Outcrops of serpentine bedrock bear as close a resemblance to islands as any terrestrial habitat, with floras often highly distinct from those of the surrounding matrix. In 30 years of research in California, I found that the sizes and spatial distribution of serpentine outcrops may influence extinction and colonization, alpha and beta diversity, reproductive success and genetics, and other facets of plant (and animal) ecology in ways consistent with spatial ecological theory. Such ‘spatial effects’ were detected primarily at scales of 10^2–103 meters separating habitat patches, whereas habitat quality and climate effects predominated at smaller and larger extents, respectively. Most regions containing sets of small outcrops (< 10 ha) also contain large ones (> 10² ha), making the detection of spatial effects dependent on careful site selection. This review highlights the interplay between testing theory and understanding the most important processes shaping biodiversity in a particular study system.