Homepage Solution manuals Ivan Niven An Introduction to the Theory of Numbers Exercise 4.1.26* (Gcd of $\binom{n}{j},\ 1 \leq j \leq n-1$)

Exercise 4.1.26* (Gcd of $\binom{n}{j},\ 1 \leq j \leq n-1$)

Let d be the greatest common divisor of the coefficients of ( x + y ) n except the first and last, where n is any positive integer > 1 . Prove that d = p if n is a power of a prime p , and that d = 1 otherwise.

Answers

Proof. Suppose that n is a power of a prime p . So n = p k for some exponent k 1 . By Problem 2.1.45 we know that

( n j ) = ( p k j ) 0 ( mod p ) ( 1 j n 1 ) .

Therefore p d , where d is the greater common divisor

d = 1 j n 1 ( n j ) = ( n 1 ) ( n 2 ) ( n n 1 )

of the coefficients of ( x + y ) n except the first and last.

We note that d ( n 1 ) = n = p k , thus d = p i is a power of p , and i 1 since p d . Moreover, by Theorem 4.2 (de Polignac’s formula),

ν p ( p k p k 1 ) = i = 1 p k p i i = 1 p k 1 p i i = 1 p k p k 1 p i , (1)

where

i = 1 p k p i = p k 1 + p k 2 + + 1 = p k 1 p 1 , i = 1 p k 1 p i = p k 2 + p k 3 + + 1 = p k 2 1 p 1 , i = 1 p k p k 1 p i = i = 1 k 1 p k 1 i ( p 1 ) ( p k 1 i ( p 1 ) < 1  if  i k ) = ( p 1 ) ( p k 2 + p k 3 + + 1 ) = ( p 1 ) p k 1 1 p 1 = p k p k 1 p + 1 p 1 .

By equation (1),

ν p ( p k p k 1 ) = 1 p 1 ( p k 1 p k 1 + 1 p k + p k 1 + p 1 ) = 1 p 1 ( p 1 ) = 1 .

Therefore ν p ( p k p k 1 ) = λp , where λ p = 1 . Here d = p i λp , where i 1 , thus p i 1 λ , where p λ , so i = 1 and d = p . If n > 1 is a power of p , then

d = 1 j n 1 ( n j ) = p .

Suppose now that n is not a power of a prime, so that n = p 1 a 1 p 2 a 2 p l a l is the decomposition of n in prime factors, with l 2 . Put p = p i any prime factor of n , and a = a i . Then n = p a w , where p w . We show that

ν p ( p a w p a ) = 0 , ( p w ) .

(The condition p w is necessary: for instance 5 ( 5 5 5 ) .) Since

( p a w p a ) = p a w ( p a w 1 ) ( p a w p a + 1 ) p a ( p a 1 ) 1 = i = 0 p a 1 p a w i p a i ,

we obtain

ν p ( p a w p a ) = i = 0 p a 1 ( ν p ( p a w i ) ν p ( p a i ) ) (2)

We compare ν p ( p a w i ) and ν p ( p a i ) for 0 i < p a .

  • If i = 0 , since p w , ν p ( p a w ) = ν p ( p a ) = a .
  • If 1 i < p a , we write i = p b λ , where 0 b < a . Then

    p a w i = p a w p b λ = p b ( p a b λ ) = p b μ ,

    where μ = p a b λ is prime to p : since a b > 0 , if p μ , then p λ , which is false by definition of λ . Therefore

    ν p ( p a w i ) = b ,

    and similarly, with w = 1 ,

    ν p ( p a i ) = b = ν p ( p a w i ) .

Then ν p ( p a w i ) = ν p ( p a i ) for all i [ [ 0 , p a [ [ . Then the equation (2) gives

ν p ( p a w p a ) = 0 (  if  p w ) . (3)

We note that d n = ( n 1 ) , thus d = p 1 b 1 p l b l , where b i 0 .

By equation (3),

p i ( n p i a i )  for  i = 1 , , l .

This shows that p i d for every index i , so b 1 = b 2 = = b l = 0 , and d = 1 .

If n > 1 is not a power of a prime, then

d = 1 j n 1 ( n j ) = 1 .

User profile picture
2024-12-29 11:26
Comments