Friday, February 6, 2009

Titanoboa and paleophidiothermometry

Yes that's right a new word to describe the measuring of ancient temperatures using fossil snakes. By now most of you will have heard of the Jason Head and colleagues' paper describing Titanoboa, the largest known snake ever. For those that might not have seen it, Titanoboa was a boa that lived 58-60 million years ago in Colombia. Many fossils (mostly vertebrae) deriving from multiple individuals have been found and the largest of these came from a snake close to 13 metres in length and probably weighing in at 1.2 tons. Fossils from eight individuals were found in this size category indicating that we aren't dealing with a single freakily large individual but a species that regularly attained this gigantic size.
What I'm going to discuss however is the second part of the paper, where the authors attempt to estimate the mean annual temperature of Titanoboa's habitat by using maximum snake size. The idea rests on the observation that the size of snakes, as animals that require external sources for their body heat (ectotherms), and cannot maintain a constant internal temperature (poikilotherms), is constrained by the temperature of their environment. Makarieva and colleagues (2005a, b)observed that the largest species of many terrestrial clades of pokilotherms occur in the tropics, while the largest in the temperate realms were smaller and the largest from the polar regions were smaller still. They argued that there is a minimum mass specific metabolic rate below which an organism simply can't function. As mass specific metabolic rate decreases with increasing size there is a maximum size over which the mass-specific metabolic rate is too low. Increasing ambient temperature allows the metabolic rate to be increased and consequently a larger size can be attained.
Head et al. then used the mathematical relationship between temperature change and the change maximum attainable body size to calculate the mean annual palaeotemperature which Titanoboa experienced. Using modern green anacondas as a model for a modern snake that is probably at its maximum acheivable size (which was taken as 7.5 metres) and the temperature data for their habitat, Head et al. calculated that Titanoboa lived in a sweltering tropical forest that averaged somewhere between 31 and 32 degrees Celcius throughout the year (compare that to the 26-27 degrees averaged by modern lowland equatorial forests that anacondas live in).
It is an interesting idea but I see all sorts of problems. Firstly the size of the fossil snake may not be accurately estimated (actually this is the least likely area of inaccuracy - Head et al.'s methods seem pretty solid in this regard). My main problems are that Makariev's formula has not yet been rigorously tested (I'm not even sure HOW you would go about testing it) and most of all I don't know how we can demostrate that modern green anacondas are up against the theoretical maximal size limit for an ambient temperature of 26-27 degrees. The mere existance of exceptional individuals of anacondas and reticulated pythons reaching length close to or even slightly over 10 metres indicates that it is possible for snakes to exceed 7.5 m in the modern tropics without their metabolic rate dropping to fatally low levels. I suspect other ecological pressures (e.g. prey size, predation, availability of suitable cryptic resting places or something else entirely) may be keeping anacondas from reaching their theoretical maximum body size.

Graph from Head et al. 2009 that shows the curve of theoretical maximum size of a boiine snake against mean annual ambient temperature and the predicted temperature derived from the size of Titanoboa.

Despite all these objections, the results are remarkable in that they appear to work. That is a MAT of 32 degrees is entirely plausible, indeed expected for the tropics of the mid Paleocene Epoch. It is well known that CO2 levels were high in the Paleocene, and the world was in a greenhouse phase. During such phases there were no icecaps and we can find fossil evidence for abundant vegetation and animal life at the poles. It used to be thought that the temperature gradients between the equator and the poles were flatter than in modern times and that although the poles were nice and warm the equatorial regions were no hotter than they are now. We now know we were wrong - the tropics were fiercely hot during such times and this fits perfectly with Head's paleophidiothermometry.


Jason J. Head, Jonathan I. Bloch, Alexander K. Hastings, Jason R. Bourque, Edwin A. Cadena, Fabiany A. Herrera, P. David Polly, Carlos A. Jaramillo (2009). Giant boid snake from the Palaeocene neotropics reveals hotter past equatorial temperatures Nature, 457 (7230), 715-717 DOI: 10.1038/nature07671

Makarieva, A. M., Gorshkov, V. G.&Li, B.-L. (2005)Gigantism, temperature and metabolic rate in terrestrial poikilotherms. Proc. R. Soc. Lond. B 272, 2325–2328.

Makarieva, A. M., Gorshkov, V. G. & Li, B.-L. (2005b)Temperature-associated upper limits to body size in terrestrial poikilotherms. Oikos 111, 425–436.


Darren Naish said...

Minor nitpickesque: in final para, make sure you change 'now icecaps' to 'no icecaps'. And then delete this comment.

Zachary Miller said...

I think that, ultimately, the point is that it's a big fricking snake.

Anonymous said...

What Zachary said!!

Anonymous said...

"My main problems are that Makarieva's formula has not yet been rigorously tested (I'm not even sure HOW you would go about testing it)"

An obvious way to test the formulae is to apply them to yet more geographic areas. So far Makarieva et al. have based their arguments on the data on the largest species from four regions: Wrangel Island (very cold, far in the North), Great Britain and Colorado State (warm enough and still a bit warmer, respectively), and the equatorial region (the warmest, world's largest species). Apparently, size spectrum data on species assemblages exist for many geographic regions. Taking the formulae predictions and comparing them with the observed relationships between maximum body sizes and relevant ambient temperature for a number of geographic regions will provide a transparent test for the theory.

Anonymous said...

What would be really helpful is to have another proxy for temperature to compare with this estimate. Oh wait, there is one - the Cerrejon leaf macroflora. I don't know how they relate stratigraphically, but both the snake and the leaves are from the same coal mine. The MAT estimate using leaf margin analysis is signficantly cooler, around 23.8degC +/- 2 degrees. I think that work was done by Fabiany Herrera with Scott Wing, though I'm not sure if it's been published in any form besides abstracts and informal reports (all floating around on the web). Anyway, it'd be interesting to see why these two estimates are so disparate.

Anonymous said...

"Anyway, it'd be interesting to see why these two estimates are so disparate."

Head et al. have something to say on that in their paper:
"These estimates exceed MATs derived from coeval Cerrejon palaeofloras by 6–8 C 20, but palaeotemperatures based on fossil leaf assemblages from riparian and wetland habitats of rain-forests are underestimates 21."
Ref 20 is the work of Herrera et al. 2005.

Myself I am very far from thinking the snake-based temperature reconstruction is the source of ultimate truth. There are many caveats in such body size analyses, e.g. using length vs width of the animal and in particular considering air-breathing vs aquatic species.