Phylogeny-Based Diversity Patterns with and without Fossils: The Limits of Historical Inference without Historical Data

Calibrated phylogenies can be used to infer the number of lineages present at different times in the past, allowing us to partially correct for incompleteness in the fossil record and to make statements about patterns of diversity over time for groups with little or no fossil record. These patterns can be compared with model-generated curves to test whether apparent increases or decreases of diversification deviate enough to support mass extinction or explosive radiation scenarios. However, comparison of extant-only approaches with those including fossils for groups with extensive living and extinct diversity indicates an inability to distinguish periods of time with low rates of background speciation from those with actual loss in diversity when the lineages responsible for diversity contraction are extinct - in other words, it may be difficult to separate acute inverted cones from bi- or multi-modal curves. Methods that model diversity over time based on external factors thought to control distribution (such as climate) may allow the generation of curves with more appropriate continuous variation in species richness over time.