Wednesday, March 21, 2007

Blue-eyed babies & the evolution of light skin   posted by Razib @ 3/21/2007 12:35:00 AM
Share/Bookmark

Over at her website Judith Rich Harris has posted her article, Parental Selection: A Third Selection Process in the Evolution of Human Hairlessness and Skin Color. When I asked Judy 10 questions I expressed some skepticism about this theory:
4) In your 2005 response to the Edge Question, "What Do You Believe Is True Even Though You Cannot Prove It?," you alluded to two things, 1) selection for light skin 2) hairlessness by parents in infants. When you pointed to these facts, did you do so in light of recent genetic work which suggests that dark skin might have evolved in humans as a response to loss of body hair? In other words, one trait would never been selected for if not for the other.

No, I hadn't heard of that work. But it doesn't matter. All humans have more or less hairless bodies, so I assume that the characteristic of hairlessness is at least as old as our species - at least 100,000 to 200,000 years old. Racial differences in skin color, on the other hand, are no more than 50,000 years old. If humans turned dark-skinned as a response to hairlessness (a theory I find dubious), then an explanation is still needed for why their skin turned white again so quickly when they inhabited Northern Europe, thousands of years later. My response to the 2005 Edge question offered a possible explanation.

By the way, I've expanded that essay into an article for a journal called Medical Hypotheses. It will be published in a few weeks.


Here is the abstract from the artice:
It is proposed that human hairlessness, and the pale skin seen in modern Europeans and Asians, are not the results of Darwinian selection; these attributes provide no survival benefits. They are instead the results of sexual selection combined with a third, previously unrecognized, process: parental selection. The use of infanticide as a method of birth control in premodern societies gave parents - in particular, mothers - the power to exert an influence on the course of human evolution by deciding whether to keep or abandon a newborn infant. If such a decision was made before the infant was born, it could be overturned in the positive direction if the infant was particularly beautiful - that is, if the infant conformed to the standards of beauty prescribed by the mother's culture. It could be overturned in the negative direction if the infant failed to meet those standards. Thus, human hairlessness and pale skin could have resulted in part from cultural preferences expressed as decisions made by women immediately after childbirth.



First, on a pedantic note, let me lodge my general protest in regards to the assumed decomposition of natural, sexual, and, Judy's putative third factor, parental selection. I'm reading some cognitive psychological work now about categorization and one thing that struck me as very apt in regards to how humans conceptualize the world is that we're always trying to break nature apart its joints, so to speak. Most regular readers know I'm a fan of consilience, and so I must reiterate and be a nag about the fact that many dynamic selective processes may bound in nature, but fundamentally they're all of a piece. As J.B.S. Haldane said: fitness is a bugger, naming selection is easy, characterizing it both accurately and precisely can often be very hard. The gene is the unit of selection, the various levels and dimension are all equal under God's eye.

In any case, to the thrust of the hypothesis. First, in the past few years a lot of work has been done on skin color. Last year I observed that Armand Leroi's afterward to Mutants where he notes that we don't understand skin color will need a coda in future editions. Here's what we know, so far....

1) In the 1960s human geneticists using classical pedigree analysis determined that 4-5 loci, genes, explained most of the intergroup variation between blacks and whites in regards to skin color.

2) In the past few years genomics has generally confirmed this view, there are a few loci of large effect (e.g., SLC24A5 in Europeans vs. non-Europeans) which explain intergroup differences in complexion.

3) But, equivalent phenotypic values can be attained via alternative genetic architectures, and this seems to have happened. In other words, light-skinned northeast Asians are not necessarily light for the same reason that Europeans are. Even if the change occurs upon the same locus, the allele or haplotype may be different.

4) Different evolutionary dynamics might affect the various loci. For example, in Europeans MC1R is highly polymorphic, suggesting either diversifying selection (e.g., frequency dependence) or a deep coalescence time (perhaps MC1R built up a great deal of variation during the residence of hominid lineages in Europe and modern humans simply assimilated the local depigmentation alleles?).

5) Also, some of these loci under selection seem to be relatively recent (e.g., perhaps within the last 10,000 years). Like LCT (lactase persistence) they leave a powerful imprint on the genome via a selective sweep.

How does this square with the hypothesis presented above? First, some prelims. Selection of all sorts can be hard to get a grip on. After all, we bandy about selection coefficients, s, in a vacuum, when they vary within their environmental contexts. Environment in this case can mean the natural world, the social world and other genetic parameters (via genetic interactions). Negative frequency dependence also throws a monkey wrench into these processes by making the selection coefficient a function of the frequency of the traits. But we need to start from simple models and build up in complexity.

The example of the !Kung woman who did not want to kill her light-skinned daughter is illustrative of Judy's hypothesis, but, it is simply a starting point. Nevertheless, I think it highlights a weakness: the genomic data is shedding light on the possibility that selection for loci which cause light skin (or, more properly explain a proportion of the intergroup variance) occurred long after the first humans settled the temperate zone. If the parental preference for light skin (which derives from the deep seated sensory bias which is also the root of sexual selection) existed prior to the arrival in the northern latitudes why is it that Eurasian populations seem to exhibit pulses of selection relatively late in history? One could make the argument, assuming that parental and sexual selection were paramount, that child and mate choice were simply not operative prior to this time period. Sexual selection works ideally through polygynous mating systems where there is a great deal of reproductive skew. Peter Frost has argued that blondeness in European females emerged through a form of sexual selection where males selected from a finite sample space of females because of the nature of the low latitude tundra, but the operative principle is the same, selection upon heritable variation. Perhaps within the last 10,000 years large boom and bust cycles in populations through the World Island has resulted in truncation selection events which reshaped the genomes (and generated high selection coefficients who show up in the long haplotypes)? I really don't know, but, I think one must say that it is more complex and contingent than a simple relaxation of functional constraint and the operation of innate preferences.

Additionally, I don't see much exposition of the details of sexual selection theory in Judy's article. For example, runaway sexual selection occurs very fast, and is often quickly checked by functional constraints. I believe that her hypothesis about light skinned preference being very deep implies some sort of sensory bias.

The intersexual difference in coloration is obviously real. There are biological reasons for this (hormone levels), but the cultural amplifiers of this tendency are rather universal. On this point the hypothesis is on strong ground, civilizations throughout the world seem to value female (relative) paleness. But, I think this point still goes back to my previous issue in regards to time scale: is this a phenomenon of the mass societies?

In any case, I'll leave it at that. There's a lot to chew on here. My main point of contention with this hypothesis is that I think the time depth is off. The most current results out of genomics suggest repeated and independent evolutionary sweeps in northern Eurasia at various times generating the phenotype of light skin. If the preference for light skin is deep within our natures, even predating bipedalism, it seems that it should have manifested immediately with the move of H. sapiens sapiens to the northern latitudes 30,000 years ago.

Labels: , , ,