Sexual selection on the American frontier


Jason Collins


November 27, 2013

It seems obvious that having multiple wives is a good thing for the fitness of a man. Similarly, having the women in a population monopolised by a small number of men is not good for the fitness of those men who miss out on a mate. In such a society, the large difference in fitness between the haves and the have-nots would be expected to result in strong sexual selection.

Having noted the obvious, an article in Evolution and Human Behavior by Moorad and colleagues presents an interesting illustration of this situation. They examined the strength of sexual selection in a population of Utah men born between 1830 and 1894, with the rate of polygamy among married men dropping from over 17 per cent among some year groups born in the 1930s to less than 1 per cent among those born at the latest dates. These men faced a ban on polygamy in 1862, among other increasing social pressures against multiple marriages. The authors’ analysis of the population data allowed them to estimate the strength of sexual selection over this period and to isolate which factors contributed to reproductive success.

The headline finding from the study is that between 1830 and 1894, the strength of sexual selection in this population dropped by 58 per cent. This authors calculated this reduction from changes in Crow’s Index, which sets an upper limit on the rate of evolutionary change. Crow’s index is an upper limit as, first, not all differences in fitness may be due to phenotype.  An environmental or random cause may be relevant. Second, phenotypic selection is only genetic to the extent that the genotype associated with the higher fitness is passed to the next generation.

Moorad and colleagues do not attempt to tease out the phenotypic or genetic selection in this sample. I am not sure how they could. However, it is reasonable to assume that changes in total selection are closely related to changes in the underlying level of genotypic and phenotypic selection.

The authors also made estimates of the costs and benefits of the polygamous mating system for men and women. They did this by examining the Bateman gradients, which are a measure of the change in reproductive success for a given change in mating success. In the 1830s, the Bateman gradient for men was 5.87, meaning that for each extra mate a male could expect almost six extra offspring. By the 1890s, this had dropped to 1.92. For women the gradient was slightly positive, increasing from 0.195 to 0.671 over the course of the study period.

The authors made a more interesting use of the Bateman gradient when they determined the gradient for women for each additional mate that a man has. This comes back to the classic trade-off question - do the reduced resources available to a woman from having to share the male’s resources with more wives outweigh the potentially higher quality of the man that a woman can obtain in the polygamous system? The answer to this question was yes (at least in terms of number of children) - there was a slightly negative gradient in the 1830s (starting at -0.06), which dropped to a low of -1.36, before ending at -1.11 in the 1890s. In the 1890s, each additional wife resulted in the other wives having, on average, one child less. The woman would want to hope that they are particularly high quality children.

A further piece of information that Moorad and colleagues teased out was how much a man’s increased fitness, due to additional wives, comes from increased reproductive rate or from a lengthened reproductive tenure. At the beginning of the sample, increasing the number of wives increased both the rate and tenure of reproduction. This switched towards the 1890s however, with increased wives only extending the tenure and coming at the cost of the rate of reproduction. This reflects the fact that towards the end of the sample, additional wives were normally the product of serial monogamy. For the few polygamists around at that stage, they still had the benefit of both an increased reproductive rate and tenure.

I enjoy studies like this. Even though they sometimes seem to be demonstrating the obvious, an empirical illustration adds some colour and robustness to the theory. Given my research interests around the economic consequences of evolutionary change in humans, illustrations of the rate of evolution in human populations are always useful. Could change genetically in a time period short enough to have economic significance. This study does not answer this, but it is another useful example.