To help or not to help?

The theory of kin selection — the notion that cooperative behavior in animals evolved because cooperation to enhance the survival of your kin also enhances the survival of your own genes — has been a dominant and widely-discussed idea for decades. Though it has recently been challenged, the response from the community has been vigorous.  Personally I plan to sit this one out.  Instead, let’s look at a fascinating new study Kevin Foster highlighted for me, which pulls together data from 267 species of birds (Cornwallis et al. 2010. Promiscuity and the evolutionary transition to complex societies, Nature 466 969-974 PMID: 20725039) to ask whether promiscuity is a major factor in the evolution of cooperative breeding.

The Pied Babbler, a monogamous and cooperative species. Picture by Sarah Knowles. Kindly sent to me by Charlie Cornwallis.

The background hypothesis here is that the decision of a female to mate either with one or with many partners is a major, though not the only, factor in determining how likely it is for family members to choose to cooperate in the arduous process of bringing up the kids.  Imagine you’re a bird in a species where females mate with only one male.  Half of your genes came from your mother, and half from your father; the same is true for all your siblings.  On average, the number of genes you share with a sibling is 1/2.  If you go off and mate with an unrelated individual, the proportion of genes you will share with your offspring is also 1/2.  So you might — depending on circumstances — choose to spare yourself the effort of nestbuilding and finding a partner you’re willing to spend the rest of your life with, and instead focus on helping your parents raise your younger siblings.  It will all come to the same thing in the end.  Species that cooperate in this way are called cooperative breeders.  The Pied Babbler, which is monogamous and a cooperative breeder, offers an example of the rewards of virtue, whereas the Cape Sugarbird is an example of the wages of sin — more than 70% of its broods contain chicks fathered by more than one male, and the females are left to raise their offspring on their own.

The Cape Sugarbird, which is both promiscuous and a solitary breeder. Photo by Mhairi McFarlane. Kindly sent to me by Charlie Cornwallis.

This is a nice simple theory with easy-to-draw moral implications, but the problem is that (as usual) real life doesn’t quite fit it.  Whatever stories you’ve heard about eternally faithful swans and wolves, real monogamy is relatively rare in vertebrates, and cooperative breeding is not restricted to those few monogamous species.  Cornwallis et al. point out that two of the highest promiscuity rates ever recorded are found in cooperative breeders: one is a bird that has the lovely romantic name “superb fairy-wren“, and the other is the Australian magpie [which sounds like a much more likely species to be promiscuous, don’t you think?].  The number of broods with the “wrong” father is around 70% for both species.  And then there is the puffin, which is not promiscuous at all but isn’t cooperative either. There does seem to be a correlation between lower levels of promiscuity and cooperation, but counterexamples to the simple story abound.  What’s going on?

The heart of this paper is a large phylogenetic tree relating the 267 bird species for which the authors were able to find data in the literature on both cooperative breeding and promiscuity.  (Actually what they did to produce this tree was to create an enormous supertree relating 6,219 species by combining 966 bird trees published by others, then prune the tree down so that only the species for which they had relevant data remained.  The publication of the supertree is promised for another day).  Using the data on cooperative breeding and promiscuity from modern-day species, and in particular the correlation between promiscuity levels in species that are descended from a common ancestor, they developed an estimate for the promiscuity level of the ancestor and a prediction of whether it was cooperative or not.  This may sound weird, but it is exactly the kind of approach used in quantitative genetics to trace back the evolution of traits through pedigrees; this analysis is treating cooperative breeding and promiscuity as heritable, not necessarily related, traits.  From the phylogenetic tree, they can also see that cooperation evolved (at least) 33 times and has been lost (at least) 20 times.  So now they can ask, are there any correlations between promiscuity levels and the gain or loss of cooperative breeding?

One way of looking at these data is to ask how promiscuous were ancestors that gave rise only to non-cooperative descendents, versus ancestors that gave rise only to cooperative descendents.  Whether the ancestor itself is predicted to have been cooperative or not, more promiscuity tends to produce more non-cooperative descendent species, and less promiscuity tends to produce a larger number of cooperative descendents. Transitions to and from cooperative behavior are also associated with a predicted change in the promiscuity of the species where the transition occurred.

This evidence is consistent with the kin selection theory, but it still doesn’t explain the highly promiscuous cooperative species, or the highly monogamous non-cooperative species.  But is cooperation really that black and white? For some birds, for example the white-winged chough, living in the harsh environment of the Australian bush, cooperation is essential: chicks don’t survive if the parents are left to go it alone.  For other birds, cooperation is not so crucial.  Sometimes the older brothers and sisters help out, other times they don’t; it probably depends on the overall environment the nest is in.  So instead of thinking of cooperative breeding as all-or-none, we need to view it as a quantitative phenomenon.

One way you can vary your level of cooperation is to help only when there’s a good chance that the chicks in the nest are highly related to you; this is called kin discrimination.  Birds seem to be able to determine which broods are related to them and which are not, using cues such as the way the chicks sound (although there’s no evidence they can identify the single chick in a brood that happens to have a different father).  The kin selection theory would suggest that kin discrimination would be important if the likelihood that you’re related to a particular brood is very variable.  There isn’t much variability at the two ends of the spectrum — maximum promiscuity (all unrelated) or maximum fidelity (all related) — but in the middle, at an intermediate level of promiscuity, you would expect it to be advantageous to evolve a way to differentiate between your full-blood brothers and sisters and your half-siblings.  And indeed, a plot of the ability to identify kin versus the level of promiscuity in different species shows a nice bell-shaped curve.

Another way of varying cooperativity would be to increase or reduce the proportion of nests that actually get a helper.  And as promiscuity increases, the benefit of kin selection needs to be larger to make cooperativity worthwhile.  So in more promiscuous cooperative species, the frequency of nests that have helpers should be lower; and it is.  Overall, then, the level of promiscuity in a species tends to correlate with a range of cooperative behaviors — lower promiscuity, higher probability of some cooperative behavior; lower promiscuity, higher probability that a given nest will get a helper; higher uncertainty about relatedness, higher chance of evolving the ability to discriminate between kin and not-kin before deciding to help.  All of this is modulated by the environment in which the species lives (and some species even vary their levels of cooperation depending on which part of their range they are living in).  If the environment is really harsh, perhaps even the chance of increasing the survival of a chick that shares 1/4 of your genes is more important than going off to try to have your own progeny.  The theory that promiscuity is a major causal factor in the development of cooperative breeding seems to explain a great deal of the data.

One last comment before I end.  In case there is someone reading this who’d like to try to use these results in birds to make arguments about human societies, let me make it extra clear that the cooperation we are talking about here is the cooperation of children with parents — not of parents with each other.  This is not a story about promiscuity breaking down society because of husbands leaving wives (it may indeed have that effect, but that’s another story) but about older brothers and sisters having more or less motivation to help bring up a new brood of offspring.  Human motivations being more complex than those of birds, I suspect that the correlation between motivation to help and relatedness will be much less marked in people.

Cornwallis CK, West SA, Davis KE, & Griffin AS (2010). Promiscuity and the evolutionary transition to complex societies. Nature, 466 (7309), 969-72 PMID: 20725039