We are grateful to Chris Hunt and Jörg Linstädter for their useful comments, and for their English revision of this text.

Introduction

1The terrestrial malacofauna from Capsian and Neolithic sites in the eastern Maghreb shows the dominance of a few species, which are present in great numbers. The occupants of these sites mainly targeted adult individuals, since this maximized the mollusc flesh gathered. In addition to the common species, other, rarer taxa are found. These are here termed minority species and they are defined as comprising <5 % of the molluscs in an assemblage. Young molluscs were also identified at all sites (Saafi 2019; Saafi et al. 2013, 2021a).

2Studies of the terrestrial malacofauna of prehistoric sites in this region, are rare and few are new. Most focus on reconstructing climatic data (Roubet 1979; Lubell et al. 1982-83). In some, researchers have tried to determine the place of land snails in the diet of human groups (Morel 1953, 1974; Lubell 2004a and b) but these accounts are not very detailed, especially concerning the distribution of gastropod age classes (adults or youngs), or the presence of minority taxa. Further, it is unclear whether minority or young taxa really contributed to the diet of the occupants of these sites, and if they were eaten, the magnitude of their contribution. For example, without an age distribution for a species the accurate estimate of the mass of meat consumed on a site cannot be computed. If consumption was not the main objective of their collection, however, different probable hypotheses to explain this phenomenon must be established.

3In our work, we study the presence of minority species and young land snails from several Capsian and Neolithic escargotière sites in Tunisia and Algeria. The results of this study will allow us to attempt to discuss the contribution of this part of the fauna to the diet of the occupants of the sites. They will also help to answer the questions mentioned above.

1 | The rammadiyet sites

4Our study area is part of the Eastern Maghreb, comprising Tunisia and Eastern Algeria. Eight sites are located in Central and Northern Tunisia and there is one site near El Eulma in Algeria (fig. 1). The rammadiya at Aïn Metherchem (Tunisia) belongs to the Typical Capsian (~ 8200 - 6200 cal. BC). At El Mekta (Tunisia), two successive occupation phases were discovered: Typical Capsian and Upper Capsian (~ 6800 and 6200-5500 cal. BC). Six other sites belong to the Upper Capsian. These are Sabkhet Halk el Menjil (SHM-1), Kef Ezzahi, El Oghrab, Aïn Oum Henda 1, Aïn Charchara (in Tunisia) and Medjez I (in Algeria). The Neolithic, dated to the first half of the 6th millennium, is represented by a single site: Kef el Agab (Northwestern Tunisia). This type of site is known as rammadiya (plural: rammadiyet: ash) or escargotières. Large quantities of ash and terrestrial shells are the two main components of this type of site. In addition, we find faunal remains, fragments of ostrich eggs, flint, charcoal and stones, sometimes burned...

2 | Methods

5The study material comes from excavations of prehistoric sites carried out recently within the framework of cooperative projects between different Maghrebian and international research institutes: National Heritage Institute, Tunisia; National Center for Prehistoric, Anthropological and Historical Research, Algeria; University of Bologna, National Research Institute, Italy. Some of the archaeological data used in this paper are briefly presented in previous studies. This concerns the sites of Kef Ezzahi, Aïn Oum Henda 1 and El Oghrab in Tunisia and Medjez I in Algeria (Saafi et al. 2021a and b, 2023).

6The identification of terrestrial shell species found in prehistoric sites is based on bibliographic references (Ktari, Rezig 1976; Abbes et al. 2009, 2011; Bouaziz-Yahyatene et al. 2017; Ezzine et al. 2017). Taxon naming updates are based on WoRMS (World Register of Marine Species; website: https://www.marinespecies.org). The determination of MNI is quantified essentially on the presence of the apex and/or the peristome of the shell. If both of these parts are present in an assemblage, the one with the higher number will be the MNI. In the absence of these two elements (the apex and peristome), the choice was made to count the samples whose size exceeds ¼ of the shell volume. Following the results obtained for each site, we can classify the identified species into two categories: majority taxa and minority taxa. By definition, minority species have a very low presence in an archaeological site. In most cases, their proportion does not exceed 5 % of the total terrestrial shells of each malacological taxon in each site. According to the ethnographic study, the size of the shell of mollusc is one of the criteria of selection of the preferred species among Tunisian consumers. We distinguish between two types. Mollusks with large shells (e.g. Helix melanostoma or Cornu aspersum) are preferred by some people (to have more meat). Medium-sized snails such as Eobania vermiculata are preferred by other consumers. These have a more delicious taste than the first type of mollusks. We can add a third type, that of small mollusks, such as Sphincterochila candidissima and Theba pisana. Tunisian consumers do not collect them because of their small flesh mass (Saafi 2019).

7For young molluscs, the distinction from adult gastropods is based mainly on shell size (measurement: length and height), and the development of the peristome which is not well formed in a young specimen (thin and fragile). Thanks to the count, we can determine the rate of young molluscs in each site, which helps us to explain their role for the occupants of the prehistoric sites studied.

3 | Results

3.1 | Minority species

8The number of minority species is limited to two taxa in the sites belonging to the Typical Capsian: Aïn Metherchem and El Mekta (T. C.: Typical Capsian). This number increases during the Upper Capsian and Neolithic (between 4 and 7 species in each site) (fig. 2). While minority species represent only half or less of the number of taxa identified in Typical Capsian sites, they predominate in number during the two most recent periods. For example, they represent six species in the Kef Ezzahi site (total species = 7) (fig. 2). In other rammadiyet of Eastern Algeria, there are three minority taxa in the Medjez II site (Camps-Fabrer, 1975) and 15 taxa in the Dra-Mta-el-Ma-el-Abiod site, near Tebessa (Morel 1974). In the Capeletti cave of Khanguet Sidi Mohamed Tahar, a Neolithic site of Capsian tradition, no minority species were reported (Roubet 1979).

9The presence of shells of minority species relative to majority taxa in a Capsian or Neolithic site is always low. With a low number (the case of Aïn Metherchem (fig. 3.1) or high (the cases of Kef Ezzahi (fig. 3.2) and Kef el Agab (fig. 3.3), their MNI rate does not exceed 10 % except in the rammadiya of SHM-1 (11.6 %) (fig. 4).

3.2 | Young molluscs

10The age distribution of molluscs suggests regional characteristics. In the Tunisian sites, the percentage of young snails is low. Except in the SHM-1 (13.4%) and El Oghrab (11.2 %) escargotières, it does not exceed 10 % of the total terrestrial malacofauna in the other sites (fig. 5). The importance of young molluscs is remarkable in the only Algerian site studied, Medjez I, where their rate is equal to 31.7 %. This high rate is due to the importance of youngs in Xerosecta sp., one of the two majority taxa (with Helix melanostoma) in this site, with youngs specimens accounting for 78.6% of the MNI of this taxons (fig. 5) (Saafi et al. 2023).

4 | Discussion

4.1 | Mollusc nutrient contribution to the diet of Capsian and Neolithic populations

11The contribution to the diet of human groups provides a distinction among minority taxa. We can distinguish between two groups: edible species and non-edible ones. The first group consists of mainly large shells such as Helix melanostoma and Cornu aspersum (fig. 6). This type of molluscs will provide consumers with more flesh, but nutrient intake will be very low from the taxa in this first group due to their limited MNI at each site.

12The second group concerns small-shelled molluscs. Xeroplana idia (Issel 1885) is one such. It is identified in the site of Kef Ezzahi (fig. 3.2; fig. 6). The height of this species (between the umbilicus and the apex) is about 6 mm (Manganelli et al. 1997). The small size of the shells of these molluscs as well as their very low MNI make it difficult for them to have contributed appreciably to the diet of human groups in the region.

13The contribution to the diet of human groups by young molluscs is possible only at Medjez I. We count 1329 young individuals of Xerosecta sp. (Total MNI = 1689 shells). Significantly, the taphonomic study shows the presence of traces of burning and anthropogenic perforations on some of these youngs (Saafi et al. 2023). This confirms their contribution to the nutrition of the occupants of Medjez I, albeit minor (MNI of all identified species = 4812 shells).

4.2 | Hypotheses on the presence of minority species and young individuals in the study sites

14The presence of a very low quantity of terrestrial shells of minority species or young individuals could be linked to certain hypotheses which will be developed below. We can classify them into four types.

4.2.1 | Methodological (collection method)

15Ethnographically, the collection of land snails is often subject to precise criteria, especially concerning the choice of the species and the age of the mollusc (young or adult). Currently, in Tunisia, each human group or family has a single preferred taxon, for example, Helix melanostoma. The expectation is that only adult land snails will be collected (Saafi 2019, 2022). Based on the ethnographic study, an inexperienced collector may, however, collect unwanted species or young molluscs. It is possible that the presence of minority taxa or young gastropods in a Capsian or Neolithic site followed collection by an inexperienced collector. The plants harvested by the occupants of a site could be contained some mollusks that take refuge there.

16A collection by necessity could also be a reason for the presence of minority or young specimens. The lack of food resources seasonally, or during extended famines, can force people to extend and diversify their normal diets, i.e. look for abnormal food resources. In this case, the collection of any other possible nutritional resource becomes an obligation. Terrestrial malacofauna are known to have been used as famine foods. Thus, the contribution of molluscs to human diets became more important in some regions of Tunisia during the crises of the 1940s (Saafi 2019; A. Brahmi, pers. comm.). In this context, the high rate of young individuals of Xerosecta sp. at Medjez I might be related to the temporary absence of other food resources (Saafi et al. 2023). Similar arguments have been advanced for the collection of very small coastal molluscs during the Epipalaeolithic and Little Ice Age in Northeast Libya (Hunt et al. 2011).

4.2.2 | Archaeological factors

17The dimensions of some rammadiyet are large, such as at El Oghrab (50×30 m) or at Aïn Oum Henda 1 (45×30 m) (Saafi et al. 2013). In some sites, just a test-pit (1×1 m) were carried out. The limited area excavated would naturally limit the samples obtained and thus might contribute to the relatively low MNI of some species, where distributions of taxa in the deposit are uneven. At other sites where the excavated area is larger, effects of spatial patterning on the relative proportions of taxa would be nullified. At Medjez I (area excavated: 9 m²), in stratigraphic subunit II209, there are 28 Helix melanostoma shells in square N43, whereas there is a higher in MNI in square N45, with 1293 individuals of the same taxon (fig. 7; Saafi et al. 2023). Even though this is a majority species, there may be very low MNI in some areas of the site. The limit of the excavated area may sometimes result in very low numbers of one or more taxa, but proof of this is difficult to demonstrate without wider excavation.

18The topography of the site directs the disposition of the refuse at a given location and thus concentration or dispersion. The removal of empty shells after consumption to a more or less flat area gives a concentration of waste on a rather limited surface. In some sites, the dump is located in a slope as is the case in El Mekta or Kef el Agab. The slope is about 45° in the rammadiya of El Mekta. This element leads to the dispersion of empty shells over a large area of the site. In some areas of the deposit, the number of shells of a given species could be restricted.

19Mollusc shells are relatively mobile but fragile sedimentary particles. After consumption, the shell becomes lighter. It could be transported by rain or strong wind from one area of the site to another. Shells may also be kicked from routeways across the midden by passers-by, or may be crushed underfoot. And thin mollusc shells may be calcined and disintegrate at relatively low temperatures by burning, for instance in the substrate of camp-fires.

4.2.3 | Ethnographic factors

20Some taxa could be used in the preparation of certain remedies for specific diseases. This has been documented ethnographically in Vietnam (Rabett et al. 2011). Others may have been used as ochre containers (Rumina decollata at Medjez II: Camps-Fabrer 1975). There is a long history of the use of molluscs as personal adornment (for instance Vanhaeren et al. 2006). Molluscs have also been documented as accompanying the deceased (e.g. Kurzawska 2010).

4.2.4 | Ecological and natural factors

21Some molluscs lived in the vicinity of the site. They were adapted to the climate and environment of the excavated area. For example, during the excavation of the site of Kef Ezzahi in December 2014, live shells of Xeroplana idia were found around the rammadiya (Saafi 2019). The presence of sometimes juvenile individuals of Rumina decollata in some sites (Oued el Akarit, south Tunisia; Iberomaurusian site) confirms the hypothesis of a natural presence since ethnographic data suggests that this species is rarely eaten (only during periods of crisis (famines)). It could be reduced to colonization by local populations of land snails (mating place) (Lubell 2004b). Some scavenger’s species (Natalina cafra for example) come to consume organic matter from prehistoric dumps or abandoned shells of their conspecifics for calcium (Appleton and Heeg 1999; Cadee 1999; Law and Thew 2015).

Conclusion

22Minority species and young molluscs are present at all sites surveyed. The number and list of taxa involved varies from one escargotière to another. The MNI rate of all minority taxa is low compared to the majority species. Only the larger molluscs may have contributed to the nutrition of the site occupants. As for young individuals, the MNI of Xerosecta sp. shells as well as the presence of anthropic traces (systematic burning and perforations) at Medjez I could be explained by a collection by necessity in order to remedy the temporary absence of certain foods. Other hypotheses could lead to the presence of these two types of molluscs in the sites concerned.

23In our future research, we plan to extend our study area to include the rest of North Africa while studying older collections. This would allow us to better interpret the archaeological data concerning this issue.