The fossil record for neoceratopsian (horned) dinosaurs in the Lower Cretaceous of North America primarily comprises isolated teeth and postcrania of limited taxonomic resolution, hampering previous efforts to reconstruct the early evolution of this group in North America. An associated cranium and lower jaw from the Cloverly Formation (?middle–late Albian, between 104 and 109 million years old) of southern Montana is designated as the holotype for Aquilops americanus gen. et sp. nov. Aquilops americanus is distinguished by several autapomorphies, including a strongly hooked rostral bone with a midline boss and an elongate and sharply pointed antorbital fossa. The skull in the only known specimen is comparatively small, measuring 84 mm between the tips of the rostral and jugal. The taxon is interpreted as a basal neoceratopsian closely related to Early Cretaceous Asian taxa, such as Liaoceratops and Auroraceratops. Biogeographically, A. americanus probably originated via a dispersal from Asia into North America; the exact route of this dispersal is ambiguous, although a Beringian rather than European route seems more likely in light of the absence of ceratopsians in the Early Cretaceous of Europe. Other amniote clades show similar biogeographic patterns, supporting an intercontinental migratory event between Asia and North America during the late Early Cretaceous. The temporal and geographic distribution of Upper Cretaceous neoceratopsians (leptoceratopsids and ceratopsoids) suggests at least intermittent connections between North America and Asia through the early Late Cretaceous, likely followed by an interval of isolation and finally reconnection during the latest Cretaceous.
Citation: Farke AA, Maxwell WD, Cifelli RL, Wedel MJ (2014) A Ceratopsian Dinosaur from the Lower Cretaceous of Western North America, and the Biogeography of Neoceratopsia. PLoS ONE 9(12): e112055. https://doi.org/10.1371/journal.pone.0112055
Editor: Peter Wilf, Penn State University, United States of America
Received: May 23, 2014; Accepted: September 22, 2014; Published: December 10, 2014
Copyright: © 2014 Farke et al. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Data Availability: The authors confirm that all data underlying the findings are fully available without restriction. All relevant data are within the paper and its Supporting Information files.
Funding: Funding was received from the National Science Foundation (DEB 9401094, 9870173, http://www.nsf.gov); National Geographic Society (5918-97, http://www.nationalgeographic.com/); and American Chemical Society (PRF #38572-AC8, http://www.acs.org). The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
Competing interests: Andrew A. Farke has read the journal's policy and the authors of this manuscript have the following competing interests: Andrew A. Farke is a volunteer section editor and academic editor for PLOS ONE. This does not alter the authors' adherence to PLOS ONE policies on sharing data and materials.
Neoceratopsia constitute one of the most taxonomically diverse and morphologically disparate clades of ornithischian dinosaurs during the Cretaceous, and are particularly known for their highly derived cranial anatomy (Figure 1; –). Early representatives, such as Liaoceratops yanzigouensis Xu et al. 2002 , exhibit expanded bony platforms for the jaw muscles that are elaborated into a broad frill in later taxa (e.g., Protoceratops andrewsi Granger and Gregory 1923 ). Ceratopsids, including Triceratops Marsh 1891 , further augmented the skull with elongate horns and spikes .
Marginocephalia and Coronosauria are node-based clades, indicated by circles. The rest of the clades shown here are stem-based, indicated by half-brackets. See text for definitions and taxonomic authorities. Silhouettes are not to scale (Acrotholus by G. Monger, Psittacosaurus by J. Headden, Triceratops by R. Amos; all others by A. Farke; all images are CC-BY and provided via www.phylopic.org).
A series of well-preserved specimens from a variety of taxa has firmly placed the origin and early diversification of Neoceratopsia in the Early Cretaceous of Asia , –. However, the timing and phylogenetic affinities of the clade's first dispersal into North America have remained unclear, due to a dearth of fossilized material that, until now, has consisted of relatively uninformative isolated teeth and postcranial elements –. These specimens indicate the presence of Neoceratopsia in the Early Cretaceous of North America, but cannot be further identified. The arrival of neoceratopsians on the continent is important for understanding the nature and timing of broader faunal interchanges during the Cretaceous. Specimens representing other dinosaur clades that potentially dispersed between Asia and North America during the Early Cretaceous—including tyrannosauroids , oviraptorosaurs , ornithomimids , therizinosaurs , and shamosaurine ankylosaurs —are similarly fragmentary in most cases and have contributed to uncertainty regarding the mode and timing of faunal exchange. Additionally, there is uncertainty on whether the dispersal was directly between North America and Asia (trans-Beringia) or via Europe , , , although the absence of important Asian clades in Europe somewhat favors a trans-Beringian dispersal hypothesis , .
A skull from the Albian part of the Cloverly Formation of Montana (Figures 2, 3) represents the first neoceratopsian from the North American Early Cretaceous that is diagnostic to the species-level. This new taxon, Aquilops americanus gen et sp. nov., exhibits definitive neoceratopsian features and is closely related to coeval Asian species. Furthermore, the discovery of Aquilops adds unambiguous support for a late Early Cretaceous (∼113–105 Ma) intercontinental migratory event between Asia and North America , , as well as support for a complex set of migratory events for organisms between North America and Asia later in the Cretaceous.
Outcrops of Cloverly Formation within the Bighorn Basin of Montana and Wyoming are shown in orange; the actual width of the outcrop band has been exaggerated for visual clarity. The approximate location of OMNH locality V1057, within Carbon County, Montana, is indicated by the blue arrow.
Partial cranium in A) dorsal and B) right lateral views. Partial lower jaw in C) right lateral and D) dorsal views. This interpretive figure is based on surface scans of the original specimen, with sutures highlighted. The lower jaw is reversed, to facilitate placement with the skull. Abbreviations: aof, antorbital fossa; d, dentary; ecp, ectopterygoid; fr, frontal; itf, infratemporal fenestra; j, jugal; la, lacrimal; m, maxilla; na, nasal; pal, palpebral; pd, predentary; pf, prefrontal; pm, premaxilla; po, postorbital; r, rostral; spl, sutural surface for splenial; sq, squamosal; stf, supratemporal fenestra.
The specimen described herein, OMNH 34557, was collected in the Cloverly Formation of Carbon County, Montana (Figure 2). Previous workers have designated subdivisions of the Cloverly Formation either numerically (from lowest/oldest to highest/youngest, Units IV–VII; ), alphabetically (from lowest, intervals A–C ), or as named members (from lowest, the Pryor Conglomerate, Little Sheep Mudstone, and Himes members ). OMNH 34557 originated from a deep red sandy claystone in the basal part of unit VII of the Cloverly Formation, as defined by Ostrom , corresponding to the Himes Member of Moberly  and within the C interval of Meyers et al. (; redefined by ). Summarizing the partly conflicting evidence at hand (below), OMNH 34557 is likely Albian in age; we provisionally regard it as about 104–109 Ma old (middle to early late Albian).
The age of the Cloverly Formation is not well constrained; indeed, as whole, it is best regarded as simply Lower Cretaceous. There is general agreement that its basal part (Pryor Conglomerate Member of Moberly ; Unit IV of Ostrom ; A interval of Meyers et al. ) is pre-Aptian (we avoid usage of the term “Neocomian” following Sames et al. ) in age, and that it may correlate with the lower Lakota (L1 interval or Chilson Member) Formation, Wyoming and South Dakota, and the lower Cedar Mountain (Yellow Cat Member) Formation, Utah , both of which may be as old as Berriasian–Valanginian –, though radiometric dates suggest a younger (Barremian–early Aptian) age for the Yellow Cat Member (see Ludvigson et al. , and references therein).
Upper age constraints for the Cloverly Formation are also open to debate. The upper parts of the Cloverly (units V–VII of Ostrom ; Little Sheep Mudstone and Himes members of Moberly ; intervals B–C of Meyers et al. ) have long been regarded as being Aptian and/or Albian in age , with recent authors favoring the younger part of this range (e.g., ; but also see ). Stratigraphically, the Cloverly Formation is overlain by the Sykes Mountain Formation, a nearshore marine unit that lacks age-diagnostic fossils . The Sykes Mountain Formation grades upward into the Thermopolis Shale, which contains the widespread bivalve Inoceramus comancheanus Cragin 1894  and is considered to date to the middle part of the Late Albian . This implies an approximate upper age constraint of between 103–105 Ma for the Cloverly Formation. Zaleha (, p. 892) reviewed then-available (2006) geochronologic and biochronologic evidence, concluding (largely based on palynomorphs) that Cloverly B–C intervals are “Albian, but no younger than middle Albian” (i.e., about 108–113 Ma). Since then, radiometric dates have appeared in two abstracts but have not been formally published. Burton et al. (, p. 52) reported a radiometric determination of 108.5±0.2 Ma (i.e., middle Albian) “from Beaver Creek near Shell, Wyoming at about 75 m above the contact with the underlying Morrison Formation” (hence presumably within units V–VII of Ostrom ). D'Emic and Britt (; see also ) presented a younger date of 103.49±0.79–1.17 Ma (late Albian) based on detrital zircons from unit VI or basal unit VII.
Materials and Methods
AMNH, American Museum of Natural History, New York, New York, United States of America; CAGS-IG, Chinese Academy of Geological Sciences-Institute of Geology, Beijing, China; CMN, Canadian Museum of Nature, Ottawa, Ontario, Canada; IVPP, Institute of Vertebrate Paleontology and Paleoanthropology, Beijing, China; OMNH, Sam Noble Oklahoma Museum of Natural History, Norman, Oklahoma, United States of America.
Within this paper, we use the following explicit phylogenetic definitions implemented by previous authors (see also Figure 1). Neoceratopsia Sereno 1986  includes all ceratopsians closer to Triceratops than to Psittacosaurus Osborn 1923 , and Coronosauria Sereno 1986  is defined as the most recent common ancestor of Protoceratops and Triceratops as well as all of the descendants of this ancestor . Ceratopsoidea Hay 1902  includes Triceratops and all taxa closer to it than to Protoceratops . Leptoceratopsidae Nopsca 1923  includes Leptoceratops gracilis Brown 1914  and all species closer to it than to Triceratops horridus Marsh 1889 , . All ages presented herein follow the Geological Time Scale 2012 .
No permits were required for the described study, which complied with all relevant regulations.
The electronic edition of this article conforms to the requirements of the amended International Code of Zoological Nomenclature, and hence the new names contained herein are available under that Code from the electronic edition of this article. This published work and the nomenclatural acts it contains have been registered in ZooBank, the online registration system for the ICZN. The ZooBank LSIDs (Life Science Identifiers) can be resolved and the associated information viewed through any standard web browser by appending the LSID to the prefix “http://zoobank.org/”. The LSID for this publication is: urn:lsid:zoobank.org:pub:C835BEAB-3A4C-47B5-8B23-1A244086B3D7. The electronic edition of this work was published in a journal with an ISSN, and has been archived and is available from the following digital repositories: PubMed Central, LOCKSS.
Dinosauria Owen 1842 
Ornithischia Seeley 1887 
Ceratopsia Marsh 1890 
Neoceratopsia Sereno 1986 
Aquilops gen. nov.
Aquilops americanus sp. nov.
OMNH 34557, a partial skull, with associated predentary, partial left dentary, and additional associated but unidentifiable fragments (Figs. 3–9; three-dimensional digital scans are contained in Files S7–S12).
A) right lateral and B) left lateral views.
A) dorsal and B) ventral views. The rostral end of the skull is to the right side of the image.
Cranium in A) right lateral and B) dorsal views; C) life restoration in right lateral view. The rendering is based on OMNH 34557 (holotype), with missing details patterned after Liaoceratops yanzigouensis and Archaeoceratops oshimai. Life restoration by Brian Engh.
Predentary and left dentary in A) medial; B) ventral; C) dorsal; and D) lateral views. The three fragments, although unattached, were placed into articulation for these photos. The rostral direction is to the right in A and to the left in B–D.
A) Upper right dentition in buccal view. B) Lower left dentition in lingual view. C) Detail of maxillary tooth in buccal view, showing selected features. The photographs shown here were taken of casts coated with ammonium chloride, to even out color variations in the specimen. Abbreviations: ar, accessory ridge; cing, cingulum; pr, primary ridge.
Note that enamel (highlighted by arrows in the inset) only occurs on the labial surface, and does not continue to the buccal extremity of the tooth at the bottom of the image. The discoloration adhering to the tooth at the top of the image (lingual side of the tooth) is sediment impregnated with consolidant resin, not enamel.
The genus name is derived from the Latin aquila, meaning “eagle,” and the Greek ops, meaning “face,” referring to the hooked beak on the skull of the animal. The species name, meaning “American” (Latin), reflects the species' status as the earliest unequivocal neoceratopsian in North America.
Locality and horizon.
OMNH locality V1057, 45°N 109°W, Carbon County, Montana, United States of America (Figure 2); basal (?middle–late Albian) part of Unit VII (as defined by ), Cloverly Formation. Precise locality data are on file at OMNH and are available to qualified investigators upon request.
A small neoceratopsian with the following autapomorphies: rostral with a ventrally-directed tip and a strongly arched keel possessing a boss; oral margin rostral to maxillary tooth row entirely concave in lateral view; antorbital fenestra more than twice as long as tall and tapering to a sharp point below the orbit.
In addition to the autapomorphies listed above, Aquilops americanus is distinguished from non-neoceratopsians (e.g., Yinlong downsi Xu et al. 2006 , Chaoyangsaurus youngi Zhao et al. 1999 , Psittacosaurus mongoliensis Osborn 1923 ) by a sharply keeled rostral with a ventral process and sharp keel on rostral edge; exclusion of the postorbital from the margin of the infratemporal fenestra; maxillary teeth with a weak and wide median ridge; and tooth crowns that are ovate in lateral view with enamel restricted to one side, among other features. Aquilops americanus is distinguished from Liaoceratops yanzigouensis by the suborbital ramus of the jugal being as deep as the orbital ramus; a rounded edge (rather than a sharp ridge) to the predentary oral margin; and the presence of pronounced cingula on the maxillary teeth. Aquilops americanus is distinguished from Yamaceratops dorngobiensis Makovicky and Norell 2006  by the sharp keel on the rostral surface of the rostral and cheek teeth with cylindrical roots and pronounced cingula. Aquilops americanus is distinguished from Helioceratops brachygnathus Jin et al. 2009  by having a less steeply inclined ventral predentary facet and a more shallow dentary. Aquilops americanus is distinguished from Archaeoceratops spp. Dong and Azuma 1997  on the restriction of enamel to one side of the tooth, and differs from Archaeoceratops oshimai Dong and Azuma 1997  in the absence of a convexity along the oral margin at the premaxilla/maxilla suture in lateral view. Aquilops americanus is distinguished from Auroraceratops rugosus You et al. 2005  by a rounded beveled edge on the cutting surface of the predentary. Aquilops americanus is distinguished from Leptoceratopsidae by a postorbital with rounded dorsal part overhanging lateral edge of supratemporal fenestra; weak and wide median ridge on maxillary teeth (except for Cerasinops hodgskissi Chinnery and Horner 2007 ); and cheek teeth with cylindrical roots. Aquilops americanus is distinguished from Coronosauria in having a flat, thin, and unmodified dorsum of the frontal; three teeth in the premaxilla; and teeth with cylindrical roots and a weak rather than distinct primary ridge.
As preserved, the sagittal plane of the cranium in OMNH 34557 (Figures 3–5) is moderately skewed to the right, along with some moderate mediolateral compression and minor dorsoventral compression (primarily affecting the dorsal portion of the skull). The right side of the skull (Figures 3B, 4A) is better preserved and more complete than the left side (Figure 4B). Elements missing from the right side include the quadratojugal, and most of the quadrate and parietal. The left side preserves only the premaxilla, prefrontal, nasal, partial maxilla, partial frontal, and partial lacrimal. The midline rostral bone is also preserved. The rostral end of the nasals, the braincase, and most of the palatal bones are missing. As preserved, the skull measures 84.2 mm long from the tip of the rostral to the tip of the jugal. Basic measurements for the cranial bones and dentition are in Tables 1 and 2.
The dorsal margin of the preorbital portion of the skull was steeply inclined as compared to the postorbital portion (Figures 3B, 4A, 6A). Despite the crushing, the breadth of the frontals relative to the surrounding parts of the skull strongly suggests that the skull was triangular in dorsal view (Figures 3A, 5A, 6B), although perhaps not as strongly so as in larger species. This reduced prominence in triangularity may in part be due to the possible juvenile nature of the specimen, as seen in Protoceratops (; see below).
The rostral bone is unelaborated compared to the condition in ceratopsids (e.g., Triceratops horridus), with short dorsal and lateral (buccal) processes (Figure 3B). Part of the dorsal process on the right side is broken away, but the sutural surface with the premaxilla remains, allowing confident reconstruction of the original morphology in the holotype. The rostral (anterior) surface is gently keeled, more prominently than seen in Yamaceratops dorngobiensis (see ) but less than the condition in Protoceratops andrewsi or Leptoceratops gracilis. A midline boss or rugosity (5.3 mm long by 1.2 mm tall) caps the leading edge of the rostral, unique among ceratopsians. Furthermore, the lateral profile of the leading edge is strongly arched, contrasting with the gently arched leading edge seen in other ceratopsians (e.g., Psittacosaurus mongoliensis, AMNH 6254; Archaeoceratops oshimai, IVPP V11114; Protoceratops andrewsi, AMNH 6466). In lateral view, the ventral margin of the rostral is strongly hooked. As articulated with the premaxilla, the tip of the rostral forms a nearly 90 degree angle with the line of the maxillary teeth. This feature is autapomorphic for Aquilops. Although Archaeoceratops oshimai has been illustrated with a similar condition (: fig. 1), direct examination of the original specimen (IVPP V11114) shows that the degree of hooking was exaggerated in the published drawings. Thus, the morphology for the holotype of A. americanus is indeed unique. In other neoceratopsians, the angle of the rostral margin relative to the maxillary teeth is much broader (e.g., ∼150 degrees in Liaoceratops yanzigouensis, IVPP V12738).
The premaxilla is roughly triangular in lateral view. The ventral border is straight, with the caudal third supporting three teeth (described below). There is no evidence of a prominent bump or convexity in the oral margin at the interface between the maxilla and premaxilla, unlike the condition in many other ceratopsians (e.g., Chaoyangsaurus youngi, CAGS-IG V371; Archaeoceratops oshimai, IVPP V11114; Leptoceratops gracilis, CMN 8887, 8889; Protoceratops andrewsi, AMNH 6466; , ), and paralleling the morphology seen in Liaoceratops yanzigouensis (IVPP V12738). The articulated maxilla, premaxilla, and rostral together produce a unique concave profile to the oral margin rostral to the maxillary teeth (Figures 3B, 4, 6). A slight lateral swelling occurs immediately dorsal to the oral margin and is confluent with the rostral bone. The swelling terminates at its tallest point just rostral to the premaxillary-maxillary suture and tapers rostrally. As in A. oshimai and other non-coronosaurian neoceratopsians, the caudal margin of the premaxilla is vertical and forms part of the rostral border of the antorbital fossa. As preserved, the external naris is 9 mm long and is therefore quite small in proportion to the orbit and antorbital fossa (Table 1), as is typical for non-ceratopsids (e.g., Liaoceratops yanzigouensis; Protoceratops andrewsi; Yinlong downsi). The naris is oriented at a 60 degree angle to horizontal, and the preserved parts of the narial margin indicate an overall elliptical shape. The naris is confluent with a shallow depression on the rostral end of the premaxilla.
In ventral view (Figure 5B), the oral margin of the paired premaxillae is broadest at its midpoint, and is pinched cranially (to receive the rostral bone) and caudally (at the suture with the paired maxillae). Accounting for distortion, the alveoli for the premaxillary teeth would have been in line with, and not outside of, many of the maxillary teeth, as in other ceratopsians with premaxillary teeth. Although the teeth in the holotype of Archaeoceratops yujingziensis (CAGS-IG-VD-003) are described as falling outside the line of the maxillary teeth , this appears to be a function of specimen orientation rather than genuine morphology; thus the anatomy in A. yujingziensis probably matches that in other ceratopsians.
The maxilla is elongate and roughly trapezoidal in lateral view, bounded dorsally by the jugal and lacrimal (Figure 3B). A prominent buccal emargination characterizes the lateral surface of the maxilla; it slopes from its highest point caudally to the rostral end, where the emargination intersects with the oral margin at the maxillo-premaxillary suture. At least four neurovascular foramina mark the lateral surface of the emargination. As typical for non-ceratopsids, the postalveolar process is quite short, extending only 5.4 mm behind the last tooth position. In ventral view (Figure 5B), the alveolar margin is slightly sinuous, with a moderate lateral concavity at the cranial third of the element. The right and left maxillae are closest to each other at their rostral ends, with a moderate lateral deflection at the premaxillary suture, as seen in many other basal neoceratopsians (e.g., Archaeoceratops spp. [CAGS-IG-VD-003; IVPP V11114]; Auroraceratops rugosus, CAGS-IG-2004-VD-001; Liaoceratops yanzigouensis, IVPP V12738; ). The tooth row is also set slightly below the premaxilla in lateral view, and the maxillary teeth are oriented horizontally (Figures 3B, 6A). Although this is slightly accentuated by crushing, the profile does not seem to be due solely to taphonomic processes.
The rostral ends of the paired nasals are missing, and crushing has slightly obscured their original shape and sutural relationships (Figure 3A). The caudal ends of the nasals contact a narrow rostral process of the paired frontals, completely separating the prefrontals from each other. The dorsal surface of the nasals is relatively flat and untextured, as seen in many other non-coronosaurs (e.g., Archaeoceratops oshimai, IVPP V11114; Yinlong downsi, IVPP V14530). Although there is some crushing on the right nasal, the dorsal surface of the left nasal is quite flat, showing that OMNH 34557 lacks a fossa on the dorsal surface, as seen in Liaoceratops yanzigouensis (IVPP V12738; ).
The palpebral is broadest at its attachment with the prefrontal and lacrimal, tapering to a gently rounded profile at its free end (Figures 3A,B, 4A, 5A). The dorsal surface of the element is flat. A loose element, interpreted as the left palpebral, shows a triangular coronal cross-section. Overall, the element in OMNH 34557 is quite similar to the morphology seen in other basal neoceratopsians.
The cranio-caudally elongate and triangular antorbital fossa is bordered caudally by the jugal, ventrally by the maxilla, rostrally by the premaxilla (for a very small extent), and dorsally by the lacrimal (Figures 3B, 4A). The fossa is tallest at its rostral end, and the caudal portion of the fossa is deepest. Its ventral border is nearly horizontal. The dorsal edge of the fossa is sharply defined at its portion immediately ventral to the orbit, with a distinct bar of bone. This distinction becomes less defined in the rostral direction. Compared to other ceratopsians, the antorbital fossa in OMNH 34557 is elongate, with a height:length ratio of 0.43 (measurements in Table 1). This compares to 0.73 in Archaeoceratops oshimai, IVPP V11114 (20.8 by 28.5 mm on the better preserved right side), and 0.72 in Auroraceratops rugosus, CAGS-IG-2004-VD-001 (20.6 by 28.7 mm on the left side), and the relatively equidimensional, circular fossae in leptoceratopsids and protoceratopsids.
Ventral to the orbit, the lacrimal inserts into the jugal along a V-shaped contact (Figures 3B, 4A), a condition that is widespread across Neoceratopsia (e.g., Liaoceratops yanzigouensis, IVPP V12738; Archaeoceratops oshimai, IVPP V11114; Leptoceratops gracilis, CMN 8889; Protoceratops andrewsi, AMNH 6429, 6637). In most specimens belonging to Psittacosaurus, the contact is fairly linear (see figures in ).
The prefrontal is roughly hatchet-shaped and is narrowest caudally, where it forms the dorsal margin of the orbit, maintaining a relatively uniform width at the portion between the nasal and lacrimal, and expanding laterally at the rostral end to contact the premaxilla (Figure 3A). The dorsal surface of the caudal end of the prefrontal is flat, and it is gently concave at the rostral end.
The preserved parts of the paired frontals are roughly diamond-shaped in dorsal view (Figure 3A), with a very slight ridge running the length of the interfrontal suture. A similarly slight (<0.5 mm deep) depression parallels this ridge. Otherwise, the dorsal surface of the frontals is quite flat, typical of ceratopsians outside of Leptoceratopsidae and Coronosauria . A narrow prong inserts between the prefrontals and contacts the nasals rostrally (as seen in other basal ceratopsians, such as Psittacosaurus spp. and Liaoceratops yanzigouensis, IVPP V12738; ). Some very faint neurovascular grooves originate in the middle of the dorsal surface of each frontal, approximately at the midpoint of the orbit, and radiate rostrally for several millimeters. Caudally, the suture between the parietals and the frontals is not discernible, assuming that it is even preserved. Thus, it cannot be determined what proportion of the rostral end of the supratemporal fenestra was formed by the frontal and parietal, respectively. The broken edges of the frontal along the caudal end of the bone uniformly approximate only a millimeter in thickness.
The postorbital is longer than tall or wide (Figures 3A,B, 4A, 5A). The dorsal and lateral surfaces of this bone are nearly at right angles to each other, separated by a prominent ridge. The dorsal surface of the postorbital is flattened and forms part of the lateral margin of the supratemporal fenestra; the narrow bone tapers caudally along the fenestra's margin. The postorbital abuts the caudal margin of the frontal. The lateral surface of the bone is also flattened, with strong tapers to both the caudal and descending limbs. This latter portion of the postorbital is sandwiched between the orbit and the ascending limb of the jugal. Although there is some cracking in this area, a thin flange of the jugal does seem to separate the postorbital from the infratemporal fenestra (Figure 3B). Consequently, the postorbital is excluded from the infratemporal fenestra as in other non-coronosaurian neoceratopsians (e.g., Yamaceratops dorngobiensis, IGM 100/1303; Liaoceratops yanzigouensis, IVPP V12738).
The jugal is almost perfectly planar along most of its lateral surface, except for a prominent ridge and concavity immediately rostral to the infratemporal fenestra (Figures 3B, 4A). The horizontally-oriented rostral limb of the element is comparatively longer and narrower than the ascending, caudal limb. The “blade” of the jugal continues the horizontal orientation of the rostral limb, terminating in a point. No epijugal scar is visible on the specimen, and we thus infer that an ossified epijugal was lacking. This condition also occurs in non-neoceratopsians, Liaoceratops yanzigouensis (IVPP V12633, V12738; ), and juvenile coronosaurs (e.g., Protoceratops andrewsi and Triceratops sp. , ). The caudal margin of the jugal, marked by a prominent concavity, is oriented at approximately 45 degrees to the horizontal. Several lightly incised neurovascular impressions occur in this area, differing from the more heavily sculptured surface seen in most other ceratopsians (e.g., Yinlong downsi, IVPP V14530; Archaeoceratops oshimai, IVPP V11114; Liaoceratops yanzigouensis, IVPP V12738). This may be a juvenile feature of OMNH 34557, as for juvenile ceratopsids .
The exact profile of the orbit is obscured by crushing, although it appears that the orbit was originally longer than tall (Figures 3B, 4A, 6A). The rostral one-third of the ventral margin of the orbit is bound by the lacrimal; the jugal bounds the caudal two-thirds. The caudo-dorsal quadrant of the orbit is bounded by the postorbital, with the remaining dorsal quarter bounded by (from caudal to cranial) the frontal, prefrontal, and palpebral. Relative to the preorbital length, the orbit is comparatively large (32 mm vs. 38 mm, or a ratio of 0.84), indicative of juvenile ontogenetic status in many vertebrates, including ceratopsians , , . By comparison, the presumably adult holotype of Archaeoceratops oshimai (IVPP V11114) has a ratio of 0.63 (47 mm vs. 74 mm) and the holotype of Auroraceratops rugosus (CAGS-IG-2004-VD-001) has a ratio estimated at 0.64 (52 mm vs. 81 mm).
The right ectopterygoid, missing its caudal end, is visible in lateral, medial, and caudal views (Figures 3B, 4A,B). As is typical for non-ceratopsid neoceratopsians (e.g., Liaoceratops yanzigouensis; IVPP V12738; Protoceratops andrewsi, AMNH 6429), the element is wrapped around the caudal surface of the alveolar process of the maxilla.
A fragment of the shaft of the right quadrate is in place near the ventral margin of the right jugal. The cross-section of the ventral edge of the fragment is D-shaped, with the convex surface pointing caudally.
The preserved fragment of the right squamosal forms part of the supratemporal fenestra and is triangular, broadening caudally. The squamosal is situated dorsal and caudal to the postorbital (Figure 3A,B).
The predentary tapers to a sharp point both in dorsal and lateral views (Figures 3C,D, 7), contrasting with the blunt, U-shaped predentary in psittacosaurids and Liaoceratops yanzigouensis, but similar to the condition in all other neoceratopsians . In Archaeoceratops oshimai (IVPP V11114) and Auroraceratops rugosus (CAGS-IG-2004-VD-001), part of the oral margin is canted dorsolaterally , , contrasting with the condition in Liaoceratops yanzigouensis (IVPP 12738) and Yamaceratops dorngobiensis (IGM 100/1867). The latter morphology occurs in OMNH 34557, in which the oral margin is relatively indistinct and faces dorsally to dorsomedially. Thus, the dorsal cutting surface of the predentary smoothly grades into the medial, shallowly scooped aspect of the bone. This shallow scooping is most similar to the condition in basal ceratopsians (e.g., Archaeoceratops oshimai, Chaoyangsaurus youngi) and differs from the deeply invaginated dorsal midline of leptoceratopsids and coronosaurs . Like all of the aforementioned neoceratopsians, but unlike non-neoceratopsians, Aquilops americanus has a blunt rather than a sharp, ridge-like oral margin for the predentary. The ventral process of the predentary is much more prominent and broader than the dorsal process. The left and right lateral surfaces of the predentary meet ventrally along a rounded keel (as in most neoceratopsians), unlike the broadly oval cross-section seen in Liaoceratops yanzigouensis (IVPP V11114) and non-neoceratopsians.
Two fragments represent this element, with a clean contact fit between each other and the predentary (Figures 3C,D, 7). The lingual surface of the incomplete left dentary is convex caudally and concave rostrally, preserving the rostral portion of the Meckelian groove (Figure 7A). This groove is quite broad, as in other neoceratopsians. In dorsal view, the tooth row on the left dentary fragment is concave laterally (Figures 3C, 7C). The dorsal portion of the lateral surface of the dentary shows a strong buccal emargination, and thus the tooth row is greatly inset. The dentition extends rostrally up to the point of the suture for the predentary, so that the length of the diastema at the rostral end of the dentary was minimal (Figures 3D, 7C). This morphology is typical of basal ceratopsians but unlike that in ceratopsids .
Although the splenial itself is not preserved, its sutural surface with the dentary is discernible on that element. The very rostral end of the splenial contacted the ventral process of the predentary (Figures 3D, 7A,C). Thus, based on the configuration of the predentary, splenial, and dentary contacts, it can be inferred that the contact between the dentaries at the midline was minimal, as is typical of non-ceratopsids . The splenials also presumably contacted each other along the midline, although this cannot be verified.
The premaxilla contains three ventrally-directed teeth, with the rostral-most tooth substantially smaller than the caudal two teeth (Figures 3B, 4A, 8A). In buccal view, the premaxillary teeth are roughly teardrop-shaped, broadest at the crown base and gently tapering toward the tip (Figure 3A). The base of each premaxillary tooth crown is mesiodistally elongate, and a carina marks the rostral and caudal edges of each tooth's crown. The crown lacks denticles (unlike the teeth in Archaeoceratops yujingziensis, Liaoceratops yanzigouensis, and Yamaceratops dorngobiensis; ,,), and both buccal and lingual surfaces bear enamel. The first premaxillary tooth on the right side is slightly rounded at its apex, suggestive of a wear facet. No wear facet is evident on the second premaxillary tooth, although this may be a function of relatively recent eruption.
The right maxilla bears nine closely-packed teeth (Figure 8A), which are separated from the premaxillary teeth by a 6 mm diastema. The teeth in the middle tooth positions are largest. In buccal view, each tooth crown tapers towards its occlusal end. A broad, distally-placed primary ridge divides the buccal surface of each crown, with at least one less pronounced accessory ridge distally and one or two mesially (Figure 8C; as seen in Archaeoceratops oshimai, IVPP V1114, and Auroraceratops rugosus, CAGS-IG-2004-VD-001) but unlike the very weakly developed accessory ridges and centrally placed primary ridge of Liaoceratops yanzigouensis (IVPP V12633) and non-neoceratopsians . The primary ridge merges with a well-defined basal cingulum on each tooth (Figure 8C). At least three “waves” (Zahnreihen of Edmund ) of teeth, moving from back to front, are preserved. The unworn crowns bear at least three denticles on each of their occlusal edges, but none of the unworn teeth is sufficiently exposed to allow a more detailed description of this feature. On worn teeth, three to four “cusps” are present, representing the occlusal expression of the buccal ridges. Only the buccal surface of each tooth crown bears enamel (Figure 9; the derived condition within Neoceratopsia, exclusive of L. yanzigouensis), and the wear surface of the maxillary teeth is angled laterally from base to tip, deviating approximately 20–30 degrees from the vertical.
Twelve incomplete, closely-packed teeth and tooth roots are preserved in the left dentary fragments (Figures 3C,D, 7, 8B). Here, the major ridge on the lingual surface is mesially placed with respect to the center of each tooth crown and is much more weakly developed than in the maxillary teeth. Weaker subsidiary ridges on each side of the primary carina terminate in fine denticles (3–4 distally, 4–5 mesially) at the occlusal edge of each dentary tooth. Similarly, the cingula are much more poorly developed in the dentary teeth than in the maxillary teeth; this condition also occurs in Archaeoceratops oshimai (IVPP V11114). Only the lingual surfaces of the tooth crowns bear enamel. As noted for Yamaceratops dorngobiensis (see ), the mesial ridges are angled relative to the primary ridge whereas the distal ridges are subparallel to the primary ridge. Replacement teeth are visible at two points ventral to the functional teeth; there does not appear to be room for a second row of underlying replacement teeth. Although no complete, unworn tooth crowns are visible, the maxillary and dentary teeth generally seem proportionately broader mesiodistally than in Archaeoceratops spp. , . Based on the diameter of the preserved portions, the rostral-most dentary teeth were smaller than those more caudally placed.
The number of dentary teeth greatly exceeds the number of maxillary teeth (9 and at least 12, respectively) in OMNH 34557. This condition is shared with Liaoceratops yanzigouensis of all known ontogenetic stages (11 or 12 vs. 15 maxillary and dentary teeth in IVPP V12738, the largest described skull; 10 vs. 12 in CAGS-IG-VD-002, the smallest described skull). In other basal neoceratopsians, the maxillary and dentary tooth counts are more closely matched (e.g., 13 vs. 12, respectively, in Auroraceratops rugosus, CAGS-IG-2004-VD-001; 13 or 14 vs. 14 in Archaeoceratops oshimai, IVPP V11114; 17 vs. 16 teeth in Leptoceratops gracilis, CMN 8889).
Ontogenetic status of OMNH 34557.
The holotype for Aquilops americanus is approximately 60 percent the size of the presumed adult holotypes for Liaoceratops yanzigouensis (IVPP V12738) and Archaeoceratops oshimai (IVPP V11114), but about the same size as a referred juvenile specimen of L. yanzigouensis (IVPP V12633), judging by preorbital length. Additionally, the orbit in OMNH 34557 is comparatively large relative to the rest of the skull, and the cranial bones are minimally textured. Based on reconstructed ontogenetic series for other ceratopsians, this suggests that OMNH 34557 is not fully grown. However, we also note that some features of the Aquilops americanus holotype are found in many adult basal neoceratopsians but not necessarily juveniles (e.g., well-defined accessory ridges on maxillary teeth; elongated contact between jugal and lacrimal). Thus, the mix of potential juvenile and adult features as well as intermediate size suggests identification of OMNH 34557 as “subadult” (comparable to assessments made for Protoceratops ). We hypothesize that the autapomorphies seen in the specimen as well as the unique combination of other features (even relative to known juveniles) support naming and diagnosis of Aquilops americanus.
In order to assess its position within Ceratopsia, Aquilops was scored using previously published matrices, with additions and revisions as noted below. Characters 1–133 were taken from Makovicky and Norell , characters 134–147 were taken from Makovicky , characters 148 and 149 were taken from characters 135 and 136 of Lee et al. , and characters 150 and 151 are from Ryan et al. . In order to help resolve Ceratopsidae, characters 152 (circumnarial depression, if deep, simple or complex) and 153 (narial spine absent or present) were added. Following Ryan et al. , character 140 was replaced with the construction and codings of character 134 from Lee et al. . A full list of characters is provided in Text S1 (File S1). Ajkaceratops kozmai Ősi et al. 2010 , Helioceratops brachygnathus, Auroraceratops rugosus You et al. 2005 , Chasmosaurus belli Lambe 1902 , Diabloceratops eatoni Kirkland and DeBlieux 2010 , Gryphoceratops morrisoni Ryan et al. 2012 , Koreaceratops hwaseongensis Lee et al. 2011 , Turanoceratops tardabilis Nessov and Kaznshkina in Nessov et a. 1989 , Unescoceratops koppelhusae Ryan et al. 2012 , and Zhuchengceratops inexpectus Xu et al. 2010  were also added to the matrix. The codings for character 138 for Centrosaurus apertus Lambe 1904  and Triceratops horridus were changed from 1 to 0. Character 6 was recoded for all taxa, to correct a misprint in the original matrix. The codings for characters 17, 52, 71, and 96 were changed from? to 1, and the coding for character 51 was changed to 1 for Zuniceratops christopheri , following examination of the original material.
The matrix (Text S2 in File S1, File S2) was run in TNT 1.1  using the tree bisection reconnection algorithm, with 10,000 replicates, up to 10,000 trees saved per replication, and branches with a minimum length of 0 collapsed. Hypsilophodon foxii Huxley 1869  was set as the outgroup relative to all other taxa, and all characters were unordered. Bremer support values were also calculated, along with bootstrap support (using sampling with replacement and 10,000 replicates).
Two equally most parsimonious trees (length = 288) were recovered, with Aquilops americanus firmly within Neoceratopsia (Figure 10; Figures S1 and S2 in File S1). Ajkaceratops kozmai, the only taxon to vary in position, was recovered as either a basal neoceratopsian (Figure S1 in File S1) or sister to Bagaceratops rozhdestvenskyi (Figure S2 in File S2). Due to the extensive ghost lineages required by the former option, we consider the latter most likely. This uncertainty appears to be due to character conflict as well as the incomplete nature of Ajkaceratops. The uncertain position of Ajkaceratops resulted in rather low bootstrap values, with most nodes in Neoceratopsia scoring less than 50 percent (Figure S3 in File S1). It took two additional steps to force Aquilops as a leptoceratopsid, but it took only one additional step to force Aquilops as sister to leptoceratopsids + coronosaurs or sister to coronosaurs alone. Five additional steps were required to force Aquilops as a ceratopsoid. The overall structure of the tree is, unsurprisingly, similar to that produced using other recent versions of the matrix , .
Some terminal taxa have been combined for space considerations, and the range bars for each taxon indicate uncertainty rather than known geological ranges. Continent icons indicate the ancestral areas reconstructed by DEC modeling. Silhouettes depict representative members of major clades and grades (Psittacosaurus by J. Headden, Zuniceratops by N. Tamura and modified by T. M. Keesey; Triceratops by R. Amos; all others by A. Farke; all images are CC-BY and provided via www.phylopic.org). Full results are presented in File S1.
Multiple synapomorphies identify Aquilops as a neoceratopsian. In the tree that recovers Ajkaceratops as a basal neoceratopsian, these include: a rostral with a ventral process and a keeled face; enlarged antorbital fossa; postorbital excluded from margin of laterotemporal fenestra; wide Meckelian groove separating dentigerous portion of mandible from external surface; predentary buccal margin rounded; maxillary teeth with median ridge; and ovate tooth crowns with enamel restricted to one side. In the tree that recovers Ajkaceratops as sister to Bagaceratops, an additional two synapomorphies were recognized as placing Aquilops within Neoceratopsia, including a convex buccal process rostral to the maxillary tooth row as well as pronounced cingula on cheek teeth.
If OMNH 34557 represents a subadult individual, this could influence the phylogenetic position recovered for Aquilops. Juvenile individuals often show primitive character states relative to adults of a species, which may pull operational taxonomic units coded from juvenile material toward the root relative to their actual position or destabilize the tree in other ways (e.g., –). We surveyed ceratopsians for which at least partial growth series were available (including Psittacosaurus, Liaoceratops, Protoceratops, Bagaceratops, Centrosaurus, and Triceratops) in order to evaluate which characters coded for Aquilops potentially were ontogenetically labile. The primary relevant, ontogenetically variable characters for the holotype of Aquilops were 29 (presence of epijugal) and 98 (presence and morphology of primary ridge on teeth) , . Coding these characters as unknown for Aquilops (along with character 30, position of epijugal, which was coded as inapplicable in the original matrix; Text S3 in File S1; File S3) resulted in 18 equally most parsimonious trees of 288 steps each (Figures S4 and S5 in File S1). In all of these trees, Aquilops was still recovered as a neoceratopsian outside of the clade Leptoceratopsidae + Coronosauria, although its relationship relative to other “basal” neoceratopsians varied. The basal position of Aquilops within Neoceratopsia is also congruent with its geological age. Full results are presented in File S1.
In order to evaluate the biogeographic patterns of North American ceratopsians relative to those from other areas, the Dispersal-Extinction-Cladogenesis (DEC) model ,  was used to reconstruct ancestral ranges within Ceratopsia. The DEC model is a likelihood model incorporating temporal data , and is increasingly used in studies of dinosaur biogeography (e.g., , ). Taxa were assigned to one of three biogeographic areas: North America, Asia, and Europe. Although the position of Ajkaceratops was variable in the phylogenetic analysis, we used the topology that placed it as sister to Bagaceratops rather than the topology that placed it as sister to Liaoceratops (Figure S6 in File S1); the former scenario was considered most likely because it minimized the lengths of ghost lineages and also is more congruent with the geological age of Ajkaceratops. Taxa outside of Ceratopsia (Hypsilophodon and Stegoceras) were trimmed from the analysis for two reasons: 1) because the outgroups used in the phylogenetic analyses represent incomplete samples of their respective group diversity and geographic/temporal range; and 2) because Neoceratopsia are the focus of the analysis. The temporal range of each taxon was estimated from the literature, and the midpoint of this range, rounded to the nearest 0.1 Ma, was used as the datum for each species unless more precise data were available (details in Table S1 in File S1; File S4). A time-calibrated phylogeny was produced using the timePaleoPhy function of paleotree 2.0, as implemented within R . Here, the branches were time-scaled using the “aba” method, so that 0.1 Ma were added to all zero-length branches; this effectively minimized ghost lineage lengths (a “strict” time calibration; File S5). A second data set was also produced using the “equal” option within timePaleoPhy, so that the lengths of ghost lineages were more equally distributed and thus longer than the “aba” method (a “smoothed” time calibration; File S6). Analyses were run using Lagrange 2.0, release 20130526 (available from https://github.com/rhr/lagrange-python).
The DEC model reconstructs an Asian origin for nearly all major ceratopsian clades (Figure 10; Table S2 in File S1 and Figure S5 in File S1). The ancestor of Aquilops dispersed from Asia into North America. The common ancestor of ceratopsoids is reconstructed as spanning North America and Asia, as for the ancestry of the clade including Turanoceratops + Ceratopsidae. Although the common ancestor of leptoceratopsids was most likely an Asian form, the ancestor of most groups within Leptoceratopsidae spanned Asia and North America together. All together, the DEC model as applied here reconstructs three different dispersals into North America from Asia—for Aquilops (and perhaps other, more poorly documented occurrences of Neoceratopsia near the Early–Late Cretaceous boundary of North America, noted below), most leptoceratopsids, and ceratopsoids. Full results are included in Table S2 in File S1 and Figure S5 in File S1.
Aquilops provides the first phylogenetically diagnostic evidence of a neoceratopsian dinosaur from the Early Cretaceous of North America. Other Early–early Late Cretaceous probable neoceratopsians from North America are represented by teeth from the Arundel Clay of Maryland , believed to be early Albian (∼113–110 Ma; ); teeth from the uppermost Cedar Mountain Formation of Utah , of earliest Cenomanian age (∼98 Ma; ) and a partial postcranial skeleton from the middle to upper Albian Wayan Formation of Idaho (∼110–101 Ma; ). Though these fossils suggest the geographically widespread presence of Neoceratopsia in North America during the late Early Cretaceous of North America, their affinities within the group are problematic. The new find from the Cloverly Formation, in concert with taxa known from elsewhere, permits a much more detailed reconstruction of biogeographic scenarios during the Cretaceous.
Somewhat surprisingly, Aquilops is not closely related to later ceratopsians from North America: instead, phylogenetic and biogeographic data suggest a complex history of interchanges between North America and Asia (Figure 10). Given the age of Aquilops and its closest relatives, the ancestor of Aquilops dispersed into North America by the late Albian (presumably sometime before 104 Ma) and possibly as early as the Aptian (∼125 Ma, based on the date for the beds containing Liaoceratops; ). The ancestors of other North American ceratopsians may have entered North America at this time (implying extensive ghost lineages) or, more probably, later (discussed below). Although the DEC model cannot completely exclude alternative possibilities, we posit that the bulk of the evidence supports at least two dispersals in addition to that for Aquilops: one for leptoceratopsids and one for ceratopsoids. This is consistent with many previous models of ceratopsian biogeography , , , , , but adds some noteworthy details.
The timing of the dispersal of leptoceratopsids into North America is uncertain. The basalmost leptoceratopsid, Asiaceratops, is early Cenomanian in age (96.2–100.5 Ma), suggesting that the ancestor of North American leptoceratopsids was in North America sometime after that interval. The oldest leptoceratopsid known from North America, Gryphoceratops, is late Santonian in age (83.6–84 Ma). Given the poor record of early Late Cretaceous vertebrates in North America and the phylogenetic position of Gryphoceratops (deeply nested within the family; see ; see also Figure 10 herein), a pre-Santonian arrival in North America appears likely for Leptoceratopsidae. Interestingly, DEC reconstructs the ancestor of most leptoceratopsid clades as spanning both Asia and North America. This result should be regarded as extremely tentative, in part due to the lability of the reconstruction of leptoceratopsid relationships. Additionally, the ancestral ranges for nodes within leptoceratopsids seem unlikely based on the apparent separation of North America and Asia from the Turonian until the late Campanian or early Maastrichtian, as suggested by coastline reconstructions and biogeographic evidence (e.g., , ). Resolution of this issue requires additional analysis and basic data improvement, in the form of new fossils from the early Late Cretaceous.
Although relationships among leptoceratopsids are subject to debate, it is worth noting that two Asian taxa (Udanoceratops and Zhuchengceratops) are nested deeply within the group, suggesting the possibility of one or two dispersals from North America into Asia. The age of Zhuchengceratops is imprecisely known, beyond the observation that it comes from an Upper Cretaceous horizon . Udanoceratops, from the Djadokhta Formation, is probably Campanian in age , perhaps 71–75 Ma , suggesting that dispersal (if any) occurred before or during that time interval. Branch support within leptoceratopsids is quite low, however, leaving open the possibility that Asian leptoceratopsids form an exclusive clade.
A biogeographic connection between North America and Asia before or during the Turonian is suggested by the occurrence of Turonian-aged Ceratopsoidea on both landmasses: Zuniceratops in New Mexico  and Turanoceratops in Uzbekistan . Thus, DEC reconstructs the clade's common ancestor as spanning Asia and North America. This connection is in concordance with inferred low sea levels ,  and paleogeographic reconstructions indicating subaerial causeways between the continents at these times . Ceratopsidae are reconstructed by DEC as originating in North America, which is consistent with fossil evidence.
In sum, the phylogenetic and biogeographic evidence suggests that the immigration event(s) for Aquilops and the ancestors of other North American clades—including ceratopsoids and leptoceratopsids—began sometime within the interval spanning the Barremian and Albian (as indicated by the probable Barremian or early Aptian age of Liaoceratops , , a close relative of Aquilops, as well as neoceratopsian teeth from the early Albian Arundel Clay of Maryland , . In the case of ceratopsoids, the dating of Zuniceratops as Turonian indicates that this clade entered North America no later than about 90 Ma. Ceratopsoids and leptoceratopsids may have entered North America simultaneously or separately, an issue that cannot be resolved with the patchy mid-Cretaceous fossil record.
Within the broader context of Cretaceous North American dinosaur biogeography, this new analysis is consistent with several previously proposed hypotheses. First, the occurrence of Aquilops in the late Albian of North America is congruent with the evidence from other clades for faunal interchange between North America and Asia in the latter part of the Early Cretaceous , , . Second, the temporal and geographic distribution of other North American neoceratopsians (leptoceratopsids and ceratopsoids) suggests at least intermittent connections between North America and Asia up to and including the Turonian, likely followed by an interval of isolation and then reconnection, potentially during the late Campanian. This, too, is consistent with records for therizinosaurs , hadrosauroids , , tyrannosauroids , and other clades, as well as with many paleogeographic reconstructions .
Given the probable origin of Neoceratopsia in Asia and their appearance in North America by the late Early Cretaceous, how did the group disperse from one landmass to the other—directly through Beringia, or via a trans-European route? As has long been recognized, Late Jurassic and earliest Cretaceous vertebrate assemblages of the western USA share points of similarity with earliest Cretaceous faunas of Europe , , –. Chinnery-Allgeier and Kirkland  suggested that trans-European dispersal explains the presence of neoceratopsian teeth in the Arundel Clay of Maryland. This scenario is uncertain due to the imprecisely known age of the Arundel (though no data known to us place it anywhere near the Barremian–Aptian boundary, as suggested by Chinnery-Allgeier and Kirkland ; an early Albian age is far more likely ) and the incomplete nature of the fossils. The Arundel specimens are almost certainly neoceratopsian, based on their morphology, but placement beyond this (particularly relative to Aquilops) is uncertain. A European dispersal route from Asia is possible biogeographically, particularly in light of European occurrences of gobiconodontid and spalacolestine mammals during the Early Cretaceous , –. However, these groups (and others not known from Europe) also occur in Asia (see  and references therein; –), and Asia is also the likely source for lizard groups appearing by the Albian–Cenomanian boundary in North America . Many other North American taxa (including dinosaurs) from the Aptian–Albian are not particularly closely allied with European species yet have plausible close Asian relatives (see summary in ). Addressing the point directly in the present context, no ceratopsians earlier than Santonian are known from Europe, and thus Europe doesn't figure in biogeographic reconstructions by the DEC model. Most critically, a denser sampling of Aptian and Albian vertebrates from Europe is necessary to establish the presence of ceratopsians there. This also does not exclude the possibility of occasional dispersals from Asia to North America via Europe as well as across Beringia. A trans-European model could be correct, but the biogeographic evidence at present is not particularly strong, as noted by Zanno and Makovicky , . Furthermore, paleogeographic reconstructions do not strongly indicate a direct subaerial connection between western North America and Europe after the Barremian (e.g., , , but see  for a contrasting interpretation; summary in ). These reconstructions are hampered, in turn, by spotty preservation of appropriately aged rocks.
We believe that the evidence in hand shows the origin of Neoceratopsia in Asia and dispersal to North America by the Albian, but cannot conclusively speak to European vs. Beringian models. At present we favor the latter, largely on the admittedly negative evidence for earlier (Barremian or Aptian) presence of Neoceratopsia in North America and the absence of phylogenetically relevant materials from the European record. In order to better elucidate the timing and mode of these events, additional fieldwork in appropriate stratigraphic intervals for Asia, Europe and North America (especially in the Aptian–Santonian) will be most helpful.
Supporting Information. Including the following: 1) Text S1, character list used for phylogenetic analysis; 2) Text S2, codings for phylogenetic analysis, in TNT format; 3) Text S3, coding for phylogenetic analysis, in TNT format, with ontogeny-dependent characters of Aquilops scored as unknown; 4) Figure S1, most parsimonious tree, recovering Ajkaceratops as most basal neoceratopsian, with synapomorphy list; 5) Figure S2, most parsimonious tree, recovering Ajkaceratops as sister to Bagaceratops, with synapomorphy list; 6) Figure S3, bootstrap and Bremer support values for phylogenetic hypothesis; 7) Figure S4, strict consensus of 18 equally parsimonious trees, with selected ontogenetic-variant characters for Aquilops americanus scored assuming that the holotype is a juvenile; 8) Figure S5, strict consensus of 18 equally parsimonious trees, with selected ontogenetic-variant characters for Aquilops americanus scored assuming that the holotype is a juvenile (Ajkaceratops excluded). 9) Figure S6, phylogenetic tree recovering Ajkaceratops as sister to Bagaceratops, used for DEC modeling, with nodal numbers referenced in results (Table S1); 10) Table S1, temporal calibrations used for DEC analysis; 11) Figure S7, Summary figure of DEC modeling results; 12) Table S2, node-by-node results for DEC analysis; 13) Literature cited.
Character matrix for phylogenetic analysis of Ceratopsia in TNT format.
Character matrix for phylogenetic analysis of Ceratopsia in TNT format, with ontogeny-dependent characters of Aquilops scored as unknown.
Files used for running biogeographic analysis of Ceratopsia. Including the following: script.txt, R script for creating time calibrated trees; ceratopsian.nex, phylogenetic tree to which time calibration was applied; ceratopsian_ages.txt, ages (in Ma) for taxa; ranges.txt, geographic ranges for use in Lagrange.
Python script for biogeographic analysis in Lagrange 2.0, for phylogeny time-scaled using “aba” option.
Python script for biogeographic analysis in Lagrange 2.0, for phylogeny time-scaled using “equal” option.
Surface scans of elements from the lower jaw of Aquilops americanus, OMNH 34557 (holotype).
Surface scan of the cranium of Aquilops americanus, OMNH 34557 (holotype).
Color surface scan of the cranium of Aquilops americanus, OMNH 34557 (holotype).
3D PDF of the lower jaw of Aquilops americanus, OMNH 34557 (holotype), based on scans from File S7.
Color 3D PDF of the cranium of Aquilops americanus, OMNH 34557 (holotype), based on scans from File S9.
It is a pleasure to offer our most heartfelt thanks to Scott K. Madsen, who found OMNH 34557 and prepared it with consummate skill. We are grateful to James Taylor, Jack Owen, the Keebler family, and the Montana Bureau of Land Management for access to outcrops of the Cloverly Formation. We thank Xu Xing (IVPP) and Hai-Lu You (formerly CAGS-IG) for facilitating access to specimens, Mark Loewen, Joseph Frederickson, Darren Naish, and Leonardo Maiorino for productive discussion and comments, and Roger Burkhalter for assistance in photography. Gary Wisser, from the scientific visualization center at Western University of Health Sciences, is gratefully acknowledged for the high resolution scan of the cranium. Reviews by Peter Makovicky, Hai-Lu You, and editor Peter Wilf improved the manuscript.
Conceived and designed the experiments: AAF WDM RLC. Performed the experiments: AAF WDM RLC. Analyzed the data: AAF WDM RLC MJW. Contributed reagents/materials/analysis tools: AAF WDM RLC MJW. Wrote the paper: AAF WDM RLC MJW.
- 1. Xu X, Makovicky PJ, Wang X-L, Norell MA, You H (2002) A ceratopsian dinosaur from China and the early evolution of Ceratopsia. Nature 416: 314–317
- 2. Dodson P, Forster CA, Sampson SD (2004) Ceratopsidae. In: Weishampel DB, Dodson P, Osmólska H, editors. The Dinosauria. Berkeley: University of California Press. pp. 494–513.
- 3. Makovicky P, Norell MA (2006) Yamaceratops dorngobiensis, a new primitive ceratopsian (Dinosauria: Ornithischia) from the Cretaceous of Mongolia. American Museum Novitates 3530: 1–42
- 4. Granger W, Gregory WK (1923) Protoceratops andrewsi, a pre-ceratopsian dinosaur from Mongolia. American Museum Novitates 72: 1–9.
- 5. Marsh OC (1889) Notice of gigantic horned Dinosauria from the Cretaceous. The American Journal of Science 38: 173–175.
- 6. Xu X, Forster CA, Clark JM, Mo J (2006) A basal ceratopsian with transitional features from the Late Jurassic of northwestern China. Proceedings of the Royal Society B: Biological Sciences 273: 2135–2140
- 7. You H, Dodson P (2003) Redescription of neoceratopsian dinosaur Archaeoceratops and early evolution of Neoceratopsia. Acta Palaeontologica Polonica 48: 261–272.
- 8. You H, Li D, Ji Q, Lammana MC, Dodson P (2005) On a new genus of basal neoceratopsian dinosaur from the Early Cretaceous of Gansu Province, China. Acta Geologica Sinica 79: 593–597.
- 9. Cifelli RL, Kirkland JI, Weil A, Deino AL, Kowallis BJ (1997) High-precision 40Ar/39Ar geochronology and the advent of North America's Late Cretaceous terrestrial fauna. Proceedings of the National Academy of Sciences 94: 11163–11167.
- 10. Chinnery BJ, Lipka TR, Kirkland JI, Parrish JM, Brett-Surman M (1998) Neoceratopsian teeth from the Lower to middle Cretaceous of North America. New Mexico Museum of Natural History and Science Bulletin 14: 297–302.
- 11. Weishampel DB, Meers MB, Akerston WA, McCrady AD (2002) New Early Cretaceous dinosaur remains, including possible ceratopsians, from the Wayan Formation of eastern Idaho. Idaho Museum of Natural History Occasional Paper 37: 5–17.
- 12. Chinnery-Allgeier BJ, Kirkland JI (2010) An update on the paleobiogeography of ceratopsian dinosaurs. In: Ryan MJ, Chinnery-Allgeier BJ, Eberth DA, editors. New Perspectives on Horned Dinosaurs: The Royal Tyrrell Museum Ceratopsian Symposium. Bloomington, Indiana: Indiana University Press. pp. 387–404.
- 13. Zanno LE, Makovicky PJ (2011) On the earliest record of Cretaceous tyrannosauroids in western North America: implications for an Early Cretaceous Laurasian interchange event. Historical Biology 23: 317–325
- 14. Makovicky PJ, Sues H-D (1998) Anatomy and phylogenetic relationships of the theropod dinosaur Microvenator celer from the Lower Cretaceous of Montana. American Museum Novitates 3240: 1–27.
- 15. Longrich N (2008) A new, large ornithomimid from the Cretaceous Dinosaur Park Formation of Alberta, Canada: implications for the study of dissociated dinosaur remains. Palaeontology 51: 983–997
- 16. Kirkland JI, Zanno LE, Sampson SD, Clark JM, DeBlieux DD (2005) A primitive therizinosauroid dinosaur from the Early Cretaceous of Utah. Nature 435: 84–87
- 17. Carpenter K, Kirkland JI (1998) Review of Lower and middle Cretaceous ankylosaurs from North America. New Mexico Museum of Natural History and Science Bulletin 14: 249–270.
- 18. Zanno LE, Makovicky PJ (2013) Neovenatorid theropods are apex predators in the Late Cretaceous of North America. Nature Communications 4: 2827
- 19. Russell DA (1995) China and the lost worlds of the dinosaurian era. Historical Biology 10: 3–12
- 20. Ostrom JH (1970) Stratigraphy and paleontology of the Cloverly Formation (Lower Cretaceous) of the Bighorn Basin area, Montana and Wyoming. Peabody Museum of Natural History Bulletin 35: 1–234.
- 21. Meyers JH, Suttner LJ, Furer LC, May MT, Soreghan MJ (1992) lntrabasinal tectonic control on fluvial sandstone bodies in the Cloverly Formation (Early Cretaceous), west-central Wyoming, USA. Basin Research 4: 315–333
- 22. Moberly R (1960) Morrison, Cloverly, and Sykes Mountain formations, northern Bighorn Basin, Wyoming and Montana. Geological Society of America Bulletin 71: 1137–1176
- 23. Zaleha MJ (2006) Sevier orogenesis and nonmarine basin filling: Implications of new stratigraphic correlations of Lower Cretaceous strata throughout Wyoming, USA. Geological Society of America Bulletin 118: 886–896
- 24. Sames B, Cifelli RL, Schudack ME (2010) The nonmarine Lower Cretaceous of the North American Western Interior foreland basin: new biostratigraphic results from ostracod correlations and early mammals, and their implications for paleontology and geology of the basin—an overview. Earth-Science Reviews 101: 207–224
- 25. Sames B (2011) Early Cretaceous Theriosynoecum Branson 1936 in North America and Europe. Micropaleontology 57: 291–344.
- 26. Sames B (2011) Early Cretaceous Cypridea Bosquet 1852 in North America and Europe. Micropaleontology 57: 345–431.
- 27. Cifelli R, Davis BM, Sames B (2014) Earliest Cretaceous mammals from the western United States. Acta Palaeontologica Polonica 59: 31–52
- 28. Ludvigson GA, Joeckel RM, Gonzalez LA, Gulbranson EL, Rasbury ET, et al. (2010) Correlation of Aptian-Albian carbon isotope excursions in continental strata of the Cretaceous foreland basin, eastern Utah, U.S.A. Journal of Sedimentary Research 80: 955–974
- 29. Jacobs LL, Winkler DA, Murry PA (1991) On the age and correlation of Trinity mammals, Early Cretaceous of Texas, USA. Newsletters on Stratigraphy 24: 35–43.
- 30. Oreska MPJ, Carrano MT, Dzikiewicz KM (2013) Vertebrate paleontology of the Cloverly Formation (Lower Cretaceous), I: faunal composition, biogeographic relationships, and sampling. Journal of Vertebrate Paleontology 33: 264–292
- 31. Cragin FW (1894) Descriptions of invertebrate fossils from the Comanche Series in Texas, Kansas and Indian Territory. Colorado College Studies 5: 49–68.
- 32. Kauffman EG (1984) Paleobiogeography and evolutionary response dynamic in the Cretaceous Western Interior Seaway of North America. In: Westermann GE, editor. Jurassic-Cretaceous Biochronology and Paleogeography of North America. Waterloo: Geological Association of Canada, Special Paper 27. pp. 273–306.
- 33. Burton D, Greenhalgh BW, Britt BB, Kowallis BJ, Elliott WS Sr, et al. (2006) New radiometric ages from the Cedar Mountain Formation, Utah and the Cloverly Formation, Wyoming: implications for contained dinosaur faunas. Geological Society of America Abstracts with Programs 38: 52.
- 34. D'Emic MD, Britt B (2008) Reevaluation of the phylogenetic affinities and age of a basal titanosauriform (Sauropoda: Dinosauria) from the Early Cretaceous Cloverly Formation of North America. Journal of Vertebrate Paleontology (Supplement to Number 3) 28: 68A.
- 35. D'Emic MD, Foreman BZ (2012) The beginning of the sauropod dinosaur hiatus in North America: insights from the Lower Cretaceous Cloverly Formation of Wyoming. Journal of Vertebrate Paleontology 32: 883–902
- 36. Sereno PC (1986) Phylogeny of the bird-hipped dinosaurs (Order Ornithischia). National Geographic Research 2: 234–256.
- 37. Osborn HF (1923) Two Lower Cretaceous dinosaurs of Mongolia. American Museum Novitates 95: 1–10.
- 38. Sereno PC (1998) A rationale for phylogenetic definitions, with application to the higher-level taxonomy of Dinosauria. Neues Jahrbuch für Geologie und Paläontologie 210: 41–83.
- 39. Hay OP (1902) Bibliography and catalogue of the fossil Vertebrata of North America. Bulletin of the United States Geological Survey 179: 1–868.
- 40. Nopsca BF (1923) Die Familien der Reptilien. Forschritte der Geologie und Palaeontologie Verlag von Gebrüder Borntraeger, Berlin 2: 1–210.
- 41. Brown B (1914) Leptoceratops, a new genus of Ceratopsia from the Edmonton Cretaceous of Alberta. Bulletin of the American Museum of Natural History 33: 567–580.
- 42. Makovicky PJ (2001) A Montanoceratops cerorhynchus (Dinosauria: Ceratopsia) braincase from the Horseshoe Canyon Formation of Alberta. In: Tanke DH, Carpenter K, editors. Mesozoic Vertebrate Life. Bloomington, Indiana: Indiana University Press. pp. 243–262.
- 43. Gradstein FM, Ogg JG, Schmitz M, Ogg G, editors (2012) The Geologic Time Scale 2012. 1st ed. Boston: Elsevier. 1176 p.
- 44. Owen R (1842) Report on British fossil reptiles. Part II. Report of the Eleventh Meeting of the British Association for the Advancement of Science 1841: 66–204.
- 45. Seeley HG (1887) On the classification of the fossil animals commonly named Dinosauria. Proceedings of the Royal Society of London 43: 165–171
- 46. Marsh OC (1890) Description of new dinosaurian reptiles. The American Journal of Science 39: 81–86.
- 47. Zhao X, Cheng Z, Xu X (1999) The earliest ceratopsian from the Tuchengzi Formation of Liaoning, China. Journal of Vertebrate Paleontology 19: 681–691
- 48. Jin L, Chen J, Zan S, Godefroit P (2009) A new basal neoceratopsian dinosaur from the Middle Cretaceous of Jilin Province, China. Acta Geologica Sinica - English Edition 83: 200–206
- 49. Dong Z, Azuma Y (1997) On a primitive neoceratopsian from the Early Cretaceous of China. Sino-Japanese Silk Road Dinosaur Expedition. Beijing: China Ocean Press. pp. 68–89.
- 50. Chinnery BJ, Horner JR (2007) A new neoceratopsian dinosaur linking North American and Asian taxa. Journal of Vertebrate Paleontology 27: 625–641
- 51. Brown B, Schlaikjer EM (1940) The structure and relationships of Protoceratops. Annals of the New York Academy of Sciences 40: 133–266.
- 52. You H, Tanoue K, Dodson P (2010) A new species of Archaeoceratops (Dinosauria: Neoceratopsia) from the Early Cretaceous of the Mazongshan area, northwestern China. In: Ryan MJ, Chinnery-Allgeier BJ, Eberth DA, editors. New Perspectives on Horned Dinosaurs: The Royal Tyrrell Museum Ceratopsian Symposium. Bloomington, Indiana: Indiana University Press. pp. 68–82.
- 53. Sereno PC (2010) Taxonomy, cranial morphology, and relationships of parrot-beaked dinosaurs (Ceratopsia: Psittacosaurus). In: Ryan M, Chinnery-Allgeier BJ, Eberth DA, editors. New Perspectives on Horned Dinosaurs: The Royal Tyrrell Museum Ceratopsian Symposium. Bloomington: Indiana University Press. pp. 21–58.
- 54. Farke AA (2010) Evolution, homology, and function of the supracranial sinuses in ceratopsian dinosaurs. Journal of Vertebrate Paleontology 30: 1486–1500
- 55. Goodwin MB, Clemens WA, Horner JR, Padian K (2006) The smallest known Triceratops skull: new observations on ceratopsid cranial anatomy and ontogeny. Journal of Vertebrate Paleontology 26: 103–112
- 56. Sampson SD, Ryan MJ, Tanke DH (1997) Craniofacial ontogeny in centrosaurine dinosaurs (Ornithischia: Ceratopsidae): taxonomic and behavioral implications. Zoological Journal of the Linnean Society 121: 293–337.
- 57. You H, Tanoue K, Dodson P (2007) A new specimen of Liaoceratops yanzigouensis (Dinosauria: Neoceratopsia) from the Early Cretaceous of Liaoning Province, China. Acta Geologica Sinica - English Edition 81: 898–904
- 58. Coombs WP (1982) Juvenile specimens of the ornithischian dinosaur Psittacosaurus. Palaeontology 25: 89–107.
- 59. Tanoue K, You H, Dodson P (2010) Mandibular anatomy in basal Ceratopsia. In: Ryan MJ, Chinnery-Allgeier BJ, Eberth DA, editors. New Perspectives on Horned Dinosaurs: The Royal Tyrrell Museum Ceratopsian Symposium. Bloomington, Indiana: Indiana University Press. pp. 234–250.
- 60. Tanoue K, You H, Dodson P (2009) Comparative anatomy of selected basal ceratopsian dentitions. Canadian Journal of Earth Sciences 46: 425–439
- 61. Edmund AG (1960) Tooth replacement phenomena in the lower vertebrates. Contributions of the Royal Ontario Museum, Life Science Division 52: 1–190.
- 62. Handa N, Watabe M, Tsogtbaatar K (2012) New specimens of Protoceratops (Dinosauria: Neoceratopsia) from the Upper Cretaceous in Udyn Sayr, southern Gobi area, Mongolia. Paleontological Research 16: 179–198
- 63. Makovicky PJ (2010) A redescription of the Montanoceratops cerorhynchus holotype, with a review of referred material. In: Ryan MJ, Chinnery-Allgeier BJ, Eberth DA, editors. New Perspectives on Horned Dinosaurs: The Royal Tyrrell Museum Ceratopsian Symposium. Bloomington, Indiana: Indiana University Press. pp. 68–82.
- 64. Lee Y-N, Ryan M, Kobayashi Y (2011) The first ceratopsian dinosaur from South Korea. Naturwissenschaften 98: 39–49
- 65. Ryan MJ, Evans DC, Currie PJ, Brown CM, Brinkman D (2012) New leptoceratopsids from the Upper Cretaceous of Alberta, Canada. Cretaceous Research 35: 69–80
- 66. Ősi A, Butler RJ, Weishampel DB (2010) A Late Cretaceous ceratopsian dinosaur from Europe with Asian affinities. Nature 465: 466–468
- 67. Lambe LM (1902) New genera and species from the Belly River series (mid-Cretaceous). Contributions to Canadian Palaeontology 3, Pt 2: 23–81.
- 68. Kirkland JI, DeBlieux DD (2010) New basal centrosaurine ceratopsian skulls from the Wahweap Formation (Middle Campanian), Grand Staircase–Escalante National Monument, southern Utah. In: Ryan MJ, Chinnery-Allgeier, B J., Eberth, D A., editors. New Perspectives on Horned Dinosaurs: The Royal Tyrrell Museum Ceratopsian Symposium. Bloomington: Indiana University Press. pp. 117–140.
- 69. Nessov LA, Kaznyshkina LF, Cherepanov GO (1989) Ceratopsian dinosaurs and crocodiles of the Mesozoic of middle Asia. In: Bogdanova TN, Khozatsky LI, editors. Theoretical and Applied Aspects of Modern Paleontology. Leningrad: Nauka. pp. 144–154.
- 70. Xu X, Wang K, Zhao X, Li D (2010) First ceratopsid dinosaur from China and its biogeographical implications. Chinese Science Bulletin 55: 1631–1635
- 71. Lambe LM (1904) On the squamoso-parietal crest of two species of horned dinosaurs from the Cretaceous of Alberta. The Ottawa Naturalist 18: 81–84.
- 72. Wolfe DG, Kirkland JI (1998) Zuniceratops christopheri n. gen. & n. sp., a ceratopsian dinosaur from the Moreno Hill Formation (Cretaceous, Turonian) of west-central New Mexico. Bulletin of the New Mexico Museum of Natural History and Science 14: 303–317.
- 73. Goloboff PA, Farris JS, Nixon KC (2008) TNT, a free program for phylogenetic analysis. Cladistics 24: 774–786
- 74. Huxley TH (1869) On Hypsilophodon, a new genus of Dinosauria. Geological Society of London, Abstracts of Proceedings 204: 3–4.
- 75. Fowler DW, Woodward HN, Freedman EA, Larson PL, Horner JR (2011) Reanalysis of “Raptorex kriegsteini”: A juvenile tyrannosaurid dinosaur from Mongolia. PLoS ONE 6: e21376
- 76. Campione NE, Brink KS, Freedman EA, McGarrity CT, Evans DC (2013) “Glishades ericksoni”. an indeterminate juvenile hadrosaurid from the Two Medicine Formation of Montana: implications for hadrosauroid diversity in the latest Cretaceous (Campanian–Maastrichtian) of western North America. Palaeobiodiversity and Palaeoenvironments 93: 65–75
- 77. Lamsdell JC, Selden PA (2013) Babes in the wood—a unique window into sea scorpion ontogeny. BMC Evolutionary Biology 13: 98
- 78. Ree RH, Moore BR, Webb CO, Donoghue MJ (2005) A likelihood framework for inferring the evolution of geographic range on phylogenetic trees. Evolution 59: 2299–2311
- 79. Ree RH, Smith SA (2008) Maximum likelihood inference of geographic range evolution by dispersal, local extinction, and cladogenesis. Systematic Biology 57: 4–14
- 80. Nesbitt SJ, Smith ND, Irmis RB, Turner AH, Downs A, et al. (2009) A complete skeleton of a Late Triassic saurischian and the early evolution of dinosaurs. Science 326: 1530–1533
- 81. Loewen MA, Irmis RB, Sertich JJW, Currie PJ, Sampson SD (2013) Tyrant dinosaur evolution tracks the rise and fall of Late Cretaceous oceans. PLoS ONE 8: e79420
- 82. Bapst DW (2012) Paleotree: an R package for paleontological and phylogenetic analyses of evolution. Methods in Ecology and Evolution 3: 803–807
- 83. Hochuli PA, Heimhofer U, Weissert H (2006) Timing of early angiosperm radiation: recalibrating the classical succession. Journal of the Geological Society 163: 587–594
- 84. He HY, Wang XL, Zhou ZH, Jin F, Wang F, et al. (2006) 40Ar/39Ar dating of Lujiatun Bed (Jehol Group) in Liaoning, northeastern China. Geophysical Research Letters 33: L04303
- 85. Chinnery B (2004) Description of Prenoceratops pieganensis gen. et. sp. nov. (Dinosauria: Neoceratopsia) from the Two Medicine Formation of Montana. Journal of Vertebrate Paleontology 24: 572–590
- 86. Smith AG, Smith DG, Funnell BM (1994) Atlas of Mesozoic and Cenozoic Coastlines. Cambridge: Cambridge University Press. 112 p.
- 87. Xu X, Wang K, Zhao X, Sullivan C, Chen S (2010) A new leptoceratopsid (Ornithischia: Ceratopsia) from the Upper Cretaceous of Shandong, China and its implications for neoceratopsian evolution. PLoS ONE 5: e13835
- 88. Gao K, Norell MA (2000) Taxonomic composition and systematics of Late Cretaceous lizard assemblages from Ukhaa Tolgod and adjacent localities, Mongolian Gobi Desert. Bulletin of the American Museum of Natural History 249: 1–118
- 89. Dashzeveg D, Dingus L, Loope DB, Swisher CC, Dulam T, et al. (2005) New stratigraphic subdivision, depositional environment, and age estimate for the Upper Cretaceous Djadokhta Formation, southern Ulan Nur Basin, Mongolia. American Museum Novitates 3498: 1–31
- 90. Sues H-D, Averianov A (2009) Turanoceratops tardabilis-the first ceratopsid dinosaur from Asia. Naturwissenschaften 96: 645–652
- 91. Haq BU, Hardenbol J, Vail PR (1987) Chronology of fluctuating sea levels since the Triassic. Science 235: 1156–1167
- 92. Miller KG, Kominz MA, Browning JV, Wright JD, Mountain GS, et al. (2005) The Phanerozoic record of global sea-level change. Science 310: 1293–1298
- 93. Chang S, Zhang H, Renne PR, Fang Y (2009) High-precision 40Ar/39Ar age for the Jehol Biota. Palaeogeography, Palaeoclimatology, Palaeoecology 280: 94–104
- 94. Sun G, Dilcher DL, Wang H, Chen Z (2011) A eudicot from the Early Cretaceous of China. Nature 471: 625–628
- 95. Zanno LE (2010) A taxonomic and phylogenetic re-evaluation of Therizinosauria (Dinosauria: Maniraptora). Journal of Systematic Palaeontology 8: 503–543
- 96. McDonald AT, Kirkland JI, DeBlieux DD, Madsen SK, Cavin J, et al. (2010) New basal iguanodonts from the Cedar Mountain Formation of Utah and the evolution of thumb-spiked dinosaurs. PLoS ONE 5: e14075
- 97. Prieto-Márquez A (2010) Global historical biogeography of hadrosaurid dinosaurs. Zoological Journal of the Linnean Society 159: 503–525
- 98. Cuenca Bescós G, Canudo JI, Gasca JM, Moreno-Azanza M, Cifelli RL Spalacotheriid symmetrodonts from the Early Cretaceous of Spain. Journal of Vertebrate Paleontology. In press.
- 99. Simpson GG (1928) A Catalogue of the Mesozoic Mammalia in the Geological Department of the British Museum. London: Trustees of the British Museum. 215 p.
- 100. Weishampel DB, Bjork PR (1989) The first indisputable remains of Iguanodon (Ornithischia: Ornithopoda) from North America: Iguanodon lakotaensis, sp. nov. Journal of Vertebrate Paleontology 9: 56–66
- 101. Pereda-Suberbiola J (1994) Polacanthus (Ornithischia, Ankylosauria), a transatlantic armoured dinosaur from the Early Cretaceous of Europe and North America. Palaeontographica Abteilung A 232: 133–159.
- 102. Kirkland JI, Britt BB, Burge DL, Carpenter K, Cifelli RL, et al. (1997) Lower to middle Cretaceous dinosaur faunas of the central Colorado Plateau: a key to understanding 35 million years of tectonics, sedimentology, evolution and biogeography. Brigham Young University Geology Studies 42: 69–103.
- 103. Cuenca-Bescós G, Canudo J (2003) A new gobiconodontid mammal from the Early Cretaceous of Spain and its palaeogeographic implications. Acta Palaeontologica Polonica 48: 575–582.
- 104. Sweetman SC (2006) A gobiconodontid (Mammalia, Eutriconodonta) from the Early Cretaceous (Barremian) Wessex Formation of the Isle of Wight, Southern Britain. Palaeontology 49: 889–897
- 105. Sweetman SC (2008) A spalacolestine spalacotheriid (Mammalia, Trechnotheria) from the Early Cretaceous (Barremian) of southern England and its bearing on spalacotheriid evolution. Palaeontology 51: 1367–1385
- 106. Kielan-Jaworowska Z, Cifelli RL, Luo Z-X (2004) Mammals from the Age of Dinosaurs: Origins, Evolution and Structure. New York: Columbia University Press. 630 p.
- 107. Cifelli RL, Davis BM Tribosphenic mammals from the Lower Cretaceous Cloverly Formation of Montana and Wyoming. Journal of Vertebrate Paleontology. In press.
- 108. Tsubamoto T, Rougier GW, Isaji S, Manabe M, Forasiepi AM (2004) New Early Cretaceous spalacotheriid “symmetrodont” mammal from Japan. Acta Palaeontologica Polonica 49: 329–346.
- 109. Li G, Luo Z-X (2006) A Cretaceous symmetrodont therian with some monotreme-like postcranial features. Nature 439: 195–200
- 110. Nydam RL (2013) Squamates from the Jurassic and Cretaceous of North America. Palaeobiodiversity and Palaeoenvironments 93: 535–565
- 111. Blakey R (2014) Paleogeography library, rectangular global maps. Colorado Plateau Geosystems. Available: http://cpgeosystems.com/rect_globe.html. Accessed 2014 Sep 1.
- 112. Hay WW, DeConto RM, Wold CN, Wilson KM, Voigt S, et al. (1999) Alternative global Cretaceous paleogeography. Geological Society of America Special Papers 332: 1–47