Schur's exponent conjecture II

Abstract

Primoz Moravec published a very important paper in 2007 where he proved that if $G$ is a finite group of exponent $n$ then the exponent of the Schur multiplier of $G$ can be bounded by a function $f(n)$ depending only on $n$. Moravec does not give a value for $f(n)$, but actually his proof shows that we can take $f(n)=ne$ where $e$ is the order of $b^{-n}a^{-n}(ab)^{n}$ in the Schur multiplier of $R(2,n)$. (Here$R(2,n)$ is the largest finite two generator group of exponent $n$, and we take $a,b$ to be the generators of $R(2,n)$.) It is an easy hand calculation to show that $e=n$ for $n=2,3$, and it is a straightforward computation with the $p$-quotient algorithm to show that $e=n$ for $n=4,5,7$. The groups $R(2,8)$ and $R(2,9)$ are way out of range of the $p$-quotient algorithm, even with a modern supercomputer. But we are able toshow that $e\geq n$ for $n=8,9$. Moravec's proof also shows that if $G$ is a finite group of exponent $n$ with nilpotency class $c$, then the exponent of the Schur multiplier of $G$ is bounded by $ne$ where $e$ is the order of $b^{-n}a^{-n}(ab)^{n}$ in the Schur multiplier of the class $c$ quotient $R(2,n;c)$ of $R(2,n)$. If $q$ is a prime power we let $e_{q,c}$ be the order of $b^{-q}a^{-q}(ab)^{q}$ in the Schur multiplier of $R(2,q;c)$. We are able to show that $e_{p^{k},p^{2}-p-1}$ divides $p$ for all prime powers $p^{k}$. If $k>2$ then $e_{2^{k},c}$ equals 2 for $c<4$, equals 4 for $4\leq c\leq11$, and equals $8$ for $c=12$. If $k>1$ then $e_{3^{k},c}$ equals 1 for $c<3$, equals 3 for $3\leq c<12$, and equals 9 for $c=12$. We also investigate the order of $[b,a]$ in a Schur cover for $R(2,q;c)$.