I was trying to prove that if $p<q$ are two prime numbers, $n$ is a positive integer and $G$ is a group of order $p^nq$, then $G$ is not simple.

What I know so far is that, if I assume $G$ simple, $n_p = q \equiv 1 \pmod p$, $n_q= p^k \equiv 1 \pmod q$, for some positive $k$ (by Sylow's theorems), and that there is an element $g$ in a $p$-Sylow whose coniugacy class $g^G$ has size $q$ (by taking a nonidentical element in the center of a $p$-Sylow). Now I'm stuck. How can I go on? Is there anything particular about coniugacy classes of prime size?


Let $G$ act on the Sylow $p$-subgroups (of which there are $q$) via conjugation. Because we assume $G$ is simple, this action embeds $G$ into $S_q$. In $S_q$, the elements of order $q$ are self-centralizing. Thus there are $p^n$ elements of order $q$ in $G$. A counting argument shows there is only one Sylow $p$-subgroup in $G$, which is a contradiction.

EDIT: Sorry, the argument above doesn't work. Here is one that does, from Isaacs's Finite Group Theory:

Let $P_n$, $1\le n\le q$ be the Sylow $p$-subgroups. Pick $K=P_i\cap P_j$ to be the intersection of two of these Sylow subgroups, such that $|K|$ is maximal. (Note a simple counting argument shows $|K|>1$.)

Now the normalizer $N_G(K)$ is such that $N_G(K)\cap P_i > K$ and $N_G(K)\cap P_j > K$, because normalizers grow in $p$-groups. By the maximality of $|K|$, $N_G(K)$ is not contained in a $p$-group. So there is a Sylow $q$-subgroup $Q$ contained in $N_G(K)$. Since $P_iQ=G$, we can write any element in $G$ as $hm$, with $h\in P_i$ and $m\in Q$. Then $P_i^{hm}=P_i^m\supset K^m=K$, since $Q\in N_G(K)$. So $K$ is contained in all Sylow $p$-subgroups, and is thus normal.

  • $\begingroup$ How do you go from the elements of order $q$ are self-centralizing to there are $p^n$ elements of order $q$ in $G$? $\endgroup$ – Human Oct 27 '16 at 7:09
  • $\begingroup$ @DerekHolt. But aren't the cycles of the same order conjugates?? $\endgroup$ – Human Oct 27 '16 at 9:14
  • $\begingroup$ Yes you are right, some of them are. So I don't think this argument works. I will delete my earlier comments. $\endgroup$ – Derek Holt Oct 27 '16 at 9:22

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.