NOTE: At the time of publication, the author Shannon J. McCauley was not yet affiliated with Cal Poly.
The growth of metacommunity ecology as a subdiscipline has increased interest in how processes at different spatial scales structure communities. However, there is still a significant knowledge gap with respect to relating the action of niche-and dispersal-assembly mechanisms to observed species distributions across gradients. Surveys of the larval dragonfly community (Odonata: Anisoptera) in 57 lakes and ponds in southeast Michigan were used to evaluate hypotheses about the processes regulating community structure in this system. We considered the roles of both niche- and dispersal-assembly processes in determining patterns of species richness and composition across a habitat gradient involving changes in the extent of habitat permanence, canopy cover, area, and top predator type. We compared observed richness patterns and species distributions in this system to patterns predicted by four general community models: species sorting related to adaptive tradeoffs, a developmental constraints hypothesis, dispersal assembly, and a neutral community assemblage. Our results supported neither the developmental constraints nor the neutral-assemblage models. Observed patterns of richness and species distributions were consistent with patterns expected when adaptive tradeoffs and dispersal-assembly mechanisms affect community structure. Adaptive tradeoffs appeared to be important in limiting the distributions of species which segregate across the habitat gradient. However, dispersal was important in shaping the distributions of species that utilize habitats with a broad range of hydroperiods and alternative top predator types. Our results also suggest that the relative importance of these mechanisms may change across this habitat gradient and that a metacommunity perspective which incorporates both niche- and dispersal- assembly processes is necessary to understand how communities are organized.