I wanted to spend a post talking about a new paper that was published recently (3 May 2013) with some colleagues from Carleton University.  It is easy to see the value of tasting bad: predators try to eat you, feel sick, then leave you alone.  Even better if you have bright colours or a strong smell (called “aposematic signals”) to go along with it – that way predators can learn to avoid your colours without having to taste you a second time.  In fact, they don’t have to taste you at all if other animals of your species also have the bad taste and the bright colours.  In theory, this chemical defence should reduce deaths due to predation which means that the prey live longer.

Now, a longer lifespan represents an opportunity for species to adapt.  Ordinarily, species are adapted to use the early part of the lifespan as efficiently as possible because lives tend to be short and you cannot pass on your genes to the next generation (the evolutionary currency) if you die before getting a chance to reproduce.  However, if you are not suffering much mortality then you can start to alter your strategy to do things when you are older, rather than having to frantically cram all of that feeding, growth and reproduction into a short period.  The result is the prediction that chemically defended species will live longer (exhibit lower levels of ageing).  Indeed, this is what has been found.  In 2005, Mario Blanco (Harvard) and Paul Sherman (Cornell) published a paper looking at almost 1200 species of fish, snakes and amphibians, classifying each species as “defended” or “undefended”.  They found that (controlling for size) chemical defence was often associated with longer maximum lifespan.

The problem with the Blanco and Sherman paper (and one that the authors acknowledged) was that they did not take into account the evolutionary relationships of the species involved when conducting their analysis.  I have mentioned this issue of “autocorrelation” in the context of spatial variation in human IQ in a previous post, but to recap: statistical tests require independent data.  Looking at multiple species (or multiple countries), some data are going to be more similar than others due to their proximity (in an evolutionary or geographical sense).  Think about comparing the heights of people from England, Wales and Japan – you would expect the English and Welsh to be more similar.  The same is true of evolutionary (or “phylogenetic”) autocorrelation – compare humans, chimpanzees and millipedes, and you would expect the two apes to be more similar due to having a greater proportion of shared evolutionary history.  This introduces a lack of independence that must be accounted for in statistical analyses, but it is missing from the Blanco and Sherman paper.  When we looked to see whether it was a problem, we noticed something worrying (click to embiggen):

For the snakes, at least, all of the chemically defended (i.e. venomous) species are clustered within one group (note the semicircle of black squares in the phylogenetic tree above).  The situation isn’t so bad in amphibians, but there are still clear clusters of related, defended species (see the right hand tree above).  Blanco and Sherman didn’t take this into account when they found that defended species lived longer.

In our paper that was recently published online in the Journal of Evolutionary Biology (also check out the lead-author’s blog on caterpillar eyespots), we checked to see whether this was an issue.  When we controlled for this phylogenetic autocorrelation, we find that amphibians that have a chemical defence live longer but this pattern does not hold for snakes.  We discuss this finding at some length in the paper, but I think it boils down to the fact that snake venom evolved for prey capture, rather than defence against predators.  This means that the effect on mortality from predators is lower, and this does not produce the same strength of selection for longer lifespan.   While snakes do bite in self defence (as many humans find out to their cost), these bites vary in the amount of venom that is injected.

The moral of the story, then, is that the recent advances in phylogenetics and systematics allow us to test existing hypotheses and theories in a new, more comprehensive (and conservative) way.  Our findings are consistent with the following: that amphibians use their chemical weaponry primarily for defence, this increases their survival, and leads to selection for greater lifespan.  The fact that snakes do not show an association between venom and longevity is more evidence for the evolution of snake venom being driven by a role in prey capture.  If you would like to ask any questions about the study then I would be happy to discuss them in the comments!

Photo credits: Dendrobatid frog photo by H. Krisp, viper photo by Pepp Cristiano, both via Wikimedia Commons (click links to visit pages).  Evolutionary trees are mine, redrawn from the paper.