A site for all seasons ? Prehistoric coastal subsistence in northwest Sicily

A site for all seasons? Prehistoric coastal subsistence in northwest Sicily Marcello Mannino & Ken Thomas There has been much debate about the role of marine resources, including shellfish, in prehistoric human diets, but there have been few studies that combine archaeological and present-day ecological investigations of coastal sites. This dual approach is being applied in a new fieldand laboratory-based project that focuses on a group of coastal sites in northwest Sicily.

A site fo r all seasons? Prehistoric coastal subsistence in northwest Sicily

Marcello Mannino & Ken Thomas
Th ere has been much debate about the role of marine resources, including shellfish, in prehistoric human diets, but there have been fe w studies that combine archaeological and present-day ecological investigations of coastal sites. Th is dual approach is being applied in a new fi eld-and laboratory-based project that fo cuses on a group of coastal sites in northwest Sicily. E vidence from stable isotope analysis1 ofhuman bones oflate Mesolithic date from a range of archaeological sites along the Atlantic coast of Europe indi cates that marine fo ods contributed signif icantly to the diet, in marked contrast to later Neolithic populations who evidently subsisted on terrestrial resources. The late Mesolithic individuals, whose bones were analyzed, probably ate mostly fish, with marine molluscs (shellfish) constituting a relatively minor part of the diet.2 This inference contrasts with evidence from the analysis of marine molluscs from the Mesolithic site of Culverwell in southern England, which suggests that they were exploited regularly, and sufficiently inten sively to have an impact on the local mol lusc populations; and analysis of the stable oxygen isotopes in shell carbonates of the top shell (Monodonta lineata), one of the principal species found at Culver well, suggested that the molluscs were exploited mainly during the autumn and winter. 3 Studies such as these have highlighted the potential significance in prehistoric subsistence of marine resources, although the available database of sites is limited mainly to the Mesolithic period and usu ally to individual sites, with the notable exception of studies of clusters of sites along the Cantabrian coastline of northern Spain. 4 There is abundant evidence of pre historic exploitation of coastal resources

Grotta • dell'Uzzo
in the Mediterranean region, but there too the work done has tended to focus on par ticular sites, such as Franchthi Cave in southern Greece and the Grotta dell'Uzzo in Sicily. 5 There is a need to take a broader view of prehistoric human coastal palaeo ecology, based on the study of subsistence change over time within well defined regions that contain many archaeological sites spanning the Upper Palaeolithic, Mesolithic and Neolithic periods. We have identified northwest Sicily as a region that meets these criteria.

The regional context
Northwest Sicily has one of the highest densities of prehistoric coastal sites in western Europe, and probably in the Medi terranean as a whole, and it offers enticing possibilities for the systematic examina tion of prehistoric human coastal adapta tion. Many of the coastal cave sites of northwest Sicily have been excavated, and most were found to contain abundant remains of shellfish, probably brought there for consumption. There are radiocar bon dates for a few of these sites, but most chronological sequences are based on stone-tool typologies.6 Several caves con tain stratigraphical sequences that span the period from the end of the Palaeolithic (Epigravettian) to the Neolithic, but in oth ers only one or two major cultural phases are represented. Apart from the stone arte facts and (in a few caves) notable rock art, and with the exception of the Grotta  dell'Uzzo, little is known about the pre historic occupants of the caves, or about their economy and environment. Excava tions in the past have yielded substantial quantities of material that have yet to be studied, among which are large collections of shells of marine molluscs.

Aims and methods of the project
Our project has two major components:7 • the analysis of assemblages of marine shellfish fr om archaeological sites • the ecological study of selected analo gous coastal localities. A first priority has been to select the sites to be studied, on the basis of reliable excavation data, the abundance and good state of preservation of shell materials, the archaeological period(s) represented, and the spatial relationships of sites. Four main clusters of sites have been selected and, within these clusters, key sites have been identified that we are either already stud ying or plan to study in the course of the project (Fig. 1, Table 1). The sites were cho sen because they have been reliably exca vated and because several of them have long sequences of occupation that span several cultural phases. One or two addi tional sites will be selected for study in each of clusters 2, 3 and 4, for comparison with the corresponding key site in each cluster. Because few of the sites have pre vious} y been dated by radiocarbon or other methods of absolute dating (Table 2), an important part of the project is to use suit ably preserved shells fr om important stratigraphical units at selected sites for accelerator mass spectrometry (AMS) radi ocarbon dating (which is capable of accu rately dating extremely small samples).
The archaeological assemblages of shells are being subjected to taxonomic and biometric analysis. Laboratory studies of shell mineralogy are also being under taken, to facilitate the selection of suitable shells from the assemblages for radio carbon dating and for the analysis of stable isotopes of oxygen. The oxygen isotopes will provide palaeoclimatic information, such as a record of variations of seawater temperatures, and also data on the seasons when molluscs were gathered at particular sites (Fig. 2). These investigations are com plemented by ecological, biometrical and isotopic studies of present-day local pop ulations of molluscs.
Two shore localities in northwestern Sicily have been selected for long-term study, one on the island of Favignana and the other on the Sicilian mainland at Monte Cofano (Fig. 3). These are visited monthly to study changes in mollusc pop ulations over two annual cycles. This includes recording species present, rela tive species abundance, changes in bio metrical variables and age profiles, and periods of recruitment to the populations (when the planktonic larvae settle on the shores and metamorphose into tiny juve nile molluscs). The species Monodonta  turbinata has been selected for intensive study at these localities and also in the archaeological material. It has been used in previous isotope studies8 and has been shown to record significant shifts in oxy gen ratios, which appear to be correlated with changes in sea water temperature. It is also a species known to be abundant in the archaeological assemblages that we are studying.
ecological questions and hypotheses to be addressed using the data collected fall into three groups: • What were the patterns of mollusc exploitation within each cultural phase, both within individual sites and within and between the clusters of sites? Sub sidiary questions that then arise are: (1) Might any differences in isotopic sea sonal patterns within and between clus ters suggest the presence of different social groups (with resource territories) in the whole region? (2) Do the molluscs from sites that have rock engravings or burials or both (i.e. sites where ritual activities may have taken place) show different patterns of shellfish exploita-lion, including seasonal differences, compared with other sites in the same cluster? (3) Do the data therefore lend support to the idea that such ritual sites could have been associated with partic ular social groups, whereby each local group in effect owned a ritual cave or caves? This inference would be re inforced if ritual sites across the whole region were found to have broadly similar patterns of seasonal use. Con versely, if the seasonality data suggest that ritual sites across the region were used at different seasons one might infer that regional patterns of ritual activity occurred independently of any local resource territories.
Live M. turbinata molluscs have been individually marked and released at both localities for the study of patterns of shell growth. Some of the molluscs will be col lected at the end of the ecological study, and shell-carbonate samples from them will be analyzed for fluctuations in stable carbon and oxygen isotope ratios over the known period of growth. The resulting data will reveal how accurate! y the isotope ratios track annual temperature changes, and therefore how indicative they may be of seasonal patterns of shell growth. In par allel with this long-term study of marked shells, other shells are being sampled at each monthly visit to investigate how stable-isotope ratios of oxygen and carbon in samples of shell-edge carbonates vary over annual cycles, for which purpose sea water temperatures are recorded monthly, and samples of sea water are taken for oxy gen isotope analysis and salinity determi nations. These various sets of data will help to calibrate the palaeo-isotope data for both palaeo-temperature reconstructions and estimates of season of death ofthe mol luscs in the archaeological assemblages.
The main archaeological and palaeo-  The results from these various investi gations are currently being gathered and analyzed, so it is premature to draw any specific conclusions at this stage. How ever, some preliminary results are availa ble and in the next section we present some of them from both the present-day season ality studies and fr om the Grotta dell'Uzzo.
Preliminary results from the seasonality studies Figure 4 shows the variation over one annual cycle in the ratio of the stable isotopes of oxygen (i.e. of 160 and 180; expressed as 8180%o) in shell-edge carbon ate samples from living Monodonta turbi nata, compared with the monthly changes in seawater temperatures, at one of the localities being studied, Monte Cofano. It is clear that the monthly 8180 values are strongly correlated with monthly sea water temperatures, suggesting that 8180 values in shell carbonates fr om archaeological specimens are likely to be good proxy indi cators of past seawater temperatures and also season of death of individual shells (i.e. when they were collected). out on shells from each of the main occu pation layers at the site, to yield a fr ame work of isotope values for each layer. These will be used to indicate general pat terns of change in seawater temperatures between layers (and, by inference, likely climatic shifts over time), as well as to provide for each layer an annual pattern of isotope values against which the shell edge isotopic values can be assessed for seasonality. Figure 6 shows the distribu tion of shell-edge isotope values in the shells analyzed so far from the Grotta  Fig. 4). Th e clustering in the higher part of the range of cr80 values shows that many shells were collected as fo od in the win ter period, wh ile those shells in the middle part of the range were collected in the autumn (as deduced fr om the cr80 values in the second and third samples taken back fr om the shell edge). No shells occur in the summer part of the range of cr80 values. Although interesting, these results must be regarded as preliminary, pending the analysis of a larger sample of speci mens.

Further research
T h e broad aims and specific objectives of this project are ambitious and it will prob ably not be possible for everything to be investigated, given the sheer quantity of material t h at exists and t h e time and resources available. T h ere is a need for fur ther analysis of material and data from t h e Grotta dell'Uzzo to be carried out before any firm suggestions can be made about the past seasonal patterns of s h ellfis h exploi tation at that site. In particular, the stable isotope database will be expanded consid erably by incorporating t h e unpublis h ed results obtained by Margaret Deith, who analyzed a large collection of shells from t h e Grotta deli'Uzzo at t h e Godwin Labo ratory (University of Cambridge) in t h e early 1980s. Work is also under way on material from several ot h er arc h aeological sites in t h e region, including the Addaura and Niscemi caves (Fig. 1), w h ere there are well known rock engravings. We are also continuing the ecological studies at the two localities selected for t h e monthly sur veys, and also at selected localities across t h e western Mediterranean part of t he present range of t h e targeted species, Monodonta turbinata. AMS radiocarbon dating of s h ells from t he Grotta dell'Uzzo is under way, and s h ells from other sites will be submitted for a furt h er 14 AMS radiocar bon dates, which will make a substantial contribution to the developing chronolog ical framework for Sicilian pre h istory. 9 Notes 1. The ratios of stable isotopes of nitrogen (expressed as o 15 N) and carbon (o 13 C) vary between organisms in marine and terres trial environments and, therefore, between terrestrial and marine food webs in which humans are consumers (as her bivores, carnivores or both). The varia tions in these isotopic ratios are complex and not always well understood, but at least in temperate regions o 13 C in marine food webs is generally significantly lower than in terrestrial ones, whereas o 15 N tends to be enriched in marine food chains, especially in animals (such as seals and, potentially, humans) that feed higher up those food chains. Analysis of 8 15 N and o 13 C in bone collagen from ancient human remains can therefore pro vide an indication of the proportion of marine and terrestrial foods in the diet of