TRUE SKULL ROOF CONFIGURATION OF ICHTHYOSAURUS AND STENOPTERYGIUS AND ITS IMPLICATIONS

Journal of Vertebrate Paleontology 25(2):338 342, June 2005 2005 by the Society of Vertebrate Paleontology TRUE SKULL ROOF CONFIGURATION OF ICHTHYOSAURUS AND STENOPTERYGIUS AND ITS IMPLICATIONS RYOSUKE MOTANI* Department of Geological Sciences, 1272 University of Oregon, Eugene, Oregon 97403-1272 U.S.A. ABSTRACT Ichthyosaurus and Stenopterygius are the two best-known ichthyosaurs, represented by numerous specimens and having a long history of paleontological study. Despite such advantages, it is shown here that a major error in skull roof reconstruction, which dates back to the time of Richard Owen, still dominates the literature. Contrary to the conventional wisdom, the two genera share a unique feature, namely the separation of the frontal and postfrontal by the prefrontal. This hitherto ignored feature is phylogenetically important for clarifying the validity, synapomorphy, and membership of the Thunnosauria, a clade defined by the two genera. The feature is also useful in basic taxonomy. Many isolated skull specimens from the lower Lias of England are identified as Ichthyosaurus. Some of the skulls lack the prefrontal feature and differ from Ichthyosaurus in several other respects, suggesting that they represent at least one undescribed genus in the Lower Lias of England. This coincides with the previously reported presence of forefin specimens that clearly do not belong to known genera. Lower Liassic ichthyosaurs were more diverse than previously thought. INTRODUCTION The study of ichthyosaurs has a long history; the first paleontological publication on the subject appeared in 1814 (Home, 1814). Of 76 genera and 235 species of ichthyosaurs formally named in the literature, 36 genera, containing 80 species, were recently considered valid by McGowan and Motani (2003). The fossil record of ichthyosaurs extends from the uppermost Lower Triassic (Olenekian) to the lowermost Upper Cretaceous (Cenomanian), representing about 145 million years. Despite such diversity, our knowledge of the group is highly biased toward taxa from the Lower Jurassic, from which numerous well-preserved specimens have been collected. Two genera of Lower Jurassic ichthyosaurs, Ichthyosaurus and Stenopterygius, are particularly well known. Ichthyosaurus (Rhaetian Sinemurian) is slightly older than Stenopterygius (Toarcian) in its stratigraphic distribution. Ichthyosaurus is known almost exclusively from England, except for isolated occurrences in Western Canada (McGowan, 1978; Dennison et al., 1990), whereas Stenopterygius is known more widely from Europe, with a vast concentration in southern Germany (e.g., McGowan, 1979, 1991). Although the exact number is unknown, specimens of Ichthyosaurus number probably on the order of hundreds, whereas the comparable figure is on the order of lower thousands for Stenopterygius (McGowan, 1991). Specimens of Stenopterygius tend to suffer from more diagenetic compaction than those of Ichthyosaurus, but there are some three-dimensionally preserved skulls that supplement our knowledge of the former genus (e.g., Owen, 1881; Godefroit, 1994). Still, specimens showing unequivocal suture patterns are rare. The temporal region of ichthyosaurs has been controversial since the 19th Century. Most controversial in recent years has been the identity of a large triradiate bone that forms the postero-lateral corner of the upper temporal fenestra. Two landmark studies, based on excellently preserved specimens of Platypterygius and Ichthyosaurus, declared in the latter half of the 20th Century that ichthyosaurs lacked the supratemporal bone, identifying the large triradiate bone as the squamosal *Present address: Department of Geology, University of California, One Shields Ave., Davis, CA 95616; motani@geology.ucdavis.edu (Romer, 1968; McGowan, 1973). More recently, other authors pointed out the presence of a supernumerary bone next to the large triradiate element, and concluded that the former was the squamosal, the latter the supratemporal (e.g., Kirton, 1983; Godefroit, 1994; Maisch, 1997; Motani, 1999a). The consensus now seems to be that the large triradiate bone is the supratemporal, although it is unusually large, and that there is taxonomic variation in the presence or absence of the squamosal. Recent reexamination of three-dimensionally preserved skulls of Ichthyosaurus and Stenopterygius revealed that their suture patterns have been reconstructed with a major error. This error first appeared in the 19th Century and has been followed by virtually all authors, obscuring the basic taxonomy of Lower Jurassic ichthyosaurs. It also hides an important synapomorphy for the Thunnosauria Motani, 1999b, for which the two genera are the basal members. This group contains the most derived ichthyosaurs that evolved a tuna-shaped body plan. The purpose of this paper is to correct this common misunderstanding of the skull configuration that has been hampering the basic taxonomy and systematics of ichthyosaurs, and to discuss its implication for the taxonomic diversity of lower Liassic ichthyosaurs. MATERIALS AND METHODS Despite the abundance of ichthyosaurian fossils, cranial suture lines are not clear in most specimens because of preservation, and this often leads to different interpretations of a single skull. Sutures may be obscured with the maturity of the individual, but this is not well established. The resulting absence of consensus hampers taxonomic and phylogenetic studies of ichthyosaurs. To minimize subjective interpretations in the present study, the suture patterns were interpreted in the following manner: first, only those skull specimens with exceptionally clear suture lines were used for primary assessment; and second, at least two skull specimens per taxon were used for cross-examination. Institutional Abbreviations BGS, British Geological Survey, London, United Kingdom; NHM, Natural History Museum, London, UK; ROM, Royal Ontario Museum, Toronto, Canada; SMNS, Staatliches Museum für Naturkunde, Stuttgart, Germany; WM, Whitby Museum, Whitby, UK. Specimens with exceptionally clear suture patterns are NHM R8177, R6697, 49203, R10021, R15943, 33157, and 32681. The 338

MOTANI ICHTHYOSAURIAN SKULL ROOF 339 first four are identified as Ichthyosaurus sp., and I support this identification as discussed later. They are clearly too large to be I. breviceps or I. conybeari, leaving I. communis as the only possibility according to the current taxonomy (McGowan and Motani, 2003), unless the specimens belong to an unknown new species. The first three are the material of McGowan (1973). The latter two specimens belong to Stenopterygius longifrons, including the holotype (NHM 33157). Other specimens studied are: SMNS 17500 for S. cuneiceps; NHM R792, SMNS 14846, and WM SIM 876.2 for S. longifrons; NHM 43006 (holotype) and 39263 for I. breviceps; NHM R1162 (neotype), 2013, 39492, and OUM J13799 for I. communis; and NHM 38523 and BGS 956 for I. conybeari. DESCRIPTION Ichthyosaurus NHM R8177, which preserves a part of the skull posterior to the external naris, is the most complete specimen examined by McGowan (1973) in his monograph on the cranial osteology of the genus. It had been acid prepared to the degree that all elements are separated from each other (Mc- Gowan, 1973), allowing researchers to examine the threedimensional overlapping patterns of suture planes, rather than just the suture lines that appear on the skull surface (note that bones overlap extensively in ichthyosaurian skulls, so that the suture plane between a pair of bones often has a large area; see Sollas, 1916, and McGowan, 1973). Because of this exceptional condition, there is no room for subjective interpretation of the cranial suture pattern in this specimen. A reconstruction of the skull roof using these elements (Fig. 1A, B) produced the same configuration as in McGowan (1973:pls. 7, 8). The paired frontals are exposed widely on both sides of the sagittal line, posteriorly enclosing the pineal foramen between them (Fig. 1A, B). The frontal is overlain by three bones: posteriorly and postero-laterally by the parietal, anteriorly and antero-laterally by the nasal, and laterally by the prefrontal, not by the postfrontal, as in most other ichthyosaurs in Parvipelvia. The postfrontal overlies the parietal posteromedially, the prefrontal medially and anteriorly, and may or may not contact the nasal antero-medially (Fig. 1A, B). Particularly noteworthy is the exposure of the prefrontal lateral to the frontal (hereafter referred to as the medial exposure of the prefrontal). This part of the prefrontal appears as the postero-medial branch of the main prefrontal exposure, which is limited to the orbital rim. When a contact exists between the postfrontal and nasal, a part of the prefrontal branch is overlain by the postfrontal, isolating the distal part of the branch from the main part of the bone. In this case, the prefrontal has two separate exposures: the main one along the orbital rim, and an island lateral to the frontal. The details of the positioning of the prefrontal and postfrontal may be difficult to discern in the published photograph of McGowan (1973:pl. 8b), but are clearly seen in an enlarged version of the same photograph kindly provided by C. McGowan. Yet another exposure of the prefrontal is FIGURE 1. Well-preserved skull roofs of Ichthyosaurus. A, B, right half of the skull roof of NHM R8177 (Ichthyosaurus sp. most likely I. communis), reassembled using separate elements; C, D, skull roof of NHM R15943 (Ichthyosaurus sp. most likely I. communis), still retaining natural articulation; E, F, skull of NHM 43006 (I. breviceps). Gray areas represent the prefrontal. Abbreviations: f, frontal; l, lacrimal; m, maxilla; n, nasal; p, parietal; pin, pineal foramen; po, postorbital; prf, prefrontal; ptf, postfrontal; q, quadrate; qj, quadratojugal; sq, squamosal; st, supratemporal. Scale bar equals 10 cm for A, 22 cm for C. Oblique hatching indicates broken surfaces.

340 JOURNAL OF VERTEBRATE PALEONTOLOGY, VOL. 25, NO. 2, 2005 hidden inside the upper temporal fenestra along its anterior margin. This exposure seems to exist not only in Ichthyosaurus but also in many other genera, including Temnodontosaurus. The same configuration, including the medial exposure of the prefrontal, is present in two more skulls of Ichthyosaurus that have been prepared with acid to reveal clear suture lines in articulation, namely NHM R10021 and R15943 (Fig. 1C, D). It should be noted that white labels on the bones of NHM R10021 could be misleading: the right medial exposure of the prefrontal is labeled as a frontal, whereas the area marked as the right prefrontal must be the anterior part of the right postfrontal. The above description is based on well-preserved yet isolated partial skulls of Ichthyosaurus, probably referable to I. communis (see Materials and Methods). Therefore, it is desirable to confirm the suture patterns in skulls of I. communis that are associated with postcranial skeletons. Four skeletons of I. communis, including the neotype, were therefore examined. As pointed out in Materials and Methods, poor preservation may prevent the suture pattern from emerging clearly in many specimens, leaving the median exposure of prefrontal poorly defined. For example, the presence of the medial prefrontal exposure is debatable in the neotype (NHM R1162) and what McGowan (1973:9) described as the finest specimen (NHM 2013), although these specimens do have bony areas medial to the postfrontal that probably represent the prefrontal. Suture patterns are clearer in other specimens, including NHM 39492 and OUM J13799, where the medial prefrontal exposure is unambiguously present. In summary, no conclusive evidence denies the presence of a prefrontal exposure between the frontal and postfrontal in I. communis, whereas such exposures are always present in those specimens whose cranial suture patterns are well preserved. The medial prefrontal exposure is also present in Ichthyosaurus breviceps, a smaller and short-snouted species of the same genus. The feature is unequivocally present in the holotype of this species (NHM 43006; Fig. 1E, F), whereas the presence is debatable in another specimen (NHM 39263). Even in the latter case, however, the situation is similar to that discussed for the neotype of I. communis: a bony area medial to the postfrontal that probably corresponds to the medial prefrontal exposure. The condition in I. conybeari, the last of the three species of the genus, is indeterminate, because the skull roof is invisible in the holotype (NHM 38523), whereas the skull roof sutures are unclear in the only referred specimen (BGS 956). Therefore, it is not possibly to firmly establish that the presence of the medial exposure of the prefrontal is common within the genus Ichthyosaurus, whereas no evidence contradicts the possibility. Stenopterygius The type specimen of Stenopterygius longifrons (NHM 33157) preserves very clear cranial sutures, as does a second specimen (NHM 32681) from the same locality (Toarcian of Curcy, France). Both specimens lack the snout anterior to the external naris (Fig. 2), and NHM 32681 is smaller and less complete than the holotype. They have not been acid prepared, unlike in the skulls of Ichthyosaurus described above, yet their preservation is so exceptional that there is no room for subjective interpretation of the suture patterns; see, e.g., Motani (1999a:fig. 6). The patterns seen in the two are almost identical to each other, and are similar to that of Ichthyosaurus described above in that a pair of prefrontals is exposed lateral to the frontals (Fig. 2). The contact between the postfrontal and frontal is also absent (Fig. 2), except on the left side of NHM 32681 where there is a point contact (not figured). Differences between Ichthyosaurus and Stenopterygius include the width of the frontal exposure, which is much wider in the former. The existence of the medial prefrontal exposure is not restricted to the specimens from the type locality. Two partial skulls from the Toarcian of Whitby (NHM R792 and WM SIM 876.2), probably belonging to Stenopterygius longifrons, exhibit the same suture pattern as the type specimen. SMNS 14846, a complete skeleton from the Toarcian of Holzmaden, Germany, also shows the feature. This specimen is labeled as S. longifrons, and while its specific assignment may be debatable, it undoubtedly belongs to the genus Stenopterygius because diagnostic features of the genus are present in its postcranial skeleton. The holotype of S. cuneiceps (SMNS 17500) also shows the feature. Despite the abundance of specimens, not many complete skeletons of Stenopterygius preserve unambiguous suture patterns of the skull roof, possibly because of the more severe diagenetic compaction than in Ichthyosaurus specimens. Therefore, it is difficult at this point to establish the presence of the medial exposure of the prefrontal in every species of the genus. However, it is still true that the feature is present in those specimens of the genus with unequivocal cranial suture lines preserved. At least, the feature does exist in some species of the genus. DISCUSSION The misinterpretation of the skull roof configuration in Stenopterygius and Ichthyosaurus dates back to two landmark studies by historical authorities, namely Owen (1881) and Sollas (1916). The medial prefrontal exposure was recognized for Stenopterygius longifrons by several authors after Owen (1881), including Fraas (1891:pl. 2, fig. 1), Kirton (1983:fig. 44 this figure is labeled Ichthyosaurus but is based on S. longifrons), and Godefroit (1994:figs 16, 18). However, these recognitions never overturned the original description of the holotype skull by Owen (1881), which did not depict the medial exposure of the prefrontal. For example, Godefroit (1994) considered that the medial exposures of the bone that he observed in his specimens were the artifacts of preservation, and the medial extent of the prefrontal was covered by the nasal and postfrontal in natural articulation. He therefore eliminated the medial prefrontal exposure from his reconstruction of the skull roof (Godefroit, 1994:fig. 19). Motani (1999b) followed this reconstruction, reproducing the error. The situation is similar in Ichthyosaurus. Sollas (1916) studied a skull of Ichthyosaurus in serial sections, and because of this FIGURE 2. Dorsal view of NHM 33157 (holotype of Stenopterygius longifrons). Scale bar equals 21 cm. See Figure 1 for abbreviations. Oblique hatching indicates unprepared space.

MOTANI ICHTHYOSAURIAN SKULL ROOF 341 procedure, his suture reconstruction has been considered unquestionable (e.g., McGowan, 1973). The identification of the specimen as the genus Ichthyosaurus seems warranted, given the very wide exposure of the frontal (Sollas, 1916:figs. 1b, 5) that is unique to this genus in the lower Lias. However, even serial sections require interpretations. A photograph given by Sollas (1916:fig 1b) of the specimen before sectioning shows that neither of the postfrontals extends medially to meet the frontals, as in the line drawing of sutures in his figure 5. The area left between the frontal and postfrontal is where the medial exposure of the prefrontal is usually found. Admittedly, the poor quality of the photograph leaves space for debate, but the medial extent of the prefrontal is usually one of the most conspicuous suture lines found on the skull roof of parvipelvian ichthyosaurs. The sections made from the specimen are no longer recognized in the Oxford collection, and the line drawings made from some of them (Sollas, 1916:fig. 2) do not have the authenticity of the lost original. I conclude that the reconstruction of Sollas (1916) does not present any strong argument against the unquestionable evidence presented by NHM 8177 and other specimens in Figure 1. It is noteworthy that McGowan (1973:fig. 37) figured the medial prefrontal exposure in the side-view reconstruction of the skull, whereas it is absent from his skull-roof reconstruction (Mc- Gowan, 1973:fig. 35). This latter line drawing has the postfrontal and frontal contacting each other without the prefrontal separating them, as in the traditional interpretations of the skull roof configuration of the genus by Sollas (1916), replicated by Romer (1956). Motani (1999b) also adopted Romer s (1956) reconstruction, further spreading the error in the literature. The observed tendency in the literature to interpret the medial prefrontal exposure as an artifact of preservation is understandable, given that the existence of the feature contradicts papers written by historical authorities, such as Owen (1881) and Sollas (1916). However, the appearance of this feature is consistent among the specimens of Ichthyosaurus and Stenopterygius, for which the suture patterns can be determined with confidence. Moreover, no specimen that can be positively identified as Ichthyosaurus or Stenopterygius unambiguously shows the absence of the feature, as far as the author has examined. Therefore, it is reasonable to conclude that this feature is real, and not a preservational artifact. The medial exposure of the prefrontal reported here for Ichthyosaurus and Stenopterygius is unique to the two genera so far as known. A specimen of an ichthyosaur from the Pliensbachian, clearly representing a new genus (McGowan and Motani, 2002), is being studied by the author (ROM 52596). This specimen also shows the feature unambiguously. If the feature is synapomorphic among the two genera and the new specimen, membership of the clade Thunnosauria Motani, 1999b, could change dramatically. However, it is premature to publish a cladogram at this point, because the critical new specimen is still under study. The feature pointed out in the present paper is taxonomically very useful, especially in identifying ichthyosaurs from the lower Lias of England. Numerous isolated ichthyosaurian skulls in the NHM collection are usually identified as Ichthyosaurus (e.g., Fig. 3). These tentative identifications seem to be based on the fact that the general design of the skulls differs from those of the other two ichthyosaurian genera from the same horizon, namely Temnodontosaurus with larger and more robust skulls, and Leptonectes with slender and delicate skulls. Examination of these skulls revealed that the medial prefrontal exposure is clearly absent in some of them (e.g., NHM 2090, R1167; see Fig. 3). Those lacking the feature are typically labeled as I. intermedius, and possess narrower snouts than I. communis (although more robust compared to Leptonectes) with relatively slender teeth (again, more robust than in Leptonectes). These skulls also tend to preserve the squamosal, a bone that is considered absent in I. communis (McGowan, 1973). Ichthyosaurus intermedius is usu- FIGURE 3. Two of the supposed Ichthyosaurus skulls in the NHM collection, probably representing at least a new genus. A, right lateral view of NHM R2090; B, C, dorsal view of skull roof of same; D, left lateral view of NHM R1164; E, F, dorsal view of skull roof of same. The lateral compaction of NHM R1164 may have altered the suture pattern. See Figure 1 for abbreviations. Broken lines indicate broken margins or the limit of the figure.

342 JOURNAL OF VERTEBRATE PALEONTOLOGY, VOL. 25, NO. 2, 2005 ally considered a junior synonym of I. communis (McGowan, 1974; McGowan and Motani, 2003). A distinctive possibility exists that these specimens represent a valid species, possibly belonging to a genus of its own. Unfortunately, these skulls are never associated with forefins, whereas the forefin of Ichthyosaurus is unique and diagnostic (Motani, 1999c). Therefore, it is difficult to pursue this taxonomic issue further at this point. At least two isolated forefins from the Lower Lias of England cannot be associated with any known species (Motani, 1999c), and the possibility that these fins belong to some of the skulls identified as I. intermedius cannot be eliminated at this point. Also, yet another new ichthyosaur from English lower Lias is being described by the author, as reported by McGowan and Motani (2002). The diversity of lower Liassic ichthyosaurs was probably higher than we currently understand, even in England alone. ACKNOWLEDGMENTS I thank S. Chapman, A. Milner, K. Seymour, and R. Wild for access to the specimens in their care. C. McGowan provided financial support to this project, made available enlarged prints of his published photographs, and discussed ichthyosaurian paleontology with the author in the NHM collection and at ROM. J. Massare, K. Padian, and R. Reisz read the manuscript and provided useful suggestions. The project was supported by a Natural Sciences and Engineering Research Council grant to C. McGowan (A9550). LITERATURE CITED Dennison, S. S., P. L. Smith, and H. W. Tipper. 1990. An Early Jurassic ichthyosaur from the Sandilands Formation, Queen Charlotte Islands, British Columbia. Journal of Paleontology 64:850 853. Fraas, E. 1891. Ichthyosaurier der süddeutschen Trias- und Jura- Ablagerungen. H. Laupp, Tübingen, 81 pp. Godefroit, P. 1994. Les reptiles marins du Toarcien (Jurassique Inférieur) Belgo-Luxembourgeois. Mémoires pour servir à l Éxplication des Cartes Géologiques et Minières de la Belgique 98:1 98. Home, E. 1814. Some account of the fossil remains of an animal more nearly allied to fishes than any other classes of animals. Philosophical Transactions of the Royal Society of London 101:571 577. Kirton, A. M. 1983. A review of British Upper Jurassic ichthyosaurs. Unpublished Ph.D. thesis, University of Newcastle-upon-Tyne, 239 pp. Maisch, M. W. 1997. A case against a diapsid origin of the Ichthyosauria. Neues Jahrbuch für Geologie und Paläontologie, Abhandlungen 205:111 127 McGowan, C. 1973. The cranial morphology of the Lower Liassic latipinnate ichthyosaurs of England. Bulletin of the British Museum (Natural History), Geology 24:1 109. McGowan, C. 1974. A revision of the latipinnate ichthyosaurs of the Lower Jurassic of England (Reptilia: Ichthyosauria). Life Sciences Contributions, Royal Ontario Museum 100:1 30. McGowan, C. 1978. Further evidence for the wide geographical distribution of ichthyosaur taxa (Reptilia: Ichthyosauria). Journal of Paleontology 52:1155 1162. McGowan, C. 1979. A revision of the Lower Jurassic ichthyosaurs of Germany with descriptions of two new species. Palaeontographica, A 166: 93 135. McGowan, C. 1991. Dinosaurs, Spitfires, and Sea Dragons. Harvard University Press, Cambridge, Massachusetts, 365 pp. McGowan, C., and R. Motani. 2002. Two new ichthyosaurs from English Lias causing a polytomy. Journal of Vertebrate Paleontology 22: 86A. McGowan, C., and R. Motani. 2003. Ichthyopterygia. Handbuch der Paläoherpetologie Part 8. Verlag Dr. Friedrich Pfeil, München. 175 pp. Motani, R. 1999a. The skull and taxonomy of Mixosaurus (Ichthyopterygia). Journal of Paleontology 73:917 928. Motani, R. 1999b. Phylogeny of the Ichthyopterygia. Journal of Vertebrate Paleontology 19:472 495. Motani, R. 1999c. On the evolution and homology of ichthyosaurian forefins. Journal of Vertebrate Paleontology 19:42 49. Owen, R. 1881. A Monograph of the Fossil Reptilia of the Liassic Formations. Part III. Palaeontographical Society, London. (Ichthyopterygia, pp. 83 134.) Romer, A. S. 1956. Osteology of the Reptiles. The University of Chicago Press, Chicago, 772 pp. Romer, A. S. 1968. An ichthyosaur skull from the Cretaceous of Wyoming. Contributions to Geology, University of Wyoming 7:27 41 Sollas, W. J. 1916. The skull of Ichthyosaurus studied in serial sections. Philosophical Transactions of the Royal Society of London, B 208: 63 126. Submitted 12 July 2004; accepted 7 August 2004.