And drepanosaurs might fly... wait, really?

Minor update (06/07/16): Thanks to Andrea Cau, a few additional citations and points of discussion have been made below - the thrust and arguments of the post are the same, but the context is improved. Thanks Andrea!

Hypuronector limnaios restored as a glider. Have palaeontologists been smoking something of variable legality, or is there some basis to this?

Assuming you've reached level 5 of palaeontological geekdom you can't fail to know of the exceptionally weird Triassic clade Drepanosauromorpha. These generally small, long-bodied reptiles are largely, but not incontrovertibly, thought to nest at the base of Archosauromorpha (so between lizards and crocs in the landscape of modern animals) and are famous for their highly aberrant anatomy. Gracile, bird-like heads and necks sit atop long, robust and tubular bodies with deepened tails and stout limbs. The hands and feet are highly modified in each species, some bearing powerful claws, others having chameleon-like opposable digits. The end of their tails are modified into either grasping, prehensile organs or sharp hooks, these being interpreted as adaptions for anchoring the tail to vegetation or substrata. Exactly what drepanosaurs did for a living has long been a subject of discussion among academics, and they are nowadays generally considered arboreal or fossorial - or a blend of both. They're pretty awesome animals.

Because the Triassic was evolution's drug-fuelled, rebellious college days, it can't be considered shocking to learn that there's a drepanosaur species which is to drepanosaurs what they are to everything else. This distinctive, strange, and controversial species is Hypuronector limnaios (above). Reasonably good fossils of this small (c. 12 cm long) animal have been known for decades from upper Triassic deposits of New Jersey, but it received its name only relatively recently (Colbert and Olsen 2001). Hypuronector is often regarded as a swimming creature because of its dorsoventrally expanded, 'leaf-shaped' tail which lacks a hooked or prehensile termination (Colbert and Olsen 2001). Its tail is remarkable for the enormous chevrons (prongs of bone projecting downwards from the underside of tail vertebrae) which extend far below and behind their vertebra of origin to create the majority of the tail depth and its 'leaf-like' profile. Some authors have likened the outline of the tail skeleton to the body shapes of gymntoid or gymnarchid fish and suggested that it propelled Hypuronector through the deep, freshwater lakes its fossils were buried in, perhaps in a newt- or crocodile-like fashion (Colbert and Olsen 2001). Although possessing unusually long legs relative to other drepanosaurs and swimming animals, it's been argued that these were also related to an aquatic lifestyle. Specifically, it's suggested that they held the long, deep tail off the ground during terrestrial bouts, the tail apparently being incapable of elevation at its base (Colbert and Olsen 2001). This aquatic Hypuronector hypothesis has been around for some time. The animal was informally known as the 'deep tailed swimmer' in the 1980s (Fraser and Renesto 2005) and this moniker was transferred more or less entirely to its scientific name in 2001: loosely translated, Hypuronector means 'deep-tailed lake swimmer'.

Hypuronector limnaios skeletal reconstruction, from Renesto et al. 2010. Scale represents 10 mm.

At first glance at least, none of this sounds too outlandish: the tail of Hypuronector certainly has an oar-like shape, and we all know that lateral undulation of a tail is the commonest means of water propulsion for vertebrates. But there are other interpretations of Hypuronector which suggest it may not have been a swimmer at all. These alternative views suggest it was more like other drepanosaurs in being suited to climbing but, more remarkably, possibly a glider (Renesto et al. 2010). Sharing early versions of my gliding drepanosaur art (above) suggests that the latter hypothesis is not well known, even among experts. However, I want to stress from the outset that this is not All Yesterdays-style artistic speculation or the bizarre opinion of a 'fringe' worker. Challenges to the aquatic Hypuronector concept and suggestions that Hypurnoector was a more 'typical' arboreal form have been made by several authors (e.g. Senter 2004; Spielmann et al. 2006; Renesto et al. 2010; Castielloa et al. 2015), and the notion that it may have been a glider has been raised on reasonable (if perhaps not yet conclusive) evidence (Renesto et al. 2010). It follows older suggestions that some drepanosaurids - Megalancosaurus specifically - were gliders (see below; Ruben 1998; Renesto 2000) and, though this all might seem bizarre, there is some genuine scientific basis to it.

The aquatic Hypuronector hypothesis under scrutiny

Aquatic drepanosaurs are were first proposed in the early 90s (Berman and Reisz 1992) and quickly received criticism from drepanosaur workers (see Renesto 2010 for history). Hypuronector perhaps remains the best candidate for an aquatic, or at least amphibious species because of its unusual tail, but somewhat ironically, it's actually this paddle-shaped organ which seems to be the main problem for this hypothesis.

Holotype of Hypuronector limnaios, a partial skeleton with the 'paddle tail' (left), disarticulated torso and bits of limb and limb girdle. From Colbert and Olsen (2001).

One thing we should address straight out is that the resemblance of the Hypuronector tail to the body of certain fishes is not a the best endorsement for swimming habits. Fish do not swim using their whole bodies (the front end of any undulating swimmers needs to be stiff), and the gymntoid or gymnarchid fish likened to the Hypuronector tail don't really move their bodies at all when swimming. Rather, they propel themselves with oscillations of long, low fins along the top of bottom of their bodies. Thus, they may be a poor shape analogue for a sculling organ, and we're better off looking at the fins and paddles of swimming animals, not their entire bodies, for clues about the aquatic potential of the Hypuronector tail.

It stands to reason that Hypuronector would have swum like a crocodylian, newt or swimming lizard, where waves of lateral undulation in the tail generate forward thrust (Colbert and Olsen 2001). This requires tail anatomy which can accommodate a lot of lateral motion, and it's here that Renesto et al. (2010) suggest we hit a major issue. The caudal vertebrae of Hypuronector seem to permit some movement at the base and tip of the tail, but the mid-tail was pretty stiff. This is because the zygaopophyeses - processes of bone that overlap neighbouring vertebrae to guide their motion - are very long and have steep articular surfaces (below). In simple terms, they seem to have 'clamped' their adjacent vertebrae rather than - as expected for an undulatory tail swimmer - provided flat, horizontal surfaces for the vertebrae to slide over.

Further rigidity is provided by those amazing chevrons (Renesto et al. 2010). These rearward-projecting bones underlie the articulations of the adjacent 7-8 vertebrae, meaning any lateral motion at the vertebral joints had to overcome the stiffness of the 7-8 bony rods hanging beneath them. Although thin bones are somewhat compliant and the Hypuronector chevrons may have been flexible to a degree, it's difficult to see their arrangement as optimised for sculling habits: they may made the tail more paddle shaped, but to obvious detriment of tail flexibility and sculling potential. Indeed, we have to note that this configuration is very similar to biological structures adapted to resist bending. Tetrapod wings are a good example: the arrangement of bat fingers, pterosaur structural fibres and bird feather shafts with respect to the wing bones echoes the chevron distribution in Hypuronector. By contrast, deep-tailed swimmers, like crocodylians and newts, have chevrons which are short, robust, and do not significantly underlie neighbouring vertebrae. They are ideal structures for anchoring tail musculature, increasing tail depth and not interfering with tail motion. I have to agree with Renesto et al. (2010) that the potential of the Hypuronector tail as a swimming organ seems limited.

Hypuronector limnaios posterior trunk (left) and tail base (right) - note elevation of the latter with respect to the former, and the significant overlap of the zygapophyses. From Renesto et al. 2010, scale represents 10 mm.

Of further relevance here are the limbs of Hypuronector, which do not have obvious aquatic signatures. Aquatic, or even semi-aquatic animals tend to have proportionally short, squat limbs, often with expanded, paddle-like bones. But the limbs of Hypuronector are elongate, gracile and hollow (Renesto et al. 2010). Its hands and feet are not well known and variably interpreted, but the elements we have suggest that they were not paddle-like. Colbert and Olsen (2001) proposed that the limbs of Hypuronector were long to lift the tail from the ground when it left the water, their work suggesting that the vertebral column was too stiff to lift the tail on its own. But this can be seen as problematic for three reasons. Firstly, as pointed out by Renesto et al. (2010), articulated fossils of Hypuronector show the tail arcing upwards with respect to the trunk vertebrae (above): this is not thought to be taphonomic or diagenetic distortion. Secondly, the forelimbs of Hypuronector are somewhat longer than the hindlimbs, which is perhaps the opposite of what we would expect if dragging the tail was a concern - surely the body would tilt backwards with this arrangement? Thirdly, since when did reptiles, aquatic or otherwise, care about dragging tails? We need to be careful that we're not providing 'empty support' for hypotheses by inventing problems for our fossil animals to solve.

Maybe Hyperonector isn't 'the weirdo drepanosaur 'after all?

Taken collectively, these points about tail shape, tail arthrology and limb size must be viewed as problematic for the aquatic Hypuronector hypothesis, and maybe we should see if there are other interpretations of Hypuronector lifestyle which are more in tune with its anatomy. A good strategy for understanding strange fossil animals is putting the controversial, weird bits of anatomy to the side and first focusing on the more reliably interpreted components. With that said, let's ignore the controversial tail of Hypuronector for a moment and look at its limbs, limb girdles and trunk anatomy. As with all drepanosaurs, the shoulder and hip bones of Hypuronector are very tall and somewhat reminiscent of the limb girdles of chameleons (Renesto et al. 2010). It is thought both limb sets were highly mobile, although the drepanosauromorph fusion of the pectoral girdle into one solid structure, as opposed to having two separate halves like chameleons, would limit forelimb reach somewhat. The limbs were likely held in a sprawling pose and, because the femora and humeri are greatly elongated, Hypuronector likely had a wide, stable base to walk and stand on.

Bits and pieces of AMNH Hypuronctor specimens, including the only known cranial material (mandible, A-C) and the ventral view of a trunk and pectoral skeleton. Note the huge, curving ribs. From Renesto et al. 2010.

Hypuronector lacks the large, fused vertebrae over the pectoral region that we see in other drepanosauromorphs, but given that these likely reflect increased forelimb muscle mass and a reinforced pectoral region for digging and prey-capture (Castielloa et al. 2015), this may not impact locomotor mechanics too much. The trunk of Hypuronector was evidently powerfully muscled all the same, the tall neural spines of the dorsal vertebrae and the presence of large, curving ribs along the entire torso suggesting large muscles enveloped most of the body.

It can be seen that Hypuronector trunk and limb anatomy matches pretty well with what we see in other drepanosaurs: powerful torsos and mobile limbs that seem well suited to walking and climbing. We might view its limb elongation as an adaptation to climbing, the increased length of the upper limb segments simultaneously increasing stability and enhancing reach while also keeping the centre of mass close to the substrate. Perhaps more surprisingly, Hypuronector is also similar to other drepanosaurs in certain aspects of tail anatomy. Although its tail has a different overall shape and lacks the derived tail-tips of true drepanosaurids, it shares the specifics of drepanosaur tail motion - flexible base and tip, rigid mid-length - with the rest of the group (Renesto et al. 2010). So perhaps the tail of Hypuronector was just a simpler, oddly-shaped variant on the drepanosauromorph tail and used for similar purposes: stability when climbing (a simple prop can aid traction, balance and recovery from accident), a brace when rearing to dig and feed, or simply for showing off (Renesto et al. 2010).

Putting these lines of evidence together, several authors have started to interpret Hypuronector as a more 'typical' drepanosaur, albeit a less-specialised species that lived like a modern arboreal lizard rather than a reptilian tree pangolin or pygmy anteater (Spielmann et al. 2006; Renesto et al. 2010). If this is true, we might view the shape of its tail as a mechanical red-herring, something which seems more important to Hypuronector behaviour than it actually was. Perhaps it had no more significance to locomotion and behaviour than do the cranial ornaments of dinosaurs and pterosaurs, structures which most now agree were more to do with communication and display than the mechanics of day-to-day life.

Yes yes yes, but we're here for the gliding stuff

Taking this idea of a climbing, generalist Hypuronector a step further, Renesto et al. (2010) note that there are several features of Hypuronector which might indicate it was a patagial glider - that is, an animal with membranes extending between its limbs to facilitate slower falls from elevated positions or glide between perches. The chief features of interest here are the the elongate limbs and, in particular, the forelimbs being as long, if not slightly longer, than the hindlimbs. This configuration is uncommon among reptiles. Well known reptiles with disproportionately long arms include canopy-browsing herbivorous dinosaurs, completely aquatic lineages like ichthyosaurs, derived sauropterygians and turtles, and flying animals like pterosaurs. It's clear that the former animals are playing an entirely different game to drepanosaurs, but the basic similarity between pterosaurs - small, gracile boned creatures which probably had climbing and gliding ancestors - and Hypuronector might be a little more intriguing. Forelimb elongation occurs again and again in patagially gliding tetrapods - pterosaurs, cologus, scaly tailed gliders etc. - and it's not unreasonable to wonder if the same phenomenon in Hypuronector betrays the presence of gliding membranes. The limb proportions of this species are not so extreme as to think it was an exemplar glider and able to travel long distances from vertical starts, but they may have housed membranes of sufficient size to cushion the fall of these small animals if they jumped or fell from high places. The deep, rounded shape of the tail becomes something to pay attention to here as well, it perhaps being well-shaped to help 'correct' a tumbling Hypuronector into the right posture for a steady glide.

Which might have been handy if the initial glide trajectory was what glider pilots call 'less than ideal'

As noted above, at least Megalancosaurus has been also posited as a potential glider in the past (Ruben 1998; Renesto 2000). These conversations were inspired (at least in part) by long-defunct (if you could ever really consider them credible!) ideas that birds may have had shared, close ancestry drepanosaurs or drepanosaur-like animals - let's quickly duck aqay further discussion of that. But why has the idea of gliding Megalancosaurus not caught on? Although not ruled out entirely (Renesto 2000), gliding doesn't seem to have stuck with this species because it its spiked tail, highly mobile wrists and ankles, and grasping appendages suggest it was quite highly adapted to climbing. While climbing and gliding are not incompatible, it also lacks features like the long, gracile limbs we would expect from flighted animals. The anatomy of Hypuronector, by contrast, is a little more generalised and ticks enough boxes in the glider column to think it could be possible.

Of course, it's worth stressing that any gliding drepanosaur is hypothetical at this stage, but we should not take this as reason to dismiss the idea out of hand. In addition to the evidence mentioned above, consider that many, perhaps all drepanosauromorphs seem to have been climbers of one kind or another, and we know from extant faunas that the step from climbing to gliding is often a short one (Renesto 2000). It's really not crazy to think extinct lineages were any less able to develop gliding forms than our modern ones, and drepanosaurs were exapted for gliding flight in many ways. Their skulls had large brains and overlapping visual fields (Renesto and Dalla Vecchia 2005) (ideal for judging distance and processing flight data); they were generally small animals with hollow limb bones (lightweight); their torsos were stiffened and reinforced (aids stability); their limbs were powerfully muscled and highly mobile (control of aerofoils) and their deep, strong tails might be ideal rudders and stabilisers. And as bizarre as it may seem to be discussing the possibility of gliding in an animal only known from bones, recall that pterosaurs were identified as flying animals in the early 1800s long before we discovered fossil remains of their wing membranes: we can identify flying animals if we look carefully enough at their bones. The challenge now is to see if we can test these ideas, perhaps carefully comparing the limb anatomy and myological signatures of Hypuronector with other drepanosaurs, modelling the effects that crazy tail has on a falling animal and so on. We can also look for Renesto et al.'s membranes on Hypuronector fossils, examining them with UV light and being extra-careful when preparing future Hypuronector specimens: experience with other delicate reptile specimens shows that it helps to know where to expect soft tissue when removing matrix.

So there we go, then: the Triassic, and drepanosaurs, might have just got even weirder/cooler/complicateder/more frustratinger than we all knew. I'm thinking we need to hang out in the Triassic even more in future blog posts - check out this label for previous conversations on Triassic topics. And note that my new art book, Recreating an Age of Reptiles, has several pages dedicated to Triassic animals - including Drepanosaurus.

This blog glides on the gentle, supportive updrafts of Patreon

The paintings and words featured here are sponsored by the organisms almost as awesome as Hypuronector: my Patreon backers. Supporting my blog from $1 a month helps me produce researched and detailed articles with paintings to accompany them, and in return you get access to bonus blog content: additional commentary, in-progress sneak-previews of paintings, high-resolution artwork, and even free prints. For this post, we'll be taking a look at a (currently unpublished) painting of a more familiar drepanosaurid.

References

• Berman, D. S., & Reisz, R. R. (1992). Dolabrosaurus aquatilis, a small lepidosauromorph reptile from the Upper Triassic Chinle Formation of north-central New Mexico. Journal of Paleontology, 66(06), 1001-1009.
• Castiello, M., Renesto, S., & Bennett, S. C. (2015). The role of the forelimb in prey capture in the Late Triassic reptile Megalancosaurus (Diapsida, Drepanosauromorpha). Historical Biology, 1-11.
• Colbert, E. H., & Olsen, P. E. (2001). A new and unusual aquatic reptile from the Lockatong Formation of New Jersey (Late Triassic, Newark Supergroup). American Museum Novitates, 1-24.
• Fraser, Nicholas C., and Silvio Renesto. Additional drepanosaur elements from the Triassic fissure infills of Cromhall Quarry, England. Virginia Museum of Natural History, 2005.
• Renesto, S. (2000). Bird-like head on a chameleon body: new specimens of the enigmatic diapsid reptile Megalancosaurus from the Late Triassic of Northern Italy. Rivista Italiana di Paleontologia e Stratigrafia (Research In Paleontology and Stratigraphy), 106(2).
• Renesto, S., & Dalla Vecchia, F. M. (2005). The skull and lower jaw of the holotype of Megalancosaurus preonensis (Diapsida, Drepanosauridae) from the Upper Triassic of Northern Italy. Rivista Italiana di Paleontologia e Stratigrafia (Research In Paleontology and Stratigraphy), 111(2).
• Renesto, S., Spielmann, J. A., Lucas, S. G., & Spagnoli, G. T. (2010). The taxonomy and paleobiology of the Late Triassic (Carnian-Norian: Adamanian-Apachean) drepnosaurs (Diapsida: Archosauromorpha: Drepanosauromorpha): Bulletin 46 (Vol. 46). New Mexico Museum of Natural History and Science.
• Ruben, R. R. (1998). Gliding adaptations in the Triassic archosaur Megalancosaurus. Journal of Vertebrate Paleontology, 18 (3), 73A.
• Senter, P. (2004). Phylogeny of Drepanosauridae (Reptilia: Diapsida). Journal of Systematic Palaeontology, 2(3), 257-268.
• Spielmann J. A., Renesto S. and Lucas S. G. (2006). The utility of claw curvature in assessing the arboreality of fossil reptiles.Bulletin of the New Mexico Museum of Natural History and Science 37: 365-368.