An even broader perspective on sex and recombination

Understanding why sexual reproduction in eukaryotes is so prevalent is a hard problem, and it has gone through a progression of stages that are typical for work on hard biological problems. First, a pioneer suggests a plausible solution, usually very general and not rigorously de®ned because the theory surrounding the problem has not been fully developed. Other pioneers may add competing but similarly broad and fuzzy theories. As interest in the problem spreads and the theory in which it is embedded matures, more detailed theories about more speci®c mechanisms are proposed; these are often presented as alternatives because of scientists' desire to be the one who solved the problem 1⁄4 the only one. The new theories become more and more detailed as the available theory space is used up. Occasionally, someone sits back and looks for a more general solution that includes all the detailed models as special cases. And often someone else comes forward and points out that many, if not most, of the models may be operating in nature, in different species or even in the same one. So it has been with the question of why sexual reproduction is so prevalent among eukaryotes (Mooney, 1992). Early hypotheses, such as Weismann's (1891) that sex facilitates evolution by increasing genetic diversity, were necessarily vague and dif®cult to evaluate because they were devised in the absence of any real understanding of transmission genetics or population and evolutionary genetics. More sophisticated hypotheses appeared after the development of Mendelian genetics and population genetics. An important example is the hypothesis of Fisher (1930) and Muller (1932) that sex facilitates natural selection for advantageous mutations, extended to selection against detrimental mutations by Muller in 1964. This was followed in the 1970s and 1980s by a proliferation of models with increasing sophistication and detail, but of decreasing generality. The books of Williams (1975), Maynard Smith (1978) and Bell (1982) contributed to the proliferation of models directly and also indirectly by making the ®eld more popular. We now have models for organisms with many different permutations of ®nite or in®nite population size, advantageous or detrimental mutations, positive or negative epistasis or no epistasis, and a variety of different reproductive patterns and ecological niches. Unfortunately, the numerous models are often presented as mutually exclusive and individually suf®cient to explain the prevalence of sex in most or all organisms. Few authors have asked if there might be a more generally applicable model that subsumes most or all of the detailed models as special cases (for two exceptions, see Felsenstein (1974) and the review by Barton & Charlesworth (1998)). Even fewer have combined the detailed models to see what happens when two or more are operating simultaneously. West, Lively and Read (1999) (WLR hereafter) have done that. They are to be applauded for emphasizing that at least some of these competing hypotheses are not mutually exclusive, and for showing that they may be more powerful, as well as more realistic, when combined.