Homepage Solution manuals Ivan Niven An Introduction to the Theory of Numbers Exercise 3.2.22* (Number of solutions $(x,y)$ of the congruence $a x^2 + by^2 \equiv 1 \pmod p$)

Exercise 3.2.22* (Number of solutions $(x,y)$ of the congruence $a x^2 + by^2 \equiv 1 \pmod p$)

Suppose that ( ab , p ) = 1 and that p > 2 . Shows that the number of solutions ( x , y ) of the congruence a x 2 + b y 2 1 ( mod p ) is p ( ab p ) .

Answers

Notation: Since ( a p ) depends only of the class of a modulo p , for α = [ a ] p pℤ , let us denote ( α p ) = ( a p ) .

(I mimiced the already done Exercises 5.6, 5.7, 5.8 of Ireland-Rosen.)

Proof. Let p be an odd prime, and N the number of solutions of the congruence a x 2 + b y 2 1 ( mod p ) . Write 𝔽 p = pℤ the field with p elements.

Let α , β 𝔽 p denote the classes of a , b in 𝔽 p : α = [ a ] p , β = [ b p ] (and u = [ x ] p , v = [ y ] p ) . We suppose that ab p = 1 , so that a 0 , b 0 ( mod p ) , thus α 0 , β 0 .

Consider the conic in the plane 𝔽 p × 𝔽 p given by

C = { ( u , v ) 𝔽 p × 𝔽 p α u 2 + β v 2 = 1 } ,

so that N = Card C . If we partition the set C in subsets where u takes a given value, we obtain

N = Card C = u 𝔽 p Card { v 𝔽 p α u 2 + β v 2 = 1 } = u 𝔽 p Card { v 𝔽 p v 2 = β 1 ( 1 α u 2 ) } .

Put δ = β 1 ( 1 α u 2 ) . Then Card { v 𝔽 p v 2 = δ } is the number of solutions in 𝔽 p of v 2 = δ , which is 1 + ( δ p ) (see Problem 3.1.23). This remains true if δ = 0 . Therefore, using ( β 1 p ) = ( β p ) ,

N = u 𝔽 p ( 1 + ( β p ) ( 1 α u 2 p ) ) ,

thus

N = p + ( β p ) u 𝔽 p ( 1 α u 2 p ) . (1)

From 1 α u 2 = α ( u 2 α 1 ) = α ( u 2 + γ ) , where γ = α 1 0 , we obtain

N = p + ( β p ) ( α p ) u 𝔽 p ( u 2 + γ p ) ,

so

N = p + ( αβ p ) u 𝔽 p ( u 2 + γ p ) . (2)

It remains to show that, for any γ 0 ,

u 𝔽 p ( u 2 + γ p ) = 1 .

Consider the particular case β = α (and as above, γ = α 1 ), so that α u 2 + β v 2 = 1 v 2 u 2 = γ . By equation (2), since ( αβ p ) = ( β 2 p ) = 1 ,

Card { ( u , v ) 𝔽 p × 𝔽 p v 2 u 2 = γ } = p + u 𝔽 p ( u 2 + γ p ) . (3)

But here we can compute directly Card { ( u , v ) 𝔽 p × 𝔽 p v 2 u 2 = γ } by the change of variables s = v + u , t = v u , equivalent to u = s t 2 , v = s + t 2 . The equation γ = v 2 u 2 = ( v + u ) ( v u ) becomes γ = st .

Let S = { ( u , v ) 𝔽 p × 𝔽 p v 2 u 2 = γ } , T = { ( s , t ) 𝔽 p × 𝔽 p st = γ } . The maps

φ { S T ( u , v ) ( v + u , v u ) , ψ { T S ( s , t ) ( s t 2 , s + t 2 ) ,

are correctly defined: if ( u , v ) S , ( v + u ) ( v u ) = v 2 u 2 = γ , so ( v + u , v u ) T , and if ( s , t ) T , then st = γ , thus ( u , v ) = ( s t 2 , s + t 2 ) satisfies v 2 u 2 = ( s + t 2 ) 2 ( s t 2 ) 2 = st = γ , so ( s t 2 , s + t 2 ) T .

Moreover ψ φ = 1 S , φ ψ = 1 T : for all ( u , v ) S , and for all ( s , t ) T ,

( ψ φ ) ( u , v ) = ψ ( v + u , v u ) = ( ( v + u ) ( v u ) 2 , ( v + u ) + ( v u ) 2 ) = ( u , v ) , ( φ ψ ) ( s , t ) = φ ( s t 2 , s + t 2 ) = ( s + t 2 + s t 2 , s + t 2 s t 2 ) = ( s , t ) .

This shows that φ is bijective (and ψ = φ 1 ) . Therefore Card S = Card T .

Similarly, the maps

ξ { 𝔽 p T s ( s , s 1 γ ) ζ { T 𝔽 p ( s , t ) s

satisfy ζ ξ = 1 𝔽 p , ξ ζ = 1 T , thus ζ is a bijection. Therefore

Card S = Card T = Card 𝔽 p = p 1 .

Then (2) gives

p 1 = Card S = p + u 𝔽 p ( u 2 + γ p ) ,

therefore

u 𝔽 p ( u 2 + γ p ) = 1 ( γ 0 ) .

By equation (2),

N = p ( αβ p ) = p ( ab p ) .

The number of solutions of the congruence a x 2 + b y 2 1 ( mod p ) is p ( ab p ) . □

User profile picture
2024-10-27 08:46
Comments