Let's put it this way: homosexuality exists in the natural world. Aside from gay penguins in the Central Park zoo, certain species of lizards have been known to have female-female sexual contact, male big horn sheep have all-male orgies, et cetera. But why should such homosexual behavior continue to exist, given that those animals that only perform homosexual sex will not pass down their genes?
A simple answer is that whatever "gay gene" or "gay genes" that may exist (I'm personally not the kind of person who believes in a "gay gene", but go with it for now) may be an advantageous trait when heterozygous, even if it is deleterious when homozygously recessive. The advantage that those possessing one copy of the gene coding for sickle cell disease have among their non-heterozygote countrymen is the classic example here. (See also the Wikipedia entry for heterozygote advantage.) In other words, if there is an advantage to having one copy of the gene, statistically-speaking it'll keep showing up even if it is harmful when it shows up in two copies -- this is one of the consequences of being a diploid organism.
More to the point, and neglecting the dubious "gay gene" concept, homosexual individuals provide "child-raisers", in social animals, without adding to the population of children to be raised. Hence they could very easily simply pass their genes along through their cousins, so long as familial groups coincide neatly with social groups.
All this is fairly old hat. Where Roughgarden goes is different:
According to Roughgarden, gayness is a necessary side effect of getting along. Homosexuality evolved in tandem with vertebrate societies, in which a motley group of individuals has to either live together or die alone. In fact, Roughgarden even argues that homosexuality is a defining feature of advanced animal communities, which require communal bonds in order to function. "The more complex and sophisticated a social system is," she writes, "the more likely it is to have homosexuality intermixed with heterosexuality."
Japanese macaques, an old world primate, illustrate this principle perfectly. Macaque society revolves around females, who form intricate dominance hierarchies within a given group. Males are transient. To help maintain the necessary social networks, female macaques engage in rampant lesbianism. These friendly copulations, which can last up to four days, form the bedrock of macaque society, preventing unnecessary violence and aggression. Females that sleep together will even defend each other from the unwanted advances of male macaques. In fact, behavioral scientist Paul Vasey has found that females will choose to mate with another female, as opposed to a horny male, 92.5% of the time. While this lesbianism probably decreases reproductive success for macaques in the short term, in the long run it is clearly beneficial for the species, since it fosters social stability. "Same-sex sexuality is just another way of maintaining physical intimacy," Roughgarden says. "It's like grooming, except we have lots of pleasure neurons in our genitals. When animals exhibit homosexual behavior, they are just using their genitals for a socially significant purpose."
In other words, homosexual behavior is to be expected in any social animal, and acts as a sort of "social lubricant" that keeps groups together and happy. And in the case of those Japanese macaques, this may in fact be true, but to generalize to the whole of homosexual behavior from these examples is a bit premature, in my book. I agree strongly with PZ Myers, quoted at the end of the article: "I think much of what Roughgarden says is very interesting. But I think she discounts many of the modifications that have been made to sexual selection since Darwin originally proposed it. So in that sense, her Darwin is a straw man. You don't have to dismiss the modern version of sexual selection in order to explain social bonding or homosexuality." (I'm sure he's happy to know that I agree.)
What's interesting is where Roughgarden sees human sexuality in all this.
"In our culture, we assume that there is a straight-gay binary, and that you are either one or the other. But if you look at vertebrates, that just isn't the case. You will almost never find animals or primates that are exclusively gay. Other human cultures show the same thing." Since Roughgarden believes that the hetero/homo distinction is a purely cultural creation, and not a fact of biology, she thinks it is only a matter of time before we return to the standard primate model. "I'm convinced that in 50 years, the gay-straight dichotomy will dissolve. I think it just takes too much social energy to preserve. All this campy, flamboyant behavior: It's just such hard work."
While I'd agree that human society (at least in the Western world today) largely uses sex as a social bonding ritual and as entertainment, only reproducing a few times in a lifetime filled with sexual encounters, I'd argue that the barrier between heterosexuality and homosexuality is a bit more stringent that Roughgarden believes. While there are many bisexual individuals in the human species, the vast majority of us are either heterosexual or homosexual, and there's some pretty good science supporting the idea that it's inborn, at least to some degree. Roughgarden's sexual experiences may have enlightened her to the idea that perhaps same-sex interactions serve social purpose in social animals, but it seems to have also blinkered her to the more common feelings of sexual desire: homosexual or heterosexual describe our ontogeny, who we are, and not what we do.
I'm keeping my eyes open on this one, though. It's a fascinating issue and one that's not likely to be solved by mostly-ignorant bloggers like me.