I am trying to give a proof of the conjugacy of the Sylow system.

One of the alternative descriptions of the Hall's theorem is that

$G$ is solvable if and only if G has a Hall $p_i'$-subgroup $H_i$ for every prime divisor of $|G|$. ($|G|=\prod_{i=1}^n p_i^{a_i} $)

Now suppose that $G$ is solvable. As $H_1 \cap H_2 \dots H_{i-1} \cap H_{i+1}\dots \cap H_n = P_i$ is a Sylow $p_i$-subgroup of $G$, we can then obtain a Sylow system $S=\{ P_1,P_2 \dots P_n \}$ generated by $\{ H_1, H_2\dots H_n \}$. We have already known that all Hall $\pi$-subgroups are conjugate, hence we have another Sylow system $\{ Q_1,Q_2 \dots Q_n \}$ generated by $\{ {H_1}^{x_1}, {H_2}^{x_2}\dots {H_n}^{x_n} \}$. And I want to prove that these two Sylow systems are conjugate.

My attempt:

First, we can simplify this problem by assuming that $\{ Q_1,Q_2 \dots Q_n \}$ generated by $\{ {H_1}^{x_1}, {H_2}\dots {H_n} \}$ for some $x_1 \in G$. (In fact, if the proof is complete under this circumstance, then we can prove that Sylow systems generated by $\{ {H_1}^{x_1}, {H_2}\dots {H_n} \}$ and $\{ {H_1}^{x_1}, {H_2}^{x_2}, {H_3}\dots {H_n} \}$ are conjugate. Then the proof is complete by repeating this progress.)

I want to prove this by induction on $n$. By the argument above, we have $P_1=Q_1$. Since $G=P_1P_2\dots P_n=Q_1Q_2\dots Q_n$, we have $H=P_2P_3\dots P_n=Q_2Q_3\dots Q_n \le G$. By induction, $(P_2, P_3 \dots P_n)=({Q_2}^x, {Q_3}^x \dots {Q_n}^x)$ for some $x \in H$. So it suffices to show that $x \in N_G(P_1)$. However, I failed to do it. And I'm starting to think that $x$ is not necessarily belongs to $N_G(P_1)$. Can anyone enlighten me with what I'm missing? Or my way isn't gonna work? Then how to prove this? Thanks in advance.


I think it might be better to use induction on $|G|$. Let $N$ be a minimal normal subgroup of $G$. Then $N$ is an elementary abelian $p$-group for some prime $p$, and we may assume that $p=p_1$.

By induction, the Sylow systems are conjugate in $G/N$. So by replacing $Q_i$ by $Q_i^g$ for the conjugating element $g$, we can assume that $Q_iN=P_iN$ for all $i$, which means that $P_1=Q_1$, since $N \le P_1 \cap Q_1$.

Now $P_2 \cdots P_n$ and $Q_2 \cdots Q_n$ are Hall subgroups of $P_2 \cdots P_nN$, so they are conjugate in $P_1 \cdots P_nN$, and in fact they are conjugate by an element of $N$ (alternatively, that follows from the Schur-Zassenhaus Theorem). Conjugating by an element of $N$ does not change $P_1$, so we can now assume in addition that $P_2 \cdots P_n = Q_2 \cdots Q_n$, and we still have $P_1=Q_1$.

Now $P_2 \cdots P_n = Q_2 \cdots Q_n$ together with $P_iN=Q_iN$ implies $P_i = Q_i$ for $2 \le i \le n$, so we are done.

  • $\begingroup$ Thanks a lot! I have tried this way by considering $G/O_p(G)$ for some prime divisor of $|G|$. And I stuck where $P_2P_3...P_n O_p(G)=Q_2Q_3...Q_n O_p(G)$. Enlightened by your answer, I realized that I stuck there because I forgot $O_p(G) \le P_1 \cap Q_1$. $\endgroup$ – Johnny May 8 '17 at 17:21
  • $\begingroup$ A reason why I gave up using induction on $|G|$ was that I thought some information of Hall $p'$-subgroups should be used and is hard to use it in this proof. So is there any relevance between the conjugacy of Hall $p'$-subgroups and the conjugacy of the Sylow system? Like the elements satisfy some conditions? $\endgroup$ – Johnny May 8 '17 at 17:31
  • $\begingroup$ Sorry this isn;t really an area I know much about, so I don''t have any deeper insight into what is going on. But (as you probably noticed) the argument in my answer would work equally well with $O_p(G)$ in place of $N$. $\endgroup$ – Derek Holt May 8 '17 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.