Way back when, the alpha male would just mate with all the females in the group, and other males would only be able to do so behind his back, usually. Females would, to my knowledge, breed with whatever they could get. This is still seen in lots of primates - baboons are a perfect example. But humans don't do that. What happened (according to Owen Lovejoy) is that we developed "epigamic differentiation". Different people are attracted to different combinations of characteristics. The alpha male, then (if such a thing existed in human society) does not mate with all females, because he doesn't find them attractive. Instead, pairs crop up, of males and females who find each other attractive. These pair bonds are very important because they allow the female to stay in one place and mind her child, while the male will retrieve food for her. By doing this, the female can nurture more than one infant at a time. Chimps never evolved to do this, and for this and other reasons are dying out. This allowed human population to explode, basically. And it's supported by my (admittedly fairly scant) knowledge of the relevant neurobiology. Pair-bonding is mediated by certain neurochemicals, including oxytocin and vasopressin. Er, wait. I actually read a paper that dealt with it recently called 'What Does Sexual Orientation Orient' by Lisa M. Diamond. She acknowledges, as I now will, that most of the data about this kind of stuff is based on non-human animals, so I suppose the point is moot. The argument in the linked article seems to stem from strict evolutionary theory wherein organisms do whatever they can solely to reproduce, and whatever happens afterward tends to have little bearing on the evolutionary path that unfolds.