Eodiscoglossus sp.

Folie, Annelise & Codrea, Vlad, 2005, New lissamphibians and squamates from the Maastrichtian of Haţeg Basin, Romania, Acta Palaeontologica Polonica 50 (1), pp. 57-71 : 61-62

Cf. Eodiscoglossus sp.

Fig. 3A .

Material.—PSMUBB V 356 (one incomplete left ilium).

Description.—Only a single posterior part of an ilium recovered at Pui Islaz can be attributed to this taxon ( Fig. 3A ). It is 2 mm long, 1.5 mm high and displays a small anterior part of the acetabulum, the base of the pars ascendens, a small but well−marked supraacetabular fossa and a shallow tubercle instead of a tuber superius. The angle between the tubercle and the pars ascendens is open, the junction between the acetabular region and the shaft is not sharply waisted, the pars descendens is lost and the pars ascendens is postero−dorsally oriented. Discussion.—A postero−dorsally elongated pars ascendens on the ilium is characteristic for the family Discoglossidae ( Rage and Hossini 2000). Moreover, an iliac crest is only present in the members of the Discoglossus group (sensu Rage and Hossini 2000) that includes the genera Discoglossus , Latonia , Eodiscoglossus , Paralatonia , and probably also Paradiscoglossus ( Duffaud and Rage 1999) .

This ilium recovered from Pui Islaz differs from Latonia in that the ilium of the latter is about 10 times larger and has a tuber superius that is mediolaterally thicker with the angle between the tuber superius and the pars ascendens that is slightly more open ( Rage and Hossini 2000). Moreover, Latonia possesses an additional fossula tuberis superioris developed behind the acetabulum, the base of the pars ascendens is narrower and the junction between the acetabulum and the shaft is waisted ( Roček 1994). The juvenile form of Latonia and the adult of Discoglossus are very similar although the latter is smaller. They share a thick tuber superius, the presence of an additional fossula tuberis superioris, a narrow base of the pars ascendens and a waisted junction between the shaft and the acetabulum ( Roček 1994; Duffaud and Rage 1999; Rage and Hossini 2000). Grigorescu et al. (1999) referred fragmentary ilia from the Sânpetru Formation to the genus Eodiscoglossus . Recently, Venczel and Csiki (2003) revised these fragments and attributed them to the new genus Paralatonia . Some characters are concordant with the ilia from Pui Islaz as the size and the great angle between the tuber superius and the pars ascendens. But some characters are different: the supraacetabular fossa is wider and deeper, the tuber superius is more crest−like and continuous with the iliac shaft, and the acetabular rim is not as prominent and does not project beyond the ventral margin of the bone. The presence of an iliac tubercule instead of a tuber superius prevent the attribution of this ilium to the genus Paradiscoglossus .

The iliac crest of PSMUBB V 356 is not well visible. Therefore, it cannot be proved that this ilium really belongs to a discoglossid frog. However, this specimen closely resembles the ilium of Eodiscoglossus Villalta, 1957 by the presence of a dorsal tubercle instead of a crest ( Evans et al. 1990). Two species are currently referred to this genus: E. santonjae Vergaud−Grazzini and Wenz, 1975 (type species) and E. oxoniensis Evans et al., 1990 . Eodiscoglossus santonjae ( Estes and Sanchiz 1982a) differs from PSMUBB V 356 by a more waisted junction between the acetabular region and the shaft (instead of less marked on the fragments from Pui Islaz), and by a lateral ridge between the iliac crest and the pars cylindroformis giving a triangular section to the bone (instead of a groove giving a drop−shaped bone section). Eodiscoglossus oxoniensis seems to be more similar to the fragments from Pui Islaz because the junction between the shaft and the acetabular area is not waisted, and the iliac crest and the pars cylindiformis are separated by a groove ( Evans et al. 1990). However, differences can be observed: in E. oxoniensis , dorsal and ventral pits separate the tubercle from the remainder of the bone and the supraacetabular fossa is longer and deeper ( Evans et al. 1990). Duffaud and Rage (1999) described very similar discoglossid ilia from the Late Cretaceous of Laño ( Spain). According to these authors, those specimens cannot be referred to Eodiscoglossus because the pars ascendens is less developed and the supraacetabular fossa of the ilium is shallower. In the specimen from Pui Islaz, the pars ascendens is not very well preserved and it is therefore impossible to assess whether it was rather strong or frail. It is also difficult to compare the depth of the supraacetabular fossa.

This ilium fragment found at Pui Islaz is very similar to Eodiscoglossus oxoniensis but the differences described above prevent the attribution of the fossil to this species.

Cf. Paradiscoglossus sp.

Fig. 3B, C .

Material.—PSMUBB V 357 (one incomplete right ilium) and V 358–V 360 (three incomplete left ilia).

Description.—These four posterior parts of ilia collected at Pui Islaz are between 2 and 3.5 mm long and 1 to 3 mm high. Two of these fragments (V 357 and V 360) are approximately twice as large as the two other fragments (V 358 and V 359) but they share several characters in common. In lateral view, the iliac shaft is not strongly curved. The junction between the acetabular region and the shaft is not sharply waisted. The pars descendens is lost. The pars ascendens is postero−dorsally oriented. The angle between the tuber superius and the pars ascendens is open. The pars ascendens is lost except on PSMUBB V 357 ( Fig. 3B ) where a small but wide part is present. The supraacetabular fossa seems to be wide and deep. The acetabulum is incomplete as it displays only a well−marked anterior edge. The tuber superius is elongate, mesiodistally compressed and does not pass the well−developed iliac crest ( Fig. 3C ). This crest is separated from the pars cylindriformis by a groove.

Discussion.—As mentioned for cf. Eodiscoglossus sp. , the presence of a postero−dorsally elongated pars ascendens on the ilium is characteristic of the Discoglossidae and the iliac crest is a characteristic of the members of the Discoglossus group ( Rage and Hossini 2000). For the reasons mentioned for the previous taxa, these ilia cannot be attributed to Latonia , Discoglossus , or Paralatonia and the presence of a crest−like tuber superius instead of a tubercle prevent the attribution of these specimens to Eodiscoglossus . However, they share several characters in common with Paradiscoglossus americanus Estes and Sanchiz, 1982 : they are similarly curved, they have well−developed dorsal crest continuous with the elongate and compressed dorsal tuber superius, which is not clearly separated from the crest ( Estes and Sanchiz 1982b), an open angle between the tuber superius and the pars ascendens, and a wide supraacetabular fossa and pars ascendens. However, the posterior part of the bones is lost on the larger fragments and the tuber superius seems to be more compressed than in P. americanus on the smaller one, thus precluding confident referral of the specimens from Pui Islaz to P. americanus .

Discoglossidae indet.

Fig. 3D, E .

Material.—PSMUBB V 353, V 354 (two incomplete urostyles), V 355 (one sacral vertebra).

Description.—Two proximal portions of urostyles were discovered in the excavated lens. Only the two oval anterior cotyles for articulation with the sacral vertebra and the anterior part of the neural canal are preserved in the first specimen. The second, better−preserved specimen ( Fig. 3D ) is 4 mm long, 2 mm high, and 2 mm wide (at the level of the cotyles). The two anterior cotyles, two broken transverse processes and a small part of the dorsal crest are preserved on this specimen. The transverse processes are triangular in cross−section and inclined posterolaterally. Behind them, two spinal foramina are posteriorly open. The dorsal crest is not complete along the fragment. Ventrally, the bone is smooth.

The sacral vertebra founded at Pui Islaz is 1.5 mm long, 2 mm wide and bears an anterior and two posterior condyles ( Fig. 3E ). The neural arch is lost. Ventrally, the posterior condyles are separated by a shallow sagittal canal.

Discussion.—A single pair of transverse processes and a poorly developed neural crest on the urostyle, and a posteriorly bicondylar sacral vertebra indicate that the specimens are referable to the family Discoglossidae . However, these elements are not specific to the generic or specific level. Moreover, the presence of two taxa referred to this family at Pui Islaz prevent the attribution of these elements to cf. Eodiscoglossus sp. or cf. Paradiscoglossus sp.

Figures

Fig. 3. A–E. Discoglossidae. A. PSMUBB V 356, cf. Eodiscoglossus sp., left ilium in lateral view. B, C. cf. Paradiscoglossus sp. B. PSMUBB V 357, great form, right ilium in lateral view. C. PSMUBB V 358, small form, left ilium in lateral view. D, E. Discoglossidae indet. D. PSMUBB V 353, urostyle in dorsal view. E. PSMUBB V 355 sacral vertebra in ventral view. F, G. Becklesius. F. PSMUBB V 365, B. cf. hoffstetteri, left dentary in lingual view. G. PSMUBB V 361, holotype of B. nopcsai sp. nov., fragment of left dentary (G1) and tooth crown (G2) in lingual view. SEM migrographs.