Journal of Archaeological Science 35 (2008) 763e772 http://www.elsevier.

com/locate/jas

Stable carbon and nitrogen isotope analysis on human remains from the Early Mesolithic site of La Vergne (Charente-Maritime, France)
Rick J. Schulting a,*, Stella M. Blockley b, Herve Bocherens c,d, e Drucker d,e, Mike Richards f,g Dorothe
School of Archaeology, University of Oxford, OX1 2PG, UK Department of Geography, Royal Holloway University of London, Egham, Surrey, TW20 0EX, UK c Institut des Sciences de lEvolution, UMR 5554, Universite Montpellier 2, France d Institut fur Ur- und Fruhgeschichte und Archaologie des Mittelalters, Universitat Tubingen, Tubingen, Germany e Archeologies Environnementales, ArScAn, UMR 7041, MAE Rene Ginouves, Nanterre, France ` f Department of Human Evolution, Max Planck Institute for Evolutionary Anthropology, Leipzig, Germany g Department of Archaeology, Durham University, South Road, Durham, DH1 3LE, UK
b a

Received 29 January 2007; received in revised form 21 May 2007; accepted 19 June 2007

Abstract We report here the results of stable carbon and nitrogen isotope analysis of human and faunal remains from La Vergne (Charente-Maritime, western France), a rare Early Mesolithic burial site (ca. 8500e8000 cal BC). The results for nine humans (average d13C 19.3&; d15N 9.4&) indicate a strongly terrestrial diet, dominated by animal protein, with the possibility of, at best, a slight contribution of marinederived protein. Given lower sea-levels in the early Holocene, the site would have been some 60e80 km from the sea at the time of its use; nevertheless, contacts with the coast are shown by the presence of numerous marine shell beads in the graves. In the light of the stable isotope results, it is suggested here that such contacts most likely took the form of exchange with coastal communities whose remains now lie underwater. 2007 Elsevier Ltd. All rights reserved.
Keywords: Early Mesolithic; France; Stable carbon and nitrogen isotopes; Palaeodiet; Territoriality

1. Introduction Human remains dating to the Early Mesolithic are rare in many parts of Europe, and are important in that they provide information concerning a period of transition, both environmentaldfrom the Preboreal to the Borealdand culturaldfrom the Epipalaeolithic to the Mesolithic. La Vergne falls within this timeframe, being an Early Mesolithic burial site (ca. 8500e8000 cal BC), located in Charente-Maritime in western France (Courtaud and Duday, 1995; Courtaud et al., 1999;

* Corresponding author. Tel.: 44 (0)1865 278246; fax: 44 (0)1865 278254. E-mail address: rick.schulting@arch.ox.ac.uk (R.J. Schulting). 0305-4403/$ - see front matter 2007 Elsevier Ltd. All rights reserved. doi:10.1016/j.jas.2007.06.008

Duday and Courtaud, 1998). Moreover, what is of particular interest at La Vergne is the presence of abundant marine shell beads, indicating some form of contact with the coast, which at this time would have been many tens of kilometres distant. This raises the question of whether the individuals buried at La Vergne were highly mobile, systematically making use of a very large territory that included the coast, or, alternatively, whether the shells were obtained through trade, or through occasional forays to the coast for primarily non-subsistence purposes. The presence of a number of individuals at the site provides the opportunity to undertake a palaeodietary study using stable carbon and nitrogen isotope analysis to address this question. In addition, fauna from La Vergne together with previously published faunal values from a near-contemporary site, enable an investigation of aspects of the terrestrial economy,

764

R.J. Schulting et al. / Journal of Archaeological Science 35 (2008) 763e772

and in particular the balance between plant and animal sources of protein, as well as the degree of variability in isotopic values within this group. The measurements provided here add to the growing body of isotopic data available for prehistoric human diets, and contribute towards a greater understanding of their temporal and spatial variability. 1.1. The site of La Vergne The site of La Vergne is located near the town of SaintJean-dAngely, Charente-Maritime, on a slope of the valley of the river Boutonne (Fig. 1; the site itself is also known by ` the name of La Grande-Piece, but we use the more common place-name as it appears in the titles of the primary publications). The site was discovered during an archaeological salvage operation in 1995; the presence of Mesolithic graves was completely unexpected, as these underlay a Gallo-Roman site, which was responsible for much disturbance, and possible destruction of part of the earlier site (Courtaud and Duday, 1995; Courtaud et al., 1999). Three surviving grave structures contained the remains of 10 individuals, including adults of both sexes and young children, with the very fragmentary and partial remains of an additional ve individuals likely deriving from two disturbed graves of the same period. Grave 10 contained the massive horn cores of two aurochs interred with the deceased (Courtaud et al., 1999, Photo 3). This grave and another contained large numbers of marine shells, many perforated for use as ornaments, as were a number of fox, deer and human teeth. 2. Materials and methods Stable carbon and nitrogen analysis was undertaken on bone collagen extracted from 13 human and ve faunal bone samples from La Vergne, in order to provide information concerning palaeodiet at the site. The human samples were analysed at two different laboratories, Bradford and Montpellier. The faunal samples were analysed at Bradford only; unfortunately very limited faunal samples are available from the site, as no associated settlement has as yet been found, and the graves themselves contained only a few species (red deer and aurochs) as part of the funerary rite. The majority of the samples are associated with three complex graves, directly dated on human bone to early in the Holocene, with three AMS estimates ranging from 9215 to 9070 BP, calibrating to the second half of the ninth millennium BC (Table 1). These three determinations derive from bone in graves with multiple individuals, and so provide good dated contexts for seven of the sampled individuals. Samples from two individuals are from disturbed contexts (lv4 and lv5), but are tentatively suggested to be of comparable age, due in part to close physical proximity to the dated Mesolithic graves, and in part to the presence of red ochre staining on at least some bones from each context (Courtaud, personal communication). The remaining four individuals are thought to considerably post-date the Mesolithic interments, and may belong to the Iron Age (lv3, lv7, lv8, lv13).

2.1. Stable isotope analysis Stable isotope analysis has been successfully applied to both human and faunal remains from an increasing number of Palaeolithic and Mesolithic sites across Europe (e.g., Bo cherens et al., 1997, 2006, 2007; Drucker and Celerier, 2001; Drucker and Henry-Gambier, 2005; Lillie and Richards, 2000; Richards et al., 2000; Schulting and Richards, 2001). Stable carbon (d13C) and nitrogen (d15N) in bone collagen can yield information about protein sources in both human and non-human animal past populations, providing a direct record of long term diet, over a period of some 5e10 years in adult humans (Schwarcz and Schoeninger, 1991). In the absence of C4 plants, which do not feature in the diet of earlier prehistoric humans in northwest Europe, d13C values reect the level of marine input into the diet, with endpoints of ca. 21 to 20& for purely terrestrial protein to ca. 12& for purely marine protein. The d15N values provide information on the trophic level at which an organism is operating: human and non-human consumer values are elevated by ca. 3e5& over the values of their diets (Bocherens and Drucker, 2003; DeNiro and Epstein, 1981; Schoeninger and DeNiro, 1984). More comprehensive local values for the major herbivores and carnivores would ideally be available to aid in the interpretation of the human data; unfortunately few data are available. Of relevance, however, are faunal values from the early Holocene component at Pont dAmbon, Dordogne, especially the material from layer 2, dated to the Preboreal Azilian period (Drucker, 2001; Drucker and Celerier, 2001). 2.2. Methods 1, Bradford Bone samples of 200e300 mg in weight were cleaned using an air abrasion system. These were demineralised at 4  C in 0.5 M HCl. Once demineralised, the samples were rinsed three times with de-ionised water. They were then introduced to a pH 3 solution and placed in a heater block at 70  C for 48 h. The samples were rst ltered in a 8 mm lter and then ultraltered, with the >30 kDa fraction taken to be frozen and freeze-dried (Brown et al., 1988). Between 0.3 and 0.5 mg of the resulting collagen was weighed into tin cups for analysis in a ThermoFinnigan Flash EA coupled to a Delta Plus XL mass spectrometer. A number of blank samples and internal standards, including three randomly assigned methionine samples of known weight (0.6e0.7 mg) were introduced with the samples to ensure instrument integrity (see Richards and Hedges, 1999). Standard measurement errors are 0.2& for d13C and 0.3& for d15N. 2.3. Methods 2, Montpellier Bone samples of 200e300 mg in weight were cleaned and ground to a powder sieved through a 0.7 mm mesh. These were demineralised at room temperature in 1 M HCl for 20 min, ltered through a 5 mm Millipore lter and the insoluble residue was soaked in a 0.125 M NaOH solution for 20 h at room temperature to remove humic components. After

R.J. Schulting et al. / Journal of Archaeological Science 35 (2008) 763e772

765

le d'Yeu

Vende

Vende Autize Svre Niortaise

Deux-Svres le de R La Grange Charente-Maritime le d'Olron

Boutonne

La Vergne
Charente

La Pierre-Saint-Louis
Seudre
20 10

Charente
Seugne

-1 00 m

e nd ro Gi

-2

m 0 -1 m 0

-3

-5 0 m

5 m

Gironde

N
0 20 40 km

Fig. 1. Map of Charente Maritime showing the location of La Vergne. The 35 m contour marks the approximate coastline at ca. 9000 BP (ca. 8300 cal BC). The actual coastline would likely be more complex than shown here. (Isobath contours from Admirality Charts 2663 and 2664.)

ltration through a 5 mm Millipore lter, the insoluble residue was heated in a pH 2 HCl solution at 100  C for 17 h, and the insoluble fraction was removed after a second ltration. The solubilized fraction containing gelatin was then freeze-dried and judged to correspond to collagen if the chemical characteristics (%C, %N, C/N) were within the range of those of collagen extracted from fresh bone using the same protocol. Around 0.2e0.3 mg of this product was weighed into tin cups for analysis in a Eurovector EA coupled to a VG-Optima mass spectrometer. A number of blank samples and internal standards, including puried alanine samples of known weight (0.2e0.3 mg) were introduced with the samples to ensure instrument integrity. The standard errors of the isotopic measurements are 0.1& and 0.2& for d13C and d15N values, respectively. 3. Results and discussion The results of the analysis on the humans and fauna are presented in Tables 2e4, and Fig. 2. For samples analyzed in both

laboratories, average d13C and d15N values were calculated for each sample when both measurements yielded C:N values within the acceptable range of 2.9 to 3.6 (DeNiro, 1985). For the Montpellier dataset, one human sample (lv4) was rejected due to its C:N ratio of 2.4 falling well outside the accepted range; therefore only the analysis performed in Bradford was retained for this specimen. C:N ratios of 2.8 were obtained for two individuals at Bradford (lv1, lv7), again outside of the accepted range; therefore, the collagen isotopic values measured in Montpellier on the same samples, with suitable C:N values, were retained. While the results from

Table 1 AMS determinations on human bone from La Vergne Grave 7 10 3 Lab. no. OxA-6699/LY-369 OxA-6700/LY-370 OxA-6698/LY-368 Date BP 9070 9215 9075 70 65 65 cal BC 8535 8605 8535 7990 8295 8005

Source: Duday and Courtaud (1998, 37). Re-calibrated using CALIB 5.0.2.

766

R.J. Schulting et al. / Journal of Archaeological Science 35 (2008) 763e772 Mesolithic, 9075 65 BP Mesolithic, w9075 BP Mesolithic, w9075 BP Mesolithic, 9070 70 BP Mesolithic, w9070 BP Mesolithic, 9215 65 BP Mesolithic, w9215 BP Mesolithic? Disturbed Mesolithic? Disturbed Upper part; likely post-Meso Disturbed; post-Mesolithic? Disturbed; post-Mesolithic? Disturbed; post-Mesolithic?

the two laboratories compare reasonably well, a number of minor discrepancies are evident at the level of specic individuals, particularly in the d13C values. The reason for this difference is not known, and warrants further investigation. However, due to the use of international standards in both laboratories, it is unlikely that the differences are due to differences in the isotope measurements themselves, but instead to the differing pretreatments, such as the additional stage of ultraltration at Bradford. Other factors may also be involved: the different isotope values seen in the infant (lv6), for example, could be due to the young age of this individual (ca. 2 years). Indeed, since two different parts of the same bone were analyzed in Bradford and Montpellier, and the turnover rate varies within a given bone, an individual with changing isotopic values in their diet may record different proportions of one diet in one part of its bone than in another part of the same bone. A child experiences changing isotopic values of its diet due to breastfeeding and weaning (e.g., Fogel et al., 1989; Fuller et al., 2003; Katzenberg et al., 1993; Schurr, 1998), changes that are recorded differently in different parts of a given bone depending on the turnover rate (e.g., Balasse et al., 1997; Herrscher and Bocherens, 2000). The ve faunal samples were measured at Bradford only. Unfortunately they show poor preservation: two failed to yield collagen altogether, while the C:N ratio of another (2.7) fell outside the acceptable range, though its d13C and d15N values are not inconsistent with the remaining two herbivore samples; it is nevertheless excluded from the summary. Ideally both the terrestrial and marine endpoints would be determined locally from contemporary fauna: in the absence of marine fauna at La Vergne, or indeed fauna from any Early Holocene coastal sites in this region (as they have been inundated by sea-level rise), this was possible only for terrestrial fauna, which yielded expected values (Table 5). The terrestrial faunal samples are augmented by the inclusion of isotopic data from the Azilian (Preboreal) site of Pont dAmbon, Dordogne (Drucker, 2001; Drucker and Celerier, 2001). Combining the faunal data, a terrestrial end-point of ca. 21& can be suggested. The human stable isotope values from La Vergne clearly represent a predominantly terrestrial diet with some individuals possibly suggesting a minor marine or estuarine component (discussed further below). The d15N signal is consistent with an emphasis on the consumption of terrestrial animal rather than plant protein. The average human d15N value of 9.4& is over 5& above the average of 4.2& for the combined faunal sample, and so is at the very top end of the suggested range of 3e5& for a trophic level shift (it is recognised that the faunal average reects the proportional representation of the species sampled and not necessarily of those hunted, but in the absence of further data from these and other species, this seems an appropriate heuristic value). As a diet of 100% terrestrial mammal protein seems improbable, this strongly suggests a certain proportion of even higher trophic level foods in the diets of some individuals, whether marine/freshwater sh or birds, or possibly marine/freshwater shellsh, some of which can show elevated d15N values. The value for the omnivorous wild boar is higher than the herbivore average by ca.

Attribution d15N

Combined

9.3 9.8 9.1 7.5 8.0 9.0 10.6 11.7 9.9 8.9 19.1 19.0 19.4 19.9 19.2 19.9 19.3 18.5 19.3 19.6 3.1 3.2 3.1 3.0 3.1 3.2 3.2 2.4 3.2 3.1 14.5 12.7 13.5 14.8 11.7 13.4 13.4 8.7 15.7 14.4 38.3 34.7 36.5 38.4 31.5 36.4 37.0 17.6 42.6 37.8 21.9 25.3 25.7 10.9 19.7 22.6 31.1 14.2 35.4 22.8

C:N

mg/g

30.7 26.0 26.6 9 lv3 lv8 lv13 F5 St.1 St.12 e 9 Tibia, L. Fibula Tibia, R. ? ? ? Adult Adult Adult 3.1 3.5 3.0 19.4 20.6 19.2 9.6 10.0 9.7 34.6 28.7 28.9 13.8 10.9 11.9 2.9 3.1 2.8 19.5 20.1 19.5 9.0 9.7 9.4 0.8 0.6 0.7

41.2 38.8 37.3 Values in italics fall outside the accepted range for good collagen preservation and are excluded from all summary calculations and graphs.

N (%)

%C

d15N

Montpellier

C:N

d15N

% Collagen

Bradford

Table 2 Human remains at la Vergne: stable isotope results

Sex

Context

Sample

lv10 lv2 lv9 lv12 lv1 lv11 lv6 lv4 lv5 lv7

St.3 St.3 St.3 St.7 St.7 St.10 St.10 St.4 St.11 St.10

63 127 585 2039 2014 668 292 e 7 152

No.

Femur, L. Cranium Tibia, R. Ulna, L. Femur, L. Tibia, R. Tibia, L. Metacarpal Fibula Ulna, L.

Element

? M? F F? F? M ? ? M ?

Adult Adult Adolescent Adult Adult Adult Infant Adult Adult Adult

Age

2.4 2.6 2.7 2.7 2.2 2.5 5.2 1.2 2.4 3.0

19.1 19.0 19.3 19.6 19.0 19.5 19.4 18.5 18.7 19.4

d13 C

9.0 9.4 8.7 7.7 7.8 8.9 9.9 11.7 10.1 9.3

33.1 34.0 29.8 32.2 10.7 31.5 35.4 33.4 32.6 31.7

%C

12.3 12.8 11.4 12.4 4.1 12.2 13.9 12.9 12.6 13.4

%N

3.1 3.1 3.0 3.0 3.1 3.0 3.0 3.0 3.0 2.8

19.1 19.1 19.5 20.2 19.4 20.2 19.2 19.3 19.9 19.6

d13 C

9.6 10.2 9.6 7.4 8.1 9.0 11.3 9.4 9.7 8.9

0.7 0.7 0.7 0.7 0.7 0.8 1.0 0.4 1.2 0.6

15.4 15.7 14.1

%N

3.1 2.9 3.1

19.5 20.4 19.5

d13 C

9.3 9.9 9.4

R.J. Schulting et al. / Journal of Archaeological Science 35 (2008) 763e772 Table 3 Faunal remains at la Vergne: stable isotope results Sample lv20 lv21 lv23 lv24 lv22 Context St.10 St.7 St.3 St.10 F5 Species Bos primigenius Cervus elaphus Bos primigenius Bos primigenius Unidentied faunal Element Astragalus Antler Astragalus Cranium Unidentied fragment % Collagen d13 C d15N %C %N

767

C:N

No collagen extracted No collagen extracted 1.8 19.2 1.1 21.0 1.4 20.6

4.8 4.1 5.2

11.9 26.4 32.9

5.1 10.1 12.3

2.7 3.1 3.1

Values in italics fall outside the accepted range for good collagen preservation are excluded from all summary calculations and graphs.

2& (Table 6), and if this species featured heavily in the diet the human values would thus also be raised. At the same time, it is important to note that some plant foods may also exhibit elevated nitrogen values. Schoeninger (1995, 98), for example, reports a value of 12.1& for modern cattail, Typha latifolia, an edible species used by Native Americans in the Great Basin, but also occurring in Europe. The possibility that this plant, or similar species, formed part of the diet of Mesolithic humans urgently requires further investigation. Nevertheless, on the currently available evidence, the most plausible scenario is that the individuals at La Vergne obtained the majority of their protein from terrestrial animal sources, though this does still leave some scope for contributions from aquatic and plant resources. The importance of large game is also indicated for one of the few Mesolithic faunal assemblages from Charente-Maritime, a small fragmentary collection from La Grange, dominated by aurochs and red and roe deer, with some boar also present. This is undoubtedly a biased sample, however, with the absence of smaller species, including birds and sh, being due to poor preservation (Laporte et al., 2000). The human values cluster reasonably tightly; the exceptions to this being individuals lv8 and lv4 (Fig. 2). Sample lv4 is tentatively considered to be Mesolithic based on archaeological data, and its high d13C and d15N values can be interpreted as possibly reecting a greater marine contribution than seen in the other Mesolithic individuals. Samples lv8, together with lv3, 7 and 13, are considered to post-date the Mesolithic interments, and may belong to a Gallo-Roman occupation on the site (Courtaud and Duday, 1995; Duday and Courtaud, 1998). Specimen lv8 shows a higher than expected d15N value in relation to its d13C value. In the absence of clear chronological information and associated fauna, it is difcult to decipher the effects of possible changes in the isotopic signatures of

terrestrial food items (e.g. Richards and Hedges, 2003) and the effects of possible changes in the proportion of different food items in the diet. Otherwise, the suspected later interments do not differ substantially from the Mesolithic burials in either dataset (Table 5), though this need not necessarily imply that their diets are equivalent, as the baseline oral and faunal values may change through time. The presumed postMesolithic samples are not discussed further here. Two samples are from disturbed contexts (lv4, lv5), but, as noted above, are thought to be Mesolithic and so are included in the following discussion. Additional AMS dates are in process to conrm this (Courtaud, personal communication). One of the key questions posed by the small Mesolithic cemetery at La Vergne is whether any consumption of marine protein is indicated. This issue arises because of the quantity of marine shells found in the graves, though in the form of pierced ornaments rather than consumption debris. It is clear from the above discussion that the overall emphasis isotopically is on terrestrial resources, and so any contribution of marine foods would be minor at best, making its detection challenging, particularly in the absence of a wider range of contemporary, local faunal values. Nevertheless, there may be some slight evidence for a contribution of marine protein in the diets of some individuals. First, the average d13C value is signicantly higher for the humans than for the fauna. Second, in the absence of C4 systems, one of the strongest lines of evidence for determining whether marine protein features in a dataset is the extent to which d13C and d15N values are positively correlated (cf. Richards and Hedges, 1999; Kelly, 2000). A linear regression analysis applied to the eight known and suspected Mesolithic adults (the infant is excluded as subject to a nursing effect, discussed below) shows a signicant positive correlation (r2 0.60, F 9.11, p 0.023) (Fig. 3). This suggests that a minor contribution of marine-derived

Table 4 Faunal remains at Pont dAmbon, layer 2, Early Preboreal age (after Drucker, 2001) Sample PAM200 PAM300 PAM400 PAM500 PAM900 PAM600 PAM700 PAM800 PAM1000 Identication PdA K5-110 c.2 PdA D8-26 c.2 PdA J5-131 c.2 PdA J5-105 c.2 PdA G5-0 c.2 PdA H6-17 c.2 Species Cervus elaphus Cervus elaphus Cervus elaphus Cervus elaphus Capreolus capreolus Equus ferus Equus ferus Equus ferus Sus scrofa Element Humerus Phalanx Radius Tibia Phalanx Phalanx Phalanx Scapula Phalanx N (%) 0.7 1.6 1.4 0.8 1.1 1.1 1.1 0.8 0.9 Yield (mg/g) 10.3 79.9 39.7 20.4 24.9 47.7 37.6 28.8 37.3 N yield 22.8 78.9 43.4 38.4 30.4 67.7 53.7 54.6 50.4 d13 C 20.3 23.4 20.2 20.5 21.0 21.6 20.6 20.7 20.1 d15N 3.9 4.0 3.1 4.5 3.4 4.1 4.3 3.4 6.3 %C 39.2 42.8 41.5 41.5 37.3 41.6 42.7 41.1 35.3 %N 14.4 15.4 15.2 15.3 13.4 15.0 15.5 15.0 12.7 C:N 3.2 3.2 3.2 3.2 3.2 3.2 3.2 3.2 3.2

PdA H8-27 c.2

768
Mesolithic ?Mesolithic

R.J. Schulting et al. / Journal of Archaeological Science 35 (2008) 763e772

fauna post-Mesolithic

14

12 lv8

lv4

10

2 -22

-21
13

-20

-19

-18

C value

Fig. 2. Plot of stable carbon and nitrogen isotope results for humans at La Vergne. The faunal data combines two samples from La Vergne and nine from Early Holocene Pont dAmbon.

protein may feature in the diet of some individuals at La Vergne, though, given the small sample size and some concerns over the dating of all the samples, it must be emphasised that even this conclusion remains tentative and subject to revision in the light of further AMS dating. Aside from the small sample size, a potential confounding factor with such borderline cases is that a slight positive correlation between d13C and d15N values can be expected in a purely terrestrial system, if the two main isotopically distinct protein sources are plants and herbivores. This is because there is a slight trophic level shift of ca. 1& in 13C (Bocherens and Drucker, 2003); when combined with the well-known trophic level shift in d15N between plants and herbivores, this can lead to a signicant positive correlation even with what are on the whole relatively depleted d13C values. The human-faunal difference in d13C observed at La Vergne is on average 1.6&, which is arguably higher than would be expected if the explanation was a trophic level enrichment. Furthermore, other inland Mesolithic sites in western Europe do not seem to show this effect, and a more typical result is to nd no signicant correlation between d13C and d15N values (Table 6). Finally, it should be emphasised here that a marine signal need not imply the consumption of shellsh or sh directly from the sea, but could also derive from anadromous sh (salmon) or
Table 5 Summary of human and faunal stable isotope results d13 C Mesolithic average Male average Female average Post-Mesolithic average Herbivores (La Vergne) Herbivores (Pont dAmbon) Omnivore (Pont dAmbon) Fauna (combined) 19.3 19.4 19.5 19.7 20.8 21.0 20.1 20.9 0.41 0.39 0.31 0.43 0.29 1.05 0.93 d15N 9.4 9.5 8.2 9.4 4.6 3.8 6.3 4.2 1.27 0.49 0.83 0.42 0.76 0.48 0.90 n 9 3 3 4 2 8 1 11

Pont dAmbon data from Drucker (2001).

from migratory birds that spend part of the year feeding offshore or along the coast but also move inland for part of the year. The rivers, lakes and estuaries of Charente Maritime today provide rich habitats for migratory birds, though the situation at 9000 BP would of course have been quite different, particularly given the distance of La Vergne from the coast. Nevertheless, the Boutonne river valley may still have provided good fowling, as well as shing, opportunities. The probable importance of waterways is demonstrated by the distribution of known Mesolithic int scatters in the region, which tend to occur along the rivers (Marchand, 2007). Owing to the fragmentary nature of the remains, a number of individuals could not be assigned a sex (Courtaud and Duday, 1995; Duday and Courtaud, 1998). While this does not leave a large sample, it is worth noting that there does appear to be a tendency for females to exhibit somewhat lower d15N values (Table 5, Fig. 3), suggesting a greater contribution of lower trophic level terrestrial foods, possibly including more plant foods. This difference fails to reach statistical signicance at the 0.05 level, and so is very tenuous, though we feel it is worthy of further exploration. A similar subtle difference in male and female diets, with the latter apparently consuming relatively less marine foods, has been suggested on the basis of a larger isotopic dataset for the coastal Late Meso lithic sites of Teviec and Hoedic, Brittany (Schulting and Richards, 2001). A very tentative suggestion for females consuming slightly lower trophic level foods has also been made for the Early Mesolithic burial site of Avelines Hole, southwest England (Schulting, 2005a). With the exception of Teviec and Hoedic, these cases are far from convincing on their own, but taken together they do at least suggest a potential pattern, and we mention it for this reason. Gender-based differences in diet are not uncommon in the ethnographic literature on hunter-gatherers, and often comprise differential access to animal meat and fat (Kelly, 1995, 23, 166). However, an important alternative possibility to consider is the effect of pregnancy and breastfeeding on d15N values. Data recently presented by Fuller et al. (2004) suggests that d15N values for pregnant and breastfeeding females can be 0.5e1.0& lower compared to other females (and presumably males). A comparison of adults and subadults is not possible, as there is only one infant (age ca. 2 years) in the sample. The results for this individual (lv6) are slightly elevated in d15N relative to d13C, as would be expected in an infant subject to the nursing effect (Fogel et al., 1989; Fuller et al., 2003; Fuller et al., 2006; Katzenberg et al., 1993; Schurr, 1998), though its position is masked to some extent by adult sample lv4, which shows the most elevated values for both isotopes (and is responsible in no small part for the positive correlation between the two noted above). There are essentially no comparative isotopic data for other humans of this age in France. An Early Mesolithic secondary burial was found at La Chaussee-Tirancourt, Somme, dating to 9020 100 BP (8530e7830 cal BC, Gif-A92523) (Ducrocq et al., 1996) and a Middle Mesolithic burial, dating to 8350 105 BP (7530e7070 cal BC, Ly-5606) is known from Auneau, Eure-et-Loir (Verjux, 1999, 2000; Verjux and

15

N value

R.J. Schulting et al. / Journal of Archaeological Science 35 (2008) 763e772 Table 6 Correlation between d13C and d15N human bone collagen values for selected inland western European Mesolithic sites Site La Vergne (Early Meso) Meuse Basin (Early Meso) Ofnet (Late Meso) Avelines Hole (Early Meso) d13C 19.3 20.5 19.6 19.9 0.40 0.40 0.18 0.59 d15N 9.4 9.6 10.8 6.6 1.26 0.50 0.33 0.93 n 8 26 20 18 r2 0.60 0.01 0.07 0.00 F 9.11 0.33 1.26 0.02 p 0.02 0.57 0.28 0.90 Source

769

This paper Bocherens et al., 2007 Bocherens et al., 1997 Schulting, 2005a

Data exclude children age less than 7, as potentially subject to the nursing effect.

Dubois, 1996). La Chaussee-Tirancourt is located at a similar distance from the modern coastline as La Vergne, while Auneau is further inland. But in neither case have associated stable isotopic analyses been reported. Further to the north, a large series of Early Mesolithic human remains (n 30) from the Meuse Basin of Belgium have been analysed isotopically, and are consistent with a fully terrestrial foodweb, though this may have included some freshwater sh (Bocherens et al., 2007, 18) (see Table 6). A useful comparison can also be made with the Late Meso lithic coastal cemeteries of Teviec and Hoedic, Brittany (Pe quart and Pequart, 1954), as these serve quart et al., 1937; Pe as a good example of the range of isotopic values that might be expected from individuals living on the Atlantic coast of western France prior to the appearance of farming. Palaeodietary studies have been undertaken on the surviving human skeletal material from these sites (Schulting and Richards, 2001). More recent work has revised some of the earlier results, and in particular has rejected the latest dates of ca. 5500e5000 BP (Schulting, 2005b). The stable isotope results have also been signicantly modied through re-analysis, but many of the previously reported general conclusions still stand. Fig. 4 uses the most recent bone collagen values from Teviec and

Hoedic, which are currently being prepared for publication (Schulting and Richards, n.d.). Teviec and Hoedic can be seen to be distinctly different from La Vergne, showing much higher d13C and d15N values. The somewhat unusual position of sample lv4 from La Vergne, with its elevated d15N value relative to its d13C value, is also seen more clearly here. While not representing a purely ma rine signal, the individuals from Teviec and Hoedic show a consistent and signicant marine component, contributing as much as 70e80% of the dietary protein for some individuals. By contrast, even the most marine individuals at La Vergne do not indicate more than 5e10% marine dietary protein at most, and it may be less than this, since early Holocene d13C values tend to be slightly depleted (although there is no evidence for this in the very limited fauna measured here) (Drucker and Celerier, 2001; Hedges et al., 2004; Richards et al., 2000; Schulting, 2005a,b). Such a difference between the sites is not unexpected. Even today La Vergne is some 50 km from the coast, and the dating of the burials to the early Holocene (ca. 9200e9000 BP or ca. 8300 cal BC) means that sea levels would have been lower by 30e45 m (Pirazzoli, 1991, Plate 22, curves A, C and D), placing the site from 60e80 km from its contemporary coastline (Fig. 1). During the much later period represented by Teviec and Hoedic (ca. 6800e6400 BP or ca. 5500e5000 cal BC), the coastline would have been within a few kilometres of these sites.

?M/M 13 12 11 10

?F/F

.Indet

r2adults= 0.60

Hodic 16 15 14 13

Tviec

La Vergne

N value

9 8

15

N value

12 11 10 9 8

lv4

7 6 5 4 -20.5

-20

-19.5
13

-19

-18.5

-18 7 6 -21 -20 -19 -18 -17

13

C value

Fig. 3. Human stable C and N isotope values for known and suspected Mesolithic individuals at La Vergne. The positive correlation (adults only) between d13C and d15N values could suggest a small input of marine protein in the diet of some individuals. Males appear to be slightly elevated in d15N relative to females.

15

-16

-15

-14

-13

-12

C value

Fig. 4. Comparison of human bone collagen stable isotope values from La Vergne (Charente-Maritime) and Teviec and Hoedic (Brittany).

770

R.J. Schulting et al. / Journal of Archaeological Science 35 (2008) 763e772

What is intriguing, then, are the numerous marine shells presentdsome 3300 in totaldin the graves at La Vergne (Fig. 5) (Courtaud and Duday, 1995; Courtaud et al., 1999; Dupont, 2003). Hinia recticulata is the most abundant, numbering over 2000 in one grave alone; also present are Dentalium sp., Semicassis saburon and Cardium norvegium (Dupont, 2003). With the exception of the latter, the shells are pierced for ornamentation, and many are coloured with red ochre (this was abundant in the graves, and the shells may have become stained by this rather than intentionally). The shells clearly indicate contact with the coast, either through direct or indirect trade with people living nearer to the coast, or that the communities represented by the burials at La Vergne themselves travelled to the coast to collect the shells. Both of these possibilitiesdwhich of course need not be mutually exclusivedhave interesting implications. If the former, then it implies the existence of separate groups living nearer the coast, and possibly exploiting marine resources more systematically. The sites, and burial places, of any such groups would now be under water. If the latter, then a high degree of mobility is indicated for at least some individuals (this need not imply residential moves by the entire group). Whether this took place at regular intervals or very intermittently cannot be said; neither pattern would be surprising for hunter-gatherers (Binford, 1982, 1991; Kelly, 1983, 1995). However, it is clear that, even if people were visiting the coast, they were not systematically exploiting marine resources for subsistence. In fact, even the marine shells used for ornamentation appear to have been collected when already dead (Dupont, 2003). On balance, given the distances involved, it might be suggested that the more likely explanation is that distinct communities were indeed present closer to the coast, and that the groups represented by the burials at La Vergne had contacts with these communities, acquiring from them marine shells for ornamentation, and perhaps also, on the rare occasion, marine foods. This implies that Mesolithic populations had an unexpected degree of subsistence and settlement specialisation, and perhaps accompanying notions of territoriality, from very early in the Holocene (and of course possibly even before this). The alternative, that no groups were systematically exploiting the coastline for subsistence resources, seems improbable, particularly in the light of the accumulating archaeological and isotopic evidence from other parts of western Europe, showing considerable use of marine resources from at least the early eighth millennium BC onwards (e.g., Fischer, 1997; Nordqvist, 1995; Schulting and Richards, 2002). An interesting question then becomes what distance from the sea marked the division between groups following primarily coastal and those following primarily inland exploitation patterns in western France, but it is not possible to address this further with the data to hand. In the absence of additional burial populations, investigation of this issue, however, can be taken forward through the analysis of lithic technologies and raw material distribution, though of course this runs against the problem of the absent assemblages from the drowned coastal plain. The present view is that Early Mesolithic lithic technology was relatively

Fig. 5. View of adult in Grave 10 showing nest of Cardium shells and aurochs bucrania. (Photograph courtesy of Patrice Courtaud.)

homogeneous over wide areas, but that raw materials were generally sourced more locally (Marchand, 2007). This could imply wide-ranging contacts, but at the same time the use of relatively restricted habitual territories.

4. Conclusions The stable isotope analyses undertaken indicates that the Mesolithic population at La Vergne essentially followed an inland way of life dominated by terrestrial resources, and in particular by animal protein, though varying contributions from freshwater sh, aquatic birds and plant foods are probable. Additional faunal values from a greater variety of species, including smaller game and freshwater plant and animal resources, are needed to further rene the interpretation of the human values. Dramatic evidence for the importance of large game, both symbolically and presumably also in terms of subsistence, is seen in the placement of the massive horn cores of two slaughtered aurochs in grave 10. There is at most slight evidence for a contribution from marine-derived protein, though it is important to acknowledge that even this could take the form of migratory sh or birds, and need not reect direct forays to the distant coast. Nevertheless, some kind of contact with the coast is indicated by the abundant marine shells, mainly in the form of beads, found in the graves. It is most likely that this contact took the form of trade with groups living on or at least nearer the coast. Such groupsdthe burial sites of which have yet to be founddmay very well have exploited marine resources for subsistence to a much greater extent. The implications of this are of more strongly differentiated regional economic and settlement adaptations than are often attributed to this early period. The tentative suggestion of a slight difference in male and female diets is intriguing, with females obtaining more of their protein from lower trophic levels, but unfortunately the very small sample size precludes the formation of stronger conclusions in this regard.

R.J. Schulting et al. / Journal of Archaeological Science 35 (2008) 763e772

771

Acknowledgements The authors would like to thank Patrice Courtaud and Henri Duday for providing the samples, and for information on the site. We also appreciate the comments and suggestions made by three anonymous reviewers. The research presented here was supported in part by a grant to RJS from the Natural Environment Research Council (NER/B/S/2000/01420). References
Balasse, M., Bocherens, H., Tresset, A., Mariotti, A., Vigne, J.-D., 1997. ` Emergence de la production laitiere au Neolithique? Contribution de lana lyse isotopique dossements de bovins archeologiques. Sciences de la Terre ` et des Planetes 325, 1000e1010. Binford, L.R., 1982. The archaeology of place. Journal of Anthropological Archaeology 1, 1e31. Binford, L.R., 1991. When the going gets tough, the tough get going: Nunamiut local groups, camping patterns and economic organization. In: Gamble, C.S., Boismier, W.A. (Eds.), Ethnoarchaeological Approaches to Mobile Campsites. International Monographs in Prehistory, Ann Arbor, pp. 25e137. Bocherens, H., Drucker, D., 2003. Trophic level isotopic enrichment of carbon and nitrogen in bone collagen: case studies from recent and ancient terrestrial ecosystems. International Journal of Osteoarchaeology 13, 46e53. Bocherens, H., Drucker, D.G., Billiou, D., Genestec, J.-M., van der Plicht, J., ` 2006. Bears and humans in Chauvet Cave (Vallon-Pont-dArc, Ardeche, France): insights from stable isotopes and radiocarbon dating of bone collagen. Journal of Human Evolution 50, 370e376. Bocherens, H., Grupe, G., Mariotti, A., Turban-Just, S., 1997. Molecular preservation and isotopy of Mesolithic human nds from the Ofnet Cave (Bavaria, Germany). Anthropologische Anzieger 55, 121e129. Bocherens, H., Polet, C., Toussaint, M., 2007. Palaeodiet of Mesolithic and Neolithic populations of Meuse Basin (Belgium): evidence from stable isotopes. Journal of Archaeological Science 34, 10e27. Brown, T.A., Nelson, D.E., Southon, J.R., 1988. Improved collagen extraction by modied Longin method. Radiocarbon 30, 171e177. ` Courtaud, P., Duday, H., 1995. Decouverte dune necropole Mesolithique a la Vergne (Charente-Maritime). Bulletins et Memoires de la Societe dAnthropologie de Paris 7, 181e184. Courtaud, P., Duday, H., Martin, H., Robin, K., 1999. La necropole Mesolithi que de La Vergne (Charente-Maratime, France). In: Bintz, P., Thevenin, A. (Eds.), LEurope des derniers chasseurs. Epipaleolithique et Mesolithique. Comite des Travaux Historiques et Scientiques, Paris, pp. 287e292. DeNiro, M.J., 1985. Post-mortem preservation and alteration of in vivo bone collagen isotope ratios in relation to palaeodietary reconstruction. Nature 317, 806e809. DeNiro, M.J., Epstein, S., 1981. Inuence of diet on the distribution of nitrogen isotopes in animals. Geochimica et Cosmochimica Acta 45, 341e351. Drucker, D., 2001. Validation methologique de lanalyse isotopique dosse ments fossiles et apports aux reconstitutions paleoecologiques du Paleoli thique superieur du sud-ouest de la France. Unpublished PhD thesis, Universite Paris 6. ` Drucker, D., Celerier, G., 2001. Teneurs en carbone-13 du collagene de grands ` mammiferes du site de Pont dAmbon (Dordogne, France): implications pour lenvironment et son exploitation au Tardiglaciaire dans le sud-ouest de la France. Paleo 13, 145e158. Drucker, D.G., Henry-Gambier, D., 2005. Determination of the dietary habits ` of a Magdalenian woman from Saint-Germain-la-Riviere in southwestern France using stable isotopes. Journal of Human Evolution 49, 19e35. Ducrocq, T., le Goff, I., Valentin, F., 1996. La sepulture secondaire Mesolithi que de la Chaussee-Tirancourt (Somme). Bulletin de la Societe Prehistorique Francaise 93, 211e216. Duday, H., Courtaud, P., 1998. Le necropole Mesolithique de la Vergne (Char ` ente-Maritime). In: Guilaine, J. (Ed.), Sepultures dOccident et geneses des ` Megalithismes (9000e3500 avant notre ere). Errance, Paris, pp. 27e37.

Dupont, C., 2003. La malacofaune de sites mesolithiques et neolithiques de la ` ` facade atlantique: contribution a leconomie et a lidentite culturelle des groupes concernes. University of Paris I-Pantheon-Sorbonne. Fischer, A., 1997. Marinarkologiske Forundersgelser Forud for Etablering af en Fast resundsforbindelse. Milj- og Energiministeriet, Skov- og Naturstyrelsen, Copenhagen. Fogel, M.L., Tuross, N., Owsley, D.W., 1989. Nitrogen isotope tracers of human lactation in modern and archeological populations. Annual Report of the Director. Geophysical Laboratory, Carnegie Institute of Washington. 89, 111e117. Fuller, B.T., Richards, M.P., Mays, S.A., 2003. Stable carbon and nitrogen isotope variations in tooth dentine serial sections from Wharram Percy. Journal of Archaeological Science 30, 1673e1684. Fuller, B.T., Fuller, J.L., Sage, N.E., Harris, D.A., OConnell, T.C., Hedges, R.E.M., 2004. Nitrogen balance and d15N: why youre not what you eat during pregnancy. Rapid Communications in Mass Spectrometry 18, 2889e2896. Fuller, B.T., Fuller, J.L., Harris, D.A., Hedges, R.E.M., 2006. Detection of breastfeeding and weaning in modern human infants with carbon and nitrogen stable isotope ratios. American Journal of Physical Anthropology 129, 279e293. Hedges, R.E.M., Stevens, R.E., Richards, M.P., 2004. Bone as a stable isotope archive for local climatic information. Quaternary Science Review 23, 959e965. Herrscher, E., Bocherens, H., 2000. Dietary behaviour in past population and its evolution over the life span. Methodological aspects of isotopic biogeochemistry (13C, 15N. European Journal of Clinical Nutrition 54, S10. Katzenberg, M.A., Saunders, S.R., Fitzgerald, W.R., 1993. Age differences in stable carbon and nitrogen isotope ratios in a population of prehistoric maize horticulturists. American Journal of Physical Anthropology 90, 267e281. Kelly, J.F., 2000. Stable isotopes of carbon and nitrogen in the study of avian and mammalian trophic ecology. Canadian Journal of Zoology 78, 1e27. Kelly, R.L., 1983. Hunter-gatherer mobility strategies. Journal of Anthropological Research 39, 277e306. Kelly, R.L., 1995. The Foraging Spectrum: Diversity in Hunter-Gatherer Lifeways. Smithonsonian Institution Press, Washington, DC. Laporte, L., Marchand, G., Sellami, F., Oberlin, C., Bridault, A., 2000. Les occupations mesolithiques et du Neolithique ancien sur le site de la Grange ` ` a Surgeres (Charente-Maritime). Revue Archeologique de lOuest 17, 101e142. Lillie, M.C., Richards, M., 2000. Stable isotope analysis and dental evidence of diet at the Mesolithic-Neolithic transition in Ukraine. Journal of Archaeological Science 27, 965e972. Marchand, G. Les derniers chasseurs prehistoriques des profondes forets dAu nis et de Saintonge, in: Glenisson, J. (Ed.), Histoire de lAunis et de la ` Saintonge, La Prehistoire et lAntiquite, Vol. I. Geste editions, La Creche (in press). Nordqvist, B., 1995. The Mesolithic settlements of the west coast of Sweden e with special emphasis on chronology and topography of coastal settlements. In: Fischer, A. (Ed.), Man and Sea in the Mesolithic. Oxbow Books, Oxford, pp. 185e196. ` Pequart, M., Pequart, S.-J., 1954. Hoedic, Deuxieme Station-Necropole du Mesolithique Cotier Armoricain. De Sikkel, Anvers. Pequart, M., Pequart, S.-J., Boule, M., Vallois, H., 1937. Teviec, Station-Ne cropole du Mesolithique du Morbihan. Archives de LInstitut de Paleontologie Humaine XVIII, Paris. Pirazzoli, P.A., 1991. World Atlas of Holocene Sea-Level Changes. Elsevier, Amsterdam. Richards, M.P., Hedges, R.E.M., 1999. Stable isotope evidence for similarities in the types of marine foods used by Late Mesolithic humans on the Atlantic coast of Europe. Journal of Archaeological Science 26, 717e722. Richards, M.P., Hedges, R.E.M., 2003. Bone collagen d13C and d15N values of fauna from Northwest Europe reect palaeoclimatic variation over the last 40,000 years. Palaeogeography, Palaeoclimatology. Palaeoecology 193, 261e267. Richards, M.P., Hedges, R.E.M., Jacobi, R., Current, A., Stringer, C., 2000. Goughs Cave and Sun Hole Cave human stable isotope values indicate

772

R.J. Schulting et al. / Journal of Archaeological Science 35 (2008) 763e772 of Teviec and Hoedic. Journal of Anthropological Archaeology 20, 314e344. Schulting, R.J., Richards, M.P., 2002. Finding the coastal Mesolithic in southwest Britain: AMS dates and stable isotope results on human remains from Caldey Island, Pembrokeshire, South Wales. Antiquity 76, 1011e1025. Schurr, M.R., 1998. Using stable nitrogen isotope ratios to study weaning behavior in past populations. World Archaeology 30, 327e342. Schwarcz, H.P., Schoeninger, M.J., 1991. Stable isotope analyses in human nutritional ecology. Yearbook of Physical Anthropology 34, 283e321. ` Verjux, C., 1999. Chronologie des rites funeraires Mesolithiques a Auneau (Eure-et-Loir, France). In: Bintz, P., Thevenin, A. (Eds.), LEurope des der niers chasseurs. Epipaleolithique et Mesolithique. Comite des Travaux Historiques et Scientiques, Paris, pp. 293e302. Verjux, C., 2000. Les fosses Mesolithiques dAuneau (Eure-et-Loir - France). In: Crotti, P. (Ed.), Epipaleolithique et Mesolithique, Actes de la Table Ronde. Cahiers dArcheologie Romande, Lausanne, pp. 129e138. Verjux, C., Dubois, J.-P., 1996. Une sepulture Mesolithique en position assise ` sur le site du Parc du Chateau a Auneau (Eure-et-Loir). Revue Archeologique de Centre de la France 35, 83e96.

a high animal protein diet in the British Upper Palaeolithic. Journal of Archaeological Science 27, 1e3. Schoeninger, M.J., 1995. Dietary reconstruction in the prehistoric Carson Desert: stable carbon and nitrogen isotopic analysis. In: Larsen, C.L., Kelly, R.L. (Eds.), Bioarchaeology of the Stillwater Marsh. American Museum of Natural History, New York, pp. 96e106. Schoeninger, M.J., DeNiro, M.J., 1984. Nitrogen and carbon isotopic composition of bone collagen from marine and terrestrial animals. Geochimica et Cosmochimica Acta 48, 625e639. Schulting, R.J., 2005a. . pursuing a rabbit in Burrington Combe: new research on the Early Mesolithic burial cave of Avelines Hole. Proceedings of the University of Bristol Spelaeological Society 23, 171e265. Schulting, R.J., 2005b. Comme la mer qui se retire: les changements dans lex ploitation des ressources marines du Mesolithique au Neolithique en Bre tagne. In: Marchand, G., Tresset, A. (Eds.), Unite et diversite des processus ` de neolithisation sur la facade atlantique de lEurope (7e4eme millenaires avant J.-C.). Memoire de la Societe Prehistorique Francaise 36, Paris, pp. 163e171. Schulting, R.J., Richards, M.P., 2001. Dating women and becoming farmers: new palaeodietary and AMS data from the Breton Mesolithic cemeteries