Let $N$ be a normal subgroup of $G$ where $G$ is finite group, then we have to prove $n_p(N)$ divides $n_p(G)$ ( here $n_p(G)$ means number of sylow $p$-subgroups of $G$)

I was able to prove that $n_p(N)$ $\leq$ $n_p(G)$ as by proving first that for any $P \in Syl_p(G)$ there is a conjugate $gPg^{-1}$ such that $gPg^{-1}\cap H \in Syl_p(H)$ for any subgroup $H$ of $G$ whether normal or not.( by considering action of $H$ on $G/P$ and then considering $stab_gP$

And by replacing $H$ by $N$ i.e. a normal subgroup I could show that for any $P \in Syl_p(G), P\cap N \in Syl_p(N)$ and all $p$-sylow subgroups of $N$ arise in this way.

It all follows from a paper of Conrad, but from here and this proof i can't seem to figure out how to show it will divide $n_p(G)$. Please help.

  • 1
    $\begingroup$ I can guess what $n_p(G)$ means, but is it really standard notation? $\endgroup$ – Derek Holt Sep 21 '14 at 18:04
  • $\begingroup$ It means number of sylow $p$-subgroups of $G$. I guess it is pretty much standard... but i am sorry i should have mentioned it. $\endgroup$ – Bhaskar Vashishth Sep 21 '14 at 19:36
  • $\begingroup$ @BhaskarVashishth, I am solving a similar question,though Sylow theory if very new to me. Can you explain further in you original post on how you proved $n_p(G) \le n_p(G)$? $\endgroup$ – cele Mar 25 '15 at 21:10

Theorem: Let $G$ act on the sets $\Omega_1,\Omega_2$ transitively, and assume that there is an onto function $\phi:\Omega_1\to \Omega_2$ such that $\phi(g.x)=g.\phi(x)$ for all $g \in G$ and $x \in \Omega_1$. Then $|\Omega_2|$ divides $|\Omega_1|$.

Proof: Let $\alpha\in \Omega_1$ and $g\in G_{\alpha}$ (stabilizer of $\alpha$). Since $$g\phi(\alpha)=\phi(g\alpha)=\phi(\alpha),$$ we have $G_{\alpha}\leq G_{\phi(\alpha)}$. Since $|\Omega_1|=|G:G_\alpha|$ and $|\Omega_2|=|G:G_{\phi(\alpha)}|$, the result follows from this.

Now, Set $\Omega_1=Syl_p(G)$ , $\Omega_2=Syl_p(N)$ and $\phi(P)=P\cap N$. Now, actions of $G$ on the sets by conjugation are transitive and $\phi$ has desired properties so we are done.

Note: If you do not want mention the theorem, you can directly use the idea in the proof to show the result.

  • 1
    $\begingroup$ very nice... thanks $\endgroup$ – Bhaskar Vashishth Sep 22 '14 at 15:24
  • $\begingroup$ @BhaskarVashishth: you are welcome. $\endgroup$ – mesel Sep 22 '14 at 15:25

@Bhaskar, I promised you a proof - did not find the time to write it down earlier. Although Mesel already gave you one, mine is of a different flavor and sticks more closely to Sylow theory. You already mentioned that you know that Sylow $p$-subgroups of a normal subgroup $N$ of a group $G$, are of the form $P \cap N$, where $P \in Syl_p(G)$. I will assume that you are familiar with the Frattini Argument and will also use the following easy lemma.

Lemma Let $G$ be a a group, $H \leq G$ and $N \trianglelefteq G$. Then for any $g \in G$ we have $(H \cap N)^g=H^g \cap N$.

Proof. Since $H \cap N \subseteq H$, $(H \cap N)^g \subseteq H^g$, and likewise $(H \cap N)^g \subseteq N^g=N$ (the latter equality follows from the fact that $N$ is normal. Hence $(H \cap N)^g \subseteq H^g \cap N$. Conversely, let $x \in H^g \cap N$, then $x \in N$ and $x=g^{-1}hg$ for some $h \in H$. Hence $h= gxg^{-1} \in N$, since $x \in N$ and $N$ is normal. So, $x=g^{-1}hg \in (H \cap N)^g$ and we are done.

Now let us get to the proof of your post.

Theorem Let $G$ be a group, $N \trianglelefteq G$, and $p$ a prime dividing the order of $G$. Then $$\#Syl_p(N) \mid\#Syl_p(G).$$

Proof. Let $P \in Syl_p(G)$, then $P \cap N \in Syl_p(N)$. Owing to the Frattini Argument we obtain $G=N_G(P\cap N)N$. Observe that $N \cap N_G(P \cap N)=N_N(P \cap N)$ and hence $$[G:N_G(P \cap N)]=[N_G(P \cap N)N:N_G(P \cap N)]=[N:N_N(P \cap N)]=\#Syl_p(N).$$ Finally let $g \in N_G(P)$, then the lemma gives $(P \cap N)^g=P^g \cap N =P \cap N.$ It follows that $N_G(P) \subseteq N_G(P \cap N)$, and thus $\#Syl_p(N)=[G:N_G(P \cap N)] \mid [G:N_G(P)]=\#Syl_p(G)$. $\square$

  • $\begingroup$ Thanks a lot, u remembered.i liked both approaches. $\endgroup$ – Bhaskar Vashishth Sep 29 '14 at 16:54

This is not so trivial to prove. Even more is true.

Theorem (M. Hall, 1967) Let $N \unlhd G$, $P \in Syl_p(G)$, then $n_p=a_pb_pc_p$, where
$a_p = \#Syl_p(G/N)$
$b_p=\#Syl_p(N)$ and
$c_p=\#Syl_p(N_{PN}(P\cap N)/(P \cap N))$.

  • $\begingroup$ I have this paper of Hall, but there must be an easier proof, my professor gave it as an assignment problem, and that too very first problem of assignment. So I am sure there is an easy proof. $\endgroup$ – Bhaskar Vashishth Sep 21 '14 at 20:29
  • $\begingroup$ Ah, good that you know this paper! And yes, I agree, there must be an easier proof (only the divisibility required), will think about it! $\endgroup$ – Nicky Hekster Sep 21 '14 at 20:35

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.