A revision of the diagnosis and affinities of the metriorhynchoids (Crocodylomorpha, Thalattosuchia) from the Rosso Ammonitico Veronese Formation (Jurassic of Italy) using specimen-level analyses

Neptunidraco ammoniticus is a thalattosuchian crocodylomorph from the Rosso Ammonitico Veronese Formation (RAVF, Middle Jurassic) of northern Italy. Erected from one partial specimen, Neptunidraco is pivotal in reconstructing thalattosuchian evolution, being it the oldest known member of Metriorhynchidae. Two additional RAVF thalattosuchians have been referred to Neptunidraco. A revised diagnosis of N. ammoniticus is provided here. Using a well-sampled phylogenetic data set of Crocodylomorpha, the affinities of all three RAVF thalattosuchian specimens are investigated simultaneously for the first time using parsimony tree-search strategies and Bayesian inference using the Fossilized Birth-Death with Sampled Ancestor (FBDSA) model. The results of the alternative analyses are not consistent in the placement of the RAVF specimens. The holotype of N. ammoniticus is consequently referred to Metriorhynchidae incertae sedis. The first referred specimen is recovered in various alternative placements among Metriorhynchoidea. The third and most fragmentary specimen is recovered as a crocodylomorph of uncertain affinities in the parsimony analysis and in the undated Bayesian analysis, and a metriorhynchoid sister taxon of the second RAVF specimen in the tip-dated Bayesian analysis. Only a subset of the results in the parsimony-based analyses supports the referral of the latter two specimens to Neptunidraco. The unusually high rate of morphological divergence for the Neptunidraco branch, inferred in previous iterations of the Bayesian inference analyses but not recovered in the novel analysis, was likely an artifact of the a priori constraint of all RAVF thalattosuchians into a single taxonomic unit, and of the arbitrarily fixed tip-age priors for the terminal taxa. These results confirm the utility of specimen-level morphological analysis and of combined tree-search strategies for inferring the affinities and the inclusiveness of fragmentary but significant fossil taxa, and reinforce the importance of incorporating stratigraphic uncertainty as prior in tip-dated Bayesian inference analyses.