Trait-based assembly and phylogenetic structure in northeast Pacific rockfish assemblages.

If natural communities are assembled according to deterministic rules, coexisting species will represent a nonrandom subset of the potential species pool. We tested for signatures of assembly rules in the distribution of species' traits in Pacific rockfish (Sebastes spp.) assemblages. We used morphology, dietary niche (estimated with stable nitrogen isotopes), and distribution data to identify traits that relate to local-scale resource use (the alpha-niche) and to environmental gradients (the beta-niche). We showed that gill raker morphology was related to trophic position (an alpha-niche axis), while relative eye size was associated with depth habitat (a beta-niche axis). We therefore hypothesized that, within assemblages of coexisting rockfish species, the gill raker trait would be overdispersed (evenly spaced) due to limiting similarity, while relative eye size would be clustered due to environmental filtering. We examined the evolutionary relatedness of coexisting species to ask whether phylogenetic community structure and trait distributions gave similar indications about the roles of assembly processes. We tested the trait distributions and phylogenetic structure of 30 published rockfish assemblages against a null model of random community assembly. As predicted, the gill raker trait tended to be more evenly spaced than expected by chance, as did overall body size, while relative eye size was more clustered than expected. Phylogenetic community structure appeared to reflect historical dispersal and speciation and did not provide consistent support for assembly rules. Our results indicate that rockfish community assembly is nonrandom with regard to species' traits and show how distinguishing traits related to the alpha- and beta-niches and incorporating functional morphology can provide for powerful tests of assembly rules.