Article: The Effects of Sampling Bias on Palaeozoic Faunas and Implications for Macroevolutionary Studies

Trilobites, a dominant component of marine faunas during the Cambrian and Ordovician and which survived until the end of the Permian (542-251 Ma) have been used in many macroevolutionary analyses. Here, we use a discovery curve to document the sampling history of trilobites, which we consider a proxy for Palaeozoic faunas in general. At higher taxonomic ranks, orders, suborders and superfamilies, the fossil record has been completely sampled, while the family rank also shows a high level of sampling completeness, having reached an asymptote in 1970. Importantly, this levelling-off occurred even though worker effort continued to increase. However, at genus level the sampling record is incomplete, indicating that families should not be used as a proxy for genera. There is little variation among the different subsets of generic data, with the sampling history of different stratigraphic periods and among different orders being very similar. However, there is noticeable variation among geographical regions, caused by variations in worker effort, and this could cause problems when comparing speciation and diversity patterns across faunal provinces. The role of synonyms on sampling history has had little effect.