Over the last seven Olympics, from 1984 through 2008, the 56 men's 100m finalists have all been of West African descent. The 56 have consisted of 17 African-Americans, 17 black West Indians, 8 West Africans (all from former British colonies, including one Nigerian running for Portugal), 7 black Britons, 6 black Canadians (1 a Haitian from Quebec), and 1 black Brazilian.
One unanswered question is: why the dominance by Anglophones? I count only 2 of the 56 coming from somewhere where English would not be a national or provincial language.
One theory is that there must be a lot of technical information about how to sprint fast that is passed down in English. Yet, the 100m dash is universally considered less technical than the 110m high hurdles, where 18 of the last 56 finalists have been non-English speakers.
So, here's a theory I concocted in 2004:
While cultural continuity is no doubt important, I've got a new complementary explanation for why Anglophone New World blacks are so dominant in the 100m over Spanish and Portuguese speaking blacks: the Iberian acceptance of mestizaje vs. the more racist English disapproval of miscegenation.
The 100m is a one dimensional sport where one skill is rewarded. For Darwinian reasons that we don't fully understand, this skill evolved to its global peak in West Africa.
Thus, top 100m men tend to be not just kind-of sort-of black in a Barack Obama sort of way, but really African-looking. While there are a number of sports, like the decathlon, where being of mixed race heritage may be of help (e.g., 1980 and 1984 gold medalist Daley Thompson and 1996 gold medalist Dan O'Brien have black fathers and white mothers), the 100m is not one of them. The 100 meter greats, like Carl Lewis, tend to be very black-looking. Frankie Fredericks of Namibia was perhaps the only top 100m man of recent years to look noticeably part-white. (He looks rather like Sir Alec Guinness.)
The slave trade spread West Africans to the New World. But population genetics studies show a sizable difference in what happened next. In the Spanish and Portuguese colonies, black and white genes tended to spread out across the population, but in the English-speaking world they tended to bunch up toward one pole or the other. Sure, there was lots of inter-breeding, but the one-drop rule in the Anglophone world, combined with restrictions on interracial marriage, had a little-understood impact: by defining a half white-half black person as socially black, it drove them toward marrying another black person rather than a white person. Thus, socially defined black people tended to not be more than half-white.
Thus, the first genetic study I've seen to address this suggests that only 10% of self-identified African-Americans are more than 50% white. In contrast, blacks in Mexico were almost completely absorbed into the general population.
Thus, in the English-speaking countries, black genes tend to be more concentrated in particular individuals.
One other population genetic aspect that could be important is that North American blacks tended to come almost solely from West Africa, while South American blacks were drawn from a wider sphere of Africa, going all the way around to East Africa. Lots of Brazilians slaves came from Portugal's East African colony of Mozambique. The weakness of Brazilians as sprinters may have to do with A] The really fast runners go into soccer. and B] Even the blackest Brazilians aren't as West African as Anglophone blacks -- they may have more South African or East African ancestors. (East Africans tend to be best at longer distances from 400m through the marathon, depending on their tribe, but none have ever been world class world-class at 100m or even 200m.) That may help explain the fact that a dark-skinned Brazilian man set the marathon record back in the 1990s.