Written in 1994 for my final-year BSc. Genetics degree dissertation, so dated and missing the references.

The Genetics of Human Homosexuality

Recent research suggesting a genetic component in homosexuality has ignited great controversy. Support is strong amongst those who regard the establishment of the cause of sexual orientation as such an immutable force as genetic predisposition as a key to greater tolerance in society. (Bailey and Pillard, 1991b). They are supported by many modern researchers of behaviour keen to use ever more powerful genetic tools to investigate previously unfathomable behaviours and place them within the context of modern evolutionary theory. In opposition are those who hold philosophical, moral and scientific objections to such deterministic arguments and are wary of the chequered past of human genetics. (Fausto-Sterling, 1985). Recent well-publicised studies by Hamer, claiming to have mapped a gene influencing male homosexuality, and LeVay, reporting an apparent anatomical difference in brain structure in homosexual men, have invigorated the debate, and many believe that the time of a genetic aetiology has come. It is unfortunate that the vast majority of research and theorising concerns male homosexuality. The lack of work on lesbianism reflects its lower profile, although the study of female homosexuality is equally valid theoretically. Fortunately, some few good studies on women do stand out, and comment on them is justified.

The scientific definition of homosexuality is a very necessary first step, but has been a subject of considerable controversy. (Friedman, 1988). Although homosexual behaviour is regarded in a different way from society to society (Dickemann, 1993), it can still be regarded as a universal human characteristic. Whilst the rate of homosexuality varies from culture to culture, there remains a homosexual pattern of behaviour or attitude, present in low frequency, that differs from the heterosexual norm. There exist individuals that have no sexual interest in members of the opposite sex, and they are born, raised and exist within a critical homophobic environment. Despite this, orientation appears to be sustained over time, (Gebhard, 1974), suggesting an innate and very powerful homosexual drive. These individuals are defined in a way important to any behavioural researcher, that of reproductive strategy. Evolutionary theory teaches that reproduction is all, and homosexuality is a common pattern of behaviour that directly and adversely affects it. Criticism is made of the investigation of homosexuality on the grounds that either the group under investigation is too small or that homosexuals are not sufficiently homogeneous in sexual orientation to define them objectively. Yet adults that can define their own sexual orientation must also define their pattern of behaviour, and to define oneself as homosexual is a significant departure from the normal heterosexual majority. This and the high frequency of homosexuality in the population compared to other non-heterosexual behaviours, such as fetishists, indicates the existence of a significant group of individuals distinctly apart from the heterosexual majority. Their existence should be explained.

The first detailed reports on human sexual practices were the 1948 and 1953 Kinsey reports. Although the results have since been criticised on the grounds of sampling error - many of those involved had been imprisoned and hence were more likely to have engaged in homosexual activity - a wide range of human sexual behaviour was revealed. Kinsey regarded sexual orientation as being a continuum of largely subjective self-evaluations of fantasy and experience, and devised a scale that attempted to define a individual's orientation relative to exclusively heterosexual (0) or homosexual (6). Exact placement depends on sexual behaviour, fantasy and self-identification. It can be assumed that those scoring five or six can be regarded as the homosexual group, almost entirely homosexual in experience, reactions and sexual feelings towards the opposite sex. Gebhard's revisions of Kinsey's data (Pillard et al, 1986), suggest a frquency of around 4% for such individuals if male and 1 to 1.5% if female. Given the personal nature of such a classification, such surveys must be regarded with scepticism, but the presence of a homosexual group, however small, is an undoubted fact. Self-identification as being sexually responsive to members of the same sex and Kinsey score are used as a basis for determining sexual orientation in modern studies, and a bimodal distribution indicating the validity of the assumed heterosexual/homosexual duality results. (Hamer et al, 1993).

Modern theories of the development of human sexual orientation are varied. It is not known whether genes can induce a specific behaviour or if they may instead affect a personality type or predisposition more likely to behave in the desired fashion, and it can only be assumed that the correlation between predisposition and behaviour is strong enough to exhibit direct genetic affects. Geneticists acknowledge that human behaviour is enormously complex, and their best models represent the interaction of the environment and many genes of varying levels of expression. In contrast, a variety of solely environmental causes have been postulated, from Freud's psychoanalysis (1905) to modern social learning theories. (Hart and Richardson, 1981). All suppose that external influences in the environment during childhood determine adult sexual behaviour. Geneticists now believe that genetic mechanisms may be as important and that the understanding of them will be at least as illuminating as untangling the complex environmental influences. Experiments designed to study the genetic basis for homosexuality will also shed light upon the kinds of environmental mechanisms influencing development, even if no genetic effect is discovered. (Reiss et al, 1991). In addition, psychological theories are thus far characterised by lack of any significant scientific evidence or predicative success. For example, the Xambia of Eastern Papau New Guinea practice universal male homosexuality from ages eight to twenty and are taught very negative attitudes to women's bodies, yet with no reported increase in adult homosexuality despite this strong behavioural influence. (Stoller and Herdt, 1985). This lack of success has led, rightly or wrongly, to an increase of interest in a genetic explanation of homosexuality.

A genetic explanation for homosexuality would be supported by evidence that homosexuality may be identified with some discernible physical phenomenon, neurological, physiological or anatomical. A gene requires some biochemical or anatomical intermediary between it and the behaviour, although some attempts have been made to link the two directly. (Macke et al, 1993). A physical sexual difference would not exclude non-genetic causes such as prenatal hormone levels or environmental influences, but could indirectly support a genetic hypothesis, since the mechanisms of gender orientation could then be reduced to a physical state that can then be explained by genetic factors. Animal models have provided a spur to such research. Sexual dimorphism has been observed in the hypothalamus of rats, where a certain group of cells undergoes development specific to sex under the control of male androgen hormones. (LeVay, 1993). The assuming of a receptive mating position by female rats, a sex-specific behaviour, appears to be hormonally controlled. The search for comparable human anatomical dimorphism or hormonal behaviour regulation has proved less successful. The theory that adult hormonal constitution determines sexual orientation has failed to be demonstrated experimentally (Downey et al, 1987; Byne and Parsons, 1993) and hormonal therapies do not change sexual orientation. There is also no evidence for changes in orientation due to alteration of hormonal level due to trauma, disease or surgery. No neurochemical differences between homosexuals and heterosexuals have been demonstrated. A more sophisticated theory suggests that prenatal hormonal levels may influence sexual orientation, although the evidence is inconclusive (Byne and Parsons, 1993). The medial preoptic region, a region of the hypothalamus, does appear to be sexually dimorphic according to more recent research on humans. (LeVay, 1993). This is an important region in the generation of male sexual behaviour in monkeys. Whilst not determined as a homologue in humans, examination of this region for physical structures pertaining to sexual behaviour was performed by LeVay (1991) in an attempt to find a morphological difference in brain structure between homosexual and heterosexual men. 41 autopsies were performed, of which 19 were homosexual men, 16 heterosexual men, and 8 heterosexual women. Analysis showed a highly significant difference in one of the interstitial nuclei of the anterior hypothalamus, INAH 3, between heterosexual men and women, the male nuclei being approximately three times the size of the female. The establishment of a neurological sex dimorphism is made more interesting by LeVay's assertion that the study showed no significant difference between the size of INAH 3 in heterosexual women and the homosexual men studied. This suggests a dimorphism related to male sexual orientation as great as that related to sex, a possibility that justifies many assumptions about the dimorphic nature of sexual orientation and the scientific methods used to investigate it, neurological and genetic. The exceptions to the pattern demonstrate that some other factor influences node size. This requires to be elucidated before LeVay's conclusion can be accepted, possibly by a systematic study relating node sizes with life-style or cause of death.

Criticisms may be made of LeVay's research and any search for a anatomical dimorphism in heterosexuals or homosexuals. The plasticity of brain structure is well known and the functional architecture only partially understood. Since the presence of a distinctive physical difference may be the result of sexual behaviour rather than the cause; given that brain structure develops with use, as exemplified by the differing structures in blind or seeing people, it does not necessarily follow that observed differences are the cause and not the result of homosexual behaviour. (Byne and Parsons, 1993). More specific criticism is made of LeVay's study of INAH3. It is observed that his specimens were obtained from AIDS victims, and the disease is well known to cause brain damage that could lead to the node size differences observed. A number of factors suggest this is not the case. Heterosexual AIDS sufferers did not differ in INAH3 size; subjects with differing INAH3 volumes did not differ in case history, such as duration of illness; and the other three INAH cell groups in the specific brain area did not differ in volume when the homosexual and heterosexual groups were compared. Corroborating evidence is provided by a subsequent examination of a homosexual who died of non-AIDS causes and showed the same INAH3 size difference. If LeVay's findings are to be believed, and they require replication and a study of hormonal sensitivity, they represent an experimental basis for the theoretical standpoint used to justify research into human behaviours on an objective and strictly scientific level.

Without a clear theory of the aetiology of sexual orientation from neuroscience, psychology or physiology, genetics has attempted to fill the gap. Studies of the extent of genetic influences in the determination of human sexual orientation have been performed for many years with differing degrees of success and credibility. Kallman's 1952 study of male homosexuals indicated a concordance rate between identical twins of 100%, a figure later believed by Kallman to be a statistical construct due to recruiting bias. (Lewontin et al, 1984). Successive studies have suffered from similar methodological problems, but a few recent investigations appear to have avoided many of the errors of earlier studies and have been meticulous in their examination of this difficult area. Especially notable are recent genetic studies of female sexual orientation, a topic not previously examined with any rigour or enthusiasm.

The first requirement of a hypothesised genetic factor in a observed phenotype must be a characterised familiarity. Pillard and Weinrich (1986) studied rates of homosexuality in the siblings of both heterosexual and homosexual men; heterosexuals had a frequency of homosexual brothers in line with accepted figures on frequency of homosexuals in the population, but homosexuals were found to be four times as likely to have a homosexual brother. Bailey and Benishay (1993b) performed a similar study on women, with homosexual sisters of homosexuals some five times more frequent than of heterosexuals. Both these figures support male and female homosexuality being familial. Interestingly, the studies were not conclusive on whether male and female homosexuality were cofamilial, whether homosexuality is familial irrespective of sex. Homosexual men or women were not significantly more likely to have a homosexual sister or brother respectively. Postulating a different form of genetic transmission of homosexuality depending on sex is not justified on the evidence available, but must be considered as a further complication in the search for a root cause. It is important to realise that familiality does not imply a genetic influence, merely that homosexuality runs in families. This is required for a genetic hypothesis, but not evidence for it; common environmental or other influences within the family may be responsible. However, having produced evidence supporting familiarity, it is feasible to attempt to ascertain the relative influence of the hypothetical genetic as opposed to environmental factors. to determine the extent to which variation in sexual preference is due to heritable factors. Such estimations are best obtained by twin studies, in which the genetic influences in identical twins are equal and comparisons can be made in the extent of variation between them. The ideal situation is that of identical twins reared completely apart from birth, but this is of course very rare. Such studies are assumed to be free of common environmental influence upon the subjects, but equal genetic influence. With these cases too rare for easy analysis, and sampling error more significant in small studies, the more feasible method adopted is to compare the concordance of dizygotic (DZ) twins, identical or monozygotic (MZ) twins, or adopted siblings. The result the extent to which variation in sexual orientation is due to a genetic cause, the heritability of the trait. No assumptions are made as to the mechanism to genetic control, other than it involves a polygenic system in which the trait is influenced by many small factors that combine additively. (Plomin et al, 1990).

A study by Buhrich et al, (1991), reinforced the familial nature of male homosexuality, but although complex analysis indicated that further research would be worthwhile, it proved inconclusive in demonstrating a genetic component. The study of male homosexual twins and adoptees by Bailley and Pillard (1991b) was indicative of a significant level of heritability of between 0.31 to 0.74 in males, depending on the assumptions made about the degree of error in selection of the probands and the actual level of homosexuality in the population as a whole. These results indicate a significant degree of genetic influences upon the development of male homosexuality. A study by Bailey et al (1993a) using similar methodology found a corresponding heritability for female homosexuality in the range 0.27 to 0.76, again supportive of a high level of genetic influence. Other twin studies of female homosexuality have largely been isolated case reports, uninformative due to small size and lack of systematic recruiting. (Bailey et al, 1993a). A report by Eckert et al (1986) found four sets of discordant MZ female twins reared apart, and attributes this to environmental influences, but describes these results as descriptive due to the sample size. King and McDonald (1992) produced a smaller study of 46 men and women, showing a higher concordance level for homosexuality in MZ as opposed to DZ twins, which would support a genetic influence, but since there is no way to differentiate the two sexes involved, no comparable figures for heredity are possible unless one assumes the aetiology of both male and female homosexuality to be the same, an assumption not justified by observation. Another study by Whitam et al (1993) found higher concordance rates for homosexuality in MZ than in DZ male and female pairs, and also found increased concordance in three sets of triplets. These figures would seem to support a genetic hypothesis. The authors also found concordance of sexual orientation in male-female DZ twins sets, which would suggest similar factors acting on each, but this hypothesis requires empirical evidence.

Criticisms of the methods and principles behind twin and adoption studies is possible. Bias on the collection of subjects is the most likely source of error, since they are commonly recruited from sources not necessarily representative of the population. Older studies often used the institutionalised as subjects. (Lewontin et al, 1984). The modern approach is commonly to recruit through the pages of homophile organisations. Samples may thus be atypical of the population in general; motives for participation may vary, and the effect on results is unpredictable. Whether twins themselves are representative of the population has not been ascertained. (Kendler, 1993). In addition, the nature of twin pairs volunteering for similar studies has often been suggested to be biased towards those twins who are both aware of and favourably inclined towards their own similarity, that in fact welcome attention drawn to their common characteristics. Twins less comfortable or unaware of this may be less common volunteers for such studies, both MZ and DZ. Such error does not invalidate an conclusion of partial heritability, but will cause error in heritability parameters. However, the two studies by Bailey et al, (1991a, 1993b), did attempt to take account of this concordance-dependent ascertainment bias, and even assuming a high degree of sampling error, significant values were calculated for heritability, albeit at the lower end of the ranges. A more unpredictable type of ascertainment bias is if the closeness of the proband's relative, whether DZ or MZ twin as opposed to how similar they may be, affects the likelihood of the proband's participation. This could lead to significant error in observed concordance rates. (Lykken et al, 1987). In Bailey et al's study of males, (1991b), it was noted that homosexual probands with discordant adoptive siblings were less likely to give permission for the siblings to be contacted than homosexual probands with concordant adoptive siblings, whilst the pattern of co-operation was similar for DZ and MZ twins. If similar factors influence participation, that homosexuals are more willing to come forward if their heterosexual siblings are biologically related, than this would cause heritability to be underestimated due to the exclusion of discordant adoptive siblings. In the female study, the similar rate for concordance in DZ cotwins (16%) and non twin sisters (14%) compared to nontwin sisters in Bailey et al (1993b) suggests that concordance-dependent bias was minimal, since the latter study recruited without reference to familiality and consideration of siblings' sexual orientation would be less likely to influence a decision to participate. An ascertainment bias was present in the over-representation of MZ twin pairs, approximately two-thirds of the sample. The implications of this are unknown, but present no obvious source of error. More seriously, the male study found that only 9% of nontwin brothers of homosexuals were themselves homosexual, whilst the genetic hypothesis would predict a rate similar to that of DZ same-sex cotwins, 22%. Similarly, the rate of homosexuality in adopted brothers of homosexual probands was 11%, above the expected population rate of around 4%, and equal to that of nontwin biological brothers. These figures cast doubt upon a simple genetic explanation. That adoptive brothers were more prone to homosexuality than biological suggests an environmental explanation. The correct diagnosis of zygosity, whether MZ or DZ, has caused problems in previous studies. Twin self-assessments are unreliable, but the studies by Bailey et al (1991b, 1993a), also involved comparison of physical characteristics and only twin pairs fully identified for zygosity and sexual orientation were used. This is sufficiently accurate, (Plomin et al, 1990), and since errors tend to be MZ twins classified as DZ, any error is likely to have falsely increased DZ concordance and reduced heritability. A serious consideration is the assumption of equal environments acting upon the sibling pairs, that the influence of their upbringing is equal for each of them, whether DZ, MZ or adopted. If this were untrue then the genetic effects cannot be differentiated from environmental ones. The equal environments assumption would seem justified for the two twin types. They share sex, a common womb and family environment, and age; however, if MZ identical twins are treated more similarly than DZ fraternal twins due to their physical or perceived similarity, or less similarly due to the deliberate exaggeration of differences between them, the assumption is invalid. What research has been done on twins mistakenly misclassified by their parents or themselves supports the equal environment assumption. However, MZ twins do appear to experience more similar environments, although the effect this will have on behaviour is unknown. Experimental studies have indicated a low correlation of measures of equal environment with differences for cognition and personality scores in identical MZ twins. Similarly behaviour differences do not appear to correlate with physical similarities. The evidence suggests that the equal environments assumption may be cautiously accepted (Plomin et al, 1991) but criticism is such that it must be accepted that the validity of these studies is reduced. (Lewontin et al, 1984). No concrete evidence exists. Similar problems arise with the unequal environments assumption in twins separated for adoption. Studies have shown a degree of correlation between the adopting families due to selection by adoption agencies. (Kendler, 1993). The studies, then, although suggestive that the genetic hypothesis is one that deserves attention and further investigation, are not sufficient proof to claim that homosexuality has been definitely demonstrated to have a significant genetic component. The observation that half the MZ twins raised in similar environments are still discordant for such a major characteristic as sexual orientation indicates considerable complexity in determination of adult behaviour. For all these reasons the genetic hypothesis requires a large definitive study free of all possible ascertainment bias and conducted on identical twins reared completely apart combined with full information about homosexual behaviour in the population as a whole. Such a high degree of rigour is necessary in such a complex and controversial field, but is probably impossible given the social stigma attached to homosexuality and the rarity of such adopted-apart twins. In the future, the divination of neurological structure development and associated genetic systems may hold some hope, but this is some way off. The evidence is persuasive, and definitely justifies further research, but is not conclusive, and cannot yet be accepted.

Given this possibility of a genetic basis for homosexuality, some studies have attempted to define the nature of the postulated genetic mechanisms behind the expression of homosexuality. Hamer et al (1993) performed a gene linkage study using family pedigrees and multiple DNA markers in an attempt to locate a causative gene or polygenic system and to ascertain the genetic system underlying the heritability figures. Unfortunately, the study examined only males; however, the results are interesting. The study of the families of homosexual men produced evidence that homosexuality is more common amongst maternal uncles and male cousins. Such a pattern of inheritance is suggestive of an X-linked dominant single-gene effect, a recessive gene located on the X chromosome. Subsequent DNA-linkage analysis suggested linkage of at least one sub type of male homosexuality to a gene locus on the subtelomeric pseudoautosomal region, Xq28, a very specific and controversial conclusion.

The location of a purported homosexual gene by Hamer et al has received particular attention and criticism. (King, 1993). It is observed that the area to which the gene is located is dismayingly large, containing many hundreds of genes; the usefulness of locating a behavioural gene to such a rich area is questionable. It remains necessary to isolate the gene exactly. Fundamental genetic problems have also been raised; only family pedigrees demonstrating definite X-linked recessive inheritance patterns were studied, intentional bias to maximise the chances of finding a gene, so the study applies to only a select group of male homosexuals at best and is methodologically questionable at worst. It is possible that maternal relatives may be more likely to reveal their sexual orientation, since they may be closer to their female relatives. Significant under-reporting of of nominally heterosexual relatives occurred. The study also contained examples of a substantial number of probands with paternally-related homosexual relatives and homosexual sisters, indicative of a more complex aetiology than simply a single-gene effect. A population rate of 2% for homosexuality was used, based on unpublished results that cannot be commented on. A classic trait of an X-linked condition, a more extreme effect in females homozygous for the gene for the condition, also remains to be demonstrated or its absence justified. It must be demonstrated that the trait is not passed from affected father to son. This is especially hard to study given the reduced fecundity of male homosexuals. The product or function of this gene and its unknown alleles remains to be investigated, to establish the mechanics of its operation. The frequency of the allele in heterosexuals is unknown, so its effects cannot be determined. Tellingly, previous studies have mapped specific genes to chromosomal locations at similar levels of certainty only to be removed from that location when their existence is proved to be untenable, for example with manic-depressive illness. (Davies et al, 1991).

Should it be supposed that there is some genetic component in human sexual orientation, how may it be reconciled with current evolutionary thought? Behaviour that reduces an organism's ability to produce progeny poses some definite problems, but the proponents of a genetic influence must provide a convincing explanation. A single gene defect is insufficiently complex to explain the behaviour. However, genetic mechanisms explaining the existence of alleles that reduce male mating success are theoretically feasible for a variety of hypothetical situations involving effects on male, female or pair parenting or mating success. (Getz 1993). These models may be used to justify a given behaviour genetics theory, but cannot provide an explanation of homosexual behaviour in themselves. Wilson (1975, 1978), hypothesised homosexual behaviour as a form of altruism, benefiting the group by providing childless individuals able to support the group's children. Theories of group selection have since fallen out of favour; however, the concept of homosexuals providing assistance in child-rearing has been retained. Rather than benefiting the group, such behaviour can be seen to assist the offspring of the homosexual's heterosexual siblings, hence restricting the beneficiaries of such genetically-inspired altruism to members of the individual's immediate family, who in theory carry the same gene. By ensuring that any loss to the individual is balanced by gain to close relatives the gene can ensure its continued propagation within the population. Another possible mechanism is heterozygote advantage; being a carrier for the condition is more advantageous than either lacking the gene or being homozygous for it. Such a pattern can be observed in the high incidence of the recessive sickle-cell anaemia gene in populations exposed to malaria, where the heterozygote enjoys reduced susceptibility to the disease. (Connor et al, 1993) This explanation allows for a continual significant rate for homosexuality in the population despite its reduced fitness. A complementary theory, (Seaborg, 1984), rests upon the observation that humans have the most complex social behaviour of any animal. Little if any of our specific behaviour is genetically determined, since our social structures and behaviour shows great variation. Complexity and adaptability are advantages in human society, and homosexuality can be said to show plasticity and flexibility in behaviour. Homosexuality is thus a desirable trait selected for in the population; however, this does not explain how these desirable homosexuals manage to propagate themselves and their desirable genes. Heterozygote advantage may be the explanation but there is no behavioural or genetic evidence for such a continuum of sexual orientation. Gallup and Suarez (1983) propose that homosexuality is an outlet for the different sexual strategies of men and women. They observe that male and female homosexuals behave in a manner predicted by current evolutionary thought on optimal sexual strategies; male homosexuals are frequently promiscuous and females frequently form close and stable monogamous couples, as predicted as the optimal heterosexual strategy. With the difficulty of finding suitable heterosexual partners fulfilling the criteria for a good mate, the opposite sex behaving in a very different manner to one's own, individuals may satisfy their desires to fulfil their innate drives with members of their own sex. Males may find release in casual homosexual sex rather than vainly pursuing coy and demanding females, and females may find the supportive and committed mates they desire in homosexual relationships. Dickemann (1993) similarly proposes that the necessary plasticity in human behaviour is utilised in societies where resources are scarce to reduce demand for mates and children by conditioning for homosexuals. These hypotheses suggest that though homosexuality has a basis in the different mating strategies employed by the different sexes, dictated by the genetic needs of their respective biologies, it represents more of an individual tactic to fulfil gentically-determined behaviour patterns rather than a successful reproductive strategy, and as such it does not provide a coherent explanation for a genetic influence for homosexual as opposed to heterosexual orientation. Neither suggests a mechanism for development of an individual committed to such behaviour, detrimental to the individual himself. All these theories suffer from a lack of conclusive evidence, either a demonstration of a mechanism for affecting behaviour or a measurement of phenotypic fitness arising from different mating or behaviour strategies in humans, and must be regarded as mere suppositions.

Should it finally be demonstrated beyond reasonable doubt that homosexuality has a significant genetic component, what are the implications? The fact that a behaviour is determined by innate characteristics is repugnant to some, but need not necessarily imply any diminution of individual choice. A model of the development of sexual orientation where one is completely a product of society is no less restrictive. Some might express sadness that the majority of the population is so constrained by biology as to be unalterably heterosexual. Without a moral judgement of different behaviours, it is impossible to judge between them if no undesirable consequences necessarily follow. In today's society, homosexuality is penalised socially, but that is a matter of the heterosexual majority's attitudes, not an unavoidable result of homosexuality. More worrying are the measures that may be taken by parents unwilling to have homosexual offspring. Genetics will provide them with the necessary technology to prevent any homosexual children from seeing the light of day. Only a change in society can prevent this scenario, and the omens are not good. Being female is alleged to be the most probable cause of being aborted in some societies, notably India and China, where the low status of women coupled with screening and abortion technologies is resulting in a huge imbalance in the number of male and female births. The use of modern genetic technology to select the characteristics of one's children must be discussed now, before the decisions are made by default by parents and doctors without the proper discussion worthy of such a complex problem. Despite trepidation about the results of homosexuality research, refusing to study the subject at all is not a solution. Problems can only be solved when all the facts are known, and investigating a genetic hypothesis for determination of sexual orientation is a vital way to obtain these facts and perhaps gain some insight into why we are as we are. If homosexuals can say that they are as nature intended, as God made them it might be said, then although the roots of prejudice are not challenged, one might optimistically hope that a beginning to the acceptance of individuals committed to a life-style different to one's own may have been made. The psychological and social theories of homosexuality have not led to increased tolerance, but to hormone treatment and electro-convulsive therapy. The risks of a genetic explanation are greater, but so is the promise.

Modern genetic techniques appear to be elucidating some of the aetiological factors involved in the determination of sexual orientation. A significant degree of genetic influence in homosexuality is suggested by recent research, and modern evolutionary theory provides some theoretical underpinnings for this counter-intuitive concept. However, much more evidence is required, not least to justify any of the evolutionary theories. It seems unlikely that such a complex problem, involving so many of our human assumptions and preconceptions about human nature, will be resolved without a full understanding of all the genetic controls on behaviour and neurology, controls as yet ill-understood. The burden of proof in such a complex science is great. It is certainly greater than what the evidence yet collected can bear, and the genetic hypothesis must be regarded as remaining not proven.