Special Issue on Sexual Differentiation of Sexual Behavior and Its Orientation.
This has a wealth of articles on the subject, many dealing with animal experiments, but also dealing with the issues in humans too. From the editorial:
In a seminal 1959 paper, William C. Young and his colleagues at the University of Kansas reported that prenatal administration of testosterone to female guinea pigs masculinized their capacity for sexual behavior in adulthood. They concluded that prenatal androgens have an ‘organizing’ effect on the brain circuitry responsible for sexual behavior: high levels of testosterone, such as are typically seen in male fetuses, promote the establishment of neural elements mediating male-typical behaviors, while low levels of androgens, typically seen in female fetuses, permit the establishment of elements mediating female-typical behaviors. The role of androgens and estrogens in adult life is simply to activate these pre-established circuits.
The ‘organizational hypothesis’ became one of the most fruitful ideas in psychobiology. Subsequent research has confirmed the basic truth of the hypothesis and elucidated many of the anatomical structures and molecular processes that underlie it. Yet this research has also highlighted complexities, limitations, exceptions, and species differences that Young (who died in 1966) could hardly have guessed at. The present set of contributions has been selected to highlight some of these complexities, but also to consider the relevance of the organizational hypothesis to two aspects of human sexuality: sexual orientation and gender identity.
 C.H. Phoenix, R.W. Goy, A.A. Gerall, W.C. Young, Organizing action of prenatally administered testosterone propionate on the tissues mediating mating behavior in the female guinea pig, Endocrinology 65 (1959) 369–382.
A greater part of this issue is on sexual orientation, and the biological basis thereof:
Melissa Hines provides a masterful survey of the evidence for an influence of prenatal gonadal steroids on human sexual orientation.Gender Identity gets relatively scant mention in the editorial, though there are several good papers in the issue.
A key observation here is that homosexuality is not an isolated trait; rather, it tends to be associated with other gendervariant cognitive and personality traits, both in childhood and in adult life. This is analogous to what has been observed in rodents and other animals subjected to prenatal manipulations of gonadal steroids, and it suggests that atypical levels of these hormones may affect a constellation of gendered traits, including sexual orientation, because many such traits are mediated by hormone-sensitive brain circuits.
Still, gay men and lesbians are not transexuals or complete gender ‘inverts’; rather, they seem to be a patchwork of gender-conformist and gender-nonconformist traits—a patchwork that varies to some extent from individual to individual. Underlying this patchwork may be differences in developmental timing between different brain systems, differences in their sensitivity to gonadal steroids, or differences in causal mechanisms (e.g., hormonal versus direct genetic effects).
It would be unethical to perform in humans the kind of animal experiments that led to the organizational hypothesis and its many subsequent ramifications, but it is possible to approach the same question in less direct ways. One approach is to take advantage of experiments of nature, such as the genetic condition congenital adrenal hyperplasia (CAH), which exposes female fetuses to higher-than-normal levels of androgens. CAH girls do display several gender-atypical childhood traits, as documented by Hines and others, and are more likely than other girls to develop same-sex attraction in adulthood. Still, even among women with the severest
form of CAH, about half are exclusively heterosexual and only a few are exclusively homosexual. Thus the CAH research supports a significant role for prenatal androgens but also leaves plenty of room for other potential factors to play a role, possibly including social ones.
Gender identity has been the subject of much less biological research than sexual orientation, but even in the absence of strong evidence there is a widespread supposition that transexuality has some biological basis. There have been reports of brain structural differences between transgendered and non-transgendered individuals of the same sex, as described by Bao and Swaab in this volume.It is... unfortunate... that such a junk-science book, largely composed of fiction and anecdote, without references or any of the hallmarks of a scientific publication, should be used as a source here. As bad as using "Moby Dick" as a textbook on cetacean biology, or one on Homeopathy to explain diabetes.
This area of study is made difficult by the relatively small numbers of transexual individuals, by the hormonal treatments that many of them undergo, and by the fact that male-to-female transexuals who are attracted to males and those who are attracted to females seem to have quite different developmental histories .
 J.M. Bailey, The Man Who would be Queen: the Science of Gender-bending and Transsexualism, Joseph Henry Press, 2003.
There's also, for the first time I can remember seeing, a mention of the Elephant in the Living Room: the climate that, in the USA at least, makes such research difficult.
As far as the political and social climate permits, researchers interested in the development of sexual orientation and gender identity would do well to take advantage of some of the large-scale longitudinal studies of childhood development now underway.Referencing Bailey's debunked book doesn't make me quite so optimistic, but the truth will out in the end. Reality always wins eventually.
As gay people become more accepted and integrated into mainstream society, public interest in the ‘‘why’’ of sexual orientation will likely wane. Yet this may be a good thing, because it will allow the study of this important aspect of human diversity to be studied in a less contentious and more genuinely scientific atmosphere than has typically been the case in the past. The contributions in the current collection point optimistically in that direction.