V.A. Krechmar-A Problem Book in Algebra-Mir Publishers, Moscow (1978)

3A~AqHHK
no
AJifEBPE
V.A. KRECHMAR
A PROBLEM
BOOK
IN
ALGEBRA
Translated from the nussum
by
Victor Shiffer
Translat10n editor Leornd Levant
MIR PUBLISHERS
MOSCOW
First published 197 4

Second printing 1978
TO THE READER
Mir Publishers would be grateful
for your comments on the content,
translation and design of this book.
We would also be pleased to receive
any other suggestions you may wish
to make ..
Our address is:
USSR, 129820, Moscow I-110, GSP
Pervy Rizhsky Pereulok, 2
MIR PUBLISHERS
Ha
English
aue.iuiicKOM J1a1>1Ke
translation, Mir Publishers, 1974
Printed in the Union of Soviet Socialist Republics
CONTENTS
1. Whole Rational Expressions
2.
3.
4.
5.
6.
7.
8.
9.
10.
Solutions to Section 1
Rational Fractions
Radicals. Inverse Trigonometric Functions. Logarithms
Equations and Systems of Equations of the First Degree
Equations and Systems of Equations of the Second Degree
Complex Numbers and Polynomials
Progressions and Sums
Inequalities
Mathematical Induction
Limits
7
117
15
136
28
174
40
208
53
247
64
285
83
361
93
396
104
450
110
480
PROBLEMS
1. WHOLE RATIONAL EXPRESSIONS

The problems presented in this section are mainly on the
identity transformations of whole rational expressions.
These are the elementary operations we have to use here:
addition, multiplication, division and subtraction of monomials and polynomials, as well as raising them to various
powers and resolving them into factors. As regards trigonometric problems, we take as known the definition of trigoncmetric functions, principal relationships between these
functions, all th::i properties connected with their periodicity, and all corollaries of the addition theorem.
Attention should be drawn only to the formulas which
enable us to transform a product of trigonometric functions
into a sum or a difference of these functions. Namely:
1
cos A cos B=-:r[cos (A+B) +cos (A-B)],

sin A cos B =+[sin (A+ B) +sin (A-B)],
sin A sin B =-}[cos (A-B)-cos (A+ B)].
1. Prove the identity

(a 2 + b2) (x 2 + y2) = (ax -
by) 2 + (bx + ay) 2.
2. Show that
(a2 + b2 + c2 + d2) (x2 + y2 + z2 + t2) =
= (ax - by - cz - dt) 2 + (bx + ay - dz + ct) 2 +
+ (ex+ dy + az - bt) 2 + (dx - cy + b~ + at) 2.
3. Prove that f~om the equalities
ax - by - cz - dt = 0, bx
ay - dz
ct = 0,
ex+ dy
+ az -
bt = 0,
dx - cy
+ bz +at =
0,
Problems
follows either a = b = c = d = 0, or x = y = z = t = 0.
4. Show that the following identity takes place
(a2
b2
c2) (x2
y2
z2) - (ax
by
cz)2 =
= (bx - ay) 2
(cy - bz) 2
(az - cx) 2.
+ +
+ +
+
5. Show that the preceding identity can be generalized in

the following way
(ai + a; + . . . + a;)
= (a1b1
+ b; + . . . + b;) =
+ a2b2 + ... + anbn) 2 + (a1b2 - a2b1) 2 +
+ (a1b3 - a3b1) 2 + ... + (an-1bn - anbn-1)
(b~
6. Let
n (a 2
b2
+ + c2 + ... + l 2) =
=
(a
+ b + c + ... + l) 2 ,
where n is the number of the quantities a, b, c, ... , l.

Prove that then
a= b = c = ... = l.
7. Prove that from the equalities

a~ + a; + ... + a; = 1, bi + b:
follows
-1 :::;;; a1b1
+ ... + b; =
+ a2b2 + ... + anbn:::;;; + 1.
8. Prove that from the equality
+ (z - x) 2 + (x - y) 2 =
= (y + z - 2x) 2 + (z + x -
(y - z) 2
follows
2y) 2
+ (x + y -
2z) 2
z.
9. Prove the following identities

(a2 -
+ (2ab)2 = (a2 + b2)2,

(6a
4ab + 4b 2) = (3a 2 + 5ab - 5b 2) +
+ (4a 2 - 4ab + 6b 2) + (5a 2 - 5ab - 3b 2)
b2)2
2 -
10. Show that

(p2 - q2)4
(2pq
+ q2)4 + (2pq + p2)4 =
2 (p2+pq+q2)".
x + XY + Y = za
q3 + 3pq2 - p3' y = -3pq (p + q)'
z = p2 + pq + q':.,
2
if
12. Prove that

(3a
+ 3b)k + (2a + 4b)k + ak + bk =

= (3a + 4b)k + (a + 3b)" + (2a -+ b)h
at k = 1, 2, 3.
13. 1 Show that if x
(ix - ky)n
+ (iy
-+ y + z
= 0, then
- kz)n
(iz - kxr =
= (iy - kx)n (iz - kyr
+ (ix -
kz)n
at n = 0, 1. 2, 4.
2 Prove that
:.en
at n
+ (x + 3)n + (x + 5)n + (x + 6r + (x + 9r +
+ (x + 1or + (x + 12r + (x + 1sr =
= (x + 1r + (x + 2r + (x + 4t + (x + 7t +
+ (x + sr + (x + 11r + (x + 13t + (x + 14r
= 0, 1, 2, 3.
14. Prove the identities

1 (a+ b + e + d) 2 +(a
+ b - e - d) +
+ (a + e - b - d) + (a + d - b - e) =
= 4 (a + b + e + d
2
2 (a
b2 + e - d2) 2 + 2 (ab - be + de + ad) 2 =
= (a 2 + b2 + e2 + d
2 (ab - ad + be + de)
2
2
2
2
2
3 (a - e + 2bd) + (d - b + 2ae) 2 =
= (a 2 - b2 + e2 - d + 2 (ab - be + de + ad) 2.
2
2 );
2 -
2) 2 -
2;
2) 2

(a+ b + e)' + (b + e - a)'+ (e +a - b) 4 +
+ (a + b - e)' = 4 (a 4 + b4 + e4) +
+ 24 (b 2e2 + e2a2 + a 2 b2).
10
Prolilems
16. Let s = a + b + c.
Prove that
s (s - 2b) (s - 2c) + s (s - 2c) (s - 2a) +
s (s - 2a) (s - 2b) = (s - 2a) (s- 2b) (b - 2c)-'-8abc.
17. Prove that if a+ b + c = 2s, then

a (s - a) 2 + b (s - b) 2 + c (s - c) 2 + 2 (s - a) X
X (s - b) (s - c) = abc.
18. Put
2s =a
b + c; 2a2 = a 2 + b2 + c2.
Show that
(a2 - a2) (a2 - b2) + (a2 - b2) (a2 - c2) +
+ (a2 - c2) (a2 - a 2) = 4s (s - a) (s - b) (s - c).
19. Factor the following expression

(x
+ Y + z)s -
x3 - ys - z9.

x9 + y9 + z9 - 3xyz.
21. Simplify the expression
(a + b + c) 9 - (a + b - c) 3 -
(b
+c-
a) 3
- (c
a-
b) 9
+ (c - a) + (a - b)
Show that if a + b + c = 0, then
a + b + c3 = 3abc.
Prove that if a + b + c = 0, then
(a 2 + b2 + c2) 2 = 2 (a' + b' + c')'.
(b -
23.
c) 9
9
24.
25. Show that

[(a- b) 2
(b - c) 2
+ (c =
a) 2]2 =
2 [(a - b)'
+ (b -
c)'
+ (c -
a)'].
26. Let a
+ b+
11
c = 0, prove that
+ + c = 5abc (a + b + c
5 (a + b + c3) (a + b + c = 6 (a + b + c
10 (a + b + c = 7 (a + b + c (a + b + c
1 2 (a
b5
2
3
5)
7)
2 );
2)
27. Given 2n numbers: a 1, a 2 ,
2)
an; b1, b 2 ,
5 );
5 ).
bn. Put
Prove that
a1b1
+ a2b2 + ... + a
11
+ (a2 -- aa) s2 +
- an) Sn-I + anSn
bn = (a1 - a 2) s,
+ + (an-I
28. Put
a1 + a 2
Prove that
(s - a 1) 2 + (s - a2) 2
+ ... + an
+ .. + (s -
n
2
s.
an) 2 =
=a~+ a~+ ..
+ 2Bxy + Cy
y = yx' + By'.
29. Given a trinomial Ax2

Put
ax'
+ a~.
+ ~y',
Then the given trinomial becomes
A'x' 2
Prove that
B' 2 30. Let
Pi
and
+ 2B'x'y' + C'y'
A'C' = (B 2
+ qi
= 1
P1+P2+ +Pn
P=
Prove that
P1q1
+ P2q2 + + Pnqn =
(i
'
~y) 2
AC) (afJ -
= 1, 2, ... ,
n)
npq -(P1 - p) 2 - (P2 - P)2- -(Pn-P) 2.
Problems
31. Prove that

1
1
1
1
1
1
T 2n-1+32n-3 + + 2n-1 T =
32. Let Sn= 1 +
!+
! (1 +
+ ... +
!.
!+
+2n~1}
Show that
1
Sn=n-(~+;+
... +n: 1 );
n- 1
r> - 2
2
1 )
2 nsn=n+ -1-+-2
-+ + n-2+ n-1
0

1
1
1
1
1
1
1 - 2 + 3 - 4 + + 2n-1 - 2n" = n + 1
n+ 2
+ -11
+2n
34. Prove
( 1+
a~1}
2a~1 ) ( 1 + 3a~1 ) X X
1
1
x(1+ (2n-1)a-1
)(1 - 2na-1
)-(1-
(n+i)a
(n+2)a.
(n+1) a-1
(n+2) a-1
(n+n)a
(n+n)a-1
35. Let [ex] denote the whole number nearest to ex which

is less than or equal to it. Thus, [ex] ~ ex< [ex) + 1.
Prove that there exists the identity
[ x) + [ x +
~ J+
[x+
! J+ ... + [
x+
n: J= [
1
nx].
36. Prove that

cos (a + b) cos (a - b) = cos 2 a - sin 2 b.
37. Show that
(cos a+ cos b) 2 +(sin a+ sin b) 2 = 4 cos 2
(cos a -cos b) 2 +(sin a - sin b) 2 = 4 sin 2
a;-b,
a;-b.
38. Given
(1
+ sin a) (1 + sin b) (1 + sin c)
cos a cos b cos c.
Simplify
(1 - sin a) (1 -
sin b) (1 -
sin c).
39. Given
+ cos a) (1 + cos P> (1 + cos y)
(1
= (1 -
=
cos a) (1 -
cos P> (1 -
cosy).
Show that one of the values of each member of this equality is

sin a sin P sin y.
40. Show that
+ cos (p + y) sin (p - y) +
+ cos (y + 6) sin (y - 6) + cos (6 + a:) sin (o - a:)
cos (a
p)
sin (a: -
p)
0.
+ b) sin (a - b) sin (c + d) sin (c - d) +

+ sin (c + b) sin (c - b) sin (d + a) sin (d - a) +
+ sin (d + b) sin (d - b) sin (a + c) sin (a - c) =
0.
41. Prove that
sin (a
42. Check the identities:

1 cos (p + y - a:) _j_ cos (y + a: - ~) +
+cos (a+ ~ - y) + cos (a:-+ ~ + v) = 4 cos a: cos ~ cosy;
2 sin (a+ ~ + y) + sin(~+ y - a:) + sin (y+a:-P)- sin (a: + ~ - y) = 4 cos a: cos ~ sin y.
43. Reduce the following <'xpression to a form convenient

for taking logarithms
+ ~ } +sin ( B + ~ } +sin { C + : } +
-+cos (A + ! )+cos ( B + ~ ) +cos ( C +
A+ B + C =:rt.
sin (A
if
: )
14
Problems
44. Reduce the following expression to a form convenient

for taking logarithms
. A
. B+.
C+ cosA+ cosB+ cosC
sm-+smsm4
4
4
4
4
4
if A+ B
C = n.
45. Simplify the product
cos a cos 2a cos 4a ... cos 2n-1a.
46. Show that
n
2n
3n
4n
5n
6n
7n
( 1 )7
cos 15 cos 15 cos T5 cos T5 cos 15cos15cos15 = 2

47. Given sin B =+sin (2A
+B).
Prove that
tan(A+B)=
~ tanA.
48. Let A and B be acute positive angles satisfying the

equalities
3 sin 2 A + 2 sin 2 B = 1,
3 sin 2A - 2 sin 2B=0.
Prove that A + 2B =
49. Show that the magnitude of the expression
cos 2 cp + cos 2 (a + cp) - 2 cos a cos cp cos (a + cp)
is independent of cp.
50. Let
a = cos cp cos' + sin cp sin ' cos t5,

a' =cos cp sin' - sin cp cos' cos tS, a" = sin cp sin tS;
b = sin cp cos' - cos cp sin' cos tS,
b' =sin cp sin' + cos cp cos' cos tS, b" = -cos cp sin tS;
c = -sin' sin 6, c' = cos' sin 6, c" = cos tS.
15
2. Rational Fractions
Prove that
+ a' + a " = 1, b + b' + b" =1,

e + e' + e" = 1,
ab+ a'b' + a"b" = 0, ae + a'e' + a"e" = 0,
be+ b'e' + b"e" = 0.
a2
2. RATIONAL FRACTIONS
Transformations of fractional rational expressions to be
considered in this section are based on standard rules of
operations with algebraic fractions.
Let us draw our attention only to one point which we have
to- use (see Problems 15, 16, 17). If we have a first-degree
binomial in x
Ax+ B
and if we know that it vanishes at two different values of x
(say, at x = a and x = b), then we may state that the
coefficients A and B are equal to zero. Indeed, from the
equalities
Aa
B = 0, Ab+ B = 0
we get
A (a - b) = 0
and since a - b =I= 0, then A = 0. Substituting this value

into one of the equalities (*), we find B = 0. Similarly,
we may assert that if a second-degree trinomial in x
Ax2 +Bx+ C
vanishes at three distinct values of x (say, at x
and x = e), then A = B = C = 0.
Indeed, we then have
Aa2
+ Ba+
C = 0, Ab 2
+ Bb + C =
0,
Ae 2 + Be
a, x = b
+C=
0.
Subtracting term by term, we have
A (a 2
b2)
+ B (a- b)
0, A (a 2
e2)
+B
(a -
e) =0.
16
Problems
Since a - b =I= 0, a - c =I= 0, we have

A (a + b) + B = 0, A (a + c) + B = 0.
Hence A = 0 (since b - c =I= 0), and then we find B = 0
and C = 0.
Analogously, we can show that if a third-degree polynomial
Ai'
Bx 2
Cx
+D
vanishes at four different values of x, then
A= B = C = D = 0,
and, in general, if an nth-degree polynomial vanishes at
n
1 different values of x, then its coefficients are equal
to zero (see Sec. 6).
Finally, considered in this section are a number of problems pertaining finite continued fractions. We take as
known the information on these fractions contained usually
in elementary textbooks.
The principal trigonometric relations used in solving
triangles are also taken here as known.
3_ (
p --- p
p3-2q3 )3
p3 q3
+ ( q 2p3-q3)3
p3 + q3
+ q3.
2. Simplify the following expression
1 (1pa+qs
1) + (p+q)4
3 (1-p2+q21) + (p+q)6
6 (1p-+q-1) .
(p+q)3
3. Simplify
1 (1P4-li'1) + (p+q)4
2 (17-qs1) +
(p+q)S
2 (1[Ji -q'i'
1) .
+ (p+q)6
4. Let
X=
a-b
a+b '
b-c
y= b+c '
c-a
Z=--.
c+a
Prove that
(1
+ x) (1 + y) (1 + z) = (1- x) (1-y) (1-z).
17
5. Show that from the equality
(a+b+c+ d) (a-b-c+d) =(a -b+c-d) (a+b-c-d)

follows
+ +
ax2 by2 ez2

be (y-z)2-j-ca (z-x)2+ab (x-y)2
if
ax+by+cz=O.
7. Prove that the following equality is true
x2y2z2
(x2-a2) (y2-a2) (z2-a2)
(x2-b2) (y2-b2) (z2-b2)
a2b2 +
a2(a2-b2)
+
b2 (b2-a2)
=
=
x2
+Y2+z2-a2-b2.
8. Put
all
(a-b) (a-e)
--.,---,..,....,--.,.-+
bll
(b-a) (b -e)
ell
(e-a) e-b)
= S II
Prove that
8 0 =81=0, 82=1, 8 3 =a+b+c,
8 4 = ab+ac +be+ a2 b2 c2 ,
8 5 = a 3 + b3 + c3 +a2 b+b2a+c 2a + a2 c + b2 c + c2 b + abc.
+ +
9. Let
b11
(b-a) (b-e) (b-d)
ell
dll
(e-a) tc-b) (e-d)
(d-a) (d-b) (d-e) = 811
all
(a-b) (a-c) (a-d)
Show that
8o=S 1 =8 2 =0, 8 3 =1,
S 4 =a+b+c+d.
10. Put
O'm =a
m (a+b) (a+e) + bm (b+e) (b--l a)+ m (e+a) (e+b)

(a-b) (a-e)
(b-e) (b-a)
c (e-a) (c- b)'
Compute
O'i.
a 2,
0' 3
and a 4
18
Problems

be (a-a.) (a-P) (a-y) +ca (b-a.)(b-P) (b-y)
(b-a) (b-c)
(a-b) (a-c)
+ab
(c-a.) (c-~) (c-y)

=abc-a~'\'
(c-a) (c-b)
12. Show that

a2b2c2
a2b2d2
(a-c) (b-c) (d-c)
a2c2d2
b2c2d2
(a-b) (c-b) (d-b)
(b-a) (c-a) (d-a)
(a-d) (b-d) (c-d)
= abc + abd + acd + brd.

13. Simplify the following expressions
10
20
+ c (c-a)1 (c-b)'
.
1
1
+ c2(c-a)(c--b).
1
a2(a-b)(a-c)+ b2(b-a)(b-c)
1
a (a-b) (a-c)
b (b-a) (b-c)

ah.
(a-b) (a-c) (x-a)
bh
(b-a) (b-c) (x-b)
+
ch
(c-a) (c-b) (x-c) '
where k= 1, 2.
15. Show that
b+c+d
(b-a) (c-a) (d-a) (x-a)
_L
a+a+~
(d--c) (a-c) (b-c) (x-c)
c+d+a
(c-b) (d-b) (a-b) (x-b)
a+b+c
(a-d) (b-d) (c-d) (x-d)
x-a-b-c-d
(x-a) (.x-b) (x-c) (x-d)

a2 (:r-b) (x-c)
b2 (x-c) (x-a)
(a-b) (a-c)
(b-c) (b-a)
z (.x-a) (x-b)
(c-a.) (c-b)
17. Prove the id entity
(x-b) (x-c)
(x-c) (.x-a)
(a-b) (a-c) (b-c) (b-a)
+ (x-a) (x-b) =
(c-11) (c-b)
1.
x2.
19
18. Show that if a+ b

( a-b
c
+ b-c
a
+c =
0, then
+~) (-c-+_a_+_b-)= 9 .
b
a-b
b-c
c-a

a-b
a+b
b-c
b+c
c-a
(a-b) (b-c)(c-a)
(a+b) (b+c)(c+a)
+ c+a +
20. Prove that

b-c
(a-b) (a-c)
c-a
a-b
+ (b-c) (b-a) + (c-a) (c-b)

=-2-+_2_+~.
a-b
b-c
c-a

a2-bc
(a+b) (a+c)
b2-ac
c2-ab
+ (b+c) (b+a) + (c-j-a) (c+b)
22. Prove that

dm (a-b) (b-c)+bm (a-d) (c-d)
cm (a-b) (a-d)+am (b-c) (c-d)
b-d
a-c
at m= 1, 2.
23. Prove that
{ 1--=-+x(x-a1)
a1
a1a2
x(x-a1)(x-a2)
a1a2a3
+ ... +
+ (_ 1)n _.:_(x-a1) (x-a2) ... (x-an-1) } X

a1a2a3. . an
-=-..+
x(x+a1)
+ -x(x+a1)(x+a2)
1
{ +
- - - - + ... +
a1
a1a2
a1a2a3
x (x+a1) (x+a2) ... (x+an-1)

a1a2a3. an
x2
x2 (x2-ar)
-1--2+
a1
a1a2
...
24. Given
b2+c2-a2
2bc
c2+ a2-h2
2ac
a2+b2-c2
2ab
= 1
20
Prove that two of the three fractions must be equal

to + 1, and the third to -1.
25. Show that from the equality
1
a+b+c= a-f b+c

follows
if n is odd.
26. Show that from the equalities
bz-j-cy
_
x (-ax+by+cz) -
cx+az
_
y (ax-by+cz) -
ay-t bx
z (ax+by-cz)
follows
x
a (b2+ c2-a2)
27. Given
a+~+y=O,
a+b+c=O,
~+1..+:l=O.
a
b
c
Prove that
aa 2 +
~b 2 +yc2 =
0.
28. If
a3+b 3+c3= (b+c) (a+c) (a+b)

and
(b2
+ c -a
2
2)
x = (c 2
+a
2 -b2 )
y = (a2 + b2 -c2 ) z,
then
x3 + y3 + z3 = (x + y) (x + z) (y + z).
29. Consider the finite continued fraction
a 0 +....!...+ 1
a1
Ii;'"+ . .
+-.
an
2. Rational
21
Fractions
Put
Q0 =1,
P 0 =a0 ,
and in general
Q1=a 1
P1=a0a1+1.
+ Pk-1>
Qk+i = ak+1Qk + Qk-1
Ptt+i = ak+tpk
Then, as is known,
0P n
1 +
=ao+a
1
+an
..
(n= 0, 1, 2, 3, ... ).
Prove the following identities
1o
2o
Pn+2 _ 1) ( 1 _ Pn-1 ) = ( On+2 _ 1) ( 1 _

Pn
Pn+I
On
Pn _.!..2_ __
On
Qo - OoQ1
010i
On-1Qn '
1___1_+
+ (-1)n-1.
3 P n+2Qn-2 - P 11-2Qn+2 --= (an+2an+1an
40
Pn
1
--=an+--+
Pn-1
an-I

On-1 ) ;
On+1
+ an+2 +an) ( -
1)n;
. +a;
__!h:__ = a,. +-1- +

On-I
an-I
+ _!__.
a1
30. Put for brevity

ao
+-1+
a1
+an
= (ao,
ai. ... , an)= POnn '
and let the fraction be symmetric, i.e.

ao =an,
a1 = an-1>
Prove that
Pn-1=Qn.
31. Suppose we have a fraction

1
a-+-+1
a
a+
+-.
a
22
Problems
Prove that
32. Let
1
X=-+
1
a
T+.
1
. +y+_!_ 1
a+T+
+-l
and let PQn and PQn- 1 be, respectively, the last and last
n
n-1
but one convergents of the fraction
Prove that
P,.Qn+PnPn-1
x= Q2n+ P nQn-1
33. Consider the continued fraction
b 0 -l-~+
'
bl
-+
b2
a
.. +-2!...
a2
bn .
Put
P 0 =b0 ,
Q0 = 1,
P 1 =b0 b1 +ai.
and in general
Ph+1
= bR+1Ph + a,.+tp,._ 17
Q1<+1 = b,.+1Q1<
Prove that
+ a,.+1Q1<-1.
Q1 =bi. ...
2. Rational
23
Fraction.~
34. Prove that

r
r+1 -
,.n+t_,.
r
r+1
-r+T
r"i-1
r
- r+1
(the number of links in the continued fraction is equal

to n).
35. Prove that

1
1
1
-+-+
... +-=
U2
Un
Ut
u2l
Ut Ut
--f
U2
ll~
u2+u3
u~-1
Un-1+un
36. Prove the equality
where c" c2 , , c11 are arbitrary nonzero quanlilies.

10
sin (n+ 1) x
sin nx
1=2cosx-1
2cosx---
2cosx -
1
-2cos x
(a total of n links);
bn
+1
24
Problems
38. Prove that

1 sin a+ sin b + sin c - sin (a+ b + c) =
. a+u . a+c . b+c .
= 4 sin - 2- sm-2sm - 2-.-,
J
2 cos a+cos b+cosc +cos (a+ b + c) =

=
a+b
b+c
a+c
4 cos - 2 - cos - 2 - cos - 2 - .
39. Show that

sin (a+b+c)
tan a+ tanb +tan c- cos a cos b cos c =tan a tan b tan c.
40. Prove that if A+ B

lowing relationships
+c = n,
then we have the fol-
A
. B+ sm
. C = 4 cos A cos B cos C ;
t c sm
+ sm
2
2
2
2 cos A +cos B +cos C = 1
+ 4 sin
~ sin ~ sin ~ ;
3 tanA+tanB+tanC=lnn.1 lanBtanC;
1
AB
AC
BC
4 tan 2 tan 2 +tan 2 tan 2 + tan 2 tan 2 =1;

5 sin 2A + sin 2B + sin 2C = 4 sin A sin B sin C.
41. Find the algebraic relations between the quantities a, b and c which satisfy the following trigonometric
equalities
. a . b c
t o cos a+ cos b + cos c = 1 4 sm
2 srn 2 sm 2 ;
2 tan a+ tan b +tan c =tan a tan b tan.c;

3 cos 2 a+cos 2 b + cos 2 c-2 cos a cos bcosc = i.
42.. Show that

x
1-x2
_Y_
_z_ _
+ 1-y2 + 1-z2 -
4xys
(1-x2)(1-y2)(1-z2)
if
x!f
+xz+ yz= 1.
25
43. Show that the sum of the three fractions

b-c
1 +be '
c-a
1+ac '
a-b
1 +ab
is equal to their product.

44. Prove that
tan 3a = tan a tan ( ~
+ a) tan ( ~
- a) .
45. Prove that from the equality

sin4 a +
cos4 a
b
follows_ the relationship

sins a +
a3
coss a
b3
46. Suppose we have
a1 cosa 1 + a 2 cosa 2 +
... +an cos an= 0,
a1 cos (a 1+ 8) + a 2 cos (a 2 + 8) + ... +an cos (an+ 8) = 0

(0 =I= kn).
Prove that for any 'A
a1 cos (a 1+'A)+ a 2 cos (a 2 +'A)+ ... +an cos (an+ 'A)= O.
sin (P-'\') + sin (v-a) +sin (a-Pl _ O
cos p cos '\'
cos '\' cos a
cos a cos p -
48. let in a triangle the sides be eriual to a, b and c,

and let
s
r=P '
ra=--,
p-a
r c =p-c
-'
where s is the area of the triangle and 2p =a+ b + c.

Prove the following relationships
2ti
Problems
r;+r;b + r;c )-- 4,

be
ac
+
(a-b) (a-c) r~ + (b-c) (b-a) rg
( a
40
ab
a2
+ (c-a) (c-b) r~
b2
+ (b-c) (b-a) rcra +

5o
ar0
brb
(a-b) (a-r)
(b+c) ra
(a-b) (a-c)
(a+b) re
(c-a) (c-b)
c2
(c-a) (c-b) r 0 rb
+ (b-r) (b-a)
(a-b) (a-c) rbrc
ere
(c-11) (c-b)-
(c+a) rb
=fl;
(b-c) (b-a)
_!!_
49. Prove the idenlily

sin (a+b-c-d) =
sin (a-c) sin (a-d)

sin (a-b)
+
+ sin (b-r) sin (b-d)
sin (b-a)
50. Given
cos e =
b+ c ,
cos cp = a+ c ,
cos ' = a+ b
(0, cp and ' lie between 0 and :n:).

Knowing that a, b and c are the sides of a triangle
whose angles are A, B and C, correspondingly, prove
that
1 tan 2
~ + tan 2 ~
+ tan
! = 1;
21
51. Prove that

1
sin(a-b)sin(a-c)
+ sin(b-a)sin(b-c) +
1
+ sin (c-a) sin (c-b)
1
2 COS -
a-b
2-
COS -
a-c
2-
C'OS -
b-c
2-
1o
sin a
sin(a-b)sin(a-c)
sin b
sin(b-a)sin(b-c)
+
~
~a
sin(a-b)sin(a-c)
sin c
sin (c-a) sin (c-b)
~b
sin(b-a)sin(b-c)
+
=0;
cosc
sin (c-a) sin (c-h)
=0.

1 sin a sin (b-c) cos (b+c-a) +
+ siQ b sin (c-a) cos (c + a-b) +
+sin c sin (a-b) cos (a+ b-c) = O;
2 cos a sin (b-c) sin (b + c-a) +
+cos b sin (c-a) sin (c+ a-b) +
+cos c sin (a- b) sin (a+ b-c) = O;
3 sin a sin (b-c) sin (b+ c-a) +
+sin bsin (c-a) sin (c+ a-b) +
+sin c sin (a - b) sin (a+ b- c) =
= 2 sin (b-c) sin (c- a) sin (a-b);
4 cosasin(b-c)cos(b+c-a}+
+cosbsin(c-a)cos(c+a- b)+
+cos c sin (a -b) cos (a+ b- c) =
= 2 sin (b-c) sin (c-a) sin (a --b).
28
Problems
54. Prove that

1 sin 3 A cos (B-C) + sin3 B cos (C-A) +
+ sin 3 C cos (A- B) = 3 sin A sin B si11 C;
2 sins A sin (B - C) + sin3 B sin (C - A) +
+ sins C sin (A - B) = 0
if A+ B
C = n.
1 sin 3A sins (B - C) + sin 3B sins (C - A)+
sin 3C sins (A - B) = O;
2 sin 3A coss (B - C) + sin 3B coss (C - A)+
sin 3C coss (A - B) = sin 3A sin 3B sin 3C
if
+
A+ B + C =
n.
3. RADICALS. INVERSE
TRIGONOMETRIC FUNCTIONS.
LOG~RITHMS
The symbol ;YA is understood here (if n is odd) as the
only real number whose nth power is equal to A. In this
case A can be either less or greater than zero. If n is even,
then the symbol ;YA is understood as the only positive
number the nth .power of which is equal to A. Here, necessarily, A ~ 0.
Under these conditions, for instance,
VA 2 =A
. VA
if
-A
A>O,
if
A<O.
All the rest of the standard rules and laws governing the
operations involving radicals, fractional and negative
exponents are considered here to be known. Let us also
remind of two formulas which sometimes turn out to be
3. Radicals. Inverse Trigonometric Function.~. Lo{{arithms
29
rather useful in performing various transformations, namely:
_1/
V-A+ V Br
A+VA2-H
2
V A- V B= -.V/
A+- V A2-H
+ -./
V
A-VA2-H
2
'
As far as trigonometric functions are concerned, let us

first of all consider the reduction formulas:
1 The functions sin x and cos x are characterized by the
period 2n, whereas tan x and cot x by the period n so that
we may write the following equalities
+ 2kn) =sin x,
+ kn)= tan x,
sin (x
tan (x
+ 2kn) =cos x,
+ kn)= cot x,
cos (x
cot (x
where k is any whole number (positive, negative or zero).

2 For the functions sin x and cos x the quantity n is the
half-period, i.e. the rejection of the quantity +n in the
argument results in a change in the sign of a function.
Consequently,
sin (x
kn) =(-1)" sin x, cos (x
kn) = (-1)" cos x,
where k is any whole number (positive, negative or zero).

3 The functions sin x, tan x and cot x are odd functions,
and cos x is an even function. Therefore
sin (-x) = -sin x, tan (-x) = -tan x,
cot (-x) = -cot x, cos (-x) =cos x.
4 If x and y are two quantities entering the relationship
then
cos x =sin y,
tan x =cot y,
sin x =cosy,
cot x =tan y.
Using these remarks, we can always reduce sine or cosine

of any argument to sine or cosine of an argument lying
in the interval between 0 and ~ . The same can be said
about tangent and cotangent.
Problems
30
Indeed, any argument a can be written in the following

form
n
a=S-+c.t
0'
2 wheres is an integer, and 0 ~ c.t 0 ~ ~, wherefrom follows
the stated proposition. Let us also mention the following
formulas (k an integer):
sin kn = 0,
tan kn = 0,
O
=
. kn
sm
2
kn
sin - -2 = ( - 1)
kn
cos 2
kn
cos 2
cos /m = (-1)\
if k is even,
k-1
2
if k is odd,
~
=(-1)2
if k is even,
=0
if k is odd.
Further, we use the symbol arcs in x to denote an arc

whose sine is equal to x and which lies in the interval
n
n
between - 2 and + 2 .
Thus, in all cases
~ ~arc sin x ~
+~ .
Similarly
n
- 2
< arc tan x < + 2n
O~arccos x~n,
0 < arccotx < n.

ln this section we also give several problems on transforming expressions containing logarithms.
1. Prove that
2+ V3
2-V3
( V2 +V 2 +-Va + -V2- V 2- -V3
= 2
3. Raaical.~. 1 nverse Trigonometric Functions. Logarithms
31
2. Show that
10 :-' V2-1 =
2
3
ii.}- Vi+ Vi;
v,YS-:14 = ~ <Y2 +,YW-,1'25);

vY28-Y"fl = ! (~-.Y28t);
4o (3+2V5)T = V5+1.
3-2V5
V5-t '
1
.''io ( ~/32_ ~J27)3 = ~/T
W'
(JI"-}+
Vi)
25+
~/3 _ ~/!):
2il
~'
=(1+2+'V8)! =
=~+VJ;+ Vci;- ~.
A
3. L e at - =b- = c- = -d .
Prove that
VAa+YBb+-VCC +VDd=
= V (a+b+c+d) (A+B--1-C+D).
4. Show that
:-' ax2+by2+cz2=ra+Yb +rc

if
a.x3 = by3 = cz 3 and
.!.x + _!_y + _!_z = 1.
5. Put
ll 11
= (1 + ~
hn= (
)n + (1- ~
1+~r-(1-
) n'
0 r.
32
Problems
Show that
bm+n =
+ 2n-"
bm-11
llm 11
6. Let
Un=
[ (
f -( i-2-v'Yi )"]
i+2VS
(n=O, 1, 2, ::l, .. ).
Prove the following relationships
1
2
Un+l=un+Un-1;
Un-I= UkUn-h
-j- Uk-tUn-h-1;
3 Uzn-1 =u~+u~-1;
4 Uan=~+u~+1-uL1;
5 U~ - Un-2Un-1Un+fUn+2 = 1;
6 Un+fUn+z-UnUn+a=(-1t;
UnUn+f -
Un-2Un-1
= Uzn-t

1
1 {2 [a2 +b 2 )"2 -a] [(a 2 +b 2)2 -bl}Z =

1
=a+b-(a2 +b2 )"2 (a> 0, b>O);

I
2 {3[(a3 +b3 )3 -a] [(a3 +b3 )3 -b]}T =

2
=(a+b)T --ta 2 -ab+b2 )3.
8. Compute the expression

1
(1-ax) (1 + axt 1 (1-+ bx)7 (1--bx)-2
at
1
x = a- 1 ( 2
.~
-1) 2
(0 <a< b
< 2a).
3. Radicals. Inverse Trigonometric Functions. Logarithms
33
9. Simplify the expressio11

n3-3n+(n2-1) ~-2
n3-3n+(n2-1) Vn2-4+2
r
l
~
vn=a-vr=a +
1-a
v1-a2-1+a
Jx
x[fl=i-!J
(O<a<1).
11. Prove that for x ~ 1
Yx+2vx=-I +Y.r-2Vx
is equal to 2 if .x~2, and to 2 Yx=t if x

12. Compute
> 2.
V a+ b + e + 2 Yae +be + V a+ b + e -2 Y ae + be
(a-,O,
b~O, e~O).
13. Prove that the trinomial x 3 + px

X=
+ q vanishes at
~
~2 + ~; .
y -t+ v ~ + ~; + v - -v'
14. Express x in terms of a new variable so that

and VxTb become rational.
15. Rationalize the denominator of the fraction
V x+a
Va+ Vb+-VC+ Va'+ Vii'+ Vc'

if
a
16. Prove that ~ cannot be represented in the form

p+
where p and q are rational (q 0 and is not
a perfect square).
yq,
>
34
Problems

3
3
10 tan ( ; -a) cos ( ; -a)
1t
+cos (n +a) sin
2 (1- sin (3n - a)+ cos (:ln
x [1-
+cos (a.-2 )sm (n-a.)+
cos (2n-a)
t- a.))
(a.-
~) =0;
sif! ( 32n - a) +cos ( 5211
a} ]
+ sin 2a = O;
3 (1-sin (n+a) +cos (n+a)J!!+
+ [ 1- sin ( 32
1t
+a)
+cos (-3;-a)J2=4-2sin2a.
<
18. Let a= 2kn+a.0 , where O~a0

2n.
Prove that there exists the following equality
. a
(
)11,/
srn2= - 1 V
Let us assume then

~a.o
< n.
1-cusa
2
Lha t a = 2kn
+a
0,
where - n ~
Show that then

a_(
),.,/1+cosa
cos2- - 1 V
2
19. If a whole number a is divisible by n leaving no
remainder, we shall write this in the following way
a== 0 (mod n)
'' l1ich i~ read: a is comparable with zero by the modulus n.

What nmainders can a whole number leave when being
clividec\ by the whole number n?
I l is obvious, that being divided by n, any whole number
can leave the following remainders
0, 1, 2, 3, ... , n - 1.
If as a result of dividing a by n we obtain a remainder k,

then we shall write
a== lr (mod n),
3. Radicals. Inverse Trigonometric Functions. Logarithms
35
since in this case

a - k
0 (mod n).
Thus, when dividing a by 2 only two cases are possible:

either a is divisible exactly, or leaves a remainder equal to 1.
In the first case we write a= 0 (mod 2), in the second
a
1 (mod 2).
The di vision by 3 can also yield a remainder (0, 1, 2),
and, consequently, only three cases are possible: a= 0
(mod 3), a== 1 (mod 3), a== 2 (mod 3) and so on.
Consider tho following problem.
We have
==
At= 1.
A 2 =cos nn.
A 3 =-= 2 cos ( ; n:rt -
118 n ) .
A 4 = 2 cos { ~ nn-
! n).
n) + 2 cos ! nn.
A5 = 2 cos (
~ n:rt - ~
A 6 = 2 cos (
~ nn -
A 1 = 2 cos (
~mt- 1~
n) + 2 cos (; nn- 1~ n) +
! nn -
7
16
f nn + 1~ n) .
n) + 2 cos ( ! nn - u n) .
A9 = 2 cos { ~ nn -
~;
n)
5
18 :rt )
-i 2 cos (
A 8 = 2 cos (
1
1
+ 2 cos ( ~ nn - ~ n) +
+ 2 cos ( ~ nn + 2; :rt) .
A 10 = 2 cos ( 51 nn- 53 n ) +2 cos 53 nn.
36
Problems
2 n:rt - 15 :rt ) + 2 cos ( :If

4 n:rt - 22
f)
A 11 = 2 cos ( tt
:rt )
22
+ 2 cos ( 161 n:rt - : 2 :rt) + 2 cos (
1~ n:rt -
3
22 :rt) +
+ 2 cos ( !~ n:rt + : 2 :rt) .

A 12 = 2 cos (
! n:rt -
~~
A 13 = 2 cos ( 123 nn-
+ 2 cos ( 163 n:rt -
~!
~ n:rt + 7~ :rt) .
:rt) + 2 cos (
n) + 2 cos (
1~ nn- 1~
n) +
1~ n) + 2 cos ( 1~ n:rt + 1~ n) +
10 nn 2 cos, ( 13
12 nn -+
4 :rt ) .
+ 2 cos 13
-13
1I
13 :rt ) j- 2 cos ( 3 nn - 3 :rt ) +

A 14 = 2 cos ( T1 nn - 14
7
14
3
-t-2 cos (; nn-14 n )

A 11; = 2 cos ( 125 n:rt -
9~
+ 2 cos ( 185 nn -
A1 6 = 2 cos (
n)
+ 2 cos ( 1;
!~
:rt) + 2 cos ( :; n:rt
n:rt - 178 :rt)
+ :8 :rt) .
! nn + ;; n) + 2 cos ( ~ nn+ ;~ n) +
+ 2 cos (
~ nn +
5
32 :rt) + 2 cos (
~ n:rt +
3
32 :rt) .
A 17 = 2 cos ( 127 n:rt + :; :rt) + 2 cos ( 1~ n:rt - 187 :rt) -!-
+ 2 cos ( 1f; nn + 2 cos (
!~ n:rt -
5
1 7 n)
+ 2 cos 187 nn +
f :rt) + 2 cos {!~ n:rt 7
5
17 :rt) +
14 nn- 1 n ) +2cos ( 16 n:rt+17n

8
).
+2cos ( 17
17
17
.11, 8 = 2 cos
(.!.n:rt+~
n)-t-2 cos (~n:rt-_!_ :rt) ...L
9
27
9
27
'
+ 2 cos ( ~ n:rt + ; 7 n)
3. Radicals. llwerse Trigonometric Functions. Logarithms
37
Prove that
As= 0 if
A 7 = 0 if
A10 = 0 if
Au= 0 if
n
n
= 1,
=
n=
n =
2
3,
2
2,
3,
3,
(mod 5),
4 (mod 7),
(mod 5),
3, 5, 7 (mod 11),
5, 7, 9, 10 (mod 13),
4 (mod7),
1,
1,
1,
2,
A13 = 0 if n
A14 = 0 if n=1,
A1 6 = 0 if n = 0 (mod 2),
A11 = 0 if n:=1, 3, 4, 6, 7, 9, 13, 14 (mod17),
and that A 2 , A 3 , A 4 , Af\, A 8 , A 9 , A 12 , .A 15 and A 18 never

vanish for any whole n (S. Ramanujan. Asymptotic formulae
in combinatory analysis).
20. Let
p(n)=A(n--j-3) 2 +B+C(-1)"1+Dcos 2 ~n (nan integer).
Prove that thorn exists the following relationship
p (n) - p (n -
1) -
p (n -
2)
+ p (n -
+p
(n -
4)
5) - p (n -
6) = 0.
21. Show that
1 sin 15 =
V6- V2
2 sin 18=
-i+4 Vs
cos 15 _ V6--1 V2 ;
-4
cos 18=
! v 10+2 vs.
22. Show that

. 6 0 -_
srn
V 30-6 V5 -V 6+2 vs
cos 50 =
v 18+6 vs + v 10--2 vs
23. Show that

cos (arcsin x) =
V 1-x
2,
tan (arccot x) = -x ,
sin (arccos x) =
V 1-x
cot (arc tan x) = x- .
Problems
38
cos(arctanx)= , 1
V
cos (arccot x) =
-V
1
1+x2
x
1~~
sin (arc tan x) =
sin (arccot x) =
-V 1+x2
x
-V
1
1~~
24. Prove that
arcsm
x + arc cos x = 2:rt .
arc tan x +arc cot x = 211 ,

x+y
arctanx+arctany=arctan 1 -xy +en,

where e = 0
e = -1
if
if
xy
xy
+1
if
.xy
e=
< 1,
>1
>1
an cl
< 0,
x > 0.
:rt
and x
26. Show that 4 arctan 5-arctan 239 = T

1
27. Show that arctan 3 +arctan5+arctan 7+
+ arctan 81 = T11
28. Show that 2arctanx+arcsin 1 ,x
~x 2
(x> 1).
=:rt
29. Prove that

1
11
arctanx+arctan -;= 2
arctan x+ arctan
!= -
if x
> o;
~ if x < O.
30. Prove that

arcsin x+ arcsin y = 'l'J arcsin (x V1-y 2 +yV1-x2 ) +en,
where 'l'J= 1,
e==O
if xy 0 or x2 +y 2 ~1,
'l'J = -1, 8 = -1 if Z 1 + y2 1, X
0, l/ 0,
'l'J= -1, e= +1 if xl+y"> 1, x>O, u>O.
<
>
<
<
3. Radicals. Inverse 1'rigonumetric Functiott3.
Lugarith111~
39
31. Check the equality

arccosx+arccos (; +} V3-3x2 )
11
if
32. If
l
A= arctan 7 altfl B =arc tan 3 ,

then prove that cos 2A =sin 4B.
33. Let a 2 + b2 = 7ab.

Prove that
1
a+b
log - 3 - = 2 (log a+ log b).

log n
34. Prove that 1 11

= 1 + loga m.
Oifom n
x(y-f-z-x) _
log x
-
y(z-t-x-y)
logy
z(x+y-z)
log z
follows xY yx = zY yz = xz .z~.
36. 1 Prove that logb aloga b = 1.
2 Simplify the expression
a
log(log a)
log a
(logarithms are taken to one and the same base).
37. Given: y=101-loi:x, z =

taken to the base 10).
Prove that
.X=
38. Given.
10
10 1 -log Y
1-loaz.
(logarithms are
40
Problems
Prove that
logb+e a+ loge-I a= 2 loge+I a loge-/, a.
39. Let a>O, c>O, b=Jlac, a, c and ac=,t=1, N>O.

Prove that
log,, N
loge N
loga N -- log,, N
logb N - loge N
40. Prove that
41. Given a geometric and an arithmetic progression

with positive terms
a, a 1 , a 2 , , a,,,
The ratio of the geometric progression and the common
difference of the arithmetic progression are positive. Prove
that there always exists a Rystem of logarithms for which
log a 11 - b11 c= log a - b (for any n).
Find the base
of this system.
4. EQUATIONS AND SYSTEMS

OF EQUATIONS
OF THE FIRST DEGREE
The general form of a first-degree equation in one unknown is
Ax+ B = 0,
where A and B are independent of x. To solve the firstdegree equation means to reduce it to this form, since then
the expreS$ion for the root becomes explicit
B
x = - A.
4. Equations and Systems of Equations of the First Degree
41
Therefore the problem of solving the first-degrC'e equation

is one of transforming the given exprPssion to the form
Ax
B = 0. In doing so great attention should he paid
to make sure that all the eq1rntions involved are equivalent.
The prc,blem of solving a system of C'quations also consists
to a considerable extent in transforming a system into
an equivalent one.
This section deals not only with equations of the first
degree in the unknown x, but also with the equations which
can be reduced to them by means of appropriate transformations (such are equations involving radicals, trigonometric
equations and ones involving exponential and logarithmic
functions). Here and in the following section we consider
a trigonometric equalion solved if we find the value of one
of the trigonometric functions of an expression linear in x.
Indeed, if it is known that
tan (mx
then we find
mx
+n
+ n)
= A,
= arctan A
+ kn,
where k is any int<'ger.

Consequently, all the required values of x are given by
formula
arctan A- n -l kn
X-=------
Likewise, if it is found that

cot (mx+ n) =A,
then
mx+ n = arccot A+ kn and
arccot A - n +kn
m
X=------
But if it is known that
sin(mx+n)=A,
then all the values of x satisfying the last equation are
found by the formula
mx
+n
= (-1) 11 arcsin A
where k, as before, is any integer.
+ kn,
42
l'roblems
Analogously, from the equation

cos (mx
+ n)
follows
mx
+n
+arccos A
+ 2/cn.
When solving exponential equations one should remembe

that the equation
ax
(a> 0 and i:-; not equal lo 1)
= 0.
has the only solution x
1. Solve the equation

x-ab
x--ac
- +a+b
a-1 c
+-.r-bc
- - -__a -!- b--j-c
bTc

x-a
x-b
.r-c
IJC+ll"C+-ab= 2
( 1 ...!.. 1
a , "b+c-1 )
6x-+ 2a 3b 1- c
6x,-2a-3b-c
2.c Ga+b+ :k
2x+6a-b-3c

a+b-x
c
a+c-x
b
b+c-x
a
4x
a+b+c

vb+x
b
pb+x
cf/x
= -a- .
x
6. Solve the equations
1 Vx+1+Vx-1=1;
2Vx+1-Vx-1=1
7. Solve the .equation
Ya+Vx+Ya-Vx=~b.
1.
-V 1-V x
-x2 = x-1.
v'ii-+. v'X-b
v'a+v'x-a
.. ;-a
Vb"
v'a:+=X+ va::x =vb

v'a+x-v'a-x
ca> O).
11. Solve the system

x+y+z=a
x+y+v=b
x+z+v=c
y + z + v = d.
Xi
Xi
+ X2 +
+ X2 X2 +
Xi Xi -
X2 -
X3 -
X4
X3 -
X4
= 2ai
= 2a2
= 2aa
X4
= 2a4.
X3
X3
X4

ax + m (y + z + v) = k
by+ m (x + z + v) = l
cz + m (x + y + v) = p
dv + m (x + y + z) = q.
z1-a 1
x1 -a1
xp-ap
m1
m2
mp
- = = - - = ...
x1+x2+ ... +xp=a.
4:~
44
Problems

1
-+-+-=a
x
y
z
_!_ + _!_ + _!. -= b

v
-;+z-+-;=c
_!.y + _!_
+ _!.v = d.
z
ay + bx = c
cx+az=b
bz + cy =a.
cy + bz = 2dyz
az + ex = 2d'zx
bx+ ay = 2d"xy.
xy
~+h
=c,
xz
n+=
=b
'
_y_z_=a.
~+cy

xyz
y+z-x=li2
xyz
z+x-y=b2
xyz
x+y-z=C2.
(b + c) (y
+ z) - ax =
+ a) (x + z) - by =
(a + b) (x + y) - cz =
a+ b + c ::p 0.
(c
if
b-
c -
a- b
+ a) y + (a + b) z (a + b) z+ (b + c) x (b +c) x + (c + a) y -
+ c) x =
(c + a) y =
(a + b) z =
(c
(b
if
b+c=;i=O, a+c=i=O, a+b=;i=O
x
a+A.
a+
b-j-
+ b-\-A.
_x_+_Y_
_x_
a+v
+ c-\-A.
z
=
z __
i __
1
1
c-\- -
__z__= 1
__Y_
+ b+v + c+v

z -1-ay+a2 x+ a3 =0
z+by+b 2 x+b3 = 0
z+cy+c 2 x +c3=0.
+ +a
+ +
z ay a 2 x a3 t
z+ by-j-b 2 x+ b3t
+ =0
+ c x+ c t + c = 0
z + dy + d x + d t + d = 0.
z-j-cy
b4
25. SolvR the syst8m

x+y+z+u=rn
ax+ by cz + du = n
+ b y-f- c z+d u =
a x + b y + c z +d u =
a 2 .r
l.
x 1 +2x2 +3x3 +
... -I- nxn

x2 + 2x3 + 3x~ + . . . -;- nx
1
= a1
=
a:?.
2a3
2b3
2c3
45
46
Problems

X3-x, +3x2- X3-:c1 x2+ 7.r3X1-X2
-Xi-
X2-
-Xn=2a
-.Tn=4a
-:r11 = 8a
X3-
x 1 +x2 +x3+ ... +xn =1

.r 1 --j-x3+ ... +xn =2
x 1 + :r2 1- x~ + ... -J-Xn =3
x 1 + x2 +
... +
Xn-t
29. Show that for the equations

ax+ b = \.,, a'x + b'
= n.
0.
to be compatible it is necessary and sufficient that
ab' - a'b = 0.
30. Show that the systems

ax+ by+ c = 0
a' x + b' y + c' = 0
and
l(ax+by+c) + l' (a'x +b'y +c) =0
m(ax+by+c) + m' (a'x+ b'y+c')=O
are equivalent if
lm' - l' m =I=- 0.

31. Prove that the system
ax+ by +c =0
a'x+b'y+c'=O
has one an<l only one solution if
ab' - a'b =I=- 0.
47
32. Prove that from the equations

ax+by =0
a'x+b'y=O,
if ab'-a'b==i=O, follows
x=y=O.
33. Show tha Lthe following three equa Lions are compatible
ax+by +c ~~O,
a'.c+b'y+c' =0,
a"x t b"y + c" =--= 0
b'c) + b" (ca' - c'a) + c" (ab' - a'b) = 0.
34. Let a, b, c be distinct numbers. Prove that from
the equations:
x + ay + a2z = 0,
x + by + b2z = 0,
x + cy -t c2z = 0
follows
x = y = z = 0.
if a" (be' -
35. Prove that from the equations

Ax+By +Cz =0,
A 1x+B 1y+C:z=O
follows
x
C1B-CH1 =
CA 1-C1A
AB1-A1B
if not all of the denominators are equal to zero.

36. Prove that the elimination of x, y, z from the equations
+ cy + bz
ex + by + az
bx+ ay + cz
ax
yield~
0,
= 0,
= 0
=
48
Problems
37. Given the system
: -+< =f- ( 1 + ~)
:-:=~(1-~)
x
z
( 1-y )
-+-=
a
c
b
Prove that the equations are compatible and determine x,

y and z.
38. Determine whether the equations of the sys Lem
(a+b).c +(ap-~ bq)y=ap'!.+bq?.
(ap + bq)
J:'
+(a p2
+ hq~) y= ap + bq;i
3
are compatible.
+ X2 = a1
X2 +
=a2
.1'1
.1'3
X;i
-f- X4
= ll3
Xn-1 -/- .Cn =Gn-1
:x,.
+ '1 =lln
a2,
a-i-~ax
a-d
by
b-d
bZy
b-d
cz
c--d
c2z
+ c-d
=O
=d(a-b)(b-c)(c-a)-
4. Equations and Systems of Equation.~ of the First Degree
49
41. Solve tho system
+ a) (y + l) = (a (y + b) (z + m) = (b (x
(z
+ c) (x +
n)
n) (l - b)
l) (m - c)
(c - m) (n - a).
42. Determine k for the system
+ (1 + k) y
(1 - k) x + ky
+ k) x + (12 - k) y
x
(t
= 0
+k
-(t + k)
= 1
=
to be compatible.
+ y sin 2a + z sin 3a =
x sin b + y sin 2b + z sin 3b =
x sin c + y sin 2c + z sin 3c =
x sin a
sin 4a
sin 4b
sin 4c.

a
sin A
sin B
c
sin C '
follows
a = b cos C
b = c cos A
c = a cos B
+B +C=
+ c cos B,
+ a cos C,
+ b cos A.
45. Show that from the given data

a = b cos C
b = c cos A
c = a cos B
0 <A
<
n,

<
0,
+ c cos B,
+ a cos C,
+ b cos A,
n,
c > 0,
foJiows
a
-;;filA -= sin JJ
c
sin C
<
and
C < n,
>
0,
50
f'roblrms
46. Given
a = b cos C + c cos B
a2 = b2 + c2
b = c cos A + a cos C (1) b2 = a 2 + c2
c = a cos B + b cos A
c2 = a 2 + b2
2bc cos.A
2ac cos B (2)
2ab cos C.
Show that systems (1) and (2) are equivalent, i.e. from
equations (1) follow equations (2) and, conversely, from
equations (2) follow equations (1).
47. Given
cos a = cos b cos c + sin b sin c cos A,
cosb = cosacosc +sin a sine cosB,
(*)
cos c = cos a cos b + sin a sin b cos C,
where a, b, c and A, B, C are between 0 and :n:.
Prove that
sin A
sinB
sinC
sin a= sinb =sine-
48. Prove that from the conditions

problem follows
1 cos A = -cos B cos C + sin B sin
cos B = -cos A cos C + sin A sin
cos C = -cos A cos B + sin A sin
0
2 Lan 4 e =
-, /
p-a
p-b
4-
tan T tan - 2 - tan -
of tne preceding
C cos a,
C cos b,
B cos c;
+ B + C - :n: and 2p =a+ b

if e =A
(b -
c) tan (.r
+ a)
p-c
tan - 2 -
+ c.
+ (c - a) .tan (x + ~) +
+ (a - b) tan (x
+ y)
= 0.
50. Prove that sin x and cos x are rational if and only if
tan ~ is rational.
sin 4 x + cos 4 x = a.
Equations and Systems of Equations of the First DPgree
51
52. Solve the following equations

1 sin x + sin 2x + sin 3x = O;
2 cos nx
cos (n - 2) :t - cos x = 0.

1 m sin (a - x) = n sin (b - x);
2 sin (x + 3a:) = 3 sin (a: - x).
sin 5x = 16 sin 5 x.
sin x + 2 sin x cos (a - x) = sin a.
sin x sin ('Y - x) = a.
sin (a: + x) + sin a sin x tan (a:
+ x)
m cos a: cos x.

cos2 a:
+ cos x + cos
2
(ci
+ x)
= 1
+ 2 cos a: cos (a:+x)

(1 - tan x) (1 + sin 2x) = 1
60. Show that if
tan x + tan 2x
+ tan x.
+ tan 3x + tan 4x
then either 5x =kn, or 8 cos 2x = 1

61. Given the expression
ax 2
+ 2bxy + cy
Make the substitution

x =
y= x
cos e - y sine,
sin e + y cos e.
0,
V 17.
52
Problems
It is required to choose the angle 0 so that to ensure tho

identity
ax 2 + 2bxy + cy 2 = AX 2 + BY 2
x
y
tan(0+a.) = tan(0+~)
z
tan(U+y)
follows
x+y sin 2 (a-~)+
x-y
y-j-z
y-z
sin 2 (~-'V)+ z+x sin 2 (y-a)=-0.

z-x
63. Solve the syslems
1o sin x = sin y = 9in z

a
x-ty-t Z=n;
2 o tan x = tan y = tan z
a
x+y+z=:n.
tanx tany= a
x+y=2b.
x-~
4x_3
2=3
x+.!.
2_22x-1.
66. Find the positive solutions of the equation

XX+}= 1.
axbY=m
x+y=n(a>O, b>O).

(ax)log a=
blog
(by) log b
x = alog u.
5. Equations and Systems of Equattuns of the Second Degree
5. EQUATIONS AND SYSTEMS

OF EQUATIONS OF THE SECOND DEGREE
The present section contains mainly problems on solving
quadratic equations and using the properties of the seconddegree trinomial.
It should be remembered that if the roots of the trinomial
ax2
bx
c* are imaginary, then this trinomial retains
its sign at any real values of x. As is easily seen in this
case the sign of the trinomial coincides with that of the
constant term (i.e. with the sign of c). Thus, if c > 0 and
the roots of the trinomial ax 2
bx
c are imaginary, then
ax2 +bx+ c > 0
for any real x.
When solving systems of equations the following proposition should be taken into account. Let a system of m equations in m unknowns be under consideration, the degrees
of these equations being, respectively,
k1' k2, ... , km.

Then our system, generally speaking, allows for k 1k 2 k.n
solution sets. To be more precise, the product of the degrees
of the equations is the maximal number of solutions.
Sometimes this limit is reached (see Problem 23), but sometimes it is not. Nevertheless, this proposition is of importance, since it pr_events the loss of solutions.
X2
(b+x)(x+c)
(x-b) (x-c)
+b2
(b+c)(b+x)
(b-c) (b-x)
+cz
(c+x)(c-t-b)
(c-x) (c-b)
--(b+c) 2
~
In this section the letters a, b, c, p, q and other constants in

the equations denote real numbers.
Problems
54

(b-c) (x-b) (x-c) + b3 (c-a) (x-c) (x- a)+
+c3 (a-b) (x-a) (x--b)=O
and show that if the roots of this equation are equal, then
exists one of the following equalities
a3
-Va+ -v11 + -v-c

(a-x) Vli=X-(b-x) V x-f1
a-x+ x-b
=a- b
.
V 4a+ b-5x+-V4b+a---5x-3 Va+b-2x= ().

5. Prove that the roots ot the equation
(x - a) (x - c) + 'A (x - b) (x - d) = 0
are real for any "A if a < b < c < d.
6. Show that the roots of the equation
~-aj~-~+~-aj~-~+~-~~-~=0
are always real.

7. Prove that at least one of the equations
x 2 + px + q = 0,
x 2 + P1X + qi = 0
has real roots if P1P = 2 (q1 + q).
8. Prove that the roots of the equation
a (x - b) (x - c) + b (x - a) (x - c) +
+ c (x - a) (x - b) = 0
are always real.
9. Find the values of p and q for which the roots of the
equation
x2 + px + q = 0
are equal to p and q.
10. Prove that for any real x, y and z there exists the
following inequality
x2
+y +z
2
2 -
xy -
xz -
yz ~ 0.
5. Equations and Systems of Equations of the Second Degree
55
11. Let
+ y + z =a.
Show that then
x2 + y2
12. Prove the inequality
x
+ z2 ~ 3
a2
+ y + z ~ V 3 (x + y + z
2
13. Let ex and
2 ).
be the roots of the quadratic equation
+ px + q = 0.
Put exk + ~k = sk.
Express sk in terms of p and q at k = 1, +2, +3, +4,
x2
5.
14. Let ex and
x2
be the roots of the quadratic equation
+ px + q =
(ex
>
0,
>
0).
Vcx+Vfl"
Express
in terms of the coefficients of the equation.
15. Show that if the two equations
Ax2 +Bx+ C = 0, A'x2 -+ B'x + C' = 0
have a common root, then
(AC' - CA') 2 =(AB' - BA') (EC' - CB').
+ y + z)
y (x + y + z)
z (x + y + z)
x (x
= a2
= b2
= c2
t 7. Solve the system
+ y + z)
y (x + y + z)
z (x + y + z)
x (x

y + 2x + z = a (y
z + 2y + x = b (z
x + 2z + y = c (y
= a - yz
= b - xz
= c - xy.
+ x) (z + x)
+ y) (x + y)
+ z} (x + z).
56
Problems

y + z + yz _.a
x+z+xz-=b
x + y + a:y
c.
==
20. Solve the syst~m

yz =ax
z.x = by (a
xy =CZ
.i:2
>
>
0, c > 0).
+ r/i =
x2
0, b
+
+
c.ry2
= b.i11z
2
z = axyz.
z2

x (y
!I (.r
+ z)
+ z)
z (x + y)
= a2
=- b2
c2

X1 =ax+ by
y 3 =bx+ ay.
x2 = a + (y - z.)2
y 2 = b + (x - z) 2
z2 = c + (x - y) 2
b(x+y)
x+y+cxy
c (y
+ z)
y+z+ayz
c(z-l x)
+ x+z-J bxz - a
a (x + y) _ b
+ x+y+cxy -
a(x+z)
...I-
b(y+z)
x+z+bxJ
v++avs
=C.
5. Equations and Systems of l!.'ouatirm1 of the Second Degree
57

-
yz =a
y2 -
xz = b
xy = c.
x1
z
y2 + z2 - (y + z) x = a
x 2 + z2 - (x + z) y = b
x2 + y2 - (x + y) z = c.
x2 + y2 + xy = cs
z2 + x2 + xz = b2
y2 + z2 + yz = a2.
2 -
x3+ys+z3=as
+ y2 + z2 = a2
x + y + z =a.
x2
x4 +
u' = a 4
xs + ya + zs + us = as
x2 + y2 + z2 + u2 = a"'
x + y + z + u =a.
31. Prove that systeII).S of equalities (1) and (2) are equivalent, i.e. from existence of (1) follows the existence of (2)
and conversely.
a2 +b2 +c2 =1, aa' +bb' +cc' =0,
(1)
a' 2 + b' 2 + c' 2 = 1, a' a" + b' b" + c' c" = 0,
a""'+ b"2 + c"2 = 1, aa" + bb" +cc" =0;
y4
+ z4 +
a2 +a'"+a2 = 1,
b1 + b' 2 + b"" = 1,
c"' +c'"'+C-1 =1,
ab+a'b' + a"b" = 0,
bc-j-b'c' +b"c" =0,
ca+c'a' +c"a" ... o.
(2)
58
Problems
32. Eliminate x, y a11d z from the equalities

x2
(y+z)=a 3 ,
z 2 (x+y)=c 3 ,
y 2 (x+z)=b3,
xyz=abc.
33. Given
Eliminate x, y and z.
34. Eliminate x, y, z from the system
+z
z +x
x2 + y
= 0
= 0
y2
2 -
2ayz
2 -
2bxz
2 -
2cxy = 0.
35. Show that the elimination of x, y and z from the system
+ yz + z2 = a2
+ xz + x =--= b
x2 + xy + y2 = c2
xy + yz + xz = 0
y2
z2
yields
(a
+ b + c)
(b
+c-
a) (a
+c-
b) (a
+b-
36. Eliminate x and y from the equations
+y
= a,
x2
+y
= b,
x3
+y
37. Eliminate a, b, c from the system

x
7=7J=-c
a2
+b +c
2
= 1
a+b+c=1.
38. Given
( ; + ; ) ( ; + : ) ( : + =) = y.
= c.
c)
0.
5. Equations and Systems of Equations of the Second Degree
59
Eliminate x, y and z.
39. Prove that if
x+y+z+w=O
+ by + cz + dw = 0
(xw + yz) + (b - d) (c - a) (yw + zx) ++ (c - d) (a - b) (zw + xy) = 0,
aJ;
d) 2
(a -
(b -
p) 2
then
y
x
(d-b) (d-c) (b-c) -
(d-c) (d-a) (c-a) -
= (d-a) (d-b) (a-b)
(b-c) (c-a) (a-b)
40. 1 Let
and
:;
cos a+ cos ~-cos (a.+~)= 2

Prove that
2 Let
and
1
cos a cos~ cos (a+~)= -3

Prove that
41. Let
cos
Compute
e + cos cp
cos (8
+ cp)
42. Given that a and

equation
a cos x
a,
sin
and
~
e + sin cp
sin (0
b.
+ cp).
are different solutions of the
+ b sin x
= c.
60
Problem.~
Prove that
2 a-~
c2
cos - 2 - = a2+b2.
43. Let
sin (0--a)
sin (0-~)
(0-a)
CO!!
cos (8-~) = (j
=ob'
Prove that
ac+ bd
COS
(a:-~)= ad+bc
44. Given
e2-1
1+2ecosa+e2
Prove that
e2 -1
_ e +ens ~
1+2ecosa+ez - e+cosa
2 0 tan a . tan ~ = 1+e
i-e.
2
2
1o
sit1 ~ _ _
sin a -
1 + e CO!'~
1 e co~ a: '
45. Prove that if

cos x-cos a
sin2 a eos ~
cos x-cos ~ = sin2 ~cos a '
then one of the values of tan ; is tan ~ . tan ~ .

46. Let
cos a= cos ~cos cp = cos'\' cos e, sin a= 2 sin ~ sin ~ .

Prove that
tan 2
= tan1
tan 2
47. Show that if

(x - a) cos 0
+ y sin 0 =
(x - a) cos 01
and
tan 20 -tan 2lft
= 2l ,
then
y1 = 2az-(1-l 2) x2
+ ysin 0 =a
1
.$. Equations antl Sylltrms of Equations of the Second Degree
(It

x cos 0 + y sin 0
= x cos cp + y sin cp = 2a
and
2 sin
~ =1
sin
follows
49. Let
cos 0 -= cos a cos p.
Prove that
0-a
2 ~
tan -0+a
2-.tan-2-=tan 2 .
50. Show that if

cos x
cos (x+O)
-a-=
b
-
cos (x
+ 20)
cos (x
+ 30)
then
a+c
b+d
-b-=-c-
51. Let
20 cosa
cos = cos~ '
Prove that
tan 2
cos y
tan a
tan 0
cos cp = cos ~ ' tan cp = tan y
tan 2
~ = tan 2 ~
52. Prove that if
qi
cos 0 =cos a cos p, cos cp =cos a 1 cos p, tan 2 tan 2 =tan 2 ,

then
sin
1
1
2A=(cos a -t) (cos--t)
a
I'
53. Let
.r cos (a+ p) +cos (a-p) = xcos (p +y)-\- cos (p-y) =
= x cos (y-t-a) +cos (v-a).
62
Problems
Prove that
tan a
1
tanB
tan 2 (B+ y)
tan 2 (a+ y)
tany
1----'---.
tan 2 (r.l +Bl
51. Prove that if

siu(H-r.)co~a
+ coa!a+0)sinB =O
cos(c:p-~)Sina
sin(rp-a)cosri
and
tan fl tan a
Ian If> Ian~
+ cos (a-Bl = O,
cos
(a+ ri)
then
1
tan e = 2 (tan~ +cot a),
tan <p = 2 (tan a-cot~).
55. Given
n 2 siri. 2 (a+~)= sin 2 a+ sin 2 ~ -2 si11 a sin~ cos (a-~).
Prove that
1+n
tan a-= 1 =-tan~

+n
!'l6. Eliminate e from the equations
cos (a - 38) = m cos~ e, sin (a - ~O) = m sina fl.
57. Eliminate 0 from the' equations
(a- b) sin (8 + cp) =(a+ b) sin (fl- er),
a tan
- b tan
c.
58. Show that the result of elimination of 0 ancl cp from

the Pq11aLions
sin~
cosO=-.-,
ilin a
sin y
cos<p=sina,
cos(fl-<p)=sinp~in1
is
tan 2 a = tan 2 ~ + tan 2 y.

59. Eliminate e and cp from the equations
a sin 2 0 + b cos 2 e = a cos 2 <p
b sin 2 cp = 1,
a ta 11 8
60. Prove that if
cos (8 - a) = a, sin (8 - ~) -= b,
b ta II <p.
5. Equations and Sy.~tems of Equations of the Second Degree
6~
then
a2
2ab sin (ex -
~)
+ b2 = cos 2 (ex -
~).

cos 3x cos3 x + sin 3x sin3 x = 0.
sin 2x + cos 2x + sin x + cos x + 1 = 0.

t 2 _ 1-cos x
an x - 1- sin x
6~.
Solve the equation

32 .cos6 x - cos 6x = 1.
65. Solve and analyze the equation

sin 3x + sin 2x = m sin x.
66. 'Solve the equation
=i+k
(1 i-. k) cosxcos(2x-a)
. cos (x-a)
cos 2 x.
sin 4 x+cos 4 x-2sin2x+
! sin 2x=0.
2

2 logx a + logax a + 3 loga2x a = 0.
69. Find the positive solutions of th.e system
xx+y=ya, yx+y=X4a (a>O).
70. Find the positive values of the unknowns x, y, u and
v satisfying the system
Problem.~
6. COMPI"'EX NUMBERS AND

POLYNOMIALS
We proceed here from the assumption that the principal
operations with complex numbers (i.e. addition, multiplication, division and evolution) are already known to the
reader. Likewise, we take as known the trigonometric form
of a complex number and de Moivre's formnla. 111 factoring
polynomials and solving certain higher-degree equations an
important role is played by the so-called remainder theorem
(stated by the French mathematician Bezout), usually
considered in textbooks of elementary algebra. Let us
recall it: if f (x) is a polynomial in x and if f (a) = 0, then
f (x) is exactly divisible by x - a. Hence (assuming that
the polynomial has one root) follows the possibility of
resolving an nth-degree polynomial into n, equal or unequal,
linear factors as well as the following propositio.n used here
repeatedly: if it is known that a certain nth-degree polynomial in x vanishes at n + 1 different values of x, then such
a polynomial identically equals zero. Consequently, if two
polynomials of the nth degree nix attain equal values at n + 1
different values of x, then such polynomials are identically
equal to each other, that is, the coefficients of equal powers
of x coincide. Finally, let us mention the relationship between the roots of an nth-degree equation and its coefficients. Let the polynomial
Xn
P1Xn-t -' P2Xn- 2
+ +
Pn-1X
Pn
have the roots x 1 , x 2 , , x 11 , so that there exists the factorization
1
xn+ p 1x"- + P2x'H + ... + Pn = (x-x 1} (x-x2) ... (x-xn)
We then have the relations:
+ + ... +
+ ... +
+
+ +
X1
X 1X2
X1X3
Xz
X1Xn
X1X2Xa
Xn =
X2X3
-p1.
+ ... +
Xn-2Xn-1Xn
Xn -1Xn
-p3,
= P2
65
6. Complex Numbers and Polynomials
1. Let x and y be two complex numbers.

Prove that
I x + y 12 + I x - y 12 = 2 {I x 12 + I y 12 }.
The symbol I ex I denotes the modulus of the complex number ex.
2. Find all the complex numbers satisfying the following
condition
1
= x2 ;
2 = x3.
The symbol denotes the number conjugate of x.
3. Prove that
x
x
V(a1 + az+ ...
+an) +(b1 +b2+ ... +bn) 2 :;;;;;Vai+b~+

2
+ -V ai+bi+ ... +Va~ +b~,

where ai and bi are any real numbers (i = 1, 2, 3, ... , n).
4. Show that
+ b + c) (a +
(a
be
+ ce
2)
(a
be2
+ ce)
if
e2
+e+1=
a3
+b +c
3
3 -
3abc
0.
5. Prove that
(a 2
+b +c
2
2 -
(x 2
ab - ac -
y2
z2 -
+ Y + Z2 ax + cy + bz,
ex + by + az,
bx + ay + cz.
= X
if
X =
Y =
Z
be) x
xy - xz -
6. Given
x+y
+z =A,
x+ ye +ze 2 =B,
x
+ yF 2 + ze
= C.
yz) =
XY -
XZ - YZ
66
Problems
Here and in the next problem e is determined by the equa

lity
8 2 + 8 + 1 = 0.
1 Express x, y, z in terms of A, B, and C.
2 Prove that
I A 12 + I B 12
+ I C 12 =
3 {I .r 12 + I y 12 + I z 12}.
7. Let
+ +
A' =x' +y' +z',

AA' =x" y" z'',
A =x+y+z,
2
2
B =X +ye +ze , B' = x' + y'e + z'e , BB'=x" + y"e+z"i:: 2,
C = x-t-ye 2 ze, C' = x' +y'e 2+z'e, CC'= x" + y"e 2 +z"e,_
Express x", y" and z" in terms of x, y, z and x', y', z'.
dt) 2 + (bx + ay -
(ax -
by -
cz -
+ (ex
+ dy
+ az -
dz + ct) 2 +
bt) 2 + (dx - cy
bz + at) 2 =
= (a2 + b2 + c2 + d2) (x2 + y2
z2
+ t2).
9. Prove the following equalities

- 1 - (
10 cos
cos"nr.p
qi -
n2 ) t an 2 cp
+ ( 4n ) t an
cp -
+A '
where
n
A= ( -1) 2 tan" cp
A= (-1)n;i (
2 ~0~ 1 ~:
= (
if n is even,
n ) tan 11 - 1 cp
n-1
if n is odd;
~ ) tan cp - ( ; ) tan3 cp + ( ~ )tan5 cp + ... +A,
where
A=(-1)
11
;
(n...:. 1 ) tan"- 1 cp
if n is even,
n-1
A= ( - 1)_2_ tan" cp
if n is odd.
67
Hero and in the following probll'ms

( n ) = c~ =
k
n (n - 1) . : . (fl - k
123 .... k
+ 1)
10. Prove the following equalities

1 22 mcos 2mx=
k=m-l
2 )
2 )
:
cos2(m-k)x+ (:: ;
k=O
k=m-1
2 22 m sin 2 mx =
2m
-1 )m+h 2 ( k ) cos 2 (m - k) x
k=O
3 2 2mcos 2m+i x =
1) cos (2m -2k +1) x;

h (2m+
k
k=m
h=O
4 2 2msin 2m+ix=
1) sin(2m-2k+1)x.
h (-1)m+h (2m+
k
k=m
k=O
11. Let
Un= cos a+r cos (a+ 0) + r 2 cos (a+ 28) +
.... +
+ rn cos (a+ n8),
Vn =sin a+ r sin (a+ 8) + r sin (a+ 28) + ... +
+ ,.n sin (a+ n8).
2
Show that
_ cos a-r cos (rx-O)-r"+1 cos [(n + 1) O+rxl + rn+ 2 cos (nO +a)
1 - 2r cos e ,.2
u n --
v _sin a
n -
'
rsin(rx-8)-rn+l sin [(n+1)0+rxl+r"+ 2 sin(n0+rx)

1 - 2r cos H+ ,.2
12. Simplify the following sums

1 S = 1 + n cos e
+ n(n-1)
12
cos 28
+ ... =
k=n
h c~ cos k8,
k=O
(C~ = t);
68
Problems
k=n
'"
'J
=II
Slll 0+
n(n-1),.
Slll
.
1 2
20 +
...
-- "1
LJ Ck.
"Sill !:O
t
h~-o

sin 2 P a+ sin 2 P 2a
+ si11
2 fl
3a
+ ... + sin
1
-- - 2
2P
I
na =
1.3.5 ..... (2p-1)

2.4.(j .... 2p
-~--:-----,,----'-
<
if a= 2: and p
2n (pa positivP integPr).
14. Prove that
1 The polynomial x (x"- 1 - nan- 1 )
an (n - 1) is divisible by (x - a) 2
2 The polynomial (1 - xn) (1
x) - 2nxn (1 - x) - n 2xn (1 - x) 2 is di visible by (1 - x) 3
15. Prove that
1 (x
y)n - xn - yn is divisible by xy (x
y) X
x (x 2 xy y2) if n is an odd number not divisible by 3.
2 (x
yt - xn - yn is divisible by xy (x
y) X
x (x2 xy y 2) 2 if n, when divided by 6, yields unity
as a remainder, i.e. if n = 1 (mod 6).
16. Show that the following identities are true
+
+ +
+
+ +
+ y)
~x + y)
(x + y)
1 (x
3 -
5 -
7 -
+
+
x3 - y 3 = 3xy (x
x 5 - y 5 = 5xy (x
x 7 - y 7 = 7xy (x
+ y);
+ y) (x + xy + y
+ y) (x + xy + y
17. Show that the expression

(x + y + z)m - xm - ym - zm
is divisible by
(x
+ y + z)3
2 );
2) 2
(m ocld)
- x3 - y3 - z3.
18. Find the condition necessary and sufficient for x:i

3
3
to be divisible by .T
y
z.
19. Deduce the coruliton at which xn - an is divisible
by xP - aP (n and p positive integers).
20. Find out whether the polyuomial x 411
x 4 b+i
x 4 c+ 2
x 4"+' 1 (a, h, c, d positive iulegNs) is divisible by
+ y + z + kxyz
+ +
x3
+ x + x + 1.
2
69
21. Find out at what n the polynomial 1 + x 2

+ ... + x2n- 2 is divisible by the polynomial 1 + x
+ ... + xn-1.
+x +
+x +
4
22. Prove that

The polynomial (cos cp + x sin cp )n - cos ncp - x sin ncp is divisible by x 2 + 1.
2 The polynomial xn sin cp - pn-l x sin ncp +
+ pn sin (n - 1) cp is divisible by x2 - 2px cos cp + p2
23. Find out at what values of p and q the binomial
x4 + 1 is divisible by x 2 + px + q.
24. Single out the real and imaginary parts in the expression Va + bi, i.e. represent this expression in the form
x + yi, where x and y are real.
25. Find all the roots of the equation
xr'
= 1.
26. Find the sum of the pth powers of the roots of the
equation
xn = 1 (p a positive integer).
27. Let
. 2:rt ( na pOSlt.lVe mteger
.
)
e =COS 2:rt + i. SID
and let
Ak = x
+ yek +
Ze2k
+ ... +
Wem-11"
(k = 0, 1, 2, ... , n - 1),
where x, y, z, ... , u, w are n arbitrary complex numbers.

Prove that
k=n-1
k=O
Ak 12 = n { I x
12
Y 12
z 12
+ +
12 }
(see Problem 6).

k=n-1
x n-1=(x -1)
2
~ (x 2 -2xcos
k=I
k=n
2 x2 "+1-1=(x-1)
II (x
k=I
-2xcos
k: +1);
2 !~ 1 +1);
Problems
70
k=n
3 x2n+ -1=(x+1)
1
II (x2 +2xcos 2 !~ 1
+1);
k=I
k=n-1
4x2n+1 =
II
x 2 -2xcos
(2k -l- 1)
Zn
+ 1) .
k=O

.
n.
2n:
.
10Slll
2n Sln 2fl ... Slil
(n-l)n
2n
-Vii.
= 2n-t ,
1L
2n:
cos 2n+1
COS
4n:
2n+1 ...
,
COS
2nn:
2n+1
(-1)2
_2_"_
if n is even.
30. Lol tho equation x = 1 have the roots 1, a, ~. y, ... , I..
Show that
(1 - a) (1 - ~) (1 - y) ... (1 - /..) = n.
11
31. Let
X1, Xz, , Xn
be the roots of the equation

x" + xn-t + ... + x + 1 = 0.
Compute the expression
1
Xj
-1
1
X2 -
+ . +
1
1
Xn -
32. Without solving the equations

x2
y2
2
+ 2-c2
z
!l2"-r
2-b2
x2
y2
v2-b2 +
_l
V2 +
z2
v2-c2
--1 '
= 1,
~-I _y_2_+
z2
_
p2 T p2-b2
p2-c2 -
find
'
71
6. Complex Numbers and Poiynomials
33. Prove that if cos a:

equation
xn + P1Xn-1
+ i sin a:
is the solution of the
+ ... + Pn
= 0,
then p 1 sin a+p 2 sin 2a+ .. +Pn sin na =0 (p1, p 2, ... ,

Pn are real).
34. If a, b, c, ... , k are the roots of the equation
xn
+ P1Xn-l + P2Xn- + ... + Pn -1X + Pn

2
(p1, p 2, ... , Pn are real), then prove that

(1 + a 2) (1 + b2) (1 + k2) =
(1 -
P2
+ p4 -
.) 2
(P1 - Pa
+ p5
) 2
35. Show that if the equations
x3
x3
+ px + q
+ p' x + q'
have a common root, then

(pq' - qp') (p - p')2
36.
= 0
= 0
= (q -
q')3.
Prove the following identities
10~+~+~=
=
2
v{
(5-3}'7);
~+~+~=V~(3.Y"-6).
37. Let a+b+c=O.

Put
a"'
+ b11 + ch = sh.
Prove the following relations

of Sec. 1)
2s4 =
6s7 = 7s3s4 ,
25s1s3 = 21s;,
s;,
(see Problems 23, 24, 26
6s5 = 5s 2ss,
10s7 = 7s2s5,
50s~ = 49s4s!,
72
Problems
38. 1 Given
x
x2
+y =
+ y2 =
+ v,
+ v2.
u
u2
Prove that
for any n.
2 Given
+y+z=
+ y2 + z2 =
+ ya + za =
x2
xa
u + v + t,
u2 + v2 + t2,
+ v3 +
ua
ta.
Prove that
for any n.
39. Let
X1
+ x 2 e + x 3 e2 ,
where
x1 + x2 e2 + x 3 e,
+e+1 =
0,
x 3 are the roots of the cubic equation
x 3 + px + q = 0.
e2
and x 1 , x 2 ,
Prove that A 3 and B 3 are the roots of the quadratic equation

z2
+ 27qz -
27p 3
0.

if
(x
+ a)
(x
+ b)
(x
a+ b
(x
+ c)
c
(x
+ d)
+ d.
+ a) 4 + (x + b) 4
c.

~+b+tj~+a+tj~+a+~~+b+tj-
- abcx
0.
73
+ 3ax + 3 (a
x3
+a +b +
be) x
2 -
3abc = 0.
c3 -
+ bx + cx + dx + e =
ax 4
if
a
+b
+c+d
= b
+ e.
= d

(a
+ b + x)
4 (a3
3 -
+ b + x3)
3
- 12abx = 0.

a2x2
x2
+ (a+x) 2 -"--"m
(a and m>O).
Deduce the condition under which all the roots are real,
and determine the number of positive and negative roots.
(5x4+ 10x2+ 1) (5a4+10a2 + 1)
(x4+ 10x2+ 1) (a4+ 10a2+5) =ax.
49. 1 Solve the equation

x3
if
xi
+ px + qx + r
2
= 0
X2:i.3.
2 Solve the equation

x3
+ px + qx + r
2
= 0 if x 1 = x 2
50. 1 Solve the system

y3
z3
x3
+ +
+ x3 + b
+ +
z3
a3
3
y3
= 3ayz
= 3bzx
c3 =
3cxy.
X3.
74
Problems
2 Solve the system

x 4 - a = y4 - b
z4
c = u4
d = xyzu
if a + b + c + d = 0.
51. In the expansion 1 + (1 + x) + ... + (1 + x)n in
powers of x find the term containing xii..
52. Prove that the coefficient of x' in the expansion in
powers of x of the expression {(s - 2) x 2 + nx - s} (x+1) 11
is equal to
s-2
nCn .
53. Prove that for x > 1 pX'l ~ qxP - p + q > 0 (p, q
positive integers and q > p).
54. Let x and a be positive numbers. Determine the
greatest term in the expansion of (x + a)n.
55. Prove that
1 im-i(i-1r+ i(~~i) (i-2)m+ ... +(-1)i- 1 i1m=o

if i
> m.
2 mm-m(m-1r+ m(~.;- 11 (m-2r+ ... +

+(-1)"'-1 m= ml
(i and m positive integers).
(x 2 + a2 )" = {xl -C~x"'-2a 2 + c;x n-4 a4 -
+ {C!x
11- 1 a -
}2 +
C~xn- 3 aa +
... }
57. Determine the coefficient of x 1 (l=O, 1, ... , 2n) in

the following products
1 {1+x+x2 + ... +xn} {1+x+x2 + ... +x"};

2 {1 +x+x2 + ... +xn} {1-x+x 2 -x3 + ... +
+ (-1)n xn};
3 {1+2x-!-3x 2 + ... +(n-!-1)x"}{1+2x+3x2 + ... +
+(n+1)xn};
2
4 {1+2x+3x + ... +(n+1)x"}{1-2x+3x2 - +
+(-1)"(n+1)x 11 }.
75
58. Prove that

1 1+c;+c~+
... =CA+c!+ ... =2n- 1 ;

C~n + dn + ... + c~; 1 = 22 "- 2 if n is even;
1 +c~n+ ... +c~; 1 =2 2 2 if n is odd.
3
1
20
con I c3n + c6n +- .. = 31 ( 2" + 2 COS 3}

n;n \ ,
C1n+ c4,.+en+
1
__ 1 (
,
(n-2):n )
... -3
2 +2cos
3
'
I
3oc2+c5+cs+
n
n
n
__ _!_(?n
1 2 COS
3
~
... -
1_
(n-4):n) .
60. Prove thaL

ft
con -t- c4n + csn + ... = 21 (21-1 + 22 C08'n;n

T
),
2 ct+c5+c9+
n n n . . =
0
21
(2"-1+22 Slll
. 4n;n '),
n
30
c2n +. c6n + c10

n + -
c3n+ c1n+cu
. 4nrc ) .
n + ... = 21 ( 2 ,_1 - 22 Sln
i_
2
(2n-1 _22 COS ~)

4 .,
n
i2+2 + .... + n =C;,+1+2 (C;; +c;_1 + ... + C~).

2
62. If at. a2 , a3 and a4 are four succes8ive coefficients in

the expansion of (1 +x)" in powers of x, then
a1
a1
+a2
aa
_
2a 2
a 3 +a~ - a~+ aa

1
1 (n-1)!
+ 3! (n-3)! + 5! (n-5)! + + (n-1)! 1!

(n even).
76
Problems
64. Find the magnitude of the sum
s=C:.-3c~+3 2 C~-3 3 C~+
....
65. Find the magnitudes of the following sums

2
4
1-Cn+Cn-Cn+
.,
CJ ' = cn1 - C3
n + C' n5 -- c'n +
fj
CJ=

1 c?,+2c;.+3c;,+4C~,+
c;.-
2''
2c;, -1- 3C:,
67. Prove that
1
+(n--t-1)C~=(n--!--2)2n- 1 ;
+ ... + (-1)"1
...
2Cn-3Cn+4Cn-J +
nC~ = 0.
( -1 )'<-1
n+i
Cn=
n+l
68. Prove that

1 1
1 2
1 1--f-2Cn+3Cn
+.
0
22c 1
20 2 c~+T
--j--
2ac 2
1
n+i
Cn=
24C 3
+T+T+ ... +
2n+1- f
n+1
2n+1c~
n+1
c;.-~c;+~c~+
... +(--!l"-1 c~=1+~+;+ ... +!.
70. Prove that

,..,..
+ ('n'n+2 +
10 C"n rI vn+t
2 C~-c;.+c;+
+enn+h =
en+! .
n+1<+1,
... +(-1)hC~=(-1)hC~-t
71. Show that t~e following equalities exist

0
, + cPc
1 cn0 cPm + cn1 cp-t
m
I
n m = cp
m+n;
ctcr+t + + vn
,..,n-rcnn = ( _ )I2n!( + )I
2 cocr
n n+ n n
n r. n r.
0

1 (C~) 2 + (C~) 2 + (C~) 2 +
... + (C~) 2 =
2 (Cgn) 2 -(Cin) 2 +(C~n) 2 -
C2n;
+(C~~) 2 =(-1)"C2n;
77
3 (C~n+1) 2 --(C~n+t) 2 +(C~n+1) 2 -
-(C~~ti) 2 =0;
2
2_
+- (C!!)
40 (Crn )2 + 2 (C2)
n + - n
n -
(2n--1)!
(n-1)! (n-1)!'
73. Let f (x) be a polynomial leaving the remainder A

when divided by x - a and the remainder B when divided
by x - b (a =I= b). Find the remaillder left by this polynomial when divided by (x-- a) (x - b).
74. Let f (x) be a polynomial leaving the remainder A
when divided by x - a, the remainder B when divided by
x - b and the remainder C when divided by x - c. Find
the remainder left by this polynomial when divided by
(x - a) (x - b) (x - c) if a, b and c are not equal to one
another.
75. Find the polynomial in x of degree (m - 1) which
at m different values of x, x 17 x 2 , , Xm, attains respectively the values Y1, Y2 ... , Ym
76. Let f (x) be a polynomial leaving the remainder A 1
when divided by x - ai. the remainder A 2 when divided
by x - a 2 , , and, finally, the remainder Am when divided by x - am. Find the remainder left by the polynomial,
when divided by (x - a 1) (x - a 2) (x - am)
77. Prove that if x 1 , x 2 , , Xm are m different arbitrary
quantities, f (x) is a polynomial of degree less than m,
then there exists the identity
78. Prove that if f (x) is a polynomial whose degree is

less than, or equal to, m - 2 and x 1 , x 2 , , Xm are m
arbitrary unequal quantities, then there exists the identity
f (x1)
(x1-x2)(x1-xa) ... (x1-Xm)
f (:r2)
(x2-X1) (x2-x3) ... (x2-Xm)
+ ... + (Xm-X1) (Xm-X2)

f (xm)
.
(Xm-Xm-1)
+
'
0.
78
Problems
79. Put
xn
+ ... + (Xm-X1 )( Xm-X2) ... (Xm-Xm-1)

(xi. x 2 , , Xm are m arbitrary unequal quantities). Show
that Sn=O if O:::;;;n<m-1, and Sm-1=1, and compute
Sn if n~m.
80. Compute the following
x-n
Sn=
-
x-n
(x1 - x2) (x1 - x3) ... (x1 -xm)

(x2- x 1) (x2-x3) .. (x2 Xm)
x-n
m
(
(n=1,2,3, ... ).
Xm-X1) Xm-X2) ... Xm-Xm-1)
+ ... +(
81. ShGw that if f (x) is a polynomial whose degree i5

less than m, then the fraction
f (x)
(x-x1) (x-x2) ... (x-xm)
(x1 , x 2 , , Xm are arbitrary quantities not equal to each

other) can be represented as a sum of m partial fractions
At
x-x1
A2
x-x2
+ ... + x-xm
Am
'
where Att A 2 , , Am are independent of x.

82. Solve the system of equations
Xt
a1-b1
Xf
a2-b1
.r,
a,,-b 1
Xn
+ a1-bn
='i
+ a2 Xn- bn =1
X2
+ a1-b2 +
X2
+ a2 -b2 +
.Tn
.T2
+ a,,-b 2 +
+ Un-bn
= 1.
83. Prove that the following identity is true

n!
--,.(.r-+-,----;1.,.-)..,-(x--f-,......,,..,2)-.-.- ..,...(.r_-f-_11)-
+2 +
=__E1_- 2q,
x
+1
+ x-f-3
3C~
.r
...
+ (-
ft+l nC~
x-f-n
79
In particular,
_1_= Ch -~c2 +~ca
n+1
2
3 n
4 n
_i.c4
+
5 n

( - i t a1a2 ... an+ (a1-b1)(a2-b1) ... (an-b1)
b1b2 ... hn
b1 (b1 - b2) ... (b1 -bn)
+ (a 1 ~b:) (a~-b2) .(b. (a~)b2) + ... +

2 ( 2- 1) 2- n
+ (a1-bn) ... (an-bn) =(-it.
bn (bn -b1) ... (bn -bn-1)

(x+~) ... (x+n~)-i
(x-~) ... \x-n~)
86. Given a series of numbers c 0 , Cti c2 , , ck, ckt 1 ,

Put 11ck = CJ:+i - ck, so that using the given series we can
form a new one
We then put
!1 2ck
11ck+ 1
/1 2c0 ,
so as to get one more series:

forth.
Prove the following formulas
11ck
/1 2c1, /1 2c2 ,
...
and so
Bo
Problems
87. Show that if f (x) is any polynomial of nth degree

in x, then there exists the following identity
f(x)=f (O)+;
~/(O)+ x<;:;- 1) ~ 2/(0)+ ... +

+x(x-1) ..~ 1 (x-n+12~n/(O),
where M (0), ~ 2/ (0), ... , ~nf (0) are obtained, proceeding

from the basic series: f (0), I (1), I (2), . . . .
88. Show that if
At (x- 1)+y
A2 ( x-1)(x- 2)+ ...
x n = A 0 +1
+ ~j
then
(x-1) (x-2) ... (x-n),
As=(s+1)"-C!sn+C!(s-1)n+ ... +(-1)"C~1".

nl
{ 1
1
1 }
x(x+1) ... (x+n) -z-+x+1+ +x+n =
1
Ch
=-;z- (x+1)2
+ (x+2)2
q +
(-
1)n
(x + n)2
90. Let
C(lh
(x)=x (x-1) (x-2) ... (x-k+ 1).
Prove that the following identity exists

((Jn (x + y) =cpn (x)
+ C~cpn-1 (x) C(l1 (y) + C~,C(ln-2 ( x) C(l2 (y) + ... +

+ c~- C(li (x) %-1 (y) +((Jn (y).
1

10 xn+yn=pn_l:_pn-2q+n(n-3) p"-4q2- ... +
1
1.2
+( - 1'"n(n-r-1)(n-r-2) ... (n-2r+1) 11 _ 2 r r+
J
20
xn+t _ y'H I
x-y
r!
... '
81
where
p=x+y,
q=xy.
92. Let x+y= 1.

Prove that
xm (1 + C~y + C~>+ 1y2 +
. . . + C~,;;-~2ym-i) +
+ ym ( f + C;,x -j- . . . + C~~2Xm-i) = 1.
93. Prove that the following identity is true

1
cai+i
CV:;:;}.2
+ (x-a)m-2 (b-a)2 + + (x-a) (b-a)m-1 +

1
{
1
~
+ +
+ (b-a)m
(x-b)m + (x-b)m-1 (a-b)
}
+ (x-b)CT,n.!.2
(a-b)m-1
94. Show that constants Ah A 2, A 3 can always be chosen

so that the following identity takes place
(x + y)" = Xn + Yn + A1xy (xn-2 + yn-2) +
+A2x2y2 (xn-4+ yn-4)
Determine these constants.

X1 +x2
=ll1
+ X2Y2 = a2
X1Y~ + X2Yi = ll3
X1Y~ + X2Y~ = ll4.
X1Y1
Show how the general system is solved

x1+x2+xa+ ... +xn-i+Xn=a1
(1)
+ X2Y2 + + XnYn = lli

+ X2Yi + . + XnY~ = a3
(2)
X1Y1
X1Y~
(3)
82
Problems

x+y+z+u+v=2
px qy rz + su tv =-- :-3
2
p x + q2 y r 2 z + s2 u + t 2 u = 1{j
+ +
+
p3x + q3 y + r 3 z + s 3 u + t 3 u =-= 31
p 4 x + q4y r 4 z + s4 u t 4 v =-c 10:)
+
r z+ s u +t v =
5
5
p 5x + qf'y + 5
235
6
6
0
pax+qsy + r z +s u + t v = 674
p'x+q 7 y+r 7 z+s 1 u+t 1 u= 1 G69

pBx+ qBy+ r 8 z +s8 u +t 8 v = 4 52(i
p 9 x+ q9 y + r 9 z + s9 u + t 9 u = 11 59.5.
97. Let m and be positive integers (:::;;; m). Put

(1-xm)(1-xm-1) ... (1-.rm--jl)
= (m, ).
(1-x) (1-.r2) ... (1-:r)
Prove that
1 (m,)=(m,m-);
2 (m,, +1)=(m-1, +1)+xn--t (m-1, );
3 (m, +1)=(, )-t-x(-t-1, )-t-x 2 (+2, )+ ...
+ xm--I (m-1, );
+-
4 (m, ) is a polynomial in x;
5 1-(m, 1) + (m, 2)-(m, 3) + ... is equal to

(1-x)(1-x 2 ) (1-xm- 1 ) if mis even,
0 if m is odd.
(Gauss, Summatio
Werke, Bd. II).
98. Prove that
quarumdam
serierum
singularium,
1 (1+xz)(1+x2z) ... (1+x"z)=

k=n
k(k+1)
... (1-xn-k+1) - 2 - k
+'~ (f-xn)(f-xn-1)
x
z.'
(1-x1) (1-x2) ... (1-xk)
k=i
83
7. Progresstons and Sums
2 (1+xz)(1+x 3z) ... (1+x 2n- 1z)=

k=n
(f-x2n) (f-x2n-2) ... (1-x2n-2k+2) xk2zk.

(1-x2) (f-x4) ... (1-x21l)
99. Let
Prove that
n(n+i)
---+
PtPn-t
Pn
x3
P2Pn-2
x~
- ... + - -Pn= 1 .
tOO. Determine the coefficients C0 , C., C2 ,

following identity
(1+xz)(1+xz- 1 )(1+x 3z)(1+~3 z- 1 )
Cn in the
(1 + x2n-1z) (1 + x2n-1z-1 ) = C0 + C1 (z + z-1 ) +

+ C2 (z 2 + z- 2) + ... + Cn
(zn
+ z-n).
tot. Let
Uk=
sin2nxsin(2n-1)x ... sin(2n-k+1)x

sin x sin 2x .. sin kx
Prove that
1 1- U1 + U2 - U3 + . + U2n =
= 2n(1-cosx) (1-cos3x) ... [1-cos (2n-1) x];
2 1-u:+u~-u~+
= (_
... +u~n=
l)n sin (2n + 2) x sin (2n + 4) x .. sin 4nx
sin 2x sin 4x ... sin 2nx
7. PROGRESSIONS AND SUMS

Solution of problems regarding the arithmetic and geometric progressions treated in the present section requires
only knowledge of elementary algebra. As far as the summing
of finite series is concerned, it is performed using the method
of finite differences. Let it be required to find the sum
Probleml
+ f (2) + ... +I (n). Find the function F (k) which

would satisfy tbe relationship
f (1)
F (k
+ 1) -
F (k) = J-(k).
Then it is obvious that
+ I (2) + ... + I (n) = [F (2) - F (1)1 +

+ [F (3) - F (2)1 + ... + [F (n + 1) - F (n)1
(1)
F (n
+ 1) -
F (1).
1. Let a2 , b 2 , c2 form an arithmetic progression. Prove

..
1
1
1
"ht h at t h e quantities
b+c,
c+a,
a+b
a 1so form an ant
metic progression.
2. ~rove that if a, b and c are respectively the pth, qth
and rth terms of an arithmetic progression, then.
(q -- r) a
+ (r -
p) b
+ (p
- q) c = 0.
3. Let in an arithmetic progression
ap
= q;
aq
= p
(an is the nth term of the progression). Find am.
4. In an arithmetic progression Sp = q; Sq = p (Sn

is the sum of the first n terms of the progression). Find Sp+q
5. Let in an arithmetic progression Sp = Sq. Prove
that Sp+q = 0.
Sm
m2
6. Given in an arithmetic progression -8 = - 2 Prove
am _ 2m-1
th a t - 2n-1 .
an
7. Show that any power nk (k ~ 2 an integer) can be

represented in the form of a sum of n successive odd numbers.
8. Let the sequence ai. a 2 , , an form an arithmetic
progression and a 1 = 0. Simplify the expression
9. Prove that in any arithmetic progression
85
7. Progressions and Sums
n-1
-V an-t +-Van - -V a1 +Van

10. Show that in any arithmetic progression
we have
11. Let S (n) be the sum of the first n terms of an arithmetic progression.
Prove that
s (n + 3)
3S (n
+ 2) + 3S (n + 1) -
2 S (3n) = 3 {S (2n) -
s(n)
= 0.
S (n)}.
12. Let the sequence a 1 , a 2, . . . , an, an+ 1, . . . be an

arithmetic progression.
Prove that the sequence S 1 , S 2 , S 3 , , where
S1
+ ... +an,
+ a2n' S3
= a1
a2
S2 =an+!+ . . .
= a2n+I
+ ... + aan,
. .. ,
is an arithmetic progression as well whose common difference is n 2 times greater than the common difference of the
given progression.
13. Prove that if a, b, c are respectively the pth, qth and
rth terms both of an arithmetic and a geometric progressions simultaneously, then
ab-c. be-a. ca-b = 1.
14. Prove that

(1 + x+x2+ ... +xn)2-xn =
= (1 + x + x 2+ ... + xn-i) ( 1 + x + x 2 + ... + xn+i).
15. Let Sn be the sum of the first n terms of a geometric
progression.
Prove that Sn (S3n - S2n) = (S211 - Sn) 2
16. Let the numbers a 1 , a 2 , a 3 , form a geometric
progression.
8fi
Problems
Knowing the sums
find the product P = a 1a 2 ... an.

17. If ab a 2, ... , an are real, then the equality
(a~+ai+ ... +a~_ 1 )(ai+a:+
... +a~)=
= (a1a2 + a2a3 + ... + an-ian)
is possible if and only if ati a 2, ... , a,, form a geometric

progression. Prove this.
18. Let a 1 , a 2 , , an be a geometric progression with
ratio q and let Sm = a1
am.
Find simpler expressions for the following sums
+ ... +
1
20
S1+S2+ ... +Sn;

1
1
+ a2-a2
+ ... + az
1
1
ah+ah + ah+ah ++ah
a2-a2
l
2
30
n-1
n-1
-azn
+ah
n
19. Prove that in any arithmetic progression, whose

common difference is not equal to zero. the product of two
terms equidistant from the extreme terms is the greater
the closer these terms are to the middle term.
20. An arithmetic and a geometric progression with
positive terms have the same number of terms and equal
extreme terms. For which of them is the sum of terms greater?
21. The first two terms of an arithmetic and a geometric
progression with positive terms are equal. Prove that all
other terms of the arithmetic progression are not greater
than the corresponding terms of the geometric progression.
22. Find the sum of n terms of the series
Sn = 1 x
2x 2
+ 3x3 + . . . +
nxn.
23. Let a 1 , a 2, ... , an form an arithmetic progression

and u 1 , u 2 , , Un a geometric one. Find the expression
for the sum
87
7. l'rogressions and S um.s
24. Find the sum

(x +
+) 2 + ( x2 +
:2 )2 +
... + ( 3;" +
x1n ) 2
25. Let
Prove that
S _ n (n+ 1)
12
l -
..., _n(n-1-1)(2n-1-1)
._,2 _
6
,
26. Prove the following general formula

(k -l-1)
I
sk +
(k+1)k
12
sIH -!-
(k-1-1)k(k-1)
123
sk-2 -t-
+ (k-l-1)S1 + S0 =(n+1)"+ 1 -1.

27. Put
1"-t-2k+ ... +n"=Sk(n).

Prove the formula
nSk (n) = Sk+dn) +Sk(n-1) + Sk(n-2)
+ ... +
+ sk (2) + sk (1).
28. 1 Prove that

1"+2"+3"+ ... +n"=An"+i+Bn"+Cn"-1 + ... +Ln,
i.e. that the sum S k (n) can be represented as a polynomial
of the (k + 1)th degree in n with coefficients independent
of n and without a constant term.
2 Show that A =
k! 1,and B= ~ .
29. Show that the following_ formulas take place

S _ n (n-1-1) (2n-1-1) (3n2+3n-1)
30
4 -
'
S _ n2(n+1)2 (2n2+ 2n-1)

;;-
12
'
S _ 6n7+21n6+21n-7n3+n _
a-
42
n (n+ 1) (2n+1) [3n2 (n+1)2-(3n2+3n-1)]
42
88
Problems
S 7 = 3n8-j 12n7+14n6-7n4+2n2 =
24
n2(n+1)2 [3n2(n+1)2-2 (2n2
=
24
+ 2n-1))
30. Prove that the following relations take place

Sa= Si, 4s: =Sa+ 3S5, 2Sd- Sa= 3s;, S5 + S1 = 2s;.
31. Consider the numbers B 0 , B 1 , B 2 , Ba, B 4 , determined by the symbolic equality
(B + 1)h +t - Bh +t = k + 1
(k = 0, 1, 2, 3, ... )
and the initial value B 0 = 1. Expanding the left member
of this equality according to the binomial formula, we have
to replace the exponents by subscripts everywhere. Thus,
the above symbolic equality is identical to the following
common equality
Bk+1 +Ck+1Bk+CL1Bk-1+ ... +C~+1B1+Bo-Bk+_1=k+1.
1 Compute B 0 , B 11 B 2 , , B 10 with the aid of this
equality.
2 Show that the following formula takes place
1h-t-2h+3h+ ... +nh=
k!i {nk~
+Ck+1B1nh-t-CL1B2 nh-t
32. Let X1i x 2 ,

It is known that
Xn
+ ... + C~+tBkn}.
form an arithmetic progression.
X1 + x 2 + ... + Xn=a,
x~+x~-t- ...
+ x;,=b
Determine this progression.

33. Determine the sums of the following series
1 1+4x+ 9x 2 + ... +n 2 xn-1;
20 i3+2ax+3ax2+ ... +naxn-1.
1 1
0
2 1 - 23 -1- 45 0
2n-1
+
7+
(
B ' + - 1)
+2+4+3+
2n-1
n-l
2n -1
2n-t '
89
1 1- 2 3 -4 + ... + ( -1 )n-l n;
20 i2-- 22 + 32_ ... + (-1)n-1 n2;
3 1-3 2+5 2-7 2 + ... -(4n-1) 2;
4 2i2+322+ ... +(n+1)n2.
36. Find the sum of n numbers of the form 1, 11, 111,
1111, ....
x4n+2 + y4n+2 =
= {x2n+l_ 2x2n-1y2 + 2x2n-3y4- ... + (-1t 2xy2n}2+

= {y2n+l _ 2y2n-lx2 + 2y2n-3x4_ ... + (-1t 2yx2n}2.
38. Find the sum of products of the numbers 1, a,
a 2 , ' an- 1 , taken pairwise.
(x"- + x!-i) +
x!-
2 ( xn-2 +
2)
+ ... + (n-1) ( x-\-
xL! (x;-=:-n
! )-t-n=
2
1 +
1
1
1
2.3+3.4++n(n+1)2
20 1.2.3+2.3.4+ ...
n(n+i)(n+2)
1 ( 1
=2
3o
1
2
1.3.5 + 3.5.7 +
1
-n+1;
z-- (n+1)(n+2));
n
+ (2n-1)(2n+ 1) (2n +3)

=
n (n+1)
2 (2n+1) (2n+3l
41. Compute the sum

14
24
31
n4
5 =n+n+5.7+ + (2n-1)(2n+1)
90
Problems
42. Let ai, a 2 ,

Prove the identity
an be an arithmetic progression
_f_+_1_+
... +-1-=
~ (-1 +-1 + ... +_!_).
a1an
a 2an-t
anal
at
an
a1
a2
an
43. Prove that
10
n
n+1
n+p
(n+1)!+(n+2)1+ ... + (n+p+1)!
20
1
+
1
(n+1)1
(n+2)1
=-;;y--
(n-tp-f-1)!'
+ + (n+p+1)!
1
<
1
1
J
<--;;-1 [ nr-(n+p+1)1
(n and p any positive integers).

1
2
4
x+1 + x2+1 + x4+1
1
2n
+ + x2n-+ 1
1
45. Let Sn=1+2+3+ .. +--;;-

Prove that
n+p+1 { n-p
n-p-1
1
}
n-p+1 n(p+1)+(n-1)(p+2)+ +n(p-11)
=Su-Sw
46. Let
1
Sn=1+ 2
+ 3 + +--;;-
s-n+1_{
1
_j_
2
n- 2
n(n-1) ' (n-1) (n-2)
+ -j- n-2}
2-3
Prove that s; = Sn
47. Let Sk be the sum of the first k terms of an arithmetic progression. What must this progression be for the
ratio ~~ to be independent of x?
48. Given that ai, a2 , , an form an arithmetic progression. Find the following sum:
S=~
i=I
a;a;+1a;+2
a; +ai+i
91
49. Find the sum

1
cos a. cos (a.-t
~)+cos (a.+~) cos (a.+ 2~) +
...
+
1
+cos [a.+ (n-1)
~)cos (a.+ n~)
50. Show that

1
a.
a.
a.
tan a+ 2 tan 2 + 4 tan T + ... + 2n-i tan 211 _1 =

=
zn-1
cot
a.
2n-1 -
2 cot a.
51. Prove the following formulas

1 sina+sin(a+h)+ ... +sin[a+(n-1)h]=
. nh sm
. ( a+n-1
sm
22
.
h}
sm 2
2 cosa+cos(a+h)+ ... +cos[a+(n-1)h]=

. nh
(
n-1 }
sm
2 cos a+-:r- h
.
sm 2
52. Find the following sums

S
Slll
. (n -1)
n
n + S. l l l2rtn + . . . + Slll
n
rt
2n
S =Cos-+
cos-+
...
n
n
(n-1) n
n
+cos--~
53. Show that

sina.+sin3a.+ ... +sin(2n-1)a.
t
cos a.+cos 3a.+ ... +cos (2n-1) a.= an
na.
54. Compute the sums
8 11 = cos 2 x + cos 2 2x + ... + cos 2 2n.r,

s; = sin 2 x f- sin 2 2x !- ... + sin 2 2n.c.
92
Problems
55. Prove that
-Pt
Pt
i=p
"'1 . mni . nni

LI Slll P+ 1 Slll P+ 1 =
i=1
1 if
m +n is divisible
by 2(p+1);
1 if m - n is divisible
by 2(p+1);
if m :f= n
and if m + n and m - n a re
not divisible by 2(p+1).
56. Find the sum

x
arctan f+f. 2x2 +arctan 1 + 2 3x2 + ..

x
+arctan f+n (n+f) x2
+
(x
> 0).
57. Find the sum

r
arctan+ ... +arctan i+ anan+i

1 + a1a2 +arctan 1- +
a2a3
if ai. a2 , form an arithmetic progression with a common
difference r (a 1 0, r 0).
58. Compute the sum
>
>

.
X1Slil-
.2n
+ X2Slil. 2 -+
n
n
2n
x1sm-+x
2sm 2 -n +
n
211:
+ X3Slil. 3 n211:+ .. -t-Xn-1Slil' ( n- 1) --;-=a2,
!)3
8. Inequalities
.3n
x 1 sm+ x 2 sm. 2 -3n:+
n
n
L
331t
-r-X3S1n
---t-...
x 1 sm
(n-1)n:
+ x2 sm 2
+ Xn-1s1n"( n- 1')3n
-=a3,
n
(n-1)n:
. 3 (n-1)n:
+ x 3 srn
+
n
... +
+ Xn-1 sm. (n -1) (n-1)n:

=an-I
n
8. INEQUALITIES
Let us recall the basic properties of inequalities.

1 If a> b and b > c, then a> c.
2 If a > b, then a + m > b + m.
3 If a> b, then am> bm form> 0 and am< bm for
m < 0, i.e., when multiplying both members of the inequality by a negative number, the sign of the inequality is
reversed.
4 If a> b > 0, then ax> bx if x > 0.
This last inequality is readily proved for a rational x.
Indeed, let us first assume that x = m is a whole positive
number. Then
am-bm = (a-b) (am-l + am- 2b + ... + bm-l).

But either of the bracketed e4pressions on the right
exceeds zero, therefore am-bm 0 and am> bm. We now
put x = _.!._ . Then ax - bx= 1;/a -y/b.
m
We have
>
(a- b) =
('V"'a-y/b)
('"' am-1 + ... +y/bm-1).
Hence, actually, it follows that
94
Problems
Let, finally, x=.!!....

We have
q
p
ax-bx=aq -bq =t/aP-.p'"bP.

But aP
>
bP (as has been proved), consequently, t/ aP
>
> ;Y bP. To pr0ve this inequality for an irrational x we may

consider x as a limit of a sequence of rational numbers and
pass to the limit.
5 If a > 1 and x > y > 0, then ax > aY; but if 0 <
<a< 1 and x > y > 0, then ax< aY. The proof is
basically reduced to that of aa > 1 if a> 0 and a> 1
and can be obtained from 4.
6 log" x > log" y if x > y and a > 1; and loga x <
< log" y if x > y and 0 < a < 1.
Out of the problems considered in this section, utmost
interest undoubtedly lies with Problem 30 both with
respect to the methods of its solution and to the number
of corollaries. Problem 50 should also be mentioned with
its inequalities useful in many cases.
1. Show that
1
1
+1>1 (
..
.
)
n+
2 n, a positive
mteger
.
1 + n+ 2 + 2n
2. Let n and p be positive integers and n ;;;::= 1, p ;;;::= 1.

Prove that
1
n+1-
n+p+1
<
1
(n-t 1)2
1
+ (n+2)2
+ (n+p)2 <
<n--n+p
3. Prove that the sum of any number of fractions taken
1
1
1
1ess
r 32 , ~ , . . . is a ways
than unity.
4. Prove that
;;;;-;i~-Vn.
5. Show that if a is a defective value of VA to within
unity (a<VA<a+1), then
1
f rom among t h e sequence ~
A-a2
a+ 2a+1
<
y-
A-a2
A< a+ 2a+1 + 4(2a+1)
95
8. Inequalities
6. Prove that
1+
1
-V12 + V"3
+ ... +
1
-Vli
<2
v-+ 1- 2.
n
7. Prove that
8. Prove that
8
cot 2 ~1
(0
+cote
< e < n).
9. Show that if A+B+C= n(A, B, C>O) and the

angle C is obtuse, then
tan A tan B < 1.
10. Let tan e = n tan cp
(n
>
0).
Prove that
tan 2 (8-cp) ~
(n-1) 2
4n
11. Show that if

1
R +tan a: tan~= tan'\', then cos 21'-< 0.
cos a cos I'
12. Let us have n fractions
(i=1,2, ... ,n).
t;;
Prove that the fraction ~~ !~:t:::

is contained between the greatest and the least of these fractions.
13. Prove that m+n+ +v ab ... l is contained between
the. greatest. and the least one of the quantities
m/-
-V a,
14. Suppose 0
Prove that
t
an a:
n/-
-v
b, ... ,
-vp/-l.
< a: < ~ < '\' < ... < '). < ~ .
sin a+ sin ~+sin y + ... +sin A. < ta ~

< cos
n ''"
a+ cos
y + ... +cos A.
~+cos
96
Problems
15. Let x 2
Prove that
= y 2 + z 2 (x, y, z > 0).

x'J..
>
x'J..
< y'>.
y'>.
+ z'>.
+ z'>.
if /.. > 2,
if/..< 2.
16. Prow that if

a2
+b
= 1,
m2
+n
=1,
then jam+ bnl ~ 1.

17. Let a, b, c and a + b - c, a + c be positive.
Prove that
abc ~(a + b - c)_ (a + c -
18. Let
19. Let
b) (b + c - a).
+C=
A+ B
Prove that
n.
+B +C =
n (A, B, C
>
0).
+B +C =
n (A, B, C
>
0).
Prove that
20. Given
b, b -t- c -
Prove that
3
1 cos A+ cos B +cos C ~ 2

0
3 V3
2 cos 2 cos 2 cos-y:s:;8 -.
21. Prove that

V<a+c) (b+d)~ -V ab+ Vcd
(a, b, c and d
22. Prove that

a3+b3
( a+b
)3
(a> 0, b > 0).
> 0).
91
8. Inequalities
23. Prove that

1 atb
20
~Vab
(a b>O);
.!_(a-b)2_a+b_Vab-~(a-b) 2
~
a
:::::::: 2
"'::::: 8
b
"f
......._b
a~
24. Prove that

a+b+c ;::::::::y
'- a/-b
a c
3
(a, b, c
> 0) .
25. Prove that
Va1a2+ Va1lia+. . +-V~<n;-i (a1+ a2+
...
(ai>O; i=1, 2, ... , n).

26. Let ai > 0 (i = 1, 2, ... , n) and a 1a 2
Prove that
(1 + a1) (1 + a2) ... (1 + an) ~ 2n.
+an)
an =
1.
27. Prove that

1 (a+b)(a+c)(b+c)~8abc
20 _a_+_b_+_c_ 2
b+c
a+c
28. Prove that
a+b
(a, b, c>O):
2'
Y(a +k) (b+ l) (c+ m)~t/abc+ :f.klm

(a, b, c, k, l, m > 0).
29. Prove that
9
..!_
..L ..!_ + .!_ . . . ._
a ' b
c r a+b-J-c
a, 'c
(x1
>
> 0)
O; i = 1, 2, ... , n),
the equality being obtained only in the case

X1 =X2= . . . =Xn
31. Let ai, a 2 ,

(a;> 0).
. ,
an form an arithmetic progression
98
Problems
ii--
Prove that v a1an

In particular
n/
a1a2 ... an~
a1+an
2
-Vn<y/n!<nt1.
32. Lel a, b, and c be positive integers.
lJ
Prove that aa+b+c.ba+b+c.ca+b+c~-}(a+b-f-c).

33. Prove thaL if a, b, c are positive, rational and such
that the surn of every two numbers exceeds a third one, then
(
b- c )
1+
a-
11 (
c- a ) b ( 1
1++a--c-b ) c <;1.
b
34. Let a, b, c, ... , l be n positive numbers and

S=
S
a+b+c+ ...
Prove that -s-a

-1--b
s-
+ l.
+ ... +-s- 1 ~-n- 1 .

S
n2
35. Prove the inequality

l lln
2)
( a1 b1 llz b2 ,I -r-' lln b n )2---'::::::: (a 2
11 a 2
2 t -,x (b~+b;+ ... -1 b~).
36. Ptove the inequality

ai-f-a 2
+ ... +an:::;;;Vn(a;+ a~+
... +a~).
37. Prove that
(x1-l-X2
t- ... +xn)(-Xj1-+-X21-
t ... +-Xn1-)~n 2 .
38. Let
.Tt-\- X2 -jX1X2 -f X1.T3 -f-

Prove that
]!_
n-\-
n-1 , /
V p
... -j- Xn
= p,
+ X1Xn -j- X2X3 -f- -f
Xn-1Xn = q.
2 _~
>-x->-_e__n-1-./ 2 _~
n-1q::o-- '"""n
n V p
n-1q.
39. Let a, b, c, ... , l be n real positive numbers and

let p and q be also two real numbers.
99
8. Inequalities
Prove that if p and q are of the same sign, then

ll
(aP+q
-f
bp+q
+ ... + [P+q) > (aP + bP + ... -f [P) X

x
(aq -t--bq+ ...
+ zq).
And if p and q have different signs, then

n (ap+q
+ bp+q + ... + zP+q)~ (aP + bP + ... + ZP) x

x (aq + bq +- ... + zq).
40. Prove that

1 (1 + a)"-> 1 + a'A (a is any positive number; 'A> 1
is rational).
2 (1 + ct)"- < f-~A. (ct > 0 real, '). , rational and positive, a'A < 1).
41. Let u,, = ( 1 + ~
n is a positive integer.
r,
1 Prove that
2 Prove that Un is a bounded quantity, i.e. there exists
a constant (independent of n) such that Un is less than this
constant for any n.
42. Prove that
V2>.Y3>Y4>~5>Y6
> >Yn>
> n+v---v n+ 1 > ...
43. Prove tl;iat
2> V3>~4 >~5 > ... >"-Jin> ;Yn-+ 1 > ...

44. Let us have
ll21X1
+ ll12X2 + + ll1nXn
+ ll22X2 + + ll2nXn
=
=
Y2
lln1X1
+ lln2X2 + + llnnXn
Yn,
ll11X1
where a;i
>
0 and rational, xu
>
0.
Y1
Problems
100
Furthermore, it is given that

ak1
a1k
+ ak2 + + akn =
+ a2k + ... + ank =
1,
1
(k = 1, 2, ... , n).
Prove that
45. Let
>
a;> 0, b;
0 (i = 1, 2, ... , n).
Prove that
{Y(a 1
+b,) (a + b2) ... (an t bn)~;Y a1a 2 ... a,,+

+ ;Y b1b2 ... b11.
2
46. Prove that

( ..r1+r2+.-I
Xn)k~ x~+x~+ ... +x~

""'
'
> 0.
47. Let the function cp (t) defined in a certain interval
possess the following property
n and k are positive integers, x;
cp ( .!..!
1t2) <
cp (t1Hn cp (t2)
for any two t 1 and t 2 not equal to each other.

Then
where t1' t 2 , , tn are n arbitrary values from the given

interval not equal to one another.
48. Find the greatest value of the sum
S = sin a1
if
a;
>
+ sin a 2 + . . . + sin an
0 and a1 + a 2
+ ... + an
= n.
49. Let x, p and q be positive, p and q being integers.

Prove that
101
8. Inequalities
if p > q (x =I= 1).

50. Let x > 0 and not equal to 1, m rational.
Prove that
mxm-l (x - 1)
>
xm - 1
>
m (x - 1)
if m does not lie between 0 and 1.

But if 0 < m < 1, then
mxm-l (x - 1) < xm - 1 < m (x -
1).
51. Prove that
(1
+ x)m ~ 1 + mx
if m does not lie in the interval between 0 and 1;
(1
+ x)m ~ 1 + mx
if 0 ~ m ~ 1 (m rational, x
52. Prove that
>
-1).
( x)
+ x~ ~- ... + x~ ) P ~ (
xj + x~ : ... + x~ ) q'
p, both q and p being positive integers.

53. Find the value of x at which the expression
(x - X1) 2 + (x - X2) 2 + ... + (x - xn) 2
takes on the least value.

54. Let x 1 + x 2 + ... + Xn = C (C constant). At what
x 1 , x 2 , . , Xn does the expression x~ + xi + . . . + x~
attdin the least value?
x1 + x 2 +
55. Let x 1 > 0 (i = 1, 2, ... , n) and
+ + Xn = C.
At what values of the variables x 1 , x 2 , , xn does the
expression
>..
.X1
+ Xz>.. + ... + Xn,..
(A. rational) attain the least value? 56. Given x 1 > 0 (i = 1, 2, ... , n) and the sum x 1 +
+ x 2 + ... + Xn = C = const. Prove that the product
XtX 2 ... Xn reaches the greatest value when Xt = x 1 =
= =
Xn
=nc
Problems
102
> 0 (i = 1, 2, ... , n) and the product

is constant, i.e., x 1x 2 . . Xn = C.
Prove that the sum x 1
x2
Xn attains the
least value when
57. Given
x 1x 2x 3 . . .
58. Let
x;
Xn
x;
>
+ x + ... +
2
+ ... +
0 (i = 1, 2, ... , n) and the sum x 1

= C = const.
Xn
Show that
x~1x~2
... x~n
takes on the greatest value when

Xj
X2
f!2
Xn
=== ==-=
d-2+ .. +n '
> 0 (i = 1, 2, ... , n) and rational.

59. Let
(i=1,2, ... ,n)
a;> 0, X; > 0
and
ll1X1
ll2X2
llnXn = C.
Prove that the product x 1x 2 xn attains the greatest
value when
;
+ ... +
>
(A.;
0 and rational).
Prove that
takes on tho greatest value when
=
61. Let x1 1 x~ 2
Show that
--
... ---
x~n = C = const.
8. Inequalities
103
attains the least value if
at!
ann
a22
{ai, x; > O; 'Ai and ; > 0 are rational).

62. Find at what values of x, y, z, ... , t the sum
x2
y2
z2
t2
+ + ... +
takes on the least value if

ax+ by
+ ... +kt =A
(a, b, ... , k and A constant).
63. At what values of x, y does the expression
u = (a1x
+ b1Y + c1) 2 + (a2x + b2Y + c2) 2 + ... +

+ (anx + bnY + cn)
take on the least value?

64. Let x 0 , x 1 , , Xn be integers and let us assume
Xo
<
X1
<
X2
< ... <
Xn.
Prove that any polynomial of nth degree xn + a 1xn-t +

+ ... +an attains at points x 0 , x 1 , , Xn the values at
least one of which exceeds or equals ;~ .
65. Let 0 ~ x ~ ~. At what value of x does the product
sin x cos x reach the greatest value?
66. Let
1t
x+y+z,..-= 2 ;
1t
O~x~ 2 ,
1t
O~y~ 2 ,
3T
O~z~2.
At what values of x, y and z does the product tan x tan y x

the greatest value?
67. Prove that
x tan z attain
+ 1 + n + 2 + ' + 3n + 1 > i
(n a positive integer).
68. Let a > 1 and n be a positive integer.
Prove that
n+i
n-1
an-1~n (a-2--az-).
104
Problems
69. Prove that

1
2< 1+-z+3+ +zn-1 < n

(n a positive integer).
70. Prove that
1
1
1
1+1
1~1
a+b c+d
(a,b,c,d>O).
a+c+b+d
9. MATHEMATICAL INDUCTION
This section contains problems which are mainly solved
using the method of mathematical induction. A certain
amount of problems is dedicated to combinatorics.
1. Given
and
Vo
= 2,
Vt
= 3.
Uo
= 0,
U1
= 1.
Prove that
2. Let
and
Prove that
Un=
3. Let a and A
>
2n - 1.
0 be arbitrc1.ry given numbers and let
... ,
=
Prove that
(an-1 +~).
an-t
9. Mathematicul Induction
105
for any whole n.

4. The series of numbers
is formed according to the following law. The first two
numbers a 0 and a 1 are given, each subsequent number being
equal to the half-sum of two previous ones. Express an in
terms of tzo, a 1 and n.
5. The terms of the series
are determined as follows
a1
= 2 and an = 3an _1
Find the sum

a1
+ 1.
+ a + ... + an.
2

are connected by the relation
an
kan-1
+ l (n
2, 3, ... ).
Express an in terms of a 1 , k, land n.

7. The sequence a 1 , a 2 , satisfies the relation an+ 1
- 2an + an -1 = 1.
Express an in terms of a 1 , a 2 and n.
8. The tern., of the series
ai, a 2 , aa, .
are related in the following way an+3 - 3an+ 2+ 3an+1-an = 1.

Express an in terms of a 1 , a 2 , a 3 and n.
9. Let the pairs of numbers
(a, b) (a1, b1) (a2, b2)
be obtained according to the following law

a+b
41-=-2-,
Problems
106
Prove that
2 (b-a)(1-4'7",
1 )
a,,~a+ 3
b,,=a+
(b-a) { 1-+
2 .~n)

are determined by the relations
Xn =
Xn-I
+ 2Yn-1 sin
2 a,
Yn
=-
Yn-t
+ 2.c,, _
cos 2
a.
Besides, it is known that x 0 = 0, y 0 = cos a.

Express xn and Yn in terms of a.
11. The n um hers
are related as follows
axn-1
+ ~Yn-1
Yn =
YX11 -I
+ 6y,, -I
Xn
~y =fa
(a6 -
Express Xn and Yn in terms of x 0 ,

0).
an<l n.
Yo
are determined by the relation

X 11
CXXn -I
+ ~Xn
Express Xn in terms of x 0 , x 1 and n.

13. The terms of the series x 0 , x 11
the relation
P.Xn-1
Xn=
-2
are counected by
+ q:tn-2 .
p+q
Express Xn in terms of x 0 , z 1 and n.

14. The terms xo, x1, x 2, .. are determined by the equality
107
9. Mathematical Induction
Express Xn in terms of x 0 and n.

Consider the particular cases
Xn-1
Xn-1
Xn=2 Xn-1
+1'
+1
Xn=---
Xn-1+3
15. The numbers:
are determined by the following law
a 0 and b 0 are given, and a 0 > b 0 > 0. Express an and b11

in terms of a 0 , b 0 and n.
n
2n+1
-f-
1
23-2
1
(2n)3-2n
+ +
1
1
n-f-1 + n+2 + +

(1 - x) (1 - x 2) ... (1 - xn) + x (1 - x 2) (1 - x 3) ... x
X (1-xn)+x 2 (1-x 3 ) . (1-xn)+ ... +
+ xh (1- xh+l) ... (1- xn) + ... + xn-l (1- xn) + x"
x
1 - x2
x2
+ 1-
x4
+ 1-
x4
xB
x 2 n-l
+ + 1 - x 2n
x-x 2n
= 1- x 1
2 ''
-x
19. Check the identity

(1-f-x)(1-f-x2 )(1-f-x4) ... (1+x2n-l)=
= 1 + x + x2 + x3 +
... +
,1'2n-1
20. Prove the validity of the identity

t+.!.+a+1+(a+1)(b+1)+ ... +
a
ab
abc
+ (a+1) (b+1) ... (s+1) (k+1) =

abc ... skl
(a+1) (b+1) ... (k+1) (l+1)

abc ... kl
Problems
f08

b+c+a+ ... +k+ l
a (a+b+ c+ ... +k+l)
b .
a (a+b)
-I
c
(a+b) (a+b+c)
+ (a+b+c) (a+b+c+d) + +
+ (a+b+ ... +k) (a+b+

l
... +k+l)
22. Let
1 q q (1-
z)
+1q q
2
2 (
1 - z) (1 - qz) + ...
+
qn
- (1-z)(1-qz) ... (1-qn- 1z)=Fn(z).
1 -qn
Prove the identity
1+Fn(z)-F,,(qz)=(1-qz)(1-q2z) ... (1-q"z).
23. Prove that

k=n
'1
L.J
(1-an)(1-an-l) ... (1-an-k+l)

1-a"
=ll.
k=l
24. Compute the sum

S
a
a (a-1)
a (a-1) (a-2) f- a (a-1)
... (a-n+ 1)
b(b-1) ... (b-n+1)
n=b+b(b-1)+ b(b-1)(b-2)
(bis not equal to 0, 1, 2, ... , n-1).

25. Let
+ ...
S,,=a1+(a1+1) a 2 + (a1+1) (~+ 1) a 3

+
+(a1+1)(~+1) ... (an-1+1)an.
I
Prove that
Sn = (a1 + 1) (~ + 1) . . . (lln + 1) - 1.
26. Prove the following identities:
X=n
~ x(x+1) . ...(x+q)=
q!
2 n(n+1) ... (n+q+1);
X=i
X=n
~ .x(x+1) ~ .. (x+q)
x-=1
; { :1-(n+i)(n+i) ... (n+qJ
f09
9. Mathematical Induction
(1 - 21- T1) + (13 - 61 - 81) +

+ (2n 1
1 - 4n
... +
~ 2-
:n ) =
1( 1 1 1
1 1)
=2 1 -2+3-4+ +2n-1 -2n
28. Let us have a sequence of numbers(Fihonacci's series)
0, 1, 1, 2, 3, 5, 8, 13, 21, ....
This sequence is determined by the following conditions
Un+I =Un+ Un-I
and u0 = 0, u 1 = L
Show that there exist the following relations
1
Un+2
Uo
2 Uzn+2 =
U1
U1
+ U2 + +
+ U3 +
U5
+ +
Un
+1;
Uzn+l;
3 U2n+1=1+u2+u4+ +uzn;
4 - Uzn-1 + 1 = U1 - llz + U3 + . . +
Uzn-\ -
5 Uzn-z+1=U1-u2+ua-u4+ ...
+Uzn-1;
6 UnUn+t = u; + u~ + ... + u~;

7 U~n = U1U2 + U2U3 + + Uzn-1Uzn;
8 Un+1Un+2 - UnUn-t3 = ( -1
9 U~, - Un+1Un-f = ( -1r+l;
10 U~ -
Un-2Un-tUn+1Un+2
r;
= 1.
29. Compute the sum

2
Un+2
f:2 + """f3 + + -Un_+_t_U_n+_3_

30. Prove the relations
1
Un+p-1 = Un-1Up-1
Uzn-1
U2n-1 = UnUn+i -
+ UnUp;
= u~ + u~- 1 ;
Un-2Un-I
U2n:
110
Problems
31. Prove that
u:, + u;,+
k~'
32. Prove lhal
Un=
u~_ 1
1-
[!:=..!)
2:
U3n
Cn
k-1
k~O
33. Find the number of whole positive solutions of the

equation x 1 x 2
x 11 = m (m a positive integer).
34. Prove that the total number of whole nonnegative
solutions of the equations
x
2y = n, 2x
3y = n - 1, ... , nx
(n
1) y = 1,
+ + ... +
+
(n + 1) x + (n + 2) y
+ +
= 0
is equal ton
1.
35. Show that the total number of whole nonnegative
solutions of the equations
x
+ !iy
c_
3n ~
1,
4x
+ 9y
= 5n - 4, 9x
7n - 9, ... , n 2x
+ (n + 1)
+ 16y
y = n (n
+ 1)
is equal Lo n.
36. There are n white and n black balls marked 1, 2, 3, ... ,
n. In how many ways can the balls be arranged in a row
so that all neighbouring balls were of different colour?
37. I 11 how many ways is it possible to distribute kn
distinct objects inlo k groups, each consisting of n elements?
38. How many permutations can be made up of n elements in which the two elements a and b never stand side
by side?
39. Find the number of permutations of n elements in
which none of the elements occupies the original position.
40. In how many ways can n distinct letters be arranged
in r squares (first, second, ... , rth square) so that each
square contains at least one letter (the order of the letters
inside the squares is disregarded)?
10. LIMITS
We take as known the concept of a variable and its limit,
as well as the basic theorems on limits which are usually
treated in elementary textbooks of algebra (the limit of a
sum, product and quotient). Let us here remind the reader
..
10. Limits
111
of one of the indications for a limit to exist: if a variable

increases but remains smaller than a certain constant, then
such a variable has a limit (likewise, a variable which,
when decreasing, remains greater than a certain constant
also has a limit). When dealing with an infinitely decreasing
geometric progression and, in general, with simple infmite
series, one should bear in mind that the symbolic notation
U1
denotes none
+ U2 + U3 + + U n + .
other than lim (u + u + ... + un)
1
if
n--.oo
such a limit exists. If there is no limit, then the series

Ut
+ U2 + U3 + + Un +
is said to be divergent, and it is useless to speak of its numerical value.

1. Let Xn = a" and I a I < 1. Prove that lim Xn = 0.
n->oo
2. Prove that
an
lim nr-=0
n-+oo
for any real a.

3. Find
4. Let
23-1
33-1
n3-1
Pn = 2a+1 33-i 1 n3 +1
Prove that lim Pn =

n-+oo
~ .
5. Prove that
]'
lffi
n--..oo
1k~2k+
... +nk_~_1_
k +1
nk+I
(k a positive integer).
n }
1
k+1
"~2
112
Problem$
(k a positive integer).
7. Let us have a sequence of numbers
the equality
Xn
determined by
Xn-1 +xn-2
Xn=
and the values x0 and x 1

Prove that
lim
Xn
n +oo
xo+2x1
3
8. Let N > 0. Let us take an arbitrary positive number x 0 and form the following sequence
X1
X2
+( + ~ ),
Xo
=; (x + ~ ),
1
1 ( Xp-1
Xp= 2
Prove that lim x,. =
+ ~),
Xp-1
V N.
n~oo
9. Generalize the result of the preceding problem for the

extracting a root of any index from a positive number.
Prove that if
m-1
X1=-m-xo+
N
mxm
I'
m-1
Xz = - m - X1
N
+ mxm-1
'
1
m-1
Xp
= -m- Xp-1 +
N
m-1 '
mxp-l
then
.
1im
n~oo
m/N
Xn=v
113
10. Limits
10. Prove that
11. Let
k= n
~ (~-1).
S,.=
k=I
Find
12. Let the variable

law of formation
Xn
be determined by the following
Xo=Va,
X1=
-v a+ Va,
x2 =Va+V a+Vll,
X3
ii a+Va +Va+Va,
Find
lim Xn.
n-->oo
13. Prove that the variable

Xn
= 1
1
1
1
+ V2
+ V3
+ ... + Vn
-
1r::
2y n
has a limit as n -+ oo.

14. Let us be given two sequences
Xo,
X1,
Xz, . . . ,
Yo, Y1, Yz,
(xo >Yo> 0),
where each subsequent term is formed from the preceding

ones in the following manner
Xn=
.Tn-1
+2 Yn-1
Yn =
VXn-IYn-1
114
Problems
Provo that x 11 and y 11 have limits which are equal to each

other.
15. Let
s 1 = 1 + q + q2 + .. . I qi< 1,
s = 1 + Q + Q2 + .. . I QI< 1.
Find
1 + qQ + q2Q2 + ....
16. Let s be the sum of terms of an infinite geometric
progression, a 2 the sum of squares of the terms. Show that
the sum of n terms of this progression is equal to
_ [ s2 - crZ
s { 1
s2+ aZ
n }
17. Prove that

1 lim nhx"= 0 if J x I< 1 and k [is a positive integer;
2
H'm ;Yn= 1.
18. Find the sums of the following series

1
1 n1 + 2-3.
0
_l_
1
3.4
20 _1_+_1_+
123 2.3.4
+ n(n+1)
l
+ ... .'
+ n(n+1)(n+2)
1
+ ....
...
19. Prove that the series

1
1+2+3+4+ +n+
is a divergent one.
20. Prove that the series
1
1 +-a+-a+-a+
2
3
4
a>
+
+ ... +
+ ...
1
... +-a+
...
n
is a convergent one if
1.
21. Find the sums of the following series
1 1 + 2x + 3x2 + ... nx"- 1 + ... ;
2 1+4x+Hx2
n2 x 11 - 1
3 1-t-23 .i+3 3 x2
+n 3 x"- 1 + ... (J:rJ<1).
+ ... ;
10. Limits
115
)n
22. 1 Prove that the variable 11,, = ( 1 I n1

(n = 1, 2,
3, ... ) has a limit.
2 Denoting the limit Un by e so that lim { 1 +
= e,
n-oo
i..)n
n
prove that
1
e=1+1+n+1.2.3+ ... + 123 ... k
+ 123 0... kk
(0<0<1).
23. Let 0
<x < ~ .
Knowing that lim sin x = 1, prove that

X-+0
.
- 1 x.
x-smx::::::::,
6 3
24. 1 Prove that the series
is a convergent one.
2 Prove that for any real number ro (0 < ro <1) it is
always-possible to find, and in the unique way, ai (0 ~ ai ~
~ 9; ai being integers), such that
(i.e. to expand the real number in decimal fractions).

3 Show that if a decimal fraction
is ftni\'.} Or periodic (i.e., for instance, an+I = ai, an+2 =

= a2, ... , a 2n = an, ... , so that the period contains n
digits: ai, a 2, , an), then ro is a rational numbPr.
25. Prove that the number~ determined by the following
series are irrational ones
1'
()=
1 + 1il
1 + ltil+
1
T1 -t -0
1n 2
116
Problems
where l is any positive integer.

20
(J) -
___!__
+ _1
_
1
z1.2 + 11.z.3
+ ... + z12! ... n + ... '
-1 1 1.z.3.~
where l is any positive integer.
26. Prove that e is an irrational number (see Problem 22).

27. Let
1
w=-z1 +-z-1-+-z-11-+
.. + z12
1 .. n
1 + .. '
12
123
where 1<l1 ~ Z2 ~ l 3 and li are integers. Prove that
w s rational only when lk (beginning with a certain k)
are all equal to one another.
28. Prove that the variable
1
Un
= 1+ 2
+ 31 + ... + n1 -
log n
has a limit.
29. Prove the following formula:
1
:rt
v~V~+~ v~ v~+~ v~+~v~+ . .
SOLUTIONS
SOLUTIONS TO SECTION 1
1. Proved immediately by a check.
2. If we remove the brackets from the right member
and apply the formula for a square of a polynomial, then
it is easily seen that all the doubled products are cancelled
out, and we get the required identity.
3. If the identity of the preceding problem is used, then
from the condition of our problem it follows that
(a2 + b2 + c2 + a2) (x2 + y2 + z2 + t2) = 0,
whence either a 2 + b2 + c2 + d 2 = 0, or x 2 + y2 + z2 +
+ t 2 = 0.
But the sum of the squares of real numbers equals zero
only when each of the numbers is equal to zero. Therefore,
from tl~e equality a 2 + b2 + c2 + d2 = 0, we get a = b =
= c = d = 0, and from the equality x 2 + y 2 + z2 + t 2 =
= 0 we have x = y = z = t = 0.
Hence follows the required result.
4. This identity can be checked directly, and also can
be obtained from identity (2) if we put in it d = t = 0
and replace y by -y and z by -z.
5. If we expand the right member of the equality, then
all doubled products are cancelled out and the validity of the
identity becomes obvious.
6. Put in identity (5) a 1 = a 2 = a 3 = ... =an= 1,
b1 =a, b2 = b, ... , bn-1 = k, bn = l.
We then get
n (a2 + b2 + c2 + ... + k2 + z2) =
=(a + b + ... +Z) 2 + (b - a) 2 +
(c - a) 2
+ ... + (k -
1) 2
118
Solutions
But since by hypothesis
n (a 2
+ b + ... + k +
l2) = (a
+ b + ... + k + l)
2,
we have
(b -
a) 2
+ (c -
+ ... + (k
a) 2
l) 2 = 0.
Hence a = b = c = . . . = k = l.
7. Make use of identity (5). By hypothesis

a; + a~ +
... + a; = 1,
b: + b; +
... + b; = 1.
Therefore we have
(a1b1
+ a2b2 + ... + a,,b,,)

= 1 -
(a1b2 - a2b1) 2
(a1b3 - a 3 b1) 2
(a,,_1bn - anb11-1)~.
Whence
< (a1b1 + a2b2 + ... + a b < 1.

-1 < a1b + a2b2 + ... + a,,b < +1.
0
Thus,
11
11 )
11
8. We have
(y
+zBut
(y
2x) 2 -
+z-
(y - z) 2
2x) 2
+ (z + x + (x + y -
2y) 2
2z) 2
(z - x) 2 +
(x-y) 2 =0.
(y - z) 2 = 4 (y -- x) (z - x)
(using the formula for a difference of squares).

Likewise we find
+x (x + y -
(z
2y) 2
2z) 2
(z -
x) 2 = 4 (z - y) (x - y),
(x - y) 2 = 4 (x - z) (y - z).
Consequently,
4 (y - x) (z - x)
+ 4 (z
y) +
+ 4 (x - z) (y - z) = 0
y) (x -
Removing the brackets, we get
2x 2
+ 2y + 2z
2
2 -
2xz - 2yz - 2xy = 0
Solutions to Sec. 1
119
or
(x -
y) 2 + (x - z) 2 + (y - z) 2 = 0,
whence
y = z = 0.
9. The first identity is obvious. Let us rewrite the second

one in tho following way
(6a 2 - 4ab + 4b 2) 3 - (4a 2 - 4ab + 6b 2)3 =
= (3a2 + 5ab - 5b 2)3 + (5a 2 - 5ab - 3b 2)3
Applying the formula for a difference of cubes to the left
member and the formula for a sum of cubes to the right
member, we find that it suffices to prove tho following identity
(3a 2 - 2ab + 2b2) 2 + (3a 2 - 2ab + 2b 2) (2a 2 - 2ab +3b 2) +
+ (2a 2 - 2ab + 3b 2) 2 = (5a 2 - 5ab - 3b 2) 2 - (5a 2 - 5ab - 3b 2) (3a 2 + 5ab - 5b2) +
+ (3a 2 + 5ab - 5b 2) 2.
This identity is proved by directly removing the brackets.
10. To see whether the identity under consideration is
valid, we may rewrite it as
(p2 - q2)4 = (p2 + pq + q2)4 - (2pq +q2)4 +
+ (p2 + pq + q2)4 - (2pq + p2)4.
It remains to simplify tho right member and to show that
it is equal to the left one.
Using the formula A 4 -B 4 = (A + B) (A - B) (A 2 +B 2),
we get the following expression for the right member
(p2 + 3pq + 2q2) (p2 - pq) [(p2 + pq + q2)2 +
(2pq + q2)2J + (2p2 + 3pq + q2) (q2 - pq) x
x [(p2 + pq + q2)2 + (2pq + p2)2] = (p + 2q) x
x p (p2 - q2) [(p2 + pq + q2)2 + (2pq + q2)2] +
+ (2p + q) q (q2 - p2) [(p2 + pq + q2)2 +
+ (2pq + p2)2J = (p2 - q2) {(p2 + pq + q2)2 x
x [p2 + 2pq2- 2pq - q2] + (p2 + 2pq) (q2 + 2pq) x
x [2pq + q2- 2pq - p2]} = (p2 - q2)2 {(p2 + pq+q2)2 - (p2 + 2pq) (q2 + 2pq)} = (p2 - q2)4.
Solutions
120
11. Check by direct substitution.

12. Check by substitution.
13. 1 The cases n = 0, 1, 2 are readily checked directly.
At n = 4 let us rewrite the identity in the following way
(ix - ky) 4
(ix - kz) 4
(iy - kz) 4 - (iy - kx) 4
(iz - kx) 4
- (iz - ky) 4 = 0.
Transform the first two terms

(ix - ky) 4 - (ix - kz) 4 = [(ix - ky) 2
(ix - kz) 2 l (2ix - ky - kz) k (z - y).
(1)
+ +
+
By virtue of the equality x

y
z = 0, we get
2ix - ky - kz = (2i
k) x.
The expression in square brackets can be rewritten as follows
(2i 2
+ 2ik) x + k
2
Thus, we have
(ix - ky) 4 - (ix - kz) 4 =
= k (2i
k) (y 2 - z2) f(2i 2
(y 2
+z
2 ).
+ 2ik) x + k
2
(y 2
+ z2)].
It remains to transform the following expressions

( iy - kz) 4 - ( iy - kx)4,
(iz - kx) 4 - (iz - ky) 4
(1')
(2)
(3)
But it is easily seen that expression (2) is obtained from the

first one, already considered, by means of a circular permutation of the letters x, y and z, i.e. when x is replaced by
y, y by z, and z by x. Expression (3) is obtained from (2)
also through such a permutation. Therefore, there is no
need to repeat computations for simplifying expressions (2)
and (3); it is sufficient only to apply appropriate permutations to the result obtained. We then have
(iy - kz) 4 - (iy - kx) 4 =
= k (2i
k) (z 2 - x 2) f(2i 2
2ik) y 2
+ k2 (z2 x2)),
(2')
+
+
121
Solutions to Sec. 1
(iz - kx) 4
(iz - ky) 4
+ k) (x
+ k2 (x2 + y2)].
= k (2i
y2) [(2i 2
2 -
+ 2ik) z +
2
(3')
Anrl adding expressions (1'), (2') and (3'), we get

+ k) {(2i 2 + 2ik) [(y 2 - z2) x 2 + (z2 - x 2) y 2 +
+ (x2 _ y2) z2) +
+ k2 (y4 _ z4 + z4 _ x4 +
k (2i
+ x4 -
y4)}
0.
2 At n = 0 the relation is obvious. Let us denote, for

brevity, the sum in the left member of the equality by
'2: (x + k)n,
and the sum in the right member by
~ (x
At n
l)n.
1 we have to prove that
8x
+ Lk =
+ L l,
8x
i.e. we have to prove that
Lk
h z.
Finally, we have to check that
Z}k
Z}l.
But
'L k
2: l
At n
+ 5 + 6 + 9 + 10 + 12 + 15 =
1 + 2 + 4 + 7 + 8 + 11 + 13 + 14
60,
= 60.
2 we have to prove that
~ (x
+ k)
(x
+ l)
2,
i.e. that
sx 2
+ 2x'L k + L k
2 =
sx 2
+ 2x 'L z + L z
And so, it remains to prove that

~k2 = ~
z2,
122
Solutions
which is easily checked directly.

Likewise, to prove the last case (n = 3) we have only to
show that
14. The first idenLity is proved in the following way
+ b + c + d) + (a + b - c - d) +
+ (a + c - b - d} + (a + d - b - c) =
= [(a + b) + (c + d)l2 + [(a + b) - (c + d )1 +
+ [(a - b) + (c - d)l2 + [(a - b) - (c - d)J2 =
= 2 (a + b) + 2 (c + d) + 2 (a - b) + 2 (c - d) =
= 2 [(a + b) + (a - b) + 2 [(c + d) +
+ (c - d) = 4 (a + b + c + d
(a
2]
2]
2 ).
The second and third identities arc also proved by a direcL

chock with some preliminary transformations.
15. Rewrite our equality as follows
[(a
+ b + c)
4 -
(a 4
(a 4
+-
b4
(a 4
b4
(a 4
b4
+ b + c + [(b + c - a)
+ c + [(c + a - b)
+ c + [(a + b - c)
+ c = 24 (a b + a c + b c
4 )]
4 -
4 )]
4 -
4 )]
4 -
4 )]
2 2
2 2
2 2 ).
Consider the first term.

\Ve have
(rt 2
+ b + c + 2ab + 2ac + 2bc) a b c =

= (ia b + 6a c + Gb c + 4ac (a + c +
+ 4ab (a + b + 4bc (b + c -j- 12a bc +
+ 12b ac + 12c ab.
2
2 2
2 2
4 -
2 -
2)
2 2
2)
4 -
2)
2
The remaining terms are obtained from the first one by

means of successive substitutions: -a for a, -b for b,
-c for c. Adding the terms, we make sure that our identity
is valid.
123
Solutions to Sec. 1
16.
s (s +
+
We have
2b) (s - 2e) + s (s - 2e) (s - 2a) +
s (s - 2a) {s - 2b) = (s - 2a) (s - 2b) (s 2a (s - 2b) (s - 2e) + s (s - 2a) (2s - 2e = (s - 2a) (s - 2b) (s - 2e) + 2a (s - 2b) (s + s (s Transform the sum
2e) +
2b) =
2e) +
2a) 2a.
2a (s - 2b) (s - 2e) + s (s - 2a) 2a =

= 2a [(s - 2b) (s - 2e) + s (s - 2a)l =
= 2a [(s - 2b) (s - 2e) + (s - 2a) (s - 2b) +
+ 2b (s - 2a)l = 2a [(s - 2b) (2s - 2e - 2a) +
+ 2b (s - 2a)l = 2a [(s - 2b) 2b + 2b (s - 2a)l =
= 2a -2b [s - 2b - 2a] = 4ab 2e = Sabe.
17. Expanding the expression in the left member in
powers of s, we get
0+b+rj~-~~+~+~+~+~+~+
+ 2s3 - 2s2 (a + b + e) +
+ 2s (ab + ae + be) -- 2abe.
Since a + b + e = 2s, we have
2s3 - 2s (a 2 + "b 2 + e2) + a 3 + b3 + c3 + 2s3 - 4.i1 +
+2s (ab + ac + be) - 2abe = -2s (a 2 + b2 + e2) +
+ aa + b:i + c:i + 2s (ab + ae + be) - 2abe =
= a 3 + b3 + e3 + (a + b + e) (ab + ae + be - a 2 - b2 - e2) - 2abe.
Directly transforming the last expression, we make sure
that it is equal to abe (see also Problem 20).
18. We have
(20 2 - 2a2) (20 2 - 2b 2) = (a 2 + e2 - b2) (b 2 + e2 - a 2) =
= e4 _ (a2 _ b2)2.
124
Solutions
Using a circular permutation, we obtain

(2a 2 -
2b 2) (2a2 -
2c2) = a 4
(b 2 -
c2) 2,
(2a 2 -
2c2) (2a2 -
2a 2) = b4
(c 2 -
a 2) 2.
Hence
4 [(a2 - a2) (a2 -
b2) + (a2 - b2) (a2 - c2) +

+(a2 _ c2) (a2 _ a2)) = a4 + b4 + c4 _ (a2 _ b2)2-
(b2 _ c2)2 _ (c2 _ a2)2
+ 2a2b2 +
= -a4 _
b4 _ c4 +
2a2c2
2b 2c2 = -[a 4 - 2 (b 2 + c2)a2
+ (b2 - c2)2l = -[a4 - 2 (b2 - c2) a2 +
(b2 _ c2)2 _ 4a2c2] = 4a2c2 _ (a2 _ b2 + c2)2 =
= (2ac + a 2 - b2 + c2} (2ac - a 2 + b2 - c2) =

= (a + b + c) (a + C-=- b} (b - a + c) (b +a - c).
But
a
a
+b+c=
+c-
2s,
b = 2 (s -
+ b - c = 2 (s - c),
b), b + c - a = 2 (s a
a)
and we see that the identity is valid.

19. We have:
+ 3x2 (y
(x + y + z) 3 = x 3 + y 3 + z3
+ z) +
+ 3y 2 (x + z) + 3z2 (x + y) + 6xyz.
Hence
+ z)a - xa - ya - z3 = 3 {x2y + x2z + yzx -j+ + z2y + 2xyz} = 3 {z (x2 + y2 + 2xy) +
(x + y
z2x
y2z +
+ z2 (x + y) + xy (x + y)} = 3 (x + y) {z (x + y)
+ z2 + xy} = 3 (x + y) (x + z) (y + z).
Thus,
(x + y + z) 3
x 3 - y 3 - z3
3 (x + y) (x + z) (y + z).
125
Solutions to Sec. 1
20. We have
(x
+ y + z)
= x3
+ y + z + 3xy (x + y + z) +
+ 3xz (x + y + z) + 3yz (x + y + z)3
3xyz.
Consequently
:i3
+y +z
3
+ y + z) 3 (x + y + z) X
X (xy + xz + yz) = (x + y + ..z) X
x (x + y + z
xy - xz - yz).
a + b - c = x, b + c - a = y, c + a - b =
readily seen that x + y + z = a+ b + c and,
3xyz = (x
3 -
3 -
21. Put
= z. It is
2 -
hence, we have to simplify the following expression

(x
+ y + z)3 -
xa -
ya -
za.
On the basis of Problem 19 we have
+ y + z)
But x + y =
(a + b + c)
(x
3 -
x3
y3
z3 = 3 (x
+ y)
+ z)
(x
(y
+ z)
+ z = 2a, y + z = 2c, therefore,

+ b - c) (b + c - a)
- (c + a - b) = 24abc.
2b, x
(a
-
3 -
3 -
22. On the basis of Problem 19 we have

x3 + y3 + z3 = (x + y + z) 3 - 3 (x + y) (x
+ z) (y + z).
Putting here x = b - c, y = c - a, z = a - b, we find

x
y
z = 0, x
y = b - a,
x
z = a - c, y
z = c - b.
+ +
Hence
(b - c) 3
+ (c -
a) 3
+ (a -
b) 3
=
3 (a -
b) (a -
c) (c -
b).
23. Readily obtained from Problem 20. But it is possible

to use the following method
(a
+ b + c) (a + b + c
2
2)
0.
126
Solutions
since
a+ b
+e=
0.
Hence,
a3
+ b + e + ab (a + b) + ae (a + e) + be (b + e)
3
But
a
+b=
-e, a
+e=
-b, b
+e=
Now the required identity is obvious.

24. We have
(a + b + e) 2 = 0,
a 2 + b2 + e2 = -2 (ab + ae
= 0.
-a.
+ be).
Squaring both members of the latter equality, we get

(a 2
+ b2 + e2) 2 = 4 [a 2b2 + a 2e2 + b2e2 + 2a be +

+ 2b ae + 2e ab] = 4 [a 2b2 + a 2e2 + b2e2 +
+ 2abe (a + b + e)l = 4 [a2b2 + a 2e2 + b2e2J.
2
On the other hand,

(a2
+ b2 + e2)2
= a4
+ b4 + e4 + 2 (a2b2 + a2e2 + b2e2).
Hence
4 (a 2b2 + a 2e2 + b2e2) = 2 (a 2 + b2 + e2) 2 - 2 (a 4
Comparing it with the equality
4 (a 2b2
+ a 2e2 + b2e2)
+b +e
4
4 ).
= (a2 + b2 + e2)2,
we get the required result.

25. Since
(a - b) + (b - e)
+ (e -
a)
0,
the result follows immediately from Problem 21.

26. 1 We have (see Problem 23)
a3
Whence
(a 3
+b +e
3
+ b + e3) (a 2 + b2 + e2)
3
= 3abe.
= = 3abe (a 2 + b2 + e2).
127
Solutions to Sec. 1
Then, transforming the left member, we obtain

a5
+ b5 + e5 + a b
+ e5 -
+ b) + a e (a + e) +
+ b e (b + e) = 3abe (a + b + e
a b e - a e b - b e a =3abe (a + b + e2).
+e
abe (ab
2 2
(a
2 2
2 2
or
a 5 + b5
2 2
2 2
2 2
2)
+b +c
2).
Hence
a5 + b5
But
5 -
+ ae +
+ ae +
-2 (ab
be) = 3abc (a 2
be) = a 2
+b +e
Hence follows the final result.

2 The answer follows immediately from Problem 23 and 1.
3 Let us write the relations
2 (a 4
+e
b4
4)
a3
(a 2 + b2 + e2 ) 2
b3 + e3 = 3abe
(Problem 24),
(Problem 23).
Multiplying these equalities, we find

2 [a 7
t-
b7
+ e + a3b3 (a + b) + a e
+ b c3 (b + e)l =
(a + e) +
3abe (a 2 + b2
3 3
Hence
2 [a7 + b7
+e
or
2 (a 7
+b +e
7
7)
2abe (a 2 b2
+ae +
2 2
3abe (a 2
b2e2)
a2 b2
+ae +
2 2
b2 e2
1(a
+ b +2
c2) 2
+ b +c
2
2) 2
= Babe (a 2
But
2) 2
b:1e3 a] =
a3 b3 e - a3 c3b -
7 -
+c
b2
+e
2) 2
(Problem 2'1).
Therefore
2 (a 7
+b +e
7
7)
7
~ abe (a 2
+b +e
2
2) 2
Using the result of 1, we llnally get tho required relation
128
Solutions
27. For the sake of convenience let us introduce the summation symbol. And so, we put
k=n
a1+a 2 + ... +an= ~ah.

k=1
Using this symbol, we can now write

k=n
a 1b1 + a 2b2 + ... + anbn
k=n
= ~ akbk = a1b1 + ~ akbk.

k=2
k=1
But it is obvious that

bk
(b1 + b 2 + . . .
+ bk)
(b1
+ bz + ... + bk-1) =
therefore our sum takes the following form

k=n
a1b1
+ k=2
~ -ah (sk -
sk-1)
~ a1b1 +
k=n-1
k=2
+ ansn -a2S1 = (a1 - az) S1

h=n-1
k=2
k=n
~ aksk-1 +
aksk -
+ anSn +
k=3
h=n-1
aksk -
k=2
k=n-1
+ k=2
~ (ah -ak+1) sh+ ansn =
a3) Sz + ... + (an-1 -a,,) Sn-I+ an Sn.
ak+1Sk = (a1 - az) S1
= (a1 -
a 2) S1
+ (a2-
28. Readily proved if we remove the brackets in the

left member and use the relation
a1
+ az + ...
+an
= z S.
29. Substituting into the given expression x' and y'

for x and y, we find that
A' = Aa 2 + 2Bay + Cy 2 ,
C' = A~ 2 + 2B~8 + C8 2 ,
B' = Aa~ + B (all + ~y) + Cyll.
Making up the expression B' 2 required identity.
A' C', we easily check the
129
Solutions to Sec. 1
30. We have
i-=n
i=n
i=n
i=n
~ Piqi= ~Pi (1-pi)= ~ Pi i=I
i=I
i=I
since
np =Pt+ P2
i=n
~PI =np- ~
i=I
i=I
py,
+ + Pn
Further
i=n
i-::-'n
~ p;qi = np-
i=I
2::
(Pi - P+ p) 2 =
ic-1
i=n
i=n
=pn - ~[(p;-p)2+2ppi-P2]=np- ~ (p;-p)2i=I
i=I
i=n
i=n
-2PL p;+nrr=np-l:i (p;-p) 2 -np 2

i=I
i=I
But
np -
np 2 = np (1
p) = npq.
Thus, we get
P1q1 + pzq2 + + Pnqn = npq - (p1 - p) 2 - (P2 - p) 2 - - (Pn - p) 2
31. Indeed
1
1
2n-1
__1_ {
- 2n
1
+ I ' 1
2n -1 T =
(2n-1)+1
(2n-3)+3 +
+ 1+(2n-1)
}-1(2n-1) + 3(2n-3)
(2n-1)1
--
+3
1 {1
=2il
1
2n- 3
T+ 2n-1 +3+ 2n-3
1
1}
+ + 2n-1
+T =
='1--( 1++++++ 2n~1 )

32. 1 It is obvious that
1
1
1
Sn=f+2+3++n=
=n+[(i-1)+({--t)+
+({--1)++( !-1)1=-
=n-({-+
;+ ... +n:-1 ).
130
Solutions
k-n
Sn=~
1<~-n
k=n
! , nsn -_L.J"'l
11-k+-k _ '\.-, ( n-k

k
- Li
k
k=I
k=l
+ l) .
k=I
Hence,
n-1
n-2
1 )
nsn=n+ ( - 1 - + - 2 - + + n-:-1 .
33. Add to and sul.itract from tho left member the following expression
We get
1
1
1
1
1 - 2 + 3 - 4 + + 2n-1
= (1 +
; +
+ +
2n
-2n=
~1 ) -
=( 1 + ; ++++zn~1
+ (
+!
+
+ ... +
2~i
) -
2 (
=
-
(1
+n1) =
~n
)+
1
1
1
1
1 + 2 + 3 + 4 + + 2n-1
1
1
+2+3+
+ {-+ +
+ ... +
2~
+2n-
n+1+n~2
+2n1
34. We have
(1+
1 ) (1-
2a
~1
) (1+
3a
~1
) X
1
x ( 1 + (2n-1)1 a-1 ) ( 1 - 2na-:--1
) -a(2a-2)3a ... (2n-1)a(2na-2) _
(a-1) (2a-1) (3a-1) ... (2na-1) 1a3a5a ... (2n-1)a(2a-2) (4a-2) ... (2na-2)
= (a-1)(2a-1) ... (na-1)[(n+ 1)a-1][(n+2)a-1] ... [(n-t-n)a-.1] =
1a3a5a ... (2n-1) a-(a.-1) (2a.-i) ... \na.-1) . 2n =
[(n+1)a.-1] ... [(n+n)a.-1] (a.-1)(2a.-1) ... (na.-1)
1a.3a.5a. ... (2n-1)a.
2n
= [(n+i)a-1) ... [(n---in)a.-1]
131
Solutions to Sec. 1
But
1 3
r:
<J
(2 -1) 2n
n
__
1.2.3.4.5 ... 2n 2n __
_
246 ... 2n
= 1 ~--~--~-~-. ~ 2n =(n+t) (n+2) ... 2n,

wherefrom we obtain the required identity.
35. Let a
x <a + 1, where a is an integer. Subdivide
the interval between a and a + 1 into n parts. Then x will
lie in one of these subintervals, i.e. we can find a whole
number p (0
p < n - 1) such that
<
<
a+L~x<a+
n"""
P+ 1 .
n
Therefore
a+
P+1 ~x+_!__
n--=
<a+
P+2
n'
Hence
[xl=[x+!J= ... =[x+ n-~- 1 ]=a

tx+ n-;:p J= ... = tx+ n-;:1 J=a+L
Consequently
[xJ+[ x++ J++[x+ n-;:1 J=

= (n-p)a+ p(a+ 1)=an+ p.
On the other hand, from the inequality
a+~ ~x<a+ p~i
Solutions
132
we get
an
+ p ~ nx <
an
+ p + 1,
hence,
[nx] =an+ p,
and the formula is proved.

36. We have
+ b) cos (a
cos (a
b)
= [cos a cos b - sin a sin bl X

X [cos a cos b + sin a sin bl = cos 2 a cos 2 b - sin 2 a sin 2 b = cos 2 a (1 - sin 2 b) - (1 - cos 2 a) sin 2 b = cos 2 a - sin 2 b.
37. Expanding the bracketed expressions in the left
members, we easily prove the equalities.
38. We have
(1-sin a) (1-sin b) (1-sin c) =
(1-sin2 a) (1-sin2 b) (1-sin2 c)
(1 +sin a) (1 +sin b) (1 +sin c)
cos2 a cos2 b cos2 c
b
= cos a cos b cos c = cos a cos cos c.
39. Multiplying both members of the given equality by

(1
+ cos a)
(1
+ cos ~)
(1
+ cosy),
we get
[(1
+ cos a)
(1
+ cos ~)
(1
+ cos y)l2
=
=
sin 2 a sin 2
40. Using the formula

sin x cosy
we get
2 cos (a
2 cos (~
~ [sin (x
+ ~) sin (a + y) sin (~ -
+ y) + sin (x -
~) ~
y)
y)],
sin 2a - sin 2~,

sin 2~ - sin 2v
and so on. Hence follows the identity.
sin 2 y.
133
Sulutions to Sec. I
41. Using the formula

sin x sin y = ~ [cos (x - y) -
cos (x + y)l,
we get the identity

(cos 2b - cos 2a) (cos 2d - cos 2c) +
+ (cos 2b - cos 2c) (cos 2a - cos 2d) +
+ (cos 2b - cos 2d) (cos 2c - cos 2a) = 0.
Let cos 2b
then
(a -
a, cos 2a
~) (y -
6)
(a - 6) (~
~) (y - 6) + (a y) (6 - ~) = (a y) (~ - y) + (y -
= (a -
+ (a
+ (a -
~.
cos 2d
cos 2c = 6,
y,
- y)
(a - y) (6 - ~) =
y + y - 6) (~ - y) +
~) (y - 6)
6) (~ - y) +
(a - y) (6 - ~) = 0.
But (a - ~) (y - 6) + (y - 6) (~ - y) = (y - 6) (a - y)
and (a - y) (~ - y) + (a - y)-(6 - ~) =(a - y) (6 -y);
hence the required sum is equal to (a - y) (y - 6) +
+ (a - y) (6 - y) = 0.
42. 1 Summing the
first two cosines, we get
2 cosy cos (~ - a); the sum of the second two cosines
yields 2 cos (a + ~) cos y. The further check is obvious.
2 Analogous to 1.
43. We have
sin
(A +
! )+cos (A +
+ sin (~-A
2
) =sin
- _!!_4 ) =
(A+
! )+
2 sin ~cos
4
(_::__A
4
_!!_
) .
4
With the aid of a circular permutation we obtain (denoting

the transformed sum by S)
~2
=- cos (
+ cos (
- A-
1)+cos ( ~ -
~ - C - ~ ) = 2 cos ( ~ -
)+
At B -
B-
xcos(A-;B +B!c)+sin (
C) X
+c+: ).
134
Solutions
Making use of the relation A+ B + C = n, we can show

that
cos ( ...::._ - A+ B - B + C ) = sin ( ...::._ + i'._
4
+ _i_8)'
Therefore we have
- 8 -=2sin (...::_+i'_+_i_) cos (
s
2
s
v2
+ 2 sin (
= 2 sin (
~ + ~ + ~ ) [cos (
A-B
+ B-C) +
2
s
+ ~ ) cos ( ;
)=
A-; B+ B-; C)+
+cos (...::.++_i_)J-4sin
8
2
8
-
(...::._+~+_!!_)
8
2
8
(n
. (n
B+C).
C+A)
xsm 3 + 2
8
sm 3 + 2
8
44. Carrying out some transformations analogous to the

previous ones, we obtain the following result
.A+.B
.C
A+ cos 4B+ cos 4C =
sm
sm 4 +sm
4
4 +cos 4
= 4 V2 cos (
) cos (
) cos (
+ ~ ).
45. We have
sin 2a = 2 sin a cos a,
sin 4a = 2 sin 2a cos 2a,
sin Sa = 2 sin 4a cos 4a,
sin 2n a = 2 sin 211 - 1 a cos 211 - 1a.
Multiplying term by term and dividing both members by
the product
sin 2a sin 4a ... sin 2n- 1 a,
we get
sin 2n a = 2 11 sin a cos a cos 2a ... cos 2n- 1a,
Solutions to Sec. 1
135
whence
sin 2na
cosacos2a ... cos2n-1a=--2n sin a

46. We have
. 2:rt
2 . :rt
:rt
sm
15 = sm 15 cos 15 ,
sm 15 =
2 sm
. 4:rt cos 4:rt ,
15
15
sm 15 =
8:rc
sm 15 =
4:rc
2 .
16:rt
2 sm
. 8:rt cos 8:rt .
15
15
sm
Multiplying the equalities and noting

.
:rt
8:rc
:2r-;
15
~os
2:rt
f;>.,
that sin \ 6; =
7:rt
= -sm 15 , cos 15 = -cos 15 , we find

:rt
2:rt
4:rt
7:rt
cos 15 cos 15 cos 15 cos 15 = 24 .

Further
5:rt
cos 15 = 2
and
.
6:rc
sm 15 =
2 sm
. 3:rt cos 3:rc ,
15
15
12:rt
sm 15 =
2 sm
. 6:rc cos 6:rc .
15
15
Hence
3:rt
6:rt
cos 15 . cos 15- = 22 .

The rest is obvious.
47. We have
tan (A+B) _sin (A+B) cos A_ sin (2A+B)+sinB _
tan A
- cos (A+ B) sin A - sin (2A+B)-sin B -
48. From the give11 :'Plations we get

sin 2B =
~ sin 2A,
3 sin 2 A = 1 - 2 sin 2 B = cos 2B,

hence
cos (A+ 2B) =cos A cos 2B - sinA sin 2B =
= cos A3sin 2 A-
~ sin A sin 2A =0.
Solution~
49. We have
2 cos a cos cp = cos (a + cp) + cos (a -
cp).
Consequently the expression under consideration is equal to

cos 2 cp
+ ros
(a + cp) -
+cos (a
+-
[cos 2 (a + cp) +
cp) cos (a - cp)l = cos 2 cp - cos 2 a cos 2 cp + sin 2 a sin 2 cp = sin 2 a.
50. We have, for instance,

a 2 + a' 2 + a" 2
cos 2 cp cos 2
1p
+ sin 2 cp sin 2 ' cos 2
+ cos 2 cp sin 2 1p + sin 2 cp cos 2 1p cos 2
o+
o+
sin 2 cp sin 2 6
(the doubled products in the first two squares are cancelled

ont). Hence
a 2 + a' 2 + a" 2 = (cos 2 cp cos 2 1p + cos 2 cp sin 2 ') +
+ (sin 2 cp sin 2 ' cos 2 6 + sin 2 cp cos 2 ' cos 2 6) +
+ sin 2 cp sin 2 o = cos 2 cp +
+ (sin 2 cp cos 2 o + sin 2 cp sin 2 6) = 1.
The remaining equalities are pro_ved similarly.
1. Rewrite the identity in the following way
3
q +q
3 (2p3 - q3)3 (p3,-q3)3 -
p - p
3 (p3 - 2q3)3
(p3+q3)3 .
It is evident that the right member can be obtained from

the left one by permuting p and q. Let us reduce the left
member to such a form, wherefrom it would be seen that
after the permutation its value remains unchanged. Then
the validity-1'.H the identity will become clear.
We have
q3
{( 3 +
(p3+q3)3.
p
3)3
+ (2 p 3 -q3)3} --
9p 3q3
( 6+
(p3+q3)3 p
q - p
6,..6)
'i
137
Solutions to Sec. 2
2. We have
p3 q3
(p+q)3 p3q3
=
=
( 1
+ (p+q)4
1 )
+ q)
6 (p
+ (p+q)5 pq =
p2 - pq - q2
3
( 1
1
2 )
(p + q)2 p3q3 + (p + q)4
pi+ lj2" + pq =
p2 - pq-+ q2
3
( p1 + _q1 ) 2 =
(p+q)2 p3q3 + (p+q)4
-p2+q2
p2-pq+ q2
'""' (p -+ q)2 p3q3
+ (p + q)2 p2q2
(p --1 q)2 p3q3
x {p2- pq+q2+3pq}=
p31-q3.
3. Grouping the last two terms of the sum, we get

2
_(_P_+_q--:-).,-4
q3-p3
p3q3
+
-
q-p
p2q2
(p -t- q)4
2 (q- p)
(p--f- q)4 p3q3
+ q -t- 2 pq) -2
'
2(q-p)
(P+ q)2 p3q3 '
Adding now the first term, we find

1
(p + q)3
q4-p4
p4q4
2(q-p)
+ (p + q)2 p3q3 =
q-p
p4q4
4. We have to prove that

1+x
1+y
1-x 1-y
1+z _
'T-Z- -
i-
Replacing x by its expression, we find \~: =

Since
y and z are obtained from x by means of a circular
permutation of the letters a, b, c, we have
1+y
1-y
=c-'
1 +z
1-z
Hence, the required identity is obvious.

5. We have
a-b+c-d
a-b-c+d
Solutions
138
c+D an d converse1y 1"f

But l. f ~
B -- .
D , then A+B
A-B = C-D,
there exists the second of these equalities, then the first
one exists as well. Reasoning in the same way (putting
A = a
b
c d, B = a
b - c - d, C =a - b c - d,
D = a - b - c
d, we find
+ + +
+
a+b
a-b
a+b
c+d
-c+d
-=- O ra-b
- ---c-d
-c-d
Hence
a
11=cr
or
-c=cr.
6. The denominator has the form

bcy 2
bcz 2 - 2bcyz
acz 2 acx 2 - 2acxz
abx 2
aby 2 - 2abxy = c (ax 2
by 2 )
b (ax 2
cz 2)
2
2
a (cz
by ) - 2bcyz - 2acxz - 2abxy =
= (a
b
c) (ax 2
by 2
cz2) - c2z 2 - b2y 2 b
c) X
- a2x 2 - 2bcyz - 2acxz - 2abxy = (a
X (ax 2
by 2
cz 2) - (ax
by
cz) 2.
+
+
+
+
+
+ +
+ +
ax + by + cz =
Since, by hypothesis,
turns out to be equal to
(a+ b
+
+ +
+ +
+ +
0, the denominator
+ c) (ax 2 + by 2 + cz2),
and our fraction is equal to

1
a-f-b+c
7. Reduce to a common denominator the expression on

the left. The numerator of the fraction obtained will be
equal to
x2y2z2 (a2 _ b2)
b2 (x2 _ a2) (y2 _ a2) (z2 _ a2) _
_ a2 (x2 _ b2) (y2 _ b2) (z2-b2).
It is obvious that
(a2 _ x2) (a2 _ y2) (a2 _ z2) =
as _ (x2
_ xzy2z2.
=
+ y2 + z2) a' + (x2y2 + x2z2 + y2z2) a2 _
139
Solutions to Sec. 2
Hence
(b2 ~ x2) (b2 _ y2) (b2 _ z2) =
= b6 _ (x2 + y2 + z2) b4 +
+ (x2y2 + x2z2 + y2z2) b2 __ x2y2z2.
Substituting these expressions into the numerator and

performing all the necessary transformations, we obtain the
required value of the fraction.
1
8. So=
(a-b)(a-c)
1
(b-a)\b-c)
1
(c-a)(c-b)
Reducing the fractions to a common denominator, we have

1
So= (a-b)(a-c)(b-c) {(b-c)-(a-c)+(a-b)}=O,

a
S1 =
(a-b) (a-c)
= Ta-b)(a-c)(b-c)
a2
(a-b) (a-c)
(a-b)
(a~c) (b-c)
b
(b-a) (b-c)
(c-a) (c-b)
{a(b-c)-b(a-c)+c(a-b)}=O,
b2
(b-a) (b-c)
c2
(c-a) (c-b)
{a2 (b- c)-b 2 (a- c)
+c
(a-b)}.
Consid'er the numerator.

We have
a 2 (b - c) - b2 (a - c) + c2 (a - b) =
= ab (a - b) - c (a 2 - b2) + c2 (a - b) =
= (a - b) (ab - ca - cb + c2) =
= (a - b} [a (b - c) - c (b - c)] =
= (a - b) (b - c) (a - c),
wherefrom it follows that S 2 = 1. S 3 , S 4 and S 5 can be
computed analogously, but we shall proceed here in a somewhat different way.
It is easily seen that there exists the following identity
(x - a) (x - b) (x - c) = x 3 - (a + b +c )x2 +
+ (ab + ac + be) x - abc.
140
Solutions
Putting, x = a, x = b and x = e, in turn, we get the following equalities
a3
+ b + e) a + (ab + ae + be) a (a + b + e) b + (ab + ae + be) b (a + b + e) e + (ab + ae + be) e (a
3 -
abe
c3 -
0,
= 0,
abe = 0.
abe
Further, divide the first of them by (a - b) (a - e),

the second by (b - e) (b - a) and the third by (e - a) X
X (e - b), and add them term by term. Then
Sa -
+ b + e)
(a
S2
+ ae + be) S 1 S = S = 0, S
abe S 0 = 0.
(ab
But since it is known that 0

1
2 = 1, we have:
Sa= a+ b e.
To compute S 4 let us take the preceding identity and
multiply its members by x. We obtain
a) (x -
x (x -
b)
e)
(x -
x4 +
(a
(ab
+ b + e) x3 +
+ ae + be) x
2 -
abex.
Proceeding analogously, we find:
4 -
(a+ b
+ e) Sa+
(ab+ ae
be)
+ e) Sa -
(ab+ ae +be)
2 -
abe
0.
Hence
--,(a+ b
(a
+ b + e)
ab -
2 -
= a
2 =
ae -
+b+
2
e2
be
+ ab + ae +
be.
Likewise, for computing S 5 (multiplying the original identity by x 2 ), we find

S5
(a
+ b + e)
S4
+ ae +
(ab
be) Sa -
abe S 2 =
0.
Consequently
+ b + e) (a2 + b2 + e + ab + ae + be) - (ab + ae + be) (a + b + e) + abe =

= (a + b + e) (a + b + e + abe =
= aa + b3 + c3 + a 2 b + a 2e + b2 a + b 2 e +
S 5 = (a
2)
+ e a + e b + abe.
2
141
Solutions to Sec. 2
9. This problem is solved analogously to the precerling

one. Namt>ly, tho equalities S 0 = S 1 -~ S 2 = 0, S 3 --- 1
are established by a direct check; and to compute S 4 we
may resort to the following identity
(x -
a) (x -
b) (x - e) (x - d)
= x
(a
4 -
-+ b + c + d) x +
3
+ (ab+ ae +ad+ be+ bd +de) x

- (abe + abd + aed + bed) x + abed
2 -
Hence
S 4 = (a
+ b + e -j- d) S
= a
+ b + e + d.
10. Put as before

am
Sm= (a-b) (a-c)
bm
(b-a) (b-c)
cm
(c-a) (c-b)
Let us take the first term of our sum am and transform it

am (a+b) (a+c) _
(a-b) (a-c) -
(a+b+c) am+l+am-1.abc
(a-b) (a-c)
Making use of a circular permutation, we get similar

expressions for the second and third terms of am. Adding
now all these terms, we find: am = (a
b
e) Sm+t
abe S m-t- Hence (after some transformations)
+ +
a1
(a
+ b + e)
S2
+ abe S
+b+e
(s = 1, s = 0)'
(a + b + e) 2 ,
S =a+ b + e, S =0,
0 2
= (a
+ Q + e)
S3
+ abe S 1 =
since
abe S 2 =
+ b + e) S 4 +
= (a + b + e) (a + b + e + ab + ae + be) + abe,
= (a + b + e) S + abe S
=
= (a + b + e) [(a + b + e) (a + b + e
+ 2abel.
a 3 = (a
a4
2)
14~
Solutions
11. Transform the left member of our identity in the

following way
abc { (a-a) (a-~) (a-y) + (b-a) (/J-~) (b-y)
(b-0) (b-a) (b-c)
\ll-0) (a-b) (a-c)
\C-a) Ir-~) (c-y)

(c-0) (c-a) (c-b)
(0-a) (0-~) (0-y) (0-c) (0-a) (c-/J)
a~y}
abc
Consider the first four terms of the sum in braces. Expanding the numerator of the first term in powers of a, we get
a3
(a
+ ~ + y) a2 + (a~ + ay +
~y) a -
a~y.
Performing an analogous operation with the remaining

three terms and adding them, we find that the sum of the
first four terms is equal to
83
(a
+ ~ + y)
82
+ (a~+ ay + ~y) 8 1 -
a~y8o,
where 8 k is the known sum (see Problem 9, where it is necessary to put d = 0). Proceeding from the results of this
problem, we find that the sum of the first four terms under
consideration is equal to unity, and, consequently, the
sought-for expression takes the form
abc { 1-
~~;
=abc-a~y.
12. Consider the following sum:

a,4
8 4 = (a-~) (a-y) (a-6)
y4
(y-a)
(y-~) (y-6)
From Problem 9 we have: 8 4

= abc, ~ = abd, y = acd, 6
a,4
(a-~)
~4
+ (~-a) (~-y) (~-II) +

64
+ (6-a)(b-~) (6-y)
=a+ ~ + y + 6. Put a
= bed. Then
a4b4c4
(a-y) (a-6) -- (abc-abd) (abc-acd) (abc-bcd) (c-d) (b-d) (a-d)
Using a circular permutation, we get analogous expression

for the remaining three terms. Thus, the given identity
is proved.
i43
Solutions to Sec. 2
13. 1 Transform one of the terms in the following way:

1
-----------
a(a-b) (a-c)
({-+)(+-+)
Then the required sum is equal to

1
(+)2
llbC
+-+) (+-
(})2
~) + -(-:-!-----:-!-)-(
----:-!-) +
-:--!
(+)2
(+-+)
(+-+)
}--1
s
abc
But (see Problem 8) S 2 = 1, and, hence, we get:

1
a (a-b) (a-c)
+ b (b-c) (b-a) + c (c-a) (c-b)
1
abc
However, this result can be obtained in a somewhat different way. Let us consider the four quantities: a, b, c and 0,
and form S 0 for them.
We then have
s 0 -
1
a (a-b) (a-c)
+ c (c-a) (c-b) +
1
+ -(.,,..0--a)-(-=-0---b-)
_,(0---c-) = O'
b (b-a) (b-c)
since S 0 = 0. Hence we get the previous result.

2 Likewise the sum can be transformed as
144
Solutions
And so
1
i
a2 (a-b) (a-c)
+ b2 (b-a) (b-c) + c2 (c-a) (c-b)

ab+ac+bc
a2b2c2.
A similar method can be applied when computing other

sums of the form
___
1 __
ah (a-b) (a--c)
bh (b-a) (b-c)
1
ch (c-a) (c-b)
14. We have
ah
(a-b) (a-c) (a-x)
bh
(b-a) (b-c) (b-x)
h
xh
c
(c-a) (c-b) (c-x)
(x-a) (x-b) (x-c)
=0
at k = 1 and at k = 2 (Problem 9).

Hence
ah
(a-b) (a-c) (x-a)
(b-a) (b-c) (x-b)
ck
(c-a) (c-b) (x-c)
(k=1,2).
(x-a) (x-b) (x-c)
15. We have
b+c+d
(b-a) (c-a) (d-a) (x-a)
=(a+b+c+d-x)
(a+b+c+ d-x)+(x-a)
(b-a) (c-a) (d-a) (x-a)
1
(b-a)(c-a)(d-a)(x-a)
1
(b-a) (c-a) (d-a)
Applying a circular permutation to the letters a, b, c, d

and adding the expressions thus obtained, we find that
the sum in the left member is equal 1to
(a
+ b -f-- c + d -
i
x) { (a-b) (a-c) (a-d) (a-x)
1
+ ~~-~~~~~~-+
(h-a) (b-c) (h-d) (b-x)
'
1 '
(c-a) (c-b) (c-d) (c-x)
1
}
(d-a) (d--b) (d-c) (d--x)
Solutions to Sec. 2
since the second sum equals zero.

It remains only to make sure that
1
(a-b) (a-c) (a-d) (a-.r)
.
1
(c-a) (c--b) (c-d) (c-x)
1
(b-a) (b-c) (b-d) (b-.r)
f-
..L
1
1
..L
(d-a) (d-b) (d-c) (d-.r) '
1
(x-a) (x-b) (x-c) (x-d)
=0.
It is possible to reduce these fractions to a common denominator and, on performing necessary transformations in
the numerator, to obtain zero. But we can, however, proceed
in a different way.
Multiplying the left member by (a - x) (b - x) (c - x) X
X (d - x), we get
1
(a-b)(a-c)(a-d)
+
+
(b-x)(c-x) (d-x)+
1
(b-a)(b-c)(b-d)
1
(c-a)(c-b)(c-d)
(a-x)(c-x)(d-x)+
(a-x)(b-x)(d-x)+
+ (d-a)(d-b)(d-c) (a-x)(b-x)(c-x)+1.
It is obvious that we deal with a third-degree polynomial
in x. It is required to prove that it is identically equal to
zero. For this purpose it is sufficient to show (see the beginning of the section) that it becomes zero at four different
particular values of x. Replacing x successively by a, b,
c, d, we make sure that our polynomial vanishes at these
four values of x, and, consequently, it is identically equal
to zero.
16. Transposing x 2 to the left, we get there a seconddegree trinornial in x. To prove that it identically equals
zero it suffices to show that it becomes zero at three different values of x. Putting x = a, b, c, we make sure that
the identity is valid.
17. Solved analogously to the preceding problem. However, Problem 16, as well as this one, can be solved by making
use of the quantities S k (see Problem 8 and the following
Olll'S).
146
Solutions
18. Put
a-b
-c-=X,
b-c
-a-=y,
c-a
-b-=z.
The left member of our equality takes the form

(x+y+z)
(.!.+.!.+
.!.)
=:3 +
.r
y
z
+ .r+z
+ x+y
y
z
y+z
:r
Consider the fraction y+z . We have

x
y+z
= (
:i:
b-c . c-a) -c-=_c_. b2-bc-l-aca T

b
a -- b
a- b
ab
a2
==
= - c - . b2-a2--c(b-a) =~ (-a-b+c) =
a-b
ab
ab
c
=ab'(-a-b-c+2c)
2~
=a;;-,
since a+ b + c = 0. Using a circular permutation, we find

y+z
x
+ x+z
+ x+y
=
y
z
2c 2
ab
_i_
1
2a 2
be
+ 2bac2
=_?__ (a3+b3+c3).
abc
But if a + b
c = 0, then a 3 + b3 + c3 = 3abc (see
Problem 23, Sec. 1). Consequently
y+ z
x
x+ z
y
+ x+
y ==
'
and the equality is solved.

19. Miltiplying the given expression by (a + b) (b + c) x
x (c + a), we get (a - b) (a + c) (b + c) + (a + c) X
X (a + b) (b - c) + (a + b) (c - a) (b + c) +
+ (a - b) (c - a) (b - c).
This expression is a second-degree trinomial in a which
becomes zero at a = b, a = c and a = 0 and, consequently,
is identically equal to zero, i.e.
a-b -!.... b-c
c-a
(a-b) (b-c) (c-a) -O
a+b ' b+c
+ c+a
+(a+b)(b+c)(c+a)-
We assume here b -=I= c. If b = c, then it is easy to make

sure directly that the identity holds true.
20. We have
b-c
lb-a)+(a-c)
1
(a-b) (a-c) = (a-h) (a-c) = a-b -
1
a-c
147
Solutions to Sec. 2
Treating the remaining two terms in a similar way, we

arrive at the proposed identity.
21. Answer. 0. Solved analogously to Problem 19.
22. It is required to prove that
dm(a-b) (b-c)+bm (a-d) (c-d) _ b-d = O.
cm (a-b) (a-d)--j-am (b-c) (c-d)
a-c
Reducing to a common denominator, Jet us prove that

the numerator equals zero. However, if the numerator
is divided by the product (a - b) (a - c) (a - d) (b - c) (b -d) x (c - d), we get the following expression
am
(a-b) (a-c) (a-d)
bm
+ (b-a) (b-c) (b-d) +

cm
dna
+ (c-a) (c-b) (c-d) + (d-a) (d-o)(d-c)
At m = 1, 2 this expression is equal to zero (see Problem 9).

23. Let us first prove that
1 _....=_+ x(x-a.t) _ x
a.1
a.1 a.2
(x-a.1) (x-a.2)
a.1 a.2a.3
+ ... +
+ (-1t x (x-a.1) (x-a.2) ... (x-a.n-1)
a.1a.2
= ( -1)n
CX.n
(x-a.1) (x-a.a) ... (x-a.n) .

a.1a.2 CX.n
()
Likewise, it is evident that the second bracketed expression

is equal to
And the product of the bracketed expressions yields
( _ 1t
(x2-a.~) (x2-a.~) ... (x2-a.~)

a.ia.~
...
a.~
Replacing here x by x 2 and a.; by a.~ and applying tho

equality ()in a reverse order, we get the required identity.
24. Given
( b2+;:c-a2
1)
+ (cz+;:c- b2 _
+(
a2--j-b2-c2
2ab
1)
+1
+
=O.
148
Solutions
The first bracketed expression is equal to

(b-c)2-a2
2bc
(b-c-a)(b-c--j a)
2bc
the second to
(a-c)2-b2
(a-c-b)(a-c-\-b)
---=-2a_c_ _ =
_ __
---2~a-c
Likewise, the third one takes the form

(a+b)2-c2
2ab
(a+b+c) (a-\-b-c)
2ab
Consider the sum of these expressions

(a+b-c)(a+c-b)
2bc
(a+b-c) (c+b-a)
2ac
+(a+b-c)2ab(a+b+c) -_
=
a+b-c
2abc
{c(a+b-t c)-b(c+b-a)-a(a+c-b )} =
(a+b-c) (c+a-b) (c-a-\-b)
2abc
Thus, we are given that

(a+b-c) (a+c-b) (c+b-a) -0
2abc
'
wherefrom follows that at least one of the factors in the

numerator equals zero. Suppose a
b - c = O; then all
the three bracketed expressions in the equality () are equal
to zero, and, consequently two of the given fractions are
equal to 1, while the third one to - 1. The remaining
two possibilities yield the same result.
25. Reducing the original equality to a common denominator and cancelling it out, we get (after some transformations)
(a+ b) (a+ c) (b+c) =.0.
(1)
But the second equality (which is to be proved) can

also be reduced to the form
(an+bn)(an+cn)(bn+c")=O.
(2)
149
Solutions to Sec. 2
It is quite obvious, that with an odd n' equality (2)

follows from (f), since if, for instance, a+b=O, then
a= -band an+bn=a"+(-a)n=an-an=O.
26. Rewrite the given proportion in the following way
(bz+cy) yz
-ax-j- by+cz =
(cx+az) xz
ax-by+cz
(ay+bx) xy
ax+by-cz '
A
C
E
But from the proportion B=n= y follows
-
C+E
D+F
C
75
A \-E
B+F
A+C
B-'-D ' A
it is easy to check, putting
-B =
= y= '),, and expressing A, C and E in terms of
A,,
F).
B, D,
Therefore we have
c (x2+ y2)+z (ax+by)
c
a (z2+ y2)
a
b (x2 j-z2)+Y (cz-f-ax)
b
Subtractingx
we get
+y +z
z (ax+ by-cz)
c
+ x (by +cz)
from each term of this equality,
x (by+ cz-ax)
a
y (cz+ax-by)
b
Take the original equalities

ay+bx
z (ax+by-cz)
cx+az
y (ax-by+cz)
bz+cy
x (-ax+by+cz)

Hence
ay+bx
bz+cy
cx+az
c = (ay + bx) ,
b = (ex + az) ,
a = (bz + cy) .
Multiplying the first of these equalities by c, the second

by b and the third by a, and forming the expression b2 +
+ c2 - a2 , we find b2 c2 - a2 = 2bcx.
Analogously, we get
c2 + a2 - b2 = 2cay, a2 + b2 - c2 = 2abz.
Solutiuns
150
Hence, finally
x
a(b2+c2-a2) =
y
z
b(a2+c2-b2) = c(a2+b2-c2)
27. Since a+ b + c = 0, we may write

(a + b + c) (aex + bp + cy) = 0.
Expanding the expression in the left member, we find
p)
a2 ex + b2 p + c2y + ab (ex +
+ ac (ex + y) +
cb (p
+ y)
0.
But ex + P = -y, ex + y = -P, ~ + y = -ex, therefore

a2ex + b2 p + c2 y - aby - acp - cbex = 0, or a2 ex + b2 p +
+c 2 y-abc(..::_-t_1.
+i)
= 0 and since_::_+ 1. + J'.. = 0
a'b
c
'
ab
c
(by hypothesis), we have: a 2 ex + b2 p + c2 y = O.
28. From the equalities
(b 2 + c2
a 2) x = (c 2 + a2
b2) y = (a 2 + b2
c2) z
follows
y
c2+a2-b2
a2+b2-c2
1
b2+c2-a2
Put for brevity
b2 + c2
a2 = A , c2 + a2
b2 = B, a2 + b2
c2 = C.
It is evident that our problem is equivalent to the following one: if the equation x3 + y3 + z3 = (x + y) (x + z) X
X (y + z) has the solution
x = a, y = b, z = c,
then it also has the following solution

1
x=A,
Y=n
z=-c
We know the following identity (see Problem 19, Sec. 1).

(x + y + z) 3
x3
y3
z3 = 3 (x + y) (x + z) (y + z).
151
Solutions to Sec. 2
Using this identity, we can easily prove that the equalities
+ y) (x + z) (y + z),
(x + y + z) = 4 (x3 + y + z =
= 4 (x + y) (x + z) (y + z),
(x + y - z) (x + z - y) (y + z - x) = -4xyz
x3
+ ya + z
= (x
(1)
3)
(2)
(3)
are equivalent, and the existence of any of them involves

the existence of the remaining ones. Thus, it is sufficient
to prove that
i.e. that
(AB +AC
But
A
+ BC)
+B
= 2c 2 ,
4 (A
A
+ B)
+C
(A
+ C) (B + C) ABC.
= 2b 2 , B
+C
= 2a 2
Therefore we have to prove

(AB +AC
+ BC)
= 32a 2 b2c2 ABC.
Let us first compute AB +AC+ BC, and then ABC.

We have
AB+ AC+ BC= A (B + C) +BC=
= (b 2 + c2 - a 2) 2a 2 + [a 2 + (b 2 - c2)] X
X [a 2 - (b 2 - c2)] = 2a2b2 + 2a 2c2 - 2a4 +
+ a4 - b4 - c4 + 2b2c2 = -a4 - b4 - c4 +
+ 2a2b2 + 2a2c2 + 2b 2c2 = 4a2b2 - (a 2 + b2 - c2 ) 2 =
= (a - b + c) (-a+ b + c) (a+ b - c) (a+ b + c).
By virtue of equality (3)
(a + c - b) (b + c - a) (a
+b-
c) = -4abc.
Therefore
AB+ AC+ BC= -4abc (a+ b
+ c).
Solutions
Compute ABC. Pnt
a2 + b2 + c2 = s,
then
ABC = (s - 2a 2) (s - 2b 2) (s - 2c2) =
= s3 - 2 (a2 + b2 + c2) s2 + 4 (a2b2 + a2c2 + b2c2) s - 8a 2b2c2 = 4 (a 2 b2 + a 2c2 + b2c2) s - s3 - 8a 2b2c2 =
= s {4a 2b2 + 4a2c2 + 4b 2c2 - (a 2 + b2 + c2) 2} -8a 2b2c2 = - s {a4 + b4 + c4 - 2a2b2 - 2a2c2 - 2b 2c2 } - 8a 2 b2c2 = s (a -1- c - b) (b + c - a) X
+ b + c) - 8a b c =
-4abc (a + b + c) (a 2 + b + c2) - 8a 2b2c2 =
-4abc {a + b + c + a (b + c) + b (a + c) +
X (a
=
=
But
(a + b) (a
-1- b - c) (a
2 2 2
+ c)
(b
+ c)
= a 2 (b
b2 (a
c2
(a+ b)
2abc}.
+ c) +
+ c) + c2 (a + b) + 2abc.
Therefore, by virtue of equality (1), the bracketed expression is equal to 2 (a 3 + b3 + c3 ).

But, by virtue of equality (2),
2 (a3
b3
+c =
3)
+ b + c)
+ b + c)
2 (a
3
Therefore ABC = -2abc (a
But, as has been deduced, AB +AC +BC =
= - 4abc (a + b + c).
Therefore,
(AB +AC + BC) 3 = 32a 2b2c2 ABC.
29. 1 We have:
Pn = anPn-i -I- Pn_ 2, Pn - P"_ 2 = anPn_ 1 ,

Qn = anQn-1 + Qn-2, Qn - Qn-2 = anOn-1
The left member of the equality in question is transformed
by the following method
Pn+2-Pn
Pn
Pn+1-Pn-1
p
n+I
Pn+I
Pn
lln+2 - p lln+l-p
n
n+I
llnt2lln+1
153
Solutions to Sec. 2
We get quite analogously that the right member also

yields an+t an+ 2. Thus, the identity is proved.
2 We have
!!.!!___
Qk
P1,-1 _ PhQ11-1-QhPk-1
Qk01i-1
011-1 -
(-1)h-1
QkOn-1
Putting here k = 1, 2, ... , n and adding termwise, we

obtain the required result.
3 We have
Pn+2Qn-2 -
Pn-2Qn+2
(an+2Pn+1
+ Pn) Qn-2 -
+ Qn) = a,,+z (Pn+1Qn-2 - Pn-20n+1) +

+ PnQn-2 - Pn-2Qn =
an+2 {(an+IPn + Pn-1) On-2 - P,,_2 (an+IQn + Qn-1)} +
+ (anPn-1 + Pn-2) Qn-2 - Pn-2 (anQn-1+Qn-2)
= an+1an+2 (PnQn-2 - Pn-2Qn) +
+ a11+2 (Pn-1Qn-2 - Pn-2Qn-1) +
Pn-2 (Qn-1-1an+2
Pn-2Qn-1)
an+1an+2 {(anPn-1
+ Pn-2) Qn-2 -
+an (Pn-IQn-2 Pn-2 (anQn-1
+ Qn-2)} + a
11
+2(-1)n +an (-1)n
+ an+2 +an) (-1)".

anPn_ 1 + Pn_ 2. Therefore
= (an+2an+1an
4 It is known
Pn
-F, =an
n-1
that Pn =
+ -p-= an +-p--=an
1
+
Pn-2
n-1
n-1
1
p
+P
lln-1 n-2
n-3
Pn-2
=an+
Pn-2
an-I+ PPn-3
=an+--+
lln-1
..
+--1__
P0
n-2
a2+-p-,
1
=an+--+
lln-1
..
+ __1_1_
a1+-
ao
The expression for
QQn
n-1
is found in a similar way.
154
Solutions
30. On the basis of the results of the preceding problem

we have
~~ 1 =(an, lln-i.
= ('lo,
... , ao)
~:
an)=
llz, ,
Consequently, Pn-1 = Qn.

31 . We have to prove that
~
Pn+1-Pn-1Pn+1 = PnPn+2-Pn,
o.r
But
PnH = aPn
Pn-t.
aPn,
Pn+z ---= aPn+t
Pn.
Consequently,
Pn+t -
Pn-t
Pn+ 2
Pn
aPn+l
Hence, follows the validity of our identity.

32. By hypothesis
x=
Pn
-~-------
(a, b, ... , l, a, b, ... , l) Qn = (a, b, .. , l)
Or
1
X=-+
a
T+
.. +-+.!2!..
l
Qn .
Thus, x is obtained from ~: if l is replaced by l +
+ 2Qn
in this fraction. But

X=
2= lPn-t + Pn- 2 .
Qn
( l + PQnn) Pn-1+Pn-2
(l
Pn )Q
+ Q;;
n-1
lQn-1
p Q
n-2
+ Qn-2
-1
n n ,-
p p
n
Therefore
n-1
Q~+PnQn-1
33. It is obvious that at k = 0, 1 our formula holds true.

Assuming that it is valid at k = n - 1, let us prove that
it takes place also at k = n. And so, we assume
bo+~
bi+
=
a
.. +2=!.
bn-t
Pn-1
Qn-1
155
Solutions to Sec. 2
However, according to the rule for composing Pk and Qk,

we have
Pn-1 = bn-1Pn-2+an-1Pn-3
On-I
bn-10n-2+an-10n-3 '
where Pn-2 Pn_3, Qn-2 Qn-3 are independent of lln-1 and

bn-1
On the other hand, it is clear that the fraction
a1
o+bi + .
.. + an-I
bn-1
an
+bn
is obtained from the fraction

a1
o+bt +.
. + an-I
bn-1
by rep 1acing bn-1 bY b11-1

Therefore
bn-IP n-2 +an-IP n-3
an
+bn
( bn-1
+ i;} Pn-2+an-1Pn-3
( bn-1
+ :: } On-2+an-10n-3
an
+bn
P n-2
bn-10n-2+ an-IOn-3+
~:
On-2
_ bnPn-1 +anPn-2
- bnOn-1
anOn-2
34. Denoting the value of our fraction

P 1 = r,
P 2 = r (r
Q1 = r
1),
Q2 =
by
~:, we have
+ 1,
r +r+
2
1.
Using the method of induction, let us prove that

rn-1
Pn=r-r- 1
Pn
On
156
Solutions
At n = 1 these formulas are valid. Assuming their validity

at n = m, let us prove that they also take place at n =
= m 1.
We have
Pm+t = bm+1Pm + am+1Pm-1
In our case we find

rm-1
Pm+t=(r+1)r---;:-=T-r
rm-1_1
r- 1
=r
rm+i-1
r- 1
Analogously we obtain that

Qm+t =
rm+2_1
r-1
35. Put
_!__+_1_=
Ur
Ur+i
1
Ur+ Xr
Then we find
u2r
Xr=
u,+ Ur+i
Therefore
1
Further
where
Thus
_!_ + _.!_ + _.!_ = ___1_2_ _
Ut
U2
U3
UI
Uf
Ut
+ +
112
X2
Using the method of induction, we also get the general

formula.
Solutions to Sec. 2
157
36. Let us denote the fraction
by ~: , and put the fraction
. d to prove t h at
It is
. require
equa1 t o P~
Q~ .
whole positive n.
We have
Pn
On=
p~ f
Q~
or any
... ,
Pi
Qi =
c101
c1b1 '
P;
Q~
c1c2a1b 2
= c1c2 (b1b2
+ a2) '
We may put P1 = a1, 01 = b1, P 2 = a1b 2, 02 = b1b2 + az,

and then the following relations take place (see Problem 33)
Pr.+I
= bn+1Pn
+ an+tPn-1
Put
P; = c1a1,
P~
= c1c2a1 b2;
o; = C1b1, 0~ = C1C2 (b1b2 -f- llz)
Let us prove that for any n we then have
p~ = C1C2 CnPn,
0~ = C1C2 Cn0n
Let us prove this assertion using the method of inducLion.

i.e. assuming its validity for a subscript smaller than, 01
equal to, n, we shall prove the validity for the subscript
n + 1.
.
We have
P~+l = C11+1bn+1P~
0
+ c,,cn+tlln+1P;, _1,
Q~+l = C11 +1b,,+1Q;,-t c,,cn 1-1an+10~-1
158
Solutions
Hence (with the asumption)

P~+I
Cn+1bn+1C1C2 CnPn
+ CnCn+1an+1C1C2 Cn-!Pn-1 =
+ a +1Pn-1)
C1C2 Cn+I (bn+1Pn
11
C1C2 Cn+1Pnt1
Likewise prove that

Q~+I
C1C2 Cn+10n+I
Now it is easy to find that
37. 1 Put
1
2 cosx-2- cos
x-
1
2_c_o_s_x - .
-2cosx
We have
P1
zcos x.
0i=
Therefore we may put
P _ sin2x
1-
sin x '
sin x
Qi=sinx
Further
!2_
Q2
Zcos x--1-= 4cos2x-1.

2 cos x
2 cos x
Com;equently, we may take

p _sin 3x
z- 5in x '
Q =sin 2.r
2
sin x
sin(n+i)x
sinnx
.
Let us prove that then Pn =
. X , Qn = -Slnx
.for any n.
Sin
Assuming that these formulas are valid for subscripts
not exceeding n, let us prove that they also take place at
n + 1. We have (see Problem 33)
sin (n + 1) x sin nx
1 .
Pn+t = 2 cos x
. x - .= - . - Slll (n+ 2) x.
sm
srn
x
sm x
159
Solutions to Sec. 2
In the same way we fmd that On+!=

fore
Pn
o;:=
sin (11+1) x
.
, and theresa1 .c
sin(n+i)x
sin nx
for any whole positive n.

2 Let us denote the continued fraction on the right by
Pn . W e have to prove that
([;;
~:
=1+b2+b2b3+ .. . +b2b3 ... b,,.
We have
Therefore we may take: P 1 = 1, Q1 = 1, P 2 = b2 + 1,

Q2 = 1. Then, using the method of induction, it is easy
to prove that
= 1 + b2 + b2b3 + ... + b2b3 ... bn,

On = 1.l
Pn
and, consequently, our equality is also true.

38. 1 We have
sin a+ sinb+ sin c =sin (a+b+c) =
=(sin a+ sin b) +[sin c- sin (a+ b + c)j =
= 2 sin
a+b cos a-b

2
2
. -a+b
= 2 Sin
2-
-2 sin a-f-b cos a+b+ 2c -
( COS a-b
- 2- -
COS
a+b+2c}
2
. a+b . a +c . b+c}
= 4 SID
-2- SID -2- sm -2- .
2 Analogous to the preceding one.

39. Consider the sum
tan a + tan b
+ tan c.
Solutions
160
We have
sin(a+b) +sine=
tan a+ tan b +tan c = cos a cos b
cos c
_sin (a-t b) cos c+sin c cos acosb _
co~ a cos b cos c
_ sin(a+b) cos c+cos(a+b) sin c-cos (a+b) sin c+sin c cos a cos b _
cos a cos b cos c
_sin (a+ b+c)+ sin c [cos a cos b-cos (a+b)] _
cos a cos b cos c
sin (a+ b+ c)+sin a sin b sin c
cos a cos b cos c
Hence follows the required equality.
40. The equalities 1, 2 and 3 are easily obtained from
Problems 38 (1, 2) and 39 putting a= A, b = B, c = C
and a + b + c = A + B + C = n.
Now let us prove 4. Rewrite the left member in the following way
C( tan 2A +tan 2B) .
S=tan 2A tan 2B +tan 2

But since
we have
A+B
tan 2C =tan ( 2:n: -A+B)
2 - =cot2-=
A+B (
tan-2
Hence
A
S = tan 2 tan
2+
tan 2 -t tan 2
A+B
tan-2-
since
A+B
tan-2-=
tan 2 +tan 2
A
B.
1-tan- tan 2
=i,
Solutions to Sec. 2
161
5 lndeed
sin 2A + sin 2B + sin 2C =
= sin 2A + 2 sin (B + C) cos (B - C)
= 2 sin A cos A + 2 sin A cos (B - C)
= 2 sin A [cos A + cos (B - C)] =
= 4 sin A sin B sin C.
41. 1 It is necessary to find how a, b, and c are related if
. a b c O
cos a+ cos b+ cos c- 1 - 4.sm 2 sm 2 sm 2 = .
To this end let us reduce the left member of the equality
to a form convenient for taking logs, i.e. try to represent
it in the form of a product of trigonometric functions of the
quantities a, b and c.
We have
.
a+b
a-b
cos a+ cos b= 2 cos2-=

2 -cos= 2
(cos 2 .!!:... cos 2 .!!_ _ sin 2 .!!:... sin2 .!!.. )

2
cos c -1
2 sin 2 ~
'
Therefore the left member takes the form

2b
2 cos 2a
2 cos 2 -
2sm
2a.
2b
srn
2sm
2c
2-
2-
. a . b . c
- 4 s111 2 sm 2 sm 2
= 2
rl cos2a2 cos 2b2 -
+ sin2
~) J=
2 [
( . 2 a . 2b
sill 2 Sill 2
cos 2 ; cos 2
. b
+ 2.sm 2asm
2
(sin ; sin
a
b +.sm 2a sm
.
b)
= 2 [(
~ cos 2 cos 2
2
b)
a
a .
X [ ( cos-cos--sm-s1n-
. c
2 ( cos -a-b
2 -+sm 2
c
sm
2 +
rJ
. CJ
+sin
+ sm 2
c] =
.
-sm2
c) =
) ( cos -a+b
.
2 --sm 2
162
Solutions
= 2 [cos
a-;
+ cos ( ~ -
~ ) J[cos
at
. n-\-b+c-a . n+a+c-b
= - 8 sill
srn
4
4
b-
~ ) J=
cos ( ~ -
. n+a+b-c . a+b-J--c-n
x srn
sm
4
4
By hypothesis, this expression must equal zero and, consequently, at least one of the factors must be equal to zero. But
from the equality sin a = 0 follow~ a = kn (where k is
any whole number). Therefore, among a, b and c, satisfying
the original relationship, there exists at least one of the
four relationships
a+ b
c = (4k
1) n, a+ b - c = (4k - 1) n,
a+ c - b = (4k - 1) n, b + c - a = (4k - 1) n.
2 We have (see Problem 30)

tan a+ tan b + tanc-tan a tan b tan c = sin (a+~+c) .
(,OS
a COS
COS
By virtue of our conditions

sin (a
+ b + c)
= 0 and a
+b+c=
kn.
3 Transform the original expression. We have

1 - cos 2 a - cos 2 b - cos 2 c + 2 cos a cos b cos c =
= 1 - cos 2 a - cos 2 b - (cos 2 c - 2 cos a cos b cos c +
cos 2 a cos 2 b) + cos 2 a cos 2 b = 1 - cos 2 a - cos 2 b - (cos c - cos a cos b) 2 + cos 2 a cos 2 b =
= (1 - cos2 a) (1 - cos2 b) - (cos c - cos a cos b) 2 =
= (sin a sin b - cos c + cos a cos b) x
x (sin a sin b cos c - cos a cos b) =
= [cos c - cos (a
b)] [cos (a - b) - cos cl =
_ 4 . a+b+c . a+b-c . a+c-b . c+b-a
Slil
Slil
z
Slil
Sill
.
Consequently, there exists at least one of the following

relations
a+ b
+c=
2kn,
42. Put
x = tan
a+ b - c = 2kn, a+ c - b = 2kn,
b + c - a= 2kn.
~,
y = tan
~,
z = tan
f.
163
Solutions to Sec. 2
Then
2x
2y
1 _ x2 = tan a,
1_
yz = tan ~,
2z
1_
zZ
= tan y,
and our problem takes the following form. Prove that

tan a+ tan~+ tan y =tan a tan
p tan y
i[
tan 2a. tan 2~ +tan 2a. tan 21' +tan 2~ tan 21' = 1.
Rewrite the last equality as
tan
(tan
+tan
f) - (1- tan
Dividing both members by 1- tan {tan
tan
n=
0.
~ , we get
tan~ tan ~ty -1 =0, tan~= cot~-;Y =tan ( ~ -~~Y).

Hence
~+~+Y _::._=kn
2
(if tal)gents are equal, the corresponding angles differ by

the multinle of n) and
a+P+v= (2k+ 1) n.
And '.'O the proposition is proved (see Problem 40, 3).
43. Put b =tan p, c =tan y, a= tan a. Then
b-c _ tan ~-tan y _tan (A- )
1-t-bc-1-t-tan~tanyPy,
and, hence, our equality is equivalent to the following one

tan(~ - v) + tan (v - a) + tan (a - ~) =
= tan (p - v) tan (v - a) tan (a - p).
Put
P- V = x, v - a = y, a - ~ = z.
Let us finally prove that
tan x + tan y + tan z = tan x tan y tan z
164
Solutions
if
+y+z=
0.
But then we have

tan (x+ y) =-tan z,
tanx+tany
1-tan xtan y
= -tanz.
Hence follows the required equality.

It is obvious, that the last two problems can be solved by
direct transformations of the considered algebraic expressions.
44. We have
t an
sin3a. sina.(3-4sin2a.)
t
3-4sin2a.
3a= -=
= an a .,---,-~
cos 3a. cos a. (1-4 sin2 a.)
1-4 sin2 a.
Divide both the numerator and denominator of this

1 - by 1 + tan 2 a.
fraction by cos 2 a and replace - cos 2 a.
We get
t an 3a = t an a 3-tan2a. = t an a -V3-t tan et. -V3-tana. .

2
1-3 tan a.
1- -V3 tan a. 1 + -V3 tan a.
Hence
tan 3a = tan a tan ( ~ t a ) tan ( ~ -
).
45. Multiplying both members of the equality by a + b

and replacing unity in the right member by (sin 2 a +
+ cos 2 a) 2 , we get
sin 4 a+cos 4 a+
~ sin4 a+: cos 4 a=

= sin 4 a + cos 4 a+ 2 sin 2 a
whence
~ sin4 a-2 sin 2 acos 2 a+: cos 4 a=0,
(V ~ sin a - V : cos a)
2
-asrn a=bcos a,
= 0,
co~''r.l,
f65
Solutions to Sec. 2
or
sin4 a
----a2 =
cos4 a
/ j 2 = A,.
Substituting it into the original equality, we find

1
'A= (a+b)2
Therefore
46. From the second equality we have

(a1 COS at
-
+ a2 COS a2 + . . . + an COS an) COS 8 + a2 sin a2 + . . . + an Sin an) Sin 8 = 0.
(a1 sin a1
On the basis of the first equality and since sin 8
a 1 sin a 1
+ a 2 sin a 2 + ... + an sin an
=I= 0, we get
= 0.
(*)
Multiplying the first equality by cos A. and the equality (*)

by sin A., and subtracting the second result from the first
one, we have
a 1 cos (a 1
+ A.) + a
cos (a 2
+ A.) + ... +
+ an cos (an + A.)
= 0.
47. It is obvious that the left member is reduced to the

following expression
tan v)
(tan
48. 1 We have
(tan ~ -
v-
tan a)
+ (tan a
Hence
____i::_ =
ra-r
ThPrPforo
ap (p-a)
tan ~)
= 0.
Solutions
166
But
s2 = p(p-a) (p-b) (p-c).

Hence
Cil=S
a
b
c
}
(p-b)(p-c)+(p-a)(p-c)+(p-a)(p-b) =
=s{(p-b)+(p-c)+ (p-a)+(p-c)
(p-b) (p-c)
(p-a) (p-c)
(p-a)+(p-b)}
(p-a) (p-b)
=2
(ra+rb+rc)
2 We have
a2r
n
-(a-b) (a-c)
(j-
=S
+ (b-c)b2rb(b-a) + (c-a)c2r(c-b)
c
b2
a2
~-aj~-~~-rj+~-~0-rj0-aj+
c2
+ (p-c) (c-a) (c-b)
But (see Problem 9)

a2
b2
+ (p-b) (b-c) (b-a) +

c2
p2
+ (p-c) (c-a) (c-b) = (p-a) (p-b) (p-c)"
(p-a) (a-b) (a-c)
Therefore
O'=
sp2
sp3
p3
p2
=-=-=(p-a)(p-b)(p-c)
s2
s
r
3 We get
(
ra+rb+rc=S
1 )
p-a+p-b+p-c
s(ab+ac+bc-p2)
=(p-a)(p-b)(p-c)"
Further
i:+~+-c
=!{a (p-a) +b (p-b) + c (p-c)} =
ra
rb
re
s
1
= -
(2 p2 - a2 - b2 - c2)
"$ ( - p 2 + ab
+ ac + be).
167
Solutions to Sec. 2

4 Consider the first sum
1 { bc(p-a)2
a= S2
__
1{ P
s2
2 [
ac(p-b)2
, ab(p-c)2}
+ (b-c) (b-a) + (c-a) (c-b) =

be
ac
+(c-a)a(c-b)
b](a-b) (a-c) + (b-c) (b-a)
(a-b) (a-c)
b [
pa c
_(a-b)(a-c)+(b-c)(b-a)+(c-a)(c-b)
]+
+ abc [ (a-b)a(a-c) + (b- c)b(b-a) + -(c-a)c(c-b) J}.

But (see Problem 8)
(a-b)(a-c)
+ (b-c)(b-a)
1
+ (c-a)(c-b)-'
1
-0
a
(a-b) (a-c)
+ (b-c) (b-a) + (c-a) (c-b) -
-0
Therefore
p2 [
a= S2
be
(a-b) (a-c)
ac
ab
+ (b-c) (b-a) + (c-a) )c-b)
J;
further
be
(a-'b) (a-c)
ac
ab
+ (b-c) (b-a) + (c-a) (c-b)

= abc {[a (a-b~ (a-c) + b (b-c~ (b-a) + c (c-a~ (c-b) +
_L
1
(0-a) (0-b) (0-c)
]+-1
}=1
abc
And so
p2
0=92=1=2.
Let us go over lo the sPcond sum. We have
rf=-1- {
a2ra
rarbrc (a-b) (a-c)
c2rc
+ (b-c)b2rb(b-a) +
{
a2
+ (c-a) (c-li) = rarbrc (a-b) (a-c) (p-a) +

~
~
}
+ (b-c) (b-a)
(p-b) + (c-a) (c-b) (p-c)
168
Solutions
But
b2
a2
(a-b) (a-c) (a-p)
+ (b-c) (b-a) (b- p) +

2
+ (c-a)(c--b)(c-p)
c~
+ (p-a)(p-b)(p-c)p
-0
'
n
Therefore
p2
p2
1
'(p-a)(p-b)(p-c) =52=-,:2
s(p-a)(p-b)(p-c)
a=
s3
5 We have
O=
a~
(a-b) (a-c)
=S
+ (b-c)b~(b-a) + (c-a)~(c--b)
a
(a-b) (a-c) (p-a)
+ (b-c) (b-a)(p-b) +
+ (c-a) (c~b) (p-c)} = - s { (a-b) (a_:_c) (a-p) +

b
c
+ (b-c)(b-a)(b-p) + (c-a)(c-b)(c-p) +
p
p
}
+ (p-a) (p-b)
(p-c)
(p-a) (p-b) (p-c) =-=
sp
p2
p
= (p-a) (p-b) (p-c) = s =-;'
Further
O=
(b+c)ra
(a-b) (a-c)
(c+a)rb
(a+b)rc
+ (b-c) (b-a) + (c-a) (c-b)

(b+cl
(c+a)
{
=S (a-b) (a-c) (p-a) + (b-c) (b-a) (p-b) +
(a-i b)
}
+(c-a)(c-b)(p-c)
=Sa+ +c)
1
(a-b)(a-c)(p-a)
1
1
}
+ (b-c) (b-a)
(p-b) + (c-a) (t:-b) (p-c)
-s {
+ (b-c) (b-a)
b
+
(a--h) (11- r) (!--a)
(p-b)
(I
+ (c-a) (c~b) (p-c)
169
Solutions to Sec. 2
But
1
(a-b)(a-c)(a-p)
+ (b-c)(b-a)(b-p) +
1
+ (p-a)(p-b)(p-c)1
-0
+ (c-a)(c-b)(c-p)
Therefore,
the
first braced expression is equal to

( p-a ) ( p- b) ( p-c ) The second braced expression is equal
1
to ~2
Hence
s(a+h-t-c)
p2
2p2
p2
p2
p
CJ=(p-a)(p-b)(p-c)-9=-s---s=-;=-;
49. Rewrite the supposed identity in the following way:

sin (a
+b-
c - fl) sin (a - b) =
= sin (a - c) sin (a - fl) - sin (b - c) sin (b - fl).
1
2 {cos (A - B) -
Using the formula sin A sin B =

B) }, we find
- cos (A
sin(a+b-c-d) sin (a-b) =
i
= 2 {cos (2b-c-d)-cos (2a-c-d)},
=
~ {cos (c-d)-cos (2a-c-d)},
sin (b-c) sin (b-d) =
~ {cos (c-d)-cos (2b-c-d)}.
sin (a-c) sin (a-d)
50. 1 We have: 1 + tan 2 ~ =~ = 1 +~os 8 =
_b+_P_c ,
cos22
where a+b+c=2p.
Hence
1 + tan 2
~ + 1 + tan 2 ~ + 1 + tan 2 ~ =
=
(b+c)+(a+c)+(a+b)
p
= 4,
Solutions
170
and, consequently, tan 2

8
2
2 tan 2
t
b+ c -1
p
et
<p t
an 2 an 2
-}
+ tan ~
2
+ tan 2
1.
p-a. Therefore
p
ljl - .. /(p-a) (p-b) (p-c)
an T -
P3
But, as is known
(p-c)
t an A t an B t an 2C -_-.V/(p-a) (p-b)
P3
2
2
ljl_
<p
Hence, tan 2 tan 2 tan 2 - tan 2 tan 2 tan 2 .
51. The left member of our equality can be rewritten as

1
sin (a-b) sin (a-c) sin (b-c) {sin (b-c)-sin (a-c) +
+sin (a-b)}.
But we have
. (b -c) -sm
. (a-c ) = 2 smb+a-2c
. b-a
sin
2 -cos
2
Therefore, the braced expression is equal to
. b-a
2 sm
- 2 -cos
b+a-2c
2
. b-a
b-a
2 sm2 -cos - 2 - =
. b-a . b-c . c-a

4 sin
- - sm - - sin - 2
But
sin (a-b) sin (a-c) sin (b-c) =
. a-b .
= 8 sm -
2-
a-c . b-c
sm -
2-
sm -
2-
a-b
cos -
2-
a-c
cos -
2-
b-c
cos -
2-

52. 1 The fraction in the left member has the form
.
Sill
.
(a- b) Sill
. (1a - c) Sill
. (b - c) { sm
+sin bsin (c-a)

=
Sill
. (b - c) +
a sm
+ f1in c sin (a- b)} =
. 1
.
(a-b) srn (a-c) sm (b-c)
~ sm
. a sm
. (b - c)
'
171
Solutions to Sec. 2
where summing is applied to all the expressions obtained

from the one under the summation sign by means of a
circular permutation. But
sinasin(b-c)=
~ [cos(a-b+c)-cos(a+b-c)].
ThereforP we have
h sin a sin (b-c) =
{{cos (a+ c-b)-r,os (a+b-c) +
+cos (b+ a-c)-cos (b+ c-a) +cos (c+ b-a)--cos (c +a-'.J)} = 0,

and our identity holds true.
2 The given identity can be proved similarly to case 1. But
we can get the same formula immediately from formula 1,
replacing a by ~ - a, b by ~ - b, and, finally, c by ~ - c.
53. 1 We have to prove that

sin a sin (b - c) X
x cos (b + c - a) = 0. Here summation is applied to all
the expressions obtained from the original one by means of
a circular permutation. But
sin a sin (b-c) ={{cos (a-b + c)-cos (a+ b-c)}.
Therefore
~sin a sin (b-c) cos (b + c-a) = {- h cos (b + c-a)

xcos (a-b+c}-
h cos(a+b-c)cos(b+c-a)=
= ~ ~ [cos 2c+ cos (2b-2a)-cos 2b-cos (2c- 2a)] =
= 41 { cos 2c - cos 2b + cos 2a - cos 2c + cos 2b - cos 2a +cos (2b- 2a)- cos (2c- 2a) +cos (2c - 2b)- cos (2a-2b) +cos (2a-2c)-cos (2b-2c)} = O.
:rt
2 Can be obtained from 1 by replacing a by 2 -a, b

n
n
by 2 -b and c by 2 -c.
Solutions
172
3 Likewise we find
L; sin a sin (b -
c) sin (b + c - a)=
=-}{sin 2 (b-a) +sin 2 (c-b) +sin 2 (a-c)}.

It only remains to show that
-} {sin 2 (b-a) +sin 2 (c-b) +sin 2 (a--c)} =

=
2 sin (b--c) sin (c-a) sin (a-b).
4 Proved analogously to 3 or by replacing a by ~ -a,

:rt
:rt
b by 2 -b and c by 2 -c.
54. 1 We have
L; sin
A cos (B-C) =
= ~
L; sin
L; sin
A sin A co~ (B-C)
A{sin(A+B-C +sin(A-B+C)}.
But since A + B + C
L;
sin 2 A cos (B - C) =
L;
n, we have
~ ~ sin 2 A (sin
2C + sin 2B)
sin 2 A (sin B cos B + sin C cos C) =
= sin2 A sin
B cos B + sin 2 A sin C cos C +
+ sin 2 B sin C cos C

+ sin 2 C sin A cos A
= sin A sin B (sin A
+ sin A sin C (sin A
+ sin B sin C (sin B
= sin A sin B sin (A
+ sin B sin C sin (B
+ sin 2 B sin A cos A +

+ sin 2 C sin B cos B =
cos B + cos A sin B) +
cos C + cos A sin C) +
cos C + cos C sin C) =
+ B) + ~in A sin C sin (A
C) = 3 sin A sin B sin C.
+ C) +
173
Solutions to Sec. 2
2 We have
~ sin3 A sin (B-C) = ~ sin 2 A sin A sin (B-C) =
=Li sin
Asin(B+C)sin(B--C)=
= ~ ~ '3in 2 A {cos 2C-cos 2B}= ~ sin 2 A (sin 2 B-sin 2 C) =

=sin 2 A sin2 B sin2 C
X
:Li (si~2 C -
SiiiZC- sin2 B + sin2 A -
si~ 2 B) =sin 2 A sin2 B sin 2 C X

1
sin2 C
+ sin21 B- sin21A} =
O.
55. 1 We have
sin 3x = 3 sin x - 4 sin3 x.
Therefore
~ sin3Asin 3 (B-C)= ~ ~ sin3A{3sin(B-C)-sin3 (B-C)} =
! ~ sin3 (B +C) sin (B-C)-
-~ ~ sin3 (B+C) sin 3 (B-C) =
={ ~ {cos(2B+4C)-cos(4B+2C)}-
- ! ~ (cos 6C -
cos 6B) =
= 8 {cos 2 (B + 2C)-cos 2 (C + 2B) +cos 2 (C + 2A)- cos 2 (A + 2C) + cos 2 (A
+ 2B) -
cos 2 (B + 2A)} -
- ! {cosGC-cos 6B+cos GA-cos 6C + cos6B-co~GA}.

But
cos (2B
+ 4C)
cos (2B + 4A),

cos (2C + 4B) =cos (2C + 4A),
cos (2A + 4C) = cos (2A
=
4.R).
174
Solutions
And so, we finally have

~ sin 3A sin3 (B -
2 Since cos 3x
h sin 3A cos
C) = 0.
= 4 cos3 x - 3 cos x, we have
(B-C) =
= ~ ~ sin3(B+C){cos3(B-C)+3cos(B-C)}=
= {
~ sin3 (B+C) cos3 (B-C) +
~ ~ sin3(B+C)cos (B-C)=
~ ~
(sin6B+sin6C)+:
+sin (2B + 4C)} =
~ {sin(4B+-2C)+
! (sin 6A +sin 6B +sin 6C) =

= sin 3A sin 3B sin 3C.
1. The validity of the given identity can be checked, for
instance, by the following method. From the formulas ()
(see the beginning of the corresponding section in "Problems")
we get
V2+V3=v;+v~' V2-V3=v;-y~.
Therefore we have
(l1<~~+ v=f )
V2 +
V ~ +V ~ -
(1+ v:i)c.
vz
2 (3 -l- V 3)
.v2
(1+ V3) 2
2V3(1+V3)
1+ V3
V6
17!::>
Solutions to Sec. 3
Likewise we get
(V1--~)2
-V2-V1+v+
(1- 1/3)2. V2
2 (3- V3)
Consequently
2+V3
+
2--V3
) 2 =(1+~3+-V3:-1) 2 =
-V2+ V 2+ -V3 -V2-V 2- V3
-V6
VB
(\Ys3)2 =
2.
2. Let us prove the proposed identities by a direct check.

1 Put J/2 =a, i.e. a 3 = 2. It is required to prove that
+a
(1 - a
We have
(1 - a +
2) 2
=1+
a2
2) 3
= 9 (a - 1).
+a +
4
2a 2
2a
2a
= 3 (a 2 -
3 -
1),
since
a,3
Hence
(1 - a
+a
2) 3
= 3
= 3
(a 2
(a
-
2, a 4 = 2a.
1) (a 2 - 1) =
a + 1) (a + 1) (a - 1) =
= 3 (a 3 + 1) (a - 1) = 9 (a - 1).
2 -
2 We have to prove that
(Y2+Y20-Y2s) 2 =9 (Ys-Y4).
Squaring the left member, we find
Y4+Y4oo+Y.62s+2 J/40-2Yso-2 ysoo=

=Y4+2 yso+s Y5+4 Y5-2yso-10Y4=
=9 (J/5-Y4).
176
S oluttons
3 Proved as in the preceding case.

4
( V5+1 ) = 3+2
vs-1
Vs
3-2Vs
Put
v4/-5=a.
We have
4
( V5+1 ) = (a.+1)4 = 1+4a+6a2+4a3+a4 =
-f/5-1
(a-1)4 1-4a+6a2-4a3+a4
3 + 2a + 3a2 + 2a3
3-2a+3a2-2a3'
since a 4 = 5.
Further
1
4
( -f/5+1 ) _ 3+2a+a2 (3+2a) _ 3+2a _ 3+2V5
V5-1
-3-2a+a2(3-2a)--3-2a - 3-2V5
5 It is required to prove that
(1+Y3-Y9) 3 =5 (2-Y21),.
Put
v5/-3 =a,
.
i.e.
a5 = 3.
We have
(1 + a - a 2) 2 = 1 + a 2 + a 4 + 2a - 2a 2 - 2a3 =
= 1 + 2a - a 2 - 2a 3 +
a,4.
Further
(1 + a - a 2) 3 = 1 + 3a - 5a3
But
+ 3a
5 -
a 6
a 6 = 3a, a 5 = 3.
Therefore
(1+a-a 2) 3 =10-5a3
5 (2 - ~27).
J"'2
6 Put
=a and prove the first equality which can be
rewritten in the following form
5 (1 + a + a 3) 2
(1 + a 2) 5
177
Solutions to Sec. 3
The right member is equal to
+ 5a.. + 10a + 10a + 5a + a =

= 5 (1 + a + 2a + 2a + a
2
10
since
ato
8),
= 4.
Further
a 5 = 2,
a 6 = 2a,
a 8 = 2a3 ,
and, consequently,
(1
+a
2) 5
= 5 (1
+ a + 2a + 4a -1- 2a:i).
4
It only remains to prove that

(1
+a +a
= 1
3) 2
+ 4a + a + 2a + 2a
3
The last equality is readily proved by removing the brackets in the left member and performing simple transformations. To prove the second equality we have to show that
or
Put
/Cl=a
a 5 =2
'V ""
'
'
a 6 =2a
a 7 =2a 2
'
a 8 =2a3
'
Then we have to prove that

(a 4
+a +a
3
+a
2).
+ 2a
4 -
1) 2 = 5 (1
Expanding the left member, we find
+ a + a + a + 2a + 2a
2
2a 4
5 -
2a3
2a.
Making use of the equalities enabling us to replace high

powers of a by lower ones, we find the required identity.
3. Put
A
-=-=-=-=')...
a
b
c
d
Then
A = at...,
B = b'A,,
ct...,
dt....
178
Solutions
Consequently
Vita+ VBb+ Vcc+ VDd= V~(a+b+c+d).

But
A +B+C +D= 'A (a+b+c+d)

and
'A=
A+B+C+D
a -t b -t- c
+d
'
i.e.
VX= -V A+B+c+v.
Va+b+c+d
Replacing
VX.
in the equality
V Aa+ V Bb+ Vee+ V LJd = 11 X. (a+ b+c+d)

by the found value, we obtain the required identity.
4. Put for brevity
~ ax 2 + by2 + cz 2 = A.
We have
A=
3vax3
-+-+-= v ax.J.(-+-+x
by3
y
cz3 .
1 )
since
ax3 -= by3 = cz3
and - x
+-1y +-1z = 1 .
Likewise we find
3/~
A = Yv
and
3/-
A =zv c.
Hence
Adding these equalities termwise, we get
=xv3/ a- '
179
Solutions to Sec. 3
Hence, finally,
A=Ya+;;-E+y-c.
5. Put
Then
an=an+~n.
bn=an-~n.
where a~= 2 .
Prove that
We have
aman - a~~n =(am+ ~m) (an+ ~n)- am-nt~m-n =
= am+n + ~m+n + an~n (am-n + ~m-n)

am-n +~m-n
2n
But.
consequently,
_ ,.,m+n -t- pAm+n -_
am a n _ am-n
2n - v .
,.,
v.m+n
The second relation is proved in the same way.

6. Put
1+2vs =a,
Then
a
+~
1,
a~ =
-1.
Furthermore
a2
a - 1 = 0,
~2
1 = 0
180
Solutions
and
Un=~5(an-~n).
Proof. 1 We have
Un+ Un-1 = ~5 (an -
~n) + ~5 (an-1 _ ~n-1) =
= ~g {(an+ an-1)-(~n + ~n-1)}.

Multiplying both members of the equality a 2 -a-1 = O
by an- 1, we get
a+ 1 =a2, an+an-1=an+i.
Analogously, it is easy to conclude that
~n
+ ~n-1 =
~n+l.
Therefore
_ 1 ( n+l i:i.n+1) _
Un + Un-1-Vg a - p
- Un+1
2 We have
UkUn-k
+ Uk-1Un-k-t =
! {(a"-~k) (an-k_pn-k) + (a"-l_~h-1) (an-11.-1_pn-h-l)}

= ! {an+~n-ak~n-k_pkan-k+an-2+pn-2_pk-lan-k-1 _
_ pn-h-lah-1} =
! {an +an-2 + ~n+ ~n-2_ ~n ( ~: + ~::~)-an ( !: + !::~)} =

=
=..!_ {an+an-2+pn+pn-2_pn a 11 ~+a 11 -l -an ~"a+~k-1 }5
~h+l
=~{an +an-2+pn+pn-2_pnak-l (a~+ 1)

5
~11+1
=
! {an+ an-2 + ~n + ~n-2},
ak+l
an ~k-1(a~+1) } ak+l
-
181
Solutions to Sec. 3
since a~+ 1 = O. Then we perform the following transformations
! {an+an-2+~n+~n-2} = !{an-1 (a+ ! )+~n-1 ( ~+ ~)} =

= ! {an-1 (a-~)+ ~n-1 (~-a)}= a-;-~ (an-1 _ ~n-1) =
=
1
-V5
(an-1 _
~n-1) =Un+
3 Obtained from 2 by putting n = 2k, and then replacing k by n.

4 We have to show that
5 (a3n_ ~3n)-(an-~n)3-(an+1 _
~n+l)3
+ (an-1_
~n-1)3
= O.
The left member is transformed in the following way

5 (asn_ Wn)-a3n ( a3+ 1_
~3) +3a2n~n ( a2~+1- :2~)
3an~2n ( a~2 + 1- a~ 2 ) + ~3n ( ~3 + 1- ; 3 ) .
It is easy to show that a 2 ~ j- 1- a!~ = 0, a~ 2 1- a~ 2 = O.

On the other hand, we can easily make sure that
a 3 + 1-aJ= ~ 3 + 1-i=a 3 +~ 3 +1 =
=(a+~) (a 2 -a~+ ~ 2 )+1 = a 2 --a~+ ~ 2 +1=5.
Hence follows the validity of our identity.

(an- ~n)4- (an-2 _ ~n-2) (an-1 _ ~n-1) (an+l _ ~n+l) X
X (an+2_ ~n+2) = 25.
First prove that

(an-2_~n-2) (a"+2_ ~"+2) _,__ cz2n+ ~2"-(-1)n (a4 +
(an-1_ ~n-1) (an+I _ ~n+l) = a2n+ ~2n + ( - ft (a2 +
But
a 2 + ~2
(a + ~) 2 - 2a~ = 3,
=
a 4 + ~4 =
(a2 + ~2)2 _
~4),
~2).
2a2~2
= 7.
182
Solutions
Therefore
( an-2 _ pn-2) (an-1 _ pn-1) (an+i _ pn+1) (an+2 _ pn+2) =
= (a2n+ p2n)2-(-1t 4 (a2n + p2n)-21.
On the other hand

(anpn)4 -= a4"-4a3npn + 4-4anp3n + p4n =
=
a"n+ p4n+ 4-4 (-1)n (a2n+p2n).
Subtracting the last-but-one equality from the last one

termwise, we find the required result.
6 and 7 are proved analogously to the previous cases.
7. 1 We have
1
2 [(a +b 2)2 -a) [(a 2+b 2)2 -b] =

2
2 (a + b )-2 (a+ b) (a 2+ b2)2 + 2ab=

2
= (a 2 + b2)-2 (a+b) Va 2 + b2+(a+ b) 2 +
+ (a + b )+2ab-(a +b)
2
(singling out a perfect square).

Consequently
1
2 [(a2 +b 2)2 -a] [(a 2 +b 2 ) 2 -b] = (a+b-V a 2+b 2) 2

Hence follows the first identity.
2 Multiplying the braced expressions on the left, we get
2
3(a +b 3)3 -3(a+b)(a 3-j--b 3)3 +3ab=

3
=3(a2 -ab+b2}3 (a+b) 3 -3(a2 -ab+b 2)3 (a+b) 3 +

2
+(a+ b} 2-(a 2-ab+ b2) = [(a+b) 3 - (a 2 -ab+b2 ) 3 ]3.
183
Solutions to Sec. 3

-. /2a-b
8. It is easily seen that ax= v - b - , hence
(1-y~)2
2a-b
1--b-
b
( 1-2 /2a-b
2 (b-a)
V b
-. /2a-b
2a-b) = a-b v - b b
b-a
Analogously, we find
1 +~-.
-. /1+bx =
V 1-bx
/2a
b
b
--av;'"2a=b
lJ-bb-.
1 +~ ., /2a-b
a
V 1
a2
---::::======-. / _~.2a-b
b
., /2a-b
-. /2a-b
a-j-b v - b -
-. /2a-b
a+b v - b -
a+b v - b -
= Va2-2ab-j-b2 =
Y(b-a)2
b-a
>
(since b-a 0). Multiplying the two obtained expressioni::, we find

a2-b2 2a-b
b
(b-a)2
9. Factor the expression

n3
3n - 2.
We have
n3
3n - 2 = n 3
- 2 (n
+ 1)
Likewise
n3
3n
n -
2n -
= (n
+2 =
2 = n (n 2
+ 1) (n
(n -
2 -
n -
1) 2) =
= (n
+ 1)
1) 2 (n
+ 2).
(n -
2).
Solutions
184
Now we may write:

n3-3n-2+(n2-1) -V~
n3-3n+2+(n2-1) Vn2-4
_(n+t)2(n- 2)--J-(n2-t)V~_ (n+1lVn=2 X
- (n-1)2 (n+2)+(n2-1) Vn2-4 - (n-1) Vn+2
(n+1) -Vn-2+(n-1) Vn+2

=
(n-1) Vn+2+(n+1)Vn-2
(n +1) Vn-2
(n-1) Vn+2
10. Consider the second one of the fractions contained in

the first brackets, namely:
t-a
t-a
-VI=a
-V1-a2-1+a = V1-a2-(1-a)
= -Vt+a-Vt-a
And so, the transformed expression takes the form
-vr=a J.-V1-li2a
l/r+ll + -vr=a -Vf-ii2-1 -
-Vr+a
-Vt+a- -Vt-a
-
V1+a- -Vt-a
- -Vt+a- -Vt-a
2a
(Vt+a--Vt-a) 2
t -
( Vt=a2- t)
'
z <Vt=a2-t)
- -1
(t+a+t-a-2Vi-a2)-
11. From the formula (*) it is easy to get:
-VA+VB+VA-VB=2 yA+ ~A 2 -B.

In our case
A=x,
B=4x-4,
A2 -B=x2 -4x+4,
x-2 if x>2,
V A 2 -B = V(x-2) 2 = { 2-x if. x < 2.
In the first case we have
V x+2V x-1+'11 x-2V x-1 =2 .,,/x+x-2

V 2 =
=2 V x-1.
f85
Solutions to Sec. 3
The second case yields
V x+;-x
V x+2V x-t+V- x-2V x-1=2

=2.
It is easy to see that at x = 2 the expression under consideration is also equal to 2.
12. In this case
A =a+ b + c, B = 4ac + 4bc,
A 2 - B = (a + b + c) 2 - 4dc - 4bc =
= a 2 + b2 + c2 + 2ab - 2bc - 2ac =
=(a+ b - c) 2
If
a+ b - c > 0,
then
VA 2 -B=a+b-c.
If
a+b-c<O,
then
VA 2 -B=c-a-b.
Hence, we easily obtain that the given expression is equal
to
a + b if a + b > c, and to 2Vc if a + b < c. At
a + b = c these values coincide.
13. Let us denote
2v
V-~-vr+~=V.
Then
x = u
Consequently
x3
= (u
But
u3
+ v)
+ v3
+ v.
= u3
+ v3 + 3uv (u + v).
q,
UV
Therefore
x3 = - q - px
or
x3
+ px + q =
which is the required result.
Solutions
186
14. We can proceed, for instance, in the following way.

Put
V x+a+ V x-t-b=z.
Then(multiplyinganddividingtheleftmemberbyV x +a- V x + b) we find.

a-b
------=Z
-V x+a- -Vx+b
a-b
V -x+a- -vx+b=-.
z
or
Hence
--
--
a-b
2 V x+b=z--z-,
2 V x+a=z+-,
z
a-b
i.e. both roots are expressed in terms of z without radicals.

15. Put
Consequently
a'= a'A,
b' =b'A, c' = c'A,
A=
a'+h'+c'
a+b+c
Therefore
Va+-Vb+-Vc+ Va'+ Vb'+ V? =

= (1 + VX) (Va+ l/"b+ Ve).
Our fraction takes the form
1
(1+ yX)
(1-Vr)
(v;i"+ Vb+ ye)=

(1-YI)
(ya+ Vb"- VC)
<1-1) (a+b-c+2Viib)
(ya+ Yb-VC) (a+b-c-2"Viib)
( 1-1) (al +bz+c-2ab-2ac-2bc)

= ("Vii+iJ-FC-Ya'+b'+c') (Va+ Yb-VC+)(a+b-c-2\liib)
(a+b+c-a' -b'-c') (aZ+b+cl-2ab-2ac-2bc)
16. Put
Hence
= p3
+ 3pq + (3p2 + q) vq-,
ya:;:;;+e
187
Solutions to Sec. 3
since q is not a perfect square, it must be 3p 2 + q = 0,

which is impossible.
17. 1 We have
~ - a) =tan ( ~ - a) =cot a,
tan ( 3; - a) =tan ( n +
cos ( 3; - a) = cos ( n +
- a) = - cos ( ; - a) =
=-sin a (20, 40),
(10, 30),
cos (2n - a)= cos ( - a)= cos a

cos
(a - ~ ) =cos ( ;
- a) =sin
(30, 40),
(20, 30),
sin(n--a.)= -sin(-a)= +sin a

cos (n+ a)= -cos a
sin
(a -
)= -
sin ( ; -
a) = -
(2),
cos
(30, 40).
Now we get
a-sin
a
. 2
1 +sm
2 ex+ cos2 a= 0 .
2
-- -cot
-cos
a + s r n a+cos a= -
2 In this case we obtain

sin(3n-a)=(-1) 3 sin(-a)= -sin(-a)=sinex (2, 3),
cos(3n+a)=(-1)3cosex= -cos ex
(2),
sin ( 32n - a) = sin ( n +
- ex) = - sin (
- a) =
=-cos ex
cos ( 5; -
a) =cos ( 2n +
a) =cos ( ~ -
(2, 4),
a) = sin
(1 or 2", 4).
Thus, we have
(1 - sin ex - cos ex) (1 + cos ex + sin ex) + sin 2ex =
= [1 - (sin a +cos a)] [1 + (sin ex + cos a)] + sin 2a =
1 - (sin a + cos a) 2 + sin 2a =
= 1 - sin 2 ex - cos2 ex - 2 sin a cos a + sin 2<X = 0.
188
Solutions
3 Analogous to the previous ones.

18. Indeed, we have
, - 2 sm
2a ,
1 -cosa2
whence
. ..::.
2
Sm
1- cos a
=+ V
But in our conditions
Then
sin ~ = sin
(Im+
~o ) = ( - 1)" sin ~0
where
. ao ......._
srn T:::::' 0 .
Therefore, indeed
. ....'.:_-(- 1)"-. / 1-cosa
Sill
2 -
JI
The second assertion is proved analogously.

19. Let us prove the validity of some of the proposed formulas. Let us, for instance, prove that A 16 = 0 if n = 0
(mod 2). Put n = 2l. Then
1
( ln
3 )
( 3l n
5 )
2 A16 =cos 4 +n- 32 n +cos - 4 -+n- 32 :n +
=-COS ( -ln
- - - :3r t ) -COS
---i-
cos ( TIn
32
( ln----:Tt
ln
5 )
4
. , ( ....91 :n + l :n -r, 325 n ) +cos
3 :n ) - - cos ( 4ln - 32
32
+ 323 :n ) -_ 5 :n ) +
( -1) 1 cos ( Tln + 32
4ln
189
Solutions to Sec. 3
Let us prove, for instance, that A 14

(mod 7). We have:
0 if n
1, 3, 4
1
( 1
13 )
( 3
3 )
2 A 14 = cos 7 nn - 14 n + cos 7 nn - 14 n +
+ cos (
nn - ;: n) .
If we replace here n by a number, which is comparable with

it by modulus 7, then all the cosines will acquire only a
common factor equal to +1. Indeed, let us assume that

n =ex (mod 7), i.e. n =ex+ 7N, where N is an integer.
Therefore
COS ( k;n
-~)=COS
k(a~7N)n -~)
= cos ( k~n + kN n - ~ )
= (-
1) kN cos (
k~n
=(-1)N COS
~)
k~n -~),
since in our case k = 1, 3, 5 and, consequently, is odd; (~is

equal either to 1~ n or to n) . Therefore, in order to prove
that A 14 = 0 at n
1, 3, 4 (mod 7), it is sufficient to prove
that it will take place at n = 1, 3, 4. The validity of this
is readily checked.
First put n = 1. Then we prove that
!!
==
13
cos ( ..!.
~ n } + cos ( ~7 n - 14
~ n) = 0
7 n - 14 n ) + cos ( ~
7 n - 14
After transformations we get:
11 n +cos 3 n +cos 7 1t =cos ( n- 3n) +cos 3 n +
cos 14
14
14
14
14
1t
+cos 2 = -cos 14 n+cos 14 n =0.

Let now n = 3. Then we have to prove that
13
cos (~n+cos ( 15
=
7
14 n) +cos (~n-~n}
7
14
7 n-~n)
14
7
27
(
1t )
=cos 14
n+cos 15
14 n+cos 14 n=cos n+ 14 +
+cos ( 2n - ~4
= - cos ~4 +cos ~ = 0.
Solutions
190
Reasoning in t.he same way, we make sure that at n = 4

we also obtain zero.
In conclusion, let us prove that A 8 never becomes zero,
i.e. at no whole values of n. We have
7 n ) +cos
21 A8 =cos (1
4 nn- 16
cos (
! nn -
~ n) +
( -
1 ) =
41 nn+nn-11.in
( - 1t cos (
! nn +
1
16 n) .
Consider the following cases:

1 Let n
0 (mod 4), n = 4N. Then
~ A8 =cos (Nn- 176 n)+(-1) 1'Ncos (Nn+ 116 n)=

= ( -1)
=(
7
N
1
cos I6 n + ( -1) cos 16 n
-1t (cos~ n +cos 176 n).
The bracketed expression is not equal to zero, since it

represents a sum of cosines of two acute angles.
2 Let n
1 (mod 4), i.e. n = 1 + 4N.
==
} A8 = cos (
= ( - 1) N
= (-
+ N n - 176 n) + cos ( 3: + 3N n - 116 n) =

{
cos (
~-
7
16
n) + cos ( 3:
1~ n) } =
t {cos 1~ n + cos / 6 n } .
It is obvious that the braced sum is not equal to zero,

and, consequently, in this case A 8 is also not equal to zero.
It only remains to consider the cases: n
3 (mod 4) and
n = 2 (mod 4), but we leave them to the reader.
==
LJp
(k) = 0
ifk=n, n-1, n-2, n-4, n-5, n-6, and the

sign before p (k) is chosen accordingly.
It is evident that
~ p(k)=A ~ (k+3) 2 +C ~ (-1) 11 +D ~cos 2 ~k.
Solutions to Sec. 3
191
The first two sums on the right are equal to zero. It remains
to prove that
"" cos -3211k = 0.
LJ
If k is a whole number, the following cases are possible:

1 k is exactly divisible by 3, k = 3l;
2 k, when divided by 3, leaves the remainder 1, k =
= 3l 1;
3 k, when divided by 3, leaves the remainder 2, k =
= 3l+2.
In case 1
211k
cos -3-= 1.
2:rt
211k
In cases 2 and 3 cos - 3- = cos 3
Let us first assume that n is divisible by 3. Then

"1
211k_cos2:rtn
2:rt(n-1)
211(n-2)+
LJ cos-cos
3
3- - -3--cos
3
+
211 (n-5)
211 (n-6)
+ cos 2:rt (n-4)
3
cos
3
- cos
3
But
2
and
= -1
2:rtk
(mod 3)
211k 1
cos -3- = cos -3-
if
= k' (mod 3).
Since by the assumption n

0 (mod 3), we have
n - 1 == -1, n - 2 = 1, n - 4 = -1,
n - 5
+1, n - 6
and our sum takes the form
211
211
1- cos -3- - cos 3
211
= 0,
211
+cos -3- +cos 3 - 1 = 0.
It remains to prove that our sum is also equal to zero in

the cases when n = +1 (mod 3). The proof is similar to the
previous case.
Solutions
f92
21. We have
sin 15=sin(45-30)=sin { ~ - ~}=sin~ cos~ n . n
V2 V3 V2 f
-cos4sm 6=-2--2---2-2=
Analogously we find cos 15.

We have
n
. 180
Slil
= Slil fO =
2n
COS
V6-4 V2
But
.n
2 Slil
5
COS 5
.2n
2 Slil
5
COS
2n
.2n
Slil
.4n
Slil
.n
Slil
5 .
Multiplying these equalities termwise, we find

n
2n
= 4 .
cos 5 cos 5
On the other hand
n
2n
cos 5 - cos 5
.3n.
Thus, if we put
. n
srn
2n
= 2 sm 10 sm 10 = 2 cos 5 cos 5
10=COS
2n
5=X,
= 2 .
COS5=y,
we have
But
1
(x+y) 2 = (x-y) 2 +4xy= 4 + 1= 4 .

Consequently,
Vs
x+y=-2-
Using this relation and the relation y-x= ~ , we get

X
Slil fO
. 180
1/5
= Slil
= -f+
4
Now cos 18 is readily found.
193
Solutions to Sec. 3
22. Indeed
sin 6 = sin (60 - 54) = sin 60 cos 54 - cos 60 sin 54.
But
sin 54 =cos 36 = 1-2sin2 18= 1-2 B-! 5-V 5 = 1 +{ 5 ,
cos54 =Vi-sin2 54=
! V10-2 V5.
To obtain the result we have to substitute these values into

the first formula; cos 6 is found in the same way.
23. Bear in mind that
(1)
- 2:rt -<arcsmx::::::::,
+ 2:rt
< arctanx < + 2

0 < arccot x < n,
- 2:rt
:rt
O::(;arccos x::(;n,
(2) sin (arcsin x) = x, cos (arccos x) = x,
tan (arctan x) = x, cot (arccot x) = x.
Let us now prove that
cos(arcsinx) =Vi - x 2
Put
arcsin x
y,
then
sin y
x.
We have got to compute cosy. But it is known that

cosy= V1-sin 2 y= V 1-.r2 ,
and the radical is taken with the plus sign, since
and, consequently,
cosy~
0.
Let us, for example, also prove that

cos (arc tan x)
Put
arctan x
= y, tan y
x.
194
Solutions
We have to find cosy. We have

-cos1-2 -y =
1+tan2 y=1
+x
Consequently
2
cos Y=
1 +x2
and
cosy =cos (arctan x) =
V1 +.r2
where the rarlical is taken with the plus sign again. sincP
cosy~
0.
The rest of the form11las are proved in the same \\ay.

24. By definition,
< arctan ;r < +2,

0 < arccot ;r < :rt.
Therefore
- 2n < arctan x + arccot x < + -3n
2
Let us compute sin (arctan x + arccot x). We
sin (arctan x + arccot x) =
= sin (arctan x) cos (arccot x) +
+ cos (arctan x) sin (arccot x) =x
x
+ 1
1t
l1
- 2
have
-1
V1+ . 2 V1+x2 -
V1+.r2 V1+32
However, if the sine of a certain arc is equal to 1, then this

arc equals
~ +2k:rt,
where k is any whole number, i.e., in other words,
arclan x
arccot x
can attain one of the following values

-
71t
31t
1t
51t
91t
... , - 2 - - 2 - T 2 ' 2
195
Solutions to Sec. 3
But only one of them, namely ~ , is contained in the

interval between -
~ and
+ 3;
. Therefore it is obliga-
tory that
arc tan x + arccot x = 2
l1
Likewise, let us prove that
. x + arccos x = 2
l1
arcsm
First of all we have
.
___. 3n
- 2n ::::;;;arcsm
x + arccos x"":::z
On the other hand,

sin (arcsin x + arccos x)
= sin (arcsin x) cos (arccos x) +
+ cos (arcsin x) sin (arccos x) =
= x2
wherefrom follows that

. x
arcsm
+ V1 -
+ arccos x
x2
V1 -
x 2 = 1,
= 21t .
25. First of all it is easy fo prove that the quantities

arctan x + arctan y
and
arc tan --=!L.
1-xy
differ from each other only by en, where e is an integer.

Indeed,
tan (arctan t+Y
)
-xy
-
x+y
1-xy '
tan (arc tan x +arc tan y) =

tan (arc tan x) +tan (arc tan y)
1-tan (arctan x) tan (arctan y)
x+y
1-xy
196
Solutions
But if two quantities have equal tangents, then they differ

from each other by a term divisible by ll.
Therefore, indeed,
t+
arc tan x +arc tan y =arc tan
-xy
+ell.
()
Let us fmd out the exact val-ue of e. Since

:re
:re
:re
- 2 < arctanx < + 2 ,
:re
- 2 < arctany< + 2 ,
we have
-ll <arc tan x +-arc tan y < + ll
and, consequently,
Iarc tan t~:Y +en I< ll.

And since
:re
- 2 < arctan
x+y
1 -xy <
+ 2:re
then lei< 2 and, consequently, e may attain only one of

the following three values
0, +1, -1.
To find the value of e let us write the following equality
cos (arc tan +arc tan~)= cos (arc tan
t~x~ +ell}.
Hence
cos (arc tan x) cos (arc tan y)-sin (arc tan x) sin (arc tan y) =
= cos (arc tan
t~;Y ) cos ell.
On the basis of the results of Problem 23 we have

1
1
V1+x2. V1+y2
V1+x2
V1+y2
1
~""T-==========;-COSell,
. , /1
Consequently
cos ell=
1-xy
Yl1+x2) (1+y2)
+(
Vi (
+
x+y
1-xy
x+y
1-xy
)2
)2
Solutions to Sec. 3
197
We have
-./ 1
+( x-t-y
1-xy
)2=-./(1-t-x2)(1-t-y2) =
(1-xy) 2
Y(1-t-x2)(1-t-y2)
Y(1-xy)2
But
V(1-xy) 2 =1-xy if 1-xy>O, i.e. if xy< 1,

and
V(1-xy) 2 = -(1-xy) if 1-xy 0, i.e. if xy
1.
Therefore, cos e:n: = 1 if xy < 1, and cos en = -1 if
xy > 1. Since e:n: can attain only the values 0, :n: and -:n:,
it follows that if xy < 1, then e = 0, and if xy > 1, then
e = +1. What sign is to be taken is decided in the following way: if xy > 1 and x > 0, then also y > 0, then
arctan x > 0 and arctan y > 0, and arctan 1x-t-y < 0.
-xy
The left member of the equality(*) is a positive quantity,
consequently, the right member must also be positive, and
therefore e:n: must exceed zero, and e = +1. Quite in the
same way we make sure that if xy > 1 and x < 0, y
0,
then e = -1.
26. We have
>
<
<
2
1
4 arc tan 5 = 2 arc tan 5 + 2 arc tan 5 =2 arctan - -1- =

1--
25
= 2 arc tan 12 = acrtan 12 +arc tan 12 =

5
=arctan
12+12
25
1 -144
120
arctan 119 .
Further
(
1 )
arctan 120
119 +arc tan - 239 =
120
1
119-239
l't
=arc tan
120 1 =arc tan 1= 4
1 119. 239
27. Using the formula of Problem 25, we easily obtain

the result,
Solutions
198
28. First of all let us notice, that since arcsin x is contained between - ~ and + ~ , and 2 arctan x lies between
-:n: and +:n:, we have
3n
3n
< +2
2x
- 2 ~2arctanx+arcsin i+x 2
Let us now compute the sine of the required arc, i.e. find
what the expression
sin ( 2 arc tan x + arcsin 1 +2xx2 )
is equal to.
We have
sin ( 2 arctan x + arcsin 1 ~x 2 ) =
=sin (2 arc tan x) cos ( arcsin i -~x2 )+
+cos (2 arc tan x) sin ( arcsin 1 ~x 2 )
First compute sin (2 arctan x). Put
arctan x = y, tan y = x.
Then
sin (2 arctan x) = sin 2y = tan 2y cos 2y.
But
2tany
1-tan2 y
tan 2y = 1- tan2 Y , cos 2y = 1 + tan2 Y
Consequently,
.
2 tan y
2x
sin( 2 arctanx)=i+tan 2 y=i+x2 .
Further
cos ( arcsin
i~x 2 ) =
y 1 - ( i~x2 )
=
--~
- -. / (1-x2)2
- V (1+x2)2
since x > 1.
Further, it is obvious that
1-x2
cos (2 arc tan x) = 1+x2
. (
.
2x )
2x
sm arcsrn 1 x2 = 1 + x 2
x 2 -1
1 +x2
'
199
Solutions to Sec. 3
therefore
sin ( 2 arc tan x + arcsin 1 ~:r 2
)=
2x
x2 - 1
1 + x2 1 + x2
1 - x2
2x
+ 1 + x2 1 + x2 = O
Thns, the sine of the required arc is equal to zero, consequently, this arc can have one of the infinite number of values:
... , -3n, -2n, -n, 0, +n, 2n, 3:n:, 4n, ....

But among these values there are only three (-n, 0 and :rt}
lying in the required interval between - 3; and+ 3; . On the
other hand, x > 1 and, consequently, 2 arctan x > 0 and
arcsin 1 ~~xx 2 > 0, and therefore the required sum
.
2 arc t an x + arcsm
2x
+
xz
will also be greater than zero and, consequently, can be

equal only to n.
29. It is evident that
-n~ arc tan
x +arc tan-~+
n.
x
Let us form
sin (arc tan x +arc tan
!)
The required sine turns out to be equal to (see Problem 23)

sin (arc tan x) cos (arc tan
.! )+cos (arc tan x) sin (arc tan ! )=

1
l/1+x2
,J;
v1+x2
y1+-1
x2
1
V1+x2
Vx2
1
V 1 + x2 + V 1 +
x
y1+-1
x2
Vx 2
x2
xY1+x2
x2
1+x2
1
1+x2
=1
200
Solutions
if x > 0 (since in this case V:?=x). And if x < 0, then
V x2 = -x and we have sin (arc tan x +arc tan
!)
1.
<
0.
Hence follows that

1
arc tan x +arc tan 7 = + 2 + 2kn,

where plus is taken when x > 0, and minus when x
But since, on the other hand, it must be
1
-n::::;; arctan x+arctan -::::;;;

+n,
x
our problem has been solved.
30. Compute the expression
sin (arcsin x + arcsin y).
We have
sin (arcsin x + arcsin y) =sin (arcsin x) cos (arcsin y) +
+cos (arcsin x) sin (arcsin y) = x V 1-y2 + y V 1-x2
Thus, considering the two arcs
arcsin x + arcsin y
and
arcsin(xV1-y 2 +y Vt-x 2 ),
we may assert that their sines are equal to each other.
However, if
R
sma.=s1np,
2s1na-~cosa+~=O
,
2
2
and, ronsequenlly, either a;~ =kn or a~~ =(2k' +1);

(k and k' in legers), i.e. either
a=~+
or
a
= -
2kn
+ (2k' + 1) n.
/herefore we may assert that

arcsin x + arcsin y = 1J arcsin (x
V 1-y2 + y V 1 -
x 2 ) +en,
Solutions to Sec. 3
201
where ri = +1 if e is even, and ri = -1 if e is odd. To determine e more accurately, let us take cosines of both
members. We get
cos (arcsin x + arcsin y) =
= cos [l] a resin ( x V 1 - y2 + y V 1
x2)
+ en] .
y2 +yV1
x 2 )].
y2 +yV1
x2 } 2
Hence
V 1-
x2 V 1
xy =
y2 -
=(-1) 8 cos[arcsin(xV1
Further
V 1- x V 1
2
y2 - xy =
=(-1)eV1-(xV1
The radicand on the right can be transformed as

1- (x
V1
y2
+ y V1
x2) 2 =
= 1 - x 2 ( 1 - y2 ) - y2 ( 1 - x 2 ) - 2xy V 1 x 2 V 1 y2 =
= (1 - x 2 ) ( 1 - y 2 ) - 2xy V 1 x2 V 1 y2 + x 2 y2 =
== (V 1-x2 V1
y2 -xy) 2
If it turns out that
V1
x 2 l/T=Y2-xy
then
> 0,
V1-(x VT=Y2 + Y VT=X2r~ =

= V (V 1 x V 1 y 2 -xy} 2 =Y1-x~YT=Y2 -xy.
2
Therefore, in thi" case

(-1)E
i.e. e is even.
And if
then
V1
x2 V 1
+1,
y2 - xy
(-1)E = -1,
and, consequently, e is odd.
< 0,
202
Solutions
Let us now consider the expression

1 - x2 - y2.
We have
1- x2 _ y2 = 1 _ x2 _ y2 + x2y2 _ x2y2 =
= (1 - x2) (1 - y2) - x2 y2 =
= (V 1-x2 V1 -y 2 -xy) (V 1-x2 V 1-y2 +.iy).
The quantity 1 - x 2 - y 2 can be greater (smaller) than or
equal to zero. Let us consider all the three cases.
1 Suppose 1 - x 2 - y 2 > 0, i.e. x 2
y 2 < 1. If the
product of two factors is positive, then these factors are
either both positive simultaneously, or both negative Rirnultaneously. And so, we have either
V 1-
> 0, V 1 -
x2 V 1 - y2 - xy
x 2 V 1 - y2
+ xy > 0
or
V 1-x
Vi-y 2 -xy
< 0, Vi-x
V1-y 2 +xy
< 0.
But the second case is impossible, since, adrling the last two
inequalities, we get
V1 -
x 2 V 1 - y 2 < 0,
which is impossible. If, however, the first two inequalities
exist, then
V 1- x 2 V1- y 2 - xy 0.
>
Consequently, in this case e is even.

Thus, if x 2
y2 < 1, then in our formula e is even.
2 Let now 1 - x 2 - y2 < 0 and, consequently, either
or
Vi-x 2 V1-y 2 -a::y
> 0,
v 1-x Vi-y
v1 - 2v1 -
< 0,
v~ 1 - x 2
:i:
y 2- xy
v1 -
+xy<O
y2 + xy
> 0.
But from the first two inequalities we easily obtain xy < 0

If this inequality is fulfilled, then it will obligatory be
V1 -
x2
V 1- y
and, consequently, e is even,
2-
xy
> 0,
Solutions to Sec. 3
203
From the second pair of inequalities we get xy > 0, and

e is odd.
3 Finally, suppose 1 - x 2 - y 2 = 0. Then again two
cases are possible: either xy :=:;;; 0 or xy > 0.
In the first case li1 - x 2 V1-y2 -xy>0, and, hence,
e is even. Likewise, the second case gives an even e (e = 0),
since there exists the following relation:
arcsin x + arcsin V 1- x 2 = ~ (x
> 0).
Thus, we can judge whether e is even or odd. Now let us

consider the value of e. We have
I arcsin x + arcsin y I < n.
Consequently
J
11arcsin (x
-V 1-y
+yV1-x2 ) +en
< n.
Hence
lei< 2.
And so, e may attain only three values: 0, +1, -1.

Comparing the results obtained, we may now assert that
< 0, then e = 0, 11 = +1,

and if x 2 + y 2 > 1 or if xy > 0, then e = +1, 11 = -1.
To find out when e = +1 and when e = -1, let us notice
that at x > 0, y > 0 arcsin x + arcsin y > 0 and, conif x 2 + y 2 :=:;;;1 or if xy
sequently,
-arcsin (x V 1-y2 + y V 1--x2 ) +en> 0,
and therefore in this case e = +1. If, however, x
y < 0, then it is obvious that e = -1.
31. We have (see Problem 24)
arccos x + arccos (
~ +
+V
3 - 3x2 ) =
= :n-arcsin x-arcsin ( ~
++ V
<
3-3x 2 };
on the other hand (Problem 30),

arcsin x + arcsin ( ; +
+V
3 - 3x2 }
0,
~ 11 arcsin + en,
Solutions
204
where
=x
x . V3v- 2 )2 +
1 -x
( 21-__
2
l/ 1 -
-2) 11-1-x.
-2
+ (2x +V3-v-1
2
-x
But
1 - ( 2x
+--V3v-)2
1- x
=
22
41 (v,11
- x 2 - ,1v 3 x )2 ,
and since x ~--}-, we have 4x 2 ~ '1: 3x2 ~ 1 - x 2 and

V3x~V1-x 2
Therefore
y1 -(; + ~V 1
x2
)2={V (V 1 x -V3x)
2
={ (V3x -V 1 -x
and
2)
-V3
s=-z.
Consequently
. t
arcsm.,,
= 3n .
The only thing which is left is to find 'l'J and e (see Problem 30).
Let us prove that
x2
2
+ ( 2x + 2-V3..,~2)
v
x- > 1.
1 -
We have
+ T + 43
x2
X2
(1- x 2) + 2
v-3 x ,v,1r- -x- ~

2
--..... 3
~T
+ yx
1 2
5
+z1 (1 -x2) =4
Consequently,
'l'J
= -1, e =
+1.
Therefore,
arccosx+arccos(
~+--}V0-3x 2 )=n-(-~
+n)=
~.
Solutions to Sec. 3
205
32. We have tan A= T, tan B = 3 . Let us compute

cos 2A. Since
1+tan2 A=_!_
A,
cos 2
we have
1
cos2 A
50
49
= 1 + 49 =- 49 and cos A= 50 .
But
98
cos 2A = 2 COS" A -1 = 50 -1
24
= 25 .
Further
sin 4B = 2 sin 2B cos 2B.
But
2
cos 2B = 2 cos 2 B-11= 1 +tan 28 -1=5'

sin 2B = 2 sin B cos B
2 tan B cos 2 B = 1 ~t::n~ B =
!.
Consequently,
.
24
sm4B=2s-s= 25
and sin4B=cos2A.
33. By hypothesis we have

(a+b) 2 =9ab or
34. Put
log an= x,
Then
(a"tb)2 =ab.
logma n = y.
Hence
ax= mY.aY,
aY==ma.
Taking logarithms of this last equality to the base a, we get

the required result.
35. Put
x(y+z-x) _
logx
y (z+x-y)
z (x+y-z)
logy
log z
206
Solutions
Then
log x = tx (y
Hence
y log x
+ z -x ),
+ x log y
logy = ty (z
x - y),
log z = tz (x
+y-
z).
+ z log y = 2txyz,
z log x + x log z = 2txyz.
= 2txyz, y log z
Consequently
y log x
+ x logy =
y log z
+ z logy
log xYyx
= z log x
+x
log z,
= log zYyz = log xzzx.
Finally
xYyX
36. 1 Put logb a
zYyZ
xzzx.
x. Then
bx= a.
Taking logarithms of this equality to the base a, we get

loga b = 1.
But x = logb a. Consequently, indeed, logb a loga b
2 We have
X
= 1.
Therefore
logb (logb
a)
(alogb a)logb(logb a)= (aloga b)logb (logb a)=
b !ogb (logb a)
= ogba
37. From the given relations 'it follows that

yl-log x
10,
z1-log y
10.
Taking logarithms of these equalities to the base 10, we get

(1 - log x) logy = 1, (1 - logy) log z = 1.
whence
1- = 1 -----=-log x = 1 -logy
__1_
1-log z
1
log z
and, consequently,
1
X = 101-log z.
207
Solutions to Sec. 3
38. The original equality yields

a 2 = (c - b) (c
Hence
2 logc+h a
+ b).
1,
2 logr-b a = logr-/J (c
logc+b (c -
b)
4 logc+ba logc-ba = logc+b (c -
+ b)
+ 1.
b) + logc-b (c + b) +
+ 1 + logc+b (c -
b) loge-b (c + b).
However,
logc-b (c + b) logc+b (c -
b) = 1.
Therefore
4 logc+b a logc-b a = 2 logc+b a - 1 + 2 logc-b a - 1+ 2.

Finally
log,+b a + logc-b a = 2 logc+b a log c-b a.
39. Put
loga N = x, loge N = y, logy-a,; N = z.
The last eq na Ii ty yields
z
(ac)2 = N.
Hence
log" N = ~ (1 +log a c),
loge N
~ (1 +loge a).
Therefore
2.r z
1 = 1ogac,
2y
--1=logra.
z
Consequent 1y
or
x
y
40. We have
x-z
z-y
208
Solutions
41. Let
Then
log an = log a+ n log q, log an - bn =
= log a + n log q - b - nd = log a - b.
Hence
n log q - nd = 0, log11 q = d, ~d = q.
And so
1. We have
( x-ab
a+b
-c) + (~b) + ( x-bc

_a) =O.
a_Lc
b--\ c
Hence
x-ab-ac-bc
a+b
x-ac-ab-bc
a+c
x-bc-ab-ac -O
b+c
-
or
1
1 )
--+--+-=0.
a+c
b+c
1
(x-ab-ac-bc) ( a+b
Assuming that
1
a+b
+ a+c + b +c
is not equal to zero, we obtain

x
If, however,
1
a+b
= ab
+ ac + be.
1
+ a+c + b+c =O,
then the given equation turns into an identity which holds

true for any value of x.
2. Rewrite the equation as follows
_ _!_ _ _!_)+(
(~
be
b
c
x-b _
ac
_!_ _ _!_) + (~-_!. _ _!_) =O.

a
ab
209
Solutions to Sec. 4
We have
x-a-b-c _ x-b-a-c __ x-c-a-b
be
-j
ac
f
ab
=O
Hence
1
1
1 )
(x-a-b-c) ( bc+ac+ab"
=0,
and, consequently,
x =a+ b + c.
It is assumed, of course, that none of the quantities a, b

and c, as also b1c
_!._
_!_b is equal to zero.
ac
a
3. If we put in our equation
+ +
6x
+ 2a
= A, 3b
+c=
B, 2x
+ 6a
C, b
+ 3c =
D,
then it is rewritten in the following way

A+B
A-B
C+D
C-D '
Adding unity to both members of the equation, we find

2A
2C
A-B .-- C-D .
Likewise, subtracting unity, we get

2B
A-B
2D
C-D'
Dividing the last equalities termwise, we have

A
7J -=-75'
i.e.
6x+2a
3b+c
2x+6a
= b+ 3c
Hence
Finally
ab
.T=-.
c
Solutions
210
4. Add 3 to both members of the equation and rewrite

it in the following way
{
a+~-x
+ 1 ) + ( a+~-x + 1 ) + ( b+:-x + 1 ) =
Hence
(a+ b+c-x) (
! + ~ + ! )=4 a~~t~~.x.
Consequently
1
(a+b+c-x) ( a-+-;;-+7and, finally,
x =a+ b
a.j-b
te
=0
+ c.
5. Taking Vb+ x outside the brackets in the left member, we get
vb+ x b +bx x =--=-a v:x.

Consequently,
1+
(b+x)
1+
!
P
!p
be
-a
Hence
P+1
b+x
(x-
)P =a,
be
b~x=
r+1
Further
p
!!..x =
( !:__ ) P+T a
'
X=-----
6. 1 Squaring both members of the given equation, we

frnd
x
1
x - 1
x 2 - 1 = 1.
+ +
+ 2v
211
Solutions to Sec. 4
Consequently,
2v x 2
4x
1 = 1 - 2x,
4 = 1
2 -
+ 4x
2 -
4x,
x =t;
Since squaring leads, generally speaking, to an equation

not equivalent to the given one, or rather to such an equation which in addition to the roots of the given equation may
have other roots different from them (so-called extraneous
roots), it is necessary to check, by substitution, whether
is really the root of the original equation. The check shows
that
does not satisfy the original equation (here, as before, we consider only principal values of the .roots).
2 Carrying out all necessary transformations similar to
the previous ones, we find that x = { is the root of our
equation.
7. Cube both members of the given equation, taking the
formula for the cube of a sum in the following form
(A
+ B)
= A3
+ B + 3AB (A + B).
3
We ha\fe
a+ Vx+a-V x+3Va2 -x(Va+ V x-1-V a-VX) =b.

Since
V a+ V x+ V -a-V x=Vb,
we have
2a-j--3 V a2 -x
v-b=b,
X=a 2 -
(b-2a)3
27 b
We assume that a and b are such that

2_
(b-2a)3 >-O
27b
~
Since the equality of cubes of two real numbers also

means the equality of the numbers themselves, the found
val11P of x satisfies the original equation as well.
212
Solutions
8. Squaring both members of the equation, we find
-V x
Hence
x4
x 2 [x2
x2
1 - x2
x2
4 -
2x.
x 2 (x - 2) 2 = 0,
x2
+ 4x]
= x2 (4x - 5)
0.
!.
Thus, the last equation has two roots x = 0 and x =

Substituting them into the original equation, we see that
the unique root of this equation is
5
X=4
9. Getting rid of the denominator, we obtain
or
Vb(x-b)=Va(x-a),
b(x-b)=a(x-a),
x=a+b.
As is easily seen, this value of x is also the root of the original equation.
10. Multiplying both the numerator and denominator by
Va + x + Va- x, we get
(Va+x+-V a-x) 2 =2xVb.

Hence
-V a2 -x2 =xVb-a.
Squaring both members of this equality, we find two roots
X=O,
2a-Vii
1+b.
However, the first of these values is not the root of the original equation, the second one will be its root if
213
Solutions to Sec. 4
Indeed, we have
1 ;-::-;-::
v u -r x
2a
l/b
V a + -r+b ~_,
-. /
v-a -.V/(1++-Vb) = vr:a 1+

Vb ,
-V H
Va--x=-./a_2a-Vil=Va
1 :-b
=Va
=
v/(-Vb-1)2
1+b
-v;;-
1
v1+b
(if Vb-1;?0).
Substituting the obtained values for Va

x and Va - x
into the original equation, we make sure that our assertion
is true.
11. Adding all the given equations, we have
, + + V= a+b+c+d
3
.r,-y
Consequently
V=(x+y+z+v)---(x-j-y-j-z)=~
a+b+c+d
-a~
3
Likewise, we obtain
Z=
a+c+d-2b
3
,
Y=
a+b+d-2c
3
,
X=
a+b+c-2d
3
.
12. Adding all the four equations, we get

4x 1 = 2a 1
+ 2a2 + 2a + 2a
3
4,
Multiplying the last two equations by -1, and then adding

all the four equations, we find
Similarly, we get
214
Solutions
13. Put x + y + z + v = s. Then the system is rewritten as follows

ax + m (s - x) = k
by + m (s - y) = l
cz + m (s - z) = p
dv + m (s - v) = q
so that
ms + x (a - m) = k, ms + y (b - m) = l,
ms+ z (c - m) = p, ms+ v (d - m) = q.
Hence
k
q
d-m
m
d-m
X = a-m - a-ms, y= b-m - b-m s, Z= c-m - c-m s,

V=-----S.
Adding these equalities termwise, we find

k
S=
a-m
+ b-m
+c=m+ d-m 1
1
1
1 )
-ms ( --+--+--+-a-m
b-m
c-m d-m .
Consequently
s [ 1 +m(
a~m
b~m + c~m
d~m)] =
=-k-+_l_+_P_+_q_.
a-m
b-m
c-m
d-m
Wherefrom we find s, and then from the equalities () we

obtain the required values of the unknowns x, y, z and v.
14. Put
'A.
Hence
ll1
+mi'),,,
X2 = a2
+ m2'A,
X1
215
Solutions to Sec. 4
Substituting these into the last one of the given equations,

we get
X1
+ X2 + . . . + Xp = a =
= (a 1 + a 2 +
+ ap) + 'A, (m1 + m 2 + ... +
mp)
Consequently,
a-a 1 -a 2 'A,=
m1+m2
-ap
+ ... +mp
'
and then we readily get the values of

15. If we put
1
-;=X,
y=y,
7=Z,
v=V,
then the solution of this system is reduced to that of Problem 11. Using the result of Problem 11, we easily obtain
3
x = a+ b + c -
2d '
Y = a+ b + d -
2c '
z = a+c+d-2b' V = b+c+d-2a
16. Dividing the first equation by ab, the second by ac
and the third by be (assuming abc =F 0), we get
Hence
~=
(.::.+JL+~)--(.::.+JL)
=_!_
(-c
+J:...+..!:..)
c
a
b
c
a
b
2
ab
ac
be
Consequently,
analogously
z
c
-=
Y=
a2+ b2-c2
2a b c
a2+c2-b2
2ac
'
i.e. z =
X=
__
c.
ab
a2+ b2-c2
and then
2 ab
b2+c2-a2
2bc
17. First of all we have an obvious solution x = y =

= z = 0. Let us now look for nonzero solutions, i.e. for
216
Solutions
such in which x, y, z are not equal to zero. Dividing the

first of the given equations by yz, the second by zx and the
third by xy, we obtain
..:.+.!;_=2d
~+!...=2d'
.!;_+~=2d".
z
y
'x
z
'y
x
Hence
..'.:+.!;_+~
=d+d' +d".
x
y
z
Therefore
!!_=d'+d" -d,
.!;_=d+d"-d'
!....=d+d' -d".
z
'
Finally
x=
y = d+d"-d' '
d'+d"-d '
z=
c
d+d'-d"
18. Rewrite the system in the following way

1
ay-1-b:r
xy
-
az+cx
1
xz
= b
c'
'
bz+ey
yz
1
a
Hence
~_i_.!;_
X
_ _!_
y -
c
1
+7=/]
y-+-;=a-
Consequently (see the preceding problem)

2a2be
2ab2e
2abe2
= ae +ab- be ' Y = be+ ab- ae ' Z = be+ ac - ab

19. The obvious solution is x = y --; z = 0. Dividing
both members of each equation of our system by xyz, we get
X
1,1
1_1
Xi'7y-Tz-~
1_1
Xy+Yz-Xz-bz'
1
y.+x;-;y=7
Adding pairwise, we find
2
7y=az+b2 g.=b2+7'
Consequently
217
Solutions to Sec. 4
Multiplying the equalities, we obtain

8a4b4c4
2 2 2
Y Z = (a2 + /J2) (b2 + c2) -(a-=-2+___,c2c:--)
Hence
xyz =
2 -V2 a2b2c2
-:-;:==::::::::::=====
-V(a2 1-b2) (b2-tc2) (a2-tc2)
Using the equality

xy
2a2b2
az+1;2'
we find for z two values which differ in the sign. By the

obtained value of z we find the corresponding values of y
and x from the equalities (*). Thus, we get two sets of
values for x, y and z satisfying our equation.
20. Adding all the three equations, we find
(x + y + z) (a + b + c) = 0.
Hence
whence
X =
a-b
a+ b + c '
a--c
Ii-a
z ~ a -i b c
Y =- a+ b -j- c '
21. Adding all the three equations termwise, we get

(b
+ c) x + (c + a) y + (a + b) z =--= 2a + 2b + 2c
3
3.
Using the given equations in succession, we fmd

2 (b + c) x = 2b3 + 2c3 , 2 (c + a) y = 2a3 + 2c3 ,
2 (a
whence
x = b2
be
+c
2,
+ b) z =
y = a2
2a3
ac
+ 2b
+c
2,
3,
z = a2
ab
+b
22. Consider the following equality

x
a+0+
y
b-t0
+T+e- 1=
(0-A,) (0-) (0-v)

(0-ta)(0-tb)(0-tc)
Let us transform the equality, by reducing its terms to a

common denominator and then rejecting the latter. We get
a second-degree polynomial in 8 with coefficients depending
on x, y, z, A., , "' a, b, c, which is equal to zero. If now we
Solutions
218
substitute successivly /.., and v for 0 into the original

expression, then, by virtue of the given equations, this
expression (and, consequently, the second-degree polynomial) vanishes. However, if a second-degree polynomial
becomes zero at three different values of the variable, then
it is identically equal to zero (see Sec. 2) and, consequently,
the equality
x
a+0
+ b+0 + c+0
(0-A.) (0-) (0-v)

(0+a) (0+b) (0+c)
- 1=
(by virtue of existence of the three given equations) is an

identity with respect to 0, i.e. it holds for any values of 0.
Multiplying both members of this equality by a + 0, put
8 = -a. Then we find
x=
(a+A.) (a+l (a+vl

(a - b) (a-c)
---,---:-,-~---
Likewise we get
y=
(b+A.) (b+) (b+v)

(b-c) (b-a)
'
Z=
(c+A.l (c+) (c+v)

(c-a) (c-/J)
Of course, we assume here that the given quantities/.., ,

v, as also a, b and c, are not equal to one another.
23. The given equations show that the polynomial
ex 3 + xex 2 + yex + z
vanishes at three different values of a, namely at ex = a, at
ex = b and at ex = c (assuming that a, b and c are not equal
to one another).
Set up a difference
ex 3 + xex 2 + yex + z - (ex - a) (ex - b) (ex - c).
This difference also becomes zero at ex equal to a, b, c.
Expanding this expression in powers of ex, we obtain
(x
+ a + b + c) ex + (y
2
- ab - ac -
be ) ex
+
+ z + abc.
This second-degree trinomial in ex vanishes at three different

values of ex, and therefore it equals zero identically and,
consequently, all its coefficients are equal to zero, i.e.
x
+a+b+c=
0,
y -
ab -
ac -
be = 0,
+ abc
= 0.
21!J
Solutions to Sec. 4
Hence
+ e),
ab + ae + be,
x = -(a+ b
y =
z = -abe
is the solution of our system.
24. We find similarly
+-
t = -(a + b + e
d),
x = ab + ae + ad + be + bd + ed,
y = -(abe + abd + aed + bed),
z =abed.
25. Multiplying the first equation by r, the second by p,
the third by q and the fourth by 1 and adding, we get
(a3
+ a q + ap + r) x + (b + b q + bp + r) y +
+ (e + e q + ep + r) z + (d + d q + dp + r) n
= mr t- np + kq + l.
3
Let us choose the quantities r, p and q so that the following equalities take place
+ b q +. bp + r = 0,
c3 + e q + ep + r = 0,
d3 + d q + dp + r = 0.
b3
Hence, we obtain (see Problem 23)

q = -(b + e. + d),
and, consequently
X=
p = be+ bd
N
a3+a2q+ap+r
+ ed,
r = -bed,
N
(a-b)(a-c)(a-d)
'
where
N = -mbed + n (be+ bd
As to the equality
+ ed)
- k (b
+ e + d) + l.
+ a2q + ap + r
= (a - b) (a - e) (a - d),
it follows readily from the identity
a3
a.3
+ qa.2 + pa.+ r =
(a. -
b) (a. -
e) (a - d).
220
Solutions
To find the variable y, the quantities q, p and rare so chosen

that the following equalities take place
a 3 + a 2q + ap + r = 0,
c3 + c2q + cp + r = 0,
d3 + d2q + dp + r = 0.
The remamrng variables are found analogously.
26. Put
X1 + X2 + . . . + Xn = S.
Adding the equations term by term, we get
s + 2s + 3s + . . . + ns
But
+n
1+2+3+
= a1 + a 2 + . . . + an.
n(n+1)
(an
arithmetic progression).
Therefore
s
2
n (n+1)
(a1 + a2 + ... + an) = A (for brevity).
Subtracting now the second equation from the first one, we

find
x 1 + x 2 + x 3 + ... + Xn - nx 1 = a 1 - a 2
Hence
and
A+a 2-a 1
X1=
n
Subtracting the third equation from the second, we get

X2=
and so on.
27. Put
Then we have
A+as-a2
n
+ X2 + . . . + Xn
-s + 2x = 2a,
-s + 8x =Ba,
X1
Hence
8
= S.
+ 4x 2 = 4a,
+ 2nxn = 2na.
-s
-s
x1 =a+ 2, x 2 ==a+ 4, x 3 =a+ 8, ... , Xn =a+ 2n.
221
Solutions to Sec. 4
Adding these equalities, we get
s=na+s ( 21 + 41
+ ... +rn1 ) .
But
Therefore
Consequently
x 2 =a+ 4 =a+ 2n- 2 na = a (1
+ n 2n-
and so on.
2)
28. Let
Then
X2
S -
2,
S -
X3
... , s -
3,
Xn -1
s-
Xn
Consequently (since s = 1)
X2=-1,
X3=-2,
... ,
Xn
1,
= n -
= n.
-(n - 1).
Hence
X2
X3
+ ... +
Xn
= - ((1
+ 2 + ... + (n -
1)1
n(n-1)
Finally
X1
= 1 - (x2
X3
29. Suppose the equations are compatible, i.e. there

exists such a value of x at which both equations are satisfied.
Substituting this value of x into the given equations, we
get the following ideqtities
ax
+ b = 0,
a'x
+ b'
0.
Multiply the first of them by b', and the second by b. Subtracting termwise the obtained equalities, we find
(ab' -
a'b) x = 0.
222
SolutionlJ
If the common solution for x is nonzero, then it actually
follows from the last equality

ab'-a'b=O.
If the common solution is equal to zero, then from the original equation it follows that
b = b' = 0,
and therefore in this case also
ab' - a'b = 0.
And so, in both cases, if the two given equations have a
common solution, then
ab' - a'b = 0.
Hence, conversely if the condition
ab' - a'b = 0
is satisfied, the two given equations have a common root
(the coefficients of the equations are proportional), and,
consequently, they are compatible.
30. To prove that the given systems are equivalent it is
necessary to prove that each solution of one of the systems
is simultaneously a solution for the other system. Indeed,
it is apparent, that each solution of the first system is at
the same time a solution for the second system. It only
remains to prove that each solution of the second system
will also be a solution for the first system. Suppose a pair of
numbers x and y is the solution of the second system, i.e.
we have identically
ls+ lT
where
o,
ms+ m's'= 0,
s=
s'
ax+ by + c,
= a'x + b'y + c'.
Multiplying the first equality by m' and the second by l',
and subtracting them termwise, we find
(lm' - ml') = 0.
Likewise, multiplying the first equality by m and the second by l, and subtracting, we get
(lm' - ml') 6' = 0.
Solutions to Sec. 4
223
But since, by hypothesis,

lm' - ml' =F 0,
it follows from the last two equalities that
~=0
and
i.e.
o,
ax+ by+ c = 0
and
a'x
+ b'y + c'
= 0.
Thus, the pair of numbers x and y, which is the solution of

the second system, is simultaneously the solution of the
first system.
31. Multiplying the first equation by b' and the second
by b, and subtracting termwise, we find
(ab' - a'b) x + cb' - c'b = 0.
We get similarly
(ab' -
a'b) y
+ c'a -
a'c = 0.
These two equations are equivalent to the original ones.

a' b =F 0, then there exists one
and only one pair of values of x and y satisfying the last
two equalities, and, consequently, the original system as
well.
32. Multiplying the first equality by b' and the second
by b, anrl subtracting, we find
(ab' - a'b) x = 0.
It is evident that if ab' -
Since, by hypothesis, ab' - a' b =F 0, it follows that x = 0.

In the same way we prove that y = 0.
33. From the first two equations we get
c'b-cb'
x-= ab' -a'b '
a'c-c'a
Y = ab' -a'b
lf the three equations are compatible, then a pair of numbers x and y being the solution of the system of the first two
equations must also satisfy the third equation. Therefore,
if the three given equations are compatible, then there
224
Solutions
exists the following relation
a"
or
a" (c'b - cb')
c'b-cb'
ab' -a'b
+ b"
+ b" (a'c -
c'a)
a'c-c'a
ab' -a'b
+c" = O
+ c" (ab'
- a'b) = 0.
()
Conversely, the existence of this relation means that a

solution, which satisfies the first two equations, satisfies the
third one as well. This relation may be rewritten in the
following ways
a' (cb" - c"b) + b' (ac" - ca") + c' (ba" - b"a) =~ 0,
a (c"b' - c' b") + b (a"c' - c"a') + c (b"a' - a"b') = 0.
Hence it follows that the solution of each pair of the three
equations is necessarily the solution of the third equation,
i.e. our system is compatible provided the condition (.)
is observed.
34. Subtracting from the first equality the second, and
then the third one, we find
(a - b) y + (a 2 - b2) z = 0, (a - c) y + (a 2 - c2) z = 0.
Since a - b =I= 0 and a - c =I= 0, we have the following
equalities
y +(a
+ b) z =
0,
+ (a + c) z
= 0.
Subtracting them term by term, we have

(b - c) z = 0.
But by hypothesis b - c =I= 0, therefore z = 0. Substituting this value into one of the last two equations, we find
y = 0. Finally, making use of one of the original equations,
we get
x = 0.
35. Multiplying the first equality by B 1 and the second
one by B, and subtracting them termwise, we get
(AB 1 - A 1B) x + (CB 1 - C1B) z = 0.
(1)
We find analogously
(AC1 - A1C) x
+ (BC
1 -
B 1C) y = 0.
(2)
225
Solutions to Sec. 4
Suppose none Qf the expressions

AB 1 -
A 1B,
CB 1 -
C 1B,
AC 1 -
A 1C
is equal to zero. Then we get

x
Z'
C1B-CB1
AH1-A1H
[multiplying both members of the first equality by the

product
and
y
Thus, in this case the required proportion really takes

place.
Let now one and only one of the expressions
AB 1
A 1B,
CB 1 -
C 1B,
AC 1
A 1C
vanish. Put, for instance, CB 1 - C 1B = 0. Then from

equalities (1) and (2) we get x = 0. Further, suppose that
two of the mentioned expressions, for instance, C1B - CB 1
and CA 1 - AC 1 are equal to zel'o, and the third one, i.e.
AB 1 - A 1B is nonzero. We then find x = y = 0. In these
cases our proportion, or, more precisely, three equalities,
= /.. (C 1B - CB 1),
y = /.. (CA 1 - AC1),
z = /.. (AB 1 - A 1B),
will also take place.

Thus, in these cases two given equations determine the
variables x, y and z "accurate to the common factor of proportionality".
If all the three quantities
AB1 -
A 1B,
CB 1 -
C 1B
and
AC 1 -
A 1C
are equal to zero, then there exists the following proportion

A
221\
Solutions
In this case the two equations (forming a system) turn

into one, and nothing definite can be said about the values
of the variables x, y and z which satisfy this equation.
36. From the first two equations (see the preceding problem) we get
x
y
bc-a2
oc-b2
ab-c2
Hence
x
= '). , (ac -
b2 ),
y = ')..,(be - a 2 ),
= ')..,(ab -
c2 ).
Substituting these values into the third equation, we find

b (ac -
b2 )
or
a3
+ a (be +b +c
3
a 2)
+ c (ab
- c2 )
3abc = 0.
3 -
37. Multiplying the first two equations, we get

x2
y2
22
a2 - -c2-- 1 - b2
The same result is obtained by multiplying the third equation by the fourth one, which shows that if there exist any
three of the given equations, then there also exists a fourth
one, i.e. the system is compatible.
To determine the values of x, y and z satisfying the given
system proceed in the following way: equating the right
members of the first and the third equations, find
Solving this equation with respect to y, we have

-A.
y=b +A. .
Substituting this into the first two equations, we get
=- _:_ _
a
+c
2A.
+A. '
:r
2
+A. .
Hence
X=a
Z= C
A.-i
+A.
227
Solutions to Sec. 4
38. Rewrite the system in the following way

a (x + py) + b (x + qy) = ap 2 + bq 2
ap (x
ap 11 - 1 (x
+ py) + bq (x + qy)
+ bq
= ap 3
+ py) + bqh-l (x + qy)
= ap 11 +1
+ bq 11
+1.
Now it is obvious that the system is equivalent to the following two equations
x +PY = p2, x + qy = q2,
and, hence, the system is compatible.
39. We have
= a1 - X1i
x 3 = a2 - x 2 = a2 x 4 = a3 - x 3 = a3 -
X2
a1
a2
+x
+ a1 1,
x 1,
Xn = an -1 - an -2
+ ll2 + ll1 + X1
It should be noted that in the last equality the upper signs
will occur when n is odd, and the lower signs when n is
even.
Consider the two cases separately.
1 Let n be odd. Then
Xn = an-I - an -2 + + a2 - al + X1
On the other hand,
Xn + X1 =an.
From these two equalities we get
an -an-1 +an-2- -a2 +a1
2
'
X1=
and, hence,
a1 -an +an_ 1- ... -a3+a 2
2
X2=
X3=
a2 -
'
a1 -t- an - ... - a4 + aa
2
'
2 Let now n be even. Then

Xn =an-I - an-2
+ -
a2
+al - X1.
Solution.~
228
On the other hand,
Consequently, for the given system of equations to be compatible the following equality must be satisfied
an -1 -
an -2
i.e.
an
+ ... -
+ an -2 + + a2
a2
= an -1
+ a1 =
an,
+ an -3 + + al
(the sum of coefficients with even subscripts must equal

the sum of coefficients with odd subscripts). It is apparent
that in this case the system will be indeterminate, i.e. will
allow an infinite number of solutions, namely:
X1
= '). ,
X2
=al -
')..,,
X3
= a2 -
al
x 4 = a3
Xn
+ '). ,
+ a 'A,,
an-2 + ... + a
a2
an-l -
1 -
3 -
a2
+ a 1-/..,,
where 'A, is an arbitrary quantity.

40. From the first two equations we find
x
-a-=-2--"'""b2,--
a-d -
='A,.
b-d
Substituting this into the third equation, we have

').., {
a~d ( b~d
+ c~ d
c~d ) + b~d ( c~2d
( a: d -
b b2 d ) } = d
a:d )
(a - b) (b - c) (c - a).
After sim plifica ti on we get

a
( b2
a-d
b-d -
c2
c-d
c
( a2
c-d
a-d -
b
+ b-d
b2 )
b-d
c2
c-d- -
a2
a-d
d (a-b) (b-c) (a-c)

(a-d) (b-d) (c-d)
Therefore
'A, =
-(a - il) (b - il) (c - il),
229
Solutions to Sec. 4
and, consequently,
x = (a -
d) (b - e) (db
y = (b - d) (e - a) (de
z = (e - d) (a - b) (ad
+ de + da + db -
be),
ae),
ab).
41. Solving the last two equations with respect to x

and y, we find
x+n= (c-m)(n--a)
z+c
(b-l) (m-c)
y+b=~---
z+m
Hence
x+a= (c-m)(n-a) --(n-a)=(a-n) z+m.
z+c
z+c
Analogously
Y+
(l - b) z + c .
z+m
Substituting the fournl values of x + a and y + l into the

first equation, we see that it is a consequence of the two
last equtions. Thus, the system is indeterminate, and all
its solutions are given by the formulas
X=
(c-m) (n-a)
z+c
-n,
(b-l) (m-c)
y = ---'--'-----'--
+m
b,
for an arbitrary z.
42. From the second and the third equations we have
(1 - k) x
+ ky
= - [(1
+ k) x + (12 -
k) y],
hence, taking into account the first equation, (5 - k) y =

= 0 wherefrom either k = 5 or y = 0 (hence x = 0), which
yields (substituting into the second equation) k = -1.
43. We have
sin 2a = 2 sin a cos a,
sin 3a = sin a (4 cos2 a - 1),
sin 4a = 4 sin a (2 cos3 a - cos a).
230
Solutions
Therefore the first of the equations of our system is rewritten in the following way
x + 2y cos a + z (4 cos 2 a - 1) = 4 (2 cos3 a - cos a).

The remaining two are similar. Expand this equation in
powers of cos a. We have
8 cos3 a - 4z cos 2 a - (2y + 4) cos a + z - x = 0.
Putting cos a = t and dividing both members by 8, we get
t3 - ~ t2 -
z-; x
t+
0.
Our system of equations is equivalent to the statement that

the equation () has three roots: t = cos a, t = cos b and
t = cos c, wherefrom follows (see Problem 23)
z
2 = cos a+ cos b +cos c,
Yt 2=
x-z
- 8-
-(cosacosb+cosacosc+cosbcosc).
=cos a cos b cos c.
Therefore the solution

x = 2 (cos a + cos b +
y = -2 - 4 (cos a cos
z = 2 (cos a + cos b +
of our system will be

cos c) + 8 cos a cos b cos c,
b + cos a cos c +cos b cos c),
cos c).
44. Put
a
- - - - - - - - k
sin A - sinB - sinC -
Since A + B + C = n, we have
sin A = sin (B + C) = sin B cos C + cos B sin C.
But from the given proportion we have
'A =-,;,
a
srn
'B
srn
b
=-,;,
cc
sm
=-,; .-
Substituting this into the last equality, we find

a = b cos C + c cos B.
The rest of the equalities are obtained similarly.
231
Solutions to Sec. 4
45. Expressing a and b in terms of c and trigonometric

functions (from the first two of the given equalities), we get
b ~. c (cos A--- cos B cos C)
-
sin2 C
ll=
c (cos B +cos A cos C)

sin2 C
(1)
'
(2)
Substituting (1) and (2) into the third equality anrl accomplishing all necessary transformations, we find
1 - cos 2 A - cos 2 B - cos 2 C - 2 cos A cos B cos C = 0.
A+ B
+C
= n.
Transform the obtained equality in the following way

cos 2 A + 2 cos A
= 1 - cos 2 B cos 2 A + 2
= 1 -
cos B cos C =
cos 2 C - cos 2 B cos 2 C + cos 2 B cos 2 C,
cos A cos B cos C + cos 2 B cos 2 C =
cos 2 B ~ cos 2 C (1 - cos 2 B),
(((os A + cos B cos C) 2 = sin 2 B sin 2 C.
But since we have obtained [see (1)1 that

cos A + cos B cos C =
h sin2 C
c
>O,
we have
cos A +cos B cos C =sin B sin C,
cos A =sin B sin C - cos B cos C =~ - cos (B -i- C),
cos AT cos (B +
c) = 2 cos A+s+c
2
cos
wherefrom follows that either
A+~+c =(2l+1) ~
or
A-B-C
- -2- - = (2z'
+ 1) T,
A-B-C
2
= 0,
232
Solutions
where l and Z' are integers. Let us first show that the second
case is impossible. In this case we would have
C = (2l'
1) n, B =A - C - (2l'
1) n,
cos B = cos (A - C - n) = -cos (A - C) =
= -cos A cos C - sin A sin C.
Consequently,
A - B -
cos B + cos A cos C = -sin A sin C
<
which is impossible, since we have obtained (2)

sin2 C
cosB+cosAcosC=
> 0.
Thus, there remains only the case

A
-f:-
+C =
(2l
+ 1) n.
However, by virtue of the inequalities, existing for A, B

and C, we have
0 < 2l
1<3,
i.e.
2l
1 = 1
and
A+ B
C = n.
+
+

a
sin A =
sin B =
We have shown that

cos A + cos B cos C
sin C
sin B sin C.
On the other hand,

cos B + cos A cos C = cos (n - A - C) + cos A cos C =
= -cos (A + C) + cos A cos C =
= sin A sin C.
Using this equality and also equalities (1) and (2), we easily
obtain the required proportion.
46. Let us first show that equation (2) follows from equations (1). Multiplying the first of equations (1) by a, the
second by b and the third by - c and adding them term-
Solutions to Sec. 4
wise we get
a2
+b
2 -
c2
233
2ab cos C,
i.e. the third of equations (2). Likewise we obtain the remaining two of equations (2).
To obtain equations (1) from equations (2) add the first
two of (2). Collecting like terms, we find
2c 2 -2 be cos A - 2 ac cos B = 0.
Hence
c = b cos A
a cos B,
i.e. we get the third of equations (1). The rest of them are
obtained similarly.
47. From the first equality we get
cos a - cos b cos c
sin b sin c
cos A = ---,--.,----,,---Hence
sin 2 A = 1 - cos 2 A =
sin2 b sin2 c-(cos a-cos b cos c)2
sin2 b sin2 c
_ (1-cos2 b) (1-cos2 c)-(cos a-cos b cos c)2
sin2 b sin2 c
1-cos2 a-cos2 b-cos2 c+ 2 cos a cos b cos c
sin2 b sin2 c
Consequently
sin2 A
sin2 a
1-cos2 a-cos2 b-cos2 c+ 2 cos a cos b cos c

sin2 a sin2 b sin2 c
Since the given formulas turn one into another by means

of a circular permutation of the letters a, b, c, A, B, C,
and as a result of this transformation the right member of
the last equality remains unchanged, we actually have
sin2A
sin2 a
sin2B
sin2 b =
sin2 C
sin2 c
But the quantities a, b, c and A, B, C are contained between

0 and n, therefore
sin A
!$ill
(1
sin B
= Sfrlb
sin C
---<
STri""C" .
234
Solutions
48. 1 Let us take the last two of the eqnalities (*) from
the preceding problem. We have
cos b - cos c cos a = sin a sin c cos B,
-cos a cos b
cos c = sin a sin b cos C.
Multiplying the first of them by cos a and the second by 1
and then adding, we find
-cos c cos 2 a
+ cos c =
sin a sin c cos B cos a

sin a sin b cos C.
Hence
cos c sin a = sin c cos a cos B
+
+ sin
b cos C.
But since it was shown in the preceding problrm that from

the equalities (*) follows the proportion
sin a
sin A
sin /J
sin H
sin c
sin C
in the last equality we can replace the quantities sin a,

sin b anrl sin c by ones proportional lo them. We get
cos c sin A - c sin C cos a cos B -t- sin B cos C.
It is apparent, that there exist six similar equalities. Let us
take one more of them, namely, the one which also contains
cos c and cos a. It will have the form
cos a sin C = sin A cos c cos B t sin B cos A .
(This equality can be obtained in the following way: multiply the second of the equalities (*) by cos c and the first
one by unity, add them, and in the obtained equality replace sin c by sin C .and so on.) Thus, we have
cos c sin A = sin C cos a cos B
sin B cos C,
cos a sin C = sin A cos c cos B
sin B cos A.
+
+
Eliminating cos c, we find

cos A = -cos B cos C
+ sin B
sin C cos a.
The rest of the equalities are obtained from this one w>ing a
circular permutation.
2 The formulas (*) of Problem 47 make it possible to
express cos A, cos B and cos C in terms of sin a, sin b,
235
Solutions to Sec. 4
sin c and cos a, cos b, cos c. Let us find the expressions for
. A an d cos A . W e h ave
sm
2
2
2 A
i
A
2 sin
- 2 = - cos = 1 -
cos a - cos b cos c
Sill
b Sill
(b-c)-cos a
sin b sin c
A 1
cos a - cos b cos c
~
sin b sin c
-
CQS
2 Cos 2 ..:!.._ = 1 ,
2
+cos
cos a-cos (b+c)

sinbsinc
Hence
A
a+b-c
a+r-b
sin
2
2
sin
sin-=
2
A
cos 2
sin h sin c
a+b+c . b+c-a
sin
s111
2
2
----s-i1_1_/J_s-i1-1c-- - -
Similar expressions are obtained for sin ~ . cos {- and

.
A+B
srn 2
A-j-B
, cos 2 . Now compute s111 - -2 - . We have

.
srn--=sin - 2 cos--L
cossin - 2
2
2
2
~=
a+b+c . a.+b-c
sin
sm
' 2
2
----s...,.in_a_s-:-i-n""""'b_ _ _ _ X
. a+c-b
2
( . sm
X
sine
b+c-a )
a-b
sin
C cos - 2
2+---s-in_c_ _ =COST
c
cos 2
Thus, we have obtained the following formula

.
Slll
a-b
cos - 2 c
-c
COST.
2- =
cosz-
A+B
236
Solution1
Likewise we find
A+B
COS - 2-
a+b
cos-2c
.
Slll
2 .
COST
Since e ---= A + B + C - n, we have

n
A+B
C-e
-2-=2--2Therefore
.
A+B
Slll - 2-
=COS -
C-e
2-
and, consequently,
C-e
cos-2-
a-b
cos-2c
C
COST
COST
Hence
C-e
C
cos - 2 - -cos 2
a-b
c
cos - 2--cos 2
C-r
C
ros-2-+cos T
a-b
cos-2
J
I
c
cos2
and, conseq1wntly,
p-b
p-a
tan 4e tan ( 2C - 4e ) =tan2- t a n -2- .
(1)
Using the formula

a+b
cos--
cos AtB =
. c
srn 2 ,
COST
we find analogous! y
p-c
tan 4e cot ( 2C - 4e ) =tan 2p tan2- .
(2)
Multiplying the equalities (1) and (2) termwise and

extracting the square root, we get
1
tan 4 e=
-. /
p-a
p-b
p-c
V tan 2p tan2-tan - 2-tan--r-.
237
Solutions to Sec. 4
49. We have
a [tan (x+y)-tan (x+
~)]
+ b [tan (x +a)- tan (x+y)] +

+c [tan (x+~) -tan (x+a)] =0.
Hence
asin(y-B)
bsin(a-y)
1
cos(x+B)cos(x+y) + cos(x+a)cos(x+y) T
_ _c_s_in-="(B_-_a-'-)__ = O.
+ cus(x+B)cos(x+a)
a sin (y - ~)cos ( x +a)+ b sin (a --y) c0s (x +~)I+ c sin(~ - a) cos (.c + y) --= 0.
Finally
a sin (y-B) cos a+ b sin (a-y) cos p+ c sin(~ -a) cosy
tan x = ---'-'-----'-'-------'------'-'-----'--'-----'.----=--~
a sin (y-p) sin a +b sin (a-'\') sin~ t- c sin (P--a) sin y
50. We have
,, x
cos 2
1+tan2 ~
2
Therefore
cos x
2 cos 2 ~
1=
sin x = tan x cos x =
1-tan 2 ~
2
1 +tan2 ~
2
2tan 2
1-tan2 _x_
2
1-tan2 ~
2
1 -1- tan2 ~
2
2tan 2
1
'
tan2 .!._
2
It is obvious that if tan

is rational, then sin x and cos x
are also rational. Show that if sin x and cos x are rational,
then tan ~ is rational too.
From the first relationship we have
( 1 + tan 2 ~
cos x
1 - tan 2 ~
Hence
tan 2 .::._ --= 1 -cos x
2
1 ..Leos x
238
Solutions
Consequently, if cos xis rational, then tan 2 ~ is rational as

well. Bnt from the second eqnality it follows that
2 tan ; =sin x ( 1 + tan 2
Hence, it is clear that if sin x and cos x are rational, then

tan ~ is also rational.
51. Since siu 2 x + cos 2 x = 1, we have
si 11 4 .c
cos~ x
2 sin 2 .c cos 2 x ~~. 1,
i.e.
sin 4 x
cos 4 x = 1-2 sin 2 x cos 2 x.
Therefore the equation is rewritten as

1 - 2 sin 2 x cos 2 x = a,
2 sin 2 x cos 2 x = 1 - a,
sin 2 2x = 2 (1 -
a), sin 2x
+ v2
(1 -
a).
For Lhe solutions to be real it is necessary and sufficient

that
+~a~i.
52. 1 Transforming the leH member of the equation,
we get
sin x
+ sin 3x + sin 2x
2 sin 2x cos x
=
sin 2x (1
Hence
(1) sin 2x = 0,
(2),
sin 2x =
2 cos x) = 0.
1
2
COS X= - - .
2 In this case the transformation of the left member

yields
cos nx
+ cos (n -
i.e.
either cos x
2) x - cos x = 2 cos (n - 1) x cos x cos x = cos x [2 cos (n - 1) x ~ 1] = 0,
=0
or cos (n -
1) x
1
=z-
239
Solutions to Sec. 4
53. 1 We have
m (sin a cos x -
cos a sin x) -
- n (sin b cos x - cos b sin x) = 0,

m sin a) cos x = 0,
(n cos b - m cos a) sin x - (n sin b -
nsinb-m sin aj
(n cos b - m cos a) cos x [ tan x b
_
n cos - m cos a
= 0.
Hence
t
nsinb-msina
an :r = n cos b - m cos a
2 We have
sin x cos 3a
+ cos x sin 3a =
3 (sin a cos x - cos a sin x).
Hence
sin x (cos 3a
+ 3 cos a) -
cos x (3 sin a - sin 3a) = 0.
But
cos 3a = 4 cos 3 a - 3 cos a, sin 3a = 3 sin a - 4 sin3 a.
Therefore the equation takes the form
sin x cos 3 a - cos x sin3 a = 0.
And so
tan x = tan 3 a.
54. It is easy to find that

sin 5x = 16 sin 5 x - 20 sin 3 x
+ 5 sin x.
Therefore our equation takes the form
-20 sin 3 x
+ 5 sin x
sin x (1 -
4 sin 2 x) = 0.
or
Thus, we have the following solutions

sin x
-=
0,
sin x
+}.
55. We have
2 sin :r cos (a - .r) - sin a
+ sin (2x -
a).
240
Solutions
The equation takes the form

sin x + sin (2x - a) = 0
or
x-a
O
. 3x-a
2 sm
- 2- cos - 2- . = .
Thus, the following is possible

. 3x-a
O an d -3x-a
sm-=
-=kn,
2
i.e.
3x= a+2kn,
X=
a-+ 2kn
3
'
where k is any integer.

Similarly, we have
cos x;a =0,.
x 2 a =(2l+1) ~ ,
X=a+(2l+1)n,
where l is any integer.

56. We have
sin x sin (y-x) =-}[cos (2x -y)-cos y].
Therefore the equation is rewritten in the following way
cos (2x - y) - cos y = 2a,
cos (2x - y) = 2a + cosy.
57. We have
. (a + x )+ smasm
.
. x sin(a+x)
O
sm
cos(a+x) --mcosacosx= .
Further
sin(a+x)
cos (a-j-x)
{cos (a+ x) +sin a sin x} - m cos a cos x = 0.
Hence
sin(a+x)
cos ( a -1,-x)
cos a cos x - m cos a cos x =

=cos a cos x{tan (a+ x)-m}= 0.
Assuming cos a =F 0, we obtain the following equalities

for determining x
cos x = 0,
tan (a + x) = m.
Solutions to Sec. 4
241
58. Rewrite the equation in the following way

cos 2 a + cos 2 (a + x) - 2 cos a cos (a + x) = 1 - cos 2 x.
Hence
[cos a - cos (a + x)]2 - sin 2 x = 0,
i.e.
[cos a - cos (a+ x) - sin xl [cos a - cos (a + x) +
+sin xl = 0.
Fnrther
[cos a (1 - cos x) + sin x (sin a - 1)1 X
X
[cos a (1 - cos x) + sin x (sin a + 1)1 = 0,
sin 2 x [cos a tan ; +sin a - 11 X

X
[cos a tan ; + sin a + 1] = 0
(if sin x =J= 0). If sin x = 0, then cos 2 a (1 - cos x) 2 = 0.

Now we easily find the following solutions:
cosx=1,
tanx=cota, i.e. x=2kn
and
2k--j-1
2
:rr.
59. We cnn readily obtain

.
srn2x=
2 tan x
1 -t-, t an2 x
Therefore
2 tan x
(1-tanx) ( 1 + i--j-tanZx
=1--t-tanx.
Hence
(1- tan x) (1 +tan x)2
1 tan2 x
_ (j
+ t an x ) =
O
'
1--j-tanx
O
-,------=--{1tan 2 x- 1- tan 2 x} = ,
1+tan2 x
tan2 :r ('J +tan x) _
1 + tan2x
-
For determining x we have: tan x = 0, tan x = -1.
Solutions
242
60. We have
tan A+ tan B = sin (A+ B)
cos A cos B
Therefore
tan x+ tan 4x+ tan 2x +tan 3x
sin 5x
+ cos 2x cos 3x
sin 5x
cos x cos 2x cos 3x cos 4x X
X
But
cos 3x
sin 5x
------..,--+
cos x cos 4x
{cos 2x cos 3x +cos x cos 4x}.
= 4 cos3 x - 3 cos x.
Thus, our equation takes the form

sin 5x
2
[cos2x(4cos x-3)+cos4x]=0.
cos 2x cos 3x cos 4x
Hence
sin 5x [4 cos2 2x-cos 2x-1] _
cos 2x cos 3x cos 4x
-
Consequently, either sin 5x = 0, i.e. 5x = kn, or

4 cos 2 2x - cos 2x - 1 = 0,
that is
s cos 2x = 1 + V IT.
61. Substituting the expressions containing X and Y for
x and y into the trinomial
ax 2
we get
+ 2bxy + cy
2,
ax 2 + 2bxy + cy 2 = a (X cos 0 - Y sin 0) 2 +

+ 2b (X cos 0 - Y sin 0) (X sin 0 + Y cos 0) +
+ c (X sin 0 + Y cos 0) 2 =
= (a cos 2 0 + 2b cos 0 sin 0 + c sin 2 0) X 2 +
+ (a sin 2 0 - 2b sin 0 cos 0 + c cos 2 0) Y 2 +
(-2a cos 0 sin 0 + 2c cos 0 sin 0 + 2b cos 2 0 -2b sin 2 0) XY.
243
Soiutions to Sec. 4
Since, by hypothesis, the coefficient of XY must be equal

to zero, we have the following equation for determining 8:
2b (cos 2
e-
sin 2 8) - 2 (a - c) sin 8 cos 8 = 0
or
2b cos 28 -
c) sin 28 = 0.
(a -
Thus,
2b
tan28 =a-c
--.
62. It is obvious that

sin(20+a+~)
x+y
x-y
sin(a-~)
Therefore
x+y
x-y
sin 2 (a-~)+ y+z sin 2 (~-y)+ z+x sin 2 (y-a)=

y-z
z-x
=sin (28 +a+~) sin (a-~)+ sin (28+~+1') sin (~-1') +

+sin (28 + y+a) sin (y-a).
But
sin (28 +a+~) sin (a -
~) = ~ {cos (28 + 2~)- cos (28-t2a)}.
Using a circular permutation, we easily check the validity of our identity.

63. 1 Put
sin x
a
sin y
sin z
c
= k
We then. have
sin x = ak,
sin y = bk,
sin z = ck.
On the other hand,

sin z = sin (n - x -
y) = sin (x + y)
=
sin x cos y + cos x sin y.
Hence
a cos y + b cos x
c, b cos z + c cos y = a,
c cos x + a cos z = b.
244
Solutions
Solving this system, we find

cos x
~+~-~
cos y =
2bc
~+~-~
2ca
a2+b2-c2
2ab
cos z =
At k = 0 we get also the following solution sin x =
= sin y = sin z = 0.
2 Put
tan_
x _
__ _
tanb_____
y _ tan
z_
_ k
_a
c_
Hence
tan x
tan y
ak,
bk,
tan z
ck.
Adding these equalities term by term, we get (see Problem

40, Sec. 2)
(a
+ b + c) k = tan x
Consequently,
+ b + c) k
(a
Thus,
k= +
k=O,
Hence either tan x

t
-. /
anx= + V
+ tan y + tan z =
(a+b+c) a
be
an
vr
tan y
'
+-. /
Z= -
k 3 abc
,-0:
tan x tan y tan z.

0.
a+b+c.
abc
tan z
0 or
-, /
any=+ V
(a+b-rc)b
ac
'
(a+b+c) c
ab
.
64. We have
tan 2b =tan (x + y) =
tan x+tan Y
1-tan x tan y
But, by hypothesis,
tan x tan y = a,
therefore
tan x + tan y = (1 - a) tan 2b.
Knowing the product and sum of the tangents it is easy
to find the tangents themselves (see Sec. 5).
245
Solutions to Sec. 4
65. Transform the equation in the following way

1
4x++ 4x=3x--z (1+3),
Hence
4x-l
_ 2_
=3
x--2
22x-3 = (V3)2x-3.
And so
Consequently,
3
2x-3=0 and x= 2 .
66. Taking logarithms of both members of our equation,
we find
(x + 1) log 10 x = 0.
Hence
x = 1.
67. Taking logarithms of the first equation, we find

x log1 0 a + y log 10 b = log 10 m.
Finally, we have to solve the system
x log 10 a + y log 10 b = log 10 m,
x + y = n.
68. Put
x = M, y =all
(from this problem on we assume that a > 0, b
b =I= 1 and find positive solutions).
Then (by virtue of the first equation):
b~Y =
But
Consequently,
aTJx.
>
0, a =I= 1,
246
Solutions
Hence
a~s
(S - 'l'J) = 0,
Thus, either x = 0 or '11 =
But at x = 0 we get y = 0,
Rejecting this solution, consider the case T) = i
Consequently,
and
But
x log a = y log b,
a'IJX,
s.
b' log a= a' log b,

Hence
I
s(logb-loga) =
log b
log-,
1oga
s=
log b
og loga
logb-loga
Therefore
x=b~=
log*
Iogb
logb
)
logb-Ioga
.
Since the ratio of logarithms of two numbers is independent

of the base chosen, in the expression
I
log b
og Toga
log b
we may consider the first logarithms as taken to the base b.

Then
I
Iogb
oglog;l
log b
log b
log a
and
X
= ( log b }
log a
log b
log b-Iog a
Y
=(
Ioga
logb }Iogb-Ioga
log a
69. Taking logarithms of the second equation, we find

logx
loga =
logy
logb
247
Solutions to Sec. 5
Putting this ratio to be equal to t we get

x =a, y = bs.
Substituting these values into the first equation and assuming a= b 1 , we find 6 = -1. Thus
1
y=T
X=7.
70. We have
Consequently,
mx
Xm=y-Y-.
Making use of the second equation, we find

mx
y-Y-=yn.
:x =
Hence, either y = 1, and then x = 1 or
n, i.e.
x = ..!!:.JL
m
Substituting into the second equation, we have:
m )m =y,n
( ny
And so
r- n'
ym-n = ( - m
n
y= ( :
x = y: = (
)m .
m
r-n .
1. We have
(b+x) (x+c)
x 2 .,.--..'--'--'---'----'"
(x-b) (x-c)
x3 (b + c + x) + xbcx
(x-b)(x-c)
Therefore the left member of our equation is equal to

(b+c+x)
x3
b3
c3
(x-b)(x-c)+(b-x)(b-c)+(c-x)(c-b)
J+
+ bcx [(x-b)x(x-c) + (b-x)b(b-c) + (c-x)c(c-bJ
248
Solutions
But (see Problem 8, Sec. 2)

x3
(x-b) (x-c)
b3
c3
+ (b-x) (b-c) + (c-x) (c-b) = b +c+x,

x
b
c
-0
(x-b) (x-c) + (b-x) (b-c) + (c-x) (c-b)-
+ c + x)
(b + c + x}
(b
Hence
+ c + x. -
b -
(b
+ c)
+ c) = 0,
c) (b + c + x + b + c)
2 -
(b
(b
= 0,
and consequently
X1
= 0,
x 2 = -2 (b
+ c).

(x -a) (x-b) (x-c) (b-c) (c -a) (a- b) { (x-a)
(c:a\ (a-b)
~
~
}
+ (x-b) (b-c)
(a-b) + (x-c) (c-a) (b-c) =
O.
As is known (see Problem. 9, Sec. 2)

a3
(a-x) (a-b) (a-c)
b3
+ (b-x) (b-a) (b-c) +

c3
x3
+ (c-x) (c-a) (c-b) + (x-a) (x-b) (x-c) = 1.
Therefore, the equation is rewritten as follows

(x-a) (x-b) (x-c) (b-c) (c-a) (a-b) X
X { 1-
x3
(x-a) (x-b) (x-c)
=0
or
(b - c) (c - a) (a -
b) [(x - a) (x -
b) (x - c) - x 3 ] = O.
Assuming that a, b, c are not equal, we get

(a
+ b + c) x
:c=
2 -
(ab
+ ac + be) x + abc
0,
ab+ac+bc V(ab+ac+bc)2-4abc (a+b+c)

2 (a+b+c)
249
Solutions to Sec. 5
For the roots to be equal it is necessary and sufficient that
(ab
+ ac + bc)
2 -
+ b + c)
4abc (a
0.
Hence
a 2 b2
+a c + bc
2 2
2a 2 bc - 2b 2ac - 2c 2 ab = 0,
(ab + ac - bc) 2 - 4a 2 bc
2 2 -
= 0,
_ _i__=O.
(..!.+..!._..!.)2
c
b
a
be
or
[(;c + ;ii )
2
! J[( ;c - ;ii ) ! J = 0
2
Finally
(
1 )(
Ve --r- Vii + Va x
1 )(
Ve - Vii+ Va
Ve + Vii - Va
x ( Ve - Vii - Va
'
1)
1)
=O.
3. Rewrite the equation in the form

3
(a-x) 2 +(x-b)2 _
1
-a
-b
'
(a-x)2 +(x-b)2
wherefrom we have
1
a-x-(a-x)2 (x-b)2 +x-b-=a-b

or
V(a-x) (x-b) = 0.
Thus, the required solutions will be
X1
=a,
X2
= b.
4. We have
V 4a+b-5x+ V4b+a-5x=3Va+b-2x.
Squaring both members of the equality and performing all
Solutions
250
the necessary transformations, we get
V 4a+b-5x V 4b+a-5x= 2 (a+b-2x).

Squaring them once again, we ftnd
(4a
+ b) (4b + a) =
Hence
x2
+ b + 4b + a) + 25x11 =
4 (a11 + b2 + 4x 2 + 2ab - 4ax - 4bx).
ax - bx + ab = 0,
5x (4a
and, consequently,
X1
=a,
X2
= b.
Substituting the found values into the original equation, we

get
Vb-a+2 Vb-a-3 Vb-a=O.
21/ a-b+ Va-b-3 Va-b=O.

Hence, if a -=I= b, then the equation has two roots: a and iJ
(strictly speaking, if the operations with complex numbers
are regarded as unknown, then there will be only one root).
5. Rewrite the given equation as
(1
+ A.) x 2 -
(a
+ e + 'J...b +
'J...d) x
ae
+ 'J...bd
= 0.
Set up the discriminant of this equation D ('J...). We have
D ('J...) = (a
+ e + 'J...b +
'J...d} 2
4 (1
+ 'J...)
(ae
+ 'J...bd).
On transformation we obtain
D ('J...) = 'J. . 2 (b -
d} 2
+ 2'J... (ab + ad + be + de - 2bd - 2ae) + (a - e}
We have to prove that D (A.) ~ 0 for any 'J.... Since D (A.) is

a second-degree trinomial in '). , and D (0) = (a - e) 2 > 0,
it is sufficient to prove that the roots of this trinomial are
imaginary. And for the roots of our trinomial to be imaginary, it is necessary and sufficient that the expression
4 (ab
+ ad + be + de -
2bd - 2ae) 2 - 4 (a -
e)ll (b - d) 2
251
Solutions to Sec, 5
be less than zero. We have
4 (ab
+ ad + be + de -
2bd - 2ae) 2 - 4 (a - e) 2 (b - d) 2 =
= 4 (ab + ad + be + de - 2bd - 2ae - ab
eb + ad - ed) X
X (ab
ad + be
de - 2bd - 2ae + ab - eb - ad + ed) =
= -16 (b - a) (d - e) (e - b) (d - a).
The last expression is really less than zero by virtue of the

given conditions
a< b < e < d.
6. The original equation can be rewritten in the following way
3x2
2 (a
+ b + e) x + ab + ac + be =
0.
Let us prove that
4 (a
+ b + e)
2 -
12 (ab
+ ae + be) ~ 0.
We have
4 (a
+ ae + be) =
= 4 (a + b + e
ab - ae - be) =
= 2 (2a + 2b + 2e
2ab - 2ae - 2be) =
= 2 {(a
2ab + b + (a
2ae + e +
+ (b 2be + e =
= 2{(a - b) + (a - e) + (b - e)
+ b + e)
12 (ab
2 -
2 -
2 2)
2 -
2 -
2)
2 -
2 )}
2}
~ 0.
7. Suppose the roots of both equations are imaginary.

Then
p 2 - 4q < 0, p~ - 4q1 < 0.
Consequently
p2
+ p= -
4q -
4q1
which is impossible.
<
0,
p2
+ p~ -
2ppl
<
0,
(p - P1) 2
<
0,
252
Solutions
8. Let us r!'Jwrite the given equation as
(a
+ b + c) x
2 (ab
2 -
+ ac + be) x + 3abc
= 0.
Prove that its discriminant is greater than or equal to zero.

We have
4 (ab
+ ac + bc)
12abc (a + b + c) =
= 2 {(ab - ac) 2 + (ab - bc) 2 + (ac 2 -
bc) 2 }
? 0.
9. By properties of the quadratic equation we have the

following system
p
+q=
-p,
pq
= q.
From the second equation we get

q (p ~ 1) = 0.
Hence, either q = 0 or p = 1. From the first one we find
if q = 0, then p = O; if p = 1, then q = -2.
Thus, we have two quadratic equations satisfying the set
requirements
x 2 = 0 and x 2
x - 2 = 0.
10. We have
x2
+y +z
2
2 -
= ~ (2x 2
xy -
xz -
+ 2y + 2z
2
= 2 {(x -
yz =
2xy - 2xz - 2yz)
2 -
y) 2
+ (x -
z) 2
+ (y
z) 2 } ? O
(see Problems 6 and 8).

But we can reason in a different way. Rearranging our
expression in powers of x, we get x 2 - (y
z) x
y2
z2 - yz. To prove that this expression is greater than,
or equal to, zero for all values of x, it is sufficient to prove
that: firstly
y2
z2 - yz? 0
and, secondly,
(y
+ z)
2 -
4 (y 2
+z
2 -
11z) ~ O.
253
Solutions to Sec. 5
It is evident that there exist the following identities
y2 + z2 - yz =
(y
+ z)
y- }
z) + : z2,
2
4 (y 2 + z2 - yz)
2 -
-3 (y - z) 2
and, consequently, our assertion is proved.

11. We have
a2
a2
x2+ y2+z2-3 =x2 + y2+ (a-x-y)2-3.

It is necessary to show that the last expression is greater
than, or equal to, zero for all values of x and y. Rearranging
this polynomial in powers of y, we get
a2
y2+ (x-a) y+x 2-ax+ 3 .
It remains only to prove that for all values of x
x 2-ax+ ; 2 ;;:?O,
We have
(x-a) 2-4 ( x 2 -ax-t- ; 2 ) ~O.
x 2 -ax + 3a2 = ( x- 2a
(x-a) 2-4 ( x 2-ax+
) 2
.....__ O
+1-12 a 2 ~
,
~) = -3 ( x- ~a ) 2~O,
which is the desired result. However, the proof can be carried out in a somewhat different way. Indeed, it is required
to prove that
3x2 + 3y 2 + 3z 2 ;;:? a2
if
x 2 + y 2 + z2 + 2xy + 2xz + 2yz = a2.
Consequently, it suffices to prove that
3x3 + 3y 2 + 3z2 ;;:? x 2 + y2 + z2 + 2xy + 2xz + 2yz
or
2x 2 + 2y 2 + 2z 2 - 2xy - 2xz - 2yz ~ 0.
And this last inequality is already known to us (see, for
instance, Problem 6).
12. See the preceding problem.
13. By the properties of quadratic equation we may write
a
~ =
-p,
a~ =
q.
254
Solutions
Therefore
St=
Since a and
-p.
are roots of the equation
+ px + q
x2
we have
~2
a 2 +pa+ q = 0,
= 0,
+ p~ + q =
0.
Adding these equalities term by term, we find
+ ps1 + 2q
S2
Hence
S2
0.
= -psi - 2q = p 2 - 2q.
Multiplying both members of our equation by xk, we get

xk+2
pxk+l
qxk = 0.
Substituting a and
and adding, we find
+ psk+ 1 + qsk
sk+2
Putting here k
= 1, we have
Further
S3 = -p (p 2 - 2q)
Likewise we find
S4 = p4 - 4p2q
To obtain
s_ 1 ,
0.
+ 2q2,
+ qp
S5
= 3pq - p 3
-p5
+ 5p3q -
let us put in our formula k

s1
But
+ pso + qs_1 =
So
Therefore
qs _1 =
+p
2,
St
5pq2.
= -1. We have
0.
= -p.
2p = - p,
S_t =
-- .
q
Likewise we get s_ 2, s_ 3, s_ 4 and s_ 5. However, we may proceed as follows

1
1
ah+t}k
sk
Sk=-+-=
=ak
t}k
(at})k
qk '
wherefrom all the desired values of s_k are readily found.
255
Solutions to Sec. 5
14. Let
Then
w4 =
a+ 4 Y a,3~ + 6 i/ a2~2 + 4 Y a,~3 + ~.
But
a
= -p,
a~
= q.
Consequently
rn 4 = -p+6Vq+4t/a~(Va+VM.
But
(lla+ -V~) 2 =a+~+2Va~= - p+2Vq,

therefore
W=V--p+6Vq+4tJ'q.V -p+2Vq.
15. Let x be the common root of the given equations.
Multiplying the first equation by A', and the second by A
and subtracting them termwise, we get
(AB' - A'B) x +AC' - A'C
0.
Likewise, multiplying the first one by B' and the second by

B and subtracting, we find
(AB' - A'B) x 2 +BC' - B'C = 0.

Take the value of x from the first obtained equality and
substitute it into the second one. Thus, we obtain the, required result.
16. Adding all the three equations termwise, we find
(x
Hence
+ y + z) 2
x+ y +z= +
a2
+ b2 + c2
V a2 + b2 + c2
Consequently
c2
z--=-====
- Va2+b2+c2
Solutions
256
17. It is obvious that the system can be rewritten in

the following way
+ z) (x + y) =
(y + z) (y + x) =
(z + x) (z + y) =
(x
a,
b,
c.
Multiplying these equations and extracting a square root

from both members of the obtained equality, we have
(x + z) (x + y) (y + z) = +
11 abc.
Hence
+ -Vabc
Y +z=
.
a
'
+ Z-- -
-Vilbc
b
+y=
;-
+ 1 abc.
c

X-'+--y
1 1)
-Vabc(1
tz-=+--+-+2
a
b
c
But since
we have
X
=+
-
V2abc
( _!_
b
+ __!.c __ _!_)
.
a
Analogously
= + -V abc (__!. + _!_ _ _!_)
Y-2
b'
simultaneously taking either pluses or minuses everywhere

18. Put
y
+x
y,
+z =
~.
Then our equations take the form

y
+~
= ay~
a+ y =bay
~+a= ca~.
+z
= a.
257
Solutions to Sec. 5
Solving this system (see Sec. 4, Problem 17), we find the

solutions of the original system
x=y=z=O
X=
~ (p~b + p~c - p~a) r
Y=
~ (p~c t p~a - p~b)'
Z=
~ (p~a+p~b-p~c)'
where
2p =a+ b
+ c.
19. Adding unity to both members of the equations,

we get
+ y + z + yz = a + 1,
+ _x + z + xz = b + 1,
+ x + y + xy = c + 1
(1 + y) (1 + z) = a + 1,
(1 + x) (1 + z) = b + 1,
(1 + y) (1 + x) = c + 1.
1
1
1
or
Multiplying these equations, we get

(1
+ x) 2 (1 + y) 2 (1 + z) 2
or
(1 x)-f1
Consequently,
-
+ a) (1 t
(1
b) (1
+ c)
+y} (1 + z) = V (1 +a) (1 + b) (1 + c).
1 + =+-./(1+b)(1-t-c)
x
1 + =+-,/(1+a)(1+c)
V
1+z=+ v(1+;~~+b).
1+a
'
1+b
'
20. MulU-ptying the given equations, we obtain

(xyz) 2 = ab ex yz.
First of all we have an obvious solution x

Then
xyz = abc.
= y = z = 0.
Solutions
258
From the original equations we find

xyz = ax2 , xyz = by 2 , xyz = cz~.
Hence
ax 2 = abc, by 2 = abc, cz 2 = abc,
x 2 = be,
y 2 = ac,
z2 = ab.
Thus, we have the following solution set
x = Vbc,
y = V ac,
x=
y=
Vbc,
x=Vbc,
z = Vab;
-Vac,
y= -Vac,
z = V ab;
z= -Vab;
-Vbc, y= Vac,
z= -Vab.
21. Adding the first two equations and subtracting the
third one, we get
2x2 = (c + b - a) xyz.
Likewise we find
2y 2 = (c + a - b) xyz, 2z 2 = (a + b - c) xyz.
Singling out the solution
x = y = z = 0,
we have
2x = (c + b - a) yz, 2y = (c + a - b) xz,
2z = (a + b - c) xy.
Then proceed as in the preceding problem.
22. The system is reduced to the form
xy + xz = a2 ,
yz + yx = b2 ,
zx + zy = c2
Adding these equations term by term, we find
X=
xy+xz
+ yz= 21 (a + b + c
2
2 ).
Taking into consideration the first three equations, we get
yz=
b2+c2-a2
a2+ c2-b2
ZX=
xy=
a2+b2-c2
259
Solutions to Sec. 5
Multiplying them, we have

(xyz) 2 =
(b2+c2--a2) (a2+ c2-b2) (a2+b2-c2)

,
8
i.e.
xyz=
-. /(b2-j- c2--a2) (a2 + c2-b2) (a2+ b2-c2)

8
+ V
Now we easily fmd

_
X-
, /(a2-r--c2-b2) (a2-j-b2-c2)
8(b2+c2-a2)
'
+ V
, /(a2+b2-c2) (b2+c2-a2)
8(a2+c2-b2)
'
Y- +
_
Z-
-, /(a2+c2-b2) (b2+c2-a2)
8 (a2+b2-c2)
.
+ V
23. Adding and subtracting the given equations termwise, we find
x3
x3
+y =
+ b) (x + y),
b) (x -
Hence
(x
+ y) (x
(x - y) (x
Thus, we
1 x +
2 x +
3 x 4 x 2 -
+b=
a (x
y)
b (x
y) = (a
y = a (x - y) - b (x - y) = (a 3
+y
+ xy + y
2 2
xy
2 -
a -
2 -
b)
+ b)
y).
= 0,
=
0.
have to consider the following systems

y = 0, x - y = O;
y = 0, x 2
xy
y2 - a
b = 0;
2
2
y = 0, x - xy
y - a - b = O;
2
xy
y - a - b = 0, x 2
xy
y2
+
+
a+
0.
The first three systems yield the following solutions

1 x=y=O;
2 x= + lla-b,
y= + Va-b;
:1 X=-=.11=+ Va+b.
260
Solutions
The last system is reduced to the following one

x 2 + y 2 = a, xy = -b.
Solving it, we get
x= ~ {e Va-2b+rt Va+2b),
y=
~ (eVa-2b-ri Va+2b),
where e and ri take on the values +1 independently of each

other. Thus, we get four more solutions.
24. Reduce the system to the following form
(x
(y
(x
+y -
z) (x
z - x) (y
z - y) (z
+
+
+z+x+y-
y) = a,
z) = b,
x) = c.
Multiplying and taking a square root, we get

(x + y - z) (x + z - y) (y + z - x) = + V abc.
Further
y+z-x=+
x+z-y=+
x+y-z=+
-./bc
a'
Vac
b'
-. /llb
v c"
Consequently
x=+
(V/"llc
T+v/lib)
c-
I
z= + (
y=
v --a+v-. ;ar;)
c-
(-./be
Vb; +Va; ).
25. Put
x+y
x+y +cxy=)',
y+z
-y+-z+'--ay-z =a,
Then the system takes the form

by + c~ = a, ca + ay
b,
a~
+ ba.
Solutions to Sec. 5
261
or
~
a
c
-+-=-.
b
a
ab
Therefore
~
'V
1. a2+b2+c2
a+b+c=2
abc
and, consequently,
b2+c2 __ a2
IX=
2bc
'
~=
a2 -l-c2-b2
0
2ac
a2+b2-c2
'
y=
2ab
Further
x+ y+cxy
x+y
y'
Finally
1
b~
-x+-;-=1-~'
wherefrom we find x, y and z.
26. Multiplying the first, second and third equations
respectively by y, z and x, we get
ex
ay
bz = 0.
Likewise, multiplying these equations by z, x and y, we
find
bx
ey
az = 0.
From these two equations (see Problem 35, Sec. 4) we obtain
+ +
_x_=_Y_=_z_=A
a2-bc
b2-ac
c2-ab
'
i.e.
x = (a 2 - be) 'A, y = (b 2 - ae) 'A, z = (e 2 - ab) 'A.

Substituting these expressions into the third equation, we
find
V=
c
_
1
(c2-ab)2-(a2-bc) (b2-ac)- a3+b3+c3-3abq
Now it is easy to find x, y and~.
262
Solutions
27. RewritP the system as follows

(y 2
(x 2
(x 2
+ (z
yz) + (z
zy) + (y
xz)
2 -
33y) = a
2 -
xy) = b
2 -
zx) = c.
Hence
x 2 - yz =
b+c-a
y 2 - xz
a+c-b
z 2 - xy =
a+b-c
i.e. we have obtained a system as in the preceding problem

28. Subtracting the equations term by term, we have
Put x
+ y + z)
z) (x + y + z)
(x - y) (x
b2
a2 ,
(x -
c2
a2
+y+z=
t, then
(x - y) t = b2 - a 2, (x - z) t = c2 - a 2.
Adding these two equations termwise, we have
[3x-(x+ Y+ z)] t = b2 +c2-2a2.
Hence
t2+ b2+ c2-2a2
X=
3t
Analogously
y=
t2.+-a2.+-c2-2b2
, 3t
t2+a2+b2-2c2
Z=
3t
Substituting these values of x, y and z in one of the equations,

we find
t4 - (a2 + b2 + c2) t2 + a4 + b4 + c4 - a2b2 - a2c2 - b2c2 = 0.
Hence
2 a2+b2+c21/3 (a+b+c) (-a+b+c) (a-b+c) (a+b-c)
t =
2
-
Knowing t, we obtain the values of x, y and z.
29. We have the following identities
(x
y + z) 2 - (x 2 + y2 + z2) = 2 (xy + xz + yz),
(x + y + z)3 - (x3 + y3 + z3) =
= 3 (x + y + z) (xy + xz + yz) - 3xyz.
263
Solutions to Sec. 5
Taking into account the second and third equations of our

system, we get from the first identity
xy
+ xz + yz
= 0.
From the second identity we have

xyz =
0.
Thus, we obtain the following solutions

x = 0, y = 0, z =
x = 0, y = a, z =
x = a, y = 0, z =
of our system
a;
O;
0.
30. Let x, y, z and u be the roots of the following fourthdegree equation

a 4 - pa 3
qa 2 - ra
t = 0.
()
Put
Then
s 4 - ps 3
But by hypothesis
+ qs
2 -
rs 1
+t=
0.
a4 , s1 = a3, s2 = a 2, s1 = a.
Therefore, the following identity must take place
a4 - pa3 + qa 2 - ra + t = 0,
i.e. the equation (*) has the root a = a, and therefore one
of the unknowns, say x, is equal to a ..
Then there must take place the equalities
s4 =
u2 + y 2 + z2 = 0,
u3 + y3 + z3 = 0,
and, consequently, (by virtue of the results of the last
problem)
u = y = z = 0.
Thus, the given s vstem has the following solutions
x = a, u = y = z = O;
y = a, x = u = z = 0:
z = a, x = y = u = O;
= a, x = y = z = 0,
+ y + z = 0,
264
Solutions
31. Equivalence of these systems follows from the identity

(a2 + b2 + c2 - 1)2 + (a'2 + b'2 + c'2 - 1)2 +
+ (a" 2 + b"2 + c" 2 - 1) 2 + 2 (aa' + bb' + cc') 2 +
+ 2 (aa" + bb" + cc") 2 + 2 (a'a" + b'b" + c'c") 2 =
= (a2 + a'2 + a"2 - 1)2 + (b2 + b'2 + b"2 - 1)2 +
+ (c 2 + c' 2 +c"2 - 1) 2 + 2 (ab+ a'b' + a"b") 2 +
+ 2 (ac + a'c' + a"c") 2 +
+ 2 (be + b' c' + b"c'') 2
It should be noted that nine coefficients: a, a', a", b,
b', b\ c, c' and c" can be (as it was established by Euler)
expressed in terms of three independent quantities p, q
and r in the following way
1+p2-q2-r2
b-2(r+pq)
2(-.,.q+pr)
a=
N
'
N
'
c=
N
'
,
2(-r+pq)
,
1-p2+q2-r2
, 2(p+qr)
a=
N
'
b=
N
' C=
N
'
b" = 2 ( -
+ rq) ,
(N=1+p2+q2+r2).
32. Multiplying the first three equalities, we get
x2y 2z2 (y + z) (x + z) (x + y) = a 3b3c3.
Using the fourth equality, we have
(y + z) (x + z) (x + y) = abc
or
x 2 (y + z) + y 2 (x + z) + z2 (x + y) + 2xyz = abc.
But adding the first three equalities, we find
x" (11 + z) + y 2 (x + z) + z2 (x + y) = a 3 + bs + c3.
Thus, finally
a8 + b8 + c3 + abc = 0.
33. Adding the three given equalities, we get
a+b+c= (y-z)(z-x) (x-y).
;I:JJ.1.
265
Solutions to Sec. 5
Similarly, we have
a-b-c=(y-~)(z+x)(x+y)'
xyz
b-c-a = (z-x) (x-f-y) (y+z)

xyz
'
c-a-b= (x-y) (y+z) (z-+ x).
xyz
Hence
(a+ b +c) (b
+c-a) (a+ c-b) (a+ b-c) =

= _ ( ; _ : )2
( : _
: ) 2, ( : _
f)=
_ a2b2c 2
Hence, we finally get the 11esult of the elimination

2b 2c2 + 2b 2a 2 + 2a2c2 - a4 - b4 - c4 + a2 b2c2
0.
34. We have
.J!..
+ _:_ = 2a,
z
y
_:_ + 3- = 2b
x
3- +
'y
J!._
x
= 2c.
Squaring these equalities and adding them, we get

~
-~ + -~ + -+
-~ + -~ + -~ + 6
~
c-=
4a2 + 4b 2 + 4c 2
On the other hand, multiplying these equalities, we find

y2
z2
z2
x2
x2
y2
zz-+ Y2+ 7 + -zz+yz-+ 7 +2 =Babe.

Consequently, the result of eliminating x, y and z from the
given system is
a 2 + b2 + c2 - 2abc = 1.
35. We have an identity
(a + b + c) (b + c -
a) (a + c -
b) (a + b - c)
= 4b2c2 - (b2 + c2 - a2)2.

Replacing in the right member a 2, b2 and c2 by their expressions in terms of x, y and z, and using the relationship
xy
we get
+ xz + yz
= 0,
266
Solutions
Thus, the actual result of eliminating x, y and z from the

given system is
(a
+ b + c) (b + c -
a) (a
+c-
b) (a
+b-
c) = O.
36. We have
(x
+ y)
= x3
=
+ y + 3xy (x + y) =
x3 + !/3 + ~ (x + y) f(x + y)2 3
(x2
+ y2)).
And so
(x
But
+ y)
x + y
3
= 3 (x
= a,
+ y) (x + y
x + y = b,
2
2) -
x1
2 (x3
+y
+y
= c.
3 ).
Consequently, the result of the elimination is

a 3 = 3ab - 2c.
37. Put
Then
a
x')..,
y')..,
(*)
z')...
On the other hand, we have
+ b + c) 2 =
(a
a2
+ b2 + c2 +
2ab + 2ac + 2bc.
Since a + b + c = 1, a 2 + b2 + c2 = 1, we obtain from

the last equality
ab + ac + be = 0.
Taking into consideration the equalities (), we find
xy
38. We have
+ xz + yz
0.
(a- :)(a-:)(a-
~)=1'
or
a 3 - (z+x+11)2
-x
-y
-z a
+ (z.x
-!J -t- -z +-Y)
ax
Hence
a~
- 1 = y.
1 ~= v.
'
267
Solutions to Sec. 5
39. From the first two equalities we find

z (d - c)
w (d - c)
+ x (d - a) + y (d + x (a - c) + y (l' -
b) = 0,}
(*)
c) = 0.
Multiplying the first equality by y, and the second by x,

and adding them, we get
(zy
+ wx) (d -
c) = x 2 (c - a)
+y
(b -
d)
+ xy (a+
c - b - d).
We find in the same way that

(zx
+ wy) (d -
c) = x 2 (a - d)
+y
zw (d - c) = x (a - d) (c - a)
2
+ y (b + xy [(a -
Ax2
d) (c -
X (b -
+ (a C = (b -
d) (a -
d) 2 (b -
+ (c -
2B
= (a
+
+
+
+ (c
d) 2 (c -
c) 2
= 0,
- d) X
a) 2
d) (c -
a - d),
+
b) + (b - d) (c - a)l.
for zy + wx, zx + wy
+ 2/Jxy + Cy
A = (c - a) (a - d) 2 (b - c) 2
+
+c-
b)
Substituting the found expressions

and zw in to the third equality, we get
where
b)
+ xy (b
d) (c -
(c -
+
a) (d -
c) (ti -
b) (b - d) 2 (c - a) 2
(b - d) (c - b) (d - c) (a c. - b - d) (a - d) 2 (b - c) 2
(b
c - a - d) (b - d) 2 (c - a) 2
(d - c) 3 (a - l;) 2
[(a - d) (c - b)
(b - d) (c - a)l (d - c) (a -
b) 2 ,
b) 2 ,
+
+
b) 2
Performing all the necessary transformations (the work can

be simplified by making use of the result of Problem 8,
268
Solutions
Sec. 2), we find

A = (a - d) 2 (c B = (d - c) (a C = (c - b) 2 (b Therefore we have
Ax2
2Bxy
Cy 2 = (c
a) 2 (c -
d),
d) (b - c) (a - c) (d d) 2 (c - d).
-
b),
d) [(a - d) (a - c) x - (b - c) (d - b) y] 2
0.
Hence
x
(b-c) (d-b)
y
(a-d) (a-c)
Substituting these values into the equality (), we get the

required proportion.
40. 1 We have
2cos a.+~ cos a.-~ -(2cos 2 a.+~
2
-1) =~2
or
4cos 2 a.t~ -4cos a.-;~ cos a.t~ +1=0.
Hence
cos a.t~ =
4cos
a.-;~ y16cos2 a.-;~
-16
Since the radicand is equal to -16sin2 a.-;~ and cos a.t~

is real, the expression -16 sin 2 ~-;~ must be greater than,
or equal to, zero. But this expression cannot exceed zero.
Therefore we have
cx-R
sin-2-" =0.
But since 0 <a< n and 0 < ~ < n, we have a = ~and,
consequently,
1
cosa= 2
and
2 Analogous to 1.
269
Solution.$ to Sec. 5
41. By hypot.hesis
0+<p
2 cos - 2-
0-cp
cos - 2-
2 sin 9
=a,
1cp cos 9 2 <p = b.
Hence
0+cp
tan-2-=-a
But
COSX=
1-tan2 .=..
2
:i;
:i;
sinx=
1+tan22
2tan 2
:i;
1+tan22
Therefore
b2
1-1.ii"
a2-b2
cos(S+<p)=
b2 =a2+b2'
1+1i2
2.!:.
sin(0+<p)=
:2
1+a2
2ab
a2+b2"
42. By hypothesis we have a cos a + b sin a = c,

a cos p + b sin p = c. Adding these equalities termwise
we find
2acos
a.tp cos a.-;P +2bsin a.tp cos a. 2 P=2c.
Hence
a.-~
cos-2- =
c
+R
a cos _a._.,_+b sin
a.+ t'
cos
a. p
(a+ btan a. t p )
Subtracting now the given equalities termwise, we obtain

. a.+P
. a.-P
a.-P
a.+P
o
- 2asm-sm
-- + 2b sm-cos-= .
2
Since a and p are different solutions of the equation, then

sin a.-;~ =I= 0. Consequently, the last equality yields
tan
a.tP =: .
270
Solutions
Let us return to computing cos 2 -a.-~

2- . We have
cos2 a.-~ =
2
cos2 a.
t~
c2
(a+ b tan a.
2 (
=c
t~)
b2 )
1 + (i2
b
a+ba
)2 =
c2
a2+h2
43. Rewl'ite the given equalities in the following way

sin e (b cos a - a cos ~) = cos e (b sin a - a sin ~),
sin e (d sin a - c sin ~) = cos e (c cos ~ - d cos a).
Eliminating 0, we find
(b cos a - a cos ~) (c cos ~ - d cos a) =
= (b sin a - a sin ~) (d sin a - e sin ~).

Hence
be cos a cos ~ - ae cos 2 ~ - bd cos 2 a + ad cos a cos ~ =
= bd sin 2 a - ad sin a sin ~ - be sin a sin ~ + ae sin 2 ~
or
(be + ad) cos a cos
Finally
+ (be + ad) sin a sin ~
bd
+ ae.
bd+ac
cos (a-~)= bc+ad
44. 1 We have
e2 -1
1+2ecosa.+e2
1+2e cos ~ + e2 _
e2-f
_
2e2+2ecos~
2e2 + 2e cos a
e+cos~
e +cos a
(by the property of proportions, from the equality : = ~

follows
+ d = ba) .
a+ c
Similarly, we have
e2-f
+ 2e cos a. + e2
1 + 2e cos ~ + e2
e2-1
-2-2e cos~
2+2e cos a.
1+e cos~
1+e cos a
271
8olutions to Sec. 5
Then
( e+cos~
e+cosa
)2 =(1-j-ecos~)2 =
e2+cos2~-1-e2cos2~
(1+ecosa;)2
e2-t-cos2a;-1-e2cos2a.
sin2~
sin2a.
Consequently,
e2 -1
1-l 2ecosa-t-e2
e cos ~
1-t-ecosa.
1+
= _
= +
-
sin ~
sin a.
2 From the given equality follows (see the result of 1)

e-t-cos ~
e+ cos a
1-t-ecos~
1-t-ecosa.
Consequently,
e +cos ~ - 1 - e cos ~ __ e +cos ~ + 1 e cos ~

e +cos a+ 1 + e cos a
e +cos a -1.- e cos a
.
a
c
( from the equality
b=d follows
a-t-c
a-c )
b+d-= b-d
Further
(1 + e) (1 +cos~)
(1-e)(1-cosa.)
(1-e) (1-cos ~)
(1 + e) (1-1- cos a)
or
(1-cosB)(1-cosa)=
g~:~~ (1+cos~)(1+cosa).
Finally
a
tan 2 tan 2 = +
1-t-e
T-=e.
45. Solving the given equation with respect to cos x,

we find
cos x (sin 2 ~ cos a - sin 2 a cos ~) =
= cos 2 a sin 2 ~ - sin 2 a cos 2 ~ = cos 2 a - cos 2 ~
But
sin 2 ~ cos a - sin 2 a cos ~ = cos a (1 - cos 2 ~) - cos ~ (1 - cos 2 a) = cos a - cos ~ +
+ cos a cos ~ (cos a - cos ~) =
= (cos a - cos ~) (1 + cos a cos ~)
therefore
cos a+ cos~
cos x = -,----__:.,1 +cos a
cos~
Solutions
272
Further
x
1 + cos a cos ~ - cos a - cos ~

i+ces a cos ~+cos a+cos ~
1-cos x
1 +cos x
tan 2 - 2 = -,----
= (1-cos a) (1-cos ~) = tan2 ~ tan21_

(1 +cos a) (1 +cos~)
and consequently
x
tan 2 =
tan 2 tan 2 .
46. We have
sin2a=4sin 2 ~ sin 2 ~ =(1-coscp)(1-cos0)=
~)
= (1_
cos~
( 1 _ cos a
) .
cos'\'
Hence
+ coscos2
a
~ cos '\'
1 - cos2 a = 1- cos a cos ~:cos '\'

cos
cos '\'
'
i.e.
cos 2 a ( 1
1
)
11.
cos I' cos '\'
= cos a
cos~+cosy
11.
cos I' cos '\'
Assuming that cos a is nonzero, we find

cos v+cos ~
1 +cosy cos~
cos a= - - - - - Now it is easy to check that

tan 2
= tan 2
tan 2
47. Put tan ~ = a, tan ~1 = ~- Then the first two equalities take the form
xa 2 - 2ya + 2a - x = 0, x~ 2 - 2y~ + 2a - x = 0.
Consequently a and
xz 2
Therefore
~
-
are the roots of the quadratic equation

2yz + 2a - x = 0.
2y
a+~=-;-,
2a-x
a~=-x-.
Furthermore
a -
~ =
2l.
273
Solutions to Sec. 5
Let us now eliminate a and

We have identically
(a
+ ~)
P from
the last three equalities.
+ 4a~.
~) 2
= (a -
Consequently,
After simplification we actually get
y2 = 2ax -
(1 -
l2) x 2
48. From the first two equalities it is obvious that 0

and cp are the roots of the equation
+ y sin a
x cos a
- 2a = 0 (unknown a).
It is clear that fl and fP are also the roots of the equation
(2a - x cos a) 2 = y 2 sin 2 a.

Transform tho last Pqnation in the following way
x2 cos 2 a - 'wx cos a + 4a 2 = y 2 (1 - cos 2 a),

(x 2 + y2) cns 2 a - 4ax cos a + 4a 2 - y 2 = 0.
Therefore the q1rn11tities cos 0 and cos cp are the roots of
the following- <'fl ua I ion
(:r 2 + y!) z2 - 4axz
4a 2 - y 2 = 0,
and therefore
('OS
We then
cos e
0 COS <f
+ cos cp =
4ax
:i:2
+Y2
h11YP
!.1 ", 112

.o
1-cos ) +cos 0 cos cp = 1. Hence, y 2 = 4a (a - x).

49. We haw
A La
tan - 2-
8-a.
tan - 2-
8
a
tan2 - - tan2 2
2
---....,8,----a.
1-tan2 2 tan2 2
274
Solutions
But
tan 2 !_= 1-cosfl = 1-cosa.cos~
2
1 + cos e
1 cos a cos ~ '
a.
__
1-cos
a
2
tan------2
1 -r cos a
Consequen Lly
COS a. COS ~
1 - CO'! a.
1 + co; a cm ~
1 -+- cos a.
___.1,...._-c_o_s_a._c-'o-s""'!3-.....,1---c-os_a._
1-
fl + a.
0 - ri
tan - 2 - tan - 2 - =
1-
+ COS a. C03 ~
+ COS a.
1-cos ~ = tan2
50.
1-t-cos~
We have
a+c
'
b-f-d =
cos (x -+- 0) cos 0

cos (x 28) cos 0
cosx+cos(x-j-20)
cos(x+0)-j-cos(.r+30)
Hence
a+c
-b-
51. We have
1 + tan 2 e = cos ~
cos a. '
cos~
. 2 (jl+ t ,m
-- .
cosy
Hence
tan2 0
tan2 <p
cos ~ - cos a.
COS a
cos y
COS~ -CO'l '\'
On the other hand, it is givPn that

tan2 0
tan2 a.
tan2 <p
tan2 y
Therefore we have
cos
cos
~-cos
a.
~-cosy
. cosy = tan2 a.
cos a.
tan2 y
From this equality we get

cos2 a. sin2 y-cos2 y sin2 a.
cos a sin2 y-sin2 a. co3 y
cos~= - - - - - - - - - -
J!.
sin2y-sin2a.
cos a. si n2 y- sin2 a. cos y
275
Solutions to Sec. 5
But
tan2!. = 1-cos ~ = cosasin2 y-sin2 a cos y-sin2 y+sin2 a =
1+cos ~
2
_
-
cos a sin2y-sin2 a cos y+sin2 y-sin2 a
sin2 a (1-c03 y)-sin2 y (1-cos a)

sin2y(1+cosa)-sin2a(1+cos.y)
__
8 sin2 ..::_ cos2 ~ sin2 _y_-8sin2 1. cos2 :L sin2..::_

2
2
2
2
2
2
8 sin2 _y_ cos2 _y_ cos2 ~-8 sin2 ~ cos2 !!:... cos2 _y_
2
2
2
2
2
2
sin2 !!:... sin2 .1.
2
2
cos2
( cos2 !!:... _ cos2 .1. )

2
tan 2 ..::_ tan 2 1.

2
cosz 1.(sin21._sin2 ~}
2
2
2
2
!:.
2'
since
52. Put
tan{=x,
tan~
=y.
Then
1-x2
1--y2
cos0= 1 +x 2 =cosacos~,
cos cp = 1+y2
=cos a 1 cos~Further
X2 =
'i-cosacos~
y2 = 1 - cos Gtt cos ~
1 -t cos a cos ~ '
1 + cos Gtt cos
'
therefore
t
2 2
(1-cosacos~)(1-cosa 1 cos~)
an 2 = x Y = (1 +co3 a cos~) (1-t cos a 1 cos~)

Add unity to both members of the equality. We find
2
_
2 (1 +cos a cos a 1 cos2 ~)
-t+cos ~ - (1 +cos a cos~) (1 +cos a 1 cos~)
Assuming cos ~ =I= 0, we obtain

cos a+ cos a 1 = 1
cos a cos 'a 1 cos2 ~.
276
Solutions
i.e.
cos a
cos a 1 = 1
cos a cos a 1 sin 2 ~ = 1
+ cos a cos a
+ cos a cos a
(1 - sin 2 ~),
cos a - cos i:x 1 =
1 = (1 - cos a) (1 - cos a 1),
1
and, consequently, indeed

sin 2
~= (cos
-1a - t)
1
(cos--1).
a1
53. We have
cos (~-a)-cos (~-y)
cos (~+y)-cos (y+a)
cos (a-~)-cos (y-a)
cos (y+a)-cos (a+~)
cos(~-y)-cos (a-~)
cos (a+~)-cos
(~+Y)
= x.
Hence
sin( ~-y)
sin(~-~)
sin ( aty
+~)
sin
(~ta
sin (
+Y)
sin(
i--a)
'\'t~
+a)
or
tan ~ - tan a+'\'
2
tan
tan y-tan
~+a
2
ta
tan a-tan
tan y+tan
tan a+tan
1'\'
----~--=
~+tan aty
~+Y
But from the equalities

a'-b'
a'+b'
a-b
a+b -
a"-b"
a"+b"
follows
a'
a"
Therefore we have
tan a
tan~
tan y
54. From the first equality we have

(tan 8 COS ~-sin~) COS a
(tan<pcosa-sina)cos~
(C03 a-tan 8 Sin a) Sin~
(cos~+tan<~sin~)sina
=0
Solutions to Sec. 5
277
Hence
sin a cos~ cos (o: - ~)"tan 0 + siu ~cos a cos (a+~) tan cp =
= 2 sin~ cos~
sin a cos a.
()
From the secoml equality we get

tan 0
cos (a. - B) tan B
tan IP
cos (a.+ B) tan a.
Therefore we may put
tan
e=
'),, cos (a -
~)
tan ~.
tan cp
-'A cos (a
Substituting the expressions for tan

equality (), we find
e and
+ ~) tan a.
tan cp into the
A = 2 sin a. sin ~
Thus
tan
cos (a. - Bl
2 sm
. a. cos ~
e=
tan cp = -
-2 (cot a
cos (a~ B) = _!_ (tan
2 cos a sm B
+ tan
a -
~ '
cot ~).
55. We have
+ sin
2 sin a sin ~cos (a - ~) =

= sin a
sin 2 ~ - 2 sin a sin ~ cos a cos ~ - 2 sin 2 a sin 2 ~ = sin 2 a - sin 2 a sin 2 ~ +
sin 2 ~ - sin 2 a sin 2 ~ - 2 sin a sin ~ cos a cos ~ =
= sin 2 a cos 2 ~
sin 2 ~ cos 2 a - 2 sin a sin ~ cos a cos ~ =
= (sin a cos ~ - cos a sin ~) 2 = sin 2 (a ~).
Therefore
sin (a - ~) =+n sin (a
~),
sin a cos ~ - cos a sin ~ = n (sin a cos ~
cos a sin ~),
tan a - tan ~ =+n (tan a
tan ~).
Finally
sin 2 a
1+n
tan a = 1 ;;; n tan ~
278
Solutions
56. Expanding the given equalities, we get

cos a cos 30 + sin a sin 30 = m cos3 0,
sin a cos 30 - cos a sin 30 = m sin3 0.
Multiplying the first equalny by cos 30, the second by
- sin 30 and adding them term by term, we find
cos a= m {cos3 0 cos 30 - sin:l 0 sin 30}.
But it is known that
cos 30 = 4 cos3 0 - 3 cos 0,
sin 30 = 3 sin El - 4 sin:i 0.
Consequently
cos3 0 cos 30 - sin3 0 sin 30 =~ 4 (cos 6 0 + sin 6 0) - 3 (sin 4 0 + cos 4 0).
But squaring the original equality and adding, we get
cos 6 0
Compute cos 4 0
COS 6
sin 6
= (cos 2 0
+ sin
+ sin
0 = _.!.._
rn2
0. We have
0 =
+ sin
+ sin
0) (cos 4 0 + sin 4 0 - cos 2 0 sin 2 0) =

= cos 4 0 + sin 4 0 - cos 2 0 sin 2 0.
Therefore
~ = (cos 2 0
rn
sin 4 0
+ cos
0) 2
3 sin 2 0 cos 2 e,
1
3 sin 2 0 cos 2 0 = 1 - -rn2
'
0 = 1 - 2 sin 2 0 cos 2 0 =
= 1--3._(1-_!.)
=_!_(1
3
rn2
3
Thus
cos a= m {4 (cos 6 0
+ sin
0) - 3 (sin 4 0
+ cos
= m { ~ _ 1 _ ~}
rn2
m2
+ rn2
~).
O)}
= 2 -m
m2
'
i.e.
m2 + m cos a = 2.
57. From the first equality we obtain
a [sin (0
+ cp)
sin (0 -
cp)l =
= b [sin (0 -
cp)
+ sin (0 + cp)].
279
Solutions to Sec. 5
Hence
a
tan cp
e.
b tan
Co nsequ en LI y
2 tan 2
-tan cp=
1 -- tan2 2
But from the second equality we have
e
tan 2
~ +c
btan
=
therefore
a
2 tan
~
a [
Putting for brevity tan

equality, we find
be (1
But
+x
2)
a2
-f = x
=
2x
r1
and transforming the last
(b 2
-+-x2-
-1
~ +
( b tan
c
1- ------.,---
+c
a 2 ) x.
2 -
=sin cp.
Finally
58. From the third equality we obtain

sin 2 e sin 2 cp = (cos e cos cp - sin ~sin y) 2
Using the first two equalities, we find
s'.n2~) ( 1 _ _;:;_'._n2y)
( 1 _ srn2
a;
s1112 a;
sin.~siny -siu~sin
sm2 a;
)2.
After some transformations this equality yields

tan 2 a = tan 2 y + tan 2 ~
59. We have
a sin 2 e b cos 2
e=
1,
a cos 2 cp
+ b sin
cp = 1.
Solutions
280
Hence
a tan 2 e + b = 1
tan 2 e, b tan 2 qi + a = 1 + tan 2
Consequently
(a - 1) tan 2 0 = 1 - b,
(b - 1) tan 2 qi = 1-a,
qi.
tan20_(1-b)2
tan2 qi 1-a
On the other hand,

tan2 0
tan2 qi
b2
li2
From the last two equalities we get (assuming that a is not
equal to b)
a+ b - 2ab = 0.
60. Rewrite the first two equalities in the following way
cos e cos a + sin 0 sin ex = a, sin e cos ~ - cos e sin ~ = b.
Multiplying first the f~rmer by sin ~ and the latter by cos ex,
and then the former by cos ~ and the latter by -sin a and
adding them, we find
sin 0 cos (a - ~) = a sin ~ + b cos a,
cos 0 cos (~ - ~) = a cos ~ - b sin a
Squaring the last two equalities and adding them, we get
cos 2 (a - ~) = a2 - 2ab sin (a - ~) + b2
61. Since
cos 3x = cos3 x - 3 sin2 x cos x,
sin 3x = -sin3 x + 3 sin x cos 2 x,
the equation takes the form
(cos 3 x - 3 sin 2 x cos x) cos3 x +
+ (-sin3 x + 3 sin x cos 2 x) sin3 x = 0,
cos6 x - 3 cos' x sin 2 x + 3 sin' x cos 2 x - sin 6 x = 0
or
(cos 2 x - sin 2 x) 3 = 0, cos 2x = 0.
62. Since
sin 2x + 1 = (sin x + cos x) 2 ,
we have
(sin x + cos x) 2 + (sin x + cos x) + cos 2 x - sin 2 x = 0.
-
281
Solutions to Sec. 5
Hence
+ cos x) (1 + 2 cos x) = 0
cos x (1 + tan x) (1 + 2 cos x) = 0.
(sin x
or
And so
tan x = -1 and cos x = -
are the required solutions of our equation.

63. We have
sin2x
1-cosx -O
cos2 x
1 - sin x --
Hence
(cos3 x-sin3 x)-(cos2 x-sin2 x) =

co32 x (1-sin x)
or
(1 - tan x) (1 - cos x)
0.
Hence
tan x
= 1 and cos
x = 1.
64. We have
cos 3ex = 4 cos3 ex - 3 cos ex.
Therefore
cos 6x = 4 cos3 2x - 3 cos 2x.
On the other hand,
cos 6 x =
( 1+cos2x
2
)3 .
The equation takes the following form

4 (1 + cos 2x) 3 - (4 cos3 2x - 3 cos 2x) = 1
or
4 cos 2 2x + 5 cos 2x + 1 = 0.
Thus
1
cos 2x = -1, cos 2x = - 4 .
65. We have
sin 2x cos x + cos 2x sin x + sin 2.c - m sin x = 0.
Hence
sin x [2 cos 2 x
sin x f4 cos 2 x
+ cos 2x + 2 cos x + 2 cos x - (m + 1)1
ml = 0,
= 0.
282
Solutions
And so, one solution is

sin x = 0.
The other is obtained by the formula
- 1
COS X=
-v4 .r,n+5
Hence, first of all, it follows that there must be
4m
+ 5 ~ 0.
Further, for one of the roots to exist it is required that
V 4m + 5 / :(; 4, i.e. that -4 :(; -1 + V 4m + 5 :(;

4 or-3 <V4m + 5 :(; 5, i.e. m :(; 5. For t 1e other
root to exist it is necessary that
I -1 +
<+
V 4m + 5 I :(; 4,
I -1 -
-4 :(; -1 -
V 4m + 5:(; 4,
m :(; 1
Thus if m
< - : , then cos x has no
it has one real value (cos x = -
real values; at m = - :
! );for -
~<
m :(; 1 cos x
4tn + 5 )
has two real va 1ues ( cos x = - 1-+ 11
and for 1
4
<
m :(;
5 cos x again has one real value (cos x
= - i+
<
Ym+5')
and at m > 5 it has no real values.

66. Rewrite the equation as
4
-c08-,(_x ___a_)_ { ( 1 + k) cos x cos (2.c - a) -
-(1+kcos2c) cos (.c- :x)} = 0.

But
1
cos x cos (2x-a) = 2 cos (3x-a) + 2 cos (x-a),

cos 2x cos (x-a)
=z-1 cos (3x-a) + 21 cos (x +a).
Solutions to Sec. 5
283
Therefore
-c-o;,-,(-x---a-)-{(1 + k) fcos (3x-a) +cos (x-a))-
-2 cos (x-rx)-k [cos (3x- a)+ cos (x+ a)]}= 0

or
cos (x-a)
{cos (3x-a)- cos (x-a)
+ k [cos (x-a)-cos (x +a)l} =

sin :r
-c0 -8 (-.r---a-)-
0,
{k sin a-sin (2x-a)} ~ 0.
Hence
sin x
0 and sin (2x - a)
k sin a.
x = 1, we have sin 4 x + cos 4 x +

+ 2 sin 2 x cos 2 x = 1 and sin 4 x cos 4 x = 1 - ; (sin 2x) 2
The equation takes the following form
sin 2 2x - 8 sin 2x + 4 = 0.
Hence
sin 2x = 4 + V16-4, sin 2x = 4 + 2l/3.
67. Since sin2 x
+ cos
Rejecting one of the solutions, we get frnally

sin 2x
= 4-
2 V3.
68. We have
1
logx a= 1oga x
log ax a= 1oga ax
log,,2xa =

2
logax
logax+1
+ logax+2
O.
Put
loga x = z.
Finally, we have to solve the following equation
Hence
z2+
1 +
z+1
3 -0
z+2
(jz2+ 11z+4
.z (z+1) (z+2)
0.
1oga a2 x
284
Solutions
The required roots are
Thus
69. We have
a
x=yx+y.
Hence
yx+y = yx+y.
.
4a2
Consequently, either y = 1 or x + y = -.~-

But at y = 1
x ,-y
x 4 a = 1 and, consequently, x = 1. Thus, we get one solution
x=1,
y=1.
Let us now find a second solution. We have
+ y)
x +y
(x
i.e.
Therefore
x2a=ya,
4a2 ,
=
=
2a.
( xy2
)a=f,
and consequently
x2 = y,
i.e.
x 2 = 2a - x.
From this quadratic equation we fmd
1 + i/14+2a.
x= -2
The positive solution is

1
I/ 1 2
x"""'-2+ V t;+ a.
The corresponding value of y is found by the formula
= x2.
285
Solutions to Sec. 6
70. Raising the first equation to the power q and the

second to p, we obLain
Dividing one of these equalities by the other termwise, we

find
and consequently
p11-qx
v =a p2-q2.
Substituting these expressions for u and v into the third and

fourth equations, we have
aP(x2+y2)-2xyq = bP2 -q2,
a2xyp-q(x2+y2) = Cp2-q2
Hence
p (x2 + y2) 2xyp -
q (x 2
2xyq = (p2 -
+y
2)
(p 2
q2) loga b,
q2) loga c.
Consequently
x 2 + y 2 = p loga b + q loga c,
2xy = q loga b + p loga c;
wherefrom we find x and y, and then u and v using the formulas (*).
1. Let x = a
+ ~i,
y = y
fli. Then
x + y = a + y + (~ + fl) i, x - y = a - y + (~ - fl) i,
Ix + y 12 + Ix - y 12 = (a + y) 2 + (~ + fl) 2 +
+ (a - y)2 + (~ - fl)2 =
= 2 (a2 + ~2) + 2 (y2 + fl2) = 2 {lxl2 +IYl2}.
28G
Solutions
x=
2. Let x =-= a
~i, hence
1 By hypothesis,
a -
~i
(2a
+ 1)
= a2
~2
Hence
Therefore
~
0,
+ 2a~i.
= a2
~i.
a -
~2.
Assume first ~ = 0, a = a or a (a - 1) = 0. And so,

first of all we have the following solutions
a = O; ~ = 0, x = O;
a = 1, ~ = 0, x = 1.
Let us now pass over to the case when 2a
1 = 0, i.e.
2
A2-~
V3
~=+-2-,
p-4'
i.e.
1
V3
X=-2+i-2-'
-v3
X=-2-i-2-.
Consequently, there exist four complex values of x satisfying the condition

namely
x=O,
X=
1,
2 Let us solve the following system

a (a 2 - 3~ 2 - 1) = 0, ~ (3a 2 - ~ 2
We find the following solutions
a= 0,
~ = O;
a= 0,
~ = +1;
a= +1,
~ = 0.
And so
x = 0, x = +1, x = +i.
3. Put
+ 1)
a1 + b1i = x,
ll2
+ b2i
y, ... ,
lln-1
= 0.
+ bn-li = u,
an+ bni =
W.
~olutions
287
to Sec. 5
Then the inequality to be proved may he rewritten as
lx+y+ ...
+u+wl:(;
<
Ix
I + I y I + ... + I u I + I w I,
i.e. we have to prove that the modulus of a sum of several

complex numbers is less than or equal to the sum of moduli
of the addends. Let us first prove this for two addends, i.e.
let us prove that
I x + y I < I x I +I y 1-
But
Ix+ YI= V(a1 + a2) 2
Ix I=
va;+ b; '
+ (b1 +b2)
Iy J =
2,
v a;+ b: .
Consequently, it is required to prove that
V(a1 + a2 )i (b1 + b2) 2 :(;Va; + b: +-Va~+ b:.

On squaring both members of this inequality and after
some simplifications we get an equivalent inequality
a1a2
+ b1b2<V (a~+ b~) (ai+ b;).
This inequality is undoubtedly true if
(a1a2
+ b1b2) 2 :(;(a; + b;} (a: + b;),
i.e. if
(a 1a 2
b1 b2) 2
(a~
bi) (a:
b~) :(; 0,
(a1b2 -
a2b1) 2
:(;
0,
which is obvious. Thus, it is proved that
lx+yl:(;lxl+ IYI
for any complex .r and y. To prove our proposition for the
general case proceed as follows. We have
lx+y+z+ ... +u-J-wl=
= I (x + y + . . . t- u) + w I < I x + y + ... +
+ u I+ I w 1.
Let us now apply an analogous operation to the first term
Ix+ y + ... + u 1.
288
Solutions
Continuing this operation, we shall prove our proposition

for the case of n terms. The above proof was carried out
by the method of mathematical induction. Let us add to it
another proof. Suppose the complex numbers are reduced to
the trigonometric form, i.e. put
x =Pi (cos cp 1
y = p2 (cos cp2
We then have
+- i sin cp2 ),
+ i sin cp
1),
w = Pn (cos (jln
. ,
x+ y+
... + w =
+ i sin cp,,).
Pk
k=I
cos cpk + i ~
Pk
k=I
sin cpk,
lxl+IYI+ +lwl= k=I

~Pk,
2
I x + y + ... + w 12= ( ~
Pk
cos cpk)
k=I
+(~
Pk
sin cpk)
k=I
It is required to prove that

n
!'!. =
~ Pk) -( ~ Pn cos cpk) -( ~ Pn sin cpk) >O.
k=I
k=I
k=I
we have
n
k=I
( ~ P1< cos cpk)
( '2j Pk)
~ P~ + 2 ~ PsP1,
11.=1
~ p~ cos 2 (jl1<
k=I
k=I
k=I
k=I
s4=t
+ 2 *t
~ PsPt cos (jls cos cp,,
( ~ P1< sin cpk) 2 = ~ pf. sin 2 <pk+ 2 ~ PsPt sin (jls sin cp,,
s*t
consequently
!'!. = 2 ~ PsPt -2 ~ PsPt cos (cps - cp, ),

s'i=t
4'-t
4. Proved by a direct check, taking into consideration

that e. 2 = - e. - 1, e.3 = 1.
289
Solutions to Sec. 6
5. IL is obvious Lhat
a
+b
+c
ac - be ~"'
=(a
eb
xy - xz - yz =
ab -
2 -
+y +z
x2
2 -
= (x
Therefore
(a 2
+ c2
+ e c) (a + e b + ec),
+ e2z)
ey
(x
+ e2y + ez).
ab - ac - be) (x 2
y 2 + z2 - xy - xz - yz) =[(ax+ cy
bz)
b2
+- +
+ (ex + by + az) e + (b.r + ay + cz) e x
X [(ax+ cy + bz) + (ex + by + az) e +
+ (bx + ay + cz) el =
= X + Y + V - XY - XZ - YZ,
where
X =ax+ cy + bz, Y =--=ex+ by + az,
Z = b.r + ay + cz.
2]
6. 1 Solving the given system with respect to x, y and z,

we get
A+B+c
3
X=
Y=
'
2 We have
I A 12 + IB 12 + I C 12
But
AA =
(x
BB = (x
'
;3
Z=
AA + BB +-
CC.
+ y + z) =
Ix 1 + I y 12 + I z 12 + x (y + z) +
+ y + z)
=
A+R+Ce2
A+JIE2-j-C
(x
ye
+ y (x + z) + z (x + y),
ze 2 )
(x
+ ye + ze)
2
+ x (ye + Ze +
+ y (xe + Ze) + z(xe + ye
2
cc = (x + ye + ze) (x + ye + Ze =
= lx\ + IYl + 1z\ + x (ye + ze) +
-+ y (xs +- zc: + z (xe + ye).
=
x \2
12
12
2)
2 ),
2)
2)
Solution.~
Adding the three equalities term by term, we find
I A 12 + I B 12 + I C 12 = AA + BB + CC =
= 3 [Ix 12 + I y 12 + I z 12 1 + x [y (1 + e + e2)
+ z (1 + e + e)l + y [x (1 + e + e) +
+ z (1 + e + e + z [x (1 + e + e +
+ y (1 + e + e)l.
since 1 + e + e = 0, the last three expressions in
2
2 )]
2)
2
But
square brackets are equal to zero and
I A 12 + IB 12 + I Cl 2 = 3 fIx 12 + I Y 12 + I z 12 1.
7, On the basis of the result obtained in 1 of Problem 6,
we have
,,
X=
AA'+BB'+CC'
,,
AA'+BB'e2+CC'e
3
'
Y=
3
'
,,
_
AA'
+Bfl'i::.-~
CC'e2
Z
3
.
Further
+ BB' + CC' = (x + y + z) (x' + y' + z') +

+ (x +ye+ ze (x' + y'e + z'e +
+ (x + ye + ze) (x' + y'e + z'e) =
= 3 (xx' + zy' + yz').
And so x" = xx' + zy' + yz'. Analogously y" = yy' +
+ xz' + zx', z" = zz' + yx' + xy' (the last two expresAA'
2)
2)
sions emerge from the first one as a result of a drcular permutation).

8. Though this formula was already proved (see Problem 2, Sec. 1), we are going to demonstrate here another
proof, using this time complex numbers.
We have the identity
(a6 - ~y) (a'6' - Wy') = (aa'
~y') (yW
M') - (aW
~6') (ya'
6y'),
let us put here
a
a'
= x
= a
-j- yi,
= z
+ bi, W =
+ ti,
+ di,
y = -(z - ti), 6 = x - yi.

y' = -(c - di), 6' = a - bi.
.~olutions
Then
= x2 -1- y2 + z2 + t2'
a' 6' - w
v' = a2 + b2 + c2 + d2'
~v
ao -
+ ~y'
aa'
2~1
to SPc. 6
= (a.r, - by - cz _:___ dt)
vW + M' =
~v'
+ aa'
+
+ i( bx + ay + dz
= (aa' + ~v').
- ct),
Therefore
(aa'
~v')
(vW + M')
by - cz - dt) 2 +
+ (bx + ay + dz - ct) 2 ,
(ax -
Further
aw + ~6'
ya'
+ 6y'
+ az + bt) +
+ i (dx + cy dy + az + bt) +
-i i (dx + cy -
(ex - dy
= -(ex -
bz
+ at),
bz
+ at),
i.e.
-(aW
~6') (ya'
+ 6y')
+ az + bt) +
+ (dx + cy - bz + at)
(ex - dy
Substituting the obtained expressions into the original

identity, we find
(a2
+ b2 + c2 + d2)
= (ax -
+ (ex
+ y2 + z2 + t2) =
by - cz - dt) 2 + (bx + ay + dz - ct) 2 +
dy + az + bt) 2 + (dx + cy - bz + at) 2.
(x2
Replacing in it d by -d and t by -t, we get the required

identity.
9. Expand the expression (cos <p
i sin cp)n, by the
binomial formula. We have
(cos <p
+ i sin <p)n
= cosn <p
+ n cosn-l <p i sin <p +
cosn-2 <p (i sin <p)2 +

+ n (n-1)
12
X
(i sin <p) 3
ll
(n-1) (n-2)
1.2.3
cosn-3 m
~
+ ... + n cos <:p (i sin <p)n-i + (i sin cpt.
292
Solutions
Separating the real part from the imaginary one in this

expansion, and using de Moivre's formula, we find
cos ncp
+ i sin ncp =
( cosn cp-
(7.-; 1) cosn- cp sin cp + ... )+

2
+ i ( n cosn-l cp sin cp- n (n-;.1J.~_- 2) cos'i-a cp sin cp + ... } .

3
Hence
sin ncp =
n (n-1)
... ,
n cosn-l cp sin cp- n (n-;.1~_(;- 2 ) cosn- 3 cp sin 3 cp + ...
n
cos ncp = cos cp -
1. 2
cos
n- 2
cp sm 2 cp +
Taking into account the parity of n and dividing both

members of these equalities by cosn cp, we get the required
formulas.
10. First prove case 1. We have
cos (jl=
Put cos cp
(cos qi+ i sin qi)+ (cos qi- i sin qi)

2
+ i sin cp =
Then cos cp- i sin cp =
e.
1::- 1 ,
Further
In the second sum put m-k= -(m-k'). Then this

sum is rewritten in the following manner.
0
"1
LJ
m-1
c2m-k' -2(m-k')
2m
k'=m-1
"1 ck
LJ
k=O
2me
-2(m-k)
And so
m-1
vm cos2m cp = ~ C~m
(1::2<m-k)
e-2cm-k))
+ C2m.
k=O
However,
1::2(m-k)
1::-2(m-k)
= 2 cos 2 (m -k).
293
Solutions to Sec. 6
Therefore,
m-1
22m-cos 2m cp =
k=O
2C~m cos 2 (m-k) cp
+ C~m
Replacing in thi.s formula cp by-~ - cp, we get formula 2.

Formulas 3 and 4 are deduced as 1 and 2.
11. Form the expression
Un + ivn = (cos a + i sin a) +
+ r [cos (a + 0) + i sin (a + 0)1 + . . . +
+ rn [cos (a + n0) + i sin (a + n0)] =
= (cos a + i sin a) {1 + r (cos 0 + i sin 0) + . . . +
+ rn (cos n0 + i sin n0) }.
Put
cos 0 + i sin 0 = e.
Then
Un+ ivn =(cos a.+ i sin a) {1 +re+ ... + (re)n}=
(re)n+1 -1
= (cos a + i sin a) rl:l- 1 .
(re)n+1-1
, separating the
Let us transform the fraction
re- 1
real part from the imaginary one.
We have
(re)n+1_ 1
re-1
[(re) 1i+1_11 [re-1)

(re-1) (re-1)
= rn+2 [cos n8+i sin n8]-r [cos8-isin 8]
1 - 2r cos 8 + r2
+--rn+l [cos (n+1) 8+isin (n-t-1) 8]+1
1 - 2r cos 8 + r2
Multiplying the last fraction by cos a + i sin a and separating the real and imaginary parts, we get the required
result
rn+ 2 [cos(n8+a)+isin(n8+a)] +
.
u n + iv
= ----"--...,..--.....,;...----'-=--~--'--~
n
1 - 2r cos 8 + r2
+ - r [cos (a-8)+i sin (a-8)] ,
-r
1 - 2r cos 8 ,.2
+ ~rn+l {cos [(n+ 1) 8+aJ+i sin [(n+ 1)"8+alf +cos a+i sin a
1 - 2r cos 8 + r2
294
Solutions
Hence
"-
cos a-r cos (a-8)-rn+l cos [(n + 1) 8+ a]+ rn+2 cos (n8+ a)
1 - 2r cos 8 + r2
'
Vn=
sin a-r sin (a-8)-rn+l sin [(n + 1) 8+ a]+ rn+z sin (110 +a)
1-2rcos8+r2
--~~~~---'-~---'-~~----=-~----=-'--,,-'~;:_;_~~----=--'--'----'
Putting in these formulas a= 0, r = 1, we find

. n+1
S!Il -
2- 8
1 + cos 0 + cos 20 + ... + cos n0 =
sm
sin 0 + sin 20 + ... + sin n0 =
COS
. (n+1)8 . n8
sm - - s1n2
2. e
S!Il
12. We have
n
S + S'i = ~
c! (cos k0 +
i sin k0)
~ C~ (cos 0 + i sin 0)h =
k=O
k=O,
= (1+cos0 + i sin 0)" = [ 2 cos 2
+ 2i sin
cos
+ ism2
. . 8 )" =
. 2n8 ) .
cos n 28 ( cos Tn8 + . sm
2nCOSn 28 ( cos 20
2n
Hence
S
2n cos n 8 cos n8
2
2
S'
2n cos" ~ sin n28
13. Put
n
sin 2 P a+ sin2P 2a + ... + sin 2 P na = ~ sin 2 l? la.

1=1
But (see Problem 10)

p-1
sin 2P
la= 22 :
1_
1(
-1)P
~ ( -1)" C~pcos 2 (p-k) la+
2!P
C~p,
k=O
therefore
p-1
( - 1)P ~
11 h '1
2
k
n CP
S = 22p-t Li ( -1) C2p Li cos (p- ) la+ 22 P 2p
k=O
/=1
295
Solutions to Sec. 6
Put 2 (p-k) a=. Then

.
n-\--1
nt
2 S
Sill - " COS - -
~ cos 2 (p- k) la= cos+ ... +cos n=
1~ 1
sin
(see the solulio11 of Problem 11).

Let us de no le
.
ll~
n+1
Sill - " COS - -
---~---
sin
CT1t.
Then we can prove that a,, = 0 if k is of the same parity as

p {k = p (mod 2)} and a,, = -1 if k and p are of different
parity {k = p + 1 (mod 2) }, and we get
p-1
S=(-1)P+1
1 CP
(-1) "C"2p+n 221
,
2p
22r-1
k=o
1t==r+1 (mocl 2)
Hence
p-1
k
C2p
+n
22p
er2r
k=O
k==p+I (mod 2)
p-1
But we can prove that
n=o
C~p = 2 2 p-z (see Pro-
k==P+l (mod 2)
blem 58 of this section) and our formula is deduced.

14. 1 Rewrite the polynomial as
xn-an-nx'an-l +nan= (xn - an)- na"- 1 (x-a) =
=-- ( x
- a) (x"- 1 + axn-l -+
... + a"-
1 -
na1.-i).
At x = a the second factor of the last product vanishes and,

consequently, is divisible by x - a; therefore the given
polynomial is divisible by (x - a) 2
2 Let us denote the polynomial by P n and set up the
difference P 11 - Pn-l Transforming this difference, we
easily prove that it is divisible by (1 - x) 3 Since it is true
296
Solutions
for any positive n, we obtain a number of equalities
Pn - Pn-1 = (1 - x) 3 cpn (x),

Pn-1 - Pn-2 = (1 - x) 3 (jln-1 (x),
P 3 - P 2 = (1 - x) 3 cp 2 (x),
P2 - P1 = (1 - x) 3 cp 1 (x),
where cp 1 (x) are polynomials with respect to x.
Hence
Pn - P1 = (1 - x) 3 'ljJ (x).
But since
P 1 = (1 - x)3,
it follows that Pn is divisible by (1 - x) 3 and our proposition is proved.
15. 1 Considering the given expression as a polynomial

in y, let us put y = 0. We see that at y = 0 the polynomial
vanishes (for any x). Therefore our polynomial is divisible
by y. Since it is symmetrical both with respect to x and y
(remains unchanged on permutation of these letters), it is
divisible by x as well. Thus, the polynomial is divisible
by xy. To prove that it is divisible by x
y, let us put in
it y = -x. It is evident that for odd n we have
(x - x)n - xn - (-x)n = 0.
Consequently, our polynomial is divisible by x
y. It only
remains to prove the divisibility of the polynomial by
x2
xy
y 2 = (y - xe) (y - xe 2),
where
e2
e
1 = 0.
For this purpose it only remains to replace y first by xe
and then by xe 2 and to make sure that with these substitutions the polynomial vanishes. Since, by hypothesis, n is not
divisible by three, it follows that n = 3l
1 or 3l
2.
At y = xe the polynomial attains the following value
(x+ xe)n-x"-(xe)n = xn {e2 n+ 1 + e"} = xn (1 + e+ e2 )=0.
Likewise we prove that at y = xe 2 the polynomial vanishes
y) x
as well, and, consequently, its divisibility by xy (x
X (x 2
xy
y 2 ) is proved.
+ +
Solutions to Sec. 6
297
2 To prove this statement let us proceed as follows.

Let the quantities -x, -y and x + y be the roots of a
cubic equation
a 3 - ra 2 - pa - q = 0.
Then, by virtue of the known relations between the roots
of an equation and its coefficients (see the beginning of
this section), we have
r = -x - y + (x + y) = 0, -p = xy - x (x + y) - y (x + y),
q = xy (x + y).
Thus, -x, -y and x + y are the roots of the following
equation
a 3 - pa - q = 0,
where
p = x 2 + xy + y 2 , q = xy (x + y).
Put
(-x)n+(-yt+(x f-yf"=S11.
Between successive values of Sn there exist the following
relationships
Sn +3 = pS n+l + qS n,
S 1 being equal to zero. Let us prove that Sn is divisible by
p 2 if n = 1 (mod 6) using the method of mathematical
induction. Suppose S 11 is divisible by p 2 and prove that
then Sn+e is also divisible by p 2 We have
Sn+6 = pSnH +qSn+3
Therefore
Sn+G = p (pSn+2
sllH
= pSn+2+ qSn+I
+ qSn+1) + q (pSn+I + qSn) =

= p2Sn+2 + 2pqSn+t + q2Sn.
Since, by supposition, S 11 is divisible by p 2 , it suffices to

prove that Sn+t is divisible by p. Thus, we only have to
prove that
( X + Yt
x)" + ( - Y)n
+ (+ xy + y if
3
is divisible by x 2
n = 2 (mod 6). Proceeding
in the same way as in 1, we easily prove our assertion. And
so, assuming that s)l is divisible by p 2 , we have proved
that Sn+e is also divisible by p 2 But S 1 = 0 is divisible
298
Solutions
by p 2 Consequently,
Sn=(x+y)n-Xn -yn
is divisible by (x 2
xy
y 2) at any n
1 (mod 6). lt
only remains to prove its divisibility by x
y and by xy.
16. Equality 1 is obvious. From Problem 15 it follows
that (x
y) 5 - x 5 - y 5 is divisible by xy (x
y) (x 2
xy
y2). Since both the polynomials (x
y) 5 - x 5 - y 5
and xy (x
y) (x 2
xy
y 2 ) are homogeneous with
respect to x and y of one and the same power, the quotient
of division (x
y) 5 - x 5 - y 5 by xy (x
y)(x 2,
xy
y2)
will be a certain quantity independent of x and y. Let us
denote it hr A. We then have
(x
y)5 - x5 - y5 = Ay (x
y) (x2
xy
y2).
Since this equality represents an identity and, hence,

holds for all values of x and y, let us put here, for instance,
x = 1, y = 1. We get
25
1 - 1 = A 2 3.
Hence A = 5, and we finally get

(x
y)5 - x5 - y5 = 5xy (x
+ y) (x2 + xy + y2).
Using the result of Problem 15 (2), we can write similarly

(x
y)7 - x 7 - y 7 = Axy (x
y) (x 2
xy
y 2) 2.
Putting h~,r~e x = y = 1, we find A = 7.

17. It is known that
(x
+ y + z)
3 -
x3
y3
z3
3 (x
+ y) (x + z) (y + z).
+ +
Let us prove that (x

y
z)m - X"' - ym - zm is divisible by x
y. Considering our polynomial rearranged
in powers of x, we put in it x = -y. We have
(-y
y
z)m - (-y)m - ym - zm -= 0,
+ +
since mis odJ.

Consequently, our polynomial is divisible by (x
y).
Likewise we make sure that it is divisible by (x
z) and
by (y
z).
.
18. The condition necessary and sufficient for a polynomial f (x) to be divisible by x - a consists in that f (a) =
299
Solutions to Se.:. 6
=
0. Put
f (x) = x3
+ kyzx
y3
z3
For this polynomial to be divisible by x
is necessary and sufficient that
f (-y - z)
+y+z
it
0.
However
z) + y3 + z3 =
3) yz (y
z),
= -(k
wherefrorn follows k = -3. Thus, for x3 + y3 + z3 +
+ kxyz to be divisible by x + y + z it is necessary and
sufficient that k = -3.
19. Divide n by p. We get n = lp
r, where l is a positive integer and 0
r < p. Consequently,
xn - an"= x 1Pxr - a1par = x 1pxr -a 1Pxr a1pxr -a 1Par =.
= xr (xlP _alp)+ alP (xr -ar).
f (-y - z)
-(y
z) 3
kyz (y
+
+
<
But xlP - a1P = (xP)l - (aP) 1 is divisible by xP - aP,

therefore for the di visibility xn - an by xP - aP it is
necessary and sufficient that xr - ar is divisible by xP - aP.
But it is possible only when r = 0, and, consequently,
n = lp. Finally, for xn - an to be divisible by xP - aP it
is necessary and sufficient that n is divisible by p.
20. Put f (x) = x 4a + x 4 b+I + x 40 + 2 + x 4d+3 , On the other
hand,
x3
x2
+x+
1 = (x
1) (x 2
1)
1) (x
f(-1) =f(i) =f(-i) =0.
21. We have
=
1 + xi+ x4 +
(x
. .
... + x2n-2
=
x 2 "-1
x2 -1
+ i)
(x -
i).
'
xn-1
1 +.c+ x~+ ... -\- xn- = --

x-- 1
'
x2n_1
xn-1
It is required to frnd out at what n x2 _ 1 : x-i

be a polynomial in x. We fmd
x2n-1
:rn-J
xn+t
x2 -1 : --:x=-1 = x 1 '
will
300
Solutions
+
+
+ + ... +
+ ... +
For xn
1 to be divisible by x
1 it is necessary and
1 = 0, i.e. that n is odd.
sufficient that .(-1)n
Thus, 1
x2
x 2n- 2 is divisible by 1
x
x2
xn-l if n is odd.
22. 1 Put
+ +
+ x sin cp) cos ncp But

x + 1 = (x + i) (x - i)
= (cos cp + i sin cp)n - (cos ncp + i sin ncp)
f (x)
(cos cp
11 -
x sin ncp.
and
f (i) =
= 0 (by de Moivre's formula). Likewise we make sure that f (-i) = 0, and
2
our supposition is proved.

2 Resolve the polynomial x 2 - 2px cos cp
p2 into
factors linear in x. For this purpose find the roots of the
quadratic equation
x 2 - 2p x cos cp
p2 = 0.
We get
x = p cos cp
+ V p2 cos 2 qi-p2 = p (cos cp + i sin cp).
Let us denote
xn sin cp- pn- 1x sin ncp
+ pn sin (n -1) cp = f (x).
We have to prove that
f [p(cos cp + i sin cp) = 0.

23. Suppose
x 4 + 1 = (x 2 + px + q) (x 2 + p' x + q') =
= x4 + (p + p') x3 + (q + q' + pp') x2 +
+ (pq' + qp') x
+ qq'.
For determining p, q, p' and q' we have four equations
P+ p' =0,
pp' -J-q+q'=O,
pq' +qp' =0,

qq' = 1.
(1)
(2)
(3)
(4)
From (1) and (3) we frncl p' = -p, p (q' - q) = 0.

1 Assumep = 0, p' = 0, q
q' = 0, qq' = 1, q2 =-1,
q = i, q' = +i.
301
Solutions to Sec. 6
The correspon<ling factorization has the form
(x 2 + i) (x 2 - i).
20 q' = q' q2 = 1, q = + 1.
Suppose first q' = q = 1. Then pp' = -2, p + p' = 0,
p 2 = 2, p = +Y2, p' = +Y2 The corrPsponding factorization is
x4
+1=
x 4 + 1 = (x 2
Assume then
q = q'
V2x +
+ p'
-1, p
1) (x 2
+ V2x +
0, pp' = 2, p
1).
+V2i,
p' = +V2i.
The factorization will be
x'1 + 1 = (x 2
24. Put.
+ V"2ix
- 1) (x 2
--
V"2ix - 1).
-va+bi = x + yi,
whence
a
+ bi
x2
+ 2xyi;
y2
consequently,
x2
y2
~=
a,
2xy
b.
To li.nd x and y it only remains to solve this system of

two equations in two unknowns.
We have
(x2
+ y2)2 = (x2-y2)2 + 4x2y2 = a2 + b2,
therefore
x2 = a+ V a2 +bi'
x= +Va+ Va2+b2,
y2 =
+{/';
x2 + y2 = V a2
a+ V a2 + b2'
y= +V -a+ Y a2+b2,
the signs of the roots being related as 2.ry = b. And so,

the following formula takes place
V a +bi =
+ {V a+ V a 2 + b2 + i
>
V - a+ V a + b2)
2
if b 0 (since then the signs of x and y must be the

same), and
Va+ bi =
if b
< 0.
+ {Va +V a2 + /J 2 -
V - a +-V a + b
2
2)
.~olutions
302
25. Tho roots of tho given equation are determined by

the formula
Xk=
2kn
. .
2kn
cos--+
ism--=
n
n
= ( cos -2n
n
+. sm. -2n }k
i
(k=O, 1,
n-1).
26. We have
where
. . 2n .
e = cos -2n+
i sm n
n
Thus
n-1
~ x~= 1 + eP
11=0
But
+ e 2P +
, .. +
e<n-1JP,
2pn +. . 2pn
e P =cos - i sm -n- .
n
It is obvious that
In this case
eP
= 1 if and only if pis divisible by n.

S=n.
And
if
enP-1
FP-f
eP
=;id,
then
s = 1+
eP
e2P
+ ... +
= 0, smce enP = 1.
Thus
n-1
2.; xf = n
k=O
if p is divisible by n,
and
n-1
~ xf = 0 if p is not divisible by n.
h=O
27. We have
n-1
n-1
2 = ~ AhA,,.
~I A1t 1
k=O
k=O
e<n-1)7> =
303
Solutions to Sec. 6
But
AkAk= (x+ye 11 + ze211 + ... +we<n- 1Jk)
x (x+ ye-k+ze-2k+ ... + we-(n-1)k) =

=xx+yy-t ... +ww+x(ye-k+ze-zk+ ...
+iDe-<n-1Jk)+ye1i (x+:Ze-2k+ ... +iVe-<n-1)k)+

+ ze2k (x + ye-k + ... + we-<n-1)k) +
+we<n-1Jk(x+ye-k+ ... +ile-<n-2>k).
Therefore
n-1
~ AkAk = n (I x 12 + IY 12 + + I w 12 ) +
k=O
n-1
+ x ~ (ye-k + :Ze-2k + ... + we-<n-1)k) +
k=O
+y
n-1
~ (xek
k=O
+w
n-1
n-1
~
k=1
+ :Ze-k + ... + we-(n-2)k) +
... +
(xe<n-1)k + ye<n-2)k + ... + ue,.).
But ~ e111 = 0 if l is not divisible by n (see Problem 26).

k=O
Therefore all the sums in the right member vanish and

we get
IAo 12 +IAi1 2 + +I An-il 2 = n {IX 12 +IY1 2 + +I W 12 }.

28. 1 Denote the roots of index 2n from unity by x 8
so that
X 8 =COS
Therefore
2n
2sn: + i.Sln
. --;:-2sn
-n-
n-1
(S
= 12
, , ... , 2)
n .
2n-1
304
Solutions
since
Xn
= -1,
x 211 =1. But x 2 n-s
=X
consequently,
8,
n-1
x2 n-1=(x 2 -1) II (x-xs)(x-xs)=

S=1
n-1
(x2 -
1)
TI (x
2 -
2x cos 51~
+ 1) .
s=1
The rest of the cases are proved similarly.

29. 1 Rewrite the equality 1 of the preceding proLkm
in the following way
n-1
x2n-2 + x2n-4 + ... + x 2 + 1 =
II (x
-2x cos
s: + 1) .
.s=1
Put in this identity x = 1. We have

n-1
n-1
S=1
S=1
n-=TI (2-2 cos s: )=II 4 sin s:

2
2 n
. ~ 2n
.
(n-1):11
22(n-1) sm
-Sm"... sm 2 - -n- - .
n
n
Hence
.
2n
sm-nsm -n- ... sin
(n-1) n
n
=
Vn
211 _ 1
2 Solved analogous! y to 1.
30. We have
x"-1
=(x-1)(x-a)(x-~)(x-y)
... (x-A.).
Hence
x11 - 1 +x11 - 2 + ...

Consequently
+x+1=(x-a)(x-~)
... (x-A.).
(1-a)(1-~)
... (1-A)=n.
31. Set up an equation whose roots are
x,-1, X2-1,
... , Xn-1.
This equation has the form
(x+1t+(x+1)''-1 + ... +(x+1)+1=0,
305
Solutions to Sec. 6
i.e.
(x+1)n+ 1 -1 _ (x+1)n+ 1 -1 -O
x+i-1
x
Then set up an equation with the roots

1
X1-1 '
1
X2-1 '
1
.,
Xn-1
It has the form

1
( x-+1
)n+t -1
(1 + x)n+l-xn+I
--'---'-----= 0.
xn
-x
Expanding the last expression in powers of x, we find

(n-t-1)xn+ (nt.~)n xn-1+ ... =0
or
n n-1
x n -+- TX
+ ....
n
The sum of the roots of this equation is equal to - 2 .

Consequently
1
X1-1
+ X2~-1 + + Xn-1
-2
32. Consider the equation (with t as an unknown)

x2
y2
-t-+ t-b2
z2
+ t-c2 = 1.
By virtue of the given equations this equation has three

roots: 2, v 2, p 2.
Expanding the last equation in powers of t, we get
t (t - b 2) (t - c2) - x 2 (t - b 2) (t - c2) - y 2 (t - c2) t - z2 (t - b 2) t = 0,
3
t + rxt 2 + . . . = 0,
where r:x, = -b 2 - c2 - x 2 - y 2 - z 2.
But as we know, the roots of this equation are 2 , v2 , p2
Therefore, it must be
2
+ v2 + p2
b2
+ c2 + x2 + y2 + z2.
306
Solutions
Hence
x2 + y2 + z2 = 2 + v2 + p2 _ b2 _ c2.
33. Since cos a + i sin a is the root of the given equation, we have
n
~ Pk(cosa+isina)n-k=O
(p 0 =1)
k=O
or
n
(cosa+isina)n ~ pR(cosa+isinath=O.
k=O
But
(cos a+ i sin at 1 = cos a - i sin a,
therefore
n
~ P11(cosa-isina)"=O,
k=O
~ p 11 (cosak-isinak)=0.
k=O
Hence, indeed,
n
~ P11 sin ka = p 1 sin a+ p 2 sin 2a + ...
k=O
+ Pn sin na = 0.
34. On the basis of the given data we have identically

xn + P1Xn-i
+ [J2Xn-
+ ...
+ Pn-1X +
Pn =
= (x-a) (x-b)
... (x-k).
Substituting for x first i and then - i and multiplying termwise, we get the required result.
35. Extracting the two given equations termwise, we
find
(p - p') x + (q - q') = 0.
(1)
Multiplying the first equation by q' and the second by q and
subtracting term by term, we have
x3 (q' - q) + x (pq' - qp') = 0
(2)
x 2 (q' - q) + pq' - qp' = 0.
Eliminating then x from equations (1) and (2), we obtain
the required result.
36. The roots of the equation
x7 = 1
307
Solutions to Sec. 6
are
(k = O, 1, 2 , .. , 6)
2kn
. . 2kn
cos77 -+ism-
Therefore, the roots of the equation
x6 +x6 +x4 +x3 +x2 +x+1=0
()
will be
2kn +. . 2kn
xism-,.- cos-7
7
(k= 1, 2, 3, 4, 5, 6).
Put
then
1
x2+ X2= y2-2,
x3+73=y3-3y.
Equation (*) may be rewritten in the following way

( x3
+ :3 ) + (x + ~) + (x + ! )+ 1 =
2
O.
It is evident that
X1 = :r5,
X2 =
X5,
X3 = X4,
X11.
1
2kn
+Xn = Xn + Xn = 2 cos -7-.
Hence, we may conclude that the quantities

2n
2 cos -7- ,
4n
2 cos -7- ,
8n
2 cos -7-
are the roots of the following equation

y 3+y 2 -2y-1 =0.
Let us set up an equation with the following roots
V2 cos
2n
-7,
4n ,
2 cos -7-
v
3
"
8n .
2 cos -7-
Let the roots of a certain cubic equation

x 3 - ax 2 +bx - c = 0
be
a, ~. y.
We then have
a+~+y=a,
a~+ay+~y=b,
a~y=c.
Solutions
308
Let the equation, whose roots are the quantities

be
x3 - Ax 2
Bx - C = 0.
Then
V1v,
;/a,
y/~,
Va+V~+Vr=A,
~'aV~+Va~/v+nVy=B,
V a~y=C.
Let us make use of the following identity

(m+ p+q)3= m3+ p3+q3+
+3 (m+ p +q) (mp+ mq+ pq)-3mpq.

Putting here instead of m, p and q first Va, V~, Vr,
and then Va~, Vay, V ~y, we find

A 3 =a+ 3AB - 3C, B 3 = b 3BCA - 3C 2
In our case we have a = -1, b = -2, c = 1, C = 1.
Hence
A3
3AB -
4,
B3 =
3AB -
5.
Multiplying these equations and putting AB = z, we

find
z3
9z2
+ 27z -
20 = 0,
(z - 3) 3
+7 =
0,
z= 3
But
-V7.
3/-
A3=3z-4=5-3 v 7,
Therefore, indeed,
Va+V~+Vv=
=~
2n
2 cos -7-+
4n
2 cos -7-+
8n
2 cos -7-=
=Vs-3V7.
The second identity is proved in the same way.
37. Since by hypothesis a
b
c = 0, we may consider that a, b and c are the roots of the following equation
+ +
+ px + q = 0,
ab + ac + be, q = -abc.
x3
where
p =
Solutions to Sec. 6
309
We have
(a
+ b + c)
= a2
+ b + c + 2 (ab + ac + be),
2
i.e.
=
82
-2p.
= a, x =
Putting in our equation in turn x

get the following equalities
a3
+ pa + q =
0,
b3
+ pb + q =
b, x
= c, we
0,
c3 +pc+ q
= 0.
Adding them term by term, we find
+ PSt + 3q = 0.
a + b + c = 0, w'e have
S3
But since s1 =
s3 = -3q.
Multiplying both members of the original equation by
xk, putting then x = a, b and c, and adding, we find
sk+3 = -psk+t - qsk.
Putting here k = 1, 2, 3, 4, we find
S4
= 2p 2 ,
S5
= 5pq,
S5
= -2p3
+ 3q
2,
S7
= -7p 2q.
Taking advantage of these relationships, we easily prove

the first six formulas. The last one is also obtained readily.
38. We have
x -
u =v - y,
x2
u 2 = v2
y2
The second equality may be rewritten as follows

(x -
Since x -
u) (x
+ u)
- (v - y) (v
+ y)
0.
u = v - y, the last equality is rewritten as

(x - u) [x
u - (v -f y)l = 0,
wherefrom follows
1 x - u = 0, v - y = 0, x = u, y = v;
2 (x + u) - (v + y) = 0, (x - u) - (v - y) = 0, x =
v, y = u.
Consequently, indeed,
xn
+ yn
= un
+ vn.
310
Solutions
Let us go over to the second case. Suppose x, y, z are the

roots of a cubic equation
a.3
+ pa + qa + r =
2
0.
Prove that u, v and tare the roots of the same equation. We

have
x + y + z = -p, xy + xz + yz = q, xyz = -r.
Hence, to prove that u, v, and t are the roots of the same
equation (whose roots are x, y and z) it is sufficient to prove
that
u + v + t = x + y + z, uv + ut + vt =
= xy + xz + yz, uvt = xyz.
The first of these equalities is true by hypothesis. The second one follows immediately from the identity
2 (xy + xz + yz) = {x + y + z) 2 - (x 2 + y2 + z2)
and from the condition
x2 + y2 + z2 = u2 + v2
+ t2.
Likewise, the third equality follows from the

3xyz = x3
+ y + z + 3 (x + y + z) X
X (xy + xz + yz)
identity
and from the condition

x3 + ys + zs = us
- (x
+ y + z)s
+ v3 + t_s.
Thus, u, v, t as well as x, y, z are the roots of the same thirddegree equation. Therefore, one of the six possibilities takes
place
11
x
y
x
y
z
z
y
x
z
z
x
y
z
z
y
x
y
x
3H
Solutions to Sec. 6
It is obvious that in all cases we have

;xn T yn+zn=un+vn+tn.
39. Squaring the first trinomial, we get
A2 = (;x: + 2;xz;x3) +(xi+ 2x1Xz) e + (x~ + 2x1x3) e2

Then
A3 = (x: + x:
+ x: + 6x 1x2x3) + (3x:x2+ 3x;x1+3x;x3) e +

+ (3x:xa + 3x~x 1 + 3xix
2)
e2
Put
a= x:x 2 + x~x 3 + x:xi.
= x 1x: + x 2 x;+ X3X~.
Now
x~
+ x:+ x: =
-(PX1 + q)-(PX2
+ q)-(px3+q) =
-3q,
since
Furthermore
therefore
A 3 = -9q + 3ae + 3~e 2

Substituting x 2 and x 3 , we also find
B 3 = -9q + 3ae2 + 3~e.
Hence
A 3 + B 3 = -18q - 3a - 3~ = -27q,
since
a + ~ = X1X2 (x1 + X2) + X2X3 (x2 + X3) +
+ x 3x 1 (x 3 + xi) = -3x1x 2x 3
Likewise we get
A 3 B3 = -27p 3 .
It should be taken into consideration that
3q.
+ (x~x:+ x~x: + x:x:) + x1xzx3 + x:x1x3 +

+ X X2X1 = 32q + X333(1+1+1),
XX
~
: J -t+ X1X2X3 (x~ + x: + x~).
a~=3x~x:x~
1 2 3
and
x:r+--;a+x:r ==
1
X1
X2
3
p3
- - q --q3 .
X3
312
Solutions
40. Put
We have
or
[(x+ ~ }2+ab-~][(x+ ~ ) 2 +cd-~]=m.

Let
Then the equation takes the form

(y
+ab- ~ ) ( y + cd- ~2 ) = m,
i.e.
y2 + (ab+ cd- ~2
y+ (ab - ~2 ) ( cd - ~2 ) -
m = 0.
It only remains to solve this quadratic equation.

41. Make the following substitution
a+b
x=y--2-'
then
a-b
a-b
x+a=y+x+ b =Y--2-.
2- ,

a-b
( y+-2-
)4 + (y--2a-b )'' =c.
But
(Y+
a;b
)4=y4+4y3 a;b + 6Y2(
a;b
+ 4y (
y4
+ 6 ( a; b ) 2 y2 + ( a; b
)2+
a; )3 + ( a;
b
r.
r +.
=
Thus, the problem is reduced to solving a biquadratic

equation.
313
Solutions to Sec. 6
42. Put for brevity

a+b+e=p
and make the substitution
x
We have
(y -
a) (y -
+p
b) (y -
= y.
e) p -
abe (y - p) = 0.
Hence
p {y3
or
y {(a
(a
+ b + e) y + (ab + ae + be) y} 2
abey = 0
+ b + e) y (a + b + e) y +
+ (ab + ae + be) (a + b + e)
2 -
- abe} = 0.
And so, we find three values for y: one of them is zero, the
other two are obtained as the roots of a quadratic equation.
Then it is easy to find the corresponding values of x.
+ a)
3be (x + a) + b3 + e3 = 0.
Put x + a = y. The equation takes the form
y3 - 3bey + b3 + e3 = 0.
But it is known (Problem 20, Sec. 1) that
(x
+b +e
3 -
3bey =
= (y + b + e) (y 2 + b2 + e2 - yb - ye - be)
Consequently, one of the roots of the last equation will be
-b - e, the other two are found by solving the quadratic
equation. Then we find the corresponding values of x.
44. The equation contains five coefficients: a, b, e, d
and e, and there exist two relationships among them. Thus,
three coefficients remain arbitrary. Let us express all the
coefficients in terms of any three.
We have
a = e + d, e = b + e.
(e + d) x 4 + bx3 + ex 2 + dx + (b + e) = 0,
e (x 4 + x 2 + 1) + dx (x 3 + 1) + b (x 3 + 1) = 0.
y3
3 -
Solution.~
314
But
i3
x4
+ +1 =
x2
+1=
(x 4
+ 1) (x2 -
(x
+ 1) -
2x 2
+ 1),
(x 2
2 =
+ 1)
x2 =--
+ 1),
+ 1) {c (x + x + 1) + dx (x + 1) +
+ b (x + 1)}
= 0.
= (x 2
+ x + 1) (x
2-
2 -
The equation is now rewritten as
(x 2
Equating the first factor to zero, we find

1 +. V3
X=2_
t-2-.
The remaining two roots are found by solving the seconrl

quadratic equation.
45. We ~ave the following formula
(a
+ b + x)
= a3
+ b + x + 3a (b + x) +
+ 3b (a + x) + 3x (a + b) + 6 ab:c.
3
Using this formula, reduce our equation to the form

x 3 - (a + b) x 2 - (a - b) 2 x + (a - b) 2 (a + b) = 0.
Hence
x 2 (x -
a -
b) (x -
b) 2 (x -
(a a-
b)
a (a -
[x 2 -
b) = 0,
b) 21 = 0,
(x - a - b) (x + a - b) (x - a + b) = 0.
Thus, the given equation has three roots:
x = a + b, x = a - b, x = b - a.
46. Rewrite the equation

as follows
<(!
2
2ax2
a+x
a2x2
(a+x)2 -
=m -
2ax2
a+x
Consequently
(x-
____!!!.__) 2 =
m2 -
a+x
2ax2 .
a+x
Hence
x4
_ m 2-
(a+x)2 -
2ax2
a+x
Solutions to Sec. 6
315
x2
a+x
Put - - = y. Then the equation takes the form

y2 +2ay-m2 =0,
wherefrom we find y and then x. For y we find the following
values
y= - a
V a2 + m"'2.
(1)
The corresponding values of x are determined by th1l

formula
(2)
Let us take the plus sign in formula (1). In this case the
value of y will exceed zero. Computing, by formula (2),
the corresponding values of x, we make sure that x has two
values: one positive, the other negative. And so, our equation always has at least two real roots, positive and negative.
Consider the case when the minus sign is taken in formula (1). Now the value of y is negative, and for x to be real
it is necessary and sufficient that y 2 + 4ay ~ 0. And, consequently, it must be
y + 4a ~ 0,
i.e.
-a-Va 2 +m 2 +4a~O,
m2 ~8a2
With this condition satisfied, all the four roots will be

real. Since ay < 0, we have
Iv--=--y4+-ay I< I ~ I
2
and, consequently, both real roots, found from formula (1)

taken with the minus sign, will be negative. Thus, if all
the four roots are real, then one of them is positive, the
remaining being negative.
47. Put for brevity
5x4+10x2+1
x4+10x2+5
= f (x).
Then the equation takes the form

f (x) f (a) = ax.
Solutions
316
Further, we have
x-f (x)=
(x-1)6
x4+10x2+5 '
Dividing the first equation by the second one, we find

x-f (x) _ ( x-1
x+f(x) x+1
)5
()
Put
x-1
x+1 -:-Y,
a-1
a+1 =b.
From the equation ( ) we get
x- f (x) = y5 x
+y
5/
(x),
x (1-y5 ) = f (x) (1 + y5 ),
(x)
1-y5
- x - = 1+ys
Likewise we have
f (a)
a
1-b5
1 +bs
Now our equation can be rewritten in the following way
whence
The last equation has five roots, namely

y,,= -be"
(k=O, 1, 2, 3, 4.);
211
8=COS5
+.ism5.
. 211
But
x- 1+11
-
1-y ,
consequently
1+Y1t
x,,=--=
1-y,,
1-bek
1+be"
(a+1)-(a-1)e"
(a+1)+(a-1)e"
Solutions to Sec. 6
31.7
Further
k
(a+1) 8-2 -(a-1) 82

k
(a+1)8-2 +(a-1)82
k
nk
cos - 5- -
a(8- 2 -82)+8-2 +82

k
,.
ia
nk
sm - 5-
nk
. . nk
a cos - 5--z sm - 5-
a(8-2 +82)+8-2 -82
In particular, at k = 0 the solution is

1
Xo=7
48. Transform the left member of the equation. Denote
the sum on the left by Sm. Then
S 1=
1 + _a_1_ t
x-a 1
a2x
(x-a1) (x-a 2)
x2
(x-a1) (x-a2)
Prove that
x2m
S m(x-a1)
=---------(x-a2) ... (x-a2m)
Suppose this equality is true at m = n, and prove that it

will be true also at m = n + 1. We have
S
x2n
n+t = (x-a1) ... (x-a2n)
a2n+1x2n
(x-a1) ... (x-a2n) (x-a2n+1)
a2n+2x2n+1
(x-a1) ... (x-a2n+2)
Reducing the right member to a common denominator

and accomplishing all the necessary transformations, we get
x2n+2
Sn+t =
(x-a1) ... (x-a2n+2)
Now our equation takes the form

x2m-2pxm+ p2
=0
(x-a1) ... (x-a2m)
or
(xm- p) (xm- p) = 0.
The equation has m double roots.
318
Solutions
49. 1 We have x 1+ x2 x 3 = - p, x 1x 2 + X1Xa

=q, X1XzX3= -r.
From the second equality we get
X1X2 + X1X3 + x: = X1 (x1+ Xz + X3)
+ x 2x3=
= q,
whence
X1= _ _']_
p
Using the first equality, we find

Xz
q-p2
+ X3=--.
p
From the third equality we have

rp
XzX3=
q
It only remains to set up
by x 2 and x 3
2 Solved analogously to
50. 1 Using the identity
we can rewrite our system in
(y
(z
(x
a quadratic equation satisfied

the preceding one.
of Problem 4 of this section,
the following way
+ z + a) (y + ze + ae
+ x + b) (z + xe + be
+ y + c) (x + ye + ce
2)
2)
2)
(y
(z
(x
+ ze + ae)
+ xe 2 + be)
+ ye 2 + ce)
2
0.
To find all the solutions of the given system it is necessary to consider all possible (27) combinations. Thus, we
get 27 systems, each containing three equations linear in
the unknowns x, y, and z.
If each of these systems is designated by a three-digit
number in which the place occupied by a certain digit
corresponds to the number of the equation and the digit itself to the number of the factor in this equation, then the 27
systems will be written as
111, 112, 113, 121, 122, 123, 131, 132, 133,
211, 212, 213, 221, 222, 223, 231, 232, 233,
311, 312, 313, 321, 322, 323, 331, 332, 333.
Let us explain, for example, system 213: taken from
the first equation is the second factor, from the second-
:i19
Solutions to Sec. 6
the first factor and from the third-the third factor. Thus,
system 213 will have the following form
y + ze + ae 2 = 0, z + x + b = 0, x + ye 2 + ce = 0.
Let us decipher some more systems
y+z+a=O,
z+x+b=O,
y+ze+ae 2 =0, z+xe2 +be=0,
y+ze2 +ae=0, z+xe 2 +be=0,
y+z+a=O, z+xe+be 2 =0,
x+y+c=O;
x -t- ye+ ce2 = O;
x+ye 2 +ce = O;
x+ye+ce 2 =0
(111)
(232)
(333)
(122)
and so on.
2 We have
x 4 = xyzu +a,
= xyzu
y4
+ b,
z4 = xyzu + c,
u 4 = xyzu
Multiplying these equations and putting xyzu

t4
= (t + a) (t + b) (t + c) (t +
+ d.
t, we find
<i).
Thus, for determining t, we have the following equation

(a + b + c + <i) t 3 + (ab + ac + ... ) t 2 +
+ (abc + acd + ... ) t + abed = 0.
However,
a+ b
c
d = 0,
+ +
therefore, for finding t we get a quadratic equation. Knowing t, we easily obtain x, y, z and u.
51. We have
2
1+(1+x)+(1+x) + ... +(1+x)
(1 +x)n+ 1 -1
(i+x)-i
n+t
! {~ C~+1X 11 -1}
k=O
n+1
~ C~+1Xk-t.
h=1
Wherefrom follows that the term containing x 11 will be

11+1 h
Cn+t.T
52. We have
320
Solutions
Since this polynomial is multiplied by the second-degree

trinomial
(s - 2) x 2 + nx - s,
it is clear that the coefficient of xs in the product will be
equal to
(s-2) c~- 2 +nc~- 1 -sC~.
Carrying out all the necessary transformations, we see that
the last expression is equal to
s-2
n Cn .
53. Put x = 1 + a, where a> 0 (since x > 1).
Then we have
'pxq-qxP- p+q= p(1+a)q -q (1+a)P- p-t q =
= p { 1+qa+ q(i~1)
-q { 1 +pa+
a2+ ... }-
P~.-;1) a2+}-p+q=
=(pC~-qC~)a 2 +(pq-qC~)a 3 +
...
Since q > p, we can prove that all the terms of the above
expansion are positive [the coefficient of ak (if k > p) will
be equal to pql. Thus, to prove the validity of our assertion, it is sufficient to prove that
~ = pC~ - qC; > 0

if q > p and k
We have
~ p.
~=pq(q-1) ... (q-k+1) _qp(p-1) ... (p-k+1)=
1.2.3 ... k
1.2.3 ... k
= ~f {(q-1)(q-2) ... (q-k+1)-(p-1)(p-2) ... x
x (p-k+1)} > 0,
since
q- 1
>
p -
1,
q- 2
>
54. Let the greatest term be

k n-h h
T 1t= CnX
a.
p - 2,
32i
Solutions to Sec. 6
This term must not be less than the two neighbouring terms
Tk-t and Tk+i Thus, there exist the following inequalities
Tk ~ Tk-1
Tk ~ Tk+to
Whence
.--=-.::;::1
n-k+1
a """' '
The first of them yields

k::;:: (n+1) a
""'
x+a
From the second one we get

k 2
(n+1)a
x+a
First assume the (n+i)a

x+a
1.
is a whole number.
Then
(n~i) a -1 is also a whole number, and since k is a

x
whole number satisfying the inequalities

(n+1) a _ 1 ::;::k::;:: (n+i) a
x+a
"""' """'
x+a
'
it can attain two values

k=(n+i)a
x+a '
k= (n+i)a
x+a
-1.
In this case there are two adjacent terms which are equal to
each other but exceed all the rest of the terms. Now consider
the case when (n++i)a is ~ot a whole number. We then have
x
(n+1)a
x+a
=[ (n+1)a
J+e
x+a
'
where 0 < 0 < 1 (for the symbol [ ] see Problem 35,

Sec. 1). In this case the inequalities take the form
k.;;:.[ (n+i)a
'"""
x+a
J+e '
k2[ (n+i)a
:;ox+a
J-(1-0).
It is apparent that in this case there exists only one value

of k at which our inequalities are satisfied, namely
k=[ (n+1)a ]
x+a
322
Solutions
And so, when (nx +1-ai)a is not a whole number, there exists
only one greatest term Tk.
55. Let i and r.i be positive integers. We have
(x+1)m-xm=mxm-i+ m(~.; 1 ) xm- 2 + ... +mx+1.

Heplacing here x by x+1, we get
(x +2)m-(x+ 1)m =
= m (x + 1)"'- 1 +
<7.;
1) (x +
1)m-z -1- ... -1- m (x + 1) + 1.
Subtracting Lhe preceding equality from the last one,

we fmd
(x-1- 2)m-2(.x+1)m -1- Xm = m (m-1) xm- 2 p 1xm-:i + ....
Analogously we obtain
(x+ 3)m __ 3 (x+ 2)m +3(x+1)m-Xm =
= m (m-1) (m-2) xm- 3 + p 2.x'11- 4 + ...
Using the method of mathematical induction, we can prove
the following general identity
(x-l-i)m- y(x-l-i-1)m + i(i1~ 1 ) (x+i-2)m+ ... +

-l-(-1)ixm=m(m-1) ... (m-i+1)xm-i+pxm-i-i+ . .. ,
wherefrom it is easy to obtain that at i = m
(x+m)m-
7(x+m-1)m+ ... -l-(-1)mxm=m!.
> m, we get
(x + i)m-T (x+ i - ~)m +
If i
0
+ i(i1--:21) (x+i--2)m+ ... +(-1)ixm=0.

Putting in the last equalities x = 0, we find the required
identities.
56. We have
(x+ai)n= xn+c~xn-i ai+C~xn- 2 a 2 i 2 +C~xn- 3 aaia+ ... =
= {xn-C!xn- 2 a 2 +C~xn-'a'- .. .
}+
+i{C~xn- 1 a--c:,xn- 3a 3
+ ... }.
323
Solutions to Sec. {j
Going over to the conjugate quantities, we get

(x-ai)n ={xn-C!xn- 2 a 2 +C~xn- 4 a'- .. . }-
.. . }.
Multiplying these equalities term by term, we find the
required result.
57. 1 We can write our product in the following way
-i{C~xn-1a-C;xn-1a3+
~
s=O
2n
~ xt= ~ A 1x 1,
X8
t=O
l=O
wherefrom it follows that

A1= ~ 1.
+t=l
O~s~r'
O~t~n
First assume l
n. Then s can attain the values s
= 0,
1, 2, ... , land, consequently,

A1
if l ~ n .
.. If n -< l ~ 2n, then we put
+1
+
l = n
l',
where 1 ~ l' ~ n, l' = l - n.
In this case s can take only the following values
s = l', l'
1, . . . , n.
The total number of values will be
n - (l' - 1) = n - (l - n - 1) = 2n - l
1.
And so,
A 1 = 2n
1 - l if n < l ~ 2n.
It is easily seen that An-k = An+k = n - k
1.
Indeed, expanding the product, we get immediately
(1 +x+x2 + ... +xn) (1+x+x 2 + ... +xn) =

=
1 +2x+3x2 + ... +nxn-1+
+ (n+ 1) xn+nxn+i+ ... +2xtn-1+xn.
2 In this case we have

n
2n
~ ( -1) 8 x 8 ~ xt = ~ A1x1
s=O
t=O
l=O
Solutions
324
Hence
Ai= ~ (-1)'.
l=s+t
0~1~n
O~t~n
Considering again separately the cases when l < n and

we arrive at the following conclusion
> n,
then A 1 = 1 +(; 1)i ,
if l-<:n,
> n,
then A 1 = 0 when l is odd and

A 1 = ( -1)n when l is even.
Thus, Ai = 0 for any odd l, i.e. the product contains
only even powers of x, and if n is even, then all the coefficients (of even powers) are equal to +1; if n is odd, then
half of them is equal to + 1, the other half to __:_ 1
if l
Ao = Az
. . . = An-t =
+1,
An+t = An+a = = Azn = -1.
3 We have
n
2n
~ (k+1) xii ~ (s+ 1) x'= ~ Aix1

s=O
k=O
Hence
Ai= ~
(k+f)(s+1)=
i=O
(ks+l+1).
k+s=I
k+s=I
O~k~n
O~a~n
O~k~n
O~a~n
Let us first assume that l ::;;;; n, then k can take on only

the following values: 0, 1, 2, ... , l, the corresponding
values of s being l, l - 1, ... , 0.
Therefore
k=O
k=O
k=O
=l ~ k- ~ k2 +(l+f) 2 =(Z+1)(Z!2)(Z+3),
taking as known that

f2+ 22+ ...
(see Problem 25, Sec. 7).
+ z2 = l (l+ 1)6(2Z+ 1)
325
Solutions to Sec. 6
Then assume n < l ~ 2n and put l = n + l', where

l' ~ n. Then k can attain only the following values
l', l' + 1, . . . , n
and, consequently,
A1=
(ks+l+1)= ~ [k(l-k)+l+1J=
k+s=l
k=l-s
=l ~ k- ~
k=l-n
k2 +(l+1)(2n-l+1)=
k=l-n
= (2n-Z+1) (l2+2z+2)
+ (l-n-1)(Z-n)(2Z--2n-1)
6
n (n+ 1) (2n+1)
6
4 Solved as the preceding case.

58. 1 We have
1 +~+~ +c~ + ... +c~- 1 +C~= (1+1)n=zn,

1-c~+c~ -C~ + ... +(-1)nc~= (1-1r=o.
Adding the two equalities and then subtracting, we get
the required identity.
2 as well as 3 is reduced to 1 if we take into account
that
Ck2n = c2n-k
2n
59. Consider the identity
(1 + xr =
C~
C~x+C~x2
C~x 3
+ ... +
c~- 1 xn- 1 + C~xn.
Putting in this identity in succession x = 1, e, e2 , where

e2 + e + 1 = 0, we get
zn = c~ + c~ + c~ + c~ + ...
(1 + e)n=C~ +c~e+C~e 2 +C~e 3 + ...

(1 + e2 r = c~ + c~ e2 + c~ e' + c~ e6 + . . . .
But 1 + ek + e2 " = 0 if k is not divisible by 3 and 1 +eh+
+ e2" = 3 if k is divisible by 3.
Consequently,
zn+ (1 +er+ (1+e2 )n = 3 {C~ +c~ +c~ + ... }.
326
Solutions
Since for e we can take the value

e=cos 32:rt
+isrn3,
.. 2:n:
we have
. 4:rt
:rt
1 + e = - e2 = - cos-4:n:
3- - i srn 3 = cos 3
+ i.srn
. 3:rt
. 2:n:
:rt
:rt
1 + e2 = -e= -cos 32:n: -ism
3 =cos 3 -ism 3 .
Therefore
2n + (1+e)n+ (1+e2 )n=2n+ 2 COS n; .

Hence, we obtain
c~ +c~ + c~ +
... =
+( +
2n
2 cos n3:n: ) '
the other two equalities are obtained similarly by considering the sums
2n+e (1+e)n+e2 (1+e 2 )n,
2n+ e2 (~ + e)n+ e (1 + e2 )n.
60. The solution is analogous to that of the preceding
problem. Consider (1 + i)n.
61. Since
C~
k(k-1)
12
k2
- 2 - 2 , we get
2C'fi_ =k 2 -k.
Consequently,
n
n
2 ~ c~ = ~ k2 k=2
k=2
wherefrom our identity is obtained.

62. Let a1 = C~, a 2 = q+ 1 , a 3 =
Then
a2
a.=
n-k
k+1 '
a4
a;=
n-k-2
k+3
n
~ k,
k=2
C~+ 2 ,
a3.
1+.!2
at
1+~
aa
n-k-1
k+2
-=--a2

1
1
---+---
a 4 = C~+ 3
327
Solutions to Sec. 6
63. If we rewrite the equality in the form

n!
1! (n-1)!
,
1
n!
3! (n-3)!
+ Sf(n--5)!
n!
+ + (11-1)!
n!
1!
= 2n-1,
then the problem is reduced to proving the following

relationship (see Problem 58)
c; c~.
c~-1 = 2n-1.
64. Consider the equality
+ + ... +
( - 21
+ i. 2V3)n = ( cos -2n3- +.ism

. 32n )n =
=
cos -2nn
3-
+. sm
. -2nn
3i
Further
(--{-+i ~3 f = (-;~) 11
(t-iV3t=
= (-;~)n ( 1 + c; (- i V3) +
+c~ (-i V3) 2 +c~ (-i
(-1)n (
12n
- i
V3) 3 + ... )=
2. 1
3Cn I
... -
V3 (C;-3c~ +3 2 C~ -3 3 C~ t ... )).

\
Equating the coefficients of i in both members of the

equality (*), we get
-V3 (C;-3c~ + 3 2C~ --3 3 C~ + ... )= (-1t2n~in 2 ~n

Hence
s -- c1n;- 3can
+32cn -
33c1n
211 sm-3 2nn '

+ ... --- ( - 1.)n+i V3
wherefrom we easily obtain

s=O
if n:=O (mod 3),
s=2n-i
if n
1 or 2 (mod 6),
s= -2n-r if n=4 or 5 (mod 6).
65. Consider the expression
{1+ir.
328
Solutions
We have
Hence
(1 + i) n = (1 - c~ + c~ - c~ +
... )+
i (c~
- c; +
c~
- ... ).
But
1+ i=
V2 (cos
~ + i sin ~ ) .
Therefore
2
C4
<J = 1 - Cn + n -
<J =
Hence, if n
c1
n-
== 0
c3
1l
csn + ... =
nn '
22 COS -4n
+ cn - c1n + ... =
. 4Mt .
22 sin
(mod 4), i.e. n = 4m, then

cr=(-1)m22m, a'=O.
If n=1 (mod4), i.e. n=4m+1, then

'
<J=<J'=(-1)m22m.
Jf n = 3 (mod 4), i.e. n = 4m + 3, then

a= ( - i)m+1 22m+1, a'= ( -
tr 22m+1.
Finally, if n==2 (mod 4), i.e. n=4m+2, then

(J = 0,
a'= (- i)m 22m+1.
66. 1 Let us write our sum in the following way
k=n
s=1C~+2c~+3c~+ ... +(n+1)C~= ~ (k+1)C~.
k=O
and introduce a new summation variable. Put k = n - k'.
Then the sum is rewritten as
k'=O
k=n
s= ~ (n-k'+1)C~-k' = ~ (n-k+1)C~=
k~n
k=O
k=n
= ~ [n+2-(k+1)]C~=
k=O
k=n
k=n
=(n+2) ~ C~- ~ (k+1)C~=(n+2)2n-s.
k=O
k=O
329
Solutions to Sec. 6
Consequently,
2s=(n+2)2n, s=(n+2)2n-1.
This sum can be computed in a somewhat different way.
Rewrite it as follows
s= (C~ +c~ + ... +C~) + (C~ +2C~ + ... + nC~) = 2n+n+

+ ( -1)+ 13 n(n-1)(n-2)+
+ 2 n(n-1)+
12
1.2.3
n n
n =2n+n { 1+(n-1)+(n-;t- 2)+
... +(n-1)-f-1} =
= 2n+ n{C~_ 1 + C~_ 1 + ... + C~.::D = 2n+n2n-1=2n-i (n+2).

2 We have
C~-2C~+3C~+ ... +(-1)n- 1 nC~=n-2 n(;.-; 1) +
+ 3
(n-;.1~-~-2) + ... + (_ 1)n-1 n =
=n { 1 _ n~1 + (n-;.~-2) + ... +(- 1 )n- 2 n~i +

+ ( - 1)n-i} = n (1-1)n-i = 0.
67. Rewrite the sum in the following manner
1 1
1 2
1 3
( -1 )n-1
n
2Cn-3Cn+4Cn-+ n+1 Cn=
n
n(n-1)
n(n-1)(n-2)
(-1)n-I
= 2 - 123 +
1.2.3.4
+ + n+1 =
=-1-{ (n+1)n _ (n+1)n(n-1) + ... -f-(- 1)n-1} =
n+1
12
123
=-1-{[ 1 - n+i + (n+1)n _ (n+1)n(n-1) +
+
n+1
_
1
1-2
1.2.3

-f-(-1)n+i]-1+ n+1} =-1-{(1-1)n+i+n}=-n-

1
n+1
n+1
68. 1 Consider the following polynomial
(1 + x)n+l = 1 + c~+ix + c~+ix 2 + ... + c~:!xn+i.
Hence
1-1=co -!- q, 2+ c~ 3+ +~ n+l

n+1
nx
2 X
3 X

n+1 x
Putting x = 1, we get the required identity.
(1+x)n+
330
Solutions
2 Obtained from the preceding identity at x = 2.

69. Put
1 c2
1 c3
Cln-2
n+3 n+
Then we have
Un
_ {
Un-I -
-{
+ (-f)n-t
enn""' Un
n
_ _!_ n (n-1) + _!_ n (n-1) (n-2) +

}2
1 2
3
1 2 3
..
__ 1 _ _!_(n-1)(n-2)+_!_(n-1)(n-2)(n-3)_
2
1.2
3
123
1 {n(n-1)
1.2
(n-f)(n-2)}+
1.2
={n-(n-- 1 )} - 2
+..!.. { n(n-1)(n-2) _(n-1)(n-2)(n-3)}-I- ...

3
123
1.2.3
'
_ 1 n-1
(n-1) (n-2)
__
-1.2+
123
+ ... -
=_!_{nn
}=
n(n-1) -'- n(n-1)(n-2)

1 2
I
1 23
=_!_{1-(1-1t}=_!_.
n
n
And so,
1
Un-Un-1=-,;:
Therefore we may write a number of equalities
1
U2-U1=2'
1
U3-U2=3'
Un -Un-I =
n1 .
Adding them term by term, we find

1
Un=1+2+3+ .. . +n.
70. 1 We may proceed as follows. The expression on
the left is the coefficient of xn in the following polynomial
s= (1 +x)n+ (1 +xr+i + (1 +xt+ 2 + ... + (1 +xt+",
331
Solutions to Sec. 6
Transforming this polynomial, we have
s= (1 +xt{1 + (1 +x) +(1 +x) 2 + ... + (1 +x)h} =

=(1+x)n (1+x)h+1_1 =_!_{(1+x)"+h+1_(1+x)n}.
x
The coefficient of xn+i in the braced polynomial is equal

to c~t~+I Thus, our proposition is proved.
2 The expression on the left is the coefficient of xn in
the following polynomial
xn(1+xr-xn- 1 (1+x)n--j- xn- 2 (1 +x)n+ ... +
+(-1)h xn-h (1 +x)n = (1 +x)"{xn-xn-i + ... +
+ ( -1)h xn-h} = (1 + x)n-1 {xn+1 + ( -1)/i xn-h}.
It is obvious that the coefficient of xn in the last
expression is equal to
( -1)" c~_1
71. 1 Consider the following polynomials
n
(1+x)n= ~ C~x",
(1+xr= ~ c:nxt.
S=O
t=O
We have
n
(1+x) 11 (1+xfn= ~ C~x ~ C~xt=

s=O
t=O
(1
+ xr+n =
wherefrom follows the required equality.

2 Follows from 1.
72. 1 Consider the product
(1+x)n(1 +x)n= (1 +x) 2 n.
We have
n
2n
~ C~x' ~ C~x 1 =
s=O
t~o
2j qnx1
1=0
Hence
C1211 --
"1L.J
~+t=I
cnCtn
m+n
~ C~+nXP,
p=O
332
Solutions
Consequently
n
c~n =
~ c~ .c~
s+t=n
~ c~c~-s = ~ (C~)2.
s=O
s=O
2 In this case we consider the following product
(1+xt(1-x)m=(1-x2)m.
()
Consequently
m
~ ( -1)" c:,.x ~ C!nx 1 =
s=O
1=0
1=0
( -1) 1 C!nx 21 ,
therefore
Let us assume first that m is even and put m = 2n. Let
l=n. Then
Hence
2n
~ ( -1)' (G;n) 2 = ( -1t c~n
s=O
3 If m is odd, then we put m = 2n

1. The coefficient
of 'x 2n+i in the left member of the equality () is equal to
2n+1
s+t=2n+1
( -1)" c~n+P~n+1 = ~
s=O
( -1)' (G;n+1) 2.
But the right member of the equality () shows that this

coefficient must equal zero (since it is evident from the
expansion that odd powers of x are absent). Therefore
2n+1
~ ( -1)" (G;n+l) 2 = 0
s=O
and equality 3 is proved.

4 We have two equalities
C~x + 2C~x2 + ... + nC~xn = nx (1 + x)n-1 ,
c~+c~x+
... +c~xn=(1-t-xt.
333
Solutions to Sec. 6
Multiplying them termwise, we find

n
+ x)
~ sC~x ~ C~x" = nx (1
s=O
k=O
2n- 1
Equating the coefficients of xn in both members of these

equalities, we get the required identity.
73. Since the product (x - a) (x - b) is a second-degree
trinomial, when divided by it, the polynomial f (x) will
necessarily leave a remainder which is a first-degree polynomial in x, rxx
~ Thus, there exists the following identity
(x) = (x -
b) Q (x)
a) (x -
It only remains to determine

tity first x =a and then x
r:x,
+ r:x.x + ~
and ~ Putting in this idenb, we get
+ ~.
rxb + ~
f (a)
= rxa
t (b)
But we know that the remainder from dividing f (x) by

a is equal to f (a), therefore,
x -
f (a) =A,
f (b) = B.
Thus, for determining rx and ~ we get the fqllowing system
of two equations in two unknowns
Hence
1
rx=--(A-B)
a-b
'
~-aB-bA
a-b
74. Reasoning as in the preceding problem, we conclude

that the remainder will have the following form
r:x,x2
+ ~x + Y
334
.Solutions
For determining a,
and y we have the following _system
aa
ab
ac
+~a+ y
=A
y = B
+ ~b +
+ ~c + y
= C.
On deterPining a, ~ and y, we may represent the required remai.nder ax 2 + ~x + y in the following symmetric
form
(x-b)(x-c) A+(x-a)(x-c) B+(x-a)(x-b)C
(a-b)(a-c)
(b-a)(b-c)
(c-a)(c-b)
75. The remainder will be

(x-x2) (x-x3) ... (x-xm)
(x1-x2) (x1-x3) ... (x1-xm)
Yt
(x-x1) (x-x3) ... (x-xm)

+ (x2-x1)(x2-x3)
... (x2-xm) Y2 + +
+ (x-x1) (x-x2). . (x-Xm-1)
Ym
76. The required polynomial (see thP preceding problem)
takes the form
(xm-x1) (xm-X2) ... (xm-Xm-1)
(x-a2) (a:-a3) ... (x-am) At+

(a 1- a 2) (a1-a3) ... (a1-am)
(x-a1) (x-a3) ... (x-am) A
+ (a2-a1)
2 + +
(a2-a3). (a2-am)
+ (x-a1) (x-a2) ... (x-am-1)
(am-a1) (am-a2) ... (am-llm-t)
77. Our equality states the identity of two polynomials. For this purpose it is sufficient to establish that
the polynomial
j ( )
Xi
(x-x2)(x-x3) ... (x-xm)

(x 1-x2) (x1-x3) ... (x1-Xm)
(x-x1) (x-x3) ... (x-xm)

(x2-x1) (x2-x3) . . (x2-xm)
+ f (xm) (x-x1) (x-x2) ... (x-Xm-1) _
(x)
(xm-xt) ... (xm-Xm-1).
is identically equal to zero. Since the degree of this polynomial is equal tom - 1, it suffices to establish that it vani-
335
Solutions to Sec. 6
shes at m different values of x. Indeed, it is easy to check

that this polynomial is really equal to zero at
X = x 1 , X2, X3, . , Xm
78. Obtained from the previous problem by equating

the coefficients of xm- 1.
79. If we put in the preceding' problem f (x) = 1, x,
x 2 , , xm- 2 , then it will be proved that sn = 0 if 0 ~
~ n < m - 1. To prove the identity
Sm-1 = 1
it is sufficient to put f (x) = xm-l in the identity of Problem
77 and to equate the coefficients of xm-l in both members
of the identity being obtained. To compute sn for n >
> m - 1 it is possible to proceed in the following way.
Suppose x 1 , x 2 , , Xm satisfy an equation of degree m
am+ P1am-i
+ P2am- 2 + + Pm-1a +Pm= 0,
where
-p1=x1+x2+ +xm,
P2 =X1Xz+X2X3
-
+ ... + Xm-1Xm,
p3=X1XzX3+ .,
Multiplying both members of our equation by a", we

get
a"'+h + P1am+k-1 + Pzam+k-2 + ... + Pm-1ai.+1+Pmak=0.
Putting in this equality successively a
and adding, we find
Sm+k
+ P1Sm+k-1 + P2Sm+k-2 +
x 1, x 2 ,
Pm-1Sk+1
+ PmSk =
At k = 0 we have
Consequently
Sm =
At k
-p1 =
X1
+ X2 +
+ Xm
1 we obtain
Sm+1
+ P1Sm + PzSm-1
= 0.
, Xm
0.
336
Solutions
Further
Sm+I = (x1
i.e.
Sm+t
+ X2 + X3 + . . . + Xm) 2 - (x1x2 + . . . + Xm-1Xm)

= x~ + x; + ... + x~ + X1X2 + X1X3 + ... ,
is equal to a sum of products of the factors
taken pairwise.
Here the factors may be both equal and unequal. Similar
results can be obtained for sm+ 21 sm+a and so on. The same
results can be obtained using a more elegant method (Gauss,
Theoria interpolationis methodo nova tractata). Put
1
-(x_1__
-x-2)-(_x_1--x-3)-.-.-.-(x_1___x_m_) = cti
- - - - - - - - - - = ct2,
(x2-x1) (x2-xa) ... (x2-xm)
1
=ctm.
(xm-Xt ) (xm-x2) ... (xm-Xm-1)
Then we have
Sn= x7a1
+ x~a2 + ... + x:iam.
Let us form the following expression
p=
ctt
1-xiz
+ 1-x2z + ... + 1-xmz

<X2
<Xm
( )
Using the formula for an infinitely decreasing geometric

progression and assuming that z is chosen so that I x1z I< 1,
I x2z I< 1, ... , I XmZ I< 1, expand the sum in an infinite series in the following way
P =a1(1 +x 1 z+x~z 2 +x~z 3 + ... ) +a2(1 +x2z +x;z2+

+x~z3 + .. .)+ ... +am(1 +xmz+x~z 2 +xi!nz 3 + ... ).
Or
P =(a1 +a2+
i.e.
... +am) +(x1a1 +x2a2+ ... +xmam) z+

+ (x~a 1 + x;a2 + ... + x~am) z2+ ... ,
337
Solutions to Sec. 6
Put for brevity

(1 - X1Z) (1 - XzZ)
Expanding Q in powers of z, we can write
Q = 1-
+ a z + ...
+ + ... +
+
+ ...
a 1z
where
CT1
X1
CT2
X1X2
2 2
X2
+amzm,
Xm,
X1X3
_.!_
Xm-!Xm,
Multiplying both members of() by (1 - x 1z) (1 - x 2z) ... X

XmZ), we have
X (1 -
PQ = a 1 (1 - XzZ) (1 - X3Z)
a2 (1 - X1Z) (1 -
+ IX3 (1
+
-
X1Z)
(1 -
XmZ)
(1 - XmZ) +
(1 - X2Z) (1 - X3Z) (1 - XmZ) + ...
1Xm (1 XiZ) (1 XzZ) (1 Xm-iZ).
X3Z)
Thus, the product PQ is an (m - 1)th-degree polynomial

in z. Let us show that it is simply equal to zm- 1 , i.e. the
following identity takes place
PQ=zm-1.
Indeed, the expression PQ-zm- 1 becomes zero at z =
1
xi
X2
Xm
xi
= - , - , ... , - . At z=- we have
Let us show in the same way that PQ - zm-l vanishes

at z = ~ .... , ..!__ But if a polynomial of degree m -1
X2
Xm
vanishes at m different values of the variable, then it is
identically equal to zero. Thus, PQ - zm-l
0. Consequently
zm-1
-Q- P
.
S ol11 lions
338
Or
zm-1
1
1-<T 1z+<T 2z2-a 3z3+
...
a,,,zm
=so+ s1z+
...
Sm_
2z"'- 2 + s,,,_lzm-t -+-
....
If we expand the left member in an infinite series in powl'rs

of z, then this series will begin only with a term containing
zm- 1. Therefore the coefficients of z0, z1, . . . , zm- 2 must
also be equal lo zero on the right, i.e. we have
So
s 1 '~' s 2 -_'
Sm _ 2 ~
0.
Besides, the coefficient at zm- 1 in the left member is equal

to 1. Therefore
Sm -1 '=
1.
Now our equality takes the followi11g form

zm-t
1-a1z
+ a2z2-03z3 + ...
amzm -
Reducing both members by z
111 -
\-S zm +s
zm-1
1,
~m1
m+l"'
l-
we find
or
1 = (1- <J1Z + CTzZ 2 - <J3Z~
+ ... +
CT111Zm) ( 1
+S Z+
+ s,,,+1Z2 --t- ... ).
1,1
Arranging the right member in powers of z and equating

the coefficients of these powers to zero (since the left member contains only 1), we find
s,.,-a1 =0,
a 2 -- 0 1s 111 + s,,,+1 = 0,
Thus, we get a possihilily lo compute sm. s 11,+1i s 111 + 2 ,
However, lo determine the general structure of sni+i le.t us
cn11sider
00
Q=
f-.r1Z
f-.t'2Z
Solutions to Sec. 6
339
But, on the other hand,

1 - 1
7j
+ SmZ + Sn1+1Z + . + Sm+hZ

2
-l-
k+l
. .,
therefore we get
S
m+k =
"'\.~
~,
s"
Li
XiX2 X3
s+.'+s"-f- ... --k-f-1
Thus, we get the following final result: sm+k is equal to a

sum of products of k
1 equal or unequal quantities taken
from the totality x 1, x 2 , , Xm. In particular
Sm+1=x;+x~+ ... +x~+x1x2 +x1xa+ ... +
+x1xm+x2xa+ ... +xm-1Xm,
Sm+2=x~+x~+ ... +x~+x:x2+ ... +x~_ 1 xm+x1x2xa+ ....
80. Let us introduce the following notation
xn
1
Sn(Xl! X2 1 , Xm)= (x1-x2 ) (x1-x3)
(
... x1-xm)
We have
x~
+ ... +
+ (ii-x1) (x2-.L'3)
.. (x2-.:rm)
n
+ (xrn-x1) (xm-x2).
Xm
(xm-Xm-1)
= xix 2 ...
x,,, (-1 __
1 ) (-1 __1 ) ... (-1 __1 )
X2
.:r~
Xj
( 1
~
( - f )m-1
( ;
;t}
Xm
Xj
Xj
)n+m-2
(+.- :
Xj
3 ) .. (
+.-
x~)
.:r1x2 .. Xm'
therefore it is obvious that

S_ 11
(x 1,
X2, . ,
Xm)=
(-
f)m-1
.r 1xa ... Xm
Sn+m-2 -
x1
, -
1
x2
, ... , -
1 )
Xm
81. The validity of the assertion follows from the identity of Problem 77. The same identity yields
Ai=
f (x1)
(x1-x2) (x1-x3) ... (x1-x 111 )
A2=
(x2)
(
(.i2-x1) X2-x3) ... (x2-xm)
'
S ulutions
340
82. Set up the expressiou

_x_1_
},-bl
.-L ---3.1.__
I
A-b~
__!__
-;-
xn
t.-bn =
= 1-
(A-a1) (A-ll2J ... (A-an)

(A.-b1) (i.-b2) ... (1.-bn)
(*)
If all the terms are transposed to the left and reduced to a

common denominator and then the latter is removed, then
the left member becomes a polynomial in 'A of degree n - 1.
By virtue of existence of the given system of equations
this polynomial vanishes at n different values of 'A, namely
at 'A = a 1 , a 2 , , an. Therefore it is identically equal to
zero, and, consequently, the original equality () is also
an identity. But then the equality () represents an expansion into partial fractions of the following fraction
(A.-b1) (A.-b 2J (A.-bn)-(A.-a1) (J.-a 2) ... O.-a 11 )
(A.-b1) (A.-b2) ... (1.-bn)
Therefore, the unknowns x 1 , x2 , . , Xn are found by the

formulas of the preceding problem, and we get
Xi= X2 =
(b1-a1) (b1-a2) .. (b1-an)

(b1 -b2) (b1-bJ) ... (b1-bn) '
(b2-a1) (b2-a2) ... (b2-an)

- (b 2-b1) (b2-b3) ... (b2-bn)'
83. Readily obtained by applying the result of Problem 81.

84. Consider the following fraction
(x-a1) (x-a2) ... (x-an)
(x-b1) (x-b2) ... (x-bn)
It is obvious that the difference

(x-a1) (x-a 2 ) (x-an)
(x-b1) (x-b2) ... (x-bn)
-1.
on reducing to a common denominator, will be a fraction in

which the power of the numerator is less than that of the
denominator. This fraction can be expanded into partial
fractions. Therefore, the following identity takes place
(x-a1) (x-a2) ... (x-an) _ 1
At
A2
An
(x-b1) (x-b2) ... (x-b,,) -
+ x-b 1 + x-b2 + + x-bn
341
Solutions to Sec. 6
Multiplying both members of this identity by x-bti we

find
In this identity we may put

X=b1.
We then have
Similar expressions are ohtained for A2 , A 3 ,

Thus, we have the following identity
(x-a 1) (x-a 2) ... (x-an) =
(:r-b1)(x-b 2 ) . . . (x-bn)
(b1-a2) ... (bi-an) X

(b1-b2) (b1-b3) ... (b1-bn)
1 +(b1-a1)
. _1_+ (b2-a1) (b2-a2) .. (b2-an). _1_

>(
x-b1
(b2-b1) (b2-b3) ... (b2-bn)

,
1
An.
x-b2
_t_
1
(bn - a1) (bn - a2) ... (bn - an)

(bn-b1) (bn-b2) ... (bn-bn-1)
'
1
0
x-b 11
At x=O we get the required identity.

85. As in the preceding prohlem, it is easy to see that
<x+~) (x+2~) ... (x-\-n~) _ 1
(x-~) (x-2B) ... (x-n~)-
+~
Ar
L..J x-rB
r=I
where
A _
r-
(r~+~) (r~-1-2~) ... (rB+n~)
(rB-~) (rB-2B) . frB-(r-1) BJ [rB-(r+1) BJ .. (r~-nB)'
It only remains to simplify this coefficient.

86. We have
342
Solutions
and formula 1 holds at n = 1. Assuming that it is true at

n, let us prove its validity at n + 1. Indeed
Ck+n+l
Ck+n
(+
ck
+ ll.ch+n,..,
n A +n(n-1) " 2
Tuck
12 u Ck
+ ll. nck) + ll. ( C1t + ~

=ck+ (
+1)
fl.ch+ n (; .-; 1) ll. 2
fl.ch+ (
=C1t +
+ n(11-1)(n-2)
123
n(;:;- 1 + ~}
>
Aa
u ch
+ ... +
c,, + ... + ll."ch} =
ll. 2 ck
+ ... + [l.n+lch =
n+1 A +(n+1)nA.,
- 1 - l..lCk
12 uch
+ . . . +An+l
L1
C11.,
and the proposition is proved.

Formula 2 is proved likewise. It is obviou~ that at
n = 1 it holds true. Let us assume that it is valid at n.
Then we have
fl.
n+l
A
n A
+n(n-1) A
Ck=L1Ch+n-yuCk+11-1
1 . 2 1..1C11+n-2-
+
+(-1) ll.ck=(Ck+n+t-C1t+n)- ~ (c1i+n-C1, ... ,._,)+
+ n(n-1)(
Ck+n-t -c1,-1-n-2) + . +< - 1n(
) C1,+1- C1t) =
1.2
11
11+1
= C/<+n+1- - 1 - C1,+n
+(n+1)n
1 . 2 C1i+n-1 -
+( -
1)n+1 C1,,.
87. It is not difficult to check the valiflitv of this formula. We see that the right member is an nth-degree polynomial in x. Let ns designate it by 
+ x (x-1) .. /x-11+1) ll."f (O)=cp (x).

II.
Let in this equality x= 0. We get q, (0) = f (0), at x = 1

we find
Cf (1) =c= I (0)
M (0) = I (1).
Using formula 1" of the preceding problem we may state
that in general
Cf (k)
=I (k) at k = 0, 1, 2, .... n.
Solutions to Sec. 6
Thus, two polynomials [q> (.1:) and j (x)J of degree n are

equal to each other at n
1 different values of the ill(iepenclent variable :r. consequenlly. they are equal ideutically, and we have
cp (x) =- j (x)
for any x.
And so, we have checked the rnlidity of the formulas.
It is not difficult to deduce this formula.
Let f (x) be an nth-degree polynomial. First of all we assert
that it is always possible to choose the coefficieJ1ts
A 0 , A 1 , A 2 . . . . . A,, such that the following identity takPs
place
'(x) = A 0
A 1.r--+-- A 2 x (x -
-+- . ~
t) (.r -
A 11 x (x -
1)(.r -
1) -;-- A 3x (x -
2) ... (x -
2)
+ 1).
Indeed, let us di vi de the polynomial j (x) by (x - 1) ><

>~ (x - 2) . . . (3 - n). Since the last polynomial is also
of degree n. the qnotient will be a constant, and the remainder a polynomial of degree not exceeding n - 1. Dividing
this polynomial hy x (x - 1) . . . (x - n
1), we find
the constant A,, _1 and so 011.
Let us llOW compute the cornitants A 0 A 1 A 2 ,
A,,_ 1, A 11
Put for brevity
x (x -
1) (x -
(x -
2)
k -'-- 1) ---' cp,, (x)
(k
~=
1, 2, 3, ... ).
Then we have
~<pk
(x) = <Jlk (x - 1) o-= (x
qi,,
(x)
-i- 1) x (x - 1) ... (x - k
x (x ~=
1) ... (.r -
(x -
+ 2)
k -i- 1)
~~
1) ... (x - k -i- 2) = kcpk-1 (x).
To determine A 0 A 1 A 2 , . . . . A,, proceed in the following

way. Put in our identity x = 0. Since cp" (0) = 0, we find
Ao
f (0).
344
Solutions
Let us now take the difference between the members of the

identity. We obtain
= A1
(x)
~'t1 (x)
+An ~cpn
Putting here x
+ A 2 ~cp 2 (x) + ... +

(x) = A1 + 2A 2(j)1 (x) + .
+ nAnfPn-1 (x).
0, we have
A1
M (0).
Further
= 2A 2 dqi 1
~ 2/ (x)
+ ... --+-- nAn :1crn _, (x)

2A + ... + n (n - 1) AnfPn-2 (x).
(x)
Hence
- "1 2! (0)
A 2-1-2.
Continuing this operation, we find all the coefficients

Ao. A1 . ... , An.
88. Replacing x by x
(.r+1)n=A 0 -;-A 1.c+
+ 1,
we have
~f x(x-1)+ ~!3 x(.r-1)
+ ~~ -'(.r-1)
+
Putting f (x) = (x
1)n and
ceding problem, we find
u~ing
(.c-2)-[- ...
... (x-n+1).
the result of the pre-
As = ~st (0).
From formula 2 of Problem 86 \Ve get the req11irPd expression for A ...
89. Putting k = 0 in formula 2 of Problem 86, we get
~
....l
11
n
--1- n (n -1.)
c0 = c,, - ye,,_,
1 . 2 c,,_2 - ...
+ ( - 1) c
n
Put
and take
1
Co= (x-i-n)2
Cn
==
1
x2.
Solutions to Sec. 6
345
to prove our identity it only remains to prove that
~n
n!
(x+n)2 - x(x+1) ... (x+n)
{..!.+_1_+
+-1-}
x
x+1
x+n
Use the method of induction. At n = 1 the formula is

true. Assuming, as usual, its validity for n, let us prove
that it is also valid for n+i. We have
~ n+ 1
(x+n+1)2
~ ( ~n
(x+n+1)2
-~{ (x+1) (x+2)nl... (x+n+1) (-1

+-1 +
x+1 x+2
+ x+!-t--1)} x (x+1) ~ 1 (x+n) {; + x~1 + + x!n}-
nl
{ 1
1
1
}
(x+1) (x+2) ... (x+n+1) x+1 + x+2+ + x+n+1 =
n!
- x (x+ 1) ... (x+n+1)
-x
{(x+n+1)(_!__!__1_,1
x ' x+1
+-1-)x+n
(x~1 + x!2+ + x+!+1)} =
(n+ 1)!
{ 1
1
1
}
=x(x+1) ... (x+n+1) x+x+1++x+n+1
At x = 1 our identity yields
_1_ { _!__!__!_
n+1 1 '2
+ +n+1
_1_ } - _1 -12
c~ +
22
_!_ ( -
'
1)n _1_
(n+1)2"
90. The expression cpn ( x y) is an nth~degree polynomial

in x. Therefore we may represent it as (see Problem 87)
where As=Mrp~(y) (since <vn(x+y) turns into (f!n(Y) at

s.
x = 0). However, it is known (Problem 87) that t.cpn (y) =
= nrpn-t (y), consequently
t. 2cpn (y) = n (n-1) {jln-2 (y),

t."cpn(Y)=n(n-1) ... (n-s+1)cpn-s(Y)
:-146
Solutions
Thus
C
A -_n(n-1)(n-2) ... (11-s+i)<pn-s!11)

s.1
~,<P11.-s y),
aud our formula is valid.

However, the validity of this formula ca11 be pro\ed
using other reaf'ons. Let x and y be positive integers
greater than n. Then the following equalities take place
(1 + z)x=, 1 _J_'
~-_1 .r(x-1) 2 -Lx(x-i)(x-2)

x~ r
1.2 z ' . 12'.1
(1-f-z)"=1+yz-\- y(~~t) z2 + y(y-:;~rr- 2 ) z3
... ,
+ .. ..
(1+zr+w=1+(x+y)z+ (x+11l~~i11-1l z2+

, (x+y) (.r+y-1)(.r-f-y-2)
-1
12'.1
z--:-
On the other hand,
(1+zf(1+z) 11 =(1 +zrw.

i. e.
~ <ii~~x) z". ~
qi:iy)
z' = ~ ~" <:~ y) zn.
Equating the coefficients of zn in both members of this

equality, we get
{j)11
(x-\- y) =
CPn
(.:r) + Chcf-11-1 (x) cp, (y)
+ ... +
+ c~- Cf'i
1
x) q, n-1 (y)
+<fn (y).
Let y 0 be a whole positive number exceeding n. Then

CJln (x+Yo) and cp,. (x)
+ Chcrn-1 (x) CJl1 (Yo)
-L
+ <P11 (Yo)
are two nth-degree polynomials in x, and they are equal to

each other at all whole \l'alues of x exceeding n. Consequently, they equal identically at all values of x. But y 0 may
attain all whole values exceeding n. Consequently, as in
the previous case, we conclude that y 0 can attain any values
and the equality
CJln (x+y) = cp,, (x) + Chcp,,_ 1(:r) cp 1 (y) + ...
I C~-tcp I ( X) <j:'11-1 (y) -i- cp,. (y)
is valid for any values of x and y.
Solutions to Sec. 6
347
91. First of all, both identities 1 and 2 can be readily

proved using the method of mathematical induction. Indeed,
at n = 1 identity 1 takes place. Suppose it takes place for
all values of the exponent, not exceeding n, so that we have
ll
11-> +n (n-3)
n- .,
x n+ y n= p n- y
p -q
12
p q--
(n-4) (n-:J)
11
123
Multiplying both members of this equality by x + y = p,

we get
xn+1
y11+1 + xy (x"-1 + yn-1) =
- p
n+i_..!!:,
n-l
1 p
+11(11-3) 71 _ 3 2 _
1 2 p
q
n(n-4)(11-5)
1 23
x p"-&q3 + ...
Hence
xn+1
yn+1
===
=p
71+1
11
-TP
ll-1
n (11-4) (n-5)
123
n- 1
- q ( p""- l - -1- p11-3q

_
- p
n+i
-3) p"-3q2 1.2
II (n
n+1
11 _
+{
11
+ (11 -
(11-1)(11-5)(11-6)
123
- -1- p
11 _ 5 3 -1
T.' ' -
1)
(11 -
1 2
4)
11-73-1-
p"-5q2 )-
"'
(tt-3) , 11--1}
1 2
T' -1-
n-a
p. q -
-{n(n-4)(n-5)+(n-1)(n-4)} 11-;;2+
123
12
p q
...
=p
11 + 1
n+ 1
--1-p
n-i
+ (n+ 1)12(n-2) p
n- 3 2
q -
(n+1)(n-3)(n-4) 11-5 2 +
1 23
p q
... '
and the theorem holds at n

1.
Proposition 2 can he proved just in the same way.
Bear in mind that if x and y are the roots of a quadratic
equation, then both formulas represent none other than
34S
Solutio11s
the expression of symmetric functions of the roots of this

equation in terms of its coefficients.
If we put in these formulas x=cosqi+isincp, y=
=cos cp- i sin rp, then
xn+Yn=2cosmp, p=x+y=2coscp, q=Xy=i,
siu (n+ 1) qi
x-y
sin qi
'fh us, we oht am
an expansion
of cos ncp a11d sin (n+ f) cp

sin qi
in powers of cos cp.
92. Put
xh+Yh = S1.., xy = q.
We have to prove that
Sm+ C~,qSm-1 + C~+1q 2 Sm- 2 +
... + Czn~~2qm-is 1 =1.
Assuming the validity of this Pqnality, let us prove thal

Sm+1+c;n+1qSm +c;;,+zq2 Sm-1+ ... +c~~1qm- 1 s 2 +
+ Czmqm S
= 1.
We may consider that x and y are the roots of the

quadratic equation a 2 -a+q=0.
Hence
for any whole k.
Consequently
Sm+i = Sm-qSm-1
Sm= Sm-1-qSrn-z,
Sm-I= Sm-2-qSm-3
S3=S 2 -qS1,
S2=S1-qSo,
S1=S1.
Let us multiply these equalities in turn by
l
Cm+1q,
and add them,
cz
in+zq,
cm-1 m-1
2m-1q
,
349
Solutions to Sec. 6
Then we obtain in the left member

S m+t
l sm--;- c2m~-zq2s m-1+ I cm-I

m-15 2--;--. C"'
+cm+1q
2m-1q
zmq ms 1
I
We 011ly have to prove that the right member is equal

to 1. The right member is equal to
l m-1-t- czm+zq-"S m-2T' -t- cm-1
m-1s 1'\
S m+C m+lq
2m-lq
II
cm2mq'''S
'
2
2S
I
S m-3-qs111-1- cm+lq
m-2- Cm-!-2q 3
m-I ms O
C2m-lq
or
Sm+(C,1,,
+ 1)qSm-1+(C~+1 +c~+i) q Sm-2+ ... +

2
cm ms
. cm-2 ) m is
TI (Cm-1
2m-2+ ~m-2 q - 1+ 2mq
1-qsm-1l
cm-2
m-1s1 - cm-1
1115 o =
- c m+1q" m-2- 2m-2q
2m-lq
={Sm+ C~qSm-1 +c~+1q 2 Sm- 2
+ ... -+ C~,;~2q111 - 1 S 1 } +
"'S 1 - cm-1
+Cm
2mq
2m-lqm5 O
But, by hypothesis, the braced expression is equal to 1

and CTmS 1 - C;',;~ 1 S 0 = 0, since S 1 = 1, and S 0 = 2. And
so, the right member is equal to 1. Furthermore, it is apparent, that at m = 1 our equality is true. Now we can assert
that it is valid for any m.
93. If u+v = 1, then
Um
(1
v2 + +cm-I
,
2m-2 vm-1) -1
+ vm (1 + C~u + C~+ 1 u 2 + ... + C2',;~2u"'- 1) =
-LC! v+Cz
1
m
m+I
Put
x-a
U=b-a'
Then u
+v =
x-b
V=a-b
1. Further
_1_=
(-1
+cl _1_+c2
umvm
vm
m vm-1
m+I _1_+
vm-2
... I
-+-(-1-+c1
_1_J_c2m+t _1_+
'
um
m um-1 '
um-2
Hence we get our identity.
cm-I
2m-2 _!__)-1-l'
. . .
+c"'-1
_..!.._)
2m-2 u
1.
Solutions
350
94. It is easily seen that we can always choose constants

2 , , so tha L the following identity takes place
A 1, A
+ t" + A1t (1 + tn- +

+ A t (1 + t"-~) + ....
Indeed, (1 + t)n is a polynomial of degree n in t. Dividing it by tn + 1, we obtain a remainder (a polynomial of
degree not exceeding n - 1). We divide it by t (tn- + 1)
(1
+ t)"
=c
2)
2 2
and so on. It is clear, that the quotients thus obtained will

be constants determined uniquely in the process of division.
Putting t = ..!!.. in the identity being formed. we find
x
(x
+ y)" = x" + y" + A1XY (xn-2+ yn-2) +
+ A 2x2y2 (.x"-1 + y'H) + ...
To di:ltermine the coefficients A 1 , A 2 ,

this identity
cos
qi
+ i sin cp,
let 11s put in
y = cos q; - i sin <p.
Then we have
(2 cos cp)u = 2 cos n<p
+ 2A
2) cp
cos (n -
+ 2A
cos (n -
4) cp
Taking advantage of the known formulas for the expansion of cosine's power in terms of the cosiiw of multiple
arcs (see Problem 10, 1 and 3), we find the expressions for
A1. A2, ...
!)5. Let y 1 and y 2 be the roots of some quadratic equation
y2 +PY+ q = 0.
Let us set up this equation, i.e. find p and q.
For this purpose we multiply the first equation by q,
the second by p, the third by unity and add the results.
We get
x1 (y~
+ PY1+q>+x2 (y; + PY2 + q) = a,q+aap + a3 = 0,
since
Y~
+ PY1 -j-q= Y~+PY2 +q =0.
Solutions to Sec.
3;)1
fj
We then 1nnltiply the second equation by q. the third by p

and the fourth by unity. We gel
Thus. for determining p and q we obtain a linPar system

a 1q
a 2q
+ a 2p + a3
+ a p + a,,
3
- U.
~~
0.
On finding p and q, we determine Yi and y 2 from the

equation y 2
py
q = 0. Knowing Yi and y 2 , we then
determine Xi and x 2 , say, from the first two equations. The
general system is solve<( in the same way ..~amely, suppose
y 1 , y 2 , , Yn are the roots of a certain equation of degree n:
Y11 1 I PiY n-1 -,-, P2Y 11-2 --:-- -t-' Pn-iY
+Pn = 0
To set up this equation multiply equation (1) by p 11 ,

equation (2) by Pn-I and so on, and, finally, equation (n
1)
by 1 and add the results. We get
ll1Pn
+ ll2Pn-I + + an+I
0.
\Ve then multiply equation (2) by p,,, equation (3) by

Pn-I and so on and, fmally, equation (n
2) by 1 and thus
obtain a second linear relationship for determining p 11 ,
p 11 _ 1, Continuing this operation, we finally get n
linear equations for determining the unknowns p 1 , p 2 , ,
p
If p 1 , p 2 , . , p,, are found, then to determine Y1t
y 2 , , y,. we have to solve the equation
11
y" + py"-l+
+ Pn-IY + Pn c-= 0.
To find .z 1 , x 2 , , x 11 it only remains to solve a system

of linear equations.
Demonstrated below is the original metbod of solving
this system belonging to S. Ramanujan. Consider the following expression
<D (0)
Xt
X2
I
,
Xn
1-Uyi 1 1-uy 2
-r -r
1-0y,,
352
Solutions
But
Xj
_
t-l:lui
-:r1 (1
X2
-1- u
- = X 2 (1
-vY2
+ eY1 + 02 Y1,-L- 83 Y11 ...

2
1
--;--
e Y21' 82 Y22+e3 Y2+

a
.),
Xn =Xn (1 -;-1 eYn+ S2 Yn2 +e3YnT

3 I
)
1-8
..
Yn
Consequently,
<D (8) = (Xi-r X2 +
+ Xn) T (X1Y1 +x2Y2 T + XnYn) 8--j+ (X1YI + + Xny;_) 8 + ... + (x1yin-i + X2Y~n-I + ... +
2
+xny~n-1)
e2n-1
+ (x1yin + ... + XnY;n)
92n
+ ....
But by virtue of the given equations we get
<D(8)=a1+a28+aa82-t-
...
+a2ne 2n-1
+ ....
Reducing the fractions to a common denominator, we find

Cl> (S) = A1+A 2e+Aae2+ ... +Anen-1
1 +s 1 e+~202-t-
... + Bnen
Hence
(a1 +a28+aa8 2 +
... +a2n8 n-i+ ... ) (1+B18+B2e +

2
... +
+Bnen) =A1+A2e+ ... +Anen-1

Therefore
A 1 =all
A 2 =a 2 +a 1B 1 ,
A 3 =a3 +a2B 1+ a 1B 2 ,
An= an+ an-1B1 -1- an-2B2 + ... a1Bn-t.

0 = an+t + anB1 + ... + a1Bn,
0 = an+ 2 + an+ 1B 1 + ... + a 2Bn,
0 = azn
+ azn-1B1 + + anBn.
Since the quantities a 1, a 2, ... , an, an+t ... , a 2n are

known, the last equations enable us to find first B 1 , B 2 , ,
Bn and then. A 1 , A 2 , , An. Knowing the quantities
353
Solutions to Sec. 6
Ai and Bio we can construct a rational fraction (8) and
then expand it into partial fractions. Let, for instance, the

following expansion take place
<l> (9) -1~
+ 1 -q2
p 2 8+1 Pa 8 + ... + 1 Pn 8 .
-q1v
-q3
-qn
Thc-n it is clear that
Xi= P11
Xz=pz,
Yi= q1;
Xn = Pn1
Yn = qn
Y2=q2;
The system is solved.

96. Eor the given case we have
2 + 8 + 392
+283 + 84
 (O) = 1-8-582+83+384-8&.

Expanding this fraction into partial ones, we get the
following values for the unknowns
3
X=-5
p=-1,
y=
1s+ Vs
10
'
q=
Z=
1s-V5
10
'
r=
u = __
8+_-V~5
2V5 '
V=
s-V5
2V5
3+-V5
2
'
3--V5
2
'
-V5-1
S=
t = _IV5+1
2
97. 1 We have
(m, )
= (1-x) (1-x2) ... (1-xm-) (1-xm-+i) ... (1-xm-1)(1-xm)
(1-x) (1-x2) ... (1-x) (1-x) (1 ~x2) ... (1-xm-)
Hence it is clear that

(m, ) = (m, m - ).
354
Solutions
2 Indeed,
1 - (1-xm)(1-xm-1) ... (1-xm-+1)(1-xm-) ( ' + )(1-x) (1-x2) ... (1-x) (1-x+ 1)
1-.T"'
(1-xm-l) ... (1-xm-) ('l-xm-- 1)
=
(1-x) (1-x2) ... (1-x+!)
'1-.r"t-- 1
Thus
1-xm
(m, +1)=(m-1, +1) 1 _xm-- 1
=(m-1, ~t+1)
1-xm--1 +xm--1
1 -xm-- 1
=(m-1, +1) [ 1..i.xm--1

1
_xm
1-x+ 1
_xm--1
1
J=
=(m-1, +1)+xm--1)(m-1, ).
3 Using the result of 2, we get a number of equalities
(m, +1)=(m-1, +1)+xm--1(m-1, ),
(m-1, +1)=(m-2, +1)+xm--2(m-2, ),
(+2, +1)=(+1, ~t+1)+x(+1, ),
(+1, +1)=(, ).
(m, +1)=(, )+x (+1, )+ ... +xm--1 (m-1, ). ()
4 It is required to prove that (m, ) is a polynomial.
We have
1
+ x
(m, 1) = 1-xm
... + xm-1 .
1 _x = +x
2+
Thus, our proposition is true at = 1 and any m. Assuming

that (m, k) is a polynomial at k ~ , by virtue of the formula (), we may assert that (m, + 1) is also a polynomial. And so, our proposition is proved by the method of
mathematical induction.
5 Introduce the notation
(x, m)
= 1-
(m, 1)
+ (m,
2) -
(m, 3)
+ ... +
+ (-1)m (m,
m).
355
Solutions to Sec. 6
First let us prove that
(x, m) = (1 -
xm- 1)
(x, m -
2).
We have
1=1,
(m, 1) = (m-1, 1) +xm-1 ,

(m, 2) = (m-1, 2)+xm- 2 (m-1, 1),
(m, 3)=(m-1, 3)+xm- 3 (m-1, 2),
(m, m-1) = (m-1, m-1)+x (m-1, m-2),

(m, m)=(m-1, m-1).
Multiplying these equalities successively by 1 and
adding the results, we get
f (x, m) = (1-xm-1 )-(m-1, 1)(1-xm-2 ) + (m-1, 2) X
X (1-xm- 3) - + (-1)m-z (m-1, m-2) (1-x).
But
(1-xm- 2 ) (m-1, 1) = (1-xm-1 ) (m-2, 1),
(1-xm-s)(m-1, 2)=(1-xm- 1)(m-2, 2),
Therefore
f(x, m)=(1-xm-1){1-(m-2, 1)+(m-2, 2)- ... +
+(-1)m-2 (m-2, m-2)}=(1-xm-1)/(x, m-2).
Thus
f(x, m)=(1-xm-1)/(x, m-2),
f (x,. m-2) =
(1-xm-a) f (x, m-4),
First let us assume that m is even. We get

f(x, m)=(1-xm-1) (1-xm-s) (1-xm-) ... (1-x 3) f(x, 2).
But
f-x2
f(x, 2)= 1 -(2, 1)+(2, 2)=2- i-x =1-x.
f(x, m)=(1-xm-1)(1-xm-s) ... (1-x3 )(1-x)
if m is even.
356
Solutions
If m is odd, we have
f (x, m) = (1-xm-i) (1-xm-a) ...

But f (x, 1) = 0, consequently f (x,
(1-x 2 ) f (x, 1).
m) = 0 for any odd

m. However, the last fact can be readily established immediately from the expression for f (x, m)
f(x, m)=1-(m, 1)+(m, 2)--(m, 3)+ ... +(-1)m(m, m).
98. 1 Put
n
k(k+1)
1+ ~
L.J
!1-xn)(1-xn-1\ ... (1-xn-k+1) x-2-zk=F(n).

(1-x) (1-x2) ... (1-xk)
Then
F (n+ 1) = 1
+ n+I
'1
L.J
k(k+I)
(1-xn+1)(1-xn) ... (1-xn-h+2)

(1-x)(1-x2) ... (1-xh)
x-2-z".
k=I
Hence
F (n+ 1)-F (n) =
~n
(1-xn) ... (1-xn-k+2)

(1-x) (1-x2) ... (1-.zh)
k (k+I)
x-2-z" {f-xnn -1
k=i
+ xn-h+l} + x
n
(n+I) (n+2)
zn+l =
(1-xn) ... (1-.:vn-h+2)
= ~ -:-:---,---:----::---,.-----';--4.J (1-x) (1-x2) ... (1-x")
h\k+I)
x-2-
zh xn-h+1 ( 1 - x 1')
k=i
+x
(n+1Hn+2)
2
zn+l=
k(k-1)
(1-xn)(1-xn-1) ... (1-xn-k+2) z"-lx~

(1-x) (1~x2) ... (1-xk-1)
'
n(n+I)
And so
+ zxn+ix-2-z" = zx"+iF (n).

F (n+ 1)-F (n) = zx"+i F (n),
i.e.
F (n+ 1) = (1 + zx"+ 1 ) F (n).
357
Solutions to Sec. 6
Therefore
F (n) = (1
+ zxn) F (n-1),
F (n-1) = (1 +zxn-t) F (n-2),
+ zx
(1 + zx
1 + xz.
F (3) = (1
3)
F (2),
F (2) =
2)
F (1),
F (1) =
Multiplying these equalities, we actually get

F (n) = (1 + xz) (1 + x2z) ... (1 + xnz).
2 is proved similarly.
From 1 it also follows that
(1-xn) (1-xn-1) ... (1-xn-h+l)
(1-.x) (1-x2) ... (1-xh)
is a polynomial in x (see Problem 97).

From the same equality we can obtain Newton's binomial
formula as well. Indeed
1-xn-h+l
1 +x+xz+ ... +xn-h
1-xh
1+x+xz+ ... +xh-1
Therefore, at x = 1 the last expression attains the value

n-k+1
Consequently, we may consider that the expresk
sion
(1-xn) (1-xn-1) ... (1-xn-h+l~
(1-x) (1-x2) ... (1-xh)
at x = 1 turns into
n (n-1) ... (n-k+1)
c~
12 ... k
and formula 1 at x = 1 yields
h=n
(1+zt=1 + ~ C~zh
h=I
(Euler).
99. Readily obtained from 1 of Problem 98 at z = -1.

100. Put
Co + C1 (z + z- 1 ) + C 2 (z 2 + z- 2 ) +
+ en (zn + z-n) = <Fn (z).
+...
Solutions
358
We then have
cp,,(x2z)=frn (z)
1 +x2n+lz
xz+x2n
(expressing cpn (z) in terms of a product). Making use of

cpn (z) expressed as a sum, we find with the aid of the last
identity
Chx2k+1 (1- x2n-2ll) = Ch+t (1- :l21i+2h+2)
(k=O, 1, 2, ... , n-1).
Furthermore, it is obvious that Cn ~= xn 2 . Putting in the
last relation the following values for k in succession:
n - 1, n - 2, .... , 0 and multiplying the equalities thus
obtained, we find
Ch= (1-x2n+2h+2) (1 -x2n+21<+4)
... (1-x4n)
(1-x2)(1-x4) ... (1-x2n-2h)
k2
(k=O; 1, .. ., n-1).
101. 1 Put
cos x
+ i sin x
cos x -
i sin x
= e.
Then
Further
cos lx
+ i sin lx
e- 1
= e1, cos lx - i sin lx = e- 1
Consequently
el
sinlx =Ti (1-s- 2
l
).
Sl:lbstituting this value of sin lx into the expression for

uh, we find
1
Uk=
(1- q2n) (1-q2n-1) ... (1-q2n-h+1)

- 9 k(2n-k)
(1-q) (1-q2) ... (1-qk)
.q -
where q = e- 2
The required sum is rewritten as follows
1- U1 + U2 2n
+ 'Y
lt=1
U3
+ .. + U2n = f +
1
h(f-q2n)(1-q2n-1) ... (f-q2n-k+1)

(- )
(1--q)(1-q2) ... (1-qh)
-2k(2n-h)
359
Solutions to Sec. 6
Now let us take advantage of formula 1 of Problem 98

1
and, replacing in it n by 2n, put x

We then have
q and z = -q-n-2.
2n
1- U1
+ U2 -
U3
+ ... + U2n =II (1- qk-n--2) =

11=1
2n
h=1
h=I
n
=II 2 [1-cos (2k-1) x] =
2n
II [1-cos (2k-1) x].

11=1
h=1
2 Put (as in Problem 97)

(1- q2n) (1-q2n-1) ... (2- q2n-h+I)
~-..,.,(1~-~q)~(~1-~q~2)-.-.-.(~1---q~h-)~
= (2n, k).
Then
- .!. h(2n-h)
Uk=
(2n, k) q
2.
where q =cos 2x- i sin 2x.

We have to compute the following sum
2n
~
2n
( -1)h Uk= ~ ( -1)" (2n, k)2 q-hc2n-h>'
h=O
k=O
where (2n, 0) = 1.
From Problem 98, 1 we have
2n
(1- qz) (1-q2z) ... (1-q2 nz) =
2'
h(h+O
(-1)h (2n, k) q-2- z 11 .
h=O
Put
(1-qz)(1-q2z) ... (1-q 2nz)=qi (z, q).
We then have
CJln (z, q) CJln ( -z, q) = CJln (q 2 , z2 ).
Hence
11
2n
k=O
k(h+1)
( -1)h (2n, k) q-2- zh.
2n
h'
s(s+t)
(2n, s) q-2-z' =
s=O
2n
2,; ( -1)m {2n, m} qmcm+uz2m 1

w=O
Solutions
360
where {2n, m} is obtained from (2n, m) by replacing q by

q2 Consider the coefficient of z2 n in both members of this
equality. On the right this coefficient is equal to
( - i t {2n, n} qn<n+I>.
In the left member we obtain the following expression
~
k(h+O + s(s+I)
( -1)h (2n, k) (2n, s) q-2-
-2-.
k+s=2n
But
(2n, 2n -k) = (2n, k),
therefore the last sum is equal to

2n
q2n2+n ~ ( -1)k (2n, k)2
And so, we have
qk2-2nk.
h=O
2n
q2n2+n ~ ( -1)h (2n, k)2 qh2-2nl< = ( -1)n {2n, n} qn2+n.

k=O
But
hence
2n
~ (-1)"uk=(-1tq-n2 {2n, n}.
Further
where Un
Finally,
h=O
!. n2
(2n, n) = U11q 2 , {2n, n} = u 11 q-n 2,
is obtained from u 11 by replacing x by 2x.
2n
'1 (-1)"u2=(-i)n sin(2ni:-2)xs~n(2n+4>_x ... sin4n.r

.L.J
k
sm 2x sm 4x ... sm 2nx
k=O
We proceeded from
2n
2J (-1)h (2n, k)
qk2- 2 nk = (
-1)n {2n, n}
q-n 2 .
h=O
Likewise we can obtain the following formula

2n+1
~ ( -1)h (2n+ 1, k)2 qk2-(2n+1)h = 0,
4=0
Solutions to Sec. 7
361
If we put q = 1, then (n, k) turns into

formulas
2n
~1
.::..... 1
k=O
( -1)h (Ck2n )2 = ( -1)n
en2n'
c~
and we get the
2n-f-1
2.,' ( -1)k (qn+l)2 = 0 .

k-=0
Likewise, if we take advantage of the identity

(j)n (z, q) cp,, (qnz, q) = \ll2n (z, q),
we get
n
~ (n, k) 2 qk 2 = (2n, n)
k=O
and hence
n
"" (Ck)2
L..J
n = C"2n
k=O
(see Problem 72).
1. \Ve have to prove that
1
c+a -
1
b-t-c
1
a+b -
1
a+c
However, this equality is equivalent to the following

b-a
c-b
or
b-a
b+c =
c-b
a+b '
i.e.
b2 - a2
c2 -
b2.
The last equality follows immediately from the condition of

the problem.
2. If an is the nth term and am the mth term of the arithmetic progression, then we have
an = a 1
am = a 1
+ d (n + d (m -
1),
1),
where d is the comi,pon difference of the progression
362
.~olntions
Hence
an - am = (n - m) d.
By hypothesis, we have the following Pqualities
b - c = (q - r) d,
c -
(r - p) d,
(p -
a- b
q) d.
l\hiltiplying the first of them by a, the sPcond by b. and

the third by c, we ~Pt
d [(q -
r) a
+ (p - q) c] =
= a (b - c) + b (c - a) + c (a - b)
r) a + (r - p) b + (p - q) c = 0.
+ (r -
whPnce
(q -
p) b
= 0,
3. We have
a,, -
al/
(p - q) d,
where d is the common difference of the progression.

Since, by hypothesis,
a =
q,
a'l
= p,
then al' -
aq =
q - p,
therefore
q- p
(p - q) d,
cooo
and, consequPntly,
d
(we assume p Further
am -
hence
am =
-1
q -=I= 0).
al'
al' =--
+ (m
(m -
p) d
p) d,
q -
ap+k =
a 11
li. We have
+ p.
+ pd.
Let k in this Pquality attain successively thP values:

1, 2, 3, . . . , q. Add termwise tho q obtained equalities. We
get
a,,+ 1
+ a,,+ + ... + ap+q =

= a1 + a 2 + . . . + a'l + pqd,
2
363
Solutions to Sec. 7
But
llp+i
+ llp+2 +
+ ap+q
= Sp+q - SP,
a1
therefore we have
+ a 2 + ... + aq
Sp+q =SP +Sq
= Sq,
+ pqd.
On the other hand, it is known that

SP
a1 +ap
p,
Sq=
a1 +aq
2
q.
Hence
2Sp
2Sq
-p - - q= ap-arz = (p-q) d
or
2 (pSp-PSq)
----=pqd.
p-q
Consequently
S
- S +S
p+q- p
q
+ 2(qSp-pSq)
(p+q)Sp-(.n+q)Sq
p-q
p-q
Finally
p-+-q
Sp+q= --(Sp-Sq)=
-(p+q).
p-q
5. Follows from Problem 4. However, the following

method may be applied. We have
a1 +a,,
SP=
a1 +aq
p,
Sq=
q,
hence
or
[2a 1
2a1 (p -
Hence
q)
+ d (p
+ d (p
- 1)1 p = [2a 1
d (q 2
p - q
q) = 0,
2a 1
d (p
2 -
1)1 q,
+q-
1) = O.
Solutions
since
ap+q = a 1
But
+ d (p + q -
a 1 _L ap+q
Sr+q=
1).
(p+q).
6. We have
S m= llt +Um
2
m,
From the given condition follows:
a1 +am
m
a1+an =-;:;:,
i.e.
2a1 +(m-1)d
2a1+(n-f) d
m
n
Hence
2a 1 (n-m) +{(m-1) n-(n-1) m}d=O,

therefore
d
llm=a1+(m-1)d=2+(n:i-1)d
2m
d,
2n-1
lln=2 -d
and fmally
2m-1
2n--1
7. It is necessary to prove that at the given n and k

(positive integers k ;;;;;:: 2) we can find a whole s such that
the following equality takes place
(2s + 1) + (2s + 3)
+ ... +
(2s
+ 2n -
1) = nk.
The left member is equal to

(2s
+ n) n.
Therefore it remains to prove that it is possible to find an

integer s such that the following equality takes place
(2s+ n) n = nk,
S= n (nl>.;2 - i ) .
But n can be either even or odd. In both cases s will be an

jnteger, and our proposition is proved,
365
Solutions to Sec. 7
8. Let a 2 = P,. Then a11. = a1

since, by hypothesis, a1 = 0.
Consequently
n-2
n-2
1) =
d (k -
d (k -
1),
n-2
= LI
~ k+1_ "1 ..!.+_1_= ~ (1+..!.)k
LI k
n-2
~
k
k=1
k=1
k=i
n-2
n-2
n-2
n-2
-~ .!.+_1_=Y 1+~ ..!__ ......, .!.+_1_=

LI k
n-2
~
LI k
LI k
n-2
k=i
k=i
k=i
k=i
an-t +~.
a2
an-t
9. Multiplying both the numerator and denominator of
each fraction on the left by the conjugate of the denominator, we get
=n-2+-1-= (n-2) d +
n-2
S= -Vii;-~+ -Vii;--Vll;
a2-at
a3-a2
=
! (Va2 -
(n-2) d
+ ... + -V~- -V~ =

an-an-t
Vat+ V a3- V a2 + ... +Van -
-V an-t) =
-V~- -Vlli
d
since
a2-at=a3-a2= ... =an-an-t=d.

Hence
S- -V~- -Vii; _
-
an-at
(Yan+ -Vat) -
n-1
-Van+ -Vat .
10. We have
a~-ai= (at-a2)
a:-a! =
(a1 +a2) = -d (a1 +a2),

(a3-a4) (a3 + a4) = -d (a3 + a4),
a~k-1 -ah= (~k-t -a2k) (~k-t +a2k) = -d (a2k-t + a2k)
Therefore
S= -d(at+a2+a3+a.+ ... +a2k-i+a2k)= -dat1;a 2" 2k.
366
Solutions
But
a2k
= a1 + d (2k-1),
a1 - a 2k =
-d (2k-1),
consequently,
S = -d (2k- 1)
a~:~~k
k = 2kk_f (a!- ah).
11. 1 We have
S (n+ 2)-S (n+ 1) =an+ 2,
S (n + 3)- S (n) =an+!+ lln+2 + lln+3
Consequently, we only have to prove that
an+!
+ an+2 + lln+3 -
3an+2 =
But it is possible to prove that

ar+as
= lls+r
2
(if r and s are of the same parity).

Indeed,
llr +as= 2a1
+ (s-1) d + (r-1) d =
=
2 [ad- ( r-~ s
-1) dJ=
2ar+s ,
2
therefore
and, consequently,
an+! + lln+2 + lln+~ - 3an+2 = O.
2 First of all
S (2n ) - S (n ) =lln+i + ... +a2n= lln+1+a2n
n.
2
Now we have
S (3n) = a1
+ llz + ... +an+ (an+1 + ... + ll2n) + a2n+1 + . +

+ a3n = lln+t 2+a2n n + ( lln + ll2n+1 ) +
+(an-I+ azn+2) + .. + (a1 +a3n)
But since the sum of two terms of an arithmetic progression equidistant from its ends is a constant, we have
all
-f-- az11+1 =
+ a3n =
an-I+ a2ii+2 = =a 1
lln+l
+ ll2n
367
Solutions to Sec. 'l
Therefore
S (3n)= an+1;ta2n n+(an+1+a2n)n=3 an+1ta2n n=
=3 (S (2n)-S (n)).
12. According to our notation we have
+ llck-1>n+2 + ... + llknt

+ llkn+2 + ... + ll1k+t)n
sk = ll1k-1)n+1
sk+I =
lllm+t
Consider the difference

We have
S k+l - S k = (llkn+n - llkn)
+ + (llkn+2 -
llck-!)n-d
+fakn+1-
+
G11t-11n+d
But since
am-llz= (m-l) d,
we have
Sk+1-S1i
= nd+ ... +nd+nd = n 2d.
13. We have
b-a=d(q-p), c-b=d(r-q),
c-a=d(r-p);
on the other hand,

a= u1wP-l,
b = U1Wq-i,
c=
u1w'- 1 ,
where u 1 is the first term of the geometric progression,

and w is its ratio.
Therefore
ab-c. be-a. ca-b = ad1q-r). bdcr-p). cdc p-q) =
ud<q-r1+dcr-p)+dcP-q>. (l)d{<q-T)(P-l>+cr-P1cq-l>+<P-Q)(T-ll}.
1
But it is easily seen that

d (q-r) +d (r- p) +a (p-q) =0,
(q-r) (p-1)
And so
14. We have
+ (r- p) (q-1) + (p-q) (r-1) =0.
3ti8
Solutions
Consequently
n2
(1 +x+x2+
... +x)
-xn ~-= (
(xn+l-1)2-xn (x-1)2
(x-1)2
xn+l -1
x- 1
)2 -Xn=
x2n+2-2xn+1+1-xn+2+2xn+l-xn
(x-1)2
(xn-1) (xn+2-1)
(x-1)(x-1) =(1+x+x2+ ... +xn-1)X
X
(1 + x + x 2 + ... + xn+i).
15. Let the considered geometric progression be

Hence
S3n - S2n
But
= U2n+I + + U3,,,
82,, - Sn= Un+I
+ + U2n
Therefore
consequently,
+ ... + U3n = q n (u1 + u2+ ... +un) = q nS

Sn= Un+I + ... + ll2n = qn (u1+u2+ ... +un) = qnSn
S3n - S2n = Uzn+t

S2n -
11 ,
Therefore
Sn (S3n -S2n) = q2 ns;, (S2n -Sn) 2 = q 2 ns~,
and the problem is solved.
16. Using the formula for the sum of terms of the
geometric progression, we get
1 1
S = anq-a1
q-1
S'
'
an q
a1
_!__1
q
Consequently
P 2 = (a1a2 ... an) 2 = (a1an)n,
Solutions to Sec. 7
369
hence
n
P=( ;, )2.
17. Let us consider Lagrange's identity mentioned in
Sec. 1 (see Problem 5)
(x~ +xi+ ... + x;_1) (y~ +Yi+ + Y~-1)
--(X1Y1+X2Y2 + + Xn-1Yn-1) 2 = (X1Y2- X2Y1) 2+

+ (X1Ya- XaY1) 2+ + (Xn-2Yn-1- Yn-2Xn-1) 2
Put
Y1=a2,
We then have
Y2=a3, .,
Yn-1=an.
(a~+ai+ ... +a;_i) (ai+a;+ ... +a;)-
-(a1~ +a2aa+ ... +an-1an) 2 = (a1aa-ai) 2+

+(a1a4-a3~) 2 + .+(an-2an-a;_1) 2.
()
The bracketed expressions on the right have the following

structure
and k + s = k' +s'. It is evident that if ai. a2 , , a 11
form a geometric progression, then (provided k + s = k'
+s')
Indeed
ak = a1qh-i,
ah=a1qh'- 1 ,
as= a1q- 1 ,
as=a 1q'- 1
Therefore
and
akas = ah'as'
Thus, if a 1 , a 2 , , an form a geometric progression,
then all the bracketed expressions in the right member of
the equality () are equal to zero, and the following rela-
mo
Solutions
tion takes place

(a;+a~+ ... +a;._1) (a~+a;+ ... +a;.)=
= (a1a2 + a2a3 + ... + lln-1Gn) 2.
Now let us assume that this relation takes place. It is
required to prove that the numbers a 1 , a 2 , , an form a
geometric progression. In this case all the bracketed expressions in the right member of the equality (*) are equal to
zero. But among these expressions there is the following one
(a1ak - a 2ak-1) 2 (k = 3, 4, ... , n).
Therefore we have
~=~
alt-1
(k=3, 4, ... , n),
a1
i.e. the numbers ai. a 2 , . , an really form a geometric

progression.
18. 1 It is known that
S - amq-a1
m-
q-1
Let us make up the required sum. We have

S +s +
+s - aiq-a1 +a2q-a1+
+anq-a1_
2
n-
(a 1
+ a2 +
20
1
U2(j2+
... + a2n-11t -
q-1
q-1
1 {1
1}
= 1-q2
li2
+(li"12 + ... +a"2""1
=
1
n_1_.
1
= 1-q2
30
q-1
a1 n
(anq-a 1) q
a1n
q-1 =
(q-1)2 - q-1
... +an) q
q-1
a2
a~-1
_!_ __
1
q2
1
q2
--1
af
(~-__!.__)
, a~
af
q (1-q2)2
qk
1-q21i
1
a~
1 )
a~
19. Let the given progression be a 1 , a 2, ... , an. Let a;;_

designate the kth term from the end of the progression. Then
al. = an - (k - 1) d, ak = a 1 + (k - 1) d.
371
Solutions to Sec. 7
Consider the product aka"k. We have

akak = a1an - (k - 1) 2 d2 + (k - 1) d (an - a 1) =a 1an - (k - 1) 2 d 2 + (k - 1) (n - 1) d2.
And soakak = a1an + d2 {(k - 1) (n - 1) - (k - 1) 2}.
It only remains to prove that the expression
Pn = (k - 1) (n - 1) - (k - 1) 2
.
. h an mcrease
.
. n f rom 1 to n or -n+i
mcreases
wit
m
2
2 -.
We have
Pk = (k - 1) (n - k),
Pk+ 1 = k (.n - k - 1).
Hence
Pk+1 -
Pk =
n - 2k.
Consequently, Pk+i >Pk if n - 2k > 0, i.e. if k <~and

our proposition is proved.
20. Let a1, a 2 , , an be an arithmetic progression,
and ui. u 2 , , Un a geometric progression. By hypothesis,
a1 = uh an = Un Let the ratio of the progression be
equal to q. Then
Put
a1
+ a2 +
Prove that
We have
a1 +an
a1
2- n =
Sn= -
+a1qn-1 n =
2
unq-u 1
On= q- 1
a1
+ qn-1
2
n,
qn-1
a1--1-.
q-
Since, by hypothesis, a 1 > 0, it only remains to prove

that
372
Solutions
Let us write the left member of the supposed inequality

in the following way
qn-11
q-
=1+q+q2+ ... +qn-3+qn-2+qn-1=
= ~ {(1 + qn-1) + (q + qn-2) + ... + (qk + qn-k-1) + ... +

+ (qn-1+1)}.
Let us prove that

Indeed
qh + qn-k-1 -1-- qn-1 == (qk -1)
+ qn-h-1 (1- qk) =

=
(qk -1) (1-qn-k-1) <: 0,
since if q > 1, then qk - 1 ~ 0, 1 - qn-1<-l ~ 0, and if

q < 1, then qk - 1 ~ 0, 1 - qn-R.-l ~ 0. At q = 1 it is
clear that the product contained in the left member of our
inequality is equal to zero. And so, indeed,
qk qn-k-1 -< 1 qn-1.
The braced expression contains n bracketed expressions each

qn- 1. Therefore
of which does not exceed 1
qn-1
~<n
1+qn-1
'
i.e.
which solves the problem.

21. Let the first common term of the progressions be a,
and the second b. Then the nth term of the arithmetic progression will be equal to
+ (b -
a) (n - 1),
and the corresponding term of the geometric progression has
the form
a
And so, we have to prove that
a+(b-a) (n-1) <a (:
r-
Solutions to Sec. 7
373
in other words, that
a+ (b-a) (n-1)-a ( ~ }n-l -:<O,

or
a{(
! - 1 ) (n -
1) __;_ [ ( : ) n- l
J} -:< 0.
Let us rewrite the left member of this inequality as follows
a ( : - 1 ) { (n-1 )-[ (
!)
n- 2 +
!)
n- 3 +
... + (
Considering separately the three cases: ~

a
>
! )+ 1J}.
1, ~
a
<
1,
~
= 1, we easily prove the validity of our inequality.
a
22. We have to compute
Sn = 1 X + 2x 2 + 3x3 + ... + nxn.
Multiplying both members of this equality by x, we have
Snx = 1 x2 + 2x3 + 3x4 + ... + (n - 1) xn + nxn+1
It is evident that the right member is equal to
Sn - x - x 2
x3
Thus, we have the identity
xn
+ nxn+1.
+ ... + x11-1),
Snx= Sn+ nxn+i_ x(1 + x+ x2
S n (x-1) = nxn+i _ xxn-1

- -1-,
x-
Sn (x-1) =x{nxn+i+1-(n + 1) xn}.

2
And, finally, we have

Sn= (x~ 1)2 {nxn+ 1 -(n + 1) xn+ 1}.
23. We have
374
Solutions
Let u~ multiply both members of this equality by q (where q

is the ratio of the geometric progression). We obtain
n
sq=~
ahuh+1
k=i
(since u 11 q = uh+ 1).

Subtract s from both members of the last equality. We
have
n
sq-s= ~
ahuh+i-
k=1
a1iuh.
k=i
Transform the right member as follows

n+1
n+i
a1i_1u1i -
=2
a1iu1i -
k=2
n+i
= -
(ah -
a 1u 1
lln+1Un+1
a1i-1l Uk -
ll1U1
k=2
=-
lln+1Un+1
n+I
= - ~
duh+ lln+1Un+1 -
ll1lli.
k=2
where d is the common difference of the arithmetic progression.

Thus
Finally
_
lln+1Un+1-a1u1
S-
q-1
ll11+1q-u2
(q-1)2
24. The required sum can be rewritten in the following

way
1 +-.
1 + ... +.1 + 2n.
x 2+ x 4+ + x 2n +2
x
x n
x~
Summing each of the geometric progressions separately and

joining the partial sums thus obtained, we have
(x
+ ! )2+ ( x2 + :2 )2+ ... + ( Xn + :n )2=

=
(x2n+2+1) (x2n-1)
(x2-1)x2n
+2 n.
375
Solutions to Sec. 7
25. The sum 8 1 is readily computed by the formula for

an arithmetic progression. Let us now compute S 2 Consider
the following identity
(x + 1)3
x3
3x2 + 3x + 1.
Putting here in succession x = 1, 2, 3, ... , n and summing up the obtained equalities termwise, we have
n
~ (x+1) 3 -
~ x3=3 ~ x 2+3 ~ x+n.
x=i
x=i
x=i
n
x=i
Or
{2a+3s+ ... +ns+(n +1)3}-{t3+2a+ ... +na}=

=3S 2 +3S1 +n.
And so 3S 2 +3S1+n=(n+1)3 -1. But
S 1_- n(n+1)
2
.
Now we find easily
Sz=n (n+1)6(2n+1).
The formula for S 3 is deduced in a similar way. We only
have to consider the identity
(x + 1) 4
x 4 = 4x3 + 6x2 + 4x + 1
and make use of the expressions for S 1 and S 2 found before.

26. We have identically
(x+ 1)h+i_ xJt+r == (k+ 1) xh +(kt.~ k xk-1+
(k+~).~.(~-1) xh-2+ ... + (k+ 1) x+ 1.
Putting here successively x = 1, 2, 3, ... , n, and summing up, we get the required formula.
376
Solutions
............
.
1h
211
311
411 ... nll
The sum of terms of each line is equal to 1 11 + 2 11 +

+ ... + n 11 = S 11 (n). Thus, the sum of all the terms of
the table will be nS 11 (n).
On the other hand, summing along the broken lines, we
get the following expression for the sum of all the terms of
the table
111 +(1 11 +22 11 ) + (111 +2 11 +33 11 ) + (1 11 +2 11 + 3 11 +44 11 ) +
+ ... + (1 11 +2 11 +3 11 + ... + (n-1) 11 + nn 11 ) =
= 1 + [S11(1)+2 11 +11+ [S11 (2) + 311 +11 [S11 (3) 4 11 +11+ ... +
+ [S11 (n-1) + n 11 +1 ] =
=S 11 (1)+S11(2)+ ... +S11(n-1)+

+ (111+1+2ii+I + 311+1 +
And so
llh+1).
nS11 (n)=S11+1 (n)+S11 (n-1)+S11 (n-2)+ ... +s11 (2)+S~(1).

28. Both 1 and 2 are readily obtained from the formula of Problem 26. Let us rewrite it as
ks
k(k-1) 8
s
s0
S k = -2
II-I 12 3
k-2 I k +1+
+ (n+k1~~1-1
n2+n
(*}
At k= 1S 1 =1+2+3+ ... +n=-2- -:-2n +2n.

Thus, both propositions (1 and 2) are valid at k = 1.
Suppose they hold true for any value of the subscript less
than k and let us prove that they are also valid at the subscript equal to k. Since, by supposition, S 11 _, is a polynomial
in n of degree k, sk-2 a polynomial of degree k - 1, and
so on, it is easily seen from the formula() that S 11 is indeed
377
Solutions to Sec. 7
a polynomial of degree k + 1. Further, since Sk_,, Sk_ 2 , ,

S 0 do not contain the term independent of n, it follows
(n+1)k+l_1
that S k also does not contain such a term (

k-l- i
,
when expanded in powers of n, will not contain a constant
term). As is evident from the same formula(*), the coefficient of the term of the highest power in the expansion of
S k in powers of n will be
1 . It only remains to prove
that the coefficient of the second term, i.e. B, is equal to
~ . In the expansion (*) there exist only two terms contai-
k!
~ Sk_,, and the
ning nk. One of them is contained in other in
(n+1) 11 +1_1
. From what has been proved we have
k+i
~ S k-t =
nk
+ ... } =
nk
+ ...
Further
(n+1)k+l_1
R+l+
k+i n
k-t 1
n '
Hence, it is obvious that

1
B=2
As to the structure of the rest of the coefficients (C, ... , L),
we may assert the following: the coefficient of n" +t-l will
be equal to
1
A
C,.+ 1 k+1'
where A is independent of k. This proposition is proved using

the method of induction with the aid of the formula (*).
29. S 4 can be computed using, for instance, the formula
from Problem 26.
However, we may also proceed in the following way.
From the result of the previous problem it follows that
1
S4 = 5
+ 21 n + Cn
+ Dn 2 +En.
It only remains to determine C, D and E. Since the last
equality is an identity, it is valid for all values of n. Putting here in succession n = 1, 2, and 3, we get a system of
equations in three unknowns C, D and E. Namely, we have
n5
378
Solutions
13
89
c+D:+-E=ro, 8C+4D+2E=5 '27+C+9D+3E=10.

Hence
C= 3 , D=O, E= -
30 .
It only remains to factor the expression

n&
n'
n3
5+2+3-30
and the required result will be found.
The remaJning three formulas are obtained similarly.
30. The validity of the identities is established by a
direct check, using the expressions for Sn obtained before.
31. Put k = 1. We have
(B + 1) 2 - B 2 = 2,
or
B2
2B,
1 - B 2 = 2.
1
Consequently, B, = 2.
Then. put k = 2. We get
(B
1)3 - B 3 = 3,
i.e.
1
B3 +3B 2 +3B,+1-B 3 =3, i.e. B2=6.
Proceeding in the same way, we get the following table
B1=2'
Bs=42'
3617
B16= -5To'
Bu=O,
691
Bi 2 = -2730'
1
Bs= - 30 , B 13 =0,
B5 =0,
B 10 = 66
B 15 =0,
Knowing this table, we may easily solve Problem 29, i.e.

arrange S 4 , 8 5 , S 6 and S 7 according to powers of n. These
numbers play quite an important role in many fields of
mathematics and possess a number of interesting properties.
They are called Bernoulli's numbers (J. Bernoulli, Ars
Conjectandi). We can show that for odd k's exceeding unity
Bk will be equal to zero. And Bernoulli's numbers with an
379
Solutions to Sec. 7
even subscript will increase rather fast. Let us consider

the value of B 19 ~. If we put B 196 = that
then it turns out
N = 171390,
z =62753 13511 04611 93672 55310 66998

93713
39107
54578
32214
31316
71217
31115
60315
01782
57614
56098
61171
06735
30541
46402
21158
29255
66687
79434
53311
41378
35788
49798
49347
16524
89530
48048
96086
68376
22145
98443
55906
46255
55345
27052
80056
87718
Thus, the numerator of this number contains 215 digits

(D. H. Lehmer, 1935).
Let us now prove relationship 2.
On the basis of the results obtained in Problem 28 we
may put
(k+1) (111 +2 11 +3 11 + .. . +nk) =
1 n 11 +cn11 - 1+Dn"-2+ ... +Ln,
= n 11 +I
+kt
where C, D, ... , L are independent of n, but undoubtedly

depend on k. Put
(k+ 1) (1 11 +2 11 +3 11 + ... +n") = n 11 +1+CJ.+1a 1n 11
2
"-1+ . . . +c"-1
+c11+1a2n
1i+1ak-1n2+c"1i+1a1in.
We may then write the following symbolic equality
(k + 1) (1 h+2h + ... + n 1') = (n + a)h+i _ ah+i.
On removing the brackets in the right member by replacing
a by CX 8 (s=O, 1, 2, ... ), we pass over from the symbolic
equality to an ordinary one.
Since this equality is an identity with respect to n, we
may put in it n + 1 instead of n and obtain
(k+1) [1h + 2h + ... + (n + 1)h] = (n + 1 + a)h+t_ah+i
Subtracting from the last equality the preceding one, we
find
(k+ 1) (n + 1)h = (n+ 1+a)h+t_ (n+a)h+t.
Putting here n = 0, we have
(a+ 1) 11 +i-cxll+i =k+i.
Solutions
380
Besides, it should be remembered (see the solution of Problem 28) that a:'s are independent of k and that a: 1 = ; .
And so, the numbers a:k and Bk are determined by one
and the same relation, and a: 1 = B 1 Therefore
a.k =Bk
for any k.
32. Let d be the common difference of our progression.
Then
Xk = X1
d (k - 1).
From the first equality we have

x1+xn
2
nx 1 +d
n=a,
n(n-1)
1.2
=a.
On the other hand,
xii= x~+2x 1 d (k-1)+d2 (k-1) 2

Therefore, from the second relation we get
n
~ x)i=nx~+2x 1 d ~ (k-1)-t d2 ~ (k-1) 2 =b2
k=1
k=1
k=1
Hence
dn(n-1)+d2 (n-1)n(2n-1)=b 2
( 1)
nx1, 2X1
12
6
(see Problem 25).
Squaring both members of the equality (*) and dividing
by n, we find
dn(n-1)+d2 n(n-1)2 =.!!!__
2 1_
(2)
nx1' 2 Xi
1.2
4
n
2 1_
Subtracting (2) from ( 1), we get

d2n(n2-1)
12
b2n-a2
n
Consequently
d=
+
-
2-V3 (b2n-a2).
n Vn2-1
Substituting d into the equality (*), we find Xi, and, consequently, we can construct the whole arithmetic progression.
Solutions to Sec. 7
381
33. 1 Put
S=
n
k 2x 11 - 1 .
Hence
XS=
k=I
k 2x 11 .
k=I
Subtracting the first equality from the second, we find

n+I
s(x-1)= ~ (k--1) 2 xll-l_ ~ k 2 x 11 - 1 .

k=I
11~2
Consequently
n
k=l
k=I
(see Problem 22).

Finally
1+4x+9x2+ ... + n 2 xn- 1 =

n2xn (x-1)2-2nxn (x-1)-+ (x"-1) (x+ 1)
(x-1)3
2 Proceed as in the previous case. Put

n
S=
i3+2ax+33x2+ ...
+naxn-1
~ kaxk-1.
k=I
Make up the difference

n
s x - s = n 3 xn_3
k 2 x"- 1 +3 ~
k=I
kx"- 1 -
k=I
x 11 - 1
k=I
Substituting the expressions obtained before for the sums

on the right, we have
S (x-1) = n3xn _ 3 n 2x"(x-1) 2 -2nxn (x-1)-+(xn-1) (x+1)
(x-1)3
+3
nxn+1-(n+1) xn+ 1
(x-1)2
xn-1
x-1
382
Solutions
Finally
s (x-1) 4 = n 3 xn (x-1) 3 -3n 2xn (x-1) 2 +
+3nxn (x 2 -1)-(xn-1) (x 2 +4x+ 1).

34. To determine the required sums first compute the
following sum
n
1 +3x+5x 2 + ... + (2n- 1) xn- 1 = ~ (2k-1) x 1H =

-2
-
n
"'i;:~
Li
k=I
k-l_
X
k=I
"1
.L.J
1<- 1
X
_2nxn(x-1)-(x+1)(xn-1)
(x-1)2
h=I
For computing the first of the sums put in the deduced

formula x = ~ . We then have
3
5
7
2n-1
1
n
1+2+-.r+s+ ... + 2n-1 = 2i-1 {3(2 -1)-2n}.
And putting x =
we fmd
1-~ +!:...-2-+
+ (- 1)n_ 1 2n-1_2n+(-1)n+1 (6n+1)
s
211-1 9.2ii-1
2 4
35. 1 First assume that n is even. Put n ==-2m. Then
1-2+3-4+ ... +(-1)n- 1 n=
= 1-2+3-4+ ... +(2m-1)-2m= (1 +3+ ... --\
+2m-1)-(2+4+ ... +2m)=-m=- ~.
Now let n be odd and put n = 2m -1. Then our sum
takes the form
[1-2 +3-4+ ... -(2m-2)] + (2m-1) =
n+- 1
= -(m-1)+2m- 1 = m=2
Thus, if we put
1-2+3-4 + ... + (-ft- 1 n =S,
then
. even, s = -n+i
. odd .
S = - 2n I"f n IS
Ir n IS
2However, this result can be obtained in a simpler way.
Indeed, if n is even, we have
S=[1- 2]+(3--4]+[5-6]-t- ... +[(2m-1)-2m]=
-
-im= -m=
Hence we also get the result for odd n.
_.!!_
2 .
Solutions to Sec. f
383
2 First assume that n is even and put n = 2m. We

have
12-22+32_ ... +(-1)n-ln2=(f2-22)+
+ (3 2-4 3) + ... + [(2m -1) 2-(2m) 2] = -(1+2)--(3 +4)- ... -(2m-1+2m)= -[1+2+3+4+ ... +
(2m+ 1} 2m
n (n+ 1)
2
=2
+2m-1+2m]=
Thus, if n is even, then
f2-22+32_ ... +(-1)n-1n2= _n(;~1).
If n = 2m 1 is odd, then
i2-22+ 32_ ... + (-1)n-1n2= i2-22+32_42_ ... _
-(2m) 2+ (2m+ 1) 2= -
2m
~m+i) + (2m+ 1)2 =

=n2
n(n-1)
n(n+1)
2
=
12
3 The required sum is equal to - 8n2 The result is

obtained as in the previous case.
4 Rewrite the required sum as
n
~ (k3 +k2) = ~ k3 + ~ k 2 =
k=1
k=1
n (n+1)
(3;;+7n+2~
k=1
(see Problem 25).

36. The. considered sum may be rewritten as
10-1
102-f
103-f
fOn-1
-9-+-9-+-9-+ ... +-9-

wherefrom we easily find its value
!{
1010n9-1 _n }
37. Consider the first bracketed expression on the right

and rewrite it in the following way
2x2n+1_2x2n-1y2 + 2x211-3y4- ... 2xy2n _ x2n+1 =
x2n+2 + y2n+2
=2X x2-ty2 - x- +1.
q
11
Solutions
384
The second bracketed expression arises from the first one

as a result of permutation of the letters x and y, therefore
x2n+2+y2n+2
it is equal to 2y x2 +Y 2
y 2n+ 1. Squaring both obtained
expressions and adding the results, we easily prove the
validity of the identity.
38. The required product is equal to
(1a+1a2 + ... +1an- 1)+(aa2 --j- --j-aan- 1)+
+ (a2a3 + ... + a2.an-1) + ... + an-2 an-1 =
= a ( 1 --j- a --j- --j- an- 2 )
--j- a 5 ( 1 +a --j-
+a
+ an-4) +
3 (
1 +a+ ... --j- an- 3 ) +
... + a 2n- 5 ( 1 +a)+ a2n- 3 =
an-1_1
an-2_1
an-3_1
=a--1-+aa--1--t-as
aaa- 1 + ... +
+ an+2+ ... + a2n-a+a2n-2)-(a+aa +as+ ... +a2n-s +

I
a2n-3)}- (an- f) (an-a)

- (a-1) (a2-1)
39. The sum on the left may be rewritten as follows

1 +-;n=2+
2
n-1) + ( Xn-1 + 2 Xn-2 --j- + (n. 1) x J+ n.
( xn-1
+-x-
The first bracketed expression is equal to

_1_( +2 2
xn
+ +( n _
1) X n-l]=x[(n-1)xn-nxn-1+1J
xn (x-1)2
(see Problem 22).

The second bracketed expression is obtained from the
first one by replacing x by_.!.._. Hence, we get the required
x
result.
385
Solutions to Sec. 7
40. 1 We have
2.3=-z-3'
1
3.4=3-4
1
n(n+1)
=-n- n+1
Adding the right and left members, we get the required

result.
2 The required sum may be rewritten in the following way
n
~ k(k+1~(k+2)
S=
"
k=i
But k(k+1)(k+2) =z-T- k+1 +2 k+2 .

Therefore
1'=1
386
Solutions
Hence
168 =
n
"\' 16k4 -1+1
LJ
= "\' (4k 2
LJ
4k2-1
k=1
k=i
+ +.!.
168 _ 4 n(n+1)(2n+1)
-
--t-
1)+.!.
"\'
(2k+1)-(2k-1)
2 LJ (2k-1) (2k+1) .
k=1
1-_}
+ '
~ (-1_ _ _
2 ..:.J
2k -1
2k 1
k=i
16s - 2n(n+1)(2n+1)
-
+ n +.!..2 { 1 ._ _!_..L_!_
_ _!_+.!.+
+3;3
5
5
+ n +-n2n+1 '
168 - 2n(n+1)(2n+1)
-
1
+ 2n-1
-
1 }
2n+1 '
Finally
1
2m'
168
where m = 2n+ 1.
42. We have
a1~n
= at +an .
a~~=n
=at+ an (
L+
:1 ),
_1___
1
a2+a,,_1 _
1
(-1-+_1_)
a2an-I - a2 +an-I a2an-1 - a2 +an-I a2
an-1 '
But
llt
lln
= ll2 + lln-1 = ll3 +lln-2 =
Therefore, adding our equalities termwise, we find
_1_-+- ... +-1-=

2
(__!__+_1
+ ... +-1
)
anal
a1 +an
a1
a
an
_1_+
a1an
a2an-t
43. 1 It is obvious that tho following identity takes

place
1
(n+k-1)!
(n+k)! =
n+k-1
(n+k)!
Putting k= 1, 2, ... , p+ 1 and adding the obtained equalities termwise, we prove that
n
, n+1
(n--f--1)! T (n--1--2)!
+ +
n1 p
(n+p+1)!
=-;;r-
1
(n+p+1)!
Solutions to Sec. 'l
387
2 We have
n
-!-n
(n+1)! ' (n+2J!
+ + (n+p+1)!
n
<
n+1
n
(n+1)!
n+p
+
1
+ (n+2)! + + (n+p+1J! =nr- (n+p+1J!

(see 1).
Therefore
1
(n+1)!
1{1
1
}
-;;r-(n+p+i)!
+ (n+2)1 + ' + (n+p+1)1 <;
44. The following identity holds true

1
-;=y--
1
z+1
2
z2-i '
In our case we have

1
x-1 -
x+i
x2-1 '
1
x2-1 -
i
x2+1
2
x4-1 '
(2)
1
x4-1 -
1
2
x4+1 = x8-1 '
(3)
1
x
2n
2
1-
2n+I
1 .
(n+ 1)
Multiply both members of equality (1) by 1, of equality (2)

by 2, of equality (3) by 2 2 and so forth, finally, multiply
both members of the equality (n + 1) by 2n. Adding the
obtained results, we find
Solutions
388
45. We have
n-p
1
~ (1-- n-p+1
p+k
1
) n- -k 1 + n-k-j-1
(
p
+ )--
h=t
n-p
"(n+1
n-p+1 ~
p+k -
p)
n-k+1
h=1
[-
= n-p+1 _(n+ 1)
1
1
p+1 + P+2 +
+-n1) -
p~1 ) J=
1
( 1
1)
= n-p+1
P+1 + +-n (n-1-1-p)=
1
1
(1 1
= P+i
+ ... +n=
+2+ ... +n1) -
- p (
~ +-;!_1
+ ... +
1
1
1)
- ( 1+2+
3 + ... +P =Sn-Sp.
46. We have
,
n+1
{
1
Sn= - 2 - - n(n-1)
+ (n-1)2(n-2)
n-2}
+~
=
n-2
_n+1 __ "1
k
2
~ (n-k+1)(n-k) =
h=\
n-2
n+1
"1
-k
= - 2 - + ~ (n-k+1)(n-k)
h=1
Let us expand the fraction (n-k+~~(n-k) into two partial

fractions. Namely, let us put
-k
(n-k+1)(n-k)
A
n-k+1
B
n-k '
-k= A (n-k) +B (n-k+ 1).

Hence, putting first k = n and then k = n 1, we find
A=n+1,
B= -n.
Solutions to Sec. 7
389
Therefore
n-2
n-2
n+ 1
"",
1
"'1
1
S=-2-+(n-j-i)L.J n-k+1-nL.J n-k=
h=1
h=l
= nt1 -j-(n+1)
(~
n~1
+ ... + !)-
-n (-1-+_1_+ ... +.!.)

n-1
n-2
! + n~1 + ... + ~) +
+ ( ~ + n~1 + + ! )-n [ ( ~ + n~1 + + ! )- ! + ~ J
= nt1
-j-n (
n+ 1
( 1'
1
n
=-2-+
n+ n-1
+ +31 ) --1- 1-z-=
1
1
1
= 1+2+3+ .. +-;.
47. Let the nth term of the required progression be an,

its common difference being equal to d. Then
S x = a1+ax
2
X,
k X.
S hx= a1+ahx
2
Hence
Skx
Sx
a1+ahx
a1+ax
k= 2a1+d(kx-1) k= 2a1-d+kxd k
2a1+d(x-1)
2a1-d+dx
For the last relation to have a value independent of x it is

necessary and sufficient that
2a 1 - d = 0,
i.e. the common difference of the required progression must
equal the doubled first term.
48. We can prove the following proposition
ah
if k
+l
k'
+ l'.
+ a1
ah'
+ a1
390
Solutions
Indeed
uk=a1+(k-1)d,
ah=a1
a1=a1+(l-1)d,
+ (k' -1) d,
a1=a1 + (l'-1) d.
Hence
+ (k+l-2)d,
2a1 + (k' + l' -2) d.
ah +a 1=2a1
ah'+ az =
But since by hypothesis
+l=
k'
+ l',
it follows from the last equalities that
And so we have
ai
+ ai+2 = ai+1 + ai+1 = 2a1+1
The given sum is therefore transformed as follows
But
ai
= ai+l -d,
ai+2
= ai+! + d,
therefore
n
S=
~ ~
(ar+ 1
-d 2)=
~ ~
[ar+2a1di+(i2-1)d2 J=
49. As is known
sin~
tan (a+k~)-tan [a+(k-1) ~]=cos(a.+k~)cos[a.+(k-1)~)
Solutions to Sec. 7
391
Therefore
n
"',
1
LJ cos(a.+k~)cos[a.+(k-1)~) =
k~1
~
SIIl p
"'1 {tan
LJ
(a+k~)-tan [a+ (k-1) ~]} =
k=-1
=-.-R {tan (a+~)-tana+tan (a+2~)-tan (a+~)+ ... +

sm p
+tan (a+n~)- tan (o:+(n-1) ~)} =
tan
(a.+.n~~ -tan a.
sm
50. We have
2cot2a-cota= -tan a,
a.
a.
a.
a.
2coto:-cot 2 = -tan 2 ,
a.
2cot 2 -cot 4 = -tan 4 ,

a.
2 COt 2n-2
COt
a.
2n-l
= - tan
a.
2n-I
! ,... ,
Multiplying these equalities in turn by 1, ~,

2 n~l
and adding termwise, we get the required result.
51. Consider the following formula
cos [a + (k - 2) h] - cos [a + kh] =
= 2 sin h sin [a + (k - 1) h].
Putting k = 1, 2, 3, ... , n - 1, n, we find
2 sin h sin a = cos (a - h) - cos (a + h),
2 sin h sin (a + h) = cos a - cos (a + 2h),
2 sin h sin (a + 2h) = cos (a + h) - cos (a + 3h),
2 sin h sin [a + (n - 2) h]
= cos [a + (n - 3) h] - cos [a + (n - 1) h],
2 sin h sin [a + (n - 1) h]
= cos [a + (n - 2) h] - cos [a + nh].
392
Solutions
Adding these equalities term by term, we fmd

2 sin h {sin a+ sin (a+ h) +sin (a+2h) + ... +sin [a+
+(n-1) hl} = cosa+cos (a-h)-cos (a+ nh)-cos [a+ (n-1) h] ={cos a-cos [a+ (n-1) h]} +
+{cos (a-h)--cos (a+ nh)} =
. n-1
. ( a +n-1
h sm
. ( a +n-1
h) + 2 sm
h) x
= 2 sin222. n +1
x sm
- 2-
h= 2 srn. ( a +-n-1
h) 2 sm. 2nh cos 2h
2-
Hence
sin a
+ sin (a + h) + sin (a +
2h) + ...
+ sin [a + (n -
1) h] =
. ( a+
n-1
h) sm2
. nh
sm
2.
sm
The second formula is obtained similarly. However, it can

also he readily obtained from the above deduced formula by
replacing a by ~ - a.
52. Putting in the previous formulas a = 0, h = _.::.,
n
we get
S =cot 2rcn , S' = 0.
53. Taking advantage of the results of Problem 51,
we have
. a srn
. ,3a+ ... sm
. [( 2n - 1) a ] = sin nasin
na
srn
.
,
S!Il a
cos a +cos 3'X + ... +cos [(2n-1) a]=
sin na cos na
.
.
sma
The rest is ohvions.

54. The required sums can be computed, for instance, in
the following way. Make up the sums S~ and s;. It is easily seen that
S~+S~=2n.
On the other hand,

S~
- S; =
COS
2x + COS 4x + ...
C05
4nx.
393
Solutions to Sec. 7
Using the second formula from Problem 51, we find

sin 2nx cos (2n
+ 1) x
cos 2x + cos 4x + ... + cos 4nx =- ----.--'--"--'"sm x

And so
S' -S" _ sin 2nx cos (2n+1) x
n
sinx
n-
S~+S~=2n.
Hence
S' _ n +sin 2nx cos (2n+ 1) x
2sin x
n-
'
S" _ n _ sin 2nx cos (2n + 1) x

2sinx
n-
55. Let us make use of tho formula

sinAsinB=; [cos(A-B)-cos(A+B)].
We then have
p
"\1 . mni
.
n:n;i
=Li Slil P+1 Slil P+i =
i=1
p
1 "\1
=2 L.l cos
(m-n) ni
P+i
1 "',
-2
i=1
L.l cos
(m+ n) ni
P+i
i=1
But if m + n is divisible by 2 (p + 1), then cos (m :_;~ ni =

=1 and
Using formula 2 from Problem .51, we easily find

p
~ cos
(m-n) ni __
t 1 --- p--
1.
i=I
Hence
__ p+1
8 -2 .
All the remaining cases are proved analogously.
394
Solutions
56. We have
arctan (k+ 1) x+ arctan ( -k.r) =
kx+x-kx
=arctan 1-(k+1)x(-kx) =arctan 1+k(k+1)x2 '

since (k + 1) x ( -kx)
Hence
<1
arctan 2x - arctan x =arc tan
x
1+12x2 '
arc tan 3x - arc tan 2x =arc tan
1 +~. 3xz
arctan (n + 1) x -arctan nx = arctan 1 + n (n + f) xi

Adding these equalities termwise, we find that
required sum is equal to
the
nx
arctan (n + 1)x-arctan x = arctan i+(n+f) xz

a11.-a1t-l
arctan a11. +arc tan ( - a11.-1) =arc tan 1 + a11.ah-t

=arc tan
1+a11.a1t-l '
Now we find easily that our sum is equal to
58. Put
1+k2 +k'= -xy,
x+y=2k.
(This is done to use the formula

arctan 1x+y =arctanx+arctany if xy < 1.)
-xy
Then
arctan (k 2 -t-k + 1) - arctan (k 2
k + 1),
395
Solutions to Sec. 7
therefore
n
~
2k
L.l arctan 2 +k 2 +k 4 =arctan3-arctan1+arctan 7k=i
-arctan 3 + ... + arctan (n 2 +n+1)-arctan(n2 -n+ 1) =

=arctan (n2 +n+1)- ~ .
59. Let k be one of the numbers 1, 2, ... , n-1. Multiply the first equation by sink.!!:....,
the second by sink~
,
n
n
the third by
sink~
and, finally, the last one by
n
(n- 1) 11
n
Adding the obtained products termwise, we
sink
find
+an-t sm k
(n-1) 11
n
And
11 . k -+sm
11
. l -sm
211 . k -+sm
211
. l -srn
311 . k -311
. l -srn
A z=Sm
n
n
n
n
n.
n
. k
+ . . . + sm. z~(n-1)11
Sm
n
(n-1)11
n
Taking advantage of formula 2 of Problem 51, let us

prove that
A1 = 0
if l ::/= k,
A1
=;
if
l =k.
Hence
2(
.
11
211
.
\n-1.)11)
x,. = a1 sm
k-+~ cos k - + ... + an-i srn k-'--"-n
n
n
n
(k= 1, 2, 3, ... , n-1).
396
Solutions
1. We have
1
2n =
1
2n '
1
2n-1
> 2n '
'
+ 2 > 2n '
+ 1 > 2n
Adding these inequalities termwise, we find

1
n+1
1
n+2
+ .+2n
> 2n+2n+ +2r1=2r1=2
1
(n+k+1)(n+k)
< (n+k)2 < (n+k-1) (n+k)
But
1
(n+k+1) (n+k)
1
(n+k-1) (n+k)
n+k -
n+k+1 '
n+k-1 -
n+k '
therefore
1
n+k -
n+k+1
< (n+k)2 < n +k-1
1
n+k
Summing these inequalities (from k= 1 to k= p), we

get the required relation.
3. Let us have n fractions (n~1)
1
a'
b'
-c a' ... ,", T
Let us assume
2~a<b<c<d<
. .. <k< l.
Then
b>a+1, c~b+1, d:;;:,c+1, ... , l~k+1.

Consequently
h~a+1, c;;;;,a+2, d;;;;,a+3, ... , l~a+n-1.
Therefore
1
az+/iZ+ +12~ll2+
1
(a+1)2
+ +
1
(a+n-1)2
<
1
a-1 -
a+n-1
397
Solutions to Sec, 8
Hence
1
ll2+-w+ +v<
(a-i)(a+n-1)
But
a-1>1, a !-n-1>n+1, (a-1) (a+n-1)>n+1

and
1
-;2+-w+ ... +v::::;;; n+1 < 1.

4. Indeed
(n!) 2 = (1 n) (2 (n - 1)) ... (n 1).
But
k (n -
since
k (n - k
Therefore
+ 1)
+ 1) > 'n,
- n = (n - k) (k - 1)
0.
1 n = n,
> n,
2) > n,
2 (n - 1)
3 (n -
n1 = n.
Hence
5. Since
a<VA<a+1,
we have
J/ A+a<2a+1,
lfence
CVA'+a) CVX-a) <VA2a+1

a,
A-a2
2a 1
+ <
i/-
A - a,
i-
JI A
A-a2
> a + 2a --1- 1
Let us now prove the second inequality.

For any x there exists the following inequality
x (1-x) = x-x2-<
!.
398
Solutions
Indeed, we have
x-x ! = -( x-; r~O.

2-
It is obvious that we have an equality only at x = ; .
Since it is possible to assume that VA - a =I= ; , we have
[1-(VA-a)] (VA-a)<! ,
1-(VA:-a) < 4T-V~-a) ,
<2a
>
+ 1) - ( V71 + a) < 4 CV 1A-a )
Multiplying both members of this inequality by VA

0, we find
(2a + 1) CV A-a)- (A-a 2 ) < -~ .
-a>
Whence finally
,-
A-a2
VA<a+ 2a+1
6. We have
v;; >
since
V n- + 1 -
+ 4(2a-t-1)
2Vn+1-2 Vii,
v-n = , /
,/
vn+i+vn
Consequently,
1 >2
~2
1
< --,
;- .
2vn
V2-2,
> 2 V3-2 v2,
~ 3 > 2 Y4-2 J/3,

1
-Vn
>2
vn + 1-2 vn.
Adding these inequalities, we obtain the required result.
399
Solutions to Sec. 8
7. Put
A=
1s C2s = ~ . ~ .
i- .. . 2s~
Then
24
2s
246
2s
A<35 2s+1 =y35 2s-1
2s+1 '
i.e.
1
A<A2s+1
Hence
A2
< 2s~1
, A< -V2:+1

1
A>2''3'5'"
2s-2
2s-1'
A =..!...~.~
2s-1
2s
6 "
Multiplying these relationships, we find
A>
8. Since
-v- .
1
8
2tan 2
tan e = - - - 8 - ,
1-tan2 2
we have
1-
8
cot2 cote= - - - -22 -18
cot T
8
cot2 2 -1
2cot
Consequently
8
1+cot 8-cot -z= 1 +
-1
2cot
8
cot2-z-1
8
8 -cot-z=
2cot 2
{ cot 28 - 2 cot 8
2
2
+ 1} =
( 1-cot
2cot 2
:::;;o,
400
Solutions
since
8
cot 2 > 0 (0
< e< rt).
9. We have
tanA-j-tan B
tan(A+B)= i-tanAtanB =tan(n-C)= -tanC>O,
since C is an obtuse angle.
And so
tanA+tanB >O
1 - tan A tan B
But since A and B are less than ~ , it follows that

tan A+ tan B > 0, and hence
1 - tan A tan B > 0, tan A tan B
1.
<
10. Indeed
tan(El- )= tan8-tancp =(n-1)tancp
<:p
1+tan8tancp
1+ntan2rp
Therefore
)
(n-1) 2
\n-1)2
....--_(n_-_1_)2
t 2 (El
an
- <:p =(cot rn"' -f-- n tan cp)2 - (cotcp-ntancp) 2+ 4n :::::::, 4n
11. We have
1-tanz '\'
cos 2 Y= 1 + tan2 '\'
To prove that cos 2y ~ 0, it is sufficient to prove that
1 - tan 2 y ~ 0.
But we have
l-tan 2 = cos2acos2~-(1-t-sinasin~)2.
Y
cos2 a cos2 ~
We only have to prove that
cos 2 a cos 2 ~ - (1 + sin a sin ~}2 ~ 0.
But
cos 2 a cos 2 ~ - (1 + sin a sin ~) 2 =
= (1 - sin 2 a) (1 - sin 2 ~) - (1 + sin a sin ~) 2 =
= - (sin a + sin ~) 2 ~ 0.
401
Solutions to Sec. 8
12. Let m be the least and M the greatest of the given

fractions.
Then
m~ :'. ~M (i= 1, 2, 3, ... , n).
I
Hence
mb;
a 1 ~ Mb;.
= 1 to
Summing all these inequalities (from

we find
n),
And so indeed
~at
m~~~M.
..! bt
13. We assume, of course, that all the quantities a, b, . .. ,

l are positive, and the principal value of the root is
taken everywhere. Besides, m, n, ... , p are positive integers. Let us take logarithms of our roots, i.e. consider the
quantities
log a
m-
.. ,
log l
p
Let be the least and M the greatest of these fractions.

On the basis of the results of Problem 12 we have
<
loga+logb+ ... +logl
m+n+ ... +p
<M.
Consequently
< logm+n+. +P Vab ... l
wherefrom follows our proposition.
14. See Problem 12.
15. We have
x1-. - y1-. - z1'. = y2 (x1-.-2 - y1-.-2)
since
x2
= y2
< M,
+ z2 (x1-.-2 -
z1-.-2),
+ z2.
From the same equality follow x > y, x

A - 2 > 0,
>
z. Therefore, if
402
Solutions
then
XA-2 -
>
yA-2
0 and
> 2,
zA > 0,
i.e.
>
zA-2
xA-2 -
0,
and, consequently, for J.

XA -
YA -
xA >YA
We prove in the same way that

XA < YA + zA if J.
<
+ zA.
2.
16. (See Problem 7, Sec. 1). It can be proved, for instance,

in the following manner. If a2 + b2 = 1, then, obviously,
We"can find an angle cp such that
a =cos cp, b =sin cp.
Likewise we can find an angle cp' such that
m = cos cp', n = sin cp'.
Then we have
I am
+ bn I =
I cos cp cos cp'
+ sin cp sin cp' I
= l cos (cp - cp') I ~ 1.
17. We have
a2
b2
c2
a2 - (b - c)2,
~ b2 - (c - a) 2 ,
~ c2 - (a - b) 2
~
Multiplying, we get
a2 b2 c2
~(a+
b -'c) 2 (a+ c -
b) 2 (b
+c-
a) 2
Hence follows the required inequality.

18. It is known that if A + B + C = n, then
A
tan 2 tan 2
+ tan 2A
tan 2 +tan 2 tan 2
(see Problem 40, 4, Sec. 2).

Put
A
tan 2 = x,
tan 2
y,,

x2+Y2+z2~1
tan 2 = z.
403
Solutions to Sec. 8
if
xy
xz
yz = 1.
But we have
2 (x 2 y2 z2) - 2 (xy
xz
yz) =
(x - z) 2
= .(x - y) 2
Hence
y2
z2) - 2 ~ 0,
2 (x 2
x2
y2
z2 ~ 1.
19. We have
+ +
+
+
+ (y -
z) 2 ~ 0.
+ +
+ +
-. /(p-b){p-c)
sm2=v
be
.
sm2=
sm2=
'
-. f(p-a) (p-c)
V
ac
-. f(p-a) (p-b)
ab
Consequently, it is sufficient to prove that

(p-a) (p-b) (p-c) ~_!_
abc
"""'8
But
p-a=
a+b+c
b+c-a
a+c-b
2
-a= : 2
'p-b=
2
'
p-C=
a+b-c
2
Therefore, we have to prove only the following

(b+c-a) (a+c-b) (a+b-c) ~ 1
abc
"""' '
provided b+c-q,>0, a+c-b>O and a+b-c>O (see

Problem 17). This inequality can be proved in a different
way. Put
.
sm 2 sm 2 sm 2 = .,,;
then we have
= 21
A+B)
A+H
cos -A-B
2 - - cos - 2 - cos - 2 - .
Hence
A+B
cos 2 - 2-
-+ 2 = 0.
A-B
A+B
- cos-2cos - 2
404
Solutions
Consequently
A-B
A+ B
cos-2- =
.
cos -_-2-
-./
JI
A-B
2- -86
cos2 -
A+B
A-B
Smee cos -2- and cos - 2-
are real, there must be
gt 0
cos 2 -A-B
2 - - .,,;;;::: '
gt.,,~cos 2 -A-B
-,
2
gt.,,~,
1 t.,,~
1 .
20. 1 We have the relationship (see Problem 40, 2,

Sec. 2)
cos A+ cos B
+ cos C = 1 + 4 sin ~
sin
sin
Using the result of the preceding problem, we get the

required inequality.
2 Since there exists the following relationship
cos
cos
cos
~ =
! (sin A+ sin B +sin C),
the given problem represents a particular case of Problem 4g

of this section.
21. It is sufficient to prove that
(a+ c) (b +d);;;=::ab + cd + 2 V abed,
i.e. that
cb+ ad~2 Vcbad.
But
cb+ad-2 V cbad= (V cb-V ad) 2 ,?!0.
22. We have
a2 + b2 -2ab = (a-b)2 ~O.
Hence
a 2 -ab+b 2 ~ab,
a3 + b3~ab (a+ b).
Consequently
3a 3 + 3b 3;;;=:: :Ja2 b
+ 3ab
Add a3 + b3 to both members of the last inequality.
Solutions to Sec. 8
We have
405
4a3 + 4b 3 ~(a+ b)3.
And so, indeed,

a3+b3
;;;:.
( a+b
)3 .
23. 1 It is required to prove that the arithmetic mean

of two positive numbers is not less than their geometric
mean. Indeed,
-a+b
2- -
v-ab=z
v-)
1 (Vab ="2
a- v-)2
b ~O.
1 ( a+b-2
2 To prove that
a-+ b -Vb~_!_
(a-b)Z
b
a --...:::: 8
(a> b),
it is sufficient to prove that
CVii- -VW
2
1 (a-b)2
:::;;;8
Consequently, it is necessary to prove the following
(-Vii+ -Vii) 2
Bb
~_!_
::::--- 2 .
We have
CVii+-Vii) 2
8b
since b
> 1.
=.!(
1 +-./a)z>-_!_
8
vb :0---2
The second inequality is proved in a similar way.

24. Put a = x 3 , b = y3 , c = z3 The only thing to be
proved is that
x 3 + y3 + z3 - 3xyz ;;;:. 0
for any non-negative x, y and z.
But we have (see Problem 20, Sec. 1)
x 1 + y3 + z3 - 3xyz = (x + y + z) X
x (x 2
+y +z
2
2 -
xy -
And so, it only remains to prove that

x2 + y 2 + z 2 xy - xz - yz ;;;:, 0.
xz -
yz).
Solutia ns
406
But we have (see Problem 10, Sec. 5)
+ 2y + 2z
2x2
2xy - 2xz - 2yz =
2 -
(x -
y) 2
25. We have
+ (x -
z) 2
+ (y
z) 2
;;;:::
0.
a1+a2
v a1a2~2 -,
i/-
Adding them termwise, we get the required inequality.

26. We have
1-t-a 1
1,11+a2 --...]/- 2 - > r alt
- a2, ... , -1+an,
-d:3 an.
9
v-
Multiplying these inequalities term by term, we have

(1.+a1) (1+a2) ... u+an)1
2n
d:3 a1a2 ... an = .
And so, indeed,
(1
at) (1
a2) . . . (1
an) ;;;::: 2n.
27. 1 Make use of the following identity
(a
+ b) (a + e) (b + e)
=
=
(ab
But
a+b+c >
3
Therefore
+ ae + be) (a + b + e ) -abe.
/-b
:::--V a e,
+ +
+ +
(a
b
e) (ab
ae
be) ;;;::: 9abe,
and consequently
(a
b) (a
e) (b :-+- e) ;;;::: 8abe.
2 We have
_a_+_b_+_c_= a+b+c _ 1 b+a+c _ 1 +
b+c
a+c
a+b
b+c
a+c
c+a+b
(. 1
1
1 )
.
1 =(a+b+e) b+c+a+c+a+b - 3
a+b
But
(b+e) +(a+ e) + (e+ b) ~3 y/(b+ e) (a+e) (a+b),

i.e.
Solutions to Sec. 8
407
Further
1
b+c
+ a+c + a+b =(b+c)(a+c)(a+b){(b+c)(a+c)+

+ (b+ c) (a+b) +(a+ b) (a+ c)}~
~ (b+c) (a!c) (a+b) :/(b + c)2 (a+ c)2 (a+ b)2.
Therefore
b;c
+ a~c + a~b ~ ~ y/(b+c) (a+c) (a+b) X

X (b+c) (a!c) (a+b) y/(b + c)2 (a+ c)2 (a+ b)2-3.
Thus
:>-:3
b+c
a+c
a+b ==--2
28. It is sufficient to prove that
(a+k) (b+ l) (c+ m)~(y/ abc+y/klm) 3
We have
(a+ k) (b + l) (c m) =
= abc+klm+ (ale t-kbc + abm) + (klc+ alm+ kbm),
(y/ abc+y/klm) 3 =abc+klm+
+ 3 y/ a b c klm + 3 y/k l m abc.

3 2 2
But
alc+kbc+abm...._ a/ 2b2 2 kl
"""'v a c m,
3
2 2
klc+alm+kbm:>-: 3 /k 2 z2
3
,,,..-v
Hence follows the validity of our inequality.

29. We have
1
a+b"+c-~ 3
v1
But
i.e.
1
a: b"c-= Vabc
-~ a+b+ c
Vabc
b
m2 a c.
408
Solutions
Therefore
_!.+_!.+_!.~3_1_>
a
Viibc
c ::;;.-
r- a-j-b-j-c
30. It is necessary to prove that the arithmetic mean

of n positive numbers is not less (>) than the geometric
mean of these numbers. We are going through several proofs
of this proposition. Let us begin with the most elegant
one which belongs to Cauchy.
Thus, we have to prove that
x1-J-x2-J- ... -J-xn--.._ n/
n
-?V X1X2
Xn
At n = 1 the validity of this inequality is obvious. At

n = 2 and n = 3 the proposition was proved in Problems 23
and 24.
Let us first show how to prove the validity of our assertion
at n = 4. We have
But
Let us now prove that, in general, if the theorem holds

at n=m, then it is valid at n=2m too.
Indeed,
X1 +x2+xa+ .. +x2m-I +x2m _
2m
x1+x2
+ X3-j-X4
+ . . + X2m-1-J-x2m
2
2
m
--..._ vx1+x2 .
-?'
X3-j-X4
...
X2m-t-J-X2m
(since we assume that the theorem is valid at n = m).
Solutions to Sec. 8
409
Further
Xt
+ + + +
X2
X3
X2rn '-.
2m
::::"'
And so, assuming that the theorem is valid at n = m,

we have proved that it is true at n = 2m as well. And
since we proved the validity of the theorem for n = 2,
it is valid for n = 4, 8, 16, ... , i.e. for n equal to any
power of two. However, we have to prove that the theorem
is true for any whole n. Let us take some value of n. If n
is a power of two, then for such a value of n the theorem
is valid, if not, then it is always possible to add a certain
q to n such that n
q will yield some power of two.
Put
We then have
x1 +x 2 +x3 + ... +xn+Xn+1+ ... +xn+q >n+q
:::cn+q/~~~~~~~~~-
V X1X2 .. XnXn+! ... Xn+q
:;?
for any positive

Put
Xn+t
Xi
Xn+2
(i=1, 2, ... , n+q).
= =
Xn+q
x1+x2+ ... +xn

n
We get
x 1+x2+ ... +xn+
x1+x2++xn
n
q
~~~~~~~~~~~~~~~>
n+q
or
410
Solutions
and finally
Xt
+x2+n .. +xn,,;:. n/
V X1X2
Xn.
And so, the theorem is valid for any whole n. It is obvious

that if x 1 = x 2 = . . . = xn, then the sign of equality
takes place in our theorem. Let us prove that the sign of
equality occurs only when all the quantities x 1 , x 2 , , Xn
are equal to one another. Suppose at least two of them,
say x 1 and x 2 , are not equal to each other. Let us prove that
in this case only the sign of inequality is possible, i.e. it
will be
Indeed
v(x11
X2 ) 2 X3 Xn
But if x 1 is not equal to x2 , then

x1+x2
> lv-
X1X21
consequently
and therefore
x1+x2+ +xn
> n/
y X1Xz
Xn
if at least two of the quantities x 1 , x 2 , , Xn are not

equal to one another.
Given befow are some more proofs of this theorem. Let
us pass over to the second one. Let n be a positive number
greater than or equal to unity (n ~ 1). We assume here
that a and bare two real positive numbers. Then the following inequality takes place
(an-t_bn-1) (a-b)"~O.
411
Solutions to Sec. 8
Hence
an+ bn >an-lb+ bn-la.
Consider n positive numbers a, b, e, ... , k, l. Let us apply

this inequality to all possible pairs of numbers made up of
the given n numbers. Adding the inequalities thus obtained,
we find
(a11+bn)+(an+en)+ ... +(an+zn)+
+ W' +en)+ ... + (bn + zn) + ... +(kn+ zn);;;::::

~ (an- 1b + bn- 1a) + (a 1e+ en- 1a) + ... +
+ (an-iz + zn-Ia) + ... + (kn-iz + l"- 1k).
11 -
Hence we 'have
+l
(n-1)(an+bn+ ...
;;;::.a (bn-1 en-1 + ...
11 );;;:::
+ zn-1) + b (an-1 + en-1 + ... + zn-1) +

... + zn-1) + ... +
+ l (an-1 + bn-1 + ... + kn-1). (*)
e (an-1 + bn-1 +
Using this inequality, it is possible to prove our theorem

on the relation between the arithmetic and geometric means
of n numbers by the method of induction. We have to prove
that
Put
... ,
Xn-1
= k 11 ,
Then it is sufficient to prove that

an+bn+ ... +kn+zn:>-, b
n
,,__a . . . kl
Let us assume that this inequality is valid at the exponent

equal to n - 1, i.e.,
bn-1
+k11- 1+zn-1;;;::.(n-1)bk ... l,
an-1+en-1+ ... +zn-1~(n-1)ae ... l,
+ ...
Solutions
412
Using the inequality (*), we find

(n - 1) (an + bn + ... + kn + ln) ~
~a(n--1)bk
+
+ l (n -1) ab ... k.
... l+b(n-1)ac ... l+ ...
Hence
(n-1)(a"+b"+ ... +k"+l")~(n-1)nabc ... kl,
i.e.
an+bn+
..._
-f--l">___
n_
_ =--abc ... ll
i .
Thus, our theorem is proved for the second time. Let us
pass over to the third proof of this theorem. It will be carried
out using the method of mathematical induction once again.
Let there be n positive numbers a, b, ... , k, l. It is required
to prove that
a+b+ ... +k+l~n;/ab ... kl.
Assuming that the theorem holds true for n - 1 numbers,

we have
a+b+ ... +k+l~(n-1)"-;/ab ... k+l.
And so, the theorem will be proved if we prove the

inequality
(n--1)
n-yab ... k+ l~nr;/ab ... kl.
Thus, we have to prove the inequality

(n -1)n-v ab ...
Zn
'-+ 1 ~n
kl
ab ... kl
/n
.
Put
ab ... kl
zn
tnln-ll
"'
Therefore, it is required to prove that

(n-1) " 1~nn-i.
And so, to prove our theorem means to prove the inequality
413
Solutions to Sec. 8
where 6 is any raal positive number and n is a positive

integer. Let us prove this inequality. At 6 = 1 we obviously
have the equality. Suppose now 6 > 1. It is required to
prove that
We have
But
Therefore
Sn-1 +6n-2+ . + + 1 < nn-1,

and, consequently, indeed
tn-1
"'~-1
If
6 < 1,
< nn-1.
we have to prove that
sn-1>
6-1
n..,
tn-1
This result is obtained as in the previous case, and, thus,

the theorem is proved.
All the considered proofs were carried out using the
method of mathematical induction. Therefore, it is desirable to get such a proof which would establish immediately
that
if a 1 , a 2 , , an are any positive quantities not equal to

one another simultaneously. Put ai = x?. Then we have
to prove that
xi'
+ x~ +n .. + x~ -X1X2
Xn
>O
i.e. the problem is reduced to finding out that a certain

function (form) of n variables x 1 , x 2 , , Xn is positive.
As is known, n letters X 1, x 2 , , Xn can be permutated
4f4
Solutions
by nl methods. If f (x 1 , x 2 , , xn) is a function of n

variables x;, x 2 , , Xn, then the symbol 'Lif(x 1, x 2, ... , xn)
will denote the sum of n! quantities obtained from
f (x 1 , x 2 , , xn), using, all possible permutations. For
example,
~ X1X2 ... Xn = n! X1X2
~ x~ = (n-1)! (x~ + x~ +
Xn,
... +x~).
Introduce the notation
It is easily seen, that whatever permutation is used, the

function cp (xi. x 2 , , Xn) remains unchanged. Therefore
we have
n! cp (xi. x 2 ,
Xn) =
! ~ (x~+x~+ .. +x~)- ~ x1x
But
~ x~
+ x~ + ... + x~ =
n! (x~ + x~
+ ... + x~).
On the other hand,

X1n+ X2n+
+ Xnn =
1
(n-1)!
""'Xn1
L.J
therefore
Let us consider the following functions

cp1_: ~ (x~- 1 - x~- 1 ) (x1 -x2),
cp 2 = ~ (x~- 2 -x~- 2 ) (x 1-x2)
cp3 = ~ (x~- 3 -
x~- 3 )
(x1 -x2)
X3,
X3X,,
..................
Xn
415
Solutions to Sec. 8
Wo have
<fln-1 = 2 ~ X~X2X3 Xn --2 ~ X1X2X3

Adding these expressions termwise, we find
+ + + ... +
Xn.
<J>1 <p2 <p3

<J>n-1 = 2 S X~ - 2 S X1X2
Comparing this with the equality ( *), we get
n! n-1).
And so
xr
+ x~ +n ... + x~
. -
-X1X2 Xn-
1 (
Zn! <p1
+ <J>2 +- + <J>n-1 )
But it is evident that n- 1 vanish if and only

if X1 = X2 = . . . = Xn.
If not all of the variables are simultaneously equal to
one another, then all <pi > 0. Indeed, we have
<J>1 = S (x1 -x2)2 (x~-i
x~- 2 ) ~O,
<p2 =
+ ... +
S (x1 -x2) 2 (x~- 3 + ... + x~- 3 ) x3 ~0,
<J>n-1= S(x1-X2)2 X3X11 ... Xn~O.

Therefore
the equality being possible only if x 1 = x 2 = . . . = Xn.

And so, the theorem is proved. This proof belongs to A. Gurwitz.
31. We have (using the preceding problem)
n/
---a1 +a2+n ... +an = a1+an

2n
-V a1a2 ... an"""'
a1 +an
2
To prove the second inequality consider the product

(a1a2 an) 2 = (a1an) (a2an-1) ... (ana1).
416
Solutions
But we can prove that
akan-k+t
a1an
(see Problem 19, Sec. 7)
Therefore
and
~/ a1a2 ... a.~V a,an.

32. Consider a quantities equal to
!,
to
and c quantities equal to
of these quantities will be
1
1
a'
b quantities equal
The arithmetic mean
a-a+bb+c7
a+b+c
The geometric mean is equal to
a+b+v_1_ . _1_. __!__

aa
bb
cc
Consequently
i.e.
33. Put
a =-;;.,
a
b=l._
m '
C=:s._,
m
where a, ~. y and m are positive integers.

Consider the product
(1+ b--;;c r(1+ c--;:a )b(1+~r=

=
V(1+ b--:c r(1+ c--;:a )t\(1+ a--:b r.
Since a, ~ and y are whole positive integers, the radicand

may be considered as a product of a factors equal to
h-c
l
c-a
1 +-a- each, ~ factors equa to 1 +-b- each, and y
Solutions to Sec. 8
factors equal to 1
a-b
+-c
each. Then we have
aH+} F(i + b :- c r (i + c ~a rl (i + a-:- b r ~

1
_ a(t+b-:-c
)+~(i+c~a
a+~+Y
)+y(t+a-:-b)
~
= 1.
Raising both members of this inequality to the power
a+ b + c, we get the required result.
34. We have
V'
sn
s
(s-a) (s--b) ... (s-l) = }I (s-a)(s-b) ... (s-1)
But
vt(s-a) (s-b) ... (s-l)~

~
(s-a)+(s--b)+ ... +(s-l)

n-1
=--S.
n
n
Therefore
1
::::;::
n
l/(s-a) (s-b) ... (s-l) ~ (n-1) s
The further proof is obvious.

35. First of all this inequality can be obtained from Lagrange's identity (see Problem 5, Sec. 1). But we shall proceed in a somewhat different way. Let us set up the following
expression
(lva1 + b1) 2 + (lva2 + b2) 2 + ... + (Ivan + bn) 2 =
= Alv 2
2Blv
C 2 ,
where
A=a~+a;+ ... +a~,
C=b~+b;-t-- ... +b~,
B = a1b1 + a2 b2 + ... + anbn.
Since the left member of this inequality represents the sum
of squares, we have
Alv2
2Blv
C 2 ~ 0.
for all values of Iv and .
118
Solutions
Consequently, the trinomial

Ax2
+ 2Bx + C
is greater than or equal to zero for all real values of x. Therefore, the roots of this trinomial are either real and equal
or imaginary, and its discriminant is less than or equal
to zero, i.e.
B2
-AC~
0.
Thus
(a1b1
+ a2b2 + ... +
anbn) 2 --(ai+a;+.:. +a~) (bi+ ...
+ b~)~O,
wherefrom also follows that the equality sign is possible

only if
a1
a2
an
b;=b;= ... =T,;
36. Put b1 = b 2 = ... = bn = 1 in the inequality of

the preceding problem. We then have
(a 1+a2+ ... +an) 2 ~n(a~+a;+ ... +a~).
Hence
a 1+ llz +
---v, n (2a + a 2 + ... +an,2\

... +an:::::::::
1
37. The result is obtained from the formula of Problem 35

if we put
... , a~ =Xn,
a; =X2,
a~=x 1 ,
b:=-1-'
b~ =-1-'
X2
Xt
... ,
b~= _1__
Xn
But we may also use the theorem on the arithmetic mean.

Then we have
,
x 1 + x2 +
... +xn~n;/x1X2 ... Xn,

_1_+_1_+ ... +-1 ~n v-1___1_ ... __!_.
Xt
Xn
X2
Xt
x2
Xn
Multiplying these inequalities, we get the required result.

38. Let us first prove that
2
2n
......._O
p -
n-1 q:;:::;;
Solutions to Sec. 8
We have
q
0 :=:;;; (x1 -
X1X2 + X1X3 + ...

X2) 2 +-(x1 - X3) 2
419
+ Xn-IXn,
+ ... + (xn-1
- xn) 2.
Consequently
(n-1)(x~+x~+ . .. +x~)--2q~O.
But
x~
+ x; + ... + x~ =
p 2- 2q,
wherefrom we get
Consider now n - 1 quantities (instead of n): x 1 , x 2 , ,

x;_ 1, xi+ 1, ... , xn, eliminating X; from the quantities
under consideration, and put
p -
q - (X;X1
X;
p',
+ X;X2 + ... + X;Xi -I + X;Xi+t + ... +
+ X;Xn)
= q'.
Using the deduced inequality, we may assert that

,2
p -
2(n-1)
q'~O.
n-2
But
q'
q -xi (x 1 +
+ ... +
X2
Xi-1
+ ...
xi+t
=
q-
X;
:rherefore
(
p-x; ) 2 -
2(n-1)(
n-2
q-px;+xi2 ) ~ O.
Consequently
nxt-2px;
+ 2 (n-1) q-(n-2) p 2:=:;;;0.
Consider the trinomial of the second degree

nx 2 - 2px
+ 2 (n -
and denote its roots by a and
1) q ~
(n - 2) p 2
+xn) =
(p - X;).
420
Solutions
~he
Solving
quadratic equation, we find
a=_!!_- n-1
n
n
p
n-1
~=-+-n
n
-./p2-~q
n-1
'
2n
p - -n-1
-q '
(~>a).
We then have an identity
+ 2 (n -
2) p 2
= n (x 1 - a) (x; - ~)
wherefrom follows that xi lies between a and ~' i.e.
nxr - 2px;
1) q -
(n -
a< X; <
0,
39. Let a and b be two real positive numbers. If p > 0,

then aP - bP > 0 for a > b; and if p < 0, then aP - bP <
< 0 for a > b. Therefore we may assert the following:
(aP - bP) (aq - bq)
0 if p and q are of the same sign;
(aP - bP) (aq - bq) ~ 0 if p and q are of different signs
and for any real a and b. Let us first consider the case when
p and q are of the same sign. We have
ap+q +bp+q>apbq +aqbp,
ap+q + cP+q ;?:aPcq +aqcp,
>
ap+q + zp+q ;?:apzq +aqz'

bP+q + cv+q > bPcq + bqcv,
Adding these inequalities term wise, we get

(n-1) (ap+q + IJP+q + ... + [P+rir:~ 2:.. apbq,
where a and b (in the last sum) attain all the values from
the series a, b, c, . .. , l. Adding 2:av+ 7 to both members
of this inequality, we get
ll (ap+q + bp+q -f- ... + lP+'l);?:(aP + bP + ... +
+ZP) (aq + bq + ... + lq).

The second inequality is obtained just in the same way.
From these inequalities wp, can easily get the results of Problems 36 and 37.
421
Solutions to Sec. 8
40. 1 Let /,= !!!.._, m

n
V( 1 +a:) (
<
1 +a
> n.
We have
: ) ... ( 1+a:)11
... 1
(1+a: )+(1+a: )++(1+a:
<
)+m-n
(the factor 1 +a :
of the radicand is taken n times.
the factor 1 is taken m - n times) . Hence

n
or
m
(1+a)n>1+a_'.'!:_.
n
2 Put
'A= -mn
and first assume that m
> n,
'A> 1.
i.e.
We have
V(
1 - a ';'. ) ( 1 - a ': ) . . . ( 1 -a : ) 1 1 ... 1
<
( 1-a : ) n+m-n
<-------m
The factor 1- a!!!...
of the radicand is taken n times, and
n
the factor 1 is taken m - n times. Hence
n
(
ni ) m
1-an
1
<1-a<--
1 +a '
1-a.'.'..:<--n
m
(1+a) n
<--1-a.!'.:
n
m < n. We have
(1+afn
Let us assume now that
y/(1 +a)m
=7(1 +a) (1 +a) ... (1+a)11 ... 1 <

= 1 +am < ___
< (1 +a) m-t-n-m
n
n
am
1
-
422
Solutions
And so, in this case also

m
(1+a)11<-1- am.
n
am.
<
Remember that we assumed - n

1.
41. 1 Put in inequality 1 of the preceding problem
1
n+1
a= n+ 1 , A=-n-. We get
n+I
Hence
1 )n+I > (1 +-n1)n '
( 1 +~n+1
i.e. Un+I >Un.

Here is one more proof. Without using the theorem on
the arithmetic mean, let us prove that
( 1 +-a-)n+I
.n+1
> ( i --1--!!:...)n
' n
if a> 0 and n is a positive integer.

Consider the identity
1+nx
1+(n-i)x
1+3x. 1+2x 1+x
1
+nx= 1+(n-1)x 1+(n-2)x 1+2x 1+x - 1 (x
But
> 0).
1+(k+1)x _ 1+-x-> 1 -+
1 + kx
-1 + kx
1 + nx
1+(n+1)x
1+nx
(k=O, 1, 2, ... ,n-1).

horefore
1+llX>r 1 ~ ~~Xf)
r,
(1 + llX)lHl > [1+(n+1)
Putting here x = n (n + 1) , we get
_a )n+I > (1 + !!:.n . )n
( 1 + n+1
xin.
423
Solutions to Sec. 8
In particular, at a= 1, we find
(1 +
~1
r+l > ( + ~ f.
1
2 We have
u.~(11 ~r~[(i+~)'J'<(-~--rr~ (~})'

Hence
1
Un<
1 h
(1-k)
for any whole positive k.

If k = 6, we find
(1 +
42. We have
r < ( ~ )6 <
3.
n(n+ll
_ n(n+1i /
1 +--;:;-
)n
n:
<
/3
V -n
n(n+1~
(see Problem 41).

But the fraction
~~1
n ""'=
if
n>-3.
-
Therefore
n+lr:::-77
-~Y-nn_1+_1,
vn
<1
if
n~3.
<1
(n = 2, 3, 4, ... ),
n(n-v (
1+-1n )n-1 -n1 --
'Yn+Tl+
i
n-}fn"
We have
n(n-1}1 (n+1)n-1
nn
n(n-v (
1 )n
1+n
1
n(n-1~3
--<
--<1.
n+1
n+1
""'=
424
Solutions
44. Let us prove that

logy;?;: llit log X1 + a; 2 log x 2 + ...
(i= 1, 2, ... , n).
To this end it suffices to prove that
+ a;n log Xn
log(a:c+by+cz+ ... +Zu)::;?-alogx-j--blogy+ ...
+llogu.
(*)
if a+ b + ... + l = 1 and a. b, ... , l are rational positive
numbers.
Put
Then
a+~+
... +'A=N.
To prove the inequality (*), it is l'ufficient to prove that
ax-t by+ cz +
... + lu ??- xal
... u 1
Bul we have
xayb ... ul
i/' xrxyf'> ... u'- =

= ~
x ... xy ... y ... u ... u::::;;;

___ ax+~Y+
... +A.u
N
::::::::
ax + by + . . . + l u.
Thus, it is proved that

logy;~ a; 1 log
x 1 + a; 2 log x2 + ...
(i= 1, 2, ... , n).
+ a;n log Xn
Hence
n
~ logy; ~(log x1)

i=i
L;
ail+ (log X2)
i=1
h a;di=1
... +
n
+(log Xn)
:2;
ll;n
i=1
or
n
~ logy;~logx 1 +logx 2 +
1=1
... +logxn=logx 1x2 Xn.
425
Solutions to Sec. 8
Finally
Y1Y2 Yn~X1X2 Xn.
45. Put .!!..!._=xi (i = 1, 2, ... , n). Then we have to prove

ai
the inequality
;.Y(1 +x1) (1 +x2) ... (1 +xn)~1 +y/x1X2
. Xn.
The theorem is valid at n = 1, 2, 3 (see Problems 21 and

28). Suppose it is true at n = m and let us prove that
it also holds at n = 2m.
We have
2 yi(1
+x1) (t +x 2) ... (1 +x2m-1) (1 +x2m) =
;vV (1 + X1) (1 + X2) v(1 + X3) (1 + X4)

... ;Y'V(1 +x2m-1) (1+x2m)~
~v (1 + ~) (1 + ~) ... (1 +
>1 + V ~ V X3X4
V x2m-1X2m) ~
V X2m-1X2m
= 1 + "{Y
2
=
X1X2 X2m
Thus, the theorem is valid for all indices equal to any

power of two. Let us now prove that it is true for any
whole n. Let n+q=2m. Then
n+y(1 +x1) (1 + X2) ... (1 +
Xn)
(1
+ Y1) (1 +
~ 1 + n-t-yX1X2
Y2) ... (1 +
Yq);;;::
... :CnY1Y2 ...
yq
Put
1+Y1=f+Y2= ... =1+yq=
=7(1+x1) (1+x2) ... (1+xn) = Y.
We have
n+JV(1+x 1)(1+x2) ... (1+xn)Yq ;;:::
~1
+n+Yx1X2 ... Xn(Y-1)q
But
(1 +x1) (1 + x2 )
(1 +xn)
= yn
426
Solutions
Therefore
n+yyriyq ;;,.1+n+yX1
. Xn
(Y -1)q,
i.e.
or
Hence
(Y --1)n>x1X2 ...
Xn,
Y-1~;Y"x1X2 ...
Xn.
Finally
Y =;Y"(1+x1) (1+x2) ... (1+xn)~1 +;Y" X1X2
Xn,
and the theorem is proved.

The equality sign is possible only if x 1 = x 2 = ... =
= Xn = f.
46. This theorem, as the previous one, is proved using
Cauchy's method. The proposition is valid at n = 1; let
us first prove that it holds true at n = 2, i.e. prove that
for any whole positive k. At k = 1 the last inequality really

takes place. Assuming the validity of this inequality at
k = l, let us prove its validity at k = l
1. And so, we
have (by supposition)
. (Xt
+ X2)I ::::::::: xr + X~
21
-.;::::
Multiplying both members of this inequality by

we find
xi
xi+ 1 +x~+i +x1x~+x2x~

4
But
x2
Solutions to Sec. 8
427
since
( x1 +x 2 )1+1 ~ x~+I +x~+I

2.
"""'
2
and the inequality () is proved for any whole k. And so,

our basic proposition is valid at n = 2. Let us now prove
that if it is true at n = m, then it is also true at n = 2m.
Indeed
( x1+x2+x3+x41~. +x2m-1+x2m
=
-
( Xj
+ x3+x4
+ . .. + X2m-1+x2m
2
2
x1+x2
2
~ X2
)Ii=
)h
+ ( X3 -~ X4
+ ... + ( X2m-!/ X2m
m
h
X1,+x2
2
+ X3h -t X4h
+h
_1_
x2m-1 - x2m
2
2
~---------------m
h
l"+
h
h
h
+ +x2m-I
+x2m
X1 +x2+x3+x4
2m
Thus, we have established that the theorem is valid

at n equal to some power of two. It remains to prove its
validity for any whole n. Put n
p = 2m.
Then
Put
Y1~~y2-=
x1 -l x2+ ... +.rn

=yp=----n---
We have
X1
+ + + + Y1 + Y2 + +
Xz
Xn
YP
=
(x1
+. . +xn) (n+ p)
n
428
Solutions
Hence
h
( x1
( x1+x2+ .. +xn )hp
X1 + +xn+
11
+ .n. .+.'Cn )Ii :::::;;-------~-------n+p
Finally
and the proposition is completely provPd. It is easy to

establish that the equality sign is possible only if
47. This proposition is the generalization of the previous

theorems (see Problems 30, 45, 46). The proof is carried
out in the same way as in the mentioned theorems. Namely,
assuming the validity of the theorem at n = m, let us
prove its validity at n = 2m. We have
<jJ (t1)
+ <jJ (t2) + 4-<JJ (t2m-1) + <jJ (t2m)

2m
(since, by hypothesis, not all of the quantities t 11 t 2 , , t 2 m

are equal to one another, they can be grouped so that, for
instance, t 1 =I= t 2). Thus, the theorem is valid at n = 2m.
Let us put now n
p = 2m. Then
m ( t1+t2+ +tn+ -r1+-r2+ .. +-rp )

't'
n+p
<
<
<p(t1)+ ... +<JJ{tn)+<JJ('t1)-t +<JJ('tp)
n+p
Solutions ta Sec. 8
(here t 1 , t 2 ,
.. ,
429
ln are not all equal to one another).
Put
td t2+ ... + tn
T1=T2= ... =Tp=--""----n---"--
1+2
+ +p=
t1 --1 t2
+n... + tn
p.
Consequently
qi
)=m(
( t1+t2+ ... +tn-i<1-1 -t-'p

n-1-p
On the other hand,

cp (ti)+ +Cf! (tnH- rp (i-i) + +Cf! (i- p)
Cf! (t1 )
-t- Cf! (tn) +pep
tj
+ n+ tn
n -+- p
From the last inequality we get

qi
Cf! Un) .
( ti-j n-'- tn ) <Cf! (ti)+-+
n
The above-deduced theorems (see Problems 30, 45, 46) are

obtained, as we already mentioned, from this more general
proposition. Let us demonstrate this.
1 Let
cp(t)= -log(1 +t),
then
Further
cp (t1H- cp (t2)
2
log (1 +ti)+ log (1 + t 2)

2
But
V(1+t1)(1+t2)< 1+t1t1+t2 =1+ t1-tt2
Therefore
(t1=;i:t2).
430
Solutions
(the base Of the logarithms being greater than nnity) and

- log
-V (1 +
f1) ( 1 + t 2)
>-
tit
log ( 1 +
t2 ) .
Thus, the function
cp (t) = - log (1 t)
really possesses the following property
ti+ t2 )
cp (
2
and therefore it must be
cp ( t1+t2+~ .. +tn )
<
<
IP (t1) +IP (t2)

2
'
1P(t1)+1P(t2): ... +IP(tn)
i.e.
-log ( f +
t1+t2+~ .. +tn ) <
<-
log(i+t1)+log(i+t2)+ ... +log(i+tn)
ra
logy/(1+t 1)(1+t;) ... (1+tn)
'
<
t1 +. ~ + tn
<log ( 1 +
).
Further
y/ (1 + t1) (1 -H2)
+ tn) <
< i + td- ..n + tn
... (1
(1+t1)+(i+t2)+ ... +<1+tn)

n
Putting 1 + t; =
n/
Xj,
we finally get
X1X2 Xn
< x1+x2+ ... +xn

n
Obviously, if we assume the possibility x 1 = x 2 = ... =

then it will be
2 If we put
cp (t) = t"'
Xn,
Solutions to
Sec. 8
= (
t11t2
431
then
<p ( t1tt2 )
Assuming that the inequality

( t1tt2
r<
ti
~t2
holds true, we get the result of Problem 46.

3 Put

cp
(ti)~cp (t 2) =log V (1 + e11) (1 + e 12).
Since
11+12
V'(1+eti)(1+et2)> 1+e-2-,
fulfilled for the function <p (t) is the inequality
cp(
t1~t2 )<cp(t1)~cp(t2)
(t1=i=t2)
Therefore
cp(
t1+t2+~ .. +tn
)<
cp(t1)+~+CJl(tn)
i.e.
et= A.,
t =loge A.
Then
7(1
+ e11) ... (1+etn)

=
;Y (1 +
=
A.1) ( 1 + A. 2) .. ( 1 + A.n) >
log A.1+ ... +Jog "-n
>
1+e
S olutioris
432
Finally
;Y(1-t-A.1)(1+A.2) . (1+1.n)> 1+;Y'l.1A2 An.
48. Let ti! t2 ,
ween 0 and n.
be contained in the interval bet-
tn
(0< ti< n).

Let us prove that
<-
-sin t1+t2+ ... +tn
sint1+sint2+ ... +sintn
For this purpose it suffices to prove that (see Problem 47)

-sin t 1+t2
2
<-
sint1+sint 2
Indeed
sin t1+sin t2
t1+t2
=sin
2
2
-sin t1+t2 cos t1-t2

2
=sin t1+t2 . 2 sin2 t1-t2

2
(in our case

Thus
qi
(t) = - sin t).
(if 0 <ti< n).

Therefore if a1 + a2 + ... +an = n, then
sin a 1 +sin a2
+ ... +sin an< n sin_::_

n
if a1 , a 2 , , an are not equal to one another.

On the other hand, if
ll1=ll2= ...
then the sum
sin a1
becomes equal to
Jt
=lln=
7,
+ ... + sin an
.
Jt
nsm-.
n
>0
Solutions to Sec. 8
433
Thus, indeed, the greatest value of the sum

sin a1
will be
provided
a1
+ sin a + . . . + sin an
2
+ a + ... + a,,
2
(a;
>
O);
and this greatest value is attained at

l1
G'[=ll2= ...
=an=---n
49. Let us prove that the difference

xP-1
xq -
(if x =I= 1 and p > q) exceeds zero. To this end it is suflicient to prove that
('). =
q (xP -1)- p (xq -1)
> 0.
First let us assume that x > 1. We have

t'l ~~ q (xP - 1) - p (xq - 1) = ( x - 1) { q ( xP-I + xP- 2 +
... +
+x+1)-p(x<1- 1 +xq- 2 + ... +x+1)}=(x-1){q(xP-i+

+ xP-2+ ... +:xq)-(p-q) (xq-1+xq-2+ ... x+ 1)}.
If x
> 1,
then
xP -1
+ xP-2 + ... + xq > (p- q) xq.
Therefore
('). =
q (xP - 1) - p ( xq - 1)
> (x -
1){ q (p - q) xq -(p-q) qxq- 1} = qxq- 1 (p-q) (x-1) 2
> 0.
Thus, if x > 1, the theorem is proved. Now let us assume

that x
1. In this case we have
<
.z:P-1
+ xP-2 + ... + xq < (p-q) xq,

... + x + 1 > qxq- 1,
xq-I + xq- 2 +
q(xP- 1
+ ... + xq)-(p-q) (xq- + ... +x-f--1) <

1
<
(p-q) qxq -q (p--q) x'l-l = q (p-q) xq-l (x-1).
434
Solutions
Consequently
!J.
> q (p-q) xq-i (x-1) > 0.

2
However, this proposition can be proved proceerling from

the theorem on the arithmetic mean. We have the following
inequality (see Problem 40)
(1 + CG)').. > 1 + CGA

('A> 1, rational, CG> 0, real).
Likewise we can deduce the following inequality
>
(1- CG)t..
1-a'A
if 0 <CG< 1; 'A> 1, rational. Csing thPse inequalities,
we shall prove that
xP-f
>
if p
xq--1
q
> q (x =fo 1).
Put
xq =
c,.
.!!_=='A.
Then we have to proYe
q
t.. - 1 > 'A (s - 1)
or
!.. - 1 - 'A( - 1) > 0.

First suppose x > 1, > 1. Puts = 1 +CG. We then have
!.. - 1 - 'A( - 1) = (1
a)t.. - 1 - 'Aa > 0.
[f
<
1, then
< 1.
S= 1-
In this case we put
CG
(0 <CG< 1).
We find easily
' - 1 - 'A( - 1) = (1 50. Let us first assume that m
a)!.. - 1 - 'A (-CG)> 0.
>
1. Put m = p_
(p
q
positive integer). We then have (see Problem 49)

P-1
p
> q-f
q
( =/= 1).
Putting
sq=X, =xq,
we get
xm - 1 > m ( x - 1).
>
q,
435
Solutions to Sec. S
1 , we frnd
.
l"1ty x b y -;;Replacing in this rnequa
1
(1 - 1).
--1>m
xm
x
Multiplying both members of this inequality by- :r"', we get

x 111 -1 < mxm-J (.r-1).
Thus, if m
> 1,
then
mxm- 1 (x-1)>xm-1>m(x-1).
Let us assume now that 0
< m < 1.
Putting
(1)
sq= x,
_J_ = m, we find
p
Replacing here x by xm, we frnd
xm - 1
< m lx -
1).
Replacing in the last inequality x by _!.._, and performing

x
all necessary transformations, we find
mx"'- 1 (x-1)<x"'-1<m(x--1)
(O<m<1).
(2)
Let us now consider negative values of m. Put m = -n,

where n > 0, rational. Let us first prove that if m is negative, then
xm-1
m (x-1).
>
>
Since n
0, it follows 'that n 1>1 and we may make
use of inequalities (1). Namely, we have
xn+i_1
(n+1)xn(x-1).
Hence
nxn ( x - 1)
xn - 1.
<
>
Replacing here n by-m, we find

-mx-m (x-1)
x-m-1.
>
Multiplying both members of this inequality by - xm,

we get
436
Solutions
And if we replace here x by
xm-1
~,
< mx"'-
then we find
(x--1).
Thus, indeed
mxm-t (x-1)
if 0 < m
< 1,
< x"'-1 < m (x-1),
1) < xm - 1< m x
m (x -
m-i (
x-
1)
if mis any rational number not lying in the interval between

0 and 1, and x is any real positive number not equal to
unity.
51. The inequalities of this problem follow immediately
from the results of the preceding problem.
52. Put
Then the inequality is rewritten as follows
)m:::::::
( Y1+Y2++Yn
n
Yf'+YJ'+ ... +y;:'
-::::
'
>
where m
1, rational. Using tho results of Problem 47,
it is sufficient to prove that
( t1+t2
)m:::;:::
tf'+tT
"""
for any rational m > 1 and for any real positive ti and t 2
In other words, it is sufficient to prove that
(
2t1
ti+t2
)m ( 2t2 )m 2
+ ti+t2 > .
(1)
Let us make use of the results of Problem 51

(1+x)m~1+mx
if m > 1 is rational and 1

lities
+x >
0. We have two inequa2ti
r~1+m (
ti +t2
r~1+m(
t, +t2
2t2
-1)'
-1).
Solutions to Sec. 8
437
Adding the:i, we get inequality (1) which is the required

result. The solution to our problem can be obtained immediately from the inequalities of Problem 51. Let us show
that, using this method, we can deduce even a more general
inequality. So let us prove that
YA+YA+.
+YAn
( Y1+Y2+ +Yn )A_,..~
l
2
n
if 'A is a rational number not lying in the interval between

zero and unity and
( Y1+Y2++Yn
)A:>-:
:::--
Y}+Y}+ ... +y~
if 0 <'A< 1. To prove the first inequality it is sufficient

to prove that
But we have (see Problem 51)

(
ny;
Y1+Y2+ +Yn
)A> 1 + '). , (
nyi
Yi +Y2+ +Yn
1 ).
Putting here i = 1, 2, ... , n and adding the inequalities

thus obtained, we actually get inequality (2). We proceed
quite analogously for the case 0 < 'A < 1.
53. Put
X1
+ X2 + ... +
Xn
p, x~
+ x; + ... + x; = p'.
We have
(x-x1) 2 +(x-x2)2 + ... +(x-xn) 2 =
2p
p'
==nx2 -2px+p'=n [_x 2 ----nx+---n
-=
=n[(x- ~ )2+ ~ - ~:].

Our expression can attain the least value only simulta2
2 (
neously with ( x- ~ )
since the quantity : - ~2 is
I
independent of
x) .
But ( x-
cannot be negative,
438
Solutions
therefore its least value will be equal to zero. Hence

p
X=-=
n
x1+ .. +xn
.
n
Thus, the sum

(x-x 1) 2 + (x-x2)2+
... + (x-xn)
attains the least value at

x = x1+x2+ ... +xn
---"---"------'-~
54. Put
Then
(x1-X2) 2+(x1-xa) 2+ ... +(x2-x3)2 + ...
(Xn-1 -Xn) 2
+
= (n-1) S2-2q,
where
q=
Further
X1X2
+ X1X3 + ... + X1Xn + XzX3 + + Xn-tXn
And so
C2 =S 2 +2q,
wherefrom we find
nS2 = C2
+ j>i
L (x;
-Xj)2
The last equality shows that S 2 takes the least value when
the least value is attained by 2,; (x; - xj) 2 The least value
i>i
of this sum is equal to zero and is attained at

But since
it follows that
x~+
... +x;
Solutions to Sec. 8
439
takes on the least value at

X1
X2
= ... =
nc
Xn
55. First let us assume that 'A does not lie in the interval
between 0 and 1. Then the following inequality takes place
x~+x~+ +x~ ~
~
( x1 +x2+ ... +xn )"'

n
'
the equality sign (as it is easy to find out) occurring only if

If it is given that
X1
+ X2 + ... +
then at all values of x 1 , x 2 ,
;\. + "'+
Xl
X2
;\.
X,,
= C,
Xn
xn, we have
~n
lIC)"'
n '
wherefrom it is seen that the least value of the expression
is n (
if 0
x~+x~+
which is reached at
< 'A< 1,
... +x~
X1
X2
= ... =
Xn
= ~ . But
then the following inequality takes place
xt+x~+
... +x~
~(x 1 +
"""
... +xn)"'.
n
Then at
X1
.'.rz
= =
Xn
we obtain the least value of the quantity
xt+xH- ... +x~.

56. We have the inequality (see problem 30)
n/X1X2 Xn--.:::
~ x1+.r2+ +xn
C
n
=n
Hence
Solutions
440
Thus, the product x 1x 2
( en
does not exceed
Xn
and reaches it only at x 1 = x 2 = ... =
Xn
)n
= .
(see Probn
lem 30). And so the greatest value is attained by the

product x 1x 2 Xn when
X1=X2= ... = X n = n .
57. We have
Consequently
The equality sign being possible if x 1 = x 2

Hence, it is clear that the sum x 1
x2
the least value if
X1
X2
= ... =
Xn
= xn.
+ ... + xn attains
;Y'c.
58. First let us assume that i (i

whole numbers. We have
1, 2, ... , n) are
c
Consequently
1 2
Xl
Xz
n...-(
Xn
::::::::,
+ C +n
+ .+n. 11, ~2 nn '
)i+ 2
and the equality sign is obtained only if

Xj
X2
Xn
- = - = ... = 1
1
IJ.n
Solutions ta Sec. 8
441
Let now ; be fractions. Reducing them to a common

denominator, we put
/.,;
;=,
where 'A; and are positive integers.
Since
X 1x2
1
xn -- 11 /xt.1xt.2
n
-y
xl.n
n'
the greatest value is reached by the product xr1x~2 ... x~n

simultaneously with the product x}1x~2 ... x~n, where 'A;
are integers. As follows from the above-proved. it happens if and only if
Xz
Xj
Xn
Ti""= T;=
T;;.
Dividing the denominators by
~t.
we get
Thus, if .r 1 >0 and x 1 x 2 + ... + Xn = C, then the product xr1x~2 ... x~n (;
0, rational) attains the greatest
value if and only if
>
Xt
X2
-=-~
... -
Xn
59. We have
wherefrom it follows that the product

reaches the greatest value only if
But since
ll1X1
a 2x 2
the product x 1x 2
if and only if
anxn =
Xn
(a 1a 2
an) (x 1x 2
Xn),
indeed reaches the greatest value
- 11
442
Solutions
60. Put
(i=1, 2, ... , n).
a;Xfi=Y;
Then
""I:'""
-a;
Xi= ( Yi )
and
Further
The problem is reduced to finding out when the product

2
Yi:l.y12
1
2
yt..n
n
"
+ + ... +
takes on the greatest value if y 1 y 2

Yn = C. Referring to the results of Problem 58, we see that it will take
place if
Thus, if
a 1 x}ta2x~2
+ ... + an.:x~n =
C,
then the product

reaches the greatest value provided
61. Put
... ,
Hence
... ,
Xn
( Yn
an
)n
'
443
Solutions to Sec. 8
and the problem is reduced to the following: under what

condition does the sum
Y1+Y2+ +y,
attain the least value if

Ai
An
A2
yn = C1
Y1ili. y2
2
n
'
where C1 is a new constant?
.
l , we pu t
.
S mce
-A.1 , ... , -A.n are rat10na
1
A.1
a1
;-=71'
a2
A.2
;-=N'
'A,n
n
... ,
an
N
Then the problem will read as follows: find out when

+ y2 + ... + Yn attains the least value if
y~1y~2 ... y~n = C2
(ai positive integers).
Y1
Finally, we put
and obtain the following problem: under what conditions
does
tX1U1
tX2U2
+ . + UnUn
attain the least value if

But
a1u1 +a2u2+ ... +an Un:>-:
a1 +a2+ .+an
~
--...__ a1+a2+ ... +an/ a 1 a 2
Hence
a1u1
Ui U2
+ ~u2 + ... + anUn
then
3.
attains the least value when
U1=U2= . . . =Un
Thus, if
an_ a1+a2+ . . +an/C
Un
444
Solutions
attains the least value provided

xt
~1 =
a11
x2
:2
xn
= ... = -t:-
ann
a22
62. Applying the Lagrange formula (see Problem 5,

Sec. 1), we have
(x2
y2
z2
t2) (a2
b2
c2
k2)
= (ax
by
kt) 2
(xb - ya) 2
(xc - za) 2
Since
a2
b2
c2
k2
is constant and
ax
by
kt = A
+ + + ... +
+ + + ... +
+ + ... + +
+
+
+ ...
+ + + ...+
+
+ ...+
(by hypothesis) and, consequently, also constant, it follows

that the sum
x2
y2
z2
t2
+ + ...+
attains the least value simultaneously with the sum

(xb - ya) 2
(xc - za) 2
+ ... .
But the least value of the latter sum is zero which is reached
when
xb - ya = 0, xc - za = 0, ... ,
i.e. when
Let us put this general ratio equal to 'A so that

x = a'A, y = b'A, z = c'A, ... , t = k'A.
Substituting these values for x, y, z, ... , t into the equality
ax + by + . . . + kt = A ,
we find
A
'A= a2+b2+ ... +k2 '
and, consequently, the required values of x, y, ... , t at

which the expression x 2 y2
t 2 takes on the least
+ + ... +
445
Solutions to Sec. 8
value will be
aA
X=-----~
a2
+ b2 + ... + k2
'
kA
bA
Y=
a2
+ b2 + ... + k2
t=
' '
a2
+ b2 + ... + k2
63. We have
il
Ax2 +2Bxy-j-Cy2 -t- 2Dx+2Ey + F,
where
A
+a b + ... + an!J,,,
D = a c + a c + ... + anc,,,
a; +a;+ ... + a;,,
C=b;+b~+ ... +b~,
E=b 1c1+b 2c2
B = a 1b1
2 2
1 1
2 2
+ ... +b,,cn,
Put
x = x'
+ a,
F=c~+c~+
y = y'
...
+c~,.
+ ~-
We then obtain
u
A (x'
+ a) + 2B (x' + a)
2
(y'
+ Bl + C (y' +
+ 2D (x' + a)
2E (y'
~)
~) 2
+ F.
Expa11ding this expression in powers of x' and y', we gPt
Ax' 2
+ 2Bx'y' + Cy' + 2 (Aa +BB + D) x' +

+ 2 (Ba + C~ + E) y' + F'.
2
Now let us choose a and ~ so that the coefficients of x' and

y' in the last expansion equal zPro. To this Pnd it is only
necessary to choose a and ~ as the solutions of the following
system
Aa + B~ + D = 0,
Ba+ C~ + E = 0.
Then we have
u
Ax' 2
Further
U=
{A 2 x' 2
+ 2Bx'y' + Cy' + F'.

2
+ 2BAx'y' + ACy' + F' =

2}
{(A:r'
+ By') + (AC-B
2
2)
y' 2 } + F'.
Solutions
446
But
AC-B 2 =(a;+ a;+ ... a~) (b; + ... b~)-
-(a1b1 + a2b2 + ... + anbn) 2 :;> 0,
Therefore, u attains the least value when

Ax' + By' = 0 and y' = 0.
Hence
x' = y' = 0 and x = a, y =
A> 0.
And so, the values of x and y at which u attains the least

value are obtained as the solution of the following system
of equations
Ax + By + D = 0, Bx + Cy + E = 0.
However, this result can be obtained in a somewhat
different way.
Put.
a1x+b 1 y+c 1 =X 1 , a 2x+b 2y+c 2 =X 2 ,

a 11 x
b11 y + c = X
11
11
Let A1, A2, ... , A11 be some constants satisfying the following conditions
a1A1 + a 2 A2
a,,A,, = 0,
+ ... +
b1'A.1 + b2A.2 + ... + bnA,, =
C1'A.1
0,
+ Cz'A.2 + -f- Cn'A.n = k,
where k is an arbitrary number.

We then have
A.1X1
+ A3X2 + ... + AnXn
and hence, we have to find the least value of the expression

provided
A.1X1
+ A. X + ... + AnXn
2
k (constant).
From the result of Problem 62 we have that the least value

is obtained if
447
Solutions to Sec. 8
Or
A1 = X 1,
A2 = X 2, ... , An = Xn.
Substituting them into the first two equalities (*), we fii,d
+ a2X2 +
a1X1
b1X1
... + anXn = 0,
+b2X2 + ... + bnXn = 0.
Hence we get the system obtained by the preceding method

of solution.
64. As is known, there exists the following identity
(see Problem 77, Sec. 6)
f (x) = f (xo)
(x-x1) (x-x2) ... (x-xn)

(xo--x1) (x 0 - x2) ... (xo-Xn)
-t-f (xi)
(x-xo) (x-x2) ... (x-xn)

(x1-xo)(x1-x2) ... (x1-xn)
+ f (xn)
'
(x-xo) (x-x1) ... (x-Xn-1)

(Xn - Xo) (Xn - Xt) ... (Xn - Xn-1)
where f (x) is any polynomial of degree n.

Equating the coefficients at xn in both members of this
equality, we find
1=
f (xo)
(xo-x1) (xo-x2) ... (xo-Xn)
+ (x1-xo)(x1-x2)
/(x1)
++
... (x1-xn)
+
f (xn)
(xn -xo) (xn -xi) ... (xn-Xn-t)
Let M denote the greatest one of the quantities

l/(xo)I,
l/(x1)I,
.,
l/(xn)I
Then
1~M
1
{ I (xo-x1) (xo-x2)
+I (x1 -xo)
. . (X1 -xn)
... (xo-Xn)
I + ... +I (xn -xo)
1
.. (xn -Xn-1)
As is easily seen, by virtue of our conditions we have

[ (xk -
Xo) (xk -
X1) ... (xk -
Xk-1) (xk -Xk+t) ... (xk -xn) I~
>k! (n-k)!.
448
Solutions
Therefore
1
I (xk -xo) (xh -x1)
... (xk-Xn)
I~
1
k! (n-k)!
Consequently
n
1 ~ M "'1
1
=
-----= ' L..J k! (n-k)!
h=O
..!!.._ "'1 Ck = M !!::_

n!
L..J
n!
h=O
Finally
M
nl
?><jn.
65. Since sin 2 x

cos 2 x = 1, i.e. the sum of the two
quantities sin 2 x and cos 2 x is constant, their product
sin 2 x cos 2 x reaches the greatest value when these quantities are equal to each other. It happens at x =-~ ~
However,
the same is easily seen from the identity

.
Slll XCOS
. ?
x = 21 sin
~x.
66. It is known that if
. +Z=2,
n
x+y
then
tan x tan y
+ tan x tan z + tan y tan z
= 1
(see Problem 40, 4, Sec. 2). Thus, the sum of the three
quantities
tan x tan y, tan x tan z, tan y tan z
is constant. Therefore, the product of these quantities
tan 2 x tan 2 y tan 2 z
reaches the greatest value if
tan x tan y = tan x tan z = tan y tan z,
i.e. if
tan x = tan y = tan z
and consequently at
Solutions to Sec. 8
449
67. We have
68. Put
It is required to prove that
a,2n_ 1 ?;:;n (an+l _an-1).
Or, which is the same,
But
a2n_f
-- =
a2-1
a,2<n-ll
-+-
,.,2<n-2>
""
...
+ a2 + 1 ~
::=--
>n ;Y-a-2.-a,-4-.-.-.-a-2-n--2
(using the theorem on the arithmetic and the geometric

mean of several numbers).
Since
2 + 4 + ... + (2n - 2) = n (n - 1),
-we have indeed
69. Rewrite the sum in the following way
+ ~ + (++ i2 )+ (++ i- +
+(
1
2n-2+1
++ ~3 )+ ... +
1 )
1
1
+ + 2n-t
+ 2n-1+
1 + +2n-1
Each of the bracketed expressions exceeds ~ and, consequently, the total sum is more than ; . On the other hand, the
Solutions
450
sum may be rewritten as
1 +(-}+-})+(
! +i-+i-+4-)+ ... +
1
1
1 )
+ ( 2n-1 + 2n-1+1 + ' + 2n-f '
But each of the bracketed expressions is less than unity,

and, consequently, the total sum is less than n.
70. On transformation we get the inequality
(a + c) (a + b) (b + d) (c + rI) - (a + b + c + rI) (c + rI) ab - (a + b + c + d) cd (a + b) ~ 0,
or the following one
(ad -
bc) 2
0.
1. Putting in the basic formula n
Vz
= 1, we find
= 3v1 - 2v 0 = 3 .3 - 2 2 = 5 = 22 + 1.
Suppose that
vk =
2k
+1
and let us prove that

Vn+t
(k = 1, 2, ... , n),
= 2n+1 + 1.
Indeed
Vn+-1=3vn-2Vn-1=3(2n+1)- 2(2n-1+1) =
= 3.2n+3-2n-2= 2n (3-1) + 1=2n+l + 1.
2. Solved as the preceding problem.
3. As is easily seen, the required relation is indeed valid
at n = 1.
Assuming its validity at the subscript equal to n, let us
prove that it is also valid at the subscript equal to n
1.
45i
Solutions to Sec. 9
Indeed
=(
But by supposition
an--VA
an+ -VA
Therefore
IZn+t-
-VA
an+1+
-VA
4. We have
Hence
Consequently
It is easy to see that there Pxists the following general

formula
- ( i)n-1 a1zn-I
-ao
lln-lln-1- -
452
Solutions
Adding term by term all the last formulas, we have
an
_ a1 =
_a1-a0 +a1-a0 _a1-a0
22
23
_1
- - a1-ao
2 ( 1-J...-t2
22
+ . +(- 1)n_
a1-ao =
2n-I
+ ' ' ' + (-1)n-2_1_)

2n-2 -_a1-ao { ( - 1)n- _1__ 1}
3
.
1
2~1
Renee, fmally,
- 2a1 + ao
an--3-
+ (-
1)n-1 a1 -ao
3.2n-I
5. Consider the relationship
a,,= 3a,,_1+1.
Putting here k equal to 2, 3, 4, ... , n, we get
n
~ ak=3 ~ ak-1+n-1.
k=2
11=2
Put
We then have
S - a1
3 (S - an)
+n-
1.
Consequently
1
S = 2{3an-a1-n+1}.
It remains to express an in terms of a1. We have
Hence
Therefore
an - an-I
3 (an-I - an-2) = 32 (an-2 - an-3)
= 33 (an-3 - an_,) = ... = 3n- 2 (a2 - a1)
453
Solutions to Sec. 9
But
And so
an - an-t = 5 .3n-2.
Putting here n equal to 2, 3, 4, ... , n, we have
a 2 - a 1 = 5 1,
aa - a2 = 5 3,
a 4 - a 3 = 53 2 ,
an - an-I = 5 .3n- 2
an - a1 = 5 (1 + 3 + 32 + ... + 3n- 2)
=
~ (3n-1 - 1).
Rewrite the expression for S in the following way

1
S = 2 {3 (an-a 1) +2a1-n+ 1}=
=+ {
1:
6. We have
Consequently
(3n-1_1)+4-n+1} ={-{5(3 11 -1)-2n}.

an = kan-1 + l,
an-1 = kan-2 + l.
an - an-I = k (an-t - an-2) = k2 (an-2 - an-a) = =

= kn- 2 (a2 - a1)
Hence
th-a1 = (~-a1),
aa-a2= k (~-a1),
a4 -aa = k2 (a 2 -a 1),
an - lln-t = kn- 2 (a2- a1)
Adding these equalities, we find
1
an= kn- a1 +
kn-1-1
k-1
l.
454
Solutions
7. Rewrite the given relationship in the following manner

an+t -
an -
an-1) = 1.
(an -
Put
an -
an-t
= Xn
(n
= 2, 3, 4,
... ).
We then have
Putting here n equal to 2, 3, ... , n - 1 and adding, we find
Xn -
X2
= n - 2.
Putting then in the equality
= 3, 4, ... ,
n and adding, we get
an -
a2
= X3
x4
+ ... +
Xn.
And so
But
n
Xk=
~ (x2 +k-2)=(n--2)x2 +(n-2)+
+ (n-3) + ... +
1 = (n-2) x2+
(n-1) (n-2)
Hence
_ + (n- 2) X2 + (n-1)2(n-2) -_
=az + (n- 2) (az-a, )+ (n-1)2(n-2) =
an-tlz
(n-1) (n-2)
1 2 -(n-)a,.
2
+(n-)a
8. Put
Then the following relationship will take place
Xn+t - 2xn + Xn-1 = 1.
Solutions to Sec. 9
455
Using the result of the preceding problem we have

Xn
(n-1) (n-2)
+(n-1) X2-(n-2) x
But it is obvious that

n-2
an-~=x,+xz+ ... +xn-2= ~ xk.

k=1
Consequently
n-2
an-a2
={
(k -
1) (k -
2)
k=1
n-2
X2
n-2
~ (k k=i
1) -
X1
~ (k - 2).
k=i
Finally
an=
(n-1) (n-2)
aa-(n- 3)(n-1)a2 +
+ (n-2)2(n-3) a1 +
(n-1) (n-2) (n-3)
9. The required formulas can be deduced by the method

of mathematical induction. It is evident that they take
place at n = 1. Since
assuming that the formulas are valid at n-1, let us

prove their validity at n. By supposition, we have
lln-t = a +
(b- a) ( 1 -
4;_ 1 } ,
bn-1=a+
i (b-a)(1+
2 .4
n-t)
Then
lln= an-11bn-1 =a+; (b-a)(1-4~)

and, consequently, this formula takes place for any whole
positive n. It only remains to prove that the formula for
bn is true for any whole positive n as well.
Solutions
456
We have
bn= an+bn-1
2
and the proof is completed.
However, this problem can be solved in quite a different
way. It is obvious that
lln-1 + bn-1
b _ lln-f + 3bn-t
an=
1
n4
2
Multiplying both members of these equalities by some
factor /..., we get
an+Mn = (-}
+{ /...) lln-1+ ( --}+ ! /. .) bn-1
Let us choose /... so that
++ ! /.. +++/. .) /. .
= (
There will be two required values of /..., and they will be

the roots of the equation
/... 2 -
/... -
2 = 0,
i.e. will be equal to /... 1 = 2 and /... 2 = -1.

And so, at these values of /... there exists the equality
lln
+Mn=
(-}++A) (an-t + Mn-1),
which holds true for all whole positive values of n. Putting here n consecutively equal to 1, 2, 3, ... , n, we get
+{-J...) (a+J...b),
~ + M2= ( -}+{ /...) (a1 + M1),
a1+f...b1= (;
lln
+ Mn =
( ;
+ {- A) (lln-1 + Abn-t).
Multiplying these equalities termwise, we find
Solutions to Sec. 9
457
for any whole positive n and at A,= 2 and -1. Substituting these values of A., we find
an+2bn = a+2b,
1
anbn =Tn(a-b),
wherefrom we have indeed

an= a+
(b - a) ( 1 -
1n ) '
bn=a+ ; (b-aH 1+
/4n )
10. We have
Xn = Xn-1 + 2 sin 2 a Yn-1'
Yn = 2 cos 2 a Xn-1 + Yn-1
Multiplying the second equality by A, and adding the first

one, we get
Xn + AYn = (1+2/.. cos 2 a) Xn-1 + (2 sin 2 a -j-A) Yn-1
Let us choose /.. so that the following equality takes place
(2 sin 2 a+/..)= A, (1+2/.. cos 2 a).
Hence
A= tan a.
We then obtain
(xn + AYn) = (1+21. cos2 a) (xn-t + AYn-1)
or
(xn + AYn) = (1+21. cos 2 a)n (xo + A.y0 ).
Substituting the values of x 0 and Yo and putting in succession A, = tan a and A. = -tan a, we find the following
two equalities
Xn + Yn tan u = (1 + sin 2a)n sin a,
Xn - Yn tan a = - (1 - sin 2a)n sin a.
Hence
Xn = {sin a {(1+sin2a)n - (1- sin 2a)n},
Yn = {cos a {(1+sin2a}n+(1- sin 2at}.
458
Solutions
11. As in the two previous problems, we get

= ~ (xo+AtYo),
Xn
Xn
+ "2Yn = ~ (xo + "2Yo),
AtYn
where 1 =a + A1Y 2 =a + A2Y A1 and A2 being the

roots of the quadratic equation
(~
+ M)
= A (a
+ A.y).
If A.1 = A2 , then we have two equations for determining

two unknowns Xn and Yn and the problem is solved.
Let us now assume that At = A2 Then t = 2 and the
two equations coincide. To determine Xn and Yn proceed
as follows.
We have
Substituting the value of

equalities, we find
Yn
Xn
into the second of the original
= '\' ( - A1Yn-t + ~-l (xo + AtYo)] + 6yn-t
Hence
Yn
Put
Yn
~zn.
+('\'At --6) Yn-1 = y~-l (xo + AtYo)

Then for
Zn
we obtain the following relation
1Zn +('\'At - 6) Zn-t = '\' (Xo + AtYo)

or
Zn=
c5-YAt Zn-t+...l.(xo+AtYo),
fJt
fJ1
wherefrom we find Zn (see Problem 6) and then Yn; Xn is

found by the formula ().
12. Rewrite the given relationship in the following way
Xn -
aXn-t -
~Xn-2 =
0.
Put
a= a+b, ~=-ab
(i.e. a and b are the roots of the quadratic equation al -
459
Solutions to Sec. 9
- as -
= 0). Then we have
Xn - axn-l - bxn-l
abxn-2 = 0,
Xn - aXn-t - b (Xn-t - axn-2) = 0,
Put
Xn -
axn-1
= Yn
The given relationship takes the form

Yn -
= 0.
bYn-1
Hence
Yn = bYn-1
Yn -1 = byn -2
Consequently
Yn = bn-1Y1
For finding Xn we now have

Xn - aXn-1 = bn-IY1
bz,. -azn-i = y 1
or
Zn
= ba
Zn-1
+ bY1
Using the result of Problem 6, we find
-(a- )n-1 Z1 +(:
Zn-
r-
~-1
b
1 Y1
-
Performing simple transformations, we finally obtain

Xn
an_ bn
a- b x 1 -
ab
an-1 _ bn-1
a- b
x 0
However, this problem can be solved by the method used

in the previous problem, if we consider two sequences x~
and Yn defined by the relationships
Xn
axn-1
+ ~Yn-1.
Yn =
iXn-1
+ OYn-1
460
Solutions
13. Solved as the preceding problem. In this case
a=1,
b=--qp+q
14. Considering the two variables

by the relationships
Yn
= r:J.Yn-1 + ~Zn-1'
Zn
Yn
and
YYn-1
Zn,
determined
+ <'>zn-1'
we put
Yn
-=Xn
Zn
Then the variable
will satisfy the given relationship
Xn
Cl.Xn-1 +~
Xn= '\'Xn-1+u" '
and the solution of our problem will be reduced to that of

Problem 11. For instance, in the given particular case
Xn =
Xn-1+1
X n-1+3
__;,;.--"-'-=-
we have
Yn
Yn-1 +
Zn-1
Zn = Yn-1 +
3zn-1
and so on.
The second particular case
Xn-1
Xn= 2xn-1+1
is readily considered in the following way.

Rewrite this relationship as follows
1
2xn-1+1 = 2 +-1- ,
Xn-1
:Xn
Xn-1
Then
1
Xn
Xn-1
---=2.
Putting here n= 1, 2, 3, ... , n and adding, we get
_1___1_=2n,
Xn
xo
Xn=
xo
2nxo+1
15. It is easily seen that

an+tbn+1
. anbn,
461
Solutions to Sec. 9
and, conseql!ently,
at any whole n.
But
Vlln"-V~
Van+Ybn
an - V~ _ an - Van-1bn-1 _
an+ Vanbn - an+ Van-1bn-1 an-1
an-1
+2 bn-1 -
, /
V an-1 n-1
+2 bn-1 +-va n-1 bn-1
,;--_ ,;b- 2
V n-1 )
( v an-1
an-1
+ v bn-1
Un-1= Un-2
2
Un-2= Un-3
u 2 = u~,
U1=
Uo.
Raising consecutively these equalities to the powers 1, 2,

22 , , 2n- 2 , we find
2 n-1
Un-1 = Uo
But
Un-1
= Van-1-Ybn-1
~----Van-1
Therefore we have
+ Ybn-1
an-1-V~
an-1
+V
aobo '
Solutlom
462
16. We have
1
1
1
1 [ 1
1
(2k)3-2k =2k (2k)Z-f =4k 2k-1 -2k+1
_..!_ { 2k-(2k-1) - 2
2k (2k-1)
J=
(2k+1)-2k } 2k (2k+1)
-
1{
1}
=2 2k-1 -2k-2k+2k+1
Therefore
n
(2k/-2k =+ { ( t +-}+
k=i
+2n~1} +
1 )
1
(1
1
1)}
1
1
+ ( 3+5+ +2n-1 +2n+1- 2 2+4+ +2n
=
-2 (
! + ... +
{ 2 ( 1+
!+
+ 2n1 1 } - 1 + 2n1+ 1 -
;,, } } = ( 1 +
! + !+
... + 2n1 1} -
1
1
1)
n
- ( 2+4+ +rn -2n+1
Hence
n
1
n
1
1
1
1
Ll (2k)3-2k+2n+1= 1 - 2 + 3 - 4 + +2n-1k=1

t 7. Let us denote our expression by <J>n (x). We have
cp1 (x) = (1 - x) + x = 1,
cpz (x) = (1 - x)(1 - x 2)
x (1 - x2)
x2 = 1,
wherefrom we can assume that IPn (x) = 1 for any n. It is

easily seen that the following relation takes place
1Pn+1 (x) = (1-xn+i) IPn (x) + xn+i.
Assuming IPn (x) = 1, from the last relation we obtain
<pn+1 (x) = 1. But since <p 1 (x) = 1, it follows that <pn (x)
= 1 for any whole positive n.
463
Solutions to Sec. 9
18. Put
x2
x2n-l
+-f-x4
+
+
-x
1-x2n =cpn(X).
x
1- - 2
Then
x2n
%+1 (x) = cpn (x)
+ 1-x2
n+i
Now it is easy to prove the required formula using the

induction method.
19. Put
(1+x)(1+x2 )(1+x2 2 )
(1+x2n-l)=X.
Multiplying both members by 1-x, we find
X (1-x) = [(1-x) (1 +x)] (1 +x2 ) (1 +x2 2 )
22
=[(1-x2)(1+x2 )](1+x) ... (1+x

=[(1-x4 ) (1+x4)) (1+x8 )
(1+x
2 n-1
(i+x2n-l) =
2 n-1
)=
n
)= ... =1-x2
Hence
2n
-1 +x+x2_j_, ;i;-+
_1
X -- 1-x
i-x -
... +x2n-1 .
20. We have
Let us assume that

1+..!..+a+b1+ ... + (a+1)(b+1) ... (s+1) _
a
abc ... sk
_ (a+1) (b+1) ... (s+1) (k+1)

-
Adding (a+ 1Hb+ 1) (s+ 1) (k+1)

abc ... skl
abc ... sk
to both members, we
464
Solutions
get
(a+1)(b+1) ... (s+1)(k--l-1)+(a--!-1)(b+1) ... (s+t)(k+t)=
abc . . . sk
abc . . . slcl
(a+t) (b+1) ... (k+1) (Z+1)
abc ... skl
and the formula is proved by the induction method.

21. We have
b
a (a--1-b)
c
(a--1-b) (a+b--1-c)
(a--1-b)-a
a (a+b)
a- a+b'
(a+b+c)-(a+b)
1
(a--1-b) (a+b--1-c) = a+b - a--f-b+c'
1
(a+b+ ... +k)(a+b+ ... +k+Z)
a+b+ .. +k+l
Adding these equalities term by term, we find

b
a (a+b)
+ (a+b) (a+b+c) + +
+ (a+b+ .. +k) (a+b--1l
... --1-k+l)
1
a
a--1-b+ ... --l-k--1- l
b+c+ ... --1-k+l

a (a+b--1-c--I- .. --1-k--l-l)
and the identity is proved.

22. We have
Fi(z) =
1 q- (1-z),
-q
Fi(qz)=1 q (1-qz).
-q
Hence
1+F1(z)-F1(qz)=1+q ( 1 - z )1-q - (1-qz)
1 -q
-q
1-qz,
i.e. the identity is true at n = 1.

But
F n (z) = F n-1(z)+ 1 ~:n (1-z) (1-qz) ... (1-qn-Iz),
F n (qz) = Fn-1 (qz) +

qn- (1-qz) (1-q2z) ... (1-qnz).
1 -qn
465
Solutions to Sec. 9
Let us assume that the identity is true at n-1, i.e.

that there exists the following equality
1+Fn-dz)-Fn_t(qz)=(1-qz)(1-q2 z) ... (1-qn-lz).
We then have
1 + Fn (z)-F n (qz) = (1-qz) (1-q2z) .. (1-qn-lz) +
+
qn- l1-z)(1-qz) ... (1-qn-Iz)1 -qn
- 1 ~:n (1-qz) (1-q2z) ... (1-qnz)=
=(1-qz) (1-q 2z) ... (1-qn- 1z) { 1+ 1 ~:n (1-z)-1 q:n (1-qnz)} = (1-qz) (1-q2z) ... (1-qn-Iz) (1-qnz),
which proves the identity for any n.

23. Put, as in the preceding problem,
q
q2
Fn (z)=-(1-z) +
2 (1-z) (1-qz)+ ... +
1 -q
1 -q
qn (1 -z) (1 -qz)
+ 1 _qn
... ( 1 -qn-1 z).
Hence
n
F n (q-n) =
_1_qhq_h ( 1- :n} ( 1- qqn) ... ( 1-
q::l}
h=1
Let us prove that

Fn (q-n) = -n.
We have (see the identity of the preceding problem)

1 +Fn (q- 1)-Fn (1) =0.
But
Fn(1)=0.
Consequently
Suppose
F n (q-n+l) = - (n-1).
466
Solutions
We have
Hence
Fn(q-n)=Fn(q-n+ 1 )-1= -(n-1)-1= -n.
And so indeed
~ 1 ~kqk
q;:
n) {1- :n ) (1 -
1- q1
= - n;
k=i
Putting here q- 1 =a, we get the required identity.

24. Put
a(a-1) ... (a-k+1)
Uk= b (b-1) ... (b-k+ 1) '

a(a-1) ... (a-k+i) (a-k)
Uk+1 = b (b-1) ... (b-k+1) (b-k) "
Hence
(b-k)uti+i=(a-k)uk
Ufi+f
a-k
u;;-=b-k'
Consequently
n
~ uk(a-k)= ~ uk+i(b+1-k-1).
k=1
k=1
But
n
~Uk'-=Sn
k=1
Therefore
n
aSn- ~ kuk=(b+-1) ~ Uk+i- ~ (k+1)uk+ti

k=1
k=1
k=1
n+1
k=1
k=2
aSn- ~ kuk=(b+1)(Sn+Un+i-U1)- ~ kuk
Hence
(a -
b-
1) Sn
(b
U1 -
+ 1) (un+t - U1) +
(n + 1) Un+t = (b -
Now Sn is readily found.
n) Un+t -
bu1.
467
Solutions to Sec. 9
25. Proved easily by the induction method.

26. Both identities are easily proved by the induction
method.
27. The left member is equal to
n
"'"'( 2k-1
1 - 4k-2
1 -4k
1) = ~
"'"'( 2k-1-2
1
1 2k-1
1 - 4k
1) =
~
k=i
k=i
~ (~
2k1 1-
~. 21k) =+ ~
k=i
(2/ 1- ;k) =
k=i
1(
1 1 1
1
1)
=2 1 -2+3-4+ +2n-1- 2n
28. If a sequence of numbers
relationship
Xn
GtXn-1
Xn
is determined by the
+ ~Xn-2
at the given initial values x 0 and x 1 , then there exists the

following general expression for Xn
Xn
an-bn
a- b X1 -
an-1-bn-l
Xo,
a- b
ab
where a and b are the roots of the quadratic equation

s2-rx.s-~ =0
(see Problem 12).
In our case we have the following relationship
Un
i.e. rx.
~ =
Un-t
+ Un-2
1 and u 0 = 0, u1 = 1. Therefore
where a and b are the roots of the equation s2 -s-1=0,

so that we may put
_1+-V5
ll-
b-1--V5
I
Finally,
Un
=-1
-V5 {(1+1/5}n-(1--V5}n}
2
2
Solution!
Using this expression for Uni we can easily check-the vaiidity

of all the proposed relations (see Problem 6, Sec. 3). However
the last expression for Un can be obtained in a different way.
We shall consider t-he quantities u 0 , u 11 ~. u 3 ,
as coefftcients of some infinite series
cp (x) = Ut U2X U3X2 U4X 3
Un-tXn-'I.
UnXn-1
or
+ ... +
00
Further
00
xcp
00
(x) = ~
Uk+txh+l
h=O
00
UkX 11 I
1&=1
00
x 2cp (x) = ~
Uk+tx 11 +2
h=O
uk-tX 11
1&=2
Therefore
q> (x)-xq> (x)-x21jl (x) =
00
Hence (since
uk+t -
uk -
uk-t
= 0)
<p (x) (1-x-x2 ) = 1
and
1
<p(x)=1-x-x2
_:_x
But the expression 1

2 can be represented in the following form (expanded into partial fractions)
1
1{
1-x-x2=a.-~
a.
~}
1+a.x -1+~x
where
Ct=
115-1
2
A-_
t'-
1/5+1
2
Solution.' to Sec. 9
469
On the other hand,
1:ax=
1:
~x =
1-ax+a2x2+ .. . ,
1- ~x + ~2x2 +
. .. .
Substituting these expressions into the equality (), we find
~ ~ { ( 1 + -V5 )11.+1 _ ( 1- -V5 )11.+1} xii..

2
2
11.=0
1
=
1-x-x2
Li -V5
Therefore, indeed
= _1 {
-V5
Uk+I
( 1
+-V5
)11.+1 -( 1- -V5 )11.+1}
2
2
By the way, all the ten identities of the present problem

can be proved using the method of mathematical induction
as well. Let us prove, for example, identities 7 and 10.
At n = 1 we have
which is really true.
Let us assume that
U1U2
+ U2U3 + +
U2n-3U2n-2
u:n-2
and prove that

U1U2
+ U2U3 + +
U2n-3U2n-2
+ U2n-1U2n
+
=
U2n-2U2n-1
u:n
Indeed, by assumption we have

(u1U2
+ +
=
+
+ U2n-1U2n
+ U2.n -2U2n + U2n U2n =
(u2n-2 +
+ U2n-1U2n =
= U2n-2U2n +
=
= U2n (U2n-2 +
U2n-3U2n-2)
u:n -2
U2n-2
U2n-2U2n-1
-1
-1
U2n-1)
U2n-1U2n
U2n-1)
u:n
Now, as far as identity 10 is concerned, it is readily checked

at n = 1.
470
Solutions
Let us assume that
U~-1 -
Un-aUn-2UnUn+1=1,
and prove that
U~ - Un-2Un-1Un+1Un+2 = f.
To this end it is sufficient to prove that
4
Un -
Un-1
+ Un-3Un-2UnUn+1 -
Un-2Un-1 Un+l Un+2
0.
But we have
4
Un- Un-1
+ Un-aUn-2UnUn+1- Un-2Un-1Un+1Un+2 =
2
=(Un+ Un-1) (Un+ Un-1) (Un -
Un-1)
-f- Un-2Un+1 (Un-aUn - Un-1Un+2) =

= Un+1Un-2 {u~ -f- u;_l -f- Un-3Un- Un-1Un+2} =
= Un+lUn-2 {u~-1- Un-1Un+2 +Un (Un+ Un-3)} =
=
Un+1Un-2 {u;_1-Un-1Un+2
+ 2unUn-1} =
= Un+1Un-2Un-1 { Un-1 since

Un-1 -
Un+2
Un+2
+ 2un} = 0
+ 2un = 0.
29. We have
n
n+f.'3+ +
Un+2
Un+tUn+3
_ 'Y
-
..:.J
k=O
n
u1i+a-uu1
UJi+fUJi+3
"1 ( 1
= L.J UJi+t -
1 }UJi+3
k=O
=(-1
+...!..+
... +-1
)-(-1
+-1
+ ... +-1
)=
U1
U2
Un+t
U3
U4
Un+3
= _1_+_1___
1___1_= u1+u2 Un+2+un+3 _
U1
U3
Un+2
u n+3
Uf U2
Un+2Un+3
U3
= -- -
30. Consider the sequence of numbers
Un+4
----"-~-
Solutions to Sec. 9
li71
determined by the following relationship

Vn+1
Vn
+ Vn-t
We then have
V2 =Vo+ Vi.
V3 = V2
v4 = v 3
v 5 = v4
+ Vt =Vo + 2vt,
+ v 2 = 2v + 3vt,
+ v3 = 3v + 5vt,
0
Using the method of induction, it is easy to get that in

general
Vn
+ UnVt
Un-t Vo
Consider the following sequence

Vo
Up-t
Vt
=Up,
., Vn
Up+n-t
Then we have
Vn
Up+n-t
Un-tUp-t
+ UnUp,
and formula 1 is proved.

Formula 2 follows from 1 at p = n. The proof of formula 3 is reduced to the proof of the following equality
2
Un
2
+ Un-1
=
UnUn+I -
Un-2Un-t
31. On the basis of formula 1 of the preceding problem

we have
Thus, it is required to prove that
3
Un-I U2n +Un U2n+1 = Un+ Un+1 -
Un-I
The proof is rather simple if only the following relations

are taken into account
U2n+1
U2n
32. Put
Un+1 +Un,
= Un-tUn + UnUn+l
Solutions
472
We have to prove that Vn =Un (where Un is the nth term of

the Fibonacci series). Let us prove that for any n there
will be
Let us first assume that n is even and put n = 2l. We have
(~]
Vn+I
(n;l]
= ~ C!-k,
Vn
k=O
(n;2]
= ~ C!-k-1,
Vn-1
k=O
= ~ C!-k-2
k=O
Since n= 2l,
[;]=l, rn-;1]=l-1, [n22J=l-1.

Therefore we have
1-1
Vn
Vn-1
1-1
= ~ C!-k-1 + ~ C!-k-2.
k=O
k=O
Put in the second sum k = k' - 1, then

1-1
~
k=I
Cn-"--1+
k'-1
Vn+Vn-1=1+ Li
Li
k'=1
Cn-k'-t=
1-1
1+ ~
k=I
(C!-k-1 +c~=L1)+c~-=-L1.
But, as is known,
k
Cn-k-1
+ ck-1
n-k-1 = ckn-lr.
Therefore
Vn +vn-1=1
1-1
k=I
k=O
+ ~ C!-lr. + c:=: = ~ C!-k =
Vn+I,
since
cl=1=1 =CL
Likewise we prove that Vn+t =
as well. But it is easy to check that
Vt =
U1,
Therefore it is obvious that

for any n.
V2
vn
U2.
+ Vn-t
for odd n's
473
Solutions to Sec. 9
33. Let us denote the number of whole positive solutions

of our equation by Nn (m). As is easily seen, N 1 (m) = 1.
Compute N 2 (m), i.e. the number of solutions to the equation
X1
X2
= m.
In this equation x 1 can attain the following values: 1, 2, 3, ... ,

m - 1 and, consequently, the equation has the following
system of solutions
(1,m-1), (2,m-2), ... , (m-1,1),
i.e.
N2 (m) = m - 1.
Let us now pass over to computing N 3 (m), i.e. to determining the number of solutions of the equation
X1
X2
X3
= m.
Let x 3 attain the values 1, 2, 3, ... , m - 2. It is clear

that
N 3 (m) =N2 (m-1) +N2 (m-2) + ... +N2 (2) =
=(m-2)+(m-3)+ ... +1 =(m-'11.~m- 2 )=C;,_ 1 :

Using the induction method, we prove that
N ( )=Cn-i=(m-1)(m-2) ... (m-n+1)
n m
m- 1
123 ... (n-1)
.
l t is obvious that
Nn (m) =Nn_t(m-1) +Nn-dm-2) +
... +Nn_t(n-1).
Assuming that
N n-1 (m) =
C:!,:~,
we have
N n (m) =
cn-2 +c11-2 +
m-2
m-3
+ c11-2
11-2 = cn-1
m-1

34. The general form of the equations under consideration will be
k;r; + (k
+ 1) y
= n - k + 1 (k = 1, 2, ... , n + 1).
()
Solutions
474
Let us rewrite this equation as follows
+ y + 1) + y = n + 1
x + y + 1 = z.
y = n + 1 - kz,
= (k + 1) z - (n + 2).
k (x
and put
Then
x
Whatever z may be these expressions yield solutions to the
equation (). Let us see what values must be attained by z
for x and y to be whole and non-negative. And so, the following inequalities must take place
(n
1) - kz ~ 0, (k
1) z - (n
2) ~ 0.
Hence
n+2 ,...
n+i
k+i ""'-z:::;;;-k-'
and z must be a whole number. If n+ 2 is not divisible

by k+1, then z takes on the following values
n+2J + 2 ... , [nt1J.

[ k+i
[ n+2J
k+i + 1
Let us denote the number of solutions of the equation ()
by N k Jn this case we have
Nk=[nt1J-[:t;J.
Jf n+2 is divisible exactly by k+1, then
N
=[n+1J_n+2+
1.
k
k+ 1
But if n + 2 is not divisible by k + 1, then
[:!~]=[~!~];
and if n+2 is divisible by k+1, then
~!~ -1 =[:!~ J.
Thus in all the cases
Nk=[nt1J-[~!:J.
475
Solutions to Sec. 9
And so, the total number of solutions is equal to
N 1+N2 +
... +N n+t = [ n 1] - [ n ! 1] +
+ [ nt1 J-r nt1 J+ ... + [ n~1 ] - [ :t~ J+

+[n+1
= [~]-[
n+1]=n+1.
_n+1 ]-[~]
_n+2
1
_n-t-2
However, this result can be obtained in a different way.
We have
00
00
1
1-qh
kx
'
x=O
1
1-qli+l
q<k+l)!I.
y=O
Therefore
qk-1
(1- qk) (1- qk+I)
00
00
qhX+(h< l>Y+k-1.
x=O Y=O
If we expand the right member of this equality in powers

of q, then it is easily seen that the. coefficient of qn in this
expansion will be equal to N k i.e. to the number of solutions of the equation
kx + (k + 1) y = n - k + 1.
Thus, the quantity
N1 + N2 + ... + Nn+t
will be the coefficient of qn in the following expansion
~
1
q
(1-q)(1-q2)+ (1-q2)(1-q3)
+ (1-q3)(1-q4) + ... +
qn
qn+I
+ (1-qn+l)(1-qn+2) + (1-qn+2) (f-qn+3) +
But it is easily seen, that this expansion is equal to

00
q(1-q)
1-qk+I -1-qk+2
k=O
00
=q(/-q)( 1 ~q-1)= ~(n+1)qn.

n=O
Solutions
476
Hence
+ ... +
N1
N2
Nn+t = n
1.
35. The general form of the equations will be
(k
1) 2 y = [(k
1) 2 - k 2 ] n - k 2
k 2x
(k = 1, 2, 3, ... , n).
+ +
A direct substitution shows that one of the solutions will be

x = -
(n
+ 1),
y = n.
Then, as is known, all the solutions will be obtained from

the expressions
x
= - (n
+ 1) + (p + 1)
t,
y = n - p 2t,
where p is one of the values attained by k.

For x and y to be non-negative it is necessary and sufficient that t attains whole values satisfying the inequalities
n+i
(p+1)2 ~t~pz.
Considering then separately two cases (n

1 is divisible
by (p
1) 2 and n
1 is not divisible by (p
1) 2), we
..come to the desired result.
36. By hypothesis the black balls alternate with the
white ones. Therefore, two suppositions are possible:
(1) the white balls occupy odd positions, i.e. the first,
third, ... , and the black balls even positions;
(2) the white balls occupy even positions, and the black
balls odd positions.
It is easily seen that the white balls numbered 1, 2, ... , n
can occupy odd positions in n! ways, likewise the black
balls can occupy even positions also in n! ways. And so,
according to the first assumption, we have (n!) 2 ways of
arrangement of all the balls.
The second assumption yields the same number of arrangements. Hence, the total number of arrangements of the balls
is 2 (n!) 2
37. Let L~k denote the number of ways in which kn distinct objects can be distributed into k groups of n objects
in each group.
Solutions to Sec. 9
477
In how many ways is it possible to make up the first

group of n objects? It is clear that the total number of the
distinct combinations is equal to C~k and it is obvious
that
ennk Lk-J
k
L nk=
nk-n
Hence
L~k = C~kC?k-t>n ... C~n
38. Let us consider the number of permutations of n

elements in which two definite elements a and b are found
side by side. The following cases are possible: (1) a occupies the first place, a occupies the second place, .. .,
finally, a occupies (n - 1)th place, and b is always on its
right, i.e. in the second, third, ... , nth place, respectively;
(2) b occupies the first place, ... , finally b occupies (n-1)th
place, in all cases followed by a. Thus, the total number of
cases amounts to 2 (n -1), each case corresponding to
(n - 2)1 permutations. Therefore the total number of the
permutations in which two definite elements a and b occur
side by side will amount to
(n - 2)! 2 (n - 1) = 2 (n - 1)!.
Consequently, the number of permutations of n elements
in which two elements a and b are not found side by side
will amount to
nl - 2 (n - 1)! = (n - 1)! (n - 2).
39. Let us denote the number of the required permutations by Qn and put nl = Pn Consider the whole totality
of the permutations Pn. Among them there exist Qn permutations in which none of the elements occupies its original
position. Let us find the number of the permutations in
which only one element retains its original position. U ndoubtedly, this number will amount to nQn-t Likewise, the
number of permutations with only two definite elements
retaining their original position will amount to n <~.-;- 1 > Qn_2 ,
and so on. Finally, the number of permutations \\There all
the elements retain the original position is Q 0 = 1. Thus,
we have
P n=Qn+nQn-1+ n (n-1)
1 . 2 Qn-2+ ... +nQ1+Qo
478
Solutions
This equality can be written symbolically as

pn = (Q
1)n.
Here after involution all the exponents (superscripts)

should be replaced by subscripts, so that Q'h. turns into
Q1i. Consequently, we can write the following symbolic
identity valid for all values of x
(P+xt= (Q+ 1
+ x)n
(since symbolically the power of P can be replaced every1).

where by the same power of Q
Putting here x = -1, we find
Qn=(P-1)".
Passing over from the symbolic equality to an ordinary one,

we have
n
Qn=Pn-TPn-t+
n(n-1)
1.2
Pn-2+
+ (-1)n-i nP + (-1)",
1
( 1
0-n=n! 21-31+41- +
(-1)n-1
(-1)n)
(n-1)1 +-nl-
40. Consider all such permutations of n letters in which

vacant squares may occur along with occupied ones. If
n = 1, then the number of ways in which one letter can be
placed in r squares is equal to r (the first square is occupied
by one letter, the rest of the squares being vacant; the second
square is occupied by one letter, the rest of the squares
being vacant, and so on). All pe:rmutations of two letters
in r squares are obtained from just considered r permutations
by placing the second letter in succession in the first, second,
... , rth square. Thus, the number of permutations of two
letters in r squares will amount to r 2 , and, as is easily seen,
the total number of permutations of n letters in r squares
will be equal to rn. Let us denote by Ar the number of
ways in which n distinct letters can be distributed in r
squares so that each square contains at least one letter.
The number of such permutations amounts to Ar. Then
we shall consider all those permutations in which one and
only one square is vacant. Their number is equal to rAr-t
Further, the number of permutations where two and only
Solutions to Sec. 9
479
two squares are vacant is equal to

r (r-1)
1.2
Ar-21
and so on.
Therefore we have
Ar+ rAr-1 + r (~~ 1 ) Ar-2+ .. +rA1+1=rn+1.

This equality can be written symbolically in the following way
(i.e. after expanding the left member A 11 should be throughout replaced by A11)
Further, we have
r
(A+1+xY = ~ c~x 11 (A+1r- 11

l&=O
This equality yields the following symbolic one which

holds true for all values of x
r
(A+1+xr = ~ C~x 11 [(r-kt+1J.

11=0
Put here x = -1. Then

r
Ar= ~ C~ (-1) 11 [(r-k)n+ 1) =

k=O
11=0
11=0
~ (-1)"(r-ktC~+ ~ (-1)"C~.
But
r
~ (-1)" C~= (1-1r =0.
11=0
Therefore
r
Ar= ~ (-1)" (r-ktC~.

11=0
480
Solutiont
Passing over from the

one, we get
s~mbolic
equality to an ordinary
A,= ~(-1) 11 (r-ktC~=

k=O
=rn- ~ (r-1)"+ r(~~ 1 ) (r-2)"- ... +(-1y-1 r

t. Put a = b1 , so that I b I > 1. Let us prove that
lbr> f+n(lbl-1)
(n> 1).
Indeed
lbln={1 +(I b l-1)}n= 1 +n(I b l-1) +
+ n <;:; 1) (j b 1-1) +
2

I b In> 1-1-n(Ib1-1)
Then
(n > 1).
lxnl=laj"= l:ln <1+n(:bl-1)

and indeed
lim Xn=O.
n-+oo
a>
2. It is easily seen, that we may assume

0. Then
0 (i = 1, 2, 3, ... ). Let k be a whole number satisfying the condition k~a < k+1, so that k~ 1 <1.
Put n > k. Then
Xi>
~
a
a
nr=
1.2.3 ... k k+1 k+2
But
Ii
481
Solutions to Sec. 10
Therefore
~
n!
But since
k-~ 1 <
(-a-)n-h
<~
kl
k+1
1, it follows
that
( k~ 1
r-h -.
O, if
n -. oo, and therefore at any real a we have

an
lim -
n-oo n.
=0,
i.e. the factorial n! increases faster than the nth power of

any real number.
3. Both the numerator and denominator of this fraction
increase without bound along with an increase in n. Consider
separately three cases: k = h, k < h and k > h.
1 k = h. Di vi de the numerator and denominator by
n,. = n". We get
ao
ah
--++11
nh-h
n
b1
bh
bo+-+ ... +-h
n
n
3 k
> h.
Analogously
w~
get in this case
a 0nh+a 1nh-1+ ... +ah

b0nh+b 1nh-I+ ... +bh -+
4. We have
But
00
=0.
482
Solutions
Therefore
++n + 1 -3
_ 2
2
n
n
. p __ 2 ] .
1lffi
n-3 lffi
n-oo
n2
5. Put
111 +2 11 +3 11 + ... +n"
nl&+l
At k = 1 we have Pn1 =
= Pn.
n- 1
--;j;;and consequently
1
. pt
1Im
n=2
n-+oo
Likewise we easily find lim
P~ =
n-+oo
lim
n-+oo
P~ =
+1 1
! . Let
us assume that
for all the values of i less than k, and
1 . p u t s; = ii
2i
. pkn = k+
prove t h at 1im
+ ni. w e
1
then have the following formula (see Problem 26, Sec. 7).
+ + ...
(k+f)s11+
(k+1)k(k-1)
(k+1)k
1. 2
1. 2.3
S1t-1+
S1t-d-
+ (k+ 1) s 1 +so= (n+ 1) 11 +1 -1.

But P~ = __!!!__, therefore we have
nk+I
pk n -
1 (1
k+1
+n1 )"+1 -
1
(k+1)
nk+!

1. p"
lffi
n-+oo
pO
p~-1
-n-n--
k+1--;;;;-'
k+ 1 .
This proposition can be proved directly. Let us make use

of the inequality (see Problem 50, Sec. 8)
m:rm-1 (x-1) > xm-1 > m (x-1)
(x > 0, not equal to 1, m is rational and does not lie
between 0 and 1).
Put here m = k 1 and replace x by ~.
We get
y
(k+1)x11 (x-y)
> x11 +1-y 11 + 1 > (k+1)y11 (x-y).
483
Put here first x=p, y=p-1 and then x=p+1, Y=P

We then fmd
(p + 1)"~1 _ pk+1>(k+1) p" > pk+1 _ (p-1)k+l.
Putting in this inequality p = 1. 2, ... , n and adding,
we obtain
(n+ 1)1i+ 1 -1>(k+1) (11i+21i+ ... +nk) > nk+i.
Dividing all members of t.he inequality by (k+ 1) nk+l,
we find
_1_{(t-:-_!_)k+1 __1_},
k+1
nk+1
_;>
>
11i+21i+ ... +nk

nk+l
1
k+1
Hence it follows that

11i+21i+ ... +nh
1
.
1Im
--n-oo
nh+1
- k+ 1 "
6. Using the notation of the preceding problem, we get

11i+21i+ ... +n"
(Pk __1_)
__n_=
k
nh
+1
+1
Making use of the expression for P~ obtained in the preceding problem, we have
n ( p~ -
k~1) =
(n+ l)h+l_nh+I
(k+ 1) nh
Hence
.
1lffi
n
(Phn _ _k +1 _1 }-- l"
lffi
{ (n+1)k+l_nk+l _..!!._ pk-i} _ _!_

(k+1)nk
d 1
pk-1
since
.
l 1m
ri-oo
(n+ 1)1<+1-nk+l
(k+1)nk
an
im
n-oo
= -1 .
7. Frorn Problem 4, Sec. 9 we have

_ 2x1 +xo
3
Xn -
+ (- i)n- 1
(x1-xo)
3.2n-l '
wherefrom follows
.
1lnl
n-oo
Xn
x 0 +2x 1
=-3- .
2,
484
Solutions
8. We have the following relationship (see Problem 3,

Sec. 9)
x11 -VN =
xn+ YN
Since
(x0 -YN)2n.
xo+ VN
-ViV I< 1, we have

I xo-xo+VN
)2n -- 0.
. ( xo- -V:N
1lnl
'
xo+ V N
n .... oo
Hence
lim Xn-VN
n-oo
Xn+ YN
and
lim Xn= VN
n-+oo
And so, we get a method for finding the square root of

a number. It consists in the following: designate any positive
number (say, the approximate value of a root accurate to
unity) by x 0 We represent N in the form of a product of
two factors, one of which is equal to x 0 so that
N
N=x0 Xo
-.
We take the arithmetic mean of these factors and denote it by x1i so that
Xt=
~ (xo+ ~ )
Then we put
and take the arithmetic mean once again
X2=; ( X1 + ~)
and so on.
The error, which we introduce when taking Xn for an
approximate value of V N, can be determined from the formula
485
9. Let us first of all prove that
x'!)>N.
Indeed
m
Xp
But
(1
Xp-1
(1 + N-x;:_I
)m
mxp-I
+ N-:;"-1
)m > 1 + N-x';-1
mxp-I
xpi_
1
x'{;-1

Therefore
x'f;>N
for any whole positive p.
prove that
Xp
Xp -
is a decreasing variable, i.e.
Xp-1
Indeed
=
Xp-Xp-1
<
0.
N-x';-1
m
mxp-I
< 0.
And so, the variable Xn decreases but remains positive.

Therefore it has a limit. Designate this limit by 'A. From
the relation
as n--+ oo, we get
A.= m-1 'A+~
'Am =N and A= ~N.
mA,m-1'
It is obvious that
m/-
Xn
> y'
> xm-1
n
'
which enables us to find the upper limit of the error introduced as a result of taking Xn for an approximate value of
y/N.
10. We have
O<
(see
Problem 4, Sec. 8).
::/T
1
nr<
-,r
.
V n
486
Solutions
Hence follows the required result.

11. It is easy to prove the following inequality
x
x
<' v-1+x-1 <2+x
2
(1+x> 0).
Putting here x = -;:;:2, we ftnd

k
2n2 + k
-. /
< V 1 + -;2- 1 < 2n2
Hence
n
"' 2n2+k
k
1
L.J
<Sn< 2n2
"' k.
L.J
k=i
k=l
The right member is equal to

n
_1_ '1k- n(n+1)

2n2 L.J 4n2
k=l
Therefore the limit of the right member is equal to 4

as n--+ oo. On the other hand,
lim ( ~
2n
n.-.oo
k=i
k=i
~ k- ~ --=--=-k--=-)
2n2+k -
lim
_,.
n
00
k
2n2 (2n2+k)
k=l
But
n
"1
k2
L.J 2n2 (2n2+k)
<
k2
-...! 4n4
k=i
12 __ z2+ ... +n2

4n4
k=i
Consequently
n
1lim {2n2
n-+oo
~k- ~
k=i
k=i
and
lim ~
n.-.oo k=i
k
__ _!_
2n2+k - 4
Thus, both variables, between which Sn is contained,

1
tend to 4 . Therefore
limS 11 =
n-+oo
!.
487
12. We have
x;= a+xn-1
It is easy to see that the variable Xn increases. Let us show

that all its values remain less than some constant number.
We have
x~_ 1 - Xn-t 0,
since Xn- 1 Xn
Hence
a<
<
{Xn-1+ -V~-1) <0.

( Xn-1- -V4a+1+1}
2
But since the second bracketed expression exceeds zero,
t must b e
Xn- 1
lim
Xn_ 1
.
, i.e.
< -V4a+1+1
2
.
.
. bl e
t h e mcreasmg
varia
is bounded, and consequently has a limit. Put

lim Xn =ex. From the original relation between
Xn-l =
n-oo
n-+oo
Xn and Xn _ 1 we get
ex 2
and since ex
ex - a = 0,
> 0, we have
-V4a+i+1
2
ex=
13. Let us prove that

Xn+i-Xn
But
= -V
Xn
is a decreasing variable. We have
n+1
-2 (V n+ 1-Vn).
- ;1
1
Vn-+ 1-l n=-Vn+1+-Vn >-2--V_n_+___i
and consequently
Xn+t
< Xn
But it is possible to prove (see Problem 6, Sec. 8) that

1
1-+ -V2
1
1
+ -V3-+
... + -Vn
>2
v-+ 1-2.
n
Therefore
xn>2(V
n+1-Vn)-2>-2.
488
Solutions
Thus, the decreasing variable Xn remains constantly

greater than-2, hence, it has a limit.
14. Let us first show that Xn
Yn Indeed
>
Xn-Yn=
Xn-1
+2 Yn-1 - y Xn-1Yn-1=21 (1r=-> 0.

r Xn-1-rlr-::---)2
Yn-1
But
Xn-1
Xn-Xn-1 =
+2 Yn-1
Xn-1
i.e. the variable

Yn -
Yn-1
>
Xn
Xn-1=
Yn-1-Xn-1
< 0;
> Xn,
is a decreasing one. On the other hand,
V Yn-1 Xn-1 -
Yn-1
= V Yn-1 (V Xn-1 -
V Yn-1) > 0,
i.e. Yn
Yn- 1 and Yn is an increasing Yariable, wherefrom
follows that each of the variables Xn and Yn has a limit.
Put lim Xn=X, lim Yn=Y We haYe
:rn =
Xn-1
+2 Yn-1 .
Hence
X=
x+y
----Z-
and consequently
X=Y
1
i-q=s 1,
15. We have
,1 _Q=s,
hence q=1-s;-,
Q=1-.!...
But
s
1 + qQ + q2Q2 +
... = 1 _! qQ
=
1
16. We have
S= u 1 +u 1q+ u 1q2+ ... = u 1 (1 +q+q2+ .. . ).
a2 = u; (1
Further
Sn=
u,,q-u1
q- 1
ui
+ q2+q4+ . . ).
1-qn
(1
U1 - 1--=S
1-q2'
-q
s =
uf
-qn) ,
(1-q)2.
489
We have
s2
+ a2 = (1-q~~t1 + q)'
s2 -
2urq
- --(1---q_,r?.'""'<..,...1+.,..---,.q)
(J2 -
Hence
and
Sn
1 - q11 ) =
S (
1-
52_+ 02Jn}
82
0' 2
17. 1 Put :r=~-.

Then /y/>1, and we may put
y
= 1 + p, where p
0.
We have
h n
nh
>
/n
:.t
/=
(1+p)n =
n(n-1) 2 ,
n(n-1) ... (n-k) hi
12 P -r- + 123 ... (k+ 1) P +
+ np+
Assuming that n
/nhxn/-
/y/=
<
nh
(1+p)n
> k,
+ +pn
we find
nh(k+1)!
n (n-1) (n-2) ... (n-k+1) (n-k) ph+I
(k+1)!
pl<+ I
( 1- ! ) ( 1-
! )...(
1 - k-;: 1 ) (n - k) .
But the expression

(k+1)!
ph+1
(1-!)(1-!)(1-k-;;- 1 )(n-k)
-o
if n-+ oo (k constant).
Therefore, indeed
lim nkxn = 0 if n-+ oo.
2 Put
Hence
;Yn-1=a
(a>O). We then have n=(1+ar
n
2
n= 1 +na+ n(n-1)
1 . 2 a+ ... +a.
490
Solutions
Consequently
> n(n-1)
12 a'
2
a2
2
4
< -n- 1 < -n
> 2).
(n
And so
2
a<
-Vn
an<l
2
< vn/-n-1 < Vil
(n
> 2).
Now it is obvious that

.
n/-
hmy n=1.
18. We have
1
1
1
1
n+2.3+ ... + n(n+1) = 1 --;:;-T
_1_
1.2.3+2-3-4+
_J_
1
_!_
n(n+1)(n+2)-- 2
(..!.1
2
(n+1)(11
2)

But
1
1
1
n+2.3+ ... + n (n-t-1) + ... =
!~1!! {1\
+ 2\ + + n (n + 1) }
!~1!! {1 - n ~-1
} = 1.
Thus
1
1 =n+2.3+ ... +
n(n i 1) + ...
Analogously
1
1
1
1
4= f.2.3+ 234 + + n (n+1) (n+2) +
We can prove a more general formula

1
1
1 23 ... (q+ 1) + 2.3.4 ... (q-t 2) + ... +
+ n (n+1) ... (q+n) + =
1
qqJ
(sre Problem 26, Sec. 9).

19. Suppose the series is a convergent one, i.e. suppose
Sn= 1 + ~ + ... +
has a limit which is equal to S as
ll-+ 00.
491
Then lim S 2 n = S. llul on the other hand,

n->oo
S2n-Sn = n+1
+ n+2 + ... +2n > 2
(see Problem 1, Sec. 8) which is impossible. Thus, the series

cannot be a convergent one. However, the divergence of
this series can be proved in a different way. Let 211 < n <
< 211 +1 We then have
Sn= 1 +
+ (
+ {) + ( ~ +
+(
1
211-Lj- 1
! + ~ + ! )+
...
1),
1
+ ... + 2h
-j- 211 + 1 +
... +
n1 .
But
1
1,1
~+T>T~2~+~+TT~>~=2
Therefore
k
Sn>1+ 2 .
But as n -+ oo, also k -+ oo, and consequently Sn -+ oo,
hence, the series is a divergent one (see also Problem 22).
1
1
1
20. Put Sn= 1 + - + - + ... +-.To prove that
2a
3a
na
the series is a convergent one it is necessary to prove that

lim Sn exists. But it is easily seen that Sn increases along
n-+oo
with an increase in n. It remains to prove that Sn is bounded. Let 211 - 1 < n::;;;; 21<. We have
s"::;;;; 1 +
1
( 2a
1)
+ 3ri"
+
+
But
(1
~+
1
(211-l)a
sa1 +
1
1
5a
+-;;a1) +
+ (21<-t+f)a + ...
...
+
1
+ (211-1)a
492
Solutions
And so
or
Sn~1 +
~
t:t-1
+ (2 2)a-1 + + (2k-1)a-1 + .,
Sn~
1--2a-1
Thus, Sn is really bounded, lim Sn exists and the series

n-+oo
converges.
21. 1 We have (see Problem 22, Sec. 7)
ix+ 2x2 + ... + nxn = (x~i) 2 {nxn+ 1-(n + 1) xn + 1},

1+2x+3x2 + ... +nxn-1+ ... =
=lim {1+2x+3x2 + ... +nxn-1}=
n-+oo
-1
1
{ n+1
- n:~ (x-1)2 nx
-
1
n + 1) xn + 1}- (x-1)2
'
since
n-+oo
(see Problem 17, 1).

2, 3 From the results of Problem 33, Sec. 7 we get
1+4 x + 9 x 2 + . . . + n 2xn-1 --J-'
1-1- 23x+ 33 x 2+ ...
+ nx
3 + ... =
n-1
+X
(11-x)3
'
1+4x+x2
( 1 -x)~
22. 1 Follows immediately from Problem 41, Sec. 8.

Hence, we can obtain one more proof of divergence of the
series
Put
lim ( 1 + _!_
n-+oo
= e.
Since the variable ( +

manner, we have
!)
493
tends to e in an increasing
for any whole positive n.

Hence
n log ( 1 +
! )< 1
if the logarithm is taken to the base e. Or
! > log ( 1 + ! ),
1
1 + 21 + 31 + ... +-;>log
2-t-log ( 1 + 21) +
+ log ( 1 + ; ) + ... + log ( 1 +
! )=
- l 234 ... (n+1)_ 1 ( + 1)

- og 123 ... n - og n
Hence
and we get a divergent series.

2 Using the binomial formula, we obtain
=1+n!+ n(n-1) _1 +
( 1+!)n
n
n
1 2
n2
, n(n-1)(n-2) .-1
123
n3
+n(n-1)(n-2) ... [n-(n-1)J._1_=

I
1 23 ... n
1- ( 1 - ! )
=2+__!_(1-!)+12
n
123
n
+ 1 2. 31 ... n
( 1-
nn
(1-~)+
n
! )(1 - ! ) ... (1 -
... +
n n 1) .
Put for brevity
1.2.t .. k
(1-!)(1-!) ... (1-k-:1)=u~.
494
Solutions
Then
(1+ ~
)n = 2+ U2 +U3 + ... + Uk + uk+1 + Uk+2 + ... +Un.
We have
1
Uk
k
n
1--
Uk+I
< 123 ... k
-Uk=
'
k+1
<
1
k+1.
Hence
1
k+ 1 ,
Uk+1 <Uk
Uk+2
1
< Uk+1 k+2
<Uk
1
(k+ 1)2 '
And so
Uk+!+ Uk+2+ +Un<
<
, _1_+
Uk
[
k+. 1 _1 I k+ 1
...
(k+ 1)'-k-1
J<!:!_Ir
Consequently .
1
Uk+1 +uk+2+ +un
< 123 ... k
Hence
0 < (1 +
! r-(2 +
U2
+ .. +Uk)< 1.2 ~ .. k ~
Let n--+ oo. Then

.
1lffi
n-+oo
Uk=
1.2
and, consequently,
O<e-(2+/2+1.~.3++ 1.2.; ... k)< 1.2~ .. k !

wherefrom follows
1
e=2+n+ 1.2.3+ ... + 123 ... k + 123 ... kk

(0<8<1).
495
Solutions to Sec. JO
Thus, we may write

e= 2 --t:-/2+1.;.3+ + 1231... k +
23. We have
. 1 x-sm
. .r = 2.
. 2xsID
1
. 2 1 x.
2 sm
sm 21 x (1 - cos 21 :r ) = 4 sm
4
2
Hence
. f x-sm.c
.
2 sm
2
since sih a <a for a
Differently
<42
.T (
)2 ,
> 0.
.
1 -1
. 1
2 sw
2 x-sm x< B;i;-.
(1)
Replacing here x by 2 x, 4 x, ... , 2n-i x, we find

2 SID
. 1 x - sm
. 1 x
4
2
?
. -1
Slll
8
X-
. -1
SID
4
)3 ,
< -81 ( -4x )3 '
<
1 ( x
(2)
8 2
(3)
(n)
Multiplying inequalities (1), (2), ... , (n) successively by
1, 2, ... , 2n- 1 and adding them, we get
1
.
1 3
2n SIDTr'x-sm
x<-gx
1
1
1 +12+42+
1
+ 2zn-2
Passing to the limit as n--+ oo, we find

.
hm {
S!Il
.
2n
x
x
x -- sin x
2n
-"::::::: 81 :r, 311111 { 1 + 4f

1 -~ 42
1+
1 }
+ 4n-1
496
Solutions
But
S!Il -
zn
lim---=1.
x
Consequently
24. 1 Put
a1
n == 10
a2
an
+ 102 + + 1on
It is required to prove that Sn has a limit as n --+ oo.

As is easily seen, Sn increases along with an increase in n
so that Sn+! ~Sn. Let us prove that Sn is bounded. We
have
S
a1
= 10 +
a2
102
1
1
1 )
+ + 1on ~ g ( 10
+ 102
+ + 10n
<
< g ( 1~ + 1~2 + + 1~n + )
an
And so, Sn < 1 and the series converges.

2 Since w lies in the interval between 0 and 1, let us
divide this interval into ten equal parts. In this event the
number w will be found either inside one of the subintervals
or at its boundary. Consequently, we can find a whole number a 1 (0 ~ a 1 ~ 9), such that
a1
a 1 +1
10~w<-10-,
i.e.
0
Thus, the number w 1
a1
~w-10<10
:h
lies in the interval between 0
and 10 . Let us divide this interval into ten equal parts.
497
Then we shall have
Hence
This operation can be continued in a similar way. Let us

prove that
. ( 10
a,
1lm
n~oo
+ 102
a2
an )
+ ... + 10n
=
(!).
Here the variable increases but remains all the time less
1 , consequently, it has a limit. Consider the
than
variable
a\6
It is easily seen that this variable decreases but remains
greater than ~~ and, consequently, also has a limit. Since

the difference
tends to zero as n - oo, both of these variables tend to

one and the same limit, which, by virtue of the inequalities
will be equal to w.
3 If the fraction is fmite, then, there is no doubt, it is
equal to a rational number. Let us pass over to the case
498
Solutions
of periodicity. In this case we have
i.e. w is a rational number.

Likewise we make sure that a mixed periodic fraction
(i.e. such a fraction whose period begins not with a 1 , but
later) will also be rational.
Making use of some arithmetic reasons, we can prove the
converse; namely, if a number is rational, then its expansion
into a decimal fraction will necessarily be either finite, or
periodic (purely periodie, or m!xed periodic).
Thus, every non-periodic infinite fraction necessarily
yields an irrational number.
z
25. Suppose w is rational, i.e. w = N , where Z and N
are whole ,numbers.
We have
N=y+v+v+ + 1n2 + z<n+1>2 +z<n+2>2
Let us multiply both members of the equality by zn 2 N

and transpose the first n terms from the right to the left.
We get
Zln2 -N (ln2-1+ 1n2 -4 + ... + 1n2-<n-1>2 + i) =

=- N
1211+1
1
1
+ l~"-H
+ l6n+9
+ }
499
Hence
I zzn2 -N (ln2-1 + zn2-4+ ... +1) I<

1
<N
1
{
12n+1
1
z2<2n+o
za<2n+1J
=N
12n+1
1 _ _1_
12n+1
And so
I zzn2 -N (ln2-1 + zn2-4 + ... + 1) I< N
z2n+11 _1 .
If n is taken sufficiently large, then the right member can

be made infinitely small, whereas the left member is an
integer not equal to zero.
2 Proved as 1.
26. We have
1
e=2+2f+m+ ... +nr+

Put
1
(n+1)1
+ ...
e=N
(where Z and N are positive integers).
Then
... +m+
F= 2+21+31+
or
Z
i, + ;
(N -1)!-( 2+
1
(N-+1)!
+ ...
+ ... + ; 1 ) N! =
_ _1_--11
N 1 I (N-+ 1) (N
+ 2)
+ ...
Hence
lz(N-1)1-(2+
<
i, + ;
1
N+1
+ ... + ;, )N!I<
1
(N+1)2
1
(N+1P
+. =N'
which is impossible, since on the right we have a regular

fraction, and on the left a whole number not equal to zero.
Thus, e is an irrational number. If e is represented as a decimal fraction, then it will be an infinite non-periodic fraction. Given below is a value of e accurate to 2 500 decimal
places.
500
Solutions
e==2.71828 18284 59045 23536 02874 71352 66249 77572

95749 66967 62772 40766 30353 54759 45713 82178
27466 39193 20030 59921 81741 35966 29043 57290
59563 07381 32328 62794 34907 63233 82988 07531
15738 34187 93070 21540 89149 93488 41675 09244
82264 80016 84774 11853 74234 54424 37107 53907
55170 27618 38606 26133 13845 83000 75204 49338
67371 13200 70932 87091 27443 74704 72306 96977
92836 81902 55151 08657 46377 21112 52389 78442
77078 54499 69967 94686 44549 05987 93163 68892
77361 78215 42499 92295 76351 48220 82698 95193
28869 39849 64651 05820 93923 98294 88793 32036
30123 81970 68416 14039 70198 37679 32068 32823
53118 02328 78250 98194 55815 30175 67173 61332
96181 88159 30416 90351 59888 85193 45807 27386
87922 84998 92086 80582 57492 79610 48419 84443
84875 60233 62482 70419 78623 20900 21609 90235
49146 31409 34317 38143 64054 62531 52096 18369
76839 64243 78140 59271 45635 49061 30310 72085
01157 47704 17189 86106 87396 96552 12671 54688
02123 40784 98193 34321 06817 01210 15627 88023
74501 58539 04730 41995 77770 93503 66041 69973
76966 40355 57071 62268 44716 25607 98826 51787
65201 03059 21236 67719 43252 78675 39855 89448
75459 18569 56380 23637 01621 12047 74272 28364
64450 78182 44235 29486 36372 14174 02388 93441
70263 75529 44483 37998 01612 54922 78509 25778
64832 62779 33386 56648 16277 25164 01910 59004
93150 56604 72580 27786 31864 15519 56532 44258
30801 91529 87211 72556 34754 63964 47910 14590
49679 12874 06870 50489 58586 71747 98546 67757
88459 20541 33405 39220 00113 78630 09455 60688
84205 58040 33637 9537{) 45203 04024 32256 61352
88386 38744 39662 53224 98506 54995 88623 42818
61717 83928 03494 65014 34558 89707 19425 86398
62953 74152 11151 36835 06275 26023 26484 72870
05958 41166 12054 52970 30236 47254 92966 69381
36450 98889 03136 02057 24817 65851 18063 03644
50704 75102 54465 01172 72115 55194 86685 08003
15219 60037 35625 27944 95158 28418 82947 87610
55990 06737 64829 22443 75287 18462 45780 36192
47564 48826 26039 03381 44182 32625 15097 48279
30899 70388 86778 22713 83605 77297 88241 25611
65070 63304 52795 46618 55096 66618 56647 09711
70462 62156 80717 48187 78443 71436 98821 85596
68620 02353 71858 87485 69652 20005 03117 34392
03293 63447 97273 55955 27734 90717 83793 42163
51326 38354 40001 86323 99149 07054 79778 05669
96690 62951 19432 47309 95876 55236 81285 90413
26029 98330 53537 08761 38939 63917 79574 54016
47093
52516
03342
95251
76146
77449
26560
20931
50569
30098
66803
25094
76464
06981
67385
63463
30436
08887
10383
95703
51930
29725
13419
96970
89613
24796
25620
91644
69829
40905
57320
16674
78369
99707
77275
39207
15137
28123
68532
85263
98197
87779
90717
34447
70959
07321
70120
78533
83241
13722
69995
64274
95260
01901
06680
92069
29760
01416
53696
79312
31825
43117
80429
12509
89422
24496
99418
07016
75051
50354
33224
08868
51246
96409
42251
35743
92622
99828
46959
86298
56812
00169
51177
73327
47109
64310
32275
14965
28183
98139
13991
96437
66394
40160
10259
13908
50054
58048
16072
36133
Solutions to
Sec. 10
501
Let us also give the logarithm of this number to base 10

accurate to 282 decimal places.
log10 e=0.43429 44819 03251 82765
18916 60508 22943 97005
65661 14453 78316 58646
87077 47292 24949 33843
18706 10674 47663 03733
92871 58963 90656 92210
81226 58521 27086 56867
93370 86965 88266 88331
77384 90514 28443 48666
65860 85135 56148 21234
43543 43573 1724 7 48049
55353 05
11289
80366
49208
17483
64167
64662
03295
16360
76864
87653
05993
27. It is easily seen, that if lk (beginning with some k)

are all equal to one another, then we deal with an infinitely
decreasing geometric progression, and w is rational indeed.
It remains to prove that if such circumstance (equality of
all lk beginning with some k) does not take place, then w
is irrational. It can be proved in the same way as in Problem 25.
28. Let us prove that the variable un decreases, i.e. that
Un+t <Un We have
1
Un+i=
Hence
Un+! -
Un=
1+2+3+ ... + 71 + n+ 1 -log (n+ 1).

n
1 - log (n + 1) + log n
n ~- 1 - log ( 1 +
! ).
Consider the variable

_ (f
Vn-
+-n1 )n+I
and prove that it decreases, i.e. that
( 1+
Vn+1
n!1 r+2 < (1 + ! r+1'
<
Vn
or that
i.e. show that

n+I
(
m
1
1+n
)n+2 >
We have (1+a)">1 +a
1+
-.
n+1
(see Problem 40, 1, Sec. 8).
502
S olutionl
1
n+f , we find
by n,
and m by n+
2
n+i
1+i.<n+f>.
( 1+i.)n+2>
n
n (n+2)
.
Replacmg
here
<X
But
n+f
1
1 + n(n+2) > 1 + n+f
And so, the variable Vn= ( 1 +~ r+i decreases. Let us
show that
. ( 1+f)n+1 =e.
hm
n
n ... oo
We have
(
1 TI - f
)ft_- __.;._(1_+ ___.;.

! r+i
__
( 1+
But
lim ( 1 +.!)n = e,
lim ( 1+.!.)=1. Thus, indeed
n-+oo
!)
n.... oo
lim ( 1 +.!)n+t = e and consequently
n ...co
( f +!)'Hi
>e.
Therefore (n+1)log(t+!)>1, log(1+!)>n!f,

and
Un+i-un<O,
and .he variable Un is a decreasing one.

On the other hand,
1
Ua=1+ 2 +3+ ... +-;;-logn> log(n+1J-log n > log ( 1 + ! ) > 0.

Since the variable Un decreases but remains greater than
zero, it has a limit. Let us denote this limit by C.
C = lim { 1 +
! +} + ... + ! - log n} .
503
C is called Euler's constant. Let us give the value of this

constant accurate to 263 decimal places.
c = 0.57721
90082
35988
66467
14631
40144
92353
37673
60087
15177
84793
56649
40243
05767
09369
44724
86542
62535
94279
35203
66115
74508
01532
10421
23488
47063
98070
83622
00333
25952
94816
28621
569
86060
59335
48677
29174
82480
41739
74293
58247
56708
19950
65120
93992
26777
67495
96050
97644
73377
09491
53233
15079
29. We have
. x = 2 sm
. x cos x ,
sm
2
2
I
sm 2
sm 22 =
2 sm
. 23
x
x
cos 23,
2 . x
x
sm 2n cos 2n .
2 sm
. x cos x ,
22
22
sm 2n-i
n
x
x
x
x
x
sm .x = 2 sm 2n cos 2 cos 22 cos 23 ... cos 2n.
Then
. x
2n SJil
Zn
sin x
1
cos
cos
22 cos 23 ... cos 2n
We have
.
sm
lim 2n sin 2xn
n~oo
Put
. ( cos x
l 1m
2
n~oo
co~
211
lim
x = x.
2n
x
x
x )
22 cos 23 ... cos 2n
=
x
= cos 2
cos 22 cos 23
504
Solutions
Then we have
x
sin x =cos 2 cos Z2 cos 23" ...
-i ,
Putting here x =
we find the required formula. The
number n, like e, is irrational and, consequently, cannot be
expressed by a finite or periodic decimal fraction. Given
below is the value of n accurate to 2035 decimal places.
n~3.14159
58209
82148
48111
44288
45648
72458
78925
33057
07446
98336
60943
00056
14684
42019
51870
50244
71010
59825
18577
38095
03530
55478
81754
85836
94482
93313
25338
67823
55706
32116
63698
81647
16136
45477
56887
82796
73929
06744
46776
94657
26535 89793
74944 59230
08651 32823
74502 84102
10975 66593
56692 34603
70066 06315
90360 01133
27036 57595
23799 62749
73362 44065
70277 05392
81271 45263
40901 22495
95611 21290
72113 49999
59455 34690
00313 78387
34904 28755
80532 17122
25720 10654
18529 68995
57242 45415
63746 49393
16035" 63707
55379-77472
67702 89891
24300 35587
54781 63600
74983 85054
53449 87202
07426 54252
06001 61452
11573 52552
62416 86251
67179 04946
79766 81454
84896 08412
27862 20391
46575 73962
64078 95126
23846
78164
06647
70193
34461
48610
58817
05305
91953
56735
66430
17176
56082
34301
21960
99837
83026
52886
46873
68066
85863
77362
06959
19255
66010
68471
52104
64024
93417
94588
7559"6
78625
49192
13347
89835
01653
10095
84886
94945
41389
94683
26433
06286
09384
85211
28475
45432
48815
48820
09218
18857
86021
29317
77857
46549
86403
29780
42522
58753
11595
13001
27886
25994
50829
06040
47101
04047
75216
74694
21641
58692
02364
51818
17321
57418
69485
46680
38837
26945
04712
08658
98352
83279
20899
46095
05559
64823
66482
20920
46652
61173
52724
39494
67523
71342
58537
44181
49951
30825
32083
62863
92787
59361
13891
53311
09277
81942
53464
20569
73263
21992
69956
80665
41757
72147
49648
56209
49886
86360
60424
37137
32645
59570
50288
86280
50582
64462
37867
13393
96282
13841
81932
89122
63952
84674
75778
10507
59813
05973
33446
81420
88235
66111
53381
24972
68617
01671
95559
62080
66024
91419
45863
90927
49911
46728
72350
43852
92192
27232
95068
19652
86960
99581
98258
41971 69399 37510

34825 34211 70679
23172 53594 08128
29489 54930 38196
83165 27120 19091
60726 02491 41273
92540 91715 36436
46951 94151 16094
61179 31051 18548
79381 83011 94912
24737 19070 21798
81846 76694 05132
96091 73637 17872
92279 68925 89235
62977 47713 09960
17328 16096 31859
85035 26193 11881
61717 76691 47303
37875 93751 95778
95909 21642 01989
82796 82303 01952
17752 83479 13151
27855 88907 50983
13900 98488 24012
61989 46767 83744
46684 25906 94912
05803 81501 93511
92726 04269 92279
15030 28618 2n145
21079 75093 02955
98818 34797 75356
90977 77279 38000
14144 19735 68548
33239 07394 14333
22184 27255 02542
79178 60857 84383
00642 25125 20511
85022 21066 11863
95636 43719 17287
33904 78027 59009

V.A. Krechmar-A Problem Book in Algebra-Mir Publishers, Moscow (1978)

Uploaded by

Document Information

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

V.A. Krechmar-A Problem Book in Algebra-Mir Publishers, Moscow (1978)

Uploaded by

Copyright:

Available Formats

3A~AqHHK

First published 197 4

translation, Mir Publishers, 1974

Printed in the Union of Soviet Socialist Republics

1. WHOLE RATIONAL EXPRESSIONS

cos A cos B=-:r[cos (A+B) +cos (A-B)],

1. Prove the identity

by) 2 + (bx + ay) 2.

5. Show that the preceding identity can be generalized in

where n is the number of the quantities a, b, c, ... , l.

7. Prove that from the equalities

+ a2b2 + ... + anbn:::;;; + 1.

8. Prove that from the equality

9. Prove the following identities

+ (2ab)2 = (a2 + b2)2,

10. Show that

+ q2)4 + (2pq + p2)4 =

1. Whole Rational Expressions

11. Prove the identity

12. Prove that

+ 3b)k + (2a + 4b)k + ak + bk =

14. Prove the identities

15. Prove the identity

17. Prove that if a+ b + c = 2s, then

19. Factor the following expression

20. Factor the following expression

22. Factor the following expression

25. Show that

1. Whole Rational Expressions

27. Given 2n numbers: a 1, a 2 ,

29. Given a trinomial Ax2

Then the given trinomial becomes

npq -(P1 - p) 2 - (P2 - P)2- -(Pn-P) 2.

31. Prove that

32. Let Sn= 1 +

33. Prove the identity

35. Let [ex] denote the whole number nearest to ex which

36. Prove that

1. Whole Rational Expressions

+ sin a) (1 + sin b) (1 + sin c)

cos a cos b cos c.

+ cos a) (1 + cos P> (1 + cos y)

Show that one of the values of each member of this equality is

+ b) sin (a - b) sin (c + d) sin (c - d) +

41. Prove that

42. Check the identities:

43. Reduce the following <'xpression to a form convenient

44. Reduce the following expression to a form convenient

cos 15 cos 15 cos T5 cos T5 cos 15cos15cos15 = 2

48. Let A and B be acute positive angles satisfying the

a = cos cp cos' + sin cp sin ' cos t5,

+ a' + a " = 1, b + b' + b" =1,

and since a - b =I= 0, then A = 0. Substituting this value

Subtracting term by term, we have

Since a - b =I= 0, a - c =I= 0, we have

vanishes at four different values of x, then

2. Simplify the following expression

+ x) (1 + y) (1 + z) = (1- x) (1-y) (1-z).

5. Show that from the equality

(a+b+c+ d) (a-b-c+d) =(a -b+c-d) (a+b-c-d)

6. Simplify the expression

ax2 by2 ez2

m (a+b) (a+e) + bm (b+e) (b--l a)+ m (e+a) (e+b)

!h: = a,. +-1- +