I'm having difficulty with Theorem 3.c from this paper.
I'll begin with some definitions for a finite $p$-group $G$. Define, for each integer $n$, $\mho_n(G) = \langle g^{p^n} \mid g \in G \rangle$ and call $G$ a $P_1$-group if $ \mho_n(S) = \{ s^{p^n} \mid s \in S \}$ for all $n$ and every section $S$ of $G$, including $S=G$.
Now suppose $G$ is a minimal non-$P_1$-group, i.e. $G$ is not a $P_1$-group but every proper section of $G$ is. Let $M$ be maximal subgroup of $G$.
Why is $|\mho_1(M)| \leq p$?
It has previously been shown:
- $G$ has exponent $p^2$ and is generated by two elements.
- $\Phi(G)$ has exponent $p$.
- If $H$ is a $P_1$-group with a subgroup $J \leq H$ such that $\exp J = p^n$ and $[H:J]=p^k$, then $|\mho_n(H)| \leq p^k$.
The paper states the result follows from 3, but how do we know $M$ contains a subgroup of exponent $p$ and index $p$?
The inequalites I can see are, $|\mho_1(M)| \leq [M:\Phi(G)]$, and $|\mho_{\log_p(\exp N)}(M)| \leq p$ and $|\mho_1(M)/\mho_1(N)| \leq p $ where $N$ is a maximal subgroup of $M$.
Reference: Mann, A. (1976). The power structure of p-groups. I. Journal of Algebra, 42(1), pp.121-135.
Since $G$ is $2$-generated, $|G:\Phi(G)| = p^2$, so $M$ has $\Phi(P)$ as a subgroup of index $p$ and exponent $p$.