The DNA

Where this came from. Two kits belonging to descendants of this line, read at MyHeritage and at 23andMe. The work was done in the Lerena archive, because the tester's father was a Lerena; everything on this page is the half of it that belongs here. The Lerena archive was asking whether a chromosome could reach an Argentine great-grandfather. It mostly could not — but the same two kits are loud about the Cape, and that half of the answer belongs here.

Of the twenty countries named across 1,500 relatives, South Africa is first by a factor of four, and more than half of those name all four grandparents there. This is what a Cape family looks like from the inside. The next countries down are the shape of the diaspora rather than of the family — the United States 144, Great Britain 135, and then Zimbabwe, Namibia and the Netherlands in the tens.

The surnames those theories name, in order of the size of the match: Booyzen, Booysen, Barry, Kolbe, Kruger, Venter, Nel, Basson, Louw, Pretorius, Botha, Jordaan, du Toit, De Villiers, Olivier, Visser, Malan.

Booyzen, Barry and Kolbe appear together, repeatedly, in trees kept by people this family has never heard of. That is not new information about the descent — those three names are the spine this archive was built on, and they were established from registers. It is something better: independent confirmation, from the other direction, that the three lines really do sit together where this archive puts them. The Barrys → · The Kolbes →

Start with cluster 7. Cluster 7 is the one to start on - six matches from 56 to 333 cM, the largest match on the whole chart among them, carrying Butler, Daniel and Daschner. Those are the trees in which Booyzen, Booysen and Barry appear together as ancestral surnames. The Barrys →

And why the fourteen can be trusted. The niece's SON was clustered on the same night as a control. His father is a Defranceski of Istria, and his chart throws up two plainly Croatian clusters - Butorac, Pavelich, Jagustin, Krkljes, Racic, Glas - beside the Cape ones. So the tool demonstrably separates a foreign paternal line from this family's when both are present. The fourteen Cape clusters on his mother's chart are not an artefact of the method.

Each cluster is a couple this archive either already has or does not. Take them one at a time, open the trees, and find where they join the spine. Fourteen leads, each one already proved to be blood.

And the shape of it is a finding in itself. Cluster 1 is enormous - 71 of her 98 clustered matches, from 75 to 158 cM, carrying Ackermann, Andre, Attwell, Brits, Calitz, Carons, Carrick, Coetsee and more. A single block of seventy-one is not one couple; it is a densely intermarried community, which is exactly what the eighteenth and nineteenth century Cape was. Splitting it is a job for a tighter threshold.

Her chart was used as the instrument in the Lerena archive: subtract it from her niece's and what is left is the niece's FATHER's side. It worked, and the one paternal cluster it found turned out to be English rather than Argentine. The method and its trap - a floor mismatch that nearly produced four false positives - are written up there. How that went →

Small, and real. West Central African and South Asian together, at trace level, is the ordinary genetic signature of the eighteenth-century Cape - the household slavery the Colony ran on, written into the families that came out of it. In the niece these are halved by her Argentine and English father, so in her mother they are roughly double, and further back they are larger again.

And the care it needs. A trace figure at half a percent is at the edge of what any panel can call, and this archive is not going to convert it into a named ancestor. It is recorded because it is consistent, because it is where this family lived, and because leaving it out would be a choice as much as putting it in.

It is also worth saying plainly why it can be attributed to this side at all. The younger tester’s other two grandparents are Argentine and English. Neither an English Devon family nor a Río de la Plata one produces a South Asian trace. The Cape does, constantly, and it is where this family has been since the eighteenth century.

1. Cluster 1 on the aunt's chart holds 71 of her 98 matches in one block. That is the Cape marrying itself, not one couple. Re-running at a tighter threshold should break it into something usable.
2. 23andMe's Parental Inheritance, which would put the Cape trace definitively on this side, is behind a paid tier.
3. Nobody has yet taken the fourteen clusters one at a time and found where each joins the spine.

The first of those is the one that matters. The aunt is on this line and no other. Subtract her match list from her niece’s and what remains is Argentine and English; keep only what they share and what remains is this family, several thousand people deep, sorted for free. Nobody has run it yet.

The same two kits are written up from the other side in the Lerena archive, where the question is an Argentine great-grandfather rather than a Cape one, and where the answer is a good deal less comfortable than it is here.