Kin Selection

Classical population genetics models of selection focus on the relative fitnesses of individual genotypes and the subsequent impact on allele frequencies.  However, the realities of the behavior of organisms often require an expanded view of the evolutionary process.  Such is the case of what has become known as kin selection.

Darwin realized that the case of sterile castes in social insects (e. g., "worker" honeybees) presented a serious challenge to a theory based on differential survival and/or reproduction among individuals.  In the case of worker honeybees, not only are the workers sterile, but also use of their stings is suicidal.  Darwin resolved the paradox by pointing out that social insect colonies were actually families, but he did not expand upon this idea.

Recent studies of animal behavior have discovered additional cases of what is often termed altruistic behavior in which individuals appear to be reducing their own fitness by performing acts which benefit others.  Some general examples are the production of alarm calls (which presumably put the caller at risk of predation), and various forms of cooperative breeding behavior in which individuals postpone reproduction in order to facilitate the reproduction of others.

The problem in all of these cases is to explain how such a trait could increase in frequency, or at least be maintained, in the population, when it involves a clear decrease in fitness.  Although several researchers, including R. A. Fisher and J. B. S. Haldane, addressed the question, the first detailed formulation of the answer was that by W. D. Hamilton in 1964.

Hamilton explained that altruistic behavior would be favored by natural selection if the beneficiaries of such acts were close relatives of the individual performing the act.  Relatives by definition have a high probability of having alleles in common with each other, and the reduction in the fitness of the individual performing the act could be offset by the benefits realized by relatives.

The parameter that quantifies the probability of sharing alleles is the coefficient of relatedness, r, which is a measure of the probability that alleles taken from two individuals are identical by descent (see Hartl (1988), p. 134-136).  If we then define c as the cost in reduced fitness to the individual performing the act and b  as the benefit of increased fitness realized by the relative, then the allele will increase in frequency as long as:

    rb-c > 0 = b/c > r

In other words, the benefit to cost ratio must be greater than the degree of relatedness.  We can then define kin selection as a form of indirect selection for alleles that operates through the realized fitnesses of relatives of carriers of the alleles.  The fitness of the individual is then redefined as its inclusive fitness: the fitness of the individual (including direct descendants) plus the effect of that individual on the fitness of relatives.

Some idea of the constraints on kin selection can be given by examining the actual expected values for the coefficient of relatedness.

      
r
descendant kin
non-descendant kin    
0.5
offspring
full siblings
0.25
grandchildren
half siblings
0.125
great-grandchildren
nephews, nieces, cousind


Thus, J. B. S. Haldane's quip that he would lay down his life for 2 of his brothers or 8 of his cousins becomes a statement of what would be necessary to maintain a completely self-sacrificial behavior pattern.

Population geneticists have attempted to formulate the basic concept of inclusive fitness into a version which can be used to predict the population genetic effects of kin selection.  For example, Wade (1980) saw kin selection as a special form of group selection (which acts via differential extinction and proliferation of groups), in which an altruistic trait would increase in frequency if the fitness differences among families (as a result of differences in the frequency of the allele among families) were greater than the increase of "selfish" alleles within families due to classical Darwinian selection for maximum individual fitness.  More explicit population genetics models have focused on the quantity know as the covariance between traits (Queller, 1992).

The practical study of kin selection is difficult because a researcher must demonstrate the degree of relatedness and attempt to quantify the costs and benefits of a particular behavior.  Even the discovery that a surprising array of organisms can distinguish between related and unrelated individuals has been difficult to combine with unambiguous data showing that this leads to a difference in social interactions (Barnard, 1991).

In spite of these practical problems, kin selection remains a useful concept in our attempts to understand the evolution of social behavior in many species.  It is not an alternative to classical Darwinian natural selection, rather it is a special case where related individuals are the unit of selection.