A new dromaeosaurid (Dinosauria: Theropoda) with Asian affinities from the latest Cretaceous of North America
- First Online:
- Cite this article as:
- Evans, D.C., Larson, D.W. & Currie, P.J. Naturwissenschaften (2013) 100: 1041. doi:10.1007/s00114-013-1107-5
- 629 Views
Dromaeosaurids from the Maastrichtian of North America have a poor fossil record and are known largely from isolated teeth, which have typically been referred to taxa based on more complete material from earlier Campanian strata. An almost complete maxilla with well-preserved dentition and an associated dentary from the Hell Creek Formation of Montana are used to establish a new dromaeosaurid taxon in the latest Maastrichtian, immediately prior to the end-Cretaceous extinction event. Acheroraptor temertyorum gen. et sp. nov. is differentiated from other dromaeosaurids on the basis of a hypertrophied postantral wall that projects posteriorly into the antorbital fenestra, a maxillary fenestra positioned low in the antorbital fossa and directly posterior to the promaxillary fenestra, and distinctive dentition with marked apicobasal ridges. The new material allows a dromaeosaurid from the Maastrichtian of North America to be placed within a phylogenetic framework for the first time. Phylogenetic analysis suggests Acheroraptor is a velociraptorine that is more closely related to Asian dromaeosaurids, including Tsaagan and Velociraptor, than it is to Dromaeosaurus, Saurornitholestes, or any other taxon from North America. As part of the Lancian Tyrannosaurus–Triceratops fauna, A. temertyorum is the latest occurring dromaeosaurid. Its relationships and occurrence suggest a complex historical biogeographic scenario that involved multiple, bi-directional faunal interchanges between Asia and North America during the Late Cretaceous.
Dromaeosaurids are a group of carnivorous theropods notable for their close phylogenetic relationship to Aves. Although dromaeosaurids can be inferred to have originated sometime prior to the Late Jurassic, their fossil record is largely restricted to the Cretaceous (Norell and Makovicky 2004; Turner et al. 2012). Dromaeosaurids had a global distribution during the Late Cretaceous, but their fossil record for this time is generally poor. Despite over a century of sustained collecting, dromaeosaurids from North America remain surprisingly poorly known; only eight species have been named, most of which are based on highly incomplete fossil remains. Two taxa are recorded from the Early Cretaceous. Deinonychus antirrhopus from the Lower Cretaceous Cloverly Formation of the western USA is the earliest dromaeosaurid from North America and is known from the most fossil material (Ostrom 1969; Maxwell and Ostrom 1995; Brinkman et al. 1998). Utahraptor ostrommaysorum is represented by fragmentary remains from Aptian sediments of the Cedar Mountain Formation of Utah (Kirkland et al. 1993). All other named dromaeosaurids from North America are based on holotype specimens that are Campanian in age. With the exception of Bambiraptor (Burnham et al. 2000), these taxa are known only from fragmentary specimens or isolated bones and teeth. Dromaeosaurus albertensis (Matthew and Brown 1922; Currie 1995), Hesperonychus elizabethae (Longrich and Currie 2009), and Saurornitholestes langstoni (Sues 1978) are based on incomplete skeletons from the Belly River Group of Alberta. Saurornitholestes robustus is known from isolated bones from the Kirtland formation of New Mexico (Sullivan 2006) and may be a nomen dubium (Turner et al. 2012), and the hypodigm of Atrociraptor marshalli from the uppermost Campanian strata of the Horseshoe Canyon Formation, Alberta, consists of an associated snout and referred isolated teeth (Currie and Varricchio 2004).
Maastrichtian dromaeosaurids from North America are known only from teeth and isolated bones of limited diagnostic value at low taxonomic levels (Longrich 2008; Sankey 2008). Teeth referred to the genera Dromaeosaurus, Saurornitholestes, and Velociraptor have been reported from Maastrichtian-aged sediments throughout North America, including Alaska, but no associated skeletons or diagnostic cranial material has been reported to date (Sloan et al. 1986; White et al. 1998; Russell and Manabe 2002; Norell and Makovicky 2004; Weishampel et al. 2004; Longrich 2008; Sankey 2008; Horner et al. 2011; Lyson and Longrich 2011). Here, we report on a new dromaeosaurid based on an almost complete maxilla with teeth and an associated dentary that were recently recovered from the same mixed faunal bonebed in the Hell Creek Formation, MT, USA. The maxilla and dentition are variable within deinonychosaurs and preserve morphology considered diagnostic at low taxonomic levels (Currie and Varricchio 2004; Turner et al. 2007; Godefroit et al. 2008; Senter et al. 2010). The new taxon is the first dromaeosaurid named on material from the Maastrichtian of North America, and it adds significant new information on the diversity and biogeography of this group immediately prior to the end-Cretaceous extinction event.
Dinosauria Owen 1842
Saurischia Seeley 1888
Theropoda Marsh 1881
Dromaeosauridae Matthew and Brown 1922
Velociraptorinae Barsbold 1983
Acheroraptor temertyorum gen. et sp. nov.
The generic name is derived from Acheron, the River of Pain in the underworld of ancient Greek mythology, in reference to the Hell Creek Formation, and Latin raptor, robber, a suffix used in previously named dromaeosaurids. The name Acheron has been deliberately shortened in the combination for euphony. The specific epithet honors James and Louise Temerty for their outstanding service and contributions to the Royal Ontario Museum.
ROM (Royal Ontario Museum) 63777, an almost complete right maxilla, missing only the posterior portion of the jugal ramus and an associated maxillary tooth complete with root.
ROM 63778, an almost complete left dentary (missing only the teeth and a small posterior portion of the bone) that may pertain to the same individual as ROM 63777.
Horizon and locality
The holotype was found in strata of the Hell Creek Formation (Upper Maastrichtian, Upper Cretaceous), approximately 45 km southwest of the town of Jordan, Garfield County, MT, USA (Fig. S1). The specimen was recovered in situ from a reddish-colored, coarse-grained sandstone containing other small vertebrate fossils and invertebrate shell fragments. The referred dentary, ROM 63778, was found approximately 4 m from the holotype maxilla in the same sandstone-hosted bonebed. Detailed locality data are on file at the ROM.
The referred dentary (ROM 63778) can be assigned to Dromaeosauridae on the basis of fused interdental plates and the curved dentary ramus with parallel dorsal and ventral margins in lateral view. The corresponding size and close physical association of the holotype maxilla and the dentary in the same bonebed suggest that they may belong to the same individual animal. However, this assignment is inconclusive, and therefore, ROM 63778 is provisionally referred to Acheroraptor.
Mid-sized dromaeosaurid that differs from other eudromaeosaurs in the following characters: relatively small anterior region of the antorbital fossa (defined as the region between the anterior borders of the antorbital fenestra and the antorbital fossa); maxillary fenestra that almost reaches the ventral margin of the antorbital fossa and is positioned directly posterior to the promaxillary fenestra; extensive, posteriorly projected postantral wall that is visible laterally through the antorbital fenestra (convergent in Austroraptor); maxillary dentition with unique configuration of prominent apicobasal ridges on both the labial and lingual surfaces of the tooth. Acheroraptor can be further differentiated from Saurornitholestes in having a relatively longer anterior ramus of the maxilla, a much smaller anterior portion of the antorbital fossa, a more robust jugal ramus of the maxilla below the antorbital fossa, and dentition with posterior denticles that are not strongly hooked apically.
The antorbital fossa occupies much of the posterolateral surface of the maxilla, but it is relatively small for a dromaeosaurid (Fig. S3). It is sharply delineated on all sides and has a horizontal ventral margin below the maxillary fenestra. This is unlike the condition in Bambiraptor and Saurornitholestes, where the ventral margin of the antorbital fossa slopes posteroventrally toward the maxillary tooth row in this region (Currie and Varricchio 2004). The anterior margin of the antorbital fossa is located above the fifth maxillary tooth position, as in Atrociraptor, Saurornitholestes, and Velociraptor. Anterior to the fossa, the anterior ramus of the maxilla (=rostral plate of Godefroit et al. (2008)) is large but shallow for a dromaeosaurid and closely resembles that of Tsaagan mangas in its proportions. The lateral lamina extends 42 mm in front of the antorbital fossa and is 37 mm high at the anterior end of the fossa, resulting in a length/height ratio of 1.14. This is similar to the condition in T. mangas (1.13) and Velociraptor spp. (1.3–1.38) but contrasts with that of Atrociraptor (0.69), Bambiraptor (0.62–0.75), Deinonychus (0.9), and Saurornitholestes (0.8), which have proportionately deeper snouts (Godefroit et al. 2008). The distance between the posterior margin of the interfenestral strut and the anterior margin of the antorbital fossa is only 20 mm. The lateral lamina of the maxilla is approximately twice as long as the latter measurement (ratio = 2.25), reflecting the relatively small size of the anterior region of the antorbital fossa. This condition is similar to T. mangas (2.1) but contrasts with the proportionately larger anterior region of the antorbital fossa of most Campanian dromaeosaurids from North America, including Bambiraptor (lamina length = 26 mm, anterior portion of the antorbital fossa = 20 mm, ratio = 0.7) and Saurornitholestes (lamina length = 27 mm, anterior portion of the antorbital fossa = 26 mm, ratio = 1.1).
The anterior portion of the antorbital fossa contains the promaxillary and maxillary fenestrae. The small maxillary fenestra is located in the anteroventral region of a deep, posterodorsally open accessory antorbital fossa, as in Atrociraptor and Saurornitholestes. The accessory antorbital fossa is separated from the antorbital fenestra proper by a narrow interfenestral strut. A similar combination of characters occurs in Bambiraptor feinbergi, D. antirrhopus, S. langstoni, and Shanag ashile (Currie and Varricchio 2004; Turner et al. 2007). The distance between the lower edge of the maxillary fenestra and the dentigerous margin is 23.5 mm, which when divided by the height of the largest maxillary tooth (13.2 mm) results in a ratio of 1.78. This proportion is similar to the condition observed in most dromaeosaurids, in which the height between the maxillary fenestra and the alveolar margin is less than twice that of the height of the largest tooth (1.5–1.6 in Bambiraptor, Deinonychus, and Dromaeosaurus, and 1.8 in Saurornitholestes) but contrasts (2.2) with the proportionately deeper maxilla in Atrociraptor (Currie and Varricchio 2004). The slit-like promaxillary fenestra is not visible in lateral view. It is located in an anterior pocket along the anterior margin of the antorbital fossa, deep to the lateral lamina of the maxilla. The promaxillary strut is a broad and plate-like sheet of bone, and the outlines of the accessory antorbital fenestrae do not overlap in the vertical plane in lateral view, as they do A. marshalli, T. mangas, and Velociraptor osmolskae. Conversely, the morphology of the promaxillary strut of ROM 63777 more closely resembles that of B. feinbergi, D. antirrhopus, and S. langstoni (Godefroit et al. 2008). ROM 63777 differs from all other known dromaeosaurids in that the ventral margin of the promaxillary fenestra is approximately level with that of the maxillary fenestra, with both situated immediately above the ventral margin of the antorbital fenestra in lateral view (Currie and Varricchio 2004).
On the lingual surface of the maxilla, a medially directed horizontal ridge, or ledge (the supraalveolar ridge, Fig. 1), up to 10 mm wide, extends along the entire preserved length of the maxilla. The ridge projects anteriorly to form the base of a well-developed anteromedial process, the anterior end of which is broken, but suggestive that it would have extended well anterior to the main premaxillary–maxillary contact when complete, as in the isolated maxilla referred to Saurornitholestes (TMP 1994.012.0844, Currie and Varricchio 2004). The medial surface of the anteromedial process is grooved for contact with the vomer. Below the horizontal ridge, the interdental plates are fused to the maxilla, and each other, as in other dromaeosaurids (Currie and Varricchio 2004). The individual interdental plates can be discerned by subtle textural differences in the surfaces, the interdental plates being more highly vascularized (Currie 1987).
Above the supraalveolar ridge, the maxillary sinus system appears to be divided into a series of chambers that connect with the antorbital, maxillary, and promaxillary fenestrae. The postantral strut (sensu Witmer 1997), a thin sheet of bone, extends dorsally from the medial ridge to the ventromedial surface of the ascending process and encloses a chamber (the maxillary antrum) medial to the maxillary fenestra. The postantral strut forms the medial border of a pocket that extends deep to the interfenestral strut and posterodorsal to the maxillary sinus system. The posterior margin of the postantral strut is broken and fragmented but is complete dorsally, along the posteromedial margin of the ascending (nasal-lacrimal) process. Although broken, it is clear from the preserved morphology that the postantral strut expanded posteriorly as a large sheet of bone that projected into the antorbital fenestra. This large apron would have been visible in the lateral view of the skull through the antorbital fenestra. A broad, posteriorly projected postantral wall is unusual in dromaeosaurids, where it is known only in Austroraptor cabazai (Novas et al. 2008) and T. mangas (Norell et al. 2006). Although incomplete, it appears that the posterior development of postantral wall in Acheroraptor is much greater than in T. mangas (Norell et al. 2006) but is likely not as extensive as in A. cabazai (Novas et al. 2008). The medial surface of the maxilla anterior to the postantral strut is crushed, reflecting the delicate nature of the thin sheet of bone that encloses the maxillary sinus system medially. As far as can be discerned, the maxillary sinus system in ROM 63777 is similar to that of Saurornitholestes (Currie and Varricchio 2004), Velociraptor (Barsbold and Osmólska 1999), and other theropods (Ostrom 1969; Witmer 1997).
The denticles on the anterior carinae are much smaller than those on the posterior carinae (Table S2), as in V. mongoliensis and all known North American dromaeosaurids other than D. albertensis and U. ostrommaysorum (Currie et al. 1990; Kirkland et al. 1993). The largest denticles range from 6.6 to 7.8 denticles per millimeter on the anterior carina, and 4.4 to 5.0 per millimeter posteriorly. Anterior denticles are low and rounded on partially erupted teeth but worn on erupted teeth (Fig. 2). As in most other dromaeosaurids, the anterior and posterior carinae lie on the midlines of the maxillary teeth, and the carinae do not twist onto the lingual side as in Dromaeosaurus (Currie et al. 1990). Posterior denticles are rounded but are very slightly asymmetrical and apically oriented (Fig. 2), unlike the strongly hooked denticles of A. marshalli and S. langstoni, but similar to those described in T. mangas (Norell et al. 2006). The largest posterior denticles are midway along the posterior carinae but become progressively smaller apically and basally, as is typical in dromaeosaurids. Additionally, the posterior denticles are largest on the second tooth (the first preserved tooth) and decrease in size on more posterior teeth. This pattern does not directly correlate with tooth size. A similar pattern may be present in A. marshalli but has not been reported in other dromaeosaurids.
Morphometric analyses (see Electronic Supplementary Information for more details) of tooth measurements from the specimen (Tables S1 and S2) indicate that the teeth are similar to previously collected isolated dromaeosaur teeth from the Hell Creek and Lance formations and that they are more similar to these teeth than any other North American dromaeosaur (Fig. S4).
ROM 63778 has 15 alveoli in the preserved section of the almost complete ramus, with the possibility of an estimated one or two additional, missing tooth positions. The alveolar lengths are generally shorter than those in the maxilla (Table S1), which is consistent with the pattern in Atrociraptor, Bambiraptor, and Dromaeosaurus and is thus consistent with the possibility that it is from same individual animal as ROM 63777. Although it lacks preserved teeth, fused interdental plates and a more ventrally positioned Meckelian groove preclude identification of the specimen as pertaining to Richardoestesia or Troodontidae, also known from the Hell Creek Formation (Larson and Currie 2013).
The lateral surface of the dentary is perforated by a series of small, irregularly distributed foramina. In lateral view, the dentary is concave dorsally with subparallel dorsal and ventral margins. The anterior terminus of the dentary is deflected from the main ramus to form a slight “chin,” unlike Velociraptor, in which this feature is strongly developed (e.g., AMNH 6515). The anterior end of the dentary is distinctive in being strongly tapered anterior to the third alveolus in lateral view, as in Linheraptor exquisitus and V. mongoliensis, but unlike the blunt termini of Atrociraptor, Bambiraptor, and Saurornitholestes, where tapering is less pronounced and restricted more anteriorly. In dorsal view (Fig. 1g), the dentary ramus is virtually straight and would have met the opposing dentary at an acute angle, as in other dromaeosaurids, including Bambiraptor and Deinonychus.
In order to assess the systematic relationships of Acheroraptor temertyorum, both holotype and referred specimens were coded into a modified dromaeosaurid data matrix (Longrich and Currie 2009) (see Electronic Supplementary Information for details). Additionally, we tested this hypothesis of relationships by including A. temertyorum in the latest analysis of Turner et al. (2012), which constitutes a broader analysis of theropod phylogenetics.
Inclusion of Acheroraptor within a broader analysis of coelurosaur relationships (Turner et al. 2012) produced 9600 MPTs of 2042 steps that resulted in a strict consensus tree where Eudromaeosauria is largely collapsed into a polytomy (Fig. S5). The results of this analysis are not inconsistent with the results presented above based on Longrich and Currie (2009), which provides greater resolution within the dromaeosaur in-group.
The Hell Creek Formation and contemporaneous deposits that date from the late Maastrichtian of North America have been extensively sampled for more than a century, in part because these strata provide the only high-resolution record of terrestrial vertebrate faunas through the last several million years of the Cretaceous and into the Palaeocene. As such, they are important for understanding dinosaur diversity and ecology just prior to the end-Cretaceous extinction event. These intensive collecting efforts have yielded over 300 associated or articulated skeletons to date (Lyson and Longrich 2011). The large-bodied dinosaur fauna is well known (White et al. 1998; Russell and Manabe 2002; Lyson and Longrich 2011; Horner et al. 2011) and is represented by dozens of exquisitely preserved skulls and skeletons of Edmontosaurus, Triceratops, and Tyrannosaurus. Despite this intense sampling, phylogenetically informative remains of small-bodied taxa are scarce (White et al. 1998; Horner et al. 2011), which may be the result of strong taphonomic biases against the preservation of small skeletons (e.g., Brown et al. 2013; Evans et al. 2013). Associated skeletons of maniraptoran dinosaurs are completely unknown (Horner et al. 2011; Lyson and Longrich 2011), and by any count, dromaeosaurids represent less than 3 % of the total dinosaur fossil assemblage (White et al. 1998; Russell and Manabe 2002; Horner et al. 2011).
Prior to the discovery of ROM 63777, dromaeosaurids from the Maastrichtian deposits of North America were known predominantly from isolated teeth. Recent studies of small theropod teeth from the Upper Maastrichtian Lance and Hell Creek formations have predicted the presence of a new, unknown dromaeosaurid species on the basis of the unique morphology of isolated teeth that are identical to those of ROM 63777 (Longrich 2008; Larson and Currie 2013). Authors (Longrich 2008; Sankey 2008; Larson and Currie 2013) have differed in opinion on how many taxa are likely represented by isolated teeth from these units. Some isolated specimens with matching size and denticle morphology lack the distinct apicobasal ridges (Longrich 2008; Sankey 2008), and it is uncertain whether a lack of ridges indicates taxonomic or individual (including positional) variation. Canonical variate analyses indicate that the maxillary teeth preserved in ROM 63777 fall within the range of variation of a large sample of dromaeosaurid teeth from the Lance and Hell Creek formations; they are more similar to these isolated dromaeosaurid teeth than to any other small theropod tooth morphotypes from these formations (Fig. S4a), indicating that most of these isolated teeth likely belong to Acheroraptor. Similar analyses conducted on a larger sample of Campanian and Maastrichtian dromaeosaurid teeth (Larson and Currie 2013) suggest Acheroraptor teeth are closer morphologically to Lance and Hell Creek dromaeosaurid teeth than to any other dromaeosaurid (Fig. S4b). However, Lance and Hell Creek teeth are dimensionally similar to teeth from the Dinosaur Park Formation referred to S. langstoni. We therefore concur with previous studies (Estes 1964; Longrich 2008) that there is little evidence for more than a single dromaeosaurid taxon, A. temertyorum, in the Hell Creek-Lance assemblages, suggesting a low diversity of this group just prior to the end-Cretaceous extinction event.
A. temertyorum is the last known dromaeosaurid in North America, and one of the latest occurring dromaeosaurids in the world, although the precise age of presumably slightly older Maastrichtian taxa from Europe, such as Balaur bondoc (Brusatte et al. 2013) and Asia, such as Adasaurus mongoliensis, are not established with certainty. As such, Acheroraptor extends our knowledge of dromaeosaurid morphological diversity in the latest Maastrichtian. The maxilla is a diagnostic bone in dromaeosaurids (Godefroit et al. 2008), and the unique morphology of ROM 63777 can be differentiated from other members of this clade on the basis of three autapomorphies: hypertrophied postantral wall that projects posteriorly into the antorbital fenestra, a maxillary fenestra located low in the antorbital fossa and directly posterior to the promaxillary fenestra, and distinctive dentition with rounded denticles and prominent apicobasal ridges. Unlike other North American dromaeosaurids, Acheroraptor has a relatively long anterior projection of the maxillary body anterior to the antorbital fossa, as in Tsaagan and Velociraptor, suggesting it also had a relatively long snout compared to other taxa from the Late Creteceous of North America.
This material allows a Maastrichtian dromaeosaur from North America to be put into a detailed phylogenetic context for the first time. Several features of the maxilla clearly set Acheroraptor apart from other North American taxa and suggest close affinities to Asian velociraptorines. Acheroraptor is posited as the sister-taxon to a clade of Adasaurus, Tsaagan, and Velociraptor and is nested within Velociraptorinae, which otherwise consists of only Asian forms. Although dromaeosaurid relationships remain controversial (Longrich and Currie 2009; Turner et al. 2012), within the context of this analysis, Velociraptorinae (sensu Longrich and Currie 2009) is inferred to have originated in Asia, and the lineage that includes Acheroraptor represents a dispersal event from Asia to North America, possibly during the Late Campanian. Fitch optimization of continental areas (Eurasia and North America) onto the resulting cladogram (Fig. 3) suggests a complex biogeographic scenario for Eudromaeosauria that involved several dispersal events in both directions across the Bering Strait in the Late Cretaceous: at least once from North America to Asia, and at least once in the opposite direction involving the Acheroraptor lineage. This is significant given that a predominant direction of dispersal from Asia to North America has been suggested for many other vertebrate groups prior to the Campanian (Russell 1993). However, recent phylogenies of ceratopsians (Xu et al. 2010b), pachycephalosaurs (Schott et al. 2009; Evans et al. 2013), tyrannosaurs (Brusatte et al. 2010), lambeosaurine hadrosaurids (Evans 2010), and saurolophine hadrosaurids (Prieto-Márquez 2010) show evidence of intercontinental dispersal back to Asia via Beringia within the Campanian (Sampson et al. 2011). This historical biogeographic analysis of dromaeosaurids provides new evidence for a more complex pattern of faunal interchange between North America and Asia than has been previously recognized in the latest Cretaceous. With its strong similarities to Asian velociraptorines, the occurrence of Acheroraptor suggests that immigration from Asia also played an important role in the assembly of North American Maastrichtian dinosaur communities.
We extend special thanks to T. Seever who facilitated the acquisition of the holotype by the ROM. The specimen was collected and prepared by CK preparations, Fort Peck, MT, USA. R. Barsbold and T. Tsogtbaatar provided access to specimens in Mongolia, and P. Holroyd (UCMP), M. Norell (AMNH), and B. Strilisky (TMP) also provided access to comparative material. C. Stoppa executed the illustrations in Fig. 1. D. Dufault executed the line art in Fig. 1. Thanks to K. Seymour for X-ray imaging and D. Scott for assistance with the SEM imaging. The Willi Hennig Society sponsored our edition of TNT. We also thank S. Brusatte, A. Turner, M. Loewen, and N. Longrich for discussions. This study was supported by Natural Sciences and Engineering Research Council of Canada Discovery Grants to DCE and PJC. Travel funds were provided by the Doris O. and Samuel P. Welles Research Fund to DWL. The specimen was acquired with help from the Louise Hawley Stone Trust administered by the Royal Ontario Museum. We thank S. Brusatte and two anonymous reviewers for comments that improved the manuscript.