They stood in a field where barley grows, Their heads did no thinking, their arms didn't move, Except when the wind cut up rough, And mice ran around on the ground

Join me if you will, my friends, for a quick trip back in time. Y’see, today marks the end of a long, long episode in my recent career, the beginning of which can be traced back almost three years. We’re journeying back to the summer of 2005 when I was a fresh-faced graduate, looking to start my Ph.D. in October of the same year. Now, with no funding for my research at that time (erm, still no funding now), I spent much of that summer living at home in the dark, dark recesses of Essex and working my ass off trying to earn some money. As anyone who has ever looked for work in a small town can testify, finding a job can be difficult at the best of times, and I ended up taking two: both working as a part-time kitchen hand. One place I worked, my local Chicago Rock Café, was pretty quiet most of the time, so I had an awful lot of free time to, well, do nothing, basically. At such times, your mind tends to wander, and you end up doing some rather strange things. Like, for instance, using your dishwater to experiment with different types of aerial predation of pelagic organisms. I did them all: skim-feeding, dip-feeding, diving… using branched fingers and food particles, I tried my best to get an insight into the mechanics of these foraging methods. It worked – kind of. And it was certainly more fun than cleaning behind the dishwasher.

There was also some semi-serious rationale behind this mildly embarrassing behaviour, too. Y’see, at various points in the summer, I’d made trips to see none-other than pterosaur guru Dave Martill and, among other things, we ended up discussing azhdarchid feeding strategies. In case anyone is unfamiliar with azhdarchids, they were a group of pterosaurs characterised by their elongate skulls, even longer necks, long hindlimbs but relatively short wings. Several azhdarchids, like Quetzalcoatlus and Hazegopteryx, were also enormous: you know, the size of a giraffe, or even large enough to swallow a modestly-proportioned pterosaur researcher whole (a little birdy tells me they may have the longest jaws of any non-marine tetrapod, dontcha know). Now, azhdarchid feeding strategies have been pretty controversial since the first good azhdarchid material was found in the 1970s, and Dave has stuck his oar into the murky swamps of azhdarchid palaeobiology more than once. Circa 2005, Dave was playing around with the idea that azhdarchids were surface gleaners (think about select terns and frigate birds – when the latter aren’t stealing food from other birds, that is), skim-feeders (this work was still in its early stages at this point, so he can be forgiven. Mostly.), or some sort of weird aerial-darter that used its long neck and jaws to grab fishes and other swimming being from the water. Now, though I didn't dare say it at the time, not one of these ideas sat right with me: long, stiff necks, drag forces, forward momentum… I couldn’t see how any of these could work, and my dishwater experiments didn’t help me resolve these issues, either. Still, I had little time to worry about such things: that crockery wasn’t going to clean itself.

Fast-forward several months to December 2005. Having volunteered to give a lunchtime seminar on my very, very early work looking at pterosaur mass, I was drawing up some pterosaur silhouettes to show pterosaur wingspans compared to our own armspans and, when drawing Quetzalcoatlus, I couldn’t stop thinking about that long, stiff neck. By this point I was working at my local cinema and spent several commutes to work, probably too many for a twenty-one year-old man, wondering about that infernal neck. Sure, its inflexibility was well known, but what was it all about? Then it hit me: it wasn’t about much at all. No, it was one of the least dynamic necks in history, permitting its owner to move its head up and down, a little-bit side to side, and not much else. What can you do with that? Wading. Azhdarchids were waders. They need next to no neck movement at all for that kind of gig. It was so clear. Their wing structure even matched up nicely with modern storks and ibises. It seemed perfect. I even drew a horrible, horrible picture of it for my talk and everything and, no, don’t ask to see it: it makes me cringe to look at it nowadays (it featured azhdarchids looking a bit like this, but worse. Much worse).

Dave wasn’t quite convinced by the idea (if memory serves, he eloquently summed it up with the word ‘bollocks’, followed by a cheeky grin), but it turns out that I wasn’t alone. Step in Darren Naish, a man with a terrifying knowledge of all things that have or ever have-had four limbs, owner of the world-famous Tetrapod Zoology and general good egg. We’d come to the same conclusion on azhdarchid lifestyles independently, but neither knew of anywhere where azhdarchid palaeobiology had been given a good, thorough review. Looking through the literature reaffirmed this observation, but, despite the lack of real research, there certainly was no shortage of proposed azhdarchid lifestyles. The earliest was that they were specialist scavengers, flying around like oversize vultures and feeding on the carcasses of dinosaurs. Others argued that they probed muddy shorelines for buried invertebrates, or pre-empted our thoughts that they patrolled the edges of water bodies to snatch up small, tasty morsels from the shallows. Some imagined azhdarchids as swimmers or aerial feeders, predating slow-flying animals in mid-air or swooping down to snatch fish from bodies of water. A healthy number of researchers, including Dave, took this latter idea even further, suggesting azhdarchids skim-fed like modern skimming birds, trawling their lower jaws through the water and grabbing any fish unfortunate enough to get in their way. Now, even an elementary survey of modern animals suggests that azhdarchids can’t do all these things: modern probers, skimmers and even scavengers are all highly specialised creatures reflecting millions of years of evolutionary fine-tuning, often leading to adaptations that close doors to other modes of life. The laterally compressed jaws of a skim-feeder, for instance, would be awful at ripping meat off a carcass, while the jaws of a carrion feeder would be of next-to-no use in probing. Applying this logic, Darren went through proposed azhdarchid lifestyles in a subsequently oft-cited blog post that summed up our thoughts on the topic as of early 2006, eventually settling on wading as the best supported. He also posted the first of several coloured azhdarchid-foraging pictures that I would produce, and, crucially, mentioned that he may write up his blog into a proper, genuine scientific paper, with me as co-author.

Well, that seemed like a good idea, and additional encouragement from Dave spurned further research into azhdarchid palaeobiology . Sitting down to write the first draft, I decided it would be neat to assess preferred azhdarchid habitat: not the easiest of things to do with fossil animals, of course, but possible if there’s A), enough fossil material and B), you keep things simple. You know: like saying whether an animal prefers an inland, coastal or marine setting. This was done by simply checking out the geological context of azhdarchid fossil sites: their sedimentology (different environments have different sedimentary characteristics), associated fossil types and so fourth. Trawling through the azhdarchid literature eventually revealed that over 50 per cent of their fossil remains (including all the most complete, articulated skeletons and several sites with associated individuals) stemmed from inland settings, while only a handful (and only generally fragmentary, scatty material, at that) of azhdarchid fossils had turned up in marine settings with no other terrestrially-derived material. What did this mean? Well, as with rivers and proverbial excrement, fossils always roll downhill, and downhill (in this case) means the sea. It is very difficult for animal remains to be transported upstream, so any fossils found in terrestrial settings indicate that the animal died more or less where it was found, or at least very close by. Additionally, the longer an animal carcass is transported, the longer it is exposed to scavenging and other phenomena that will break the body apart. Hence, complete animals have been buried quickly, while isolated bones may have been knocking around for some time. These observations suggest that live azhdarchids were hanging around inland floodplains and woodlands, not at the beach or out to sea. Now, while some authors were of this opinion before we came along to test it, other workers stuck to the dogma that azhdarchids, like all pterosaurs [cough], were shorebird-like animals and that their occurrences inland were the result of deaths occurring en-route between marine habitats. This is like suggesting that tyrannosaurs were actually beach-combers, but their numerous occurences in terrestrially-derived sediments occured because they seasonally trudged cross-country: it's just plain daft. A much simpler and better supported explanation is that animals like tyrannosaurs and azhdarchids simply inhabited inland environments, and speculation like the 'inland occurence via migration' theory can be smacked-down with the simple argument of ‘don’t be silly’.

Next, we looked at the mechanics of the azhdarchid skeleton and compared it against modern animals filling the same ecological roles as those proposed for azhdarchids. Now, good azhdarchid skeletons aren’t that well known: there are a handful of complete skeletons, but most of their remains are pretty fragmentary. Luckily for us, what remains we do have suggest a relative conservative body plan, allowing us to draw some conclusions that should apply to the group as a whole. We started with the earliest proposed azhdarchid lifestyle: could azhdarchids have been giant vulture-like creatures? Using data from other bits of my research into pterosaur lifestyles, it appears that their wings, being relatively short and broad, were well adapted to flying around inland settings, generating excellent lift forces when taking off in environments with variable wind conditions or lots of vegetation to snag longer wings onto. It is quite possible that azhdarchids exploited thermal updrafts of warm air as modern raptorial birds do today. Hence, there is little issue with azhdarchids being able to soar around in search of a carcass, but how would they handle eating it when they got there? Well, this may not have been so easy: although the large size of many azhdarchids would scare the pants of many other scavengers and give them dominance at carcasses, their long, stiff necks and skulls would limit their ability to reach deep into a dead body and leave much of the food inaccessible. Their jaws also lack the ‘meathook’ of modern avian scavengers, suggesting they were not specifically adapted for pulling bodies apart. Hence, we didn’t rate azhdarchids much as specialist scavengers, although the readiness of even maladapted modern animals to eat carrion means we couldn’t rule out the odd limited dabble in scavenging.

Probing was put under the spotlight next: would azhdarchids have wandered around Mesozoic swamps and lakes sticking their beaks into mud looking for buried shellfish? Well, they would’ve had a hard time figuring out just what was in said mud, what with their bills lacking any indication of sensory pits (so-called Herbst corpuscles) that are so important to modern probers and all. These often densely packed pits form a sensory battery that informs modern probers all about subterranean conditions, enabling them to detect potential prey items without seeing them. What’s more, forcing your beak into sand or mud is much easier with a narrow, streamlined bill than it is with the deep rostrum of azhdarchids and, crucially, how does one grab a morsel of food if, as in azhdarchids, the only joint along your jaw is right at the back? Modern probers have secondary jaw articulations that enable them to open their jaw tips while keeping their mouths closed, thus forgoing the need to open their entire mouths against the preussue of the surrounding sediment. Probing for food, therefore, seemed beyond azhdarchids.

How about swimming? Well, if anything, azhdarchids seemed to be among the least aquatic of all pterosaurs, bearing long, lanky limbs with small hands and feet that would be of little use in propelling themselves through water. Their anatomy similarly ruled out aerial hawking of less-acrobatic critters: seeing as azhdarchids could not use their hands or feet to grab aerial prey (both tied up in the wing membrane, y'see), they would rely entirely on their mouths to catch their food. The long neck and skull of azhdarchids are not what one would expect from an aerial predator: modern hawkers relying on their mouths to catch food have wide, gaping maws and short necks combined with high aerial agility: azhdarchids did not possess any of these adaptations, so we crossed this one off the list, too.

Then we hit the people's choice: aerial fishing. Now, in spite of its popularity among pterosaur workers, we found this to be rather flimsily supported. Our research verified countless hours of playing in kitchen sinks: that crazy, stiff azhdarchid neck simply cannot flex in the manner essential for grabbing fish with jaws in flight. The problem is momentum: the bodies of aerial fishers – be they dip-feeders or skim-feeders - continue to move forward while the head has to rotate beneath the body to grasp a relatively stationary prey item. This requires a flexible neck, and, despite its length, azhdarchid necks simply could not do this. The need for a flexible neck is greater in skim-feeders than dip-feeders: when impacting a prey item (or submerged branch, rock or shoreline – lots of scope for accidents in skim-feeding), the head is forcibly rotated beneath the body with the impact. The neck therefore acts as a cushion for both intentional and unintentional skim-feeding collisions, and azhdarchids would literally find both collisions a real pain in the neck if they attempted them. Furthermore, modern skim-feeding is only habitually practised by a few highly specialised bird species that have been through all sorts of evolutionary loops to forage efficiently in this manner. Most importantly, the lower jaw has been streamlined to a knife-like thickness to minimise drag while simultaneously remaining deep and robust to absorb skimming impacts, the jaw articulation has been reinforced for the same reason, and the jaw musculature is appropriately swollen to resist drag and impact effects. Not one azhdarchid bears any anatomical features like these, instead having slender jaws with flattened biting surfaces, relatively small jaw muscles and a pretty meagre jaw joint. Combined with their stiff neck, it seemed wholly unlikely that azhdarchids could skim-feed, and the aforementioned study carried out in parallel with our azhdarchid work reinforced this.

Wading around water courses is certainly a less demanding lifestyle and, to us, seemed more tangible than any other hypothesis. Azhdarchids belong to a group of pterosaurs that appear to have pretty competent terrestrial abilities, so lots of walking wouldn’t be much of an issue. Their long limbs would be dead handy for wading in moderately deep water, and their stiff necks – the bugbear of many other hypotheses - is no issue here, either: the animal simply has to lower the neck slightly to reach the substrate where the food lies, and nothing else. The long jaws help here, too, minimising the neck action required to put the jaw tips in the water. However, just as things were looking like being wrapped up in time for dinner, a 5 m tall azhdarchid put its 35 cm long foot right through the wading hypothesis. Having looked at lots of azhdarchid bones, we thought it only right to check their trace fossil record –currently limited to a few gigantic footprints and trackways in Korea (and maybe Mexico). Simply put, for their body size, azhdarchid feet and hands are tiny. Ask any animal wading around on soft substrates like those adorning river channels and lakes for their footwear of choice, and they'll tell you that they want large, splayed organs to spread their weight across the maximum possible area, thus stopping them sinking. This doesn't sound much like the small appednages of azhdarchids, suggesting they wouldn't have been great waders. This is especially true of the biggest forms: if they did weigh a quarter-ton, weight spreading is an issue they would certainly have to face when wading. Bummer: bang goes the only hypothesis we had any stock in.

So what, what, what, could azhdarchids do? We’d been through the list: no-one had suggested anything that fit all the evidence. A fresh idea was clearly needed. Then it struck home: all these other ideas were anchored to the two commandments of the Pterosaur Bible: 1), Thy volant reptiles shalt feed only in or around environments of total saturation and dampness, and 2), Thy primordial dragons of long ago shalt perform all activities suspended above the ground. What if, what if we took azhdarchids out of the ponds and streams, tethered them to a stick and forced them to stand around a bit? Keep them there long enough and eventually they’d get hungry: then they’d have to forage on the ground - how well would they get on doing that? Well, returning to the same footprints that walked all over the wading hypothesis, it occurred to us that small feet are mechanically advantageous in walking over dry land, actually decreasing the amount of effort required in every step. Their heel and toes were padded, too, making for a comfortable ride on long trips or when standing for long periods of time. Long legs are a bonus in walking, too, as they mean more ground is covered per-step, again increasingly walking efficiency. One azhdarchid trackway (the longest pterosaur trackway in the world at 7 m, trivia fans) revealed a particularly interesting fact about azhdarchid gaits: it recently became apparent that derived pterosaurs strutted around with their limbs held mostly underneath the body, but azhdarchids went even further than most, converging with the highly-efficient, upright postures of mammals and birds. Their ability to hold their limbs directly beneath the body would be extremely advantageous for a walking azhdarchid, providing the most efficient and stable platform to support their (often massive) bodies. Hence, when it comes to looking for food, the azhdarchid tethered-up on the lawn would be perfectly happy despite having lost its ability to fly and its separation from water. So, azhdarchids were perfectly competent on land - moreso than any other pterosaurs, so far as we could see.

Well, you know, the point made about the necks of wading azhdarchids applies here, too: an azhdarchid standing on firm ground requires as little neck motion to get its jaws to the floor as a wader does to stick its jaws in the water. Could this mean that azhdarchids fed on land? Damn right: far from being just another group of fishing pterosaurs, we reckon that azhdarchids were specialist terrestrial stalkers, patrolling ancient plains on foot like reptilian equivalents of H. G. Wells enormous tripods, their watchful eyes and jaws mounted atop long necks to give them superior views of the runty animals scampering around their feet. We expect azhdarchids to have picked up all manner of bite-size morsels, from insects and fruits in smaller azhdarchids to, in the case of the largest, small or baby dinosaurs: after all, 'bite-size’ can be pretty big when your skull is over two metres long. Some vindication that such a lifestyle is possible is provided by modern storks and ground hornbills, birds that patrol around African grasslands looking for small vertebrates to terrorise. This hypothesis is a world away from the conventional view of pterosaurs as shorebird-like critters, but really is the only one that uses all the pieces of the puzzle.

But that’s not quite the end. Y’see, we’d more-or-less come to this conclusion by the end of 2006 (see this picture for proof), but, as is often the case, the project spent a lot of time on the proverbial back-burner while Darren and I worked on other things. Then, finally, at the beginning of this year, the text, figures and tables were finally finished, and the project was submitted. It took a few weeks to find a home (we can’t deny its rather anal, specialist nature), but eventually, PLoS One welcomed us on board. A quick turnaround from the reviewers and ourselves saw the paper accepted for publication on the 24th of April, and today, the 28th of May 2008, heralds its release to the world. Being part of PLoS, absolutely anyone can check out the paper, so why not grab the online version here, or check out the official Azhdarchid Paleobiology website Darren and I set up (including some nice high-resolution images, quality fans)? Be sure to see Darren’s side of the story at Tetrapod Zoology, too, and, judging from the amount of phone calls and E-mails Darren and I have been fielding today, there's no shortage of press interest around either.

So, there you go, then. Almost three years of progression from playing in dishwater, casually knocking around ideas with colleagues and finally making the project something worth talking about are concluded today. Only one last thing remains: as per usual, I’ve rambled on for ages without mentioning the image way, way, above all this writing. What you’ve got there is the ‘official’ press release image for the paper (actually Figure 9 from the paper itself), and, actually, it almost came as an afterthought. With just about everything else done, it occurred to me that many papers discussing the ecology of extinct animals conclude with a restoration of their animal in question feeding in whatever manner has been deduced by the authors, so I pitched the idea to Darren that we follow suit. The answer was an unhesitant ‘hell yes’, a decision no-doubt helped by our unashamedly juvenile attitude towards pretty-pictures in scientific papers. In our concluding image, then, you’ve got 750 kg of monster azhdarchids, Quetzalcoatlus, marauding around a Cretaceous fern-prairie in search of food. The one on the right has grabbed a baby titanosaur, the parents of whom are known from the same deposits that yield these giant pterosaurs. Personally, I reckon this is much better than my last reconstruction of azhdarchids engulfing a baby dinosaur, and, on reflection, it’s vaguely reminiscent of the iconic ‘Hills are Alive’ scenes in The Sound of Music. You know, with the rolling hills and mountains in the background and all that. Replace Julie Andrews with 5 m tall pterosaurs and it’s like you’re watching the same movie.

With that last comment probably having diminished any respect people may have gained for me for achieving my first publication where I’ve been granted the honour of being first author, I’d best finish. Thanks to all those who helped on the project, with a particular tip of the hat to Darren and Dave (who, I believe, now finally agrees with me), and to anyone who’s made it from beginning to end on this novella of a picture comment.