mm, we need more hypotheses? f(x) = -x and g(x) = 1-x have order 2 in the group of bijections from Z to Z. but f(g(x)) = -(1-x) = x-1 has infinite order.
even in finite groups, e.g. S_3, (12)(23) doesn't have order 2.
the subgroup generated by the elements of order dividing p would be a characteristic subgroup, i guess. if G is finite abelian this is just the set of elements having order some power of p (including 0).
also the elements whose p-th power themselves are of course characteristic (fully characteristic even) if they are a subgroup, but they needn't be in general
being abelian is an easy sufficient condition for them constituting a subgroup
i'm trying to show that a minimal normal subgroup of a finite solvable group is an elementary abelian $p$-group. to my understanding, this is a group of the form $\{x \in G : x^p = 1\}$ (i.e. all nonidentity elements having order $p$) that is abelian
then perhaps i'm looking at the wrong subgroup (let's call it S)? i know i need to use the minimality of a minimum normal subgroup $M$ to show that $M = 1$ or $M = S$. which thus calls for showing that $S$ is normal in $M$
gosh, i'm reeling in the points today. looking forward to my mse swag private jet. i think i will go for the global 7500 so i can back to ireland without stopping.
well, since things are $C^2$ on either side of the boundary you just need to show that the (1st & 2nd) partials of $v$ are continuous at that boundary. then you can conclude that $v$ is $C^2$.
then since $u,v$ are harmonic and agree on an open set you can conclude more..