# Using Fermat's Little Theorem Prove if $p$ is prime, prove $1^p + 2^p + 3^p +…+(p-1)^p \equiv 0 \bmod{p}$

Using Fermat's Theorem Prove if $p$ is prime, prove $1^p + 2^p + 3^p +...+(p-1)^p \equiv 0 \bmod{p}$

The two definitions of Fermat's Little Theorem is $a^p \equiv a \bmod{p}$ and $a^{p-1} \equiv 1 \bmod{p}$ but I don't know how to use this solve the problem

-

Hint: Use Fermat's theorem for each $k$ ($1\le k\le p-1$) and add all them up. Then use Gauss's formula for the sum $1+2+\cdots +(p-1)$.

-

$1^p + ... + (p-1)^p \equiv 1 + 2 + ... (p-1) \equiv (1 + (p-1)) + (2 + (p-2)) + ... +(\frac{p-1}{2} + \frac{p+1}{2}) \equiv p + ... + p \equiv 0 (\mod p)$

Edit: Assuming p is an odd prime, of course

-
I understand the 1+(p-1)+2+(p-2)+.... part but how did you come up with (p-1/2 + p+1/2)? –  Cindy Dec 10 '12 at 4:54
Because $p$ is odd, that'll be the last two pairs in the sum. Write out some of these sums for various primes and you'll see where it comes from. –  anonymous Dec 10 '12 at 16:39
Since $1^p \equiv 1$, and $2^p \equiv 2$, $\ldots$ $(p-1)^p \equiv p-1$, we have that
$$1^p + 2^p + \ldots (p-1)^p \equiv 1 + 2 + \ldots + p-1$$
However, we know this sum: $\sum_{i = 1}^{p-1} i = \frac{p (p-1)}{2}$, which for odd primes would be a multiple of $p$, and we are done by reducing modulo $p$, then we only must check the case for $2$, which doesn't hold, and it's awkward...