The number of species (species richness) and 13 soil variables were determined in 30 0.1 ha plots in the Duke Forest, Durham and Orange Counties, North Carolina, in order to test whether variability in species richness can be accounted for by the environmental variability hypothesis, or the sampling effect hypothesis. The environmental variability hypothesis predicts a positive correlation between the variance in factors which are ecologically important and species richness. However, the variabilities of environmental factors are unrelated to richness. The sampling effect hypothesis predicts that more rare species will be found in species-rich sites. However, as richness increases, the number of species in all abundance classes increases. Both the environmental variability hypothesis and the sampling effect hypothesis are therefore unlikely to account for the substantial variability in species richness which exists in the Duke Forest.