You are on page 1of 10

Diet and ethnicity during the Viking colonization

of northern Scotland: evidence from fish bones


and stable carbon isotopes
JAMES H. BARRETT,ROELFP. BEUKENS& REBECCA
A. NICHOLSON*
Diet and ethnicity are strongly related. Recent work on fish-bone ratios and stable carbon
isotopes suggest fhat the Vikings increased the fish contribution to the diet of Orkney and
Shetland by a greater investment in deep-sea fishing.

Key-words: cod, Vikings, fish bones, Orkney, Shetland, carbon isotopes

Introduction thesis (built on observations first made by


The legitimacy of migration in archaeological Wheeler 1977) that a more maritime-oriented
explanation waned in the mid to late 20th cen- subsistence strategy was introduced to Viking
tury, at least in the Anglophone scholarly tradi- Age northern Scotland by Norse farmer-fisher-
tion (Hiirke 1998: 19).Revisionist interpretations men (Barrett et al. 1999: 369-70). The evidence
of Viking colonization in northern Scotland includes zooarchaeological assemblages and
emerged in step with this trend. In the 1950s, new stable carbon isotope data for human bone
F.T. Wainwright (1962: 125-6) was confident of ‘Pictish’ (c. 300-800 AD) and ‘Viking Age’
that ‘Scandinaviansettlement amounted to amass- (c. 800-1050 AD) date (see FIG~JRE
migration. The Scandinavians arrived in num-
bers sufficient to overwhelm the earlier inhabitants Methods and potential biases
politically, socially, culturally and linguistically’. Zooarchaeology
By the mid 1970s, however, AnnaRitchie (1974; A shift in the role of marine resources at the
see also Ritchie 1993: 25-6) proposed what can Pictish-Viking Age transition was discussed pre-
be described as the integration or acculturation viously in a study of the northern Scottish fish-
model (see Morris 1996: 77) - which assumes bone record from the Neolithic to the Middle
considerable continuity of native culture -origi- Ages (Barrett et al. 1999).To address this ques-
nally based on the discovery of Pictish artefacts tion in detail, however, it is useful to focus
in presumed Viking Age houses. This model of exclusively on fish assemblages of Pictish and
cultural continuity - with implicit or explicit Viking Age date.
acceptance of a Norse speaking Blite stratum - TABLE1 includes data regarding a series of
has since become quite widely accepted for at variables chosen to isolate trends in both the
least the earliest phase of Scandinavian immi- intensity and kind of fishing conducted:
gration (e.g. Buteux 1997: 2 6 1 4 ; Hunter 1997: the total number of identified fish specimens
249-53; Sharples & Parker Pearson 1999). (NISP) recovered,
Given a renewed appreciation of migration the ratio of identified fish to mammal bone,
in the wider archaeological community (e.g.
Anthony 1990; Chapman & Hamerow 1997; 1 The conflation of cultural and chronological labels is
Burmeister ZOOO), however, it is appropriate to unavoidable given their long tradition of use. Late Iron Age
is a n alternative term for the Pictish period in Scotland
reassess the acculturation model. This paper (e.g. Foster 1990: 143), but is also imperfect in a wider
will do so from a single perspective - that of European context where the same terminology can refer to
subsistence and diet. It revisits a recent hypo- the earlier 01later centuries.
* Barrett, Department of Archaeology, The King’s Manor, University of York, York yo1 7EP, England. jhb58york.ac.uk
Beukens, IsoTrace Laboratory, University of Toronto, Toronto, ONT M5S 1A7,Canada. roelf.beukens@utoronto.ca
Nicholson, Department of Archaeological Sciences, University of Bradford, Bradford BD7 l D P , England.
r.a.nicholson@bradford.ac.uk
Received 24 May 2000, accepted 27 July 2000, revised 13 December 2000
ANTIQUITY75 (2001): 145-54
146 JAMES H. BARRETT, ROELF P. BEUKENS & REBECCA A. NICHOLSON

SHETLAND

B
0
D

FIGURE1. Location of
sites mentioned.
1 Sealloway;
2 Old Scatness;
3 St Boniface;
4 Pool;
5 Westness;
6 Brough Road sites
and Buckquoy;
7 Saevar Howe;
a Howe;
9 NewarkBay;
10 Smoo;
11 Freswick Links.

the ratio of cod family (Gadidae) to other to measure the relative importance of ‘offshore’
fishes, versus littoral fishing, require some explana-
the ratio of cod (Gadus morhua) to saithe tion. During recent centuries in northern Scot-
(Pollachius virens) and land saithe of a size common in the assemblages
the ratio of ling (Molvu molva) and torsk under consideration (the majority of bones rep-
(Brosme brosme) to rocklings (Ciliata or resent individuals between 100 and 400 mm
Gaidropsurus species), wrasse (Labridae) total length, see Barrett et al. 1999: 361-2) were
and cottids (Cottidae). caught from the shore and from boats in shal-
The total NISP and the ratio of fish to mam- low water using a net or simple rod and line
mal are both intended to measure the absolute (Fenton 1978: 527-30; Baldwin 1982). Con-
intensity of fishing activity and, by implica- versely, cod fishing, often for larger individu-
tion, the role of fish in the diet. Conversely, als (the majority of specimens from the Viking
the ratio of cod family to other fishes is included Age assemblages considered represent fish over
to detect any shift in prey selection towards 800 mm total length, see Barrett et al. 1999:
the gadid species known to have been of para- 361-2), was conducted principally from boats
mount importance in Iron Age, Viking Age and in open water using weighted ‘hand lines’ of
medieval Norway (Enghoff 1999: 55; Perdikaris c. 60 fathoms or ‘long lines’ of even greater length
1999).The last two ratios, which are intended (e.g. Goodlad 1971: 107-9; 130). If these pat-
period & assemblage1 proportion mesh fish fish: cod cod: ling& references
sieved size NISP mammal family: saithe torsk
other rocklings,
fish wrasse &
cottids2 E
m
4
>
Pic fish 2,
U
Brough Rd Area 3 unknown 10 mm 85 1.25 8.44 0.07 0.20 Colley 1989; Rackham 1989 m
Freswick Links SCA total 1mm 0.7 Jones et al. 1995b 2
Howe Phase 8 substantial? unknown 1147 0.11 1.18 0.15 0.02 Locker 1994; C. Smith 1994 5
Pool Phase 6 minimal 59 0.01 10.8 1.77 7.00 Nicholson 1998a; Bond pers. comm. E!
-3
l m ,3 m
Saevar Howe Phase 1 total 5 mm 115 0.68 2.97 0.31 0.55 Colley 1983; Rowley-Conwy 1983 4
Scalloway Early Phase 3 total &substantial?’ 1 mm 588 0.14 5.00 0.14 Cerh-Carrasco 1998a; S h q l e s 1998;pers comm.
St Boniface Phase 7 total 1mm 1324 69.68 146.11 0.4 0.14 Cer6n-Carrasco 1998b; McCormick 1998
2E
z
6,
mean 553 11.98 29.08 0.51 1.58 4
mean (excluding St Boniface) 398.8 0.44 5.68 0.53 1.94

median 351.5 0.41 6.72 0.31 0.2


median (excluding St Boniface) 115 0.14 5.00 0.23 0.38
n
Viking Age
Brough Rd Areas 1 & 2 partial 10 mm 4432 1.16 44.69 2.68 1.88 Colley 1989; Rackham 1989
Pool Phase 7 minimal 1mm, 3mm 4596 0.42 25.72 3.2 12.33 Nicholson 1998a; Bond pers. comm. 8
Saevar Howe Phase 2 total 5 mm 869 3.41 56.93 5.00 2.25 Colley 1983; Rowley-Conwy 1983 2
Scalloway Late Phase 3 total & ~ubstantial?~1 mm 5682 1.41 3.13 2.6 2.5 Cer6n-Carrasco 1998a;Sharples 1998;pers comm.
Smoo Phase 5 total 4 mm 1116 38.48 20.88 0.87 5.00 Barrett 1996

mean 3339 8.98 30.27 2.87 4.79


r:
median 4432 1.41 25.72 2.68 2.5
Bz
1 Where possible disturbed and poorly dated contexts have been excluded. See Barrett (1995: 528-41) for details. rn
n
2 Zero values rounded to 1 to facilitate division unless no data available for either group. 0
3 Only sieved contexts have been included for the total fish NISP and the ratio of fish:mamnial. For the remaining variables the sieved and hand-collected material can not
be quantified separately. j:
3
TABLE1.A summary of fish bone data for Pictish and Viking Age assemblages from northern Scotland. The variables have been chosen to reflect the
intensity of fishing and the ratio of ‘offshore’ to littoral activity The quality of recovery is indicated on an ordinal scale: n o sieving, minimal sieving
(<I a%), partial sieving (10-50%], substantial sieving (>50%) or total sieving. Other potential biases are discussed in the text.
+
4
u
148 JAMES H. BARRETT, ROELF P. BEUKENS & REBECCA A. NICHOLSON

terns can be extrapolated back in time, the ra- ardize the measure by unit volume. Large- and
tio of cod to saithe should measure the inten- small-scale excavations are represented in both
sity of boat based vis-a-vis shore-based fishing. periods (Hedges 1983; Morris 1989; Pollard
Similarly, the final ratio contrasts taxa with rela- 1992; Hunter e f a]. 1993;B.B. Smith 1994;Morris
tively well defined deep and shallow water habitats et al. 1995; Sharples 1998; Lowe 1998),but there
(Whitehead et a1 1986a; 1986b). If Norse colo- can be no rigorous control of this particular bias.
nists did introduce a more maritime-oriented It is also impossible to impose retrospective
subsistence strategy, some or all of the variables control on the proportion of sediment sieved
considered should have larger values in the Vi- and the mesh size employed at the excavations
king Age than in the preceding Pictish period. under consideration. In these cases, however,
The logic behind the choice of these five cri- the recovery methods are known and vary as
teria is relatively unambiguous, but the data are much within each period as between them (TA-
susceptible to biases: uncertain dating, variable BLE 1).No systematic bias is evident. Preserva-
recovery, taphonomy and inconsistent analyti- tion is difficult to quantify for the assemblages
cal methods. The first issue to consider is that of under consideration, but there is no a priori
dating. Most of the zooarchaeological assemblages reason to suspect consistent differences between
discussed have been grouped into broad chrono- the Pictish period and the Viking Age.
logical phases (with cultural labels)based on radio- The fourth potential bias is variability in ana-
carbon dates,artefact typology and the morphology lytical methods. These are broadly comparable
of buildings. However, middens are notoriously in most of the assemblages under consideration
difficult to date precisely. At least one bone as- (see Barrett et al. 1999: 357-358), but a few ex-
semblage tentatively dated as Pictish or Late Iron ceptions require comment. Firstly, only four
Age based on structural remains almost certainly anatomical elements of the gadid skeleton were
belongs to the Viking Age. The material in ques- systematically quantified for Freswick Links
tion is from late phase 3 at Scalloway, Shetland. due to the large size of this assemblage (Jones
It is undeniable that the distinctive cellular ar- et al. 1995a: 153). It has therefore been neces-
chitecture of this sub-phase is likely to be Pictish sary to exclude all data from this site except
( e g Sharples 1998: 63-5), bul the date of the the ratio of cod to saithe (both gadids).2 Sec-
middens which overlie and infill these structures ondly, the ratios of different fish taxa are based
are better interpreted as Viking Age on the basis on NISP data for all assemblages except Scallo-
of both radiocarbon date ranges (three of five dates way, for which only minimum number of indi-
from late phase 3 include the 9th or loth centu- viduals (MNI)estimates were available. Given the
ries in their 2-sigma error range) and the occur- well-established correlation between NISP and
rence of a Viking style comb, comb case and steatite MNI (Grayson 1984: 62), however, these ratios
fishing-weight in block 7.1 (Sharples 1998: 84; have been included in the analyses to follow.
186). At the least, late phase 3 includes a mix of For all of the variables considered one-tailed
material from before and after Norse contact. It Mann-Whitney tests (Minitab Inc. 2000) have
is classified as Viking Age for the purposes of been used to investigate whether the values of
this study, a conservative approach given that the Viking Age assemblages exceed those of the
Pictish material in the assemblage will dilute any Pictish samples at the 0.05 significance level.
changes introduced by Scandinavian colonists. A nonparametric test was chosen because out-
Early phase 3 thus provides the only unambigu- liers, particularly phase 7 at St Boniface (Cerbn-
ous Late Iron Age assemblage from this site. Carrasco 1998b), suggest that the samples are
Recovery and taphonomy are well-known unlikely to be drawn from normal distributions.
biasing factors in fish-bone assemblages (e.g. The tests were run both with and without this
Jones 1982; Nicholson 1996). Inter-site differ- unusual assemblage.
ences in sample size, the degree to which sieving
was employed, mesh size and preservation could Stable carbon isotopes
all swamp any cultural signals if there is a sys- The dietary implications of stable carbon iso-
tematic bias between Pictish and Viking Age tope signatures (expressed as 6I3Cin parts per
assemblages. The total amount of sediment mil relative to the standard Pee Dee Belemnitella)
excavated will clearly affect the fish NISP and
there are insufficient data available to stand- 2 T h e NISP for this assemblage is 289.
DIET AND ETHNICITY DURING THE VIKING COLONIZATION OF NORTHERN SCOTLAND 149

-16.2 7
I -1
I
-16.7 -- I
, --
tI
I
I

-17.2 -- I
I I
I
2
-17.7 - - I 8
I
I I

-18.2 -- I
I
I
I
I

$j) -18.7 -.
1
I
I
- 1
I

z
I I

-7-
-19.2 -- I _ _ _ ; _ _

-19.7 - - - I

Date AD
(atmospheric calibration)
2. Temporal
FIGURE
trends in the marine Pictish Viking Age Middle Ages
contribution to the -16.2
Orcadian diet based
-16.7
on 6*’C and I4C assays
of human bone. More
+
I
-17.2 I
I

positive 6% values 81

-
-17.7
imply a greater I
I
II
I
I
I

reliance on marine -1 8.2 !

!
protein. FIGURE20 $j) -18.7 ,
I
I
I I
I
t
shows calibrations
using the 1998 -19.2
decadal atmospheric -1 9.7
(terrestrial] data set.
FIGURE2b shows -20.2

1-
calibrations using the -20.7
1998 ‘mixed’
-21.2
atmosph eric/marine
data set and . :II , , >!I , , ; , ,
--+--- r
-21.7
3

estimates of %
I ~

4
marine carbon 0 200 400 600 800 1000 1200 1400
(Stuiver b Reimer
1993; Stuiver et al. Date AD
1998; see Barrett et
al. 2000a). (mixed atmospheric/marine calibration)

for human bone collagen are relatively well ues for 100% marine and 100°/~terrestrial di-
understood (Tauber 1981;Ambrose & Norr 1993; ets appropriate for Scotland have been estimated
Richards &Mellars 1998).In environments such as -12.0% and -Z0.6%0 respectively (Barrett et
as Scotland, which lack C, plants, the more al. 2000a). Precise dietary reconstruction is
positive the 8% value the greater the amount problematic, but more positive 8 ’ T values do
of marine protein in a diet. Endpoint 6I3C Val- indicate a greater reliance on marine foods.
150 JAMES H. BAKKETT, KOELF P. BEUKENS & REBECCA A. NICHOLSON

FIGURES 2a & 2b present 613C values for 22 excavated at Buckquoy (Ritchie 1 9 7 7 ) , the
skeletons of Pictish and Viking Age date from Brough Road, Birsay (Morris 1989),Newark Bay
Orkney. They are plotted against radiocarbon (Barrett et al. 2000a) and Westness (Kaland 1996;
assays calibrated at 2-sigma using both the at- Sellevold 1999). The radiocarbon dates were
mospheric curve and a mixed atmospheric/ determined at the IsoTrace accelerator mass
marine calibration with percent marine carbon spectrometry facility, University of Toronto, and
estimates (Stuiver & Reimer 1993; Stuiver et 6'% was measured by the Environmental Iso-
al. 1998).The latter procedure corrects the dates tope Laboratory of the University of Waterloo.
for marine reservoir effects introduced by the Marine reservoir correction aside, the most
consumption of fish and other marine resources, important potential biasing factor is bone dia-
either directly or indirectly via livestock (Barrett genesis. As samples with less than 5% of their
et al. 2000a). Only material From Orkney has original protein (equivalent to less than 1%
been considered in order to minimize geographi- collagen yield by bone weight) surviving after
cal variability. The samples are from burials biochemical purification can produce anoma-

variable Pictish Viking Age W significance difference in


level distributions?'
fish NISP n=6 n=5 25 0.028 Yes
median = 351.50 median = 4432.00
fish:mammal n=6 n=5 29 0.118 110
median = 0.41 median = 1.41
cod family: n=6 n=5 29 0.118 no
other fish median = 6.72 median = 25.72
cod:saithe n=7 n=5 29 0.005 yes
median = 0.31 median = 2.68
ling & torsk: n=5 n=5 19 0.047 Yes
rocklings, median = 0.20 median = 2.50
wrasse & cottids
= One-tailed test, Viking Age>Pictish

TABLE
2a.

variable Pictish Viking Age W significance difference in


level distributions?l
fish NISP n=5 n=5 17 0.018 Yes
median = 115.00 median = 4432.00
fish:niammal n=5 n=5 18 0.030 Yes
median = 0.14 median = 1.41
cod family: n=5 n=5 18 0.030 Yes
other fish median = 5.00 median = 25.72
cod:saithe n=6 n=5 22 0.007 Yes
median = 0.23 median = 2.68
ling & torsk: n=4 n=5 14 0.089 no
rocklings, median = 0.38 median = 2.50
wrasse & cottids
= One-tailed test, Viking Age>Pictish
2b.
TABLE

2. One-tailed Mann-Whitney test comparisons of the zooarchaeological variables in TABLE1 .


TABLE
TABLE2a includes all sites whereus TABLE2b omits the anomalous assemblage from St Boniface phase 7.
DIET AND ETHNICITY DURING THE VIKING COLONIZATION OF NORTHERN SCOTLAND 151

lous isotopic results (Taylor 1997: 90-91), six is also a notable shift from small to large saithe,
specimens with yields below this threshold have a circumstance almost certainly due to local
been omitted from consideration. Three sam- factors; the waters around Sumburgh Head are
ples with 6% values of <-22 have also been particularly rich in this species.
left out. They are probably contaminated by The lack of a significant difference in the
humic acids (van Klinken 1999: 691). Full de- remaining variables is noteworthy, however, and
tails regarding analysis of these samples, in- requires some consideration. It is due to anoma-
cluding all 6l”Cvalues and radiocarbon dates, lous values for the St Boniface phase 7 assem-
are available elsewhere (see Barrett et al. 2000a; blage. If this site is excluded, four of the five
Barrett et al. Zooob). Carbon isotope data from variables considered have significantly higher
the published literature produced along with values in the Viking Age - implying both in-
radiometric radiocarbon dates (e.g. Sellevold creased fishing intensity and a shift from litto-
1999: 7) have also been omitted as a small ral to open-water activity (TABLEah). The
fractionation effect can be induced during com- non-significant results for ling and torsk ver-
bustion of the large samples used (G. Cooke sus rockling, wrasse and cottids in this case
pers. comm.; S. Gulliksen pers. comm.). may simply reflect the decrease in sample size
The individual burials analysed have been due to the exclusion of St Boniface.
classified as ‘Pictish’ or ‘Viking’ based on the The outlying values for some variables of St
presence of grave-goods, a Norse tradition, and Boniface phase 7 are particularly odd given that
the modes of calibrated I4C dates. The distribu- the ratio of cod to saithe for this sample is clearly
tions were then analysed using one-tailed t-tests consistent with the remaining Pictish assem-
(Minitab Inc. 2000) in order to determine whether blages. It is possible that the elevated fish NISP
the 6% values of the Viking Age samples are sta- and ratio of fish to mammal bone at this site
tistically higher (at the 0.05 significance level) are due to the unusually rigorous procedure of
than the Pictish ones - implying an increase in quantifying all fish bone recovered from a 1-
the consumption of marine resources. mm mesh (Cer6n-Carrasco 1994: 207). However,
the extremely high ratio of gadid to non-gadid
Results species remains anomalous. It may relate to the
Zooarchaeology context from which the bone derives. Most of
Statistical comparison of the Pictish and Vi- the other assemblages are from settlement sites
king Age assemblages indicates significant dif- and associated middens. Conversely, phase 7
ferences between the two periods for the NISP at St Boniface was a buried ground surface which
of fish bone, the ratio of cod to saithe and the developed over a considerable period of time
ratio of ling and torsk to rocklings, wrasse and (Lowe 1998: 86-8).
cottids (TABLE2a). These variables have nota-
bly larger values in the Viking Age, possibly Stable carbon isotopes
implying a n increase in the amount of fish Student’s t-tests comparing the 6I3C values of
caught and a greater emphasis on offshore fish- Pictish and Viking Age samples confirm that
ing. Similar patterns have also been recognized the Viking Age skeletons do have higher val-
in material from Old Scatness, Shetland, al- ues (TABLE3) - a difference which can be de-
though quantitative data are not yet available tected visually in FIGURES
2a-zb.This distinction
(Nicholson 1998b;unpublished data). Here there remains significant regardless of whether the

calibration Pictish Viking Age f-value degrees of significance difference


method freedom level in means?‘
atmospheric n=7 n=15 3.89 19 <0.001 Yes
mean = -20.46 mean = -19.00
atmospheridmarine n = 7 n=13 3.68 16 0.001 Yes
mean = -20.46 mean = -1 9.38

= One-tailed test, Viking Age>Pictish

TABLE3 . One-tailed t-test comparison of 613Cvalues for Pictish and Viking Age burials from Orkney.
152 JAMES H. BARRETT. ROELF P. BEUKENS & REBECCA A. NICHOLSON

burials are classified using I 4 C dates calculated cussed here could be interpreted as a result of
with the atmospheric curve or with the mixed production of obligatory food renders or emu-
atmospheric/marine calibration (in which case lation of a high-status practice. These options
two samples become post-Viking Age and are seem unlikely, however, given the almost com-
excluded from analysis). It would appear that plete dominance of Norse place-names in the
Viking Age Orcadians ate more marine protein Northern Isles (Fellows-Jensen 1984: 151) and
than their predecessors, a result consistent with evidence that fish was probably of inferior sta-
the zooarchaeological evidence for an increase tus to meat and grain in Norse Scotland (Barrett
in the intensity of fishing activity. Although 1995: 280). Alternatively, the present evidence
this isotopic study is restricted to Orkney, it is may imply that Norse farmer-fishermen immi-
worth noting that the small number of 6I3C as- grated on a large scale and displaced or gained
says from Late Iron Age and Viking burials in direct economic control over their predeces-
the Hebrides illustrate the same pattern (Neigh- sors.
bour & Montgomery forthcoming). This hypothesis may be supported by other
ecofactual evidence. It has been recognised, for
Conclusion example, that seabirds such as gannets ( S u l a
The available zooarchaeological and stable iso- bassana),cormorants (Phalacrocorax carbo)and
topic data from northern Scotland suggest that shags (Phalacrocorax aristotelis) are more abun-
the Viking Age was characterized by an increase dant in Viking Age than in Pictish assemblages
in the intensity of fishing and in the relative from Orkney (Mead 1999: 67,75,85;Serjeantson
importance of boat fishing for large gadids, in press). Given that the closest nesting colony
particularly cod. What implications do these of gannets is currently Sule Stack, c. 60 km from
observations have for the Viking colonization the archipelago (Booth et al. 1984: 12), and the
of Scotland? The relationship between ethnic- observation that all three species could be caught
ity and material culture is complex. Simple at sea (Mead 1999: 85), an increase in their
correlations have been shown to be nai've for exploitation is consistent with an intensifica-
suites of material culture, and are likely to be tion of fishing and maritime activity in gen-
even more so for a single facet of social life eral. It may also be relevant that there was a
such as subsistence (see Jones 1997: 106-27). concurrent transformation of the terrestrial
Nevertheless, food is an important mechanism economy. At Pool, Orkney, for example, Julie
for expressing identity (Crabtree 1990: 177-91; Bond (1998; see also Bond & Hunter 1987) has
Gumerman 1997: 126). Moreover, successful recognized changes in cattle management and
open-water fishing in the Northern Atlantic an increase in the importance of flax between
would require a corpus of traditional knowl- Pictish and Viking Age levels.
edge (see Morrison 1992: 128)which, based on
the cod dominated middens of Iron Age and Acknomdsdgsrnents. An earlier version of this paper was
presented at the 64th annual meeting of the Society for
Viking Age Norway (Perdikaris 1999), could American Archaeology, Chicago, 24-28 March 1999. The
easily have been introduced by Norse colonists. study was funded by the Social Sciences and Humanities
Other possible explanations for the increase Research Council of Canada as a component of the Viking
in offshore fishing at the Pictish-Viking Age Age Transitions Project (Phase 1). Human bone samples
transition, such as development of the Euro- were kindly donated by Anne Brundle of the Tankerness
House Museum, Orkney, Theya Molleson of the Natural
pean fish trade, only become relevant with a History Museum, London and Alison Sheridan of the Na-
further increase in fishing intensity in the 11th- tional Museums of Scotland. Sarah King assisted with col-
12th centuries (Barrett 1997;Barrett et al. 2ooob). lecting the samples. Julie Bond, Ruby Cer6n-Carrasco,
If one begins from an assumption that the Picts Gordon Cook, lnge Bodker Enghoff, Steinar Gulliksen,
Jennifer Harland, Coralie Mills, Tim Neighbour, Terry
and their culture were largely undisturbed in O'Connor, Berit Sellevold, Dale Serjeantson and Niall
the Viking Age, except for the immigration of Sharpies are all owed thanks for providing helpful infor-
a small Norse Blite, the economic changes dis- mation and offprints.
DIET AND ETHNICITY DIJRING THE VIKING COLONIZATION OF NORTHERN SCOTLAND 153

References
AMBROSE, S.H. 1990. Preparation and characterization of hone ENCHOFF.I.B. 1999.Fishing in the Baltic region from the 5th
and tooth collagen for isotopic analysis, Journal of Ar- century BC to the 16th century AD: Evidence from fish
chaeological Science 17:431-51. bones, Archaeofauna 8: 41-85.
AMBROSE, S.H. & L. NORR.1993. lsotopic composition of di- FELLOWS-JENSEN, G. 1984. Viking settlement in the Northern
etary protein and energy versus bone collagen a i d apa- and Western Isles - the place name evidence as seen
tite: Purified diet growth experiments, in J.B. Lamhert & from Denmark and the Danelaw, in A. Fenton & H. Pilsson
G . Grupe (ed.), Prehistoric human bone: Archaeology a t [ed.), The Northern and Western Islesin the Viking World
the molecular level: 1-37. Berlin: Springer-Verlag. 146-68. Edinburgh: John Donald.
ANTHONY, D.W. 1990. Migrations in archaeology: The baby and FENTON,A. 1978. The Northern Isles: Orkney and Shetlond.
the bathwater, American Anthropologist 92: 895-914. Edinburgh: John Donald.
BALDWIN, J.R. 1982.Fishing the sellag: Hand netting traditions FOSTER,S.M. 1990. Pins, combs and the chronology of later
from Caithness, the Northern and Western Isles, in J.R. Atlantic Iron Age settlement, in I. Armit (ed.), Beyond
Baldwin (ed.),Caithness: A cultural crossroads: lGl-212. the brochs: 143-74. Edinburgh: Edinburgh University
Edinburgh: Edina Press. Press.
BARRETT, J.H. 1995.‘Few know an earl in fishing-clothes.’Fish GOODLAL), C.A. 1971. Shetland fishing saga. Lerwick: Shetland
middens and the economy of the Viking Age and Late Times.
Norse earldoms of Orkney and Caithness, northern Scot- GKAYSON, D.K. 1984. Quantitative zoonrchneology. London:
land. Ph.D thesis, University of Glasgow. Academic Press.
1996.A summary report on animal bone from excavations GUMERMAN IV, G. 1997. Food and complex societies, Journal
at Smoo Cave, Sutherland, Scotland, 1992. Department of Archaeological Method and Theory 4: 105-39.
of Archaeology, University of Glasgow. HARKE, H. 1998.Archaeologists and migrations: A problem of
1997. Fish trade in Norse Orkney and Caithness: a zoo- attitude, Current Anthropology 39: 19-45.
archaeological approach, Antiquity 71: 616-38. HEDGES,J.W. 1983. Trial excavations on Pictish and Viking
BAKRETT, J.H., R.A. NICHOLSON& R. CER~N-CARKASCO. 1999. settlements at Saevar Howe, Birsay, Orkney, Glasgow
Archaeo-ichthyological evidence for long-term socioeco- Archaeological Journal 10: 73-124.
nomic trends in northern Scotland: 3500 BC to AD 1500, HUNTER, J.R. 1997.The early Norse period, in K.J. Edwards &
Journal ofilrchaeological Science 26: 353-88. I.B. Ralston (ed.), Scotland: environment a n d archaeol-
BAKRET’I’, J.H., R.P. BEUKENS & D.R. BROTHWELL. 2000a. Radiocar- ogy, ROO0 BC-AD 1000: 241-54. New York (NY):John Wiley
bon dating and marine reservoir correction of Viking Age & Sons.
Christian burials from Orkney, Antiquity 74:53743. HUNTER, J.R., J.M. BOND& A.M. SMITH. 1993. Some aspects of
BARRETI’, J., R. BEUKENS, I. SIMPSON, P. ASHMORE, S. POAPS & J. Viking settlement in Orkney, in C.E. Batey, J. Jesch & C.D.
HUNTLEY. 2000b.What was the Viking Age and when did Morris (ed.). The Viking Age in Caithness, Orkney a n d
it happen? A view from Orkney, Norwegian Archaeological the North Atlantic: 272-84. Edinburgh: Edinburgh IJni-
Review 33: 1-39. versity Press.
BOND,J.M. 1998. Beyond the fringe? Recognising change and JONES,A.K.G. 1982. Bulk-sieving and the recovery of fish re-
adaptation in Pictish and Norse Orkney, in C.M. Mills & mains from urban archaeological sites, in A.R. Hall & H.
C,. Coles (ed.). Hziman settlement in marginal areas: 81- Kenward (ed.), Environment01 archaeology in the urban
90. Oxford: Oxbow. Monograph 100. context. 79-85. London: Council for British Archaeol-
BOND,J.M. & J.R. HUNTER. 1987.Flax-growing in Orkney from ogy. Research report 43.
the Norse period to the 18th century, Proceedings of the JONES, A.K.G., C.D. MORRIS& D.J. RACKHAM. 1995a.Environ-
Society of Antiquaries of Scotland 117: 175-81. mental sampling: methods and materials, in Morris et
BOOTH,C., M. CIJTHBERT & P. REYNOLDS. 1984. The Birds of al. (ed.): 149-53.
Orkney. Kirkwall: The Orkney Press. 1995b.The fish material, in Morris et ai. (ed.): 154-91.
BURMEISTER, S. 2000. Archaeology and migration: Approaches JONES, S. 1997.The archaeologyofethnicity. London: Routledge.
to an archaeological proof of migration, Current Anthro- KALAND, S.H.H. 1996. En vikingtidsgird og -gravplass pB
pology 41: 539-69. Orknayene, in Krager & Naley (ed.): 63-8.
BIJTEUX, S. (ed.). 1997. Settlements a t Skaill, Deerness, Ork- KRBGER. J.F. & H. NALEY(ed.). 1996. Nortfsjoen: Handel, reli-
my. Oxford: Tempvs Reparatvm. BAR British series 260. gion og politikk. Karmay Kommune: Vikingfestivalen.
CERON-CARRASCO, R. 1994. The investigation of fish remains LOCKER,A. 1994. The fish remains, in Smith (ed.): 157-9.
from an Orkney farm mound, in W. Van Neer (ed.), Fish LOWE, C. (ed.). 1998.S t Boniface Church, Orkney: Coastal ero-
exploitation in the past: Proceedings of the 7th meeting sion a n d archaeological assessment. Edinburgh: Sutton
of th e ICAZ fish remains working group: 2 0 7-1 0. Tervuren: PublishingiHistoric Scotland.
Musee Royal de L‘Afrique Centrale. MCCORMICK, F. 1998. Mammal bone, in Lowe (ed.): 14R-9.
1998a. Fish, i n N. Sharples (ed.), Scalloway: A hroch, Late MEAD,P. 1999. Exploitation of the wild bird resource in the
Iron Age settlement a n d medieval cemetery. Oxford: Northern Isles over three prehistoric time periods. L J m
Oxbow. Monograph 8 2 . published undergraduate dissertation, University of Brad-
1998b. Fish bone, in C. Lowe (ed.), St Boniface Church, ford.
Orkney: Coastol erosion a n d archaeological assessment: MINITABINC. 2000.Minitab Statistical Software Release 13.1.
149-55. Edinburgh: Sutton Pnblishing/Historic Scotland. State College (PA): Minitab Inc.
CHAPMAN, J. & H. HAMEROW (ed.1. 1997. Migrations a n d inva- MORRIS,C.D. (ed.]. 1989. The Birsaj~BoyProject 1:Brough Road
sion in archaeological explanation. Oxford: Tempvs Excavations 1976-1 982. Durham: University of Durham,
Reparatvm. BAR International series S 6 6 4 . Department of Archaeology. Monograph series 1.
COLLEY, S.M. 1983.Marine resource exploitation, in J.W. Hedges, 1996.The Norse impact in the Northern Isles of Scotland,
Trial excavations o n Pictish and Viking settlements at in Krager & Naley (ed.): 69-83,
Saevar Howe, Birsay, Orkney, 10: 111-13, MF 87-92, MF MORRIS, C.D., C.E. BATEY&D.J.RACKHAM.1995.Freswick Links,
94-102. Caithness: Excavation a n d survey of a Norse Settlement.
1989.The fish hones, in Morris (ed.): 248-59. Inverness: Highland LihrariedNorth Atlantic Biocultural
CKABTREE,
P. 1990.Zooarchaeology and complex societies: some Organisation.
uses of faunal analysis for the study of trade, social sta- MORRISON, 1.1992. Traditionalism and innovation in the maritime
tus, and ethnicity, in M.B. Schiffer (cd.), Archaeological technology of Shetland and other North Atlantic com-
method a n d theory 2: 155-205. Tucson (AZ): University munities, in T.C. Smout (ed.),Scotland a n d the sea: 114-
of Arizona Press. 36.Edinburgh: John Donald.
154 JAMES H. BARRETT, ROELF P. BEUKENS & REBECCA A. NICHOLSON

NEIGHBOUR, T. & J. MONTGOMERY. Forthcoming. 'You are what historic perspective on the extinction of the great auk,
you eat: Skeletal clues to the impact of the Norse on the International Journal of Osteoarchaeology.
Western Isles', in M. MacLeod, Gall-Ghaidheil: The Western SHARPLES, N. 1998. Scalloway: A broch, Late Iron Age settle-
Isles in the Viking World, Stornoway, Lewis, 2000. ment a n d medieval cemetery in Shetland. Oxford: Ox-
NICHOLSON, R.A. 1996. Fish bone diageriesis in different soils, bow. Monograph 82.
Archaeofauna 5: 79-91. SHARPLES, N. & M. PARKER PEARSON. 1999. Norse settlement in
1998a. Fishing in the Northern Isles: A case study based on the Outer Hebrides, Norwegian Archaeological Review
fish bone assemblages from two multi-period sites on 32: 41-62.
Sanday, Orkney, Environmental Archaeology: The Jour- SMITH, B.B. (ed.). 1994. Howe: Fourmillenia of Orkneyprehis-
nal of Human Palaeoecology 2: 15-28. tory. Edinburgh: Society of Antiquaries of Scotland. Mono-
1998b. Fishing for facts: A preliminary view of the fish re- graph series 9.
mains from Old Scatness Broch, in R.A. Nicholson & S.J. SMITH,C. 1994. Animal bone report, in Smith (ed.): 139-53.
Dockrill (ed.),Old Scatness Broch, Shetland: Retrospect S ~ U I V E M. R , & P.J. REIMER.1993. Extended ''C database and
a n d prospect. 97-110. Bradford: Iiniversity of Bradford/ revised CALIB radiocarbon calibration program, Radio-
Shetland Amenity Trust/North Atlantic Biocultural Or- carbon 35: 215-30.
ganisation. S r U I V E R , M., P.J. REIMER& T.F. BRAZIUNAS. 1998. High-preci-
PERDIKARIS, S. 1999. From chiefly provisioning to commercial sion radiocarbon age calibration for terrestrial and ma-
fishery: Long-term economic change in Arctic Norway, rine samples, Radiocorbon 40: 1127-51.
World Archaeology 30: 388-402. TAUBER, H. 1981. "C evidence for dietary habits of prehistoric
POLLARD, T. 1992. Smoo Cave. Glasgow: Glasgow University man in Denmark, Nature 292: 332-3.
Archaeological Research Division. TAYLOR, R.E. 1997. Radiocarbon dating, in R.E. Taylor & M.J.
RACKHAM, D.J. 1989. Excavations beside the Brough Road, Birsay: Aitken [ed.), Chronometric dating in archaeology: 65-
The biological assemblage, in Morris (ed.): 231-72. 96. New York (NY): Plenum Press.
RICHARDS, M.P. & P.A. MELLARS. 1998. Stable isotopes and the VAN KLINKEN, G.J. 1999. Bone collagen quality indicators for
seasonality of the Oronsay middens, Antiquity 72: 178- palaeodietary and radiocarbon measurements, Journal of
84. Archaeological Science 26: 687-96.
RITCHIE, A. 1974. Pict and Norseman in northern Scotland, WAINWKIGHT, F.T. (ed.). 1964. The Northern Isles. Edinburgh:
Scottish Archaeological Forum 6 : 23-36. Thomas Nelson.
1993. Viking Scotland. London: B.T. Batsford/Historic Scot- WHEELER, A. 1977. The fish bones from Buckquoy, Orkney, in
land. A. Ritchie, Excavation of Pictish and Viking-age farm-
ROWLEY-CONWY, P. 1983. Theanimal and bird hones, in Hedges steads at Buckquoy, Orkney, Proceedings of the Society
(1983): fiche 70-91. of Antiquaries of Scotland 108:211-14.
SELLEVOLD, B. J. 1999. Picts and Vikings at Westness: Anthro- WHITEHEAD, P.J. P., M.L. BAUCHOT, J.C. HUREAU, J. NIELSEN& E.
pological investigations of the skeletal material from the TORTONESE (ed.). 1986a. Fishes of the northeastern At-
cemetery at Westness, Rousay, O r h e y Islands. Oslo: NIKU lantic and the Mediterranean 2. Paris: UNESCO.
Scientific Report 010: 1-62. 198fib. Fishes ofthe northeastern Atlantic a n d the Mediter-
SBRIEANTSON, D. I n prcss. The great auk and the gannet: A pre- ranean 3 . Paris: UNESCO.

You might also like