A finite group $G$ satisfies property $P_n$ if for every prime integer $p$, $G$ has at most $(n−1)$ non-central conjugacy classes the order of the representative element of which is a multiple of $p$. Let for $x\in G$, $x^G$ be a conjugacy classe of $G$. Also Let $G$ be a solvable group satisfying property $P_5$, $M=G^{'}Z(G)$, $|\frac{G}{M}|=2$ and $G-M=x^G\cup y^G\cup z^G$, where ‎$|C_G(x)|=|C_G(y)|=|C_G(z)|=6$‎, ‎$o(x)=2$‎, ‎$o(y)=6$ and $z=y^{-1}$‎.‎ Also ‎$‎M‎$‎ is of odd order and ‎$‎M‎$‎ has a normal and abelian 3-complement ($G^{'}$ is the commutator subgroup and $Z(G)$ is the center of $G$). Can we say that $Z(G)=1$?

  • 2
    $\begingroup$ Perhaps you could give some motivation as to why this is interesting to you. $\endgroup$ – Chris Ramsey Feb 29 '16 at 15:33
  • 1
    $\begingroup$ I edited tags. As a rule, you should always have some "arXiv-tag", if possible, in this case gr.group-theory. $\endgroup$ – Frieder Ladisch Feb 29 '16 at 16:45
  • 2
    $\begingroup$ Cross-posted from MathSE: math.stackexchange.com/questions/1677046 $\endgroup$ – YCor Feb 29 '16 at 17:51

There are no finite groups $G$ such that all your assumptions are met (so in a sense, the answer is "Yes").

Let $A$ be the normal $3$-complement of $G$, so that $M/A$ is a $3$-group. I claim that $|M/A| = 3$. Otherwise, the number of conjugacy classes of $M/A$ is at least $9$. Since $|C_G(x)|=6$, we have $|C_M(x)|=3$ and thus $M/A$ has at least $6$ conjugacy classes which lift to conjugacy classes not contained in $Z(G)$. Since $|G/M|=2$, the subgroup $M$ contains at least $3$ non-central conjugacy classes of $G$, such that its elements have order divisible by $3$. Since also $y^G$ and $z^G$ are such conjugacy classes, this contradicts property $P_5$.

Thus $|M/A| = 3$, and $G = \langle y \rangle A$. Therefore, $G' \subseteq A$ and we must have $y^2 \in Z(G)$ (because otherwise $M\subseteq A$.) But this also leads to a contradiction: For each $a\in A\setminus 1$ we get two further non-central conjugacy classes of order-$3$-elements, namely $(y^2a)^G = \{y^2a, y^2a^x\}$ and $(y^4a)^G=\{y^4a,y^4a^x\}$. So if $A\neq 1$, then $A\setminus 1 = \{a, a^x\}$ and $|A| = 3$, contradiction. If $A=1$, then $G$ is cyclic and $G=M$, contradiction.

  • $\begingroup$ In line 3, why does $M/A$ have at least 9 conjugacy classes? Is $M/A$ abelian? In line 4, why does $M/A$ have at least 6 non-central conjugacy classes? Also in line 8, why is $G/A$ abelian? $\endgroup$ – M. R Mar 4 '16 at 15:21
  • $\begingroup$ @MehdiRezaei: line 3: $M/A$ is a $3$-group, and for a $p$-group $P$, we always have $|P/P'| \geq p^2$. And we always have $x^P \subseteq xP'$. line 4: By your assumptions, $|Z(G)\cap M| \leq 3$. line 8: You assumed that $y$ has order $6$, and we now have $|G/A|=6$, and $|A|$ is not divisible by $2$ or $3$. Thus $G=\langle y \rangle A$ and $G/A = \langle yA \rangle$ is cyclic. $\endgroup$ – Frieder Ladisch Mar 4 '16 at 16:43

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.