The Origins of the British: The New Prehistory of Britain (77 page)

35
. Amos et al. (2006).

36
. i.e with a sufficient number of new, mutationally novel genetic haplotypes related to an identifiable founding line.

37
. The numerous (278) singleton haplotypes were all assigned to their statistically/phylogeographically most likely source. Haplotypes that were unique to the
British Isles were assigned as such, and in nearly every case could be assigned to uniquely British clusters over 3,000 years old. The identification of the latter separate class is the main reason why the overall degree of intrusion from Continental sources is greater than that for the haplotype matches. I tabulated 490 haplotypes belonging to 40 clusters from 11 haplogroups. There were 278 singleton haplotypes, leaving 212 types with potential for matching. Statistical testing has not been performed in this analysis, since error calculation would be difficult.

38
. Present study: overall 29.6% (319/1,077); lowest is Faversham, Kent with 14.5% (8/55); highest is Fakenham, Norfolk with 41.5% (22/53). Admixture proportion estimate for England from Capelli et al. (2003) ‘over 40%’, from North Germany/Denmark 37.5%, from Norway 24.3%. These figures are not mutually exclusive, and the overall intrusion from north-west Europe would not be as much as their sum, but certainly more than 40%.

39
. ‘Anglo-Saxon intrusions’ to British Isles: England (including Cornwall) 5.5% (59/1,077), British Isles overall 3.8% (79/2,082), Isle of Man 4.5% (2/44), Scottish Isles 2.9% (10/344), Scotland 1.7% (3/178), Wales 1.5% (3/201), Ireland 0.8% (2/238).

40
. Härke (2002), p.150.

41
. Procopius,
History of the Wars
8.20.9–10.

42
. J1a (J-16231); see Forster et al. (2004).

43
. To dissociate Dark Ages Continental ‘Anglo-Saxon’ women from English women, Forster et al. (2004) note that ‘the Low-German-speaking (“Saxon”) areas of the North German Plain harbour a “Saxon” mtDNA marker H/16189 at about 25 per cent … which is rare in England where there is a frequency of only 3.5 per cent … This low proportion indicates a contribution of zero to maximally 25 per cent of north German women to the native population of England.’

1
. Cunliffe (2004), p. 489.

2
. Cunliffe (2004), pp. 488–93.

3
. i.e. as opposed to Neolithic contacts, which can be inferred from the genetic record discussed in this book.

4
. Wrenn and Bolton (1988), pp. 41–5.

5
. The three historic clusters are I1a-3, British age 1,200 years (rooted on Ht. 398,
n
= 68, SD ±400); R1b-8a, age 2,137 years (rooted on Ht. 152,
n
= 12, SD ±1,510); and R1a1-3b, British age 1,830 years (rooted on Ht. 96,
n
= 21, SD ±1,830). Overall rate in British Isles 4.9% (101/2,082).

6
. Danish matches 4.4% (91/2,082), Anglo-Saxon matches 3.8% (79/2,082).

7
. Fakenham has the highest percentage intrusion (20%) of I1a-3 of all British geographical samples.

8
. I1a-3 British age 1,200 years (rooted on Ht. 398,
n
= 68, SD ±400). Distribution: particularly in Norfolk, with six individuals of Ht. 398 in Fakenham.

9
. The root type of I1a-3, haplotype 398, is identical to the common haplotype 2.49 in the dataset of Wilson et al. (2001).

10
. Neither of the two founding clusters of the historical period from Scandinavia (total
n
= 93/2,082, 4.5%) is particularly associated with Norfolk except for I1a-3 root haplotype 398, which has six representatives in Fakenham and one in Sheringham (R1a1-3b has one representative in Sheringham as well). I1a-3 otherwise tends to follow the full Viking dispersal, while R1a1-3b is distributed more to the north of Britain and its islands and specifically with Norwegian Viking dispersals.

11
. 18.9% (10/53). For haplotype matching, Ht. 398 is assigned to Denmark rather than Norway on the basis of overall frequency; this, however, splits cluster I1a-3.

12
. (1) R1a1-3c (total in cluster
n
= 10), age in Britain 3,660 years (rooted on Ht. 93,
n
= 7, SD ±3,660); (2) R1a1-3a, 4,080 years (rooted on Ht. 95,
n
= 22, SD ±2,260); (3) R1a1-1 (total
n
= 10), undated and derives from northern Gemany/Denmark and Frisia; (4) R1a1-2 (total in cluster
n
= 52), 5,600 years (rooted on Ht. 87,
n
= 38, SD ±2,650); (5) R1a1-2a, age in Britain 7,700 years (rooted on Ht. 86,
n
= 10, SD ±5,440), derives from Norway; (6) R1a1-2b, age in Norway 5,700 years (rooted on Ht. 87,
n
= 38, SD ±2,160).

‘Two Bronze Age clusters’, respectively (1) and (2) above: R1a1-3c centres on the Isle of Man; R1a1-3a centres on Shetland, Orkney and the Western Isles.

13
. R1a1-3b (total
n
= 22), age in British Isles 1,830 years (rooted on Ht. 96,
n
= 21, SD ±1,831). Source cluster in Norway: age 3,200 years (rooted on Ht. 96,
n
= 20, SD ±1,920).

14
. 17.5% (22/126).

15
. R1b-8a, age 2,137 years (rooted on Ht. 152,
n
= 12, SD ±1,510).

16
. R1b haplogroup in Shetland 63.7% (86/135); ‘indigenous/British’ Shetland haplotypes 71.8% (97/135). Other studies have arrived at similar figures: for instance, Gooadcre et al. (2005) found a patrilineal ancestry admixture in Shetland of 55.5% Irish/British. They also found an equally balanced male– female Scandinavian admixture in Shetland and Orkney, unlike the Western Isles and Iceland, in each of which the male component was twice as large. They suggest that this implies that Orkney and Shetland may have been settled from Scandinavia by families (see p. 398). See also table 4 in Helgason et al. (2001), who estimate a 64.5% ‘Gaelic’ mtDNA component in Orkney and a 62.5% ‘Gaelic’ component in Iceland.

17
. Admixture estimates from Goodacre et al. (2005), table 3. See also Helgason et al. (2001), table 4, where the Norse : Gaelic ancestral mtDNA ratio in Iceland is given as 3:5. For earlier estimates of male/female, British/Scandinavian components in Iceland, see Helgason et al. (2000a,b). Male slaves are recorded in the sagas, but 25% seems a large figure.

18
. R1b in Iceland: 37.0% (67/181).

19
. R1b in Trondheim: 28.9% (26/90); I1a rates are 24.4% (22/90) in Trondheim and 29.3% (53/181) in Iceland. More specifically, I1a-3 constitutes 76.9% (40/52) of I1a in Iceland and only 18.2% (4/22) of I1a in Trondheim. The highest equivalent figure elsewhere in Scandinavia is in Bergen, farther south: 52.6% (20/38). These figures suggest both a greater contribution from southern Scandinavia and a strong I1a-3 founder event in Iceland. Orkney and Shetland show no such I1a-3 founding event; rather they have a diverse mix of I1a and R1a1 haplotypes (in addition to indigenous R1b) belonging to lineages which arrived earlier in Britain. The same applies to a lesser extent to the Western Isles, where Viking founder events (I1a-3 and R1a1-3b) account collectively for 45.4% (10/22) of I1a and R1a1 types, but there are also roughly equal representatives of nine Neolithic Scandinavian founders, in particular I1a-5, which has its main British presence there (present study). For genetic drift in Iceland, see also Helgason et al. (2003).

20
. Danish haplotype matches in Shetland and Orkney: 3% (8/256). For a graphic display of the relative importance of mtDNA from northern Britain and southern Norway in the colonization of Iceland, see Forster et al. (2004), figure 8.4.

21
. 13.3% (18/135).

22
. Goodacre et al. (2005), table 3.

23
. Cunliffe (2004), p. 502; but see Smith (2001), Crawford (1981/2) and Wainwright (1962).

24
.
Orkneyinga Saga
(12th century), translation in Cunliffe (2004), p. 493.

25
. Present study.

26
. Goodacre et al. (2005). The change in sex ratio in the Western Isles may, of course, partly reflect distance from Scandinavia. The Western Isles have roughly equal Neolithic and Viking male intrusions from Scandinavia – see note 16 in this chapter.

27
. e.g. Smith (2001).

28
. But see also Forsyth (1995).

29
. Norwegian haplotype intrusion: to Shetland 20% (27/135), Orkney 17% (21/121), Western Isles 20% (18/88), Isle of Man 16% (10/62), Channel Islands 6.2% (8/128).

30
. Norwegian haplotype intrusion based on exact matches: Oban 12% (5/42), Durness 13.7% (7/51). For the western British locations (excluding the Channel Islands) I have combined southern and northern Scandinavian haplotype matches as ‘Norwegian’ since, owing to the distribution of I1a-3, these would otherwise appear as ‘Danish’, which is historically less likely. Results are based on the present study and are generally lower than estimated by Goodacre et al. (2005), which is to be expected, given the use of exact matching.

31
. 6.2% (129/2,082) (ignoring the Danish-to-Norwegian adjustment made in the previous note).

32
. Härke (2002) has suggested 2–4%, invasion based on a population denominator of 1–2 million. The latter figure is rather larger than estimated elsewhere for the late Anglo-Saxon period, so 2–4% could be nearer 4–6%, although archaeological estimates during this period are really guestimates. An alternative estimate for the ninth-century Anglo-Saxon population is 750,000–850,000: see Coleman and Salt (1992).

33
. Two estimates: 1.75 and 1.1–2.6 million: see Coleman and Salt (1992), table 1.1 and pp. 2–8.

34
. Härke (2002).

35
. Bragg (2003), p. 42.

36
. Golding (2001), pp. 61–2.

37
. Hinde (2004).

38
. There are other sources of demographic measurement and estimation from this period.

39
. Bragg (2003).

40
. Douglas (1942).

41
. Cunliffe (2004), p. 498.

1
. Much of the best detail of colonization of the British Isles comes from my reanalysis of the Y-chromosome evidence; so most of the figures summarized here relate to that evidence. But published work on mitochondrial DNA inherited from our mothers generally supports that picture.

2
. British Isles 72.6% (1,511/2,082), Irish 87.9% (233/265), Welsh 81.1% (180/222), Cornwall 78.8% (41/52), Scotland and islands 70.1% (366/522), England 67.7% (691/1,021).

3
. Pryor (2004), pp. 143–4.

4
. Forster et al. (2006).

5
. National Statistics, 2001 Census,
http://www.statistics.gov.uk/census2001/profiles/commentaries/ethnicity.asp
>.

6
. Gildas,
De excidio Brittaniae
3.27.

1
. If one takes the HVS 2 segment of the control region normally studied – see methods in Forster et al. (1996).

2
. See methods in Forster et al. (1996) and Saillard et al. (2000).

3
. See e.g. Forster et al. (2000), Zhivotovsky (2001) and Zhivotovsky et al. (2004).

4
. See methods and calibration in Forster et al. (2000).

5
. For the consensus nomenclature, see The Y-Chromosome Consortium (2002).

6
. See
figure 3.3
in Oppenheimer (2003).