Osteology and reassessment of Dineobellator notohesperus, a southern eudromaeosaur (Theropoda: Dromaeosauridae: Eudromaeosauria) from the latest Cretaceous of New Mexico

Abstract

Dromaeosaurids (Theropoda: Dromaeosauridae), a group of dynamic, swift predators, have a sparse fossil record, particularly at the end of the Cretaceous Period. The recently described Dineobellator notohesperus, consisting of a partial skeleton from the Upper Cretaceous (Maastrichtian) of New Mexico, is the only diagnostic dromaeosaurid to be recovered from the latest Cretaceous of the southwestern United States. Reinterpreted and newly described material include several caudal vertebrae, portions of the right radius and pubis, and an additional ungual, tentatively inferred to be from manual digit III. Unique features, particularly those of the humerus, unguals, and caudal vertebrae, distinguish D. notohesperus from other known dromaeosaurids. This material indicates different physical attributes among dromaeosaurids, such as use of the forearms, strength in the hands and feet, and mobility of the tail. Several bones in the holotype exhibit abnormal growth and are inferred to be pathologic features resulting from an injury or disease. Similar lengths of the humerus imply Dineobellator and Deinonychus were of similar size, at least regarding length and/or height, although the more gracile nature of the humerus implies Dineobellator was a more lightly built predator. A new phylogenetic analysis recovers D. notohesperus as a dromaeosaurid outside other previously known and named clades. Theropod composition of the Naashoibito Member theropod fauna is like those found in the more northern Late Cretaceous North American ecosystems. Differences in tooth morphologies among recovered theropod teeth from the Naashoibito Member also implies D. notohesperus was not the only dromaeosaurid present in its environment.

1 INTRODUCTION

Dromaeosaurids (Theropoda: Dromaeosauridae) are a group of small- to medium-sized theropods that lived during the Cretaceous Period, mainly on the northern paleocontinent of Laurasia. They have been found in North America throughout the Cretaceous, from present day Alaska to Maryland (e.g., Jasinski & Dodson, 2015; Matthew & Brown, 1922; Turner et al., 2012). However, their fossil record is poor near the end of the Cretaceous and the Cretaceous-Paleogene boundary in North America. Early Cretaceous North American dromaeosaurids include Deinonychus antirrhopus, Utahraptor ostrommaysi, and Yurgovuchia doellingi (e.g., Brinkman et al., 1998; Kirkland et al., 1993; Ostrom, 1969; Senter et al., 2012). Several other taxa are known from the Late Cretaceous, but almost all are from the Campanian, although it is noted that Atrociraptor marshalli comes from the Maastrichtian (probably early Maastrichtian) portion of the Horseshoe Canyon Formation (e.g., Burnham, 2004; Burnham et al., 2000; Currie & Varricchio, 2004; Longrich & Currie, 2009; Larson et al., 2010; Jasinski, 2015a; Jasinski & Dodson, 2015; Jasinski et al., 2015; Matthew & Brown, 1922; Sues, 1978). Recently, two taxa (Acheroraptor temertyorum and Dakotaraptor steini) were named from the upper Maastrichtian Hell Creek Formation, but, aside from these two skeletal fossil specimens, non-tooth material of Maastrichtian taxa is rare (e.g., DePalma et al., 2015; Evans et al., 2013; Jasinski & Dodson, 2015). Although isolated dromaeosaurid teeth are somewhat common in Campanian age strata of North America, these teeth reveal relatively little information about the paleoecology of this group (Jasinski et al., 2020).

The dromaeosaurid Dineobellator notohesperus was recently described from the Maastrichtian (latest Cretaceous) of New Mexico (Jasinski et al., 2020). The holotype specimen of D. notohesperus (SMP VP-2430), collected from the Naashoibito Member of the Ojo Alamo Formation in the San Juan Basin in northwestern New Mexico, consists of over 20 identifiable skeletal elements, including parts of the skull, axial skeleton, and fore- and hindlimbs. Since its original publication, additional observations and reinterpretations have been made and are presented below, allowing for a more robust understanding of its place within Dromaeosauridae. This more detailed account allows for a reassessment of the phylogenetic relationships of D. notohesperus and further provides insight into the paleoecology of this late surviving dromaeosaurid in what is now New Mexico at the end of the Cretaceous.

Institutional abbreviations: AMNH FARB, American Museum of Natural History, New York, New York; IGM, Mongolian Institute of Geology, Ulaan Bataar, Mongolia; IVPP, Institute of Vertebrate Paleontology and Paleoanthropology, Beijing, China; MNHN, Muséum National d'Histoire Naturelle, Paris; MPC, Paleontological and Geological Center of the Mongolian Academy of Sciences, Ulaan Bataar, Mongolia; OMNH, Oklahoma Museum of Natural History, Norman, Oklahoma; SMP, State Museum of Pennsylvania, Harrisburg, Pennsylvania; TMP, Royal Tyrrell Museum of Palaeontology, Drumheller, Alberta; UALVP, University of Alberta Laboratory of Vertebrate Palaeontology, Drumheller, Alberta.

Anatomical abbreviations: ab, abnormal bone growth; acdl, anterior centrodiapophyseal lamina; al, alveolus; as, articular surface; avg, accessory vascular groove; bptp, basipterygoid process; bspr, basisphenoid recess; bt, basal tubera; capt, capitulum; cg, costal groove; con, contact; dci, distal (or caudal) indentation; dcs, distal (or caudal) centrum surface; dpc, deltopectoral crest; eg, extensor groove; fhc, curvature of femoral head; fm, foramen magnum; for, foramen; ft, flexor tubercle; ind, indentation; lcd, distal lateral condyle; lct, lateral cotyle; ld, m. latissimus dorsi scar; lg, lateral groove; lr, lacrimal recess; mc, medial crest; mcd, distal medial condyle; mcr, middle cotylar ridge; mct, medial cotyle; mg, medial groove; na, neural arch; nc, neural canal; ns, neural spine; oc, occipital condyle; op, olecranon process; or, orbital rim; pci, proximal (or cranial) centrum indentation; pcs, proximal (or cranial) centrum surface; pocdf, postzygapophyseal centrodiapophyseal fossa; prcdf, prezygapophyseal centrodiapophyseal fossa; prdl, prezygadiapophyseal lamina; rs, rugose surface; sdf, spinodiapophyseal fossa; spof, spinopostzygapophyseal fossa; sprf, spinoprezygapophyseal fossa; ss, sutural surface; tp, transverse process; tub, tuberculum; up, ulnar papillae; ur, ulnar ridge; vg, ventral groove; vr, ventral ridge; zr, zygapophyseal rod.

1.1 Geologic setting

The holotype of Dineobellator notohesperus (SMP VP-2430) was discovered and collected a few meters above the base of the Naashoibito Member of the Ojo Alamo Formation, San Juan Basin, New Mexico. The specimen was found weathering out of a relatively poorly consolidated sandstone (Figure 1). ⁴⁰Ar/³⁹Ar dates acquired from detrital sanidines give a maximum depositional age for the Naashoibito Member at 66.5 ± 0.2 Ma (upper Maastrichtian) (Greene et al., 2018; Heizler et al., 2013; Mason et al., 2013; Peppe et al., 2013; Williamson & Brusatte, 2014). However, biostratigraphic evidence suggests an early late Maastrichtian age of about 70.0–68.0 Ma (Fowler, 2017).

FIGURE 1

(a) Geographic map showing Late Cretaceous through Eocene strata of the San Juan Basin, where the type specimen of Dineobellator notohesperus (SMP VP-2430) was collected (marked by a star). (b) Stratigraphic chart showing significant strata of the San Juan Basin. It is noted that within the NALMA/NALVA column, most of the ages are the North American Land Mammal Ages, however, the Kirtlandian was defined by Sullivan and Lucas (2003, 2006) as a North American Land-Vertebrate Age based on (mainly) non-mammal components. Additionally, the age of the Naashoibito Member of the Ojo Alamo Formation has been debated as to whether it is from the Edmontonian or Lancian (e.g., Jasinski, Sullivan, & Lucas, 2011). Figure pools information from multiple sources, including the Geologic Map of New Mexico (2003), Sullivan and Lucas (2003, 2006), Williamson and Weil (2008a, 2008b), Lucas et al. (2009, 2016), Jasinski and Sullivan (2011, 2016), Jasinski, Sullivan, and Lucas (2011); Jasinski, Lucas, and Moscato (2011); Koenig et al. (2012), Sullivan and Jasinski (2012), Williamson and Brusatte (2014, 2016), Jasinski et al. (2015, 2020), Jasinski and Dodson (2015), and Ksepka et al. (2017). Clarkfork, Clarkforkian; Ed, Edmontonian; Lanc, Lancian; Maast, Maastrichtian; NALMA, North American Land Mammal Age; NALVA, North American Land-Vertebrate Age; Sel, Selandian; Tiff, Tiffanian

2 MATERIALS AND METHODS

SMP VP-2430 was first reported by Jasinski, Sullivan, and Lucas (2011), who provided a brief description of some of the material and identified it as an indeterminate dromaeosaurid and possibly a new species. Jasinski et al. (2020) described the material further while naming the taxon Dineobellator notohesperus and conducted phylogenetic analyses investigating its evolutionary relationships. Because the description provided by Jasinski et al. (2020) was relatively short, a more thorough description of this specimen is warranted and provided here. We also reidentify some of the previously described elements along with descriptions of newly identified material. This includes not only the original material of the holotype specimen that was largely collected in 2008 (see Jasinski et al., 2020), but also additional elements subsequently collected from the type locality in 2009, 2015, and 2016. Anatomical terminology utilizes rostral, cranial (=anterior), and caudal (=posterior). For caudal vertebrae, we utilize proximal (=anterior or cranial) and distal (=posterior or caudal) directions or surfaces.

A new phylogenetic analysis, including Dineobellator notohesperus, was run using the recent dataset of Powers et al. (2022) to further investigate the evolutionary relationships of dromaeosaurids, with a focus on eudromaeosaurs (Eudromaeosauria). Our results are compared to those of Jasinski et al. (2020) in two other datasets [those of Currie and Evans (2019) and the Theropod Working Group dataset which was more recently utilized by Cau et al. (2015, 2017) and Turner et al. (2021)] to further compare those relationships within Dromaeosauridae. The Powers et al. (2022) dataset originally ran 175 characters on 25 operational taxonomic units (24 ingroup OTUs). “Velociraptor sp.” was removed from their dataset as this is a potentially new taxon that they are working on further. It was replaced with Dineobellator notohesperus, resulting in the same number of characters and OTUs. The Powers et al. (2022) dataset is a modified version of the dataset of Powers (2020), which is a further modification of that of Currie and Evans (2019). Current data were run with TNT version 1.5 (Goloboff & Catalano, 2016). The dataset was subjected to a New Technology search (with default parameters for sectorial search, ratchet, tree drift, and tree fusion). Bootstrap resampling was done using 100 replicates to compare the results more closely to those of Powers et al. (2022).

2.1 Systematic paleontology

Dinosauria Owen, 1842

Theropoda Marsh, 1881

Coelurosauria Huene, 1914

Dromaeosauridae Matthew & Brown, 1922

Dineobellator Jasinski et al., 2020

Dineobellator notohesperus Jasinski et al., 2020

2.1.1 Holotype

SMP VP-2430 is a disarticulated, associated individual (Figure 2) consisting of a rostromedial (anteromedial) portion of the right premaxilla, a left maxillary fragment, ?maxillary tooth, dorsolateral process of the left lacrimal, left? nasal fragment, incomplete right jugal, incomplete left basipterygoid, incomplete occipital condyle, isolated prezygapophyses, isolated vertebral processes, caudal vertebra 1, middle caudal vertebra, three more distal caudal vertebrae (distal or posterior to mid-caudal vertebra), four fused distal caudal vertebrae, several vertebral fragments, nearly complete rib and rib fragments, nearly complete right humerus, nearly complete right ulna, incomplete right radius, incomplete right metacarpal III, nearly complete right manual ungual II, nearly complete right? manual ungual? III, fragmentary right pubis, incomplete right femur, incomplete right metatarsals II and III, right pedal phalanx I-I, nearly complete right pedal ungual III, and various other cranial and post-cranial bone fragments.

FIGURE 2

Skeletal reconstruction of Dineobellator notohesperus (SMP VP-2430) with known elements colored in white. Figured bones are as follows: Fused distal caudal vertebra (a); distal caudal vertebra (b); middle caudal vertebra (c); caudal vertebra 1 (d); right femur (e); rib (f); right basipterygoid (g); left lacrimal (reversed) (h); right jugal (i); right humerus (j); right ulna (k); right metacarpal III (l); right manual ungual II (m); right metatarsal II (n); right metatarsal III (o); right pubic boot (p). Individual scale bars, 2 cm. Figure modified from Jasinski et al. (2020, figure 2)

2.1.2 Type locality and horizon

The type locality, SMP 410b, Bisti/De-na-zin Wilderness, New Mexico. Precise locality information is on file at the State Museum of Pennsylvania, Section of Paleontology and Geology, and is available to qualified researchers. The holotype (SMP VP-2430) was collected in a weathered sandstone a few meters above the base of the Naashoibito Member (Ojo Alamo Formation). ⁴⁰Ar/³⁹Ar dates acquired from detrital sanidines give a maximum depositional age for the Naashoibito Member at 66.5 ± 0.2 Ma (upper Maastrichtian; Greene et al., 2018; Heizler et al., 2013; Mason et al., 2013; Peppe et al., 2013; Williamson & Brusatte, 2014). Biostratigraphic evidence, however, suggests an early late Maastrichtian age, approximately 70.0–68.0 Ma (Fowler, 2017; Jasinski, Sullivan, & Lucas, 2011).

2.1.3 Revised diagnosis

A mid-sized dromaeosaurid theropod that differs from other eudromaeosaurs by the following characters: offset of medial and lateral grooves (or blood grooves) on manual ungual; distinct and conspicuous mediodorsal groove proximally, dorsal to the articular surface of the manual ungual; sharp angle of distal deltopectoral crest of the humerus; opisthocoelous proximal caudal vertebrae, with gently convex proximal (anterior) centrum surface; short and robust neural spines on proximal caudal vertebrae; gracile and sub-rectangular transverse processes on proximal caudal vertebrae; proximal caudal vertebrae with curved ventral surface and oval to sub-rectangular proximal and distal (posterior) centrum surfaces; distinct round concavities on proximal and distal centrum surfaces in mid-caudal vertebrae; enlarged flexor tubercles on manual ungual II and pedal ungual III; and secondary lateral grooves (or accessory vascular grooves) on pedal unguals on both the medial and lateral surfaces, with those on the lateral surfaces more defined and conspicuous.

2.2 Descriptions and comparisons

Jasinski et al. (2020) briefly described much of the material considered here for the holotype specimen of Dineobellator notohesperus (SMP VP-2430). Here we provide a more detailed account of the holotype, along with comparisons to other known dromaeosaurid taxa and material. Moreover, we have re-identified some of material described by Jasinski et al. (2020), along with documenting several elements not mentioned in their work. This present contribution provides the opportunity to figure and illustrate all the pertinent elements, thus giving us a clearer picture of this unique dromaeosaurid dinosaur.

2.3 Skull

A few small fragments of SMP VP-2430 are from the skull of Dineobellator notohesperus. A small, rostromedial fragment of the right premaxilla is preserved bearing two empty alveoli, which are visible caudomedially (Figure 3a,b). Rostrally the surface is smooth where the rostral parts of the left and right premaxillae would contact, roughly parallel to the midline of the skull. The remnants of the alveoli are quite small labially (medially; see Table 1 for all measurements) and closely spaced, similar to the condition found in other dromaeosaurids. Portions of the lateral surface of the premaxilla are rugose and may be pathologic. The fragment implies a more vertical, rather than sloped, orientation of the premaxilla. This more vertical orientation would help distinguish Dineobellator from several other dromaeosaurids that tend to have more sloping premaxillae (Halszkaraptor, Sinornithosaurus, Utahraptor).

FIGURE 3

Premaxilla, maxilla, and tooth material of Dineobellator notohesperus (SMP VP-2430). (a, b) Rostromedial fragment of right premaxilla in rostral (a) and caudal (b) views. (c, d) left maxilla fragment in lateral (c) and medial (d) views. (e, f) ?Maxillary tooth showing full tooth (e), and magnified image showing denticle morphology (f). Scale bars corresponding to (a–d) and (e) equal 1 cm. Scale bar for (f) equals 1 mm

TABLE 1. Anatomical measurements of the holotype of Dineobellator notohesperus (SMP VP-2430)

Selected measurements of Dineobellator notohesperus (SMP VP-2430)
Element	Length (in mm)	Notes
Premaxilla (right)
Length	8.86^a
Depth	12.38^a
Alveolus “a” length	2.28
Alveolus “b” length	1.81
Maxilla
Length (rostrocaudal)	6.34^a
Depth (dorsoventral)	9.09^a
Alveolus “a” length	3.03
?Maxillary tooth
Apical length	12
Crown height	11.3
Fore-aft basal length	7.78
Basal width	4.15
Wear facet length	6.7
Lacrimal (left)
Length	9.00^a
Depth	6.96^a
Nasal (left)
Length	18.51^a
Width	6.71^a
Jugal (right)
Length (craniocaudal)	37.17^a
Depth (dorsoventral)	20.66^a	Maximum preserved depth measured near caudal-most portion of bone
Basioccipital
Length (rostrocaudal)	14.79^a
Width (mediolateral)	16.04^a
Basipterygoid Process (right)
Length (rostrocaudal)	31.67	Between the basal tubera and the basipterygoid process
Height (dorsoventral)	13.8	Measured posteriorly at the basal tubera
Width A (mediolateral)	22.06	Measured across the caudal portion with the basal tubera
Width B (mediolateral)	11.58	Measured across the middle constriction just rostral to the carotid canal
Width C (mediolateral)	8.81	Measured across the rostral portion with the basipterygoid process
Anterior Caudal Vertebra #1
Length (proximodistal)	26.4
Height (dorsoventral)	30.32	From ventral edge to dorsal edge of neural spine
Width (mediolateral)	15.9
Proximal centrum surface height (dorsoventral)	12.75
Proximal centrum surface width (mediolateral)	19.3
Distal centrum surface height (dorsoventral)	14.1
Distal centrum surface width (mediolateral)	21.95
Neural spine height (dorsoventral)	15.3
Midcaudal Vertebra (~8 to ~12)
Length (proximodistal)	32.56
Proximal width (mediolateral)	15.23
Distal width (mediolateral)	14.71
Mid-length width (mediolateral)	10.86
Neural arch length (proximodistal)	28.33
Caudal Vertebra (third identified)
Length (proximodistal)	34.4	This is the preserved length as the proximal portion is not preserved
Caudal Vertebra (fourth identified)
Length (proximodistal)	46.26
Caudal Vertebra (fifth identified)
Length (proximodistal)	22.35	This is the preserved length as the distal portion is not preserved
Proximal centrum surface height (dorsoventral)	7.21
Proximal centrum surface width (mediolateral)	9.96
Distal Caudal Vertebrae
Fused section length (proximodistal)	19.95^a
Fused section height (dorsoventral)	9.5	Maximum height
Caudal vertebra a length (proximodistal)	4.12	This represents one of the two complete distal caudal vertebrae in the fused section
Caudal vertebra b length (proximodistal)	5.1	This represents one of the two complete distal caudal vertebrae in the fused section
Rib A		Largest incomplete rib
Length (proximodistal)	100.55^a
Width A (craniocaudal)	25.12	At proximal end across the “D”
Width B (craniocaudal)	10.35	At distally preserved end
Humerus (right)
Length (proximodistal)	185.78^a
Shaft diameter	17.25–18.52
Shaft width	17.1	proximal, just distal to deltopectoral crest
Shaft length	91.27^a	distal to deltopectoral crest
Ulna (right)
Length (proximodistal)	100.96^a
Width	15.71	Maximum width measured across trochlea and olecranon process
Shaft width	12	Average shaft width
Radius (right)
Height (craniocaudal)	15.16	Measured at distal end
Metacarpal III (right)
Length (proximodistal)	72.75^a
Width (craniocaudal)	16.68^a	Measured at proximal end
Manual Ungual II (right)
Length (proximodistal)	45.64^a	Measured from ventral edge of proximal articulation surface to preserved distal tip
Length (proximodistal)	46.97^a	Measured from ventroproximal edge of flexor tubercle to preserved distal tip
Height (dorsoventral)	29.95	Measured at the proximal end (from dorsal edge of articulation surface to ventral edge of flexor tubercle)
Height (dorsoventral)	16.46	Measured at the midlength of the claw
Height (dorsoventral)	9.1	Measured at the preserved distal tip
Articulation surface height (dorsoventral)	18
Articulation surface width (mediolateral)	8.33
Flexor tubercle height (dorsoventral)	16.75
Flexor tubercle width (mediolateral)	6.7
?Manual Ungual? III (right)
Length (proximodistal)	30.38	Measured from broken proximal surface to distal tip
Height (dorsoventral)	8.16	Measured at proximal broken surface
Height (dorsoventral)	6.71	Measured at midlength of the preserved portion of the claw
Width (mediolateral)	4.01	Measureed at broken proximal surface dorsal to medial groove
Width (mediolateral)	3.61	Measureed at broken proximal surface ventral to medial groove
Pubis (right)
Length (proximodistal)	21.33	Measured along what would be the distal pubic shaft to the proximodistal-most point of pubic boot
Boot length (craniocaudal)	19.5	Measured along the craniocaudal segment of the pubic boot
Femur (right)
Length (proximodistal)	68.83^a
Width	20.18^a
Metatarsal I (right)
Length (proximodistal)	11.43^a
Width	7.2	Maximum width measured distally
Metatarsal II (right)
Length (proximal section, proximodistal)	46.79^a
Shaft width (proximodistal section)	18.14	Maximum width measured at preserved distal end
Shaft width (proximodistal section)	8.75	Proximal-most shaft width
Length (distal section, proximodistal)	33.3^a
Shaft width (distal section)	11.52	Maximum width measured at distal end
Metatarsal III (right)
Length (proximodistal)	47.14^a
Width	14.04	Maximum width measured at preserved distal end
?Astragalus (left)
Length (proximodistal)	10.6^a
Width	6.82^a
Pedal Ungual III (right)
Length (proximal section, proximodistal)	13.23^a
Height (proximal section, dorsoventral)	21.3
Articulation surface height (dorsoventral)	10.65
Articulation surface width (mediolateral)	4.12
Flexor tubercle height (dorsoventral)	7.1
Flexor tubercle width (mediolateral)	3.53
Length (distal section, proximodistal)	32.03^a
Height (distal section, dorsoventral)	11.95	Measured at preserved proximal end
Height (distal section, dorsoventral)	9.05	Measured at preserved distal end

• ^a Incomplete element, measurement as preserved.

The left maxilla is represented by a small, sub-rectangular fragment (Figure 3c,d). Although a few small, questionable, randomly distributed foramina are present, it is otherwise smooth on its lateral surface. Two partial alveoli are present lingually (medially). The more complete alveolus on the maxillary fragment is larger and this is more prominent than any preserved on the premaxillary fragment. Regardless, the small alveoli mean the fragment is from the caudal portion of the left maxilla.

A single tooth, possibly pertaining to the maxilla, is small and relatively gracile (Figure 3e,f). The apical length measures 12.0 mm in total, with a crown height of 11.3 mm (Table 2). There are approximately 18–20 denticles per 5 mm on the distal (or posterior) carina (distal basal denticles), but no denticles on the mesial (or anterior) carina. The presence or absence of denticles on the mesial carina (see Smith, 2005; Torices et al., 2014) is highly variable among dromaeosaurids but, in addition to Dineobellator, they are absent in Bambiraptor feinbergi and several Asian taxa including Linheraptor exquisitus, Tsaagan mangas, and Velociraptor osmolskae (e.g., Burnham, 2004; Godefroit et al., 2008; Norell et al., 2006; Xu et al., 2010). The presence or absence of denticles on carinae can also sometimes vary along the tooth row in some taxa, particularly in earlier diverging dromaeosaurids (e.g., Poust et al., 2020), as well as the potential development of denticles in the later ontogenetic stages of some theropods (e.g., Carr & Williamson, 2004; Currie et al., 1990). There are 3.7–4.3 denticles per 1 mm, most easily seen close to the base of the distal carina where the denticles are best preserved (Figure 3f). The angle between the lines of 10% and 90% of the length of the exposed denticles normally falls between 86°–95° for denticles that are not worn down, broken, or deformed, with most falling just under 90°. This measurement was done in a similar fashion to that of Larson (2008) to evaluate denticle shape, with the lines of 10% and 90% used to remove any irregularities in the extreme ends of the denticles. The denticles on the distal carina are mostly rounded off and show no traces of hooks. Additionally, the denticles tend to be proximodistally (or dorsoventrally) shallow in relation to craniocaudal depth. Additional dromaeosaurids possessing distal (caudal) denticles that are not apically hooked (and instead are rounded) are Acheroraptor temertyorum and Tsaagan mangas whereas others, such as Atrociraptor marshalli and Saurornitholestes langstoni, are apically hooked (Currie & Evans, 2019; Currie & Varricchio, 2004; Larson, 2008). On the apical end of the rostral edge of the tooth, there is a wear facet that measures approximately 7 mm along the curvature of the tooth from the distal tooth tip where it would occlude with a dentary tooth. The tooth is not constricted between root and crown, has a concave-curve caudally, and would not have been strongly raked in the alveolus due to the orientation of the alveoli. Raked here refers to the tooth curving caudally, as in the motion of sweeping with a rake or broom. This latter aspect of the teeth differs from those of Atrociraptor, Bambiraptor, and Deinonychus, where the teeth are strongly raked caudally (e.g., Burnham, 2004; Burnham et al., 2000; Currie & Varricchio, 2004; Ostrom, 1969).

TABLE 2. Tooth measurements of dromaeosaurid teeth from the late cretaceous of the San Juan Basin, New Mexico

Specimen number	Taxonomic identity	Element identity	Apical length*	Crown height*	Fore-aft basal length*	Basal width*	Denticles per 5 mm
SMP VP-2430	Dineobellator notohesperus	?Maxillary tooth	12	11.3	7.78	4.15	18–20
SMP VP-2595	Dromaeosauridae indeterminate	Tooth	14.25	12.59	8.41	4.71	17
NMMNH P-32814	Dromaeosauridae morphotype A	Tooth	6.86	5.13	3.46	1.66	25
SMP VP-1901	cf. Saurornitholestes sullivani	Tooth	14.85	12.62	6.75	3.36	14–15

• Note: *Measurements in mm. Identity of NMMNH P-32814 from Williamson and Brusatte (2014).

The right jugal (Figure 4a,b), is represented by a flat, trapezoidal fragment suggesting it was a deep and plate-like element. It is slightly curved laterally toward its rostral and caudal ends. The element is flat and as there is no evidence of any medial or lateral projections on the preserved portions, it is likely that none were present on the complete element as well. The curvature of the dorsal edge of the jugal indicates the orbit was circular. Most of the margin of the bone is incomplete, although, in addition to a ventral portion of the orbit, there is also a region caudodorsally where the thin ventral margin is complete and gracile. The dorsoventral extent of the caudal portion of the jugal implies it was relatively deep, like some dromaeosaurids such as Adasaurus mongoliensis (Barsbold, 1983; Turner et al., 2012, figure 10a), but distinguishing it from taxa with shallower jugals such as Bambiraptor and Halszkaraptor (Burnham, 2004; Cau, 2020; Cau et al., 2017). There is also a partial foramen present rostrally along the broken surface. This foramen may have been close to the contact between the jugal and the maxilla, although that is uncertain due to its poor preservation. It is similar to foramina present on the lateral surface of the jugal of Linheraptor exquisitus (IVPP V 16923) and Velociraptor mongoliensis (AMNH FARB 6515; Xu et al., 2010, see figure 2; Xu et al., 2015).

FIGURE 4

Lateral and dorsal cranial elements of Dineobellator notohesperus (SMP VP-2430). (a, b) Incomplete right jugal in lateral (a) and medial (b) views. (c, e) Left ?nasal fragment in dorsal (c), ventral (d), and medial (e) views. (f, g) Dorsolateral process of left lacrimal in ventrolateral (f) and dorsomedial (g) views. Scale bars corresponding to (a,b), (c–e), and (f,g) equal 1 cm

Another small fragment is tentatively identified as part of the left nasal (Figure 4c–e). It is sub-rectangular, with an enlarged medial surface that may be the sutural surface between the two nasals, and a flat dorsal surface. If identified correctly, it shows the nasals are flat dorsally, like in Bambiraptor and Deinonychus (e.g., Burnham, 2004; Burnham et al., 2000; Ostrom, 1969).

A fragment of the left lacrimal is small, robust, and sub-triangular with a rounded point laterally (Figure 4f,g). It preserves the lateral process with a small concavity on the rostroventral portion of the rostrolateral edge (lacrimal fenestra or lacrimal recess) and a smooth surface along the caudolateral edge. The element projects rostrally beyond the extent of the “sub-triangular” form, and the lacrimal presumably would have been “T”-shaped if complete. The bone is incomplete with no sutural surfaces preserved.

The braincase is incomplete (Figure 5). The basioccipital condyle is preserved (Figure 5d). It is subcircular and the entire element is obliquely twisted. The shape of the dorsocaudal portion means the foramen magnum was circular to subcircular, which would distinguish it from the dorsoventrally tall oval foramen magnum in Tsaagan mangas (MPC 100/1015). However, the shape in the latter may be due, at least partially, to post-depositional mediolateral deformation (Norell et al., 2006). As only the most caudal portion of the occipital condyle is preserved, no cranial nerve passages are present.

FIGURE 5

Braincase elements of Dineobellator notohesperus (SMP VP-2430). (a–c) Incomplete basisphenoid, including portions of the left basipterygoid in ventral (a), dorsolateral (b), and caudal (c) views. (d) Occipital condyle in caudal view. Scale bar equals 1 cm

A portion of the basisphenoid, including the left basipterygoid process, is preserved (Figure 5a–c). A small, but prominent, subcircular basipterygoid recess is visible medially. This recess is subcircular whereas it tends to be more craniocaudally elongate in other dromaeosaurids such as Dromaeosaurus (Currie, 1995) and Velociraptor (Norell et al., 2004). The left basal tuber is robust but is incomplete medially. The hypophyseal fossa may be present but incomplete, as breakage makes its identification tentative. The basisphenoid extends rostrally with a portion of the well-developed left basipterygoid process preserved. The rostral-most edge of this process is directed rostrolaterally. The caudal surface is smooth and flat, and the processes are separated by a deep U-shaped notch, which is partially visible medially. This deep U-shaped notch further distinguishes Dineobellator from dromaeosaurids with a smaller and shorter notch between the basipterygoid processes such as Dromaeosaurus (AMNH FARB 5356). Furthermore, the flattened aspect of the caudal surface of the basal tuber further distinguishes it from Tsaagan mangas (Norell et al., 2006) and Velociraptor (Norell et al., 2004; Norell & Makovicky, 1999). Dorsomedial to this notch lies part of the basipterygoid recess. This recess is prominent externally, continues as a thin canal internally, and is directed caudodorsally. Portions of the carotid canal are present on the medial edge of the basipterygoid recess and extend rostrocaudally. Dorsomedially, the carotid canal contains multiple small foramina. Ventrally, the opening for the carotid canal is 6.5 mm long and oval. It is noted that Jasinski et al. (2020) incorrectly identified the basisphenoid process as being from the right side, but their other observations were correct. Several other problematic bone fragments are tentatively identified as coming from the skull, but their exact identifications are uncertain.

2.4 Vertebrae

Several vertebrae and vertebral fragments are preserved in SMP VP-2430, although all material identifiable to a particular part of the vertebral column is part of the caudal series (Figures 6-9). A nearly complete proximal (anterior) caudal vertebra is inferred to be the first caudal (caudal vertebra 1; Figure 6). The centrum of the vertebra is complete, and portions of the prezygapophyses and transverse processes are preserved. The neural arch and spine are nearly complete and robust, and the neural spine is dorsoventrally short. The proximal (anterior) and distal (posterior) centrum surfaces are both oval and wider than high. The centrum surfaces of the caudal vertebrae in other dromaeosaurids tend to be sub-rectangular (e.g., Dakotaraptor, Deinonychus, Saurornitholestes), leading to the more box-like centra of these vertebrae, although some other taxa (e.g., Velociraptor) have been known to have slightly more rounded proximal vertebral centrum surfaces. The dorsal vertebrae of dromaeosaurids tend to have rounded centrum surfaces, so the oval surfaces in Dineobellator potentially mark part of the transition between these morphologies in the vertebral column, although the sacrum may account for more of this transition. Caudal vertebra 1 is opisthocoelous, with a gently convex proximal surface of the centrum and a concave distal surface. The proximal caudal vertebrae of dromaeosaurids are usually acoelous or amphiplatyan, making the opisthocoelous nature of the vertebra unique among known dromaeosaurids, although opisthocoelous proximal caudal vertebrae have been noted in one other theropod, the caenagnathid theropod Gigantoraptor (Xu et al., 2007). Several dromaeosaurids with preserved proximal caudal vertebrae do not show this condition (e.g., Achillobator giganticus, Deinonychus antirrhopus, Kuru kulla, Pyroraptor olympius, Yurgovuchia doellingi; Ostrom, 1969; Perle et al., 1999; Allain & Taquet, 2000; Senter et al., 2012; Napoli et al., 2021). However, it is also noted that most dromaeosaurid specimens with proximal caudal vertebrae are preserved either flattened and in matrix (e.g., Changyuraptor yangi, Halszkaraptor escuilliei, Microraptor gui, M. hanqingi, M. zhaoianus, Tianyuraptor ostromi, Wulong bohaiensis, Zhenyuanlong suni, Zhongjianosaurus yangi; Hwang et al., 2002; Xu et al., 2003; Zheng et al., 2010; Gong et al., 2012; Turner et al., 2012; Han et al., 2014; Cau et al., 2017; Xu & Qin, 2017; Cau, 2020; Poust et al., 2020) and/or articulated (e.g., Bambiraptor feinbergi, Buitreraptor gonzalezorum, Linheraptor exquisitus, Mahakala omnogovae, Shri devi, Sinornithosaurus, Velociraptor mongoliensis; Burnham, 2004; Burnham et al., 2000; Gianechini et al., 2018; Liu et al., 2004; Lü & Brusatte, 2015; Norell & Makovicky, 1997; Norell & Makovicky, 1999; Turner et al., 2011; Turner et al., 2012; Turner et al., 2021; Turner, Pol, et al., 2007; Xu et al., 2010), so comparative material is limited and this condition may be more widely spread than currently known. Additionally, an isolated vertebra (MNHN BO 016) referred to Pyroraptor olympius was mentioned by Allain and Taquet (2000) and described as a proximal caudal vertebra. It was described with concave centrum surfaces, making it amphicoelous. The vertebra figured by Allain and Taquet (2000, figure 2E), which is mislabeled as MNHN BO 017 in the figure caption, agrees morphologically with a proximal caudal vertebra, potentially also the first caudal. Additionally, the recently named velociraptorine Kuru kulla includes the description of a proximal caudal vertebra with a “concave anterior articular surface and weakly concave posterior articular surface” (Napoli et al., 2021, p. 19), also making it amphicoelous, implying there is more variation in proximal caudal vertebrae centrum surfaces in dromaeosaurids than previously noted.

FIGURE 6

Caudal vertebra 1 of Dineobellator notohesperus (SMP VP-2430) in dorsal (a), ventral (b), right lateral (c), left lateral (d), distal (e), and proximal (f) views. Scale bar equals 1 cm

FIGURE 7

Mid-caudal vertebra of Dineobellator notohesperus (SMP VP-2430) in lateroventral (a), ventral (b), right lateral (with ventral surface to top of page) (c), dorsal (d), left lateral (e), proximal (f), and distal (g, h) views. (c) Has the ventral portion toward the top of the page and the vertebra has been flipped with the distal centrum surface to the left. (g) and (h) are both distal views of the vertebra but using different lighting to highlight the impression on the distal (or caudal) centrum surface. Scale bar equals 1 cm

FIGURE 8

Caudal vertebrae of Dineobellator notohesperus (SMP VP-2430). (a–e) Caudal vertebra in left lateral (a), distal (b), right lateral (c), dorsal (d), and ventral (e) views. (f–j) Caudal vertebra in left lateral (f), right lateral (g), dorsal (h), proximal (i), and distal (j) views. (k–o) Distal caudal vertebra in proximal (i), dorsal (l), left lateral (m), ventral (n), and right lateral (o) views. Scale bar equals 1 cm

FIGURE 9

Distal caudal vertebrae and various vertebral fragments of Dineobellator notohesperus (SMP VP-2430). (a, b) Caudal vertebrae in left lateral (a) and ventral (b) views. (c–f) Prezygapophyseal fragments. Isolated right prezygapophysis in lateroventral (c) and mediodorsal (d) views. Isolated left prezygapophysis in dorsal (e) and medial (f) views. Scale bar equals 1 cm

Laminae are present on the lateral surfaces of the first caudal vertebra, particularly close to and dorsal to the transverse process, with more on the right side, probably due to it being the better preserved of the two sides. These include the anterior centrodiapophyseal lamina ventrally and the prezygadiapophyseal lamina dorsally (see Wilson, 1999). The ventral surface is mediolaterally flat but has a distinct craniocaudal (or proximodistal) curvature when viewed laterally. This curvature, with the distal portion of the centrum distinctly ventral to the proximal portion, is distinct among the caudal vertebrae of dromaeosaurids. This characteristic morphology is found in the cervicals of dromaeosaurids (e.g., Deinonychus), where it likely results in a sigmoidal curvature of the neck (see Ostrom, 1969). Additionally, while there is some curvature to the ventral surface of the centrum of the proximal caudal vertebrae in some dromaeosaurids, the proximal portion lies ventral to the distal portion (e.g., Velociraptor, IGM 100/985). We infer the tail of Dineobellator was, therefore, distinct, with the tail dipping down, at least slightly, at its proximal portion before probably becoming more straightened and rod-like, as in other known dromaeosaurids. Dorsal on the centrum, the transverse processes, particularly the more complete right side, project laterally more than dorsolaterally, are sub-rectangular, short, and gracile, and situated midway through to just caudal to the midpoint of the centrum. The proximal or medial portion of the right transverse process is well preserved but broken distally. The right transverse process is more gracile and does not fan out as much as those in the proximal caudals of Velociraptor (IGM 100/985). The neural canal is marked by a depression on the dorsal surface of the centrum ventral to the neural spine. A medial depression or ventral groove on the ventral surface of the centrum has a width of 5.1 mm. Craniolaterally, there is a shallow depression or fossa on the right transverse process, likely the prezygapophyseal centrodiapophyseal fossa (see Wilson et al., 2011), which becomes more inconspicuous laterally on the process. The spinoprezygapophyseal fossa lies dorsoventrally on the proximal surface of the neural spine. The neural spine is low, robust, and flares laterally toward its distal (posterior) portion. Caudally, there is a deep foramen just dorsal to the neural canal representing the spinopostzygapophyseal fossa.

Another nearly complete caudal vertebra is probably from midway through the caudal series (likely eighth through twelfth caudal vertebra; Figure 7). The centrum is complete, and a portion of the left neural arch is preserved, although none of the neural spine is preserved. The neural arch spans most of the proximodistal (craniocaudal) length of the centrum. The centrum underwent some shear deformation. It is platycoelous with well-defined and conspicuous circular to subcircular indentations on both the proximal and distal centrum surfaces (Figure 7f–h). These concavities are symmetrical, and both lie just dorsal to the center of the centrum on their respective surfaces. These subcircular indentations have not been seen in other dromaeosaurid caudal vertebrae, and their purpose is uncertain. Both the proximal and distal centrum faces are sub-rectangular to sub-trapezoidal, and the centrum is thinner midway between the two centrum faces.

Several other incomplete and fragmentary caudal vertebrae were collected but not reported by Jasinski et al. (2020) (Figure 8). These include a caudal vertebra that underwent taphonomic deformation via slight mediolateral compression (Figure 8a–e). The distal portion of the centrum is preserved while the proximal portion along with the neural arches and spine and processes are not. While incomplete, the vertebra is longer than the mid-caudal vertebra described above, and so is probably from more distal (posterior) in the caudal series. The distal end of the centrum is ventrally lower than the middle, which is common among dromaeosaurid caudal vertebrae. The distal surface is slightly concave, although most of this is not well-preserved due to deformation but probably was similar to the platycoelous condition of the more complete mid-caudal vertebra. Ventral ridges and a ventral groove are preserved and are particularly visible near the caudal end.

Another caudal vertebra also underwent significant mediolateral compression (Figure 8f–j), although it is more complete than the one described above. Ventral portions of the neural arches are preserved, with more of the left neural arch preserved, where it spans most of the length of the centrum. The neural arches, particularly the left neural arch, help distinguish part of the neural canal. The proximal surface of the centrum, while incomplete, preserves part of an indentation, like the more complete mid-caudal vertebra discussed above. This means the indentations on the centrum surfaces of these caudal vertebrae are not taphonomic features. The vertebra is platycoelous with the proximal and distal centrum surfaces gently concave. This caudal vertebra is also the longest (46.3 mm) of the preserved caudal vertebrae, although some of the others are incomplete.

Another incomplete caudal vertebra consists of most of the centrum, although the distal portion is missing (Figure 8k–o). While the dorsal surface of the vertebra preserves the broken surface where the neural arch would have been, the neural arch and spine are missing. The floor of the neural canal is present between the broken neural arch bases. Additionally, there is a small indentation in the concave proximal centrum surface, similar to those on several of the other caudal vertebrae. The shorter preserved length of the centrum (22.4 mm) means this vertebra was located farther distal in the caudal series where vertebra lengths decreased as the overall dimensions of the vertebral centra also become smaller.

A small section of at least four fused caudal centra is preserved with SMP VP-2430 (Figure 9a,b). The lateral surfaces of the vertebrae are generally flat to slightly convex. Lateral ridges represent the contact surfaces between adjacent caudals. Two of the vertebrae are complete with lengths of 4.1 and 5.1 mm, respectively. Thin longitudinal “lines” are present toward the dorsal half of the caudal section that may represent the zygapophyseal rods found in various dromaeosaurid caudal sections that help give the tail added strength. The vertebrae are quite small and are inferred to be from the more distal (or posterior) part of the tail. Although the fusion of distal caudal vertebrae can form a pygostyle in some theropods such as the oviraptorosaur Nomingia (e.g., Barsbold, Currie, et al., 2000; Barsbold, Osmólska, et al., 2000), the caudal fusion in Dineobellator may be pathologic as they do not resemble a pygostyle and the tail does not seem to be shortened compared to other eudromaeosaurs. We note, however, that the sizes of these fused caudal vertebrae are smaller than would be expected of the individual represented by SMP VP-2430 given the dimensions of the other elements, and it is possible they do not belong with SMP VP-2430. They do not possess any of the features used to diagnose Dineobellator notohesperus. Several additional small elements are identified as portions of vertebrae and agree with the known caudal vertebral morphology of other dromaeosaurids, including several isolated zygapophyseal fragments (Figure 9c–f).

2.5 Ribs

Several bone fragments are identified as parts of dorsal ribs, are fairly nondescript, and do not offer much morphological information. There is one more complete left dorsal rib that is missing its distal portion, exhibits some folding dorsally, and has a long, thin costal groove laterally (Figure 10). Proximal to the costal groove is a ridge that wraps toward the caudolateral surface and continues to the proximal edge. The proximal surface is abnormally expanded and has a “D”, or semi-circular, shape to it (Figure 10d,e). This abnormal surface incorporates both the capitulum and tuberculum, with both forming one irregular articular surface with the corresponding vertebra. Just ventrolateral to this proximal expansion is another expansion of the element, which wraps medially to the proximal edge. The rib preserves several areas of irregular morphology on its surface, with areas of slight expansion or depression along the rib shaft. This irregular morphology shows bone restructuring and is inferred to be pathologic in origin.

FIGURE 10

Incomplete pathologic left dorsal rib of Dineobellator notohesperus (SMP VP-2430) in caudal (a), medial/proximal (b), and cranial (c) views. (d, e) Oblique proximal view in different lighting conditions to highlight the abnormal bone growth on the proximal section of the rib. Scale bars for (a–c) and (d, e) equal 1 cm

2.6 Appendicular elements

The nearly complete right humerus (Figure 11) is missing parts of both the distal and proximal ends, including parts of the deltopectoral crest. The preserved portion of the humerus measures 18.6 cm, with an estimated total length of 21.5 cm. This is similar in size to Deinonychus (Table 3), although the humerus of Dineobellator is more gracile in proportion. This is particularly evident when comparing the mediolateral width of the humerus (diameter in cranial or caudal view) around its midlength, which is approximately 11 mm in Dineobellator and 21 mm in Deinonychus. The degree to which these widths differ may be exaggerated due to some taphonomic deformation in SMP VP-2430 mediolaterally, resulting in a somewhat oval cross-section, although this deformation does not account for the entirety of the differences (i.e., midshaft circumference) present and still results in Dineobellator being more gracile than Deinonychus. Taken along with the morphology and size of the ulna described below, Dineobellator had a relatively elongate humerus and forelimb. The relatively elongate humerus distinguishes Dineobellator from most other known eudromaeosaurs except Bambiraptor and Saurornitholestes, although an elongate humerus is also present in some microraptorines and unenlagiines. The proximal end of the right humerus of Dineobellator is thin, gracile, relatively flat, and is bent somewhat medially. The shape of the proximal end is distinct from the sigmoidal shape present in other dromaeosaurids (e.g., Bambiraptor, Deinonychus, Saurornitholestes).

FIGURE 11

Humerus in dromaeosaurids. (a, b) Nearly complete right humerus of Dineobellator notohesperus (SMP VP-2430) in lateral (a) and medial (b) views. (c) Outline drawing of the complete left humerus of Saurornitholestes langstoni (TMP 1988.121.0039), in medial view. Bone has been reflected for easier comparison with that of D. notohesperus. (d, e) Schematic drawing of the distal deltopectoral crest; (d) angles between the distal border of the deltopectoral crest and the long axis of the humerus in medial view, and (e) hypothetical orientation of the muscle bodies and fibers (e.g., m. brachialis) from its origin marked with an arrow. The arrows are perpendicular to the distal border of the deltopectoral crest. (f, g) Highlighting the distal angle of the deltopectoral crest in the dromaeosaurids Dineobellator notohesperus (SMP VP-2430) (f) and Saurornitholestes langstoni (TMP 1988.121.0039) (g) with the dashed arrows showing approximately equivalent points separating the deltopectoral crest from the humerus shaft, and solid arrows showing hypothetical angles of muscle bodies with the deltopectoral crest. Dn, Dineobellator notohesperus; Sl, Saurornitholestes langstoni. Scale bar for (a–c) equals 2 cm. (f–g) are not to scale

TABLE 3. Measurements of the humerus of several dromaeosaurid taxa focusing on the deltopectoral crest and its size relative to the full length of the humerus

Eudromaeosaur humerus (and deltopectoral crest) measurements
Taxon	Humerus total length	Deltopectoral crest length	Deltopectoral crest/humerus total length
Dineobellator notohesperus	21.5 cm^a	6.67 cm^a	31%^a
Deinonychus antirrhopus	20.4 cm	5.71 cm	28%
Saurornitholestes langstoni	17.7 cm	4.43 cm	25%
Dakotaraptor steini	31.8 cm^a	7.47 cm	23.5%
Bambiraptor feinbergori	10.5 cm	2.15 cm	20.5%

• Note: The measurements for Dakotaraptor steini are for the larger and more robust morph discussed by DePalma et al. (2015).
• a Estimated total values based on preserved portions.

The deltopectoral crest is thinner than the shaft, projects cranially, and lies closer to perpendicular to the long axis of the humeral head (68° in D. notohesperus versus 23° in S. langstoni; Figure 11d,f,g). The deltopectoral crest is relatively elongate at approximately 31% of the total length of the humerus, making it proportionally longer in Dineobellator than in several other dromaeosaurids with preserved humeri (e.g., Bambiraptor, Dakotaraptor, Deinonychus, Saurornitholestes, Table 3, Figure 12). It is unknown if the relative length of the deltopectoral crest changed through ontogeny. The holotype of Bambiraptor feinbergi (AMNH FARB 30556) likely represents a juvenile dromaeosaurid that may indicate a relatively smaller deltopectoral crest in juvenile dromaeosaurids, which would further indicate the holotype of Dineobellator (SMP VP-2430) was at a more advanced ontogenetic age at the time of its death. The distal edge of the deltopectoral crest creates a sharp angle with the shaft of the humerus. This sharp-angled curvature of the distal portion of the deltopectoral crest is unique among dromaeosaurids, although the gentle curvature of the humerus can become sharper farther away from the humeral shaft (e.g., Dakotaraptor, Deinonychus; Figure 12). Furthermore, the orientation and thin structure of the deltopectoral crest (in relation to the humeral shaft) also help distinguish Dineobellator from other dromaeosaurids, particularly from members of Unenlagiinae.

FIGURE 12

Comparison of eudromaeosaur humeri in medial view. Note that the farthest right humerus referred to Dakotaraptor steini by DePalma et al. (2015) includes proximal and distal reconstructed portions (see their figure 2). All humeri have been scaled to the same 5 cm scale bar

The muscle scar of the m. latissimus dorsi is present proximally on the lateral surface of the humerus. The medial depression near this muscle attachment is pronounced and continues distally from the proximal portion for approximately 1/3 of the humerus length. On the lateral surface, and cranial to the depression, lies a ridge that runs longitudinally down the bone. The raised ridge on the proximolateral surface of the bone is approximately 9.3 cm long. It terminates just beyond the distal extent of the deltopectoral crest. There is a slight bend midway along the humeral shaft that is inferred to be, at least partially, due to taphonomic deformation. Distal to the bend, the shaft is sub-circular to oval in cross-section. Distally, the humerus expands craniocaudally and there is a depression running longitudinally up the shaft on the medial surface. The humerus is also hollow in cross-section. There is slight longitudinal crushing of the shaft toward its middle, but this is undoubtedly due to the hollow nature of the bone.

The right ulna is a long, thin, bowed bone (Figure 13). The distal end is missing, but the proximal portion is complete. The preserved portion has a total proximodistal length of 10.1 cm, giving the bone a total estimated length of approximately 14.0 cm based on the ulnae of Deinonychus, Saurornitholestes, and Velociraptor. The ulna has a shallow trochlea and a diminutive, transversely broad, olecranon process. The process is nearly complete and sub-triangular. The ulna flares out mediolaterally just distal to the olecranon process and the sigmoid notch of the ulna. The shaft of the ulna is curved and flares becoming thin and wide distally. There is an uneven texture along the ventral ulnar ridge that bears at least six protuberances, identified as ulnar papillae or quill knobs (Figure 13d). This implies approximately 12–14 secondary remiges. These protuberances start approximately 4.7 mm from the proximal end. Several dromaeosaurid taxa are known to possess feathers, or feather-like structures, such as the Barremian–early Aptian Changyuraptor (Han et al., 2014), the Aptian Sinornithosaurus (Liu et al., 2004; Xu et al., 1999), Zhenyuanlong (Lü & Brusatte, 2015), Wulong (Poust et al., 2020), and the Albian Microraptor (Gong et al., 2012; Xu et al., 2000, 2003). Indirect evidence of feathers in several dromaeosaurids in the form of quill knobs (i.e., ulnar papillae) suggests the presence of feathers in the Campanian Asian velociraptorine Velociraptor mongoliensis (Turner, Makovicky, & Norell, 2007), the Maastrichtian North American dromaeosaurine Dakotaraptor (DePalma et al., 2015), and Dineobellator notohesperus (Jasinski et al., 2020). The 12–14 secondary feathers of D. notohesperus (Jasinski et al., 2020) is similar to that of V. mongoliensis (14 secondaries; Turner, Makovicky, & Norell, 2007) and between the estimates for the Maastrichtian Rahonavis (10 secondaries; Forster et al., 1998), the Tithonian Archaeopteryx (12 or more secondaries; Elzanowski, 2002), and the Albian Microraptor (18 secondaries; Xu et al., 2003).

FIGURE 13

Nearly complete right ulna of Dineobellator notohesperus (SMP VP-2430) in lateral (a), cranial (b), and medial (c) views. (d) Shows a closer image of right ulna in lateral view with feathers placed where ulnar papillae are located along the ulnar ridge; feathers used are from Megascops kennicottii. Scale bars for (a–c) and (d) equal 1 cm

The distal 2.2 cm of the right radius is preserved (Figure 14a,b). The element is craniocaudally longer than mediolaterally wide, at least at its distal extremity. It is wider cranially, becomes thinner caudally, and results in a triangular shape distally. This triangular distal surface is like that in Bambiraptor [compare with Burnham (2004, figure 3.20)], although the surface is thinner in Dineobellator making it more similar to that of the Asian Kuru kulla (Napoli et al., 2021).

FIGURE 14

Radius and metacarpal III of Dineobellator notohesperus (SMP VP-2430). (a, b) Distal portion of right radius in medial (a), and distal (b) views. (c–e) Proximal portion of right metacarpal III in medial (c) and lateral (d) views (with dorsal to top of page), and drawing of the element in proximal (e) view. Scale bars for (a, b) and (c, d) equals 1 cm while (e) is not to scale

The proximal end of the right metacarpal III is preserved, although the shaft has been partially crushed mediolaterally and thins mediolaterally away from the proximal end (Figure 14c–e). Proximally the bone is also concavely curved. On the distally preserved surface, the bone is wider (mediolaterally) dorsally than ventrally, giving the bone somewhat of a “P” shape. In Deinonychus, the ventral portion remains wider, resulting in a sub-trapezoidal shape, with only the ventral-most portion thinning out, while the tapering ventrally is more gradual in Dineobellator. There is a narrow groove on the lateral surface of the shaft, although this may be due to deformation. The entire bone is slightly curved with the proximal and distal ends being slightly more cranial than the rest of the shaft.

The nearly complete right manual ungual II is missing only the distal tip (Figure 15). It measures 4.6 cm long from the ventral edge of the articular surface to the preserved distal tip, with an estimated total length of approximately 5 cm when complete. The claw is strongly curved, with a pronounced flexor tubercle along its proximoventral edge. However, when viewed medially or laterally, the claw is prominently arched, with the dorsal surface oriented approximately 114° in relation to the articular surface. The articular surface preserves the medial and lateral cotyles, separated by a cotylar ridge. The lateral cotyle accounts for most of the mediolateral width of the articular surface, as is often present in dromaeosaurids. There is a lateral depression or groove that runs along its length toward the distal tip. This groove starts between the articular surface and flexor tubercle, and travels along the length of the claw toward the dorsal surface distally. Medially a similar groove is present, although there is a prominent foramen in the groove just distal to the articular surface. The two grooves are offset on both sides of the ungual as the groove on the medial surface does not get closer to the dorsal surface, unlike the lateral groove. The distinct offset nature of the longitudinal grooves (Figure 15f) is often present on pedal unguals of various dromaeosaurid taxa (e.g., Senter, 2007), and these are barely offset in the manual ungual of one other taxon (Boreonykus certekorum), but not in other dromaeosaurids. Other studies have called these grooves various terms including blood grooves (e.g., Brilhante et al., 2022), vascular grooves (e.g., Longrich & Currie, 2009), or corial grooves (e.g., DePalma et al., 2015). Near the proximal end of the groove on the medial surface is a distinct and prominent indentation or gouge mark (Figure 15g,h). This mark starts as a furrow proximoventrally and terminates in a prominent, but small, depression closer to the dorsal edge. The indentation has an approximate width of 3 mm and extends for a length of 9 mm. This feature does not show abnormal bone regrowth and is likely not due to infection or pathology on and under the keratinous sheath.

FIGURE 15

Right manual ungual II of Dineobellator notohesperus (SMP VP-2430) in medial (a), proximal (b), lateral (c), dorsal (d), and ventral (e) views. (f) Distal view showing the cross-section of the distal portion of the preserved claw, highlighting the offset of the medial and lateral grooves. (g, h) Magnified image (g) and drawing (h) of the proximal portion of the claw in medial view highlighting the relevant features, including the indentation and the pronounced extensor groove above the articular surface

The flexor tubercle is perpendicular to, and nearly vertical or angled almost straight down (ventrally) from, the articular surface of the ungual. It is also distinctly larger relative to the articular surface than those of other dromaeosaurids [flexor tubercle approximately 93% height of articular surface in Dineobellator, compared to those of Microraptor (56%), Bambiraptor (55%), Deinonychus (55%), Boreonykus (60%), and Velociraptor mongoliensis (77%)]. Immediately dorsal to the articular surface is a ridge running up the middle of the ungual, with a depression present on either side (medially and laterally). On its proximodorsal-most surface, there is a slight lip. This feature results in a distinct mediodorsal groove (= extensor groove) dorsal to the articular surface and is deeper and more prominent than in other dromaeosaurids. This surface is the attachment area for the m. extensor digitorum brevis, and the prominent and enlarged aspect of the area corresponds to the enlarged attachment surfaces for the digit extensors. Directly ventral to the articular surface is another slight lip or ridge that comes to two lateral points or projections. Ventrally, the ungual widens near the mid-length of the element and thins again as it progresses distally (Figure 15e).

An incomplete right ungual is tentatively identified as from manual digit III (Figure 16). The distal portion is preserved while the proximal portion, including the articular surface with the penultimate phalanx and the flexor tubercle, are missing. The ungual underwent slight dorsoventral deformation, making it less curved than it was in life. The medial and lateral grooves of the ungual are offset (Figure 16e,f), with the groove most well-preserved near its mid-length. There is some abnormal bone growth present ventral to the grooves, which is likely pathologic. The offset nature of the grooves is like that of the right manual ungual II (Figure 15f), which is partially why this ungual is tentatively identified as from manual digit III. The medial and lateral grooves appear to be on opposite sides because the former is viewed proximally, and the latter is viewed distally (see Figures 15f and 16f). We note that pedal unguals of eudromaeosaurs tend to have an accessory vascular groove medially, and this groove is most often visible proximoventrally on the medial surface (e.g., Brilhante et al., 2022; Longrich & Currie, 2009; Turner et al., 2021).

FIGURE 16

Right ?Manual ungual ?III of Dineobellator notohesperus (SMP VP-2430), distal portion, in lateral (a), medial (b), dorsal (c), ventral (d), and proximal (e, f) views. Panel (f) shows the proximal broken portion of the claw in proximal view, highlighting the offset medial and lateral grooves. Scale bar equals 1 cm for (a–e) while (f) is not to scale

A portion of the right pubic boot is preserved (Figure 17a,b). While incomplete, the preserved part shows the pubic boot had a semi-rounded and only slightly projecting cranial portion. The pubic boot projects caudally, although it is incomplete, making comparison with other eudromaeosaurs difficult. The medial sutural surface between the two pubes is wide and shows a sutural contact. There are some muscle attachments on the pubis, most notably the M. puboischiofemoralis internus and M. p. externus, although these tend to attach to the shaft of the pubis in dromaeosaurids (e.g., Rhodes et al., 2021).

FIGURE 17

Pubis and femur of Dineobellator notohesperus (SMP VP-2430). (a, b) Distal right pubic boot in lateral (a) and medial (b) views. (c–f) Incomplete right femur in caudal (c), medial (d), cranial (e), and lateral (f) views. The dashed line in (c) estimates the proximal end of the femur and, based on the distal preserved curvature of the femoral head, the femoral head orientation and shape. Scale bars equals 1 cm for (a, b) and (c–f)

The right femur is a relatively robust, incomplete element missing the majority of the femoral head and a significant part of the distal end (Figure 17c–f). The preserved portion of the element is 6.9 cm long, with a total estimated length of approximately 27.5 cm based on more complete femora of other dromaeosaurids such as Dakotaraptor, Saurornitholestes, and Velociraptor. Proximally, the preserved portion projects medially making up a small medioventral portion of the femoral head. This projection indicates the femoral head was directed dorsomedially. The shaft is oval in cross-section and flares craniocaudally. It is thinner than that of Deinonychus (OMNH 50268) and more similar to the dimensions of Velociraptor (e.g., IGM 100/986). A slight raised portion of the bone along the caudomedial margin is inferred to be part of the fourth trochanter.

Several elements of the foot were recovered (Figure 18), including several incomplete metatarsals. The right metatarsal II is represented by the proximal (Figure 18e–g) and distal (Figure 18h–j) portions, with the middle part of the shaft missing. The preserved shaft of metatarsal II is relatively thin and sub-circular to sub-rectangular in cross-section. Proximally, the bone is sub-triangular, with a distinct groove (or furrow) from the proximal edge that extends approximately halfway down the proximal section. The proximal portion of this furrow, which results in a lateral concavity of metatarsal II, corresponds with the contact surface for metatarsal I. While the proximal surface of metatarsal II is sub-trapezoidal in Deinonychus, the contact surface for metatarsal I is angled but flat rather than slightly concave in Dineobellator. There is a caudolateral depression near the distal end. The cranial portion of the distal end tapers cranially to a relatively sharp point. There is a depression on the dorsal surface of the proximal end that runs distally from between the two condyles. Another depression is present on the distomedial surface of the bone, lying just proximal to the left condyle. The distal portion of the right metatarsal II is incomplete due to deformation and erosion and is sub-rectangular to sub-triangular. Distally, the metatarsal flares out into two separated projections, or condyles. On the medial surface, just proximal to the distal end, lies a prominent foramen. On the lateral surface, approximately even with the foramen, lies a small, raised protuberance. There are no traces of a tuber where the extensor surface lies. The overall proportions imply a relatively short metatarsal, the common condition in eudromaeosaurs, rather than the relatively elongate metatarsals in unenlagiines (Gianechini et al., 2020).

FIGURE 18

Right pedal phalanx I–I and metatarsals of Dineobellator notohesperus (SMP VP-2430). (a–d) Right pedal phalanx I–I in medial (a), craniolateral (b), lateral (C), and distal (d) views. (e–g) Proximal portion of right metatarsal II in medial (e), lateral (f), and line drawing of proximal (g) views, with dorsal to top of page in (g). (h–j) Distal portion of right metatarsal II in cranial (h), caudal (i), and distal (j) views, with cranial to top of page in (j). (k, l) Proximal portion of right metatarsal III in medial (k), lateral (l), and line drawing of proximal (m) views, with dorsal to top of page in (m). (n) Outline drawing of proximal view of the articulation between metatarsals II–III in D. notohesperus. Scale bars for (a–d), (e–m), and (n) equal 1 cm

Right metatarsal III is represented by the proximal end, with the shaft thinner proximally and flaring out steadily toward the distal surface (Figure 18k–m). The dorsal edge, while having been crushed somewhat mediolaterally, is still wider than the ventral edge. The dorsal and ventral edges of the proximal surface also taper to a ridge that runs mediolaterally along the middle of this surface. The bone flares slightly toward the dorsolateral and ventromedial regions on the distal surface. Relative to metatarsal II, a slight concavity on the medial side of metatarsal III abuts with the former element, resulting in metatarsal II shifting slightly caudally (Figure 18n). This is also relative to the condition in Deinonychus, where the caudal portion of metatarsal III is not distinctly thinner than the cranial portion. Proximally metatarsals II and III articulate tightly (Figure 18n). Although metatarsal IV is not preserved, the curvature of metatarsal III suggests it also would have been situated slightly caudal in relation to the latter element, distinct from the more mediolateral (or coronal or frontal) orientation of these elements in Deinonychus. This provides a slightly crescent orientation among the metatarsals of Dineobellator.

The relatively small and nearly complete right pedal phalanx I-I is preserved (Figure 18a–d). The bone was identified by Jasinski et al. (2020) as right metatarsal I and was briefly described. The bone exhibits a slight twist about the shaft axis. There is also a relatively large foramen present toward its distal end, measuring 3.6 mm by 2.3 mm. Enclosing this foramen is a relatively pronounced and rounded rim. The outer surface near the proximal end is eroded and incomplete.

Two bone fragments are identified as portions of a pedal ungual, inferred to be from right digit III (Figure 19). A small, middle portion is missing, but the proximal edge and most of the distal portions are present. The proximal fragment measures 1.3 cm (proximodistally) and the distal fragment is 3.2 cm long, resulting in a total length of approximately 5.5 cm when complete. The dorsal portion of the proximal surface preserves the articular surface, including both the medial and lateral cotyles. The cotylar ridge is shifted slightly medially, making the lateral cotyle slightly wider than the medial cotyle, as is the common condition in dromaeosaurids. Ventral to the articular surface is the flexor tubercle, which is significantly smaller and less pronounced than that of the manual ungual II. Although smaller than in the manual ungual, the flexor tubercle of the pedal ungual is still enlarged relative to other dromaeosaurid taxa (e.g., Boreonykus, Dakotaraptor, Deinonychus), and is probably most similar in relative proportions to Utahraptor pedal unguals. While there is a slight concave curvature below the articular surface, the flexor tubercle again projects perpendicular to the articular surface. The distal portion of the pedal ungual is missing the tip. A set of grooves are present on both the lateral and medial surfaces of the claw and offset from each other, like those of the manual ungual. There is also a ridge dorsal to the lateral groove, and the groove proceeds toward the dorsal surface as it extends distally, while the medial groove or depression does not, as in manual ungual II. There is a second, less conspicuous groove (accessory vascular groove) ventral to the main one on both the medial and lateral surfaces (Figure 19d). Longrich and Currie (2009) mentioned an accessory vascular groove on the medial surface of pedal ungual II in eudromaeosaurs and figured this feature in Saurornitholestes langstoni (see their figure 2e), although this feature has also been noted in other studies (e.g., Brilhante et al., 2022). Its presence on the medial and lateral surfaces indicates it is a derived morphological feature based on our phylogenetic analysis. It is also noted that the curvature and asymmetry of the articular surface, along with the wider lateral cotyle compared to the medial one, suggest that the lateral surface has more well-defined grooves than the medial surface. The pedal ungual is relatively thinner (dorsoventrally) compared to the manual ungual and is not nearly as strongly curved.

FIGURE 19

Right pedal ungual III of Dineobellator notohesperus (SMP VP-2430) in medial (a), proximal (b), lateral (c), and cross-section outline drawing of ungual in distal (d) views. (d) Represents the cross-section of the claw from the broken distal end, highlighting the placement of the medial, lateral, and accessory vascular grooves. Scale bar equals 1 cm for (a–c) while (d) is not to scale

A small bone was previously identified as a possible left astragalus by Jasinski et al. (2020). The bone is incomplete and distinctly smaller than would be expected for a theropod the size of Dineobellator, at least that of the holotype (SMP VP-2430). This bone is not identifiable and is currently conservatively re-identified as an indeterminate element. Numerous other bone pieces are too fragmentary for identification.

3 DISCUSSION

3.1 Phylogenetic analysis

The phylogenetic analysis resulted in 43 most parsimonious trees, each with a tree length of 394 steps, a Consistency Index of 0.460, and a Retention Index of 0.508 (Figure 20). The 50% majority rule consensus tree is reported here as the strict consensus tree resulted in a large polytomy between most dromaeosaurid species (see our Supporting Information). Bootstrap resampling was also done using 100 replicates to compare the results more closely to those of Powers et al. (2022).

FIGURE 20

Time-calibrated phylogeny of dromaeosaurid relationships illustrating the major relationships within the clade including their paleobiogeography. 50% majority rule consensus phylogenetic tree resulting in 43 most parsimonious trees, each with a tree length of 394 steps, a consistency index of 0.460, and a retention index of 0.508. Bootstrap resampling values are provided for each clade below the nodes. Archaeopteryx lithographica is the outgroup. Temporal positions and biogeographic locations of dromaeosaurid taxa are provided. Silhouettes are taken from phylopic.org and are freely available for reuse under a public domain or creative commons license (www.phylopic.org). See Supporting Information File S1 material for more information about individual silhouettes. Credits for silhouettes and references for temporal positions and biogeographic locations provided in the Supporting Information

The analysis resulted in the recovery of Dineobellator notohesperus at an unresolved node outside the three main recognized dromaeosaurid clades (e.g., Dromaeosaurinae, Saurornitholestinae, Velociraptorinae). This is distinct from the findings of Jasinski et al. (2020), who recovered Dineobellator as a velociraptorine sister to Tsaagan mangas + Linheraptor exquisitus. The analysis also recovered Acheroraptor temertyorum as sister to Atrociraptor marshalli, similar to the results of Powers et al. (2022), but different from those of Jasinski et al. (2020), who also recovered it in Velociraptorinae as the basal-most member.

Comparatively, the position of several taxa other than Dineobellator change between the studies, particularly among North American taxa. While Dineobellator and Acheroraptor are not found within Velociraptorinae in the current phylogenetic analysis, others such as Deinonychus (within Saurornitholestinae) and Utahraptor (within Dromaeosaurinae) are now explicitly recovered within previously known clades. As character states and scores were not altered, it is not surprising that the results of the current analyses and those of Powers et al. (2022) largely agree. The placement of Dineobellator as a potentially more basal eudromaeosaur outside of previously known and named clades is probably due to a lack of scorable characters rather than its proper phylogenetic placement. More material is needed to gain a better understanding of its evolutionary relationships, or more material is needed from other taxa to properly compare the recovered elements with those of other taxa.

The present results and those of Powers et al. (2022) largely follow biogeographic regions. This led the latter to conclude that similarities among the eudromaeosaurs are due more to convergence between clades than faunal exchange between North America and Asia. While this may be the case, the analysis of Powers et al. (2022) did not include several taxa that Jasinski et al. (2020) did, including the eudromaeosaurs Dakotaraptor steini, Boreonykus certekorum, and the Baynshire Formation dromaeosaurid, in addition to several species of unenlagiines, microraptorines, and some basal dromaeosaurid species. Some of these taxa have fewer scorable characters than others but their inclusion or exclusion could affect the results. Additionally, while the current results and those of Powers et al. (2022) largely follow biogeographic regions, the recovery of Achillobator giganticus as a dromaeosaurine with Utahraptor and Dromaeosaurus is contrary to this pattern. Jasinski et al. (2020) recovered the former as sister to Utahraptor in an unnamed sister-clade to Velociraptorinae. Its recovery here as sister to Dromaeosaurus means there is still more to learn about the evolution of eudromaeosaurs. While convergence may be part of the evolutionary story of Eudromaeosauria, it is likely not the whole reason behind the resulting relationships. As suggested by Jasinski et al. (2020), there may still be some vicariance between biogeographical regions. There is a clear connection between many of the dinosaurs (e.g., ankylosaurids, hadrosaurids, pachycephalosaurids, ornithomimids, tyrannosaurids) of Asia and North America during the Late Cretaceous (e.g., Arbour et al., 2014; Brusatte and Carr, 2016; Dalman et al., 2017; Jasinski & Sullivan, 2011, 2016; Lucas et al., 2016; McFeeters et al., 2017; Prieto-Márquez, 2010; Sullivan, 1999, Sullivan & Lucas, 2003, Sullivan & Lucas, 2006) and how these came about, including possibilities for migration patterns, vicariance, and/or convergence, still need to be investigated.

3.2 Behavioral implications

Several of the features of Dineobellator have possible paleobiological implications. However, hypothesizing paleobiological aspects of extinct animals is inherently difficult, particularly when there are no closely related animals alive today to use as analogs. Dromaeosaurids (particularly Deinonychus) have been interpreted as group predators or pack animals (e.g., Maxwell & Ostrom, 1995; Ostrom, 1994), although others do not necessarily agree with this interpretation (e.g., Brinkman et al., 1998). In drawing analogies between modern raptorial birds and dromaeosaurids, Fowler et al. (2011) hypothesized the pedal unguals of dromaeosaurids were used during predation and prey capture to hold down prey of smaller body sizes in what has been called the prey immobilization behavior model. Still, others have suggested dromaeosaurids focused on larger prey, at least for eudromaeosaurs (e.g., Gianechini et al., 2020). Some studies suggest that dromaeosaurids were arboreal due to pelvic and manual anatomy (e.g., Chatterjee, 1997), while still others have proposed the claws were used for climbing (Manning et al., 2009). Regardless, while their potential paleoecology has been investigated, morphological differences between taxa imply at least some differences in behavior and/or ecological niches.

Several features of the humerus can be used to infer possible behavioral differences between Dineobellator and other dromaeosaurids (Figures 10 and 12). The deltopectoral crest of the humerus is the attachment site for several muscles in the forelimb, including the m. brachialis, which originates on its distal edge (e.g., Burch, 2014, 2017). This muscle aids in the flexion of the forearm. As mentioned by Jasinski et al. (2020), enlarging the distal portion of the deltopectoral crest also enlarges the origin of the m. brachialis. Changing the angle of the distal portion of the deltopectoral crest may cause a shift of the origin of the m. brachialis, subsequentially resulting in the orientation of the muscle adhering more closely (i.e., more parallel) to the long axis of the radius and ulna, in addition to the humerus (Figure 12e–g). This change in angle can clearly be seen when comparing the distal edge of the deltopectoral crest with the long axis of the humerus (68° in D. notohesperus vs. 23° in S. langstoni, Figure 12d). Furthermore, this orientation may allow the use of lower muscle forces for flexion of the forearm. Utilizing the full space available for muscle origination on a larger deltopectoral crest would allow similar or larger muscle bodies that could provide greater strength capabilities of forearm flexion and greater mechanical advantage, particularly compared with other dromaeosaurids (Figure 12). While the humeri of Dineobellator and Deinonychus are of similar lengths (Table 3), that of Dineobellator is more gracile than Deinonychus, particularly evidenced in the mediolateral width of the midlength of the shaft being nearly twice as wide in the latter, suggesting further differences in the forearm usage and capabilities between these two taxa. The morphology of the humerus of Dakotaraptor is highly distinct, particularly as it is wider and more robust just distal to the deltopectoral crest. This difference in humerus morphology between it and other eudromaeosaurs also implies further distinctions in its forearm function and usage in comparison. The proximal expansion of the humeral shaft near the level of the deltopectoral crest is more like some other theropods such as therizinosauroids (e.g., Hedrick et al., 2015; Li et al., 2007) and allosauroids (e.g., Cuesta et al., 2018; Currie & Carpenter, 2000; Madsen, 1976).

The forelimb claws of some dromaeosaurids have been hypothesized to be important predatory weapons, notably as tools to dismember prey (e.g., Fowler et al., 2011). The more conspicuous extensor groove and relatively larger flexor tubercle of the manual ungual (Figure 15) results in corresponding differences in the potential uses of this element by Dineobellator. The enlarged extensor groove implies larger digital extensors (m. extensor digitorum brevis) in the manus. Whether this resulted in faster or stronger movements in the extension of the manus is unknown, although it is possible it allowed simply for more extension of the ungual and within the digit. The changes in digital extension could have also been counteracted by tighter grip strength within the ungual and digit. This is further implied by the enlarged flexor tubercle of the manual ungual relative to other dromaeosaurids (flexor tubercle approximately 93% height of articular surface in Dineobellator), including those of Microraptor (56%), Bambiraptor (55%), Deinonychus (55%), Boreonykus (60%), and Velociraptor mongoliensis (77%). In addition to an enlarged flexor tubercle implying enlarged flexor tendons, shortened metacarpals and metatarsals would increase the mechanical advantage (and grip strength; see Fowler et al., 2011; Gianechini et al., 2020).

In the largest measured grip strengths in extant Aves, Bubo virginianus, weighing around 1.4 kg, produced force values of 300 N (Newtons; Ward et al., 2002). While there is no special relationship with Bubo, this provides a way to approximate potential grip strength in these theropods. Campione et al. (2014) estimated body masses for Velociraptor mongoliensis of around 20 kg and Deinonychus antirrhopus of up to around 100 kg. While Dineobellator notohesperus was potentially of similar body size (i.e., length and/or height) to Deinonychus, it was more gracile and, therefore, likely had a lower body mass. Regardless, taking the two values to provide a range and using B. virginianus to calculate potentially maximum or higher-end theropod grip strengths in relation to extant birds, we get grip strength ranges of 4.28–21.4 kPa if the relationship was isometric. While the relationship is almost certainly not isometric, this still provides a generalized range that may account for Dineobellator if it was similar to the abilities of the more extreme living avian examples. Further restraints are present on the higher potential values, such as the failure strength and flexibility and strength of the living tissues (e.g., Manning et al., 2009; Patek & Caldwell, 2005; Walilko, Viano, & Bir, 2005), so the overall strength must be made relative to other comparable taxa for relative values. However, some of these elements, such as metacarpals, metatarsals, and soft-tissues, are incomplete or unknown in Dineobellator, and cannot add support for or against this hypothesis.

The pedal ungual of Dineobellator shows a similar condition with a pronounced flexor tubercle (Figure 16), although not to the same degree as that of the manual ungual. The pedal claws of dromaeosaurids have been hypothesized to be capable of serving both defensive and predatory functions (e.g., Adams, 1987; Colbert & Russell, 1969; Gianechini et al., 2020; Manning et al., 2006, 2009; Ostrom, 1969, 1990; Parsons & Parsons, 2009; Senter, 2009). The flexor tubercle in the pedal ungual of Dineobellator is approximately 67% the height of the articular surface. This is more than several other eudromaeosaurs with known pedal unguals (e.g., 50% in Dakotaraptor, 40% in Utahraptor, 36% in Deinonychus, 30% in Dromaeosaurus, 22% in Boreonykus, 20% in Velociraptor mongoliensis, and 17% in Bambiraptor) and implies greater flexion in the pedal digits of Dineobellator.

The possession of opisthocoelous proximal caudal vertebrae in Dineobellator (Figure 6) implies a difference in local joint behavior and, in turn, global mobility of the tail. Dromaeosaurids have zygapophyseal rods that would have aided in keeping the tail stiff and mainly parallel to the long axis of the body, increasing overall rigidity of the tail (e.g., Holtz Jr., 2003; Norell & Makovicky, 2004; Ostrom, 1969). While a stiffened tail could have aided in maintaining balance, as mentioned by Ostrom (1969), its ability to help with agility in these dinosaurs may have been overstated. Increased agility refers to a greater ability to apply acceleration in different (or changing) directions rapidly. Increased agility was hypothesized by Ostrom (1969) in Deinonychus. Increased agility was also inferred in the dromaeosaurid Velociraptor mongoliensis due to increased floccular lobes in the endocranium by King et al. (2020). This structure is used to maintain head and eye stability during movement and has been linked to agility by others (e.g., Witmer & Ridgely, 2009). Cheetahs and other tetrapods use their tails for maneuvering, balancing, and reorienting their center of mass during dynamic behaviors (e.g., Libby et al., 2012; Patel et al., 2016). Given that much of the center of mass for the tail is at the base, changes to the range of motion of the proximal (or anterior) caudal vertebrae may induce more dramatic changes in these righting behaviors. Carrier et al. (2001) discussed some theropod dinosaurs having potentially high rotational inertia, and corresponding lower agility, due to longer and rigid tails, keeping more of their body mass away from the trunk of their body. However, increased mobility at the base of the tail would allow Dineobellator to move the entire tail more easily, potentially bringing more mass nearer the body center if the tail was angled toward the body, thereby decreasing rotational inertia and increasing its agility. Opisthocoelous proximal caudal vertebrae could have resulted in more tail mobility (Figure 21), as compared to platycoelous caudal vertebrae (or even amphiplatyan vertebrae), even if that angular increase was only based on the ability of the first caudal vertebra to move mediolaterally compared to the caudal end of the sacrum. Increased range of movement near the base while keeping the rest of the tail stiff could also allow it to act as a rudder (center of mass reorientation) or counterweight (balance; e.g., Eaton, 1972; Patel et al., 2016; Patel & Braae, 2014; Thompson, 1998). However, without the preservation of caudal zygapophyses or neural arches, it is impossible to estimate muscle forces or global range of motion for the spinal column. Despite its only definitive presence at the proximal (or cranial) end of the tail, the opisthocoelous morphology may have increased the agility of Dineobellator and thus may have implications for its predatory behavior, particularly with respect to the pursuit of prey. However, understanding changes in agility does not add clarity to the reconstruction of the hunting style of Dineobellator; including whether Dineobellator pursued small, quick prey solitarily or was a social predator hunting in packs and pursuing larger prey, or a combination of styles. Additionally, as noted above, the caudal vertebral morphology of dromaeosaurids seems to vary more than previously noted, inferred here to mean there is more variation in their behavior than has also been previously noted.

FIGURE 21

Difference in movement of the tail of Dineobellator notohesperus based on that movement taking place near the base of the tail versus smaller angular increments throughout the tail (in dorsal view). Panels (a) and (b) shows movement of the tail when the majority of the angular shift is proximally at the base of the tail and the rest of the tail can be held more stiffly and act as a rudder and counterbalance. Panel (a) shows movement at 5° from the sagittal plane, while (b) is 15° from the sagittal plane. Panel (c) shows 15° offset from the sagittal plane, but is shown as being gentler, with slight turning within all the vertebrae, causing a curvature of the entire tail and less ability to act as a rudder for agile movements and quick changes in direction

The indentation on the manual ungual (Figure 15a,g,h) is only present on the medial side of the element, with no mediolateral crushing or deformation present throughout the rest of the element. We infer the indentation is not due to postmortem deformation. There is no evidence of remodeling or retexturing of the bone, which would imply the feature was due to an infection or disease causing a pathology (e.g., Gutherz et al., 2020; Hanna, 2002; Sullivan et al., 2011e). However, there is evidence of pathologies occurring under the keratinous scutes of some turtles due to bacterial or fungal growth (e.g., Garner et al., 1997; Hutchison & Frye, 2001; Rothschild et al., 2013; Scheyer & Sánchez-Villagra, 2007). While this occurs under a keratinous tissue in turtles and, therefore, could also occur under the keratinous claw sheath, it still leads to abnormal growth and remodeling of the bone in the former condition. The absence of remodeling or retexturing implies external trauma caused by these features at or near the time of death. The size and morphology of the indentation is approximately similar to the morphology one would get from the impression of the distal claw of a dromaeosaurid of similar size to Dineobellator (Jasinski et al., 2020). While it is possible the feature was made through an altercation, the force necessary to create this is probably greater than would be possible in their manual grip strength. Indeed, cortical bone can undergo compressional stresses of over 190 MPa (megapascals) longitudinally and over 130 MPa transversely (e.g., Hart et al., 2017), converting to forces of 19.37 MN (meganewtons) and 13.25 MN/m², respectively. However, fracturing from tensile loading can occur with 2–6 MPa/m (e.g., Morgan et al., 2018), or 203.94 kN and 611.83 kN/m². As mentioned above, even estimating grip strength forces using measurements from the strongest extant bird (Bubo virginianus) results in grip strengths of 4.28–21.4 N for D. notohesperus, well below the force necessary to cause transverse compressional deformation in cortical bone. Higher forces can be generated through impacts or strikes. While normally generated forces would not be large enough to cause bone deformation, there are many other factors affecting potential damage to bone, including angle of the force applied, age of the individual, density of the bone, and type of bone, among others (e.g., Morgan et al., 2018). Bite force estimates for Deinonychus antirrhopus [between 4,100 and 8,200 N by Gignac et al. (2010)] are higher, allowing for fracture and breakage of bone, but still well below the thresholds considered for deformation of bone. It is clear, though, that forces placed on bone, particularly concentrated forces focused on a single point, can more often lead to fractures (e.g., Karr & Outram, 2012). Permanent deformation (plasticity) in bone is difficult without inducing fracture as the deformation of protein molecules, fibrils, and fibers is usually reversible when applied forces are removed (e.g., Launey et al., 2010). Due to the concentration of forces and the higher forces possible from impact strikes, it is possible this feature is from an impact from the claw of a different individual. However, it is also still highly possible this feature is due to some infection or other abnormality occurring underneath the keratinous claw sheath, but one that did not result in remodeling and retexturing of the bone.

Conversely, the more complete rib of Dineobellator shows evidence of abnormal growth and remodeling, implying a break of the rib that subsequently healed (Figure 10). Other elements of Dineobellator also indicate abnormal bone growth, including the right premaxillary fragment (Figure 3a,b), distal (posterior) caudal vertebrae (Figure 9a,b), right ?manual ungual ?III (Figure 16), and potentially the right manual ungual II (Figure 15). Healed injuries resulting in pathologies are well-known in the fossil record (e.g., Dalman & Lucas, 2018; Robinson et al., 2015; Sullivan et al., 2000, 2011e; Tanke & Currie, 1998). More data are necessary, including further study of the internal structure of the rib, to more definitively determine whether it does represent a pathology and, if so, what kind.

The presence of ulnar papillae on the ulnar ridge of the ulna of Dineobellator is evidence that it had winged forelimbs (Figure 13d). While there have been suggestions of the winged forelimbs being used for stabilization during predatory attack in dromaeosaurids (Fowler et al., 2011), this would have been less important for larger-bodied taxa such as Dakotaraptor, unless larger-bodied taxa focused on prey of similar proportions relative to their own. It has been shown that coloration and patterns highly discernible within taxa may not have the same effect on prey (e.g., Outamuro et al., 2017). This implies that feathers can act as bright markers, species-recognition markers, and/or sexual display features without being visual signals that call attention of predators or prey. Modern raptorial birds show that color patterns can still be intricate and serve to both camouflage the predator and be part of the sexual selection process (e.g., Hill & McGraw, 2006; Holt et al., 1990; Snyder & Snyder, 2006), and similar feather styles may have been present in dromaeosaurids, including Dineobellator.

3.3 San Juan Basin Late Cretaceous vertebrate diversity

The known diversity of Late Cretaceous vertebrates from the San Juan Basin of New Mexico has undergone significant revision, with taxon count increasing over approximately the last two decades. This increase in known diversity has been driven by continued collecting in the Upper Cretaceous strata of the region. Newly named species from the Fruitland, Kirtland, and Ojo Alamo formations include the nanhsiungchelyid turtle Basilemys gaffneyi (Sullivan et al., 2013); the? azhdarchid pterosaur Navajodactylus boerei (Sullivan and Fowler, 2011); and several dinosaurs including the tyrannosaurid Bistahieversor sealeyi (Carr & Williamson, 2010); the dromaeosaurids Saurornitholestes sullivani (Jasinski, 2015a), Dineobellator notohesperus (Jasinski et al., 2020), and the proposed dromaeosaurid “Saurornitholestes” robustus (Sullivan, 2006; now inferred to be a troodontid by Evans et al., 2014); the caenagnathid Ojoraptorsaurus boerei (Sullivan, Jasinski, & van Tomme, 2011); the saurolophine hadrosaurids Anasazisaurus horneri and Naashoibitosaurus ostromi (Hunt & Lucas, 1993); the ankylosaurids Nodocephalosaurus kirtlandensis (Sullivan, 1999), Ahshislepelta minor (Burns and Sullivan, 2011), and Ziapelta sanjuanensis (Arbour et al., 2014); the nodosaurid Glyptodontopelta mimus [Ford, 2000; Edmontonia australis was also named by Ford (2000) but later found to be a junior synonym of G. mimus by Burns, 2008]; the ceratopsids Ojoceratops fowleri (Sullivan & Lucas, 2010), Navajoceratops sullivani and Terminocavus sealeyi (Fowler & Freedman Fowler, 2020), Titanoceratops ouranos (Longrich, 2011), and Bisticeratops froeseorum (Dalman, Jasinski, & Lucas, 2022); and the pachycephalosaurids Sphaerotholus goodwini (Williamson and Carr, 2002) and Stegoceras novomexicanum (Jasinski & Sullivan, 2011, 2016). Newly named genera for existing species include Denazinemys for the baenid turtle D. nodosa (Lucas and Sullivan, 2006); Scabremys for the baenid turtle S. ornata (Sullivan et al., 2013); and Denazinosuchus for the goniopholidid mesoeucrocodylian D. kirtlandicus (Lucas and Sullivan, 2003).

Additional material, including new specimens of already known taxa, have been reported from the Late Cretaceous of New Mexico by numerous workers. These studies span the early Campanian through the end of the Cretaceous, and provide a clearer picture of this region leading up to (and after) the end-Cretaceous mass extinction (e.g., Dalman et al., 2021; Dalman, Jasinski, & Lucas, 2022; Dalman & Lucas, 2018; Dalman, Lucas, et al., 2022; Gates, Evans, & Sertich, 2021; Kirkland & Wolfe, 2001; Lichtig & Lucas, 2016; Lucas, Spielmann, Sullivan, & Lewis, 2006; McDonald et al., 2006, 2010, 2018, 2021; Sullivan, Lucas, Jasinski, & Tanke, 2011; McDonald & Wolfe, 2018; Nesbitt et al., 2019; Sullivan & Lucas, 2015; Williamson, 1997; Wolfe & Kirkland, 1998). Fossils from the rocks of the Upper Cretaceous Fruitland Formation through Naashoibito Member (Ojo Alamo Formation) evince a world teeming with life during the time of the retreating Western Interior Seaway (Figure 1). The Ojo Alamo Formation, as presently accepted, straddles the Cretaceous–Paleocene boundary, with the Kimbeto Member considered Paleocene age (see Jasinski, Sullivan, & Lucas, 2011), although these younger strata are not believed to contain in-situ dinosaur fossils (Lucas et al., 2009; Koenig et al., 2012; contra Fassett & Lucas, 2000; Fassett et al., 2002, 2011; Fassett, 2009). It is noted that some have considered the Naashoibito Member as the uppermost member of the Kirtland Formation (e.g., Baltz et al., 1996; Carr & Williamson, 2000; D'Emic et al., 2011; Farke & Williamson, 2006; Flynn et al., 2020; Flynn, 1986; Hobbs & Fawcett, 2021; Hunt & Lucas, 1992; Lucas & Sullivan, 2000; Mason et al., 2013; Sullivan & Lucas, 2000; Sullivan & Williamson, 1999; Williamson, 1996; Williamson & Brusatte, 2014; Williamson & Weil, 2008a, 2008b). Conversely, others have maintained the Naashoibito Member as a member of the Ojo Alamo Formation (e.g., Bauer, 1916; Fassett, 2009; Fassett et al., 2002; Jasinski et al., 2015, 2018, 2020; Jasinski & Dodson, 2015; Jasinski, Lucas, & Moscato, 2011; Jasinski & Sullivan, 2016; Jasinski, Sullivan, & Lucas, 2011; Lucas et al., 2009; Powell, 1973; Reeside, 1924; Sullivan et al., 2013; Sullivan, Boere, & Lucas, 2005; Sullivan & Jasinski, 2012; Sullivan, Jasinski, Guenther, & Lucas, 2011; Sullivan, Jasinski, & van Tomme, 2011; Sullivan & Lucas, 2003, 2006; Sullivan & Lucas, 2010, 2014, 2015; Sullivan, Lucas, & Braman, 2005). Regardless, most of the recently named taxa have come from the Upper Campanian Fruitland and Kirtland formations, as these are more richly fossiliferous than the younger Maastrichtian Naashoibito Member of the Ojo Alamo Formation (e.g., Jasinski et al., 2020; Jasinski, Sullivan, & Lucas, 2011; Sullivan & Lucas, 2015).

Dromaeosaurid material has been reported from various Upper Cretaceous strata in the San Juan Basin. Armstrong-Ziegler (1978, 1980) was the first to report material potentially referable to Dromaeosauridae from the San Juan Basin, although it was identified as an indeterminate ?dromaeosaurid from the Upper Campanian Fruitland Formation. Lucas et al. (1987) and Hunt and Lucas (1993) listed the presence of indeterminate dromaeosaurids in the Fruitland, Kirtland, and Ojo Alamo formations, although this material is made up almost exclusively of isolated teeth. Sullivan and Lucas (2000) described the first potentially diagnostic material, an isolated left frontal (SMP VP-1270) from the underlying Upper Campanian De-na-zin Member (Kirtland Formation), which they attributed to Saurornitholestes langstoni. This specimen was later re-studied and determined to represent a distinct dromaeosaurid, which Jasinski (2015a) named S. sullivani. It is noted, however, that S. sullivani was referred to Saurornitholestes largely based on the general triangular morphology of the frontal (in dorsal aspect) similar to that of S. langstoni (see Jasinski, 2015a; Sullivan & Lucas, 2000). However, there are several differences between S. sullivani and S. langstoni, similar in number to those with other known dromaeosaurids, and it may be that S. sullivani may not represent Saurornitholestes and instead may be referable to another dromaeosaurid genus or represent a distinct genus. Sullivan (2006a) later described another isolated left frontal from the De-na-zin Member which he referred to a new species, S. robustus. However, this was restudied by Evans et al. (2014), who determined it represented an indeterminate troodontid. Additionally, isolated teeth (SMP VP-1901) from the De-na-zin Member (Kirtland Formation) differ from those of Dineobellator. They are gently curved, have slightly apically hooked denticles, less dense denticles (14–15 distal denticles per 5 mm; ~18–21 mesial denticles per 5 mm), and possess mesial denticles. The most complete tooth in SMP VP-1901 has an apical length of 14.9 mm, crown height of 12.6 mm, fore-aft basal length of 6.8 mm, and basal width of 3.4 mm. While these teeth may be from S. sullivani, they are distinct from those of Dineobellator. Furthermore, Williamson and Brusatte (2014) identify teeth as their “Dromaeosauridae Morphotype A” from the Santonian Point Lookout Sandstone, early Campanian Menefee, late Campanian Fruitland and Kirtland, and Maastrichtian Ojo Alamo formations (Naashoibito Member). They consider this morphotype to be very similar to, if not the same as, the generalized “saurornitholestine-type” tooth identified from farther north. The apical-hook of the denticles common in this morphotype is distinct from Dineobellator, and while they can have relatively large denticles, usually they are smaller (~22–25 denticles per 5 mm). Even so, SMP VP-1901 likely represents Dromaeosauridae Morphotype A, and a saurornitholestine-like dromaeosaurid, making its identity as S. sullivani as a more distinct possibility. “Dromaeosauridae Morphotype B” of Williamson and Brusatte (2014) is represented by a single tooth (NMMNH P-33148) from the Hunter Wash Member of the Kirtland Formation, although an incomplete tooth (NMMNH P-30225) was mentioned as potentially also being representative. Williamson and Brusatte (2014) inferred the tooth to represent a Dromaeosaurus-like dromaeosaurine. Distinct from Dineobellator, these teeth are noted for having mesial and distal denticles with an asymmetrical, D-shaped cross-section, and usually have larger and less numerous denticles (~15–18 denticles per 5 mm). This implies more than one dromaeosaurid species is present in these strata and likely lived in the region at the same time.

Morphological variation in the fossil record is important to consider but difficult to fully understand. This variation is often thought of in terms of inter- versus intraspecific variation (e.g., Griffin & Nesbitt, 2016; Plavcan & Cope, 2001). However, this variation can be covered by ontogenetic variation (e.g., Farke et al., 2013; Hone et al., 2016; Lichtig et al., 2021) and sexual dimorphism (e.g., Mallon, 2017), and these must be particularly considered with the latter. Additionally, some of this variation is particularly hard to definitively identify in the fossil record, especially sexual dimorphism (e.g., Mallon, 2017). Many studies have been conducted on morphological variation, particularly within fossil taxa (e.g., Arbour et al., 2016; Bell, 2011; Burns et al., 2015; Carter et al., 2021; Currie, 2003a, 2003b; Dalman et al., 2017, 2021; Dalman, Jasinski, & Lucas, 2022; Dalman, Lucas, et al., 2022; Delcourt & Iori, 2018; Dodson, 1976; Evans et al., 2013, 2014; Fabrezi et al., 2017; Gee & Jasinski, 2021; Grillo & Delcourt, 2017; Jasinski, 2011, 2013, 2015b, 2018; Jasinski et al., 2018, 2022; Jasinski & Moscato, 2014, 2017; Jasinski & Wallace, 2014, 2015; Ji et al., 2011; Johnson, 2020; Johnson et al., 2021; Lacovara et al., 2014; Lehman, 1987; 2001; Longrich, 2014; Lucas et al., 2011, 2016; Machado et al., 2013; Moscato & Jasinski, 2016; Osborn, 1923; Rivera-Sylva et al., 2012; Rowe, Colbert, & Nations, 1981; Sampson et al., 2010; Sullivan et al., 2013; Sullivan & Jasinski, 2012; Sullivan, Jasinski, Guenther, & Lucas, 2011; Sullivan, Lucas, & Jasinski, 2011c, 2011d; Vamberger et al., 2020; Voris et al., 2019). While the only definitive specimen of Dineobellator notohesperus is the holotype (SMP VP-2430), other specimens from the Naashoibito Member argue for the presence of more than one dromaeosaurid taxon in this stratigraphic unit. Although this hypothesis is reliant on tooth variation, the amount of variation in tooth morphology necessary to determine the difference between intra- and interspecific variation is difficult to discern and undoubtedly varies between taxa (including interspecific variation and higher taxonomic levels). Other strata (e.g., Hell Creek, Dinosaur Park, Two Medicine, Judith River, Oldman, and Djadochta formations) are already known to possess more than one dromaeosaurid taxon (see Holtz, 2021, table 1), so the presence of more than one dromaeosaurid species in the Naashoibito Member is not unexpected.

3.4 Naashoibito community

The vertebrate fauna of the Naashoibito Member was originally discussed by Lehman (1981) but was most recently reviewed in full detail by Jasinski, Sullivan, and Lucas (2011). While the latter study reviewed previously known material and reported new fossil material from the strata, little has been done since that time. Jasinski et al. (2020) fully described Dineobellator notohesperus from the strata, although the presence of a dromaeosaurid was previously known. A more recent addition to the known taxa known is a Torosaurus-like ceratopsid that has yet to be described. The fauna itself is less complete and well-known than in the underlying strata, but nevertheless includes at least 32 distinct vertebrate taxa, including chondrichthyans, actinopterygians, amphibians, turtles, crocodylians, mammals, and various dinosaurs including theropods, sauropods, ankylosaurians, hadrosaurids, and ceratopsids (Figure 22a, Table 4).

FIGURE 22

Vertebrate fauna of the Naashoibito Member during the Maastrichtian (Late Cretaceous) in New Mexico. (A) Makeup of the Naashoibito vertebrate ecosystem based on recovered fossils. (B) Theropod fauna of the Naashoibito Member scaled to estimated body sizes. Richardoestesia is not identified to species. Silhouettes are mostly taken from phylopic.org and are freely available for reuse under a public domain or creative commons license (www.phylopic.org). See Supporting Information for more information about individual silhouettes

TABLE 4. Vertebrate fauna of the Naashoibito Member (Ojo Alamo Formation) from the Maastrichtian of New Mexico

Ojo Alamo Formation (Naashoibito Member) fauna (Alamo Wash local fauna)
Major clade	Subclade	Genus	Species	References
Chondrichthyes	Anacoracidae	Myledaphus	Myledaphus sp.	Jasinski, Sullivan, and Lucas (2011)
Chondrichthyes	Orectolobidae	?Squatirhina	?Squatirhina sp.	Jasinski, Sullivan, and Lucas (2011)
Actinopterygii	Lepisosteidae	Lepisosteidae indeterminate	Lepisosteidae indeterminate	Jasinski, Sullivan, and Lucas (2011)
Amphibia	?Batrachosauroididae	?Batrachosauroididae indeterminate	?Batrachosauroididae indeterminate	Jasinski, Sullivan, and Lucas (2011)
Testudines	Pleurosternidae	Compsemys	Compsemys sp.	Jasinski, Sullivan, and Lucas (2011)
Testudines	Baenidae	Baenidae indeterminate	Baenidae indeterminate	Jasinski, Sullivan, and Lucas (2011)
Testudines	?Kinosternidae	Hoplochelys	Hoplochelys sp.	Jasinski, Sullivan, and Lucas (2011)
Testudines	Adocidae	Adocidae indeterminate	Adocidae indeterminate	Jasinski, Sullivan, and Lucas (2011)
Testudines	Nanhsiungchelyidae	Basilemys	Basilemys sp.	Jasinski, Sullivan, and Lucas (2011); Sullivan et al. (2013)
Testudines	Trionychidae	cf. Plastomenus	cf. Plastomenus sp.	Jasinski, Sullivan, and Lucas (2011)
Squamata	Teiidae	?Chamops	?Chamops sp.	Jasinski, Sullivan, and Lucas (2011)
Squamata	Teiidae	Peneteius	Peneteius sp.	Jasinski, Sullivan, and Lucas (2011)
Crocodylia	Alligatoridae	cf. Brachychampsa	cf. Brachychampsa sp.	Jasinski, Sullivan, and Lucas (2011)
Crocodylia	Crocodyloidea indeterminate	Crocodyloidea indeterminate	Crocodyloidea indeterminate	Jasinski, Sullivan, and Lucas (2011)
Theropoda	Coelurosauria incertae sedis	Richardoestesia	Richardoestesia sp.	Jasinski, Sullivan, and Lucas (2011)
Theropoda	Tyrannosauridae	Tyrannosauridae indeterminate	Tyrannosauridae indeterminate	Jasinski, Sullivan, and Lucas (2011)
Theropoda	Ornithomimidae	Ornithomimidae indeterminate	Ornithomimidae indeterminate	Jasinski, Sullivan, and Lucas (2011)
Theropoda	Caenagnathidae	Ojoraptorsaurus	Ojoraptorsaurus boerei	Jasinski, Sullivan, and Lucas (2011)
Theropoda	Dromaeosauridae	Dineobellator	Dineobellator notohesperus	Jasinski, Sullivan, and Lucas (2011, 2020)
Theropoda	Troodontidae	Troodontidae indeterminate	Troodontidae indeterminate	Jasinski, Sullivan, and Lucas (2011)
Sauropoda	Titanosauridae	Alamosaurus	Alamosaurus sanjuanensis	Sullivan and Lucas (2000); Fowler and Sullivan (2011); Jasinski, Sullivan, and Lucas (2011)
Thyreophora	Ankylosauridae	Ankylosauridae indeterminate	Ankylosauridae indeterminate	Jasinski, Sullivan, and Lucas (2011)
Thyreophora	Nodosauridae	Glyptodontopelta	Glyptodontopelta mimus	Ford (2000); Jasinski, Sullivan, and Lucas (2011); Burns (2008)
Ornithopoda	Hadrosauridae	Lambeosaurini indeterminate	Lambeosaurini indeterminate	Jasinski, Sullivan, and Lucas (2011); Sullivan, Jasinski, Guenther, and Lucas (2011); Sullivan and Lucas (2014)
Marginocephalia	Ceratopsidae	Ojoceratops	Ojoceratops fowleri	Sullivan and Lucas (2010); Jasinski, Sullivan, and Lucas (2011)
Marginocephalia	Ceratopsidae	Chasmosaurinae indeterminate	Chasmosaurinae indeterminate	New record; is Torosaurus-like
Mammalia	Neoplaguaulacidae	Mesodma	Mesodma formosa	Jasinski, Sullivan, and Lucas (2011)
Mammalia	Eucosmodontidae	cf. Essonodon	cf. Essonodon sp.	Jasinski, Sullivan, and Lucas (2011)
Mammalia	Taeniolabidae	aff. Meniscoessus	aff. Meniscoessus sp.	Jasinski, Sullivan, and Lucas (2011)
Mammalia	cf. Pediomyidae	cf. Pediomyidae indeterminate	cf. Pediomyidae indeterminate	Jasinski, Sullivan, and Lucas (2011)
Mammalia	Didelphidae	Alphadon	Alphadon marshi	Jasinski, Sullivan, and Lucas (2011)
Mammalia	Glasbiidae	aff. Glasbius	aff. Glasbius sp.	Jasinski, Sullivan, and Lucas (2011)

• Note: References are not necessarily all of those that mention the given taxa in the strata.

At least six theropod taxa are known from the Naashoibito Member, including the dromaeosaurid Dineobellator notohesperus, the caenagnathid Ojoraptorsaurus boerei, an indeterminate tyrannosaurid, an indeterminate ornithomimid, an indeterminate troodontid, and the coelurosaur Richardoestesia (tooth morphogenus; e.g., Gilmore, 1916; Jasinski et al., 2020; Jasinski, Sullivan, & Lucas, 2011; Lehman, 1981; Sullivan, Jasinski, & van Tomme, 2011). Utilizing the size classes of Holtz (2021), several of these taxa fall into size class 2 (= 11–50 kg), including Dineobellator, Richardoestesia, and the troodontid (Figure 22B, Table 5), although Dineobellator probably falls on the higher end of this range, and may even reach slightly into class 3. Ojoraptorsaurus may have gotten slightly larger, possibly between size classes 2 and 3 (latter = 51–100 kg), while the ornithomimid is inferred to be part of size class 4 (101–500 kg). Finally, the tyrannosaurid would have easily been the largest terrestrial predator in the ecosystem, potentially being part of size class 7 (>5,000 kg), although material reported by Jasinski, Sullivan, and Lucas (2011) were not representative of any individuals reaching the higher size ranges of Tyrannosaurus rex. The Naashoibito tyrannosaur is undergoing further study to determine its evolutionary place among tyrannosauroids.

TABLE 5. Theropod fauna of the Naashoibito Member (Ojo Alamo Formation) from the Maastrichtian of New Mexico.

Ojo Alamo Formation (Naashoibito Member) theropods
Major clade	Subclade	Genus	Species	Size class	Predation type	References
Coelurosauria	Coelurosauria indeterminate	Richardoestesia	sp.	2	Hypercarnivore	Jasinski, Sullivan, and Lucas (2011)
Coelurosauria	Tyrannosauridae	Tyrannosauridae indeterminate	Tyrannosauridae indeterminate	7	Hypercarnivore	Jasinski, Sullivan, and Lucas (2011)
Coelurosauria	Ornithomimidae	Ornithomimidae indeterminate	Ornithomimidae indeterminate	4	?Omnivore	Jasinski, Sullivan, and Lucas (2011)
Coelurosauria	Caenagnathidae	Ojoraptorsaurus	boerei	2–3	Carnivore	Jasinski, Sullivan, and Lucas (2011); Sullivan, Lucas, and Jasinski (2011c)
Coelurosauria	Dromaeosauridae	Dineobellator	notohesperus	2	Hypercarnivore	Jasinski, Sullivan, and Lucas (2011); Jasinski et al. (2020)
Coelurosauria	Troodontidae	Troodontidae indeterminate	Troodontidae indeterminate	2	Carnivore	Jasinski, Sullivan, and Lucas (2011)

• Note: The size classes correspond to other studies on carnivore guilds (e.g., Holtz, 2021). Size class 1: <10 kg; class 2: 11–50 kg; class 3: 51–100 kg; class 4: 101–500 kg; class 5: 501–1,000 kg; class 6: 1,001–5,000 kg; class 7: >5,000 kg.

In addition to the holotype of Dineobellator notohesperus (SMP VP-2430), several other dromaeosaurid specimens are known from the San Juan Basin and the Naashoibito Member of the Ojo Alamo Formation in particular. Jasinski, Sullivan, and Lucas (2011) reviewed the fossil vertebrates from the Naashoibito Member and reported the presence of indeterminate dromaeosaurids from both tooth and non-tooth material. SMP VP-2430, the specimen described herein, was also first reported by Jasinski, Sullivan, and Lucas (2011) as an indeterminate dromaeosaurid, and later named by Jasinski et al. (2020). Jasinski, Sullivan, and Lucas (2011) noted that there may be multiple dromaeosaurid taxa present even though they were focused on the Naashoibito Member. Williamson and Brusatte (2014) studied theropod teeth recovered from the Upper Cretaceous strata of the San Juan Basin. They determined there were at least two types of dromaeosaurids based on differing tooth morphotypes, although they noted only one of these types (“Dromaeosauridae Morphotype A”) in the Naashoibito Member. Jasinski, Sullivan, and Lucas (2011) reported on a large, isolated dromaeosaurid tooth that was later referred to a small tyrannosaurid by Williamson and Brusatte (2014), concurred with by Jasinski et al. (2020). SMP VP-2595, an isolated tooth, was referred to Dromaeosauridae by Jasinski, Sullivan, and Lucas (2011). This tooth does represent a dromaeosaurid, with rounded denticles, slightly less dense denticle density than Dineobellator notohesperus with approximately 17 per 5 mm on the distal carina, and lacking denticles on the mesial carina. While the tip is missing, SMP VP-2595 is larger than the tooth preserved with SMP VP-2430, with an apical length of 14.3 mm, a crown height of 12.6 mm, a fore-aft basal length of 8.4 mm, and a basal width of 4.7 mm. As noted above, Williamson and Brusatte (2014) reported the presence of dromaeosaurids in the member as well based on NMMNH P-32814, which they referred to their “Dromaeosauridae Morphotype A.” This tooth has rounded denticles that are not hooked, as in Acheroraptor, has 25 denticles per 5 mm on the distal carina, and has small mesial denticles. It has an apical length of 6.9 mm, a crown height of 5.1 mm, a fore-aft basal length of 3.5 mm, and a basal width of 1.7 mm. While SMP VP-2595 may represent Dineobellator notohesperus, NMMNH P-32814 appears distinct, and argues for the presence of at least two dromaeosaurid taxa in the Naashoibito Member. Lehman (1981) noted a complete right metatarsal I similar to Stenonychosaurus that he identified as dromaeosaurid or troodontid from the Naashoibito Member. However, the specimen is part of a private collection and was not considered further by Jasinski, Sullivan, and Lucas (2011); Jasinski et al. (2020), nor in the present study.

3.5 Maastrichtian North American dromaeosaurids

Dineobellator notohesperus is the only diagnostic dromaeosaurid from the Maastrichtian south of Montana and northern South Dakota (i.e., southern Laramidia). Features of the forelimb, unguals, caudal vertebrae, and teeth help distinguish it from other known North American Maastrichtian dromaeosaurids. Based on similar humerus lengths (Table 3), Dineobellator would have been similar in size to Deinonychus, and larger than some taxa such as Acheroraptor and Saurornitholestes, but smaller than the large-bodied Dakotaraptor. While DePalma et al. (2015) noted the presence of two different-sized individuals of Dakotaraptor, they attribute this to sexual dimorphism. The only potential overlap in material between the stratigraphically co-occurring Acheroraptor and Dakotaraptor are teeth, with estimates of 22–25 and 22–23 denticles per 5 mm for the distal denticles of each, respectively. This still leaves open the possibility of them being different ontogenetic stages of the same species, although little work has investigated ontogeny in dromaeosaurids (e.g., Parsons & Parsons, 2015), and lack of overlapping material makes this difficult to determine in the case of these two taxa. Comparatively, Dineobellator has larger and less condensed denticles (18–20 denticles per 5 mm) than either Hell Creek dromaeosaurid. Additionally, based on the current available material and phylogenetic analysis of Jasinski et al. (2020), these two Hell Creek taxa are found in two distinct clades, with Dakotaraptor recovered as a basal member of the Dromaeosaurinae and Acheroraptor as a basal member of Velociraptorinae. Jasinski et al. (2020, figure 3) found the North American dromaeosaurids from the Maastrichtian in different areas of their phylogenetic analysis. While Dakotaraptor was recovered basally within Dromaeosaurinae, Acheroraptor and Dineobellator were recovered in Velociraptorinae. Additionally, Acheroraptor was recovered as the basal-most velociraptorine, while Dineobellator was sister to Tsaagan + Linheraptor. Conversely, the recent study by Powers et al. (2022) recovered Acheroraptor within Saurornitholestinae as did our phylogenetic analysis. While Powers et al. did not include Dineobellator or Dakotaraptor in their analysis, we found Dineobellator to lie in an unresolved clade with currently recognized eudromaeosaur clades (Dromaeosaurinae, Saurornitholestinae, Velociraptorinae; Figure 20), probably due to the less complete nature of the holotype (SMP VP-2430). The phylogenetic placement of Acheroraptor in relation to Dakotaraptor, and whether this latter taxon is still more closely aligned with other dromaeosaurines, must be further investigated.

3.6 Ecological niche/theropod communities

A recent study by Holtz (2021) suggested that theropod guilds from the Late Cretaceous tend to be relatively similar, particularly in the common size classes between them. The Naashoibito theropods are similar to many of the other theropod communities, particularly those with tyrannosaurids. Holtz (2021), when discussing the various theropod communities in the north and size classes, often found several theropods within size class 2, with tyrannosaurids often classified as size class 6 (501–1,000 kg) or class 7. Even faunas that have more medium-sized theropods usually have these classes filled by smaller tyrannosaurids or tyrannosauroids (Holtz, 2021). One of the hypotheses for many of these faunas missing medium-sized theropods is that juvenile and sub-adult tyrannosaurids filled those niches, outcompeting other potentially similar-sized taxa (e.g., Holtz, 2021; Schroeder et al., 2021). This ontogenetic niche shift among the ecological niches of large-bodied theropods may lead to less opportunities for small- to medium-sized theropod taxa within theropod communities. Erickson et al. (2004) hypothesized that Tyrannosaurus remained below approximately 1,000 kg for around 14 years, making this part of the lag cycle (see their figure 2). This was followed by a log phase of very high growth rates that lasted a little over 5 years, further followed by lower growth rates in a stationary phase. While Tyrannosaurus was shown to go through a log phase marked by incredibly high growth rates, other tyrannosaurids such as Daspletosaurus, Gorgosaurus, and Albertosaurus saw much lower growth rates through their log phase and reached only a fraction of the total adult size (Erickson et al., 2004). Regardless, theropod communities with these latter tyrannosaurids are found to contain similar theropod guilds and theropod community structures (see Holtz, 2021; Schroeder et al., 2021).

There are several theropod communities that have several large taxa (e.g., from the Anacleto, Bajo Barreal, Huincul, Bahariya, Aoufous, Elrhaz, Sao Khua, and Wessex formations) although these tend to be prior to the Campanian, and lack derived tyrannosaurids (Holtz, 2021). However, some other Maastrichtian theropod communities, such as those from the Lameta Formation (India), also have multiple larger-bodied theropods (e.g., Indosaurus matleyi, Rajasaurus narmadensis, Rahiolisaurus gujaratensis) but lack derived tyrannosaurids. This may imply that abelisaurids grew and behaved differently than tyrannosaurids, allowing more large-bodied taxa to co-exist in the same ecosystems. However, while Erickson et al. (2004) suggested that tyrannosaurids had different growth rates during their exponential growth stages, particularly with Tyrannosaurus rex exhibiting significantly higher growth rates during these stages, Holtz (2021) found ecosystems with different tyrannosaurid taxa possessed similar theropod guilds. Rather, what may be important in tyrannosaurids is not then how they grow, but how they behave, with similar behavior between members of the clade. Behavioral differences between tyrannosaurids and other large Late Cretaceous theropods like abelisaurs may then be the reason that non-tyrannosaurid-inhabited Late Cretaceous terrestrial ecosystems more often had multiple large terrestrial predators.

Dakotaraptor occupied a size class not often found in other theropod communities with tyrannosaurids. While Dineobellator was not as large, it was still larger than some other Late Cretaceous eudromaeosaurs such as Dromaeosaurus, Saurornitholestes, and Velociraptor, and was similar in size to Deinonychus based on similar dimensions of the humerus. This argues there was more variation in the members of these theropod communities than previously thought. It is likely that some of these other size classes had different theropod species within them, but if they were rarer components of the landscape, or if there were other preservation or collecting biases against their fossils, it will require greater samples to better understand their respective faunas. Finding whether Dakotaraptor, as a medium-sized theropod within a community including Tyrannosaurus, represents a body class of theropods that was present but rare in other tyrannosaurid-dominated theropod communities, or if it represents an endemic theropod class, will need further data. This may also argue for more variation in theropod communities in the Late Cretaceous, with more medium-sized theropods competing with juvenile and sub-adult tyrannosaurids in some ecosystems while others, like the Naashoibito with Dineobellator, may still lack distinct taxa in these size classes.

4 CONCLUSION

Dineobellator notohesperus is the most complete North American dromaeosaurid known from south of the 43rd latitude (South Dakota). The holotype (SMP VP-2430) consists of at least 27 identifiable elements that provide data on the morphology of this unique southern North American eudromaeosaur species. Some material previously identified in the holotype (right metatarsal I and left astragalus of Jasinski et al., 2020) are re-identified, while the more thorough description of other material allows for more comparison with other known dromaeosaurid material.

A new phylogenetic analysis recovers Dineobellator notohesperus in an unresolved polytomy with other recognized eudromaeosaur clades (Dromaeosaurinae, Saurornitholestinae, Velociraptorinae) rather than as a velociraptorine. Previously known and named clades such as Saurornitholestinae and Velociraptorinae are recovered with taxa known exclusively from North America and Asia, respectively. Dromaeosaurinae, however, is recovered with both North American and Asian taxa, implying either a more complicated biogeographic history for this clade or that some of its taxa will eventually be found in different clades.

Morphological differences between Dineobellator and other dromaeosaurids implies possible behavioral differences. Based on similar lengths of the humerus, Dineobellator would have been close to the body length of Deinonychus, but a more gracile humerus in the former indicates less body mass in Dineobellator. Another possibility is that Dineobellator may have been a smaller animal with relatively longer forelimbs than Deinonychus. Abnormal bone growth on several bones point to pathologies derived from injuries or maladies during life for the holotype individual. The sediments of the San Juan Basin preserve diverse terrestrial vertebrate communities that existed during the last 10 million years of the Cretaceous Period. The Naashoibito Member, in particular, preserves a theropod community similar to more well-studied faunas from farther north in North America. Dineobellator is the only diagnosed dromaeosaurid from the Maastrichtian in the southwestern United States, and as such occupies an important biogeographic and temporal position in the evolution of Dromaeosauridae. Dineobellator also provides important new insights regarding the total morphological variation seen in Dromaeosauridae, particularly at the end of the Cretaceous. More southern North American dromaeosaurid specimens may yet determine that if biogeography and convergence are important factors in dromaeosaurid evolution, a southern, North American eudromaeosaur clade will eventually be found.

AUTHOR CONTRIBUTIONS

Steven Jasinski: Conceptualization (lead); formal analysis (lead); investigation (lead); methodology (lead); software (lead); writing – original draft (lead); writing – review and editing (lead). Robert Sullivan: Conceptualization (equal); data curation (supporting); supervision (supporting); writing – original draft (supporting); writing – review and editing (equal). Aja Carter: Methodology (equal); writing – original draft (supporting); writing – review and editing (equal). Erynn Johnson: Methodology (equal); writing – original draft (supporting); writing – review and editing (equal). Sebastian Dalman: Methodology (supporting); writing – original draft (supporting); writing – review and editing (equal). Juned Zariwala: Methodology (supporting); writing – original draft (supporting); writing – review and editing (equal). Philip J. Currie: Supervision (supporting); writing – original draft (supporting); writing – review and editing (equal).

ACKNOWLEDGMENTS

We thank many colleagues for discussion on dromaeosaurids, including Stephen Brusatte, David Evans, Denver Fowler, Derek Larson, Mark Norell, David Varricchio, Spencer Lucas, and Peter Dodson. Thanks also go to those who we had discussions with on San Juan Basin stratigraphy and fossils, including Spencer Lucas, and Thomas Williamson. Joan Buccilli, Robert Giegengack, and Hermann Pfefferkorn offered immense support throughout the project. Phylopics was used for a majority of the silhouettes used in the phylogenetic reconstruction in Figures 20 and 22, and the artists whose work these represent are gratefully acknowledged, including Dmitry Bogdanov, Arthur Brum, Danny Cicchetti, Andy Farke, FunkMonk, Scott Hartman, Jaime Headden, T. Michael Keesey, Matthew Martyniuk, Brad McFeeters, Lankester Edwin Ray, Nobu Tamura, Milton Tan, Emily Willoughby, and zoosnow. SMP VP-2430 was collected under Bureau of Land Management Paleontological Resources Use Permit NM07-001 S. Phil Gensler and Sherrie Landon provided help with acquiring and renewing collecting permits. James Nikas III helped in the initial collection of the specimen. Robert M. Sulluvian led the fieldwork that recovered the first elements of the type specimen. Steven E. Jasinski thanks Tony Fiorillo, Cathy Forster, and Dave Weishampel for the invitation to contribute to this volume. Tony Fiorillo provided editorial help and comments, and two anonymous reviewers provide thorough and helpful comments and suggestions that greatly improved this manuscript. Peter Dodson is thanked for the guidance and immense support he provided throughout the years, including as the doctoral advisor to several of us (Steven E. Jasinski, Aja M. Carter, and Erynn H. Johnson). He has contributed in numerous ways to the study and our knowledge of dinosaurs, and we are all proud to be able to continue with this endeavor. We are all honored to be able to contribute this article in honor of an incredible scientist and wonderful human being.