The description of Tetrapodophis amplectus, a four-limbed Early Cretaceous reptile from the Crato Formation of Brazil, as the first four-legged snake made it a central piece of evidence in the debate on snake origins — until first-hand restudy concluded that it is a dolichosaurid lizard with aquatic adaptations and no evidence of constriction. That reversal closed the specimen-specific questions raised in the original 2015 description (limb function, a four-limbed stage in snake limb loss, early constriction), but left fully open the general problem the discovery was meant to illuminate: the origin and early evolution of snakes. The surviving open problem comprises three linked questions: (1) the phylogenetic position of Serpentes within Squamata — including whether snakes are allied with mosasauroid lizards as morphology-based analyses repeatedly suggest, or placed elsewhere within Toxicofera as molecular data indicate — and which Mesozoic taxa genuinely lie on the snake stem; (2) the geographic origin of stem versus crown snakes (Laurasian stem versus Gondwanan crown) and the timing, route, and mode of dispersal between Gondwanan and Laurasian faunas; and (3) the ecology of the earliest snakes and the origin of constriction, specifically whether the characteristically high precloacal vertebral counts of snakes evolved in association with constricting prey. A consilient resolution would bring the fossil record, phylogenetic inference, and functional evidence together in one coherent account of how, where, and why the lizard-to-snake transition occurred.

Background

Snakes (Serpentes) are limbless, extremely elongate squamate reptiles — Squamata being the clade comprising lizards and snakes, with roughly 12,000 living species. Their origin from lizard ancestors is one of the most debated evolutionary transitions in vertebrate biology. A few distinctions frame the debate. The crown group of a clade comprises all living members and their most recent common ancestor; the stem comprises extinct lineages closer to that crown than to any other clade but outside it. Living snakes split into Scolecophidia (small, burrowing, insectivorous blind snakes) and Alethinophidia (all other snakes, mostly carnivorous and capable of macrostomy, the wide-gape skull mechanism allowing ingestion of large prey). Constriction — subduing prey by encircling it with the body and tightening during the prey's exhalation — is the signature predatory strategy of many alethinophidians, and snakes are characterized by very high precloacal (trunk) vertebral counts, from roughly 120 to over 300, versus a few dozen in typical lizards.

Three ecological scenarios compete for the ancestral snake: a fossorial (burrowing) origin, a marine origin, and a terrestrial origin. The fossil record bearing on this is thin but growing. Cretaceous snakes with hind limbs — Najash from Patagonia, and Pachyrhachis, Haasiophis and Eupodophis from the eastern Tethys margin — document that the lizard-to-snake transition passed through limbed stages, and Dinilysia (Late Cretaceous, Patagonia) is a well-known stem snake. Caldwell et al. (2015) extended the record back to the Middle Jurassic–Lower Cretaceous with fragmentary snakes from both Laurasia and Gondwana, implying an early and geographically widespread radiation; these include the parviraptorids (e.g., Portugalophis, Diablophis), small aquatic forms whose snake affinities remain debated. Morphological phylogenetic analyses have often placed snakes near the mosasauroids — mosasaurs being large extinct marine lizards, with dolichosaurs and aigialosaurs their small, elongate Cretaceous relatives — whereas molecular data place snakes within Toxicofera (the clade also comprising iguanian and anguimorph lizards) with unstable exact relationships.

In 2015, Martill, Tischlinger and Longrich described Tetrapodophis amplectus from the Early Cretaceous (Aptian) Crato Formation of Brazil — a laminated limestone deposit famous for preserving delicate soft-bodied fossils — as the first snake retaining all four limbs. They interpreted it as fossorial, with small limbs adapted for grasping (seizing prey or claspers in mating), a flexible spine capable of constricting prey, vertebrate remains in the gut, and argued for a Gondwanan origin of crown Serpentes. The unresolved issues raised by that description are: the function of the limbs; whether they represented a transient stage in limb loss; the affinities of Tetrapodophis and their bearing on the morphology-based snake–mosasaur association; the contrast between possible Laurasian stem snakes and a Gondwanan crown origin; the timing and mode of dispersal between Gondwana and Laurasia; and the hypothesis that high precloacal vertebral counts evolved for constriction.

Later first-hand restudy overturned the specimen's snake identity. Lee et al. (2016) reported aquatic adaptations in all four limbs, and Caldwell et al. (2021), examining the holotype and its previously undescribed counterpart, found no snake-diagnostic anatomy, recovered Tetrapodophis as a dolichosaur in every dataset-method combination tested, and found no evidence for constriction or macrostomy. The specimen-specific sub-questions are therefore settled negatively, and what remains open is the general problem the discovery was meant to illuminate. Recent work keeps every component of that problem in play: skull-shape analyses recover a fossorial ancestor of crown snakes within a surface-terrestrial broader clade (Da Silva et al. 2018); brain-endocast evidence supports burrowing at the crown-snake origin (Macrì et al. 2023); the Middle Jurassic parviraptorids still oscillate between stem-snake and non-snake placements (Benson et al. 2025); and a new exceptionally preserved Brazilian stem snake, Tametara mirim, is described by Simões et al. (2026) against what they call the century-old challenge of snake ecological origins.

Problem Statement

Resolve the origin and early evolution of snakes (Serpentes). The original formulation (Martill et al. 2015) raised a bundle of unresolved issues around the four-limbed Crato Formation reptile Tetrapodophis amplectus; later first-hand restudy established that Tetrapodophis is an Early Cretaceous dolichosaurid lizard with aquatic adaptations and no evidence of constriction (Lee et al. 2016; Caldwell et al. 2021), so the specimen-specific sub-questions — the function of its limbs, their status as a four-limbed stage in snake limb loss, and its value as evidence of early constriction — are settled negatively and are not part of this problem. The surviving open core is the general snake-origins problem the discovery was meant to illuminate, which the subsequent literature continues to treat as open in its full generality (Caldwell et al. 2015; Benson et al. 2025; Simões et al. 2026); it comprises three linked questions that jointly serve one objective, a consilient reconstruction of how, where, and why the lizard-to-snake transition occurred:

(1) Phylogenetic position and stem lineage: where does Serpentes attach within the squamate tree — in particular, is it allied with mosasauroid lizards (dolichosaurs, aigialosaurs, mosasaurs) as morphology-based analyses have repeatedly suggested, or elsewhere within Toxicofera as molecular data indicate — and which Mesozoic taxa (e.g., parviraptorids, Najash, Dinilysia, Tametara) genuinely lie on the snake stem?

(2) Biogeography and timing: did stem snakes originate in Laurasia while crown Serpentes originated in Gondwana, and — under whatever placement emerges from (1) — when did snakes originate and by what route and mode did they disperse between Gondwanan and Laurasian faunas?

(3) Ecology and feeding: what was the ancestral ecology of the snake stem lineage (fossorial, terrestrial, or aquatic), and when and in what order did constriction enter snake feeding ecology relative to the evolution of the characteristically high precloacal vertebral counts — specifically, did axial elongation evolve in association with constriction?

All three sub-questions are mandatory and must be integrated into a single consilient account, each resolved at the level of an explicit reconstruction — a recovered position of Serpentes on the squamate tree with the Mesozoic stem taxa placed (1), a dated geographic and dispersal scenario (2), and an ancestral-ecology reconstruction ordering constriction relative to axial elongation (3) — not a partial answer to one sub-question in isolation. Where the supporting evidence conflicts — as it does for the placement of Serpentes (mosasauroid morphology versus Toxicoferan molecules) and for the parviraptorid affinities — a sub-question counts as resolved by a committed reconstruction that states the conclusion the total evidence most supports and explicitly adjudicates the competing signal, and (2) must state what it inherits from the placement adjudicated in (1); merely reporting that competing scenarios remain possible does not resolve the sub-question.

References

  1. David M. Martill, Helmut Tischlinger, Nicholas R. Longrich (2015), 'A four-legged snake from the Early Cretaceous of Gondwana', Science 349(6246):416–419. DOI: 10.1126/science.aaa9208. https://doi.org/10.1126/science.aaa9208
  2. Susan E. Evans (2015), 'Four legs too many?', Science 349(6246):374–375. DOI: 10.1126/science.aac5672. https://doi.org/10.1126/science.aac5672
  3. Michael W. Caldwell, Randall L. Nydam, Alessandro Palci, Sebastián Apesteguía (2015), 'The oldest known snakes from the Middle Jurassic-Lower Cretaceous provide insights on snake evolution', Nature Communications 6:5996. DOI: 10.1038/ncomms6996. https://doi.org/10.1038/ncomms6996
  4. Michael S. Y. Lee, Alessandro Palci, Marc E. H. Jones, Michael W. Caldwell, James D. Holmes, Robert R. Reisz (2016), 'Aquatic adaptations in the four limbs of the snake-like reptile Tetrapodophis from the Lower Cretaceous of Brazil', Cretaceous Research 66:194–199. DOI: 10.1016/j.cretres.2016.06.004. https://doi.org/10.1016/j.cretres.2016.06.004
  5. Robert R. Reisz, Michael W. Caldwell (2016), 'Harness passion of private fossil owners', Nature 537(7620):307. DOI: 10.1038/537307a. https://doi.org/10.1038/537307a
  6. Michael W. Caldwell, Tiago R. Simões, Alessandro Palci, Fernando F. Garberoglio, Robert R. Reisz, Michael S. Y. Lee, Randall L. Nydam (2021), 'Tetrapodophis amplectus is not a snake: re-assessment of the osteology, phylogeny and functional morphology of an Early Cretaceous dolichosaurid lizard', Journal of Systematic Palaeontology 19(13):893–952. DOI: 10.1080/14772019.2021.1983044. https://doi.org/10.1080/14772019.2021.1983044
  7. Filipe O. Da Silva, Anne-Claire Fabre, Yoland Savriama, Joni Ollonen, Kristin Mahlow, Anthony Herrel, Johannes Müller, Nicolas Di-Poï (2018), 'The ecological origins of snakes as revealed by skull evolution', Nature Communications 9:376. DOI: 10.1038/s41467-017-02788-3. https://doi.org/10.1038/s41467-017-02788-3
  8. Michael C. Grundler, Daniel L. Rabosky (2021), 'Rapid increase in snake dietary diversity and complexity following the end-Cretaceous mass extinction', PLOS Biology 19(10):e3001414. DOI: 10.1371/journal.pbio.3001414. https://doi.org/10.1371/journal.pbio.3001414
  9. Simone Macrì, Ida-Maria Aalto, Rémi Allemand, Nicolas Di-Poï (2023), 'Reconstructing the origin and early evolution of the snake brain', Science Advances 9(39):eadi6888. DOI: 10.1126/sciadv.adi6888. https://doi.org/10.1126/sciadv.adi6888
  10. Roger B. J. Benson, Stig A. Walsh, Elizabeth F. Griffiths, Zoe T. Kulik, Jennifer Botha, Vincent Fernandez, Jason J. Head, Susan E. Evans (2025), 'Mosaic anatomy in an early fossil squamate', Nature 647:673–679. DOI: 10.1038/s41586-025-09566-y. https://doi.org/10.1038/s41586-025-09566-y
  11. Tiago R. Simões, Gabriela Sobral, Simone Macrì, Roy Ebel, Thiago S. Fachini, Agustín G. Martinelli, William R. Nava, Giovanna M. X. Paixão, Luis M. Chiappe, Nicolas Di-Poï, Annie S. Hsiou (2026), 'Exceptional brain and ecological diversity in the earliest snakes', Nature. DOI: 10.1038/s41586-026-10809-9. https://doi.org/10.1038/s41586-026-10809-9