123

12 Prehistoric shell middens along the French Atlantic façade

12 Prehistoric shell middens along the French Atlantic façade: the use of marine and terrestrial resources in the diets of coastal human populations Catherine Dupont, Rick Schulting and Anne Tresset

Introduction The current study of French shellmiddens is linked to a number of recent and ongoing multidisciplinary projects, funded by different French institutions (CNRS, Ministry of Research and Sustainable Development, Ministry of Culture). They include: ‘Beginnings of herding on the coastline and small islands of north-western Europe’ (Action Thématique et Incitative sur Programme, directed by A. Tresset); ‘The neolithisation of the European Atlantic façade: cultural interactions, technical transfers and interactions with the natural environment’ (Action Concertée Incitative, directed by G. Marchand); ‘Archaeological survey of the archipelago of Molène, Brittany’ (directed by Y. Pailler); ‘The Mesolithic in Brittany’ (Programme Collectif de Recherches, directed by G. Marchand); and ‘Prehistoric population on the islands and the coastline of the Charente’ (Programme Collectif de Recherches, directed by L. Laporte). Their purpose is on the one hand to combine the raw data from the archaeological sites with surveys and quantitative data on bones and shells, and on the other to interpret our studies within the broader context of the European Atlantic façade. These programmes bring together researchers working on different subjects, including material culture (and in particular lithics), subsistence, and the settlements and territoriality of Mesolithic and Neolithic groups. As in other European countries along the Atlantic façade, shellmidden research begins with the Mesolithic period. Indeed, at the European

scale, the definition of the Mesolithic period by archaeologists was from the outset greatly influenced by the seeming appearance of shellmiddens following the Palaeolithic but preceding the Neolithic (Zvelebil and Rowley-Conwy 1986). Since the first half of the 20th century, the shellmiddens of Téviec and Hoëdic (Morbihan) have, due to their association with elaborate burials (Péquart et al. 1937; Péquart and Péquart 1954), overshadowed other Mesolithic shellmiddens of the French Atlantic façade, such as Beg-anDorchenn (Finistère), Saint-Gildas (LoireAtlantique) (Giot 1947; Bellancourt 1980; Kayser 1985; 1990a), and Beg-er-Vil (Morbihan) (Kayser and Bernier 1988; Kayser 1990b). These other sites are less well-known, and admittedly are also less well-preserved, but nevertheless they contribute to our knowledge of the pattern of coastal exploitation at this time. The same can be said for shellmiddens of more recent (post-Neolithic) date which have been neglected. For the past five years or so, research linked to the study of the processes of neolithisation have resulted in both the revisiting of older excavations on Mesolithic shellmiddens, and on comparisons with coastal Neolithic sites (Tresset 2000; 2002; 2003; Dupont 2003; Dupont and Gruet 2005). Unfortunately, the faunal remains of later prehistoric and historic shellmiddens are rarely studied, and so we do not discuss them here. We begin with an inventory of features of the geographical distribution of prehistoric shellmiddens, discussed in terms of the sedimentary

124

Catherine Dupont, Rick Schulting and Anne Tresset history and the geology distinctive to the study area. The aim here is to know whether the absence of archaeological sites in parts of the study area is the result of human choice or a taphonomical bias. Following this, we present an overview of the available information on the subsistence of Mesolithic and Neolithic coastal populations in the area, through qualitative and quantitative analysis of the shellmidden fauna, and stable isotope analyses of human remains. Finally, we discuss the impact of neolithisation on the subsistence of coastal populations. The comparison of zooarchaeological and stable isotopic data provides an opportunity to compare the two independent lines of evidence, though a number of problems arise in terms of the representativeness of the faunal collections from earlier excavations, and the relatively poor and sporadic geographic and chronological coverage for both datasets. Nevertheless, the combination of the two lines of evidence permits observations and inferences that would not be possible using each on its own.

The spatial distribution of shellmiddens on the French Atlantic façade: explaining the gaps. Research on shellmiddens is a relatively recent phenomenon in France, and differs in this respect from the long and continuing history of research in, for example, Denmark (Andersen 2004, this volume). This fact is not linked to their absence. Indeed, surveys of different Breton islands show the presence of a number of unexcavated shellmiddens (SRA, Regional Institution of Archaeology files), though some of them are presumably more recent. Most archaeological attention has been and remains focused on Mesolithic and Neolithic shellmidden sites. The earliest known shellmiddens of the French Atlantic façade date to the Mesolithic, the earliest dates so far obtained (late 7th/early 6th millennium BC) being from Beg-an-Dorchenn and SaintGildas, though these are complicated because they are taken on marine shell. The absence of evidence for a coastal Palaeolithic is of course explained by the much lower sea-levels at that time. The Early and Middle Mesolithic periods are also not represented by shellmidden sites, as even in the latter period sea-levels remained some 20 to 30 m lower than at present (Pirazzoli 1991). Indeed, Late Mesolithic sea-levels along the French Atlantic coast were still between six and ten metres lower than at present (Pirazzoli 1991), so that the surviving shellmidden sites would have been at some distance from the actual coast at the time of their occupation (see below). Thus, the appearance of shellmiddens in the 6th millennium BC should

not be interpreted as a sudden shift in economic practices by Late Mesolithic groups. Variations in sea-level are the most obvious, but not the only consequence of the Flandrian transgression. Sedimentary deposits formed at this time have covered entire mid-Holocene landscapes. For example, the Landes area in southwest France is covered by a sand dune of more than ten metres thickness in places. In addition, between the rivers Gironde and the Loire, during the Neolithic there were large bays which are now marshes filled by sedimentary deposits that are again no doubt covering archaeological sites. Taken together, changing sea-levels and the coastal accretion of sedimentary deposits partly explain the heterogenous distribution of Mesolithic and Neolithic shellmiddens along the French Atlantic façade (Fig. 12.1) (Dupont 2003). The other relevant factor in the distribution of sites involves the cultural processes of neolithisation, which were variable both geographically and chronologically (Dupont 2003). Indeed, the near-absence of Mesolithic shellmiddens south of the Loire river may be explained by the earlier transition to the Neolithic way of life there as compared with Brittany. Any earlier Mesolithic sites in this area would have been lost to rising sealevels; indeed, it is noteworthy that the single Mesolithic shellmidden currently known south of the Loire – the fortuitously placed Saint-Gildas – is also among the earliest known sites. The relative lack of Neolithic shellmiddens in Brittany can be linked to the greater acidity of the soils here compared with those occurring between the Loire and the Gironde. This lack of sites probably does not necessarily mean the absence of shellfish consumption. In order that shells are preserved, the quantity of shells initially discarded at one location needs to be sufficient to change the pH of the soil to permit the survival of subsequent deposits of shells. In this respect, the situation in France differs from that in Denmark, where a degree of continuity in the use of a shellmidden is not infrequently seen between the Mesolithic and Neolithic (e.g. Andersen 1989, 1991, this volume). Only Beg-an-Dorchenn has a Neolithic component overlying a Mesolithic shellmidden, but this represents a change in the use of the site, as it is a chambered tomb (Giot 1947). Finally, both Mesolithic and Neolithic midden sites in Brittany are also subject to being covered by Holocene sediments, reducing their visibility to surveys, and relying on accidental discovery or erosion, such as was the case with Beg-er-Vil (Kayser 1990b). The appearance of the shellmiddens on the landscape is relevant not only to how the sites were used and perceived in the past, but also to how easily they are found during archaeological survey today. The thickness of the known Mesolithic

12 Prehistoric shell middens along the French Atlantic façade

Fig. 12.1: Map of the distribution of Mesolithic and Neolithic sites cited in the text.

125

126

Fig. 12.2: The stratigraphy of Mesolithic shellmiddens.

Catherine Dupont, Rick Schulting and Anne Tresset

shellmiddens is less than one metre (Fig. 12.2). Thus, even before becoming covered by later Holocene sand dunes, the middens were not in themselves notable features in the landscape. Their very survival today does indicate that they were situated at relatively high points in the local landscape, overlooking the coast, but it is impossible to say whether this is an intentional choice or, more likely, reflects a preservation bias. Known Neolithic shellmiddens are even slighter, with a thickness of less than 0.1 m (Dupont 2003). The only departure from this pattern of low, horizontal accumulations is found in mounded middens of Medieval sites such as Saint-Michelen-L’Herm (Vendée), a centre of oyster (Ostrea edulis) exploitation (Lecointre 1958). As implied above, post-Neolithic shellmiddens are certainly known from France, and indeed are numerous from the Gallo-Roman period, when they extend into the interior (Coulon 1990). But knowledge of these sites remains largely anecdotal (Labrousse 1959), and they are not discussed further in this paper.

A stone’s throw? The distance between the sites and their contemporary coastlines As discussed above, our knowledge of the presence of human populations is filtered by natural factors. These factors are also responsible for the different distances of the shellmiddens from the coastline

today compared with that existing during the period of their use. One of the present authors (C. Dupont) addressed this theme through compiling various lines of evidence, including the different sea-level curves, isobath maps, submarine topographic data, and paleoenvironmental boring. There remain many problems with the resulting reconstructions (Dupont 2003) and the distances shown in Table 12.1 should be seen as approximate and theoretical. The table does show, however, that many Mesolithic shellmiddens must have been destroyed by coastal erosion. During their use, these sites appear to have been located less than one kilometre from the shoreline: essentially, the Mesolithic communities exploiting marine resources were situated on or adjacent to the littoral. For the Neolithic period in west central France, the distance of the sites to the sea is more heterogeneous, with some groups further inland also having access to shellfish. Generally this distance was still less than 10 km, but in one case, the Late Neolithic site of Montagant, marine shells occur some 50 km inland.

Marine resources in the Mesolithic and the Neolithic: from exploitation to consumption The exploitation of marine shellfish As marine shells are the basic component of the archaeological sites that we are discussing in this

127

12 Prehistoric shell middens along the French Atlantic façade

NEOLITHIC

MESOLITHIC

Periods

Archaeological site

Beg-an-Dorchenn Téviec Late/ Hoëdic Final Beg-er-Vil Saint-Gildas IB Saint-Gildas IC Auzay Middle Les Gouillauds Champ-Durand Le Rocher Pied-Lizet La Mastine La Sauzaie La Garenne Late/ Montagant Final Chez Reine Ponthezières La Perroche Er Yoh La grotte du Phare Jardinet

Present distance from the Estimated distance from the coast coast at time of occupation 0m From 10 m to 500 m 0m From 20 m to 1 km 0m From 20 m to 1 km 0m From 150 to 500 m ca. 10 m ca. 1 km ca. 10 m ca. 1 km 27 km Less than 5 km 0m Less than 1 km 38 km Less than 9 km 3 km Less than 500 m 10 km Less than 1 km 14 km Less than 5 km 7 km Less than 500 m 15 km Less than 5 km 84 km 50 km 3 km 3 km 2 km 2 km 5 km Less than 1 km 50 m Less than 500 m 0m Less than 1 km 23 km Less than 5 km

paper, we first summarise some of the characteristics of the exploitation of this resource (Tables 12.2 and 12.3) (Dupont 2004). For both the Mesolithic and the Neolithic periods, human selection appears to have been designed to optimise the exploitation of the shellfish, in the sense that they gathered those available in the immediate environment of the site (with the above-noted exception of Montagant). Where the population had access to rocky shores, these species are the ones predominantly exploited. The limpet (Patella sp.) is the most frequently gathered species in such situations. The bivalves are also abundant (oysters Ostrea edulis, mussels Mytilus edulis, cockles Cerastoderma sp., carpetshells Tapes decussatus, peppery furrow shells Scrobicularia plana). Other gastropods seem to have been less frequently consumed. Those species most easily accessible in the littoral zone, and those from which the flesh can be rapidly removed from the shell, form the focus of human exploitation. At least theoretically, the shellfish gathering could have been a daily activity during the Mesolithic as well as during the Neolithic, in spite of some of the Neolithic sites being at a greater distance from the shore. Once seasonality studies have been undertaken, there will probably be found a strong seasonal component to collection, as has been noted elsewhere (Deith 1983; Milner 2002). Some differences are also observed between the two periods. During the Mesolithic, people seem

to have more fully exploited the available diversity in shellfish, while in the Neolithic there is a tendency to focus on a single substrate, with a limited number of species. This can be seen in the species diversity at two of the best-documented Mesolithic (Beg-an-Dorchenn and Beg-er-Vil) and Late Neolithic (Ponthezières and La Perroche) sites, with comparatively large sample sizes and adequate recovery through sieving (Fig. 12.3). For the two Late Neolithic assemblages, over 90% of the total weight of shells is comprised of only two species. For the two Mesolithic sites, the combined total of seven species is required to reach the same level. This difference could correspond to a lower diversity in the marine environment utilised by the Neolithic groups. However, the island locations of Ponthezières and La Perroche, as well as palaeoenvironmental coring next to the settlement of La Perroche raise doubts over this explanation (Dupont 2003). These cores show that the site was adjacent to a bay with estuarine conditions during its utilization. So theoretically, some shell species like peppery furrow, cockles or clams could have been accessible in the nearby environment of the site. The low consumption of these three species at La Perroche seems to be more human choice than environmental absence. What is needed to confirm or refute these observations on the differences in Mesolithic and Neolithic shellfish exploitation is additional new excavations on other sites from the two periods and in the same region.

Table 12.1: The distance of Mesolithic and Neolithic shellmiddens to the coast.

Sieving (Dupont 2003)

-fishes -crabs Minor except for birds -sea birds (ducks and waders) - grey seal (Tresset submitted) (Péquart and Péquart 1930, 1954; Desse and Granier 1976; Tresset submitted). (Tessier 1984) (Tessier 1984)

Abundant (Péquart and Péquart 1954).

829 (MNI) (Tessier 1984)

172 (MNI) (Tessier 1984)

Hoëdic

Téviec

Saint-Gildas IB

Saint-Gildas IC

This deposit is composed almost exclusively of shells, with occasional remains of fish and crustaceans. The mammal remains seem to be rare (Péquart and Péquart 1954)

+ (Dupont, 2005) (Tessier 1984)

Shells are the only economic component known from the midden (Tessier 1984; Dupont 2005)

Shells are the only economic component known from the midden (Dupont 2005)

Xҏį13C = -14.3‰; Xҏį15N =12.6‰ ; n = 11 c. 65–85% marine protein (Schulting and Richards 2001) *Note that ongoing reanalyses will modify these values somewhat Xҏį13C = -15.3‰; Xҏį15N =11.9‰ ; n = 10 ca. 55–75% marine protein (Schulting and Richards 2001) *Note that ongoing reanalyses will modify these values somewhat

-

Table 12.2: Quantitative and qualitative data on marine resources in the diet of Mesolithic coastal populations (n: number of fragments, MNI: minimum number of individuals).

(Tessier 1984)

(Tessier 1984)

-fishes + -crabs (but mainly Minor -sea birds in burial (Péquart and Péquart 1954). (Péquart and Péquart 1930 1954, Desse and contexts) Granier 1976).

Abundant (Péquart and Péquart 1954).

-

The marine shell deposit permitted the preservation of rare faunal remains (Kayser 1985)

Bibliographical information about the Isotopic data part of the marine resources in the diet

The major part of the deposit is + composed of shells with some terrestrial (but mainly landsnails and a small proportion of in burial terrestrial mammals remains (Péquart contexts) and Péquart 1930)

Minor n = 56 (Schulting et al. 2004)

-fishes: n= 75 (old collections) and analysis in progress (excavation Kayser 1984–1988 and excavation Dupont 2001) -crabs: n = 38 excavation Dupont 2001 on 1m² + -sea birds: n=3 (old collections) and n = 3 (excavation Kayser 1984–1988 on 53m²). (Schulting et al. 2004; Dupont 2003; Dupont and Gruet 2005) -fishes: analysis in progress -crabs: n = 101 -sea birds: n = 41 + -marine mammals, seal: n = 12 (Dupont 2003; Schulting et al. 2004 ; Tresset, submitted)

Other marine resources discovered

Abundant n = 86360 (Dupont 2003)

Minor Old collections, n= 34 (Tresset 2000) excavation Kayser 1984– 1988 on 53m² n = 74: (Schulting et al. 2004)

Terrestrial resources

Beg-er-Vil

Abundant n = 170958: excavation Beg-an-Dorchenn Dupont 2001 on 1m² (Dupont 2003)

Sites

Quantities of shells studied

128 Catherine Dupont, Rick Schulting and Anne Tresset

Terrestrial mammal resources n = 408 MNI = 338 (Large and Birocheau (Gruet unpublished in Dupont 2003) 2004)

n = 998 (Tresset in Laporte et al. 1998) n=50 (Lehnbach 2003) n = 1400 (approx) (Le Rouzic 1930 ; Tresset in progress)

n = 18330 (Dupont 2003)

n = 331814 (Laporte 1994)

n = 1523 (Dupont 2003)

Present (Le Rouzic 1930)

La Perroche

Ponthezières

La Grotte du Phare

Er Yoh

-fishes n = 1 -birds n = 10 (marine origin?) (Sicard et al. 2003) -fishes : analysis in progress -cetaceans: analysis in progress -fishes n = about 300 -crabs n = 77 (Gruet and Laporte 1996; Laporte et al, 1998) -seabirds, n=7 - fishes? (Lehnbach 2003) -fishes (mainly sea bream) – crabs -sea birds n = 334 -marine mammals (grey seal) (Le Rouzic 1930; Boyle 2005; Tresset submitted)

-

+

+

+

+

-

-

-

-

-

-

-

-

-fishes n = 4 -cuttle-fish -crab n = 1 -birds n = 89 (marine origin ?) (Joussaume 1981 ; Braguier 2000)

-

-

-

+

-

-

-

-

Only 160 fragments of shells have been counted for 15000 bony fragments associated (Debénath et al. 1984)

-

-

-

-

-

Almost 100 kg of animals remains littered the bottom of the hut with which the shells were associated (Gabet and Massaud 1965)

-

C = -19.7‰; į15N =10.7‰ ; n = 1, c. 10– 20% marine protein (Schulting et al., 2004)

į13

-

-

-

-

-

-

-

-

Shells are the only dietary component of the midden excavated, but terrestrial resources could exist in other unknown parts of the site (Rousseau et al. 2001)

-

Bibliographical informations about the part of the Isotopic data marine resources in the diet

Table 12.3: Quantitative and qualitative data on marine resources in the diet of Neolithic coastal populations (n: number of fragments, MNI: minimum number of individuals).

analysis in progress

n = 420 (Dupont unpublished)

Jardinet

n=75 (Boujot et al. 1997) n = 533 (Braguier 2000) n = 15594 Terrestrial birds n = 3 (Debénath et al. 1984) n = 2016 (Sicard et al. 2003)

n=328 (Boujot et al. 1997)

n = 173 (Debénath et al. 1984)

MNI = 13 (Gruet unpublished in Dupont 2003) MNI = 22 (Gruet unpublished in Dupont 2003) n = 4203 (Gruet unpublished in Dupont 2003)

Chez Reine

Le Rocher

Pied-Lizet

La Mastine

n = 28975 MNI = 206 (Braguier 2000) (Gruet unpublished in Dupont 2003)

n = 594 MNI =1435 (Gruet unpublished in Dupont 2003) (Braguier 2000)

La Sauzaie

ChampDurand

-

Abundant (Gabet and Massaud, 1965)

n = 162 (Dupont 2003)

La Garenne -birds n =2 (marine origin ?) (Braguier 2000)

-

n = 119 (Boujot et al. 1996)

n = 31 (Boujot et al. 1996)

Montagant

-

-

-fishes n = 2 -crabs -sea birds (Birocheau and Large 1987)

-

Sieving

Other marine resources discovered

n = 11087 (Dupont 2003)

Quantities of shells studied

Les Gouillauds

Auzay

Sites

12 Prehistoric shell middens along the French Atlantic façade 129

130

Catherine Dupont, Rick Schulting and Anne Tresset

Fig. 12.3: Comparison of the shellfish diversity at selected Mesolithic and Neolithic sites.

Thus, on current evidence it seems that Mesolithic communities exploited a wider range of the available molluscan species compared with the Neolithic. If this can be sustained, then it seems that Mesolithic populations practiced a more opportunistic, or ‘broad spectrum’ (Flannery 1969) strategy in this respect. The interpretation of the use of shellfish on the Neolithic sites is uncertain. First of all, a distinction must be made between the insular sites on Île de Ré and Île d’Oléron, and those on the mainland. For the insular Neolithic sites, shellfish collection does not seem to be a casual activity, but rather one focussed on a small range of species, although overall a wider range of species is exploited compared to the mainland sites (though there are issues with recovery bias also to be considered, as the insular sites were sieved). Partly this emphasis may be due to the use of the shells for the manufacture of beads, which appears to have been an important activity at Ponthezières and La Perroche (Laporte et al. 1998; Dupont 2003; Schulting et al. 2004). In light of the additional data to be discussed below, the more selective shellfish collection strategy of Neolithic communities could also be related to their overall reduced reliance on the marine environment. While the insular shellmiddens are large, those on

the adjacent mainland are typically far smaller, often only with pockets or thin scatterings of shells (Table 12.3). In spite of the constant attrition to which the French shellmiddens are subjected from various destructive agents, these sites can still contain a significant amount of deposits. The weight of the shell in the Beg-er-Vil shellmidden has been estimated to be more than three tons, at Beg-anDorchenn about five tons, and Hoëdic more than 14 tons (Dupont and Gruet 2002). Though the values are approximate, these deposits would seem to reflect more than incidental exploitation of this resource by Mesolithic populations. Comparable calculations for the Neolithic sites are currently underway. But what can already be noted is that the concentration of shells is much denser in the Mesolithic sites. Even the sites of La Perroche and Ponthezières have a shell density in the range of 2–3 kg of shell per cubic metre, as compared to 6 kg/m3 for Beg-er-Vil and more than 29 kg/ m3 for Beg-an-Dorchenn (Dupont 2003). The wider exploitation of the marine environment Shellfish of course often represent only part of the marine economy evidenced in shellmiddens. In some cases they may indeed represent the

131

12 Prehistoric shell middens along the French Atlantic façade principal use of marine resources at the site, while in other cases they may be overshadowed, in nutritional terms, by the remains of fish, sea birds and marine mammals. Any quantitative comparison of these different categories of zooarchaeological remains presents a host of wellknown obstacles, including preservation and recovery biases. The difficulties are exacerbated when it is considered that, even aside from these biases, the shellmidden may never have contained a representative picture of the marine economy (let alone the terrestrial economy) of the group using it, since there may be a more complex spatial structure to the occupation. In other words, the preparation, consumption and disposal of nonshellfish foods may take place away from the shellmidden. With these caveats firmly in mind, summary information concerning marine subsistence resources from Mesolithic and Neolithic sites is presented in Tables 12.2 and 12.3. The larger Mesolithic shellmiddens (Beg-anDorchenn, Beg-er-Vil, Téviec and Hoëdic) contain evidence for many different aspects of a marine economy, including varied fishes, crabs, marine mammals, sea birds and other bird species – such as anatids and waders – that can be found on the shore at some time of the year. As seen with the shellfish, a wide range of the available marine resources has been exploited at each location, for example at least four crab species have been identified at à Beg-an-Dorchenn and at Beg-er-Vil (Dupont and Gruet 2005). These sites contain evidence for a wide range of activities (Marchand 2005), fortuitously preserved by the presence of the shells acting as a soil pH buffer, and protecting structures such as hearths and other features from later disturbance through burial. People appear to be living or at least camping on, or near, the shellmiddens. There is a strong indication from various lines of evidence that these sites represent more than temporary camps, with seasonal indicators of more than one season, though yearround occupation cannot be firmly demonstrated (if indeed this is ever possible). To these large shellmiddens may be added the information from the two smaller Mesolithic sites at Saint-Gildas. The shellmiddens here differ in containing only shellfish remains (Table 12.2), rather than the range of faunal remains present at the other sites, suggesting that they operated as specialized shellfish collection locations (Dupont 2005; Dupont et al. submitted). The Saint-Gildas sites may be linked to residential sites located elsewhere through a logistical organization of the territory (Dupont 2004, 2005). The morphology of the shellmidden deposits differs in the Neolithic sites (Dupont 2005). Many are thin layers of shells where the malacofauna is not necessarily the major component of the

deposits. The largest quantities of shells observed during the Neolithic are found on island sites (Ponthezières, La Perroche, Les Gouillauds and Er Yoh) or sites next to the shoreline (Le Rocher, La Sauzaie). But some Neolithic sites located at less than 1 km from the littoral are composed of low quantities of shells (Pied-Lizet, La Garenne, Chez Reine). Unlike the Mesolithic sites, the Neolithic shellmiddens showing a greater diversity of marine resources are rare and are mainly linked to insular sites. On the mainland, only Auzay shows some diversity, with the rare remains of crabs. However, the absence of sieving for many of these sites can partly account for the absence of remains of fish and crabs, though if these were present in any quantity, the larger species at least should have been noted (as for example with the fish at Er Yoh). The data on the proportions of excavated shells in comparison with the terrestrial resources are often qualitative. When quantitative data are present, they show the same result, that is to say, a smaller proportion of marine resources. The terrestrial component Terrestrial faunal remains are certainly present in most of the Mesolithic shellmiddens, but are difficult to quantify. Recent new excavations have only looked at small areas, while many of the remains from the earlier excavations have been lost. The most obvious point that can be made on the basis of the available evidence is that the terrestrial mammalian fauna does seem to be far better represented than marine mammalian fauna (Table 12.2), which consists of only a few remains of grey seal at Téviec and at Beg-er-Vil. The re-analysis of the mammalian remains of these Mesolithic deposits shows no evidence of domestic species (Tresset 2002). The indications of domestic fauna quoted in older publications are likely to be misidentifications or later intrusions, as it seems to have occurred at Beg-anDorchenn, where an Iron-Age site is located on top of the shellmidden. Although virtually no data documents the beginning of the Neolithic, Mid to Late Neolithic shellmiddens suggest an economy heavily reliant on domestic fauna, namely cattle, sheep/goat and pig. Marine molluscs were probably not a negligible component of the diet, but crustaceans, fishes, birds and wild mammals all seem to have been of minor importance. A notable exception is the Late Neolithic site of Er Yoh, where grey seal and sea bird remains (especially gulls, shags and cormorants) were quite abundant. This specificity is very likely linked to the location of the site on what was already an island (the site nowadays corresponds to a small islet, peripheral to Houat island). Demographic data obtained on seal remains, characterised by an

132

Catherine Dupont, Rick Schulting and Anne Tresset important number of very young animals, point toward the exploitation of neighbouring colonies (Boyle 2005), while the lack of clear cut marks and breakage questions the anthropogenic origin of the bird assemblage (Tresset submitted). Stable isotopic analysis Stable carbon and nitrogen isotope analysis undertaken on human bone collagen confirm the importance of marine resources in the diet of Late Mesolithic coastal populations at Téviec and Hoëdic. The previously reported average values for these sites indicate a high reliance on marine protein, estimated at 60–90% (Table 12.2; Fig. 12.4) (Schulting and Richards 2001). Work in

Fig. 12.4: Plot of AMS date and d13C values (division into ‘terrestrial’, ‘intermediate’, and ‘marine’ is based on 3‰ divisions between assumed endpoints of -21 and -12‰) (Source: Schulting and Richards 2001; Schulting 2005). Individual

Table 12.4: AMS dates and stable isotope values for Neolithic human bone from Brittany. Source: Schulting 2005, Table 12.1; Giot et al. 1994.

Port Blanc R82.31.1 Port Blanc, Broca 265 Port Blanc R82.31.25 Conguel, Broca 282 Conguel, Broca 283 Er Yoh, EY.HS.4 Er Yoh, EY.HS.2 Beg an Dorchenn Ty Floc'h

Lab. No. OxA-10615 OxA-10936 OxA-10699 OxA-10937 OxA-10938 OxA-10843 OxA-10810 OxA-5363 OxA-5974

Date BP BP 5070 ± 50 5050 ± 40 4200 ± 45 3630 ± 35 3610 ± 35 4240 ± 55 3130 ± 40 4140 ± 55 5270 ± 80

Calibration: CALIB 5.0.2 (Reimer et al. 2004) Sources: Schulting 2005, Table 1; Giot et al. 1994

progress utilising ultrafilters in sample preparation is revising these values in the direction of slightly more elevated results for both carbon and more especially nitrogen, indicating an even higher (and more consistent) contribution of marine protein (Schulting and Richards n.d.; the AMS dates are more radically affected, Schulting 2005). Of the other Mesolithic sites, only Beg-er-Vil has yielded human remains, consisting of two isolated elements, one of which has been lost. Unfortunately, the collagen in the remaining specimen was poorly preserved, and did not yield any usable results. The great majority of the available Neolithic stable isotope measurements are on human remains from chambered tombs rather than the shellmidden sites that form the focus of this discussion (Fig. 12.4; Table 12.4). The only exception to this is Er Yoh, with values for two humans, one Late Neolithic and the other Late Bronze Age. The results for the former (Table 12.3) do suggest a small input of marine protein, which can be seen as consistent with the available faunal data. The remains of shellfish, fish (dominated by sea bream), seabirds and seals (Boyle 2005) here are still secondary to the domestic faunal remains, which comprise c. 80% of the mammalian fauna (Schulting et al. 2004; Tresset 2003 submitted and see above), though of course direct comparison with the other categories of faunal remains is problematic, particularly in the absence of sieving. The stable nitrogen value for the Final Neolithic human is high (δ15N = 12.0‰), which on the face of it would suggest that the major contribution of marine protein was in the form of seals. This is complicated by the fact that the specimen in question is a child, and so may be subject to a residual breastfeeding effect. The stable carbon isotope (δ13C) values on human remains from Neolithic chambered tombs at Beg-an-Dorchenn (Finistère) and Port Blanc Date cal BC (± 2V) 3950–3710 3930–3660 2870–2580 2120–1890 2120–1840 2890–2590 1410–1140 2840–2480 4330–3960

G13C

G15N

-19.9 -19.5 -19.6 -20.6 -20.7 -19.3 -18.7 -19.5 -21.6

10.9 10.3 11.0 7.9 9.8 12.0 10.7 -

Estimated % marine (± 10%) 12 17 16 4 3 19 26 17 0

133

12 Prehistoric shell middens along the French Atlantic façade and Conguel (Morbihan) range between –19.5‰ and –20.7‰, while δ15N values range between 7.9‰ and 12.0‰. From the direct AMS determinations, two individuals from Port Blanc are earliest (c. 3900–3700 BC, Middle Neolithic), while the others belong to either the Late Neolithic or, in the case of Conguel, the Early Bronze Age (Schulting 2005, Table 12.1). The lowest δ13C and δ15N values are on the two late individuals from Conguel, though there is no consistent pattern of a decrease through time in the contribution of marine protein, since the Late Bronze Age individual from Er Yoh has the highest value (-18.7‰) seen among the post-Mesolithic humans. It should also be noted that the contribution of marine protein need not always be direct, but may come from feeding marine foods to domestic animals (e.g. seaweed to sheep; almost anything to pigs), as it has been recently demonstrated that these practices did occur as early as the Neolithic in Orkney (Balasse et al. 2005). There are unfortunately no human remains available for stable isotope analysis from the Late Neolithic shellmiddens of west central France. Some human bones are available from the Middle Neolithic site of Auzay, but these have yet to be analysed isotopically (Large et al. 2004). In particular, it would be interesting to compare the insular sites, with their large middens, with the mainland sites. How important an element of the subsistence economy are the shellfish that are so visible at sites like Ponthezières and La Perroche?

Conclusions In the first part of this paper we addressed the biases acting on the surviving distribution of shellmidden sites along the French Atlantic façade. A combination of rising sea-levels and sedimentation along the coastal zone has contributed to the archaeological invisibility of many sites, under the sea in the first instance (affecting mainly Mesolithic sites), and under sand and/or estuarine sediments in the second instance (affecting mainly Neolithic sites). These taphonomic factors in part account for the spatial separation of Mesolithic and Neolithic shellmidden sites in the study area. The Mesolithic sites fall late within this period, and are mainly observed in Brittany, while the Neolithic sites, again mainly late within the period, are primarily found south of the river Loire. For both periods, but particularly for Brittany, with its more acid soils, there is likely to be a strong bias against the survival (and identification) of small shellmiddens. A buffer needs to accumulate before a shellmidden will survive to the extent where it can be identified

millennia later. This would not only affect our understanding of the full range of Mesolithic shellmidden sites (and Saint-Gildas serves to show that special-purpose shellfish collection was also practiced at this time), but may also partly account for the hiatus of shellmidden sites observed between the Late Mesolithic and the Late Neolithic. The Neolithic exploitation of shellfish may have been more sporadic and incidental, and not repeatedly focussed at the same location. Cultural factors are also relevant. At the large scale, the process of neolithisation was not synchronous along the French Atlantic façade. A comparison between the species consumed and the palaeoenvironment shows that a wide range of marine resources was exploited in the Mesolithic, most if not all of which could probably be taken from the in-shore region. Although some food debris could have been discarded away from the settlement areas (and this would apply of course to both periods), several lines of evidence show a lower consumption of marine resources by Neolithic populations in comparison with those of the Mesolithic. The volumes of the shellmiddens are generally significantly larger during the Mesolithic, while in the Neolithic they tend to be small deposits within the settlements. Groups in the Mesolithic appear to have utilised the larger of the available species in approximate proportion to their expected prevalence near the site. For the Neolithic, a greater element of selection is implied, with a focus on fewer species. The faunal data from the excavations are not inconsistent with the stable isotope data on human bone. Both lines of evidence seem to show a dietary change in the Neolithic, towards an economy based mainly on terrestrial resources. The speed of this process is difficult to assess with zooarchaeological data in the study area, since the sites that do preserve faunal remains are all late in the Neolithic. Many of the earliest Neolithic settlements are today likely to be located underwater or, if on shore, buried under metres of sand. However, stable isotope measurements on dated human bone from Neolithic monuments in Brittany does bridge some of this gap, and suggests that the change in economic orientation was rapid and fairly comprehensive, though some individuals do continue to exhibit a slight contribution from marine resources. The sample size for the Neolithic remains low, and further work is underway to analyse additional humans from coastal Brittany and Normandy. Thus, the acquisition of domestic plants and animals, and the associated techniques of farming and stock-rearing had major repercussions on the diet of coastal populations, leading to a partial forsaking of marine resources that in the Mesolithic were used to a far greater extent. The

134

Catherine Dupont, Rick Schulting and Anne Tresset little information that is available suggests that for the more recent, post-Neolithic periods, marine resources continued to be of secondary importance in the diet, though much more research needs to be undertaken on their use in the later prehistoric and historic periods along the French Atlantic façade.

References Andersen, S. H. (1989) Norsminde: A “Køkkenmødding” with Late Mesolithic and Early Neolithic occupation. Journal of Danish Archaeology 8, 13–40. Andersen, S. H. (1991) Bjørnsholm. A stratified Køkkenmødding on the central Limfjord, north Jutland. Journal of Danish Archaeology 10, 59–96. Andersen, S. H. (2004) Danish shell middens revisited. In A. Saville (ed.), Mesolithic Scotland and its Neighbours, Edinburgh: Society of Antiquaries of Scotland, 393–411. Balasse, M., Tresset, A., Dobney, K. and Ambrose, S. (2005) Investigating seaweed grazing from tooth enamel ratio: an exploratory study on the North Ronaldsay sheep (Orkney, Scotland). Journal of Zoology 266, 283–291. Bellancourt, G. (1980) Le kjoddenmodding de la Pointe Saint-Gildas et les sites Néolithiques à microlithes de l’interieur et des rivages de la Loire-Atlantique. Société Nantaise de Préhistoire 1980, 5–28. Birocheau, P. and Large, J. M. (1987) L’éperon des Châtelliers du Vieil-Auzay, In Préhistoire de PoitouCharente : problèmes actuels, 111ème congrès national des sociétés savantes, Poitiers 1986, Paris: Éditions du Comité des travaux historiques et scientifiques, 399–432. Boujot, C., Cassen, S., Chambon, Ph. and Gruet, Y. (1996) Matignons et Moulin-de-Vent à Montagant (Le Brandard, Mainxe, Charente), Bulletin de la Société Préhistorique Française 93, 63–83. Boujot C., Hoad I. and Serjeantson, D. (1997) Les vestiges osseux des enceintes de La Mastine et Pied-Lizet. In S. Cassen and C. Scarre (eds.) Les enceintes néolithiques de la Mastine et Pied-Lizet (Charente-Maritime). Association des Publications Chauvinoises, Mémoire XIII, 111–122. Boyle, K. (2005) Late Neolithic seal hunting in southern Brittany: a zooarchaeological study of the site of Er Yoh (Morbihan). In G. Monks (ed.) The Exploitation and Cultural Importance of Sea Mammals, Oxford: Oxbow Books, 77–94. Braguier, S. (2000) Économie alimentaire et gestion des troupeaux au Néolithique récent / final dans le centre-ouest de la France, PhD. University Toulouse Le Mirail. Coulon, G. (1990) Les gallo-romains Métiers, vie quotidienne et religion. Tome 2, Paris: Armand Colin. Deith, M. R. (1983) Molluscan calendars: the use of growthline analysis to establish seasonality of shellfish collection at the Mesolithic site of Morton, Fife. Journal of Archaeological Science 10, 423–440. Débenath, A., Flies, and Pautreau, J. P. (1984) Gisement chalcolithique de l’enclos de Tonnay-charente, Bulletin de la Société Préhistorique Française 81, 117–121.

Desse, G. and Granier, J. (1976) Les poissons, In H. de Lumley (ed.) La Préhistorie Française, Civilisations paléolithiques et mésolithiques 1, Paris: C N R S, 437–443. Dupont, C. (2003) La malacofaune de sites mésolithiques et néolithiques de la façade atlantique: Contribution à l’économie et à l’identité culturelle des groupes concernés, PhD. University Paris I. Dupont, C. (2004) L’exploitation de la malacofaune marine dans l’économie de subsistance des populations mésolithiques et néolithiques de la façade atlantique française: une activité secondaire? In Actes des XXIVème rencontres internationales d’archéologie et d’histoire d’Antibes ‘Petits animaux et sociétés humaines: du complément alimentaire aux ressources utilitaires’, 15–27. Dupont, C. (2005) Les coquillages alimentaires des dépôts et amas coquilliers du Mésolithique récent / final de la façade atlantique de la France: de la fouille à un modèle d’organisation logistique du territoire, Préhistoire, Anthropologie Méditerranéennes 2003 12, 221–238. Dupont, C. and Gruet, Y. (2002) Estimation de la ressource alimentaire en masse de chair d’après les restes de coquilles: applications aux berniques Patella sp. et au ‘bigorneau’ Monodonta lineata de sites mésolithiques et néolithiques, Revue d’Archéométrie 26, 93–112. Dupont, C. and Gruet, Y. (2005) Malacofaune et crustacés marins des amas coquilliers mésolithiques de Beg-anDorchenn (Plomeur, Finistère) et de Beg-er-Vil (Quiberon, Morbihan). In G. Marchand, and A. Tresset (eds.) Unité et diversité des processus de néolithisation sur la façade atlantique de l’Europe (6e–4e millénaires avant J.-C.), Table ronde de Nantes 26–27 avril 2002, Mémoire de la Société Préhistorique Française 36, 139–161. Dupont, C., Marchand, G., Gruet, Y. and Tessier, M. (submitted) La Pointe Saint-Gildas (Préfailles, LoireAtlantique): lieu témoin des passages de populations humaines du Mésolithique et de modifications environnementales, Gallia Préhistoire. Flannery, K. (1969) Origins and ecological effects of early domestication in Iran and Near East. In P. J. Ucko and G. W. Dimbleby (eds.) The Domestication and Exploitation of Plants and Animals, London: Duckworth, 73–100. Gabet, C. and Massaud, J. (1965) Le gisement peurichardien de la Garenne, commune de St-Hippolyte (charente-Maritime), Bulletin de la Société Préhistorique Française 62, 159–195. Joussaume, R. (1981) Le néolithique de L’Aunis et du Poitou occidental dans son cadre atlantique. Travaux du laboratoire d’anthropologie-préhistoire-protohistoire et quaternaire armoricains, Rennes: Université de Rennes1. Giot, P.-R. (1947) Le tumulus mégalithique de Beg-anDorchenn (Finistère). Gallia 5, 167–170. Giot, P.-R., D. Marguerie and H. Morzadec. (1994) About the age of the oldest passage-graves in western Brittany. Antiquity 68, 624–626. Gruet, Y. and Laporte, L. (1996) Crabes pêchés au Néolithique Final à Ponthezières: Identification des modes de pêche et application de la mètrique, Supplément de la revue Archéométrie, 197–201. Kayser, O. (1985) A propos de la fin du Mésolithique en Bretagne: l’amas coquillier de Beg-an-Dorchenn

135

12 Prehistoric shell middens along the French Atlantic façade (Finistère). Note préliminaire. Travaux de l’Institut d’art préhistorique, Université de Toulouse-Le-Mirail, 80–92. Kayser, O. (1990a) Beg-an-Dorchenn. Gallia information 64, 30–31. Kayser, O. (1990b) Beg-er-vil. Gallia information 64, 64. Kayser, O. and Bernier, P. (1988) Nouveaux objets décorés du Mésolithique Armoricain. Bulletin de la Société Préhistorique Française 85, 45–47. Labrousse, M. (1959) Recherches sur l’alimentation des populations gallo-romaines : escargots, huîtres et fruits de mer à Montmaurin. Pallas, VIII fasc.2, 57–84. Laporte, L. (1994) Parures et centres de production dans le CentreOuest de la France au Néolithique final, PhD. University Paris I. Laporte, L., Desse-Berset, N., Gruet, Y. and Tresset, A. (1998) Un lieu de production de parure au Néolithique final et son économie de production: le site de Ponthezières à Saint-Georges-d’Oléron (17). In Actes du 21ème Colloque Interrégional sur le Néolithique, Poitiers, oct. 1994, Associations des Publications Chauvinoises, 237– 255. Large, J. M., Birocheau, P., Cros, J.-P. and Duday, H. (2004) Les Châtelliers du Vieil-Auzay (Vendée). Une archéologie d’un site exceptionnel de la préhistoire récente, La Roche-sur-Yon: Groupe Vendéen d’Etudes Préhistoriques. Lecointre, G. (1958) Sur les buttes coquillères de St-Michel en l’Herm (Vendée), Bulletin de la Société Préhistorique Française 55, 466. Lehnbach, C. (2003) La grotte du Phare (Biarritz, Pyrénées Atlantiques). Origine des assemblages fauniques du Néolithique récent/final et de l’Age du Fer. Etude archéozoologique. Mémoire de Maîtrise, Université de Paris I – PanthéonSorbonne. Le Rouzic, Z. (1930) Carnac, fouilles faites dans la région de l’îlot d’Er-Yoh (Le Mulon). Commune de Houat 1924–1925. Vannes: Lafolye and J. De Lamarzelle. Marchand, G. (2005) Les niveaux coquilliers du Mésolithique finalen Bretagne: fonctionnement des habitats côtiers et intégration territoriale, Préhistoire, Anthropologie Méditerranéennes 2003 12, 209–220. Milner, N. (2002) Incremental Growth of the European Oyster Ostrea edulis. Oxford: Archaeopress, British Archaeological Reports, International Series 1057. Péquart, M. and Péquart, S.-J. (1930) La nécropole mésolithique de Téviec, Bulletin de l’Association Lorraine d’Etudes Anthropologiques 1, 35–46. Péquart, M. and Péquart, S.-J. (1954) Hoëdic, deuxième stationnécropole du Mésolithique côtier Armoricain, Anvers: De Sikkel. Péquart, M., Péquart, S.-J., Boule, M. and Vallois, H. (1937) Téviec: station nécropole mésolithique du Morbihan, Archives de l’Institut de Paléontologie humaine 18. Pirazzoli, P. A. (1991) World atlas of Holocene sea level changes. Oceanography Series 58. Reimer, P. J., Baillie, M. G. L., Bard, E., Bayliss, A., Beck, J. W., Bertrand, C., Blackwell, P. G., Buck, C. E., Burr, G., Cutler, K. B., Damon, P. E., Edwards, R. L., Fairbanks, R. G., Friedrich, M., Guilderson, T. P., Hughen, K. A., Kromer, B., McCormac, F. G., Manning, S., Bronk

Ramsey, C., Reimer, R. W., Remmele, S., Southon, J. R., Stuiver, M., Talamo, S., Taylor, F. W., van der Plicht, J. and Weyhenmeyer, C. E. (2004) Intcal version 4.14. Radiocarbon 46, 1029–1058. Rousseau, J., Robert, P., Cantin, N., Dupont, C., Gruet, Y., Huet, N. and Perrichot, V. (2001) Exploitation d’un territoire insulaire au Néolithique Moyen d’après la fouille d’un dépôt coquillier Les Gouillauds à BoisPlage-en-Ré (Charente-Maritime – France), Supplément à la Revue d’Archéométrie, 221–242. Schulting, R. J. (2005) Comme la mer qui se retire: les changements dans l’exploitation des ressources marines du Mésolithique au Néolithique en Bretagne. In G. Marchand and A. Tresset (eds.) Unité et diversité des processus de néolithisation sur la façade atlantique de l’Europe (7–4ème millénaires avant J.-C.), Paris, Mémoire de la Société Préhistorique Française XXXVI, 163–171. Schulting, R. J. and Richards, M. P. (2001) Dating women and becoming farmers: new paleodietary and AMS dating evidence from the Breton Mesolithic cemeteries of Téviec and Hoëdic, Journal of Anthropological Archaeology, 1–31. Schulting, R. J., Tresset, A. and Dupont, C. (2004) From harvesting the sea to stock rearing along the Atlantic façade of North-Western Europe, Environmental Archaeology 9, 143–154. Sicard, S., Braguier, S., Dupont, C., Goin, R., Riche, C., Rousseau, R., Sellami, F. and Sidéra, I. (2003) Résultats de la fouille préventive de l’enceinte néolithique du Jardinet aux Magnils-Reigniers (Vendée), In Actes de la Journée d’information du 16 novembre 2002 Paris Internéo 4, 131–146. Tessier, M. (1984) Les industries préhistoriques à microlithes du Pays de Retz, in Collectif, les sites à microlithes entre Vilaine et Marais Poitevin, Études Préhistoriques et Protohistorique des Pays de la Loire 7, 73–132. Tresset, A. (2000) Early husbandry in Atlantic areas. Animal introductions, diffusions of techniques and native acculturation at the north-western fringe of Europe, In J. C. Henderson (ed.) The Prehistory and Early History of Atlantic Europe. Papers from session held at the European Association of Archaeologists Fourth Annual Meeting in Göteborg 1998, British Archaeological Reports, International Series 861, 17–32. Tresset, A. (2002) De la mer au bétail en domaine atlantique: unité et diversité des processus d’apparition de l’élevage à la marge nord-ouest de l’Europe, Anthropozoologica 36, 13–35. Tresset A. (2003) French connections II: of cows and men. In I. Armit, E. Murphy, E. Nelis, and D. Simpson (eds.) Neolithic Settlement in Ireland and Western Britain. Oxford: Oxbow Books, 18–30. Tresset A. (submitted) L’avifaune des sites mésolithiques et néolithiques de Bretagne (5500 à 2500 av. J.-C.): implications ethnologiques et biogéographiques. Revue de Paléobiologie. Zvelebil, M., and Rowley-Conwy, P. (1986) Foragers and farmers in Atlantic Europe, In M. Zvelebil (ed.) Hunters in transition. Cambridge: Cambridge University Press, 67–93.