Homepage Solution manuals Ivan Niven An Introduction to the Theory of Numbers Exercise 5.6.4 (Condition for a circle to be included in an algebraic curve)

Exercise 5.6.4 (Condition for a circle to be included in an algebraic curve)

Let f ( x , y ) be a polynomial of degree d with real coefficients, and set p ( t ) = f ( 2 t ( 1 + t 2 ) , ( 1 t 2 ) ( 1 + t 2 ) ) ( 1 + t 2 ) d . Show that p ( t ) is a polynomial of degree at most 2 d . Deduce that if 𝒞 f ( ) has more than 2 d distinct points in common with the circle x 2 + y 2 = 1 , then this circle is a subset of 𝒞 f ( ) .

Answers

Proof. By definition

p ( t ) = f ( 2 t 1 + t 2 , 1 t 2 1 + t 2 ) ( 1 + t 2 ) d . (1)

If we write f ( x , y ) = ( i , j ) A a i , j x i y j , where A 2 is a finite set such that i + j d for all ( i , j ) A , then

p ( t ) = ( 1 + t 2 ) d ( i , j ) A a i , j ( 2 t 1 + t 2 ) i ( 1 t 2 1 + t 2 ) j = ( i , j ) A a i , j ( 2 t ) i ( 1 t 2 ) j ( 1 + t 2 ) d i j ,

so p ( t ) is a polynomial, because in every term i + j d . Moreover

deg [ ( 2 t ) i ( 1 t 2 ) j ( 1 + t 2 ) d i j ] = i + 2 j + 2 ( d i j ) = 2 d i 2 d .

Therefore deg p 2 d .

Let 𝒞 denote the circle x 2 + y 2 = 1 . Suppose that 𝒞 f ( ) has more than 2 d distinct points in common with the circle x 2 + y 2 = 1 . Assume for the sake of contradiction that 𝒞 𝒞 f ( ) , so that there is some point A = ( a , b ) 𝒞 such that A 𝒞 f ( ) . By a rotation of the axes, we can choose a system of coordinates such that the coordinates of A are ( 0 , 1 ) . The equation of 𝒞 is the same, and the new equation of 𝒞 f is g , where deg ( g ) = deg ( f ) , and g ( 0 , 1 ) 0 .

(Explicitely, if ( e 1 , e 2 ) is the old basis, and ( e 1 , e 2 ) the new basis, where e 1 = b e 1 + a e 2 , e 2 = a e 1 b e 2 , then g ( X , Y ) = f ( bX aY , aX bY ) , so that g ( 0 , 1 ) = f ( a , b ) 0 .)

Replacing g by f in the notations, this brings the problem back to the case where f ( 1 , 0 ) 0 , which we will assume from now on.

Then the parametrization of 𝒞 starting from the point A = ( 0 , 1 ) with x = t ( y + 1 ) gives

( x , y ) 𝒞 ( x , y ) = ( 0 , 1 )  or  t , x = 2 t 1 + t 2 , y = 1 t 2 1 + t 2 . (2)

Let M = ( x , y ) 𝒞 j ( ) 𝒞 . Then M A since f ( 0 , 1 ) 0 , thus there exists some t such that ( x , y ) = ( 2 t 1 + t 2 , 1 t 2 1 + t 2 ) satisfies f ( x , y ) = 0 , so that

f ( 2 t 1 + t 2 , 1 t 2 1 + t 2 ) = 0 .

By equation (1), t is a root of p . By (2), two distinct points correspond to distinct values of t .

By hypothesis, 𝒞 f ( ) has more than 2 d distinct points in common with the circle 𝒞 , which are distinct from A because A 𝒞 f ( ) , thus p has more that 2 d roots. Since deg p 2 d , this implies p = 0 .

Then, by equation (1), for all t ,

f ( 2 t 1 + t 2 , 1 t 2 1 + t 2 ) = 0 .

This shows that

𝒞 { A } 𝒞 f ( ) .

Since f is a continuous function,

f ( 0 , 1 ) = lim t f ( 2 t 1 + t 2 , 1 t 2 1 + t 2 ) = 0 ,

so A 𝒞 f ( ) . But this is a contradiction, since by hypothesis A 𝒞 f ( ) . This contradiction shows that 𝒞 𝒞 f ( ) . □

User profile picture
2025-05-05 08:36
Comments