Daphoenus is an extinct genus of amphicyonids, a group colloquially known as "bear-dogs". It includes not just some of the best preserved material out of any amphicyonid, but also the earliest members of the family, first appearing in the middle Eocene and surviving into the Early Oligocene. The members of the genus are rather small compared to some of its later relatives, such as Amphicyon or Ysengrinia, ranging in size from comparable to a house cat to a small wolf. It was widely distributed across North America, with most of its remains being discovered in the White River Group of the Great Plains, though the John Day Beds of Oregon and the Cypress Hills Formation in Saskatchewan also represent important fossil sites. Other specimens have been described from localities in Texas, Georgia, Florida, and California.
Taxonomy
The first fossils of Daphoenus vetus were collected by John Evans in 1849, before they were described by Joseph Leidy in 1853, making them among the first White River mammals to be named. One year later, however, Leidy synonymized his newly erected genus with Amphicyon, though it was later resurrected by Scott in 1898. Hough considered their body shape to be more feline than canine. Two decades after the description of D. vetus, the smaller D. hartshornianus was named by Edward Drinker Cope on the basis of fossils he discovered in Cedar Creek, Colorado. Originally, the various species of the genus were considered to be primitive members of the Canidae. It is the most common genus of North American amphicyonids, with over sixty skulls, various rostra, and a number of postcranial skeletons having been discovered.
Evolution and classification The origin of the amphicyonids is a highly debated topic, with their continent of origin remaining unresolved. This is partly a result of the plesiomorphic and homoplastic traits of early members of the family, as well as the poor fossil record of Middle Eocene Asia. As the genus Daphoenus includes the oldest known amphicyonid, the diminutive Daphoenus demilo from early Duchesnean Wyoming, it is integral to the understanding of the issue. The 2016 description of Angelarctocyon and Gustafsonia includes a review of the earliest Amphicyonidae, and a phylogenetic analysis that recovers Daphoenus as more derived than the aforementioned taxa, as well as the Eurasian Cynodictis. The authors also discuss the various proposals of the amphicyonid place of origin. They consider it unlikely that Europe was the cradle of their evolution, as their earliest securely dated appearance on the continent is Cynodictis lacustris from MP18 (~36-35 Ma). While Middle Eocene remains of Cynodoctis have been mentioned in faunal lists of the Lushi Basin and the Ulan Shireh Formation, which would make them the earliest known bear dogs, as well as a possible tooth of similar age from Myanmar, the validity of the assignment of these fossils to the Amphicyonidae could not be verified. Thus, the oldest known Asian occurrences of amphicyonids, belonging to Guangxicyon from China and an unnamed taxon from Mongolia, date to the Late Eocene and belong to animals more derived than both Angelarctocyon and Gustafsonia. Therefore, in the middle Eocene, amphicyonids are only known from North America, where they are represented by at least four different genera. This, in addition to the basal standing of A. australis and G. cognita in the phylogenetic analysis, supports a possible North American origin of the family. The lack of amphicyonid fossils from the late Uintan and early Duchesnean faunas west of the Rocky Mountains further supports this hypothesis, as they include some of the oldest known reports of other North American immigrants of Eurasian origin. Indeed, the earliest bear dogs are only known from the midst of the continent, rather than in high latitudes closer to Beringia. However, the position of the European Cynodictis as basal most amphicyonid in the consensus tree is puzzling, and may speak against a North American origin of the family.
The cladogram below follows the cladistic analysis of Tomiya & Tseng (2016).
Both the dental features and skull morphology of the earliest Daphoenus are reminiscent of Cynodictis, indicating a common origin between the two genera, though it differs from its European relative in the presence of a "closed" trigonid, the first two molars being rather robust, and the fact that its third molars are merely reduced, not absent. It is likely that Daphoenus was ancestral to the later daphoenines of North America, such as Daphoenictis, Paradaphoenus, and Daphoenodon, whereas Cynodictis gave rise to the diversity of Old World bear dogs. The daphoenines were a likely monophyletic group endemic to North America, and ranged from the middle Eocene to early Miocene epoch. Originally considered to be canids, Hough raised them to family status in 1948, including genera such as Amphicyon, evidently being unaware that Schlosser had created the family Amphicyonidae in 1888, before Hunt demoted them back to subfamily status in 1974, though as amphicyonids, rather than canids. Daphoenus was also closely related to Brachyrhynchocyon, and the later members of the genus, such as D. vetus & D. hartshornianus, probably descend from the early D. lambei.
Species D. demilo Described by Dawson on the basis of a jaw, a fragmentary maxilla and jaw fragment, this small, slender-jawed species was discovered at the Wood locality 20 and locality 20 of the Badwater Creek area in central Wyoming. It is more primitive than other species of the genus, as shown by the well-developed buccal shelf on its M1, a feature typical of miacids, as well as the less open trigonid of M1 and the greater development of the paraconid on M2. Furthermore, the gaps between its anterior premolars is smaller than in later members of Daphoenus. It dates to the early Duchesnean, making it the oldest known amphicyonid, with the Badwater 20 locality dating to ~40 Ma. A tooth possibly referable to this species is also known from the Lac Pelletier Lower Fauna of the Cypress Hills Formation in Saskatchewan. Hunt considers it to be conspecific with D. lambei.
D. vetus This is the most well-known species of the genus, being widely distributed across the White River faunas from ~36-35 Ma to 28.4 Ma, with the lineage continuing into the Whitneyan, where it is known from the Protoceras channels of South Dakota, represented by a form larger than the Orellan D. vetus, but smaller than the terminal Great Plains species. It is similar in size to a coyote, with a skull length of 17–20 cm.
D. hartshornianus In the White River Group, D. vetus coexisted with D. hartshornianus, a much smaller relative that reached perhaps 11 kg, and had a skull length of 14–16 cm. It was common throughout the Orellan, from ~36-35 Ma to ~32-30 Ma.
D. lambei This species is even smaller than D. hartshornianus, from which it also differs in its prominent deuterocone and well-developed cingulum. It is known from the Conglomerate Creek and Calf Creek valleys of the Cypress Hills Formation, Chadronian, Saskatchewan, and possibly the Texan Porvenir Local Fauna in the lower part of the Chambers Formation. Both of these localities date to the Chadronian. Hunt suggests that D. demilio might be a junior synonym of D. lambei, which would make the latter the oldest amphicyonid known.
D. ruber Discovered in the Tecuya beds of California, this species was described by Stock on the basis of a fragmentary left ramus. It is slightly smaller than D. vetus, and the original author suggested that it might prove to be a connecting link between early Daphoenus and D. socialis. Hunt does not mention the species in his review of North American amphicyonids, but agrees with the referral of the Tecuya specimem to Daphoenus.
D. transversus This species from the John Day beds is quite similar to D. hartshornianus, and its holotype was initially even assigned to the latter species. However, in 1899 Wortman and Matthew created Paradaphoenus transversus on the basis of the holotype rostrum, though later authors do not follow the assignment to this genus. Even though it is similar to D. hartshornianus in its size, frontal inflation less pronounced than in D. vetus, and a narrow snout, it differs in its more sectorial dentition, which is reminiscent of Oligocene canids. The dating of the fossils from the John Day beds is difficult, as the localities where many of them were discovered have been lost.
D. socialis Whereas D. tranversus occupies a similar morphospace as D. hartshornianus, D. socialis is more similar to D. vetus in its larger size, broader snout, and inflated frontal region, though it also deviates from its Great Plains relative in its dentition. It was originally described as Pericyon socialis by Thorpe, on the basis of a specimen from the Haystack valley area of the John Day beds. In addition to fragmentary remains, and several isolated teeth, it is also known from a partial cranium (YPM 10064).
Undescribed species In addition to the mentioned species, Daphoenus also includes two as of yet undescribed forms. Both of them are among the last and largest members of the genus. The first is known from a single cranium, discovered in the uppermost White River Group of Wildcat Ridge, located in western Nebraska, dating to 28.6 Ma, with a 24 cm long skull. Furthermore, a large maxilla with unusually big teeth (JODA 1411) is known from Fremd's Unit C from the Blue Basin, correlating with the earliest Arikareean, indicating that a third species of Daphoenus was present in the John Day beds.
Description
Daphoenus was a plesiomorphic, rather unspecialized early member of the amphicyonids. This genus varied greatly in size, though it was much smaller than many other bear dogs, with the earliest members being among the smallest known members of the family, reaching the size of a house cat (2–4 kg), while the common D. vetus was a coyote-sized, and the terminal species approaching a small wolf in size (20–30 kg).
It is characterized through a dolichocephalic skull, with a long snout, and narrow, elongated premolars, which increase in size posteriorily. Whereas in other amphicyonids the first upper molar as well as the premolars tend to be reduced, while the second and third molar form are enlarged to form a crushing platform, this is not the case among daphoenines. Daphoenus possessed generally unspecialized teeth, their shearing molars had a plesiomorphic pattern, as indicated by the prominent tricuspid trigonid with well-developed paraconid-protoconid shear and a distinct metaconid. Daphoenus tooth formula was . As with its teeth, the auditory bulla was among the most plesiomorphic of all amphicyonids, similar to Cynodictis and the modern African palm civet.
The postcranial skeleton of Daphoenus was gracile, with an anatomy possibly suggesting a scansorial animal, and even the earliest members of the genus show subcursorial adaptions, possessing a paraxonic digitigrade to subdigitigrade gate and the ability to evert and invert their hindfoot. Footprints belonging to the ichnogenus Axiciapes were most likely made by Daphoenus. If these animals were digitigrade, as indicated by their anatomy, suggests that their front feet possessed thick digital pads or semi-retractible claws as only the tips of their claws are preserved in the footprints. the tracks of their hind feet are puzzling, as no trace of their first digit is visible. This only supports the fact that their weight was spread unequally on its outer side, with digit I possibly being held clear of the ground while walking.
Paleobiology