Why sex? It's for repair, stupid. Michod (Ecology and Evolutionary Biology/Univ. of Arizona) says that sex is not for diversity in the gene pool (the conventional wisdom), but rather to repair genetic damage and rid the genome of unwanted mutations. Remember all that business you learned in biology about sexual division (meiosis), that complicated process by which chromosomes split various times, then come together at fertilization to produce an offspring with genes from all four grandparents? Well, that certainly makes for diversity, argues Michod, but it's secondary to keeping the gene lineage pure: That chromosome activity can repair damage. In defense of this provocative idea, the author reviews the course of evolution from asexual and sexual reproduction in single cells on to complex organisms, explaining the increasingly sophisticated means by which DNA replication is controlled and mistakes are corrected. Using mathematical models and examples drawn from nature he illustrates the high cost of sex (energy consumed in searching and wooing, chance of disease, etc.) to demonstrate that sex must be doing something vital. That something turns out to be preserving the genome. Shades of Richard Dawkins (The Selfish Gene, 1977): Sex is not for the pleasure of thee and me, it's just the genes' way not only of making other genes but of making sure those genes are clean. Michod attempts to clarify by way of diagrams and chapter notes that may challenge the general reader, as does his soaring last chapter, in which he argues for both the unity of life and the distinctiveness of species. No doubt many will respond that there must be more to sex than repair, and some will raise the issue of such phenomena as transduction in bacteria and viral infection as ways in which nature mixes genomes for better or worse. But Michod's ideas surely merits a hearing. Sure to spark a lively debate.