Supplementary Information for “ The evolution of eusociality ”

Kin selection theory based on the concept of inclusive fitness is often presented as a general approach that can deal with many aspects of evolutionary dynamics. Here we show that this is not the case. Instead, inclusive fitness considerations rest on fragile assumptions, which do not hold in general. The quantitative analysis of kin selection relies completely on inclusive fitness theory. No other theory has been proposed to discuss kin selection. We show the limitations of inclusive fitness theory. We do not discuss implications of kin selection that might exist independent of inclusive fitness theory. We set up a general calculation for analyzing mutation and selection of two strategies, A and B, and then derive the fundamental condition for one strategy to be favored over the other. This condition holds for any mutation rate and any intensity of selection. Subsequently, we limit our investigation to weak selection, because this is the only ground that can be covered by inclusive fitness theory. For weak selection, we show that the natural selection interpretation is appropriate for all cases, whereas the kin selection interpretation, although possible in several cases, cannot be generalized to cover all situations without stretching the concept of “relatedness” to the point where it becomes meaningless. Therefore we have a general theory, based on natural selection and direct fitness, and a specific theory based on kin selection and inclusive fitness. The general theory is simple and covers all cases, while the specific theory is complicated and works only for a small subset of cases. Whenever both theories work, inclusive fitness does not provide any additional insights. Criticisms of inclusive fitness theory have already been raised by population geneticists and mathematicians (CavalliSforza and Feldman 1978 and Karlin and Matessi 1983). The present criticism is based on a game theoretic perspective in structured populations which has been developed recently. The extra complication of inclusive fitness theories arises from the attempt to bring into the discussion increasingly abstract notions of ‘relatedness’ when it is not natural to do so. This situation is not particular to theory. Hunt (2007) citing Mehdiabadi et al (2003) points out that “increasingly complex scenarios are required to keep recent empirical data within the theoretical construct of haplodiploidy-based maximization of inclusive fitness.” The fact that inclusive fitness calculations are more complicated than direct fitness calculations has been accepted by theoreticians such as Rousset and Billiard (2000) and Taylor et al (2006). As complicated as inclusive fitness is to calculate, it is even more complicated to measure empirically. Only very few studies have attempted to do this (Queller and Strassman 1989, Gadagkar 2001)