Algebra 395
holds true forα= 2 nπ+ 1 , 2 n^2 π+ 1 ,..., 2 nnπ+ 1. Hence the equation
(
2 n+ 1
1
)
xn−
(
2 n+ 1
3
)
xn−^1 +···+(− 1 )n
(
2 n+ 1
2 n+ 1
)
= 0
has the roots
xk=cot^2
kπ
2 n+ 1
,k= 1 , 2 ,...,n.
The product of the roots is
x 1 x 2 ···xn=
( 2 n+ 1
2 n+ 1
)
( 2 n+ 1
1
)=
1
2 n+ 1
.
So
cot^2
π
2 n+ 1
cot^2
2 π
2 n+ 1
···cot^2
nπ
2 n+ 1
=
1
2 n+ 1
.
Because 0< 2 kπn+ 1 <π 2 ,k= 1 , 2 ,...,n, it follows that all these cotangents are posi-
tive. Taking the square root and inverting the fractions, we obtain the identity from the
statement.
(Romanian Team Selection Test for the International Mathematical Olympiad, 1970)
163.A good guess is thatP(x)=(x− 1 )n, and we want to show that this is the case. To
this end, letx 1 ,x 2 ,...,xnbe the zeros ofP(x). Using Viète’s relations, we can write
∑
i
(xi− 1 )^2 =
(
∑
i
xi
) 2
− 2
∑
i<j
xixj− 2
∑
i
xi+n
=n^2 − 2
n(n− 1 )
2
− 2 n+n= 0.
This implies that all squares on the left are zero. Sox 1 =x 2 = ··· =xn=1, and
P(x)=(x− 1 )n, as expected.
(Gazeta Matematica ̆(Mathematics Gazette, Bucharest))
164.Letα, β, γbe the zeros ofP(x). Without loss of generality, we may assume that
0 ≤α≤β≤γ. Then
x−a=x+α+β+γ≥0 and P(x)=(x−α)(x−β)(x−γ).
If 0≤x≤α, using the AM–GM inequality, we obtain
−P(x)=(α−x)(β−x)(γ−x)≤
1
27
(α+β+γ− 3 x)^3