The Ribesalbes-Alcora Basin is an intramontane basin located in the eastern Iberian Peninsula (Fig. 1A), covering an area of 150 km² and is bounded by ENE-WSW to NNE-SSW normal faults (Anadón et al. 1989). The sedimentary infill of the Basin can be divided into two main sequences (Agustí et al. 1988). The lower one, named the Ribesalbes sequence, is composed of alluvial and lacustrine deposits. It has a thickness of up to 600 m, which originated in the Early to Middle Miocene. The Alcora sequence is constituted by alluvial deposits up to 200 m thick. Most probably this was formed during the Middle and Late? Miocene (Anadón et al. 1989). Anadón (1983) defined five depositional units (A to E) in the Ribesalbes sequences, in which only Units B and C have yielded palaeontological localities. Thus, Unit B is characterised by oil-shales and laminated bituminous dolomicrite where the Konservat-Lagertätten sites of La Rinconada (Ribesalbes) and San Chils (L’Alcora) are located (Peñalver et al. 2016; Álvarez-Parra and Peñalver 2019). The Araia/Mas de Antolino outcrop (Araia d’Alcora, Castelló, Spain), which includes the studied sites, belongs to Unit C in the Ribesalbes sequence (Agustí et al. 1988; Crespo et al. 2019c). This side of the ravine is formed by grey and yellow mudstones, limestones and sandstones, ∼ 100 m thick. It comprises seven sections, from oldest to youngest (Fig. 1B): Mas dels Coixos (MCX), Mas de Torner (MTR), Araia Cantera Sud (ACS), Barranc de Campisano (BC), Foieta la Sarra (FS), Mas d’Antolino B (MAB) and Corral de Brisca (CBR) (Crespo et al. 2019c). Recently, Álvarez-Parra et al. (2021) described a new Konservat-Lagerstätte site in the section of FS, Foieta la Sarra A (FSA).

Crespo et al. (2019c) studied 45 new sites with mammal remains, which had been correlated with the local biozone C (MN 4, Early Miocene) from the Calatayud-Montalbán Basin (Spain), the type area of the Aragonian stage, with an estimated age between 16.49 and 15.98 Ma as established by Van der Meulen et al. (2012). An alternative interpretation has been proposed by Casanovas-Vilar et al. (2022), who discuss that the extinction boundary of Ligerimys florancei Stehlin & Schaub, 1951 falls within the MN4–MN5 transition. Under this framework, the sites of the Ribesalbes-Alcora Basin, associated with the local biozone L. florancei, would correspond to MN 4. In contrast, those linked to the local biozone Ligerimys ellipticus Daams, 1976, would be assigned to MN 5.

ACS, Araia Cantera Sud; BC, Barranc de Campisano; CBR, Corral de Brisca; FS, Foieta la Sarra; HEG, Herpetological Ecohysiological Groups method; LAD, last appearance datum; MAB, Mas d’Antolino B; MAP, mean annual precipitation; MCX, Mas dels Coixos; MGUV, Museu de la Universitat de València d’Història Natural; MN, Mammal Neogene Stages; MNCN-CSIC, Museo Nacional de Ciencias Naturales-CSIC (Madrid, Spain); MSC, Messinian Salinity Crisis; MTR, Mas de Torner; PLA, preacetabular lateral angle; UV, Universitat de València (Valencia, Spain).

Caudata Scopoli, 1777

Urodela Duméril, 1806

Salamandridae Goldfuss, 1820

Lissotriton Bell, 1839

cf. Lissotriton sp.

Fig. 2A–I

Localities. MCX7, MTR1, MTR2, ACS2, BC1, BC2, MAB0A, MAB2, MAB4, MAB8, CBR0B and CBR0D.

Material. One right dentary (MGUV-40685), one left dentary (MGUV-40445); two dentaries (MGUV-40347 and 40358), 43 tooth-bearing bones (MGUV-40342, MGUV-40350, 40384, 40394, 40606, 40609, 40655 and 40686), two trunk vertebrae (MGUV-40349), one caudal vertebra (MGUV-40688), 48 vertebral centra (MGUV-40343, 40348, 40389, 40410, 40440, 40497, 40573, 40605, 40634, 40684, 40645 and 40652), one ilium (MGUV-40352) and five phalanges (MGUV-40344, 40345, 40687 and 40646).

Description. The preserved dentaries are elongated and curved bones in medial view. Although no teeth are present, they were densely packed, pleurodont and positioned on the labial side (Fig. 2A). The crista dentalis is wide. The Meckelian groove is open, but not in the anteriormost portion. A wide incisura dentalis is present and it seems to be located in the posterior half (Fig. 2B). The symphysis is wide and kidney-shaped. Although it is broken, a slight medial expansion in the posterior direction develops, forming the medial surface that acts as a contact area with the other dentary. No sulcus dentalis is present in dorsal view. The dentary is slightly curved in a medial direction. In lingual view, two small lingual foramina are present.

The two most complete vertebrae are opisthocoelous and small-sized. In dorsal view, the prezygapophyses are rounded in shape and project anteriorly. Their medial edges are parallel (Fig. 2C). In lateral view, the neural arch rises upwards caudally. The zygapophyseal and ventral crests are well developed. Thus, the posterior zygapophyseal crest contacts the dorsal border of the diapophysis. On its behalf, the anterior ventral crest contacts the ventral transverse process (Figs 2D, 2E). The condyle has a flat anterior surface and it is slightly anteroventrally inclined. In ventral view, the centrum is elongated. The centrum shows a relatively thick and short condyle clearly separated from the vertebral centrum by a well-developed pericondylar constriction. The subcentral foramina are large and some vertebral centra show supplementary, more or less smaller ones (Fig. 2F). The ventral lamina is more or less trapezoidal (Fig. 2G). In anterior view, the neural arch is laterally compressed and slightly vaulted. The condyle has a circular or elliptical shape, with the major axis orientated horizontally. The foramen at the basis of the parapohyses is small.

MGUV-40688 is an amphicoelous vertebra. Although it is broken, the presence of the basis of a pair of haemapophyses denotes that it is a caudal vertebra. The haemapophysis starts posteriorly in the centrum, which indicates that this bone may be the first or second caudal vertebra (Fig. 2H). In dorsal view, MGUV-50688 has a not pronounced interzygapophyseal constriction. The neural arch is short and rectangular. The transverse process projects posterolaterally. The prezygapophyseal process is subovoid and projects laterally. In lateral view, the neural crest, although it is broken, is higher in its posterior region than in the anterior one. The anterior zygapophyseal crest is horizontal, prominent and contacts the prezygapophysis with the dorsal edge of the diapophysis. The anterior ventral crest contacts the ventral edge of the parapophysis. The incisura ventralis caudalis is well marked. The posterior ventral crest is prominent. In ventral view, the centrum is triangular. No pericondylar constriction is present. The ventral lamina is straight. The haemapophysis is bifurcated and its base runs along the entire length of the centrum. The latter presents a pair of large subcentral foramina. In anterior view, the neural arch is dorsoventrally compressed. Its roof is vaulted and overpasses the edge of the prezygapophysis. The latter, which projects laterally, forms an angle of approximately 30° with respect to the horizontal axis. The condyle is rounded in this view. In the posterior view, the neural arch is laterally compressed and vaulted. The cotyle is broken.

The ilial shaft is rod-like. Its surface is smooth. The acetabular region is large and well-defined. It has a well-developed ventrolateral tuberosity, which projects laterodorsally. No tuber superior is present. The pars ascendens and pars descendens are short. The pars ascendens runs straight at the same height as the pars cilindriformis. The acetabulum is large, flat, and wide. The pre-acetabular fossa is large and well-defined. The pubic process is straight and projects ventrally. The articulation area with both the pubic and the ischium is damaged. The Pre-acetabular Lateral Angle (PLA) is approximately 135° (Fig. 2I).

Remarks. The described traits of the vertebra are consistent with a taxonomical attribution to the genus Lissotriton: a small opisthocoelous vertebra with a well-defined pericondylar constriction, well-developed crests and the trapezoidal shape of the ventral lamina (Macaluso et al. 2023). The latter trait is coincident with the extant species Lissotriton boscai (Lataste, 1879) (see Sanchiz (1998a); Macaluso et al. (2023)). Despite all the above, the poor state of preservation of the remains studied does not allow for a precise attribution.

Palaeoecology. Smooth newts inhabit Europe, from the Iberian Peninsula to the Caucasus and some areas of the Middle East (Dufresnes 2019). Lissotriton currently comprises 10 species that inhabit lentic environments without strict habitat preferences.

Nowadays, two species inhabit the Iberian Peninsula (Speybroeck et al. 2016): the palmate newt, Lissotriton helveticus (Razoumovsky, 1789), in the northern realm of Spain and the northern half of Portugal and Bosca’s newt, Lissotriton boscai (Lataste, 1879), in western Iberia. However, a possible third species, Maltzan’s newt Lissotriton maltzani (Boettger, 1879), inhabits south-western Portugal (Dufresnes 2019). All species are highly adaptable, using a wide range of different waterbodies (Speybroeck et al. 2016).

Bufonidae Gray, 1825

Bufonidae indet.

Fig. 2J, K

Locality. BC1.

Material. One centrum of a sacral vertebra (MGUV-40353).

Description. In ventral view, the procoelous centrum is wider than long and trapezoidal in shape. Posteriorly, it has two condyles, which articulate with the urostyle (Fig. 2J). In anterior and posterior views, the cotyle and the condyles are dorso-ventrally compressed. Although they are eroded, the condyles seem to contact in their medial region. In the posterior view, the walls of the neural arch are short and wide (Fig. 2K).

Remarks. Amongst the European anuran families, only Pelodytidae, Hylidae and Bufonidae have bicondylar procoelous vertebrae. However, Pelodytidae has elongated centra and the condyles are small and clearly separated (Sanchiz 1978; Bailon 1991). Regarding Hylidae, the centrum is more gracile and longer than the studied remains.

Palaeoecology. The taxonomical attribution of that taxon precludes any palaeocological interpretation.

Anura indet.

Fig. 2L, M

Localities. MAB2.

Material. One vertebral centrum (MGUV-40574).

Description. MGUV-40574 is an isolated procoelous vertebral centrum. In ventral view, the centrum is slightly longer than wide. A large foramen is present. The lateral walls of the neural arch are wide (Fig. 2L). In posterior view, the condyle is dorsoventrally compressed (Fig. 2M).

Remarks. The poor conservation state of the studied material precludes any attempt to propose a taxonomical attribution at the order level.

Palaeoecology. The taxonomical attribution of that taxon precludes any palaeocological interpretation.

Lissamphibia indet.

Fig. 2N

Localities. MCX3, BC1, MAB2 and MAB3.

Material. One tooth-bearing bone (premaxilla?) (MGUV-40674) and seven phalanges (MGUV-40351, 40575, and 40576).

Description. The scarce taxonomical value and/or the poor state of conservation of the included material preclude any tentative taxonomical attribution.

Crocodylia Owen, 1842

Crocodylia indet.

Localities. MCX3, MTR2, BC1, BC2, MAB0A, MAB0B, MAB-3, MAB3A, FS1, MAB9, MAB10, MAB11, MAB11A, MAB11B, CBR0A, CBR0B, CBR0D, MAB13, CBR1, CBR3 and CBR4. Fig. 3A–E.

Material. 298 teeth (MGUV-25293, 25294, 25295, 25340, 25341, 25342, 26256, 26257, 26258, 36658, 36669, 36738, 36739, 36740, 36807, 36811, 36905, 36908, 38379, 38404, 38540, 38557, 38558, 38568, 38788, 38789, 38792, 40390, 40457, 40477, 40528 and 40542) and 17 osteoderms (MGUV-40381, 40432, 40478, 40504, 40541, 40635 and 40670).

Description. Several teeth have been recovered from the different studied sites. All of them only preserve the crown, which is subconical and variably pointed apically (Figs 3A, 3B). Lingually, the teeth are slightly concave, whereas they show a convex surface in labial view (Figs 3A, 3B). In both views, their surfaces are generally smooth, but some specimens exhibit wrinkles and slight vertical striations. The mesiodistal carina is unserrated (Fig. 3C). Amongst the teeth, two morphotypes can be identified: the first consists of slender, long and more pointed teeth (Figs 3A, 3B), whereas the second features more massive teeth with a bulbous aspect and apically blunt (Fig. 3D).

The recovered osteoderms are fragmentary. On its external surface, large, rounded pits ornament these remains (Fig. 3E). MGUV-40543 is a small osteoderm with poorly-developed ornamentation, which could be interpreted as belonging to a juvenile individual (Pereda Suberbiola et al. 2001).

Remarks. As Brochu (2000) notes, isolated teeth and osteoderms have little taxonomic value. Thus, the recovered remains can only be assigned to an indeterminate crocodylian. However, some observations can be made. The two tooth morphotypes correspond to different positions along the tooth row. The slender, elongated and pointed teeth likely belonged to the anterior region, whereas the second morphotype—characterised by massive, blunt and bulbous teeth—was located in the posterior region of the tooth-bearing bones (Martin et al. 2014). The presence of bulbous posterior teeth is a trait observed in Diplocynodon Pomel, 1847, but is absent in the gavialid Gavialosuchus Toula & Kail, 1885 (e.g. Górka et al. (2025)). Unfortunately, the recovered crocodilian osteoderms are too fragmentary to determine whether a longitudinal keel is present, a typical feature of Diplocynodon, but generally absent in gavialids (Martin et al. (2014); Walter et al. (2026)). In summary, the poor preservation of the material precludes a precise taxonomic identification of the crocodylian remains from the Ribesalbes–Alcora Basin.

Palaeoecology. Traditionally, crocodylians have been considered indicators of warm climatic conditions. In this context, extant crocodilian genera are primarily distributed in low latitudes, generally within 30° of the Equator (Marwick 1998). However, during the Eocene–Oligocene boundary, Diplocynodon appeared at higher—and therefore cooler—latitudes, coinciding with the development of a continental ice sheet in Europe. Its presence may serve as indirect evidence of a possible ability to withstand low temperatures, perhaps through thermoregulatory strategies (Martin 2010). For example, Alligator sinensis (Fauvel, 1879) ceases feeding and enters hibernation during the winter when temperatures drop, a process involving complex molecular mechanisms (Lin et al. 2020).

Testudinidae Linnaeus, 1758

cf. Testudinidae indet.

Fig. 3F, G

Localities. MAB11.

Material. Two fragments of plates (MGUV-40567 and 40568).

Description. The remains are relatively large, thick fragment of an undetermined plates (Fig. 3F). Undoubtedly, these remains belong to a large/medium turtle. The surface of the remains presents a delicate ornamentation consisting of small circular hollows and vermicular pits. These plate fragments are relatively thick in lateral view (Fig. 3G).

Remarks. The relative thickness of the plates is one of the criteria to differentiate Testudinidae from the rest of the European chelonian families during the Neogene (Hervet 2000). Probably this remains belonged to an undetermined Testudinidae. However, this attribution is highly tentative.

Palaeoecology. The taxonomical attribution of that taxon precludes any palaeocological interpretation.

Testudinoidea indet. (Geomydidae or Emydidae)

Fig. 3H–N

Localities. MAB5 and CBR0B.

Material. One peripheral (MGUV-40620 and 40643) and three fragments of plate (MGUV-40414, and 40644).

Description. MGUV- 40414 is a trapezoidal fragment of a relatively thick plate. Both the dorsal and ventral surfaces exhibit a well-marked straight sulcus (Figs 3H, 3I). Its anterior border is convex, forming a small curvature of approximately 120°. The ornamentation of the plates is composed of small circular hollows and vermicular pits (Fig. 3H). In lateral view, the external surface is narrower than the innermost region (Fig. 3N).

MGUV-40643 is a peripheral plate, probably of the right side of the carapace. This plate is rectangular, clearly longer that wide. The sulcus between the marginal scutes is present, both in dorsal and ventral views (Figs 3L, 3M). The ornamentation of the plate is similar to the one observed in MGUV- 40414. This plate is very flat (Fig. 3N).

Remarks. The presence of a pleuromarginal sulcus that does not coincide with the costal–peripheral suture allows us to rule out some chelonian taxa present in Europe during the Miocene. This feature is absent in both Testudinidae and Chelydridae, in which the sulcus typically overlaps the suture between the costal and peripheral plates (Hervet (2000); Joyce (2016)). Moreover, the presence of a sulcus between the scutes suggests that these remains did not belong to Trionychidae, as these marks are absent in this group (Georgalis and Joyce 2017). Taking into account the fossil record of the group during the Miocene in the Iberian Peninsula (Pérez-García 2017) and after discarding the aforementioned clades, only two taxa could potentially be attributed to the described remains: Geomydidae and Emydidae. The distinction between the two groups is challenging, based on such fragmentary shell elements (de Lapparent de Broin 2001). Traditionally, the location of the pleuro-marginal sulcus with respect to the costo-peripheral suture has been used as a criterion for the discrimination between both groups, at least in Quaternary deposits (i.e. Bailon (2010)). Thus, under that premise, MAB5-941 could belong to the family Emydidae; however, the studied remains are in very poor condition, precluding any close taxonomic attribution. Moreover, emydids appear in the European fossil record only during the Late Miocene of Eastern Europe (Fritz 1995). Despite this, Böhme and Ilg (2003) include undetermined emydids from some Spanish Early Miocene localities. If future work could confirm the presence of that taxon in that chronology, it would imply a remarkable change in the palaeobiogeography of the group, which is supposed to having reached Eurasia during the Late Miocene (Fritz 1995). Most probably, the attribution of these remains may be related with an older conception of Emydidae, in which Geomydidae were included inside this taxon (Hirayama 1985).

Palaeoecology. These remains could belong to an undetermined terrapin species. These kinds of turtles inhabit different types of waterbodies.

Gekkota Cuvier, 1817

Gekkota indet.

Fig. 4A–D

Localities. BC1, MAB3, MAB5, and FS1.

Material. One probable maxilla (MGUV-40386), four left dentaries (MGUV-40479, 40480, 40597 and 40624).

Description. A probable maxilla has been recovered from BC1. This fossil is very fragmentary. In a lingual view, MGUV-40386 has eight packed pleurodontal tooth positions, but only one tooth is preserved. It is slender and conical, lacking its apex. Its base presents a small, rounded resorption pit. The palatal shelf is straight and highly damaged (Fig. 4A). In labial view, the surface of the fossil has one large labial foramen and another small one in a more dorsal position relative to the former (Fig. 4B).

The fragmentary dentaries exhibit a slightly convex ventral edge. In lingual view, the bone has a tall alveolar shelf that supports a large number, probably ten, of closely-packed pleurodont dental positions. The Meckelian canal is closed (Fig. 4C). In a labial view, the dentary has two large, elliptical labial foramina (Fig. 4D).

Remarks. The morphology of the teeth and the dentary, with a convex ventral border and a closed Meckelian canal, are characteristics of Gekkota (Bailon 1991). This clade is represented in the European Miocene by three genera: a gekkonoid, Palaeogekko Schleich, 1987 and two sphaerodactylids, Gerandogekko Hoffstetter, 1946 and the extant Euleptes Fitzinger, 1843 (Daza et al. 2014; Villa et al. 2022; Villa 2023). Unfortunately, the studied material is too fragmentary for a genus-level assignment. Due to its preservation, the maxillary fragment can be referred to this squamate clade only tentatively.

Palaeoecology. The indeterminate taxonomic attribution of that taxon precludes any palaeocological interpretation. However, gekkotans are considered indicators of warm/temperate conditions in palaeontological sites.

Pyrenasaurus Bolet & Augé, 2014

cf. Pyrenasaurus sp.

Fig. 5A–D

Locality. CBR1.

Material. One left dentary (MGUV-40463).

Description. The material studied is fragmentary and only involves the posterior region of a left dentary, in which the processes are absent (Fig. 5A–E). This remnant shows four dental positions, with only two of them being present. The teeth are triangular in shape, taller than wide and labio-lingually compressed in occlusal view. The only two preserved teeth have a robust appearance and a marked triangular shape in labial view (Fig. 5A). These are pleurodont, with the alveolar crest exceeding half the height of the tooth. The interdental gap is narrow. The crown is tricuspid, with a medial cusp more developed than the lateral ones. The crown presents a deflection in its convexity. Lingually, five or six faint striae run parallel to the tooth axis (Fig. 5B). A hypothetical rounded resorption pit is located in the central region at the base of the tooth (Fig. 5A). Labially, the central part of the tooth is bulging and lingually to it, it is somewhat flattened. The lamina horizontalis is narrow and has a uniform thickness along its entire preserved length. A well-developed intermandibular septum is present, lying ventromedially from the oval-shaped alveolar canal and located in a posterior position at the level of the last preserved teeth of the dental shelf. The lamina horizontalis finishes posteriorly close to the base of the most posterior dental position. The ventral region of the dentary is broken; thus, it is not possible to discern if the Meckelian groove was open or closed. No foramen is present on the labial surface of the dentary (Fig. 5D).

Remarks. MGUV-40463 shows a particular morphology (Fig. 5A–D). Amongst the different squamate taxa described in the Palaeogene and Neogene in Eurasia, the general tooth morphology of the remain is similar to Pyrenasaurus evansae Bolet & Augé, 2014, from the sites of Escamps B (MP 19, late Eocene, Phosphorites du Quercy, France) and Sossís (MP17a, late Eocene, Central Southern Pyrenean Basin, Spain) (Fig. 5F, G). Thus, in this taxon, the posterior teeth are large, labiolingually compressed, with a characteristic deflection in the convexity of the crown, so resembling a short obelisk (less developed), with its faces converging close to the tip and striation is only present on the lingual side of the crown (Bolet and Augé 2014). These authors have tentatively assigned P. evansae to Scincomorpha (a paraphyletic clade, for example, Simões and Pyron (2021)). Another taxon that shares some characteristics with MGUV-40463 is “Scincomorpha indéterminé Forme B” (USTL, MAL 605) from the French site of Malpérié (MP 17; middle Eocene) (Augé 2005) (Fig. 5H–I). In contrast to P. evansae, MGUV-40463 and USTL, MAL 605 have a more pronounced tricuspid condition (Augé 2005; Bolet and Augé 2014). As Bolet and Augé (2014) pointed out, it is possible that the differences between P. evansae and USTL, MAL 605 and, by extension, MGUV-40463, may represent different, but related genera.

Palaeoecology. Pyrenasaurus evansae has been interpreted as a burrowing or surface-dwelling lizard, based on the short dentary, low number of teeth, high coronoid process and dentary miniaturisation (Bolet and Augé 2014).

Chalcides Laurenti, 1768

Chalcides sp.

Fig. 6

Localities. MCX3, BC1 and MAB3.

Material. One left maxilla (MGUV-40577), one left dentary (MGUV-40679), three right dentaries (MGUV-40578 and 40579), two dentaries (MGUV-40360) and three tooth-bearing bones (MGUV-40580).

Description. The preserved maxilla (MGUV-40577) only presents its anteriormost region (Fig. 6A). In dorsal view, it presents a deep U-shaped concavity. Neither the anterolateral process nor the anteromedial process are preserved. In the lingual view, a well-defined foramen is present. In labial view, only three teeth are present. These are pleurodont, closely spaced, cylindrical and bicuspid (one cusp labially and a second lingually, smaller than the former). No expansion of the crown of the teeth is observed (Fig. 6A). In lingual view, the tooth crown bears striations (Fig. 6B).

The most complete dentary (MGUV-40579) shows 16 tooth positions (Fig. 6C, D). The morphology of the teeth is similar to that described in the maxilla, but their striation is more marked than in the maxilla. Although the teeth appear to be monocuspid, in lingual view, underneath the labial cusp, a smaller lingual cusp may be present. The antrum intercristatum is a well-marked groove located on one side and on the other, between the lingual and labial cusps. They are prominent and pointed and are connected by a short crista intercuspidalis in occlusal view. The crista distalis and mesialis have the same length and are bow-shaped, being narrower in the anterior teeth. The culmines lateres are prominent. Seven-eight pronounced, straight and parallel striae run in the flattened lingual surface of the teeth (Fig. 6E). The Meckelian canal opens along the entire length of the dentaries. The location and width of this structure vary; in the posterior region (MGUV-40579), it is laterally expanded (Fig. 6C), whereas in a more anterior position, it is narrow and located ventrally (MGUV-40578) (Fig. 6F). The symphysis is large, oval, almost horizontal and located in a ventral position and the ventral end of the maxilla is straight. In labial view, the most complete dentary (MGUV-40578) has five dental foramina (Fig. 6D). The studied remains do not show any impression of the coronoid. In the occlusal view, a sulcus dentalis is present. It is narrow and not well developed (Fig. 6G).

Remarks. The overall morphology of the dentary (ventral end straight, the Meckelian canal is fully open, narrowing anteriorly and its anterior half opens in ventral direction, both the sulcus dentalis and the subdental ridge are patent, absence of an impression of the coronoid and the lack of bi- or tricuspid teeth) are characters shared by Scincidae and Cordylidae (Augé and Smith 2009), which are related clades. Both were present in Europe during the Miocene (Villa and Delfino 2019a). Additionally, as Gao and Fox (1996) pointed out, both groups are often difficult to distinguish. However, in cordylids, the symphysis is more horizontal than in scincids (Čerňanský 2012; Yadav et al. 2023) and, although the teeth are very similar, cordylids often have their base slightly expanded related to their crown (Roček 1984; Georgalis et al. 2019a). The presence of an open Meckelian canal is only present in some scincid genera. In the case of current European taxa, this condition is present in Chalcides, Ophiomorus Duméril & Bibron, 1839 and Eumeces Wiegmann, 1834. Thus, Heremites Gray, 1845 and Ablepharus Fitzinger in Eversmann, 1823 are the unique scincids of the continent with a closed Meckelian canal (Bailon (1991); Camaiti et al. (2019); Čerňanský and Syromyatnikova (2021)). The studied remains have gracile teeth more similar to Chalcides than to Ophiomorus and especially to Eumeces, in which these teeth are inflated with rounded tips (Čerňanský and Syromyatnikova 2021). However, the incompleteness of the studied remains precludes the specific attribution of the taxon.

Palaeoecology. Chalcides is a genus of scincid lizards found in Macaronesia, the Mediterranean Basin, western and eastern Africa and the Middle East through to Pakistan. In general, the species included in that genus show a generalist ecology and perhaps occur in a broad range of habitats (Caputo et al. 1995).

Lacertidae Oppel, 1811

Lacertidae indet.

Fig. 7

Localities. MCX1, MCX2, MCX3, MTR1, MTR2, BC1, MAB0A, MAB0B, MAB2, MAB3, MAB3A, MAB4, MAB5, FS1, MAB11, MAB0A, CBR0B, CBR0G, CBR1 and CBR4.

Material. Ten premaxillae (MGUV-26252, 40599, 40615, 40625 and 40663); four left maxillae (MGUV-26264 and 40601), one right maxilla (MGUV-36906), six maxillae (MGUV-26263, 40435, and 40454), five left dentaries (MGUV-40492 and 40582), 29 frontals (MGUV-26265, 40355, 40356, 40357, 40408, 40411, 40462, 40472, 40489, 40503, 40532, 40545, 40603, 40647 and 40677), six right dentaries (MGUV-40581 and 40692), one left dentary (MGUV-40614), eight right dentaries (MGUV-23382, 26259 and 26262), eight dentaries (MGUV-26261 and 40433), 143 tooth-bearing bones (MGUV-25296, 25299, 25300, 26253, 26260, MGUV-36907, 40412, 40434, 40460, 40461, 40476, 40529, 40572, 40585, 40613, 40623, 40666, 40667, 40656 and 40657), one presacral vertebra (MGUV-40368), two caudal vertebrae (MGUV-40371), 21 vertebral centra (MGUV-40365, 40425, 40499, 40553, 40610, 40620 and 40668) and one left coxal (MGUV-40354).

Description. Frontals are anteroposteriorly elongated bones. Within the studied material, there are both unfused (MGUV-26265, 40355, 40356, 40408, 40462, 40472, 40489, 40503, 40532, 40545, 40603, 40635, 40647 and 40677; Fig. 7A–C) and fused frontals (MGUV-40357 and 40411; Fig. 7D, E). In the latter cases, a not well-marked suture line is present in the middle region of its ventral surface (Fig. 6E). Apart from that, all the elements have a very similar morphology; for this reason, they probably belong to different ontogenetic stages of the taxon. Frontals have a subtriangular shape with a relatively well-marked constriction. The dorsal surface is totally covered by fused osteodermal shields that show a vermicular ornamentation. The sulcus interfacialis, which separates frontal and frontoparietal scutes, is slightly convex anteriorly and is located approximately at the point of maximum constriction (Figs 7A, 7I). Unfortunately, neither the medial nor the lateral process nor the posterolateral is preserved. The posterior margin of these frontals is slightly interdigitated. In lateral view, the prefrontal facet is broad and wedge-shaped. Posteriorly, the facet for the postfrontal is thin and low-marked. Neither facet is in contact; thus, the frontal may have been exposed on the orbit when the animal was alive (Fig. 6C). In ventral view, the crista cranii is broad and dorsoventrally flattened (Figs 6B, 6E). At least three well-defined foramina pierce this structure. The subolfactory process is broken.

Premaxillae are small and preserve only their alveolar region. Thus, nasal processes are not preserved. However, a pair of foramina is present at the base of this broken process (Fig. 7F). In labial view, the premaxillae have a broad and short incisive process. Both 7–8 tooth positions are present, teeth are pleurodont, isodont and cylindrical (Figs 7G). The lingual surface of the premaxillae is smooth (Fig. 6H).

The recovered maxillae are in a very poor state of preservation, with none of the processes present. In lingual view, the palatal shelf is slightly convex. The teeth are pleurodont, isodont, cylindrical, mono- and bicuspid, and have a blunt apex (Fig. 6I). In labial view, the maxillae have 1–2 labial foramina (Fig. 6J).

In lingual view, the dentaries show an open Meckelian groove along their whole length. The most complete dentary (MGUV-26259) bears 14 tooth positions, with only four teeth in their life positions. Teeth are similar to those described for the maxilla. The ventral end of the dentary is convex (Fig. 6K). In labial view, the dentaries show seven labial foramina. The posterior end of the dentary is not preserved (Fig. 6L).

Remarks. Although the middle constriction is well-marked in these frontals, it is less pronounced than the genera included in the tribe Eremiadini. Moreover, this clade has a typically fused frontal lacking a longitudinal suture line in the ventral view (Arnold (1989); Villa and Delfino (2019b)). Some specimens of our sample (MGUV-40357 and 40411) show that line. This character has been interpreted as resulting from the fusion of the two halves during the late stages of ontogeny. As a result, few examples of current Lacertini genera show their frontals fused. Amongst Iberian genera, Barahona (1996) has reported that in large specimens of Psammodromus algirus (Linnaeus, 1758) and Podarcis hispanicus s.l. Steindachner, 1870, the frontal can be fused, even if the suture is not visible. Regarding extinct lacertid taxa present in Eurasia, frontals are known in a few cases and, in those, they are from large species. Some of them seem to have unfused frontals, such as Janosikia ulmensis (Gerhardt, 1903) (see Čerňanský et al. (2016)) and Plesiolacerta eratosthenesi Čerňanský & Augé, 2013 (see Čerňanský and Auge (2013)). On the other hand, taxa with fused frontals with ventral sutures are Dracaenosaurus croizeti Gervais, 1848 (see Čerňanský et al. (2017)); Dormaalisaurus sp. (see Augé and Smith (2002)); and Pseudeumeces cadurcensis Filhol, 1877 (see Augé and Hervet (2009); Georgalis et al. (2021a)). All these taxa not only differ from our material by their larger size, but also by some morphological traits. Thus, Dormaalisaurus sp. has more pronounced interdigitations than the studied frontals (Augé and Smith 2002). The prefrontal facet is clearly wider in Dr. croizeti than in our material (Čerňanský et al. 2017). Finally, the parietal tab is wider and more marked in Pseudeumeces cadurcensis than the studied frontals (Augé and Hervet 2009).

The rest of the material included within this taxon has characters attributable to Lacertidae (i.e. fused premaxilla with an undivided palatal process distinct and developed; dentaries and maxillae with pleurodont teeth, which are mono-, bi- or tricuspid; dentaries with the Meckelian groove totally open and the ventral border of the bone convex). However, due to the osteology of the lacertid still being poorly understood, it is not possible to propose a more confident taxonomic attribution.

In conclusion, although the studied frontal bones exhibit certain similarities with the extant genera Psammodromus and Podarcis, it is not possible to propose an attribution below the family level. This is due to the significant osteological homogeneity of lacertids and the fact that the osteological differences between their various taxa are not yet fully understood. Furthermore, the possibility that the studied material corresponds to more than one taxon cannot be ruled out, given the high diversity of lacertids in modern ecosystems and their status as a dominant element of the European herpetofauna since the spectacular radiation of the group during the Neogene (Arnold et al. 2007).

Palaeoecology. The indeterminate taxonomical attribution of that taxon precludes any palaeocological interpretation.

Blanidae Kearney, 2003

Blanidae indet.

Fig. 8

Localities. BC1, MAB5 and MAB11.

Material. One left maxilla (MGUV-40361) and three left dentaries (MGUV-40552 and 40616).

Description. MGUV-40361 is a left maxilla with a poor state of conservation. It only preserves three fragmentary teeth. The largest of them is the central one (Fig. 8A). In labial view, two large foramina are present (Fig. 8B).

MGUV-40616 is a robust, short left dentary with eight dental positions, probably 9 and 7 teeth present (Fig. 6C). The latter are pleurodont, conical, laterally compressed, unicuspid and closely spaced. Moreover, the teeth are slightly inclined anteriorly. The first one is markedly expanded, with a broad base, being the most robust tooth of the dentary. The distal edge of the bone is broken. Thus, the mandibular symphysis is lost. The Meckelian groove is fully open along the entire length of the dentary. Although its proximal end has some mineral concretions, the surface of the bone seems smooth, without any foramina present, in labial view. Under the repository number MGUV-40552, two left dentaries are included, which only preserve their proximal ends (Fig. 6D–G). The subdental shelf shows a constant depth and at the last tooth position, it is narrow and turns dorsally. The intramandibular septum is well preserved. Its ventral margin is fused with the medial wall of the dentary. Its posterior edge has a wide, subtriangular notch. At the base of the septum, a narrow and elongated incision is present. It might correspond to the impression of the splenial.

Remarks. The presence of a short, robust dentary indicates the presence of an amphisbaenian amongst the material recovered from MAB5. Despite the poor state of preservation of the fossils, some considerations about its taxonomy may be noted. Amongst the European fossil record of this order, the remains differ from the family Polyodontobaenidae in having, in contrast to the latter, the anterior teeth more developed than the posterior ones (Folie et al. 2013). Regarding Blanidae, the general morphology resembles the genus Palaeoblanus, in which the most anterior tooth is clearly more robust than the rest. Nevertheless, in MGUV-40616, this tooth is not the tallest; the tallest one is probably the second one, but it is eroded. Moreover, neither the number of teeth (seven in Palaeoblanus, eight or nine in MGUV-40616) nor the tiny interdental gap has been reported in that blanid genus (Schleich (1988); Böhme (1999)). The presence of enlarged anterior teeth has been reported in other Paleogene blanid genera, for example, Cuvieribaena carlgansi Čerňanský et al., 2015 from the locality of Le Bretou (Phosphorites du Quercy, late middle Eocene) (Čerňanský et al. 2015). Our material differs from the latter by a large number of teeth and, apparently, in MGUV-40616, the teeth do not reach a uniform height. Regarding Blanus, our material does not fully match the general morphology of this genus. The different species of the genus Blanus have seven or eight teeth, with the ninth position only recorded in very exceptional cases (Bolet et al. (2014); Georgalis et al. (2018); Villa et al. (2019b)). The enlargement of the third tooth is a typical trait of the genus (Bailon 1991), but, in our material, that tooth is not present. Thus, the genus-level attribution is unclear for this material.

Palaeoecology. Amphisbaenians are fossorial reptiles that currently inhabit tropical and subtropical regions (Augé 2012). However, the broad taxonomical attribution precludes any palaeocological interpretation.

Chamaeleonidae Rafinesque, 1815

?Chamaeleonidae indet.

Fig. 6R, S

Locality. CBR0B.

Material. One undetermined dental bone (MGUV-40413).

Description. MGUV-40413 is a very fragmentary, tooth-bearing bone. The only two preserved teeth have an acrodont implantation, being fused to the upper margin of the bone. The teeth are mediolaterally compressed, large and triangular in shape, with pointed apices. Both teeth are clearly well-spaced, with gaps measuring between 10% and 15% of the tooth length. An accessory cusp is apparently present on the anterior edge of the larger tooth (Fig. 6A). In lingual view, the interdental grooves are ventrally directed and the lateral surface of the teeth appears concave (Fig. 6B). In the occlusal view, the interdental gap is well developed. The cutting edge of the teeth projects anterolaterally with respect to the central axis of the bone.

Remarks. Some features observed in the specimen, such as the acrodont dentition fused to the dorsal surface of the bone and the labiolingual compression of the teeth with a flattened lingual surface, are characteristic of Chamaeleonidae (Moody and Roček 1980). However, as Rage and Bailon (2011) and Georgalis et al. (2023b) pointed out, these traits are shared with some agamid taxa. Thus, under this premise and due to the poor preservation of the specimen, the attribution of this remain as a chameleon needs to be met with some caution.

Chamaeleonids are a group of highly specialised and charismatic squamates, with 228 described species distributed across Africa, the Middle East, southern India and Sri Lanka, as well as some Mediterranean islands and parts of southern Europe (Speybroeck et al. 2016). Molecular data suggest a Cretaceous origin of chamaeleonids in Africa (Tolley et al. 2013); however, their fossil record remains sparse (Bolet and Evans 2013). The earliest fossils have been reported from the Early Miocene of Europe and Kenya (Bolet and Evans 2013). Chamaeleonids are thought to have reached Europe during MN 3 (Čerňanský (2010a); Čerňanský et al. (2015); Georgalis et al. (2016, 2023b)). Traditionally, the European fossil record of the group has been considered limited primarily to the Early and Middle Miocene of Central Europe. However, in recent years, new discoveries have expanded their known distribution within Europe for this time period. For instance, Georgalis et al. (2016) described the presence of fossil chamaeleonids in Greece.

In the Iberian Peninsula, Böhme and Ilg (2003) noted the presence of Chamaeleo aff. caroliquarti at the Agramón site (MN3 or MN3/4, see Szyndlar and Alférez (2005)), but provided no figure or description or collection number that could allow evaluation of such claim. Those fossils had been considered lost (Georgalis et al. 2016) until their recent re-discovery, which proved the presence of an undetermined chameleonid within the assemblage (A. Villa, pers. comm.). More recently, Bolet et al. (2022) reported, in a conference abstract, Chamaeleonidae indet. from the site of Turó de les Forques 1 (Vallès-Penedès Basin, Catalonia, north-eastern Spain; MN3). Together with the material described in this study, these findings confirm the presence of chameleons in the Early Miocene of the Iberian Peninsula. In addition, Talavera and Sanchiz (1983) described remains of Chamaeleo chamaeleon from the Bronze Age site of Rincón de la Victoria (Málaga, southern Spain). Currently, this taxon occurs discontinuously along coastal regions in the south and east of the Iberian Peninsula, from the Algarve (Portugal) to Valencia (Spain). The origin of these Iberian populations was likely anthropogenic, resulting from recent introductions (Salvador 2014).

Palaeoecology. To date, all Eurasian chamaeleonid fossils have been tentatively assigned to the genus Chamaeleo (Čerňanský (2010b, 2011); Sankhyan and Čerňanský (2016)). Although, as Georgalis et al. (2016) noted, the absence of a particular group in the fossil record does not preclude its existence at a given time, there is currently no evidence for the presence of other chamaeleonid genera in Eurasia during the Miocene.

Most species within Chamaeleo are typically arboreal. However, some exhibit less arboreal behaviour, such as Chamaeleo arabicus Matschie, 1893 (van der Kooij 2000) and Chamaeleo namaquensis Smith, 1831 (Tolley and Burger 2007). Čerňanský (2010b) hypothesised that chameleons with non-contacting tooth bases inhabited closed habitats. Following this hypothesis, the studied specimen may belong to a chameleon that lived in forested or densely vegetated shrubland environments.

Anguidae Gray, 1825

Ophisaurus Linnaeus, 1766

Ophisaurus (sensu lato) sp.

Fig. 10A–G

Localities. MCX4, BC1, MAB0B, MAB3, MAB6, MAB11, MAB0A, MAB0B, MAB0G and CBR3.

Material. 1555 osteoderms (MGUV-40403, 40417, 40418, 40419, 40427, 40431, 40455, 40466, 40601, 40629 and 40680), two right maxillae (MGUV-40537 and 40681), one maxilla (MGUV-40451), one left dentary (MGUV-40565), seven tooth-bearing bones (MGUV-38352 40388, 40436 and 40596), three isolated teeth (MGUV-40494), one axis (MGUV-40450) and 58 vertebral centra (MGUV-40420, 40421, 40422, 40452, 40453, 40456, 40554, 40556, 40586 and 40630).

Description. A large number of osteoderms have been recovered from numerous sites within the studied Basin. The most complete specimens are rectangular in shape and longer than wide. These osteoderms display two clearly distinct regions. The anterior region has a smooth surface (gliding surface sensuHoffstetter (1962)), initially covered by the preceding osteoderm in the living animal. The remaining surface exhibits vermiculate ornamentation and a low longitudinal keel running centrally across it.

Four fragments of potential maxillae were recovered. All are small in size. The highest number of tooth positions is six in specimen MGUV-40681. The preserved teeth are pleurodont, conical and slightly recurved both distally and lingually. The apices possess well-developed and sharp mesial and distal cutting edges. No striations are visible on the tooth apices. The interdental gaps measure approximately one third of the total tooth width. The sulcus dentalis is poorly developed (Fig. 10A). In lingual view, only one third of the preserved tooth extends above the crista dentalis in MGUV-40596, which, based on its size, likely belonged to the middle region of the tooth row (Fig. 10B).

As with the maxillae, the dentaries are also fragmentary. The teeth are similar to those described above. In lingual view, the Meckelian groove is open. MGUV-40405 presents the Meckelian groove in lateral position.

The centrum of the axis is large and robust. The odontoid process is fused to the centrum without any suture. This process is short and wide, with a blunt border. In ventral view, the odontoid process slightly projects beyond the anterior edge of the centrum. In this view, the margo lateralis is straight, giving the centrum a rhomboid shape. The centrum is of similar width and length. The anterior region of the centrum is clearly convex. The synapophysis, which is wide and short, projects posterolaterally, forming an angle of about 45° with respect to the axis of the centrum. This inclination gives the front part of the centrum a rounded appearance. The axis shows the bases of two robust intercentra, which are not fused. A protruding keel connects both intercentra, formed by the robust intercentrum 3. A well-defined pair of subcentral foramina is present (Fig. 10C). In lateral view, the odontoid process projects anterodorsally and is inclined at approximately 130° relative to the horizontal plane (Fig. 10D). In anterior view, this structure is ovoid in shape. The anterior articular surface is wider than tall. Due to the anterior projection of the odontoid process on the dorsal part of the facet, the entire surface is dorsoventrally concave (Fig. 10E). In the posterior view, the condyle is dorsoventrally flattened and broader than it is tall, having an elliptical shape. In this view, the dorsal margin of the condyle is slightly convex (Fig. 10F).

Fifty-eight vertebral centra have been recovered in five sites of the Ribesalbes-Alcora Basin (CBR0B, CBR0G, MAB3, MAB6 and MAB11). All are procoelous and longer than wide (Fig. 10G). In ventral view, the margo lateralis of the centrum is slightly concave. In its anterior region, the centra show a pair of well-defined subcentral foramina. Some centra have a faintly defined haemal keel occupying two thirds of their total length. However, the studied material presents a high degree of variability. Thus, there are centra with clearly concave margo lateralis, whereas other fossils show parallel borders (MGUV-40451). Moreover, defining this character is difficult because of damage to the anterior region in some specimens. In the anterior and posterior views, the cotyle and condyle are dorsoventrally depressed, respectively.

Remarks. Some traits are consistent with an attribution to the genus Ophisaurus sensu lato, i.e. presence of recurved, sharply pointed and broadly based teeth; elliptical condyle of the axis with a convex dorsal edge in posterior view and dorsal vertebra with convex margo lateralis (Klembara et al. 2014; Čerňanský et al. 2019). Ophisaurus traditionally includes three highly divergent lineages: Ophisaurus s.s. from North America, Dopasia from Southeast Asia and Hyalosaurus from North Africa (Macey et al. 1999). However, osteological differentiation amongst these taxa is still poorly understood. To these genera must be added the recently described Smithosaurus Vasilyan et al., 2022 and Fontisaurus Villa et al., 2025, which has a dentition similar to some species of Ophisaurus s.l. (Vasilyan et al. 2022; Villa et al. 2025). Amongst the various axis vertebrae identified by Čerňanský et al. (2019) for the extant species of Ophisaurus s.l., the studied material resembles that of Dopasia gracilis (Gray, 1845). Nevertheless, the tooth morphology is not consistent with any extant species of the genus, as they typically exhibit lingual striations (Klembara et al. 2014). Although most tooth-bearing bones studied are small and may belong to juveniles (where such striation is apparently absent or weakly developed, as noted by Klembara et al. (2014)), the presence of a large isolated tooth (MAB3-890) demonstrates that adult specimens also lacked striations. In this sense, taxa having unstriated teeth have reported from the Miocene of Europe, such as F. tarumbaire (Villa et al. 2025).

Numerous fossil remains of Ophisaurus-like taxa with unstriated teeth have been described from localities of similar age, including Ophisaurus sp. from Béon 1 (France) (Rage and Bailon 2005), Ophisaurus cf. spinari, Ophisaurus sp. 1 and 2 from Dolnice (Czechia) (Roček 1984), Ophisaurus from Oberdorf 4 (Germany) (Čerňanský et al. 2015), Ophisaurus morphotype 2 from Ulm-Westtangente (Germany) (Klembara et al. 2019), Anguine morphotype I from Merkur-Nord (Czechia) (Klembara 2015) and cf. Ophisaurus sp. from Karydia-3 (Greece) (Georgalis et al. 2019b). Some differences between our material and these taxa concern the width of the tooth base relative to the apex, the degree of curvature and the extent to which the tooth height exceeds the crista dentalis. However, Roček (1984) and Georgalis et al. (2019b) noted that such traits can exhibit intraspecific variability.

Nowadays, Ophisaurus s.l. is represented in Asia by six species included in the genus Dopasia, which are distributed from northern India through the Indochinese Peninsula southwards to Indonesia: Dopasia buettikoferi (Lidth de Jeude, 1905); Dopasia gracilis; Dopasia hainanensis Yang, 1983; Dopasia harti (Boulenger, 1899); Dopasia sokolovi (Darevsky & Nguyen, 1983); and Dopasia wegneri (Mertens, 1959) (Nguyen et al. 2011). Regarding North Africa, only Hyalosaurus koellikeri (Günther, 1873) is recorded in northern Morocco (Macey et al. 1999).

Palaeoecology. Dopasia and Hyalosaurus include forest-dwelling species, although little is known about their ecology (Nguyen et al. 2011). Hyalosaurus koellikeri, the sole species in its genus, is better understood and considered a forest-dwelling lizard (Escoriza and Comas 2015). According to the latter authors, the occurrence of the species is mainly explained by the presence of forest land, but H. koellikeri does not have a strong preference for a specific habitat composition and was found occupying both dry forests of junipers and evergreen oaks and humid deciduous forests.

Members of Ophisaurus s.l. inhabited Europe until the latest Middle Pleistocene, coinciding with the disappearance of the subtropical humid forest biome around 1.2 Ma. The extirpation of the genus in this area occurred during the Early-Middle Pleistocene transition, an episode of major climatic change associated with the intensification of Northern Hemisphere glaciation between 1.4 and 0.7 Ma (Blain and Bailon 2019). In this sense, the Ophisaurus taxa with unstriated teeth recovered in European localities have been interpreted as forest species (i.e. Ophisaurus acuminatus Jörg, 1965; see Klembara and Čerňanský (2020)).

cf. Pseudopus sp.

Fig. 10H, I

Locality. MAB0A and MAB11.

Material. 197 osteoderms (MGUV-40490, 40502, 40506, and 40516), one pterygoid (MGUV-40550), two tooth-bearing bones (MGUV-40496 and 40549) and six vertebral centra (MGUV-40501 and 40517).

Description. MGUV-40550 is probably a fragmentary pterygoid. This bone shows three to four tooth rows on the ventral surface of the bone. The teeth are low, stout and blunt, but some of them bear a small cusp. The most prominent teeth are located in the lateral row, while the remaining ones are smaller (Fig. 10H).

The recovered tooth-bearing bones have pleurodont, monocuspid, cylindrical, bulbous and robust teeth. Their crowns are blunt without distinguishable cusps. The teeth are well-spaced and non-striated (Fig. 10I).

Remarks. Nowadays, the genus is limited to a single species, Pseudopus apodus (Pallas, 1775). Its distribution includes the Balkans, the Crimean Peninsula and the Ciscaucasia Region in Europe, as well as Asia Minor and the Middle East (Speybroeck et al. 2016). So far, four extinct species have been recognised: Pseudopus confertus Klembara & Rummel, 2018 (Early Miocene, MN3, Ahníkov I, Merkur North opencast mine, Czechia), Pseudopus ahnikoviensis Klembara, 2012 (Early Miocene, MN3, Central Europe), Pseudopus laurillardi (Lartet, 1851) (Early–Middle Miocene of Europe) and Pseudopus pannonicus (Kormos, 1911) (Late Miocene–Middle Pleistocene of central and eastern Europe; Klembara and Rummel (2018); Loréal et al. (2025)). The Iberian fossil record of Pseudopus is scarce and mainly limited to the Early–Middle Pleistocene of eastern Spain (Blain et al. 2016). During the Miocene, the genus is only reported in the Iberian Peninsula from the localities of els Casots (cf. Pseudopus; earliest MN5, Middle Miocene; Vallès-Penedès Basin, Catalonia, Spain), Crevillente 2 (MN11, Late Miocene; Bajo Segura Basin, Comunitat Valenciana, Spain) , and Libros (cf. Pseudopus; middle-late Tortonian, Late Miocene; Teruel Basin, Teruel, Spain) (Casanovas-Vilar et al. 2022; Marquina-Blasco et al. 2025; Linares-Martín et al. 2026).

Palaeoecology. Although the only extant species of the genus Pseudopus, Pseudopus apodus, inhabits dry and well-vegetated rocky slopes, extinct species have been interpreted as subhumid to humid forest dwellers (Klembara et al. 2010).

Anguidae indet.

Fig. 10J

Localities. MCX1, MCX2, MCX3, MTR2, MCX5, MCX6, MTR3, MCX7, BC1, BC2, MAB0A, MAB0B, MAB1, MAB3A, MAB4, MAB5, FS1, MAB7, MAB8, MAB9, MAB10, MAB11A, MAB11B and MAB13.

Material. 1574 osteoderms (MGUV-38555, 38556, 40363, 40379, 40380, 40391, 40395, 40409, 40437, 40446, 40447, 40448, 40470, 40471, 40473, 40474, 40475, 40488, 40530, 40542, 40543, 40570, 40571, 40602, 40607, 40621, 40622, 40632, 40636, 40638, 40639, 40648, 40653, 40658, 40659, 40660, 40662, 40665, 40669, 40678, 40682, 40683, and 40689), one pterygoid (MGUV-40675), two tooth-bearing bones (MGUV-38370) and 55 vertebral centra (MGUV-40366, 40367, 40387, 40392, 40407, 40438, 40443, 40449, 40464, 40468, 40483, 40536, 40541, 40617, 40619, 40635, 40640, 40641, 40664, 40671 and 40673).

Description. This taxon includes all the material exhibiting characters attributable to Anguidae, but which cannot be confidently assigned to either of the two genera described in the present study, due to the poor state of preservation of the material or its limited taxonomic value beyond the family level.

Remarks. Despite the above, the presence of subrectangular osteoderms bearing a longitudinal keel is a trait found in non-Anguis anguine taxa (i.e. Pseudopus or Ophisaurus s.l.; Hoffstetter (1962)).

Palaeoecology. The taxonomical attribution of that taxon precludes any palaeocological interpretation.

Scolecophidia Duméril & Bibron, 1844

Scolecophidia indet.

Fig. 11

Locality. CBR0B.

Material. One trunk vertebra (MGUV-40416).

Description. The vertebra has a short centrum. In dorsal view, the anterior edge of the neural arch is missing. Although it is partially broken, it is appreciable that the prezygapophysis projects anterolaterally. No neural spine is present. The postzygapophysis is broad and short (Fig. 11A). In ventral view, the centrum is elongated. Its surface is relatively concave and has a pair of subcentral foramina. The centrum is devoid of any structure, with no signs of haemal keel or hypapophysis. The posterior edge of the centrum is broken (Fig. 11B). In anterior view, the neural arch is dorsoventrally flattened. Likewise, the cotyle is dorsoventrally flattened and elliptical (Fig. 11C). In posterior view, the neural arch is again flattened dorsoventrally, although to a lesser extent than in anterior view.

Remarks. The described characters are typical of Scolecophidia (Szyndlar and Georgalis 2023). However, vertebrae lack taxonomically valuable information below the subordinal level (Bailon (1991); Szyndlar and Georgalis (2023)).

Palaeoecology. Scolecophidians are a group of fossorial, primitive snakes that inhabit tropical and subtropical regions, with greater diversity on the Gondwanan continents (Vidal et al. 2010). Currently, the unique autochthonous scolecophidian in Europe is the typhlopid Xerotyphlops vermicularis. This species occupies open, dry habitats, characterised by both heavily degraded vegetation and dense stands of grasses and herbs (Akman and Göçmen 2019). The main factors affecting the distribution of the species are the annual mean temperature, the driest-month precipitation and annual precipitation (Afsar et al. 2016).

Colubroidea Oppel, 1811

Colubridae Oppel, 1811

Natricidae Bonaparte, 1838

Palaeonatrix Szyndlar in Młynarski et al. 1982

Palaeonatrix sp.

Fig. 12

Localities. BC1, MAB0B, MAB3, MAB5 and MAB11.

Material. Eight vertebrae (MGUV-40377, 40533, 40543, 40544, 40557 and 40583) and 40 vertebral centra (MGUV-40376, 40378, 40382, 40383, 40404, 40558, 40590 and 40618).

Description. The studied material includes vertebrae (Fig. 12A–P) and isolated vertebral centra (Fig. 12Q–S). The studied vertebrae are procoelous and small-sized (CL_med = 3.74 ± 0.57 mm; NAW_med = 2.47 ± 0.30 mm; CL/NAW_med = 1.51 ± 0.05 mm; n = 2) (Fig. 12A–O). In dorsal view, the neural spine runs along the entire length of the neural arch, except for its anterior third. The interzygapophyseal constriction is poorly marked (Fig. 12A, F, H, L). The prezygapophyses is only partially preserved in MGUV-40543. Its articular surface is large and oval (Fig. 12F). In lateral view, the neural spine is low and does not overhang posteriorly. The margo lateralis is well defined and reaches the condyle (Fig. 12B). Behind the prezygapophyses, there is a large, well-defined lateral foramen. The parapophyseal process, only present in MGUV-40577, is wide and short, projecting antero-ventrally (Fig. 12I). The centrum has a wide hypapophysis with a strong developed base, anterior keel sensuMłynarski et al. (1982). However, the end of the hypapophysis is broken, except in the case of the isolated centrum MGUV-40382, in which the tip of the hypapophysis does not overpasses the cotylar tip (Fig. 12Q). It does not contact the condyle (Fig. 12B, Q, R). In ventral view, the centrum is elongated and narrow. Prominent and blunt subcentral ridges are present. The subcentral grooves are well-defined and relatively narrow. They extend almost parallel to the long axis of the vertebra (Fig. 12C, J, S). It is not possible to discern whether this is due to deformation by sediment pressure. In anterior view, the neural arch is high, flattened laterally and shows its upper limit slightly domed (Fig. 12D, G, K). The prezygapophysis is horizontal (Fig. 12G). Paracotylar foramina are present in wide depressions on each side of the relatively small and rounded cotyle, whose ventral end is straight (Fig. 12D, K, P). The parapophysis overpasses the ventral limit of the centrum (Fig. 12K). In dorsal view, the condyle is relatively small and rounded (Figs 12D, 12E), although, in some specimens, it appears slightly dorsoventrally flattened (Fig. 11E).

Remarks. The presence of a relatively well-developed hypapophysis accompanied by prominent anterior keel distinguishes our material from other natricid genera described in the European Miocene, i.e. Natrix Laurenti, 1768 and Neonatrix Holman, 1973. Regarding the recently described Wintershofia robusta Ivanov et al., 2025, our material differs from this taxon by the less elongated centrum, the hypapophysis overpasses the caudal rim of the condyle, the anterior keel is higher and the parapophyseal processes are shorter (Ivanov et al. 2025). In summary, the presence of a prominent anterior keel accompanying the hypapophysis, a strongly developed margo lateralis reaching the condyle, parapophysis projected below the ventral limit of the centrum and a vestigial neural spine are diagnostic characters of the genus Palaeonatrix (Szyndlar (1987); Rage and Bailon (2005); Ivanov et al. (2025)).

Two species are recognised within Palaeonatrix, Palaeonatrix silesiaca Szyndlar, 1982 in Młynarski et al. (1982) and Palaeonatrix lehmani (Rage & Roček, 1983), which are distinguishable, based on minor features, such as the degree of lateral expansion of the postzygapophysis, presence of epizygapophyseal spine in P. lehmani, but not in P. silesiaca and the different size of the condyle (Szyndlar (1991); Ivanov et al. (2025)). The fossil record of Palaeonatrix is primarily known from the Miocene of Eastern Europe, whereas its record in the Western region is scarcer (Ivanov 2022). A probably new taxon, morphologically very close to P. lehmani, has been reported from the French localities of Béon 1 (MN4) and Sansan (MN6) (Augé and Rage (2000); Rage and Bailon (2005)). In our material, the size of the cotyle and the neural arch are similar. In contrast, in P. lehmani from Petersbuch 2, the second is slightly larger than the first, as is the case in P. aff. lehmani from Sansan (Augé and Rage 2000). However, the poor state of preservation of the material studied here precludes any specific attribution.

This genus is known mainly from the Early and Middle Miocene of Eastern Europe (Szyndlar 1991), but only two previous reports from Western Europe, the above-mentioned French localities of Béon 1 and Sansan (Augé and Rage (2000); Rage and Bailon (2005)). Thus, the material from the Ribesalbes-Alcora Basin represents the first mention of Palaeonatrix from the Iberian Peninsula.

Palaeoecology. Most extant natricid species are linked to aquatic habitats (Deepak et al. 2022). Palaeonatrix is presumed to have inhabited diverse aquatic environments based on taxonomic inference (Ivanov et al. 2025).

cf. Viperidae indet.

Fig. 13A–E

Localities. MAB3, MAB11 and CBR0D.

Material. One vertebra (MGUV-40592) and three vertebral centra (MGUV-40441, 40442 and 40560).

Description. MAB3-886 is a small-sized procoelous vertebra. It is as long as it is wide (however, MAB11-168 is a centrum slightly longer than wide). The interzygapophyseal constriction is well developed (Fig. 13A). In lateral view, the diapophysis is rounded and the parapophysis projects antero-ventrally. The centrum is transversally convex. The hypapophysis preserves only its base, which is in contact with the condyle (Fig. 13B). In ventral view, the centrum is triangular in shape, wider anteriorly than posteriorly, with a poorly-defined margo ventralis. A pair of small subcentral foramina is located in the anterior part of the centrum (Fig. 13C). The condyle and the cotyle are large and rounded, respectively, in anterior and posterior views (Figs 13D, 13E). In this latter view, the neural arch is slightly dorsoventrally flattened (Fig. 13E).

Remarks. The studied remains share some characteristics typical of Viperidae, such as a dorsoventrally flattened neural arch, a transversely convex centrum with a vague margo ventralis, well-developed cotyle and condyle and a condyle ventrally connected to the base of the hypapophysis (Bailon 1991). However, the poor state of preservation of these remains allows us to propose only a very tentative taxonomic attribution for them.

Palaeoecology. The taxonomic attribution of this taxon precludes any palaeocological interpretation.

Alethinophidia indet.

Localities. MCX3, MTR2, MTR3, MCX7, BC1, BC2, MAB0A, MAB0B, MAB1, MAB3, MAB3A, MAB5, FS1, MAB6, MAB7, MAB8, MAB9, MAB10, MAB11, MAB11A, CBR0A, CBR0B, CBROC, CBR0D, CBR0G and CBR1.

Material. one left dentary (MGUV-40628), one right dentary (MGUV-40469), two dentaries (MGUV-40428 and 40589), two tooth-bearing bones (MGUV-40359), 32 isolated teeth (MGUV-25343, 36660, 36672, 36741, 36764, 40485, 40491, 40586, 40626 and 40627), seven precaudal vertebrae (MGUV-40509, 40510, 40512, 40513, 40527 and 40676), 23 caudal vertebrae (MGUV-40362, 40369, 40370, 40372, 40373, 40424, 40439, 40456, 40493, 40508, 40589, 40520, 40521, 40522, 40527, 40535, 40590, 40633, 40637, 40690, 40692, 40651 and 40654), one isolated caudal vertebral centrum (MGUV-40519), three isolated neural arches (MGUV-40507), 262 vertebral centra (MGUV-40364, 40374, 40393, 40415, 40423, 40430, 40457, 40466, 40467, 40481, 40482, 40498, 40500, 40511, 40514, 40518, 40525, 40526, 40534, 40540, 40559, 40561, 49562, 40563, 40584, 40585, 40587, 40588, 40591, 40604, 40626, 40631, 40641, 40642, 40650, 40661 and 40691) and 27 isolated condyles (MGUV-40375). Fig. 13F–Q.

Description. This taxonomic attribution includes material which, either because of its poor state of preservation or its low value for taxonomic determination, could not be included in any of the taxa described above. However, three distinct caudal vertebra morphologies are present in this assemblage, most probably corresponding to positional variation along the vertebral column rather than taxonomic differences (Szyndlar and Georgalis 2023). The first morphology includes 22 vertebrae (MGUV-40370, 40424, 40493, 40520, 40521, 40522, 40535, 40590, 40633, 40637, 40690, 40692, 40651 and 40654). They are procoelous, small and elongated. In dorsal view, the interzygapophyseal constriction is poorly marked. The neural spine extends over the entire length of the neural arch, except for the region of the zygosphene. It is broad and trilobate. The prezygapophysis projects anteriorly and presents a large, elliptical articular surface. The prezygapophyseal process is broken (Fig. 13F). In lateral view, the neural spine is low. The margo lateralis is not well defined. The pleurapophyses are well developed, triangular in shape and extend through the anterior three quarters of the total centrum length. They project anteroventrally. The haemal processes are low and rounded. The lateral foramen is small and poorly defined. In ventral view, the centrum is triangular and elongated (Fig. 13G). In anterior view, the neural canal is dorsoventrally compressed, with a convex roof (Fig. 13H). Cotyle and condyle are small and rounded in the anterior and posterior views, respectively.

The second morphology is represented by MGUV-40527, which preserves only its left side. In dorsal view, the interzygapophyseal constriction is well marked (Fig. 13I). The prezygapophysis is distinguishable from the rest of the vertebra. The prezygapophyseal process is short and acuminate. The centrum appears slightly longer than wide in ventral view (Fig. 13J). The articular surface of the postzygapophysis is rounded. The lymphapophyses are broken; however, the dorsal ramus retains its base, which is broad and dorsoventrally flattened. The presence of lymphapophyses suggests that the ventral structure corresponds to the left branch of the haemal process, which is wide, prominent and has parallel margins; it also appears tall in lateral view (Fig. 13K). The vertebra has a cotylar neck distinct from the centrum (Fig. 13L).

Finally, the third morphology is represented by 11 small-sized and procoelous vertebrae (MGUV-40362, 40369, 40372, 40373, 40439, 40456, 40508 and 40589). In dorsal view, the interzygapophyseal constriction is poorly marked. The zygosphene is slightly concave. Prezygapophyses are indistinct from the neural arch and project anterolaterally (Fig. 13M). In the lateral view, the neural spine is low. The margo lateralis is well marked. The lateral foramen is large and a small accessory foramen is present behind the paradiapophysis. The margo ventralis is weakly defined. In ventral view, the centrum is short and slightly longer than it is wide. A pair of large, well-defined subcentral foramina is present. The centrum bears paired haemal processes which, in lateral view, are short and project ventrally. The pleurapophyses are directed anteroventrally (Fig. 13N). In the anterior view, the neural canal is large and dorsoventrally flattened (Fig. 13O). Both cotyle and condyle are small and dorsoventrally compressed in anterior and posterior views, respectively.

Remarks. The first morphology has some similarities with the caudal vertebra recently attributed to Palaeonatrix aff. lehmani from the Early Miocene locality of Wintershof-West, Germany by Ivanov et al. (2025). These authors proposed this attribution, although tentatively, on the basis of the low neural spine and large diameter of the cotyle. Our material differs from this specimen by its lower neural spine and the convex shape of the roof of the neural canal in anterior view.

Additionally, this type of remains is known from two taxa of the genus Texasophis: Texasophis bohemiacus Szyndlar, 1987 from the Early Miocene of Czechia and Texasophis sp. from the Turkish site Kilçak 3B (MN1, Early Miocene) (Szyndlar (1994); Syromyatnikova et al. (2019)). The present material differs from Texasophis sp. mainly in the shape of the pleurapophyses in ventral view. In our specimens, the pleurapophyses are longer (exceeding half the centrum length) and have straight, inclined edges. In contrast, in Texasophis sp., the edges are obtuse, giving the pleurapophysis a trapezoidal appearance. Furthermore, in Texasophis sp., both cotyle and condyle are dorsoventrally flattened (fig. 10 in Syromyatnikova et al. (2019)), whereas in our specimens, both are rounded. However, these differences may be due to the vertebral position within the tail, as the Kilçak material likely corresponds to the anterior caudal region. Regarding T. bohemiacus, the described vertebrae are indistinguishable from those of this species from the Ehrenstein 12 site (middle Oligocene, Germany; Szyndlar (1994)). Nevertheless, two reasons prevent a genus-level assignment: (1) the absence of additional remains corroborating the presence of Texasophis in the studied assemblage and (2) the limited comparative data on caudal vertebrae in fossil colubrids.

Palaeoecology. The taxonomical attribution precludes any palaeocological interpretation.

Squamata indet.

Localities. ACS2, MAB0A, MAB0C, MAB3, MAB4, MAB8 and CBR0B.

Material. Three osteoderms (MGUV-40505, 40538, and 40608), two tooth-bearing bones (MGUV-25298 and 40633) and six vertebral centra (MGUV-40346 and 40426).

No more precise attribution is possible for these materials due to their poor state of preservation.

Palaeoecology. The taxonomical attribution of this taxon precludes any palaeocological interpretation.

A total of 4,690 remains of reptiles and amphibians were recovered from various sites within the Ribesalbes-Alcora Basin (Fig. 14). By far, the most common elements are anguid osteoderms (3,317, 70.73% of the overall) (Table 1). The studied herpetofauna comprises 21 taxa. There are no significant changes in the faunal composition throughout the sequence. However, based on the appearance of specific taxa, some palaeoecological implications can be inferred. Amongst the localities studied, the following can be considered as intensively sampled (from the oldest to youngest): MCX3, MTR2, BC1, MAB0A, MAB0B, FS1, MAB3, MAB5, MAB11, CBR0B and CBR1 (Crespo et al. 2019c). Therefore, the palaeoecological analysis is primarily based on these sites.

Taxa linked to aquatic habitats (i.e. lissamphibians and crocodylians) are common at most sites, although they are absent from MCX1, 2, 4, 5 and 6; MTR3; MAB0C, 1, 6, 7 and 8; and CBR0C and 2. The lithologies of the section indicate a depositional context interpreted as a distal deltaic and shallow lacustrine environment (Anadón (1983); Crespo (2017); Crespo et al. (2019c)). Based on this and following the palaeoecological inferences proposed by Crespo (2017) and Albesa et al. (2022), we observe variability in “wet” and “dry” conditions throughout the Ribesalbes–Alcora Basin sequence. In this context, we have applied the HEG method to the herpetological assemblages of the MCX3, MTR2, BC1, MAB0A, MAB0B, FS1, MAB3, MAB5, MAB11, CBR0B and CBR1 sites (Table 2). Nowadays, the Mean Annual Precipitation (MAP) of the Basin is 432.19 ± 7.21 mm (MAP_max: 454 mm; MAP_min: 418 mm). Estimated MAP values indicate higher precipitation than today for all the sites, except for MAB0A.

For comparison with other proxies in the outcrop—namely, small mammals, gastropods and isotopes (Ríos (2013); Crespo (2017); Álvarez-Parra et al. (2021); Albesa et al. (2022); Crespo et al. (2025) and references therein)—we interpret “humid (or dry) conditions” based solely on the ecological preferences of the identified taxa, as indicating high precipitation. High annual rainfall generally implies humid environments, but this is not always the case, as it depends on several factors, including the annual distribution of precipitation, actual evapotranspiration, local soil conditions, microclimate and vegetation cover and structure (Ritchie et al. (1999); Beck et al. (2018); Wright and Francia (2024); Kim and Johnson (2025)). Nonetheless, we adopt a simplified approach to allow comparison with the aforementioned proxies.

The estimated MAP at the oldest site, MCX3, is 564.94 mm, higher than the current average for the Basin and the second-lowest rainfall value in the overall sequence (Fig. 15, Table 2). The oldest site, MCX3, includes both aquatic and terrestrial taxa. Amongst the latter, the presence of Chalcides sp. and Anguidae indet., taxa linked to forest or shrubland vegetation, is notable. The small mammal fauna from MCX3 seems to corroborate the presence of open forest (Crespo 2017), as indicated by the occurrence of Galerix symeonidisi Doukas, 1986, which is often found at lignite-rich sites (Crespo et al. 2020a). However, the rodent assemblage includes Heteroxerus rubricati Crusafont et al., 1955, a ground squirrel typical of warm and dry habitats (Crespo et al. 2021a), suggesting the presence of open areas around the lake.

Regarding MTR2, the site yields a high estimated MAP of 800.23 mm (Fig. 15, Table 2), which contrasts with values inferred from small-mammal data (Crespo 2017). Thus, our proposed palaeoclimatic scenario indicates an increase in annual rainfall between MCX3 and MTR2. Unfortunately, the open taxonomic attribution of most taxa from MTR2 limits palaeoecological interpretations. However, the presence of Anguidae indet. implies some forest/bush cover in the area.

With an estimated rainfall of 473.76 mm, BC1 is considered the second driest site of the overall sequence. Although higher than the current MAP in the region, BC1 shows the smallest difference between the latter and the observed value (+ 41.57 mm) (Fig. 15, Table 2). No xerophilic taxa have been recovered, but some of them have closely-related species that currently inhabit this kind of habitat, such as cf. Lissotriton sp. and Bufonidae indet. On the other hand, particular mammal and gastropod taxa, such as the soricid Paenelimnoecus micromorphus (Doben-Florin, 1964) (Crespo et al. 2019b), the lagomorph Lagopsis penai Royo, 1928 (Crespo et al. 2022) and the helicid Palaeotachea (Albesa et al. 2022), are indicative of open landscapes. Paradoxically, the occurrence of L. penai in the site is residual, ~ 3%, whereas in other localities, such as MTR2, its presence is higher, ~ 24% (Crespo et al. 2022). The isotopic analysis by Ríos (2013) estimated a MAT of 20.09 °C.

MAB0A records the lowest rainfall value of all studied sites. In fact, it is the only locality with lower value than the one recorded nowadays in the region, with a differences of -30.74 mm. However, remains attributable to cf. Lissotriton sp. have been found in MAB0A, supporting the higher humidity at this site than at the other ones. In line with our result, Crespo et al. (2019c) documented gypsum layers occurring at the top of the ferruginous MAB0A layer and at the bottom of MAB0B, which differ markedly from the aforementioned MAB0A layer. This aridity trend was accompanied by an increase in the MAT for MAB0A (21.16 °C) with respect to BC1 (20.09 °C), in accordance with the preliminary results of Ríos (2013).

MAB0B and FS1 yielded similar palaeoherpetological assemblages. These sites represent a new humidity peak within the sequence, with high MAP values of 765.12 and 620.00 mm, respectively, although, a progressive decrease of the MAP value from MAB0B to FS1 is perceptible (Fig. 15, Table 2). In this sense, the herpetofauna assemblages are composed of highly generalist taxa, as Gekkota indet. and Lacertidae indet. Compared to the previous site, MAT has recorded a significant decline (21.16 °C for MAB0A and 17.95 °C for MAB0B) (Ríos 2013).

Located at the base of section FS, the Konservat-Lagerstätte site FSA has been interpreted as a shallow, perennial and quiet palaeolake, based on its palaeofaunal assemblage, which comprises arthropods, lymnaeid and planorbid molluscs, ostracods, cladocerans, charophytes and fishes (Álvarez-Parra et al. 2021). Nonetheless, its extension could have fluctuated probably cyclically. A key proxy for this supposition is the cladoceran crustacean Daphnia (Ctenodaphnia), which usually breeds partenogenetically. However, under unfavourable environmental conditions, they produce diploid eggs encapsulated in a protective, saddle-like structure called an ephippium, which is strongly melanised and contains two large eggs (Ebert 2005). The mass records in the rock slabs of Daphnia (Ctenodaphnia) ephippia suggest ecologically stressful conditions, such as temporary desiccation or a lack of food resources (Álvarez-Parra et al. 2021). Although La Rinconada is slightly older than the studied outcrops, Barrón et al. (2010) pointed out that the palaeoclimatology was characterised by drought seasonality. Climatic conditions during the Early Miocene in the region have been interpreted as tropical or subtropical (Peñalver et al. (2016); Crespo (2017)). In this sense, shallow tropical lakes are influenced by seasonal cycles of run-off and water input. In a few cases, the loss of water by evaporation from a lake is larger than that by surface outflow. Thus, only shallow tropical lakes of the arid and semi-arid tropics are also liable to at least temporary extinction, due to long-term rainfall deficits (Talling 2001).

The downward trend in annual rainfall finishes in MAB3. The obtained value, 488.9 mm, is the second-lowest MAP amongst the studied sites. The presence of Chalcides sp. could be a proxy of some degree of dry conditions (Caputo et al. 1995). In MAB3, the small mammal assemblage shows affinities with forested environments (Crespo et al. 2019c). Amongst the herpetofauna, the only clearly forest-associated taxa are Ophisaurus s.l. sp. (Escoriza and Comas 2015).

MAB5, MAB11 and CBR0B represent a positive trend from the oldest (MAB5) to the youngest (CBR0B) of these sites, with MAP values of 532.93, 682.19 and 889.72 mm, respectively (Fig. 15, Table 2). It is relevant to note that this phase ended with the highest MAP value of the overall sequence, at the CBR0B site. Some taxa described in this site seem to corroborate this assumption. Ophisaurus s.l. sp. is interpreted as a woodland inhabitant. Another potential forest-dwelling member of the palaeoherpetofaunal assemblage from CBR0B is ?Chamaeleonidae indet. As mentioned previously, the presence of a tooth with its bases not in contact could be a proxy for closed-canopy forest or dense shrubland habitats (Čerňanský 2010b). At first glance, our results do not appear to contradict the palaeoecological reconstruction proposed by Crespo (2017), who identified an expansion of closed forests, based on small vertebrate faunas. In this regard, the abundance of the dimylid Plesiodimylus ilercavonicus in MAB5 has been interpreted as indicative of very humid conditions (Crespo et al. 2018). Moreover, the presence of the flying squirrel Aliveria cf. luteyni in CBR0B indicates wooded formations in the surrounding area of the site (Crespo et al. 2021a). Paradoxically, in CBR0B ground squirrels, Heteroxerus rubricati and Aragoxerus sp., reached one of the highest frequencies of the overall palaeomammal assemblages (Crespo et al. 2021a). Both taxa have been linked to warm, relatively dry and open habitats with scarce vegetation (Van Dam and Weltje 1999). The gastropod assemblage appears to support this inference (Albesa et al. 2022). This phase was accompanied by a decline in MAT in MAB5 (13.31 °C), followed by an increase in MAB11 to 19.02 °C—values comparable to those at MAB0A and MAB0B (Ríos 2013).

Finally, MAP decreases again at the CBR1 site, reaching 765.12 mm (Fig. 15, Table 2). However, the obtained MAP is still high, given that it has the same value as MAB0B.

CBR1 presents an assemblage composed of forest-associated taxa (cf. Pseudopus sp.) or generalist (Lacertidae indet.) taxa. In this sense, the small mammal assemblages from the sites indicate drier environmental conditions (Crespo 2017). CBR1 represents the second locality with the highest frequency of ground squirrels of the overall sequence. Specifically, Aragoxerus sp. has been recovered from CBR1, which has been interpreted as an ecology linked to warm, dry open habitats by taxonomic inference (Crespo et al. 2021a).

The palaeoclimatic analysis, based on small mammals for the same localities, as proposed by Crespo (2017), relies on the palaeoecological preferences of different taxa inferred from dental morphology. As a result, that reconstruction is qualitative, whereas our results are quantitative. Moreover, both approaches yield significantly different outcomes. Crespo (2017) suggested a continuous increase in relative humidity from MCX3 to MAB5, followed by a slight decrease in humidity at FS1 and MAB11 and then a renewed increase at CBR0B. A final decrease in moisture availability was proposed towards the end of the sequence. In contrast, our results show a more variable scenario, with an alternation of at least three “humid” phases (MTR2; MAB0B to FS1; and MAB5 to CBR1) and three “dry” peaks (MCX3; BC1 to MAB0A; and CBR1).

These discrepancies amongst the three methods stem from differences in the proxies used. Both Crespo’s (2017) approach and ours are grounded in actualism—that is, reconstructing the ecology of extinct organisms through analogies with closely-related extant species. This approach, however, entails certain limitations. For small mammals, the correlation between climatic ranges and dental morphology in modern taxa may not fully apply to fossil forms (Hernández-Fernández and Peláez‐Campomanes 2005). Regarding reptiles and amphibians, the HEG method assumes that the physiological strategies related to thermoregulation, water balance and gas exchange have remained largely consistent amongst related taxa (Böhme et al. 2006). Their distribution is primarily shaped by climatic factors (Rodríguez et al. 2005), although they exhibit a broad range of adaptations to cope with diverse environmental conditions (Vitt and Caldwell 2009). One such adaptation is the restriction of their presence to habitat patches where factors such as topography or soil type create particularly favourable microclimates (Guisan and Hofer 2003). This enables the persistence of species in regions with generally unfavourable climates (Brito et al. (2011); Peterman and Semlitsch (2013)). The method we applied estimates MAP, based on the climatic preferences of modern relatives of fossil taxa. However, local rainfall variability is strongly influenced by topography and other factors, such as proximity to the sea (Gao et al. 2007). On the other hand, the implemented method is based solely on the presence of the identified taxa, not on their relative abundance. MAB11 is a perfect example of that. In this site, Anguidae indet. and Crocodylia indet. make up 57.89% and 21.56% of the total numbers of remains, respectively; whereas cf. Testudinidae indet. is represented by just two remains, that is, 0.17% of the overall. Despite this, the three taxa have the same importance in the calculation of the palaeoclimatic parameter. Moreover, in some localities, our results differ from other proxies. In this sense, MAB11 is a thin bed of black lignitous mudstones with a ferrugenous bottom (Crespo et al. 2019c). In contrast, isotopic data are subject to biases related to plant and animal physiology and behaviour, environmental chemistry and diagenetic processes (Hayes et al. (1990); Ehleringer and Monson (1993); Koch (1998)).

Vasilyan et al. (2022) applied the HEG method to the palaeoherpetofaunal assemblage from the locality of Echzell (Germany), correlated with MN4. The authors estimated a MAP value of 791 ± 254 mm, indicating more humid conditions than those currently prevailing. This rainfall value exceeds those recorded in the Ribesalbes-Alcora Basin localities, except for MTR2 (800.23 mm) and CBR0B (889.72 mm). However, other sites show similar values to the estimated for Echzell, such as MAB0B (765.12 mm) and CBR0B (765.12 mm). When considered individually, these values seem to contradict previous studies that proposed a lack of pronounced latitudinal regionalisation in Europe during the Burdigalian (Böhme et al. (2011); Bruch et al. (2011)). These differences may reflect palaeolatitudinal or palaeoecological factors that warrant further investigation. Notably, the Ribesalbes-Alcora Basin is part of the Ibero-Levantine bioprovince, which is considered more humid than other parts of the Iberian Peninsula.

Although the assemblages from the Ribesalbes-Alcora Basin do not show significant differences compared to other coeval European and Iberian localities, suggesting certain degree of homogeneity in the palaeoherpetological composition at the family level during the Miocene (Marquina-Blasco et al. 2025). However, it is important to note that this apparent homogeneity may be partly biased by the prevailing taxonomic precision, as identifications are often possible only at supraspecific or suprageneric ranks, potentially masking finer-scale endemism that would require species-level recognition (Villa et al. 2017). Therefore, taking into account the aforementioned limitations, the only probable endemism is cf. Pyrenasaurus sp., which has also been reported from the Eocene of northern Spain and south-eastern France (Bolet and Augé 2014). However, due to the limited number of studies focused on amphibians and reptiles from this time period, particularly in the Iberian Peninsula, further research is needed to confirm or refute this observation.

In general, all sites of the Basin share common elements such as anguids and crocodylians. However, other taxa are only recovered in specific localities and/or sections, which may reflect ecological and/or taphonomic biases or sampling deficiencies. Urodeles are one such group, represented by only one form: cf. Lissotriton sp. The first fossils attributable to that genus were cranial and vertebral material from the extinct species Lissotriton rohrsi (Herre, 1955) at the German locality of Oberleichtersbach (MP30, lower Oligocene; Böhme (2008)). Thus, this genus is a salamandrid taxon with a long natural history in Europe, whose fossil remains have been cited in several sites (Macaluso et al. 2022). During the MN4, remains of the genus have been cited in numerous sites of central Europe. The taxa identified at these sites are related to extant species. Böhme (2003) pointed out the presence of Lissotriton vulgaris (Linnaeus, 1758) in Petersbuch 38B and 36II (MN4) and Wintershof Ost (MN3a) and Lissotriton helveticus from Petersbuch 7, 2 (MN4a) and 4 (?MN4), Stubersheim 2 and 3 (MN 3b) and, finally, Weißenburg 6 (MN1). In the case of the Iberian Peninsula, the fossil record of the genus is really scarce. Until now, during the Late Miocene, ocurrences were cited only in Agramón (Lissotriton aff. vulgaris) and Rubielos de Mora (Lissotriton sp.) (Böhme and Ilg 2003). Our finding thus represents a new record for Lissotriton in the region, extending its known Early Miocene distribution to the easternmost Iberian Peninsula.

The classical site of La Rinconada (Ribesalbes, Castellón), also located in the Ribesalbes-Alcora Basin, is recognised as a Konservat-Lagerstätte due to its exceptionally well-preserved fossils. Amongst the herpetofauna, Sanchiz (1977b) described material assigned to three urodele taxa: Chelotriton ogygius (Goldfuss, 1831), cf. Triturus sp. and Megalotriton filholi Zittel, 1890. Posteriorly, in the nowadays lost online database, Lisanfos KMS Martín and Sanchiz revised the assemblage and identified Chelotriton sp., Salamandra sp. and Pleudelinae indet (Peñalver et al. 2016). In addition, Martín and Sanchiz reported the presence of Ranidae indet. in La Rinconada (see Peñalver et al. (2016)). Although there are no direct fossils of urodeles, the other Konservat-Lagerstätte locality, San Chils, has yielded coprolites that could have been produced by adult and larval urodeles (Álvarez Parra and Peñalver 2019).

Although both sites have been considered slightly older than the outcrop studied in the present work (Peñalver et al. (2016); Álvarez Parra and Peñalver (2019)), its exact chronology is not fully known. According to Schoch et al. (2015), Chelotriton paradoxus Pomel, 1853 has a predominantly aquatic lifestyle, based on taphonomical studies, based mainly on the large number of remains and the preservation of many more or less completely preserved adult specimens recovered from the late Oligocene Enspel locality and the Miocene Randeck Maar. However, Pogoda et al. (2020) proposed a more terrestrial ecology at least for Enspel Chelotriton, based on the similarities of the cranial morphology with the extant Eastern-Asian Echinotriton Nussbaum & Brodie, 1982. In this genus, the adults are fully terrestrial and females deposit egg clutches on land. The larvae develop in lentic waterbodies (Utsunomiya et al. 1978). Nowadays, Echinotriton inhabits peri-humid, broadleaved forests at low to moderate elevations (Hernández et al. 2017). The original attribution of some articulated premetamorphic specimens by Sanchiz (1977b) to Megalotriton filholi was questioned by Estes (1981), who tentatively proposed its attribution to Salamandra sansanensis Lartet, 1851. Thus, the posterior correction by Martin and Sanchiz to Salamandra sp. included a new taxon, a clearly terrestrial newt that breeds on land, which relies only on aquatic habitats during its larval stage (Dufresnes 2019). Regarding ranids, only adult remains are known from that locality (Peñalver et al. 2016). Therefore, the amphibian assemblage from La Rinconada is mainly composed of terrestrial taxa that only use waterbodies for breeding. In this sense, this assemblage is ecologically similar because we have described a clearly terrestrial anuran, as Bufonidae indet. and a urodelan taxon, cf. Lissotriton sp., in which extant species experienced a yearly terrestrial phase (Dufresnes 2019).

Bufonids are a common element in the present-day European batrachofauna. Three genera with six species currently inhabit the continent: the common toads Bufo Garsault, 1764; green toads Bufotes Rafinesque, 1815 and natterjack toads Epidalea Cope, 1864 (Dufresnes 2019). The earliest known bufonid remains come from the localities of Itaboraí (middle Palaeocene, Brazil) and Cernay (late Palaeocene; France) (Rage 2003). Fossil remains of Bufo and Bufotes are already present in the Early Miocene of Europe. For instance, Bufo sp. has been reported from Mokrá-Western Quarry, 2/2003 Reptile Joint (MN4; Czechia) (Ivanov 2008), Erkertshofen 1 and Sandelzhausen (MN4; Germany) (Böhme 2003, 2010). In Spain, records include Buñol (MN4), Córcoles and Valdemoros IIIb (MN4b) (Sanchiz (1977a); Alférez and Brea (1981)). The oldest record of Bufotes comes from the Vieux-Collonges site (MN4; France) (Bailon and Hossini 1990). Regarding Epidalea, it has been cited in the Manchones II and Manchones localities (Sanchiz 1977a), although Böhme and Ilg (2003) reassigned those remains to Bufotes. Therefore, the oldest unambiguous Epidalea comes from the Late Miocene (MN12) Spanish localities of Concud’98 (Epidalea aff. calamita) and Los Mansuetos (Epidalea cf. calamita) (Sanchiz (1977b, 1998b); Böhme and Ilg (2003)). Nowadays, only two species of toads are present in the Iberian Peninsula: Bufo spinosus Daudin, 1803 and Epidalea calamita (Laurenti, 1768). However, Bufotes was still present in eastern Spain until the late Early Pleistocene (Blain et al. 2010).

During the Early Miocene, Europe was inhabited by at most three genera of crocodylians: Diplocynodon and the gavialids Tomistoma Müller, 1846 and Gavialosuchus Toula & Kail, 1885 (Böhme and Ilg 2003). The latter is mostly found in marine or estuarine deposits (Buffetaut et al. 1984). Although various European Tomistoma species are mainly associated with such contexts, they have also been recorded in fluvial deposits (Antunes and Pais 1993). In contrast, Diplocynodon is considered a freshwater crocodile (Delfino and Rossi 2013). Given the lacustrine nature of the studied sites (Anadón et al. 2004) and the abundance of similar fossil records in coeval Iberian localities (Pereda Suberbiola et al. (2001); Böhme and Ilg (2003); Casanovas-Vilar et al. (2022); Palazuelos et al. (2025); Walter et al. (2026)), it is probable that the remains described here belong to an undetermined member of the genus Diplocynodon. Today, crocodiles are absent from the Mediterranean Basin, though they were present in the region until the early 20^th century, from Palestine to Morocco (Trape et al. 2012; Bentley 2022). The only species currently found in the Sahel-Sahara Region, the West-African crocodile (Crocodylus suchus Geoffroy, 1807), was extirpated due to a combination of palaeogeological events, climatic shifts and human pressure (De Smet 1998). In the Iberian Peninsula, crocodylians survived until the Early Pliocene (Castillo and Agustí (1996); Guerra-Merchán et al. (2013)).

The fossil record of gekkotans is generally poor (Augé 2005), although it is relatively well represented in Europe compared to other continents (Daza et al. 2014). The oldest record dates to the earliest Eocene (MP7), with Dollogekko dormaalensis Čerňanský et al., 2022 from the Belgian site of Dormaal (Čerňanský et al. 2022), and Gekkota indet. from the Portuguese locality of Silveirinha (Rage and Augé 2003). Unfortunately, most European Early Miocene gekkotans fossils consist of isolated and fragmentary elements, which prevent attribution below the order level (Čerňanský and Bauer (2010); Georgalis et al. (2021a)). As a result, the group’s evolution remains poorly understood (Augé 2005). During MN4 and early MN5, indeterminate gekkotans have been reported from the Iberian localities of Córcoles (Alférez and Brea 1981), els Casots (Casanovas-Vilar et al. 2022), Agramón (probably with two distinct forms), Artesilla, San Marcos, San Roque 3 and 5 and Vargas 1A (Böhme and Ilg 2003).

Diverse scincoids are known from various Palaeogene and Neogene European sites (Böhme and Ilg 2003). In the Ribesalbes-Alcora Basin, two forms have been identified, one potentially related to the Palaeogene genus Pyrenasaurus. If confirmed, Pyrenasaurus would qualify as a Lazarus taxon sensuJablonski (1986). The only known species, P. evansae, is known from two late Eocene sites in Spain and France (Bolet and Augé 2014). Its survival beyond the Grande Coupure—a major faunal turnover at the Eocene-Oligocene boundary (Stehlin 1909)—is significant, as many endemic taxa disappeared, while Asian immigrants colonised Europe (Hooker et al. 2004). Although traditionally viewed as a European phenomenon, similar turnovers have been reported in North America, Asia and Africa. The event likely correlates with the Oi-1 glaciation (~ 33.5 Ma), the largest Palaeogene cooling pulse (Costa et al. 2011). Lizard taxa that survived mostly vanished shortly after, replaced by Asian lineages (Rage 2013). A parallel case is the choristoderan Lazarussuchus Hecht, 1992. Until the discovery of that taxon choristoderans were considered as disappearing at the end of the Eocene. However, Lazarussuchus inexpectatus Hecht, 1992 was described from the late Oligocene of Armissan (France) and, later, further material of Lazarussuchus was recovered from deposits of the Oligocene and Miocene of Central Europe (Matsumoto et al. (2013); Kirscher et al. (2016)). The youngest known choristoderan comes from Hammerschmiede 5 (MN 9, Late Miocene, Germany; Kirscher et al. (2016)). The extinction of Lazarussuchus has been linked to the climatic deterioration throughout the Miocene, including increased seasonality and a decline in global temperatures (Evans and Klembara 2005). If P. evansae or a related form survived into the Miocene in the Ribesalbes-Alcora Basin, it could indicate ecological stability in the region. Supporting this, Crespo et al. (2020b) described a bat assemblage dominated by molossids—common in the early Oligocene of Western Europe, but rare by the Early Miocene—suggesting the persistence of tropical forests in the area.

Molecular data suggest that modern Chalcides began to diversify around 10 Ma, probably in the Moroccan Region (Carranza et al. 2008). However, fossils attributable to that genus have been reported from Eurasia in chronologies older than the proposed diversification. The biogeographical scenario of the dispersal of Chalcides into Europe was proposed by Carranza et al. (2008). Thus, they identified at least five major colonisation waves that colonised Europe. The first ones took place at about 5 Ma, coinciding with the Messinian Salinity Crisis (MSC), when the Mediterranean Sea was isolated from the Atlantic Ocean by the closure of the Betic and Rifian Strait at ~ 7.6 and ~ 6.7–6.2 My, respectively. As a result, the Mediterranean practically dried out and large evaporite deposits accumulated in its Basin (Booth-Rea et al. 2018). According to the chronologies proposed by Carranza et al. (2008), the dispersal of Chalcides in Europe was coincident with the last of the major dispersal waves in the Western Mediterranean, the named Gerbil Event (~ 5.9–5.3 My), when gerbils dispersed into Europe from northern Africa (Agustí et al. 2006). From the Iberian Peninsula, the clade colonised the Balearic Islands. Fossils of Chalcides have been reported from Na Burguesa-1 (Late Miocene/Early Pliocene, Majorca) (Bover et al. 2014) and from a currently lost site from Menorca (Boulenger in Bate (1909)). The posterior dispersal event may be transmarine from north-western Africa and it is supposed that the invasion of south-western Europe by Chalcides striatus (Cuvier, 1829) and Chalcides chalcides (Linnaeus, 1758) occurred at 2.6 and 1.4 Ma, respectively (Carranza et al. 2008).

The new material from the MCX3, BC1 and MAB3 sites draw a new possible biogeographical scenario. Chalcides evolved in Europe, where it is recorded by the oldest fossils of the genus and later dispersed to Asia and Africa. In this sense, a potential occurrence of the genus has been reported from the locality of Kavakdere (MP 25, Rupelian-Chattain boundary, late Oligocene; Edirne-Keşan, East Thrace, Turkey). However, this material is very fragmentary, as a consequence its attribution to Chalcides and even Scincidae is very tentative (Georgalis et al. 2021b). The only fossil species of the group, Chalcides augei Čerňanský et al., 2020 from the East Siberian Tagay locality (Baikal Lake, Russia), has been linked with the current Chalcides ocellatus (Forskål, 1775) on the basis of its tooth morphology (basically the tooth necks are slightly swollen lingually) and the frontal morphology and type of sculpture in its dorsal surface (Čerňanský et al. 2020a). Moreover, these characters are shared with the other fossil taxa of Central and Eastern Europe from the Early Miocene, such as cf. Chalcides sp. from Echzell (MN4, Germany) (Vasilyan et al. 2022) and C. augei from Oberdorf (MN4, Austria), which was originally described as an undetermined form of scincid by Čerňanský (2016) and posteriorly attributed to C. augei by Čerňanský et al. (2020a) and from the Middle and Late Miocene from Eastern Europe, as it is denoted by the described remains attributed to C. augei, previously described as Chalcides cf. ocellatus by Venczel and Hír (2013), from Felsötárkány 1 and 3/10 sites (Hungary) (Čerňanský et al. 2020a) and Chalcides sp. from Creţeşti-1 site (MN9, Romania) (Codrea et al. 2017). On its behalf, the record from the Iberian Peninsula and the Balearic Islands presents different tooth morphology; basically, these taxa presented teeth without enlarged crowns. Böhme and Ilg (2003) include remains attributed to Chalcides from the La Gloria 4 site (MN14, Early Pliocene); however, they are still neither described nor figured, precluding the corroboration or refutation of this taxonomical attribution and, therefore, it is not possible to include them in any of the two morphotypes. In this sense, in the Eurasian fossil record of Chalcides, it is recognisable two possible morphotypes: the first one includes the “expanded crown” taxa and it is located in Central and Eastern Europe and Central Asia, and the second one, limited to south-western Europe, presents “not-expanded crown” (Fig. 16). As of this date, the African fossil record of Chalcides is limited to the Quaternary (Bailon et al. 2017). Thus, palaeontological studies of northern Africa remain scarce, yet some remarks can be made. Scincids do not appear in the Moroccan palaeoherpetofaunal assemblages until the Middle Miocene. Rage (1976) studied the squamate remains from the Beni Mellal site. He described the presence of material attributable to the skink Eumeces Wiegmann, 1834. No scincid fossils from other areas of North Africa are known, except Scincidae indet. from the Egyptian site of Bir Tarwafi (Chibanian, Middle Pleistocene) (Szyndlar 1993).

The relationship of our material with current Chalcides species is unknown until the recovery of more complete material. The fossil record of these species is, until now, limited to Quaternary sites (Villa and Delfino 2019a). Under the proposed scenario, Chalcides must have dispersed into Africa at some point during the Miocene and, probably, disappeared from Eurasia. After that, ancestors of the current species of the genus re-colonised Europe, dispersed from North Africa and reached the Canary Islands at different events during the Late Miocene and Pleistocene (Carranza et al. 2008). Future molecular works combined with fossil data must shed light on the palaeogeography of the genus. Similar scenarios exist in which the fossil record contrasts with molecular data, indicating a recent and rapid re-colonisation of southern Europe. Examples include the gekkotans Tarentola Gray, 1825 and Hemidactylus Oken, 1817, as well as the freshwater turtles Mauremys leprosa (Schweigger, 1812) and Emys orbicularis (Linnaeus, 1758) (Stuckas et al. (2014); Veríssimo et al. (2016); Villa and Delfino (2019a)).

Europe is probably the centre of origin for lacertids (Arnold et al. (2007); Čerňanský et al. (2020b)), given that the basal fossil taxa closest to this family have been found on that continent (Borsuk-Bialynicka et al. (1999); Čerňanský and Auge (2013)). During the European Palaeogene, the lizard fauna was dominated by a mixture of lacertids and related primitive forms (Čerňanský et al. (2016); Čerňanský and Smith (2018); Čerňanský et al. (2020b)). Over time, the latter became extinct and lacertids emerged as the dominant group in the European herpetofauna. The European fossil record of lacertids includes forms related to extant species since the beginning of the Neogene. In this context, the oldest representative of an extant genus (Lacerta Linnaeus, 1758) dates from the Early Miocene (MN 4b) of the Czech locality Dolnice, with the description of remains attributable to the green lizard group (Lacerta viridis group) (Roček 1984; Čerňanský 2010b). Recently, Marquina-Blasco et al. (2025) described remains tentatively attributable to the genus Timon Tschudi, 1836, in the locality of Crevillente 2 (MN11, Late Miocene; Crevillente, Alicante, Spain). Many extant genera are already cited in the fossil record from the Early Pliocene (Villa and Delfino 2019b). Accordingly, possible remains of the genus Podarcis have been described from the Early Pliocene of Poland and the Late Pliocene of Germany (Böhme and Ilg 2003); however, the attribution of these specimens remains uncertain. Regarding Psammodromus, its entire fossil record is limited to the Iberian Pleistocene (Böhme and Ilg 2003). Therefore, confirmation that the studied material belongs to either of the two extant genera (Psammodromus or Podarcis) would constitute their oldest known record.

Čerňanský et al. (2015) suggested that chamaeleonids reached Europe during the MN3 biozone. Based on this and the recently described putative remains of chamaeleonids in Anatolia, chamaeleons reached Europe through Anatolia, probably via the Gomphotherium Land Bridge (Georgalis et al. 2016, 2023b), which connected Europe with Arabia and Asia approximately between 19 and 14 Ma ago, with several interruptions before closing definitively (Harzhauser et al. (2007); Bialik et al. (2019)). The oldest Anatolian record, at the Sabuncubeli site (MN3, dated at approximately 20 Ma), described by Georgalis et al. (2023b), is chronologically similar to the oldest European record by the report of Chamaeleo caroliquarti Moody & Roček, 1980, at the German site of Wintershof West (Moody and Roček 1980) and Chamaeleonidae indet. in the Czech locality of Merkur-North (Čerňanský 2010b) and the Spanish Turó de les Forques 1 (Bolet et al. 2022). All these localities have been correlated to MN3a (Daams and Freudenthal (1981); Ivanov (2002); Casanovas-Vilar et al. (2016)). Thus, the colonisation of Europe by chamaeleonids may have been very fast from the Anatolia Peninsula (Georgalis et al. 2023b), although the apparently limited dispersal capacity of the different members of that clade (Cuadrado 2001) and it coincided with the Miocene Climatic Optimum (Georgalis et al. 2016). At the Iberian context, all the regional records are younger than Turó de les Forques 1 and located geographically further south than the first one. This indicates a probable colonisation of the Iberian Peninsula via a pass to the east of the Pyrenees. Thus, Chamaeleo aff. caroliquarti from Agramón site is dated as MN3/4 by Böhme and Ilg (2003) or MN4 by Van der Meulen in Szyndlar and Alférez (2005). These remains were considered lost (Georgalis et al. 2016), until they were recently located, allowing the presence of chameleons in this association to be corroborated (A. Villa, pers. comm.). The described remains in the present work, recovered from CBR0B, have been correlated with the local biozone Ligerimys ellipticus of the latest MN4 (or earliest MN5, after Casanovas-Vilar et al. (2021); latest Early Miocene), due to the co-ocurrence of the cricetids Democricetodon Fahlbusch, 1964 and Megacricetodon Fahlbusch, 1964, the presence, but in low numbers, of the eomyid L. ellipticus and the occurrence of the insectivore Heterosorex Hutchison, 1966 (Crespo et al. 2019c, 2025).

Both anguid genera described in the Basin, Ophisaurus s.l. and Pseudopus, are cited in several Iberian localities in the Early Miocene. Thus, Ophisaurus s.l. is recorded in Iberia at the localities of Agramón, Barranco de Tudela 1 and 3, La Nasa 1 (Trapiazul), Rincón del Bu, Cetina de Aragón, els Casots, Olmo Redondo 4A, San Roque 2 and 3, Vargas 4A, Villafeliche 2A, La Retama and the Loranca Basin (Murelaga et al. (2002); Böhme and Ilg (2003); Blain and Bailon (2019)). Regarding Pseudopus, the fossil record of the genus is mainly restricted to its north-eastern region (Blain et al. 2016). Until now, only two records have been reported outside this area: Pseudopus sp. from the Late Miocene site of Crevillente 2 (Marquina-Blasco et al. 2025) and cf. Pseudopus sp. from Zújar site in the Guadix-Baza Basin (Bailon 1991), although the latter is considered questionable (Marquina-Blasco et al. 2025). Therefore, the new material presented here not only represents a novel record for southern Iberia, but also constitutes the southernmost known occurrence of Pseudopus in the Early Miocene of Western Europe.

The fossil record of blind snakes is rather scarce. The oldest record of a scolecophidian, Boipeba tayasuensisFachini et al. 2020, comes from Late Cretaceous sediments in Brazil (Fachini et al. 2020); however, Head et al. (2022) considered it as a stem scolecophidian or a proximal sister taxon to the snake crown. In Europe, the group does not appear until the earliest Eocene (MP7) of Belgium. From the latest Eocene (MP19) to the earliest Early Miocene (MN1-2), blind snake remains are apparently absent, with the exception of still unpublished remains from La Colombière, France (MP30; Late Oligocene) (Rage and Bailon (2005); Smith and Georgalis (2022)). Scolecophidians re-appear in Europe during the Early Miocene (MN4) at, for example, Dolnice, Czechia (Szyndlar 1987); Béon-1, France (Rage and Bailon 2005); and Córcoles (Alférez and Brea 1981). Therefore, the discovery of an isolated vertebra in CBR0B is a new record at this stage. During the rest of the Miocene and the Pliocene, scolecophidians have been reported from some localities of southern Europe (Mead 2013). Regarding the Iberian Peninsula, the group seems to disappear during the Early Pleistocene (upper Villanyian, MN17) (Bailon and Blain (2007); Blain et al. (2016)). At present, Xerotyphlops vermicularis (Merrem, 1840) is the only autochthonous European species of blind snakes, which inhabits the southern regions of the Balkans. In the latter years, the Brahminy blind snake, Indotyphlops braminus (Daudin, 1803) has been introduced in soil imported with potted plants in a few localities of Europe: Aguadulce (Almería) and Majorca (Balearic Islands) both in Spain, Malta and Ischia Island (southern Italy), as well as in other regions of the Mediterranean Basin, such as north-western Libya, Egypt (around Cairo and in the Sinai) and Macaronesia, Gran Canaria (Canary Islands, Spain) and Madeira (Portugal) (Geniez (2018); Paolino et al. (2019)).

During the Early to Middle Miocene, the European colubrid fauna was dominated by natricids—often referred to as the “Age of Natricinae”—which persisted in Western Europe until MN6 (Rage and Bailon (2005); Szyndlar (2012); Ivanov et al. (2025)). Amongst the four genera described in this chronology, Natrix, Neonatrix, Wintershofia and Palaeonatrix, only the last two seem to have disappeared at the Middle Miocene (Ivanov et al. (2025); Marquina-Blasco et al. (2025)). Regarding the Iberian Peninsula, the fossil record of this clade from the Miocene is scarce and has been described only sporadically in literature. Thus, during the Early Miocene, an undetermined form of Natricidae has been reported from the Agramón site (Szyndlar and Alférez 2005). In this sense, the finding of Palaeonatrix remains in MAB11 not only represents the first record of the taxon in the Iberian Peninsula, but also enhances our understanding of the group’s history in the region.

The attribution of seven vertebrae to Viperidae from the sites CBR0D, MAB3 and MAB11 is putative, although they are of great interest. These snakes, together with Colubridae, are the current dominant element of the European ophidiofauna. The oldest record of the group consists of isolated fangs from the German site of Oppenheim/Nierstein (MN2, Early Miocene) (Kuch et al. 2006). Burdigalian was a key moment in the composition of the snake palaeocommunities. During this time, the ophidian faunas of MN 2 and early MN 3 were markedly different from those of late MN 3 and MN 4 (Ivanov 2022). For most of the Paleogene, Constrictores (sensuGeorgalis and Smith (2020)) were the dominant element of the ophidiofauna (Smith and Georgalis 2022), though basal caenophidians were scarce, but still present (Georgalis and Jones 2025). Gradually, this old fauna was displaced by modern snake lineages, specifically the families Colubridae, Viperidae and Elapidae, during the Early Miocene (Ivanov 2002). However, this replacement finishes in a dramatic event named “MN-4-event”, when this archaic fauna was “literally flooded” by the mentioned modern ophidiofauna in Central and Western Europe (Szyndlar 2012). This event has been linked to climatic change from cool, arid conditions to warm, humid ones. Regarding the Iberian Peninsula, the composition of the snake fauna was similar to that of the rest of Europe (Szyndlar and Alférez 2005). Thus, the absence of ancient ophidian taxa indicates that the studied sites belonged to the late MN4, which is in accordance with the biochronological framework proposed by Crespo et al. (2019c), based on rodent assemblages from different localities in the Basin.