Palaeozoic Tetrapods as a Case Study

Studies of diversity, whether of species richness within regions (alpha diversity) or faunal turnover between regions (beta diversity), will depend heavily on the “bioregions” into which a study area is divided. However, such studies in the palaeontological literature have often been extremely arbitrary in their definition of bioregions and have employed a wide variety of spatial scales, from individual localities to formations/basins to entire continents. Such bioregions will not necessarily be separated by biologically meaningful boundaries, and results obtained at different spatial scales will not be directly comparable. In many neontological studies, however, bioregions are defined more rigorously, usually as areas of endemicity. Here a procedure is proposed whereby this principal may be applied to palaeontological datasets. In each time bin/assemblage localities are subjected to two hierarchical cluster analyses, the first grouping the localities by geographic distance, the second by taxonomic distance. Clusters shared between the two will represent geographically continuous areas of endemicity and so may be used as bioregions. When calculating alpha or beta diversity through time, the spatial scale at which the bioregions are defined needs to be standardized between each time bin. This is done by grouping clusters of localities below a predefined geographic cluster node height. This approach is used to assess changes in beta diversity of Palaeozoic tetrapods and resolve disagreements regarding changes in faunal provinciality across the Carboniferous/Permian boundary. When the bioregions are defined at a smaller spatial scale, splitting the globe into many small regions, beta diversity decreases substantially during the earliest Permian. However, when the bioregions are defined at larger spatial scales, representing areas roughly the size of continents, beta diversity remains high. This result indicates that local environmental barriers to dispersal were decreasing in importance, rejecting previous suggestions that the rainforest collapse caused an “island biogeography” effect. Instead, dispersal at this time is restricted by continental-scale barriers, with the increased orogenic uplift as a possible control.Peer Reviewe