Two weeks ago I and colleagues published a new paper in the august open-access online pages of PLoS ONE. Led by Aleksandra Birn-Jeffery of the Royal Veterinary College, and co-authored by Charlotte Miller, Emily Rayfield, Dave Hone and myself, the paper is titled ‘Pedal claw curvature in birds, lizards and Mesozoic dinosaurs – complicated categories and compensating for mass-specific and phylogenetic control’ (Birn-Jeffery et al. 2012).

This study is the latest in a long-running series of contributions on the hypothesised relationship that seemingly exists between behaviour and claw curvature in birds and other tetrapods; though restricted to foot (or pedal) claws alone (we ignore manual – or hand – claws entirely), it represents the grandest of such studies in scope and amount of data. While we certainly have results that are worth talking about, one of the take-home points of the study concerns the ambiguity and overlap that exists when we try to match claw curvature with behaviour.

Incidentally, this is another of those studies that was achingly long in the making. Things started with Charlotte’s Masters thesis in 2004-2005 (produced under Dave’s supervision) and it grew in scope and content over the intervening years. I feel that we all brought something useful to the table and the end result, with Aleksandra at the helm, is a strong paper. It goes without saying that this paper represents another triumph as goes open-access publishing. I think this is the third paper I’ve published in PLoS ONE. I have two more in press or in review.

A history of ideas on claw curvature

The idea that claw curvature might represent some sort of special and reliable indicator with respect to behaviour and lifestyle is a fairly old one, perhaps best known due to its mention in the literature on Archaeopteryx. In 1974, John Ostrom illustrated the pedal unguals of various perching, predatory, climbing and ground-running birds alongside those of four of the Archaeopteryx specimens (Ostrom 1974) [note to theropod nerds: I will skilfully avoid the thorny topic of archaeopterygid systematics in this article]. Note that Ostrom illustrated pedal unguals – the bones that support foot claws – and not the keratinous claws themselves.

Because the Archaeopteryx unguals look weakly curved and have weakly developed flexor tubercles (the convex lumps on the undersides of phalanges that anchor the muscles and ligaments involved in digit flexion), Ostrom (1974) argued that the ungual morphology of Archaeopteryx was most consistent with a cursorial lifestyle, not one involving perching or climbing. Yalden (1985) also illustrated claws from a variety of living animals (this time, mammals as well as birds). However, he argued that comparison with living animals showed that claw curvature in Archaeopteryx is more like that of climbers.

As just mentioned, Ostrom used pedal ungual shape, not the shape of the keratinous claws themselves. We can see from living animals and well-preserved fossils that keratinous claws typically represent horny ‘extensions’ of the underlying bony cores: they make the claw in the living animal longer, more curved, and more pointed than the ungual bone is on its own. In some living animals (fruit bats, apparently, are a good example) the keratinous claw is substantially different in form to the ungual.

However, there’s some uncertainty over whether this is universally the case, since there are other living animals where the keratin sheath isn’t that much longer, or that much more curved, than the ungual (examples: some crocodylians, some ratites). This raises an interesting question: are unguals and their overlying keratinous sheaths similar enough in shape that they provide the same ‘signal’ as goes behaviour? We were only able to test this relationship preliminarily and found a significant difference between unguals and keratinous sheaths. Unguals are shallower and less curved than keratinous sheaths (Birn-Jeffery et al. 2012, p. 10).

Anyway, as is obvious from Ostrom’s (1974) and Yalden’s (1985) comparisons and conclusions pertaining to Archaeopteryx, several generalisations can be made about how claw shape (both ungual shape and keratinous sheath shape) relates to behaviour and lifestyle. In general, animals that run or walk on the ground exhibit weak claw curvature, those that climb up or cling to vertical surfaces have laterally compressed claws with needle-like points, and those that perch in trees have strongly curved, conical claws. Markedly curved claws are also (typically) seen in animals that use their extremities in apprehending prey.

Making generalisations ‘by eye’ is one thing, but there’s an undeniable degree of subjectivity when it comes to deciding how ‘curved’ a given claw is. Feduccia (1993) added rigour to this area by introducing a measure of how curved a claw is: it involved measuring the degree of curvature present across the inner (or ventral) curve of the claw.

Feduccia (1993) included a sample of over 500 bird species, but deliberately excluded raptors*, long-legged birds and various other ‘unusual’ kinds, presumably because they might cloud the results. Birds grouped into ‘ground’, ‘percher’ and ‘climber’ clusters, and Archaeopteryx – the main focus of his study – grouped among ‘perchers’ according to its foot claw curvature and among ‘climbers’ according to its hand claw curvature (Feduccia 1993) [see the figure below]. The three categories overlapped extensively and a criticism is that his grouping obscured the continuum that exists between them (Glen & Bennett 2007): many ‘perchers’ are actually generalists that spend much of their foraging time on the ground, for example (Pike & Maitland 2004). Another criticism (Dececchi & Larsson 2011) is that Feduccia’s exclusion of predatory birds disallowed fossil taxa (like Archaeopteryx) from falling into anything other than the 'percher' or 'climber’ clusters: raptorial birds overlap with perchers and climbers (Pike & Maitland 2004) in claw curvature yet the species concerned are not ordinarily specialised for perching or climbing.

* Yes, I mean proper raptors… how I so hate the fact that a perfectly good, oft-used name is now hopelessly ambiguous whenever birds and non-bird maniraptorans are referred to in the same discussion.

Since the publication of Feduccia’s 1993 paper, other authors have argued that claw curvature of the sort seen in Archaeopteryx is not demonstrative of a scansorial or arboreal lifestyle (Peters & Görgner 1992, Chiappe 1997, Pike & Maitland 2004, Glen & Bennett 2007, Dececchi & Larsson 2011). Other workers have applied claw curvature analysis to non-birds, including dromaeosaurids (Xu et al. 2000, Manning et al. 2006, 2009, Parsons & Parsons 2009) and the Triassic archosauromorph Trilophosaurus (Spielmann et al. 2005) (there's also a short paper out there on pterosaur claw curvature but, despite having seen it once and allocated its existence to memory, I cannot find it today, nor does anyone I know own it). And interest in the claw anatomy of lizards means that there is now at least some literature on claw curvature and its correlation (or lack of correlation) with behaviour in these animals too (Zani 2000, Tulli et al. 2009).

What we did

We aimed to advance our understanding of this area by analysing an increased number of taxa, by seeing if the inclusion of a distantly related group (namely, lizards) affected the analysis, by looking at different kinds of claw measurement, and by testing claw curvature against phylogenetic control (Birn-Jeffery et al. 2012). A huge amount of data was collected: Aleksandra and Charlotte measured over 830 specimens belonging to 331 species. And several individuals of the sampled species (as many as six) were measured and included in the study.

Inner claw curvature (the measurement analysed by Feduccia) and outer claw curvature (the measurement analysed by Pike and Maitland) were both recorded, as was the vertical height of the claw at mid-length. The claw of pedal digit III is preferred for these studies, since it’s longest and is the one that interacts for the longest time with the substrate during locomotion. However, we also took data from the claws on the other digits and compared this with the digit III measurements when appropriate. We also included body mass data because one hypothesis we aimed to test was whether mass had any link with claw curvature (note that Pike & Maitland (2004) tested for the effects of mass on their dataset too). And we also mapped species onto phylogenies in order to see if relatedness had any special control over claw morphology. Animals were placed into four behavioural categories determined by dominant aspects of their lifestyle: we recognised ‘predatory’, ‘climber’, ‘percher’ and ‘ground-dweller’ categories (Birn-Jeffery et al. 2012).

Some basic (and, perhaps, predictable) conclusions were arrived at, by which I mean that we found some of the generalisations made about claw shape to be reliable… approximately. The claws of ‘ground-dwellers’ are less curved than those of animals in other categories, and ‘ground-dweller’ claws are also deeper at mid-length than the claws present in animals belonging to other categories. Animals in the ‘ground-dweller’ category were more variable in both inner and outer curvature measurements and in mid-length claw height than members of other categories (Birn-Jeffery et al. 2012), perhaps showing that animals with phylogenetically distinct backgrounds and histories become ground-adapted via numerous different evolutionary ‘routes’.

There are caveats

Otherwise, things turn out to be more complicated than we might prefer and our results are, to a degree, confusing and even conflicting. The animals we measured did not group into neat sets, there was substantial overlap between sets, there were various outliers and exceptions, and different data sets sometimes gave different conclusions.

But this isn’t wholly a surprise. As discussed above, we already know that grouping animals into behavioural categories is difficult and relies on making broad (and perhaps unacceptable, even erroneous) generalisations. We also know that animals consist of more than just claws, so conclusions based on claw morphology alone need to be taken as provisional: there’s a lot of other data that needs to be considered if we want to properly develop hypotheses about an organism’s way of life*. And claw morphology may represent a compromise between different aspects of an animal’s lifestyle. We included all of these caveats within the paper (Birn-Jeffery et al. 2012): they might be missed since they’re buried in various sections of text so are worth emphasising.

* For a fuller discussion of many of the anatomical considerations that need to be taken into account here, see Dececchi & Larsson (2011). They mostly concluded that birds and other theropods did not possess the anatomical features expected for a climbing lifestyle. However, they relied heavily on comparison with mammals and we are concerned that this may have confused their analysis.

Conclusions, categories, confidence and contradictions

Anyway, what did we find? The inclusion of numerous lizard taxa seemingly did nothing to alter the robustness of the conclusions for birds, allowing us “to be more confident in asserting that trends in claw morphology occur across tetrapods” (Birn-Jeffery et al. 2012, p. 7). In other words, our analysis might help us be more confident in linking claw morphology with behaviour and lifestyle in fossil species where behaviour and lifestyle can’t be observed.

What about the relationship between body mass and claw morphology, and between phylogeny and claw morphology? Because we analysed inner and outer claw curvature as well as claw mid-length height, we recovered numerous different correlations. Firstly, there is a correlation between inner and outer claw curvature and claw mid-length height, which is good. Secondly, the claws of different digits differed significantly in morphology – digit III was always least curved, for example. This at least raises the possibility that different behavioural ‘signals’ might be obtained if different claws from the same taxon (or same individual) were analysed and emphasises the need for standardisation across studies.

Thirdly, both inner and outer claw curvature, and claw mid-length height, are correlated with body mass: claws are both less curved, and deeper, in heavier animals than in lighter ones, though the strength of these relationships was extremely low. These relationships were previously discussed by Pike & Maitland (2004). Fourthly, if we account for the impact of phylogeny – if we use a statistical method that removes the effects of relatedness – we find little relationship (or poor relationships) between claw curvature and behaviour (Birn-Jeffery et al. 2012). This result contradicts the general relationship we recovered between claw curvature and behaviour, and it’s difficult to know how to interpret it.

Most living taxa plotted where predicted, but there were a few surprises. The Satin bowerbird Ptilonorhynchus violaceus [adjacent photo by Tatiana Gerus] is part of the ‘ground-dweller’ group according to behavioural data, but it’s a consistent outlier in inner claw curvature, better plotting in the region expected for animals in the 'climber' category. I wonder if this is because the literature we used was misleading with respect to how much of a ‘ground-dweller’ the Satin bowerbird really is; then again, it does seem to spend more time on the ground than other bowerbirds.

Among the 'climber' group, the Frill-necked lizard Chlamydosaurus kingii had lower outer curvature than expected. Again, this might be because our characterisation of it as a ‘climber’ is too much of a generalisation: Frill-necked lizards walk and run on the ground a fair bit (their remarkable bipedal terrestrial behaviour is well known); they certainly aren’t restricted to a life of climbing. Several bird species we analysed (including stilts, bustards, shore larks and rheas) had especially low outer curvature measurements compared to other members of the ‘ground-dweller’ category.

Climbing, perching and ground-running Mesozoic birds and other theropods

What about fossil taxa? As we saw at the start of this article, much of the interest in claw curvature has been driven by a desire to better understand the lifestyle and biology of Archaeopteryx. While work on lizard claw curvature comes from a desire to better understand lizard anatomy and behaviour, work on claw curvature in modern birds has been instigated by palaeontological interest: yet another case where people interested in fossils – as opposed to people who work on modern animals – were the ones who had to get the ball rolling. Insert comment about lazy biologists and their obsession with genetics (I kid, I kid).

It isn’t just the lifestyle of Archaeopteryx that’s been controversial. A general debate about the lifestyles of all archaic Mesozoic birds has encouraged the study of their claws and other anatomical details, and a number of authors have suggested that dromaeosaurids and other non-bird coelurosaurian theropods might have been climbers too. Study of claw curvature and claw function in dromaeosaurids has also been driven by a desire to better understand the function of the remarkable giant, hyperextendable pedal digit II present in these animals (Manning et al. 2006, Fowler et al. 2011); in turn, this has resulted in some excellent research on claw anatomy and function in raptors and owls (Fowler et al. 2009).

My personal take on these studies is that they indicate a probable prey-immobilisation role in dromaeosaurid (and other deinonychosaur) digit II claws, but they also allow for the possibility that the claws had a role in occasional climbing too. The idea that dromaeosaurids clambered up the bodies of giant prey animals – a scenario promoted by Manning et al. (2006) – still seems absurd to me.

I have a special interest in suggestions that non-bird theropods might have climbed trees and have published two articles on the history of this idea (Naish 2000a, b). Note that these are reviews of what people have said across history, not functional analyses. I’m wary of this part of the article turning into an opinion-fest, but I have to say that I think that the whole idea of climbing abilities in Archaeopteryx, in other Mesozoic birds, and in non-bird theropods has been afflicted by an undue amount of polarisation. On the one hand, some people want Mesozoic birds to be tree-dwellers because they argue that bird flight, and birds themselves, must have arisen in an arboreal setting. These are often (but not always) the same people who argue that birds cannot be dinosaurs. On the other hand, some people insist that Mesozoic birds and bird-like theropods were ground-dwelling animals without any climbing ability at all, and these are often the people who argue that bird flight and birds themselves evolved within a terrestrial, cursorial context.

Based on what I know about living animals, I find it hard to look at Archaeopteryx, at an enantiornithine, or even at Deinonychus and Velociraptor without coming away with the idea that these animals very probably could climb if they wanted to (this contention is based on body size, limb proportions, forelimb orientation, and hand and foot anatomy). That doesn’t mean that they didn’t walk, run or forage on the ground for much or most of their time, in fact animals like Velociraptor- and Deinonychus-sized dromaeosaurids were obviously predominantly terrestrial. But if they needed to climb a tree – if it was a good idea in avoiding predators, or when finding food or shelter – I’m confident that they could do it. When it comes to Microraptor and some other small deinonychosaurs, their curved claws, insane hindlimb plumage and small body size all render it likely that climbing was a frequent activity. The inference that there’s an incompatability between the dinosaurian origin of birds and a ‘trees-down’ origin for flight is flatly incorrect, though I’m not necessarily saying that maniraptoran flight originated in a strictly arboreal context.

With all of this in mind, we included claw data from various Mesozoic birds and other theropods in our analysis. Most non-avialan maniraptorans plotted within the space shared by ground-dwellers, climbers, perchers and predators, though inner curvature and outer curvature measurements sometimes gave different results. Theropods that I would predict to have little to no climbing ability – ornithomimids, compsognathids and Caudipteryx, for example – grouped in the ‘ground-dweller’ space (Birn-Jeffery et al. 2012). Velociraptor did too, perhaps showing that those ideas about hypothetical climbing abilities are sometimes not supported once you start looking at the details (incidentally, Parsons & Parsons (2009) argued that pedal claw shape in Deinonychus resembles that of climbing modern birds while pedal claw shape in Velociraptor does not). Based on outer curvature, certain other dromaeosaurids (including Deinonychus and Microraptor) and Archaeopteryx bavarica fell into the ‘climber’ cluster, as did some unambiguous Mesozoic birds like Changchengornis (Archaeopteryx is an “ambiguous bird”, since it doesn’t fall on the same branch as modern birds in all phylogenies) (Birn-Jeffery et al. 2012).

Anchiornis – an elaborately feathered Jurassic maniraptoran suggested by some authors to be a troodontid (Hu et al. 2009) – plotted in different places according to the sort of data analysed, and two different specimens plotted in different places anyway. When inner curvature was measured, and when claw mid-length depth was accounted for, one specimen plotted within the ‘ground-dweller’ region, but the other one plotted within the shared ‘climber’-‘predatory’-‘percher’-‘ground-dweller’ space. The Cretaceous troodontid Sinornithoides did likewise (Birn-Jeffery et al. 2012). Finding Anchiornis to fall within an exclusive ‘ground-dweller’ region of the graph might seem in agreement with comments that Anchiornis had a “strong cursorial capability” (Hu et al. 2009, p. 462); however, the long feathers growing off the metatarsus and rest of the hindlimb are problematic for a terrestrial lifestyle.

Intriguingly, the different Archaeopteryx species plotted in different places. A. bavarica was in the area shared by members of the ‘climber’, ‘predatory’, ‘percher’ and ‘ground-dweller’ groups while A. lithographica was in the ‘ground-dweller’ space (Birn-Jeffery et al. 2012). This could highlight the unreliability of claw curvature for inferring lifestyle. Or it could show that different archaeopterygid taxa were ecologically and behaviourally very different - a wholly reasonable possibility.

These results mostly conform to what I said above. Theropods that appear to be truly cursorial grouped as ‘ground-dwellers’, while those that look like ‘ability generalists’ (like Deinonychus) plot in the overlapping space shared by members of all behavioural groups according to some claw data, but in the ‘climber’ category according to other data (Birn-Jeffery et al. 2012).

We didn’t find any strong or convincing evidence for arboreality in any of the Mesozoic taxa we analysed (Birn-Jeffery et al. 2012), a discovery which is in keeping with the fact that none of them exhibit morphological features suggestive of such a lifestyle. There are no woodpecker or treecreeper analogues among Mesozoic dinosaurs, for example... at least, not yet.

All in all, I would say that our study seems to confirm via statistics what we already suspected: that claw curvature provides only an approximate view of lifestyle, that the behavioural categories used in claw curvature studies are overlapping, overly simplistic and perhaps inaccurate, that claw curvature doesn’t provide a single, simple guide to behaviour but needs to be considered alongside other lines of anatomical information, and that the results of claw curvature analysis are affected by body mass, phylogeny and even on how claw curvature itself is measured.

As always, there’s much more that’s worthy of discussion in the paper itself: it’s open access, so you can obtain it for free should you want to.

For previous Tet Zoo articles relevant to some of the themes discussed here, see...

Refs - -

Birn-Jeffery, A. V., Miller, C. E., Naish, D., Rayfield, E. J., Hone, D. W. E. 2012. Pedal claw curvature in birds, lizards and Mesozoic dinosaurs – complicated categories and compensating for mass-specific and phylogenetic control. PLoS ONE 7(12): e50555. doi:10.1371/journal.pone.0050555

Chiappe, L. M. 1997. Climbing Archaeopteryx? A response to Yalden. Archaeopteryx 15, 109-112.

Dececchi, T. A. & Larsson, H.C. E. 2011. Assessing arboreal adaptations of bird antecedents: testing the ecological setting of the origin of the avian flight stroke. PLoS ONE 6(8): e22292. e22292. doi:10.1371/journal.pone.0022292

Feduccia A. 1993. Evidence from claw geometry indicating arboreal habits of Archaeopteryx. Science 259, 790-793.

Fowler, D. W., Freedman, E. A. & Scannella, J. B. 2009. Predatory functional morphology in raptors: interdigital variation in talon size is related to prey restraint and immobilisation technique. PLoS ONE 4(11): e7999. doi:10.1371/journal.pone.0007999

- ., Freedman, E. A., Scannella, J. B. & Kambic, R. E. 2011. The predatory ecology of Deinonychus and the origin of flapping in birds. PLoS ONE 6(12): e28964. doi:10.1371/journal.pone.0028964

Glen, C. L. & Bennett, M. B. 2007. Foraging modes of Mesozoic birds and non-avian theropods. Current Biology 17, 911-912.

Hone, D. W. E., Sullivan, C. & Bennett, S. C. 2009. Interpreting the autopodia of tetrapods:interphalangeal lines hinge on too many assumptions. Historical Biology 21, 67-77.

Hu, D., Hou, L., Zhang, L. & Xu, X. 2009. A pre-Archaeopteryx troodontid theropod from China with long feathers on the metatarsus. Nature 461, 640-643.

Manning, P. L., Margetts, L., Johnson, M. R., Withers, P. J., Sellers, W. I., Falkingham, P. L., Mummery, P. M., Barrett, P. M., Raymont, D. R. 2009. Biomechanics of dromaeosaurid dinosaur claws: application of x-ray microtomography, nanoindentation, and finite element analysis. The Anatomical Record: Advances in Integrative Anatomy and Evolutionary Biology 292, 1397-1405.

- ., Payne, D., Pennicott, J., Barrett, P. M., & Ennos, R. A. 2006. Dinosaur killer claws or climbing crampons? Biology Letters 2, 110-112.

Morell, V. 1993. Archaeopteryx: early bird catches a can of worms. Science 259, 764-765.

Naish, D. 2000a. 130 years of tree-climbing dinosaurs: Archaeopteryx, ‘arbrosaurs’ and the origin of avian flight. The Quarterly Journal of the Dinosaur Society 4 (1), 20-23.

- . 2000b. Theropod dinosaurs in the trees: a historical review of arboreal habits amongst nonavian theropods. Archaeopteryx 18, 35-41.

Ostrom, J. H. 1974. Archaeopteryx and the origin of flight. Quarterly Review of Biology 49, 27-47.

Parsons, W. L. & Parsons, K. M. 2009. Further descriptions of the osteology of Deinonychus antirrhopus (Saurischia, Theropoda). Bulletin of the Buffalo Society of Natural Sciences 38, 43

Peters, S. F. & Görgner, E. 1992. A comparative study on the claws of Archaeopteryx. In Campbell, K. E. (ed.) Papers in avian palaeontology Honoring Pierce Brodkorb. Los Angeles: Natural History Museum of Los Angeles County 36, 29-37.

Pike, A. V. L. & Maitland, D. P. 2004. Scaling of bird claws. Journal of Zoology 262, 73-81.

Spielmann, J. A., Heckert, A. B. & Spencer, G. L. 2005. The late Triassic archosauromorph Trilophosaurus as an arboreal climber. Rivista Italiana di Paleontologia e Stratigrafia 111, 395-412.

Tulli, M. J., Cruz, F. B., Herrel, A., Vanhooydonck, B. & Abdala, V. 2009. The interplay between claw morphology and microhabitat use in neotropical iguanian lizards. Zoology 112, 379-392.

Xu, X., Zhou, Z. H., & Wang, X.-L. 2000. The smallest known non-avian theropod dinosaur. Nature 408, 705-708.

Yalden, D. W. 1985. Forelimb function in Archaeopteryx. In Hecht, M. K., Ostrom, J. H., Viohl, G. & Wellnhofer, P. (eds) The Beginnings of Birds – Proceedings of the International Archaeopteryx Conference, Eichstätt 1984, pp. 91-97.

Zani, P. A. 2000. The comparative evolution of lizard claw and toe morphology and clinging performance. Journal of Evolutionary Biology 13, 316-325.