In order to find and define any assembly rules for communities, we mus
t first investigate the patterns among species assemblages. We used a
series of null models to test for patterns in wetland plant compositio
n al the level of species, functional guilds, and traits. At the speci
es level, we found significant checkerboard and nestedness patterns. T
hree functional guilds had some tendency to contribute a constant perc
entage to species richness, but after Bonferroni correction there was
no significant pattern. Coexisting plant species showed no consistent
overall pattern of morphological dispersion. However, when we consider
ed each of 11 traits in turn, we found that 4 traits were overdisperse
d and 3 were underdispersed. Thus there are morphological assembly rul
es that constrain wetland plant community composition. These results r
econcile contrasting views of community assembly. Communities can be s
imultaneously structured by a tension between two forces: abiotic exte
rnal forces that constrain certain traits within limits and biotic int
ernal forces that tend to keep coexisting species from being too simil
ar. Because our sites vary along a fertility/disturbance gradient, we
also investigated how trait dispersion varies in space. Trait dispersi
on increases with soil Fertility; soil phosphorus explains about 36% o
f the variance in mean nearest neighbor distance. Species richness ten
ds to decline with mean nearest neighbor distance, which contrasts wit
h the general pattern for animal assemblages.