Mathematics Stack Exchange is a question and answer site for people studying math at any level and professionals in related fields. It's 100% free, no registration required.

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

I have seen a proof for FLT, $n=3$ using factorisation in the ring of Eisenstein integers, but it's quite long and convoluted; I am wondering if there is a more 'advanced' proof which avoids infinite descent/messy repeated calculations.

What's your favorite proof of Fermat's last theorem for $n=3$, and where can it be found written down?

share|cite|improve this question
I don't think I've ever seen a proof that didn't use factorization in the Eisenstein integers. Mess is in the eyes of the beholder. – Gerry Myerson Nov 15 '12 at 5:24
Suppose $a^3 + b^3 = c^3$ where $a, b, c$ are nonzero integers. Let $x = 4bc/a^3$ and $y = 4(a^3+2b^3)/a^3$. Then $y^2 = x^3 + 16$, which is the equation for an elliptic curve. If you could show the only rational points on this curve (besides the point at infinity) are $(0,\pm 4)$, then we get a contradiction since from our example $x \not= 0$. That would settle FLT for $n = 3$. – KCd Nov 18 '12 at 9:04

Recently I have found two simple proofs of FLT for n=3. They use the ring $\mathbb{Z}[\sqrt[3]{2}]$. If $x^3+y^3+z^3=0$ then

(1) $x^6-4 (yz)^3=a^2$,

where $a=y^3-z^3$.

The idea is to deduce from (1) that the algebraic number $x^2-\sqrt[3]{4} yz$ is a square, more precisely:

(2) $x^2-\sqrt[3]{4} yz=u \beta^2$,

where $\beta \in \mathbb{Z}[\sqrt[3]{2}]$, and $u$ is a unit.

Once this is established, then it is not difficult to get a contradiction via infinite descent. It is similar to Euler's proof, but another number field is used.

The (2) follows from the unique factorization in the Euclidean ring $\mathbb{Z}[\sqrt[3]{2}]$

My second proof deduces (2) without using the unique factorization.

The idea is simple: find a number $\alpha \in \mathbb{Z}[\sqrt[3]{2}]$, such that $\alpha^2$ is divided by the $x^2-\sqrt[3]{4} yz$ and $\alpha^2/(x^2-\sqrt[3]{4} yz)$ has a small norm. It turns out that the last number may have small norm indeed: less than 9. Then the norm of it can be either 1 or 4, and (2) easily follows. The method can be used to prove FLT for other concrete $n$ as well.

It is easy to get such $\alpha$ because $\alpha$ belongs to $I$, where $I$ is the ideal of norm $a$, generated by $a$ and $x^2-\sqrt[3]{4} yz$. We know from the theory of algebraic numbers, that there is an $\alpha \in I$, such that $N(\alpha)<m N(I)$, where $m$ is easily calculated constant, that turns to be less than 3. Then $N(\alpha^2/(x^2-\sqrt[3]{4} yz))<(3 a)^2/a^2=9$.

This proof is shorter and easier than that of Euler and Kummer.

Now I'll explain how to get contradiction from (2). We may assume, that $x$ is odd, otherwize take $y$ to play the role of $x$.

Units in the ring $\mathbb{Z}[\sqrt[3]{2}]$ have the form: $u=\pm(\sqrt[3]{2}-1)^k$, where $k$ is integer. Therefore, from (2) follows, that either

(3) $x^2−\sqrt[3]{4} yz=\beta_1^2$


(4) $(\sqrt[3]{2}-1)(x^2−\sqrt[3]{4} yz)=\beta_2^2$,

where $\beta_1$ and $\beta_2$ belong to the ring $\mathbb{Z}[\sqrt[3]{2}]$.

However (4) is impossible, because the coefficient at $\sqrt[3]{2}$ turns out to be odd in the left part and even in the right part of (4). Therefore (3) takes place, that is:

(5) $x^2−\sqrt[3]{4} yz=(a_0+a_1 \sqrt[3]{2}+a_2 (\sqrt[3]{2})^2)^2$, where $a_0$, $a_1$ and $a_2$ are integers.

From (5) we get:

(6) $a_0^2+4 a_1 a_2=x^2$

(7) $2 a_2^2+2 a_0 a_1=0$

(8) $a_1^2+2 a_0 a_2=-yz$

It follows, that all three numbers $a_0$, $a_1$ and $a_2$ are not zero. We may assume, that $a_2>0$, otherwise change sign of all three numbers.

It follows from (7) that $a_0=\pm m b_0^2$, $a_1=\mp m b_1^2$, $a_2=m b_0 b_1$, where $b_0$ and $b_1$ are relatively prime positive integers and $m$ is a positive integer.

Now it follows from (6) that:

(9) $m^2 b_0 (b_0^3 \mp 4 b_1^3)=x^2$

From (9) it follows that $b_0$ is a square, that is $b_0=x_1^2$, and $(b_0^3 \mp 4 b_1^3)$ is a square, that is:

(10) $x_1^6 \mp 4 b_1^3$ is a square

It can be shown, that $x_1^6 \mp 4 b_1^3 \neq 1$ (a separate lemma). Then from (9) it follows, that $b_0<x^2$, that is $x_1<x$ and we've got the infinite descent.

share|cite|improve this answer

Proof based on method of infinite descent: Method of infinite descent and proof of Fermat's last theorem for $n=3$. Canadian Journal on Computing in Mathematics, Natural Sciences, Engineering and Medicine (CMNSEM), Vol.1, No.6, Septermber 2010, pp. 181-186.

share|cite|improve this answer
I'd be careful about anything by that author and/or in that publication. See, for example,… – Gerry Myerson Oct 13 '13 at 0:34

protected by Community Oct 27 '13 at 1:57

Thank you for your interest in this question. Because it has attracted low-quality or spam answers that had to be removed, posting an answer now requires 10 reputation on this site.

Would you like to answer one of these unanswered questions instead?

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