by Bryan Sykes
From a genetic point of view, I wanted to see whether I could find a parallel in the living descendants of the ancient Picts. Was there, hidden deep within the cells of Scots still living in the Pictish heartland, a signal of their ancient identity every bit as real, or perhaps more so, as the Stone of Destiny itself? We began our search for Pictish genes at Auchterarder, 15 miles south-west of Perth and temptingly close to the famous golfing hotel of Gleneagles, but I am sad to report that our research budget did not stretch to that level of subsistence. Auchterarder was the first of many visits that my research team paid to blood-donor sessions.
Three months before, in the spring of 1996, I had spent a week travelling all round Scotland visiting the directors of all the Scottish Blood Transfusion Service centres, enlisting their help in our project. It was never difficult to explain why we wanted to do this work, but there were a lot of details to be sorted out in getting permission from the donors’ representatives, as well as formal permission from the Transfusion Service itself, ensuring we did not compromise the confidentiality of the donors. We also had to agree a way of collecting the blood that would not interfere with the smooth running of the donor sessions. There was one thing both I and the directors were agreed on. We must attend the sessions in person. Too many researchers ask for blood to be collected on their behalf, without actually going to the sessions. This makes extra work for the donor nurses. I also wanted to be sure we were there to explain our project to the donors and get their consent, and also to talk to them about their own backgrounds and to get the feel of the place.
You may be a blood donor yourself, in which case you will know how the sessions work. As each donor arrives, they wait to be checked in. This, we jointly decided, was the best time for us to introduce ourselves and to ask for their consent to having us analyse their DNA. I was extremely fortunate in having a team of researchers in my group at the time who were absolutely brilliant in the art of persuasion. To the waiting donors, we explained that we were creating a new genetic map of Britain and trying to work out from it our Celtic, Pictish and Viking roots. That was about all the explanation anyone needed before agreeing to take part, especially when they realized we would only be taking a sample of blood from their donation, so there was no need for another needle. Donors told us where they were from, as far back as their grandparents. Those who didn’t know, generally the men, took a purple form home with them to ask someone in the family, invariably a woman, and sent it on to us.
By the end of the session in Auchterarder we had collected 187 blood samples, a wonderful start. Over the next two years we visited almost all of the donor sessions throughout Scotland, from Galashiels and Thornhill in the Borders, to Thurso at the very top of Caithness, to Stornoway in the Western Isles, to Campbeltown on the tip of the Kintyre peninsula. Everyone in my research lab joined in, even if they were working on different projects. Itineraries were prepared so that, with luck, several different sessions in different regions could be covered in one trip. We travelled to Scotland by road, by air or by the Highland sleeper from Euston. We set aside a small office at the Institute as a planning room, with donor-session schedules on the wall and a large map of Scotland next to them. By the time we went to our last Scottish session, two years later in Fort William, we had collected over 5,000 blood samples and clocked up over 50,000 miles between us. It is a testament to the team’s powers of persuasion that we only ever had one person decline to take part in the project–a farmer from Callander in the Trossachs, north of Stirling–who had to rush off because one of his cows was about to give birth and he didn’t have time to fill in our form. It says a lot about him, and donors in general, that they take their blood donations very seriously. None of the donors is paid a penny and the sessions have a tangible atmosphere of selfless community service. Most sessions are entirely run by women, with the only men present being the drivers of the vans that bring the teams and the equipment. It was friendly, calm and efficient. Very impressive all round.
To begin looking for the genetic signatures of the Picts on the mainland I began by dividing Scotland into regions. It was easy to decide where to draw the line on Orkney, Shetland and the Hebrides–they were islands–but on the mainland I needed to draw boundaries. These are shown on the map on page xiii. Pictland was covered by two of these: the Grampian region and Tayside and Fife, which for convenience we will henceforth refer to simply as Tayside.
Since we began our analysis of the Northern Isles with the Y-chromosome, and also in deference to the Pictish tradition of matrilineal inheritance, we began the search with mitochondrial DNA. The clan proportions for both Pictish regions were remarkably similar to one another–again you can see this in the Appendix. When I put the results through a statistical test, the only clans that had a significantly different frequency in the two regions were Jasmine, higher in Grampian, lower in Tayside, and Tara, higher in Tayside, lower in Grampian. Otherwise there was no difference. It looked, at this level of scrutiny, as if the maternal ancestry of the two Pictish regions was almost indistinguishable one from the other.
When it came to analysing the detailed sequences, though, I could see plenty of differences. That is always to be expected, because a large proportion of mitochondrial sequences are unique to one region. At the risk of being technical, of the 170 different sequences we found in the two Pictish regions of Grampian and Tayside, 70 occurred only once. I used our experience with disentangling the Norse and Pictish components in the Northern Isles to devise a simple score between 0 and 100 which would summarize the similarity between the two regions. If all the sequences in one region have an exact match in the other, then the score is 100. If none of them matches, the score is 0. On this scale the match score for Grampian and Tayside was 77. Of course that doesn’t mean a great deal without anything else to compare it with but, as we shall see, 77 is a very high score compared to most others. It is high enough to consider both regions as one for the purpose of searching for Pictish genes.
When I began to look in detail, it was immediately obvious that the DNA from these two regions was not really all that different from much else in Britain. There was no sign at all of the exotic sequences one might associate with a truly relic population that had been somehow isolated from the rest of mankind. Lots of sequences were unique to the two regions, but that isn’t unusual, as we have just seen. Looking through these unique sequences, I could see they were closely related to mDNA sequences from other regions, differing from them by just a single mutation. There was nothing very special about the Pictish DNA, at least on the maternal side. It didn’t seem to me that, on this evidence, a case could be made for treating the Picts of Tayside and Grampian as being particularly unusual. But that was just how they appeared to me at the time. We would certainly need to compare them with the rest of the Isles to gauge their true nature. That was my impression from the maternal signal, but what would the Y-chromosomes look like?
Once again the overall Y-chromosome clan structures in Grampian and Tayside were, like the maternal signals, remarkably similar to each other. The clan of Oisin predominated in both, rising to 84 per cent in Grampian–not quite as high as the west of Ireland, but much higher than in Orkney or Shetland. Wodan was quite high in both, at 12 per cent in Grampian and 18 per cent in Tayside respectively, but Sigurd was very low indeed. Only 2 per cent of men in both Pictish regions belonged to Sigurd’s clan. You will recall that, in Orkney and Shetland, we assigned all the Sigurds to a Norse Viking origin. On the evidence from the Pictish regions, with low numbers in the clan of Sigurd, it looked as if Grampian and Tayside had virtually no Viking ancestry. This is precisely what we would have expected from the history and the archaeology of both regions. There are no remains of Viking longhouses and no Norse place-names. In fact, some of the place-names have recognizably Pictish origins, notably Pitlochry on the River Tummel a few miles north of Dunkeld. In Orkney and Shetland the reverse is the case. All the place-names have Norse origins.
In Pi
ctland, the genetics suggests a very low level of Viking ancestry among the men. However, if we accept that, as I think we should, what can explain the substantial percentages of Wodan in both regions? In the Northern Isles the proportions of Sigurd and Wodan were roughly equal. If, as we had done, we attribute both clans in the Northern Isles to Viking settlement, based on the close affinities with Y-chromosomes we know exist in Norway, how do we explain the Wodan presence in Pictland? If Viking settlers in Orkney and Shetland were composed of roughly equal numbers of Wodans and Sigurds and these reflected the composition of a typical Viking immigration anywhere in the Isles, then only 2 per cent of the Pictland Wodans could have a Viking origin, leaving the other 12–16 per cent unaccounted for. When I checked through the detailed signatures of Pictland and Northern Isles Y-chromosomes from the clan of Wodan, there were plenty that matched–and plenty that didn’t. This was a puzzle. The Pictland Wodans could not all have arrived as Vikings, but where had they come from? Certainly not Gaelic Ireland, where they are almost unknown. Perhaps these men in the clan of Wodan really were the surviving descendants of the Picts.
I made the settlement date calculations, as I had in Ireland, for both the paternal and the maternal ancestors. There was still a wide gap between the male and the female side. The ages of the mDNA clans varied between Ursula at 9,200 years, slightly older than in Ireland, and Jasmine, again the youngest at 5,000 years. The paternal clans were slightly older than in Ireland, but still much younger than the maternal dates. There was no immediate answer to what this meant; indeed by this time I had more or less decided to wait until I had surveyed the entire Isles before trying to make sense of it all. But to what extent had the Picts been replaced by the Dalriadan Celts, the Gaels from Ireland? To try to find out we need to move away over the mountains that separate Pictland from our next destination–the Celtic west.
The gradual colonization of the west from the Irish kingdom of Dál Riata during the first half of the first millennium AD, and the consolidation of their Gaelic kingdom in Scotland following their defeat by the Ui Neill, had an immense cultural impact in Scotland. As we have seen, the language changed from the P-Gaelic of the Picts to the Q-Gaelic of the Irish and, on the accession of Kenneth MacAlpin in 843, Gaelic political dominance was complete.
In the west, one question which our genetic analysis could hope to answer was the degree to which the arrival of the Dalriadan Celts from Ireland displaced the Pictish inhabitants of the region. But there is also another factor to bear in mind, and that is our old friends the Vikings. The whole of the west coast and the Hebrides were repeatedly raided by Vikings from the first attack on the monastery of Iona in 795. If the experience in Orkney and Shetland is anything to go by, we should expect to see evidence of a Viking presence among today’s inhabitants of the west coast and the islands.
When we were collecting our samples from the west coast and from the Hebrides, there was a distinctly different response to my questions about people’s own ancestry. In Shetland the last thing men, in particular, wanted to see in their genes was the signal of an Irish ancestry. In the Western Isles, and along the west coast, there was still a certain arcane thrill at the possibility of a Viking ancestry, but this was eclipsed by an affiliation with a Celtic past, whatever that actually meant. People were keen to expose an Irish ancestry, if there was one, but most showed no real interest in the prospect of being of Pictish descent. And yet that was the most likely outcome, since it is almost always, in my experience, the earliest occupants who dominate the gene pool of a region. The later arrivals may get all the headlines, but it takes a lot to displace indigenous genes, especially on the female side. Thanks to the Scottish Blood Transfusion Service, we travelled to donor sessions from Thurso in the far north, along the west coast to Ullapool, Gairloch and Fort William, then south to Oban, Lochgilphead at the top of the Kintyre peninsula and right down to Campbeltown at the very end. We travelled over the sea to the Western Isles and across the bridge to Skye. We saw the land in all its moods, from brilliant sunny days when the bright hills shone in the sunlight to furious tempests when wind and rain lashed along streets and through doorways.
Not even the foulest weather prevented the calm progression of the donor sessions, even when the rain was coming in through the windows of community centres that had seen better days. Attendance by the donors was just as high in the bad weather as in the good. In Grampian, almost everyone coming to the donor sessions had been born nearby, and so had their grandparents. In this respect, the most stable place we visited was Huntly in rural Aberdeenshire, where 78 per cent of donors had all four grandparents born close by. In the west this figure was quite a bit lower, and there was a noticeable proportion who had moved into the area in the recent past, mainly from the towns of the central lowlands or from England. Since the project covered the whole of Britain, practically everybody could contribute to the outcome, even if they had only recently arrived in their current locality. It also worked both ways. In and around Glasgow, Edinburgh and London we often encountered donors whose ancestors had come from the west of Scotland and where, for the purposes of the genetic map, they could be confidently placed.
What of the results? We were becoming very adept at identifying Viking DNA and, sure enough, we found plenty of it. In Caithness and along the stunningly beautiful north coast from the Kyle of Tongue to Loch Eribol and Durness we found, by the same tests we had used in Orkney and Shetland, that 15 per cent of the DNA was Norse in origin. Like the Northern Isles, this was true both of Y-chromosomes and of mitochondrial DNA, so it looked as though it was by establishing family-based communities that the Vikings came to settle here, however unlikely this sounds in relation to their folk memory as bloodthirsty plunderers. However, in the Western Isles and Skye, the genetic evidence for a more typecast male-dominated Viking colonization began to emerge when we looked at the results.
There are twice as many Norse Y-chromosomes in Skye and the Western Isles as there are Norse mitochondria; 22 per cent of Hebridean Y-chromosomes, but only 11 per cent of mitochondrial DNAs, had a Norse origin. The further down the west coast, the lower the Viking component became until, in Argyll, it was down to 7 per cent for Y-chromosomes and only 2 per cent for mitochondria.
The diminishing Viking input and its increasing asymmetry between the sexes as we travel down the west coast seems to me best explained by a gradual process of Viking settlement from the early bases in Orkney and Shetland. Some men took their women with them, or returned to Orkney to bring their families once they had laid claim to a plot of land; others intermarried with local Gaelic or Pictish women. In general we found the same detailed Norse Y-chromosomes along the west coast and in the Hebrides as we had already discovered in the Northern Isles. It really didn’t look as though there had been a rush of fresh arrivals from Scandinavia.
If that was the level of Norse DNA, what of the rest? Could we assign this to Celtic or Pictish origin? For this, I made a start by comparing our results from the Pictish regions of Grampian and Tayside with the west coast locations. I could tell straightaway that they were substantially different. Not only that–there was also a big difference between the three regions of the west. Much as I had divided Pictland into Grampian and Tayside, so I split the west into three. They were the Highlands from Durness to Fort William, then from Oban south to Kintyre, which I grouped together as Argyll, and, thirdly, the Hebrides, which combined Skye and the Western Isles.
On clan comparisons alone, the Hebrides stood out as very different from the other two. Argyll, at this crude level, was far more like the Pictland regions of Grampian and Tayside than the Hebrides. The Highland coast was somewhere in between. Even when I removed the Norse DNA, the picture was the same. At the greater level of detail revealed by the precise sequences, in the case of the mitochondria or the profiles of the Y-chromosomes, the stark differences between the regions still stood out. For the mitochondrial comparisons, on the scoring system of similarity that I introduced in Pictl
and, which goes from 0–100 (the higher the score, the greater the similarity), Argyll vs Pictland scored 60, the same score as Argyll vs Highlands but much higher than the Hebrides scored in this equation of similarity with either of them. However, the Y-chromosomes told a different story. The Argyll Y-chromosomes were much more like their Hebridean counterparts than those in the Highlands.
If your head is spinning, you are feeling just as I did when I first tried to decipher these results. It seemed to be going so well. We had identified the genetic legacy of the Vikings and we had found that, just as the archaeology and history leads us to expect, they did not settle in Pictland to any extent. We had seen their diminishing genetic impact as we travelled further and further away from their forward bases in Orkney and Shetland. Until then everything made good sense. But then, the simple story, based on our historical assumption, began to unravel. Far from Pictland being genetically distinct from the Celtic heartlands of Dalriada and Argyll, they were remarkably close, on the maternal side at least. However, this similarity is not reproduced by the Y-chromosome, where Argyll has a low gene-sharing score with Grampian, even after the Norse component has been subtracted.
