Thursday 9 April 2015

The weird, awesome, and weirdly awesome, Triassic hindlimb-glider Sharovipteryx mirabilis

Sharovipteryx mirabilis, a tiny reptilian 'hindlimb glider' from the Late Triassic of Kyrgyzstan. A Triassic spider is thrown in for fun (spiders are a very ancient group, and were almost certainly present along the ancient lakes frequented by Sharovipteryx). Prints of this painting can be purchased here.
Sharovipteryx mirabilis is a mainstay of books on prehistoric animals, a Triassic Kyrgyz species mentioned frequently as a weird and wonderful, non-dinosaurian Mesozoic species. It is best known for its hindlimb-dominated approach to gliding, but also achieves some popularity via suggestions that it might be closely related to pterosaurs. This relationship is suggested by the possession of membranous flight organs in both lineages and some similarity in hindlimb structure. Recent studies have not looked favourably on this suggestion however, because the detailed anatomy of pterosaurs, distorted as it is by their flight adaptations, shows much greater similarity to that of dinosaurs and their immediate ancestors to that of Sharovipteryx. Moreover, with membranous gliding aids appearing and disappearing multiple times in the evolution of gliding species (including other fossil reptiles) their shared presence in Sharovipteryx and Pterosauria may have little evolutionary significance. Although the relationships of Sharovipteryx to other reptiles remain somewhat ambiguous (see below), most suggest they are part of Protorosauria, an early offshoot from the archosauromorphs which also includes weirdo taxa like Tanystropheus and the drepanosaurs.

The anatomy and proportions of Sharovipteryx are remarkable and unique. We only know of this species from a fossil discovered in 1965, later named and described in 1971 by Alexander G. Sharov. Initially called Podopteryx, it became Sharov’s namesake in 1981 when it was realised that a damselfly already existed with the former name. Our only specimen of Sharovipteryx is variably preserved – sometimes excellently, sometimes considerably less so. Although most of the skeleton is represented, some of the bones are crushed and the two bony slabs sharing the skeleton are split through the middle of some bones – most notably the skull. This has resulted in some controversy concerning detailed interpretations of Sharovipteryx anatomy and hindered discussions of its relationships to other animals. Some of these disagreements are significant, not the least being whether or not any material from the forelimbs is known. Some authors claim to have not only found elements of the arms, but enough to ascertain that that they were proportionally short. Others suggest there is no trace of them at all, and further interpretations suggest the arms were present, but damaged and modified during specimen preparation.

Holotype and only known specimen of Sharovipteryx mirabilis. From Gans et al. (1987).
Such details aside, many elements of Sharovipteryx appearance are relatively clear and can be crudely summarised as resembling a small (< 250 mm long, including tail), long-bodied lizard with enormous, membrane-bound legs. Skin impressions adjacent to the skull, torso and feet show the body was covered with small scales, some of which overlapped. The head was small (19 mm long) and narrow with a short snout, large eyes, small nostril openings and at least 15 small, sharp and widely-spaced teeth in each jaw. The neck and trunk are long and narrow, and of approximately equal length. The narrow tail was at least as long as the body and neck combined. Arm and shoulder bones, as indicated above, are at best very poorly known, or not known at all - depending who you ask. It is likely they were - like those of other protorosaurs - relatively short. It is a shame that we do not know them better: some protorosaur arms are adapted for unusual habits in truly bizarre ways, and given how strange many details of Sharovipteryx are, odd hand or forelimb anatomy would not be unexpected.

The legs of Sharovipteryx are, of course, the most striking feature of the animal. The crus is slightly longer than the thigh, with each bone being approximately as long, if not a little longer, than the torso. The five-toed feet are not especially slender however, although they are proportionally large and broad. Much of their size is devoted to the toes, which increase in size from digit I - V, while the body of the foot is rather short and robust.

Behind the legs and feet is a set of expansive, scale-less membranes extending from the base of the tail, along the posterior margin of the leg, and down to the large fifth toe. These membranes bear striations and folds radiating from the tightly-folded legs. The striations may be fibres within the membranes themselves, perhaps acting to control membrane flutter and aid neat membrane collapse, but the folds suggest Sharovipteryx membranes did not shrink entirely when the legs were folded in. Similarly folded membranes occur in some modern gliders, such as flying squirrels. It is widely thought that these membranes could be deployed as a gliding apparatus simply by extending the legs sideways, a posture clearly attainable in life given the fossilised pose of our sole Sharovipteryx specimen. It is unlikely that Sharovipteryx could flap its hindlimb wings however, its pelvis and hindlimb bones lacking suitable room and reinforcement for flapping muscle attachment.

Whether other membranes were present in front of the legs - or even along the arms - is debated. Some suggest they can be observed on the specimen, but this is not universally agreed on. There are aerodynamic reasons to suspect the posterior hindlimb membrane was not the sole flight surface, however. Studies modelling the glide performance of Sharovipteryx find that the centre of lift is located too far back along the body to safely glide and land without the aid of additional membranes. In other words, there is too much weight in front of the wing, and the animal would risk toppling forward in flight or landing at hazardous speeds. An anterior flight surface, perhaps along the front of the thigh or associated with the arms, would negate these issues and permit safer landings.

Alternative reconstructions of the flight apparatus of Sharovipteryx. Reconstruction 'd' not only provides the best glide path, but also makes flight safe enough to ensure happy landings. Scale bar is 20 mm. From Dyke et al. (2006).

How effective was hindlimb-dominated gliding? It is tempting to take the absence of this gliding mechanic from modern animals as a sign of inefficiency. After all, we see gliding and parachuting in fish, rodents, flying ‘lemurs’, frogs, snakes, lizards and more groups, and none rely on a gliding system approximating that of Sharovipteryx. Indeed, key similarities in gliding anatomies have developed within these groups, perhaps indicating certain 'optimal' gliding adaptations exist for vertebrates. For instance, both frogs and geckos glide on splayed tissues around their hands and feet. All gliding mammals sport extensive membranes between all four limbs. Gliding snakes and agamid lizards have mobile ribs which can expand their bodies into flattened aerofoils. Although the details of these structures differ – as would be expected given their development in such distantly related animals – it is nevertheless interesting that the same anatomical components have developed repeatedly into gliding organs. Perhaps the uniqueness of hindlimb-dominated gliding to Sharovipteryx indicates that it arose via an especially lax interval of natural selection and competition, and that it just doesn't cut the evolutionary mustard outside of the Triassic.

Flight models of the Sharovipteryx glide path conflict with this assessment, however. Indeed, some models of Sharovipteryx glide paths indicate more successful gliding abilities than skilled modern reptile gliders such as Draco, and potentially more manoeuvrability. The ability to control the principle flight membrane with movements of the legs and tail, along with additional (and hypothetical) assistance from forelimb membranes, likely conferred tight control over the glide path. It is notable that the uniqueness of Sharovipteryx gliding apparatus among animals is not carried over to human technology, its wing shape having been likened to delta wing aircraft - the sort of designs we see in fast, nimble fighter jets. In short, no experiments have suggested that the gliding apparatus of Sharovipteryx was ineffective or clumsy, and instead indicated quite the opposite.
Perhaps instead of viewing Sharovipteryx as an evolutionary oddball, we might wonder why more animals have not attained similar anatomies. I suspect the answer may lie in the need for a specific ancestral body form to become a hindwing-dominated glider. The legs must be relatively large to provide a suitable wing surface and possess greater potential for use as wings than the arms, as well as being able to project sideways. While there are plenty of animals with proportionally large hindlimbs, most are bipeds with hip joints particularly restrictive against lateral leg rotation. A small, lightweight body and head, probably reduced arms, long, balancing tail and perhaps climbing habits might also be ideal. It's difficult to think of species other than protorosaurs which might present such a combination of these characteristics, so the greatest potential for hindlimb gliders may have disappeared along with them at the end of the Triassic. Whatever the reason, the take home message is that Sharovipteryx as really a perfect recipe of genetic resources, adaptive pressures and animal behaviour to create a remarkable, unique and effective gliding animal, and not the result of an evolutionary wrong turn.

Of course, Sharovipteryx would not be gliding all the time: how did it move on land? Relatively little commentary exists on this point, but we might assume that terrestrial locomotion and climbing was used when foraging, perhaps for insects, given the absence of adaptations for aerial prey capture. Because the legs of Sharovipteryx are of such length, possibly much longer than the arms and support a large amount of soft-tissue, it is not unreasonable to wonder how walking and running was performed. Unfortunately, our lack of data on Sharovipteryx forelimb bones becomes a real issue here: modelling the terrestrial abilities of any extinct animal really has to start with knowing basic limb proportions. Protorosaurs seem to have been lizard-like quadrupeds, but the forelimbs of Sharovipteryx would have to be very long to operate in this fashion. Shorter forelimbs, as controversially-interpreted by some, do not strictly rule out quadrupedality, although they might require the adoption of frog- or rabbit-like hopping to work harmoniously with the enormous hindlimbs. Bipedality, of course, also cannot be excluded.

Quadrupedal or otherwise, the short, splaying foot of Sharovipteryx contrasts with the generally elongate feet of fast running animals. Despite its long limbs, Sharovipteryx was probably not adapted for routinely sprinting. This doesn’t completely rule out fast terrestrial locomotion of course – we see similar foot structure and limb proportions in rapidly running lizards like basilisks and frilled lizards – but we might assume it was not a regular part of Sharovipteryx habits. Climbing, however, like was: strongly built, asymmetrical feet with increasing lateral toe length like those seen in Sharovipteryx are common in climbing species. We might imagine Sharovipteryx as spending much of its time in tree canopies, and it is perhaps noteworthy that the fossil locality which yielded Sharovipteryx is also known for a diverse fossil flora. Artists might want to consider reconstructing Sharovipteryx with arms showing adaptations to climbing, given that climbing with hindlimbs alone might be very difficult - even for something as strange as Sharovipteryx. Strong feet may also be consistent with powerful leaping and impact absorption during landing, although – again – without knowing much about the Sharovipteryx forelimbs, reconstructing its landing cycle is difficult.


  • Dyke, G. J., Nudds, R. L., & Rayner, J. M. V. (2006). Flight of Sharovipteryx mirabilis: the world's first delta‐winged glider. Journal of Evolutionary Biology, 19(4), 1040-1043.
  • Gans, C., Darevski, I., & Tatarinov, L. P. (1987). Sharovipteryx, a reptilian glider?. Paleobiology, 415-426.


  1. "Perhaps instead of viewing Sharovipteryx as an evolutionary oddball, we might wonder why more animals have not attained similar anatomies." Would you consider azhdarchids to be something of a convergence? Sure, they were powered flyers and their flight surfaces were attached to the forelimbs, but they seem somewhat front-heavy at first glance, and approximate that delta-wing shape to my mind.

    Also, on an unrelated note, I saw this and immediately thought you might like to spread it around. Nothing beats a good meme-worthy image for getting the word out.

    1. I know of a couple of pterosaur workers who've compared azhdarchid planforms to delta wings in the past. I'm not sure I share that view because the wing may have been quite tapered towards the leg, at least based on what we see in pterosaur wing membrane fossils, preventing them from attaining a truly triangular shape. Plus, the front-heaviness of pterosaurs likely necessitated slightly more anteriorly-swept wings than most illustrations show, moving pterosaurs even further from delta-wing-like planforms.

  2. Mark you go to such lengths to avoid the literature and online resources for Sharovipteryx as a biped and as a relative not only to pterosaurs but to Cosesaurus, Kyrgzysaurus and Longisquama (its contemporary). If you have issues with the controversial bone identifications, ignore them. Use what everyone agrees to. You'll know which clade Sharovipteryx belongs to when you simply compare its feet to those of all other reptiles. You did not mention the elongate pelvis and extended sacrum (way more than two sacrals there). Also common to this clade, as are the attenuated tail and extra dermal membranes. Nice looking art though.

  3. Aeronautically, it seems odd to put delta wings on a gliding creature. I strongly suspect that it was a fairly unswept wing, simply because deltas are actually not very efficient at low Reynold's number and mach. For practical purposes, the mach number of any flying animal but a nosediving azhdarchid would be essentially zero. Even extremely optimistic estimates for azhdarchids full of dinner diving with wings folded in from obscene altitudes put them safely below the transonic drag regime at 5000 meters and under.