Wednesday, October 15, 2008

Polygyny or patrilocality?

Have all humans been more or less equally polygynous? The answer seems to be yes if we believe a team of researchers from the University of Arizona. They found that genetic diversity is higher on the maternally inherited X chromosome than on chromosomes inherited by both sexes (autosomes) in samples from five different populations: Biaka (Central African Republic), Mandenka (Senegal), San (Namibia), Basques (France), Han (China), and Melanesians (Papua New Guinea). Their conclusion: “our results point to a systematic difference between the sexes in the variance in reproductive success; namely, the widespread effects of polygyny in human populations.” In other words, proportionately more women than men have been contributing to the gene pool (Hammer et al., 2008).

It’s no surprise that polygyny has existed in the five populations under study. Almost all human populations are polygynous to some degree. The surprise is the relative lack of difference between the European and African subjects. Indeed, according to this study, the Basques have been more polygynous than the Mandenka have been. This is truly counterintuitive. Among the Basques, polygyny is normally limited to its serial form (marriage to a second wife upon the death of the first), as well as occasional cuckoldry. Among the Mandenka, polygyny is the preferred marriage type.

For some people in the blogosphere, this is simply scientific truth and we just have to accept it, however counterintuitive it may seem. There is nonetheless an alternate explanation: patrilocality. In many societies, a wife goes to live in her husband’s community after marriage. This has the effect of inflating the genetic diversity of women in any one community.

These two confounding levels of explanation, polygyny and patrilocality, bedeviled the previous methodology of comparing maternally inherited mtDNA with the paternally inherited Y chromosome. With the new methodology, patrilocality biases the results even more because the Y chromosome is no longer a point of reference.

The University of Arizona researchers do not mention patrilocality in their paper although they do discuss ‘sex-biased forces.’ Under this heading, they tested a model where only females migrate between communities (‘demes’) and at such a rate that panmixia eventually results. They concluded that this factor could not be significant. To my mind, the model is unrealistic, partly because the assumed migration rate is far too high and partly because two demes are used to represent a real world where brides are exchanged among many communities separated by varying genetic distances. To be specific, the more genetically different a bride is from her host community, the further away will be her community of origin, and the lower will be the probability of panmixia between the two.

To the extent that the methodology is biased toward patrilocality effects, any polygyny effects will be less apparent. If this new methodology primarily tracks differences in patrilocality, no major differences would be observable among the different population samples.

In addition, there may be a weak inverse relationship between patrilocality and polygyny. Patrilocality correlates with patriarchy, which correlates with high paternal investment, which inversely correlates with polygyny. If so, the two effects – polygyny and patrilocality – would tend to cancel each other out in the data.

Finally, the burden of proof is on those who propose new methodologies, especially one that produces inconsistent results. The University of Arizona researchers themselves say as much: “Our findings of high levels of diversity on the X chromosome relative to the autosomes are in marked contrast to results of previous studies in a wide range of species including humans.” More importantly, their findings run counter to the comparative literature on human mating systems. To cite only one authority, Pebley and Mbugua (1989) note:

In non-African societies in which polygyny is, or was, socially permissible, only a relatively small fraction of the population is in polygynous marriages. Chamie's (1986) analysis of data for Arab Muslim countries between the 1950s and 1980s shows that only 5 to 12 percent of men in these countries have more than one wife. … Smith and Kunz (1976) report that less than 10 percent of nineteenth-century American Mormon husbands were polygynists. By contrast, throughout most of southern West Africa and western Central Africa, as many as 20 to 50 percent of married men have more than one wife … The frequency is somewhat lower in East and South Africa, although 15 to 30 percent of husbands are reported to be polygynists in Kenya and Tanzania.


Hammer, M.F., Mendez, F.L., Cox, M.P., Woerner, A.E., & Wall, J.D. (2008). Sex-biased evolutionary forces shape genomic patterns of human diversity. PLoS Genet, 4(9), e1000202. doi:10.1371/journal.pgen.1000202

Pebley, A. R., & Mbugua, W. (1989). Polygyny and Fertility in Sub-Saharan Africa. In R. J. Lesthaeghe (ed.), Reproduction and Social Organization in Sub-Saharan Africa, Berkeley: University of California Press, pp. 338-364.


Tod said...

Any validity in the idea of a brief period of extreme polygamy in the early medieval period masking a largely monogamous past. Genghis Khan is said to be the ancestor of 0.5% of living men. One in five men in Northwest Ireland share a common ancestor.

Would the techniques used by Hammer et al be able to separate the enduring effects of such an overwhelming but transient hegemony derived from kingship from the practices of the vast majority over thousands of years?

ziel said...

Would presumably high levels of cuckoldry in tropical cultures also perhaps be an explanation?

Peter Frost said...


The short answer is 'no'. It's impossible to fully separate polygyny effects from patrilocality effects. It might be better to compare Y chromosome genetic diversity with autosomal genetic diversity. Any difference between that methodology and Hammer et al.'s would probably reflect patrilocality.


Not really. The puzzle is why the results were so similar between 'tropical' (West African, Melanesian) and 'non-tropical' (Basque, Han Chinese) subjects. Higher levels of cuckoldry in the first group would have created more of a difference

Tod said...

A frequent Y-chromosome b2/b3 subdeletion shows strong association with male infertility in Han-Chinese population.(Wu et al)
"Our data showed a higher frequency of deletion events in this Han-Chinese population than in populations elsewhere in the world"

Tod said...

Sex-Specific Genetic Structure and Social Organisation in Central Asia: Insights from a Multi-Locus Study. (Laure Segurel et al)

"We develop a new multi locus approach to jointly analyze autosomal and X-linked markers in order to aid the understanding of sex-specific contributions to population differentiation ... this study suggests that differences in sex-specific migration rates may not be the only cause of contrasting male and female differentiation in humans".

Anonymous said...

Patrilocality is certainly a problem, but so also is serial monogamy. A rich man will remarry when his wife dies. A poor man is happy to have gotten his one shot at marriage.

Women also had fewer chances at multiple fathers. During childbearing years, women were far more likely to die than men. It was only after they finish having children that their survival rates improved. At that point, a well-off woman could certainly remarry. In some cultures, remarriage of a young wife would be largely avoided as it could rob a woman of influence over the sons of her first marriage, which was her best shot at power in some cultures, in return for a theoretical but uncertain future son with the new husband. This was true from the beginning of recorded history into modern times pretty much all of East Asia and into the Middle East, with a strong divide with Jewish culture where wives were more important than mothers--only in a minority of cultures in Eurasia were wife-women more important than mother-women of powerful men.

Anonymous said...

Here's another thought, too. A woman's children are most likely to have kids of their own if her husband is high ranking. High ranking men are more likely to marry women from far away. With hereditary nobility, this leads to an incestuous nobility, but strict heredity of that type seems to be pretty recent. So the women from the farthest genetic distance are most likely to have surviving kids, while somewhat close relatives of the high-ranking man are also most likely to have kids who have kids.