Introduction
The first terrestrial mammal fossils known from the Chesapeake Group on the mid-Atlantic coast of the United States were proboscidean remains identified by Harlan (Reference Harlan1842). In the over 170 years of collecting since then, hundreds of specimens of terrestrial mammals have been recovered from the marine deposits of the Chesapeake Group (e.g., Harlan, Reference Harlan1842; Cope, Reference Cope1867; Hayden, Reference Hayden1873; Shattuck, Reference Shattuck1904; Berry, Reference Berry1938; Gazin and Collins, Reference Gazin and Collins1950; Blackwelder and Ward, Reference Blackwelder and Ward1976; Tedford and Hunter, Reference Tedford and Hunter1984; Wright and Eshelman, Reference Wright and Eshelman1987). The early Barstovian Calvert Formation has had a wide array of both marine and terrestrial mammals collected from its strata (e.g., Gazin and Collins, Reference Gazin and Collins1950; Wright and Eshelman, Reference Wright and Eshelman1987), including carnivorans (Berry, Reference Berry1938; Tedford and Hunter, Reference Tedford and Hunter1984). However, only a few terrestrial mammals have been identified from the younger late Barstovian Choptank Formation, including the tayassuids Desmathyus sp. and “Prosthenops” niobrarensis (“Hesperhys” of Gazin and Collins, Reference Gazin and Collins1950), the proboscidean cf. Gomphotherium calvertensis (Gazin and Collins, Reference Gazin and Collins1950), and the protoceratid artiodactyls Prosynthetoceras sp. and ?Synthetoceras sp. (e.g., Gazin and Collins, Reference Gazin and Collins1950; Tedford and Hunter, Reference Tedford and Hunter1984; Wright and Eshelman, Reference Wright and Eshelman1987). No carnivoran material has been previously reported from the Choptank Formation.
In the older strata of the underlying Calvert Formation, a partial lower jaw (USNM 15561) of a borophagine dog was found and named Tomarctus marylandica by Berry (Reference Berry1938). Tedford and Hunter (Reference Tedford and Hunter1984) reviewed various aspects of the Miocene in eastern North America and suggested that Tomarctus marylandica may be referable to Cynarctus. They did, however, also feel that the species would represent a primitive member of Cynarctus and was comparable to other early Barstovian species. In their thorough review of the phylogenetic systematics of the Borophaginae, Wang et al. (Reference Wang, Tedford and Taylor1999) found that the only distinct feature of ?C. marylandica (Berry, Reference Berry1938) that is indicative of Cynarctus is a protostylid (or eoconulid) on the m1. However, the distinct crest on the posterior (or distal) face of the m1 trigonid has not been seen in other borophagines, making it distinct (Wang et al., Reference Wang, Tedford and Taylor1999). Even so, Wang et al. (Reference Wang, Tedford and Taylor1999) felt that the identity of ?C. marylandica as a medium-sized borophagine was likely and maintained the questionable referral of ?Cynarctus marylandica to Cynarctus as was proposed by Tedford and Hunter (Reference Tedford and Hunter1984). Regardless of its generic affinity, ?C. marylandica represents one of only three known borophagine dogs from the northeastern United States, with the other two being Metatomarctus canavus (Simpson, Reference Simpson1932) from the late early Hemingfordian of Delaware (and Florida) and Paracynarctus kelloggi (Merriam, Reference Merriam1911) from the late early Hemingfordian of Delaware (Wang et al., Reference Wang, Tedford and Taylor1999).
Due to the location of the Chesapeake Group in the Calvert Cliffs, much of the more recent collecting has been done by amateurs. Stratigraphically, the Chesapeake Group is made up of the older Calvert Formation, the middle Choptank Formation, and the younger St. Mary’s Formation (Fig. 1). Fossils collected in situ come from the cliffs themselves. However, fossils frequently fall from the cliffs and are found as float below the cliffs. In addition, fossils can be moved around by wave action. Often, when fossils fall from the cliffs, they are found on the beach below the formation they originated from. While there is not always a certainty of a fossil originating in the closest formation, is does provide a probable geologic formation for fossils found as float, and not in situ, from Calvert Cliffs.
Figure 1 Stratigraphy of the Chesapeake Group formations (modified from Vogt and Eshelman, Reference Vogt and Eshelman1987; Petuch and Drolshagen, Reference Petuch and Drolshagen2010; Weems and George, Reference Weems and George2013). ?Cynarctus marylandica was collected from Zone 10 in the Calvert Formation and is marked by ?Cm. Cynarctus wangi n. sp. (USNM 534040) is believed to be from the Choptank Formation, which means somewhere between Zones 17 and 19, and is marked by Cw. Note that Petuch and Drolshagen (Reference Petuch and Drolshagen2010) mentioned a small unit that is found along part of the St. Mary’s River called the Chancellor Point Member of the St. Mary’s Formation, which is not shown in the figure. BCM=Boston Cliffs Member; CM=Conoy Member; DrM=Drumcliff Member; Fm.=formation; Mbr.=member; SLM=St. Leonard Member; WPM=Windmill Point Member.
Locality and repository information
Specimens examined for this study and abbreviated in the text are as follows: AMNH, American Museum of Natural History, New York, New York; UCMP, Museum of Paleontology, University of California at Berkeley; USNM, United States National Museum, Washington, DC. The material discussed within (USNM 534040) comes from the Chesapeake Group of Maryland. Because of where it was collected in the Calvert Cliffs area, it is believed to be from the middle Miocene Choptank Formation.
Systematic paleontology
Dental terminology used herein follows Hershkovitz (Reference Hershkovitz1971) and Dahlberg (Reference Dahlberg1971). Systematic paleontology follows McKenna and Bell (Reference McKenna and Bell1997) and Wang et al. (Reference Wang, Tedford and Taylor1999).
Class Mammalia Linnaeus, Reference Linnaeus1758
Order Carnivora Bowdich, Reference Bowdich1821
Suborder Caniformia Kretzoi, Reference Kretzoi1943
Infraorder Cynoidea Flower, Reference Flower1869
Family Canidae Fischer de Waldheim, Reference Fischer de Waldheim1817
Subfamily Borophaginae Simpson, Reference Simpson1945
Tribe Borophagini Wang, Tedford, and Taylor, Reference Wang, Tedford and Taylor1999
Subtribe Cynarctina McGrew, Reference McGrew1937
Genus Cynarctus Matthew, Reference Matthew1902
Type species
Cynarctus saxatilis Matthew, Reference Matthew1902.
Included species
Cynarctus crucidens Barbour and Cook, Reference Barbour and Cook1914; Cynarctus galushai Wang et al., Reference Wang, Tedford and Taylor1999; ?Cynarctus marylandica (Berry, Reference Berry1938); Cynarctus saxatilis Matthew, Reference Matthew1902; Cynarctus voorhiesi Wang et al., Reference Wang, Tedford and Taylor1999; Cynarctus wangi, n. sp.
Distribution
Early Barstovian of California and Maryland; late Barstovian of California, Colorado, Maryland, and Nebraska; early Clarendonian of Nebraska, South Dakota, and Texas; late Clarendonian of Nebraska; and Clarendonian of Texas.
Diagnosis
Derived characters that distinguish Cynarctus from Paracynarctus are an auditory meatus of small diameter with a small lip, M1 transversely narrow and subquadrate, c1 strongly recurved, and p4 posterolingual shelf with weak ‘metastylid’ occasionally present. In addition, Cynarctus and Paracynarctus share several derived dental features distinguishable from those of Metatomarctus and other, more primitive taxa: a weak subangular lobe, a high mandibular condyle, short P4 relative to M1, enlarged M2, presence of an m1 protostylid, widened m1 talonid, and elongated m2 and m3. Diagnosis after Wang et al. (Reference Wang, Tedford and Taylor1999).
Remarks
Wang et al. (Reference Wang, Tedford and Taylor1999) thoroughly reviewed Cynarctus. For further information, see their discussion of the genus and included species.
Cynarctus wangi new species
Figure 2 Holotype right M2 of Cynarctus wangi n. sp. (USNM 534040). (A) Occlusal view. (B) Illustration of occlusal view. c=cleft; hyp=hypocone; labc=labial (external) cingulum; linc=lingual (internal) cingulum; met=metacone; mts=metastyle; mtl=metaconule; par=paracone; posc=posterior (distal) cingulum; prs=parastyle; pro=protocone. Scale bar=1 cm.
Holotype
USNM 534040, right M2.
Diagnosis
Member of Cynarctus based on having enlarged M2s with posterior (distal) expansions (Wang et al., Reference Wang, Tedford and Taylor1999), in turn making it a member of the Borophaginae and the Canidae. Further diagnosed from other species of Cynarctus by the prominent degree of posterior expansion of the M2 concentrated with the posterior cingulum; reduction of the labial cingulum lateral to the metacone; lack of expansion labially and/or posteriorly in the metacone; enlarged metaconule; enlarged lingual cingulum and hypocone; less pronounced posterior expansion near the hypocone; less developed metastyle; relatively weak posterior cingulum; and slightly pronounced labial cingulum near the anterior of the tooth.
Type locality and horizon
Chesapeake Group, believed to be from the Choptank Formation (late Barstovian), Calvert County, Maryland.
Etymology
Named in honor of Xiaoming Wang for his work with borophagines, other fossil canids, and carnivorans in general.
Description and remarks
The right M2 (USNM 534040) of Cynarctus wangi n. sp. is subquadrate in shape and anteroposteriorly (or mesiodistally) elongate, particularly near the lingual portion. Maximum length is 13.44 mm (anteroposteriorly or mesiodistally) and maximum width is 11.77 mm (labiolingually). Paracone is pronounced, with the metacone being slightly reduced in comparison to the former. A distinct valley is present between the paracone and metacone (and between the postparacrista and premetacrista), completely separating them. Metastyle is reduced and less distinct, while only slightly inflated posteriorly (or distally). A labial cingulum is distinct around the paracone and is believed to be reduced around the metacone according to the morphology that is preserved; however, a portion of the cingulum is missing from the specimen lateral to the metacone. Parastyle is relatively indistinct mesial (or anterior) to the paracone. Protocone is pronounced and exhibits a gentle curve. Anteriorly (or mesially), the curved ridge of the protocone reaches the parastyle via the protocrista (or protoloph). Lingual to the protocrista lies another portion of the cingulum, although the protocone creates a disjunct in the elevations between the labial and lingual portions of the cingulum on the anterior (mesial) edge of the tooth. A metaconule is present posterior to the protocone, with a valley separating the two. Metaconule is pronounced, with a distinct curvature posteriorly (distally) toward the metastyle (= transcrista), which is also distinct. In addition, the hypocone is pronounced, with its anteroposterior (or mesiodistal) ridge curving toward the lingual ridge of the metaconule, forming the distocrista. However, the distocrista is separated from the metaconule by a valley posterior (distal) to the post-talon basin. A posterior (distal) cingulum lies distal to the metaconule and hypocone as well. Lingual (or internal) cingulum is pronounced, especially as it nears the hypocone, with a cleft present that separates a nearly cusp-like development of the anterolingual (or mesiolingual) cingulum (= entostyle or pericone) from the hypocone. A strong posterior (distal) expansion of the M2 is most pronounced near the hypocone.
Cynarctus is characterized by having enlarged M2s with posterior (distal) expansions (Wang et al., Reference Wang, Tedford and Taylor1999). According to AMNH 27543 (holotype) and AMNH 27550, Cynarctus galushai has a strongly curved M2, with a more distinct posterior (distal) expansion of the lingual portion than that in C. wangi (USNM 534040). The M2 in AMNH 27550 is more complete than that of the holotype (AMNH 27543), allowing for a more thorough comparison. A strong labial cingulum is present on the M2 in AMNH 27543. The protocone in C. galushai is relatively gently curved, similar to that in C. wangi. However, the metaconule and hypocone are less pronounced in C. galushai. As mentioned by Wang et al. (Reference Wang, Tedford and Taylor1999, p. 118), the lingual cingulum of the “M2 in C. galushai is thickened to suggest an initiation of a conate hypocone,” a condition also present in C. wangi.
?Cynarctus marylandica is only known from lower jaw material from USNM 15561 (holotype) and USNM 299471, and so cannot be compared directly to C. wangi (USNM 534040).
Cynarctus saxatilis is known from multiple M2s from several referred specimens (see Wang et al., Reference Wang, Tedford and Taylor1999). Similar to USNM 534040, the labial cingulum is absent or weak lateral to the metacone in C. saxatilis. In addition, the metacone in C. saxatilis is relatively small and not expanded posteriorly (distally). In focusing on a referred specimen (UCMP 29891), the M2 is quadrate, with the anteroposterior (mesiodistal) length roughly equal to the labiolingual width. The metacone is more pronounced than that of USNM 534040, being almost equal in size to the paracone. In addition, the protocone is more constricted and more sharply curved, and the metaconule is less pronounced than in USNM 534040.
Cynarctus voorhiesi has an isolated left M2 known from a referred specimen (AMNH 49143). The tooth is heavily worn, but as was mentioned by Wang et al. (Reference Wang, Tedford and Taylor1999), the M2 is greatly expanded posteriorly (distally), although the hypocone is less posteriorly (distally) expanded than in C. crucidens (discussed in the following).
Cynarctus crucidens is known from multiple M2s from several referred specimens (see Wang et al., Reference Wang, Tedford and Taylor1999). The M2 is anteroposteriorly (mesiodistally) elongate (and may even exceed the length of the M1). The hypocone is large, aiding in the anteroposterior (mesiodistal) elongation, and posteriorly (distally) situated. The metastyle is well developed. A strong posterior (distal) cingulum is also present. Compared to C. wangi (USNM 534040), the protocone is more constricted and strongly curved in C. crucidens (particularly AMNH 49312). A slightly smaller and less pronounced metacone and hypocone also differentiate C. crucidens from C. wangi.
Similar to the enlarged M2s present in Cynarctus, several other borophagine genera have M2s that are at least somewhat enlarged and/or quadrate to subquadrate in shape, including Cormocyon, Metatomarctus, Oxetocyon, Paracynarctus, Phlaocyon, Psalidocyon, and Tomarctus.
Most of the borophagine genera with enlarged M2s tend to exhibit that expansion in the labiolingual, rather than the anteroposterior (mesiodistal), direction. (e.g., Cormocyon, Metatomarctus, Oxetocyon, Paracynarctus, Phlaocyon, Psalidocyon, Tomarctus). Paracynarctus, however, is more expanded anteroposteriorly (mesiodistally), making it more quadrate in shape and more similar to both Cynarctus and USNM 534040 (with both former genera members of the Cynarctina). When posterior (distal) expansion of the M2 does occur within borophagines, it is frequently focused near the lingual portion of the tooth, giving the tooth a general concave curvature on the posterior (distal) surface (e.g., Cormocyon, Metatomarctus, Oxetocyon, Phlaocyon, Psalidocyon, Tomarctus). This tends to make the teeth less quadrate in shape and often more ovular. Again, Paracynarctus exhibits more expansion posteriorly (distally) with less posterior (distal) curvature, similar to Cynarctus and USNM 534040. The protocone is often strongly curved and connects to the metaconule, which is sometimes offset from the plagiocrista (or metaloph) of the protocone. In many borophagines the metacone is differentiated and expanded in the labial and/or posterior (distal) directions (e.g., Cormocyon, Oxetocyon, Paracynarctus, Phlaocyon, Psalidocyon, Tomarctus). Metatomarctus has a slight expansion of the metacone, less than many other borophagines, but still more prominent than in C. wangi (USNM 534040). In addition, protocones are often at least somewhat reduced and strongly bent or curved (e.g., Cormocyon, Oxetocyon, Paracynarctus, Phlaocyon, Psalidocyon, Tomarctus). The metaconule is commonly less prominent and offset from the protocone (e.g., Cormocyon, Metatomarctus, Oxetocyon, Psalidocyon, Tomarctus). Sometimes the hypocone, particularly the apex, is situated more lingually (e.g., Metatomarctus, Paracynarctus, Psalidocyon, Tomarctus). In Cormocyon, the internal (lingual) cingulum is more symmetrically distributed. In Phlaocyon, the metaconule and hypocone tend to be situated more closely, while the hypocone is more restricted and not nearly as wide. Similar to USNM 534040, a cleft or valley is present in Paracynarctus that separates the entostyle (or pericone) from the hypocone.
Discussion
The enlarged M2, with a majority of the expansion in the anteroposterior (or mesiodistal) direction, shows that USNM 534040 compares most closely with M2s known from Cynarctina (includes Cynarctus and Paracynarctus). However, USNM 534040 can be distinguished from Paracynarctus by several features in the latter, including an expansion of the metacone; a somewhat reduced and strongly bent or curved protocone; and a hypocone, particularly the apex, that is situated more lingually. Overall morphology suggests that the tooth most closely resembles teeth attributed to Cynarctus and is why it is here referred to a distinct species within said genus.
Cynarctus is thought to be more hypocarnivorous, or consuming less meat and more nonmeat foods, than Paracynarctus and other borophagines (Wang et al., Reference Wang, Tedford and Taylor1999). In addition, the hypocarnivorous features of Paracynarctus and Cynarctus are believed to be derived features with respect to similar conditions present in Metatomarctus and Desmocyon. Similarly, various parts of the dentition of Cynarctus are rather ursid-like, just as the name implies. Originally, Matthew (Reference Matthew1902) believed Cynarctus belonged to the Amphicyonidae, and in later studies maintained that it was not a canid (Matthew, Reference Matthew1924, Reference Matthew1930). However, McGrew (Reference McGrew1937, Reference McGrew1938) believed Cynarctus belonged to the Procyonidae because of its hypocarnivorous dentition. Finally, Wang et al. (Reference Wang, Tedford and Taylor1999) found Cynarctus to be a canid according to features of the middle ear and a member of the Borophaginae according to various cranial and dental features.
Cynarctus galushai, the oldest species of the genus, is known from the early to late Barstovian of California. ?Cynarctus marylandica is known from the early Barstovian of Maryland. Cynarctus saxatilis is known from the late Barstovian of Colorado and Nebraska. Cynarctus wangi is now known from the late Barstovian of Maryland. Cynarctus voorhiesi is known from the late late Barstovian of Nebraska. Cynarctus crucidens, the youngest species, is known from the early Clarendonian of South Dakota and the early to late Clarendonian of Nebraska and Texas.
Cynarctus galushai exhibits the most primitive, non-hypocarnivorous, features of the genus (Wang et al., Reference Wang, Tedford and Taylor1999). These are especially evident when compared to C. saxatilis, C. voorhiesi, and C. crucidens. Species of Cynarctus exhibit a general decrease in size, coupled with a general reduction of the premolars and an increase in the size, particularly the posterior (or distal) expansion, of the M2. While other features are discussed by Wang et al. (Reference Wang, Tedford and Taylor1999), it is the features of the M2 that are vital to the taxonomic identification of USNM 534040 (C. wangi). While these comparisons were discussed previously, it is evident that several features of C. wangi n. sp. are considered derived features by Wang et al. (Reference Wang, Tedford and Taylor1999). The posterior (distal) enlargement of C. wangi is prominent, more so than C. galushai and C. saxatilis, although this is exhibited more with the posterior (distal) cingulum than the hypocone. The labial cingulum appears to be reduced lateral to the metacone in C. wangi, along with a lack of expansion labially and/or posteriorly (distally) in the metacone, conditions similar to those in C. saxatilis, C. voorhiesi, and C. crucidens. The metaconule is also enlarged, more similar to C. voorhiesi and C. crucidens. In addition, the lingual cingulum and hypocone are both enlarged, similar to C. saxatilis, C. voorhiesi, and C. crucidens. Wang et al. (Reference Wang, Tedford and Taylor1999) mentioned that C. saxatilis lacks a posterior (distal) expansion near the hypocone on the M2, a feature that implies it is more primitive than C. voorhiesi and C. crucidens, which both exhibit this feature. Cynarctus wangi lies somewhat between these two groups, with general posterior (distal) expansion more than C. saxatilis but less prominent than the expansion exhibited by C. voorhiesi and C. crucidens. General shape of the tooth is more anteroposteriorly (mesiodistally) elongate than C. saxatilis and more similar to C. voorhiesi and C. crucidens. Conversely, the primitive characters of C. wangi include a less pronounced posterior (distal) expansion near the hypocone, less developed metastyle, and a relatively weak posterior (distal) cingulum (in comparison to C. crucidens). Cynarctus wangi also possesses a slightly pronounced labial cingulum toward the anterior (mesial) of the tooth; this is not as pronounced as in C. galushai, but more so than in C. saxatilis, C. voorhiesi, and C. crucidens. Cynarctus wangi seems to be more derived than C. galushai and less derived than C. voorhiesi and C. crucidens in terms of Cynarctus M2 features. In general, it shows more features in common with C. voorhiesi and C. crucidens than with C. saxatilis, implying C. wangi is slightly more derived than the latter taxon.
Comparison with the geographically similar ?C. marylandica is important to consider, although this is difficult as the upper dentition of this taxon is not known. Although only an upper M2 is known from C. wangi (USNM 534040), there is a close relationship between the occlusal surfaces of upper M2s and lower m2s that can give insights into what the hypothetical morphology of a lower m2 from C. wangi could look like. As such, occlusal relationships between upper M2s and lower m2s can be used to aid in comparing C. wangi and ?C. marylandica. In members of Cynarctus that exhibit anteroposterior (mesiodistal) expansion of the upper M2, commonly with a pronounced lingual cingulum near the protocone, the corresponding lower m2 often has an expanded talonid basin permitted by a relatively anteriorly situated (or mesially situated) metaconid. The enlarged grinding surface occludes with the upper M2 protocone and/or lingual cingulum. While this morphology in C. wangi (USNM 534040) seems to be distinct from what would be expected in ?C. marylandica and is distinct from C. galushai, it is more similar to C. saxatilis and C. crucidens. Indeed, a hypothetical lower m2 of C. wangi would have an enlarged hypoconid that would occlude with the large trigon basin in USNM 534040. Given the morphological similarities present of the upper M2 morphology of C. wangi (USNM 534040) with C. saxatilis and C. crucidens (particularly the latter taxon), one would expect the corresponding lower m2 morphology to also show similarities. If this is the case, than the lower m2 of C. wangi would be expected to possess more accessory cusps (namely around and including the protostylid [or eoconulid] and metastylid [or postmetaconulid]) than are present in the ?C. marylandica holotype. Indeed, based on occlusal relationships between upper M2s and lower m2s, the morphological differences between the M2 of C. wangi and the m2 of ?C. marylandica are as different as those observed across different Cynarctus species. This agrees with the argument that USNM 534040 represents a distinct cynarctin taxon.
Phylogenetic constraints on canid tooth morphology may lead to the addition and/or complication of cusp morphology of molars, particularly of the M2 in the upper dentition. Rather than re-evolving an upper M3, it may be easier to add complexity to the M2 that is already present, leading to improved and complicated grinding surfaces. This would be especially important for hypocarnivorous animals and in the more hypocarnivorous members of Cynarctus and borophagine canids in general.
If C. wangi is from the Choptank Formation, as it is believed to be, than it would be roughly late Barstovian in age (see Fig. 1). Regardless, it appears to be younger than the holotype of ?C. marylandica (early Barstovian, Calvert Formation). Therefore, there are two possible scenarios with respect to USNM 534040. Due to geographic and stratigraphic provenance, there is a possibility that USNM 534040 may be referable to ?C. marylandica. On the other hand, USNM 534040 could be a younger and more derived species of Cynarctus or cynarctin borophagine.
If the first scenario were correct, then the generic placement of ?Cynarctus marylandica would also be correct. It would also suggest that ?C. marylandica has M2s that are more derived than C. galushai and C. saxatilis but less derived than C. voorhiesi and C. crucidens. If the second scenario is correct, then the generic placement of ?C. marylandica cannot be commented on. However, here we argue for the second scenario where a closely related species of cynarctin borophagine is present in slightly younger strata in the same geographic region as ?C. marylandica. In addition, it could be that ?C. marylandica may have given rise to C. wangi, which exhibits M2 features more derived than C. saxatilis but less derived than C. voorhiesi and C. crucidens, potentially providing an example of anagenesis. While it is speculative, such a situation could possibly be showing parallel evolution (more hypocarnivorous behavior) between closely related species, one set in western North America and another set in the east. It is noted that multiple borophagine taxa seem to be present in the late early Hemingfordian of Delaware at the Pollack Farm Site (Wang et al., Reference Wang, Tedford and Taylor1999), so multiple borophagine taxa potentially living together is known from another site in the northeastern United States.
Summary
USNM 534040 represents a right M2 of a cynarctin borophagine (Canidae: Borophaginae: Cynarctina), here identified as Cynarctus wangi, and is the first carnivoran reported from the middle Miocene Choptank Formation (Chesapeake Group). However, the presence of another cynarctin borophagine from the Chesapeake Group, ?Cynarctus marylandica from the older Calvert Formation, has potential implications. As ?C. marylandica is known only from lower dentition, its generic affinities are uncertain. Indeed, USNM 534040 advocates two possible scenarios regarding its identification. In one, its geographic and stratigraphic provenance imply that the specimen belongs to ?C. marylandica. If this were correct, then the generic placement of ?C. marylandica would also be correct and the taxon would be more derived toward hypocarnivory than C. galushai and C. saxatilis yet less derived than C. voorhiesi and C. crucidens. However, the second possibility is that the specimen represents a distinct cynarctin borophagine that is also present in the Chesapeake Group, but in strata younger than the type specimen of ?C. marylandica. This latter possibility is argued here and may imply that C. wangi represents the next evolutionary stage of ?C. marlyandica and that anagenesis was occurring in Cynarctus during this time. Regardless, the presence of C. wangi expands the sparse fossil record of borophagine canids in northeastern North America and furthers our knowledge of the fossil record of terrestrial taxa in this region.
Acknowledgments
D. Bohaska brought attention to and loaned the specimen to SCW. C. Mehling provided images of AMNH 49143. This research was supported in part by the National Science Foundation (EAR-0958985). P. Dodson provided a helpful review of an earlier version of the manuscript. We thank Lars Werdelin and an anonymous reviewer for helpful comments that improved the manuscript. We also thank Editor Mary Silcox for her comments and suggestions, which helped improve the manuscript.
