We examined the relationship of habitat spatial characteristics to local and regional species diversity patterns in salt marshes of the Point Reyes peninsula.  Marshes in this area consist of  pocket marshes of varying size in a variety of matrix types, providing a unique opportunity to test the components of species-area relationship theories (e.g. island biogeography, metacommunity) in a naturally occurring  habitat at a scale where confounding factors (e.g. different evolutionary histories) can be avoided.   We examined three marsh complexes of  large continuous sites matched to smaller sites in order to test the response of species diversity to habitat discontinuity.   For each sampling area, we quantified vegetation composition, cover and vertical complexity, spider composition (with pit traps and vacuuming sampling), and spatial characteristics (using GPS).  Our research questions included:  Is species diversity across trophic levels (e.g. plants and spiders) related to area and isolation?  How might this relationship be affected by habitat discontinuity (e.g. small distinct patches vs. small patches in a continuous habitat)?  For example, will high marsh communities in patchy environments show different patterns of diversity (local, differentiation and regional) than in continuous habitats, as predicted by metacommunity models?  Will smaller, isolated habitats have increased generalists and decreased specialists?

Results/Conclusions

When all sites were grouped together across the region, small patches and continuous sites were significantly different in spider and plant species multivariate space.  This pattern was most pronounced in more “closed” sites, those surrounded extensively by marsh plain, and least in sites more contiguous with the surrounding upland matrix.  Specialist, web-building spiders were more strongly correlated to patches in continuous habitats and vegetation complexity than those in small, isolated patches.  In contrast, generalist, hunting spiders were stronger indicators of small habitat patches.  Patterns in spider diversity reflected patterns of community organization more likely predicted by metacommunity theory.  Sites that were the most open, in terms of increased contact with upland communities, had the greatest Beta diversity.  Sites with lower regional spider diversity had similarly low beta diversity.  Plant richness was positively correlated to area, as predicted by island biogeography theory, but spider richness was not.  Our study suggests that multiple processes influence species diversity patterns in response to habitat spatial characteristics and life-history traits.