At the beginning of “The Evolution of Sex” John Maynard Smith outlined a possible general theory for the evolution of sex. He suggested that if the reason for the existence of diploidy is the complementation of recessive deleterious mutations, and if this complementation was abandoned for some reasons at some point of a life cycle, sex would be useful to recover it. His main problem, however, was to explain why the very reason for diploidy – complementation – is abandoned at all, just to be recovered later using sexual reproduction; this seemed a circular argument, and he abandoned the idea. I suggest that intragenomic conflict might provide not only a solution to this problem, and a basis for Maynard Smith’s idea, but also lead to a general theory that explains other problems related to the evolution of sex and recombination. First, due to the molecular mechanism of the initiation of recombination, if an allele manages to induce the initiation of recombination on the homolog, it will have a strong advantage due to gene conversion, which would promote recombination even in the absence of an advantage for the individual. Second, recombination during asexual reproduction induces loss of complementation of recessive deleterious mutations, and the deleterious effects of this can be quantified. I show theoretically that the fitness cost of this loss of complementation in asexual species can easily be larger than the twofold cost of meiosis so that sexual reproduction is stable against invasion by apomixis for a large range of parameters in diploid species. Asexual reproduction can sometimes invade a sexual population but asexual individuals are then replaced by sexual individuals in very few generation; in some cases asexual mutants do not even invade. Triploidy, however, slows down loss of complementation enough to allow apomictic species to replace sexual species; all known apomictic species, at least in plants, are actually triploid. The results are similar for automixis and endomitosis. Third, loss of complementation may explain why meiosis is usually two-step (a duplication followed by two reductions), or has some complicated alternatives but it is very rarely, if ever, a one-step process (just one reduction). My suggestion is to discuss whether these intragenomic conflicts that I describe can lead to a comprehensive theory for the evolution of sex and recombination. In particular, I need to discuss data on the number of lethal equivalents and the occurrence of recombination in apomictic species.