Homepage Solution manuals Ivan Niven An Introduction to the Theory of Numbers Exercise 4.3.11 (If $S(n) = \sum_{j \in [\![1, n]\!], \, j \wedge n = 1} j^2$, then $\sum_{j=1}^n j^2 = \sum_{d \mid n} d^2 S \left(\frac{n}{d}\right)$)

Exercise 4.3.11 (If $S(n) = \sum_{j \in [\![1, n]\!], \, j \wedge n = 1} j^2$, then $\sum_{j=1}^n j^2 = \sum_{d \mid n} d^2 S \left(\frac{n}{d}\right)$)

Let S ( n ) denote the sum of the squares of the positive integers n and prime to n . Prove that

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

Answers

Proof. Let

S ( n ) = j [ [ 1 , n ] ] , j n = 1 j 2

denote the sum of the squares of the positive integers j n and prime to n .

If d is some divisor of n , we define the set

A d = { j [ [ 1 , n ] ] j n = d } .

Since

[ [ 1 , n ] ] = d n A d (disjoint sum) ,

we obtain

j [ [ 1 , n ] ] j 2 = d n j A d j 2 ,

that is

j = 1 n j 2 = d n j [ [ 1 , n ] ] , j n = d j 2 .

Put m = n d . If j n = d , then there is a unique integer i [ [ 1 , n d ] ] such that j = di , n = dm and i m = i m 2 = 1 . Conversely, if i ( n d ) = 1 , then j = di satisfies j n = d . Therefore, the change of index j = di gives

j = 1 n j 2 = d n j [ [ 1 , n ] ] , j n = d j 2 = d n d 2 i [ [ 1 , n d ] ] , i ( n d ) = 1 i 2 = d n d 2 S ( n d ) .

We have proved

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

We know that φ { { d : d n } { d : d n } d n d . is bijective (see Problem 6), therefore the change of index δ = n d gives

d n d 2 S ( n d ) = δ n n 2 δ 2 S ( δ ) ,

In conclusion,

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

User profile picture
2025-01-27 10:42
Comments