One of the few important empirical generalizations regarding herbaceous plant systems has been the demonstration that species richness is related to standing crop with maximum richness occurring at moderate levels of standing crop. This relationship is normally demonstrated by comparing among vegetation types (i.e., vegetation with different dominants). We undertook this study to test whether the species richness-standing crop relationship was evident at a finer-groined level of organization, the within vegetation type level. Fifteen wetland sites were sampled in eastern Canada and species richness and standing crop determined in each of 224 0.25 m 2 quadrats. Each site was relatively homogeneous in terms of the dominant species present and were therefore categorized as single vegetation types. However, as a group, the sites comprised a wide range of vegetation types.
A second order polynomial regression indicated a significant bitonic relationship between species richness and standing crop at the among-vegetation types scale, that is, when all 15 sites were combined. At the withinvegetation type level, however, no significant relationships were observed (p > 0.05). The results indicate that the model of species richness proposed by Grime has predictive power at a coarse-grained level of organization, among vegetation types, but does not survive the transition to a finer-grained level of organization, the within vegetation type level. Therefore, the higher level processes which structure species richness patterns among vegetation types are not the same processes which determine richness patterns within a vegetation type.