Big cat: small cat

So it’s been a while since my last post, but this one comes as an exciting advancement in my scientific career (and a day late for my birthday). My first postdoc paper was accepted and has now been published:

Cuff et al., 2015.Big cat, small cat: Reconstructing body size evolution in living and extinct Felidae. Journal of Evolutionary Biology. doi: 10.1111/jeb.12671.

This post isn’t to say how amazing the work is, more a way of me distilling and simplifying the information so that my family, non-scientific (or at least phylogenetics based) friends and interested others may be able to understand what I have published on (see last post). If you are interested in a copy and do not have access to it online (we unfortunately could not justify the £2000 for open access), and can't wait a year, please do email me and I can get you a copy.

The postdoctorate I am working on is part of a larger project trying to understand how all living cat (felid) species vary particularly with respect to their muscles, bones and scaling with body size. In modern species this size range is from 1ish kilos in the black-footed cat and rusty spotted cat, to 3-4kgs in domestic cats, to the largest male lions and tigers pushing 300kgs.

Body mass ranges of living felids.

If we look back in time there were even bigger cat species, with some of the sabre toothed cats (belonging to the Machiarodontidae) and largest cave lions pushing 4-500kgs. Despite work being done on other groups’ evolutionary history (e.g. dogs – Valkenburgh et al., 2004) no-one had yet looked at it in felids, and this is where this paper comes in.

Body mass range of living and extinct felids.

So the first step in trying to understand the evolution of body mass in the felids, is getting a family tree (phylogeny) of all living and extinct species. There are some great phylogenies of modern taxa (e.g. Johnson et al., 2006), but the problem with these trees based on genetic material is that very few contain fossil taxa (there are some exceptions including cave lions and the American lion, for which some DNA has been preserved. The challenge then became tracking down an extensive phylogeny for both modern and extinct taxa. The best available at present is that from Piras et al., (2013) which has a very thorough sampling of modern and fossil taxa. From this phylogeny it should be stated here that we used a variety of permutations that affect particularly the fossil ages: first occurrence (when the first fossil appeared, or at least the oldest estimate for the fossil is), mid (midpoint between first and last), last occurrence (when the species died out or oldest estimate for a fossil), as well as looking at only the modern clade of felids (both including and excluding fossil taxa).I will caveat here that a few modern species have moved relations compared to the genetic information (particularly those of the Panthera genus – lions, tigers leopards etc.). There are also some newer fossils belonging to the Panthera genus that were not included (e.g. P. blythaea: Tseng et al., 2014). However, I am hopeful that even when a new, bigger Felidae phylogeny is made, the results will hold true. We’ve also included all of our materials and methods in the supplementary information so it should be easy enough to replicate.

The next step was finding a database of felid body masses. For most of the living taxa there is a lot of data known on the body masses (or at least a range for male and female). These were used to calculate an average for each species (nearly all of my data came from a coauthor’s previous paper – Randau et al., 2013). For the remaining species where the data wasn’t readily available, estimates for body mass were taken from their describing papers, or from an average calculated from skull length (condylobasalar length – from snout to vertebral attachment) using an equation calculated from living species.

Now we have the data for family tree, and for each of their masses. The next step was to remove all of the species from the tree for which we didn’t have body masses. When this tree pruning was done, the next step was to assess the amount of phylogenetic signal in the data - the amount the shape of the tree, and the position of the species on the tree affect the data. In simplest terms, you’d expect the most closely related species to have masses more similar to each other than species that are less closely related. In our data it turns out there is a lot of phylogenetic signal allowing us to carry out the next tests, testing mode of evolution that family was undergoing. When I say mode of evolution, I really mean the way body mass evolves. Initially we tested for Brownian motion, white, trend, OU and early burst.

Brownian motion is a random walk pattern. Imagine flipping a coin, heads you increase in body mass, tails you decrease. Over time you could have all heads, all tails, but more likely a relatively even mix of both the longer the length of time studied. A white model has no change at all through the tree. A trend model is best described as a Brownian motion pattern where there is a directional pattern (e.g. selection that meant only heads were flipped if going back to our coin analogy). There are some famous models e.g. Cope’s “rule” which suggests there is an increase in body mass through lineages in time (not going to discuss the joys of Cope’s rule here as that would be as long as this post is too). OU (Orstein-Uhlenbeck) models are similar to trend models initially, so there is a selection pressure encouraging animals to evolve in a particular direction (e.g. all heads), however once they reach an optimal position they stay there (i.e. there is stabilising pressure so that masses stop increasing or decreasing from the optimum). This is often best described in an adaptive landscape (I am changing analogies here), where fitness of an animal is described as a hill (or island depending on preference), if you are at the bottom, you want to get to the top where you are more optimally adapted for the environment. But once at the top (or above sea level), it’s disadvantageous for the species to leave this hilltop/island, so they stay there. Early burst is the final model, where there is a rapid evolutionary pulse near the origin of the group where all major morphospaces (hills/islands) are occupied, with then some further expanding (into the small islands) of the range across the rest of the group’s history. The Cambrian explosion often is cited as a good example of this.

The test for which model all of these is best is called the Akaikes information criterion (AIC). A more recent version corrects for finite sample sizes (as we do not have infinite numbers of samples) and is perhaps understandably known as the corrected Akaikes information criterion (AICc). This method compares the probability that a model fits the data and then gives a likelihood of any model being best (normally displayed as a percentage as in our results). From this there was the suggestion that an OU model best described the data for the first occurrence phylogeny, and Brownian models best explained the mid-, last and modern occurrences. However, with the AICc we could only test single OU optimum models, and this is where SURFACE and bayou come in. Both of these packages are plug-ins for R (which is rapidly becoming the go-to stats program online) independently developed and tested. Both of these packages allow for testing of multiple OU optima (e.g. a big size and small size) and whether there is convergence between them.

Using these programs, SURFACE recovered 2 optima for modern felids, with the Panthera lineages and Puma evolving to convergent large body masses, and the rest of the felids staying at smaller sizes. bayou did not recover any pattern different to that of Brownian motion. The first occurrence data was probably the most entertaining as far as things I’ve ever written into results with SURFACE finding a range of optima, including two ridiculous ones: a large body mass (near the size of Juipiter); and a small body mass (close to carbon atom size). These are obviously not real optima, although they are entertaining to consider, and the crazy scale is most likely associated with: 1) the optima being evolved towards have not been reached; 2) the strength of the selection across the tree (i.e. how quickly things walk or run up their hills) varies across the tree. Because bayou runs many simulations (I ran 1,000,000 per model) multiple selection strengths could be tested, and the results found again two optima, a small one and a larger one. The mean and last occurrence data, both found two convergent optima supporting a large and small body masses in SURFACE, but this is only also recovered for the last occurrence data in bayou.

From Cuff et al., 2015. Phylogeny of all extant and extinct felid taxa using last occurrence dates (modified from Piras et al., 2013) showing the results from ‘SURFACE’ and ‘bayou’. (a) ‘SURFACE’ and ‘bayou’ phylogenies with shifts shown. ‘SURFACE’ shifts shown on the branches (red and blue), whereas ‘bayou’ rates are shown on the nodes with the colours representing increases and decreases, and the size of the circles showing the probability. (b) Phenogram showing distribution of taxa body masses against their phylogeny for posterior probabilities >0.2 (Table S4). Convergence shows the puma/cheetah lineage mostly being in the large body mass optima, whereas the clouded leopard species converge into the small body mass optima.

What does this all mean? Well there is some data for Smilodon from the La Brea tar pits suggesting they do attain larger body masses through evolutionary time. So despite using average masses (which would hide this signal), there is reason to believe that the last occurrence results are most realistic and best match what we see in the modern world. If this is the case, felids evolve two body mass optima, with large body forms and small body forms. The exact value for these optima varies depending on the method used, but generally they are divided somewhere around 5kg and >25kg ranges. The upper body mass limit fits with previous biomechanical and ecological data showing that large felids (>25kgs) have to take prey as large or larger than themselves in general to maintain their energy levels, whilst smaller species tend to take small prey. From this it may also be able to extend our understanding to some of the extinct species and what their ecologies were. Our results differed from what has been found in canids (dogs, foxes, wolves etc.) where there seems to be a trend towards continued larger body sizes (i.e. Cope’s rule), except in the foxes which show smaller sizes (Van Valkenburgh et al., 2004; Finarelli, 2007). It should still be mentioned that despite canids evolving increases in body size, the largest (at 70kgs in wolves), do not match even the largest living felids, let alone the incredible size (500kg) found in some of the extinct species.

References

Cuff et al., 2015.Big cat, small cat: Reconstructing body size evolution in living and extinct Felidae. Journal of Evolutionary Biology. doi: 10.1111/jeb.12671.

Finarelli, J.A. & Goswami, A., 2013. Potential pitfalls of reconstructing deep time evolutionary history with only extant data, a case study using the canidae (Mammalia, Carnivora). Evolution 67, 3678-3685. doi:10.1111/evo.12222

Johnson, W.E., Eizirik, E., Pecon-Slatter, J., Murphy, W.J., Antunes, A., Teeling E., et al., 2006. The late Miocene radiation of modern Felidae: a genetic assessment. Science 311:73-77

Piras, P., Maiorino, L., Teresi, L., Meloro, C., Lucci, F., Kotsakis, T., et al., 2013. Bite of the cats: Relationships between functional integration and mechanical performance as revealed by mandible geometry. Syst. Biol. 62: 879-900

Randau, M., Carbone, C. & Turvey, S.T. 2013. Canine evolution in sabretoothed carnivores: natural selection or sexual selection? PLoS ONE 8: e72868

Tseng, Z.J., Wang, X., Slater, G.J., Takeuchi, G.T., Li, Q. & Liu, J. et al. 2014. Himalayan fossils of the oldest known pantherine establish ancient origin of big cats. P Roy Soc B-Biol Sci 281 (1774): 20132686.

Van Valkenburgh, B., Wang, X. & Damuth, J. 2004. Cope’s Rule, hypercarnivory and extinction in North American canids. Science 306: 101-104.