Problems of the interrelationships of crown and stem amniotes

Amniota is a tetrapod group that was recognized originally in the 19th century on the basis of developmental features (extra-embryonic membranes), but since the widespread adoption of phylogenetic systematics in the late 20th century, it has been recast as a crown clade. The oldest amniotes are Late Carboniferous in age (ca. 318 million years ago), and they are preserved in coal beds and lycopod tree stumps that have yielded rich faunas of temnospondyls, anthracosaurs, and other early tetrapods. Numerous phylogenetic studies of the past three decades have focused on the relationships of amniotes to other early tetrapods, resulting in the prevailing picture that Amniota is the crown of a total group with many extinct Paleozoic clades comprising its stem group. The content and the sequence of branching among stem amniotes are vigorously debated topics in early tetrapod paleontology. The 'traditional' stem amniote group Diadectomorpha has been proposed as the sister group of synapsid amniotes in a series of papers by one research group. If correct, the placement of Diadectomorpha within Amniota implies that high-fiber herbivory, a key component of terrestrial vertebrate ecosystems of modern aspect, is an entirely amniote phenomenon. Another stem group, Recumbirostra, has been recovered by other researchers within Amniota as reptiles. Recumbirostrans exhibit cranial features that are correlated with a fossorial lifestyle, and most species for which we have postcrania reveal the phenomena of body elongation and many of these exhibit limb reduction. If correct, the identification of recumbirostrans as early reptiles not only greatly augments the content of Paleozoic Amniota but also marks the evolution of skeletal features indicative of a burrowing lifestyle that predates the appearance of unequivocal fossorial amniotes (cistecephalid dicynodonts) by 40 million years. However, it is premature to accept either diadectomorphs or recumbirostrans as amniotes, given issues with scorings for amniote taxa, character argumentation, and poor homologization of braincase features between amniote and anamniote taxa.