Tracking natural and anthropogenic Pb exposure to its geological source

Human Pb exposure comes from two sources: (i) natural uptake through ingestion of soils and typified by populations that predate mining activity and (ii) anthropogenic exposure caused by the exposure to Pb derived from ore deposits. Currently, the measured concentration of Pb within a sample is used to discriminate between these two exposure routes, with the upper limit for natural exposure in skeletal studies given as 0.5 or 0.7 mg/kg in enamel and 0.5/0.7 μg/dL in blood. This threshold approach to categorising Pb exposure does not distinguish between the geological origins of the exposure types. However, Pb isotopes potentially provide a more definitive means of discriminating between sources. Whereas Pb from soil displays a crustal average 238U/204Pb (μ) value of c 9.7, Pb from ore displays a much wider range of evolution pathways. These characteristics are transferred into tooth enamel, making it possible to characterize human Pb exposure in terms of the primary source of ingested Pb and to relate mining activity to geotectonic domains. We surmise that this ability to discriminate between silicate and sulphide Pb exposure will lead to a better understanding of the evolution of early human mining activity and development of exposure models through the Anthropocene.

forming events are generally related to major geological mountain building processes of which three dominate European geology 13 ; the Alpine event of c. 60-2.5 Ma which is most evident in circum-Mediterranean geology; the Hercynian c. 280-380 Ma, which mostly affects northern continental Europe and southern Britain and the Caledonian event of c. 390-490 Ma seen in the Palaeozoic and older rocks of Britain and Scandinavia. The calculation of Pb model age gives an estimate of the age and hence geological episode to which mineralisation is associated. Mu (μ) provides evidence of the geochemical nature of the source rock of the mineralization. For example, deposits such as those in Tunisia 14 , source their Pb from uranium rich granite domains and hence have elevated 238 U/ 204 Pb (μ) values, whereas Pb derived from more basic/ultrabasic deposits, such as are found in Cyprus, reflect the low uranium nature of the host with low 238 U/ 204 Pb (μ) values 14 . The combination of the model age (T) and 238 U/ 204 Pb (μ) thus provide geological, and hence geographic, constraints on the origin of the Pb without recourse to large reference datasets.
In this study, the transfer of labile soil Pb into fauna is primarily demonstrated using Neolithic (pre-anthropogenic Pb) pigs teeth. Pigs ingest soil while grubbing for food and hence provide a simple transfer model. The animals are from the Neolithic feasting site of Durrington Walls in southern England. These data are supplemented by two human 'natural exposure' populations: (i) a dataset of Neolithic individuals from British archaeological sites, and (ii) 10 th century individuals, all typified by very low Pb concentration levels (0.11 ± 0.18 mg/kg, 2 SD, n = 34) 15,16 . These are then compared with data from three Early (5-7 th century) Anglo Saxon and Anglian sites in England, where elevated Pb concentrations are suggestive of anthropogenic Pb exposure. The sites, and the average Pb concentrations in the tooth enamel, are as follows: Berinsfield 17 in central England, where individuals have average tooth enamel Pb concentrations of 2.5 ± 4 mg/kg (2 SD, n = 11); Eastbourne in southern England, where individuals have average tooth enamel Pb concentrations of 6 mg/ kg ± 22 mg/kg (2 SD, n = 21) and West Heslerton, north eastern England, which straddles the natural/anthropogenic Pb exposure boundary (0.7 ± 2.8 mg/kg; 2 SD, n = 33) 18 .

Method Section
Tooth enamel samples were prepared as follows: The enamel surface of the tooth was abraded from the surface to a depth of >100 microns using a diamond coated dental bur and the removed material discarded. An enamel sample was cut from the tooth using a flexible diamond edged rotary dental saw. All surfaces were mechanically cleaned with a diamond bur to remove adhering dentine. The resulting sample was transferred to a clean (class 100, laminar flow) working area for further preparation. In a clean laboratory, the sample was cleaned ultrasonically in high purity water to remove dust. It was then rinsed twice in de-ionized water, and soaked for an hour at 60 °C, before rinsing again and then leaching for 5 minutes with Teflon distilled 0.2 M HCl,. After a final rinse, the sample was dried and transferred into a pre-cleaned Teflon beaker where it was dissoslved in Teflon distilled 8MHNO3, evaporated to dryness and converted to bromide form using Romil© UpA HBr. Soils were leached with deionised water for 24 hours, centrifuged and the supernatant decanted into clean Savillex Beakers, evaporated to dryness and converted to bromide form as before. Separation of Pb from samples was undertaken using standard ion exchange techniques. The data in this paper has been acquired over a number of years and includes lead isotope compositions that were determined by either thermal ionisation mass spectrometry (TIMS) using a Finnigan Mat 262, or multi-collector inductively coupled plasma mass spectrometry (MC-ICP-MS) using a Nu Plasma HR or a Thermo Fisher Scientific Neptune Plus. TIMS Pb was run using rhenium filaments in a silica gel-phosphoric acid. Lead blanks were c 70 pg. Lead isotope ratios were normalised to values of NBS 981 19 , which gave the following reproducibility during the period of analysis: 206 Pb/ 204 Pb = 0.20%, 207 Pb/ 204 Pb = 0.29%, 208 Pb/ 204 Pb = 0.40% (2σ, n = 31). Samples analysed by MC-ICP-MS were spiked with a thallium (Tl) solution and introduced into the instrument via an ESI 50 μl/min PFA micro-concentric nebuliser attached to a de-solvating unit (Nu Instruments DSN 100 or Cetac Aridus II) and normalised to NBS981 20 . Average 2 SD reproducibility for the following ratios is 206 Pb/ 204 Pb = 0.008%; 207 Pb/ 204 Pb = 0.008%; 208 Pb/ 204 Pb = 0.009%. The pig enamel samples, which were analytically challenging due to low Pb yields, were run on the Neptune using a high sensitivity jet cone and reproducibility was 206 Pb/ 204 Pb = 0.027% %; 207 Pb/ 204 Pb = 0.031%; 208 Pb/ 204 Pb = 0.041%. Lead concentrations, where documented, were measured by either isotope dilution 8 or solution plasma 21 . Details of the samples and sites and extended methodology are supplied in the supplementary information section. Data is presented in Table 1.
Bio-available Pb from modern British soils defines a broadly horizontal field of data with 238 U/ 204 Pb = 9.74 ± 0.18 (2 SD, n = 29) Fig. 1. The bio-available Pb from ancient soils, as represented by archaeological bone and dentine composition, give a comparable result of 238 U/ 204 Pb = 9.70 ± 0.08 (2 SD, n = 34). Both these results are in agreement with the average crust composition of 238 U/ 204 Pb = 9.7 22 and synonymous with recycled sedimentary rocks. Some samples give negative model ages which is common in samples from limestone terrains and caused by a disproportionate uptake of U compared to Pb in marine carbonates 23 .
The transfer of the bio-available soil Pb into fauna is shown in Fig. 2. Data from the Neolithic pigs tooth enamel range between model ages (T) of 209 and 471 Ma with 238 U/ 204 Pb = 9.78 ± 0.05 (2 SD, n = 23). Human tooth enamel data yields a similar 238 U/ 204 Pb = 9.73 ± 0.06 (2 SD, n = 56). The coincidence of the soil and faunal data fields provides firm evidence that natural Pb exposure is consistent with the ingestion of the bioavailable component of Pb in silicate based soil. Figure 3 shows the Pb isotope composition of tooth enamel from 5-7 th century individuals whose elevated Pb concentrations is taken as evidence of anthropogenic Pb exposure. The figure includes data from galena (PbS) in British deposits of the Mendips, Pennines and central Wales for comparison. The most obvious aspect of the diagram is the steeply sloping data fields created by both the tooth data and the galena compositions. The central Wales data best highlights the highly correlated nature of the ore composition, created during the process of mineralization. Similar arrays can be seen in many galena datasets 14  Continued exposure of these individuals was dominated by British ore. However, six of the Eastbourne samples, and most of the Berinsfield data, extend beyond the range of the British deposits suggesting that some individuals carry a component of non-British Pb. The Pb is locked into tooth enamel during mineralization, which for the M2 teeth of this study, occurs between two and eight years age 24 . There are a number of options for the Pb source and ingestion route. The main route for modern children's Pb exposure is though hand to mouth soil ingestion 25 . However non-local Pb isotope signatures can arise from a number of routes: (1) The individual was exposed to Pb somewhere other than where they were found ie they are not of local origin (2) They were exposed to Pb from a non-local source 1   The data from the modern soils was produced by leaching modern soil samples with deionised water. This modern data ( ) is compared with the Pb isotope composition of bone and dentine from archaeological sites. The assumption made is that the bone and dentine re-equilibrated with the labile soil component close to the time of burial and thus provide a measure of labile Pb that predates modern pollutants ( ).  inherited a non-local Pb composition from their mother via placental 26 or lacational 27 transfer that was available or re-mobilized during tooth mineralization.

, and (3) They
Strontium and oxygen isotope analysis has also been undertaken on these samples but does not support a non-British childhood for the majority of individuals from the Berinsfield and Eastbourne sites 17,28 and so we rule out an immigrant population.
Thus the most likely exposure routes would appear to be either from imported goods or, that these are first generation arrivals whose mothers carried and transferred a Pb isotope signature from her homeland 29 ; or it may be a combination of both. Grave goods from Berinsfield highlight continental connections 17 .
Some constraints can be placed on the geological origin of the Pb these people were exposed to: the Berinsfield array indicates an end-member of a geologically young, low-U Pb terrain, whereas the Eastbourne upper end-member indicates a U-rich terrain that is c. 600 Ma old. Lead and Pb-bearing silver deposits with isotope compositions similar to those seen in the Eastbourne and Berinsfield populations can be found in Europe 14 and hence this signature could have been introduced to England either by early Anglo-Saxon groups arriving in England or through trade and exchange of coins, ornaments or weaponry with continental populations.
This study shows that naturally derived, bio-available Pb from ingested soil is characterized by a horizontal data array in 238 U/ 204 Pb-T space, which is mimicked by fauna exposed to this type of Pb. In contrast, sulphide ore deposits define steeply dipping data arrays, a trend that is also reflected in the tooth enamel of people who have been exposed to anthropogenic Pb. This difference in the orientations of the fields can thus be used to distinguish between natural and anthropogenic exposure.
It is proposed that this approach to characterizing the origin of human Pb exposure provides an alternative method to examining Pb sources, regardless of exposure levels, and allows new insights into the rise of mining during the Anthropocene 30 , development of metal working and trade in the ancient world and its impact on human health.