Analyses of how fossil record sampling influences detection of past biodiversity change, and geographic patterns of animal diversity, are central to understanding macroevolution, deep time macroecology, and interactions between the rock and fossil records. The diversification and establishment of complex trophic structure among terrestrial vertebrates (tetrapods) and three hypothesized mass extinctions occurred in the Permian (299.0–252.3 Ma), during a predominantly icehouse global climate regime. Prominent, short-lived features of Permian species diversity curves, such as a possible end-Guadalupian mass extinction, are robust to sampling biases. However, multivariate and time series approaches show that longer term trends in the accumulation of terrestrial biodiversity are more sensitive. Olson’s extinction, believed to have occurred at the end of the Early Permian, might be an artifact of the abrupt geographic shift from predominant sampling of Lower Permian, amphibian-rich, paleotropical everwet biome in North America, to more tetrapod-depauperate Middle–Late Permian, cool paleotemperate biomes in Russia and South America. Thus, geographic heterogeneity in fossil record sampling, superimposed on a prominent, icehouse latitudinal biodiversity gradient, might have created an artifactual macroevolutionary event. This apparent event remains even after correcting data for uneven temporal sampling. Thus, a complex interplay between spatial and temporal biodiversity trends, and geographic and temporal sampling factors, is responsible for generating observed past biodiversity patterns.
