Suspected limited mobility of a Middle Pleistocene woman from Southern Italy: strontium isotopes of a human deciduous tooth

We present the Sr isotopic composition of enamel of the most ancient deciduous tooth ever discovered in Italy to assess human mobility in Middle Pleistocene. Reconstructing ancient mobility is crucial for understanding human strategy at exploiting temporally and spatially patchy resources, with most studies focusing on indirect evidences, ultimately affecting our interpretation on hominin territoriality and energetic costs invested by hominin groups. Here, we use the high spatial resolution and micro-destructivity options offered by the Laser Ablation Multi-Collector Inductively Coupled Plasma Mass Spectrometry technique, to determine the 87Sr/ 86Sr intra-tooth variability of a human deciduous incisor from the Middle Pleistocene layers of the Isernia La Pineta site (Italy). We compared these data with the Sr isotopic signature of local micro-mammals, the broadest home-range of the macro-mammals and with modern plant samples. Our study reveals that while macro-mammals have possibly migrated through the landscape for up to 50 km, the pregnant woman from Isernia was probably local, given that the isotopic ratio of the enamel falls within the local range and is comparable with the signature of the local plants in a radius of 10 km. This is the first case study of Sr isotopic composition determination in such ancient deciduous tooth.

During the last decade, strontium isotope ( 87 Sr/ 86 Sr) analyses have been exploited in the study of human past mobility and landscape exploitation because of the special relationship between the Sr isotopic fingerprint stored in human bones and teeth and the living location of the individual (e.g. refs [6][7][8][9][10]. For example, Copeland et al. 9 studied the Sr isotopic composition of permanent enamel in early hominins (~2 million years ago) from South Africa (Sterkfontein and Swartktrans) showing that only about 30% of the analyzed individuals (Paranthropus robustus and Australopithecus africanus) were non-consistent with the local dolomite Sr isotope composition and, therefore, were considered non-local. Their comparison between the proportion of non-local individuals between small (probably female) and large hominins (probably male) shows that at least 50% of the female were non-local in contrast with the 11% of non-local male, suggesting a different landscape use between adult males and females 9 . In contrast, Balter et al. 10 Sr isotope analyses of three different hominin groups (P. robustus, A. africanus and early Homo) from the same region, including also browser and grazers, revealed that the groups show similar home range areas, indistinguishable from the local fauna.
Richard et al. 11 defined a supply region of about 20 km for a Homo neanderthalensis from Greece (Lakonis site). The enamel of this individual presented a quite large range of Sr isotope ratios (on average 0.7107 ± 0.0010; recalculated from the data of Richard et al. ). The more radiogenic values observed in the enamel, compared with lower local isotope ratios, and the great variability within the tooth led the authors to hypothesize that the individual spent part of its childhood far away from the site/region in which it was found, at least 20 km, and that this Neanderthal group practiced residential mobility.
Recently, Willmes et al. 12 reported three Sr isotope ratios of a Homo neanderthalensis enamel (unerupted molar) from Payre (France) ranging from 0.71084 to 0.70902 suggesting a high mobility of the studied individual.
Despite the great potential of the Sr isotope method application to mobility studies, there has been no attempt to use it in Middle Pleistocene specimens. To our knowledge, mobility patterns in Middle Pleistocene hominins have only been inferred by observations of raw material exploitation for lithic artefacts, from butchery testimonies and from environmental data 3 . Overall, these evidences suggest that the presence of a rich and diversified fauna coupled to a less cold environment in central and southern Italy compared to northern regions, may have led to the local adaptation of hominin groups 3 .
In this paper, we analyze the strontium isotopic composition of a deciduous human incisor from the Middle Pleistocene site of Isernia La Pineta (Molise, Italy) through a micro-destructive technique, namely LA-MC-ICP-MS (Laser Ablation Multi-Collector Inductively Coupled Plasma Mass Spectrometry). Given that deciduous teeth form within the uterus of pregnant women, the aim of this work is try to determine the mobility pattern of the mother of the individual, and ultimately the role that (pregnant) women may have played within archaic human groups.

Results
The strontium isotope ratios of the human tooth and animal teeth are summarized in Table 1 and Table S1. Analyses by LA-MC-ICP-MS of the human tooth IS42 (Fig. 1), yielded an average 87 Sr/ 86 Sr ratio of 0.70914 ± 0.00038 (2σ; 5 ablation lines) and a Sr concentration of 87 ± 20 ppm.
Several rodent teeth were analyzed by dissolution MC-ICPMS and their isotopic composition varies between 0.70915 and 0.70938 (average 0.70924 ± 0.00013; 2σ; n = 16), whilst their Sr concentration varies between 166 and 392 ppm. Concerning the macro-fauna, we have analyzed by dissolution MC-ICP-MS seven bison teeth. Their 87 Sr/ 86 Sr ratio ranges between 0.70927 and 0.70954 (average 0.70938 ± 0.00018; 2σ; n = 7), while their Sr concentration between 94 and 649 ppm. Furthermore, we have analyzed nine rhino teeth (two specimens represent different portions of the same tooth) by dissolution MC-ICPMS. These specimens show a 87 Sr/ 86 Sr ratio ranging from 0.70941 and 0.70979 (average 0.70958 ± 0.00030; 2σ; n = 9) and a Sr concentration ranging from 209 to 514 ppm. No significant statistical difference is observed between the 87 Sr/ 86 Sr ratios of the human tooth and the rodent samples (p = 0.7726; two-tailed Mann-Withney), suggesting that the isotopic signal of IS42 falls within the local variability. Contrariwise, both bisons (p = 0.0012; two-tailed Mann-Withney) and rhinos (p < 0.0001; two-tailed Mann-Withney) show a Sr isotope ratio statistically different from the local 87 Sr/ 86 Sr ratio (~0.7096 vs. ~0.7092). The difference between the animal average 87 Sr/ 86 Sr ratio and the local average 87 Sr/ 86 Sr ratio (∆ local ) is 0.00014 for the bison and 0.00034 for the rhino.
The baseline of the bioavailable strontium of the surrounding area has been inferred from modern plant specimens and their isotopic ratios range from 0.7096 to 0.7083 (see Table S2). Plant samples have been divided in two groups based on the distance from Isernia La Pineta (Table S3). The group of samples collected at less than 15 km from the site shows a 87 Sr/ 86 Sr average value of 0.7087 ± 0.0005 (2σ), whilst the group of samples at more than 15 km shows an average value of 0.7088 ± 0.0011 (2σ). One archaeological rodent tooth from Grotta Reali (c.a. 20 km from Isernia) has a Sr isotope ratio of 0.70884. In addition, Sr isotopic data of lavas from Conticelli et al. 13 have been reported as the local value for the area of Roccamonfina at about 35-40 km southward, with the highest radiogenic values of about 0.7100.

Discussion
The average 87 Sr/ 86 Sr ratio obtained for the human tooth IS42 (0.70914) fits within the local home isotopic range as defined by the 2σ and the average of the rodent teeth (0.70924 ± 0.00013). The Isernia La Pineta site ( Fig. 2) is locally dominated by Holocene and Pleistocene fluvial-lacustrine deposits, while the surrounding area (about 30 km away) consists mainly of carbonate and terrigenous deposits formed between the Mesozoic and the Cenozoic 14 . Unfortunately, no isotope data are available in the literature from waters, sediments or carbonate rocks of the area. However, we can indirectly infer them from the seawater 87 Sr/ 86 Sr curve of McArthur et al. 15 . The McArthur curve derives from measurements of 87 Sr/ 86 Sr ratio of biostratigraphically well-dated marine sediments and fossils (e.g. shells, sedimentary rocks, evaporite rocks) and is broadly exploited in dating of marine strata, through the so-called Strontium Isotope Stratigraphy methodology 15 . Therefore, the Sr isotopic signature of sedimentary rocks formed during the time-stretch of geologic interest to the Isernia La Pineta site should lie somewhere between 0.70730 (140 Ma) and modern seawater 0.70917 15 . Animal bones from the Gargano promontory 16 , about 200 km east of Isernia, with outcropping rocks from the Trias to the Miocene, show an 87 Sr/ 86 Sr range from 0.70817 (caprovine from Grotta Scaloria, Neolithic site) to 0.70863 (caprovine from La Torretta/Poggio Imperiale, Neolithic site). These isotopic values fit the Sr isotope ratios predicted by the McArthur curve for the age range of the sedimentary rocks outcropping at the two sites. Therefore, these animals have lived in a substratum with Oligocene-Miocene rocks and/or have ingested food with such a provenance, which means that they lived locally in the Gargano area. One shell from Poggio Imperiale, a locality close to the sea, shows, as expected, a Sr isotope ratio of 0.70909, close to modern seawater 16 . A rodent tooth from Grotta Reali, a Mousterian settlement located about 20 km west of Isernia La Pineta, has an 87 Sr/ 86 Sr ratio of 0.70884. This area is characterized by carbonate (Jurassic-Lower Miocene) and flysch (Upper Miocene) sedimentary sequences 16 with a predicted Sr isotope range between ~0.7070 and 0.7090 15 . Again, the rodent tooth Sr isotopic composition fits within this range and corresponds to a Tortonian age, characteristic of the flysch deposits. The rodents of Grotta Reali and the caprovine of La Torretta/Poggio Imperiale confirm the fact that the local Sr isotopic composition fits the local geological background.
Given the small home range of rodents that reflects the local geology, as also extensively reported in literature (e.g. refs 8, 9, 17), we should expect the Sr isotope ratios of the local geology of Isernia La Pineta inferred from the McArthur curve (and therefore ranging between 0.70730 and modern seawater 0.70917) to reflect those of the rodents. Instead, we found that rodents from Isernia La Pineta have a mean 87 Sr/ 86 Sr ratio of 0.70924 ± 0.00013, slightly more radiogenic than expected. Peretto et al. 14 have shown that pyroclastic deposits are intercalated within the sedimentary layers of the site, although this is not reported in the official geological maps of the area available in the literature. Moreover, several pyroclastic layers (Fig. 2) have also been documented by 18 . Therefore, volcanic material could have contaminated the Sr isotopic composition of local water and soil. The closest volcano to the Isernia site is the extinct Roccamonfina volcano, at about 40 km to the southwest, which was active between 630 and 50 ka. Its lavas have petrologic characteristic consistent with a convergent plate boundary with the low K series lavas showing 87 Sr/ 86 Sr ratios ranging from 0.7064 to 0.7083 and the high K series lavas from 0.7085 to 0.7100 19,20 . The oldest volcanic activity (630 ka) was dominated by leucite-bearing lava flows, with an average 87 Sr/ 86 Sr ratio of 0.70969 ± 0.00059 (2σ; calculated from 13 ). The Roccamonfina lavas are also characterized by very high Sr concentration, ranging from 774 to 1188 ppm in the low K series, from 1348 to 2358 ppm in the high K series and from 1670 to 2190 (average 1875 ± 388 ppm) in the oldest leucite-lavas 13,19,20 . Another nearby volcano showing high 87 Sr/ 86 Sr ratios is the Colli Albano volcano, located at about 150 km to the northeast of Isernia. Its lavas have 87 Sr/ 86 Sr ratios higher than 0.7100 21,22 and Sr concentrations up to ~2200 ppm. The highly radiogenic values of these volcanic rocks may explain the slightly higher ratios observed in the rodent teeth with respect to the local geology. Given the high Sr concentration of the volcanic rocks, a small amount of volcanic material can strongly contribute to the contaminations of the soils and therefore transfer the radiogenic isotopic signature to the local fauna.
The rhinos teeth of Isernia La Pineta have the highest 87 Sr/ 86 Sr ratios and isotopic variability of the analyzed fauna (2σ = 0.00030). Their relative Sr concentration (1/Sr) shows a strong negative correlation with their Sr isotope ratios (n = 9; r 2 = 0.64; p = 0.001), indicating that the highest-Sr reservoir has also the more radiogenic 87 Sr/ 86 Sr ratios (Fig. S1). The radiogenic Sr isotopic composition of the rhinoceros can be explained by the intake of water and food "contaminated" with the 87 Sr/ 86 Sr volcanic signature of the nearby Roccamonfina volcano. Recently, Lugli et al. 23 detected cyclic variations in the enamel Sr isotopic composition of two rhinos teeth of Isernia La Pineta by in situ LA-ICP-MS high resolution investigations, with values ranging from 0.70937 to 0.70997. These variations have been correlated to the tooth growth features, suggesting that large mammals may have seasonally migrated southward in search of food, being attracted by the more fertile ground offered by the volcanic rocks of Roccamonfina 23 . Although farther away, even the Colli Albani area could have been reached by mammals during seasonal migration, as showed by the high 87 Sr/ 86 Sr ratios and Sr concentration of the lavas 22 .
Bisons have lower Sr isotope ratios and smaller variability than rhinos but their values are still higher than the human and rodents and are certainly also correlated to a larger or different area of supply that most likely reached the volcanic grounds.
The Sr isotopic values of the Isernia human tooth do not reflect a strong contamination by the volcanic radiogenic isotope signature of Roccamonfina (or Colli Albani).
This suggests either a smaller supply area for the Isernia hominin compared to the large mammals of the site and/or a supply area not contaminated by volcanic deposits. More detailed inferences can be deduced from the isotope data of the local modern plants as a proxy for the bioavailable strontium. Considering together the local rodents (c.a. 1-2 km) and the within-15 km bioavailable strontium variability, the isotopic values observed in the human tooth fit within this range. The distant (>15 km) plants Sr isotope ratios and the Roccamonfina lava data are highly variable, and this variability seems not reflected in the human tooth (Fig. 3). The interpretations of the Sr homogeneous isotopic composition signal observed in this human tooth can be diverse. First, given that Central-Southern Italy is characterized by relatively young homogeneous limestone sediments 24 , even large human movements can be isotopically invisible if occurring on the same type of rock. Second, our knowledge on the Sr isotope composition of deciduous teeth in general is still far from being completely understood. In fact, it is not clear whether enamel tissue retains specific time-resolved information or if it is homogeneous because of the buffering of the Sr reservoirs of the mother bone (see Supplementary Information; refs 25, 26). Third, the low variability of the isotopic signal can be simply the result of a very limited supply area. The limits of the LA technique and the type of samples do not allow us to unequivocally choose between one of the hypotheses. Nevertheless, given that our data show macro-mammals moving through the landscape back and forth from marine sedimentary soils to volcanic soils, our preferred explanation for the local Sr isotopic composition of the human deciduous tooth of Isernia lies with a limited supply area and a suspected limited mobility of a pregnant woman in the Middle Pleistocene because of group strategy.
Human occupation at the Isernia La Pineta site occurred at the transition between the Marine Isotope Stage 15 (MIS15) and MIS14 14 . According to several climate and orbital proxies, since 700 kyr (mid-Pleistocene transition) a shift occurred from 41 kyr climatic cycles to strong 100 kyr cycles. The increase in amplitude in glacial-interglacial cycles caused an increase in the severity of glaciations. The first of these large-amplitude deglacial episodes is Termination VII (Marine Isotope Stage16-15 boundary), when global climate transitioned from the first very cold, extended glacial episode (MIS 16), to a relatively warm interglacial (MIS 15). In many sedimentary records, MIS 14 stands out as a short and much warmer glacial period than other glacial epochs of the last 800 kyr. It has been proposed that the extra-long duration of interglacial/mild stadial climates during MIS 15−13 may have provided favorable conditions over 100 kyr for the dispersal of African hominins into the Eurasia region during the Middle Pleistocene, leading to sustainable occupation 27 . In this climatic context, the human group of Isernia La Pineta, as most of the contemporary ones, seems to be not very mobile, probably exploiting local resources, a strategy likely similar to the modern collectors 3,5 .
The suspected limited mobility suggested by strontium isotopes is in agreement with technological choices and hunting strategies 28 , which also attest for an evolutionary innovation in terms of mental templates. In fact, it seems that technical criteria employed in small débitage are not opportunistic and unstructured as previously thought, but reflect a well-established production method 14,[29][30][31] . Moreover, raw materials were collected by Middle Pleistocene Homo heidelbergensis mainly in fairly narrow areas and, seemingly, no large displacements were made for this purpose 2,28,30,32 .
In terms of group strategy, the likely limited mobility of the pregnant women/mothers could also have had a huge impact on their role within the society and the gendered division of labor 33 . In recent hunter-gatherer communities, the general form in which the labor is subdivided can be summarized as men hunt and search for exotic resources, while women and children gather and exploit local resources. Moreover, women generally stay at camp to care for children, seldom following the men during long travels 33 . However, it seems that the gendered division of labor evolutionary appears with Upper Palaeolithic humans and was not exploited, for example, by Neanderthals 33 . Thus, the suspected reduced mobility of the Middle Pleistocene pregnant woman of Isernia could be in agreement with the division of labor observed in modern hunter-gatherer groups, a behavior risen perhaps earlier than we thought in human evolution.
The work of Wall-Scheffler and Myers 34 also suggests that pregnant females of hominin groups might have changed the travel speed of the entire group and its mobility pattern. In particular, the persistent loads during pregnancy and child carrying slow the optimal walking speed of women and raise the energetic cost to walk with the group. Murray et al. 35 state that Pan troglodytes pregnant females spent less time traveling than other females, reducing their physical activity, in order to store energy and fat for the forthcoming lactation. This strategy, as the authors say, is common in both humans and nonhuman primates, which can accumulate considerable fat reserves.
Despite the limited possibility of comparison with other human data of the same period, the tooth from Isernia shows the lowest Sr isotope internal variability yet presented in the literature in ancient hominins. Few studies on other species suggest that early hominins 9 and Neanderthals 11, 12 were characterized by larger mobility patterns, likely related to an adaptation strategy similar to modern foragers.

Conclusion
In this paper, we present the first direct Sr isotope composition of a deciduous Middle Pleistocene (c.a. 570 ka) hominin tooth enamel. The 87 Sr/ 86 Sr ratios of this tooth are indistinguishable from the Sr isotope ratios of local rodents, possibly suggesting a local living location for the mother of the Isernia child. A limited mobility pattern hypothesis, similar to the logistical mobility of the modern hunter-gatherer collectors, is cautiously put forward, even if the data presented here are definitely not unambiguous. We cannot exclude mobility over geologically similar sedimentary rocks along the central portion of the Apennine chain. However, the comparison with modern plant specimens reveals the likely exploitation of local resources. Contrariwise, macro-fauna, namely rhinoceros and bisons, shows more radiogenic Sr isotope ratios, probably due to contamination with volcanic soils and likely related to seasonal migrations in the south-west direction, toward the Roccamonfina volcano area.

Materials and Methods
Laser Ablation MC-ICP-MS. The 87 Sr/ 86 Sr ratio of the IS42 human tooth was quantified by LA-MC-ICPMS technique on enamel micro-sampling carried out on the external surface of the tooth. We sampled 5 portions (linear transects) of the external enamel surface on the lingual side, following a cervical-incisal direction. Sr data were acquired with a double focusing MC-ICPMS Neptune ™ (Thermo Fisher Scientific), coupled to a 213 nm Nd:YAG laser ablation system (New Wave Research ™ ), housed at the Centro Interdipartimentale Grandi Strumenti (CIGS) of the University of Modena and Reggio Emilia, following the protocol described in Lugli et al. 23 . Laser parameters used during the analysis are those reported in Lugli et al. 23 . In order to maximize the ion beam size stability, we chose to use dynamic linear ablation (500 × 100 µm), which provides more precise results compared to static spot ablation 36 . To detect all masses required for the analysis ( 88 Sr, 87 Sr, 86.5, 86 Sr, 85 Rb, 85.5, 84 Sr, 83 Kr, 82 Kr), nine Faraday detectors were employed. A 60 s background signal for each mass was collected before the analysis to correct for Kr interferences. 83 Kr background signal was always lower than 1 mV. Signals on mass 88 Sr of the IS42 sample ranged from 1.1 to 1.5 V. Instrument was tuned for low oxide production, constantly monitoring the 87 Sr/ 86 Sr ratio of reference materials. Moreover, before analysis, the 83 Kr/ 82 Kr ratio in the background was kept approximately equal to 1. Laser data were acquired in single blocks of 200 cycles, with an integration time of 0.5 s. To correct for the 40 Ca 13 P 16 O formation we used our set of in-house bioapatite standards 23,37 . The analysis of our human tooth in-house reference material (ROCht42) yielded an average (n = 10) LA 87 Sr/ 86 Sr ratio of 0.70865 ± 0.00040 (2σ) and an accuracy of 0.005% (16 ppm) compared with the previously obtained dissolution MC-ICP-MS value of 0.70864 ± 0.00001 (2se). Bison schoetensacki (n = 7) and Stephanorhinus hundsheimensis (n = 9), were collected from the same archaeological level of the human tooth and from the two adjacent layers (immediately above and below; Table S1).
Modern plant specimens (n = 13) were collected in the area surrounding Isernia La Pineta, in a radius of about 30 km. Each specimen represents a pool of different plants (arboreal and herbaceous), collected in natural areas away from roads and cultivated fields, reflecting the average bioavailable strontium of the area. Before Sr extraction, plant samples were ashed at 500 °C 9 . One modern snail was also sampled from one of the modern plant site (Scapoli, see Table S2). Moreover, one archaeological rodent tooth from Grotta Reali (c.a. 20 km far from Isernia) was analysed.
Animal enamel samples, modern plant specimens and snail were analyzed after chemical dissolution and chromatographic separation of Sr. Samples preparation follows the protocol presented in Lugli et al. 23 . 5-10 mg of enamel and about 15 mg of plant ashes were digested in 1 ml of 14 N HNO 3 and re-dissolved in 3 N HNO 3 after evaporation. The Sr separation uses columns with a 300 µl volume filled with Eichrom Sr spec-resin. After several washes with MilliQ ® water and the conditioning of the resin with 3 N HNO 3 , samples were loaded into the columns and washed repeatedly with 3 N HNO 3 . Strontium was eluted with several reservoirs of MilliQ ® water. The whole procedure was conducted in a clean lab, with a Sr blank typically lower than 100 pg, at the Department of Chemical and Geological Sciences of the University of Modena and Reggio Emilia. Sr isotope ratios were determined by means of a MC-ICP-MS Neptune, housed at the CIGS (UNIMORE). A [blank/standard/blank/sample/blank] bracketing sequence was adopted to monitor any instrumental drift. Data were normalized through exponential law to a 88 Sr/ 86 Sr ratio of 8.375209 23 . During this work, the standard yielded a 87 Sr/ 86 Sr ratio of 0.71025 ± 0.00002 (2σ; n = 33). The Sr ratios were corrected for instrumental bias to the NBS-987 value of 0.71026 ± 0.00002 reported by 38 .