This site is not maintained. Click here for the new website of Richard Dawkins.


← A misguided attack on kin selection

s.k.graham's Avatar Jump to comment 18 by s.k.graham

Ok so with C=0, the equation basically says "a trait can develop if it has a positive influence on the queen". Duh, that's a triviality.

Note that you can make an equation like "rB > C" always hold in any case by just making up values for B or C; I don't see the insight that gives you in cases like this where it resolves to a triviality like "B > 0".

C is only (approximately) zero for some genes and not others. The equation applies to any and all genes. Also you don't get to "make up values" for B and C... the values are what they are -- you may or may not be able to easily measure them, but you don't just get to "make them up". Some genes will have rB>C and other genes will not. The latter genes will die out. The former may or may not flourish depending on whether which among competing genes has the largest rB-C.

rB>C is no more "trivial" when C is zero than when C is any other number. Who cares if it is a simple trivial equation? If it describes a (minimum) requirement for natural selection to favor a gene -- the so be it. Actually my expression rB-C is really the more important thing, or more precisely sum(rB)-C -- this is basically the "inclusive fitness" of a give gene (really we should be saying allele, here, to get technical), and it is this notion of inclusive fitness that is the real insight. Between two alleles competing for the same spot on the genome, the one with higher sum(rB)-C will be favored (but if they are close, chance plays a role, or neither becomes dominant in the population).

The genes responsible for sterile worker castes were selected because they had higher sum(rB)-C than competing genes within the population. In the context of already sterile worker caste, genes which have negligible effect on reproduction of the worker have approximately C=0 in Hamilton's equation, but that does not mean the equation does not apply, that just means C=0 for those genes.

Finally if the workers did not (high high probability due to kinship) share their genes with the queen, then it would not matter which traits they had because those traits would not be passed down to future generation of workers. This is why we say 'colony selection' or 'group selection' is not enough. It is the success of the genes that matter. Sterility of the workers is not an absolute. If some mutation (or novel combination of genes) causes a worker to have a small chance of reproducing under certain circumstances (negative C), without having an overly negative impact on the queen's reproduction (negative B), then rB-C will be a net gain for the genes in the worker and since there is a chance the queen has the same mutation, there is a good chance that that gene will flourish.

So it is the inclusive fitness sum(rB)-C of a gene (allele) that always matters, and it is sufficient explanation for 'altruistic' and 'cooperative' behaviors. Simply enhancing 'group fitness' is neither necessary nor sufficient for a gene to thrive. The same is true of 'individual fitness'. There is an overlap between group fitness and sum(rB), but group fitness can contain irrelevancies (because groups do not necessarily comprise kin), and sum(rB) includes things that group fitness does not (because kin do not all necessarily form cohesive groups). Individual fitness (individual reproductive success) is reflected in C, but in terms of a reduction of individual fitness

Tue, 31 Aug 2010 19:02:45 UTC | #508786