2012 D.E. Nelson, N. Lynnerup, and J. Arneborg 51 Introduction In the first half of the second millennium AD, Greenland was home to two human populations with very different cultures and very different adaptations to the conditions of the land. The European Norse had arrived and established settlements at the southern tip by the late A.D. 980s, and shortly thereafter, people of the Thule (or proto-Inuit) Culture arrived on the northern Greenlandic coast after their rapid sweep across the arctic from Siberia. The Norse were agriculturalists who stretched the limits of North European pastoralism in Greenland. In contrast, the Thule Culture people were sea-mammal hunters with a specialized technology adapted to Arctic conditions. Their dietary economy was based entirely on seasonal exploitation of a wide range of the resources within an extensive area, including terrestrial game (e.g., caribou) when and where available (e.g. Arneborg 2004, Gulløv 2004). While the primary focus of the Greenland Isotope Project as a whole is an isotopic analysis of the Norse adaptation and diet in Greenland, we make a diversion here into that of the Thule Culture people. There are good reasons for so doing. First, our analyses of the prey species in Greenland (Nelson et al. 2012 [this volume]) has made it clear that the application of isotopic dietary analyses to Greenland is complicated by the observed isotopic differences within the marine biosphere. We can thus use the Greenlandic Thule Culture as a test of the isotopic method, in that we know they were heavily dependent on marine resources, and should thus provide a direct measure of the isotopic values for humans consuming almost entirely marine protein. This study can also be seen from a more basic archaeological perspective. The question, “To what extent did the Norse eventually modify their dietary economy to one similar to that of the Thule Culture?” can be addressed directly by comparing the isotopic values of members of both groups. These people can thus serve as an ideal counter-point or means of calibrating the isotopic method for our study of the Norse. Having a second human population against which to compare is most unusual in applications of the isotopic dietary method. Second, the Thule Culture people are of great interest in their own right. With the exception of carbon measurements on a few individuals made as a first test of the isotopic method (Tauber 1981, 1989) and some undertaken for dating purposes (e.g., Heinemeier and Rud 1997), no isotopic dietary studies have been undertaken on them. As a spinoff, we thus present here a first overview isotopic study of the Greenlandic Thule Culture. With the aid of H.C. Gulløv, National Museum of Denmark, who is a specialist in Thule Culture archaeology and ethnology, a range of individuals in the repository at the Laboratory for Biological Anthropology, University of Copenhagen was sampled. The samples included sites from as much of the periphery of Greenland as possible and were from very approximately the same time period as that of the Norse colonies. In the analysis to follow, the only archaeological information used was that from a review of the faunal lists as given by the excavators of the sites under study (see Gulløv 2012 [this volume]), and we then asked H.C. Gulløv to evaluate and comment on these interpretations. Are the isotopic observations in accord with existing knowledge? Can the method add anything new to the field? His independent analysis and commentary is to be found in Gulløv 2012 (this volume). A First Isotopic Dietary Study of the Greenlandic Thule Culture D. Erle Nelson1, Niels Lynnerup2, and Jette Arneborg3,4,* Abstract - The isotopic dietary method has been applied to samples of some 65 Thule Culture individuals from existing archaeological collections of Greenlandic human skeletal material. The aim was to use the Greenlandic Thule Culture as a test of the isotopic method, in that we know they were heavily dependent on marine resources, and should thus provide a direct measure of the isotopic values for humans consuming almost entirely marine protein. The sites from which the material was originally obtained encompass almost the entire periphery of Greenland. Isotopic data from a study of animals of Greenland was used as the basis for analysis. As expected, the results indicate that these people were almost entirely dependent on the marine biosphere for their protein. An exception is those from Northeast Greenland, whose isotopic signatures show evidence for consumption of terrestrial protein as well. Not expected were the observed differences at the regional and local levels. Special Volume 3:51–64 Greenland Isotope Project: Diet in Norse Greenland AD 1000–AD 1450 Journal of the North Atlantic 1FRSC Professor Emeritus, Simon Fraser University, Department of Archaeology. Burnaby, BC, Canada. 2University of Copenhagen, Faculty of Health Sciences, Department of Forensic Medicine, Laboratory of Biological Anthropology. Blegdamsvej 3B, DK-2200 Copenhagen N, Denmark. 3National Museum of Denmark, Danish Middle Ages and Renaissance. Frederiksholms Kanal 12, DK-1220 Copenhagen, Denmark. 4School of GeoSciences, University of Edinburgh, Scotland UK. *Corresponding author - Jette.arneborg@natmus.dk. 2012 52 Journal of the North Atlantic Special Volume 3 The Samples Samples of 65 individuals from 14 Greenlandic sites were selected for measurement from the collection. No great care was taken in choice, as we simply attempted to measure every available individual from each site. As will be seen, this had a strange consequence. A shortlist of the sites, their locations, and references to the reports from which the faunal lists were obtained are given in Table 1. Gulløv (2012 [this volume]) provides more archaeological detail on these topics. Figure 1 places these sites on the map. For most, the excavations were undertaken early in the 20th century. An important exception is the site Asummiut, which was excavated during a salvage operation a few years before this study was undertaken. A list of all individuals and their recorded sex is given in the first five columns of Table 2, which contains the basic data for this study. All are adults unless otherwise noted, and all samples were taken from crania, with the exception of those from Asummiut, which were sampled at mid-shaft femur. Only 12 individuals were identified as males in the suite of 65 individuals. The bones were sampled at the University of Copenhagen so that only the small amounts of material destined for measurement had to be moved from the collection to the analytical laboratory. At the beginning of the project, the sample sizes were typically 100–150 mg of bone; later we reduced the sample size to 20–40 mg. Collagenextraction procedure and isotopic measurement was as for all other samples in this overall project, as described in Takahashi and Nelson (2012 [Appendix 1, this volume]). The data obtained are also listed in Table 2. We have every reason to believe that these measures are reliable, as the process of collagen extraction, the extract yields, and the carbon and nitrogen concentrations were as expected for wellpreserved bone. The Sites Greenland is not a homogeneous whole, and for this first analysis, we divide the island into regions that reflect the different environmental conditions as described in a recent overview ecological text (Born and Böcher 2001). In the descriptions that follow, this information, as well as that provided by the comprehensive review of Greenlandic mammals given by Vibe (1981), is used to discuss the prey species that were likely to have been available in each region. A summary review of the species listed in the excavation faunal reports is also given for each site. To the extent possible, the animal isotopic data (Nelson et al. 2012 [this volume]) for each region then provides the basis for interpretation of the human data. We thus define the “Northwest region” as the area on the west coast having Disko Bay as its southernmost limit. Disko Bay is presently at about the lower limit of west coast pack ice in the winter. This region includes the sites Tunnungassoq, Illutalik, and Nuugaaq (Mathiassen 1930, 1934; see also Gulløv 2012 [this volume]). These people would have had direct access to large numbers of ringed seals (Pusa hispida) both in winter and spring, as this seal favors pack ice. Migrating harp seals (Pagophilus groenlandicus) would appear in spring and fall, as might much lesser numbers of hooded seals (Cystophora cristata), which tend not to go so far north. Harbor seals do not like ice and are not numerous. The solitary bearded seals (Erignathus barbatus) are rare but perhaps present in limited numbers in the summer. There may have been plentiful Table 1. The sites studied. “FM Site Number” refers to the “Fortidsminde Nummer” site designation system of the Greenland National Museum and Archives Estimated Site name FM site number Latitude Longitude period (AD) References Tunnungassoq 72V1-IV-022 72°57' 56°05' 16–1700s Mathiassen 1930 Illutalik 69V2-I-017 69°55' 50°36' 15–1600s Mathiassen 1934 Nuugaaq 69V2-IV-014 69°58.5' 51°16' 15–1600s Mathiassen 1934 Asummiut 66V1-IV-066 66°57.5' 53°44.0' 16–1700s Grummegaard-Nielsen 1997 Lynnerup et al. 1997 Heinemeier and Rud 1997 Utoqqaat 65V1-IV-023 65°51' 53°14' 14–1600s Mathiassen 1931 Qoornoq 64V1-I-009 64°32' 51°06' 16–1700s Gulløv 1983 Uunartoq 60V2-IV-001 60°31' 45°20' 1700s Mathiassen 1936 Ruinnæsset 63Ø1-IV-001 63°30' 41°40' 16–1700s Mathiassen 1936 Suukkersit 66Ø2-III-001 66°02' 37°56' 17–1800s Mathiassen 1933 Skærgårdshalvø 68Ø3-III-001 68°08.5' 31°46' 15–1700s Larsen 1938 Uunarteq 70Ø1-II-003 70°25' 21°58' 16–1700s Amdrup 1909 Kap Harry 72Ø2-IV-003 72°47' 24°52' 15–1600s Glob 1935 Suess Land 73Ø2-III-002 73°00' 25°00' 16–1700s Glob 1935 Dødemandsbugten 74Ø2-II-005 74°07' 20°53' 15–1600s Larsen 1934 Stormbugt 76Ø1-034 76°49' 18°50' 15–1600s Thostrup 1911 Rypefjeldet 76Ø1-025 76°56.0' 20°23.0' 15–1600s Thostrup 1911 2012 D.E. Nelson, N. Lynnerup, and J. Arneborg 53 Figure 1. Map of site locations. 54 Journal of the North Atlantic Special Volume 3 Table 2. Summary of Neo-Eskimo sample data. δ13C given in ‰ v PBD; δ15N given in ‰ v AIR. LBA ID refers to the Laboratory for Biological Anthropology identification number. Project Other Sex Extract Site ID LBA ID information (age) C% N% C/N δ13C δ15N Northwest region Tunnungassoq #325 KAL-0001X02 Grave 6 B F 44.4 15.3 2.9 -13.0 21.4 " #328 KAL-0001X01 Grave 6 A M 44.1 15.4 2.9 -13.1 21.3 " #331 KAL-0001X03 Grave 6 (II)C F 43.9 15.6 2.8 -13.3 21.9 Illutalik #334 KAL-1548X01 Grave 34 c F 44.2 15.3 2.9 -13.4 21.6 " #335 KAL-1547X01 Grave 34 b F 44.5 16.7 2.7 -12.7 21.4 " #352 KAL-1546X01 Grave 34 a M 44.4 15.3 2.9 -13.3 20.2 " #353 KAL-1549X01 Grave 34 a F 44.9 16.2 2.8 -13.5 21.8 Nuugaaq #332 KAL-0160X01 Grave 1 F 44.6 15.8 2.8 -13.5 19.7 " #340 KAL-0159X01 Grave 1 F 45.0 15.7 2.9 -13.0 21.0 " #342 KAL-0158X01 Grave 1 F 45.0 16.1 2.8 -12.9 20.9 " #347 KAL-0157X01 Grave 1 F 45.0 16.1 2.8 -13.1 22.6 Southwest region Asummiut #96 AS 23/95 SIK 892 D7 x 621 F 46.5 17.1 2.7 -12.8 19.9 " #97 " SIK 892 D7 x 729 F 47.3 17.4 2.7 -12.6 18.2 " #98 " SIK 892 D7 x 119 F 46.3 17.1 2.7 -12.4 19.5 " #99 " SIK 892 D7 x 5 F 47.1 17.4 2.7 -12.4 19.9 " #100 " SIK 892 D7 x 40 F 46.1 17.3 2.7 -12.3 19.7 " #101 " SIK 892 D7 x 596 F 46.7 17.2 2.7 -12.6 19.6 " #102 " SIK 892 D7 x 322 F 46.9 17.5 2.7 -12.1 19.1 " #103 " SIK 892 D7 x 28 F 46.9 17.3 2.7 -12.9 19.2 " #124 AS 21/94 SIK 887: 1a 4–5 y 43.9 15.0 2.9 -12.9 19.9 " #125 " SIK 887: grave 1b F 44.1 14.7 3.0 -12.4 18.0 " #126 " SIK 887: grave 1c F 43.7 14.5 3.0 -12.4 19.2 " #127 " SIK 887: grave 2a F 44.4 14.7 3.0 -12.8 18.2 " #128 " SIK 887: grave 2b F 44.0 14.8 3.0 -12.7 19.2 " #129 " SIK 887: grave 2c 7 y 44.5 14.8 3.0 -13.4 18.2 " #130 AS 23/95 SIK 892 D1: #1 F 44.0 14.6 3.0 -12.2 18.9 " #131 " SIK 892 D1: #2 F 44.4 15.0 3.0 -11.8 19.7 " #132 " SIK 892 D1: #3 M 44.5 14.8 3.0 -12.1 19.0 " #133 " SIK 892 D1: #4 12–16 y, F? 44.2 14.9 3.0 -12.3 20.1 " #134 " SIK 892 D1: #5 F 44.0 14.9 3.0 -12.5 19.1 " #135 " SIK 892 D1: #6 F 45.0 15.0 3.0 -12.7 18.4 " #136 " SIK 892 D1: #7 F 44.4 14.9 3.0 -12.1 19.3 " #137 " SIK 937 D23: X9 10–12 y 44.6 15.0 3.0 -11.7 19.9 " #138 " SIK 939 D26 F 44.0 14.8 3.0 -12.6 20.2 Utoqqaat #338 KAL-0011X01 Grave 1 M 44.4 15.0 3.0 -13.6 18.6 Qoornoq #346 KAL-1440X01 House 3 skel. 7 F 44.8 15.7 2.9 -12.8 19.4 " #351 KAL-1449X01 House 3 skel. 7 F 44.9 16.0 2.8 -13.2 19.1 Southwest tip region Uunartoq #329 KAL-1460X01 Grave 8 A F 44.3 15.8 2.8 -13.4 19.6 " #336 KAL-1469X01 Grave 8 F 44.6 15.8 2.8 -13.2 19.4 " #345 KAL-1468X01 Grave 8 D F 44.7 15.9 2.8 -14.0 18.9 " #349 KAL-1461X01 Grave 8 B M 45.1 15.7 2.9 -13.3 19.7 " #350 KAL-1467X01 Grave 8 C F 44.7 15.9 2.8 -13.2 19.0 Southeast region Ruinnæsset #327 KAL-1422X01 Grave 4 B F 44.2 15.5 2.8 -13.8 19.3 " #348 KAL-1425X01 Grave 4 A F 45.1 15.8 2.9 -13.3 19.4 Suukkersit #337 KAL-0122X01 House ruin 5 F 45.1 16.4 2.8 -13.7 18.8 Skærgårdeshalvø #339 KAL-1410X01 Grave 9 F 44.7 15.2 3.0 -13.7 20.3 " #344 KAL-0048X01 Grave 9 F 45.4 15.5 2.9 -13.4 19.0 Northeast region Uunarteq #333 KAL-0707X01 Grave 5 10–11 y 44.4 15.5 2.9 -14.5 20.1 Kap Harry #321 KAL-0067X01 Grave 16 F 44.1 15.3 2.9 -15.6 18.0 Suess Land #322 KAL-0043X01 Grave 24 F 44.2 16.4 2.7 -15.9 18.1 " " #324 KAL-0044X01 Grave 25 F 43.9 15.0 2.9 -16.1 16.8 " " #388 KAL-0041X01 Grave 21 M 44.4 14.8 3.0 -15.3 17.2 " " #389 KAL-0065x01 M 44.4 14.5 3.1 -15.7 17.4 Dødemandsbugten #326 KAL-0055X01 Grave 6 M 44.0 15.7 2.8 -16.1 17.3 " #379 KAL-0723X01 Grave 1 F 44.7 15.5 2.9 -15.6 16.8 " #380 KAL-0030X01 Grave 4 F 45.1 15.2 3.0 -15.5 19.3 " #381 KAL-0059X01 Grave 14 F 44.7 15.1 3.0 -14.8 19.6 " #382 KAL-0057X01 Grave 11 M 44.1 14.6 3.0 -14.8 19.2 " #383 KAL-0061X01 Grave 16 F 44.3 14.9 3.0 -15.3 17.1 " #384 KAL-0058X01 Grave 12 M 44.1 14.7 3.0 -14.9 19.3 " #385 KAL-0060X01 Grave 15 M 43.9 14.7 3.0 -15.3 17.2 " #386 KAL-1243X01 Grave 13 juvenile 44.2 14.8 3.0 -15.1 19.4 " #387 KAL-1244X01 Grave 13 M 44.3 15.5 2.9 -14.8 18.8 Rypefjeldet #323 KAL-0077X01 529, 530, 531 F 44.3 15.7 2.8 -15.5 18.7 Stormbugt #320 KAL-0026X01 321 F 45.3 15.8 2.9 -15.1 19.1 2012 D.E. Nelson, N. Lynnerup, and J. Arneborg 55 walrus (Odobenus rosmarus), as the Disko Bay area and that immediately to the south supported a large walrus population until they were virtually exterminated by commercial hunting in recent times (Born et al. 1994). However, to quote a present walrus expert, walrus are hard to kill, hard to butcher, hard to eat, and ridden with the parasite trichina (Trichinella spiralis, a species of roundworm), (M. Aquarone, University of Copenhagen Zoological Museum, Copenhagen, Denmark, pers. comm.). They were likely important to the Thule Culture for their hides, for ivory, and perhaps for dog food, but not as a human dietary staple. Narwhal (Monodon monocerus) and beluga (Delphinapterus leucas) would be seasonally available during their migrations, and may have congregated at the southern pack-ice limit in the winter, where they occasionally can be killed in large numbers when trapped at a shrinking openwater lead. The much larger baleen whales (e.g., bowhead [Balaena mysticetus] and North Atlantic right [Eubalaena glacialis] whales) would also have been seasonally present. Caribou (Rangifer tarandus) were cyclically available in abundance at least in the southern part of this region. Fish would also have been available, and there are several known large colonies of sea-birds in the area. The excavators’ faunal reports for the three sites in this region list seal, beluga, narwhal, caribou, and bird bones as occurring in these sites. Of the seal bones, those from ringed seals were predominant, but harp and harbor seal were also represented (Gulløv 2012 [this volume]; Mathiassen 1930, 1934). To the immediate south, the “Southwest region” encompasses the sites Asummiut, Utoqqaat and Qornoq (Grummesgaard-Nielsen 1997; Gulløv 1983, 2012 [this volume]; Heinemeier and Rud 1997; Lynnerup et al. 1997; Mathiassen 1931), all of which are on a coastline which at present has only local landfast ice in the winter, beyond which is open ocean. If the climatic circumstances in the past were similar, this region would be less favorable for ringed seals than that immediately to the north, but the boundary between pack ice and open water would be attractive to beluga and narwhal in the winter. In the spring and fall, the large migrating herds of harp seals would be available. As discussed above, hooded seals do not tend to get so far north in their migrations, but some may have been present, and one of the large Greenlandic congregations of walrus was to be found on this coast. Caribou cyclically occur, sometimes in great abundance in the interior areas, and Thule Culture use of these herds is well documented (Meldgaard 1986). Fish were present, and there are some large sea-bird colonies in the area. The excavators’ faunal reports for the sites studied in this region list harp seal, ringed seal, harbor seal (Phoca vitulina) and a few bearded seal bones, as well as walrus, sea-bird, fish, and numerous caribou bones (Degerbøl 1931, Gulløv 2012 [this volume]). We place the single site Uunartoq (Mathiassen 1936) in a region of its own, as it is on the very “Southwest tip” of the island, where pack ice from the east coast drifts around with the East Greenland Current to affect both winter and summer conditions. In this region, both harp and hooded seals appear in abundance during their spring and fall migrations. Narwhal, beluga, and walrus would not have been dietary factors, as their distribution does not go so far south. Summer would see the appearance of other whales and porpoises from the south. Caribou were present in this locale, but were few in number. Once again, fish (especially the great numbers of spawning capelin [Mallotus villosus]) were available, but there are no large sea-bird colonies. These observations are strongly reflected in the excavation faunal lists, which mention harp and hooded seals, cod, capelin, sea-bird, and caribou bone (Gulløv 2012 [this volume], Mathiassen 1936). The Uunartoq site is close to the Norse farm ruin site Ø149 at Narsarsuaq. The three sites Ruinnæsset, Suukkersit, and Skærgårdshalvø (Larsen 1938; Mathiassen 1933, 1936) are in the “Southeast region”, which is choked with pack ice for most the year. Ringed seal should then have been an important staple, and to a much lesser extent, the east coast populations of harp and hooded seals may have been as well. Both narwhal and beluga would have been present, with beluga predominating. There were no caribou and no resident walrus. There are no sea-bird colonies in this locale. The faunal lists from the sites mention seal, narwhal, beluga, and walrus bones (Degerbøl 1934; Gulløv 2012 [this volume]; Larsen 1938; Mathiassen 1933, 1936). The “Northeast region” includes the coast north of Ittoqqortoormiit/Scoresbysund, which is also choked by pack ice most of the year. It includes the sites Uunarteq, Kap Harry, Suess Land, Dødemandsbugten, Stormbugt, and Rypefjeldet (Amdrup 1909, Glob 1935, Larsen 1934, Thostrup 1911). This is the only locale studied here in which muskoxen (Ovibos moschatus) occurred, as did caribou, although the past distribution of both within the region is not well understood. This would be an area favored by local ringed seals, and northeast harp seals may have been available. The bearded seal is also found off these coasts. There are several large sea-bird colonies on the coast at Scoresby Land. Walrus occur, especially in the summer when males wander down the coast from the north. Narwhal are present, and some beluga would be expected in the southern parts of this region. These predictions from present observations are also reflected in the excavators’ faunal lists, which include ringed seal in abundance, some harp 56 Journal of the North Atlantic Special Volume 3 seal, a few bearded seal and walrus, narwhal, musk oxen, and caribou. It is interesting to note that the polar bear (Ursus maritimus) is also frequently mentioned. (Amdrup 1909, Degerbøl 1934, Gulløv 2012 [this volume], Larsen 1934,Thostrup 1911). This description of the protein resources in each of these locales is obviously greatly over-simplified, making no distinctions between some species and ignoring others. While recognizing this limitation, we proceed here to examine the human data with respect to the limited data set we have for Greenlandic animals. The intention here is not to provide a definitive dietary analysis, but to determine the extent to which the present animal data can provide explanations for the human results. Results and Interpretations The human data are presented in detail by locale (color) and by site (shape of symbol) in Figure 2. In the interpretations that follow, the reader should bear in mind that the isotopic data reflect protein intake by an individual over a long time period. In Greenland, animal fat and protein are the only sources of metabolic energy. A first observation is that the dataset as a whole certainly does confirm that these people were heavily reliant on marine protein. Most individuals have δ13C values within the range -12 to -14‰ and δ15N values from +18 to +22‰. In general, both are consistent with the values measured for Greenlandic marine animals Figure 2. The human data for all sites. 2012 D.E. Nelson, N. Lynnerup, and J. Arneborg 57 differences would also be most desirable, especially given the differences seen internally within this dataset and between the Greenlandic data and that for the Canadian Arctic (Coltrain et al. 2004). The bird data are limited to a few measures of thick-billed murre (Uria lomvia L.) from the Southwest, and we have no direct measures of fish bone. Indirect information on the two very abundant fish species capelin and arctic char (Salvenius alpinus) was obtained from measures of the flesh of modern specimens (D.E. Nelson and J. Møhl, unpubl. data) from which bone values for capelin (δ13C = -14.0‰; δ15N = +12.0‰) and char (δ13C = -13.0‰; δ15N = +14.0‰) can be inferred. These are very similar to those of the seals studied, although with lower δ15N values, as expected, and so to a first approximation, the seal data is a surrogate for the fish. While we can thus make no attempt to provide quantitative determinations of the relative proportions of dietary protein consumed from the various species, it should be possible to determine which were of fundamental importance and which were peripheral. For ease of reference, Table 3 reproduces the animal averages as determined in the companion study (Nelson et al. 2012 [this volume]). We begin with the sites in the Northwest region, and proceed counter-clockwise around the island. In the discussions that follow, the reader should note that, in order to include the data for the terrestrial animals, the scale of the plots is expanded to include a much larger range of values than that in Figure 1. For clarity, this scale changes a little from plot to plot. Northwest region Figure 3 plots the individual human data for the settlements at Illutalik, Nuugaaq, and Tunnungas- (Nelson et al. 2012 [this volume]). At this coarse interpretive level, the isotope dietary method does then provide reliable information. The nitrogen data are particularly distinctive. The δ15N values of the Greenlandic caribou and musk oxen are +2 to +3‰, while those of the seals are typically +15 to +18‰ (Nelson et al. 2012 [this volume]). As a second generalization, we can say that these people obtained only a very small fraction of their protein from terrestrial animals. Further, although there are far too few human males in the dataset to make a good comparison, all have isotopic values well within the ranges of those of the females at the same sites, and so there were apparently no male/female differences in the sorts of protein consumed. Do these data provide more detailed information? It would seem so. Immediately evident in Figure 2 are the differences between the different locales and even within the same locale; there is patterning within the data both on a larger and on a smaller scale. From a naive perspective, this observation is not expected for groups of hunters making fluid seasonal rounds within the same region, and certainly not for people engaging in regular long-distance movement between regions to exploit different resources. Such transhumance should lead to a homogeneous distribution of isotopic values—a more random scatter of values than is observed here. These data suggest that in Greenland, the Thule Culture population at a given site had a home territory within which they adapted their procurement strategies to exploit the local resources, and within which they stayed. These observations can be examined in more detail by discussing the regions one by one, using the animal isotopic data at hand as a basis for interpretation. In such interpretations, experience shows that there are trophic-level isotopic shifts of about + 0.8–1‰ for δ13C and 2–6‰ for δ15N between the bone collagen of the animals consumed and those of the consumer (for discussion, see the several contributions in Ambrose and Katzenberg 2000). For this first analysis, we chose the values 0.8‰ and 4‰ and use them to predict the isotopic values for humans consuming only protein from a given species. These values are sufficiently accurate that any corrections will be minor, and will not affect the interpretations to be drawn here. If the human has consumed a mixture of species, the shift will be from the weighted average for the different species. In making these interpretations, we note that the animal data we have available are not sufficient to provide a detailed isotopic dietary analysis, as they were not obtained for the purpose and are incomplete. In particular, we do not have adequate data for the whales, with only five measurements of large whales (all baleen?) and a single measure of a juvenile beluga or narwhal. A more detailed understanding of possible regional Table 3. Animal averages and standard errors from Nelson et al. 2012 (this volume). n = estimated number of individuals. The uncertainties are one standard error. δ13C δ15N n (‰ v PDB) (‰ v AIR) West coast Caribou 17 to 21 -18.2 ± 0.1 2.0 ± 0.2 Ringed seal 15 to 18 -14.1 ± 0.2 16.6 ± 0.4 Harp seal 9 to 13 -14.3 ± 0.2 14.5 ± 0.2 Hooded seal 11 or 12 -13.6 ± 0.2 15.8 ± 0.3 Bearded seal 4 or 5 -12.6 ± 0.2 13.5 ± 1.0 Harbor seal 8 or 9 -12.6 ± 0.1 17.0 ± 0.3 Walrus 9 -12.7 ± 0.1 11.7 ± 0.3 Narwhal or beluga 1 -12.6 16.1 Baleen whales 2 to 5 -14.4 ± 0.3 12.6 ± 0.8 Thick-billed murre 6 -16.1 ± 0.2 14.6 ± 0.5 East coast Caribou 6 -19.3 ± 0.1 1.5 ± 0.2 Muskoxen 6 -20.2 ± 0.1 2.8 ± 0.2 Ringed seal 5 -15.2 ± 0.1 14.3 ± 0.6 Harp seal 5 -15.7 ± 0.2 11.8 ± 0.9 58 Journal of the North Atlantic Special Volume 3 admixtures of harbor seal, beluga or narwhal, and perhaps harp seal would take the description into greater detail and provide an explanation for the observed scatter. On the basis of these data, caribou did not provide much protein to these people, nor did murre, bearded seal, or walrus. The human δ15N values are actually slightly higher (about 1‰) than is predicted from the average for the ringed seal. A possible explanation comes from an observation made in measuring these animals (Nelson et al. 2012 [this volume]) in which it was noted that juvenile animals had higher nitrogen values than their elders. Were these people targeting juvenile ringed seal? This speculation should be testable from an analysis of soq, the first two of which are neighboring sites in the Disko Bay area, and the third is on the coast to the north. The data for the 11 individuals form a well-defi ned group, all within a range of 1‰ (δ13C) and 3‰ (δ15N). The average animal isotopic values and their standard errors are also plotted. The arrow drawn on the plot represents the trophic level shift and, as discussed above, points to the result that is predicted for a human consuming nothing but meat from animals with the isotopic values given at the arrow’s origin. In this case, it is drawn from the average value found for the west coast ringed seals. As is seen, to a first approximation, this single species provides a suffi cient explanation for the human data. Adding small Figure 3. Human isotopic data for the three sites in the Northwest region. Each red point is the result for an individual. The average values and the standard errors for the wild animals in the region are given for comparison. The arrow represents a trophic level shift of + 0.8 ‰ (δ13C) and + 4.0 ‰ (δ15N), as discussed in the text. 2012 D.E. Nelson, N. Lynnerup, and J. Arneborg 59 the ringed seal bones in the middens. If the observed scatter within the human data represents real differences, we note that two individuals (those with the lowest δ15N values) have isotopic values similar to the people from the south (see Fig. 2). Southwest region Figure 4 gives a similar plot of the human data for the sites Asummiut, Qoornoq, and Utoqaat. Asummiut (Grummesgaard-Nielsen 1997) stands out as the only site in the project for which we have a large number of individuals, some 23 in all. Of these, eight were of women buried sequentially in a common gravesite (SIK 892 D7) over a period from about A.D. 1450 to A.D. 1650; another seven are also from a common grave (SIK 892 D1) in the same locale. There are two graves with three individuals each nearby (SIK 887) and two containing single individuals (SIK 937 and 939). For the three sites as a whole, the human isotopic data are quite different from those of their neighbors to the immediate north (Fig. 1). As a whole, they have higher δ13C values and lower δ15N values, and the range of C is twice that of their neighbors to the north. The observed differences are not random. In particular, the people at Asummiut form a tight, distinctive group with a few outliers. These outlying values include two of the three young children, and three of the six people Figure 4. Southwest region humans, plotted as described in Figure 3. 60 Journal of the North Atlantic Special Volume 3 indicate a dietary mixture of beluga or narwhal and harp seal is sufficient to explain the human data. Certainly, there may have been an admixture of ringed seal, bearded seal, and harbor seal, but their nitrogen values are a little too high for them to be primary contributors. Caribou provided very little protein to the diets of these individuals. The murre data were obtained on ancient bird bone from a Thule Culture site in this region and so were directly applicable here. Again, it is obvious that they are not primary contributors of protein. Southwest tip The data for the five individuals found at Uunartoq are given in Figure 5 as are the averages for the local species that are likely to have been of importance. This locale provides a good test of the method, as the situation is a little less complicated than in the two graves at SIK 887, the most recent of the sites. While we can make no more detailed attempt to correlate the isotopic data with archaeological data other than that given in Table 1, we note that those with a detailed understanding of this site may find such an examination interesting. One of the two individuals from Qoornoq has isotopic values within the range of those at Asummiut, while the other has a slightly more negative δ13C value, as does the single individual from Utoqqaat, and another from Asummiut. The available animal data for this region are also given in Figure 4. As before, the arrows represent the expected trophic level difference between the isotopic values of dietary species and consumer. In this case, no one species is compatible with the human data. Two arrows drawn from the averages for the harp seals and the single narwhal/beluga datum Figure 5. The Southwest Tip region, plotted as described in Figure 3. 2012 D.E. Nelson, N. Lynnerup, and J. Arneborg 61 data from this area and so must use that from other regions. Figure 6 plots the human data as well as the average isotopic values for the ringed and harp seals from both west coast and northeast coast sites. Although hooded seals were not mentioned in the excavation reports for the sites in this region, we include the west coast average for this species as well. The human data are very similar to those from the Southwest tip (see Fig. 1 for more detail) and they correspond much more closely to the averages for the west coast seals than they do to those for the Northeast animals. We do not have sufficient data on either humans or animals to explore this interesting observation further. However, one can use the Southeast human data to turn the arrow the other way around and thus predict that the Southeast seals will have isotopic values similar to those from the Southwest. that in the others. Neither caribou nor walrus were present in significant numbers, nor are there large sea-bird colonies in the area. Ringed seals would be available in the winter, as would large numbers of harp and hooded seals during their spring and fall migrations. The migratory patterns of both narwhal and beluga fall well to the north. The contribution of other migrating whale species is unclear, but need not be invoked. In this case, the base of the arrow is drawn at the average for the three seal species plotted, indicating that a human population obtaining equal amounts of protein from each of these species would have an isotopic value at the arrow’s head. Again, we need consider no other species to explain the human data. Southeast region We cannot make a direct interpretation for the people from the Southeast, as we do not have animal Figure 6. The Southeast region, plotted as described in Figure 3. 62 Journal of the North Atlantic Special Volume 3 Northeast region The data for the individuals from the six sites in the Northeast Region are plotted in Figure 7. This is the only region in which there were two terrestrial species, the caribou and the musk ox, which could have contributed significant amounts of protein to human diet. The limited animal data available for the Northeast is also plotted in Figure 7. As can be seen, it is insufficient to allow a good evaluation; in particular, we lack information on the local whales and walrus. Both ringed seal and harp seal are represented by only five individuals each (Nelson et al. 2012 [this volume]), and within each of these sets, there was evidence for differences, perhaps of a regional nature. On the other hand, the musk ox and caribou averages are relatively secure, even if they were made on only a few animals each, as the data fit well with the measures of both and domestic herbivores for the west coast (Nelson et al. 2012 [this volume]). As usual, the arrowhead in the plot points to the value expected for individuals consuming only protein from the animal at the arrow’s base. Once again, we observe that there is a range of human values and that there is patterning evident in the data. There is some correlation with site, and the data are not randomly scattered about a mean, but seem to be distributed about a line, unlike the data for other locales. The 10 individuals from Dødemandsbugten seem to fall into two distinct groups (see Fig. 1 for more detail) differing in both δ13C and Figure 7. The Northeast region, plotted as described in Figure 3. 2012 D.E. Nelson, N. Lynnerup, and J. Arneborg 63 δ15N by an amount corresponding to almost a trophic level, giving further evidence for intra-site differences. The human data are far too high in δ15N to reflect much protein from the harp seal. Even though we have no isotopic data for the Northeast, walrus, baleen whales, and bearded seal would not be high dietary contributors; as the west coast data on these animals so clearly illustrates, they feed lower on the food chain and thus will have even lower δ15N values than the local harp seals. Even the ringed seal data don’t fit as well here as they did for the people on the Northwest coast. Is our ringed seal sample biased because we have consciously selected adult animals, while the Thule Culture targeted more juveniles with higher δ15N values? Or would the missing narwhal isotopic data resolve this issue, as it is a toothed whale, and will thus be at a high trophic level? Or can one go a step farther and argue that polar bear formed a significant part of the local diet? These bears eat ringed seals and are thus a trophic level higher. We can draw no present conclusions other than that the individuals with the highest nitrogen values must have consumed marine protein of a very high trophic level, while those whose values fall progressively lower were hunting and eating significant amounts of caribou and/or musk oxen. While we do not have sufficient data to provide reliable quantitative determinations of amounts, these data are approximately consistent with a marine-to-terrestrial consumption ratio of about four or five to one for the individuals with the lowest isotopic values. Conclusions What has been learned from this exercise? We can draw various conclusions which range from those that are quite firm to some that are much more speculative and will require further research to examine. First, the human data indicate that the isotopic dietary method can provide reliable information on Thule Culture diet in Greenland. It indicates that the people studied here were very marine-adapted and perhaps surprisingly so; with the possible exception of the Northeast people, there is no need to include any terrestrial species to explain the present data. Next, the method appears to provide useful information at more detailed levels of analysis. While the varying isotopic signatures of the marine animal species will make determinations of simple marine/ terrestrial protein consumption ratios very complicated (e.g., for studies of the Norse in Greenland), this same observation may be very useful in studies of the Thule Culture and perhaps of their predecessors, the Dorset- and Saqqaq-culture people. The marine mammals have characteristic isotopic values that appear to be reflected in the people who eat them. In this study, there is clear patterning in the human data at the regional level, which must reflect differences in diet. Even the limited isotopic information we have for the animals can to a large extent provide explanations for the human data. For the most part, these explanations tend to be in accord with ecological descriptions of the regions and with the lists of faunal remains given by the excavators at the various sites. In the latter case, however, some re-adjustments in archaeological interpretation may be required. Were some animals hunted for something other than their meat? There is also patterning in the human data within regions and possibly within sites. The within-site patterning is most clearly evident at the two sites for which we have measurements on many individuals (23 at Asummiut and 10 at Dødemandsbugten). For each, the data are not all randomly distributed about some mean. As well, at some sites a few people have isotopic values that would fit better within regions other than those in which they were buried. These observations strongly suggest that the isotopic data reflect real differences at the level of the individual, and one is led to speculate whether there is evidence here for social differentiation or for changes over time, or perhaps for migration by individuals from one region to another. Are these observations in accord with those from traditional archaeology? We can only provide a naive answer here, but at first glance, this patterning seems to be at odds with that which might be expected for groups of generalist hunters making wide use of all the food resources available within a large region. For all but one of the regions, even the very limited animal isotopic data could provide a sufficient explanation for the human data, indicating that to a good first approximation, we need consider only a few species of marine mammals (one or more of ringed, harp, and hooded seal and likely narwhal or beluga) to predict the human results. This result seems contradictory to the concept of generalist hunters. In particular, this observation is at odds with the archaeological studies of the west coast Thule Culture, which indicate that these people routinely exploited the local caribou herds (Gulløv 2012 [this volume]). Were the individuals we have measured fortuitously from a period of low caribou abundance, or were the caribou primarily important for their antler, hides, and fat? The regional exception to the statement above is the Northeast, for which the human data and the limited sea-mammal data do not correspond so well. The best present explanation includes significant consumption of musk oxen and caribou and possibly of higher trophic level animals such as bears. There are also apparent differences between and within sites. Of the regions studied, this one remains the most enigmatic from an isotopic point of view. In summary, we conclude 64 Journal of the North Atlantic Special Volume 3 Heinemeier, J., and N. Rud. 1997. Kulstof-14 datering med accelerator-metoden. Tidsskriftet Grønland 5–7:232–237. Larsen, H. 1934. Dødemandsbugten: An eskimo settlement on Clavering Island. Meddelelser om Grønland 102(1):68–70. Larsen, H. (Ed.). 1938. Archaeological Investigations in Knud Rasmussens Land. Meddelelser om Grønland 119(8). 77 pp. Lynnerup N, J.C. Brings Jacobsen, J. Thorsen, and T. Kludt. 1997. Menneskeknoglerne fra Asummiut. Tidsskriftet Grønland 5–7:227–32. Mathiassen, T. 1930. Inugsuk, a medieval Eskimo settlement in Upernavik District West Greenland. Meddelelser om Grønland 77(4):165–166. Mathiassen, T. 1931. Ancient Eskimo Settlements in the Kangamiut area. Meddelelser om Grønland 91(1). 29 pp. Mathaissen, T. 1933. Prehistory of the Angmagssalik Eskimoes. Meddelelser om Grønland 92(4). 37 pp. Mathiassen, T. 1934. Contributions to the Archaeology of Disko Bay. Meddelelser om Grønland 93(2). 192 pp. Mathiassen, T. 1936. The former Eskimo settlements on Frederik VIs Coast. Meddelelser om Grønland 109(2). 58 pp. Meldgaard, M. 1986. The Greenland caribou: Zoogeography, taxonomy, and population dynamics. Meddelelser om Grønland, Bioscience 20:1–88. Nelson, E., J. Møhl, and J. Arneborg. 2012. Stable carbon and nitrogen isotopic measurements of the wild animals hunted by the Norse and the Neo-Eskimo people of Greenland. Journal of the North Atlantic Special Volume 3:40–50. Tauber, H. 1981. 13C evidence of dietary habits of prehistoric man in Denmark. Nature 292:332–333. Tauber, H. 1989. Age and diet of the mummified Eskimos from Qilakitsoq, Meddelelser om Grønland, Man and Society 12:137–138. Thostrup, C.B. 1911. Ethnographic description of the Eskimo settlements and stone remains in Northeast Greenland. Meddelelser om Grønland 44(4):183–355. Vibe, C. 1981. Pattedyr (Mammalia). Pp. 363–459, In F. Salomonsen (Ed). Grønlands Fauna. Nordisk Forlag, Copenhagen, Denmark. that the isotopic dietary method can provide dietary information at the regional, site, and individual levels. The data obtained here for these people provides an excellent isotopic counter-point for our studies of their Greenland Norse neighbors in the same regions. Literature Cited Ambrose, S.H., and A. Katzenberg. (Eds.). 2000. Biogeochemical Approaches to Paleodietary Analysis. Advances in Archaeological and Museum Science Vol. 5. Kluwer Academic/Plenum Publishers, New York, NY, USA. 286 pp. Amdrup, G. 1909. The former Eskimo settlements on the East Coast of Greenland between Scoreseby Sund and Angmagsalik District. Meddelelser om Grønland 28(6):314–315. Arneborg 2004. Det europæiske landnam–Nordboerne I Grønland 985–1450 e.v.t. Pp. 219–278, In H.C.Gulløv (Ed.). Grønlands Forhistorie. Copenhagen, Denmark. Born, E.W., and J. Böcher (Eds.). 2001. The Ecology of Greenland. Atuakkiorfik Undervisning, Nuuk, Grønland. 429 pp. Born, E.W., M.P. Heide-Jørgensen, and R.A. Davis. 1994. The Atlantic walrus (Odobenus rosmarus rosmarus) in West Greenland. Meddelelser om Grønland 40:1–33. Coltrain, J.B., M.G. Hayes, and D.H. O’Rourke. 2004. Sealing, whaling, and caribou: The skeletal isotope chemistry of Eastern Arctic foragers. Journal of Archaeological Science 31:39–57. Degerbøl, M. 1934. Animal Bones from the Eskimo Settlement in Dødemandsbugten, Clacering Island. Pp. 173–180, In H. Larsen (Ed.). Archaeological Investigations in Knud Rasmussens Land. Meddelelser om Grønland 119(8). 77 pp. Degerbøl, M. 1935. Animal bones from King Oscar Fjord Region in East Greenland. Pp. 93–97, In P.V. Glob (Ed.). Eskimo settlements in Kempe Fjord and Kong Oscar Fjord. Meddelelser om Grønland 102(2):5–92. Degerbøl, M., R. Hørring, and I.R. Pfaff. 1931. Animal bones from the Sukkertoppen District 1930. Pp. 134–139, In T. Mathiassen (Ed.). Ancient Eskimo Settlements in the Kangamiut area. Meddelelser om Grønland 91(1). 29 pp. Grummesgaard-Nielsen, S. 1997. Thulekuturens Grave. Tidsskriftet Grønland 5–7:198–226. Glob, P.V. (Ed.). 1935. Eskimo settlements in Kempe Fjord and Kong Oscar Fjord. Meddelelser om Grønland 102(2):5–92. Gulløv, H.C. 1983. Nuup kommuneani qangarnitsanik eqqaasutit inuit-kulturif nunaqarfu - Fortidsminder i Nuuk kommune inuit-kulturen bopladser. Kalaallit Nunaata Katersugasivia, and The Danish National Museum, Nuuk, Greenland. 240 pp. Gulløv, H.C. 2004. Nunarput, Vort Land–Thulekulturen 1200–1900 e.v.t. Pp. 281–343, In H.C. Gulløv (Ed.). Grønlands Forhistorie. Gyldendal, Copenhagen, Denmark. Gullov, H.C. 2012. Archaeological commentary on the isotopic study of the Greenland Thule culture. Journal of the North Atlantic Special Volume 3:65–76.