Back to the African drawing boardLast night, I found a lot of new matches who share my great-uncle's West African segment, and it turns out that my paper trail connection to at least two of them (KN and her daughter) does not fit my theory of how the African DNA got into us. KN is a great-great-granddaughter of Juliane Davidsdatter (1813-1879), a sister of my 3rd great-grandfather Daniel Davidsen (1819-1866). Juliane and Daniel came from Vanse parish on the Lista peninsula at the extreme southwestern tip of Norway.
The thing is, FTDNA gave my great-uncle another match on the exact same segment, a man who also descends from Juliane, but according to FTDNA, this man did not match the two Haitians with whom my great-uncle shares his African segment. So I concluded that the match with this Juliane descendant had to be located on the other copy of the chromosome, the copy without the African segment. For this reason, I was completely sure that the African DNA could not have come from the Davidsen side (or from my great-uncle's paternal side at all). This led me to formulate a theory of Haitian descent on my great-uncle's maternal side, and it is this theory that has now gone bust.
Because KN and her daughter do match the Haitians, and they share the West African segment with them as well as with my great-uncle. The full shared segment between KN and the Haitians is over 20 cM, so this is not a tiny trace. With the other Juliane descendant, there seems to be something wrong with FTDNA's matching algorithm. I have double-checked, and my great-uncle still has zero ICW matches with this man, which seems suspiciously low.
At least I now have a very strong hint as to which line the African DNA comes from - the Davidsen line - but I still have no idea how it happened. Especially since the match with the Haitians suggests a rather close West Indian connection. Our ancestor might yet be from Haiti. However, it cannot be Dérival Lévêque, because he was born as late as 1788, and it is extremely unlikely that he would father two illegitimate children (Juliane and Daniel) with the same woman in the same distant land six years apart. And if he did, both my great-uncle and his matches would have had much more West African DNA than they actually do. (My old theory had an additional generation between Dérival and my great-uncle, which would account for the small amount of African DNA.)
Alternatively, perhaps Daniel Davidsen or one of his close relatives travelled to Haiti and fathered children there??? We know there was contact between Haiti and this part of Norway in the mid-1800s... Daniel's wife's first husband died in Port-au-Prince in 1845. It would be a bit ironic for a Norwegian sailor to bring African DNA into a Haitian family line, but it is not unthinkable!
The good thing about this is that the truth came out, and that I've now got a total of seven matches who share my great-uncle's West African segment. This means that the segment is more substantiated than ever, and that we have a good number of individuals to compare in order to pinpoint our common African ancestor exactly.
Additionally, KN, who is one generation "below" my great-uncle (i.e. one generation further removed from the MRCA couple), has tested at 23andMe and shows up with 0.1% West African there. This means that my great-uncle should get at least that amount if he tested at 23andMe, probably more, since his segment looks somewhat larger than KN's at GEDmatch. This is encouraging. 23andMe estimates KN to have a West African ancestor between 6 and 9 generations back, probably born between 1690 and 1810.
In other words, the siblings Juliane (born 1813) and Daniel (born 1819) must have been of quite recent West African descent. They probably had a West African 3rd great-grandparent or closer. There might even be records documenting it, if we can only find them!
Bye bye EnglandThe match with KN and her daughter not only overturns my theory about how we got our African DNA. It also overturns the family story that my great-great-grandfather was the son of an Englishman. Why? Because now that my great-uncle matches three descendants of Juliane Davidsdatter on the same segment (coming from two different lines of descent from Juliane), I can no longer chalk it up to chance due to shared geography. I have to accept that her brother, my "official" and very much Norwegian great-great-great-grandfather Daniel Davidsen, is in fact my great-great-grandfather's biological father.
There was (and, indeed, still is) genetic evidence for a close connection to England (see this blog post), and an FTDNA project admin even told me he believed our family story to be true based on Y-DNA, but the evidence of the autosomal matches is now very strong, and can only be interpreted as debunking the family story. It seems my great-uncle's paternal line might well have English origins, but much further back in time.
Unfortunately, the loss of the English Lister connection means that several of my blog posts are no longer valid.
Summary1. I've "regained" my 3rd great-grandfather Daniel Davidsen (1819-1866).
2. Although my Haitian theory is shown to be wrong, it is now more clear than ever that our West African DNA is both real and recent (definitely within the last 500 years, and probably no earlier than the late 1600s), and that it comes from none other than Daniel himself. Our West African ancestor is hiding somewhere in Daniel's ancestral tree...
Stay tuned for updates, because I will solve this puzzle.