Azhdarchid pterosaurs: 'terrestrial stalkers' or pelican-esque 'scoop-feeders'?

This week saw the pre-publication of a new paper by myself and Darren Naish on one of our favourite topics, azhdarchid pterosaur* feeding habits. The article is now available in proof format in the Open Access journal Acta Palaeontologica Polonica, with the final, fully typeset version following sometime next year. Darren and I are no strangers to the long-necked, frequently gigantic azhdarchids of course, having discussed azhdarchid foraging habits at length in a 2008 paper and concluding that previously proposed lifestyles - skim-feeding, sediment probing, obligate scavenging - were inconsistent with azhdarchid functional anatomy. Instead, we proposed a novel hypothesis, that azhdarchids were 'terrestrial stalkers', basically just a sexy way of saying 'they wandered around on the ground and ate whatever they could find'. Hey, half of selling an idea is a snappy name, baby.

*Surely no-one here needs to be told what an azhdarchid is? You do? Then check out this article for a primer.

Why do we think azhdarchids were 'terrestrial stalkers'? A handy infographic explaining our reasoning, from Witton and Naish (2013).  The greyed cervical vertebrae indicate the range of azhdarchid neck motion according to Averianov (2013), which we are pleased to see meeting our expectations of ground-reaching ability (see Witton and Naish 2008; Fig. 8 and caption).

We're not the only folks with opinions on azhdarchid palaeoecology of course. Although I think the terrestrial stalker idea has been generally well received, Alexander Averianov (2013) disagreed with the idea. Earlier this year, he proposed that the terrestrial stalker hypothesis is flawed for three major reasons, which can be summarised as:

1. Azhdarchid remains are always found in ancient lake and river deposits, which indicates they were feeding there as well.
2. Grounded azhdarchids were vulnerable to predation from theropod dinosaurs, being ill-suited to rapid takeoff or other means of quick escape.
3. We overlooked the helical jaw joint of azhdarchids in our 2008 paper. Azhdarchids possess a skewed jaw joint which laterally displaces the mandibular rami when the jaw is opened, expanding the throat region marginally. According to Averianov (2013), this is a sign of expanding, pelican-like jaws, which permitted fish to be scooped from water in flight, which is a superior hypothesis to terrestrial stalking.

After some discussion between ourselves, Darren and I decided that we should respond formally to these points - Witton and Naish (2013) is the result. In doing so, we were able to explore some aspects of azhdarchid palaeobiology a little more, as well as put some comments into print on the way we interpret the lifestyles of fossil animals. Hopefully, a lot of the things we have to say on this will be of interest to readers here, so I thought I'd provide a quick summary.

Taphonomy is not destiny
Averanov's (2013) first 'flaw' is problematic for pretty elementary reasons. It's common knowledge that all manner of fossil terrestrial animals occur in aquatic environments because that's where the majority of continental sediments accumulate. Azhdarchids routinely occur in aquatic deposits with the likes of dinosaurs, reptiles, birds and so on, but we don't assume the latter are tied to water simply because their fossils are found in ancient rivers and lakes. Ergo, we shouldn't assume this for azhdarchids either. Taphonomy does not necessarily correlate with palaeobiology. Moreover, it's not true that all azhdarchids are found in remnants of aquatic settings: some occur in ancient deserts and ash beds. There's not much else to say on this fairly basic point (check out the paper if you want to read our full response), so we'll get onto the more interesting stuff.

Killer storks, giant pterosaurs, and the Age of MurderDeathReptiles
A number of folks have asked us about the vulnerability of grounded pterosaurs to predators, and Averianov (2013) specifically mentions the problems azhdarchids would have taking off when faced with attackers ("It is hardly probable that huge azhdarchids could take wing in one go and running for acceleration is difficult in marshland conditions” - Averianov 2013, p. 207). As we note in our new paper, palaeobehaviour is hard to discuss in a truly scientific manner and we are wary of just making bold, arm-wavy comments about ancient predator-prey interactions. There are some comments we can make, however, which do not rely on crass speculation.

Firstly, modern ideas of pterosaur takeoff (which regular pterosaurophiles will know means quadrupedal launching) suggest these animals could become airborne in seconds from a standing start (contra Averianov 2013). Thus, there is little reason to think that azhdarchids - or any other pterosaurs - would have to engage in panicked running to escape predators. Quad launches also permit greater acceleration and power than bipedal launches. This may make pterosaurs actually more adept at turning tail from predators than large modern birds, which do have to engage in a little taxiing before becoming airborne. We therefore do not envisage that grounded pterosaurs - even giant azhdarchids - would struggle to escape predators when startled.

According to some, this is pretty much what the Mesozoic looked like all the time. Background borrowed from here.

We also note that while terrestrial stalking is considered an unusual lifestyle for pterosaurs, comparable ecologies are actually pretty common among modern birds. Indeed, a lifestyle of walking around and eating stuff found on the ground seems to be the 'default' foraging strategy for many bird groups, and there's no indication that this makes them atypically vulnerable to predation. This even applies to large birds which live in predator-rich environments, where big cats, dogs, hyenas and other predatory species are real dangers. We have to ask why Mesozoic ecosystems would be any different? Is it because ancient reptiles are generally portrayed as aggressive monsters who're constantly pitched in battle (above)? Maybe, but this is almost certainly wrong. Darren communicates this very clearly in our new paper:

"...the idea of azhdarchids may have been highly vulnerable to terrestrial predation labours under several probably erroneous assumptions, including viewing theropods as unstoppable killing machines, immediately pouncing on and devouring any grounded pterosaur. In point of fact, the behaviour of living predators indicates that theropods large and small likely exploited easy prey (Hone and Rauhut 2010), ignored or avoided large or awkward prey, and were not a perpetual, 24-hour menace across all environments, worldwide." Witton and Naish 2013 (In Press)

I've discussed the over-statement of aggressive behaviour of Mesozoic animals several times before, and I'm sure I'm not alone in finding portrayal of dinosaurs as angry murder/death/kill machines irritating. It's frustrating enough when seen in popular media, but particularly irksome when it seemingly influences scientific discussions. I don't want to understate predation risks, but modern animals demonstrate that behaviours like extended bouts of foraging, resting and socialising can be performed without being ripped to pieces by passing predators. Assuming the Mesozoic operated under the same basic principles, it almost certainly wasn't the 190 million year bloodbath it's often made out to be.

A giant pterosaur compared to top theropod carnivores of giant azhdarchid-bearing Late Cretaceous ecosystems. A, Tyrannosaurus rex, representing the largest known predator in Maastrichtian North America; B, Balaur bondoc, largest predatory theropod of Maastrichtian Romania; C, Arambourgiania philadelphiae, standing in for the similarly-sized azhdarchids which lived alongside A and B, respectively; D, human sleuth for scale. From Witton and Naish (2013).

The composition of azhdarchid-bearing faunas is also of interest here (above). In some parts of time and space, enormous, 10 m wingspan azhdarchids lived alongside large predators like tyrannosaurids and spinosaurids. In others, the biggest theropods were turkey-sized. In fact, in latest Cretaceous European deposits, azhdarchids are the biggest predatory animals by a huge margin, and unlikely to be bothered by any theropods once they grew beyond a certain size. In these settings, azhdarchids weren't in perpetual trouble from theropods: they were perpetual trouble for theropods. Heck, the sheer size of an adult giant azhdarchid is impressive even alongside the very largest carnivores, and we wonder if this alone would dissuade less ambitious predators. Of course, there are plenty of small azhdarchid species which may be somewhat more easily subdued by theropods, and there are plenty of faunas were azhdarchids are not large, dominant species, but it's worth stressing that some azhdarchids lived in settings devoid of serious predator risk.

Of course, there were likely some occasions when azhdarchids were caught out by predators: would this spell instant doom for the pterosaur? Not necessarily. Again, this is hard to say with confidence, but we note that large modern storks - which resemble azhdarchids more than any other modern species - can be far more dangerous than most folks realise. These birds can inflict severe, sometimes fatal injuries with their beaks when panicked and cornered. Children are seriously wounded or even killed by marabou storks when trying to harvest soft white contour feathers from these usually calm birds (Mackay 1950). Zoo staff routinely arm themselves against attack from captive jabiru storks because attacks are so frequent and vicious (Shannon 1987). Indeed, even relatively large animals like tapirs are no match for angry jabirus. These storks are not armed with razor-sharp, hooked beaks: they deliver this damage with their simple, long, pointed bills. Whether this means azhdarchids used their jaws as similarly formidable weapons is anyone's guess, but it demonstrates that azhdarchid-like bills can be used as fearsome predator deterrents if wielded properly. Remember, of course, that some azhdarchids probably had beaks over 2 m long, 6-8 times longer than those of even the largest modern storks. An giant azhdarchid in a bad mood may be well worth avoiding.

We have some additional discussion on this point in our MS, but I think you get the gist of what we're saying. Our bottom line is not that azhdarchids could wander about Cretaceous plains without a care in the world, just that there is no reason to assume they were overtly vulnerable to predation risks. Indeed, there is evidence to quite the opposite in several cases, and there is no reason to think this is a flaw in the terrestrial stalker hypothesis.

The scoop-feeding pelican-mimic thing
This does not mean, of course, that azhdarchids had to be terrestrial stalkers just because they could walk around without being eaten immediately: water-trawling 'scoop feeding' could still be a viable alternative to terrestrial stalking. Citing the helical jaw joint of azhdarchids as evidence for a pelican-like expanding throat region, Averianov (2013)'s summation of his azhdarchid feeding hypothesis reads:

"...azhdarchids flied [sic] slowly above the water surface of large inland water bodies… looking out for fish or small fish shoals. As prey is detected, they opened the mouth, expanding the throat sac due to the spiral jaw joint, and captured fish in this scoop net, formed by the jaw rami and throat sac. Then, the head was thrown abruptly back by extension of the neck in the posterior region and prey was swallowed.” Averianov 2013, p. 209 

Although far from the first author to compare pterosaur and pelican jaws favourably, this is the first time (to my knowledge) that specifically pelican-like throat expansion has been proposed for pterosaurs and linked to a certain foraging strategy. The exact method of foraging suggested here - a mix of 'scoop' and skim-feeding - does not have a modern representative but is clearly an 'extreme' lifestyle, likely to incur considerable loading on azhdarchid skulls, jaws and neck. As with some other proposed 'extreme' azhdarchid lifestyles, like skim-feeding, we'd expect to see considerable specialisation in azhdarchid anatomy to reflect this but, unfortunately, we don't. Indeed, our assessment of this feeding mechanism suggests it is fraught with biomechanical and functional problems, in addition to failing tests offered by comparative anatomy.

Extending jaw area measurements of the brown pelican and select azhdarchid pterosaurs. Note the pelican is being rather lazy with it's jaw bowing, and yet still achieves much greater area increase than the azhdarchids. From Witton and Naish (2013).

We investigated the plausibility of 'scoop-feeding' in several ways. Firstly, we measured flexed and unflexed jaw areas of azhdarchids and pelicans to compare their range of jaw expansion (above). It turns out that azhdarchid jaws achieve pretty negligible amounts of jaw area increase even when an unrealistic amount of jaw flexion is permitted. By contrast, a bowed pelican jaw achieves an enormous area increase even when not trying very hard: we could only measure a partially bowed pelican jaw, but even this left pterosaur jaw expansion looking pretty pathetic. We utilised the same area measurements of azhdarchid jaws to calculate drag forces incurred on an azhdarchid neck during the 'scoop' phase of foraging, when the entire mandible has to be ploughed through the water. Unsurprisingly, the resultant drag forces were pretty huge, and are several hundreds times higher than the strain permissible by an azhdarchid fifth neck vertebra (hat tip to Mike Habib for suggesting using our jaw area data in this way). An azhdarchid that lowered its jaw into the water to try 'scoop feeding' would die a horrible, horrible death. This, of course, has further negative implications on the idea that azhdarchids were skim-feeders: even partial submersion of their mandibles was likely to snap their necks.

Brown pelican jaws in action. From Schreiber et al. (1975)

As if it didn't look bleak enough for 'scoop feeding', things took a turn for the worse when we compared azhdarchid and pelican jaw anatomy. Pelican mandibles and throats are amazingly freaky and specialised, and these adaptations directly relate to their manner of grabbing prey (above). Their foraging adaptations include differentially mineralised jaw bones which create distinct 'hinges' at points along the jaw; short mandibular symphyses; loosely-jointed posterior jaw bones; super-elastic throat tissues; reduction of the tongue, and skin-like beak tissues which permit jaw flexion. At least some of these features should be detectable in jaw fossils, but no indication of similar adaptations are found in azhdarchid jaws. In fact, directly opposing anatomies are seen in most instances. But what of the helical jaw joint? Isn't that functionally significant? Probably not, because helical jaw joints are far from unique to azhdarchids, being seen across all manner of archosaurs. Given the range of ecologies encapsulated by archosaurs with helical jaw joints, they're clearly of questionable, if any, significance to foraging strategies. It seems that the potential for azhdarchid jaws to perform expanding actions are limited at best, and we should stop referring to their gently-bowing mandibular rami as being 'pelican-like': they're really nothing of the sort. Indeed, the only animals we know of with even remotely pelican-like jaws are rorqual whales. I could go on (and we do in the paper), but I guess it's already clear that we don't consider 'scoop feeding' a viable alternative to terrestrial stalking at all.

Extreme lifestyles require extreme anatomies. Here's a summary of what you need to be a skim-feeding species, according to the modern skimming bird, Rynchops. From this post.

A closing point
In sum, we more-or-less go full circle in our new study, coming back to terrestrial stalking as the most likely current interpretation of azhdarchid palaeecology. Reflecting on this study, and the other studies into pterosaur palaeoecology I've been involved with (Humphries et al. 2007; Witton and Naish 2008, 2013; Witton 2012), it strikes me that proposed 'extreme' foraging methods are almost always inferred from a few anatomical characteristics rather than entire bauplans. This is certainly the case for 'scoop feeding' and skim-feeding (e.g. Kellner and Langston 1996; Martill 1997; Averianov 2013). Why do we keep doing this? It almost seems that our default assumption for pterosaurs is that they lived crazy, outlandish lives which we select evidence to verify. This is a completely backwards and unscientific way of assessing ancient animal habits. Modern animals with 'extreme' lifestyles wear their adaptations across their bodies, suggesting that we need to look at the entire picture of extinct species before we propose our palaeoecological interpretations (see details on skim-feeding adaptations, above). Folks like myself and Darren currently champion the terrestrial stalker hypothesis not because it's our 'pet idea', but because it's currently the only hypothesis which considers the entire azhdarchid bauplan (see our infographic at the top of the post), is consistent with biomechanical or functional parameters of azhdarchid anatomy and matches lifestyle predictions made through comparative anatomy. It may well not be the last word on this topic, but at least there's a foundation of science to it, which is more than can be said for a lot of proposed pterosaur lifestyles (see Witton 2013 for a review). If we're expecting to understand the palaeoecology of these animals in detail, we really have to move away from our rather basic, selective interpretations of their anatomy and provide more detailed, dedicated assessments.

I'll have to stop there for now. Be sure to check out the rest of Witton and Naish (2013) for further details on this study and, for more on pterosaur palaeoecology and azhdarchids in general, you may want to check my book (Witton 2013).

References

• Averianov, A. O. (2013). Reconstruction of the neck of Azhdarcho lancicollis and lifestyle of azhdarchids (Pterosauria, Azhdarchidae). Paleontological Journal, 47(2), 203-209.
• Humphries, S., Bonser, R. H., Witton, M. P., & Martill, D. M. (2007). Did pterosaurs feed by skimming? Physical modelling and anatomical evaluation of an unusual feeding method. PLoS biology, 5(8), e204.
• Kellner, A. W., & Langston Jr, W. (1996). Cranial remains of Quetzalcoatlus (Pterosauria, Azhdarchidae) from Late Cretaceous sediments of Big Bend National Park, Texas. Journal of Vertebrate Paleontology, 16, 222-231.
• Mackay, H. (1950). The quaint Marabou stork. Zoo Life 5, 91-92.
• Martill, D. M. (1997). From hypothesis to fact in a flight of fancy: The responsibility of the popular scientific media. Geology Today, 13, 71-73.
• Schreiber, R. W., Woolfenden, G. E. & Curtsinger, W. E. (1975). Prey capture by the Brown Pelican. The Auk, 92(4), 649-654.
• Shannon, P. W. (1987) The Jabiru Stork (Jabiru mycteria) in zoo collections in the United States. Colonial Waterbirds 10, 242-250.
• Witton, M. P. (2012). New insights into the skull of Istiodactylus latidens (Ornithocheiroidea, Pterodactyloidea). PloS One, 7(3), e33170.
• Witton, M. P. (2013). Pterosaurs: Natural History, Evolution, Anatomy. Princeton University Press.
• Witton, M. P., & Naish, D. (2008). A reappraisal of azhdarchid pterosaur functional morphology and paleoecology. PLoS One, 3(5), e2271.
• Witton, M. P. & Naish, D. (2013) Azhdarchid pterosaurs: water-trawling pelican mimics or "terrestrial stalkers? Acta Palaeontologica Polonica (in press)