Is natural selection the sole driver of evolution at the level of genes?

Steven Rose: Evolution is a fact; natural selection is a theory about how evolution occurs. But there are many factors that lead to evolutionary change and natural selection - and competition for scarce resources (the very Darwinian mechanism) - is only one of them.

Natural selection is an excellent mechanism for explaining species getting better at what species do. It is poor (as Darwin himself recognised) at explaining speciation - how a species becomes another species. For that, we need to include explanations such as founder effects, genetic drift and contingencies.

Further, Neo-Darwinism insists on one level of selection only in the individual gene. To understand evolution we need to consider multiple levels of selection, from genes to populations. I am astonished at the naive propositions being advanced by people who call themselves sociobiologists. The crudity of the way in which sociobiology is applied at present is equivalent to the crudity of vulgarly economic reductionist interpretations of Karl Marx.

- Steven Rose is professor of biology, the Open University.

John Maynard Smith: Neo-Darwinism is almost, but not quite, universally accepted by the biological community. But I do not accept that "natural selection is the sole driver of evolution" if by evolution you mean genetic change, because there is quite a lot of change at molecular level occurring randomly that makes no difference, and is all very neutral. It is not the sole driver of evolution because some changes are selectively neutral and have no effect one way or another. But if you were to ask, "Is natural selection the only cause of adaptation", then I would have to say yes, it is, by and large. But such blank universal truths are very hard to pin down.

- John Maynard Smith is emeritus professor of biology at the University of Sussex Gabriel Dover: We absolutely require natural selection but we have evidence that it cannot itself be the only explanation of the complex process of evolution. There is a variety of internal genetic mechanisms that can increase the frequencies of certain mutations: the general instability of DNA can actively promote or demote some mutations. All such mechanisms are termed molecular drive. So conventional genetics would say that a mother with the genes Aa coding for a particular characteristic would produce eggs, half of which contain the gene "A" and half "a". But molecular drive means she may produce more of "A" and fewer of "a" as a consequence of one or other mechanism of DNA flux. This increase in "A" is distinct from that of natural selection and adaptation. We have evidence of how all these internal processes interact with natural selection. Many people working on evolution have not been educated in the new molecular genetics. They make observations in the wild and have recourse to Kipling-type Just So Stories to explain them. They assume the sole role of natural selection all too glibly, armed only with the wacky pretensions of selfish gene computer games and without recourse to rigorous proof. Ignorance and received wisdom make a dangerous cocktail in science.

- Gabriel Dover is professor of genetics, Leicester University.

Marion Lamb: Neo-Darwinism is almost totally accepted by the biological community. There is no way to explain complex adaptations except by natural selection, but I would not be happy with the statement that natural selection is the sole driver of evolution at the level of genes. The classic Neo-Darwinian view is that there can be no transfer of information from generation to generation other than through genes, which are usually equated with DNA, but there are other ways, for example, through behavioural transfer and through inherited epigenetic marks. In genomic imprinting, for example, the same gene, when inherited from the male, is interpreted in the offspring differently from when inherited from the female. You only have to think of the way that behaviour is transmitted, to understand that DNA is not the be-all and end-all. Neo-Darwinism is fundamental. But what is wrong with the present version is that it says that all variation is random and it is all in DNA sequence.

- Marion Lamb, formerly of Birkbeck College, is co-author of Epigenetic Inheritance and Evolution.