One incomplete right humerus (MNHN.F.IBC1603; Fig. 1).

This humerus was regarded as a possible Ranidae in the first publication of the In Becetèn fauna (Broin et al. 1974). MNHN.F.IBC1603 was later illustrated and shortly described in 1984 (Rage 1984: fig. 1A) and considered to be a “Ranoid”. The specimen is the distal half of a right humerus (Fig. 1). Most of the distal head is preserved (Fig. 1). The diaphysis is straight and not curved ventrally (Fig. 1A-F). The crista ventralis is well developed (Fig. 1D, F). Most of the preserved region of the crista ventralis is divided into two ridges, with an elongate groove between the two; only the distalmost portion forms a single crest (i.e., no crista paraventralis; Fig. 1A, D). The crista medialis is thin (Fig. 1B, C, D, F). The fossa cubitalis ventralis is triangular and distally broader than its proximo-distal diameter (width greater than height; Fig. 1A, D). The fossa is shallow (although it seems deeper in Fig. 1A) but well delimited by a thin, distinct ridge (Fig. 1A, D). The eminentia capitata is well developed and not shifted laterally from the diaphysis axis (Fig. 1A, D). The epicondylus ulnaris is well prominent, reaching distally the level of the eminentia capitata (Fig. 1A, B, D, E), while the epicondylus radialis appears to be confluent with the eminentia (Fig. 1A, B, D, E). The olecranon scar extends proximally, with a pointed and tapered proximal margin (Fig. 1B, E).

This humerus is similar to the neobatrachian morphology in having: 1) a well-developed eminentia capitata; and 2) asymmetrical epicondyles (Prasad & Rage 2004; Rage et al. 2021). Furthermore, the crista ventralis is divided into two ridges in numerous neobatrachians (this may denote muscular insertion; Otero et al. 2014). Moreover, the presence of: 1) a straight diaphysis; 2) a shallow fossa cubitalis ventralis; 3) an eminentia capitata not shifted laterally; and 4) an epicondylus ulnaris more developed than the epicondylus radialis, is similar to ranoid morphology (Rage et al. 2013; Lapparent de Broin et al. 2020). In addition, a well-delimited fossa cubitalis ventralis, although rare (Rage 1984), is observed in several ranoid taxa from the Eocene of the Quercy Phosphorites (Rage 2016) and among extant ranoids (Worthy 2001). MNHN.F.IBC1603 is therefore similar to ranoids and can be attributed to the Ranoidea. The presence of a distinct crista medialis is often linked to sexual and/or ontogenetic dimorphism (Duellman & Trueb 1994). This suggests that this humerus might belong to a male ranoid. A humerus from the Cenomanian of the Kem Kem (Morocco) with ranoid affinities has also been recently described (UCRC-PV104; see Lemierre & Blackburn 2022: fig. 6A-C). Both humeri share: 1) a straight diaphysis; 2) a well-developed eminentia capitata; 3) asymmetrically developed epicondyles; 4) an eminentia capitata not shifted from the diaphysis axis; and 5) a shallow fossa cubitalis ventralis. However, MNHN.F.IBC1603 differs from UCRC-PV104 in: 1) having a more developed crista ventralis divided into two ridges; 2) a well-delimited fossa cubitalis; 3) a more protruding epicondylus ulnaris; 4) an olecranon scar more extended proximally; and 5) in having a discrete crista medialis. Thus, we think that these humeri likely represent different taxa. Furthermore, MNHN.F.IBC1603 possesses more ranoid characters than UCRC-PV104. In conclusion, MNHN.F.IBC1603 is attributed to a new unidentified male ranoid, and thus represent the oldest occurrence of Ranoidea.

22 fragmentary maxillae (MNHN.F. IBC1983, IBC1984 and IBC1985) and eight fragments of exocranium (MNHN.F.IBC1986) that include incomplete frontoparietal, nasals and squamosals.

Ornamentation on dermal bones in anurans occurs in various anuran clades (Paluh et al. 2020). However, combination of osteological characters can be here used to exclude certain major clades. Latonia ‌Meyer, 1843 (Alytidae ‌Fitzinger, 1843) is the only known alytid exhibiting dermal ornamentation (Roček 1994). However, its current stratigraphic range is limited only to Oligocene and Neogene of Europe and Northern Africa (Gardner & Rage 2016; Syromyatnikova et al. 2019). In addition, our ornamented taxon can be differentiated from Latonia in having a rounded medial margin of the lamina horizontalis (pointed in Latonia). Pelobatids are known to exhibit dermal ornamentation (Roček 2013) that resembles the ornamentation of the In Becetèn specimens (tubercles closely spaced together). However, our specimens differ from all pelobatids in lacking an elongate palatine process of the maxilla (Roček 2013). Furthermore, no pelobatids are known within the fossil record of Gondwana (a “Pelobatidae” was mentioned in the Upper Cretaceous of Madagascar, but the attribution is considered erroneous; Asher & Krause 1998; Gardner & Rage 2016) (Gardner & Rage 2016). The In Becetèn ornamented specimens can also be differentiated from Pipimorpha in: 1) having a well-developed squamosal (small and reduced squamosal in Pipimorpha); and 2) a rather thick maxilla (slender maxillae in Pipimorpha). Two anurans of indeterminate phylogenetic position from North America possess similar ornamentation: Scotiophryne ‌Estes, 1969 and Theatonius ‌Fox, 1976. Scotiophryne is known from the Late Cretaceous and might also occur in Early Cretaceous and Paleocene (Roček 2013; Gardner & DeMar 2013). Scotiophryne differs from our taxon in having a narrow horizontal lamina (Gardner 2008). Theatonius is known from Late Cretaceous deposits (Gardner & DeMar 2013). Theatonius differs from our taxon in lacking teeth on its maxillae. Thus, our taxon can be excluded from Leiopelmatoidea, Alytoidea, Pipimorpha, and Pelobatoidea. The only major clade left is the Neobatrachia. This includes several extinct and extant hyperossified taxa.

Two neobatrachians are known in the Mesozoic of Africa, Beelzebufo ‌ampinga ‌Evans, Jones and Krause, 2008 from the Maastrichtian of Madagascar and Cretadhefdaa ‌taouzensis ‌Lemierre & Blackburn, 2022 from the Cenomanian of Morocco. Beelzebufo ‌ampinga, Cretadhefdaa ‌taouzensis and our specimens share: 1) ornamentation extending on most of the lateral surface of the maxilla (Fig. 2B); and 2) the presence of a tectum supraorbitale (Fig. 2I-J). The latter is also shared with Cretadhefdaa ‌Lemierre & Blackburn, 2022. However, Beelzebufo does not possess a distinct processus frontalis (present in our taxon; Fig. 2F) and all known Beelzebufo specimens are far larger than our specimens. Our specimens can be differentiated from Cretadhefdaa in: 1) having an ornamentation made of small tubercles resembling a honeycomb pattern (deep pits and ridges in Cretadhefdaa); and 2) lacking ornamentation on the lateral surface of the pars dentalis (present in Cretadhefdaa; Lemierre & Blackburn 2022: fig. 3E-G).

Another Mesozoic hyperossified neobatrachian is Hungarobatrachus ‌szukacsiSzentesi & Venczel, 2010 from the Santonian of Hungary. Our specimens are similar to Hungarobatrachus in having: 1) ornamentation extending on most of the lateral surface of the maxilla (Fig. 2B); 2) a well-developed pterygoid process of the maxilla (Fig. 2C, F); and 3) a tectum supraorbitale (Fig. 2I-J). However, our specimens differ from H. ‌szukacsi in: 1) having a well-developed lamina horizontalis (moderately developed in H. ‌szukacsi); and 2) in having the maxillary tooth row extending posteriorly beyond the pterygoid process.

One last hyperossified neobatrachian from the Mesozoic is Baurubatrachus ‌Báez & Perí, 1989 from the Maastrichtian of Brazil. Baurubatrachus and our specimens differ in the ornamentation on their cranial remains (pits and ridges in Baurubatrachus; Báez & Gómez 2018). However, both are similar in having: 1) the medial margin of the lamina horizontalis rounded (in cross-section, Fig. 2D); and 2) a crista dentalis ending slightly beyond the pterygoid process (Fig. 2F).

In Africa, the next oldest hyperossified neobatrachian is the ranoid Rocekophryne ‌ornata ‌Rage et al., 2021 from the Early Eocene of Algeria. Our specimens can be differentiated from Rocekophryne in having: 1) a different ornamentation (deep pits and ridges in Rocekophryne; Rage et al. 2021); 2) ornamentation extending ventrally to cover the entire lateral surface of the pars dentalis; and 3) in having an anterodorsal process (Fig. 2A). However, our specimens and Rocekophryne are similar in having a rounded lamina horizontalis (Fig. 2H). Another extinct hyperossified ranoid is Thaumastosaurus ‌De Stefano, 1903, of which at least three to five taxa are known (Lemierre et al. 2021; Georgalis et al. 2023). Our specimens differ from Thaumastosaurus in: 1) having an ornamentation made of small tubercles and pits (deep pits and ridges in Thaumastosaurus; Fig. 2B); and 2) having a well-developed pterygoid process (Fig. 2D, F).

Thus, the In Becetèn ornamented anuran can be mainly differentiated from all extinct hyperossified anurans with the following characters: 1) a dermal ornamentation made of small closely spaced tubercles forming a honeycomb pattern (shared with Theatonius and Scotiophryne); and 2) a well-developed pterygoid process (shared with Hungarobatrachus).

Known extant hyperossified and ornamented neobatrachians mainly occur among hyloids, except for three genera, Aubria ‌Boulenger, 1917, Cornufer ‌Tschudi, 1838 and Pyxicephalus ‌Tschudi, 1838. Aubria and Pyxicephalus (in the African endemic family Pyxicephalidae) both possess the same characteristics on their maxillae as Thaumastosaurus ‌De Stefano, 1903, except for having a more developed pterygoid process (Sheil 1999). Thus, our specimens can be differentiated from all Pyxicephalidae (Thaumastosaurus, Aubria and Pyxicephalus). Cornufer is a hyperossified ranoid (family Ceratobatrachidae) from Indonesia. Its maxilla bears no anterodorsal process, and its anterior region is low and mostly devoid of ornamentation. Furthermore, the lamina horizontalis projects lingually into a thin blade in Cornufer. Thus, our specimens likely differ from all known hyperossified ranoids.

All remaining ornamented taxa are found within the Hyloidea, mainly among Ceratophryidae ‌Tschudi, 1838 and Calyptocephalellidae ‌Reig, 1960. Our taxon resembles the Ceratophryidae in (likely) having an anterodorsal process on its maxilla. Our taxon differs from all ceratophryids in having a lamina horizontalis slightly protruding from the lingual surface of the maxilla (flattened in Ceratophryidae).

Calyptocephalellidae and its type genus, Calyptocephalella, are known throughout the South American fossil record since the Late Cretaceous (Agnolin 2012; Muzzopappa 2019; Muzzopappa et al. 2020). Incomplete specimens attributed to this clade have all been referred to Calyptocephalella, an attribution considered equivocal (Nicoli et al. 2017). Hence, we will consider all materials referred to Calyptocephalella as Calyptocephalellidae in this comparison. Our specimens resemble calyptocephalellids in having: 1) ornamentation covering the lateral surface of the maxilla, except for a thin strip of bone ventrally along the margin of the crista dentalis (Fig. 2B); and 2) in having an anterodorsal process (Fig. 2A). However, our specimens differ from calyptocephalellids in lacking a distinct lamina horizontalis on most of the length of the maxilla (in calyptocephalellids, the lamina horizontalis is flattened lateromedially only at the level of the orbit; Muzzopappa & Báez 2009: fig. 3.2).

To summarize, we can exclude our specimens from all non-neobatrachian anuran clades, but we did not recover known synapomorphies to firmly attribute these fossils to Neobatrachia. However, they seem to share several characters on the maxilla with neobatrachian clades, like the presence of an anterodorsal process. Thus, these ornamented specimens represent a new neobatrachian taxon, here informally called “Neobatrachia gen. et sp. indet.”. We refrain from erecting a formal taxon in the absence of better-preserved materials. To assess its affinities, we included this unnamed neobatrachian in a morphological dataset composed of extant and extinct anurans with emphasis on neobatrachians.

Nine fragments of maxillae (MNHN.F.IBC1989a, IBC1989b, IBC1991a-IBC1991f, IBC2063).

All three maxillary morphotypes are distinct from the unnamed neobatrachian and appear to represent three other distinct taxa. It should be noted that the presence of teeth does not exclude an attribution to the four pipimorph taxa known (Lemierre et al. 2023, 2025). Although most pipids lack teeth (Trueb et al. 2000), xenopodines and extinct pipimorphs are known to have teeth (Henrici & Báez 2001; Báez & Púgener 2003). However, the ornamentation present in morphotypes A and C differs from that of Pachycentrata ‌Báez & Rage, 2004 and the unnamed pipimorph 2 (Báez & Rage 1998; Lemierre et al. 2025), while Inbecetenanura Lemierre, Bailon, Folie & Laurin, 2023 and the unnamed pipimorph 1 lack ornamentation. Hence, maxillary morphotypes A, B and C can be assigned neither to Pachycentrata nor to unnamed pipimorph 2. Furthermore, the presence of small patches of ornamentation in the maxillary morphotype B around the frontal process renders an attribution to Inbecetenanura unlikely, as such ornamentation would likely be present on the frontoparietal. Hence, the maxillary morphotype B could be attributed to the unnamed pipimorph 1 or to another non-pipid taxon in In Becetèn. Maxillary morphotypes A and C are assigned to indeterminate non-pipid anurans.

Thus, based on maxilla elements, between three and four non-pipid anuran taxa are present in In Becetèn.

Five incomplete angulosplenials (MNHN.F.IBC1987, IBC1988a, IBC1988b, IBC1990, IBC2064).

The angulosplenials are all represented by posterior portions that lack their posteriormost parts. The sulcus pro cartilagine Meckeli is present in the coronoid portion, excluding pipid affinities (Gómez 2016). Two morphotypes can be distinguished, morphotype A (MNHN.F.IBC1987, 1988a, 1988b) and morphotype B (MNHN.F.IBC1990, IBC2064). However, it cannot be excluded that their differences are of ontogenetic nature.

All specimens attributed to angulosplenial morphotype A are large and thick (Fig. 3J). Although only the anterior region of the coronoid process is preserved, it can be inferred that an anteroposteriorly elongate coronoid crest was present (Fig. 3J, K). This crest did not extend as a flange (such as in Pipidae). A shallow depression is present lateral to the coronoid process (Fig. 3J). The sulcus pro cartilagine Meckeli extends on the lateral region of the angulosplenial and narrows posteriorly (Fig. 3J).

Angulosplenial morphotype B differs from morphotype A in: 1) being smaller and medio-laterally thinner; and 2) having a broader and deeper sulcus pro cartilagine Meckeli (Fig. 3L, M). However, as in morphotype A, the sulcus pro cartilagine Meckeli extends on the lateral surface of the bone (Fig. 3K, M) and the coronoid process forms an anteroposteriorly elongated crest (Fig. 3J-M).

One incomplete sacral vertebra (MNHN.F.IBC1992).

MNHN.F.IBC1992 only preserves the centrum and the base of the neural arch (Fig. 4A-D). The vertebra bears an anterior condyle and two posterior condyles. Most of the anterior condyle is eroded, but it seems laterally wide (Fig. 4A). The posterior condyles are circular and well separated from each other (Fig. 4B, D).

This vertebra was attributed to Ranidae in the original study on In Becetèn (Broin et al. 1974). The presence of a bicondylar sacro-urostylar articulation and a bicondylar (anterior and posterior condyles) sacral vertebra is known in only Alytoidea and Ranoidea (Duellman & Trueb 1994). Alytoidea are not known in Africa before the Neogene (Gardner & Rage 2016). One taxon from the Late Jurassic/Early Cretaceous of Morocco, aff.Enneabatrachus ‌Evans & Milner, 1993, possesses an amphicoelous sacral vertebra with two posterior condyles (Jones et al. 2003). However, its affinity, and the referral of non-ilium elements to this taxon, are not certain (Jones et al. 2003). In addition, MNHN.F.IBC1992 differs from aff. Ennabatrachus in lacking a remnant of notochordal canal separating the posterior condyle. Most characters used for attribution to Ranoidea are found in the neural arch, which is missing. In conclusion, the scant data available are insufficient to refer this element to a specific anuran clade, as it could belong to an unidentified clade (as for the Anoual’s specimen; Jones et al. 2003) or to an unnamed neobatrachian.

Two urostyles (MNH.F. IBC1993, IBC1994).

The two urostyles represent two morphotypes, A (MNHN.F.IBC1993) and B (MNHN.F.IBC1994), both attributed to distinct taxa. As they are not coalesced to the sacral vertebra, they are excluded from Pipidae.

Urostylar morphotype A is represented by an incomplete urostyle preserving most of its anterior portion (Fig. 4E-G). It bears two anterior cotyles for sacral articulation. The cotyles are subcircular and closely spaced, separated by a thin dorso-ventrally oriented crest (Fig. 4E). Transverse processes are absent (Fig. 4F, G). Most of its neural arch is missing, but it clearly extended into a thin dorsal crest (Fig. 4F). However, neither its height nor posterior extension are known.

Urostylar morphotype B (MNHN.F. IBC1994) also preserves only its anterior portion (Fig. H-J). It bears a single anterior cotyle for sacro-urostylar articulation (Fig. 4H). This cotyle is wide and slightly compressed dorsoventrally. The dorsal crest is low and reduced to a thin ridge around the level of the spinal foramen (Fig. 4I, J).

Urostylar morphotype A differs from morphotype B in: 1) having a bicondylar sacro-urostylar articulation; and 2) having a dorsal crest not reduced to a ridge. These differences are not ontogenetic, so they represent two distinct taxa. Morphotype A can also be differentiated from MNHN.F. IBC1992 (the sacral vertebra) in having closely spaced articular facets (cotyles and condyles) for the sacro-urostylar articulation. Thus, MNHN.F. IBC1992 (sacral vertebra), IBC1993 (urostyle morphotype A) and IBC1994 (urostyle morphotype B) are all attributed to three distinct taxa. Morphotype A cannot be attributed to a specific anuran clade. The presence of monocotylar articulation in Morphotype B is uncommon among anurans. It has been recovered in several unidentified anurans from the Late Cretaceous of North America (Roček et al. 2010), in several Bombinatoridae (used as a diagnostic character; Folie et al. 2012), Altyidae (Rage & Hossini 2000) and in some Pipimorphs. Within Bombinatoridae ‌Gray, 1825 and Alytidae, urostyles bear a low dorsal ridge, as in morphotype B, but possess transverse processes (unknown in morphotype B). Thus, the scant data available are insufficient to attribute Morphotype B to any anuran clade less inclusive than Neobatrachia. However, it should be noted that monocotylar articulation has been associated with a fairly aquatic lifestyle (Roček et al. 2010).

One incomplete scapula (MNHN.F.IBC1996).

The scapula preserves the broken base of its pars acromialis (Fig. 4K, L). The pars glenoidalis is not distinct from the scapular shaft and pars acromialis, forming a large articular facet Fig. 4K). The suprascapular region is elongated dorsoventrally. The anterior margin bears an extended crest (Fig. 4K, L). No assignment among anurans is possible.

Three incomplete left ilia (MNHN.F.IBC1600, IBC1601, IBC1995).

These three ilia represent three different morphotypes. All three are excluded from Pipimorpha by: 1) having a dorsal crest (iliac morphotypes A and C); 2) presenting a dorsal and ventral acetabular expansion (iliac morphotypes A and B); and 3) lacking an interiliac tubercle (iliac morphotypes A and B).

Iliac morphotype A (MNHN.F.IBC1600; Fig. 4M-P) bears a low dorsal crest (Fig. 4M, N). The iliac shaft is strongly compressed lateromedially. The dorsal prominence is low and forms an elongate oval bulge (Fig. 4M, O) projecting laterally and located anterodorsal to the acetabular rim (Fig. 4M). The ventral acetabular expansion is incomplete but reduced (Fig. 4M). The preacetabular zone is expanded (Fig. 4M). A small depression beneath the ventral acetabular margin could be a preacetabular fossa. The dorsal acetabular expansion is not preserved. No interiliac tubercle seems to be present (Fig. 4P).

Ilia morphotype B (MNHN.F.IBC1601) lacks a dorsal crest (at least on its distal region; Fig. 4Q-R). The dorsal prominence is scarcely distinct from the rest of the bone, forming an ovoid bulge that projects neither medially nor laterally (Fig. 4Q). The acetabular region is too poorly preserved to distinguish any features.

Iliac morphotype C (MNHN.F.IBC1995) is missing most of its acetabular region. The preserved iliac shaft shows a well-developed dorsal crest (Fig. 4S, T). Although its dorsalmost portion is not preserved, the crest is at least the height of the iliac shaft. The dorsal prominence is an anteroposteriorly elongate ovoid bulge (Fig. 4S). The dorsal protuberance projects lateroposteriorly (Fig. 4S, U). Anterior to the dorsal prominence, the notch for the attachment of the musculus gluteus magnus is present (Fig. 4T). The dorsal prominence is located anterior to the acetabular rim.

Iliac morphotypes A and B are puzzling. The presence of a weakly developed dorsal prominence in morphotype B is reminiscent of several neobatrachians, like Telmatobius ‌(Gómez & Turazzini 2016) and Myobatrachidae ‌(Tyler 1976), and the presence of a low dorsal ridge is also known in some neobatrachians (Gómez & Turazzini 2016). However, both ilia are far too incomplete to propose any precise attribution. In addition, these two morphotypes seem to differ from all known ilia from the Cretaceous of North America (Roček et al. 2010). Regarding morphotype C, the presence of a high dorsal crest has often been used as a characteristic of ranoids (Rage 1984). However, a high dorsal crest also occurs within Alytoidea, like the Cenozoic Latonia ‌(Roček 2013). Thus, we cannot attribute Morphotype C to any clade within Anura.

All three ilia represent three distinct morphotypes, and likely three taxa. However, they are too incomplete to identify them more precisely than Anura, and we cannot determine if they represent any of the other taxa in In Becetèn.

25 radioulnae (MNHN.F.IBC1999), five femora (MNHN.F.IBC2034), 22 tibiofibulae (MNHN.F.IBC2035) and three metatarsals (MNHN.F.IBC2002).

Radioulnae, femora, tibiofibulae and metatarsals, all fragmentary, are assigned to anurans (Fig. 4V, W). However, they do not display peculiar features (such as additional ridges); consequently, no further identification is possible.