Adorable Triassic Pseudosuchians

Erpetosuchus sp. from North America

Erpetosuchids are one of those groups that I saw reference to somewhere, filed it away in the recesses of my brain, and didn’t remember again until I read the new paper by Nesbitt et al. (2018) about the brain and endocast of Parringtonia gracilis. There’s a gorgeous skeletal restoration on page 124 and that skull looked awfully familiar to me. Oh yes, it looks a whole lot like the skull of Erpetosuchus grandi, a taxon I first read about thanks to Benton & Walker (2002) probably a decade ago. With a little more digging, I found that the Erpetosuchidae is a small but charismatic family of mid-sized pseudosuchians with unusual dentition and surprisingly croc-like armor. This is one of those small, obscure groups of Triassic hellasaurs that I like to put in the spotlight, if even briefly.

Little Erpetosuchus grandi was named and described in 1893 by E. T. Newton and redescribed in 2002 by Benton & Walker. Newton recognized that Erpetosuchus was related to crocodilians but considered it closer to phytosaurs or aetosaurs. While the original specimen was found in northeastern Scotland, a partial skull has turned up in Connecticut (Oslen et al. 2000), which may represent a different species. Newton’s fossil represented essentially the front third of the animal, including a complete skull. The skull and mandible, however, was a natural mold in the rock and Newton used gutta percha material (which is what golf balls are made of) to produce casts from these molds. The results are extraordinarily detailed, even capturing fine details of the outer surface of the skull bones.

The skull of Erpetosuchus from Benton & Walker

The skull of Erpetosuchus is interesting in a number of ways. For me, the most surprising detail is how few teeth it has. There are four small, recurved teeth in the premaxilla and just four larger recurved teeth in the anterior-most section of the maxilla. The rest of the maxilla is toothless, although Benton & Walker suspect that at least two teeth were broken off behind the first four. The situation is similar on the dentary—several recurved teeth of varying size are found in the anterior section of the dentary, but then there is a shot gap before three larger recurved teeth appear. It’s not clear whether that gap represents an actual toothless gap or if there are missing teeth. However, it should be noted that the dentary has more teeth than the premaxilla and maxilla.

Erpetosuchus has a particularly large fossa surrounding the antorbital fenestrae. Both the maxilla and the jugal have distinct, large ventral portions that are flattened or at the very least diagonally-facing towards the outside edge of the skull. As a result of this, the palatine bones are narrow—packed in between the flattened ventral surfaces of the maxilla and anterior jugal. The ectopterygoids are quite deep, again due in part to the weirdly diagonally-directed ventral surface of the jugal. In lateral view, the orbits look a bit like an oval that’s in the process of tipping over, but in dorsal view they are more or essentially circles. If you squint, the back of the skull (in lateral view) looks a bit like that of my alligator skull in that the quadrate is strongly diagonally-directed, the squamosal overhangs it, and there’s a small gap between the quadrate and the quadratojugal. Unlike my alligator, however, the lower temporal fenestra of Erpetosuchus is enormous and deeply recessed.

Skeletal restoration from Benton & Walker

Erpetosuchus probably has seven cervical vertebrate—the transition between neck and torso is not clear. In any case, it had a very short neck. Its neck and back, however, were covered by two parallel rows of rectangular scutes. Two additional rows may have begun behind the pectoral girdle. Given the proportions of the pectoral elements, Benton & Walker suspect that Erpetosuchus was a bipedal cursor, and its dentition suggests an insectivorous diet. Their phylogenetic analysis (keep in mind this is 2002), Erpetosuchus was found to be a sister group of Crocodylomorpha among then-known croc-line archosaurs. The authors suggest that fragmentary African taxon Parringtonia may represent another erpetosuchid, but more material will be needed to make that determination. They also note that Germany’s Dyoplax was initially referred to the Erpetosuchidae but it has since been reassigned as a genuine basal crocodylomorph. Well, we’ll quickly be revisiting both of these genera.

Looks like a mummy, right? More like a natural mold.

Dyoplax arenaceus was initially described by Fraas in 1867. Unlike Erpetosuchus, Dyoplax is known from a single essentially complete skeleton. Fraas did not attempt to classify it, and over the years it wound up in various families—the Aetosauria in 1902, the Erpetosuchidae in 1961, and the Crocodylomorpha in 1998 after Lucas et al. redescribed it for the first time. In 2013, Maisch et al. re-evaluated the skeleton in light of Benton & Walker’s revision of Erpetosuchus. While virtually complete apart from most of the limbs and tip of the tail, the holotype of Dyoplax is, like Newton’s Erpetosuchus specimen, preserved as a natural cast. Unfortunately, this cast is not as finely-detailed as Newton’s, so many details of Dyoplax are lost or at least ambiguous.

A "best guess" Dyoplax skull from Maisch et al.

The skull is restored as roughly resembling that of Erpetosuchus, but shallower, with a less-exaggerated antorbital fossa. The authors don’t comment on the postcrania aside from noting that Dyoplax also has median osteoderms running down the back, and seem unique to Erpetosuchus among crurotarsians. Honestly, the specimen looks like a cheaply-made plaster cast so I’m not surprised nobody has been able to do a detailed description. Maisch et al. note that only Walker suspected that Dyoplax may be an erpetosuchid. Even Benton & Walker accepted Lucas’ reference to the Sphenosuchia. However, Maisch et al. point out several features that readily link Dyoplax and Erpetosuchus, but given the poor preservation of the specimen, the authors refrain from definitely assigning Dyoplax to the Erpetosuchidae but they strongly suspect that it belongs there.

More interestingly, Maisch et al. consider the possibility that Dyoplax is actually a German specimen of Erpetosuchus. If they are the same, then Dyoplax has priority, as it was described almost thirty years before Erpetosuchus.

“Both taxa are from at least roughly coeval beds, and they are also strikingly similar in size. This is particularly important because it excludes the possibility that observed differences may be attributable to ontogeny.”

However, after considering that differences in armor preservation may be taphonomic rather than taxonomic, the authors note that many of the differences in the two animals’ skulls—especially regarding the size and shape of the antorbital fenestra—appear to be genuine. Although it remains an intriguing possibility, Maisch et al. conclude that without better specimens of Dyoplax, synonymizing the two is premature.

Fragment of the skull table of Tarjadia

In 1998, Acrucci & Marsicano described a “distinctive new archosaur,” Tarjadia ruthae, from Argentina. Sadly, the holotype consists of six “almost complete” dorsal osteoderms, fragments of a few more, and all found in association with fragmentary vertebrae. They refer a bunch of skull fragments to the same taxon. Unfortunately, but maybe not surprisingly, the authors could not classify Tarjadia as anything other than a “basal crurotarsal archosaur.” This is another taxon we’ll be revisiting soon.

First, let’s take a hard left to the Doswelliidae, a group of basal archosauriforms that also had distinctive armor. Desojo et al. (2011) described Archeopelta arborensis, a Brazilian taxon known from a braincase, a long string of dorsal vertebrae, a sacrum with attached pelvis (and right femur), a right humerus and several osteoderms. Those authors found that Archeopelta, Doswellia, and Arcurri & Marsicano’s Tarjadia share two features related to the structure of the osteoderms. Admitting that the holotype of Tarjadia hardly provides a robust set of characters, Desojo et al. removed it from their analysis and found eight non-osteoderm characters that united Doswellia and Archeopelta.

Desojo et al. used a Doswellia skeletal in their Archeopelta description.

However, just two years later, in their description of Texan doswellid Ankylosuchus chinlegroupensis, Lucas et al. (2013) assert that most of the diagnostic characters of Archeopelta cannot be distinguished from those of Tarjadia, simply because Tarjadia is so poorly-represented. They thus determine that Archeopelta is a junior synonym of Tarjadia—at least until better remains of the latter are found....HINT HINT.

In his massive 2016 phylogeny of basal archosaurs, Ezcurra found that the two really do differ in important ways, resurrecting Archeopelta, and placing them all, once again, in a monophyletic Doswelliidae.

The maxilla of Parringtonia

Now before we get to the thrilling conclusion, let’s introduce Parringtonia gracilis, pretty much the only other erpetosuchid that was recognized as such relatively quickly. Huene (1939) described this fragmentary beasty from Tanzania. Represented by a worn-down maxilla, scapula, ischium, some dorsal and caudal vertebrae, and osteoderms, Parringtonia did not present a terribly exciting fossil and Huene could not decide whether it was an ornithosuchid, basal crocodylomorph, or new lineage. A couple decades later, Krebs (1965) noted the similarities between the scapulae of Parringtonia and Erpetosuchus. Walker (1970) ran with that idea, including both—almost with Dyoplax—in a new family, the Erpetosuchidae, but six years later, Krebs removed Dyoplax and, as you’ll recall, Benton & Walker (2002) were not convinced that Parringtonia really was an erpetosuchid. So, once again, Erpetosuchidae is left with just Erpetosuchus.

Tarjadia revealed! Yellow bones remain unknown.

Now then, let’s take a quick detour back to Tarjadia. As reported in Ezcurra et al. (2017), new specimens of Tarjadia now represent almost every bone of the body excluding only the feet, hands, and most of the tail. And it looks an awful lot like Erpetosuchus. In their analysis, Tarjadia joins Erpetosuchus, Dyoplax, Archeopelta, and Parringtonia in a monophyletic Erpetosuchidae, which is the sister group of Ornithosuchidae.* Erpetosuchidae + Ornithosuchidae is the sister group to the heavily-armored, snub-nosed Aetosauria. I should note, however, that Nesbitt & Butler (2013) could not determine exactly where Erpetosuchidae fell on the Archosauria tree, and could be either a very early-diverging member or fairly derived relative of aetosaurs and Revueltosaurus. Based on braincase characters, Nesbitt et al. (2018) found Erpetosuchidae as an outgroup to (Aetosauria + Revueltosaurus)—essentially the earliest-diverging suchians.

It turns out that Huene’s Parringtonia is in the same camp as Tarjadia: Nesbitt et al. (2018) report that new specimens of Parringtonia have revealed virtually its entire skeleton and guess what? It’s very clearly an erpetosuchid.

Parringtonia revealed! Look at how big its head is compared to the body.

And finally, literally as I was getting read to press "Publish," another erpetosuchid appeared in the literature: Pagosvenator delariensis from Brazil, described by Lacerda et al (2018). The holotype is a mostly complete but pretty beat up skull with mandibles and two incomplete vertebrae and five osteoderms. Like its relatives, Pagosvenator isn't particularly large but the skull looks a whole lot like every other erpetosuchid where good skull material is known.

The authors state that Pagosvenator is the first erpetosuchid from South America, but that doesn't really jive with the existence of Tarjadia (Argentina) or Archeopelta (also Brazil). In fact, if you look through their paper they only mention Tarjadia once, where they compare one of its osteoderms to that of Pagosvenator, and Archeopelta twice--once in the same osteoderm context and a second time where the authors call it a doswellid, citing Desojo et al. (2011). Clearly, this paper must have been in press while Ezcurra et al. (2017) and Nesbitt et al. (2018) were being prepped. The authors run a phylogenetic analysis to try and work out where erpetosuchids (minus Tarjadia and Archeopelta) fit into the rest of Archosauria. In general, their results reflect other recent phylogenies: they're either a sister group to Ornithosuchidae or they're one step above Aetosauria, as a sister group to (Gracilisuchidae + Paracrocodylomorpha).

Offhand, one wonders if Pagosvenator is potentially a junior synonym of Archeopelta, as both are Brazilian and the latter is so poorly known. Desojo et al. (2011) write that Archeopelta is Landian or Carnian--and so is Pagosvenator. Lacerda et al. (2018) compare the osteoderms of Archeopelta and Pagosvenator but make no other osteological comparisons or even bring Archeopelta up, content to leave it as an unrelated doswellid. However, I think it's possible that the two are synonymous now that Ezcurra et al. (2017) have made a good case for Archeopelta being an erpetosuchid. As I always wind up saying in these essays, more material will be necessary for a more conclusive, uh, conclusion.

Now, what kinds of animals are these cute little crocs? While Benton & Walker believed (based on the proportions of the forelimbs) that Erpetosuchus could have run around on its hind limbs if it wanted to, Parringtonia and Tarjadia present much more traditionally-proportioned quadrupeds. They are also clearly impressively armored. In addition to the osteoderms running down the back, Tarjadia at least has osteoderms on its hind limbs and possibly its forelimbs. In Argentina, Tarjadia shared its habitat with rhynchosaurs, small mammals, and mid-to-large-sized carnivorous pseudosuchians. These were dog-sized cursors going after small game.

There are several features of the skull that I find interesting: the enormous fossa surrounding the anteorbital fenestrae; the enlarged, posteriorly-directed jugal (which gives the lateral temporal fenestra an odd shape); and the fact that the dentary has significantly more teeth than the maxilla. It's also strange that the group is so homogeneous--everyone is roughly the same size, with roughly the same skull shape and dentition, and differ mainly in terms of osteoderm compliment (assuming the differences between Tarjadia and Parringtonia are genuine and not the result of missing data).

 This group did not last terribly long: Parringtonia is the earliest member (Ansian) and most of them were extinct by the beginning of the Norian (although North America’s Erpetosuchus appears to be from the late Norian). Prior to the five-minutes-ago publication of Lacerda et al. (2018), a large ghost lineage existed between Parringtonia and everyone else, who turn up in the late Carnian. However, Pagosvenator is from the late Landian/early Carnian, which shortens the temporal gap somewhat. At any rate, erpetosuchids were certainly extinct by the Rhaetian, and thus were spared from whatever traumatic event caused the Triassic/Jurassic extinction.

*Funnily enough, in his original description of Erpetosuchus, Newton also described Ornithosuchus woodwardi. However, he thinks that the latter is a very primitive dinosaur or, at least, an immediate ancestor of dinosaurs.