If $G$ is finite with a unique maximal subgroup, then $G$ is cyclic.

If $G$ is infinite, not. The example I'm studying is the following: let $Z_{p^{\infty}}$ be the Prufer $p$-group and $C_p$ the cyclic group of order $p$ (obviously $p$ is prime).

Consider now $\Gamma:=Z_{p^{\infty}}\times C_p$. This new group is infinite and not cyclic, in which $M:=Z_{p^{\infty}}\times1$ is clearly maximal.

In order to show that it's unique, we argue by contradiction and suppose that there exists $M'\neq M$ another maximal subgroup of $\Gamma$.

Then my teacher wrote that $\Gamma/M'$ has prime order and $M\cap M'$ is finite: I can't understand why, and I have no idea where to start.

Moreover I'm asking myself if is this true in general for infinite groups.

Thank you all.

  • 1
    $\begingroup$ Sea turtles answer tells why it works. If G is not abelian, indeed if it is not nilpotent, then maximal subgroups need not have prime index. Examples of nonnilpotent p groups are tarski monsters and McClain char simple p groups. $\endgroup$ May 1, 2014 at 23:32

1 Answer 1


Clearly $\Gamma/M'$ is abelian, since $\Gamma$ is abelian. As $M'$ is assumed maximal, $\Gamma/M'$ is simple. Abelian and simple implies cyclic of prime order. (Proof. Good exercise. Hint: lattice correspondence.)

Since $M$ and $M'$ have finite index, their intersection does as well. (Proof. This holds generically, not just in this situation. Let $H$ and $K$ be finite index subgroups of some $G$. The Cartesian product of coset spaces $G/H\times G/K$ is finite and also equipped with a left $G$-action, so the orbit of $H\times K$ is also finite, so its stabilizer $H\cap K$ must have finite index by the orbit-stabilizer theorem.)

This reasoning applies to any abelian group: maximal subgroups, if they exist, must have prime index. Maximal subgroups need not exist (just look at the Prufer $p$-group) and need not be unique, for instance ${\bf Z}_p\times{\bf Z}_p$ (we use $p$-adic integers to get an infinite group, even though this example is perfectly analogous to $C_p\times C_p$). I don't think anything is different for nonabelian groups.

  • $\begingroup$ Thanks a lot! $M\cap M'$ has finite index and I agree. However I asked why it has finite order too. I'm arguing about this, but I can't see the light! $\endgroup$
    – Joe
    May 2, 2014 at 17:26
  • $\begingroup$ @Joe $M$ is the only proper subgroup of $\Gamma$ that isn't finite. How to prove this? In any direct product $G=H\times K$, if $\pi_H$ and $\pi_K$ are the projection maps and $M\le G$ is any subgroup, we have the containment $M\le \pi_H(M)\times\pi_K(M)$. The Prufer $p$-group's proper subgroups are all cyclic of prime power order, in particular they are finite. $\endgroup$
    – blue
    May 2, 2014 at 17:34
  • $\begingroup$ Thanks a lot! The finitness of $M\cap M'$ is trivial, I've seen this after your comment answer. However your answer was really clear and useful, thank you! :-) $\endgroup$
    – Joe
    May 2, 2014 at 19:18

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.