Understanding human races: the retreat of neutralism.
Discussion and debate about human races has been dominated for decades by neutral theory and statistics. Since this literature never posed a real question, it has never produced an answer. Lewontin's 1972 paper with its claim that a value of 1/8 of a statistic like Fst is “small” and that this means that human race differences are insignificant is a staple of our textbooks.
Recently geneticists have had a closer look and pointed out that Fst of 1/8 describes differences among sets of half sibs and few claim that half sibs are insignificantly related. Anthony Edwards has shown that the significance of differences is in the correlation structure of a large number of traits, again denying the Lewontin assertion that human differences are small. Alan Templeton in 1998 claimed that human races were less differentiated that races of some other large mammals, but he compared human nuclear DNA statistics with statistics from mtDNA in the other species. An appropriate comparison shows that human are more, not less, differentiated than other large mammal species.
Since neutral differences are a passive record of demographic history they are not very significant for issues of functional biology. Newly available data sources allow us to study the natural selection of race differences instead of their drift. It appears that there is a lot of ongoing evolution in our species and the loci under strong selection on different continents only partially overlap. Human race differences may be increasing rapidly.
I wrote about Harpending's change of mind about Lewontin's celebrated statistic in VDARE.com in 2004.
Acceleration of adaptive evolution in modern humans.
J. Hawks and G. Cochran
Humans vastly increased in numbers during the past 40,000 years. Recent surveys of human genomic variation have suggested a large surplus of recent positive selection, indicated by excess linkage disequilibrium and skewed SNP frequency spectra. We applied estimates of prehistoric and historic population sizes to estimate the importance of population growth in explaining the number of recent adaptive mutations. Our estimates are consistent with genomic evidence in suggesting that the rate of generation of positively selected genes has increased as much as a hundredfold during the past 40,000 years.
Do skeletal features reflect this genomic evidence of selection? Under positive selection, rapid appearance of new variants during the terminal Pleistocene and early Holocene would cause maximal phenotypic change during the last 2000-4000 years. We compared original and published series of Holocene cranial data from Europe, Jordan, Nubia, South Africa, and China, in addition to Late Pleistocene samples from Europe and West Asia, to test the hypothesis that the genomic acceleration in positive selection correlates with phenotypic evolution during this time period. A constellation of features in the face and cranial vault, notably including endocranial volume, changed globally during this time period and documents common patterns of selection in different regions. Holocene changes were similar in pattern and chronologically faster than those at the archaic-modern transition, which themselves were rapid compared to earlier hominid evolution. In genomic and craniometric terms, the origin of modern humans was a minor event compared to more recent evolutionary changes.
For example, here's the forensic reconstruction of the 14,000-year-old skull of a woman from Sicily (via Dienekes). What modern race would she be? She's obviously human, but doesn't look particularly like any large group around today. You run into this a lot with older skulls.
Admixture in Mexico City: implications for admixture mapping.
E. Cameron et al.
"The average proportions of Native American, European and West African admixture were estimated as 65%, 30% and 5% respectively."
"In a logistic model with higher educational status as dependent variable, the odds ratio for higher educational status associated with an increase from 0 to 1 in European admixture proportions was 9.4 (95% credible interval 3.8 – 22.6). This association of socioeconomic status with individual admixture proportion shows that genetic stratification in this population is paralleled, and possibly maintained, by socioeconomic stratification."
I'm not sure how to interpret this "odds ratio" but this certainly points in the direction I've been arguing since 2000.
Patterns of admixture in Mexican Americans assessed from 101,150 SNPs.
M.G. Hayes et al.
"No significant differences were observed between the 10 subsets, allowing us to average the admixture estimates across the subsets: 68% European, 27% Asian (as a proxy for Native American), and 6% African."
So, this is the reverse of the Mexico City data above. Are the populations different, or is admixture analysis still unreliable? The people at the illegal immigrant rally in Van Nuys, CA last spring were a lot shorter and more Indian-looking than the Mexican-Americans I grew up with, so perhaps the Mex-Am population is changing, with new immigrants being drawn from a more Indian background. Well, the African percentages in both are in line with earlier estimates of 3% to 8% that I cited in my 2002 article "Where Did Mexico's Blacks Go?"
Intracontinental Distribution of Haplotype Variation: Implications for Human Demographic History.
M.C. Campbell et al.
"These results suggest that diverse African populations were more subdivided with lower levels of gene flow during human history."
I suspect poor transportation and the lack of large states in Africa helped keep gene flow low between regions within sub-Saharan Africa relative to, say, Europe.