The prehistory of the Mediterranean region has long been a subject of considerable interest, particularly the links between human groups and regions of origin. We utilize the spatial variation in the δ2H and δ18O values of precipitation (isoscapes) to develop proxies for geographic locations of fauna and humans. Bone collagen hydrogen isotope ratios (δ2H) in cattle (and to a lesser extent, ovicaprids) across the Mediterranean reflect the isotopic differences observed in rainfall (but δ18O values do not). We conclude that δ2H in herbivore bone collagen can be used as a geolocation tracer and for palaeoenvironmental studies such as tracing past isotopic variations in the global hydrological cycle. In contrast, human bone δ2H values are relatively tightly grouped and highly distinct from precipitation δ2H values, likely due to human-specific food practices and environmental modifications. Given the inter-species variability in δ2H, care should be taken in the species selected for study.
Migration is a profoundly important part of human existence, from the dispersal of Homo sapiens out of Africa to the past and present movement of humans over the globe. The new tools available in the biological and physical sciences have contributed to a lively interest in migration studies of past populations and individuals. The scale at which migration can be studied is quite varied. For ancient DNA, continental and global-scale studies have examined the genomic fine structure of populations and movement or inter-mixing between them1. However, ancient DNA studies lack power to resolve the timing and duration of some migration(s). Separately, variation in stable isotope ratios of tissues can be used to identify migrating individuals or groups directly.
The Mediterranean region has a long history of exchanges and migration, which substantially intensified from the first millennium BCE2,3. We focus here on the Bronze and Iron Ages and the Phoenician-Punic periods (ca. 3000–500 BCE) – a time of increasing social complexity and pan-Mediterranean linkages and exchange. Until now it has been difficult to use stable isotopes to identify migration in this region. In general, latitude and altitude changes can be useful in tracking migrations due to the environmental and dietary differences that often accompany these trajectories. In contrast, the variation in carbon isotopes (δ13C) through broad regions of the latitudinally similar Mediterranean are very slight and trend predominantly along a north-south gradient in Europe, making migrations around the Mediterranean invisible to this biomarker4. Nitrogen isotopes (δ15N) are related to soil composition and environment at a very local scale and are not useful for geolocation5. No comprehensive isoscape (isotopic landscape) of bio-available strontium isotope (87Sr/86Sr) values covering the entire Mediterranean region has been established. Given the large spatial extent of the region, overlapping Sr isotope values of geographically distant regions are found6, rendering this approach an inadequate tool when used alone for pan-Mediterranean migration studies. Instead, we investigate variations in hydrogen (δ2H) and oxygen (δ18O) isotopes in bone collagen of fauna and humans and their relationship with a well-developed isoscape across the region based on precipitation δ2H and δ18O.
Hydrogen and oxygen isotope ratio variations are products of the global hydrological cycle, influenced by temperature and atmospheric transport from source to precipitation location; as a result, δ2H and δ18O in precipitation can vary significantly with geography. The distribution of these isotopes across the Mediterranean basin was first described by Gat and Carmi7. Globally, δ2H and δ18O in precipitation co-vary, forming the Global Meteoric Water Line (GMWL)8. Deuterium excess (d), defined as d = δ2H – 8 δ18O, varies across the globe and is influenced by local aridity and the relative humidity of originating air masses9. The isotopic variation in precipitation can be incorporated into biological tissues and serve as a tracer of location, as exemplified by δ2H in bird feather keratin used for geo-location of the origin of migrant birds10,11, δ2H and δ18O variation in human hair keratin with water variation12, and δ18O in tooth enamel apatites used to distinguish non-local individuals13. Herding strategies and vegetation variation have also been investigated with time-resolved δ18O values of tooth enamel14. In contrast, organic H and organic O in bone collagen have not been developed as a geo-location tool, with the exception of an early study on deer15. Feeding and observational studies show that ingested water isotopic composition is incorporated in bone collagen H and O16,17,18, and ovicaprid dentin collagen19, and thus bone collagen δ2H and δ18O should reflect the environmental source water and the origin of the individual.
For applications to palaeoenvironmental and past migration questions, bone collagen offers advantages over other tissues: i) bones are often preserved in the geological and archaeological record (in contrast to keratin), allowing direct isotopic study of the individual of interest; ii) bone collagen reflects an integrated, albeit variable time period20,21 in contrast to enamel apatite, which is formed at discrete ages and may not capture migration ‘events’; iii) collagen may be directly dated by radiocarbon. Further, ideal sources of δ2H and δ18O of the past, waters and plants, are largely unavailable due to lack of preservation, so that bone collagen may be useful in recording and tracing past environmental variation22.
Here, we present sub-fossil bone collagen δ2H and δ18O from a range of archaeological sites spanning 3500 km across the Mediterranean basin (Fig. 1). Precipitation δ2H, d excess, and to a lesser degree δ18O vary substantially from west (lowest values) to east (highest values), with δ2H ranging over 25‰, δ18O over 2‰, and d excess over 9‰ from Spain to Israel (Fig. 1, Table S1). We investigate whether bone collagen δ2H and δ18O values vary spatially, in concert with the longitudinal variation in precipitation δ2H, δ18O, and d excess values.
Results and Discussion
Bone collagen δ2H and δ18O were measured across 5 archaeological sites (Figs. 1–2, Tables S4-S6). δ2H values in cattle collagen are lowest at the westernmost site (1, Los Berrocales), higher at the central site (2, Sardinia), and highest at the easternmost site (5, Megiddo), and ovicaprid δ2H values are lower in the centre/west than the easternmost site (Fig. 2), matching the isotopic pattern in precipitation. Collagen δ2H in cattle is significantly correlated with mean precipitation δ2H (p = 0.001, r2 = 0.46, slope=0.75 ± 0.19, 1 se, Table S3) and d excess (p = 1.2 × 10–8, r2 = 0.84, Table S3). Ovicaprids also show a significant but weaker correlation between collagen δ2H and precipitation δ2H and d (Table S3). Given d is a parameter related to precipitation δ2H (d = δ2H – 8 δ18O), the correlation of collagen δ2H and rainfall d values is consistent. The bone collagen of these animals is incorporating the isotopic variation in precipitation across the Mediterranean basin, both directly from water and through food sources. The faunal δ2H data are consistent with the modern pattern of precipitation, with lower δ2H and d excess, increased rainfall, and lower aridity in the west compared to the east (Fig. 2, Table S1). The climate in the Mediterranean region in the late Holocene was generally stable after mid-Holocene aridification (see Supplementary Information). If there were any shift in the isotopic composition of precipitation between ~3000–4000 years ago and the present, it has not obscured this collagen-precipitation relationship. In contrast, neither cattle nor ovicaprid collagen δ18O is correlated with mean precipitation δ18O (p = 0.09–0.90, r2 = 0.00–0.09, Table S3).
The covariation of δ2H in collagen and precipitation is consistent with the incorporation of H from drinking water. Additionally, food hydrogen (plant material) and plant water hydrogen also contribute to H in collagen16,17. If the isotopic fractionation from precipitation to plant water and/or plant tissue is roughly similar across the environments sampled here, the same isotopic offset should be translated into the herbivore bone collagen via food and plant water inputs. The large shift in the absolute values of δ2H and δ18O in collagen vs. meteoric water indicates that there is significant fractionation from water to animal tissue (or from water to plant to animal tissue) as noted in previous studies16,17, but despite this, the resulting bone collagen δ2H maintains a relationship with precipitation δ2H and d excess values.
The stronger relationship between bone collagen δ2H and precipitation δ2H in cattle than in ovicaprids may be due to differences in water and plant consumption. Water flux scales with body size23,24 and larger-bodied cattle consume more drinking water than smaller ovicaprids; as a result, cattle may be more directly reflecting precipitation water. Plants can vary in the depth at which they draw water and the relative amount of evapotranspiration25, which may result in variation in plant water δ2H, so that different plants eaten by cattle and ovicaprids may also be affecting the collagen δ2H values seen here. Additionally, differences in the amount of water ingested from plant matter (e.g. leaves) vs surface water may influence inter-species δ2H patterns. In using bone collagen δ2H and δ18O values for geo-location or palaeoenvironmental studies, caution must be used in selecting the species for comparison; δ2H in collagen shows clear differences between some animal species26,27.
Human bone collagen contains the least variation in both δ2H and δ18O across the Mediterranean. Although there is a difference in the δ2H of human bone collagen from the westernmost site in Spain compared with the more easterly sites, reflecting lower precipitation δ2H in the west (Fig. 2; Tables S2-S3), human bone collagen δ18O does not vary across the transect from Spain to Israel.
In contrast to the spread in δ2H values in collagen in cattle and ovicaprids (41 and 47‰ ranges, respectively), human bone collagen δ2H values from these sites cluster relatively tightly (27‰ range, Fig. 2). Similarly, δ18O values show higher ranges in the herbivores (5.4–5.5‰) than in the humans (3.2‰, Fig. 2). The reasons for this clustering in human collagen δ2H and δ18O values are likely due to human manipulation of food and environment, including agricultural practices (c.f. nitrogen isotopic changes28,29), animal management, and food preparation techniques. Cooking and boiling have been shown to increase δ2H and δ18O, for example30,31. Further work on the isotopic differences in human foods vs animal foods is warranted.
In the Mediterranean, hydrogen in bone collagen tracks precipitation δ2H values, especially so in cattle, while oxygen does not (Fig. 2, Table S3). In other studies, similarly, feather keratin also shows a weaker correlation of δ18O values with precipitation compared to δ2H values32. Across the Mediterranean region sampled here, from our westernmost to easternmost site, δ18O in precipitation increases by ~2‰ (−6.7‰ (site 1, west), −4.3‰ (sites 2–4, central), −4.8‰ (site 5, east), Table S1). This range is relatively small given the large range of δ18O values seen at a single site (e.g. 2.3–2.4‰ range in cattle at Los Berrocales and in Sardinia). This latter ‘biological noise’ may arise from variation in fractionation in δ18O values within a single species, perhaps due to food selection or physiological/biochemical parameters (c.f. different species-specific ‘calibration’ lines relating drinking water δ18O and biological apatite δ18O). In this case δ18O lacks geospatial discriminating power in bone collagen. Given the non-trivial analytical challenge in measuring collagen δ18O16,33, it is noteworthy that δ2H can be measured separately and more quickly with a Cr-packed reactor34,35, and that δ2H values in general have a larger range of variation relative to analytical uncertainty. In sum, the use of hydrogen isotopes is a more promising approach relative to δ18O for determination of migrants in the past.
To our knowledge, we present here the first combined δ2H and δ18O values of collagen from different localities, from multiple species. These data demonstrate that δ2H in collagen varies with precipitation δ2H, most strongly in cattle, but also to a lesser extent in ovicaprids and humans. Hydrogen isotopes in bone collagen may be a powerful tool as a geographic discriminator over large geographic scales where the isotopic composition of meteoric water varies significantly. Given that δ2H can also vary between species, care should be taken when including and comparing different species26,27.
Humans demonstrate starkly different δ2H values from co-local fauna, and a smaller range in both δ2H and δ18O than in the fauna. We posit that human-specific dietary and cultural factors (e.g. cooking, agricultural modification) are at play, and the isotopic separation between humans and their habitat indicates a high degree of environmental manipulation.
Bone samples were obtained from five sites dated to the Bronze and Iron Ages: Los Berrocales (Spain, lat. 40.375, long. −3.578, 22–16th century cal BCE)36, S’Urachi (Sardinia, lat. 40.015, long. 8.583, 10–5th century cal BCE), Villamar (Sardinia, lat. 39.619, long. 8.96, 4–2th century cal BCE), Monte Sirai (Sardinia, lat. 39.179, long. 8.488, 7–4th century cal BCE), and Megiddo (Israel lat. 32.585, long. 35.184, 20–13th century cal BCE, pers. comm. Melissa Cradic, Robert Homsher, Mario A.S. Martin). Site descriptions and the archaeological contexts of the samples are given in Supplementary Information. Bone fragments (human and faunal) were demineralized in 0.5 M EDTA, rinsed in distilled water numerous times, and freeze-dried.
Freeze-dried collagen samples (~300 μg) were weighed into silver capsules, and introduced via a zero-blank autosampler into a Thermal Conversion Elemental Analyzer (TCEA) for pyrolysis to H2 and CO gases. The gases were separated using a gas chromatograph (1.8 m length, 5 Å molecular sieve) and were introduced into an isotope-ratio mass spectrometer for δ2H and δ18O determination. δ2H and δ18O values were normalized on the VSMOW-SLAP scale, using VSMOW and SLAP in silver divots as references. The TCEA was packed with a Cr-metal powder filling, as described previously34,35, which has been shown to result in quantitative conversion of N-containing organics to H2 gas34. Consequently the δ2H data reported here is as obtained by the more newly-developed Cr-packing method. Our long-term reproducibility on organic samples is 3.6‰ for δ2H (Cr method) and 0.9‰ for δ18O.
We did not correct δ2H for exchangeable H, given that it is typically low in non-gelatinized collagen, all samples were analyzed in our laboratory over a few months, and there is still discordance between laboratories in how to perform water-collagen exchange experiments to carry out this correction (see ref. 34 for further discussion). It is worth noting that all bone collagen is “reset” with easily exchangeable hydrogen during the decalcification process. This does not affect our interpretation and conclusions.
We excluded putative diagenetically altered samples by the criteria described in ref. 35. Excluded samples had (one or more) of the following: lower mass fraction H, lower mass fraction O, or higher O/H ratios (Table S8). We also exclude from further analysis one sample from S’Urachi with a highly outlying radiocarbon date (Table S4), and juvenile (non-adult) humans from all sites (Table S7).
Water isotope data
Water isotope data are from the Global Network of Isotopes in Precipitation (GNIP) of the International Atomic Energy Agency37. To approximate rainfall isotope values at each archaeological site, we used the data from two modern reporting sites near each archaeological site: Madrid-Retiro (lat. 40.41, long. −3.68) and Puerto de Navacerrada (40.79, −4.01) for Los Berrocales (Spain); Capo Caccia (40.57, 8.17) and Cagliari-Elmas (39.25, 9.07) for S’Urachi, Monte Sirai, Villamar (Sardinia, Italy); and Har Kna’an (32.97, 35.5) and Bet Dagan (32.00, 34.82) for Megiddo (Israel).
The datasets generated during and analyzed during the current study are available in Supplementary Information and are available from the corresponding author on reasonable request.
Sarno, S. et al. Ancient and recent admixture layers in Sicily and southern Italy trace multiple migration routes along the Mediterranean. Sci. Rep. 7, 1984 (2017).
Broodbank, C. The Making of the Middle Sea: A History of the Mediterranean from the Beginning to the Emergence of the Classical World (Thames & Hudson, 2013).
van Dommelen, P. Colonialism and migration in the ancient Mediterranean. Ann. Rev. Anthropol. 41, 393–409 (2012).
Van Klinken, G. J., Van der Plicht, H. & Hedges, R. E. M. Bone 13C/12C ratios reflect (palaeo-)climatic variations. Geophys. Res. Lett. 21, 445–448 (1994).
Szpak P. Complexities of nitrogen isotope biogeochemistry in plant-soil systems: implications for the study of ancient agricultural and animal management practices. Front. Plant Sci. 5, article 288 (2014).
Hoogewerff, J. A. et al. Bioavailable 87Sr/86Sr in European soils: A baseline for provenancing studies. Sci. Tot. Environ. 672, 1033–1044 (2019).
Gat, J. R. & Carmi, I. Evolution of the isotopic composition of atmospheric waters in the Mediterranean Sea area. J. Geophys. Res. 75, 3039–3048 (1970).
Craig, H. Isotopic variations in meteoric waters. Science 133, 1702–1703 (1961).
Pfahl, S. & Sodemann, H. What controls deuterium excess in global precipitation? Clim. Past 10, 771–781 (2014).
Rubenstein, D. R. et al. Linking breeding and wintering ranges of a migratory songbird using stable isotopes. Science 295, 1062–1065 (2002).
Hobson, K. A., Van Wilgenburg, S. L., Wassenaar, L. I. & Larson, K. Linking hydrogen (δ2H) isotopes in feathers and precipitation: sources of variance and consequences for assignment to isoscapes. PLoS ONE 7, e35137 EP (2012).
Ehleringer, J. R. et al. Hydrogen and oxygen isotope ratios in human hair are related to geography. PNAS 105, 2788–2793 (2008).
Pellegrini, M., Pouncett, J., Jay, M., Pearson, M. P. & Richards, M. P. Tooth enamel oxygen “isoscapes” show a high degree of human mobility in prehistoric Britain. Sci. Rep. 6, 34986 (2016).
Vaiglova, P. et al. Climate stability and societal decline on the margins of the Byzantine empire in the Negev Desert. Sci. Rep. 10, 1512 (2020).
Cormie, A. B., Schwarcz, H. P. & Gray, J. Relation between hydrogen isotopic ratios of bone collagen and rain. Geochim. Cosmochim. Acta 58, 377–391 (1994).
Tuross, N., Warinner, C., Kirsanow, K. & Kester, C. Organic oxygen and hydrogen isotopes in a porcine controlled dietary study. Rapid Commun. Mass Spectrom. 22, 1741–1745 (2008).
Kirsanow, K. & Tuross, N. Oxygen and hydrogen isotopes in rodent tissues: Impact of diet, water and ontogeny. Palaeogeog. Palaeoclimatol. Palaeoecol. 310, 9–16 (2011).
Topalov, K., Schimmelmann, A., Polly, P. D., Sauer, P. E. & Viswanathan, S. Stable isotopes of H, C and N in mice bone collagen as a reflection of isotopically controlled food and water intake. Isotopes Environ. Health Stud. 55, 129–149 (2019).
Kirsanow, K., Makarewicz, C. & Tuross, N. Stable oxygen (δ18O) and hydrogen (δD) isotopes in ovicaprid dentinal collagen record seasonal variation. J. Archaeol. Sci. 35, 3159–3167 (2008).
Hedges, R. E. M., Clement, J. G., Thomas, C. D. L. & O’Connell, T. C. Collagen turnover in the adult femoral mid-shaft: modeled from anthropogenic radiocarbon tracer measurements. Amer. J. Phys. Anthropol. 133, 808–816 (2007).
Koon, H. & Tuross, N. The Dutch whalers: a test of a human migration in the oxygen, carbon and nitrogen isotopes of cortical bone collagen. World Archaeol. 45, 360–372 (2013).
Reynard, L. M., Meltzer, D. J., Emslie, S. D. & Tuross, N. Stable isotopes in yellow-bellied marmot (Marmota flaviventris) fossils reveal environmental stability in the late Quaternary of the Colorado Rocky Mountains. Quatern Res. 83, 345–354 (2015).
Nagy, K. A. & Peterson, C. C. Scaling of water flux rate in animals (University of California Press, 1988).
Bryant, J. D. & Froehlich, P. N. A model of oxygen isotope fractionation in body water of large animals. Geochim. Cosmochim. Acta 59, 4523–4537 (1995).
Ehleringer, J. R. & Dawson, T. E. Water uptake by plants: perspectives from stable isotope composition. Plant Cell Environ. 15, 1073–1082 (1992).
Reynard, L. M. & Hedges, R. E. M. Stable hydrogen isotopes of bone collagen in palaeodietary and palaeoenvironmental reconstruction. J. Archaeol. Sci. 35, 1934–1942 (2008).
Topalov, K., Schimmelmann, A., Polly, P. D., Sauer, P. E. & Lowry, M. Environmental, trophic, and ecological factors influencing bone collagen δ2H. Geochim. Cosmochim. Acta 111, 88–104 (2013).
Bogaard, A. et al. Crop manuring and intensive land management by Europe’s first farmers. PNAS 110, 12589–12594 (2013).
Styring, A. K., Fraser, R. A., Bogaard, A. & Evershed, R. P. The effect of manuring on cereal and pulse amino acid δ15N values. Phytochemistry 102, 40–45 (2014).
Brettell, R., Montgomery, J. & Evans, J. Brewing and stewing: the effect of culturally mediated behaviour on the oxygen isotope composition of ingested fluids and the implications for human provenance studies. J. Anal. At. Spectrom. 27, 778–785 (2012).
Tuross, N., Reynard, L. M., Harvey, E., Coppa, A. & McCormick, M. Human skeletal development and feeding behavior: the impact on oxygen isotopes. Archaeol. Anthropol. Sci. 9, 1–7 (2017).
Hobson, K. A. & Koehler, G. On the use of stable oxygen isotope (δ18O) measurements for tracking avian movements in North America. Ecol. Evol. 5, 799–806 (2015).
von Holstein, I. C. C. et al. Collagen proteins exchange oxygen with demineralisation and gelatinisation reagents and also with atmospheric moisture. Rapid Commun. Mass Spectrom. 32, 523–534 (2018).
Reynard, L. M. & Tuross, N. Hydrogen isotopic analysis with a chromium-packed reactor of organic compounds of relevance to ecological, archaeological, and forensic applications. Rapid Commun. Mass Spectrom. 30, 1857–1864 (2016).
Reynard, L. M., Ryan, S. E. & Tuross, N. The interconversion of δ 2H values of collagen between thermal conversion reactor configurations. Rapid Commun. Mass Spectrom. 33, 678–682 (2019).
Díaz-del-Río, P. et al. Diet and mobility patterns in the late prehistory of central Iberia (4000-1400 cal BC): the evidence of radiogenic (87Sr/86Sr) and stable (δ18O, δ13C) isotope ratios. Archaeol. Anthropol. Sci. 9, 1–14 (2017).
IAEA/WMO. Global Network of Isotopes in Precipitation. The GNIP Database. Accessible at, http://www.iaea.org/water (2019).
Bowen, G. J. & Revenaugh, J. Interpolating the isotopic composition of modern meteoric precipitation. Water Resources Research 39, 1299 (2003).
IAEA/WMO. Global Network of Isotopes in Precipitation. The GNIP Database. Accessible at, http://www.iaea.org/water (2015).
Bowen, G. J. Gridded maps of the isotopic composition of meteoric waters. Accessible at, http://www.waterisotopes.org (2019).
Wessel, P. & Smith, W. H. F. Free software helps map and display data. EOS Trans. AGU 72, 441–448 (1991).
Wessel, P., Smith, W. H. F., Scharroo, R., Luis, W. H. F. & Wobbe, F. Generic Mapping Tools: Improved version released. EOS Trans. AGU 94, 409–410 (2013).
We thank Mike Floyd for assistance with map generation, the American School for Prehistoric Research for funding, and the Max Planck-Harvard Center for the Archaeoscience of the Ancient Mediterranean for a fellowship to SER. We also thank Enrique Baquedano for allowing and assisting sample collection at the Regional Archaeological Museum of Madrid, the Megiddo Expedition for help with samples from Megiddo, and the members of the excavation team at S’Urachi for their assistance, and Richard Waldbauer for sample collection support. We also thank the director and staff at the Soprintendenza Archeologia Belle Arti e Paesaggio per la città metropolitana di Cagliari e le province di Oristano e sud Sardegna for the permission and collaboration to enable us to sample finds from three excavations in Sardinia.
The authors declare no competing interests.
Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
About this article
Cite this article
Reynard, L.M., Ryan, S.E., Guirguis, M. et al. Mediterranean precipitation isoscape preserved in bone collagen δ2H. Sci Rep 10, 8579 (2020). https://doi.org/10.1038/s41598-020-65407-0