The Mesolithic

Martin Bell and Graeme Warren with Hannah Cobb, Simon Fitch, Antony J Long, Garry Momber, Rick J Schulting, Penny Spikins, and Fraser Sturt

Described by Mithen (1999) as being the period of British prehistory most in need of new research, the Mesolithic represents half of the postglacial (Fig 2.1) but it has arguably received significantly less research attention than any other period in British archaeology, particularly with regard to maritime themes. Thus, while in the last 30 years the archaeology of the Palaeolithic has been totally transformed, stimulated in part by the discovery of well-stratified coastal sediment sequences at Boxgrove, so far there has only been geographically patchy attention to the coastal and riverine sequences which have so much potential for Mesolithic research.

The Prehistoric Society (1999) recently celebrated advances in hunter-gatherer archaeology but it is notable that many of their examples are in Scotland (eg Mithen 2000) and Ireland (Woodman et al 1999) rather than England. Moreover, the broad interdisciplinary projects that they advocated have since produced outstanding rewards in the Palaeolithic, but have been less developed in the Mesolithic. The British Mesolithic still tends to be dominated by the lacustrine edge site at Star Carr (Clark 1954; 1972), the focus of significant ongoing work and deteriorating preservation conditions (Conneller et al 2009; 2010; Milner et al 2011), and by lithic scatters on dry land, albeit many of them clustered near the coast (Wymer 1977). As important as these sites are, from as early as 1976 Clarke urged investigation of wetland contexts in southern Britain to fill fundamental gaps in our knowledge of this period. Whilst the potential for exceptional organic and settlement site preservation in wetland or submerged contexts is clearly shown by recent work in Denmark (Andersen 1985; 2009), the Netherlands (Louwe Kooijmans 2001a; 2001b; Peeters 2007), Belgium (Crombé 2005), and Ireland (Mossop 2009; McQuade and O’Donnell 2007; 2009), more than 30 years after Clark’s comments it has yet to be fully realised in England. As such, the archaeology of the British Mesolithic stands poised as a period rich in potential but still heavily in need of further research.

Figure 2.1 Timeline for Chapter 2 Temperature anomaly from Davis et al 2003; sea-level data from Rholing et al 2009

Maritime, wetland and coastal archaeology in a Mesolithic context

In the British Isles the Mesolithic is generally defined as beginning with the Holocene and ending with the appearance of the Neolithic. This provides a date range from 9700 cal BC (Walker et al 2009) to c 3800 cal BC (Whittle et al 2011), the later date being somewhat contentious, regionally specific, and often rounded to c 4000 cal BC. Of course, considerable continuities link the Mesolithic and its preceding and following periods. Notably in the post-Last Glacial Maximum settlement of northern Europe, the ebb and flow of human occupation of the northern European lowlands, including England, was closely related to climate change, made extensive use of now flooded landscapes, and showed considerable continuity between the ‘final’ Palaeolithic and the earliest Mesolithic, with, for example, technical and typological links between Ahrensburgian, Federmesser, and early Mesolithic lithic industries (De Bie and Vermeesch 1998, 39).

The coast, especially marine resources, is  considered to be a key determinant of hunter-gatherer settlement patterns in Europe. Based on broad ethnographic observations and ecological principles, Simmons argues that ‘coasts exert a very strong pull force in terms of available resources, to the point where no society would ignore them unless prevented by other human groups from gaining access to them’ (Simmons 1996, 194). It is important to note that such suggestions (eg Bonsall 1981) are often based on assessment of the resources likely  to have been available during particular seasons, rather than on a detailed examination of the biological or sedimentary evidence. Given that our current understanding of the scales of territoriality, mobility, and exchange for the Mesolithic is limited, it is hypothetically possible to argue that almost all aspects of Mesolithic archaeology in England may have been influenced by marine factors. Indeed, assessing the degree to which coasts did influence Mesolithic settlement is one of the central research challenges for the period. As such, some parameters are required to put coherent limits on this chapter. The modern political construct of England is meaningless in the context of a radically different early Holocene geography that saw Britain linked to continental Europe. A great deal of relevant archaeological and palaeoenvironmental material is now underwater or found in neighbouring European countries. For the purposes of this discussion, sites that were located in areas of direct marine influence during the Mesolithic, including land now reclaimed and sites beneath modern sea level, are included. Other sites with direct evidence of the exploitation of marine resources are discussed as appropriate. Figure 2.2 shows the key sites discussed.

Figure 2.2 Map of key coastal Mesolithic sites discussed in Chapter 2 Basemap data derived from GEBCO 08 (www gebco net)

Mesolithic research has to take particular account of dramatic environmental changes over a timespan of c 10,000 years. This applies to understanding site location, the resources that would have been available, and predicting the probable locations of buried or submerged sites. Research in this period inevitably draws on a wide range of scientific disciplines including oceanography, marine ecology, geophysics and geology, as well as the whole array of environmental and geoarchaeological techniques. Each discipline conducts research on different scales relevant to available data and disciplinary research questions. A particular challenge in understanding the topography and landscape of the Mesolithic is, therefore, identifying appropriate spatial and temporal scales of analysis and critical use and meaningful integration of diverse data sets. Research needs to be appropriate both to understanding the broad-scale evolution of the topography, coastline and environments through time, but also how that landscape and its changing nature would have been encountered at a human scale.

Inherent in this is consideration of the time depth of a Mesolithic community’s environmental knowledge from oral historical sources (including no doubt song, dance and art), their spatial knowledge (derived from individual and group movements and communication with other groups), and the overall rate of change. Our knowledge of these issues is limited. Little or no serious research has investigated concepts of time in Mesolithic Britain and whilst models of Mesolithic seasonal movement proposed by Clark (1972), Jacobi (1979), Simmons (1996), Barton et al (1995), and Bell (2007) exist, these are not universally accepted (see Spikins 2000). Most of these models envisage movement from the coasts  to upland exploitation in summer. Evidence from Ireland, Wales and elsewhere in Europe from human bone isotopic analysis is beginning to challenge this movement by identifying some individuals with a mainly coastal-based diet, and others with a mainly terrestrial diet. Some researchers are also questioning the concept of coastal/inland movement on the basis of typological contrasts between lithic assemblages in north-west (R Cowell, pers comm) and north-east England (P Spikins, pers comm). Change over time in patterns of movement also remains rather obscure, and we must be careful not to allow static models to dominate a diverse record.

Elsewhere in Europe considerable investment in maritime archaeology has been made, and has led in Denmark to spectacular Mesolithic discoveries on submerged sites (Andersen 1987; 2009; Skaarup and Grøn 2004). Despite the extent of landscape submergence, Bouldnor Cliff is the first permanently submerged Mesolithic site in UK waters that has been subject to archaeological and palaeoenvironmental investigation, demonstrating the potential of the submerged heritage resource (Momber et al 2011). However, most of the Danish sites which have been extensively investigated are submerged relatively shallowly in calm, Baltic waters with a limited tidal range. This stands in stark contrast to a large proportion of the British coast.

Broad research issues

In order for our understanding of the Mesolithic   to move forward there is a general need for the following:

  • Increased absolute dating of sites, assemblages and environmental sequences to help refine understanding of change through time.
  • Additional isotopic and dietary analysis to resolve further the degree to which coastal resources influenced Mesolithic settlement patterns and ways of life.
  • Greater communication and collaboration with Quaternary scientists to aid multiscalar palaeoenvironmental and archaeological interpretations.
  • Investigation of how the changing nature of the sea and seaways over this period affected Mesolithic communities.

Theme 2 1: Coastal change

2.1.1 Mesolithic sea-level change

The defining characteristic of the Quaternary period is the cycle of glacials and interglacials that alternately locked up large volumes of water on land in extensive ice sheets before releasing it back into the oceans under warmer climates. The most recent glacial termination, which began c 18,000 BC and resulted in rapid sea-level rise until c 5000 BC, was the fastest and most sustained rise in sea level in at least the last 120,000 years (Lambeck et al 2000; Yokoyama et al 2000a; Clark et al 2009). Evidence for the rise in global sea level is recorded most precisely by coral reef and mangrove deposits now submerged in the Bonaparte Gulf, north Australia (Yokoyama et al 2000b), the Sunda Strait, Java-Sumatra (Hanebuth et al 2009), in the waters surrounding Tahiti (Bard et al 1996) and Barbados (Fairbanks 1989; Peltier and Fairbanks 2006), and in the raised corals of Papua New Guinea (Edwards et al 1993). Rapid sea-level changes and shifting shorelines were the norm during the early–mid- Holocene across much of the globe.

The rise in global sea level was not a smooth, steady process (Fig 2.3). Generally rates were fastest during the initial period of rapid ice sheet melt, decaying to slower rates after c 8000 BC, but the rise was interrupted by several jumps caused by meltwater pulses (MWPs), sudden steps in sea level caused by the collapse of melting ice sheets in the northern and southern hemispheres (Fairbanks 1989). The first widely recognised jump, MWP1A, saw a rise in sea level of 16–24m between 12,600 and 11,500 BC. The second (MWP1B) involved a rise of up to 28m, this time dated to the Younger Dryas, c 10,900–9700 BC. A third, smaller, jump of c 1–3m was probably caused by the final drainage of the glacial meltwater lakes that surrounded the former Laurentide Ice Sheet in North America at c 6200 BC (Törnqvist et al 2004; Hijma and Cohen 2010). The actual magnitude of these jumps varied around the world. For example, MWPs sourced from Antarctica would have had a larger impact on the British Isles compared with ones from the Laurentide Ice Sheet of North America because of the global readjustment in the gravity field resulting from the transfer of significant mass from land to sea.

Figure 2.3 Sea-level curve 20,000–0 BP showing meltwater pulses (MWP) (data derived from Gornitz 2007)

Around the British Isles, the global rise in Mesolithic sea level was moderated by regional-scale changes in land uplift and subsidence, notably due to the British and Irish Ice Sheet and the much larger Fenoscandanavian Ice Sheet (Shennan 1989; Shennan and Horton 2002; Shennan et al 2006). In general terms, land uplift in northern England partly offset the rise in sea level. In southern England, subsidence, due to the collapse of a peripherial bulge that once surrounded these ice sheets, caused sea level to rise at a rate equal to, or slightly faster than, the global value (Lambeck 1995a).

The rising sea levels of the Late Glacial and Mesolithic flooded the continental shelf that surrounds the British Isles and caused the progressive isolation of the mainland from Ireland and then continental Europe (Fig 2.4). A combination of field data and geophysical modelling suggests that a depression in the floor of the Irish Sea was the first region to be inundated by tidal waters, with Ireland separated from Britain at c 12,000 BC by a narrow channel (eg Devoy 1995; Wingfield 1995; Lambeck 1995a; 1996a; Uehara et al 2006). Geological evidence of this initial inundation is fragmentary and reflects the limited sea-based survey work in the area (Eyles and McCabe 1989; Gallagher et al 2004; Kelley et al 2006). This has led to considerable debate regarding the validity of different sea-level models in the Irish Sea since the last glacial maximum (eg McCabe et al 2007; Roberts et al 2007; Brooks et al 2008; McCabe 2008; Shennan et al 2008).

Figure 2.4 Palaeogeographic maps of British Isles (data from Brooks et al 2011). Basemap data derived from GEBCO 08 (

The next land bridge to be flooded was the English Channel, several millennia after the flooding  of  the Irish Sea. Tidal waters spread up the Western Approaches  and  progressively  flooded  the former Fleuve Manche river system after c 11,000 BC,  first penetrating the Strait of Dover at c 7000 BC (Lambeck 1996a; 1997; Shennan et al 2000; Waller and Long 2003). Examples of drowned coastal peats that date from this initial inundation are reported from the Devon, Hampshire, Sussex, and Kent coasts of England (Devoy 1979; Jennings and Smyth 1987; Waller and Kirby 2002; Momber 2000; Gupta et al 2004; Massey et al 2008) and from the Seine Estuary in France (Frouin et al 2007).

The melting of the remnants of the northern ice sheets was nearly complete by the time the third and final land bridge connecting Great Britain to mainland Europe was breached. The so-called ‘Doggerland’ of the southern North Sea (Coles 1998; 1999a; 2000) has been a point of discussion for almost a century (eg Reid 1913; Godwin and Godwin 1933; Behre et al 1979), stimulated by the discovery of a Mesolithic antler harpoon and other artefacts dredged from the drowned surface of the former land mass. Figure 2.4 shows a series of palaeogeographic maps (Brooks et al 2011) revealing the inundation of this landmass, constrained by seabed sediment cores as well as geophysical model predictions that account for differential crustal motions and changes in tidal range. A recent investigation of the drowned landscapes of the southern North Sea by Gaffney et al (2007; 2009) has used extensive geophysical survey data. These studies vividly demonstrate the complex nature of the drowned landscapes in the region. The cause of the final inundation of Doggerland is uncertain. It could be that the last low-lying areas were simply overwhelmed by the rise in sea level that had typified much of the Mesolithic, or it may have finally been flooded by the Storegga slide tsunami which originated off Norway and dated to c 6000 BC (Weninger et al 2008) or by MWP1C that occurred c 6200 BC.

In summary, the Mesolithic in England was characterised by rapid sea-level rise, including several jumps that varied between several tens of metres in a few centuries to abrupt events such as the Storegga slide tsunami. In general, shorelines retreated inland as the land mass available  to Mesolithic peoples diminished. In some areas of Britain, such as Scotland, uplift has preserved a more significant proportion of Mesolithic coastlines and the archaeology of these areas has seen considerable emphasis, to the detriment of those areas where the Mesolithic coastal landscape has since been submerged. Mainland Britain was isolated first from Ireland, then the French coast and finally mainland Europe by tidal waters flooding across the continental shelf from the west and north.

Our ability to reconstruct these changes is improving rapidly. Early models assumed a global sea-level rise and simply superimposed this on current seabed topographic data (eg Behre et al 1979), but recent reconstructions are more sophisticated, using detailed geophysical survey data (Gaffney et al 2009) as well as complex glacial isostatic rebound models that account for changes in crustal elevation (Lambeck 1995a; Shennan et al 2000). The latter are also now able to reconstruct changes in tidal range, tidal flows (Uehara et al 2006), and wave regimes (Neill et al 2009), allowing ever finer detail to be added to our knowledge of  the changing coastlines and coastal processes  of the Mesolithic. In addition, this helps us to consider the medium with which Mesolithic seafarers would have had to engage. This, combined with our increasing knowledge of changes in climate throughout the Holocene, allows us to consider in more detail the prospects of interaction and communication by sea.

2.1.2 Coastal ecology and succession

Sea-level change resulted in the incursion of wetland conditions and coastal sediments over terrestrial landscapes including submerged forests (Fig 2.6) and peats. Such transgressive sequences are preserved in submarine contexts, eg in the Solent, and widely in the intertidal zone, and are buried below areas of reclaimed coastal wetland. Transgressive sediments seal and preserve old land surfaces and Mesolithic sites (see Reid 1913 and Gaffney et al 2009 for discussion of their archaeological and palaeoenvironmental significance). Most are only episodically exposed as a result of storm events and, when this happens, intertidal Mesolithic sites may be revealed; such sites are highly susceptible to erosion. A rapid literature survey (Bell 1997) recorded 95 submerged forests and intertidal peat sites in England (Fig 2.7). A more detailed survey of Wales and adjoining areas of western Britain has increased the number of sites in that area from 47 to 75 (Bell 1997). Hazell (2008) has likewise increased the number of known sites in the Solent and has compiled a database of submerged peats in England for English Heritage. The MALSF- funded ‘Waterlands’ Project has generated a GIS data layer on UK submerged palaeoenvironments (including coastal peats and forests).

Figure 2.6 Submerged forest, Stolford, Somerset
(source: M Bell)
Figure 2.7 Submerged forests and intertidal peats in England and Wales (after Bell 1997). Note, boats marked are Bronze and Iron Age not Mesolithic.

As sea level rose the basal woodland was drowned, providing evidence of the original pre-incursion ecology, revealing trees, often unbranched to 10m+, demonstrating growth in dense climax woodland. The earliest, intertidal submerged forests are exposed in areas with a high tidal range, such as the Severn Estuary which reaches up to 14.8m. Here large oaks date to c 6400 cal BC and floating dendrochronological sequences have been developed for the basal forest, cross-matching with the basal forest at Bouldnor Cliff in the Solent; both have been subject to wiggle-match AMS radiocarbon dating. Forest horizons above the Holocene basement generally represent episodes of negative marine tendency when woodlands could colonise coastal wetlands. The negative tendencies most probably relate to reductions in the rate of sea-level rise, although local factors, including the development of sand and shingle barriers, will have played a part in some areas. In western Britain many submerged forests date between 6000 and 3500 cal BC, thus spanning the critical transition to the Neolithic (Bell 2007). In Essex recorded submerged forests are later, dating to the Neolithic and Early Bronze Age (Wilkinson and Murphy 1995), perhaps because the Thames basin is sinking with the result that later submerged forests are exposed within the present intertidal zone. The basal submerged forest is sometimes covered in reed peat. There is often evidence for lithic scatters and charcoal in these basal peats and some occupation surfaces exhibit organic preservation, eg the Severn Estuary and Pembrokeshire (Bell 1997; Leach 1918; Gordon Williams 1926). Subsequent sea-level rise led to saltmarsh and rapid minerogenic silt deposition that has preserved human and animal footprint-tracks in the Severn Estuary (Aldhouse-Green et al 1992; Allen 1997; Bell 2007) with some similar evidence at Westward Ho!, Rhyl, and Hartlepool (Waughman 2005). Some human and animal tracks at Formby may also be as early as the Mesolithic, although most are later (Huddart et al 1999). As the rate of sea-level rise declined, c 5000 cal BC, there was a transition from minerogenic silts to peats, initially reed peat, then fen woodland, and in places raised mire development. Thus, in estuaries and other coastal wetlands there would have been a distinct sequence of vegetation zones from seaward to dryland: mudflats, saltmarsh, reed and sedge peat, fen woodland and, in places, raised mires and dry woodlands inland. The boundaries of these zones varied according to the fluctuating extent of marine influence, creating a highly dynamic coastal fringe throughout the Mesolithic. Each of the vegetation zones would have produced valued resources for Mesolithic communities, although current evidence indicates that sites are often concentrated on the immediately pre- transgression surface and at saltmarsh edges, at the limits of marine influence (Bell 2007). This points to the significance of the maritime edge both in terms of resources and, perhaps, its social significance as a liminal place (Pollard 1996; Cobb 2008). Locations on the immediate edges of wetland/dryland zones are also important in Ireland where occupation is associated with fen peat at lake edges (eg Fredengren 2009; Mitchell 1972) and likewise in Scandinavia (Welinder 1978).

2.1.3 Site formation, spatial scales, and

Given the complexity of the cumulative effects of sea-level change and the pattern of change in coastal ecology and succession mapped out above, it is not surprising that this raises a number of key methodological and interpretive questions.

First among them are questions of site formation and geoarchaeology (discussed in more detail in Appendix 4). One of the key questions for Mesolithic archaeology concerns the location and survival of sites across the period. The increasing use of a combination of geoarchaeological methods, from investigating sedimentary structures on a range of scales to utilising palaeobiological evidence, is beginning to address this in some areas. For example, in areas where former Mesolithic estuaries or coastline have subsequently been subject to sedimentation as a result of coastal progradation, the buried topography can be reconstructed using borehole records. Allen (2001) used hundreds of commercial boreholes to reconstruct the Holocene basement of the Severn Estuary Levels. Buried topography and coastal change has been reconstructed using boreholes and dated palaeoenvironmental sequences in the East Anglian Fenland (Waller 1994), the Somerset Levels (Kidson and Heyworth 1976), the Humber Estuary (Metcalfe et al 2000), and the Thames in central London (Sidell et al 2000). In London, for instance, geomorphological evidence is increasingly used to predict prehistoric site location (Nixon et al 2002).

However, in  Denmark, where there has been a greater focus on Mesolithic archaeology, the distribution maps of sites are correspondingly more complete as a result of survey both above and below water. Sites are often marked by prominent shell middens and artefact scatters, and it has been possible to develop models of those coastal contexts favoured by Mesolithic communities (Fischer 1997) which have had a predictive value (Pedersen et al 1997). There remains, however, a relative dearth of sites that relate to either Mesolithic coastlines or marine transgression in the UK. This means that researchers must ask more innovative questions of the known sites, employing new methodologies and approaches. For example, where Mesolithic sites are related to coastal sediments (Haslett 2000) their investigation increasingly draws on geoarchaeology to understand their context and formation processes (English Heritage 2004). On the other hand, as Mesolithic coastal survey becomes more widespread in Britain, the development of predictive models of site location will be increasingly important, especially in submerged landscapes such as Doggerland, the Solent, and areas where sites are buried by coastal sediments. This is discussed in more detail in Appendix 4, including six types of sites particularly important for Mesolithic research in England.

The second key issue is that of spatial scale. The geomorphological drivers that reshaped the landscape throughout the Mesolithic were fluvial and marine. These processes occurred before, during, and after inundation. A broad-scale understanding of the geological canvas upon which the morphology was fashioned is necessary to interpret the working and potential reworking of terrestrial and marine deposits.

Mesolithic people were attracted to economically, spiritually or logistically favoured locations. These areas of high occupation potential can only be understood fully where their relationship to the broader environment is known. Therefore there is a need to read the palaeolandscape at a scale that will enable physical relationships to be characterised. This also has to take into account subsequent reworking of the landscape. To achieve this, the early Holocene palaeolandscape needs to be defined on a scale that is sufficient to model the geomorphological processes and impacts of change through time. Once the progression is modelled it will be possible to detect areas with the greatest potential for Mesolithic activity at fixed points in time. This can be particularly challenging where the relic land surfaces may now be buried in metres of sediment, or partly eroded.

To model  submerged  prehistoric landscapes and identify areas with the greatest potential for human habitation, mapping of the seabed is fundamental. This has been undertaken by a number of recent projects, notably the North Sea Palaeolandscapes project. Comparable projects, at differing scales, include the Rising Tide research programme (Dawson and Wickham-Jones 2009), investigations of the archaeological applications of the Joint Irish Bathymetric Survey Data undertaken at Coleraine (Westley  et  al  2011),  and research associations such as the Submerged Landscapes Archaeological Network. That investigations are becoming increasingly feasible in offshore contexts is demonstrated by a number of areas investigated as part of the Seabed Prehistory Project (Wessex Archaeology 2007a; 2007b; 2008a–d). Seismic survey of the Arun palaeovalley, 18km off Littlehampton, Sussex, was accompanied by marine coring and grab sampling which revealed a peat-covered Mesolithic surface at –34.5m OD with boreal pollen and some worked flints and charcoal (Wessex Archaeology 2008a; 2008d). Archaeological assessment prior to construction of offshore wind farms has identified early Holocene submarine peats, in one case with charcoal, in the Thames Estuary and offshore palaeochannels off the north Norfolk coast (Wessex Archaeology, pers comm, 2010). In many cases, however, questions remain as to how we relate specific, sometimes dated, deposits or individual landsurfaces to broader reconstruction of the stratigraphy evident in sub-bottom seismic data.

To interpret the societies that lived in these landscapes, sites need to be interrogated at a much higher resolution than broad-scale mapping. Locating such sites in the open waters around the English coast is challenging but an understanding of their relationship with the natural and physical environment is essential to inform models drawn from broad-scale surveys. Currently, such sites exist in the preserved landscape at the foot of Bouldnor Cliff, off the Isle of Wight. Interestingly, the parallels with some of the landscapes identified in the Outer Silver Pit of Doggerland are strikingly comparable, particularly those associated with wetland. The same can be seen in relation to the Mesolithic sites in the Severn Estuary which have parallels with palaeoestuaries recorded in the North Sea Palaeolandscapes project.

Finally, questions of temporality and abrupt change events need to be considered. Improvements in radiocarbon dating, alongside Bayesian statistical methods, provide the potential for an increasingly precise chronology for earlier prehistory. In the Neolithic, it has in some cases become possible to speak of generational time, rather than the more usual centuries (Whittle et al 2007). Application of these methods to Mesolithic archaeology remains limited (Bayliss and Woodman 2009; Schulting 2005), in part because of the paucity of appropriate sites. Yet a more precise chronology is fundamental to investigating not only the tempo of cultural change, and hence of lived experience, but also the impact of climatic events and changes in coastal landscape.

The most notable climatic event is the ‘8200 cal BP’ (6250 cal BC) downturn, the impact of which on Mesolithic activity in England remains elusive (Weninger et al 2009). This rapid cooling episode was probably associated with a period of faster than normal sea- level rise, as meltwater from the large proglacial lakes that surrounded the former Laurentide Ice Sheet discharged, in two stages, into the North Atlantic. Detailed reconstructions from Rotterdam, the Netherlands, suggest that the first stage of this sea-level jump began at 8450 ± 44 cal BP (6500 cal BC) and that, over a 200-year period, there was a rise in relative sea level of 2.11 ± 0.89m, in addition to the ongoing background relative sea-level rise (1.95 ± 0.74m) (Hijma and Cohen 2010, 275). These estimates are considerably larger than the values assumed by Weninger et al (2008) of 0.25–0.50m. The effects of this rise on Mesolithic coastlines in England is not known, but in Rotterdam it caused extensive coastal flooding and it is reasonable to expect similar changes in many coastal lowlands in the southern North Sea, English Channel, and Irish Sea.

A second potentially dramatic event was associated with the Storegga slide tsunami which may also have contributed to the final drowning of Doggerland and other North Sea lowlands and the creation of the English Channel. The extent to which the latter was an abrupt event is crucial to understanding the context of the marked insularity of the British Late Mesolithic. Geological evidence for the Storegga tsunami is extensive in the northern North Sea, with deposits tentatively attributed to this event observed as far south as Howick, Northumberland (Boomer et al 2007). Wave run-up in north-east Scotland was 3– 5m and archaeological sites such as Broughty Ferry, Dundee, and Castle Street, Inverness, may preserve evidence of the impact of the tsunami (Smith et al 2004). The timing of this event overlaps with that of the 8200 cal BP (c 6250 cal BC) cold event and so it is not yet possible to determine the relative importance of each in shaping coastal evolution during this period of rapid change.

Finally, rapidly rising sea levels during the Mesolithic would also have triggered periods of abrupt shoreline retreat, especially for areas of coastline that were protected by coastal barriers of sand or gravel that were either inundated or breached. Barrier step- back is observed in early Holocene coastal sequences in Start Bay, Devon, although the precise timing of these instabilities is not known (R H Clarke 1970). However, much of the evidence from the Mesolithic is deeply buried and reconstructing exact chronologies of coastal change remains a challenge. A key point here is that vertical changes in sea level, however abrupt, are not necessarily synonymous with lateral shifts in shoreline position. Sediment supply is the crucial mediator, so that those coasts with abundant sediment supply would have been more resilient than those with limited or diminishing quantities.

2.1.4 Key research questions for coastal

Coastal evolution is the critical context for understanding the interplay of the coast and Mesolithic societies. We have  successful  large-scale  models of sea-level and coastal change but these differ depending on the data sets, criteria, and assumptions used. Detailed local sequences are generally better at reconstructing vertical changes rather than a spatial picture of coastal change. There is a good knowledge of the English coastal lowland sequences in some areas, and some sites that are likely to cover specific time intervals in the Mesolithic, and indeed specific depositional environments, can be identified with a reasonable degree of precision.

This includes submerged forests and the full range of freshwater to brackish and marine environments. This is critical for identifying and targeting areas of potential preservation.

Not all parts of England preserve this potential and in some instances natural processes of erosion and various anthropogenic activities have caused their destruction. The knowledge now exists to develop a directed programme of research that might focus on specific wetlands and their hinterlands in different regions of England. Given the expected close link between hinterlands and coast, it would seem sensible to develop a sampling strategy that is driven by patterns of behaviour in the former and then explore the archaeological and environmental potential of the latter. Potential is particularly great at the interfaces between coastal drylands and Mesolithic wetlands, especially where river channels are nearby. In terms of the key issues outlined below, a sensible strategy might be developed that considers maritime/hinterland settlement (eg based on population density and known Mesolithic sites), likely local and regional seafaring, maritime networks, identity, and space.

Sea-level and coastline change

  • How can we best reconcile the divergent models for coastal evolution in the Late Glacial and Early Holocene?
  • How did regional variations in sea-level change, erosion, and deposition reconfigure the coastline of England during the Mesolithic?
  • Can palaeogeographic models be combined with other proxy data to create an understanding of the changing texture of seaways in the early Holocene?
  • What steps can be taken to extend British Isles dendrochronology to the mid- and late Mesolithic facilitating precise comparison of regional sea- level change, palaeoenvironmental sequences, and patterns of human activity?
  • How can we best make use of large-scale submarine survey data originally gathered for other purposes (eg oil exploration or wind farms) to improve our understanding of palaeogeographical changes?
  • To what extent did extreme events (such as the 8200 cal BP cold event) impact on Mesolithic communities in Britain?
  • What was the response of Mesolithic communities to known large-scale changes in shoreline position, such as the inundation of Doggerland?

Palaeoenvironments and Mesolithic ecologies

  • How can statistical modelling of palaeoenvironmental data be used to improve our understanding of change through time at a range of temporal and spatial scales?
  • How best can we refine our knowledge of the distribution, date, and archaeological potential of submerged forests and associated sediment sequences representing old land surfaces, reed swamps, saltmarshes, mudflats, estuarine and marine environments?
  • To what extent did coastal evolution and changes in the marine environment influence the nature/ timing the Mesolithic/Neolithic transition? How far has this affected the availability of evidence for the transition?
  • What impacts did changes in coastline configuration and associated environments have on Mesolithic groups?

Theme 2 2: Maritime settlement and marine exploitation

2.2.1 Material culture

Given the likely significance of the coast to the lives of Mesolithic people in England it is somewhat surprising that little material culture provides a direct link to the sea. The lack of evidence for boats is discussed in Section 2.3.1, but in comparison to elsewhere in Europe, there is currently no direct evidence for the exploitation of marine resources through the use of fish traps or weirs, fish hooks or nets (see Crumlin-Pedersen 1995). This is probably connected to preservation; elsewhere in Europe wetland excavations (either submarine or in deep estuarine deposits) have recovered substantial evidence for the exploitation of marine resources. The interpretative potential of such finds must be stressed. In Dublin, for example, the identification of Mesolithic fishing weirs is interpreted as demonstrating routines of coppicing the local woodlands (McQuade and O’Donnell 2009). At a European level the evidence for marine exploitation suggests that inshore fishing using traps/nets was common, though there is little evidence for deep sea fishing (Pickard and Bonsall 2004).

In terms of direct evidence for the use of marine resources there is, again, comparatively little data at a meaningful scale of resolution. The dietary evidence considered below is limited in scope. Recent reviews (Mears and Hillman 2007) have emphasised the potential significance of the coast in terms of starchy and other plant foods. However, little work can substantiate this at present. Zvelebil (1994) has shown that the distribution of perforated antler mattocks has a particularly coastal and riverine focus, possibly reflecting their use in obtaining plant resources, but other interpretations are possible. Functional analyses of stone tools, especially microwear and starch residue analysis, may hold some potential in this regard but are not commonly employed in England; the exception, in the case of microwear, is a pilot study at Goldcliff (van Gijn 2007). Such techniques are routinely used on the Continent (van Gijn 1990), and are becoming more common in Scotland, as part of the Scotland’s First Settlers project (Wickham-Jones and Hardy 2009).

At the Sands of Forvie, Aberdeenshire, a pilot study successfully extracted starch grains from stone tools and identified different groups of starches, allowing tentative links between tool types and starch types to be made (Warren 2005).

Some sites feature ‘bevel-ended pebbles’, sometimes described as limpet scoops. These simple artefacts are also manufactured in bone and antler (see Saville 2004; Waddington 2007, 193–6). They form part of a suite of objects found in Britain, Ireland, and Brittany (see Pailler and Dupont 2007). Bevel-ended pebbles are frequently, but not always, found in coastal locations. Their function has long been debated, with a role in limpet processing, either as hammers for removing limpets from rocks or scoops for removing limpets from their shells, often suggested. Alternative interpretations include their use in hide processing (Finlayson 1995), possibly (but increasingly speculatively) in particular association with the processing of seal hides (see Waddington 2007, 193–6).

Marine shells were perforated, presumably being used as personal adornments of some kind, possibly in turn signifying something about personal identity or status. Examples at Culverwell include limpets, dogwhelk, cockle, and an artificially shaped and perforated oyster (Palmer 1999). Perforated cowrie shells are also known from Scotland (Saville 2004). Some perforated marine mollusc shells are found inland, suggesting the movement of people or the existence of exchange networks linking the coast and inland. Some evidence from Scotland suggests the use of marine mollusc shells as scoops or other containers including large scallop shells (ibid). At Culverwell a perforated scallop forms part of a deliberate deposit with a chert flake axe and a smoothed limestone pebble (Palmer 1999, 26).

Coasts were also important sources for lithic raw materials, many of which are also found at some distance inland, implying networks of movement or exchange (see Section 2.2.2). This may have been a very substantial reason for the exploitation of the coasts. However, many discussions of this kind are somewhat generalised, and detailed modelling of the availability of raw materials has not always been undertaken. In contrast, in zones where flint is rare much more archaeological attention has been devoted to beach surveys which attempt to identify key areas for procurement (eg Mithen 2000; Dolan 2005).

Much of the influence of the sea on Mesolithic material culture may be indirect. Issues of insularity and seafaring are dealt with in Sections 2.3 and 2.4, but the fact that Britain does not participate in the broad European shift to trapeze-shaped projectile points has often been linked to sea-level rise and the final breaching of the north European land bridges.

2.2.2 Subsistence, mobility and sedentism

Understanding the maritime environment during the Palaeolithic and Mesolithic faces the problem of rising sea levels, resulting in the loss of much of the coastal zone. This zone is particularly important for hunter-gatherers, as is clear both from ethnographic sources and from the archaeological record of those parts of northern Europe with intact early and mid- Holocene coastlines (eg Pedersen et al 1997). While the coastal environment is, of course, itself variable, certain locales have the potential for abundant subsistence resources, to the extent of facilitating a significant reduction in mobility.

The challenge is how to investigate this lost landscape. Whilst many key areas for understanding coastal Mesolithic adaptations in England are, most likely, submerged, a very small number of locations have been preserved, including near-coastal sites with subsistence evidence, such as Culverwell in Dorset, and Blashenwell and Westward Ho! in Devon (Balaam et al 1987; Palmer 1999; Preece 1980). Their exact position vis-à-vis the coast is difficult to determine from the general sea-level curves that are available, but coastline reconstruction and the presence of shellfish in all three cases suggests that it was close. There is an intriguing suggestion, based on a study of the shellfish remains of Monodonta lineata at Culverwell, that this resource may have been overexploited, suggesting intensive and persistent use (Mannino and Thomas 2001). Perhaps surprisingly, these sites contain little evidence for the exploitation of other marine species (fish and sea mammals), although methods of recovery and the position of sites at some distance from the actual coastline are, no doubt, factors. Of course the coastal zone is about more than the exploitation of marine foods: the often more open conditions can also be attractive for terrestrial resources (cf Jacobi 1978). Coastal marshes and freshwater wetlands behind coastal barriers provide particularly rich habitats for edible plants, birds (including migratory species), as well as large game, most notably aurochs. It should also be noted that some ‘marine’ resources, particularly salmon, would be accessible to inland communities as they ascended spawning rivers and streams. The problem here is that there are so few Mesolithic sites with faunal preservation. The exceptions, sites like Star Carr and Thatcham, while on waterside locations, have minimal evidence for fishing. Recent excavations at Howick, on the Northumberland coast, recovered a very fragmentary bone assemblage including seal, boar, and dog, and a range of shell species represented by small and scattered fragments. No midden is present and it is postulated that one may have existed closer to the contemporary shoreline, which is now lost through erosion (Waddington 2007). Such arguments remind us that many of our assessments of subsistence are based on very small spatial samples of Mesolithic landscapes.

The occurrence of perforated marine mollusc shells in inland cave sites in the Wye  valley and   at Torbryan, as well as the occurrence of lithic raw materials of coastal origin on inland and upland sites such as Waun-Fignen-Felen (Barton et al 1995; Barton and Roberts 2004) and also upland sites in north  Wales  (Bell  2007),  has  been  interpreted in terms of seasonal movement up valleys from coast to uplands in summer. This provides an indirect means of assessing the use of coastal resources more generally, as well as potentially offering insights into mobility, although disentangling the movement of people and that of materials through exchange will always prove difficult.

Finally, stable carbon and nitrogen isotope analysis of human and dog remains from near-coastal contexts provides another means of investigating the use of marine subsistence resources, as well as informing on mobility and territoriality (Schulting 2009). Oronsay, on the west coast of Scotland, and Ferriter’s Cove on the south-west coast of Ireland, have proved instrumental in demonstrating that people there were focused on the coast year-round and were not moving inland seasonally (Richards and Mellars 1998; Woodman 2008). A series of sites on Caldey Island, in south Wales, show more variable human values ranging from moderate to high reliance on marine protein. They also push back the significant use of marine resources to at least c 7500 BC (Schulting and Richards 2002). Unfortunately, the paucity of human bone from near-coastal sites in England has meant that this technique has so far made less of a contribution here. An important exception is seen in an early Mesolithic dog, from Seamer Carr, with elevated δ 13C values suggesting a contribution of some 50% marine protein (Schulting and Richards 2009). This may very well indicate coast/inland movements, although on a relatively small scale. Measurements on humans found further inland, most notably from the early Mesolithic (c 8300 BC) site of Aveline’s Hole in the Mendips, Somerset, show no use of marine protein (Schulting 2005), nor does an isolated late Mesolithic (c 5700 BC) femur from Staythorpe, on the River Trent, some 60km from the east coast. The ‘terrestrial’ result from Staythorpe also suggests that salmon (which have a marine isotope signature) did not feature strongly in the diet of at least this individual.

Evidence from elsewhere in Britain and Europe suggests that the scale of movement for Mesolithic groups may have been more limited than often thought, raising interesting questions regarding the relationships between groups on the coast and those more inland (and of the movement and significance of ‘coastal’ materials found inland). Certainly the assumption that all hunter-gatherers moved in predictable seasonal rounds from the coast to inland areas is challenged by these recent analyses.

2.2.3 Key research questions for maritime
settlement and marine exploitation

Some commentators have supposed that coastal resources were so attractive to Mesolithic communities that they would always be utilised if available. In practice, the nature and extent of exploitation  of coastal resources and the impact this has on settlement is not entirely clear. Moreover, coastal resources are themselves highly variable, both in absolute terms and in terms of the ease of access with a given technology.

Settlement patterns

  • Can we create reliable predictive models of Mesolithic activity patterns and settlement to facilitate the investigation of submerged landscapes?
  • Is the oft-quoted hypothesis of coastal sedentism supported by clear evidence from England, and do regional differences exist?
  • How does evidence for the role of human agency  in coastal environments compare between the Mesolithic and Neolithic, and what contribution does that make to understanding the nature of the transition?
  • What was the nature of territoriality in the Mesolithic and how significant was coastal/inland mobility? In what ways did access to the coast influence patterns of territoriality?
  • What were the key factors influencing the distribution of coastal settlement? Did distinctive rocky headlands or ‘edge’ locations provide particular focal points?
  • How might targeted work on specific landscapes, of known date and depositional type, facilitate the development of a sampling strategy enabling the relationships between coasts and hinterlands in settlement patterns to be examined?

Subsistence practices

  • How has coastal evolution and other taphonomic factors influenced the preservation and visibility of the Mesolithic archaeological record?
  • Is evidence for fire in coastal environments the result of natural wild fire or human agency? If the latter, what does the context tell us about why it was done?
  • How widely distributed amongst Mesolithic sites are marine resources (eg shells, fish, geological materials from coastal exposures) and how are we best to understand the presence of coastal materials on inland sites?
  • To what extent did Mesolithic subsistence practices focus on marine foods and how did this change across space and time within the Mesolithic?
  • What were the key marine resources and how were they exploited?
  • Can use-wear analysis of Mesolithic tools better inform us as to subsistence practices?

Theme 2 3: Seafaring

2.3.1 Seafaring and insularity

Our understanding of seafaring links between places in the Mesolithic is at present relatively limited. Dugout canoes (logboats) are well known from Mesolithic contexts in mainland Europe, particularly in Denmark (Pedersen et al 1997) and the Netherlands (Louwe Kooijmans 2001a; 2001b). One old find of a logboat from the estuarine carse clays of the Firth of Forth, at Perth, has been argued to be Mesolithic on stratigraphic grounds (Geikie 1880), but this is far from certain (McGrail 1978). Another possible logboat was found with submerged trees and estuarine sediments in the foreshore near a core and tranchet axe at Thurlestone, Devon, in the 1920s (Winder 1924), and a fragmentary, but very dubious, possible logboat from Lough Neagh, Northern Ireland, dates to the late Mesolithic (Woodman 2003). A single possible paddle has been identified at the site of Star Carr (Clark 1954, 177) which, despite being beside a lake, was nonetheless an inland site. Indeed the paddle itself is relatively small and it is uncertain whether it would have been of practical use. As well as logboats, it is highly likely that Mesolithic communities also used skin boats, and coastal archaeologists need always  to be vigilant for the remains of these and logboats in appropriate stratigraphic contexts and palaeochannels. At present there is no firm evidence of boats from the British Mesolithic and most discussions of seafaring technology are therefore speculative.

Evidence for Mesolithic seafaring includes the colonisation of islands involving long sea journeys, such as Shetland and the Outer Hebrides (Warren 2005), as well as the location of sites on islands involving very difficult sea crossings, such as Ynys Enlli (Bardsey Island) in North Wales, where one must negotiate the treacherous Swnt Enlli (Edmonds et al 2009). Recent attempts at modelling journey times through the Western Seaways (Callaghan and Scarre 2009) are an interesting means of understanding the scale of seafaring in the past, although many need to be nuanced in the light of different sea conditions/tidal regimes etc, not least in light  of different sea levels. Evidence for fish exploitation provides some insight into sailing capacity, with recent reviews suggesting that there is little evidence for the practice of deep-sea fishing in the European Mesolithic (Pickard and Bonsall 2004). The nature of Mesolithic seafaring thus remains subject to much conjecture (eg Warren 2000) and is often assessed through indirect evidence, especially suggested typological links between artefacts or the use of raw materials specific to particular regions, such as Arran pitch stone or Rhum Bloodstone (Wickham-Jones 2005).

2.3.2 Key research questions for seafaring

A range of evidence from around Europe in the Mesolithic shows that seafaring was relatively commonplace and could include difficult sea crossings; yet little direct evidence for seafaring exists in England.

  • Can we find concrete evidence of what kind of boats were in use in the Mesolithic period in Britain? Which contexts have particular potential for discoveries?
  • Can experimental archaeology help us assess the sea-worthiness of logboats (of different kinds of wood) and skin boats?
  • Can experimental archaeology help identify proxy evidence for boat use and manufacture?
  • What can models of sea-level change and coastal evolution tell us of changing conditions/constraints on past sea crossings?
  • To what extent can subsistence evidence, especially that pertaining to fishing, inform us of the nature of seafaring?
  • Is there any evidence for a change in the nature of seafaring at the Mesolithic/Neolithic transition, when a variety of evidence suggests more contact between areas separated by water?

Theme 2 4: Maritime networks

Our limited understanding of seafaring connections between Mesolithic maritime communities has been further compounded, in some areas, by our modern interpretations of material culture.

A particularly good example is found in studies  of the late Mesolithic in the Irish Sea basin. Here there are distinct differences in lithic typologies over time and across this broad area. In England, Wales, and Scotland, the use of broad-blade non-geometric microliths in the early Mesolithic is replaced by the use of narrow-blade geometric microlith forms in the later Mesolithic. In contrast, in Ireland the use of narrow-blade microliths in the Irish early Mesolithic (technologically equivalent to the English later Mesolithic, with the English early Mesolithic unknown in Ireland) was replaced in the Irish later Mesolithic by broad-blade and flake technologies, sometimes but not always including the ‘Bann flake’ (Woodman 2004). A similar pattern, though with some stylistic differences, is also present on the Isle of Man (McCartan 2002; 2004). These contrasting patterns have been interpreted as representing a lack of seafaring connections across the Irish Sea in the late Mesolithic, and narratives of the period have been dominated by the idea that the coastal communities around the Irish Sea were relatively insular and unconnected (eg Sheridan 2007). Yet it may be the case that to interpret such distinct differences in material culture as the product of insularity is too simplistic. Indeed a variety of works in both archaeology and anthropology since the late 1970s have shown that material culture styles have the potential to demonstrate identities and group affiliations (cf Conkey 2006 for a detailed review). In particular,  as Hodder’s (1982) ethnoarchaeological work in the Lake Baringo area of Kenya illustrated, distinct differences in material culture styles arose amongst groups who were in regular contact but who sought to differentiate group identity by using such different styles. Thus it should not be automatically anticipated that material culture can provide an index of levels of contact and, with this perspective in mind, it could be argued that, far from being insular, the late Mesolithic communities in the Irish Sea basin may have had strong seafaring links but may have deliberately sought to differentiate group identities through different technical choices and lithic styles. Such arguments are not necessarily new: Woodman (1981, 107) argued that the differences in technology did not imply an absence of contact, but that such contacts were at a comparatively low level. Some possible material links between places on different sides of the Irish Sea (Cobb 2007a; 2007b; 2008; 2009a; 2009b; Saville 2003) have been suggested, and might support such models of low-level contact. Edmonds et al (2009, 389) review a range of themes beyond typology linking communities on either side of the Irish Sea and argue that the problem is not one of whether or not there was contact, nor of the existence of a stark opposition between Ireland and Britain, but ‘how we might construct a more nuanced understanding of the material basis of contact over time across the Irish Sea’.

Similar issues relate to contact across the English Channel and across the North Sea. Such issues have been thrown into focus in the context  of debates about the origins of the Neolithic, especially Sheridan’s arguments for relatively discrete episodes of contact and colonisation, again argued to have taken place against a background of Mesolithic insularity (see Section 2.2.1) (Sheridan 2007; Thomas 2004). Certainly European research, with detailed models of the movement of materials in networks of trade and exchange across the sea (eg Zvelebil 1998; 2008), suggests radically different possibilities for Mesolithic seafaring and contact than is currently evidenced in England.

2.4.1    Key research questions for maritime networks

Two primary networks are  relevant  here:  the first, the extent to which marine materials move inland (see Section 2.2.2), and second, the extent to which the sea facilitates contact between adjacent islands/mainlands. The nature and extent of marine networks of both kinds have been significant in a European context.

  • To what extent can isotopic analyses of diet and lifetime mobility help us understand subsistence, coastal/inland movement and other patterns of mobility?
  • To what extent do changing sea levels over time influence connections across the sea?
  • Did the English Channel and the Irish Sea act as a complete barrier to movement, as suggested by the lithic typologies, or are these typologies poor proxies for understanding the level and nature of contact?
  • What does the absence/presence of evidence for contact indicate about the nature of the societies involved?
  • To what extent do these networks change at the Mesolithic/Neolithic transition?
  • To what extent does the archaeological understanding of this material compare to the claims from modern and ancient DNA for routes of colonisation and contact during the Mesolithic?

Theme 2 5: Maritime identities and perceptions of maritime space

2.5.1 Belief, ritual, and perception

With relatively limited structural evidence compared to the following Neolithic period, and a material record dominated by stone tools, evidence for Mesolithic belief systems in the UK may seem relatively sparse. Recently, it has been Britain’s shell midden sites that have been the focus of much literature about Mesolithic belief, ritual, and ontology (Chatterton 2006; Cobb 2007a, b; Cummings 2003; T Pollard 1996; J Pollard 2000; Warren 2007) and this places the archaeology of Mesolithic coasts at the centre of our attempts to reconstruct belief, ritual, and perception. There are many reasons why midden sites can be argued to enable a greater understanding of Mesolithic belief systems. Middens on the island of Oronsay, on the west coast of Scotland, have yielded fragments of human bone from the very end of the Mesolithic and some comparable data are available from other middens: a single human femur was recovered from a shell midden at Rockmarshall, Co. Louth (Mitchell 1947; 1949), whilst fragmentary human bones were recovered from Ferriter’s Cove, Co. Kerry (Woodman et al 1999), which included small spreads/deposits of shells. No middens from England or Wales have included human bone. The presence of human bone on some later Mesolithic middens is in contrast to the general absence of funerary evidence for the British later Mesolithic.

Recent reviews of the human bone from the Oronsay middens have identified two broad processes leading to the presence of these materials. The first is the occasional appearance of ‘isolated loose bones’, considered to be quite common on Mesolithic sites where conditions for faunal preservation are present and to result from a ‘random taphonomic phenomenon’, perhaps relating to a number of possible reasons for depositing human bone (Meiklejohn et al 2005). Recent research by Gray Jones (2011) has demonstrated that loose bones on Mesolithic sites are most likely to relate to funerary processes that were extended across the landscape, and which often involved the fragmentation of bodies. On the Oronsay middens small clusters of hand and foot bones form the second broad grouping, and have sometimes been seen as implying that human bodies may have been excarnated upon the middens, with the remaining bones being missed when defleshed bodies were removed. Meiklejohn et al (2005, 100–1) question how the tight spatial grouping of these hand and foot bones would result from these processes. Instead, they emphasise the deliberate deposition of groups of hand and foot bones. At Cnoc Coig this includes one instance with human hand and foot bones, from more than one individual, placed on top of a seal flipper. In both cases, what is seemingly demonstrated is the meaningful manipulation of human bodies. Consideration of the Oronsay middens, and the human bones from them, reminds us of the considerable extent to which our discussions of the British Mesolithic are dominated by a very small number of sites (see for example Meiklejohn et al 2011).

The funerary evidence, coupled with the sheer size of some these sites (but see Finlayson 2006) and their shoreline location during the Mesolithic, have been fundamental to the argument that some shell middens may have possessed important transformative properties which derived from their position between land and water. Indeed a number of authors have argued that the transformative powers of such locations extended not simply to transforming humans but to transforming animals and mediating human/animal relationships as well (Cobb 2008; T Pollard 1996; J Pollard 2000). The association of human and seal bones at Cnoc Coig, for example, adds strength to this kind of interpretation. Beyond the specifics of these accounts, the preservational qualities of shell middens, like those at wetland sites elsewhere in Europe, enable a much closer understanding of the interrelationships between different types of material practices, objects, and contexts that greatly facilitates the complex task of unpicking Mesolithic ritual and belief. Whilst this makes them a useful focus for exploring belief and ritual amongst the maritime communities of Mesolithic Britain, it is important to be aware that such sites are the exception rather than the norm, not least because relative sea-level changes means that Mesolithic shorelines in England are frequently now submerged.

Beyond the importance of some individual locations, the experience of the daily round amongst maritime communities is likely to have been ritualised. From the daily observation of the changing tides to the processes of going to sea, of fishing, and of hunting and moving through tidal mud flats, all of these tasks would have taken place according to Mesolithic beliefs and understandings of the world. These beliefs would have been exhibited in a number of ways. Activities may have been organised around identity categories such as age, gender, and sexuality (which, it is important to remember, may not correspond to our own modern understandings of these categories), and for which another perspective is provided by footprint tracks, such as those at Uskmouth and Goldcliff. Many of these routines may have been highly temporally specific, with the rhythms of the tides and, ultimately, the cycles of the moon likely to have been a key determinant (Pollard 1996). In this, some possible distinctions between marine foragers and later farmers, more likely to have been influenced by the diurnal cycle, may be noted.

Moreover, whilst midden sites may represent specific locales that were venerated, it  may  also be the case that visual connections between places along coastlines or across the sea were important aspects of Mesolithic belief systems for people both on the land and at sea. Tilley (1994), Cobb (2008) and Cummings (2000) have argued from a phenomenological perspective for the significance of particular coastal landforms and rock outcrops in both the Mesolithic and Neolithic. As noted in Section 2.1.3, it is interesting that some locations remained significant despite undergoing substantial landscape change, including the transgression of the sea.

In this, it may be worth noting the need to assess critically the role that islands play in Mesolithic Britain. Particularly dense concentrations of Mesolithic activity occur on some rocky islands. Sometimes activity takes the form of middens of marine molluscs as at Culverwell, Portland (Palmer 1999), or Oronsay (Mellars 1987), Risga (Pollard 2000) and Morton (Coles 1971) in Scotland. Sometimes, as at Goldcliff, Wales, there are concentrations of activity on the edges of a former island (Bell 2007), often   in the form of lithic scatters and even in locales which would have involved difficult sea crossings, as on Lundy and Bardsey. Such concentrations pose the question: do they reflect the maritime resources offered by coastal islands, including perhaps the opportunities beaches present for lithic procurement, as suggested at Morton  (Deith  1983),  or  did social factors also play a part in these concentrations of activity? Ethnographies of northern hunter-gatherer communities suggest that islands may have had an important cosmological role, perhaps acting as intermediaries between different tiers of the cosmos, whilst the location of cemeteries on some islands in the European Mesolithic has been tentatively suggested to imply an association between death and certain islands.

At  times,  it appears that peninsulas may also have held some significance. The location of sites, especially middens, on these islands and peninsulas appears to emphasise the extreme margins  of dryland, and similar themes  may  be  present  in lacustrine settlement in England and Ireland. Understanding the reasons underlying site location in coastal environments is a significant challenge. It is important to be able to assess the real extent of Mesolithic activity, as opposed to archaeologically highly visible dumps of material, which may have taken place at the water’s edge (and in this the ongoing work at Star Carr provides valuable lessons). Similarly, it is important to assess whether the large amount of Mesolithic evidence from small offshore islands genuinely reflects a focus of Mesolithic activity or is a product of biasing factors, such as the tendency for these islands to have seen less intensive agriculture and development, thereby preserving Mesolithic evidence, or the attractiveness of islands as a research focus. Consideration of the role of islands needs to include submarine outcrops and  topographic  rises,  such  as  one  identified off the Norfolk coast, which may have been significant islands before submergence (Murphy 2007).

2.5.2 Landscape change and social change

As outlined above, during the Mesolithic period climatic fluctuations drove environmental change and sea-level rise. At the outset of the Mesolithic, Britain was a peninsula of northern Europe (Fig 2.4) where the North Sea and eastern English Channel would have been a low-lying plain interspersed with rivers, wetlands and hills (Gaffney et al 2007; Shennan et al 2000; Lambeck and Chappell 2001). This enabled an unhindered transfer of knowledge and a common culture across territories from western Russia to Scotland (Clark 1936; Bailey and Spikins 2008). A few thousand years later the human landscape had changed markedly. As the sea encroached, available land area was reduced but productive estuaries, sheltered archipelagos and maritime coastlines increased (Coles 1998). These rich ecosystems, calculated by Rowley-Conwy (1983) as being three times more productive than inland areas, could have attracted a greater number of people. This happened in the Baltic where similar processes led to an increase in maritime exploitation (Grøn 2003; Fischer 2004; Lübke 2009). It is reasonable to assume that such a comparable environment drew people to the growing estuaries between Britain and the Continent, possibly increasing population densities in the centuries prior to inundation.

The consequences of sea-level changes during the Later Mesolithic would have been increasingly noticeable by local communities with, on occasion, a rise in sea level being quite marked. Moreover occasional rapid changes, such as the Storegga tsunami (see Section 2.1.3), are likely  to  have  been particularly significant, felt both in terms of pressure on resources and in terms of the perception of the sea. Indeed Weninger et al (2008) speculate that 700–3000 individuals may have died with the tsunami, particularly those concentrated on the rich resources of the Outer Silver Pit (Gaffney et   al 2007), with marked social impact on those communities. A gradual rise in sea level would have a less catastrophic, but not less significant, impact on changes in landscape and land use.

Whilst  sea  levels  clearly  inundated  Doggerland and caused other large-scale changes in the landscape (both slowly and in more rapid events), the debate remains unresolved over the longer- term social influences of such change. Substantial land losses and catastrophic events, particularly during and after final severance, would have placed pressure on resources, perhaps leading to increased levels of competition or even overt violence. Jacobi (1976, 78) goes as far as to suggest a marked social effect was caused by the pressure of rising sea levels, with an increasing regionalisation of settlement patterns in Britain and marked isolation from different processes happening on the Continent. Certainly, settlement would have been affected by these changes, and distinct lithic styles developed in the British Isles, in contrast to the Continent, at the time of the flooding of Doggerland. Rising seas would have led to larger areas of coastline and more temperate climates and vegetation within the British Isles with subsequent influences on resources and settlement pattern. Thomas (2007, 429), however, notes that increasing regionalisation is a feature   of the later Mesolithic on the Continent, as well as in Britain, so it remains open to question whether these effects were related to sea-level change alone rather than broader environmental changes or other historical trajectories (Spikins 2000). Potentially, highly maritime societies on Mesolithic coastlines may have maintained some contacts with the Continent and Ireland (Thomas 2007, 429; Bailey 2004; Waddington 2007) and these links are also likely to have been influenced by changing sea levels and sea behaviour.

Aside from the decoupling of lithic traditions between England and the rest of Europe, the main long-term effect of sea-level change was likely to have been a change in the distribution and focus   of Mesolithic communities, with potentially new opportunities being created, particularly along the newly produced eastern coast, and old landscapes, such as large riverine and estuary systems in Doggerland, being lost. Many other impacts are likely: for example on the nature of myth, legend, and oral tradition, but it is uncertain whether archaeological analysis will provide meaningful detail of such consequences.

2.5.3 Key research questions for maritime identities and perceptions of maritime space

  • To what extent were changes in sea level perceptible to Mesolithic communities?
  • Can we consider the Mesolithic of England as a ‘maritime culture’?
  • How did the marine environment influence belief and social identity?
  • Did shell middens play a key ‘transformative’ or ‘liminal’ role as claimed by some?
  • How did coastal change impact on belief and society?
  • To what extent do concentrations of Mesolithic activity reflect resource availability and/or the social significance of place?