Abstract: Using Whittaker’s concepts of alpha, beta, and gamma diversity, Sepkoski asked how global diversity was assembled at scales ranging from the community to the province. In the years since, ecologists have recast diversity in terms of additive partitions where total diversity can be decomposed into sample-level alpha diversity plus the sum of a series of beta diversity terms that reflect progressively larger spatial scales. Given that marine alpha diversity represents a tiny fraction of global diversity, Phanerozoic global diversity patterns must be dominated by changes in beta diversity at one or more scales. A ballooning ecological literature demonstrates wide variation in beta diversity among ecosystems, regions, and taxa, suggesting that large changes in beta diversity on evolutionary timescales are likely. But the question is which scales are the most important. Several recent palaeontological studies help to constrain beta diversity within sedimentary basins, and the emergence of sample-based databases puts an answer to Sepkoski’s question within reach. A new method for calculating diversity partitions for richness is introduced, which allows the calculation of each species’ contribution to alpha and beta diversity, as well as the contribution of each sampling unit to beta diversity.