Now this, this is a tricky one. Like some of the other areas I have (and indeed will) write about, this is one of the ones where many people seem to have dipped their oar in over the years and everyone seems to think they know what’s going on, when the truth is in fact quite a bit more complex when you start digging around under the surface.
Well, I have done a lot of digging and in some rather unusual ways as this subject made up the bulk of my PhD at Bristol and of course as a result I was pretty much immersed in this problem for the best part of three years. I also still have a few things ongoing and have been able to discuss this problem with just about everyone involved at the Wellnhofer pterosaur meeting in Munich. So hopefully this will be as definitive as possible, but prepare for some disappointment – it does not end well.
To cut a long story short, pterosaurs are damned difficult to place in the reptile tree. Skipping over a few historical anomalies (pterosaurs as flying marsupials, pterosaurs as bird ancestors and far worse) we are left with actually quite a limited, but complex, rage of possibilities. Thanks to the two posterior holes in the skull (the upper and lower temporal fenestrae for those in the know) we know pterosaurs are diapsid reptiles, but after this it gets really tricky, and the diapsids are a big group (it includes snakes, crocodile, dinosaurs, lizards, various marine reptiles, and all kinds of odds and sods like rhynchosaurs).
Pterosaurs, rather obviously, are highly adapted for their flying lifestyle and as a result their anatomy has changed a huge amount. Not just the obviously elongate fourth finger, but the skull bones have kind of merged, the humerus has changed dramatically to accommodate the flight muscles and the vertebrae have reshaped themselves to provide better support and plenty more besides. What this means in practice is that all those little anatomical details we rely on to link clades to each other have either disappeared, or changed so profoundly we are not sure what they have evolved from.
Normally we use a method called cladistics (it sort of tots up the similarities and differences of these anatomical changes to give you a family tree) to link together taxa but when the differences are few and far between things get tricky. Basically we don’t have enough information to link the pterosaurs to any of the diapsids with great confidence. Cutting out the details, we essentially have two options – pterosaurs either evolved from the dinosauromorphs (basically dinosaur-like reptiles) or from the base of the prolacertiforms (a group of early diapsids that includes all kinds of odd-balls). However, there is only a very limited amount of evidence to tie them to either group, so the whole thing hangs in the balance.
There are problems with both possible solutions – the prolacertiforms (and their relatives) are all pretty chunky, long-necked, sprawling lizard-like beasts which are hardly the best stock to produce small, short-necked, arboreal flying basal pterosaurs. With the dinosauromorphs they tend to have long legs and short arms (not very pterosaurian) and are bipedal (very un-pterosaurian – despite what you might have read). Hmm, not much to go on there then. When pterosaurs fall out with the dinosauromorphs, they almost inevitably end up in close association with Scleromochlus. Sadly this reptile is exceptionally poorly known and there is a great limit to what we can accurately determine about its skeleton which does not help matters.
How can we solve this problem? Well, we can add more characters to the analysis in the hope to accumulate more evidence (which is not really possible as we have pretty much exhausted them already), we can try new methods of cladistics to look for new solutions (the basis of my PhD and it didn’t work – ***Important point alert*** – Science does not always work!!!) or we can try and find new fossil material. Oddly enough the last of these actually looks the most reasonable right now, which should tell you a lot about the others! There is no guarantee that any possible ‘transitional’ pterosaur even fossilised, let alone that we will ever find one. Right now we are stuck without a pterosaurian Archaeopteryx (something that would unambiguously link pterosaurs to another diapsid group with unique characters) and given the size and age of such a creature it will probably never be found.
So there you go, placing pterosaurs is really tricky. But what is the answer? I have avoided giving you any hints until now, so here it is: drum-roll please……
Pterosaurs are……..
DINOSAUROMORPHS!
Probably.
We think.
Look, I told you it didn’t end well. The truth of the matter is that currently the best supported hypothesis is that pterosaurs derived from the dinosauromorphs and thus are very close relatives of the dinosaurs. However, it would not take much to swing the balance the other way and move pterosaurs to a more basal position in the reptile tree to link them with the prolacertiforms. As I said above, this one hangs in the balance, and at the moment, just tips towards dinosauromorphs.
However, at the risk of being really irritating, I do know of a couple of analyses that are slowly making their way towards publication that might actually reinforce the ‘dinosauromorph’ argument and make things a bit more clear cut. There are also a few anecdotal links between pterosaurs and dinosaurs that suggest a link and which would firm up this relationship still further. Nothing concrete for now, sadly, but the scales are tipping and right now the money is on dinosauromorphs.
PS Anyone in the know should feel free to not mention Sharovipteryx – it is simply not worth the effort. It has nothing to do with pterosaurs. Really.
This is a revised Mk.1 post, to see the original with comments etc., go here.
Dave Hone's Archosaur Musings 
David, I was curious about your latest writings on pterosaur origins, and this seems to be it, unless there are some pdfs floating around that can be emailed. Please also copy me off-blog (email provided in text below).
From this post, you report, “the prolacertiforms (and their relatives) are all pretty chunky, long-necked, sprawling lizard-like beasts which are hardly the best stock to produce small, short-necked, arboreal flying basal pterosaurs.”
And, “Anyone in the know should feel free to not mention Sharovipteryx – it is simply not worth the effort. It has nothing to do with pterosaurs. Really.”
First, I appreciate your comment that basal pterosaurs were small, had short necks and were arboreal (i.e. MPUM 6009 and various anurognathids and proto-anurognathids), as opposed to Austriadactylus, Eudimorphodon and company. That’s a good start.
Second, I have to object to your disparaging comment characterizing ALL the so-called “prolacertiforms” as chunky, long-necked srawlers. Although much of what you say is indeed correct, you seem to be missing a point.
Let’s take this word at at time:
1. prolacertiforms. > In order to be a prolacertiform, you have to be related to Prolacerta. You saw my presentation at the Wellnhofer conference in 2007. From it, the only relatively complete taxa related to Prolacerta from the traditional list of prolacertiforms are Protorosaurus and Boreopricea. And this clade is indeed basal to Proterosuchidae. Dilkes 1998 and Rieppel, Fraser and Nosotti 2003 have also found a division between these three and the rest of the so-called prolacertiforms(tansystropheids, etc.) which we are more interested in because therein lie the sister taxa of pterosaurs. RF&N 2003 could not resolve the issue because their inclusion set did not have a sufficiently wide gamut (more on this below.
2. chunky. > Pareiasaurs and turtles are chunky. Tanystropheus, Langobardisaurus, Cosesaurus, Macrocnemus are certainly not. Sharovipteryx is the polar opposite. Which taxa do you have in mind when you say ‘chunky’?
3. long-necked. > Certainly true of most so-called prolacertiforms–but not all. As you listed in your above cladogram, Longisquama is the closest sister taxon to pterosaurs (Peters 2000). You’ve evidently forgotten or overlooked the fact that it has as short a neck too. The first five cervicals are shorter than the next three in sum, as in MPUM 6009, the Milano specimen attributed (incorrectly) to Eudimorphodon. Cosesaurus does not have a particularly long neck either.
With regard to cervicals, Scleromochlus also has a short series, but it is composed of only six, or at the most seven, cervicals. That’s one (or two depending on how you count them)short of the pterosaur/cosesaur/longsiquamid number. The proportions and rib shapes are different in Scleromochlus, which is more like Gracilisuchus.
4. Sprawling. > Brings up a good question: weren’t all pterosaurs sprawling? That’s all I’ve ever heard, except for Padian 1983, etc. MPUM 6009 and several other taxa do not have a distinct femoral neck with a head inturned 90 degrees, although some early pterosaur taxa did. Even so, Peters 2000 (the Ichnos paper) demonstrated that a sprawling Cosesaurus (a pterosaur ancestral sister) exactly matched certain Rotodactylus tracks, which are, narrow-gauge. Then there’s Renesto, Dalla Vecchia & Peters 2002 Morphological evidence for bipedalism in the Late Triassic Prolacertiform reptile Langobardisaurus, another sprawler. I would also encourage you to visit
http://bioweb.ad.uc.edu/faculty/jayne/bruce.htm
to see some modern sprawling lizards get up on their hind legs. Check out the narrow gauge and digitigrady.
Manipulable skeletons of pterosaurs further demonstrate that even with sprawling femora, so long as the knees are below the plane of the acetabula, a right angle at the knee will return the ankles closer to the parasagittal plane. Heck, you can even
Lizard-like. > This is excellent, because, as you’ll remember from the Wellnhofer symposium, I found the basal lizard Huehuecuetzpalli (Reynoso 1998) to be at the base of all the rest of the so-called “prolacertiforms,” far removed from the real prolacertiforms (i.e. Prolacerta) and archosauriforms. Like pterosaurs and all the intervening taxa, this lizard does not have a terminal naris, but an elongated one displaced posteriorly. It has a short neck. It has all of its fingers and toes, without any of those vestiges that plague other archosaurian candidates. In particular pedal digit V has an elongated proximal phalanx, exactly what one sees in all the intervening taxa leading to pterosaurs (Macrocnemus and drepanosaurs lose the length in this phalanx and are not on the line leading to pteros). In Huehuecuetzpalli the tail is attenuated, as in all intervening taxa, without those deep chevrons that plague other archosaurian candidates. Unlike all other lizards, but like pterosaurs and all intervening taxa, the tarsus was not fused, and had a simple hinge ankle joint! It’s a basal form. The intervening taxa (Cosesaurus, etc.) fill in most of the other characters (like an interclavicle keel, sternal complex, antorbital fenestra, multi-cusped posterior teeth, strap-like pectoral girdle, as I wrote in 2000 (the Rivisita paper).
Sharovipteryx. > Not worth the effort? Why make such a blanket statement without filling in with a few juicy details? As you know, it was a contemporary of Longisquama and the basalmost pterosaurs in the Late Triassic. These three taxa were certainly as distinct from one another as three Dr. Seuss characters, but when compared to hundreds of other reptiles in cladistic analysis, these three oddballs, along with Cosesaurus, do indeed find each other in a single clade, the Fenestrasauria (Peters 2000). Cosesaurus preceded the other three by several tens of millions of years and sister taxa (trackmakers of Rotodactylus) were wide ranging across North America, Africa and Europe, so Sharovipteryx, Longisquama and pterosaurs represent divergent sisters, like Coelosphysis, Fabrosaurus and Plateosaurus from roughly the same era.
While Sharovipteryx and pterosaurs were polar opposites in the same way that a kiwi and an albatross are polar opposites, they shared a suite of unique synapomorphies including that famous elongated fifth toe, attenuated tail with hemals parallel to centra, hyper-elongated ilial processs both anterior and posterior together with an imperforate puboischiatic plate with prepubes and uropatagia. Otherwise the tibia is longer than the femur, the lateral toes and metatarsals are longer than the medial ones, the torso is relatively short, the scapula and coracoids are strap-like, the sacral series is expanded and such, but all these are shared with other taxa too.
You can’t dismiss Sharoviptyerx and Longisquama as pterosaur ancestors because they did not have longer forelimbs. Scleromochlus falls into the same trap. So do any predecessors of Archaeopteryx.
So, enough with the mystery. What about Sharovipteryx removes it from candidacy as a sister taxon to pterosaurs? And what is it closer to?
Finally, you report: “Pterosaurs are…….. DINOSAUROMORPHS! Probably. We think.”
Sounds like you’re not sure. Hopefully you’re not lumping Lagerpeton in with dinos, because it nests much better with Tropidosuchus, as you may know already if your inclusion set was broad enough.
If you’re really interested in the answer to what pterosaurs are, send me your latest paper. I’ll tell you if I see any problems. I’ll also congratulate you if I see any insights. You should know that much has been discovered in fenestrasaurs since 2000,2002, and for that matter, 2007, some of which is scheduled for publication. I think what I can share with you will resolve your cladistic turmoil.
All I want to do is learn. And you can help.
David Peters
davidpeters@att.net
ps: not sure where that happy face icon came from.
I will keep this short because I have no desire to get drawn into a very dull discussion on things that frankly lie outside of the technical literature and thus are not really relevant to this forum. First off though I will say that this post is aimed at a general audience which explains several of the points below (e.g. I simply did not want to talk about Sharovipteryx and to confuse the issue, so ignored it, however I mentioned it so that non-interested people would know there are other taxa out there, and the knowledgeable would not think I had missed it, I say “all prolacertidforms” since I don’t want to discuss the 15 different taxa that have been placed in there at various poins). Also, please don’t post things like this online and then ask me to reply in person, what is the point of that?
1. Plenty of people (including me) have noted that Prolacerta is not a prolacertifrom. Thus the clade needs a new name an definition.
2. Chunky. Big. Scleromochlus is not, Longisquama and Drepanosaurus are.
3. See above. Also, this statement has been repeated in the literature plenty of times before and indeed has been considered a characteristic of prolacertifroms (e.g. See Benton’s Vertebrate Palaeontology book).
4. Pterosaurs are not sprawlers according to a great many people – indeed they are considered to have an upright posture that has ‘spread out’ if you like. Humans are upright, but if you tied our ankles to a pool cue and made us squt we could still walk in a sprawling gait.
I can indeed not discuss Sharovipteryx since it’s a) my blog, b) I wrote it, and c) I did not want to commet on it at this time, especially in the light of d) having papers on this subject (at the time this was written) in prep or in press.
“You can’t dismiss Sharoviptyerx and Longisquama as pterosaur ancestors because they did not have longer forelimbs. Scleromochlus falls into the same trap. So do any predecessors of Archaeopteryx. “
No, I can dismiss them because they have very little in common with pterosaurs, just as I won’t bother to consider rhynchosaurs or snakes either. And check out recent papers on bird ancestors / sister taxa if you think thay have short arms. Shorter, obviously, but not short.
“Sounds like you’re not sure.”
I’m not sure, just cautious (espeially in a public forum and relating to unpublished material). I go with the data and analyses, not assumptions, and those are finely balanced but on the side of dinosauromorphs.
“If you’re really interested in the answer to what pterosaurs are, send me your latest paper. I’ll tell you if I see any problems. I’ll also congratulate you if I see any insights.”
The arrogance of that statement stuns me.
David,
Sorry to raise your ire. I simply wanted a cc offline in order to access your new email address, which I do not have and cannot find. Your BSPG address bounced back. With your email I can provide you with unpublished and pre-published data of interest.
I note a common thread in all your arguments: “plenty of people” “the literature” “a great many people” as sources for your arguments. Wouldn’t it be much better if we could talk about taxa and characters? Since this clade has such a variety wouldn’t it be more scientific to acknowledge the exceptions to the rule? Like the short neck in Longisquama and to a lesser extent, Cosesaurus? On the other hand, if a long neck is indeed a requirement for memberships within “prolacertiforms” then do these two need to be excluded?
Next, your comment about arrogance comes unexpectedly. Having had numerous reviews of my manuscripts, it’s easy to spot referees with personal bias by their complete lack of positive comments. Those who wish to be helplful combine criticisms with congratulations. By making the comment I did, I only wished you to see me as one of the latter sort, a helpful peer.
Finally, with regard to pterosaur origins, Sobral and Langer 2008 wrote an abstract in the Journal of Vertebrate Paleontology based on a supertree that found Sharovipteryx, Longisquama and pterosaurs to be sister taxa. Would you like to comment on that, either privately or on this blog?
Best regards and many apologies,
David Peters
If you want to contact me, ask then. Don’t fire off a huge comment. My address is online (here, and elsewhere), and on my recently published manuscripts, and in Chris Bennett’s mailing list and people like Dave Unwin have it.
I use such abbreviations as, as I mentioned above, I am writing a general essay for a general audience. I am not doing detailed analyses of taxa and characters for that very reason. I am not avoiding discussing them per se, I just don’t want to do it here. A I say, I think that is a technical discussion for technical papers, or a techncial blog / mailing list. This is not that forum. Of course I am interested in taxa and characters, but one must start with the exisitng data and work from the fossils. Clades are based on multiple characters, not individual ones, though of course it is useful to pick near universal ones to characterise a clade (mammals have fur, birds have feathers).
As for arrogance, if that is your concept fine, but if you word in in such a way that you offer to solve my problem and tell me my failings in an area in which I am an expert and you, frankly, are not, how exactly did you expect me to see it?
As for the supertree, I have not seen the paper, so I won’t comment in detail. But if you read my latest paper on supertrees and pterosaurs, and indeed other papers I have produced on the subject, and notes on supertrees on this blog it comes as no surprise at all. It does not make it right, supertrees as odd things that handle data in unusual ways and can be highly odd in the wa they combine and construst output trees.
I personally think a place within Ornithodira is the best for pterosaurs. Both pterosaurs and dinosaurs have pulmonary air sacs, which likely only evolved once. Plus, it would be interesting if the dinosaurian feathers and the pterosaurian “fur” had a same origin – likely something like the ceratopsian quills.