A new peculiar early diverging caenophidian snake (Serpentes) from the late Eocene of Hordle Cliff, England

palevol comptes rendus 2025  24  25 Comptes Rendus Palevol est une revue en flux continu publiée par les Publications scientifiques du Muséum, Paris et l’Académie des sciences, Paris Comptes Rendus Palevol is a fast track journal published by the Museum Science Press, Paris and the Académie des sciences, Paris Les Publications scientifiques du Muséum publient aussi / The Museum Science Press also publish: Adansonia, Geodiversitas, Zoosystema, Anthropozoologica, European Journal of Taxonomy, Naturae, Cryptogamie sous-sections Algologie, Bryologie, Mycologie. L’Académie des sciences publie aussi / The Académie des sciences also publishes: Comptes Rendus Mathématique, Comptes Rendus Physique, Comptes Rendus Mécanique, Comptes Rendus Chimie, Comptes Rendus Géoscience, Comptes Rendus Biologies. Diffusion – Publications scientifiques Muséum national d’Histoire naturelle CP 41 – 57 rue Cuvier F-75231 Paris cedex 05 (France) Tél. : 33 (0)1 40 79 48 05 / Fax : 33 (0)1 40 79 38 40 [email protected] / https://sciencepress.mnhn.fr Académie des sciences, Institut de France, 23 quai de Conti, 75006 Paris. © This article is licensed under the Creative Commons Attribution 4.0 International License (https://creativecommons.org/licenses/by/4.0/) ISSN (imprimé / print) : 1631-0683/ ISSN (électronique / electronic) : 1777-571X Directeurs De la publication / Publication directors : Gilles Bloch, Président du Muséum national d’Histoire naturelle Étienne Ghys, Secrétaire perpétuel de l’Académie des sciences réDacteurs en chef / editors-in-chief (*, took charge of the editorial process of the article/a pris en charge le suivi éditorial de l’article) : Michel Laurin* (CNRS), Philippe Taquet (Académie des sciences) assistante De réDaction / assistant editor : Adenise Lopes (Académie des sciences ; [email protected]) Mise en page / Page layout : Audrina Neveu (Muséum national d’Histoire naturelle ; [email protected]) révisions linguistiques Des textes anglais / english language revisions : Kevin Padian (University of California at Berkeley) réDacteurs associés / associate editors : Micropaléontologie/Micropalaeontology Lorenzo Consorti (Institute of Marine Sciences, Italian National Research Council, Trieste) Paléobotanique/Palaeobotany Cyrille Prestianni (Royal Belgian Institute of Natural Sciences, Brussels) Anaïs Boura (Sorbonne Université, Paris) Métazoaires/Metazoa Annalisa Ferretti (Università di Modena e Reggio Emilia, Modena) Paléoichthyologie/Palaeoichthyology Philippe Janvier (Muséum national d’Histoire naturelle, Académie des sciences, Paris) Amniotes du Mésozoïque/Mesozoic amniotes Hans-Dieter Sues (Smithsonian National Museum of Natural History, Washington) Tortues/Turtles Walter Joyce (Universität Freiburg, Switzerland) Lépidosauromorphes/Lepidosauromorphs Hussam Zaher (Universidade de São Paulo) Oiseaux/Birds Jingmai O’Connor (Field Museum, Chicago) Paléomammalogie (mammifères de moyenne et grande taille)/Palaeomammalogy (large and mid-sized mammals) Grégoire Métais (CNRS, Muséum national d’Histoire naturelle, Sorbonne Université, Paris) Paléomammalogie (petits mammifères sauf Euarchontoglires)/Palaeomammalogy (small mammals except for Euarchontoglires) Robert Asher (Cambridge University, Cambridge) Paléomammalogie (Euarchontoglires)/Palaeomammalogy (Euarchontoglires) K. Christopher Beard (University of Kansas, Lawrence) Paléoanthropologie/Palaeoanthropology Aurélien Mounier (CNRS/Muséum national d’Histoire naturelle, Paris) Archéologie préhistorique (Paléolithique et Mésolithique)/Prehistoric archaeology (Palaeolithic and Mesolithic) Nicolas Teyssandier (CNRS/Université de Toulouse, Toulouse) Archéologie préhistorique (Néolithique et âge du bronze)/Prehistoric archaeology (Neolithic and Bronze Age) Marc Vander Linden (Bournemouth University, Bournemouth) référés / reviewers : https://sciencepress.mnhn.fr/fr/periodiques/comptes-rendus-palevol/referes-du-journal couverture / cover : Made from the Figures of the article. Comptes Rendus Palevol est indexé dans / Comptes Rendus Palevol is indexed by: – Cambridge Scientific Abstracts – Current Contents® Physical – Chemical, and Earth Sciences® – ISI Alerting Services® – Geoabstracts, Geobase, Georef, Inspec, Pascal – Science Citation Index®, Science Citation Index Expanded® – Scopus®. Les articles ainsi que les nouveautés nomenclaturales publiés dans Comptes Rendus Palevol sont référencés par / Articles and nomenclatural novelties published in Comptes Rendus Palevol are registered on: – ZooBank® (http://zoobank.org) 505 COMPTES RENDUS PALEVOL • 2025 • 24 (25) © Publications scientifiques du Muséum et/and Académie des sciences, Paris. www.cr-palevol.fr Georgios L. GEORGALIS Institute of Systematics and Evolution of Animals, Polish Academy of Sciences, Sławkowska 17, 31-016 Kraków (Poland) [email protected] (corresponding author) Marc E. H. JONES Science Group: Fossil Reptiles, Amphibians and Birds Section, Natural History Museum, Cromwell Road, SW7 5BD London (United Kingdom) Submitted on 4 December 2024 | Accepted on 9 August 2025 | Published on 7 November 2025 A new peculiar early diverging caenophidian snake (Serpentes) from the late Eocene of Hordle Cliff, England urn:lsid:zoobank.org:pub:182D1817-2CD6-4289-B9DA-7144AEC882DD Georgalis G. L. & Jones M. E. H. 2025. — A new peculiar early diverging caenophidian snake (Serpentes) from the late Eocene of Hordle Cliff, England, in Georgalis G. L., Zaher H. & Laurin M. (eds), Snakes from the Cenozoic of Europe – towards a macroevolutionary and palaeobiogeographic synthesis. Comptes Rendus Palevol 24 (25): 505-530. https:// doi.org/10.5852/cr-palevol2025v24a25 ABSTRACT We here describe a new genus and species of snake, based on several trunk and caudal vertebrae, originating from the late Eocene (MP 17a) of Hordle Cliff, England. We investigated the fossil material via both visual microscopy and micro-computed tomography (μCT) scanning, focused on its intracolumnar variation, and extensively compared it with other Paleogene snake taxa from England and continental Europe. The new small taxon is characterized by an array of bizarre and distinctive vertebral features that can differentiate it from all other snakes. Its morphology is similar to that of russellophiids; however, some anatomical features differ drastically from those observed in the latter group and therefore, defy such family-level placement. In addition, the new English taxon bears some striking resemblance with extant acrochordids, particularly with the species Acrochordus granulatus (Schneider, 1799). Accordingly, we consider that the new taxon most likely represents an early diverging caenophidian, potentially being even a member of Acrochordidae Bonaparte, 1831, far outside the so far known stratigraphic and geographic distribution of the latter group. It further adds to the astonishing diversity of vertebral morphologies of European Paleogene snakes. RÉSUMÉ Un nouveau serpent caenophidien (Serpentes) particulier à divergence précoce de la fin de l’Éocène de Hordle Cliff, Angleterre. Nous décrivons ici un nouveau genre et une nouvelle espèce de serpents, basés sur plusieurs vertèbres troncales et caudales, originaires de la fin de l’Eocène (MP 17a) de Hordle Cliff, en Angleterre. Nous avons étudié le matériel fossile à la fois par microscopie visuelle et par tomodensitométrie (μCT), en nous concentrant sur sa variation intracolonnaire et en le comparant de manière approfondie à d’autres taxons de serpents du Paléogène d’Angleterre et d’Europe continentale. Le nouveau petit taxon est caractérisé par un ensemble de caractéristiques vertébrales bizarres et distinctives qui peuvent le différencier de tous les autres serpents. Sa morphologie se rapproche quelque peu de celle des russellophiidés, cependant, certaines de ses caractéristiques anatomiques sont radicalement différentes de celles observées dans ce dernier groupe et défient donc un tel placement au niveau de la KEY WORDS Squamata, Caenophidia, Russellophiidae, Acrochordidae, vertebral morphology, new genus, new species. 506 COMPTES RENDUS PALEVOL • 2025 • 24 (25) Georgalis G. L. & Jones M. E. H. INTRODUCTION Snakes from the Eocene of England have been known for almost two centuries. They comprise some of the earliest important finds in the history of palaeophidiology, described by one of the most prominent palaeontologists and anatomists of the time, Sir Richard Owen. Indeed, the early Eocene of England yielded the first remains of the iconic giant snake Palaeophis Owen, 1841 (see Owen 1841, 1850), whereas the late Eocene of Hordle Cliff, yielded the first fossil constrictor snake to be described, Paleryx Owen, 1850 (see Owen 1850). Since then, Eocene English fossil snakes have been the focus of several studies, including the description of new material, the establishment of new taxa, and/or the redocumentation of important species that were named in the 19th century (Lydekker 1888a, b; Rage & Ford 1980; Holman 1993, 1996; Holman & Harrison 1998a, b; Holman etal. 2006; Georgalis etal. 2021b). Here we describe a new genus and species of a very peculiar, small snake from the late Eocene of Hordle Cliff. This material was briefly mentioned and discussed for a couple of sentences (but not figured) in Milner etal. (1982) as “ Caenophidian 2”. It was noted to have some general resemblance with the genus Russellophis Rage, 1975, which back then was known only from early Eocene of France. We conduct a detailed anatomical documentation of the material and provide extensive comparisons with other coeval snakes from Europe. MATERIAL AND METHODS All fossil specimens of the new taxon, described here, are permanently curated at the collections of the Natural History Museum, London (NHMUK). All specimens of the new taxon were collected by Roy Gardner in 1981. He collected matrix and processed it using screen washing techniques and most microvertebrate remains were separated using sieves with a 1 mm mesh. Digital models were made for five of the specimens: NHMUK PV R 10795 (holotype); NHMUK PV R 10796; NHMUK PV R 10797; NHMUK PV R 10798 (paratype); NHMUK PV R 38946. These models were made at the Natural History Museum UK using X-ray micro-Computed Tomography with the Zeiss Xradia 520 Versa. Specimens were packed individually in small plastic tubes and scanned separately. Settings comprised 140 kV and 72 μA, exposure 3 ms, to produce TIFF slices with a voxel size of 0.00447 × 0.00447 × 0.00447 mm. The Tiffs stacks were then opened as volumes in Avizo 3D 2021.2 (Thermo Fisher Scientific, Waltham, Massachusetts, United States). The volumes were cropped before segmentation. Segmentation initially involved selecting a threshold which closely represented the external surface of the bone. Manual segmentation tools (e.g. lasso, draw) were then used to remove remaining matrix. Typically, every other slide was segmented and the selection was interpolated. For each bone two surface files were made: to document the segmentation a surface file was made with Smoothing set to none; for the figures a surface file was made with Smoothing set to Unconstrained smoothing. These bone surface files were exported as a Stanford PLY. All these models are available at the online repository of Morphosource: https://www.morphosource. org/projects/000658862. Taxonomy follows Smith & Georgalis (2022). Anatomical terminology follows Georgalis etal. (2021b), Head (2021), and Szyndlar & Georgalis (2023). InstItutIonal abbrevIatIons ISEA Institute of Systematics and Evolution of Animals, Polish Academy of Sciences, Kraków; MNHN Muséum national d’Histoire naturelle, Paris; MSUVP Michigan State University Museum of Vertebrate Paleontology, East Lansing, Michigan; NHMUK Natural History Museum, London. LOCALITY Hordle (also known in the older literature as Hordwell) Cliff is a locality of the Headon Hill Formation, in southern England (Edwards & Daley 1997). Besides snakes, among reptiles, the locality has yielded a diverse assemblage of turtles, crocodylians, and lizards (Owen & Bell 1849; Huxley 1859; Seeley 1876; Lydekker 1887, 1888a, b, 1889; Hooley 1905; Milner etal. 1982; Benton & Spencer 1995; Klembara & Green 2010; Georgalis & Joyce 2017). The locality is particularly known for its rich and considerably diverse mammal fauna (e.g. Hooker etal. 2005; Hooker & Harrison 2008). More precisely, the material described herein, as well as all fossil specimens collected by Roy Gardner (from Fareham, Hampshire) in the 1980’s originates from the Mammal Bed, which is the same site that yielded also some of the material collected by the Marchioness of Hastings during the 19th century (see Milner etal. 1982; Benton & Spencer famille. De plus, le nouveau taxon anglais présente une ressemblance frappante avec les acrochordidés existants, particulièrement avec l’espèce Acrochordus granulatus (Schneider, 1799). En conséquence, nous considérons que le nouveau taxon représente très probablement un caenophidien divergent précoce, pouvant même être un membre des Acrochordidae Bonaparte, 1831, bien en dehors de la distribution stratigraphique et géographique connue jusqu’à présent de ce dernier groupe. Il ajoute en outre à l’étonnante diversité des morphologies vertébrales des serpents européens du Paléogène. MOTS CLÉS Squamata, Caenophidia, Russellophiidae, Acrochordidae, morphologie vertébrale, genre nouveau, espèce nouvelle. 507 A new peculiar early diverging caenophidian snake COMPTES RENDUS PALEVOL • 2025 • 24 (25) 1995). The exact spot where the new snake taxon was found is in Christchurch Bay, on the Hampshire coast, in the stretch of the Hordle Cliff, between Becton Bunny in the west and Long Mead End in the east (see Milner etal. 1982). The age of the Mammal Bed locality of Hordle Cliff correspond to the MP 17a zone, i.e., with an age of approximately 36.91 ± 0.326 Ma (Biochrom’97 1997; Escarguel etal. 1997). For more information on the geology of the locality, see Milner etal. (1982) and Edwards & Daley (1997). SYSTEMATIC PALAEONTOLOGY SERPENTES Linnaeus, 1758 ALETHINOPHIDIA Nopcsa, 1923 CAENOPHIDIA Hoffstetter, 1939 Paradoxophidion n. gen. urn:lsid:zoobank.org:act:447C8D95-2A3D-4A4A-A41C-D068C458797A type specIes. — Paradoxophidion richardoweni n. gen., n. sp. DIagnosIs. — As for the type and only known species. etymology. — The new genus name derives from the Greek words “παράδοξος” (“paradoxos”) meaning “paradox”/“weird”, and “ὀφίδιον” (“ophidion”) meaning “snake”, in referral of the weird combination of vertebral features present in the new taxon. Gender of the new genus name is neuter. Paradoxophidion richardoweni n gen., n. sp. (Figs 1-14; Appendices 1; 2) urn:lsid:zoobank.org:act:DF7D9EEA-52E4-4E8E-8DFC-AC001D0D2E64 type materIal. — Holotype. England • 1 specimen (a trunk vertebra); Christchurch Bay, Hordle Cliff; Headon Hill Formation; 36.91 ± 0.326 Ma; MP 17, Priabonian (late Eocene); NHMUK PV R 10795 (Figs 1; 2). Paratype. England • 1 specimen (a trunk vertebra); Christchurch Bay, Hordle Cliff; Headon Hill Formation; 36.91 ± 0.326 Ma; MP 17, Priabonian (late Eocene); NHMUK PV R 10798 (Figs 3; 4). referreD specImens. — 18 trunk vertebrae (NHMUK PV R 10792, NHMUK PV R 10794, NHMUK PV R 10797, NHMUK PV R 38879, NHMUK PV R 38942, NHMUK PV R 38943, fig. 1. — Holotype trunk vertebra NHMUK PV R 10795 of Paradoxophidion richardoweni n. gen., n. sp. in anterior (A), posterior (B), left lateral (C), dorsal (D), and ventral (E) views. Scale bar: 1 mm. AB C DE 508 COMPTES RENDUS PALEVOL • 2025 • 24 (25) Georgalis G. L. & Jones M. E. H. NHMUK PV R 38944, NHMUK PV R 38945, NHMUK PV R 38946, NHMUK PV R 38950, NHMUK PV R 38952-NHMUK PV R 38954, and NHMUK PV R 38956-NHMUK PV R 38960), two ?cloacal vertebrae (NHMUK PV R 38947 and NHMUK PV R 38948), and nine caudal vertebrae (NHMUK PV R 10791, NHMUK PV R 10793, NHMUK PV R 10796, NHMUK PV R 10799, NHMUK PV R 10800, NHMUK PV R 38941, NHMUK PV R 38949, NHMUK PV R 38951, and NHMUK PV R 38955). type localIty anD age. — All specimens originate from the locality of Christchurch Bay at Hordle Cliff, Hampshire, England; Mammal Bed, Totland Bay Member, MP 17 (approximately 36.91 ± 0.326 Ma), late Eocene. geographIc anD stratIgraphIc range. — Taxon known exclusively from the type locality. DIagnosIs. — Paradoxophidion richardoweni n. gen., n. sp. is unique among all snakes in possessing the following characters: fig. 2. — 3D μCT images of holotype trunk vertebra NHMUK PV R 10795 of Paradoxophidion richardoweni n. gen., n. sp. in anterior (A), left anterodorsolateral (B), right anterolateral (C), right ventrolateral (D), dorsal (E), posterodorsal (F), left posterolateral (G), left lateral (H), posterior (I), posteroventral (J), posterodorsal (K), right lateral (L), ventral (M), anteroventral (N), right ventrolateral (O), and left posterolateral (P) views. Scale bar: 1 mm. AB CD EF GH IJ KL MN OP neural spine prezygapophyseal articular facet neural arch neural canal paradiapophysi s zygantral roof prezygapophyseal buttress postzygapophyseal articular facet subcentral foramen zygosphene condyle cotyle hypapophysis 509 A new peculiar early diverging caenophidian snake COMPTES RENDUS PALEVOL • 2025 • 24 (25) hypapophysis with a triangular shape in lateral view; vaulted neural arch with a peculiar dome-shaped arching; thick and transversely expanded zygantral roof (i.e., “thickened posterior margin of the neural arch” sensu Head 2021); and the contact of the paradiapophyses with the lateral margins of cotyle well dorsal to the level of its ventral margin. Paradoxophidion richardoweni n. gen., n. sp. can be further differentiated from all other snakes by the following combination of features: small vertebral size, with centrum length (CL) ranging around 2 mm; moderately (dorsoventrally) short neural spine that is almost strictly confined to the posteriormost portion of the neural arch; presence of hypapophysis across all trunk vertebrae, that is prominent and much ventrally expanded in lateral view and narrow in ventral view; presence of prominent prezygapophyseal buttresses that are compressed and form a more or less vertical ridge; presence of a small, pterapophysislike, tubercle on each side of the neural arch; paradiapophyses not clearly divided into diapophyses and parapophyses; paradiapophyses facing ventrally in anterior view and reaching below the ventral level of the cotyle; absence of paracotylar foramina; very small and circular cotyle and condyle; absence of paracotylar ventrolateral processes; occasional presence of a short median ventral process/ tubercle in the ventral margin of the cotyle; zygosphene very thin and always larger than the cotyle; very large neural canal; absence of parazygosphenal foramina; caudal vertebrae with elongated and slender pleurapophyses; anterior caudal vertebrae with a haemal keel; and mid-caudal and posterior caudal vertebrae with paired haemapophyses. For detailed comparisons with other snakes, see Discussion below. etymology. — The new species name honours Sir Richard Owen (1804-1892), one of the greatest palaeontologists and comparative anatomists of all time. Richard Owen was the first to describe fossil remains of snakes from Hordle Cliff, as well as one of the first researchers that provided a thorough documentation of the snake vertebral column, focusing on its anatomical structures and the intracolumnar variation. Richard Owen was also the founder of the Natural History Museum in London (1881), where all known specimens of the new taxon are permanently curated. The name also alludes to the Richard Owen Fund of the Palaeontographical Society that was awarded to the first author (GLG) in 2023 and enabled him to study the fossil specimens of this new taxon at NHMUK. DescrIptIon Holotype (NHMUK PV R 10795) The holotype trunk vertebra NHMUK PV R 10795 of Paradoxophidion richardoweni n. gen., n. sp. is practically complete, with only some erosion observed in its right paradiapophysis and (perhaps also) the dorsalmost part of the neural spine (Figs 1; 2; Appendix 1). In anterior view (Figs 1A; 2A), the zygosphene is very thin and is much wider than the cotyle. The dorsal roof of the zygosphene is slightly convex. The neural canal is very large. The prezygapophyses are only slightly dorsally inclined. Prominent prezygapophyseal buttresses are present, forming a somehow vertical ridge; these commence dorsally from the level of the prezygapophyses and terminating ventrally towards the level of the paradiapophyses. The cotyle is very small and circular. There are no paracotylar foramina. There are no paracotylar ventrolateral processes. The paradiapophyses fig. 3. — Paratype trunk vertebra NHMUK PV R 10798 of Paradoxophidion richardoweni n. gen., n. sp. in anterior (A), posterior (B), right lateral (C), dorsal (D), and ventral (E) views. Scale bar: 1 mm. ABC DE 510 COMPTES RENDUS PALEVOL • 2025 • 24 (25) Georgalis G. L. & Jones M. E. H. are distinct from the centrum and extend ventrally, below the level of the ventral lip of the cotyle; the paradiapophyses contact the lateral margins of the cotyle well above the level of its ventral margin. The hypapophysis is relatively thick and is very elongated, extending much ventrally. In posterior view (Figs 1B; 2I), the zygantral roof is very thick (i.e., “thickened posterior margin of the neural arch” sensu Head 2021). The neural arch is moderately vaulted, with a vaulting ratio (sensu Georgalis etal. 2021b) equal to 0.42, and possesses a peculiar dome-shaped arching. The condyle is small and circular. The paradiapophyses (only the left one is almost complete, with a slight breakage in its parapophyseal surface; the right one is eroded), are massive and they seem not to be divided into diapophyseal and parapophyseal portions. In dorsal view (Figs 1D; 2E), the base of the neural spine is strictly confined to the posterior half of the neural arch. The zygosphene possesses two distinct lateral lobes. The prezygapophyses are small, slightly extending anterolaterally. The prezygapophyseal articular facets are relatively small and oval. The interzygapophyseal constriction is moderately deep. A posterior median notch of the neural arch is almost absent. In ventral view (Figs 1E; 2M), the centrum is slightly wider than long (CL/NAW = 0.95). The hypapophysis extends across most of the midline of the centrum, commencing anteriorly from the base of the cotyle and terminating posteriorly well before the condyle. fig. 4. — 3D μCT images of paratype trunk vertebra NHMUK PV R 10798 of Paradoxophidion richardoweni n. gen., n. sp. in anterior (A), right anterolateral (B), left anterolateral (C), right lateral (D), left lateral (E), left anterolateral (F), right posterodorsolateral (G), right posteroventrolateral (H), dorsal (I), left dorsolateral (J), ventral (K), left ventrolateral (L), posterior (M), posterodorsal (N), posteroventral (O), and left posteroventrolateral (P) views. Scale bar: 1 mm. AB CD EF GH IJKL MNOP 511 A new peculiar early diverging caenophidian snake COMPTES RENDUS PALEVOL • 2025 • 24 (25) The condyle is situated in a distinct condylar neck. The subcentral grooves are deep. The postzygapophyses are moderately large and somewhat triangular. There are no prezygapophyseal accessory processes, however, small, russellophiid-like spines are present below the prezygapophyseal articular facets. In lateral view (Figs 1C; 2H, L), the neural spine is relatively dorsoventrally short, while it is slender and anteroposteriorly relatively short; it abruptly commences augmenting in height in the posterior half of the centrum and is much posteriorly inclined. A small, pterapophysis-like, tubercle is present posteriorly on each side of the neural arch. A large lateral foramen is situated below the interzygapophyseal ridges. The subcentral ridges are strongly convex. The hypapophysis is large and almost triangular; it commences anteriorly from the tip of the cotyle but it is mostly developed towards the posterior half of the centrum, where it becomes prominent and extends rather ventrally. Paratype and referred specimens – intracolumnar variation All available trunk vertebrae of Paradoxophidion richardoweni n. gen., n. sp. are rather small, with CL ranging around 2 mm or even less (Figs 3-9). The paratype (NHMUK PV R 10798; Figs 3; 4) and all referred specimens share with the holotype a rather similar vertebral morphology, though some differences can be observed, which can be attributed to intracolumnar variation. The centrum is anteroposteriorly short and relatively wide in all trunk vertebrae. Among them, notably, NHMUK PV R 38943 (Fig. 8K-O) is rather anteroposteriorly short and laterally wide, i.e., wider than every other available vertebra of the species (CL/NAW = 0.77). This shape is probably due to intracolumnar variation, as this vertebra (NHMUK PV R 38943) is also characterized by a much vaulted neural arch (vaulting ratio sensu Georgalis etal. 2021b equal to 0.48). Still though, even this vertebra has a prominent hypapophysis, while all other distinctive features (e.g. shape and thickness of the zygantrum, paradiapophyses being distinct from the centrum and ventrally expanded, posterior confinement and size of the neural spine) perfectly match the morphology of the remaining vertebrae. The zygosphene in trunk vertebrae most usually possesses two (more or less) distinct lateral lobes and no median lobe in dorsal view, while in anterior view, it is slender and slightly arched (e.g. NHMUK PV R 38944 [Fig. 9A, D] and the holotype [NHMUK PV R 10795; Figs 1A, D; 2A, E]) to very arched (e.g. NHMUK PV R 38942 [Fig. 8F], NHMUK PV R 38943 [Fig. 8K]). Nevertheless, there are cases, where a median lobe is present and the lateral lobes are incipient (NHMUK PV R 10797; Figs 7E; 8D). Although prezygapophyseal accessory processes are absent in other specimens, tall, blade-like fig. 5. — Trunk vertebra NHMUK PV R 38946 of Paradoxophidion richardoweni n. gen., n. sp. in anterior (A), posterior (B), left lateral (C), dorsal (D), and ventral (E) views. Scale bar: 1 mm. ABC DE 518 COMPTES RENDUS PALEVOL • 2025 • 24 (25) Georgalis G. L. & Jones M. E. H. fig. 13. — Anterior caudal vertebrae of Paradoxophidion richardoweni n. gen., n. sp.: A-E, NHMUK PV R 10799 in anterior (A), posterior (B), left lateral (C), dorsal (D), and ventral (E) views; F-J, NHMUK PV R 10800 in anterior (F), posterior (G), right lateral (H), ventral (I), and dorsal (J) views; K-L, NHMUK PV R 10791 in ventral (K) and posterior (L) views; M-Q, NHMUK PV R 10793 in anterior (M), posterior (N), left ventrolateral (O), dorsal (P), and ventral (Q) views; R-V, NHMUK PV R 38951 in anterior (R), posterior (S), right lateral (T), dorsal (U), and ventral (V) views. Scale bars: 1 mm. AB C D E F G H I J K L MN O P Q R ST U V short median ventral tubercle pleurapophysis R-V M-Q K, L F-J A-E 519 A new peculiar early diverging caenophidian snake COMPTES RENDUS PALEVOL • 2025 • 24 (25) DISCUSSION comparIsons wIth russellophIIDs Paradoxophidion n. gen. resembles Russellophiidae Rage, 1978a, an enigmatic group of aquatic or semiaquatic snakes, probably lying somewhere within caenophidians (Rage 1978a, 1984; Rieppel 1988; Rage etal. 2008; Head etal. 2016, 2022; McCartney & Seiffert 2016; Zaher etal. 2019; Smith & Georgalis 2022). The type genus of this family, Russellophis, is known from the early Eocene of France, Belgium, and India (Rage 1975, 1978a; Rage etal. 2008; Smith & Georgalis 2022), but other potential russellophiids are also known from the Late Cretaceous (Campanian) of Sudan, the early Eocene of France and Brazil, the middle Eocene of France, and the late Eocene of Egypt (Rage & Werner 1999; Rage 2008; McCartney & Seiffert 2016; Head etal. 2022; Smith & Georgalis 2022). Paradoxophidion n. gen. shares with russellophiids certain characteristic features, such as the moderately (dorsoventrally) short neural spine that is strictly confined to the posteriormost portion of the neural arch, the vaulted neural arch with a peculiar dome-shaped arching, the shape and thickness of the zygantral roof (i.e., thickened posterior margin of the neural arch), the presence of small spines below the prezygapophyseal articular facets, the prominent prezygapophyseal buttresses that are compressed and form a more or less vertical ridge, the paradiapophyses that are not clearly divided into diapophyses and parapophyses and that face ventrally in anterior view and reach below the ventral level of the cotyle, the absence of paracotylar foramina, and a very large neural canal (characters from Rage 1984; Rage etal. 2008). For comparative purposes, photographs of the holotype (MNHN.F.CB1603) and a referred vertebra (MNHN.F.CB1623) of Russellophis tenuis Rage, 1975, from the early Eocene (MP 8/9) type locality of Condé-en-Brie in the Paris Basin, France, are presented here in Figure 15. However, there are some important differences between the new taxon from Hordle and russellophiids. Most importantly, the English taxon possesses a prominent hypapophysis (particularly large, broad, and ventrally inclined in some vertebrae) instead of haemal keel across all trunk vertebrae, whereas in Russellophiidae hypapophyses are confined solely to the anterior trunk vertebrae and then substituted by haemal keels in succeeding trunk vertebrae (Rage 1984; Rage etal. 2008). Moreover, Paradoxophidion n. gen. lacks a supposedly diagnostic feature of Russellophis, i.e., the ventrolateral inclination of the prezygapophyseal articular facets in anterior view (which is otherwise visible in several [but not all] known specimens of Russellophis; see Rage 1975, 1984; Rage etal. 2008). Instead, in the new English taxon, the prezygapophyses are almost horizontal to very slightly dorsally inclined. Other major differences of Paradoxophidion n. gen. from russellophiids, include the elongation of the centrum and the distinctiveness of the subcentral ridges: Russellophis and other known russellophiid records are all characterized by a very elongated centrum, that is much longer than wide, and distinct subcentral ridges (Rage 1975, 1984; Rage etal. 2008; Georgalis etal. 2025; this paper, Fig. 15E, K). Instead, the centrum of Paradoxophidion n. gen. is relatively anteroposteriorly short and laterally wide, compared to any published russellophiids. Among published russellophiids, only perhaps the unnamed form from the late Eocene of Fayum, Egypt, has such an anteroposteriorly short and laterally narrow centrum (see McCartney & Seiffert 2016: fig. 7A). Finally, Paradoxophidion richardoweni n. gen., n. sp. is smaller than all known russellophiids: it has a CL around 2 mm, while this value is 3.6 mm in Russellophis tenuis from fig. 14. — Posterior caudal vertebrae of Paradoxophidion richardoweni n. gen., n. sp.: A-E, NHMUK PV R 38941 in anterior (A), posterior (B), right lateral (C), dorsal (D), and ventral (E) views; F-J) NHMUK PV R 38949 in anterior (F), posterior (G), left lateral (H), dorsal (I), and ventral (J) views. Scale bars: 1 mm. ABCD E F G HI J 520 COMPTES RENDUS PALEVOL • 2025 • 24 (25) Georgalis G. L. & Jones M. E. H. fig. 15. — Russellophis tenuis Rage, 1975 from the early Eocene (MP 8/9) type locality of Condé-en-Brie in the Paris Basin, France: A-F, holotype trunk vertebra (MNHN.F.CB1603) in anterior (A), posterior (B), left lateral (C), right (ventro)lateral (D), ventral (E), and dorsal (F) views; G-K, trunk vertebra (MNHN.F.CB1623) in anterior (G), posterior (H), right lateral (I), ventral (J), and dorsal (K) views (this specimen has been previously figured in Rage 1983: fig. 10). Scale bar: 2 mm. ABC DEF GHI JK 521 A new peculiar early diverging caenophidian snake COMPTES RENDUS PALEVOL • 2025 • 24 (25) the early Eocene of France and Belgium (Rage 1975, 1984), more than 4 mm in Russellophis crassus Rage, Folie, Rana, Singh, Rose & Smith, 2008, from the early Eocene of India (Rage etal. 2008), and 3.3 mm in Krebsophis thobanus Rage & Werner, 1999, from the Late Cretaceous (Campanian) of Sudan (Rage & Werner 1999; age after Head etal. 2022). Comparisons regarding caudal vertebral morphology among Paradoxophidion richardoweni n. gen., n. sp. and russellophiids are hard to make because of the scarcity of such descriptions for the latter group. As a matter of fact, the sole known caudal vertebrae for russellophiids are: one caudal vertebra of Russellophis tenuis described (but not figured) by Rage (1975) and two caudal vertebrae of Russellophis crassus described (but again not figured) by Rage etal. (2008) as possessing pleurapophyses and haemapophyses. In any case, the relatively high proportion of caudal vertebrae (compared to the available trunk vertebrae; i.e., 9 out of 31) of the new taxon in our NHMUK sample is interesting; it evokes the suggestion of Smith (2013) that high percentages of caudal vertebrae of one species in a single fossil locality denote snakes with longer tails, however, certainly more material is required in order to assess such proportions. comparIsons wIth other paleogene englIsh anD contInental european taxa Paradoxophidion richardoweni n. gen., n. sp. differs significantly from all other named snake taxa from Hordle Cliff and other late Eocene localities in England, i.e., Cadurceryx pearchi Holman, Harrison & Ward, 2006, Paraplatyspondylia batesi Holman & Harrison, 1998b, Hordleophis balconae Holman, 1996, Totlandophis thomasae Holman & Harrison, 1998a, Vectophis wardi Rage & Ford, 1980, Paleryx rhombifer Owen, 1850, and Headonophis harrisoni Holman, 1993. More particularly, Paradoxophidion richardoweni n. gen., n. sp. can be readily differentiated from Cadurceryx pearchi by the absence of complex structures on the caudal vertebrae (see Holman etal. 2006). Paradoxophidion richardoweni n. gen., n. sp. can be differentiated from Paraplatyspondylia batesi (type and only known species of the genus Paraplatyspondylia Holman & Harrison, 1998b) by its neural spine being confined solely to the posteriormost portion of the neural arch (unlike the dorsoventrally short but anteroposteriorly long neural spine of Paraplatyspondylia), the presence of hypapophysis instead of haemal keel on trunk vertebrae, and the prominent prezygapophyseal buttresses (see Holman & Harrison 1998a). Paradoxophidion richardoweni n. gen., n. sp. can be differentiated from Hordleophis balconae (type and only known species of the genus Hordleophis Holman, 1996a) by its neural spine being extremely small and confined solely to the posteriormost portion of the neural arch (unlike the dorsoventrally very high and anteroposteriorly long neural spine of Hordleophis), the presence of hypapophysis instead of haemal keel on trunk vertebrae, and the prominent prezygapophyseal buttresses (see Holman 1996). Moreover, the validity of Hordleophis needs to be further assessed as this taxon bears some resemblance with the genus Cadurcoboa Rage, 1978b, from the middle and late Eocene of France (Rage 1978b; Rage & Augé 2010), a resemblance that was already suggested by Rage & Augé (2010), mainly in reference to the high neural spine and the depressed neural arch. However, the sole available illustrations of the holotype trunk vertebra (MSUVP 1361) of Hordleophis balconae and of one (among the three) paratype trunk vertebrae (MSUVP 1362) presented in Holman (1996: figs 1, 2) hinder any definite conclusions. Even more, long before the establishment of Hordleophis, Milner etal. (1982) had listed material that they assigned to ?Cadurcoboa sp. from Hordle Cliff, however, this was never figured or described. Paradoxophidion richardoweni n. gen., n. sp. can be differentiated from Totlandophis thomasae (type species of Totlandophis Holman & Harrison, 1998a) by being much smaller and its neural spine being extremely small and confined solely to the posteriormost portion of the neural arch (unlike the dorsoventrally higher and anteroposteriorly much longer neural spine of Totlandophis thomasae), the presence of hypapophysis instead of haemal keel in trunk vertebrae, the much slenderer prezygapophyseal articular facets, and the less shallow posterior median notch of the neural arch (see Holman & Harrison 1998a). Paradoxophidion richardoweni n. gen., n. sp. can be differentiated from Vectophis wardi (type and only known species of the genus Vectophis Rage & Ford, 1980) by its neural spine being extremely small and confined solely to the posteriormost portion of the neural arch (unlike the dorsoventrally higher and anteroposteriorly much longer neural spine of Vectophis), the more depressed neural arch, the less shallow posterior median notch of the neural arch, the not so prominent subcentral grooves, the presence of hypapophysis instead of haemal keel across trunk vertebrae, the prominent prezygapophyseal buttresses, and the different shape of the zygantrum and, the less vaulted neural arch (see Rage & Ford 1980). Paradoxophidion richardoweni n. gen., n. sp. nov. can be differentiated from Paleryx rhombifer (sole currently recognized valid species of the genus Paleryx) according to its much smaller overall size, the shape and size of the neural spine that is confined solely to the posteriormost portion of the neural arch (instead of the much more prominent neural spine of Paleryx that is dorsoventrally higher and anteroposteriorly longer), the presence of hypapophysis instead of haemal keel, the prominent prezygapophyseal buttresses, the thinner zygosphene, and the thinner prezygapophyseal articular facets (see Georgalis etal. 2021b). Paradoxophidion richardoweni n. gen., n. sp. can be differentiated from Headonophis harrisoni (type species of the genus Headonophis Holman, 1993) by its shorter centrum, the paradiapophyses being distinct from the centrum and much ventrally directed, the neural spine being extremely small and confined solely to the posteriormost portion of the neural arch (unlike the dorsoventrally higher and anteroposteriorly longer neural spine of Headonophis), the shallow posterior median notch of the neural arch, the absence of paracotylar foramina, the less thick (and usually) dorsoventrally higher hypapophysis, the less dorsally inclined prezygapophyses, and the prominent prezygapophyseal buttresses (see Holman 1993). 522 COMPTES RENDUS PALEVOL • 2025 • 24 (25) Georgalis G. L. & Jones M. E. H. Nevertheless, it has to be highlighted that the sole published figure of the holotype trunk vertebra of Headonophis harrisoni (MSUVP 1342) is poorly illustrated (Holman 1993: fig. 1), the holotype trunk vertebra is rather incomplete, and the only supposedly referred material are four cloacal or caudal vertebrae that were briefly described in Holman etal. (2006) but never figured. Outside the English Eocene, Paradoxophidion richardoweni n. gen., n. sp. greatly differs from other Paleogene small snakes from continental Europe (with the exception of the abovediscussed russellophiids). More particularly, Paradoxophidion can be differentiated from Eoanilius europae Rage, 1974 (type species of the genus Eoanilius Rage, 1974), from the late Eocene of France, and Eoanilius oligocenicus Szyndlar, 1994, from the early Oligocene to Early Miocene of Germany, by being larger and dorsoventrally taller, and possessing a more slender zygosphene, hypapophyses throughout the trunk region, the prominent prezygapophyseal buttresses, the paradiapophyses being distinct from the centrum and much ventrally directed, the smaller prezygapophyseal articular facets, and a thinner neural spine in dorsal view that covers an even smaller area onto the neural arch (see Rage 1974, 1984). Paradoxophidion richardoweni n. gen., n. sp. can be differentiated from the two European species of the genus Dunnophis Hecht in McGrew etal., 1959, i.e., Dunnophis cadurcensis Rage, 1974, from the middle and late Eocene of France, and Dunnophis matronensis Rage, 1973, from the early and middle Eocene of France and Portugal, by its shorter centrum, its thinner neural spine in dorsal view that covers an even smaller area onto the neural arch, the presence of hypapophysis instead of haemal keel across trunk vertebrae, the paradiapophyses being distinct from the centrum and much ventrally directed, and the prominent prezygapophyseal buttresses (see Rage 1973, 1974, 1984; Augé etal. 1997; Rage & Augé 2003, 2010). Paradoxophidion richardoweni n. gen., n. sp. can be differentiated from Szyndlaria aureomontensis Rage & Augé, 2010, from the middle Eocene of France, and Cadurcoboa insolita Rage, 1978b, from the late Eocene of France, by its much smaller size, the shape and size of the neural spine that is confined solely to the posteriormost portion of the neural arch (instead of the much prominent and tall neural spine of Szyndlaria Rage & Augé, 2010, and Cadurcoboa Rage, 1978b), the more vaulted neural arch, the prominent hypapophyses, and the prominent prezygapophyseal buttresses (see Rage 1978b; Rage & Augé 2010). Paradoxophidion richardoweni n. gen., n. sp. can be differentiated from Falseryx neervelpensis Szyndlar, Smith & Rage, 2008, from the early Oligocene of Belgium, by the shape and size of the neural spine that is confined solely to the posteriormost portion of the neural arch (instead of the short but still anteroposteriorly long neural spine of Falseryx) and the paradiapophyses being distinct from the centrum and much ventrally directed (see Szyndlar etal. 2008). Paradoxophidion richardoweni n. gen., n. sp. can be differentiated from “Calamagras” gallicus Rage, 1977, from the early Eocene of France, by the shape and size of the neural spine that has its base commencing more posteriorly on the neural arch and being confined solely to the posteriormost portion of the neural arch, the presence of hypapophysis instead of haemal keel on trunk vertebrae, the less thick neural spine in caudal vertebrae, and the absence of complex structures on the caudal vertebrae (see Rage 1977). Paradoxophidion richardoweni n. gen., n. sp. can be readily differentiated from Cadurceryx filholi Hoffstetter & Rage, 1972, from the middle and late Eocene of France, and Rageryx schmidi Smith & Scanferla, 2021, from the latest early to earliest middle Eocene of Germany, by the absence of complex structures on its caudal vertebrae and a completely different morphology of the trunk vertebrae (see Hoffstetter & Rage 1972; Rage 2013; Smith & Scanferla 2021; Szyndlar & Georgalis in press). Paradoxophidion richardoweni n. gen., n. sp. can be readily differentiated from Rottophis atavus (Meyer, 1855), from the late Oligocene of Germany, by its smaller size, the shape and size of the neural spine that is confined solely to the posteriormost portion of the neural arch (instead of the short but still anteroposteriorly long neural spine of Rottophis Szyndlar & Böhme, 1996), the slenderer prezygapophyseal articular facets, the less dorsally inclined prezygapophyses, and the much thinner zygosphene (see Szyndlar & Böhme 1996). Paradoxophidion richardoweni n. gen., n. sp. can be differentiated from the three known species of Platyspondylia Rage, 1974, by the shape and size of the neural spine that is confined solely to the posteriormost portion of the neural arch (instead of the short but still anteroposteriorly long neural spine of Platyspondylia), the presence of hypapophyses across trunk vertebrae, and the paradiapophyses being distinct from the centrum and much ventrally directed (see Rage 1974, 1988a). Paradoxophidion richardoweni n. gen., n. sp. can be differentiated from Messelophis variatus Baszio, 2004, and Rieppelophis ermannorum (Schaal & Baszio, 2004), from the latest early to earliest middle Eocene of Germany, by the presence of hypapophyses across all trunk vertebrae and the more vaulted neural arch, its thinner zygosphene, and its neural spine that is confined to the posteriormost portion of the neural arch (see Baszio 2004; Schaal & Baszio 2004; Scanferla etal. 2016). Paradoxophidion richardoweni n. gen., n. sp. can be differentiated from Anomalophis bolcensis (Massalongo, 1859), from the early Eocene of Italy by its less elongated centrum, the neural spine that is shorter and confined solely to the posterior portion of the neural arch, the more massive and prominent paradiapophyses that are distinct from the centrum and much ventrally directed, and the presence of hypapophyses (instead of haemal keels) across all trunk vertebrae (see Auffenberg 1959). Paradoxophidion richardoweni n. gen., n. sp. can be differentiated from Woutersophis novus Rage, 1980, from the early Eocene of Belgium, by its not so elongated centrum, the presence of hypapophyses across trunk vertebrae, the much thinner zygosphene, and the much shorter neural spine that is confined is the posterior portion of the neural arch (see Rage 1980). Finally, other Paleogene taxa from Europe are much greater in size compared to Paradoxophidion n. gen. (e.g. members of the genera Palaeopython Rochebrune, 1880, Phosphoroboa Georgalis, Rabi & Smith, 2021b, Eoconstrictor Scanferla & Smith 2020, Messelopython Zaher & Smith, 523 A new peculiar early diverging caenophidian snake COMPTES RENDUS PALEVOL • 2025 • 24 (25) 2020, Bransateryx Hoffstetter & Rage, 1972, and Bavarioboa Szyndlar & Schleich, 1993; see Hoffstetter & Rage 1972; Rage 1984; Szyndlar & Rage 2003; Georgalis & Scheyer 2019; Scanferla & Smith 2020; Zaher & Smith 2020; Georgalis etal. 2021b; Smith & Scanferla 2022; Palci etal. 2024) and/or differ significantly in their vertebral morphology (e.g. palaeophiids, archaeophiids; see Rage 1984; Georgalis etal. 2020, 2021a; Georgalis 2023). taxonomIc affInItIes of ParadoxoPhidion richardoweni n. gen., n. sp. The new taxon Paradoxophidion richardoweni n. gen., n. sp. exhibits a mosaic of features that are present across distantly related groups of alethinophidian snakes. Even though certain features of Paradoxophidion n. gen. approach the morphology of russellophiids, there are also other features that preclude such taxonomic referral. It appears actually challenging to assign Paradoxophidion n. gen. to a particular group of snakes with confidence, as its unique vertebral morphology differs from known groups of either non-caenophidians (Szyndlar & Georgalis 2023) and caenophidians (see Hoffstetter & Gasc 1969; Zaher etal. 2019). The proportion of CL/NAW <1 could at first glance be indicative of constrictors (see Georgalis & Smith 2020; Szyndlar & Georgalis 2023) but this character is further observed across other groups of non-constrictor snakes (see Szyndlar & Georgalis 2023), including the caenophidian group of acrochordids (Hoffstetter & Gayrard 1964; Head 2005). Similarly, the presence of prominent hypapophyses across the trunk vertebral column is observed across an array of distantly related snake groups, including both noncaenophidian (e.g. tropidophiids, bolyeriids, candoiids, palaeophiids) and caenophidian (e.g. acrochordids, xenodermids, natricids, elapids, homalopsids, viperids) snakes (see Zaher etal. 2019; Szyndlar & Georgalis 2023). The paradiapophyses not clearly divided into diapophyseal and parapophyseal portions is more typical of non-caenophidian snakes (see Szyndlar & Georgalis 2023), as in caenophidians, paradiapophyses are most usually clearly divided into distinct diapophyses and parapophyses (Zaher etal. 2019); in fact, this very feature, coupled with the absence of prezygapophyseal accessory processes, had led Rage (1984) to conclude that russellophiids lie more basally within the lineage of Colubroides Zaher, Grazziotin, Cadle, Murphy, Cesar de Moura-Leite & Bonatto, 2009 (Colubroidea in his terminology). Nevertheless, as will be highlighted in detail below, the paradiapophyses of Paradoxophidion n. gen. are much strongly reminiscent of acrochordids, in terms of their shape and ventral extension. Szyndlar & Georgalis (2023) recently emphasized the taxonomic importance and utility of the subcentral structures of the vertebrae around the trunk to caudal transition (“pericloacal vertebrae” sensu Smith 2013). Paradoxophidion n. gen. exhibits a haemal keel (or “short hypapophysis”) in its posteriormost trunk, cloacal, and anterior caudal vertebrae, which ultimately becomes transitioned into paired short haemapophyses into succeeding posterior caudal vertebrae. A similar pattern can be observed in some extant snake groups, such as tropidophiids and some constrictors (Szyndlar & Georgalis 2023). Aside from the above discussed similarities to russellophiids, it should be highlighted that the greatest degree of resemblance between Paradoxophidion n. gen. and some particular snake group is shared with Acrochordidae Bonaparte, 1831. Acrochordids are a group of fully aquatic snakes, currently distributed across marine and freshwater areas stretching from India to the western edge of the Pacific Ocean (McDowell 1979; Sanders etal. 2010). The family comprises solely a single valid genus, Acrochordus Hornstedt, 1787, with only three extant species (Acrochordus arafurae McDowell, 1979, Acrochordus granulatus [Schneider, 1799], and the type species, Acrochordus javanicus Hornstedt, 1787; see McDowell 1979; Sanders etal. 2010) and a fossil record confined to the Neogene of southern Asia, including a single, massive, extinct species, Acrochordus dehmi Hoffstetter, 1964 ( Hoffstetter 1964; West etal. 1991; Rage etal. 2001; Head 2005; Head etal. 2007; Sanders etal. 2010; Kapur etal. 2021). The lineage likely represents one of the basalmost groups (if not the basalmost one) of extant caenophidians, being recovered as the sister group of Colubroides in recent phylogenies (Slowinski & Lawson 2002; Lawson etal. 2005; Lee etal. 2007; Vidal etal. 2007; Zaher etal. 2009, 2019, 2023; Sanders etal. 2010; Pyron etal. 2011, 2013; Hsiang etal. 2015; Figueroa etal. 2016; Streicher & Wiens 2016; Zheng & Wiens 2016; Burbrink etal. 2020). Accordingly, despite the relatively young so far existing fossil record, the lineage of Acrochordidae is considered rather old, with divergence date estimates varying between the early Eocene (56 Ma; Sanders etal. 2010; Zaher etal. 2019) and late Oligocene (Head etal. 2016). Acrochordids are characterized by extreme and unique anatomical features and life history traits, including distinctive features in their skull and vertebrae (Hoffstetter & Gayrard 1964; McDowell 1979; Rieppel & Zaher 2001; Head 2005; Head etal. 2007; Sanders etal. 2010). Paradoxophidion n. gen. shares with Acrochordus a number of important vertebral features: presence of hypapophyses throughout the trunk column; the paradiapophyses are distinct from the centrum, narrow, more ventrally than laterally oriented and extend well below the ventral margin of the cotyle; a large neural canal; presence of a small, pterapophysislike, tubercle on each side of the neural arch; the (occasional) presence of a short median ventral process/tubercle in the ventral margin of the cotyle; and prezygapophyseal accessory processes consisting of vertically oriented blades, as observed in at least one specimen (NHMUK PV R 10797) of the new English taxon (Hoffstetter & Gayrard 1964; McDowell 1979; Rage 1984; Head 2005; Head etal. 2007; Sanders etal. 2010); for further comparative purposes, we present here the intracolumnar variation of the trunk vertebrae of a skeleton (ISEA R/497) of Acrochordus arafurae (Fig. 16). Moreover, the size, cross-sectional shape, and orientation of the reduced neural spine of Paradoxophidion n. gen. possesses a striking similarity to that of Acrochordus granulatus (see Hoffstetter & Gayrard 1964: fig. 6), a species that was for long time placed in its own genus, Chersydrus Cuvier, 1817, and which possesses significant vertebral differences from other Acrochordus spp. (most notably, the absence 524 COMPTES RENDUS PALEVOL • 2025 • 24 (25) Georgalis G. L. & Jones M. E. H. of parazygosphenal foramina in A. granulatus; Hoffstetter & Gayrard 1964; Rage 1984; Head 2005; Sanders etal. 2010). It is further worth highlighting that two striking shared features of Paradoxophidion n. gen. and Acrochordus, i.e., the presence of small, pterapophysis-like, tubercles on the neural arch and the occasional presence of a short median ventral process/tubercle in the ventral margin of the cotyle (potentially serving as an attachment site for perilymphatic ligaments), were interpreted by McDowell (1979) as being homologous features with the pterapophyses and anterior hypapophyses of palaeophiids. These being said, it seems rather likely that Paradoxophidion represents a basal caenophidian (Fig. 17). It is even possible that the English taxon belongs to Acrochordidae, as an early member of the group, much away from its so far known stratigraphic and geographic (extant and fossil) distribution. Nevertheless, we cannot exclude the alternative possibility that it belongs to some other group of early diverging caenophidians. Among extant taxa, the subsequently early diverging caenophidians after acrochordids according to most phylogenetic analyses (e.g. Zaher etal. 2019), are xenodermids and pareids, both of which exhibit much different fig. 16. — Intracolumnar variation of trunk vertebrae in the extant acrochordid Acrochordus arafurae McDowell, 1979 (specimen ISEA R/497): A, 41st vertebra; B, 71st vertebra; C, 116th vertebra; D, posterior trunk (158th) vertebra. Views in each row correspond to anterior, posterior, right lateral, dorsal, and ventral respectively. Scale bars: 2 mm. A B C D 525 A new peculiar early diverging caenophidian snake COMPTES RENDUS PALEVOL • 2025 • 24 (25) vertebral morphology than Paradoxophidion n. gen. (particularly pareids, which even lack hypapophyses throughout the trunk column; Zaher etal. 2019). In any case, the discovery of Paradoxophidion n. gen. as an early diverging caenophidian in the Eocene of England, is of particular significance, as only a few Eocene taxa from Europe have been regarded as true members of Caenophidia (Head etal. 2016; Zaher etal. 2019; Smith & Georgalis 2022). These records are the above discussed Russellophis from the early Eocene of France and Belgium (Rage 1975, 1983) and Headonophis from the late Eocene of England (Holman 1993; Holman etal. 2006), plus the genera Anomalophis Auffenberg, 1959, from the early Eocene of Italy (Auffenberg 1959; Seghetti etal. 2022), and potentially also, Vectophis from the late Eocene of England (Rage & Ford 1980; Zaher etal. 2019) and the nigerophiid Woutersophis Rage, 1980 from the early Eocene of Belgium (Rage 1980, 1983). In addition, there are indeterminate finds of russellophiids from the middle Eocene (MP 16) of Le Bretou, Quercy, France (Rage 1988a), and an indeterminate caenophidian from the early Eocene (MP 10) of Prémontré, France (Augé etal. 1997). In the “nearby” area of northern Africa, Eocene caenophidians are also known, represented by cf. Procerophis sp. from Fayum, Egypt (El-Hares etal. 2022), a genus otherwise known from the early Eocene of India (Rage etal. 2008), and the thaumastophiid Renenutet McCartney & Seiffert, 2016, also from Fayum (McCartney & Seiffert 2016; Zaher etal. 2021; El-Hares etal. 2022); in addition, an as yet unnamed colubriform has been described from the early-middle Eocene of Glib Zegdou HGL50 ( Hammada Gour Lazib50), Algeria (Rage etal. 2021). Besides these records, a few other Eocene caenophidian remains are known from other continents (Zaher etal. 2021; Smith & Georgalis 2022). Despite this low number of fossil occurrences, it is evident that Eocene caenophidians had already achieved some relative diversity, as it is attested by the disparity of vertebral morphologies, sometimes combining a mosaic of features that is otherwise observed across different extant caenophidian groups (Head etal. 2016; Zaher etal. 2019, 2021). It is further worth noticing that the area of England seems to be an “epicenter” of this Eocene caenophidian diversity (but this fig. 17 . — Hypothetical life reconstruction of Paradoxophidion richardoweni n. gen., n. sp. Credits: Artwork by Jaime Chirinos. 526 COMPTES RENDUS PALEVOL • 2025 • 24 (25) Georgalis G. L. & Jones M. E. H. could also well reflect sampling biases), as three different taxa are now known from the region (Vectophis, Headonophis, and Paradoxophidion n. gen.); however, as highlighted above, we cast doubt on the status and even the validity of Headonophis, while we should leave open the possibility that Vectophis could alternatively represent some non-caenophidian snake, as it has been suggested also elsewhere (e.g. Rage & Ford 1980; Rage 1984; Rage etal. 2008; Smith & Georgalis 2022). The imprecise affinities of the new taxon from Hordle Cliff within Caenophidia hinder direct biogeographic inferences. Nevertheless, if Paradoxophidion n. gen. belongs indeed to Russellophiidae, then the new English taxon, with an age of MP 17a (i.e., 37.8-37.5 Ma), would represent the last occurrence of the European continent, as Russellophiidae is supposed to have become extinct in Europe around the middle Eocene (MP 16 [40-37.8 Ma], locality of Le Bretou, France; Rage 1988b). It could also represent the most recent global record, with only potential exception being the unnamed form from Fayum, Egypt, which has an age approximately 37 Ma (McCartney & Seiffert 2016). If Paradoxophidion n. gen. belongs to Acrochordidae, it would then support divergence date estimates that place the split of that group already by the Eocene and moreover indicate a rather extraordinary geographic expansion for acrochordids, which are otherwise known exclusively from the Neogene and Quaternary of southern Asia and the western Pacific Ocean. Furthermore, the identification of Paradoxophidion n. gen. casts some doubt on the supposed occurrence of Russellophis from England. Indeed, the presence of a caenophidian resembling Russellophis from the late Eocene of Christchurch Bay at Hordle Cliff, England was first reported in Milner etal. (1982), but that was apparently based on the same NHMUK material here designated as the new taxon Paradoxophidion richardoweni n. gen., n. sp. Subsequently, Milner (1986) mentioned such supposed presence of Russellophis sp. from Christchurch Bay at Hordle Cliff, but also from two additional English localities: the middle Eocene of Huntingbridge Division and the latest Eocene/earliest Oligocene (MP 20/21) of the Bembridge Limestone (Isle of Wight), however, again all these occurrences were not accompanied by any kind of figure, description, or collection number that could verify these records and identifications. Eventually, Holman etal. (2006), mentioned (but not figured) one further anterior trunk vertebra (MSUVP 2065) from the Hordle Cliff, which he considered as identical to Russellophis tenuis. Furthermore, Holman etal. (2006) erroneously mentioned that the English localities from which Milner (1986) had reported Russellophis were the middle Eocene (MP 16) of Creechbarrow in Dorset and the Lower Headon Beds (i.e., Hordle Cliff). Moreover, Milner (1986) mentioned the presence of a form dubbed as “Acrochordid 1” from the Lower Headon Beds, however, this was not accompanied by any kind of description, figure or collection number. We here cannot comment about the taxonomic status of the other supposed British occurrences of russellophiids and/ or acrochordids mentioned in Milner (1986) and Holman etal. (2006), but at least for the case of the material from Hordle Cliff from the NHMUK collection, it does not belong to Russellophis but instead represents the new taxon Paradoxophidion richardoweni n. gen., n. sp. Finally, Paradoxophidion represents an important addition to our knowledge of Eocene faunas from England, which have been predominantly studied for their mammals (e.g. Hooker etal. 2005; Vasileiadou etal. 2022), and offers a small step forwards in tracing reptile communities and diversity across the Paleogene and especially around the “Grande Coupure” at the Eocene-Oligocene transition (Hooker 2010), and how they may have changed due to ecological and environmental factors and perturbations. Acknowledgements GLG acknowledges funding from the research project no. 2023/49/B/ST10/02631 financed by the National Science Center of Poland (Narodowe Centrum Nauki). GLG also acknowledges travel support from the 2023 Richard Owen Fund of the Palaeontographical Society, which enabled him to travel to London and study the NHMUK collection. We thank Agnese Lanzetti and Brett Clark for μCT scanning five of the specimens of the new taxon; Andrew Milner and Jerry Hooker for encouragement and discussion; Olaf Lizak for help with photographing the Acrochordus specimen from the ISEA collection; and Brett Clark for advice on videos of μCT models. For access to comparative material of Russellophis, Vectophis, and other snakes at the collections of MNHN, GLG is grateful to Nour-Eddine Jalil and acknowledges travel support from SYNTHESYS FR-TAF_Call4_035 (MNHN) that enabled him to study these specimens. We also thank Jaime Chirinos for the life reconstruction of the new taxon, presented in Figure 17. We finally thank the reviewers, Krister Smith, Jason Head, and two anonymous ones, for their useful comments and suggestions, which enhanced the quality of the paper. REFERENCES auffenberg w. 1959. — Anomalophis bolcensis (Massalongo), a new genus of fossil snake from the Italian Eocene. Breviora 114: 1-16. https://www.biodiversitylibrary.org/page/3190823 a ugé m., D uffauD s., l apparent De b roIn f. De , r age J.-c. & v asse D. 1997. — Les amphibiens et les reptiles de Prémontré (Cuisien, Bassin parisien) : une herpétofaune de référence pour l’Eocène inférieur. Géologie de la France 1 (2): 23-33. baszIo s. 2004. — Messelophis variatus n. gen. n. sp. from the Eocene of Messel: a tropidopheine snake with affinities to Erycinae (Boidae). Courier Forschungsinstitut Senckenberg 252: 47-66. benton m. J. & spencer p. s. 1995. — Fossil Reptiles of Great Britain. Geological Conservation Review Series 10. Chapman and Hall, London, 386 p. https://doi.org/10.1007/978-94-011-0519-4 b Iochrom ’97 1997. — Synthèses et Tableaux de corrélations, in aguIlar J.-p., legenDre s. & mIchaux J. (eds), Actes du Congrès BiochroM’97, Montpellier, 14-17 avril: biochronologie mammalienne du cénozoïque en Europe et domaines reliés. École pratique des hautes études, Institut de Montpellier (Mémoires et travaux de l’Institut de Montpellier; 21), Montpellier: 768-805. 527 A new peculiar early diverging caenophidian snake COMPTES RENDUS PALEVOL • 2025 • 24 (25) bonaparte c. l. 1831. — Saggio di una distribuzione metodica degli animali vertebrati. Giornale Arcadico di Scienze, Lettere, ed Arti, Roma 49: 3-77. https://doi.org/10.5962/bhl.title.48624 burbrInk f. t., grazzIotIn f. g., pyron r. a., cunDall D., Donnellan s., IrIsh f., keogh J. s., kraus f., murphy r. w., noonan b., raxworthy c. J., ruane s., lemmon a. r., lemmon e. m. & zaher h. 2020. — Interrogating genomic-scale data for Squamata (lizards, snakes, and amphisbaenians) shows no support for key traditional morphological relationships. Systematic Biology 69 (3): 502-520. https://doi.org/10.1093/sysbio/syz062 eDwarDs n. & Daley b. 1997. — Stratigraphy of the Totland Bay Member (Headon Hill Formation, Late Eocene) at Hordle Cliff, Hampshire, southern England. Tertiary Research 18: 35-50. el-hares m. a., zaher h., el-mekkawy D., el-sayeD s., seIffert e. r. & sallam h. m. 2022. — New records of legless squamates from the lowest upper Eocene deposits of the Fayum Depression, Egypt. Journal of Vertebrate Paleontology 41 (4): e1992770. https://doi.org/10.1080/02724634.2021.1992770 escarguel g., maranDat b. & legenDre s. 1997. — Sur l’âge numérique des faunes de mammifères du Paléogène d’Europe Occidentale, en particulier celles de l’Éocène inférieur et moyen. in aguIlar J.-p., legenDre s. & mIchaux J. (eds), Actes du Congrès BiochroM’97, Montpellier, 14-17 avril: biochronologie mammalienne du cénozoïque en Europe et domaines reliés. École pratique des hautes études, Institut de Montpellier (Mémoires et travaux de l’Institut de Montpellier; 21), Montpellier: 442-460. fIgueroa a., mckelvy a. D., grIsmer l. l., bell c. D. & laIlvaux s. p. 2016. — A species-level phylogeny of extant snakes with description of a new colubrid subfamily and genus. PLoS ONE 11 (9): e0161070. https://doi.org/10.1371/journal.pone.0161070 georgalIs g. l. 2023. — First potential occurrence of the large aquatic snake Pterosphenus (Serpentes, Palaeophiidae) from Nigeria, with further documentation of Pterosphenus schweinfurthi from Egypt. Alcheringa 47 (3): 327-335. https://doi.org/10.1080/03 115518.2023.2217874 georgalIs g. l. & Joyce w. g. 2017. — A review of the fossil record of Old World turtles of the clade Pan-Trionychidae. Bulletin of the Peabody Museum of Natural History 58 (1): 115-208. https://doi.org/10.3374/014.058.0106 g eorgalIs g. l. & s cheyer t. m. 2019. — A new species of Palaeopython (Serpentes) and other extinct squamates from the Eocene of Dielsdorf (Zurich, Switzerland). Swiss Journal of Geosciences 112: 383-417. https://doi.org/10.1007/s00015-019-00341-6 georgalIs g. l. & smIth k. t. 2020. — Constrictores Oppel, 1811 – the available name for the taxonomic group uniting boas and pythons. Vertebrate Zoology 70 (3): 291-304. https://doi. org/10.26049/VZ70-3-2020-03 georgalIs g. l., Del favero l. & DelfIno m. 2020. — Italy’s largest snake: Redescription of Palaeophis oweni from the Eocene of Monte Duello, near Verona. Acta Palaeontologica Polonica 65 (3): 523-533. https://doi.org/10.4202/app.00711.2019 georgalIs g. l., guInot g., kassegne k. e., amouDJI y. z., Johnson a. k. c., cappetta h. & hautIer l. 2021a. — An assemblage of giant aquatic snakes (Serpentes, Palaeophiidae) from the Eocene of Togo. Swiss Journal of Palaeontology 140 (20), 18 p. https://doi.org/10.1186/s13358-021-00236-w georgalIs g. l., rabI m. & smIth k. t. 2021b. — Taxonomic revision of the snakes of the genera Palaeopython and Paleryx (Serpentes, Constrictores) from the Paleogene of Europe. Swiss Journal of Palaeontology 140 (18), 140 p. https://doi.org/10.1186/ s13358-021-00224-0 georgalIs g. l., zaher h. & laurIn m. 2025. — Introduction to: Snakes from the Cenozoic of Europe – towards a macroevolutionary and palaeobiogeographic synthesis, in g eorgalIs g. l., z aher h. & l aurIn M. (eds), Snakes from the Cenozoic of Europe – towards a macroevolutionary and palaeobiogeographic synthesis. Comptes Rendus Palevol 24 (3): 45-49. https://doi. org/10.5852/cr-palevol2025v24a3 heaD J. J. 2005. — Snakes of the Siwalik Group (Miocene of Pakistan): systematics and relationship to environmental change. Palaeontologia Electronica 8.1.18A: 1-33. heaD J. J. 2021. — A South American snake lineage from the Eocene Greenhouse of North America and a reappraisal of the fossil record of ‘‘anilioid” snakes. Geobios 66-67: 55-65. https:// doi.org/10.1016/j.geobios.2020.09.005 heaD J. J., mohabey D. m. & wIlson J. a. 2007. — Acrochordus Hornstedt (Serpentes, Caenophidia) from the Miocene of Gujarat, Western India: Temporal constraints on dispersal of a derived snake. Journal of Vertebrate Paleontology 27 (3): 720-723. https://doi.org/10.1671/0272-4634(2007)27[72 0:AHSCFT]2.0.CO;2 heaD J. J., mahlow k. & müller J. 2016. — Fossil calibration dates for molecular phylogenetic analysis of snakes 2: Caenophidia, Colubroidea, Elapoidea, Colubridae. Palaeontologia Electronica 19.2.2FC: 1-21. https://doi.org/10.26879/625 h eaD J. J., h owarD a. f. c. & m üller J. 2022. — The first 80 million years of snake evolution: the Mesozoic fossil record of snakes and its implications for origin hypotheses, biogeography, and mass extinction, in gower D. & z aher H. (eds), The Origin and Early Evolution of Snakes. Cambridge University Press, Cambridge: 26-54. https://doi. org/10.1017/9781108938891.005 h echt m. k. 1959. — Amphibians and Reptiles, in m cgrew p. o., berman J. e., hecht m. k., hummel J. m., sImpson g. g. & wooD A. E. (eds), The geology and paleontology of the Elk Mountain and Tabernacle Butte area, Wyoming. Bulletin of American Museum of Natural History 117: 130-146. h offstetter r. 1939. — Contribution à l’étude des Elapidæ actuels et fossiles et de l’ostéologie des Ophidiens. Archives du Muséum d’Histoire Naturelle de Lyon 15: 1-78. h offstetter r. 1964. — Les Serpents du Néogène du Pakistan (couches des Siwaliks). Bulletin de la Société Géologique de France, 7esérie 6 (4): 467-474. https://doi.org/10.2113/gssgf - bull.S7-VI.4.467 h offstetter r. & g asc J. 1969. — Vertebrae and ribs of modern reptiles, in gans c., bellaIrs a. D’a. & parsons T. S. (eds), Biology of the Reptilia. Volume 1. Morphology A. Academic Press, London and New York: 201-310. hoffstetter r. & gayrarD y. 1964. — Observations sur l’ostéologie et la classification des Acrochordidae (Serpentes). Bulletin du Muséum national d’Histoire Naturelle 36: 677-696. hoffstetter r. & rage J.-c. 1972. — Les Erycinæ fossiles de France (Serpentes, Boidæ). Compréhension et histoire de la sous-famille. Annales de Paléontologie 58: 81-124. holman J. a. 1993. — A new genus of primitive colubroid snake from the Upper Eocene, Isle of Wight, England. Tertiary Research 14: 151-154. h olman J. a. 1996. — A new genus of diminutive boid snake from the upper Eocene of Hordle Cliff, Hampshire, England. Tertiary Research 17: 11-13. holman J. a. & harrIson D. l. 1998a. — A new genus of small boid snake from the upper Eocene of Hordle Cliff, Hampshire, England. Acta Zoologica Cracoviensia 41: 29-33. holman J. a. & harrIson D. l. 1998b. — A new genus of snake (Serpentes: Boidae) from the Upper Eocene of Hordle Cliff, Hampshire, England. Acta Zoologica Cracoviensia 41: 23-27. h olman J. a., h arrIson D. l. & w arD D. J. 2006. — Late Eocene snakes from the Headon Hill Formation, southern England. Cainozoic Research 5: 51-62. hooker J. J. 2010. — The ‘Grande Coupure’ in the Hampshire Basin, UK: taxonomy and stratigraphy of the mammals on either side of this major Palaeogene faunal turnover, in whIttaker J. e. & hart M. B. (eds), Micropalaeontology, sedimentary environments and stratigraphy: a tribute to Dennis Curry (1912–2001). The Geological Society of London, London:147-215. https:// doi.org/10.1144/TMS004.8