http://psyed.org/r/nn/nnd/s/sxd/kamikaze.html
In 1988 Baker and Bellis first proposed the "Kamikaze sperm" hypothesis (Baker and Bellis 1988; Baker and Bellis 1989; Baker and Bellis 1993; Baker and Bellis 1993) . This was an extention of Parker's ideas on the role of plugs and blocking agents in sperm competition (Parker 1970; Parker 1984; Parker and Begon 1993) . In essence, they envisaged two types of human sperm, "egg-getters" and "kamikaze" forms specialized either to block competitors or to roam around the female tract seeking to destroy competitors (using acrosomal enzymes?). This was fleshed out in a highly entertaining and best-selling book, "Human Sperm Competition. Copulation, Masturbation and Infidelity" (Baker and Bellis 1995) , and has been widely publicised including a TV documentary, "Sperm Wars". Roger Short has a devastating review from the 47th newsletter of the European Sociobiological Society
The book actually provides a wonderful read, even though many of its conclusions are pretty wild and have not been published in peer-reviewed journals (sperm heads changing shape in mixed ejaculates, for example). My major problem is that they produce a most complicated explanation for what may be a straightforward problem and there is no convincing direct evidence for adaptive roles for the various morphs (Gomendio and Roldan 1991; Gomendio and Roldan 1993) . If you accept that humans produce abnormal sperm simply because we are not very fertile and the quality control mechanisms in the testis are deficient, then there is no need for an obsessive need to find an "adaptive" explanation for every aspect of sperm biology. In fact Occam's razor demands that we should always prefer the simplest explanation even if it doesn't grab the popular imagination.
Harcourt (1991) has shown that the numbers of abnormal sperm in mammals -- and especially in Primates -- do not correlate with mating pattern. If the "kamikaze sperm" hypothesis were correct one would expect to see more abnormal forms in species with high intensity of sperm competition, whereas if anything the reverse is true. Baker and Bellis responded by suggesting that variation is likely to be greater in species that have cryptic estrus (like humans) but of course this doesn't explain high levels of pleiomorphism in sperm of gorillas where the chances of sperm competition are pretty minimal, nor in a diverse range of species ranging from thoroughbred stallions (cryptic estrus in mares?) to koalas, cheetahs and other big cats (Cummins 1990) .
I suspect that Baker and Bellis have fallen into the Panglossian trap of assuming that every aspect of sperm biology is adaptive. Indeed abnormal sperm can't even be considered to be exaptations (features that may enhance fitness but that were not designed by natural selection for their current role) (Gould and Vrba 1982) because they do not normally participate in fertilization. While abnormally shaped sperm (in mice) can generate normal embryos when microinjected into oocytes (Burruel, Yanagimachi et al. 1996) , in normal circumstances barriers in the reproductive tract such as cervical mucus and the zona pellucida actively select against abnormal forms (Katz, Morales et al. 1990; Liu and Baker 1994) and are probably the major selective mechanisms for defining the heredity of sperm form (Bedford 1991) .
Kamikaze sperm? I'm not convinced. I'm happy to post any comments or rebuttals, though.