The danger of assumptions

From looking at the tree attached to a DNA match, I hoped / presumed / assumed that I'd found yet another piece of corroboration for our connection to the Jenkins/Buchanan family we have so many matches to.
Which got me quite excited.
These brickwalls (Archibald Henderson and and his daughter-in-law Amelia Millar) have been standing a long time.

So I spent a happy day or two researching and validating the line back from South Africa to Perthshire.
I've never read a will before where someone was divvying up a town amongst his children: to x the railway station to y the post office to z the store.....

Between them, the two sibling kits involved had shared matches to a satisfactory selection of my Henderson/Millar cousins. Both kits were managed by a person we didn't match, so her relationship to the matches wasn't readily visible.

Thankfully this was a responsive tester and the flaw in my assumption was quickly revealed.
She was their mother, and it was her line that led back to Jenkins/Buchanan family via a branch in South Africa.
And we didn't match her. 

The matches were over 30cMs, so it was extremely unlikely this was a testing company glitch unfortunately.

So back to the drawing board and over to the paternal side of their tree, which was a bit sparse given this was why they had tested.

It didn't take me too long to realise that a Smith born in Central Otago around that time period just might be the Smith branch, well one of the Smith branches, in my tree.
All Smiths are related? Right?

Sure enough, he was already in my Henderson/Millar data.
I used to hold all these names / dates / places in my memory cells, but the explosion of people that you get interested in when exploring DNA matches means some just have to drop out to make room!

I feel old, the matches' grandmother is my 3rd cousin.

I was lacking the crucial link in my database down to the new matches as this 3rd cousin's sister had omitted to tell me the rest of the story back in 2003 and I would never have found this branch without family hints.
I feel some updates coming on.

Moral of the story: 

Shared matches and hope are not always enough to prove a theory.
If the mother had not also tested I'd likely never have realised the actual source of this DNA as at first glance nothing looked very promising to explain the link on the father's side.

The irony is that the matches will be related to the Jenkins/Buchanan family from both their parents - once we do figure out how I and my cousins connect to them given so many matches are telling us we do.
If only the right branch to test could be found, or pop out of the woodwork!

What you can find when you aren't even looking...

...  or 

I hope this stroke of serendipity portends other breakthroughs on the gaps in my tree in 2021
starting the year as I hope to go on - well apart from interrupting the chat on our Talking Family History zoom session last night with an excited "I've just cracked a new DNA match".
Thankfully for the group, which I admit I was listening to rather than watching, my internet connection dropped out.
They were spared the journey I'd been on most of the day, culminating in the final piece of the jigsaw being found whilst I was listening, but also documenting the newfound connection on FamilySearch and in my genie program, dotting the is and crossing the ts ready for adding to WikiTree to see what further DNA clues may pop out as connections are found there.

The excitement being that not only had I cracked where a new DNA match likely fitted, but also added two ancestors to my tree at 5* great grandparent level - bringing that level up to 61 out of 128 identified at least by name.

DNAPainter Tree / TreeCompleteness stats

The journey

Most days I quickly check a basic set of Ancestry kits I manage* for:

- new matches predicted as 4th cousin or closer, and 

- any more distant unviewed matches that now show up in the Common ancestors filter

* myself, a maternal 1st cousin, a paternal 2nd cousin and a 3rd cousin on a brickwall Borders line

Top of the unviewed list yesterday (New Year's day) was a new predicted 4th cousin match who looked to be the father or son of a match over on 23andme which I had explored a while ago because my chromosome maps on DNAPainter had shown he matched me at the same spot I matched a known 5th cousin on MyHeritage. But he had no tree and only listed a couple of names, neither known to me.

The 5th cousin and I (and another match on Ancestry) share common papertrail  ancestors of John Hanham and Ann Rawlins who married in 1777 at Sutton Montis. This being my relatively recently found branch of my Somerset ancestry from when the mystery of my 2* great grandmother Jane Gibson was cracked (The Lazy Vicar refers).

I wasn't actively researching for Ann's parents, having run out of steam back when Jane's origins had been identified, and the thrill of the new chase abated once the lack of further success kicked in.

In recent times I've been having, yet another, concentrated effort mining all the DNA clues at my disposal (currently re-reviewing shared matches with assorted 2nd and 3rd cousin pairs) to try and crack my two very solid, chip resistant brickwalls of Archibald Henderson and his daughter-in-law Amelia Millar.
And once again, getting nowhere beyond realising that just possibly there may indeed be a reason I, and my paternal cousins have always shown about 10-16% Irish ancestry, and starting to notice a growing number of admittedly distant matches from around Mull / Argyll which seem likely to extend the search from Perthshire/Stirlingshire.

Research into this new English match appealed as light relief, and went rather more smoothly given the added bonus of now having a tree to work with from the Ancestry match - which included a John Rawlings born in Somerset about 1781, having children in West Lydford.

A working assumption was that by age alone my Ann and his John might be aunt/nephew. Sutton Montis and West Lydford aren't that far apart, at least as the proverbial crow flies.

Somerset: Yeovil / Martock / West Lydford / Sutton Montis

 John Rawlings and wife Jane were easily enough found in census and the Somerset records (images on Ancestry).
The assorted versions of the families on the FamilySearch tree were tidied up by reconciling discrepancies and either merging duplicates or  splitting out the waifs and strays to make what looked to be a sourced coherent whole, including the two families of the two sons, James and Henry with whose descendants my matches were.

Which left  John's parents and siblings to identify. That wasn't as easy until I found an 1853 death that led me to question John's age somewhat. Aged 80 being somewhat older than the 70 he'd showed as in the 1851 census which age most trees used.
It must have been a tough life as a yeoman farmer in those last 3 years to age him that much!
With the revised earlier birth,  hindsight showing this as rather more within range for the rounded down 1841 census age of 65 there was a good candidate of the right age and in the right place, to an Edward and Frances Rawlings, with very irregularly spaced siblings ranging from 1759 to 1774, but no Ann who could well be a sibling with these earlier timeframes being thrown into the mix.

My Ancestry searches, I thought, exhausted, I switched back to FamilySearch to pull the final pieces together there too, and did a last search for any more baptisms.
And there was an Ann, baptized in West Lydford in 1755, sibling to John.
Which point is where I excitedly interrupted the Zoom session chat as that just had to be way too much of a coincidence.

My working assumption has been quickly revised to siblings.

 Overnight, Ancestry's ThruLines hints kicked in showing the four matches from the two lines down from John, as hoped / expected.
Today I've checked the maternal 1st cousin's ThruLines, and they now show a match descended from John, and Ann's sister Jane.
This is looking good.

Welcome to the family Edward and Frances Rawlings.

The lessons

There's always one more record to be found - try different sites, their search algorithms will pop up different results

Remember to check and re-evaluate all records - I'd have found John's baptism quicker if I'd re-evaluated the 1841 census earlier

Keep good notes of your DNA matches.

DNAPainter imports of segment data are invaluable for a cross-company chromosome map.
This gave me the initial lead for this family