At the world's edge: Reconstructing diet and geographic origins in medieval Iceland using isotope and trace element analyses

Abstract Objectives A multi‐isotope study was conducted on individuals buried at Skriðuklaustur monastery (AD 1493–1554) to investigate their geographic origins and dietary composition. Comparative material from individuals excavated from Skeljastaðir, an inland farm site was also analyzed. Materials and methods Bone collagen was extracted from 50 humans (Skriðuklaustur and Skeljastaðir) and 25 animals (Skriðuklaustur) and analyzed for δ13C, δ15N, and δ34S. Dental enamel samples from 31 individuals (Skriðuklaustur) were also analyzed for 87Sr/86Sr, δ18O, δ13C, and trace elements (Pb, Sr, Zn, Ba). Results The mean value determined from individuals from Skriðuklaustur (n = 36) was δ13C = −18.7 ± 0.8‰, δ15N = 12.8 ± 1.1‰, and δ34S = 9.0 ± 1.6‰, whereas at Skeljastaðir (n = 14), it was δ13C = −20.5 ± 0.8‰, δ15N = 7.8 ± 0.9‰, and δ34S = 9.4 ± 1.6‰. At Skriðuklaustur, human dental enamel samples (n = 31) provided a 87Sr/86Sr range of 0.7060–0.7088, δ18Ophosphate from 13.9 to 16.1‰ and δ13Ccarbonate from −16.6 to −12.9‰. Inferred drinking water (δ18Odw) values range from −12.3 to −8.9‰. Sr concentrations range from 25.8 to 156.7 ppm, Ba from 0.11 to 0.81 ppm, Zn from 43.8 to 145.8 ppm, and Pb from 0.13 to 9.40 ppm. Discussion A combination of results indicates that the people from Skriðuklaustur were born in Iceland, but some lived inland during childhood while others lived closer to the coast. Since Skriðuklaustur was a hospital, these individuals may have sought medical treatment at the monastery. The δ13C and δ15N values determined from bone collagen indicate that the people residing at Skriðuklaustur consumed a diet high in marine protein, while those residing at Skeljastaðir exhibit values more consistent with terrestrial resources.


| INTRODUCTION
Recent excavations at the site of Skriðuklaustur (AD 1493-1554), located in an inland valley in eastern Iceland, demonstrated that this monastery functioned not only as a place of scholarly work and monastic activities, but also a place where the community could seek medical care and treatment (Figure 1). Although oral history suggests the presence of post-Reformation burials marked by gravestones at Skriðuklaustur, these graves have never been located. Radiocarbon dating and other archaeological dating methods demonstrated that the site was in use during the monastic period, specifically between AD 1493 and 1554 (Kristjánsdóttir, 2012, pp. 153-154). Historical and archaeological evidence has revealed specialized medical knowledge, surgical tools, diverse medicinal plants and herbs, and imported objects and food, indicative of extensive involvement with foreign trade and monastic networks. In 2002-2012, the skeletal remains of around 300 individuals were excavated from the site, and these presented with a vast array of pathological conditions, including infectious diseases, traumatic injuries, and congenital anomalies (Kristjánsdóttir, 2008;Kristjánsdóttir, 2011;Kristjánsdóttir, 2012;Kristjánsdóttir & Collins, 2011). Comparative material was sought from a subset of skeletons excavated from a cemetery belonging to a single, inland farm site at Skeljastaðir in the Þjórsárdalur valley in southern Iceland ( Figure 1). This site is around 60 km from the coast; however, there are numerous lakes and rivers with freshwater fish in the region. The farm was occupied earlier than Skriðuklaustur, originally believed to be from around~AD 1000 until at least AD 1104 when the enormous eruption of Hekla occurred, likely leading to the abandonment of the site (Gestsdóttir, 2014). Radiocarbon dating of a subset (n = 7) of the human skeletal remains suggested a date range between AD 890 and 1220 (Sveinbjörnsdóttir et al., 2010). However, some human activity and animal grazing persisted in the valley at least until another eruption occurred in AD 1300 (Dugmore et al., 2007). In 1939, 63 skeletons were excavated from the early Christian cemetery at Skeljastaðir, although only 56 skeletons are available for study today (Gestsdóttir, 2014;Steffensen, 1943). This site was selected due to the extensive research conducted on it (e.g., archaeology, palaeoclimatology, multiple stable isotope analyses, palaeopathology), its well-preserved material, sample size, and geographic location. Both are inland sites without direct access to coastal resources; Skeljastaðir is located in the highlands, a mostly uninhabitable volcanic desert, while Skriðuklaustur is located in a mountainous area.
The geographic origin of the individuals buried at Skriðuklaustur is unknown; these individuals may have traveled to the monastery during pilgrimages, for trade or to seek treatment from elsewhere in the country or from abroad. The aim of this research was to undertake a multi-isotope study of human and animal remains from the site to establish the origins of the individuals buried there: did the monastery serve predominantly the local community, or had the inhabitants traveled from further afield? Considering the monastic context, this research also sought to evaluate dietary composition and the possible dietary differences between men, women, and children, or between religious and secular individuals buried at the monastery. Stable isotope analyses of δ 13 C, δ 15 N, and δ 34 S were obtained from samples of human and animal bone collagen and of δ 13 C , δ 18 O, and 87 Sr/ 86 Sr from human dental enamel samples. Trace element (Zn, Ba, Sr, and Pb) concentrations were also measured in the dental enamel samples to establish baselines for Icelandic humans and to attempt to further examine geographic provenance. Regarding Skeljastaðir, the previous isotopic study of δ 13 C and δ 15 N in bone collagen samples (n = 13) demonstrated a diet predominately derived from terrestrial resources (see Sveinbjörnsdóttir et al., 2010). Since δ 34 S values have been used in mobility and palaeodietary studies to differentiate between freshwater, marine, and terrestrial resources, we aimed to refine these F I G U R E 1 A map of Iceland depicting the locations of Skriðuklaustur and Skeljastaðir. The major routes of trade and travel to and from Skriðuklaustur are also shown ( © Kristjánsdóttir and Gunnarsdóttir) interpretations by evaluating possible freshwater fish consumption at Skeljastaðir. The other previous analysis conducted on the Skeljastaðir assemblage identified only one non-Icelandic migrant according to strontium isotope analysis (see Gestsdóttir & Price, 2003). The current study investigates the utility of δ 34 S analyses to improve interpretations of local versus nonlocal origins within Iceland given the potential relative homogeneity of strontium isotopes deriving from the predominantly igneous basalt bedrock. Finally, we wished to compare the diet between individuals buried in these two very different settlement types (monastic and farm). Isotopic analyses have not been conducted on humans or fauna from Skriðuklaustur prior to this research and this is the first in-depth multi-isotope study of its kind undertaken on a skeletal collection dated to the Late Medieval Period from Iceland. This research thereby provides unique data that considers the link between diet and geographic provenance from an isotopic perspective. It provides previously unavailable Icelandic human and animal isotope baselines and comparative data from the Late Medieval Period that will be useful for future research concerning Iceland and other locales in the North Atlantic.

| CURRENT EVIDENCE FOR DIET AND MOBILITY IN HISTORICAL ICELAND
Iceland is located just below the Arctic Circle. Despite the challenging, subpolar climate, people were proficient in supplying themselves with a variety of food from the beginning of the Settlement period (AD 871 ± 2) (Karlsson, 2000). During the Middle Ages, sheep, dairy products, and fish, especially dried fish, were the primary staples of the diet and a substantial proportion of terrestrial foods were acidified, soured, fermented, or salt preserved (Mehler, 2011). It was common to consume bone marrow as well, particularly from sheep long bones (Outram, 2003). Other protein sources included cow, pig, duck, goose, and various sea birds as well as eggs (Mehler, 2011;Svanberg & AEgisson, 2012). Domestic fowl (Gallus gallus sp.) have been only very rarely identified from Icelandic archaeological sites prior to the 17th century (Hamilton-Dyer, 2010). Additionally, an important trade economy was in place that provided the country with access to a variety of nonlocal food products. For example, one historical record dated AD 1567 states that for over 50 years traders from Bremen, Germany imported tons of flour, salt, beer, vinegar, grains, and bread, although most of these commodities did not become dietary staples until the 20th century (DI 15, nr. 12;Gísladóttir, 1999;Mehler, 2011). Despite ample local production, fish and meats were also imported to Iceland and exchanged for goods, such as walrus ivory and refined sulfur (Mehler, 2011).
People residing in high latitude regions generally attain much of their protein from animal sources (Cordain et al., 2000). For example, isotopic studies conducted on the Greenland Norse revealed that in the early period of settlement approximately 20% of the diet was marine-based, increasing to 50-80% in later periods (Arneborg et al., 1999;Nelson, Heinemeier, Lynnerup, Sveinbjörnsdóttir, & Arneborg, 2012). A study of proto-Inuit individuals revealed δ 15 N values reached as high as 20‰ due to the frequent consumption of whales and seals (Coltrain, Hayes, & O'Rourke, 2004). Seals were a notable meat and fat source in historical Iceland, but whales were only consumed if stranded or beached (Mehler, 2011;Outram, 2003;Riddell, 2015).
In Iceland, historical and archaeological evidence suggests that subsidiary traditions primarily came from Scandinavia; however, the longstanding tradition of gathering and consuming seaweed (Dulse) likely came from Celtic settlers during the Settlement Period (Sigurðsson, 1988). Isotopic studies have demonstrated that the consumption of seaweed can raise δ 13 C values, resulting in a dietary signature that appears to be more marine-based (Schulting, Vaiglova, Crozier, & Reimer, 2017). Marine isotope signals could also be introduced in humans who consume terrestrial animals that graze upon seaweed (Schulting & Richards, 2009). In historical Iceland, sheep that were kept near seaweed beaches regularly grazed upon seaweeds and farmers also deliberately fed dried seaweed to their sheep for up to 18 weeks a year (Hallsson, 1964). Fishmeal, made of Icelandic herring, is often mixed with hay to supplement protein for sheep, particularly ewes (Sveinbjörnsson & Einarsson, 1998). Although it is not clear when the use of fishmeal as sheep fodder began, the steady exploitation of herring started as early as the 9th century AD (McGovern, Perdikaris, Einarsson, & Sidell, 2006). The heavy dietary reliance on meat from fish-eating sheep in historical Iceland must therefore be considered as a potential confounding factor in isotopic signals determined from human bone collagen.
2.1 | Dietary reconstruction in Iceland: Carbon (δ 13 C carbonate and δ 13 C collagen ), nitrogen (δ 15 N collagen ), and sulfur (δ 34 S collagen ) isotopes Pagan and early Christian bone samples selected from all over Iceland by Sveinbjörnsdóttir et al. (2010) produced δ 15 N collagen values between 6.5 and 15.5‰ and δ 13 C collagen values as ranging between −20.3 and −16.4‰, with most of the data in the range of −20 to −18‰. Recently, Price and Gestsdóttir (2018) (Sayle et al., 2016). This combination of isotopic analyses enabled the researchers to differentiate between terrestrial, freshwater, and marine dietary signals as well as to identify possible migrants and examine aspects of livestock trade and animal husbandry (see Sayle et al., 2016). However, due to differences in geology and volcanic activity, it must be noted that δ 34 S values measured in skeletal material from northern Iceland may not necessarily entirely correspond with δ 34 S values determined from skeletal material from southern Iceland. Isotopic value ranges established from Icelandic material (i.e., animals, water, and geology) to date are presented in Table 2. 2.2 | Residential mobility in Iceland: Strontium ( 87 Sr/ 86 Sr) and oxygen (δ 18 O) analysis of dental enamel Price and Gestsdóttir (2006) analyzed strontium isotope ( 87 Sr/ 86 Sr) ratios in dental enamel from 83 pre-Christian individuals buried across the country, identifying at least 32 nonlocal migrants, that is, those having 87 Sr/ 86 Sr values above~0.7092 -the value for rain and seawater -which defines the upper 87 Sr/ 86 Sr end member for the basaltic biosphere of Iceland. Recently, Price and Gestsdóttir (2018) expanded this study by bringing the total sample size to 127 individuals and by analyzing 87 Sr/ 86 Sr, δ 18 O, and δ 13 C from dental enamel and δ 13 C and δ 15 N from bone collagen. Icelandic inhabitants 87 Sr/ 86 Sr ratios range between 0.7055 and 0.7092 (Price & Gestsdóttir, 2018). Icelandic geologic and bioavailable strontium isotope baselines ( 87 Sr/ 86 Sr) are presented in Table 3. Price et al. (2015) noted that bioavailable strontium isotope ratios in Iceland are higher than those from whole rock (~0.7030-0.7037; Sigmarsson et al., 1992) due to the sea spray that occurs over much of the country. Such a phenomenon has been observed in other basaltic and particularly high-rainfall environments T A B L E 1 Examples of some of the most commonly consumed edible wild plants in historical Iceland (see Gísladóttir, 1999;Mehler, 2011;Svanberg & AEgisson, 2012) such as the Isle of Skye (Evans, Montgomery, & Wildman, 2009) and Hawaii (Capo, Stewart, & Chadwick, 1998). The swamping of geological 87 Sr/ 86 Sr by atmospheric deposition via rainwater and marine seasplash and spray has also been observed in maritime granitic and sandstone island biospheres such as the Western and Northern Isles of Scotland where the geology contributes 87 Sr/ 86 Sr values considerably higher than 0.7092 Montgomery, Evans, & Cooper, 2007). Price et al. (2015) and Gestsdóttir (2006, 2018) also suggested that variations in 87 Sr/ 86 Sr may reflect dietary differences between inland (values closer to 0.7030 reflecting the basalt geology) and coastal (values closer to 0.7092 reflecting the input of seawater) dwelling individuals. Strontium isotope ratios close to those of seawater in archaeological coastal and island populations are often coupled with unusually high strontium concentrations and such a combination may be explained by the high concentration of diet-derived strontium from marine plant-based resources and the extent of aerial sea spray deposition (Montgomery et al., 2007;Montgomery & Evans, 2006). Cultural culinary and husbandry practices, such as using seaweed for food, fodder, or fertilizer, grazing animals on coastal floodplains or preserving food in sea salt, may thus elevate animal and human strontium concentrations and provide a dietary 87 Sr/ 86 Sr value of 0.7092 even when humans are not consuming fish or marine mammal meat which are both low in bioaccessible strontium (unless bones are consumed) compared to plant-based foods (Montgomery et al., 2007;Montgomery & Evans, 2006).
The first published oxygen isotope ratios were measured from a small subset (n = 5) of the same Viking Age samples subjected to strontium isotope analysis by Price and Gestsdóttir (2006) (see   (Price & Gestsdóttir, 2018). Additionally, bone and tooth samples of one early Settlement Period female migrant, Bláklaedda Konan (LKS 1), were subjected to 87 Sr/ 86 Sr, δ 18 O , δ 15 N, and δ 13 C isotope analyses, as well as lead (Pb) and strontium (Sr) trace element analyses (Montgomery & Jakob, 2015). The results of both studies provide a limited pool of T A B L E 2 Isotope value ranges from both modern flora, seawater, freshwater, geology, and archaeological and modern fauna from Iceland  Sayle et al. (2013), and references therein, report the range of δ 34 S‰ in Icelandic geology (volcanic/basaltic) and water sources (ground, flood, river, and spring water).
T A B L E 3 Strontium ratios ( 87 Sr/ 86 Sr) determined in bedrock, grass grown on volcanic soil, modern barley, and seawater and from modern and archaeological faunal dental enamel. The sheep samples were collected from inland sheep that did not graze upon seaweed, which can raise strontium isotope ratios (Price, Frei, & Naumann, 2015) Enamel  Åberg (1995) comparative data. Modern precipitation of δ 18 O values in Iceland ranges from −13 to −8‰ (see Price et al., 2015, figure 20;Bowen, 2018) and in modern Icelandic groundwater from −8.8 to −8.2‰ (n = 11) (Friedrich & Schlosser, 2013).

| Trace element analysis of humans in Iceland:
Zinc (Zn), lead (Pb), barium (Ba), and strontium (Sr) Assessing trace element variability between individuals within or at differing archaeological sites can indicate intrapopulation and interpopulation differences in geographic provenance, particularly when the elements are nonessential and not subject to homeostatic control. These applications are made possible because living organisms absorb elements through the consumption of water and food (Jaouen & Pons, 2017 in Icelandic geology, groundwater, soil, and plants (see Table 4). For example, Panek and Kepinska (2002) found no evidence for anthropogenic lead input in Icelandic soil and plants, particularly in comparison with the concentrations found in Sweden and Poland.

| METHODS
In this study, bone samples from 50 humans (36 from Skriðuklaustur and 14 from Skeljastaðir) and 25 animals (from Skriðuklaustur) were subjected to isotope analysis for carbon (δ 13 C), nitrogen (δ 15 N) and sulfur (δ 34 S). The sample selection was informed by state of preserva-  Price and Gestsdóttir (2006) were available.
The enamel samples were prepared following the procedure given in Montgomery (2002 4 | RESULTS

| Carbon, nitrogen, and sulfur isotope analysis in bone collagen
All human and animal skeletal samples provided well preserved bone collagen with C:N atomic ratios (SKR C:N atomic mean of 3.2, ÞSK C: N atomic mean of 3.3) falling between 3.0 and 3.4 (see Ambrose, 1990;DeNiro 1985). No notable differences were observed between sex and age groups. However, two male individuals from Skriðuklaustur (SKR 135, lower, and SKR 172, higher) and one male from Skeljastaðir (ÞSK 44, higher) had outlying δ 34 S values. Descriptive statistics are presented in Table 5. Overall, at Skriðuklaustur (n = 36), the range and overall δ 13 C mean value indicate a marine dietary signal (closer to the marine dietary end member of −12.5‰) (see Table 5 and Figure 2), while at Skeljastaðir (see Table 5  both studies. Therefore, the overall δ 13 C mean reported in the present analysis at Skeljastaðir is raised by approximately 1‰ when the two datasets are combined and calculated together. This offset in the stable isotope results is difficult to explain but may be due to differences in procedures in extraction methods, stable isotope analytical protocols, and instrumentation. Overall, the carbon and nitrogen isotope ratios in adult bone collagen from Skriðuklaustur show a diet dominated by mixed marine and C 3 terrestrial protein resources, which is significantly different to the predominately C 3 terrestrial protein diet seen at Skeljastaðir. The mean values of δ 13 C, δ 15 N, and δ 34 S determined in animal bones (n = 25) from Skriðuklaustur are presented in Table 6 and plotted in Figures 1 and 2. The isotope values determined in the human bone samples can only be compared with animal baselines from Skriðuklaustur ( Figure 2 and Table 6) and elsewhere in Iceland (see Table 2).  Sayle et al. (2013) reported sheep/goat samples with means of δ 13 C −21.2 ± 0.4‰, δ 15 N 2.5 ± 1.1‰, and δ 34 S 6.7 ± 1.9‰ but did not report any outliers in δ 15 N values. This provides evidence that notably different husbandry feeding practices were being followed for domestic animals.
4.2 | Strontium, oxygen, and carbon isotope and trace element analysis in dental enamel The δ 18 O, δ 13 C, and 87 Sr/ 86 Sr (n = 31) determined from the dental enamel samples are presented in Table 7 and isotope ratio means and ranges are presented in Table 8. The humans from Skriðuklaustur T A B L E 5 Mean and SD for δ 13 C, δ 15 N, and δ 34 S analyses among sampled individuals from Skriðuklaustur and Skeljastaðir. At Skriðuklaustur, the δ 34 S means does not include SKR 174 due to an insufficient quantity of collagen for analysis, therefore for δ 34 S values n = 11 for males, n = 29 for adults, and n = 35 for all individuals. At Skeljastaðir, the δ 34 S means do not include the individuals sampled by Sveinbjörnsdóttir et al. (2010), therefore n = 8 for males, n = 6 for females, and n = 14 for all individuals in the determined δ 34 S values. The values for ÞSK 16, 34, and 48 from Sveinbjörnsdóttir et al. (2010) were not included in these means because they were also analyzed in this study determined from the dental enamel samples are presented in Table 7 and Supplementary Figures S1 and S2 and the medians and ranges in  Gestsdóttir andPrice (2003, 2006) and Price and Gestsdóttir (2006     Strontium is a nonessential trace element that enters biological tissues via calcium pathways following the ingestion and metabolization of food and drink. In omnivores, such as humans, known antagonisms, and synergisms with other elements and food types result in skeletal strontium being predominately derived from the plant part of the diet. It passively enters and remains in the skeleton by substituting for calcium and is not subject to homeostatic control. The amount of strontium incorporated into skeletal tissues is dependent on dosage but is thought to reflect the amount of strontium that is bioavailable from the local environment, diet, dietary calcium, and any strontium released from the skeleton through calcium homeostasis (Montgomery, Evans, Chenery, Pashley, & Killgrove, 2010). Similarly, barium is a sensitive dietary indicator that, like strontium, is not subject to homeostatic control but is absorbed by plants in smaller amounts and may be discriminated according to rising trophic chain position (Szostek, Głąb, & Pudło, 2009). In this study, the strontium concentrations range from 25.8 to 156.7 ppm with a median of 64.1 ppm and the barium concentrations determined in the same individuals range from 0.11 to 0.81 ppm with a median of 0.40 ppm.
Ancient and modern dental enamel concentrations of zinc reportedly range from 58 to 2,100 ppm according to a combination of studies reviewed by Ezzo (1994). Some studies have identified a positive correlation between zinc and lead concentrations and areas of increased urbanization and industrialization (see Ezzo, 1994;Tvinnereim, Eide, Riise, Fosse, & Wesenberg, 1999). Lead concentrations in dental enamel as associated with geological or environmental exposure, T A B L E 8 Medians and ranges of trace elements and means and ranges of isotope ratios determined in the human dental enamel samples from Skriðuklaustur. Samples SKR 100, 115, and 189 were run during the second analytical session which gave the average 87 Sr/ 86 Sr value and reproducibility for the international isotope reference material NBS987 as 0.710266 ± 0.000009 (2 SD; n = 6) while all others were run in the first analytical session which gave 0.710258 ± 0.000013 (2 SD; n = 12)  , including LKS 1 (Bláklaedda konan) (Montgomery & Jakob, 2015), are indicated by the yellow points. Icelandic bedrock (0.7030-0.7037), sheep (0.7059-0.7069), and seawater (0.7092) are indicated by the horizontal lines (Sigmarsson et al., 1992;; also see Table 3). Error bar for 2 SD for δ 18 O at ±0.5‰ and 87 Sr/ 86 Sr at 0.002 rather than from anthropogenic exposure, are generally less thañ 0.7 ppm (Millard et al., 2014;Montgomery et al., 2010). The zinc concentrations in this study range from 43.8 to 145.8 ppm with a median of 95.3 ppm, while the lead concentrations range from 0.13 to 9.4 ppm with a median of 0.61 ppm.

| Monasteries, travel, and trade
Medieval Icelandic monasteries were normally located on major travel routes or near coastal settlements where most of the population resided at the time. Although Skriðuklaustur now appears to be situated in a remote inland valley, during its occupation it was centrally located on a major routeway between the northern and southern parts of the country (Kristjánsdóttir, 2012, p. 296;Kristjánsdóttir, 2016). Until the 17th century, when the route closed due to climate change, pilgrims, patients, fish traders and other individuals easily traveled over the Vatnajökull glacier to reach Skriðuklaustur in the Fljótsdalur valley (Björnsson, 2009, p. 243;Kristjánsdóttir, 2016) ( Figure 1). Severe environmental and epidemiological conditions were likely to have been major catalysts for the movement of people to Skriðuklaustur, which represents one of the largest skeletal assemblages excavated in Iceland and a high prevalence of pathological conditions. The Black Death first came to Iceland at the beginning of the 15th century, killing more than half of the population (Karlsson, 2000, pp. 114-117;Kristjánsdóttir, 2016;Júlíusson, 2018). The second wave of the Plague occurred around AD 1495-1496, just after the establishment of monastery (Kristjánsdóttir, 2016). Furthermore, the only confidently diagnosed cases of treponemal disease in Iceland were found at Skriðuklaustur (Kristjánsdóttir, 2012;Walser, Kristjánsdóttir, Gowland, & Desnica, 2018), indicating that F I G U R E 5 (a) Plot of 87 Sr/ 86 Sr and concentration (ppm) from the dental enamel samples from individuals buried at Skriðuklaustur. The teal marker labeled LKS 1 represents an early Settlement Period migrant to Iceland known as Bláklaedda konan (Montgomery & Jakob, 2015 Walker, Power, Connell, & Redfern, 2014). The Black Death and other plagues in Iceland also coincided with climate changes such as sustained summer rains and cooling weather, which lead to grass and crop failure, increased disease burden, caused food shortages and increased the number of homeless people (Kristjánsdóttir, 2016).
It is known that the brethren residing at Skriðuklaustur aimed to buy farms near the coast: access to the valuable resources (e.g., fish, driftwood, whales, and seals) that coastal sites offered being the probable driver. Those running monasteries found a variety of ways for earning money, including participation in local and international trade (Steinsson, 1965, p. 108;Steinsson, 1966;Kristjánsdóttir, 2016). For example, refined sulfur, an important commodity in medieval trade was found at Skriðuklaustur (Kristjánsdóttir, 2012;Mehler, 2011), possibly for medicinal uses or the production of vermilion (Mehler, 2015). Other notable imports include an effigy of Saint Barbara, a monastic trumpet and rare ceramic pottery imported from France (Kristjánsdóttir, 2012;Mehler, Kristjánsdóttir, & Kluttig-Altmann, 2018). Other sources of income included donations from benefactors from the local community, the sale of books and payment for medicinal treatment and community charity (Kristjánsdóttir, 2016). The Skriðuklaustur monastery also partly depended upon foreign commerce for dietary resources. A large amount of fish bones, primarily from cod (Gadus morhua), ling (Molva molva), haddock (Melanogranmus aeglefinus), shark, and rays were also found during the excavation at Skriðuklaustur, indicating that marine fish were indeed an important dietary component. Smaller fish (60-80 cm in length) are generally found in the Greenland Sea and around the northern and eastern coasts of Iceland, while larger fish (often over 100 cm), such as those found at Skriðuklaustur, are normally caught around the southern and western coasts (Kristjánsdóttir, 2016). It is important to note that larger fish tend to reach higher trophic levels and elevated isotope values (Schoeninger & DeNiro, 1984;Häberle et al., 2016). Fish were both dietarily and culturally important at Skriðuklaustur, where religious fasting was practiced (Kristjánsdóttir, 2017).
Zooarchaeological research demonstrated that fresh fish were regularly consumed at the monastery unlike at most inland sites where only dried fish are normally found (Hamilton-Dyer, 2010;Pálsdóttir, 2006). This notable difference was undoubtedly connected with the religious fasting practiced at the monastery (Pálsdóttir, 2006). However, bipedal animals such as poultry were also permitted for consumption during fasting periods among some Augustinian and other monastic orders (Kristjánsdóttir, 2017) and seals were also still consumed during the fast.
The results of bone collagen stable isotope analyses for carbon and nitrogen on individuals from Skriðuklaustur align with the historical and zooarchaeological evidence, demonstrating a diet substantially derived from marine and potentially freshwater fish protein in some individuals (see Figure 2). No sex-based differences in dietary intake were noted. Based on a trophic shift of +1‰ for carbon and 5.5 ± 0.5‰ for nitrogen (see Fernandes, 2015;Sayle et al., 2016) resulting in a δ 13 C value of −20.7‰ and a δ 15 N value of 7.7 ± 0.5‰, none of the individuals at Skriðuklaustur subsisted on an entirely terrestrial protein diet in adulthood. However, if the two sheep that grazed on seaweed and/or fishmeal (see Figure 2)

| Mobility and geographic provenance
Prior to this research, the origin of the individuals buried at Skriðuklaustur was unknown. Due to the site's hospital, monastic, and trade network functions, it was hypothesized that the individuals buried there could be represented by locals, foreign traders, pilgrims or patients seeking treatment from elsewhere in the country or from abroad (Kristjánsdóttir, 2012;Kristjánsdóttir, 2017). The strontium F I G U R E 6 Cleft maxilla and premaxilla from SKR 22, inferior view. © Joe W. Walser III / National Museum of Iceland and oxygen isotope results demonstrate that all the individuals sampled during this study were likely to be of indigenous origin (see Table 7 and Figure 5). None of the individuals exceed the 87 Sr/ 86 Sr value of rain and seawater (0.7092), which implies an Icelandic origin.
The possibility that some individuals had originally resided on chalk, which has a range of 0.708-0.709, (e.g., Denmark and Southern Brit-  et al., 2014). Furthermore, the δ 18 O phosphate range is~2‰, which is normal for a single temporally contemporaneous population, indicating that these individuals most likely represent a local group of people.
Overall, the results suggest that the individuals sampled from the temporally constrained (1493-1554 AD) cemetery at Skriðuklaustur were of Icelandic origin. As the 87 Sr/ 86 Sr value approaches the value of rain and/or seawater, δ 13 C carbonate values move away from a wholly terrestrial value thereby suggesting a higher input of marine-derived strontium and carbon into the diet ( Figure 6). However, it is important to note that δ 13 C carbonate reflects whole diet (e.g., fat, carbohydrates, and protein) rather than just the protein component of the diet. This positive correlation is strengthened by the Sr concentrations which are also positively correlated with δ 13 C carbonate and 87 Sr/ 86 Sr: as Sr isotope ratios approach the seawater value the amount of Sr in the enamel increases (see Figure 6). This suggests that the higher values are indicative of individuals that grew up in a coastal area where the food web was impacted by marine sea spray and splash, consumption of seaweed or seaweed-eating fauna, while those with lower ratios and concentrations resided further inland where the basalt dominated the food chain. The correlation between these three F I G U R E 7 A medieval map of Iceland depicting the four governmental quarters of Iceland until the 17 th century ( © National and University Library of Iceland). Skriðuklaustur (blue) served the south-eastern quadrant depicted in green. These districts were not active during the occupation of Skeljastaðir (red) parameters is logical but rarely observed so clearly in human populations-highlighting the advantages of studying populations inhabiting the geologically homogenous islands providing two-end member systems-and suggests that it is possible to use isotope analysis to identify residence within different residential zones in Iceland.

One older adult male (SKR 174) with bone changes indicative of
Paget's disease (e.g., expansion of cranial diploë, endocranial "cotton wool" appearance and diffuse, irregular new bone formation throughout the cranium and all long bones) (see Ortner, 2003) had the lowest 87 Sr/ 86 Sr value, which may imply that he resided at a site further inland during childhood (see Figure 7). According to clinical research, individuals with Paget's disease involving the tibia, femur, or acetabular portion of the ilium have clinically and statistically significant functional and mobility impairments (Lyles et al., 1995). Aside from limited mobility, the condition often causes muscular diseases (e.g., atrophy) and sensory or psychological impairments (e.g., hearing loss, dementia) (Kimonis et al., 2008;Monsell, 2004) that can lead to social disability (see Roberts, 2000). This individual possibly moved to Skriðuklaustur from a site further inland for treatment or hospice care, potentially with the aid of his community. According to written documents, medieval Icelandic monasteries each served specific regions or districts of the country, meaning that people traveled to their regional monastery when in need of their services (Kristjánsdóttir, 2016;Kristjánsdóttir, 2017). In some cases, bodies were even moved from the place of death to the local monastery for proper burial and funerary services (Kristjánsdóttir, 2017, pp. 134, 248-249). While not all monasteries in Iceland served as hospitals, they were generally obligated to provide hospitality to travelers and the poor (Kristjánsdóttir, 2016;Kristjánsdóttir, 2017). As Skriðuklaustur served the south-eastern area of the country (Kristjánsdóttir, 2016) (see Figure 7), the isotope results might suggest that some of the sampled individuals migrated there from other places within its district. On the other hand, according to the 87 Sr/ 86 Sr results of Price and Gestsdóttir (2006), only one individual (ÞSK 39) out of 33 analyzed from Skeljastaðir was of nonlocal origin. This individual also exhibits a δ 15 N value significantly higher than other individuals measured from Skeljastaðir both in this study and as reported in Sveinbjörnsdóttir et al. (2010) (see Figure 2). The difference in this individual's diet may therefore be correlated with their foreign provenance prior to their migration to Skeljastaðir.

| Dietary reconstruction
The results of this study indicate that a primarily terrestrial protein diet was consumed at Skeljastaðir, thus confirming the results by Sveinbjörnsdóttir et al. (2010). Within a few decades of AD 1000, during the Medieval Warm Period, sea fishing in the North Atlantic increased dramatically (Barrett, Locker, & Roberts, 2004). The lack of significant marine dietary input at Skeljastaðir may possibly relate to a reduced reliance on sea fishing during the occupation of the site (1000-1104 AD). However, this may also simply relate to the long distance (~60 km) between Skeljastaðir and the coast. In parallel, using δ 13 C and δ 15 N isotope analysis on British skeletons, Müldner and Richards (2007) (Dugmore et al., 2007;Gestsdóttir, 2014;Steffensen, 1943;Þórðarson, 1943). When compared with the δ 13 C and δ 15 N values from bone collagen, the results of the sulfur isotope (δ 34 S) analysis suggest that some individuals at Skeljastaðir were consuming more freshwater protein than others, while those from Skriðuklaustur appear to have been consuming both saltwater and freshwater resources (see Table 7 and likely other sites in the east as well (Rafnsson, 1990, p. 93, 100).
Skriðuklaustur was established just 16 years later (Kristjánsdóttir, 2016) implying that the surrounding environment may have still been considerably altered by volcanic emissions. As sulfur isotope ratios may reflect both diet and geological provenance, it is important to consider that δ 34 S values may increase in populations residing close to active, or erupting, volcanic systems.

| Trace element analyses
The trace element analyses of dental enamel determined low Ba concentrations, which probably results from the low Ba content in seawater and Icelandic groundwater and basalt (see Naimy, 2008) (see Supplementary Figure S1). The low values, small range, and little variation in Ba concentrations also corroborate the interpretation that the people residing at Skriðuklaustur represented a local population of individuals that grew up in Iceland. The means for Pb indicate that at least some anthropogenic exposure to lead occurred at Skriðuklaustur (Supplementary Figures S1 and S2). Only 12 individuals exhibited Pb concentrations greater than~0.7 ppm (see Table 7), which is normally considered the threshold in archaeological human dental enamel between natural and anthropogenic Pb exposure (Millard et al., 2014;Montgomery et al., 2010 were used both as clothing and currency (vaðmál). Due to the importance of cloth currency for its use to pay tithes, taxes and other legal or economic transactions, cloth production was intensely standardized. Some evidence even suggests that legal controls may have regulated occupational roles and simultaneously that cloth currency production was also an important form of female agency to Icelandic society (Norrman, 2008;Smith, 2014). Additionally, the accumulation, retention, and susceptibility to the health effects of exposure to toxic metals have been demonstrated to differ between men and women, particularly during pregnancy or menopause (Vahter, Åkesson, Lidén, Ceccatelli, & Berglund, 2007). Regarding children, it is well established that they not only absorb a far greater amount of ingested lead than adults but are also more frequently exposed to it due to hand-tomouth activities and other behavioral tendencies such as outdoor play (see Jacobs & Nevin, 2006;Wittmers, Aufderheide, Rapp, & Alich, 2002). Women and children may have been more regularly exposed to lead at home due to more frequent contact with lead objects and structures within the household. A young adult female (SKR 189) with cystic echinococcosis (hydatid disease) had a Pb concentration of 9.40 ppm, significantly higher than all others in the sample set, indicating substantial childhood anthropogenic exposure to lead (see Montgomery et al., 2010). The individual (SKR 23) with the second highest Pb concentration (4.1 ppm) was a young adult (c. 17-25) female with treponemal disease, exhibiting skeletal changes consistent with venereal syphilis, according to the criteria described by Hackett (1976) and Ortner (2003). Clinical research has also demonstrated that deficiencies in essential minerals, such as calcium, can result in the abnormal and rapid absorption of toxic heavy metals or trace elements, such as Pb, particularly in malnourished children (Talpur, Afridi, Kazi, & Talpur, 2018). It is thereby possible that some individuals with elevated lead concentrations may have had low calcium intake during childhood. It is evident that parts of the cemetery were contaminated with anthropogenic lead, which is likely to be associated with the infrastructure of the monastery (e.g., lead window frames; door hinges) and objects found on site (Kristjánsdóttir, 2012). By the 17th century, lead-glazed kitchenware significantly increased in availability in Iceland, particularly among high status individuals, but it has also been found at sites dating to the early medieval period (Þorgeirsdóttir, 2010). For context, Rasmussen, Skytte, Jensen, and Boldsen (2015) measured Pb concentrations in individuals that resided in rural, monastic, and urban sites around Denmark and northern Germany.
Their results indicate that higher status, urban dwellers were more likely to live among lead structures (e.g., window frames, roof tiles possibly in contact with drinking water) and be able to afford lead or lead-glazed kitchenware. Another female (SKR 65) exhibited a Pb concentration of 3.51 ppm. She was one of a few individuals found buried within the church itself, potentially indicating that she was a benefactor or had a special status at the monastery (Kristjánsdóttir, 2010;Walser et al., 2018). It is therefore possible that the individuals with elevated anthropogenic Pb concentrations were exposed to lead within their households or the monastic grounds, if they resided there during childhood. Nonetheless, with a small number of exceptions the range of Pb concentrations determined in dental enamel is largely below~0.7 ppm, suggesting that those analyzed from Skriðuklaustur represent individuals that grew up in an unpolluted environment, such as Iceland (see Montgomery et al., 2014).
The highest Zn concentration was 145.8 ppm (SKR 10), which is still well within the lower spectrum of expected Zn concentrations in dental enamel (9.9-1,550 ppm) (see Jaouen, Herrsher, & Balter, 2017). Clinical studies have noted a relationship between malnutrition and lower enamel zinc concentrations (Brown et al., 2004) and that enamel Zn concentrations of <90 ppm may reflect marginal zinc supply during childhood (e.g., Tvinnereim et al., 1999). The individuals with the lowest zinc concentrations include an adult female (SKR 195) (43.8 ppm) and an adult male (SKR 150) (47.3 ppm), possibly implying limited zinc supply during childhood (see Supplementary Figure S2).
These two individuals also exhibit dental enamel hypoplasia, a pathological indicator of metabolic or health stress during childhood (see Ortner, 2003). However, zinc is an essential trace element under homeostatic control and its concentrations are altered by numerous and complex interactions including diet, disease, individual variation, digestion, and absorption. As a result, zinc concentrations determined in dental enamel may not accurately reflect palaeodiet (Dolphin & Goodman, 2009;Ezzo, 1994  approximately 12 individuals from Skriðuklaustur may have been exposed to anthropogenic sources of lead during childhood (Pb concentrations >0.7 ppm), possibly from the lead objects and window frames that were found during the excavation of the monastery ruins.
Overall, the dietary differences noted between the two noncontemporary inland sites may reflect cultural changes in trade, subsistence strategies, and environment (e.g., the Little Ice Age, volcanic eruptions) in medieval Iceland. Furthermore, the results of isotope analyses conducted on individuals excavated from Skriðuklaustur indicates that the monastery was operated, visited and inhabited by the local population of brethren, pilgrims, patients, and other local individuals. Considering the functions of the monastery, these findings also provide further evidence for the movement of disease (e.g., syphilis), goods (e.g., lead wares), food (e.g., fish, fruit), and information (e.g., medicine) from other parts of Iceland and abroad, implying that these past people were never found in isolation, even at the edge of the world.