50 & 100 YEARS AGOEVOLUTIONARY BIOLOGYGenetics and bisexualityVincent Savolainen and Laurent LehmannA population-genetic model indicates that if there is a gene responsible for homosexual behaviour it can readily spread in populations. The model also predicts widespread bisexuality in humans.For human societies at large, homosexuality is a sensitive issue. For biologists it is an intriguing one1,2. How can genes influencing homosexual — and so non-reproductive — behaviour be favoured by natural selection? An answer is offered by Gavrilets and Rice in a paper that has just appeared in Proceedings of the Royal Society3. They provide a population-genetic analysis that explains why, in theory, a gene pre-disposing an individual to homosexual behav-iour would spread in a population, and that predicts its widespread occurrence in humans and other sexually reproducing species. No predisposing gene for homosexual behav-iour has been identified, but there is evidence that genetic controls are involved: for example, human twins are more likely both to be gay compared with non-identical brothers; and male homosexuality is more often inherited maternally, indicating that heritable maternal effects and/or genes linked to the X chromo-some are in operation2,3. However, unlike heterosexuals, who devote a significant amount of time to reproductive sex, homosexuals are involved in non-reproductive sex, hampering the direct transmission of any gene under lying this behaviour. Homosexuality has a cost to fit-ness — that is, the ability of an individual to produce offspring that survive and reproduce — and it can only evolve if it otherwise provides indirect benefits to reproduction. Three main mechanisms have been pro-posed in which variety in genes controlling homosexuality could be maintained in a popu-lation: overdominance, sexually antagonistic selection, and kin altruism2–4. For simplifica-tion, we will consider here male homosexual-ity, but these mechanisms also apply to female homosexuality. They also apply no matter how many genes contribute, but Gavrilets and Rice’s analysis deals with a single theoretical gene and its two variants (alleles). First, in the case of overdominance, a ‘gay allele’ would result in homosexual behaviour in an individual who has received this allele from both parents (homozygous), but would provide an advantage to the heterozygote (where only one parent has transmitted the gay allele). This situation would be similar to the renowned example of sickle-cell anaemia in Africa, a genetically inherited disease controlled by a deficient allele. Homozygotes for this allele suf-fer severe disorders. But because this allele con-fers resistance to malaria when heterozygous, it is maintained in human populations exposed to malaria. Under this scenario, heterozygotes for the gay allele may have higher success in attracting females and/or their sperm may have some competitive advantage5. In the second case, sexually antagonistic selection, a gay allele would result in a cost when expressed in males (‘feminization’ and loss of fitness), which would be counter-balanced by a fitness advantage when it is expressed in females. In the third hypothesis, kin altruism, homo-sexuals would help their own family members, increasing the fitness of their relatives and therefore the probability that a gay allele is passed on to the next generation2,4. These hypotheses have previously been speculative, but they have now been mod-elled and formalized by Gavrilets and Rice3. The authors adapted the classical popula-tion-genetic equations established by J. B. S. Haldane6,7 and describe the evolution of the frequency of two alleles at one locus, in large populations for which each allele may result in sex-specific effects on fitness. Considering hypothetical straight and gay alleles, Gavrilets and Rice document the conditions of relative costs and benefits to fitness under which the gay allele can enter a population of straight alleles and be maintained subsequently. They establish the conditions under both the over-dominance and sexually antagonistic-selection hypotheses for a homosexual gene that would be located on autosomes (non-sexual chro-mosomes) or on the X chromosome. These conditions still remain to be evaluated in the kin-altruism hypothesis. Crucially, in these population-genetic mod-els, a gay allele will produce variable degrees of homosexual behaviour, which is equivalent to the fitness cost of that behaviour (which, for example, could be interpreted as the propor-tion of time devoted to homosexual rather than reproductive sex). If one homozygous individual is not at all involved in reproductive sex, then the cost of homosexuality is maximal and this individual’s phenotype is obviously strictly gay; however, in all other combinations, homozygous individuals exhibit a degree of bisexual behaviour depending on the costs.Gavrilets and Rice show that, for a large set of costs and benefits, the gay allele can invade a population. Under overdominance, once a gay allele has entered a population it will be main-tained in a polymorphic equilibrium, and this is easier if the homosexual gene is autosomal rather than X-linked. Further, under sexually antagonistic selection, the gay allele may even 50 YEARS AGOAn almost unexplored record of the Earth’s history is preserved in the sediment under the great oceans. Profiles of the uppermost layers have been obtained by many different types of corer, and to this list Prof. B. Kullenberg has now added a modified version of his elegant piston corer which should reach deeper than ever before... The new corer carries the same weight as the old model, 1,500 kgm., but has greater penetrating power, because with a length of only 2.3 metres there is a relatively small area of wall in frictional contact with the sediment... The corer sinks down through the sediment to a depth determined by the length of cable wrapped around its upper end. When all this cable has been unwound, the piston, which up to now has been locked at the lower end of the corer, is released and a core is taken. Although this ingenious corer has reached down to 29 m., and is expected to reach 50 m. in soft sediment... there still remains the ultimate problem of how to reach and take a continuous profile of sediment which can be a full kilometre thick. From Nature 12 January 1957.100 YEARS AGOThe issue of Science for November 23 contains an article by Prof.