Tupilakosaurus heilmani Nielsen

J. Li)ua. Soc. (Zool.) 47, 31 1, pp. 2223-2229 With 3 jgures Printed iii Greut Britrrw October. 1967 New observations on the skull-roof of the holotype of Tupilakosaurus heilmani Nielsen BY EIGIL NIELSEN Niiieralogisk Mux~rcitr, CopeTLhagen, Denmark Tn 1954 I erected the genus Tupilnko.wums (Nielsen, 1954) based on the scattered remnants of the skeleton of a batrachomorplr (specimen A), preserved in a fairly large concretion collected by myself in 1932, from Fish-zone V in the Eotriassic beds wwt of River VII near Kap Stosch, in Kortheast Greenland (cf. Nielsen, 1935: 104-105). After a careful manual preparation on one face of the concretion a number of relatively 15 ell-preserved bones were exposed, including several vertebral centra, neural arches, ribs and parts of dermal bones from the head. TI? o other specimens from my Triassir collections of vertebrates made in 1932 and 1933 (specimens B and C) I doubtfully referrccl to the same genus. The vertebral centra in all three sperinicns were extremely short and their end faces concave, the centra therefore looked nruvh as those of the Ichthyosaurla. Of special interest, however, was, the fact that Tupiltckosrcurus evidently had two vertebral centra below each neural arch, an intercentruni and a plcurocentrum, and thus was an ernbolomerous labyrinthodont. These animals were previously known only from Carboniferous and Lou er Permian deposits. The most interesting dermal bones oft hr skull (most of which nere only expo\ed from their inner side) were interpreted by me as belonging to the posterior part of the skull-roof, and as the following articulated elements the parietal, the postparietal, the postorbital, and the supratemporal; on the basis of these clermal bones I attempted a reconstruction of this part of the skull-roof. Of other dermal bones of the skull 1 itlr~ntifictl the left quadratojugal and, TI ith some doubt, a jugal and an angular. Part of the parietal was preserved as an impression shon ing the sculpture of the external face of the bone, and this sculpture was ctrcidedly of labyrinthodont type. I therefore suggested, although with mncli doubt, that Tupilakosaurus or a related form might be ancestral to the ichthyosaurians, the oldest member of which appeared a short tinir later in the Triassic. This suggestion was received in differcnt ways. Most scientists considered the evidence much too slender but others, as for instance Professor I?. v. Huene, accepted or almost accepted the hypothesis (v. Huene, 1954. 1956, 1959). In his paper in 1959-Neues iiber Tuprlrrkosaurus-v. Huene described some material from the Lower Triassic (zone Va) at Wctluga in northern Russia, which he interpreted as fragments of Tupilnkosaurus This material included several vertebral bodies, a parasphenoid, and articulated with that the right and left pterygoid. The parasphenoid and the pterygoids showed especially clost resemblance to the Middle Triassic brachyopid Batrachosuchus. Based on this material as ndl as on tlir material from Northeast Greenland, v. Hunee made tvr o attempted restorations of tlic sltull of Tupilakosaurus, in dorsal ant1 ventral view, one of which is shown in Fig. 1. In 1961, M. A. Shishkin published a short paper on Tupilakosaurus based on the Russian material, part of which had already been tlmx4xd by v. Huene. In addition to the skeletal 15

826 EIGIL XIELSEN Fig. 1. Attempted restoration of the skull of Tupilakosaurus heilrnani in dorsal view. After v. Huene, 1959. The supposed posterior part of the skull-roof of the Greenlandic holotype indicated by heavy lines. P, Frontal; J, jugal; L, lacrimal; M, maxillary; N, nasal; P, parietal; Pi, postparietal; Pm, premaxillary ; Pj, postfrontal; Po, postorbital; Prf, prefrontal; Qj, quadratojugal; Sq, squamosal; St, supratemporal: Tb, tabular. elements mentioned by v. Huene, Shishkin described a fairly complete skull which was very different from v. Huene s attempted restoration of Tupilukosaurus heilmani, being much shorter and of decidedly brachyopid type (Fig. 2). For good reasons, as we shall see later, Shishkin had some difficulties in attempting to compare the skull fragment of the Greenland form with the much more complete skullroof of the Russian form, for which he erected the new species Tupilakosaurus wetlugensis. However, he considered the difficulties to be due to the fact that the skull-roof in Tupiluko-

The skull roof (4 Tupilakosaurus 227 saurus heilmani was mainly exposed from its ventral fa.ce and that of Tupzlnkosaurus wetlugensis from its dorsal face. Both species he placed in a new family, the Tupilakosauridae. After having read Shishkin s paper I dccided to make a reinvestigation of the holotype of Tupilakosaurus heilmni by the now well-knownacetic acid method, in order to see the preserved parts of the skull-roof in dorsal view. With valuable assistance from mag. scient. Niels Bonde a part of the concretion ~liich mainly contained vertebral centra, Pig. 2. Tiipilnkosnurus ~cctlugoisis Shishkirr. The skull in dorsal view. After Shishkin, 1961. The part of the skull-roof corresponding to that, of Tupilnkostrurus heilnimi indicated by heavy lines. B, Basioccipital; E, esoccipital; F, frontal: F.p., foramen paraquadratum :.I, jngal; I,, lacrimal; hf, maxille.rp; S, nasal: P. parietal; Pf, postfrontal; Pm, premasillary; PO, post,orbital; PP, postpariet,al; Prf, prefroiitd ; Qj, quadratojugal; Sq, squamosal : Sf, supratemporal; T, tabular. neural arches and ribs, mas partially ernliedded in clear resin, and the uncovered part treated with 15% acetic acid until all or almost all of the matrix had been dissolved. This first experiment was highly successful and we therefore treated the rest of the holotype in the same way. Thus we exposed many skeletal elements which we knew from the manual preparation and in addition several bones 15 hich had been completely embedded in matrix. We obtained a good view of the upper face of the skull-roof showing in addition to the characteristic sculpture also the groores for the supraorbital sensory canals (Fig. 3). Moreover we exposed three articulated bones not previously known, the right tabular, the right postparictal, and the right exoccipital and in addition a complete scapula, a fragment of the lower jaw with remnants of three teeth, some vertebral centra and a neural arch, all of which had been conipletely embedded in matrix. Tn this short paper I am only dealing with that part of the skull-roof which I had studied previously, mainly from its rentral face, but which is now completely exposed from above as well.

228 ETCIL XIELSEN Fig. 3. Tupilukosaurus heilmani. Reconstruction of part of the skull-roof in dorsal view. P, Frontal; N, nasal; Pf, postfrontal; Prf, prefrontal. A comparison with the skull of Tupilakosaurus wetlugensis showed that my first interpretation of this part of the skull-roof was completely wrong. What I had considered as a posterior part of the skull-roof was in fact anterior, the evidence for this being the course of the grooves for the sensory canals, i.e. the supraorbital sensory canals, shown distinctly in Fig. 3. Also, what I had considered as the posterior margin of the cranial roof was in fact the anterior one. According to the new interpretation of the articulated part of the skull-roof, we are thus dealing with the frontals, the postfrontals, the prefrontals, and the nasals. In that way the resemblance with the other short-faced brachyopids is much more convincing, and the Tupilakosauridae undoubtedly are closely related to the Brachyopidae. There can be no doubt that Tupilakosaurus heilmani and Tupilakosaurus wetlugensis are both embolomerous forms and the question naturally arises, what about the vertebral structure in the other short-faced labyrinthodonts included in the family Brachyopidae. Most of these genera are described on the basis of one or a few specimens of which only the skull or parts of the skull are known. Only in very few cases, viz. the genera Tungussogyrinus (Efremov, 1939) and Brachyops (Owen, 1854), do we know something of the vertebral structures. Tungussogyrinus has solid, elongate vertebrae reminiscent of the vertebrae in the Plagiosauridae, but in Brachyops from the Lower Triassic Mangali beds of Central India we suspect that the vertebral centra found in the contemporaneous beds of the Panchet group in fact are vertebrae of Bruchyops (Huxley, 1865, pl. 6, figs 9 to 10; Romer, 1947: 236). These vertebrae or vertebral centra remind one very much of those in Tupilakosaurus, and therefore quite probably Brachyops also was an embolomerous form. REFERENCES EFREMOV, J. A., 1939. First representative of Siberian early Tetrapoda. Dokl. Akad. NaukSSSR, 23, 1. HUENE, F. v., 1954. Dse Saurierwelt tcnd ihre geschachliclien Zusammenhange. Jena,. HUENE, F. v., 1956. Palaontologie und Phylogenze der Niederen Tetrapoden. Jena. HUENE, F. v., 1959a. Xeues nber T1~pilnl~o,m,trir9. -\'\'pues Jb. Oeol. Pnlaont., Nh., 230-233.

The skull roqf oj Tupilakosaurus 229 HUEXE, F. v., 1939b. Palaontologie uwd Phylogrnic der Kiederen Tetrupoden, Kcxchtrage tind Erganzungen. Jena. Huxi.n-, T. H., 1865. On vertebrate fossils frorn the Parichet roclis near Kauigiiiij, Bengal. Mem. geol. Stcrv. IPLdiu Pa,laeont. indica, (a), 1, 1 : 1-24. NIELSEN, E., 1935. The Permian and Eotriassic \-ert,ebrate-bearing Beds at Godthetb Gulf. Meddr Grenlatrtl, 98, 1. NIELSEK, E., 1054. Tupilukosuurzrs heilmcmi, 1144. el n.sp. An interesting bat,racliomori)i! from the Triassic of East Greenland. Afeddr Gronkund, 72 (2), 8. OWK.?-, R., 1834. Description of the oraniiiiii of a labyrinthodont rept,ile (Urucltyops Ircticcps) from hlangnli, Central India. Q. Jl geol. Soc. Lontl., 10. Rokik;it, A. S., 1947. Review of the Labyrint,liodonlia. Bull. Mus. comp. 2002. Hurv.. 99, 1. SHISHliIS, J1. A., 1961. New data on Tupilnkos~rnrns [Russian]. Dokl. Akacl. Xnitk. SSS'X, 136, 4: 938-94l.