Friday 11 September 2015

The life aquatic with flying reptiles

Pteranodon sternbergi dives for a school of panicked fish. So, what, pterosaurs are super good at swimming now? Read on... Reworked version of an image from Witton (2013). Click here to buy prints of this image (and join my Patreon campaign for a discount!). And yes, I'm calling this animal Pteranodon, not Geosternbergia.

Whether or not pterosaurs could swim, or how well they could swim, is a recurrent discussion among those interested in flying reptiles. For the most part, palaeontologists have seemed happy to assume that pterosaurs were aquatically capable, at least long enough to permit their escape from water, because so many pterosaur fossils occur in coastal or marine sediments. Moreover, some long-known specimens show evidence of pterosaurs feeding on aquatic prey. Odds are that pteroaurs would end up in water some of the time, even if only by accident, so it makes sense that they could at least keep themselves afloat for a while. Plus, virtually all tetrapods can swim one way or another, including bats and bird species which, on first principles, seem ill-suited to aquatic locomotion. Pterosaurs might be a bit strange, but they'd have to be very strange not to be capable of at least limited aquatic locomotion.

Proof that pterosaur workers of old thought swimming was possible: Bramwell and Whitfield's (1974) landmark paper on Pteranodon flight depicts one attempting to take off from water.
In recent years, pterosaur researchers have taken a more in depth look at pterosaur swimming, with three main lines of inquiry offering perspectives on how, and how well, pterosaurs took to water. The first concerns pterosaur swim tracks, scratch marks made in upper Jurassic sediments of North America made by pterosaurs paddling across a shallow lake. First described by Lockley and Wright (2003), these tracks record pterosaur feet scraping narrow gouges into sediment, sometimes with toe pad impressions, as buoyant pterosaurs propelled themselves over lake margins. At least locally, such tracks are not rare: Lockley and Wright (2003) report bedding planes covered with hundreds of parallel scratch marks attributed to swimming pterosaurs. Toe pad impressions are only seen occasionally, suggesting that most of these track makers were more or less entirely supported by water, only the very tips of their toes scraping the lake bed. Notably absent altogether from the same slabs are imprints from pterosaur forelimbs. Neither wingtips or walking fingers left impressions when these pterosaurs were swimming or punting about. This helps us work out what these swimming pterosaurs might have looked like, as well as how they propelled themselves: their arms must been held higher than their legs, and at least some of their locomotion was achieved through peddling feet. Quite what this means for mysterious 'manus only' pterosaur tracks (pterosaur track sites where only hand impressions are recorded) is a discussion for another day.

Select examples of pterosaur swim tracks from the Jurassic Summerville (A) and Sundance (B) formations, North America. There are many more examples of tracks like this - some bedding planes are covered in them. Traced illustrations from Lockley and Wright (2003), published in Witton (2013).
Pterosaur swim tracks provide a compelling answer to the basic question over whether pterosaurs could swim at all: clearly, some could, and swim track abundance suggests this behaviour was not unusual, in at least some species. But were all pterosaurs equally adept at swimming, and what - beyond relative positions of the limbs - was their likely floating posture? We have typically assumed that floating pterosaurs might look a bit like floating birds, sitting high on the water surface with their heads well clear of the anything wet. To test this, Hone and Henderson (2014) threw sophisticated, 3D virtual models of pterosaurs into buckets of digital water to see how they floated. These models had variable density for major body components, so account for the distribution of airsacs throughout pterosaur bodies. Some readers may be familiar with Don Henderson's digital water experiments concerning other species: we've seen sauropods and giraffes given the same treatment to understand their floating mechanics (Henderson 2004; Henderson and Naish 2010; and yes, that reads 'giraffes': as explained here, no-one really knows how well they swim!). The Hone and Henderson study included multiple pterosaur species, and grounded itself with convincingly replicating the floating postures of birds (as, indeed, other Henderson studies have done with the floating postures of other extant animals).

How did pterosaurs fare? Although they assumed a stable floating posture, it was not quite as expected. As explained by Dave Hone at The Guardian, pterosaurs were incapable of assuming a bird-like pose when floating. Playing around with postures and body component densities made little difference: the digital pterosaurs consistently floated with their heads close to, or somewhat submerged, in water. Crucially, their nostrils always ended up close to the digital water line, suggesting that anything but a motionless pterosaur in the calmest water was going to be struggling for a clear airway. The problem, it seems, is that pterosaurs are very front heavy. Pterosaurs combine large heads, necks and shoulders with comparatively slender hindquarters so that, even accounting for their pneumatic features and denser hindlimbs, they consistently pitch forward when floating. This condition is more pronounced in pterodactyloids than other pterosaurs, but the general problem applies across the group.

So, sorry, 2013 Ornithocheirus-as-a-bird-like-floater-image-that-I-have-a-little-soft-spot-for, you're out of date.
Does the inability for pterosaurs to float like birds make them unlikely to enter water? To answer that, we might need to think about bird anatomy for just a minute. Because we see swimming birds so often, we don't think their floatation skills are especially remarkable, but they actually are. Bird anatomy is ideal for stable, effortless floating: not only are their necks and heads smaller than those of pterosaurs, but their heavy shoulder regions are counterbalanced by large, strongly muscled legs. This, of course, is a reflection of the hindlimb-dominated launch strategies employed by birds. The only reason their legs are so large and heavy is because of the demands of bipedal launch. Other tetrapod fliers, like pterosaurs, launch using different strategies have no need for heavy hindquarters, which means they overbalance easily in water. Bird bodies are thus exapted for floating, their comparatively large, well-balanced and pneumatised torsos providing stable, lightweight platforms to rest on water. Birds are so good at floating that they can be entirely passive when doing so, their heads sufficiently clear to avoid issues with respiration and light enough that they can move them around freely without overbalancing. This is taken to extreme in aquatic birds like swans and ducks, which perform much of their daily activities on water surfaces. They're essentially flying barges, alighting on water to collect food with lightweight, crane-like necks and heads. The next time you see a floating swan, consider that you're looking at an evolutionary champion as goes floating and foraging from the water-surface.

What does this mean for pterosaurs? Their inability to float like birds probably rules out some behaviours, such as prolonged bouts of sitting on water to rest or forage. However, most swimming animals - even those which routinely travel or forage in water - also can't float like birds. Indeed, predicted pterosaur floating postures are actually pretty consistent with those of other non-avian tetrapods. We might therefore surmise that pterosaur floating abilities are not atypically bad, but simply not at the 'advanced' avian level. As with most tetrapods, pterosaurs might have been quite happy in water, the caveat being that it would never be a passive, restful act. Water-borne pterosaurs were likely either were there for a reason (e.g. finding food, moving through an environment) or, if they had no business there, sought to escape it as soon as possible.

This brings us to the third string of recent work on aquatic pterosaur habits: the biomechanics of entering and exiting water. Earlier this year I discussed pterosaur water launch at some length, so will only provide a brief summary here. Calculations by pterosaur biomechanicists Michael Habib and Jim Cunningham (2010) suggest that pterosaur quadrupedal launching also works on water, albeit in a modified, and slightly more energy intensive form. For some pterosaurs, the effort needed to escape water necessitates a series of hops across the water/air interface to escape surface tension and build up velocity, but some - like the big, powerful azhdarchids - could hulk smash water powerfully enough to escape in one go.

Ornithocheiroid Ornithocheirus simus achieves launch velocity from a coastal sea. Prints of this painting are available from my shop.
Pterosaur water launch becomes especially relevant to our discussion of aquatic habits when we consider the adaptations it imposed on pterosaur anatomy. Certain pterosaurs - ornithocheirids, pteranodontids, rhamphorhynchids - possess features which are unusual among pterosaurs until viewed in light of aquatic launching. Reconfiguring the shoulder muscles to optimise for aquatic launches favours warped or hatchet-shaped deltopectoral crests (the flange on the humerus which anchors flight muscle), robust shoulders, reduced hindlimbs and broad wing joints: most or all of these characters occur in these lineages (Habib and Cunningham 2010). Other pterosaurs - like the aforementioned azhdarchids - seem capable of water launching without these features, suggesting they are not strictly essential for water launch. We might therefore consider some pterosaurs as specifically adapated for aquatic takeoff, implying that some taxa were routinely entering aquatic realms rather than just casually dropping in, or suffering the odd accident.

As with terrestrially-based pterosaurs, it seems takeoff strains put a cap on the maximum size of aquatic-adapted forms. In his Flugsaurier 2015 talk, Mike Habib explained how animals larger than Pteranodon (biggest wingspans around 5-6 m) would struggle with water launching. The largest ornithocheiroids (8-9 m wingpan, c. 160kg in mass) seem to require significant energetic investment and space to take off from water, to the extent that entering aquatic settings resulted in a net loss of energy unless food was particularly plentiful (Habib 2015). This is not to say water launching was impossible for very large or giant pterosaurs, but that the energy demands make it an unlikely routine behaviour. Pterosaur aficionados will note that this size constraint is lower than those proposed for terrestrial launchers (Habib 2013): as might be expected, this reflects the complexity of launching from a fluid substrate instead of hard ground.

Nevertheless, most pterosaurs were not operating at those enormous proportions, and so could theoretically enter water with less concern. Intriguingly, early calculations suggest that some pterosaurs were well-suited to rapid water entry. Qualitative assessments of Pteranodon anatomy indicate that it might be capable of performing shallow dives because, in general construction, it is no less robust than diving birds like pelicans (Bennett 2001; note this is not advocating pelican-like feeding for Pteranodon per se, but simply that Pteranodon anatomy was robust enough to dive into water from a flighted position). Mike's Flugsaurier 2015 talk suggested that this observation is borne out in some basic assessments of skeletal strength. Diving actions would not exceed safety factors of the Pteranodon skeleton, and its streamlined head and air sacs anterior to the torso would aid force dissipation as the animal penetrated the water surface. I must admit to finding the concept of diving Pteranodon quite appealing. Pteranodon skulls are especially streamlined and pointy compared to many other marine pterosaurs (not the least because they lack teeth and anterior crests), and we know that at least some individuals predated relatively tiny fish (Bennett 2001) which may have been difficult to snag during flight. Thus, some sort of shallow diving to get Pteranodon into water where it can pursue prey makes sense to me (as depicted above, see Witton 2013 for more discussion of this concept).

Pteranodon sp. jaw specimen AMNH 5098. That mass of random crap between the manidibular rami is a heap of small, half-digested fish vertebrae. Scale bar represents 100 mm.
To tie this together, our understanding of pterosaurian aquatic locomotion has moved on a lot in just over a decade. While it would be remiss to pretend we have anything more than a basic understanding of their aquatic skills, we nevertheless have some basic hypotheses in place for further work: footprints tell us pterosaurs did swim; digital models give us some idea of likely floating postures and constraints on behaviour; and biomechanical studies hint at anatomical parameters suited to aquatic locomotion. This allows us to start asking more refined questions: which species regularly entered water, and why? How did they propel themselves? Were any pterosaurs specifically adapted for aquatic lifestyles? There are several projects in the works which have bearing on these questions, and those interested in pterosaur lifestyles will definitely want to keep an eye out for them.

This post was brought to you by Patreon

If you've enjoyed this post, and would like to see more original artwork and articles on cool fossil animals, please consider contributing to my Patreon campaign. For as little as $1 a month you can help keep this enterprise ticking over, and you'll get access to exclusive content, discount print prices and other rewards for your troubles. A huge thanks to those who have signed on in my first week of Patreon - your support is really appreciated and encouraging.


  • Bennett, S. C. (2001). The osteology and functional morphology of the Late Cretaceous pterosaur Pteranodon Part I. General description of osteology. Palaeontographica Abteilung A, 1-112.
  • Bramwell, C. D., & Whitfield, G. R. (1974). Biomechanics of Pteranodon. Philosophical Transactions of the Royal Society B: Biological Sciences, 267(890), 503-581.
  • Habib, M. (2013). Constraining the air giants: limits on size in flying animals as an example of constraint-based biomechanical theories of form. Biological Theory, 8(3), 245-252.
  • Habib, M. 2015. Size limits of marine pterosaurs and energetic considerations of plunge versus pluck feeding. Flugsaurier 2015 Portsmouth, Abstract Volume, 24-25.
  • Habib, M. B., & Cunningham, J. (2010). Capacity for water launch in Anhanguera and Quetzalcoatlus. Acta Geoscientica Sinica, 31, 24-25.
  • Henderson, D. M. (2004). Tipsy punters: sauropod dinosaur pneumaticity, buoyancy and aquatic habits. Proceedings of the Royal Society of London B: Biological Sciences, 271(Suppl 4), S180-S183.
  • Henderson, D. M., & Naish, D. (2010). Predicting the buoyancy, equilibrium and potential swimming ability of giraffes by computational analysis. Journal of theoretical biology, 265(2), 151-159.
  • Hone, D. W., & Henderson, D. M. (2014). The posture of floating pterosaurs: Ecological implications for inhabiting marine and freshwater habitats. Palaeogeography, Palaeoclimatology, Palaeoecology, 394, 89-98.
  • Lockley, M. G., & Wright, J. L. (2003). Pterosaur swim tracks and other ichnological evidence of behaviour and ecology. Geological Society, London, Special Publications, 217(1), 297-313.
  • Witton, M. P. (2013). Pterosaurs: Natural History, Evolution, Anatomy. Princeton University Press.


  1. 160 kg? I had no idea any pterosaurs besides Azhdarchids got anywhere near that size. Is there really any evidence that they achieved 9 meter wingspans?

    1. Yep - some ornithocheirids really do seem to get that large. A recently described specimen from Brazil makes a compelling case for >8 m wingspans in this lineage:

      Of course, ornithocheirids are tiny bodied, short-legged and large-winged, so an 8 m wingspan doesn't equate to the same overall magnitude that we'd expect in a comparably sized azhdarchid.

  2. I find it weird that pterosaurs were front heavy, yet only left hindlimb tracts. They may not have been bird-like in floating posture, but they were probably more competent at that than the Henderson study suggested.

    After all, of the tested species, only Rhamphorhynchus is aquatic and it actually bears a stable posture.

  3. "Notably absent altogether from the same slabs are imprints from pterosaur forelimbs [...] Quite what this means for mysterious 'manus only' pterosaur tracks (pterosaur track sites where only hand impressions are recorded) is a discussion for another day."

    Surely it means only that some members of this 160-million-year clade had different behaviour from others? That shouldn't come as a surprise.

  4. Something I noticed about the swimming pterosaur models: On the Pteranodon one, the area where the nostrils would have been do appear to be above the water line.
    That made me remember that the Pteranodon's nostrils seem to be higher up on its face than those of other pterosaurs...which makes me wonder if maybe that was an adaptation specifically to help their swimming abilities. Like how crocodiles or spinosaurids have high up nostrils.
    I mean that's just what I'm thinking.

  5. Given so very many animals fit and fill the same niches throughout time, I'd be astounded if some pterosaurs did not float, swim and dive to exploit food sources the same as birds. Not necessarily via convergent evolution, given the design but no doubt in some ways, in some cases.