Phenotypic and phylogenetic evidence for the role of food and habitat in the assembly of communities of marine amphipods

The study of community assembly processes currently involves (a) long-standing questions about the relative importance of environmental filtering vs. niche partitioning in a wide range of ecosystems, and (b) more recent questions about methodology. The rapidly growing field of community phylogenetics has generated debate about the choice between functional traits and phylogenetic relationships for understanding species similarities, and has raised additional questions about the contribution of experimental vs. observational approaches to understanding evolutionary constraints on community assembly. In this study, we use traits, a phylogeny, and field surveys to identify the forces structuring communities of herbivorous marine amphipods and isopods living in adjacent seagrass and macroalgae. In addition, we compare our field results to a recently published mesocosm experiment that tested the effects of both trait and phylogenetic diversity on coexistence using the same species and system. With respect to community assembly processes, we found that environmental filtering was the dominant process in macroalgae habitats, that niche partitioning was the dominant process in seagrass habitats, and that the strength of these assembly mechanisms varied with seasonal fluctuations in environmental conditions and resource availability. These patterns are indicated by both phylogenetic relationships and trait distances, but the type of resources being partitioned in seagrass habitats can only be deciphered using trait data. Species coexisting in seagrass in the field differed not in their feeding niche but in traits related to microhabitat use, providing novel evidence of the relative importance of competition for food vs. habitat in structuring communities of phytophagous invertebrates. With respect to methodology, the results for seagrass habitats conflict with those obtained in mesocosms, where feeding trait diversity did promote coexistence and phylogenetic diversity had no effect. This contrast arises because a greater range of traits (some of which have much stronger phylogenetic signal than feeding traits) contribute to community assembly in the field. This highlights a mismatch between the processes that drive community assembly in the field and the processes we isolated in experimental tests, and illustrates that using phylogeny as a single proxy in both contexts may impede the synthesis of observational and experimental results.