CONTRASTING DIVERSIFICATION PATTERNS IMPLIED BY FOSSILIZED BIRTH-DEATH ANALYSES AMONG THREE AMMONOID CLADES OVER THE LATE CRETACEOUS

Bayesian phylogenetic analyses of fossil taxa such as Fossilized Birth-Death (FBD) analyses infer not only relationships among taxa, but also evolutionary histories, including diversification rates and changes in diversification rates over time. To date, few multiclade studies have been conducted to assess whether contemporaneous clades show similar reactions to the same major evolutionary and/or ecological events. Similarly, there are few studies contrasting how well diversification implied by FBD analyses corresponds to diversification implied by traditional taxonomic analyses, particularly for clades with good fossil records. Here, we examine dynamics among three different Late Cretaceous ammonoid clades (Acanthoceratidae, Baculitidae and Desmoceratoidea). For each clade, we ran reversible jump Markov Chain Monte Carlo (rjMCMC) FBD analyses with “skyline models” of varying complexity allowing for one or more shifts in origination and extinction rates corresponding to major events (e.g., beginning and ends of OAE1c or OAE2, the Santonian-Campanian boundary.). For both acanthoceratids and baculitids, rjMCMC analyses support the maximum number of diversification shifts; however, for desmoceratoids, analyses support rate shifts only at the onset and end of the OAE2 event. This corroborates prior suggestions that the evolutionary dynamics of desmoceratoids are unusually static relative to other ammonoids. All three groups show repressed origination rates during OAE2, consistent with taxonomic analyses suggesting that diversity loss at that time reflects depletion from low origination rates rather than elevated extinction rates. It is noteworthy that we find these results despite the fact that the FBD analyses are somewhat handicapped with regard to extinction rate estimation because the analyses do not take into account the survivorship of observed taxa, as taxonomic analyses do; instead, only the effect of “death rates” on expected divergence times is used. Implementation of birth-death-range methods might therefore be expected to bring diversification histories implied by Bayesian phylogenetic analyses into further accord with those implied by survivorship & reverse-survivorship analyses.