Gorden L. Bell Jr.
Museum of Geology
South Dakota School of Mines & Technology
Rapid City, SD 57701

Fragmentary remains of mosasauroids have recently been recovered in Big Bend National Park from lower portions of the Ernst Member of the Boquillas Formation. Some of these are equivalent in age to the previously oldest known mosasaurs from North America but others are possibly older. The youngest of these remains are associated with Collignoniceras woolgari which indicates an early Middle Turonian age. Foraminiferal correlations suggest that the oldest material may be Early Turonian age but inoceramid correlations suggest a slightly younger age. Preserved material presents two characters which suggest the represented taxon is more closely related to a Tylosaurinae + Plioplatecarpini clade than to a Mosasaurinae clade. This also indicates that a divergence between these two major clades occurred during or before the early Middle Turonian. Presence of very small vertebrae in this interval suggest that some conservative mosasauroid or possibly a dolichosaur was also present. This is the only report of vertebrate fossils from the Ernst Member other than indeterminate chondrichthyans.

The purpose of this paper is to describe and discuss some new finds of early mosasaurs and closely related specimens of a less derived nature. The latter show some morphological similarity with taxa and specimens previously referred to the family Aigialosauridae [a group that includes Aigialosaurus dalmaticus Kornhuber, Opetiosaurus bucchichi Kramburger, Carsosaurus marchesetti Kramburger, and the "Trieste aigialosaur" (Carroll and DeBraga, 1992)], and perhaps to the Dolichosauridae (Romer, 1956:562). The new finds are very fragmentary but provide information concerning a minimum age for diversification and radiation of the two most successful clades of mosasaurs. They also indicate that lizards with features similar to the above specifically named taxa were more widespread than their previously known distribution in Croatia, Slovenia, and England.

Camp (1923:321) included Aigialosauridae within the superfamily Varanoidea and mosasaurs as the superfamily Mosasauroidea, even though several authors including Camp (1923), Dollo (1904), Knopcsa (1903), and Williston (1904) believed aigialosaurs to be the ancestors of mosasaurs. A recent phylogenetic analysis (Bell, 1993:243) provides evidence that Aigialosauridae is only an artificial grouping, the members of which are associated by possessing numerous plesiomorphic characters, especially in the limbs. Because some of these more conservative taxa are imbedded within phylogenetic lineages which include acknowledged typical mosasaurian taxa, this author chooses to include those basic taxa previously referred to Aigialosauridae (Romer, 1956:560) within Mosasauroidea.

In the following discussion, use of informal names will follow these criteria: mosasauroid refers to members of the most generalized taxon, Mosasauroidea, which includes the less derived forms (previous members of paraphyletic Aigialosauridae) and Mosasauridae; mosasaurid or mosasaur refers to members of Mosasauridae, which includes all of the more derived, fully marine forms (including Halisaurus); and mosasaurine refers to members of a diverse lineage which includes Mosasaurus (as opposed to a lineage which includes Tylosaurus and Platecarpus). (See Russell, 1967:202.)

Mosasauroid fossils have a worldwide distribution and are known from marine Cretaceous rocks ranging from Late Cenomanian to terminal Maestrichtian age (Russell, 1967). However, very little is known of the representatives of this group before Coniacian time. Basal members with relatively underived terrestrial limbs have been found in Slovenian and Croatian rocks of Late Cenomanian to Turonian age (Langer, 1961; Carroll and DeBraga, 1992). The oldest record of derived mosasaurids consist of small vertebral fragments from the latest Cenomanian (Holaster subglobosus Zone) of England (Woodward, 1905:187). The latter have never been figured or described, therefore verification of this record has not been established. Cenomanian mosasaurids are also recorded from Texas (Stenzel, 1944:443; Thurmond, 1969), but examination of the two jaw portions has resulted in their reinterpretation as Pachyrhizodus leptopsis, a teleost (Stewart and Bell, 1989; Stewart and Bell, 1994).

The next oldest mosasaurids are known from a few fragments from the Fairport Chalk Member and the Blue Hill Shale Member of the Carlile Shale in Kansas (Martin and Stewart, 1977). A single maxillary fragment was recovered from 6 meters below the top of the Fairport Chalk. The entire Fairport Chalk is included within the Collignoniceras woolgari Zone, which is assigned by Cobban (1984:76) to the earliest part of the Middle Turonian. The jaw fragment is unidentifiable. Martin and Stewart (1977) reported also two posterior caudal vertebrae from the lower Carlile Shale. Stewart (pers. com., 1994) indicated this specimen is actually from the Blue Hill Shale Member. Martin and Stewart remarked that these caudals resemble those of Clidastes, but stop short of referring them to that taxon. In that same paper, a specimen consisting of four vertebrae and fragments, also collected from the Blue Hill Shale Member, was referred to Clidastes. The Blue Hill Shale Member is considered to be late Middle Turonian age (Hattin, 1962:79).

Within the past three years, the author has discovered fragmentary remains of mosasauroids in the eastern end of Big Bend National Park, Texas. These occurrences are near the bottom of a shaley sequence about 30-40 meters above the base of the Ernst Member of the Boquillas Formation. These are also the first vertebrate fossils other than unidentified chondrichthyans reported from the Ernst Member (Langston et al., 1989:19). Ammonites directly associated with the stratigraphically highest fragments are identified as Collignoniceras woolgari. Therefore, the minimum age for these remains is equivalent to the oldest mosasaurids from Kansas.

All Big Bend National Park specimens reported herein are reposited at the Vertebrate Paleontology Laboratory of the Texas Memorial Museum at Balcones Research Center, University of Texas at Austin. Abbreviation for that facility is TMM.


Mosasauroid remains were recovered from two localities in Big Bend National Park. The first locality is on bluffs along the Rio Grande River from a point 50 meters downstream of the Hot Springs to a point about 150 meters upstream. Here, a fragment of a mosasaur prootic (TMM 43352-1) was found within a vertebrate hash bed situated 38 meters above the formational contact with the Buda Limestone.

At the same locality, 42 meters above the base of the Boquillas Formation, two caudal vertebrae (TMM 43352-2, 43352-3) were recovered from thin calcarenite beds. These were also associated with fragments of other vertebrates, including mostly elasmobranchs and osteichtyans. Only 0.4 meters above this level, a single tooth (TMM 43352-4) was found near the top of a bed of fine grained, finely laminated, argillaceous limestone. This same horizon also provided two specimens of Collignoniceras woolgari (Mantell).

Frush and Eicher (1975) sampled the foraminiferans from this locality and found (fig. 3) that the Cenomanian-Turonian boundary occurred 37 meters above the base of the Boquillas Formation, based on the last occurrence of Rotalipora. These results are also supported by Sanders (1988) in an unpublished masters thesis. Sanders found essentially the same 18O and 13C excursions noted near the Cenomanian-Turonian boundary in the mid-continental Greenhorn Limestone (Pratt, 1985) between 33 and 42 meters above the base of the Boquillas section at the Hot Springs locality.

The Cenomanian-Turonian boundary as established by foraminiferans occurs only about 1 meter below the lowest mosasaurid fragment found. Only 5.4 meters above the boundary, C. woolgari indicates an early Middle Turonian age: thus, the entire Lower Turonian could be represented within about 5 meters or less of limestone and marlstone.

However, at Pico Etereo just east of Big Bend N. P., Powell (1965:521) picked the base of the Turonian no higher than 22.5 meters (75 ft.) above the lower contact with the Buda Limestone, based on the molluscan fauna, especially Tragodesmoceroides and Inoceramus labiatus. By these criteria, the Lower Turonian may be as much as 15 meters (4X) thicker than indicated by foraminiferans.

The second mosasauroid-bearing locality is along the north wall of the canyon about 200 meters below (west of) Ernst Tinaja, which is about 11.5 km (7.5 mi) north of the Hot Springs locality. Two very small mosasauroid vertebrae were found in a thick (0.4m) bed of calcarenite, the base of which lies 32.4 meters above the Buda-Boquillas contact. This is approximately equivalent to the mosasaur-bearing interval at the Hot Springs, based on position within the shaly interval and several associated bentonite layers as well as a rusty weathering, siliceous limestone marker bed that occurs uniformly about 12 meters above this horizon in both outcrops. Sanders (1988, fig. 2.1) also measured this section but did not report isotope data. These two vertebrae, a posterior trunk (TMM 43056-3) and an anteriormost caudal (TMM 43056-4), were found only centimeters apart and could possibly belong to the same individual. Their age is assumed to be essentially equivalent to that of the specimens from the Hot Springs section.


The prootic fragment (TMM 43352-1) is from the right side but the paroccipital process as well as the anteroventral and anterodorsal rami are broken. The lateral side is fairly well preserved and shows several structures of the inner ear, including the fenestra ovale, the anterior vertical and horizontal semicircular canals, and entrances for the seventh and eighth cranial nerves. Numerous small nutrient canals and foramina are preserved within these structures, indicating that the individual was not fully grown. Its size is approximately equivalent to that of a medium sized Clidastes propython, but the morphology clearly is unlike that taxon or any other mosasaur for which this morphology is known. The material cannot be assigned to any taxon less general than Mosasauridae.

One of the two caudal vertebrae (TMM 43352-2) found at the Hot Springs locality is relatively well preserved (Fig. 2A) and measures 36 mm. in length. It has the bases of broadly ovate transverse processes and part of the neural spine. There is no indication of a postzygopophysis on the preserved part of the neural spine. In posterior aspect, the centrum has a hexagonal outline that is higher than wide, although some lateral compression is evident. Ventrally, two facets are present for articulation of the hemal arches. This vertebra is from immediately posterior to the pygal series.

Lack of fusion of the hemal arches to the caudal centra (until late in life) is a plesiomorphic character present in all the basal Slovenian and Croatian mosasauroids and retained throughout a clade of relatively derived mosasaurs including all the known basic taxa within Tylosaurus, Ectenosaurus, Platecarpus, and Plioplatecarpus (Bell, 1993:125). Fusion of the arch to the centrum is a diagnostic character of Mosasaurinae and is convergent in Halisaurus; therefore, TMM 43352-2 cannot represent a member of either of these latter two clades.

TMM 43352-4 is a marginal tooth (Fig. 2B) that measures 14 mm. high and about 6.5 mm. near the base of the enamel. The portions of the tooth below the crown are not preserved. A distinct, non-serrate carina is exposed on the posterior side but the anterior side is imbedded within marlstone matrix and the fragile nature of the specimen prevents removal. A cross section of the crown would seem to be essentially round, but with the lingual side slightly more convex. Several striations are apparent on the lingual surface of the crown near the base and extend upward more than half the distance to the tip.

Lingual (medial) tooth striations are a synapomorphy contributing to the diagnosis of a major mosasaur clade that includes all the known basic taxa within Tylosaurus, Ectenosaurus, Platecarpus, and Plioplatecarpus (Bell, 1993:105). This specimen then indicates that the divergence of this clade from mosasaurine mosasaurs had already occurred by early Middle Turonian time.

The two vertebrae found near Ernst Tinaja (Fig. 2C) are quite small. TMM 43056-3 is 8.5 mm long and TMM 43056-4 is approximately 5 mm long. The former is still partially imbedded within fine calcarenite matrix with the ventral surface exposed. This surface is essentially flat with two small nutrient foramina near the midline. The lateral surfaces of the centrum are also relatively flattened, which results in a relatively sharp ventrolateral ridge that trends posteriorly from the synapophysis to the posterior condylar surface. Pre- and postzygopophyses are present, and a zygosphene seems to be exposed on a broken surface of the matrix. The posterior and anterior outlines cannot be seen without further preparation. Size of the vertebra and presence of zygosphenes implies relationships with the most conservative mosasauroids, dolichosaurs, or Coniasaurus. However, this vertebra is unlike Opetiosaurus bucchichi, the Trieste specimen, and Coniasaurus crassidens in that those taxa have a pair of rounded ventral ridges immediately adjacent to the ventral midline of the posterior trunk vertebrae. The condition is unknown in Aigialosaurus dalmaticus and, in the dolichosaurs where the ventral posterior trunk vertebrae are visible, those seem to be more convex. Another possibility is that this vertebra belongs to a ontogenetically young mosasaur individual; however, the sharp ventrolateral ridge has not been observed on any mosasaur specimen regardless of age.

The small caudal vertebra is exposed in dorsal aspect with the ventral side and parts of transverse processes imbedded within matrix. Posterior edges of the transverse processes are exposed on the broken rock surface and identify this as a pygal vertebra. These transverse processes are quite elongate laterally and very thin dorsoventrally. The thinness is similar to the condition seen in Coniasaurus crassidens, but the processes of this vertebra are not as long relative to the centrum diameter. The neural arch and spinous process are broken away. The posterior profile is subrounded and slightly dorsoventrally depressed. The posterior condyle is 3.9 mm. wide and 3.1 mm. high. A single large nutrient foramen penetrates the dorsal surface of the centrum within the neural canal. The overall proportions and general structure of the neural canal are quite unlike that of Coniasaurus crassidens, but similarity to any of the Yugoslavian taxa, including dolichosaurs, cannot be ruled out.

These two small vertebrae are distinctly unlike Coniasaurus crassidens Owen, which is the only New World taxon referred to Dolichosauridae (Bell et al., 1982) and otherwise found only in England. They are also distinctly unlike the basal mosasauroid from Dallas, Texas, reported by Bell (1993:239) in which a cross section of the posterior vertebral centra is essentially round, not trapezoidal as would be the Big Bend vertebra. Thus, of the two most similar taxa known from Texas (and North America), these specimens cannot be referred to either. The dolichosaurs from Yugoslavia or Dolichosaurus from England and Aigialosaurus dalmaticus from Yugoslavia cannot be definitely eliminated from a list of taxa to which these vertebrae might be referred. It is also quite possible that they represent an unknown new taxon.


New specimens of mosasaurs from Big Bend National Park are at least as old as the oldest previously reported from North America and some are potentially older, but not as old as the English specimens. These demonstrate that small amounts of material can have important consequences. The parts that were recovered do not provide extensive of information but are sufficient to indicate that a major divergence between the mosasaurine lineage and the lineage including Tylosaurinae + Plioplatecarpini took place during or before the Early Turonian.

Additional tiny vertebrae from Ernst Tinaja suggest that smaller semi-aquatic forms similar to those from the Cenomanian-Turonian of Slovenia, Croatia, and England were also present during the Early or early Middle Turonian in west Texas. Whether these remains represent basal mosasauroids or dolichosaurs is unknown, but they cannot with certainty be referred to any of the previously described forms. It is hoped that discovery of more complete specimens will help resolve these problems.


I extend thanks to Ernest Lundelius and Melissa Winans of the Vertebrate Paleontology Laboratory at Balcones Research Center, University of Texas at Austin for allowing collection using their permit and for expediting applications with the U. S. Park Service. Thanks also to the U. S. Park Service personnel at Big Bend National Park, especially Phil Kepp, who not only arranged permits and logistics, but was also a successful member of our prospecting crew. Thanks to the rest of the crew: my wife Patti, James Lamb, Jr., and Kathy Wallace. Ken Barnes provided a place to store numerous samples, an occasional crash pad, and always interesting discussions about Big Bend fossils. W. A. Cobban verified the ammonite identifications. James E. Martin and J. D. Stewart reviewed and provided constructive comments for an early version of this paper. Thanks also to the Foundation office and the Haslem Postdoctoral Endowment at the South Dakota School of Mines and Technology for funding the field work. And last, but not least, thanks for the Hot Springs.

Bell, B.A., Murray, P.A. and Osten, L.W., 1982. Coniasaurus Owen 1850, from North America: Journal of Paleontology, v. 56(2), p. 520-524.

Bell, G.L., Jr., 1993. A Phylogenetic Revision of Mosasauroidea (Squamata), <Ph.D. Dissertation>: Austin, University of Texas, 293p.

Camp, C.L., 1923. Classification of the lizards: American Museum of Natural History Bulletin, v. 49, p. 289-241.

Carroll, R.L. and DeBraga, M., 1992. Aigialosaurs: mid-Cretaceous varanoid lizards: Journal of Vertebrate Paleontology, v. 12(1), p. 66-86.

Cobban, W.A., 1984. Mid-Cretaceous ammonite zones, Western Interior, United States: Geological Society of Denmark   Bulletin, v. 33, p.71-89.

Dollo, L., 1904. L'Origine des mosasauriens: Bulletin de la Societe belge de Geologie, Paleontologie d'Hydrologie, v.18, p. 217-222.

Frush, M.P. and Eicher, D.L., 1975. Cenomanian and Turonian foraminifera and paleoenvironments in the Big Bend region of Texas and Mexico: Geological Association of Canada Special Paper No. 13, p. 277-301.

Hattin, D.E., 1962. Stratigraphy of the Carlile Shale (Upper Cretaceous) in Kansas: State Geological Survey of Kansas Bulletin No. 156, 155 p.

Knopcsa, F., 1903. Uber die Varanus-artigen Lacerten Istriens: Beitrag zur Palaeontologie und Geologie Osterreich-Ungarns und des Orients, v. 15, p. 31-42.

Langer, W., 1961. Uber das Alter der Fischschiefer von Hvar-Lesina (Dalmatien): Neues Jahrbuch fur Geologie und   Paleontologie, Monatshefte 1961, p. 329-331.

Langston, W., Jr., Standhart, B.R. and Stevens, M.S., 1989. Fossil vertebrate collecting in the Big Bend - history and retrospective; in Busbey, A.B., III, and Lehman, T.M., eds., Vertebrate Paleontology, Biostratigraphy and Depositional Environments, Latest Cretaceous and Tertiary, Big Bend Area, Texas: Field Trip Guidebook for the 49th Annual Meeting of the Society of Vertebrate Paleontology, Austin, Texas, 90p.

Martin, L.D. and Stewart, J.D., 1977. The oldest (Turonian) mosasaurs from Kansas: Journal of Paleontology, v. 51(5), p. 973-975.

Powell, J.D., 1965. Late Cretaceous platform-basin facies, northern Mexico and adjacent Texas: Bulletin of the American Association of Petroleum Geologists, v. 49(5), p. 511-525.

Pratt, L., 1985. Isotopic studies of organic matter and carbonate in rocks of the Greenhorn marine cycle: Society of Economic Paleontologists and Mineralogists Field Trip Guidebook 4:38-48.

Romer, A.S., 1956. Osteology of the Reptiles: Chicago, University of Chicago Press, 772p.

Russell, D.A., 1967. Systematics and Morphology of American Mosasaurs: Peabody Museum of Natural History, Yale University Bulletin No. 23, 240p.

Sanders, R.B., 1988. Sediment and Isotope Geochemistry of the Upper Cretaceous Ernst Member of the Boquillas Formation, Big Bend National Park, Texas <Masters Thesis>: Lubbock, Texas Tech University, 170p.

Stenzel, H.B., 1944. Decapod crustaceans from the Cretaceous of Texas: University of Texas Bulletin No. 4401, p. 401-476.

Stewart, J.D., and Bell, G.L., Jr., 1989. The earliest reputed North American mosasaur records are not mosasaurs: Journal of Vertebrate Paleontology, Supl. Vol. 9, p. 39A.

Stewart, J.D., and Bell, G.L., Jr., 1994. North America's oldest mosasaurs are teleosts: Natural History Museum of Los Angeles County, Contributions in Science No. 441.

Thurmond, J.T., 1969. Notes on mosasaurs from Texas: Texas Journal of Science, v. 21(1), p. 69-80.

Williston, S.W., 1904. The relationships and habits of the mosasaurs: Journal of Geology, v. 12, p. 43-51.

Woodward, A.S., 1905. Note on some portions of mosasaurian jaws obtained by Mr. G.E. Dibley from the Middle Chalk of Cuxton, Kent: Proceedings of the Geological Association, v. 19, p. 185-187.

Fig. 1 - Map of state of Texas with enlargement of Big Bend National Park showing localities and measured sections mentioned in text: ET = Ernst Tinaja, HS = Hot Springs.

Fig. 2 - Mosasauroid specimens recovered from Ernst Member of Boquillas Formation: A - TMM 43352-2, anterior caudal vertebra from Hot Springs locality (arrow designates facet for hemal arch articulation); B - TMM 43352-4, marginal tooth from Hot Springs locality; C - TMM 43056-4, pygal vertebra on left and TMM 43056-3, posterior trunk vertebra on right.

Back to Table of Contents

United States Department of the Interior, National Park Service

Go to Nature Net Home Page Go to Park Net Home Page