RootsChat.Com

What site did you take the test with? In my experience, very-very few people from Germany have tested with Ancestry, and not many more with MyHeritage.

What are your largest cM matches? And are you able to sort out between your mother's paternal and maternal matches?

You'd generally need matches through the actual unknown father to determine who he is. If there are cousin and/or 2nd cousin matches, you might be able to work out his grandparents, but probably not him. If there are no useful matches within that range then it gets a bit tricky, especially if there happens to be more non-paternal events.

The basic process is to go through all your matches from the largest down; go through the shared matches (in a tab), noting groups of matches that tend to group in a cluster. Generally I require a match to cluster with at least 3 other tests, but there can be overlaps that we can't really flesh out with Ancestry's limited tools. Add those to a group. Look at their trees, open the dead-end ancestors and press the [Search] button to see if you can continue the line in public trees (obvious accuracy warning). When you find common ancestors among the matches, note them down. If the match is a dead-end, note that so you don't keep going over it. Note it if it's possibly useful. Note if it you've added it to a file with MRCAs.

Once you've done that, if you have close matches, you will now have your mother's documentable maternal line, and unknown groups that will probably relate to the unknown father. You can then look through those groups, looking for a possible marriage between descendants of common ancestors in the matches.

That's the basics, but there are plenty of caveats, things to understand, and a few pitfalls.

Generally speaking, DNA tests are just an extra coordinate and require a lot of leg work.

(№ 12) I had a German man (1) August in India and Singapore; owing to him being a Lutheran none of the records of his marriage or baptisms of his four children are to hand. I'd noted from a grave inscription and later marriage that he was almost definitely married to a Miss Kinsgley, but I thought he might have been married twice, and I couldn't find any Kinglsey DNA matches, so that hypothetical line was pushed into the background.

A few days ago a DNA match comes up and I find Kingsleys in India. She overlapped with a total of five descendants of (1) August, which seemed a bit of a thin overlap. No other shared matches. The new testee gave me access to their matches and I found three further Kingsley relatives in India in her matches, which just graze several other tests.

Unfortunately there are no shared matches other than those descendants of a man probably born in India in 1804, so no easy route back to England. The testee is now getting their father and his cousin tested, who are on the same generational level as my great-grandmother - so that may provide enough surface area to graze some Kingsley matches in England.

This new match says her family beleived they are relatives of the author Charles Kingsley, whose paternal grandfather was in India - though I can't see a simple link, unless it was on the wrong side of the sheets.

This line of ancestry also includes a William Jones, surgeon in the army. Luckily the East India Company records listed his date of death, so after some after some rummaging around I've found him in the 1841 + 51 censuses, his will, death entry, and 1838 marriage, but no baptism - so he is the target for brick wall №13.

It would be lovely to hear how others feel/see this quandary of mine. I would love to be convinced it is worthwhile continuing, though exactly where I would fit it in, I don't know.

Jane

PS If you think this is bad, you should see my emails!

If you mean if it's worth pursuing the small clusters, generally I'd say - no. If you do find anything, it usually takes a lot of legwork. I've found that clusters when the largest match is not more than ~29cM (not sure of the exact threshold) are difficult to pin down. That's probably going to be a bit more difficult if your ancestry is Irish and to a lesser extent Scottish, owing to poorer records.

I do have one cluster where the largest matchs are around 12-13cM. That cluster does just graze another cluster in two matches. And of the matches, the point the common line goes back to my ancestral area is in the late 1700s; and then beyond that there is a man born c. 1720 who has a few lines within the matches. So it's quite clear that's the line.

The MRCA is born c. 1660. I do have several other clusters with MRCAs in the 1660 window - none earlier. But those are clusters that have multiple intersections wither more recent generations.

But generally speaking, I've not been able to locate how I'm related to these more distant clusters. It's obviously a bit of pot luck. Of these clusters I've tended to find the MRCAs are further back than you might think. If the largest cM match in a cluster is 20-29cM, I'd expect the MRCAs to be more in the window 1700-1725. While that level also tends to be the ballpark for larger 5th cousins, it just happens that these smaller chunks can end up not getting broken up over a couple of generations. I forget the exact amount, but the average size of a segment from recombination is about 30cM. So while it would be very uncommon to have a 5th cousin with a full 70cM segment in common, a segment of about 25cM can pass unchanged across multiple generations. And there are areas of DNA where recombination is less likely to be recombined.

If you are lucky, you might find an obvious surname in such a cluster. But the odds are against it. These clusters will take you to MRCAs born around 1700. So to match them you'd need to know a total of 256 pairs of 6th GGPs from your tree and your match. And that's not likely, since gaps and dodgy trees proliferate in this period. I've got 64% of my 6th GGPs - let's say 75% wither surnames - so a 25% blind spot where I'm just not going to be able to make the match. Let's say you have a 50% blind spot in your matches' tree. That means there's a 37.5% chance you will be able to locate the MRCAs. Then throw in NPEs, dodgy trees, the issue I mention below, and whatever else - and you're down to a low chance of finding a match.

It would be easiest if your tree is very diverse in localities and your ancestors are all from wealthier parts of England. Most of my ancestors are from the Welsh borders, so it becomes quite difficult to figure out where another Edwards might fit in.

These more distant clusters also tend to have MRCAs shared in the matches which are removed from your actual MRCAs by more than one generation. The chances of matching a 7th cousin on Ancestry is about 1%. So you're looking at people whose more distant ancestors probably happened to have a bigger chunk from one set of MRCAs that remained bigger than average for a few generations. So your matches can cluster on descendants of one nth cousin. The result is you can have a cluster where, say, several matches have MRCAs Robert York (1799) and Catherine Patrickson (1805). But your MRCA might be George York (1727), and you don't have any matches that show that flow. So with such clusters, you often end up with several to a dozen possible MRCAs.

As mentioned above, you can find matches where about 29cM is inherited by a shared match and their child? Over how many generations can 29cM be inherited unbroken? The answer is likely quite large, though increasingly small. And what about 12cM? It's likely some of our shared matches at that level are from the medieval period.

For the past few days, I've noted that almost every single match I looked at had shared matches, including low cM matches. Previously once you got into those lower cM matches, more would have no shared matches than not.

I just went and checked my 19cM matches that were not assigned to a group; and the reason for that would be that they previously had no shared matches. I would have otherwise grouped them.

I've not got ProTools at the moment. Does anyone have that and able to check if Ancestry now shows shared matches in the 8-19cM range?

Never really made sense why Ancestry didn't provide this. If so, they really need to increase the groups from 64 to 256.

Does anyone have any idea where this place is in Montgomeryshire? It's probably in the parish of Aberhafesp, Tregynon or Bettws Cedawain. I think it's probably the name of a building. And I'm referring to the first one, not Llanfihangel.

The best I can come up with is Abberhew in Bettws.

https://ibb.co/N66wm8x7