The debate between niche and neutral models of community assembly is vigorous and far from settled. Using six years of data (2001-2006) comprising plant species records within 105 self-contained depressions at Mount St. Helens, Washington, I examined whether niche or neutral models better predict the characteristics of these sites. These depressions, which are located on the north flank of the mountain in an area of severe disturbance as a result of pumice flow, act as safe-site islands for establishing plants resulting in classic metapopulation structure. This structure, and the suggestion from earlier research that sites were dispersal limited, suggests that this may be an appropriate system to test for neutral dynamics. I tested predictions of extinction and dispersal from MacArthur and Wilson's Theory of Island Biogeography as well as Hubbell's Unified Theory of Biodiversity and Biogeography against predictions of species sorting implicated by niche assembly. I used Mantel tests to study the relationship between a ecological distance in a primary matrix of species presence/absence (Sorenson distance) and Euclidean distance in a secondary matrix containing either soil factors (pH, organic material, particle size) or UTM coordinates. I found a significant correlation between distance between plots and ecological distance for all years, suggesting that dispersal limitation may be an important factor structuring these sites. R values ranged from 0.125 to 0.247 and P values ranged from <0.001 to 0.003. Conversely, I found that the relationship between species presence/absence and soil factors was no greater than that expected by chance. I will also present the results of analyses testing whether extinction is related to abundance, site characteristics, and species characteristics.