22
$\begingroup$

A friend asked me this question:

If $p$ is a prime, prove that $(p - 1)! + 1$ is a power of $p$ if and only if $p = 2, 3$ or $5$.

Clearly one direction is obvious, namely that $p=2,3,5$ implies $(p - 1)! + 1$ is a power of $p$.

The other direction is not clear to me. Since by Wilson's theorem $p$ divides $(p - 1)! + 1$ so we need to show that if there are no other prime factors then $p=2,3,5$. Can someone give me a hint for establishing this?

Thanks

$\endgroup$

2 Answers 2

Reset to default
26
$\begingroup$

Let $p^k - 1 = (p-1)!$. Cancel the factor of $(p-1)$, we get $$p^{k-1} + \cdots + 1 = (p-2)!$$ Now consider mod $p-1$ to get information about $k$. When can the equality ever hold?

$\endgroup$
2
  • $\begingroup$ Can you be a little more explicit? What do you mean by considering mod $p-1$ to get information about $k$? $\endgroup$
    – user10575
    Jan 27, 2013 at 13:36
  • 2
    $\begingroup$ @shahab, show that when $p > 5$, left hand side = k mod $p-1$ while RHS = 0 mod $p-1$. Show that this implies $k \ge p-1$. $\endgroup$
    – user27126
    Jan 27, 2013 at 19:54
5
$\begingroup$

Not sure if only one hint would help, you could try using more hints as needed. First, the statement holds for all natural numbers $n \notin \{2, 3, 5\} $, not just for primes. One way to proceed is to show that for $n > 5$:
1) $n$ cannot be even if it satisfies the equation
2) for $n$ odd, $(n-1)^2 \space | \space (n-1)!$
3) from (2) above and the given equation, show $(n-1) \space | \space k$ which implies $n-1 \le k$
4) show that $\forall n > 2, (n-1)! < n^{n-1} - 1$

Hope that helps!

$\endgroup$

You must log in to answer this question.