Shellfish from the Bronze Age Site of Clos des Châtaigniers ( Mathieu , Normandy , France )

1. University of Rennes 1, UMR 6566 CReAAH, Campus de Beaulieu, Bât 24-25, 263 avenue de Général Leclerc, CS 74205, 35042 Rennes CEDEX. Email: caroline.mougne@gmail.com 2. CNRS, UMR 6566 CReAAH, Campus de Beaulieu, Bât 24-25, 263 avenue de Général Leclerc, CS 74205, 35042 Rennes CEDEX.catherine.dupont@univ-rennes1.fr 3. INRAP, UMR 6566 CReAAH, Base INRAP de Bourguébus, 4 boulevard de l'Europe, 14 540 Bourguébus. Email: david.giazzon@inrap.fr 4. CNRS, UMR 6566 CReAAH, Campus de Beaulieu, Bât 24-25, 263 avenue de Général Leclerc, CS 74205, 35042 Rennes CEDEX. Email:laurent.quesnel@univ-rennes1.fr


Introduction
This article presents the archaeomalacological study of the site of Clos des Châtaigniers, at Mathieu (Normandy, France).The interest of such a study is that it is one of the first on the Channel coast dated to the Bronze Age.In fact, for this period, there are few studies on the exploitation of shellfish by the coastal populations in France (Weydert 1994;Dupont 2008;2011;2013;Mougne in prep), and especially on the Channel coast (Mougne and Dupont 2012;Mougne and Dupont in press).
Five objectives guide this study.The first is to characterise the marine invertebrates discovered on the site, then to establish whether their gathering was intentional (e.g. for human consumption) or accidental, in order to generate some data on the diet of the inhabitants.These data make it possible to define the environments exploited and the gathering methods used.The third objective tries to identify some culinary practices linked to the preparation and cooking of marine shells.Then, it seems interesting to carry out a spatial distribution of the shells within the settlement, in order to obtain information on the management of marine waste and on the possible presence of some preparation or consumption areas.The final objective aims at analysing a particular set of small fragments of shellfish, which could have originated from the stomach contents of aquatic animals.

Description of the Site
The Clos des Châtaigniers was excavated in 2010 by the French National Institute of Preventive Archaeological Research (INRAP), under the direction of David Giazzon (Giazzon 2013).This domestic settlement is located approximately 10km from the seashore, close to the River Dan, a tributary of the Orne.The archaeologists discovered a vast semi-circular structure, both pottery and radiocarbon dating assigning it to the Late Bronze Age (Cal BC 1190-980, Cal BC 1000-840, Cal BC 1050-900, Cal BC 1020-900 and Cal BC 970-820).This enclosure consists of two joined ditches, with a V-shaped profile, reaching 1.6m to 2m in depth.This structure is open on its northern side and is associated with numerous structures, mostly post-holes, but also some pits and a large combustion structure.The archaeological material is not very abundant.A small amount of slag and some fragments of crucible seem to indicate an activity linked with metallurgy.This hypothesis is reinforced by the discovery of ceramic moulds in the ditch of the enclosure and also in some pits.Even fragmented, these remains indicate the making of swords, spearheads, chapes and possibly axes.Eight shell accumulations have been identified in the ditch of the enclosure (Figure 1).

Methods
All the information and results presented here have been obtained from sedimentary samples (45 litres in total) collected during the excavation.
Sieved with fine meshes (4 and 2 mm), these samples were sorted and studied at the Archaeosciences Laboratory of the University of Rennes.
The identification of marine invertebrates (gastropods, bivalves and sea urchins) was carried out using the comparative collection from the same laboratory (comparative collection Gruet and Dupont, UMR 6566-CReAAH), and was confirmed using several reference works on marine biology (Tebble 1966;Poppe and Goto 1991;1993;Hayward and Ryland1995;Quéro and Vayne 1998;Audibert and Delemarre 2009).The scientific names used for the identified marine molluscs correspond to the international standards of the CLEMAM (2013).Several counting methods were used to calculate the relative proportions of each species.The NISP (Number of Identified Specimens) corresponds to all the identified remains greater than 2mm.The MNI (Minimum Number of Individuals) was also used.For bivalves, an MNI was obtained after assigning valves to the left or right according to the position of the teeth, the ligament and the pallial sinus line (Dupont 2006a;McCarthy et al. 1999).For spiral gastropods, the MNI is linked to the presence of the peristome (Dupont 2006a).Finally, the remains of each species were weighed (in grams).
The shells were then measured using a digital calliper graduated in millimetres (mm) (0.01) in accordance with the procedures described by Dupont (2006a).Some specific biometrical studies were also performed on the common mussel (Mytilus edulis).Shells found at the site are highly fragmented and consequently the whole length is not measurable.In fact, their fragile nature meant that the shells were found crushed as a result of the mechanical pressure exerted by the overlying sediments.Some authors highlight the fact that a good correlation exists between the height and the length of the shell (Buchanan 1985;Dupont 2006a, 86-87;Campbell 2013) and that it is therefore possible to reconstruct a mussel original size from measurements of shell height.However, mussels discovered at the Clos des Châtaigniers were so fragmented that shell height could not be measured.Therefore, a different method was employed, relying on a correlation established between original mussel size and a small part near the teeth (corresponding to the not pearly part, located just after the anterior adductor muscle scar; see Figure 4) (Mougne in prep).Through taphonomic analysis, modifications due to physical and chemical factors on the archaeomalacological material were identified (Claassen 1998;Dupont 2006a;Gutiérrez Zugasti 2008).The associated faunas in the thickness of the shell and in or on the shells were also observed and recorded (Hayward and Ryland 1995).

Faunal spectrum
Thirty species of marine shells (20 bivalves, 10 gastropods), one sea urchin and one crustacean were identified (Figure 2) at the site of Clos des Châtaigniers.Mussels (Mytilus edulis) represent the majority of marine shell remains in all three types of quantification used (98% of the NISP, 75% of the MNI and 94% of total weight) (Figure 3).The high percentage of NISP for this species is linked to high levels of fragmentation.As a consequence, the relation between the Number of Identified Specimens and the Minimum Number of Individuals (221752/1490) shows that, for a single mussel individual, 149 remains were counted.It is thus often difficult to obtain total lengths and to describe the collecting strategies.From these correlations between total length and a small part near the teeth, 214 total lengths (RV = 105; LV = 109) have been reconstructed.These measurements ranged between 27 and 65mm, with a mean of 43.5mm, corresponding to adult individuals (Figure 4).represented by only 1 to 13 individuals (Figure 3).The small quantities of these marine invertebrates are unlikely to indicate intentional collecting for human consumption.In addition, some of these species are small in size and evidence shows that deposited specimens were already dead when they arrived at the site.Only one fragment of sea urchin, species unidentified, was found (Figure 2, 31).Finally, some barnacles, belonging to the order of Cirripedia crustaceans, were observed (Figure 2, 32) and probably brought to the site during the transport of the mussels.
In summary, the size and quantity of mussel remains show that this shellfish was consumed by past inhabitants.However, the consumption of the other thirty-one identified marine invertebrates cannot be proved because of their presence in small quantities (Figure 3), small sizes and/or observed taphonomic modifications.

Gathering area
The site of Clos des Châtaigniers is located 10km from the Channel coast.
The gathering strategies reflected the values attributed to each species (e.g.flavour), and other environmental parameters such as accessibility and abundance at the seashore.The inhabitants seem to have exploited mainly rocky to muddy/rocky shores (84% of the MNI of marine shells) (Figure 5).This is evidenced by the numerical dominance of the mussel, the most frequently consumed shellfish on the site, which lives in the intertidal zone, down to 10m in depth along the coast.Mussel gathering is quite straightforward since these are readily visible and can be collected by hand or with a tool to cut the byssus, in clusters or individually.
Mussels can be consumed all year round but are fleshier during the reproductive period, between March and October (Poppe and Goto 1993).
Those mussels living in sandy to muddy/sandy shore substrates were more rarely exploited (11% of the MNI) (Figure 5).It is interesting to note that the spectrum of species coming only from sandy to muddy/sandy environment mostly consists of small individuals, with a shell size smaller than 20mm (Donax sp., Spisula sp., Abra sp., Macoma balthica).

From gathering to consumption
After being gathered, the shells were then transported to the site of Clos des Châtaigniers.Transport may have either been by boat, given that the River Orne was navigable during protohistory, or by land.The return trip would have taken at least a half-day.The large number of barnacles (5,927 specimens) and small shells found in shell deposits seem to indicate that the preparation of the shells was probably done on the site.
The barnacles were probably attached to the mussels and transported with them.It is likely that some of the barnacles detached from the mussels during the preparation or preservation phases.Traces of burning are evident on some mussels, representing on average 87% of NISP of the mussels.These marks are perhaps related to cooking methods or secondary activities occurring after consumption.Archaeological studies of burnt shell remains are not numerous in the archaeological literature.For this reason, we undertook research on the methods of cooking mussels and their effects on the shell, as described in the ethnohistorical literature, in order to gain some information (Waselkov 1987).Mussels can be cooked by placing them in the heart of a fire, in a steam oven (Best 1924, 417;Greengo 1952, 77;Duguet 1995, 367;Kroeber and Barrett 1960, 113;Meehan 1977, 366;Waselkov 1987, 101-2), placing them on hot stones (Terrell 1967, 44;Oberg 1973, 67) or roasting them around fires (Waselkov 1987, 101).Other hypotheses could also explain the high rate of burned archaeological mussels.After the mussels had been eaten, the shells could have been thrown into a fire.The shells could also have been used eventually to maintain the fire, or even to put it out.
This activity could reflect hygienic and/or marine waste management strategies.Nevertheless, it is difficult to select one of these possibilities given the lack of available bibliographical references.However, experimentation on shell material could make it possible to identify specific criteria for each type of cooking method, as has already been undertaken for bones (Costamagno et al. 2010;Lebon 2010;Zazzo 2010).The results obtained would provide some real answers regarding cooking methods and heat treatments and seems a promising area for future studies.
Previous research demonstrated, through experimental archaeology, that it was possible to dry a very large number of mussels in a relatively short period of time, with minimum effort, and using simple technology (Henshilwood et al. 1994).The most effective experiments in terms of time/calories related to the cooking and drying of mussel flesh inside the shell.Once cooked and dried, the flesh can be easily removed from the shell in only two or three seconds, so that fewer people and less time are needed to process large quantities of mussels.This method also reduces the need for firewood, since the mussels are cooked within around seven minutes.This technique eliminates the need for drying frames or posts and suspended lines, and minimises labour input (Henshilwood et al. 1994).
Nevertheless, we lack the archaeological data (organic residues in pottery, artefacts linked with preservation such as baskets, etc.) to state definitely that such a preservation technique was used at the site of Clos des Châtaigniers.The distribution of the shell accumulations seems to be confined to specific areas within the enclosure.Shells have only been found inside the ditch of this enclosure and in a single outside pit (F28) (Figure 1 and Figure 6).Four sedimentary samples have been extracted from inside this ditch (accumulations 1, 2, 7 and 8) and one in pit 28 (Figure 6).The quantification used for this distribution is the NISP expressed as a percentage.Inside the pit, accumulations and the percentage of mussel NISP are quite similar, demonstrating that this mussel was consumed in different parts of the site (Figure 6).The distribution of Balanus sp., however, shows an important difference.Barnacles represent between 1 and 2% of the NISP in pit 28 and in accumulations 1 and 2, as opposed to 8.5% in accumulation 8 and 23% in accumulation 7 (Figure 6).These barnacles, found in large quantities, might have become detached when the mussels were prepared or cooked, as previously mentioned.

Spatial distribution of archaeological remains
Considering the large quantities of barnacles, the preparation, cleaning and cooking of this shellfish could have taken place at the site, and perhaps in an area close to accumulations 7 and 8 (Figure 6).Burnt shell remains represent 60% of the total NISP and 87% of mussel NISP (Figure 7).The distribution of these remains is, however, heterogeneous at the site.The percentage of burnt mussel remains is significantly higher in the ditch of the enclosure (between 35 and 88% of the NISP) than at pit 28 (1.5% of the NISP) (Figure 8).This seems to indicate either a spatial organisation of the burnt mussel remains or differential treatment of the shells during or after their cooking.

The presence of marine shells as an indicator of the consumption of fish?
Numerous small marine shells (smaller than 11mm) have been found at the site.These remains are not burnt, unlike the great majority of mussels with which they were mixed in the ditch of the enclosure and in the pit (Figure 9).These shells do not seem to have been consumed by past humans (see section 4.1).In order to understand the origin of these   pollachius) or dogfish (Scyliorhinus sp.) (Figure 10) (Muus et al. 2005;Teletchea 2009).One should bear in mind that the presence of such remains could also result from the accumulation of dead aquatic birds or from faecal waste (Erlandson and Moss 2001;Van Leeuwen 2012).Thus, shell remains could also originate from animal stomach contents.It is important to note here that remains of three unidentified fish and one bird have been discovered at Clos des Châtaigniers.Analysis of bivalves other than mussels allowed us to understand the distribution of these small shells within the settlement (Figure 11).Pit 28 is unique, with a strong presence of Semelidae (44% of the NISP) while Tellinacea is almost absent (3% of the NISP).The distribution of Tellinacea, Pholadidae, Mactridae and Cardiidae is uniform in the ditch of the enclosure.By contrast, Semelidae are totally absent from accumulation 1.This distribution could potentially correspond to the preparation of fish or birds that have different diets.
Acanthocardia sp.Theodoropoulou (2007, 178) points out that, on some archaeological sites, small-sized marine shells could originate from stomach contents of mammals and aquatic birds or fish.However, she highlights the fact that it is very difficult to establish such an origin given the lack of reference studies.Van Neer and Pieters (1997) mentioned that, in a 15th-century coastal village in Belgium, remains of plaice (Pleuronectes platessa, 130 individuals) have been discovered in association with bean clam shells (Donax vittatus), interpreted as the stomach contents of these fish.
Likewise, in waste areas of the medieval castle of Boves (Somme), all the identified marine shells (8 at species level and 9 at genus level) originate from the preparation of fish used in human diet (Dupont 2005).
In order to answer this type of interrogation more precisely, it would be interesting to propose some criteria for the identification of aquatic animals from their stomach contents, through the development of studies and bibliographical research on the diet of these faunas.The results could demonstrate the consumption of other aquatic animals, whose skeletal remains were not preserved, marine shells being the only indicators of their initial presence.

Conclusion
In conclusion, the archaeomalacological study shows that the shell remains discovered at Clos des Châtaigniers represent mainly food waste.Despite the size and the diversity of the samples, mussels seem to have the seashore.This sorting would reduce the number of mussels to be transported, thus less effort was required.The presence of numerous barnacles may signify that the mussels were washed and prepared at the site of Clos des Châtaigniers.The high ratio of burnt mussel remains at the site could be associated with a particular cooking method or the management of marine waste.
The presence of common mussels (Mytilus edulis) is frequent in protohistorical sites in Normandy (Dupont 2006b;Mougne and Dupont 2012;Mougne et al. 2013;Mougne in prep).Therefore, the consumption of mussels at the site of Clos des Châtaigniers is not an isolated phenomenon but corresponds to an activity practiced in Normandy during the Bronze and Iron Ages.

Figure 1 :
Figure 1: Location and map of the archaeological site of Clos des Châtaigniers at Mathieu.(Image credit: D. Giazzon, L. Quesnel, C. Mougne)

Figure 3 :
Figure 3: Spectrum of the marine invertebrates at Clos des Châtaigniers

Figure 6 :
Figure 6: Spatial distribution of the marine invertebrates from the percentage of the NISP.(Image credit: D. Giazzon, L. Quesnel, C. Mougne)

Figure 7 :
Figure 7: Traces of burning on left valves of mussels (Mytilus edulis) in the shell accumulation no.7.(Image credit: C. Mougne)

Figure 8 :
Figure 8: Spectrum of the burned or unburned remains of mussels and barnacles by shell accumulations in percentage of the NISP.(Image credit: D. Giazzon, L. Quesnel, C. Mougne) shell fragments, they have been compared with other remains from modern mussel-beds.Bronze Age populations could have accidentally transported these small shell remains fixed to and mixed with collected mussels.Mussels originating from four modern-day coastal areas (Sangette, Nord-Pas-de-Calais; Saint-Honorine-des-Pertes, Basse-Normandie; Quiberon, Brittany and Lérat, Pays-de-La-Loire) located at the Channel and Atlantic coasts have been collected.These mussels were cleaned in a strainer and associated shell remains compared with archaeological remains.Fragment sizes in modern mussels vary (between 2 and 20mm) and these fragments show rounded outlines caused by marine erosion.By contrast, the sizes of archaeological remains identified at Clos des Châtaigniers are more uniform (between 5 and 10mm), and show angular and regular breaks.

Figure 11 .Figure 11 :
Figure 11.Spectrum of the bivalve remains (excepted the mussels) by shell accumulations.(Image credit: D. Giazzon, L. Quesnel, C. Mougne) been the only shellfish to have really played a role in the diet of local inhabitants during the Bronze Age.The choice of this shellfish can be explained as a result of easy access and abundance at the coastline of Normandy.Other marine invertebrates may have been brought to the site during the transport of mussels or represent the collection of empty shells.The study also reveals the consumption of other marine animals such as fish or birds, whose stomach contents contained small amounts of shells.The site is characterised by high numbers of mussels with medium to large sizes.Gathered in a rocky or muddy/rocky environment, they were then transported 10km inland.The absence of small-sized specimens of mussels might indicate prior selection during collection at Figure 10: Examples of fishes present in Normandy nowadays, consumed by men and who have a diet based partially on marine shells (+: consumption of marine shellfish; Muus et al. 2005; Teletchea 2009).