Journal of the North Atlantic M. Law and N. Thew 2015 Special Volume 9 125 Introduction Dune systems with wind-blown shell sands occur at regular intervals along the whole of the west coast of Britain. Archaeological remains have often been found within these sands, associated with wellpreserved assemblages of land snails. In the Western Isles and the Northern Isles of Scotland, these dune systems, known as the machair, were the subject of a number of informative studies of sub-fossil molluscan assemblages in the 1970s and 1980s, notably by John Evans (1971, 1979, 2004), Penny Spencer (1975), Michael Vaughan (1976), Annie Milles (1991), and Nigel Thew (2003), from which a number of themes began to emerge. John Evans’ work also had a considerable influence on subsequent investigations by Paul Davies and Martin Bell of land snails in coastal deposits in South Wales and South West England (e.g., Bell and Johnson 1990, Davies 2008). Since the 1990s, however, work on assemblages of terrestrial molluscs from Scottish dune contexts has largely ceased. John Evans did return briefly to working on material from Sligeanach in the Outer Hebrides in 2003, which was included in an important paper on the relationship between two common taxa (Evans 2004); sadly, the complete study was only recently published posthumously (Evans et al. 2012). This paper is intended to reawaken an interest in the analysis of sub-fossil land snails associated with archaeological sites in the sand dunes of northern Britain, by revisiting and highlighting some of the patterns that emerged from the earlier work, proposing new lines of enquiry, and assessing how these ideas might be applied to Hebridean archaeology today. Themes discussed will include the use of land snail assemblages as a biostratigraphic tool that can help in the relative dating of layers and structure infills, and information that molluscan faunas can provide about the local environment. Indeed, the analysis of series of snail assemblages from a number of contexts that represent both site stratigraphy and spatial variation across a site, allows conclusions to be drawn about changes in the local environment and land use through time and about site-formation processes. In turn, the comparison of data from a number of sites scattered across the Western Isles should help to develop a larger picture of environmental change (Fig. 1). Land snails thus represent an excellent palaeoenvironmental indicator that can provide high-quality information for archaeological investigations of machair sites. The Taphonomy of Hebridean Snail Assemblages Before the remains of terrestrial snails from any context can be interpreted, the taphonomic processes that have transformed populations of live molluscs into assemblages of sub-fossil shells have to be understood. Land snails from machair contexts can be considered to be more or less autochthonous (Evans 1972, 1979). This determination depends upon whether they come from buried soils indicative of surface stability, in which case the sub-fossil molluscan faunas will represent the immediate local environment, or from deposits of wind-blown sand indicative of unstable surfaces, where the shells could come from a much wider catchment area. Rapidly accumulating wind-blown deposits may also be subject to a high degree of temporal mixing, as shells can be eroded from much older deposits during blowouts. In addition, some allochthonous shells may be introduced into machair deposits in plant material brought in for thatching, fodder, or manuring or carried in by flood waters from nearby freshwater marshes and lakes (Smith 1994, Thew 2003). Land Snails, Sand Dunes, and Archaeology in the Outer Hebrides Matt Law1,* and Nigel Thew2 Abstract - Although the Western Isles have been subject to a number of recent archaeological investigations, there has been limited recent work on molluscan assemblages, despite the very good degree of preservation to be expected in a number of deposits and the significant work in the 1970s and 1980s on a number of sites such as Northton, Baleshare, and Hornish Point. In the meantime, land snail analysis has flourished in southern England and elsewhere in Europe, with the development of new techniques of numerical analysis such as the taxocene framework, the use of land snail assemblages in climate reconstruction, and recent refinements in amino acid dating. This paper provides a brief summary of the work to date on Hebridean snail assemblages, and presents preliminary results from work in progress, exploring aspects of site-formation processes, middening, land use, and relative dating that can be explored using land snails. Special Volume 9:125–133 2010 Hebridean Archaeology Forum Journal of the North Atlantic 1School of History, Archaeology, and Religion, Cardiff University, John Percival Building, Colum Drive, Cardiff CF10 3EU, UK. 2Neuchâtel, Switzerland. *Corresponding author - LawMJ@cf.ac.uk. 2015 Journal of the North Atlantic 126 M. Law and N. Thew 2015 Special Volume 9 Another problem with the interpretation of snail assemblages is the relationship between the shells and the layers in which they are found. Thus, layers interpreted as ploughsoils may include shells indicative of stable, shaded conditions, because the non-intensive nature of ploughing with ards until post-Viking times left pre-existing shells reasonably intact. Even in modern times, ploughs in the Outer Hebrides are often set to shallow to preserve the root mat and maintain soil stability in the face of frequent wind erosion (Barber 2011:50). Furthermore, the basal part of wind-blown sands often includes shells indicative of more stable conditions eroded from the underlying layers, while the summit of wind-blown sand deposits that are succeeded by a buried soil, often include shells brought down from the overlying soil by earthworms. In practice, however, most published profiles show a reasonably coherent ecological sequence, which suggests that many of these potential problems do not impair the integrity of land snail data sets (Davies 2008:131). Furthermore, taking series of spatially separated sample columns and spot samples across a site will reduce problems of mixing or redeposition by making aberrations in the molluscan sequence easier to detect (Thew 2003). Biostratigraphy and Relative Dating Studies of blown-sand deposits have revealed that a number of land snail species have arrived in western Britain during the course of the last part of the Holocene, since the start of the Neolithic (Davies 2008:130, Evans 1979:20, Thew 2003:163). These include Cernuella virgata, Cornu aspersum (Helix aspersa), Cochlicella acuta, and Helicella itala. Recent arrivals in coastal dune systems include Candidula intersecta, Cochlicella barbara, and Theba pisana. In the Western and Northern Isles, the arrival and then the expansion of Cochlicella acuta and Helicella itala provide clear biostratigraphic markers that can act as a relative dating tool (Thew 2003). The precise dating of the arrival and expansion of these two species is still rather uncertain, though a number of radiocarbon dates from sites such as Northton, Hornish Point, and Baleshare (Fig. 1) seem to suggest that Helicella itala appeared near the end of the Late Bronze Age or early in the Iron Age, and Cochlicella acuta arrived during the earlier part of the Iron Age. Both species became widespread and much more abundant at some stage during the Iron Age. This relative dating tool will be improved as more assemblages from sites with 14C dates are studied. The expansion of Helicella itala and Cochlicella acuta seems to have caused a major decline in several other species (including Pupilla muscorum, Cochlicopa spp., Vallonia spp. [especially Vallonia costata], and Cepaea hortensis) by competing with them for available ecological niches (Thew 2003:167). The use of amino acid (racemization) dating might also be used to establish relative chronologies with a good resolution, especially in sites or layers where there is little material available for 14C dating, although this method still needs to be developed in Figure 1. Map showing key sites mentioned in the text. the context of machair sites. Journal of the North Atlantic M. Law and N. Thew 2015 Special Volume 9 127 Site Formation Processes: Stability, Flux, and Grazing As machair systems develop, intervals of instability cause wind-blown sands to bury earlier land surfaces. Thus, profiles through machair deposits typically include layers of clean wind-blown sand over more organic, buried soil surfaces, with ironstained deposits beneath these organic horizons that have also been affected by pedogenetic processes (Fig. 2). Erosion surfaces are often present, which coincide with hiatuses in the depositional sequence. Most of the numerical variations within molluscan assemblages are thus attributable to differences in the original populations coupled with the rate of layer accumulation. Sand dunes and blown sand are, by their nature, rather unstable habitats, which tend to have restricted molluscan faunas. Interpretation of the land snails therefore depends upon discerning variations among faunas of low diversity that indicate environments with a greater or lesser degree of stability, herbaceous cover, moisture, and sand accumulation (Thew 2003:163). Differences could also be due to human activities such as ploughing, grazing, the spread of fertilizer (including seaweed), or the deposition of rubbish (middening). Ancient ploughsoils tend to be thick, rather homogenous, fairly organic layers, caused by physical mixing together with the addition of organic material. Associated snail assemblages have a tendency to resemble the more varied faunas found in buried soils, but with higher frequencies of taxa favored by middening (Thew 2003:169-171). Most specimens from dune deposits have lost their periostracum, which is the colored protein layer that covers the surface of the shell (Evans 1972:22). In rare cases, however, the periostracum may be preserved where rapid burial has taken place and the sandy deposits have remained both dry and undisturbed, as in some of the sample assemblages from Mound 2A at Bornais (M. Law and N. Thew, unpubl. data). Following its arrival in the Outer Hebrides, probably during the Early Iron Age, Cochlicella acuta came to dominate the more unstable contexts on the machair, as they were able to flourish among tall, deep-rooted marram grass (Ammophila arenaria) Figure 2. Buried organic layer with cultural material sealed by wind-blown sand, Baleshare, September 2010. Journal of the North Atlantic 128 M. Law and N. Thew 2015 Special Volume 9 that is adapted to dune habitats with accumulating sand (Thew 2003:167). When strong winds cause the sand to drift, they climb the stems of the marram grass and wait till conditions improve. Cochlicella acuta, together with Helicella itala, has also become the principal species of short-turf grazed machair grassland in the Western Isles (Thew 2003:167). Before the expansion of Cochlicella acuta and Helicella itala, blown-sand contexts were often dominated by low frequencies of Pupilla muscorum, which seems to have been able to survive in the rather dry, exposed, unstable conditions, where most other species could not (Thew 2003:170). It was much more abundant, however, in relatively stable conditions associated with short-sward grazed grassland, accompanied by Vallonia excentrica and Vallonia costata (Thew 2003:170). Greater surface stability is associated with greater biodiversity, with snail populations being influenced by different grazing regimes, which include varying intensities of grazing by sheep and/or by cattle. In especially stable, rather short-turf, fairly moist, sheep-grazed grassland, for example, peaks in Vallonia excentrica are accompanied by a number of less common species, including Vertigo pygmaea, Punctum pygmaeum, Euconulus fulvus, Vitrina pellucida, Vitrea crystallina, Nesovitrea hammonis, and Oxychilus alliarius (Thew 2003:170). In modern dry grassland habitats, Pupilla muscorum prefers open sward lacking in litter accumulation, whereas Vallonia excentrica, Vallonia pulchella and Vallonia costata favor a denser sward and the presence of leaf litter, conditions that allow Vertigo pygmaea, Punctum pygmaeum and Euconulus fulvus to also be present (Jakupec 1997). Vallonia costata, Cochlicopa spp., and Lauria cylindracea are pioneer species that are able to flourish in situations where the biotope is changing. Such suitable conditions include both situations where there is moderate surface instability, with broken ground, as long as there is some humidity, pioneer vegetation, and limited sand accumulation, and also environments where there is increasing shade due to the development of denser vegetation such as tall herbs and possibly scattered shrubs (Thew 2003:170). These three species can sometimes take advantage of middens and the presence of walled structures too (Fig. 3). Peaks in these taxa often Figure 3. Lauria cylindracea on an upturned fence post, South Uist, September 2010. Journal of the North Atlantic M. Law and N. Thew 2015 Special Volume 9 129 coincide, sometimes in conjunction with Vitrina pellucida (Thew 2003:170). Relationships between certain key species may thus be especially revealing. In a review of studies from wind-blown sand contexts from throughout the British Isles, for example, Evans (2004) suggested that the relationship between Pupilla muscorum and Lauria cylindracea may provide a useful indicator of prehistoric farming practices. Pupilla was encouraged by the relative stability of short-turf grazed grassland, while Lauria, by contrast, was favored by changing conditions, such as the instability associated with broken ground caused by more intensive land use, by more shaded conditions after land was abandoned or left fallow, and by the effects of middening or the formation of “plaggen” soils. Unfortunately, over the last 2000 to 1000 years, Lauria has extended its ecological range into more open grazed grassland and dune habitats in much of northern and northwestern Britain, replacing Pupilla muscorum to a considerable extent and thus making environmental reconstruction more difficult. In the Orkneys, for example, which were never colonized by Cochlicella acuta, Lauria has replaced Pupilla as the main snail species of machair grassland (Cameron 2002, Evans 2004:371–372). To summarize, variations in snail assemblages reflect differences in surface stability, vegetation cover, and dampness in the terrestrial environment surrounding habitation sites on the machair, which in turn reflect differences in grazing regimes, precipitation, wind activity, and the altitude above the water table. In the area immediately surrounding the walled structures of habitation sites, new species associations occur that reflect human activities such as middening or the spreading of seaweed, in addition to rupestral habitats offered by walled structures. These associations will be discussed below. Work by Evans (1971) at Northton, South Harris, also revealed evidence for previously existing woodland in the Outer Hebrides, which despite partial Neolithic and Bronze Age clearance, persisted into the Iron Age before finally disappearing. This persistence has been confirmed by pollen analyses across the Western Isles, which showed that birch-hazel woodland, with smaller frequencies of pine, oak, and elm must have been locally present in many parts of the islands at the start of the Neolithic (Brayshay and Edwards 1996). As Thew (2003) has noted, phases with shaded conditions, either open woodland or tall herbs mixed with scattered shrubs, are usually associated with species assemblages consisting of at least 15 non-marsh land snail species that include true shade-demanding taxa. At Northton, the woodland phases are marked by the presence of up to 22 non-marsh species (Evans 1971), and at Buckquoy in Orkney by up to 20 non-marsh taxa (Evans and Spencer 1977). In comparison, counts of non-marsh species at Baleshare, Hornish Point, Newtonferry, and Balelone are lower than 15 in all samples, suggesting rather open environments (Thew 2003). Similarly, the highest species count for non-marsh taxa for Mound 2A at Bornais, is 12 (M. Law and N. Thew, unpubl. data). Low frequencies of marsh and freshwater aquatic species were present at Baleshare, Hornish Point, Newtonferry, Balelone, and also at Bornais (M. Law and N. Thew, unpubl. data; Thew 2003). These results are probably the result of winter flooding from nearby marshes and freshwater lakes due to rising water tables (Ritchie 1979), though as the water table at these sites is fairly high, patches of marshy ground may well have persisted in hollows between the dunes. The evidence for episodic winter flooding at these sites raises questions about site location and site-formation processes. Some of these sites may only have been used seasonally, while strategies may have had to be developed to cope with seasonal or sporadic flooding at other permanently occupied locations. Evans (2004) has speculated, following ideas advanced by Pollard (1996), that midden sites in the Scottish islands may have been created deliberately as visible monuments. The reality of the situation may have been far more prosaic, however, with the build up of mounds being encouraged to avoid winter flooding. Site-Formation Processes: Middening, Walled Structures, and the Spreading of Seaweed At Baleshare, Hornish Point, Balelone, and Newtonferry, Thew (2003:170) reported peaks (from ~10% to 20% of total molluscs) in the relative frequency of Oxychilus alliarius in contexts where domestic organic refuse had been dumped, although several stratigraphic sub-blocks with concentrations of archaeological refuse and a fairly high organic content had no concomitant rise in the numbers of Oxychilus at Baleshare. Although Oxychilus alliarius can be found in low numbers in stable machair grassland habitats (Paul 1976), high frequencies of this species, often associated with Vallonia costata, were noted from the island of Flat Holm (Bristol Channel) in habitats with a high organic content and the presence of much decomposing flesh (Young and Evans 1991). It seems likely, therefore, that peaks in Oxychilus alliarius may reflect the deposition of fresh domestic waste near Journal of the North Atlantic 130 M. Law and N. Thew 2015 Special Volume 9 to the walls of the byre. The byre and an abandoned kiln also revealed shells from species such as Lauria cylindracea, Cochlicopa spp., Punctum pygmaeum, and Nesovitrea hammonis, which had taken advantage of the locally shaded conditions. Climate Data from Land Snail Assemblages Although there have been a number of significant climatic changes over the last 6000 years, since the beginning of the Neolithic (Magny 1995), none of these seem to have influenced the presence or absence of molluscan species in the Outer Hebrides. This lack of correlation with climate can be attributed to the fact that all of the snail taxa recovered from the machair deposits have geographical ranges that extend well beyond these islands, so the limited temperature changes that took place during the second half of the Holocene seem unlikely to have affected their presence or absence. Instead, the occurrence of non-marsh land snails has been largely determined by environmental changes linked with human activities. Nevertheless, as Dawson et al. (2011) have shown, increased deposition of wind-blown sand can be correlated with periods of climatic downturn, such as the Little Ice Age; thus, the frequency and thickness of deposits of wind-blown sand in sites across the Western Isles may give indications of intervals with a poorer climate. Moreover, as Evans (2004) has pointed out, molluscan evidence of freshwater flooding, caused by fluctuations in the height of the water table, or of significantly damper surface conditions, represents a good indirect indicator of climatic changes that affected the precipitation regime. Numerical Methods for Modelling Ecological Change The application of numerical techniques to subfossil land snail assemblages has not been fully explored, although two particular types of technique are reasonably common in snail reports, and are thus worthy of comment here. The first of these is the use of diversity indices. The diversity of an assemblage can be measured simply by counting the total number of taxa present; revealing patterns do emerge from these simple counts, especially if used in conjunction with indicator species such as the presence of true shade-demanding taxa. More sophisticated is the use of mathematical diversity indices, which assess the evenness of spread of species throughout a series of assemblages. The use of these indices is suspect, however, because in addition to the diversity the center of a site, while an absence of such peaks in contexts with concentrations of archaeological material and a relatively high soil organic content might represent rubbish being removed from near the center of a site and redeposited nearer the edge of a settlement. Alternatively, if fresh waste was rapidly buried by new refuse or by wind-blown sand, Oxychilus may not have had time to colonize. As Vallonia costata is also well known from gardens and rubbish-midden locations (Evans 1972), sudden peaks in this species might also, in some cases, indicate the dumping of domestic rubbish. Although Clausilia bidentata can occasionally be found in relatively damp, tall, stable, grassland, or grassland over rocks, this species is normally rupestral and as such probably reflects the nearby presence of standing walls or perhaps isolated bushes or trees (Paul 1976, 1992). Some peaks in Lauria cylindracea, which often lives in rupestral habitats (Paul 1976, 1992), may also be linked with the nearby presence of stone built structures, as seems possible for the northern stratigraphic blocks at Iron Age Baleshare (Thew 2003:170). This association becomes more likely if there are no concomitant peaks in the pioneer species Vallonia costata or Cochlicopa spp. The use of seaweed for animal fodder and as a fertilizer that also stabilizes soil surfaces is wellattested historically, including in the Outer Hebrides (Smith 1994). Marine snails are often introduced with the seaweed (Bell 1981:121), as seems to have been the case at Hornish Point, Baleshare, Balelone, and Newtonferry (Pain and Thew 2003), as well as at Bostadh Beach (Cerón-Carrasco 2005:32) and Bornais (M. Law and N. Thew, unpubl. data). The seaweed snails at Newtonferry were often most frequent in contexts with low numbers of land snails indicative of middening, suggesting that seaweed was primarily brought in as a fertilizer away from the center of the site, possibly in areas that were cultivated. Even in locations where snails may not be expected to be conserved in the archaeological record, such as the acidic “blacklands” (peaty areas behind the machair), certain socioeconomic practices may create conditions suitable for their preservation. This was the case at Howmore, South Uist, where Smith (1994) attributed the presence in the byre of snail species typical of machair grassland, to the cutting of turves from the machair for use as animal bedding, although turves could equally well have been used for roofing material. She also attributed shells of the synanthropic Oxychilus cellarius to the presence of organic material and of rupestral taxa Journal of the North Atlantic M. Law and N. Thew 2015 Special Volume 9 131 tive associations of snails in sub-fossil assemblages, and provide a method for identifying patterns in the land snail data without resorting to analogies drawn from the modern ecology of the species present. For obvious reasons, taxocenes can only be deduced from contexts dominated by autochthonous shells. The taxocene framework has not so far been applied to Hebridean snail assemblages, primarily because of a relative lack of data from appropriate contexts, such as buried soils. Moreover, it is unlikely that the same taxocenes identified in southern Britain will occur on the islands. Ultimately, however, as Evans (2004:366) has stated, numerical methods often mask finer details. This insight would suggest that the use of numerical methods is appropriate as a means of exploring data, but much depends on the taphonomic variables affecting individual assemblages. Pielou (1975:34) has argued that working back from statistical hypotheses to ecological hypotheses is a very difficult task. In sub-fossil assemblages, where the contextual information are much less complete, the difficulties of drawing palaeoenvironmental, palaeoecological, and archaeological inferences is even greater. Conclusions Land snails are sensitive indicators of site histories and of changes in land-management regimes. The ways they respond to changes in the vegetation cover, to agricultural practices such as grazing and ploughing, and to changes in surface humidity linked to climate change are recurrent from site to site. Land snails can also act as indicators for archaeological site-formation processes, such as middening, the spread of seaweed, or the presence of nearby walled structures. Within such structures, the snail faunas can give some idea of their previous function and possibly of construction materials such as the use of turves from the machair for roofing. Although land snail assemblages can be affected by a number of taphonomic processes, particularly in the unstable context of sand dunes, analyzing multiple samples distributed spatially across a site, the application of statistical techniques, and an understanding of the changes that occur in molluscan sequences in response to different depositional environments, can help to distinguish allochthonous and autochthonous elements in the snail assemblages and to thus interpret faunal variations in terms of archaeoenvironmental events. Further work is needed to establish a precise chronology for the arrival and spread of new snail species in the Western Isles in order to strengthen and density of live populations, linked to habitat suitability, assemblages are also subject to a number of taphonomic variables, including the concentration or dispersal of shells during transport, the rate of sediment accumulation, and differential preservation of shell material linked to physical and chemical conditions post deposition. In a review of the use of diversity indices in archaeology, Ringrose (1993) states that their use is not appropriate for inter-site comparisons where there are likely to be taphonomic differences, and Ken Thomas (Institute of Archaeology, University College London, UK, pers. comm.) adds that the same is true for intra-site comparisons where taphonomic biases are suspected. Diversity indices are thus most useful for highlighting taphonomic issues, but not for drawing palaeoecological conclusions. In his analysis of the infills of two treethrow hollows from Ascott-under-Wychwood, for example, Evans (2005) used the Shannon diversity index to show that the higher number of species in the upper fill was at least partially a reflection of a greater frequency of shells. By using a combination of rank-order curves of species abundance and Fisher’s alpha diversity measures, Kenward (1978) has been able to distinguish between autochthonous and allochthonous insect taxa in urban assemblages, with important consequences for the interpretation of the contexts from which they derived. This same technique could be used for land snail assemblages. The second major type of numerical analysis applied to sub-fossil terrestrial molluscs consists of ordination techniques, which allow data to be explored on the basis of the relative abundance of species in a set of samples. These are descriptive rather than explanatory techniques, so their effectiveness is dependent on subjective interpretations. Detrended correspondence analysis has been the most commonly used for molluscan assemblages, as seen in work by Bush (1988), Davies (1998), and Peacock and Gerber (2008). This technique organizes the data into a set of axes which account for different percentages of variation between samples. This variation is usually explained in terms of environmental factors such as moisture, pH, or shade. The influences of taphonomy, however, should also be considered, as should biogeography in the case of assemblages from islands. Detrended correspondence analysis also can be used to identify taxocenes, although this can be done more simply but subjectively by examining molluscan results tables and diagrams. The concept of taxocenes was applied to dryground and wet-ground assemblages in central southern England in the 1990s by Evans (1991). Taxocenes are used to define recurring and distincJournal of the North Atlantic 132 M. Law and N. Thew 2015 Special Volume 9 their use as a relative dating tool and to detect changes in the representation of pre-existing taxa in response to adventive species. Gathering snail data from larger numbers of archaeological excavations in the machair area will enable further refinements in the degree of sensitivity with which molluscan responses to environmental and anthropogenic changes can be interpreted. Literature Cited Barber, J. 2011. Characterising archaeology in machair. Pp. 37–54, In D. Griffiths and P. Ashmore (Eds.). Aeolian Archaeology: The Archaeology of Sand Landscapes in Scotland. Scottish Archaeology Internet Reports 48. Society of Antiquaries of Scotland, Edinburgh, Scotland, UK. Available on line at http://www. sair.org.uk. Accessed 17 May 2011. Bell, M. 1981. Seaweed as a Prehistoric Resource. Pp. 117–126, In D. Brothwell and G. Dimbleby (Eds.). Environmental Aspects of Coasts and Islands Symposia of the Association for Environmental Archaeology No. 1. BAR International Series 94. British Archaeological Reports, Oxford, UK. Bell, M., and S. Johnson. 1990. Mollusca and Other Zoological Evidence. Pp. 246–252, In M. Bell (Ed.). Brean Down: Excavations 1983–1987. English Heritage, London, UK. Brayshay, B., and K. Edwards, 1996. Late-glacial and Holocene vegetational history of South Uist and Barra. Pp. 13–26, In D. Gilbertson, M. Kent, and J. Grattan (Eds). The Outer Hebrides: The Last 14,000 Years. Sheffield Academic Press, Sheffield, UK. Bush, M.B. 1988. The use of multivariate analysis and modern analogue sites as an aid to the interpretation of data from fossil mollusc assemblages. Journal of Biogeography 15(5/6):849–861. Cameron, R.A.D. 2002. The land mollusks of North Ronaldsay, Orkney: Human intervention and island faunal diversity. Journal of Conchology 37:445–453. Cerón-Carrasco, R.N. 2005. “Of Fish and Men” (“De iasg agus dhaoine”): A Study of the Utilization of Marine Resources as Recovered from Selected Hebridean Archaeological Sites. BAR British Series 400. Archaeopress, Oxford, UK. 258 pp. Davies, P. 1998. Numerical analysis of subfossil wetground Molluscan taxocenes from overbank alluvium at Kingsmead Bridge, Wiltshire. Journal of Archaeological Science 25:39–52. Davies, P. 2008. Snails: Archaeology and Landscape Change. Oxbow, Oxford, UK. 199 pp. Dawson, S., A.G. Dawson, and J.T. Jordan, 2011. North Atlantic climate change and Late Holocene windstorm activity in the Outer Hebrides, Scotland. Pp. 25–36, In D. Griffiths and P.Ashmore (Eds.). Aeolian Archaeology: The Archaeology of Sand Landscapes in Scotland. Scottish Archaeology Internet Reports 48.Society of Antiquaries of Scotland, Edinburgh, UK. Available online at http://www.sair.org.uk Accessed 17 May 2011. Evans, J.G. 1971. Habitat change in the calcareous soils of Britain: The impact of Neolithic man. Pp 27–74, In D.D.A. Simpson (Ed.). Economy and Settlement in Neolithic and Early Bronze Age Britain and Europe. Leicester University Press, Leicester, UK. Evans, J.G. 1972. Land Snails in Archaeology. Seminar Press, London, UK. 436 pp. Evans, J.G. 1979. The palaeo-environment of coastal blown-sand deposits in western and northern Britain. Scottish Archaeological Forum 9:16–26. Evans, J.G. 1991. An approach to the interpretation of dry-ground and wet-ground Molluscan taxocenes from central-southern England. Pp 75–91. In D.R Harris and K.D. Thomas (Eds.). Modelling Ecological Change. Institute of Archaeology, London, UK. Evans, J.G. 2004. Land snails as a guide to the environments of wind-blown sand: The case of Lauria cylindracea and Pupilla muscorum. Pp 366–379, In A. Gibson and A. Sheridan. (Eds.). From Sickles to Circles: Britain and Ireland at the Time of Stonehenge. Tempus, Stroud, UK. Evans, J.G. 2005. The snails. Pp 55–70, In D. Benson and A. Whittle (Eds.). Building Memories: The Neolithic Cotswold Long Barrow at Ascott-Under-Wychwood, Oxfordshire. Cardiff Studies in Archaeology. Oxbow, Oxford, UK. Evans, J.G., and P. Spencer 1977. The Mollusca and environment, Buckquoy, Orkney. Pp. 174–224, In A. Ritchie (Ed.). Excavation of Pictish and Viking-age Farmsteads at Buckquoy, Orkney, Proceedings of the Society of Antiquaries of Scotland 108. Evans, J.G., M. Law, and N. Thew 2012. Stability and flux in the dune environment. Pp. 250–253, In N. Sharples. The Beaker-period and Early Bronze Age settlement at Sligeanach, Cill Donnain, In M. Parker Pearson (Ed.). From Machair to Mountains: Archaeological Survey and Excavation in South Uist. Oxbow, Oxford, UK. Jakupec, M. 1997. Detection of the main ecological factors influencing the compositions of snail communities inhabiting dry grasslands in Lower Austria. Pp 111– 112, In G.Falkner (Ed.) Contributions to Palaearctic Malacology, Heldia 4(5). Kenward, H.E.K. 1978. The Analysis of Archaeological Insect Assemblages: A New Approach. York Archaeological Trust, York, UK. Magny M. 1995. Une Histoire du Climat, des Derniers Mammouths au Siècle de L'automobile. Errance, Paris, France. Milles, A. 1991. The Molluscan biostratigraphy and archaeology of Holocene coastal blown sand in the British Isles. Unpublished Ph.D. Thesis. Cardiff University, Cardiff, UK. Pain, C., and N. Thew 2003. The microscopic marine Mollusca. Pp 173–177, In J. Barber (Ed.). Bronze Age Farms and Iron Age Farm Mounds of the Outer Hebrides. Scottish Archaeological Internet Reports 3,. Available online at http://www.sair.org.uk/sair3/. Accessed 17 May 2011. Paul, C.R.C. 1976. The non-narine Mollusca of Colonsay and Oronsay. Journal of Conchology 29:107–110. Journal of the North Atlantic M. Law and N. Thew 2015 Special Volume 9 133 Paul, C.R.C. 1992. The non-marine Mollusca of Ulva, Inner Hebrides. Journal of Conchology 34:175–178. Peacock, E., and J. Gerber 2008. Using land snails and freshwater mussels to chart human transformation of the landscape: An example from North Mississippi, USA. Pp 123-142, In E.J. Reitz, C.M. Scarry, and S.J. Scudder (Eds.). Case Studies in Environmental Archaeology, 2nd Edition. Springer, New York, NY, USA Pielou, E.C. 1975. Ecological Diversity. Wiley, London, UK.176 pp. Pollard, T. 1996. Time and tide: coastal environments, cosmology, and ritual practice in early prehistoric Scotland. Pp. 198–210, In T. Pollard and A. Morrison (Eds.). The Early Prehistory of Scotland. Edinburgh Academic Press, Edinburgh, UK. Ringrose, T.J. 1993. Diversity indices and archaeology. Pp. 279–285, In J. Andresen, T. Madsen, and I. Scollar (Eds.). Computing the Past CAA92. Aarhus University Press, Aarhus, Denmark. Ritchie, W. 1979. Machair development and chronology of the Uists and adjacent Islands. Proceedings of the Royal Society of Edinburgh B 77:107–122. Smith, H. 1994. Middening in the Outer Hebrides: An ethnoarchaeological investigation. Unpublished Ph.D. Thesis. University of Sheffield, Sheffield, UK. Spencer, P.J. 1975. Habitat change in coastal sand-dune areas: The molluscan evidence. Pp 96–103, In J.G. Evans, S. Limbrey, and H. Cleere (Eds.). The Effect of Man on the Landscape: The Highland Zone. CBA Research Report 11. Council for British Archaeology, York, UK. Thew, N. 2003. Chapter 14: The molluscan assemblage. Pp. 163–177, In J. Barber (Ed.). Bronze Age Farms and Iron Age Farm Mounds of the Outer Hebrides. Scottish Archaeological Internet Reports 3. Available online at http://www.sair.org.uk/sair3/. Accessed 17 May 2011. Vaughan, M.P. 1976. Environmental change in areas of blown sand on the western coasts of the British Isles: Land snail analysis from sites in Ireland, Pembrokeshire, Somerset, the Outer Hebrides, and the Orkneys. Unpublished B.A. Dissertation. University College Cardiff, Cardiff, UK. Young, M.S., and J.G. Evans. 1991. Modern land mollusc communities from Flat Holm, South Glamorgan. Journal of Conchology 34:63–70.