The metacommunity abundances are sampled logseries randomly, without respect to the metacommunity phylogeny, which is maybe wierd. I implemented a couple different ways of simulating abundances on the tree in my PIED package, so this could be integrated into MESS to give more biologically meaningful metacommunity abundances.