Marine reptiles from the Nanaimo Group (Upper Cretaceous) of Vancouver Island

1591 Marine reptiles from the Nanaimo Group (Upper Cretaceous) of Vancouver Island Elizabeth L. Nicholls and Dirk Meckert Abstract: A new fauna of fossil marine reptiles is described from the Late Cretaceous Nanaimo Group of Vancouver Island. The fossils are from the Haslam and Pender formations (upper Santonian) near Courtenay, British Columbia, and include elasmosaurid plesiosaurs, turtles, and mosasaurs. This is only the second fauna of Late Cretaceous marine reptiles known from the Pacific Coast, the other being from the Moreno Formation of California (Maastrichtian). The new Nanaimo Group fossils are some 15 million years older than those from the Moreno Formation. However, like the California fauna, there are no polycotylid plesiosaurs, and one of the mosasaurs is a new genus. This reinforces the provinciality of the Pacific faunas and their isolation from contemporaneous faunas in the Western Interior Seaway. Résumé : Une nouvelle faune de reptiles marins est décrite en provenance du Groupe de Nanaimo (Crétacé tardif) de l île de Vancouver. Les fossiles proviennent des formations de Haslam et de Pender (Santonien supérieur), près de Courtenay, en Colombie-Britannique et ils comprennent des plésiosaures Elasmosauridae, des tortues et des mosasaures. Cette faune est seulement la deuxième de reptiles marins du Crétacé tardif connue sur la côte du Pacifique, l autre provenant de la Formation de Moreno (Maastrichtien) en Californie. Les nouveaux fossiles du Groupe de Nanaimo sont quelque 15 millions d années plus âgés que ceux de la formation de Moreno. Toutefois, tout comme la faune de la Californie, il n y a pas de plésiosaures Polycotylidae et l un des mosasaures représente un nouveau genre. Cela donne du poids à l idée de provincialisme des faunes du Pacifique et leur isolation des faunes contemporaines de la «Western Interior Seaway». [Traduit par la Rédaction] Nicholls and Meckert 1603 Introduction The Late Cretaceous Western Interior Seaway of North America has an abundant and well-studied record of fossil marine reptiles. In contrast to this, few fossil marine reptiles are known from the Pacific coast. Only the Moreno Formation of California (Maastrichtian) has a well-documented fauna of mosasaurs (Camp 1942), plesiosaurs (Welles 1943), and turtles (Foster 1980). More recently, Hilton and Antuzzi (1997), Hilton et al. (1999), and Parham and Stidham (1999) have reported fragmentary specimens of mosasaurs, birds, pterosaurs and turtles from the Chico Formation (Campanian) of California. In Canada, the only published record of a Late Cretaceous marine reptile from the Pacific Coast is a fragmentary turtle from the Nanaimo Group of Vancouver Island (Nicholls 1992), although unidentified mosasaurs and plesiosaurs are know to occur in this unit (Ludvigsen and Beard 1994; Ludvigsen 1996; Nicholls and Meckert 2001). Study of these Nanaimo Group specimens is hampered by the very hard, concretionary matrix in which they are preserved. Only a few of them are adequately prepared to permit a detailed, anatomical study. In this paper, we describe some of these new specimens and their paleobiogeographical significance. All specimens were collected from the Haslam and Pender formations (upper Santonian) along the Puntledge, Browns, and Trent rivers near Courtenay, British Columbia (Fig. 1) and are in the collections of the Courtenay and District Museum (CDM). As in the California faunas, there are no polycotylid plesiosaurs, and the mosasaur is a new genus and species, unknown in the Western Interior. This reinforces the provinciality of the Pacific faunas and their isolation from contemporaneous faunas in the Western Interior Seaway. Geological background The Late Cretaceous Nanaimo Group is an alternating sequence of Upper Cretaceous marine sandstones, mudstones, and conglomerates exposed along the east coast of Vancouver Island and the islands of the Georgia Strait. The sediments are Turonian through early Maastrichtian in age, and biostratigraphic subdivision and correlation has been based on molluscs (Muller and Jeletzky 1970; Ward 1978; Haggart 1991) and Foraminifera (McGugan 1962, 1964, 1990). Its lithology, depositional history and stratigraphy have recently been reviewed by Mustard (1994), and Mustard et al. (1999). The stratigraphic nomenclature of the Nanaimo Group has a Received 15 February 2002. Accepted 9 August 2002. Published on the NRC Research Press Web site at http://cjes.nrc.ca on 18 November 2002. Paper handled by Associate Editor H.-D. Sues. E.L. Nicholls. 1 Royal Tyrrell Museum of Palaeontology, P.O. Box 7500, Drumheller, AB T0J 0Y0, Canada. D. Meckert. 615-4th Street, Courtenay, BC V9N 1H4, Canada. 1 Corresponding author (e-mail: Betsy.Nicholls@gov.ab.ca). Can. J. Earth Sci. 39: 1591 1603 (2002) DOI: 10.1139/E02-075

1592 Can. J. Earth Sci. Vol. 39, 2002 Fig. 1. Map of Vancouver Island, with enlargement of the area around Courtenay, where the specimens were collected. Shaded areas indicate exposures of the Nanaimo Group. complex history. It has been proposed that the sediments of the Nanaimo Group were deposited in two separate sedimentary subbasins the Comox Basin to the north and the Nanaimo Basin to the south. In a number of previous works, different formational names have been applied to the sediments in the two basins (England 1989; Bickford et al. 1990; England and Calon 1991; Ludvigsen and Trask 1995). The Courtenay and District Museum has used the nomenclature associated with the Comox Basin. Consequently, field notes and stratigraphic records associated with the vertebrate fossils discussed in this work indicate that they were collected in the lower part of the Trent River Formation (Nicholls 1992; Ludvigsen and Beard 1994; Ludvigsen and Trask 1995; Ludvigsen 1996; Nicholls and Meckert 2001). Recent work by Mustard (1994) supports the view that the two subbasins are, in fact, the erosional remnants of a single basin, and it is therefore more appropriate to apply a single terminology. This is the terminology adopted here, and the names Haslam and Pender formations are used, following the work of Ward (1978), Haggart (1989), and Mustard (1994). Stratigraphic work by the junior author (DM) has established that one of the specimens discussed in this paper (CDM 003) was collected from the uppermost part of the Haslam Formation. This specimen was originally described as coming from the Browns Member of the Trent River Formation (Nicholls 1992). The remainder of the specimens are found higher in the section and come from the Pender Formation. This was previously considered to be the Royston Member of the Trent River Formation (Ludvigsen and Trask 1995). The Haslam and Pender formations are often considered to be upper Santonian lower Campanian in age (Ward 1978; Haggart 1989; Haggart and Ward 1989). One of us (DM) has found numerous specimens of the crinoid Uintacrinus socialis in the Pender Formation along the Puntledge River. This taxon is generally considered to be restricted to the Santonian (Hancock and Gale 1996; Lamolda and Hancock

Nicholls and Meckert 1593 1996; Kennedy and Kaplan 2000). For this reason, we believe this vertebrate fauna to be upper Santonian in age. Fig. 2. CDM 002, Elasmosauridae gen. et sp. indet. Fragment of left maxilla. Institutional abbreviations AMNH, American Museum of Natural History, New York; CDM, Courtenay and District Museum, Courtenay, British Columbia; CNMH, Chicago Field Museum of Natural History, Chicago, Illinois; KUVP, Museum of Natural History, University of Kansas, Lawrence, Kansas; SMUSMP, Southern Methodist University, Shuler Museum of Paleontology, Dallas, Texas; UCMP, University of California Museum of Paleontology, Berkeley, California. Systematic paleontology Sauropterygia Owen, 1860 Plesiosauria De Blainville, 1835 Family Elasmosauridae Cope, 1869 Genus and species Indeterminate Referred specimen CDM 002. Mandible, cranial fragments, 28 isolated teeth, 66 vertebral centra from all parts of the column, ribs, pubes, ischia, fragments of a coracoid, 4 propodials, 3 epipodials, and numerous gastroliths (Figs. 2 4). Locality and horizon South bank of Puntledge River, 400 m down stream from its confluence with the Browns River. Courtenay, British Columbia. 49 41 07 N, 125 02 50 W. Pender Formation, upper Santonian. Specimen found by Mike and Heather Trask. Description The specimen is poorly preserved. Numerous bite marks are present on some of the bones, and all of the bone surface is heavily eroded and pitted. This suggests that the carcass lay on the sea floor for some time and was subjected to both predation and chemical dissolution. Skull and mandible. Two pieces of the left maxilla are preserved. One piece, 12 cm long, includes the alveolar border. One tooth is preserved in place, but evidence of additional sockets, if any, has been destroyed (Fig. 2). The tooth is straight, circular in cross section, and has a prominent ornamentation of fine striations. The striations are uniform in length and thickness and extend almost to the tip of the crown, around the entire circumference of the tooth. The part of the basicranium preserved includes the occipital condyle, the left exoccipital-opisthotic, the posterior part of the pterygoid plates, and part of the left quadrate. It is badly crushed, and the quadrates are twisted out of shape. The pterygoids meet in the ventral midline for most of their preserved length (10 cm), separating only in the region of the basioccipital. An interpterygoid vacuity, if present, must have been anterior to the preserved part of the pterygoids. This is contrary to the condition in other elasmosaurs, where the pterygoids meet ventral to the basioccipital and posterior to a large interpterygoid vacuity (Welles 1962; Carpenter 1997). The very unusual condition in CDM 002 may be the result of the extensive crushing. The complete right mandibular ramus is 521.3 mm long. On the left ramus, only the dentary is preserved (Fig. 3). The mandibular symphysis is short, measuring 77.7 mm along the dorsal midline, although no sutures are visible. It is impossible to determine the original number of teeth. The bone is so badly corroded that even many of the alveoli have been destroyed. On the right dentary there are four teeth, two of which are broken at the base. There are also two incoming replacement teeth and alveoli for three additional teeth. Gaps in the tooth row leave space for at least six more teeth, suggesting that there could have been any where from 12 16 teeth in each dentary. This is low. Thalassomedon haningtoni has 14 16 dentary teeth (Welles 1943), but most elasmosaurs have 18 22 teeth in the dentary (Brown 1993; Carpenter 1999). The left dentary has four erupting replacement teeth and one large, fully developed, procumbent tooth preserved at the mandibular symphysis. A large tooth socket on the left dentary and a broken root of a very large tooth on the right dentary indicate the presence of enlarged caniniform teeth at a point adjacent to the posterior end of the mandibular symphysis. Twenty eight isolated teeth were found in the concretion containing the skull. Many of these still have the root intact (Ludvigsen 1996, fig. 13.4). All are circular in cross-section and have ornamentation identical to that of the maxillary and dentary teeth. On most teeth, the striations are closely spaced and continuous around the entire crown. On some of the larger teeth, the striations extend the length of the crown only on the lingual side. Postcranial skeleton. Of the 90 preserved vertebral centra, only 32 can be clearly identified as cervicals. The field map made during collection of the specimen (Ludvigsen and Beard 1994, fig. 109) indicates that significant parts of the neck are missing. All of the cervical centra are longer than high and many have the lateral, longitudinal ridge characteristic of the anterior and middle cervicals of many long-necked plesiosaurs (Welles 1943; Brown 1981). All of the centra have a prominent ventral midline keel. However, there is no evidence of the usual ventral, paired foramina. This is probably due to the poor preservation of the bone surface. There is a single rib articulation. In all cases, it is low on the side of the centrum and very elongate extending most of the length of the centrum.

1594 Can. J. Earth Sci. Vol. 39, 2002 Fig. 3. CDM 002. Elasmosauridae gen. et sp. indet. Mandible. BT, broken tooth base; F, isolated fragment; S, socket. Fig. 4. CDM 002, Elasmosauridae gen. et sp. indet. Pelvic girdle. Pu, pubis; Is, ischium; L, ischium length; W, ischium width. The pelvic girdle is almost complete (Fig. 4). The anteroposterior length of the right ischiac plate (Fig. 4, L) is short, even for an elasmosaur, and is slightly less than the mediolateral width of the element along the acetabular neck (Fig. 4, W). The ischiac plate has a pronounced anterior process, which may have contributed to a median pelvic bar. The presence of a median pelvic bar is supported by a posterior extension from the midline of the pubic plates, and suggests the individual was probably an adult. In both the left humerus and right femur, the dorsal tuberosity (or trochanter) is well separated from the articular head, providing additional evidence that the animal was an adult (Brown 1981). Both humerus and femur have a well developed anterodistal knee (Welles 1943, p. 146), and are only moderately expanded distally. The breadth:length ratio of the right humerus is 66 and that of the right femur is 59. This is in the low range for Cretaceous elasmosaurs (Welles 1952, 1962). As in most elasmosaurs, the humerus is longer than the femur. Of the two femora, only the right femur is well enough preserved to retain any anatomical detail. Distally, the fibular facet is extended into a pointed posterior projection. Two of the three epipodials were found associated with the right humerus. One of these is longer than wide but this may be due to crushing, as the other two epipodials are slightly wider than long. However, none has the short, rectangular shape characteristic of most Late Cretaceous elasmosaurs. Both lateral and medial edges of all epipodials are formed of smoothly finished, periosteal bone, and they are gently concave, suggesting the presence of an epipodial foramen. Discussion In spite of the poor preservation, the specimen retains several distinctive characters. These include (1) minimal distal expansion of the propodials; (2) epipodials not greatly shortened, with some shaft retention; (3) short ischiac plates; (4) tooth ornamentation, in which striations are present around the complete circumference of the tooth and extend to the tip of the crown. The first two characters may be considered primitive within the Late Cretaceous Elasmosauridae. Primitively, the epipodials in elasmosaurs are longer than wide, and both sides of the elements are slightly concave, retaining some semblance of the original periosteal shaft (Brown 1981). In the more derived condition, the shaft has been totally lost and the epipodials are short, rectangular blocks, with no foramen between them. In the Courtenay elasmosaur, the epipodials are only a little wider than long, but there is still some evidence of a shaft and the presence of an epipodial foramen. The very short ischiac plate in CDM 002 is a distinctive character. Usually, in elasmosaurs, the length of the ischiac plate exceeds the width. The other elasmosaurs in which this condition is known are juveniles (e.g., Fresnosaurus drescheri Welles 1943; Leurospondylus ultimus Brown 1913). CDM 002 is not a juvenile, so this may be an apomorphy for the species. Tooth ornamentation in this specimen is distinctive, and this is known to be of some taxonomic value in plesiosaurs (Tarlo 1960; Brown 1981; Mulder 1990). The tooth orna-

Nicholls and Meckert 1595 mentation in CDM 002 differs from that in most elasmosaurs in that striations are of uniform length and extend to the tip of the crown. In most teeth, ornamentation is present around the entire circumference. In Libonectes morgani (SMUSMP 69120), Hydrotherosaurus alexandrae (UCMP 33912), and Hydralmosaurus serpentinus (AMNH 5835), well-developed striations are present only on the smaller teeth and only on the lingual side. In these elasmosaurs, the larger teeth lack striations, and the enamel is smooth, although it may have some horizontal wrinkles. Due to these characters, CDM 002 may well prove to be a new taxon. However, the skull is fragmentary, and the specimen is so poorly preserved that it would be unwise to create a name for it at this time. It is left as Elasmosauridae, gen. et sp. indet. until better preserved, or more complete material is found. Referred specimen CDM 018. 16 18 vertebral centra, ribs, and fragments Locality and horizon North bank of Trent River, 300 m upstream from Inland Island Highway bridge, 49 35 45 N, 125 59 00 W. Pender Formation, upper Santonian. Specimen found by Mike Trask and Lynne Roux. Description Only four centra have been prepared. In all of these, the length of the centrum exceeds both the height and bredth, and a long, low rib articulation is present on the ventrolateral sides. The centra are too poorly preserved to tell if there was a ventral keel or notch. Discussion These centra are anterior cervical vertebrae of an elasmosaur, but can not be identified further. Lepidosauria Squamata Lacertilia Family Mosasauridae Gervais, 1853 Subfamily Mosasaurinae Williston, 1897, revised Bell 1997 Tribe Leiodontini Lingham-Soliar, 1995 Genus Kourisodon n. gen. Etymology Kouris, from the Greek κουρις (razor), and odon, from the Greek δονς (tooth), referring to the razor-like, laterally compressed teeth. Type species Kourisodon puntledgensis Diagnosis Small mosasaur, diagnosed by the following combination of characters: prominent supraorbital process, extensive edentulous prow on dentary, 14 16 teeth on the maxilla, teeth laterally compressed shearing blades, humerus with well-developed ectepicondyle and entepicondyle, radius distally expanded, metacarpal I expanded proximally. Differs from other members of the Mosasaurinae in the short maxillary premaxillary suture and the slender shape of the surangular. Differs from Leiodon in its small adult size, the presence of an edentulous prow on the dentary, and a short maxillary premaxillary suture. Kourisodon puntledgensis n. sp. Etymology After the Puntledge River, British Columbia, where the specimen was found. Holotype CDM 022. Incomplete skull and mandible, vertebral column and ribs, right humerus, left pectoral girdle, and front limb (Figs. 5 13). Locality and horizon South bank of Puntledge River, 910 m down stream from confluence with Browns River, 740 m upstream from power house. 49 41 22 N, 125 02 17 W. Pender Formation, upper Santonian. Specimen found by Joe Zanbilowicz. Diagnosis As for genus. Description The specimen was somewhat disarticulated, and cranial elements had disassociated. Consequently, the exact number of vertebrae and the identification of some carpal elements are uncertain. The specimen is small for a mosasaur, with an estimated length of only about 3.75 metres. It is assumed to be an adult as neural arches are fused to the centra, articular ends of the limb bones are well formed, and the carpus is fully ossified. Skull (Fig. 5). The rostral portion of the paired premaxillae is V-shaped in dorsal profile. There is a short predental rostrum, but this may be exaggerated by crushing. No teeth are preserved on the premaxillae, but there are alveoli for the usual two teeth on each element. The anterior alveoli are much smaller than the posterior ones, and there is no indication that the teeth were procumbent. The left maxilla is incomplete. The right maxilla, however, is almost complete (Fig. 6), missing only the posterior tip ventral to the orbit. Four complete teeth are preserved in place. Three teeth are broken off at the base, and there are seven positions for replacement teeth. This results in 14 preserved tooth locations. Assuming that no more than one to two teeth are missing posteriorly, the tooth count for the right maxilla is 14 16. The suture for the premaxilla rises quite sharply extending from the first to fourth maxillary tooth. The concave border for the anterior lobe of the external naris extends from the fourth to the seventh tooth positions. The left prefrontal is badly crushed, but the right prefrontal is complete and well preserved. The impressions for the overlapping of the frontal are clearly impressed on its dorsomedial surface. It is not possible to tell whether or not the prefrontal enters the naris.

1596 Can. J. Earth Sci. Vol. 39, 2002 Only the anterior tip of the frontal is damaged. There is a low midline ridge on the dorsal surface of the frontal, which commences at the level of the supraorbital crest and extends anteriorly. The frontal parietal suture closely resembles that of Clidastes (Bell 1997, fig. 6A), with two crescentic prongs of the frontal extending into the parietal. Laterally, the suture extends out to the smoothly rounded ala. The frontal enters the dorsal border of the orbit. On the ventral surface of the frontal, the prefrontal and postorbitofrontal do not meet, but are separated by a well-defined buttress. The olfactory tract is open ventrally. The parietal foramen is small (smaller than the stapedial foramen) and the parietal table is roughly rectangular in shape. While the sides of the parietal table converge slightly in the middle, they do not meet. Both jugals are preserved. The two rami of the jugals make an obtuse angle of about 120, and there is a well-developed posteroventral process. The quadrate does not differ from that of Clidastes. The suprastapedial process is constricted medially, and there is a small, distinct infrastapedial process. There is no groove in the rim of the tympanic ala, and the stapedial foramen is oval (Fig. 7). In the basioccipital basisphenoid unit, the basal tubera are parallel sided and extend out about 45 from the horizontal. There is no canal for the basilar artery. Mandible Most elements of both mandibular rami are present, although they had disassociated and many are broken. Neither dentary is complete enough to determine the number of mandibular teeth. In the left dentary, the rostral tip and five teeth are preserved. A distinctive feature is the presence of an edentulous prow, anterior to the first tooth (Fig. 8A). This edentulous space is equivalent to a full tooth space and is comparable to that in Tylosaurus (Russell 1967), although it is rounded rather than rectangular in lateral view. The coronoid is very narrow, both anteriorly and posteriorly, and its posterior ascending process is not prominently developed (Fig. 8B) The surangular is not as broadly expanded as in some mosasaurines, and posteriorly its dorsal and ventral edges are subparallel. The tip of the retroarticular process is broken. (Fig. 8C) Dentition The number of teeth is known only on the premaxilla, which has alveoli for the usual two teeth. The maxilla has at least 14 teeth and may have had up to 16. The numbers of teeth on the dentary and pterygoid are unknown. The teeth are very distinctive. All teeth are posteriorly curved and laterally compressed, sometimes to the point of being blade-like. They have two prominent carinae that lack serrations. Enamel is smooth, although there is a poorly developed, irregular faceting on a few teeth. The largest tooth crown (maxillary tooth 9, Figs. 6, 9, 10) has a midline height of 18 mm. The degree of lateral compression increases posteriorly along the tooth row. On both the maxilla and the dentary, anterior tooth crowns are only moderately compressed. In these anterior teeth, the carinae are slightly offset from the midline and the labial surface of the crown is flat, while the lingual surface is convex (Fig. 10A). By tooth 9 on the Fig. 5. CDM 022, Kourisodon puntledgensis n. gen., n. sp. Reconstruction of dorsal skull roof. Shaded areas not preserved. F, frontal; J, jugal; M, maxilla; P, parietal; Pm, premaxilla; PrF, prefrontal; PoF, Postorbitofrontal. maxilla and tooth 4 on the dentary, the teeth have become laterally compressed shearing blades, and both labial and lingual surfaces of the crown are subparallel (Fig. 10B). This differs significantly from the teeth of most mosasaurs, which are much more robust and have somewhat inflated crowns (Russell 1967). The only other mosasaur, in which this type of dentition is reported, is Leiodon sectorius (Russell 1967; Lingham-Soliar 1995). Postcranial skeleton The preserved vertebral column includes 7 cervicals, 36 dorsals, and 22 caudals. Judging from the quarry map, many sections of the vertebral column were articulated or semiarticulated, and the number of cervicals is assumed to be complete. While there are no obvious gaps in the vertebral column, some dorsal vertebrae may be missing, and the tail is clearly incomplete. All cervicals have well-developed peduncles for the articulation of hypapophyses. On the axis, this extends along the entire ventral surface of the centrum. On more posterior cervicals, the peduncle is restricted to a large swelling on the posterior ventral surface of the centrum. The articular faces of the cervical centra are depressed and oval in outline. On the anterior dorsal vertebrae, there are well-developed zygosphenes and zygantra, although these are not uniform throughout the column. Many of the dorsal vertebrae have been preserved fused together, so their articular faces are not

Nicholls and Meckert 1597 Fig. 6. CDM 022, Kourisodon puntledgensis n. gen., n. sp. Right maxilla. Numbers refer to tooth positions. Fig. 7. CDM 022, Kourisodon puntledgensis n. gen., n. sp. (A) Right quadrate, posterior view. (B) Left quadrate, medial view. visible. However, anterior dorsal centra have a well-developed ventral keel, and their articular faces are circular, or slightly heart-shaped. Posterior dorsal centra lack the ventral keel, and their articular faces are circular. Posterior dorsal vertebrae have very unusual neural spines. The spines are flared distally, and are much wider (anterioposteriorly) at their tips, than at their bases. These distally flared neural spines overlap the spines of adjoining vertebrae in a manor that must have greatly reduced mobility. At the base of the neural spines, zygapophyses and accessory articulations have fused together, and there is a considerable amount of accessory bone, which extends in sheets across the intervertebral articulations. This is probably the result of a form of ligamentous ossification known as DISH (diffuse idiopathic skeletal hyperostosis; Tanke and Rothschild 1997). It is not uncommon in mosasaurs (Mulder 2001). There are no clearly identifiable sacral vertebrae. Among the 22 preserved caudals, 8 are considered pygals. These have transverse processes, but no haemal arch. The articular faces of these pygals are circular to pentagonal. The remaining caudals retain the remnants of haemal arches, which are fused to the base of the centra. These centra have circular articular faces. Appendicular skeleton The hind limbs and pelvic girdle have not been preserved. The left pectoral girdle and front limb are almost complete. In the right front limb, only the humerus is present. The scapula and coracoid are approximately the same size. The coracoid has a well developed neck, resulting in a symmetrical, fan-shaped element (Fig. 11). The coracoid foramen is centrally located, and there is no anterior emargination to the coracoid plate. The humerus has a prominent postglenoid process (Fig. 12). The articular surface of the head slightly overhangs the flexor surface of the shaft and is confluent with the anteriorly situated pectoral crest. The pectoral crest is high and well developed, extending diagonally, almost half the length of the shaft. Distally, there is a narrow, pointed ectepicondyle and a prominent entepicondyle. On the extensor surface of the shaft, there is a subdued ectepicondylar groove. This groove appears to be a variable character, as it is present on the left, but not the right, humerus. The humerus closely resembles that of Clidastes (Russell 1967, text-fig. 50) except for the variable presence of the ectepicondylar groove and the confluence of the head with the pectoral crest. The epipodials also conform to those of Clidastes. The ulna is hourglass shaped, with a well-developed olecranon. The radius has a narrow proximal end and is broadly expanded distally. Six carpal bones and five metacarpal bones were recovered. Unfortunately their original orientation can only be inferred. One of the carpals has a saddle-shaped notch, separating two articular surfaces. This is assumed to be the intermedium, and implies that the intermedium contacted the ulna. Another is elongate and is assumed to be the pisiform. The metacarpals are more easily identifiable, and metacarpal I is broadly expanded proximally. The carpus is restored in Fig. 13. While the exact position and identity of the other carpals is uncertain, it is certain that at least six well-ossified carpals were present. Discussion This taxon is referred to the subfamily Mosasaurinae, as recently revised by Bell (1997) on the basis of the following characters: (1) pronounced supraorbital process on the prefrontal, (2) condyles of anterior trunk vertebrae round,

1598 Can. J. Earth Sci. Vol. 39, 2002 Fig. 8. CDM 022, Kourisodon puntledgensis n. gen., n. sp. (A) Left dentary. (B) Left coronoid. (C) Left surangular. Fig. 9. CDM 022, Kourisodon puntledgensis n. gen., n. sp. Photograph of right maxilla, showing ninth and tenth maxillary tooth. Anterior to the right. Scale in mm. (3) numerous presacral vertebrae at least 43, (4) haemal arches fused to centra, (5) humerus ectepicondyle present as a distinct prominence, (6) humerus entepicondyle present as a distinct prominence, (7) humerus postglenoid process prominent, (8) ulna contacts intermedium, (9) metacarpal I broadly expanded. The genus has many morphological similarities to Clidastes. In fact, had only the front limb, dorsal skull roof and quadrate been found, we would have had no trouble referring this specimen to Clidastes. However, CDM 022 differs from Clidastes in the prominent edentulous prow on the dentary and the laterally compressed, blade-like teeth. It differs from

Nicholls and Meckert 1599 Fig. 10. CDM 022, Kourisodon puntledgensis n. gen., n. sp. Cross-sectional diagram of maxillary teeth. Anterior (mesial) to the left. (A) Maxillary tooth number 3. (B) Maxillary tooth number 9. Fig. 11. CDM 022, Kourisodon puntledgensis n. gen, n. sp. Left scapula and coracoid. Clidastes, and all other members of the Mosasurinae, in the slender, parallel-sided surangular, the short premaxillary maxillary suture, which terminates above the 3rd maxillary tooth, and the variable presence of an ectepicondylar groove on the humerus. The only other mosasaur with laterally compressed, blade-like teeth is the genus Leiodon. This genus, which has a complex taxonomic history, was recently reviewed by Lingham-Soliar (1995). Direct comparison with Leiodon is difficult, as the genus is based on isolated teeth and jaw fragments. No postcranial material of Leiodon is known. Lingham-Soliar (1995) recognized three species of Leiodon, differentiated from each other on the degree of lateral compression of the teeth. This character must be used with caution, as it is hard to quantify, and varies along the length of the tooth row. In degree of lateral compression of the teeth, CDM 022 closely compares to Leiodon mosasauroides, as defined by Lingham-Soliar (1995, p. 493), and is consequently referred to his tribe Leiodontini within the subfamily Mosasaurinae. However, CDM 022 differs from Leiodon in its much smaller size, in the edentulous prow on the dentary, and in the very short maxillary premaxillary suture. Family Mosasauridae Gervais, 1853 Subfamily, genus, and species indeterminate Referred specimen CDM 19. Twenty one caudal vertebrae, partially articulated. Locality and horizon North bank of Puntledge River, 300 m upstream from power house, 49 41 30 N, 125 02 25 W. Pender Formation, upper Santonian. Found by Richard Boldt and Tim Obear. Description This specimen is still in its plaster field jacket and only partly prepared. Sixteen of the vertebrae are pygals, with transverse processes, but no haemal arch. The articular faces of these centra are circular to subtriangular in shape. The remaining two centra are more posterior caudals and have distinct nodular facets for the articulation of the haemal arch. This specimen has been illustrated in a local field guide (Ludvigsen and Trask 1995, fig. 10) and in Ludvigsen (1996, fig. 13.7). Discussion Due to the articular facets for the haemal arch, this specimen is clearly not referable to the subfamily Mosasaurinae and indicates that there is more than one mosasaur taxon present in the fauna. However, it is not identifiable below the family level. Referred specimen CDM 025. Mandible, teeth, possibly skull. Locality and horizon One km north of Dove Creek Road and New Island Highway, base of Mount Washington Road intersection, 49 44 15 N, 125 06 30 W. Horizon uncertain, uppermost Haslam Formation? upper Santonian. Found by Rick Ross. Description This specimen is in a huge block and is unprepared. The mandibular symphysis and a number of anterior teeth are exposed. This is a very large mosasaur. Crown height of one of the exposed, anterior teeth is 27.2 mm. The teeth are moderately compressed, with an anterior, but no posterior carina. There are no serrations on the carina, and the labial and lingual sides of the teeth are equal. There does not appear to be an edentulous prow anterior to the dentary teeth. Discussion This specimen may be distinguished from Kourisodon by its larger size, the structure of the teeth, and the apparent lack of an anterior edentulous prow on the dentary. However further identification will have to await preparation.

1600 Can. J. Earth Sci. Vol. 39, 2002 Fig. 12. CDM 022, Kourisodon puntledgensis n. gen., n. sp. Left humerus. (A) flexor surface. (B) Extensor surface. Ect, ectepicondyle; EG, ectepicondylar groove; Ent, entepicondyle; H, humeral head; PC, pectoral crest; PgP, postglenoid process. Fig. 13. CDM 022, Kourisodon puntledgensis n. gen., n. sp. Restored left front limb, flexor view. H, humerus; I, intermedium; MC, metacarpal; P, pisiform; R, radius; Ra, radiale; U, ulna; Ul, ulnare. Locality and horizon South bank of the Puntledge River, 3.5 km west of Courtenay. 49 41 00 N, 125 03 24 W, Haslam Formation, 129 m above the base of the underlying Comox Formation, upper Santonian. Chelonioidea Family, genus, and species Indeterminate Referred specimen CDM 021. Mandible, cervical vertebrae. Locality and horizon Southeast side of Trent River, 150 m east of Inland Island Highway, 49 36 00 N, 125 58 30 W. Pender Formation, upper Santonian. Found by Tim Obear. Description The cervicals of this specimen have not been prepared. The mandible appears to be complete, but is only partially prepared. No sutures are visible. It has a short mandibular symphysis and a smooth, flat titurating surface, lacking any ridges. There is a high coronoid process. The distance between the coronoid process and the mandibular symphysis is 131 mm. Testudines Chelonioidea Family Desmatochelyidae Williston, 1894? Desmatochelys lowi Williston, 1894 Referred specimen CDM 003. Mandible, humerus, isolated girdle and carapace fragments. Described by Nicholls (1992). Discussion The genus Desmatochelys was defined by Williston (1894) on cranial characters, and the validity of the family has been established primarily on cranial characters (Smith 1989). However, a combination of mandibular and postcranial characters are distinctive for the genus. These include the short mandibular symphysis, with a single, reduced labial ridge; the large head of the humerus, with a pronounced medial process; and a divided, triangular lateral process restricted to the anterior side of the humeral shaft (Hirayama 1992a, 1997). On the basis of these characters, CDM 003 is tentatively referred to Desmatochelys, until more diagnostic cranial material can be found (Nicholls 1992). The mandible

Nicholls and Meckert 1601 of CDM 021 differs slightly from that of CDM 003. It lacks the low sagittal ridge along the mandibular symphysis, as well as the low labial and lingual ridges. In the absence of any other material, it is left as Chelonioidea indeterminate. Biogeographic significance of the Nanaimo Group fauna Looking at the fauna as a whole, there are very significant differences between the marine reptile fauna described here and contemporaneous marine reptile faunas in the Cretaceous Western Interior Seaway. Being upper Santonian in age, the Nanaimo Group fauna is contemporaneous with the Niobrara fauna of the Western Interior. The most notable difference is the absence of polycotylid plesiosaurs in the Nanaimo Group. In the Niobrara fauna, polycotylids are very abundant, outnumbering elasmosaurs by almost five to one (Russell 1993; Carpenter 1996). Polycotylids are also very abundant in the Sharon Springs Member of the Pierre Shale, which overlies the Niobrara in the Western Interior. In the Sharon Springs Member (early Campanian), polycotylid plesiosaurs are three to four times more abundant than elasmosaurs (Nicholls 1989; Nicholls and Russell 1990). In contrast, all of the plesiosaurs found in the Nanaimo Group are elasmosaurs. It should be noted, however, that the sample size from the Nanaimo Group is still very small. Consequently the absence of polycotylid plesiosaurs from this fauna should be considered a hypothesis, to be tested by the collection of additional material. It is also noteworthy that the one identifiable mosasaur in the Nanaimo Group (CDM 022) is a new taxon, not known in the Western Interior. The same may be true for the elasmosaur (CDM 002), although more complete material is needed to confirm this. No leiodontinine mosasaurs are known from the Western Interior Seaway, in spite of the very extensive collections of fossil marine reptiles from this region. Leiodontinine mosasaurs are known from Campanian Maastrichtian deposits of Europe, North Africa, and the Atlantic seaboard of North America (Russell 1967; Gallagher 1993; Lingham-Soliar 1995). The only faunal similarity between the marine reptile faunas of the Nanaimo Group and the Western Interior Seaway is the common occurrence of the turtle Desmatochelys. All known specimens of Desmatochelys from the Western Interior Seaway are late Cenomanian to early Turonian in age (Elliott et al. 1997). However, Desmatochelys is also known from the Aptian of Colombia (Smith 1989) and the Turonian of Japan (Hirayama 1992b). Clearly Desmatochelys was widely distributed and not restricted to the Interior Seaway. However, the new Nanaimo Group fauna does share many faunal similarities with the Moreno Formation of California. These include the absence of polycotylid plesiosaurs and an endemic fauna of mosasaurs and elasmosaurs. In the Moreno Formation, the absence of polycotylids could be explained by the stratigraphic age of the deposits. The Moreno Formation is Maastrichtian, and polcotylids disappear from the Western Interior Seaway in the mid-late Campanian (Sullivan 1987; Russell 1988; Sato 2001). In the Western Interior, late Campanian Maastrichtian marine deposits are represented by the Bearpaw Formation (Caldwell 1968; Lerbekmo and Braman 2002). Polycotylids are very rare in the Bearpaw, and most of the plesiosaur fauna is elasmosaur (Sato 2001). However, the absence of polycotylids from the Santonian Nanaimo Group suggests that polycotylids were not part of the Late Cretaceous marine vertebrate fauna of the Pacific coast. The only other Late Cretaceous marine reptile fauna known from the north Pacific is from the Yezo and Hakobuchi groups of Japan. These units have an abundant fauna of elasmosaurs, mosasaurs, and turtles (Obata et al. 1989; Suzuki 1985; Nakaya 1989; Hirayama and Chitoku 1994; Tanimoto et al. 1998). This fauna is hard to evaluate, as much of the material is too incomplete to identify below the family level. However, polycotylid plesiosaurs are rare. Only a single polycotylid specimen has been reported (Sato and Storrs 2000) in contrast to the numerous elasmosaurs. Lingham-Soliar (1995) suggested that Mosasaurus hobetsuensis Suzuki 1985 from the Hakobuchi Group (Maastrichtian), may be referable to Leiodon. The Japanese specimen, however, is known only from postcranial material and a single tooth crown. Tanimoto et al. (1998) illustrated some very laterally compressed mosasaur teeth from the Izumi Group (Maastrichtian) of southwest Japan. They referred these to Mosasaurus sp., although they may, in fact, be leiodontinine. The taxonomic status of these specimens must await the discovery of additional material. The marine reptile fossils from the Nanaimo Group represent only the second extensive Cretaceous marine reptile fauna from the west coast of North America. In spite of its greater age, the similarity between the marine reptile fauna of the Nanaimo Group of Vancouver Island and the Moreno Formation of California reinforces the provinciality of the Pacific faunas and their isolation from the contemporaneous faunas of the Western Interior Seaway Acknowledgments We thank the Courtenay and District Museum for making these specimens available for study. Many of the specimens in this study have been found and collected by local residents of the Courtenay and Comox district. These include Michael and Heather Trask, Lynne Roux, Richard Boldt, Tim Obear, Rick Ross, and Joe Zanbilowicz. We appreciate their enthusiasm and dedication. We thank Jim Haggart and Peter Mustard for helpful discussions on Nanaimo Group stratigraphy, and Gorden Bell for sharing unpublished data on mosasaurs. Figures 2, 3, 6, 7, and 12 were drawn by Irene Hughes, and Figs. 8 and 11 were drawn by Joe Morin. Two of the specimens (CDM 002, CDM 022) were prepared by Prehistoric Animal Structures (PAST). We are grateful to Colin Cooke of that organization for his assistance and discussions. This paper has benefitted from the reviews of Robert Holmes and Glenn Storrs, and from the Associate Editor, H.-D. Sues. We thank them for their thoughtful comments. References Bell, G.A. 1997. A phylogenetic revision of North American and Adriatic Mosasauroidea. In Ancient Marine Reptiles. Edited by J.M. Callaway and E.L. Nicholls. Academic Press, San Diego, Calif., pp. 293 332. Bickford, C.G.C., Hoffman, G., and Kenyon, C. 1990. Geological investigations in the coal measures of the Oyster River, Mount Washington and Cumberland areas, Vancouver Island. Province of British Columbia, Ministry of Energy, Mines and Petroleum

1602 Can. J. Earth Sci. Vol. 39, 2002 Resources, Mineral Resources Division, Geological Fieldwork 1989, Paper 1990-1, pp. 431 437. Brown, D.S. 1981. The English Upper Jurassic Plesiosauroidea (Reptilia) and a review of the phylogeny and classification of the Plesiosauria. Bulletin of the British Museum (Natural History), Geology Series, 35: 253 347. Brown, D.S. 1993. A taxonomic reappraisal of the families Elasmosauridae and Cryptoclididae (Reptilia: Plesiosauroidea). Revue de Paléobiologie, 7: 9 16. Caldwell, W.G.E. 1968. The Late Cretaceous Bearpaw Formation in the South Saskatchewan River Valley. Saskatchewan Research Council, Geology Division, Report 8. Camp, C.L. 1942. California mosasurs. University of California, Memoir 13. Carpenter, K. 1996. A review of short-necked plesiosaurs from the Cretaceous of the Western Interior, North America. Neues Jahrbuch für Geologie und Paläontologie Abhandlungen, 20: 259 287. Carpenter, K. 1997. Comparative cranial anatomy of two North American Cretaceous plesiosaurs. In Ancient Marine Reptiles. Edited by J.M. Callaway and E.L. Nicholls. Academic Press, San Diego, Calif., pp. 191 216. Carpenter, K. 1999. Revision of North American elasmosaurs from the Cretaceous of the Western Interior. Paludicola, 2: 148 173. Elliott, D.K., Irby, G.V., and Hutchison, J.H. 1997. Desmatochelys lowi, a marine turtle from the Upper Cretaceous. In Ancient Marine Reptiles. Edited by J.M. Callaway and E.L. Nicholls. Academic Press, San Diego, Calif., pp. 243 258. England, T.D.J. 1989. Lithostratigraphy of the Nanaimo group, Georgia Basin, southwestern British Columbia. In Current research, part E, Geological Survey of Canada, Paper 89-1E, pp. 103 108. England, T.D.J., and Calon, T.J. 1991. The Cowichan fold and thrust system, Vancouver Island, southwestern British Columbia. Geological Society of America Bulletin, 103: 336 362. Foster, D.E. 1980. Osteopygis sp., a marine turtle from the Late Cretaceous Moreno Formation of California. PaleoBios, 34: 1 14. Gallagher, W.B. 1993. The Cretaceous/Tertiary mass extinction event in the northern Atlantic coastal plain. The Mosasaur, 5: 75 154. Haggart, J.W. 1989. New and revised ammonites from the Upper Cretaceous Nanaimo Group of British Columbia and Washington State. In Contributions to Canadian paleontology. Geological Survey of Canada, Bulletin 396, pp. 181 221. Haggart, J.W. 1991. A new assessment of the age of the basal Nanaimo Group, Gulf Islands, British Columbia. In Current research, part E. Geological Survey of Canada, Paper 91-1E, pp. 77 82. Haggart, J.W., and Ward, P.D. 1989. New Nanaimo Group ammonites (Cretaceous, Santonian Campanian) from British Columbia and Washington state. Journal of Paleontology, 63: 218 227. Hancock, J.M., and Gale, A.S. 1996. The Campanian stage. Bulletin de L Institut Royal des Sciences Naturelles de Belgique, Sciences de la Terre, 66(Supplement): 103 109. Hilton, R.P., and Antuzzi, P.J. 1997. Chico Formation yields clues to Late Cretaceous paleoenvironment in California. California Geology, September October: 135 144. Hilton, R.P., Göhre, E.S., Embree, P.G., and Stidham, T.A. 1999. California s first fossil evidence of Cretaceous winged vertebrates. California Geology, July August: 4 10. Hirayama, R. 1992a. Humeral morphology of chelonioid sea turtles: its functional analysis and phylogenetic implications. Hobetsu Museum, Bulletin 8: 17 57. (Japanese, with English abstract.) Hirayama, R. 1992b. Desmatochelys lowi (Family Protostegidae) from Upper Cretaceous (Saku Fm., Middle Yezo Group; Middle Turonian) of Yuubari, Hokkaido, North Japan. 141th regular meeting of the Paleontological Society of Japan, June 20 21, 1992, Iwate, Japan, Abstracts, p. 12. Hirayama, R. 1997. Distribution and diversity of Cretaceous Chelonioids. In Ancient Marine Reptiles. Edited by J.M. Callaway and E.L. Nicholls. Academic Press, San Diego, Calif., pp.225 241. Hirayama, R., and Chitoku, T. 1994. Fossil turtles from the Japanese Cretaceous marine strata. In Evolution and adaptation of marine vertebrates 2. Edited by I. Kobayashi. Association for the Geological Collaboration in Japan, Monograph 43, pp. 17 24. (Japanese, with English abstract.) Kennedy, W.J., and Kaplan, U. 2000. Ammoniten Faunen des hohen Oberconiac und Santon in Westfalen. Geologie und Palaeontologie in Westfalen, 57: 1 131. Lamolda, L., and Hancock, J.M. 1996. The Santonian stage and substages. Bulletin de L Institut Royal des Sciences Naturelles de Belgique, Sciences de la Terre, 66(Supplement): 95 102. Lerbekmo, J.F., and Braman, D.R. 2002. Magnetostratigraphic and biostratigraphic correlation of late Campanian and Maastrichtian marine and continental strata from the Red Deer Valley to the Cypress Hills, Alberta, Canada. Canadian Journal of Earth Sciences, 39: 539 557. Lingham-Soliar, T. 1995. The mosasaur Leiodon bares its teeth. In Vertebrate fossils and the evolution of scientific concepts. Edited by W.A.S. Sarjeant. Gorden and Breach, Amsterdam, The Netherlands, pp. 483 498. Ludvigsen, R. 1996. Ancient Saurians: Cretaceous Reptiles of Vancouver Island. In Life in stone. A natural history of British Columbia s fossils. Edited by R. Ludvigsen. The University of British Columbia Press, Vancouver, B.C., pp. 156 166. Ludvigsen, R., and Beard, G. 1994. West Coast Fossils. Whitecap Books, Vancouver. Ludvigsen, R., and Trask, M. 1995. A Field Guide to the Trent River Formation near Courtenay, B. C. The 1st British Columbia Paleontological Symposium, Courtenay, B.C., April 7 9, 1995. British Columbia Paleontological Alliance, pp. 1 44. McGugan, A. 1962. Upper Cretaceous foraminiferal zones, Vancouver Island, British Columbia, Canada. Journal of the Alberta Society of Petroleum Geologists, 16: 585 592. McGugan, A. 1964. Upper Cretaceous zone Foraminifera, Vancouver Island, British Columbia, Canada. Journal of Paleontology, 38: 933 951. McGugan, A. 1990. Upper Cretaceous (Santonian lower Campanian) foraminiferal biostratigraphy of the Nanaimo Group, subsurface of the Parksville area, eastern Vancouver Island. Bulletin of Canadian Petroleum Geology, 38: 28 38. Mulder, E.W.A. 1990. Ein Elasmosaurierzahn aus der oberen Kreide des St. Pietersberges bei Maastrict, Süd-Limburg, Niederlande. Paläontologische Zeitschrift, 64: 145 151. Mulder, E.W.A. 2001. Co-ossified vertebrae of mosasaurs and cetaceans: implications for the mode of locomotion of extinct marine reptiles. Paleobiology, 27: 724 734. Muller, J.E., and Jeletzky, J.A. 1970. Geology of the Upper Cretaceous Nanaimo Group, Vancouver Island and Gulf Islands, British Columbia. Geological Survey of Canada, Paper 69-25. Mustard, P.S. 1994. The Upper Cretaceous Nanaimo Group, Georgia Basin. In Geology and geological hazards of the Vancouver region, southwestern British Columbia. Edited by J.W.H. Monger. Geological Survey of Canada, Bulletin 481, pp. 27 95. Mustard, P.S., Katnick, D.C., Baker, J. Enkin, R.J., and Mahoney, J.B. 1999. Multidisciplinary studies of the Upper Cretaceous Nanaimo Group, Hornby and Denman Islands, British Columbia. In Current research 1999-A. Geological Survey of Canada, Paper 1999-A, pp. 231 238.