I've been going through the Wikipedia article on cyclotomic polynomials, trying to prove the various identities that can be used as tools to compute $\Phi_n(x)$.
One such identity that I'm not sure how to prove is the following, with $p$ prime and $p \nmid n$:
$\Phi_{np}(x) = \Phi_n(x^p)/\Phi_n(x)$.
A proof is given in Trygve Nagell's Introduction to Number Theory, but it is difficult to follow.
How might I go about proving this?
We need to prove that $\Phi_n(x)\Phi_{np}(x)=\Phi_n(x^p)$. The left side is squarefree and its zeros are the primitive $n$-th and $np$-th roots of unity. As $p\nmid n$ then $\zeta^p$ is a primitive $n$-th root of unity if $\zeta$ is, and it is clear that $\zeta^p$ is a primitive $n$-th root of unity if $\zeta$ is a primitive $np$-th root of unity. Conversely if $\zeta^p$ is a primitive $n$-the root of unity, then $\zeta$ has multiplicative order $n$ or $np$.
Summarising, the zeros of $\Phi_n(x)\Phi_{np}(x)$ are the $p$-th roots of the primitive $n$-th roots of unity. But these are precisely the zeros of $\Phi_n(x^p)$. As this is monic, with the same degree as $\Phi_n(x)\Phi_{np}(x)$ (why?) then these polynomials are equal.