1
$\begingroup$

For which primes $p\not=2$ is $5$ a square mod $p$?


Using the Legendre symbol, $5$ is a square modulo $p$ if

$$\left(\frac{5}{p}\right)=5^{\dfrac{p-1}{2}} \equiv 1 \pmod{p}$$

Now we have

$$5^{\dfrac{p-1}{2}} = (5^{p-1})^{1/2} \equiv 1\pmod{p}$$

So $5$ is a square for any $p$. But this doesn't seem correct to me. Can anyone check this and if it's wrong, explain why.

$\endgroup$
  • $\begingroup$ What is the $1/2$-power supposed to mean ? $\endgroup$ – Captain Lama Apr 28 '16 at 17:58
  • $\begingroup$ Can you use Quadratic Reciprocity? $\endgroup$ – André Nicolas Apr 28 '16 at 18:02
  • $\begingroup$ @AndréNicolas Yes I can use the three laws. $\endgroup$ – user2850514 Apr 28 '16 at 18:09
2
$\begingroup$

You can't take the square root like this. By Fermat's little theorem, we know that for $p$ prime we have that $5^{p-1} \equiv 1 \mod p$. This means that $p \mid 5^{p-1} -1$, and hence $$p \mid \left(5^{\tfrac{p-1}{2}} -1\right)\left(5^{\tfrac{p-1}{2}} +1\right)$$

Therefore $5^{\tfrac{p-1}{2}} \equiv 1 \mod p$ or $5^{\tfrac{p-1}{2}} \equiv -1 \mod p$.

Note that if $p$ is not prime, there can be even more possibilities.


Note that for every odd prime $p$ we have $2 \mid p-1$. Also $4 \mid 5-1$. So $2 \mid \left(\frac{p-1}{2} \cdot \frac{5-1}{2} \right)$. Hence by quadratic reciprocity we have $$\left( \frac{p}{5} \right)\left( \frac{5}{p} \right) = (-1)^{\frac{p-1}{2} \cdot \frac{5-1}{2}} = 1$$

So $\left( \frac{p}{5} \right) =1$ if $\left( \frac{5}{p} \right)=1$, and the latter happens when $p\equiv 1 \mod 5$ or $p \equiv 4 \mod 5$.

Note that one can not use the law if $p=5$, since the law of quadratic reciprocity requires $p$ and $q$ to be different primes. Whether $p=5$ is a quadratic residue depends on the definition of quadratic residue.

$\endgroup$
  • $\begingroup$ Am I correct in thinking $(p/5)=1\iff (5/p)=1$ since if $(5/p)=-1$ then their product would be $-1$? $\endgroup$ – user2850514 Apr 28 '16 at 18:25
  • 1
    $\begingroup$ @user2850514 Yes. $\endgroup$ – wythagoras Apr 28 '16 at 18:32
  • $\begingroup$ Please forgive this question as I am a rank amateur. I was wondering about your line "the latter happens." Might it not be that (5/p)=1 if (p/5)=1 and that happens if $p\equiv 1,4 \pmod 5$. Thanks and regards, $\endgroup$ – user12802 May 22 '16 at 12:49
  • $\begingroup$ @Andrew I'm sorry, but I don't understand your question. The latter means $(5/p)=1$. $(5/p)=1$ if and only if $p \equiv 1, 4 \mod 5$. $\endgroup$ – wythagoras May 22 '16 at 12:52
  • $\begingroup$ I was thinking that first one shows that (p/5) =1 for $p\equiv 1,4\pmod 5$ and then it follows that (5/p) =1 subsequently holds for $p\equiv 1,4\pmod 5$? $\endgroup$ – user12802 May 22 '16 at 13:08
3
$\begingroup$

The number $5$ is a square modulo the prime $5$, though I would not call it a quadratic residue of $5$. Now assume that $p$ is an odd prime other than $5$.

By Quadratic Reciprocity, since $5$ is of the form $4k+1$, we have $(5/p)=(p/5)$.

Note that $(p/5)=(r/5)$, where $r$ is the remainder when we divide $p$ by $5$.

It is easy to check that $1$ and $4$ are quadratic residues of $5$, and $2$ and $3$ are not.

Thus $(5/p)=1$ if and only if the odd prime $p$ is congruent to $1$ or $4$ modulo $5$. Equivalently, $(5/p)=1$ if and only if the rightmost decimal digit of the prime $p$ is $1$ or $9$.

$\endgroup$
  • $\begingroup$ +1, but you forgot $p=5$ itself. $\endgroup$ – wythagoras Apr 28 '16 at 18:20
  • $\begingroup$ @wythagoras: I dealt separately with $5$ in the first paragraph. $\endgroup$ – André Nicolas Apr 28 '16 at 18:22
  • $\begingroup$ Yes, I see. But why wouldn't you call $5$ a quadratic residue of $5$? It satisfies the definition. $\endgroup$ – wythagoras Apr 28 '16 at 18:24
  • $\begingroup$ @wythagoras: There are marginally different definitions of quadratic residue modulo $p$. The standard textbooks I have used exclude $0$. That way, for odd primes $p$ there are exactly as many quadratic residues as non-residues, and the product of a residue and a non-residue is a non-residue. (That result would fail if we allow $0$ as a QR.) $\endgroup$ – André Nicolas Apr 28 '16 at 18:28

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

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