Smitten with middens
Dr Katrin Fenech (Research Support Officer)
Prof Patrick J. Schembri
Dr Nicholas Vella
Dr Katrin Fenech (Research Support Officer)
Prof Patrick J. Schembri
Dr Nicholas Vella
Research Questions
During the summer of 2014 FRAGSUS excavations at Taċ-Ċawla in Gozo, several very interesting deposits were discovered and sampled for environmental investigation. Eight environmental samples came from “gully” fills and seven from snail-rich deposits. The snail-rich deposits were of particular interest: they consisted of snail shells densely packed in pits and on excavation resulted to consist of several thousands of adult individuals, mainly the common red-banded snail Eobania vermiculata (Figs. 1-3). Covered with loose soil that also contained a few pottery sherds and some fragmented bones, the heavily soil-encrusted shells survived in a surprisingly good state. |
Although shell middens have been found in many archaeological sites in the Mediterranean and beyond, dating from the Palaeolithic up to Neolithic times, this may well be the first time that they have been found and recorded in the Maltese Islands, despite extensive excavations at numerous archaeological sites since the 19th century. To find out whether these snail-rich deposits are really middens, their characteristics have been compared to those of the “gully” fills.
Material and Methods
The environmental samples were taken between 9th April and 30th May 2014. The vast majority of samples were collected on site on April 22nd by Patrick J. Schembri and Katrin Fenech. All samples were weighed prior to wet-sieving to establish their individual weight and calculate the average weight of all samples combined for standardisation. Wet-sieving easily released the light organic material, which was caught in a 500μ sieve, while soil particles <1 mm were washed out. After drying, the flot was passed through graded sieves of 8 mm, 4 mm, 2 mm and 1 mm mesh sizes. All fractions were sorted for any contents other than stone/sand particles. The dried 500μ organic material was analysed under a stereomicroscope for shells, charcoal and plant remains.
The environmental samples were taken between 9th April and 30th May 2014. The vast majority of samples were collected on site on April 22nd by Patrick J. Schembri and Katrin Fenech. All samples were weighed prior to wet-sieving to establish their individual weight and calculate the average weight of all samples combined for standardisation. Wet-sieving easily released the light organic material, which was caught in a 500μ sieve, while soil particles <1 mm were washed out. After drying, the flot was passed through graded sieves of 8 mm, 4 mm, 2 mm and 1 mm mesh sizes. All fractions were sorted for any contents other than stone/sand particles. The dried 500μ organic material was analysed under a stereomicroscope for shells, charcoal and plant remains.
Results
The samples were treated as two separate groups: “gully” fills (eight bulk samples) and snail-rich samples (seven bulk samples).
Granulometry
Fig. 4 shows the percentage distribution of the different particle sizes from the “gully” samples and the snail-rich deposits. All the sediment samples analysed were well-sorted, and particles <1 mm predominate, varying between 60% (TCC14/83 BT3) and 87% (TCC14/100-1). Stones >8 mm were found in all deposits and were most prominent in TCC14/83 BT3, where they reached nearly 30%. The highest amount of stones in the snail-rich samples was 17%, while the lowest amount was in TCC14/100-1. The fractions >1 mm to >4 mm are fairly evenly present in each sample, varying between 3% and 9%.
The samples were treated as two separate groups: “gully” fills (eight bulk samples) and snail-rich samples (seven bulk samples).
Granulometry
Fig. 4 shows the percentage distribution of the different particle sizes from the “gully” samples and the snail-rich deposits. All the sediment samples analysed were well-sorted, and particles <1 mm predominate, varying between 60% (TCC14/83 BT3) and 87% (TCC14/100-1). Stones >8 mm were found in all deposits and were most prominent in TCC14/83 BT3, where they reached nearly 30%. The highest amount of stones in the snail-rich samples was 17%, while the lowest amount was in TCC14/100-1. The fractions >1 mm to >4 mm are fairly evenly present in each sample, varying between 3% and 9%.
Anthropogenic and biological contents
All samples contained an amount of anthropogenic and biological material, with pottery sherds being the dominant component in terms of weight in the “gully” samples and shells in the snail-rich deposits (Fig. 5). On removing the shells from the analysis of the snail-rich deposits, the anthropogenic contents of samples TCC14/95-2 to -4 and TCC14/100-1 were by far the least abundant (Fig. 5, bottom right). Bone was more frequently found in the “gully” samples than in the snail-rich deposits; the pieces were all fragmented and showed no sign of fire. One sample, TCC14/69 BT1, which also contained the highest amount of pottery sherds, contained a comparatively large amount of chert, flint and obsidian – material that was also found in much smaller amounts in most other samples from the fills. In the snail-rich samples, flint, chert and obsidian were only found in one sample (TCC14/63 SS28), Shells were present in all samples, but only in small amounts in the “gully” samples. Organic remains mainly consisted of root matter, very little charcoal and a few seeds.
All samples contained an amount of anthropogenic and biological material, with pottery sherds being the dominant component in terms of weight in the “gully” samples and shells in the snail-rich deposits (Fig. 5). On removing the shells from the analysis of the snail-rich deposits, the anthropogenic contents of samples TCC14/95-2 to -4 and TCC14/100-1 were by far the least abundant (Fig. 5, bottom right). Bone was more frequently found in the “gully” samples than in the snail-rich deposits; the pieces were all fragmented and showed no sign of fire. One sample, TCC14/69 BT1, which also contained the highest amount of pottery sherds, contained a comparatively large amount of chert, flint and obsidian – material that was also found in much smaller amounts in most other samples from the fills. In the snail-rich samples, flint, chert and obsidian were only found in one sample (TCC14/63 SS28), Shells were present in all samples, but only in small amounts in the “gully” samples. Organic remains mainly consisted of root matter, very little charcoal and a few seeds.
Molluscan analyses
By broad habitat categories:
All samples contained snails from each broad habitat category, but the standardized amount of land snails differs greatly between the “gully” fills and the snail-rich samples (Fig. 6). As expected, the ubiquitous group is the largest one in all samples, particularly in the snail-rich samples. These land snails do not have any very specific habitat requirements and are hence the least useful in the reconstruction of environments. There were more open country species than leaf-litter species in the “gully” fills, except in TCC14/93 BT1, where slightly more leaf-litter species were present, while in the snail-rich deposits leaf-litter species were more prominent.
The main open country species were Trochoidea spratti and Papillifera papillaris, while the leaf-litter species consisted mainly of Vitrea spp., indicating a possibly garigue environment with some leaf litter on the ground (possibly a high garigue).
By broad habitat categories:
All samples contained snails from each broad habitat category, but the standardized amount of land snails differs greatly between the “gully” fills and the snail-rich samples (Fig. 6). As expected, the ubiquitous group is the largest one in all samples, particularly in the snail-rich samples. These land snails do not have any very specific habitat requirements and are hence the least useful in the reconstruction of environments. There were more open country species than leaf-litter species in the “gully” fills, except in TCC14/93 BT1, where slightly more leaf-litter species were present, while in the snail-rich deposits leaf-litter species were more prominent.
The main open country species were Trochoidea spratti and Papillifera papillaris, while the leaf-litter species consisted mainly of Vitrea spp., indicating a possibly garigue environment with some leaf litter on the ground (possibly a high garigue).
Edible/non-edible snails:
Although all land snails are theoretically edible, they are not equally palatable. It is mainly the large helicid snails that are suitable for consumption as well as the large decollate snail Rumina decollata. The shells of these species have been reported as food remains from all around the Mediterranean, in archaeological deposits (e.g. Hunt in Barker et al. 1995, p. 151; Lubell 2004). However, they also occur in archaeological deposits in the Maltese Islands alongside smaller land snails as part of natural assemblages (e.g. Fenech and Schembri, in press).
Fig. 7 shows a stark contrast in the occurrence of potentially edible snails relative to non-edible snails found in the “gully” fills and the snail-rich deposits. In the “gully” fills, the number of edible snails is generally very small, and in most samples the majority belong to the non-edible category. Furthermore, all edible land snail species found in the snail assemblages included a large number of juveniles, indicating a natural assemblage. In the snail-rich samples, the large majority of snails are edible, and they are also present in significantly higher numbers as opposed to the “gully” fills assemblages. The high amount of non-edible species here is due to the presence of the abundant and ubiquitous Cochlicella acuta, which consisted of around 90% juveniles, and Vitrea spp., also with 55% juveniles, while an average of 98% of the edible land snails were adult species. Marine snails were very few in both groups of sample; in total, fragments of only seven specimens were found in the “gully” fills, of which only three species were edible (Patella sp., Phorcus sp. and Cerastoderma sp.), and four marine molluscs were found in the snail-rich deposits, two specimens each of Patella sp. and Phorcus sp., a number too small to be taken into consideration here.
Although all land snails are theoretically edible, they are not equally palatable. It is mainly the large helicid snails that are suitable for consumption as well as the large decollate snail Rumina decollata. The shells of these species have been reported as food remains from all around the Mediterranean, in archaeological deposits (e.g. Hunt in Barker et al. 1995, p. 151; Lubell 2004). However, they also occur in archaeological deposits in the Maltese Islands alongside smaller land snails as part of natural assemblages (e.g. Fenech and Schembri, in press).
Fig. 7 shows a stark contrast in the occurrence of potentially edible snails relative to non-edible snails found in the “gully” fills and the snail-rich deposits. In the “gully” fills, the number of edible snails is generally very small, and in most samples the majority belong to the non-edible category. Furthermore, all edible land snail species found in the snail assemblages included a large number of juveniles, indicating a natural assemblage. In the snail-rich samples, the large majority of snails are edible, and they are also present in significantly higher numbers as opposed to the “gully” fills assemblages. The high amount of non-edible species here is due to the presence of the abundant and ubiquitous Cochlicella acuta, which consisted of around 90% juveniles, and Vitrea spp., also with 55% juveniles, while an average of 98% of the edible land snails were adult species. Marine snails were very few in both groups of sample; in total, fragments of only seven specimens were found in the “gully” fills, of which only three species were edible (Patella sp., Phorcus sp. and Cerastoderma sp.), and four marine molluscs were found in the snail-rich deposits, two specimens each of Patella sp. and Phorcus sp., a number too small to be taken into consideration here.
Edible snail species:
Five different edible land snail species were found in the “gully” fills and in the snail-rich deposits. Fig. 8 shows the species composition of the edible land snails in both groups. While in the “gully” fills the most common species found was Theba pisana, followed by Cernuella caruanae, in the snail-rich deposits the dominant species was Eobania vermiculata, with the other species only forming a minor component of the assemblages. The edible species in the “gully” fills reached 70 specimens maximally; the minimum number of edible species found in the snail-rich deposits was over 500 (TCC14/63 SS28), reaching well over 2500 individuals in TCC14/100-1.
Five different edible land snail species were found in the “gully” fills and in the snail-rich deposits. Fig. 8 shows the species composition of the edible land snails in both groups. While in the “gully” fills the most common species found was Theba pisana, followed by Cernuella caruanae, in the snail-rich deposits the dominant species was Eobania vermiculata, with the other species only forming a minor component of the assemblages. The edible species in the “gully” fills reached 70 specimens maximally; the minimum number of edible species found in the snail-rich deposits was over 500 (TCC14/63 SS28), reaching well over 2500 individuals in TCC14/100-1.
Adult/juvenile ratio:
All deposits contained a mixture of adult and juvenile land snail species. The presence of juveniles in a deposit may point to a natural assemblage, but in the case of the samples analysed, there are several problems associated with the results. For one, the amount sampled is unlikely to represent the original assemblage accurately. Another factor is that the subfossil assemblages may incorporate several generations, and lastly, the mobility of the living animals or lack thereof, once they have died in the sediment (Evans, 1978). Nonetheless, important information may be gained from the results obtained.
Fig. 9 shows that there were more juveniles than adult individuals in both the edible and non-edible categories from the “gully” trench samples. The snail-rich samples show a different picture. Here, adult edible species by far outweigh the number of edible juveniles present. Furthermore, the size of the edible juveniles found, where often only the last whorl was missing, points to an accidental collection rather than a natural occurrence. Within the non-edible snails, again there were more juveniles than adult individuals, indicating that they were part of the soil that covered the snails.
All deposits contained a mixture of adult and juvenile land snail species. The presence of juveniles in a deposit may point to a natural assemblage, but in the case of the samples analysed, there are several problems associated with the results. For one, the amount sampled is unlikely to represent the original assemblage accurately. Another factor is that the subfossil assemblages may incorporate several generations, and lastly, the mobility of the living animals or lack thereof, once they have died in the sediment (Evans, 1978). Nonetheless, important information may be gained from the results obtained.
Fig. 9 shows that there were more juveniles than adult individuals in both the edible and non-edible categories from the “gully” trench samples. The snail-rich samples show a different picture. Here, adult edible species by far outweigh the number of edible juveniles present. Furthermore, the size of the edible juveniles found, where often only the last whorl was missing, points to an accidental collection rather than a natural occurrence. Within the non-edible snails, again there were more juveniles than adult individuals, indicating that they were part of the soil that covered the snails.
Pottery
Various amounts of pottery fragments were found in most of the samples analysed; only two samples (TCC14/95-4 and TCC14/100-1) contained no pottery at all. So far, only the material from the snail-rich deposit TCC14/63 SS28 has been analysed. It was found to contain a mixture of prehistoric pottery (27 sherds, including Ġgantija phase and possibly Żebbuġ phase) with later Punic/Roman material (15 sherds). TCC14/102 pit 3 contained six pottery sherds, of which one is probably Punic and clearly not prehistoric (pers. comm. Maria Elena Zammit, Heritage Malta). The pottery from the “gully” fills is likely to be a similar admixture.
Future Research
From the evidence available so far it would appear that the snail-rich deposits are escargotières and these may be the first ever land snail middens found and recorded in the Maltese Islands in an archaeological context. The people who created the middens had a clear bias to collecting the red-banded snail Eobania vermiculata, a species that is no longer eaten in the Maltese Islands, despite being very common and abundant. Eobania vermiculata is still sold for food in the markets of Crete (e.g. Welter-Schultes 1998) and collected for export in Turkey (Yıldırım et al. 2004). Interestingly, the only presently consumed indigenous snail in the Maltese Islands, Cantareus aspersus, is absent from the middens, which may be an indication that the middens pre-date the Roman period (when it is presumed that this species was re-introduced to the islands). The date of the middens is so far unknown, but the presence of Punic/Roman pottery sherds among the prehistoric pottery in TCC14/63 SS28 would point to a Punic/Roman date. It is hoped that the pottery analysis will help to date these middens more closely and perhaps also shed more light on the cultural background of the snail-eaters, whose food remains are as mysterious as their cultural identity.
Various amounts of pottery fragments were found in most of the samples analysed; only two samples (TCC14/95-4 and TCC14/100-1) contained no pottery at all. So far, only the material from the snail-rich deposit TCC14/63 SS28 has been analysed. It was found to contain a mixture of prehistoric pottery (27 sherds, including Ġgantija phase and possibly Żebbuġ phase) with later Punic/Roman material (15 sherds). TCC14/102 pit 3 contained six pottery sherds, of which one is probably Punic and clearly not prehistoric (pers. comm. Maria Elena Zammit, Heritage Malta). The pottery from the “gully” fills is likely to be a similar admixture.
Future Research
From the evidence available so far it would appear that the snail-rich deposits are escargotières and these may be the first ever land snail middens found and recorded in the Maltese Islands in an archaeological context. The people who created the middens had a clear bias to collecting the red-banded snail Eobania vermiculata, a species that is no longer eaten in the Maltese Islands, despite being very common and abundant. Eobania vermiculata is still sold for food in the markets of Crete (e.g. Welter-Schultes 1998) and collected for export in Turkey (Yıldırım et al. 2004). Interestingly, the only presently consumed indigenous snail in the Maltese Islands, Cantareus aspersus, is absent from the middens, which may be an indication that the middens pre-date the Roman period (when it is presumed that this species was re-introduced to the islands). The date of the middens is so far unknown, but the presence of Punic/Roman pottery sherds among the prehistoric pottery in TCC14/63 SS28 would point to a Punic/Roman date. It is hoped that the pottery analysis will help to date these middens more closely and perhaps also shed more light on the cultural background of the snail-eaters, whose food remains are as mysterious as their cultural identity.
References
Fenech, K. and Schembri, P.J. (in press) “Environmental analyses based on molluscan and other sedimentological remains,” in Bonanno, A. and Vella, N. (eds.) Tas-Silġ, Marsaxlokk (Malta) I: Archaeological Excavations Conducted by the University of Malta 1996-2005, Ancient Near Eastern Studies Supplement Series 48. Leuven: Peeters.
Hunt, C.O. (1995) in Barker, G., Hodges, R. and Clark, G., A Mediterranean valley. Landscape archaeology and annals history in the Biferno Valley. Leicester: Leicester University.
Lubell, D. (2004) Prehistoric edible land snails in the circum-Mediterranean: the archaeeological evidence. In Brugal, J.-J. and Desse, J. (eds.), Petits Animaux et Societés Humaines. Du Complément Alimentaires Aux Ressources Utilitaires. XXIVe rencontres internationals d’archéologie d’Antibes, pp. 41-62.
Welter-Schultes, F. W. (1998) Human-dispersed land snails in Crete, with special reference to Albinaria (Gastropoda: Clausiliidae). Biologia Gallo-hellenica 24 (2): pp. 83-106.
Yıldırım, M.Z., Kebapçı, Ü. and Gümüs, B.A (2004) Edible snails (terrestrial) of Turkey. Turkish Journal of Zoology 28: pp. 329-335.
Website design and maintenance: Gianmarco Alberti