Homepage Solution manuals Ivan Niven An Introduction to the Theory of Numbers Exercise 3.7.3 (Number of proper representations of $n$ by $x^2 + xy + y^2$)

Exercise 3.7.3 (Number of proper representations of $n$ by $x^2 + xy + y^2$)

Show that any positive definite binary quadratic form of discriminant 3 is equivalent to f ( x , y ) = x 2 + xy + y 2 . Show that a positive integer n is properly represented by f if and only if n is of the form n = 3 α p β , where α = 0 or 1 and all the primes p are of the form 3 k + 1 . Show that for n of this form, r f ( n ) = 6 2 s , where s is the number of distinct primes p 1 ( mod 3 ) that divide n .

Answers

Proof.

(a)
First we search the reduced forms ( a , b , c ) of discriminant 3 . By Theorem 3.19, 0 < a d 3 = 1 , thus a = 1 . Since d = 3 = b 2 4 ac , b is odd, and | a | < b | a | , so we obtain b = 1 , and c = ( 3 + b 2 ) ( 4 a ) = 1 . This shows that the only reduced form of discriminant 3 is f ( x , y ) = x 2 + xy + y 2 , which is primitive. Therefore any form with discriminant 3 is properly equivalent to f ( x , y ) = x 2 + xy + y 2 (Theorem 3.18), and H ( 3 ) = h ( 3 ) = 1 .
(b)
Let n be any positive integer. By Theorems 3.13 and 3.17, n is properly represented by f if and only if 3 u 2 ( mod 4 n ) for some integer u .

Since 3 is not a square modulo 8 , n is odd, so the decomposition of n in prime factors is of the form n = 3 α p A p β ( p ) ( α 0 , β ( p ) > 0 ) , where the primes p A are odd. Note that 3 9 = 3 2 ( mod 12 ) , but 3 33 ( mod 36 ) is not a square modulo 36 (the squares modulo 36 are 0 , 1 , 4 , 9 , 13 , 16 , 25 , 28 ). If α 2 , then 3 u 2 ( mod 36 ) . This is a contradiction, so α = 0 or α = 1 .

If p A is an odd prime divisor of n , then 3 u 2 ( mod p ) , thus ( 3 p ) = 1 . Moreover, by the Law of Quadratic Reciprocity,

( 3 p ) = ( 1 p ) ( 3 p ) = ( 1 ) p 1 2 ( 1 ) 3 1 2 p 1 2 ( p 3 ) = ( p 3 ) .

Therefore ( p 3 ) = 1 , thus p 1 ( mod 3 ) . We have proven that if a positive integer n is properly represented by f , then n = 3 α p A p β ( p ) ( β ( p ) > 0 ) , where α = 0 or α = 1 , and all the primes p A are of the form 3 k + 1 .

Conversely, suppose that n = 3 α p A p β ( p ) ( β ( p ) > 0 ) , where α = 0 or α = 1 , and all the primes p A are of the form 3 k + 1 .

For every p A , ( 3 p ) = ( p 3 ) = 1 , thus there is some integer u p 0 ( mod p ) such that 3 u p 2 ( mod p ) . Put g ( x ) = x 2 + 4 [ x ] . Then g ( x ) 0 ( mod p ) has a solution u p , such that g ( u p ) = 2 u p 0 ( mod p ) . By the Hensel’s Lemma (Theorem 2.23), this solution lifts to a unique solution v p modulo p β ( p ) , so that 3 v p 2 ( mod p β ( p ) ) . Moreover, since 3 0 ( mod 3 ) is a square modulo 3 , the congruence x 2 3 ( mod 3 a ) has a solution v 3 . By the Chinese Remainder Theorem, there is some integer v such that v 1 ( mod 4 ) , v v p ( mod p β ( p ) ) ( p A ) , v v 3 ( mod 3 α ) . Then v 2 3 ( mod p β ( p ) ) ( p A ) , v 2 3 ( mod 3 α ) , so v 2 3 ( mod n ) , and v 2 1 3 ( mod 4 ) , thus 3 v 2 ( mod 4 n ) . This proves that n is properly represented by f ( x , y ) = x 2 + xy + y 2 .

A positive integer n is properly represented by f if and only if n is of the form 3 α p A p β ( p ) ( β ( p ) > 0 ) , where α = 0 or α = 1 , and all the primes p A are of the form 3 k + 1 .

(c)
Suppose now that n is of this form, so that r f ( n ) > 0 . By Theorem 3.26, since a = b = c , w ( f ) = 6 , and by Theorem 3.27, r f ( n ) = 6 H f ( n ) . (1)

Here H f ( n ) is the number of integers h , 0 h < 2 n such that h 2 3 ( mod 4 n ) (the condition n x 2 + hxy + h 2 d 4 n + x 2 + xy + y 2 is always satisfied by part (a), since the discriminant of n x 2 + hxy + h 2 d 4 n is d = 3 ). Thus H f ( n ) = N 3 ( n ) , where N 3 ( n ) denotes the number of integers h for which h 2 3 ( mod 4 n ) and 0 h < 2 n (see p. 175).

Since h is a solution of the congruence u 2 3 ( mod 4 n ) if and only if h + 2 n is a solution, it follows that N 3 ( n ) is precisely one-half the total number of solutions of the congruence u 2 3 ( mod 4 n ) . To summarize:

H f ( n ) = 1 2 Card { u [ [ 0 , 4 n [ [ u 2 3 ( mod 4 n ) } .

As in Theorem 2.20, put N ( 4 n ) = Card { u [ [ 0 , 4 n [ [ u 2 3 ( mod 4 n ) } , so N ( 4 n ) is the number of solutions of the congruence x 2 3 ( mod 4 n ) , and

H f ( n ) = 1 2 N ( 4 n ) . (2)

By Theorem 2.20,

N ( 4 n ) = N ( 4 ) N ( 3 α ) p A N ( p β ( p ) ) .

If p A , p 1 ( mod 3 ) , so ( 3 p ) = ( p 3 ) = 1 , thus N ( p ) = 2 , and by Hensel’s Lemma, N ( p β ( p ) ) = 2 . Moreover N ( 3 ) is the number of solutions of x 2 3 0 ( mod 3 ) , thus N ( 3 ) = 1 . The number of solutions of x 2 3 1 ( mod 4 ) is N ( 4 ) = 2 . Then

N ( 4 n ) = 2 2 | A | . (3)

If s = | A | is the number of distinct primes p 1 ( mod 3 ) that divide n , then equations (1), (2) and (3) give

r f ( n ) = 6 2 s .

User profile picture
2024-12-03 09:27
Comments