Comment 21 by kriton

Zeuglodon, the unit of replication would be either the individual gene or the whole bacterial genome.

In Comment 9, you were talking about bacteria and organisms. At least concede you've just changed tracks.

But the individual gene is expressed as a protein

Do you know that the cellular processes inside and outside the nucleus are regulated by these things? Again, you can't have the reproducing organism without genes.

and the genome is expressed as an individual organism. I argue that genomes can make identical copies of themselves.

Except that the genome suffers the same problem as the organism - recombination quickly "destroys" it, and the process is only an extension of genetic-level replication in the first place. How would you explain viruses and non-inherited gene-exchange (I believe it is called epigenetics) on this genome-centric model?

Why would the resulting organism have to be identical atom for atom? As I said before, I'm not exactly the same as one year ago, but I'm still the same individual from an evolutionary point of view.

Because that's what replication means. The errors per replication can't be so severe that they'd never have identical copies yet distinct generations, otherwise there's no selecting going on at any point. It would be a random process and the resulting organisms would be increasingly likely to die off. Vehicles have less requirement for atomic exactitude because their job is simply to square off with the environment, and that involves changing and taking changes. A replicator needs to be stable over several generations. Jos Gibbons made the same point when he described acquired characteristics earlier.

And that's not the empirical evidence I'm asking for. I'm asking about altruistic behaviour.

What a strange request. The altruistic behaviour was what the theory was invoked to explain in the first place.

Parental care, sibling care, mobbing, food distribution among relatives, self-sacrifice et al. are all examples of altruistic behaviour. Hymenopterans are, again, a particularly rich source of such evidence, but mammalian caring behaviour, avian roosting behaviour, crocodilian caring behaviour, and several kinds of fish and marine species' maternal/paternal behaviour (e.g. that of seahorses and octopus mothers) would also be fine examples. Family units in nature are generally the examples of most interest. If you wanted to vindicate Hamilton's theory, however, you'd have to perform a genetic analysis at some point, and the obvious animals I would look at are hymenopterans.

If the 1/2, 1/4 and so on only applies when genes or gene variants are not already widespread in the population, then I think I get it. But if that is the case, I would say this is something that should be pointed out MUCH more. If altruistic gene variants have been successful, they should be widespread in the human population already.

Well, duh. Most of them will be. That's the point of the competition between alleles - that one of them eventually supplant the other. This isn't always clear cut, however. Sometimes, alleles are highly conditional on what's going on elsewhere in the gene, and this can hamper their performance. For instance, Gene A1 might beat its allele Gene A2 only if Gene B1 isn't around.

And then those 1/2 and 1/4 numbers should be irrelevant for the current human situation. I don't think I'm the only one who gets confused. It seems you are explaining things I'm not asking about.

Ah, I think I see where we're getting confused. You have to take into account the process by which a gene reaches fixation in the first place.

Imagine: you're a new mutant on the scene, and the process of reproduction goes like this. Copies of your altruistic allele are created in the gametes. But your reproductive partner is also shuffling their genes, including your rival allele. The shuffling means that, when copulation occurs, you have no idea whether the offspring will have a copy of your own allele or the rival. There's a fifty-fifty chance. Your offspring is born. What's the best way to treat them? The optimal would be to go along with the fifty-fifty chance that the offspring inherited your allele, so they are a half a priority compared to yourself. If you have more children, then obviously your chances have gone up that any one of them will have your gamete. In these cases, it would be worth behaving more altruistically than usual towards them, but there's still some holding back and there will be tensions between parent and child.

If the mutant allele in this case is successful, (its success is conditional), then the population of the behaviour will increase generation by generation. Obviously, the strategy can be refined by additional genetic inputs on other genetic loci, but the principle is the same because each gene on each locus is following the same general rule. Even when it reaches fixation, the allele will still be programming its host to use the rules of thumb that ensured it got there in the first place.

Your scenario can't come about by the introduction of a new pan-species altruism allele, because that would suffer against less suicidal rivals that were measured against the mendelian probabilities of being passed on to the next generation. The only way for your scenario (of pan-species altruism) to come about would be if every last copy of the fixated allele miraculously mutated such that the resulting phenotype treated every species member it met as though they were genetic clones. This is, to say the least, incredibly unlikely.

Remember, this whole competition is against rival alleles. The strategy that gave an allele an edge over its rival will stay there when it drives the rival allele to extinction. And don't forget that multiple gene interactions can make the process very complicated in the real world.

Does this make it clearer? Please say if it doesn't, as it's important you get these points.