1

For a Diophantine equation $x^2+py^2=z^2$ where $p$ is a prime of the form $p\equiv 1(mod4)$ and $(x,y,z)=1$.

Show that $z$ is necessarily odd.

2 Answers2

2

Suppose to the contrary that $z$ is even.

If $y$ is even, then $x^2$ is even, so $x$ is even, contradicting the fact that $\gcd(x,y,z)=1$. Similarly, if $x$ is even, then $py^2$ is even, and since $p$ is odd $y^2$ is even, and therefore $y$ is even, and again $\gcd(x,y,z)\gt 1$.

So $x$ and $y$ are both odd. Thus $x^2\equiv 1\pmod{4}$ and $y^2\equiv 1\pmod{4}$, and therefore $x^2+py^2\equiv 2\pmod{4}$. This is impossible, since any even square is congruent to $0$ modulo $4$.

André Nicolas
  • 507,029
0

For the equation:

$$x^2+py^2=z^2$$

There is a solution.

$$x=t^2-pk^2$$

$$y=2tk$$

$$z=t^2+pk^2$$

Means it is necessary to rewrite the equation in the form. $p=4s+1$

$$x^2+(4s+1)y^2=z^2$$

This is possible if the number of $x,y - $ odd.

Means: $$z=\frac{t^2+(4s+1)k^2}{2}$$

and the numbers $t,k - $ always odd. Imagine the numbers.

$$t=2t+1$$

$$k=2k+1$$

We substitute in the formula and find out what the parity of the number $z$.

$$z=2s(2k+1)^2+2(t^2+k^2)+2(t+k)+1$$

This number is always odd! With the high value of the coefficient.

individ
  • 4,301
  • How do you know that $t$ and $k$ are integers? – barak manos Dec 24 '14 at 14:07
  • I do not see how this answers the question, which in its current form is false. – user133281 Dec 24 '14 at 14:07
  • Как только мы получим формулу, которая может дать вам пример: $1^2+73^2=8^2$ или еще. $29^2+733^2=92^2$ – individ Dec 24 '14 at 14:15
  • 1
    This works for all solutions of the form $x=t^2-pk^2$ etc, however, I fail to see where you show that there can be no other solutions. – Some Math Student Dec 24 '14 at 14:59
  • @SomeMathStudent All solutions of this equation are described by the formula which I have brought. A special case of this Pythagorean triples. You want to look at the formula in General? It is very long. Don't see the point of her show. – individ Dec 24 '14 at 15:03
  • @ individ Yes, I'm aware of that, and I know that proof. However, I'm not sure if the op is, and in that case this part might seem a bit mysterious. I believe a link to a proof or at the very least the statement that not there is a solution of this form, but all solutions are of this form would be helpful here. – Some Math Student Dec 24 '14 at 15:06
  • @SomeMathStudent for this equation this formula gives all solutions. Couldn't argue with that. If you want to see the whole formula it can be seen there. http://math.stackexchange.com/questions/738446/solutions-to-ax2-by2-cz2/738527#738527 – individ Dec 24 '14 at 15:12