Mutual altruism and long-term optimization of the inclusive fitness in multilocus genetic systems.

The dynamics of long-term evolution in a complex genetically-structured population with a flux of random mutations is employed here to study the evolution of mutual altruism between relatives that are encountered repeatedly, where the level of altruism is measured by the risk one is willing to accept in order to save the life of one's relative. It is shown that regardless of the number of loci involved, of the rates of recombination among them, and of the intensity of the selection forces, the long-term dynamics can phenotypically converge only to a level of altruism that maximizes the individual inclusive fitness as it ha previously defined by students of the individual approach to evolution. Except for the widely studied case of weak selection, however, the convergence to such a level of altruism is not necessarily generation-to-next monotone. It is further shown that, unlike the case of the one-shot encounter, repeated encounters between relatives allows for more than one level of altruism which may maximize the inclusive fitness, in which case not all such levels of altruism are evolutionarily accessible.