THE TRIVERS-WILLARD EFFECT
Dan Willard was a graduate student in mathematics at Harvard University, and he wanted to meet some women. There were very few women in mathematics at Harvard in those days (1970), so Dan started to prowl the halls in search of courses with more agreeable sex ratios. He soon found one in Professor DeVore's renowned primate behavior course. This course attracted almost three hundred students each year, about two-thirds of them women. Dan developed an interest in primate behavior. I was a graduate teaching assistant in Professor DeVore's course, and my duties included giving a few lectures to the class. One lecture was on the sex ratio at birth, and another one was on the logic of mate choice. Dan Willard put pieces of the two lectures together and came up with a brand new argument for adaptive variation in the sex ratio at birth.
In the lecture on the sex ratio at birth, I explained to the class Fisher's sex ratio theory, which, under outbreeding, resulted in 50/50 sex ratios. This is because any deviations from a 1: 1 sex ratio are unstable: whichever sex is produced in greater numbers becomes less valuable to its parents than offspring of the opposite sex, because it has a lower chance of reproducing. For example, an excess of males means that each male has less than one female on average to fertilize. I emphasized that as long as sex ratios canceled out across a population to give an overall sex ratio of 1: 1, each was equally adaptive. Thus, a 5: 1 sex ratio might cancel out with a 1: 5 sex ratio in a population of individuals otherwise tending to produce 1: 1 sex ratios, and 1:5, 5:1, and 1:1 would be equally adaptive. This turned out to be a key point in Willard's argument.
In the lecture on mate choice I had noted that in the human species, a more or less monogamous one, there was evidence that women tended on average to marry up the socioeconomic scale. A consequence of this bias