The regulation of plant species diversity was studied for 14 islands, four island groups, and 10 coastal mainland areas of California and Baja California. Native and total species numbers for these areas were regressed on five independent variables using several kinds of models. The results suggest the following for these areas: 1. Area is the best single predictor of species diversity. 2. Curvilinear models are superior to linear models and mixed linear-curvilinear models. 3. The addition of other variables significantly contributes to the prediction of species numbers. 4. In the islands, ecological diversity, richness of the environment, and isolation contribute significantly to the regulation of species abundance. 5. For mainland areas, ecological diversity and richness of the environment contribute significantly.