Firstly, I’m using the term “African American” in the restricted sense of Black Americans whose ancestors trace back to pre-Civil War US slavery days (whether slaves or Free Men of Color). Recent immigrants from Africa would be excluded, for obvious reasons as you see what my questions is. Not really interested in any Native American admixture (which is much, much lower than commonly thought, btw).
It’s generally known that African Americans average about 20% European ancestry. But the standard deviation is quite large, and there are folks with as much as 80% European ancestry who self-identify as black or African American.
I’m interested in the other end of the bell curve. Are there any documented cases of African Americans with 0% European ancestry? I’ve seen lots of genealogy shows and not once have I seen an African American’s DNA analyzed that gave such results. I’ve seen as low as 5%, but never less. And I’m assuming that 5% is above the “background noise”.
Also of interest, if it exists, would be a statistical analysis of the likelihood of such an individual. My own guess is that the odds would be vanishingly small.
This study looks at the genetic admixture of African Americans. [PDF]
Based off this we have two sample means and two standard deviations:
14% mean, 10% SD
17.7% mean, 15% SD
Assuming a normal distribution, we can calculate two estimates for the probability of an African American having no European ancestry: 8.08% and 11.9%. Given an estimated 42,020,743 African Americans [PDF], there are approximately 3.5 million to 5 million people who identify as African Americans with no European ancestry. I don’t think data exists to analyze solely African Americans with pre-Civil War ancestry; that doesn’t seem to be a distinction any demographers make.
True, as far as it goes, but I don’t think it’s a normal distribution. But that’s certainly a reasonable, first order approximation even if it turns out to be inaccurate.
I would argue that you cannot use a normal approximation here. You are interested in 0% which is at the endpoint of the distribution where I suspect normality is not even close to correct. I assume that the probability distribution for genes is close to some type of beta distribution at least approximately, but I’m not sure what. But even that is a continuous distribution.
On a somewhat unrelated note: Comedian and TV host Conan O’Brian says that he has only Irish-American and Irish ancestors (who came to the US before the American Civil War).
Wouldn’t there be “European” genes floating around Africa anyway? There’s been plenty of contact between Europeans and Africans of all sorts for thousands of years. Similarly you’d expect "African " genes turning up in Europeans. Then you get into a weird space with trying to figure out what a “pure” African genome is.
If that’s true, then maybe the question is more of a historical than a genetic one.
Specifically: Are there any African Americans whose ancestry in North America we could could trace all the way back to the pre-Civil War period with no evidence of European ancestry during that time?
It’s widely understood among historians that sexual relations or sexual assault (i’m not going to delve into the thorny issue of consent here) between masters and slaves was extremely common during the colonial period and the national period before the Civil War. Thomas Jefferson and Sally Hemmings constitute a well-known example, and it appears in a bunch of famous slave narratives, including Frederick Douglass’s, where he tells us right at the beginning of his book that his master was widely understood to be his father. Harriet Jacobs talks about how her master pursued her relentlessly, and how she only managed to avoid him by taking up with another (white) man. Solomon Northup talks about Edwin Epps’ frequent raping of enslaved woman Patsey.
Of course, because these inter-racial sexual relationships were illicit—banned by law in most slave states, and considered socially and culturally unacceptable in all of them—we don’t always have a lot of historical evidence for them, outside of those narratives. Planters did not generally talk in their diaries or letters about the times they assaulted their female slaves, or about their own mixed-race progeny among the children of their households. It’s not really clear to me what percentage of slave-owning households this would have applied to.
And even if it happened on the majority of slave-owning properties, it’s worth noting that the majority of slaves in the south were concentrated on a relatively small number of large plantations. The quoted numbers vary a little depending on the source, but about 50-55 percent of slaves were held on properties with twenty or more slaves, and about 70-75 percent were on properties with ten or more. Presumably, even planters who sexually assaulted (at least one of) their female slaves might not necessarily have assaulted ALL of them.
Leaving aside the issue of genetic testing, and focusing on historical evidence and the mathematics of the population, i wonder what the chances would be of a modern African American having no person of European ancestry in his or her family tree in North America, going back to the antebellum or the colonial period?
If you want to ignore actual genetics data then this just becomes a probability problem.
You can make assumptions about number of generations, percentages of inter-race shagging and so on an so fourth and come up with some sort of estimate.
I ain’t doing it though (the calculations, not the shagging).
Probably, but this is at least part of what I meant when I said “background noise” in the OP. The other part would be the simple accuracy of the DNA testing.
I think the point mhendo is making is that it’s impossible to test everyone (at least now), and so population statistics would be a way to make up for that. For instance, the type of modeling use to determine the Identical Ancestor Point.