14
$\begingroup$

For every odd prime $q \geq 3$, does there exist a prime $p < q$ and integers $x,y$ such that

$$\displaystyle x^2 + py^2 = q?$$

One can easily show that all primes $q \not \equiv -1 \pmod{3}$ can be written as such a sum. Indeed, if $q \equiv 1 \pmod{3}$ then $q$ is the norm of an element in the ring of Eisenstein integers, so there exist integers $x,y$ such that $q = x^2 + xy + y^2$. If $xy \equiv 0 \pmod{2}$, say $2 | y$, then we can write

$$\displaystyle q = \left(x + \frac{y}{2} \right)^2 + 3 \left(\frac{y}{2} \right)^2,$$

so we are done. Otherwise we must have $x,y$ are both odd. Now note that $q = x^2 + xy + y^2$ implies that

$$\displaystyle q = (-x - y)^2 + y(-x - y) + y^2,$$

and $x + y$ is even.

Playing similar games for other small primes of class number one gives that at least half of odd primes can be expressed in the form above.

$\endgroup$
15
  • 12
    $\begingroup$ $x = 0, y = 1, p = q$? $\endgroup$ Commented Jan 1, 2023 at 20:44
  • 3
    $\begingroup$ @QiaochuYuan yes that is a careless oversight. I added the restriction that $p < q$. $\endgroup$ Commented Jan 2, 2023 at 0:33
  • 3
    $\begingroup$ There is a book by D Cox quite relevant to your question... $\endgroup$ Commented Jan 2, 2023 at 0:42
  • 4
    $\begingroup$ For what it's worth, there are 15 primes up to $10^9$ which are not of the form $x^2+p$ (for a smaller prime $p$), the largest being $7549$. It's plausible that these examples run out (although it may be hard to prove), see also math.stackexchange.com/questions/3710032/… $\endgroup$ Commented Jan 2, 2023 at 4:52
  • 3
    $\begingroup$ "odd prime $q\ge 3$" is a nice example of pleonasm $\endgroup$ Commented Jan 2, 2023 at 8:53

0

You must log in to answer this question.