In M. Ram Murty's paper Artin's Conjecture on Primitive Roots I am not able to understand a statement.
Let $k$ be the order of $a \bmod p$ where $p$ is a prime and also we have $a^{p-1} \equiv 1 \mod p$ then $k \mid (p-1)$ if $k \neq p-1$ then $k\mid(p-1)/q$ for some prime divisor $q$ of $p-1$.
I am not able to understand how the author was able to make the second claim i.e if $k \neq p-1$ then $k\mid(p-1)/q$ for some prime divisor $q$ of $p-1$.