plane is 36 (r 30) + 1(j 3) + 0(: + 2) = 0 or j 36r = .
T
0
(t) =
1
I
37
h318 sin3t. 0. 318 cos 3ti i |T
0
(t)| =
18
2
sin
2
3t + 18
2
cos
2
3t
I
37
=
18
I
37
i
N(t) =
T
0
(t)
|T
0
(t)|
= h3sin3t. 0. 3cos 3ti. So N() = h0. 0. 1i and B() =
1
I
37
h36. 1. 0i h0. 0. 1i =
1
I
37
h1. 6. 0i.
Since B() is a normal to the osculating plane, so is h1. 6. 0i.
An equation for the plane is 1(r 30) + 6(j 3) + 0(: + 2) = 0 or r + 6j = 6.
47. The ellipse 9r
2
+ 4j
2
= 36 is given by the parametric equations r = 2 cos t, j = 3 sint, so using the result from
Exercise 40,
r(t) =
| r j 3 r j|
[ r
2
+ j
2
]
3'2
=
|(32 sint)(33 sint) 3(3 cos t)(32 cos t)|
(4 sin
2
t + 9 cos
2
t)
3'2
=
6
(4 sin
2
t + 9 cos
2
t)
3'2
.
At (2. 0), t = 0. Now r(0) =
6
27
=
2
9
, so the radius of the osculating circle is
1r(0) =
9
2
and its center is
3
5
2
. 0
. Its equation is therefore
r +
5
2
2
+j
2
=
81
4
.
At (0. 3), t =
r
2
, and r
r
2
=
6
8
=
3
4
. So the radius of the osculating circle is
4
3
and
its center is
0.
5
3
. Hence its equation is r

2
+
j 3
5
3
2
=
16
9
.
49. The tangent vector is normal to the normal plane, and the vector h6. 6. 38i is normal to the given plane.
But T(t) k r
0
(t) and h6. 6. 38i k h3. 3. 34i, so we need to nd t such that r
0
(t) k h3. 3. 34i.
r(t) =
t
3
. 3t. t
4
i r
0
(t) =
3t
2
. 3. 4t
3
k h3. 3. 34i when t = 31. So the planes are parallel at the point (31. 33. 1).
51. r =
dT
dc
dTdt
dcdt
=
|dTdt|
dcdt
and N =
dTdt
|dTdt|
, so rN =
dT
dt
dT
dt
dT
dt
dc
dt
=
dTdt
dcdt
=
dT
dc
by the Chain Rule.
53. (a) |B| = 1 i B B = 1 i
d
dc
(B B) = 0 i 2
dB
dc
B = 0 i
dB
dc
z B
(b) B = TN i
dB
dc
=
d
dc
(TN) =
d
dt
(TN)
1
dcdt
=
d
dt
(TN)
1
|r
0
(t)|
= [(T
0
N) + (TN
0
)]
1
|r
0
(t)|
=

T
0

T
0
|T
0
|
+ (TN
0
)
1
|r
0
(t)|
=
TN
0
|r
0
(t)|
i
dB
dc
z T
(c) B = TN i T z N, B z Tand B z N. So B, T and Nform an orthogonal set of vectors in the three-
dimensional space R
3
. From parts (a) and (b), dBdc is perpendicular to both Band T, so dBdc is parallel to N.
Therefore, dBdc = 3t(c)N, where t(c) is a scalar.
(d) Since B = TN, T z Nand both Tand Nare unit vectors, Bis a unit vector mutually perpendicular to both Tand
N. For a plane curve, T and Nalways lie in the plane of the curve, so that Bis a constant unit vector always
perpendicular to the plane. Thus dBdc = 0, but dBdc = 3t(c)Nand N 6= 0, so t(c) = 0.
55. (a) r
0
= c
0
T i r
00
= c
00
T+c
0
T
0
= c
00
T+c
0
dT
dc
c
0
= c
00
T+r(c
0
)
2
Nby the rst Serret-Frenet formula.
(b) Using part (a), we have
r
0
r
00
= (c
0
T) [c
00
T+r(c
0
)
2
N]
= [(c
0
T) (c
00
T)] +
(c
0
T) (r(c
0
)
2
N)
(by Property 3 of Theorem 13.4.8 [ ET 12.4.8])

= (c
0
c
00
)(TT) +r(c
0
)
3
(TN) = 0 +r(c
0
)
3
B = r(c
0
)
3
B
(c) Using part (a), we have
r
000
= [c
00
T+r(c
0
)
2
N]
0
= c
000
T+c
00
T
0
+r
0
(c
0
)
2
N+ 2rc
0
c
00
N+r(c
0
)
2
N
0
= c
000
T+c
00
dT
dc
c
0
+r
0
(c
0
)
2
N+ 2rc
0
c
00
N+r(c
0
)
2
dN
dc
c
0
= c
000
T+c
00
c
0
rN+r
0
(c
0
)
2
N+ 2rc
0
c
00
N+r(c
0
)
3
(3rT+t B) [by the second formula]
= [c
000
3r
2
(c
0
)
3
] T+ [3rc
0
c
00
+r
0
(c
0
)
2
] N+rt(c
0
)
3
B
(d) Using parts (b) and (c) and the facts that B T = 0, B N = 0, and B B = 1, we get
(r
0
r
00
) r
000
|r
0
r
00
|
2
=
r(c
0
)
3
B
[c
000
3r
2
(c
0
)
3
] T+ [3rc
0
c
00
+r
0
(c
0
)
2
] N+rt(c
0
)
3
B
|r(c
0
)
3
B|
2
=
r(c
0
)
3
rt(c
0
)
3
[r(c
0
)
3
]
2
= t.
57. r =

t.
1
2
t
2
.
1
3
t
3
i r
0
=

1. t. t
2
, r
00
= h0. 1. 2ti, r
000
= h0. 0. 2i i r
0
r
00
=

t
2
. 32t. 1
i
t =
(r
0
r
00
) r
000
|r
0
r
00
|
2
=
t
2
. 32t. 1
h0. 0. 2i
t
4
+ 4t
2
+ 1
=
2
t
4
+ 4t
2
+ 1
59. For one helix, the vector equation is r(t) = h10 cos t. 10 sint. 34t(2)i (measuring in angstroms), because the radius of each
helix is 10 angstroms, and : increases by 34 angstroms for each increase of 2 in t. Using the arc length formula, letting t go
from 0 to 2.9 10
8
2, we nd the approximate length of each helix to be
1=

2.910
8
2r
0
|r
0
(t)| dt =

2.910
8
2r
0
(310 sint)
2
+ (10 cos t)
2
+
34
2r
2
dt =

100 +
34
2r
2.910
8
2r
= 2.9 10
8
2
100 +
34
2r
2
E 2.07 10
10
more than two meters!
14.4 Motion in Space: Velocity and Acceleration ET 13.4
1. (a) If r(t) = r(t) i +j (t) j +:(t) k is the position vector of the particle at time t, then the average velocity over the time
interval [0. 1] is
v
ave
=
r(1) 3r(0)
1 30
=
(4.5 i + 6.0 j + 3.0 k) 3(2.7 i + 9.8 j + 3.7 k)
1
= 1.8 i 33.8 j 30.7 k. Similarly, over the other
intervals we have
[0.5. 1] : v
ave
=
r(1) 3r(0.5)
1 30.5
=
(4.5 i + 6.0 j + 3.0 k) 3(3.5 i + 7.2 j + 3.3 k)
0.5
= 2.0 i 32.4 j 30.6 k
[1. 2] : v
ave
=
r(2) 3r(1)
2 31
=
(7.3 i + 7.8 j + 2.7 k) 3(4.5 i + 6.0 j + 3.0 k)
1
= 2.8 i + 1.8 j 30.3 k
[1. 1.5] : v
ave
=
r(1.5) 3r(1)
1.5 31
=
(5.9 i + 6.4 j + 2.8 k) 3(4.5 i + 6.0 j + 3.0 k)
0.5
= 2.8 i + 0.8 j 30.4 k
SECTION 14.4 MOTION IN SPACE: VELOCITY AND ACCELERATION ET SECTION 13.4

153
(b) We can estimate the velocity at t = 1 by averaging the average velocities over the time intervals [0.5. 1] and [1. 1.5]:
v(1) E
1
2
[(2 i 32.4 j 30.6 k) + (2.8 i + 0.8 j 30.4 k)] = 2.4 i 30.8 j 30.5 k. Then the speed is
|v(1)| E
(2.4)
2
+ (30.8)
2
+ (30.5)
2
E 2.58.
3. r(t) =
3
1
2
t
2
. t
i At t = 2:
v(t) = r
0
(t) = h3t. 1i v(2) = h32. 1i
a(t) = r
00
(t) = h31. 0i a(2) = h31. 0i
|v(t)| =
I
t
2
+ 1
5. (t) = 3 cos t i + 2 sint j i At t = 3:
v(t) = 33 sint i + 2 cos t j v
r
3
= 3
3
I
3
2
i +j
a(t) = 33 cos t i 32 sint j a
r
3
= 3
3
2
i 3
I
3 j
|v(t)| =

9 sin
2
t + 4 cos
2
t =

4 + 5 sin
2
t
Notice that r
2
9 +j
2
4 = sin
2
t + cos
2
t = 1, so the path is an ellipse.
7. r(t) = t i +t
2
j + 2 k i At t = 1:
v(t) = i + 2t j v(1) = i + 2 j
a(t) = 2 j a(1) = 2 j
|v(t)| =
I
1 + 4t
2
Here r = t, j = t
2
i j = r
2
and : = 2, so the path of the particle is a
parabola in the plane : = 2.
9. r(t) =

t
2
+ 1. t
3
. t
2
31
i v(t) = r
0
(t) =

2t. 3t
2
. 2t
, a(t) = v
0
(t) = h2. 6t. 2i,
|v(t)| =

(2t)
2
+ (3t
2
)
2
+ (2t)
2
=
I
9t
4
+ 8t
2
= |t|
I
9t
2
+ 8.
11. r(t) =
I
2 t i +c
I
j +c
3I
k i v(t) = r
0
(t) =
I
2 i +c
I
j 3c
3I
k, a(t) = v
0
(t) = c
I
j +c
3I
k,
|v(t)| =
I
2 +c
2I
+c
32I
=

(c
I
+c
3I
)
2
= c
I
+c
3I
.
13. r(t) = c
I
hcos t. sint. ti i
v(t) = r
0
(t) = c
I
hcos t. sint. ti +c
I
h3sint. cos t. 1i = c
I
hcos t 3sint. sint + cos t. t + 1i
a(t) = v
0
(t) = c
I
hcos t 3sint 3sint 3cos t. sint + cos t + cos t 3sint. t + 1 + 1i
= c
I
h32 sint. 2 cos t. t + 2i
|v(t)| = c
I
cos
2
t + sin
2
2
t + cos
2
t + 2 sint cos t +t
2
+ 2t + 1
= c
I
I
t
2
+ 2t + 3
15. a(t) = i + 2 j i v(t) =
a(t) dt =

(i + 2 j) dt = t i + 2t j +Cand k = v(0) = C,
so C = k and v(t) = t i + 2t j +k. r(t) =
v(t) dt =

(t i + 2t j +k) dt =
1
2
t
2
i +t
2
j +t k +D.
But i = r (0) = D, so D = i and r(t) =
1
2
t
2
+ 1
i +t
2
j +t k.
17. (a) a(t) = 2t i + sint j + cos 2t k i
v(t) =
(2t i + sint j + cos 2t k) dt = t

2
i 3cos t j +
1
2
sin2t k +C
and i = v(0) = 3j +C, so C = i +j
and v(t) =
t
2
+ 1
i + (1 3cos t) j +
1
2
sin2t k.
r(t) =
t
2
+ 1
i + (1 3cos t) j +
1
2
sin2t k]dt
=

1
3
t
3
+t
i + (t 3sint) j 3
1
4
cos 2t k +D
But j = r (0) = 3
1
4
k +D, so D = j +
1
4
k and r(t) =
1
3
t
3
+t
i + (t 3sint + 1) j +
1
4
3
1
4
cos 2t
k.
(b)
19. r(t) =

t
2
. 5t. t
2
316t
i v(t) = h2t. 5. 2t 316i, |v(t)| =

I
4t
2
+ 25 + 4t
2
364t + 256 =
I
8t
2
364t + 281
and
d
dt
|v(t)| =
1
2
(8t
2
364t + 281)
31'2
(16t 364). This is zero if and only if the numerator is zero, that is,
16t 364 = 0 or t = 4. Since
d
dt
|v(t)| < 0 for t < 4 and
d
dt
|v(t)| 0 for t 4, the minimum speed of
I
153 is attained
at t = 4 units of time.
21. |F(t)| = 20 N in the direction of the positive :-axis, so F(t) = 20 k. Also i = 4 kg, r(0) = 0 and v(0) = i 3j.
Since 20k = F(t) = 4 a(t), a(t) = 5 k. Then v(t) = 5t k +c
1
where c
1
= i 3j so v(t) = i 3j + 5t k and the
speed is |v(t)| =
I
1 + 1 + 25t
2
=
I
25t
2
+ 2. Also r(t) = t i 3t j +
5
2
t
2
k +c
2
and 0 = r(0), so c
2
= 0
and r(t) = t i 3t j +
5
2
t
2
k.
23. |v(0)| = 500 ms and since the angle of elevation is 30
, the direction of the velocity is

1
2
I
3 i +j
. Thus
v(0) = 250
I
3 i +j
and if we set up the axes so the projectile starts at the origin, then r(0) = 0. Ignoring air resistance, the
only force is that due to gravity, so F(t) = 3io j where o E 9.8 ms
2
. Thus a(t) = 3o j and v(t) = 3ot j +c
1
. But
250
I
3 i +j
= v(0) = c
1
, so v(t) = 250
I
3 i + (250 3ot) j and r(t) = 250
I
3 t i +
250t 3
1
2
ot
2
j +c
2
where
0 = r(0) = c
2
. Thus r(t) = 250
I
3 t i +
250t 3
1
2
ot
2
j.
(a) Setting 250t 3
1
2
ot
2
= 0 gives t = 0 or t =
500
g
E 51.0 s. So the range is 250
I
3
500
g
E 22 km.
(b) 0 =
d
dt
250t 3
1
2
ot
2
= 250 3ot implies that the maximum height is attained when t = 250o E 25.5 s.
Thus, the maximum height is (250)(250o) 3o(250o)
2 1
2
= (250)
2
(2o) E 3.2 km.
(c) From part (a), impact occurs at t = 500o E 51.0. Thus, the velocity at impact is
v(500o) = 250
I
3 i + [250 3o(500o)] j = 250
I
3 i 3250 j and the speed is |v(500o)| = 250
I
3 + 1 = 500 ms.
SECTION 14.4 MOTION IN SPACE: VELOCITY AND ACCELERATION ET SECTION 13.4

155
25. As in Example 5, r(t) = (
0
cos 45
)t i +
(
0
sin45
)t 3
1
2
ot
2
j =
1
2
0
I
2 t i +
0
I
2 t 3ot
2
. Then the ball

lands at t =

0
I
2
o
s. Now since it lands 90 m away, 90 =
1
2
0
I
2

0
I
2
o
or
2
0
= 90o and the initial velocity
is
0
=
I
90o E 30 ms.
27. Let c be the angle of elevation. Then
0
= 150 ms and from Example 5, the horizontal distance traveled by the projectile is
d =

2
0
sin2c
o
. Thus
150
2
sin2c
o
= 800 i sin2c =
800o
150
2
E 0.3484 i 2c E 20.4
or 180 320.4 = 159.6
.
Two angles of elevation then are c E 10.2
and c E 79.8
.
29. Place the catapult at the origin and assume the catapult is 100 meters from the city, so the city lies between (100. 0)
and (600. 0). The initial speed is
0
= 80 ms and let 0 be the angle the catapult is set at. As in Example 5, the trajectory of
the catapulted rock is given by r (t) = (80 cos 0)t i +
(80 sin0)t 34.9t

2
j. The top of the near city wall is at (100. 15),

which the rock will hit when (80 cos 0) t = 100 i t =
5
4 cos 0
and (80 sin0)t 34.9t
2
= 15 i
80 sin0
5
4 cos 0
34.9
5
4 cos 0
2
= 15 i 100 tan0 37.65625 sec
2
0 = 15. Replacing sec
2
0 with tan
2
0 + 1 gives
7.65625 tan
2
0 3100 tan0 + 22.62625 = 0. Using the quadratic formula, we have tan0 E 0.230324, 12.8309 i
0 E 13.0
, 85.5
. So for 13.0
< 0 < 85.5
, the rock will land beyond the near city wall. The base of the far wall is
located at (600. 0) which the rock hits if (80 cos 0)t = 600 i t =
15
2 cos 0
and (80 sin0)t 34.9t
2
= 0 i
80 sin0
15
2 cos 0
3 4.9
15
2 cos 0
2
= 0 i 600 tan0 3275.625 sec
2
0 = 0 i
275.625 tan
2
0 3600 tan0 + 275.625 = 0. Solutions are tan0 E 0.658678, 1.51819 i 0 E 33.4
, 56.6
. Thus the
rock lands beyond the enclosed city ground for 33.4
< 0 < 56.6
, and the angles that allow the rock to land on city ground
are 13.0
< 0 < 33.4
, 56.6
< 0 < 85.5
. If you consider that the rock can hit the far wall and bounce back into the city, we
calculate the angles that cause the rock to hit the top of the wall at (600. 15): (80 cos 0)t = 600 i t =
15
2 cos 0
and
(80 sin0)t 34.9t
2
= 15 i 600 tan0 3275.625 sec
2
0 = 15 i 275.625 tan
2
0 3600 tan0 + 290.625 = 0.
Solutions are tan0 E 0.727506, 1.44936 i 0 E 36.0
, 55.4
, so the catapult should be set with angle 0 where

13.0
< 0 < 36.0
, 55.4
< 0 < 85.5
.
31. (a) After t seconds, the boat will be 5t meters west of point . The velocity
of the water at that location is
3
400
(5t)(40 35t) j. The velocity of the
boat in still water is 5 i. so the resultant velocity of the boat is
v(t) = 5 i +
3
400
(5t)(40 35t) j = 5i +
3
2
t 3
3
16
t
2
j. Integrating, we obtain
r(t) = 5t i +
3
4
t
2
3
1
16
t
3
j +C. If we place the origin at (and consider j

to coincide with the northern direction) then r(0) = 0 i C = 0 and we have r(t) = 5t i +
3
4
t
2
3
1
16
t
3
j. The boat
reaches the east bank after 8 s, and it is located at r(8) = 5(8)i +
3
4
(8)
2
3
1
16
(8)
3
j = 40 i +16 j. Thus the boat is 16 m

downstream.
(b) Let c be the angle north of east that the boat heads. Then the velocity of the boat in still water is given by
5(cos c) i + 5(sinc) j. At t seconds, the boat is 5(cos c)t meters from the west bank, at which point the velocity
of the water is
3
400
[5(cos c)t][40 35(cos c)t] j. The resultant velocity of the boat is given by
v(t) = 5(cos c) i +
5 sinc +
3
400
(5t cos c)(40 35t cos c)
j = (5 cos c) i +
5 sinc +
3
2
t cos c 3
3
16
t
2
cos
2
c
j.
Integrating, r(t) = (5t cos c) i +
5t sinc +
3
4
t
2
cos c 3
1
16
t
3
cos
2
c
j (where we have again placed

the origin at ). The boat will reach the east bank when 5t cos c = 40 i t =
40
5 cos c
=
8
cos c
.
In order to land at point 1(40. 0) we need 5t sinc +
3
4
t
2
cos c 3
1
16
t
3
cos
2
c = 0 i
5
8
cos c
sinc +
3
4
8
cos c
2
cos c 3
1
16
8
cos c
3
cos
2
c = 0 i
1
cos c
(40 sinc + 48 332) = 0 i
40 sinc + 16 = 0 i sinc = 3
2
5
. Thus c = sin
31
3
2
5
E 323.6
, so the boat should head 23.6
south of
east (upstream). The path does seem realistic. The boat initially heads
upstream to counteract the effect of the current. Near the center of the river,
the current is stronger and the boat is pushed downstream. When the boat
nears the eastern bank, the current is slower and the boat is able to progress
upstream to arrive at point 1.
33. r(t) = (3t 3 t
3
) i + 3t
2
j i r
0
(t) = (3 3 3t
2
) i + 6t j,
|r
0
(t)| =

(3 33t
2
)
2
+ (6t)
2
=
I
9 + 18t
2
+ 9t
4
=

(3 33t
2
)
2
= 3 + 3t
2
,
r
00
(t) = 36t i + 6 j, r
0
(t) r
00
(t) = (18 + 18t
2
) k. Then Equation 9 gives
o
T
=
r
0
(t) r
00
(t)
|r
0
(t)|
=
(3 33t
2
)(36t) + (6t)(6)
3 + 3t
2
=
18t + 18t
3
3 + 3t
2
=
18t(1 +t
2
)
3(1 +t
2
)
= 6t

or by Equation 8,
o
T
=
0
=
d
dt
3 + 3t
2
= 6t
and Equation 10 gives o

^
=
|r
0
(t) r
00
(t)|
|r
0
(t)|
=
18 + 18t
2
3 + 3t
2
=
18(1 +t
2
)
3(1 +t
2
)
= 6.
35. r(t) = cos t i + sint j +t k i r
0
(t) = 3sint i + cos t j +k, |r
0
(t)| =

sin
2
t + cos
2
t + 1 =
I
2,
r
00
(t) = 3cos t i 3sint j, r
0
(t) r
00
(t) = sint i 3cos t j +k.
Then o
T
=
r
0
(t) r
00
(t)
|r
0
(t)|
=
sint cos t 3sint cos t
I
2
= 0 and o
^
=
|r
0
(t) r
00
(t)|
|r
0
(t)|
=
sin
2
t + cos
2
t + 1
I
2
=
I
2
I
2
= 1.
37. r(t) = c
I
i +
I
2 t j +c
3I
k i r
0
(t) = c
I
i +
I
2 j 3c
3I
k, |r(t)| =
I
c
2I
+ 2 +c
32I
=

(c
I
+c
3I
)
2
= c
I
+c
3I
,
r
00
(t) = c
I
i +c
3I
k. Then o
T
=
c
2I
3c
32I
c
I
+c
3I
=
(c
I
+c
3I
)(c
I
3c
3I
)
c
I
+c
3I
= c
I
3c
3I
= 2 sinht
and o
^
=
I
2c
3I
i 32 j 3
I
2c
I
k
c
I
+c
3I
=
2(c
32I
+ 2 +c
2I
)
c
I
+c
3I
=
I
2
c
I
+c
3I
c
I
+c
3I
=
I
2.
39. The tangential component of a is the length of the projection of a onto T, so we sketch
the scalar projection of a in the tangential direction to the curve and estimate its length to
be 4.5 (using the fact that a has length 10 as a guide). Similarly, the normal component of
a is the length of the projection of a onto N, so we sketch the scalar projection of a in the
normal direction to the curve and estimate its length to be 9.0. Thus o
T
E 4.5 cms
2
and
o
^
E 9.0 cms
2
.
41. If the engines are turned off at time t, then the spacecraft will continue to travel in the direction of v(t), so we need a t such
that for some scalar c 0, r(t) +c v(t) = h6. 4. 9i. v(t) = r
0
(t) = i +
1
t
j +
8t
(t
2
+ 1)
2
k i
r(t) +c v(t) =

3 +t +c. 2 + lnt +
c
t
. 7 3
4
t
2
+ 1
+
8ct
(t
2
+ 1)
2
i 3 +t +c = 6 i c = 3 3t,
so 7 3
4
t
2
+ 1
+
8(3 3t)t
(t
2
+ 1)
2
= 9 C
24t 312t
2
34
(t
2
+ 1)
2
= 2 C t
4
+ 8t
2
312t + 3 = 0.
It is easily seen that t = 1 is a root of this polynomial. Also 2 + ln1 +
3 31
1
= 4, so t = 1 is the desired solution.
14 Review ET 13
1. A vector function is a function whose domain is a set of real numbers and whose range is a set of vectors. To nd the derivative
or integral, we can differentiate or integrate each component of the vector function.
2. The tip of the moving vector r(t) of a continuous vector function traces out a space curve.
3. The tangent vector to a smooth curve at a point 1 with position vector r(t) is the vector r
0
(t). The tangent line at 1 is the line
through 1 parallel to the tangent vector r
0
(t). The unit tangent vector is T(t) =
r
0
(t)
|r
0
(t)|
.
4. (a) (f ) See Theorem 14.2.3 [ET 13.2.3].
5. Use Formula 14.3.2 [ ET 13.3.2], or equivalently, 14.3.3 [ ET 13.3.3].
6. (a) The curvature of a curve is r =
dT
dc
where Tis the unit tangent vector.

(b) r(t) =
T
0
(t)
r
0
(t)
(c) r(t) =
|r
0
(t) r
00
(t)|
|r
0
(t)|
3
(d) r(r) =
|1
00
(r)|
[1 + (1
0
(r))
2
]
3'2
7. (a) The unit normal vector: N(t) =
T
0
(t)
|T
0
(t)|
. The binormal vector: B(t) = T(t) N(t).
(b) See the discussion preceding Example 7 in Section 14.3 [ ET 13.3].
8. (a) If r(t) is the position vector of the particle on the space curve, the velocity v(t) = r
0
(t), the speed is given by |v(t)|,
and the acceleration a(t) = v
0
(t) = r
00
(t).
(b) a = o
T
T+o
^
Nwhere o
T
=
0
and o
^
= r
2
.
9. See the statement of Keplers Laws on page 880 [ET page 844].
1. True. If we reparametrize the curve by replacing & = t
3
, we have r(&) = &i + 2&j + 3&k, which is a line through the origin
with direction vector i + 2 j + 3 k.
3. False. By Formula 5 of Theorem 14.2.3[ ET 13.2.3],
d
dt
[u(t) v(t)] = u
0
(t) v(t) +u(t) v
0
(t).
5. False. r is the magnitude of the rate of change of the unit tangent vector Twith respect to arc length c, not with respect to t.
7. True. At an inection point where 1 is twice continuously differentiable we must have 1
00
(r) = 0, and by Equation 14.3.11
[ET 13.3.11], the curvature is 0 there.
9. False. If r(t) is the position of a moving particle at time t and |r(t)| = 1 then the particle lies on the unit circle or the unit
sphere, but this does not mean that the speed |r
0
(t)| must be constant. As a counterexample, let r(t) =

t.
I
1 3t
2
, then
r
0
(t) =

1. 3t
I
1 3t
2
and |r(t)| =
I
t
2
+ 1 3t
2
= 1 but |r
0
(t)| =

1 +t
2
(1 3t
2
) = 1
I
1 3t
2
which is not
constant.
11. True. See the discussion preceding Example 7 in Section 14.3 [ ET 13.3].
1. (a) The corresponding parametric equations for the curve are r = t,
j = cos t, : = sin t. Since j
2
+:
2
= 1, the curve is contained in a
circular cylinder with axis the r-axis. Since r = t, the curve is a helix.
(b) r(t) = t i + cos t j + sin t k i
r
0
(t) = i 3 sin t j + cos t k i
r
00
(t) = 3
2
cos t j 3
2
sin t k
3. The projection of the curve C of intersection onto the rj-plane is the circle r
2
+j
2
= 16. : = 0. So we can write
r = 4 cos t, j = 4 sint, 0 $ t $ 2. From the equation of the plane, we have : = 5 3r = 5 34 cos t, so parametric
equations for C are r = 4 cos t, j = 4 sint, : = 5 34 cos t, 0 $ t $ 2, and the corresponding vector function is
r(t) = 4 cos t i + 4 sint j + (5 34 cos t) k, 0 $ t $ 2.
5.

1
0
(t
2
i +t cos t j + sin t k) dt =

1
0
t
2
dt
i +
1
0
t cos t dt
j +
1
0
sin t dt
k
=

1
3
t
3
1
0
i +
I
r
sin t
1
0
3
1
0
1
r
sin t dt
j +
3
1
r
cos t
1
0
k
=
1
3
i +
1
r
2
cos t
1
0
j +
2
r
k =
1
3
i 3
2
r
2
j +
2
r
k
where we integrated by parts in the j-component.
7. r(t) =

t
2
. t
3
. t
4
i r
0
(t) =

2t. 3t
2
. 4t
3
i |r
0
(t)| =
I
4t
2
+ 9t
4
+ 16t
6
and
1 =

3
0
|r
0
(t)| dt =

3
0
I
4t
2
+ 9t
4
+ 16t
6
dt. Using Simpsons Rule with 1(t) =
I
4t
2
+ 9t
4
+ 16t
6
and n = 6 we
have {t =
330
6
=
1
2
and
1 E
{I
3
1(0) + 41
1
2
+ 21(1) + 41
3
2
+ 21(2) + 41
5
2
+1(3)
=
1
6
I
0 + 0 + 0 + 4
1
2
2
+ 9
1
2
4
+ 16
1
2
6
+ 2
4(1)
2
+ 9(1)
4
+ 16(1)
6
+ 4
3
2
2
+ 9
3
2
4
+ 16
3
2
6
+ 2
4(2)
2
+ 9(2)
4
+ 16(2)
6
+ 4
5
2
2
+ 9
5
2
4
+ 16
5
2
6
+
4(3)
2
+ 9(3)
4
+ 16(3)
6
E 86.631
9. The angle of intersection of the two curves, 0, is the angle between their respective tangents at the point of intersection.
For both curves the point (1. 0. 0) occurs when t = 0.
r
0
1
(t) = 3sint i + cos t j +k i r
0
1
(0) = j +k and r
0
2
(t) = i + 2t j + 3t
2
k i r
0
2
(0) = i.
r
0
1
(0) r
0
2
(0) = (j +k) i = 0. Therefore, the curves intersect in a right angle, that is, 0 =
r
2
.
11. (a) T(t) =
r
0
(t)
|r
0
(t)|
=
t
2
. t. 1
|ht
2
. t. 1i|
=
t
2
. t. 1
I
t
4
+t
2
+ 1
(b) T
0
(t) = 3
1
2
(t
4
+t
2
+ 1)
33'2
(4t
3
+ 2t)
t
2
. t. 1
+ (t
4
+t
2
+ 1)
31'2
h2t. 1. 0i
=
32t
3
3t
(t
4
+t
2
+ 1)
3'2
t
2
. t. 1
+
1
(t
4
+t
2
+ 1)
1'2
h2t. 1. 0i
=
32t
5
3t
3
. 32t
4
3t
2
. 32t
3
3t
2t
5
+ 2t
3
+ 2t. t
4
+t
2
+ 1. 0
(t
4
+t
2
+ 1)
3'2
=
2t. 3t
4
+ 1. 32t
3
3t
(t
4
+t
2
+ 1)
3'2
|T
0
(t)| =
I
4t
2
+t
8
32t
4
+ 1 + 4t
6
+ 4t
4
+t
2
(t
4
+t
2
+ 1)
3'2
=
I
t
8
+ 4t
6
+ 2t
4
+ 5t
2
(t
4
+t
2
+ 1)
3'2
and N(t) =
2t. 1 3t
4
. 32t
3
3t
I
t
8
+ 4t
6
+ 2t
4
+ 5t
2
.
(c) r(t) =
|T
0
(t)|
|r
0
(t)|
=
I
t
8
+ 4t
6
+ 2t
4
+ 5t
2
(t
4
+t
2
+ 1)
2
13. j
0
= 4r
3
, j
00
= 12r
2
and r(r) =
|j
00
|
[1 + (j
0
)
2
]
3'2
=
12r
2
(1 + 16r
6
)
3'2
, so r(1) =
12
17
3'2
.
15. r(t) = hsin2t. t. cos 2ti i r
0
(t) = h2 cos 2t. 1. 32 sin2ti i T(t) =
1
I
5
h2 cos 2t. 1. 32 sin2ti i
T
0
(t) =
1
I
5
h34 sin2t. 0. 34 cos 2ti i N(t) = h3sin2t. 0. 3cos 2ti. So N = N() = h0. 0. 31i and
B = TN =
1
I
5
h31. 2. 0i. So a normal to the osculating plane is h31. 2. 0i and an equation is
31(r 30) + 2(j 3) + 0(: 31) = 0 or r 32j + 2 = 0.
17. r(t) = t lnt i +t j +c
3I
k, v(t) = r
0
(t) = (1 + lnt) i +j 3c
3I
k,
|v(t)| =

(1 + lnt)
2
+ 1
2
+ (3c
3I
)
2
=

2 + 2 lnt + (lnt)
2
+c
32I
, a(t) = v
0
(t) =
1
I
i +c
3I
k
19. We set up the axes so that the shot leaves the athletes hand 7 ft above the origin. Then we are given r(0) = 7j,
|v(0)| = 43 fts, and v(0) has direction given by a 45
angle of elevation. Then a unit vector in the direction of v(0) is

1
I
2
(i +j) i v(0) =
43
I
2
(i +j). Assuming air resistance is negligible, the only external force is due to gravity, so as
in Example 14.4.5 [ET 13.4.5] we have a = 3o j where here o E 32 fts
2
. Since v
0
(t) = a(t), we integrate, giving
v(t) = 3ot j +Cwhere C = v(0) =
43
I
2
(i +j) i v(t) =
43
I
2
i +
43
I
2
3ot
j. Since r
0
(t) = v(t) we integrate
again, so r(t) =
43
I
2
t i +
43
I
2
t 3
1
2
ot
2
j +D. But D = r(0) = 7 j i r(t) =

43
I
2
t i +
43
I
2
t 3
1
2
ot
2
+ 7
j.
(a) At 2 seconds, the shot is at r(2) =
43
I
2
(2) i +
43
I
2
(2) 3
1
2
o(2)
2
+ 7
j E 60.8 i + 3.8 j, so the shot is about 3.8 ft above

the ground, at a horizontal distance of 60.8 ft from the athlete.
(b) The shot reaches its maximum height when the vertical component of velocity is 0:
43
I
2
3ot = 0 i
t =
43
I
2 o
E 0.95 s. Then r(0.95) E 28.9 i + 21.4 j, so the maximum height is approximately 21.4 ft.
(c) The shot hits the ground when the vertical component of r(t) is 0, so
43
I
2
t 3
1
2
ot
2
+ 7 = 0 i
316t
2
+
43
I
2
t + 7 = 0 i t E 2.11 s. r(2.11) E 64.2 i 30.08 j, thus the shot lands approximately 64.2 ft from the
athlete.
21. (a) Instead of proceeding directly, we use Formula 3 of Theorem 14.2.3 [ ET 13.2.3]: r(t) = t R(t) i
v = r
0
(t) = R(t) +t R
0
(t) = cos .t i + sin.t j +t v
u
.
(b) Using the same method as in part (a) and starting with v = R(t) + t R
0
(t), we have
a = v
0
= R
0
(t) +R
0
(t) +t R
00
(t) = 2 R
0
(t) +t R
00
(t) = 2 v
u
+t a
u
.
(c) Here we have r(t) = c
3I
cos .t i +c
3I
sin.t j = c
3I
R(t). So, as in parts (a) and (b),
v = r
0
(t) = c
3I
R
0
(t) 3c
3I
R(t) = c
3I
[R
0
(t) 3R(t)] i
a = v
0
= c
3I
[R
00
(t) 3R
0
(t)] 3c
3I
[R
0
(t) 3R(t)] = c
3I
[R
00
(t) 32 R
0
(t) +R(t)]
= c
3I
a
u
32c
3I
v
u
+c
3I
R
Thus, the Coriolis acceleration (the sum of the extra terms not involving a
u
) is 32c
3I
v
u
+c
3I
R.
PROBLEMS PLUS
1. (a) r(t) = 1cos .t i +1sin.t j i v = r
0
(t) = 3.1sin.t i +.1cos .t j, so r = 1(cos .t i + sin.t j) and
v = .1(3sin.t i + cos .t j). v r = .1
2
(3cos .t sin.t + sin.t cos .t) = 0, so v z r. Since r points along a
radius of the circle, and v z r, v is tangent to the circle. Because it is a velocity vector, v points in the direction of motion.
(b) In (a), we wrote v in the form .1u, where u is the unit vector 3sin.t i + cos .t j. Clearly |v| = .1|u| = .1. At
speed .1, the particle completes one revolution, a distance 21, in time T =
21
.1
=
2
.
.
(c) a =
dv
dt
= 3.
2
1cos .t i 3.
2
1sin.t j = 3.
2
1(cos .t i + sin.t j), so a = 3.
2
r. This shows that a is proportional
to r and points in the opposite direction (toward the origin). Also, |a| = .
2
|r| = .
2
1.
(d) By Newtons Second Law (see Section 14.4 [ET 13.4]), F = ia, so |F| = i|a| = i1.
2
=
i(.1)
2
1
=
i|v|
2
1
.
3. (a) The projectile reaches maximum height when 0 =
dj
dt
=
d
dt
[(
0
sinc)t 3
1
2
ot
2
] =
0
sinc 3ot; that is, when
t =

0
sinc
o
and j = (
0
sinc)
0
sinc
o
3
1
2
o
0
sinc
o
2
=

2
0
sin
2
c
2o
. This is the maximum height attained when
the projectile is red with an angle of elevation c. This maximum height is largest when c =
r
2
. In that case, sinc = 1
and the maximum height is

2
0
2o
.
(b) Let 1 =
2
0
o. We are asked to consider the parabola r

2
+ 21j 31
2
= 0 which can be rewritten as j = 3
1
21
r
2
+
1
2
.
The points on or inside this parabola are those for which 31 $ r $ 1 and 0 $ j $
31
21
r
2
+
1
2
. When the projectile is
red at angle of elevation c, the points (r. j) along its path satisfy the relations r = (
0
cos c) t and
j = (
0
sinc)t 3
1
2
ot
2
, where 0 $ t $ (2
0
sinc)o (as in Example 14.4.5 [ET 13.4.5]). Thus
|r| $
0
cos c
2
0
sinc
o
2
0
o
sin2c
2
0
o
= |1|. This shows that 31 $ r $ 1.

For t in the specied range, we also have j = t
0
sinc 3
1
2
ot
=
1
2
ot
2
0
sinc
o
3t
D 0 and
j = (
0
sinc)
r
0
cos c
3
o
2
0
cos c
2
= (tanc) r 3
o
2
2
0
cos
2
c
r
2
= 3
1
21cos
2
c
r
2
+ (tanc) r. Thus
j 3
31
21
r
2
+
1
2
=
31
21cos
2
c
r
2
+
1
21
r
2
+ (tanc) r 3
1
2
=
r
2
21
1 3
1
cos
2
c
+ (tanc) r 3
1
2
=
r
2
(1 3sec
2
c) + 21(tanc) r 31
2
21
=
3(tan
2
c) r
2
+ 21(tanc) r 31
2
21
=
3[(tanc) r 31]
2
21
$ 0
We have shown that every target that can be hit by the projectile lies on or inside the parabola j = 3
1
21
r
2
+
1
2
.
161
162

PROBLEMS PLUS
Now let (o. /) be any point on or inside the parabola j = 3
1
21
r
2
+
1
2
. Then 31 $ o $ 1 and 0 $ / $ 3
1
21
o
2
+
1
2
.
We seek an angle c such that (o. /) lies in the path of the projectile; that is, we wish to nd an angle c such that
/ = 3
1
21cos
2
c
o
2
+ (tanc) o or equivalently / =
31
21
(tan
2
c + 1)o
2
+ (tanc) o. Rearranging this equation we get
o
2
21
tan
2
c 3o tanc +
o
2
21
+/
= 0 or o
2
(tanc)
2
32o1(tanc) + (o
2
+ 2/1) = 0 (W) . This quadratic equation
for tanc has real solutions exactly when the discriminant is nonnegative. Now 1
2
34C D 0 C
(32o1)
2
34o
2
(o
2
+ 2/1) D 0 C 4o
2
(1
2
3o
2
32/1) D 0 C 3o
2
32/1 +1
2
D 0 C
/ $
1
21
(1
2
3o
2
) C / $
31
21
o
2
+
1
2
. This condition is satised since (o. /) is on or inside the parabola
j = 3
1
21
r
2
+
1
2
. It follows that (o. /) lies in the path of the projectile when tanc satises (W), that is, when
tanc =
2o1
4o
2
(1
2
3o
2
32/1)
2o
2
=
1
I
1
2
32/13o
2
o
.
(c) If the gun is pointed at a target with height / at a distance 1 downrange, then
tanc = /1. When the projectile reaches a distance 1 downrange (remember
we are assuming that it doesnt hit the ground rst), we have 1 = r = (
0
cos c)t,
so t =
1
0
cos c
and j = (
0
sinc)t 3
1
2
ot
2
= 1tanc 3
o1
2
2
2
0
cos
2
c
.
Meanwhile, the target, whose r-coordinate is also 1, has fallen from height / to height
/ 3
1
2
ot
2
= 1tanc 3
o1
2
2
2
0
cos
2
c
. Thus the projectile hits the target.
5. (a) a = 3o j i v = v
0
3ot j = 2 i 3ot j i s = s
0
+ 2t i 3
1
2
ot
2
j = 3.5 j + 2t i 3
1
2
ot
2
j i
s = 2t i +
3.5 3
1
2
ot
2
j. Therefore j = 0 when t =

7o seconds. At that instant, the ball is 2
7o E 0.94 ft to the
right of the table top. Its coordinates (relative to an origin on the oor directly under the tables edge) are (0.94. 0). At
impact, the velocity is v = 2 i 3
I
7o j, so the speed is |v| =
I
4 + 7o E 15 fts.
(b) The slope of the curve when t =

7
o
is
dj
dr
=
djdt
drdt
=
3ot
2
=
3o
7o
2
=
3
I
7o
2
. Thus cot 0 =
I
7o
2
and 0 E 7.6
.
(c) From (a), |v| =
I
4 + 7o. So the ball rebounds with speed 0.8
I
4 + 7o E 12.08 fts at angle of inclination
90
3 0 E 82.3886
. By Example 14.4.5 [ET 13.4.5], the horizontal distance traveled between bounces is
d =

2
0
sin2c
o
, where
0
E 12.08 fts and c E 82.3886
. Therefore, d E 1.197 ft. So the ball strikes the oor at

about 2
7o + 1.197 E 2.13 ft to the right of the tables edge.

PROBLEMS PLUS

163
7. The trajectory of the projectile is given by r(t) = ( cos c)t i +
( sinc)t 3
1
2
ot
2
j, so
v(t) = r
0
(t) = cos ci + ( sinc 3ot) j and
|v(t)| =

( cos c)
2
+ ( sinc 3ot)
2
=

2
3(2o sinc) t +o
2
t
2
=
o
2
t
2
3
2
o
(sinc) t +

2
o
2
= o
t 3

o
sinc
2
+

2
o
2
3

2
o
2
sin
2
c = o
t 3

o
sinc
2
+

2
o
2
cos
2
c
The projectile hits the ground when ( sinc)t 3
1
2
ot
2
= 0 i t =
2r
g
sinc, so the distance traveled by the projectile is
1(c) =
(2r'g) sin o
0
|v(t)| dt =

(2r'g) sin o
0
o
t 3

o
sinc
2
+

2
o
2
cos
2
cdt
=o
t 3(o) sinc
2
t 3

o
sinc
2
+
o
cos c
2
+
[(o) cos c]
2
2
ln
t 3

o
sinc +
t 3

o
sinc
2
+
o
cos c
(2r'g) sin o
0
[using Formula 21 in the Table of Integrals]
=
o
2
o
sinc
o
sinc
2
+
o
cos c
2
+
o
cos c
2
ln
o
sinc +
o
sinc
2
+
o
cos c
+

o
sinc
o
sinc
2
+
o
cos c
2
3
o
cos c
2
ln
o
sinc +
o
sinc
2
+
o
cos c
=
o
2
o
sinc

o
+

2
o
2
cos
2
cln
o
sinc +

o
+

o
sinc

o
3

2
o
2
cos
2
cln
o
sinc +

o
=

2
o
sinc +

2
2o
cos
2
cln
(o) sinc +o
3(o) sinc +o
=

2
o
sinc +

2
2o
cos
2
cln
1 + sinc
1 3sinc
We want to maximize 1(c) for 0 $ c $ 2.

1
0
(c) =

2
o
cos c +

2
2o
cos
2
c
1 3sinc
1 + sinc

2 cos c
(1 3sinc)
2
32 cos c sinc ln
1 + sinc
1 3sinc
=

2
o
cos c +

2
2o
cos
2
c
2
cos c
32 cos c sinc ln
1 + sinc
1 3sinc
=

2
o
cos c +

2
o
cos c
1 3sinc ln
1 + sinc
1 3sinc
=

2
o
cos c
2 3sinc ln
1 + sinc
1 3sinc
1(c) has critical points for 0 < c < 2 when 1

0
(c) = 0 i 2 3sincln
1 +sin o
1 3sin o
= 0 (since cos c 6= 0).

Solving by graphing (or using a CAS) gives c E 0.9855. Compare values at the critical point and the endpoints:
1(0) = 0, 1(2) =
2
o, and 1(0.9855) E 1.20
2
o. Thus the distance traveled by the projectile is maximized
for c E 0.9855 or E 56
.
15 PARTIAL DERIVATIVES ET 14
15.1 Functions of Several Variables ET 14.1
1. (a) From Table 1, 1(315. 40) = 327, which means that if the temperature is 315
C and the wind speed is 40 kmh, then the

air would feel equivalent to approximately 327
C without wind.
(b) The question is asking: when the temperature is 320
C, what wind speed gives a wind-chill index of 330
C? From
Table 1, the speed is 20 kmh.
(c) The question is asking: when the wind speed is 20 kmh, what temperature gives a wind-chill index of 349
C? From
Table 1, the temperature is 335
C.
(d) The function W = 1(35. ) means that we x T at 35 and allow to vary, resulting in a function of one variable. In
other words, the function gives wind-chill index values for different wind speeds when the temperature is 35
C. From
Table 1 (look at the row corresponding to T = 35), the function decreases and appears to approach a constant value as
increases.
(e) The function W = 1(T. 50) means that we x at 50 and allow T to vary, again giving a function of one variable. In
other words, the function gives wind-chill index values for different temperatures when the wind speed is 50 kmh . From
Table 1 (look at the column corresponding to = 50), the function increases almost linearly as T increases.
3. If the amounts of labor and capital are both doubled, we replace 1. 1 in the function with 21. 21, giving
1(21. 21) = 1.01(21)
0.75
(21)
0.25
= 1.01(2
0.75
)(2
0.25
)1
0.75
1
0.25
= (2
1
)1.011
0.75
1
0.25
= 21(1. 1)
Thus, the production is doubled. It is also true for the general case 1(1. 1) = /1
o
1
1 3o
:
1(21. 21) = /(21)
o
(21)
13o
= /(2
o
)(2
13o
)1
o
1
13o
= (2
o+13o
)/1
o
1
13o
= 21(1. 1).
5. (a) According to Table 4, 1(40. 15) = 25, which means that if a 40-knot wind has been blowing in the open sea for 15 hours,
it will create waves with estimated heights of 25 feet.
(b) / = 1(30. t) means we x at 30 and allow t to vary, resulting in a function of one variable. Thus here, / = 1(30. t)
gives the wave heights produced by 30-knot winds blowing for t hours. From the table (look at the row corresponding to
= 30), the function increases but at a declining rate as t increases. In fact, the function values appear to be approaching a
limiting value of approximately 19, which suggests that 30-knot winds cannot produce waves higher than about 19 feet.
(c) / = 1(. 30) means we x t at 30, again giving a function of one variable. So, / = 1(. 30) gives the wave heights
produced by winds of speed blowing for 30 hours. From the table (look at the column corresponding to t = 30), the
function appears to increase at an increasing rate, with no apparent limiting value. This suggests that faster winds (lasting
30 hours) always create higher waves.
7. (a) 1(2. 0) = 2
2
c
3(2)(0)
= 4(1) = 4
(b) Since both r
2
and the exponential function are dened everywhere, r
2
c
3iu
is dened for all choices of values for r and j.
Thus the domain of 1 is R
2
.
165
166 CHAPTER 15 PARTIAL DERIVATIVES ET CHAPTER 14
(c) Because the range of o(r. j) = 3rj is R, and the range of c
i
is (0. "), the range of c
g(i.u)
= c
3iu
is (0. ").
The range of r
2
is [0. "), so the range of the product r
2
c
3iu
is [0. ").
9. (a) 1(2. 31. 6) = c
I
632
2
3(31)
2
= c
I
1
= c.
(b) c
I
:3i
2
3u
2
is dened when : 3r
2
3j
2
D 0 i : D r
2
+j
2
. Thus the domain of 1 is
(r. j. :) | : D r
2
+j
2
.
(c) Since
: 3r
2
3j
2
D 0, we have c
I
:3i
2
3u
2
D 1. Thus the range of 1 is [1. ").
11.
I
r +j is dened only when r +j D 0, or j D 3r. So
the domain of 1 is {(r. j) | j D 3r}.
13. ln(9 3r
2
39j
2
) is dened only when
9 3r
2
39j
2
0, or
1
9
r
2
+j
2
< 1. So the domain of 1
is
(r. j)

1
9
r
2
+j
2
< 1
, the interior of an ellipse.

15.
I
1 3r
2
is dened only when 1 3r
2
D 0, or r
2
$ 1
C 31 $ r $ 1, and
1 3j
2
is dened only when
1 3j
2
D 0, or j
2
$ 1 C 31 $ j $ 1. Thus the
domain of 1 is {(r. j) | 31 $ r $ 1. 31 $ j $ 1}.
17.

j 3r
2
is dened only when j 3r
2
D 0, or j D r
2
.
In addition, 1 is not dened if 1 3r
2
= 0 i
r = 1. Thus the domain of 1 is
(r. j) | j D r
2
. r 6= 1
.
19. We need 1 3r
2
3j
2
3:
2
D 0 or r
2
+j
2
+:
2
$ 1,
so 1 =

(r. j. :) | r
2
+j
2
+:
2
$ 1
(the points inside

or on the sphere of radius 1, center the origin).
21. : = 3, a horizontal plane through the point (0. 0. 3).
SECTION 15.1 FUNCTIONS OF SEVERAL VARIABLES ET SECTION 14.1

167
23. : = 10 34r 35j or 4r + 5j +: = 10, a plane with
intercepts 2.5, 2, and 10.
25. : = j
2
+ 1, a parabolic cylinder
27. : = 4r
2
+j
2
+ 1, an elliptic paraboloid with vertex
at (0. 0. 1).
29. : =

r
2
+j
2
so r
2
+j
2
= :
2
and : D 0, the top half
of a right circular cone.
31. The point (33. 3) lies between the level curves with :-values 50 and 60. Since the point is a little closer to the level curve with
: = 60, we estimate that 1(33. 3) E 56. The point (3. 32) appears to be just about halfway between the level curves with
:-values 30 and 40, so we estimate 1(3. 32) E 35. The graph rises as we approach the origin, gradually from above, steeply
from below.
33. Near , the level curves are very close together, indicating that the terrain is quite steep. At 1, the level curves are much
farther apart, so we would expect the terrain to be much less steep than near , perhaps almost at.
35. 37.
39. The level curves are (j 32r)
2
= I or j = 2r
I
I,
I D 0, a family of pairs of parallel lines.
41. The level curves are j 3lnr = I or j = lnr +I.
43. The level curves are jc
i
= I or j = Ic
3i
, a family of
exponential curves.
45. The level curves are
j
2
3r
2
= I or j
2
3r
2
= I
2
,
I D 0. When I = 0 the level curve is the pair of lines
j = r. For I 0, the level curves are hyperbolas with
axis the j-axis.
47. The contour map consists of the level curves I = r
2
+ 9j
2
, a family of
ellipses with major axis the r-axis. (Or, if I = 0, the origin.)
The graph of 1 (r. j) is the surface : = r
2
+ 9j
2
, an elliptic paraboloid.
If we visualize lifting each ellipse I = r
2
+ 9j
2
of the contour map to the plane
: = I, we have horizontal traces that indicate the shape of the graph of 1.
49. The isothermals are given by I = 100(1 +r
2
+ 2j
2
) or
r
2
+ 2j
2
= (100 3I)I [0 < I $ 100], a family of ellipses.

169
51. 1(r. j) = c
3i
2
+c
32u
2
53. 1(r. j) = rj
2
3r
3
The traces parallel to the j:-plane (such as the left-front trace in the graph above) are parabolas; those parallel to the r:-plane
(such as the right-front trace) are cubic curves. The surface is called a monkey saddle because a monkey sitting on the surface
near the origin has places for both legs and tail to rest.
55. (a) C (b) II
Reasons: This function is periodic in both r and j, and the function is the same when r is interchanged with j, so its graph is
symmetric about the plane j = r. In addition, the function is 0 along the r- and j-axes. These conditions are satised only by
C and II.
57. (a) F (b) I
Reasons: This function is periodic in both r and j but is constant along the lines j = r +I, a condition satised only by F and
I.
59. (a) B (b) VI
Reasons: This function is 0 along the lines r = 1 and j = 1. The only contour map in which this could occur is VI. Also
note that the trace in the r:-plane is the parabola : = 1 3r
2
and the trace in the j:-plane is the parabola : = 1 3j
2
, so the
graph is B.
61. I = r + 3j + 5: is a family of parallel planes with normal vector h1. 3. 5i.
63. I = r
2
3j
2
+:
2
are the equations of the level surfaces. For I = 0, the surface is a right circular cone with vertex the origin
and axis the j-axis. For I 0, we have a family of hyperboloids of one sheet with axis the j-axis. For I < 0, we have a
family of hyperboloids of two sheets with axis the j-axis.
65. (a) The graph of o is the graph of 1 shifted upward 2 units.
(b) The graph of o is the graph of 1 stretched vertically by a factor of 2.
(c) The graph of o is the graph of 1 reected about the rj-plane.
(d) The graph of o(r. j) = 31(r. j) + 2 is the graph of 1 reected about the rj-plane and then shifted upward 2 units.
67. 1(r. j) = 3r 3r
4
34j
2
310rj
Three-dimensional view Front view
It does appear that the function has a maximum value, at the higher of the two hilltops. From the front view graph, the
maximum value appears to be approximately 15. Both hilltops could be considered local maximum points, as the values of 1
there are larger than at the neighboring points. There does not appear to be any local minimum point; although the valley shape
between the two peaks looks like a minimum of some kind, some neighboring points have lower function values.
69. 1(r. j) =
r +j
r
2
+j
2
. As both r and j become large, the function values
appear to approach 0, regardless of which direction is considered. As
(r. j) approaches the origin, the graph exhibits asymptotic behavior.
From some directions, 1(r. j) <", while in others 1(r. j) <3".
(These are the vertical spikes visible in the graph.) If the graph is
examined carefully, however, one can see that 1(r. j) approaches 0
along the line j = 3r.
71. 1 (r. j) = c
ci
2
+u
2
. First, if c = 0, the graph is the cylindrical surface
: = c
u
2
(whose level curves are parallel lines). When c 0, the vertical trace
above the j-axis remains xed while the sides of the surface in the r-direction
curl upward, giving the graph a shape resembling an elliptic paraboloid. The
level curves of the surface are ellipses centered at the origin.
c = 0
For 0 < c < 1, the ellipses have major axis the r-axis and the eccentricity increases as c <0.
c = 0.5 (level curves in increments of 1)

171
For c = 1 the level curves are circles centered at the origin.
c = 1 (level curves in increments of 1)
When c 1, the level curves are ellipses with major axis the j-axis, and the eccentricity increases as c increases.
c = 2 (level curves in increments of 4)
For values of c < 0, the sides of the surface in the r-direction curl downward and approach the rj-plane (while the vertical
trace r = 0 remains xed), giving a saddle-shaped appearance to the graph near the point (0. 0. 1). The level curves consist of
a family of hyperbolas. As c decreases, the surface becomes atter in the r-direction and the surfaces approach to the curve in
the trace r = 0 becomes steeper, as the graphs demonstrate.
c = 30.5 (level curves in increments of 0.25)
c = 32 (level curves in increments of 0.25)
73. : = r
2
+j
2
+crj. When c < 32, the surface intersects the plane : = I 6= 0 in a hyperbola. (See graph below.) It intersects
the plane r = j in the parabola : = (2 +c)r
2
, and the plane r = 3j in the parabola : = (2 3c)r
2
. These parabolas open in
opposite directions, so the surface is a hyperbolic paraboloid.
When c = 32 the surface is : = r
2
+j
2
32rj = (r 3j)
2
. So the surface is constant along each line r 3j = I. That
is, the surface is a cylinder with axis r 3j = 0, : = 0. The shape of the cylinder is determined by its intersection with the
plane r +j = 0, where : = 4r
2
, and hence the cylinder is parabolic with minima of 0 on the line j = r.
c = 35, : = 2 c = 310 c = 32
When 32 < c $ 0, : D 0 for all r and j. If r and j have the same sign, then
r
2
+j
2
+crj D r
2
+j
2
32rj = (r 3j)
2
D 0. If they have opposite signs, then crj D 0. The intersection with the
surface and the plane : = I 0 is an ellipse (see graph below). The intersection with the surface and the planes r = 0 and
j = 0 are parabolas : = j
2
and : = r
2
respectively, so the surface is an elliptic paraboloid.
When c 0 the graphs have the same shape, but are reected in the plane r = 0, because
r
2
+j
2
+crj = (3r)
2
+j
2
+ (3c)(3r)j. That is, the value of : is the same for c at (r. j) as it is for 3c at (3r. j).
c = 31, : = 2 c = 0 c = 10
So the surface is an elliptic paraboloid for 0 < c < 2, a parabolic cylinder for c = 2, and a hyperbolic paraboloid for c 2.
75. (a) 1 = /1
o
1
13o
i
1
1
= /1
o
1
3o
i
1
1
= /
1
1
o
i ln
1
1
= ln
1
1
i
ln
1
1
= ln/ +cln
1
1
(b) We list the values for ln(11) and ln(11) for the years 18991922. (Historically, these values were rounded to
2 decimal places.)
SECTION 15.2 LIMITS AND CONTINUITY ET SECTION 14.2

173
Year r = ln(11) j = ln(11)
1899 0 0
1900 30.02 30.06
1901 30.04 30.02
1902 30.04 0
1903 30.07 30.05
1904 30.13 30.12
1905 30.18 30.04
1906 30.20 30.07
1907 30.23 30.15
1908 30.41 30.38
1909 30.33 30.24
1910 30.35 30.27
Year r = ln(11) j = ln(11)
1911 30.38 30.34
1912 30.38 30.24
1913 30.41 30.25
1914 30.47 30.37
1915 30.53 30.34
1916 30.49 30.28
1917 30.53 30.39
1918 30.60 30.50
1919 30.68 30.57
1920 30.74 30.57
1921 31.05 30.85
1922 30.98 30.59
After entering the (r. j) pairs into a calculator or CAS, the resulting least squares regression line through the points is
approximately j = 0.75136r + 0.01053, which we round to j = 0.75r + 0.01.
(c) Comparing the regression line from part (b) to the equation j = ln/ +cr with r = ln(11) and j = ln(11), we have
c = 0.75 and ln/ = 0.01 i / = c
0.01
E 1.01. Thus, the Cobb-Douglas production function is
1 = /1
o
1
13o
= 1.011
0.75
1
0.25
.
15.2 Limits and Continuity ET 14.2
1. In general, we cant say anything about 1(3. 1)! lim
(i.u)<(3.1)
1(r. j) = 6 means that the values of 1(r. j) approach 6 as
(r. j) approaches, but is not equal to, (3. 1). If 1 is continuous, we know that lim
(i.u)<(a.l)
1(r. j) = 1(o. /), so
lim
(i.u)<(3.1)
1(r. j) = 1(3. 1) = 6.
3. We make a table of values of
1(r. j) =
r
2
j
3
+r
3
j
2
35
2 3rj
for a set
of (r. j) points near the origin.
As the table shows, the values of 1(r. j) seem to approach 32.5 as (r. j) approaches the origin from a variety of different
directions. This suggests that lim
(i.u)<(0.0)
1(r. j) = 32.5. Since 1 is a rational function, it is continuous on its domain. 1 is
dened at (0. 0), so we can use direct substitution to establish that lim
(i.u)<(0.0)
1(r. j) =
0
2
0
3
+ 0
3
0
2
35
2 30 0
= 3
5
2
, verifying
our guess.
5. 1(r. j) = 5r
3
3r
2
j
2
is a polynomial, and hence continuous, so lim
(i.u)<(1.2)
1(r. j) = 1(1. 2) = 5(1)
3
3(1)
2
(2)
2
= 1.
7. 1(r. j) =
4 3rj
r
2
+ 3j
2
is a rational function and hence continuous on its domain.
(2. 1) is in the domain of 1, so 1 is continuous there and lim
(i.u)<(2.1)
1(r. j) = 1(2. 1) =
4 3(2)(1)
(2)
2
+ 3(1)
2
=
2
7
.
9. 1(r. j) = j
4
r
4
+ 3j
4
. First approach (0. 0) along the r-axis. Then 1(r. 0) = 0r

4
= 0 for r 6= 0, so 1(r. j) <0.
Now approach (0. 0) along the j-axis. Then for j 6= 0, 1(0. j) = j
4
3j
4
= 13, so 1(r. j) <13. Since 1 has two different
limits along two different lines, the limit does not exist.
11. 1(r. j) = (rj cos j)(3r
2
+j
2
). On the r-axis, 1 (r. 0) = 0 for r 6= 0, so 1(r. j) <0 as (r. j) <(0. 0) along the
r-axis. Approaching (0. 0) along the line j = r, 1(r. r) = (r
2
cos r)4r
2
=
1
4
cos r for r 6= 0, so 1(r. j) <
1
4
along this
line. Thus the limit does not exist.
13. 1(r. j) =
rj
r
2
+j
2
. We can see that the limit along any line through (0. 0) is 0, as well as along other paths through
(0. 0) such as r = j
2
and j = r
2
. So we suspect that the limit exists and equals 0; we use the Squeeze Theorem to prove our
assertion. 0 $
rj
r
2
+j
2
$ |r| since |j| $
r
2
+j
2
, and |r| <0 as (r. j) <(0. 0). So lim
(i.u)<(0.0)
1(r. j) = 0.
15. Let 1(r. j) =
r
2
jc
u
r
4
+ 4j
2
. Then 1(r. 0) = 0 for r 6= 0, so 1(r. j) <0 as (r. j) <(0. 0) along the r-axis. Approaching
(0. 0) along the j-axis or the line j = r also gives a limit of 0. But 1
r. r
2
=
r
2
r
2
c
i
2
r
4
+ 4(r
2
)
2
=
r
4
c
i
2
5r
4
=
c
i
2
5
for r 6= 0, so
1(r. j) <c
0
5 =
1
5
as (r. j) <(0. 0) along the parabola j = r
2
. Thus the limit doesnt exist.
17. lim
(i.u)<(0.0)
r
2
+j
2
r
2
+j
2
+ 1 31
= lim
(i.u)<(0.0)
r
2
+j
2
r
2
+j
2
+ 1 31

r
2
+j
2
+ 1 + 1
r
2
+j
2
+ 1 + 1
= lim
(i.u)<(0.0)
r
2
+j
2
r
2
+j
2
+ 1 + 1
r
2
+j
2
= lim
(i.u)<(0.0)
r
2
+j
2
+ 1 + 1
= 2
19. c
3iu
and sin(:2) are each compositions of continuous functions, and hence continuous, so their product
1(r. j. :) = c
3iu
sin(:2) is a continuous function. Then
lim
(i.u.:)<(3.0.1)
1(r. j. :) = 1 (3. 0. 1) = c
3(3)(0)
sin( 12) = 1.
21. 1(r. j. :) =
rj +j:
2
+r:
2
r
2
+j
2
+:
4
. Then 1(r. 0. 0) = 0r
2
= 0 for r 6= 0, so as (r. j. :) <(0. 0. 0) along the r-axis,
1(r. j. :) <0. But 1(r. r. 0) = r
2
(2r
2
) =
1
2
for r 6= 0, so as (r. j. :) <(0. 0. 0) along the line j = r, : = 0,
1(r. j. :) <
1
2
. Thus the limit doesnt exist.
SECTION 15.2 LIMITS AND CONTINUITY ET SECTION 14.2

175
23. From the ridges on the graph, we see that as (r. j) <(0. 0) along the
lines under the two ridges, 1(r. j) approaches different values. So the
limit does not exist.
25. /(r. j) = o(1(r. j)) = (2r + 3j 36)
2
+
I
2r + 3j 36. Since 1 is a polynomial, it is continuous on R
2
and o is
continuous on its domain {t | t D 0}. Thus / is continuous on its domain.
1 = {(r. j) | 2r + 3j 36 D 0} =

(r. j) | j D 3
2
3
r + 2
, which consists of all points on or above the line j = 3

2
3
r +2.
27. From the graph, it appears that 1 is discontinuous along the line j = r.
If we consider 1(r. j) = c
1'(i3u)
as a composition of functions,
o(r. j) = 1(r 3j) is a rational function and therefore continuous except
where r 3j = 0 i j = r. Since the function /(t) = c
I
is continuous
everywhere, the composition /(o(r. j)) = c
1'(i3u)
= 1(r. j) is
continuous except along the line j = r, as we suspected.
29. The functions sin(rj) and c
i
3j
2
are continuous everywhere, so 1(r. j) =
sin(rj)
c
i
3j
2
is continuous except where
c
i
3j
2
= 0 i j
2
= c
i
i j =
I
c
i
= c
1
2
i
. Thus 1 is continuous on its domain
(r. j) | j 6= c
i'2
.
31. 1(r. j) = arctan
r +
= o(1(r. j)) where 1(r. j) = r +
j, continuous on its domain {(r. j) | j D 0}, and

o(t) = arctant is continuous everywhere. Thus 1 is continuous on its domain {(r. j) | j D 0}.
33. G(r. j) = ln
r
2
+j
2
34
= o(1(r. j)) where 1(r. j) = r

2
+j
2
34, continuous on R
2
, and o(t) = lnt, continuous on
its domain {t | t 0}. Thus G is continuous on its domain
(r. j) | r
2
+j
2
34 0
=

(r. j) | r
2
+j
2
4
, the
exterior of the circle r
2
+j
2
= 4.
35.

j is continuous on its domain {j | j D 0} and r
2
3j
2
+:
2
is continuous everywhere, so 1(r. j. :) =
j
r
2
3j
2
+:
2
is
continuous for j D 0 and r
2
3j
2
+:
2
6= 0 i j
2
6= r
2
+:
2
, that is,
(r. j. :) | j D 0, j 6=
I
r
2
+:
2
.
37. 1(r. j) =
r
2
j
3
2r
2
+j
2
if (r. j) 6= (0. 0)
1 if (r. j) = (0. 0)
The rst piece of 1 is a rational function dened everywhere except at the
origin, so 1 is continuous on R
2
except possibly at the origin. Since r
2
$ 2r
2
+j
2
, we have
r
2
j
3
(2r
2
+j
2
)
j
3
. We
know that
j
3
<0 as (r. j) <(0. 0). So, by the Squeeze Theorem, lim

(i.u)<(0.0)
1(r. j) = lim
(i.u)<(0.0)
r
2
j
3
2r
2
+j
2
= 0.
But 1(0. 0) = 1, so 1 is discontinuous at (0. 0). Therefore, 1 is continuous on the set {(r. j) | (r. j) 6= (0. 0)}.
39. lim
(i.u)<(0.0)
r
3
+j
3
r
2
+j
2
= lim
r<0
+
(v cos 0)
3
+ (v sin0)
3
v
2
= lim
r<0
+
(v cos
3
0 +v sin
3
0) = 0
41. lim
(i.u)<(0.0)
c
3i
2
3u
2
31
r
2
+j
2
= lim
r<0
+
c
3r
2
31
v
2
= lim
r<0
+
c
3r
2
(32v)
2v
[using lHospitals Rule]
= lim
r<0
+
3c
3r
2
= 3c
0
= 31
43. 1(r. j) =
sin(rj)
rj
if (r. j) 6= (0. 0)
1 if (r. j) = (0. 0)
From the graph, it appears that 1 is continuous everywhere. We know
rj is continuous on R
2
and sint is continuous everywhere, so
sin(rj) is continuous on R
2
and
sin(rj)
rj
is continuous on R
2
except possibly where rj = 0. To show that 1 is continuous at those points, consider any point (o. /) in R
2
where o/ = 0.
Because rj is continuous, rj <o/ = 0 as (r. j) <(o. /). If we let t = rj, then t <0 as (r. j) <(o. /) and
lim
(i.u)<(a.l)
sin(rj)
rj
= lim
I<0
sin(t)
t
= 1 by Equation 3.4.2 [ET 3.3.2]. Thus lim
(i.u)<(a.l)
1(r. j) = 1(o. /) and 1 is continuous
on R
2
.
45. Since |x 3a|
2
= |x|
2
+ |a|
2
32 |x| |a| cos 0 D |x|
2
+ |a|
2
32 |x| |a| = (|x| 3|a|)
2
, we have
|x| 3|a|
$ |x 3a|. Let
c 0 be given and set c = c. Then if 0 < |x 3a| < c,
|x| 3|a|
$ |x 3a| < c = c. Hence lim

x<a
|x| = |a| and
1 (x) = |x| is continuous on R
n
.
15.3 Partial Derivatives ET 14.3
1. (a) uTur represents the rate of change of T when we x j and t and consider T as a function of the single variable r, which
describes how quickly the temperature changes when longitude changes but latitude and time are constant. uTuj
represents the rate of change of T when we x r and t and consider T as a function of j, which describes how quickly the
temperature changes when latitude changes but longitude and time are constant. uTut represents the rate of change of T
when we x r and j and consider T as a function of t, which describes how quickly the temperature changes over time for
a constant longitude and latitude.
(b) 1
i
(158. 21. 9) represents the rate of change of temperature at longitude 158
W, latitude 21
N at 9:00 AM when only

longitude varies. Since the air is warmer to the west than to the east, increasing longitude results in an increased air
temperature, so we would expect 1
i
(158. 21. 9) to be positive. 1
u
(158. 21. 9) represents the rate of change of temperature
at the same time and location when only latitude varies. Since the air is warmer to the south and cooler to the north,
increasing latitude results in a decreased air temperature, so we would expect 1
u
(158. 21. 9) to be negative. 1
I
(158. 21. 9)
represents the rate of change of temperature at the same time and location when only time varies. Since typically air
SECTION 15.3 PARTIAL DERIVATIVES ET SECTION 14.3

177
temperature increases from the morning to the afternoon as the sun warms it, we would expect 1
I
(158. 21. 9) to be
positive.
3. (a) By Denition 4, 1
T
(315. 30) = lim
I<0
1(315 +/. 30) 31(315. 30)
/
, which we can approximate by considering / = 5
and / = 35 and using the values given in the table:
1
T
(315. 30) E
1(310. 30) 31(315. 30)
5
=
320 3(326)
5
=
6
5
= 1.2,
1
T
(315. 30) E
1(320. 30) 31(315. 30)
35
=
333 3(326)
35
=
37
35
= 1.4. Averaging these values, we estimate
1
T
(315. 30) to be approximately 1.3. Thus, when the actual temperature is 315
C and the wind speed is 30 kmh, the

apparent temperature rises by about 1.3
C for every degree that the actual temperature rises.

Similarly, 1
r
(315. 30) = lim
I<0
1(315. 30 +/) 31(315. 30)
/
which we can approximate by considering / = 10 and
/ = 310: 1
r
(315. 30) E
1(315. 40) 31(315. 30)
10
=
327 3(326)
10
=
31
10
= 30.1,
1
r
(315. 30) E
1(315. 20) 31(315. 30)
310
=
324 3(326)
310
=
2
310
= 30.2. Averaging these values, we estimate
1
r
(315. 30) to be approximately 30.15. Thus, when the actual temperature is 315
C and the wind speed is 30 kmh, the

apparent temperature decreases by about 0.15
C for every kmh that the wind speed increases.

(b) For a xed wind speed , the values of the wind-chill index W increase as temperature T increases (look at a column of
the table), so
uW
uT
is positive. For a xed temperature T, the values of W decrease (or remain constant) as increases
(look at a row of the table), so
uW
u
is negative (or perhaps 0).
(c) For xed values of T, the function values 1(T. ) appear to become constant (or nearly constant) as increases, so the
corresponding rate of change is 0 or near 0 as increases. This suggests that lim
r<"
(uWu) = 0.
5. (a) If we start at (1. 2) and move in the positive r-direction, the graph of 1 increases. Thus 1
i
(1. 2) is positive.
(b) If we start at (1. 2) and move in the positive j-direction, the graph of 1 decreases. Thus 1
u
(1. 2) is negative.
7. (a) 1
ii
=
O
Oi
(1
i
), so 1
ii
is the rate of change of 1
i
in the r-direction. 1
i
is negative at (31. 2) and if we move in the
positive r-direction, the surface becomes less steep. Thus the values of 1
i
are increasing and 1
ii
(31. 2) is positive.
(b) 1
uu
is the rate of change of 1
u
in the j-direction. 1
u
is negative at (31. 2) and if we move in the positive j-direction, the
surface becomes steeper. Thus the values of 1
u
are decreasing, and 1
uu
(31. 2) is negative.
9. First of all, if we start at the point (3. 33) and move in the positive j-direction, we see that both / and c decrease, while o
increases. Both / and c have a low point at about (3. 31.5), while o is 0 at this point. So o is denitely the graph of 1
u
, and
one of / and c is the graph of 1. To see which is which, we start at the point (33. 31.5) and move in the positive r-direction.
/ traces out a line with negative slope, while c traces out a parabola opening downward. This tells us that / is the r-derivative
of c. So c is the graph of 1, / is the graph of 1
i
, and o is the graph of 1
u
.
11. 1(r. j) = 16 34r
2
3j
2
i 1
i
(r. j) = 38r and 1
u
(r. j) = 32j i 1
i
(1. 2) = 38 and 1
u
(1. 2) = 34. The graph
of 1 is the paraboloid : = 16 34r
2
3j
2
and the vertical plane j = 2 intersects it in the parabola : = 12 34r
2
, j = 2
(the curve C
1
in the rst gure). The slope of the tangent line
to this parabola at (1. 2. 8) is 1
i
(1. 2) = 38. Similarly the
plane r = 1 intersects the paraboloid in the parabola
: = 12 3j
2
, r = 1 (the curve C
2
in the second gure) and
the slope of the tangent line at (1. 2. 8) is 1
u
(1. 2) = 34.
13. 1(r. j) = r
2
+j
2
+r
2
j i 1
i
= 2r + 2rj, 1
u
= 2j +r
2
Note that the traces of 1 in planes parallel to the r:-plane are parabolas which open downward for j < 31 and upward for
j 31, and the traces of 1
i
in these planes are straight lines, which have negative slopes for j < 31 and positive slopes for
j 31. The traces of 1 in planes parallel to the j:-plane are parabolas which always open upward, and the traces of 1
u
in
these planes are straight lines with positive slopes.
15. 1(r. j) = j
5
33rj i 1
i
(r. j) = 0 33j = 33j, 1
u
(r. j) = 5j
4
33r
17. 1(r. t) = c
3I
cos r i 1
i
(r. t) = c
3I
(3sinr) () = 3c
3I
sinr, 1
I
(r. t) = c
3I
(31) cos r = 3c
3I
cos r
19. : = (2r + 3j)
10
i
u:
ur
= 10(2r + 3j)
9
2 = 20(2r + 3j)
9
,
u:
uj
= 10(2r + 3j)
9
3 = 30(2r + 3j)
9
21. 1(r. j) =
r 3j
r +j
i 1
i
(r. j) =
(1)(r +j) 3(r 3j)(1)
(r +j)
2
=
2j
(r +j)
2
,
1
u
(r. j) =
(31)(r +j) 3(r 3j)(1)
(r +j)
2
= 3
2r
(r +j)
2
23. n = sinc cos o i
un
uc
= cos c cos o,
un
uo
= 3sinc sino
25. 1(v. c) = v ln(v
2
+c
2
) i 1
r
(v. c) = v
2v
v
2
+c
2
+ ln(v
2
+c
2
) 1 =
2v
2
v
2
+c
2
+ ln(v
2
+c
2
),
1
s
(v. c) = v
2c
v
2
+c
2
+ 0 =
2vc
v
2
+c
2

179
27. & = tc
u'I
i
u&
ut
= t c
u'I
(3nt
32
) +c
u'I
1 = c
u'I
3
n
t
c
u'I
= c
u'I
1 3
n
t
,
u&
un
= tc
u'I
1
t
= c
u'I
29. 1(r. j. :) = r: 35r
2
j
3
:
4
i 1
i
(r. j. :) = : 310rj
3
:
4
, 1
u
(r. j. :) = 315r
2
j
2
:
4
, 1
:
(r. j. :) = r 320r
2
j
3
:
3
31. n = ln(r + 2j + 3:) i
un
ur
=
1
r + 2j + 3:
,
un
uj
=
2
r + 2j + 3:
,
un
u:
=
3
r + 2j + 3:
33. & = rj sin
31
(j:) i
u&
ur
= j sin
31
(j:),
u&
uj
= rj
1
1 3(j:)
2
(:) +sin
31
(j:) r =
rj:
1 3j
2
:
2
+rsin
31
(j:),
u&
u:
= rj
1
1 3(j:)
2
(j) =
rj
2
1 3j
2
:
2
35. 1(r. j. :. t) = rj:
2
tan(jt) i 1
i
(r. j. :. t) = j:
2
tan(jt),
1
u
(r. j. :. t) = rj:
2
sec
2
(jt) t +r:
2
tan(jt) = rj:
2
t sec
2
(jt) +r:
2
tan(jt),
1
:
(r. j. :. t) = 2rj: tan(jt), 1
I
(r. j. :. t) = rj:
2
sec
2
(jt) j = rj
2
:
2
sec
2
(jt)
37. & =

r
2
1
+r
2
2
+ +r
2
n
. For each j = 1, . . ., n, &
i
l
=
1
2
r
2
1
+r
2
2
+ +r
2
n
31'2
(2r
.
) =
r
.
r
2
1
+r
2
2
+ +r
2
n
.
39. 1(r. j) = ln
r +
r
2
+j
2
i
1
i
(r. j) =
1
r +
r
2
+j
2
1 +
1
2
(r
2
+j
2
)
31'2
(2r)
=
1
r +
r
2
+j
2
1 +
r
r
2
+j
2
,
so 1
i
(3. 4) =
1
3 +
I
3
2
+ 4
2
1 +
3
I
3
2
+ 4
2
=
1
8
1 +
3
5
=
1
5
.
41. 1(r. j. :) =
j
r +j +:
i 1
u
(r. j. :) =
1(r +j +:) 3j(1)
(r +j +:)
2
=
r +:
(r +j +:)
2
,
so 1
u
(2. 1. 31) =
2 + (31)
(2 + 1 + (31))
2
=
1
4
.
43. 1(r. j) = rj
2
3r
3
j i
1
i
(r. j) = lim
I<0
1(r +/. j) 31(r. j)
/
= lim
I<0
(r +/)j
2
3(r +/)
3
j 3(rj
2
3r
3
j)
/
= lim
I<0
/(j
2
33r
2
j 33rj/ 3j/
2
)
/
= lim
I<0
(j
2
33r
2
j 33rj/ 3j/
2
) = j
2
33r
2
j
1
u
(r. j) = lim
I<0
1(r. j +/) 31(r. j)
/
= lim
I<0
r(j +/)
2
3r
3
(j +/) 3(rj
2
3r
3
j)
/
= lim
I<0
/(2rj +r/ 3r
3
)
/
= lim
I<0
(2rj +r/ 3r
3
) = 2rj 3r
3
45. r
2
+j
2
+:
2
= 3rj: i
u
ur
(r
2
+j
2
+:
2
) =
u
ur
(3rj:) i 2r + 0 + 2:
u:
ur
= 3j
r
u:
ur
+: 1
C
2:
u:
ur
33rj
u:
ur
= 3j: 32r C (2: 33rj)
u:
ur
= 3j: 32r, so
u:
ur
=
3j: 32r
2: 33rj
.
u
uj
(r
2
+j
2
+:
2
) =
u
uj
(3rj:) i 0 + 2j + 2:
u:
uj
= 3r
j
u:
uj
+: 1
C 2:
u:
uj
33rj
u:
uj
= 3r: 32j C
(2: 33rj)
u:
uj
= 3r: 32j, so
u:
uj
=
3r: 32j
2: 33rj
.
47. r 3: = arctan(j:) i
u
ur
(r 3:) =
u
ur
(arctan(j:)) i 1 3
u:
ur
=
1
1 + (j:)
2
j
u:
ur
C
1 =

j
1 +j
2
:
2
+ 1
u:
ur
C 1 =

j + 1 +j
2
:
2
1 +j
2
:
2
u:
ur
, so
u:
ur
=
1 +j
2
:
2
1 +j +j
2
:
2
.
u
uj
(r 3:) =
u
uj
(arctan(j:)) i 0 3
u:
uj
=
1
1 + (j:)
2

j
u:
uj
+: 1
C
3
:
1 +j
2
:
2
=

j
1 +j
2
:
2
+ 1
u:
uj
C 3
:
1 +j
2
:
2
=

j + 1 +j
2
:
2
1 +j
2
:
2
u:
uj
C
u:
uj
= 3
:
1 +j +j
2
:
2
.
49. (a) : = 1(r) +o(j) i
u:
ur
= 1
0
(r),
u:
uj
= o
0
(j)
(b) : = 1(r +j). Let & = r +j. Then
u:
ur
=
d1
d&
u&
ur
=
d1
d&
(1) = 1
0
(&) = 1
0
(r +j),
u:
uj
=
d1
d&
u&
uj
=
d1
d&
(1) = 1
0
(&) = 1
0
(r +j).
51. 1(r. j) = r
3
j
5
+ 2r
4
j i 1
i
(r. j) = 3r
2
j
5
+ 8r
3
j, 1
u
(r. j) = 5r
3
j
4
+ 2r
4
. Then 1
ii
(r. j) = 6rj
5
+ 24r
2
j,
1
iu
(r. j) = 15r
2
j
4
+ 8r
3
, 1
ui
(r. j) = 15r
2
j
4
+ 8r
3
, and 1
uu
(r. j) = 20r
3
j
3
.
53. n =
I
&
2
+
2
i n
u
=
1
2
(&
2
+
2
)
31'2
2& =
&
I
&
2
+
2
, n
r
=
1
2
(&
2
+
2
)
31'2
2 =

I
&
2
+
2
. Then
n
uu
=
1
I
&
2
+
2
3&
1
2
(&
2
+
2
)
31'2
(2&)
I
&
2
+
2
2
=
I
&
2
+
2
3&
2
I
&
2
+
2
&
2
+
2
=
&
2
+
2
3&
2
(&
2
+
2
)
3'2
=

2
(&
2
+
2
)
3'2
,
n
ur
= &
3
1
2

&
2
+
2
33'2
(2) = 3
&
(&
2
+
2
)
3'2
, n
ru
=
3
1
2

&
2
+
2
33'2
(2&) = 3
&
(&
2
+
2
)
3'2
,
n
rr
=
1
I
&
2
+
2
3
1
2
(&
2
+
2
)
31'2
(2)
I
&
2
+
2
2
=
I
&
2
+
2
3
2
I
&
2
+
2
&
2
+
2
=
&
2
+
2
3
2
(&
2
+
2
)
3'2
=
&
2
(&
2
+
2
)
3'2
.
55. : = arctan
r +j
1 3rj
i
:
i
=
1
1 +
i+u
13iu
2

(1)(1 3rj) 3(r +j)(3j)
(1 3rj)
2
=
1 +j
2
(1 3rj)
2
+ (r +j)
2
=
1 +j
2
1 +r
2
+j
2
+r
2
j
2
=
1 +j
2
(1 +r
2
)(1 +j
2
)
=
1
1 +r
2
.
:
u
=
1
1 +
i+u
13iu
2

(1)(1 3rj) 3(r +j)(3r)
(1 3rj)
2
=
1 +r
2
(1 3rj)
2
+ (r +j)
2
=
1 +r
2
(1 +r
2
)(1 +j
2
)
=
1
1 +j
2
.
Then :
ii
= 3(1 +r
2
)
32
2r = 3
2r
(1 +r
2
)
2
, :
iu
= 0, :
ui
= 0, :
uu
= 3(1 +j
2
)
32
2j = 3
2j
(1 +j
2
)
2
.
57. & = rsin(r + 2j) i &
i
= r cos(r + 2j)(1) + sin(r + 2j) 1 = rcos(r + 2j) + sin(r + 2j),
&
iu
= r(3sin(r + 2j)(2)) + cos(r + 2j)(2) = 2 cos(r + 2j) 3 2rsin(r + 2j),
&
u
= rcos (r + 2j) (2) = 2rcos(r + 2j),
&
ui
= 2r (3sin(r + 2j)(1)) + cos (r + 2j) 2 = 2 cos(r + 2j) 32rsin(r + 2j). Thus &
iu
= &
ui
.

181
59. & = ln
r
2
+j
2
= ln(r
2
+j
2
)
1'2
=
1
2
ln(r
2
+j
2
) i &
i
=
1
2
1
r
2
+j
2
2r =
r
r
2
+j
2
,
&
iu
= r(31)(r
2
+j
2
)
32
(2j) = 3
2rj
(r
2
+j
2
)
2
and &
u
=
1
2
1
r
2
+j
2
2j =
j
r
2
+j
2
,
&
ui
= j(31)(r
2
+j
2
)
32
(2r) = 3
2rj
(r
2
+j
2
)
2
. Thus &
iu
= &
ui
.
61. 1(r. j) = 3rj
4
+r
3
j
2
i 1
i
= 3j
4
+ 3r
2
j
2
, 1
ii
= 6rj
2
, 1
iiu
= 12rj and
1
u
= 12rj
3
+ 2r
3
j, 1
uu
= 36rj
2
+ 2r
3
, 1
uuu
= 72rj.
63. 1(r. j. :) = cos(4r + 3j + 2:) i
1
i
= 3sin(4r + 3j + 2:)(4) = 34 sin(4r + 3j + 2:), 1
iu
= 34 cos(4r + 3j + 2:)(3) = 312 cos(4r + 3j + 2:),
1
iu:
= 312(3sin(4r + 3j + 2:))(2) = 24 sin(4r + 3j + 2:) and
1
u
= 3sin(4r + 3j + 2:)(3) = 33 sin(4r + 3j + 2:),
1
u:
= 33 cos(4r +3j +2:)(2) = 36 cos(4r +3j +2:), 1
u::
= 36(3sin(4r +3j +2:))(2) = 12 sin(4r +3j +2:).
65. & = c
r0
sin0 i
u&
u0
= c
r0
cos 0 + sin0 c
r0
(v) = c
r0
(cos 0 + v sin0),
u
2
&
uv u0
= c
r0
(sin0) + (cos 0 +v sin0) c
r0
(0) = c
r0
(sin0 + 0 cos 0 + v0 sin0),
u
3
&
uv
2
u0
= c
r0
(0 sin0) + (sin0 +0 cos 0 +v0 sin0) c
r0
(0) = 0c
r0
(2 sin0 +0 cos 0 +v0 sin0).
67. n =
r
j + 2:
= r(j + 2:)
31
i
un
ur
= (j + 2:)
31
,
u
2
n
uj ur
= 3(j + 2:)
32
(1) = 3(j + 2:)
32
,
u
3
n
u: uj ur
= 3(32)(j + 2:)
33
(2) = 4(j + 2:)
33
=
4
(j + 2:)
3
and
un
uj
= r(31)(j + 2:)
32
(1) = 3r(j + 2:)
32
,
u
2
n
uruj
= 3(j + 2:)
32
,
u
3
n
ur
2
uj
= 0.
69. By Denition 4, 1
i
(3. 2) = lim
I<0
1(3 +/. 2) 31(3. 2)
/
which we can approximate by considering / = 0.5 and / = 30.5:
1
i
(3. 2) E
1(3.5. 2) 31(3. 2)
0.5
=
22.4 317.5
0.5
= 9.8, 1
i
(3. 2) E
1(2.5. 2) 31(3. 2)
30.5
=
10.2 317.5
30.5
= 14.6. Averaging
these values, we estimate 1
i
(3. 2) to be approximately 12.2. Similarly, 1
i
(3. 2.2) = lim
I<0
1(3 +/. 2.2) 31(3. 2.2)
/
which
we can approximate by considering / = 0.5 and / = 30.5: 1
i
(3. 2.2) E
1(3.5. 2.2) 31(3. 2.2)
0.5
=
26.1 315.9
0.5
= 20.4,
1
i
(3. 2.2) E
1(2.5. 2.2) 31(3. 2.2)
30.5
=
9.3 315.9
30.5
= 13.2. Averaging these values, we have 1
i
(3. 2.2) E 16.8.
To estimate 1
iu
(3. 2), we rst need an estimate for 1
i
(3. 1.8):
1
i
(3. 1.8) E
1(3.5. 1.8) 31(3. 1.8)
0.5
=
20.0 318.1
0.5
= 3.8, 1
i
(3. 1.8) E
1(2.5. 1.8) 31(3. 1.8)
30.5
=
12.5 318.1
30.5
= 11.2.
Averaging these values, we get 1
i
(3. 1.8) E 7.5. Now 1
iu
(r. j) =
u
uj
[1
i
(r. j)] and 1
i
(r. j) is itself a function of two
variables, so Denition 4 says that 1
iu
(r. j) =
u
uj
[1
i
(r. j)] = lim
I<0
1
i
(r. j +/) 31
i
(r. j)
/
i
1
iu
(3. 2) = lim
I<0
1
i
(3. 2 +/) 31
i
(3. 2)
/
.
We can estimate this value using our previous work with / = 0.2 and / = 30.2:
1
iu
(3. 2) E
1
i
(3. 2.2) 31
i
(3. 2)
0.2
=
16.8 312.2
0.2
= 23, 1
iu
(3. 2) E
1
i
(3. 1.8) 31
i
(3. 2)
30.2
=
7.5 312.2
30.2
= 23.5.
Averaging these values, we estimate 1
iu
(3. 2) to be approximately 23.25.
71. & = c
3o
2
I
2
I
sinIr i &
i
= Ic
3o
2
I
2
I
cos Ir, &
ii
= 3I
2
c
3o
2
I
2
I
sinIr, and &
I
= 3c
2
I
2
c
3o
2
I
2
I
sinIr.
Thus c
2
&
ii
= &
I
.
73. & =
1
r
2
+j
2
+:
2
i &
i
=

3
1
2
(r
2
+j
2
+:
2
)
33'2
(2r) = 3r(r
2
+j
2
+:
2
)
33'2
and
&
ii
= 3(r
2
+j
2
+:
2
)
33'2
3r
3
3
2
(r
2
+j
2
+:
2
)
35'2
(2r) =
2r
2
3j
2
3:
2
(r
2
+j
2
+:
2
)
5'2
.
By symmetry, &
uu
=
2j
2
3r
2
3:
2
(r
2
+j
2
+:
2
)
5'2
and &
::
=
2:
2
3r
2
3j
2
(r
2
+j
2
+:
2
)
5'2
.
Thus &
ii
+&
uu
+&
::
=
2r
2
3j
2
3:
2
+ 2j
2
3r
2
3:
2
+ 2:
2
3r
2
3j
2
(r
2
+j
2
+:
2
)
5'2
= 0.
75. Let = r +ot, n = r 3ot. Then &
I
=
u[1() +o(n)]
ut
=
d1()
d
u
ut
+
do(n)
dn
un
ut
= o1
0
() 3oo
0
(n) and
&
II
=
u[o1
0
() 3oo
0
(n)]
ut
= o[o1
00
() +oo
00
(n)] = o
2
[1
00
() +o
00
(n)]. Similarly, by using the Chain Rule we have
&
i
= 1
0
() +o
0
(n) and &
ii
= 1
00
() +o
00
(n). Thus &
II
= o
2
&
ii
.
77. : = ln(c
i
+c
u
) i
u:
ur
=
c
i
c
i
+c
u
and
u:
uj
=
c
u
c
i
+c
u
, so
u:
ur
+
u:
uj
=
c
i
c
i
+c
u
+
c
u
c
i
+c
u
=
c
i
+c
u
c
i
+c
u
= 1.
u
2
:
ur
2
=
c
i
(c
i
+c
u
) 3c
i
(c
i
)
(c
i
+c
u
)
2
=
c
i+u
(c
i
+c
u
)
2
,
u
2
:
uruj
=
0 3c
u
(c
i
)
(c
i
+c
u
)
2
= 3
c
i+u
(c
i
+c
u
)
2
, and
u
2
:
uj
2
=
c
u
(c
i
+c
u
) 3c
u
(c
u
)
(c
i
+c
u
)
2
=
c
i+u
(c
i
+c
u
)
2
. Thus
u
2
:
ur
2
u
2
:
uj
2
3
u
2
:
uruj
2
=
c
i+u
(c
i
+c
u
)
2

c
i+u
(c
i
+c
u
)
2
3
3
c
i+u
(c
i
+c
u
)
2
2
=
(c
i+u
)
2
(c
i
+c
u
)
4
3
(c
i+u
)
2
(c
i
+c
u
)
4
= 0
79. If we x 1 = 1
0
. 1(1. 1
0
) is a function of a single variable 1, and
d1
d1
= c
1
1
is a separable differential equation. Then
d1
1
= c
d1
1
i

d1
1
=

c
d1
1
i ln|1| = cln|1| +C (1
0
), where C(1
0
) can depend on 1
0
. Then
|1| = c
oln|1| +C(1
0
)
, and since 1 0 and 1 0, we have 1 = c
oln 1
c
C(1
0
)
= c
C(1
0
)
c
ln 1
= C
1
(1
0
)1
o
where
C
1
(1
0
) = c
C(1
0
)
.
81. By the Chain Rule, taking the partial derivative of both sides with respect to 1
1
gives
u1
31
u1
u1
u1
1
=
u [(11
1
) + (11
2
) + (11
3
)]
u1
1
or 31
32
u1
u1
1
= 31
32
1
. Thus
u1
u1
1
=
1
2
1
2
1
.

183
83. By Exercise 82, 1\ = i1T i 1 =
i1T
\
, so
u1
uT
=
i1
\
. Also, 1\ = i1T i \ =
i1T
1
and
u\
uT
=
i1
1
.
Since T =
1\
i1
, we have T
u1
uT
u\
uT
=
1\
i1

i1
\

i1
1
= i1.
85.
u1
ui
=
1
2
2
,
u1
u
= i,
u
2
1
u
2
= i. Thus
u1
ui

u
2
1
u
2
=
1
2
2
i = 1.
87. 1
i
(r. j) = r + 4j i 1
iu
(r. j) = 4 and 1
u
(r. j) = 3r 3j i 1
ui
(r. j) = 3. Since 1
iu
and 1
ui
are continuous
everywhere but 1
iu
(r. j) 6= 1
ui
(r. j), Clairauts Theorem implies that such a function 1(r. j) does not exist.
89. By the geometry of partial derivatives, the slope of the tangent line is 1
i
(1. 2). By implicit differentiation of
4r
2
+ 2j
2
+:
2
= 16, we get 8r + 2: (u:ur) = 0 i u:ur = 34r:, so when r = 1 and : = 2 we have
u:ur = 32. So the slope is 1
i
(1. 2) = 32. Thus the tangent line is given by : 32 = 32(r 31), j = 2. Taking the
parameter to be t = r 31, we can write parametric equations for this line: r = 1 +t, j = 2, : = 2 32t.
91. By Clairauts Theorem, 1
iuu
= (1
iu
)
u
= (1
ui
)
u
= 1
uiu
= (1
u
)
iu
= (1
u
)
ui
= 1
uui
.
93. Let o(r) = 1(r. 0) = r(r
2
)
33'2
c
0
= r|r|
33
. But we are using the point (1. 0), so near (1. 0), o(r) = r
32
. Then
o
0
(r) = 32r
33
and o
0
(1) = 32, so using (1) we have 1
i
(1. 0) = o
0
(1) = 32.
95. (a) (b) For (r. j) 6= (0. 0),
1
i
(r. j) =
(3r
2
j 3j
3
)(r
2
+j
2
) 3(r
3
j 3rj
3
)(2r)
(r
2
+j
2
)
2
=
r
4
j + 4r
2
j
3
3j
5
(r
2
+j
2
)
2
and by symmetry 1
u
(r. j) =
r
5
34r
3
j
2
3rj
4
(r
2
+j
2
)
2
.
(c) 1
i
(0. 0) = lim
I<0
1(/. 0) 31(0. 0)
/
= lim
I<0
(0/
2
) 30
/
= 0 and 1
u
(0. 0) = lim
I<0
1(0. /) 31(0. 0)
/
= 0.
(d) By (3), 1
iu
(0. 0) =
u1
i
uj
= lim
I<0
1
i
(0. /) 31
i
(0. 0)
/
= lim
I<0
(3/
5
30)/
4
/
= 31 while by (2),
1
ui
(0. 0) =
u1
u
ur
= lim
I<0
1
u
(/. 0) 31
u
(0. 0)
/
= lim
I<0
/
5
/
4
/
= 1.
(e) For (r. j) 6= (0. 0), we use a CAS to compute
1
iu
(r. j) =
r
6
+ 9r
4
j
2
39r
2
j
4
3j
6
(r
2
+j
2
)
3
Now as (r. j) <(0. 0) along the r-axis, 1
iu
(r. j) <1 while as
(r. j) <(0. 0) along the j-axis, 1
iu
(r. j) <31. Thus 1
iu
isnt
continuous at (0. 0) and Clairauts Theorem doesnt apply, so there is
no contradiction. The graphs of 1
iu
and 1
ui
are identical except at the
origin, where we observe the discontinuity.
15.4 Tangent Planes and Linear Approximations ET 14.4
1. : = 1(r. j) = 4r
2
3j
2
+ 2j i 1
i
(r. j) = 8r, 1
u
(r. j) = 32j + 2, so 1
i
(31. 2) = 38, 1
u
(31. 2) = 32.
By Equation 2, an equation of the tangent plane is : 34 = 1
i
(31. 2)[r 3(31)] +1
u
(31. 2)(j 32) i
: 34 = 38(r + 1) 32(j 32) or : = 38r 32j.
3. : = 1(r. j) =
rj i 1
i
(r. j) =
1
2
(rj)
31'2
j =
1
2
jr, 1
u
(r. j) =
1
2
(rj)
31'2
r =
1
2
rj, so 1
i
(1. 1) =
1
2
and 1
u
(1. 1) =
1
2
. Thus an equation of the tangent plane is : 31 = 1
i
(1. 1)(r 31) +1
u
(1. 1)(j 31) i
: 31 =
1
2
(r 31) +
1
2
(j 31) or r +j 32: = 0.
5. : = 1(r. j) = j cos(r 3 j) i 1
i
= j(3sin(r 3 j)(1)) = 3j sin(r 3 j),
1
u
= j(3sin(r 3j)(31)) + cos(r 3j) = j sin(r 3j) + cos(r 3j), so 1
i
(2. 2) = 32 sin(0) = 0,
1
u
(2. 2) = 2 sin(0) + cos(0) = 1 and an equation of the tangent plane is : 32 = 0(r 32) + 1(j 32) or : = j.
6. : = 1(r. j) = c
i
2
3u
2
i 1
i
(r. j) = 2rc
i
2
3u
2
, 1
u
(r. j) = 32jc
i
2
3u
2
, so 1
i
(1. 31) = 2, 1
u
(1. 31) = 2.
By Equation 2, an equation of the tangent plane is : 31 = 1
i
(1. 31)(r 31) +1
u
(1. 31)[j 3(31)] i
: 31 = 2(r 31) + 2(j + 1) or : = 2r + 2j + 1.
7. : = 1(r. j) = r
2
+rj + 3j
2
, so 1
i
(r. j) = 2r +j i 1
i
(1. 1) = 3, 1
u
(r. j) = r + 6j i 1
u
(1. 1) = 7 and an
equation of the tangent plane is : 35 = 3(r 31) + 7(j 31) or : = 3r + 7j 35. After zooming in, the surface and the
tangent plane become almost indistinguishable. (Here, the tangent plane is below the surface.) If we zoom in farther, the
surface and the tangent plane will appear to coincide.
9. 1(r. j) =
rj sin(r 3j)
1 +r
2
+j
2
. A CAS gives 1
i
(r. j) =
j sin(r 3j) +rj cos (r 3j)
1 +r
2
+j
2
3
2r
2
j sin(r 3j)
(1 +r
2
+j
2
)
2
and
1
u
(r. j) =
rsin(r 3j) 3rj cos (r 3j)
1 +r
2
+j
2
3
2rj
2
sin(r 3j)
(1 +r
2
+j
2
)
2
. We use the CAS to evaluate these at (1. 1), and then
substitute the results into Equation 2 to compute an equation of the tangent plane: : =
1
3
r 3
1
3
j. The surface and tangent
plane are shown in the rst graph below. After zooming in, the surface and the tangent plane become almost indistinguishable,
SECTION 15.4 TANGENT PLANES AND LINEAR APPROXIMATIONS ET SECTION 14.4

185
as shown in the second graph. (Here, the tangent plane is shown with fewer traces than the surface.) If we zoom in farther, the
surface and the tangent plane will appear to coincide.
11. 1(r. j) = r
j. The partial derivatives are 1

i
(r. j) =
j and 1
u
(r. j) =
r
2
j
, so 1
i
(1. 4) = 2 and 1
u
(1. 4) =
1
4
. Both
1
i
and 1
u
are continuous functions for j 0, so by Theorem 8, 1 is differentiable at (1. 4). By Equation 3, the linearization of
1 at (1. 4) is given by 1(r. j) = 1(1. 4) +1
i
(1. 4)(r 31) +1
u
(1. 4)(j 34) = 2 + 2(r 31) +
1
4
(j 34) = 2r +
1
4
j 31.
13. 1(r. j) =
r
r +j
. The partial derivatives are 1
i
(r. j) =
1(r +j) 3r(1)
(r +j)
2
= j(r +j)
2
and
1
u
(r. j) = r(31)(r +j)
32
1 = 3r(r +j)
2
, so 1
i
(2. 1) =
1
9
and 1
u
(2. 1) = 3
2
9
. Both 1
i
and 1
u
are continuous
functions for j 6= 3r, so 1 is differentiable at (2. 1) by Theorem 8. The linearization of 1 at (2. 1) is given by
1(r. j) = 1(2. 1) +1
i
(2. 1)(r 32) +1
u
(2. 1)(j 31) =
2
3
+
1
9
(r 32) 3
2
9
(j 31) =
1
9
r 3
2
9
j +
2
3
.
15. 1(r. j) = c
3iu
cos j. The partial derivatives are 1
i
(r. j) = c
3iu
(3j) cos j = 3jc
3iu
cos j and
1
u
(r. j) = c
3iu
(3sinj) + (cos j)c
3iu
(3r) = 3c
3iu
(sinj +rcos j), so 1
i
(. 0) = 0 and 1
u
(. 0) = 3.
Both 1
i
and 1
u
are continuous functions, so 1 is differentiable at (. 0), and the linearization of 1 at (. 0) is
1(r. j) = 1(. 0) +1
i
(. 0)(r 3) +1
u
(. 0)(j 30) = 1 + 0(r 3) 3(j 30) = 1 3j.
17. Let 1(r. j) =
2r + 3
4j + 1
. Then 1
i
(r. j) =
2
4j + 1
and 1
u
(r. j) = (2r +3)(31)(4j +1)
32
(4) =
38r 312
(4j + 1)
2
. Both 1
i
and 1
u
are continuous functions for j 6= 3
1
4
, so by Theorem 8, 1 is differentiable at (0. 0). We have 1
i
(0. 0) = 2, 1
u
(0. 0) = 312
and the linear approximation of 1 at (0. 0) is 1(r. j) E 1(0. 0) +1
i
(0. 0)(r 30) +1
u
(0. 0)(j 30) = 3 + 2r 312j.
19. 1(r. j) =

20 3r
2
37j
2
i 1
i
(r. j) = 3
r
20 3r
2
37j
2
and 1
u
(r. j) = 3
7j
20 3r
2
37j
2
,
so 1
i
(2. 1) = 3
2
3
and 1
u
(2. 1) = 3
7
3
. Then the linear approximation of 1 at (2. 1) is given by
1 (r. j) E 1(2. 1) +1
i
(2. 1)(r 32) +1
u
(2. 1)(j 31) = 3 3
2
3
(r 32) 3
7
3
(j 31) = 3
2
3
r 3
7
3
j +
20
3
.
Thus 1(1.95. 1.08) E 3
2
3
(1.95) 3
7
3
(1.08) +
20
3
= 2.84
6.
21. 1(r. j. :) =

r
2
+j
2
+:
2
i 1
i
(r. j. :) =
r
r
2
+j
2
+:
2
, 1
u
(r. j. :) =
j
r
2
+j
2
+:
2
, and
1
:
(r. j. :) =
:
r
2
+j
2
+:
2
, so 1
i
(3. 2. 6) =
3
7
, 1
u
(3. 2. 6) =
2
7
, 1
:
(3. 2. 6) =
6
7
. Then the linear approximation of 1 at
(3. 2. 6) is given by
1(r. j. :) E 1(3. 2. 6) +1
i
(3. 2. 6)(r 33) +1
u
(3. 2. 6)(j 32) +1
:
(3. 2. 6)(: 36)
= 7 +
3
7
(r 33) +
2
7
(j 32) +
6
7
(: 36) =
3
7
r +
2
7
j +
6
7
:
Thus
(3.02)
2
+ (1.97)
2
+ (5.99)
2
= 1(3.02. 1.97. 5.99) E
3
7
(3.02) +
2
7
(1.97) +
6
7
(5.99) E 6.9914.
23. From the table, 1(94. 80) = 127. To estimate 1
T
(94. 80) and 1
1
(94. 80) we follow the procedure used in Section 15.3
[ET 14.3]. Since 1
T
(94. 80) = lim
I<0
1(94 +/. 80) 31(94. 80)
/
, we approximate this quantity with / = 2 and use the
values given in the table:
1
T
(94. 80) E
1(96. 80) 31(94. 80)
2
=
135 3127
2
= 4, 1
T
(94. 80) E
1(92. 80) 31(94. 80)
32
=
119 3127
32
= 4
Averaging these values gives 1
T
(94. 80) E 4. Similarly, 1
1
(94. 80) = lim
I<0
1(94. 80 +/) 31(94. 80)
/
, so we use / = 5:
1
1
(94. 80) E
1(94. 85) 31(94. 80)
5
=
132 3127
5
= 1, 1
1
(94. 80) E
1(94. 75) 31(94. 80)
35
=
122 3127
35
= 1
Averaging these values gives 1
1
(94. 80) E 1. The linear approximation, then, is
1(T. 1) E 1(94. 80) +1
T
(94. 80)(T 394) +1
1
(94. 80)(1 380)
E 127 + 4(T 394) + 1(1 380) [or 4T +1 3329]
Thus when T = 95 and 1 = 78, 1(95. 78) E 127 + 4(95 394) + 1(78 380) = 129, so we estimate the heat index to be
approximately 129
F.
25. : = r
3
ln(j
2
) i d: =
u:
ur
dr +
u:
uj
dj = 3r
2
ln(j
2
) dr +r
3
1
j
2
(2j) dj = 3r
2
ln(j
2
) dr +
2r
3
j
dj
27. i = j
5
q
3
i di =
ui
uj
dj +
ui
uq
dq = 5j
4
q
3
dj + 3j
5
q
2
dq
29. 1 = co
2
cos ~ i d1 =
u1
uc
dc +
u1
uo
do +
u1
u~
d~ = o
2
cos ~ dc + 2co cos ~ do 3co
2
sin~ d~
31. dr = {r = 0.05, dj = {j = 0.1, : = 5r
2
+j
2
, :
i
= 10r, :
u
= 2j. Thus when r = 1 and j = 2,
d: = :
i
(1. 2) dr + :
u
(1. 2) dj = (10)(0.05) + (4)(0.1) = 0.9 while
{: = 1(1.05. 2.1) 31(1. 2) = 5(1.05)
2
+ (2.1)
2
35 34 = 0.9225.
33. d =
u
ur
dr +
u
uj
dj = j dr +rdj and |{r| $ 0.1, |{j| $ 0.1. We use dr = 0.1, dj = 0.1 with r = 30, j = 24;
then the maximum error in the area is about d = 24(0.1) + 30(0.1) = 5.4 cm
2
.
35. The volume of a can is \ = v
2
/ and {\ E d\ is an estimate of the amount of tin. Here d\ = 2v/dv +v
2
d/, so put
dv = 0.04, d/ = 0.08 (0.04 on top, 0.04 on bottom) and then {\ E d\ = 2(48)(0.04) +(16)(0.08) E 16.08 cm
3
.
Thus the amount of tin is about 16 cm
3
.
SECTION 15.5 THE CHAIN RULE ET SECTION 14.5

187
37. The area of the rectangle is = rj, and { E d is an estimate of the area of paint in the stripe. Here d = j dr +rdj,
so with dr = dj =
3+3
12
=
1
2
, { E d = (100)
1
2
+ (200)
1
2
= 150 ft
2
. Thus there are approximately 150 ft
2
of paint
in the stripe.
39. First we nd
u1
u1
1
implicitly by taking partial derivatives of both sides with respect to 1
1
:
u
u1
1
1
1
=
u [(11
1
) + (11
2
) + (11
3
)]
u1
1
i 31
32
u1
u1
1
= 31
32
1
i
u1
u1
1
=
1
2
1
2
1
. Then by symmetry,
u1
u1
2
=
1
2
1
2
2
,
u1
u1
3
=
1
2
1
2
3
. When 1
1
= 25, 1
2
= 40 and 1
3
= 50,
1
1
=
17
200
C 1 =
200
17
l.
Since the possible error for each 1
.
is 0.5%, the maximum error of 1 is attained by setting {1
.
= 0.0051
.
. So
{1 E d1 =
u1
u1
1
{1
1
+
u1
u1
2
{1
2
+
u1
u1
3
{1
3
= (0.005)1
2
1
1
1
+
1
1
2
+
1
1
3
= (0.005)1 =
1
17
E 0.059 l.
41. The errors in measurement are at most 2%, so
{n
n
$ 0.02 and
{/
/
$ 0.02. The relative error in the calculated surface

area is
{o
o
E
do
o
=
0.1091(0.425n
0.42531
)/
0.725
dn + 0.1091n
0.425
(0.725/
0.72531
) d/
0.1091n
0.425
/
0.725
= 0.425
dn
n
+ 0.725
d/
/
To estimate the maximum relative error, we use
dn
n
=
{n
n
= 0.02 and
d/
/
=
{/
/
= 0.02 i
do
o
= 0.425 (0.02) + 0.725 (0.02) = 0.023. Thus the maximum percentage error is approximately 2.3%.
43. {: = 1(o +{r. / +{j) 31(o. /) = (o +{r)
2
+ (/ +{j)
2
3(o
2
+/
2
)
= o
2
+ 2o {r + ({r)
2
+/
2
+ 2/ {j + ({j)
2
3o
2
3/
2
= 2o {r + ({r)
2
+ 2/ {j + ({j)
2
But 1
i
(o. /) = 2o and 1
u
(o. /) = 2/ and so {: = 1
i
(o. /) {r +1
u
(o. /) {j +{r{r +{j {j, which is Denition 7
with -
1
= {r and -
2
= {j. Hence 1 is differentiable.
45. To show that 1 is continuous at (o. /) we need to show that lim
(i.u)<(a.l)
1(r. j) = 1(o. /) or
equivalently lim
({i.{u)<(0.0)
1(o +{r. / +{j) = 1(o. /). Since 1 is differentiable at (o. /),
1(o +{r. / +{j) 31(o. /) = {: = 1
i
(o. /) {r +1
u
(o. /) {j +-
1
{r +-
2
{j, where c
1
and c
2
<0 as
({r. {j) <(0. 0). Thus 1(o +{r. / +{j) = 1(o. /) +1
i
(o. /) {r +1
u
(o. /) {j +-
1
{r +-
2
{j. Taking the limit of
both sides as ({r. {j) <(0. 0) gives lim
({i.{u)<(0.0)
1(o +{r. / +{j) = 1(o. /). Thus 1 is continuous at (o. /).
15.5 The Chain Rule ET 14.5
1. : = r
2
+j
2
+rj, r = sint, j = c
I
i
d:
dt
=
u:
ur
dr
dt
+
u:
uj
dj
dt
= (2r +j) cos t + (2j +r)c
I
3. : =

1 +r
2
+j
2
, r = lnt, j = cos t i
d:
dt
=
u:
ur
dr
dt
+
u:
uj
dj
dt
=
1
2
(1+r
2
+j
2
)
31'2
(2r)
1
t
+
1
2
(1+r
2
+j
2
)
31'2
(2j)(3sint) =
1
1 +r
2
+j
2
r
t
3j sint

5. n = rc
u':
, r = t
2
, j = 1 3t, : = 1 + 2t i
dn
dt
=
un
ur
dr
dt
+
un
uj
dj
dt
+
un
u:
d:
dt
= c
u':
2t +rc
u':
1
:
(31) +rc
u':
3
j
:
2
2 = c
u':
2t 3
r
:
3
2rj
:
2
7. : = r
2
j
3
, r = c cos t, j = c sint i
u:
uc
=
u:
ur
ur
uc
+
u:
uj
uj
uc
= 2rj
3
cos t + 3r
2
j
2
sint
u:
ut
=
u:
ur
ur
ut
+
u:
uj
uj
ut
= (2rj
3
)(3c sint) + (3r
2
j
2
)(c cos t) = 32crj
3
sint + 3cr
2
j
2
cos t
9. : = sin0 cos c, 0 = ct
2
, c = c
2
t i
u:
uc
=
u:
u0
u0
uc
+
u:
uc
uc
uc
= (cos 0 cos c)(t
2
) + (3sin0 sinc)(2ct) = t
2
cos 0 cos c 32ct sin0 sinc
u:
ut
=
u:
u0
u0
ut
+
u:
uc
uc
ut
= (cos 0 cos c)(2ct) + (3sin0 sinc)(c
2
) = 2ct cos 0 cos c 3c
2
sin0 sinc
11. : = c
r
cos 0, v = ct, 0 =
I
c
2
+t
2
i
u:
uc
=
u:
uv
uv
uc
+
u:
u0
u0
uc
= c
r
cos 0 t +c
r
(3sin0)
1
2
(c
2
+t
2
)
31'2
(2c) = tc
r
cos 0 3c
r
sin0
c
I
c
2
+t
2
= c
r
t cos 0 3
c
I
c
2
+t
2
sin0
u:
ut
=
u:
uv
uv
ut
+
u:
u0
u0
ut
= c
r
cos 0 c +c
r
(3sin0)
1
2
(c
2
+t
2
)
31'2
(2t) = cc
r
cos 0 3c
r
sin0
t
I
c
2
+t
2
= c
r
c cos 0 3
t
I
c
2
+t
2
sin0
13. When t = 3, r = o(3) = 2 and j = /(3) = 7. By the Chain Rule (2),

d:
dt
=
u1
ur
dr
dt
+
u1
uj
dj
dt
= 1
i
(2. 7)o
0
(3) +1
u
(2. 7) /
0
(3) = (6)(5) + (38)(34) = 62.
15. o(&. ) = 1(r(&. ). j(&. )) where r = c
u
+ sin, j = c
u
+ cos i
ur
u&
= c
u
,
ur
u
= cos ,
uj
u&
= c
u
,
uj
u
= 3sin. By the Chain Rule (3),
uo
u&
=
u1
ur
ur
u&
+
u1
uj
uj
u&
. Then
o
u
(0. 0) = 1
i
(r(0. 0). j(0. 0)) r
u
(0. 0) +1
u
(r(0. 0). j(0. 0)) j
u
(0. 0) = 1
i
(1. 2)(c
0
) +1
u
(1. 2)(c
0
) = 2(1) + 5(1) = 7.
Similarly,
uo
u
=
u1
ur
ur
u
+
u1
uj
uj
u
. Then
o
r
(0. 0) = 1
i
(r(0. 0). j(0. 0)) r
r
(0. 0) +1
u
(r(0. 0). j(0. 0)) j
r
(0. 0) = 1
i
(1. 2)(cos 0) +1
u
(1. 2)(3sin0)
= 2(1) + 5(0) = 2
17. & = 1(r. j), r = r(v. c. t), j = j(v. c. t) i
u&
uv
=
u&
ur
ur
uv
+
u&
uj
uj
uv
,
u&
uc
=
u&
ur
ur
uc
+
u&
uj
uj
uc
,
u&
ut
=
u&
ur
ur
ut
+
u&
uj
uj
ut

189
19. n = 1(v. c. t), v = v(r. j), c = c(r. j), t = t(r. j) i
un
ur
=
un
uv
uv
ur
+
un
uc
uc
ur
+
un
ut
ut
ur
,
un
uj
=
un
uv
uv
uj
+
un
uc
uc
uj
+
un
ut
ut
uj
21. : = r
2
+rj
3
, r = &
2
+n
3
, j = & +c
u
i
u:
u&
=
u:
ur
ur
u&
+
u:
uj
uj
u&
= (2r +j
3
)(
2
) + (3rj
2
)(1),
u:
u
=
u:
ur
ur
u
+
u:
uj
uj
u
= (2r +j
3
)(2&) + (3rj
2
)(c
u
),
u:
un
=
u:
ur
ur
un
+
u:
uj
uj
un
= (2r +j
3
)(3n
2
) + (3rj
2
)(c
u
).
When & = 2, = 1, and n = 0, we have r = 2, j = 3,
so
u:
u&
= (31)(1) + (54)(1) = 85,
u:
u
= (31) (4) + (54)(1) = 178,
u:
un
= (31)(0) + (54)(1) = 54.
23. 1 = ln(&
2
+
2
+n
2
), & = r + 2j, = 2r 3j, n = 2rj i
u1
ur
=
u1
u&
u&
ur
+
u1
u
u
ur
+
u1
un
un
ur
=
2&
&
2
+
2
+n
2
(1) +
2
&
2
+
2
+n
2
(2) +
2n
&
2
+
2
+n
2
(2j)
=
2& + 4 + 4nj
&
2
+
2
+n
2
,
u1
uj
=
u1
u&
u&
uj
+
u1
u
u
uj
+
u1
un
un
uj
=
2&
&
2
+
2
+n
2
(2) +
2
&
2
+
2
+n
2
(31) +
2n
&
2
+
2
+n
2
(2r)
=
4& 32 + 4nr
&
2
+
2
+n
2
.
When r = j = 1 we have & = 3, = 1, and n = 2, so
u1
ur
=
9
7
and
u1
uj
=
9
7
.
25. & = r
2
+j:, r = jv cos 0, j = jv sin0, : = j +v i
u&
uj
=
u&
ur
ur
uj
+
u&
uj
uj
uj
+
u&
u:
u:
uj
= (2r)(v cos 0) + (:)(v sin0) + (j)(1) = 2rv cos 0 +:v sin0 +j,
u&
uv
=
u&
ur
ur
uv
+
u&
uj
uj
uv
+
u&
u:
u:
uv
= (2r)(j cos 0) + (:)(j sin0) + (j)(1) = 2rj cos 0 +:j sin0 +j,
u&
u0
=
u&
ur
ur
u0
+
u&
uj
uj
u0
+
u&
u:
u:
u0
= (2r)(3jv sin0) + (:)(jv cos 0) + (j)(0) = 32rjv sin0 +:jv cos 0.
When j = 2, v = 3, and 0 = 0 we have r = 6, j = 0, and : = 5, so
u&
uj
= 36,
u&
uv
= 24, and
u&
u0
= 30.
27.

rj = 1 +r
2
j, so let 1(r. j) = (rj)
1'2
31 3r
2
j = 0. Then by Equation 6
dj
dr
= 3
1
i
1
u
= 3
1
2
(rj)
31'2
(j) 32rj
1
2
(rj)
31'2
(r) 3r
2
= 3
j 34rj
rj
r 32r
2
rj
=
4(rj)
3'2
3j
r 32r
2
rj
.
29. cos(r 3j) = rc
u
, so let 1(r. j) = cos(r 3j) 3rc
u
= 0.
Then
dj
dr
= 3
1
i
1
u
= 3
3sin(r 3j) 3c
u
3sin(r 3j)(31) 3rc
u
=
sin(r 3j) +c
u
sin(r 3j) 3rc
u
.
31. r
2
+j
2
+:
2
= 3rj:, so let 1(r. j. :) = r
2
+j
2
+:
2
33rj: = 0. Then by Equations 7
u:
ur
= 3
1
i
1
:
= 3
2r 33j:
2: 33rj
=
3j: 32r
2: 33rj
and
u:
uj
= 3
1
u
1
:
= 3
2j 33r:
2: 33rj
=
3r: 32j
2: 33rj
.
33. r 3: = arctan(j:), so let 1(r. j. :) = r 3: 3arctan(j:) = 0. Then
u:
ur
= 3
1
i
1
:
= 3
1
31 3
1
1 + (j:)
2
(j)
=
1 +j
2
:
2
1 +j +j
2
:
2
u:
uj
= 3
1
u
1
:
= 3
3
1
1 + (j:)
2
(:)
31 3
1
1 + (j:)
2
(j)
= 3
:
1 +j
2
:
2
1 +j
2
:
2
+j
1 +j
2
:
2
= 3
:
1 +j +j
2
:
2
35. Since r and j are each functions of t, T(r. j) is a function of t, so by the Chain Rule,
dT
dt
=
uT
ur
dr
dt
+
uT
uj
dj
dt
. After
3 seconds, r =
I
1 +t =
I
1 + 3 = 2, j = 2 +
1
3
t = 2 +
1
3
(3) = 3,
dr
dt
=
1
2
I
1 +t
=
1
2
I
1 + 3
=
1
4
, and
dj
dt
=
1
3
.
Then
dT
dt
= T
i
(2. 3)
dr
dt
+T
u
(2. 3)
dj
dt
= 4
1
4
+ 3
1
3
= 2. Thus the temperature is rising at a rate of 2
Cs.
37. C = 1449.2 + 4.6T 30.055T
2
+ 0.00029T
3
+ 0.0161, so
uC
uT
= 4.6 30.11T + 0.00087T
2
and
uC
u1
= 0.016.
According to the graph, the diver is experiencing a temperature of approximately 12.5
C at t = 20 minutes, so
uC
uT
= 4.6 30.11(12.5) + 0.00087(12.5)
2
E 3.36. By sketching tangent lines at t = 20 to the graphs given, we estimate
d1
dt
E
1
2
and
dT
dt
E 3
1
10
. Then, by the Chain Rule,
dC
dt
=
uC
uT
dT
dt
+
uC
u1
d1
dt
E (3.36)
3
1
10
+ (0.016)
1
2
E 30.33.
Thus the speed of sound experienced by the diver is decreasing at a rate of approximately 0.33 ms per minute.
39. (a) \ = /n/, so by the Chain Rule,
d\
dt
=
u\
u/
d/
dt
+
u\
un
dn
dt
+
u\
u/
d/
dt
= n/
d/
dt
+//
dn
dt
+/n
d/
dt
= 2 2 2 + 1 2 2 + 1 2 (33) = 6 m
3
s.
(b) o = 2(/n +// +n/), so by the Chain Rule,
do
dt
=
uo
u/
d/
dt
+
uo
un
dn
dt
+
uo
u/
d/
dt
= 2(n +/)
d/
dt
+ 2(/ +/)
dn
dt
+ 2(/ +n)
d/
dt
= 2(2 + 2)2 + 2(1 + 2)2 + 2(1 + 2)(33) = 10 m
2
s
(c) 1
2
= /
2
+n
2
+/
2
i 21
d1
dt
= 2/
d/
dt
+ 2n
dn
dt
+ 2/
d/
dt
= 2(1)(2) + 2(2)(2) + 2(2)(33) = 0 i
d1dt = 0 ms.
41.
d1
dt
= 0.05,
dT
dt
= 0.15, \ = 8.31
T
1
and
d\
dt
=
8.31
1
dT
dt
38.31
T
1
2
d1
dt
. Thus when 1 = 20 and T = 320,
d\
dt
= 8.31
0.15
20
3
(0.05)(320)
400
E 30.27 Ls.
43. Let r be the length of the rst side of the triangle and j the length of the second side. The area of the triangle is given by
=
1
2
rj sin0 where 0 is the angle between the two sides. Thus is a function of r, j, and 0, and r, j, and 0 are each in

191
turn functions of time t. We are given that
dr
dt
= 3,
dj
dt
= 32, and because is constant,
d
dt
= 0. By the Chain Rule,
d
dt
=
u
ur
dr
dt
+
u
uj
dj
dt
+
u
u0
d0
dt
i
d
dt
=
1
2
j sin0
dr
dt
+
1
2
rsin0
dj
dt
+
1
2
rj cos 0
d0
dt
. When r = 20, j = 30,
and 0 = 6 we have
0 =
1
2
(30)
sin
r
6
(3) +
1
2
(20)
sin
r
6
(32) +
1
2
(20)(30)
cos
r
6
d0
dt
= 45
1
2
320
1
2
+ 300
I
3
2

d0
dt
=
25
2
+ 150
I
3
d0
dt
Solving for
d0
dt
gives
d0
dt
=
3252
150
I
3
= 3
1
12
I
3
, so the angle between the sides is decreasing at a rate of
1
12
I
3
E 0.048 rads.
45. (a) By the Chain Rule,
u:
uv
=
u:
ur
cos 0 +
u:
uj
sin0,
u:
u0
=
u:
ur
(3v sin0) +
u:
uj
v cos 0.
(b)

u:
uv
2
=

u:
ur
2
cos
2
0 + 2
u:
ur
u:
uj
cos 0 sin0 +
u:
uj
2
sin
2
0,
u:
u0
2
=

u:
ur
2
v
2
sin
2
0 3 2
u:
ur
u:
uj
v
2
cos 0 sin0 +
u:
uj
2
v
2
cos
2
0. Thus
u:
uv
2
+
1
v
2
u:
u0
2
=
u:
ur
2
+
u:
uj
(cos
2
0 + sin
2
0) =
u:
ur
2
+
u:
uj
2
.
47. Let & = r 3j. Then
u:
ur
=
d:
d&
u&
ur
=
d:
d&
and
u:
uj
=
d:
d&
(31). Thus
u:
ur
+
u:
uj
= 0.
49. Let & = r +ot, = r 3ot. Then : = 1(&) +o(), so u:u& = 1
0
(&) and u:u = o
0
().
Thus
u:
ut
=
u:
u&
u&
ut
+
u:
u
u
ut
= o1
0
(&) 3oo
0
() and
u
2
:
ut
2
= o
u
ut
[1
0
(&) 3o
0
()] = o
d1
0
(&)
d&
u&
ut
3
do
0
()
d
u
ut
= o
2
1
00
(&) +o
2
o
00
().
Similarly
u:
ur
= 1
0
(&) +o
0
() and
u
2
:
ur
2
= 1
00
(&) +o
00
(). Thus
u
2
:
ut
2
= o
2
u
2
:
ur
2
.
51.
u:
uc
=
u:
ur
2c +
u:
uj
2v. Then
u
2
:
uv uc
=
u
uv
u:
ur
2c
+
u
uv
u:
uj
2v
=
u
2
:
ur
2
ur
uv
2c +
u
uj
u:
ur
uj
uv
2c +
u:
ur
u
uv
2c +
u
2
:
uj
2
uj
uv
2v +
u
ur
u:
uj
ur
uv
2v +
u:
uj
2
= 4vc
u
2
:
ur
2
+
u
2
:
uj ur
4c
2
+ 0 + 4vc
u
2
:
uj
2
+
u
2
:
uruj
4v
2
+ 2
u:
uj
By the continuity of the partials,
u
2
:
uvuc
= 4vc
u
2
:
ur
2
+ 4vc
u
2
:
uj
2
+ (4v
2
+ 4c
2
)
u
2
:
uruj
+ 2
u:
uj
.
53.
u:
uv
=
u:
ur
cos 0 +
u:
uj
sin0 and
u:
u0
= 3
u:
ur
v sin0 +
u:
uj
v cos 0. Then
u
2
:
uv
2
= cos 0
u
2
:
ur
2
cos 0 +
u
2
:
uj ur
sin0
+ sin0
u
2
:
uj
2
sin0 +
u
2
:
uruj
cos 0
= cos
2
0
u
2
:
ur
2
+ 2 cos 0 sin0
u
2
:
uruj
+ sin
2
0
u
2
:
uj
2
and
u
2
:
u0
2
= 3v cos 0
u:
ur
+ (3v sin0)
u
2
:
ur
2
(3v sin0) +
u
2
:
uj ur
v cos 0
3v sin0
u:
uj
+v cos 0
u
2
:
uj
2
v cos 0 +
u
2
:
uruj
(3v sin0)
= 3v cos 0
u:
ur
3v sin0
u:
uj
+v
2
sin
2
0
u
2
:
ur
2
32v
2
cos 0 sin0
u
2
:
uruj
+v
2
cos
2
0
u
2
:
uj
2
Thus
u
2
:
uv
2
+
1
v
2
u
2
:
u0
2
+
1
v
u:
uv
= (cos
2
0 + sin
2
0)
u
2
:
ur
2
+
sin
2
0 + cos
2
0
u
2
:
uj
2
3
1
v
cos 0
u:
ur
3
1
v
sin0
u:
uj
+
1
v
cos 0
u:
ur
+ sin0
u:
uj
=
u
2
:
ur
2
+
u
2
:
uj
2
as desired.
55. (a) Since 1 is a polynomial, it has continuous second-order partial derivatives, and
1(tr. tj) = (tr)
2
(tj) + 2(tr)(tj)
2
+ 5(tj)
3
= t
3
r
2
j + 2t
3
rj
2
+ 5t
3
j
3
= t
3
(r
2
j + 2rj
2
+ 5j
3
) = t
3
1 (r. j).
Thus, 1 is homogeneous of degree 3.
(b) Differentiating both sides of 1(tr. tj) = t
n
1(r. j) with respect to t using the Chain Rule, we get
u
ut
1(tr. tj) =
u
ut
[t
n
1(r. j)] C
u
u(tr)
1(tr. tj)
u(tr)
ut
+
u
u(tj)
1(tr. tj)
u(tj)
ut
= r
u
u(tr)
1(tr. tj) +j
u
u(tj)
1(tr. tj) = nt
n31
1(r. j).
Setting t = 1: r
u
ur
1(r. j) +j
u
uj
1(r. j) = n1(r. j).
57. Differentiating both sides of 1(tr. tj) = t
n
1(r. j) with respect to r using the Chain Rule, we get
u
ur
1(tr. tj) =
u
ur
[t
n
1(r. j)] C
u
u (tr)
1(tr. tj)
u (tr)
ur
+
u
u (tj)
1(tr. tj)
u (tj)
ur
= t
n
u
ur
1(r. j) C t1
i
(tr. tj) = t
n
1
i
(r. j).
Thus 1
i
(tr. tj) = t
n31
1
i
(r. j).
15.6 Directional Derivatives and the Gradient Vector ET 14.6
1. We can approximate the directional derivative of the pressure function at K in the direction of S by the average rate of change
of pressure between the points where the red line intersects the contour lines closest to K (extend the red line slightly at the
left). In the direction of S, the pressure changes from 1000 millibars to 996 millibars and we estimate the distance between
these two points to be approximately 50 km (using the fact that the distance from K to S is 300 km). Then the rate of change of
pressure in the direction given is approximately
996 31000
50
= 30.08 millibarkm.
SECTION 15.6 DIRECTIONAL DERIVATIVES AND THE GRADIENT VECTOR ET SECTION 14.6

193
3. 1
u
1(320. 30) = Q1 (320. 30) u = 1
T
(320. 30)
1
I
2
+ 1
r
(320. 30)
1
I
2
.
1
T
(320. 30) = lim
I<0
1(320 +/. 30) 31(320. 30)
/
, so we can approximate 1
T
(320. 30) by considering / = 5 and
using the values given in the table: 1
T
(320. 30) E
1(315. 30) 31(320. 30)
5
=
326 3(333)
5
= 1.4,
1
T
(320. 30) E
1(325. 30) 31(320. 30)
35
=
339 3(333)
35
= 1.2. Averaging these values gives 1
T
(320. 30) E 1.3.
Similarly, 1
r
(320. 30) = lim
I<0
1(320. 30 +/) 31(320. 30)
/
, so we can approximate 1
r
(320. 30) with / = 10:
1
r
(320. 30) E
1(320. 40) 31(320. 30)
10
=
334 3(333)
10
= 30.1,
1
r
(320. 30) E
1(320. 20) 31(320. 30)
310
=
330 3(333)
310
= 30.3. Averaging these values gives 1
r
(320. 30) E 30.2.
Then 1
u
1(320. 30) E 1.3
1
I
2
+ (30.2)
1
I
2
E 0.778.
5. 1(r. j) = jc
3i
i 1
i
(r. j) = 3jc
3i
and 1
u
(r. j) = c
3i
. If u is a unit vector in the direction of 0 = 23, then
from Equation 6, 1
u
1(0. 4) = 1
i
(0. 4) cos
2r
3
+1
u
(0. 4) sin
2r
3
= 34
3
1
2
+ 1
I
3
2
= 2 +
I
3
2
.
7. 1(r. j) = sin(2r + 3j)
(a) Q1(r. j) =
u1
ur
i +
u1
uj
j = [cos(2r + 3j) 2] i + [cos(2r + 3j) 3] j = 2 cos (2r + 3j) i + 3 cos (2r + 3j) j
(b) Q1(36. 4) = (2 cos 0)i + (3 cos 0)j = 2 i + 3 j
(c) By Equation 9, 1
u
1(36. 4) = Q1(36. 4) u = (2 i + 3 j)
1
2
I
3 i 3j
=
1
2
2
I
3 33
=
I
3 3
3
2
.
9. 1(r. j. :) = rc
2u:
(a) Q1(r. j. :) = h1
i
(r. j. :). 1
u
(r. j. :). 1
:
(r. j. :)i =

c
2u:
. 2r:c
2u:
. 2rjc
2u:
(b) Q1(3. 0. 2) = h1. 12. 0i

(c) By Equation 14, 1
u
1(3. 0. 2) = Q1(3. 0. 2) u = h1. 12. 0i
2
3
. 3
2
3
.
1
3
=
2
3
3
24
3
+ 0 = 3
22
3
.
11. 1(r. j) = 1 + 2r
j i Q1(r. j) =
j. 2r
1
2
j
31'2
=

2
j. r
, Q1(3. 4) =
4.
3
2
, and a unit vector in

the direction of v is u =
1
4
2
+ (33)
2
h4. 33i =

4
5
. 3
3
5
, so 1
u
1(3. 4) = Q1(3. 4) u =

4.
3
2
4
5
. 3
3
5
=
23
10
.
13. o(j. q) = j
4
3j
2
q
3
i Qo(j. q) =

4j
3
32jq
3
i +
33j
2
q
2
j, Qo(2. 1) = 28 i 312 j, and a unit

vector in the direction of v is u =
1
I
1
2
+3
2
(i + 3 j) =
1
I
10
(i + 3 j), so
1
u
o(2. 1) = Qo(2. 1) u = (28 i 312 j)
1
I
10
(i + 3 j) =
1
I
10
(28 336) = 3
8
I
10
or 3
4
I
10
5
.
15. 1(r. j. :) = rc
u
+jc
:
+:c
i
i Q1(r. j. :) = hc
u
+:c
i
. rc
u
+c
:
. jc
:
+c
i
i, Q1(0. 0. 0) = h1. 1. 1i, and a unit
vector in the direction of v is u =
1
I
25+1+4
h5. 1. 32i =
1
I
30
h5. 1. 32i, so
1
u
1(0. 0. 0) = Q1(0. 0. 0) u = h1. 1. 1i
1
I
30
h5. 1. 32i =
4
I
30
.
17. o(r. j. :) = (r + 2j + 3:)
3'2
i
Qo(r. j. :) =

3
2
(r + 2j + 3:)
1'2
(1).
3
2
(r + 2j + 3:)
1'2
(2).
3
2
(r + 2j + 3:)
1'2
(3)
=

3
2
I
r + 2j + 3:. 3
I
r + 2j + 3:.
9
2
I
r + 2j + 3:
, Qo(1. 1. 2) =
9
2
. 9.
27
2
,
and a unit vector in the direction of v = 2 j 3 k is u =
2
I
5
j 3
1
I
5
k, so
1
u
o(1. 1. 2) =
9
2
. 9.
27
2
0.
2
I
5
. 3
1
I
5
=
18
I
5
3
27
2
I
5
=
9
2
I
5
.
19. 1(r. j) =

rj i Q1(r. j) =

1
2
(rj)
31'2
(j).
1
2
(rj)
31'2
(r)
j
2
rj
.
r
2
rj
, so Q1(2. 8) =

1.
1
4
.
The unit vector in the direction of
33<
1Q = h5 32. 4 38i = h3. 34i is u =

3
5
. 3
4
5
, so
1
u
1(2. 8) = Q1(2. 8) u =

1.
1
4
3
5
. 3
4
5
=
2
5
.
21. 1(r. j) = j
2
r = j
2
r
31
i Q1(r. j) =
3j
2
r
32
. 2jr
31
=

3j
2
r
2
. 2jr
.
Q1(2. 4) = h34. 4i, or equivalently h31. 1i, is the direction of maximum rate of change, and the maximum rate
is |Q1(2. 4)| =
I
16 + 16 = 4
I
2.
23. 1(r. j) = sin(rj) i Q1(r. j) = hj cos(rj). rcos(rj)i, Q1(1. 0) = h0. 1i. Thus the maximum rate of change is
|Q1(1. 0)| = 1 in the direction h0. 1i.
25. 1(r. j. :) =

r
2
+j
2
+:
2
i
Q1(r. j. :) =

1
2
(r
2
+j
2
+:
2
)
31'2
2r.
1
2
(r
2
+j
2
+:
2
)
31'2
2j.
1
2
(r
2
+j
2
+:
2
)
31'2
2:
r
2
+j
2
+:
2
.
j
r
2
+j
2
+:
2
.
:
r
2
+j
2
+:
2
.
Q1(3. 6. 32) =

3
I
49
.
6
I
49
.
32
I
49
=

3
7
.
6
7
. 3
2
7
. Thus the maximum rate of change is

|Q1(3. 6. 32)| =

3
7
2
+
6
7
2
+
3
2
7
2
=

9 +36 +4
49
= 1 in the direction
3
7
.
6
7
. 3
2
7
or equivalently h3. 6. 32i.

27. (a) As in the proof of Theorem 15, 1
u
1 = |Q1| cos 0. Since the minimum value of cos 0 is 31 occurring when 0 = , the
minimum value of 1
u
1 is 3|Q1| occurring when 0 = , that is when u is in the opposite direction of Q1
(assuming Q1 6= 0).
(b) 1(r. j) = r
4
j 3r
2
j
3
i Q1(r. j) =
4r
3
j 32rj
3
. r
4
33r
2
j
2
, so 1 decreases fastest at the point (2. 33) in the

direction 3Q1(2. 33) = 3h12. 392i = h312. 92i.
29. The direction of fastest change is Q1(r. j) = (2r 32) i + (2j 34) j, so we need to nd all points (r. j) where Q1(r. j) is
parallel to i +j C (2r 32) i + (2j 34) j = I (i +j) C I = 2r 32 and I = 2j 34. Then 2r 32 = 2j 34 i
j = r + 1. so the direction of fastest change is i +j at all points on the line j = r + 1.

195
31. T =
I
r
2
+j
2
+:
2
and 120 = T(1. 2. 2) =
I
3
so I = 360.
(a) u =
h1. 31. 1i
I
3
,
1
u
T(1. 2. 2) = QT(1. 2. 2) u =

3360
r
2
+j
2
+:
2
33'2
hr. j. :i
(1.2.2)
u = 3
40
3
h1. 2. 2i
1
I
3
h1. 31. 1i = 3
40
3
I
3
(b) From (a), QT = 3360
r
2
+j
2
+:
2
33'2
hr. j. :i, and since hr. j. :i is the position vector of the point (r. j. :), the
vector 3hr. j. :i, and thus QT, always points toward the origin.
33. Q\ (r. j. :) = h10r 33j +j:. r: 33r. rji, Q\ (3. 4. 5) = h38. 6. 12i
(a) 1
u
\ (3. 4. 5) = h38. 6. 12i
1
I
3
h1. 1. 31i =
32
I
3
(b) Q\ (3. 4. 5) = h38. 6. 12i, or equivalently, h19. 3. 6i.
(c) |Q\ (3. 4. 5)| =
I
38
2
+ 6
2
+ 12
2
=
I
1624 = 2
I
406
35. A unit vector in the direction of
3<
1 is i and a unit vector in the direction of
3<
C is j. Thus 133<
/T
1(1. 3) = 1
i
(1. 3) = 3 and
133<
/C
1(1. 3) = 1
u
(1. 3) = 26. Therefore Q1(1. 3) = h1
i
(1. 3). 1
u
(1. 3)i = h3. 26i, and by denition,
133<
/T
1(1. 3) = Q1 u where u is a unit vector in the direction of
33<
1, which is
5
13
.
12
13
. Therefore,
133<
/T
1 (1. 3) = h3. 26i
5
13
.
12
13
= 3
5
13
+ 26
12
13
=
327
13
.
37. (a) Q(o& +/) =
u(o& +/)
ur
.
u(o& +/)
uj
=

o
u&
ur
+/
u
ur
. o
u&
uj
+/
u
uj
= o
u&
ur
.
u&
uj
+/
u
ur
.
u
uj
= o Q& +/ Q
(b) Q(&) =
u&
ur
+&
u
ur
.
u&
uj
+&
u
uj
u&
ur
.
u&
uj
+&
u
ur
.
u
uj
= Q& +&Q
(c) Q
&
u&
ur
3&
u
ur
2
.

u&
uj
3&
u
uj
u&
ur
.
u&
uj
3&
u
ur
.
u
uj
2
=
Q& 3&Q
2
(d) Q&
n
=

u(&
n
)
ur
.
u(&
n
)
uj
=

n&
n31
u&
ur
. n&
n31
u&
uj
= n&
n31
Q&
39. Let 1(r. j. :) = 2(r 32)
2
+ (j 31)
2
+ (: 33)
2
. Then 2(r 32)
2
+ (j 31)
2
+ (: 33)
2
= 10 is a level surface of 1.
1
i
(r. j. :) = 4(r 32) i 1
i
(3. 3. 5) = 4, 1
u
(r. j. :) = 2(j 31) i 1
u
(3. 3. 5) = 4, and
1
:
(r. j. :) = 2(: 33) i 1
:
(3. 3. 5) = 4.
(a) Equation 19 gives an equation of the tangent plane at (3. 3. 5) as 4(r 33) + 4(j 33) + 4(: 35) = 0 C
4r + 4j + 4: = 44 or equivalently r +j +: = 11.
(b) By Equation 20, the normal line has symmetric equations
r 33
4
=
j 33
4
=
: 35
4
or equivalently
r 33 = j 33 = : 35. Corresponding parametric equations are r = 3 +t, j = 3 +t, : = 5 +t.
41. Let 1(r. j. :) = r
2
32j
2
+:
2
+j:. Then r
2
32j
2
+:
2
+j: = 2 is a level surface of 1
and Q1(r. j. :) = h2r. 34j +:. 2: +ji.
(a) Q1(2. 1. 31) = h4. 35. 31i is a normal vector for the tangent plane at (2. 1. 31), so an equation of the tangent plane
is 4(r 32) 35(j 31) 31(: + 1) = 0 or 4r 35j 3: = 4.
(b) The normal line has direction h4. 35. 31i, so parametric equations are r = 2 + 4t, j = 1 35t, : = 31 3t, and
symmetric equations are
r 32
4
=
j 31
35
=
: + 1
31
.
43. 1(r. j. :) = 3: +rc
u
cos : i Q1(r. j. :) = hc
u
cos :. rc
u
cos :. 31 3rc
u
sin:i and Q1(1. 0. 0) = h1. 1. 31i.
(a) 1(r 31) + 1(j 30) 31(: 30) = 0 or r +j 3: = 1
(b) r 31 = j = 3:
45. 1(r. j. :) = rj +j: +:r, Q1(r. j. :) = hj +:. r +:. j +ri, Q1(1. 1. 1) = h2. 2. 2i, so an equation of the tangent
plane is 2r + 2j + 2: = 6 or r +j +: = 3, and the normal line is given by r 31 = j 31 = : 31 or r = j = :. To graph
the surface we solve for :: : =
3 3rj
r +j
.
47. 1(r. j) = rj i Q1(r. j) = hj. ri, Q1(3. 2) = h2. 3i. Q1(3. 2)
is perpendicular to the tangent line, so the tangent line has equation
Q1(3. 2) hr 33. j 32i = 0 i h2. 3i hr 33. r 32i = 0 i
2(r 33) + 3(j 32) = 0 or 2r + 3j = 12.
49. Q1(r
0
. j
0
. :
0
) =

2r
0
o
2
.
2j
0
/
2
.
2:
0
c
2
. Thus an equation of the tangent plane at (r

0
. j
0
. :
0
) is
2r
0
o
2
r +
2j
0
/
2
j +
2:
0
c
2
: = 2
r
2
0
o
2
+
j
2
0
/
2
+
:
2
0
c
2
= 2(1) = 2 since (r
0
. j
0
. :
0
) is a point on the ellipsoid. Hence
r
0
o
2
r +
j
0
/
2
j +
:
0
c
2
: = 1 is an equation of the tangent plane.

197
51. Q1(r
0
. j
0
. :
0
) =

2r
0
o
2
.
2j
0
/
2
.
31
c
, so an equation of the tangent plane is

2r
0
o
2
r +
2j
0
/
2
j 3
1
c
: =
2r
2
0
o
2
+
2j
2
0
/
2
3
:
0
c
or
2r
0
o
2
r +
2j
0
/
2
j =
:
c
+ 2
r
2
0
o
2
+
j
2
0
/
2
3
:
0
c
. But
:
0
c
=
r
2
0
o
2
+
j
2
0
/
2
, so the equation can be written as
2r
0
o
2
r +
2j
0
/
2
j =
: +:
0
c
.
53. The hyperboloid r
2
3j
2
3:
2
= 1 is a level surface of 1(r. j. :) = r
2
3j
2
3:
2
and Q1 (r. j. :) = h2r. 32j. 32:i is a
normal vector to the surface and hence a normal vector for the tangent plane at (r. j. :). The tangent plane is parallel to the
plane : = r +j or r +j 3: = 0 if and only if the corresponding normal vectors are parallel, so we need a point (r
0
. j
0
. :
0
)
on the hyperboloid where h2r
0
. 32j
0
. 32:
0
i = c h1. 1. 31i or equivalently hr
0
. 3j
0
. 3:
0
i = I h1. 1. 31i for some I 6= 0.
Then we must have r
0
= I, j
0
= 3I, :
0
= I and substituting into the equation of the hyperboloid gives
I
2
3(3I)
2
3I
2
= 1 C 3I
2
= 1, an impossibility. Thus there is no such point on the hyperboloid.
55. Let (r
0
. j
0
. :
0
) be a point on the cone [other than (0. 0. 0)]. Then an equation of the tangent plane to the cone at this point is
2r
0
r + 2j
0
j 32:
0
: = 2
r
2
0
+j
2
0
3:
2
0
. But r
2
0
+j
2
0
= :
2
0
so the tangent plane is given by r
0
r +j
0
j 3:
0
: = 0, a plane
which always contains the origin.
57. Let (r
0
. j
0
. :
0
) be a point on the surface. Then an equation of the tangent plane at the point is
r
2
I
r
0
+
j
2
j
0
+
:
2
I
:
0
=
I
r
0
+
j
0
+
I
:
0
2
. But
I
r
0
+
j
0
+
I
:
0
=
I
c, so the equation is
r
I
r
0
+
j
j
0
+
:
I
:
0
=
I
c. The r-, j-, and :-intercepts are
I
cr
0
,
cj
0
and
I
c:
0
respectively. (The r-intercept is found
by setting j = : = 0 and solving the resulting equation for r, and the j- and :-intercepts are found similarly.) So the sum of
the intercepts is
I
c
I
r
0
+
j
0
+
I
:
0
= c, a constant.
59. If 1(r. j. :) = : 3r
2
3j
2
and o(r. j. :) = 4r
2
+j
2
+:
2
, then the tangent line is perpendicular to both Q1 and Qo
at (31. 1. 2). The vector v = Q1 Qo will therefore be parallel to the tangent line.
We have Q1(r. j. :) = h32r. 32j. 1i i Q1(31. 1. 2) = h2. 32. 1i, and Qo(r. j. :) = h8r. 2j. 2:i i
Qo(31. 1. 2) = h38. 2. 4i. Hence v = Q1 Qo =
i j k
2 32 1
38 2 4
= 310 i 316 j 312 k.

Parametric equations are: r = 31 310t, j = 1 316t, : = 2 312t.
61. (a) The direction of the normal line of 1 is given by Q1, and that of G by QG. Assuming that
Q1 6= 0 6= QG, the two normal lines are perpendicular at 1 if Q1 QG = 0 at 1 C
hu1ur. u1uj. u1u:i huGur. uGuj. uGu:i = 0 at 1 C 1
i
G
i
+1
u
G
u
+1
:
G
:
= 0 at 1.
(b) Here 1 = r
2
+ j
2
3 :
2
and G = r
2
+ j
2
+ :
2
3 v
2
, so
Q1 QG = h2r. 2j. 32:i h2r. 2j. 2:i = 4r
2
+ 4j
2
34:
2
= 41 = 0, since the point (r. j. :) lies on the graph of
1 = 0. To see that this is true without using calculus, note that G = 0 is the equation of a sphere centered at the origin and
1 = 0 is the equation of a right circular cone with vertex at the origin (which is generated by lines through the origin).
At any point of intersection, the spheres normal line (which passes through the origin) lies on the cone, and thus is
perpendicular to the cones normal line. So the surfaces with equations 1 = 0 and G = 0 are everywhere orthogonal.
63. Let u = ho. /i and v = hc. di. Then we know that at the given point, 1
u
1 = Q1 u = o1
i
+/1
u
and
1
v
1 = Q1 v = c1
i
+d1
u
. But these are just two linear equations in the two unknowns 1
i
and 1
u
, and since u and v
are not parallel, we can solve the equations to nd Q1 = h1
i
. 1
u
i at the given point. In fact,
Q1 =

d 1
u
1 3/ 1
v
1
od 3/c
.
o 1
v
1 3c 1
u
1
od 3/c
.
15.7 Maximum and Minimum Values ET 14.7
1. (a) First we compute 1(1. 1) = 1
ii
(1. 1) 1
uu
(1. 1) 3[1
iu
(1. 1)]
2
= (4)(2) 3(1)
2
= 7. Since 1(1. 1) 0 and
1
ii
(1. 1) 0, 1 has a local minimum at (1. 1) by the Second Derivatives Test.
(b) 1(1. 1) = 1
ii
(1. 1) 1
uu
(1. 1) 3[1
iu
(1. 1)]
2
= (4)(2) 3(3)
2
= 31. Since 1(1. 1) < 0, 1 has a saddle point at (1. 1)
by the Second Derivatives Test.
3. In the gure, a point at approximately (1. 1) is enclosed by level curves which are oval in shape and indicate that as we move
away from the point in any direction the values of 1 are increasing. Hence we would expect a local minimum at or near (1. 1).
The level curves near (0. 0) resemble hyperbolas, and as we move away from the origin, the values of 1 increase in some
directions and decrease in others, so we would expect to nd a saddle point there.
To verify our predictions, we have 1(r. j) = 4 +r
3
+j
3
33rj i 1
i
(r. j) = 3r
2
33j, 1
u
(r. j) = 3j
2
33r. We
have critical points where these partial derivatives are equal to 0: 3r
2
33j = 0, 3j
2
33r = 0. Substituting j = r
2
from the
rst equation into the second equation gives 3(r
2
)
2
33r = 0 i 3r(r
3
31) = 0 i r = 0 or r = 1. Then we have
two critical points, (0. 0) and (1. 1). The second partial derivatives are 1
ii
(r. j) = 6r, 1
iu
(r. j) = 33, and 1
uu
(r. j) = 6j,
so 1(r. j) = 1
ii
(r. j) 1
uu
(r. j) 3[1
iu
(r. j)]
2
= (6r)(6j) 3(33)
2
= 36rj 39. Then 1(0. 0) = 36(0)(0) 39 = 39,
and 1(1. 1) = 36(1)(1) 39 = 27. Since 1(0. 0) < 0, 1 has a saddle point at (0. 0) by the Second Derivatives Test. Since
1(1. 1) 0 and 1
ii
(1. 1) 0, 1 has a local minimum at (1. 1).
5. 1(r. j) = 9 32r + 4j 3r
2
34j
2
i 1
i
= 32 32r, 1
u
= 4 38j,
1
ii
= 32, 1
iu
= 0, 1
uu
= 38. Then 1
i
= 0 and 1
u
= 0 imply
r = 31 and j =
1
2
, and the only critical point is
31.
1
2
.
1(r. j) = 1
ii
1
uu
3(1
iu
)
2
= (32)(38) 30
2
= 16, and since
1
31.
1
2
= 16 0 and 1
ii
31.
1
2
= 32 < 0, 1
31.
1
2
= 11 is a
local maximum by the Second Derivatives Test.
SECTION 15.7 MAXIMUM AND MINIMUM VALUES ET SECTION 14.7

199
7. 1(r. j) = r
4
+j
4
34rj + 2 i 1
i
= 4r
3
34j, 1
u
= 4j
3
34r,
1
ii
= 12r
2
, 1
iu
= 34, 1
uu
= 12j
2
. Then 1
i
= 0 implies j = r
3
,
and substitution into 1
u
= 0 i r = j
3
gives r
9
3r = 0 i
r(r
8
31) = 0 i r = 0 or r = 1. Thus the critical points are (0. 0),
(1. 1), and (31. 31). Now 1(0. 0) = 0 0 3(34)
2
= 316 < 0,
so (0. 0) is a saddle point. 1(1. 1) = (12)(12) 3(34)
2
0 and
1
ii
(1. 1) = 12 0, so 1(1. 1) = 0 is a local minimum. 1(31. 31) = (12)(12) 3(34)
2
0 and
1
ii
= (31. 31) = 12 0, so 1(31. 31) = 0 is also a local minimum.
9. 1(r. j) = (1 +rj)(r +j) = r +j +r
2
j +rj
2
i
1
i
= 1 + 2rj +j
2
, 1
u
= 1 +r
2
+ 2rj, 1
ii
= 2j, 1
iu
= 2r + 2j,
1
uu
= 2r. Then 1
i
= 0 implies 1 + 2rj +j
2
= 0 and 1
u
= 0 implies
1 +r
2
+ 2rj = 0. Subtracting the second equation from the rst gives
j
2
3r
2
= 0 i j = r, but if j = r then 1 + 2rj +j
2
= 0 i
1 + 3r
2
= 0 which has no real solution. If j = 3r then
1 + 2rj +j
2
= 0 i 1 3r
2
= 0 i r = 1, so critical points are (1. 31) and (31. 1).
1(1. 31) = (32)(2) 30 < 0 and 1(31. 1) = (2)(32) 30 < 0, so (31. 1) and (1. 31) are saddle points.
11. 1(r. j) = r
3
312rj + 8j
3
i 1
i
= 3r
2
312j, 1
u
= 312r + 24j
2
,
1
ii
= 6r, 1
iu
= 312, 1
uu
= 48j. Then 1
i
= 0 implies r
2
= 4j and
1
u
= 0 implies r = 2j
2
. Substituting the second equation into the rst
gives (2j
2
)
2
= 4j i 4j
4
= 4j i 4j(j
3
31) = 0 i j = 0 or
j = 1. If j = 0 then r = 0 and if j = 1 then r = 2, so the critical points
are (0. 0) and (2. 1). 1(0. 0) = (0)(0) 3(312)
2
= 3144 < 0, so (0. 0) is a saddle point.
1(2. 1) = (12)(48) 3(312)
2
= 432 0 and 1
ii
(2. 1) = 12 0 so 1(2. 1) = 38 is a local minimum.
13. 1(r. j) = c
i
cos j i 1
i
= c
i
cos j, 1
u
= 3c
i
sinj.
Now 1
i
= 0 implies cos j = 0 or j =
r
2
+n for n an integer.
But sin
r
2
+n
6= 0, so there are no critical points.

15. 1(r. j) = (r
2
+j
2
)c
u
2
3i
2
i
1
i
= (r
2
+j
2
)c
u
2
3i
2
(32r) + 2rc
u
2
3i
2
= 2rc
u
2
3i
2
(1 3r
2
3j
2
),
1
u
= (r
2
+j
2
)c
u
2
3i
2
(2j) + 2jc
u
2
3i
2
= 2jc
u
2
3i
2
(1 +r
2
+j
2
),
1
ii
= 2rc
u
2
3i
2
(32r) + (1 3r
2
3j
2
)
2r
32rc
u
2
3i
2
+ 2c
u
2
3i
2
= 2c
u
2
3i
2
((1 3r
2
3j
2
)(1 32r
2
) 32r
2
),
1
iu
= 2rc
u
2
3i
2
(32j) + 2r(2j)c
u
2
3i
2
(1 3r
2
3j
2
) = 34rjc
u
2
3i
2
(r
2
+j
2
),
1
uu
= 2jc
u
2
3i
2
(2j) + (1 +r
2
+j
2
)
2j
2jc
u
2
3i
2
+ 2c
u
2
3i
2
= 2c
u
2
3i
2
((1 +r
2
+j
2
)(1 + 2j
2
) + 2j
2
).
1
u
= 0 implies j = 0, and substituting into 1
i
= 0 gives
2rc
3i
2
(1 3r
2
) = 0 i r = 0 or r = 1. Thus the critical points are
(0. 0) and (1. 0). Now 1(0. 0) = (2)(2) 30 0 and 1
ii
(0. 0) = 2 0,
so 1(0. 0) = 0 is a local minimum. 1(1. 0) = (34c
31
)(4c
31
) 30 < 0
so (1. 0) are saddle points.
17. 1(r. j) = j
2
32j cos r i 1
i
= 2j sinr, 1
u
= 2j 32 cos r,
1
ii
= 2j cos r, 1
iu
= 2 sinr, 1
uu
= 2. Then 1
i
= 0 implies j = 0 or
sinr = 0 i r = 0, , or 2 for 31 $ r $ 7. Substituting j = 0 into
1
u
= 0 gives cos r = 0 i r =
r
2
or
3r
2
, substituting r = 0 or r = 2
into 1
u
= 0 gives j = 1, and substituting r = into 1
u
= 0 gives j = 31.
Thus the critical points are (0. 1),
r
2
. 0
, (. 31),
3r
2
. 0
, and (2. 1).

1
r
2
. 0
= 1
3r
2
. 0
= 34 < 0 so
r
2
. 0
and
3r
2
. 0
are saddle points. 1(0. 1) = 1(. 31) = 1(2. 1) = 4 0 and

1
ii
(0. 1) = 1
ii
(. 31) = 1
ii
(2. 1) = 2 0, so 1(0. 1) = 1(. 31) = 1(2. 1) = 31 are local minima.
19. 1(r. j) = r
2
+ 4j
2
34rj + 2 i 1
i
= 2r 34j, 1
u
= 8j 34r, 1
ii
= 2, 1
iu
= 34, 1
uu
= 8. Then 1
i
= 0
and 1
u
= 0 each implies j =
1
2
r, so all points of the form
r
0
.
1
2
r
0
are critical points and for each of these we have

1
r
0
.
1
2
r
0
= (2)(8) 3(34)
2
= 0. The Second Derivatives Test gives no information, but
1(r. j) = r
2
+ 4j
2
34rj + 2 = (r 32j)
2
+ 2 D 2 with equality if and only if j =
1
2
r. Thus 1
r
0
.
1
2
r
0
= 2 are all local

(and absolute) minima.

201
21. 1(r. j) = r
2
+j
2
+r
32
j
32
From the graphs, there appear to be local minima of about 1(1. 1) = 1(31. 1) E 3 (and no local maxima or saddle
points). 1
i
= 2r 32r
33
j
32
, 1
u
= 2j 32r
32
j
33
, 1
ii
= 2 + 6r
34
j
32
, 1
iu
= 4r
33
j
33
, 1
uu
= 2 + 6r
32
j
34
. Then
1
i
= 0 implies 2r
4
j
2
32 = 0 or r
4
j
2
= 1 or j
2
= r
34
. Note that neither r nor j can be zero. Now 1
u
= 0 implies
2r
2
j
4
32 = 0, and with j
2
= r
34
this implies 2r
36
32 = 0 or r
6
= 1. Thus r = 1 and if r = 1, j = 1; if r = 31,
j = 1. So the critical points are (1. 1), (1. 31),(31. 1) and (31. 31). Now 1(1. 1) = 1(31. 1) = 64 316 0 and
1
ii
0 always, so 1(1. 1) = 1(31. 1) = 3 are local minima.
23. 1(r. j) = sinr + sinj + sin(r +j), 0 $ r $ 2, 0 $ j $ 2
From the graphs it appears that 1 has a local maximum at about (1. 1) with value approximately 2.6, a local minimum
at about (5. 5) with value approximately 32.6, and a saddle point at about (3. 3).
1
i
= cos r + cos(r +j), 1
u
= cos j + cos(r +j), 1
ii
= 3sinr 3sin(r +j), 1
uu
= 3sinj 3sin(r +j),
1
iu
= 3sin(r +j). Setting 1
i
= 0 and 1
u
= 0 and subtracting gives cos r 3cos j = 0 or cos r = cos j. Thus r = j
or r = 2 3j. If r = j, 1
i
= 0 becomes cos r + cos 2r = 0 or 2 cos
2
r + cos r 31 = 0, a quadratic in cos r. Thus
cos r = 31 or
1
2
and r = ,
r
3
, or
5r
3
, yielding the critical points (. ),
r
3
.
r
3
and
5r
3
.
5r
3
. Similarly if
r = 2 3j, 1
i
= 0 becomes (cos r) + 1 = 0 and the resulting critical point is (. ). Now
1(r. j) = sinr sinj +sinr sin(r +j) +sinj sin(r +j). So 1(. ) = 0 and the Second Derivatives Test doesnt apply.
However, along the line j = r we have 1(r. r) = 2 sinr + sin2r = 2 sinr + 2 sinr cos r = 2 sinr(1 + cos r), and
1(r. r) 0 for 0 < r < while 1(r. r) < 0 for < r < 2. Thus every disk with center (. ) contains points where 1 is
positive as well as points where 1 is negative, so the graph crosses its tangent plane (: = 0) there and (. ) is a saddle point.
1
r
3
.
r
3
=
9
4
0 and 1
ii
r
3
.
r
3
< 0 so 1
r
3
.
r
3
=
3
I
3
2
is a local maximum while 1
5r
3
.
5r
3
=
9
4
0 and
1
ii
5r
3
.
5r
3
0, so 1
5r
3
.
5r
3
= 3
3
I
3
2
is a local minimum.
25. 1(r. j) = r
4
35r
2
+j
2
+3r +2 i 1
i
(r. j) = 4r
3
310r +3 and 1
u
(r. j) = 2j. 1
u
= 0 i j = 0, and the graph
of 1
i
shows that the roots of 1
i
= 0 are approximately r = 31.714, 0.312 and 1.402. (Alternatively, we could have used a
calculator or a CAS to nd these roots.) So to three decimal places, the critical points are (31.714. 0), (1.402. 0), and
(0.312. 0). Now since 1
ii
= 12r
2
310, 1
iu
= 0, 1
uu
= 2, and 1 = 24r
2
320, we have 1(31.714. 0) 0,
1
ii
(31.714. 0) 0, 1(1.402. 0) 0, 1
ii
(1.402. 0) 0, and 1(0.312. 0) < 0. Therefore 1(31.714. 0) E 39.200 and
1(1.402. 0) E 0.242 are local minima, and (0.312. 0) is a saddle point. The lowest point on the graph is approximately
(31.714. 0. 39.200).
27. 1(r. j) = 2r + 4r
2
3j
2
+ 2rj
2
3r
4
3j
4
i 1
i
(r. j) = 2 + 8r + 2j
2
34r
3
, 1
u
(r. j) = 32j + 4rj 34j
3
.
Now 1
u
= 0 C 2j(2j
2
32r + 1) = 0 C j = 0 or j
2
= r 3
1
2
. The rst of these implies that 1
i
= 34r
3
+ 8r + 2,
and the second implies that 1
i
= 2 + 8r + 2
r 3
1
2
34r
3
= 34r
3
+ 10r + 1. From the graphs, we see that the rst
possibility for 1
i
has roots at approximately 31.267, 30.259, and 1.526, and the second has a root at approximately 1.629
(the negative roots do not give critical points, since j
2
= r 3
1
2
must be positive). So to three decimal places, 1 has critical
points at (31.267. 0), (30.259. 0), (1.526. 0), and (1.629. 1.063). Now since 1
ii
= 8 312r
2
, 1
iu
= 4j,
1
uu
= 4r 312j
2
, and 1 = (8 312r
2
)(4r 312j
2
) 316j
2
, we have 1(31.267. 0) 0, 1
ii
(31.267. 0) 0,
1(30.259. 0) < 0, 1(1.526. 0) < 0, 1(1.629. 1.063) 0, and 1
ii
(1.629. 1.063) < 0. Therefore, to three decimal
places, 1(31.267. 0) E 1.310 and 1(1.629. 1.063) E 8.105 are local maxima, and (30.259. 0) and (1.526. 0) are saddle
points. The highest points on the graph are approximately (1.629. 1.063. 8.105).

203
29. Since 1 is a polynomial it is continuous on 1, so an absolute maximum and minimum exist. Here 1
i
= 4, 1
u
= 35 so
there are no critical points inside 1. Thus the absolute extrema must both occur on the boundary. Along 1
1
: r = 0 and
1(0. j) = 1 35j for 0 $ j $ 3, a decreasing function in j, so the maximum value is 1 (0. 0) = 1 and the minimum value
is 1(0. 3) = 314. Along 1
2
: j = 0 and 1(r. 0) = 1 + 4r for 0 $ r $ 2, an
increasing function in r, so the minimum value is 1(0. 0) = 1 and the maximum
value is 1(2. 0) = 9. Along 1
3
: j = 3
3
2
r + 3 and 1
r. 3
3
2
r + 3
=
23
2
r 314
for 0 $ r $ 2, an increasing function in r, so the minimum value is
1(0. 3) = 314 and the maximum value is 1(2. 0) = 9. Thus the absolute
maximum of 1 on 1 is 1(2. 0) = 9 and the absolute minimum is 1(0. 3) = 314.
31. 1
i
(r. j) = 2r + 2rj, 1
u
(r. j) = 2j +r
2
, and setting 1
i
= 1
u
= 0
gives (0. 0) as the only critical point in 1, with 1(0. 0) = 4.
On 1
1
: j = 31, 1(r. 31) = 5, a constant.
On 1
2
: r = 1, 1(1. j) = j
2
+j + 5, a quadratic in j which attains its
maximum at (1. 1), 1(1. 1) = 7 and its minimum at
1. 3
1
2
, 1
1. 3
1
2
=
19
4
.
On 1
3
: 1(r. 1) = 2r
2
+ 5 which attains its maximum at (31. 1) and (1. 1)
with 1(1. 1) = 7 and its minimum at (0. 1), 1(0. 1) = 5.
On 1
4
: 1(31. j) = j
2
+j + 5 with maximum at (31. 1), 1(31. 1) = 7 and minimum at
31. 3
1
2
, 1
31. 3
1
2
=
19
4
.
Thus the absolute maximum is attained at both (1. 1) with 1(1. 1) = 7 and the absolute minimum on 1 is attained at
(0. 0) with 1(0. 0) = 4.
33. 1(r. j) = r
4
+j
4
34rj +2 is a polynomial and hence continuous on 1, so
it has an absolute maximum and minimum on 1. In Exercise 7, we found the
critical points of 1; only (1. 1) with 1(1. 1) = 0 is inside 1. On 1
1
: j = 0,
1(r. 0) = r
4
+ 2, 0 $ r $ 3, a polynomial in r which attains its maximum
at r = 3, 1(3. 0) = 83, and its minimum at r = 0, 1(0. 0) = 2.
On 1
2
: r = 3, 1(3. j) = j
4
312j + 83, 0 $ j $ 2, a polynomial in j
which attains its minimum at j =
3
I
3, 1
3.
3
I
3
= 83 39
3
I
3 E 70.0, and its maximum at j = 0, 1(3. 0) = 83.
On 1
3
: j = 2, 1(r. 2) = r
4
38r + 18, 0 $ r $ 3, a polynomial in r which attains its minimum at r =
3
I
2,
1
3
I
2. 2
= 18 36
3
I
2 E 10.4, and its maximum at r = 3. 1(3. 2) = 75. On 1
4
: r = 0, 1(0. j) = j
4
+ 2, 0 $ j $ 2, a
polynomial in j which attains its maximum at j = 2, 1(0. 2) = 18, and its minimum at j = 0, 1(0. 0) = 2. Thus the absolute
maximum of 1 on 1 is 1(3. 0) = 83 and the absolute minimum is 1(1. 1) = 0.
35. 1
i
(r. j) = 6r
2
and 1
u
(r. j) = 4j
3
. And so 1
i
= 0 and 1
u
= 0 only occur when r = j = 0. Hence, the only critical point
inside the disk is at r = j = 0 where 1(0. 0) = 0. Now on the circle r
2
+j
2
= 1, j
2
= 1 3r
2
so let
o(r) = 1(r. j) = 2r
3
+ (1 3r
2
)
2
= r
4
+ 2r
3
32r
2
+ 1,31 $ r $ 1. Then o
0
(r) = 4r
3
+ 6r
2
34r = 0 i r = 0,
32, or
1
2
. 1(0. 1) = o (0) = 1, 1
1
2
.
I
3
2
= o
1
2
=
13
16
, and (32. 33) is not in 1. Checking the endpoints, we get
1(31. 0) = o(31) = 32 and 1(1. 0) = o(1) = 2. Thus the absolute maximum and minimum of 1 on 1 are 1(1. 0) = 2 and
1(31. 0) = 32.
Another method: On the boundary r
2
+j
2
= 1 we can write r = cos 0, j = sin0, so 1(cos 0. sin0) = 2 cos
3
0 + sin
4
0,
0 $ 0 $ 2.
37. 1(r. j) = 3(r
2
31)
2
3(r
2
j 3r 31)
2
i 1
i
(r. j) = 32(r
2
31)(2r) 32(r
2
j 3r 31)(2rj 31) and
1
u
(r. j) = 32(r
2
j 3r 31)r
2
. Setting 1
u
(r. j) = 0 gives either r = 0 or r
2
j 3r 31 = 0.
There are no critical points for r = 0, since 1
i
(0. j) = 32, so we set r
2
j 3r 31 = 0 C j =
r + 1
r
2
[r 6= 0],
so 1
i
r.
r + 1
r
2
= 32(r
2
31)(2r) 32
r
2
r + 1
r
2
3r 31
2r
r + 1
r
2
31
= 34r(r
2
31). Therefore
1
i
(r. j) = 1
u
(r. j) = 0 at the points (1. 2) and (31. 0). To classify these critical points, we calculate
1
ii
(r. j) = 312r
2
312r
2
j
2
+ 12rj + 4j + 2, 1
uu
(r. j) = 32r
4
,
and 1
iu
(r. j) = 38r
3
j + 6r
2
+ 4r. In order to use the Second
Derivatives Test we calculate
1(31. 0) = 1
ii
(31. 0) 1
uu
(31. 0) 3[1
iu
(31. 0)]
2
= 16 0,
1
ii
(31. 0) = 310 < 0, 1(1. 2) = 16 0, and 1
ii
(1. 2) = 326 < 0, so
both (31. 0) and (1. 2) give local maxima.
39. Let d be the distance from (2. 1. 31) to any point (r. j. :) on the plane r + j 3: = 1, so
d =

(r 32)
2
+ (j 31)
2
+ (: + 1)
2
where : = r + j 3 1, and we minimize
d
2
= 1(r. j) = (r 32)
2
+ (j 31)
2
+ (r +j)
2
. Then 1
i
(r. j) = 2(r 32) + 2(r +j) = 4r + 2j 34,
1
u
(r. j) = 2(j 31) +2(r +j) = 2r +4j 32. Solving 4r +2j 34 = 0 and 2r +4j 32 = 0 simultaneously gives r = 1,
j = 0. An absolute minimum exists (since there is a minimum distance from the point to the plane) and it must occur at a
critical point, so the shortest distance occurs for r = 1, j = 0 for which d =

(1 32)
2
+ (0 31)
2
+ (0 + 1)
2
=
I
3.
41. Let d be the distance from the point (4. 2. 0) to any point (r. j. :) on the cone, so d =

(r 34)
2
+ (j 32)
2
+:
2
where
:
2
= r
2
+j
2
, and we minimize d
2
= (r 34)
2
+ (j 32)
2
+r
2
+ j
2
= 1 (r. j). Then
1
i
(r. j) = 2 (r 34) + 2r = 4r 38, 1
u
(r. j) = 2 (j 32) + 2j = 4j 34, and the critical points occur when 1
i
= 0 i
r = 2, 1
u
= 0 i j = 1. Thus the only critical point is (2. 1). An absolute minimum exists (since there is a minimum
distance from the cone to the point) which must occur at a critical point, so the points on the cone closest
to (4. 2. 0) are
2. 1.
I
5
.

205
43. r +j +: = 100, so maximize 1(r. j) = rj(100 3r 3j). 1
i
= 100j 32rj 3j
2
, 1
u
= 100r 3r
2
32rj,
1
ii
= 32j, 1
uu
= 32r, 1
iu
= 100 32r 32j. Then 1
i
= 0 implies j = 0 or j = 100 32r. Substituting j = 0 into
1
u
= 0 gives r = 0 or r = 100 and substituting j = 100 32r into 1
u
= 0 gives 3r
2
3100r = 0 so r = 0 or
100
3
.
Thus the critical points are (0. 0), (100. 0), (0. 100) and
100
3
.
100
3
.
1(0. 0) = 1(100. 0) = 1(0. 100) = 310,000 while 1
100
3
.
100
3
=
10,000
3
and 1
ii
100
3
.
100
3
= 3
200
3
< 0. Thus (0. 0),
(100. 0) and (0. 100) are saddle points whereas 1
100
3
,
100
3
is a local maximum. Thus the numbers are r = j = : =

100
3
.
45. Center the sphere at the origin so that its equation is r
2
+j
2
+:
2
= v
2
, and orient the inscribed rectangular box so that its
edges are parallel to the coordinate axes. Any vertex of the box satises r
2
+j
2
+:
2
= v
2
, so take (r. j. :) to be the vertex
in the rst octant. Then the box has length 2r, width 2j, and height 2: = 2
v
2
3r
2
3j
2
with volume given by
\ (r. j) = (2r)(2j)
v
2
3r
2
3j
2
= 8rj
v
2
3r
2
3j
2
for 0 < r < v, 0 < j < v. Then
\
i
= (8rj)
1
2
(v
2
3r
2
3j
2
)
31'2
(32r) +
v
2
3r
2
3j
2
8j =
8j(v
2
32r
2
3j
2
)
v
2
3r
2
3j
2
and \
u
=
8r(v
2
3r
2
32j
2
)
v
2
3r
2
3j
2
.
Setting \
i
= 0 gives j = 0 or 2r
2
+j
2
= v
2
, but j 0 so only the latter solution applies. Similarly, \
u
= 0 with r 0
implies r
2
+ 2j
2
= v
2
. Substituting, we have 2r
2
+j
2
= r
2
+ 2j
2
i r
2
= j
2
i j = r. Then r
2
+ 2j
2
= v
2
i
3r
2
= v
2
i r =

v
2
3 = v
I
3 = j. Thus the only critical point is
v
I
3. v
I
3
. There must be a maximum

volume and here it must occur at a critical point, so the maximum volume occurs when r = j = v
I
3 and the maximum
volume is \
r
I
3
.
r
I
3
= 8
r
I
3
r
I
3
v
2
3
r
I
3
2
3
r
I
3
2
=
8
3
I
3
v
3
.
47. Maximize 1(r. j) =
rj
3
(6 3r 32j), then the maximum volume is \ = rj:.
1
i
=
1
3
(6j 32rj 3j
2
) =
1
3
j(6 32r 32j) and 1
u
=
1
3
r(6 3r 34j). Setting 1
i
= 0 and 1
u
= 0 gives the critical point
(2. 1) which geometrically must yield a maximum. Thus the volume of the largest such box is \ = (2)(1)
2
3
=
4
3
.
49. Let the dimensions be r, j, and :; then 4r + 4j + 4: = c and the volume is
\ = rj: = rj
1
4
c 3r 3j
=
1
4
crj 3r
2
j 3rj
2
, r 0, j 0. Then \
i
=
1
4
cj 32rj 3j
2
and \
u
=
1
4
cr 3r
2
32rj,
so \
i
= 0 = \
u
when 2r +j =
1
4
c and r + 2j =
1
4
c. Solving, we get r =
1
12
c, j =
1
12
c and : =
1
4
c 3r 3j =
1
12
c. From
the geometrical nature of the problem, this critical point must give an absolute maximum. Thus the box is a cube with edge
length
1
12
c.
51. Let the dimensions be r, j and :, then minimize rj + 2(r: +j:) if rj: = 32,000 cm
3
. Then
1(r. j) = rj + [64,000(r +j)rj] = rj + 64,000(r
31
+j
31
), 1
i
= j 364,000r
32
, 1
u
= r 364,000j
32
.
And 1
i
= 0 implies j = 64,000r
2
; substituting into 1
u
= 0 implies r
3
= 64,000 or r = 40 and then j = 40. Now
1(r. j) = [(2)(64,000)]
2
r
33
j
33
31 0 for (40. 40) and 1
ii
(40. 40) 0 so this is indeed a minimum. Thus the
dimensions of the box are r = j = 40 cm, : = 20 cm.
53. Let r. j, : be the dimensions of the rectangular box. Then the volume of the box is rj: and
1 =

r
2
+j
2
+:
2
i 1
2
= r
2
+j
2
+:
2
i : =

1
2
3r
2
3j
2
.
Substituting, we have volume \ (r. j) = rj
1
2
3r
2
3j
2
, (r. j 0).
\
i
= rj
1
2
(1
2
3r
2
3j
2
)
31'2
(32r) +j
1
2
3r
2
3j
2
= j
1
2
3r
2
3j
2
3
r
2
j
1
2
3r
2
3j
2
,
\
u
= r
1
2
3r
2
3j
2
3
rj
2
1
2
3r
2
3j
2
. \
i
= 0 implies j(1
2
3r
2
3j
2
) = r
2
j i j(1
2
32r
2
3j
2
) = 0 i
2r
2
+j
2
= 1
2
(since j 0), and \
u
= 0 implies r(1
2
3r
2
3j
2
) = rj
2
i r(1
2
3r
2
32j
2
) = 0 i
r
2
+ 2j
2
= 1
2
(since r 0). Substituting j
2
= 1
2
32r
2
into r
2
+ 2j
2
= 1
2
gives r
2
+ 21
2
34r
2
= 1
2
i
3r
2
= 1
2
i r = 1
I
3 (since r 0) and then j =

1
2
32
1
I
3
2
= 1
I
3. So the only critical point is
1
I
3. 1
I
3
which, from the geometrical nature of the problem, must give an absolute maximum. Thus the maximum
volume is \
1
I
3. 1
I
3
=

1
I
3
1
2
3
1
I
3
2
3
1
I
3
2
= 1
3
3
I
3
cubic units.
55. Note that here the variables are i and /, and 1(i. /) =
n
. =1
[j
.
3(ir
.
+/)]
2
. Then 1
r
=
n
. =1
32r
.
[j
.
3(ir
.
+/)] = 0
implies
n
. =1
r
.
j
.
3ir
2
.
3/r
.
= 0 or
n
. =1
r
.
j
.
= i
n
. =1
r
2
.
+/
n
. =1
r
.
and 1
l
=
n
. =1
32[j
.
3(ir
.
+/)] = 0 implies
n
. =1
j
.
= i
n
. =1
r
.
+
n
. =1
/ = i
. =1
r
.
+n/. Thus we have the two desired equations.

Now 1
rr
=
n
. =1
2r
2
.
, 1
ll
=
n
. =1
2 = 2n and 1
rl
=
n
. =1
2r
.
. And 1
rr
(i. /) 0 always and
1(i. /) = 4n
. =1
r
2
.
34
. =1
r
.
2
= 4
. =1
r
2
.
. =1
r
.
0 always so the solutions of these two

equations do indeed minimize
n
. =1
d
2
.
.
15.8 Lagrange Multipliers ET 14.8
1. At the extreme values of 1, the level curves of 1 just touch the curve o(r. j) = 8 with a common tangent line. (See Figure 1
and the accompanying discussion.) We can observe several such occurrences on the contour map, but the level curve
1(r. j) = c with the largest value of c which still intersects the curve o(r. j) = 8 is approximately c = 59, and the smallest
value of c corresponding to a level curve which intersects o(r. j) = 8 appears to be c = 30. Thus we estimate the maximum
value of 1 subject to the constraint o(r. j) = 8 to be about 59 and the minimum to be 30.
3. 1(r. j) = r
2
+j
2
, o(r. j) = rj = 1, and Q1 = `Qo i h2r. 2ji = h`j. `ri, so 2r = `j, 2j = `r, and rj = 1.
From the last equation, r 6= 0 and j 6= 0, so 2r = `j i ` = 2rj. Substituting, we have 2j = (2rj) r i
j
2
= r
2
i j = r. But rj = 1, so r = j = 1 and the possible points for the extreme values of 1 are (1. 1) and
(31. 31). Here there is no maximum value, since the constraint rj = 1 allows r or j to become arbitrarily large, and hence
1(r. j) = r
2
+j
2
can be made arbitrarily large. The minimum value is 1(1. 1) = 1(31. 31) = 2.
SECTION 15.8 LAGRANGE MULTIPLIERS ET SECTION 14.8

207
5. 1(r. j) = r
2
j, o(r. j) = r
2
+ 2j
2
= 6 i Q1 =

2rj. r
2
, `Qo = h2`r. 4`ji. Then 2rj = 2`r implies r = 0 or

` = j. If r = 0, then r
2
= 4`j implies ` = 0 or j = 0. However, if j = 0 then o(r. j) = 0. a contradiction. So ` = 0 and
then o(r. j) = 6 i j =
I
3. If ` = j, then r
2
= 4`j implies r
2
= 4j
2
, and so o(r. j) = 6 i
4j
2
+ 2j
2
= 6 i j
2
= 1 i j = 1. Thus 1 has possible extreme values at the points
0.
I
3
, (2. 1), and

(2. 31). After evaluating 1 at these points, we nd the maximum value to be 1(2. 1) = 4 and the minimum to be
1(2. 31) = 34.
7. 1(r. j. :) = 2r + 6j + 10:, o(r. j. :) = r
2
+j
2
+:
2
= 35 i Q1 = h2. 6. 10i, `Qo = h2`r. 2`j. 2`:i. Then
2`r = 2, 2`j = 6, 2`: = 10 imply r =
1
`
, j =
3
`
, and : =
5
`
. But 35 = r
2
+j
2
+:
2
=

1
`
2
+
3
`
2
+
5
`
2
i
35 =
35
`
2
i ` = 1, so 1 has possible extreme values at the points (1. 3. 5), (31. 33. 35). The maximum value of 1 on
r
2
+j
2
+:
2
= 35 is 1(1. 3. 5) = 70, and the minimum is 1(31. 33. 35) = 370.
9. 1(r. j. :) = rj:, o(r. j. :) = r
2
+ 2j
2
+ 3:
2
= 6 i Q1 = hj:. r:. rji, `Qo = h2`r. 4`j. 6`:i. If ` = 0 then at
least one of the coordinates is 0, in which case 1(r. j. :) = 0. (None of these ends up giving a maximum or minimum.)
If ` 6= 0, then Q1 = `Qo implies ` = (j:)(2r) = (r:)(4j) = (rj)(6:) or r
2
= 2j
2
and :
2
=
2
3
j
2
. Thus
r
2
+ 2j
2
+ 3:
2
= 6 implies 6j
2
= 6 or j = 1. Thus the possible remaining points are
I
2. 1.
2
3
I
2. 1. 3
2
3
3
I
2. 1.
2
3
3
I
2. 1. 3
2
3
. The maximum value of 1 on the ellipsoid is

2
I
3
, occurring when
all coordinates are positive or exactly two are negative and the minimum is 3
2
I
3
occurring when 1 or 3 of the coordinates are
negative.
11. 1(r. j. :) = r
2
+j
2
+:
2
, o(r. j. :) = r
4
+j
4
+:
4
= 1 i Q1 = h2r. 2j. 2:i, `Qo =

4`r
3
. 4`j
3
. 4`:
3
.
Case 1: If r 6= 0, j 6= 0 and : 6= 0, then Q1 = `Qo implies ` = 1(2r
2
) = 1(2j
2
) = 1(2:
2
) or r
2
= j
2
= :
2
and
3r
4
= 1 or r =
1
4
I
3
giving the points
1
4
I
3
.
1
4
I
3
.
1
4
I
3
1
4
I
3
. 3
1
4
I
3
.
1
4
I
3
1
4
I
3
.
1
4
I
3
. 3
1
4
I
3
1
4
I
3
. 3
1
4
I
3
. 3
1
4
I
3
all with an 1-value of

I
3.
Case 2: If one of the variables equals zero and the other two are not zero, then the squares of the two nonzero coordinates
are equal with common value
1
I
2
and corresponding 1 value of
I
2.
Case 3: If exactly two of the variables are zero, then the third variable has value 1 with the corresponding 1 value of 1.
Thus on r
4
+j
4
+:
4
= 1, the maximum value of 1 is
I
3 and the minimum value is 1.
13. 1(r. j. :. t) = r +j +: +t, o(r. j. :. t) = r
2
+j
2
+:
2
+t
2
= 1 i h1. 1. 1. 1i = h2`r. 2`j. 2`:. 2`ti, so
` = 1(2r) = 1(2j) = 1(2:) = 1(2t) and r = j = : = t. But r
2
+j
2
+:
2
+t
2
= 1, so the possible points are
1
2
.
1
2
.
1
2
.
1
2
. Thus the maximum value of 1 is 1
1
2
.
1
2
.
1
2
.
1
2
= 2 and the minimum value is

1
3
1
2
. 3
1
2
. 3
1
2
. 3
1
2
= 32.
15. 1(r. j. :) = r + 2j, o(r. j. :) = r +j +: = 1, /(r. j. :) = j
2
+:
2
= 4 i Q1 = h1. 2. 0i, `Qo = h`. `. ì
and jQ/ = h0. 2jj. 2j:i. Then 1 = `, 2 = ` + 2jj and 0 = ` + 2j: so jj =
1
2
= 3j: or j = 1(2j), : = 31(2j).
Thus r +j +: = 1 implies r = 1 and j
2
+:
2
= 4 implies j =
1
2
I
2
. Then the possible points are
1.
I
2. ~
I
2
and the maximum value is 1
1.
I
2. 3
I
2
= 1 + 2
I
2 and the minimum value is 1

1. 3
I
2.
I
2
= 1 32
I
2.
17. 1(r. j. :) = j: +rj, o(r. j. :) = rj = 1, /(r. j. :) = j
2
+:
2
= 1 i Q1 = hj. r +:. ji, `Qo = h`j. `r. 0i,
jQ/ = h0. 2jj. 2j:i. Then j = `j implies ` = 1 [j 6= 0 since o(r. j. :) = 1], r +: = `r + 2jj and j = 2j:. Thus
j = :(2j) = j(2j) or j
2
= :
2
, and so j
2
+:
2
= 1 implies j =
1
I
2
, : =
1
I
2
. Then rj = 1 implies r =
I
2 and
the possible points are
I
2.
1
I
2
.
1
I
2
I
2.
1
I
2
. 3
1
I
2
. Hence the maximum of 1 subject to the constraints is

1
I
2.
1
I
2
.
1
I
2
=
3
2
and the minimum is 1
I
2.
1
I
2
. ~
1
I
2
=
1
2
.
Note: Since rj = 1 is one of the constraints we could have solved the problem by solving 1(j. :) = j: + 1 subject to
j
2
+:
2
= 1.
19. 1(r. j) = c
3iu
. For the interior of the region, we nd the critical points: 1
i
= 3jc
3iu
, 1
u
= 3rc
3iu
, so the only
critical point is (0. 0), and 1(0. 0) = 1. For the boundary, we use Lagrange multipliers. o(r. j) = r
2
+ 4j
2
= 1 i
`Qo = h2`r. 8`ji, so setting Q1 = `Qo we get 3jc
3iu
= 2`r and 3rc
3iu
= 8`j. The rst of these gives
c
3iu
= 32`rj, and then the second gives 3r(32`rj) = 8`j i r
2
= 4j
2
. Solving this last equation with the
constraint r
2
+ 4j
2
= 1 gives r =
1
I
2
and j =
1
2
I
2
. Now 1
1
I
2
. ~
1
2
I
2
= c
1'4
E 1.284 and
1
1
I
2
.
1
2
I
2
= c
31'4
E 0.779. The former are the maxima on the region and the latter are the minima.
21. (a) 1(r. j) = r, o(r. j) = j
2
+r
4
3r
3
= 0 i Q1 = h1. 0i = `Qo = `
4r
3
33r
2
. 2j
. Then
1 = `(4r
3
33r
2
) (1) and 0 = 2`j (2). We have ` 6= 0 from (1), so (2) gives j = 0. Then, from the constraint equation,
r
4
3r
3
= 0 i r
3
(r 31) = 0 i r = 0 or r = 1. But r = 0 contradicts (1), so the only possible extreme value
subject to the constraint is 1(1. 0) = 1. (The question remains whether this is indeed the minimum of 1.)
(b) The constraint is j
2
+r
4
3r
3
= 0 C j
2
= r
3
3r
4
. The left side is non-negative, so we must have r
3
3r
4
D 0
which is true only for 0 $ r $ 1. Therefore the minimum possible value for 1(r. j) = r is 0 which occurs for r = j = 0.
However, `Qo(0. 0) = `h0 30. 0i = h0. 0i and Q1(0. 0) = h1. 0i, so Q1(0. 0) 6= `Qo(0. 0) for all values of `.
(c) Here Qo(0. 0) = 0 but the method of Lagrange multipliers requires that Qo 6= 0 everywhere on the constraint curve.
23. 1(1. 1) = /1
o
1
13o
, o(1. 1) = i1 +n1 = j i Q1 =

c/1
o31
1
13o
. (1 3c)/1
o
1
3o
, `Qo = hì. `ni.

Then c/(11)
13o
= ì and (1 3c)/(11)
o
= `n and i1 +n1 = j, so c/(11)
13o
i = (1 3c)/(11)
o
n or
nc[i(1 3c)] = (11)
o
(11)
13o
or 1 = 1nc[i(1 3c)]. Substituting into i1 +n1 = j gives 1 = (1 3c)jn
and 1 = cji for the maximum production.
SECTION 15.8 LAGRANGE MULTIPLIERS ET SECTION 14.8

209
25. Let the sides of the rectangle be r and j. Then 1(r. j) = rj, o(r. j) = 2r + 2j = j i Q1(r. j) = hj. ri,
`Qo = h2`. 2ì. Then ` =
1
2
j =
1
2
r implies r = j and the rectangle with maximum area is a square with side length
1
4
j.
27. Let 1(r. j. :) = d
2
= (r 32)
2
+ (j 31)
2
+ (: + 1)
2
, then we want to minimize 1 subject to the constraint
o(r. j. :) = r +j 3: = 1. Q1 = `Qo i h2(r 32). 2(j 31). 2(: + 1)i = `h1. 1. 31i, so r = (` + 4)2,
j = (` + 2)2, : = 3(` + 2)2. Substituting into the constraint equation gives
` + 4
2
+
` + 2
2
+
` + 2
2
= 1 i
3` + 8 = 2 i ` = 32, so r = 1, j = 0, and : = 0. This must correspond to a minimum, so the shortest distance is
d =

(1 32)
2
+ (0 31)
2
+ (0 + 1)
2
=
I
3.
29. Let 1(r. j. :) = d
2
= (r 34)
2
+ (j 32)
2
+:
2
. Then we want to minimize 1 subject to the constraint
o (r. j. :) = r
2
+j
2
3:
2
= 0. Q1 = `Qo i h2 (r 34) . 2 (j 32) . 2:i = h2`r. 2`j. 32`:i, so r 34 = `r,
j 32 = `j, and : = 3`:. From the last equation we have : +`: = 0 i : (1 +`) = 0, so either : = 0 or ` = 31.
But from the constraint equation we have : = 0 i r
2
+j
2
= 0 i r = j = 0 which is not possible from the rst two
equations. So ` = 31 and r 34 = `r i r = 2, j 32 = `j i j = 1, and r
2
+j
2
3:
2
= 0 i
4 + 1 3:
2
= 0 i : =
I
5. This must correspond to a minimum, so the points on the cone closest to (4. 2. 0)
are
2. 1.
I
5
.
31. 1(r. j. :) = rj:, o(r. j. :) = r +j +: = 100 i Q1 = hj:. r:. rji = `Qo = h`. `. ì. Then ` = j: = r: = rj
implies r = j = : =
100
3
.
33. If the dimensions are 2r, 2j, and 2:, then maximize 1(r. j. :) = (2r)(2j)(2:) = 8rj: subject to
o(r. j. :) = r
2
+j
2
+:
2
= v
2
(r 0, j 0, : 0). Then Q1 = `Qo i h8j:. 8r:. 8rji = `h2r. 2j. 2:i i
8j: = 2`r, 8r: = 2`j, and 8rj = 2`:, so ` =
4j:
r
=
4r:
j
=
4rj
:
. This gives r
2
: = j
2
: i r
2
= j
2
(since : 6= 0)
and rj
2
= r:
2
i :
2
= j
2
, so r
2
= j
2
= :
2
i r = j = :, and substituting into the constraint
equation gives 3r
2
= v
2
i r = v
I
3 = j = :. Thus the largest volume of such a box is
1
r
I
3
.
r
I
3
.
r
I
3
= 8
r
I
3
r
I
3
r
I
3
=
8
3
I
3
v
3
.
35. 1(r. j. :) = rj:, o(r. j. :) = r + 2j + 3: = 6 i Q1 = hj:. r:. rji = `Qo = h`. 2`. 3ì.
Then ` = j: =
1
2
r: =
1
3
rj implies r = 2j, : =
2
3
j. But 2j + 2j + 2j = 6 so j = 1, r = 2, : =
2
3
and the volume
is \ =
4
3
.
37. 1(r. j. :) = rj:, o(r. j. :) = 4(r +j +:) = c i Q1 = hj:. r:. rji, `Qo = h4`. 4`. 4ì. Thus
4` = j: = r: = rj or r = j = : =
1
12
c are the dimensions giving the maximum volume.
39. If the dimensions of the box are given by r, j, and :, then we need to nd the maximum value of 1(r. j. :) = rj:
[r. j. : 0] subject to the constraint 1 =

r
2
+j
2
+:
2
or o(r. j. :) = r
2
+j
2
+:
2
= 1
2
. Q1 = `Qo i
hj:. r:. rji = `h2r. 2j. 2:i, so j: = 2`r i ` =
j:
2r
, r: = 2`j i ` =
r:
2j
, and rj = 2`: i ` =
rj
2:
. Thus
` =
j:
2r
=
r:
2j
i r
2
= j
2
[since : 6= 0] i r = j and ` =
j:
2r
=
rj
2:
i r = : [since j 6= 0].
Substituting into the constraint equation gives r
2
+r
2
+r
2
= 1
2
i r
2
= 1
2
3 i r = 1
I
3 = j = : and the
maximum volume is
1
I
3
3
= 1
3
3
I
3
.
41. We need to nd the extreme values of 1(r. j. :) = r
2
+j
2
+:
2
subject to the two constraints o(r. j. :) = r +j + 2: = 2
and /(r. j. :) = r
2
+j
2
3: = 0. Q1 = h2r. 2j. 2:i, `Qo = h`. `. 2ì and jQ/ = h2jr. 2jj. 3ji. Thus we need
2r = ` + 2jr (1), 2j = ` + 2jj (2), 2: = 2` 3j (3), r +j + 2: = 2 (4), and r
2
+j
2
3: = 0 (5).
From (1) and (2), 2(r 3j) = 2j(r 3j), so if r 6= j, j = 1. Putting this in (3) gives 2: = 2` 31 or ` = : +
1
2
, but putting
j = 1 into (1) says ` = 0. Hence : +
1
2
= 0 or : = 3
1
2
. Then (4) and (5) become r +j 33 = 0 and r
2
+j
2
+
1
2
= 0. The
last equation cannot be true, so this case gives no solution. So we must have r = j. Then (4) and (5) become 2r +2: = 2 and
2r
2
3: = 0 which imply : = 1 3r and : = 2r
2
. Thus 2r
2
= 1 3r or 2r
2
+r 31 = (2r 31)(r + 1) = 0 so r =
1
2
or
r = 31. The two points to check are
1
2
.
1
2
.
1
2
and (31. 31. 2): 1
1
2
.
1
2
.
1
2
=
3
4
and 1(31. 31. 2) = 6. Thus
1
2
.
1
2
.
1
2
is
the point on the ellipse nearest the origin and (31. 31. 2) is the one farthest from the origin.
43. 1(r. j. :) = jc
i3:
, o(r. j. :) = 9r
2
+ 4j
2
+ 36:
2
= 36, /(r. j. :) = rj +j: = 1. Q1 = `Qo +jQ/ i
jc
i3:
. c
i3:
. 3jc
i3:
= `h18r. 8j. 72:i +jhj. r +:. ji, so jc

i3:
= 18`r +jj, c
i3:
= 8`j +j(r +:),
3jc
i3:
= 72`: +jj, 9r
2
+ 4j
2
+ 36:
2
= 36, rj +j: = 1. Using a CAS to solve these 5 equations simultaneously for r,
j, :, `, and j (in Maple, use the allvalues command), we get 4 real-valued solutions:
r E 0.222444, j E 32.157012, : E 30.686049, ` E 30.200401, j E 2.108584
r E 31.951921, j E 30.545867, : E 0.119973, ` E 0.003141, j E 30.076238
r E 0.155142, j E 0.904622, : E 0.950293, ` E 30.012447, j E 0.489938
r E 1.138731, j E 1.768057, : E 30.573138, ` E 0.317141, j E 1.862675
Substituting these values into 1 gives 1(0.222444. 32.157012. 30.686049) E 35.3506,
1(31.951921. 30.545867. 0.119973) E 30.0688, 1(0.155142. 0.904622. 0.950293) E 0.4084,
1(1.138731. 1.768057. 30.573138) E 9.7938. Thus the maximum is approximately 9.7938, and the mininum is
approximately 35.3506.
45. (a) We wish to maximize 1(r
1
. r
2
, . . . , r
n
) =
q
I
r
1
r
2
r
n
subject to o(r
1
. r
2
, . . . , r
n
) = r
1
+r
2
+ +r
n
= c
and r
.
0.
Q1 =

1
n
(r
1
r
2
r
n
)
1
q
31
(r
2
r
n
) ,
1
n
(r
1
r
2
r
n
)
1
q
31
(r
1
r
3
r
n
) , . . . ,
1
n
(r
1
r
2
r
n
)
1
q
31
(r
1
r
n31
)
and `Qo = h`. `, . . . , ì, so we need to solve the system of equations

1
n
(r
1
r
2
r
n
)
1
q
31
(r
2
r
n
) = ` i r
1'n
1
r
1'n
2
r
1'n
n
= n`r
1
1
n
(r
1
r
2
r
n
)
1
q
31
(r
1
r
3
r
n
) = ` i r
1'n
1
r
1'n
2
r
1'n
n
= n`r
2
.
.
.
1
n
(r
1
r
2
r
n
)
1
q
31
(r
1
r
n31
) = ` i r
1'n
1
r
1'n
2
r
1'n
n
= n`r
n
This implies n`r
1
= n`r
2
= = n`r
n
. Note ` 6= 0, otherwise we cant have all r
.
0. Thus r
1
= r
2
= = r
n
.
But r
1
+r
2
+ +r
n
= c i nr
1
= c i r
1
=
c
n
= r
2
= r
3
= = r
n
. Then the only point where 1 can
have an extreme value is
c
n
.
c
n
, . . . ,
c
n
. Since we can choose values for (r

1
. r
2
. . . . . r
n
) that make 1 as close to
zero (but not equal) as we like, 1 has no minimum value. Thus the maximum value is
1
c
n
.
c
n
, . . . ,
c
n
=
q
c
n

c
n

c
n
=
c
n
.
(b) From part (a),
c
n
is the maximum value of 1. Thus 1(r
1
. r
2
, . . . , r
n
) =
q
I
r
1
r
2
r
n
$
c
n
. But
r
1
+r
2
+ +r
n
= c, so
q
I
r
1
r
2
r
n
$
r
1
+r
2
+ +r
n
n
. These two means are equal when 1 attains its
maximum value
c
n
, but this can occur only at the point
c
n
.
c
n
, . . . ,
c
n
we found in part (a). So the means are equal only

when r
1
= r
2
= r
3
= = r
n
=
c
n
.
15 Review ET 14
1. (a) A function 1 of two variables is a rule that assigns to each ordered pair (r. j) of real numbers in its domain a unique real
number denoted by 1(r. j).
(b) One way to visualize a function of two variables is by graphing it, resulting in the surface : = 1(r. j). Another method for
visualizing a function of two variables is a contour map. The contour map consists of level curves of the function which are
horizontal traces of the graph of the function projected onto the rj-plane. Also, we can use an arrow diagram such as
Figure 1 in Section 15.1 [ET 14.1].
2. A function 1 of three variables is a rule that assigns to each ordered triple (r. j. :) in its domain a unique real number
1(r. j. :). We can visualize a function of three variables by examining its level surfaces 1(r. j. :) = I, where I is a constant.
3. lim
(i.u)<(a.l)
1(r. j) = 1 means the values of 1(r. j) approach the number 1 as the point (r. j) approaches the point (o. /)
along any path that is within the domain of 1. We can show that a limit at a point does not exist by nding two different paths
approaching the point along which 1(r. j) has different limits.
4. (a) See Denition 15.2.4 [ET 14.2.4].
(b) If 1 is continuous on R
2
, its graph will appear as a surface without holes or breaks.
5. (a) See (2) and (3) in Section 15.3 [ET 14.3].
(b) See Interpretations of Partial Derivatives on page 917 [ET 881].
(c) To nd 1
i
, regard j as a constant and differentiate 1(r. j) with respect to r. To nd 1
u
, regard r as a constant and
differentiate 1(r. j) with respect to j.
6. See the statement of Clairauts Theorem on page 921 [ET 885].
7. (a) See (2) in Section 15.4 [ET 14.4]
(b) See (19) and the preceding discussion in Section 15.6 [ET 14.6].
8. See (3) and (4) and the accompanying discussion in Section 15.4 [ET 14.4]. We can interpret the linearization of 1 at (o. /)
geometrically as the linear function whose graph is the tangent plane to the graph of 1 at (o. /). Thus it is the linear function
which best approximates 1 near (o. /).
9. (a) See Denition 15.4.7 [ET 14.4.7].
(b) Use Theorem 15.4.8 [ET 14.4.8].
10. See (10) and the associated discussion in Section 15.4 [ET 14.4].
11. See (2) and (3) in Section 15.5 [ET 14.5].
12. See (7) and the preceding discussion in Section 15.5 [ET 14.5].
13. (a) See Denition 15.6.2 [ET 14.6.2]. We can interpret it as the rate of change of 1 at (r
0
. j
0
) in the direction of u.
Geometrically, if 1 is the point (r
0
. j
0
. 1(r
0
. j
0
)) on the graph of 1 and C is the curve of intersection of the graph of 1
with the vertical plane that passes through 1 in the direction u, the directional derivative of 1 at (r
0
. j
0
) in the direction of
u is the slope of the tangent line to C at 1. (See Figure 5 in Section 15.6 [ET 14.6].)
(b) See Theorem 15.6.3 [ET 14.6.3].
14. (a) See (8) and (13) in Section 15.6 [ET 14.6].
(b) 1
u
1(r. j) = Q1(r. j) u or 1
u
1(r. j. :) = Q1(r. j. :) u
(c) The gradient vector of a function points in the direction of maximum rate of increase of the function. On a graph of the
function, the gradient points in the direction of steepest ascent.
15. (a) 1 has a local maximum at (o. /) if 1(r. j) $ 1(o. /) when (r. j) is near (o. /).
(b) 1 has an absolute maximum at (o. /) if 1(r. j) $ 1(o. /) for all points (r. j) in the domain of 1.
(c) 1 has a local minimum at (o. /) if 1(r. j) D 1(o. /) when (r. j) is near (o. /).
(d) 1 has an absolute minimum at (o. /) if 1(r. j) D 1(o. /) for all points (r. j) in the domain of 1.
(e) 1 has a saddle point at (o. /) if 1(o. /) is a local maximum in one direction but a local minimum in another.
16. (a) By Theorem 15.7.2 [ET 14.7.2], if 1 has a local maximum at (o. /) and the rst-order partial derivatives of 1 exist there,
then 1
i
(o. /) = 0 and 1
u
(o. /) = 0.
(b) A critical point of 1 is a point (o. /) such that 1
i
(o. /) = 0 and 1
u
(o. /) = 0 or one of these partial derivatives does not
exist.
17. See (3) in Section 15.7 [ET 14.7]
18. (a) See Figure 11 and the accompanying discussion in Section 15.7 [ET 14.7].
(b) See Theorem 15.7.8 [ ET 14.7.8].
(c) See the procedure outlined in (9) in Section 15.7 [ET 14.7].
19. See the discussion beginning on page 970 [ET 934]; see Two Constraints on page 974 [ET 938].
1. True. 1
u
(o. /) = lim
I<0
1(o. / +/) 31(o. /)
/
from Equation 15.3.3 [ET 14.3.3]. Let / = j 3/. As / <0, j </. Then by
substituting, we get 1
u
(o. /) = lim
u<l
1(o. j) 31(o. /)
j 3/
.
3. False. 1
iu
=
u
2
1
uj ur
.
5. False. See Example 15.2.3 [ET 14.2.3].
7. True. If 1 has a local minimum and 1 is differentiable at (o. /) then by Theorem 15.7.2 [ET 14.7.2], 1
i
(o. /) = 0 and
1
u
(o. /) = 0, so Q1(o. /) = h1
i
(o. /). 1
u
(o. /)i = h0. 0i = 0.
9. False. Q1(r. j) = h0. 1ji.
11. True. Q1 = hcos r. cos ji, so |Q1| =

cos
2
r + cos
2
j. But |cos 0| $ 1, so |Q1| $
I
2. Now
1
u
1 (r. j) = Q1 u = |Q1| |u| cos 0, but u is a unit vector, so |1
u
1(r. j)| $
I
2 1 1 =
I
2.
1. ln(r +j + 1) is dened only when r +j + 1 0 i j 3r 31,
so the domain of 1 is {(r. j) | j 3r 31}, all those points above the
line j = 3r 31.
3. : = 1(r. j) = 1 3j
2
, a parabolic cylinder 5. The level curves are
4r
2
+j
2
= I or 4r
2
+j
2
= I
2
,
I D 0, a family of ellipses.
7.
9. 1 is a rational function, so it is continuous on its domain. Since 1 is dened at (1. 1), we use direct substitution to evaluate the
limit: lim
(i.u)<(1.1)
2rj
r
2
+ 2j
2
=
2(1)(1)
1
2
+ 2(1)
2
=
2
3
.
11. (a) T
i
(6. 4) = lim
I<0
T(6 +/. 4) 3T(6. 4)
/
, so we can approximate T
i
(6. 4) by considering / = 2 and
using the values given in the table: T
i
(6. 4) E
T(8. 4) 3T(6. 4)
2
=
86 380
2
= 3,
T
i
(6. 4) E
T(4. 4) 3T(6. 4)
32
=
72 380
32
= 4. Averaging these values, we estimate T
i
(6. 4) to be approximately
3.5
Cm. Similarly, T
u
(6. 4) = lim
I<0
T(6. 4 +/) 3T(6. 4)
/
, which we can approximate with / = 2:
T
u
(6. 4) E
T(6. 6) 3T(6. 4)
2
=
75 380
2
= 32.5, T
u
(6. 4) E
T(6. 2) 3T(6. 4)
32
=
87 380
32
= 33.5. Averaging these
values, we estimate T
u
(6. 4) to be approximately 33.0
Cm.
(b) Here u =

1
I
2
.
1
I
2
, so by Equation 15.6.9 [ ET 14.6.9], 1

u
T(6. 4) = QT(6. 4) u = T
i
(6. 4)
1
I
2
+T
u
(6. 4)
1
I
2
.
Using our estimates from part (a), we have 1
u
T(6. 4) E (3.5)
1
I
2
+ (33.0)
1
I
2
=
1
2
I
2
E 0.35. This means that as we
move through the point (6. 4) in the direction of u, the temperature increases at a rate of approximately 0.35
Cm.
Alternatively, we can use Denition 15.6.2 [ ET 14.6.2]: 1
u
T(6. 4) = lim
I<0
T
6 +/
1
I
2
. 4 +/
1
I
2
3T(6. 4)
/
,
which we can estimate with / = 2
I
2. Then 1
u
T(6. 4) E
T(8. 6) 3T(6. 4)
2
I
2
=
80 380
2
I
2
= 0,
1
u
T(6. 4) E
T(4. 2) 3T(6. 4)
32
I
2
=
74 380
32
I
2
=
3
I
2
. Averaging these values, we have 1
u
T(6. 4) E
3
2
I
2
E 1.1
Cm.
(c) T
iu
(r. j) =
u
uj
[T
i
(r. j)] = lim
I<0
T
i
(r. j +/) 3T
i
(r. j)
/
, so T
iu
(6. 4) = lim
I<0
T
i
(6. 4 +/) 3T
i
(6. 4)
/
which we can
estimate with / = 2. We have T
i
(6. 4) E 3.5 from part (a), but we will also need values for T
i
(6. 6) and T
i
(6. 2). If we
use / = 2 and the values given in the table, we have
T
i
(6. 6) E
T(8. 6) 3T(6. 6)
2
=
80 3 75
2
= 2.5, T
i
(6. 6) E
T(4. 6) 3T(6. 6)
32
=
68 3 75
32
= 3.5.
Averaging these values, we estimate T
i
(6. 6) E 3.0. Similarly,
T
i
(6. 2) E
T(8. 2) 3T
i
(6. 2)
2
=
90 3 87
2
= 1.5, T
i
(6. 2) E
T(4. 2) 3T(6. 2)
32
=
74 3 87
32
= 6.5.
Averaging these values, we estimate T
i
(6. 2) E 4.0. Finally, we estimate T
iu
(6. 4):
T
iu
(6. 4) E
T
i
(6. 6) 3T
i
(6. 4)
2
=
3.0 3 3.5
2
= 30.25, T
iu
(6. 4) E
T
i
(6. 2) 3T
i
(6. 4)
32
=
4.0 3 3.5
32
= 30.25.
Averaging these values, we have T
iu
(6. 4) E 30.25.
13. 1(r. j) =
2r +j
2
i 1
i
=
1
2
(2r +j
2
)
31'2
(2) =
1
2r +j
2
, 1
u
=
1
2
(2r +j
2
)
31'2
(2j) =
j
2r +j
2
15. o(&. ) = &tan
31
i o
u
= tan
31
, o
r
=
&
1 +
2
17. T(j. q. v) = j ln(q +c
r
) i T
= ln(q +c
r
), T
c
=
j
q +c
r
, T
r
=
jc
r
q +c
r
19. 1(r. j) = 4r
3
3rj
2
i 1
i
= 12r
2
3j
2
, 1
u
= 32rj, 1
ii
= 24r, 1
uu
= 32r, 1
iu
= 1
ui
= 32j
21. 1(r. j. :) = r
I
j
I
:
r
i 1
i
= Ir
I31
j
I
:
r
, 1
u
= |r
I
j
I31
:
r
, 1
:
= ir
I
j
I
:
r31
, 1
ii
= I(I 31)r
I32
j
I
:
r
,
1
uu
= |(| 31)r
I
j
I32
:
r
, 1
::
= i(i31)r
I
j
I
:
r32
, 1
iu
= 1
ui
= I|r
I31
j
I31
:
r
, 1
i:
= 1
:i
= Iir
I31
j
I
:
r31
,
1
u:
= 1
:u
= |ir
I
j
I31
:
r31
23. : = rj +rc
u'i
i
u:
ur
= j 3
j
r
c
u'i
+c
u'i
,
u:
uj
= r +c
u'i
and
r
u:
ur
+j
u:
uj
= r
j 3
j
r
c
u'i
+c
u'i
+j
r +c
u'i
= rj 3jc
u'i
+rc
u'i
+rj +jc
u'i
= rj +rj +rc
u'i
= rj +:.
25. (a) :
i
= 6r + 2 i :
i
(1. 32) = 8 and :
u
= 32j i :
u
(1. 32) = 4, so an equation of the tangent plane is
: 31 = 8(r 31) + 4(j + 2) or : = 8r + 4j + 1.
(b) A normal vector to the tangent plane (and the surface) at (1. 32. 1) is h8. 4. 31i. Then parametric equations for the normal
line there are r = 1 + 8t, j = 32 + 4t, : = 1 3t, and symmetric equations are
r 31
8
=
j + 2
4
=
: 31
31
.
27. (a) Let 1(r. j. :) = r
2
+ 2j
2
33:
2
. Then 1
i
= 2r, 1
u
= 4j, 1
:
= 36:, so 1
i
(2. 31. 1) = 4, 1
u
(2. 31. 1) = 34,
1
:
(2. 31. 1) = 36. From Equation 15.6.19 [ET 14.6.19], an equation of the tangent plane is
4(r 32) 34(j + 1) 36(: 31) = 0 or, equivalently, 2r 32j 33: = 3.
(b) From Equations 15.6.20 [ET 14.6.20], symmetric equations for the normal line are
r 32
4
=
j + 1
34
=
: 31
36
.
29. (a) r(&. ) = (& +) i +&
2
j +
2
k and the point (3. 4. 1) corresponds to & = 2, = 1. Then r
u
= i + 2&j i
r
u
(2. 1) = i + 4 j and r
r
= i + 2 k i r
r
(2. 1) = i + 2 j. A normal vector to the surface at (3. 4. 1) is
r
u
r
r
= 8 i 32 j 34 k, so an equation of the tangent plane there is 8(r 33) 32(j 34) 34(: 31) = 0 or equivalently
4r 3j 32: = 6.
(b) A direction vector for the normal line through (3. 4. 1) is 8 i 32 j 34 k, so a vector equation is
r(t) = (3 i + 4 j +k) +t (8 i 32 j 34 k), and the corresponding parametric equations are r = 3 + 8t, j = 4 32t,
: = 1 34t.
31. The hyperboloid is a level surface of the function 1(r. j. :) = r
2
+ 4j
2
3:
2
, so a normal vector to the surface at (r
0.
j
0
. :
0
)
is Q1(r
0.
j
0
. :
0
) = h2r
0
. 8j
0
. 32:
0
i. A normal vector for the plane 2r + 2j +: = 5 is h2. 2. 1i. For the planes to be
parallel, we need the normal vectors to be parallel, so h2r
0
. 8j
0
. 32:
0
i = I h2. 2. 1i, or r
0
= I , j
0
=
1
4
I, and :
0
= 3
1
2
I.
But r
2
0
+ 4j
2
0
3:
2
0
= 4 i I
2
+
1
4
I
2
3
1
4
I
2
= 4 i I
2
= 4 i I = 2. So there are two such points:
2.
1
2
. 31
and
32. 3
1
2
. 1
.
33. 1(r. j. :) = r
3
j
2
+:
2
i 1
i
(r. j. :) = 3r
2
j
2
+:
2
, 1
u
(r. j. :) =
jr
3
j
2
+:
2
, 1
:
(r. j. :) =
:r
3
j
2
+:
2
,
so 1(2. 3. 4) = 8(5) = 40, 1
i
(2. 3. 4) = 3(4)
I
25 = 60, 1
u
(2. 3. 4) =
3(8)
I
25
=
24
5
, and 1
:
(2. 3. 4) =
4(8)
I
25
=
32
5
. Then the
linear approximation of 1 at (2. 3. 4) is
1(r. j. :) E 1(2. 3. 4) +1
i
(2. 3. 4)(r 32) +1
u
(2. 3. 4)(j 33) +1
:
(2. 3. 4)(: 34)
= 40 + 60(r 32) +
24
5
(j 33) +
32
5
(: 34) = 60r +
24
5
j +
32
5
: 3120
Then (1.98)
3
(3.01)
2
+ (3.97)
2
= 1(1.98. 3.01. 3.97) E 60(1.98) +
24
5
(3.01) +
32
5
(3.97) 3120 = 38.656.
35.
d&
dj
=
u&
ur
dr
dj
+
u&
uj
dj
dj
+
u&
u:
d:
dj
= 2rj
3
(1 + 6j) + 3r
2
j
2
(jc
+c
) + 4:
3
(j cos j + sinj)
37. By the Chain Rule,
u:
uc
=
u:
ur
ur
uc
+
u:
uj
uj
uc
. When c = 1 and t = 2, r = o(1. 2) = 3 and j = /(1. 2) = 6, so
u:
uc
= 1
i
(3. 6)o
s
(1. 2) +1
u
(3. 6) /
s
(1. 2) = (7)(31) + (8)(35) = 347. Similarly,
u:
ut
=
u:
ur
ur
ut
+
u:
uj
uj
ut
, so
u:
ut
= 1
i
(3. 6)o
I
(1. 2) +1
u
(3. 6) /
I
(1. 2) = (7)(4) + (8)(10) = 108.
39.
u:
ur
= 2r1
0
(r
2
3 j
2
),
u:
uj
= 1 3 2j1
0
(r
2
3j
2
)

where 1
0
=
d1
d(r
2
3j
2
)
. Then
j
u:
ur
+r
u:
uj
= 2rj1
0
(r
2
3j
2
) +r 32rj1
0
(r
2
3j
2
) = r.
41.
u:
ur
=
u:
u&
j +
u:
u
3j
r
2
and
u
2
:
ur
2
= j
u
ur
u:
u&
+
2j
r
3
u:
u
+
3j
r
2
u
ur
u:
u
=
2j
r
3
u:
u
+j
u
2
:
u&
2
j +
u
2
:
u u&
3j
r
2
+
3j
r
2
u
2
:
u
2
3j
r
2
+
u
2
:
u&u
j
=
2j
r
3
u:
u
+j
2
u
2
:
u&
2
3
2j
2
r
2
u
2
:
u&u
+
j
2
r
4
u
2
:
u
2
Also
u:
uj
= r
u:
u&
+
1
r
u:
u
and
u
2
:
uj
2
= r
u
uj
u:
u&
+
1
r
u
uj
u:
u
= r
u
2
:
u&
2
r +
u
2
:
u u&
1
r
+
1
r
u
2
:
u
2
1
r
+
u
2
:
u&u
r
= r
2
u
2
:
u&
2
+ 2
u
2
:
u&u
+
1
r
2
u
2
:
u
2
Thus
r
2
u
2
:
ur
2
3j
2
u
2
:
uj
2
=
2j
r
u:
u
+r
2
j
2
u
2
:
u&
2
32j
2
u
2
:
u&u
+
j
2
r
2
u
2
:
u
2
3r
2
j
2
u
2
:
u&
2
32j
2
u
2
:
u&u
3
j
2
r
2
u
2
:
u
2
=
2j
r
u:
u
34j
2
u
2
:
u&u
= 2
u:
u
34&
u
2
:
u&u
since j = r =
&
j
or j
2
= &.
43. Q1 =

:
2
I
j c
i
I
u
,
r:
2
c
i
I
u
2
I
j
, 2:c
i
I
u
= :c
i
I
u
:
I
j,
r:
2
I
j
, 2
45. Q1 = h1
I
r. 32ji, Q1(1. 5) = h1. 310i, u =
1
5
h3. 34i. Then 1
u
1(1. 5) =
43
5
.
47. Q1 =

2rj. r
2
+ 1
2
I
j
, |Q1(2. 1)| =

4.
9
2
. Thus the maximum rate of change of 1 at (2. 1) is

I
145
2
in the
direction
4.
9
2
.
49. First we draw a line passing through Homestead and the eye of the hurricane. We can approximate the directional derivative at
Homestead in the direction of the eye of the hurricane by the average rate of change of wind speed between the points where
this line intersects the contour lines closest to Homestead. In the direction of the eye of the hurricane, the wind speed changes
from 45 to 50 knots. We estimate the distance between these two points to be approximately 8 miles, so the rate of change of
wind speed in the direction given is approximately
50 345
8
=
5
8
= 0.625 knotmi.
51. 1(r. j) = r
2
3rj +j
2
+ 9r 36j + 10 i 1
i
= 2r 3j + 9,
1
u
= 3r + 2j 36, 1
ii
= 2 = 1
uu
, 1
iu
= 31. Then 1
i
= 0 and 1
u
= 0
imply j = 1, r = 34. Thus the only critical point is (34. 1) and
1
ii
(34. 1) 0, 1(34. 1) = 3 0, so 1(34. 1) = 311 is a local minimum.
53. 1(r. j) = 3rj 3r
2
j 3rj
2
i 1
i
= 3j 32rj 3j
2
,
1
u
= 3r 3r
2
32rj, 1
ii
= 32j, 1
uu
= 32r, 1
iu
= 3 32r 32j. Then
1
i
= 0 implies j(3 32r 3j) = 0 so j = 0 or j = 3 32r. Substituting into
1
u
= 0 implies r(3 3r) = 0 or 3r(31 +r) = 0. Hence the critical points are
(0. 0), (3. 0), (0. 3) and (1. 1). 1(0. 0) = 1(3. 0) = 1(0. 3) = 39 < 0 so
(0. 0), (3. 0), and (0. 3) are saddle points. 1(1. 1) = 3 0 and
1
ii
(1. 1) = 32 < 0, so 1(1. 1) = 1 is a local maximum.
55. First solve inside 1. Here 1
i
= 4j
2
32rj
2
3j
3
, 1
u
= 8rj 32r
2
j 33rj
2
.
Then 1
i
= 0 implies j = 0 or j = 4 32r, but j = 0 isnt inside 1. Substituting
j = 4 32r into 1
u
= 0 implies r = 0, r = 2 or r = 1, but r = 0 isnt inside 1,
and when r = 2, j = 0 but (2. 0) isnt inside 1. Thus the only critical point inside
1 is (1. 2) and 1(1. 2) = 4. Secondly we consider the boundary of 1.
On 1
1
: 1(r. 0) = 0 and so 1 = 0 on 1
1
. On 1
2
: r = 3j + 6 and
1(3j + 6. j) = j
2
(6 3j)(32) = 32(6j
2
3j
3
) which has critical pointsat j = 0 and j = 4. Then 1(6. 0) = 0 while
1(2. 4) = 364. On 1
3
: 1(0. j) = 0, so 1 = 0 on 1
3
. Thus on 1 the absolute maximum of 1 is 1(1. 2) = 4 while the
absolute minimum is 1(2. 4) = 364.
57. 1(r. j) = r
3
33r +j
4
32j
2

From the graphs, it appears that 1 has a local maximum 1(31. 0) E 2, local minima 1(1. 1) E 33, and saddle points at
(31. 1) and (1. 0).
To nd the exact quantities, we calculate 1
i
= 3r
2
33 = 0 C r = 1 and 1
u
= 4j
3
34j = 0 C
j = 0, 1, giving the critical points estimated above. Also 1
ii
= 6r, 1
iu
= 0, 1
uu
= 12j
2
34, so using the Second
Derivatives Test, 1(31. 0) = 24 0 and 1
ii
(31. 0) = 36 < 0 indicating a local maximum 1(31. 0) = 2;
1(1. 1) = 48 0 and 1
ii
(1. 1) = 6 0 indicating local minima 1(1. 1) = 33; and 1(31. 1) = 348 and
1(1. 0) = 324, indicating saddle points.
59. 1(r. j) = r
2
j, o(r. j) = r
2
+j
2
= 1 i Q1 =

2rj. r
2
= `Qo = h2`r. 2`ji. Then 2rj = 2`r and r

2
= 2`j
imply ` = r
2
(2j) and ` = j if r 6= 0 and j 6= 0. Hence r
2
= 2j
2
. Then r
2
+j
2
= 1 implies 3j
2
= 1 so j =
1
I
3
and
r =
2
3
. [Note if r = 0 then r
2
= 2`j implies j = 0 and 1 (0. 0) = 0.] Thus the possible points are
2
3
.
1
I
3
and
the absolute maxima are 1
2
3
.
1
I
3
=
2
3
I
3
while the absolute minima are 1
2
3
. 3
1
I
3
= 3
2
3
I
3
.
61. 1(r. j. :) = rj:, o(r. j. :) = r
2
+j
2
+:
2
= 3. Q1 = `Qo i hj:. r:. rji = `h2r. 2j. 2:i. If any of r, j, or : is
zero, then r = j = : = 0 which contradicts r
2
+j
2
+:
2
= 3. Then ` =
j:
2r
=
r:
2j
=
rj
2:
i 2j
2
: = 2r
2
: i
j
2
= r
2
, and similarly 2j:
2
= 2r
2
j i :
2
= r
2
. Substituting into the constraint equation gives r
2
+r
2
+r
2
= 3 i
r
2
= 1 = j
2
= :
2
. Thus the possible points are (1. 1. 1), (1. 31. 1), (31. 1. 1), (31. 31. 1). The absolute maximum
is 1(1. 1. 1) = 1(1. 31. 31) = 1(31. 1. 31) = 1(31. 31. 1) = 1 and the absolute
minimum is 1(1. 1. 31) = 1(1. 31. 1) = 1(31. 1. 1) = 1(31. 31. 31) = 31.
63. 1(r. j. :) = r
2
+j
2
+:
2
, o(r. j. :) = rj
2
:
3
= 2 i Q1 = h2r. 2j. 2:i = `Qo =

`j
2
:
3
. 2`rj:
3
. 3`rj
2
:
2
.
Since rj
2
:
3
= 2, r 6= 0, j 6= 0 and : 6= 0, so 2r = `j
2
:
3
(1), 1 = `r:
3
(2), 2 = 3`rj
2
: (3). Then (2) and (3) imply
1
r:
3
=
2
3rj
2
:
or j
2
=
2
3
:
2
so j = :
2
3
. Similarly (1) and (3) imply
2r
j
2
:
3
=
2
3rj
2
:
or 3r
2
= :
2
so r =
1
I
3
:. But
rj
2
:
3
= 2 so r and : must have the same sign, that is, r =
1
I
3
:. Thus o(r. j. :) = 2 implies
1
I
3
:
2
3
:
2
:
3
= 2 or
: = 3
1'4
and the possible points are (3
31'4
. 3
31'4
I
2. 3
1'4
), (3
31'4
. 33
31'4
I
2. 3
1'4
). However at each of these
points 1 takes on the same value, 2
I
3. But (2. 1. 1) also satises o(r. j. :) = 2 and 1(2. 1. 1) = 6 2
I
3. Thus 1 has an
absolute minimum value of 2
I
3 and no absolute maximum subject to the constraint rj
2
:
3
= 2.
Alternate solution: o(r. j. :) = rj
2
:
3
= 2 implies j
2
=
2
r:
3
, so minimize 1(r. :) = r
2
+
2
r:
3
+:
2
. Then
1
i
= 2r 3
2
r
2
:
3
, 1
:
= 3
6
r:
4
+ 2:, 1
ii
= 2 +
4
r
3
:
3
, 1
::
=
24
r:
5
+ 2 and 1
i:
=
6
r
2
:
4
. Now 1
i
= 0 implies
2r
3
:
3
32 = 0 or : = 1r. Substituting into 1
u
= 0 implies 36r
3
+ 2r
31
= 0 or r =
1
4
I
3
, so the two critical points are
1
4
I
3
.
4
I
3
. Then 1
1
4
I
3
.
4
I
3
= (2 + 4)
2 +
24
3
6
I
3
2
0 and 1
ii
1
4
I
3
.
4
I
3
= 6 0, so each point
is a minimum. Finally, j
2
=
2
r:
3
, so the four points closest to the origin are
1
4
I
3
.
I
2
4
I
3
.
4
I
3
1
4
I
3
. 3
I
2
4
I
3
.
4
I
3
.
65. The area of the triangle is
1
2
co sin0 and the area of the rectangle is /c. Thus, the
area of the whole object is 1(o. /. c) =
1
2
co sin0 +/c. The perimeter of the object
is o(o. /. c) = 2o + 2/ +c = 1. To simplify sin0 in terms of o, /, and c notice
that o
2
sin
2
0 +
1
2
c
2
= o
2
i sin0 =
1
2o
I
4o
2
3c
2
. Thus
1(o. /. c) =
c
4
I
4o
2
3c
2
+/c. (Instead of using 0, we could just have used the
Pythagorean Theorem.) As a result, by Lagranges method, we must nd o, /, c, and ` by solving Q1 = `Qo which gives the
following equations: co(4o
2
3c
2
)
31'2
= 2` (1), c = 2` (2),
1
4
(4o
2
3c
2
)
1'2
3
1
4
c
2
(4o
2
3c
2
)
31'2
+/ = ` (3), and
2o + 2/ +c = 1 (4). From (2), ` =
1
2
c and so (1) produces co(4o
2
3c
2
)
31'2
= c i (4o
2
3c
2
)
1'2
= o i
4o
2
3c
2
= o
2
i c =
I
3 o (5). Similarly, since
4o
2
3c
2
1'2
= o and ` =
1
2
c, (3) gives
o
4
3
c
2
4o
+/ =
c
2
, so from
(5),
o
4
3
3o
4
+/ =
I
3 o
2
i 3
o
2
3
I
3 o
2
= 3/ i / =
o
2
1 +
I
3
(6). Substituting (5) and (6) into (4) we get:

2o +o
1 +
I
3
+
I
3 o = 1 i 3o + 2
I
3 o = 1 i o =
1
3 + 2
I
3
=
2
I
3 33
3
1 and thus
/ =
2
I
3 33
1 +
I
3
6
1 =
3 3
I
3
6
1 and c =

2 3
I
3
1.
PROBLEMS PLUS
1. The areas of the smaller rectangles are
1
= rj,
2
= (1 3r)j,
3
= (13r)(W 3j),
4
= r(W 3j). For 0 $ r $ 1, 0 $ j $ W, let
1(r. j) =
2
1
+
2
2
+
2
3
+
2
4
= r
2
j
2
+ (1 3r)
2
j
2
+ (1 3r)
2
(W 3j)
2
+r
2
(W 3j)
2
= [r
2
+ (1 3r)
2
][j
2
+ (W 3j)
2
]
Then we need to nd the maximum and minimum values of 1(r. j). Here
1
i
(r. j) = [2r 3 2(1 3r)][j
2
+ (W 3j)
2
] = 0 i 4r 321 = 0 or r =
1
2
1, and
1
u
(r. j) = [r
2
+ (1 3r)
2
][2j 32(W 3j)] = 0 i 4j 3 2W = 0 or j = W2. Also
1
ii
= 4[j
2
+ (W 3j)
2
], 1
uu
= 4[r
2
+ (1 3r)
2
], and 1
iu
= (4r 321)(4j 32W). Then
1 = 16[j
2
+ (W 3j)
2
][r
2
+ (1 3r)
2
] 3(4r 321)
2
(4j 32W)
2
. Thus when r =
1
2
1 and j =
1
2
W, 1 0 and
1
ii
= 2W
2
0. Thus a minimum of 1 occurs at
1
2
1.
1
2
W
and this minimum value is 1
1
2
1.
1
2
W
=
1
4
1
2
W
2
.
There are no other critical points, so the maximum must occur on the boundary. Now along the width of the rectangle let
o(j) = 1(0. j) = 1(1. j) = 1
2
[j
2
+ (W 3j)
2
], 0 $ j $ W. Then o
0
(j) = 1
2
[2j 32(W 3j)] = 0 C j =
1
2
W.
And o
1
2
=
1
2
1
2
W
2
. Checking the endpoints, we get o(0) = o(W) = 1
2
W
2
. Along the length of the rectangle let
/(r) = 1(r. 0) = 1(r. W) = W
2
[r
2
+ (1 3r)
2
], 0 $ r $ 1. By symmetry /
0
(r) = 0 C r =
1
2
1 and
/
1
2
1
=
1
2
1
2
W
2
. At the endpoints we have /(0) = /(1) = 1
2
W
2
. Therefore 1
2
W
2
is the maximum value of 1.
This maximum value of 1 occurs when the cutting lines correspond to sides of the rectangle.
3. (a) The area of a trapezoid is
1
2
/(/
1
+/
2
), where / is the height (the distance between the two parallel sides) and /
1
, /
2
are
the lengths of the bases (the parallel sides). From the gure in the text, we see that / = rsin0, /
1
= n 32r, and
/
2
= n 32r + 2rcos 0. Therefore the cross-sectional area of the rain gutter is
(r. 0) =
1
2
rsin0 [(n 32r) + (n 32r + 2rcos 0)] = (rsin0)(n 32r +rcos 0)
= nrsin0 32r
2
sin0 +r
2
sin0 cos 0, 0 < r $
1
2
n, 0 < 0 $
r
2
We look for the critical points of : uur = nsin0 34rsin0 + 2rsin0 cos 0 and
uu0 = nrcos 0 32r
2
cos 0 +r
2
(cos
2
0 3sin
2
0), so uur = 0 C sin0 (n 34r + 2rcos 0) = 0 C
cos 0 =
4r 3n
2r
= 2 3
n
2r
(0 < 0 $
r
2
i sin0 0). If, in addition, uu0 = 0, then
0 = nrcos 0 32r
2
cos 0 +r
2
(2 cos
2
0 31)
= nr
2 3
n
2r
32r
2
2 3
n
2r
+r
2
2 3
n
2r
2
31
= 2nr 3
1
2
n
2
34r
2
+nr +r
2
8 3
4n
r
+
n
2
2r
2
31
= 3nr + 3r
2
= r(3r 3n)
221
222

PROBLEMS PLUS
Since r 0, we must have r =
1
3
n, in which case cos 0 =
1
2
, so 0 =
r
3
, sin0 =
I
3
2
, I =
I
3
6
n, /
1
=
1
3
n, /
2
=
2
3
n,
and =
I
3
12
n
2
. As in Example 15.7.6 [ET 14.7.6], we can argue from the physical nature of this problem that we have
found a local maximum of . Now checking the boundary of , let
o(0) = (n2. 0) =
1
2
n
2
sin0 3
1
2
n
2
sin0 +
1
4
n
2
sin0 cos 0 =
1
8
n
2
sin20, 0 < 0 $
r
2
. Clearly o is maximized when
sin20 = 1 in which case =
1
8
n
2
. Also along the line 0 =
r
2
, let /(r) =
r.
r
2
= nr 32r
2
, 0 < r <
1
2
n i
/
0
(r) = n 34r = 0 C r =
1
4
n, and /
1
4
n
= n
1
4
n
32
1
4
n
2
=
1
8
n
2
. Since
1
8
n
2
<
I
3
12
n
2
, we conclude that
the local maximum found earlier was an absolute maximum.
(b) If the metal were bent into a semi-circular gutter of radius v, we would have n = v and =
1
2
v
2
=
1
2
2
=
n
2
2
.
Since
n
2
2

I
3 n
2
12
, it would be better to bend the metal into a gutter with a semicircular cross-section.
5. Let o(r. j) = r1
j
r
. Then o
i
(r. j) = 1
j
r
+r1
0
j
r
3
j
r
2
= 1
j
r
3
j
r
1
0
j
r
and
o
u
(r. j) = r1
0
j
r
1
r
= 1
0
j
r
. Thus the tangent plane at (r

0
. j
0
. :
0
) on the surface has equation
: 3 r
0
1
j
0
r
0
=

1
j
0
r
0
3j
0
r
31
0
1
0
j
0
r
0
(r 3r
0
) + 1
0
j
0
r
0
(j 3j
0
) i
j
0
r
0
3j
0
r
31
0
1
0
j
0
r
0
r +
1
0
j
0
r
0
j 3: = 0. But any plane whose equation is of the form or +/j +c: = 0

passes through the origin. Thus the origin is the common point of intersection.
7. Since we are minimizing the area of the ellipse, and the circle lies above the r-axis,
the ellipse will intersect the circle for only one value of j. This j-value must
satisfy both the equation of the circle and the equation of the ellipse. Now
r
2
o
2
+
j
2
/
2
= 1 i r
2
=
o
2
/
2
/
2
3j
2
. Substituting into the equation of the

circle gives
o
2
/
2
(/
2
3j
2
) +j
2
32j = 0 i

/
2
3o
2
/
2
j
2
32j +o
2
= 0.
In order for there to be only one solution to this quadratic equation, the discriminant must be 0, so 4 34o
2
/
2
3o
2
/
2
= 0 i
/
2
3o
2
/
2
+o
4
= 0. The area of the ellipse is (o. /) = o/, and we minimize this function subject to the constraint
o(o. /) = /
2
3o
2
/
2
+o
4
= 0.
Now Q = `Qo C / = `(4o
3
32o/
2
), o = `(2/ 32/o
2
) i ` =
/
2o(2o
2
3/
2
)
(1),
` =
o
2/(1 3o
2
)
(2), /
2
3o
2
/
2
+o
4
= 0 (3). Comparing (1) and (2) gives
/
2o(2o
2
3/
2
)
=
o
2/(1 3o
2
)
i
2/
2
= 4o
4
C o
2
=
1
I
2
/. Substitute this into (3) to get / =
3
I
2
i o =
3
2
.
16 MULTIPLE INTEGRALS ET 15
16.1 Double Integrals over Rectangles ET 15.1
1. (a) The subrectangles are shown in the gure.
The surface is the graph of 1(r. j) = rj and { = 4, so we estimate
\ E
3
. =1
2
, =1
1(r
.
. j
,
) {
= 1(2. 2) {+1(2. 4) {+1(4. 2) {+1(4. 4) {+1(6. 2) {+1(6. 4) {
= 4(4) + 8(4) + 8(4) + 16(4) + 12(4) + 24(4) = 288
(b) \ E
3
. =1
2
, =1
1
r
.
. j
,
{ = 1(1. 1) {+1(1. 3) {+1(3. 1) {+1(3. 3) {+1(5. 1) {+1(5. 3) {

= 1(4) + 3(4) + 3(4) + 9(4) + 5(4) + 15(4) = 144
3. (a) The subrectangles are shown in the gure. Since { =
2
4, we estimate
T
sin(r +j) d E
2
.=1
2
,=1
1
r
W
.,
. j
W
.,
{
= 1(0. 0) {+1
0.
r
2
{+1
r
2
. 0
{+1
r
2
.
r
2
{
= 0
r
2
4
+ 1
r
2
4
+ 1
r
2
4
+ 0
r
2
4
=
r
2
2
E 4.935
(b)

T
sin(r +j) d E
2
.=1
2
,=1
1(r
.
. j
,
) {
= 1
r
4
.
r
4
{+1
r
4
.
3r
4
{+ 1
3r
4
.
r
4
{+1
3r
4
.
3r
4
{
= 1
r
2
4
+ 0
r
2
4
+ 0
r
2
4
+ (31)
r
2
4
= 0
5. (a) Each subrectangle and its midpoint are shown in the gure. The area of each
subrectangle is { = 2, so we evaluate 1 at each midpoint and estimate
T
1(r. j) d E
2
. =1
2
, =1
1
r
.
. j
,
{
= 1(1.5. 1) {+1(1.5. 3) {
+ 1(2.5. 1) {+1(2.5. 3) {
= 1(2) + (38)(2) + 5(2) + (31)(2) = 36
223
224 CHAPTER 16 MULTIPLE INTEGRALS ET CHAPTER 15
(b) The subrectangles are shown in the gure. In each subrectangle, the sample point
farthest from the origin is the upper right corner, and the area of each subrectangle
is { =
1
2
. Thus we estimate
T
1(r. j) d E
4
. =1
4
, =1
1(r
.
. j
,
) {
= 1(1.5. 1) {+1(1.5. 2) {+1(1.5. 3) {+1(1.5. 4) {
+1(2. 1) {+1(2. 2) {+1(2. 3) {+1(2. 4) {
+1(2.5. 1) {+1(2.5. 2) {+1(2.5. 3) {+1(2.5. 4) {
+1(3. 1) {+1(3. 2) {+1(3. 3) {+1(3. 4) {
= 1
1
2
+ (34)
1
2
+ (38)
1
2
+ (36)
1
2
+ 3
1
2
+ 0
1
2
+ (35)
1
2
+ (38)
1
2
+ 5
1
2
+ 3
1
2
+ (31)
1
2
+ (34)
1
2
+ 8
1
2
+ 6
1
2
+ 3
1
2
+ 0
1
2
= 33.5
7. The values of 1(r. j) =
52 3r
2
3j
2
get smaller as we move farther from the origin, so on any of the subrectangles in the
problem, the function will have its largest value at the lower left corner of the subrectangle and its smallest value at the upper
right corner, and any other value will lie between these two. So using these subrectangles we have l < \ < 1. (Note that this
is true no matter how 1 is divided into subrectangles.)
9. (a) With i = n = 2, we have { = 4. Using the contour map to estimate the value of 1 at the center of each subrectangle,
we have
T
1(r. j) d E
2
. =1
2
, =1
1
r
.
. j
,
{ = {[1(1. 1) +1(1. 3) +1(3. 1) +1(3. 3)] E 4(27 + 4 + 14 + 17) = 248

(b) 1
ave
=
1
/(T)
T
1(r. j) d E
1
16
(248) = 15.5
11. : = 3 0, so we can interpret the integral as the volume of the solid o that lies below the plane : = 3 and above the
rectangle [32. 2] [1. 6]. o is a rectangular solid, thus
T
3 d = 4 5 3 = 60.
13. : = 1(r. j) = 4 32j D 0 for 0 $ j $ 1. Thus the integral represents the volume of that
part of the rectangular solid [0. 1] [0. 1] [0. 4] which lies below the plane : = 4 32j.
So
T
(4 32j) d = (1)(1)(2) +
1
2
(1)(1)(2) = 3
SECTION 16.2 ITERATED INTEGRALS ET SECTION 15.2

225
15. To calculate the estimates using a programmable calculator, we can use an algorithm
similar to that of Exercise 5.1.7 [ET 5.1.7]. In Maple, we can dene the function
1(r. j) =
I
1 +rc
3u
(calling it f), load the student package, and then use the
command
middlesum(middlesum(f,x=0..1,m),
y=0..1,m);
to get the estimate with n = i
2
squares of equal size. Mathematica has no special
Riemann sum command, but we can dene f and then use nested Sum commands to
calculate the estimates.
n estimate
1 1.141606
4 1.143191
16 1.143535
64 1.143617
256 1.143637
1024 1.143642
17. If we divide 1 into in subrectangles,
T
I d E
r
. =1
n
, =1
1
r
W
.,
. j
W
.,
{ for any choice of sample points
r
W
.,
. j
W
.,
.
But 1
r
W
.,
. j
W
.,
= I always and
r
. =1
n
, =1
{ =area of 1 = (/ 3o)(d 3c). Thus, no matter how we choose the sample
points,
r
. =1
n
, =1
1
r
W
.,
. j
W
.,
{ = I
r
. =1
n
, =1
{ = I(/ 3o)(d 3c) and so
T
I d = lim
r.n<"
r
. =1
n
, =1
1
r
W
.,
. j
W
.,
{ = lim
r.n<"
I
r
. =1
n
, =1
{ = lim
r.n<"
I(/ 3o)(d 3c) = I(/ 3o)(d 3c).
16.2 Iterated Integrals ET 15.2
1.

5
0
12r
2
j
3
dr =

12
r
3
3
j
3
i=5
i=0
= 4r
3
j
3
i=5
i=0
= 4(5)
3
j
3
34(0)
3
j
3
= 500j
3
,
1
0
12r
2
j
3
dj =

12r
2
j
4
4
u=1
u=0
= 3r
2
j
4
u=1
u=0
= 3r
2
(1)
4
33r
2
(0)
4
= 3r
2
3.

3
1
1
0
(1 + 4rj) drdj =

3
1
r + 2r
2
j
i=1
i=0
dj =

3
1
(1 + 2j) dj =

j +j
2
3
1
= (3 + 9) 3(1 + 1) = 10
5.

2
0
r'2
0
rsinj dj dr =

2
0
rdr
r'2
0
sinj dj [as in Example 5] =

r
2
2
2
0
3cos j
r'2
0
= (2 30)(0 + 1) = 2
7.

2
0
1
0
(2r +j)
8
drdj =

2
0
1
2
(2r +j)
9
9
i=1
i=0
dj [substitute & = 2r +j i dr =
1
2
d&]
=
1
18
2
0
[(2 +j)
9
3(0 +j)
9
] dj =
1
18
(2 +j)
10
10
3
j
10
10
2
0
=
1
180
[(4
10
32
10
) 3(2
10
30
10
)] =
1,046,528
180
=
261,632
45
9.

4
1
2
1
r
j
+
j
r
dj dr =

4
1
rln|j| +
1
r

1
2
j
2
u=2
u=1
dr =

4
1
rln2 +
3
2r
dr =

1
2
r
2
ln2 +
3
2
ln|r|
4
1
= 8 ln2 +
3
2
ln4 3
1
2
ln2 =
15
2
ln2 + 3 ln4
1'2
=
21
2
ln2
11.

1
0
1
0
(& 3)
5
d&d =

1
0
1
6
(& 3)
6
u=1
u=0
d =
1
6
1
0
(1 3)
6
3(0 3)
6
d
=
1
6
1
0
(1 3)
6
3
6
d =
1
6
3
1
7
(1 3)
7
3
1
7
1
0
= 3
1
42
[(0 + 1) 3(1 + 0)] = 0
13.

2
0
r
0
v sin
2
0 d0 dv =

2
0
v dv
r
0
sin
2
0 d0 [as in Example 5] =

2
0
v dv
r
0
1
2
(1 3cos 20) d0
=

1
2
v
2
2
0
1
2
0 3
1
2
sin20
r
0
= (2 30)
1
2
3
1
2
sin2
0 3
1
2
sin0
= 2
1
2
[( 30) 3(0 30)] =
15.

T
(6r
2
j
3
35j
4
) d =

3
0
1
0
(6r
2
j
3
35j
4
) dj dr =

3
0
3
2
r
2
j
4
3j
5
u=1
u=0
dr =

3
0
3
2
r
2
31
dr
=

1
2
r
3
3r
3
0
=
27
2
33 =
21
2
17.

T
rj
2
r
2
+ 1
d =

1
0
3
33
rj
2
r
2
+ 1
dj dr =

1
0
r
r
2
+ 1
dr

3
33
j
2
dj =

1
2
ln(r
2
+ 1)
1
0
1
3
j
3
3
33
=
1
2
(ln2 3ln1)
1
3
(27 + 27) = 9 ln2
19.

r'6
0
r'3
0
rsin(r +j) dj dr
=

r'6
0
3rcos(r +j)
u =r'3
u =0
dr =

r'6
0
rcos r 3rcos
r +
r
3
dr
= r
sinr 3sin
r +
r
3
r'6
0
3
r'6
0
sinr 3sin
r +
r
3
dr [by integrating by parts separately for each term]

=
r
6
1
2
31
3cos r + cos
r +
r
3
r'6
0
= 3
r
12
3
3
I
3
2
+ 0 3
31 +
1
2
=
I
331
2
3
r
12
21.

T
rjc
i
2
u
d=

2
0
1
0
rjc
i
2
u
drdj =

2
0
1
2
c
i
2
u
i=1
i=0
dj =
1
2
2
0
(c
u
31) dj =
1
2
c
u
3j
2
0
=
1
2
[(c
2
32) 3(1 30)] =
1
2
(c
2
33)
23. : = 1(r. j) = 4 3r 32j D 0 for 0 $ r $ 1 and 0 $ j $ 1. So the solid
is the region in the rst octant which lies below the plane : = 4 3r 32j
and above [0. 1] [0. 1].
25. \ =

T
(12 33r 32j) d =

3
32
1
0
(12 33r 32j) drdj =

3
32
12r 3
3
2
r
2
32rj
i=1
i=0
dj
=

3
32
21
2
32j
dj =

21
2
j 3j
2
3
32
=
95
2
27. \ =

2
32
1
31
1 3
1
4
r
2
3
1
9
j
2
drdj = 4
2
0
1
0
1 3
1
4
r
2
3
1
9
j
2
drdj
= 4
2
0
r 3
1
12
r
3
3
1
9
j
2
r
i=1
i=0
dj = 4
2
0
11
12
3
1
9
j
2
dj = 4
11
12
j 3
1
27
j
3
2
0
= 4
83
54
=
166
27
SECTION 16.3 DOUBLE INTEGRALS OVER GENERAL REGIONS ET SECTION 15.3

227
29. Here we need the volume of the solid lying under the surface : = rsec
2
j and above the rectangle 1 = [0. 2] [0. 4] in the
rj-plane.
\ =

2
0
r'4
0
rsec
2
j dj dr =

2
0
rdr
r'4
0
sec
2
j dj =

1
2
r
2
2
0
tanj
r'4
0
= (2 30)(tan
r
4
3tan0) = 2(1 30) = 2
31. The solid lies below the surface : = 2 +r
2
+ (j 32)
2
and above the plane : = 1 for 31 $ r $ 1, 0 $ j $ 4. The volume
of the solid is the difference in volumes between the solid that lies under : = 2 +r
2
+ (j 32)
2
over the rectangle
1 = [31. 1] [0. 4] and the solid that lies under : = 1 over 1.
\ =

4
0
1
31
[2 +r
2
+ (j 32)
2
] drdj 3
4
0
1
31
(1) drdj =

4
0
2r +
1
3
r
3
+r(j 32)
2
i =1
i =31
dj 3
1
31
dr

4
0
dj
=

4
0
(2 +
1
3
+ (j 32)
2
) 3(32 3
1
3
3(j 32)
2
)
dj 3[r]
1
31
[j]
4
0
=

4
0
14
3
+ 2(j 32)
2
dj 3[1 3(31)][4 30] =

14
3
j +
2
3
(j 32)
3
4
0
3(2)(4)
=

56
3
+
16
3
0 3
16
3
38 =
88
3
38 =
64
3
33. In Maple, we can calculate the integral by dening the integrand as f
and then using the command int(int(f,x=0..1),y=0..1);.
In Mathematica, we can use the command
Integrate[f,{x,0,1},{y,0,1}]
We nd that
T
r
5
j
3
c
iu
d = 21c 357 E 0.0839. We can use plot3d
(in Maple) or Plot3D (in Mathematica) to graph the function.
35. 1 is the rectangle [31. 1] [0. 5]. Thus, (1) = 2 5 = 10 and
1
ave
=
1
(1)
T
1(r. j) d =
1
10
5
0
1
31
r
2
j drdj =
1
10
5
0
1
3
r
3
j
i=1
i=31
dj =
1
10
5
0
2
3
j dj =
1
10
1
3
j
2
5
0
=
5
6
.
37. Let 1(r. j) =
r 3j
(r +j)
3
. Then a CAS gives
1
0
1
0
1(r. j) dj dr =
1
2
and
1
0
1
0
1(r. j) drdj = 3
1
2
.
To explain the seeming violation of Fubinis Theorem, note that 1 has an innite discontinuity at (0. 0) and thus does not
satisfy the conditions of Fubinis Theorem. In fact, both iterated integrals involve improper integrals which diverge at their
lower limits of integration.
16.3 Double Integrals over General Regions ET 15.3
1.

4
0
I
u
0
rj
2
drdj =

4
0
1
2
r
2
j
2
i=
I
u
i=0
dj =

4
0
1
2
j
2
[(
j )
2
30
2
]dj =
1
2
4
0
j
3
dj =
1
2
1
4
j
4
4
0
=
1
2
(64 30) = 32
3.

1
0
i
i
2
(1 + 2j)dj dr =

1
0
j +j
2
u=i
u=i
2
dr =

1
0
r +r
2
3r
2
3(r
2
)
2
dr
=

1
0
(r 3r
4
)dr =

1
2
r
2
3
1
5
r
5
1
0
=
1
2
3
1
5
30 + 0 =
3
10
5.

r'2
0
cos 0
0
c
sin 0
dv d0 =

r'2
0
vc
sin 0
r=cos 0
r=0
d0 =

r'2
0
(cos 0) c
sin 0
d0 = c
sin 0
r'2
0
= c
sin(r'2)
3c
0
= c 31
7.

T
j
2
d =

1
31
u
3u32
j
2
drdj =

1
31
rj
2
i=u
i=3u32
dj =

1
31
j
2
[j 3(3j 32)] dj
=

1
31
(2j
3
+ 2j
2
)dj =

1
2
j
4
+
2
3
j
3
1
31
=
1
2
+
2
3
3
1
2
+
2
3
=
4
3
9.

T
rd=

r
0
sin i
0
rdj dr =

r
0
[rj]
u=sin i
u=0
dr =

r
0
rsinrdr

integrate by parts
with u = r, Ju = sin rJr
=

3rcos r + sinr
r
0
= 3 cos + sin + 0 3sin0 =
11.

T
j
2
c
iu
d =

4
0
u
0
j
2
c
iu
drdj =

4
0
jc
iu
i=u
i=0
dj =

4
0
jc
u
2
3j
dj
=

1
2
c
u
2
3
1
2
j
2
4
0
=
1
2
c
16
38 3
1
2
+ 0 =
1
2
c
16
3
17
2
13.

1
0
i
2
0
rcos j dj dr =

1
0
rsinj
u =i
2
u =0
dr =

1
0
rsinr
2
dr = 3
1
2
cos r
2
1
0
=
1
2
(1 3cos 1)
15.

2
1
2u31
23u
j
3
drdj =

2
1
rj
3
i=2u31
i=23u
dj =

2
1
[(2j 31) 3(2 3j)] j
3
dj
=

2
1
(3j
4
33j
3
) dj =

3
5
j
5
3
3
4
j
4
2
1
=
96
5
312 3
3
5
+
3
4
=
147
20
17.

2
32
I
43i
2
3
I
43i
2
(2r 3j) dj dr
=

2
32
2rj 3
1
2
j
2
u=
I
43i
2
u=3
I
43i
2
dr
=

2
32
2r
I
4 3r
2
3
1
2
4 3r
2
+ 2r
I
4 3r
2
+
1
2
4 3r
2
dr
=

2
32
4r
I
4 3r
2
dr = 3
4
3
4 3r
2
3'2
2
32
= 0
[Or, note that 4r
I
4 3r
2
is an odd function, so
2
32
4r
I
4 3r
2
dr = 0.]
19. \ =

1
0
i
i
4
(r + 2j) dj dr
=

1
0
rj +j
2
u=i
u=i
4
dr =

1
0
(2r
2
3r
5
3r
8
) dr
=

2
3
r
3
3
1
6
r
6
3
1
9
r
9
1
0
=
2
3
3
1
6
3
1
9
=
7
18
21. \ =

2
1
7 33u
1
rj drdj =

2
1
1
2
r
2
j
i =7 33u
i =1
dj
=
1
2
2
1
(48j 342j
2
+ 9j
3
) dj
=
1
2
24j
2
314j
3
+
9
4
j
4
2
1
=
31
8

229
23. \ =

2
0
3 3
3
2
i
0
(6 33r 32j) dj dr
=

2
0
6j 33rj 3j
2
u =3 3
3
2
i
u =0
dr
=

2
0
6(3 3
3
2
r) 33r(3 3
3
2
r) 3(3 3
3
2
r)
2
dr
=

2
0
9
4
r
2
39r + 9
dr =

3
4
r
3
3
9
2
r
2
+ 9r
2
0
= 6 30 = 6
25.
\ =

2
32
4
i
2
r
2
dj dr
=

2
32
r
2
u=4
u=i
2
dr =

2
32
(4r
2
3r
4
) dr
=

4
3
r
3
3
1
5
r
5
2
32
=
32
3
3
32
5
+
32
3
3
32
5
=
128
15
27.
\ =

1
0
I
1 3i
2
0
j dj dr =

1
0
j
2
2
u =
I
1 3i
2
u =0
dr
=

1
0
1 3r
2
2
dr =
1
2
r 3
1
3
r
3
1
0
=
1
3
29. From the graph, it appears that the two curves intersect at r = 0 and
at r E 1.213. Thus the desired integral is
T
rdE
1.213
0
3i3i
2
i
4
rdj dr =

1.213
0
rj
u =3i 3i
2
u =i
4
dr
=

1.213
0
(3r
2
3r
3
3r
5
) dr =

r
3
3
1
4
r
4
3
1
6
r
6
1.213
0
E 0.713
31. The two bounding curves j = 1 3r
2
and j = r
2
31 intersect at (1. 0) with 1 3r
2
D r
2
31 on [31. 1]. Within this
region, the plane : = 2r + 2j + 10 is above the plane : = 2 3r 3j, so
\ =

1
31
13i
2
i
2
31
(2r + 2j + 10) dj dr 3
1
31
13i
2
i
2
31
(2 3r 3j) dj dr
=

1
31
13i
2
i
2
31
(2r + 2j + 10 3(2 3r 3j)) dj dr
=

1
31
13i
2
i
2
31
(3r + 3j + 8) dj dr =

1
31
3rj +
3
2
j
2
+ 8j
u=13i
2
u=i
2
31
dr
=

1
31
3r(1 3r
2
) +
3
2
(1 3r
2
)
2
+ 8(1 3r
2
) 33r(r
2
31) 3
3
2
(r
2
31)
2
38(r
2
31)
dr
=

1
31
(36r
3
316r
2
+ 6r + 16) dr =

3
3
2
r
4
3
16
3
r
3
+ 3r
2
+ 16r
1
31
= 3
3
2
3
16
3
+ 3 + 16 +
3
2
3
16
3
33 + 16 =
64
3
33. The solid lies below the plane : = 1 3r 3j
or r +j +: = 1 and above the region
1 = {(r. j) | 0 $ r $ 1. 0 $ j $ 1 3r}
in the rj-plane. The solid is a tetrahedron.
35. The two bounding curves j = r
3
3r and j = r
2
+r intersect at the origin and at r = 2, with r
2
+r r
3
3r on (0. 2).
Using a CAS, we nd that the volume is
\ =

2
0
i
2
+i
i
3
3i
: dj dr =

2
0
i
2
+i
i
3
3i
(r
3
j
4
+rj
2
) dj dr =
13,984,735,616
14,549,535
37. The two surfaces intersect in the circle r
2
+j
2
= 1, : = 0 and the region of integration is the disk 1: r
2
+j
2
$ 1.
Using a CAS, the volume is
T
(1 3r
2
3j
2
) d =

1
31
I
13i
2
3
I
13i
2
(1 3r
2
3j
2
) dj dr =

2
.
39. Because the region of integration is
1 = {(r. j) | 0 $ j $
I
r. 0 $ r $ 4} =

(r. j) | j
2
$ r $ 4. 0 $ j $ 2
we have
4
0
I
i
0
1(r. j) dj dr =

T
1(r. j) d =

2
0
4
u
2
1(r. j) drdj.
1 =

(r. j) | 3
9 3j
2
$ r $
9 3j
2
. 0 $ j $ 3
=

(r. j) | 0 $ j $
I
9 3r
2
. 33 $ r $ 3
we have
3
0
I
93u
2
3
I
93u
2
1(r. j) drdj =

T
1(r. j) d
=

3
33
I
93i
2
0
1(r. j) dj dr
1 = {(r. j) | 0 $ j $ lnr, 1 $ r $ 2} = {(r. j) | c
u
$ r $ 2, 0 $ j $ ln2}
we have
2
1
ln i
0
1(r. j) dj dr =

T
1(r. j) d =

ln 2
0
2
c
|
1(r. j) drdj
45.

1
0
3
3u
c
i
2
drdj =

3
0
i'3
0
c
i
2
dj dr =

3
0
c
i
2
j
u=i'3
u=0
dr
=

3
0
r
3
c
i
2
dr =
1
6
c
i
2
3
0
=
c
9
31
6

231
47.
4
0
2
I
i
1
j
3
+ 1
dj dr =

2
0
u
2
0
1
j
3
+ 1
drdj
=

2
0
1
j
3
+ 1
i=u
2
i=0
dj =

2
0
j
2
j
3
+ 1
dj
=
1
3
ln
j
3
+ 1
2
0
=
1
3
(ln9 3ln1) =
1
3
ln9
49.

1
0
r'2
arcsin u
cos r
1 + cos
2
rdrdj
=

r'2
0
sin i
0
cos r
I
1 + cos
2
rdj dr
=

r'2
0
cos r
I
1 + cos
2
r
u=sin i
u=0
dr
=

r'2
0
cos r
I
1 + cos
2
rsinrdr

Let u = cos r, Ju = sin rJr,
Jr = Ju/(sin r)
=

0
1
3&
I
1 +&
2
d& = 3
1
3
1 +&
2
3'2
0
1
=
1
3
I
8 31
=
1
3
2
I
2 31
51. 1 = {(r. j) | 0 $ r $ 1, 3r + 1 $ j $ 1} {(r. j) | 31 $ r $ 0, r + 1 $ j $ 1}

{(r. j) | 0 $ r $ 1, 31 $ j $ r 31} {(r. j) | 31 $ r $ 0, 31 $ j $ 3r 31}, all type I.
T
r
2
d=

1
0
1
1 3i
r
2
dj dr +
0
31
1
i+1
r
2
dj dr +
1
0
i 31
31
r
2
dj dr +
0
31
3i 31
31
r
2
dj dr
= 4
1
0
1
1 3i
r
2
dj dr [by symmetry of the regions and because 1(r. j) = r
2
D 0]
= 4
1
0
r
3
dr = 4
1
4
r
4
1
0
= 1
53. Here Q =

(r. j) | r
2
+j
2
$
1
4
. r D 0. j D 0
, and 0 $ (r
2
+j
2
)
2
$
1
4
2
i 3
1
16
$ 3(r
2
+j
2
)
2
$ 0 so
c
31'16
$ c
3(i
2
+u
2
)
2
$ c
0
= 1 since c
I
is an increasing function. We have (Q) =
1
4
1
2
2
=
r
16
, so by Property 11,
c
31'16
(Q) $
C
c
3(i
2
+u
2
)
2
d $ 1 (Q) i
r
16
c
31'16
$
C
c
3(i
2
+u
2
)
2
d $
r
16
or we can say
0.1844 <

C
c
3(i
2
+u
2
)
2
d < 0.1964. (We have rounded the lower bound down and the upper bound up to preserve the
inequalities.)
55. The average value of a function 1 of two variables dened on a rectangle 1 was
dened in Section 16.1 [ET 15.1] as 1
ave
=
1
/(T)
T
1(r. j)d. Extending
this denition to general regions 1, we have 1
ave
=
1
/(T)
T
1(r. j)d.
Here 1 = {(r. j) | 0 $ r $ 1. 0 $ j $ 3r}, so (1) =
1
2
(1)(3) =
3
2
and
1
ave
=
1
/(T)
T
1(r. j)d =
1
3'2
1
0
3i
0
rj dj dr
=
2
3
1
0
1
2
rj
2
u=3i
u=0
dr =
1
3
1
0
9r
3
dr =
3
4
r
4
1
0
=
3
4
57. Since i $ 1(r. j) $ A,
T
id $
T
1(r. j) d $
T
A d by (8) i
i
T
1 d $
T
1(r. j) d $ A
T
1 d by (7) i i(1) $
T
1(r. j) d $ A(1) by (10).
59.

T
(r
2
tanr +j
3
+ 4) d =

T
r
2
tanrd+
T
j
3
d+
T
4 d. But r
2
tanr is an odd function of r and 1 is
symmetric with respect to the j-axis, so
T
r
2
tanrd = 0. Similarly, j
3
is an odd function of j and 1 is symmetric with
respect to the r-axis, so
T
j
3
d = 0. Thus
T
(r
2
tanr +j
3
+ 4) d = 4
T
d = 4(area of 1) = 4
I
2
2
= 8
61. Since
1 3r
2
3j
2
D 0, we can interpret
1 3r
2
3j
2
d as the volume of the solid that lies below the graph of
: =

1 3r
2
3j
2
and above the region 1 in the rj-plane. : =

1 3r
2
3j
2
is equivalent to r
2
+j
2
+:
2
= 1, : D 0
which meets the rj-plane in the circle r
2
+j
2
= 1, the boundary of 1. Thus, the solid is an upper hemisphere of radius 1
which has volume
1
2
4
3
(1)
3
=
2
3
.
16.4 Double Integrals in Polar Coordinates ET 15.4
1. The region 1 is more easily described by polar coordinates: 1 =

(v. 0) | 0 $ v $ 4, 0 $ 0 $
3r
2
.
Thus
T
1(r. j) d =

3r'2
0
4
0
1(v cos 0. v sin0) v dv d0.
3. The region 1 is more easily described by rectangular coordinates: 1 =

(r. j) | 31 $ r $ 1, 0 $ j $
1
2
r +
1
2
.
Thus
T
1(r. j) d =

1
31
(i+1)'2
0
1(r. j) dj dr.
5. The integral
2r
r
7
4
v dv d0 represents the area of the region
1 = {(v. 0) | 4 $ v $ 7, $ 0 $ 2}, the lower half of a ring.
2r
r
7
4
v dv d0 =

2r
r
d0
7
4
v dv
=

0
2r
r
1
2
v
2
7
4
=
1
2
(49 316) =
33r
2
7. The disk 1 can be described in polar coordinates as 1 = {(v. 0) | 0 $ v $ 3, 0 $ 0 $ 2}. Then
T
rj d =

2r
0
3
0
(v cos 0)(v sin0) v dv d0 =

2r
0
sin0 cos 0 d0
3
0
v
3
dv
=

1
2
sin
2
0
2r
0
1
4
v
4
3
0
= 0.
9.

T
cos(r
2
+j
2
) d =

r
0
3
0
cos(v
2
) v dv d0 =

r
0
d0
3
0
v cos(v
2
) dv
=

0
r
0
1
2
sin(v
2
)
3
0
=
1
2
(sin9 3sin0) =
r
2
sin9
11.

T
c
3i
2
3u
2
d =

r'2
3r'2
2
0
c
3r
2
v dv d0 =

r'2
3r'2
d0
2
0
vc
3r
2
dv
=

0
r'2
3r'2
3
1
2
c
3r
2
2
0
=
3
1
2
(c
34
3c
0
) =
r
2
(1 3c
34
)
13. 1 is the region shown in the gure, and can be described
by 1 = {(v. 0) | 0 $ 0 $ 4. 1 $ v $ 2}. Thus
T
arctan(jr) d =

r'4
0
2
1
arctan(tan0) v dv d0 since jr = tan0.
Also, arctan(tan0) = 0 for 0 $ 0 $ 4, so the integral becomes
r'4
0
2
1
0 v dv d0 =

r'4
0
0 d0

2
1
v dv =

1
2
0
2
r'4
0
1
2
v
2
2
1
=
r
2
32

3
2
=
3
64
2
.
SECTION 16.4 DOUBLE INTEGRALS IN POLAR COORDINATES ET SECTION 15.4

233
15. One loop is given by the region
1 = {(v. 0) |36 $ 0 $ 6, 0 $ v $ cos 30 }, so the area is
T
d =

r'6
3r'6
cos 30
0
v dv d0 =

r'6
3r'6
1
2
v
2
r=cos 30
r=0
d0
=

r'6
3r'6
1
2
cos
2
30 d0 = 2
r'6
0
1
2
1 + cos 60
2
d0
=
1
2
0 +
1
6
sin60
r'6
0
=

12
17. By symmetry,
= 2
r'4
0
sin 0
0
v dv d0 = 2
r'4
0
1
2
v
2
r=sin 0
r=0
d0
=

r'4
0
sin
2
0 d0 =

r'4
0
1
2
(1 3cos 20) d0
=
1
2
0 3
1
2
sin20
r'4
0
=
1
2
r
4
3
1
2
sin
r
2
30 +
1
2
sin0
=
1
8
( 32)
19. \ =

i
2
+u
2
$4
r
2
+j
2
d =

2r
0
2
0
I
v
2
v dv d0 =

2r
0
d0

2
0
v
2
dv =

0
2r
0
1
3
v
3
2
0
= 2
8
3
=
16
3

21. The hyperboloid of two sheets 3r
2
3j
2
+:
2
= 1 intersects the plane : = 2 when 3r
2
3j
2
+4 = 1 or r
2
+j
2
= 3. So the
solid region lies above the surface : =

1 +r
2
+j
2
and below the plane : = 2 for r
2
+j
2
$ 3, and its volume is
\ =

i
2
+u
2
$3
2 3
1 +r
2
+j
2
d =

2r
0
I
3
0
(2 3
1 +v
2
) v dv d0
=

2r
0
d0

I
3
0
(2v 3v
I
1 +v
2
)dv =

0
2r
0
v
2
3
1
3
(1 +v
2
)
3'2
I
3
0
= 2
3 3
8
3
30 +
1
3
=
4
3
23. By symmetry,
\ = 2

i
2
+u
2
$a
2
o
2
3r
2
3j
2
d = 2
2r
0
a
0
o
2
3v
2
v dv d0 = 2
2r
0
d0

a
0
v
o
2
3v
2
dv
= 2
2r
0
3
1
3
(o
2
3v
2
)
3'2
a
0
= 2(2)
0 +
1
3
o
3
=
4r
3
o
3
25. The cone : =

r
2
+j
2
intersects the sphere r
2
+j
2
+:
2
= 1 when r
2
+j
2
+
r
2
+j
2
2
= 1 or r
2
+j
2
=
1
2
. So
\ =

i
2
+u
2
$1'2
1 3r
2
3j
2
3
r
2
+j
2
d =

2r
0
1'
I
2
0
1 3v
2
3v
v dv d0
=

2r
0
d0

1'
I
2
0
v
I
1 3v
2
3v
2
dv =

0
2r
0
3
1
3
(1 3v
2
)
3'2
3
1
3
v
3
1'
I
2
0
= 2
3
1
3
1
I
2
31
=
r
3
2 3
I
2

27. The given solid is the region inside the cylinder r
2
+j
2
= 4 between the surfaces : =

64 34r
2
34j
2
and : = 3
64 34r
2
34j
2
. So
\ =

i
2
+u
2
$4
64 34r
2
34j
2
3
64 34r
2
34j
2
d =

i
2
+u
2
$4
2
64 34r
2
34j
2
d
= 4
2r
0
2
0
I
16 3v
2
v dv d0 = 4
2r
0
d0

2
0
v
I
16 3v
2
dv = 4
2r
0
3
1
3
(16 3v
2
)
3'2
2
0
= 8
3
1
3
(12
3'2
316
2'3
) =
8r
3
64 324
I
3
29.

3
33
I
93i
2
0
sin(r
2
+j
2
)dj dr =

r
0
3
0
sin
v
2
v dv d0
=

r
0
d0

3
0
v sin
v
2
dv = [0]
r
0
3
1
2
cos
v
2
3
0
=
3
1
2
(cos 9 31) =
r
2
(1 3cos 9)
31.

r'4
0
I
2
0
(v cos 0 +v sin0) v dv d0 =

r'4
0
(cos 0 + sin0) d0

I
2
0
v
2
dv
= [sin0 3cos 0]
r'4
0
1
3
v
3
I
2
0
=

I
2
2
3
I
2
2
30 + 1

1
3
2
I
2 30
=
2
I
2
3
33. The surface of the water in the pool is a circular disk 1 with radius 20 ft. If we place 1 on coordinate axes with the origin at
the center of 1 and dene 1(r. j) to be the depth of the water at (r. j), then the volume of water in the pool is the volume of
the solid that lies above 1 =

(r. j) | r
2
+j
2
$ 400
and below the graph of 1(r. j). We can associate north with the
positive j-direction, so we are given that the depth is constant in the r-direction and the depth increases linearly in the
j-direction from 1(0. 320) = 2 to 1(0. 20) = 7. The trace in the j:-plane is a line segment from (0. 320. 2) to (0. 20. 7).
The slope of this line is
7 32
20 3(320)
=
1
8
, so an equation of the line is : 37 =
1
8
(j 320) i : =
1
8
j +
9
2
. Since 1(r. j) is
independent of r, 1(r. j) =
1
8
j +
9
2
. Thus the volume is given by
T
1(r. j) d, which is most conveniently evaluated
using polar coordinates. Then 1 = {(v. 0) | 0 $ v $ 20, 0 $ 0 $ 2} and substituting r = v cos 0, j = v sin0 the integral
becomes
2r
0
20
0
1
8
v sin0 +
9
2
v dv d0 =

2r
0
1
24
v
3
sin0 +
9
4
v
2
r =20
r =0
d0 =

2r
0
1000
3
sin0 + 900
d0
=

3
1000
3
cos 0 + 9000
2r
0
= 1800
Thus the pool contains 1800 E 5655 ft
3
of water.
35.

1
1'
I
2
i
I
1 3i
2
rj dj dr +
I
2
1
i
0
rj dj dr +
2
I
2
I
4 3i
2
0
rj dj dr
=

r'4
0
2
1
v
3
cos 0 sin0 dv d0 =

r'4
0
v
4
4
cos 0 sin0
r =2
r =1
d0
=
15
4
r'4
0
sin0 cos 0 d0 =
15
4
sin
2
0
2
r'4
0
=
15
16
SECTION 16.5 APPLICATIONS OF DOUBLE INTEGRALS ET SECTION 15.5

235
37. (a) We integrate by parts with & = r and d = rc
3i
2
dr. Then d& = dr and = 3
1
2
c
3i
2
, so
"
0
r
2
c
3i
2
dr = lim
I<"
I
0
r
2
c
3i
2
dr = lim
I<"
3
1
2
rc
3i
2
I
0
+
I
0
1
2
c
3i
2
dr
= lim
I<"
3
1
2
tc
3I
2
+
1
2
"
0
c
3i
2
dr = 0 +
1
2
"
0
c
3i
2
dr [by lHospitals Rule]
=
1
4
"
3"
c
3i
2
dr [since c
3i
2
is an even function]
=
1
4
I
[by Exercise 36(c)]
(b) Let & =
I
r. Then &
2
= r i dr = 2&d& i
"
0
I
rc
3i
dr = lim
I<"
I
0
I
rc
3i
dr = lim
I<"
I
I
0
&c
3u
2
2&d& = 2
"
0
&
2
c
3u
2
d& = 2
1
4
I
[by part(a)] =
1
2
I
.
16.5 Applications of Double Integrals ET 15.5
1. Q =

T
o(r. j) d =

3
1
2
0
(2rj +j
2
) dj dr =

3
1
rj
2
+
1
3
j
3
u=2
u=0
dr
=

3
1
4r +
8
3
dr =

2r
2
+
8
3
r
3
1
= 16 +
16
3
=
64
3
C
3. i =

T
a(r. j) d =

2
0
1
31
rj
2
dj dr =

2
0
rdr

1
31
j
2
dj =

1
2
r
2
2
0
1
3
j
3
1
31
= 2
2
3
=
4
3
,
r =
1
r
T
ra(r. j) d =
3
4
2
0
1
31
r
2
j
2
dj dr =
3
4
2
0
r
2
dr

1
31
j
2
dj =
3
4
1
3
r
3
2
0
1
3
j
3
1
31
=
3
4

8
3

2
3
=
4
3
,
j =
1
r
T
ja(r. j) d =
3
4
2
0
1
31
rj
3
dj dr =
3
4
2
0
rdr

1
31
j
3
dj =
3
4
1
2
r
2
2
0
1
4
j
4
1
31
=
3
4
2 0 = 0.
Hence, (r. j) =
4
3
. 0
.
5. i =

2
0
33i
i'2
(r +j) dj dr =

2
0
rj +
1
2
j
2
u=33i
u=i'2
dr =

2
0
3 3
3
2
r
+
1
2
(3 3r)
2
3
1
8
r
2
dr
=

2
0
3
9
8
r
2
+
9
2
dr =

3
9
8
1
3
r
3
+
9
2
r
2
0
= 6,
A
u
=

2
0
33i
i'2
(r
2
+rj) dj dr =

2
0
r
2
j +
1
2
rj
2
u=33i
u=i'2
dr =

2
0
9
2
r 3
9
8
r
3
dr =
9
2
,
A
i
=

2
0
33u
i'2
(rj +j
2
) dj dr =

2
0
1
2
rj
2
+
1
3
j
3
u=33i
u=i'2
dr =

2
0
9 3
9
2
r
dr = 9.
Hence i = 6, (r. j) =
A
u
i
.
A
i
i
=

3
4
.
3
2
.
7. i =

1
0
c
{
0
j dj dr =

1
0
1
2
j
2
u=c
{
u=0
dr =
1
2
1
0
c
2i
dr =
1
4
c
2i
1
0
=
1
4
(c
2
31),
A
u
=

1
0
c
{
0
rj dj dr =
1
2
1
0
rc
2i
dr =
1
2
1
2
rc
2i
3
1
4
c
2i
1
0
=
1
8
(c
2
+ 1),
A
i
=

1
0
c
{
0
j
2
dj dr =

1
0
1
3
j
3
u=c
{
u=0
dr =
1
3
1
0
c
3i
dr =
1
3
1
3
c
3i
1
0
=
1
9
(c
3
31).
Hence i =
1
4
(c
2
31), (r. j) =
1
8
(c
2
+ 1)
1
4
(c
2
31)
.
1
9
(c
3
31)
1
4
(c
2
31)
=

c
2
+ 1
2(c
2
31)
.
4(c
3
31)
9(c
2
31)
.
9. Note that sin(r1) D 0 for 0 $ r $ 1.
i =

1
0
sin(ri'1)
0
j dj dr =

1
0
1
2
sin
2
(r1) dr =
1
2
1
2
r 3
1
4r
sin(2r1)
1
0
=
1
4
1,
A
u
=

1
0
sin(ri'1)
0
r j dj dr =
1
2
1
0
rsin
2
(r1) dr

integrate by parts with
u = r, Ju = sin
2
(tr/J) Jr
=
1
2
r
1
2
r 3
1
4r
sin(2r1)
1
0
3
1
2
1
0
1
2
r 3
1
4r
sin(2r1)
dr
=
1
4
1
2
3
1
2
1
4
r
2
+
1
2
4r
2
cos(2r1)
1
0
=
1
4
1
2
3
1
2
1
4
1
2
+
1
2
4r
2
3
1
2
4r
2
=
1
8
1
2
.
A
i
=

1
0
sin(ri'1)
0
j j dj dr =

1
0
1
3
sin
3
(r1) dr =
1
3
1
0
1 3cos
2
(r1)
sin(r1) dr
[substitute & = cos (r1)] i d& = 3
r
1
sin(r1)]
=
1
3
3
1
r
cos(r1) 3
1
3
cos
3
(r1)
1
0
= 3
1
3r
31 +
1
3
31 +
1
3
=
4
9r
1.
Hence i =
1
4
, (r. j) =
1
2
8
14
.
41(9)
14
=

1
2
.
16
9
.
11. a(r. j) = Ij = Iv sin0, i =

r'2
0
1
0
Iv
2
sin0 dv d0 =
1
3
I
r'2
0
sin0 d0 =
1
3
I
3cos 0
r'2
0
=
1
3
I,
A
u
=

r'2
0
1
0
Iv
3
sin0 cos 0 dv d0 =
1
4
I
r'2
0
sin0 cos 0 d0 =
1
8
I
3cos 20
r'2
0
=
1
8
I,
A
i
=

r'2
0
1
0
Iv
3
sin
2
0 dv d0 =
1
4
I
r'2
0
sin
2
0 d0 =
1
8
I
0 + sin20
r'2
0
=
r
16
I.
Hence (r. j) =
3
8
.
3r
16
.
13. a(r. j) = I
r
2
+j
2
= Iv,
i=

T
a(r. j)d =

r
0
2
1
Iv v dv d0
= I
r
0
d0

2
1
v
2
dv = I()
1
3
v
3
2
1
=
7
3
I,
A
u
=

T
ra(r. j)d =

r
0
2
1
(v cos 0)(Iv) v dv d0 = I
r
0
cos 0 d0

2
1
v
3
dv
= I
sin0
r
0
1
4
v
4
2
1
= I(0)
15
4
= 0
[this is to be expected as the region and density
function are symmetric about the y-axis]
A
i
=

T
ja(r. j)d =

r
0
2
1
(v sin0)(Iv) v dv d0 = I
r
0
sin0 d0

2
1
v
3
dv
= I
3cos 0
r
0
1
4
v
4
2
1
= I(1 + 1)
15
4
=
15
2
I.
Hence (r. j) =
0.
15I'2
7rI'3
=

0.
45
14r
.
15. Placing the vertex opposite the hypotenuse at (0. 0), a(r. j) = I(r
2
+j
2
). Then
i =

a
0
a 3i
0
I
r
2
+j
2
dj dr = I
a
0
or
2
3r
3
+
1
3
(o 3r)
3
dr = I
1
3
or
3
3
1
4
r
4
3
1
12
(o 3r)
4
a
0
=
1
6
Io
4
.
By symmetry,
A
u
= A
i
=

a
0
a 3i
0
Ij(r
2
+j
2
) dj dr = I
a
0
1
2
(o 3r)
2
r
2
+
1
4
(o 3r)
4
dr
= I
1
6
o
2
r
3
3
1
4
or
4
+
1
10
r
5
3
1
20
(o 3r)
5
a
0
=
1
15
Io
5
Hence (r. j) =
2
5
o.
2
5
o
.

237
17. 1
i
=

T
j
2
a(r. j)d =

1
0
c
{
0
j
2
j dj dr =

1
0
1
4
j
4
u=c
{
u=0
dr =
1
4
1
0
c
4i
dr =
1
4
1
4
c
4i
1
0
=
1
16
(c
4
31),
1
u
=

T
r
2
a(r. j) d =

1
0
c
{
0
r
2
j dj dr =

1
0
r
2
1
2
j
2
u=c
{
u=0
dr =
1
2
1
0
r
2
c
2i
dr
=
1
2
1
2
r
2
3
1
2
r +
1
4
c
2i
1
0
[integrate by parts twice] =
1
8
(c
2
31),
and 1
0
= 1
i
+1
u
=
1
16
(c
4
31) +
1
8
(c
2
31) =
1
16
(c
4
+ 2c
2
33).
19. As in Exercise 15, we place the vertex opposite the hypotenuse at (0. 0) and the equal sides along the positive axes.
1
i
=

a
0
a3i
0
j
2
I(r
2
+j
2
) dj dr = I
a
0
a3i
0
(r
2
j
2
+j
4
) dj dr = I
a
0
1
3
r
2
j
3
+
1
5
j
5
u=a3i
u=0
dr
= I
a
0
1
3
r
2
(o 3r)
3
+
1
5
(o 3r)
5
dr = I
1
3
1
3
o
3
r
3
3
3
4
o
2
r
4
+
3
5
or
5
3
1
6
r
6
3
1
30
(o 3r)
6
a
0
=
7
180
Io
6
,
1
u
=

a
0
a3i
0
r
2
I(r
2
+j
2
) dj dr = I
a
0
a3i
0
(r
4
+r
2
j
2
) dj dr = I
a
0
r
4
j +
1
3
r
2
j
3
u=a3i
u=0
dr
= I
a
0
r
4
(o 3r) +
1
3
r
2
(o 3r)
3
dr = I
1
5
or
5
3
1
6
r
6
+
1
3
1
3
o
3
r
3
3
3
4
o
2
r
4
+
3
5
or
5
3
1
6
r
6
a
0
=
7
180
Io
6
,
and 1
0
= 1
i
+1
u
=
7
90
Io
6
.
21. Using a CAS, we nd i =

T
a(r. j) d =

r
0
sin i
0
rj dj dr =

2
8
. Then
r =
1
i
T
ra(r. j) d =
8
r
0
sin i
0
r
2
j dj dr =
2
3
3
1
and
j =
1
i
T
ja(r. j) d =
8
r
0
sin i
0
rj
2
dj dr =
16
9
, so (r. j) =
2
3
3
1
.
16
9
.
The moments of inertia are 1
i
=

T
j
2
a(r. j) d =

r
0
sin i
0
rj
3
dj dr =
3
2
64
,
1
u
=

T
r
2
a(r. j) d =

r
0
sin i
0
r
3
j dj dr =

2
16
(
2
33), and 1
0
= 1
i
+1
u
=

2
64
(4
2
39).
23. 1
i
=

T
j
2
a(r. j)d =

I
0
l
0
aj
2
drdj = a
l
0
dr

I
0
j
2
dj = a
l
0
1
3
j
3
I
0
= a/
1
3
/
3
=
1
3
a//
3
,
1
u
=

T
r
2
a(r. j)d =

I
0
l
0
ar
2
drdj = a
l
0
r
2
dr

I
0
dj = a
1
3
r
3
l
0
[j]
I
0
=
1
3
a/
3
/,
and i = a (area of rectangle) = a// since the lamina is homogeneous. Hence r
2
=
1
u
i
=
1
3
a/
3
/
a//
=
/
2
3
i r =
/
I
3
and j
2
=
1
i
i
=
1
3
a//
3
a//
=
/
2
3
i j =
/
I
3
.
25. In polar coordinates, the region is 1 =

(v. 0) | 0 $ v $ o. 0 $ 0 $
r
2
, so
1
i
=

T
j
2
a d =

r'2
0
a
0
a(v sin0)
2
v dv d0 = a
r'2
0
sin
2
d0

a
0
v
3
dv
= a
1
2
0 3
1
4
sin20
r'2
0
1
4
v
4
a
0
= a
r
4

1
4
o
4
=
1
16
ao
4
,
1
u
=

T
r
2
a d =

r'2
0
a
0
a(v cos 0)
2
v dv d0 = a
r'2
0
cos
2
d0

a
0
v
3
dv
= a
1
2
0 +
1
4
sin20
r'2
0
1
4
v
4
a
0
= a
r
4

1
4
o
4
=
1
16
ao
4
,
and i = a (1) = a
1
4
o
2
since the lamina is homogeneous. Hence r
2
= j
2
=
1
16
ao
4
1
4
ao
2
=
o
2
4
i r = j =
o
2
.
27. (a) 1(r. j) is a joint density function, so we know
R
2
1(r. j) d = 1. Since 1(r. j) = 0 outside the
rectangle [0. 1] [0. 2], we can say
R
2
1(r. j) d=

"
3"
"
3"
1(r. j) dj dr =

1
0
2
0
Cr(1 +j) dj dr
= C
1
0
r
j +
1
2
j
2
u=2
u=0
dr = C
1
0
4rdr = C
2r
2
1
0
= 2C
Then 2C = 1 i C =
1
2
.
(b) 1 (A $ 1. Y $ 1) =
1
3"
1
3"
1(r. j) dj dr =

1
0
1
0
1
2
r(1 +j) dj dr
=

1
0
1
2
r
j +
1
2
j
2
u =1
u =0
dr =

1
0
1
2
r
3
2
dr =
3
4
1
2
r
2
1
0
=
3
8
or 0.375
(c) 1 (A +Y $ 1) = 1 ((A. Y ) M 1) where 1 is the triangular region shown in
the gure. Thus
1(A +Y $ 1) =

T
1(r. j) d =

1
0
1 3i
0
1
2
r(1 +j) dj dr
=

1
0
1
2
r
j +
1
2
j
2
u =13i
u =0
dr =

1
0
1
2
r
1
2
r
2
32r +
3
2
dr
=
1
4
1
0
r
3
34r
2
+ 3r
dr =
1
4
i
4
4
34
i
3
3
+ 3
i
2
2
1
0
=
5
48
E 0.1042
29. (a) 1(r. j) D 0, so 1 is a joint density function if
R
2
1(r. j) d = 1. Here, 1(r. j) = 0 outside the rst quadrant, so
R
2
1(r. j) d=

"
0
"
0
0.1c
3(0.5i +0.2u)
dj dr = 0.1
"
0
"
0
c
30.5i
c
30.2u
dj dr = 0.1
"
0
c
30.5i
dr

"
0
c
30.2u
dj
= 0.1 lim
I<"
I
0
c
30.5i
dr lim
I<"
I
0
c
30.2u
dj = 0.1 lim
I<"
32c
30.5i
I
0
lim
I<"
35c
30.2u
I
0
= 0.1 lim
I<"
32(c
30.5I
31)
lim
I<"
35(c
30.2I
31)
= (0.1) (32)(0 31) (35)(0 31) = 1

Thus 1(r. j) is a joint density function.
(b) (i) No restriction is placed on A, so
1(Y D 1) =

"
3"
"
1
1(r. j) dj dr =

"
0
"
1
0.1c
3(0.5i+0.2u)
dj dr
= 0.1
"
0
c
30.5i
dr

"
1
c
30.2u
dj = 0.1 lim
I<"
I
0
c
30.5i
dr lim
I<"
I
1
c
30.2u
dj
= 0.1 lim
I<"
32c
30.5i
I
0
lim
I<"
35c
30.2u
I
1
= 0.1 lim
I<"
32(c
30.5I
31)
lim
I<"
35(c
30.2I
3c
30.2
)
(0.1) (32)(0 31) (35)(0 3c

30.2
) = c
30.2
E 0.8187
(ii) 1(A $ 2. Y $ 4) =

2
3"
4
3"
1(r. j) dj dr =

2
0
4
0
0.1c
3(0.5i+0.2u)
dj dr
= 0.1
2
0
c
30.5i
dr
4
0
c
30.2u
dj = 0.1
32c
30.5i
2
0
35c
30.2u
4
0
= (0.1) (32)(c
31
31) (35)(c
30.8
31)
= (c
31
31)(c
30.8
31) = 1 +c
31.8
3c
30.8
3c
31
E 0.3481
(c) The expected value of A is given by
j
1
=

R
2
r1(r. j) d =

"
0
"
0
r
0.1c
3(0.5i+0.2u)
dj dr
= 0.1
"
0
rc
30.5i
dr

"
0
c
30.2u
dj = 0.1 lim
I<"
I
0
rc
30.5i
dr lim
I<"
I
0
c
30.2u
dj
To evaluate the rst integral, we integrate by parts with & = r and d = c
30.5i
dr (or we can use Formula 96

239
in the Table of Integrals):

rc
30.5i
dr = 32rc
30.5i
3
32c
30.5i
dr = 32rc
30.5i
34c
30.5i
= 32(r + 2)c
30.5i
.
Thus
j
1
= 0.1 lim
I<"
32(r + 2)c
30.5i
I
0
lim
I<"
35c
30.2u
I
0
= 0.1 lim
I<"
(32)
(t + 2)c
30.5I
32
lim
I<"
(35)
c
30.2I
31
= 0.1(32)
lim
I<"
t + 2
c
0.5I
32
(35)(31) = 2 [by lHospitals Rule]

The expected value of Y is given by
j
2
=

R
2
j 1(r. j) d =

"
0
"
0
j
0.1c
3(0.5 +0.2u)
dj dr
= 0.1
"
0
c
30.5i
dr

"
0
jc
30.2u
dj = 0.1 lim
I<"
I
0
c
30.5i
dr lim
I<"
I
0
jc
30.2u
dj
To evaluate the second integral, we integrate by parts with & = j and d = c
30.2u
dj (or again we can use Formula 96 in
the Table of Integrals) which gives
jc
30.2u
dj = 35jc
30.2u
+
5c
30.2u
dj = 35(j + 5)c
30.2u
. Then
j
2
= 0.1 lim
I<"
32c
30.5i
I
0
lim
I<"
35(j + 5)c
30.2u
I
0
= 0.1 lim
I<"
32(c
30.5I
31)
lim
I<"
35
(t + 5)c
30.2I
35
= 0.1(32)(31) (35)
lim
I<"
t + 5
c
0.2I
35
= 5 [by lHospitals Rule]

31. (a) The random variables A and Y are normally distributed with j
1
= 45, j
2
= 20, o
1
= 0.5, and o
2
= 0.1.
The individual density functions for A and Y , then, are 1
1
(r) =
1
0.5
I
2
c
3(i345)
2
'0.5
and
1
2
(j) =
1
0.1
I
2
c
3(u320)
2
'0.02
. Since A and Y are independent, the joint density function is the product
1(r. j) = 1
1
(r)1
2
(j) =
1
0.5
I
2
c
3(i345)
2
'0.5
1
0.1
I
2
c
3(u320)
2
'0.02
=
10
r
c
32(i345)
2
350(u320)
2
.
Then 1(40 $ A $ 50, 20 $ Y $ 25) =

50
40
25
20
1(r. j) dj dr =
10
r
50
40
25
20
c
32(i345)
2
350(u320)
2
dj dr.
Using a CAS or calculator to evaluate the integral, we get 1(40 $ A $ 50, 20 $ Y $ 25) E 0.500.
(b) 1(4(A 345)
2
+ 100(Y 320)
2
$ 2) =
T
10
r
c
32(i345)
2
350(u320)
2
d, where 1 is the region enclosed by the ellipse
4(r 345)
2
+ 100(j 320)
2
= 2. Solving for j gives j = 20
1
10
2 34(r 345)
2
, the upper and lower halves of the
ellipse, and these two halves meet where j = 20 [since the ellipse is centered at (45. 20)] i 4(r 345)
2
= 2 i
r = 45
1
I
2
. Thus
T
10
r
c
32(i345)
2
350(u320)
2
d =
10
r
45+1'
I
2
4531'
I
2
20+
1
10
I
2 34(i345)
2
203
1
10
I
2 34(i345)
2
c
32(i345)
2
350(u320)
2
dj dr.
Using a CAS or calculator to evaluate the integral, we get 1(4(A 345)
2
+ 100(Y 320)
2
$ 2) E 0.632.
33. (a) If 1(1. ) is the probability that an individual at will be infected by an individual at 1, and I d is the number of
infected individuals in an element of area d, then 1(1. )I d is the number of infections that should result from
exposure of the individual at to infected people in the element of area d. Integration over 1 gives the number of
infections of the person at due to all the infected people in 1. In rectangular coordinates (with the origin at the citys
center), the exposure of a person at is
1 =

T
I1(1. ) d = I
T
20 3d(1. )
20
d = I
1 3
(r 3r
0
)
2
+ (j 3j
0
)
2
20
drdj
(b) If = (0. 0), then
1 = I
1 3
1
20
r
2
+j
2
drdj
= I
2r
0
10
0
1 3
v
20
v dv d0 = 2I
v
2
2
3
v
3
60
10
0
= 2I
50 3
50
3
=
200
3
I E 209I
For at the edge of the city, it is convenient to use a polar coordinate system centered at . Then the polar equation for
the circular boundary of the city becomes v = 20 cos 0 instead of v = 10, and the distance from to a point 1 in the city
is again v (see the gure). So
1 = I
r'2
3r'2
20 cos 0
0
1 3
v
20
v dv d0 = I
r'2
3r'2
v
2
2
3
v
3
60
r=20 cos 0
r=0
d0
= I
r'2
3r'2
200 cos
2
0 3
400
3
cos
3
0
d0 = 200I
r'2
3r'2
1
2
+
1
2
cos 20 3
2
3
1 3sin
2
0
cos 0
d0
= 200I
1
2
0 +
1
4
sin20 3
2
3
sin0 +
2
3

1
3
sin
3
0
r'2
3r'2
= 200I
r
4
+ 0 3
2
3
+
2
9
+
r
4
+ 0 3
2
3
+
2
9
= 200I
r
2
3
8
9
E 136I
Therefore the risk of infection is much lower at the edge of the city than in the middle, so it is better to live at the edge.
16.6 Triple Integrals ET 15.6
1.

T
rj:
2
d\ =

1
0
3
0
2
31
rj:
2
dj d: dr =

1
0
3
0
1
2
rj
2
:
2
u=2
u=31
d: dr =

1
0
3
0
3
2
r:
2
d: dr
=

1
0
1
2
r:
3
:=3
:=0
dr =

1
0
27
2
rdr =
27
4
r
2
1
0
=
27
4
3.

1
0
:
0
i+:
0
6r: dj drd: =

1
0
:
0
6rj:
u=i+:
u=0
drd: =

1
0
:
0
6r:(r +:) drd:
=

1
0
2r
3
: + 3r
2
:
2
i=:
i=0
d: =

1
0
(2:
4
+ 3:
4
) d: =

1
0
5:
4
d: = :
5
1
0
= 1
5.

3
0
1
0
I
13:
2
0
:c
u
drd: dj =

3
0
1
0
r:c
u
i=
I
13:
2
i=0
d: dj =

3
0
1
0
:c
u
I
1 3:
2
d: dj
=

3
0
3
1
3
(1 3:
2
)
3'2
c
u
:=1
:=0
dj =

3
0
1
3
c
u
dj =
1
3
c
u
3
0
=
1
3
(c
3
31)
7.

r'2
0
u
0
i
0
cos(r +j +:)d: drdj =

r'2
0
u
0
sin(r +j +:)
:=i
:=0
drdj
=

r'2
0
u
0
[sin(2r +j) 3sin(r +j)] drdj
=

r'2
0
3
1
2
cos(2r +j) + cos(r +j)
i=u
i=0
dj
=

r'2
0
3
1
2
cos 3j + cos 2j +
1
2
cos j 3cos j
dj
=

3
1
6
sin3j +
1
2
sin2j 3
1
2
sinj
r'2
0
=
1
6
3
1
2
= 3
1
3
9.

T
2rd\ =

2
0
I
43u
2
0
u
0
2rd: drdj =

2
0
I
43u
2
0
2r:
:=u
:=0
drdj =

2
0
I
43u
2
0
2rj drdj
=

2
0
r
2
j
i=
I
43u
2
i=0
dj =

2
0
(4 3j
2
)j dj =

2j
2
3
1
4
j
4
2
0
= 4
SECTION 16.6 TRIPLE INTEGRALS ET SECTION 15.6

241
11. Here 1 = {(r. j. :) | 0 $ r $ 1. 0 $ j $
I
r. 0 $ : $ 1 +r +j}, so
T
6rj d\ =

1
0
I
i
0
1+i+u
0
6rj d: dj dr =

1
0
I
i
0
6rj:
:=1+i+u
:=0
dj dr =

1
0
I
i
0
6rj(1 +r +j) dj dr
=

1
0
3rj
2
+ 3r
2
j
2
+ 2rj
3
u=
I
i
u=0
dr =

1
0
(3r
2
+ 3r
3
+ 2r
5'2
) dr =

r
3
+
3
4
r
4
+
4
7
r
7'2
1
0
=
65
28
13. 1 is the region below the parabolic cylinder : = 1 3j
2
and above the
square [31. 1] [31. 1] in the rj-plane.
T
r
2
c
u
d\ =

1
31
1
31
13u
2
0
r
2
c
u
d: dj dr
=

1
31
1
31
r
2
c
u
(1 3j
2
) dj dr
=

1
31
r
2
dr

1
31
(c
u
3j
2
c
u
) dj
=

1
3
r
3
1
31
c
u
3(j
2
32j + 2)c
u
1
31
integrate by
parts twice
=
1
3
(2)[c 3c 3c
31
+ 5c
31
] =
8
3c
15. Here T = {(r. j. :) | 0 $ r $ 1. 0 $ j $ 1 3r. 0 $ : $ 1 3r 3j}, so
T
r
2
d\ =

1
0
13i
0
13i3u
0
r
2
d: dj dr =

1
0
13i
0
r
2
(1 3r 3j)dj dr
=

1
0
13i
0
(r
2
3r
3
3r
2
j)dj dr =

1
0
r
2
j 3r
3
j 3
1
2
r
2
j
2
u=13i
u=0
dr
=

1
0
r
2
(1 3r) 3r
3
(1 3r) 3
1
2
r
2
(1 3r)
2
dr
=

1
0
1
2
r
4
3r
3
+
1
2
r
2
dr =

1
10
r
5
3
1
4
r
4
+
1
6
r
3
1
0
=
1
10
3
1
4
+
1
6
=
1
60
17. The projection 1 on the j:-plane is the disk j
2
+:
2
$ 1. Using polar
coordinates j = v cos 0 and : = v sin0, we get
T
rd\ =

T
4
4u
2
+4:
2
rdr
d =
1
2
4
2
3(4j
2
+ 4:
2
)
2
d
= 8
2r
0
1
0
(1 3v
4
) v dv d0 = 8
2r
0
d0
1
0
(v 3v
5
) dv
= 8(2)
1
2
v
2
3
1
6
v
6
1
0
=
16r
3
19. The plane 2r +j +: = 4 intersects the rj-plane when
2r +j + 0 = 4 i j = 4 32r, so
1 = {(r. j. :) | 0 $ r $ 2, 0 $ j $ 4 32r, 0 $ : $ 4 32r 3j} and
\ =

2
0
432i
0
432i3u
0
d: dj dr =

2
0
432i
0
(4 32r 3j) dj dr
=

2
0
4j 32rj 3
1
2
j
2
u=432i
u=0
dr
=

2
0
4(4 32r) 32r(4 32r) 3

1
2
(4 32r)
2
dr
=

2
0
(2r
2
38r + 8) dr =

2
3
r
3
34r
2
+ 8r
2
0
=
16
3
21. \ =

3
33
I
93i
2
3
I
93i
2
53u
1
d: dj dr =

3
33
I
93i
2
3
I
93i
2
(5 3j 31) dj dr =

3
33
4j 3
1
2
j
2
u=
I
93i
2
u=3
I
93i
2
dr
=

3
33
8
I
9 3r
2
dr = 8
i
2
I
9 3r
2
+
9
2
sin
31
i
3
3
33
using trigonometric substitution or

Formula 30 in the Table of Integrals
= 8
9
2
sin
31
(1) 3
9
2
sin
31
(31)
= 36
r
2
3
3
r
2
= 36
Alternatively, use polar coordinates to evaluate the double integral:
3
33
I
93i
2
3
I
93i
2
(4 3j) dj dr=

2r
0
3
0
(4 3v sin0) v dv d0
=

2r
0
2v
2
3
1
3
v
3
sin0
r=3
r=0
d0
=

2r
0
(18 39 sin0) d0
= 180 + 9 cos 0
2r
0
= 36
23. (a) The wedge can be described as the region
1 =

(r. j. :) | j
2
+:
2
$ 1, 0 $ r $ 1, 0 $ j $ r
=

(r. j. :) | 0 $ r $ 1, 0 $ j $ r, 0 $ : $
1 3j
2
So the integral expressing the volume of the wedge is
T
d\ =

1
0
i
0
I
1 3u
2
0
d: dj dr.
(b) A CAS gives
1
0
i
0
I
1 3u
2
0
d: dj dr =
r
4
3
1
3
.
(Or use Formulas 30 and 87 from the Table of Integrals.)
25. Here 1(r. j. :) =
1
ln(1 +r +j +:)
and {\ = 2 4 2 = 16, so the Midpoint Rule gives
T
1(r. j. :) d\ E
I
.=1
r
,=1
n
I=1
1
r
.
. j
,
. :
I
{\
= 16[1(1. 2. 1) +1(1. 2. 3) +1(1. 6. 1) +1(1. 6. 3)
+ 1(3. 2. 1) +1(3. 2. 3) +1(3. 6. 1) +1(3. 6. 3)]
= 16
1
ln 5
+
1
ln 7
+
1
ln 9
+
1
ln 11
+
1
ln 7
+
1
ln 9
+
1
ln 11
+
1
ln 13
E 60.533
27. 1 = {(r. j. :) | 0 $ r $ 1, 0 $ : $ 1 3r, 0 $ j $ 2 32:},
the solid bounded by the three coordinate planes and the planes
: = 1 3r, j = 2 32:.

243
29.
If 1
1
, 1
2
, 1
3
are the projections of 1 on the rj-, j:-, and r:-planes, then
1
1
=

(r. j) | 32 $ r $ 2, 0 $ j $ 4 3r
2
=

(r. j) | 0 $ j $ 4, 3
I
4 3j $ r $
I
4 3j
1
2
=

(j. :) | 0 $ j $ 4, 3
1
2
I
4 3j $ : $
1
2
I
4 3j
=

(j. :) | 31 $ : $ 1, 0 $ j $ 4 34:
2
1
3
=

(r. :) | r
2
+ 4:
2
$ 4
Therefore
1 =

(r. j. :) | 32 $ r $ 2, 0 $ j $ 4 3r
2
, 3
1
2
4 3r
2
3j $ : $
1
2
4 3r
2
3j
=

(r. j. :) | 0 $ j $ 4, 3
I
4 3j $ r $
I
4 3j, 3
1
2
4 3r
2
3j $ : $
1
2
4 3r
2
3j
=

(r. j. :) | 31 $ : $ 1, 0 $ j $ 4 34:
2
, 3
4 3j 34:
2
$ r $
4 3j 34:
2
=

(r. j. :) | 0 $ j $ 4, 3
1
2
I
4 3j $ : $
1
2
I
4 3j, 3
4 3j 34:
2
$ r $
4 3j 34:
2
=

(r. j. :) | 32 $ r $ 2, 3
1
2
I
4 3r
2
$ : $
1
2
I
4 3r
2
, 0 $ j $ 4 3r
2
34:
2
=

(r. j. :) | 31 $ : $ 1, 3
I
4 34:
2
$ r $
I
4 34:
2
, 0 $ j $ 4 3r
2
34:
2
Then
T
1(r. j. :) d\ =

2
32
43i
2
0
I
43i
2
3u'2
3
I
43i
2
3u'2
1(r. j. :) d: dj dr =

4
0
I
43u
3
I
43u
I
43i
2
3u'2
3
I
43i
2
3u'2
1(r. j. :) d: drdj
=

1
31
434:
2
0
I
43u34:
2
3
I
43u34:
2
1(r. j. :) drdj d: =

4
0
I
43u'2
3
I
43u'2
I
43u34:
2
3
I
43u34:
2
1(r. j. :) drd: dj
=

2
32
I
43i
2
'2
3
I
43i
2
'2
43i
2
34:
2
0
1(r. j. :) dj d: dr =

1
31
I
434:
2
3
I
434:
2
43i
2
34:
2
0
1(r. j. :) dj drd:
31.
If 1
1
, 1
2
, and 1
3
are the projections of 1 on the rj-, j:-, and r:-planes, then
1
1
=

(r. j) | 32 $ r $ 2. r
2
$ j $ 4
=

(r. j) | 0 $ j $ 4. 3
j $ r $
,
1
2
=

(j. :) | 0 $ j $ 4. 0 $ : $ 2 3
1
2
j
=

(j. :) | 0 $ : $ 2. 0 $ j $ 4 32:
, and
1
3
=

(r. :) | 32 $ r $ 2. 0 $ : $ 2 3
1
2
r
2
=

(r. :) | 0 $ : $ 2. 3
I
4 32: $ r $
I
4 32:
Therefore
1 =

(r. j. :) | 32 $ r $ 2, r
2
$ j $ 4, 0 $ : $ 2 3
1
2
j
=

(r. j. :) | 0 $ j $ 4, 3
j $ r $
j, 0 $ : $ 2 3
1
2
j
=

(r. j. :) | 0 $ j $ 4, 0 $ : $ 2 3
1
2
j, 3
j $ r $
=

(r. j. :) | 0 $ : $ 2, 0 $ j $ 4 32:, 3
j $ r $
=

(r. j. :) | 32 $ r $ 2, 0 $ : $ 2 3
1
2
r
2
, r
2
$ j $ 4 32:
=

(r. j. :) | 0 $ : $ 2, 3
I
4 32: $ r $
I
4 32:, r
2
$ j $ 4 32:
Then

T
1(r. j. :) d\ =

2
32
4
i
2
23u'2
0
1(r. j. :) d: dj dr =

4
0
I
u
3
I
u
23u'2
0
1(r. j. :) d: drdj
=

4
0
23u'2
0
I
u
3
I
u
1(r. j. :) drd: dj =

2
0
432:
0
I
u
3
I
u
1(r. j. :) drdj d:
=

2
32
2 3i
2
'2
0
432:
i
2
1(r. j. :) dj d: dr =

2
0
I
432:
3
I
432:
432:
i
2
1(r. j. :) dj drd:

245
33.
The diagrams show the projections
of 1 on the rj-, j:-, and r:-planes.
Therefore
1
0
1
I
i
1 3u
0
1(r. j. :) d: dj dr =

1
0
u
2
0
13u
0
1(r. j. :) d: drdj =

1
0
13:
0
u
2
0
1(r. j. :) drdj d:
=

1
0
13u
0
u
2
0
1(r. j. :) drd: dj =

1
0
13
I
i
0
13:
I
i
1(r. j. :) dj d: dr
=

1
0
(13:)
2
0
13:
I
i
1(r. j. :) dj drd:
35.
1
0
1
u
u
0
1(r. j. :) d: drdj =

T
1(r. j. :) d\ where 1 = {(r. j. :) | 0 $ : $ j, j $ r $ 1, 0 $ j $ 1}.
If 1
1
, 1
2
, and 1
3
are the projections of 1 on the rj-, j:- and r:-planes then
1
1
= {(r. j) | 0 $ j $ 1, j $ r $ 1} = {(r. j) | 0 $ r $ 1, 0 $ j $ r},
1
2
= {(j. :) | 0 $ j $ 1, 0 $ : $ j} = {(j. :) | 0 $ : $ 1, : $ j $ 1}, and
1
3
= {(r. :) | 0 $ r $ 1, 0 $ : $ r} = {(r. :) | 0 $ : $ 1, : $ r $ 1}.
Thus we also have
1 = {(r. j. :) | 0 $ r $ 1, 0 $ j $ r, 0 $ : $ j} = {(r. j. :) | 0 $ j $ 1, 0 $ : $ j, j $ r $ 1}
= {(r. j. :) | 0 $ : $ 1, : $ j $ 1, j $ r $ 1} = {(r. j. :) | 0 $ r $ 1, 0 $ : $ r, : $ j $ r}
= {(r. j. :) | 0 $ : $ 1, : $ r $ 1, : $ j $ r} .
Then
1
0
1
u
u
0
1(r. j. :) d: drdj =

1
0
i
0
u
0
1(r. j. :) d: dj dr =

1
0
u
0
1
u
1(r. j. :) drd: dj
=

1
0
1
:
1
u
1(r. j. :) drdj d: =

1
0
i
0
i
:
1(r. j. :) dj d: dr
=

1
0
1
:
i
:
1(r. j. :) dj drd:
37. i =

T
a(r. j. :) d\ =

1
0
I
i
0
1+i+u
0
2 d: dj dr =

1
0
I
i
0
2(1 +r +j) dj dr
=

1
0
2j + 2rj +j
2
u=
I
i
u=0
dr =

1
0
2
I
r + 2r
3'2
+r
dr =

4
3
r
3'2
+
4
5
r
5'2
+
1
2
r
2
1
0
=
79
30
A
u:
=

T
ra(r. j. :) d\ =

1
0
I
i
0
1+i+u
0
2rd: dj dr =

1
0
I
i
0
2r(1 +r +j) dj dr
=

1
0
2rj + 2r
2
j +rj
2
u=
I
i
u=0
dr =

1
0
(2r
3'2
+ 2r
5'2
+r
2
) dr =

4
5
r
5'2
+
4
7
r
7'2
+
1
3
r
3
1
0
=
179
105
A
i:
=

T
ja(r. j. :) d\ =

1
0
I
i
0
1+i+u
0
2j d: dj dr =

1
0
I
i
0
2j(1 +r +j) dj dr
=

1
0
j
2
+rj
2
+
2
3
j
3
u=
I
i
u=0
dr =

1
0
r +r
2
+
2
3
r
3'2
dr =

1
2
r
2
+
1
3
r
3
+
4
15
r
5'2
1
0
=
11
10
A
iu
=

T
:a(r. j. :) d\ =

1
0
I
i
0
1+i+u
0
2: d: dj dr =

1
0
I
i
0
:
2
:=1+i+u
:=0
dj dr =

1
0
I
i
0
(1 +r +j)
2
dj dr
=

1
0
I
i
0
(1 + 2r + 2j + 2rj +r
2
+j
2
) dj dr =

1
0
j + 2rj +j
2
+rj
2
+r
2
j +
1
3
j
3
u=
I
i
u=0
dr
=

1
0
I
r +
7
3
r
3'2
+r +r
2
+r
5'2
dr =

2
3
r
3'2
+
14
15
r
5'2
+
1
2
r
2
+
1
3
r
3
+
2
7
r
7'2
1
0
=
571
210
Thus the mass is
79
30
and the center of mass is (r. j. :) =
A
u:
i
.
A
i:
i
.
A
iu
i
=

358
553
.
33
79
.
571
553
.
39. i =

a
0
a
0
a
0
(r
2
+j
2
+:
2
) drdj d: =

a
0
a
0
1
3
r
3
+rj
2
+r:
2
i=a
i=0
dj d: =

a
0
a
0
1
3
o
3
+oj
2
+o:
2
dj d:
=

a
0
1
3
o
3
j +
1
3
oj
3
+oj:
2
u=a
u=0
d: =

a
0
2
3
o
4
+o
2
:
2
d: =

2
3
o
4
: +
1
3
o
2
:
3
a
0
=
2
3
o
5
+
1
3
o
5
= o
5
A
u:
=

a
0
a
0
a
0
r
3
+r(j
2
+:
2
)
drdj d: =

a
0
a
0
1
4
o
4
+
1
2
o
2
(j
2
+:
2
)
dj d:
=

a
0
1
4
o
5
+
1
6
o
5
+
1
2
o
3
:
2
d: =
1
4
o
6
+
1
3
o
6
=
7
12
o
6
= A
i:
= A
iu
by symmetry of 1 and a(r. j. :)
Hence (r. j. :) =
7
12
o.
7
12
o.
7
12
o
.
41. 1
i
=

1
0
1
0
1
0
I(j
2
+:
2
) d: dj dr = I
1
0
1
0
1j
2
+
1
3
1
3
dj dr = I
1
0
2
3
1
4
dr =
2
3
I1
5
.
By symmetry, 1
i
= 1
u
= 1
:
=
2
3
I1
5
.
43. 1
:
=

T
(r
2
+j
2
) a(r. j. :) d\ =

i
2
+u
2
$a
2
I
0
I(r
2
+j
2
) d:
d =

i
2
+u
2
$a
2
I(r
2
+j
2
)/d
= I/
2r
0
a
0
(v
2
) v dv d0 = I/
2r
0
d0

a
0
v
3
dv = I/(2)
1
4
v
4
a
0
= 2I/
1
4
o
4
=
1
2
I/o
4
45. (a) i =

3
33
I
93i
2
3
I
93i
2
53u
1
r
2
+j
2
d: dj dr
(b) (r. j. :) =
A
u:
i
.
A
i:
i
.
A
iu
i
where
A
u:
=

3
33
I
93i
2
3
I
93i
2
53u
1
r
r
2
+j
2
d: dj dr, A
i:
=

3
33
I
93i
2
3
I
93i
2
53u
1
j
r
2
+j
2
d: dj dr, and
A
iu
=

3
33
I
93i
2
3
I
93i
2
53u
1
:
r
2
+j
2
d: dj dr.
(c) 1
:
=

3
33
I
93i
2
3
I
93i
2
53u
1
(r
2
+j
2
)
r
2
+j
2
d: dj dr =

3
33
I
93i
2
3
I
93i
2
53u
1
(r
2
+j
2
)
3'2
d: dj dr

247
47. (a) i =

1
0
I
13i
2
0
u
0
(1 +r +j +:) d: dj dr =
3r
32
+
11
24
(b) (r. j. :) =
i
31
1
0
I
13i
2
0
u
0
r(1 +r +j +:) d: dj dr.
i
31
1
0
I
13i
2
0
u
0
j(1 +r +j +:) d: dj dr.
i
31
1
0
I
13i
2
0
u
0
:(1 +r +j +:) d: dj dr
=

28
9 + 44
.
30 + 128
45 + 220
.
45 + 208
135 + 660
(c) 1
:
=

1
0
I
13i
2
0
u
0
(r
2
+j
2
)(1 +r +j +:) d: dj dr =
68 + 15
240
49. (a) 1(r. j. :) is a joint density function, so we know
R
3
1(r. j. :) d\ = 1. Here we have
R
3
1(r. j. :) d\ =

"
3"
"
3"
"
3"
1(r. j. :) d: dj dr =

2
0
2
0
2
0
Crj: d: dj dr
= C
2
0
rdr

2
0
j dj

2
0
: d: = C
1
2
r
2
2
0
1
2
j
2
2
0
1
2
:
2
2
0
= 8C
Then we must have 8C = 1 i C =
1
8
.
(b) 1(A $ 1. Y $ 1. Z $ 1) =

1
3"
1
3"
1
3"
1(r. j. :) d: dj dr =

1
0
1
0
1
0
1
8
rj: d: dj dr
=
1
8
1
0
rdr

1
0
j dj

1
0
: d: =
1
8
1
2
r
2
1
0
1
2
j
2
1
0
1
2
:
2
1
0
=
1
8
1
2
3
=
1
64
(c) 1(A +Y +Z $ 1) = 1((A. Y. Z) M 1) where 1 is the solid region in the rst octant bounded by the coordinate planes
and the plane r +j +: = 1. The plane r +j +: = 1 meets the rj-plane in the line r +j = 1, so we have
1(A +Y +Z $ 1) =

T
1(r. j. :) d\ =

1
0
13i
0
13i3u
0
1
8
rj: d: dj dr
=
1
8
1
0
13i
0
rj
1
2
:
2
:=13i3u
:=0
dj dr =
1
16
1
0
13i
0
rj(1 3r 3j)
2
dj dr
=
1
16
1
0
13i
0
[(r
3
32r
2
+r)j + (2r
2
32r)j
2
+rj
3
] dj dr
=
1
16
1
0
(r
3
32r
2
+r)
1
2
j
2
+ (2r
2
32r)
1
3
j
3
+r
1
4
j
4
u=13i
u=0
dr
=
1
192
1
0
(r 34r
2
+ 6r
3
34r
4
+r
5
) dr =
1
192
1
30
=
1
5760
51. \ (1) = 1
3
i 1
ave
=
1
1
3
1
0
1
0
1
0
rj: drdj d: =
1
1
3
1
0
rdr

1
0
j dj

1
0
: d:
=
1
1
3
r
2
2
1
0
j
2
2
1
0
:
2
2
1
0
=
1
1
3
1
2
2
1
2
2
1
2
2
=
1
3
8
53. The triple integral will attain its maximum when the integrand 1 3r
2
32j
2
33:
2
is positive in the region 1 and negative
everywhere else. For if 1 contains some region 1 where the integrand is negative, the integral could be increased by excluding
1 from 1, and if 1 fails to contain some part G of the region where the integrand is positive, the integral could be increased
by including G in 1. So we require that r
2
+ 2j
2
+ 3:
2
$ 1. This describes the region bounded by the ellipsoid
r
2
+ 2j
2
+ 3:
2
= 1.
16.7 Triple Integrals in Cylindrical Coordinates ET 15.7
1. (a)
r = 2 cos

4
=
I
2, j = 2 sin

4
=
I
2, : = 1,
so the point is
I
2.
I
2. 1
in rectangular coordinates.
(b)
r = 4 cos
3
r
3
= 2, j = 4 sin
3
r
3
= 32
I
3,
and : = 5, so the point is
2. 32
I
3. 5
in rectangular
coordinates.
3. (a) v
2
= r
2
+j
2
= 1
2
+ (31)
2
= 2 so v =
I
2; tan0 =
j
r
=
31
1
= 31 and the point (1. 31) is in the fourth quadrant of
the rj-plane, so 0 =
7r
4
+ 2n; : = 4. Thus, one set of cylindrical coordinates is
I
2.
7r
4
. 4
.
(b) v
2
= (31)
2
+
3
I
3
2
= 4 so v = 2; tan0 =
3
I
3
31
=
I
3 and the point
31. 3
I
3
is in the third quadrant of the

rj-plane, so 0 =
4r
3
+ 2n; : = 2. Thus, one set of cylindrical coordinates is
2.
4r
3
. 2
.
5. Since 0 =
r
4
but v and : may vary, the surface is a vertical half-plane including the :-axis and intersecting the rj-plane in the
half-line j = r, r D 0.
7. : = 4 3v
2
= 4 3(r
2
+j
2
) or 4 3r
2
3j
2
, so the surface is a circular paraboloid with vertex (0. 0. 4), axis the :-axis, and
opening downward.
9. (a) r
2
+j
2
= v
2
, so the equation becomes : = v
2
.
(b) Substituting r
2
+j
2
= v
2
and j = v sin0, the equation r
2
+j
2
= 2j becomes v
2
= 2v sin0 or v = 2 sin0.
11. 0 $ v $ 2 and 0 $ : $ 1 describe a solid circular cylinder with
radius 2, axis the :-axis, and height 1, but 32 $ 0 $ 2 restricts
the solid to the rst and fourth quadrants of the rj-plane, so we have
a half-cylinder.
13. We can position the cylindrical shell vertically so that its axis coincides with the :-axis and its base lies in the rj-plane. If we
use centimeters as the unit of measurement, then cylindrical coordinates conveniently describe the shell as 6 $ v $ 7,
0 $ 0 $ 2, 0 $ : $ 20.
SECTION 16.7 TRIPLE INTEGRALS IN CYLINDRICAL COORDINATES ET SECTION 15.7

249
15. The region of integration is given in cylindrical coordinates by
1 = {(v. 0. :) | 0 $ 0 $ 2, 0 $ v $ 4, v $ : $ 4}. This represents the solid
region bounded below by the cone : = v and above by the horizontal plane : = 4.
4
0
2r
0
4
r
v d: d0 dv =

4
0
2r
0
v:
:=4
:=r
d0 dv =

4
0
2r
0
v(4 3v) d0 dv
=

4
0
(4v 3v
2
) dv

2r
0
d0 =

2v
2
3
1
3
v
3
4
0
2r
0
=

32 3
64
3
(2) =
64r
3
17. In cylindrical coordinates, 1 is given by {(v. 0. :) | 0 $ 0 $ 2. 0 $ v $ 4. 35 $ : $ 4}. So
r
2
+j
2
d\ =

2r
0
4
0
4
35
I
v
2
v d: dv d0 =

2r
0
d0

4
0
v
2
dv

4
35
d:
=

0
2r
0
1
3
v
3
4
0
4
35
= (2)
64
3
(9) = 384
19. In cylindrical coordinates 1 is bounded by the paraboloid : = 1 +v
2
, the cylinder v
2
= 5 or v =
I
5, and the rj-plane,
so 1 is given by
(v. 0. :) | 0 $ 0 $ 2. 0 $ v $
I
5. 0 $ : $ 1 +v
2
. Thus
T
c
:
d\ =

2r
0
I
5
0
1+r
2
0
c
:
v d: dv d0 =

2r
0
I
5
0
v
c
:
:=1+r
2
:=0
dv d0 =

2r
0
I
5
0
v(c
1+r
2
31) dv d0
=

2r
0
d0

I
5
0
vc
1+r
2
3v
dv = 2
1
2
c
1+r
2
3
1
2
v
2
I
5
0
= (c
6
3c 35)
21. In cylindrical coordinates, 1 is bounded by the cylinder v = 1, the plane : = 0, and the cone : = 2v. So
1 = {(v. 0. :) | 0 $ 0 $ 2. 0 $ v $ 1. 0 $ : $ 2v} and
T
r
2
d\ =

2r
0
1
0
2r
0
v
2
cos
2
0 v d: dv d0 =

2r
0
1
0
v
3
cos
2
0 :
:=2r
:=0
dv d0 =

2r
0
1
0
2v
4
cos
2
0 dv d0
=

2r
0
2
5
v
5
cos
2
0
r=1
r=0
d0 =
2
5
2r
0
cos
2
0 d0 =
2
5
2r
0
1 + cos 20
2
d0 =
1
5
0 +
1
2
sin20
2r
0
=
2
5
23. (a) The paraboloids intersect when r
2
+j
2
= 36 33r
2
33j
2
i r
2
+j
2
= 9, so the region of integration
is 1 =

(r. j) | r
2
+j
2
$ 9
. Then, in cylindrical coordinates,

1 =

(v. 0. :) | v
2
$ : $ 36 33v
2
, 0 $ v $ 3, 0 $ 0 $ 2
and
\ =

2r
0
3
0
36 33r
2
r
2
v d: dv d0 =

2r
0
3
0
36v 34v
3
dv d0 =

2r
0
18v
2
3v
4
r=3
r=0
d0 =

2r
0
81 d0 = 162.
(b) For constant density 1, i = 1\ = 1621 from part (a). Since the region is homogeneous and symmetric,
A
u:
= A
i:
= 0 and
A
iu
=

2r
0
3
0
3633r
2
r
2
(:1) v d: dv d0 = 1
2r
0
3
0
v
1
2
:
2
:=3633r
2
:=r
2
dv d0
=
1
2
2r
0
3
0
v((36 33v
2
)
2
3v
4
) dv d0 =
1
2
2r
0
d0

3
0
(8v
5
3216v
3
+ 1296v) dv
=
1
2
(2)
8
6
v
6
3
216
4
v
4
+
1296
2
v
2
3
0
= 1(2430) = 24301
Thus (r. j. :) =
A
u:
i
.
A
i:
i
.
A
iu
i
=

0. 0.
2430r1
162r1
= (0. 0. 15).
25. The paraboloid : = 4r
2
+ 4j
2
intersects the plane : = o when o = 4r
2
+ 4j
2
or r
2
+j
2
=
1
4
o. So, in cylindrical
coordinates, 1 =

(v. 0. :) | 0 $ v $
1
2
I
o. 0 $ 0 $ 2. 4v
2
$ : $ o
. Thus
i=

2r
0
I
a'2
0
a
4r
2
1v d: dv d0 = 1
2r
0
I
a'2
0
(ov 34v
3
) dv d0
= 1
2r
0
1
2
ov
2
3v
4
r=
I
a'2
r=0
d0 = 1
2r
0
1
16
o
2
d0 =
1
8
o
2
1
Since the region is homogeneous and symmetric, A
u:
= A
i:
= 0 and
A
iu
=

2r
0
I
a'2
0
a
4r
2
1v: d: dv d0 = 1
2r
0
I
a'2
0
1
2
o
2
v 38v
5
dv d0
= 1
2r
0
1
4
o
2
v
2
3
4
3
v
6
r=
I
a'2
r=0
d0 = 1
2r
0
1
24
o
3
d0 =
1
12
o
3
1
Hence (r. j. :) =
0. 0.
2
3
o
.
27. The region of integration is the region above the cone : =

r
2
+j
2
, or : = v, and below the plane : = 2. Also, we have
32 $ j $ 2 with 3
4 3j
2
$ r $
4 3j
2
which describes a circle of radius 2 in the rj-plane centered at (0. 0). Thus,
2
32
I
43u
2
3
I
43u
2
2
I
i
2
+u
2
r: d: drdj =

2r
0
2
0
2
r
(v cos 0) : v d: dv d0 =

2r
0
2
0
2
r
v
2
(cos 0) : d: dv d0
=

2r
0
2
0
v
2
(cos 0)
1
2
:
2
:=2
:=r
dv d0 =
1
2
2r
0
2
0
v
2
(cos 0)
4 3v
2
dv d0
=
1
2
2r
0
cos 0 d0
2
0
4v
2
3v
4
dv =
1
2
[sin0]
2r
0
4
3
v
3
3
1
5
v
5
2
0
= 0
29. (a) The mountain comprises a solid conical region C. The work done in lifting a small volume of material {\ with density
o(1) to a height /(1) above sea level is /(1)o(1) {\ . Summing over the whole mountain we get
W =

C
/(1)o(1) d\ .
(b) Here C is a solid right circular cone with radius 1 = 62,000 ft, height 1 = 12,400 ft,
and density o(1) = 200 lbft
3
at all points 1 in C. We use cylindrical coordinates:
W =

2r
0
1
0
T(13:'1)
0
: 200v dv d: d0 = 2
1
0
200:
1
2
v
2
r=T(13:'1)
r=0
d:
= 400
1
0
:
1
2
2
1 3
:
1
2
d: = 2001
2
1
0
: 3
2:
2
1
+
:
3
1
2
d:
= 2001
2
:
2
2
3
2:
3
31
+
:
4
41
2
1
0
= 2001
2
1
2
2
3
21
2
3
+
1
2
4
=
50
3
1
2
1
2
=
50
3
(62,000)
2
(12,400)
2
E 3.1 10
19
ft-lb
v
1
=
1 3:
1
= 1 3
:
1
SECTION 16.8 TRIPLE INTEGRALS IN SPHERICAL COORDINATES ET SECTION 15.8

251
16.8 Triple Integrals in Spherical Coordinates ET 15.8
1. (a)
r = a sinccos 0 = (1) sin0 cos 0 = 0,
j = a sincsin0 = (1) sin0 sin0 = 0, and
: = a cos c = (1) cos 0 = 1 so the point is
(0. 0. 1) in rectangular coordinates.
(b)
r = 2 sin
r
4
cos
r
3
=
I
2
2
, j = 2 sin
r
4
sin
r
3
=
I
6
2
,
: = 2 cos
r
4
=
I
2 so the point is
I
2
2
.
I
6
2
.
I
2
in
rectangular coordinates.
3. (a) a =

r
2
+j
2
+:
2
=
I
1 + 3 + 12 = 4, cos c =
:
a
=
2
I
3
4
=
I
3
2
i c =

6
, and
cos 0 =
r
a sinc
=
1
4 sin(6)
=
1
2
i 0 =

3
[since j 0]. Thus spherical coordinates are
4.

3
.

6
.
(b) a =
I
0 + 1 + 1 =
I
2, cos c =
31
I
2
i c =
3
4
, and cos 0 =
0
I
2 sin(34)
= 0 i 0 =
3
2
[since j < 0].
Thus spherical coordinates are
I
2.
3
2
.
3
4
.
5. Since c =
r
3
, the surface is the top half of the right circular cone with vertex at the origin and axis the positive :-axis.
7. a = sin0 sinc i a
2
= a sin0 sinc C r
2
+j
2
+:
2
= j C r
2
+j
2
3j +
1
4
+:
2
=
1
4
C
r
2
+ (j 3
1
2
)
2
+:
2
=
1
4
. Therefore, the surface is a sphere of radius
1
2
centered at
0.
1
2
. 0
.
9. (a) r = a sinccos 0, j = a sincsin0, and : = a cos c, so the equation :
2
= r
2
+j
2
becomes
(a cos c)
2
= (a sinccos 0)
2
+ (a sincsin0)
2
or a
2
cos
2
c = a
2
sin
2
c. If a 6= 0, this becomes cos
2
c = sin
2
c. (a = 0
corresponds to the origin which is included in the surface.) There are many equivalent equations in spherical coordinates,
such as tan
2
c = 1, 2 cos
2
c = 1, cos 2c = 0, or even c =
r
4
, c =
3r
4
.
(b) r
2
+:
2
= 9 C (a sinccos 0)
2
+ (a cos c)
2
= 9 C a
2
sin
2
ccos
2
0 +a
2
cos
2
c = 9 or
a
2
sin
2
ccos
2
0 + cos
2
c
= 9.
11. a = 2 represents a sphere of radius 2, centered at the origin, so a $ 2 is this
sphere and its interior. 0 $ c $
r
2
restricts the solid to that portion of the
region that lies on or above the rj-plane, and 0 $ 0 $
r
2
further restricts the
solid to the rst octant. Thus the solid is the portion in the rst octant of the
solid ball centered at the origin with radius 2.
13. a $ 1 represents the solid sphere of radius 1 centered at the origin.
3r
4
$ c $ restricts the solid to that portion on or below the cone c =
3r
4
.
15. : D
r
2
+j
2
because the solid lies above the cone. Squaring both sides of this inequality gives :
2
D r
2
+j
2
i
2:
2
D r
2
+j
2
+:
2
= a
2
i :
2
= a
2
cos
2
c D
1
2
a
2
i cos
2
c D
1
2
. The cone opens upward so that the inequality is
cos c D
1
I
2
, or equivalently 0 $ c $
r
4
. In spherical coordinates the sphere : = r
2
+j
2
+:
2
is a cos c = a
2
i
a = cos c. 0 $ a $ cos c because the solid lies below the sphere. The solid can therefore be described as the region in
spherical coordinates satisfying 0 $ a $ cos c, 0 $ c $
r
4
.
17. The region of integration is given in spherical coordinates by
1 = {(a. 0. c) | 0 $ a $ 3. 0 $ 0 $ 2. 0 $ c $ 6}. This represents the solid
region in the rst octant bounded above by the sphere a = 3 and below by the cone
c = 6.
r'6
0
r'2
0
3
0
a
2
sincda d0 dc =

r'6
0
sincdc

r'2
0
d0

3
0
a
2
da
=

3cos c
r'6
0
r'2
0
1
3
a
3
3
0
=

1 3
I
3
2
(9) =
9
4
2 3
I
3
19. The solid 1 is most conveniently described if we use cylindrical coordinates:

1 =

(v. 0. :) | 0 $ 0 $
r
2
. 0 $ v $ 3. 0 $ : $ 2
. Then
T
1(r. j. :) d\ =

r'2
0
3
0
2
0
1(v cos 0. v sin0. :) v d: dv d0.
21. In spherical coordinates, 1 is represented by {(a. 0. c) | 0 $ a $ 5. 0 $ 0 $ 2. 0 $ c $ }. Thus
T
(r
2
+j
2
+:
2
)
2
d\ =

r
0
2r
0
5
0
(a
2
)
2
a
2
sincda d0 dc =

r
0
sincdc

2r
0
d0

5
0
a
6
da
=

3cos c
r
0
2r
0
1
7
a
7
5
0
= (2)(2)
78,125
7
=
312,500
7
E 140,249.7
23. In spherical coordinates, 1 is represented by
(a. 0. c)
1 $ a $ 2. 0 $ 0 $
r
2
. 0 $ c $
r
2
. Thus
T
: d\ =

r'2
0
r'2
0
2
1
(a cos c) a
2
sincda d0 dc =

r'2
0
cos csincdc
r'2
0
d0
2
1
a
3
da
=

1
2
sin
2
c
r'2
0
r'2
0
1
4
a
4
2
1
=

1
2
r
2
15
4
=
15r
16
SECTION 16.8 TRIPLE INTEGRALS IN SPHERICAL COORDINATES ET SECTION 15.8

253
25.

T
r
2
d\ =

r
0
r
0
4
3
(a sinccos 0)
2
a
2
sincda dcd0 =

r
0
cos
2
0 d0

r
0
sin
3
cdc

4
3
a
4
da
=

1
2
0 +
1
4
sin20
r
0
3
1
3
(2 + sin
2
c) cos c
r
0
1
5
a
5
4
3
=

r
2
2
3
+
2
3
1
5
(4
5
33
5
) =
1562
15

27. The solid region is given by 1 =

(a. 0. c) | 0 $ a $ o. 0 $ 0 $ 2.
r
6
$ c $
r
3
and its volume is

\ =

T
d\ =

r'3
r'6
2r
0
a
0
a
2
sincda d0 dc =

r'3
r'6
sincdc

2r
0
d0
a
0
a
2
da
= [3cos c]
r'3
r'6
[0]
2r
0
1
3
a
3
a
0
=

3
1
2
+
I
3
2
(2)
1
3
o
3
=
I
331
3
o
3
29. (a) Since a = 4 cos c implies a
2
= 4a cos c, the equation is that of a sphere of radius 2 with center at (0. 0. 2). Thus
\ =

2r
0
r'3
0
4 cos c
0
a
2
sincda dcd0 =

2r
0
r'3
0
1
3
a
3
p=4 cos c
p=0
sincdcd0 =

2r
0
r'3
0
64
3
cos
3
c
sincdcd0
=

2r
0
3
16
3
cos
4
c
c=r'3
c=0
d0 =

2r
0
3
16
3
1
16
31
d0 = 50
2r
0
= 10
(b) By the symmetry of the problemA
u:
= A
i:
= 0. Then
A
iu
=

2r
0
r'3
0
4 cos c
0
a
3
cos csincda dcd0 =

2r
0
r'3
0
cos csinc
64 cos
4
c
dcd0
=

2r
0
64
3
1
6
cos
6
c
c=r'3
c=0
d0 =

2r
0
21
2
d0 = 21
Hence (r. j. :) = (0. 0. 2.1).
31. By the symmetry of the region, A
iu
= 0 and A
u:
= 0. Assuming constant density 1,
i =

T
1\ = 1
r
0
r
0
4
3
a
2
sincda dcd0 = 1
r
0
d0

r
0
sincdc

4
3
a
2
da
= 1
3cos c
r
0
1
3
a
3
4
3
= 21
37
3
=
74
3
1
and A
i:
=

T
j 1 d\ = 1
r
0
r
0
4
3
(a sincsin0) a
2
sincda dcd0 = 1
r
0
sin0 d0
r
0
sin
2
cdc
4
3
a
3
da
= 1

3cos 0
r
0
1
2
c 3
1
4
sin2c
r
0
1
4
a
4
4
3
= 1(2)
r
2
1
4
(256 381) =
175
4
1
Thus the centroid is (r. j. :) =
A
u:
i
.
A
i:
i
.
A
iu
i
=

0.
17514
7413
. 0
=

0.
525
296
. 0
.
33. (a) The density function is a(r. j. :) = 1, a constant, and by the symmetry of the problem A
i:
= A
u:
= 0. Then
A
iu
=

2r
0
r'2
0
a
0
1a
3
sinc cos cda dcd0 =
1
2
1o
4
r'2
0
sinc cos cdc =
1
8
1o
4
. But the mass is 1(volume of
the hemisphere) =
2
3
1o
3
, so the centroid is
0. 0.
3
8
o
.
(b) Place the center of the base at (0. 0. 0); the density function is a(r. j. :) = 1. By symmetry, the moments of inertia about
any two such diameters will be equal, so we just need to nd 1
i
:
1
i
=

2r
0
r'2
0
a
0
(1a
2
sinc) a
2
(sin
2
c sin
2
0 + cos
2
c) da dcd0
= 1
2r
0
r'2
0
(sin
3
c sin
2
0 + sinc cos
2
c)
1
5
o
5
dcd0
=
1
5
1o
5
2r
0
sin
2
0
3cos c +
1
3
cos
3
c
3
1
3
cos
3
c
c=r'2
c=0
d0 =
1
5
1o
5
2r
0
2
3
sin
2
0 +
1
3
d0
=
1
5
1o
5
2
3
1
2
0 3
1
4
sin20
+
1
3
0
2r
0
=
1
5
1o
5
2
3
( 30) +
1
3
(2 30)
=
4
15
1o
5

35. In spherical coordinates : =

r
2
+j
2
becomes cos c = sinc or c =
r
4
. Then
\ =

2r
0
r'4
0
1
0
a
2
sincda dcd0 =

2r
0
d0

r'4
0
sincdc

1
0
a
2
da =
1
3
2 3
I
2
,
A
iu
=

2r
0
r'4
0
1
0
a
3
sinccos cda dcd0 = 2
3
1
4
cos 2c
r'4
0
1
4
=
r
8
and by symmetry A
u:
= A
i:
= 0.
Hence (r. j. :) =
0. 0.
3
8
2 3
I
2
.
37. In cylindrical coordinates the paraboloid is given by : = v
2
and the plane by : = 2v sin0 and they intersect in the circle
v = 2 sin0. Then
T
: d\ =

r
0
2 sin 0
0
2r sin 0
r
2
v: d: dv d0 =
5r
6
[using a CAS].
39. The region 1 of integration is the region above the cone : =

r
2
+j
2
and below the sphere r
2
+j
2
+:
2
= 2 in the rst
octant. Because 1 is in the rst octant we have 0 $ 0 $
r
2
. The cone has equation c =
r
4
(as in Example 4), so 0 $ c $
r
4
,
and 0 $ a $
I
2. So the integral becomes
r'4
0
r'2
0
I
2
0
(a sinccos 0) (a sincsin0) a
2
sincda d0 dc
=

r'4
0
sin
3
cdc

r'2
0
sin0 cos 0 d0

I
2
0
a
4
da =

r'4
0
1 3cos
2
c
sincdc

1
2
sin
2
0
r'2
0
1
5
a
5
I
2
0
=

1
3
cos
3
c 3cos c
r'4
0

1
2

1
5
I
2
5
=

I
2
12
3
I
2
2
3
1
3
31
2
I
2
5
=
4
I
235
15
41. In cylindrical coordinates, the equation of the cylinder is v = 3, 0 $ : $ 10.
The hemisphere is the upper part of the sphere radius 3, center (0. 0. 10), equation
v
2
+ (: 310)
2
= 3
2
, : D 10. In Maple, we can use the coords=cylindrical option
in a regular plot3d command. In Mathematica, we can use ParametricPlot3D.
43. If 1 is the solid enclosed by the surface a = 1 +
1
5
sin60 sin5c, it can be described in spherical coordinates as
1 =

(a. 0. c) | 0 $ a $ 1 +
1
5
sin60 sin5c. 0 $ 0 $ 2. 0 $ c $
. Its volume is given by

\ (1) =

T
d\ =

r
0
2r
0
1 +(sin 60 sin 5c)'5

0
a
2
sincda d0 dc =
136r
99
[using a CAS].
45. (a) From the diagram, : = v cot c
0
to : =
I
o
2
3v
2
, v = 0
to v = o sinc
0
(or use o
2
3v
2
= v
2
cot
2
c
0
). Thus
\ =

2r
0
a sin c
0
0
I
a
2
3r
2
r cot c
0
v d: dv d0
= 2
a sin c
0
0
v
I
o
2
3v
2
3v
2
cot c
0
dv
=
2r
3
3(o
2
3v
2
)
3'2
3v
3
cot c
0
a sin c
0
0
=
2r
3
o
2
3o
2
sin
2
c
0
3'2
3o
3
sin
3
c
0
cot c
0
+o
3
=
2
3
o
3
1 3
cos
3
c
0
+ sin
2
c
0
cos c
0
=
2
3
o
3
(1 3cos c
0
)
SECTION 16.9 CHANGE OF VARIABLES IN MULTIPLE INTEGRALS ET SECTION 15.9

255
(b) The wedge in question is the shaded area rotated from 0 = 0
1
to 0 = 0
2
.
Letting
\
.,
= volume of the region bounded by the sphere of radius a
.
and the cone with angle c
,
(0 = 0
1
to 0
2
)
and letting \ be the volume of the wedge, we have
\ = (\
22
3\
21
) 3(\
12
3\
11
)
=
1
3
(0
2
30
1
)
a
3
2
(1 3cos c
2
) 3a
3
2
(1 3cos c
1
) 3a
3
1
(1 3cos c
2
) +a
3
1
(1 3cos c
1
)
=
1
3
(0
2
30
1
)
a
3
2
3a
3
1
(1 3cos c
2
) 3
a
3
2
3a
3
1
(1 3cos c
1
)
=
1
3
(0
2
30
1
)
a
3
2
3a
3
1
(cos c
1
3cos c
2
)
Or: Show that \ =

0
2
0
1
p
2
sin c
2
p
1
sin c
1
r cot c
1
r cot c
2
v d: dv d0.
(c) By the Mean Value Theorem with 1(a) = a
3
there exists some a with a
1
$ a $ a
2
such that
1(a
2
) 31(a
1
) = 1
0
( a)(a
2
3a
1
) or a
3
1
3a
3
2
= 3 a
2
{a. Similarly there exists c with c
1
$

c $ c
2
such that cos c
2
3cos c
1
=

3sin

c
{c. Substituting into the result from (b) gives

{\ = ( a
2
{a)(0
2
30
1
)(sin

c) {c = a
2
sin

c{a {c{0.
16.9 Change of Variables in Multiple Integrals ET 15.9
1. r = 5& 3, j = & + 3.
The Jacobian is
u(r. j)
u(&. )
=
uru& uru
uju& uju
5 31
1 3
= 5(3) 3(31)(1) = 16.

3. r = c
3r
sin0, j = c
r
cos 0.
u(r. j)
u(v. 0)
=
uruv uru0
ujuv uju0
3c
3r
sin0 c
3r
cos 0
c
r
cos 0 3c
r
sin0
= c
3r
c
r
sin
2
0 3c
3r
c
r
cos
2
0 = sin
2
0 3cos
2
0 or 3cos 20
5. r = &, j = n, : = n&.
u(r. j. :)
u(&. . n)
=
uru& uru urun

uju& uju ujun
u:u& u:u u:un
1 3&
2
0
0 1n 3n
2
3n&
2
0 1&
=
1
1n 3n
2
0 1&
3
&
0 3n
2
3n&
2
1&
+ 0
0 1n
3n&
2
0
=
1
1
&n
30
+
&
0 3

&
2
n
+ 0 =
1
&n
3
1
&n
= 0
7. The transformation maps the boundary of o to the boundary of the image 1, so we rst look at side o
1
in the &-plane. o
1
is
described by = 0 [0 $ & $ 3], so r = 2& +3 = 2& and j = & 3 = &. Eliminating &, we have r = 2j, 0 $ r $ 6. o
2
is the line segment & = 3, 0 $ $ 2, so r = 6 + 3 and j = 3 3. Then = 3 3j i r = 6 + 3(3 3j) = 15 33j,
6 $ r $ 12. o
3
is the line segment = 2, 0 $ & $ 3, so r = 2& + 6 and j = & 32, giving & = j + 2 i r = 2j + 10,
6 $ r $ 12. Finally, o
4
is the segment & = 0, 0 $ $ 2, so r = 3 and j = 3 i r = 33j, 0 $ r $ 6. The image of
set o is the region 1 shown in the rj-plane, a parallelogram bounded by these four segments.
9. o
1
is the line segment & = , 0 $ & $ 1, so j = = & and r = &
2
= j
2
. Since 0 $ & $ 1, the image is the portion of the
parabola r = j
2
, 0 $ j $ 1. o
2
is the segment = 1, 0 $ & $ 1, thus j = = 1 and r = &
2
, so 0 $ r $ 1. The image is
the line segment j = 1, 0 $ r $ 1. o
3
is the segment & = 0, 0 $ $ 1, so r = &
2
= 0 and j = i 0 $ j $ 1. The
image is the segment r = 0, 0 $ j $ 1. Thus, the image of o is the region 1 in the rst quadrant bounded by the parabola
r = j
2
, the j-axis, and the line j = 1.
11.
u(r. j)
u(&. )
=
2 1
1 2
= 3 and r 33j = (2& +) 33(& + 2) = 3& 35. To nd the region o in the &-plane that
corresponds to 1 we rst nd the corresponding boundary under the given transformation. The line through (0. 0) and (2. 1) is
j =
1
2
r which is the image of & + 2 =
1
2
(2& +) i = 0; the line through (2. 1) and (1. 2) is r +j = 3 which is the
image of (2& +) + (& + 2) = 3 i & + = 1; the line through (0. 0) and (1. 2) is j = 2r which is the image of
& + 2 = 2(2& +) i & = 0. Thus o is the triangle 0 $ $ 1 3&, 0 $ & $ 1 in the &-plane and
T
(r 33j) d =

1
0
13u
0
(3& 35) |3| d d& = 33
1
0
& +
5
2
r=13u
r=0
d&
= 33
1
0
& 3&
2
+
5
2
(1 3&)
2
d& = 33
1
2
&
2
3
1
3
&
3
3
5
6
(1 3&)
3
1
0
= 33
1
2
3
1
3
+
5
6
= 33
13.
u(r. j)
u(&. )
=
2 0
0 3
= 6, r
2
= 4&
2
and the planar ellipse 9r
2
+ 4j
2
$ 36 is the image of the disk &
2
+
2
$ 1. Thus
T
r
2
d=

u
2
+r
2
$1
(4&
2
)(6) d&d =

2r
0
1
0
(24v
2
cos
2
0) v dv d0 = 24
2r
0
cos
2
0 d0
1
0
v
3
dv
= 24
1
2
r +
1
4
sin2r
2r
0
1
4
v
4
1
0
= 24()
1
4
= 6
SECTION 16.9 CHANGE OF VARIABLES IN MULTIPLE INTEGRALS ET SECTION 15.9

257
15.
u(r. j)
u(&. )
=
1 3&
2
0 1
=
1
, rj = &, j = r is the image of the parabola

2
= &, j = 3r is the image of the parabola
2
= 3&, and the hyperbolas rj = 1, rj = 3 are the images of the lines & = 1 and & = 3 respectively. Thus
T
rj d =

3
1
I
3u
I
u
&
d d& =

3
1
&
ln
I
3& 3ln
I
&
d& =

3
1
&ln
I
3 d& = 4 ln
I
3 = 2 ln3.
17. (a)
u(r. j. :)
u(&. . n)
=
o 0 0
0 / 0
0 0 c
= o/c and since & =

r
o
, =
j
/
, n =
:
c
the solid enclosed by the ellipsoid is the image of the
ball &
2
+
2
+n
2
$ 1. So
T
d\ =

u
2
+r
2
+u
2
$1
o/c d&d dn = (o/c)(volume of the ball) =
4
3
o/c
(b) If we approximate the surface of the earth by the ellipsoid
r
2
6378
2
+
j
2
6378
2
+
:
2
6356
2
= 1, then we can estimate
the volume of the earth by nding the volume of the solid 1 enclosed by the ellipsoid. From part (a), this is
T
d\ =
4
3
(6378)(6378)(6356) E 1.083 10
12
km
3
.
19. Letting & = r 32j and = 3r 3j, we have r =
1
5
(2 3&) and j =
1
5
( 33&). Then
u(r. j)
u(&. )
=
315 25
335 15
=
1
5
and 1 is the image of the rectangle enclosed by the lines & = 0, & = 4, = 1, and = 8. Thus
T
r 32j
3r 3j
d =

4
0
8
1
&
1
5
d d& =
1
5
4
0
&d&

8
1
1
d =
1
5
1
2
&
2
4
0
ln||
8
1
=
8
5
ln8.
21. Letting & = j 3r, = j +r, we have j =
1
2
(& +), r =
1
2
( 3&). Then
u(r. j)
u(&. )
=
312 12
12 12
= 3
1
2
and 1 is the
image of the trapezoidal region with vertices (31. 1), (32. 2), (2. 2), and (1. 1). Thus
T
cos
j 3r
j +r
d =

2
1
r
3r
cos
&
3
1
2
d&d =
1
2
2
1
sin
&
u=r
u=3r
d =
1
2
2
1
2 sin(1) d =
3
2
sin1
23. Let & = r +j and = 3r +j. Then & + = 2j i j =
1
2
(& +) and & 3 = 2r i r =
1
2
(& 3).
u(r. j)
u(&. )
=
12 312
12 12
=
1
2
. Now |&| = |r +j| $ |r| + |j| $ 1 i 31 $ & $ 1, and
|| = |3r +j| $ |r| + |j| $ 1 i 31 $ $ 1. 1 is the image of the square
region with vertices (1. 1), (1. 31), (31. 31), and (31. 1).
So
T
c
i+u
d =
1
2
1
31
1
31
c
u
d&d =
1
2
c
u
1
31
1
31
= c 3c
31
.
16 Review ET 15
1. (a) A double Riemann sum of 1 is
r
. =1
n
, =1
1
r
W
.,
. j
W
.,
{, where { is the area of each subrectangle and
r
W
.,
. j
W
.,
is a
sample point in each subrectangle. If 1(r. j) D 0, this sum represents an approximation to the volume of the solid that lies
above the rectangle 1 and below the graph of 1.
(b)

T
1(r. j) d = lim
r.n<"
r
. =1
n
, =1
1
r
W
.,
. j
W
.,
{
(c) If 1(r. j) D 0,
T
1(r. j) d represents the volume of the solid that lies above the rectangle 1 and below the surface
: = 1(r. j). If 1 takes on both positive and negative values,
T
1(r. j) d is the difference of the volume above 1 but
below the surface : = 1(r. j) and the volume below 1 but above the surface : = 1(r. j).
(d) We usually evaluate
T
1(r. j) d as an iterated integral according to Fubinis Theorem (see Theorem 16.2.4
[ET 15.2.4]).
(e) The Midpoint Rule for Double Integrals says that we approximate the double integral
T
1(r. j) d by the double
Riemann sum
r
. =1
n
, =1
1
r
.
. j
,
{ where the sample points
r
.
. j
,
are the centers of the subrectangles.

(f ) 1
ave
=
1
(1)
T
1(r. j) d where (1) is the area of 1.
2. (a) See (1) and (2) and the accompanying discussion in Section 16.3 [ET 15.3].
(b) See (3) and the accompanying discussion in Section 16.3 [ET 15.3].
(c) See (5) and the preceding discussion in Section 16.3 [ET 15.3].
(d) See (6)(11) in Section 16.3 [ET 15.3].
3. We may want to change from rectangular to polar coordinates in a double integral if the region 1 of integration is more easily
described in polar coordinates. To accomplish this, we use
T
1(r. j) d =

c
o
l
a
1(v cos 0. v sin0) v dv d0 where 1 is
given by 0 $ o $ v $ /, c $ 0 $ o.
4. (a) i =

T
a(r. j) d
(b) A
i
=

T
ja(r. j) d, A
u
=

T
ra(r. j) d
(c) The center of mass is (r. j) where r =
A
u
i
and j =
A
i
i
.
(d) 1
i
=

T
j
2
a(r. j) d, 1
u
=

T
r
2
a(r. j) d, 1
0
=

T
(r
2
+j
2
)a(r. j) d
5. (a) 1(o $ A $ /. c $ Y $ d) =
l
a
u
c
1(r. j) dj dr
(b) 1(r. j) D 0 and
R
2
1(r. j) d = 1.
(c) The expected value of A is j
1
=

R
2
r1(r. j) d; the expected value of Y is j
2
=

R
2
j1(r. j) d.
6. (a)

T
1(r. j. :) d\ = lim
I.r.n<"
I
. =1
r
, =1
n
I =1
1
r
W
.,I
. j
W
.,I
. :
W
.,I
{\
(b) We usually evaluate
T
1(r. j. :) d\ as an iterated integral according to Fubinis Theorem for Triple Integrals
(see Theorem 16.6.4 [ET 15.6.4]).
(c) See the paragraph following Example 16.6.1 [ET 15.6.1].
(d) See (5) and (6) and the accompanying discussion in Section 16.6 [ET 15.6].
(e) See (10) and the accompanying discussion in Section 16.6 [ET 15.6].
(f ) See (11) and the preceding discussion in Section 16.6 [ET 15.6].
7. (a) i =

T
a(r. j. :) d\
(b) A
u:
=

T
ra(r. j. :) d\ , A
i:
=

T
ja(r. j. :) d\ , A
iu
=

T
:a(r. j. :) d\ .
(c) The center of mass is (r. j. :) where r =
A
u:
i
, j =
A
i:
i
, and : =
A
iu
i
.
(d) 1
i
=

T
(j
2
+:
2
)a(r. j. :) d\ , 1
u
=

T
(r
2
+:
2
)a(r. j. :) d\ , 1
:
=

T
(r
2
+j
2
)a(r. j. :) d\ .
8. (a) See Formula 16.7.4 [ET 15.7.4] and the accompanying discussion.
(b) See Formula 16.8.3 [ET 15.8.3] and the accompanying discussion.
(c) We may want to change from rectangular to cylindrical or spherical coordinates in a triple integral if the region 1 of
integration is more easily described in cylindrical or spherical coordinates or if the triple integral is easier to evaluate using
cylindrical or spherical coordinates.
9. (a)
u (r. j)
u (&. )
=
uru& uru
uju& uju
=
ur
u&
uj
u
3
ur
u
uj
u&
(b) See (9) and the accompanying discussion in Section 16.9 [ET 15.9].
(c) See (13) and the accompanying discussion in Section 16.9 [ET 15.9].
1. This is true by Fubinis Theorem.
3. True by Equation 16.2.5 [ET 15.2.5].
5. True:

T
4 3r
2
3j
2
d = the volume under the surface r
2
+j
2
+:
2
= 4 and above the rj-plane
=
1
2
the volume of the sphere r

2
+j
2
+:
2
= 4
=
1
2

4
3
(2)
3
=
16
3

7. The volume enclosed by the cone : =

r
2
+j
2
and the plane : = 2 is, in cylindrical coordinates,
\ =

2r
0
2
0
2
r
v d: dv d0 6=

2r
0
2
0
2
r
d: dv d0, so the assertion is false.
1. As shown in the contour map, we divide 1 into 9 equally sized subsquares, each with area { = 1. Then we approximate
T
1(r. j) d by a Riemann sum with i = n = 3 and the sample points the upper right corners of each square, so
T
1(r. j) dE
3
. =1
3
, =1
1(r
.
. j
,
) {
= {[1(1. 1) +1(1. 2) +1(1. 3) +1(2. 1) +1(2. 2) +1(2. 3) +1(3. 1) +1(3. 2) +1(3. 3)]
Using the contour lines to estimate the function values, we have
T
1(r. j) d E 1[2.7 + 4.7 + 8.0 + 4.7 + 6.7 + 10.0 + 6.7 + 8.6 + 11.9] E 64.0
3.

2
1
2
0
(j + 2rc
u
) drdj =

2
1
rj +r
2
c
u
i=2
i=0
dj =

2
1
(2j + 4c
u
) dj =

j
2
+ 4c
u
2
1
= 4 + 4c
2
31 34c = 4c
2
34c + 3
5.

1
0
i
0
cos(r
2
) dj dr =

1
0
cos(r
2
)j
u=i
u=0
dr =

1
0
rcos(r
2
) dr =
1
2
sin(r
2
)
1
0
=
1
2
sin1
7.

r
0
1
0
I
13u
2
0
j sinrd: dj dr =

r
0
1
0
(j sinr):
:=
I
13u
2
:=0
dj dr =

r
0
1
0
j
1 3j
2
sinrdj dr
=

r
0
3
1
3
(1 3j
2
)
3'2
sinr
u=1
u=0
dr =

r
0
1
3
sinrdr = 3
1
3
cos r
r
0
=
2
3
9. The region 1 is more easily described by polar coordinates: 1 = {(v. 0) | 2 $ v $ 4, 0 $ 0 $ }. Thus
T
1(r. j) d =

r
0
4
2
1(v cos 0. v sin0) v dv d0.
11. The region whose area is given by
r'2
0
sin 20
0
v dv d0 is
(v. 0) | 0 $ 0 $
r
2
. 0 $ v $ sin20
, which is the region contained in the

loop in the rst quadrant of the four-leaved rose v = sin20.
13.

1
0
1
i
cos(j
2
) dj dr =

1
0
u
0
cos(j
2
) drdj
=

1
0
cos(j
2
)
i=u
i=0
dj =

1
0
j cos(j
2
) dj
=

1
2
sin(j
2
)
1
0
=
1
2
sin1
15.

T
jc
iu
d =

3
0
2
0
jc
iu
drdj =

3
0
c
iu
i=2
i=0
dj =

3
0
(c
2u
31) dj =

1
2
c
2u
3j
3
0
=
1
2
c
6
33 3
1
2
=
1
2
c
6
3
7
2
17.

T
j
1 +r
2
d =

1
0
I
i
0
j
1 +r
2
dj dr =

1
0
1
1 +r
2
1
2
j
2
u=
I
i
u=0
dr
=
1
2
1
0
r
1 +r
2
dr =

1
4
ln(1 +r
2
)
1
0
=
1
4
ln2
19.

T
j d =

2
0
83u
2
u
2
j drdj
=

2
0
j
i=83u
2
i=u
2
dj =

2
0
j(8 3j
2
3j
2
) dj
=

2
0
(8j 32j
3
) dj =

4j
2
3
1
2
j
4
2
0
= 8
21.

T
r
2
+j
2
3'2
d =

r'3
0
3
0
(v
2
)
3'2
v dv d0
=

r'3
0
d0

3
0
v
4
dv =

0
r'3
0
1
5
v
5
3
0
=

3
3
5
5
=
81
5
23.

T
rj d\ =

3
0
i
0
i+u
0
rj d: dj dr =

3
0
i
0
rj
:=i+u
:=0
dj dr =

3
0
i
0
rj(r +j) dj dr
=

3
0
i
0
(r
2
j +rj
2
) dj dr =

3
0
1
2
r
2
j
2
+
1
3
rj
3
u=i
u=0
dr =

3
0
1
2
r
4
+
1
3
r
4
dr
=
5
6
3
0
r
4
dr =

1
6
r
5
3
0
=
81
2
= 40.5
25.

T
j
2
:
2
d\ =

1
31
I
13u
2
3
I
13u
2
1 3u
2
3:
2
0
j
2
:
2
drd: dj =

1
31
I
13u
2
3
I
13u
2
j
2
:
2
(1 3j
2
3:
2
) d: dj
=

2r
0
1
0
(v
2
cos
2
0)(v
2
sin
2
0)(1 3v
2
) v dv d0 =

2r
0
1
0
1
4
sin
2
20(v
5
3v
7
) dv d0
=

2r
0
1
8
(1 3cos 40)
1
6
v
6
3
1
8
v
8
r=1
r=0
d0 =
1
192
0 3
1
4
sin40
2r
0
=
2r
192
=
r
96
27.

T
j: d\ =

2
32
I
43i
2
0
u
0
j: d: dj dr =

2
32
I
43i
2
0
1
2
j
3
dj dr =

r
0
2
0
1
2
v
3
(sin
3
0) v dv d0
=
16
5
r
0
sin
3
0 d0 =
16
5
3cos 0 +
1
3
cos
3
0
r
0
=
64
15
29. \ =

2
0
4
1
(r
2
+ 4j
2
) dj dr =

2
0
r
2
j +
4
3
j
3
u=4
u=1
dr =

2
0
(3r
2
+ 84) dr = 176
31.
\ =

2
0
u
0
(23u)'2
0
d: drdj =

2
0
u
0
1 3
1
2
j
drdj
=

2
0
j 3
1
2
j
2
dj =
2
3
33. Using the wedge above the plane : = 0 and below the plane : = ir and noting that we have the same volume for i < 0 as
for i 0 (so use i 0), we have
\ = 2
a'3
0
I
a
2
39u
2
0
irdrdj = 2
a'3
0
1
2
i(o
2
39j
2
) dj = i
o
2
j 33j
3
a'3
0
= i
1
3
o
3
3
1
9
o
3
=
2
9
io
3
.
35. (a) i =

1
0
13u
2
0
j drdj =

1
0
(j 3j
3
) dj =
1
2
3
1
4
=
1
4
(b) A
u
=

1
0
1 3u
2
0
rj drdj =

1
0
1
2
j(1 3 j
2
)
2
dj = 3
1
12
(1 3j
2
)
3
1
0
=
1
12
,
A
i
=

1
0
1 3u
2
0
j
2
drdj =

1
0
(j
2
3j
4
) dj =
2
15
. Hence (r. j) =
1
3
.
8
15
.
(c) 1
i
=

1
0
13u
2
0
j
3
drdj =

1
0
(j
3
3 j
5
) dj =
1
12
,
1
u
=

1
0
13u
2
0
jr
2
drdj =

1
0
1
3
j(1 3j
2
)
3
dj = 3
1
24
(1 3j
2
)
4
1
0
=
1
24
,
1
0
= 1
i
+1
u
=
1
8
, j
2
=
1'12
1'4
=
1
3
i j =
1
I
3
, and r
2
=
1'24
1'4
=
1
6
i r =
1
I
6
.
37. The equation of the cone with the suggested orientation is (/ 3:) =
I
a
r
2
+j
2
, 0 $ : $ /. Then \ =
1
3
o
2
/ is the
volume of one frustum of a cone; by symmetry A
u:
= A
i:
= 0; and
A
iu
=

i
2
+u
2
$a
2
I3(I'a)
I
i
2
+u
2
0
: d: d =

2r
0
a
0
(I'a)(a3r)
0
v: d: dv d0 =
a
0
v
/
2
o
2
(o 3v)
2
dv
=
/
2
o
2
a
0
(o
2
v 32ov
2
+v
3
) dv =
/
2
o
2
o
4
2
3
2o
4
3
+
o
4
4
=
/
2
o
2
12
Hence the centroid is (r. j. :) =
0. 0.
1
4
/
.
39.

3
0
I
93i
2
3
I
93i
2
(r
3
+rj
2
) dj dr =

3
0
I
93i
2
3
I
93i
2
r(r
2
+j
2
) dj dr
=

r'2
3r'2
3
0
(v cos 0)(v
2
) v dv d0
=

r'2
3r'2
cos 0 d0

3
0
v
4
dv
=

sin0
r'2
3r'2
1
5
v
5
3
0
= 2
1
5
(243) =
486
5
= 97.2
41. From the graph, it appears that 1 3r
2
= c
i
at r E 30.71 and at
r = 0, with 1 3r
2
c
i
on (30.71. 0). So the desired integral is
T
j
2
dE
0
30.71
13i
2
c
{
j
2
dj dr
=
1
3
0
30.71
[(1 3r
2
)
3
3c
3i
] dr
=
1
3
r 3r
3
+
3
5
r
5
3
1
7
r
7
3
1
3
c
3i
0
30.71
E 0.0512
43. (a) 1(r. j) is a joint density function, so we know that
R
2
1(r. j) d = 1. Since 1(r. j) = 0 outside the rectangle
[0. 3] [0. 2], we can say
R
2
1(r. j) d=

"
3"
"
3"
1(r. j) dj dr =

3
0
2
0
C(r +j) dj dr
= C
3
0
rj +
1
2
j
2
u=2
u=0
dr = C
3
0
(2r + 2) dr = C
r
2
+ 2r
3
0
= 15C
Then 15C = 1 i C =
1
15
.
(b) 1(A $ 2. Y D 1) =

2
3"
"
1
1(r. j) dj dr =

2
0
2
1
1
15
(r. j) dj dr =
1
15
2
0
rj +
1
2
j
2
u=2
u=1
dr
=
1
15
2
0
r +
3
2
dr =
1
15
1
2
r
2
+
3
2
r
2
0
=
1
3
(c) 1(A +Y $ 1) = 1((A. Y ) M 1) where 1 is the triangular region shown in
the gure. Thus
1(A +Y $ 1) =
T
1(r. j) d =

1
0
13i
0
1
15
(r +j) dj dr
=
1
15
1
0
rj +
1
2
j
2
u=13i
u=0
dr
=
1
15
1
0
r(1 3r) +
1
2
(1 3r)
2
dr
=
1
30
1
0
(1 3r
2
) dr =
1
30
r 3
1
3
r
3
1
0
=
1
45
45.

1
31
1
i
2
13u
0
1(r. j. :) d: dj dr =

1
0
13:
0
I
u
3
I
u
1(r. j. :) drdj d:
47. Since & = r 3j and = r +j, r =
1
2
(& +) and j =
1
2
( 3&).
Thus
u(r. j)
u(&. )
=
12 12
312 12
=
1
2
and
T
r 3j
r +j
d =

4
2
0
32
&
1
2
d&d = 3
4
2
d
= 3ln2.
49. Let & = j 3r and = j +r so r = j 3& = ( 3r) 3& i r =
1
2
( 3&) and j = 3
1
2
( 3&) =
1
2
( +&).
u(r. j)
u(&. )
ur
u&
uj
u
3
ur
u
uj
u&
3
1
2
1
2
3
1
2
1
2
3
1
2
=
1
2
. 1 is the image under this transformation of the square
with vertices (&. ) = (0. 0), (32. 0), (0. 2), and (32. 2). So
T
rj d =

2
0
0
32
2
3&
2
4
1
2
d&d =
1
8
2
0
2
& 3
1
3
&
3
u=0
u=32
d =
1
8
2
0
2
2
3
8
3
d =
1
8
2
3
3
3
8
3
2
0
= 0
This result could have been anticipated by symmetry, since the integrand is an odd function of j and 1 is symmetric about
the r-axis.
51. For each v such that 1
r
lies within the domain, (1
r
) = v
2
, and by the Mean Value Theorem for Double Integrals there
exists (r
r
. j
r
) in 1
r
such that 1 (r
r
. j
r
) =
1
v
2
T
u
1(r. j) d. But lim
r<0
+
(r
r
. j
r
) = (o. /),
so lim
r<0
+
1
v
2
T
u
1 (r. j) d = lim
r<0
+
1(r
r
. j
r
) = 1(o. /) by the continuity of 1.
PROBLEMS PLUS
1. Let 1 =

5
. =1
1
.
, where
1
.
= {(r. j) | r +j D j + 2. r +j < j + 3. 1 $ r $ 3. 2 $ j $ 5}.
T
[[r +j]] d =
5
. =1
T
l
[[r +j]] d =
5
.=1
[[r +j]]
T
l
d, since
[[r +j]] = constant = j + 2 for (r. j) M 1
.
. Therefore
T
[[r +j]] d=

5
.=1
(j + 2) [(1
.
)]
= 3(1
1
) + 4(1
2
) + 5(1
3
) + 6(1
4
) + 7(1
5
)
= 3
1
2
+ 4
3
2
+ 5(2) + 6
3
2
+ 7
1
2
= 30
3. 1
ave
=
1
/ 3o
l
a
1(r) dr =
1
1 30
1
0
1
i
cos(t
2
) dt
dr
=

1
0
1
i
cos(t
2
) dt dr =

1
0
I
0
cos(t
2
) drdt [changing the order of integration]
=

1
0
t cos(t
2
) dt =
1
2
sin
t
2
1
0
=
1
2
sin1
5. Since |rj| < 1, except at (1. 1), the formula for the sum of a geometric series gives
1
1 3rj
=
"
n=0
(rj)
n
, so
1
0
1
0
1
13iu
drdj =

1
0
1
0
"
n=0
(rj)
n
drdj =
"
n=0
1
0
1
0
(rj)
n
drdj =
"
n=0
1
0
r
n
dr
1
0
j
n
dj
=
"
n=0
1
n+1

1
n+1
=
"
n=0
1
(n+1)
2
=
1
1
2
+
1
2
2
+
1
3
2
+ =

"
n=1
1
n
2
7. (a) Since |rj:| < 1 except at (1. 1. 1), the formula for the sum of a geometric series gives
1
1 3rj:
=
"
n=0
(rj:)
n
, so
1
0
1
0
1
0
1
1 3rj:
drdj d: =

1
0
1
0
1
0
"
n=0
(rj:)
n
drdj d: =
"
n=0
1
0
1
0
1
0
(rj:)
n
drdj d:
=
"
n=0
1
0
r
n
dr
1
0
j
n
dj
1
0
:
n
d:
=
"
n=0
1
n + 1

1
n + 1

1
n + 1
=
"
n=0
1
(n + 1)
3
=
1
1
3
+
1
2
3
+
1
3
3
+ =
"
n=1
1
n
3
(b) Since |3rj:| < 1, except at (1. 1. 1), the formula for the sum of a geometric series gives
1
1 +rj:
=
"
n=0
(3rj:)
n
, so
1
0
1
0
1
0
1
1 +rj:
drdj d: =

1
0
1
0
1
0
1
1 +rj:
"
n=0
(3rj:)
n
drdj d: =
"
n=0
1
0
1
0
1
0
(3rj:)
n
drdj d:
=
"
n=0
(31)
n
1
0
r
n
dr
1
0
j
n
dj
1
0
:
n
d:
=
"
n=0
(31)
n
1
n + 1

1
n + 1

1
n + 1
=
"
n=0
(31)
n
(n + 1)
3
=
1
1
3
3
1
2
3
+
1
3
3
3 =
"
n=0
(31)
n31
n
3
To evaluate this sum, we rst write out a few terms: c = 1 3
1
2
3
+
1
3
3
3
1
4
3
+
1
5
3
3
1
6
3
E 0.8998. Notice that
265
266

PROBLEMS PLUS
o
7
=
1
7
3
< 0.003. By the Alternating Series Estimation Theorem from Section 12.5 [ ET 11.5], we have
|c 3c
6
| $ o
7
< 0.003. This error of 0.003 will not affect the second decimal place, so we have c E 0.90.
9. (a) r = v cos 0, j = v sin0, : = :. Then
u&
uv
=
u&
ur
ur
uv
+
u&
uj
uj
uv
+
u&
u:
u:
uv
=
u&
ur
cos 0 +
u&
uj
sin0 and
u
2
&
uv
2
= cos 0
u
2
&
ur
2
ur
uv
+
u
2
&
uj ur
uj
uv
+
u
2
&
u: ur
u:
uv
+ sin0
u
2
&
uj
2
uj
uv
+
u
2
&
uruj
ur
uv
+
u
2
&
u: uj
u:
uv
=
u
2
&
ur
2
cos
2
0 +
u
2
&
uj
2
sin
2
0 + 2
u
2
&
uj ur
cos 0 sin0
Similarly
u&
u0
= 3
u&
ur
v sin0 +
u&
uj
v cos 0 and
u
2
&
u0
2
=
u
2
&
ur
2
v
2
sin
2
0 +
u
2
&
uj
2
v
2
cos
2
0 32
u
2
&
uj ur
v
2
sin0 cos 0 3
u&
ur
v cos 0 3
u&
uj
v sin0. So
u
2
&
uv
2
+
1
v
u&
uv
+
1
v
2
u
2
&
u0
2
+
u
2
&
u:
2
=
u
2
&
ur
2
cos
2
0 +
u
2
&
uj
2
sin
2
0 + 2
u
2
&
uj ur
cos 0 sin0 +
u&
ur
cos 0
v
+
u&
uj
sin0
v
+
u
2
&
ur
2
sin
2
0 +
u
2
&
uj
2
cos
2
0 32
u
2
&
uj ur
sin0 cos 0
3
u&
ur
cos 0
v
3
u&
uj
sin0
v
+
u
2
&
u:
2
=
u
2
&
ur
2
+
u
2
&
uj
2
+
u
2
&
u:
2
(b) r = a sinccos 0, j = a sincsin0, : = a cos c. Then
u&
ua
=
u&
ur
ur
ua
+
u&
uj
uj
ua
+
u&
u:
u:
ua
=
u&
ur
sinc cos 0 +
u&
uj
sinc sin0 +
u&
u:
cos c, and
u
2
&
ua
2
= sinccos 0
u
2
&
ur
2
ur
ua
+
u
2
&
uj ur
uj
ua
+
u
2
&
u: ur
u:
ua
+ sincsin0
u
2
&
uj
2
uj
ua
+
u
2
&
uruj
ur
ua
+
u
2
&
u: uj
u:
ua
+ cos c
u
2
&
u:
2
u:
ua
+
u
2
&
uru:
ur
ua
+
u
2
&
uj u:
uj
ua
= 2
u
2
&
uj ur
sin
2
c sin0 cos 0 + 2
u
2
&
u: ur
sinc cos c cos 0 + 2
u
2
&
uj u:
sinc cos c sin0
+
u
2
&
ur
2
sin
2
ccos
2
0 +
u
2
&
uj
2
sin
2
csin
2
0 +
u
2
&
u:
2
cos
2
c
Similarly
u&
uc
=
u&
ur
a cos ccos 0 +
u&
uj
a cos csin0 3
u&
u:
a sinc, and
u
2
&
uc
2
= 2
u
2
&
uj ur
a
2
cos
2
c sin0 cos 0 32
u
2
&
uru:
a
2
sinc cos c cos 0
3 2
u
2
&
uj u:
a
2
sinc cos c sin0 +
u
2
&
ur
2
a
2
cos
2
c cos
2
0 +
u
2
&
uj
2
a
2
cos
2
c sin
2
0
+
u
2
&
u:
2
a
2
sin
2
c 3
u&
ur
a sinc cos 0 3
u&
uj
a sinc sin0 3
u&
u:
a cos c
PROBLEMS PLUS

267
And
u&
u0
= 3
u&
ur
a sincsin0 +
u&
uj
a sinccos 0, while
u
2
&
u0
2
= 32
u
2
&
uj ur
a
2
sin
2
c cos 0 sin0 +
u
2
&
ur
2
a
2
sin
2
c sin
2
0
+
u
2
&
uj
2
a
2
sin
2
ccos
2
0 3
u&
ur
a sinc cos 0 3
u&
uj
a sinc sin0
Therefore
u
2
&
ua
2
+
2
a
u&
ua
+
cot c
a
2
u&
uc
+
1
a
2
u
2
&
uc
2
+
1
a
2
sin
2
c
u
2
&
u0
2
=
u
2
&
ur
2
(sin
2
c cos
2
0) + (cos
2
c cos
2
0) + sin
2
0
+
u
2
&
uj
2
(sin
2
c sin
2
0) + (cos
2
c sin
2
0) + cos
2
0
+
u
2
&
u:
2
cos
2
c + sin
2
c
+
u&
ur
2 sin
2
c cos 0 + cos
2
c cos 0 3sin
2
c cos 0 3cos 0
a sinc
+
u&
uj
2 sin
2
c sin0 + cos
2
c sin0 3sin
2
c sin0 3sin0
a sinc
But 2 sin
2
ccos 0 +cos
2
ccos 0 3sin
2
ccos 0 3cos 0 = (sin
2
c+cos
2
c31) cos 0 = 0 and similarly the coefcient of
u&uj is 0. Also sin
2
ccos
2
0 + cos
2
ccos
2
0 + sin
2
0 = cos
2
0 (sin
2
c + cos
2
c) + sin
2
0 = 1, and similarly the
coefcient of u
2
&uj
2
is 1. So Laplaces Equation in spherical coordinates is as stated.
11.

i
0
u
0
:
0
1 (t) dt d: dj =

T
1 (t) d\ , where
1 = {(t. :. j) | 0 $ t $ :, 0 $ : $ j, 0 $ j $ r}.
If we let 1 be the projection of 1 on the jt-plane then
1 = {(j. t) | 0 $ t $ r, t $ j $ r}. And we see from the diagram
that 1 = {(t. :. j) | t $ : $ j, t $ j $ r, 0 $ t $ r}. So
i
0
u
0
:
0
1(t) dt d: dj =

i
0
i
I
u
I
1(t) d: dj dt =

i
0
i
I
(j 3t) 1(t) dj
dt
=

i
0
1
2
j
2
3tj
1(t)
u =i
u =I
dt =

i
0
1
2
r
2
3tr 3
1
2
t
2
+t
2
1(t) dt
=

i
0
1
2
r
2
3tr +
1
2
t
2
1(t) dt =

i
0
1
2
r
2
32tr +t
2
1(t) dt
=
1
2
i
0
(r 3t)
2
1(t) dt
17 VECTOR CALCULUS ET 16
17.1 Vector Fields ET 16.1
1. F(r. j) =
1
2
(i +j)
All vectors in this eld are identical, with length
1
I
2
and
direction parallel to the line j = r.
3. F(r. j) = j i +
1
2
j
The length of the vector j i +
1
2
j is
j
2
+
1
4
. Vectors
are tangent to parabolas opening about the r-axis.
5. F(r. j) =
j i +rj
r
2
+j
2
The length of the vector
j i +rj
r
2
+j
2
is 1.
7. F(r. j. :) = k
All vectors in this eld are parallel to the :-axis and have
length 1.
9. F(r. j. :) = rk
At each point (r. j. :), F(r. j. :) is a vector of length |r|.
For r 0, all point in the direction of the positive :-axis,
while for r < 0, all are in the direction of the negative
:-axis. In each plane r = I, all the vectors are identical.
269
270 CHAPTER 17 VECTOR CALCULUS ET CHAPTER 16
11. F(r. j) = hj. ri corresponds to graph II. In the rst quadrant all the vectors have positive r- and j-components, in the second
quadrant all vectors have positive r-components and negative j-components, in the third quadrant all vectors have negative r-
and j-components, and in the fourth quadrant all vectors have negative r-components and positive j-components. In addition,
the vectors get shorter as we approach the origin.
13. F(r. j) = hr 32. r + 1i corresponds to graph I since the vectors are independent of j (vectors along vertical lines are
identical) and, as we move to the right, both the r- and the j-components get larger.
15. F(r. j. :) = i + 2 j + 3 k corresponds to graph IV, since all vectors have identical length and direction.
17. F(r. j. :) = ri +j j + 3 k corresponds to graph III; the projection of each vector onto the rj-plane is ri +j j, which points
away from the origin, and the vectors point generally upward because their :-components are all 3.
19.
The vector eld seems to have very short vectors near the line j = 2r.
For F(r. j) = h0. 0i we must have j
2
32rj = 0 and 3rj 36r
2
= 0.
The rst equation holds if j = 0 or j = 2r, and the second holds if
r = 0 or j = 2r. So both equations hold [and thus F(r. j) = 0] along
the line j = 2r.
21. 1(r. j) = rc
iu
i
Q1(r. j) = 1
i
(r. j) i +1
u
(r. j) j = (rc
iu
j +c
iu
) i + (rc
iu
r) j = (rj + 1)c
iu
i +r
2
c
iu
j
23. Q1(r. j. :) = 1
i
(r. j. :) i +1
u
(r. j. :) j +1
:
(r. j. :) k =
r
r
2
+j
2
+:
2
i +
j
r
2
+j
2
+:
2
j +
:
r
2
+j
2
+:
2
k
25. 1(r. j) = r
2
3j i Q1(r. j) = 2ri 3j.
The length of Q1(r. j) is
I
4r
2
+ 1. When r 6= 0, the
vectors point away from the j-axis in a slightly downward
direction with length that increases as the distance from
the j-axis increases.
27. We graph Q1 along with a contour map of 1.
The graph shows that the gradient vectors are
perpendicular to the level curves. Also, the gradient
vectors point in the direction in which 1 is increasing and
are longer where the level curves are closer together.
SECTION 17.2 LINE INTEGRALS ET SECTION 16.2

271
29. 1(r. j) = r
2
+j
2
i Q1(r. j) = 2ri + 2j j. Thus, each vector Q1(r. j) has the same direction and twice the length of
the position vector of the point (r. j), so the vectors all point directly away from the origin and their lengths increase as we
move away from the origin. Hence, Q1 is graph II.
31. 1(r. j) = (r +j)
2
i Q1(r. j) = 2(r +j) i + 2(r +j) j. The r- and j-components of each vector are equal, so all
vectors are parallel to the line j = r. The vectors are 0 along the line j = 3r and their length increases as the distance from
this line increases. Thus, Q1 is graph II.
33. At t = 3 the particle is at (2. 1) so its velocity is V(2. 1) = h4. 3i. After 0.01 units of time, the particles change in
location should be approximately 0.01 V(2. 1) = 0.01 h4. 3i = h0.04. 0.03i, so the particle should be approximately at the
point (2.04. 1.03).
35. (a) We sketch the vector eld F(r. j) = ri 3j j along with
several approximate ow lines.The ow lines appear to be
hyperbolas with shape similar to the graph of j = 1r,
so we might guess that the ow lines have equations
j = Cr.
(b) If r = r(t) and j = j(t) are parametric equations of a ow line, then the velocity vector of the ow line at the
point (r. j) is r
0
(t) i +j
0
(t) j. Since the velocity vectors coincide with the vectors in the vector eld, we have
r
0
(t) i +j
0
(t) j = ri 3j j i drdt = r, djdt = 3j. To solve these differential equations, we know
drdt = r i drr = dt i ln|r| = t +C i r = c
I +C
= c
I
for some constant , and
djdt = 3j i djj = 3dt i ln|j| = 3t +1 i j = c
3I +1
= 1c
3I
for some constant 1. Therefore
rj = c
I
1c
3I
= 1 = constant. If the ow line passes through (1. 1) then (1) (1) = constant = 1 i rj = 1 i
j = 1r, r 0.
17.2 Line Integrals ET 16.2
1. r = t
3
and j = t, 0 $ t $ 2, so by Formula 3
C
j
3
dc =

2
0
t
3
ui
uI
2
+
uu
uI
2
dt =

2
0
t
3
(3t
2
)
2
+ (1)
2
dt =

2
0
t
3
I
9t
4
+ 1 dt
=
1
36

2
3
9t
4
+ 1
3'2
2
0
=
1
54
(145
3'2
31) or
1
54
145
I
145 31
3. Parametric equations for C are r = 4 cos t, j = 4 sint, 3

r
2
$ t $
r
2
. Then
C
rj
4
dc =

r'2
3r'2
(4 cos t)(4 sint)
4
(34 sint)
2
+ (4 cos t)
2
dt =

r'2
3r'2
4
5
cos t sin
4
t
16(sin
2
t + cos
2
t) dt
= 4
5
r'2
3r'2
(sin
4
t cos t)(4) dt = (4)
6
1
5
sin
5
t
r'2
3r'2
=
2 4
6
5
= 1638.4
5. If we choose r as the parameter, parametric equations for C are r = r, j =
I
r for 1 $ r $ 4 and
r
2
j
3
3
I
r
dj =

4
1
r
2
(
I
r)
3
3
I
r
1
2
I
r
dr =
1
2
4
1
r
3
31
dr
=
1
2
1
4
r
4
3r
4
1
=
1
2
64 34 3
1
4
+ 1
=
243
8
7. C = C
1
+C
2
On C
1
: r = r, j = 0 i dj = 0 dr, 0 $ r $ 2.
On C
2
: r = r, j = 2r 34 i dj = 2 dr, 2 $ r $ 3.
Then
C
rj dr + (r 3j) dj =

C
1
rj dr + (r 3j) dj +
C
2
rj dr + (r 3j) dj
=

2
0
(0 + 0) dr +
3
2
[(2r
2
34r) + (3r + 4)(2)] dr
=

3
2
(2r
2
36r + 8) dr =
17
3
9. r = 2 sint, j = t, : = 32 cos t, 0 $ t $ . Then by Formula 9,
C
rj: dc =
r
0
(2 sint)(t)(32 cos t)
ui
uI
2
+
uu
uI
2
+
u:
uI
2
dt
=

r
0
34t sint cos t
(2 cos t)
2
+ (1)
2
+ (2 sint)
2
dt =

r
0
32t sin2t
4(cos
2
t + sin
2
t) + 1 dt
= 32
I
5
r
0
t sin2t dt = 32
I
5
3
1
2
t cos 2t +
1
4
sin2t
r
0
integrate by parts with

u = |, Ju = sin 2| J|
= 32
I
5
3
r
2
30
=
I
5
11. Parametric equations for C are r = t, j = 2t, : = 3t, 0 $ t $ 1. Then
C
rc
u:
dc =

1
0
tc
(2I)(3I)
I
1
2
+ 2
2
+ 3
2
dt =
I
14
1
0
tc
6I
2
dt =
I
14
1
12
c
6I
2
1
0
=
I
14
12
(c
6
31).
13.

C
r
2
j
I
: d: =

1
0
(t
3
)
2
(t)
I
t
2
2t dt =

1
0
2t
9
dt =
1
5
t
10
1
0
=
1
5
15. On C
1
: r = 1 +t i dr = dt, j = 3t i
dj = 3 dt, : = 1 i d: = 0 dt, 0 $ t $ 1.
On C
2
: r = 2 i dr = 0 dt, j = 3 + 2t i
dj = 2 dt, : = 1 +t i d: = dt, 0 $ t $ 1.
Then
C
(r +j:) dr + 2rdj +rj: d:
=

C
1
(r +j:) dr + 2rdj +rj: d: +
C
2
(r +j:) dr + 2rdj +rj: d:
=

1
0
(1 +t + (3t)(1)) dt + 2(1 +t) 3 dt + (1 +t)(3t)(1) 0 dt
+

1
0
(2 + (3 + 2t)(1 +t)) 0 dt + 2(2) 2 dt + (2)(3 + 2t)(1 +t) dt
=

1
0
(10t + 7) dt +
1
0
(4t
2
+ 10t + 14) dt =

5t
2
+ 7t
1
0
+
4
3
t
3
+ 5t
2
+ 14t
1
0
= 12 +
61
3
=
97
3

273
17. (a) Along the line r = 33, the vectors of F have positive j-components, so since the path goes upward, the integrand F Tis
always positive. Therefore
C
1
F dr =

C
1
F Tdc is positive.
(b) All of the (nonzero) eld vectors along the circle with radius 3 are pointed in the clockwise direction, that is, opposite the
direction to the path. So F Tis negative, and therefore
C
2
F dr =

C
2
F Tdc is negative.
19. r(t) = 11t
4
i +t
3
j, so F(r(t)) = (11t
4
)(t
3
) i + 3(t
3
)
2
j = 11t
7
i + 3t
6
j and r
0
(t) = 44t
3
i + 3t
2
j. Then
C
F dr =

1
0
F(r(t)) r
0
(t) dt =

1
0
(11t
7
44t
3
+ 3t
6
3t
2
) dt =

1
0
(484t
10
+ 9t
8
) dt =

44t
11
+t
9
1
0
= 45.
21.

C
F dr =

1
0
sint
3
. cos(3t
2
). t
4
3t
2
. 32t. 1
dt
=

1
0
(3t
2
sint
3
32t cos t
2
+t
4
) dt =

3cos t
3
3sint
2
+
1
5
t
5
1
0
=
6
5
3cos 1 3sin1
23. F(r(t)) = (c
I
)
c
3I
2
i + sin
c
3I
2
j = c
I3I
2
i + sin
c
3I
2
j, r
0
(t) = c
I
i 32tc
3I
2
j. Then
C
F dr =

2
1
F(r(t)) r
0
(t) dt =

2
1
c
I3I
2
c
I
+ sin
c
3I
2
32tc
3I
2
dt
=

2
1
c
2I3I
2
32tc
3I
2
sin
c
3I
2
dt E 1.9633
25. r = t
2
, j = t
3
, : = t
4
so by Formula 9,
C
rsin(j +:) dc =

5
0
(t
2
) sin(t
3
+t
4
)
(2t)
2
+ (3t
2
)
2
+ (4t
3
)
2
dt
=

5
0
t
2
sin(t
3
+t
4
)
I
4t
2
+ 9t
4
+ 16t
6
dt E 15.0074
27. We graph F(r. j) = (r 3j) i +rj j and the curve C. We see that most of the vectors starting on C point in roughly the same
direction as C, so for these portions of C the tangential component F Tis positive. Although some vectors in the third
quadrant which start on C point in roughly the opposite direction, and hence give negative tangential components, it seems
reasonable that the effect of these portions of C is outweighed by the positive tangential components. Thus, we would expect
C
F dr =

C
F Tdc to be positive.
To verify, we evaluate
C
F dr. The curve C can be represented by r(t) = 2 cos t i + 2 sint j, 0 $ t $
3r
2
,
so F(r(t)) = (2 cos t 32 sint) i + 4 cos t sint j and r
0
(t) = 32 sint i + 2 cos t j. Then
C
F dr =

3r'2
0
F(r(t)) r
0
(t) dt
=

3r'2
0
[32 sint(2 cos t 32 sint) + 2 cos t(4 cos t sint)] dt
= 4
3r'2
0
(sin
2
t 3sint cos t + 2 sint cos
2
t) dt
= 3 +
2
3
[using a CAS]
29. (a)

C
F dr =

1
0
c
I
2
31
. t
5
2t. 3t
2
dt =

1
0
2tc
I
2
31
+ 3t
7
dt =

c
I
2
31
+
3
8
t
8
1
0
=
11
8
31c
(b) r(0) = 0, F(r(0)) =
c
31
. 0
;
r
1
I
2
=

1
2
.
1
2
I
2
, F
1
I
2
=

c
31'2
.
1
4
I
2
;
r(1) = h1. 1i, F(r(1)) = h1. 1i.
In order to generate the graph with Maple, we use the PLOT command
(not to be confused with the plot command) to dene each of the vectors.
For example,
v1:=PLOT(CURVES([[0,0],[evalf(1/exp(1)),0]]));
generates the vector from the vector eld at the point (0. 0) (but without an arrowhead) and gives it the name v1. To show
everything on the same screen, we use the display command. In Mathematica, we use ListPlot (with the
PlotJoined - True option) to generate the vectors, and then Show to show everything on the same screen.
31. r = c
3I
cos 4t, j = c
3I
sin4t, : = c
3I
, 0 $ t $ 2 .
Then
dr
dt
= c
3I
(3sin4t)(4) 3 c
3I
cos 4t = 3c
3I
(4 sin4t + cos 4t),
dj
dt
= c
3I
(cos 4t)(4) 3c
3I
sin4t = 3c
3I
(34 cos 4t + sin4t), and
d:
dt
= 3c
3I
, so
dr
dt
2
+
dj
dt
2
+
d:
dt
2
=

(3c
3I
)
2
[(4 sin4t + cos 4t)
2
+ (34 cos 4t + sin4t)
2
+ 1]
= c
3I
16(sin
2
4t + cos
2
4t) + sin
2
4t + cos
2
4t + 1 = 3
I
2 c
3I
Therefore

C
r
3
j
2
: dc =

2r
0
(c
3I
cos 4t)
3
(c
3I
sin4t)
2
(c
3I
) (3
I
2 c
3I
) dt
=

2r
0
3
I
2 c
37I
cos
3
4t sin
2
4t dt =
172,704
5,632,705
I
2 (1 3c
314r
)
33. We use the parametrization r = 2 cos t, j = 2 sint, 3
r
2
$ t $
r
2
. Then
dc =

ui
uI
2
+
uu
uI
2
dt =

(32 sint)
2
+ (2 cos t)
2
dt = 2 dt, so i =

C
I dc = 2I
r'2
3r'2
dt = 2I(),
r =
1
2rI
C
rI dc =
1
2r
r'2
3r'2
(2 cos t)2 dt =
1
2r
4 sint
r'2
3r'2
=
4
r
, j =
1
2rI
C
jI dc =
1
2r
r'2
3r'2
(2 sint)2 dt = 0.
Hence (r. j) =
4
r
. 0
.
35. (a) r =
1
i
C
ra(r. j. :) dc , j =
1
i
C
ja(r. j. :) dc, : =
1
i
C
:a(r. j. :) dc where i =

C
a(r. j. :) dc.
(b) i =

C
I dc = I
2r
0
4 sin
2
t + 4 cos
2
t + 9 dt = I
I
13
2r
0
dt = 2I
I
13,
r =
1
2I
I
13
2r
0
2I
I
13 sint dt = 0, j =
1
2I
I
13
2r
0
2I
I
13 cos t dt = 0,
: =
1
2I
I
13
2r
0
I
I
13
(3t) dt =
3
2
2
2
= 3. Hence (r. j. :) = (0. 0. 3).


275
37. From Example 3, a(r. j) = I(1 3j), r = cos t, j = sint, and dc = dt, 0 $ t $ i
1
i
=

C
j
2
a(r. j) dc =

r
0
sin
2
t [I(1 3sint)] dt = I
r
0
(sin
2
t 3sin
3
t) dt
=
1
2
I
r
0
(1 3cos 2t) dt 3I
r
0
(1 3cos
2
t) sint dt

Let u = cos |, Ju = sin | J|
in the second integral
= I
r
2
+
31
1
(1 3&
2
) d&
= I
r
2
3
4
3
1
u
=

C
r
2
a(r. j) dc = I
r
0
cos
2
t (1 3sint) dt =
I
2
r
0
(1 + cos 2t) dt 3I
r
0
cos
2
t sint dt
= I
r
2
3
2
3
, using the same substitution as above.

39. W =

C
F dr =

2r
0
ht 3sint. 3 3cos ti h1 3cos t. sinti dt
=

2r
0
(t 3t cos t 3sint + sint cos t + 3 sint 3sint cos t) dt
=

2r
0
(t 3t cos t + 2 sint) dt =

1
2
t
2
3(t sint + cos t) 32 cos t
2r
0
integrate by parts
in the second term
= 2
2
41. r(t) = h1 + 2t. 4t. 2ti, 0 $ t $ 1,
W =

C
F d r =
1
0
h6t. 1 + 4t. 1 + 6ti h2. 4. 2i dt =

1
0
(12t + 4(1 + 4t) + 2(1 + 6t)) dt
=

1
0
(40t + 6) dt =

20t
2
+ 6t
1
0
= 26
43. Let F = 185 k. To parametrize the staircase, let r = 20 cos t, j = 20 sint, : =
90
6r
t =
15
r
t, 0 $ t $ 6 i
W =

C
F dr =

6r
0
h0. 0. 185i
320 sint. 20 cos t.

15
r
dt = (185)
15
r
6r
0
dt = (185)(90) E 1.67 10
4
ft-lb
45. (a) r(t) = hcos t. sinti, 0 $ t $ 2, and let F =ho. /i. Then
W =

C
F d r =
2r
0
ho. /i h3sint. cos ti dt =

2r
0
(3o sint +/ cos t) dt =

o cos t +/ sint
2r
0
= o + 0 3o + 0 = 0
(b) Yes. F(r. j) = I x = hIr. Iji and
W =

C
F d r =
2r
0
hI cos t. I sinti h3sint. cos ti dt =

2r
0
(3I sint cos t +I sint cos t) dt =

2r
0
0 dt = 0.
47. The work done in moving the object is
C
F dr =

C
F Tdc. We can approximate this integral by dividing C into
7 segments of equal length {c = 2 and approximating F T, that is, the tangential component of force, at a point (r
W
.
. j
W
.
) on
each segment. Since C is composed of straight line segments, F Tis the scalar projection of each force vector onto C.
If we choose (r
W
.
. j
W
.
) to be the point on the segment closest to the origin, then the work done is
C
F Tdc E
7
. =1
[F(r
W
.
. j
W
.
) T(r
W
.
. j
W
.
)] {c = [2 + 2 + 2 + 2 + 1 + 1 + 1](2) = 22. Thus, we estimate the work done to
be approximately 22 J.
17.3 The Fundamental Theorem for Line Integrals ET 16.3
1. C appears to be a smooth curve, and since Q1 is continuous, we know 1 is differentiable. Then Theorem 2 says that the value
of
C
Q1 dr is simply the difference of the values of 1 at the terminal and initial points of C. From the graph, this is
50 310 = 40.
3. u(2r 33j)uj = 33 = u(33r + 4j 38)ur and the domain of F is R
2
which is open and simply-connected, so by
Theorem 6 F is conservative. Thus, there exists a function 1 such that Q1 = F, that is, 1
i
(r. j) = 2r 33j and
1
u
(r. j) = 33r + 4j 38. But 1
i
(r. j) = 2r 33j implies 1(r. j) = r
2
33rj +o(j) and differentiating both sides of this
equation with respect to j gives 1
u
(r. j) = 33r +o
0
(j). Thus 33r + 4j 38 = 33r +o
0
(j) so o
0
(j) = 4j 38 and
o(j) = 2j
2
38j +1 where 1 is a constant. Hence 1(r. j) = r
2
33rj + 2j
2
38j +1 is a potential function for F.
5. u(c
i
sinj)uj = c
i
cos j = u(c
i
cos j)ur and the domain of F is R
2
. Hence F is conservative so there exists a function 1
such that Q1 = F. Then 1
i
(r. j) = c
i
sinj implies 1(r. j) = c
i
sinj +o(j) and 1
u
(r. j) = c
i
cos j +o
0
(j). But
1
u
(r. j) = c
i
cos j so o
0
(j) = 0 i o(j) = 1. Then 1(r. j) = c
i
sinj +1 is a potential function for F.
7. u(jc
i
+ sinj)uj = c
i
+ cos j = u(c
i
+rcos j)ur and the domain of F is R
2
. Hence F is conservative so there
exists a function 1 such that Q1 = F. Then 1
i
(r. j) = jc
i
+ sinj implies 1(r. j) = jc
i
+r sinj +o(j) and
1
u
(r. j) = c
i
+rcos j +o
0
(j). But 1
u
(r. j) = c
i
+rcos j so o(j) = 1 and 1(r. j) = jc
i
+rsinj +1 is a potential
function for F.
9. u(lnj +2rj
3
)uj = 1j +6rj
2
= u(3r
2
j
2
+rj)ur and the domain of F is {(r. j) | j 0} which is open and simply
connected. Hence F is conservative so there exists a function 1 such that Q1 = F. Then 1
i
(r. j) = lnj + 2rj
3
implies
1(r. j) = rlnj +r
2
j
3
+o(j) and 1
u
(r. j) = rj + 3r
2
j
2
+o
0
(j). But 1
u
(r. j) = 3r
2
j
2
+rj so o
0
(j) = 0 i
o(j) = 1 and 1(r. j) = rlnj +r
2
j
3
+1 is a potential function for F.
11. (a) F has continuous rst-order partial derivatives and
u
uj
2rj = 2r =
u
ur
(r
2
) on R
2
, which is open and simply-connected.
Thus, F is conservative by Theorem 6. Then we know that the line integral of F is independent of path; in particular, the
value of
C
F dr depends only on the endpoints of C. Since all three curves have the same initial and terminal points,
C
F dr will have the same value for each curve.
(b) We rst nd a potential function 1, so that Q1 = F. We know 1
i
(r. j) = 2rj and 1
u
(r. j) = r
2
. Integrating
1
i
(r. j) with respect to r, we have 1(r. j) = r
2
j +o(j). Differentiating both sides with respect to j gives
1
u
(r. j) = r
2
+o
0
(j), so we must have r
2
+o
0
(j) = r
2
i o
0
(j) = 0 i o(j) = 1, a constant.
Thus 1(r. j) = r
2
j +1. All three curves start at (1. 2) and end at (3. 2), so by Theorem 2,
C
F dr = 1(3. 2) 31(1. 2) = 18 32 = 16 for each curve.
13. (a) 1
i
(r. j) = rj
2
implies 1(r. j) =
1
2
r
2
j
2
+o(j) and 1
u
(r. j) = r
2
j +o
0
(j). But 1
u
(r. j) = r
2
j so o
0
(j) = 0 i
o(j) = 1, a constant. We can take 1 = 0, so 1(r. j) =
1
2
r
2
j
2
.
SECTION 17.3 THE FUNDAMENTAL THEOREM FOR LINE INTEGRALS ET SECTION 16.3

277
(b) The initial point of C is r(0) = (0. 1) and the terminal point is r(1) = (2. 1), so
C
F dr = 1(2. 1) 31(0. 1) = 2 30 = 2.
15. (a) 1
i
(r. j. :) = j: implies 1(r. j. :) = rj: +o(j. :) and so 1
u
(r. j. :) = r: +o
u
(j. :). But 1
u
(r. j. :) = r: so
o
u
(j. :) = 0 i o(j. :) = /(:). Thus 1(r. j. :) = rj: +/(:) and 1
:
(r. j. :) = rj +/
0
(:). But
1
:
(r. j. :) = rj + 2:, so /
0
(:) = 2: i /(:) = :
2
+1. Hence 1(r. j. :) = rj: +:
2
(taking 1 = 0).
(b)

C
F dr = 1(4. 6. 3) 31(1. 0. 32) = 81 34 = 77.
17. (a) 1
i
(r. j. :) = j
2
cos : implies 1(r. j. :) = rj
2
cos : +o(j. :) and so 1
u
(r. j. :) = 2rj cos : +o
u
(j. :). But
1
u
(r. j. :) = 2rj cos : so o
u
(j. :) = 0 i o(j. :) = /(:). Thus 1(r. j. :) = rj
2
cos : +/(:) and
1
:
(r. j. :) = 3rj
2
sin: +/
0
(:). But 1
:
(r. j. :) = 3rj
2
sin:, so /
0
(:) = 0 i /(:) = 1. Hence
1(r. j. :) = rj
2
cos : (taking 1 = 0).
(b) r(0) = h0. 0. 0i, r() =
2
. 0.
so
C
F dr = 1(
2
. 0. ) 31(0. 0. 0) = 0 30 = 0.
19. Here F(r. j) = tanj i +rsec
2
j j. Then 1(r. j) = rtanj is a potential function for F, that is, Q1 = F so
F is conservative and thus its line integral is independent of path. Hence
C
tanj dr +rsec
2
j dj =

C
F d r =1
2.
r
4
31(1. 0) = 2 tan
r
4
3tan0 = 2.
21. F(r. j) = 2j
3'2
i + 3r
j j, W =

C
F d r. Since u(2j
3'2
)uj = 3
I
j = u(3r
j )ur, there exists a function 1 such

that Q1 = F. In fact, 1
i
(r. j) = 2j
3'2
i 1(r. j) = 2rj
3'2
+o(j) i 1
u
(r. j) = 3rj
1'2
+o
0
(j). But
1
u
(r. j) = 3r
I
j so o
0
(j) = 0 or o(j) = 1. We can take 1 = 0 i 1(r. j) = 2rj
3'2
. Thus
W =

C
F d r = 1(2. 4) 31(1. 1) = 2(2)(8) 32(1) = 30.
23. We know that if the vector eld (call it F) is conservative, then around any closed path C,
C
F dr = 0. But take C to be a
circle centered at the origin, oriented counterclockwise. All of the eld vectors that start on C are roughly in the direction of
motion along C, so the integral around C will be positive. Therefore the eld is not conservative.
25. From the graph, it appears that F is conservative, since around all closed
paths, the number and size of the eld vectors pointing in directions similar
to that of the path seem to be roughly the same as the number and size of the
vectors pointing in the opposite direction. To check, we calculate
u
uj
(sinj) = cos j =
u
ur
(1 +rcos j). Thus F is conservative, by
Theorem 6.
27. Since F is conservative, there exists a function 1 such that F = Q1, that is, 1 = 1
i
, Q = 1
u
, and 1 = 1
:
. Since 1,
Q and 1 have continuous rst order partial derivatives, Clairauts Theorem says that u1uj = 1
iu
= 1
ui
= uQur,
u1u: = 1
i:
= 1
:i
= u1ur, and uQu: = 1
u:
= 1
:u
= u1uj.
29. 1 = {(r. j) | r 0, j 0} = the rst quadrant (excluding the axes).
(a) 1 is open because around every point in 1 we can put a disk that lies in 1.
(b) 1 is connected because the straight line segment joining any two points in 1 lies in 1.
(c) 1 is simply-connected because its connected and has no holes.
31. 1 =

(r. j) | 1 < r
2
+j
2
< 4
= the annular region between the circles with center (0. 0) and radii 1 and 2.
(a) 1 is open.
(b) 1 is connected.
(c) 1 is not simply-connected. For example, r
2
+j
2
= (1.5)
2
is simple and closed and lies within 1 but encloses points that
are not in 1. (Or we can say, 1 has a hole, so is not simply-connected.)
33. (a) 1 = 3
j
r
2
+j
2
,
u1
uj
=
j
2
3r
2
(r
2
+j
2
)
2
and Q =
r
r
2
+j
2
,
uQ
ur
=
j
2
3r
2
(r
2
+j
2
)
2
. Thus
u1
uj
=
uQ
ur
.
(b) C
1
: r = cos t, j = sint, 0 $ t $ , C
2
: r = cos t, j = sint, t = 2 to t = . Then
C
1
F dr =

r
0
(3sint)(3sint) + (cos t)(cos t)
cos
2
t + sin
2
t
dt =

r
0
dt = and
C
2
F dr =

r
2r
dt = 3
Since these arent equal, the line integral of F isnt independent of path. (Or notice that
C
3
F dr =

2r
0
dt = 2 where
C
3
is the circle r
2
+j
2
= 1, and apply the contrapositive of Theorem 3.) This doesnt contradict Theorem 6, since the
domain of F, which is R
2
except the origin, isnt simply-connected.
17.4 Green's Theorem ET 16.4
1. (a) Parametric equations for C are r = 2 cos t, j = 2 sint, 0 $ t $ 2. Then
C
(r 3j) dr + (r +j) dj =

2r
0
[(2 cos t 32 sint)(32 sint) + (2 cos t + 2 sint)(2 cos t)] dt
=

2r
0
(4 sin
2
t + 4 cos
2
t) dt =

2r
0
4 dt = 4t
2r
0
= 8
(b) Note that C as given in part (a) is a positively oriented, smooth, simple closed curve. Then by Greens Theorem,
C
(r 3j) dr + (r +j) dj =

T
O
Oi
(r +j) 3
O
Ou
(r 3j)
d =

T
[1 3(31)] d = 2
T
d
= 2(1) = 2(2)
2
= 8
3. (a) C
1
: r = t i dr = dt, j = 0 i dj = 0 dt, 0 $ t $ 1.
C
2
: r = 1 i dr = 0 dt, j = t i dj = dt, 0 $ t $ 2.
C
3
: r = 1 3t i dr = 3dt, j = 2 32t i dj = 32 dt, 0 $ t $ 1.
Thus

C
rj dr +r
2
j
3
dj =

C
1
+C
2
+C
3
rj dr +r
2
j
3
dj
=

1
0
0 dt +
2
0
t
3
dt +
1
0
3(1 3t)(2 32t) 32(1 3t)

2
(2 32t)
3
dt
= 0 +
1
4
t
4
2
0
+
2
3
(1 3t)
3
+
8
3
(1 3t)
6
1
0
= 4 3
10
3
=
2
3
SECTION 17.4 GREENS THEOREM ET SECTION 16.4

279
(b)

C
rj dr +r
2
j
3
dj =

T
O
Oi
(r
2
j
3
) 3
O
Ou
(rj)
d =

1
0
2i
0
(2rj
3
3r) dj dr
=

1
0
1
2
rj
4
3rj
u=2i
u=0
dr =

1
0
(8r
5
32r
2
) dr =
4
3
3
2
3
=
2
3
5. The region 1 enclosed by C is given by {(r. j) | 0 $ r $ 2. r $ j $ 2r}, so
C
rj
2
dr + 2r
2
j dj =

T
O
Oi
(2r
2
j) 3
O
Ou
(rj
2
)
d
=

2
0
2i
i
(4rj 32rj) dj dr
=

2
0
rj
2
u=2i
u=i
dr
=

2
0
3r
3
dr =
3
4
r
4
2
0
= 12
7.

C
j +c
I
i
dr + (2r + cos j
2
) dj =

T
O
Oi
(2r + cos j
2
) 3
O
Ou
j +c
I
i
d
=

1
0
I
u
u
2
(2 31) drdj =

1
0
(j
1'2
3j
2
) dj =
1
3
9.

C
j
3
dr 3r
3
dj =

T
O
Oi
(3r
3
) 3
O
Ou
(j
3
)
d =

T
(33r
2
33j
2
) d =

2r
0
2
0
(33v
2
) v dv d0
= 33
2r
0
d0

2
0
v
3
dv = 33(2)(4) = 324
11. F(r. j) =
I
r +j
3
. r
2
+
I
j
and the region 1 enclosed by C is given by {(r. j) | 0 $ r $ . 0 $ j $ sinr}.

C is traversed clockwise, so 3C gives the positive orientation.
C
F d r = 3
3C
I
r +j
3
dr +
r
2
+
I
j
dj = 3
O
Oi
r
2
+
3
O
Ou
I
r +j
3
d
= 3
r
0
sin i
0
(2r 33j
2
) dj dr = 3
r
0
2rj 3j
3
u=sin i
u=0
dr
= 3
r
0
(2rsinr 3sin
3
r) dr = 3
r
0
(2rsinr 3(1 3cos
2
r) sinr) dr
= 3
2 sinr 32rcos r + cos r 3

1
3
cos
3
r
r
0
[integrate by parts in the rst term]
= 3
2 32 +
2
3
=
4
3
32
13. F(r. j) =
c
i
+r
2
j. c
u
3rj
2
and the region 1 enclosed by C is the disk r

2
+j
2
$ 25.
C is traversed clockwise, so 3C gives the positive orientation.
C
F d r = 3
3C
(c
i
+r
2
j) dr + (c
u
3rj
2
) dj = 3
O
Oi
(c
u
3rj
2
) 3
O
Ou
(c
i
+r
2
j)
d
= 3
T
(3j
2
3r
2
) d =

T
(r
2
+j
2
) d =

2r
0
5
0
(v
2
) v dv d0
=

2r
0
d0

5
0
v
3
dv = 2
1
4
v
4
5
0
=
625
2

15. Here C = C
1
+C
2
where
C
1
can be parametrized as r = t, j = 1, 31 $ t $ 1, and
C
2
is given by r = 3t, j = 2 3t
2
, 31 $ t $ 1.
Then the line integral is
C
1
+C
2
j
2
c
i
dr +r
2
c
u
dj =

1
31
[1 c
I
+t
2
c 0] dt
+
1
31
[(2 3t
2
)
2
c
3I
(31) + (3t)
2
c
23I
2
(32t)] dt
=

1
31
[c
I
3(2 3t
2
)
2
c
3I
32t
3
c
23I
2
] dt = 38c + 48c
31
according to a CAS. The double integral is
uQ
ur
3
u1
uj
d =

1
31
23i
2
1
(2rc
u
32jc
i
) dj dr = 38c + 48c
31
, verifying Greens Theorem in this case.
17. By Greens Theorem, W =

C
F dr =

C
r(r +j) dr +rj
2
dj =

T
(j
2
3r) dj dr where C is the path described in the
question and 1 is the triangle bounded by C. So
W =

1
0
13i
0
(j
2
3r) dj dr =

1
0
1
3
j
3
3rj
u =13i
u =0
dr =

1
0
1
3
(1 3r)
3
3r(1 3r)
dr
=

3
1
12
(1 3r)
4
3
1
2
r
2
+
1
3
r
3
1
0
=

3
1
2
+
1
3
3
1
12
= 3
1
12
19. Let C
1
be the arch of the cycloid from (0. 0) to (2. 0), which corresponds to 0 $ t $ 2, and let C
2
be the segment from
(2. 0) to (0. 0), so C
2
is given by r = 2 3t, j = 0, 0 $ t $ 2. Then C = C
1
C
2
is traversed clockwise, so 3C is
oriented positively. Thus 3C encloses the area under one arch of the cycloid and from (5) we have
= 3
3C
j dr =

C
1
j dr +
C
2
j dr =

2r
0
(1 3cos t)(1 3cos t) dt +
2r
0
0 (3dt)
=

2r
0
(1 32 cos t + cos
2
t) dt + 0 =

t 32 sint +
1
2
t +
1
4
sin2t
2r
0
= 3
21. (a) Using Equation 17.2.8 [ ET 16.2.8], we write parametric equations of the line segment as r = (1 3t)r
1
+tr
2
,
j = (1 3t)j
1
+tj
2
, 0 $ t $ 1. Then dr = (r
2
3r
1
) dt and dj = (j
2
3j
1
) dt, so
C
rdj 3j dr =

1
0
[(1 3t)r
1
+tr
2
](j
2
3j
1
) dt + [(1 3t)j
1
+tj
2
](r
2
3r
1
) dt
=

1
0
(r
1
(j
2
3j
1
) 3j
1
(r
2
3r
1
) +t[(j
2
3j
1
)(r
2
3r
1
) 3(r
2
3r
1
)(j
2
3j
1
)]) dt
=

1
0
(r
1
j
2
3r
2
j
1
) dt = r
1
j
2
3r
2
j
1
(b) We apply Greens Theorem to the path C = C
1
C
2
C
n
, where C
.
is the line segment that joins (r
.
. j
.
) to
(r
.+1
. j
.+1
) for j = 1, 2, . . ., n 31, and C
n
is the line segment that joins (r
n
. j
n
) to (r
1
. j
1
). From (5),
1
2
C
rdj 3j dr =

T
d, where 1 is the polygon bounded by C. Therefore
area of polygon = (1) =
T
d =
1
2
C
rdj 3j dr
=
1
2
C
1
rdj 3j dr +
C
2
rdj 3j dr + +
C
q1
rdj 3j dr +
C
q
rdj 3j dr
To evaluate these integrals we use the formula from (a) to get

(1) =
1
2
[(r
1
j
2
3r
2
j
1
) + (r
2
j
3
3r
3
j
2
) + + (r
n31
j
n
3r
n
j
n31
) + (r
n
j
1
3r
1
j
n
)].
SECTION 17.4 GREENS THEOREM ET SECTION 16.4

281
(c) =
1
2
[(0 1 32 0) + (2 3 31 1) + (1 2 30 3) + (0 1 3(31) 2) + (31 0 30 1)]
=
1
2
(0 + 5 + 2 + 2) =
9
2
23. We orient the quarter-circular region as shown in the gure.
=
1
4
o
2
so r =
1
o
2
2
C
r
2
dj and j = 3
1
o
2
2
C
j
2
dr.
Here C = C
1
+C
2
+C
3
where C
1
: r = t, j = 0, 0 $ t $ o;
C
2
: r = o cos t, j = o sint, 0 $ t $
r
2
; and
C
3
: r = 0, j = o 3t, 0 $ t $ o. Then
C
r
2
dj =

C
1
r
2
dj +
C
2
r
2
dj +
C
3
r
2
dj =

a
0
0 dt +
r'2
0
(o cos t)
2
(o cos t) dt +
a
0
0 dt
=

r'2
0
o
3
cos
3
t dt = o
3
r'2
0
(1 3sin
2
t) cos t dt = o
3
sint 3
1
3
sin
3
t
r'2
0
=
2
3
o
3
so r =
1
o
2
2
C
r
2
dj =
4o
3
.
C
j
2
dr =

C
1
j
2
dr +
C
2
j
2
dr +
C
3
j
2
dr =

a
0
0 dt +
r'2
0
(o sint)
2
(3o sint) dt +
a
0
0 dt
=

r'2
0
(3o
3
sin
3
t) dt = 3o
3
r'2
0
(1 3cos
2
t) sint dt = 3o
3
1
3
cos
3
t 3cos t
r'2
0
= 3
2
3
o
3
,
so j = 3
1
o
2
2
C
j
2
dr =
4o
3
. Thus (r. j) =
4o
3
.
4o
3
.
25. By Greens Theorem, 3
1
3
a
C
j
3
dr = 3
1
3
a
T
(33j
2
) d =

T
j
2
a d = 1
i
and
1
3
a
C
r
3
dj =
1
3
a
T
(3r
2
) d =

T
r
2
a d = 1
u
.
27. Since C is a simple closed path which doesnt pass through or enclose the origin, there exists an open region that doesnt
contain the origin but does contain 1. Thus 1 = 3j(r
2
+j
2
) and Q = r(r
2
+j
2
) have continuous partial derivatives on
this open region containing 1 and we can apply Greens Theorem. But by Exercise 17.3.33(a) [ ET 16.3.33(a)],
u1uj = uQur, so
C
F dr =

T
0 d = 0.
29. Using the rst part of (5), we have that
T
drdj = (1) =
OT
rdj. But r = o(&. ), and dj =
u/
u&
d& +
u/
u
d,
and we orient uo by taking the positive direction to be that which corresponds, under the mapping, to the positive direction
along u1, so
OT
rdj =

OS
o(&. )
u/
u&
d& +
u/
u
d
=

OS
o(&. )
u/
u&
d& +o(&. )
u/
u
d
=
O
Ou
o(&. )
OI
Or
3
O
Or
o(&. )
OI
Ou
d [using Greens Theorem in the &-plane]

=
Og
Ou
OI
Or
+o(&. )
O
2
I
OuOr
3
Og
Or
OI
Ou
3o(&. )
O
2
I
Or Ou
d [using the Chain Rule]

=
Oi
Ou
Ou
Or
3
Oi
Or
Ou
Ou
d [by the equality of mixed partials] =
S
O(i.u)
O(u.r)
d&d
The sign is chosen to be positive if the orientation that we gave to uo corresponds to the usual positive orientation, and it is
negative otherwise. In either case, since (1) is positive, the sign chosen must be the same as the sign of
u (r. j)
u(&. )
.
Therefore (1) =
T
drdj =

S
u(r. j)
u(&. )
d&d.
17.5 Curl and Divergence ET 16.5
1. (a) curl F = QF =
i j k
uur uuj uu:
rj: 0 3r
2
j
= (3r
2
30) i 3(32rj 3rj) j + (0 3r:) k
= 3r
2
i + 3rj j 3r: k
(b) div F = Q F =
u
ur
(rj:) +
u
uj
(0) +
u
u:
(3r
2
j) = j: + 0 + 0 = j:
i j k
uur uuj uu:
1 r +j: rj 3
I
:
= (r 3j) i 3(j 30) j + (1 30) k

= (r 3j) i 3j j + k
(b) div F = Q F =
u
ur
(1) +
u
uj
(r +j:) +
u
u:
rj 3
I
:
= : 3
1
2
I
:
5. (a) curl F= QF =
i j k
uur uuj uu:
r
r
2
+j
2
+:
2
j
r
2
+j
2
+:
2
:
r
2
+j
2
+:
2
=
1
(r
2
+j
2
+:
2
)
3'2
[(3j: +j:) i 3(3r: +r:) j + (3rj +rj) k] = 0
(b) div F= Q F =
u
ur
r
2
+j
2
+:
2
+
u
uj
r
2
+j
2
+:
2
+
u
u:
r
2
+j
2
+:
2
=
r
2
+j
2
+:
2
3r
2
(r
2
+j
2
+:
2
)
3'2
+
r
2
+j
2
+:
2
3j
2
(r
2
+j
2
+:
2
)
3'2
+
r
2
+j
2
+:
2
3:
2
(r
2
+j
2
+:
2
)
3'2
=
2r
2
+ 2j
2
+ 2:
2
(r
2
+j
2
+:
2
)
3'2
=
2
r
2
+j
2
+:
2
i j k
uur uuj uu:
lnr ln(rj) ln(rj:)
=

r:
rj:
30
i 3
j:
rj:
30
j +
j
rj
30
k =

1
j
. 3
1
r
.
1
r
(b) div F = Q F =
u
ur
(lnr) +
u
uj
(ln(rj)) +
u
u:
(ln(rj:)) =
1
r
+
r
rj
+
rj
rj:
=
1
r
+
1
j
+
1
:
9. If the vector eld is F = 1 i +Qj +1k, then we know 1 = 0. In addition, the r-component of each vector of F is 0, so
1 = 0, hence
u1
ur
=
u1
uj
=
u1
u:
=
u1
ur
=
u1
uj
=
u1
u:
= 0. Q decreases as j increases, so
uQ
uj
< 0, but Q doesnt change
in the r- or :-directions, so
uQ
ur
=
uQ
u:
= 0.
(a) div F =
u1
ur
+
uQ
uj
+
u1
u:
= 0 +
uQ
uj
+ 0 < 0
(b) curl F =

u1
uj
3
uQ
u:
i +
u1
u:
3
u1
ur
j +
uQ
ur
3
u1
uj
k = (0 30) i + (0 30) j + (0 30) k = 0

SECTION 17.5 CURL AND DIVERGENCE ET SECTION 16.5

283
11. If the vector eld is F = 1 i +Qj +1k, then we know 1 = 0. In addition, the j-component of each vector of F is 0, so
Q = 0, hence
uQ
ur
=
uQ
uj
=
uQ
u:
=
u1
ur
=
u1
uj
=
u1
u:
= 0. 1 increases as j increases, so
u1
uj
0, but 1 doesnt change in
the r- or :-directions, so
u1
ur
=
u1
u:
= 0.
(a) div F =
u1
ur
+
uQ
uj
+
u1
u:
= 0 + 0 + 0 = 0
(b) curl F =

u1
uj
3
uQ
u:
i +
u1
u:
3
u1
ur
j +
uQ
ur
3
u1
uj
k = (0 30) i + (0 30) j +
0 3
u1
uj
k = 3
u1
uj
k
Since
u1
uj
0, 3
u1
uj
k is a vector pointing in the negative :-direction.
13. curl F = QF =
i j k
uur uuj uu:
j
2
:
3
2rj:
3
3rj
2
:
2
= (6rj:
2
36rj:
2
) i 3(3j
2
:
2
33j
2
:
2
) j + (2j:
3
32j:
3
) k = 0
and F is dened on all of R
3
with component functions which have continuous partial derivatives, so by Theorem 4,
F is conservative. Thus, there exists a function 1 such that F = Q1. Then 1
i
(r. j. :) = j
2
:
3
implies
1(r. j. :) = rj
2
:
3
+o(j. :) and 1
u
(r. j. :) = 2rj:
3
+o
u
(j. :). But 1
u
(r. j. :) = 2rj:
3
, so o(j. :) = /(:) and
1(r. j. :) = rj
2
:
3
+/(:). Thus 1
:
(r. j. :) = 3rj
2
:
2
+/
0
(:) but 1
:
(r. j. :) = 3rj
2
:
2
so /(:) = 1, a constant.
Hence a potential function for F is 1(r. j. :) = rj
2
:
3
+1.
15. curl F = QF =
i j k
uur uuj uu:
2rj r
2
+ 2j: j
2
= (2j 32j) i 3(0 30) j + (2r 32r) k = 0, F is dened on all of R

3
,
and the partial derivatives of the component functions are continuous, so F is conservative. Thus there exists a function 1
such that Q1 = F. Then 1
i
(r. j. :) = 2rj implies 1(r. j. :) = r
2
j +o(j. :) and 1
u
(r. j. :) = r
2
+o
u
(j. :). But
1
u
(r. j. :) = r
2
+ 2j:, so o(j. :) = j
2
: +/(:) and 1(r. j. :) = r
2
j +j
2
: +/(:). Thus 1
:
(r. j. :) = j
2
+/
0
(:) but
1
:
(r. j. :) = j
2
so /(:) = 1 and 1(r. j. :) = r
2
j +j
2
: +1.
17. curl F = QF =
i j k
uur uuj uu:
jc
3i
c
3i
2:
= (0 30) i 3(0 30) j + (3c

3i
3c
3i
) k = 32c
3i
k 6= 0,
so F is not conservative.
19. No. Assume there is such a G. Then div(curl G) =
u
ur
(rsinj) +
u
uj
(cos j) +
u
u:
(: 3rj) = sinj 3sinj + 1 6= 0,
which contradicts Theorem 11.
21. curl F =
i j k
uur uuj uu:
1(r) o(j) /(:)
= (0 30) i + (0 30) j + (0 30) k = 0. Hence F = 1(r) i +o(j) j +/(:) k

is irrotational.
For Exercises 2329, let F(r. j. :) = 1
1
i +Q
1
j +1
1
k and G(r. j. :) = 1
2
i +Q
2
j +1
2
k.
23. div(F+G) = divh1
1
+1
2
. Q
1
+Q
2
. 1
1
+1
2
i =
u(1
1
+1
2
)
ur
+
u(Q
1
+Q
2
)
uj
+
u(1
1
+1
2
)
u:
=
u1
1
ur
+
u1
2
ur
+
uQ
1
uj
+
uQ
2
uj
+
u1
1
u:
+
u1
2
u:
=

u1
1
ur
+
uQ
1
uj
+
u1
1
u:
u1
2
ur
+
uQ
2
uj
+
u1
2
u:
= divh1
1
. Q
1
. 1
1
i + divh1
2
. Q
2
. 1
2
i = div F+ div G
25. div(1F) = div(1 h1
1
. Q
1
. 1
1
i) = divh11
1
. 1Q
1
. 11
1
i =
u(11
1
)
ur
+
u(1Q
1
)
uj
+
u(11
1
)
u:
=

1
u1
1
ur
+1
1
u1
ur
1
uQ
1
uj
+Q
1
u1
uj
1
u1
1
u:
+1
1
u1
u:
= 1
u1
1
ur
+
uQ
1
uj
+
u1
1
u:
+ h1
1
. Q
1
. 1
1
i
u1
ur
.
u1
uj
.
u1
u:
= 1 div F+F Q1
27. div(FG) =Q (FG) =
uur uuj uu:

1
1
Q
1
1
1
1
2
Q
2
1
2
=
u
ur
Q
1
1
1
Q
2
1
2
3
u
uj
1
1
1
1
1
2
1
2
+
u
u:
1
1
Q
1
1
2
Q
2
Q
1
u1
2
ur
+1
2
uQ
1
ur
3Q
2
u1
1
ur
31
1
uQ
2
ur
1
1
u1
2
uj
+1
2
u1
1
uj
31
2
u1
1
uj
31
1
u1
2
uj
1
1
uQ
2
u:
+Q
2
u1
1
u:
31
2
uQ
1
u:
3Q
1
u1
2
u:
1
2
u1
1
uj
3
uQ
1
u:
+Q
2
u1
1
u:
3
u1
1
ur
+1
2
uQ
1
ur
3
u1
1
uj
1
1
u1
2
uj
3
uQ
2
u:
+Q
1
u1
2
u:
3
u1
2
ur
+1
1
uQ
2
ur
3
u1
2
uj
= G curl F3F curl G

29. curl(curl F) = Q (QF) =
i j k
uur uuj uu:
u1
1
uj 3uQ
1
u: u1
1
u: 3u1
1
ur uQ
1
ur 3u1
1
uj
=

u
2
Q
1
ujur
3
u
2
1
1
uj
2
3
u
2
1
1
u:
2
+
u
2
1
1
u:ur
i +
u
2
1
1
u:uj
3
u
2
Q
1
u:
2
3
u
2
Q
1
ur
2
+
u
2
1
1
uruj
j
+
u
2
1
1
uru:
3
u
2
1
1
ur
2
3
u
2
1
1
uj
2
+
u
2
Q
1
uju:
k
Now lets consider grad(div F) 3Q
2
F and compare with the above.
(Note that Q
2
F is dened on page 1102 [ ET 1066].)
SECTION 17.5 CURL AND DIVERGENCE ET SECTION 16.5

285
grad(div F) 3Q
2
F =

u
2
1
1
ur
2
+
u
2
Q
1
uruj
+
u
2
1
1
uru:
i +
u
2
1
1
ujur
+
u
2
Q
1
uj
2
+
u
2
1
1
uju:
j +
u
2
1
1
u:ur
+
u
2
Q
1
u:uj
+
u
2
1
1
u:
2
u
2
1
1
ur
2
+
u
2
1
1
uj
2
+
u
2
1
1
u:
2
i +
u
2
Q
1
ur
2
+
u
2
Q
1
uj
2
+
u
2
Q
1
u:
2
j
+
u
2
1
1
ur
2
+
u
2
1
1
uj
2
+
u
2
1
1
u:
2
=

u
2
Q
1
uruj
+
u
2
1
1
uru:
3
u
2
1
1
uj
2
3
u
2
1
1
u:
2
i +
u
2
1
1
ujur
+
u
2
1
1
uju:
3
u
2
Q
1
ur
2
3
u
2
Q
1
u:
2
j
+
u
2
1
1
u:ur
+
u
2
Q
1
u:uj
3
u
2
1
1
ur
2
3
u
2
1
2
uj
2
k
Then applying Clairauts Theorem to reverse the order of differentiation in the second partial derivatives as needed and
comparing, we have curl curl F = graddiv F3Q
2
F as desired.
31. (a) Qv = Q
r
2
+j
2
+:
2
=
r
r
2
+j
2
+:
2
i +
j
r
2
+j
2
+:
2
j +
:
r
2
+j
2
+:
2
k =
ri +j j +: k
r
2
+j
2
+:
2
=
r
v
(b) Qr =
i j k
u
ur
u
uj
u
u:
r j :
=

u
uj
(:) 3
u
u:
(j)
i +
u
u:
(r) 3
u
ur
(:)
j +
u
ur
(j) 3
u
uj
(r)
k = 0
(c) Q
1
v
= Q
r
2
+j
2
+:
2
=
3
1
2
r
2
+j
2
+:
2
(2r)
r
2
+j
2
+:
2
i 3
1
2
r
2
+j
2
+:
2
(2j)
r
2
+j
2
+:
2
j 3
1
2
r
2
+j
2
+:
2
(2:)
r
2
+j
2
+:
2
k
= 3
ri +j j +: k
(r
2
+j
2
+:
2
)
3'2
= 3
r
v
3
(d) Qlnv = Qln(r
2
+j
2
+:
2
)
1'2
=
1
2
Qln(r
2
+j
2
+:
2
)
=
r
r
2
+j
2
+:
2
i +
j
r
2
+j
2
+:
2
j +
:
r
2
+j
2
+:
2
k =
ri +j j +: k
r
2
+j
2
+:
2
=
r
v
2
33. By (13),
C
1(Qo) ndc =

T
div(1Qo) d =

T
[1 div(Qo) +Qo Q1] d by Exercise 25. But div(Qo) = Q
2
o.
Hence
T
1Q
2
o d =

C
1(Qo) ndc 3
T
Qo Q1 d.
35. Let 1(r. j) = 1. Then Q1 = 0 and Greens rst identity (see Exercise 33) says
T
Q
2
o d =

C
(Qo) ndc 3
T
0 Qo d i

T
Q
2
o d =

C
Qo ndc. But o is harmonic on 1, so
Q
2
o = 0 i

C
Qo ndc = 0 and
C
1
n
o dc =

C
(Qo n) dc = 0.
37. (a) We know that . = d, and from the diagram sin0 = dv i = d. = (sin0)v. = |wr|. But v is perpendicular
to both w and r, so that v = wr.
(b) From (a), v = wr =
i j k
0 0 .
r j :
= (0 : 3.j) i + (.r 30 :) j + (0 j 3r 0) k = 3.j i +.rj

(c) curl v = Qv =
i j k
uur uuj uu:
3.j .r 0
=

u
uj
(0) 3
u
u:
(.r)
i +
u
u:
(3.j) 3
u
ur
(0)
j +
u
ur
(.r) 3
u
uj
(3.j)
k
= [. 3(3.)] k = 2. k = 2w
39. For any continuous function 1 on R
3
, dene a vector eld G(r. j. :) = ho(r. j. :). 0. 0i where o(r. j. :) =
i
0
1 (t. j. :) dt.
Then div G =
u
ur
(o(r. j. :)) +
u
uj
(0) +
u
u:
(0) =
u
ur
i
0
1(t. j. :) dt = 1(r. j. :) by the Fundamental Theorem of
Calculus. Thus every continuous function 1 on R
3
is the divergence of some vector eld.
17.6 Parametric Surfaces and Their Areas ET 16.6
1. 1(7. 10. 4) lies on the parametric surface r(&. ) = h2& + 3. 1 + 5& 3. 2 +& +i if and only if there are values for &
and where 2& + 3 = 7, 1 + 5& 3 = 10, and 2 +& + = 4. But solving the rst two equations simultaneously gives
& = 2, = 1 and these values do not satisfy the third equation, so 1 does not lie on the surface.
Q(5. 22. 5) lies on the surface if 2& + 3 = 5, 1 + 5& 3 = 22, and 2 +& + = 5 for some values of & and . Solving the
rst two equations simultaneously gives & = 4, = 31 and these values satisfy the third equation, so Q lies on the surface.
3. r(&. ) = (& +) i + (3 3) j + (1 + 4& + 5) k = h0. 3. 1i +&h1. 0. 4i + h1. 31. 5i. From Example 3, we recognize
this as a vector equation of a plane through the point (0. 3. 1) and containing vectors a = h1. 0. 4i and b = h1. 31. 5i. If we
wish to nd a more conventional equation for the plane, a normal vector to the plane is a b =
i j k
1 0 4
1 31 5
= 4 i 3 j 3 k
and an equation of the plane is 4(r 30) 3(j 33) 3(: 31) = 0 or 4r 3j 3: = 34.
5. r(c. t) =
c. t. t
2
3c
2
, so the corresponding parametric equations for the surface are r = c, j = t, : = t

2
3c
2
. For any
point (r. j. :) on the surface, we have : = j
2
3r
2
. With no restrictions on the parameters, the surface is : = j
2
3r
2
, which
we recognize as a hyperbolic paraboloid.
SECTION 17.6 PARAMETRIC SURFACES AND THEIR AREAS ET SECTION 16.6

287
7. r(&. ) =
&
2
+ 1.
3
+ 1. & +
, 31 $ & $ 1, 31 $ $ 1.
The surface has parametric equations r = &
2
+ 1, j =
3
+ 1, : = & +,
31 $ & $ 1, 31 $ $ 1. In Maple, the surface can be graphed by entering
plot3d([u2+1,v3+1,u+v],u=-1..1,v=-1..1);. In
Mathematica we use the ParametricPlot3D command. If we keep &
constant at &
0
, r = &
2
0
+ 1, a constant, so the corresponding grid curves must
be the curves parallel to the j:-plane. If is constant, we have j =
3
0
+ 1,
a constant, so these grid curves are the curves parallel to the r:-plane.
9. r(&. ) =
&cos . &sin. &

5
.
The surface has parametric equations r = &cos , j = &sin,
: = &
5
, 31 $ & $ 1, 0 $ $ 2. Note that if & = &
0
is constant
then : = &
5
0
is constant and r = &
0
cos , j = &
0
sin describe a
circle in r, j of radius |&
0
|, so the corresponding grid curves are
circles parallel to the rj-plane. If =
0
, a constant, the parametric
equations become r = &cos
0
, j = &sin
0
, : = &
5
. Then
j = (tan
0
)r, so these are the grid curves we see that lie in vertical
planes j = Ir through the :-axis.
11. r = sin, j = cos &sin4, : = sin2&sin4, 0 $ & $ 2, 3
r
2
$ $
r
2
.
Note that if =
0
is constant, then r = sin
0
is constant, so the
corresponding grid curves must be parallel to the j:-plane. These
are the vertically oriented grid curves we see, each shaped like a
gure-eight. When & = &
0
is held constant, the parametric
equations become r = sin, j = cos &
0
sin4,
: = sin2&
0
sin4. Since : is a constant multiple of j, the
corresponding grid curves are the curves contained in planes
: = Ij that pass through the r-axis.
13. r(&. ) = &cos i +&sin j + k. The parametric equations for the surface are r = &cos , j = &sin, : = . We look at
the grid curves rst; if we x , then r and j parametrize a straight line in the plane : = which intersects the :-axis. If & is
held constant, the projection onto the rj-plane is circular; with : = , each grid curve is a helix. The surface is a spiraling
ramp, graph I.
15. r(&. ) = sin i + cos & sin2 j + sin& sin2 k. Parametric equations for the surface are r = sin, j = cos & sin2,
: = sin& sin2. If =
0
is xed, then r = sin
0
is constant, and j = (sin2
0
) cos & and : = (sin2
0
) sin& describe a
circle of radius |sin2
0
|, so each corresponding grid curve is a circle contained in the vertical plane r = sin
0
parallel to the
j:-plane. The only possible surface is graph II. The grid curves we see running lengthwise along the surface correspond to
holding & constant, in which case j = (cos &
0
) sin2, : = (sin&
0
) sin2 i : = (tan&
0
)j, so each grid curve lies in a
plane : = Ij that includes the r-axis.
17. r = cos
3
& cos
3
, j = sin
3
& cos
3
, : = sin
3
. If =
0
is held constant then : = sin
3
0
is constant, so the
corresponding grid curve lies in a horizontal plane. Several of the graphs exhibit horizontal grid curves, but the curves for this
surface are neither circles nor straight lines, so graph III is the only possibility. (In fact, the horizontal grid curves here are
members of the family r = o cos
3
&, j = o sin
3
& and are called astroids.) The vertical grid curves we see on the surface
correspond to & = &
0
held constant, as then we have r = cos
3
&
0
cos
3
, j = sin
3
&
0
cos
3
so the corresponding grid curve
lies in the vertical plane j = (tan
3
&
0
)r through the :-axis.
19. From Example 3, parametric equations for the plane through the point (1. 2. 33) that contains the vectors a = h1. 1. 31i and
b = h1. 31. 1i are r = 1 +&(1) +(1) = 1 +& +, j = 2 +&(1) +(31) = 2 +& 3,
: = 33 +&(31) +(1) = 33 3& +.
21. Solving the equation for j gives j
2
= 1 3r
2
+:
2
i j =
I
1 3r
2
+:
2
. (We choose the positive root since we want the
part of the hyperboloid that corresponds to j D 0.) If we let r and : be the parameters, parametric equations are r = r, : = :,
j =
I
1 3r
2
+:
2
.
23. Since the cone intersects the sphere in the circle r
2
+j
2
= 2, : =
I
2 and we want the portion of the sphere above this, we
can parametrize the surface as r = r, j = j, : =

4 3r
2
3j
2
where r
2
+j
2
$ 2.
Alternate solution: Using spherical coordinates, r = 2 sinccos 0, j = 2 sincsin0, : = 2 cos c where 0 $ c $
r
4
and
0 $ 0 $ 2.
25. Parametric equations are r = r, j = 4 cos 0, : = 4 sin0, 0 $ r $ 5, 0 $ 0 $ 2.
27. The surface appears to be a portion of a circular cylinder of radius 3 with axis the r-axis. An equation of the cylinder is
j
2
+:
2
= 9, and we can impose the restrictions 0 $ r $ 5, j $ 0 to obtain the portion shown. To graph the surface on a
CAS, we can use parametric equations r = &, j = 3 cos , : = 3 sin with the parameter domain 0 $ & $ 5,
r
2
$ $
3r
2
.
Alternatively, we can regard r and : as parameters. Then parametric equations are r = r, : = :, j = 3
I
9 3:
2
, where
0 $ r $ 5 and 33 $ : $ 3.
29. Using Equations 3, we have the parametrization r = r, j = c
3i
cos 0, : = c
3i
sin0, 0 $ r $ 3, 0 $ 0 $ 2.

289
31. (a) Replacing cos & by sin& and sin& by cos & gives parametric equations
r = (2 + sin) sin&, j = (2 + sin) cos &, : = & + cos . From the graph, it
appears that the direction of the spiral is reversed. We can verify this observation by
noting that the projection of the spiral grid curves onto the rj-plane, given by
r = (2 + sin) sin&, j = (2 + sin) cos &, : = 0, draws a circle in the clockwise
direction for each value of . The original equations, on the other hand, give circular
projections drawn in the counterclockwise direction. The equation for : is identical in
both surfaces, so as : increases, these grid curves spiral up in opposite directions for
the two surfaces.
(b) Replacing cos & by cos 2& and sin& by sin2& gives parametric equations
r = (2 + sin) cos 2&, j = (2 + sin) sin2&, : = & + cos . From the graph, it
appears that the number of coils in the surface doubles within the same parametric
domain. We can verify this observation by noting that the projection of the spiral grid
curves onto the rj-plane, given by r = (2 + sin) cos 2&, j = (2 + sin) sin2&,
: = 0 (where is constant), complete circular revolutions for 0 $ & $ while the
original surface requires 0 $ & $ 2 for a complete revolution. Thus, the new
surface winds around twice as fast as the original surface, and since the equation for :
is identical in both surfaces, we observe twice as many circular coils in the same
:-interval.
33. r(&. ) = (& +) i + 3&
2
j + (& 3) k.
r
u
= i +6&j +k and r
r
= i 3k, so r
u
r
r
= 36&i +2 j 36&k.
Since the point (2. 3. 0) corresponds to & = 1, = 1, a normal vector
to the surface at (2. 3. 0) is 36 i + 2 j 36 k, and an equation of the
tangent plane is 36r + 2j 36: = 36 or 3r 3j + 3: = 3.

35. r(&. ) = &
2
i + 2&sin j +&cos k i r(1. 0) = (1. 0. 1).
r
u
= 2&i + 2 sin j + cos k and r
r
= 2&cos j 3&sin k,
so a normal vector to the surface at the point (1. 0. 1) is
r
u
(1. 0) r
r
(1. 0) = (2 i + k) (2 j) = 32 i + 4 k.
Thus an equation of the tangent plane at (1. 0. 1) is
32(r 31) + 0(j 30) + 4(: 31) = 0 or 3r + 2: = 1.
37. The surface o is given by : = 1(r. j) = 6 33r 32j which intersects the rj-plane in the line 3r + 2j = 6, so 1 is the
triangular region given by
(r. j)

0 $ r $ 2. 0 $ j $ 3 3
3
2
r
. By Formula 9, the surface area of o is

(o) =

T
1 +
u:
ur
2
+
u:
uj
2
d
=

T
1 + (33)
2
+ (32)
2
d =
I
14
T
d =
I
14 (1) =
I
14
1
2
2 3
= 3
I
14.
39. : = 1(r. j) =
2
3
(r
3'2
+j
3'2
) and 1 = {(r. j) | 0 $ r $ 1. 0 $ j $ 1}. Then 1
i
= r
1'2
, 1
u
= j
1'2
and
(o) =

T
1 + (
I
r)
2
+
I
j
2
d =

1
0
1
0
I
1 +r +j dj dr
=

1
0
2
3
(r +j + 1)
3'2
u=1
u=0
dr =
2
3
1
0
(r + 2)
3'2
3(r + 1)
3'2
dr
=
2
3
2
5
(r + 2)
5'2
3
2
5
(r + 1)
5'2
1
0
=
4
15
(3
5'2
32
5'2
32
5'2
+ 1) =
4
15
(3
5'2
32
7'2
+ 1)
41. : = 1(r. j) = rj with 0 $ r
2
+j
2
$ 1, so 1
i
= j, 1
u
= r i
(o) =

T
1 +j
2
+r
2
d =

2r
0
1
0
I
v
2
+ 1 v dv d0 =

2r
0
1
3
(v
2
+ 1)
3'2
r=1
r=0
d0
=

2r
0
1
3
2
I
2 31
d0 =
2r
3
2
I
2 31
43. : = 1(r. j) = j
2
3r
2
with 1 $ r
2
+j
2
$ 4. Then
(o) =

T
1 + 4r
2
+ 4j
2
d =

2r
0
2
1
I
1 + 4v
2
v dv d0 =

2r
0
d0

2
1
v
I
1 + 4v
2
dv
=

0
2r
0
1
12
(1 + 4v
2
)
3'2
2
1
=
r
6
17
I
17 35
I
5
45. A parametric representation of the surface is r = r, j = 4r +:

2
, : = : with 0 $ r $ 1, 0 $ : $ 1.
Hence r
i
r
:
= (i + 4 j) (2: j +k) = 4 i 3j + 2: k.
Note: In general, if j = 1(r. :) then r
i
r
:
=
u1
ur
i 3j +
u1
u:
k and (o) =
1 +
u1
ur
2
+
u1
u:
2
d. Then
(o) =

1
0
1
0
I
17 + 4:
2
drd: =

1
0
I
17 + 4:
2
d:
=
1
2
:
I
17 + 4:
2
+
17
2
ln
2: +
I
4:
2
+ 17
1
0
=
I
21
2
+
17
4
ln
2 +
I
21
3ln
I
17
47. r
u
= h2&. . 0i, r
r
= h0. &. i, and r
u
r
r
=

2
. 32&. 2&
2
. Then
(o) =

T
|r
u
r
r
| d =

1
0
2
0
I
4
+ 4&
2
2
+ 4&
4
d d& =

1
0
2
0
(
2
+ 2&
2
)
2
d d&
=

1
0
2
0
(
2
+ 2&
2
) d d& =

1
0
1
3
3
+ 2&
2
r=2
r=0
d& =

1
0
8
3
+ 4&
2
d& =

8
3
& +
4
3
&
3
1
0
= 4
49. : = 1(r. j) = c
3i
2
3u
2
with r
2
+j
2
$ 4.
(o) =
1 +
32rc
3i
2
3u
2
2
+
32jc
3i
2
3u
2
2
d =

T
1 + 4(r
2
+j
2
)c
32(i
2
+u
2
)
d
=

2r
0
2
0
1 + 4v
2
c
32r
2
v dv d0 =

2r
0
d0

2
0
v
1 + 4v
2
c
32r
2
dv = 2
2
0
v
1 + 4v
2
c
32r
2
dv E 13.9783

291
51. (a) (o) =
1 +
u:
ur
2
+
u:
uj
2
d =

6
0
4
0
1 +
4r
2
+ 4j
2
(1 +r
2
+j
2
)
4
dj dr.
Using the Midpoint Rule with 1(r. j) =
1 +
4r
2
+ 4j
2
(1 +r
2
+j
2
)
4
, i = 3, n = 2 we have
(o) E
3
. =1
2
, =1
1
r
.
. j
,
{ = 4 [1(1. 1) +1(1. 3) +1(3. 1) +1(3. 3) +1(5. 1) +1(5. 3)] E 24.2055

(b) Using a CAS we have (o) =
6
0
4
0
1 +
4r
2
+ 4j
2
(1 +r
2
+j
2
)
4
dj dr E 24.2476. This agrees with the estimate in part (a)
to the rst decimal place.
53. : = 1 + 2r + 3j + 4j
2
, so
(o) =
1 +
u:
ur
2
+
u:
uj
2
d =

4
1
1
0
1 + 4 + (3 + 8j)
2
dj dr =

4
1
1
0
14 + 48j + 64j
2
dj dr.
Using a CAS, we have
4
1
1
0
14 + 48j + 64j
2
dj dr =
45
8
I
14 +
15
16
ln
11
I
5 + 3
I
14
I
5
3
15
16
ln
3
I
5 +
I
14
I
5
or
45
8
I
14 +
15
16
ln
11
I
5 +3
I
70
3
I
5 +
I
70
.
55. (a) r = o sin&cos , j = / sin&sin, : = c cos & i
r
2
o
2
+
j
2
/
2
+
:
2
c
2
= (sin&cos )
2
+ (sin&sin)
2
+ (cos &)
2
= sin
2
& + cos
2
& = 1
and since the ranges of & and are sufcient to generate the entire graph,
the parametric equations represent an ellipsoid.
(b)
(c) From the parametric equations (with o = 1, / = 2, and c = 3), we calculate
r
u
= cos & cos i + 2 cos & sin j 33 sin&k and r
r
= 3sin& sin i + 2 sin& cos j. So
r
u
r
r
= 6 sin
2
& cos i + 3 sin
2
& sin j + 2 sin& cos &k, and the surface area is given by
(o) =
2r
0
r
0
|r
u
r
r
| d&d =

2r
0
r
0
36 sin
4
& cos
2
+ 9 sin
4
& sin
2
+ 4 cos
2
& sin
2
&d&d
57. To nd the region 1: : = r
2
+j
2
implies : +:
2
= 4: or :
2
33: = 0. Thus : = 0 or : = 3 are the planes where the
surfaces intersect. But r
2
+j
2
+:
2
= 4: implies r
2
+j
2
+ (: 32)
2
= 4, so : = 3 intersects the upper hemisphere.
Thus (: 32)
2
= 4 3r
2
3j
2
or : = 2 +
4 3r
2
3j
2
. Therefore 1 is the region inside the circle r
2
+j
2
+(3 32)
2
= 4,
that is, 1 =

(r. j) | r
2
+j
2
$ 3
.
(o) =

T
1 + [(3r)(4 3r
2
3j
2
)
31'2
]
2
+ [(3j)(4 3r
2
3j
2
)
31'2
]
2
d
=

2r
0
I
3
0
1 +
v
2
4 3v
2
v dv d0 =

2r
0
I
3
0
2v dv
I
4 3v
2
d0 =

2r
0
32(4 3v
2
)
1'2
r=
I
3
r=0
d0
=

2r
0
(32 + 4) d0 = 20
2r
0
= 4
59. Let (o
1
) be the surface area of that portion of the surface which lies above the plane : = 0. Then (o) = 2(o
1
).
Following Example 10, a parametric representation of o
1
is r = o sinccos 0, j = o sincsin0,
: = o cos c and |r
c
r
0
| = o
2
sinc. For 1, 0 $ c $
r
2
and for each xed c,
r 3
1
2
o
2
+j
2
$
1
2
o
2
or
o sinccos 0 3
1
2
o
2
+o
2
sin
2
csin
2
0 $ (o2)
2
implies o
2
sin
2
c 3o
2
sinccos 0 $ 0 or
sinc(sinc 3cos 0) $ 0. But 0 $ c $
r
2
, so cos 0 D sinc or sin
r
2
+0
D sinc or c 3
r
2
$ 0 $
r
2
3c.
Hence 1 =

(c. 0) | 0 $ c $
r
2
, c 3
r
2
$ 0 $
r
2
3c
. Then
(o
1
) =

r'2
0
(r'2) 3c
c3(r'2)
o
2
sincd0 dc = o
2
r'2
0
( 32c) sincdc
= o
2
[(3 cos c) 32(3ccos c + sinc)]
r'2
0
= o
2
( 32)
Thus (o) = 2o
2
( 32).
Alternate solution: Working on o
1
we could parametrize the portion of the sphere by r = r, j = j, : =

o
2
3r
2
3j
2
.
Then |r
i
r
u
| =
1 +
r
2
o
2
3r
2
3j
2
+
j
2
o
2
3r
2
3j
2
=
o
o
2
3r
2
3j
2
and
(o
1
) =

0 $(i 3(a'2))
2
+u
2
$(a'2)
2
o
o
2
3r
2
3j
2
d =

r'2
3r'2
a cos 0
0
o
I
o
2
3v
2
v dv d0
=

r'2
3r'2
3o(o
2
3v
2
)
1'2
r =a cos 0
r =0
d0 =

r'2
3r'2
o
2
[1 3(1 3cos
2
0)
1'2
] d0
=

r'2
3r'2
o
2
(1 3|sin0|) d0 = 2o
2
r'2
0
(1 3sin0) d0 = 2o
2
r
2
31
Thus (o) = 4o
2
r
2
31
= 2o
2
( 32).
Notes:
(1) Perhaps working in spherical coordinates is the most obvious approach here. However, you must be careful
in setting up 1.
(2) In the alternate solution, you can avoid having to use |sin0| by working in the rst octant and then
multiplying by 4. However, if you set up o
1
as above and arrived at (o
1
) = o
2
, you now see your error.
17.7 Surface Integrals ET 16.7
1. The faces of the box in the planes r = 0 and r = 2 have surface area 24 and centers (0. 2. 3), (2. 2. 3). The faces in j = 0 and
j = 4 have surface area 12 and centers (1. 0. 3), (1. 4. 3), and the faces in : = 0 and : = 6 have area 8 and centers (1. 2. 0),
(1. 2. 6). For each face we take the point 1
W
.,
to be the center of the face and 1(r. j. :) = c
30.1(i+u+:)
, so by Denition 1,
S
1(r. j. :) do E [1(0. 2. 3)](24) + [1(2. 2. 3)](24) + [1(1. 0. 3)](12)
+ [1(1. 4. 3)](12) + [1(1. 2. 0)](8) + [1(1. 2. 6)](8)
= 24(c
30.5
+c
30.7
) + 12(c
30.4
+c
30.8
) + 8(c
30.3
+c
30.9
) E 49.09
SECTION 17.7 SURFACE INTEGRALS ET SECTION 16.7

293
3. We can use the r:- and j:-planes to divide 1 into four patches of equal size, each with surface area equal to
1
8
the surface
area of a sphere with radius
I
50, so {o =
1
8
(4)
I
50
2
= 25. Then (3. 4. 5) are sample points in the four patches,
and using a Riemann sum as in Denition 1, we have
1
1(r. j. :) do E 1(3. 4. 5) {o +1(3. 34. 5) {o +1(33. 4. 5) {o +1(33. 34. 5) {o
= (7 + 8 + 9 + 12)(25) = 900 E 2827
5. : = 1 + 2r + 3j so
u:
ur
= 2 and
u:
uj
= 3. Then by Formula 4,
S
r
2
j: do =

T
r
2
j:
u:
ur
2
+
u:
uj
2
+ 1 d =

3
0
2
0
r
2
j(1 + 2r + 3j)
I
4 + 9 + 1 dj dr
=
I
14
3
0
2
0
(r
2
j + 2r
3
j + 3r
2
j
2
) dj dr =
I
14
3
0
1
2
r
2
j
2
+r
3
j
2
+r
2
j
3
u=2
u=0
dr
=
I
14
3
0
(10r
2
+ 4r
3
) dr =
I
14
10
3
r
3
+r
4
3
0
= 171
I
14
7. o is the part of the plane : = 1 3r 3j over the region 1 = {(r. j) | 0 $ r $ 1. 0 $ j $ 1 3r}. Thus
S
j: do =

T
j(1 3r 3j)
(31)
2
+ (31)
2
+ 1 d =
I
3
1
0
13i
0
j 3rj 3j
2
dj dr
=
I
3
1
0
1
2
j
2
3
1
2
rj
2
3
1
3
j
3
u=13i
u=0
dr =
I
3
1
0
1
6
(1 3r)
3
dr = 3
I
3
24
(1 3r)
4
1
0
=
I
3
24
9. r(&. ) = &
2
i + &sin j + &cos k, 0 $ & $ 1, 0 $ $ 2, so
r
u
r
r
= (2&i + sin j + cos k) (&cos j 3&sin k) = 3&i + 2&
2
sin j + 2&
2
cos k and
|r
u
r
r
| =

&
2
+ 4&
4
sin
2
+ 4&
4
cos
2
=

&
2
+ 4&
4
(sin
2
+ cos
2
) = &
I
1 + 4&
2
(since & D 0). Then by
Formula 2,
S
j: do =

T
(&sin)(&cos ) |r
u
r
r
| d =

r'2
0
1
0
(&sin)(&cos ) &
I
1 + 4&
2
d&d
=

1
0
&
3
I
1 + 4&
2
d&

r'2
0
sin cos d

let t = 1 + 4&
2
i &
2
=
1
4
(t 31) and
1
8
dt = &d&
=

5
1
1
8

1
4
(t 31)
I
t dt

r'2
0
sin cos d =
1
32
5
1
t
3'2
3
I
t
dt

r'2
0
sin cos d
=
1
32
2
5
t
5'2
3
2
3
t
3'2
5
1
1
2
sin
2
r'2
0
=
1
32
2
5
(5)
5'2
3
2
3
(5)
3'2
3
2
5
+
2
3
1
2
(1 30) =
5
48
I
5 +
1
240
11. o is the portion of the cone :
2
= r
2
+j
2
for 1 $ : $ 3, or equivalently, o is the part of the surface : =

r
2
+j
2
over the
region 1 =

(r. j) | 1 $ r
2
+j
2
$ 9
. Thus
S
r
2
:
2
do =

T
r
2
(r
2
+j
2
)
r
2
+j
2
2
+
r
2
+j
2
2
+ 1 d
=

T
r
2
(r
2
+j
2
)
r
2
+j
2
r
2
+j
2
+ 1 d =

T
I
2 r
2
(r
2
+j
2
) d =
I
2
2r
0
3
1
(v cos 0)
2
(v
2
) v dv d0
=
I
2
2r
0
cos
2
0 d0

3
1
v
5
dv =
I
2
1
2
0 +
1
4
sin20
2r
0
1
6
v
6
3
1
=
I
2 ()
1
6
(3
6
31) =
364
I
2
3

13. Using r and : as parameters, we have r(r. :) = ri + (r
2
+:
2
) j +: k, r
2
+:
2
$ 4. Then
r
i
r
:
= (i + 2rj) (2: j +k) = 2ri 3j + 2: k and |r
i
r
:
| =
I
4r
2
+ 1 + 4:
2
=

1 + 4(r
2
+:
2
). Thus
S
j do =

i
2
+:
2
$4
(r
2
+:
2
)
1 + 4(r
2
+:
2
) d =

2r
0
2
0
v
2
I
1 + 4v
2
v dv d0 =

2r
0
d0

2
0
v
2
I
1 + 4v
2
v dv
= 2
2
0
v
2
I
1 + 4v
2
v dv

let & = 1 + 4v
2
i v
2
=
1
4
(& 31) and
1
8
d& = v dv
= 2
17
1
1
4
(& 31)
I
&
1
8
d& =
1
16
17
1
(&
3'2
3&
1'2
) d&
=
1
16
2
5
&
5'2
3
2
3
&
3'2
17
1
=
1
16
2
5
(17)
5'2
3
2
3
(17)
3'2
3
2
5
+
2
3
=

60
391
I
17 + 1
15. Using spherical coordinates and Example 17.6.10 [ ET 16.6.10] we have r(c. 0) = 2 sinccos 0 i + 2 sincsin0 j + 2 cos ck
and |r
c
r
0
| = 4 sinc. Then
S
(r
2
: +j
2
:) do =

2r
0
r'2
0
(4 sin
2
c)(2 cos c)(4 sinc) dcd0 = 16 sin
4
c
r'2
0
= 16.
17. o is given by r(&. ) = &i + cos j + sin k, 0 $ & $ 3, 0 $ $ 2. Then
r
u
r
r
= i (3sin j + cos k) = 3cos j 3sin k and |r
u
r
r
| =

cos
2
+ sin
2
= 1, so
S
(: +r
2
j) do =

r'2
0
3
0
(sin +&
2
cos )(1) d&d =

r'2
0
(3 sin + 9 cos ) d
= [33 cos + 9 sin]
r'2
0
= 0 + 9 + 3 30 = 12
19. F(r. j. :) = rj i +j: j +:rk, : = o(r. j) = 4 3r
2
3j
2
, and 1 is the square [0. 1] [0. 1], so by Equation 10
S
F dS =

T
[3rj(32r) 3j:(32j) +:r] d =

1
0
1
0
[2r
2
j + 2j
2
(4 3r
2
3j
2
) +r(4 3r
2
3j
2
)] dj dr
=

1
0
1
3
r
2
+
11
3
r 3r
3
+
34
15
dr =
713
180
21. F(r. j. :) = r:c
u
i 3r:c
u
j +: k, : = o(r. j) = 1 3r 3j, and 1 = {(r. j) | 0 $ r $ 1. 0 $ j $ 1 3r}. Since o has
downward orientation, we have
S
F dS = 3
T
[3r:c
u
(31) 3(3r:c
u
)(31) +:] d = 3
1
0
13i
0
(1 3r 3j) dj dr
= 3
1
0
1
2
r
2
3r +
1
2
dr = 3
1
6
23. F(r. j. :) = ri 3 : j + j k, : = o(r. j) =

4 3r
2
3j
2
and 1 is the quarter disk
(r. j)
0 $ r $ 2. 0 $ j $
I
4 3r
2
. o has downward orientation, so by Formula 10,
S
F dS = 3
3r
1
2
(4 3r
2
3j
2
)
31'2
(32r) 3(3:)
1
2
(4 3r
2
3j
2
)
31'2
(32j) +j
d
= 3
r
2
4 3r
2
3j
2
3
4 3r
2
3j
2

j
4 3r
2
3j
2
+j
d
= 3
T
r
2
(4 3(r
2
+j
2
))
31'2
d = 3
r'2
0
2
0
(v cos 0)
2
(4 3v
2
)
31'2
v dv d0
= 3
r'2
0
cos
2
0 d0

2
0
v
3
(4 3v
2
)
31'2
dv

let & = 4 3v
2
i v
2
= 4 3& and 3
1
2
d& = v dv
= 3
r'2
0
1
2
+
1
2
cos 20
d0

0
4
3
1
2
(4 3&)(&)
31'2
d&
= 3
1
2
0 +
1
4
sin20
r'2
0
3
1
2
8
I
& 3
2
3
&
3'2
0
4
= 3
r
4
3
1
2
316 +
16
3
= 3
4
3
SECTION 17.7 SURFACE INTEGRALS ET SECTION 16.7

295
25. Let o
1
be the paraboloid j = r
2
+:
2
, 0 $ j $ 1 and o
2
the disk r
2
+:
2
$ 1, j = 1. Since o is a closed
surface, we use the outward orientation.
On o
1
: F(r(r. :)) = (r
2
+:
2
) j 3: k and r
i
r
:
= 2ri 3j +2: k (since the j-component must be negative on o
1
). Then
S
1
F dS =

i
2
+:
2
$1
[3(r
2
+:
2
) 32:
2
] d = 3
2r
0
1
0
(v
2
+ 2v
2
cos
2
0) v dv d0
= 3
2r
0
1
4
(1 + 2 cos
2
0) d0 = 3
r
2
+
r
2
= 3
On o
2
: F(r(r. :)) = j 3: k and r
:
r
i
= j. Then
S
2
F dS =

i
2
+:
2
$1
(1) d = .
Hence
S
F dS = 3 + = 0.
27. Here o consists of the six faces of the cube as labeled in the gure. On o
1
:
F = i + 2j j + 3: k, r
u
r
:
= i and
S
1
F dS =

1
31
1
31
dj d: = 4;
o
2
: F = ri + 2 j + 3: k, r
:
r
i
= j and
S
2
F dS =

1
31
1
31
2 drd: = 8;
o
3
: F = ri + 2j j + 3 k, r
i
r
u
= k and
S
3
F dS =

1
31
1
31
3 drdj = 12;
o
4
: F = 3i + 2j j + 3: k, r
:
r
u
= 3i and
S
4
F dS = 4;
o
5
: F = ri 32 j + 3: k, r
i
r
:
= 3j and
S
5
F dS = 8;
o
6
: F = ri + 2j j 3 3 k, r
u
r
i
= 3k and
S
6
F dS =

1
31
1
31
3 drdj = 12.
Hence
S
F dS =

6
.=1
S
l
F dS = 48.
29. Here o consists of four surfaces: o
1
, the top surface (a portion of the circular cylinder j
2
+:
2
= 1); o
2
, the bottom surface
(a portion of the rj-plane); o
3
, the front half-disk in the plane r = 2, and o
4
, the back half-disk in the plane r = 0.
On o
1
: The surface is : =

1 3j
2
for 0 $ r $ 2, 31 $ j $ 1 with upward orientation, so
S
1
F dS =

2
0
1
31
3r
2
(0) 3j
2
3
j
1 3j
2
+:
2
dj dr =

2
0
1
31
j
3
1 3j
2
+ 1 3j
2
dj dr
=

2
0
1 3j
2
+
1
3
(1 3j
2
)
3'2
+j 3
1
3
j
3
u=1
u=31
dr =

2
0
4
3
dr =
8
3
On o
2
: The surface is : = 0 with downward orientation, so
S
2
F dS =
2
0
1
31
3:
2
dj dr =

2
0
1
31
(0) dj dr = 0
On o
3
: The surface is r = 2 for 31 $ j $ 1, 0 $ : $
1 3j
2
, oriented in the positive r-direction. Regarding j and : as
parameters, we have r
u
r
:
= i and
S
3
F dS =
1
31
I
13u
2
0
r
2
d: dj =

1
31
I
13u
2
0
4 d: dj = 4(o
3
) = 2
On o
4
: The surface is r = 0 for 31 $ j $ 1, 0 $ : $
1 3j
2
, oriented in the negative r-direction. Regarding j and : as
parameters, we use 3(r
u
r
:
) = 3i and
S
4
F dS =

1
31
I
13u
2
0
r
2
d: dj =

1
31
I
13u
2
0
(0) d: dj = 0
Thus
S
F dS =
8
3
+ 0 + 2 + 0 = 2 +
8
3
.
31. : = rj i u:ur = j, u:uj = r, so by Formula 4, a CAS gives
S
rj: do =

1
0
1
0
rj(rj)
j
2
+r
2
+ 1 drdj E 0.1642.
33. We use Formula 4 with : = 3 32r
2
3j
2
i u:ur = 34r, u:uj = 32j. The boundaries of the region
3 32r
2
3j
2
D 0 are 3
3
2
$ r $
3
2
and 3
I
3 32r
2
$ j $
I
3 32r
2
, so we use a CAS (with precision reduced to
seven or fewer digits; otherwise the calculation may take a long time) to calculate
S
r
2
j
2
:
2
do =

I
3'2
3
I
3'2
I
3 32i
2
3
I
3 32i
2
r
2
j
2
(3 32r
2
3j
2
)
2
16r
2
+ 4j
2
+ 1 dj dr E 3.4895
35. If o is given by j = /(r. :), then o is also the level surface 1(r. j. :) = j 3 /(r. :) = 0.
n =
Q1(r. j. :)
|Q1(r. j. :)|
=
3/
i
i +j 3/
:
k
I
/
2
i
+ 1 +/
2
:
, and 3n is the unit normal that points to the left. Now we proceed as in the
derivation of (10), using Formula 4 to evaluate
S
F dS =

S
F ndo =

T
(1 i +Qj +1k)
u/
ur
i 3j +
u/
u:
k
u/
ur
2
+ 1 +
u/
u:
u/
ur
2
+ 1 +
u/
u:
2
d
where 1 is the projection of o onto the r:-plane. Therefore
S
F dS =

T
1
u/
ur
3Q+1
u/
u:
d.
37. i =

S
1do = 1 4
1
2
o
2
= 2o
2
1; by symmetry A
i:
= A
u:
= 0, and
A
iu
=

S
:1do = 1
2r
0
r'2
0
(o cos c)(o
2
sinc) dcd0 = 21o
3
3
1
4
cos 2c
r'2
0
= 1o
3
.
Hence (r. j. :) =
0. 0.
1
2
o
.
39. (a) 1
:
=

S
(r
2
+j
2
)a(r. j. :) do
(b) 1
:
=

S
(r
2
+j
2
)
10 3
r
2
+j
2
do =

1 $i
2
+u
2
$16
(r
2
+j
2
)
10 3
r
2
+j
2
I
2 d
=

2r
0
4
1
I
2 (10v
3
3v
4
) dv d0 = 2
I
2
4329
10
=
4329
5
I
2
41. The rate of ow through the cylinder is the ux
S
av ndo =

S
av dS. We use the parametric representation
r(&. ) = 2 cos &i + 2 sin&j + k for o, where 0 $ & $ 2, 0 $ $ 1, so r
u
= 32 sin&i + 2 cos &j, r
r
= k, and the
outward orientation is given by r
u
r
r
= 2 cos &i + 2 sin&j. Then
S
av dS = a
2r
0
1
0
i + 4 sin
2
&j + 4 cos
2
&k
(2 cos &i + 2 sin&j) d d&

= a
2r
0
1
0
2 cos & + 8 sin

3
&
d d& = a
2r
0
cos & + 8 sin

3
&
d&
= a
sin& + 8
3
1
3
(2 + sin
2
&) cos &
2r
0
= 0 kgs
SECTION 17.8 STOKES THEOREM ET SECTION 16.8

297
43. o consists of the hemisphere o
1
given by : =

o
2
3r
2
3j
2
and the disk o
2
given by 0 $ r
2
+j
2
$ o
2
, : = 0.
On o
1
: E = o sinc cos 0 i + o sinc sin0 j + 2o cos ck,
T
c
T
0
= o
2
sin
2
c cos 0 i +o
2
sin
2
c sin0 j +o
2
sinc cos ck. Thus
S
1
E dS =

2r
0
r'2
0
(o
3
sin
3
c + 2o
3
sinc cos
2
c) dcd0
=

2r
0
r'2
0
(o
3
sinc +o
3
sinc cos
2
c) dcd0 = (2)o
3
1 +
1
3
=
8
3
o
3
On o
2
: E = ri +j j, and r
u
r
i
= 3k so
S
2
E dS = 0. Hence the total charge is q = -
0
S
E dS =
8
3
o
3
-
0
.
45. 1Q& = 6.5(4j j + 4: k). o is given by r(r. 0) = ri +
I
6 cos 0 j +
I
6 sin0 k and since we want the inward heat ow, we
use r
i
r
0
= 3
I
6 cos 0 j 3
I
6 sin0 k. Then the rate of heat ow inward is given by
S
(31Q&) dS =

2r
0
4
0
3(6.5)(324) drd0 = (2)(156)(4) = 1248.
47. Let o be a sphere of radius o centered at the origin. Then |r| = o and F(r) = cr |r|
3
=

co
3
(ri +j j +: k). A
parametric representation for o is r(c. 0) = o sinc cos 0 i +o sinc sin0 j +o cos ck, 0 $ c $ , 0 $ 0 $ 2. Then
r
c
= o cos c cos 0 i +o cos c sin0 j 3o sinck, r
0
= 3o sinc sin0 i +o sinc cos 0 j, and the outward orientation is given
by r
c
r
0
= o
2
sin
2
c cos 0 i +o
2
sin
2
c sin0 j +o
2
sinc cos ck. The ux of F across o is
S
F dS =

r
0
2r
0
c
o
3
(o sinc cos 0 i +o sinc sin0 j +o cos ck)
o
2
sin
2
c cos 0 i +o
2
sin
2
c sin0 j +o
2
sinc cos ck
d0 dc
=
c
o
3
r
0
2r
0
o
3
sin
3
c + sinc cos
2
c
d0 dc = c
r
0
2r
0
sincd0 dc = 4c
Thus the ux does not depend on the radius o.
17.8 Stokes' Theorem ET 16.8
1. Both 1 and 1 are oriented piecewise-smooth surfaces that are bounded by the simple, closed, smooth curve r
2
+j
2
= 4,
: = 0 (which we can take to be oriented positively for both surfaces). Then 1 and 1 satisfy the hypotheses of Stokes
Theorem, so by (3) we know
1
curl F dS =

C
F dr =

T
curl F dS (where C is the boundary curve).
3. The paraboloid : = r
2
+j
2
intersects the cylinder r
2
+j
2
= 4 in the circle r
2
+j
2
= 4, : = 4. This boundary curve C
should be oriented in the counterclockwise direction when viewed from above, so a vector equation of C is
r(t) = 2 cos t i + 2 sint j + 4 k, 0 $ t $ 2. Then r
0
(t) = 32 sint i + 2 cos t j,
F(r(t)) = (4 cos
2
t)(16) i + (4 sin
2
t)(16) j + (2 cos t)(2 sint)(4) k = 64 cos
2
t i + 64 sin
2
t j + 16 sint cos t k.
and by Stokes Theorem,
S
curl F dS =

C
F dr =

2r
0
F(r(t)) r
0
(t) dt =

2r
0
(3128 cos
2
t sint + 128 sin
2
t cos t + 0) dt
= 128
1
3
cos
3
t +
1
3
sin
3
t
2r
0
= 0
5. C is the square in the plane : = 31. By (3),
S
1
curl F dS =

C
F dr =

S
2
curl F dS where o
1
is the original cube
without the bottom and o
2
is the bottom face of the cube. curl F = r
2
: i + (rj 32rj:) j + (j 3r:) k. For o
2
, we choose
n = k so that C has the same orientation for both surfaces. Then curl F n = j 3r: = r +j on o
2
, where : = 31. Thus
S
2
curl F dS =

1
31
1
31
(r +j) drdj = 0 so
S
1
curl F dS = 0.
7. curl F = 32: i 32rj 32j k and we take the surface o to be the planar region enclosed by C, so o is the portion of the plane
r +j +: = 1 over 1 = {(r. j) | 0 $ r $ 1, 0 $ j $ 1 3r}. Since C is oriented counterclockwise, we orient o upward.
Using Equation 17.7.10 [ ET 16.7.10], we have : = o(r. j) = 1 3r 3j, 1 = 32:, Q = 32r, 1 = 32j, and
C
F dr =

S
curl F dS =

T
[3(32:)(31) 3(32r)(31) + (32j)] d
=

1
0
13i
0
(32) dj dr = 32
1
0
(1 3r) dr = 31
9. curl F = (rc
iu
32r) i 3(jc
iu
3j) j + (2: 3:) k and we take o to be the disk r
2
+j
2
$ 16, : = 5. Since C is oriented
counterclockwise (from above), we orient o upward. Then n = k and curl F n = 2: 3: on o, where : = 5. Thus
F dr =
S
curl F ndo=
S
(2: 3:) do =

S
(10 35) do = 5(area of o) = 5( 4
2
) = 80
11. (a) The curve of intersection is an ellipse in the plane r +j +: = 1 with unit normal n =
1
I
3
(i +j +k),
curl F = r
2
j +j
2
k, and curl F n =
1
I
3
(r
2
+j
2
). Then
C
F dr =

S
1
I
3
r
2
+j
2
do =

i
2
+u
2
$9
r
2
+j
2
drdj =

2r
0
3
0
v
3
dv d0 = 2
81
4
=
81r
2
(b) (c) One possible parametrization is r = 3 cos t, j = 3 sint,
: = 1 33 cos t 33 sint, 0 $ t $ 2.
13. The boundary curve C is the circle r
2
+j
2
= 1, : = 1 oriented in the counterclockwise direction as viewed from above.
We can parametrize C by r(t) = cos t i + sint j +k, 0 $ t $ 2, and then r
0
(t) = 3sint i + cos t j. Thus
F(r(t)) = sin
2
t i + cos t j +k, F(r(t)) r
0
(t) = cos
2
t 3sin
3
t, and
C
F dr =

2r
0
(cos
2
t 3sin
3
t) dt =

2r
0
1
2
(1 + cos 2t) dt 3
2r
0
(1 3cos
2
t) sint dt
=
1
2
t +
1
2
sin2t
2r
0
3
3cos t +
1
3
cos
3
t
2r
0
=
Now curl F = (1 32j) k, and the projection 1 of o on the rj-plane is the disk r
2
+j
2
$ 1, so by Equation 17.7.10
[ ET 16.7.10] with : = o(r. j) = r
2
+ j
2
we have
S
curl F dS =

T
(1 32j) d =

2r
0
1
0
(1 32v sin0) v dv d0 =

2r
0
1
2
3
2
3
sin0
d0 = .
15. The boundary curve C is the circle r
2
+:
2
= 1, j = 0 oriented in the counterclockwise direction as viewed from the positive
j-axis. Then C can be described by r(t) = cos t i 3sint k, 0 $ t $ 2, and r
0
(t) = 3sint i 3cos t k. Thus
SECTION 17.9 THE DIVERGENCE THEOREM ET SECTION 16.9

299
F(r(t)) = 3sint j + cos t k, F(r(t)) r
0
(t) = 3cos
2
t, and
C
F dr =

2r
0
3cos
2
t dt = 3
1
2
t 3
1
4
sin2t
2r
0
= 3.
Now curl F = 3i 3j 3k, and o can be parametrized (see Example 17.6.10 [ ET 16.6.10]) by
r(c. 0) = sinc cos 0 i + sinc sin0 j + cos ck, 0 $ 0 $ , 0 $ c $ . Then
r
c
r
0
= sin
2
c cos 0 i + sin
2
c sin0 j + sinc cos ck and
S
curl F dS =

i
2
+:
2
$1
curl F (r
c
r
0
) d =

r
0
r
0
(3sin
2
c cos 0 3sin
2
c sin0 3sinc cos c) d0 dc
=

r
0
(32 sin
2
c 3 sinc cos c) dc =

1
2
sin2c 3c 3
r
2
sin
2
c
r
0
= 3
17. It is easier to use Stokes Theorem than to compute the work directly. Let o be the planar region enclosed by the path of the
particle, so o is the portion of the plane : =
1
2
j for 0 $ r $ 1, 0 $ j $ 2, with upward orientation.
curl F = 8j i + 2: j + 2j k and
C
F dr =

S
curl F dS =

T
38j (0) 32:
1
2
+ 2j
d =

1
0
2
0
2j 3
1
2
j
dj dr
=

1
0
2
0
3
2
j dj dr =

1
0
3
4
j
2
u=2
u=0
dr =

1
0
3 dr = 3
19. Assume o is centered at the origin with radius o and let 1
1
and 1
2
be the upper and lower hemispheres, respectively, of o.
Then
S
curl F dS =

1
1
curl F dS +
1
2
curl F dS =

C
1
F dr +
C
2
F dr by Stokes Theorem. But C
1
is the
circle r
2
+j
2
= o
2
oriented in the counterclockwise direction while C
2
is the same circle oriented in the clockwise direction.
Hence
C
2
F dr = 3
C
1
F dr so
S
curl F dS = 0 as desired.
17.9 The Divergence Theorem ET 16.9
1. div F = 3 +r + 2r = 3 + 3r, so
T
div Fd\ =

1
0
1
0
1
0
(3r + 3) drdj d: =
9
2
(notice the triple integral is
three times the volume of the cube plus three times r).
To compute
S
F dS, on
o
1
: n = i, F = 3 i +j j + 2: k, and
S
1
F dS =

S
1
3 do = 3;
o
2
: F = 3ri +rj + 2r: k, n = j and
S
2
F dS =

S
2
rdo =
1
2
;
o
3
: F = 3ri +rj j + 2rk, n = k and
S
3
F dS =

S
3
2rdo = 1;
o
4
: F = 0,
S
4
F dS = 0; o
5
: F = 3ri + 2rk, n = 3j and
S
5
F dS =

S
5
0 do = 0;
o
6
: F = 3ri +rj j, n = 3k and
S
6
F dS =

S
6
0 do = 0. Thus
S
F dS =
9
2
.
3. div F = r +j +:, so
T
div Fd\ =

2r
0
1
0
1
0
(v cos 0 +v sin0 +:) v d: dv d0 =

2r
0
1
0
v
2
cos 0 +v
2
sin0 +
1
2
v
dv d0
=

2r
0
1
3
cos 0 +
1
3
sin0 +
1
4
d0 =
1
4
(2) =
r
2
Let o
1
be the top of the cylinder, o
2
the bottom, and o
3
the vertical edge. On o
1
, : = 1, n = k, and F = rj i +j j +rk, so
S
1
F dS =

S
1
F ndo =

S
1
rdo =

2r
0
1
0
(v cos 0) v dv d0 =

sin0
2r
0
1
3
v
3
1
0
= 0.
On o
2
, : = 0, n = 3k, and F = rj i so
S
2
F dS =

S
2
0 do = 0.
o
3
is given by r(0. :) = cos 0 i + sin0 j +: k, 0 $ 0 $ 2, 0 $ : $ 1. Then r
0
r
:
= cos 0 i + sin0 j and
S
3
F dS =

T
F (r
0
r
:
) d =

2r
0
1
0
(cos
2
0 sin0 +: sin
2
0) d: d0
=

2r
0
cos
2
0 sin0 +
1
2
sin
2
0
d0 =

3
1
3
cos
3
0 +
1
4
0 3
1
2
sin20
2r
0
=
r
2
Thus
S
F dS = 0 + 0 +
r
2
=
r
2
.
5. div F =
O
Oi
(c
i
sinj) +
O
Ou
(c
i
cos j) +
O
O:
(j:
2
) = c
i
sinj 3c
i
sinj + 2j: = 2j:, so by the Divergence Theorem,
S
F dS =

T
div Fd\ =

1
0
1
0
2
0
2j: d: dj dr = 2
1
0
dr

1
0
j dj

1
0
: d: = 2
1
0
1
2
j
2
1
0
1
2
:
2
2
0
= 2.
7. div F = 3j
2
+ 0 + 3:
2
, so using cylindrical coordinates with j = v cos 0, : = v sin0, r = r we have
S
F dS =

T
(3j
2
+ 3:
2
) d\ =

2r
0
1
0
2
31
(3v
2
cos
2
0 + 3v
2
sin
2
0) v drdv d0
= 3
2r
0
d0

1
0
v
3
dv

2
31
dr = 3(2)
1
4
(3) =
9r
2
9. div F = j sin: + 0 3j sin: = 0, so by the Divergence Theorem,
S
F dS =

T
0 d\ = 0.
11. div F = j
2
+ 0 +r
2
= r
2
+j
2
so
S
F dS =

T
(r
2
+j
2
) d\ =

2r
0
2
0
4
r
2
v
2
v d: dv d0 =

2r
0
2
0
v
3
(4 3v
2
) dv d0
=

2r
0
d0

2
0
(4v
3
3v
5
) dv = 2
v
4
3
1
6
v
6
2
0
=
32
3

13. div F = 12r
2
: + 12j
2
: + 12:
3
so
S
F dS =

T
12:(r
2
+j
2
+:
2
) d\ =

2r
0
r
0
T
0
12(a cos c)(a
2
)a
2
sincda dcd0
= 12
2r
0
d0

r
0
sinc cos cdc

T
0
a
5
da = 12(2)
1
2
sin
2
c
r
0
1
6
a
6
T
0
= 0
15.

S
F dS =

T
I
3 3r
2
d\ =

1
31
1
31
2 3i
4
3u
4
0
I
3 3r
2
d: dj dr =
341
60
I
2 +
81
20
sin
31
I
3
3
17. For o
1
we have n = 3k, so F n = F (3k) = 3r
2
: 3j
2
= 3j
2
(since : = 0 on o
1
). So if 1 is the unit disk, we get
S
1
F dS =

S
1
F ndo =

T
(3j
2
) d = 3
2r
0
1
0
v
2
(sin
2
0) v dv d0 = 3
1
4
. Now since o
2
is closed, we can use
the Divergence Theorem. Since div F =
O
Oi
(:
2
r) +
O
Ou
1
3
j
3
+ tan:
+
O
O:
(r
2
: +j
2
) = :
2
+j
2
+r
2
, we use spherical
coordinates to get
S
2
F dS =

T
div Fd\ =

2r
0
r'2
0
1
0
a
2
a
2
sincda dcd0 =
2
5
. Finally
S
F dS =

S
2
F dS 3
S
1
F dS =
2
5
3
3
1
4
=
13
20
.
19. The vectors that end near 1
1
are longer than the vectors that start near 1
1
, so the net ow is inward near 1
1
and div F(1
1
) is
negative. The vectors that end near 1
2
are shorter than the vectors that start near 1
2
, so the net ow is outward near 1
2
and
div F(1
2
) is positive.
21. From the graph it appears that for points above the r-axis, vectors starting near a
particular point are longer than vectors ending there, so divergence is positive.
The opposite is true at points below the r-axis, where divergence is negative.
F(r. j) =
rj. r +j
2
i div F =
O
Oi
(rj) +
O
Ou
r +j
2
= j + 2j = 3j.
Thus div F 0 for j 0, and div F < 0 for j < 0.
23. Since
x
|x|
3
=
ri +j j +: k
(r
2
+j
2
+:
2
)
3'2
and
u
ur
r
(r
2
+j
2
+:
2
)
3'2
=
(r
2
+j
2
+:
2
) 33r
2
(r
2
+j
2
+:
2
)
5'2
with similar expressions
for
u
uj
j
(r
2
+j
2
+:
2
)
3'2
and
u
u:
:
(r
2
+j
2
+:
2
)
3'2
, we have
div
x
|x|
3
=
3(r
2
+j
2
+:
2
) 33(r
2
+j
2
+:
2
)
(r
2
+j
2
+:
2
)
5'2
= 0, except at (0. 0. 0) where it is undened.
25.

S
a ndo =

T
div ad\ = 0 since div a = 0.
27.

S
curl F dS =

T
div(curl F) d\ = 0 by Theorem 17.5.11 [ ET 16.5.11].
29.

S
(1Qo) ndo =

T
div(1Qo) d\ =

T
(1Q
2
o +Qo Q1) d\ by Exercise 17.5.25 [ ET 16.5.25].
31. If c = c
1
i +c
2
j +c
3
k is an arbitrary constant vector, we dene F = 1c = 1c
1
i +1c
2
j +1c
3
k. Then
div F = div 1c =
u1
ur
c
1
+
u1
uj
c
2
+
u1
u:
c
3
= Q1 c and the Divergence Theorem says
S
F dS =

T
div Fd\ i
S
F ndo =

T
Q1 c d\ . In particular, if c = i then
S
1i ndo =

T
Q1 i d\ i
S
1n
1
do =

T
u1
ur
d\ (where n = n
1
i +n
2
j +n
3
k). Similarly, if c = j we have
S
1n
2
do =

T
u1
uj
d\ ,
and c = k gives
S
1n
3
do =

T
u1
u:
d\ . Then
S
1ndo =

S
1n
1
do
i +
S
1n
2
do
j +
S
1n
3
do
k
=

T
u1
ur
d\
i +
T
u1
uj
d\
j +
T
u1
u:
d\
k =

T
u1
ur
i +
u1
uj
j +
u1
u:
k
d\
=

T
Q1 d\ as desired.
17 Review ET 16
1. See Denitions 1 and 2 in Section 17.1 [ ET 16.1]. A vector eld can represent, for example, the wind velocity at any location
in space, the speed and direction of the ocean current at any location, or the force vectors of Earths gravitational eld at a
location in space.
2. (a) A conservative vector eld F is a vector eld which is the gradient of some scalar function 1.
(b) The function 1 in part (a) is called a potential function for F. that is, F = Q1.
3. (a) See Denition 17.2.2 [ ET 16.2.2].
(b) We normally evaluate the line integral using Formula 17.2.3 [ ET 16.2.3].
(c) The mass is i =

C
a (r. j) dc, and the center of mass is (r. j) where r =
1
r
C
ra (r. j) dc, j =
1
r
C
ja (r. j) dc.
(d) See (5) and (6) in Section 17.2 [ ET 16.2] for plane curves; we have similar denitions when C is a space curve
(see the equation preceding (10) in Section 17.2 [ ET 16.2]).
(e) For plane curves, see Equations 17.2.7 [ ET 16.2.7]. We have similar results for space curves
(see the equation preceding (10) in Section 17.2 [ ET 16.2]).
4. (a) See Denition 17.2.13 [ ET 16.2.13].
(b) If F is a force eld,
C
F dr represents the work done by F in moving a particle along the curve C.
(c)

C
F dr =

C
1 dr +Qdj +1d:
5. See Theorem 17.3.2 [ ET 16.3.2].
6. (a)

C
F dr is independent of path if the line integral has the same value for any two curves that have the same initial and
terminal points.
(b) See Theorem 17.3.4 [ ET 16.3.4].
7. See the statement of Greens Theorem on page 1091 [ ET 1055].
8. See Equations 17.4.5 [ ET 16.4.5].
9. (a) curl F =

u1
uj
3
uQ
u:
i +
u1
u:
3
u1
ur
j +
uQ
ur
3
u1
uj
k = QF
(b) div F =
u1
ur
+
uQ
uj
+
u1
u:
= Q F
(c) For curl F, see the discussion accompanying Figure 1 on page 1100 [ ET 1064] as well as Figure 6 and the accompanying
discussion on page 1132 [ ET 1096]. For div F, see the discussion following Example 5 on page 1102 [ ET 1066] as well
as the discussion preceding (8) on page 1139 [ ET 1103].
10. See Theorem 17.3.6 [ ET 16.3.6]; see Theorem 17.5.4 [ ET 16.5.4].
11. (a) See (1) and (2) and the accompanying discussion in Section 17.6 [ ET 16.6] ; See Figure 4 and the accompanying
discussion on page 1107 [ ET 1071] .
(b) See Denition 17.6.6 [ ET 16.6.6 ].
(c) See Equation 17.6.9 [ ET 16.6.9].
12. (a) See (1) in Section 17.7 [ ET 16.7].
(b) We normally evaluate the surface integral using Formula 17.7.2 [ ET 16.7.2].
(c) See Formula 17.7.4 [ ET 16.7.4].
(d) The mass is i =

S
a(r. j. :) do and the center of mass is (r. j. :) where r =
1
r
S
ra(r. j. :) do,
j =
1
r
S
ja(r. j. :) do, : =
1
r
S
:a(r. j. :) do.
13. (a) See Figures 6 and 7 and the accompanying discussion in Section 17.7 [ ET 16.7]. A Mbius strip is a nonorientable
surface; see Figures 4 and 5 and the accompanying discussion on page 1121 [ ET 1085].
(b) See Denition 17.7.8 [ ET 16.7.8].
(c) See Formula 17.7.9 [ ET 16.7.9].
(d) See Formula 17.7.10 [ ET 16.7.10].
14. See the statement of Stokes Theorem on page 1129 [ ET 1093.].
15. See the statement of the Divergence Theorem on page 1135 [ ET 1099].
16. In each theorem, we have an integral of a derivative over a region on the left side, while the right side involves the values of
the original function only on the boundary of the region.
1. False; div F is a scalar eld.
3. True, by Theorem 17.5.3 [ ET 16.5.3] and the fact that div 0 = 0.
5. False. See Exercise 17.3.33 [ ET 16.3.33]. (But the assertion is true if 1 is simply-connected; see Theorem 17.3.6
[ ET 16.3.6].)
7. True. Apply the Divergence Theorem and use the fact that div F = 0.
1. (a) Vectors starting on C point in roughly the direction opposite to C, so the tangential component F Tis negative.
Thus
C
F dr =

C
F Tdc is negative.
(b) The vectors that end near 1 are shorter than the vectors that start near 1, so the net ow is outward near 1 and
div F(1) is positive.
3.

C
j: cos rdc =

r
0
(3 cos t) (3 sint) cos t
(1)
2
+ (33 sint)
2
+ (3 cos t)
2
dt =

r
0
(9 cos
2
t sint)
I
10 dt
= 9
I
10
3
1
3
cos
3
t
r
0
= 33
I
10 (32) = 6
I
10
5.

C
j
3
dr +r
2
dj =

1
31
j
3
(32j) + (1 3j
2
)
2
dj =

1
31
(3j
4
32j
2
+ 1) dj
=

3
1
5
j
5
3
2
3
j
3
+j
1
31
= 3
1
5
3
2
3
+ 1 3
1
5
3
2
3
+ 1 =
4
15
7. C: r = 1 + 2t i dr = 2 dt, j = 4t i dj = 4 dt, : = 31 + 3t i d: = 3 dt, 0 $ t $ 1.
C
rj dr +j
2
dj +j: d: =

1
0
[(1 + 2t)(4t)(2) + (4t)
2
(4) + (4t)(31 + 3t)(3)] dt
=

1
0
(116t
2
34t) dt =

116
3
t
3
32t
2
1
0
=
116
3
32 =
110
3
9. F(r(t)) = c
3I
i +t
2
(3t) j + (t
2
+t
3
) k, r
0
(t) = 2t i + 3t
2
j 3k and
C
F dr =

1
0
(2tc
3I
33t
5
3(t
2
+t
3
)) dt =

32tc
3I
32c
3I
3
1
2
t
6
3
1
3
t
3
3
1
4
t
4
1
0
=
11
12
3
4
c
.
11.
O
Ou
[(1 +rj)c
iu
] = 2rc
iu
+r
2
jc
iu
=
O
Oi
c
u
+r
2
c
iu
and the domain of F is R

2
, so F is conservative. Thus there
exists a function 1 such that F = Q1. Then 1
u
(r. j) = c
u
+r
2
c
iu
implies 1(r. j) = c
u
+rc
iu
+o(r) and then
1
i
(r. j) = rjc
iu
+c
iu
+o
0
(r) = (1 +rj)c
iu
+o
0
(r). But 1
i
(r. j) = (1 +rj)c
iu
, so o
0
(r) = 0 i o(r) = 1.
Thus 1(r. j) = c
u
+rc
iu
+1 is a potential function for F.
13. Since
O
Ou
(4r
3
j
2
32rj
3
) = 8r
3
j 36rj
2
=
O
Oi
(2r
4
j 33r
2
j
2
+ 4j
3
) and the domain of F is R
2
, F is conservative.
Furthermore 1(r. j) = r
4
j
2
3r
2
j
3
+j
4
is a potential function for F. t = 0 corresponds to the point (0. 1) and t = 1
corresponds to (1. 1), so
C
F dr = 1(1. 1) 31(0. 1) = 1 31 = 0.
15. C
1
: r(t) = t i +t
2
j, 31 $ t $ 1;
C
2
: r(t) = 3t i +j, 31 $ t $ 1.
Then
C
rj
2
dr 3r
2
j dj =

1
31
(t
5
32t
5
) dt +
1
31
t dt
=

3
1
6
t
6
1
31
+
1
2
t
2
1
31
= 0
Using Greens Theorem, we have
C
rj
2
dr 3r
2
j dj =

T
u
ur
(3r
2
j) 3
u
uj
(rj
2
)
d =

T
(32rj 32rj) d =

1
31
1
i
2
34rj dj dr
=

1
31
32rj
2
u=1
u=i
2
dr =

1
31
(2r
5
32r) dr =

1
3
r
6
3r
2
1
31
= 0
17.

C
r
2
j dr 3rj
2
dj =

i
2
+u
2
$4
O
Oi
(3rj
2
) 3
O
Ou
(r
2
j)
d =

i
2
+u
2
$4
(3j
2
3r
2
) d = 3
2r
0
2
0
v
3
dv d0 = 38
19. If we assume there is such a vector eld G, then div(curl G) = 2 + 3: 32r:. But div(curl F) = 0 for all vector elds F.
Thus such a Gcannot exist.
21. For any piecewise-smooth simple closed plane curve C bounding a region 1, we can apply Greens Theorem to
F(r. j) = 1(r) i +o(j) j to get
C
1(r) dr +o(j) dj =

T
O
Oi
o(j) 3
O
Ou
1(r)
d =

T
0 d = 0.
23. Q
2
1 = 0 means that
u
2
1
ur
2
+
u
2
1
uj
2
= 0. Now if F = 1
u
i 31
i
j and C is any closed path in 1, then applying Greens
Theorem, we get
C
F dr =

C
1
u
dr 31
i
dj =

T
O
Oi
(31
i
) 3
O
Ou
(1
u
)
d = 3
T
(1
ii
+1
uu
) d = 3
T
0 d = 0
Therefore the line integral is independent of path, by Theorem 17.3.3 [ ET 16.3.3].
25. : = 1 (r. j) = r
2
+ 2j with 0 $ r $ 1, 0 $ j $ 2r. Thus
(o) =
T
I
1 + 4r
2
+ 4 d =

1
0
2i
0
I
5 + 4r
2
dj dr =

1
0
2r
I
5 + 4r
2
dr =
1
6
(5 + 4r
2
)
3'2
1
0
=
1
6
27 35
I
5
.
27. : = 1(r. j) = r
2
+j
2
with 0 $ r
2
+j
2
$ 4 so r
i
r
u
= 32ri 32j j +k (using upward orientation). Then
S
: do =

i
2
+u
2
$4
(r
2
+j
2
)
4r
2
+ 4j
2
+ 1 d =

2r
0
2
0
v
3
I
1 + 4v
2
dv d0 =
1
60
391
I
17 + 1
(Substitute & = 1 + 4v
2
and use tables.)
29. Since the sphere bounds a simple solid region, the Divergence Theorem applies and
S
F dS =

T
(: 32) d\ =

T
: d\ 32
T
d\ = i: 32
4
3
2
3
= 3
64
3
.
Alternate solution: F(r(c. 0)) = 4 sinc cos 0 cos ci 34 sinc sin0 j + 6 sinc cos 0 k,
r
c
r
0
= 4 sin
2
c cos 0 i + 4 sin
2
c sin0 j + 4 sinc cos ck, and
F (r
c
r
0
) = 16 sin
3
c cos
2
0 cos c 316 sin
3
c sin
2
0 + 24 sin
2
c cos c cos 0. Then
S
1 do =

2r
0
r
0
(16 sin
3
c cos c cos
2
0 316 sin
3
c sin
2
0 + 24 sin
2
c cos c cos 0) dcd0
=

2r
0
4
3
(316 sin
2
0) d0 = 3
64
3

31. Since curl F = 0,
S
(curl F) dS = 0. We parametrize C: r(t) = cos t i + sint j, 0 $ t $ 2 and
C
F dr =

2r
0
(3cos
2
t sint + sin
2
t cos t) dt =
1
3
cos
3
t +
1
3
sin
3
t
2r
0
= 0.
33. The surface is given by r +j +: = 1 or : = 1 3r 3j, 0 $ r $ 1, 0 $ j $ 1 3r and r
i
r
u
= i +j +k. Then
C
F dr =

S
curl F dS =

T
(3j i 3: j 3rk) (i +j +k) d =

T
(31) d = 3(area of 1) = 3
1
2
.
35.

T
div Fd\ =

i
2
+u
2
+:
2
$1
3 d\ = 3(volume of sphere) = 4. Then
F(r(c. 0)) (r
c
r
0
) = sin
3
c cos
2
0 + sin
3
c sin
2
0 + sinc cos
2
c = sinc and
S
F dS =

2r
0
r
0
sincdcd0 = (2)(2) = 4.
37. Because curl F = 0, F is conservative, and if 1(r. j. :) = r
3
j: 33rj +:
2
, then Q1 = F. Hence
C
F dr =

C
Q1 dr = 1(0. 3. 0) 31(0. 0. 2) = 0 34 = 34.
39. By the Divergence Theorem,
S
F ndo =

T
div Fd\ = 3(volume of 1) = 3(8 31) = 21.
41. Let F = a r = ho
1
. o
2
. o
3
i hr. j. :i = ho
2
: 3o
3
j. o
3
r 3o
1
:. o
1
j 3o
2
ri. Then curl F = h2o
1
. 2o
2
. 2o
3
i = 2a,
and
S
2a dS =

S
curl F dS =

C
F dr =

C
(a r) dr by Stokes Theorem.
PROBLEMS PLUS
1. Let o
1
be the portion of l(o) between o(o) and o, and let uo
1
be its boundary. Also let o
1
be the lateral surface of o
1
[that
is, the surface of o
1
except o and o(o)]. Applying the Divergence Theorem we have
OS
1
r n
v
3
do =

S
1
Q
r
v
3
d\ .
But
Q
r
v
3
=

u
ur
.
u
uj
.
u
u:
r
(r
2
+j
2
+:
2
)
3'2
.
j
(r
2
+j
2
+:
2
)
3'2
.
:
(r
2
+j
2
+:
2
)
3'2
=
(r
2
+j
2
+:
2
33r
2
) + (r
2
+j
2
+:
2
33j
2
) + (r
2
+j
2
+:
2
33:
2
)
(r
2
+j
2
+:
2
)
5'2
= 0
i

OS
1
r n
v
3
do =

S
1
0 d\ = 0. On the other hand, notice that for the surfaces of uo
1
other than o(o) and o,
r n = 0 i
0 =
OS
1
r n
v
3
do =

S
r n
v
3
do +
S(a)
r n
v
3
do +
S
O
r n
v
3
do =

S
r n
v
3
do +
S(a)
r n
v
3
do i
S
r n
v
3
do = 3
S(a)
r n
v
3
do. Notice that on o(o), v = o i n = 3
r
v
= 3
r
o
and r r = v
2
= o
2
, so
that 3
S(a)
r n
v
3
do =

S(a)
r r
o
4
do =

S(a)
o
2
o
4
do =
1
o
2
S(a)
do =
area of o (o)
o
2
= |l(o)|.
Therefore |l(o)| =

S
r n
v
3
do.
3. The given line integral
1
2
C
(/: 3cj) dr +(cr 3o:) dj +(oj 3/r) d: can be expressed as
C
F dr if we dene the vector
eld F by F(r. j. :) = 1 i +Qj +1k =
1
2
(/: 3cj) i +
1
2
(cr 3o:) j +
1
2
(oj 3/r) k. Then dene o to be the planar
interior of C, so o is an oriented, smooth surface. Stokes Theorem says
C
F dr =

S
curl F dS =

S
curl F ndo.
Now
curl F=

u1
uj
3
uQ
u:
i +
u1
u:
3
u1
ur
j +
uQ
ur
3
u1
uj
k
=

1
2
o +
1
2
o
i +
1
2
/ +
1
2
/
j +
1
2
c +
1
2
c
k = o i +/ j +c k = n
so curl F n = n n = |n|
2
= 1, hence
S
curl F ndo =

S
do which is simply the surface area of o. Thus,
C
F dr =
1
2
C
(/: 3cj) dr + (cr 3o:) dj + (oj 3/r) d: is the plane area enclosed by C.
307
308

PROBLEMS PLUS
5. (F Q) G=

1
1
u
ur
+Q
1
u
uj
+1
1
u
u:
(1
2
i +Q
2
j+1
2
k)
=

1
1
u1
2
ur
+Q
1
u1
2
uj
+1
1
u1
2
u:
i +
1
1
uQ
2
ur
+Q
1
uQ
2
uj
+1
1
uQ
2
u:
j
+

1
1
u1
2
ur
+Q
1
u1
2
uj
+1
1
u1
2
u:
k
= (F Q1
2
) i +(F QQ
2
) j + (F Q1
2
) k.
Similarly, (G Q) F = (G Q1
1
) i +(G QQ
1
) j +(G Q1
1
) k. Then
F curl G =
i
1
1
u1
2
uj 3uQ
2
u:
j
Q
1
u1
2
u: 3u1
2
ur
k
1
1
uQ
2
ur 3u1
2
uj
=

Q
1
uQ
2
ur
3Q
1
u1
2
uj
31
1
u1
2
u:
+1
1
u1
2
ur
i +
1
1
u1
2
uj
31
1
uQ
2
u:
31
1
uQ
2
ur
+1
1
u1
2
uj
j
+
1
1
u1
2
u:
31
1
u1
2
ur
3Q
1
u1
2
uj
+Q
1
uQ
2
u:
k
and
G curl F =

Q
2
uQ
1
ur
3Q
2
u1
1
uj
31
2
u1
1
u:
+1
2
u1
1
ur
i +
1
2
u1
1
uj
31
2
uQ
1
u:
31
2
uQ
1
ur
+1
2
u1
1
uj
j
+
1
2
u1
1
u:
31
2
u1
1
ur
3Q
2
u1
1
uj
+Q
2
uQ
1
u:
k.
Then
(F Q) G+F curl G=

1
1
u1
2
ur
+Q
1
uQ
2
ur
+1
1
u1
2
ur
i +
1
1
u1
2
uj
+Q
1
uQ
2
uj
+1
1
u1
2
uj
j
+
1
1
u1
2
u:
+Q
1
uQ
2
u:
+1
1
u1
2
u:
k
and
(G Q) F+G curl F=

1
2
u1
1
ur
+Q
2
uQ
1
ur
+1
2
u1
1
ur
i +
1
2
u1
1
uj
+Q
2
uQ
1
uj
+1
2
u1
1
uj
j
+
1
2
u1
1
u:
+Q
2
uQ
1
u:
+1
2
u1
1
u:
k.
Hence
(F Q) G+F curl G +(G Q) F +G curl F
=

1
1
u1
2
ur
+1
2
u1
1
ur
Q
1
uQ
2
ur
+Q
2
uQ
1
uj
1
1
u1
2
ur
+1
2
u1
1
ur
i
+
1
1
u1
2
uj
+1
2
u1
1
uj
Q
1
uQ
2
uj
+Q
2
uQ
1
uj
1
1
u1
2
uj
+1
2
u1
1
uj
j
+
1
1
u1
2
u:
+1
2
u1
1
u:
Q
1
uQ
2
u:
+Q
2
uQ
1
u:
1
1
u1
2
u:
+1
2
u1
1
u:
k
= Q(1
1
1
2
+Q
1
Q
2
+1
1
1
2
) = Q(F G).
18 SECOND-ORDER DIFFERENTIAL EQUATIONS ET 17
18.1 Second-Order Linear Equations ET 17.1
1. The auxiliary equation is v
2
3v 36 = 0 i (v 33)(v +2) = 0 i v = 3, v = 32. Then by (8) the general solution is
j = c
1
c
3i
+c
2
c
32i
.
2
+ 16 = 0 i v = 4j. Then by (11) the general solution is
j = c
0i
(c
1
cos 4r +c
2
sin4r) = c
1
cos 4r +c
2
sin4r.
5. The auxiliary equation is 9v
2
312v + 4 = 0 i (3v 32)
2
= 0 i v =
2
3
. Then by (10), the general solution is
j = c
1
c
2i'3
+c
2
rc
2i'3
.
2
3v = v(2v 31) = 0 i v = 0, v =
1
2
, so j = c
1
c
0i
+c
2
c
i'2
= c
1
+c
2
c
i'2
.
2
34v + 13 = 0 i v =
4
I
336
2
= 2 3j, so j = c
2i
(c
1
cos 3r +c
2
sin3r).
2
+ 2v 3 1 = 0 i v =
32
I
12
4
= 3
1
2

I
3
2
, so
j = c
1
c
(31'2+
I
3'2)I
+c
2
c
(31'23
I
3'2)I
.
2
+ 200v + 101 = 0 i v =
3200
I
3400
200
= 31
1
10
j, so
1 = c
3I
c
1
cos
1
10
t
+c
2
sin
1
10
t
.
2
32v 33 = (5v + 3)(v 31) = 0 i v = 3
3
5
,
v = 1, so the general solution is j = c
1
c
33i'5
+c
2
c
i
. We graph the basic
solutions 1(r) = c
33i'5
, o(r) = c
i
as well as j = c
33i'5
+ 2c
i
,
j = c
33i'5
3c
i
, and j = 32c
33i'5
3c
i
. Each solution consists of a single
continuous curve that approaches either 0 or "as r <".
17. 2v
2
+ 5v + 3 = (2v + 3)(v + 1) = 0, so v = 3
3
2
, v = 31 and the general solution is j = c
1
c
33i'2
+c
2
c
3i
.
Then j(0) = 3 i c
1
+c
2
= 3 and j
0
(0) = 34 i 3
3
2
c
1
3c
2
= 34, so c
1
= 2 and c
2
= 1. Thus the solution to the
initial-value problem is j = 2c
33i'2
+c
3i
.
19. 4v
2
34v +1 = (2v 31)
2
= 0 i v =
1
2
and the general solution is j = c
1
c
i'2
+c
2
rc
i'2
. Then j(0) = 1 i c
1
= 1
and j
0
(0) = 31.5 i
1
2
c
1
+c
2
= 31.5, so c
2
= 32 and the solution to the initial-value problem is j = c
i'2
32rc
i'2
.
309
310 CHAPTER 18 SECOND-ORDER DIFFERENTIAL EQUATIONS ET CHAPTER 17
21. v
2
+ 16 = 0 i v = 4j and the general solution is j = c
0i
(c
1
cos 4r +c
2
sin4r) = c
1
cos 4r +c
2
sin4r. Then
j
r
4
= 33 i 3c
1
= 33 i c
1
= 3 and j
0
r
4
= 4 i 34c
2
= 4 i c
2
= 31, so the solution to the
initial-value problem is j = 3 cos 4r 3sin4r.
23. v
2
+ 2v + 2 = 0 i v = 31 j and the general solution is j = c
3i
(c
1
cos r +c
2
sinr). Then 2 = j(0) = c
1
and
1 = j
0
(0) = c
2
3c
1
i c
2
= 3 and the solution to the initial-value problem is j = c
3i
(2 cos r + 3 sinr).
25. 4v
2
+ 1 = 0 i v =
1
2
j and the general solution is j = c
1
cos
1
2
r
+c
2
sin
1
2
r
. Then 3 = j(0) = c
1
and
34 = j() = c
2
, so the solution of the boundary-value problem is j = 3 cos
1
2
r
34 sin
1
2
r
.
27. v
2
33v + 2 = (v 32)(v 31) = 0 i v = 1, v = 2 and the general solution is j = c
1
c
i
+c
2
c
2i
. Then
1 = j(0) = c
1
+c
2
and 0 = j(3) = c
1
c
3
+c
2
c
6
so c
2
= 1(1 3c
3
) and c
1
= c
3
(c
3
31). The solution of the
boundary-value problem is j =
c
i+3
c
3
31
+
c
2i
1 3c
3
.
29. v
2
36v + 25 = 0 i v = 3 4j and the general solution is j = c
3i
(c
1
cos 4r +c
2
sin4r). But 1 = j(0) = c
1
and
2 = j() = c
1
c
3r
i c
1
= 2c
3r
, so there is no solution.
31. v
2
+ 4v + 13 = 0 i v = 32 3j and the general solution is j = c
32i
(c
1
cos 3r +c
2
sin3r). But 2 = j(0) = c
1
and 1 = j
r
2
= c
3r
(3c
2
), so the solution to the boundary-value problem is j = c
32i
(2 cos 3r 3c
r
sin3r).
33. (a) Case 1 (` = 0): j
00
+`j = 0 i j
00
= 0 which has an auxiliary equation v
2
= 0 i v = 0 i j = c
1
+c
2
r
where j(0) = 0 and j(1) = 0. Thus, 0 = j(0) = c
1
and 0 = j(1) = c
2
1 i c
1
= c
2
= 0. Thus j = 0.
Case 2 (` < 0): j
00
+`j = 0 has auxiliary equation v
2
= 3` i v =
I
3` [distinct and real since ` < 0] i
j = c
1
c
I
3Ai
+c
2
c
3
I
3Ai
where j(0) = 0 and j(1) = 0. Thus 0 = j(0) = c
1
+c
2
(W) and
0 = j(1) = c
1
c
I
3A1
+c
2
c
3
I
3A1
().
Multiplying (W) by c
I
3A1
and subtracting () gives c
2
c
I
3A1
3c
3
I
3A1
= 0 i c
2
= 0 and thus c
1
= 0 from (W).
Thus j = 0 for the cases ` = 0 and ` < 0.
(b) j
00
+`j = 0 has an auxiliary equation v
2
+` = 0 i v = j
I
` i j = c
1
cos
I
`r +c
2
sin
I
`r where
j(0) = 0 and j(1) = 0. Thus, 0 = j(0) = c
1
and 0 = j(1) = c
2
sin
I
`1 since c
1
= 0. Since we cannot have a trivial
solution, c
2
6= 0 and thus sin
I
`1 = 0 i
I
`1 = n where n is an integer i ` = n
2
2
1
2
and
j = c
2
sin(nr1) where n is an integer.
SECTION 18.2 NONHOMOGENEOUS LINEAR EQUATIONS ET SECTION 17.2

311
18.2 Nonhomogeneous Linear Equations ET 17.2
2
+ 3v + 2 = (v + 2)(v + 1) = 0, so the complementary solution is j
c
(r) = c
1
c
32i
+c
2
c
3i
.
We try the particular solution j
(r) = r
2
+1r +C, so j
0
= 2r +1 and j
00
= 2. Substituting into the differential

equation, we have (2) + 3(2r +1) + 2(r
2
+1r +C) = r
2
or 2r
2
+ (6+ 21)r + (2+ 31 + 2C) = r
2
.
Comparing coefcients gives 2 = 1, 6+ 21 = 0, and 2+ 31 + 2C = 0, so =
1
2
, 1 = 3
3
2
, and C =
7
4
. Thus the
general solution is j(r) = j
c
(r) +j
(r) = c
1
c
32i
+c
2
c
3i
+
1
2
r
2
3
3
2
r +
7
4
.
2
32v = v(v 32) = 0, so the complementary solution is j
c
(r) = c
1
+c
2
c
2i
. Try the particular
solution j
(r) = cos 4r +1sin4r, so j

0
= 34sin4r + 41cos 4r and j

00
= 316cos 4r 3161sin4r. Substitution

into the differential equation gives (316cos 4r 3161sin4r) 32(34sin4r + 41cos 4r) = sin4r i
(316381) cos 4r + (83161) sin4r = sin4r. Then 316381 = 0 and 83161 = 1 i =
1
40
and
1 = 3
1
20
. Thus the general solution is j(r) = j
c
(r) +j
(r) = c
1
+c
2
c
2i
+
1
40
cos 4r 3
1
20
sin4r.
2
34v + 5 = 0 with roots v = 2 j, so the complementary solution is
j
c
(r) = c
2i
(c
1
cos r +c
2
sinr). Try j
(r) = c
3i
, so j
0
= 3c
3i
and j
00
= c
3i
. Substitution gives
c
3i
34(3c
3i
) + 5(c
3i
) = c
3i
i 10c
3i
= c
3i
i =
1
10
. Thus the general solution is
j(r) = c
2i
(c
1
cos r +c
2
sinr) +
1
10
c
3i
.
2
+ 1 = 0 with roots v = j, so the complementary solution is j
c
(r) = c
1
cos r +c
2
sinr.
For j
00
+j = c
i
try j
1
(r) = c
i
. Then j
0
1
= j
00
1
= c
i
and substitution gives c
i
+c
i
= c
i
i =
1
2
,
so j
1
(r) =
1
2
c
i
. For j
00
+j = r
3
try j
2
(r) = r
3
+1r
2
+Cr +1. Then j
0
2
= 3r
2
+ 21r +C and
j
00
2
= 6r + 21. Substituting, we have 6r + 21 +r
3
+1r
2
+Cr +1 = r
3
, so = 1, 1 = 0,
6+C = 0 i C = 36, and 21 +1 = 0 i 1 = 0. Thus j
2
(r) = r
3
36r and the general solution is
j(r) = j
c
(r) +j
1
(r) +j
2
(r) = c
1
cos r +c
2
sinr +
1
2
c
i
+r
3
36r. But 2 = j(0) = c
1
+
1
2
i
c
1
=
3
2
and 0 = j
0
(0) = c
2
+
1
2
36 i c
2
=
11
2
. Thus the solution to the initial-value problem is
j(r) =
3
2
cos r +
11
2
sinr +
1
2
c
i
+r
3
36r.
2
3v = 0 with roots v = 0, v = 1 so the complementary solution is j
c
(r) = c
1
+c
2
c
i
.
Try j
(r) = r(r +1)c

i
so that no term in j
is a solution of the complementary equation. Then

j
0
= (r
2
+(2+1)r +1)c
i
and j
00
= (r
2
+(4+1)r +(2+21))c
i
. Substitution into the differential equation
gives (r
2
+ (4+1)r + (2+ 21))c
i
3(r
2
+ (2+1)r +1)c
i
= rc
i
i (2r + (2+1))c
i
= rc
i
i
=
1
2
, 1 = 31. Thus j
(r) =
1
2
r
2
3r
c
i
and the general solution is j(r) = c
1
+c
2
c
i
+
1
2
r
2
3r
c
i
. But
2 = j(0) = c
1
+c
2
and 1 = j
0
(0) = c
2
31, so c
2
= 2 and c
1
= 0. The solution to the initial-value problem is
j(r) = 2c
i
+
1
2
r
2
3r
c
i
= c
i
1
2
r
2
3r + 2
.
2
+ 3v + 2 = (v + 1)(v + 2) = 0, so v = 31, v = 32 and j
c
(r) = c
1
c
3i
+c
2
c
32i
.
Try j
= cos r +1sinr i j
0
= 3sinr +1cos r, j
00
= 3cos r 31sinr. Substituting into the differential

equation gives (3cos r 31sinr) + 3(3sinr +1cos r) + 2(cos r +1sinr) = cos r or
(+ 31) cos r + (33+1) sinr = cos r. Then solving the equations
+ 31 = 1, 33+1 = 0 gives =
1
10
, 1 =
3
10
and the general
solution is j(r) = c
1
c
3i
+c
2
c
32i
+
1
10
cos r +
3
10
sinr. The graph
shows j
and several other solutions. Notice that all solutions are

asymptotic to j
as r <". Except for j
, all solutions approach either "

or 3"as r <3".
13. Here j
c
(r) = c
1
cos 3r +c
2
sin3r. For j
00
+ 9j = c
2i
try j
1
(r) = c
2i
and for j
00
+ 9j = r
2
sinr
try j
2
(r) = (1r
2
+Cr +1) cos r + (1r
2
+ 1r +G) sinr. Thus a trial solution is
j
(r) = j
1
(r) +j
2
(r) = c
2i
+ (1r
2
+Cr +1) cos r + (1r
2
+1r +G) sinr.
15. Here j
c
(r) = c
1
+c
2
c
39i
. For j
00
+ 9j
0
= 1 try j
1
(r) = r (since j = is a solution to the complementary equation)
and for j
00
+ 9j
0
= rc
9i
try j
2
(r) = (1r +C)c
9i
.
17. Since j
c
(r) = c
3i
(c
1
cos 3r +c
2
sin3r) we try j
(r) = r(r
2
+1r +C)c
3i
cos 3r +r(1r
2
+1r +1)c
3i
sin3r
(so that no term of j
is a solution of the complementary equation).

Note: Solving Equations (7) and (9) in The Method of Variation of Parameters gives
&
0
1
= 3
Gj
2
o (j
1
j
0
2
3j
2
j
0
1
)
and &
0
2
=
Gj
1
o (j
1
j
0
2
3j
2
j
0
1
)
We will use these equations rather than resolving the system in each of the remaining exercises in this section.
19. (a) Here 4v
2
+ 1 = 0 i v =
1
2
j and j
c
(r) = c
1
cos
1
2
r
+c
2
sin
1
2
r
. We try a particular solution of the form

j
(r) = cos r +1sinr i j

0
= 3sinr +1cos r and j

00
= 3cos r 31sinr. Then the equation

4j
00
+j = cos r becomes 4(3cos r 3 1sinr) + (cos r +1sinr) = cos r or
33cos r 331sinr = cos r i = 3
1
3
, 1 = 0. Thus, j
(r) = 3
1
3
cos r and the general solution is
j(r) = j
c
(r) +j
(r) = c
1
cos
1
2
r
+c
2
sin
1
2
r
3
1
3
cos r.
(b) From (a) we know that j
c
(r) = c
1
cos
i
2
+c
2
sin
i
2
. Setting j
1
= cos
i
2
, j
2
= sin
i
2
, we have
j
1
j
0
2
3j
2
j
0
1
=
1
2
cos
2 i
2
+
1
2
sin
2 i
2
=
1
2
. Thus &
0
1
= 3
cos r sin
i
2
4
1
2
= 3
1
2
cos
2
i
2
sin
i
2
= 3
1
2
2 cos
2 i
2
31
sin
i
2
and &
0
2
=
cos r cos
i
2
4
1
2
=
1
2
cos
2
i
2
cos
i
2
=
1
2
1 32 sin
2 i
2
cos
i
2
. Then
&
1
(r) =

1
2
sin
i
2
3cos
2 i
2
sin
i
2
dr = 3cos
i
2
+
2
3
cos
3 i
2
and
SECTION 18.2 NONHOMOGENEOUS LINEAR EQUATIONS ET SECTION 17.2

313
&
2
(r) =

1
2
cos
i
2
3sin
2 i
2
cos
i
2
dr = sin
i
2
3
2
3
sin
3 i
2
. Thus
j
(r) =

3cos
i
2
+
2
3
cos
3 i
2
cos
i
2
+
sin
i
2
3
2
3
sin
3 i
2
sin
i
2
= 3
cos
2 i
2
3sin
2 i
2
+
2
3
cos
4 i
2
3sin
4 i
2
= 3cos
2
i
2
+
2
3
cos
2 i
2
+ sin
2 i
2

cos
2 i
2
3sin
2 i
2
= 3cos r +
2
3
cos r = 3
1
3
cos r
and the general solution is j(r) = j
c
(r) +j
(r) = c
1
cos
i
2
+c
2
sin
i
2
3
1
3
cos r.
21. (a) v
2
32v + 1 = (v 31)
2
= 0 i v = 1, so the complementary solution is j
c
(r) = c
1
c
i
+c
2
rc
i
. A particular solution
is of the form j
(r) = c
2i
. Thus 4c
2i
34c
2i
+c
2i
= c
2i
i c
2i
= c
2i
i = 1 i j
(r) = c
2i
.
So a general solution is j(r) = j
c
(r) +j
(r) = c
1
c
i
+c
2
rc
i
+c
2i
.
(b) From (a), j
c
(r) = c
1
c
i
+c
2
rc
i
, so set j
1
= c
i
, j
2
= rc
i
. Then, j
1
j
0
2
3j
2
j
0
1
= c
2i
(1 +r) 3rc
2i
= c
2i
and so
&
0
1
= 3rc
i
i &
1
(r) = 3
rc
i
dr = 3(r 31)c
i
[by parts] and &
0
2
= c
i
i &
2
(r) =

c
i
dr = c
i
. Hence
j
(r) = (1 3r)c
2i
+rc
2i
= c
2i
and the general solution is j(r) = j
c
(r) +j
(r) = c
1
c
i
+c
2
rc
i
+c
2i
.
23. As in Example 5, j
c
(r) = c
1
sinr +c
2
cos r, so set j
1
= sinr, j
2
= cos r. Then j
1
j
0
2
3j
2
j
0
1
= 3sin
2
r 3cos
2
r = 31,
so &
0
1
= 3
sec
2
r cos r
31
= sec r i &
1
(r) =

sec rdr = ln(sec r + tanr) for 0 < r <
r
2
,
and &
0
2
=
sec
2
r sinr
31
= 3sec r tanr i &
2
(r) = 3sec r. Hence
j
(r) = ln(sec r + tanr) sinr 3sec r cos r = sinrln(sec r + tanr) 31 and the general solution is
j(r) = c
1
sinr +c
2
cos r + sinr ln(sec r + tanr) 31.
25. j
1
= c
i
, j
2
= c
2i
and j
1
j
0
2
3 j
2
j
0
1
= c
3i
. So &
0
1
=
3c
2i
(1 +c
3i
)c
3i
= 3
c
3i
1 +c
3i
and
&
1
(r) =

3
c
3i
1 +c
3i
dr = ln(1 + c
3i
). &
0
2
=
c
i
(1 +c
3i
)c
3i
=
c
i
c
3i
+c
2i
so
&
2
(r) =

c
i
c
3i
+c
2i
dr = ln
c
i
+ 1
c
i
3 c
3i
= ln(1 + c
3i
) 3c
3i
. Hence
j
(r) = c
i
ln(1 + c
3i
) + c
2i
[ln(1 + c
3i
) 3 c
3i
] and the general solution is
j(r) = [c
1
+ ln(1 +c
3i
)]c
i
+ [c
2
3c
3i
+ ln(1 +c
3i
)]c
2i
.
27. v
2
32v + 1 = (v 31)
2
= 0 i v = 1 so j
c
(r) = c
1
c
i
+c
2
rc
i
. Thus j
1
= c
i
, j
2
= rc
i
and
j
1
j
0
2
3j
2
j
0
1
= c
i
(r + 1)c
i
3rc
i
c
i
= c
2i
. So &
0
1
= 3
rc
i
c
i
(1 +r
2
)
c
2i
= 3
r
1 +r
2
i
&
1
= 3
r
1 +r
2
dr = 3
1
2
ln
1 +r
2
, &
0
2
=
c
i
c
i
(1 +r
2
)
c
2i
=
1
1 +r
2
i &
2
=

1
1 +r
2
dr = tan
31
r and
j
(r) = 3
1
2
c
i
ln(1 +r
2
) +rc
i
tan
31
r. Hence the general solution is j(r) = c
i
c
1
+c
2
r 3
1
2
ln(1 +r
2
) +rtan
31
r
.
18.3 Applications of Second-Order Differential Equations ET 17.3
1. By Hookes Law I(0.25) = 25 so I = 100 is the spring constant and the differential equation is 5r
00
+ 100r = 0.
The auxiliary equation is 5v
2
+ 100 = 0 with roots v = 2
I
5 j, so the general solution to the differential equation is
r(t) = c
1
cos
2
I
5 t
+c
2
sin
2
I
5 t
. We are given that r(0) = 0.35 i c

1
= 0.35 and r
0
(0) = 0 i
2
I
5 c
2
= 0 i c
2
= 0, so the position of the mass after t seconds is r(t) = 0.35 cos
2
I
5 t
.
3. I(0.5) = 6 or I = 12 is the spring constant, so the initial-value problem is 2r
00
+ 14r
0
+ 12r = 0, r(0) = 1, r
0
(0) = 0.
The general solution is r(t) = c
1
c
36I
+c
2
c
3I
. But 1 = r(0) = c
1
+c
2
and 0 = r
0
(0) = 36c
1
3c
2
. Thus the position is
given by r(t) = 3
1
5
c
36I
+
6
5
c
3I
.
5. For critical damping we need c
2
34iI = 0 or i = c
2
(4I) = 14
2
(4 12) =
49
12
kg.
7. We are given i = 1, I = 100, r(0) = 30.1 and r
0
(0) = 0. From (3), the differential equation is
d
2
r
dt
2
+c
dr
dt
+ 100r = 0
with auxiliary equation v
2
+cv + 100 = 0.
If c = 10, we have two complex roots v = 35 5
I
3 j, so the motion is underdamped and the solution is
r = c
35I
c
1
cos
5
I
3 t
+c
2
sin
5
I
3 t
. Then 30.1 = r(0) = c

1
and 0 = r
0
(0) = 5
I
3 c
2
35c
1
i c
2
= 3
1
10
I
3
,
so r = c
35I
30.1 cos
5
I
3 t
3
1
10
I
3
sin
5
I
3 t
.
If c = 15, we again have underdamping since the auxiliary equation has roots v = 3
15
2

5
I
7
2
j. The general solution is
r = c
315I'2
c
1
cos
5
I
7
2
t
+c
2
sin
5
I
7
2
t
, so 30.1 = r(0) = c
1
and 0 = r
0
(0) =
5
I
7
2
c
2
3
15
2
c
1
i c
2
= 3
3
10
I
7
.
Thus r = c
315I'2
30.1 cos
5
I
7
2
t
3
3
10
I
7
sin
5
I
7
2
t
.
For c = 20, we have equal roots v
1
= v
2
= 310, so the oscillation is critically damped and the solution is
r = (c
1
+c
2
t)c
310I
. Then 30.1 = r(0) = c
1
and 0 = r
0
(0) = 310c
1
+c
2
i c
2
= 31, so r = (30.1 3t)c
310I
.
If c = 25 the auxiliary equation has roots v
1
= 35, v
2
= 320, so we have overdamping and the solution is
r = c
1
c
35I
+c
2
c
320I
. Then 30.1 = r(0) = c
1
+c
2
and 0 = r
0
(0) = 35c
1
320c
2
i c
1
= 3
2
15
and c
2
=
1
30
,
so r = 3
2
15
c
35I
+
1
30
c
320I
.
If c = 30 we have roots v = 315 5
I
5, so the motion is
overdamped and the solution is r = c
1
c
(315 +5
I
5 )I
+c
2
c
(315 35
I
5 )I
.
Then 30.1 = r(0) = c
1
+c
2
and
0 = r
0
(0) =

315 + 5
I
5
c
1
+
315 35
I
5
c
2
i
c
1
=
35 33
I
5
100
and c
2
=
35 +3
I
5
100
, so
r =

35 33
I
5
100
c
(315 +5
I
5)I
+
35 +3
I
5
100
c
(315 35
I
5)I
.
SECTION 18.3 APPLICATIONS OF SECOND-ORDER DIFFERENTIAL EQUATIONS ET SECTION 17.3

315
9. The differential equation is ir
00
+Ir = 1
0
cos .
0
t and .
0
6= . =

Ii. Here the auxiliary equation is iv
2
+I = 0
with roots
Iij = .j so r
c
(t) = c
1
cos .t +c
2
sin.t. Since .
0
6= ., try r
(t) = cos .
0
t +1sin.
0
t.
Then we need (i)
3.
2
0
(cos .
0
t +1sin.
0
t) +I(cos .
0
t +1sin.
0
t) = 1
0
cos .
0
t or
I 3i.
2
0
= 1
0
and
1
I 3i.
2
0
= 0. Hence 1 = 0 and =
1
0
I 3i.
2
0
=
1
0
i(.
2
3.
2
0
)
since .
2
=
I
i
. Thus the motion of the mass is given
by r(t) = c
1
cos .t +c
2
sin.t +
1
0
i(.
2
3.
2
0
)
cos .
0
t.
11. From Equation 6, r(t) = 1(t) +o(t) where 1(t) = c
1
cos .t +c
2
sin.t and o(t) =
1
0
i(.
2
3.
2
0
)
cos .
0
t. Then 1
is periodic, with period
2r
.
, and if . 6= .
0
, o is periodic with period
2r
.
0
. If
.
.
0
is a rational number, then we can say
.
.
0
=
a
l
i o =
l.
.
0
where o and / are non-zero integers. Then
r
t +o
2r
.
= 1
t +o
2r
.
+o
t +o
2r
.
= 1(t) +o
t +
l.
.
0

2r
.
= 1(t) +o
t +/
2r
.
0
= 1(t) +o(t) = r(t)

so r(t) is periodic.
13. Here the initial-value problem for the charge is Q
00
+ 20Q
0
+ 500Q = 12, Q(0) = Q
0
(0) = 0. Then
Q
c
(t) = c
310I
(c
1
cos 20t +c
2
sin20t) and try Q
(t) = i 500 = 12 or =
3
125
.
The general solution is Q(t) = c
310I
(c
1
cos 20t +c
2
sin20t) +
3
125
. But 0 = Q(0) = c
1
+
3
125
and
Q
0
(t) = 1(t) = c
310I
[(310c
1
+ 20c
2
) cos 20t + (310c
2
320c
1
) sin20t] but 0 = Q
0
(0) = 310c
1
+ 20c
2
. Thus the charge
is Q(t) = 3
1
250
c
310I
(6 cos 20t + 3 sin20t) +
3
125
and the current is 1(t) = c
310I
3
5
sin20t.
15. As in Exercise 13, Q
c
(t) = c
310I
(c
1
cos 20t + c
2
sin20t) but 1(t) = 12 sin10t so try
Q
(t) = cos 10t + 1sin10t. Substituting into the differential equation gives
(3100 + 2001 + 500) cos 10t + (31001 3200 + 5001) sin10t = 12 sin10t i
400 + 2001 = 0 and 4001 3200 = 12. Thus = 3
3
250
, 1 =
3
125
and the general solution is
Q(t) = c
310I
(c
1
cos 20t +c
2
sin20t) 3
3
250
cos 10t +
3
125
sin10t. But 0 = Q(0) = c
1
3
3
250
so c
1
=
3
250
.
Also Q
0
(t) =
3
25
sin10t +
6
25
cos 10t +c
310I
[(310c
1
+ 20c
2
) cos 20t + (310c
2
320c
1
) sin20t] and
0 = Q
0
(0) =
6
25
3 10c
1
+ 20c
2
so c
2
= 3
3
500
. Hence the charge is given by
Q(t) = c
310I
3
250
cos 20t 3
3
500
sin20t
3
3
250
cos 10t +
3
125
sin10t.
17. r(t) = cos(.t +c) C r(t) = [cos .t cos c 3sin.t sinc] C r(t) =
c
1
cos .t +
c
2
sin.t
where
cos c = c
1
and sinc = 3c
2
C r(t) = c
1
cos .t +c
2
sin.t. [Note that cos
2
c + sin
2
c = 1 i c
2
1
+c
2
2
=
2
.]
18.4 Series Solutions ET 17.4
1. Let j(r) =
"
n=0
c
n
r
n
. Then j
0
(r) =
"
n=1
nc
n
r
n31
and the given equation, j
0
3j = 0, becomes
"
n=1
nc
n
r
n31
3
"
n=0
c
n
r
n
= 0. Replacing n by n + 1 in the rst sum gives
"
n=0
(n + 1)c
n+1
r
n
3
"
n=0
c
n
r
n
= 0, so
"
n=0
[(n + 1)c
n+1
3c
n
]r
n
= 0. Equating coefcients gives (n + 1)c
n+1
3c
n
= 0, so the recursion relation is
c
n+1
=
c
n
n + 1
, n = 0. 1. 2. . . .. Then c
1
= c
0
, c
2
=
1
2
c
1
=
c
0
2
, c
3
=
1
3
c
2
=
1
3

1
2
c
0
=
c
0
3!
, c
4
=
1
4
c
3
=
c
0
4!
, and
in general, c
n
=
c
0
n!
. Thus, the solution is j(r) =
"
n=0
c
n
r
n
=
"
n=0
c
0
n!
r
n
= c
0
"
n=0
r
n
n!
= c
0
c
i
.
3. Assuming j(r) =
"
n=0
c
n
r
n
, we have j
0
(r) =
"
n=1
nc
n
r
n31
=
"
n=0
(n + 1)c
n+1
r
n
and
3r
2
j = 3
"
n=0
c
n
r
n+2
= 3
"
n=2
c
n32
r
n
. Hence, the equation j
0
= r
2
j becomes
"
n=0
(n + 1)c
n+1
r
n
3
"
n=2
c
n32
r
n
= 0
or c
1
+ 2c
2
r +
"
n=2
[(n + 1)c
n+1
3c
n32
] r
n
= 0. Equating coefcients gives c
1
= c
2
= 0 and c
n+1
=
c
n32
n + 1
for n = 2. 3, . . . . But c
1
= 0, so c
4
= 0 and c
7
= 0 and in general c
3n+1
= 0. Similarly c
2
= 0 so c
3n+2
= 0. Finally
c
3
=
c
0
3
, c
6
=
c
3
6
=
c
0
6 3
=
c
0
3
2
2!
, c
9
=
c
6
9
=
c
0
9 6 3
=
c
0
3
3
3!
, . . ., and c
3n
=
c
0
3
n
n!
. Thus, the solution
is j (r) =
"
n=0
c
n
r
n
=
"
n=0
c
3n
r
3n
=
"
n=0
c
0
3
n
n!
r
3n
= c
0
"
n=0
r
3n
3
n
n!
= c
0
"
n=0
r
3
3
n
n!
= c
0
c
i
3
'3
.
5. Let j (r) =
"
n=0
c
n
r
n
i j
0
(r) =
"
n=1
nc
n
r
n31
and j
00
(r) =
"
n=0
(n + 2)(n + 1)c
n+2
r
n
. The differential equation
becomes
"
n=0
(n + 2)(n + 1)c
n+2
r
n
+r
"
n=1
nc
n
r
n31
+
"
n=0
c
n
r
n
= 0 or
"
n=0
[(n + 2)(n + 1)c
n+2
+nc
n
+c
n
]r
n
= 0
since
"
n=1
nc
n
r
n
=
"
n=0
nc
n
r
n
. Equating coefcients gives (n + 2)(n + 1)c

n+2
+ (n + 1)c
n
= 0, thus the
recursion relation is c
n+2
=
3(n + 1)c
n
(n + 2)(n + 1)
= 3
c
n
n + 2
, n = 0. 1. 2. . . . . Then the even
coefcients are given by c
2
= 3
c
0
2
, c
4
= 3
c
2
4
=
c
0
2 4
, c
6
= 3
c
4
6
= 3
c
0
2 4 6
, and in general,
c
2n
= (31)
n
c
0
2 4 2n
=
(31)
n
c
0
2
n
n!
. The odd coefcients are c
3
= 3
c
1
3
, c
5
= 3
c
3
5
=
c
1
3 5
, c
7
= 3
c
5
7
= 3
c
1
3 5 7
,
and in general, c
2n+1
= (31)
n
c
1
3 5 7 (2n + 1)
=
(32)
n
n! c
1
(2n + 1)!
. The solution is
j (r) = c
0
"
n=0
(31)
n
2
n
n!
r
2n
+c
1
"
n=0
(32)
n
n!
(2n + 1)!
r
2n+1
.
7. Let j (r) =
"
n=0
c
n
r
n
i j
0
(r) =
"
n=1
nc
n
r
n31
=
"
n=0
(n + 1)c
n+1
r
n
and j
00
(r) =
"
n=0
(n + 2)(n + 1)c
n+2
r
n
. Then
(r31)j
00
(r) =
"
n=0
(n+2)(n+1)c
n+2
r
n+1
3
"
n=0
(n+2)(n+1)c
n+2
r
n
=
"
n=1
n(n+1)c
n+1
r
n
3
"
n=0
(n+2)(n+1)c
n+2
r
n
.
SECTION 18.4 SERIES SOLUTIONS ET SECTION 17.4

317
Since
"
n=1
n(n + 1)c
n+1
r
n
=
"
n=0
n(n + 1)c
n+1
r
n
, the differential equation becomes
"
n=0
n(n + 1)c
n+1
r
n
3
"
n=0
(n + 2)(n + 1)c
n+2
r
n
+
"
n=0
(n + 1)c
n+1
r
n
= 0 i
"
n=0
[n(n + 1)c
n+1
3(n + 2)(n + 1)c
n+2
+ (n + 1)c
n+1
]r
n
= 0 or
"
n=0
[(n + 1)
2
c
n+1
3(n + 2)(n + 1)c
n+2
]r
n
= 0.
Equating coefcients gives (n + 1)
2
c
n+1
3(n + 2)(n + 1)c
n+2
= 0 for n = 0. 1. 2, . . . . Then the recursion relation is
c
n+2
=
(n + 1)
2
(n + 2)(n + 1)
c
n+1
=
n + 1
n + 2
c
n+1
, so given c
0
and c
1
, we have c
2
=
1
2
c
1
, c
3
=
2
3
c
2
=
1
3
c
1
, c
4
=
3
4
c
3
=
1
4
c
1
, and
in general c
n
=
c
1
n
, n = 1. 2. 3, . . . . Thus the solution is j(r) = c
0
+c
1
"
n=1
r
n
n
. Note that the solution can be expressed as
c
0
3c
1
ln(1 3r) for |r| < 1.
9. Let j(r) =
"
n=0
c
n
r
n
. Then 3rj
0
(r) = 3r
"
n=1
nc
n
r
n31
= 3
"
n=1
nc
n
r
n
= 3
"
n=0
nc
n
r
n
,
j
00
(r) =
"
n=0
(n + 2)(n + 1)c
n+2
r
n
, and the equation j
00
3 rj
0
3 j = 0
becomes
"
n=0
[(n + 2)(n + 1)c
n+2
3nc
n
3c
n
]r
n
= 0. Thus, the recursion relation is
c
n+2
=
nc
n
+c
n
(n + 2)(n + 1)
=
c
n
(n + 1)
(n + 2)(n + 1)
=
c
n
n + 2
for n = 0. 1. 2, . . . . One of the given conditions is j(0) = 1. But
j(0) =
"
n=0
c
n
(0)
n
= c
0
+ 0 + 0 + = c
0
, so c
0
= 1. Hence, c
2
=
c
0
2
=
1
2
, c
4
=
c
2
4
=
1
2 4
, c
6
=
c
4
6
=
1
2 4 6
, . . . ,
c
2n
=
1
2
n
n!
. The other given condition is j
0
(0) = 0. But j
0
(0) =
"
n=1
nc
n
(0)
n31
= c
1
+ 0 + 0 + = c
1
, so c
1
= 0.
By the recursion relation, c
3
=
c
1
3
= 0, c
5
= 0, . . . , c
2n+1
= 0 for n = 0, 1, 2, . . . . Thus, the solution to the initial-value
problem is j(r) =
"
n=0
c
n
r
n
=
"
n=0
c
2n
r
2n
=
"
n=0
r
2n
2
n
n!
=
"
n=0
(r
2
2)
n
n!
= c
i
2
'2
.
11. Assuming that j(r) =
"
n=0
c
n
r
n
, we have rj = r
"
n=0
c
n
r
n
=
"
n=0
c
n
r
n+1
, r
2
j
0
= r
2
"
n=1
nc
n
r
n31
=
"
n=0
nc
n
r
n+1
,
j
00
(r) =
"
n=2
n(n 31)c
n
r
n32
=
"
n=31
(n + 3)(n + 2)c
n+3
r
n+1
[replace n with n + 3]
= 2c
2
+
"
n=0
(n + 3)(n + 2)c
n+3
r
n+1
,
and the equation j
00
+r
2
j
0
+rj = 0 becomes 2c
2
+
"
n=0
[(n + 3)(n + 2)c
n+3
+nc
n
+c
n
] r
n+1
= 0. So c
2
= 0 and the
recursion relation is c
n+3
=
3nc
n
3c
n
(n + 3)(n + 2)
= 3
(n + 1)c
n
(n + 3)(n + 2)
, n = 0. 1. 2, . . . . But c
0
= j(0) = 0 = c
2
and by the
recursion relation, c
3n
= c
3n+2
= 0 for n = 0, 1, 2, . . . . Also, c
1
= j
0
(0) = 1, so c
4
= 3
2c
1
4 3
= 3
2
4 3
,
c
7
= 3
5c
4
7 6
= (31)
2
2 5
7 6 4 3
= (31)
2
2
2
5
2
7!
, . . . , c
3n+1
= (31)
n
2
2
5
2
(3n 31)
2
(3n + 1)!
. Thus, the solution is
j(r) =
"
n=0
c
n
r
n
= r +
"
n=1
(31)
n
2
2
5
2
(3n 31)
2
r
3n+1
(3n + 1)!
.
18 Review ET 17
1. (a) oj
00
+/j
0
+cj = 0 where o, /, and c are constants.
(b) ov
2
+/v +c = 0
(c) If the auxiliary equation has two distinct real roots v
1
and v
2
, the solution is j = c
1
c
r
1
i
+c
2
c
r
2
i
. If the roots are real and
equal, the solution is j = c
1
c
ri
+c
2
rc
ri
where v is the common root. If the roots are complex, we can write v
1
= c+jo
and v
2
= c 3jo, and the solution is j = c
oi
(c
1
cos or +c
2
sinor).
2. (a) An initial-value problem consists of nding a solution j of a second-order differential equation that also satises given
conditions j(r
0
) = j
0
and j
0
(r
0
) = j
1
, where j
0
and j
1
are constants.
(b) A boundary-value problem consists of nding a solution j of a second-order differential equation that also satises given
boundary conditions j(r
0
) = j
0
and j(r
1
) = j
1
.
3. (a) oj
00
+/j
0
+cj = G(r) where o, /, and c are constants and Gis a continuous function.
(b) The complementary equation is the related homogeneous equation oj
00
+/j
0
+cj = 0. If we nd the general solution j
c
of the complementary equation and j
is any particular solution of the original differential equation, then the general
solution of the original differential equation is j(r) = j
(r) +j
c
(r).
(c) See Examples 15 and the associated discussion in Section 18.2 [ ET 17.2].
(d) See the discussion on pages 11581160 [ ET 11221124].
4. Second-order linear differential equations can be used to describe the motion of a vibrating spring or to analyze an electric
circuit; see the discussion in Section 18.3 [ ET 17.3].
5. See Example 1 and the preceding discussion in Section 18.4 [ ET 17.4].
.
1. True. See Theorem 18.1.3 [ ET 17.1.3].
3. True. coshr and sinhr are linearly independent solutions of this linear homogeneous equation.
2
32v 315 = 0 i (v 35)(v + 3) = 0 i v = 5, v = 33. Then the general solution
is j = c
1
c
5i
+c
2
c
33i
.
2
+ 3 = 0 i v =
I
3 j. Then the general solution is j = c
1
cos
I
3 r
+c
2
sin
I
3 r
.
5. v
2
34v + 5 = 0 i v = 2 j, so j
c
(r) = c
2i
(c
1
cos r +c
2
sinr). Try j
(r) = c
2i
i j
0
= 2c
2i
and j
00
= 4c
2i
. Substitution into the differential equation gives 4c
2i
38c
2i
+ 5c
2i
= c
2i
i = 1 and
the general solution is j(r) = c
2i
(c
1
cos r +c
2
sinr) +c
2i
.
7. v
2
32v + 1 = 0 i v = 1 and j
c
(r) = c
1
c
i
+c
2
rc
i
. Try j
(r) = (r +1) cos r + (Cr +1) sinr i

j
0
= (C 3r 31) sinr + (+Cr +1) cos r and j

00
= (2C 31 3r) cos r + (32313Cr) sinr. Substitution

gives (32Cr + 2C 32321) cos r + (2r 32+ 21 32C) sinr = rcos r i = 0, 1 = C = 1 = 3
1
2
.
The general solution is j(r) = c
1
c
i
+c
2
rc
i
3
1
2
cos r 3
1
2
(r + 1) sinr.
9. v
2
3v 36 = 0 i v = 32, v = 3 and j
c
(r) = c
1
c
32i
+c
2
c
3i
. For j
00
3j
0
36j = 1, try j
1
(r) = . Then
j
0
1
(r) = j
00
1
(r) = 0 and substitution into the differential equation gives = 3
1
6
. For j
00
3j
0
36j = c
32i
try
j
2
(r) = 1rc
32i
[since j = 1c
32i
satises the complementary equation]. Then j
0
2
= (1 321r)c
32i
and
j
00
2
= (41r 341)c
32i
, and substitution gives 351c
32i
= c
32i
i 1 = 3
1
5
. The general solution then is
j(r) = c
1
c
32i
+c
2
c
3i
+j
1
(r) +j
2
(r) = c
1
c
32i
+c
2
c
3i
3
1
6
3
1
5
rc
32i
.
2
+ 6v = 0 and the general solution is j(r) = c
1
+c
2
c
36i
= I
1
+I
2
c
36(i31)
. But
3 = j(1) = I
1
+I
2
and 12 = j
0
(1) = 36I
2
. Thus I
2
= 32, I
1
= 5 and the solution is j(r) = 5 32c
36(i31)
.
2
35v + 4 = 0 and the general solution is j(r) = c
1
c
i
+c
2
c
4i
. But 0 = j(0) = c
1
+c
2
and 1 = j
0
(0) = c
1
+ 4c
2
, so the solution is j(r) =
1
3
(c
4i
3c
i
).
15. Let j(r) =
"
n=0
c
n
r
n
. Then j
00
(r) =
"
n=0
n(n 31)c
n
r
n32
=
"
n=0
(n + 2)(n + 1)c
n+2
r
n
and the differential equation
becomes
"
n=0
[(n + 2)(n + 1)c
n+2
+ (n + 1)c
n
]r
n
= 0. Thus the recursion relation is c
n+2
= 3c
n
(n + 2)
for n = 0. 1. 2, . . . . But c
0
= j(0) = 0, so c
2n
= 0 for n = 0. 1. 2, . . . . Also c
1
= j
0
(0) = 1, so c
3
= 3
1
3
, c
5
=
(31)
2
3 5
,
c
7
=
(31)
3
3 5 7
=
(31)
3
2
3
3!
7!
, . . . , c
2n+1
=
(31)
n
2
n
n!
(2n + 1)!
for n = 0. 1. 2. . . . . Thus the solution to the initial-value problem
is j(r) =
"
n=0
c
n
r
n
=
"
n=0
(31)
n
2
n
n!
(2n + 1)!
r
2n+1
.
17. Here the initial-value problem is 2Q
00
+ 40Q
0
+ 400Q = 12, Q(0) = 0.01, Q
0
(0) = 0. Then
Q
c
(t) = c
310I
(c
1
cos 10t +c
2
sin10t) and we try Q
(t) = . Thus the general solution is

Q(t) = c
310I
(c
1
cos 10t +c
2
sin10t) +
3
100
. But 0.01 = Q
0
(0) = c
1
+ 0.03 and 0 = Q
00
(0) = 310c
1
+ 10c
2
,
so c
1
= 30.02 = c
2
. Hence the charge is given by Q(t) = 30.02c
310I
(cos 10t + sin10t) + 0.03.
19. (a) Since we are assuming that the earth is a solid sphere of uniform density, we can calculate the density a as follows:
a =
mass of earth
volume of earth
=
A
4
3
1
3
. If \
r
is the volume of the portion of the earth which lies within a distance v of the
center, then \
r
=
4
3
v
3
and A
r
= a\
r
=
Av
3
1
3
. Thus 1
r
= 3
GA
r
i
v
2
= 3
GAi
1
3
v.
(b) The particle is acted upon by a varying gravitational force during its motion. By Newtons Second Law of Motion,
i
d
2
j
dt
2
= 1
u
= 3
GAi
1
3
j, so j
00
(t) = 3I
2
j (t) where I
2
=
GA
1
3
. At the surface, 3io = 1
T
= 3
GAi
1
2
, so
o =
GA
1
2
. Therefore I
2
=
o
1
.
(c) The differential equation j
00
+I
2
j = 0 has auxiliary equation v
2
+I
2
= 0. (This is the v of Section 18.1 [ ET 17.1],
not the v measuring distance from the earths center.) The roots of the auxiliary equation are jI, so by (11) in
Section 18.1 [ ET 17.1], the general solution of our differential equation for t is j(t) = c
1
cos It +c
2
sinIt. It follows that
j
0
(t) = 3c
1
I sinIt +c
2
I cos It. Now j (0) = 1 and j
0
(0) = 0, so c
1
= 1 and c
2
I = 0. Thus j(t) = 1cos It and
j
0
(t) = 3I1sinIt. This is simple harmonic motion (see Section 18.3 [ ET 17.3]) with amplitude 1, frequency I, and
phase angle 0. The period is T = 2I. 1 E 3960 mi = 3960 5280 ft and o = 32 fts
2
, so
I =

o1 E 1.24 10
33
s
31
and T = 2I E 5079 s E 85 min.
(d) j(t) = 0 C cos It = 0 C It =
r
2
+n for some integer n i j
0
(t) = 3I1sin
r
2
+n
= I1. Thus the

particle passes through the center of the earth with speed I1 E 4.899 mis E 17,600 mih.

Solutions Manual (JJC)

Uploaded by

Document Information

Original Description:

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Solutions Manual (JJC)

Uploaded by

Copyright:

Available Formats

1 FUNCTIONS AND MODELS

1.1 Four Ways to Represent a Function

must be 1. Since o is atter than / near the origin, we must have

matched with o and (a)

C. (Equivalently, 1 increases by 9 when C increases by 5

we stretch the graph of 1

N curve in Figure 9 on June 21) is

. March 31 is the 90th day of the year, so the model gives

for n = 0, 1, . . . , 94. But

[by periodicity of sine]. n = 0, 1, . . . , 94

are decreasing. The graph of

= ln3 C (r 35) ln2 = ln3 C r 35 =

. Using formula (10), we can write

; this is the time

= ln81 E 4.4 years, as in part (a).

= |(r 31)(r 33)|.

. The intersection of these intervals is

40(2 +/) 316(2 +/)

o(t) = 31 (b) lim

o(t) = 2 (e) lim

1(r) = 3" (e) lim

1(r) = 3"and lim

1(r) = 3"and lim

is the reference angle for 30

are also equations of

= 0 by the Squeeze Theorem.

(2r + |r 33|) = lim

(2r + 3 3r) = lim

(r + 3) = 3 + 3 = 6. Since the left and right limits are equal,

1(r) = 31 and lim

< 1 C 1 < 4 +r 33r

< 0.1 C 1.9 < 4 +r 33r

< -. By the denition

< - C |r + 3 35| < - [r 6= 2] C |r 32| < -. So choose c = -.

< -. By the denition of a limit, lim

< -. Thus, lim

= |r|. So this is true if we pick c = -.

= 1 by the denition of a limit.

< - or upon simplifying we need

. If 0 < |r 33| < c, then |r 33| < 2 i 32 < r 33 < 2 i

-. This suggests that we let

. If 0 < |r 3o| < c, then

1(r) = 1(32). 1 is continuous from the right at 2 and 4 since

= {r | (r + 3)(r + 2) 6= 0} = {r | r 6= 33, 32} or (3". 33) (33. 32) (32. ").

is continuous on its domain, which is

= {t | |t| 1} = (3". 31) (1. ")

= 8 32c. So 1 is continuous C 4c +4 = 8 32c C 6c = 4 C c =

= (cos o)(1) 3(sino)(0) = cos o

1(r) = " (c) lim

= 0, so by the Squeeze Theorem, lim

1(r) = 3", and lim

o(r) = 3"and lim

1(r) = "and lim

4(1 +/) 3(1 +/)

. The initial rate of change is greater in magnitude than the

5000 + 10(100 +/) + 0.05(100 +/)

C. This means that as the temperature increases past 16

5(t +/) 39(t +/)

|r| is differentiable at r = 31. But no

1 + 4(r +/) 3(r +/)

4(r +/) 33(r +/)

(31) = 31. Since these limits are not equal,