Many oceanic islands are notable for their high endemism, suggesting that islands may promote unique assembly processes. However, mainland assemblages sometimes harbour comparable levels of endemism, suggesting that island biotas may not be as unique as often assumed. Here, we test the uniqueness of island biotic assembly by comparing the rate of species turnover among islands and the mainland, after accounting for distance decay and environmental gradients. We modeled species turnover as a function of geographic and environmental distance for mainland (M-M) communities of Anolis lizards and Terrarana frogs, two clades that have diversified extensively on Caribbean islands and the mainland Neotropics. We compared mainland-island (M–I) and island-island (I–I) species turnover to predictions of the M–M model. If island assembly is not unique, then the M–M model should successfully predict M–I and I–I turnover, given geographic and environmental distance. We found that M–I turnover and, to a lesser extent, I–I turnover were significantly higher than predicted for both clades. Thus, in the first quantitative comparison of mainland–island species turnover, we confirm the long-held but untested assumption that island assemblages accumulate biodiversity differently than their mainland counterparts.