In our final post in this series my own PhD student Ross Elgin takes us through the work we did (with others I must hastily add) on the aerodynamics of pterosaur crests. This is one of those much discussed but little researched areas with (perhaps ironically) lots of hot air, but little moving air (like in a wind tunnel for example). So read on as we take to the skies with those crested pterosaurs.
Aside from the opportunity to congratulate Peter Wellnhofer, celebrate his excellent work on pterosaurs and enjoy a number of lectures on all manner of pterosaur goodies, the meeting in Munich highlighted another important thing…. just how many people we have now have with an interest in pterosaur research! And why not after all!? As I am sure anyone who has followed the posts on “the Musings” with any sort of continuity will be aware the pterosaurs are just one of those groups which despite over 200 odd years of constant scientific research we still don’t know all that much about them: where did they come from?; how did they grow and what did they eat?; were they efficient walkers on the ground or did they spend much of their life on the wing?; The list could get pretty extensive if I chose to include everything we still aren’t sure about, and I haven’t even mentioned taxonomy yet. Still the one thing I am certain of is that number of people looking to take up the challenge of “pterosaurology” is perhaps greater then ever and maybe just maybe, with the right degree of dedication and luck, we might be able to pry some more secrets from this ever more intriguing group. However I digress.
Ignoring all the intricate bits and bobs that we palaeontologists usually concern ourselves it wouldn’t be too far removed from the truth if I was to describe a pterosaur as little more than a flying head, and pretty spot on if we only consider the more derived pterodactyloids. Thus pick any derived pterosaur and you will be confronted with (hopefully) a large pair of wings and head often larger than the actual body itself, which is so reduced at times that I have often heard our engineers musing as to where they could possibly keep the fish after catching it. Of course in considering the size of the head we also must account for the large and extravagantly coloured, if Luis’s lovely renditions (on the banner above) are to be believed, crest that crowns the majority of pterosaur taxa. And here begins the small paradox that I found myself investigating many moons ago when I was but a lowly MSc. student back in Bristol. “Why would an animal go to such great lengths reducing its body size and hollowing out its bones, only to mount whopping sheet of bone/soft tissue on top of its head?”. My colleague at the time had a saying that “if anything in nature doesn’t make sense, blame it on the females,” talking, of course, about the function of sexual selection (before he or I are misinterpreted) and while he was certainly not the first person to have suggested this idea it has come to be more or less accepted by modern pterosaurologists.
This idea behind the sexual selection of the pterosaur crest is pretty simple, those males pterosaurs with the biggest, most brightly coloured or elaborate crests get all the girl pterosaurs and so pass their genes onto the next generation. Those with a less spectacular crest however are found wanting and more often than not simply have to do without, thus their genes disappear from the population. With such a strong positive selection acting on the crest (i.e. you aren’t going to mate without one) it is little wonder that as time marched on and old species split into newer species, larger and ever more elaborate crests were passed along as well. And thus we find the situation observed towards the end of the pterosaur lineage where the pressure to show off to the opposite sex was so strong that the crests themselves could easily end up dwarfing the original size of the skull, think Nyctosaurus and Tupandactylus navigans/imperator here. So, problem solved? Probably – after all many other types of animals have developed all manner of weird and wonderful ways to get noticed by the opposite sex (Bristol social life aside). Also several lines of evidence indeed suggest that the crest were sexually dimorphic and absent, or at least reduced, in female pterosaurs although the fossil record isn’t quite complete enough for this to be certain in many taxa. The obvious point being therefore that if it was performing some function that was not related to gender differences, we would see it in both genders.
However flying animals, and particularly large ones, have to be a bit more careful where they attach any extra bits and bobs and it is certainly no coincidence that all pterosaurian crests are extremely thin and streamlined from the front. Thus while sexual selection may, and probably does, explain the why of the crest it doesn’t consider the aerodynamic effects of attaching a whopping big sheet of bone/tissue on top of an already large head and this was what we wanted to test, and others in the past have suggested biomechanical benefits of the crest which we could look to support or refute with our analyses.
While such a feature might be imagined to impede a pterosaur during flight, for much of the last 100 years it has been an ongoing suggestion that the crest actually assisted it in one of several ways by a simple turn of the head and thus providing a possible alternative explanation to the origin of the crest. To test these predictions palaeontology (for a while anyway) regressed to an almost play school-like level and through the tireless effort of moulding play dough-like putty into banana shaped representations of pterosaurs and casting with the help of a rather large tub of latex rubber, several scaled down representations of Pteranodon were born. Things subsequently got a bit more scientific and the models strapped into a fairly sophisticated wind tunnel courtesy of City University, London and subjected to a number varying wind speeds to determine their aerodynamic profiles and test all the major aerodynamic possibilities that a crest might produce. So, ignoring the role of sexual selection in the development of the pterosaur crest what are these alternate theories and how exactly would a pterosaur benefit from having a crest mounted on the top of its head?
1) The crest as an airbrake.
Simply put if a large headed, large crested pterosaur turned its head against the air flow it would have been expected to produce a large amount of drag (the force that resists the movement of an object through a fluid) which, in turn, should have helped to slow the animal during landing or other manoeuvres. This however is one of these ideas that far sense in theory rather in practice. To produce the most drag possible the head would have to be turned broadside to the airflow so the animal wouldn’t be looking where it was flying and have big forces pushing on the fairly stiff neck. Also for maximum effect the crest should ideally have an aspect ratio (i.e. square) of 1 rather than the typical high aspect ratio crest (long and thin) associated with P. longiceps.
The other question that must be asked is could the crest compete with the drag production of the wings……probably not! Cambering the main wing by depressing the propatagium (fore wing) or lowering the hind limbs, which in turn cambers both the wing and brings the uropatagium (hind wing) into the air flow would have been far superior in this respect. Though the results of the wind tunnel experiments unsurprisingly confirm that a crested Pteranodon produces around 25-50% more drag than an uncrested form (confirming that crested models produce larger forces) they also unusually indicates that if the animals pitched their heads downwards (as suggested for some derived pterosaurs based on other research into the orientation of the ear) the profiles would collapse onto one another and all the models ended up producing similar coefficients of drag. (The coefficient of a force just being a convenient way of comparing the force production of models by removing any differences in size).
While a crest must have undoubtedly helped increase drag during a head turn, providing the head itself was not depressed, it is often of the wrong shape to maximise drag and the resulting force is actually very small, only a few newtons (N) – made even worse by the assumption that a pterosaur must have been travelling slower than 10m/s when such a tactic might have been employed to any use. As noted above, if the feature was also genuinely effective and useful one must wonder why only putative males had large crests and not the females too (a point that applies to the other situations below). Use of the crest as an airbrake, slim to none.
2) The crest as a forward rudder.
Put an object into a flow, turn a feature acting as a rudder and the object will naturally develop a torque (force producing a torsion or rotation) yawing (turning) it in the desired direction thus, as the theory goes, why could a large crest not be used in a similar manner to assist the animal during sharp turns? Bats and birds however it seems have not evolved such a feature for assisting in turns, relying on their wings instead (and note that large headed birds like hornbills and toucans don’t seem to use their heads in this way, though of course structurally these are very different features), and while it is a valid argument to state that there is no such need for a crest to play a role in such aerodynamic manoeuvres it is useful to see exactly how much more efficient the wings are. Bypassing the actual maths a pterosaur (or flying animal/vehicle of your choice) of weight 4.3kg that banks at an angle of 45° will produce a turning force of 30 N. That is more than three times the maximum generated by the head and crest at an angle of yaw of 40° and six times that generated at 15° yaw. Thus we can confidently state that the crest would have been of little additional use in helping the animal to turn. More likely the crest would have been a nuisance to the animal being simply one extra torque that would have had to have been balanced by the wings, resulting in an increase in drag (and again an unwanted pull on the neck).
The effect the crest, however, both for better or worse, it seems was ultimately limited by a fairly low flight speed. Falcons have been noted to have some problems with developing undesired torques as they usually approach their prey with the head turned out to the side, partially explaining why they choose to approach in a circling pattern rather than diving directly along the shortest possible path. With a flight speed of around 10-12m/s however even large pterosaurs were unlikely to have generated any seriously detrimental torques with a turn of the head. Use of the crest as a forward rudder, unnecessary and in fact probably slightly detrimental to stability.
3) The crest as a counter balance.
When not busy developing extra weight in the form of crests pterosaur took a number of weight saving measures and there was obviously a strong positive selection for reducing weight as much as possible. As muscle is a good deal heavier than bone perhaps pterosaurs developed the crest to balance the head and protect the neck against unwanted torques, thus becoming of evolutionary value by allowing for a reduction in neck muscles. So was this the case? Surprisingly enough, no. While all manners of variations in crest shape and form have been observed, most located forward of the neck joint, even the caudally directed crest of P. longiceps proved pretty lousy at counterbalancing the head. Other tests with a large variety of Brazilian pterosaurs confirm that pterosaur crests across the lineage just can’t counterbalance the head. So was the crest useful as a counterbalance, not in any of the studied taxa (probably including Tupandactylus imperator).
Thus in a nutshell while the cranial crest of Pteranodon succeeds in lowering the yawing moment of the head and limiting the movement of the centre of pressure, the overall aerodynamic effect is rather modest. Were the animal to try to use its head to assist in steering, braking, or manoeuvring, it would have instead created aerodynamic instabilities (which would have to be countered by the wings) and imposed additional twisting loads on the neck attachment (which is also pretty bad). So in the end while sexual selection seems to have won out in being the primary cause of the pterosaur crest it also appears that aerodynamic influences on the development of the crest were very few indeed. Other than having to maintain the compulsory streamlined front profile for drag reduction, the slow flight speeds of the pterosaurs allowed the crest almost unlimited freedom to develop in whatever way was possible from lateral view. It should probably therefore come as no surprise if more pterosaurs turn up in the future with ever larger, stranger and more elaborate crests, the latest batch of fossils out of Brazil unveiling the true extent of the crests of Ludodactylus and T. imperator for example as well as others like Raeticodactylus or Dsungaripterus.
With the aerodynamics of the head neatly tied up it is hopefully on to the much more taxing and interesting matter of determining the flight characteristics of the wings, bodies and pro/uropatagia. Hopefully these might eventually warrant another post in the not to distant future and we may well on our way to understanding the aerodynamics of a fully integrated pterosaur.
Elgin, R.A., Grau, C., Palmer, C., Hone, D.W.E., Greenwell, D., & Benton, M.J. 2008. Aerodynmic characters of the cranial crest in Pteranodon. In: Hone, D.W.E & Buffetaut, E (eds.) Flugsaurier: pterosaur papers in honour of Peter Wellnhofer. Zitteliana B, 28, 169-176.
Dave Hone's Archosaur Musings 
I note that you have not discussed a role for the crest in aiding flight stability and coordination, despite that this is the role of the analogous structure on modern airframes.
Excellent post, David!
Nathan, as this entry previously asked, if the crests were somehow beneficial with regards to flight, why would only the males have them?
Mark: It could be that only males were obliged to perform stunt flying, to impress females, and thus have greater need for stabilization. In that case they would still be used for display, in a sense, but not (solely) as display surfaces. Of course once it became a visible distinguishing feature, it would tend to hypertrophy, as such things do.
It’s easy to imagine a proto-moose’s antlers having been useful, once.
But again, the sheer range of crest shapes and sizes would still argue against this, at least in many taxa. One cannot expect something like Pteranodon, Tapejara, Tropegnathus, Austriadactylus and others to all be acting in simalr ways. Nor for that matter does it explain why most often, all speciemns of a given species are equally crested which strongly suggests presence on both sexes.
When you find yourself arguing that something potentially very useful was not actually used at all, you must pause and consider whether you have really strong evidence.
There is no question but that the crests were often far larger than any practical need would dictate. Pimpmobiles are, too, and so is Air Force One, but one doesn’t argue that they are not also transportation.
Presence of equal crests on both sexes of most taxa argues for at least some usefulness.
Yes, and it does argue in favour of sexual seletion / communication / display. A lack of *mechanical* function is what we found, not a lack of function full stop. This is the one function of course where although size can be important, sape generally isn’t. I’m not suggesting it is true of every taxon, but it does explain why there can be gross differences beween putative genders, or incredible similarities between them, and that none of the so-far offered suggestions for a mechancial function hold up to scrutiny, or suffer form only being applicable to a few taxa and not others. I have a paper on this very subject which hopefully will be submitted soon.
But you didn’t find a lack of mechanical function. You found a display function that is entirely compatible with mechanical function. To disprove mechanical function you would need to demonstrate that taxons which lacked a crest also lacked sex-linked flight behaviors. That seems, on its face, hard to demonstrate.
That’s “taxons in which one sex lacked a crest…”.
That’s not what we found though. The mechanical benefits were incredibly minor and negative in some cases (and the other, as yet unpublished taxa that were reviewed were often suffering profound negative effects) so it’s hardly ‘compatible’, it’s effectively neutral or nagative, when it had been claimed as strongly positive. Indeed this is highly analagoius to many extant display structures where a minor secondary function is present (e.g. heat loss through deer antlers) but it is clear that this is not their primary purpose, or the one which led to their initial evolution and developemnt. There might well be a minor mechanical gain, but the lack in females and variation argues against anything major (which is what we found) or even a sex-linked function in the manner in which you describe. Males might well have extra disaplys, but the differences are so minor that it would take a huge increse in crest size to allow only the slightest increse in agility which would be therefore of little consideration toa female. Do you notice that one makle can turn 2 degrees better than another, or that one has a crest 3 times the size? If the two are linked in this manner, the secondary aspect simply becomes part of the first in any case (i.e. a disaply flight backed by a crest, still requires the crest for disaply and this is the primary function and feature).
A male that can fly a little better than its neighbor might be much better-fed, and a better-fed male may be much more attractive to a female via any number of signals, including growing a fancier crest display.
I notice you did not consider, or did not mention, a non-aerodynamic role: increasing the roll moment of inertia of the head. Such increased moment would enable converting sudden shocks such as snatching items with the beak, below the center line, into compression along the major axis of the neck, rather than flexion. Again, in taxa where males performed the food gathering of that sort, they might have greater need for such a feature.
This sort of inquiry does not seem subject to conclusion, because there is really no way to know whether you have actually identified all the alternatives.
But that then takes us back to the origial argument, if it was generally beneficial (i.e. increases abilty to catch prey) then it would be selected for in both genders and probably be of a simialr type in all species, or of a few limited types and this is not the case.
No, we can’t check every possibility, but we can check all the likely ones. More research is needed, as ever, but we did test pretty much everyhting that had previously been suggested in the literature and demonstrated that none of it really held up to scruitny as with the great skimming debate.
Few features are “generally beneficial”. Tradeoffs vary, feeding styles vary, sex roles vary. Will you argue now that men’s stronger muscles had no functional value, on the basis that women find stronger men more attractive? This kind of logic just doesn’t work very well, because you have to know everything, and we know very little.
The conclusion that the crests were not air brakes is undoubtedly sound. The conclusion that they were not used to apply strong yaw forces is undoubtedly sound. That they must have had a big role in mating display is unimpeachable. To conclude, scrupulously, that they had no substantial functional, “mechanical” role would take more evidence than still exists. The alternatives are to give up the conclusion or to give up on scruples.
The best you can do is pick off proposed functional roles, one at a time, as you started out doing.
Oops..looks like i have been leaving Dave to defend my work for me so i’ll try and add a little onto what has already been said.
Firstly pterosaur crests have always been a bit of a sticky issue and people seem to be compelled to find some kind of mechanical means of justifying them. The point of this study was to show that the theories of crests acting as novel structures to increase the aerodynamic performance of the animal were simple not true. While i know Chris Bennett, and others im sure, have said this since the early 90’s the aim was to get some experimental data to support this. In the end the limited effects of the crests and the greater efficiency of the wings for anything the animal wanted to do ment that we rejected the “aerodynamical” function of the crest.
While Nathan points out we would need more evidence were we to rule out ALL possible mechanical functions that could be attributed to the crest….. however, at this point in time, it seems unlikely that any force other than sexual selection was at work here.
In truth however there are not that many possible mechanical uses for a crest build on top of your head once areodynamic enhancement has been thrown out the window: thermoregulation is one, however outside T.sethi this has no support, perhaps we could also argue that the ornithocheiroids developed their rostral crest as a “wave cutter” or to increase their bite acceleration in the same way that gharials or Sarcosuchus developed an additional weight on the end of their snouts.
In the end however all mechanical arguments have to accept that, other than an increase in size and material make up, pterosaur crests did not evolve towards a particular form. By this i mean that there was no drive towards all later pterosaurs having rostral crests or enlarging them so that the majority of the crest was, say, caudal to the occipital condyle. The SHEER diversity in crests argues against any “universal” mechanical advantage they might have contributed!
Thus while it is conceivable that the crests may have performed some mechanical advantage, this advantage would have had to have been very different from another closely related group with a different shape of crest. Whatever this advantage was, however, it was not deemed important/necessary enough to be retained by a more derived group that evolved later on. When we consider this, sexual selection makes far more sence as the driving force for crest evolution.
By all means i think experimental testing of such mechanical theories are a great idea, however, until i see any data suggesting this i will remain unconvinced that the crests truely provided any mechanical advantage of evolutionary importantance.
Returning to Nathan, I think you will find that is largely what I was saying. We tested a bunch of mechanical functions and found no (real) support for them, I have been talking contextually, not literally. Yes, there may be other functions, but we tested a fair few and more importantly, all the ones that *had* been postulated in the literature and found they did not stand up, there was ‘no mechanical function’ observable.
The reason for promoting sexual selection as the primary role, and indeed originator of many, if not all, crests is as I stated above. Since the mechnical effects that we can see are typically minor if they are positibve and are generally negative, soemthing else must be going on. Even with the examples where there are positive effects they are minor – it takes a big crest to produce any effect and thus the original small crests when they evolved must have had a slective advantage, in the absence of mechanical ones which don’t kick in util they are much bigger, it seems that SS / display etc. is the obvious candidate. Once more, the variation seen across genders / species certainly argues against a mechnical functin for most pterosaurs crests or they would almost certainly be more homogenous. Pterosaurs are incredibly conservative organisms and yet we see huge numbers of incredibly varied crests. If there was a big mechanical advantage at play, why would it not spread rapidly? They clearly have enormous cranial plasticity, and yet everyhting about their wing structure, limb proportions, body shape etc. etc. was incredibly conservative, and one assumes as a result of the strong pressures of powered flight. It’s hard to explain how something like the ornithocheirds which can barely be distinguised *apart* form the crests would have gained an advatage from them given just how incredibly unifrom they are othrrwise.
Thank you, Ross, for your explanation.
It is obvious that the common hypertrophy and radical variation must be a consequence of sexual selection, and it would be foolish to argue that the structures were not otherwise heavily influenced by display demands. Anything that can be a display structure becomes one in short order.
It’s a long way from there, though, to demonstrating an entire lack of practical value. Too many of the conceivable uses are not subject to verification, and too many are not even mentioned in the exposition. For example, were sensory uses even thought of? Did no one else mention moment of inertia, before me? Why not? Do you understand the difference between steering and stabilization?
We see a similar process around the narwhal’s tusk, and they’re not even extinct yet. We know the tusk is heavily innervated, which argues for a sensory function in addition to the obvious display / fitness-demonstration function. Were the crests innervated?
While Nathan and David are apparently correct with regard to generic display and species i.d., there is currently no evidence, that I am aware of, of any gender differences in pterosaurs. Bennett’s oft-quoted pelvic differences can be attributed to comparing a Pteranodon with a large nyctosaur pelvis, and scaled correctly, yield subequal cloacal openings, obviating the need for different pelvic morphologies. The rise and fall of the crest in Pteranodon are phylogenetic, as any cladogram test will show. And gender-based crest differences have not been noted in other taxa.
Can you provide references to peer-reviewed publciations for those assertions. And how exactly can you test the rise and fall of a crest phylogenetically when only two species are recognised? (that’s rheorical by the way). I have asked you before not to refer to unpublished research and frankly, unsubstantiated ideas and concepts. It does no one any favours and it not what this blog is for.
The problem is, David, Chris Bennett’s work with crests, ontogeny, species and pelvic morphology is unsubstantiated. No cladistic analyses attended these studies. There are no other pterosaurs that have male/female crests and male/female pelvises. Anyone who has ever seen a nyctosaur pelvis, and Chris is among them, should recognize the ‘female’ pelvis as belonging to one of those.
Chris’s hypothesis of just two species is just that: an hypothesis, and a poor one at that. Earlier workers, from Cope to Eaton to Williston and Miller have found more species, but they are forgotten in the modern era following the Bennett juggernaut. Chris is a lumper.
I know it’s important to ‘go with the flow’ of current literature, but it’s out there begging to be tested.
It would be great if you could follow with some gender differences in other pterosaurs. If you have them, trumpet them.
You don’t ‘test’ ontogeny with cladistic analysis. Nor do you ‘test’ crests, pelvic morphology or diagnoses of species. That is not how these things work. And the reason you can’t see these kinds of differences in other species is that either we have very few individuals of a given taxon, or we have no good 3-D material to compare, or both.
Chris may be a lumper, but not much of one. Species (and other ranks) as diagnosed by taxonomists are based on observation. He decided (and other agree) that the variation seen is intraspecific or a result of ontogeny, not speciation. You simply do not seem to understand how taxonomy works. How would you ‘test’ species with cladistics? That makes no sense? Cladistics does not and cannot do that kind of analysis.