THE MISSING DIVERSITY OF THE DINOSAUR FOSSIL RECORD

A large suite of research is predicated upon the assumption that the incomplete fossil record is representative of the actual diversity that existed in the past. Based on several lines of evidence, I argue that this assumption is unjustifiable. Approximately 1,300 species of non-avian dinosaurs are known from the entire Mesozoic Era. This diversity pales in comparison to that of modern diverse amniote clades. The mean species richness of diverse, disparate amniote clades in the present is ~9000, while an average of only 7 non-avian dinosaurs are known from any 1 Ma interval in the Mesozoic. Controlling for body size, global non-avian dinosaur species richness at any time should be no less than 10³, indicating that our total knowledge of dinosaur diversity comprises at most 1% of their actual diversity, which was on the order of 100,000 total species. This diversity is asymmetrically distributed in time and space, with 20% of the Mesozoic Era containing over 50% of known dinosaur species, and each time interval represented by regional localities rather than a global sample. Even the most productive strata and localities preserve clear size bias, usually (but not exclusively) favoring the preservation of large taxa. The fossil records of non-avian dinosaurs and extant amniote clades show a linear relationship between the number of species recognized and the number of fossils sampled per geologic stage, indicating that even putatively well-sampled intervals do not approach an adequate summary of regional or global biodiversity. This and coupled diversity trends among amniote clades suggests the recognized species diversity in any geological stage is purely a function of sampling. The vast quantity of missing species in the fossil record is problematic for macroevolutionary and paleoecological research, as the influence of unknown taxa is unquantifiable and consistent with an infinite number of scenarios. Field exploration and the discovery of new fossils remain critical components of paleontological research. As methods for morphological species delimitation improve, it is likely that collections of individuals currently referred to a single species will prove to represent assemblages of multiple taxa, especially when the specimens come from a wide spatial or temporal distribution. As species are continually discovered, phylogenetic topologies will likely change considerably, particularly within clades that comprise mostly small taxa such as Paraves.