The Origins of the British: The New Prehistory of Britain (20 page)

Figure 3.7
Ivan (I gene group): male re-expansions into north-west Europe from the Balkans and Ukraine after the Ice Age. Ivan is the second largest male gene group in Europe. This map shows his likely Ice Age refuge in the Balkans and his main regions of dispersal – Slavic countries, north-west Europe and Sardinia. While these dispersals most likely occurred during the Mesolithic and Neolithic, they may have started even before the Younger Dryas freeze-up 13,000 years ago.

A number of researchers have used detailed Y-chromosome markers known as
STRs
(
single tandem repeat sequences
) in an attempt to overcome this problem, and over the past few years have accumulated a huge British dataset of STR gene types within
the established Y gene groups in order to break them up into more detail.
⁴¹
STR testing adds much more information for each individual than just the group markers because STR markers are more rapidly mutating than the group markers, but the information is less easy to use in constructing unambiguous gene trees. Lacking this last step, of breaking the gene groups into smaller branches, the data, especially for British males, are under-analysed with respect to defining more genetic branches.
⁴²

Early on in writing this book, I decided to address this problem and use these STR gene-type data to map the detail of British colonization, by defining more geographically useful clusters of gene types, each closely related to identifiable founding gene types in Western Europe within each of the main gene groups.
⁴³
I have re-analysed a huge Y dataset of 3,084 individuals, collected mainly from British and continental European populations this century. The Y-chromosome results have been published, as comparable but separate datasets in a number of different papers which are in the public domain, covering Spain, the Basque Country, Turkey, north-west Europe, Scandinavia, Iceland and the British Isles. I pooled all these datasets for reanalysis. My aim was to chart the post-glacial re-expansion from the Iberian and East European refuges. My methods for this analysis and for other European male lines entering the British Isles are described in the Appendices, where the naming and definition of groups and clusters are clarified, and there are dated trees both for maternal and paternal lines.
⁴⁴

For the British Isles, I have collected all the available published data on the Ruisko, Rostov and Ivan gene groups and compared them with mainland Europe.
⁴⁵
This reveals an even more dramatic British dominance of Ruisko than on much of the neighbouring
mainland: he accounts for around 90% of all males in Ireland, north and south Wales, and Cornwall (
Figure 3.6a
).
⁴⁶

Time and the massive expansion of Ruisko into Western Europe from the Iberian Peninsula over the past 15,000 years have resulted in a complex network of interrelated gene lines whose origin, ancestry and temporal relationships have to be sorted out before we can draw more specific conclusions about the male colonization of the British Isles.
⁴⁷
When all the individual Ruisko gene types are assessed for the similarity of their STR markers, they initially split up neatly into sixteen gene clusters, all having very closely related gene types, and each with a central or modal (most common) gene type. These clusters each have a very clear geographical localization, so it is likely that their modal types represent ancestral founders. I have labelled the clusters R1b-1 to R1b-16 (see Appendix C). All clusters are descended from one of the two main gene types from Spain, the oldest of which I shall discuss next.
⁴⁸
(I should emphasize at this stage that I have extended this cluster approach to the entire dataset, i.e. breaking up Rostov and the Ivan subgroups as well.)

Further analysis revealed that, ultimately, all clusters and their constituent gene types can be traced back to one root type found in cluster R1b-9 which expanded northwards from the Basque refuge around 16,500 years ago;
⁴⁹
for this reason I label it the
Basque Haplotype
(
Figure 3.6b
).
⁵⁰
This root type differs by one STR mutational from its nearest relative, known as the
Atlantic Modal Haplotype
, which resides in cluster R1b-10, and did not expand until much later (see
Chapter 4
).

In fact, the major early expanding male clusters can be divided fairly clearly on the Y tree into those dating in the British Isles from between 16,000 and 13,000 years ago, which derive directly from the Basque Haplotype in R1b-9, and those expanding from Iberia after 11,500 years, which all derive from the Atlantic Modal Haplotype in R1b-10 (see
Figure A4
in Appendix C, and
Figures 3.6b
and
4.5
). The root male gene cluster R1b-9 can be called
Rox
, and his son, R1b-10,
Ruy
; both are listed as Basque names although, curiously, traditional Basque has no word-initial ‘r-’. Rox and his descendants show significant expansions from their refuge at least twice if not three times since the last Ice Age. The first such expansion seems to have kicked off smartly from the southern refuge around 15,600 years ago, since a number of Rox’s sons found in Europe or Britain (or both) date to that time (see
Figure A4
in Appendix C, and
Figures 3.6c
–
3.6f
).
⁵¹
Six of the sixteen British male founding clusters (R1b-4 to R1b-6 and R1b-14 to R1b-16) descend directly from Rox and date from before 13,000 years ago. In total, around 27%
⁵²
of modern British men can claim descent through their fathers from the seven clusters arriving in this early post-LGM period. This is certainly within the bracket of the 25–42% I estimated for maternal descent, but obviously nearer the lower limit of 25%. But even a 30% contribution of Basque Late Upper Palaeolithic male and female ancestors for modern British imposes a completely different balance on our ‘roots’ perspectives.

I shall return to the events taking place after the 13,000-year threshold, but it is likely that this genetic watershed between the initial Late Upper Palaeolithic recolonization period and what came later, during the Mesolithic, is not just a genetic accident. The watershed may reflect the profound climatic reversal that
occurred 13,000 years ago, known as the Younger Dryas Event, a short worldwide freeze-up which ended abruptly around 11,500 years ago with another equally dramatic warm-up (see below).

During the LGM, Ivan (group I) may have diversified in isolated local refuges in the Balkan region, into at least four descendent gene groups,
⁵³
followed by several post-LGM dispersals into Northern Europe. Ivan arrived much later than Rox in northwest Europe and the British Isles. But one minority branch of the Ivan gene group, I1c (named here
Ingert
, a German name to fit his location of highest frequency), appears to have placed at least one of his three clusters in Britain as long ago as did Ruy, around the time of the Younger Dryas. This raises a question whether Ingert arrived in Western Europe and Britain before or after the Younger Dryas, or a bit of both;
⁵⁴
I shall return to this later.

As the ice receded in the north, the earliest Ruisko clusters began to move northwards up the Atlantic coast. As beachcombers, they would have met no English Channel, and no Irish Sea – they could have walked in a straight line from the tip of Brittany across to the south-west coast of Ireland. To the east, the continental shelf was dry land (
Figure 3.3
), and the Channel Islands, Land’s End and the Isles of Scilly were all well inland. The majority of our Atlantic coastal beachcombers would then have been able to walk onwards, round the west coast of Ireland, where the coastline has not changed much since. From there they could have spread straight on towards the Western Isles, Scotland’s north coast and Orkney, which at that time all
formed one landmass, whose west-facing coastlines are also still mostly unchanged. After the initial deglaciation 15,000 years ago, there was still a small ice cap in the Scottish Grampians spreading just across into Antrim in Northern Ireland.

The lower sea level also exposed the bed of the North Sea continental shelf in a straight line across from Scotland to Denmark. The Scottish ice cleared over the next couple of thousand years, and the North Sea coast receded somewhat, allowing access to Aberdeenshire. Although it is possible that both Ireland and Scotland were colonized as early as 15,000 years ago, there is no archaeological evidence for this. The ultimate spread of these gene lines north along the Atlantic fringe to Scotland from Ireland, as evidenced in the genetic record, could have been somewhat delayed.

In spite of this delay, the distributions of the eight earliest Ruisko clusters, in particular the main founder Rox, still actually reflect the extended beachscape and coastline of the Greater British Atlantic coast before the sea level rose to create the English Channel and the Irish Sea split off Ireland from Wales (
Figures 3.3
and 3.6). Largely shunning Wales, and southern and eastern England (except for pockets on the south coast), these early founders distributed themselves round the less changed parts of the old western coastline, including western Ireland and the Western Isles, right up to Scotland and Orkney. They also feature in northern England, for instance in Nottinghamshire, near the Late Upper Palaeolithic site of Creswell Crags (
Figure 3.8
).

The highest rates for Rox, of up to 27% of all gene lines, are found in Scotland, with a flow round to the Scottish east coast to Stonehaven, where 23% are Rox. This distribution perhaps
reflects the coastline at that time (
Figure 3.6b
), while the high rates most likely reflect genetic ‘founding events’ along a beachcombing route – which is not entirely surprising if Scotland was their terminus. Rox is notable by his low rates on the nearby Continental coast (2–5%), which would have been joined to Britain, but well inland at that time and inaccessible to Scottish beachcombers because of the Scandinavian ice cap.

The westerly distribution of Rox is reflected in that of his
derived
branches: cluster R1b-5 features in Ireland, Wales and northern Scotland (
Figure 3.6c
); cluster R1b-16 characterizes north-east Scotland, Ireland and south-west Britain (
Figure 3.6f
). There was some later regional re-expansion and spread during the Mesolithic in several other Rox clusters (see also
Chapter 4
); so cluster R1b-4 features in Scotland and north Wales, being absent from Ireland, while R1b-6 expanded in eastern England and R1b-15 re-expanded in Cornwall, central Wales and, to a lesser extent, in Ireland, Scotland and the Channel Islands. Three R1b-15 sub-clusters focus on the Atlantic fringe regions, each with slightly different emphasis. However, only one of these three actually expanded on arrival (
Figure 3.6e
); one more re-expanded immediately after the Younger Dryas, while a third re-expanded during the Neolithic.
⁵⁵

Cluster R1b-14 is perhaps the most interesting migrant gene line of this early post-LGM period. He has recently been strongly associated with Irishmen with Gaelic surnames, for which reason, and that of his location, he has been given the nickname
Rory
(Gaelic for ‘little king’). Rory’s distribution correlates with that of ancestral Rox,
⁵⁶
missing out on Wales and southern England. But while featuring in Scotland and Orkney in the far north, Rory is characteristic of Ireland, being found at
rates of up to 30% there (
Figure 3.6d
). This distribution again fits with the dry Irish Sea and English Channel at that time.

As I have included the same dataset that gave rise to this dramatic claim in my analysis, it is natural that I should make the same gen etic association with the ‘Gaelic’ sub-sample, and that I should find the frequency of Rory to be higher than in other Irish samples unselected for surname. However, my re-dating of Rory’s arrival in the British Isles to around 16,000–15,000 years ago is quite differ ent from the date arrived at in the original report in
Nature
by Emmeline Hill and colleagues at Trinity College Dublin, which suggested a Neolithic age. Before my new dating is taken as a suggestion that Gaelic arrived in Ireland 15,000 years ago, I should point out that there are much simpler explanations for the association.
⁵⁷
I shall come back to this and to the resurgence of three descendent clusters of Rory during the Mesolithic and Neolithic, but for the time being it should be clear that their male ancestors arrived in the west long before then.

Apart from Rory, several other derived Rox clusters (R1b-4, 5, 6 and 15) re-expanded and split up again thousands of years later, and will be revisited in
Chapters 4
and
5
, when I come to discuss the Mesolithic and Neolithic periods. The fact that their ancestors arrived so early obviously reduces estimates of the size of later immigration.