The Diophantine equation \(x^ 2\pm \ell y^ 2=z^{\ell}\) connected with Fermat's Last Theorem (Q1107561): Difference between revisions

Let \(\ell\) be an odd prime number and put \(\ell^*=(-1)^{(\ell - 1)/2}\ell\). Fermat's Last Theorem was proved by Euler for the exponent \(\ell =3\) and by Dirichlet for the exponent \(\ell =5\). Their proofs, which are reproduced here in modern terms, are based on the fact that a certain implication is justified for \(\ell =3\) or \(\ell =5\). It is often said that their success is due to the unique factorization property in the maximal order of the quadratic field \({\mathbb Q}(\sqrt{\ell^*})\) for \(\ell =3\) or \(\ell =5\), respectively. But this point of view is not exact; for the above implication is true for virtually any prime \(\ell\). The examples given here will show that the difficulty lies in finding the step of ``infinite descent,'' not in the failure of unique factorization.