Megistotherium: Not a gigantic hyaenodont creodont. More like a very basal seal.

Traditionally described as a gigantic hyaenodont creodont,
Megistotherium (Savage 1973; Miocene; Fig. 1) nests in the large reptile tree (LRT, 1544 taxa) at the base of the clade within Carnivora that ultimately produced extant seals and extinct Palaeosinopa (Figs. 1,2). In the LRT creodonts are large marsupial predators, convergent with members of the clade Carnivora.

Therefore 
Megistotherium is also a sister to the Machaeroides clade (which gave rise to Stylinodon) both derived from the Kerbos and Gulo (wolverine) clades (which gave rise to terrifying short face bears, like Arctodus). So, several gigantic, fearless bear-like taxa arise from this branch within Carnivora.

Figure 1. Megistotherium skull in several views. It is 2/3 of a meter in length. Don't overlook the skull of tiny relative, Palaeosinopa with a 10cm skull length.

Figure 1. Megistotherium skull in several views. It is 2/3 of a meter in length. Don’t overlook the skull of tiny relative, Palaeosinopa with a 10cm skull length.

Megistotherium osteothlastes (Savage 1973; Miocene, 23mya; 66cm skull length) was originally considered a giant hyaenodontid creodont based on tooth data. Here, because it nests basal to Palaeosinopa and seals it was probably semi-aquatic. Premaxillary teeth were weak and disappearing. The jaw muscles were enormous judging by the widespread cheek arches and tall cranial crest. The large diameter canines were housed in large, laterally expanded maxillae. The braincase was narrow. Note the prefrontal and lacrimal are no longer fused to one another.

Earlier mistakes in nesting Megistotherium
may be assigned to an over reliance on dental traits, which tend to converge more often than traditionally realized, and to taxon exclusion, something the LRT tends to minimize due to its wide gamut, getting bigger every week.

Figure y. Palaeosinopa in situ with tail reconstructed from disturbed elements.

Figure 2. Palaeosinopa in situ with tail reconstructed from disturbed elements. This taxon provides clues to the post-crania of Megistotherium by way of phylogenetic bracketing.

Side note:
In overall size and general features, the skull of Megistotherium was similar, by convergence, to the giant Eocene elephant shrew, the mighty Andrewsarchus.

Figure 1. Harpagolestes macrocephalus compared to sisters Sinonyx and Andrewsarchus to scale.

Figure 1. Harpagolestes macrocephalus compared to sisters Sinonyx and Andrewsarchus to scale. Compare these elephant shrews to Megistotherium (Fig. 1)/

In the past,
several mammal taxa achieved gigantic proportions not found in today’s relatives.


References
Savage RJ 1973. Megistotherium, gigantic hyaeonodont from Miocene of Gebel Zelten, Libya. Bulletin of the British Museum (Natural History) Geology 22(7):483–511.

wiiki/Megistotherium

Resurrecting extinct taxa: Creodonta, Mesonychidae, Desmostylia and Gephyrostegidae

Taxonomy
“the branch of science concerned with classification, especially of organisms; systematics.”  Taxon: a taxonomic group of any rank, such as a species, family, or class.

The large reptile tree
(LRT, 1366 taxa) has resurrected several taxa (in this case, clades) long thought to be extinct.

Figure 1. Adding Sinopa to the LRT nests it here, between the extant quoll (Dasyurus) and the extant Tasmanian devil (Sarcophilus).

Figure 1. Members of the traditionally extinct Creodonta include the extant quoll (Dasyurus) and the extant Tasmanian devil (Sarcophilus).

Creodonta
According to Wikipedia: “Creodonta” was coined by Edward Drinker Cope in 1875. Cope included the oxyaenids and the viverravid Didymictis but omitted the hyaenodontids. In 1880. he expanded the term to include MiacidaeArctocyonidaeLeptictidae (now Pseudorhyncocyonidae), OxyaenidaeAmbloctonidae and Mesonychidae. Cope originally placed creodonts within the Insectivora. In 1884, however, he regarded them as a basal group from which both carnivorans and insectivorans arose. Hyaenodontidae was not included among the creodonts until 1909. Over time, various groups were removed, and by 1969 it contained, as it does today, only the oxyaenids and the hyaenodontids.

In the LRT, Oxyaena and Hyaenodon are members of an extinct clade. However, Sinopa is considered a hyaenodontid, and it nests between the extant quoll (genus: Dasyurus) and the extant Tasmanian devil (genus: Sarcophilus). Sarkastodon is considered an oxyaenid and it nests as a sister to Sarcophilus. So… either the quoll and Tasmanian devil are living members of the Creodonta, or we’ll have to redefine the Creodonta.

Figure 1. Rorqual evolution from desmostylians, Neoparadoxia, the RBCM specimen of Behemotops, Miocaperea, Eschrichtius and Cetotherium, not to scale.

Figure 1. Rorqual evolution from desmostylians, Neoparadoxia, the RBCM specimen of Behemotops, Miocaperea, Eschrichtius and Cetotherium, not to scale.

Desmostylia
According to Wikipedia: “Desmostylians are the only known extinct order of marine mammals. The Desmostylia, together with Sirenia and Proboscidea (and possibly Embrithopoda), have traditionally been assigned to the afrotherian clade Tethytheria, a group named after the paleoocean Tethys around which they originally evolved. The assignment of Desmostylia to Afrotheria has always been problematic from a biogeographic standpoint, given that Africa was the locus of the early evolution of the Afrotheria while the Desmostylia have only been found along the Pacific Rim. That assignment has been seriously undermined by a 2014 cladistic analysis that places anthracobunids and desmostylians, two major groups of putative non-African afrotheres, close to each other within the laurasiatherian order Perissodactyla.”

In the LRT, desmostylians are indeed derived from anthracobunids, which, in turn, are derived from hippos and mesonychids. Mysticeti, the clade of baleen whales are derived from desmostylians. So… baleen whales are extant desmostylians.

Figure 3. Four mesonychids to scale. Here Mesonyx, Anthracobune, Paleoparadoxia and Hippopotamus are compared.

Figure 3. Four mesonychids to scale. Here Mesonyx, Anthracobune, Paleoparadoxia and Hippopotamus are compared.

Mesonychidae
According to Wikipedia, “Mesonychidae is an extinct family of small to large-sized omnivorouscarnivorous mammals closely related to cetartiodactyls (even-toed ungulates & cetaceans) which were endemic to North America and Eurasia during the Early Paleocene to the Early Oligocene. The mesonychids were an unusual group of condylarths with a specialized dentition featuring tri-cuspid upper molars and high-crowned lower molars with shearing surfaces. They were once viewed as primitive carnivores, like the Paleocene family Arctocyonidae, and their diet probably included meat and fish. In contrast to this other family of early mammals, the mesonychids had only four digits furnished with hooves supported by narrow fissured end phalanges.”

In the LRT, mesonychids include hippos and baleen whales. So, they are extant mesonychids. On the other hand, Arctocyonidae includes Arctocyon, which nests in the unrelated marsupial clade, Creodonta (see above). Certain other traditional mesonychids, like Sinonyx and Andrewsarchus, are not mesonyhids, but nest with the elephant shrew, Rhychocyon, close to tenrecs.

Figure 1. Silvanerpeton and Gephyrostegus to the same scale. Each of the two frames takes five seconds. Novel traits are listed. This transition occurred in the early Viséan, over 340 mya. Gephyrostgeus is more robust and athletic with a larger capacity to carry and lay eggs.

Figure 1. Silvanerpeton and Gephyrostegus to the same scale. Each of the two frames takes five seconds. Novel traits are listed. This transition occurred in the early Viséan, over 340 mya. Gephyrostgeus is more robust and athletic with a larger capacity to carry and lay eggs.

Gephyrostegidae
According to Wikipedia, “Gephyrostegidae is an extinct family of reptiliomorph tetrapods from the Late Carboniferous including the genera GephyrostegusBruktererpeton, and Eusauropleura.”

In the LRT, Gephyrostegus is the last common ancestor of the Amniota (= Reptilia). So… gephyrostegids include all living mammals, archosaurs (crocs + birds) and lepidosaurs.

References

wiki/Gephyrostegidae
wiki/Mesonychidae
wiki/Desmostylia

Marsupial sabertooth taxa are polyphyletic

Traditionally
sabertooth marsupials nest together with other carnivorous marsupials in a clade Cope 1875 called Creodonta and Ameghino 1895 called Saprassodonta (back when creodonts were considered archaic placentals).

Figure 1. Patagosmilus to scale alongside Hadrocodium. These sabetooth taxa are not directly related to Thylacosmilus in the LRT.

Figure 1. Patagosmilus to scale alongside Hadrocodium. These sabetooth taxa are not directly related to Thylacosmilus in the LRT. Note the shallow rooted canine. Note the first molar is now the last premolar, contra the original drawing interpretation. The premaxilla is hypothetical based on phylogenetic bracketing and not scored. At a screen resolution of 72 dpi (standard) these are full scale images.

Patagosmilus goini (Fig. 1; Forasiepi and Carlini 2010) is a large sabertooth marsupial from the Middle Miocene of south America traditionally considered a sister to the more famous and distinctly different and more famous sabertooth from South America, Thylacosmilus (Fig. 2). After testing in the large reptile tree (LRT, 1361 taxa), Patagosmilus nests with the ultra tiny basal sabertooth metatherian, Hadrocodium (Figs. 1,3) from the Early Jurassic. Evidently transitional taxa have yet to be discovered (or tested).

Figure 2. Thylacosmilus skull. Note the deep maxillae in dorsal contact containing giant canine roots. These are not present in Patagosmilus.

Figure 2. Thylacosmilus skull. Note the deep maxillae in dorsal contact containing giant canine roots. These are not present in Patagosmilus.

A little backstory:
Thylacosmilus atrox (Riggs 1933; Miocene, 40-3 mya; 1 m long) was a leopard-like predator, but plantigrade. Thylacosmilus was a sister to Early Cretaceous Vincelestes. The premaxillary teeth are either absent or were taphonomivally lost here. The canines are enlarged to curved fangs. Their long roots extend above the orbits. The dentary has deep ventral processes that guide and protect the canines. The lower canines are round stubs. The coronoid process is small. The mandible was able to open nearly 90º.

Figure 3. Tiny Hadrocodium (Early Jurassic) nests as a sister to Patagosmilus (middle Miocene) in the LRT.

Figure 3. Tiny Hadrocodium (Early Jurassic) nests as a sister to Patagosmilus (middle Miocene) in the LRT.

Hadrocodium wui (Luo, Crompton and Sun 2001; Early Jurassic; skull length: 1.2cm), known only from a tiny skull about the size of a thumbnail, Hadrocodium was originally considered a juvenile basal mammal, but later a tiny adult. Hadrocodium has a relatively larger brain size and more advanced ear structure than MegazostrodonHadrocodium nests with other basal metatherians with three (not two) molars, Morganucodon and Volaticotherium. The first molar was originally considered the third premolar.

Figure 4. Vincelestes soul showing the separation of the nasals and frontals by the conjoined maxillae housing giant canine roots, as in sister Thylacosmilus.

Figure 4. Vincelestes soul showing the separation of the nasals and frontals by the conjoined maxillae housing giant canine roots, as in sister Thylacosmilus.

Contra traditional studies
none of these taxa are related to marsupial creodonts and/or borhyaenids, like Borhyaena and Hyaenodon (Fig. 5), all of which have large canine teeth, none of which have saber teeth.

Figure 5. Subset of the LRT focusing on the Metatheria. Here sabertooth Patagosmilus nests far apart from sabertooth Thylacosmilus, which nests apart from the clade of borhyaenid marsupials.

Figure 5. Subset of the LRT focusing on the Metatheria. Here sabertooth Patagosmilus nests far apart from sabertooth Thylacosmilus, which nests apart from the clade of borhyaenid marsupials. A red square is placed next to Carnivora to indicate the presence of Smilodon, Haplophoneus and other sabertooth cats.

Oddly,
elephants, walruses, deer and other taxa with hyper-elongated teeth are never considered sabertooth proboscidians, sabertooth seals and sabertooth deer. Perhaps this is so because such teeth have a round cross-section, not a narrow, sword/saber-like morphology.

References
Ameghino F 1892. (Issued in 1894.) Enumeration synoptique deses pesces de mammiferes fossilesdes formation sócénes de Patagonie. Boletindela, Academia Nacionalde Cienciasen Cordoba, XIII, p.259 (p.108 in reprint).
Cope ED 1875. On the Supposed Carnivora of the Eocene of the Rocky Mountains. Proceedings of the Academy of Natural Sciences, Philadelphia. pp. 444–449.
Forasiepi AM and Carlini AA 2010. A new thylacosmilid (Mammalia, Metatheria, Sparassodonta) from the Miocene of Patagonia. Zootaxa. 2552, ss. 55–68.

wiki/Patagosmilus
wiki/Volaticotherium
wiki/Morganucodon
wiki/Hadrocodium

The Creodonta revisited in the LRT

Keep in mind the concept of convergence
whenever reviewing purported members of the Creodonta. Several purported creodonts have been added to the large reptile tree (LRT, 1342 taxa) recently.

According to Wikipedia,
“Creodonts were the dominant carnivorous mammals from 55 to 35 million years ago, peaking in diversity and prevalence during the Eocene.”

McKenna1975 considered the Creodonta
the sister taxa to the Carnivora within the clade Ferae (Carnivora + Pholidota (= pangolins)). The LRT finds pretty much the same relationship, but with creodonts on the marsupial side of the node and carnivores on the placental side of the node. The arboreal didelphid Caluromys is the only taxon that nests between marsupial creodonts and placental carnivores at present. (In the LRT pangolins nest with currently dissimilar bats. Ancestors of both, Chriacus and Zhangheotherium were much more similar.)

Halliday et al. 2015 nested creodonts
as sisters to pangolins in a cladogram that bore little to no resemblance to the LRT.

Figure 1. Oxyaena, a traditional creodont. This is a cat-like member of the carnivorous Marsupialia.

Figure 1. Oxyaena, a traditional creodont. This is a cat-like member of the carnivorous Marsupialia.

According to Wikipedia
“Creodonta
 was coined by Edward Drinker Cope in 1875. Cope included the oxyaenids and the viverravid Didymictis but omitted the hyaenodontids. In 1880. he expanded the term to include MiacidaeArctocyonidaeLeptictidae (now Pseudorhyncocyonidae), OxyaenidaeAmbloctonidae and Mesonychidae. Cope originally placed creodonts within the Insectivora. In 1884, however, he regarded them as a basal group from which both carnivorans and insectivorans arose. Hyaenodontidae was not included among the creodonts until 1909. Over time, various groups were removed, and by 1969 it contained, as it does today, only the oxyaenids and the hyaenodontids.”

Figure 1. Hyaenodon horrid us was the size of a large dog. This carnivorous marsupial was formerly considered a creodont.

Figure 2. Hyaenodon horridus was the size of a large dog. This carnivorous marsupial is considered a traditional creodont.

The LRT recovers
members of the traditional Creodonta in the carnivorous clade of the Marsupialia (Fig. 3). Earlier we looked at similar situation with members of the Didelphidae.

Figure 1. Subset of the LRT focusing on Basal Mammalia including Creodonta.

Figure 3. Subset of the LRT focusing on Basal Mammalia including Creodonta. Members of the Didelphidae and Creodonta are sprinkled throughout this subset.

Here (Fig. 3) the traditional creodont Sinopa (Fig. 4) nests with the extant dasyurids, Dasyurus and Sarcophilus (Fig. 4).

Figure 1. Adding Sinopa to the LRT nests it here, between the extant quoll (Dasyurus) and the extant Tasmanian devil (Sarcophilus).

Figure 4. Adding Sinopa to the LRT nests it here, between the extant quoll (Dasyurus) and the extant Tasmanian devil (Sarcophilus).

References
Andrews CW 1906. Descriptive Catalogue of the Tertiary Vertebrata of the Fayum, British Museum.
Cope ED 1880. On the Genera of the Creodonta. Proceedings of the American Philosophical Society. 19(107): 76–82.
Halliday TJD, Upchurch P and Goswami A 2017. Resolving the relationships of Paleocene placental mammals. Biological Reviews: n/a–n/a. doi:10.1111/brv.12242. ISSN 1464-7931.
Matthew WD 1901. Additional Observations on the Creodonta.  Bulletin of the American Museum 14:1.
McKenna MC 1975. Toward a phylogenetic classification of the Mammalia. Pp. 21–46 in Luckett WP and Szalay FS. Phylogeny of the Primates. New York: Plenum.
Morlo M, Gunnell G and Polly PD 2009. What, if not nothing, is a creodont? Phylogeny and classification of Hyaenodontida and other former creodonts. Journal of Vertebrate Paleontology 29(Supplement 3): 152A.
Polly PD 1994. What, if anything, is a creodont?. Journal of Vertebrate Paleontology. 14: 42A.
Sinclair WJ 1905. The Marsupial Fauna of the Santa Cruz Beds, Proceedings of the American Philosophical Society 49:73.
Wortman JL 1901-1902. Eocene Mammalia in the Peabody Museum, pt. i. Carnivora,” American Journal of Science 11–14.

Encyclopædia_Britannica/Creodonta
wiki/Creodonta

A fourth tested ‘bear-dog’ also nests with hyaenas

Bone crushing jaws.
That’s what they say about the mighty and ferocious ‘bear-dog’ Amphicyon (Fig. 1). Unfortunately only one species (Amphicyon longiramus) of four tested bear-dogs actually nests with dogs and wolves in the large reptile tree (LRT, 1348 taxa). Two (Amphicyon galushi and A. idoneus) nest with the hyaena (genus: Crocuta). A fourth nests with Thylophoropsa carnivorous marsupial we looked at a few days ago here.

Figure 1. Amphicyon idoneus, from the Yale Museum fossil collection, nests not with dogs, but with hyaenas.

Figure 1. Amphicyon idoneus, from the Yale Museum fossil collection, nests not with dogs, but with hyaenas. The convergence is spectacular, but phylogenetic analysis splits these taxa three ways.

Amphicyon longiramus (Lartet 1836, Blainville 1841; Mid-Miocene to Pliocene, 16–9mya; up to 2.5m) is the wide ranging bone-crushing, bear-dog. This is a wastebasket taxon with some specimens (A. longiramus) related to dogs, others (A. galushi) related to hyaenas, while still others (lower right above) nesting with carnivorous marsupials, traditionally considered creodonts.

?Amphicyon idoneus (YPM VP.013134) was originally considered a bear dog, but here nests with A. galushi a type of hyaena.

From what I can gather
there are several dozen specimens attributed to Amphicyon. It would make a good PhD dissertation to untangle this wastebasket genus.

References
Blainville HM 1841. Osteographie et description iconographique des Mammiferes récentes et fossiles (Carnivores) 1, 2 Paris.
Lartet E 1836. Nomenclature des mammife`res et des coquilles qu’il a trouve´s dans un terrain d’eau douce pre`s de Simorre et de Sansan (Gers). Bulletin de la Socie´te´ Ge´ologique de France 7: 217–220.

wiki/Amphicyon

Sinopa: not related to Palaeosinopa

Sinopa rapax (Leidy 1871; Eocene-Early Oligocene; 50mya) was traditionally considered a creodont or sometimes ‘provivverine’ relative of Hyaenodon, but here nests between Dasyurus and Sarcophilus (Fig. 1). It is similar in size to these extant taxa. Not sure about those premaxillary teeth. The dentary canines are larger than the maxillary canines.

BTW,
otter-like Palaeosinopa (Matthew 1901) was first considered a creodont, then a fish-eating pantolestid insectivore (it does get crazy here), but now in the large reptile tree (LRT) nests with eutherian (placental) aquatic seals. like Phoca.

Figure 1. Adding Sinopa to the LRT nests it here, between the extant quoll (Dasyurus) and the extant Tasmanian devil (Sarcophilus).

Figure 1. Adding Sinopa to the LRT nests it here, between the extant quoll (Dasyurus) and the extant Tasmanian devil (Sarcophilus).

The marsupial Sinopa is not related
to Palaeosinopa, the ancestor to seals. Instead Sinopa is one more creodont nesting in the Marsupialia, still… very close to the basalmost clade in the Placentalia (Eutheria), the Carnivora. We’ll discuss creodonts soon.

References
Leidy J 1871. Remains of extinct mammals from Wyoming. Proceedings of the Academy of Natural Sciences Philadelphia1871:113–117.

wiki/Sinopa

Bishops enters the LRT

Figure 1. The dentary of Bishops compared to its Late Cretaceous sister, Asioryctes, which has fewer and larger premolars.

Figure 1. The dentary of Early Cretaceous Bishops (1.5cm long) compared to its Late Cretaceous sister, Asioryctes, which has fewer and larger premolars and one more molar.

The genus Bishops whitmorei
(Rich et al. 2001; Early Cretaceous, Australia; Fig.1) is represented by a small mandible with a high coronoid process, six premolars and only three molars. In the LRT it nests basal to the much larger carnivorous marsupials (= creodonts), starting with the wolf-sized Arctocyon. It is a sister to Asioryctes (Fig. 1) which is basal to the herbivorous marsupials of Australia.

What makes this important?
It is the only tiny creodont known. All others are dog to wolf-sized. Cenozoic descendants of Bishops include the following carnivorous marsupials: Thylacinus, Thylacosmilus, Borhyaena, Hyaenodon and Vincelestes.

References
Rich TH, Flannery TF, Trusler P. Kool L, van Klaveren NA and Vickers-Rich P 2001. A second tribosphenic mammal from the Mesozoic of Australia. Records of the Queen Victoria Museum 110: 1-9.

Oops! What’s wrong with this picture?

So far you’ve learned so much about the skeletons of vertebrates. Now, can you tell what is wrong with the published image below? It will be obvious once you know what to look for. Scroll down for the solution.

Figure 1. Can you tell what is wrong with this picture of a museum mount of Ernanodon published in Vickers-Rich and Rich 1993?

Figure 1. Can you tell what is wrong with this picture of a museum mount of Ernanodon published in Vickers-Rich and Rich 1993?

Earlier we looked at and nested the basal marsupial, Ernanodon (Figs. 1, 2). The museum mount published in Vickers-Rich and Rich 1993, has one glaring error. Can you spot it?

Ernanodon anteilos (Ting [Ding] 1979; Paleocene; 50 cm in length) was originally considered placental mammal, perhaps a primitive anteater, then regarded as a primitive pangolin, like Manis. Here Ernanodon nests with Hyaenodon and Deltatheridum as a creodont marsupial, sharing large canines with both.

The skull was robust with a jaw joint nearly as far back as the occiput. The claws were broad and long, ideal for digging. The tail was long, but very slender.

Figure 2. Here is the same museum mount repaired in Photoshop. The pelvis was originally installed backwards. Here the pelvis is correctly mounted.

Figure 2. Here is the same museum mount repaired in Photoshop. The pelvis was originally installed backwards. Here the pelvis is correctly mounted.

Answer
The pelvis of the museum mount was installed backwards. Here (Fig. 2) the pelvis has been flipped in Photoshop to its correct position.

References
Ding SY 1979. A new edentate from the Paleocene of Guangdong. Vertebrata PalAsiatica 17:57–64. [Chinese 57–61; English 62–64].
Vickers-Rich P and Rich TH 1993. Wildlife of Gondwana. REED, Chatswood, Australia. 276 pp.

wiki/Ernanodon

The Tasmanian wolf has a new sister: Hyaenodon

Here’s an enigmatic mammal
that has been aching to be nested in a recognized clade for over 150 years. It should not have been this difficult to nest Hyaenodon (Fig. 1).

Figure 1. Hyaenodon horrid us was the size of a large dog. This carnivorous marsupial was formerly considered a creodont.

Figure 1. Hyaenodon horrid us was the size of a large dog. This carnivorous marsupial was formerly considered a creodont.

It’s interesting to see how
Wikipedia plays down the affinities of Hyaenodon (Laizer and Parieu, 1838; Eocene-Miocene, Figs. 1-3): “a group of extinct carnivorous fossil mammals from Eurasia, North America and Africa…Some species of this genus were among the largest terrestrial carnivorous mammals of their time; others were only of the size of a marten.” The Wiki authors do not place Hyaenodon into the Eutheria nor the Metatheria. They don’t create a family tree for Hyaenodon. Most authors consider Hyaenodon a member of the Creodonta, a clade considered a ‘wastebasket’ by Wikipedia. That clade may have to be revised or deleted in the future, but at present only one creodont has been tested and it has dropped out.

About Creodonts and Carnivorans
Wikipedia reports, “creodonts and carnivorans were once thought to have shared a common ancestor, but given that different teeth are involved in making up the carnassials (both between creodonts and carnivorans and between the main groups of creodonts), this appears to be a case of evolutionary convergence. Creodonta was coined by Edward Drinker Cope in 1875. Cope included the oxyaenids and the viverravid Didymictis but omitted the hyaenodontids. In 1880. he expanded the term to include MiacidaeArctocyonidaeLeptictidae (now Pseudorhyncocyonidae), OxyaenidaeAmbloctonidaeand Mesonychidae.[12] Cope originally placed creodonts within the Insectivora. In 1884, however, he regarded them as a basal group from which both carnivorans and insectivorans arose.[13]Hyaenodontidae was not included among the creodonts until 1909.[14] Over time, various groups were removed, and by 1969 it contained, as it does today, only the oxyaenids and the hyaenodontids.”

Figure 1. Hyaenodon skull cast showing in orange the lacrimal - septomaxilla common to most, if not all, metatherians.

Figure 2. Hyaenodon skull cast showing in orange the lacrimal – septomaxilla common to most, if not all, metatherians.

When added to the large reptile tree (now 796 taxa)
Hyaenodon (chimaera taxon, based on several specimens and authors) nested with another carnivorous mammal, Thylacinus, the Tasmanian wolf, a basal marsupial. They scored nearly identically.

Unique for marsupials,
Thylacinus had largely cartilaginous epipubic bones with a highly reduced osseous elements. Perhaps that’s why epipubes were never found with Hyaenodon.

Figure 4. Thylacinus, the Tasmanian wolf, recently extinct, was a sister to Hyaenodon.

Figure 3. Thylacinus, the Tasmanian wolf, recently extinct, was a sister to Hyaenodon. Note the three molars and lack of epipubes.

In both taxa
only three molars were present. That’s one less than in most marsupials and the number typical, but not universal, in placentals. In both taxa the jugal extends nearly to the back of the skull where the jaw joint is. That’s a typical marsupial trait. Likewise, the septomaxilla appears on the snout (Fig. 2), as in the marsupials Vombatus and Vintana. The occiput (Fig. 4) is also very metatherian.

Figure 2. Hyaenodon occiput. Note the strong resemblance to the occiput of Vintana, with the post parietal above the supraoccipital and framed by tabulars and tall squamosals.

Figure 4. Hyaenodon occiput. Note the strong resemblance to the occiput of Vintana, with the post parietal above the supraoccipital and framed by tabulars and tall squamosals.

This link to Scott (1895, p183)
discusses “M. Gaudrey’s (1878) view as to the marsupial  character of the genus [Hyaenodon] is definitely disproved by the abundant material now at command.” Good job, Gaudrey! You’ve been vindicated! (…about 150 years too late, unfortunately).

Figure 6. Subset of the large reptile tree: the marsupials featuring the newest taxon, Hyaenodon.

Figure 6. Subset of the large reptile tree: the marsupials featuring the newest taxon, Hyaenodon.

Hyaenodon species
According to Wikipedia, H. gigas was the size of a bear (est. 1,100 lbs, 500 kgs, 3 m). H. horridus was the size of a large dog (est. 88 lbs, 40 kgs). H. microdon and H. mustelinus were much smaller, about the size of Eomaia, another basal marsupial. There were several mid-sized taxa, too. Hyaenodon leptorhynchus was the type species.

Once again,
this discovery was made without ever having seen the fossil first hand. The LRT and a computer monitor are all the tools one needs in many cases, such as this one.

References
Gaudrey A 1878. Les enchaînements du monde animal dans les temps géologiques  mammifères tertiaires. F. Savy. (ed.) Paris 28pp.  online here.
Laizier L and de Parieu J 1838. Description et determination d’une machoire fossile appartenant a un mammifere jusqu’a pressent inconnu, Hyaenodon leptorhynchus. Comptes-rendus hebdomadaires des séances de l’Académie des Sciences, Paris 7:442.
Scott WM 1895. The osteology of Hyaenodon. Academy Natural Sciences Philadelphia Journal 9:499-536. online here.

wiki/Hyaenodon