Wednesday, September 24, 2014

Gender Differences In Effective Population Size

Genetic evidence shows that the effective population of men in prehistoric populations was smaller pretty much across the board than the effective population size of women in those populations, although the extent of the discrepancy varied significantly in different places and times.

While I agree that this likely involves men having children with more than one woman in their life, the press release connected to the study confounds the notion of census population, which is boy heavy for children, about equal at sexual maturity, and woman heavy at older ages in modern populations, with effective population size, which is typically several times smaller than census population and has a different technical definition. In particular, effective population size is increased relative to census population when each parent has the same number of children, and is decreased when the number of children per parent is highly varied.

The abstract of the open access paper reads as follows (in selected parts):
Our results confirm the controversial assertion that genetic differences between human populations on a global scale are bigger for the NRY than for mtDNA, although the differences are not as large as previously suggested. More importantly, we find substantial regional variation in patterns of mtDNA versus NRY variation. Model-based simulations indicate very small ancestral effective population sizes (less than 100) for the out-of-Africa migration as well as for many human populations. We also find that the ratio of female effective population size to male effective population size (Nf/Nm) has been greater than one throughout the history of modern humans, and has recently increased due to faster growth in Nf than Nm.
Later on, the body of the study elaborates on the ancestral effective population sizes and notes that:
The simulation results suggest a small founding size in Africa of about 60 females and 30 males (all population sizes are effective population sizes); migration out of Africa about 75 kya associated with a bottleneck of around 25 females and 15 males; migrations from this non-African founding population to Oceania 61 kya, to Europe 49 kya, to Central and East Asia 37 kya, and from East Asia to the Americas about 15 kya.
Thus, all humans can be traced to a founding population with an effective size of only about 90 individuals, and all non-Africans derived from a subset a population with an effective size of about 40 ancestral individuals. Even if a rule of thumb factor of three was used to convert effective population sizes to census population sizes, in each case were are talking about huge populations derived from tiny villages at quite specific dates in the past.

If Neanderthal admixture, for example, took place at around the time of this minimum sized ancestral Out of African population, the number of instances of Neanderthal admixture that accounts for the lion's share of Neanderthal ancestry in modern humans today could have arisen from just a couple of human-Neanderthal hybrid births in all of human history at just the right time and place.

The study also notes that some substantial regional subsets of Native American populations have substantially less mtDNA diversity than the New World as a whole.

The likely scenario is not that there were substantially fewer men than women outside isolated moments in the aftermath of wars and the like, although bad hunts in hunter-gatherer societies may have served a similar function, but that a higher percentage of men than of women were excluded from reproduction entirely.

For example, if women, on average have roughly the same number of children per lifetime, but some men have only one or two, while other many have many children, the effective population size of men will be smaller than the effective population size of women, even if on a census basis there are the same number of men and women with children.  This could happen, for example, if men were equally likely to have a first wife, but if high status widowers were more likely to remarry upon a wife's death in childbirth, than low status widowers.

Also, my analysis earlier this year at sister blog Wash Park Prophet, of the way that Southern Civil War widows coped historically with the loss of their men, suggests that the capacity of serial monogamy, rather than polygny to cope with imbalances between the number of women and the number of men in a population is greatly underestimated. In that situation:

About 18% of white men in the South aged 13 to 43 died in the American Civil War (1861-1865) and Reconstruction, and many more were crippled as a result, for example, with amputated limbs. Yet, by 1890, there was no excess in the number of women never married in the South.
The way that people did cope with this imbalance was complex, but de facto polygamy was very rarely the solution adopted.

I suspect that remarriage by men following the losses of their wives in childbirth may have been an important source of the enduring nature of this imbalance across so many different eras and cultures of prehistory.  Widowers outnumbered widows during the colonial and revolutionary war eras of the United States, for example.

3 comments:

Maju said...

... "the effective population of women in prehistoric populations was smaller pretty much across the board than the effective population size of men in those populations"...

There seems to be a critical typo or misunderstanding in this sentence, Andrew: it should read LARGER, not "smaller". Not only that's what the simulations' results of the mentioned study strongly suggest (although with important variations in time and space) but also what allows your later derived statement implying polygyny.

Maju said...

"If Neanderthal admixture, for example, took place at around the time of this minimum sized ancestral Out of African population, the number of instances of Neanderthal admixture that accounts for the lion's share of Neanderthal ancestry in modern humans today could have arisen from just a couple of human-Neanderthal hybrid births in all of human history at just the right time and place".

Yes but only if we consider only those who left a significant legacy. A problem I see with "effective population size" is that, because of whatever random or fitness factors, different individuals could have left more or less legacy in the long run and the legacy of many could well have been cut at some point long after their lifetime.

This is particularly true if we judge on haploid lineages, as the study does. Patri- and matrilineages can come extinct with relative ease, while a good deal of that genetic legacy of people whose haploid lineages went extinct can still be around. In fact the Neanderthal legacy is a good example of that: there are no Neanderthal matri- or patrilineages alive anymore but their overall genetic legacy in the OoA-derived population is still quite significant (~2.4%)

So overall the effective population size was necessarily quite bigger than what can be inferred from surviving haploid lineages.

"some substantial regional subsets of Native American populations have substantially less mtDNA diversity than the [pre-colonial] New World as a whole".

That just makes good common sense: the parts cannot have more diversity than the whole ever. Some parts can perfectly have much less diversity because of whatever "bottlenecks" involved in their formation - but in general all parts should have less diversity than the whole.

"This could happen, for example, if men were equally likely to have a first wife, but if high status widowers were more likely to remarry upon a wife's death in childbirth, than low status widowers".

Polygamy, especially polygyny but sometimes also all kind of multiple marriages, seems to have a high prevalence in the anthropological data, also in Native America. I see no reason to assume serial monogamy as default, although it is true that some societies favor that model instead.

Your example from a Christian Industrial society doesn't seem to be of any relevance. I would go to the anthropological literature in order to find out.

In this sense I am under the impression from this study that the gender imbalance grew dramatically with Neolithic and Metals, and it is probably true that in those "Ages of Farming", certain men could effectively displace others based on their economic and socio-political power, adopting many wives and concubines.

andrew said...

Critical typo fixed. Thanks for catching that. It sucks not to have a copy editor.

The fact that haploid lineages are prone to noise and lineage loss is figured into the analysis when using the total modern haploid lineage mix to calculate effective ancestral population size, which eliminates this systemic bias. You can mathematically infer how many lineages you lost along the way from the mix that you have today within reasonable margins of error. While your haploid lineage comprises only a small percentage of your total ancestry, the collective haploid lineages of the entire population provides a data set that can accurately measure ancestral population sizes and compare them differentially by gender. This is a case where the macro-analysis is qualitatively different than the micro-analysis.

Polygamy certain has existed in many cultures (many daughter's big history project this year involves a non-Mormon white man who had three simultaneous Native American wives (who were sisters) in 18th century Colorado and then later married a white woman from the Eastern U.S. after his first wife die). But, it is easy to overestimate the extent to which polygamy is necessarily a major driver of effective population size imbalance.