Cteniogenys, Gilmore, 1928

Guillaume, Alexandre R. D., Puértolas-Pascual, Eduardo & Moreno-Azanza, Miguel, 2025, Revisiting the choristodere and stem-lepidosaur specimens of the Guimarota Beds (Kimmeridgian, Portugal): taxonomic implications, Acta Palaeontologica Polonica 70 (1), pp. 77-96 : 80-89

Cteniogenys aff. Cteniogenys antiquus Gilmore, 1928

Figs. 2 , 3 .

1973 Cteniogenys reedi ; Seiffert 1973: 14, text-figs. 3–8.

1983 Cteniogenys reedi Seiffert, 1973 ; Estes 1983: nomen dubium.

Material. —One premaxilla, MG 27600; 2 maxillae, MG 27417, 27419; 8 highly fragmented specimens; 18 dentaries, MG 27605, 27609, 27627, 27628, 27638, 27646, 27652, 27653, 27662, 27664, 27666, 28761, 28826, 28831, 28843, 28882, 28901, 28914; 33 specimen holders with highly fragmented specimens; 2 surangulars, MG 28420, 28423; 5 vertebra, MG 28284, 28287, 28512, 28514, 28552; 2 humeri, MG 28890, 28928; 5 ilia, MG 28282, 28288–28791; block of associated bones, MG 28792, vertebra, frontal, and putative jugal and postorbitofrontal ( Evans 1991b). All from Guimarota beds (District of Leiria); Alcobaça Formation (Kimmeridgian), Upper Jurassic of Portugal ( Seiffert 1973; Evans 1990).

Description.— Premaxilla ( Fig. 2A ): Only one specimen, a left premaxilla, was identified, but the very short anteromedial sutured border suggests it was paired in the live animal. The border surface is rather smooth, with two small pits ( Fig. 2A 2 ). The dorsal process indents the anterodorsal margin of the external nares ( Fig. 2A 1 , A 2 ). It is anteroposteriorly elongated, projecting strongly with an acute posterior apex. Although posteriorly broken, it extends up to 2.5 tooth loci ( Fig. 2A 3 ). Its smooth dorsal margin is bevelled and diverges medially from the anterior midline of the bone. The alveolar border displays three tooth loci ( Fig. 2A 3 ). Locus 1 and 3 preserved the root and part of the crown. The teeth are subcircular, subthecodont with the labial wall higher than the lingual wall. Locus 2 is empty and seems larger than the other ones. Although only three tooth loci are preserved, the bone seems to be almost complete. Indeed, the premaxilla displays a strong posterior suture to articulate with the maxilla, as well as a deep, short lateral facet ( Fig. 2A 1 ). The palatal shelf is not preserved. In lateral view, the premaxilla is relatively low, without anterior extension. Instead, it is subvertical, facing anteroventrally ( Fig. 2A 1 ). Two anterior sensory foramina are preserved, one dorsal and one ventral, followed posteriorly by fainter foramina organized in line.

Maxilla ( Fig. 2B, C ): Maxillary elements are mostly fragments from the posterior ramus. The maxilla is a long, slender bone with at least 21 teeth based on MG 27419 Fig. 2C ). The dorsal process is flabellate, like a crest, with a dorsomedial facet ornamented by striae ( Fig. 2C 1 ). The labial surface displays heavy ornamentations similar to the dentary ( Fig. 2B 1 ): nutritive foramina extending into short anteroposterior striae. The anterior process is not preserved in MG 27419, despite what was previously illustrated Seiffert 1973), and could not be identified among other fragments. The bone tapers posteriorly into a long, low orbital process, forming a dorsoventrally compressed ramus, as a plate ( Fig. 2C 1 ). It displays a very shallow posterodorsal facet, which extends across the full width. The alveolar bor- der exhibits a broad, elongated medial facet for the vomer Fig. 2C 1 ), with a large gap between vomerine and palatine facets to form the internal nares. It is associated with a large foramen connected to a deep anteroposterior groove, posterior to the dorsal process, for the opening of the superior alveolar canal.

Dentary ( Fig. 2D, E ): Dentaries are by far the most common bone, although most of the specimens consist of highly fragmented small elements of which any assessment cannot be safely confirmed. The dentary is a long, slender, mostly straight bone. Due to the preservation of the specimens, the tooth row displays between 24 and 33 teeth. The symphyseal surface is posteriorly elongated, dorsally and ventrally encapsulating the Meckelian canal ( Fig. 2D 1, E 1 ). Its surface is rugose, but not ornamented. The facet for the splenial is visible, posteriorly to the symphysis ( Fig. 2E 1 ). Its anterior apex is acute. It participates ventrally, then medially, to the subdental shelf, with a bevelled rim. The facet keeps the Meckelian canal open dorsally and extends to the end of the tooth row based on MG 28914 ( Fig. 2E 1 ). The subdental ridge is strong and robust, more marked than the ventral part of the dentary. Its mediolateral surface is blunt anteriorly but becomes posteriorly a horizontal shelf which separates the coronoid facet to the ventral prearticular facet.

The Meckelian canal is open all along the dentary, although the opening becomes wider towards the posterior end of the tooth row ( Fig. 2D 1, E 1 ). MG 28914 does not display an angular facet, but a posterior alveolar opening on the ventral wall. The labial surface is highly ornamented with anteroposterior short striae extending nutritive foramina in a double line ( Fig. 2D 3 , E 2 ). The teeth are conical, with an acute apex, and subthecodont. They are striated mostly on their lingual surface, although this is restricted to the top of the crown. The striations are absent in MG 28901, of which teeth are more triangular with a blunt apex, but they are well pronounced in MG 28914. The teeth have a cylindric base, with a slight anteroposterior compression. The anteriormost teeth are slightly curved distally.

Surangular ( Fig. 2F ): Two elements could be identified as surangulars, with MG 27420 being still articulated with the posterior part of a dentary. It is a robust, elongated, subrectangular bone. The lateral surface is broad, with few foramina extending into short grooves in its anterior section. In MG 27423, the large articular facets have an acute posterior apex ( Fig. 2F 2 ). The dorsal facet is obscured in MG 27420 because of the dentary, and the ventral one could not be confirmed. The ventral margin displays a deep anteroposterior groove, the angular facet ( Fig. 2F 3 ). Its medial flange is shorter and can be seen in medial view. The posterior edge is acute to blunt, displaying a small process ( Fig. 2F 1 ).

The surangular exhibits a medially extended process, which is anteroposteriorly compressed to form a prominent prong. This process delimits a posterior, more rugose fossa for the articular ( Fig. 2F 1 ). Dorsally, it is connected to a strong wide ridge forming the dorsal border of the bone ( Fig. 2F 4 ). It participates in forming the thick, round adductor region and opens anteriorly into a deep slot with an acute posterior apex ( Fig. 2F 1 ). Ventrally, the process extends into a shallow ridge participating in the ventral border of the surangular with the angular facet ( Fig. 2F 1 ). It displays anteriorly a triangular facet for the prearticular, which is less marked than the dorsal slot.

Dorsal/sacral vertebra ( Fig. 3A, C ): MG 28512 and 28552 both possess a short amphicoelous to platycoelous centrum, with subcircular anterior and posterior articular surfaces. The notochordal canal can be seen in MG 28552 ( Fig. 3C 4 , C 5 ), suggesting it is an immature individual. Relatively square-shaped in dorsoventral view, the centrum is wider than long. Its dorsal surface is excavated on either side of strong median crest ( Fig. 3A 1, C 1 ). Although the excavation is clogged by coal in MG 28512, the dorsal ridge of the median crest, as well as the margin of the excavation, can be clearly seen.

Both vertebrae display large, raised lateral articular facets. These facets are roughly triangular, with the apex projecting laterally. In MG 28512, it dips lateroventrally, with the apex more ventral than the base ( Fig. 3A 2 ). The facets are immediately posterior to the anterior margin of the centrum, being separated by a shallow groove. Neither ventral surface exhibits a sharp, midventral keel, as they are rather convex. However, MG 28512 is smooth and broad, which would suggest it might be more likely dorsal than sacral ( Fig. 3A 3 ); while MG 28552 is smooth, but narrower, almost constricted ( Fig. 3C 3 ), suggesting it might be sacral or transitional. However, the lack of material to compare with, and the fact that MG 28552 seems to be more immature, prevent us from confirming the exact identification.

Caudal vertebra ( Fig. 3B, D ): Only three caudal vertebrae could be safely identified. The morphology of MG 28514 ( Fig. 3B ) suggests it is an anterior caudal vertebra. It displays a short amphicoelous to platycoelous centrum, relative square-shaped in dorsal and ventral views ( Fig. 3B 1 , B 3 ). The opening of the notochordal canal is visible ( Fig. 3B 4 , B 5 ), suggesting it is an immature specimen. The anterior and posterior articular surfaces are subcircular. The dorsal surface of the centrum is excavated on either side of a strong median crest ( Fig. 3B 1 ); while the ventral surface displays a large, deep midventral groove with paired ventral flanges ( Fig. 3B 3 ). The morphology of MG 28284 and 28287 suggests they were caudal and more posterior than MG 28514, maybe mid-caudal. They share an amphicoelous to platycoelous subcircular centrum, slightly compressed laterally ( Fig. 3D 2 ). They display a deep ventral groove flanked by two lateral flanges extending anteroposteriorly ( Fig. 3D 2 ). They are preserved with the neural arch ( Fig. 3D 1 ), but the dorsal surface is obscured by sediments. The neural arch exhibits a basal longitudinal thin crest. The neural spine and the zygapophyses are not preserved.

Humerus ( Fig. 3E ): Two left humeri could be identified, although they are not complete. Based on overlapping material, the humerus displayed a broad proximal and distal head for a slender shaft that looks dense in cross-section and rotates around 90°. The proximal head is anteroposteriorly flattened. It extends on the proximal ridge as a convex bar, with a faint condyle in the middle. The ectotuberosity is well marked on the radial border ( Fig. 3E 1 ). It is separated from the head by a long proximodistal crest. The mediodorsal border displays a small peg, as well as a long groove on the dorsal surface where it joins the shaft in MG 28928. The fossa is well marked. In medial view, along the radial border and distal to a shallow surface, a small dorsoradial rough area can be observed ( Fig. 3E 2 ).

The distal head is mediolaterally flattened and flares dorsoventrally, although the epicondyle is poorly developed as no clear differentiation of articular facet can be observed. The distal head also exhibits a deep but narrow ectepicondylar groove in dorsal view. The lateral supinator region barely extends, and is almost not separated from the rest of the bone. There is no sign of an entepicondylar groove. Both ventral and dorsal surfaces are smooth. The latter displays a shallow fossa, while the former displays a fossa which has a strong anterior rim and becomes fainter posteriorly.

Ilium ( Fig. 3F ): Only the ventralmost part of the ilia has been preserved. The acetabular region, fully preserved only in MG 28289, is mediolaterally flattened. The anterior and posterior margins flare out, making a triangular general outline. The ventral border is wide V-shaped, highly flattened and obtuse, with a bulbous apex. Both ischial and pubic articular facets are preserved, the former being hooked in MG 28289 (seemingly more mature than other specimens). The acetabulum is subovoid to triangular, rather shallow, no matter the size of the specimens, but with a strong rim in MG 28289 ( Fig. 3F 2 ).

The supraacetabular ridge thickens in the middle, creating a distinct boss which merges with the ridge running along the iliac shaft; and is bordered ventrally in MG 28282 by a round cavity that may yield a small foramen. The ilium narrows dorsally into a posteriorly curved blade, with a lateral shallow ridge running along. There is no sign of a dorsal tuberosity. The iliac shaft is mediolaterally flattened with a sharp posterior edge and a blunter anterior edge. Its width remains similar in all the sections, but the dorsalmost part is not preserved. It bears a medial facet ( Fig. 3E 1 ). However, its ventral margin can be either convex, making it like a strip, or concave, making it circular.

Remarks.— The dentaries here described display the typical morphology and ornamentation previously observed in the Cteniogenys sp. material from Kirtlington in the UK ( Evans 1989, 1990), as well as in the material found in the Morrison Formation ( Chure and Evans 1998; Foster and Trujillo 2000; Foster et al. 2020) and to the type species Cteniogenys antiquus ( Gilmore 1928; Evans 1989), based on holotype and referred material currently housed in the Smithsonian National Museum of Natural History and the Yale Peabody Museum USA; ARDG, personal observation 2022). They display the characteristic double row of nutritive foramina extended in striae on the labial surface, visible splenial facets tapered anteriorly, the medially directed symphysis, the broad conical teeth with subthecodont implantation—the base of each tooth is sitting in a shallow, rounded socket, but there is not root—and striae restricted to the apical part of the lingual surface of the tooth crown ( Seiffert 1973; Evans 1989, 1990).

In his original work, Seiffert (1973) used the depth of the subdental ridge and the compression of the tooth crowns to distinguish the Guimarota taxon as C. antiquus Gilmore, 1928 . However, it was later argued that the designated holotype MG 27628 (Gui A. 33 in the original publication, see Fig. 2D ) does not exhibit any significant difference in depth, neither does it preserve any tooth crown ( Estes 1983); and being relatively small, it could belong to a juvenile ( Evans 1989). Our examination of the Portuguese material, as well as holotype and referred material of C. antiquus , confirmed the limitation of the dentary to distinguish the species.

Nevertheless, previous work on the material from Guimarota and Kirtlington ( Seiffert 1973; Evans 1989, 1990, 1991b) allows us identification of more elements with affinity between both localities, suggesting they were indeed sharing this faunal component. Notably, Seiffert (1973) could identify premaxillae and maxillae aside from the dentaries, although some of the premaxillae were later referred to the lepidosauromorph Marmoretta ( Evans 1991a; Griffiths et al. 2021). Although the attribution of the maxilla MG 27419 to Cteniogenys (Gui L. 234 in the original publication, see Fig. 2C ) was questioned ( Evans 1989), the other maxillae ( MG 27417, and other fragmentary specimens) described in this work could confirm the original assessment of Seiffert (1973) regarding MG 27419: the maxilla displays the characteristic prominent dorsal process, confined to the anterior part of the maxilla, and slightly inrolled dorsally Evans 1989, 1990; Gao and Fox 1998).

No complete maxillae were found associated with C. antiquus material in North America, and fragments were only referred to Cteniogenys sp. ( ARDG personal observation, 2022). None seem to have preserved the flabellate dorsal process, but the material would require a deeper revision to confirm this statement as the isolated specimens are still embeded in their matrix. Nevertheless, their robustness, their ornamentation, and their tooth implantation match the European material.

Only one premaxilla ( MG 27600) assignment to Cteniogenys could be safely confirmed, based on its differences with other premaxillae attributed to Marmoretta and the presence of the dorsal process observed in the Cteniogenys sp. from Kirtlington ( Evans 1990, 1991a). However, the Portuguese material slightly differs from the English material: the dorsal process is more acute, projecting not only further posteriorly (2.5 tooth loci instead of 1.5) but also more posteriorly rather than dorsally. The alveolar border from the Guimarota specimen is similar to the one observed in Kirtlington ( Evans 1990: fig. 5A), although it preserves only three tooth loci instead of four and seems rather complete, suggesting it bore fewer teeth than the Kirtlington specimen. Finally, the Kirtlington premaxillae display a small anterior extension to the dorsal process, that can form a small ventral sickle ( Evans 1990: fig. 5C, D); but that feature is not observed in Guimarota specimen. The well-preserved anterior section of the bone is instead subvertical, facing anteroventrally.

In addition to these bones, here we report for the first time the presence of surangulars, isolated vertebrae, humeri, and ilia. All of them display strong affinities with the material from Kirtlington. Surangulars were never reported previously, and they were actually labelled as Albanerpetontidae in the Guimarota collection. Albanerpetontids are an enigmatic group of extinct lissamphibians ( Fox and Naylor 1982), but the surangulars from Guimarota do not resemble to any bone elements known in this group ( Guillaume et al. 2023b), while their morphology is strikingly similar to the surangulars from Kirtlington ( Evans 1990).

The vertebrae MG 28512 and 28514 were also labelled as Albanerpetontidae . However, they display the dorsal excavations in either side of the median crest that is absent in albanerpetontids, and albanerpetontid vertebrae are hourglass shaped ( Guillaume et al. 2023b). On the contrary, the square overall morphology, the dorsal excavations in either side of the median crest, the morphology of the articular facet, and the presence of facets to articulate with the rib match what has been described in dorsal and sacral vertebrae from Kirtlington ( Evans 1991b). MG 28552 was only labelled as Amphibia, but the square overall morphology, the dorsal excavations in either side of the median crest, the morphology of the articular facet, and ventral groove flanked by paired flanges also fit the description of caudal vertebrae from Kirtlington ( Evans 1991b). The vertebra MG 28287 was labelled as “Anura? or Cteniogenys ?” but was not reported by Seiffert (1973). It closely resembles the vertebra MG 28284, labelled Cteniogenys sp. , although the latter was not reported by Seiffert (1973) either. Furthermore, MG 28287 resembles other mid-caudal vertebrae from Kirtlington ( Evans 1991b), and it is seemingly more elongated than MG 28514.

The humerus MG 28890 is labelled as Cteniogenys sp. but not mentioned by Seiffert (1973); while the humerus MG 28928 was labelled as Squamata and neither referred in Seiffert work (1973). Regarding the ilia, only MG 28289 and 28290 were labelled as Cteniogenys sp. , while MG 28291 was labelled as Anura, and MG 28282 and 28288 as “Anura? or Cteniogenys ?”; but none was mentioned nor reported previously. However, none of the five ilia displays the dorsal tuberosity characteristic of anurans. Rather, their morphology shares strong similarities with other choristoderes, especially Cteniogenys ( Evans 1991b; Matsumoto et al. 2009). Furthermore, even if the ilia from the Guimarota beds exhibit different stages of preservation, they can be assigned to the same taxon.

However, some differences can be noted between the Portuguese and the English material, which would suggest there were several Cteniogenys species. The surangulars from the Guimarota beds display a small process on the posterior edge of the ventral margin which was not reported in the Kirtlington specimens ( Evans 1990). Also, the bones from Guimarota seem longer than the ones from Kirtlington, although this could be due to a preservation bias in the English material. The ectotuberosity in the proximal head of the humerus appears relatively lower than the one observed in Kirtlington specimens. Furthermore, the ectotuberosity in Kirtlington is separated from the humeral head by a groove ( Evans 1991b), not a crest as observed in Guimarota. The humeri from Kirtlington also display a proximal roughened region ( Evans 1991b), but this one is absent in Guimarota specimens. Finally, the ectepicondylar grove is less marked in the Portuguese specimens than in the English ones ( Evans 1991b).

Previous statements regarding the validity of the Guimarota species were correct ( Estes 1983; Evans 1989; Chure and Evans 1998), as the dentaries alone do not allow differentiation of species from North America and Europe. A deeper revision of the material from the type locality of C. antiquus (Quarry 9, Como Bluff, Wyoming), as well as a more complete redescription of the type species with associated material, is needed to confirm, thus preventing us from erecting a new species or supporting previous taxonomic claims of its condition as a new species ( Seiffert 1973). Nevertheless, our study confirms that although they are similar, the Portuguese and English species are different, hinting toward a higher diversity of early choristoderes in the Jurassic.

Lepidosauromorpha Gauthier et al., 1988

Genus Marmoretta Evans, 1991a

Type species: Marmoretta oxoniensis Evans, 1991a ; Old Cement Works Quarry , Kirtlington, ( Oxfordshire, England, UK); upper Bathonian .

Diagnosis.—Small lepidosauromorph; large upper and lower temporal fenestrae; premaxillae paired, each with deep posterolateral maxillary facet; specialized maxillary/ premaxillary overlap; small posteroventral process of the jugal; narrow fused frontals; fused parietal forming a broad parietal table, parietal foramen absent, large midline sagittal crest; dorsoventrally wide posterior (squamosal) process of the postorbital that overlaps on to a broad shallow facet on the squamosal; palatine with small teeth that decrease in size medially from a larger row along the medial choana margin to smaller scattered teeth on the ventral surface; pterygoids bear three rows of teeth which radiate anteriorly; long and slender dentary with subpleurodont teeth; coronoid with prominent coronoid process having a smooth concave posterior surface that emerges through the lower temporal fenestra (from Griffiths et al. 2021).

Stratigraphic and geographic range. —Kilmaluag Fm from the Isle of Skye (Bathonian), Middle Jurassic of Scotland ( Waldman and Evans 1994; Griffiths et al. 2021); Kirtlington Old Cement Works and Leigh Delamere (all Bathonian), Middle Jurassic of England, UK ( Evans 1991a; Evans and Milner 1994; Evans 1998a); and Guimarota beds (Kimmeridgian), Upper Jurassic of Portugal ( Evans 1991a).

Marmoretta drescherae sp. nov.

Figs. 4 , 5 .

ZooBank LSID: urn:lsid:zoobank.org:act:.

Etymology: In honour of Ellen Eggert (born Drescher), the German preparator from the IPFUB who oversaw and prepared the thousands of specimens from the Guimarota beds. Her tremendous and priceless work allowed and keeps allowing researchers to study one of the most important parts of the Portuguese palaeontological heritage.

Type material: Holotype: MG28841 , right maxilla ( Fig. 4B ) . Paratypes: 2 premaxillae, MG27588 , 27644 ; 6 maxillae, MG28771 , 28816 , 28842 , 28907 , 28908 , 28925 ; 3 dentaries, MG28785 , 28794 , 28910 ; 4 frontals, MG28740–28743 .

Type locality: Guimarota mine, Leiria municipality, central Portugal . Type horizon: Guimarota beds; Alcobaça Formation , Kimmeridgian, Upper Jurassic .

Material. — Type material and one premaxillae, MG27593 ; 6 maxillae, MG28751 , 28757 , 28798 , 28815 , 28887 , 28888 ; 50 dentaries, MG27626 , 27633 , 27641 , 27643 , 27659 , 27665 , 27667 , 28763 , 28778 , 28779 , 28784 , 28791 , 28797 , 28803 , 28828 , 28833 , 28857 , 28874 , 28905 , 28942 , and 30 specimen holders and sample boxes with highly fragmented specimens. All from the type locality and horizon.

Diagnosis.—Small lepidosauromorph differing from Marmoretta oxoniensis by having highly constricted margins of the maxillary facet in the paired premaxillae; a more curved anterior rim of the dorsal process in the maxilla, making it rise more sharply; a lacrimal facet not extending dorsally on the dorsal process of the maxilla; maxillary teeth without apicolingual curvature; postfrontal facets not meeting posterodorsally in the fused frontals.

Description.— Premaxilla ( Fig. 4A ): None of the three premaxillae is complete, as the specimens from the Guimarota beds only preserved the anterior part. The alveolar tooth bearing plate is curved and formed by the alveolar border and crest. The premaxilla did not bear more than five teeth. Although none is preserved, the alveolar border suggests they were pleurodont, being attached via the labial wall only by the root, and were subcircular, being mesiodistally constricted ( Fig. 4A 1 ). MG 27644 does not preserve the maxillary facet. In other specimens, the facet is located in the posterolateral region of the premaxilla and is anteroposteriorly elongated, with an acute anterior apex ( Fig. 4A 2 ). It is ventrally deep and V-shaped, with margins highly constricted, making it look like a groove or a slot. In MG 27588 Fig. 4A ), the facet exhibits a shallow, small anterodorsal emargination which is not present in other specimens. The narial process extends posteriorly from the anteromedial part of the bone and participates in the alveolar tooth bearing plate, but it is poorly preserved ( Fig. 4A ). None of the specimens preserves the anterodorsal process (or nasal process sensu Evans 1991b).

Maxilla ( Fig. 4B ): The maxilla is a long, slender bone, tapered posteriorly. It is overall quite gracile, but its anterior section appears more robust. It is mediolaterally constricted, forming a long crest in dorsal view. The long, bulked anterior process is anteriorly acute with a round apex ( Fig. 4B 2 , B 4 ). It is best preserved in MG 28841 and 28908, and partially in MG 28771, 28816, 28907, and 28925; although, based on the shape of the maxillary facets observed in the premaxillae, it seems that no specimens fully preserved the anterior tip. The pars dentalis anteriorly extends up to 1/3 of the process ( Fig. 4B 4 ). MG 28841 and 28908 do not display a clear premaxillary facet, but MG 28925 displays a narrow slot, although its medial surface is convex. The dorsal surface of the anterior process is rather smooth, with faint anteroposterior ridges extending to the dorsal process. The dorsal margin itself is a thin crest ( Fig. 4B 1 ).

The tip of the dorsal process (or facial process sensu Evans 1991b) is preserved in none of the Portuguese specimens, as it is always dorsally broken ( Fig. 4B 2 , B 4 ). Therefore, the development of the apex is unknown. Only the ventral section has been preserved in some specimens, allowing some overlapping material for that part. The base is rather large and its unparallel anterior and posterior edges suggest the dorsal process was either triangular or trapezoidal, but not much developed. Also, it does not seem to be medially inclined ( Fig. 4B 1 ). The anterior rim is thicker than the posterior rim, which looks more like a crest. The anterior rim outline is concave, being extremely marked in MG 28816, 28841, and 28842. Medially, there is a small foramen at the base of the dorsal process. The lacrimal facet is not clearly marked: there is a deep medial depression posterior to the dorsal process, but it is ventrally restricted and does not extend dorsally.

The long posterodorsal jugal facet exhibits an acute anterior apex on the medial surface. The margin is well marked. It is located in the groove formed by the rim and the alveolar shelf. The palatine facet is well preserved only in the medial surface MG 28908. It displays an anterodorsal-posteroventral shallow groove with two facets on the alveolar shelf: the dorsal facet is rather short while the ventral one is longer. It is ventral to the posterior rim of the dorsal process and is pierced by two foramina in its dorsal facet. The lateral surface of the bone is rather smooth. MG 28842 and 28908 display occasional nutritive foramina. Posteriorly, the midline of the surface can be marked by a deep anteroposterior grove. MG 28841 exhibits a tooth row with 17 teeth but is not complete and is missing its anterior section. Teeth are tubular to conical, with an acute apex. They are pleurodont, being attached only via the labial wall only by the root. No apicolingual curvature could be observed. The enamel covers the entire crown and can be slightly apically striated on the lingual side.

Dentary ( Fig. 4C ): Most of the specimens are broken, although several preserved all the parts, making MG 28794 and 28910 the only ones seemingly complete and not broken. However, the former is embedded in coal matrix, and part of the anterior section of the latter is obscured. The dentary is a long, slender bone, relatively gracile. It deepens dorsoventrally in its posterior section. The labial surface is relatively smooth, although it can be covered by small nutritive foramina ( Fig. 4C 3 ). They are usually aligned, restricted to the anterodorsal part of the dental parapet. The anterior symphysis is small and short, forming a subvertical wall ( Fig. 4C 2 ). Lingually, the symphysis is divided in an upper and a lower rim by the Meckelian canal. The symphyseal surface is mostly flat, either in right or left specimens, although more rugose than the rest of the bone, and is U to V-shaped, the apex being more or less round. Also, the dentary is medially curved in that part ( Fig. 4C 1 ).

The lingual surface of the dentary is divided in two parts by the Meckelian canal: the dorsal subdental ridge and the ventral lip ( Fig. 4C 2 ). This canal is anteriorly widely opened at the symphysis and immediately posteriorly. Then, around the 4 th and 5 th tooth loci, the canal narrows drastically, to the point to form a thin groove or even contacting both the subdental ridges and the ventral lips, making a fold. However, some specimens can be dorsoventrally crushed, making one part looks like it overlaps the other. Posteriorly, the canal reopens broadly, with an acuminate apex. It broadens posteriorly up to the full height of the bone. Due to various preservations and intraspecific variation from one specimen to another, it is difficult to safely determine where this posterior opening starts, but it seems to be around seven or eight tooth loci before the end of the tooth row.

The tooth row is supported by the alveolar border. Posteriorly, where the Meckelian canal fully opens, it becomes a shelf. By the posterior end of the tooth row, it becomes weaker and bears a shallow facet for the coronoid ( Fig. 4C 1 , C 2 ). Teeth are tubular to conical, sharply pointed, and more apically elongated than in the maxilla. They are pleurodont, being attached via the taller labial wall only by the root; although, the implantation seems weak. The enamel covers most of the crown, with the apex being striated on the lingual surface. Based on specimens with tooth row the most complete, there were at least 19 teeth and up to 26.

Frontal ( Fig. 5 ): All frontal bones are fused, suggesting they are all from mature individuals ( Evans 1991a). MG 28740 is broken midway in two pieces, and MG 28743 did not preserve the anterior part. In dorsoventral view, their general outline is trapezoidal to subrectangular, with the posterior border being wider than the anterior border, especially in MG 28742 ( Fig. 5B ). The lateral margins are embayed for the orbits.

Despite having similar length, MG 28741 is broader than MG 28742, probably due to intraspecific variation. Only MG 28740 exhibits an ornamented dorsal surface, being sculptured by fine furrow and drumlin-shaped knobs. The fused frontals are anteroposteriorly convex in lateral views, being more marked in MG 28742 but with upturned lateral edges (medial orbital margin) stronger in MG 28741 ( Fig. 5A 3 , A 4 , B 3 , B 4 ). Unfortunately, the anterior border is not well preserved in any specimen. MG 28742 retains a short process that could be the median process ( Fig. 5B 1 , B 2 ), but no anterolateral process can be seen.

In dorsal view, the anterior section of the dorsal surface is rather flat and shallow, while the posterior section appears more bulbous in its centre. The postfrontal facets are at the posterolateral corners of the bone ( Fig. 5A 3 , A 4 , B 3 , B 4 ). They are restricted to the lateral border, with a lenticular to triangular outline, and face lateroventrally. The blunt ventral vertex, where the anterior ridge of the facet meets the ventral surface of the bone, can be located at the midpoint of the facet to slightly posteriorly ( MG 28741, Fig. 5A 3 , and A 4), or more anteriorly ( MG 28742, Fig. 5B 3 , B 4 and MG 28743, not figured).

The ventral margin is more marked than the dorsal. The posterior margin of the fused frontals is not well preserved, but the postfrontal facets do not seem to meet posterodorsally ( Fig. 5A 1 , A 2 , B 1 , B 2 ). The anterolateral prefrontal facets participate to the anterior section of the fused frontal lateral edges ( Fig. 5A 3 , A 4 , B 3 , B 4 ). They are quite deep and wide, forming a posteroventral right angle, with posterior and ventral margins strongly marked. The ventral surface of the fused frontals bears distinct paired anteroposterior crista cranii ( Fig. 5A 1, B 1 ), ventral extensions of the bone that act as dorsolateral walls for the olfactory tract. They lengthen the surface of the fused frontals’ lateral edges and participate in the orbits. They extend anteriorly after the lateral edges, and border the prefrontal facets ventrally. Both crista cranii form an hourglass outline, being lateromedially restricted at the centre of the bone. Their width is similar between all specimens. The ventral surface of the fused frontals displays a paired depression, separated by a ridge ( Fig. 5A 1, B 1 ). The ridge can be marked, as in MG 28742, or fainter, as in MG 28741 and 28743, in which the crista cranii are posteriorly stronger. These paired depressions might house the dorsal region of the brain and would indicate that both cerebral hemispheres would be well differentiated by a sagittal sulcus interhemispheric fissure sensu Barrios et al. 2023).

Remarks.—Thanks to the previous works on Marmoretta from the Middle Jurassic of the British Isles ( Evans 1991a; Waldman and Evans 1994; Griffiths et al. 2021), more substantial material could be identified from the Guimarota beds and used to confirm the presence of the genus in the Upper Jurassic of Portugal. Indeed, all bones identified share similarities with Marmoretta oxoniensis material from Kirtlington and the Isle of Skye. The premaxillae are paired and exhibited a deep posterolateral maxillary facet, the maxilla displayed the characteristic anterior process to articulate with the premaxilla, the frontal bones are fused and narrow with shallow hourglass crista cranii, and the dentaries are long and slender with subpleurodont teeth.

However, some differences suggest the material from the Guimarota beds belongs to another species, as previously suggested ( Evans 1991a; Evans and Milner 1994). Notably, the curvature of the anterior rim of the dorsal process in the maxilla in Marmoretta drescherae sp. nov. is more marked than in M. oxoniensis , where the rim rise smoothly instead ( Evans 1991a; Waldman and Evans 1994; Griffiths et al. 2021). Marmoretta drescherae sp. nov. does not exhibit a dorsally extended lacrimal facet in the maxilla as represented in the reconstruction based on the specimens from Kirtlington ( Evans 1991b: fig. 2D); and the maxillary teeth do not display apicolingual curvature as observed in the specimen from the Isle of Skye ( Griffiths et al. 2021).

Furthermore, the dorsal process itself is not as medially inclined as observed in the Isle of Skye NMS G1992.47.1a, although this has been suggested to be a postmortem deformation ( Griffiths et al. 2021). In the premaxilla, the characteristic maxillary facet exhibits margins more constricted than in M. oxoniensis , making the V-shaped facet narrower and acuter. Finally, the postfrontal facets do not meet posterodorsally as observed in the fused frontals from Kirtlington ( Evans 1991a). When the anterior ridge of the postfrontal facets meets the ventral surface of the bone, they form a blunt ventral vertex. This vertex can be located at the midpoint of the facet, as observed in the specimen from the Isle of Skye ( Griffiths et al. 2021) or in MG 28741; but can also appear more anterior as in MG 28742 and 28743. Unfortunately, the position of the vertex could not be confirmed in frontals from Kirtlington based on figured specimens, but it seems to be also located at the midpoint ( Evans 1991a). It remains unclear if these different positions of the ventral vertex represent some intraspecific variation or are due to a preservational bias.

Stratigraphic and geographic range. — Type horizon and locality only.

Figures

Fig. 2. Jaws of the choristoderan diapsid Cteniogenys aff. C. antiquus Gilmore, 1928, from the Guimarota beds (Guimarota mine, Leiria; Kimmeridgian). A. Left premaxilla MG27600, in labial (A1), lingual (A2), and ventral (A3) views. B. Right maxilla MG27417, in lateral (B1) and ventral (B2) views. C. Right maxilla MG27419 (Seiffert 1973: Gui L. 234, fig. 4), in lingual (C1) view. D. Right dentary MG27628 (holotype, Seiffert 1973: Gui A33, fig. 3), in lingual (D1), dorsal (D2), and labial (D3) views. E. Left dentary MG28514, in lingual (E1) and labial (E2) views. F. Left surangular MG27423, in medial (F1), lateral (F2), ventral (F3), and dorsal (F4) views.

Fig. 3. Axial and appendicular skeleton of the choristoderan diapsid Cteniogenys aff. C. antiquus Gilmore, 1928, from the Guimarota beds (Guimarota mine, Leiria; Kimmeridgian). A. Dorsal vertebra MG28512, in dorsal (A1), lateral (A2), ventral (A3), anterior (A4), and posterior (A5) views. B. Anterior caudal vertebra MG28514, in dorsal (B1), lateral (B2), ventral (B3), anterior (B4), and posterior (B5) views. C. Dorsal/sacral vertebra MG28552, in dorsal (C1), lateral (C2), ventral (C3), anterior (C4), and posterior (C5) views. D. Mid-caudal vertebra MG28287 in anterior (D1) and ventral (D2) views. E. Left humerus MG28928, in lateral (E1) and medial (E2) views. F. Right ilium MG28289, in medial (F1) and lateral (F2) views.

Fig. 1. Geographical and geological context of the vertebrate microfossil assemblages localities from the Alcobaça Formation. A. Map of Europe (modified from Erin Dill <fr.vecteezy.com>). B. Geological sketch of the Iberian Peninsula, showing the location of the study area (star). C. Geological sketch of the Leiria area, showing the location of the assemblage (modified from Teixeira et al. 1966). The entrance of the Guimarota mine, together with the outline of the pit and the approximate location of the fossil beds (blue area) are shown. Note that the fossil locality is located 30 meters below the surface, in the Upper Jurassic formations overlain by the Cretaceous. D. Stratigraphic position of the Guimarota beds and stratigraphic log of the Guimarota beds. General stratigraphic sketch derived from data in Teixeira et al. 1968. The colours for each stage correspond to the ones used in the map. Detailed stratigraphic log of the Alcobaça Formation from Schudack (2000a). Abbreviations: Apt–Cen, Aptian–Cenomanian; Kim–Tith, Kimmeridgian–Tithonian.

Fig. 4. Jaws of the lepidosauromorph diapsid Marmoretta drescherae sp. nov. from the Guimarota beds (Guimarota mine, Leiria; Kimmeridgian). A. Right premaxilla MG27588, in ventral (A1) and dorsal (A2) views. B. Right maxilla, holotype MG28841, in dorsal (B1), lateral (B2), ventral (B3), and medial views (B4). C. Right dentary MG28785, in dorsal (C1), medial (C2), and lateral views (C3); the specimen is broken, but both parts could be put together for the lateral and medial views.

Fig. 5. Frontals of the lepidosauromorph diapsid Marmoretta drescherae sp. nov. from the Guimarota beds (Guimarota mine, Leiria; Kimmeridgian). A. MG28741, in ventral (A1), dorsal (A2), lateral right (A3), and lateral left (A4) views. B. MG28742, in ventral (B1), dorsal (B2), lateral right (B3), and lateral left (B4) views.