Mathematics Stack Exchange is a question and answer site for people studying math at any level and professionals in related fields. Join them; it only takes a minute:

Sign up
Here's how it works:
  1. Anybody can ask a question
  2. Anybody can answer
  3. The best answers are voted up and rise to the top

For odd $n$ the Fermat equation $x^n + y^n = z^n$ factorises as $$(x + y)(x + \zeta y) \cdots (x + \zeta^{n-1}y) = z^n,$$

where $\zeta = e^{2 \pi i/n}$. I tried seeing this was true by multiplying the factors out by hand but got into a mess. Is there a neat way of showing this (e.g. is there a useful identity that can be used involving products of the $\zeta^i$ etc.?

share|cite|improve this question
up vote 5 down vote accepted

We will make use of the following theorem

Theorem Let $P(x)$ be a polynomial and $P(z)=0$, then $x-z$ divides $P(x)$.

Let $n$ be odd and write $$P(x) = x^n + y^n,$$ then $P(-\zeta^r y)=0$ for all $n$ powers of $r$.

Therefore $(x+y)(x+\zeta y)(x+\zeta^2 y)\cdots$ divides $P(x)$ but it has the same degree so its equal (up to a constant factor which is easily seen to be 1).

share|cite|improve this answer

Fix $x \neq 0$. Let

$$P(y)=(x + y)(x + \zeta y) \cdots (x + \zeta^{n-1}y) -x^n-y^n$$

Then $P(-x)=P(-\zeta x)=P(-\zeta^2 x)=..=P(-\zeta^{n-1} x)=0$ and $P(0)=x^n-x^n=0$.

Thus $P$ is a polynomial of degree at most $n$ which has at least $n+1$ roots. This proves that

$$(x + y)(x + \zeta y) \cdots (x + \zeta^{n-1}y) =x^n+y^n$$

for all $y$ and all $x \neq 0$.

The case $x=0$ is trivial.

share|cite|improve this answer
Rats, this is beautiful! +1 – DonAntonio May 2 '13 at 17:39

Let $n$ be odd. The $n$ distinct roots of the polynomial $X^n-1$ are $1,\zeta,\dots,\zeta^{n-1}$ $$ X^n-1=(X-1)(X-\zeta)\cdots(X-\zeta^{n-1}) $$ Apply the transformation $X\mapsto -\frac{x}{y}$ $$ \left(-\frac{x}{y}\right)^n-1=\left(-\frac{x}{y}-1\right)\left(-\frac{x}{y}-\zeta\right)\cdots\left(-\frac{x}{y}-\zeta^{n-1}\right) $$ Finally, multiply through by $-y^n$ $$ x^n+y^n=(x+y)\cdots(x+\zeta^{n-1}y) $$

share|cite|improve this answer
You have to multiply it with $(-y)^n $ and not $-y^n $ to get the RHS, but then your LHS is $ x^n-(-y)^n $, so this can't be done unless $n$ is odd. – smiley06 May 2 '13 at 15:42
@smiley06 Yes, I assumed that the OP knew this identity only held for old $n$ (as stated in the question), so they are the same thing. – Warren Moore May 2 '13 at 15:52
I like your way very much! +3 :) – neofoxmulder Mar 1 '14 at 18:36

Your Answer


By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.