In my response to the OP in Embedding $S_3$ into $Aut(F_2)$ (I continue notation from my response in that thread), I indicated the possibility that lifting elements from $GL_{2}(\mathbb{Z}/(p))$ to $Aut(B(2,p))$ could extend the ability of character-free methods in proving congruences relating the order of a group and its number of conjugacy classes. I can now report a small bit of concrete progress in this direction. Specifically, what I can prove is:

Theorem. If $G$ is a finite group of exponent 3, then $|G| \equiv c(G) \mod{16}$.

Proof. We regard the Burnside group $B(2,3) = < x,y|(*)^{3}=1 >$ (here $ * $ ranges over every word in the generators) as $< x,y>$ and specify elements of $Aut(B(2,3))$ by their action on $x$ and $y$. Specifically, consider the following elements of $Aut(B(2,3))$:

$\alpha: (x,y) \to (xy,x^{-1}y)$
$\beta: (x,y) \to (x,y^{-1})$

By checking that $\beta^{2} = 1$ and $\beta\alpha\beta = \alpha^{3}$, we see that $<\alpha,\beta>$ is a quotient of the semidihedral group of order $16$. By checking that $\alpha^{4}$ sends $(x,y)$ to $(x^{-1},xy^{-1}x^{-1})$ so that $\alpha^{4} \neq 1$, we see that $<\alpha,\beta>$ is semidihedral of order $16$.

Now let $<\alpha,\beta>$ act on non-commuting pairs of elements of our group $G$ of exponent 3. We wish to show that every non-identity element of $<\alpha,\beta>$ has no fixed points in its action on the set of non-commuting pairs of elements of $G$. Every non-identity element of $<\alpha,\beta>$ has a power equal to $\alpha^{4}$ or is conjugate to $\beta$, so it suffices to check this property for just $\alpha^{4}$ and $\beta$.

If $(x,y)$ is a fixed point of $\alpha$, then $x = x^{-1}$ so $x^{2} = 1$ and, since $x^{3} = 1$, $x = 1$ and $(x,y)$ is not a non-commuting pair.
If $(x,y)$ is a fixed point of $\beta$, then $y = y^{-1}$ so $y^{2} = 1$ and, since $y^{3} = 1$, $y = 1$ and $(x,y)$ is not a non-commuting pair.

Therefore every orbit for the action of $<\alpha,\beta>$ on the set of non-commuting pairs of elements of $G$ has exactly 16 elements and we conclude that 16 divides $|G|(|G| - c(G))$ . Since $|G|$ is odd, we conclude that $|G| \equiv c(G) \mod{16}$.

I do not know how to give a character-free proof of this congruence (either for general $p$-groups or just those of exponent $p$) for any prime $p$ with $p \equiv 3 \mod{4}$ and $p > 3$. Complicating hopes of extending this is the fact that $B(2,n)$ is not known to be finite for any larger odd value of $n$.

(i) I'm probably just not thinking clearly enough right now, but how does one use this to prove the congruence when $G$ is an arbitrary 3-group?
(ii) Is it hopeless to expect to extend this to larger primes congruent to 3 mod 4?

  • $\begingroup$ I believe $\alpha$ is not an automorphism! $\endgroup$
    – Steve D
    Oct 8, 2011 at 19:20
  • $\begingroup$ Thanks! I mistranscribed the formula for $\alpha$. Will edit now. $\endgroup$ Oct 8, 2011 at 23:24
  • $\begingroup$ When I first read this I missed the "character free" part, so I thought it's pretty simple. I only realized this after clicking on the linked question. :) (This explains why there is a deleted answer of mine pointing to the obvious proof.) $\endgroup$ Oct 22, 2011 at 1:34

3 Answers 3


Burnside's classical formula $|G| \equiv c(G) \text{ mod } 16$ has been generalized by Hirsch in the paper

"On a theorem of Burnside, Quart. J. Math. Oxford 1 (1950), 97-99".

By elementary methods (without using characters or other techniques from representation theory) he shows in the main theorem:

Let $p_1,...,p_k$ be the prime-divisors of $|G|$ and let $d$ be the greatest common divisor of the numbers $p_i^2-1$ $(i=1,...,k)$. Then $$\begin{array}{ll} |G| \equiv c(G) \text{ mod } 2d, & \text{ if } |G| \text{ is odd} \newline |G| \equiv c(G) \text{ mod } 3, & \text{ if } |G| \text{ is even and } 3, |G| \text{ are coprime } \end{array}$$

This also generalizes Geoff Robinson's note for $p$-groups with $p \equiv \pm 1(8)$.


It is not hopeless to extend this argument so that it applies to all finite groups of odd prime power order. Let $p$ be an odd prime, and let $K$ be a $p$-group. The congruence $|K| \equiv c(K) \mod{16}$ can be proven as a consequence of the stronger (when $p > 3$) congruence $|K| \equiv c(K) \mod{(p^{2}-1)(p-1)}$. This congruence is the best possible in the sense that there are $p$-groups (nonabelian groups of order $p^{3}$) for which $|K| - c(K) = (p^{2}-1)(p-1)$.

Let the exponent of $K$ be $p^{k}$, and let $L$ be the nilpotence class of $K$. It is difficult to work with $B(2,p^{k})$ or $Aut(B(2,p^{k}))$ when $p^{k} > 3 $ and neither of those groups is known to be finite, or, worse, when $p^{k} \geq 673$ and they are known to be infinite.
Instead, we build from $K$ the group

$P = < x,y| w^{p^{k}}=1, [w_{1}, \ldots , w_{L+1}]=1 >$,
where $w, w_{1}, \ldots , w_{L+1}$ range independently over every possible word in the generators.

Claim. $P$ is finite.
Proof of Claim. We regard $k$ as fixed and proceed by induction on $L$. When $L = 1$, $P$ is abelian and generated by two elements of order dividing $p^{k}$ so $|P| \leq p^{2k}$ and the base case is proved.
If $L > 1$, it suffices to prove that the group $P_{L}$ in the descending central series of $P$ is finitely generated, since it is abelian and all its elements have finite order dividing $p^{k}$. Then every $(L-1)$-fold commutator $[w_{1}, \ldots , w_{L}]$ is central so $P/P_{L}$ is a quotient of the group which has the same definition as $P$, except for the replacement of $L$ by $L-1$. Since that group is finite by induction, $P/P_{L}$ is finite and the finiteness of $P_{L}$ implies the finiteness of $P$. To prove this, we use the

Commutator Lemma. If $G = < g_{1}, \ldots , g_{m} >$ and $H = < h_{1}, \ldots , h_{r} > \leq G$ for some positive integers $m$ and $r$, then the commutator subgroup $[G,H]$ has a generating set consisting of iterated commutators $[e_{1}, \ldots , e_{C}]$, where each $e_{i} \in $ {$g_{1}^{\pm 1}, \ldots , g_{m}^{\pm 1}, h_{1}^{\pm 1}, \ldots , h_{r}^{\pm 1}$}.

Proof of Commutator Lemma.
Recall the commutator identities $[t, uv] = [t,v][t,u][t,u,v]$ and $[tu, v] = [t,v][t,v,u][u,v]$.
An arbitrary element of $[G,H]$ is a product of commutators of the form $[g,h]$, where $g = \gamma_{1} \ldots \gamma_{s_{1}}$, $h = \eta_{1} \ldots \eta_{s_{2}}$, each $\gamma_{i} \in$ {$g_{1}^{\pm 1}, \ldots , g_{m}^{\pm 1}$} and each $\eta_{i} \in$ {$h_{1}^{\pm 1}, \ldots , h_{r}^{\pm 1}$}. We prove it for commutators of this form by reducing the length of $h$ and then reducing the length of $g$ as follows:

If $s_{2} > 1$,

$[g, \eta_{1} \ldots \eta_{s_{2}}] = [g, \eta_{2} \ldots \eta_{s_{2}}][g, \eta_{1}][g, \eta_{1}, \eta_{2} \ldots \eta_{s_{2}}]$.

If $s_{2} = 1$ and $s_{1} > 1$,

$[\gamma_{1} \ldots \gamma_{s_{1}}, \eta_{1}] = [\gamma_{1}, \eta_{1}][\gamma_{1}, \eta_{1}, \gamma_{2} \ldots \gamma_{s_{1}}][\gamma_{2} \ldots \gamma_{s_{1}}, \eta_{1}]$.
The middle factor on the right hand side can be written as a product of commutators of the desired form by repeatedly applying the equation cited for handling the $s_{2} > 1$ case. Then the final factor has the length of $g$ reduced by 1, as desired. The Commutator Lemma is now proven.

Then $[P,P]$ is generated by a set of such iterated commutators involving $x$, $y$, $x^{-1}$ and $y^{-1}$. Since these have at most $L$ arguments, this gives a generating set of at most $\frac{4^{L+1} - 16}{3}$ elements for $[G,G]$. Then, if one has a $\rho$-element generating set for the subgroup $P_{i}$ in the descending central series of $G$, the Commutator Lemma gives a generating set for $P_{i+1}$ consisting of at most $\frac{(2\rho + 2)^{L+1} - (2\rho + 2)^{2}}{2\rho + 1}$ elements (these finite geometric series arise from summing over the possible lengths of such commutators). Then, continuing in this way, all of the subgroups in the descending central series for $P$ are finitely generated, and the finiteness of $P$ is proven.

Here $\Phi$ denotes the Frattini subgroup.
Denote $|P|$ by $p^{E}$. Let $W_{1}(x,y)$ and $W_{2}(x,y)$ be elements of $P$, considered as words in $x$ and $y$ (well-defined up to the relations). Any substitution $(x,y) \to (W_{1}(x,y),W_{2}(x,y))$ turns relations into other relations (due to the way the relations defining $P$ involve all possible words), so any such substitution defines an endomorphism of $P$. Such a substitution defines an automorphism of $P$ if and only if it is invertible. Since $P$ is finite, an endomorphism from $P$ to itself is invertible if and only if it is surjective. If an endomorphism $\psi$ of $P$ is not surjective, then $\psi(P)$ is contained in some maximal subgroup of $P$ and $\psi(P)\Phi(P)/\Phi(P)$ is a subspace of $P/\Phi(P)$ of positive $\mathbb{Z}/(p)$-codimension. So the endomorphism $\psi$ of $P$ is an automorphism if and only if its action on $P/\Phi(P)$ is via an element of $GL(2,\mathbb{Z}/(p))$. This means that the proportion of endomorphisms of $P$ which are automorphisms is the proportion of 2x2 matrices over $\mathbb{Z}/(p)$ which are invertible. Therefore, we obtain the result that $|Aut(P)| = p^{2E-3}(p^{2}-1)(p-1)$.
Now we let $Aut(P)$ act on the set of non-commuting pairs of elements of $K$. This is well-defined since the relations that hold in $P$ also hold in $K$, and in the subgroup of $K$ generated by two non-commuting elements. To get information about the size of an orbit of a non-commuting pair under this action, it suffices to get information about the order of the stabilizer in $Aut(P)$.
If $a,b \in K$ form a non-commuting pair such that $a = W_{1}(a,b)$ and $b = W_{2}(a,b)$, then $1 = a^{-1}W_{1}(a,b) = b^{-1}W_{2}(a,b)$. $| < a,b > /\Phi(< a,b >)| = p^{2}$, since $< a,b >$ is noncyclic ($< a,b >$ is nonabelian since $(a,b)$ is a non-commuting pair) and generated by 2 elements. This means that, passing to $< a,b >/\Phi(< a,b >)$, we see that $(a,b) \to (W_{1}(a,b),W_{2}(a,b))$ is mapped to the identity matrix in $GL(2,\mathbb{Z}/(p))$. Then likewise, in $Aut(P)$, $(x,y) \to (W_{1}(x,y),W_{2}(x,y))$ must be in the subgroup of order $p^{2E-4}$ which is the kernel of the homomorphism from $Aut(P)$ to $GL(2,\mathbb{Z}/(p))$. Therefore any orbit for the action of $Aut(P)$ on the set of non-commuting pairs of elements of $K$ has size equal to a multiple of $(p^{2}-1)(p^{2}-p)$, and therefore $(p^{2}-1)(p^{2}-p)$ divides $|K|(|K| - c(K))$. Since $|K|$ is a power of $p$, we obtain the congruence $|K| \equiv c(K) \mod{(p^{2}-1)(p-1)}$ as claimed.

  • $\begingroup$ I would just say that to prove $P$ finite, you can simply remark that periodic=locally finite for nilpotent groups. $\endgroup$
    – Steve D
    Oct 26, 2011 at 11:28

Removed- missed "character-free" part. In response to comment below: note that if $p$ is a prime congruent to $\pm 1$ (mod 8), then a finite $p$-group $G$ has $|G| \equiv c(G)$ (mod 32) by an easy adaptation of Burnside's argument.


Your Answer

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

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