Guest post: The shape of pterosaur evolution

It would not be a huge surprise if you had missed the release of this paper, since it has only been published online so far, and does not seem to have been picked up by the media or any other bloggers as yet:

Dyke, G.J., McGowan, A.J., Nudds, R.L. and Smith, D. 2009. The shape of pterosaur evolution: evidence from the fossil record. Journal of Evolutionary Biology (Online February 2009: DOI: 10.1111/j.1420-9101.2008.01682.x).

The crux of this paper is the idea that, contrary to oft published belief, there was no great battle between birds and pterosaurs in the Mesozoic and that they co-existed with no major ecological problems. My old friend Al McGowan (one of the authors, and a real AAB mainstay) has kindly penned this extended guest post to cover in some detail, what they did, how and why, and what it means.

Pterosaurs and birds not competing. Image courtesy of Todd Marshall, used with permission.

Pterosaurs and birds not competing. Image courtesy of Todd Marshall, used with permission.

What is the paper about?

For many decades there has been a debate about the role of the rise and radiation of birds in the decline and eventual extinction of pterosaurs. Many readers may have a nagging feeling that birds must have some been ‘better’ at being flying animals. Just think of how the word ‘dinosaur’ is used in English to describe people whose ways or views are ‘out-of-date’. Such views are embedded in most cultures, a topic which the late Steven J. Gould wrote about extensively. Most previous studies into pterosaur evolution have focused on describing and interpreting the anatomy of the group, detailed and highly skilled work. However, we used a different approach. We combined a number of lines of evidence on the evolution of diversity (how many species) and disparity (how much morphological variety) throughout the history of the Pterosauria between the Late Triassic (around 220 million years ago) until the Cretaceous/Tertiary extinction (a small aside, geologists now talk about the Cretaceous/Palaeogene boundary, due to changes in the way we use names of intervals of geological time). We found good evidence that, contrary to previous views, pterosaurs continued to produce new species and innovative morphologies until at least the end of the Early Cretaceous. Previous work had highlighted the appearance of birds during the Late Jurassic (Archaeopteryx is around 145 million years old) and implicated birds in the decline of pterosaurs.

What was your contribution to the paper? My background and expertise as a systematist is in ammonids (extinct molluscs that look like Nautilus, but more closely related to living squid and cuttlefishes), but in the course of my research career I have developed a broad range of skills in data analysis and statistics. So anyone reading this who is thinking of being a palaeontologist, remember that maths, statistics and computer programming are as important as a knowledge of fossils and geology. I contributed analyses on the changes in pterosaur morphological variety (‘disparity’) throughout their history. One way to estimate these changes numerically is through the use of mathematical spaces called morphospaces. Morphospaces are a sort of graph that specimens plot into as a function of aspects of their morphology. As an example, think about the people in your class or work group. You could measure the height and weight of all those people and plot them in a morphospace based on those two measurements, a familiar x-y type graph. Even with these two measurements you could do some interesting things. If the group is comprised mostly of adults, then you would probably be able to work out who was male and who was female as the males will tend to be taller and heavier. With the pterosaur data we used measurements of the major elements of the fore- and hind limbs and then used a statistical technique called Principal Components Analysis (PCA) to produce a series of summary plots like the x-y graph example. I was then able to use some other statistical methods to express the changes in morphospace occupation through time with numbers, although making and examining plots is an important part of data analysis too. The values I got from these analyses indicated that pterosaurs continued to diversify morphologically until the Early Cretaceous. An interesting observation about the way that pterosaurs filled the morphospace was that it followed a similar pattern to extant birds. First the centre of the morphospace was filled, then new pterosaur groups would appear at the edges of the morphospace. A problem with just leaving the analysis at this point is that there is variation in how many specimens are sampled from each time interval. Differences in sample sizes can result in apparent differences in the values you are interested in that are could be due only to the differences in sample size. To return to the example of the people in your class or workplace, imagine picking a sample of five people then a sample of fifty people. The second sample is more likely to pick out the biggest and smallest people in the group, resulting in differences in your estimates of the morphospace occupation: yet both samples are from the same group. A statistical technique that can compensate for this problem is called rarefaction. Using a computer you can repeatedly draw samples of uniform size (usually equal to one less than the size of your smallest samples) and this will indicate whether an observed difference could (not is) the result of differences in sample size. Using this technique we were able to show that Early Cretaceous pterosaur disparity was unusually high. My other major contribution to the paper was to use rarefaction techniques to compare the observed diversity to the number of specimens available from each interval to check if differences in sample size could be responsible for differences in diversity in both birds and pterosaurs. Using this technique we were able to demonstrate that, based on the currently available data, pterosaurs were more diverse than birds during both the Early and Late Cretaceous.

What did your co-authors do?

Gareth Dyke (University College Dublin) and Robert Nudds (University of Manchester) are specialists in bird research, both living and extinct. Gareth has broad experience in vertebrate palaeontology and his student Donal Smith helped collect some of the data we used in the paper. Robert is a functional morphologist with interests in animal locomotion, the evolution of flight and bird evolution in general. Gareth put together the team and kept things moving along. We were working from three different institutions, but email makes such collaborations much easier than they were before. Gareth and Robert focussed on using two other data sets to provide independent lines of evidence about changes in pterosaur diversity through time. Independent evidence is important, as the more analyses that are done that lead to similar conclusions, the more confident we can be that our findings are correct. Gareth and Robert used a family tree, called a phylogeny or evolutionary tree, of pterosaurs to provide another means of counting pterosaur diversity in each time slice. As evolutionary trees do not rely on recovering specimens in each time slice, they can compensate for intervals where the fossil record of a group is poor. Pterosaurs and birds both have poor fossil records, particularly compared to marine seashell fossils. As the evolutionary tree makes estimates about when groups diverged based on fossil occurrences and the geometry of the evolutionary tree, we can arrive at reasonable estimates of diversity during each interval. An introduction to this area can be found in a series of columns by Dr Peter Forey, an expert on reconstructing the evolutionary relationships of fossil fishes, in the Palaeontological Association Newsletter. The other data they handled dealt with a range of methods for estimating how our knowledge of the pterosaur fossil record has improved over time, and how confident we can be that the classification of pterosaur fossils into species, genera and families is likely to remain the same over time.

What inspired you to become involved with the research? Partly because I enjoy working with Gareth, who I have known for over ten years. We would hatch research plans in the pub, and sometimes these plans actually work out! I have a lot of fun working with Gareth, which is another important part of research. We are, after all, paid to do our hobby. The reason I have been particularly involved with these projects on flying vertebrates is that they are excellent systems to test ideas about the influence of competition among major groups in shaping their evolutionary fates. The first project I worked on with Gareth examined all three groups of flying vertebrates; birds, bats and pterosaurs, which was also published in the Journal of Evolutionary Biology in 2007. We were interested in testing ideas about competition among these three groups. I was very inspired by a paper by Penny & Phillips (2004) in the journal Trends in Ecology and Evolution. That paper took the classic example of dinosaurs and mammals at the Cretaceous/Palaeogene boundary and showed how to use a combination phylogenetic, morphological and ecological data to place such events on a graded scale of models that ranged from a genuine role for mass extinctions in triggering radiations to cases in which the radiations started much earlier in the fossil record, and the supposed beneficiaries of the mass extinction event had, in fact occupied many ecological niches before the boundary. We were able to modify this approach in our earlier paper. This paper focusing on the pterosaurs is an example of palaeontology’s unique contribution to our understanding of the history of life. As birds and bats, the other two groups of flying vertebrates, have living members there have been excellent studies of the diversification of these groups based on molecular family trees. As pterosaurs are completely extinct, the task of understanding their evolutionary dynamics falls to the palaeobiologists.

Figure 2 (Dyke et al., 2009) showing A - numbers of species for birds and pterosaurs, and B - diversification rates

Figure 2 (Dyke et al., 2009) showing A - numbers of species for birds and pterosaurs, and B - diversification rates

So if birds and pterosaurs weren’t competing, what were birds doing differently? Many Mesozoic birds were apparently similar to modern birds such as divers (loons to North American readers) and cormorants which dive for fish from the surface, something we doubt pterosaurs were capable of doing. Other birds known from the Cretaceous are about the size of modern song-birds (the most diverse group of birds) and may have eaten seeds, something we don’t think pterosaurs did. Both groups ate other organisms, including shellfish, insects and fish. Ecologists working on living taxa are currently involved in work on a theory that most competition experienced by individuals is with other members of their own species. More support for this idea might make palaeobiologists less willing to look for interactions among major groups as a reason for the rise and fall of these groups.

What are the main points you want people to take from the paper? Three things: Using a wide range of data and statistical techniques is very important for understanding the evolutionary dynamics of groups in the fossil record. The evolutionary history of pterosaurs was probably not greatly influenced by Mesozoic bird radiations. Our knowledge of the fossil record and evolutionary classifications is continuing to improve, especially in the last twenty years. The money and effort spent on such research is really improving our knowledge.

Where can I find out more? Unfortunately, the paper is published by the Journal of Evolutionary Biology. They do make papers freely available after two years, so you can get the earlier (McGowan & Dyke 2007) paper from their website soon. Many sites exist on the web that deal with the ideas discussed. Wikipedia and Google are your friends. Remember, articles from university websites and museums are more likely to be accurate and up-to-date. You can also ask questions through Dave’s blog, or through the Ask-a-Biologist web forum.

I’d just like to add a few words of my own to the end of this to say that while the data and analyses are convincing, don’t be trapped into thinking this is the end of the debate. Certainly I think there are perhaps some other factors at play which might prove to be difficult, if not impossible to test. One can certianly argue that there is at least some ecological separation simple based on the typical size of the pterosaurs in the Cretaceous being substantially larger than that of the birds, and there would doubltess be some direct local competition between some species in some places. As Al says, this suggests there is no great impact overall of birds on pterosaurs in term sof diversity, not that there was none, or even very stong pressures in some places. It is however a great start, and shows the importance of testing ideas emperically (where possible – palaeontologists often have onyl a single data point to play with) and how interesting apaprently simple and ‘fixed’ ideas can be. There will be plenty more to come on this I am sure as our knowldge of both birds and pterosaurs increase in the future.

3 Responses to “Guest post: The shape of pterosaur evolution”

  1. 1 220mya 01/03/2009 at 8:19 am


    This is a great plain-language summary of your paper. I know that some of your other work has focused on the impact of the rock record and outcrop area on paleodiversity curves. I haven’t had a chance to read the paper yet, but I was wondering if you addressed the potential impact of outcrop area on pterosaur through time?

    A second question – has there been any work to see if the rock record affects patterns in disparity through time? This might seem like an unusual question, but if disparity correlates with ecology, specific paleoenviroments would be predisposed to preserving taxa from a certain part of the morphospace.

    Randy Irmis

  2. 2 Will Baird 12/03/2009 at 1:06 am

    I have a query as well, and I apologize for the lateness which it comes. However, I have to wonder about something that Dave stated. pterosaurs and birds weren’t in the same niches, I think[1], due to their size differences. Could you have restricted your analysis to where their body mass overlapped and how would that have changed (if at all) your results?

    1. Not being a pterosaur guy (or really a paleo type in general, jsut an interested party), I got the impression that the pterosaurs were rather different in their practices of raising their young. Unwin mentioned in his pop sci book that when pterosaurs hatch, they’re just about ready to fly. If that’s the case, could the uber sized pterosaurs filled multiple niches as they grew and competed with the birds that way? It really depends on the rate of growth though, I’d think…

  3. 3 Al McGowan 14/04/2009 at 9:31 pm

    Sorry, just got round to checking back for comments.

    Randy, we didn’t address the link between rock availability and pterosaur diversity in our paper, as I knew of a couple of other projects that were working on this issue. When we started this paper I was concentrating mainly on the morphospace side of things, but then had to move to cover rarefaction issues as well to satisfy a reviewer. Gareth and Robert did a huge amount of work on the phylogenetic side of the paper. What I hope is clear from my summary is that we need multiple approaches to understanding biodiversity patterns.

    The closest papers that I can think of that address the question about disparity are the work by Roy et al. (2001) and work by Pascal Neige (2003). These papers were more about spatial packing of disparity, but pointed up an interesting pattern where you can have high diversity but low disparity and vice versa. Molecular phylogenies were critical in both cases, as they allowed the researchers to sort out whether the lineages invovled were closely or distantly related without reference to morphology. So rapid radiations from a single lineage (think African Rift Lake cichlids) could be distinguished from cases where many distantly related lineages co-exist in the same area.

    It is somewhat tricky to work out how disparity might respond to flucutations in the rock record. Shanan Peters and Michael Foote, who really got people excited about this issue again, used geological formations as an attempt to cover a range of sedimentary habitats, and reasoned that more formations would equal a wider range of taxa and higher diversity. Trouble is that disparity isn’t clearly linked to diversity, as I have explained above. So it is actually an interesting empirical question. Any ecologists reading this who can shed light on what patterns are like in extant taxa, please do.

    I would also watch for any publications from Marcello Ruta at Bristol, as I think he may have something in the works.

    Will, thanks for the question. Your reasoning is good, and the notion that rather than dividing niche space up by species, pterosaurs might do this among age-classes is a valid hyptothesis that I have also heard discussed about dinosaurs and mammals. However, I tend to agree with the view that most competition organisms experience is with members of their own species. The trouble with testing these sort of hypotheses in detail is that abundance data (how many individuals of each species of interest) are the best way to do this. Such data would be very difficult to gather about pterosaurs and birds. But someone may come up with a way to do it.

Comments are currently closed.

@Dave_Hone on Twitter


Enter your email address to follow this blog and receive notifications of new posts by email.

Join 567 other followers

%d bloggers like this: