Perennials on rock islands: testing for patterns of colonization and competition.

The density and presence of 18 saxicolous (occurring among rocks) perennial plant species were measured on 61 habitat islands in Baja California. These data were compared with a random colonization simulation that placed species on islands in direct proportion to species densities and island areas. The initial simulation placed too many species on the islands. This problem was rectified by introducing the effects of within island seed dispersal. Both the data and the simulations showed patterns of density compensation. Density compensation is usually perceived as being evidence of interspecific competition; however, it could be caused by other factors, such as lack of predators or habitat differences between the islands. Since the simulation included no differences between the species in their competitive ability, I used the deviations of the simulations from the data as a measure of relative competitive ability. A conservative requirement for demonstrating that competition affects the densities of these species is finding an ecological tradeoff between colonizing and competitive ability. There was no evidence for a tradeoff between competitive and colonizing ability. The dispersal method of the plants (either animal or wind) had a larger but nonsignificant effect on the deviations. Thus there is no evidence for competition affecting the deviations in density and occurrence of these species from the simulations. Random colonization including the effects of both multiple source pools and perhaps the dispersal method of the species is the most parsimonious explanation for the density and diversity patterns.