Search

The origin of snakes remains one of the most contentious evolutionary transitions in vertebrate evolution. The discovery of snake fossils with well-formed hind limbs provided new insights into the phylogenetic and ecological origin of snakes. In 2015, a fossil from the Early Cretaceous Crato Formation of Brazil was described as the first known snake with fore- and hind limbs (Tetrapodophis amplectus), and was proposed to be fossorial, to exhibit large gape feeding adaptations (macrostomy) and to possess morphologies suggesting constriction behaviours. First-hand examination of T. amplectus, including its undescribed counterpart, provides new evidence refuting it as a snake. We find: a long rostrum; straight mandible; teeth not hooked zygosphenes/zygantra absent; neural arch and spines present and tall with apical epiphyses; rib heads not tubercular; synapophyses simple; and lymphapophyses absent. Claimed traits not preserved include: braincase/descensus parietalis; ���L���-shaped nasals; intramandibular joint; replacement tooth crowns; haemal keels; tracheal rings; and large ventral scales. New observations include: elongate retroarticular process; apex of splenial terminating below posterior extent of tooth row; >10 cervicals with hypapophyses and articulating intercentra; haemapophyses with articulating arches; reduced articular surfaces on appendicular elements; rows of small body scales; and reduced mesopodial ossification. The axial skeleton is uniquely elongate and the tail with >100 vertebrae is not short as previously claimed, although overall the animal is small (���195 mm total length). We assessed the relationships of Tetrapodophis using a revised version of the original morphological dataset, an independent morphological dataset, and these two datasets combined with molecular data. All four were analysed under parsimony and Bayesian inference and unambiguously recover Tetrapodophis as a dolichosaur. We find that Tetrapodophis shows aquatic adaptations and there is no evidence to support constricting behaviour or macrostomy.

Supplementary Text; Figures S1 to S13; Tables S1 to S3.

Additional file 1. Text document containing additional methodological information and supplementary figures.

Additional file 2. Text document containing list of sampled taxa, along with their occurrence data, stratigraphic interval, age, anatomical bibliography, and personally observed specimen numbers.

Additional file 3. Text document including phylogenetic morphological characters list with individual character descriptions.

additional information, figures, and line drawings of the study specimen

Log file of the model-based analysis

List of the characters used for the phylogenetic analysis, list of resulting apomorphies, and additional resulting trees of the parsimony analyses

Landmarking procedure, replacement taxa, and supplementary figures S1-3.

Figure S1, two phylogenies inclusive of fossil taxa that were used in this study; Figure S2, accuracy test of landmarking procedure; Figure S3, distribution of variance of the principal components; Figure S4, distribution of variance for the first 10 phylogenetic principal components derived from both phylogenies adopted (i.e. with Dinilysia as a stem alethinophidian and as a stem ophidian).