Jared Diamond has a short review article in Nature on the evolution of skin color. This is an old story in anthropology, but it has taken some interesting twists lately with the assessment of genetic variability in some of the genes related to pigmentation. This short review covers recent papers by Nina Jablonski (2004) and George Chaplin (2004). After a short history of how skin reflectance has been measured, Diamond cuts to the selective point:
Jablonski and Chaplin prefer a combination of two selective factors involving several costs and one benefit of UVR [ultraviolet radiation]. The costs involve the destructive photolysis of many compounds, of which Jablonski and Chaplin attach particular importance to the B vitamin folate. Everybody requires folate, so everybody would have dark skins (to screen out UVR and reduce photolysis) if there were no other selective factors. However, UVR also provides a benefit: catalysing the synthesis of vitamin D. Hence skin colour evolves as a compromise between skins light enough to permit UVR penetration for vitamin D synthesis, but dark enough to reduce folate photolysis (283).
The folate explanation is probably the most distinctive aspect of Jablonski and Chaplin's take on skin coloration, and Diamond naturally wonders what happened to the hypothesis that skin cancer was a primary selective factor:
Although the harmful effects of UVR are often taken to be especially associated with skin cancer, Jablonski and Chaplin minimize the importance of this aspect on the grounds of the late age of onset, after some or most of a couple's children have been morn. Most geneticists similarly minimize the selective importance of damage or disease late in life.
Here I am sceptical. This common view may be valid for mammal species in which parent-offspring ties are severed soon after weaning, but it overlooks three of the most distinctive features of human life-histories in traditional societies: the long post-weaning dependence of offspring on parents for learning and then for social status, the large contribution of hunter-gatherer grandmothers to their grandchildren's food supply, and the dependence of an entire band or village on its oldest people as the repositories of knowledge in a preliterate society. It would be interesting to try to estimate the selective importance of skin cancers for skin-colour evolution from this perspective... (283).
Before my reaction, it should be noted that the great importance of folate in the hypothesis of Jablonski and Chaplin is also a distinctive character of our species. Since folate is of paramount importance for normal neural growth during fetal development, the uniquely large and complex brains of humans are a greater target of selection to avoid folate photolysis.
But Diamond's objection about skin cancer is also interesting in light of recent genomic findings concerning targets of positive selection in human evolution. Nielsen et al. (2005) (reviewed in a previous post) found that several of the genes showing the strongest evidence for positive selection in the human lineage were related to cancer risk.
I think it likely that these genes are positively selected in order to reduce cancer risk in humans, because of our uniquely long lifespans compared to most primates, and because of the great importance of older individuals to the survival and reproduction of their younger kin. It is possible that some of these genes are particularly selected because of skin cancer itself, since the human adaptation to resist skin cancer must have been initiated as hominids lost their body fur. This time was likely substantially earlier than either any great increase in longevity or any great increase in the importance of older adults. For this reason, skin cancer risk may have been one of the earliest factors leading to selection on genes associated with tumor growth and other aspects of cancer etiology.
Diamond also discusses exceptions to the pattern of correlation of skin color with UVR intensity:
Of the 12 most negative residuals (skins unexpectedly dark), the highest and four others are for Bantu-speaking southern African populations that migrated from the Equator towards high southern latitudes only 2000 years ago. These populations have not yet had enough time for evolutionary loss of their equatorially adapted dark skins. Conversely, three of the nine most positive residuals (skins unexpectedly pale) are for peoples of the Philippines, Vietnam and Cambodia, who migrated towards the equator from high latitudes only in recent millennia and have not yet evolved appropriately dark skins. But why do the people of Bougainville Island in the South Pacific, and why did the Aboriginal Tasmanians, have such dark skins, after more than 10,000 years of in situ adaptation? (284)
An answer to this question might be found in the nature of genetic alterations that lead to pigmentation levels in humans. Diamond's short review does not discuss this research, but an entrée may be found in an article on MC1R variation by Harding et al. (2000). The structure of global variation in this gene shows a relatively low level of allelic diversity in Africans and a higher level outside of Africa. The Eurasian population in particular has a high number of unrelated alleles, with many of the alleles functionally related to lighter skin tones and hair colors. The apparent pattern for this gene is a strong selective constraint within Africa for a functionally dark pigmentation, with a contrasting pattern outside Africa. This contrast may be based on a relative relaxation of selection at higher latitudes, or conversely it may reflect positive selection for many different mutations, all of which result in lighter skin color. In either case, the striking aspect of the pattern is the many ways in which skin color may be light, corresponding to the spread of several distinct mutations in the non-African human population.
When considering small island populations, a natural hypothesis is that the necessary mutations to lighten skin color were rare enough that they may not have occurred. But the range of different mutations effectively resulting in lighter skin (if the variation of MC1R may provide a model for at least some of the genes related to skin color) tends to make me think that adaptive mutations (resulting in lighter skin) were not unlikely. For this reason, I think we must turn to the level of selection on skin color, which was either absolutely low in these populations or relatively low compared to the rate of genetic drift.
So Diamond's question is an important one, because it essentially asks why selection upon skin color should be high in some populations and low in others. This variable nature of selection in different human populations is a venue in which we may see many more similar questions in the future.
Chaplin G. 2004. Geographic distribution of environmental factors influencing human skin coloration. Am J Phys Anthropol 125:292-302. PubMed
Diamond J. 2005. Geography and skin color. Nature 435:283-284. Nature online
Harding RM et al. 2000. Evidence for variable selective pressures at MC1R. Am J Hum Genet 66:1351-1361.
Jablonski N. 2004. The evolution of human skin and skin color. Annu Rev Anthropol 33:585-623.
Nielsen R, et al.. 2005. A Scan for Positively Selected Genes in the Genomes of Humans and Chimpanzees. PLoS Biol 3:e170. PLoS Online