2012 D.E. Nelson, J. Heinemeier, J. Møhl, and J. Arneborg 77 Introduction In the late first millennium AD, the Scandinavian Norse introduced agriculture into the islands of the North Atlantic, settling in the Faeroes, the Shetlands, Iceland, and eventually Greenland, bringing with them all they needed to establish a pastoral economy in areas that are very marginal for agriculture. This was especially the case in Greenland, where the only suitable land is on the Southwest coast, at the very edge of the largest glacier in the Northern Hemisphere. As is described in detail in Arneborg et al. 2012a [this volume], they brought with them cattle, sheep, goats, and even pigs; horses were included for transport and dogs for herding and hunting. These domesticates formed the basis of an agrarian economy which was to persist for almost five centuries. In Greenland, plant husbandry was not possible other than on a very limited scale, and even animal husbandry was insufficient to provide for the dietary needs of the settlements. As a broad generalization, the faunal remains from Norse sites include approximately equal numbers of wild and domesticate bones (cf. McGovern 1985). As by far the greater numbers of the wild animals hunted were marine, we have the possibility to use isotope analysis to follow the domestic and wild protein through to the human diet. This approach requires establishing the isotope compositions for the animals consumed, so that the human isotope compositions may be interpreted using the animal values as a basis. A preceding paper in this volume (Nelson et al. 2012a) examined the isotope ratios of the wild prey species in Greenland. Here, we carry this analysis on to the domesticates. While limited carbon isotope information has been provided from past radiocarbon dating of Norse domesticates, this work was not done with dietary reconstruction as a goal, nor with methods designed for the purpose. No previous measurements of the stable nitrogen isotopes have been made. In this paper, we present and examine the isotopic data obtained from analysis of a large number of bones of the Norse Greenlandic domesticates. No attempt is made here to integrate these results into a broader context than that of direct interest to this project. In particular, we wished to address several inter-related technical and economic issues at varying levels of detail. The technical issues are directly concerned with the application of carbon and nitrogen isotope analysis to reconstruct human diet in Greenland. An introduction to the methodology is given in Arneborg et al. 2012a [this volume]. Other papers in this volume have attempted to characterize the isotope composition of non-domesticated fauna of Greenland (Nelson et al. 2012a [this volume]). Here, by pursuing the following questions, we aim to extend the focus to the domesticates that were introduced to Greenland by the Norse: - What are the isotope ratios of the animals introduced by the Norse into Greenland? In general, do these domestic herbivores have ratios typical of herbivores in a C3 plant environment, or are they different? - How variable are these ratios? - Are there differences between the domestic species? Isotopic Analyses of the Domestic Animals of Norse Greenland D. Erle Nelson1, Jan Heinemeier2, Jeppe Møhl3, and Jette Arneborg4,5,* Abstract - To provide a basis for the isotopic dietary study of the Greenland Norse, and as an interesting study in itself, measures of the stable carbon (δ13C) and nitrogen (δ15N) isotope ratios were obtained for 118 samples of archaeological bone from 6 species of the Norse domestic animals. These samples were obtained from museum faunal collections representing archaeological excavations of 10 Norse sites, five in each of the two Norse settlements in Greenland. In general, the carbon isotope values for the herbivores of dietary importance (cattle, sheep, and goats) were as expected for animals living in a C3 environment. The nitrogen isotope data hint at differing field management practices between the two settlements. There is no isotopic evidence for any unusual pastoral adaptation to conditions in Greenland, or for any change in animal management over the lifetime of the settlements. A few pigs form an exception to this statement, but they are peripheral to the Norse dietary economy. These data provide a solid first data set on which to base isotopic dietary analyses of the Norse settlers themselves. Special Volume 3:77–92 Greenland Isotope Project: Diet in Norse Greenland AD 1000–AD 1450 Journal of the North Atlantic 1FRSC Professor Emeritus, Simon Fraser University, Deptartment of Archaeology. Burnaby, BC Canada. 2AMS 14C Dating Centre, Department of Physics and Astronomy, Aarhus University, Ny Munkegade 120, DK-8000 Aarhus C, Denmark. 3Zoological Museum, Natural History Museums of Denmark, University of Copenhagen, Copenhagen, Denmark. 4Danish Middle Ages and Renaissance, Research and Exhibitions, The National Museum of Denmark Frederiksholms Kanal 12, DK-1220 Copenhagen. Denmark. 5Institute of Geography, School of GeoSciences, University of Edinburgh, Scotland UK. *Corresponding author - Jette.arneborg@natmus.dk. 2012 78 Journal of the North Atlantic Special Volume 3 - How do the isotope ratios of the domesticates compare to those of the wild terrestrial herbivores? - Since the climatic zones of Greenland range from the subarctic to the high arctic, and the annual precipitation can vary by a factor of ten between regions (e.g., Hansen 2001), can we see evidence for natural regional differences in the isotopes of Greenlandic terrestrial herbivores? The human economic issues center around the Norse practice of animal husbandry under the conditions present in this new land. Information on the Norse domestic economy is critical both for its own sake and for the technical background it provides. We sought to answer the following questions: - Is there evidence that Norse husbandry is refl ected in the isotope ratios of their animals? - Can we detect any differences between the domestic animals in the two settlements which reflect different management practices in the subarctic Eastern Settlement and the low arctic Western Settlement? - Is there any evidence that the early Norse colonists adjusted their animal husbandry to suit Greenlandic conditions? - Are there any long-term differences for the domesticates in either of the two settlements which might reflect changing practice or changing climate? Some of these questions may seem a little naive to those familiar with dietary isotope studies, but there is a potential unusual complication here. We know that even until recent times, farmers in the far north of Norway, on the North Atlantic islands, and on the Northern British Islands routinely used heather, bushes, horse manure, sea-weed, and even fish remains to supplement the hay fed to their cattle during the long winter (Amorosi et al. 1998:46-47 [with references], Balasse et al. 2009:12 [with references], Mørkved 1996). This “Vårknipa” or hungerfeeding practice has a long history (ibid), and there is thus good reason to believe that it might have been in use in Greenland. If so, it could certainly have had an impact on the isotope ratios of the animals. Answers to these various questions should allow us to examine the basic question that underlies interpretations of human isotopic data: - To what extent can we define characteristic carbon and nitrogen isotope ratios for the terrestrial protein consumed by the Norse? Knowledge of the average isotopic composition of the protein consumed by humans allows prediction of the human values, and so we attempt with this study to shed additional light on the answer to that question. The Sites and the Samples For this first study, we addressed these questions by selecting bone samples of each of the important domestic species from archival faunal remains at the Zoological Museum of the University of Copenhagen. While these collections are extensive, there were limitations. Most result from archaeological excavations undertaken over the past century, and there was often little or no information available on context or stratigraphic placement. Further, many bones were fragmentary. It was therefore not always possible to be certain that each fragment was that of a different individual, nor was it always possible to estimate the age of the individual animal. For some sites, it was necessary to choose samples that could only be identified as either sheep or goat. These two animals are very difficult to distinguish between by bone morphology, and the samples available did not always permit choice of those that were clearly one or the other. Table 1 gives a summary list of the sites from which samples were obtained. For a detailed review of these sites, the reader is referred to the introductory paper (Arneborg et al. 2012a [this volume]). Table 2 lists the numbers of specimens (n) of each species taken for study. An attempt to determine the number of individuals (N) represented was Table 1. Norse sites from which faunal samples were obtained. Norse place names are in italics. Site codes Danish National Greenland National Site name Museum Museum and Archives Western Settlement GUS 64V2-III-555 Naajaat Kuuat V63 64V2-III-506 Niaquusat V48 64V2-III-507 Nipaatsoq V54 64V2-III-502 Kilaarsarfik / Sandnes V51 64V2-III-511 Eastern Settlement Qassiarsuk / Brattahlid Ø29a 61V3-III-539 Igaliku / Gardar Ø47 60V2-0IV-621 Narsaq Ø17a 60V1-00I-518 VatnahverfiØ71 60V2-0IV-602 VatnahverfiØ167 60V2-0IV-603 Table 2. Number of specimens (n) and estimated number of individuals (N) of each species studied. Estimated Number of # of Common name Latin name specimens (n) individuals (N) Cattle Bos taurus 52 40 to 52 Sheep Ovis aries 25 16 to 25 Goat Capra hircus 17 11 to 17 Sheep or goat - 10 3 to 10 Pig Sus scrofa 4 4 Horse Equus caballus 6 4 to 6 Dog Canis familiaris 4 3 or 4 2012 D.E. Nelson, J. Heinemeier, J. Møhl, and J. Arneborg 79 made by taking into account the bone element, any information on animal biological age, find context, and any radiocarbon or derived archaeological dating information. We decided against attempting to use differences in isotopic values to further differentiate between individuals. Of the total of 118 samples, a minimum number of 88 individuals have been identified. Details of the samples themselves are given later in Tables 4–11. All samples are those of adult animals if not otherwise noted. The question of whether the Norse changed their husbandry practices to adapt them to Greenlandic circumstances and eventually to changing climate requires chronological control for these samples. While the overall chronology for Norse Greenland is well established from both historical and archaeological data, the details are not, as many of the Norse sites were excavated before stratigraphic methods were developed. To test for change over time, we could thus not rely on excavated information to provide adequate chronological detail. Nor was it practical to consider radiocarbon dating every sample, and so a combined approach was taken. The purpose was not to provide detailed chronologies for the Norse settlements, but simply to test the isotopic data for evidence of change over time. The time-span of Norse occupation of Greenland was divided into three parts, defined as the Early (I) ca. 980–ca. 1160 AD, Middle (II) ca. 1160–1300 AD and Late (III) ca. 1300–ca. 1450 AD periods of settlement. For those sites for which there was stratigraphic chronological information, the measured samples were then sorted into these three groups. A further 30 samples were radiocarbon dated (Arneborg et al. 2011a [this volume]) and could then be placed reliably into these same periods. Here, we make use of the results of these determinations for the cattle, the sheep, and the goats, as these are the animals of particular interest in this context. Measurement Procedures and Results For both stable isotope analysis and radiocarbon dating, the method used for collagen extraction is the same as that used throughout this project, as described in Takahashi and Nelson (2012 [this volume]). The specimens were sampled at the Zoological Museum using a slow speed drill to sample underlying material after the surface had been removed by milling. As with the other sampling, the size of the holes drilled and the amount of material removed decreased over the sampling period. Typically, 200– 300 mg of drillings were taken at the outset, decreasing later to 50–80 mg. All samples were taken to the Isotope Archaeology Laboratory at Simon Fraser University for further processing. For the stable isotope determinations, aliquots of about 50 mg of bone were used and the extracted collagen sent to the Earth and Ocean Sciences Department, University of British Columbia, for direct measurement of the carbon and nitrogen concentrations of the extract, the C/N ratio, and both δ13C and δ15N values. As usual, these latter are reported relative to the respective standards VPDB and AIR. As discussed in Takahashi and Nelson (2012 [this volume]), the uncertainties in this measurement procedure are about ±0.1‰ for δ13C and ±0.2‰ for δ15N. These are sufficiently small that measurement uncertainty does not impact on our basic interpretations. Even so, another source of measurement variability needs to be considered. Bone re-modeling during the lifetime of the animal could result in different isotopic compositions for different parts of the bone if the animal’s diet is not isotopically constant. To get an idea of the potential magnitude of such an effect, we made two measurements each on bones from three cattle, two adults and a neonate. For each of the three bones, both samples were taken from the cortex of the diaphysis with a spacing of about 1–2 cm. For Adult 1, the samples were from the proximal portion of the tibia; for Adult 2, the central part of the metatarsus; and for the Neonate, the proximal portion of the ulna. One might thus expect to see the greatest differences in Adult 1, less in Adult 2 and little or none in the Neonate. Further, one would expect the effect to be more strongly seen in the nitrogen isotope data than in the carbon isotope data. The results obtained are given in Table 3. The observed isotopic differences between samples from the same bone are not significant in view of the basic measurement uncertainty, except for the δ15N values of Adult 2 (1.0‰) and both the δ13C and δ15N values for Adult 1 (0.3 and 2.0, respectively). The large differences for both carbon isotope ratios Table 3. Comparisons of bone sampling location versus δ13C and δ15N. Sample ID Animal Bone element δ13C (‰ v PDB) δ15N (‰ v AIR) Individual #53 Adult 1 Ttibia -20.5 10.2 Identical to sample #354 #354 Adult 1 Tibia -20.8 8.2 Identical to sample #53 #155 Adult 2 Metatarsus -21.2 11.5 Identical to sample #355 #355 Adult 2 Metatarsus -21.3 10.5 Identical to sample #155 #159 Neonate Uulna -20.8 12.2 Identical to sample #356 #356 Neonate Ulna -20.9 11.9 Identical to sample # 159 80 Journal of the North Atlantic Special Volume 3 and nitrogen isotope ratios in Adult 1 likely reflect bone re-modeling. Since it was not possible to control for this extra variable in our sampling, we thus estimate that, at the level of the individual animal, differences in δ13C values of ≤0.3‰ and δ15N differences of ≤2‰ may not reflect significant mean dietary differences between individuals. As will be seen in the coming sections, differences of slightly greater magnitude are seen when comparing different populations. These observations will complicate detailed statistical comparisons of the data. Fortunately, such comparisons are not necessary for our purposes. For samples also destined for radiocarbon dating, the aliquot size was increased to about 200 mg, but the extraction procedure remained the same. The extracted collagen was then sent to the Aarhus University AMS 14C Dating Centre for age determination (see, e.g., Arneborg et al. 1999 and Sveinbjörnsdottir et al. 2010 concerning reservoir correction and Andersen et al. 1989 for technical details). Three of these extracted samples were also dated at the Lawrence Livermore National Laboratory (CAMS). In general, we found collagen preservation to be excellent, as can be seen in the measured C and N concentrations and the C/N ratios given for each sample in Tables 4 to 11. The very few exceptions to this were simply eliminated from further consideration. The time period to which the sample has been assigned (Arneborg et al. 2012a [this volume]) and the stable isotope data obtained is given in the remaining columns of Tables 4 to 11. Discussion The species of primary interest here are those that would have formed the basis of Norse dietary economy: the cattle, sheep, and goats (e.g., Enghoff 2003, McGovern 1985). Of these, the cattle were of great importance to the Norse, both in terms of the prestige that they proffered to the owner and the foodstuffs they could provide. They also required the greatest amount of care, in comparison to the more hardy sheep and the very hardy goats, and they were likely given the best pasture lands in the summer and the best fodder in the winter. The cattle The data obtained for the cattle from sites in the subarctic Eastern Settlement are given in Table 4 and Figure 1. It is immediately apparent that the δ13C values are remarkably constant, with a mean and standard deviation for the entire suite of <δ13C> = -20.2 ± 0.5‰. While it is tempting to examine this data for evidence of internal patterning (e.g., removing the young animals decreases the variability even Table 4. The Eastern Settlement cattle. The estimated number of individuals is based on archaeological information, bone element, and biological age of the animal. Dates are based on either radiocarbon dating (C14) or archaeological information (A). Extract Extract Extract δ13C δ15N Site Sample ID Bone element Period C% N% C/N (‰ v PDB) (‰ v AIR) Individual/context Qassiarsuk/Brattahlid, Ø29a #17 I (C14) 3.0 -20.2 4.2 Unique Igaliku/Gardar, Ø47 #152 Phalange 43.7 15.1 2.9 -20.3 4.9 Without information #153 Phalange 43.0 15.0 2.9 -20.6 5.1 Without information #161 Metatarsus, juv. 43.3 14.7 2.9 -19.2 1.6 Unique Vatnahverfi/Vatnahverfi, Ø71 #110 Scapula II (C14) 45.8 15.3 3.0 -20.1 4.9 Unique #111 Metatarsus II (C14) 43.8 16.7 2.6 -20.2 2.8 Unique #117 Metatarsus, juv. I (C14) 48.2 16.2 3.0 -19.5 4.4 Unique Vatnahverfi/Vatnahverfi, Ø167 #87 Cranium, neonate? 43.5 15.8 2.8 -20.0 6.1 Unique #91 Astragalus sin. 40.4 14.9 2.7 -20.5 3.5 House 1, room III #92 Astragalus sin. 44.5 16.5 2.7 -20.6 3.8 House 1, room III #93 Metacarpus sin, neonate 44.5 15.7 2.8 -19.1 4.2 House 1, room III #94 Phalange 44.6 16.1 2.8 -20.8 2.5 House 1, room III #108 Metacarpus, juv. I (C14) 46.6 15.9 2.9 -20.2 4.4 Unique #109 Phalange I (C14) 46.2 16.0 2.9 -20.3 3.2 Unique #112 Metatarsus 43.8 16.1 2.7 -20.1 3.3 Unique #113 Metacarpus sin. II (C14) 44.7 16.6 2.7 -20.4 2.7 House 1, room V #114 Calcarius III (C14) 43.8 16.6 2.6 -20.7 2.1 Unique #115 Radius, neonate? 46.8 15.4 3.0 -19.5 4.7 Unique #116 Metatarsus, neonate? 47.4 16.0 3.0 -20.1 3.5 Unique #118 Phalange 47.9 16.6 2.9 -21.0 2.2 Unique #119 Astragalus, juv. II (C14) 47.8 16.2 3.0 -19.9 3.7 House 1, room V #120 Costa I C14) 47.9 15.8 3.0 -20.4 2.6 House 1, room III n (number of specimens) = 22 Average -20.2 3.7 N (number of individuals) = 17 to 22 1 σ 0.5 1.1 2012 D.E. Nelson, J. Heinemeier, J. Møhl, and J. Arneborg 81 further to ±0.3‰), that is pushing the data farther than warranted, as measurement uncertainty (as discussed above) at such a level of statistical analysis becomes an important consideration. Such detail goes well beyond the needs of this study, and pursuing it would require more consideration in sample choice. In contrast, the nitrogen isotopes are more variable, with a mean and standard deviation of <δ15N> = 3.7 ±1.1‰. Here, we expect some of this variation to reflect age, as a newborn animal has the same δ15N as its mother, but the youngster’s value rises about 3‰ (a trophic level) higher while it suckles and then falls to the same value again after weaning if mother and offspring are eating the same food (cf. Fogel et al. 1989). There seems to be evidence of this in the data here. One can go through the list of juvenile animals and speculatively interpret their values in this light, which could be a factor in the reconstruction of Norse dietary habits. There is no evidence for significant isotopic differences between the animals at the different Eastern Settlement sites, and this limited data set thus suggests that the circumstances at each were similar. Farther north, the cattle from the low arctic sites of the Western Settlement (Table 5 and Figure 1) have δ13C values very similar to those in the south. Although the mean value <δ13C> = -20.5 ± 0.5‰ is probably different from that of the Eastern Settlement (Students t: P = 0.02), the difference is so small (0.3‰) that measurement uncertainty becomes an issue, and for our present purposes, this is not important. The same is not true for the nitrogen isotope data. While the mean (<δ15N> = 6.1 ± 2.2‰) differs from that of the southern animals, it is the range of values that is of greater interest. Some have much higher values than those normally expected for a terrestrial herbivore, and it is immediately clear from Figure 1 that there is patterning within the data set. Especially noteworthy in this respect are the animals from Niaquusat, V48. This anomaly mirrors that found in a study of modern Greenlandic caribou and plants (E. Nelson and J. Møhl, unpubl. data) in which a few specimens of modern grasses also exhibited very high nitrogen isotope ratios. The speculative explanation there was that those anomalies were due to a fertilization— deliberate or not—effect by local midden accumulations on the plants growing on them. Subsequent research actually indicates that high δ15N values could be due to fertilization practices applied to the infields (Commisso and Nelson 2010 and references therein). Figure 1. δ15 N values versus δ13 C values for cattle. Data points for juvenile/neonate individuals are marked with J. 82 Journal of the North Atlantic Special Volume 3 although these exhibit almost twice as much variability in their nitrogen values, as compared to the other groups. We had only five goat samples from the Eastern Settlement, all from the site Ø47. These are slightly different, but 3 to 5 individuals comprise too small a group from which to form statistically valid conclusions, especially as several were young animals. With this one small caveat, we can state that for the population means, there are no differences between these two species or locales that can be detected with present methods. Given these observations, it is not surprising that those samples which could only be identified as either sheep or goat also have these isotopic compositions. All sheep and goats can then be lumped into a single isotopic group with means of <δ13C> = -19.7 ± 0.4‰ and <δ15N> = 4.1 ± 1.1‰. As these values The sheep and goats Distinguishing between the bones of these two domestic species can be difficult, especially when the material at hand is fragmentary. The samples listed in Tables 6 and 7 are those for which identifi- cation could be made with some certainty. In order to increase sample size somewhat, a separate category of either sheep or goat was established, as given in Table 8. The data obtained for these two species at both Eastern Settlement and Western Settlement sites are so similar that we discuss them as a single group (Fig. 2). No statistical comparisons are necessary to note that the mean isotopic compositions for sheep from Eastern Settlement sites are identical to those from the Western Settlement. This observation can be extended to the goats from the Western Settlement, Table 5. The Western Settlement cattle. The estimated number of individuals is based on archaeological information, bone element, and biological age of the animal. Dates are based on either radiocarbon dating (C14) or archaeological information (A). Extract Extract Extract δ13C δ15N Site Sample ID Bone element Period C% N% C/N (‰ v PDB) (‰ v AIR) Individual/context GUS #54.2 Metatarsus dex. III (A) 43.2 15.7 2.8 -20.4 6.6 X1487, Room I #58 Carpal III (A) 45.4 16.7 2.7 -20.1 3.9 X1432, Room I #60 Metacarpus sin. III (C14) 44.9 16.2 2.8 -20.0 4.8 X0678, Room I #61 Humerus dex. 44.7 16.3 2.7 -20.9 8.6 X0577, Room I #104 Phalange I (A) 45.6 16.6 2.7 -20.4 4.2 Unique #265 Mandible dex., neonate 43.0 14.4 3.0 -19.6 3.4 Unique Naajaat Kuuat, V63 #154 Phalange 42.9 14.9 2.9 -20.9 5.2 Unique Niaquusat, V48 #32 Phalange I (C14) 44.2 16.0 2.8 -21.3 7.3 Can be same individual as #53/354 #37 Metatarsus I (C14) 43.0 15.5 2.8 -20.5 4.5 Unique #38 Metacarpus 43.4 15.8 2.7 -20.6 6.1 Unique #53/354 Tibia sin. 44.2 15.6 2.8 -20.6 9.2 Can be same individual as #32 #56 Radius sin. I (C14) 44.7 16.4 2.7 -20.8 5.6 Can be same individual as #59 #59 Costa 44.0 16.0 2.8 -21.3 6.4 Can be same individual as #56 #155/355 Metatarsus II (C14) 43.3 15.0 2.9 -21.3 11.0 Unique #157 Metacarpus III (A) 43.2 14.7 2.9 -20.6 5.8 Unique #159/356 Ulna, foetus/neonate III (A) 43.6 14.6 3.0 -20.8 12.1 Unique Nipaatsoq, V54 #151 Phalange III (C14) 43.4 15.3 2.8 -20.4 7.7 Unique #163 Radius, juv. III (C14) 43.7 14.3 3.1 -19.5 6.9 Unique Kilaarsarfik / Sandnes, V51 #156 Pelvis II (A) 42.9 14.7 2.9 -20.5 6.0 Unique #158 Metacarpus, neonate 43.4 14.6 3.0 -20.5 8.1 Unique #160 Metacarpus, juv. 43.8 14.3 3.1 -21.1 7.9 Unique #390 Metacarpus sin. II (A) 43.9 17.0 2.6 -20.4 5.5 Unique #395 Metacarpus sin. II or III (A) 44.5 17.0 2.6 -19.9 4.4 Unique #396 Femur sin. I (C14) 44.9 16.8 2.7 -20.2 3.3 Unique #399 Humerus dex. I (A) 44.8 16.7 2.7 -20.6 5.3 Unique #402 Phalange, juv. II (A) 45.2 17.1 2.6 -20.6 4.1 Unique #404 Tibia frag. I (C14) 45.1 17.1 2.6 -20.1 4.8 Unique #408 Cranial vault I (C14) 45.3 17.0 2.7 -20.4 3.1 Unique #409 Phalange I (C14) 44.9 17.0 2.6 -20.4 4.9 Unique #415 Metacarpus I (C14) 45.0 17.0 2.6 -20.0 7.4 Unique n (number of specimens) = 30 Average -20.5 6.1 N (number of individuals) = 25 to 30 1 σ 0.5 2.2 2012 D.E. Nelson, J. Heinemeier, J. Møhl, and J. Arneborg 83 the Norse diet. The pigs were simply too few, and both horses and dogs could not be eaten without offending dietary taboos (Egardt 1981). Nevertheless, we have measured a few samples of each for comparative purposes. The four pigs from the Eastern Settlement sites (Table 9) are clearly not exclusively terrestrial feeding animals, as both their carbon and nitrogen isotopes indicate that they consumed significant quantities of marine protein. Pigs are notorious omnivorous scavengers, and we can only speculate what these ate; perhaps their menu included the seal offal in the Norse middens. Investigations on pigs from the Western Isles of Scotland (Mulville et al. 2009:55), Gotland, Sweden (Kosiba et al. 2007:405), and Orkney (Richards et al. 2006:125) indicate that pigs’ fodder include marine sources. Actually, the pigs from Viking Age –Early Middle Ages coastal incorporate 52 measurements on 32 to 52 individuals, they provide a very good first estimate for these two species. Having made that observation, one can also attempt to detect patterning even within these small ranges of values. Only one stands out: two of the three goats from V48 have nitrogen isotope compositions higher than those of any of the other 49 samples (see Fig. 2). This observation complements and extends that for the cattle from this site. As these goats were fully-grown animals, this difference cannot be due to the suckling effect, and we can again postulate a fertilization effect for this farm. The pigs, horses and dogs Judging from the zooarchaeological records, these animals did not form a significant portion of Table 6. The sheep: the estimated number of individuals is based on archaeological information, bone element, and biological age of the animal. Dates are based on either radiocarbon dating (C14) or archaeological information (A). Extract Extract Extract δ13C δ15N Site Sample ID Bone element Period C% N% C/N (‰ v PDB) (‰ v AIR) Individual/context Eastern Settlement Narsaq, Ø17a #82 Metatarsus I (C14) 44.1 15.3 2.9 -20.4 3.8 Ruin 4, by water channel, lower layer #83 Metatarsus sin. I (A) 44.2 15.7 2.8 -20.3 4.2 Ruin 4, by water channel, lower layer #84 Metatarsus I (C14) 44.1 15.3 2.9 -20.1 3.6 Ruin 4, by water channel, lower layer Vatnahverfi, Vatnahverfi, Ø167 #95 Metacarpus 45.2 16.2 2.8 -20.2 2.5 Unique Igaliku, Gardar, Ø47 #281 Metatarsus dex. 43.8 14.6 3.0 -18.7 4.5 Without information #282 Metatarsus dex. 43.8 14.8 3.0 -19.5 3.9 Without information #283 Metatarsus dex. 44.1 14.7 3.0 -19.5 2.8 Without information #284 Metacarpus sin. 44.2 14.0 3.2 -20.0 5.1 Without information #285 Metacarpus 44.1 14.9 3.0 -20.1 2.9 Without information #286 Metacarpus sin. 44.3 14.9 3.0 -19.3 4.5 Without information #287 Metacarpus sin. 44.8 14.4 3.1 -19.2 3.9 Without information #288 Metacarpus 44.3 14.9 3.0 -19.9 2.9 Without information n (number of specimens) = 12 Average -19.8 3.7 N (number of individuals) = 8 to 12 1 σ 0.5 0.8 Western Settlement Nipaatsoq, V54 #41 Metatarsus III (C14) 44.6 16.1 2.8 -19.7 4.3 Unique #43 Metacarpus sin. I (C14) 44.0 16.0 2.8 -19.2 4.6 Unique GUS #62 Cornus sin. III (A) 43.7 15.8 2.8 -19.6 4.2 Unique #64 Metacarpus dex. II or III (A) 44.4 16.3 2.7 -19.5 3.6 Unique #65 Metacarpus dex. II or III (A) 44.8 16.2 2.8 -19.7 5.1 Unique #105 Astragalus, sin. I (A) 44.8 16.5 2.7 -20.4 5.0 Unique #275 Cornus sin II (A) 44.0 14.8 3.0 -20.0 2.7 Unique #276 Cornus sin II (C14) 44.0 14.6 3.0 -19.9 4.3 Unique #277 Pelvis sin 44.0 14.5 3.0 -19.7 2.2 Unique #278 Metacarpus dex 44.0 14.5 3.0 -19.9 4.7 Unique #279 Metatarsus dex 44.8 14.8 3.0 -20.1 4.0 Unique #280 Cornus dex 43.8 15.0 2.9 -19.6 3.7 Without information Kilaarsarfik, Sandnes, V51 #397 Tibia dist. sin. I (A) 44.4 16.8 2.6 -20.1 4.5 Unique n = 13 Average -19.8 4.1 N = 12 to 13 1 σ 0.3 0.8 84 Journal of the North Atlantic Special Volume 3 Table 7. The goats: the estimated number of individuals is based on archaeological information, bone element, and biological age of the animal. Dates are based on either radiocarbon dating (C14) or archaeological information (A). Extract Extract Extract δ13C δ15N Site Sample ID Bone element Period C% N% C/N (‰ v PDB) (‰ v AIR) Individual/context Eastern Settlement Igaliku, Gardar, Ø47 #270 Metacarpus sin. 44.2 14.7 3.0 -18.8 4.4 Without information #289 Metacarpus dex. juv. 43.2 15.8 2.7 -19.1 4.3 Without information #290 Metacarpus dex. juv. 43.0 15.5 2.8 -19.2 3.5 Without information #291 ??? juv. 43.2 15.7 2.8 -18.8 4.8 Without information #292 Metacarpus dex. 44.0 15.6 2.8 -18.9 6.1 Without information n (number of specimens) = 5 Average -19.0 4.6 N (number of individuals) = 2 to 5 1 σ 0.2 1.0 Western Settlement Niaquusat, V48 #33 Cornus III (A) 44.0 16.0 2.8 -19.8 6.5 Unique #39 Metatarsus II (A) 44.1 15.6 2.8 -19.9 5.3 Unique #40 Metacarpus 45.0 15.2 3.0 -20.0 7.9 Unique Nipaatsoq, V54 #44 Metatarsus sin. III (A) 45.0 16.5 2.7 -19.7 2.6 Unique GUS #63 Humerus sin. 40.5 14.8 2.7 -19.2 4.2 Without information #66 Metacarpus sin. 44.4 15.9 2.8 -19.2 4.5 Without information #268 Cornus sin. 44.1 14.5 3.0 -19.8 2.9 Without information #269 Humerus dex. 44.7 14.1 3.2 -19.7 2.8 Without information #271 Pelvis sin 44.1 14.4 3.1 -19.6 2.7 Without information #272 Humerus sin. 44.6 14.3 3.1 -20.4 5.5 Unique #273 Femur sin II or III (C14) 44.2 14.6 3.0 -19.5 3.5 Unique Kilaarsarfik, Sandnes, V51 #414 Metacarpus dist. II (A) 45.2 17.1 2.6 -19.2 4.0 Unique n (number of specimens) = 12 Average -19.7 4.4 N = 8 to 12 1 σ 0.3 1.7 Figure 2. δ15 N values versus δ13 C values for sheep and goats. 2012 D.E. Nelson, J. Heinemeier, J. Møhl, and J. Arneborg 85 domestic herbivores. While the δ15N values fit well within the range of the ruminant cattle, sheep, and goats, the δ13C values are all ≈1‰ lower. Is this due to their diet, or is it a result of the difference between Ridanäs on the Swedish island of Gotland very much resemble the Norse Greenland pigs. Even though only a few horses were measured (Table 10), they are slightly different from the other Table 8. Sheep or goat: the estimated number of individuals is based on archaeological information, bone element, and biological age of the animal. Dates are based on either radiocarbon dating (C14) or archaeological information (A). Extract Extract Extract δ13C δ15N Site Sample ID Bone element Period C% N% C/N (‰ v PDB) (‰ v AIR) Individual/context Eastern Settlement Narsaq. Ø17a #85 Tibia sin. I (A) 44.8 15.5 2.9 -20.2 4.5 Can be from same individual as #86 #86 Femur dex. I (A) 45.4 15.8 2.9 -20.1 4.8 Can be from same individual as #85 N (number of individuals) = 1 or 2 Western Settlement Kilaarsarfik, Sandnes, V51 #391 Phalange II (A) 44.3 17.1 2.6 -20.1 4.5 Can be from same individual as #392 #392 Humerus sin. II (A) 44.7 17.1 2.6 -19.8 3.8 Can be from same individual as #391 #393 Phalange I (A) 44.7 16.6 2.7 -19.8 2.8 Can be from same individual as #394 #394 Humerus sin. I (A) 44.9 16.6 2.7 -19.8 3.9 Can be from same individual as #393 #400 Scalpula dex. I (A) 44.6 16.9 2.6 -19.7 5.9 Unique #403 Metacarpus II (A) 44.7 16.7 2.7 -19.7 4.7 Unique #407 Calcaneus dex. I (A) 44.7 17.0 2.6 -19.4 3.8 Unique #410 Tibia dex. I A) 44.5 16.8 2.6 -19.5 4.4 Unique n (number of specimens) = 8 Average -19.7 4.2 N = 7 to 8 1 σ 0.2 0.9 Table 9. The pigs: the estimated number of individuals is based on archaeological information, bone element, and biological age of the animal. Extract Extract Extract δ13C δ15N Site Sample ID Bone element C% N% C/N (‰ v PDB) (‰ v AIR) Individual/context Narsaq, Ø17a #80 Mandibula 44.7 15.0 3.0 -17.2 11.7 Unique Vatnahverfi. Vatnahverfi, Ø167 #81 Tibia, dex. 43.7 15.4 2.8 -17.7 9.2 Unique Igaliku, Gardar, Ø47 #308 Long bone 42.7 15.1 2.8 -17.2 11.6 Without information #309 Long bone 42.9 15.5 2.8 -16.2 12.3 Without information n (number of specimens) = 4 Average -17.1 11.2 N (number of individuals) = 3 to 4 1 σ 0.7 1.4 Table 10. The horses: the estimated number of individuals is based on archaeological information, bone element, and biological age of the animal. Dates are based on either radiocarbon dating (C14) or archaeological information (A). Extract Extract Extract δ13C δ15N Site Sample ID Bone element Period C% N% C/N (‰ v PDB) (‰ v AIR) Individual/context Western Settlement GUS #57 Cranial vault III (A) 44.4 16.4 2.7 -21.2 5.6 Unique #67 Mandibula, dex. III (C14) 44.2 15.9 2.8 -21.2 4.6 Unique #266 Astragalus 44.1 14.7 3.0 -21.2 1.3 Unique #267 Mandible, dex. III (C14) 44.2 14.5 3.1 -21.0 3.8 Unique #307 Mandibula, sin 43.1 15.8 2.7 -21.2 2.3 Without information Eastern Settlement Igaliku, Gardar, Ø47 #149 Phalange, juv. 44.2 14.9 3.0 -21.1 4.5 Unique n (number of specimens) = 6 Average -21.2 3.7 N (number of individuals) = 5 to 6 1 σ 0.1 1.6 86 Journal of the North Atlantic Special Volume 3 sample as a whole, e.g., a sample dated as “late II or III” will be grouped with other Period III samples. As will be seen below, these choices do not affect the outcome. There are no differences in the mean Eastern Settlement cattle δ13C or δ15N values for Periods I and II. The two Eastern Settlement cattle of Period III are very slightly different, but this difference is far too small to enable any conclusions to be drawn. We do not have the data to compare the Eastern Settlement sheep/goats over time. Sufficient data are at hand to make more defi nitive statements about the Western Settlement animals. For both the cattle and the sheep/goats, the differences between the δ13C means for the three Periods are at most 0.1‰, and so there was no change for either group. While it is intriguing that the mean δ15N values for the Western Settlement cattle do apparently increase over the three Periods, it is evident from the high variabilities observed that these means mask non-random differences within the data. A review of the data in Table 5 shows that the largest differences are correlated to locale; in particular, the V48 cattle are unusual. (Two of the V48 goats also have the highest observed δ15N values in the Western Settlement). At all Western Settlement farms, there are some cattle with high δ15N values in the 7–9‰ range. It is clear that a large part of the variability seen in the nitrogen data for these cattle is linked to locale and to circumstance. There is insufficient data here to separate these factors from any change over time. We can only re-state that the Western Settletheir hindgut fermentation digestive system and the foregut system of the ruminants? Again, this observation is interesting but based on very limited data and is beyond the scope of this study. Like the pigs, the dogs (Table 11) had much marine protein in their diet. For some, their isotope ratios approach those of Greenlandic seals (Nelson et al. 2012a [this volume]). One from V48 is the most distinctive in this respect. Clearly, the protein from the terrestrial animals was too precious to waste on the dogs. Also, marine food sources, especially fish, deteriorate rapidly and are therefore likely to be have been thrown to the dogs in significant quantities. It is a common observation that dogs in association with humans with a marine diet exhibit similar marine isotope composition (see, e.g., Fischer et al. 2007, Noe-Nygaard 1988). Changes over time Can we find evidence in the isotopic data for any changes in husbandry practices? Of the domestic animals of dietary importance, those most likely to be affected are the cattle, which require more care than do the free-ranging sheep and goats. Given the narrow ranges already noted in the previous section for all three species, it seems unlikely that any detectable changes did take place. Table 12 gives the mean values for the data available for the cattle and sheep/goats (as a group) of the Early (I), Middle (II), and Late (III) Periods at each settlement. Those samples which span more than one Period are grouped with the one that most closely reflects the Table 11. The dogs: the estimated number of individuals is based on archaeological information, bone element, and biological age of the animal. Extract Extract Extract δ13C δ15N Site Sample ID Bone element C% N% C/N (‰ v PDB) (‰ v AIR) Individual/context Niaquusat, V48 #29 Metatarsal 45.1 16.5 2.7 -14.2 13.8 Unique #30 Pelvis 43.7 15.8 2.8 -13.2 16.6 Unique Igaliku, Gardar, Ø47 #305 Ulna 42.8 15.2 2.8 -15.8 12.7 Without information #306 Humerus 42.9 15.5 2.8 -14.8 14.1 Without information n (number of specimens) = 4 Average -14.5 14.3 N (number of individuals) = 3 to 4 1 σ 1.1 1.7 Table 12. Stable isotopic averages as a function of time. Eastern Settlement Western Settlement Period n <δ13C> ± 1σ <δ15Ν> ± 1σ n <δ13C> ± 1σ <δ15Ν> ± 1σ Cattle I 5 -20.1 ± 0.4 3.8 ± 0.8 10 -20.5 ± 0.4 5.0 ± 1.5 II 4 -20.2 ± 0.2 3.5 ± 1.0 5 -20.5 ± 0.5 6.2 ± 2.8 III 1 -20.7 2.1 7 -20.3 ± 0.4 6.8 ± 2.7 Sheep and goats I 5 -20.2 ± 0.1 4.2 ± 0.5 8 -19.7 ± 0.4 4.4 ± 0.9 II 7 -19.8 ± 0.3 4.2 ± 0.8 II or III and III 7 -19.6 ± 0.1 4.3 ± 1.7 2012 D.E. Nelson, J. Heinemeier, J. Møhl, and J. Arneborg 87 ment cattle were unusual, as compared to those in the Eastern Settlement, and that there may be an increase in their δ15N values over time, which could be due to cumulative effects of throwing out refuse or deliberate fertilization practices. Much more data will be required to test this tentative observation. Summary Discussion The data obtained provide an over-view of the isotopic compositions of the animals which formed the basis of the Greenlandic Norse domestic economy. As is evident from the tables of values presented in the preceding section, the results for the different species are consistent and definitive, and allow us to provide detailed responses to the questions posed at the outset. We address these questions here in turn. For reference, Table 13 gives a summary of the mean values for the domesticates as well as for their contemporary wild herbivores (Nelson et al. 2012a [this volume]). Questions: What are the isotopic compositions of the animals introduced by the Norse into Greenland? In general, do these domestic herbivores have compositions typical of herbivores in a C3 plant environment, or are they different? How variable are these compositions? The data for the Norse herbivores are very similar to those obtained for herbivores from most other areas of the world (Kosiba et al. 2007, Müldner and Richards 2005, Mulville et al. 2009, Noe-Nygaard et al. 2005, Richards et al. 2006), where typical δ13C values are around -22 to -21‰. The cattle, sheep, goats, and horses all have the characteristic carbon isotope composition (e.g., within the range of about -20 ±1‰) expected for herbivores in a C3 environment or 1–2‰ more positive. This result fits very well with that of a separate study of modern Greenlandic plants (E. Nelson and J. Møhl, unpubl. data) in which the mean δ13C value <δ13C> = -25.2‰ (corrected for the atmospheric impact of the Industrial Revolution) was obtained, giving the expected 5‰ difference between an animal’s diet and its bone collagen. For all these Greenlandic herbivores, the observed δ13C standard deviations of ≤0.5‰ are very small, indicating that the measurement procedures are reliable, that there is little dietary variation between individual animals, and that the calculated mean for each species is a robust determination. With some exceptions, the δ15N values for the individual animals are also within the range observed in other studies for herbivores in temperate climates (see, e.g., Noe-Nygaard et al. 2005). These exceptions, especially as is seen in the Western Settlement cattle, suggest real differences between animals and sites. Again, these nitrogen data mirror those obtained for modern Greenlandic plants (ibid) for which <δ15N> = - 0.4‰, but for which there also were a few samples with high anomalous values. Question: Are there differences between the domestic species? The mean carbon isotope data for the different domestic herbivores and locales all fall within a range of about 1.3‰. Of these, the important food animals—cattle, sheep, and goats—have means differing by ≤0.8‰. The horses may be 1‰ lower, but as they were not a part of the human diet, that is only of indirect interest here. This δ13C range is so small that measurement uncertainty will become an important consideration in more detailed comparative analyses of the species. Even so, the small differences are statistically significant and have also been of practical value, as in a few instances when the identity of a sample was challenged by the isotopic data and zoological re-examination of the bone resulted in re-assignment to another species. With the exception of the Western Settlement cattle, the mean δ15N values for the different species (Table 13) are identical within observed variation. In short, while there are some differences observed between species, they are very small. Questions: How do the isotopic compositions of the domesticates compare to those of the wild terrestrial herbivores? Can we see evidence for natural regional differences in the isotopes of Greenlandic terrestrial herbivores? As seen in Table 13, the mean values for the domestic animals are close to those of their wild Table 13. The average isotopic values for domestic and wild herbivores. Species Locale Number (N) of individuals δ13C ± 1σ (‰ v PDB) δ15N ± 1σ (‰ v AIR) Domestic Cattle Eastern Settlement 17 to 22 -20.2 ± 0.5 3.7 ± 1.1 Western Settlement 23 to 30 -20.5 ± 0.5 6.1 ± 2.2 Sheep / goats ES & WS 32 to 52 -19.7 ± 0.4 4.1 ± 1.1 Horses ES & WS 4 to 6 -21.2 ± 0.1 3.7 ± 1.6 Wild Musk oxen Northeast 6 -20.2 ± 0.2 2.8 ± 0.4 Hare Western Settlement 1 -20.6 3.8 Caribou Northeast 6 -19.3 ± 0.2 1.5 ± 0.5 Southwest 17 to 21 -18.2 ± 0.4 2.0 ± 0.7 88 Journal of the North Atlantic Special Volume 3 contemporaries. In particular, the δ13C means for the cattle from both settlements are identical within measurement variability to those of the musk oxen from Northeast Greenland and the single hare measured from a Western Settlement site. As the musk oxen are the closest wild equivalent of the cattle, it would appear that natural climatic or regional differences have little effect on the carbon isotope values (Nelson et al. 2012a [this volume]). Sheep, goats, and caribou forage more selectively than do cattle and musk oxen, and their δ13C values suggest that they chose vegetation with a slightly different mean value. The isotopic compositions for the Southwest caribou reflect their consumption of lichen (Nelson et al. 2012a [this volume]) which other herbivores cannot easily digest. It would be interesting to know whether sheep and goats are also consuming and digesting small amounts of lichen, or whether the difference from the cattle reflects a more selective foraging of other graze and browse species foraging. Other than the single hare measurement, the stable nitrogen isotopic compositions for the wild animals are lower than those observed for the domesticates. This is expected for the Southwest caribou, as lichen nitrogen values are also lower than those of the C3 plants (ibid). Is there a real difference in δ15N between plants in the high arctic Northeast and those in the low arctic and subarctic Southwest? We don’t have the data to directly answer this question, but it would appear from these results that any such difference between these locales is only about 1‰. With the exception of some Western Settlement δ15N values, the isotopic data for the different domestic species, the different sites, and the two settlements are very homogeneous. Both isotopes for the sheep and goats from all sites were so similar that they were combined into one group. As these two species were the most likely to have been left to fend for themselves, we can argue that any natural isotopic differences between Eastern Settlement and Western Settlement are smaller than our measurement resolution, and so the different climatic regimes in which these settlements were found are not reflected in the animals’ isotopic compositions. Questions: Is there evidence that Norse husbandry is reflected in the isotopic compositions of their animals? Can we detect any differences between the domestic animals in the two settlements which refl ect different management practices in the subarctic Eastern Settlement and the other low arctic Western Settlement? Of the domestic species of dietary importance, the cattle are most likely to reflect any such differences, but their carbon isotope ratios for the Eastern Settlement and Western Settlement differ by only 0.3‰. While this difference is statistically signifi cant, it is far too small to allow any substantive conclusions to be drawn about differences in cattle husbandry between Eastern Settlement and Western Settlement. The nitrogen data offer a first tentative indication that Norse farming practice may have affected the isotopic data of their animals. A possible explanation for the unusual Western Settlement δ15N values is that they result from fertilization of some Norse fields. Such fertilization could either be a natural occurrence due to animals grazing in the same areas and gradually fertilizing the soil through their urine and feces, as has been postulated for the caribou greens in Southwest Greenland (Fredskild and Holt 1993, Thing 1984) or it could be anthropogenic if the Norse were actively adding fertilizer to their hay meadows. As mentioned above, this interpretation seems to be confirmed by recent studies (Commisso and Nelson 2010). The strongest evidence for this comes from the small farm at Niaquusat W48, where cattle with high δ15N values were noted. While this farm was not representative of the mainstream Norse domestic economy (Arneborg et al. 2012a [this volume]), it is interesting to note that the isotopic results indicate, in accordance with the archeological evidence, that the domestic animals were confined to a limited grazing area close to the farm, leading to concentrated fertilization from refuse and from the grazing animals. Were the farmers using “Vårknipa” methods to feed their cattle in the winters? While feeding bushes, leaves and perhaps even horse manure to cattle would not have a large effect on the isotopic compositions of their bone collagen, any substantial inclusion of marine protein (e.g., from seaweed, fish, or seals) would leave an imprint on both isotopes. At least, one might expect to see the practice reflected in higher carbon and nitrogen isotope ratios for neonates if their mothers were fed marine protein during the winter in which the fetus was formed. The data available here indicate that marine protein was not a significant factor in winter-feeding of gestating cows. While the carbon data for the neonates and juveniles are very slightly higher than those for the adult animals at each farm, the difference is small, inconsistent, and not mirrored in the nitrogen data (Tables 4, 5; Fig. 1). Any consumption of marine protein by a gestating cow was at most a very small part of its diet. A single animal illustrates this conclusion: At Niaquusat W48, the Late Period fetus/ neonate with the extremely high δ15N value has the same δ13C value as the other animals from the farm, and so we can state with certainty that the high nitrogen value was not the result of its mother being fed marine protein. Questions: Is there any evidence that the early Norse colonists adjusted their animal husbandry to suit 2012 D.E. Nelson, J. Heinemeier, J. Møhl, and J. Arneborg 89 goats significant amounts of protein derived from the sea, as has been done in times of need elsewhere in far Northern Scandinavia. This statement includes Western Settlement cattle of the Late period. In summary, neither the carbon nor nitrogen isotope data provide evidence for unusual adaptations or for changes over time in the Eastern Settlement. For the Western Settlement, the carbon isotope data are also as expected and constant over time, while the variable nitrogen isotope data indicate different circumstances from those to the South. Although the nature of these differences is not known, there is no evidence in the isotopic data for any change with time. The few pigs form an exception to (and a confirmation of) these statements, as both their carbon and nitrogen isotope ratios clearly show that their diet included marine protein. Was this a conscious Norse adaptation to Greenlandic conditions or did the pigs work this out for themselves? We cannot say, but as pigs were an early curiosity rather than a dietary staple, this interesting observation has no impact on the analyses of the Greenlandic Norse dietary economy. Question: To what extent can we define characteristic carbon and nitrogen isotope end-members for the terrestrial protein consumed by the Norse? Greenlandic conditions? Are there any long-term differences for the domesticates in either of the two settlements which might reflect changing practice or changing climate? The isotopic data for the primary Norse domesticates in the Eastern Settlement provide no indication of unusual adaptations to suit Greenlandic circumstances. Both the carbon and nitrogen isotope data for these domesticates are as expected for herbivores consuming the local plants. The same holds true for the carbon isotope data for Western Settlement animals. The only evidence for differences in farming practice is that given by the high δ15N values for some of the Western Settlement cattle and goats, especially those at V48. We noted that higher nitrogen values occur in animals from the Early through to the Late settlement Periods, and that there may perhaps have been an increase over time in the values at V48. Do these differences reflect early or later changes in field management in the Western Settlement? Differences between the larger and smaller farms? Natural differences in the environment? More data will be required to answer these questions, but we can speculate that the Western Settlement animal data may reflect an outcome of field fertilization. The available data indicates that the Norse farmers were not winter-feeding their cattle, sheep, or Figure 3. Distribution of δ13 C values. 90 Journal of the North Atlantic Special Volume 3 form a homogeneous whole, and the δ13C mean value and standard error are thus very well characterized for both settlements at <δ13C> = -20.01 ± 0.06‰, i.e., with a precision which is well beyond the measuring accuracy. The situation may not be quite so clear-cut for a δ15N mean value, as even for the herbivores of interest here; these measures are not necessarily normally distributed about a central mean. Young suckling animals have higher values, as will those consuming unusual fodder, such as that from fertilized fields. How homogeneous are the nitrogen isotope data for these animals? Figure 4 plots the measured values for all the cattle, sheep, and goats from both settlements. Overlaying these data is a calculated Gaussian distribution with a mean and standard deviation of 4.0 ± 1.0‰. With the exception of a few Western Settlement cattle, this mean and standard deviation describe the observed distribution well. To a good first approximation then, the mean nitrogen isotope ratio of 4.0 ± 0.1‰ (at 1 standard error) can be applied to the primary domesticates of Norse Greenland, if it is borne in mind that there could be unusual scatter in the values for human consumers in the Western Settlement. We have a clear answer to this question. The isotope data obtained here for all Norse domestic animals of dietary importance form a very homogeneous data set. While there is evidence for very small isotopic differences between the cattle as one group and the sheep and goats as another, and perhaps even between the cattle in the different settlements, we can define meaningful isotope diet mean values for both carbon and nitrogen which are applicable to the domesticates as a whole. Figure 3 shows the δ13C results for the domestic species of dietary importance plotted as a histogram. The mean and standard deviation for the entire data set (cattle, sheep, and goats) is <δ13C> = -20.01 ± 0.57‰. While there is some internal patterning within this food resource (e.g., any humans who ate only beef would have very slightly different δ13C values than those who ate only mutton), the animal data as a whole are normally distributed. The solid curve drawn over the data in Figure 3 is the calculated Gaussian distribution with the observed average and standard deviation. The good fit is obvious. For humans randomly sampling from this resource, it is the standard error that is of importance in dietary calculations, as a single human will have consumed protein from many more individual animals than those represented in this study. These data Figure 4. Distribution of δ15 N values. 2012 D.E. Nelson, J. Heinemeier, J. Møhl, and J. Arneborg 91 Literature Cited Amorosi, T., P.C. Buckland, K.J. Edwards, I. Mainland, T. McGovern, J.P. Sadler, and P. Skidmore. 1998. They did not live by grass alone: The politics and palaeocology of animal fodder in the North Atlantic region. Environmental Archaeology 1:41–54. Andersen, G.J., J. Heinemeier, H.L. Nielsen, N. Rud, M.S. Thomsen, S. Johnsen, Á.E. Sveinbjörnsdóttir, and A. Hjartarson. 1989. Ams C-14 Dating on the Fossvogur sediments, Iceland. Radiocarbon 31(3):592–600. Arneborg, J., J. Heinemeier, N. Lynnerup, H.L. Nielsen, N. Rud, and Á.E. Sveinbjörnsdóttir. 1999. Change of diet of the Greenland Vikings determined from stable carbon isotope analysis and C-14 dating of their bones. Radiocarbon 41(2):157–168. Commisso, R.G., and D.E. Nelson. 2010. Stable nitrogen isotopic examination of Norse sites in the Western settlement of Greenland. Journal of Archaeological Science 37:1233–1240. Malasse, M., I. Mainland, and M.P. Richards. 2009. Stable Isotope evidence for seasonal consumption of marine seaweed by modern and archaeological sheep in the Orkney archipelago (Scotland). Environmental Archaeology 14(1):1–14. Egardt, B. 1981. Hästkött. Kulturhistorisk Leksikon for Nordisk Middelalder 7:280–281. Enghoff, I.B. 2003. Hunting, fishing and animal husbandry at The Farm Beneath The Sand, Western Greenland. Meddelelser om Grønland – Man & Society 28. The Danish Polar Center, Copenhagen, Denmark. 104 pp. Fischer, A., J. Olsen, M. Richards, J. Heinemeier, Á.E. Sveinbjörnsdóttir, and P. Bennike. 2007. Coastinland mobility and diet in the Danish Mesolithic and Neolithic: Evidence from stable isotope values of humans and dogs. Journal of Archaeological Science 34:2125–2150. Fredskild, B., and S. Holt. 1993. The West Greenland “greens”: Favourite caribou summer grazing areas and late Holocene climatic changes. Geografisk Tidsskrift 93:30–38. Fogel, M.L., N. Tuross, and D.W. Owsley. 1989. Nitrogen isotopes tracers of human lactation in modern and archaeological populations. Annual Report of the Director, Carnegie Institution of Washington Yearbook 89:111–117. Hansen, B.U. 2001. The Climate. Pp. 66–83, In E.W. Born and J. Böcher. (Eds). The Ecology of Greenland. Atuakkiorfi k, Nuuk, Greenland. 429 pp. Kosiba, S.B., R.H. Tykot, and D. Carlsson. 2007. Stable isotopes as indicators of change in the food procurement and food preference of Viking Age and Early Christian populations on Gotland (Sweden). Journal of Anthropological Archaeology 26(3):394–411. McGovern, T.H., 1985. Contribution to the paleoeconomy of Norse Greenland. Acta Archaeolgica 54:73–122. Muldner G., and M.P. Richards. 2005. Fast or feast: Reconstructing diet in later medieval England by stable isotope analysis. Journal of Archaeological Science 32(1):39–48. Mulville, J., R. Madwick, R. Stevens, T. O´Connell, O. Craig, A. Powell, N. Sharples, and M.P. Pearson. 2009. Isotopic Analysis of Faunal Material from South Uist, Western Isles, Scotland. Journal of the North Atlantic 2:51–59. Conclusions Stable isotope measurements of bones of the domestic animals from the two Norse settlements in Greenland have provided answers to the basic technical questions underlying application of stable isotope analyses to dietary studies. For the domestic species of importance to Norse diet, it is possible to clearly define characteristic mean values for both δ13C and δ15N of bone collagen. While there are small isotopic differences between species and there may be small differences between locales, these are not large enough to impact dietary reconstruction for the humans who consumed these animals. The results also provide archaeological information on Norse animal husbandry in Greenland. First, we find no evidence that the Norse were in general required to adopt any unusual farming practices to maintain their cattle, sheep, and goats. Second, while the small differences between the animals in the subarctic Eastern Settlement and the low arctic Western Settlement may indicate slight differences in practice, it will require detailed studies to further investigate this possibility. In particular, there is tantalizing evidence that some of the Western Settlement animals reflect direct or indirect field fertilization. Third, the domesticates of marginal or no dietary importance had isotopic compositions unlike the other groups. It is evident that marine protein formed a significant and sometimes dominant part of the diets of the pigs and dogs. Horses had isotopic compositions a little different from those of the other herbivores, and it is not clear whether this is due to their fundamentally different digestive system or whether they fed on a slightly different selection of plants. Last, other than the possibility of fertilization effects, we can find no evidence for any chronological changes in the isotopic compositions of the animals. Our data include measurements on animals from the early settlement period through to the mid-14th century, by which time the people in the Western Settlement are thought to have been leaving. Present reconstruction of settlement abandonment places emphasis on the impact of worsening climate on the Norse life-style and on their apparent unwillingness to change from an economy based on animal husbandry to one based on hunting. If this were the case, one might have expected to see evidence for change in the animals’ isotopic compositions, as one would expect that the Norse would make increasingly desperate attempts to keep their domestic stock and thus their life-style in good health. Either the isotopic method cannot detect any such changes that were made, or the Norse simply gave up without such a struggle, or the emphasis on climate change is exaggerated in present reconstruction. 92 Journal of the North Atlantic Special Volume 3 Mørkved, B., 1996. Vårknipa i nordnorsk tradisjon, med hovedvekt på forholdene i Troms og Loftoen/Vesterålen. Polarflokken 20(1):9–18. Noe-Nygaard, N., 1988. δ13C-values of dog bones reveal the nature of changes in man’s food resources at the Mesolithice-Neolithic transition, Denmark. Chemical Geology (Isotope Geoscience Section) 73:87–96. Noe-Nygaard, N., T.D. Price, and S.U. Hede. 2005. Diet of aurochs and early cattle in southern Scandinavia: Evidence from 15N and 13C stable isotopes. Journal of Archaeological Science 32(6):855–871. Richards M.P., B.T. Fuller, and T.I. Molleson. 2006. Stable isotope palaeodietary study of humans and fauna from the multi-period (Iron Age, Viking and Late Medieval) site of Newark Bay, Orkney. Journal of Archaeological Science 33(1):122–131. Sveinbjörnsdóttir Á.E., J. Heinemeier, J. Arneborg, N. Lynnerup, G. Ólafsson, and G. Zoëga. 2010. Dietary reconstruction and reservoir correction of 14C dates on bones from pagan and early Christian graves from Iceland. Radiocarbon 52(2–3): 682–696. Thing, H. 1984. Feeding ecology of the West Greenland caribou (Rangifer tarandus groenlandicus) in the Sisimiut – Kangerlussuaq region. Danish Review of Game Biology 12(3):1–53.