The Origins of the British: The New Prehistory of Britain (70 page)

41
. Key papers for the British as opposed to Continental data are Wilson et al. (2001), Capelli et al. (2003), Weale et al. (2002) and Hill et al. (2000). Additional indigenous Orkney and Shetland STR data courtesy Jim Wilson (from D.K. Faux and J. Wilson, unpublished) at
http://www.davidfaux.org/shetlandislandsY–DNA3
>. Methods and protocols used for typing STR Y-chromosome microsatellites DYS388, 393, 392, 19, 390 and 391 can be found in Thomas et al. (1999) and Goodacre et al. (2005).

42
. The main papers listing British unique event polymorphism (UEP) and STR data used in this book are as given in the previous note.

43
. This was a difficult decision, since in spite of the volume of data, the information provided by STR markers, unlike the UEPs used to identify haplogroups, is ambiguous and difficult to resolve even by using network software programs. The differences between the ‘bi-allelic’ haplogroup markers (i.e. dichotomous unambiguous ‘UEPs’) and STRs is that the latter are unstable, rapidly mutating,
multi-allelic and thus potentially ambiguous in interpretation. The amount of work involved in network building for such large haplogroups is horrendous, without the availability of more UEPs, which is why the field has previously stayed fallow. It was necessary not only to resolve branches but also to date them. The whole task took me about six months.

44
. For source papers on the British and other Western European populations, see methods in the Appendices. For comparison I have also referred to papers giving additional STR data on Ireland and Iceland (Helgason et al. 2000a,b, 2003), other European areas including various parts of Spain (Galicia, Valencia, and additional Basque: Brion et al. (2003, 2004)); Catalan (Bosch et al. 2001), the Saami (Tambets et al. 2004), and Hb ‘I’ in the Balkans and the rest of Europe (Rootsi et al. 2004). Wilson et al. (2001) also have data for Friesland, Basques, Syria and Turkey. Weale et al. (2002) also include Friesland and Norway.

45
. When all the information is collected together, the 1,947 samples with the R1b Y group can be split into 180 individual STR haplotypes of varying frequency. Two of these haplotypes (Ht. 157 and Ht. 155 in the present study – see Appendix C) are the most common in the northern Atlantic coastal region, and are distributed throughout Western Europe, accounting on their own for respectively 15% and 9% of all male gene types in the entire dataset that I use. One paper has even given the former founder type the grand name of ‘Atlantic Modal Haplotype’ (AMH) (Wilson et al. (2002); see also Capelli et al. (2003), who also use the term ‘AMH’).

Note the rates of two most common Atlantic founding lines: 15% (467/3,084, Ht. 157 in the present study) and 9% (272/3,084, Ht. 155 in the present study), where 3,084 is the number of individuals in the entire dataset in my analysis, and the former percentage (i.e. 15%) refers to the AMH type.

46
. Unfortunately, in spite of the plethora of papers recording British male genes, there are none which report having tested for the R1b type-marker, P25. However, they have all tested for R1a1, and most have done enough other tests to identify R1b types of the British Isles to the exclusion of other R groups. See note 38 for references.

47
. I have used a similar approach to that described in Martin Richards’ European mitochondrial DNA paper as Founder Analysis (Richards et al. 2000).

48
. For the methods, see Appendix C. In this preliminary description I have labelled these clusters R1b-1 to R1b-16 according to their rank in the sorting process. The geographical distribution of these clusters and their individual members
helps a lot in tracing the process of that initial recolonization and of subsequent migrations from Iberia. My re-analysis of the Y-tree confirmed the validity and integrity of these clusters as real branches, with minor misclassifications and rearrangements mainly found in clusters R1b-1 and R1b-16. In my subsequent discussion, I therefore use this 1–16 numbering convention, but with all the corrections included.

49
. R1b-9 root: Ht. 155/6. R1b-9’s age in the Iberian refuge: 20,600 years (rooted on Ht. 155/R1b-9,
n
= 138, SD ±6,830); overall R1b-9 age in database, but excluding R1b-10 and descendants: 19,200 years (
n
= 513, SD ±5,790); northward expansion from the refuge towards the British Isles dated to 15,600 years ago (
n
= 1,513, SD ±4,800).

50
. Ht. 155 and 156 in present study (i.e. = R1b-9 = Basque Haplotype).

51
. For archaeological dating of early re-expansion, see note 68.

52
. 27.4% = 570/2,082 in the British database (i.e. the British part of the present study).

53
. I1a–c and I* in Rootsi et al. (2004).

54
. See note 39.

55
. R1b-15a, age 11,539 years (rooted on Ht. 253,
n
= 10, SD ±6,014); R1b-15b, age 5,729 years (rooted on Ht. 250,
n
= 46, SD ±3,037); R1b-15c, age 14,958 years (rooted on Ht. 247,
n
= 30, SD ±4,623).

56
. Hill et al. (2000). The distribution of R1b-14 correlates with that of R1b-9: regression (correlation) of R1b-14 with R1b-9 in the present study:
y
= 0.4385
x
+ 0.029;
R
= 0.376.

57
. Age of R1b-14 in the British Isles: 15,760 years (
n
= 170, SD ±8,440).

58
. I1c: maximum frequency in Germany 12.5%, Netherlands 10% (table 1 and figure 1D in Rootsi et al. 2004). For eastern England, the maximum in York, for example, is 13% (present study).

59
. My estimated age for I1c in my dataset as a whole (rooted on Ht. 346) is 20,800 years (SD ±4,550). See Appendix C.

60
. I1c-1: 12,800 years (
n
= 12, SD ±7,400), I1c-2: 14,200 years (
n
= 19, SD ±3,880) and I1c-3: 11,700 years (
n
= 25, SD ±5,470) – see Appendix C.

61
. Two-thirds of I1c types in Britain are unique. This is an unusual picture for any haplogroup in this dataset, and is a measure of its antiquity there reflected in the ages given in the previous note.

62
. Vennemann (2003).

63
. Barton et al. (2003), figure 1. Note that unless otherwise stated, these dates are corrected or ‘calibrated’ and represent calendar years.

64
. Barton et al. (2003).

65
. Barton et al. (2003).

66
. Oppenheimer (1998), pp. 29–32.

67
. Figures 1 and 4 in Gamble et al. (2004), and figure 6 in Barton et al. (2003).

68
. Figure 1 in Barton et al. (2003). Note that not all these types of human evidence are present simultaneously at the same sites during these climatic phases, possibly because of changes in the locations of humans’ settlement and their lifestyle.

69
. Oppenheimer (2003), pp. 248–53 and Figure 6.2.

70
. Richards et al. (2000) and present study (see Appendix and Figure A3).

71
. Oppenheimer (1998), pp. 30–48, figure 1.

72
. Stuart and Lister (2001).

73
. Schirrmeister et al. (2002): ‘After that time [LGM, 24,000–18,000 years ago] fossil insect assemblages point to a far more continental climate with much warmer summers (Sher et al. 2001). The decrease of mammoth fossils dated to 20 kyr–15 kyr
BP
and their subsequent increase indicates less favorable environmental conditions for large animals during the first substage, and better conditions during the second ….’

74
. Oppenheimer (1998), pp. 38, 228, 234, 255, 260.

1
. Oppenheimer (2003), pp. 24–7, 76–80.

2
. Barton and Roberts (2004).

3
. Oppenheimer (1998), pp. 29–32.

4
. Cunliffe (2004), pp. 119–38.

5
. Cunliffe (2004), pp. 123–6.

6
. Cunliffe (2004), pp. 123–6.

7
. Cunliffe (2004), pp. 126–38.

8
. Cunliffe (2004), pp. 126–38.

9
. Cunliffe (2004), pp. 138–9.

10
. On the maternal Mesolithic contribution and Founder Analysis, see Richards et al. (2000). Note that there are problems in trying to match the Founder Analysis paper too closely to the peopling of the British Isles. First, it measures intrusion
only of putative Near Eastern lineages, which ignores potential repeat Mesolithic/Neolithic expansions from the Basque refuge. Second, apportioning founding lineages between the LUP and the Mesolithic is affected not only by the foregoing (i.e. multiple expansions and/or non-Near Eastern lineages) but also by poor resolution of haplogroup H. Richards et al. (2000) acknowledge these problems; for further discussion of relative Mesolithic component see Richards (2003) and McEvoy et al. (2004).

11
. For instance, a large part of the LUP component in that Founder Analysis (Richards et al. 2000) was made up of unresolved H types, so potentially recurrent migrations from the same Iberian source up the Atlantic coast (i.e. not from the Near East) could not be excluded. Such issues were mentioned in that paper, and Richards has since discussed the possible underestimation of the size of West European Mesolithic populations (Richards 2003) and, as mentioned, has subsequently resolved the large H group somewhat more (Pereira et al. 2005).

12
. Pereira et al. (2005), tables 1 and 2. Ages estimated using coding-region data: H3, 9,000 years (SE ±3,000), also H2, 11,600 years (SE ±8,000), H5a, 8,000 years (SE ±4,000), but also H4, 7,000 years (SE ±4,000). In all, these four putative Mesolithic lineages account for 10.8% of Irish and 10.6% of UK maternal lines.

13
. T, T2, T4 and K dates straddling the YD: Richards et al. (2000), figure 1. T, T2, T4 and K found in the Basque Country, Spain and Cornwall: Richards et al. (1996). See also discussions on the size of the Mesolithic component in Richards (2003) and McEvoy et al. (2004).

14
. British 17.57% (366/2,082), Basque 19.0% (15/79); age of R1b-10 cluster and derivatives overall in British Isles: 9,800 years (
n
= 693, SD ±3,410). Present study.

15
. When Principal Components Analysis (see Chapters 6 and 11) is performed, the PC plot for R1b-10 links Wales strongly with the Basque Country. Using the coordinates of each cluster, it is possible to plot the geographical region in the British Isles of highest relative frequency for each cluster.

16
. 26.1% (23/88).

17
. 10.5% (8/76) and 17.8% (26/146).

18
. 25.6% (10/39) and 21.6% (19/58).

19
. Respectively: R1b-11 and 13: 9.6% (200/2,082); R1b-7, 8 and 12: 14.9% (311/2,082); and R1b-1 to R1b-3: 2.2% (46/2,082). Ages: age of combined
cluster R1b-11 to R1b-13: 9,500 years (rooted in Ht. 172,
n
= 225, SD ±4,740); R1b-2a, 8,500 years (
n
= 9, SD 7,264); R1b-2b, 10,300 years (
n
= 30, SD ±5,040); R1b-7 and R1b-8 – see notes 54–56 to
Chapter 5
.

20
. R1b-13: frequency 3.8% (79/2,082), age 7,800 years (SD ±4,090).

21
. R1b-13 in North Wales (Llangefni) 11.4% (10/88), and Cumbria (Penrith) 8.9% (8/90). Present study.

22
. Age of R1b-11 to 13: 9,540 years (
n
= 225, SD ±4,740); R1b-13, 7,790 years (
n
= 79, SD ±4090); R1b-11, 4,560 years (
n
= 121, SD ±3,370); R1b-12, 4,620 years (
n
= 25, SD ±4,140). Rate of R1b-11: 5.8% (121/2,082).

23
. 10.8% (225/2,082).