Cooperation Trumps Selfishness in the Foundress's Dilemma

Summer time in Pine Valley, California, and the ant queens are flying. With summer rains impending, queen ants are embarking on nuptial flights and trying to start colonies of their own. But it’s not easy being a queen. To gain an edge, many queens in this region band together with unrelated co-foundresses to start colonies (a phenomenon known as pleometrosis). And yet not all of the queens are cooperative. Some are aggressive, founding solitarily. Why is it that in some places one behavior is dominant and in other locations another behavioral type holds sway? Such variation in ant founding behavior has long interested myrmecologists, but the basic themes of cooperation and conflict may speak to a broader mystery in the natural world – the idea that natural selection may act upon levels greater than the individual.In our recent paper, "The Foundress’s Dilemma" I and my co-authors apply just such a perspective to the question of pleometrosis in a species of seed-harvester ant.¹ The idea that natural selection can act upon whole groups or family units actually goes back to Charles Darwin himself.² But this idea, called group selection, has been controversial in recent decades.^3,4 Examples of group (more generally termed ‘multilevel’) selection have been well documented in both experimental and agricultural systems (check out the recent interview in This View of Lifebetween David Sloan Wilson and William M. Muir on the fascinating and practical application of this concept to chicken breeding).^5,6 But documented examples from the natural world have been in shorter supply.^7,8

Needle in a HaystackWhat would group selection look like in a foundress aggregation? We collected queens during the height of the nuptial flights from Pine Valley and neighboring Lake Henshaw and assembled them into ant farms in differently sized groups. We found that very quickly many of the queens dug in and began sharing the work of colony-founding: foraging, excavating, and laying eggs. But some of the queens behaved very differently, very aggressively. Within a few days, these aggressive queens were attacking their nest-mates and in short order the less aggressive queens were killed (often by the complete severing of their heads or abdomens). In the meantime, the groups composed of purely cooperative queens went about their business, quietly tending to their brood and getting along. While aggressive queens were the ‘winners’ in their groups, purely cooperative groups outlasted those containing aggressive queens. As David Sloan Wilson and E.O, Wilson put it a 2007 paper, "Selfishness beats altruism within groups. Altruistic groups beat selfish groups. Everything else is commentary."⁹ Thus, within the foundress associations of P. californicus we can make a reasonable case for just such a process playing out in nature. The tricky thing about social evolution is that there may be different ways to frame a question that may be more or less useful in different circumstances.¹⁰ Here, one could also phrase it as individual benefits that feedback via group effects. But we felt that the clearest way to understand this process was by invoking group selection.But there were still questions to be resolved. If group selection was truly at work, why did aggressive queens exist at all in the natural environment? In order to get at these mysteries, several of my colleagues designed an agent-based evolutionary model to look for testable ecological hypotheses explaining the survival of both queen types in this species. Initially, we built the model with the idea that there was some sort of ecological constraint (drought, for example) that forces the queens together. In some species of ants that show pleometrosis, this may indeed play a role. In an early version of the model we saw this expected result (no surprise). But what did surprise us was that independent of such a planned effect the model suggested that density of colonies (and queens) could drive the success or failure of one or the other type. This was purely an emergent property of the model that we had not anticipated. We stripped the simulation down, removing many imposed assumptions – and were left with our final simple model. It must be said that the prediction of our model that queen density predicts pleometrosis has long been the view of researchers such as the eminent myrmecologist Walter Tschinkel (who tested it quite nicely in fire ants).¹¹ My collaborator (and mentor in this work), Jennifer Fewell, has spent years with her various students studying queen ant behavior in San Diego County, California. It has been found that the field site of Pine Valley has mostly cooperative queens while nearby Lake Henshaw has mostly non-cooperating queens.¹² With the density hypothesis of our model in mind another colleague of ours returned to the field to see if the predictions of the model were supported. He found that, indeed, colonies in Pine Valley were more clustered than colonies in nearby Lake Henshaw. When collecting queens at the height of the colony-founding season in Pine Valley, it is typical to find queens huddling together (often more than twenty) in every available hole or weed on the ground. In contrast, in Lake Henshaw, most newly founded nests are occupied by a single queen. Thus, a queen attempting to found a colony in Pine Valley is very likely to be forced into a social living arrangement whether she prefers such accommodations or not. In addition, the Pine Valley queens must contend with what is likely intense local competition for resources. Not only are they hunted by the nearby mature colonies (who are not exactly welcoming toward new colonies setting up shop on their doorstep) but they are also in competition with every other nascent colony to produce the first batch of tiny workers (called minims). Producing workers early is important as it allows the new colony to forage and begin making territorial claims of its own. Thus, it seems likely that ecological conditions where foundress density is increased (and interspecific competition is likewise more intense) leads to a natural analog of Muir’s famous chicken experiment.⁵ In Muir’s experiment, groups of related hens were forced into confined living conditions and artificially selected based upon their group productivity (as opposed to their individual productivity). In contrast, in the case of P. californicus in Pine Valley – it is the ecological circumstances that force ant queens into close living conditions and natural selection that is the agent of change. Similar to a team of athletes, success in such an environment is all about how the group performs. To use the language of Richard Dawkins - the group is the vehicle and target of selection. Thus, individual queens (like the aforementioned team of athletes) are selected by evolution in such circumstances to work well with others. Ball-hogs (or murderous ant queens) need not apply.1. Shaffer, Z. et al. The foundress’s dilemma: group selection for cooperation among queens of the harvester ant, Pogonomyrmex californicus. Scientific Reports 6, 29828 (2016).2. Darwin, C. The Origin of Species by Means of Natural Selection. (D. Appleton and Company, 1882).3. Nowak, M. A., Tarnita, C. E. & Wilson, E. O. The evolution of eusociality. Nature 466, 1057–1062 (2010).4. Pinker, S. "The false allure of group selection." The Edge (2012).5. Muir, W. M. Group Selection for Adaptation to Multiple-Hen Cages: Selection Program and Direct Responses. Poultry Science 75, 447–458 (1996).6. Goodnight, C. J. & Stevens, L. Experimental Studies of Group Selection: What Do They Tell US About Group Selection in Nature? The American Naturalist 150, S59–S79 (1997).7. Pruitt, J. N. & Goodnight, C. J. Site-specific group selection drives locally adapted group compositions. Nature 514, 359–362 (2014).8. Eldakar, O. T. et al. The role of multilevel selection in the evolution of sexual conflict in the water strider Aquarius remigis. Evolution 64, 3183–3189 (2010).9. Wilson, D. S. & Wilson, E. O. Rethinking the theoretical foundation of sociobiology. Q Rev Biol 82, 327–348 (2007).10. Foster, K. R. A defense of sociobiology. Cold Spring Harb Symp Quant Biol 74, 403–418 (2009).11. Tschinkel, W. R. The Fire Ants. Harvard University Press (2006).12. Overson, R., Fewell, J. & Gadau, J. Distribution and origin of intraspecific social variation in the California harvester ant Pogonomyrmex californicus. Insectes Soc 63, 531–541 (2016).