I asked a similar question yesterday. But I found out that I should make some major changes to it. So I decided to ask a new version of my previous question.

Prove or find a counterexample to the following statement.

Let $G$ be a finite group with at least two minimal subgroups and $I_1, I_2, ... ,I_n$ be a family of minimal subgroups of $G$. Suppose that $H_1, H_2, ... ,H_m$ are all the subgroups of $G$ such that for each $i,j$, $I_i\leq H_j$ and there is not any minimal subgroup other than $I_i$s contained in $H_j$ for all $1\leq j\leq m$ (i.e. $H_j$s have the same minimal subgroups and there is not any other subgroup such that it's minimal subgroups are as same as $H_j$s). Suppose also that every $H_j$ is a non-maximal subgroup. Then there exists at least one minimal subgroup say $L$ such that $L\nsubseteq H_1\cup H_2\cup ... \cup H_m$ and $\langle H_j,L\rangle \neq \langle H_k,L \rangle$ for all $j\neq k$. Furthermore $\langle H_1,L\rangle, \langle H_2,L\rangle, ... \langle H_m,L\rangle$ are proper subgroups of $G$ which have the same minimal subgroups.


1 Answer 1


There are so many conditions that it is hard to find examples satisfying the hypotheses, which makes it harder to find counterexamples, but I am sure they exist.

I believe there is a counterexample that is an extension of $P=C_{17^2} \times C_{17^2}$ by $C_9$ with fixed point free action of $C_9$ on $P$. For the $H_i$ take all subgroups of $P$ that contain $C_{17} \times C_{17}$. The only minimal subgroups of $G$ outside of $P$ are of order $3$, and there is a unique conjugacy class of these. Any such subgroup $L$ of order $3$ generates $\langle P,L \rangle$ with all $H_i$ except for $C_{17} \times C_{17}$.

  • $\begingroup$ Recall that I said "Suppose also that every $H_j$ is a non-maximal subgroup". But in your example the subgroup $P$ is a maximal subgroup and also contains all minimal subgroups of order 17 as same as other $H_j$s. Am I right? $\endgroup$ Sep 4, 2016 at 21:25
  • $\begingroup$ P is not a maximal subgroup. $\endgroup$ Sep 4, 2016 at 22:40
  • $\begingroup$ Index $G/P \cong C_9$, so $P$ is not maximal. The reason this example is a as large as it is, is that $G/P$ had to have non-prime order, and I couldn't find examples with $|G/P|$ even. But it is possible that there are smaller examples. $\endgroup$
    – Derek Holt
    Sep 5, 2016 at 7:33
  • $\begingroup$ Dear professors Holt and Robinson. Thank you for your helpful answers. $\endgroup$ Sep 8, 2016 at 11:11
  • $\begingroup$ If you believe that this is indeed a counterexample, then you should accept the answer. $\endgroup$
    – Derek Holt
    Sep 8, 2016 at 12:54

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

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