Journal of the North Atlantic T.D. Price and H. Gestsdóttir 2018 Special Volume 7 146 Introduction As noted in previously in this special volume, the Icelandic Sagas recorded that the first settlers came to Iceland after AD 874, that they came from western Norway, and that they came during the first 60 years and then stopped. As Vésteinsson has reported (e.g., 1998; Vésteinsson and Gestsdóttir (2016, this volume), several lines of evidence dispute these facts. Reinterpretation of these documents and source criticism brought revisions to the rather specific time and the place of origin of the settlers as stated in the Sagas (Lárusson 1929, Vésteinsson 1998, Vésteinsson and McGovern 2012). The dating of the landnam tephra to the 870s AD reaffirmed that there were no earlier traces of settlement on the island (Grönvold et al. 1995, Zielinski et al. 1997). Recent genetic studies of the origins of the Icelanders suggest that substantial numbers of the early Icelanders were born in parts of the northern British Isles. These investigations have examined both modern (Helgason et al. 2000, 2001) and ancient (Helgason et al. 2009) DNA and arrived at similar conclusions. The project reported in these pages is intended to examine the questions of the arrival and homeland of the first settlers of Iceland from another direction. We measured strontium, oxygen, and carbon isotopes in the teeth of archaeological burials from Iceland to determine whether these individuals were local or foreign born. In addition, carbon and nitrogen in bone collagen were measured in the radiocarbon samples that were obtained and are also reported in this article. Iceland was settled in the late 9th and the 10th century AD, with most of the settlers probably arriving from Norway, the Scottish isles, and Ireland. The earliest settlers were pagan, although there are references to Christians among them. The written sources suggest that the Pagan period covered approximately the first 130 years of the settlement of Iceland until AD 1000. The sources stated that the Christianization of Iceland took place by public consent at the Alþing around the year AD 1000 (Karlsson 2000:12–16, 33). Burial archaeology in Iceland then tends to be split into two, probably over-simplified, groups. These are the Pagan Viking Age burials on the one hand (referred to as kuml in Icelandic literature) and Christian cemeteries on the other. The Viking Age burials date to the early period of settlement in Iceland and are replaced by the first Christian cemeteries following the conversion of the Icelanders to Christianity. We have analyzed virtually all the available adult skeletal remains from this time interval where enamel is preserved and report our results here. The initial project involved the study of the settlement of Iceland, and part of the data has already been published (Gestsdóttir and Price 2006). Further sampling was undertaken from Icelandic archaeological skeletal material following the promising results from the initial analysis. In addition, we have obtained radiocarbon dates on a number of the skeletons and also report that information here in the context of the timing of colonization. We report the latest results of our investigations of the place of birth of settlers, the period of colonization, and other aspects of human movement to, on, and from Iceland. The study concludes with a discussion of the nature of Icelandic burial archaeology in light of the results of the strontium isotope analysis. 1Laboratory for Archaeological Chemistry, University of Wisconsin-Madison. 2Institute of Archaeology, Bárugata 3, 101 Reykjavík, Iceland. *Corresponding author - tdprice@wisc.edu. Viking Settlers of the North Atlantic: An Isotopic Approach 2018 Journal of the North Atlantic Special Volume 7:146–163 The Peopling of the North Atlantic: Isotopic Results from Iceland T. Douglas Price1,* and Hildur Gestsdóttir2 Abstract - Iceland was colonized by settlers from the North Atlantic rim of Europe near the end of the first millennium AD. This ws a remarkable achievement and the subject of much discussion. Historical documents, the Sagas, suggest that the settlers came from western Norway and all arrived within a brief period after which no further settlement took place. Genetic data, both modern and ancient, suggests that the settlers came from several places in Scandinavia and the northern British Isles and Ireland. We have used isotopic proveniencing, focusing on strontium, oxygen, and carbon in tooth enamel from early burials on Iceland to examine questions of place of origin. We have dated a number of these burials to discuss questions of the timing and pace of arrivals. Carbon and nitrogen isotope ratios in bone collagen were measured along with the radiocarbon date. Our data indicate that the settlers came from several different places, that settlement continued until around AD 1000 and stopped around the time of the transition to Christianity. We can also suggest that there was movement in both directions across the Atlantic. Changes in diet are suggested with greater consumption of marine foods over time. Some differences in diet are also related to the location of settlement, whether coastal or inland. Journal of the North Atlantic T.D. Price and H. Gestsdóttir 2018 Special Volume 7 147 Early Burials on Iceland The kuml burial grounds are, in all instances, small sites. They tend to be a cluster of graves, with the largest found to date containing 14 excavated burials. They are always inhumations, frequently aligned in a linear pattern, with the individuals buried either in a supine or a fetal position. They are often outlined with rows of stones, and were probably never covered with more than a small mound. The location of the plots seems quite significant. They are usually found just outside the home field of farms or on the junction of farm boundaries and/or roads. Grave goods are common, most often containing animal sacrifices (usually horses), but weapons, knives, combs, or beads are also common finds. Although single inhumations are frequent finds, the current thought is that these are probably not the norm, because in most instances where sites with single burials have been re-visited, more burials have been discovered at the site (Eldjárn 2000:258–288, Friðriksson 2009:12–20). Kuml have frequently been disturbed in antiquity. These disturbances, often referred to in the literature as “robbing”, have received surprisingly little attention. They have been dismissed as being due to people searching for loot. However, the frequency and varied nature of these disturbances is such that it is more likely that they may have been a part of the funerary process (Pétursdóttir 2009). The Christian cemeteries are of a different nature. The earliest excavated sites tend to be circular, usually surrounded by some sort of boundary no more than 30 m in diameter. The burials are aligned east–west, surrounding and aligned with a central church. The graves themselves are quite homogenous. The individuals buried there are in a supine position, some are found within simple wooden coffins, but there are no grave goods (Zoëga 2008:66–68, Þórðarson 1943:134–136). These early medieval cemeteries were most likely farm -based burial plots, serving the farm they stood on, or perhaps the neighboring farms (Vésteinsson and Gestsdóttir 2011). By the 13th century, these smaller home-plots seem to mostly go out of use, with increased centralization of the church and the establishment of parish-based cemeteries (Vésteinsson 2000:92). The nature of the burials within these later sites is similar to the earlier cemeteries. However, they often have a much longer period of use, and many are still in use today. Thus the complexity of the burials there is often much greater than within the earlier smaller plots (cf., Hallgrímsdóttir 1991:119–120). The dating of the earliest Christian cemeteries is often problematic. Documentary sources and/or tephrochronology (volcanic ash dating) can often give us a clear date as to when they went out of use, but the date of when they came into use is often less certain, and is often not based on any more solid evidence than the documented date of the conversion of Icelanders to Christianity. Recently, there has been an increase in the radiocarbon dates available; however, there are problems associated with these dates, as will be discussed below as well as in the paper by Ascough et al. (2014, this volume). The Samples In selecting the skeletal material, focus was primarily on adult individuals from the early Viking Age as the aim of the project was to identify the early settlers of the island. Approximately 200 individuals from the pagan period or kuml have been excavated on Iceland (Gestsdóttir 2004). A total of 127 skeletons were eventually sampled for the present study, including 99 from the pagan period. Of these specimens, about 30% are well preserved. There were no children included in the sample; the youngest individuals fell in the 18–25 age group. Approximately 68% of the sampled individuals are male and 32% are female (Gestsdóttir 1998). These burials are found around Iceland, often within a few kilometers of the coast (Fig. 1). The burials come from 54 sites. A large number of the samples, 25 skeletons, came from single inhumations. Where there were multiple burials from the same site, an attempt was made to sample all the individuals. In all instances where skeletons from burials within grave groups have been left out, it has been because either there was no preserved dental enamel, or the skeletons were inaccessible for sampling (Gestsdóttir and Price 2006). An attempt was made to obtain an even geographical distribution of samples. However, this strategy was hampered by concentrations of the burials, with 2 major clusters of excavated gravesites, one in the north of Iceland and the other in the southwest of the country. Thus a large proportion of the samples are from these areas. As previously noted, the pagan period or kuml in Iceland is traditionally considered to last from the first settlement towards the end of the 9th century until AD 1000. Of the pagan period skeletons sampled here, 22 (47.8%) have datable grave goods. In most instances where there are no grave goods, there is no clear archaeological dating for these burials (Eldjárn 2000). An additional 44 individuals were sampled from 2 early Christian cemetery sites; 33 skeletons from the site of Skeljastaðir in southern Journal of the North Atlantic T.D. Price and H. Gestsdóttir 2018 Special Volume 7 148 Iceland dated to the 9th to the early 11th century (Sveinbjörnsdóttir et al. 2010, Þórarinsson 1968, Þórðarson, 1943) and 11 skeletons from the site of Haffjarðarey in western Iceland dated to the 11th– 16th century (Steffensen 1945). For the Christian burials, the remains from 2 cemeteries, Skeljastaðir in Þjórsárdalur and Haffjarðarey in Haffjörður (Fig. 1), were analyzed. The cemetery at Skeljastaðir was excavated in 1939 by the then state antiquarian Matthías Þórðarson, as a part of a Nordic project involving the excavation of 8 Viking age farms in Þjórsárdalur (Þórðarson 1943). The cemetery had been greatly disturbed before the excavation. Records from last decades of the 19th century detail the erosion of the site (Jónsson 1885), and in 1935 Eiður Kvaran, an anthropologist at the University of Greifswald, Germany, carried out an excavation at the site. He excavated at least 30 skeletons and took them to Germany, where they were subsequently lost during the Second World War (Þórarinsson 1968). Skeljastaðir is not mentioned in any documentary sources, but the oral tradition states that Skeljastaðir cemetery had served all of Þjórsárdalur (Jónsson 1885). Radiocarbon dating suggests that Skeljastaðir cemetery may have come into use as early as the late 9th century AD, continuing into the early 13th century (Sveinbjörnsdóttir et al 2010:688). Previous work suggested that burials in the cemetery ceased when Þjórsárdalur was abandoned due to the AD 1104 eruption of Mt. Hekla (Þórarinsson 1968). On the other hand, there are 2 previously radiocarbon dated skeletons from the site with a date range of 1155–1220 (68.2% probability), which appear to contradict this (Sveinbjörnsdóttir et al. 2010:688). Fifty-six skeletons from the cemetery at Skeljastaðir are preserved in the Icelandic National Museum, and 33 of these were sampled. The samples used in this study from Skeljastaðir are listed with more information in Supplemental Table 1 (available online at https://eaglehill.us/jonaonline/supplfiles/ jona0072-price&gestsdottir-s1.pdf, and, for BioOne subscribers, at http://dx.doi.org/10.1656/ jona0072.s1). The earliest documented reference to the cemetery in Haffjarðarey dates to AD 1223. Although it is not known when the cemetery first came into use, it is known is that it went out of use in AD 1563, and so it was probably in use for approximately five centuries, ca. AD 1200–1563 (Steffensen 1945). The island is severely affected by erosion, and sources from the early 18th century Figure 1. Sites sampled for human tooth enamel for isotopic analysis on Iceland. Journal of the North Atlantic T.D. Price and H. Gestsdóttir 2018 Special Volume 7 149 mention exposure of human skeletal remains in the cemetery. Bones were first removed from the cemetery in Haffjarðarey in 1905, when Vilhjálmur Stefánsson recovered at least 50 skulls that lay on the surface and took them with him to the United States. There are also records of medical students removing bones from the site for comparative material. In AD 1945 Jón Steffensen and Kristján Eldjárn excavated the bones from the Haffjarðarey cemetery that are today preserved in the Icelandic National Museum. They excavated a total of 24 in situ skeletons and removed skeletal material representing at least 34 individuals, so the total collection represents 58 individuals. The extent of the erosion of the cemetery prior to the excavations in 1945 means that it is difficult to determine how large a proportion of the original population was recovered (Steffensen 1945). The lack of any dating evidence from the excavation means that it is not known when in the period of use of the cemetery the excavated skeletons date. A total of ten skeletons from Haffjarðarey were sampled. For further information on the Haffjarðarey skeletons sampled, see Supplemental Table 1 (available online at https://eaglehill.us/jonaonline/supplfiles/ jona0072-price&gestsdottir-s1.pdf, and, for BioOne subscribers, at http://dx.doi.org/10.1656/ jona0072.s1). Radiocarbon Dates Radiocarbon dates and light isotope (carbon and nitrogen) ratios were obtained for 37 samples from the Iceland burials (Supplemental Table 2 (available online at https://eaglehill.us/jonaonline/suppl-files/ jona0072-price&gestsdottir-s2.pdf, and, for BioOne subscribers, at http://dx.doi.org/10.1656/ jona0072. s2). These samples were measured at the Scottish Universities Environmental Research Centre AMS Facility (SUERC). The atomic ratio of carbon to nitrogen (C:N) is often used as one criterion to assess the quality of collagen preservation (Ambrose and DeNiro 1999, van Klinken 1999). The C:N ratio in modern collagen is approximately 3.2 (Ambrose and DeNiro 1999). The C:N values in the 37 samples from Iceland averaged 3.4 ± 0.2, with 4 values at 3.6 or above, 2 at 3.6, a single value at 3.8, and another at 4.3. The Oxford Radiocarbon Laboratory uses values between 3.1 and 3.5 for acceptable samples. This range would indicate that all but 4 of the Iceland samples had well-preserved collagen and that those dates could be considered as reliable. For other labs and scientists, values of 2.9–3.6 are generally accepted as indicative of unaltered collagen. Ambrose and DeNiro (1999) and DeNiro (1985) found a C:N values of 2.9–3.6 in well-preserved prehistoric human bone and fresh collagen, respectively. This range would include all but 2 of the Iceland samples. The radiocarbon dates were calibrated using the OxCal program, version 3. The 37 radiocarbon dates have a wide range of years. There are, however, some problematic dates within the set, including a handful of dates that predate the settlement of Iceland. These are from the sites of Vatnsdalur (640–730), Tyrðilmýri (650–780), Hringsdalur (660–780), and Kálfskinn (660–820). There are several problems associated with radiocarbon dating on Iceland (see for example Ascough et al. 2007; 2010; 2012; 2014, this volume). One of these is the effects of the East Greenland Current, which carries 14C depleted waters relative to the Atlantic currents. This means that any individuals with a high marine diet derived from these waters will always be dated too old (Ascough et al. 2007; 956). The effects of the East Greenland current are most likely to have been felt in the northern and western part of Iceland. The 4 skeletons in the current sample are from the northern (Kálfskinn), and western (Vatnsdalus, Tyrðilmýri and Hringsdalur) coastal regions, and all exhibit δ13C values that suggest a high marine diet, perhaps as much as 50%. It is therefore quite likely that the radiocarbon dates for these individuals are too early. There are also 3 burials within the collection which are dated to the post-Viking period. Two of these, Smyrlaberg (1030–1220) and Urriðaá (1180–1275), are sites that were both recorded as kuml during the initial reporting, although in both cases it is noted that the lack of grave goods (Urriðaá), and the layout of the grave (Smyrlaberg) do not appear as typical Viking age Pagan burials (Eldjárn 2000: 121–122, 127–129). It is therefore likely that these are not Viking Age kuml, although it is not certain whether they are burials from Christian cemeteries, deviant burials, or represent other burial traditions. The third burial that post–dates the Viking period is more problematic, Hafurbjarnarstaðir (1215–1285). These bones are from one of the richer Viking age kuml grave fields in Iceland, with graves with burial goods that are undoubtedly from the Viking Age (for example, spears, combs, beads, a sword). A total of 9 kuml graves were excavated at the site, and the bones sampled for this date came from a disturbed burial containing several individuals. It is possible that this disturbed burial represents a later intrusion at the site. The probability distributions for these dates are shown in Figure 2. The radiocarbon data will be considered again later in this article. Journal of the North Atlantic T.D. Price and H. Gestsdóttir 2018 Special Volume 7 150 Analyses Analytical methods have been described in detail in a previous article on Isotopic Proveniencing (Price 2018, this volume). Specific sampling procedures are outlined here. Where possible the first molar was sampled, with approximately ¼ of the tooth removed for analysis. In those instances where the first molar was unavailable, an incisor, canine, or premolar was sampled and half the tooth was used for the analysis. Teeth displaying pathological lesions or non-metric traits were avoided, as were teeth that remained fixed in the alveolar bone, even if teeth other than the first molar were sampled. More information on the teeth that were sampled appears in Supplemental Table 1 (available online at https://eaglehill.us/jonaonline/ suppl-files/jona0072-price&gestsdottirs1. pdf, and, for BioOne subscribers, at http://dx.doi.org/10.1656/ jona0072.s1). This section is organized to present the isotopic data we have collected from Iceland. Radiocarbon dates provide some fundamental information on the chronology of the samples we have taken and are very useful for examining variation in isotope ratios over time. Next the isotopic analyses are summarized, including strontium, oxygen, carbon in tooth enamel, and carbon and nitrogen in bone collagen. In conclusion, the results of the isotopic investigations of the human remains from Iceland are synthesized and put into a broader framework of the North Atlantic colonization. Comparison of this information with conditions on Greenland provides an intriguing look at differences in the adaptations of colonists in these distinct environments. Strontium Isotope Analysis Prior to a presentation of the results of the strontium isotope analysis of the Iceland material, we briefly summarize the bioavailable baseline for this area. Iceland is one of the youngest landmasses on earth, a large volcanic island created over the last 20–25 million years. The strontium isotope ratios of these new rocks are very low, and numerous studies have reported values that average approximately 0.703–0.704 (e.g., Figure 2. Radiocarbon probability distributions for dates on 37 human bones from Iceland. Journal of the North Atlantic T.D. Price and H. Gestsdóttir 2018 Special Volume 7 151 Christianity came to Iceland around 1000 AD. The 2 sets of graves are almost exclusively from separate chronological periods. A number of observations can be made from the graph in Figure 3. It is clear that a high proportion of the pagan individuals are non-local (32 of 83; 39%) with values greater than 0.7092. However, no more than a single Christian individual (3%) appears to be non-local. There is also a significant difference between pagan and Christian burials in terms of the mean 87Sr/86Sr for the 2 groups. Pagan 87Sr/86Sr values average 0.7095 ± 0.0035, while the Christian graves have a mean value of 0.7074 ± 0.0012. It seems to be the case that the migration from Scandinavia and the northern British Isles had largely stopped by the 11th century, at least in the areas served by the Skeljastaðir and Haffjarðarey cemeteries. It also needs to be reiterated that strontium isotopes cannot distinguish individuals coming from foreign areas with 87Sr/86Sr ratios similar to the local values on Iceland. Fortunately, there are few places around the eastern North Atlantic rim with values below 0.709. Denmark, western Scania in Sweden, parts of the North European Plain and the southern areas of the Britain would be the possible candidates. It is also the case in Figure 3 that the non-local pagan individuals exhibit a wide range of 87Sr/86Sr values above 0.7092, indicative of multiple places Moorbath and Walker 1965, Sun and Jahn 1975, Wood et al. 2004). As previously noted, we also have data for the local baseline for Iceland from the geology as well as modern plants and both modern and archaeological animals. We have measured 1 sample of modern barley (0.7068), modern sheep bones from 3 difference locations on the island (0.7059–0.7069; Price and Gestsdottir 2006), and archaeological cattle and pig bones from northern Iceland (average of 0.7042). Thus, we expect that the baseline for 87Sr/86Sr sources in foods for the inhabitants of Iceland varies between approximately 0.7040 and 0.7092. This range in tooth enamel should identify individuals born on Iceland. With this background in mind, the results of the 87Sr/86Sr analysis of the tooth enamel from human burials on Iceland are presented in Supplemental Table 1 (available online at https://eaglehill.us/jonaonline/ suppl-files/jona0072-price&gestsdottir-s1. pdf, and, for BioOne subscribers, at http://dx.doi. org/10.1656/jona0072.s1). The 127 values have a mean and standard deviation of 0.7087 ± 0.03 and vary from 0.7056 to 0.7257. These data are initially presented in a bar graph of rank-ordered values shown in Figure 3. The baseline data from the basalt, barley, sheep, and cows and pigs appears on the left side of the graph, and the human samples are sorted and ranked by pagan and Christian graves in the main body of the graph. As previously noted, Figure 3. A bar graph of 87Sr/86Sr values from rock, plants, animals, and both pagan and Christian humans on Iceland. The red line marks the 87Sr/86Sr of seawater and marine foods at 0.7092. Journal of the North Atlantic T.D. Price and H. Gestsdóttir 2018 Special Volume 7 152 in individual diet varied from about 10% to 55%, depending largely on the geographical position (distance from the sea) of the excavation site. A plot of isotope ratios from human bone against distance of the burial site from the coast (Fig. 5) reveals a dramatic decrease in δ13C values as distances increase. This pattern is true for both the Christian and pagan graves. Moreover the range of δ13C values is much greater along the coast than at the inland sites, documenting greater diversity in coastal diets. The same pattern can be observed in the strontium isotope data by comparing inland and coastal cemeteries from the early Christian period on Iceland. We have samples from 2 cemeteries: Skeljastaðir (33 samples) and Haffjarðarey (11 samples). Haffjarðarey is located on the west coast of Iceland and Skeljastaðir lies approximately 50 km inland in southwestern Iceland. We have categorized the samples as marine, “mostly marine,” terrestrial, “mostly terrestrial,” or foreign (the one foreigner among the Christian burials) based on the 87Sr/86Sr values. The lowest 87Sr/86Sr values are assumed to reflect terrestrial diets, and the values just below 0.7092 to be marine diets. Diets designated as mostly marine or mostly terrestrial lie between the two extremes. A bar graph of frequency of occurrence of these categories is shown in Figure 6. Boundaries for the intervals for Skeljastaðir are foreign 0.7097, marine 0.7090, mostly marine 0.7075, mostly terrestrial 0.7060, and terrestrial. As can be seen from this graph, all of the individuals sampled from the churchyard at the coastal site of Haffjarðarey are characterized by a marine diet. The burials from Skeljastaðir show a variety of diets from marine to terrestrial, but “mostly terrestrial” diets are most common by far. The mostly marine diet, observed in 5 individuals from the cemetery, may well belong to individuals of origin. It seems very unlikely that such variation in 87Sr/86Sr is present in a small region of Norway. In all likelihood the high 87Sr/86Sr values represent a number of different places in Norway and the northern British Isles and Ireland. We will return to the question of the homelands of these non-local individuals in a later section of this article. It is also the case that the 87Sr/86Sr values among the local Icelandic individuals show a very smooth and continuous line from values around 0.7055 to values around 0.7092. These values reflect the range of 87Sr/86Sr sources on Iceland, from fully terrestrial to fully marine. An important question concerns the reasons for this variation. On Greenland, for example, increasing marine resources in the diet were correlated with time as the climate deteriorated and more food had to be obtained from the sea (e.g., Arneborg et al. 1999, 2012). The distribution of the 87Sr/86Sr values in the pagan burials is shown in a histogram in Figure 4. The distribution is clearly bimodal and heavily skewed to the right. The lower border of 87Sr/86Sr bioavailable values on Iceland is clearly indicated by the absence of values below 0.7055. The low frequency of values above 0.709 records individuals who are not native to Iceland. It is also the case that there may be non-local individuals among some of the 87Sr/86Sr values below 0.7092, but these samples cannot be distinguished from local Icelanders on the basis of strontium isotope ratios alone. The 2 modes in the histogram likely distinguish diets dominated by terrestrial foods vs. diets where marine foods play an important role. On Iceland it is probable that this distinction is related to the location of settlement, that is distance from the sea, rather than change over time. A recent study by Sveinbjörnsdóttir et al. (2010) demonstrated the relationship between diet and settlement location using δ13C values in bone collagen. The range of marine protein percentage Figure 4. Histogram of 87Sr/86Sr values in human tooth enamel of pagan burials from Iceland. Figure 5. A plot of distance from shore vs. δ13C for human collagen from burials on Iceland (from Sveinbjörnsdóttir et al. 2010). Journal of the North Atlantic T.D. Price and H. Gestsdóttir 2018 Special Volume 7 153 who were born on the coast and later moved inland. There is also a significant difference between the mean 87Sr/86Sr values for these 2 cemeteries, 0.7086 ± 0.0005 at Haffjarðarey and 0.7070 ± 0.0011 at Skeljastaðir. These data point to site location as the primary determinant of diet during the Viking Period on Iceland. Change over time may also be reflected here as the coastal cemetery at Haffjarðarey is 100 years younger than Skeljastaðir. Diet on Greenland is known to have become more marine-oriented over time (Arneborg et al. 1999), and the same may be expected for Iceland. At the same time, the pronounced differences between the 2 sites suggest more for a geographic cause. There is also some information regarding the role of the sexes in migration in the 87Sr/86Sr data. In the sample of pagan graves, 64 of the skeletons could be determined as to the sex of the buried individual. Of that number, 21 were female and 43 were male. Among the non-local individuals identified by 87Sr/86Sr values greater than 0.7092, 11 were female and 12 were male. Thus, 52% of the females in our sample of pagan burials were non-local, whereas only 28% of the males were of foreign birth. There was clearly a higher incidence of mobility among the females. Finally, there is also information on the length of the period of settlement on Iceland. As noted previously, the Sagas state that the period of colonization lasted 60 years. Comparison of 87Sr/86Sr values with the 37 individuals who were radiocarbon dated is particularly informative. This information is presented in graphic form in Figure 7. The dotted line in this scatterplot marks the upper boundary of local 87Sr/86Sr values. Individuals above this line were not born on Iceland and are considered as arriving colonists lie above this line. Although the sample size is limited, a number of non-local individuals are seen throughout the pagan period. There may be a slight decrease in the number of individuals over time, but it is clear that the migration to Iceland continued throughout the pagan period. Collagen Carbon and Nitrogen As discussed in a previous article in this special volume, carbon isotopes can be measured in a number of different tissues in the human body. Carbon and nitrogen isotope ratios in bone collagen are normally included as part of the process of radiocarbon dating. Both of these ratios contain information about past human diet and the trophic level of the individual under investigation. Bone, and specifically collagen, is continuously remodeled during the life of the individual so that the ratios obtained refer to the last years of life, in contrast to the early years of life in Figure 6. A bar graph of 43 enamel samples categorized by diet based on 87Sr/86Sr values. Figure 7. Scatterplot of enamel 87Sr/86Sr vs. years AD for 37 human bone samples from Iceland. The dotted line in this plot marks the upper limit of local values on Iceland at 0.7092. Figure 8. Histogram of δ13C in collagen from 37 bone samples from Iceland. Journal of the North Atlantic T.D. Price and H. Gestsdóttir 2018 Special Volume 7 154 on 37 individuals, presumed to belong to the pagan period. Included with the resulting 14C measurements were data on 13C and 15N isotope ratios. The δ13C values for these 37 individuals averaged -19.86‰ ± 1.02, with a minimum of -17.3‰ and a maximum of 22.0‰. These data are shown in a histogram in Figure 8 and are generally normally distributed with a left skew. Only a few values higher than -19.‰ or lower than -21‰. These values point to a largely terrestrial diet among the individuals sampled for radiocarbon dating. Values for 37 δ15N ratios averaged 10.15‰ ± 2.03 with a minimum of 3.7‰ and a maximum of 14.5‰. The distribution of values is shown in a histogram in Figure 9. The histogram appears bimodal with 2 very low values to the left. There is a primary mode between 10‰ and 11‰ and a secondary mode between 12‰ and 13‰. These modes likely mark the norms for terrestrial and marine diets, respectively. A plot of δ13C vs. δ15N (Fig. 10) for these data shows a generally linear trend with a significant positive correlation (r = 0.705). Specifically, as δ13C values become more positive (more marine), nitrogen values also increase indicating a higher trophic level, i.e., a more carnivorous diet. There is also a cluster of points between δ13C -19‰ and -21‰ and between δ15N 9‰ and 13‰ where a linear relationship is less pronounced. These values likely reflect the role of the meat of domestic animals in more terrestrial diets. Sveinbjörnsdóttir et al. (2010) also report δ15N measurements of the human bone collagen spanning 6.5‰ to 15.5‰. They observe a similar trend between δ15N and δ13C, where the bone collagen with the more negative δ13C values also has the less positive δ15N values, and vice versa. the chemistry of enamel. It is also the case that the carbon in the protein collagen comes largely from the protein in consumed foods and thus contains information largely on that component of diet. Carbohydrates and lipids are not normally represented in collagen carbon. Several studies of carbon and nitrogen isotope ratios for early Icelandic burials have been reported (Ascough et al. 2012, Sveinbjörnsdóttir et al. 2010). Sveinbjörnsdóttir et al. (2010) report δ13C and δ15N from 79 human samples from pagan and early Christian graves on Iceland. The bone collagen samples show substantial variation in δ13C, from -16.4‰ to -20.3‰, although most of the data lie between -18.0‰ and -20.0‰. The mean δ13C value of human bone collagen of the 45 skeletons studied from the Christian gravesites was -19.39‰ ± 0.46. The mean δ13C value of human bone collagen of the 30 skeletons studied from pagan was -18.73‰ ± 1.05. As part of our investigations of the Icelandic burials, we obtained radiocarbon dates Figure 9. Histogram of δ15N in collagen from 37 bone samples from Iceland. Figure 10. Scatterplot of δ13C vs. δ15N for 37 samples of bone collagen from pagan burials on Iceland. Figure 11. Scatterplot of calibrated years AD vs. δ13C for 37 human bone samples from Iceland. Journal of the North Atlantic T.D. Price and H. Gestsdóttir 2018 Special Volume 7 155 It is also possible to compare the Christian and pagan burials using δ13C in tooth enamel. Mean values are actually very similar, whereas the pagan burials show much more variable values as reflected by the standard deviation. Average δ13C values for the Christian samples were -15.4‰ ± 0.61, while the pagan burials had a mean and standard deviation of -15.3‰ ± 1.10. A scatterplot of δ13C vs. 87Sr/86Sr confirms this pattern with the Christian values clustered and the pagan values rather widely spread (Fig.13). There are 11 samples of tooth and bone from the same individual where enamel and collagen carbon isotope ratios can be compared. The mean offset between these values is -4.4‰ ± 1.06, with a maximum of -2.3‰ and a minimum of -6.6‰. These values point to a terrestrial C3 source for food, It is also possible to examine changes in diet using δ13C as a proxy by plotting these values vs. calibrated calendar years (Fig. 11). This plot shows a gradually decreasing value for δ13C over time—earlier inhabitants have more positive values suggesting a more marine diet. A plot of δ15N produces a similar picture, with trophic level decreasing slightly over time. Enamel Carbon Isotopes The carbon isotope ratio, δ13C, was also measured in 116 samples of tooth enamel from pagan and early Christian graves on Iceland. The mean value for this ratio was -15.3‰ ± 0.95 with a maximum and minimum of -12.4‰ and -17.3‰, respectively. Enamel and collagen δ13C ratios have varied ranges of values with an average offset of approximately 7‰ (Tykot 2004). A histogram for the values from Iceland is shown in Figure 12. The histogram exhibits some positive kurtosis with slight right skew in the direction of more marine diets. Figure 12. Histogram of enamel δ13C for 116 samples from Iceland. Figure 13. Scatterplot of δ13C vs. 87Sr/86Sr for all Iceland samples. Red = Christian; blue = pagan. Figure 14. A diagram with a plot of δ13C collagen vs. enamel for 11 human bone samples from Iceland with predicted dietary contribution from Froehle et al. (2012). Figure 15. Histogram of enamel δ18O for 117 samples from Iceland. Journal of the North Atlantic T.D. Price and H. Gestsdóttir 2018 Special Volume 7 156 including protein. A scatterplot of δ13C collagen vs. enamel on a standardized diagram (Froehle et al. 2012) indicates the primarily terrestrial orientation of the diet of these individuals (Fig. 14). The high nitrogen values associated with these carbon isotope ratios points to a diet where meat is a major component. Strontium isotope ratios indicate that all of the individuals in this group are native Icelanders and radiocarbon confirms their date from the pagan period. Enamel Oxygen Isotopes Oxygen isotope ratios were measured on 117 samples of human enamel from pagan and early Christian graves on Iceland. The mean δ18O value was -4.86 ‰ ± 0.97 with a maximum and minimum of between -2.23‰ and -6.94‰. A histogram of these values is provided in Figure 15 and shows a generally normal distribution. These oxygen values generally fit with the expected range for Iceland. Values reported by Fricke et al. (1995), LeColle (1985), and others suggest an oxygen isotope ratio with a carbonate δ18OPDB value comparable to tooth enamel of ca. -4.4‰, which fits well with our data. These values are also found in large parts of southern and central Norway and across the northern British Isles and Ireland. It is also possible to compare averages and ranges between the pagan and early Christian samples from Iceland. The mean δ18O value for the 74 pagan samples was -4.91‰ ± 1.11, whereas the 32 Christian samples had a mean and standard deviation of -4.76‰ ± 0.67. The difference in means, while not significant, is of interest and may reflect the fact that there are a large number of samples from the inland Christian cemetery where oxygen isotope ratios may be somewhat less negative. A plot of δ18O vs. 87Sr/86Sr for Christian and pagans (not shown) closely resembles the plot of δ13C vs. 87Sr/86Sr with a cluster of Christian samples and a widespread scatter of pagan individuals. Oxygen isotopes should vary geographically as discussed previously. A plot of δ18O vs. 87Sr/86Sr for the pagan burial samples from Iceland provides some insight on this variation (Fig. 16). The dotted red line in this plot marks an 87Sr/86Sr value of 0.7092 and the upper boundary for Icelandic natives. All of the individuals to the right of this line are non-local to Iceland. For the native individuals to the left, the variation in oxygen isotope ratios is considerable and comparable with the variation seen in the non-locals. This pattern suggests that the range of δ18O on Iceland lies between roughly -4.0‰ and -7.0‰, with a few individuals more positive than -4.0‰. Summary of the Isotope Analyses These isotopic investigations of diet and mobility among the early Icelandic settlers have revealed substantial new information. Individuals born outside of Iceland could be readily identified by strontium isotope ratios above 0.7092. There was substantial migration during the Pagan period and almost none during the early Christian period on Iceland. Comparison of radiocarbon dates and strontium isotope ratios for 37 individuals indicated that while the majority of individuals in the sample came during the early years of settlement, new migrants continued to arrive throughout the Pagan period. A large proportion (39%) of the pagan burials we analyzed were not born in Iceland, implying a very high rate of migration. A much higher proportion of females (52%) were identified as migrants compared to males (28%), although there were only half as many female as male graves. a fact that has important implications. Investigations of modern and ancient DNA have indicated that females from northern Britain and Ireland are heavily represented among the colonists of Iceland. Given this it is also the case that among the nonnative Icelanders that multiple places of origin are indicated by diverse 87Sr/86Sr values, suggesting that Western Norway was not the only source of the colonists who traveled to Iceland. The strontium isotope data also revealed that the inhabitants of Iceland exhibited values varying from approximately 0.706 to 0.7092, the highest ratio of isotopes available. This continuous range of values reflects varying diets from fully terrestrial to almost fully marine among the inhabitants. While there is Figure 16. Scatterplot of δ18O vs. 87Sr/86Sr for the pagan burial samples from Iceland. The dotted red line marks an 87Sr/86Sr value of 0.7092, the boundary for local vs. nonlocal on Iceland. Journal of the North Atlantic T.D. Price and H. Gestsdóttir 2018 Special Volume 7 157 a slight tendency for diets to be more marine over time, the terrestrial component of diet appears to vary more directly with the location of settlement as coastal sites have strongly marine diets and inland localities generally lack foods from the sea. Carbon and nitrogen isotope ratios were also measured in bone collagen from 37 individuals as part of the radiocarbon dating procedure. The carbon isotope ratios averaged -20‰ and the δ15N 10‰. As noted previously, human consumers of terrestrial plants and animals typically have δ15N values in bone collagen of 6–10‰ whereas consumers of freshwater or marine fish and mammals may have δ15N values of 15–20‰ (Schoeninger and DeNiro 1984, Tykot 2004). The carbon in the Icelandic bone collagen points to terrestrial food sources and the nitrogen isotopes emphasize a predominance of animal protein in the diet. Carbon isotope ratios in enamel apatite reiterate the importance of terrestrial foods. Both the offset in the ratios between collagen and apatite carbon values the plot of these values on a standardized diagram (Froehle et al. 2012) emphasizes the terrestrial diet of these individuals. The high nitrogen values associated with these carbon isotope ratios points to a diet where meat is a major component. The Christian and pagan burials have very similar mean δ13C enamel values, but the pagan burials exhibit much more variation. Oxygen isotopes are variable, and values generally fall in a range that encompasses those found in Iceland, southern Norway, and northern Britain and Ireland. The variation in these values that is observed likely reflects both the variety of places that these individuals came from and the natural variation in oxygen isotope ratios in human tooth enamel. Comparison of the Icelandic isotopic information with the data from Norway, Greenland, and elsewhere can be found in the concluding chapter of this volume. Beyond the details of life and death provided by the isotopic data, there are larger implications for the archaeology of Iceland and the North Atlantic in terms of the characterization of burials and the identification of a large segment of the population, which does not appear among the burials that have been excavated (Vésteinsson and Gestsdóttir 2016, this volume). Perhaps an even greater lesson is the power of archaeological sciences to elucidate new information from the past. Implications of the Isotope Analyses Based on this new information it is possible to examine some of the patterning of settlement on Iceland and how the results reflect on the nature of burial archaeology. The results for the Christian cemeteries have been discussed elsewhere (Price and Gestsdóttir 2006:139–142), so the focus of the discussion here will be on the results of the analysis of the Viking Age remains. The Book of the Settlements chronicles the settlement of Iceland, the earliest versions of which date to the 13th century. It records that Iceland was settled in ca. AD 874 and states that the entire country was settled simultaneously and that the process was rather swift, lasting only 60 years (Karlsson 2000:12). Archaeological evidence supports this late 9th-century date of the first settlement, with little evidence of human activity in Iceland preceding the settlement layer (landnámslag), a volcanic tephra layer dated to AD 872 ± 2 (Grönvold et al. 1995, Vésteinsson 1998:2–4). Various estimates of the size of the population of Iceland for the settlement period have been put forward. The number of immigrants into the country has been estimated at approximately 20,000 and that the population at the end of the settlement period (930) was approximately 30,000–35,000 (Eldjárn 2000:256). Earliest documented population figures in Iceland are the first national census carried out in 1703. At the time the population was approximately 50,000 people, and remained so until the early 19th century (Garðarsdóttir 2002:36). There have, however, been suggestions that the population during the medieval period may have reached as high as 70,000 (Vasey 1996:367). Two plague epidemics during the 15th century partially explain the decrease. The first was in 1402, which resulted in the death of approximately 50–60% of the population and the second in 1494, during which an estimated 30–50% of the population died (Karlsson 1996:265). It is of course impossible to know what the population growth rate would have been in 10th-century Iceland. However, it has been documented that short-term population growth in small populations is often very high, reaching as much as 0.5–2% per year (Chamberlain 2009:281). A population growth rate of 0.5% would mean that that by the year 1000 the population of Iceland would have been approximately 60,000 people. As with data on population size, the earliest data on crude death rates in Iceland are from the 18th century. In that period and towards the last decade of the 19th century, the crude death rate in Iceland was 20–40 per 1000 population (Garðarsdóttir 2002:37). Based on such figures, it can be estimated that 95,000–100,000 people died during the first 130 years of the settlement of Iceland. If these figures of Journal of the North Atlantic T.D. Price and H. Gestsdóttir 2018 Special Volume 7 158 from Viking Age burials in Iceland. These burials, the kuml, rather than being representative of the burial tradition of the Viking Age period in Iceland as a whole, are most likely rather representative of a subgroup of that population, they are mostly the adult males, and probably the adult males of high standing. People of lower standing—slaves, indebted free men, and even children—were not provided with the kuml graves. In preferably selecting to sample the skeletons from burials which can be dated to the Viking Age with certainty, that is those burials with typologically datable grave goods, it is very likely that this bias has been emphasized even further—a biased sampling process of a biased sample. The bias within the Icelandic Viking burial record is reiterated when one looks at the ratio of immigrant women versus immigrant men. Of the 83 skeletons in the sample 65 could be sexed (in general the preservation of skeletal material from Icelandic kuml is quite poor and does therefore not always allow for sexing of the remains). Of these 44 were male and 21 were female reflecting the 2:1 ratio of men to women in the burials. However, 11 of the females were immigrants (52% of the total females sampled), as opposed to 12 (27%) of the males. So although males represent a larger proportion of those buried in Viking age burials, a higher percentage of the females were immigrants. This could also be a reflection of the nature of immigration as discussed above. That is, that a large proportion of the women living in Iceland during the first 130 years of settlement, are represented by a trickle of female immigrants over a long period of time being brought in to balance the ratio of men to women. However, if we look at the sex ratio in Viking Age burials in Norway, it appears to be even more imbalanced than within the Viking age burials in Iceland, closer to a ratio of 3:1 males versus females. In Norway this is explained as being due to a status difference, that is, that a majority of the Viking age burials excavated in Norway are high-status burials rather than representative of the population as a whole (Dommasnes 1992:5). It must be noted here that in discussing the burials in Kaupang, where 10th-century female burials only make up for 24% of the total. Stylegar (2007:82– 83) points out that it is important to differentiate between gendered burials (based on grave goods) and sexed burials (based on osteological diagnosis of sex). It is not possible to ignore the fact that the imbalance between the 2 sexes may be created to some extent by the fact that gendering male burials based on artefacts (for example, weapons) may be easier than female burials because the latter may be the number of settlers, the population growth rate, and crude death rate are true, one would expect the immigrants to make up for approximately 20% of the population that lived in Iceland during the first 130 years of the settlement, a much lower figure than the very minimum of 39% immigrants demonstrated by the results of the strontium analysis. These numbers do not take into account the effect that return migration would have had on the demography. However, as return migrants would most likely have been the immigrants themselves who in turn would be replaced by new migrants, (Gmelch 1980:136) these figures are unlikely to have had an effect on the proportion of immigrants within the skeletal population. This discrepancy needs explaining. One factor that needs to be considered is the sample selection process. As already stated the burials that could be securely dated to the Viking period were preferably selected, and a large proportion of these were from graves that contained grave goods. It must therefore be questioned whether in selecting the individuals from burials with datable grave artifacts a biased sample towards the immigrants was created? To answer this question, it is important to consider the demography of the skeletal population from the Icelandic kuml as a whole. The samples taken for this project were all from adult individuals, and it is of course very likely that a large proportion of the deaths in settlement period Iceland would have been due to high infant mortality during the settlement period. Infant mortality numbers are not known for this period; however, data from the early 19th century in Iceland suggests that at that time infant mortality was 300 per 1000 live births (Garðarsdóttir 2002:20). It is therefore striking that only 2 infants (under 2 years of age) have been found in Viking age burials making up only 1% of the skeletal population (Gestsdóttir 1998:10–17). Similarly, there is an imbalance between the sexes. Of the 109 individuals for whom sex could be determined, only 35 (32%) were female (Gestsdóttir 1998:5). It is expected within a migrant population to see a higher percentage of males than females; for example, in transatlantic migration during the 19th-century women were only 40% of the population (Harzig et al. 2009:119). However, it is unlikely that such an uneven ratio would be seen in a sample representing a 130-year period of a population. It would be expected that by the second or third generation, the ratio would have evened out to something closer to 1:1. These discrepancies perhaps indicate that the bias seen in the results of the strontium isotope analysis is merely the bias within the existing skeletal collection Journal of the North Atlantic T.D. Price and H. Gestsdóttir 2018 Special Volume 7 159 this volume, Sveinbjörnsdóttir et al. 2010:686); Daðastaðir, where a late 10th-century trefoil broach in the Borre style was found (Eldjárn 2000:363– 365); and Sílastaðir which included 10th-century beads (Hreiðarsdóttir 2005:165). The fact that immigrants are found at all the sites that can be dated to the latter half of the 10th century would suggest that a significant percentage of immigrants were arriving in Iceland well into the 10th century. These results indicate that rather than the settlement process involving an influx of people for a relatively short period during the late 9th and early 10th century, the settlement “period” seems to have lasted much longer, into at least the latter part of the 10th century. The fact that only one immigrant was found in the sample of 44 (2%) taken from early Christian cemeteries, Skeljastaðir and Haffjarðarey (Price and Gestsdóttir 2006:140), would indicate that the period of intensive settlement had ended by the time they came into use, at least in the areas which these cemeteries served. Further supporting this are the results of the strontium analysis of some of the larger kuml burial grounds, in particular Dalvík and Sílastaðir, in the north of Iceland. These burial grounds have been considered as farm-based plots, that is, that the individuals buried there all belong to the same farm (Eldjárn 1974:133), although more recent work has opened up the possibility of there being more than one kuml cemetery on each farm, or the possibility of a single plot representing burials from more than one farm (Friðriksson 2009:10–11). The results from the 2 above-mentioned sites can perhaps shed some light on what would have been the nature of the occupation of these sites. From Dalvík there are 8 samples (5 men and 3 women). Of these 6 individuals (3 of the men and all of the women) were immigrants into Iceland. However, if one looks at the results of the strontium analysis there is a great variation in the signature, from 0.7099 to 0.7257. The same pattern is seen at Sílastaðir. Four individuals (3 men and 1 woman) were buried at the site, and all were found to be immigrants. The variation in the strontium ratio signature was from 0.7093 to 0.7182., which indicates that the immigrants buried at these sites originated from different areas. In other words, within these farm-based plots we are not dealing with people who originated from the same place, settled together on the same farm and were buried with their family or descendents. One issue is that these burials are from early excavations, which means that the stratigraphy of the burials is not known, so it is not possible to tell how contemporary the graves in the group are. more likely to not contain gender-specific artifacts. What we do know is that the demography of the Norwegian Viking Age graves would not have been affected by immigration patterns as the Icelandic graves would have, and so it is questionable as to whether the high ratio of men to women in the Icelandic kuml can be explained focusing only on the effects of immigration on the demography of the first years of the settlement of Iceland. Perhaps a more likely explanation is simply that immigrant women of a high enough standing were given burial in kuml, much more commonly than men were. One of the main problems with dealing with these issues is associated with the dating of the burials. Based on documented sources, the practice of burying in kuml dates from the first settlement until the Christianization of Iceland in AD 1000 (Karlsson 2000:33–36). Only a handful of kuml have been dated through tephrochronology, and all of them fit into this period. Similarly, those graves that have datable artifacts mostly date to the 10th century (Eldjárn 2000:473–475). Radiocarbon dates do exist for kuml burials, some of which extend into the 11th century, whereas others predate the late 9th century time of settlement by as much as 200 years (Ascough et al. 2007:948, Price at al. 2018, Sveinbjörnsdóttir et al. 2010:686–688). However, recent work on radiocarbon dating from kuml and middens in northern Iceland indicates that the effect of both the marine reservoir (which has been affected by the East Greenland current) and geothermally derived CO2 in the groundwater are distorting the radiocarbon dates and ageing those affected as much too old (Ascough et al. 2007, 2010). Recent work in trying to solve the issue of the freshwater reservoir effect suggests that this is an issue that will not be resolved (Ascough et al. 2014, this volume). On the other hand, due to the short period during which people were being buried in kuml (approximately 130 years) coupled by the plateau in the calibration curve between ca.780 and ca. 980 (Sveinbjörnsdóttir 2010:23), it is unlikely that even with secure radiocarbon dating a good chronology could be achieved. However, what does become clear when surveying those burials sampled for this project is that the immigrants are just as likely to be found in the burials that can be dated to the latter part of the period. Three sites, all from northern Iceland can be dated to the latter part of the 10th century, and immigrants are found in all of these sites. These are Dalvík, where 2 independent radiocarbon dates exist for one of the immigrant skeletons which indicate a late 10th-century date for the site (Price et al. 2015, Journal of the North Atlantic T.D. Price and H. Gestsdóttir 2018 Special Volume 7 160 the Christianization of Iceland in AD 1000 burial in Christian cemeteries took over. Conclusions The questions raised by the results of this project are several. What should be considered an Icelandic Viking Age burial? Because of the problem with dating the burials, in many instances only those burials that include grave goods are considered to be Viking Age graves. For example, Kuml og haugfé (Eldjárn 2000), a catalogue of all Viking Age burials in Iceland, contains several finds of isolated graves which are topographically akin to kuml and are clearly not associated with a traditional Christian cemetery. Because of the lack of grave goods, these burials are not included in the main catalogue, but are reported in a smaller font within the book as a sort of anomalous group. Thus, the current way of dealing with burials in Iceland is to try and fit all finds into either the kuml or the Christian cemetery category, and anything that does not fit neatly into either group is considered as a sort of anomaly of one or the other. What the demography of the kuml burials coupled with the results of the strontium analysis indicates is that perhaps it is time to abandon such simplistic views, that Icelandic burial archaeology cannot be simply split into 2 groups. The typical Viking age kuml burials, the burials with grave goods and frequent episodes of disturbance in antiquity, are representative of a subgroup of the population, consisting mainly of males of high standing and immigrant women, rather than being representative of the population of Iceland as a whole. So, at this stage it becomes a question of semantics, what is it that we mean when we talk about kuml burials within Icelandic archaeology: is it a Viking burial, or is it a specific type of a Viking burial? In other words, within the Icelandic burial record all kuml are Viking Age burials, but all Viking Age burials are not necessarily kuml. Such ideas are not new to the burial archaeology of the Norse. In his paper, Dying and the dead: Viking Age mortuary behavior, Neil Price wrote, “After more than a century of excavations there can remain no doubt whatever that we cannot speak of a standard orthodoxy of burial practice common to the whole Norse world” (Price 2008:257). He goes on to discuss local variation in Norse burial practice in Scandinavia at the community, village, or even farmstead level (Price 2008:257–258). To assume that these local variations disappear once Iceland is settled would seem a drastic oversimplification. The results of the strontium isotope analysis of the Viking Age material can also inform the pattern of settlement at regional levels. By the 10th-century Iceland had been divided into 4 quarters: north–Norðlendingafjórðungur (33 individuals were sampled for the strontium project); south – Sunnlendingafjórðungur (14 individuals sampled); east—Austfirðingafjórðungur (29 individuals sampled) and west–Vestfirðingafjórðungur (7 individuals sampled) for political and administrative purposes (cf., Karlsson 2000:22– 25). The number of samples from each quarter is varied and in no instance very high, and so any conclusion reached as to their meaning have to be made cautiously. However, there is one factor that is quite noteworthy if one looks at these results; the findings from Norðlendingafjórðungur, where 55% of the individuals sampled are clearly not born in Iceland, a much higher percentage than the 39% of the country as a whole. The significance of this must be questioned. It has long been noted that a large proportion of Viking Age burials have been found in the north of Iceland. A popular explanation for this is that much more intensive roadwork, associated with a denser modern population, means that more kuml are found in these areas, and that those in the west for example where very few kuml burials are known, simply have not been found yet (Eldjárn 1974:133). However, the high percentages of kuml in the north could be explained by the fact that there simply are more kuml burials located there. That is, that the type of burial we most easily recognize as kuml, are most common in the north of Iceland. Others have put similar ideas forward, for example, Sigurðsson (2000:28) mentions in a footnote that there is “something dubious” about the lack of kuml burials in the Westfjords of Iceland. As discussed earlier, the Viking age burial in Iceland, kuml, is a well-identified archaeological feature, or so we would like to believe. A recent M.A. thesis points out the discrepancy in recent works as to how to classify a kuml, resulting in very varied figures as to the number of known Viking kuml in Iceland (Sigurðarson 2009:15). At the same time, very little focus is put on trying to explain those burials that are not included in the count of kuml burials each time. However, as has been demonstrated here, whatever method is used to count the kuml, the demography of these burials indicates that they are not representative of the entire population of Iceland during the 9th–10th century, there are only a handful of children and the ratio of men versus women is imbalanced. There remains therefore the issue of the nature of burials in the settlement period in Iceland. The traditional, perhaps simplistic view is that the first settlers buried their dead in these kuml, but after Journal of the North Atlantic T.D. Price and H. Gestsdóttir 2018 Special Volume 7 161 This of course raises the question, where are the rest? The short answer to that is we don’t know. One issue is the dating of the earliest Christian cemeteries. It is recorded in the Book of the Settlements that Christians were among the first settlers (Karlsson 2000:16), and, undoubtedly, they would have buried their dead according to Christian customs. Recently published radiocarbon dates from the early Christian cemeteries at Skeljastaðir and Keldudalur cannot exclude an early 9th-century date for some of the burials there (Sveinbjörnsdóttir et al. 2010:687–688), although it must be noted that these dates do not take into account the possibility of a freshwater reservoir affect as pointed out by Ascough et al. (2010). On the other hand, as has already been mentioned, conclusive dating of the start of use of many medieval cemeteries is often lacking. For example, the previously published dates of the start of use of the cemetery at Skeljastaðir in Þjórsárdalur to ca. 1000 (Þórðarson 1943:134) seems to be largely based on the documented date of the Christianization of Iceland, so it cannot be excluded that some of the Christian cemeteries were in use during the settlement period. Another issue which possibly has even a greater effect is the fact that most of the Viking Age burials found to date belong to the typical kuml group as already described plainly because we recognize them. We know where they are located — just outside home-field boundaries or on the junction of farm boundaries and roads. We know what they look like; low mounds sometimes with low stone-lined borders. Even more frequently, as has been demonstrated in the past few years by a kuml project in the north of Iceland, the kuml are found due to the fact that when the graves were “robbed” in antiquity they were left open and so they are easily recognizable in the landscape because of shallow depressions in the ground (Friðriksson 2009:10–16). We know what to look for and where to find the kuml, and so we go out and we find them. The same can be said for the early Christian cemeteries. They are located near farm mounds, bordered by a low circular boundary surrounding a central structure, so we go out and find them. It has been theorized that these early Christian cemeteries may have been modeled on the Pagan burial grounds (Vésteinsson and Gestsdóttir 2016, this volume). However, if these early Pagan burial grounds are not as easily recognized in the landscape, it could be that we simply are not finding them, or even not recognizing them when we do find them. Therefore, we place the graves found within them in an “anomalous” group. From this perspective, any burial that does not fit into a previously defined burial group gets dismissed as an anomaly and is often forgotten as a footnote in an archive somewhere. So, this is where the search needs to start looking for the rest. Within the Icelandic archaeological archive there are countless records of burial finds, whether the records are textual in nature or skeletal finds. To begin to answer the question of where the rest of the Viking population of Iceland is buried, we need to go back to this data, and interpret it on its own merit rather than trying to fit it into our preconceived ideas as to what an Icelandic burial should look like. Acknowledgments Hildur would like to thank the Icelandic Centre for Research (RANNIS) for funding parts of this study. We would also like to thank Phillipa Ascough for her help with the radiocarbon dates. Literature Cited Ambrose, S.H., and M.J. DeNiro. 1999. Climate and habitat reconstruction using stable carbon and nitrogen isotope ratios of collagen in prehistoric herbivore teeth from Kenya. Quaternary Research 31:1–16. Arneborg, J., J. Heinemeier, N. Lynnerup, H.L. Nielsen, N. Rude, and A.E. Sveinbjornsdotir. 1999. Change of diet of the Greenland Vikings determined from stable carbon isotope analysis and 14C dating of their bones. Radiocarbon 41:157–168. Arneborg, J., N. Lynnerup, J. Heinemeier, J. Møhl, and A.E. Sveinbjörnsdóttir. 2012. Norse Greenland dietary economy ca. A.D. 980–ca. A.D. 1450. Pp. 1–39, In J. Arneborg, J. Heinemeier, and N. Lynnerup. Greenland Isotope Project: Diet in Norse Greenland AD 1000–AD 1450. Journal of the North Atlantic. Special Volume 3. Ascough, P.L., M.J. Church, G.T. Cook, Á. Einarsson, T.H. McGovern, A.J. Dugmore, and K.J. Edwards. 2014. Stable isotopic (δ13C and δ15N) characterization of key faunal resources from Norse period settlements in North Iceland. Pp. 25–42, In D. Price, Viking Settlers of the North Atlantic: An Isotopic Approach. Special Volume 7, Journal of the North Atlantic. Ascough, P., G. Cook, M. Church, A. Dugmore, T.H. McGovern, E. Dunbar, Á. Einarsson, A. Friðriksson and H. Gestsdóttir. 2007. Reservoirs and radiocarbon: 14C dating problems in Mývatnssveit Northern Iceland. Radiocarbon 49:947–961. Ascough, P., G. Cook, M. Church, E. Dunbar, Á. Einarsson, T.H. McGovern, A. Dugmore, S. Perdiakris, H. Hastie, A. Friðriksson and H. Gestsdóttir. 2010. Temporal and spacial variations in 14C reservoir effects: Lake Mývatn, northern Iceland. Radiocarbon 52:1098–1112. Ascough, P.L., M.J. Church, G.T. Cook, E. Dunbar, H. Gestsdóttir, T.H. McGovern, A.J. Dugmore, A. Friðriksson, and K.J. Edwards. 2012. Radiocarbon Journal of the North Atlantic T.D. Price and H. Gestsdóttir 2018 Special Volume 7 162 reservoir effects in human bone collagen from northern Iceland. Journal of Archaeological Science 39:2261–2271. Chamberlain, A.T. 2009. Archaeological Demography. Human Evolution 81:275–286. DeNiro, M.J. 1985. Postmortem preservation and alteration of in vivo bone collagen isotope ratios in relation to palaeodietary reconstruction. Nature 317:806–809. Dommasnes, L.H. 1992. Two decades of women in prehistory and in archaeology in Norway. A review. Norwegian Archaeological Review 25:1–14. Eldjárn, K. 1974. Fornþjóð og minjar. Pp. 101–152, In S. Líndal (Ed.), Saga Íslands I. Hið íslenzka bókmenntafélag. Sögufélagið. Reykjavík, Iceland. Eldjárn, Kristján. 2000. Kuml og haugfé úr heiðnum sið á Íslandi. 2nd edition. Mál og menning, Reykjavík, Iceland. Fricke, H.C., J.R. O’Neil, and N. Lynnerup. 1995. Oxygen isotope composition of human tooth enamel from medieval Greenland: Linking climate and society. Geology 23:869–872. Friðriksson, A. 2009. Social and symbolic landscapes in late Iron Age Iceland. Archaeologia Islandica 7:9–21. Froehle, A.W., C.M. Kellner, and M.J. Schoeninger. 2012. Multivariate carbon and nitrogen stable isotope model for the reconstruction of prehistoric human diet. American Journal of Physical Anthropology 147:352–369. Garðarsdóttir, Ó. 2002. Saving the Child. Regional, Cultural and Social Aspects of the Infant Mortality Decline in Iceland, 1770–1920. Umeå University Press, Umeå, Sweden. Gestsdóttir, H. 1998. Kyn-og lífaldursgreiningar á beinum úr íslenskum kumlum. Fornleifastofnun Ísland, Reykjavík, Iceland. Gestsdóttir, H. 2004. Mannabein í þúsund ár. Vitnisburður um lífskjör og lifnaðarhætti. Pp. 78–85. In Á. Björnsson and H. Róbertsdóttir (Eds.), Hlutavelta tímans. Menningararfur í Þjóðminjasafni. National Museum of Iceland, Reykjavík, Iceland. Gmelch, G. 1980. Return migration. Annual Review of Anthropology 9:135–159. Grönvold, K., N. Óskarsson, S.J. Johnsen, H. Clausen, C.U. Hammer, G. Bond, and E. Bard. 1995. Ash layers from Iceland in the Greenland GRIP ice core correlated with oceanic and ash sediments. Earth and Planetary Science Letters 135:149–55. Hallgrímsdóttir, M. 1991. Klaustur, spítali og kirkjustaður. Fornleifarannsókn í Viðey 1987–1989. Landnám Ingólfs 4:109–133. Harzig, C., D. Hoerder, and D. Gabaccia. 2009. What is Migration History? Polity Press, Cambridge, UK. Helgason, A., S. Sigurdardottir, J. Nicholson, B. Sykes, E.W. Hill, D.G Bradle, V. Bosnes, J.R. Gulcher, R. Ward, and K. Stefánsson. 2000. Estimating Scandinavian and Gaelic ancestry in male settlers in Iceland. American Journal of Human Genetics 67:697–717. Helgason, A., E. Hickey, S. Goodacre, V. Bosnes, K. Stefánsson, R. Ward, and B. Sykes. 2001. mtDNA and the islands of the North Atlantic: Estimating the proportions of Norse and Gaelic ancestry. American Journal of Human Genetics 68:723–737. Helgason, A., C. Lalueza-Fox, S. Ghosh, S. Sigurðardóttir, et al. 2009. Sequences from first settlers reveal rapid evolution in Icelandic mtDNA pool. PLoS Genetics 5:e1000343. doi:10.1371/journal.pgen.1000343. Hreiðarsdóttir, E.Ó. 2005. Íslenskar perlur frá víkingaöld—með viðauka um perlur frá síðari öldum. M.A. Dissertation. University of Iceland, Reykjavík, Iceland. Jónsson, Brynjólfur. 1885. Um Þjórsárdal. Árbók Hins íslenzka fornleifafélags 1884–1885:38–60. Karlsson, G. 2000. Iceland’s 1100 years. Mál og menning, Reykjavík, Iceland. Lárusson, Ó. 1929. Úr byggðarsögu Íslands. Vaka 3:319–369. Lecolle, P. 1985. The oxygen isotope composition of land snail shells as a climatic indicator: Applications to hydrogeology and paleoclimatology. Chemical Geology 58:157–181. Moorbath, S., and G.P.L. Walker. 1965. Strontium isotope investigation of igneous rocks from Iceland. Nature 207:837–840. Pétursdóttir, Þ. 2009. Icelandic Viking Age graves: Lack of material—lack of interpretation? Archaeologia Islandica 7:22–40. Price, N. 2008. Dying and the dead. Pp. 257–273, In S. Brink and N. Price (Eds.). Viking Age Mortuary Behaviour. Routledge, London, UK. Price, T.D., K. Margarita Frei, and E. Naumann. 2015. Isotopic Baselines in the North Atlantic Region, Pp. 103–136, In D. Price, Viking Settlers of the North Atlantic: An Isotopic Approach. Journal of the North Atlantic Special Volume 7:103–136. Price, T.D., and H. Gestsdóttir. 2006. The first settlers of Iceland: An isotopic approach to colonisation. Antiquity 80:130–144. Sigurðsson, G.S. 2000. Gaelic Influence in Iceland: Historical and Literary Contacts: A Survey of Reseach. University of Iceland Press, Reykjavík, Iceland. Sigurðarson, G.S. 2009. Hvar fornmenn hvíla: Staðfræði kumla og kerfisbundin leit þeirra. M.A. Dissertation. University of Iceland, Reykjavík, Iceland. Steffensen, J. 1945. Rannsóknir á kirkjugardinum í Haffjardarey sumarid 1945. Skírnir CXX:144–162. Steffensen, J. 1946. Rannsóknir á kirkjugarðinum í Haffjarðarey sumarið 1945. Skírnir 70:144-162. Stylegar, F.A. 2007. The Kaupang cemeteries revisited. Pp. 65–128, In D. Skre (Ed.), Kaupang in Skiringssal. Aarhus University Press, Aarhus, Denmark. Sun, S.-S., and Bor–ming Jahn. 1975. Lead and strontium isotopes in post-glacial basalts from Iceland. Nature 255:527–528. 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 in Iceland. Radiocarbon 52(2–3):682–696. Journal of the North Atlantic T.D. Price and H. Gestsdóttir 2018 Special Volume 7 163 Tykot, R.H. 2004. Stable isotopes and diet: You are what you eat. Pp. 204–220, In Physics Methods in Archaeometry, Martini, M. Milazzo and M. Piacentini (Eds.). Società Italiana di Fisica, Bologna, Italy. Vasey, D. 1996. Population regulation, ecology, and political economy in preindustrial Iceland. American Ethnologist 23:366–392. Vésteinsson, O. 1998. Patterns of Settlement in Iceland. A Study in Prehistory. Saga-Book of the Viking Society 25:1–29. Vésteinsson, O. 2000. The Christianization of Iceland. Priests, Power and Social Change 1000–1300. Oxford, Oxford University Press, Oxford, UK. Vésteisson, O., and H. Gestsdóttir. 2011. Kirkjur og kirkjugarðar. Pp. 72–89, In Mannvist. Sýnibók íslenskra fornleifa, Birna Lárusdóttir (Ed.). Opna, Reykjavík, Iceland. Vésteisson, O. and H. Gestsdóttir. 2016. The colonization of Iceland in light of isotope analyses, Pp. 137–145, In D. Price, Viking Settlers of the North Atlantic: An Isotopic Approach. Special Volume 7, Journal of the North Atlantic. Vésteinsson, O., and T.H. McGovern. 2012. The peopling of Iceland. Norwegian Archaeological Review 45:206–218. Wood, D.A., J.-L. Joron, M. Treuil, M. Norry, and J. Tarney. 2004. Elemental and Sr isotope variations in basic lavas from Iceland and the surrounding ocean floor. Contributions to Mineralogy and Petrology 70:319–339. Zielinski, G.A., M.S. Germani, G. Larsen, M.G.L. Baille, S. Whitlow, M.S. Twickler, and K.C. Taylor. 1997. Volcanic aerosol records and tephrochronology of the Summit, Greenland, ice cores. Journal of Geophysical Research 102:26625–26640. Zoëga, G. 2008. Keldudajur í Hegranesi. Skagfirðingabók 31:57–77. Þórarinsson, S. 1968. Beinagrindur og bókarspennsli. Árbók Hins íslenzka fornleifafélags 1967:50–58. Þórðarson, M. 1943. Skeljastaðir, Þjórsárdalur. Pp. 121–136, In M. Stenberger (Ed.). Forntida gårdar i Island: Meddelanden från den nordiska arkeologiska undersökningen i Island sommaren 1939. Ejnar Munksgaard, Copenhagen, Denmark.