Caucasoid vs. Negroid phenotypic variation

It is well known that the populations of Africa have an excess of neutral genetic variation. This observation, in addition to paleoanthropological data has led to the conclusion that mankind has originated in Africa.

Neutral variation is just one type of human variation. It is the type of variation that doesn't seem to be doing anything, thus making it useful to assess the relationships between populations. On the other hand, genes that actually do something may be under selection, and hence their patterns of variation may reflect environmental selective pressure and not genetic relationship.

We can't at present quantify the amount of genetic variation for most human traits of interest, because we don't have a clue about which genes are responsible for these traits. Therefore, we have to turn to the study of human phenotypes thought to be under genetic control.

I have carried a simple experiment using Howells' craniometric data to measure the level of phenotypic diversity for 45 different measurements.

I limit myself to the three medieval European populations (Norse, Berg and Zalavar) and the three Sub-Saharan African populations (Dogon, Teita, Zulu). It's interesting that the European populations are spatially closer to each other than the African populations. Therefore, we should expect them to be more similar to each other than the Africans are. In addition, if African neutral genetic diversity extends also to the functional diversity underlying human cranial development, we would expect the Africans to be even more diverse.

I limit myself to the male samples and compute the standard deviation for each trait separately for the Europeans, the Africans, and the entire human sample.

First, we can compute what the average standard deviation is, i.e., averaged over all 45 traits. It turns out that this is (normalized so that humans have SD=100):

Humans (100) : Africans (88) : Europeans (85)

Predictably, humans in total are most diverse, followed by Africans, and then Europeans. It is however surprising that Europeans are almost as diverse as Africans, given that they are drawn from three medieval locations of Central-Northern Europe compared to the three widely separated African populations. It would be interesting to see how these numbers look if we sample from the entire range of the Caucasoid and the Negroid race, rather than just these convenient publically available populations.

Of course, the above ratio is averaged over all 45 traits. What about individual traits? I list the three traits in which Africans are most diverse compared to Europeans and vice versa. Here we set the African standard deviation to be 100:

Simotic chord (least nasal breadth): Europeans (73)
Bregma-subtense fraction: Europeans (77)
Orbit breadth (left): Europeans (77)
Interorbital breadth: Europeans (116)
Nasio-occipital length: Europeans (118)
Biauricular breadth: Europeans (140)

It is obvious that for some traits Africans are more diverse compared to Europeans and for others Europeans are more diverse than Africans.

In conclusion, it turns out that we should not generalize from neutral genetic variation to functional genetic variation, as the latter may show a different pattern of variation than the former. Moreover, each group may be more diverse in terms of some traits but not in terms of others.