COS 96-2
Mechanisms of species diversity in serpentine meadows: An experimental decoupling of local and regional processes
Ecologists have long sought to explain the high diversity of species observed in many biological communities, given that classic coexistence theory predicts that diversity is limited by the number of available niches. Recent work, however, also recognizes the importance of regional processes, such as dispersal limitation and environmental filtering, which promote diversity by limiting species distributions across landscapes. Because regional processes are difficult to test experimentally in natural systems, most studies to date have evaluated the relative role of local versus regional processes statistically using sampling data. Here, we used a novel ‘hay transfer’ technique to experimentally decouple environment from dispersal limitation in serpentine meadows. This technique involved vacuuming all seed and other loose material from plots, pooling the material from plots according to five spatial scaling treatments (plots located within <1 m to >10 km away), and then redistributing it. We predicted that environmental conditions would increasingly explain variation in species composition as the spatial scaling treatment increases. This result would support our hypothesis that species diversity patterns reflect dispersal limitation, and that given the opportunity, species sort deterministically according to their environmental requirements.
Results/Conclusions
Prior to manipulation, species composition at the study site showed clear signatures of spatial autocorrelation, a result that could be explained by either environmental filtering or dispersal limitation. We found that species richness significantly increased with the spatial scale of the species pool (P < 0.001), with nearly twice as many species in the largest (22.4 species) compared to the smallest (12.8 species) spatial scaling treatment on average. If this increase in species richness reflected a homogenizing effect due to transient source-sink dynamics, then sites would be expected to become more similar in species composition as the scale of the species pool increased; this was not the case. Instead, dissimilarity was high both within and among treatments, even in comparing plots found in close proximity that received different spatial scaling treatments, and did not differ between spatial scaling treatments (P > 0.1). Together, our preliminary results are consistent with our hypothesis that species distributions are strongly limited by dispersal, and that by removing this limitation local species diversity roughly doubles. The experimental removal of this dispersal limitation might allow species sorting along environmental gradients, a hypothesis that will be confirmed by incorporating site-level environmental information into our analyses.