[ print ]

Senior Lecturer in Behavioural Biology School, University of Bristol
Reassessing Relatedness

I have been a true disciple of kin-selection theory ever since I discovered the wonders of social evolution as a young graduate student. Kin selection theory emphasizes the importance of relatedness (i.e. kin) in the evolution of social behavior. The essence of social living lies in sharing tasks amongst group members; for example some individuals may end up monopolizing reproduction (dominants or queens), whilst others defend or forage for the group (helpers or altruists). The key to understanding how sociality evolves rests on finding a watertight explanation for altruism. Why should any individual sacrifice their reproductive rights in order to help another individual reproduce?

When groups consist of families, there is an intuitive basis for the evolution of altruism. Helping relatives, with whom I share many of my genes, is potentially a lucrative strategy for passing my genes on to future generations. This reasoning led W.D. Hamilton to his theory of inclusive fitness, or kin selection, in 1964: a social action evolves if the benefit (b) weighted by the relatedness (r) between group members exceeds the costs (c) of that action (i.e. br>c). Evolution is satisfyingly parsimonious, so it is only natural that an apparently complicated thing like sociality can be explained in such simple terms.

I am certain I speak for many students of sociality in being eternally grateful to Hamilton in providing such an elegant theory with such clear predictions to test. Off we go, armed with Hamilton’s Rule to settle our quest for understanding what makes an animal social. There are three things to measure: relatedness between group members (or actors and recipients), costs (to the actor in being altruistic) and benefits (to the recipient in receiving help). Happily, the molecular revolution has brought gene-level analytical tools to behavioral ecologists, allowing relatedness to be quantified accurately. Costs and benefits are more problematic to quantify, as they might vary over an individual’s life-time. Relatedness, therefore could be a fast-track route to securing the secrets of sociality in a kin-selected context.

The social Hymenoptera (ants, bees and wasps) are an excellent group for studying sociality because they live in large groups and have a peculiar genetic sex-determination system (haplodiploidy), engendering high levels of relatedness. If relatedness predisposes any animal to be social, it will be the Hymenoptera. As an altruist in a social insect colony there are several ways by which you could favor your most closely related group members. You could selectively feed sibling brood that share the same parents as you. Or, you could eat the eggs laid by your siblings in preference to those laid by your mother (the queen). On planning an elopement from the homestead to start a new colony, you might choose full-sisters as your companions rather than a random relative. The predictions are elegant, simple and depend on the kin structure of a specific colony.

With insect cadavers mounting up in university freezers all over the world, we raced to test these predictions. The results were disheartening: worker wasps headed by multiply mated mother queens were not maximizing their indirect fitness by laying heaps of parthenogenetic male eggs; worker ants were unable to optimally manipulate the brood sex ratios (and hence their inclusive fitness) in relation to how many times their mother had mated; social wasps were feeding larvae in regard to their need rather than relatedness; swarming wasps were indiscriminate in who they founded new colonies with.

On the back of robust experiments like these, I have changed my mind about relatedness being the primary dictator of social evolution. Insects are unable to discriminate relatedness on an individual level. Instead, relatedness may act at the colony/population level, or simply in distinguishing kin from non-kin. This make sense. An individual-level kin discrimination mechanism is vulnerable to invasion by an occasional nepotist, who would favor its closest relatives over others. As the gene for nepotism spreads, the variation on which the kin-discrimination is based (e.g. chemical or visual cues) will disappear and individuals will no longer be able to tell kin from non-kin, let alone full siblings from half siblings: sociality breaks down. We knew this long before many of the kin-discrimination experiments were done, but optimism perseveres until enough evidence pervades.

Does this mean kin selection theory is wrong? Absolutely not! The reason for this is that relatedness is only one (albeit important) component of kin selection theory. The key is likely to be the interaction of a high (and variable) benefit to cost ratio from helping, and a positive relatedness between actors and recipients: relatedness does not have to be high for altruism to evolve, it just needs to be greater than the population average. I still believe you cannot hope to understand sociality unless you put relatedness at the top of your list. But, we need to complement the huge amount of data generated by the molecular hamster wheel with some serious estimates of the costs and benefits of social actions.