Albanerpeton pannonicum Venczel et Gardner, 2005

Albanerpeton pannonicum Venczel et Gardner, 2005

( Figs 2–15 View Figs 2–5 View Figs 6–11 View Figs 12–15 )

Referred material – 3 right premaxillae (VER 2015.162.1.; 2015.241.1.; 2015.242.1.) , 2 right maxillae (VER 2015.161.1.; 2015.245.) and 4 left maxillae (VER 2015.163.1.–2.; 2015.238.2.–3.) , 1 left maxilla fused with lacrimal and jugal (VER 2015.239.1.) , 12 right dentaries (VER 2015.161.2.; 2015.238.4.–8.; 2015.239.4.–6.; 2015.240.2.; 2015.241.2.; 2015.242.2.) and 2 left dentaries (VER 2015.240.3.; 2015.241.3.) , 3 frontals (VER 2015.239.2.–3.; 2015.240.1.).

TAPHONOMICAL OBSERVATIONS

The specimens are isolated and incomplete ones, and show different qualities of fossilization. Most albanerpetontid fossils have clear, smooth and hard surfaces with red clay filling in their pores. Nevertheless, the surface of some bones is mostly glossy and encrusted with fine-grained sediment, which partially fills the grooves and foramina. Consequently, these structures on these latter mentioned bones are difficult to interpret. Some bone coatings of manganese oxides form more or less large spots. The soluble manganese compounds are mobilized at the same time as the disintegrating of the limestone is produced and they are transported with water that infiltrates the sediment, but these precipitate with the variation of pH, in this case on the surface of the fossilized bones. The environment of karst cavities (as for example Csarnóta 3) is characterized as wet, mildly alkaline and oxidizing, so the stable forms of iron and manganese oxide are easily transported with water. These precipitate on the surfaces of bones affected by pH changes due to variable amounts of rainfall (e.g. LÓPEZ-GONZÁLEZ et al. 2006).

DESCRIPTION

Premaxillae – Three isolated premaxillae are known from Csarnóta 3/6. The best preserved of these is the VER 2015.162.1. ( Figs 2–3 View Figs 2–5 ). The bone is small but it is relatively robust in build compared to similar sized albanerpetontids, for example Albanerpeton arthridion (e.g. GARDNER 1999, text-fig. 2, A-D). The grooves and flanges along the medial surface of this bone suggest that it was paired with another one in life (e.g. VENCZEL & GARDNER 2005). The pars dorsalis is relatively low and broad. The laterodorsal notch is partially broken, but it is wide and was probably deep. The dorsal edge of the pars dorsalis is damaged; the laterodorsal edge bears low, indistinct ridges and grooves due to sutured contact with the nasal.

In labial view, the pars dorsalis is covered by a low dorsal ‘boss’, which is ornamented with irregular shallow pits, grooves and low ridges. On the labial surface of the premaxilla below the ‘boss’ the surface is rugged and external nutritive foramina are scattered below. Labially, the lateroventral corner of the premaxilla bears a seam surface for the lateral process of the maxilla.

In lingual view, the suprapalatal pit opens in moderate size on the medial part of the pars dorsalis, above the pars palatinum, and its outline is elliptical. A labiolingually low internal strut is equally present lateral and medial to the suprapalatal pit. The pars palatinum is a lingually broad shelf, and its lateral end is developed into a flange-like maxillary process. The pars palatinum bears a prong-like vomerine process near its medial end. The oval palatal foramen opens ventrally, midway on the pars palatinum and it joins the suprapalatal pit dorsally. The size of the palatal foramen is less than the diameter of the shaft of the teeth. The pars dentalis is moderately deep and it becomes shallower laterally.

Maxillae ( Figs 4–7 View Figs 2–5 View Figs 6–11 ) – Six maxillae are preserved as isolated bones and one left maxilla is preserved fused with an incomplete lacrimal and jugal. All maxillae are covered lingually with fine sediment crust. These bones are moderately elongated and low. Near its anterior end, the pars dorsalis bears a dorsally protruding nasal process ( Fig. 4 View Figs 2–5 ) with a posteriorly concave margin. The pars dorsalis becomes shallower posteriorly and its dorsal surface from the posterior base of the nasal process is flattened toward the posterior end (the posteriodorsal end is broken on all bones) which were fused, in life, anteriorly with the lacrimal and posteriorly with the jugal (e.g. VENCZEL & GARDNER 2005). The labial surface of the maxilla is plain and on its anterior part, from the base of the premaxillary lateral process to about the middle of the bone, there are tiny, oval nutritive foramina. The anteriorly projecting premaxillary lateral process is relatively short equally in lingual and labial views, and its anterior end is blunt ( Fig. 4 View Figs 2–5 ), or not too acuminate ( Fig. 5 View Figs 2–5 ). Its length is subequal to its depth at the base. The pars palatinum is a lingually broad shelf, which anteriorly has a flange-like premaxillary dorsal process ( Fig. 5 View Figs 2–5 ; it is not figured in lingual view). The medial edge of the pars palatinum is slightly concave that formed, in life, the lateral margin of the narial opening (e.g. VENCZEL & GARDNER 2005). The posterior part of the pars palatinum is broken on all maxillae. The pars dentalis is anteriorly deepest and becomes shallow posteriorly. Labially, the ventral margin of the pars dentalis is nearly straight. The anterior end of the tooth row is approximately in line with the anterior edge of the nasal process.

Dentaries ( Figs 8–11 View Figs 6–11 ) – 16 dentaries were unearthed from the locality but none of them is completely preserved. The dentaries are conspicuously curved labially in dorsal and ventral views. These bones are elongated and slightly taper anteriorly. The vertical symphysis is flattened anteriorly and posteriorly, and usually bears two symphyseal prongs that project medially and slightly posteriorly ( Figs 10–11 View Figs 6–11 ). The end of the symphysis is a swelling-like form, which protrudes labially. The dentary has mesiodistally compressed, highly pleurodont teeth. The tooth row extends for about the anterior three-quarters of the dentary. A dentary has a relatively tall dental parapet; an anteriorly relatively shallow subdental shelf. The dental parapet bears a slightly convex dorsal margin. The labial surface of the dentaries is not ornamented but bears external nutritive foramina that extend along the anterior half of the bone. Under these foramina, on the ventral surface of the bone there is a shallow, anteroposteriorly elongated recess, which is delimited with a low ridge. The Meckelian canal is closed lingually at the end of tooth row, while anteriorly it opens with a small foramen on the ventral part of the symphysis.

Frontals ( Figs 12–15 View Figs 12–15 ) – Only three incompletely preserved fused pairs of frontals are known from the studied deposits. The frontal which is not figured contains only the right part of the bone from the ventrolateral crest to the anterolateral process. All pairs of frontals are solidly fused along the midline, which is visible as a faint line ventrally on the bone. The fused frontals are approximately equilateral triangles in dorsal or ventral views. The spike-like, elongated internasal process projects anteriorly. On both sides, the shorter, prong-like anterolateral process is well developed. In the anterior part of the frontal, on the edge of the bone two slots open to either side of the midline. From these, the anteromedial slot opens between the internasal and the anterolateral processes, whereas the anterolateral slot, which has deeply recessed dorsal and ventral margins, lies more posteriorly and laterally, behind the anterolateral process.

Lacrimal ( Figs 6–7 View Figs 6–11 ) – A left maxilla from Csarnóta 3 is in articulation with a jugal and a lacrimal (VER 2015.239.1.). This incomplete left lacrimal is approximately U-shaped on its side, with an anteriorly directed apex. The dorsal and ventral arms of the lacrimal are separated by a broad, horizontally extending trough (filled with fine grained deposits) on the labial face. In life, the lower surface of the dorsal arm and the upper surface of the ventral arm collectively formed the anterior rim of the orbit (e.g. VENCZEL & GARDNER 2005).

The dorsal arm of the lacrimal projects dorsally and slightly posteriorly, while the ventral arm of this bone extends posteriorly and slightly ventrally. The ventral side of the ventral arm of the lacrimal articulates with the dorsal margin of the pars dorsalis of the maxilla, from the distal end of the nasal process posteriorly to a point about two-thirds along the maxilla. The ventral arm of the lacrimal wraps ventrally on the labial surface of the maxilla and extends lingually onto the maxilla through the dorsal surface of the pars palatinum. The posterior end and the medioposterior edge of the ventral arm of the lacrimal do not connect with the maxilla, forming a gap lingually between these bones.

Jugal ( Figs 6–7 View Figs 6–11 ) – The jugal is not complete, its dorsal flange and posterior end are damaged. Based on its preserved part, the jugal is elongated. The lateral flange is low; its labial surface is smooth. The anteriormost end of the jugal is solidly wedged between the underlying maxillary pars dorsalis and the overlying posterior end of the ventral arm of the lacrimal.

Remarks – The symphyseal prongs on the dentaries are autapomorphic for albanerpetontids (e.g. FOx & NAYLOR 1982, GARDNER 2001). Based on the synapomorphies of the frontals (triangular in outline and moderately elongated) that is diagnostic for the genus Albanerpeton . The narrow and flattened median keel on the fused frontals is autapomorphic of A. pannonicum ( VENCZEL & GARDNER 2005) . Beside this the robustly constructed premaxilla (compared to its size), the well-developed vomerine process on this bone, the relatively lateral process on the maxillae (with length subequal to depth at the base), the ventrally relatively straight dental parapet and the nearly aligned maxillary tooth row and anterior edge of the nasal processes also suggest that these bones are close to this species.

PALAEOECOLOGICAL AND PALAEOBIOGEOGRAPHICAL NOTES

The fossil record of the Albanerpeton genus extends from the middle or late Campanian to the late Pliocene in Europe (summarized by GARDNER & BÖHME 2008, Table 2 View Table 2 ). This suggests that it is a very conservative genus preserved in relatively conservative skeletal forms during this long geological time. The studies of the fossiliferous layers of Csarnóta 2 suggest that the palaeoenvironment changed from forest to grassland during the deposition of its 25 layers (e.g. JÁNOSSY 1986). Despite this process, the albanerpetontids appeared in all layers ( VENCZEL & GARDNER 2005). Conversely, the fauna composition of Csarnóta 2 is typical of humid forests ( KRETZOI 1959, KRETZOI & PÉCSI 1982) with small mammals of Southeast Asian origin, while KRETZOI (1962) wrote in another work that the upper levels of Csarnóta 2, as well as Csarnóta 1, 3 and 4 represent grassland-steppe faunal assemblages.

Compared with Csarnóta 2 the number of vertebrate fossils at Csarnóta 3 is very low but represents many taxa. Besides unidentifiable bone chips of larger mammals only few bones of small mammals are known, as well as bufonid remains and snake vertebrae described previously from the brecciate sediments. Despite the fact that few remains were found in the site, the small mammal fauna is quite substantial ( Table 1 View Table 1 ). Insectivores, squirrel, dormouse, mole-rat, hamster, voles and mice equally occurred.

Ten Soricidae species were determined in the Csarnóta 3 fossil assemblage; all of them belong to the subfamily Soricinae . Five species ( Sorex minutus Linnaeus, 1766 ; Deinsdorfia kordosi Reumer, 1984 ; Petenyia hungarica Kormos, 1934 ; Blarinella europaea Reumer, 1984 ; Zelceina soriculoides Sulimski, 1959 ) belong to the tribe Soricini . Blarinoides mariae Sulimski, 1959 and Mafia csarnotensis Reumer, 1984 can be classified as Blarinini . The other three forms appertain to different tribes ( Episoriculus gibberodon (Petényi, 1864) – Soriculini; Beremendia fissidens (Petényi, 1864) – Beremendiini; Paenelimnoecus pannonicus (Kormos, 1934) – Allosoricini).

In the relative groups of the shrews several Chiroptera, Erinaceomorpha and Talpidae species occurred in the Csarnóta 3 sample, but they are not discussed here. However, the study of the soricid species has a great importance providing relevant data on the stratigraphic position and the palaeoecology of the site.

In addition to the rich insectivore fauna, Arvicolinae and Muridae materials are also significant ( Table 1 View Table 1 ). Six Arvicolinae species were determined in this assemblage, among which Mimomys stehlini Kormos, 1931 indicates the MN 16A ( Mimomys hassiacus - Mimomys stehlini ) Neogene mammal zone (after MEIN 1975). However, the two most common voles are Dolomys nehringi Kretzoi, 1956 and Propliomys hungaricus Kormos, 1934 , similarly to the Csarnóta 2 site. The other voles ( Baranomys loczyi Kormos, 1933 ; Cseria gracilis Kretzoi, 1956 ; Dolomys milleri Nehring, 1898 ) are represented by only few specimens in the material. The same three mice species ( Micromys praeminutus Kretzoi, 1959 ; Apodemus dominans Kretzoi, 1959 ; Rhagapodemus frequens Kretzoi, 1959 ) found in this site were described by KRETZOI (1959) from the Csarnóta 2 locality.

The specific composition of the Csarnóta 3 Soricidae fauna is also very similar to that of the Csarnóta 2 locality. We have found all the well-determinable species which were mentioned by REUMER (1984) from that site ( Table 2 View Table 2 ). Though the stratigraphical occurrences of some forms (for example Mafia csarnotensis , in RZEBIK-KOWALSKA & POPOV 2005, etc.) have been emended since Reumer’s studies, we can be sure of the similar ages of the two sites on the basis of the high number of concomitant shrews.

On the other hand, the ratio of the Csarnóta 3 shrew species is markedly different from that of Csarnóta 2 ( Table 3 View Table 3 ; Fig. 16 View Fig ). The main difference is the higher number of the big sized forms ( Blarinoides mariae and Beremendia fissidens ) and the lower rate of Episoriculus gibberodon and Paenelimnoecus pannonicus at site 3 compared to site 2. The variance can be caused by the fact that the layers of Csarnóta 3 are contemporaneous with the upper part of the strata of site 2, where the aforementioned rates occur, as well (see REUMER 1984, fig. 20; upwards from layer 12). This presumption is supported by the presence of Mafia csarnotensis and the taphonomical features.

Some of the shrew genera reported here are useful in the palaeo-ecotype reconstruction ( REUMER 1984, RZEBIK-KOWALSKA 1995). Sorex and Blarinella indicate a humid forest environment. The presence of open water is marked by Paenelimnoecus and probably Beremendia ( BOTKA & MÉSZÁROS 2014) . Blarinoides and Petenyia are typified by REUMER (1984) as opportunist and ubiquitous genera.

While the typical steppe indicator Crocidura is not present in this sample, the real forests species ( Sorex and Blarinella ) are not frequent either ( Fig. 16 View Fig ). The high rate of water-preferring forms indicates an open surface of a lake or a river, perhaps with forest or shrubby vegetation. At the same time we cannot exclude the presence of more open ecotypes in the wider surroundings because of the great number of opportunist forms.

The same can be observed in the case of rodents. The ratio of the Csarnóta 3 rodent families and subfamilies ( Sciuridae , Gliridae , Spalacidae , Cricetinae, Arvicolinae, Rhagapodemus and Muridae ) is very similar to the upper part of the Csarnóta 2 site ( Fig. 17 View Fig ). Although, Muridae (Apodemus-Micromys) are still dominant, the ratio of voles and Rhagapodemus frequens is also significant. The ratio of the other groups ( Sciuridae , Gliridae , Spalacidae , and Cricetinae) is insignificant, only the proportion of Spalacidae (layers 6–9) and Gliridae (layers 4–5) exceeds 10 %. This fauna composition indicates a rather closed ecotype, but it was more open than that typical in the bottom layers of Csarnóta 2. The accumulation of this fauna is probably a snapshot from a transitional period between the forest and the grassland vegetation.

This study confirms as a new result the presence of albanerpetontids in sediments of Csarnóta 3. Compared with Csarnóta 2, from where several hundred albanerpetontid bones are known, this palaeovertebrate site with its 27 isolated skulls and jaw bones is really poor. Nevertheless, despite the small number of albanerpetontid remains, the fossils represent the diagnostic features of Albanerpeton pannonicum .

Among the studied 10 layers of Csarnóta 3 albanerpetontid fossils have not been unearthed from the strata 1–2 (top of the series) and 4–5. Even though, the levels 1–2 do not have albanerpetontid fossils, it cannot be claimed that this group became extinct by that time because its remains are also absent from the strata 4–5. This fact can be more readily explained by the small number of fossils, which is probably due to the different depositional environment of Csarnóta 2.

KRETZOI (1962) previously mentioned remains of bufonid toads and snakes from Csarnóta 3. Preliminary analysis of the herpetofauna of this site revealed that the bones of the lizards Lacerta and Ophisaurus are present in all layers but the former genus is significantly more abundant. The lacertids are present in the largest numbers in layers 3–7, while the presence of Ophisaurus is uniformly low. The abundance of Lacerta bones in the above mentioned layers may be the result of extraordinary events, for example, an abrupt climate change (e.g. too rapid melting of snow or early heavy rains) during their hibernation ( BEHRENSMEYER 1984, ANDREWS 1990). At times of such accumulation a given population is overrepresented ( BEHRENSMEYER 1991). Snakes and colubrids are the most abundant in the five uppermost layers. In the tenth layer the vertebrae of the Elaphe genus dominate but from the fifth level the bones of the Natrix genus prevail. Bones of turtles as well as some shell fragments were unearthed from strata 1–6. Salamanders are rare but appear in all layers except 2 and 8. The bones of the toad Bufo and the spadefoot Pelobates are present in relatively large numbers in all layers, whilst ranid frogs are rare (layers 1, 5 and 10). It was mentioned above that the bones of Albanerpeton pannonicum are not known from all layers (the reason for this is not known exactly), but it is the most frequent in about the middle of the section (layer 6).