The first record of albanerpetontid amphibians (Amphibia: Albanerpetontidae) from East Asia

PLOS ONE, Jan 2018

Albanerpetontids are an enigmatic fossil amphibian group known from deposits of Middle Jurassic to Pliocene age. The oldest and youngest records are from Europe, but the group appeared in North America in the late Early Cretaceous and radiated there during the Late Cretaceous. Until now, the Asian record has been limited to fragmentary specimens from the Late Cretaceous of Uzbekistan. This led to speculation that albanerpetontids migrated into eastern Asia from North America in the Albian to Cenomanian interval via the Beringian land bridge. However, here we describe albanerpetontid specimens from the Lower Cretaceous Kuwajima Formation of Japan, a record that predates their first known occurrence in North America. One specimen, an association of skull and postcranial bones from a single small individual, permits the diagnosis of a new taxon. High Resolution X-ray Computed Microtomography has revealed previously unrecorded features of albanerpetontid skull morphology in three dimensions, including the presence of a supraoccipital and epipterygoids, neither of which occurs in any known lissamphibian. The placement of this new taxon within the current phylogenetic framework for Albanerpetontidae is complicated by a limited overlap of comparable elements, most notably the non-preservation of the premaxillae in the Japanese taxon. Nonetheless, phylogenetic analysis places the new taxon closer to Albanerpeton than to Anoualerpeton, Celtedens, or Wesserpeton, although Bootstrap support values are weak. The results also question the monophyly of Albanerpeton as currently defined.

A PDF file should load here. If you do not see its contents the file may be temporarily unavailable at the journal website or you do not have a PDF plug-in installed and enabled in your browser.

Alternatively, you can download the file locally and open with any standalone PDF reader:

http://journals.plos.org/plosone/article/file?id=10.1371/journal.pone.0189767&type=printable

The first record of albanerpetontid amphibians (Amphibia: Albanerpetontidae) from East Asia

January The first record of albanerpetontid amphibians (Amphibia: Albanerpetontidae) from East Asia Ryoko Matsumoto 0 1 Susan E. Evans 1 ☯ These authors contributed equally to this work. 1 1 0 Department of Zoology, Kanagawa Prefectural Museum of Natural History , Odawara, Kanagawa Prefecture , Japan , 2 Department of Cell and Developmental Biology, University College London , London, England 1 Editor: Thierry Smith, Royal Belgian Institute of Natural Sciences , BELGIUM Albanerpetontids are an enigmatic fossil amphibian group known from deposits of Middle Jurassic to Pliocene age. The oldest and youngest records are from Europe, but the group appeared in North America in the late Early Cretaceous and radiated there during the Late Cretaceous. Until now, the Asian record has been limited to fragmentary specimens from the Late Cretaceous of Uzbekistan. This led to speculation that albanerpetontids migrated into eastern Asia from North America in the Albian to Cenomanian interval via the Beringian land bridge. However, here we describe albanerpetontid specimens from the Lower Cretaceous Kuwajima Formation of Japan, a record that predates their first known occurrence in North America. One specimen, an association of skull and postcranial bones from a single small individual, permits the diagnosis of a new taxon. High Resolution X-ray Computed Microtomography has revealed previously unrecorded features of albanerpetontid skull morphology in three dimensions, including the presence of a supraoccipital and epipterygoids, neither of which occurs in any known lissamphibian. The placement of this new taxon within the current phylogenetic framework for Albanerpetontidae is complicated by a limited overlap of comparable elements, most notably the non-preservation of the premaxillae in the Japanese taxon. Nonetheless, phylogenetic analysis places the new taxon closer to Albanerpeton than to Anoualerpeton, Celtedens, or Wesserpeton, although Bootstrap support values are weak. The results also question the monophyly of Albanerpeton as currently defined. - Data Availability Statement: The data is within the paper in the form of description and images. The data used in the phylogenetic analysis can be found in S1 Supporting information file. Funding: The first author received funding from the Linnean Society (GB), named as the Anne Sleep award, which funded travel to the UK for collaboration (to RM). The second author received no specific funding for this work. Competing interests: The authors have declared that no competing interests exist. Introduction Albanerpetontidae form a distinct and highly derived extinct amphibian clade with a long fossil record. Albanerpetontids share the double occipital condyle of lissamphibians but are characterised by a unique set of skeletal features including: a complex `mortise and tenon' interdentary joint; non-pedicellate, slightly tricuspid, teeth; a sculptured median (fused) frontal; and an amniote-like `atlas-axis' involving three anterior cervical components. Albanerpetontids have been variously considered to be caudates or stem-caudates [1±5], stem-batrachians [6±9], the sister group of Gymnophionomorpha [10±11]; or stem-lissamphibians [5,11±13]. This uncertainty is due partly to differing opinions on the origins and monophyly of Lissamphibia [14], but also to the fragmentary nature of most albanerpetontid material. Where complete specimens exist, from the Early Cretaceous Spanish locality of Las Hoyas [6, 8, 15], they are two-dimensionally compressed, leaving many skull features difficult to interpret. Some three-dimensional skull associations have been described [13, 16], but this material (Albanerpeton pannonicum) is fragmentary and comparatively recent (Pliocene), so not necessarily representative of primitive albanerpetontid morphology. Albanerpetontidae are predominantly Laurasian in distribution [17], the exceptions being specimens from the Middle Jurassic (Bathonian) [18] and Early Cretaceous (Berriasian) [19± 20] of Morocco. The longest record for albanerpetontids is in Europe and extends from the Middle Jurassic (Bathonian) of France [21] and England [20, 22±23], to the Pliocene of Hungary [16] and Italy [24], albeit with an unexplained hiatus in the Eocene [25]. In North America, albanerpetontids are first recorded from the uppermost Aptian or lower Albian [26±28] and then remained a consistent element of North American microvertebrate assemblages through to the late Palaeocene [17, 26±27, 29±30]. By contrast, the Asian record of the group is poor and, until now, was restricted to rare and fragmentary material from the Late Cretaceous (Cenomanian/Coniacian) of Uzbekistan, Central Asia [17, 31±33]. The relative ages of the first Asian and American records have influenced discussions as to the direction, timing, and route of possible albanerpetontid dispersal between the two continents [26, 32±34]. Here we extend the record of albanerpetontids in Asia with the description of a new taxon represented by an associated specimen and two isolated dentaries from the Early Cretaceous (Barremian) of Japan. This is the earliest confirmed record from Asia and the first from eastern Asia. Moreover, the three-dimensional preservation of individual elements, as revealed through High Resolution X-ray Computed Microtomography (μCT), provides important new information on albanerpetontid morphology. Given that most recent phylogenetic analyses find albanerpetontids to be related to crown Lissamphibia, a fuller understanding of their morphology has the potential to inform the debate on lissamphibian origins. Geological setting The Mesozoic (Middle Jurassic-Lower Cretaceous) deposits of the Tetori Group are widely distributed in the western part of central Japan, and present a gradual transition from marine to freshwater conditions [35±38]. The Tetori Group is traditionally subdivided into three ascending units: the Kuzuryu Subgroup, dominated by marine deposits; the Itoshiro Subgroup, containing marine and terrestrial deposits; and the Akaiwa Subgroup, consisting mainly of terrestrial sediments [36±37]. The Kuwajima Formation forms the upper part of the Itoshiro Subgroup in the Tetodori River District, and it is mainly composed of non-marine sandstones and mudstones. The albanerpetontid specimens described herein were collected from an outcrop of the Kuwajima Formation forming the Fossil Cliff ("Kaseki-Kabe") in the Kuwajima district, Hakusan City (formerly Shiramine Village), Ishikawa Prefecture (Fig 1). Isajii et al. [39] identified three facies at this locality: Facies I, a carbonaceous swamp; Facies II, a shallow lake; and Facies III, a vegetated swamp. Together, these are thought to represent the wide, stable, vegetated floodplain of a meandering river system, with a humid environment [40±41]. All of these facies have yielded vertebrate remains, including dinosaurs [42±43], pterosaurs [44], mammals [45±46], tritylodonts [47±49], lizards [50±54], fish [55±57], turtles [58], turtle eggshells [59], choristoderes 2 / 58 Fig 1. Map of western Japan showing position of the type locality. Areas marked by hashed lines indicate outcrops of the Tetori Group in this region. [60±61], and rare frogs [62]. To date, more than 2500 specimens have been recorded from this locality. The greatest abundance of terrestrial taxa has been recovered from Facies III, as were the albanerpetontid specimens described herein. Most researchers have dated the Kuwajima Formation to the Early Cretaceous, but age estimates have varied from early Neocomian [40] to Hauterivian [38], Valanginian [39], or Berriasian-Hauterivian [42, 60], based on biostratigraphic correlations with other Tetori Group strata. However, zircon U-Pb from a tuff that intruded into the lower part of the Kuwajima Formation gave a date of 130.7 ± 0.8 (2SE) Myr [63], and zircon U-Pb ages of 132 ± 0.9 (2 SE) and 117 ± 0.7 (2 SE) Myr have been reported [64] for the Okurodani Formation in neighbouring Gifu Prefecture, which shares faunal components with the Kuwajima Formation (e.g. the lizard Sakurasaurus and the choristodere Monjurosuchus). The most recent work [65] dated the Kuwajima Formation to the Barremian, and this is the age estimate we are using herein. The deposit may be older than this but is unlikely to be younger. Material and methods The construction of a road tunnel through the "Kaseki-Kabe" fossil cliff in 1997 yielded almost 17, 000 m3 of fossiliferous matrix. Much of this was attributed to the plant rich Facies I, but a sample of 210 m3 of Facies II and Facies III was separated out and retained. The matrix is resistant to chemicals and each large block is carefully split into smaller pieces, while examining the exposed surfaces for bone. Any exposed bone is then prepared manually by collections staff and volunteers, with the resulting specimens also being examined periodically by specialists. However, this approach does make it almost impossible to trace the component parts of an 3 / 58 original block once it has been broken into smaller pieces. No permits were required for the described study, which complied with all relevant regulations. The available albanerpetontid material comprises two referred dentaries and a small block of matrix, containing a bone association. All three specimens are held in the Shiramine Institute of Paleontology, Hakusan Board of Education, Hakusan City, Ishikawa Prefecture, Japan (SBEI, Shiramine Board of Education). They are accessible to researchers on application to the curators. Comparative albanerpetontid material from the UK (NHMUK PV R 36956, a parietal) and Spain (MCCM-LH-15710, a partial skeleton) was on loan from the permanent collections of the Natural History Museum, London, and the Museo de las Ciencias de Castilla-LaMancha, Cuenca, Spain, respectively. The matrix block containing the bone association is very small. Initially, only a small area of sculptured bone was exposed on one surface, providing no indication that it originally formed part of a larger specimen, the remains of which cannot be traced. Subsequent manual preparation of the block revealed not only several skull bones on the surface (Fig 2A), but also further underlying bones that could not be accessed without damaging neighbouring elements. Courtesy of the Tokyo Metropolitan Industrial Technology Research Institute, we were able to scan the whole block using a CT scanner (Toscaner 30000 micro CN) at a slice distance of 0.029 mm (100 kV, 30 μA). The braincase region was then rescanned on the same machine at a slice thickness of 0.016 mm (100.0 kV, 30 μA) to achieve greater resolution. One dentary specimen (SBEI 2405) was scanned at the Nagoya Municipal Industrial Research Institute (Toscanner 30000) at a slice width of 0.008 mm (100 kv, 35 μA). A second dentary (SBEI 2462), recovered later, was scanned at the Tokyo Metropolitan Industrial Technology Research Fig 2. Shirerpeton isajii gen. et sp. nov., SBEI 2459, holotype block. A, digital photograph showing surface view of the specimen after manual preparation; B, rendered view of the surface from μCT data showing identifications of exposed elements. Abbreviations: Br, braincase elements; Fr, frontal; L.La, left lacrimal; L.Mx, left maxilla; L.N, left nasal; L.Pa, left parietal; LPf, left prefrontal; L.Sm, left septomaxilla; L.Sq, left squamosal; R.La, right lacrimal; R.Pa, right parietal; R.Pf, right prefrontal; R.Sq, right squamosal;?, unidentified element. Scale bars = 5 mm. 4 / 58 Fig 3. Shirerpeton isajii gen. et sp. nov., skull reconstruction. A-D, Model construction. 3-D models constructed using printouts of the individually segmented elements from the μCT data (mirrored as needed: nasal, parietal, possible supratemporal) and fitted into modelling clay. A, dorsal; B, right lateral; C, left lateral; D, left anterolateral showing the relations of the nasal, lacrimal, and maxilla in the narial margin. The tip of the rostrum is roughly reconstructed in modelling clay. E, outline reconstruction of the skull in dorsal view, based on the 3-D model in A-D. Note that the suspensorial elements are omitted as their positions are uncertain. Abbreviations: Fr, frontal; J, jugal; La, lacrimal; Mx, maxilla; N, nasal; Pa, parietal; Pf, prefrontal; S.O, supraoccipital;? St, possible supratemporal. Institute at a slice thickness of 0.006 mm (100 kV, 65 μA). Image reconstruction in all cases used AVISO v. 8 (Fig 2B), although the small size of individual elements rendered segmentation of features like foramina, tooth tips, and fine edges difficult. Individual elements segmented out from the scanned block were then 3-D printed (Objet350 Connex) at 20x original size, mirrored where necessary (to provide pairs), and fitted together manually using modelling clay to determine bone positions and articulations in three dimensions (Fig 3A±3D). The physical model generated was then used for the skull reconstruction shown in Fig 3E. Nomenclatural acts The electronic edition of this article conforms to the requirements of the amended International Code of Zoological Nomenclature, and hence the new names contained herein are available under that Code from the electronic edition of this article. This published work and the nomenclatural acts it contains have been registered in ZooBank, the online registration system for the ICZN. The ZooBank LSIDs (Life Science Identifiers) can be resolved and the associated information viewed through any standard web browser by appending the LSID to the prefix ªhttp://zoobank.org/º. The LSID for this publication is: urn:lsid:zoobank.org:pub: C8490EBE-E927-4CD6-BEA6-BEFE1A71DF77. The electronic edition of this work was published in a journal with an ISSN, and has been archived and is available from the following digital repositories: PubMed Central, LOCKSS, UCL Discovery Publications database. 5 / 58 Note on nomenclature Marjanović and Laurin [11] noted that as erpeton is neuter, a number of original species names have a grammatically incorrect ending e.g. Anoualerpeton unicus [20] should be An. unicum. We have therefore used the corrected species spelling throughout this manuscript. Systematic palaeontology Lissamphibia [ 66 ] Albanerpetontidae [12] Shirerpeton gen. nov. Etymology urn: lsid: zoobank.org: act:C3726120-0EF5-4407-8624-05A4C0B28EFD From the Japanese Shiro, white, partly for Shiramine, the type locality, but also because the family name, Albanerpetontidae, derives from the original French locality of La Grive-SaintAlban, with Alba/Alban (Latin) meaning white. Type species Shirerpeton isajii sp. nov. urn:lsid:zoobank.org:act:86783E12-F44E-4F12-8C96-81327BA47DB6 Diagnosis. As for type and only species Shirerpeton isajii sp. nov. Holotype. Shiramine Board of Education Ishikawa Prefecture, SBEI 2459, a small block bearing most of a disarticulated but associated skull with some postcranial elements (Fig 2A). The specimen is housed in the Shiramine Institute of Paleontology, Hakusan Board of Education, Hakusan City, Ishikawa Prefecture, Japan. Etymology. Species name honours Dr Shinji Isaji, Chiba Prefecture Museum, Japan, for his longstanding work on the fossils, geology, and palaeoenvironment of the Kuwajima Formation. Locality and horizon. "Kaseki-Kabe" (fossil cliff), Kuwajima, Hakusan City (formerly Shiramine Village), Ishikawa Prefecture, Honshu, Japan. Lower Cretaceous, (Barremian), Kuwajima Formation, Tetori Group. Both the holotype and the referred specimen came from Facies III, the more terrestrial of the three facies at the type locality. Differential diagnosis. A genus of albanerpetontid that resembles Albanerpeton spp. and Wesserpeton evansae [34], and differs from Anoualerpeton spp. [20] and Celtedens ibericus [6], in having a frontal that is triangular rather than bell-shaped; differs from all but Albanerpeton arthridion [12], in that the frontal body is short but anteriorly quite wide, but differs from A. arthridion in having a longer, more pointed internasal process; resembles Wesserpeton and A. arthridion in its very small body size, but the frontal differs from that of Wesserpeton in the more tapered internasal process, the anterior contact between the ventrolateral crests, and the fact that the posterior margins of the prefrontal facets lie posterior to the mid-length of the bone; differs from Wesserpeton, Albanerpeton inexpectatum [3], Celtedens ibericus [6], and Anoualerpeton priscum [20], in having parietals with proportionally longer postorbital wings, the dorsal surfaces of which remain completely unsculptured; further differs from A. inexpectatum in lacking fusion of the prefrontal and lacrimal, having a bifurcate occipital shelf of each parietal, and in having a nasal that enters the narial margin. In the latter feature it resembles 6 / 58 A. pannonicum [16], but differs in that the nasal makes a larger contribution to the narial margin; differs from both Neogene taxa in the lack of fusion of the basicranial and otic elements, the discrete supraoccipital, and the long anterodorsal process of the latter bone. The dentary and maxilla resemble Anoualerpeton spp. and differ from Wesserpeton, Celtedens, unattributed material from the Cretaceous of Uzbekistan (`Nukusurus'), and Albanerpeton (except A. nexuosum [ 2 ]) in showing size heterodonty with large anterior teeth supported by a convex profile of the labial alveolar margin. However, the maxilla of Shirerpeton is distinguished from that of Anoualerpeton spp and A. nexuosum by the combination of a pointed (rather than rounded) anterior premaxillary process and a weakly concave anterior narial margin. The dentary of Shirerpeton is distinguished from that of Anoualerpeton spp and A. nexuosum by the greater sinuosity of the labial dental margin (with a concave-convex-concave profile), the positioning of the prearticular facet behind the tooth row (rather than extending forward beneath it), and the shallow posterior inclination of the subdental shelf such that the posterior teeth are not markedly smaller than those in the symphyseal region. Note that direct comparison with the type species of Celtedens, C. megacephalus [67], is not possible as the holotype specimen (Instituto Geologico dell'Universita de Napoli, Italy, IGUN M542: the anterior part of a skeleton) does not show comparable features. Referred specimens. SBEI 2405, an almost complete right dentary (Fig 4A±4D); and SBEI 2462 (Fig 4E), a second right dentary, both from the type locality. Anatomical description The holotype of Shirerpeton isajii, SBEI 2459 (Fig 2), is a small block of grey mudstone bearing three-dimensionally preserved but partly disarticulated skull and postcranial elements. From the anatomical relationships of the bones, lack of duplication, and equivalent size, the remains clearly pertain to a single individual. The reconstructed midline skull length was 8-10mm, with an estimated snout-pelvis length (SPL) of roughly 45 mm (based Gardner [26], with SPL approximately 10x frontal length). As exposed on the surface of the block (Fig 2A), the most distinctive elements are a fused frontal bearing the raised polygonal sculpture typical of albanerpetontids, and paired sculptured parietals. Despite the small size of the bones, the sculpture is of strong relief. Displaced to the left of the frontal are several other bones including the left and right lacrimal, the left prefrontal, and the left maxilla preserved in partial dorsal view. The right prefrontal, crosses the anterior end of the frontal. Anterior to the prefrontal is the left nasal, inverted but complete. Posterior to the frontal are the left and right parietals. An anteriorly directed, rod-like element adjacent to the edge of the right parietal is interpreted as the displaced squamosal of that side. There are further partially exposed elements posterior to the parietals. These were difficult to interpret in surface view but were subsequently found to be vertebrae and parts of the braincase. The surface view has been supplemented by the μCT data which has revealed important details of the exposed bones as well as those of elements that are fully or partially embedded within the matrix (Figs 5±7). These additional bones include the right maxilla, both septomaxillae, quadrates, epipterygoids, and jugals, as well as many parts of the endocranium, several vertebrae, a few limb elements, and several additional elements, some of which have a distinctive structure but have yet to be identified. Unfortunately there is no trace of the premaxillae on the holotype block, nor the lower jaws, and these parts of the block were probably removed inadvertently during trimming of a larger block. Skull. There are few published accounts of the albanerpetontid nasal [3,8], leading to a degree of uncertainty about its presence, size, and relations in most taxa. Venczel and Gardner [16] described the first three-dimensionally preserved albanerpetontid nasal in the Pliocene Albanerpeton pannonicum, revealing that it was larger than previously estimated [3]. 7 / 58 Fig 4. Shirerpeton isajii gen. et sp. nov., referred right dentaries. A-D, SBEI 2405, in A, labial view as preserved, digital photograph; B-C, specimen as segmented from μCT slice data in B, labial, and C, lingual views; D, lingual view, interpretative drawing. E, SBEI 2462, labial view as preserved, digital photograph. Abbreviations:? an, possible angular facet; mk.f, Meckelian fossa; pra, facet for prearticular; sds, subdental shelf; sym, symphyseal prongs. Scale bars = 1 mm. 8 / 58 Fig 5. Shirerpeton isajii gen. et sp. nov., SBEI 2459, holotype block in dorsal view showing elements segmented from μCT slice data. Abbreviations: Ept, epipterygoid; Ex.O, exoccipital plate; Fr, frontal; L.J, left jugal; L.La, left lacrimal; Lm.B, limb element; L.Mx, left maxilla; L.N, left nasal; L.Pa, left parietal; LPf, left prefrontal; L.Q, left quadrate; L.Sm, left septomaxilla; L.Sq, left squamosal;? Opis, possible opisthotic;? Ot, otic associated element;? Pal, possible palatal element;? Pro, possible prootic; R.J, right jugal; R.La, right lacrimal; R.Mx, right maxilla; R.Pa, right parietal; R.Pf, right prefrontal; R.Q, right quadrate; R.Sq, right squamosal; Sph, sphenoid;? St, possible supratemporal;?, unidentified elements. Scale bar = 1 mm. Note that the vertebrae are not labelled in this figure but are described and figured later in the text. Mechanical preparation of SBEI 2459 exposed a single complete nasal in ventral view, and further details have been revealed by the μCT data (Fig 8). Based on its morphology and fit with the frontal, we interpret the bone as a left nasal. It is roughly rhomboidal. The posterior tip is tapered whereas the anterior tip is blunt and thickened. The facetted posteromedial edge fitted against the internasal process of the frontal. The complementary edges (of nasal and frontal) are of similar length, suggesting that the nasals met only, at most, for a short distance at their anteromedial angle. The posterolateral edge of the nasal bears a narrower surface that contacted the medial edges of the prefrontal and the lacrimal. Our reconstruction demonstrates 9 / 58 Fig 6. Shirerpeton isajii gen. et sp. nov., SBEI 2459, holotype block in ventral view showing elements segmented from μCT slice data. Abbreviations: Ex.O, exoccipital plate; Fr, frontal; L.J, left jugal; L.La, left lacrimal; Lm.B, limb element; L.Mx, left maxilla; L.N, left nasal; L.Pa, left parietal; L.Pf, left prefrontal; L.Q, left quadrate; L.Sm, left septomaxilla; L.Sq, left squamosal;? St, possible left supratemporal;? Opis, possible opisthotic;? Ot, otic associated element;? Pal, possible palatal element;? Pro, possible prootic; R.J, right jugal; R.La, right lacrimal; R.Mx, right maxilla; R.Pa, right parietal; R.Q, right quadrate; R.Sq, right squamosal; S.O, supraoccipital;?Sk1, unidentified skull element; Sph, sphenoid;? St, possible supratemporal;?, unidentified elements. Scale bar = 1 mm. Note that the vertebrae are not labelled in this figure but are described and figured later in the text. that the nasal-lacrimal contact excluded the prefrontal from the narial margin. Of the two anterior nasal borders, the thin unfacetted anterolateral edge clearly entered the narial opening. The anteromedial edge is thicker, inflected ventrally, and slightly buttressed. It supports a dorsal facet for the nasal process of the premaxilla (Pm.ft). The thickened edge is squared-off 10 / 58 Fig 7. Shirerpeton isajii gen. et sp. nov., SBEI 2459, holotype block. Lateral view of skull elements as segmented from μCT slice data. Abbreviations: Ex.O, exoccipital plate; Fr, frontal; L.La, left lacrimal; Lm.B, limb element; L.Mx, left maxilla; L.N, left nasal; L.Pa, left parietal; LPf, left prefrontal; Q, quadrates; L.Sm, left septomaxilla;? Opis, possible opisthotic;? Ot, otic associated element;? Pal, possible palatine;? Pro, possible prootic; R.La, right lacrimal; R.Pa, right parietal; R.Pf, right prefrontal; R.Sq, right squamosal; S.O, supraoccipital; Sph+ Ept, sphenoid and epipterygoid;? St, possible supratemporal;?, unidentified elements. Scale bar = 1 mm. Note that the vertebrae are not labelled in this figure but are described and figured later in the text. and bears several rugosities. In other albanerpetontid taxa, the naso-premaxillary contact is usually described as being an abutment and/or a variably developed suture in which the premaxilla overlaps the anterior margin of the nasal [16]. The Japanese specimen appears to combine these features, with the premaxilla overlapping the nasal, and the tip of the nasal abutting against a ridge or tuberosity on the underside of the premaxillary nasal process. However, a second anteroventral facet on the Japanese nasal suggests the articulation may have been more complex, perhaps with a pocket facet on the premaxilla (?Pm.ft). Without the premaxilla, however, we cannot speculate further. The μCT scans revealed the presence of a very small element lying adjacent to the anteromedial edge of the nasal. This has a smooth, rounded, hemispherical external surface and an internal surface divided between a deep concavity and a rugose portion that tapers to a point. It appears to be a single complete element rather than a broken part of a larger bone (Fig 9). From its position in relation to the nasal, and its convex-concave shape, a septomaxilla is the most plausible identification. This element, associated with the nasal capsule, is present in at least some representatives of both extinct and extant amphibian lineages [68, 69], and therefore its presence in albanerpetontids would not be unexpected. The frontal is well preserved. In surface view the internasal process is obscured by the overlying right prefrontal (Figs 2 and 5), but this has been digitally removed. As reconstructed from μCT data (Fig 10), the frontal has a midline length of c. 4.5 mm and a posterior width of c. 3.7 mm, giving a midline length to posterior width proportion of 1.21. The internasal process is relatively long and tapering (length = 1.26x basal width), with the edge recessed in its posterior half by the nasal facet. Further anteriorly, the naso-frontal articulation appears to have been more of an abutment but allowance must be made for possible artefacts introduced during the segmentation of these very small bones. Small anterolateral processes separate the nasal facets from the prefrontal facets. However, these processes do not appear to have reached the level of the dorsal surface, so that the prefrontal and nasal would have been in contact in a dorsal view of the skull. The anterior end of the orbital margin (demarcated by the posterior end of the slot for reception of the prefrontal) lies posterior to the mid-point of the frontal long 11 / 58 Fig 8. Shirerpeton isajii gen. et sp. nov., SBEI 2459, nasal. Left nasal as segmented from μCT slice data in A, dorsal; B, ventral; C, medial; D, lateral; and E, anterior views. Abbreviations: Fr.ft, frontal facet; La.ft, lacrimal facet; N.bd, narial border; Pf.ft, prefrontal facet; Pm.ft, anterodorsal facet for nasal process of premaxilla;? Pm.ft, possible anteroventral facet for the premaxilla. Scale bar = 1 mm. axis. The lateral edges of the bone are only slightly concave, but the dorsal surface of the bone (Fig 10A) is slightly narrower than the ventral one (Fig 10B), so that a narrow gutter runs along the edge from the posterior margin of the prefrontal facet to the posterolateral tip of the frontal (Fig 10C and 10D). Here there is a small dorsal slot facet for the parietal (more completely preserved on the left than the right). The posterior margin of the frontal (Fig 10A, 10B and 10E) is weakly W-shaped, with shallow bilateral emarginations flanking a short posteromedian process. The ventral surface of the frontal has been rendered digitally (Fig 10B). The surfaces of the ventrolateral crests are weakly concave. The crests are widest at the posterior margin of the prefrontal facet and then narrow slightly toward the frontoparietal suture. In the posterior two-thirds of the bone, the crests are separated by a deep concavity, but this is closed off further anteriorly where the ventrolateral crests are joined in the midline. There is a weak ventromedian crest. Seen in lateral view, the prefrontal facets are deep and extend along the full length of the lateral surface of the anterolateral processes. As noted above, the nasal facets are shallower. At the posterodorsal corner of the frontal (preserved more completely on the left than the right, Fig 10A), the frontal bears a small facet for the parietal. This is continuous with a larger ventral facet that extends medial to the ventrolateral crest. The posterolateral corner of the frontal thus slots into a recess in the anterior face of the parietal. Albanerpetontid parietals have rarely been described and, to date, the only nearly complete bones that have been figured are those of the Miocene Albanerpeton inexpectatum [3] and the Early Cretaceous Celtedens ibericus [6, 8]. The parietals are well preserved in the Tetori specimen. The left bone is roughly in situ relative to the frontal (Figs 5, 6 and 11), but the right has been rotated clockwise so that its anterior margin faces laterally (Fig 2). Most of the important details of the dorsal surface can be seen on the original block (Fig 11), but the description is supplemented by images from μCT scans which also allow visualisation of the ventral surface (Figs 12 and 13). 12 / 58 Fig 9. Shirerpeton isajii gen. et sp. nov., SBEI 2459, septomaxilla. Left septomaxilla as segmented from μCT slice data in A, dorsal; B, ventral; C, posterior; D, anterior; E, lateral; and F, medial views. Scale bar = 0.5 mm. Each parietal is roughly triangular. The lateral margin of the parietal is curved and extends from the lateral tip of the postorbital process to the tip of the occipital shelf. The anterior margin is smooth along the tapering postorbital process, then irregular where it bears the frontal facet. The μCT scan images reveal a deep pocket (Fr.ft) that received the posterolateral tip of the frontal (Fig 12C). Anteroventral and anterolateral flanges match the facets on the corresponding surfaces of the frontal, contributing to a firm articulation. The central part of the frontoparietal suture and the straight interparietal suture appear weaker with only small interdigitations. However, the medial edge of the occipital shelf bears an incised slot facet for the anterior ramus of the supraoccipital (Figs 12B, 12C and 13B±13D). The dorsal surface of the parietal (Figs 12A and 13A) is subdivided into three areas: a sculptured triangular anteromedial surface that forms the posterior part of the skull table; an unsculptured lateral postorbital wing; and an unsculptured occipital shelf to which neck muscles presumably attached. On the right parietal the shelf is extensive, as in A. inexpectatum, but it differs in being bifurcated. This does not appear to be an artefact of breakage (although a vertebral transverse process projects between the two rami) or crushing, as the edges bordering the emargination appear intact, and the completely preserved posteromedial processes on both parietals match one another in shape and size. On the left parietal the more lateral of the posterior processes has broken off at its base, as is clear from a comparison of the ventral surfaces of the right and left bones. On both elements, a sharp anteromedial to posterolateral transverse crest is seen to divide the main, medial, part of the bone from the recessed postorbital wing (Figs 12B and 13B). A second anteroposterior crest runs smoothly from the midpoint of the transverse crest to the tip of the posterolateral process. On the left parietal, this second crest is truncated posteriorly, marking where the posterolateral process has broken away. On the right parietal, the area lateral to the second crest is seen to bear a lateral recess, with a more oblique anterolateral recess towards its anterior end. Estes and Hoffstetter's image of the parietal in Albanerpeton inexpectatum ([3] plate 8) shows a similar lateral recess, which the authors interpreted as accommodating the dorsal surface of the otic capsule. However, the larger lateral 13 / 58 Fig 10. Shirerpeton isajii gen. et sp. nov., SBEI 2459, frontal. Frontal as segmented from μCT slice data in A, dorsal; B, ventral; C, right lateral; D, left lateral; and E, posterior views. Abbreviations: alp, anterolateral process; inp, internasal process; N.ft, nasal facet; Pa.ft, parietal facet; Pf.ft, prefrontal facet; vlc, ventrolateral crest; vmc, ventromedian crest. Scale bar = 1 mm. recess in Shirerpeton more closely resembles a facet for a dermal element, possibly a supratemporal (see below), whereas the more anterior recess matches the size and position of the dorsal end of an element that we interpret as the epipterygoid (see below). The strongly concave anteroventral surface of the postorbital wing presumably provided an area of origin for jaw adductor muscles (possibly the deep part of the internal adductor). As reconstructed, therefore, the occipital shelves appear to have been deeply emarginated in the Japanese taxon. With the supraoccipital in articulation between the posteromedial parietal processes, this creates large ovoid openings in the skull roof on each side of the midline, medial to the otic capsules (Fig 3E). If genuine (i.e. not an artefact), these openings could have been covered by a thick sheet of fascia or an additional dermal element such as a postparietal. 14 / 58 Fig 11. Shirerpeton isajii gen. et sp. nov, SBEI 2459, digital photograph showing details of skull roofing elements. A, frontal and left parietal; B, right parietal. Abbreviations: Fr, frontal; Fr,ft, frontal facet; L. La, left lacrimal; L.Mx, left maxilla; L.Pa, left parietal; L.Pf, left prefrontal; plp, posterolateral process; pmp, posteromedial process; PoW, postorbital wing; R.Pf, right prefrontal;? St, possible supratemporal; V, vertebra. Scale bars = 1 mm. Both maxillae are preserved, with the left roughly in situ and the right lying under the right parietal (Figs 5 and 6). The maxilla has the shape of a shallow scalene triangle (Fig 14A and 14B), with a tapering premaxillary process, a low rounded facial process, and a long tapering jugal process (Fig 14A±14F and 14H). The anterior, narial, margin of the facial process is almost straight (rather than strongly concave as in some other taxa). Allowing for artefacts in 15 / 58 Fig 12. Shirerpeton isajii gen. et sp. nov., SBEI 2459, right parietal. Right parietal as segmented from μCT slice data in A, dorsal; B, ventral; C, oblique anteroventral; D, anterior views. Abbreviations: al.re, anterolateral recess; ap.c, anteroposterior crest; Fr.ft, frontal facet; l.re, lateral recess; Occ.s, occipital shelf; plp, posterolateral process; pmp, posteromedial process; PoW, postorbital wing; SO.ft, supraoccipital facet; tr.c, transverse crest; Scale bar = 1 mm. segmentation and some damage, the labial surface of the maxilla appears unsculptured, but at least two neurovascular foramina open on this surface. The premaxillary process is bifurcated (Fig 14C±14E and 14G). The medial ramus bears a ventral flange that is almost vertical in orientation on the left bone (Fig 14G), but ventromedial on the right. It appears to bear a facet, either for the premaxilla or the vomer. The posterodorsal edge of the facial process is recessed both labially and lingually where it was straddled by the lacrimal bone (Fig 14B±14E and 14H). Further posteriorly, the dorsomedial surface bears a slightly flattened surface that supported the jugal. Lacrimal and jugal facets meet, indicating that the maxilla was excluded from the ventral orbital margin, and this is confirmed when the bones are rearticulated (Fig 3A±3D). A medial view of the right maxilla (Fig 14H) also shows a distinct embayment in the lingual edge of the maxillary shelf posterior to the medial premaxillary process. We interpret this as the lateral margin of the choana. It is less obvious on the left due to damage. Posterior to the embayment, the edge of the shelf is straighter and bears a facet for a palatal element. A large medial foramen perforates the posterior base of the right facial process and probably carried neurovascular structures into the bone. There is a second, more posterior, excavation but we cannot determine whether this also opened as a foramen. Anteriorly, the maxillary tooth row reaches the bifurcation of the premaxillary process, a point anterior to the narial margin of the facial 16 / 58 Fig 13. Shirerpeton isajii gen. et sp. nov., SBEI 2459, left parietal. Left parietal as segmented from μCT slice data in A, dorsal; B, ventral; C, anteroventral; D, ventromedial; and E, anterior views. Abbreviations: al. re, anterolateral recess; ap.c, anteroposterior crest; br.ed, broken edge; Fr.ft, frontal facet; occ.s, occipital shelf; pmp, posteromedial process; PoW, postorbital wing; So.ft, supraoccipital facet; tr.c, transverse crest. Scale bar = 1 mm. process. Allowing for empty positions, the tooth row seems to accommodate 18±22 teeth, those below the facial process being longer than those at the anterior and posterior ends of the tooth row. This heterodonty is also reflected by the downward curvature of the ventrolateral edge of the maxilla which is deepest level with the apex of the facial process (Fig 14A and 14B). At its anterior end, the right maxilla is in contact with a fragment of another bone. This could be part of a right septomaxilla or premaxilla (Fig 14H). The left prefrontal lies lateral to the frontal (Figs 2 and 5±7). It is a slender bone with narrow medial shelf facets for the nasal and frontal, and a lateral, or ventrolateral, lacrimal facet visible on its exposed margin (Fig 15). The medial and lateral facets converge anteriorly and, at the tip of the bone, are separated only by a low ridge. Aligning this element to the edge of the frontal indicates that the posterior part of the medial facet met the frontal, but the more horizontal anteromedial part was overlapped by the nasal. The nasal and lacrimal met anterior to the 17 / 58 Fig 14. Shirerpeton isajii gen. et sp. nov., SBEI 2459, maxilla. A-G, left maxilla as segmented from μCT slice data, in A, lingual; B, labial; C, dorsal; D, ventral; E, oblique anterior; F, oblique posterior; G, anterior views. H, right maxilla in lingual view. Abbreviations: a.ft, anterior facet; fa.p, facial process; J.ft, jugal facet; J. p, jugal process; La.ft, lacrimal facet; m.e, medial emargination; Mx.s, maxillary shelf; n.m, narial margin; p.ft, posterior facet; Pm.p, premaxillary process;? Sm/Pm, bone fragment, part of premaxilla or septomaxilla; v.fl, ventral flange; Scale bar = 1 mm. prefrontal and excluded it from the narial margin (Fig 3E). The right prefrontal lies across the anterior tip of the frontal and has the same morphology as the left. The dorsal edge of the left lacrimal, bearing prefrontal and nasal facets, is exposed lateral to the maxilla and left prefrontal. The remainder of the bone is buried almost vertically in matrix (Figs 5±7). The right lacrimal is fully exposed in lateral view on the matrix block (Fig 2). Both bones are biradiate with tapering ventral and anterodorsal edges processes that lie almost at right angles to one another. Between them, the long, curved, posterior border framed the orbit, whereas the shorter, almost vertical, anterior edge contributed to the narial border. The extracted left lacrimal (Fig 16) resembles the right but reveals rather more detail. Just anterior to the orbital margin, the ventrolateral surface is perforated by two lacrimal duct foramina (Fig 16B). These run anteromedially into a short canal that opens on to the medial surface through a single foramen (Fig 16A and 16C). The lateral foramina are visible only as indentations on the exposed right bone, but they lie in the same position as those on the lateral surface of the right bone. 18 / 58 Fig 15. Shirerpeton isajii gen. et sp. nov., SBEI 2459, prefrontal. Left prefrontal as segmented from μCT slice data in A, dorsal; B, ventral; C, lateral; and D, medial views. Abbreviations: Fr.ft, frontal facet; La.ft, lacrimal facet; N.ft, nasal facet. Scale bar = 1 mm. Fig 16. Shirerpeton isajii gen. et sp. nov., SBEI 2459, lacrimal. Left lacrimal as segmented from μCT slice data in A, medial; B, lateral; C, anterodorsomedial; D, anterior; E, posterior; F, dorsal; and G, ventral views. Abbreviations: J.ft, jugal facet; La.f, lacrimal duct foramen; l.lm, lateral lamina; m.lm, medial lamina; Mx.ft, maxillary facet; N.bd, narial border; N.ft, nasal facet; Or.bd, orbital border; Pf.ft, prefrontal facet; Scale bar = 1 mm. 19 / 58 The medial surface of the lacrimal is exposed most clearly on the left bone, with dorsal and ventral facets separated by a short body (Fig 16A and 16C). The body is divided into a thin, medially concave anterior part that walled part of the nasal chamber (Fig 16D) and a thickened posterior orbital margin (Fig 16E). The single lacrimal duct foramen opens into the posterior edge of the nasal concavity. The dorsomedial border of the bone bears a large tapering facet for the prefrontal. Anterior to this facet, the smooth bone surface changes slightly in its orientation and may represent the point at which the lacrimal contacted the nasal (Fig 16A, 16C and 16F). The ventral margin of the bone bears a deep groove divided into anterior and posterior articular surfaces for the maxilla and jugal respectively. The ventral groove is flanked by medial and lateral laminae (Fig 16G). However, whereas the medial lamina is deepest anteriorly, the lateral lamina is deeper posteriorly. This arrangement ensures that the lacrimal, maxilla, and jugal are locked. The ventral groove accommodates the nasal process of the maxilla anteriorly, leaving this process exposed in lateral view but braced medially by the medial lacrimal lamina, and the jugal posteriorly, with the lateral lacrimal lamina bracing the jugo-maxillary contact. Overall, the lacrimal of the Japanese albanerpetontid is broadly similar to that figured for both Albanerpeton inexpectatum [3] and A. pannonicum [16], although the detailed shape (lateral recess, dorsal and ventral process dimensions) is different in each taxon. The reported fusion of the prefrontal to the lacrimal in A. inexpectatum [3, 7] obscures the dorsal relationship between these elements. Unlike extant lissamphibians, albanerpetontids are known to have retained a jugal [6, 8,16] but dorsal reconstructions of the skull show it as a sliver of bone running between the maxilla and suspensorium. McGowan [8] reconstructed the jugal as overlapping the lateral surface of the squamosal, but this has not been confirmed from three-dimensional material of the squamosal. Venczel and Gardner [16] described several partial jugals of Albanerpeton pannonicum in association with the maxilla and lacrimal. Their most complete specimen is shown as extending only a short distance beyond the end of the maxillary tooth row, with its slightly expanded posterior edge bearing both medial and lateral facets. This would require that the tip slotted into a recess on the squamosal or was wedged between the squamosal and quadrate. However, in the Tetori specimen, the scans have revealed complete left and right jugals, the former lying under the frontal and the latter lying within the matrix in association with the right maxilla (Figs 2, 5, 6 and 17). The bone has the form of a curved bar, divided into a shallow anterior section and a vertical posterior blade. The anterior portion is tapered at its tip and consists of narrow lateral and medial flanges that straddled the dorsal surface of the maxilla. Dorsally, the anterior end bears a short facet for the lacrimal which overlapped both maxilla and jugal at this point. The posterior half of the bone is deeper in its mid section, but also tapers to a point posterodorsally, as seen on the left bone. Precisely how the jugal articulated with the suspensorium, if at all, is uncertain, as we could find no evidence of an articular surface on the posteroventral tip of the bone (as recorded for A. pannonicum [16]). From the reconstruction, it appears likely that the posterodorsal tip of the jugal approached the lateral tip of the parietal postorbital wing. However, there are no obvious facets on the postorbital wing or the tip of the jugal, and the latter may have been connected to the skull roof by a ligament. A rod-like element lying to the right of the frontal and right parietal (Figs 5±7) is interpreted as the right squamosal. It broadly resembles the squamosal preserved in the Las Hoyas Celtedens ibericus skull MCCM-LH15710 (Fig 18). As exposed on the Tetori block, the bone has the appearance of a slightly tapering blade but the μCT scans reveal a somewhat compressed cylinder, open along its medial side, with the posteroventral lamina slightly narrower than the anterodorsal one (Fig 19). The lateral margin is narrower than the open medial one and appears rugose but this edge is exposed on the block and the surface may be damaged. The 20 / 58 Fig 34. Shirerpeton isajii gen. et sp. nov., SBEI 2459, rib-like elements. Rib-like elements adjacent to vertebrae as segmented from μCT slice data in five views (original orientation uncertain). The dark grey elements are probably ribs or hyoid elements. The pale grey element is unidentified. Scale bar = 0.5 mm. 16±17, 20, 32±34] has led to a detailed understanding of many aspects of albanerpetontid morphology and a relatively stable phylogenetic framework for known taxa. Recently used data matrices [16, 34] coded 31 characters. Unfortunately, a majority of these (58%) relate to the premaxilla which is commonly preserved in microvertebrate deposits but is not preserved in the Japanese specimen. From the structure of the anterior margin of the nasal we infer that the nasal abutted the premaxilla medially but overlapped part of the premaxilla Fig 35. Shirerpeton isajii gen. et sp. nov., SBEI 2459, limb element. Limb element as segmented from μCT slice data in six views (original orientation uncertain). Scale bar = 1 mm. 37 / 58 laterally. This does not match any of the states described [16] for character 4. Given that we lack the premaxilla of the Japanese albanerpetontid, we have coded this as unknown rather than create an additional autapomorphic state. Of the remaining, non-premaxillary, characters, we were able to score all but one for Shirerpeton. Shirerpeton has a long premaxillary process of the maxilla (ch.15:0) (Fig 36C), lacks a dorsally projecting process on the dentary behind the tooth row (ch.16:0), and has the maxillary tooth row extending anterior to the level of the facial process (ch.20:0), all regarded as primitive states [34]. We were unable to score the presence or absence of labial sculpture on the maxilla with confidence as the irregularities on the surface of the maxilla may be artefacts of damage or reconstruction (ch.17:?), but sculpture seems to be absent on the dentary. However, Shirerpeton shares the presence of a convexity of the maxillary and dentary alveolar margin (ch.18:1) with Anoualerpeton and Albanerpeton nexuosum, although it differs from both in having the convexity preceded by a concavity that renders the dentary very shallow anteriorly. It also resembles Anoualerpeton spp. and Albanerpeton nexuosum in having larger teeth anteriorly (ch.19:1). With an estimated snout-vent length of around 45 mm, Shirerpeton also shows the derived condition (ch.25:1), as do A. arthridion and Wesserpeton. On frontal characters (Fig 37E), Shirerpeton shows the derived state for frontal shape (ch.21:1), having a frontal that is approximately triangular rather than bell-shaped; a proportion of frontal length to posterior width that lies at the borderline between long (ch.22:0) and moderate (ch.22:1); a long derived internasal process (ch.23:1); a derived condition of the ventrolateral crests (ch.24:1); a derived condition for the anterior limit of the orbital margin (ch.28:1), located posterior to the midpoint of the anteroposterior long axis; a derived tapered form for the internasal process (ch.29:1); and a weak median keel along the ventral surface (ch.31:1). However, character 29 in the matrix of Sweetman and Gardner [34] stems from the original description of Celtedens ibericus as having a `bulbous' nasal process [8, 74], a description based on interpretation of the articulated, but split, holotype specimen. An attributed specimen (MCCM-LH-15710) from the type locality, Las Hoyas, has the anterior end of the frontal disarticulated and clearly preserved (Fig 37B). It shows the nasal process to be a rounded taper rather than bulbous, and there is also greater separation between nasal and prefrontal facets than originally proposed. This raises doubts as to the original description. Given that the most recent data matrices for albanerpetontids are heavily dependent on premaxillary characters [16, 34], it is important to begin to identify additional features that allow a more comprehensive characterisation of the skeleton as a whole. As a start, we have added three parietal characters. The first (ch.32) relates to the length of the postorbital wing compared to the frontoparietal suture width. The second (ch.33) relates to the extent to which sculpture does or does not cover the postorbital wing. The third (ch.34) relates to whether the occipital shelf is single or bifurcate. We ran a parsimony analysis, using the character matrix of Sweetman and Gardner [34] with our additional characters (S1 File). The first analysis was run with the Branch and Bound option of PAUP (for direct comparison with the analysis of Sweetman and Gardner [34]) (see S1±S3 Figs). A second analysis used TNT (version 1.5, [75]), New Technology search option with Ratchet (20 iterations) and 1000 Random addition sequences, followed by a scrutiny of the resulting trees using Traditional search. In each case, we ran the analysis (a) with and without the all-0 hypothetical ancestor of Sweetman and Gardner [34], rooting either with a basal node (PAUP) or with Anoualerpeton priscum (TNT), and also (b) with and without character 29 as the derived state is an uncertain autapomorphy of Celtedens. Despite the differences between analyses, and the programme used (PAUP, TNT), the resulting tree topologies were closely similar. Fig 38 shows the strict consensus of 19 trees from a TNT analysis of the full matrix (34 characters) and with the hypothetical ancestor as the outgroup (L = 51), but the tree 38 / 58 Fig 36. Lateral profiles of maxillae in albanerpetontids (not to scale). A, Anoualerpeton priscum (Middle Jurassic, UK); B, An. unicum (Early Cretaceous, Morocco); C, Shirerpeton isajii (Early Cretaceous, Japan); D, Wesserpeton evansae (Early Cretaceous, UK); E, Albanerpeton arthridion (Early Cretaceous, North America); F, Albanerpeton gracile (Late Cretaceous, North America); G, Albanerpeton nexuosum (Late Cretaceous, North America); H, Albanerpeton galaktion (Late Cretaceous, North America); I, Albanerpeton inexpectatum (Miocene, France); J, Albanerpeton pannonicum (Pliocene, Hungary). Outlines redrawn or reconstructed from: A-B [20]; C, original, SBEI 2459; D [34]; E [26]; F-H [29]; I [72]; J [16]. Images C-H have been "mirror-imaged" for ease of comparison. topology was the same with character 29 deleted, and with the hypothetical ancestor replaced with Anoualerpeton. Shirerpeton is placed in an unresolved polytomy with Albanerpeton species. The same analyses run with PAUP (Branch and Bound), yielded 53 trees (L = 57) of which the Strict Consensus was identical to that in Fig 38. Examination of the individual trees 39 / 58 Fig 37. Comparison of frontal shape in albanerpetontids, scaled so that posterior widths are roughly equal. A, Anoualerpeton unicum (Early Cretaceous, Morocco); B, Celtedens ibericus (Early Cretaceous, Spain), based on [28]; C, cf. Celtedens ibericus (Early Cretaceous, Spain); D, Wesserpeton evansae (Early Cretaceous, UK); E, Shirerpeton isajii (Early Cretaceous, Japan); F, Albanerpeton arthridion (Early Cretaceous, North America); G, Albanerpeton gracile (Late Cretaceous, North America); H, Albanerpeton pannonicum (Pliocene, Hungary); I, Albanerpeton inexpectatum (Miocene, France). Figures redrawn from: A [20]; B [28]; C, original, MCCM-LH-15710; D [34]; E, original, SBEI 2459; F [26]; G [29]; H [16]; I [72]. 40 / 58 Fig 38. Strict consensus of 19 MPT's from an analysis of the full matrix (NT search in TNT with Ratchet followed by a Traditional search of the resulting trees in RAM). Of the 19 MPT's, 11% placed Shirerpeton as the sister taxon to a monophyletic Albanerpeton; 42% placed it as the sister taxon to A. arthridion; and 47% placed it crownward of A. arthridion (see Fig 39). An analysis with the limited matrix and without Hypothetical Ancestor yields trees with the same topology for the in-group taxa. from both analyses identified three alternative positions for Shirerpeton (Fig 39): i) as sister group to Albanerpeton as a whole; ii) as sister taxon to A. arthridion; or iii) in variable positions crownward of A. arthridion. The first of these arrangements was the rarest (15%), with the remaining trees showing the second or third topology in roughly equal numbers. However, a Bootstrap analysis (2000 replicates) using TNT (NT + Ratchet [20 iterations]) yielded a more resolved tree (Fig 40) resembling that of Sweetman and Gardner [34] in recovering a `robustsnouted clade' comprising A. nexuosus, A. pannonicum, A. inexpectatum, and the Paskapoo taxon, and a `gracile-snouted clade' comprising A. gracile, A. galaktion, and A. cifellii, with A. arthridion as the sister taxon to both clades, and Wesserpeton, the Uña taxon, Celtedens, and Anoualerpeton as successive outgroups. Shirerpeton is nested within Albanerpeton as the sister taxon of the `robust-snouted clade', but Bootstrap support values for all nodes are very low and only the `robust-snouted clade' survived a decay analysis beyond one step. Moreover, when synapomorphies are followed on the tree, none unequivocally supports the placement of Shirerpeton as the sister taxon of the `robust-snouted clade' and only two frontal characters (ch.22, 28) support the placement of Shirerpeton crownward of Wesserpeton. 41 / 58 Fig 39. The three configurations recovered within the individual trees of the different analyses described in the text. In (C) the position of Shirerpeton is unstable with respect to individual Albanerpeton sp. The `A. inexpectatum clade' consistently comprises A. inexpectatum, A. pannonicum and the Paskapoo taxon, with the variable addition of A. nexuosum. In summary, the results of the phylogenetic analyses consistently support the placement of Shirerpeton crownward of Anoualerpeton and Celtedens, and provide more limited support for a position crownward of the Uña taxon and Wesserpeton, in a closer relationship with species currently referred to the genus Albanerpeton. This is discussed further below. Discussion Novel morphology SBEI 2459 preserves the disarticulated, but associated, bones of a single small individual. Although a few surface elements have suffered damage, individual bones are largely uncrushed 42 / 58 Fig 40. Bootstrap analysis (1000 replicates) using TNT (NT search + Ratchet [20 iterations]) of the full matrix, with the hypothetical ancestor as outgroup. The same analysis run using the limited matrix yielded the same topology, with minor differences in the Bootstrap values. and the μCT scans have revealed significant new information on albanerpetontid cranial morphology. The specimen preserves several elements that were previously either unknown or incompletely known from other albanerpetontid taxa±notably the nasal, septomaxilla, prefrontal, lacrimal, jugal, quadrate, squamosal, parietal, supraoccipital, sphenoid, and other braincase components. Moreover, some of the unidentified elements suggest that the albanerpetontid skull may have retained other elements subsequently lost in crown Lissamphibia (e.g. the supratemporal). The albanerpetontid rostrum was first reconstructed by Estes & Hoffstetter [3] on the basis of disarticulated, but well-preserved, bones of the Miocene species, A. inexpectatum. The nasals were not preserved but, based on surrounding bones, the authors reconstructed them as small oval elements lying between the frontal, premaxillae, and prefrontals, but excluded from the narial margins. The prefrontal was inferred to be fused to the lacrimal, with some uncertainty as to the limits of each bone. Gardner [72±73] broadly agreed with this interpretation. However, on the basis of articulated Early Cretaceous material of Celtedens ibericus from Las Hoyas, Spain, McGowan [8] concluded that small ovoid nasals `did not appear to exist', and 43 / 58 re-interpreted Estes and Hoffstetter's [3] fused prefrontals as elongate, slender nasals. In McGowan's interpretation, the posterolateral facet on the frontal accommodated the lacrimal, not a prefrontal, with the nasal articulating further anteriorly and entering the narial margin. However, the Las Hoyas specimens are flattened and are usually split between part and counterpart, complicating interpretation (SEE pers. obs.). The rostral morphology of Celtedens ibericus is therefore uncertain. More recently, three-dimensionally preserved and partly articulated material of A. pannonicum [16] clarified the relationships of the nasals, prefrontals, lacrimals, and frontal to one another, at least in Neogene Albanerpeton. SBEI 2459 supports the interpretations of Estes & Hoffstetter [3] and of Venczel & Gardner [16], rather than McGowan [8]. It demonstrates the presence of a discrete prefrontal that articulated with the frontal medially and lacrimal laterally, and of a separate nasal. As in A. pannonicum, the nasals of Shirerpeton were relatively large. Our reconstruction (Fig 3) suggests that the nasals were separated for most of their length by the internasal process of the frontal, but entered the narial margins and contacted the lacrimals to exclude the prefrontals from those margins. In shape (relative angles of the dorsal and lateral processes, position and depth of the lateral groove), the lacrimal of Shirerpeton more closely resembles that of A. inexpectatum than A. pannonicum, allowing for the non-fusion of the prefrontal (Fig 41). A small septomaxilla has also been revealed in Shirerpeton. This bone is variably present in members of all extant lissamphibian clades [68] but has not previously been described in an albanerpetontid. The quadrate of Albanerpeton inexpectatum was mentioned briefly by previous authors [ 2, 3 ], and figured at small scale. In Shirerpeton, the facets suggest that the quadrate had a firm articulation with the squamosal and pterygoid. SBEI 2459 has also confirmed the presence of discrete jugals in albanerpetontids, as in Eocaecilia [76] but not extant lissamphibians. However, unlike previous reconstructions [8] that present the jugal as a straight bar between the maxilla and suspensorium, in Shirerpeton the jugal has a posterodorsal curvature and may have had a ligamentous attachment with the skull roof. There are no obvious posteroventral facets or processes to indicate how (or whether) the jugal contacted the suspensorium. Although the albanerpetontid parietal is probably as taxonomically distinctive in its shape and ornamentation as the frontal, it has only rarely been preserved and/or described (Fig 42). Estes & Hoffstetter [3] provided images of the almost complete parietals of A. inexpectatum, showing the bones to be paired, with an acuminate lateral postorbital wing, a sculptured triangular central area, and unsculptured occipital shelves (Fig 42E). The parietals of Celtedens ibericus [6,8] have a similar overall shape, although the details are less clearly preserved (Fig 42C). For the remaining taxa, however, the parietal is either unknown or fragmentary. McGowan [23] figured the anterior portion of a right parietal from the Middle Jurassic of Britain, referable to Anoualerpeton priscum (Fig 42A), and Wiechmann [71] figured fragments of two left bones of `Celtedens' guimarotae (not reproduced here as the figures are difficult to interpret). Sweetman & Gardner [34] did not describe the parietal of Wesserpeton, but a partial (anterior) left bone was recovered from the type deposit (Natural History Museum, London, NHMUK PV R 36956, Fig 42B). The Tetori holotype specimen (Figs 3 and 42D) is therefore exceptional in preserving both parietals almost in their entirety. They differ from the known parietals of other taxa in three major respects: the postorbital wing is relatively longer; there is a more limited lateral extension of the sculpture so that it does not encroach on the postorbital wing; and the occipital shelf is shorter and bifurcated. This bifurcation is clearest on the right bone where part of a vertebra extends between the posteromedial and posterolateral rami. The detailed three-dimensional preservation of the parietals in Shirerpeton also demonstrates, for the first time, how they articulated with the supraoccipital and epipterygoids. 44 / 58 Fig 41. Skull roofing bones in albanerpetontids (not to scale). A, Shirerpeton isajii; B, Celtedens ibericus; C, Albanerpeton inexpectatum; D, Albanerpeton pannonicum. B-D, redrawn from [16], but with B based on [8]. Perhaps the greatest differences in morphology between Shirerpeton and the European Neogene taxa are found in the braincase. Estes and Hoffstetter [3] figured stereopairs of an isolated three-dimensional braincase of A. inexpectatum, but these images are too small to yield any detail. Moreover, this specimen, formerly in the collections of the Museum National d'Histoire Naturelle, Paris, is now apparently lost [13]. Nonetheless, it is clear from the images and the description that the component elements of the braincase were fully conjoined. A second three-dimensional braincase, attributed to the Pliocene A. pannonicum, has also been described [13]. Again, the components are co-ossified, leaving the homologies of some regions uncertain. Although Maddin et al. [13] reported the presence of an ossified tectum synoticum in A. pannonicum, with an anterior projection, the lack of sutures left the authors unable to determine whether this ossification represented a supraoccipital. In Shirerpeton, however, it is clear that the supraoccipital is an independent ossification. Neither of the Neogene braincases 45 / 58 Fig 42. Parietal morphology in albanerpetontids (not to scale). A, Anoualerpeton priscum (Middle Jurassic, UK); B, c.f. Wesserpeton evansae (Early Cretaceous, UK); C, Celtedens ibericus (Early Cretaceous, Spain); D, Shirerpeton isajii, Early Cretaceous, Japan); E, Albanerpeton inexpectatum, Miocene, France). In each case, the small asterisk marks the lateral limit of the frontal facet, lateral to which is the postorbital wing. Abbreviations: Occ.s, occipital shelf; PoW, postorbital wing. Figures redrawn from: A [23] (mirror-imaged for comparison); B, original, NHMUK PV R 36956; C [6]; D, original, SBEI 2459; E [3]. has an anteromedian process as long as that of Shirerpeton, but A. pannonicum preserves a short process that may be incomplete. No living lissamphibian has a discrete supraoccipital ossification, although De Beer [70] suggested that it might be included within the os basale of caecilians. A supraoccipital ossification is also absent in putative stem-lissamphibians like the temnospondyl Doleserpeton [77], but it is present in the lepospondyl Brachydectes [78] where the element is remarkably similar to that of Shirerpeton. Ventrally, Shirerpeton lacks a median basioccipital but the paired exoccipital plates presumably had the same embryonic origin. Whether the notochord was enclosed by the plates, or overlay them in the midline is unclear. Maddin et al. [13] postulated the presence of a transverse suture between anterior and posterior parts of the braincase floor in both A. inexpectatum and Celtedens ibericus. This is also the case in Shirerpeton, although our examination of the key Las Hoyas specimen (Museo de las Ciencias de Castilla-La-Mancha, Cuenca, Spain, MCCM-LH-15710) showed that labelled suture in this specimen [13] is actually a break within the exoccipital plate, with the other part of this element preserved on the counterpart (blue outlines in Fig 43). Moreover, the more anterior element interpreted by [13] as the dorsum of the braincase, is actually a sphenoid like that of Shirerpeton (red outlines in Fig 43). In 46 / 58 Fig 43. Celtedens ibericus, Las Hoyas specimen, Museo de las Ciencias de Castilla-La-Mancha, MCCM-LH-15710, digital photograph of skull of part and counterpart. A,B MCCM-LH-15710b, with A, whole skull and B, enlargement of braincase region; C,D, MCCM-LH-15710a, with C, whole skull and D, enlargement of braincase region. In B and D, blue outlined element is the exoccipital plate, and red outlined element is the sphenoid. Scale bar = 1 mm. A. pannonicum, Maddin et al. [13] concluded that the ventral floor of the neurocranium was `composed largely of the parasphenoid' with `minor contributions at the posterolateral corners from the otic capsules'. This is very different from the condition in Shirerpeton where the floor is clearly formed by the exoccipital plates. The configuration of the parasphenoid remains 47 / 58 unknown. In living frogs and salamanders, the exoccipital plates do not generally meet in the midline, but this may not apply to the solidly fused os basale of caecilians [70]. Both previous accounts of the albanerpetontid braincase [3, 13] recorded the presence of one or more hypoglossal foramina piercing the exoccipital. Foramina are also present in the Tetori specimen but we were unable to follow a clear path through the bone and therefore the identification of these foramina remains uncertain. The albanerpetontid palate remains an enigma. One small element in the Tetori skull has been tentatively identified as a possible ectopterygoid or palatine but the other ventral components of the skull (parasphenoid, pterygoids, vomers, elongated median hyoid element) are not preserved. It seems likely that these ventral skull components were displaced either laterally or anteriorly and were on the part of the block that was lost before the dorsal elements were exposed. Systematic position of Shirerpeton within Albanerpetontidae As most albanerpetontid material is disarticulated, the elements most frequently used in diagnosis are the premaxilla and the distinctive median frontal [6, 30, 34, 74]. The oldest, and apparently more plesiomorphic, taxa (Anoualerpeton, Celtedens) have a frontal that has been described as `hour-glass' or `bell-shaped' [6, 8, 73±74], in that the bone is significantly longer than it is wide, and the lateral orbital borders are concave (Fig 37A±37C). The anterior limit of the frontal orbital margin (posterior edge of the prefrontal facet) lies anterior to the mid-point of the anteroposterior long axis. The second frontal morphotype, which is considered to characterise the derived genus Albanerpeton [28, 30, 34, 72], has lateral borders that are straight (or nearly so) and a posterior margin that is significantly wider than the anterior one (Fig 37F± 37I). The anterior margin of the orbit lies at or posterior to the mid-point of the long axis of the frontal. The frontal length and width are roughly equal, so that the bone is strongly triangular, a shape enhanced by an acuminate internasal process that may be flanked by smaller anterolateral processes. However, length/width proportions differ considerably between early species like A. arthridion and derived Neogene species such as A. inexpectatum and A. pannonicum. The Barremian genus Wesserpeton [34] has a frontal that is intermediate in morphology between the two main types, in that it is triangular and acuminate like that of Neogene Albanerpeton, but is somewhat longer in relation to its posterior width, and the anterior margin of the orbit is roughly level with the mid-point of the frontal long axis (Fig 37D). As reconstructed, the frontal of Shirerpeton has a length/posterior width proportion of 1.23, as compared to 0.87±1 for A. inexpectatum [72]; 0.88±0.95 for A. pannonicum [16]; ~1.2 in A. gracile, A. nexuosum, A. galaktion [29, 30], and A. arthridion [26]; 1.0±1.3 for Wesserpeton [34]; 1.3±1.4 for Celtedens ibericus (depending on the reconstruction used); and 1.57 for Anoualerpeton unicum and An. priscum [20]. Of all species currently known, the reconstructed frontal shape and proportions of Shirerpeton are most like those of the North American A. arthridion (Fig 37F), the oldest (uppermost Aptian or lower Albian) species referred to Albanerpeton, although the internasal process is longer in the Japanese taxon. Estes & Hoffstetter [3] figured the parietals of A. inexpectatum which, although slightly damaged posteriorly, clearly differ in shape from those of the Japanese taxon in that the unsculptured posterior occipital shelf is long and wide rather than bifurcate. There is also a difference in proportion in relation to the width of the fronto-parietal contact and the length of the postorbital wing. In A. inexpectatum, the postorbital wing is relatively short (Fig 42E) and the sculpture extends almost to its tip. In Shirerpeton, the wing forms almost half the anterior width of the bone, and is unsculptured (Fig 42D). The parietals are also preserved in Celtedens ibericus [6, 8], but are crushed. The postorbital wings of Celtedens, as reconstructed (Fig 42C), 48 / 58 are proportionally longer than those of Shirerpeton (more than 58% parietal width), but bear sculpture as in A. inexpectatum. The same condition is seen in the few fragmentary parietals preserved for the Middle Jurassic Anoualerpeton priscum [23], Fig 42A), and in the parietal fragment attributed to Wesserpeton evansae (Fig 42B). Unfortunately, the tip of the wing is missing in both UK taxa, but (from the angle of the converging margins) it does not appear to have been long. A long, unsculptured postorbital wing may therefore be a diagnostic feature of Shirerpeton, but without comparable elements from A. arthridion and members of the `gracilesnouted' clade, this remains tentative. Of the jaw elements, comparison between albanerpetontid species has focused mainly on the premaxilla, an element that remains unknown in Shirerpeton. As listed [34], maxillary characters have been limited to the presence or absence of labial sculpture; the position of the anteriormost tooth relative to point of maximum indentation of the narial margin; and the presence or absence of an expansion of the labial dental margin in association with enlarged teeth. Shirerpeton shows the primitive state in the first two of these, and shares the heterodonty and expanded dental margin with species of Anoualerpeton and Albanerpeton nexuosum. However, as Fig 36 demonstrates, there are other maxillary features that vary between species. These include the length and shape (rounded or pointed) of the anterior premaxillary process; the degree of concavity of the anterior narial margin; and the shape of the dorsal facial process in terms of angulation and profile. The difficulty is that few complete maxillae have been described and most sample sizes are small. On the basis of existing material, the maxilla of Shirerpeton (Fig 36C) appears to be distinct in combining heterodonty with a relatively short, pointed premaxillary process, little curvature of the narial margin, and a low, rounded facial process. Even fewer dentary characters have been identified, other than those relating to labial sculpture and heterodonty, and the presence of a dorsal process posterior to the tooth row, currently reported only in Albanerpeton inexpectatum [34]. Again, Shirerpeton most closely resembles Anoualerpeton and Albanerpeton nexuosum in having enlarged anterior teeth supported by an expansion of the labial dental margin, although it differs in the greater degree of sinuosity of the margin due to the concave-convex-concave profile. The dentary of Shirerpeton also appears to differ in two other ways. The anterior limit of the prearticular facet seems to be in line with the rear of the tooth row and posterior to the opening of the Meckelian fossa, whereas it extends forward below the tooth row to a point roughly in line with the opening of the Meckelian fossa in most other taxa where this region is known. Secondly, in many albanerpetontids (including A. nexuosum and the two species of Anoualerpeton), the subdental shelf rises steeply at the posterior end of the tooth row such that the posterior tooth loci are significantly smaller (25% or less) than those close to the symphysis. In Shirerpeton, the subdental shelf is shallower and, apart from the group of enlarged teeth, there is less of a difference in size between anterior and posterior teeth. Thus the combined characters of the maxilla and dentary in Shirerpeton allow it to be distinguished from other albanerpetontid taxa known only from dental elements. This includes elements from the Upper Cretaceous of Central Asia attributed to `Nukusurus' [31], which lack heterodonty and have an anteriorly extended prearticular facet and very small posterior teeth. Taken together with the morphological features, the results of the phylogenetic analyses provide some support for the placement of the Japanese albanerpetontid crownward of the European Wesserpeton and the Uña taxon (sensu [34]). Although the Strict Consensus (PAUP, TNT) is partially unresolved (Fig 38), only a minority of individual trees placed the Japanese taxon as the sister group to a monophyletic Albanerpeton. Of the remaining trees, roughly half placed the Japanese albanerpetontid as the sister taxon to A. arthridion, which it resembles in frontal morphology and small size, but not dentary morphology (proportionally deeper anteriorly in A. arthridion, without the strongly sinuous alveolar margin [26]). The other half of the 49 / 58 individual trees placed the Japanese taxon crownward of A. arthridion, but of unstable position in relation to other included species. The Bootstrap analysis (Fig 40) placed Shirerpeton on the stem of the `robust-snouted' clade. These topologies raise the question as to whether the Japanese species should be attributed to Albanerpeton. The type species of Albanerpeton is the European Miocene A. inexpectatum [3]. Although the first albanerpetontids (represented by dentaries) recovered in North America were placed in the genus Prosiren [ 1 ], the designated holotype of Prosiren was a vertebra later shown to belong to a salamander [12]. Based on their unique structure (tooth and symphysial morphology) the dentaries originally attributed to Prosiren [ 1 ] were re-assigned to the distinctive European genus Albanerpeton [12], as was all subsequent North American material. In the interim, however, there have been both temporal range extensions (Middle Jurassic to Pliocene), and a recognition of generic diversity among European Mesozoic albanerpetontids (Anoualerpeton, Celtedens, Wesserpeton). Compared to these relatively short-lived genera, Albanerpeton as currently defined extends from the? Aptian-Albian to the late Pliocene, an interval of more than 100 Myr. Given the difference in frontal shape between, for example, A. arthridion and A. inexpectatum (Fig 37), it seems likely that there is greater taxonomic diversity within Albanerpeton than is currently recognised by the nomenclature. This applies particularly to the earliest North American species, and the unstable placement of the Japanese taxon in relation to these species (especially A. arthridion) may partly reflect this. Resolution of this problem would be advanced by the recovery and description of additional elements (especially parietals) for the North American species, as current diagnoses and phylogenies rely too heavily on premaxillary characters. Based on the above discussion, we consider it preferable to give the new Japanese albanerpetontid separate generic status. This reflects i) its distinctive character combination; ii) the striking differences in braincase morphology between Shirerpeton and the Neogene Albanerpeton species; iii) the unstable phylogenetic position of the Japanese taxon in relation to existing Albanerpeton species; iv) a lack of overlap of preserved elements, resulting in more than 50% missing data when coded into existing matrices (mainly due to a lack of premaxillae); v) the very limited knowledge of albanerpetontid parietal and braincase morphology; and vi) the fragmentary nature of most of the North American material. The separate taxonomic identity also facilitates an objective discussion of albanerpetontid biogeography. Systematic position of Albanerpetontidae among tetrapods Discussion of the systematic position of albanerpetontids amongst tetrapods is complicated by the lack of agreement on the evolutionary relationships and monophyly of lissamphibians [14, 79±80], and by a lack of information on the early history of each major group. Triadobatrachus [77, 81±82] and Czatkobatrachus [83±85] provide some information on early frog evolution, but it is very incomplete, as is that for early caudates [86±89]. Eocaecilia is more completely known [5, 76] but, until the recent description of the Triassic Chinlestegophis [78] as a possible stem-caecilian, it had been isolated in time. The recovery of new material pertaining to the early history and divergence of lissamphibians is clearly crucial. Albanerpetontids have the potential to shed light on early lissamphibian morphology, but in the current state of phylogenetic flux, a detailed analysis of albanerpetontid relationships to other tetrapods is beyond the scope of the current paper. Existing data matrices differ quite strikingly in their coding of the same taxa (e.g. compare the codings in [5, 14] vs [79]) and yield conflicting results, even when the same approaches are used. Nonetheless, the presence of epipterygoids and a separate supraoccipital in albanerpetontids argues against a nested position within Batrachia (as a caudate sister group) as some previous authors have suggested 50 / 58 [1±5]. Moreover, although there is no question as to the monophyly of Albanerpetontidae, the differences in braincase anatomy between Shirerpeton and the Neogene species of Albanerpeton provide a clear argument against representing Albanerpetontidae as a single operational unit in phylogenetic analyses as has often been the case [5, 10, 79, 90]. Shirerpeton is separated from A. inexpectatum and A. pannonicum by around 100 million years, but is itself separated by a similar time interval from the likely Permo-Triassic ancestral albanerpetontid of which nothing is known. Biogeography Whatever the precise relationship between albanerpetontids and the three living lissamphibian clades, their last common ancestor must have lived in the Paleozoic. This leaves a long hiatus before the first recorded appearance of albanerpetontids in the Middle Jurassic, limiting discussion as to their place of origin. The Jurassic record is currently restricted to Europe (France, England, Portugal [21, 22, 71, 91]) and North Africa (Morocco [18]). Most Early Cretaceous records are also European (Britain [34, 92], France [ 93 ], Italy [ 94 ], Spain [6], Sweden [ 95 ]) and the group is represented in Europe for much of the Late Cretaceous (Spain, France, Hungary, Romania, [ 96 ]), as well as intervals of the Palaeogene (Oligocene, Germany [17]) and Neogene (France, Germany, Czech Republic, Austria, Hungary, Italy [17]). Albanerpetontids survived in Europe until at least the late Pliocene (Italy [24]) where they are associated with the remains of extant reptile and amphibian genera. There are no confirmed records of albanerpetontids in North America before the late Early Cretaceous and this coincides with the first occurrence of specimens currently referred to Albanerpeton [17, 26]. Gardner's phylogenetic work [16±17, 28, 34] recognises A. arthridion from the Cloverley and Antlers formations of Oklahoma and Texas, USA, as the sister taxon to all other (younger) Albanerpeton species and, Shirerpeton aside, our analyses support this placement. The precise age of the oldest horizons bearing A. arthridion are uncertain, as radiometric dates are lacking, but they are generally given as uppermost Aptian or lowermost Albian [ 26, 97 ]. Albanerpeton then apparently radiated in North America through the Late Cretaceous (at least seven attributed species), with records in every stage from the Cenomanian to the Maastrichtian [17]. Specimens from the middle to late Palaeocene of Canada [17] show that the group survived the end-Cretaceous extinction in North America, but there are currently no younger records. In Europe, the genus Albanerpeton is first reported from the Late Cretaceous (Campanian) of France, but it is important to note that many Late Cretaceous European albanerpetontid records are based on fragmentary specimens that are really diagnostic only at family level. The two named European species, the Miocene A. inexpectatum and Pliocene A. pannonicum, fall within a single well-supported clade [17] with material from the Palaeocene Paskapoo Formation of Canada. Up until recently, therefore, the distribution pattern was consistent with the hypothesis that the genus Albanerpeton originated in North America, around the mid-Cretaceous, and subsequently spread to Europe. However, as Gardner & Bőhme [17] observed, the expanding record of albanerpetontids in Europe weakens that argument, as does the absence of any confirmed record of albanerpetontids in North America prior to the latest Aptian/earliest Albian. This apparent absence may be an artefact of the fossil record, there being no Middle Jurassic or Berriasian-Barremian microvertebrate faunas known from North America. However, albanerpetontids are also conspicuously absent from the Upper Jurassic Morrison Formation despite the existence of several rich microvertebrate localities where they would be expected to occur (based on other components of the faunal assemblage). It is therefore possible that albanerpetontids first reached North America during the Early Cretaceous but, if so, from where [26]? 51 / 58 One possibility is that derived albanerpetontids spread to North America from Europe via a North Atlantic route, as suggested for some mammals [98±99]. This hypothesis has received support from the description of the European Wesserpeton and its placement as a sister taxon to Albanerpeton. Sweetman & Gardner [34] suggested that the split between the ancestral lineages of these two genera might have occurred in western Eurasia, with a Wesserpeton-like ancestor subsequently migrating into North America via Europe and giving rise to Albanerpeton. They further proposed (based on Wesserpeton) that `acquisition of a triangular (frontal) shape preceded shortening of the bone and that this evolutionary trait was acquired in western Eurasia some time prior to the Barremian' [34]. An alternative route into western North America, where records are concentrated, would be from eastern Asia via the Beringian land bridge. This route has been proposed for several groups of dinosaurs, lizards, mammals and choristoderes that appeared in North America toward the end of the Early Cretaceous [100±103]. However, some authors [17, 31±32] have argued against that scenario for albanerpetontids, on the basis that the first record in North America (A. arthridion) both predated the opening of the Beringian land bridge and predated any record of albanerpetontids in Asia. The first objection is moot as there is uncertainty both on the precise dating of the relevant American deposits (e.g. [ 97 ]) and of the opening of the land bridge (e.g. [101±104]). Nonetheless, up until now, the second objection stood, because the earliest confirmed Asian records of albanerpetontids were from the early Cenomanian of Uzbekistan [31±33,105±106]). To explain their apparently late appearance in Asia, Gardner & Averianov [31] and Skutschas [32] favoured a mid-Cretaceous dispersal of albanerpetontids from North America into Asia. Clearly, the recovery of a relatively derived albanerpetontid in the Early Cretaceous of Japan is not consistent with that scenario. It is plausible that the ancestral Wesserpeton-like stock proposed by Sweetman & Gardner [34] could have spread eastwards rather than (or as well as) westwards from western Eurasia, although the potential for interchange may have been limited by the opening of the Turgai Straits between Asia and western Eurasia from the late Middle Jurassic onward [ 31, 107 ]. Moreover, none of the Middle Jurassic assemblages of Russia and Central Asia has yielded conclusive evidence of albanerpetontids [ 33, 108 ]. This is surprising given the general resemblance of these assemblages to the contemporaneous faunas of Britain where albanerpetontids do occur [22]. An albanerpetontid frontal was reported [ 106 ] from the Middle Jurassic Balabansai Formation of Kirghizia, but this observation cannot be confirmed as the specimen was neither figured nor described and cannot be located ([31]. It remains possible, of course, that the apparent absence of the group at some localities is an artefact of local environmental or taphonomic factors [ 109 ]. Albanerpetontids have yet to be recovered from the fossil rich Chinese Daohugou Beds (Jurassic) or the Jehol Biota (Early Cretaceous), yet the discovery of Shirerpeton in Japan, then part of the eastern Asian mainland, renders it highly likely that the group was present in China and other parts of Asia at this time. Recent isotope analysis of reptile bone suggests that the Early Cretaceous Yixian Formation, at least, was deposited under similar (cool) climatic conditions (mean air temperature of ~10ÊC, [ 110 ]) to those of the Kuwajima Formation, but perhaps the lacustrine environments that have been richly sampled in China were less suitable for small terrestrial albanerpetontids. Most Mesozoic albanerpetontids are found in lowland estuarine or swampy environments [17]. The Japanese Kuwajima Formation conforms to this pattern (swampy inland wetland), but albanerpetontid remains are extremely rare (to date, 3 out of 2459 catalogued specimens) and it seems likely that they entered the deposit from further afield, possibly in the gut of a predator. One further chapter in the history of Asian albanerpetontids is currently being written. Daza et al. [111] recently described a collection of lizard fossils from an amber deposit in Myanmar (Burma) dated to Albian-Cenomanian age. The image of one very small specimen, 52 / 58 tentatively identified [111] as an early chamaeleonid, was re-identified by one of us (SEE) as an albanerpetontid. μCT has confirmed this identification and work on that specimen is ongoing. It provides an important temporal link between the Japanese specimen and the younger material from Uzbekistan, and offers support for the idea that albanerpetontids were well established and relatively widespread across at least eastern and south-eastern Asia during the Cretaceous. Conclusions The recovery of the new Japanese specimens sheds new light on albanerpetontid morphology and biogeography, but raises as many questions as it resolves. There is clearly much more to discover about these enigmatic little tetrapods, in terms of their morphology, relationships, and evolutionary history. Recent discoveries have demonstrated that the group had a more extensive temporal and geographical distribution in Asia than previously understood. Awareness of this among researchers may lead to further discoveries. Supporting information S1 Fig. Strict consensus of 53 trees run in PAUP using the full matrix, with the hypothetical ancestor as outgroup. This tree topology matches that recovered from the TNT analysis shown in Fig 34. Of the 53 individual MPTs, 15% placed Shirerpeton as the sister taxon to a monophyletic Albanerpeton; 45% placed it as the sister taxon to A. arthridion; and 40% placed it crownward of A. arthridion. (TIF) S2 Fig. PAUP analysis of the full matrix run without the hypothetical ancestor. Left, Bootstrap Analysis; right, 70% Majority Rule Tree of 53 individual trees. (TIF) S3 Fig. Bootstrap analysis using PAUP of the limited matrix. There is less resolution with respect to Wesserpeton and the Uña taxon. (TIF) S1 File. Characters and data matrix used in the phylogenetic analysis. (DOCX) Acknowledgments Our thanks to Yuki Iwama (Nagoya Municipal Industrial Research Institute) and Akira Monkawa (Tokyo Metropolitan Industrial Technology Research Institute) for μCT scanning; Reiko Kohno, Archaeology Department National Science Museum, Tokyo, for access to a 3D printer; Akemi Wakimoto, Kento Otsuka, and Tsuyoshi Hibino (Hakusan City Board of Education, Ishikawa Prefecture) for access to the specimens described here and much of the initial preparation; Steven Sweetman (University of Portsmouth) for access to the partial parietal of c.f. Wesserpeton and for donating it to the Natural History Museum, London; Filip Weichmann (Free University Berlin) and Pavel Skutschas (St Petersburg State University) for information on Guimarota and Russian material respectively; Angela Delgado (Autonoma University Madrid) for arranging the loan and transportation of MCCM-LH-15710 from the Museo de las Ciencias de Castilla-La-Mancha (Cuenca); Juan Daza (Sam Houston State University) for confirming our identification of the Myanmar amber specimen; Nick Crumpton (University College London) and Masatoshi Okura (Aichi Prefecture) for taking the digital images; and three reviewers (Annelise Folie, Marton Venczel, James Gardner) for helpful comments on an 53 / 58 earlier version of the manuscript. Yui Inagaki, a Junior High School student from Kanagawa City, found the second albanerpetontid jaw while the original manuscript version was in review: our special thanks to her. We are also indebted to the mayors and administrations of Kuwajima District (Hakusan City, Ishikawa Prefecture, Japan) and Ishikawa Prefecture for their generous hospitality and support over many years. We also acknowledge the Willi Hennig Society for access to the phylogeny programme TNT. Author Contributions Conceptualization: Susan E. Evans. Investigation: Ryoko Matsumoto, Susan E. Evans. Methodology: Ryoko Matsumoto. Writing ± original draft: Susan E. Evans. Writing ± review & editing: Ryoko Matsumoto, Susan E. Evans. 3. Estes R, Hoffstetter R. Les urodeÂles du Miocène de la Grive-St.Alban (Isère, France). Bull Mus Natl Hist Nat, Paris 1976; 57:297±343. 4. Trueb L, Cloutier R. A phylogenetic investigation of the inter- and intra-relationships of the Lissamphibia (Amphibia, Temnospondyli). In: Schultze H-P, Trueb L, editors. Origins of the higher groups of tetrapods: controversy and consensus. Ithaca: Cornell University Press;1991. p.223±313. 5. Maddin HC, Jenkins FA Jr, Anderson JS. The braincase of Eocaecilia micropodia (Lissamphibia, Gymnophiona) and the origin of caecilians. PLoS ONE 2012 Dec 5; 7(12):e50743. https://doi.org/10. 1371/journal.pone.0050743 PMID: 23227204 6. McGowan G, Evans SE. Albanerpetontid amphibians from the early Cretaceous of Spain. Nature 1995; 373:143±45. 7. Gardner JD. Monophyly and the affinities of albanerpetontid amphibians (Temnospondyli; Lissamphibia). Zool J Linn Soc. 2001; 131:309±52. 8. McGowan GJ. Albanerpetontid amphibians from the Lower Cretaceous of Spain and Italy: a description and reconsideration of their systematics. Zool J Linn Soc. 2002; 135:1±32. 9. Ruta M, Jeffery JE, Coates MI. A supertree of early tetrapods. Proc R Soc Lond B Biol Sci. 2003; 270: 2507±16. 10. Ruta M, Coates MI. Dates, nodes and character conflict: addressing the lissamphibian origin problem. J Syst Palaeontol. 2007; 5:69±122. 11. Marjanović D, Laurin M. A re-evaluation of the evidence supporting an unorthodox hypothesis on the origin of extant amphibians. Contr Zool. 2008; 77(3): 149±199. 12. Fox RC, Naylor BG. A reconsideration of the relationships of the fossil amphibian Albanerpeton. Can J Earth Sci. 1982; 19:118±28. 13. Maddin HC, Venczel M, Gardner JD, Rage J-C. Micro-computed tomography study of a three-dimensionally preserved neurocranium of Albanerpeton (Lissamphibia, Albanerpetontidae) from the Pliocene of Hungary. J Vertebr Paleontol 2013; 33:568±87. 14. Anderson JS. Focal Review: the origin(s) of modern amphibians. Evol Biol. 2008; 35:231±47. 15. Evans SE. Albanerpetontidae. In: Poyato-Ariza FJ, Buscalioni AD, editors. Las Hoyas: a Cretaceous wetland. Munich: Verlag Dr Friedrich Pfeil; 2016. p.133±7. 16. Venczel M, Gardner JD. The geologically youngest albanerpetontid amphibians from the lower Pliocene of Hungary. Palaeontology 2005; 48:1273±1300. 17. Gardner JD, Bőhme M. Review of the Albanerpetontidae (Lissamphibia) with comments on the paleoecological preferences of European Tertiary albanerpetontids. In: Sankey JT, Baszio S, editors. Vertebrate microfossil assemblages. Bloomington and Indianapolis: Indiana University Press; 2008.p.178± 218. 54 / 58 18. Haddoumi H, Allain R, Meslouh S, Metais G, Monbaron M, Pons D, et al. Guelb el Ahmar (Bathonian, Anoual Syncline, eastern Morocco): first continental flora and fauna including mammals from the Middle Jurassic of Africa. Gondwana Res. 2016; 29:290±319. 19. Sigogneau-Russell D, Evans SE, Levine JF, Russell DA. The Early Cretaceous microvertebrate locality of Anoual, Morocco: a glimpse at the small vertebrate assemblages of Africa. Bull New Mex Mus Nat Hist Sci. 1998; 14:177±81. 20. Gardner JD, Evans SE, Sigogneau-Russell D. New albanerpetontid amphibians from the Early Cretaceous of Morocco and Middle Jurassic of England. Acta Palaeontol Pol. 2003; 48:301±19. 21. Seiffert J. Urodelan atlas aus dem obersten Bajocien von S.E. Aveyron (SuÈdfrankreich). Palaontol Z. 1969; 43:32±6. 22. Evans SE, Milner AR. Microvertebrate faunas from the Middle Jurassic of Britain. In: Fraser N, Sues H-D, editors. In the shadow of the dinosaurs: early Mesozoic tetrapods. Cambridge: Cambridge University Press; 1994. p.303±21. 23. McGowan GJ. Albanerpetontid amphibians from the Jurassic (Bathonian) of southern England. Mus North Ariz Bull. 1996; 60:227±34. 24. Delfino M, Sala B. Late Pliocene Albanerpetontidae (Lissamphibia) from Italy. J Vertebr Paleontol. 2007; 27: 716±19. 25. Rage J-C. The lower vertebrates from the Eocene and Oligocene of the Phosphorites du Quercy (France): an overview. Strata 2006; 13:161±73. 26. Gardner JD. The amphibian Albanerpeton arthridion and the Aptian-Albian biogeography of albanerpetontids. Palaeontology 1999; 42:529±44 27. Gardner JD. New albanerpetontid amphibians from the Albian to Coniacian of Utah, USAÐbridging the gap. J Vertebr Paleontol 1999; 19:632±38. 28. Gardner JD. Monophyly and intra-generic relationships of Albanerpeton (Lissamphibia; Albanerpetontidae). J Vertebr Paleontol. 2002; 22:112±22. 29. Gardner JD. Albanerpetontid amphibians from the Upper Cretaceous (Campanian and Maastrichtian) of North America. Geodiversitas 2000; 22:349±88. 30. Gardner JD. Revised taxonomy of albanerpetontid amphibians. Acta Palaeontol Pol. 2000; 45:55±70. 31. Gardner JD, Averianov AO. Albanerpetontid amphibians from the Upper Cretaceous of Middle Asia. Acta Palaeontol Pol. 1998; 43:453±67. 32. Skutchas PP. New specimens of albanerpetontid amphibian from the Upper Cretaceous of Uzbekistan. Acta Palaeontol Pol. 2007; 52:819±21. 33. Skutschas PP. Mesozoic salamanders and albanerpetontids of Middle Asia, Kazachstan, and Siberia. Palaeobio Palaeoenv. 2013; 93:441±57. 34. Sweetman SC, Gardner JD. A new albanerpetontid amphibian from the Barremian (Early Cretaceous) Wessex Formation of the Isle of Wight, southern England. Acta Palaeontol Pol. 2013; 58:295±324. 35. Maeda S. [A stratigraphical study on the Tetori Series in the upper Shiokawa district in Gifu Prefecture]. J Geol Soc Japan 1952; 58:145±53. Japanese with English abstract. 36. Maeda S. On the geological history of the Mesozoic Tetori Group in Japan. J College Arts Sci, Chiba Univ. 1961; 3: 369±426. 37. Kusuhashi N, Matsuoka H, Kamiya H, Setoguchi T. Stratigraphy of the late Mesozoic Tetori Group in the Hakusan Region, central Japan: an overview. Mem Fac Sci, Kyoto Univ, Geol Min Ser. 2002; 59:9±31. 38. Fujita AM. Geological age and correlation of the vertebrate-bearing horizons in the Tetori Group. Mem Fukui Prefectural Dinosaur Mus. 2003; 2:3±14. 39. Isajii S, Okazaki H, Hirayama H, Matsuoka H, Barrett P, Tsubamoto T, et al. Depositional environments and taphonomy of the bone−bearing beds of the Lower Cretaceous Kuwajima Formation, Tetori Group, Japan. Bull Kitakyushu Mus Nat Hist Human Hist. 2005; A3:123±33. 40. Isajii S. [Geological setting of the Kuwajima ªKaseki−kabeº, the Kuwajima Formation]. In: Matsuoka H, editor. Fossils of the Kuwajima ªKaseki−kabeº (Fossil−bluff): Scientific Report on a Neocomian (Early Cretaceous) Fossil Assemblage from the Kuwajima Formation, Tetori Group, Shiramine, Ishikawa, Japan. Shiramine: Shiramine Village Board of Education; 2000. p.14±16. Japanese. 41. Okasaki H, Isajii S. Comparative study of sedimentary processes forming bone-bearing beds between the Early Cretaceous Kuwajima Formation, Central Japan, and Wonthaggi Formation, South Australia. Nat Hist Res. 2008; 10:1±8. 42. Barrett PM, Hasegawa Y, Manabe M, Isajii S, Matsuoka H. Sauropod dinosaurs from the Lower Cretaceous of eastern Asia: taxonomic and biogeographical implications. Palaeontology 2002; 45: 1197± 217. 55 / 58 43. Ohashi T, Barrett PM. A new ornithischian dinosaur from the Lower Cretaceous Kuwajima Formation of Japan. J Vertebr Paleontol. 2009; 29:748±57. 44. Unwin DM, Matsuoka H. [Pterosaurs and birds]. In: Matsuoka H, editor. Fossils of the Kuwajima "Kaseki-Kabe" (Fossil-bluff), Scientific report on a Neocomian (Early Cretaceous) fossil assemblage from the Kuwajima Formation, Tetori Group, Shiramine, Ishikawa, Japan. Shiramine: Shiramine Village Board of Education; 2000. p.99±104. Japanese. 45. Rougier GW, Isajii S, Manabe M. An Early Cretaceous mammal from the Kuwajima Formation (Tetori Group), Japan, and a reassessment of triconodont phylogeny. Ann Carnegie Mus. 2007; 76:73±115. 46. Kusuhashi N. Early Cretaceous multituberculate mammals from the Kuwajima Formation (Tetori Group), Central Japan. Acta Palaeontol Pol. 2008; 53:379±90. 47. Setoguchi T, Matsuoka H, Matsuda M. New discovery of an Early Cretaceous tritylodontid (Replilia. Therapsida) from Japan and the phylogenetic reconstruction of Tritylodontidae based on the dental characters. In: Wang Y, Deng T, editors. Proceedings of the Seventh Annual Meeting of the Chinese Society of Vertebrate Paleontology. Beijing: China Ocean Press; 1999. p.117±24. 48. Matsuoka H, Setoguchi T. Significance of Chinese tritylodonts (Synapsida, Cynodontia) for the systematic study of Japanese materials from the Lower Cretaceous Kuwajima Formation, Tetori Group of Shiramine, Ishikawa, Japan. Asian Paleoprimatol. 2000; 1:161±76. 49. Matsuoka H, Kusuhashi N, Corfe IJ. 2016. A new Early Cretaceous tritylodontid (Synapsida, Cynodontia, Mammaliamorpha) from the Kuwajima Formation (Tetori Group) of Central Japan. J Vertebr Paleontol. 2016 Mar 22; 36(4):e1112289. 50. Evans SE, Manabe M. Early Cretaceous lizards from the Okurodani Formation of Japan. Geobios 1999; 32:889±99. 51. Evans SE, Manabe M. A herbivorous lizard from the Early Cretaceous of Japan. Palaeontology 2008; 51:487±98. 52. Evans SE, Manabe M. The Early Cretaceous lizards of Eastern Asia: new material of Sakurasaurus from Japan. Spec Pap Palaeontol. 2009; 81:1±17. 53. Evans SE, Manabe M, Noro M, Isajii S, Yamaguchi M. A long−bodied aquatic varanoid lizard from the Early Cretaceous of Japan. Palaeontology 2006; 49:1143±65. 54. Evans SE, Matsumoto R. An early Cretaceous lizard fauna from Japan. Palaeontol electronica 2015 Jul; 18(2):1±36. 55. Yabumoto Y. Early Cretaceous freshwater fishes from the Tetori Group, central Japan. Bull Kitakyushu Mus Nat Hist Human Hist. 2005; 3:135±43. 56. Yabumoto Y. Sinamia kukurihime, a new Early Cretaceous amiiform fish from Ishikawa, Japan. Paleontol Res. 2014; 18:211±23. 57. Yabumoto Y, Yang SY, Kim TW. Early Cretaceous freshwater fishes from Japan and Korea. J Palaeontol Soc Korea 2006; 22:119±312. 58. Hirayama R, Isajii S, Hibino T. Kappachelys okurai gen. et sp. nov., a new stem soft-shelled turtle from the Early Cretaceous of Japan. In: Brinkman DB, Holroyd PA, Gardner JD, editors. Morphology and evolution of turtles. Vertebrate Paleobiology and Paleoanthropology Series. Dordrecht: Springer; 2013. p.179±85. 59. Isajii S, Matsushita A, Hirayama R. Chelonian eggshells from the Lower Cretaceous Kuwajima Formation of the Tetori Group, Central Japan. Paleontol Res. 2006; 10:29±36. 60. Matsumoto R, Evans SE, Manabe M. The choristoderan reptile Monjurosuchus from the Early Cretaceous of Japan. Acta Palaeontol Pol. 2007; 52:329±50. 61. Matsumoto R, Manabe M, Evans SE. The first record of a long-snouted choristodere (Reptilia, Diapsida) from the Early Cretaceous of Ishikawa Prefecture, Japan. Hist Biol. 2015; 27:583±94. 62. Matsuoka H. [A frog fossil]. In: Matsuoka H, editor, Fossils of the Kuwajima "Kaseki-Kabe" (Fossilbluff), Scientific report on a Neocomian (Early Cretaceous) fossil assemblage from the Kuwajima Formation, Tetori Group, Shiramine, Ishikawa, Japan. Shiramine: Shiramine Village Board of Education; 2000. p.50±2. Japanese. 63. Matsumoto A, Kusuhashi N, Murakami M, Tagami T, Hirata T, Iizuka T, et al. [LA−ICPMS U−Pb zircon dating of tuff beds of the upper Mesozoic Tetori Group] Abstracts with Programs of the 155th Regular Meeting of the Palaeontological Society of Japan. Kyoto; 2006. p.110. Japanese. 64. Kusuhashi N, Matsumoto A, Murakami M, Tagami T, Hirata T, Iizuka T, et al. Zircon U−Pb ages from tuff beds of the upper Mesozoic Tetori Group in the Shokawa district, Gifu Prefecture, central Japan. The Island Arc 2006; 15:378±90. 65. Sano S. New view of the stratigraphy of the Tetori Group in Central Japan. Mem Fukui Prefect Dinosaur Mus. 2015; 14:25±61. 56 / 58 67. Costa O. Paleontologia del regno di Napoli contenente la descrizione e figura di tutti gli avanzi organici fossili racciusi nel suolo di questo regno. Atti Accad pont Naples. 1864; 8:1±198. 68. Lapage EO. The septomaxillary of the Amphibia Anura and of the Reptilia. J Morphol Physiol. 1928; 46:399±430. 69. Duellman WE, Trueb L. Biology of amphibians. New York: McGraw Hill Book Company;1973. 70. De Beer GR. The development of the vertebrate skull. Chicago: The University of Chicago Press; 1937. 71. Wiechmann MF. Albanerpetontidae (Lissamphibia) aus dem Mesozoikum der Iberischen Halbinsel und dem Neogen von SuÈddeutschland [PhD dissertation]. Freie UniversitaÈt Berlin; 2003. 72. Gardner JD. Redescription of the geologically youngest albanerpetontid (?Lissamphibia): Albanerpeton inexpectatum Estes and Hoffstetter, 1976, from the Miocene of France. Ann PaleÂontol. 1999; 85:57±84. 73. Gardner JD. Comments on the anterior region of the skull in the Albanerpetontidae (Temnospondyli: Lissamphibia). Neues Jahrb Geol PalaÈontol Monats. 2000c;1±14. 74. McGowan GJ. Frontals as diagnostic indicators in fossil albanerpetontid amphibians. Bull Natl Sci Mus, Tokyo, C, 1998; 24:185±94. 75. Goloboff PA, Catalano SA. TNT version 1.5, including a full implementation of phylogenetic morphometrics. Cladistics 2016; 32:221±38. 76. Jenkins FA Jr, Walsh DM, Carroll RL. Anatomy of Eocaecilia micropodia, a limbed caecilian of the Early Jurassic. Bull Mus Comp Zool 2007 Aug 20; 158(6);285±365. 77. Sigurdsen T. The otic region of Doleserpeton (Temnospondyli) and its implications for the evolutionary origin of frogs. Zool J Linn Soc. 2008; 154:738±51. 78. Pardo JD, Small BJ, Huttenlocker AK. Stem caecilian from the Triassic of Colorado sheds light on the origins of Lissamphibia. Proc Natl Acad Sci USA, epub Jun 19 2017; https://doi.org/10.1073/pnas. 1706752114 PMID: 28630337 79. Marjanović D, Laurin M. The origin(s) of modern amphibians: a commentary. Evol Biol. 2009; 36:336± 38. 80. Marjanović D, Laurin M. The origin(s) of extant amphibians: a review with emphasis on the ªlepospondyl hypothesisº. Geodiversitas 2013; 35 (1):207±72. 81. Rage J-C, Roček Z. Redescription of Triadobatrachus massinoti (Piveateau, 1936) an anuran amphibian from the Early Triassic. Palaeontographica Abt A 1989; 206:1±16. 82. Ascarrunz E, Rage JC, Legreneur P, Laurin M. Triadobatrachus massinoti, the earliest known lissamphibian (Vertebrata: Tetrapoda) re-examined by μCT scan, and the evolution of trunk length in batrachians. Contrib. Zool. 2016; 85(2):201±34. 83. Borsuk-Biaøynicka M, Cook E, Evans SE, Maryanska T. A microvertebrate assemblage from the Early Triassic of Poland. Acta Palaeontol Pol. 1999; 44:1±22. 84. Borsuk-Biaøynicka M, Evans SE. The scapulocoracoid of an Early Triassic stem-frog from Poland. Acta Palaeontol Pol. 2002; 47:79±96. 85. Evans SE, Borsuk-Biaøynicka M. The Early Triassic stem-frog Czatkobatrachus from Poland. Palaeontol Pol. 2009; 65:79±105. 86. Nessov LA, Fedorov PV, Potanov DO, Golovyeva LS. The structure of the skulls of caudate amphibians collected from the Jurassic of Kirgizstan and the Cretaceous of Uzbekistan. Vest Sankt-Peters Univ Geol. 1996; 7:3±11. 87. Skutschas PP, Martin T. Cranial anatomy of the stem salamander Kokartus honorarius (Amphibia: Caudata) from the Middle Jurassic of Kyrgyzstan. Zool J Linn Soc. 2011; 161:816±38. 88. Ivachnenko M. Urodeles from the Triassic and Jurassic of Soviet Central Asia. Paleontol J. 1978; 12:362±68. 89. Evans SE, Milner AR, Mussett F. The earliest known salamanders (Amphibia: Caudata): a record from the Middle Jurassic of England. Geobios 1988; 21:539±52. 90. Anderson JS, Reisz RR, Scott D, FroÈbisch NB, Sumida SS. A stem batrachian from the Early Permian of Texas and the origin of frogs and salamanders. Nature 2008; 453:515±18. https://doi.org/10.1038/ nature06865 PMID: 18497824 91. Wiechmann MF. The albanerpetontids from the Guimarota mine. In: Martin T, Krebs B. editors. Guimarota: a Jurassic ecosystem. MuÈnchen: Verlag Dr Friedrich Pfeil; 2000. p.51±8. 92. Evans SE, McGowan GJ. Lissamphibian remains from the Purbeck Limestone Group, Southern England. Spec Pap Palaeontol. 2002; 68:103±19. 57 / 58 1. Estes R. Prosirenidae , a new family of fossil salamanders . Nature 1969 ; 224 : 87 ± 8 . 2. Estes R. Gymnophiona , Caudata. Wellnhofer P, editor. Handbuch der Palaoherpetologie , 2. Stuttgart, New York: Gustav Fischer Verlag; 1981 . 66. Haeckel E. Generelle morphologie der organismen. 2 volumes . Berlin: Reimer; 1866 . 93. Mazin J.-M. , Billon-Bruyat J.-P. , Pouech J. , Hantzpergue P. The Purbeckian site of Cherves-deCognac (Berriasian , Early Cretaceous, southwest France) : a continental ecosystem accumulated in an evaporitic littoral depositional environment . In: Barrett PM , Evans SE , editors. Abstracts and Proceedings, Ninth International Symposium on Mesozoic Terrestrial Ecosystems and Biota , Manchester, UK. 2006 . p. 84 - 8 . 94. D'Erasmo G. La fauna e l'eta dei calcari a ittioliti di Pietraroia, in provincial di Benevento . Palaeontogr Ital . 1914 ; 20 : 29 ± 36 . 95. Rees J , Evans SE . Amphibian remains from the Lower Cretaceous of Sweden: the first Scandinavian record of the enigmatic group Albanerpetontidae . GFF . 2002 ; 124 : 87 ± 91 . 96. Szentesi Z , Gardner JD , Venczel M. 2013 . Albanerpetontid amphibians from the Late Cretaceous (Santonian) of Iharkut, Hungary, with remarks on regional differences in late Cretaceous Laurasian amphibian assemblages . Can J Earth Sci . 2013 ; 50 : 268 ± 81 . 97. Cifelli RL . Cretaceous mammals of Asia and North America . Paleontol Soc Korea Sp Pub . 2000 ; 4 : 49 ± 84 . 98. Sweetman SC . A gobiconodontid (Mammalia, Eutriconodonta) from the Early Cretaceous (Barremian) Wessex Formation of the Isle of Wight, Southern Britain . Palaeontology 2006 ; 49 : 889 ± 97 . 99. Sweetman SC . A spalacolestine spalacotheriid (Mammalia, Trechnotheria) from the Early Cretaceous (Barremian) of southern England and its bearing on spalacotheriid evolution . Palaeontology 2008 ; 51 : 1367 ± 85 . 100. Cifelli RL , Kirkland JI , Weil A , Deino AL , Kowallis BJ . High-precision 40Ar/39Ar geochronology and the advent of North America's Late Cretaceous terrestrial fauna . Proc Natl Acad Sci USA 1997 ; 94 : 11163 ± 67 . PMID: 9326579 101. Fiorillo AR . Dinosaurs of Alaska: implications for the Cretaceous origin of Beringia . Geol Soc Am Sp Pap . 2008 ; 442 : 313 ± 26 . 102. Zanno LE , Makovicky PJ . On the earliest record of Cretaceous tyrannosauroids in western North America: implications for an Early Cretaceous Laurasian interchange event . Hist Biol . 2011 ; 23 : 317 ± 25 . 103. Farke AA , Maxwell WD , Cifelli RL , Wedel MJ . A ceratopsian dinosaur from the Lower Cretaceous of western North America and the biogeography of neoceratopsians . PLoS ONE 2014 Dec 10 ; 9 ( 12 ): e112055. https://doi.org/10.1371/journal.pone. 0112055 PMID: 25494182 104. Zakharov YD , Shigeta Y , Popov AM , Velivetskaya TA , Afanasyeva TB . Cretaceous climatic oscillations in the Bering area (Alaska and Koryak Upland): isotopic and palaeontological evidence . Sediment Geol . 2011 ; 235 : 122 ± 31 . 105. Nessov LA . [Cretaceous salamanders and frogs of Kizylkum Desert] Trudy Zool Inst Akad Nauk SSSR 1981 ; 101 : 57 ± 88 . Russian . 106. Nessov LA . Late Mesozoic amphibians and lizards of Soviet Middle Asia . Acta Zool Cracov . 1988 ; 31 : 475 ± 86 . 107. Sanmartin I , Enghoff H , Ronquist F . Patterns of animal dispersal, vicariance and diversification in the Holarctic . Biol J Linn Soc . 2001 ; 73 : 345 ± 90 . 108. Skutschas PP , Krasnolutskii SA . A new genus and species of basal salamanders from the Middle Jurassic of Western Siberia, Russia . Proceedings ZIN 2011 ; 315 : 167 ± 75 . 109. Sweetman SC . A comparison of Barremian±early Aptian vertebrate assemblages from the Jehol Group, north-east China and the Wealden Group, southern Britain: the value of microvertebrate studies in adverse preservational settings . Palaeobio Palaeoenv . 2016 ; 96 ( 1 ): 149 ± 167 . 110. Amiot R , Wang X , Zhou Z , Buffetaut E , LeÂcuyer C , Ding Z , et al. Oxygen isotopes of East Asian dinosaurs reveal exceptionally cold Early Cretaceous climates . Proc Natl Acad Sci USA . 2011 ; 108 : 5179 ± 83 . https://doi.org/10.1073/pnas.1011369108 PMID: 21393569 111 . Daza JD , Stanley EL , Wagner P , Bauer AM , Grimaldi DA . Mid-Cretaceous amber fossils illuminate the past diversity of tropical lizards . Sci Adv . 2016 Mar 4 ; 2 ( 3 ):e1501080. https://doi.org/10.1126/sciadv. 1501080 PMID: 26973870


This is a preview of a remote PDF: http://journals.plos.org/plosone/article/file?id=10.1371/journal.pone.0189767&type=printable

Ryoko Matsumoto, Susan E. Evans. The first record of albanerpetontid amphibians (Amphibia: Albanerpetontidae) from East Asia, PLOS ONE, 2018, DOI: 10.1371/journal.pone.0189767