INSTRUCTOR’S SOLUTIONS MANUAL APPENDIX I.

(PAGE A-10)

APPENDICES 4π 4π
15. z = 3 cos + 3i sin
5 5
4π
|z| = 3, Arg (z) =
5
Appendix I. Complex Numbers 3π π
(page A-10) 16. If Arg (z) = and Arg (w) = , then
4 2
3π π 5π
arg (zw) = + = , so
1. z = −5 + 2i, Re(z) = −5, Im(z) = 2 4 2 4
y 5π −3π
z = −5 + 2i Arg (zw) = − 2π = .
z-plane 4 4
5π π
z = −6 17. If Arg (z) = − and Arg (w) = , then
6 4
x 5π π 13π
z = 4−i arg (z/w) = − − =− , so
6 4 12
13π 11π
Arg (z/w) = − + 2π = .
z = −πi 12 12
18. |z| = 2, arg (z) = π ⇒ z = 2(cos π + i sin π ) = −2

Fig. .1 19. |z| = 5, θ = arg (z) = π ⇒ sin θ = 3/5, cos θ = 4/5

z = 4 + 3i
2. z = 4 − i, Re(z) = 4, Im(z) = −1

3π 3π 3π
20. |z| = 1, arg (z) = ⇒ z = cos + i sin
3. z = −πi, Re(z) = 0, Im(z) = −π 4 4 4
1 1
4. z = −6, Re(z) = −6, Im(z) = 0 ⇒ z = −√ + √ i
2 2
√ 
5. z = −1 + i, |z| = 2, Arg (z) = 3π/4 π π π
√ 21. |z| = π, arg (z) = ⇒ z = π cos + i sin
z = 2 (cos(3π/4) + i sin(3π/4)) √ 6 6 6
π 3 π
⇒ z= + i
6. z = −2, |z| = 2, Arg (z) = π 2 2
z = 2(cos π + i sin π ) 22. |z| = 0 ⇒ z = 0 for any value of arg (z)
7. z = 3i, |z| = 3, Arg (z) = π/2 1 π 1 π π
23. |z| = , arg (z) = − ⇒ z= cos − i sin
z = 3(cos(π/2) + i sin(π/2)) 2 √ 3 2 3 3
1 3
8. z = −5i, |z| = 5, Arg (z) = −π/2 ⇒ z= − i
4 4
z = 5(cos(−π/2) + i sin(−π/2))
24. 5 + 3i = 5 − 3i
√
9. z = 1 + 2i, |z| = 5, θ = Arg (z) = tan−1 2 25. −3 − 5i = −3 + 5i
√
z = 5(cos θ + i sin θ ) 26. 4i = −4i
√
10. z = −2 + i, |z| = 5, θ = Arg (z) = π − tan−1 (1/2) 27. 2 − i = 2 + i
√
z = 5(cos θ + i sin θ ) 28. |z| = 2 represents all points on the circle of radius 2
centred at the origin.
11. z = −3 − 4i, |z| = 5, θ = Arg (z) = −π + tan−1 (4/3)
z = 5(cos θ + i sin θ ) 29. |z| ≤ 2 represents all points in the closed disk of radius 2
centred at the origin.
12. z = 3 − 4i, |z| = 5, θ = Arg (z) = −tan−1 (4/3)
30. |z − 2i | ≤ 3 represents all points in the closed disk of
z = 5(cos θ + i sin θ ) radius 3 centred at the point 2i .
√
13. z = 3 − i, |z| = 2, Arg (z) = −π/6 31. |z − 3 + 4i | ≤ 5 represents all points in the closed disk of
z = 2(cos(−π/6) + i sin(−π/6)) radius 5 centred at the point 3 − 4i .
√ √ 32. arg (z) = π/3 represents all points on the ray from the
14. z = − 3 − 3i, |z| = 2 3, Arg (z) = −2π/3
√ origin in the first quadrant, making angle 60◦ with the
z = 2 3(cos(−2π/3) + i sin(−2π/3)) positive direction of the real axis.

661
 APPENDIX I. (PAGE A-10) R. A. ADAMS: CALCULUS

33. π ≤ arg (z) ≤ 7π/4 represents the closed wedge-shaped 48. cos(3θ ) + i sin(3θ ) = (cos θ + i sin θ )3
region in the third and fourth quadrants bounded by the = cos3 θ + 3i cos2 θ sin θ − 3 cos θ sin2 θ − i sin3 θ
ray from the origin to −∞ on the real axis and the ray Thus
from the origin making angle −45◦ with the positive
direction of the real axis. cos(3θ ) = cos3 θ − 3 cos θ sin2 θ = 4 cos3 θ − 3 cos θ
34. (2 + 5i ) + (3 − i ) = 5 + 4i sin(3θ ) = 3 cos2 θ sin θ − sin3 θ = 3 sin θ − 4 sin3 θ.
35. i − (3 − 2i ) + (7 − 3i ) = −3 + 7 + i + 2i − 3i = 4
36. (4 + i )(4 − i ) = 16 − i 2 = 17 49. a) z = 2/z can be rewritten |z|2 = zz = 2, so is √
satisfied by all numbers z on the circle of radius 2
37. (1 + i )(2 − 3i ) = 2 + 2i − 3i − 3i 2 = 5 − i centred at the origin.
38. (a + bi )(2a − bi) = (a + bi )(2a + bi ) = 2a2 − b2 + 3abi b) z = −2/z can be rewritten |z|2 = zz = −2, which
39. (2 + i )3 = 8 + 12i + 6i 2 + i 3 = 2 + 11i has no solutions since the square of |z| is nonnega-
tive for all complex z.
2−i (2 − i )2 3 − 4i √
40. = 2
= 50. If z = w =√−1, then zw = 1, so zw = 1. √But if we
√
2+i 4−i 5
use√ z = −1 = i √
√ and the same value for w, then
1 + 3i (1 + 3i )(2 + i ) −1 + 7i z w = i 2 = −1 = zw.
41. = =
2−i 4 − i2 5
51. The three cube roots of −1 = cos π + i sin π are of the
1+i 1+i (1 + i )(−3 − 2i ) −1 − 5i form cos θ +i sin θ where θ = π/3, θ = π , and θ = 5π/3.
42. = = =
i (2 + 3i ) −3 + 2i 9+4 13 Thus they are
(1 + 2i )(2 − 3i ) 8+i √ √
43. = =1 1 3 1 3
(2 − i )(3 + 2i ) 8+i +i , −1, −i .
2 2 2 2
44. If z = x + yi and w = u + vi , where x, y, u, and v are
real, then


3π 3π
z + w = x + u + (y + v)i 52. The three cube roots of −8i = 8 cos + i sin
2 2
= x + u − (y + v)i = x − yi + u − vi = z + w. are of the form 2(cos θ + i sin θ ) where θ = π/2,
θ = 7π/6, and
θ = 11π/6. Thus they are
45. Using the fact that |zw| = |z||w|, we have
√ √

 2i, − 3 − i, 3 − i.
z zw zw zw z
= = = = .
w |w|2 |w|2 ww w

53. The three cube


√ √  π π √ 3π 3π
46. z = 3 + i 3 = 2 3 cos + i sin roots of −1 + i = 2 cos + i sin are of the

6 6  4 4
√ 2π 2π 1/6
form 2 (cos θ + i sin θ ) where θ = π/4, θ = 11π/12,
w = −1 + i 3 = 2 cos + i sin
3 3 and θ = 19π/12.


√ 5π 5π
zw = 4 3 cos + i sin 54. The four
√ fourth roots of 4 = 4(cos 0 + i sin 0) are of the
6 6 form 2(cos θ + i sin θ ) where θ = 0,√θ = π/2, π√ , and

 √ √
z √ −π −π √ θ = 3π/2. Thus they are 2, i 2, − 2, and −i 2.
= 3 cos + i sin = −i 3 √
w 2 2 55. The equation z 4 + 1 − i 3 = 0 has
solutions that are the 
√
3π 3π
 √ 2π 2π
47. z = −1 + i = 2 cos + i sin four fourth roots of −1 + i 3 = 2 cos + i sin .
4 4 3 3
 π π Thus they are of the form 21/4 (cos θ + i sin θ ), where
w = 3i = 3 cos + i sin θ = π/6, 2π/3, 7π/6, and 5π/3. They are the complex

2 2 
√ 5π 5π numbers
zw = 3 2 cos + i sin = −3 − 3i
4 4 √   √ 
√  3 1
z 2 π π 1 1 1/4 1/4 1 3
±2 + i , ±2 − i .
= cos + i sin = + i 2 2 2 2
w 3 4 4 3 3

662
 INSTRUCTOR’S SOLUTIONS MANUAL APPENDIX II. (PAGE A-19)

56. The equation z 5 + a 5 = 0 (a > 0) has solutions that are 8. The function w = z 2 = x 2 − y 2 + 2x yi transforms the
−a5
the five fifth roots of = a (cos π + i sin π ); they are line x = 1 to u = 1 − y 2 , v = 2y, which is the parabola
of the form a(cos θ + i sin θ ), where θ = π/5, 3π/5, π , v 2 = 4 − 4u with vertex at w = 1, opening to the left.
7π/5, and 9π/5.
9. The function w = z 2 = x 2 − y 2 + 2x yi transforms the
57. The n nth roots of unity are
line y = 1 to u = x 2 − 1, v = 2x, which is the parabola
v 2 = 4u + 4 with vertex at w = −1 and opening to the
ω1 = 1 right.
2π 2π
ω2 = cos + i sin
n n 10. The function w = 1/z = (x − yi )/(x 2 + y 2 ) transforms
4π 4π the line x = 1 to the curve given parametrically by
ω3 = cos + i sin = ω22
n n
6π 6π 1 −y
ω4 = cos + i sin = ω23 u= , v= .
n n 1 + y2 1 + y2
..
.
This curve is, in fact, a circle,
2(n − 1)π 2(n − 1)π
ωn = cos + i sin = ω2n−1 .
n n 1 + y2
u 2 + v2 = = u,
Hence (1 + y 2 )2

ω1 + ω2 + ω3 + · · · + ωn = 1 + ω2 + ω22 + · · · + ω2n−1 with centre w = 1/2 and radius 1/2.

1 − ω2n 0 11. The function w = e z = e x cos y + i e x sin y transforms
= = = 0.
1 − ω2 1 − ω2 the horizontal strip −∞ < x < ∞, π/4 ≤ y ≤ π/2 to
the wedge π/4 ≤ arg (w) ≤ π/2, or, equivalently, u ≥ 0,
v ≥ u.
Appendix II. Complex Functions 12. The function w = ei z = e−y (cos x + i sin x) transforms
(page A-19) the vertical half-strip 0 < x < π/2, 0 < y < ∞ to the
first-quadrant part of the unit open disk |w| = e−y < 1,
In Solutions 1–12, z = x + yi and w = u + vi , where x, 0 < arg (w) = x < π/2, that is u > 0, v > 0, u2 + v 2 < 1.
y, u, and v are real.
1. The function w = z transforms the closed rectangle 13. f (z) = z 2 = (x + yi )2 = x 2 − y 2 + 2x yi
0 ≤ x ≤ 1, 0 ≤ y ≤ 2 to the closed rectangle 0 ≤ u ≤ 1, u = x 2 − y 2, v = 2x y
−2 ≤ v ≤ 0. ∂u ∂v ∂u ∂v
= 2x = , = −2y = −
2. The function w = z transforms the line x + y = 1 to the ∂x ∂y ∂y ∂x
line u − v = 1. ∂u ∂v
f  (z) = +i = 2x + 2yi = 2z.
3. The function w = z 2 transforms the closed annular sector ∂x ∂x
1 ≤ |z| ≤ 2, π/2 ≤ arg (z) ≤ 3π/4 to the closed annular
14. f (z) = z 3 = (x + yi )3 = x 3 − 3x y 2 + (3x 2 y − y 3 )i
sector 1 ≤ |w| ≤ 4, π ≤ arg (w) ≤ 3π/2.
u = x 3 − 3x y 2 , v = 3x 2 y − y 3
4. The function w = z 3 transforms the closed quarter- ∂u ∂v ∂u ∂v
circular disk 0 ≤ |z| ≤ 2, 0 ≤ arg (z) ≤ π/2 to the closed = 3(x 2 − y 2 ) = , = −6x y = −
three-quarter disk 0 ≤ |w| ≤ 8, 0 ≤ arg (w) ≤ 3π/2. ∂x ∂y ∂y ∂x
∂u ∂v
5. The function w = 1/z = z/|z|2 transforms the closed f  (z) = +i = 3(x 2 − y 2 + 2x yi ) = 3z 2 .
∂x ∂x
quarter-circular disk 0 ≤ |z| ≤ 2, 0 ≤ arg (z) ≤ π/2
to the closed region lying on or outside the circle 1 x − yi
15. f (z) = = 2
|w| = 1/2 and in the fourth quadrant, that is, having z x + y2
−π/2 ≤ arg (w) ≤ 0. x −y
u= 2 2
, v= 2
6. The function w = −i z rotates the z-plane −90◦ , so trans- x +y x + y2
∂u 2
y −x 2 ∂v ∂u −2x y ∂v
forms the wedge π/4 ≤ arg (z) ≤ π/3 to the wedge
= 2 = , = 2 =−
−π/4 ≤ arg (z) ≤ −π/6. ∂x (x + y 2 )2 ∂y ∂y (x + y 2 )2 ∂x
√ ∂u ∂v −(x 2 − y 2 ) + 2x yi −(z)2 1
7. The function w = z transforms the ray arg (z) = −π/3 f  (z) = +i = = = − 2.
(that is, Arg (z) = 5π/3) to the ray arg (w) = 5π/6. ∂x ∂x 2
(x + y ) 2 2 (zz) 2 z

663
 APPENDIX II. (PAGE A-19) R. A. ADAMS: CALCULUS

2 2 −y 2
16. f (z) = e z = e x (cos(2x y) + i sin(2x y)) 21. cos z = 0 ⇔ e zi = −e−zi ⇔ e2zi = −1
u=e x 2 −y 2
cos(2x y), v=e x 2 −y 2
sin(2x y) ⇔ e−2y [cos(2x) + i sin(2x)] = −1
∂u 2 2 ∂v ⇔ sin(2x) = 0, e−2y cos(2x) = −1
= e x −y (2x cos(2x y) − 2y sin(2x y)) =
∂x ∂y ⇔ y = 0, cos(2x) = −1
∂u 2 2 ∂v π 3π
= −e x −y (2y cos(2x y) + 2x sin(2x y)) = − =⇔ y = 0, x = ± , ± , . . .
∂y ∂x 2 2
∂u ∂v Thus the only complex zeros of cos z are its real zeros at
f  (z) = +i
∂x ∂x z = (2n + 1)π/2 for integers n.
2 2
= e x −y [2x cos(2x y) − 2y sin(2x y) 22. sin z = 0 ⇔ e zi = e−zi ⇔ e2zi = 1
+ i (2y cos(2x y) + 2x sin(2x y))]
⇔ e−2y [cos(2x) + i sin(2x)] = 1
x 2 −y 2 z2
= (2x + 2yi )e (cos(2x y) + i sin(2x y)) = 2ze . ⇔ sin(2x) = 0, e−2y cos(2x) = 1
17. e yi= cos y + i sin y (for real y). Replacing y by −y, we ⇔ y = 0, cos(2x) = 1
get e−yi = cos y − i sin y (since cos is even and sin is =⇔ y = 0, x = 0, ±π, ±2π, . . .
odd). Adding and subtracting these two formulas gives
Thus the only complex zeros of sin z are its real zeros at
z = nπ for integers n.
e yi + e−yi = 2 cos y, e yi − e−yi = 2i sin y.
23. By Exercises 20 and 21, cosh z = 0 if and only if
e yi + e−yi e yi − e−yi cos(i z) = 0, that is, if and only if z = (2n + 1)πi /2
Thus cos y = and sin y = . for integer n.
2 2i
Similarly, sinh z = 0 if and only if sin(i z) = 0, that is, if
18. e z+2π i = e x (cos(y + 2π ) + i sin(y + 2π )) and only if z = nπi for integer n.
= e x (cos y + i sin y) = e z .
24. e z = e x+yi = e x cos y + i e x sin y
Thus e z is periodic with period 2πi . So is e−z = 1/e z .
Since ei(z+2π ) = e zi+2π i = e zi , therefore ezi and also e−z = e−x−yi = e−x cos y − e −x sin y
e−zi are periodic with period 2π . Hence e z + e−z e x + e−x e x − e−x
cosh z = = cos y + i sin y
2 2 2
e zi + e−zi e zi − e−zi = cosh x cos y + i sinh x sin y
cos z = and sin z =
2 2i Re(cosh z) = cosh x cos y, Im(cosh z) = sinh x sin y.
are periodic with period 2π , and e z − e−z e x − e−x e x + e−x
25. sinh z = = cos y + i sin y
2 2 2
e z + e−z e z − e−z = sinh x cos y + i cosh x sin y
cosh z = and sinh z =
2 2 Re(sinh z) = sinh x cos y, Im(cosh z) = cosh x sin y.

are periodic with period 2πi . 26. e i z = e−y+xi = e−y cos x + i e−y sin x

d d e zi + e−zi i e zi − e−zi e−i z = e y−xi = e y cos x − i e y sin x

19. cos z = = = − sin z ei z + e−i z e−y + e y e−y − e y
dz dz 2 2 cos z = = cos x + i sin x
d d e zi − e−zi i e zi + e−zi 2 2 2
sin z = = = cos z = cos x cosh y − i sin x sinh y
dz dz 2i 2i
d d e z + e−z e z − e−z Re(cos z) = cos x cosh y, Im(cos z) = − sin x sinh y
cosh z = = = sinh z
dz dz 2 2 ei z − e−i z e−y − e y e−y + e y
sin z = = cos x + i sin x
d d e z − e−z e z + e−z 2i 2i 2i
sinh z = = = cosh z
dz dz 2 2 = sin x cosh y + i cos x sinh y
ei z + e−i z Re(sin z) = sin x cosh y, Im(sin z) = cos x sinh y.
20. cosh(i z) = = cosh z
2 27. z 2 + 2i z = 0 ⇒ z = 0 or z = −2i
1 e −e z
i z −i
−i sinh(i z) = = sin z 28. z 2 − 2z + i = 0 ⇒ (z − 1)2 = 1 − i
i 2

−z z √ 7π 7π
e +e = 2 cos + i sin
cos(i z) = = cosh z 4 4
2

−z
e −e z −e−z + e z 7π 7π
sin(i z) = =i = i sinh z ⇒ z = 1 ± 21/4 cos + i sin
2i 2 8 8

664
 INSTRUCTOR’S SOLUTIONS MANUAL APPENDIX II. (PAGE A-19)

29. z 2 + 2z + 5 = 0 ⇒ (z + 1)2 = −4 35. Since P(z) = z 5 + 3z 4 + 4z 3 + 4z 2 + 3z + 1 has real

⇒ z = −1 ± 2i coefficients, if z1 = i is a zero of P(z), then so is
z 2 = −i . Now
30. z 2 − 2i z − 1 = 0 ⇒ (z − i )2 = 0
⇒ z = i (double root) (z − z 1 )(z − z 2 ) = (z − i )(z + i ) = z 2 + 1.

31. z 3 − 3i z 2 − 2z = z(z 2 − 3i z − 2) = 0 By long division (details omitted) we discover that

⇒ z = 0 or z 2 − 3i z − 2 = 0


3 2 1 z 5 + 3z 4 + 4z 3 + 4z 2 + 3z + 1
⇒ z = 0 or z − i =− = z 3 + 3z 2 + 3z + 1
2 4 z2 + 1


3 1 = (z + 1)3 .
⇒ z = 0 or z = ± i
2 2
⇒ z = 0 or z = i or z = 2i Thus P(z) has the five zeros: i , −i , −1, −1, and −1.

32. z 4 − 2z 2 + 4 = 0 ⇒ (z 2 − 1)2 = −3
√ √
z2 = 1 − i 3 or z2 = 1 + i 3

 
5π 5π π π
z 2 = 2 cos + i sin , z 2 = 2 cos + i sin 36. Since P(z) = z 5 − 2z 4 − 8z 3 + 8z 2 + 31z − 30 has real
3 3 3 3 coefficients, if z1 = −2 + i is a zero of P(z), then so is


√ 5π 5π z 2 = −2 − i . Now
z = ± 2 cos + i sin , or
6 6
√  π π 
z = ± 2 cos + i sin (z − z 1 )(z − z 2 ) = z 2 + 4z + 5.
 6  6  
3 i 3 i
z=± −√ , z=± +√ By long division (details omitted) we discover that
2 2 2 2

33. z 4 + 1 = 0 ⇒ z 2 = i or z 2 = −i z 5 − 2z 4 − 8z 3 + 8z 2 + 31z − 30
1+i 1−i z 2 + 4z + 5
⇒ z=± √ , z=± √ 3 2
= z − 6z + 11z − 6.


2  2
4 1+i 1−i
z +1 = z− √ z− √
Observe that z3 = 1 is a zero of z 3 − 6z 2 + 11z − 6. By

2 
2 
1+i 1−i long division again:
× z+ √ z+ √
 2  2 

1 2 1 1 2 1 z 3 − 6z 2 + 11z − 6
= z−√ + z+ √ + = z 2 − 5z + 6 = (z − 2)(z − 3).
2 2 2 2 z−1
2
√ 2
√
= (z − 2z + 1)(z + 2z + 1)
Hence P(z) has the five zeros −2 + i , −2 − i , 1, 2, and
3.
34. Since P(z) = z4 − 4z 3
+
√ 12z 2
− 16z + 16 has real
coefficients, if z1 = 1 − 3i is a zero of P(z), then so is
z 1 . Now

(z − z 1 )(z − z 1 ) = (z − 1)2 + 3 = z 2 − 2z + 4.
37. If w = z 4 + z 3 − 2i z − 3 and |z| = 2, then |z 4 | = 16 and
By long division (details omitted) we discover that
|w − z 4 | = |z 3 − 2i z − 3| ≤ 8 + 4 + 3 = 15 < 16.
z 4 − 4z 3 + 12z 2 − 16z + 16
= z 2 − 2z + 4.
z 2 − 2z + 4 By the mapping principle described in the proof of Theo-
rem 2, the image in the w-plane of the circle |z| = 2 is a
Thus z 1 and z 1 are both double zeros of P(z). These closed curve that winds around the origin the same num-
are the only zeros. ber of times that the image of z4 does, namely 4 times.

665
 APPENDIX II. (PAGE A-19) R. A. ADAMS: CALCULUS

Appendix III. Continuous Functions 7. Suppose n is a positive integer and a > 0.

(page A-25)
1. To be proved: If a < b < c, f (x) ≤ g(x) for a ≤ x ≤ c,
limx→b f (x) = L, and limx→b g(x) = M, then L ≤ M.
Proof: Suppose, to the contrary, that L > M. Let
Let  > 0 be given. Let b = a1/n , and let
δ = min{a(1 − 2−n ), bn−1 }.
If |x − a| < δ, then x > a/2n , and if y = x 1/n , then
y > b/2. Thus

1/n

 = (L − M)/3, so  > 0. There exist numbers δ1 > 0

x − a 1/n
= |y − b|
and δ2 > 0 such that if a ≤ x ≤ b, then |y n − bn |
=
y n−1
+ y n−2 b + · · · + bn−1
|x − b| < δ1 ⇒ | f (x) − L| < 
|x − a| bn−1 
|x − b| < δ2 ⇒ |g(x) − M| < . < n−1 < n−1 = .
b b
Thus if |x − b| < δ = min{δ1 , δ2 , b − a, c − b}, then
Thus lim x→a x 1/n = a 1/n , and x 1/n is continuous at
L−M x = a.
f (x)− g(x) > L − − M − = L − M −2 = > 0.
3 8. By Exercise 5, x m is continuous everywhere. By Exer-
cise 7, x 1/n is continuous at each a > 0. Thus for a > 0
This contradicts the fact that f (x) ≤ g(x) on [a, b]. we have
Therefore L ≤ M.  m  m
2. To be proved: If f (x) ≤ K on [a, b) and (b, c], and if lim x m/n = lim x 1/n = lim x 1/n
x→a x→a x→a
limx→b f (x) = L, then L ≤ K .
= (a 1/n )m = a m/n ,
Proof: If L > K , then let  = (L − K )/2; thus  > 0.
There exists δ > 0 such that δ < b − a and δ < c − b, and and x m/n is continuous at each positive number.
such that if 0 < |x − b| < δ, then | f (x) − L| < . In this
case
9. If m and n are integers and n is odd, then
L−K (−x)m/n = cx m/n , where c = (−1)m/n is either −1 or 1
f (x) > L −  = L − > K, depending on the parity of m. Since x m/n is continuous
2
at each positive number a, so is cx m/n . Thus (−x)m/n is
which contradicts the fact that f (x) ≤ K on [a, b) and
continuous at each positive number, and x m/n is continu-
(b, c]. Therefore L ≤ K .
ous at each negative number.
3. Let  > 0 be given. Let δ =  1/r , (r > 0). Then
If r = m/n > 0, then limx→0+ x r = 0 by Exercise
3. Hence limx→0− x r = (−1)r limx→0+ x r = 0, also.
0<x <δ ⇒ 0 < x r < δr = .
Therefore limx→0 x r = 0, and x r is continuous at x = 0.
Thus lim x→0+ x r = 0. 10. Let  > 0 be given. Let δ = . If a is any real number
then
4. a) Let f (x) = C, g(x) = x. Let  > 0 be given and let

δ = . For any real number x, if |x − a| < δ, then

|x| − |a|
≤ |x − a| <  if |x − a| < δ.
| f (x) − f (a)| = |C − C| = 0 < ,
Thus lim x→a |x| = |a|, and the absolute value function is
|g(x) − g(a)| = |x − a| < δ = .
continuous at every real number.
Thus lim x→a f (x) = f (a) and limx→a g(x) = g(a), 11. By the definition of sin, Pt = (cos t, sin t), and
and f and g are both continuous at every real num- Pa = (cos a, sin a) are two points on the unit circle
ber a. x 2 + y 2 = 1. Therefore
5. A polynomial is constructed by adding and multiplying
|t − a| = length of the arc from Pt to Pa
finite numbers of functions of the type of f and g in
Exercise 4. By Theorem 1(a), such sums and products > length of the chord from Pt to Pa

are continuous everywhere, since their components have = (cos t − cos a)2 + (sin t − sin a)2 .
been shown to be continuous everywhere.
6. If P and Q are polynomials, they are continuous every- If  > 0 is given, and |t − a| < δ = , then the above
inequality implies that
where by Exercise 5. If Q(a) = 0, then
P(x) P(a)
limx→a = by Theorem 1(a). Hence P/Q is | cos t − cos a| ≤ |t − a| < ,
Q(x) Q(a)
continuous everywhere except at the zeros of Q. | sin t − sin a| ≤ |t − a| < .

666
 INSTRUCTOR’S SOLUTIONS MANUAL APPENDIX IV. (PAGE A-30)

Thus sin is continuous everywhere. Thus g is continuous and increasing on Ê .

If f is continuous on [a, b], then
12. The proof that cos is continuous everywhere is almost
  f (x)
identical to that for sin in Exercise 11.
 a a h(x) = g f (x) =
1 + | f (x)|
13. Let a > 0 and  > 0. Let δ = min , .
2 2 is also continuous there, being the composition of contin-
a 1 2
If |x − a| < δ, then x > , so < whenever t is uous functions. Also, h(x) is bounded on [a, b], since
2 t a
 

between a and x. Thus

| f (x)|

g f (x)
≤ ≤ 1.
1 + | f (x)|
| ln x − ln a|
By assumption in this problem, h(x) must assume max-
1 imum and minimum values; there exist c and d in [a, b]
= area under y = between t = a and t = x
t such that
2 2 a      
< |x − a| < = . g f (c) ≤ g f (x) ≤ g f (d)
a a 2
for all x in [a, b]. Since g is increasing, so is its inverse
Thus lim x→a ln x = ln a, and ln is continuous at each
g −1 . Therefore
point a in its domain (0, ∞).
f (c) ≤ f (x) ≤ f (d)
14. Let a be any real number, and let  > 0 be given. As-
sume (making  smaller if necessary) that  < ea . Since for all x in [a, b], and f is bounded on that interval.



     2
ln 1 − a + ln 1 + a = ln 1 − 2a < 0, Appendix IV. The Riemann Integral
e e e
(page A-30)
      
we have ln 1 + a < − ln 1 − a . 1 if 0 ≤ x ≤ 1
e e 1. f (x) =
 0 if 1 < x ≤ 2
 
Let δ = ln 1 + a . If |x − a| < δ, then Let 0 <  < 1. Let P = {0, 1 − 3 , 1 + 3 , 2}. Then
e   
      L( f, P) = 1 1 − +0+0=1−
ln 1 − a < x − a < ln 1 + a 3
 3
e e   2 
  U ( f, P) = 1 1 − +1 +0 =1+ .
1 − a < e x−a < 1 + a 3 3 3
e e

x−a
 Since U ( f, P) − L( f, P) < , f is integrable on [0, 2].

e − 1
< a
e Since L( f, P) < 1 < U ( f, P) for every , therefore
 2
|e x − ea | = ea |e x−a − 1| < . f (x) d x = 1.
0
Thus lim x→a e x = ea and e x is continuous at every point 
1 if x = 1/n (n = 1, 2, 3, . . .)
a in its domain. 2. f (x) =
0 otherwise
If P is any partition of [0, 1] then L( f, P) = 0. Let
15. Suppose a ≤ xn ≤ b for each n, and lim xn = L. Then
0 <  ≤ 2. Let N be an integer such that
a ≤ L ≤ b by Theorem 3. Let  > 0 be given. Since 2
f is continuous on [a, b], there exists δ > 0 such that if N + 1 > ≥ N . A partition P of [0, 1]

a ≤ x ≤ b and |x − L| < δ then | f (x) − f (L)| < . Since can be constructed so that the first two points of P
lim xn = L, there exists an integer N such that if n ≥ N  1
then |xn − L| < δ. Hence | f (xn ) − f (L)| <  for such n. are 0 and , and such that each of the N points
2 n
Therefore lim( f (xn ) = f (L). (n = 1, 2, 3, . . . , n) lies in a subinterval of P having
 1
t length at most . Since every number with n a pos-
16. Let g(t) = . For t = 0 we have 2N   n
1 + |t|
itive integer lies either in 0, or one of these other
2
1 + |t| − t sgn t 1 + |t| − |t| 1 N subintervals of P, and since max f (x) = 1 for these
g  (t) = = = > 0. subintervals and max f (x) = 0 for all other subintervals
(1 + |t|)2 (1 + |t|)2 (1 + |t|)2
 
of P, therefore U ( f, P) ≤ + N = . By Theorem
If t = 0, g is also differentiable, and has derivative 1: 2 2N
3, f is integrable on [0, 1]. Evidently
 1
g(h) − g(0) 1
g  (0) = lim = lim = 1. f (x) d x = least upper bound L( f, P) = 0.
h→0 h h→0 1 + |h| 0

667
 APPENDIX IV. (PAGE A-30) R. A. ADAMS: CALCULUS


1/n if x = m/n in lowest terms
3. f (x) = Let P be the common refinement of P1 and
0 otherwise P2 . Then the above inequalities hold with P re-
Clearly L( f, P) = 0 for every partition P of [0, 1]. placing P1 and P2 . If m 1 ≤ f (x) ≤ M1
Let  > 0 be given. To show that f is integrable we and m 2 ≤ g(x) ≤ M2 on any interval, then
must exhibit a partition P for which U ( f, P) < . We m 1 + m 2 ≤ f (x) + g(x) ≤ M1 + M2 there. It
can assume  < 1. Choose a positive integer N such follows that
that 2/N < . There are only finitely many integers n
such that 1 ≤ n ≤ N . For each such n, there are only U ( f + g, P) ≤ U ( f, P) + U (g, P),
finitely many integers m such that 0 ≤ m/n ≤ 1. There- L( f, P) + L(g, P) ≤ L( f + g, P).
fore there are only finitely many points x in [0, 1] where
f (x) > /2. Let P be a partition of [0, 1] such that all Therefore
these points are contained in subintervals of the partition
having total length less than /2. Since f (x) ≤ 1 on U ( f + g, P) −  ≤ I + J ≤ L( f + g, P) + .
these subintervals, and f (x) < /2 on all other subinter-  b 
vals P, therefore U ( f, P) ≤ 1 × (/2) + (/2) × 1 = , f (x) + g(x) d x = I + J .
1 Hence
and f is integrable on [0, 1]. Evidently 0 f (x) d x = 0, a
since all lower sums are 0. d) Assume a < b < c; the other cases are similar. Let
 > 0. If
I∗ − I ∗  b  c
4. Suppose, to the contrary, that I∗ > I ∗ . Let  = ,
3 f (x) d x = I, and f (x) d x = J,
so  > 0. By the definition of I∗ and I ∗ , there exist a b
partitions P1 and P2 of [a, b], such that L( f, P1 ) ≥ I∗ − 
and U ( f, P2 ) ≤ I ∗ + . By Theorem 2, then there exist partitions P1 of [a, b], and P2 of
L( f, P1 ) ≤ U ( f, P2 ), so [b, c] such that

L( f, P1 ) ≤ I < L( f, P1 ) +
3 = I∗ − I ∗ ≤ L( f, P1 ) +  − U ( f, P2 ) +  ≤ 2. 2

L( f, P2 ) ≤ J < L( f, P2 ) +
Since  > 0, it follows that 3 ≤ 2. This contradiction 2
shows that we must have I∗ ≤ I ∗ . (with similar inequalities for upper sums). Let P be
the partition of [a, c] formed by combining all the
5. Theorem 3 of Section 6.4: Proofs of parts (c)–(h). subdivision points of P1 and P2 . Then

c) Multiplying a function by a constant multiplies all L( f, P) = L( f, P1 )+L( f, P2 ) ≤ I + J < L( f, P)+.

its Riemann sums by the same constant. If the con-
stant is positive, upper and lower sums remain upper Similarly, U ( f, P) −  < I + J ≤ U ( f, P). Therefore
and lower; if the constant is negative upper sums  c
become lower and vice versa. Therefore f (x) d x = I + J.
a
 b  b
A f (x) d x = A f (x) d x.
a a e) Let
 b  b
It therefore remains to be proved only that the in- f (x) d x = I, and g(x) d x = J,
tegral of a sum of functions is the sum of the inte- a a
grals. Suppose that
where f (x) ≤ g(x) on [a, b]. We want to show that
 b  b I ≤ J . Suppose, to the contrary, that I > J . Then
f (x) d x = I, and g(x) d x = J. there would exist a partition P of [a, b] for which
a a
I−J I−J
If  > 0, then there exist partitions P1 and P2 of I < L( f, P) + , and U (g, P) − < J.
2 2
[a, b] such that
I+J
Thus L( f, P) > > U (g, P) ≥ L(g, P).
  2
U ( f, P1 ) − ≤ I < L( f, P1 ) + However, f (x) ≤ g(x) on [a, b] implies that
2 2
  L( f, P) ≤ L(g, P) for any partition. Thus we have
U (g, P2 ) − ≤ J < L(g, P2 ) + . a contradiction, and so I ≤ J .
2 2

668
 INSTRUCTOR’S SOLUTIONS MANUAL APPENDIX IV. (PAGE A-30)

f) Since −| f (x)| ≤ f (x) ≤ | f (x)| for any x, we have 7. Suppose f is uniformly continuous on [a, b]. Taking
by part (e), if a ≤ b,  = 1 in the definition of uniform continuity, we can find
a positive number δ such that | f (x)− f (y)| < 1 whenever
 b  b  b x and y are in [a, b] and |x − y| < δ. Let N be a positive
− | f (x)| d x ≤ f (x) d x ≤ | f (x)| d x. integer such that h = (b − a)/N satisfies h < δ.
a a a
If xk = a + kh, (0 ≤ k ≤ N ), then each of the subintervals


 of the partition P = {x0 , x1 , . . . , x N } has length less than

b
b
Therefore

f (x) d x

≤ | f (x)| d x. δ. Thus
a a

g) By parts (b), (c) and (d), | f (xk ) − f (xk−1 )| < 1 for 1 ≤ k ≤ N.

 a  0  a By repeated applications of the triangle inequality,

f (x) d x = f (x) d x + f (x) d x
−a
−a
a
0
 a | f (xk−1 ) − f (a)| = | f (xk−1 ) − f (x0 )| < k − 1.
= f (−x) d x + f (x) d x
0 a 0 If x is any point in [a, b], then x belongs to one of the
= [ f (−x) + f (x)] d x. intervals [xk−1 , xk ], so, by the triangle inequality again,
0
| f (x)− f (a)| ≤ | f (x)− f (xk−1 )|+| f (xk−1 )− f (a)| < k ≤ N.
If f is odd, the last integral is 0. If f is even, the
a
last integral is 2 f (x) d x. Thus both (g) and (h) Thus | f (x)| < | f (a)| + N , and f is bounded on [a, b].
0
are proved. 8. Suppose that | f (x)| ≤ K on [a, b] (where K > 0), and
that f is integrable on [a, b]. Let  > 0 be given, and let
δ = /K . If x and y belong to [a, b] and |x − y| < δ,
6. Let  > 0 be given. Let δ =  2 /2. Let 0 ≤ x ≤ 1 and then
0√≤ y ≤ 1. If √x <  2 /4 and y <  2 /4 then
√ √
 x  y

| x − y| ≤ x + y < .

If |x − y| < δ and either x ≥ 2 /4 or y ≥  2 /4 then |F(x) − F(y)| =

f (t) dt − f (t) dt

a x
a



√ √ |x − y| 2 2 =

f (t) dt

≤ K |x − y| < K = .
| x − y| = √ √ < × = . y K
x+ y  2

√ (See Theorem 3(f) of Section 6.4.) Thus F is uniformly

Thus f (x) = x is uniformly continuous on [0, 1]. continuous on [a, b].

669