Intra‐ and inter‐tooth variation in strontium isotope ratios from prehistoric seals by laser ablation multi‐collector inductively coupled plasma mass spectrometry

Rationale Strontium isotope ratios (87Sr/86Sr) in modern‐day marine environments are considered to be homogeneous (~0.7092). However, in the Baltic Sea, the Sr ratios are controlled by mixing seawater and continental drainage from major rivers discharging into the Baltic. This pilot study explores if variations in Sr can be detected in marine mammals from archaeological sites in the Baltic Sea. Methods 87Sr/86Sr ratios were measured in tooth enamel from three seal species by laser ablation multi‐collector inductively coupled plasma mass spectrometry (LA‐MC‐ICP‐MS). The method enables micro‐sampling of solid materials. This is the first time that the method has been applied to marine samples from archaeological collections. Results The analyses showed inter‐tooth 87Sr/86Sr variation suggesting that different ratios can be detected in different regions of the Baltic Sea. Furthermore, the intra‐tooth variation suggests possible different geographic origin or seasonal movement of seals within different regions in the Baltic Sea through their lifetime. Conclusions The method was successfully applied to archaeological marine samples showing that: (1) the 87Sr/86Sr ratio in marine environments is not uniform, (2) 87Sr/86Sr differences might reflect differences in ecology and life history of different seal species, and (3) archaeological mobility studies based on 87Sr/86Sr ratios in humans should therefore be evaluated together with diet reconstruction.


| INTRODUCTION
To establish migratory patterns in prehistoric seals in the Baltic is important for understanding both prehistoric human and seal behaviour and the interaction between them. In this pilot study we make use of the fact that the present-day Baltic Sea displays large variations in 87 Sr/ 86 Sr in different parts of the basin. Based on this, and the fact that there are distinct differences in social, foraging and breeding behaviour of the three seal species included in this pilot study, we predict that sequential in situ measurement of 87 Sr/ 86 Sr isotope ratios in seal teeth by laser ablation multi-collector inductively coupled plasma mass spectrometry (LA-MC-ICP-MS) can provide data on life cycle and migratory patterns within and among species.
The 87 Sr/ 86 Sr in oceanic seawater is globally homogeneous within the limits of analytical precision, averaging 0.70918. This is due to a long residence time of Sr in the oceans (5.0 × 10 6 years) that causes Sr to homogenize isotopically by mixing on a timescale of about 1000 years. 1 However, in coastal areas, and in smaller basins such as the Baltic Sea, the 87 Sr/ 86 Sr ratios can vary considerably. The present-day Baltic Sea is a brackish water body, but its complex natural history has involved fluctuations between freshwater and brackish/marine conditions during the Holocene and it is not clear to what extent this has affected the Sr isotope composition. Today, there is a major inflow of freshwater as approximately 50 rivers discharge into the basin. 2 The rivers drain various sediments ranging from different Sr isotope ratios ( Figure 1). 3,4 Accordingly, the rocks in the north have low Sr concentrations and high 87 Sr/ 86 Sr ratios (0.7111-0.7452), whereas in the south they have high Sr concentrations and low 87 Sr/ 86 Sr ratios, closer to that of seawater (0.7095-0.7105). 2,3 In the water areas where rivers with high Sr concentrations discharge, a bias towards Sr river water concentrations is expected, as deviations from perfect mixing with seawater have been observed. 4  Strontium has been extensively used in archaeology to elucidate questions of human and animal mobility. Understanding mobility patterns helps us to reconstruct resource acquisition strategies, social networks and interactions between groups, e.g. trading with goods. 7 Although previous studies of mobility in prehistoric people have assumed that the marine signature is uniform in all environments, 8 the variation seen in 87 Sr/ 86 Sr in the Baltic suggests that studies of people in coastal areas who feed extensively on marine mammals may be affected by this. Different seal species have different ecological requirements and display distinctive foraging and migrating patterns. In the present-day Baltic Sea, there are three species of seal: the grey seal (Halichoerus grypus), the ringed seal (Phoca hispida) and the harbour seal (Phoca vitulina). However, in prehistory there was also the harp seal (Phoca groenlandica). In the present study, teeth from grey seal, ringed seal and harp seal are investigated.
The Baltic Sea population of grey seals is one of three modern grey seal populations that exist independently from each other and have different breeding areas. The other two are the grey seal populations of the western North Atlantic and eastern North Atlantic. 9 The grey seal population of the Baltic Sea breeds in groups at the margin of the fast ice in the Gulfs of Bothnia, Riga and Finland, but breeding on rocky islets has also been recorded during extremely mild winters, 9,10 demonstrating highly adaptive behavior. The grey seal is a non-migratory species, spending most of its time in the proximity of the breeding area but it has been recorded that some, especially subadult, individuals can cover long distances in search of food. 9,11,12 The Baltic Sea population of grey seals breeds in February-March and lactation lasts approximately 3-4 weeks. 13 Harp seal, by contrast, is an ice-dependent and highly migratory species. Today, there are three populations of harp seals with breeding grounds in the White Sea, the Jan Mayen Islands and around Newfoundland. Every year they form large moulting rookeries during spring, and after moulting is completed they start their annual migration northwards to the summer feeding grounds to return to their breeding grounds in the late autumn. 14 They aggregate in many thousands to give birth on pack ice and lactation lasts 10-12 days. 14 During the Litorina stage of the Baltic Sea, harp seals were breeding probably in the region between Gotland and the Åland islands, and in the southwestern Baltic Sea. 15 1000-800 cal BC, and they appear sporadically during the Iron Age. [18][19][20] The ringed seal is a circumpolar species, found along the Arctic coasts, which preferably inhabits inshore waters.

| Sample collection
We sampled six seal teeth in total (five canines and one molar) from grey seal, harp seal and ringed seal ( Table 1) Table 1). One sample is from a modern grey seal from the Baltic Sea. The selection of the samples was based on the grounds that different species have different ecological requirements and display different foraging and migration patterns. In addition, it enabled testing if regional differences in 87 Sr/ 86 Sr can be detected in seals from different sites.

| Ageing of the seals according to teeth
In most phocid seals the deciduous teeth are shed in utero, so that the permanent dentition is already erupted at the time of birth, or erupts shortly after birth. 23,24 In harp seals, the lower canines erupt a week after birth. 25 The canines grow rapidly from an average length at birth of 15.3 ± 0.53 mm to an average length of 26.1 ± 0.84 mm at the age of 1 year. The pulp canal has an average diameter of 6.6 ± 0.6 mm at birth and decreases rapidly within the first 12 months as it becomes filled with dentine. Based on the canine length of 22 mm, the seal from Ajvide (sample 029) was estimated to be younger than 6 months, since harp seal canines within 6 months reach 93% of their length at the age of 1 year. However, as the pulp canal measured 6 mm, corresponding to measurements of newborns, the estimated age is probably considerably less than 6 months. The second harp seal canine, from Korsnäs, has a total length of 28 mm and the pulp cavity is closed, suggesting that it belonged to an animal older than 12 months (Table 1).
The length of canines of newborn grey seals is about 16-22 mm.
The tooth is not fully erupted until the pups are weaned at approximately 3-4 weeks; the canines have by then grown another 5-7 mm. 26 The canine length at 6-8 months is 32.3-38.5 mm. 26 The pulp cavity is broadly open at a newborn stage and it closes gradually during the first 4-5 years of life, until only a narrow canal for the blood vessels remains open. 26 Both the grey seal canines (the modern sample 004 and sample 015 from Stora Förvar) were estimated to belong to seals younger than 3 months, as they measured 26 and 25 mm, respectively, and both had broadly open pulp cavities (8 and 9 mm, respectively). The grey seal molar from Stora Förvar (sample 013) was not possible to age, since there are no ageing data on molars in grey seals, but the fact that the pulp cavity is closed suggests that the seal was older than 12 months (Table 1).
Ringed seals are born with permanent canines 27 and tooth eruption completes shortly after birth. The canine measured 19 mm, which, together with the closed pulp cavity, suggests an adult animal. 28 Table. 29-31 Strontium in tooth enamel is resistant to taphonomic diagenesis, in contrast to strontium in bones and dentine, 32,33 and is therefore suitable for analysis.

| Strontium and incremental growth in tooth enamel
Enamel formation, or amelogenesis, is a prolonged and slow process which consists of two main stages: the secretory phase and the maturation phase. [34][35][36] Enamel formation is controlled and Tooth enamel is suitable for intra-tooth sequential sampling because after its formation there is no remodeling of the tissue. Thus, by analyzing the isotopic composition of enamel, it is possible to study mobility, dietary and climatic factors sequentially from the whole duration of its formation.
The deciduous dentition in phocid seals is shed in utero and the permanent teeth are formed prior to birth. The enamel of the permanent dentition is laid down before birth. 13 Therefore, for this study we assume that the isotopic ratios reflect only the mother's dietary and mobility patterns during gestation.
Although intra-tooth variation in enamel analysis might be biased by the pattern and duration of enamel mineralization, oxygen isotope analyses on enamel apatite from herbivores have shown that the sequential sampling of enamel reflects a chronological sequence, which makes it suitable for detecting seasonal variations. 40,41 Thus, sequential sampling of teeth can provide insights into certain environmental events, diet and mobility.

| Sample preparation
The surface of the teeth was gently cleaned with a brush and deionized water. The samples were then sonicated in deionized water for approximately 1 min and subsequently rinsed with deionized water; this procedure was repeated twice. The samples were then left to dry at room temperature. Before laser ablation analysis was performed, the enamel surface was cleaned with ethanol.

| Horizontal sequential sampling
Sr isotope ratios of enamel were measured in situ on the teeth at the  Table S1 (supporting information). A sample cell with a flexible mounting table allows insertion of the specimens without embedding or cutting them.
After cleaning, the teeth were mounted as parallel as After these corrections are applied, a fractionation factor is calculated based on the stable 86 Sr/ 88 Sr ratio and is used to quantify the contribution of Rb to m/z 87 (stable 87 Rb/ 85 Rb is assumed to be 0.3861). All corrections are applied online to each integration allowing interference-and fractionation-corrected Sr ratios to be determined for every data point ( Table 2).
An additional polyatomic interference on m/z 87 has been reported (e.g. 42,43 ) and is described as 40  corrected online but has to be reduced by a thorough low oxide tuning of the gases. 43 We achieve this through the introduction of nitrogen (~8 mL/min) into the He-Ar-Sample mixture prior to injection into the plasma using a CETAC Aridus II, while simultaneously aspirating pure 0.3 M nitric acid. This allows UO + /U + oxide production to be kept well below 1%.
To verify that all interferences are successfully removed, a tooth reference material (Otomys 26-r52 44

| RESULTS
To facilitate interpretation, we have colour-coded the 87 Sr/ 86 Sr ratios in five intervals in plots and in the map (Figures 1-4).    (26). Dotted line indicates a higher value which is displayed in Figure 5. Error bars are the external precision 2SD (see Table S2, supporting information) [Color figure can be viewed at wileyonlinelibrary.com] Measurements 2-9 (blue) show a gradual increase, until at measurement 10 they enter the green interval, which corresponds to the discharge areas of rivers into the southern Baltic (Table S2, supporting information; Figure 1). Measurements 20-24 (blue) show a slight decrease, whereas the last measurement increases slightly to 0.71008 (green).
The harp seal canine from Korsnäs (sample 009, Figures 2 and 7) yielded 87 Sr/ 86 Sr ratios that were all higher than those of the Ajvide sample, and show much larger fluctuations; the total range is 0.7120-0.7241 (from green to yellow). Measurements 27-29 are omitted because they were made on dentine (Table S2,  The grey seal canine from Stora Förvar (sample 015, Figure 3;  A grey seal molar (sample 013, Figure 3; Table S2, supporting information) from Stora Förvar yielded small-scale fluctuations in the 87 Sr/ 86 Sr ratio between~0.7138 and~0.7155 (green). The first seven FIGURE 3 87 Sr/ 86 Sr ratios for grey seal teeth. The x-axis represents the sampling lines from the tip of the tooth (1) towards the cervix (40). Error bars are the external precision 2SD (see Table S2, supporting information) [Color figure can be viewed at wileyonlinelibrary.com] FIGURE 4 87 Sr/ 86 Sr ratios for ringed seal tooth. The x-axis represents the sampling lines from the tip of the tooth (1) towards the cervix (21). Error bars are the external precision 2SD (see Table S2, supporting information) [Color figure can be viewed at wileyonlinelibrary.com] sampling lines yield ratios that fluctuate around 0.7139, whereas the measurements from the lower half of the tooth oscillate around 0.7150, with a slightly more scattered distribution.
The canine from a modern grey seal from the Baltic Sea (sample 004, Figure 3; The ringed seal canine (sample 006, Figure 4;

| DISCUSSION
All the individuals in this study were sampled in the same way, and the study compares individuals from species of the same family, taking into account any species-dependent differences in the pattern and duration of enamel mineralization that might cause biases regarding interpretation. 41 Harp seals are nowadays highly migratory. They start their annual migration northwards as soon as the ice clears. In general harp seals stay close to pack ice all year round and return to their breeding grounds in late autumn to give birth. The two harp seal canines that were analyzed in this study yielded isotope ratios that were significantly different from each other (Figures 2 and 5 This is in accordance with the life pattern of grey seals as they are non-migratory but can undertake long foraging trips in search of food, as evidenced by the larger variation in 87 Sr/ 86 Sr ratios in the canine from the modern grey seal (sample 004). Recent satellite tracking studies also demonstrate that grey seals can cover long distances within the Baltic Sea in search of food. 11,12 All three grey seal samples indicated foraging in the southwestern Baltic Sea. This is in accordance with the known distribution of the modern grey seal population in the Baltic Sea which has its westernmost distribution around Bornholm. 46 The ringed seal has 87 Sr/ 86 Sr ratios within the same range as the grey seal, indicating a similar geographic distribution in the southwestern Baltic Sea and possible seasonal movements between water sources in the same region as the grey seals. This is in sharp contrast to the present-day distributions in the Baltic Sea, where ringed seals mainly occur in the Gulf of Bothnia, the Gulf of Finland and the Gulf of Riga, where the ice conditions allow proper breeding grounds. 47

| CONCLUSIONS
This study demonstrates that 87 Sr/ 86 Sr ratios in some marine environments such as the Baltic are not necessarily uniform. Differences in 87 Sr/ 86 Sr ratios can therefore be detected in marine mammals to study migration patterns. Further analysis should be performed in order to identify regional differences within the Baltic Sea basin at a more detailed level.
The results of this study also have implications for the archaeological interpretation of human mobility or migration. Strontium in tooth enamel has been used as an indicator of human mobility between regions with different Sr isotope ratios. Our study shows that for prehistoric coastal cultures, especially in regions with freshwater influx such as the Baltic Sea, where marine mammal hunting and consumption have been significant for human subsistence, such an interpretation should be applied with caution. We suggest that the differences in 87 Sr/ 86 Sr ratios between different regions should be reflected in humans, as consumers of seals and fish, and thus interpretation of such data should be ideally performed in combination with carbon and nitrogen isotope analysis as indicators for diet.