Mesolithic middens and molluscan ecology : a view from southern Britain

H uman subsistence in the mesolithic period (c. 10,0005000 be) of western Europe has been characterized as broadly based. It is thought that people lived by hunting large game animals, fishing, gathering wild plant foods and foraging for other food resources, such as shellfish in coastal and estuarine environments. Large accumulations of the remains of shellfish, often in the form of shell mounds or middens, are a character­ istic feature of the coastal archaeological record for this period. Within Britain, the distribution of coastal midden sites is mark­ edly biased to the north, because uplift of the land in the north following the melting of the most recent ice sheet outpaced the post-glacial rise of sea level. For example, in both eastern and western Scotland, many mesolithic middens have been pre­ served.1 However, in southern Britain the post-glacial rise of sea level coincided with downwarping of the Earth's crust; as a result, many coastal areas were drowned, and it is likely that many mesolithic sites were likewise submerged. The shell mid­ den site of Westward Ho! in north Devon, which is accessible to archaeologists only at very low spring tides/ is an example of this process. On the high cliffs of the Isle of Portland (Dorset) there are shell-midden sites that are exceptional in southern Brit­ ain, being both well preserved and easily accessible for archaeological investigation. One such is the large midden site of Cul­ verwelP (Fig. 1 ) , which has been excavated over many years, although our investiga­ tions began there only in the summer of 1995.

which is accessible to archaeologists only at very low spring tides/ is an example of this process.On the high cliffs of the Isle of Portland (Dorset) there are shell-midden sites that are exceptional in southern Brit ain, being both well preserved and easily accessible for archaeological investigation.One such is the large midden site of Cul verwelP (Fig. 1), which has been excavated over many years, although our investiga tions began there only in the summer of 1995.

The Culverwell shell midden
Culverwell shell midden is located some 230m from the eastern cliffs of the island and at about 30m above mean sea level.The mesolithic deposits date, by radio carbon, to c. 7000 bp and are preserved under a medieval field system.Excava tions at the site by Susann Palmer since 1967 have revealed the presence of a very extensive shell midden.Its full dimen sions are not yet known, but trenches show that it extends over an area of at least 700m2• Although extensive, the midden is not very thick over most of its extent, except at the western end of the site where  it reaches 50-60 cm in thickness.Our work has fo cused on this thick midden, which is located in a deep underlying feature that may be a natural gully.In the summers of 1995 to 1997 we excavated a 1.3m-deep trench through the deposits in this area (Fig. 2) and revealed an upper sequence of deposits (known as colluvium) that had been moved down slope, sealing the dark midden beneath (Fig. 3).Column samples of midden material were collected and processed in the Wolfson Archaeological Science Laboratories at the Institute of Archaeology.The main objective of this work, in addition to the field investigation of the gully-like feature, was to gain an understanding of the composition of the midden, and especially to quantify the assemblages of shellfish (mollusc) remains in the various stratigraphic layers.This would in turn provide information about the range of species exploited (presumably for fo od); the nature of the coastal environ ments exploited by mesolithic foragers; and perhaps other information such as whether shellfish were exploited at one season or more frequently through the year and the possible impact of exploitation on the shellfish and other resources.

The shellfish remains
We recovered fr om the midden the shells of marine bivalves (two-shell molluscs such as mussels) and of marine snails (gastro pods) that had been collected in mesolithic times from both rocky and sandy shores, although gastropods from rocky shores were by far the most abundant in all layers of the midden.The three dominant species are limpets (Patella spp.), edible peri winkles (Littorina littorea) and the toothed top shell (Mon odonta lineata), followed by the dog whelk (Nucella lapillus).All these are intertidal species of rocky shores and they have overlapping distributions on the shore.Apart from molluscs, other marine resources recovered from the midden in clude many fragments of claws of the edi ble crab (Cancer pagurus), but only two fishbones.A few small highly fragmented pieces of mammal bone were also found.Thus, in bioarchaeological terms, the mid den contents are overwhelmingly dom inated by shellfish, especially gastropods.
The three most abundant species show significant changes in frequency through the mid den (Fig. 4).The decline in all three species through the samples from layer 8 and into layer 7, which together represent the most intensive phase of human occu pation at the site, is particularly interest ing.These changes could be interpreted as the result of environmental change, of the impact of human exploitation, of changes in human food choice, or possibly of more complex interactions of these factors.
As an explanation for the parallel decline in the three main mollusc species, environ mental change (involving either climatic change or local shore changes associated with rising sea levels) is unlikely for the following reasons: • Mon odonta lineata is a southern species near the northern edge of its range at Portland, yet the other two (more north erly) species decline by at least as much and sometimes more.
• If rising sea levels during the occupation of the site had led to shore morphologies similar to those of today, Mon odonta and Littorina would have declined, but Patella would have remained unaf fected.
• If rising sea levels had led to greater exposure of shores to wave action, Patella, again, should have been less affe cted and other species, such as the dog whelk, whose shell morphology is known to vary according to shore expo sure, might also reflect this.In fact, changes in the morphology of shells of dog whelk from the midden show that, if anything, shores became progres sively less exposed during this period.
If human exploitation had led to a decline in these species, one might expect this to result in changed age profiles in the mid den samples.Of the three main species only the age of Monodonta shells can be determined (by counting the annual growth lines).Age-profile histograms of this spe cies, for one series of samples (sample column 1) through the midden (the other two series are still being analyzed), show a progressive shift from older to younger age groups (Fig. 5).These results are consistent with the overexploitation resulting from  ---------------------� ---- -----------------� --� ----

ARC HAE OLOGY INT ERNATIONAL
foragers initially gathering the larger (there fore older) snails living on the shore, lead ing to a shift in the age distribution of the living shore population towards the younger age categories, which ultimately become the dominant age categories in the mid den.
Of course, it could be argued that the decline in abundance of the three gastropod species through the midden reflects a pro gressive shift away from shellfish and to wards other dietary items (although few other fo od resources were found preserved in the midden).The role of human food choice as a factor that might cause change in the representation of species is notoriously difficult to demonstrate archaeologically and therefore should not be dismissed out of hand.But the changes in abundance, cou pled with the progressive shifts in the age profiles of Monodonta described above, are difficult to account for solely on the basis of changing human fo od preferences.
We believe, therefore, that we have dem onstrated that, during the period of most intense occupation at the midden site, human foraging in the shore zone is likely to have had a significant impact on mol luscan food resources, causing at least three species to decline in local abundance.As yet, we do not know the time period rep resented by the series of samples, although radiocarbon dates on them are awaited, but it seems clear that exploitation of intertidal molluscs from the site was suffi ciently intensive, in scale and frequency, to prevent the local shore populations of them from reverting to their natural age structures.

Further research
Our investigations ofthe molluscs from the Culverwell midden have led us to under take research on four related topics that we hope will make valuable contributions to the archaeological study of shell middens, as well as to molluscan ecology: • Morphological analyses of the archaeo logical material, including changes in the average sizes of shells in different age classes over time.
o Analyses of the oxygen-isotope compo sition of shell carbonates of Mon odonta from the mid den, to gain palaeotempera ture data and evidence of season of col lection, based on the ratio of the lighter to the heavier oxygen isotopes, which varies with seawater temperature.
o Ecological studies of modern popula tions of Monodonta lineata to provide information on population age struc tures, recruitment of young individuals to the shore populations from the plank ton, and their movements and seasonal patterns of aggregation on the shore.o Oxygen-isotope analyses of modern shells, collected from various locations (at various times in the year) with known seawater temperatures, to provide a "calibration curve" for interpreting, in terms of past sea temperatures, the oxy gen-isotope ratios in the archaeological shells. 19

Conclusions
With much work still in progress, few firm conclusions can be reached at this stage.The most interesting conclusion to be drawn from the data collected so far is that human exploitation of shellfish resources on Port land may have been sufficiently intensive and frequent to have had significant im pacts on the local shore populations of molluscs.It will be interesting to find out whether these patterns of intensive exploi tation had a seasonal component -wheth er the shellfish were gathered at particular seasons or throughout the year.This, in turn, will yield insights into how people exploited intertidal resources in southern Britain, in comparison with other studies of shellfish exploitation in Britain and along the Atlantic coastline of western Europe. Notes

Figure 1
Figure 1 Location of the Isle of Portland and the Culverwell midden site.

Figure 3
Figure3Vi ew of the stratigraphic se quence in the Culverwell mid den, show ing, below the dashed white line, the darker shell-rich mid den deposits beneath the overlying ligh ter colluvial deposits (scale bar 15cm).

Figure 5
Figure 5 Ch anges in the age profiles of the top shell {Monodonta lineata) through the mid den deposits atCulverwell (sample column 1]from the top of layer B (LB) down to layers 9-12; n = the number of shells analyzed in each sample.