Non-technical Summary
The Indian Eocene snake fauna is of great importance for understanding the nature of snake diversification on the Indian subcontinent after the Cretaceous-Palaeogene boundary (K-Pg) approximately 66 million years ago. This boundary represents a dramatic time interval in India’s geological history because it witnessed one of the most extensive episodes of volcanic activity—the Deccan Traps volcanism. The Indian subcontinent was still drifting northward during this interval before colliding with Asia. Here, we describe a diverse snake fauna from early and middle Eocene of India (between 42 and 56 million years before present), comprising both aquatic (palaeophiid and nigerophiid) and giant terrestrial/semiaquatic (madtsoiid) snakes. The palaeophiids include Pterosphenus rannensis n. sp. and represent the geologically youngest record of the family Palaeophiidae from the subcontinent. Additionally, a previously described middle Eocene species, Pterosphenus biswasi, is reexamined here and retained as a valid taxon. The palaeophiid snakes described here suggest a considerable diversity of habitats occupied by these fossil snakes. The new finds also highlight the importance of the Indian fossil record in understanding the origin and diversification of the genus Pterosphenus. The madtsoiid snake described here lived alongside the recently described gigantic madtsoiid snake Vasuki indicus. The nigerophiid snake described here is among the oldest known post-K-Pg extinction record of the family Nigerophiidae globally.
Introduction
The Eocene strata in the Gujarat state of western India preserve a rich diversity of fossil snakes (Rage et al., Reference Rage, Bajpai, Thewissen and Tiwari2003, Reference Rage, Folie, Rana, Singh, Rose and Smith2008; Bajpai and Head, Reference Bajpai and Head2007; Smith et al., Reference Smith, Kumar, Rana, Folie, Solé, Noiret, Steeman, Sahni and Rose2016; Smith and Georgalis, Reference Smith, Georgalis, Gower and Zaher2022; Datta and Bajpai, Reference Datta and Bajpai2024). When seen in conjunction with the Late Cretaceous record (Wilson Mantilla et al., Reference Wilson Mantilla, Mohabey, Peters and Head2010; Rage et al., Reference Rage, Prasad, Verma, Khosla, Parmar, Prasad and Patnaik2020), the Paleogene snake fauna has the potential to provide significant insights into the biotic effect of India’s geographic isolation during its northward drift and its subsequent collision with Asia (Chatterjee et al., Reference Chatterjee, Scotese and Bajpai2017). The influence on contemporary biota is especially significant from the perspectives of: (1) possible endemism during India’s physical isolation and its evolving faunal affinities from essentially Gondwanan to cosmopolitan as a result of the Indian-Asia collision, and (2) the nature of faunal diversification and turnover across the interval of Deccan volcanic activity at the Cretaceous-Paleogene (K-Pg) boundary.
In India, fossil snakes predating the K-Pg mass extinction are exclusively Maastrichtian in age and known from the terrestrial Lameta Formation and intertrappean beds associated with the Deccan Trap volcanics (Jain and Sahni, Reference Jain, Sahni and Maheshwari1983; Gayet et al., Reference Gayet, Rage, Rana, Buffetaut, Jaeger and Rage1984; Prasad and Sahni, Reference Prasad and Sahni1987; Rage and Prasad, Reference Rage and Prasad1992; Prasad and Rage, Reference Prasad and Rage1995; Rage et al., Reference Rage, Prasad and Bajpai2004, Reference Rage, Prasad, Verma, Khosla, Parmar, Prasad and Patnaik2020; Head et al., Reference Head, Howard, Müller, Gower and Zaher2022). This fauna is represented by both stem snakes (Coniophis sp., Sanajeh indicus Wilson Mantilla et al., Reference Wilson Mantilla, Mohabey, Peters and Head2010; Rage et al., Reference Rage, Prasad and Bajpai2004; Wilson Mantilla et al., Reference Wilson Mantilla, Mohabey, Peters and Head2010; Zaher et al., Reference Zaher, Mohabey, Grazziotin and Wilson Mantilla2023) and members of the crown group Serpentes, e.g., madtsoiids (e.g., Madtsoia pisdurensis Mohabey et al., Reference Mohabey, Head and Wilson Mantilla2011) and nigerophiids (Indophis sahnii Rage and Prasad, Reference Rage and Prasad1992; Rage et al., Reference Rage, Prasad and Bajpai2004, Reference Rage, Prasad, Verma, Khosla, Parmar, Prasad and Patnaik2020). The post-K-Pg fauna in India is characterized by the appearance of palaeophiids—e.g., Palaeophis vastaniensis (Bajpai and Head, Reference Bajpai and Head2007); Pterosphenus kutchensis Rage et al., Reference Rage, Bajpai, Thewissen and Tiwari2003; Pterosphenus schucherti Lucas, Reference Lucas1898 (Rage et al., Reference Rage, Bajpai, Thewissen and Tiwari2003; Bajpai and Head, Reference Bajpai and Head2007; Natarajan et al., Reference Natarajan, Dasgupta, Rakshit and Kashyap2024)—constrictors (Rage et al., Reference Rage, Folie, Rana, Singh, Rose and Smith2008; see Pyron et al., Reference Pyron, Reynolds and Burbrink2014; Georgalis and Smith, Reference Georgalis and Smith2020), and caenophidians (colubroideans and Thaumastophis Rage et al., Reference Rage, Folie, Rana, Singh, Rose and Smith2008; Zaher et al., Reference Zaher, Folie, Quadros, Rana, Kumar, Rose, Fahmy and Smith2021), along with holdovers from the Cretaceous (madtsoiids, e.g., Platyspondylophis Smith et al., Reference Smith, Kumar, Rana, Folie, Solé, Noiret, Steeman, Sahni and Rose2016 and Vasuki Datta and Bajpai, Reference Datta and Bajpai2024). These taxa are almost exclusively from the early Eocene (Ypresian) Cambay Shale Formation and from the early middle Eocene (Lutetian) horizons of the Naredi Formation (Rage et al., Reference Rage, Bajpai, Thewissen and Tiwari2003, Reference Rage, Folie, Rana, Singh, Rose and Smith2008; Datta and Bajapi, Reference Datta and Bajpai2024). Conversely, fossil records from the Paleocene and the later parts of the Eocene are poor or nonexistent and include the recently described Pterosphenus schucherti from the Harudi Formation that constitutes the sole record from the late middle Eocene of India. Younger records include those from the late Oligocene of Ladakh molasse (e.g., Wasir et al., Reference Wazir, Sehgal, Čerňanský, Patnaik, Kumar, Singh, Uniyal and Singh2021).
The present study describes new material collected from the early and middle Eocene of Kutch (= Kachchh), Gujarat, western India. This discovery provides new insights into the diversity and phylogenetic relationships of the Indian Eocene snakes. The taxa described in this study significantly improve our understanding of the diversity and pattern of faunal turnover across the K-Pg boundary in India.
Geological setting
The Kutch Basin is a pericratonic rift basin situated along the western continental margin of India and includes one of the most well-preserved marine Tertiary sequences on the subcontinent (Biswas, Reference Biswas1992; Catuneanu and Dave, Reference Catuneanu and Dave2017). These Tertiary sequences are exposed as crescentic belts in the southwestern region of the Kutch mainland along the periphery of the Mesozoic highs. The Paleogene succession of Kutch includes the Matanomadh, Naredi, Harudi, Fulra limestone, and Maniyara Fort formations, in successive order (Biswas, Reference Biswas1992). The present study deals with fossil snakes recovered from the Naredi and Harudi formations (Fig. 1).
Geological map of Kutch Basin (after Biswas, Reference Biswas1992) showing snake fossil yielding localities. Light purple patches represent discontinuous outcrops of the Harudi Formation.
Within the Naredi Formation, fossils were collected from the type section and the lignite-bearing Panandhro area of western Kutch. The type section (23°34.602’N, 68°38.628’E), exposed in the cliffs along the Kakdi River, rests unconformably on weathered basalts of the Deccan Traps (Biswas, Reference Biswas1992). It includes three members, the basalmost of which is a gypsiferous Shale Member comprising glauconitic shale and claystone, succeeded by a Ferruginous Claystone Member, and the Assilina-bearing Limestone Member on top. The snake-yielding horizon in the type section of the Naredi Formation consists of grayish green silty clays and is located 2 m above the base of the section. This horizon is Ypresian in age and ~1 m below the interval dated as SBZ8 corresponding to planktic foraminiferal zone P6 (= E3−E4; Serra-Kiel et al., Reference Serra-Kiel, Hottinger, Caus, Drobne and Ferrandez1998; Keller et al., Reference Keller, Khozyem, Saravanan, Adatte, Bajpai and Spangenberg2013). Based on small benthic and rare planktic foraminiferans, a normal marine inner-shelf environment was suggested for this horizon (Keller et al., Reference Keller, Khozyem, Saravanan, Adatte, Bajpai and Spangenberg2013).
Elsewhere, the snake-yielding unit in the lignite-bearing succession of the Panandhro area of Kutch (23°41.897’N, 68°46.531’E) lies above the top lignite seam. Besides snakes (Pterosphenus Lucas, Reference Lucas1898; Madtsoiidae/Boidae), this unit previously produced a diverse array of taxa including rays and sharks (Myliobatis Cuvier, Reference Cuvier1816; Galeocerdo Müller and Henle, Reference Müller and Henle1837), bony fishes (e.g., the catfish Arius Cuvier and Valenciennes, Reference Cuvier and Valenciennes1840), crocodilians, and archaic whales (Andrewsiphius Sahni and Mishra, Reference Sahni and Mishra1975; Kutchicetus Bajpai and Thewissen, Reference Bajpai and Thewissen2000) (Bajpai and Thewissen, Reference Bajpai and Thewissen2002; Rage et al., Reference Rage, Bajpai, Thewissen and Tiwari2003). Initially this unit was thought to be early Eocene (Ypresian) in age following Biswas (Reference Biswas1992), but more recent investigations proposed an early middle Eocene (early Lutetian) age for the snake-bearing strata above the topmost lignite seam based on age-diagnostic palynomorphs and dinoflagellate cysts (Agrawal et al., Reference Agrawal, Verma, Rao, Garg, Kapur and Bajpai2017), a proposal which was followed later by Datta and Bajpai (Reference Datta and Bajpai2024). A back-swamp environment has been suggested for this unit and the lignite succession (Mukhopadhyay and Shome, Reference Mukhopadhyay and Shome1996; Agrawal et al., Reference Agrawal, Verma, Rao, Garg, Kapur and Bajpai2017).
The Harudi Formation is a shale-dominated unit bearing yellow limonitic partings in the lower part and calcareous claystone and siltstone in the upper part (Biswas, Reference Biswas1992). Snake fossils were recovered from horizons exposed at Godhatad (23°39.1’N, 68°45.2’E), Dhedidi North (23°46’42”N, 68°47’16”E), and Babia Hill (23°41’N, 68°45.4’E) localities that are also known to yield a host of archaic whales (protocetids, remingtonocetids, andrewsiphiines, and remingtonocetines) and sirenians (Bajpai et al., Reference Bajpai, Thewissen, Kapur, Tiwari and Sahni2006; Thewissen and Bajpai, Reference Thewissen and Bajpai2009). A late Lutetian age was assigned to the whale-bearing horizons of the Harudi Formation based on 87Sr/86Sr values (Ravikant and Bajpai, Reference Ravikant and Bajpai2010). The depositional setting of the snake-yielding localities at Godhatad and Babia Hill were interpreted to represent tidal and swampy/marshy environments, respectively (Mukhopadhyay and Shome, Reference Mukhopadhyay and Shome1996; Thewissen and Bajpai, Reference Thewissen and Bajpai2009).
Material and methods
Material
A total of 14 vertebrae from the five fossil localities (see Geological setting) were studied, the details of which are given in Appendix 1. All specimens are housed in the Vertebrate Paleontology Laboratory of the Department Earth Sciences, Indian Institute of Technology Roorkee. For comparative purposes the following taxa/families were studied: Palaeophiidae—Palaeophis vastaniensis Bajpai and Head, Reference Bajpai and Head2007; Palaeophis maghrebianus Arambourg, Reference Arambourg1952 (see Houssaye et al., Reference Houssaye, Rage, Bardet, Vincent, Amaghzaz and Meslouh2013); Palaeophis africanus Andrews, Reference Andrews1924 (see Parmley and DeVore, Reference Parmley and DeVore2005; Folie et al., Reference Folie, Mees, De Putter and Smith2021; Georgalis et al., Reference Georgalis, Guinot, Kassegne, Amoudji, Johnson, Cappetta and Hautier2021a); Palaeophis casei Holman, Reference Holman1982; Palaeophis oweni Zigno, Reference Zigno1881 (see Georgalis et al., Reference Georgalis, Del Favero and Delfino2020); Palaeophis typhaeus Owen, Reference Owen and Owen1850 (see Rage, Reference Rage1983a); Palaeophis colossaeus Rage, Reference Rage1983a (see McCartney et al., Reference McCartney, Roberts, Tapanila and Oleary2018; Georgalis et al., Reference Georgalis, Guinot, Kassegne, Amoudji, Johnson, Cappetta and Hautier2021a); Palaeophis nessovi Averianov, Reference Averianov1997 (see Snetkov, Reference Snetkov2011); Palaeophis toliapicus Owen, Reference Owen1841 (see Owen, Reference Owen and Owen1850; Rage, Reference Rage1983b; Zvonok and Snetkov, Reference Zvonok and Snetkov2012); Pterosphenus schucherti Lucas, Reference Lucas1898 (see Calvert et al., Reference Calvert, Mead and Parmley2022; Natarajan et al., Reference Natarajan, Dasgupta, Rakshit and Kashyap2024); Pterosphenus schweinfurthi Andrews, Reference Andrews1901 (see McCartney and Seiffert, Reference McCartney and Seiffert2016; Georgalis, Reference Georgalis2023); Pterosphenus sheppardi Hoffstetter, Reference Hoffstetter1958; Pterosphenus muruntau Averianov, Reference Averianov1997 (Averianov, Reference Averianov2023); Pterosphenus kutchensis Rage et al., Reference Rage, Bajpai, Thewissen and Tiwari2003; Pterosphenus biswasi Rage et al., Reference Rage, Bajpai, Thewissen and Tiwari2003; Archaeophis proavus Janensch, Reference Janensch1906; and ‘Archaeophis’ turkmenicus Tatarinov, Reference Tatarinov1963 (Tatarinov, Reference Tatarinov1963; Rage, Reference Rage1984; Rage et al., Reference Rage, Bajpai, Thewissen and Tiwari2003); Madtsoiidae—Platyspondylophis tadkeshwarensis Smith et al., Reference Smith, Kumar, Rana, Folie, Solé, Noiret, Steeman, Sahni and Rose2016 (Smith et al., Reference Smith, Kumar, Rana, Folie, Solé, Noiret, Steeman, Sahni and Rose2016); and Vasuki indicus Datta and Bajpai, Reference Datta and Bajpai2024; and Nigerophiidae—Nigerophis mirus Rage, Reference Rage1975; Kelyophis hechti LaDuke et al., Reference LaDuke, Krause, Scanlon and Kley2010; Indophis fanambinana Pritchard et al., Reference Pritchard, McCartney, Krause and Kley2014; and Indophis sahnii (see Rage, Reference Rage1975; Rage and Prasad, Reference Rage and Prasad1992; LaDuke et al., Reference LaDuke, Krause, Scanlon and Kley2010; Pritchard et al., Reference Pritchard, McCartney, Krause and Kley2014).
Methods
The osteological description of the vertebrae follows the terminology of Rage (Reference Rage1984), Rage et al. (Reference Rage, Bajpai, Thewissen and Tiwari2003), LaDuke et al. (Reference LaDuke, Krause, Scanlon and Kley2010), Rio and Mannion (Reference Rio and Mannion2017), and Szyndlar and Georgalis (Reference Szyndlar and Georgalis2023). Different parameters (following Datta and Bajpai, Reference Datta and Bajpai2024; Natarajan et al., Reference Natarajan, Dasgupta, Rakshit and Kashyap2024) of the specimens were measured (Fig. 2) using Mitutoyo digital callipers with a precision of 0.01 mm and explanatory line drawings have been used where necessary. The terminology of the vaulting ratio of the neural arch (VR) follows Georgalis et al. (Reference Georgalis, Rabi and Smith2021b). The holotype and referred specimens of Pterosphenus kutchensis, Pterosphenus biswasi, and Vasuki indicus were studied first-hand for comparative and phylogenetic analysis. Published literature and photographs were referred for information on other ophidian taxa.
Measured parameters on schematic diagrams: (1–5) basic snake vertebra, (6) generalized Pterosphenus vertebra: (1) anterior, (2, 6) lateral, (3) posterior, (4) dorsal, and (5) ventral views. Vertebral measurements follow Datta and Bajpai (Reference Datta and Bajpai2024) and Natarajan et al. (Reference Natarajan, Dasgupta, Rakshit and Kashyap2024). CL = centrum length; CNH = condyle height; CNW = condyle width; COH = cotyle height; COW = cotyle width; NAW = neural arch width; NCH = neural canal height; NCW = neural canal width; NSH = neural spine height; NSL = neural spine width; POFL = postzygapophyseal facet length; POFW = postzygapophyseal facet width; POW = postzygapophyseal width; POα = postzygapohyseal angle; PRFL = prezygapophyseal facet length; PRFW = prezygapophyseal facet width; PRW = prezygapophyseal width; PRα = prezygapophyseal angle; PTH = height of pterapophysis; PT.VH = vertebral height from base of condyle to dorsal tip of pterapophysis; SYα = synapophyseal angle; SYW = synapophyseal width; TVH = total vertebral height; ZS.VH = vertebral height from base of condyle to dorsal tip of zygosphene; ZSFL = zygosphenal facet length; ZSFW = zygosphenal facet width; ZSH = zygosphene height; ZSW = zygosphene width; ZSβ = zygosphene angle; ZYH, height of zygantrum; ZYW, width of zygantrum. Schematic diagrams not to scale and after Datta and Bajpai (Reference Datta and Bajpai2024) (1–5), Rage et al. (Reference Rage, Bajpai, Thewissen and Tiwari2003) and Garberoglio et al. (Reference Garberoglio, Gómez and Caldwell2024) (6).
Abbreviations for measured parameters
CL, centrum length; CNH, condyle height; CNW, condyle width; COH, cotyle height; COW, cotyle width; NAW, neural arch width; NCH, neural canal height; NCW, neural canal width; NSH, neural spine height; NSL, neural spine length; POFL, postzygapophyseal facet length; POFW, postzygapophyseal facet width; POW, postzygapophyseal width; POα, postzygapohyseal angle; PRFL, prezygapophyseal facet length; PRFW, prezygapophyseal facet width; PRW, prezygapophyseal width; PRα, prezygapophyseal angle; PTH, height of pterapophysis; PT.VH, vertebral height from base of condyle to dorsal tip of pterapophysis; SYα, synapophyseal angle; SYW, synapophyseal width; TVH, total vertebral height; VR, vaulting ratio; ZS.VH, vertebral height from base of condyle to dorsal tip of zygosphene; ZSFL, zygosphenal facet length; ZSFW, zygosphenal facet width; ZSH, zygosphene height; ZSW, zygosphene width; ZSβ, zygosphene angle; ZYH, height of zygantrum; ZYW, width of zygantrum.
Other abbreviations
ATV, anterior trunk vertebrae; MTV, middle trunk vertebrae; PTV, posterior trunk vertebrae; TV, trunk vertebrae.
Repository and institutional abbreviations
CGM, Egyptian Geological Museum, Cairo, Egypt; DPC, Duke Lemur Center, Durham, North Carolina, USA; GCVP, Georgia College Vertebrate Paleontology Collection, Georgia, USA; GU/RSR/VAS, Department of Geology, H.N.B. Garhwal University, Uttaranchal, India; IITBR1, Indian Institute of Technology Bombay, Mumbai, Maharashtra, India; IITR/VPL/SB, Vertebrate Paleontology Laboratory, Indian Institute of Technology Roorkee, Roorkee, India; MGP-PD, Museo di Geologia e Paleontologia dell’Università di Padova, Italy; MNHN, Muséum National d’Histoire Naturelle, Paris, France; MSUVP, Michigan State University Museum, East Lansing; NHMUK, The Natural History Museum, London, UK; NHMW, Naturhistorisches Museum Wien, Vienna, Austria; OCP, Office Chérifien des Phosphates, Khouribga, Morocco; PU, Museum of Natural History of Princeton University, Princeton, New Jersey; RMCA-RGP, Royal Museum for Central Africa—Registre Général Paléontologie, Tervuren, Belgium; RUSB, Roorkee University/Sunil Bajpai/Vertebrate Paleontology Laboratory; USNM, National Museum of Natural History (= United States National Museum), Washington, D.C.; VPL/JU, Vertebrate Paleontology Laboratory, University of Jammu; ZIN PC, Paleontological collection, Zoological Institute, Russian Academy of Sciences, Saint Petersburg, Russia; ZIN PH, Paleoherpetological collection, Zoological Institute, Russian Academy of Sciences, Saint Petersburg, Russia.
Systematic paleontology
Squamata Oppel, Reference Oppel1811
Order Serpentes Linnaeus, Reference Linnaeus1758
Family Palaeophiidae Lydekker, Reference Lydekker1888
Subfamily Palaeophiinae Lydekker, Reference Lydekker1888
Genus Pterosphenus Lucas, Reference Lucas1898
Type species
Pterosphenus schucherti Lucas, Reference Lucas1898 from the late Lutetian or early Bartonian, and Priabonian, USA.
Pterosphenus rannensis new species
Pterosphenus rannensis n. sp., holotype IITR/VPL/SB 3014-1, nearly complete anterior trunk vertebra shown by photographs and interpretative line drawings (in successive order): (1, 2) anterior; (3, 4) left lateral (white arrowhead and black arrow indicate fossae below interzygapophyseal ridge and at the base of neural spine, respectively); (5, 6) posterior; (7, 8) dorsal; (9, 10) ventral views. Red dashed line marks the dorsal margin of the zygosphene. Gray arrows indicate anterior direction. an.hy = anterior hypapophysis; co = cotyle; cn = condyle; izr = interzygapophyseal ridge; kl = keel; nc = neural canal; ns = neural spine; po = postzygapophysis; po.hy = posterior hypapophysis; pr = prezygapophysis; pt = pterapophysis; sy = synapophysis; zg = zygantrum; zgf = zygantral foramen; zgfa = zygantral facet; zgfo = zygantral fossa; zs = zygosphene. Scale bar = 10 mm.
Pterosphenus rannensis n. sp.: (1–5) holotype IITR/VPL/SB 3014-3, partial anterior trunk vertebra; (6–10) holotype IITR/VPL/SB 3014-2, partial ?middle trunk vertebra; and (11–13) holotype IITR/VPL/SB 3014-4, partial precloacal neural arch-spine complex: (1, 6, 11) anterior, (2, 7, 12) left lateral, (3, 8, 13) posterior, (4, 9) dorsal, and (5, 10) ventral views. Red dashed line marks the dorsal margin of the zygosphene. Gray arrows indicate anterior direction. an.hy = anterior hypapophysis; co = cotyle; cn = condyle; izr = interzygapophyseal ridge; nal = neural arch lamina; nc = neural canal; ns = neural spine; po = postzygapophysis; po.hy = posterior hypapophysis; pr = prezygapophysis; pt = pterapophysis; sy = synapophysis; zg = zygantrum; zgfa = zygantral facet; zgfo = zygantral fossa; zs = zygosphene. Scale bars = 10 mm.
Pterosphenus rannensis n. sp., holotype IITR/VPL/SB 3014-5, nearly complete middle trunk vertebra shown by photographs and interpretative line drawings (in successive order): (1, 2) anterior; (3, 4) right lateral (black arrow indicates fossa at the base of neural spine); (5, 6) posterior; (7, 8) dorsal; and (9, 10) ventral views. Red dashed line marks the dorsal margin of the zygosphene. Gray arrows indicate anterior direction. co = cotyle; cn = condyle; izr = interzygapophyseal ridge; kl = keel; nc = neural canal; ns = neural spine; pcofo = paracotylar fossa; po = postzygapophysis; po.hy = posterior hypapophysis; pr = prezygapophysis; pt = pterapophysis; sy = synapophysis; zg = zygantrum; zgf = zygantral foramen; zgfa = zygantral facet; zgfo = zygantral fossa; zs = zygosphene. Scale bar = 10 mm.
Pterosphenus rannensis n. sp.: (1–3) holotype IITR/VPL/SB 3014-6, partial middle trunk vertebra; (4–8) holotype IITR/SV 3014-7, partial middle trunk vertebra; (9–13) holotype IITR/SV 3014-8, partial vertebra; and (14–16) IITR/VPL/SB 2980, partial trunk vertebra: (1, 4, 9, 14) anterior; (2, 15) left lateral; (3, 6, 11) posterior; (5, 10) right lateral; (7, 12) dorsal; and (8, 13, 16). Gray arrows indicate anterior direction. co = cotyle; cn = condyle; izr = interzygapophyseal ridge; kl = keel; nc = neural canal; ns = neural spine; pcofo = paracotylar fossa; po = postzygapophysis; po.hy = posterior hypapophysis; pr = prezygapophysis; sy = synapophysis; zs = zygosphene. Scale bars = 10 mm.
Holotype
A partial vertebral column comprising 11 vertebrae (IITR/VPL/SB 3014-1–10, 3015) pertaining to the precloacal region (Appendix 1) from Godhatad, district Kutch, Gujarat state, western India.
Differential diagnosis
Pterosphenus rannensis n. sp. is assigned to Palaeophiinae based on the following combination of characters: presence of pterapophyses, posterior hypapophysis in anterior and middle trunk vertebrae, anterior hypapophysis in ATV, extension of synapophyses ventral to the centrum. The following combination of characters further designates this snake to Pterosphenus: triangular zygosphene in anterior view, neural spine originating from the anterior margin of the zygosphenal roof, weakly inclined prezygapophyseal facets. Additionally, Pterosphenus rannensis n. sp. is diagnosed based on a unique combination of following characters: differs from all other Pterosphenus species in ventral positioning of the pterapophysis relative to the dorsal roof of the zygosphene and shallow ventral extension of the synapophyses in MTV; differs from Pterosphenus kutchensis and Pterosphenus biswasi, and Pterosphenus schweinfurthi in having larger and smaller MTV, respectively; differs from Pterosphenus kutchensis, Pterosphenus schweinfurthi, and Pterosphenus schucherti in weak lateral compression of vertebrae; shares with Pterosphenus biswasi and Pterosphenus schucherti the presence of paracotylar fossae, but differs from Pterosphenus schweinfurthi in lacking paracotylar foramina; differs from Pterosphenus muruntau in having synapophyses ventral to the centrum in MTV; differs from Pterosphenus kutchensis in the absence of paired synapophyses originating from a common base; differs from Pterosphenus schweinfurthi, Pterosphenus kutchensis, Pterosphenus schucherti, and Pterosphenus muruntau in having a trapezoidal neural canal; shares with Pterosphenus kutchensis and Pterosphenus biswasi, but differs from Pterosphenus schweinfurthi, Pterosphenus schucherti, and Pterosphenus muruntau in having zygosphene as wide as the cotyle; differs from all other Pterosphenus species, except Pterosphenus kutchensis, in termination of ridge extending from pterapophysis dorsal to the interzygapophyseal ridge and a low PTH/CL ratio; differs from Pterosphenus biswasi, Pterosphenus kutchensis, Pterosphenus schweinfurthi, and Pterosphenus muruntau in lacking subcentral foramina.
Occurrence
Harudi Formation, middle Eocene (late Lutetian); Godhatad and Dhedidi North locality, district Kutch, Gujarat state, western India.
Description
The collection comprises 11 well-preserved, associated, partially to nearly complete vertebrae (Figs. 3–6). These are assigned to the precloacal region based on the presence of a hypapophysis and the absence of hemapophyses, pleurapophyses, and lymphapophyses (sensu Rage, Reference Rage1984). Furthermore, following Parmley and Reed (Reference Parmley and Reed2003), all specimens in the present study are considered as skeletally mature because of a smaller neural canal relative to the condyle.
The vertebrae, in general, are large (CL = 20–26 mm; Appendix 2) and mediolaterally compressed (PRW/CL = 1–1.1). They are dorsoventrally high (IITR/VPL/SB 3014-1, TVH/CL = 2.1) with the cotyle and condyle arranged along a horizontal axis. The cotyle is deeply concave in anterior view and mediolaterally wider than dorsoventrally high (IITR/VPL/SB 3014-6, COW/COH = 1.2; Figs. 3.1, 3.2, 5.1, 5.2; Appendix 2). It is cordiform in outline with a horizontal to weakly concave dorsal margin and a tapering ventral margin. This ventral tapering of the cotyle is most prominent in ATV (Figs. 3.1, 4.1). Furthermore, the ventral cotylar rim projects slightly anteriorly relative to the dorsal rim in ATV but is at the same level in MTV. The cotyle is flanked laterally by slender subtriangular paracotylar fossae, which lack paracotylar foramina (Fig. 5.1, 5.2). The condyle, like the cotyle, is mediolaterally wider than high (IITR/VPL/SB 3014-5, CNW/CNH = 1.2; Appendix 2) with a horizontal to weakly concave dorsal and a tapered ventral margin (Fig. 6.3, 6.6).
The synapophysis extends beneath the ventral cotylar rim in all specimens, but its form and position progressively change across the precloacals. In ATV, the synapophysis descends well below the ventral cotylar rim and bears a ventrolaterally directed facet (Fig. 3.1–3.4), whereas in MTV, it is closer to the cotyle with a strongly ventrally directed facet (Figs. 5.1–5.4, 6.1, 6.2). The synapophysis is reniform in lateral/ventrolateral view and lacks a distinct parapophysis and diapophysis (Figs. 3.3, 3.4, 3.9, 3.10, 6.5, 6.8). The ventral half of the synapophysis is deflected anteriorly but does not extend beyond the cotyle.
The prezygapophyseal buttress is vertical and ridge-like, extending from the anterolateral edge of the prezygapophysis to the dorsal margin of the synapophysis. Near its connection with the synapophysis, the buttress is laterally concave (Figs. 3.1, 3.2, 4.6). A prezygapophyseal accessory process is absent and instead, the buttress bears an anterolaterally directed convexity/bulge immediately ventral to the prezygapophyseal articular facet (Figs. 5.1, 5.2, 6.1, 6.2). The orientation of this bulge differs slightly from the anterior to middle trunk vertebrae, being more anteriorly directed in the former.
The prezygapophysis shows a weak dorsolateral lateral deflection in anterior view, with the angle of inclination with the horizontal decreasing from ATV (IITR/VPL/SB 3014-1, PRα = 25°; Fig. 3.1, 3.2) to MTV (IITR/VPL/SB 3014-5, PRα = 9°; Fig. 6.1, 6.4). It is reduced and extends laterally beyond the synapophysis. Relative to the neural canal, the prezygapophysis remains at the same level as the ventral margin of the canal or rests marginally above the latter (Fig. 5.1, 5.2). The position of the prezygapophysis relative to the neural canal, however, appears higher due to its dorsolateral deflection. In dorsal view, the prezygapophyseal articular facet is triangular (IITR/VPL/SB 3014-1, PRFL/PRFW = 1.5) and anterolaterally deflected from the sagittal plane (Figs. 4.9, 6.7, 6.12). Posterior to the prezygapophysis, the lateral surface of the centrum bears a fossa that is roofed by the interzygapophyseal ridge (Fig. 3.3, 3.4). The postzygapophysis also bears a dorsolateral deflection, but in contrast to the prezygapophysis, the angle of its inclination with the horizontal increases from anterior (IITR/VPL/SB 3014-1, POα = 6°; Fig. 3.5, 3.6) to middle trunk vertebrae (IITR/VPL/SB 3014-7, POα = 16°; Fig. 5.5, 5.6). The postzygapophyseal articular facet is longer than wide (IITR/VPL/SB 3014-1, POFL/POFW = 1.4). In lateral view, the postzygapophysis extends below the level of the prezygapophysis at least in ATV (Fig. 3.3, 3.4). The pre- and postzygapophysis are connected by a sharp and horizontal interzygapophyseal ridge which remains distinct from the pterapophysis throughout its length (Figs. 3.3, 3.4, 5.3, 5.4).
The neural canal is trapezoidal in both anterior and posterior views, and transversely narrower than both the cotyle and zygosphene (Fig. 6.3, 6.4, 6.9, 6.11). The latter is as wide as the cotyle, triangular in anterior view and dorsoventrally high (IITR/VPL/SB 3014-5, MTV, ZSW/COW = 1; ZSH/ZSW = 0.7; Figs. 3.1, 3.2, 4.1, 7, 5.1, 5.2). The anterior surface of the zygosphene accommodates a prominent fossa resulting in a deeply V-shaped embayment, as seen in dorsal view (Figs. 5.7, 5.8, 6.7, 6.12). The concavity of this embayment, however, weakens from ATV (IITR/VPL/SB 3014-1, ZSβ = 115°) to MTV (IITR/VPL/SB 3014-5, ZSβ = 125°). The dorsolateral margin of the zygosphene is concave and anteroventrally directed (Fig. 6.4). The zygosphenal facet is dorsoventrally higher than mediolaterally wide (IITR/VPL/SB 3014-7, ZSFL/ZSFW = 1.2) and anteroventrally deflected so that much of the facet is visible in anterior view. The facet, however, remains posterior to the prezygapophysis. The lateral surface of the neural arch accommodates a shallow fossa between the zygosphene and the pterapophysis at the base of the neural spine (Figs. 3.3, 3.4, 4.12, 5.3, 5.4). This fossa extends slightly below the level of the zygosphene and is flanked ventrally by an anteroventrally directed ridge emanating for the pterapophysis.
The neural arch is strongly vaulted in posterior view with the vaulting ratio (sensu Georgalis et al., Reference Georgalis, Rabi and Smith2021b) across the precloacals ranging from 0.8 (ATV) to 0.6 (MTV). The zygantrum is partially roofed and transversely wider than dorsoventrally high (IITR/VPL/SB 3014-4, ZYH/ZYW = 0.2; Figs. 4.3, 4.13, 5.5, 5.6). It accommodates deep bilateral fossae that face posteromedially and are separated by a low median wall (Figs. 3.5, 3.6, 5.5, 5.6). Each fossa comprises an oval, dorsomedially directed zygantral facet. Immediately dorsal to these fossae, a small zygantral foramen in present along the sagittal plane (Figs. 3.5, 3.6, 5.5, 5.6).
The pterapophysis is low and located at ~77% of the total vertebral height (at least in ATV). It is posterodorsally directed in lateral view and remains below the dorsal tip of the zygosphene (Figs. 3.3, 3.4, 4.12, 5.3, 5.4). A mediolaterally compressed and dorsolaterally inclined ridge connects the pterapophysis to the postzygapophysis. Sharp neural arch laminae extend dorsomedially from the pterapophyses to the neural spine. These laminae are concave dorsolaterally with the degree of concavity ranging from 95° (IITR/VPL/SB 3014-3) to 100° (IITR/VPL/SB 3014-4).
The neural spine is mediolaterally compressed and extends to the anterior border of the zygosphene, forming a part of the zygosphenal roof (Figs. 3.1–3.4, 5.1–5.4). It is dorsoventrally high and forms 39% of the total vertebral height in ATV. The spine is subtriangular in lateral view with both the anterior and posterior margins posterodorsally directed.
The ventral surface of the vertebra comprises a large posterior hypapophysis, which is blunt and ventrally directed (Figs. 3.9, 3.10, 6.8). In lateral view, the base of the posterior hypapophysis extends from midlength of the centrum nearly to the ventral condylar rim. A sharp carina/keel extends anteriorly from the posterior hypapophysis and gives way to an anteroventrally directed anterior hypapophysis in ATV (Fig. 3.9, 3.10). The anterior hypapophysis is morphologically similar to the posterior hypapophysis, albeit smaller, and is replaced in MTV by a low knob (Figs. 5.1, 5.2, 5.9, 5.10, 6.4, 6.8). Across the precloacals, this ridge connecting the two hypapophyses varies from being dorsal to the synapophysis in ATV (Fig. 3.3, 3.4, 3.9, 3.10) to somewhat ventral in MTV (Fig. 5.3, 5.4, 5.9, 5.10).
In addition to the specimens described above from Godhatad, we report another vertebra (IITR/VPL/SB 2980) of Pterosphenus rannensis n. sp. from the Dhedidi locality of the Harudi Formation. IITR/VPL/SB 2980 is missing both the pre- and postzygapophyses, the pterapophyses, and much of the zygantrum and neural spine. Its generic identification is based on the presence of a triangular zygosphene with the preserved anterior margin of the neural spine extending from the dorsal zygosphenal margin (Fig. 6.14, 6.15). The specific attribution is based on the overall morphological similarity of IITR/VPL/SB 2980 with Pterosphenus rannensis n. sp. from Godhatad, with the shared presence of a cordiform cotyle that is transversely wider than dorsoventrally high (COW/COH = 1.2), and a zygosphene that is nearly as wide as the cotyle (ZSW/COW = 0.9). Furthermore, IITR/VPL/SB 2980 is most likely from the middle trunk region because the ventral surface of this specimen, although poorly preserved, suggests the presence of a posterior hypapophysis succeeded anteriorly by a low knob rather than an anterior hypapophysis (sensu McCartney and Seiffert, Reference McCartney and Seiffert2016; Fig. 6.16).
Etymology
The specific name is after the Rann of Kutch, referring to the area of geographic origin.
Referred specimen
IITR/VPL/SB 2980, an isolated precloacal vertebra.
Remarks
The diagnostic characters of the new Indian taxon that support its placement within the subfamily Palaeophiinae and the genus Pterosphenus are listed above in the Differential diagnosis section (see Rage, Reference Rage1984; Rage et al., Reference Rage, Bajpai, Thewissen and Tiwari2003; Houssaye et al., Reference Houssaye, Rage, Bardet, Vincent, Amaghzaz and Meslouh2013; McCartney and Seiffert, Reference McCartney and Seiffert2016; Georgalis, Reference Georgalis2023).
In comparison with the previously recognized species of Pterosphenus, Pterosphenus rannensis n. sp. (IITR/VPL/SB 3014-5, MTV, CL = 21 mm; PRW = 23.1 mm) is larger than the Indian forms Pterosphenus kutchensis (RUSB 2721-1, holotype, MTV, CL = 8.3 mm; PRW = 6.6 mm; Rage et al., Reference Rage, Bajpai, Thewissen and Tiwari2003) and Pterosphenus biswasi (RUSB 2784-4, holotype, MTV, CL = 18.9 mm; PRW = 19.7 mm; Rage et al., Reference Rage, Bajpai, Thewissen and Tiwari2003), but smaller than the African Pterosphenus schweinfurthi (NHMUK PV R 3009b, CL = 33.3 mm, PRW = 29.7 mm). This new taxon is, however, comparable in size to Pterosphenus schucherti known from North America (GCVP 19863, MTV, CL = 24.8 mm; PRW = 22.2 mm; Calvert et al., Reference Calvert, Mead and Parmley2022) and now from India (IITBR1/MT1, MTV, CL = 26.72 mm; Natarajan et al., Reference Natarajan, Dasgupta, Rakshit and Kashyap2024), as well as to the central Asian Pterosphenus muruntau (ZIN PC 2/34, paratype, MTV, CL = 23.7; Averianov, Reference Averianov2023).
Contrary to the marked lateral flattening characteristic of Pterosphenus vertebrae (Rage et al., Reference Rage, Bajpai, Thewissen and Tiwari2003; Head et al., Reference Head, Holroyd, Hutchison and Ciochon2005; Georgalis, Reference Georgalis2023) (e.g., Pterosphenus kutchensis [RUSB 2721-1, holotype, MTV, PRW/CL = 0.7; Rage et al., Reference Rage, Bajpai, Thewissen and Tiwari2003], Pterosphenus schweinfurthi [CGM 10194, holotype, TV, PRW/CL = 0.8], Pterosphenus schucherti [GCVP 19863, MTV, PRW/CL = 0.8; Calvert et al., Reference Calvert, Mead and Parmley2022]), Pterosphenus rannensis n. sp. shows weaker lateral compression (IITR/VPL/SB 3014-6, MTV, PRW/CL = 1.1). The latter feature is reminiscent of ‘primitive’-grade Palaeophis Owen, Reference Owen1841 (sensu Rage et al., Reference Rage, Bajpai, Thewissen and Tiwari2003), e.g., Palaeophis colossaeus (MNHN.F.TGE 615, holotype, MTV, PRW/CL = 1.2) and Palaeophis maghrebianus (MNHN APH 5, holotype, MTV, PRW/CL = 1.1). A weak lateral compression, however, is also seen in the ‘advanced’-grade palaeophiid Palaeophis typhaeus ([catalogue number not cited], MTV, PRW/CL = 1.2; Owen, Reference Owen and Owen1850), Pterosphenus muruntau (ZIN PH 4/287, ATV, PRW/CL = 1.1; Averianov, Reference Averianov2023) and Pterosphenus biswasi (RUSB 2784-4, holotype, MTV, PRW/CL = 1; Rage et al. Reference Rage, Bajpai, Thewissen and Tiwari2003). Pterosphenus rannensis n. sp. shares with Pterosphenus biswasi and Pterosphenus schucherti (Calvert et al., Reference Calvert, Mead and Parmley2022) the presence of paracotylar fossae. Although this feature is also present in Pterosphenus schweinfurthi (see McCartney and Seiffert, Reference McCartney and Seiffert2016), it differs from Pterosphenus rannensis n. sp. in bearing paracotylar foramina.
We recognize the weak ventral extension of the synapophyses in MTV of Pterosphenus rannensis n. sp. as a distinguishing feature from other Pterosphenus species. In taxa such as Pterosphenus kutchensis, Pterosphenus biswasi, Pterosphenus schucherti, and Pterosphenus schweinfurthi, the synapophyses in MTV show marked ventral extension relative to the centrum (Rage et al., Reference Rage, Bajpai, Thewissen and Tiwari2003; Mccartney and Seiffert, Reference McCartney and Seiffert2016; Georgalis, Reference Georgalis2023; Natarajan et al., Reference Natarajan, Dasgupta, Rakshit and Kashyap2024). In Pterosphenus muruntau, the synapophyses remain dorsal to the ventral margin of the centrum (Averianov, Reference Averianov2023). Furthermore, the presence of paired synapophyses originating from a common base in Pterosphenus kutchensis (see Rage et al., Reference Rage, Bajpai, Thewissen and Tiwari2003) differentiates the latter from Pterosphenus rannensis n. sp. in which this feature is absent. Rage et al. (Reference Rage, Bajpai, Thewissen and Tiwari2003) mentioned the slight ventral displacement of the synapophysis as a characteristic of ‘primitive’ Palaeophis snakes.
The dorsolateral deflection of the pre- and postzygapophysis in Pterosphenus rannensis n. sp. contrasts with horizontal zygapophyses in other Pterosphenus species, including Pterosphenus kutchensis, Pterosphenus biswasi, Pterosphenus schweinfurthi, and Pterosphenus schucherti (Rage et al., Reference Rage, Bajpai, Thewissen and Tiwari2003; McCartney and Seiffert, Reference McCartney and Seiffert2016; Calvert et al., Reference Calvert, Mead and Parmley2022; Georgalis, Reference Georgalis2023). In some specimens of Pterosphenus schucherti, however, the prezygapophyses are dorsolaterally deflected (Lucas, Reference Lucas1898; Holman, Reference Holman1977; Natarajan et al., Reference Natarajan, Dasgupta, Rakshit and Kashyap2024). Pterosphenus rannensis n. sp. shares with Pterosphenus biswasi (see Rage et al., Reference Rage, Bajpai, Thewissen and Tiwari2003) in anterior view a trapezoidal neural canal. This contrasts with the other Pterosphenus spp. (Pterosphenus schweinfurthi, Pterosphenus kutchensis, Pterosphenus schucherti, and Pterosphenus muruntau) in which the canal is semicircular (Rage et al., Reference Rage, Bajpai, Thewissen and Tiwari2003; McCartney and Seiffert, Reference McCartney and Seiffert2016; Calvert et al., Reference Calvert, Mead and Parmley2022; Averianov, Reference Averianov2023; Georgalis, Reference Georgalis2023).
Pterosphenus rannensis n. sp. (IITR/VPL/SB 3014-5, MTV, NAW/CL = 1) shows weak interzygapophyseal constriction in the middle trunk region, similar to Pterosphenus schweinfurthi (DPC 25680, MTV, NAW/CL = 0.9) and Pterosphenus biswasi (RUSB 2784-4, holotype, MTV, NAW/CL = 0.9). On the contrary, a stronger interzygapophyseal constriction is seen in Pterosphenus kutchensis (RUSB 2721-1, holotype, MTV, NAW/CL = 0.6) and Pterosphenus schucherti (GCVP 19863, MTV, NAW/CL = 0.7). A ridge extending anteroventrally from the pterapophysis to the interzygapophyseal ridge distinguishes all species of Pterosphenus (McCartney and Seiffert, Reference McCartney and Seiffert2016; Zouhri et al., Reference Zouhri, Khalloufi, Bourdon, de Broin, Rage, M’haïdrat, Gingerich and Elboudali2018; Calvert et al., Reference Calvert, Mead and Parmley2022; Averianov, Reference Averianov2023; Georgalis, Reference Georgalis2023; Natarajan et al., Reference Natarajan, Dasgupta, Rakshit and Kashyap2024) except for Pterosphenus kutchensis from Pterosphenus rannensis n. sp. In the latter Indian forms, this ridge terminates well above the interzygapophyseal ridge.
In Pterosphenus rannensis n. sp., the zygosphene is as wide as the cotyle in MTV (IITR/VPL/SB 3014-5, MTV, ZSW/COW = 1), resembling Pterosphenus kutchensis (RUSB 2721-1, holotype, MTV, ZSW/COW = ~1) and Pterosphenus biswasi (RUSB 2784-4, holotype, MTV, ZSW/COW = ~1). In Pterosphenus schweinfurthi (DPC 25680, MTV, ZSW/COW = 1.1) and Pterosphenus schucherti (GCVP 19863, MTV, ZSW/COW = 1.1; USNM V 4047, holotype, TV, ZSW/COW = 1.1; IITBR1/MT1, ZSW/ COW = 1.3), the zygosphene is transversely wider, whereas, in Pterosphenus muruntau it is narrower (ZIN PC 2/34, holotype, MTV, ZSW/COW = 0.6). Pterosphenus kutchensis, however, differs from Pterosphenus rannensis n. sp. in the sigmoidal anterior zygosphenal margin and lacking a fossa on the anterior surface of the zygosphene (Rage et al., Reference Rage, Bajpai, Thewissen and Tiwari2003; McCartney and Seiffert, Reference McCartney and Seiffert2016).
A high vaulting ratio of the neural arch in MTV distinguishes Pterosphenus kutchensis (RUSB 2721-1, VR = 1.2) and Pterosphenus schweinfurthi (DPC 25680, VR = 0.9) from Pterosphenus rannensis n. sp. (IITR/VPL/SB 3014-5, VR = 0.6), Pterosphenus biswasi (RUSB 2784-4, VR = 0.7) and Pterosphenus schucherti (GCVP 19863, VR = 0.6; IITBR1/MT1, MTV, VR = 0.6) in which this ratio is noticeably lower. A major distinguishing feature between Pterosphenus rannensis n. sp. and other Pterosphenus species is the height of the pterapophysis relative to both the zygosphene and CL. In Pterosphenus rannensis n. sp., the pterapophysis remains below the dorsal tip of the zygosphene on all precloacal vertebrae, and the PTH/CL ratio (0.59–0.62) is low. In all other named species of Pterosphenus, the pterapophysis extends above the zygosphene (Holman, Reference Holman1977; Rage et al., Reference Rage, Bajpai, Thewissen and Tiwari2003; McCartney and Seiffert, Reference McCartney and Seiffert2016; Averianov, Reference Averianov2023; Georgalis, Reference Georgalis2023; Natarajan et al., Reference Natarajan, Dasgupta, Rakshit and Kashyap2024), and the PTH/CL ratio is significantly higher: Pterosphenus schucherti (PTH/CL = 1.3–0.71; see Natarajan et al., Reference Natarajan, Dasgupta, Rakshit and Kashyap2024, table S3); Pterosphenus schweinfurthi (PTH/CL = 1.31–0.82; see Natarajan et al., Reference Natarajan, Dasgupta, Rakshit and Kashyap2024, table S3), Pterosphenus muruntau (PTH/CL = 0.68; see Natarajan et al., Reference Natarajan, Dasgupta, Rakshit and Kashyap2024, table S3); Pterosphenus biswasi (RUSB 2784-4, holotype, MTV, PTH/CL = 0.69; this study). Only in Pterosphenus kutchinesis is the PTH/CL ratio (RSUB 2721-1, holotype, MTV, PTH/CL = 0.59; this study) comparable to that in Pterosphenus rannensis n. sp. It is noteworthy that the PTH/CL ratio was considered an important distinguishing feature between Pterosphenus schucherti and Pterosphenus muruntau in a recent study on middle Eocene palaeophiids from India by Natarajan et al. (Reference Natarajan, Dasgupta, Rakshit and Kashyap2024).
The pterapophysis-postzygapohysis connection in Pterosphenus schucherti, Pterosphenus schweinfurthi, Pterosphenus kutchensis, and Pterosphenus muruntau shows a well-developed, posteriorly directed concavity in lateral view (Lucas, Reference Lucas1898; McCartney and Seiffert, Reference McCartney and Seiffert2016; Averianov, Reference Averianov2023; Georgalis, Reference Georgalis2023; Natarajan et al., Reference Natarajan, Dasgupta, Rakshit and Kashyap2024). On the contrary, in Pterosphenus rannensis n. sp. and Pterosphenus biswasi, this connection is vertical/incipiently concave. Furthermore, Pterosphenus rannensis n. sp. differs from Pterosphenus biswasi, Pterosphenus kutchensis, Pterosphenus schweinfurthi, and Pterosphenus muruntau in lacking subcentral foramina.
It can be noted here that the position of the pterapophysis relative to the zygosphene in Pterosphenus rannensis n. sp. is also seen in some ‘primitive’- and ‘advanced’-grade Palaeophis spp. including Palaeophis colossaeus (MNHN.F.TGE 615; Georgalis et al., Reference Georgalis, Guinot, Kassegne, Amoudji, Johnson, Cappetta and Hautier2021a), Palaeophis oweni (MGP-PD 6981Za, 6981Zc; Georgalis et al., Reference Georgalis, Del Favero and Delfino2020), Palaeophis nessovi (ZIN PH 16/153; Zvonok and Snetkov, Reference Zvonok and Snetkov2012), Palaeophis casei (PU 23488; Holman, Reference Holman1982), Palaeophis tamdy Averianov, Reference Averianov1997 (ZIN PH 18/153; Zvonok and Snetkov, Reference Zvonok and Snetkov2012), and Palaeophis littoralis Cope, Reference Cope1868 (MSUVP 1212; Parmley and Case, Reference Parmley and Case1988).
The three ecological grades traditionally recognized within Palaeophiinae (‘primitive’ and ‘advanced’ Palaeophis grade and Pterosphenus grade) illustrate a morphological series ranging from species that are weakly adapted to an aquatic habitat to those that are strongly adapted (Rage and Wouters, Reference Rage and Wouters1979; Rage et al., Reference Rage, Bajpai, Thewissen and Tiwari2003). Houssaye et al. (Reference Houssaye, Rage, Bardet, Vincent, Amaghzaz and Meslouh2013), however, considered the segregation of Palaeophis species into ‘primitive’ and ‘advanced’ grades of no systematic value. Other workers highlighted the lack of a distinct boundary between the two phenotypic genera (Palaeophis and Pterosphenus) and considered this distinction as possibly artificial (Rage et al., Reference Rage, Bajpai, Thewissen and Tiwari2003; McCartney et al., Reference McCartney, Roberts, Tapanila and Oleary2018; Georgalis et al., Reference Georgalis, Guinot, Kassegne, Amoudji, Johnson, Cappetta and Hautier2021a; Smith and Georgalis, Reference Smith, Georgalis, Gower and Zaher2022; Garberoglio et al., Reference Garberoglio, Gómez and Caldwell2024). Fossil evidence showing this indistinctness was previously presented based on material from India and Morocco, in which palaeophiine forms assigned to Pterosphenus also preserve features characteristic of Palaeophis (Rage et al., Reference Rage, Bajpai, Thewissen and Tiwari2003; Zouhri et al., Reference Zouhri, Khalloufi, Bourdon, de Broin, Rage, M’haïdrat, Gingerich and Elboudali2018). The palaeophiine Palaeophis nessovi from Kazakhstan presents a similar pattern because it shows morphological similarity with Pterosphenus (Averianov, Reference Averianov1997; Rage et al., Reference Rage, Bajpai, Thewissen and Tiwari2003; Zouhri et al., Reference Zouhri, Khalloufi, Bourdon, de Broin, Rage, M’haïdrat, Gingerich and Elboudali2018). In this context, Pterosphenus rannensis n. sp. further adds to the list of palaeophiine taxa that do not completely fit into one of the three ecological grades. The new Indian taxon, although showing features that are diagnostic of Pterosphenus, also preserves morphological traits reminiscent of ‘primitive’ Palaeophis species. Nonetheless, based on the zygosphenal and neural spine morphology, we favor placing the new species within Pterosphenus. Increased sampling of Eocene deposits in Kutch and other coeval horizons could provide greater anatomical coverage of Pterosphenus rannensis n. sp., leading to improved understanding of palaeophiine intracolumnar variation. This, in turn, could provide a better resolution for distinguishing between Pterosphenus and Palaeophis.
We recognize that the intermediate morphology of the palaeophiines mentioned above and the lack of a clear distinction between Pterosphenus and Palaeophis present a morphological conundrum. It appears likely that the palaeophiine taxa with such intermediate morphology represent transitional forms between Palaeophis and Pterosphenus, as suggested by Zouhri et al. (Reference Zouhri, Gingerich, Khalloufi, Bourdon and Adnet2021). Alternatively, the occurrence of Palaeophis features among Pterosphenus species and vice versa points to the possibility that some of these morphological features represent ancestral traits of the clade Palaeophiinae. However, in light of the poor understanding of palaeophiine intracolumnar variation, it currently seems premature to prefer one explanation over the other.
Genus Palaeophis Owen, Reference Owen1841
Type species
Palaeophis toliapicus Owen, Reference Owen1841, from the Eocene (Ypresian) of England.
?Palaeophis sp. indet.
?Palaeophis sp. indet., IITR/VPL/SB 2632, nearly complete middle trunk vertebra shown by photographs and interpretative line drawings (in successive order): (1, 2) anterior; (3, 4) right lateral (white arrow indicates fossa at the base of the neural spine); (5, 6) posterior; (7, 8) dorsal; and (9, 10) ventral views. Gray arrows indicate anterior direction. co = cotyle; cn = condyle; izr = interzygapophyseal ridge; kl = keel; nc = neural canal; ns = neural spine; pcofo = paracotylar fossa; po = postzygapophysis; po.hy = posterior hypapophysis; pr = prezygapophysis; pt = pterapophysis; sy = synapophysis; zg = zygantrum; zgfa = zygantral facet; zgfo = zygantral fossa; zs = zygosphene. Scale bar = 10 mm.
Occurrence
Harudi Formation, middle Eocene (late Lutetian); Babia Hill, Kutch district, western India.
Description
IITR/VPL/SB 2632 is an incomplete, isolated vertebra missing the neural spine, left prezygapophysis, left postzygapophysis, and the dorsal tip of the left pterapophysis (Fig. 7). This procoelous vertebra is large (CL = 28 mm; Appendix 2), mediolaterally compressed (PRW/CL = 0.8), and from the middle trunk region. The latter assessment is based on the presence of a large posterior and a smaller peg-like anterior hypapophysis (Fig. 7.1, 7.2, 7.9, 7.10), along with the absence of pleurapophyses, hemapophyses, and lymphapophyses (sensu Rage, Reference Rage1984; McCartney and Seiffert, Reference McCartney and Seiffert2016).
Both the cotyle and condyle are transversely wider than dorsoventrally high (COW/COH = 1.1, CNW/CNH = 1.1; Appendix 2) and arranged along a horizontal axis. The cotyle is cordiform in outline whereas the condyle is somewhat triangular (Fig. 7.1, 7.2, 7.5, 7.6). Well-developed paracotylar fossae flank the cotyle laterally but lack paracotylar foramina. The synapophysis is dorsoventrally high and extends well below the ventral margin of the cotyle. It is anteriorly deflected and lacks distinct parapophyseal and diapophyseal facets (Fig. 7.1–7.4).
The prezygapophysis is reduced and strongly deflected dorsolaterally in anterior view (PRα = 39°; Fig. 7.1, 7.2). It extends laterally beyond the synapophysis and is connected to the latter by a mediolaterally compressed prezygapophyseal buttress. The buttress is posteroventrally directed in lateral view and bears an anterolaterally directed, pronounced convexity immediately ventral to the prezygapophyseal articular facet (Fig. 7.1–7.4). The latter is longer than wide in dorsal view (PRFL/PRFW = 1.7) and anterolaterally deflected from the sagittal plane. The prezygapophyseal articular facet rests above the floor of the neural canal for most of its length (in anterior view). However, the medialmost part of the articular facet remains below the canal floor owing to the dorsolateral deflection of the facet. The postzygapophysis is medioventrally deflected (IITR/VPL/SB 2632, POα = 12°), in posterior view (Fig. 7.5, 7.6) and bears a triangular articular facet (IITR/VPL/SB 2632, POFL/POFW = 1.4). The interzygapophyseal ridge is weakly developed and slightly deflected posterodorsally in lateral view (Fig. 7.3, 7.4). The interzygapophyseal constriction is moderate (NAW/CL = 0.8) with the interzygapophyseal ridge being laterally concave in dorsal view.
The anterior surface of the neural arch is badly eroded, obscuring the morphology of the zygosphene and the anterior exit of the neural canal. The eroded zygosphene appears triangular although this could be a preservation artifact. The posterior surface of the neural arch is better preserved and accommodates a subtriangular neural canal, which is succeeded dorsally by a partially roofed zygantrum (Fig. 7.5, 7.6). The zygantral fossae, separated by a low median wall, are deep and posteromedially directed. The zygantral facets, although poorly preserved, appear oval in outline and face dorsomedially. The vaulting ratio (sensu Georgalis et al., Reference Georgalis, Rabi and Smith2021b) of the neural arch was found to be high (0.8).
The pterapophysis is well developed, posterodorsally directed, and appears to have extended above the zygosphene, judging from preserved dorsal extent of the latter (Fig. 7.3–7.6). A sharp ridge extends on the anterior surface of the pterapophysis. This ridge turns anteroventrally toward the interzygapophyseal ridge but does not reach it and flanks a shallow fossa medially (Fig. 7.3, 7.4). The pterapophysis is connected to the neural spine by sharp and dorsally concave (θ = 85°) neural arch laminae. Only the base of the neural spine is preserved, which is triangular in cross section and appears to terminate well posterior to the zygophene (Fig. 7.7, 7.8).
The ventral surface of the centrum bears a sharp midline keel that extends between the posterior and anterior hypapophyses (Fig. 7.9, 7.10). The keel is ventrally concave in lateral view and remains dorsal to the ventral margin of the synapophysis (Fig. 7.3, 7.4). The anterior hypapophysis is small and peg-like and bordered laterally by the synapophyses. The large posterior hypapophysis is triangular in lateral view, with a convex posterior and a concave anterior margin. This hypapophysis extends from a short distance anterior to the ventral condylar rim and terminates slightly anterior to the midlength of the centrum.
Materials
IITR/VPL/SB 2632, an isolated middle trunk vertebra.
Remarks
IITR/VPL/SB 2632 is identified as a palaeophiine based on the presence of well-developed pterapophysis, paired hypapophyses, and strong ventral deflection of the synapophyses relative to the centrum (Rage, Reference Rage1984; Houssaye et al., Reference Houssaye, Rage, Bardet, Vincent, Amaghzaz and Meslouh2013). When compared with the three palaeophiinae ecological grades proposed by Rage (Reference Rage1984) and Rage et al. (Reference Rage, Bajpai, Thewissen and Tiwari2003), IITR/VPL/SB 2632 shows similarity with ‘advanced’-grade Palaeophis and Pterosphenus. The strong lateral compression of the vertebrae and ventral extent of the synapophysis seen in IITR/VPL/SB 2632 (MTV, PRW/CL = 0.8) resembles those in ‘advanced’ Palaeophis (e.g., Palaeophis nessovi, ZIN PH 119, MTV, PRW/CL = 0.8), Pterosphenus (e.g., Pterosphenus kutchensis [RUSB 2721-1, holotype, MTV, PRW/CL = 0.7; Rage et al., Reference Rage, Bajpai, Thewissen and Tiwari2003], and Pterosphenus schweinfurthi [DPC 25680, MTV, PRW/CL = 0.9; McCartney and Seiffert, Reference McCartney and Seiffert2016; Georgalis, Reference Georgalis2023]). This Indian palaeophiine (IITR/VPL/SB 2632, MTV, PRFL/CL = 0.3) also shares a marked reduction of the prezygapophyses with these two ecological grades (e.g., Pterosphenus schweinfurthi [e.g., DPC 25680, MTV, PRFL/CL = ~0.3], Pterosphenus schucherti [GCVP 19863, MTV, PRFL/CL = 0.3], Palaeophis nessovi [ZIN PH 119, MTV, PRFL/CL = ~0.3], Palaeophis cf. Palaeophis toliapicus from Crimea [ZIN PH 14/153, PRFL/CL = 0.2]). The dorsal extent of the pterapophysis in IITR/VPL/SB 2632 is similar to that of Pterosphenus (e.g., Pterosphenus schweinfurthi, Pterosphenus schucherti, Pterosphenus muruntau, Pterosphenus kutchensis), whereas its neural spine morphology is characteristic of Palaeophis because it remains distinct from the zygosphene (Rage, Reference Rage1984; Rage et al., Reference Rage, Bajpai, Thewissen and Tiwari2003; Houssaye et al., Reference Houssaye, Rage, Bardet, Vincent, Amaghzaz and Meslouh2013; McCartney and Seiffert, Reference McCartney and Seiffert2016; Georgalis et al., Reference Georgalis, Guinot, Kassegne, Amoudji, Johnson, Cappetta and Hautier2021a; Averianov, Reference Averianov2023; Georgalis, Reference Georgalis2023).
The discussion above suggests that IITR/VPL/SB 2632 had clear adaptations for an aquatic life considering its similarities with ‘advanced’-grade Palaeophis and Pterosphenus snakes (sensu Rage et al., Reference Rage, Bajpai, Thewissen and Tiwari2003). Although most features characterizing IITR/VPL/SB 2632 are common to the two aforementioned ecological grades, the termination of the neural spine posterior to the zygosphenal roof suggests referral of this Indian palaeophiine to Palaeophis (sensu Rage, Reference Rage1984; Rage et al., Reference Rage, Bajpai, Thewissen and Tiwari2003; Houssaye et al., Reference Houssaye, Rage, Bardet, Vincent, Amaghzaz and Meslouh2013). In light of the poorly preserved zygosphene, the neural spine morphology of IITR/VPL/SB 2632 is possibly the only feature indicative of its generic affinity. Furthermore, the vertebral morphology of IITR/VPL/SB 2632 also highlights its distinctiveness from Pterosphenus rannensis n. sp. Although IITR/VPL/SB 2632 is of the same geological age as Pterosphenus rannensis n. sp., it differs from the latter taxon in having stronger lateral vertebral compression, greater ventral extent of synapophysis, pterapophysis extending above zygosphene, and the neural spine distinct from the zygosphene. Although the morphological comparison of IITR/VPL/SB 2632 with the palaeophiine ecological grades and Pterosphenus rannensis n. sp. suggests the presence of a second middle Eocene palaeophiine taxon in Kutch, we refrain from naming IITR/VPL/SB 2632 because of limited material currently available and instead refer to it as ?Palaeophis sp. indet.
Genus Pterosphenus Lucas, Reference Lucas1898
Type species
Pterosphenus schucherti Lucas, Reference Lucas1898, from the late Lutetian or early Bartonian, and Priabonian, USA.
Pterosphenus biswasi Rage et al., Reference Rage, Bajpai, Thewissen and Tiwari2003
Pterosphenus biswasi Rage et al., Reference Rage, Bajpai, Thewissen and Tiwari2003: (1–3) holotype RUSB 2784-4, partial middle trunk vertebra; (4, 5) RUSB 2565-1, partial trunk vertebra: (1, 4) anterior; (2, 5) right lateral; and (3) posterior. Gray arrow indicates anterior direction. co = cotyle; cn = condyle; nc = neural canal; po = postzygapophysis; pr = prezygapophysis; pt = pterapophysis; sy = synapophysis; zg = zygantrum; zs = zygosphene. Scale bar = 10 mm.
Holotype
RUSB 2784-4, partial trunk vertebra.
Revised Diagnosis
Pterosphenus biswasi is diagnosed based on a unique combination of the following features: transversely wide trunk vertebrae (differs in this respect from Pterosphenus schucherti, Pterosphenus schweinfurthi, and Pterosphenus kutchensis); synapophyses separated by anterior hypapophysis (differs from Pterosphenus kutchensis); synapophyses extendind below ventral cotylar rim (differs from Pterosphenus muruntau); pterapophysis dorsoventrally low, extending slightly above zygosphene (differs from Pterosphenus schucherti, Pterosphenus schweinfurthi, and Pterosphenus rannensis n. sp.); weakly concave pterapophyseal-postzygapophyseal connection (differs from Pterosphenus muruntau, Pterosphenus schucherti, and Pterosphenus schweinfurthi).
Referred specimen
Partial trunk vertebrae (RUSB 2565-1, 2790-21).
Remarks
Pterosphenus biswasi was originally described by Rage et al. (Reference Rage, Bajpai, Thewissen and Tiwari2003), together with the highly distinct Pterosphenus kutchensis, from the early Lutetian lignite deposits of Kutch (see Datta and Bajpai, Reference Datta and Bajpai2024, supplementary note 1). Rage et al. (Reference Rage, Bajpai, Thewissen and Tiwari2003) distinguished the two taxa based on the presence of an anterior hypapophysis in Pterosphenus biswasi along with short and separated synapophyses and a concave anterior margin of the zygosphene. The latter taxon was further distinguished from Pterosphenus schucherti and Pterosphenus schweinfurthi by a shallower concavity of the anterior zygosphenal margin and a higher zygosphenal plane. Recently, however, Natarajan et al. (Reference Natarajan, Dasgupta, Rakshit and Kashyap2024) synonymized Pterosphenus biswasi with Pterosphenus schucherti. These authors, while describing vertebrae attributed to Pterosphenus schucherti from the late Lutetian of Kutch, considered the previously listed features distinguishing Pterosphenus biswasi from Pterosphenus schucherti as nondiagnostic due to intra- and intercolumnar variation.
Here, we reassess the holotype and referred specimens of Pterosphenus biswasi (Fig. 8), considering its validity in the diversity of Indian palaeophiines and the stratigraphic range and evolution of Pterosphenus schucherti. Our observations find marked differences between Pterosphenus biswasi and Pterosphenus schucherti/Pterosphenus schweinfurthi with regard to the pterapophyseal morphology and overall form. Unlike Pterosphenus schucherti (IITBR1/AT5, ATV, PRW/CL = 0.8; GCVP 19863, MTV, PRW/CL = 0.8; Calvert et al., Reference Calvert, Mead and Parmley2022) and Pterosphenus schweinfurthi (CGM 10194, holotype, TV, PRW/CL = 0.8) in which the vertebrae show strong lateral compression, the vertebrae of Pterosphenus biswasi are transversely wider (RUSB 2784-4, holotype, MTV, PRW/CL = 1; RUSB 2565-1, TV, PRW/CL = 1.12; Rage et al., Reference Rage, Bajpai, Thewissen and Tiwari2003). The pterapophysis in Pterosphenus biswasi is dorsoventrally low and extends only for a short distance above the zygosphene (RUSB 2784-4, holotype, MTV, PTH/CL = 0.69, PT. VH/ZS. VH = 1; see Fig. 2). In both Pterosphenus schucherti (PTH/CL = 1.3–0.71, mean 1.1; see Natarajan et al., Reference Natarajan, Dasgupta, Rakshit and Kashyap2024, table S3; IITBR1/AT1–MT1, PT. VH/ZS. VH = 1.2–1.6) and Pterosphenus schweinfurthi (PTH/CL = 1.31–0.82, mean 1.2; see Natarajan et al., Reference Natarajan, Dasgupta, Rakshit and Kashyap2024, table S3; CGM 10194, holotype, TV, PT. VH/ZS. VH = 1.2; NHMW 2010/0188/0001d, TV, PT. VH/ZS. VH = 1.2), the pterapophyses are high and extend well above the zygosphene. Additionally, a weakly concave pterapophyseal-postzygapophyseal connection in Pterosphenus biswasi distinguishes it from Pterosphenus schucherti/Pterosphenus schweinfurthi in which this connection is markedly concave (Lucas, Reference Lucas1898; Andrews, Reference Andrews1901; McCartney and Seiffert, Reference McCartney and Seiffert2016; Georgalis, Reference Georgalis2023; Natarajan et al., Reference Natarajan, Dasgupta, Rakshit and Kashyap2024). Although a low PTH/CL ratio similar to that in Pterosphenus biswasi is also seen in Pterosphenus muruntau (PTH/CL = 0.68; see Natarajan et al., Reference Natarajan, Dasgupta, Rakshit and Kashyap2024, table S3), the latter taxon is distinguished by a higher PT. VH/ZS. VH ratio (1.4), zygosphene narrower than cotyle (ZIN PC 2/34, holotype, MTV, ZSW/COW = 0.6), and synapophysis placed dorsal to ventral cotylar rim in MTV and marked concavity of pterapophyseal-postzygapophysis connection (see Averianov, Reference Averianov2023).
In light of these morphological differences, we reassert the validity of Pterosphenus biswasi. Although the validity of some features (e.g., orientation of synapophysis, shape of posterior hypapophysis, inclination of pterapophysis and neural spine, concavity of zygosphene, width of neural canal, and height of zygapophyseal plane) traditionally used to differentiate palaeophiine taxa have been questioned (Georgalis et al., Reference Georgalis, Del Favero and Delfino2020; Georgalis, Reference Georgalis2023; Natarajan et al., Reference Natarajan, Dasgupta, Rakshit and Kashyap2024), we consider the pterapophyseal morphology and the extent of lateral vertebral compression in Pterosphenus biswasi as robust diagnostic features. Indeed, both features have been repeatedly used by previous workers to differentiate palaeophiine taxa (e.g., Georgalis et al., Reference Georgalis, Del Favero and Delfino2020; Georgalis, Reference Georgalis2023; Aniny et al., Reference Aniny, Georgalis, Gingerich and Zouhri2024; Garberoglio et al., Reference Garberoglio, Gómez and Caldwell2024). The extent of vertebral compression is also considered among the key features defining the three palaeophiine ecological grades proposed by Rage (Reference Rage1984) and Rage et al. (Reference Rage, Bajpai, Thewissen and Tiwari2003). It is noteworthy that the height of the pterapophysis and its ratio to centrum length (PTH/CL), in particular, were used by Natarajan et al. (Reference Natarajan, Dasgupta, Rakshit and Kashyap2024) to not only assign their specimens to Pterosphenus schucherti but also to distinguish the latter from Pterosphenus muruntau and Pterosphenus sheppardi. Natarajan et al. (Reference Natarajan, Dasgupta, Rakshit and Kashyap2024), however, did not evaluate the pterapophyseal height or the degree of vertebral compression of Pterosphenus biswasi when synonymizing it with Pterosphenus schucherti.
Family Madtsoiidae Hoffstetter, Reference Hoffstetter1961
Madtsoiidae gen. indet. sp. indet.
Madtsoiidae gen. indet. sp. indet., IITR/VPL/SB 2782, nearly complete middle trunk vertebra: (1) anterior; (2) left lateral (red arrows indicate fosse above and below bony strut extending posteriorly from the zygosphene; white arrowhead indicates fossa medial to diapophysis); (3) posterior (red arrow indicates endozygantral foramen); (4) inset of 9.3 at higher magnification showing paired parazygantral foramina on right parazygantral fossa; (5) dorsal (white arrowhead indicates fossa above bony strut extending posteriorly from the zygosphene); and (6) ventral views. Gray arrows indicate anterior direction. co = cotyle; cn = condyle; da = diapophysis; hk = hemal keel; izr = interzygapophyseal ridge; nal = neural arch lamina; nc = neural canal; ns = neural spine; pa = parapophysis; pcof = paracotylar foramen; pcofo = paracotylar fossa; pcon = paracotylar notch; po = postzygapophysis; pr = prezygapophysis; pt.hk = posterior terminus of hemal keel; pzgf = parazygantral foramen; pzgfo = parazygantral fossa; scf = subcentral foramen; scfo = subcentral fossa; zg = zygantrum; zs = zygosphene. Scale bars = 50 mm (except as noted).
Occurrence
Naredi Formation, middle Eocene (early Lutetian); Panandhro Lignite Mine, Kutch district, Gujarat state, western India.
Description
IITR/VPL/SB 2782 is an isolated nearly complete vertebra (Fig. 9) missing only the left zygosphenal facet and parapophysis. The specimen possibly represents a middle trunk vertebra based on the absence of a hypapophysis, pleurapophyses, and lymphapophyses, and a wider neural arch relative to centrum length (sensu Rage, Reference Rage1984; LaDuke et al., Reference LaDuke, Krause, Scanlon and Kley2010; Rio and Mannion, Reference Rio and Mannion2017).
The vertebra is massive with a centrum length (CL) and prezygapophyseal width (PRW) of 67.9 mm and 95.8 mm, respectively (Appendix 2). The centrum is procoelous with the cotyle and condyle deflected anteroventrally and posterodorsally, respectively. The cotyle is concave in anterior view and transversely wider than dorsoventrally high (COW/COH = 1.2; Fig. 9.1). It is bordered laterally by well-developed paracotylar fossae. The dorsal and ventral margins of the fossa are defined by bony struts extending from the dorsolateral and lateral cotylar margins, respectively. A secondary, weak bony strut, arising from the dorsolateral margin of the cotyle, further divides the paracotylar fossa into a deeper ventral and a shallower dorsal portion. The latter accommodates a large paracotylar foramen beside the lateroventral margin of the neural canal (Fig. 9.1). The condyle, similar to the cotyle, is mediolaterally wider than dorsoventrally high (CNW/CNH = 1.1; Fig. 9.3).
The synapophysis is dorsoventrally high and ventrolaterally inclined in anterior view (SYα = 64°). In lateral view, the synapophysis is reniform in outline and inclined at 30° from the vertical plane. A distinct diapophysis and parapophysis are present (Fig. 9.2). The parapophyseal facet is subrectangular in lateral view and remains dorsal to the ventral cotylar rim. A prominent paracotylar notch separates the parapophysis from the cotyle (Fig. 9.1). The diapophysis is bulbous and posterodorsally directed in lateral view. It extends beyond the prezygapophysis in anterior view. The dorsal margin of the diapophysis remains ventral to the dorsal cotylar rim. A large fossa is present on the lateral surface of the centrum posterior to the paradiapophysis (Fig. 9.3).
The prezygapophyseal buttress is massive and bears a weakly raised oblique ridge on its anterior surface (Fig. 9.1). A prezygapophyseal accessory process is absent and the buttress is succeeded posteriorly by a deep, elliptical fossa (Fig. 9.2). The latter is roofed by the interzygapophyseal ridge and lies medial to the diapophysis on the lateral surface of the centrum. Both the pre- and postzygapophyses are medioventrally inclined (IITR/VPL/SB 2782, PRα = 35°, POα = 31°) in anterior and posterior views (Fig. 9.1, 9.3), respectively, and bear subelliptical facets (IITR/VPL/SB 2782, PRFL/PRFW = 1.2, POFL/POFW = 1.1). In dorsal view, the prezygapophyseal articular facets are obliquely oriented at an angle of 46° to the sagittal plane, and, in anterior view, the prezygapophyses extend well above the dorsal margin of the neural canal. The interzygapophyseal ridge is robust and extends posterodorsally from the pre- to the postzygapophysis. In dorsal view. the ridge is largely straight except for a weak lateral convexity near the prezygapophysis (Fig. 9.5).
The neural canal is transversely wider than dorsoventrally high (IITR/VPL/SB 2782, NCW/NCH = 1.9; Fig. 9.1) and reniform in outline. The width of the canal resembles that of the zygosphene but is slightly less than that of the cotyle. The trapezoidal zygosphene (Fig. 9.1) is mediolaterally wider than dorsoventrally high (IITR/VPL/SB 2782, ZSW/ZSH = 1.4) and comprises a large oval facet (IITR/VPL/SB 2782, ZSFL/ZSFW = 0.9). The dorsal margin of the zygosphene is concave in anterior view, and the facet is steeply inclined (22° from the vertical). In dorsal view, the anterior zygosphenal margin is notched (ZSβ = 92°; Fig. 9.5). The zygantrum is transversely wide and comprises subelliptical facets (Fig. 9.3). The latter are steeply inclined (30° from the vertical) and accommodate paired endozygantral foramina that open posterodorsally. The zygantral roof above each facet is dorsomedially convex, and the two roofs converge along the midline to form a U-shaped embayment. The zygantrum is flanked laterally by prominent parazygantral fossae, where the left fossa accommodates two small parazygantral foramina (Fig. 9.3, 9.4).
The neural spine is dorsoventrally low (IITR/VPL/SB 2782, NSH/TVH = 0.09; Fig. 9.2) and vertically oriented. The anterior spinal margin is posterodorsally directed, whereas the posterior margin is straight and the cross section of the spine approximates the shape of an arrowhead (Fig. 9.5). The spine is buttressed posteriorly by neural arch laminae that extend posteroventrally from the dorsal margin of the neural spine to the dorsolateral margin of the postzygapophyses and enclose a deep median embayment (Fig. 9.3). The latter is mediolaterally compressed in dorsal view, with the laminae diverging at an angle of 51°. The neural spine is bound on either side by a shallow, elongated fossa (Fig. 9.2, 9.5). The latter is placed immediately posterior to the zygosphene and bordered by ventrally rounded bony struts extending from the posterolateral margin of the zygosphene. Ventral to these struts, the dorsal surface of the neural arch bears triangular fossae, with the apex of the triangle directed posteriorly (Fig. 9.2).
The centrum is triangular in ventral view and widest across the parapophyses (Fig. 9.6). Much of the ventral surface of the centrum is occupied by anteroposteriorly elongated paired subcentral fossae, which are bordered laterally by subcentral ridges. The ridges are robust and sinuous, extending posteromedially from the parapophyses to the dorsoventral midpoint of the condyle. A broad and transversely convex hemal keel separates the subcentral fossae. This keel terminates anterior to the precondylar constriction and is connected to the condyle by a sharp ridge (Fig. 9.6). At its posterior terminus, the hemal keel is chisel-shaped with sharp posteroventrally directed protuberances. Small subcentral foramen is present on either side of the hemal keel. The hemal keel, in lateral view, descends well below the level of the parapophysis (Fig. 9.2).
Material
IITR/VPL/SB 2782, a nearly complete middle trunk vertebra.
Remarks
IITR/VPL/SB 2782 is identified as a madtsoiid based on a unique combination of the following characters: presence of paracotylar and parazygantral fossae and foramina, diapophysis extending beyond the lateral margin of the prezygapophysis, prominent hemal keel with paired projections at posterior terminus, and prezygapophyseal accessory process absent. The generic affinity of this specimen was also evaluated, and in view of its large size (CL = 67.9 mm, PRW = 96.3 mm), comparisons were made with the only two large-bodied madtsoiids (Vasuki indicus and Platyspondylophis tadkeshwarensis) known from the Eocene of India (Smith et al., Reference Smith, Kumar, Rana, Folie, Solé, Noiret, Steeman, Sahni and Rose2016; Datta and Bajpai, Reference Datta and Bajpai2024).
We consider the referral of IITR/VPL/SB 2782 to the stratigraphically older Platyspondylophis (~55 Ma) unlikely because of the significantly smaller size (CL = 18–21 mm; PRW = 26–43 mm; Smith et al., Reference Smith, Kumar, Rana, Folie, Solé, Noiret, Steeman, Sahni and Rose2016) and the lack of both paracotylar and parazygantral foramina in the latter taxon. Platyspondylophis further differs from IITR/VPL/SB 2782 in bearing arcuate interzygapophyseal ridges and transverse prezygapophyseal facets, visible in dorsal view, a trilobate neural canal, an unnotched zygosphene, and a dorsoventrally higher neural spine (see Smith et al., Reference Smith, Kumar, Rana, Folie, Solé, Noiret, Steeman, Sahni and Rose2016). On the other hand, referral of IITR/VPL/SB 2782 to Vasuki seems most likely because the two were found at the same fossil locality. This referral is further supported by the following features shared by IITR/VPL/SB 2782 with the MTV of Vasuki: comparable vertebral size; cotyle and condyle transversely wider than high (IITR/VPL/SB 2782, COW/COH = 1.2, CNW/CNH = 1.1); moderately deep paracotylar fossae that are weakly subdivided into deeper ventral and shallower dorsal parts; high synapophyseal angle (IITR/VPL/SB 2782, SYα = 64°) and distinct para- and diapophyseal facts; moderately inclined pre- and postzygapophysis in anterior and posterior views, respectively (PRα = 35°, POα = 31°); reniform neural canal; trapezoidal, transversely wider than high zygosphene (IITR/VPL/SB 2782, ZSW/ZSH = 1.4); deep fossa medial to the diapophysis; large subcentral fossae with prominent subcentral foramina.
Nevertheless, a cohort of features distinguish this middle trunk vertebra (IITR/VPL/SB 2782) from those of Vasuki indicus. These include the presence of two parazygantral foramina on the right parazygantral fossa, larger paracotylar foramina, ventral position of the diapophysis relative to the dorsal cotylar rim, strongly notched anterior zygosphenal margin (IITR/VPL/SB 2782, ZSβ = 92°), weaker angle of divergence of the neural arch laminae (51°), absence of a prespinal lamina, and extension of the hemal keel below the parapophyses. These differences suggest the possible presence of a second large madtsoiid in the middle Eocene of Kutch. We consider the differences between IITR/VPL/SB 2782 and Vasuki as taxonomically significant and not intracolumnar variation because these differences are consistently present in all MTV of Vasuki indicus. In view of the morphological similarities and differences between Vasuki and IITR/VPL/SB 2782, the latter possibly represents a new species of Vasuki. Alternatively, the morphological differences in IITR/VPL/SB 2782 might reflect sexual dimorphism. LaDuke (Reference LaDuke1991) reported sexual dimorphism in Thamnophis sirtalis sirtalis Linnaeus, Reference Linnaeus1758 based on the relative size of the cotyle and condyle, with these features being larger in females. In IITR/VPL/SB 2782, however, the proportions of the cotyle and condyle resemble those in Vasuki. For these reasons and in view of the limited material currently available as well as a lack of knowledge of sexual dimorphism in madtsoiids, we refrain from naming a new species at this time.
Family ?Nigerophiidae Rage, Reference Rage1975
?Nigerophiidae gen. indet. sp. indet.
?Nigerophiidae gen. indet. sp. indet., IITR/VPL/SB 97, nearly complete anterior trunk vertebra shown by actual photographs and interpretative line drawings (in successive order): (1, 2) anterior; (3, 4) posterior; (5, 6) right lateral; (7, 8) left lateral; (9) left ventrolateral; (10, 11) dorsal; and (12, 13) ventral views. White arrows and black/white arrowheads indicate anteroposteriorly oriented fossae on lateral surface of centrum and ornamentation of neural arch, respectively. Gray arrows indicate anterior direction. co = cotyle; cn = condyle; hk = hemal keel; hy = hypapophysis; lf = lateral foramen; izr = interzygapophyseal ridge; nc = neural canal; ns = neural spine; pcof = paracotylar foramen; pcofo = paracotylar fossa; po = postzygapophysis; pr = prezygapophysis; scf = subcentral foramen; sy = synapophysis; zg = zygantrum; zs = zygosphene. Scale bars = 2 mm.
Comparison of IITR/VPL/SB 97 (?Nigerophiidae gen. indet. sp. indet.) with precloacal vertebrae of other nigerophiids and palaeophiids: (1, 2) Pterosphenus kutchensis Rage et al., Reference Rage, Bajpai, Thewissen and Tiwari2003; (3, 4) Palaeophis nessovi Averianov, Reference Averianov1997; (5, 6) Palaeophis colossaeus Rage, Reference Rage1983a; (7, 8) Indophis sahnii Rage and Prasad, Reference Rage and Prasad1992; and (9, 10) IITR/VPL/SB 97 (red dashed lines indicate missing areas of neural arch laminae dorsal to postzygapophysis): (1, 3, 5, 7, 9) posterior; (2, 4, 6, 8) left lateral; and (10) right lateral views. Schematic diagrams based on the following sources: (1, 2) RUSB 2721-1 (holotype), RUSB 2790-1 (Rage et al., Reference Rage, Bajpai, Thewissen and Tiwari2003; Garberoglio et al., Reference Garberoglio, Gómez and Caldwell2024); (3, 4) ZIN PH 119/1 (Snetkov, Reference Snetkov2011); (5, 6) MNHN.F. TGE615 (holotype; Rage, Reference Rage1983a; Garberoglio et al., Reference Garberoglio, Gómez and Caldwell2024); (7, 8) VPL/JU/500 (holotype; Rage and Prasad, Reference Rage and Prasad1992). cn = condyle; nal = neural arch lamina; nc = neural canal; ns = neural spine; po = postzygapophysis; pt = pterapophysis; zg = zygantrum. Scale bars = 2 mm.
Occurrence
Naredi Formation, Kutch district, Gujarat state, western India; early Eocene (Ypresian).
Description
IITR/VPL/SB 97 (Fig. 10) is an isolated vertebra missing the right prezygapophysis, left postzygapophysis, and part of the zygosphene and neural spine-arch complex. Based on the presence of a hypapophysis, and absence of hemapophyses, pleurapophyses, and lymphapophyses, this specimen is assigned to the precloacal region (sensu Rage, Reference Rage1984).
The vertebra is small (CL = 10.16 mm, PRW = ~10.9 mm; Appendix 2), mediolaterally compressed (PRW/CL = 1.1), and dorsoventrally high (TVH/PRW = 1.2). The cotyle is subcircular (COW/COH = 1.1) and bordered laterally by crescentic paracotylar fossae (Fig. 10.1, 10.2). These fossae are moderately deep and separated from the ventrolateral margins of the neural canal by prominent ridges. A small paracotylar foramen is present only in the left paracotylar fossa. The condyle, similar to the cotyle, is nearly circular (CNW/CNH = ~1; Fig. 10.3, 10.4). The synapophysis is ventrolaterally directed and extends slightly beyond the ventral margin of the centrum (Fig. 10.1, 10.2). Distinct parapophyseal and diapophyseal facets are not discernible.
The prezygapophysis is nearly horizontal in anterior view (PRα = 5°) and rests slightly above the ventral rim of the neural canal (Fig. 10.1, 10.2). It is laterally directed, although in dorsal view, the prezygapophysis is strongly deflected anterolaterally. The prezygapophyseal articular facet is triangular (PRFL/PRFW = 2.4). The prezygapophyseal buttress is a sharp, anterolaterally convex ridge. The postzygapophysis is weakly inclined medioventrally in posterior view and bears a subelliptical facet (POFL/POFW = 1.1; Fig. 10.3, 10.4). It is connected to the prezygapophysis by a thick and rounded interzygapophyseal ridge. Small to medium-sized foramina are present on the right lateral surface of the centrum ventral to interzygapophyseal ridge, whereas the left lateral surface bears a series of anteroposteriorly aligned large fossae (Fig. 10.5–10.9). Some of the fossae are pierced by a foramen.
Both the neural canal and zygosphene are trilobate/spade-shaped in anterior view, with the canal being slightly narrower than the cotyle (NCW/COW = 0.9; Fig. 10.1, 10.2). The dorsal margin of the zygosphene bears a sharp crest along the midline flanked by a lateral trough. It appears that the dorsal zygosphenal margin, when complete, might have had two troughs separated by a crest in anterior view. The zygosphenal facet is subtriangular and inclined at an angle of 63° from the horizontal. The zygantrum is spade-shaped in posterior view and appears partially roofed, although the zygantral roof is not well preserved (Fig. 10.3, 10.4).
The neural spine is dorsoventrally low (~6% of TVH) and tubular, with a sharp lamina extending from its anterior margin to the dorsal zygosphenal roof (Fig. 10.5–10.11). The neural spine is buttressed posteriorly by thin, posterodorsally directed neural-arch laminae, forming a shallow median embayment. Ventral to the neural spine, multiple shallow fossae that are separated by thin ridges ornament the lateral surface of the neural arch (Fig. 10.5–10.9).
The centrum is triangular in ventral view and slightly narrower posteriorly (Fig. 10.12, 10.13). It bears a sharp hemal keel that gives way posteriorly to a dorsoventrally high hypapophysis. The latter is trapezoidal in lateral view and does not extend to the posterior condyle. A subcentral foramen is present posteromedial to each synapophysis.
Material
IITR/VPL/SB 97, a nearly complete precloacal vertebra.
Remarks
IITR/VPL/SB 97 presents a taxonomic conundrum because many of the features characterizing this snake are commonly found in the aquatic groups Palaeophiidae and Nigerophiidae. Some of these features include laterally flattened vertebra, ventrally deflected synapophysis, reduced prezygapophysis, and cotyle and condyle aligned on a horizontal axis (sensu Rage et al., Reference Rage, Bajpai, Thewissen and Tiwari2003; Houssaye et al., Reference Houssaye, Rage, Bardet, Vincent, Amaghzaz and Meslouh2013). Compared with Palaeophiinae, IITR/VPL/SB 97 can be distinguished from Pterosphenus based on a trilobate/spade-shaped zygosphene and a neural spine that is separated from the dorsal zygosphenal roof. Furthermore, the lack of a prominent or reduced (knob-like) anterior hypapaophysis differentiates IITR/VPL/SB 97 from ATV and MTV of all Palaeophis species. It must be noted, however, that a single (posterior) hypapophysis characterizes Palaeophis PTV, although some morphological differences do exist between Palaeophis PTV and IITR/VPL/SB 97. In many Palaeophis species, the PTV are characterized by ovoid cotyles (e.g., Palaeophis maghrebianus [OCP DEK/ GE 517, COW/COH = 1.4], Palaeophis africanus [RMCA-RGP 16028a, COW/COH = 1.4], Palaeophis casei [COW/COH = 1.4]) and a hypapophysis that is small or even reduced to a hemal keel (Holman, Reference Holman1982; Houssaye et al., Reference Houssaye, Rage, Bardet, Vincent, Amaghzaz and Meslouh2013; Folie et al., Reference Folie, Mees, De Putter and Smith2021; Garberoglio et al., Reference Garberoglio, Gómez and Caldwell2024). This contrasts with the subcircular cotyle (COW/COH = 1.1) and large hypapophysis seen in IITR/VPL/SB 97. Archaeophiinae, which includes only two known species, Archaeophis proavus and ‘Archaeophis’ turkmenicus, can be distinguished from IITR/VPL/SB 97 by the presence of more elongated vertebrae with synapophyses that remain dorsal to the ventral surface of the centrum and the replacement of hypapophysis by hemal keel in PTV and perhaps also in MTV (Janensch, Reference Janensch1906; Tatarinov, Reference Tatarinov1963; Rage, Reference Rage1984; Rage et al., Reference Rage, Bajpai, Thewissen and Tiwari2003).
On the other hand, IITR/VPL/SB 97 shows a combination of features present in Nigerophiidae (e.g., Nigerophis mirus, Kelyophis hechti, Indophis fanambinana, I. sahnii), such as the vertebra being dorsoventrally deeper posteriorly than anteriorly, centrum narrowing posteriorly in ventral view, ventrolaterally positioned synapophyses, low, tubercular neural spine, and a weak median embayment of the neural arch (sensu Rage, Reference Rage1975, 1984; Rage and Prasad, Reference Rage and Prasad1992; LaDuke et al., Reference LaDuke, Krause, Scanlon and Kley2010; Pritchard et al., Reference Pritchard, McCartney, Krause and Kley2014). Another feature that could further support the possible attribution of IITR/VPL/SB 97 to Nigerophiidae as opposed to Palaeophiidae, is the absence of a pterapophysis based on comparison with the corresponding regions in known nigerophiids and palaeophiids (Fig. 11). Although the neural arch laminae are partially preserved in IITR/VPL/SB 97, the left lamina is strongly ventrolaterally directed and missing only a small portion dorsal to the postzygapophysis. On the other hand, the vertebral dimensions of IITR/VPL/SB 97 are significantly larger than those of any known nigerophiid, and the anterior extent of the neural spine is reminiscent of the condition in palaeophiids. If IITR/VPL/SB 97 is indeed a nigerophid, it represents an ATV as suggested by the presence of a single posterior hypapophysis (sensu Rage, Reference Rage1984). Limited material currently precludes secure referral of this specimen.
Phylogenetic analysis
Taxa
We only tested the phylogenetic position of Pterosphenus rannensis n. sp. and Pterosphenus biswasi because they are known from multiple specimens representing the precloacal region whereas the other taxa described in this study are represented by single vertebrae. A modified version of the dataset from Snetkov (Reference Snetkov2011) was used in the present analysis and comprises 15 taxa including four ‘primitive’-grade Palaeophis (i.e., Palaeophis colossaeus, Palaeophis africanus, Palaeophis maghrebianus, and Palaeophis vastaniensis) and three ‘advanced’-grade Palaeophis (i.e., Palaeophis toliapicus, Palaeophis typhaeus, Palaeophis nessovi) species along with Palaeophis oweni, and five species of Pterosphenus (i.e., Pterosphenus schucherti, Pterosphenus schweinfurthi, Pterosphenus kutchensis, Pterosphenus biswasi, and Pterosphenus muruntau). The basal ophidian Dinilysia patagonica Smith-Woodward, Reference Smith-Woodward1901 was used as the outgroup. Sources of information are as follows: D. patagonica (see Caldwell and Albino, Reference Caldwell and Albino2003; Scanferla and Canale, Reference Scanferla and Canale2007), Palaeophis colossaeus (Rage, Reference Rage1983a; Snetkov, Reference Snetkov2011; McCartney et al., Reference McCartney, Roberts, Tapanila and Oleary2018; Georgalis et al., Reference Georgalis, Guinot, Kassegne, Amoudji, Johnson, Cappetta and Hautier2021a), Palaeophis maghrebianus (Snetkov, Reference Snetkov2011; Houssaye et al., Reference Houssaye, Rage, Bardet, Vincent, Amaghzaz and Meslouh2013; Georgalis et al., Reference Georgalis, Guinot, Kassegne, Amoudji, Johnson, Cappetta and Hautier2021a), and Palaeophis vastaniensis (Bajpai and Head, Reference Bajpai and Head2007; Snetkov, Reference Snetkov2011); Palaeophis toliapicus (Owen, Reference Owen1841; Rage, Reference Rage1983b; Snetkov, Reference Snetkov2011; Zvonok and Snetkov, Reference Zvonok and Snetkov2012), Palaeophis typhaeus (Owen, Reference Owen and Owen1850; Rage, Reference Rage1983b); Palaeophis oweni (Georgalis et al., Reference Georgalis, Del Favero and Delfino2020); Palaeophis africanus (Folie et al., Reference Folie, Mees, De Putter and Smith2021; Georgalis et al., Reference Georgalis, Guinot, Kassegne, Amoudji, Johnson, Cappetta and Hautier2021a), Palaeophis nessovi (Snetkov, Reference Snetkov2011), Pterosphenus schucherti (Holman, Reference Holman1977; Snetkov, Reference Snetkov2011; Calvert et al., Reference Calvert, Mead and Parmley2022), Pterosphenus schweinfurthi (Snetkov, Reference Snetkov2011; McCartney and Seiffert, Reference McCartney and Seiffert2016; Georgalis, Reference Georgalis2023), Pterosphenus kutchensis (Rage et al., Reference Rage, Bajpai, Thewissen and Tiwari2003; Snetkov, Reference Snetkov2011), Pterosphenus biswasi (Rage et al., Reference Rage, Bajpai, Thewissen and Tiwari2003; Snetkov, Reference Snetkov2011), and Pterosphenus muruntau (Averianov, Reference Averianov2023).
Characters
The analysis was based exclusively on vertebral characters because knowledge of the palaeophiine skull is nonexistent (Folie et al., Reference Folie, Mees, De Putter and Smith2021). We emphasize that although snake vertebrae are highly useful for taxonomic identification, phylogenetic considerations based on them should be treated with caution because snake vertebrae are strongly influenced by convergence in which many features are repeatedly and independently acquired across diverse lineages (Smith and Georgalis, Reference Smith, Georgalis, Gower and Zaher2022; Szyndlar and Georgalis, Reference Szyndlar and Georgalis2023). The phylogenetic analysis presented here is subject to possible uncertainties arising from our limited understanding of intracolumnar variation in palaeophiines (Houssaye et al., Reference Houssaye, Rage, Bardet, Vincent, Amaghzaz and Meslouh2013; Folie et al., Reference Folie, Mees, De Putter and Smith2021) and should therefore be treated with caution. We still tested the phylogenetic position of Pterosphenus rannensis n. sp. (considering its intermediate morphology between Palaeophis and Pterosphenus) and that of Pterosphenus biswasi in view of the concerns raised about its validity. Thirty vertebral characters (Appendix 3) were used of which 26 were from Snetkov (Reference Snetkov2011). Four of these characters were modified (characters [chs.] 1, 11, 12, 14) and four new characters were added (chs. 27−30). Furthermore, to have at least twice the number of characters than species analyzed (sensu Hammer and Harper, Reference Hammer and Harper2006; Folie et al., Reference Folie, Mees, De Putter and Smith2021), the analysis included only 15 taxa.
Results
The character-taxon matrix comprising 15 taxa and 30 characters (Supplementary Dataset 1) was analyzed using TNT version 1.6 (Goloboff and Morales, Reference Goloboff and Morales2023) with the software memory set to retain 10,000 trees and a display buffer of 10 Mb. The Traditional Search option was used in which the analysis constraints included 50 replications of Wagner trees, with bisection reconnection as the swapping algorithm, and 10 trees saved per replication. Robustness of nodes was determined from (1) Bremer support values, calculated using script bremer.run in which only trees suboptimal by 20 steps were retained, and (2) symmetric resampling values based on 1,000 replicates and expressed as frequency differences. Three most parsimonious trees (MPTs; Fig. 12.1, 12.2) were recovered with a tree length of 69, consistency index (CI) of 0.449, and retention index (RI) of 0.6. In both the strict and majority-rule trees, Pterosphenus was well resolved and a later-diverging relative to all Palaeophis snakes. Within Pterosphenus, Pterosphenus rannensis n. sp. was the earliest-diverging taxon followed by Pterosphenus biswasi, whereas Pterosphenus schucherti and Pterosphenus schweinfurthi constituted the latest-diverging subclade. Four unambiguous synapomorphies support the recovery of Pterosphenus rannensis n. sp. as the basalmost Pterosphenus: zygosphene with dorsoventrally high midline crest (ch. 1), thick zygosphene (ch. 2), anterior margin of neural spine extends to the dorsal zygosphenal roof (ch. 14), and distance between prezygapophysis less than the distance measured from the ventral cotylar rim to the dorsal zygosphenal margin (ch. 22). Furthermore, two local autapomorphies diagnose Pterosphenus rannensis n. sp.: absence of subcentral foramina (ch. 15) and pterapophysis remaining below the dorsal zygosphenal margin (ch. 28).
(1) Strict consensus tree of three MPTs showing palaeophiinae inter-relationships, and (2) 50% majority-rule tree showing inter-relationships of palaeophiine snakes; Pterosphenus rannensis n. sp. given in bold. Numbers below nodes indicate the frequency that a clade is represented in the MPTs (left) and Bremer support values (right). Numbers above nodes indicate group present/contradicted (GC) symmetric resampling values.
Unlike Pterosphenus, a fully resolved picture of Palaeophis interrelationships is only seen in the majority-rule tree (Fig. 12.2). Here, Palaeophis vastaniensis is the basalmost palaeophiine with the other ‘primitive’-grade paleophiids (Palaeophis maghrebianus and Palaeophis africanus) also restricted to the basal part of the tree. An exception to this is Palaeophis colossaeus, which, together with the ‘advanced’-grade Palaeophis toliapicus, constituted the latest-diverging clade of Palaeophis closest to Pterosphenus. As for the other ‘advanced’-grade palaeophiids, Palaeophis typhaeus was a sister taxon to Palaeophis oweni, whereas Palaeophis nessovi was recovered in a more basal position, being phylogenetically bracketed by Palaeophis maghrebianus and Palaeophis africanus.
Discussion
The in-group relationships of Palaeophiinae obtained here failed to recover a monophyletic Palaeophis, consistent with the results of Snetkov (Reference Snetkov2011) and Folie et al. (Reference Folie, Mees, De Putter and Smith2021). Rather, the Palaeophis species formed successive outgroups to Pterosphenus, on which the earliest branching were, mostly, the ‘primitive’-grade Palaeophis taxa, followed by the ‘advanced’-grade forms (Fig. 12.1, 12.2). Such an arrangement of taxa essentially mirrors the three ecological grades identified within Palaeophiinae (sensu Rage et al., Reference Rage, Bajpai, Thewissen and Tiwari2003) because it presents a morphological series ranging from snakes that are poorly adapted to aquatic life to those that are strongly adapted. Some of the major evolutionary trends defining this series are consistent with the morphological changes associated with the ecological grades by Rage et al. (Reference Rage, Bajpai, Thewissen and Tiwari2003) and include progressive thickening of the zygosphene and development of a dorsoventrally high central crest on the dorsal zygosphenal surface (chs. 1, 2), increase in the ventral extent of the synapophyses (ch. 11), increased lateral flattening of vertebrae (ch. 12), anterior margin of the neural spine extending to the zygosphene (ch. 14), progressive increase in height of the pterapophysis relative to the dorsal margin of zygosphene and the length of the centrum (chs. 28, 29), and development of a marked concavity along the pterapophysis-postzygapophysis connection (ch. 30). The only taxon that diverges from this morphological series is Palaeophis colossaeus, which has a sister-taxon relationship with the ‘advanced’-grade Palaeophis toliapicus despite being identified as a ‘primitive’-grade form (sensu Rage et al., Reference Rage, Bajpai, Thewissen and Tiwari2003). A possible explanation for this could be the presence of some shared features between Palaeophis colossaeus and Palaeophis toliapicus, e.g., a thick zygosphene (ch. 2), a weak interzygapophyseal ridge (ch. 6), angle formed by V-shaped incision on the posterior part of the neural arch > 105° (ch. 25), and the pterapophysis extending slightly above the zygosphene (ch. 28).
The Pterosphenus in-group relationships obtained here corroborate our taxonomic inferences concerning Pterosphenus rannensis n. sp. and Pterosphenus biswasi. The new taxon Pterosphenus rannensis n. sp. presents a host of morphological features that characterize both Palaeophis [e.g., shallow ventral extent of the synapophyses (ch. 11); weak lateral flattening of the vertebrae (ch. 12); pterapophysis-postzygapophysis connection lacking a marked concavity (ch. 30)] and Pterosphenus (e.g., dorsoventrally high central crest on the dorsal surface of the zygosphene [ch. 1]; anterior margin of the neural spine extending to the zygosphene [ch. 14]). This reflects the intermediate phylogenetic position of this taxon relative to Pterosphenus and Palaeophis and supports our osteological study, which highlighted the intermediate morphology of Pterosphenus rannensis n. sp. The anatomical and phylogenetic position of Pterosphenus rannensis n. sp. and African Pterosphenus of similar nature (see Zouhri et al., Reference Zouhri, Khalloufi, Bourdon, de Broin, Rage, M’haïdrat, Gingerich and Elboudali2018) hint at the possibility of the distinction between Pterosphenus and Palaeophis being artificial and highlights the need for a revision of the subfamily Palaeophiinae, as suggested by Rage et al. (Reference Rage, Bajpai, Thewissen and Tiwari2003) and Smith and Georgalis (Reference Smith, Georgalis, Gower and Zaher2022). The early-diverging position of Pterosphenus biswasi, distinct from Pterosphenus schucherti and Pterosphenus schweinfurthi, supports its validity. Its distinctiveness from Pterosphenus schucherti and Pterosphenus schweinfurthi was supported by a combination of features, including a prominent interzygapophyseal ridge (ch. 6), weak lateral compression of vertebrae (ch. 12), centrum length less than the distance between the lateral margin of synapophyses (ch.13), pterapophysis extending slightly above the zygosphene (ch. 28), PTH/CL < 0.7 (ch. 29), and pterapophysis-postzygapophysis connection without marked concavity (ch 30). It should be noted that Snetkov (Reference Snetkov2011) in his phylogenetic analysis of Palaeophiidae also recovered Pterosphenus biswasi as distinct from Pterosphenus schucherti and Pterosphenus schweinfurthi.
Significance of the fauna
Paleobiogeogrpahy and paleoecology
Fossil snakes from the Eocene of Kutch described herein represent a varied fauna that has elements that are both cosmopolitan and predominantly Gondwanan in nature. The madtsoiid (IITR/VPL/SB 2782) and possible nigerophiid (IITR/VPL/SB 97) from the Naredi Formation represent Gondwanan elements because these ophidian families are largely known from Gondwanan landmasses (Rage and Prasad, Reference Rage and Prasad1992; LaSuke et al., 2010; Rio and Mannion, Reference Rio and Mannion2017; Datta and Bajpai, Reference Datta and Bajpai2024). The cosmopolitan elements of this fauna are represented by the palaeophiines, Pterosphenus rannensis n. sp., and IITR/VPL/SB 2632 from the stratigraphically younger Harudi Formation.
Palaeophiinae
The presence of palaeophiine fossils in Eocene deposits is not unexpected considering their prolific diversification during this period (McCartney and Seiffert, Reference McCartney and Seiffert2016; Folie et al., Reference Folie, Mees, De Putter and Smith2021; Smith and Georgalis, Reference Smith, Georgalis, Gower and Zaher2022; Garberoglio et al., Reference Garberoglio, Gómez and Caldwell2024). The oldest record of Palaeophiinae comes from the Late Cretaceous of Africa, followed by their restricted occurrence in the Paleocene (Rage and Wouters, Reference Rage and Wouters1979; Erickson, Reference Erickson1998; Garberoglio et al., Reference Garberoglio, Gómez and Caldwell2024). These snakes became truly cosmopolitan only during the Eocene, with reports from Africa, the Americas, Asia, Europe, and India (Rage, Reference Rage1983b; Rage et al., Reference Rage, Bajpai, Thewissen and Tiwari2003, Reference Rage, Folie, Rana, Singh, Rose and Smith2008; Folie et al., Reference Folie, Mees, De Putter and Smith2021; Georgalis et al., Reference Georgalis, Guinot, Kassegne, Amoudji, Johnson, Cappetta and Hautier2021a; Smith and Georgalis, Reference Smith, Georgalis, Gower and Zaher2022). Among the two phenotypic genera nested within this subfamily (sensu Rage et al., Reference Rage, Bajpai, Thewissen and Tiwari2003), the Palaeophis-type snakes show a wider temporal range extending from the Late Cretaceous to the late Eocene (Rage et al., Reference Rage, Folie, Rana, Singh, Rose and Smith2008; Head et al., Reference Head, Howard, Müller, Gower and Zaher2022; Smith and Georgalis, Reference Smith, Georgalis, Gower and Zaher2022). Pterosphenus-type snakes, on the other hand, are restricted to the Eocene, with most taxa reported from horizons that are either early Lutetian or Bartonian-Priabonian in age (Rage et al., Reference Rage, Folie, Rana, Singh, Rose and Smith2008; Folie et al., Reference Folie, Mees, De Putter and Smith2021; Smith and Georgalis, Reference Smith, Georgalis, Gower and Zaher2022; Georgalis, Reference Georgalis2023; Datta and Bajpai, Reference Datta and Bajpai2024, supplementary note 1). There appear to be no known Ypresian and middle/late Lutetian occurrences globally, possibly as a result of sampling bias, except for an isolated vertebra from the Cambay Shale Formation of India, which is the only record of this genus from the Ypresian (Rage et al., Reference Rage, Folie, Rana, Singh, Rose and Smith2008). The late Lutetian palaeophiine taxa described here (Pterosphenus rannensis n. sp. and ?Palaeophis sp.), along with the recently described remains of Pterosphenus schucherti from Kutch (Natarajan et al., Reference Natarajan, Dasgupta, Rakshit and Kashyap2024), bridge this temporal gap. These taxa also represent stratigraphically the youngest record of Palaeophiidae in the Indian subcontinent. The phylogenetic position of Pterosphenus rannensis n. sp. presents some interesting biogeographic scenarios. The basal position of Pterosphenus rannensis n. sp. relative to the other Indian Pterosphenus-grade snakes (Pterosphenus biswasi, Pterosphenus kutchensis) needs to be reconciled in view of the younger age of Pterosphenus rannensis n. sp. The latter taxon appears to be the relic of a long-surviving, early-diverging lineage that gave rise to the more-derived Pterosphenus biswasi and Pterosphenus kutchensis. Moreover, the phylogenetic placement of Pterosphenus rannensis n. sp. also highlights the significance of the Indian record to understanding the origin and diversification of the genus, mainly due to the intermediate morphology of Pterosphenus rannensis n. sp., which is reminiscent of both Palaeophis and Pterosphenus. In this scenario, the Ypresian Pterosphenus sp. indet. (GU/RSR/VAS 1009) from the Cambay Shale Formation becomes important because it represents the earliest global record of this genus (Rage et al., Reference Rage, Folie, Rana, Singh, Rose and Smith2008) although its poor preservation does not allow for specific identification at present. Alternatively, a dispersal event from Africa to explain the occurrence of Pterosphenus rannensis n. sp. cannot be ruled out considering the marine adaptations of palaeophiine snakes and the recovery of the oldest (Late Cretaceous) member of this subfamily (Palaeophis sp.) from that continent. Faunal exchanges between Africa and the Indian subcontinent during the Paleogene have been previously suggested for a diverse array of aquatic taxa including pelomedusoid turtles, dyrosaurid crocodyliforms, and madtsoiid and colubroid snakes (Smith et al., Reference Smith, Kumar, Rana, Folie, Solé, Noiret, Steeman, Sahni and Rose2016; Zaher et al., Reference Zaher, Folie, Quadros, Rana, Kumar, Rose, Fahmy and Smith2021; Datta and Bajpai, Reference Datta and Bajpai2024). Instances of an African-North American transoceanic dispersal has been observed in Palaeophis africanus and Pterosphenus schweinfurthi (Andrews, Reference Andrews1924; Parmley and DeVore, Reference Parmley and DeVore2005; McCartney and Seiffert, Reference McCartney and Seiffert2016; Georgalis et al., Reference Georgalis, Guinot, Kassegne, Amoudji, Johnson, Cappetta and Hautier2021a; Smith and Georgalis, Reference Smith, Georgalis, Gower and Zaher2022; Georgalis, Reference Georgalis2023). A similar dispersal event from India, across marine waters, could also explain the phylogenetic position of the Asian Pterosphenus muruntau (Bartonian; Averianov, Reference Averianov1997, Reference Averianov2023), which is phylogenetically bracketed by the stratigraphically older Pterosphenus biswasi and Pterosphenus kutchensis. Nonetheless, the biogeographic scenarios presented here should be treated as tentative considering that the present phylogenetic results are based solely on vertebrae. We also stress the need for more extended collection efforts in the Cretaceous and Palaeogene deposits across the globe because it could lead to recovery of more phylogenetically informative material allowing a better understanding of Palaeophiinae biogeography.
An aquatic habitus is typically inferred for palaeophiine snakes based on vertebral morphology and because their fossil remains have been recovered from various aquatic environments ranging from lagoons, estuaries, mangrove forests, and near-coastal fluvial to shallow/nearshore marine settings (Hoffstetter, Reference Hoffstetter1958; Westgate and Ward, Reference Westgate and Ward1981; Holman, Reference Holman1982; Rage, Reference Rage1983a; Hutchison, Reference Hutchison1985; Westgate and Gee, Reference Westgate and Gee1990; Rage et al., Reference Rage, Bajpai, Thewissen and Tiwari2003; Parmley and DeVore, Reference Parmley and DeVore2005; Houssaye et al., Reference Houssaye, Rage, Bardet, Vincent, Amaghzaz and Meslouh2013; McCartney et al., Reference McCartney, Roberts, Tapanila and Oleary2018; Folie et al., Reference Folie, Mees, De Putter and Smith2021; Georgalis et al., Reference Georgalis, Guinot, Kassegne, Amoudji, Johnson, Cappetta and Hautier2021a). However, based on the degree of development of vertebral features associated with aquatic adaptations (e.g., laterally compressed vertebrae, ventrally shifted synapophyses with facets facing ventrally to place the ribs beneath the vertebrae, and the presence of pterapophyses), Pterosphenus snakes are considered better adapted to this mode of life than Palaeophis-grade snakes. Accordingly, the latter taxa are considered as nonpelagic, whereas remains of Pterosphenus have been reported from an open marine environment (Westgate, Reference Westgate and Gunnel2001; Rage et al., Reference Rage, Bajpai, Thewissen and Tiwari2003; Houssaye et al., Reference Houssaye, Rage, Bardet, Vincent, Amaghzaz and Meslouh2013).
From a paleoecological standpoint, the two palaeophiines described here, Pterosphenus rannensis n. sp. and ?Palaeophis sp. (IITR/VPL/SB 2632), together with the recently described Pterosphenus schucherti by Natarajan et al. (Reference Natarajan, Dasgupta, Rakshit and Kashyap2024), present an ecological conundrum. All three taxa are from the Harudi Formation, albeit from different localities, and present a scenario rarely seen among palaeophiines. The co-occurrence of Palaeophis and Pterosphenus in the same stratigraphic context is known only from the early Eocene Cambay Shale Formation of India, and the late Eocene of the Hardie Mine in the USA (Parmley and DeVore, Reference Parmley and DeVore2005; Rage et al., Reference Rage, Folie, Rana, Singh, Rose and Smith2008). Ecological segregation/niche partitioning seems to be a likely explanation for this conundrum in view of the hypothesized differences in aquatic adaptaions between Pterosphenus and Palaeophis (Westgate, Reference Westgate and Gunnel2001; Rage et al., Reference Rage, Bajpai, Thewissen and Tiwari2003; Houssaye et al., Reference Houssaye, Rage, Bardet, Vincent, Amaghzaz and Meslouh2013). Natarajan et al. (Reference Natarajan, Dasgupta, Rakshit and Kashyap2024) considered a marginal marine habitat for Pterosphenus schucherti based on nearshore bivalves, gastropods, and nautiloids from the fossil locality at Rato River (the type section of the Harudi Formation). Pterosphenus schucherti is generally considered to have lived in coastal to shallow shelf waters, similar to the extant yellow-bellied sea snake, and sporadically moved into the open ocean (Hecht et al., Reference Hecht, Kropach and Hecht1974; Natarajan et al., Reference Natarajan, Dasgupta, Rakshit and Kashyap2024). As for Pterosphenus rannensis n. sp. and ?Palaeophis sp., the inferred depositional setting of the fossil localities indicates contrasting habitats. Based on sedimentology, tidal and lagoonal lithofacies (Mukhopadhyay and Shome, Reference Mukhopadhyay and Shome1996; Thewissen and Bajpai, Reference Thewissen and Bajpai2009) were suggested for the Godhatad and Dhedidi North localites from which came the fossils of Pterosphenus rannensis n. sp. Evidence for the depositional conditions also comes from the associated fauna of fossil whales (Thewissen and Bajpai, Reference Thewissen and Bajpai2009) and mollusks including oysters (Banerjee et al., Reference Banerjee, Das, Halder and Chakrabarti2019). Protocetids, which occur abundantly at the Pterosphenus rannensis n. sp.-yielding beds at Godhatad and Dhedidi North, support a tidal to shallow coastal marine environment (Thewissen and Bajpai, Reference Thewissen and Bajpai2009; Thewissen et al., Reference Thewissen, Cooper, George and Bajpai2009; Bajpai and Thewissen, Reference Bajpai and Thewissen2014). Regarding ?Palaeophis sp., a marshy/swampy environment was proposed for the Babia Hill locality (Mukhopadhyay and Shome, Reference Mukhopadhyay and Shome1996). The abundance of remingtonocetines and near absence of protocetids at this locality corroborate the inferred paleoenviroment, because the small orbits in Remingtonocetus Kumar and Sahni, Reference Kumar and Sahni1986 have been considered an adaptation to the muddy waters expected in a swamp/marsh (Thewissen and Bajpai, Reference Thewissen and Bajpai2009; Bajpai et al., Reference Bajpai, Thewissen and Conley2011). The rarity of ?Palaeophis (single vertebra) at this locality, however, might suggest that this specimen was transported.
Interestingly, the vertebral morphology of the two new Harudi palaeophiines is not consistent with their inferred paleohabitats. The vertebrae of Pterosphenus rannensis n. sp. do not show strong near-shore aquatic adaptations commonly associated with Pterosphenus, e.g., strongly laterally compressed vertebrae with high pterapophyses (sensu Rage et al., Reference Rage, Bajpai, Thewissen and Tiwari2003). ?Palaeophis sp., which possibly lived in a swamp/marsh, shows pronounced aquatic adaptations, e.g., strong lateral vertebral compression and a high pterapophysis (see McCartney et al., Reference McCartney, Roberts, Tapanila and Oleary2018). Another example reminiscent of Pterosphenus rannensis n. sp. is Palaeophis colossaeus, which comes from a nearshore marine environment, although this snake is not considered to have strong adaptations to aquatic life (Rage, Reference Rage1983a; McCartney et al., Reference McCartney, Roberts, Tapanila and Oleary2018). Our study thus highlights the need for re-evaluation of morphological correlates of aquatic adaptations in fossil snakes. Although the present observations seemingly support the presence of niche partitioning among the two Harudi palaeophiines described here, we still consider this inference tentative and subject to confirmation from additional material. We recognize that the burial site of fossils, which in our case includes disarticulated/isolated vertebrae, does not always coincide with the habitat area. Nonetheless, the Kutch specimens of Pterosphenus rannensis n. sp. do not appear to have suffered long transportation because they are not strongly abraded or deformed, as evident from the high angularity and low sphericity of the specimens (see Mukherjee and Ray, Reference Mukherjee and Ray2012; Datta et al., Reference Datta, Mukherjee and Ray2020).
Madtsoiidae
IITR/VPL/SB 2782 potentially enriches the known diversity of Indian madtsoiids, the temporal range of which extends from the latest Cretaceous to the late Oligocene (Rage et al., Reference Rage, Prasad and Bajpai2004; Mohabey et al., Reference Mohabey, Head and Wilson Mantilla2011; Smith et al., Reference Smith, Kumar, Rana, Folie, Solé, Noiret, Steeman, Sahni and Rose2016; Wazir et al., Reference Wazir, Sehgal, Čerňanský, Patnaik, Kumar, Singh, Uniyal and Singh2021; Head et al., Reference Head, Howard, Müller, Gower and Zaher2022; Smith and Georgalis, Reference Smith, Georgalis, Gower and Zaher2022; Datta and Bajpai, Reference Datta and Bajpai2024). Although an isolated vertebra does not allow a confident generic or specific designation, the morphological similarities and differences of IITR/VPL/SB 2782 with the contemporaneous (early Lutetian) Vasuki indicus highlight the uniqueness of this taxon. It is likely that IITR/VPL/SB 2782 and Vasuki belong to the same lineage. Datta and Bajpai (Reference Datta and Bajpai2024) argued for an Indian origin of the madtsoiid lineage leading to Vasuki, and we think that IITR/VPL/SB 2782 provides further evidence for the evolutionary diversification of this lineage. The vertebral morphology of IITR/VPL/SB 2782 also allows for paleoecological interpretations. An ecological habitus similar to that of Vasuki (back-swamp marsh; Datta and Bajpai, Reference Datta and Bajpai2024) is envisaged for IITR/VPL/SB 2782 based on their shared morphology, consistent with previous paleoenvironmental interpretations based on sedimentological data (Mukhopadhyay and Shome, Reference Mukhopadhyay and Shome1996) and associated vertebrates (abundant mylobatids, catfishes, sharks, crocodilians, trionychid turtles, and archaic whales) (Bajpai and Thewissen, Reference Bajpai and Thewissen2002). IITR/VPL/SB 2782 was a large, slow-moving snake with a cylindrical body, given that the vertebra is longitudinally short and transversely wide, with laterally facing synapophyses that would have supported laterally directed ribs (sensu Mosauer, Reference Mosauer1932; LaDuke et al., Reference LaDuke, Krause, Scanlon and Kley2010; McCartney et al., Reference McCartney, Roberts, Tapanila and Oleary2018; Datta and Bajpai, Reference Datta and Bajpai2024). These features argue against an aquatic lifestyle for IITR/VPL/SB 2782 (Rage et al., Reference Rage, Bajpai, Thewissen and Tiwari2003; Houssaye et al., Reference Houssaye, Rage, Bardet, Vincent, Amaghzaz and Meslouh2013; McCartney et al., Reference McCartney, Roberts, Tapanila and Oleary2018; Georgalis, Reference Georgalis2023). We, however, cannot completely rule out an aquatic lifestyle for IITR/VPL/SB 2782 because some features associated with aquatic adaptations show inconsistencies in some aquatic snakes. For example, true lateral vertebral compression is only seen in the caudal region of hydrophiine sea snakes, whereas in the Cretaceous aquatic snake Simoliophis Sauvage, Reference Sauvage1880, the synapophyses shift from ventrally to laterally facing across the vertebral column (Rage et al., Reference Rage, Vullo and Neraudeau2016; Datta and Bajpai, Reference Datta and Bajpai2024). An arboreal or fossorial lifestyle also seems unlikely for IITR/VPL/SB 2782 because of its large size, obliquely directed prezygapophyseal articular facets in dorsal view, and a broad hemal keel (see Szyndlar and Georgalis, Reference Szyndlar and Georgalis2023). Furthermore, the vertebrae of arboreal snakes are often longer and have shorter zygapophyses, whereas fossorial snakes tend to have depressed neural arch-spine complexes that place the epaxial muscles close to the sagittal plane (Johnson, Reference Johnson1955; Auffenberg, Reference Auffenberg1961; LaDuke et al., Reference LaDuke, Krause, Scanlon and Kley2010). Therefore, a terrestrial/semiaquatic mode of life is more likely for IITR/VPL/SB 2782, as previously suggested for Vasuki indicus (sensu Datta and Bajpai, Reference Datta and Bajpai2024). Moreover, like the latter taxon, this snake would have been an ambush predator that subdued prey by constriction.
Nigerophiidae
IITR/VPL/SB 97 is tentatively assigned to Nigerophiidae and is among the oldest fossil snakes from the Paleogene of India (Ypresian). If this assignment is accurate, then this specimen represents one of the oldest Cenozoic records of Nigerophiidae globally, and possibly the first record of this family from the Cenozoic of South Asia. Unfortunately, given its tentative taxonomic status, IITR/VPL/SB 97 does not provide reliable insights into the biogeography of Nigerophiidae. The same can also be said for other known Late Cretaceous/Paleogene nigerophiids because of the uncertainties surrounding the monophyly of this family and its in-group relations (LaDuke et al., Reference LaDuke, Krause, Scanlon and Kley2010; Pritchard et al., Reference Pritchard, McCartney, Krause and Kley2014). However, the discovery of the Indian nigerophiid Indophis Rage and Prasad, Reference Rage and Prasad1992 from the Late Cretaceous of Madagascar has led some authors to propose Indo-Madagascar biotic links during this geological interval (Pritchard et al., Reference Pritchard, McCartney, Krause and Kley2014; Rage et al., Reference Rage, Prasad, Verma, Khosla, Parmar, Prasad and Patnaik2020). Paleoecologically, nigerophiids have traditionally been regarded as aquatic snakes based on vertebral morphology (e.g., ventrally shifted synapophyses, narrow prezygapophyseal buttress with anterolateral convexity and high vertebrae), with fossils recovered from freshwater lacustrine to marginal/shallow marine deposits (Rage and Prasad, Reference Rage and Prasad1992; Prasad and Khajuria, Reference Prasad and Khajuria1996; Rage et al., Reference Rage, Prasad and Bajpai2004, Reference Rage, Prasad, Verma, Khosla, Parmar, Prasad and Patnaik2020; LaDuke et al., Reference LaDuke, Krause, Scanlon and Kley2010; Houssaye et al., Reference Houssaye, Rage, Bardet, Vincent, Amaghzaz and Meslouh2013; McCartney et al., Reference McCartney, Roberts, Tapanila and Oleary2018; Smith and Georgalis, Reference Smith, Georgalis, Gower and Zaher2022). We also envisage an aquatic lifestyle for IITR/VPL/SB 97 based on the ventrally deflected synapophyses that extend below the centrum, the sharp and anterolaterally convex prezygapophyseal buttress, and the dorsoventral height of the vertebra. Supporting evidence comes from the depositional environment of the horizon yielding IITR/VPL/SB 97, which was reconstructed as a brackish to marine inner shelf based on benthic foraminiferans, bivalves, and fishes including gars (Keller et al., Reference Keller, Khozyem, Saravanan, Adatte, Bajpai and Spangenberg2013; Khozyem et al., Reference Khozyem, Adatte, Keller, Spangenberg, Saravanan and Bajpai2013).
Conclusions
A diverse snake fauna, comprising both cosmopolitan taxa and Gondwana holdovers from the Late Cretaceous, is described from the Eocene of Kutch, western India. The Cretaceous relics include a large madtsoiid and a possible nigerophiid, the former being sympatric to the early Lutetian giant madtsoiid Vasuki indicus. The nigerophiid, on the other hand, is stratigraphically among the oldest Paleogene snakes known from India. The cosmopolitan elements, represented by late Lutetian palaeophiines, are represented by Pterosphenus (Pterosphenus rannensis n. sp.) and ?Palaeophis, the latter sharing affinities with ‘advanced’-grade Palaeophis. Pterosphenus rannensis n. sp. is characterized by an intermediate vertebral morphology that is reminiscent of both Pterosphenus and Palaeophis.
Pterosphenus biswasi, originally described by Rage et al. (Reference Rage, Bajpai, Thewissen and Tiwari2003) from the middle Eocene of Kutch and subsequently synonymized with Pterosphenus schucherti (Natarajan et al., Reference Natarajan, Dasgupta, Rakshit and Kashyap2024), is re-examined here. Based on the presence of a low pterapophysis and weak lateral vertebral flattening, we retain Pterosphenus biswasi as a valid taxon.
Phylogenetic analysis corroborated our anatomical observations because it recovered Pterosphenus rannensis n. sp. and Pterosphenus biswasi at the base of the clade comprising all Pterosphenus snakes. Pterosphenus rannensis n. sp. is the earliest-diverging member of this clade and occupies an intermediate position between Pterosphenus- and Palaeophis-grade snakes. Furthermore, the biogeography of palaeophiines, seen in conjunction with the phylogenetic position and stratigraphic age of Pterosphenus rannensis n. sp., highlights the importance of India as a center for the diversification of Pterosphenus.
Based on vertebral morphology, an aquatic habitus is suggested for most taxa in this fauna except for the madtsoiid, which was terrestrial/semiaquatic. Niche partitioning is observed among the two palaeophiine snakes, based on depositional environment of the fossil-yielding horizons.
The present study provides, for the first time, insights into ophidian diversity during the middle Eocene of India, besides augmenting the known diversity of Indian Eocene snakes. When compared with the subcontinent’s Late Cretaceous record, the Eocene snake diversity shows marked differences in faunal composition and ecological preferences. The Cretaceous fauna is largely known from terrestrial/freshwater settings and includes stem snakes (Coniophis Marsh, Reference Marsh1892; Sanajeh Wilson Mantilla et al., Reference Wilson Mantilla, Mohabey, Peters and Head2010), madtsoiids (indeterminate forms; Madtsoia pisdurensis), and a nigerophiid (Indophis sahnii) (Rage and Prasad, Reference Rage and Prasad1992; Rage et al., Reference Rage, Prasad and Bajpai2004, Reference Rage, Prasad, Verma, Khosla, Parmar, Prasad and Patnaik2020; Wilson Mantilla et al., Reference Wilson Mantilla, Mohabey, Peters and Head2010; Mohabey et al., Reference Mohabey, Head and Wilson Mantilla2011). On the contrary, the Eocene fauna is marked by the appearance of palaeophiids, constrictors, and caenophidians, aside from nigerophiids and madtsoiids that survived from the Cretaceous (Rage et al., Reference Rage, Bajpai, Thewissen and Tiwari2003, Reference Rage, Folie, Rana, Singh, Rose and Smith2008; Smith et al., Reference Smith, Kumar, Rana, Folie, Solé, Noiret, Steeman, Sahni and Rose2016; Datta and Bajpai, Reference Datta and Bajpai2024). The paleoecological contexts of these snakes range from aquatic (swamp to marginal marine) to amphibious and terrestrial. The observed faunal turnover was most likely related to the end-Cretaceous extinction event, which extinguished countless terrestrial and marine vertebrates. The niches vacated by the K-Pg extinction allowed snakes to not only occupy these ecological spaces but also to venture into new morphospaces.
Acknowledgments
The authors acknowledge with thanks the helpful comments, suggestions, and a constructive critique of the manuscript by J. Calede (Editor), H.-D. Sues (Associate Editor), J. Kastigar (Managing Editor), G. Georgalis, and an anonymus reviewer. We thank R. Sharma, D. Das, V.V. Kapur, N. Saravanan, L. Cooper, L. Stevens, and H. Thewissen for help during fieldwork and A. Saha and A. Rautela for help and discussions. The Science and Engineering Research Board (now ANRF) (Grant no. PDF/2021/000468 for National Post-doctoral Fellowship to DD), IIT Roorkee for Institute Post Doctoral Fellowship to DD, and the Department of Science and Technology (SB), Government of India are acknowledged for financial support. DD would like to acknowledge IIT Roorkee for providing infrastructural facilities. SB would like to acknowledge support obtained from IIT Roorkee as part of his Institute Chair Professorship.
Competing interests
Authors declare no competing interests.
Data availability statement
All appendices are available from Dryad Digital repository: https://doi.org/10.5061/dryad.z8w9ghxq2. Supplementary data are available from Morphobank (http://morphobank.org/permalink/?P5887).
Appendix 1. Holotype and referred specimens examined. All specimens are partial or nearly complete, if not mentioned otherwise
Appendix 2. Measurement of vertebral specimens (IITR/VPL/SB). All measurements in mm unless otherwise indicated. b. = measurement of broken regions; β = angle (when added to anatomical abbreviations); H = height; L = length; W = width. CL = centrum length; CNH = condyle height; CNW = condyle width; COH, cotyle height; COW, cotyle width; NAW, neural arch width; NCH, neural canal; NCW, neural canal width; NSH, neural spine height; POFL = postzygapophyseal facet length; POFW = postzygapophyseal facet width; POW = postzygapophyseal width; POα = postzygapohyseal angle; PRFL = prezygapophyseal facet length; PRFW = prezygapophyseal facet width; PRW = prezygapophyseal width; PRα = prezygapophyseal angle; TVH = total vertebral height; ZSFL = zygosphenal facet length; ZSFW = zygosphenal facet width; ZSH = zygosphene height; ZSW = zygosphene width; ZSβ = zygosphene angle
Appendix 3. Characters used in the phylogenetic analysis. A brief explanation follows the new characters used, whereas for a character already used in previous literature, its citation follows the last character state. S = Snetkov (Reference Snetkov2011). Numerals indicate the character number of each citation
-
1. Shape of the zygosphene: dorsal surface devoid of dorsoventrally high central crest (0); dorsal surface bearing dorsoventrally high central crest (1) [modified after S1]. The character was modified to adequately cover the difference/variation in shape of the zygosphene between Palaeophis and Pterosphenus-grade snakes.
-
2. Thickness of the zygosphene: thin (width more than twice as great as the thickness) (0); thick (width is less than twice as great as the thickness) (1) [S2].
-
3. Presence of the anterior hypapophysis on the anterior trunk vertebrae: vertebrae in the anterior part of the vertebral column lack anterior hypapophyses, which are much more developed than hypapophysis on other trunk vertebrae (0); vertebrae in the anterior part of the vertebral column have anterior hypapophyses, which are much more developed than that of other trunk vertebrae (1) [S3].
-
4. Presence of the anterior hypapophysis on the vertebrae of the middle and posterior trunk region: absent (0); present (1) [S4].
-
5. Relative height of the neural canal in the middle trunk vertebrae in anterior view: height ratio of the neural canal and cotyle > 0.4 (0); ratio < 0.4 (1) [S5].
-
6. Presence of the interzygapophyseal ridge: absent or very weak (0); well developed (1) [S6].
-
7. Shape of the centrum: condyle positioned much lower than cotyle (0); condyle positioned at the same level as cotyle (1) [S8].
-
8. Size of the prezygapophyses: large (ratio of prezygapophysis length to height of the cotyle > 0.7) (0); small (ratio < 0.7) (1) [S10].
-
9. Position of the prezygapophyses: medial part of the articular surface of the prezygapophyses positioned above the dorsal surface of the centrum (0); medial part of the articular surface of the prezygapophyses positioned at the level of the dorsal surface of the centrum (1) [S11].
-
10. Shape of the prezygapophyses: articular surfaces strongly inclined (0); articular surfaces positioned almost horizontally (1) [S12].
-
11. Position of synapophyses on middle trunk vertebrae: dorsal to ventral cotylar rim (0); extending slightly for a short distance below ventral cotylar rim (1); extending strongly below ventral cotylar rim (2) [modified after S13]. Modifications were introduced to adequately cover the variation in morphology seen in the examined taxa.
-
12. Extent of lateral flattening of the middle trunk vertebrae: slightly flattened prW/cL ≥ 1 (0); strongly flattened prW/cL < 1 (1) [modified after S14]. Modifications were made to evaluate the lateral flattening of the vertebrae in a quantitative manner.
-
13. Length of the centrum in the vertebrae that do not belong to the posterior trunk region in ventral view: less than the distance between the lateral margins of the synapophyses (0); more than the distance between the lateral margins of the synapophyses (1) [S15].
-
14. Position of the neural spine: anterior margin not extending to the dorsal margin of the zygosphene (0); anterior margin extending to the dorsal margin of the zygosphene (1) [modified after S16]. The character was modified to adequately cover the variation in neural spine morphology seen in palaeophiine snakes.
-
15. Presence of subcentral foramina: absent (0); present (1) [S17].
-
16. Development of the haemal keel on the middle trunk vertebrae: well developed, gradually passing into the posterior hypapophysis, which projects slightly more strongly than the haemal keel (0); poorly developed, projecting to a considerably lesser extent than the posterior hypapophysis (1) [S18].
-
17. Shape of the condyle: inclined strongly anteriorly (0); almost lacking anterior inclination (1) [S19].
-
18. Shape of the cross section of the centrum: ovate (0); almost triangular (1) [S20].
-
19. Width of the neural canal in the middle part of the vertebra: not wider than the centrum (0); much wider than the centrum (1) [S21].
-
20. Presence of subcentral crests on the anterior and middle trunk vertebrae: present (0); absent (1) [S22].
-
21. Width of the zygosphene: wide (width more than half the distance between the lateral ends of the prezygapophyses) (0); narrow (width less than half of the distance between the lateral ends of the prezygapophyses) (1) [S23].
-
22. Height of the vertebra: distance between the lateral margins of prezygapophyses more than the vertebral height from the ventral side of the cotyle to the dorsal margin of the zygosphene (0); distance between the lateral margins of prezygapophyses less than the vertebral height from the ventral side of the cotyle to the dorsal margin of the zygosphene (1) [S24].
-
23. Height of the neural arch: ratio of the distance from the dorsal side of the centrum to the dorsal margin of the zygosphene to the distance between the lateral margins of prezygapophyses < 0.4 (0); ratio > 0.4 (1) [S25].
-
24. Shape of the neural canal in anterior view: lateral margins strongly inclined (0); lateral margins almost vertical (1) [S26].
-
25. Angle formed by the margins of the V-shaped incisure in the posterior part of the neural arch in the vertebrae that do not belong to the posterior trunk region: > 105° (0); < 105° (1) [S27].
-
26. Shape of the posterior part of the neural arch in posterior view: distance from the dorsal surface of the centrum to the dorsal margin of the zygantrum greater than the distance from the dorsal margin of the zygantrum to the bases of pterapophyses (0); distance from the dorsal surface of the centrum to the dorsal margin of the zygantrum less than the distance from the dorsal margin of the zygantrum to the bases of pterapophyses (1) [S28].
-
27. Prezygapophyseal buttress: buttress lacking convexity or weakly convex (0); buttress strongly convex (1). In basal forms, e.g., Dinillysia Smith-Woodward, Reference Smith-Woodward1901, the prezygapophyseal buttress lacks convexity (coded as 0). In most palaeophiines the buttress bears a convexity (coded as 1).
-
28. Pterapophysis in precloacal vertebrae: absent (0); present but remaining below or at par with the dorsal zygosphenal margin (1); present and extending slightly above the dorsal zygosphenal margin (2); present and extending well above the dorsal zygosphenal margin (3). In basal forms such as Dinilysia, a pterapophysis is absent (coded as 0). Within Palaeophiinae, in Palaeophis snakes, the pterapophysis extends slightly above the zygosphene (coded as 2). In most Pterosphenus snakes, the pterapophysis extends well above the zygophene (coded as 3). The pterapophysis remains below the zygosphene in Pterosphenus rannensis n. sp. (coded as 1).
-
29. Height of pterapophysis (PTH) relative to centrum length (CL): PTH/CL ≤ 0.4 (0); 0.4 < PTH/CL ≤ 0.7 (1); PTH/CL > 0.7 (2). In ‘primitive’-grade palaeophiids (e.g., Palaeophis maghrebianus and Palaeophis africanus), PTH/CL is ≤ 0.4 (coded as 0), whereas in ‘advanced’-grade palaeophiids (e.g., Palaeophis toliapicus and Palaeophis typhaeus), PTH/CL ranges between 0.4 and 0.7 (coded as 1). In Pterosphenus taxa (e.g., Pterosphenus schucherti and Pterosphenus schweinfurthi), PTH/CL is > 0.7 (coded as 2).
-
30. Pterapophysis-postzygapophysis connection in lateral view: straight/weakly concave (0); markedly concave (1). In Palaeophis snakes (e.g., Palaeophis maghrebianus and Palaeophis typhaeus), the pterapophysis-postzygapophysis connection is weakly concave or devoid of it (coded as 0), whereas in most Pterosphenus taxa, the pterapophysis-postzygapophysis connection is markedly concave (coded as 1).
