The Origins of the British: The New Prehistory of Britain (69 page)

62
. Forster and Toth (2003).

63
. But see the critique in McMahon and McMahon (2006).

64
. Forster and Toth (2003).

65
. 3200
BC
± 1,500 years (Forster and Toth 2003).

66
. 8100
BC
± 1,900 years (Forster and Toth 2003).

67
. Renfrew (1989), Forster and Toth (2003).

68
. Based on the Roll of the Kings of Ireland after Ireland was divided up between the invading Milesian-Gaelic lords Éber and Éremón, as given in O’Donovan (1848–51) and Macalister (1938–56).

69
. Roll of the Kings of Ireland, taken from
http://www.ucc.ie/celt/online/T100005A
>.

70
. Roll of the Kings of Ireland, taken from
http://www.ucc.ie/celt/online/T100005A
>, Annal M3266.1.

71
. According to O’Rahilly (1946), the Cruithne or Priteni arrived c.700–500
BC
, the Builg or Érainn c.500
BC
, and the Lagin, the Domnainn and the Gálioin, c.300
BC
.

72
. Cunliffe (2004) – discussed in
Part 2
of this book.

73
.
Lebor Gabála Érenn
(Macalister 1938–56).

74
.
Lebor Gabála Érenn
(Macalister 1938–56).

75
. Gildas,
De excidio Britanniae
.

76
. Cunliffe (2004), figure 8.3.

77
. Cunliffe (1988), figure 15.

78
. Cunliffe (2004), pp. 222–6, 296.

79
. Cunliffe (2004), pp. 225–6.

80
. Cunliffe (2004), p. 243.

81
. Cunliffe (2004), p. 296.

82
. Renfrew (1989).

1
. Sims-Williams (1998a, b, 2003).

2
. Gray and Atkinson (2003), Atkinson et al. (2005).

3
. 5,200 (±1,500) years (Forster and Toth 2003).

4
. 6,100 years using Dyen’s cognate data: Gray and Atkinson (2003); 5,600 years using Ringe’s cognate data: Atkinson et al. (2005). The Cambridge–Zurich group (Forster and Toth 2003) cannot estimate the date of this node (the root of celtic), but bracket it above by an estimate of the date of the break-up of Indo-European as a whole in Europe at about 10,000 years ago (10,100 ± 1,900), and below by a date of 5,200 (±1,500) years for the fragmentation of Gaulish, Goidelic and Brythonic from their most recent common ancestor.

1
. Oppenheimer (2003).

2
. See figure 6 in Barton et al. (2003). For the British Isles at the Late Glacial see also Barton (1999).

3
. Figure 6 in Barton et al. (2003); see also figure 1 in Gamble et al. (2004).

4
. i.e. Solutrean and Magdalenian/Badegoulian – Gamble et al. (2004); see pp. 108–9.

5
. Gamble et al. (2004), p. 247.

6
. Oppenheimer (2003), pp. 248–53 and Figure 6.2.

7
. Oppenheimer (2003), pp. 113, 276. I qualified this observation by adding ‘at least until the last five hundred years’, indicating that the reason for this conservatism could be the truism that it is easier to invade unoccupied than occupied lands – unless one has guns.

8
. Archaeological: Barton et al. (2003) and Gamble et al. (2004). Genetic: see pp. 111–13, the expansion is greater when estimated from mtDNA than from the Y-chromosome.

9
. Barton et al. (2003).

10
. Barton et al. (2003).

11
. Barton et al. (2003).

12
. Ripoll et al. (2004). See also images at
http://www.bradshawfoundation.com/creswell.html
>.

13
. Oppenheimer (2003), Appendix 1, pp. 365–9.

14
. Table 4 in Richards et al. (2000).

15
. Present study (i.e. my own re-analysis on a pooled dataset carried out for this book: see the section ‘Re-analysis of the Y-chromosome data’ later in this chapter, and the Appendices).

16
. 52.7% (±4.5%): table 5 in Richards et al. (2000).

17
. 58.8% (±4.7%): table 5 in Richards et al. (2000).

18
. Semino et al. (2000).

19
. Diamond and Bellwood (2003).

20
. Oppenheimer (2003), pp. 214–15.

21
. Gamble et al. (2004).

22
. Gamble et al. (2004).

23
. Gamble et al. (2004); for other genetic evidence see also Torroni et al. (1998) and Pereira et al. (2005).

24
. i.e. during the Early Upper Palaeolithic; see Oppenheimer (2003), pp. 130–41, and Richards et al. (2000).

25
. Oppenheimer (2003), pp. 144–50, and Torroni et al. (1998, 2001).

26
. When H was characterized using the short-sequence (i.e. HVS1 and RFLP) tools of the 1990s, the branch was monolithic and poorly resolved.

27
. Torroni et al. (1998).

28
. Torroni et al (2001).

29
. Although haplogroup V’s highest population frequency (52%) is among the Skolt Saami of Norway, followed by 12% among the Basques, the former high rate, lacking diversity, is almost certainly a founder effect. This would have resulted from a small founding population, possibly carrying only V and a couple of others from the south-west, which then amplified V by drift and growth. Her young age among the Saami and the complex aspects of her phylogeography suggest that she may have reached the Fenno-Scandian Peninsula more recently via a circuitous route through Central Europe (Torroni et al. 2001, Tambets et al. 2004).

30
. Torroni et al. (2001), figure 4.

31
. It should be noted, however, that while H1 seems to date to the Late Upper Palaeolithic, c.14,000 years ago, the smaller H3 group appears to be younger, possibly post-Younger Dryas, and is dated to c.9,000 years ago. H1: 14,000 years ago (SE ±4,000) using coding-region data, and c.16,000 years ago (SE ±3,500) using HVS-I; H3: c.9,000 years ago (SE ±3,000) based on coding-region data and c.11,000 years ago (SE ±3,000) using HVS-I. See Pereira et al. (2005).

32
. As some measure of this, in the Founder Analysis published in 2000, Martin Richards and colleagues inferred that up to 53% of today’s female lines arrived in north-west Europe (for ‘north-west Europe’ see Richards et al. (2000), table 5) during the Late Upper Palaeolithic, shortly after 16,000 years ago. In this estimate, they included most of H anyway, which covered a large proportion of what were at that time unresolved H types. Their figure 1 shows e.g. 38% + 3.9% = c.42% of all lineages for the whole of Europe in the LUP contributed by H, plus a further 10% addition from other gene groups ( T, T2 and K); the last three groups, however, overlap with the Mesolithic.

33
. Oppenheimer (2003), pp. 150–4, 250–3.

34
. R1b designation in The Y Chromosome Consortium (2002). It is equivalent to Eu-18 in Semino et al. (2000).

35
. Distribution of ‘R1b’ (Ruisko) haplogroup in Europe: present study, Oppenheimer (2003), Underhill et al. (2000), Wilson et al. (2001), Capelli et al. (2003), Semino et al. (2000).

36
. Unless otherwise indicated in this note, data are collated from table 3 in Tambets et al. (2004): Spanish Basque 88.9% and Dutch 70.4% (Semino et al. 2000); French 59%; Germans 50% (as in Semino et al. (2000) – note the typographic error of ‘0’ in published table 3, Tambets et al. (2004)); Danes 36.1%; Norwegians 27.8%; Swedes 22.0%; northern Russians 21.3%; Estonians 18.2%; Poles 16.1%; Latvians 7.0% and Finns 0%.

37
. In Central Europe and the Balkans there are varying mixtures of the two European members of the R clan, R1a1 and R1b. In the west, including the British Isles, Germany, the Netherlands, Belgium and France, it is more straightforward: R1b dominates, especially south and west of the Rhine. R1a1, defined by a unique mutation known and referred to in my last book (Oppenheimer 2003) as M17, is nicknamed Rostov in this book.

38
. The overall age of R1a1 (rooted on Ht. 87/R1a1–2 in database) is 19,500 years (
n
= 291, SD ±5,540) (see Appendix C). This implies post-LGM re-expansion.
For pre-LGM history, note Ruslan/M17 in Oppenheimer (2003), p. 152, which is likely of great antiquity in South Asia. Data from several sources: Tambets et al. (2004), table 3; Rosser et al. (2004), Hg1 in table 1, and figures 3 and 4; Weale et al. (2002), Hg1 in tables 2 and 3; Cappelli et al. (2003), Hg R1xR1a1 in table 1; Semino et al. (2000), Hg Eu18 in table 1.

39
. Age estimates for subgroup I1c from table 3 in Rootsi et al. (2004): time since sub-clade divergence 14,600 (±3,800) years, age of STR variation 13,200 (±2,700) years, time since population divergence 11,200 (±2,300) years. The second and third figures agree with my own estimate for I1c in Western Europe of 12,000–14,000 years, although I estimate sub-clade divergence as older, at 21,000 years.

40
. For example, eleven haplogroups as given in Capelli et al. (2003); these are extended using STR clusters to fourteen. Relevant published Y-chromosome datasets for mainland Western Europe include all those cited by Tambets et al. (2004), and also Rosser et al. (2000), who subdivide the dataset into twelve haplogroups, but only seven of these are relevant to the British Isles; Scozzari et al. (2001), who attempt to break up other haplogroups using STR markers, but unfortunately not the R haplogroup; Gusma et al. (2003), who look specifically at STR haplotypes within the R1b haplogroup and find geographical structure, but do not proceed to create any phylogeny; and Roewer et al. (2005), who look at STR variation geographically but do not use it to resolve the relevant haplogroups.