Earldom of Orkney DNA during the Late Middle Ages
by Craig Kirtland-Sinclair
Sinclairs in the North
St Clairs were first documented in Caithness as early as 1321 when Sir Henry Sinclair of Roslin was the Scots baillie to Caithness; then in 1364 we find Thomas Sinclair similarly employed in Orkney where he was ‘Baillie of the King of Norway’. But it wasn’t until 1379, when the St Clairs of Roslin became earls of Orkney, that this geographic location really became associated with the name St Clair and Sinclair. What is interesting, is that the lineage defining R-S5246 SNP, a subclade of S21/U106 > Z346, had been given an estimated date very close to the time that the St Clairs of Roslin became earls of Orkney in 1379; however, the "best guess" forming of this SNP has now been adjusted to 1045 A.D (Iain McDonald of Jodrell Bank).
The Earldom group is very interesting because it includes a small number of members with the surname Cummings. They are positive for both S5246+, S5629+, S5632+, S5634+, S5636+, and FGC15253+, but negative for FGC15254-; so possibly an early split before the foundation of the St Clairs of Orkney, when they were barons of Roslin in Lothian.
However, YFull recently analysed 4 Cummings raw data samples and 1 McDonald whom are S5246. But there is something interesting here. One of the Cummings samples (from England, but with a papertrail back to the 'Earnside' branch of Cummings, who are the same patrilineal line as the 'Altyre' Gordon-Cummings and "the" Comyns) and McDonald split from the main Cummings cluster (which share a TMRCA during the early part of 16th century) at S5246, FGC15253 (YFull groups them S5246*), with an estimated TMRCA of 1050ybp (969 A.D.). If this age estimate holds up then it could suggest they were patrilineally Cummings as early as the 11th century, during a period of fluid surname usage.
To test this age estimate, all the earldom samples were invited to upload their Big Y raw data to YFull, and a high percentage did so. The YFull tree (July 2019) estimated that the S5246 block formed approx 950ybp (1069 A.D.), and the TMRCA for the single Cummings kit, single McDonald, Cummings cluster, and Sinclair earldom cluster, dates to approx 900ybp (1119 A.D.).
So to simplify things, there are three branches splitting off from one another at the S5246 block: 2 Cummings lines* and the Sinclair earldom line. Estimated to share a patrilineal common ancestor during the latter part of the 11th to early 12th century.
The question you may ask, with 2 lines being Cummings and another being Sinclair, were they patrilineally Cummings or Sinclair? Probability would obviously lead us to favour one over the other, that being Cummings, but both scenarios are possible. We must also consider this, that if the TMRCA was just a slightly earlier, then we are back into the period of fluid surname usage and could be looking merely at mutual origins on the Continent.
Historically, Sir Robert Sinclair (whose origins are still to be confirmed*) was part of William Comyn's (1163-1233) entourage during the reign of Alexander II (1214-1249). Robert first appeared in charters from 1190 onwards and by 1203 he had progressed to royal court records during the reign of Alexander's father, William the Lion, but disappeared from records after 1240. Comyn himself was one of the most powerful men at the court of King William I (reign 1165-1214) and witnessed no less than 88 of the king's charters.
*It has been suggested, from studying charters, that Robert may have been of the Herdmanston family, and that he possibly married into the Graham family whom held lands in Pentland (including Roslin). The earliest surviving documents to place a Sinclair at Roslin dates to 1279 with Sir William Sinclair of Roslin, but this may not be when they first arrived on the land.
So the Cummings name and Sinclair are part of the same narrative within a presentable timeframe. This began with Sir Robert Sinclair in the 1190s and extended until the end of the 1200s with Sir William Sinclair of Roslin, who was also a close associate of the Comyn earls of Buchan. But the relevance of this is highly dependent on both the accuracy of the TMRCA and the papertrail of the 'Earnside' branch of Cummings.
Taking everything into account, the Sinclair Earldom "lineage defining" SNP is now adjusted to *FGC15254*, which has now been estimated by YFull (as dated July 2019) to have formed approximately 900ybp (1119 A.D.).
*continuing analysis will further define the S5246* group with the single Cummings and single McDonald sample. YFull is reliant on a good volume of data to enable them to accurately age SNPs and MRCAs, so it would be useful if all S5245 and S5246 Big Y testees uploaded their .BAM files for this independent NGS interpretation service.
Subgroups / Lineages
The “Orkney and Shetland Cluster” is additionally positive for FGC15254+, FGC15257+, FGC23570+, FGC15256+ SNPs. However, there are a small number of members with the surname Spence, who descend from this Sinclair line, and other Sinclairs who have Orcadian ancestry and are positive for the FGC15254 SNP (formed approx 1119 A.D. Yfull) but negative for the FGC15256 SNP (formed approx 1544 AD YFull & 1453 A.D. McDonald).
One of the kits in this cluster has their earliest known ancestor as being born about 1600 in East Lothian. They don’t have any known ancestry much further north. If there were considerably more, then perhaps geographies for this cluster would expand further south; although it may only suggest a post Orkney migration.
The “Caithness Cluster” differs, they are negative for the FGC15256- SNP, but are further identified by being positive for the FGC35613+ SNP (formed approx 1544 AD YFull & 1472 A.D. McDonald/jb.man)
Another interesting thing to observe is the DYS487 YSTR marker: “Orkney and Shetland Cluster” is “13” and “Caithness Cluster” is “12”. Earlier matches are DYS487=13, so we could be looking at the “Caithness Cluster” mutating away on this marker.
There are a small number of members who think they may descend from the Broynach line of Sinclairs and are negative for the FGC35613, but more research is required to confirm.
Where did they come from?
From the frequency of matches, the origins of Sinclair Earldom DNA are possibly Frankish - pre-medieval Germanic - and were in this region up until 1,000ybp, as suggested by their S5627 match (upstream of S5246) from Baden-Württemberg in Germany with the surname Ruf.
However, one of the difficult things to establish is the probable DNA of the Saint Clairs found in 11th century Normandy and England. There's no verifiable records to suggest the Anglo-Norman and Scottish branches are related; but that's not to say this isn't the case. Sadly, the English line died out by the latter part of the 15th century, so no clues there.
So would they have a Northern Germanic or classic Norse type DNA? But it's not as simple as that. Many pre Norman families of Northern France and the surrounding region became absorbed into the Norman culture, so this makes pinpointing a definitive DNA type incredibly difficult. They could have classic Norse or perhaps an indigenous French DNA type of the pre medieval period. Also, the Germanic S21/U106 haplogroup has a relatively low frequency in France or those with recent French heritage; but a greater volume of testing is required in this region.
It would be quite impossible to answer this question without a-DNA (ancient DNA samples). There are possible options in southern England (Danbury St Clers) but it's a moral dilemma regarding the disturbance of non archaeological sites and would require special permissions from the Church of England and a sizeable budget. And the answers may still be inconclusive.
What to do if you match?
The Z8 SNP Pack is the recommended option for those matching close on the Y-STRs to the Sinclair Earldom DNA lineage. However, the Big Y-700 is still the ultimate test to do, and those who have done so should consider sending their Big Y raw data (.vcf file) to Ytree and (.BAM file) to YFull. This will help to identify genetic splits in the Sinclair line and possible earlier origins.
If you are R1b-P312, search out your haplogroup in Alex Williamson's The Big Tree, which is a draft phylogeny for NGS (next generation sequencing) raw data from tests like the Big Y. S5627 and subclades below.
And for Iain McDonald's U106 tree and age estimates
If you don't know where that SNP belongs, search for it in the NGS YFull Tree. S5627 and subclades below.
'Years Before Present' (ybp) is calculated to begin at 1950 in the archaeological discipline for the purpose of radiocarbon dating. However, YFull uses a formula of 144.41 years being the average time in between SNPs and 60 for the assumed age of the person. YFull also round off TMRCAs to 25, 50 & 100 years, depending on the estimated age range.