Current and historical perspectives on the completeness of the fossil record of pelycosaurian-grade synapsids

• Diversity of pelycosaurs correlates negatively with a skeletal completeness metric.
• This suggests species have been named based on poor material.
• A low relative completeness index implies that much of the record is missing.
• There has been little improvement in the record over the past two decades.
• The same biases affecting the record in the past centuries still affect it today.

Understanding of the early evolution of tetrapods may have been influenced by the effects of geological and anthropogenic sampling biases, as suggested by previous studies. Here we make a thorough assessment of how these and other biases affect the record of pelycosaurian-grade synapsids, a group of tetrapods that dominated the terrestrial realm during the late Carboniferous and early Permian. Modified versions of the character and skeletal completeness metrics are used to assess the completeness of the known specimens, while a recently published supertree of basal synapsids is tested for congruence between the stratigraphic and phylogenetic hypotheses. A historical analysis is also applied to the record in order to show how the biases in the study of this group over the past centuries may have influenced the current record. The results show that the completeness of basal synapsid specimens shows very little variation throughout the Early Permian. A peak in skeletal and character completeness metric values in the Middle Permian (Wordian) is attributed to exceptional preservation of the Mezen fauna. A significant negative correlation is found between the skeletal completeness metric and a taxic diversity curve of pelycosaurs, implying a tendency to name many species based on poor material. The lack of correlation between the character completeness metric and diversity is attributed to the history of discovery in the group: the majority of pelycosaur species were named between the 1930s and 1960s, when character based assignments were secondary to size, location and stratigraphy. A strong correspondence between the phylogeny and stratigraphy implies a reliable phylogenetic hypothesis, but the low relative completeness index score suggests that a great deal of the fossil record is missing. While the relative completeness index has improved over the past century, the rate of improvement has slowed in the last two decades, and the slowing rate of discovery of new species suggests that the limit of what the fossil record of pelycosaurs has to offer is being approached. A consistent shape of diversity curves over the past century implies that the same biases which affected the record in the past are still affecting it now. This study highlights the need for sampling correction methods to be employed in studies of early synapsids.