ABSTRACT
Island biogeographic and stochastic modeling techniques are applied to the afroalpine moss flora of East Africa. A total of 277 species is recorded from seven regions; Mount Kilimanjaro, Mount Kenya, Aberdare Mountains, Mount Meru, Mount Elgon, Ruwenzori Mountains, and Virunga Volcanoes. Using Jaccard’s coefficient of similarity, most floristic relationships between these regions do not depart from ones based on random distributions of the moss flora. Significant departures from random expectation are not correlated with known physical or historical features.
Island biogeographic analyses were done to test the hypothesis that species number on each region reflects equilibrium achieved during full glacial times, when timberline decreased and afroalpine enclaves were larger than at present. Regression analyses do not support this hypothesis. Rather, habitat diversity is revealed to be the most important predictor of species richness, followed by present area. When the effects of habitat diversity are removed, residual species richness in afroalpine-restricted species is most strongly correlated with an unadjusted wind dispersal immigration index. Distance between regions is shown to be the best predictor of residual species richness in circumboreal disjuncts and afromontane endemics.
