Estimating minimum global species diversity for groups with a poor fossil record: a case study of Late Jurassic–Eocene lissamphibians

Abstract

Estimates of past diversity using palaeontological evidence can be achieved within two main analytical frameworks. The traditional ‘taxic approach’ makes a straightforward use of stratigraphic ranges at a given taxonomic level, whereas the ‘phylogenetic approach’ derives palaeodiversity estimates from both phylogenetic topologies and stratigraphic occurrences. At lower taxonomic levels, the taxic and phylogenetic approaches have limitations for several groups because of the incompleteness of the fossil record and the frequent absence of stable, comprehensive phylogenetic hypotheses. In particular, fossil specimens unidentifiable at lower taxonomic levels are frequently discarded from both kinds of diversity analysis. However, these specifically or generically indeterminate occurrences may prove crucial for recovering minimal lineage-level diversity patterns, especially for groups with poor fossil records. Diversity counts can be achieved by a simple, intuitive protocol that incorporates all fossil occurrences and minimal principles of evolutionary continuity. This method, intermediate between the taxic and the phylogenetic approaches, has received little attention among palaeobiologists so far although it holds great promises for diversity studies. Here this alternative method is extended, generalised, and tested. When empirical occurrence data are artificially degraded, the technique captures an equal or a better part of the original diversity signal than the taxic approach. This method is applied here to an original locality-level compilation of fossil lissamphibians occurring in the Late Jurassic–Eocene interval. This empirical study provides the first global lineage-level palaeodiversity estimate ever made for this group, and it shows that specifically indeterminate occurrences and evolutionary continuity can increase species diversity estimates by 400%. The major implications of this work are that (1) palaeontological reports and datasets should record all fossil occurrences, regardless of the taxonomic level at which they are diagnostic; (2) it seems preferable to abandon the old commitment of using only taxa assigned to the same rank in diversity analyses; (3) the clade Lissamphibia has experienced a virtually extinction-free gradual rise during the Late Jurassic–Eocene interval; (4) long-term climatic disruptions for the K/T bio-event and hypothetical megabiases in the Late Cretaceous fossil record of small terrestrial vertebrates are unlikely.