In the last chapter I broadly outlined the history of research of the Mesolithic and Early Neolithic, highlighting the key concerns of archaeologists over the years. This present chapter will set out more comprehensively the evidence and prominent interpretations of the Mesolithic and Neolithic transition in Ireland in general, in order to put the evidence available for the west in perspective. The evidence for the area covered by this thesis, the west of Ireland, will be discussed in chapter 5 when I will deal with the various areas more specifically.
To begin, the evidence for the post-glacial arrival of the flora and fauna (including humans) will be discussed, looking at the evidence for the existence and timing of the land bridge connection to Ireland from Britain. I will then move on to the Early Mesolithic, and discuss the sites of Mt. Sandel and Lough Boora, and how these have been interpreted, and then look at the general distribution of evidence in Ireland. This is followed by a look at the debate on the reasons for the transition from the Early to Later Mesolithic, and then on to the Later Mesolithic itself. I will then outline the evidence for burial practices in the Early and Later Mesolithic. Finally, I will outline the evidence for the Mesolithic-Neolithic transition, looking at the interpretations put forward for the arrival of farming.
The discussion of the post-glacial arrival of flora and fauna has centred on glacial refugia, and the theories of the landbridge which connected Britain and Ireland during periods of lower sea levels. While it is apparent that there was a land bridge at some stage, the evidence suggests that this cannot account for all the arrivals, including humans; Mesolithic communities, possibly both accidentally and deliberately, brought various species with them, possibly from as far away as the Pyrenees; however, this is open to question.
The postglacial colonisation of Ireland by Mesolithic communities is certain, but the means and timing are not. The ponderings concerning the arrival of humans onto the island have been mediated on an understanding of the processes whereby the flora and fauna recolonised the island as well, an issue that since the mid-nineteenth century has been hotly debated (Preece et al. 1986, 488). At that time the discovery in Ireland of organisms that were only known from Iberia and the Mediterranean – the Lusitanian elements – led to the formulation of three theories concerning their arrival: the first maintained that there was a survival of species in the south of Ireland; the second envisaged an early post-glacial terrestrial route; and the third, an accidental, trans-marine post-glacial introduction (ibid.). Added to these is of course the human aspect of people accidentally or intentionally introducing species (Mitchell and Ryan 2001, 99).
Mitchell and Ryan contended that the landbridge theory was most plausible as the arrival of flora and fauna was not sporadic, but rather showed distinct patterning which implies an orderly advance, cut off with the disappearance of the landbridge; the landbridge was considered to run from the north-east of Ireland or from the mid-east (ibid.). Preece et al’s. (1986) analysis of non-marine mollusca from a deposit of calcareous tufa in Co. Dublin suggested that while some of the species may have used a landbridge, the later arrivals are only present after 8000 BP, some two millennia after the postulated landbridge would have disappeared. This led to the difficulty of “how it is that many vertebrates with active powers of dispersal have been unable to reach Ireland whereas most of the British land snails, with proverbially poor powers of dispersal, have done so in a sequential order post-dating the existence of any postulated landbridge”; they offered two explanations: glacial refugia; or that “the existence of the Irish Sea has not posed such as insurmountable barrier as has been supposed” (ibid., 504). They argued however, that the glacial refugia notion was untenable, and left open the possibility of a landbridge, but suggested the mollusca evidence did not support a northern route, but rather a more southerly one, and one that remained open for later (ibid., 506).
Ten years later, Lynch looked at the colonisation of Ireland by mammals, with his results suggesting for his sampling of otters, stoats, and badgers, that the Irish populations were more like English populations than Scottish, therefore showing weak evidence for a northern landbridge, but also commenting that these findings could not “disprove the theory of a recent origin for the otter and badger populations” (Lynch 1996, 182). Moreover, Lynch argued that the badger’s general ecological niche would not suggest the animal using a soggy landbridge, which led to the postulation that people may have introduced the badger, as well as the pine martin, commenting
“our knowledge of the past history of Irish mammals is remarkably scant…Naturally, to state that humans could have introduced a species to Ireland, does not imply that they did in fact introduce species. However, given man’s propensity for the introduction of species, the possibility must be addressed” (ibid., 183).
Using genetic evidence from pygmy shrews, Mascheretti et al. (2003) have suggested that their results point to an arrival to Ireland of a pygmy shrew population from around the Pyrenees region; further, as the pygmy shrew population on the Isle of Man had similarities with Britain, and not with Ireland, a landbridge connection at that area, as postulated by Wingfield, could be discounted. Moreover, they comment that the lack of other species such as other shrews and voles in Ireland would also discount a land bridge, and maintain that the arrival of the pygmy shrew, as well as other of the Lusitanian elements, suggests human transportation during the Mesolithic colonisation (ibid., 1598). However, recent work on the pygmy shrew, which used a larger data set, has suggested that the Irish pygmy shrew was of the same ancestry as the Isle of Man and British populations, but that later the Irish type was displaced in Britain and the Isle of Man, but remained in Ireland (Searle 2006). This, therefore, discounted the necessity of the human transportation link to Ireland.
Wickham-Jones and Woodman (1998) have outlined the available evidence for the possible presence of Palaeolithic communities in Ireland, and suggest that the miniscule evidence of possible early lithic types – found in both the north and the south of the island – could mean a pre-12000 BP date for human occupation, but c. 7900 cal. BC is the earliest definitive evidence (Woodman, pers. comm.).Top of Page
Understandings of the Irish Early Mesolithic are dominated by the excavations of Mt. Sandel and Lough Boora; these two sites, a site-type base camp, and site-type temporary camp, have conjured up a lucid picture of the Early Mesolithic for us. These sites will be discussed as well as looking at the other evidence for the Early Mesolithic communities.
The Early Mesolithic communities in Ireland are traditionally defined by their use of microliths, a composite tool type indicative of the Mesolithic throughout Europe. Particular to Ireland was the use of surface-retouched needle points, flake axes, and polished stone axes (Costa et al. 2005, 19). As noted in chapter two, the distribution of Early Mesolithic material was traditionally centred in the northeast. The following two maps (Fig. 3-2) relay this conspicuously, with Woodman’s move to Munster expanding the known Early Mesolithic sites to the far south.
Two main sites have dominated discussions of the Early Mesolithic: Mt. Sandel, Co. Antrim, and Lough Boora, Co. Offaly. The Mt. Sandel excavations dominate the picture of the Early Mesolithic as so few other sites have been excavated and fully published, let alone found. Not only that, but here was evidence for dwellings – until recently it was not until the Neolithic that there was again evidence for houses in Ireland. The Mt. Sandel structures were interpreted as having been egg-shaped, approx. 6m in diameter, and formed by bending over c. 20cm thick saplings. The use of sod walling was suspected, with an estimate that this could have covered c. 50% of the house, the remainder covered with hides (Woodman 1985, 133). While numerous posthole patterns could be ascertained, it was argued that one house was in use at any one time; slighter structures were also identified. In some cases the hearths in the interior of the houses were recut, suggesting reuse of the site. The dating of the site ranged mostly from c. 7900 cal. BC to c. 7500 cal. BC, with a number of later dates (Woodman 2003, 8). There was also Neolithic activity, as well as evidence for later periods (Woodman 1985, 121).
The settlement at Mt. Sandel produced a limited amount of faunal remains. These consisted of calcinated bone fragments with the acidic nature of the soil precluding the survival of uncalcinated bone. Therefore, the faunal record was argued as being unrepresentative (Woodman 1985, 71). What was recorded showed an overwhelming dominance of fish bones (81% of total), of which salmonoids accounted for c. 84%, bass and eel for 15%, and the remaining 1% was plaice or flounder. The authors noted that none of the fish species would have been available for catching during the winter, with the modern salmon run beginning in April and continuing for five months (ibid.). However, Warren (2001, 183) has noted that to transfer this nineteenth and twentieth century data back to early prehistory is predicated on the water temperature being the same, as this would affect the salmon seasonality – an unknown variable. While the salmon run is intensified during a shorter period, salmon is today available for half the year.
The concentration of fish bones in one hearth was taken as signs that the hut at that spot had been used for smoking fish, and the storage of other foods was also suggested. While it was also argued that the number of occupants may have been smaller in winter than in summer due to the dispersed nature of pigs’ habitats (ibid.) – in contradiction to the usual surmising on winter camps – this leaves out the evidence they had surmised for the storage of foods, and hence the ready supply available during winter. The mammal bones (15% of total) were 98% pig bones, with the interpretation for winter hunting of the pigs. The remaining mammal bones were hare and either dog or wolf. While birds accounted for only 4% of the total of the bones, 13 species were present, ranging from woodland, to river, to coastal types (Woodman 1985, 72). The presence of at least two types of bird of prey was noted, including the goshawk, and Woodman noted that goshawks were also found at the Mesolithic site at Dalkey Island, and also the Beaker period assemblage at Newgrange (ibid.). While Woodman interpreted the goshawk bones as present on the site as the bird would have been attracted by bird traps set while hunting for pigs (Woodman 1985, 75), Mitchell and Ryan (2001, 115) have commented on the possibility of the goshawk having been used in falconry.
Due to the poor preservation of organic material at Mt. Sandel, the evidence for flora was negligible. The usual suspect, the ubiquitous hazelnut, played a dominant part in the material, with evidence also for wild pear/crab apple, and water lily (Woodman 1985, 80). Over the years Woodman has oscillated in terms of the importance placed on flora in the diet during the Mesolithic, recently suggesting that the coastal and riverine locations of Mesolithic sites does not support such a theory of a plant dominated diet, as well as the lack of plant processing equipment (cf. Woodman 1985; Woodman et al. 1999, 138). However, contra Woodman, it is precisely these locations that are concentrated with flora (Brown 1997, 283). Tools are multifunctional; axes may have been digging tools as well, for example for tubers. Moreover, the collecting of seaweeds, seeds, nuts, berries, lichen, fungi, and greens do not need an elaborate, clearly identifiable tool kit for collecting and processing: indeed, many do not need tools at all.
Woodman (1985, 171) has argued that the Mt. Sandel lithic assemblage has insular characteristics, and therefore this site may not be indicative of an early date for the arrival of people on the island. The dominant lithic types found were microliths (needle points, rods, and scalene triangles), along with core and flake axes, blades, awls, and scrapers. There was also evidence of ochre stained blades and flakes, with the possibility that the true number is under recorded as the ochre had since been washed off; the ochre on a borer may represent the colouring or preserving of hides (ibid., 51; 64). The use of red ochre as a dye and paint is well recorded in prehistory. Found in Mesolithic funerary contexts, it is interpreted as a material suitable as symbolic of life-blood (Bradley 1998, 24). But this was but one slice of the spectrum of colours possibly used. Other mineral, floral, and faunal substances were probably used for the painting and dying of materials, as well as peoples’ bodies. While it is suggested that ochre on some of the tools represents the preparation of other materials, it may be that ochre was placed on the lithics as an end in itself.
The Lough Boora excavation produced evidence of a lake shore site which was eventually encroached by bog in the Later Mesolithic. The original post-glacial lake was of a substantial size, and it has been conjectured that it would have encompassed the current lakes of Boora, Derg and Ree (Fig. 3-3).
The site was on a ridge overlooking a lagoon which lay in between it and the lakeshore storm beach (Mitchell and Ryan 2001, 115). The dates returned for the site were broadly contemporaneous with those of Mt. Sandel, but no dwelling structures were identified. The site was interpreted as a temporary camp site, with the dominance of eel bones suggesting a summer occupation. The range of mammals was similar to Mt. Sandel, again with a preponderance of pig bones, with the additional evidence for wild cat (Waddell 2000, 14). The flint in a chert dominated assemblage at Lough Boora, and vice versa for Mt. Sandel, was interpreted as signs of peoples’ movement in the landscape (Woodman 1985, 166) or signatures of exchange and social contact (Cooney and Grogan 1999, 24).
Substantial amounts of the lithics were burnt, as were the bones that survived. The fact that the bones showed evidence of having been burnt at a high temperature for a long time was interpreted as suggesting that they had been boiled in a broth after the meat had been eaten; after being boiled the bones would then be dumped in the fire (Woodman 1985, 75). In a seminar presentation, Nyree Finlay (2003A) noted that at Lough Boora there was also evidence for burnt lithics. She suggested that the lithics were being treated like bone – the heating of the lithics changed their colour akin to bone, and therefore the lithics were metamorphosed; and a similar pattern may be apparent for Mt. Sandel. What may be occurring here is evidence of a ritual aspect to the hearth – the purposive deposition of bone and lithics in the fire, rather than simply a pragmatic use of the fireplace as a means of disposal of rubbish.
While for many years it was presumed that an indirect percussion technique was used in the production of the lithics, it is now considered that a direct percussion technique, using a soft hammer, was used (Costa et al., 2005, 25). An analysis of an assemblage found on a beach near Greencastle, Co. Donegal has shown that two strategies were used in blade production (Costa et al. 2001, 2). A hard hammer, direct percussion technique was used to make large blades and blade-like flakes, while a soft hammer, direct percussion technique was used to make small regular blades – the former were used to make notched pieces or pieces with lateral retouch, and the latter to make microliths and backed pieces (ibid.).
In terms of the distribution of Early Mesolithic material apart from the sites of Mt. Sandel and Lough Boora, Fig. 3-2 shows that the majority finds are from the Bann Valley (where Mt. Sandel is located) and along the northeast coast, with another cluster on Strangford Lough, Co. Antrim, and more from Co. Louth and Co. Dublin (Woodman 1978). A reappraisal of excavated material of a ring barrow at Rathjordan near Lough Gur, Co. Limerick turned up microliths (Woodman 1986), and further Early Mesolithic presence in Co. Limerick has been noted at Hermitage and Killuragh (see below, Section 3.6). Finlay (pers. comm.) identified some early cores amongst the predominantly Later Mesolithic assemblage from Lough Gara and Fredengren’s (2002, 120) work on L. Gara dated a piece of brushwood to 7330-7050 BC. Fieldwork on the Blackwater, Co. Waterford revealed early lithic scatters there (Woodman 1986, 10), as has the Barrow Valley (Ramsden et al. 1995, 331). Recent finds in the north include those found on a beach at Greencastle, Co. Donegal – here microliths were found along with flint axes and “a number of small pointed bone artefacts” (McNaught 1998, 64; Costa et al. 2001). Another two assemblages have been found 300m apart, overlooking the Liffey, at Leixlip, Co. Dublin (Mullins 1999). In general, the evidence so far for the Early Mesolithic material tends to be found on higher ground over looking water, or on the coast.Top of Page
A key change in the lithic strategies in the Mesolithic is apparent, from the transformation from the microlithic tools to the macrolithic tools. The various suggestions for this change, such as population replacement, differing economic practices, and changing social decisions, are presented and assessed.
A significant transformation in lithic production occurred some time in the Mesolithic, with a change from the production and use of microliths to macroliths. This innovation is usually cited as exceptional in Europe apart from the Isle of Man (e.g. Gibbons et al. 2005; Fredengren 2001; Kimball 2000A, 3; McCarton 2000, 26), however somewhat comparative changes are noted in areas of Portugal (Warren, pers. comm.) and the Tyrrhenian Islands in the Mediterranean (Costa et al. 2005, 31). In the case of Ireland Mitchell and Ryan (2001, 119) maintain that the change can be rationalised by asserting a population influx – their Sandelians being replaced by their Larnians. However, this notion is based on a strict reading of culture groups as epitomised by lithic types – that lithics equal peoples – and also on the notion of diffusion as the cause of, and reasons for, change in material culture. It is generally agreed, however, that a population change is not needed to explain the transformation in the Irish Mesolithic.
Mallory and Hartwell (1997, 4-5), commenting on the fashionable turn towards social explanations rather than economic ones, have put in their few pence worth, suggesting that the change in lithic strategies can be seen as the switch from a male-dominated microlithic society with men as hunters and fishers, to a female-inclusive society where food procurement methods changed to allow women to participate, while minding the kids – akin, I guess, to today’s job sharing. Finlay (2003B, 92) has succinctly remarked that “this misogynistic view of the Mesolithic needs to be redressed…we can dismiss the exclusive male identification with microliths as a legacy of traditional ‘boys and arrows’ narratives”.
Warren (2003, 23) has suggested that the climatic deterioration that is reported to have occurred c. 6200 cal. BC – which led to a drop in temperatures for 200-400 years – should be looked at in terms of the change in lithic strategies. While allowing that this would not have caused the change, he argues that it may have contributed to its timing. However this begs the question as to whether one can view this teleologically – can a change be implicit amongst people, and be waiting for a trigger of some sort. It is unclear as whether this is what he is implying. His thesis rests on a clarification of the dating of the last microlithic assemblage, and the first macrolithic assemblage, which at the moment is not certain. Costa et al. (2005, 23) have outlined the known dates for the two assemblages, and maintain that the change in lithics occurs much earlier than 6200 cal. BC, possibly by 7000 cal. BC. However, this latter date points to the last known date of microliths, not the start point of macroliths and should be noted in the light of the dearth of reliable datable contexts for the transition, especially the insecure contexts of the dates from Newferry and Cushendun, Co. Antrim (Warren, pers. comm.).
Woodman and Anderson (1990) suggested that a change in procurement methods over the course of the Mesolithic could account for the change, whereby the focus of lithic production and use was more geared towards the building and maintenance of traps, weirs and so forth, rather than a change in what was being caught. Costa et al. (2005, 29) have generally agreed with Woodman and Anderson’s position – that a change in procurement methods may have led to the change in lithic types – but suggest this would not account for a change in the entire repertoire of tools. They argue that the lack of a change in subsistence practices rules out this as an explanation. They have argued that the change led to a more flexible approach to stone tool making, as the larger flakes were then less geared towards specific tasks and they state that they
“believe that the greater flexibility in lithic production could be interpreted as a social rather than technological or economic revolution. This liberation in lithic production could be the result of a shift in the perception of the importance of lithic tools in the activities of Later Mesolithic society” (Costa et al. 2005, 30).
However, this neglects anthropological work as discussed by Ingold – he has argued that the great tool-use fallacy, which has separated social and technical domains, “has blinded us to the fact that one of the outstanding features of technical practices lies in the embeddedness in the current of sociality” (2000, 195). In their conclusion Costa et al. (2005, 31) seem to fall back on a change of subsistence and differing methods of procurement as the reasons, having previously discounted these.
The Later Mesolithic has been contrasted to the Early Mesolithic in terms of its perceived lack of base camps, leading to speculations on whether this is an artefact of a more mobile society, or simply a lack of evidence for similar sites such as Mt. Sandel. This led to debates on mobility and social complexity in the Mesolithic, and how these concepts are evidenced in the archaeological record. The later Mesolithic evidence is often characterised by single finds, or specialist sites, and frequently sites with no lithics but returning Later Mesolithic dates. The Later Mesolithic has produced numerous coastal sites, and there has been a focus of research on coastal areas, such as the excavations at Ferriter’s Cove. I will discuss the more recent research projects that have been carried out – the coastal site of Ferriter’s Cove, the Bally Lough landscape survey project, and the island platform site at Lough Kinale, and other constructed Mesolithic sites.
As we have seen, the timing and reasons for the beginning of what we call the Later Mesolithic are uncertain. With the model of a Mt. Sandel base camp, and a Lough Boora temporary camp, a picture of a sedentary or semi-sedentary Early Mesolithic could be posited, and what was perceived as the general European model of a highly mobile society veering to greater degrees of sedentism throughout the Mesolithic turned on its head. The lack of Later Mesolithic base-camps or similar settlements has been read in various ways. Woodman queried (1986, 13; 1992, 302) whether this lack meant that comparative sites simply have not been found, or whether it meant that the Later Mesolithic communities were highly mobile and sparsely populated. Cooney and Grogan (1999, 21-3), however, questioned this idea of an increase in mobility, arguing that the Atlantic period would have been more productive and hence suitable for permanent settlement – they also suggested that situation in Ireland would more than likely have followed the European model of the sedentary complex hunter-gatherer.
Woodman has been at pains to counteract what he sees as various researchers misinterpreting him, arguing that Cooney, Fredengren, and Kimball had got the wrong end of the stick, or only looked at part of the stick, and that in fact he had only suggested that
“we could not presume that the Early Mesolithic was followed by a period of increased population levels and sedentism. Indeed it was a call for a more open attitude to the possibility of different solutions rather than accepting a convenient conventional wisdom prevalent elsewhere” (2003, 15).
Furthermore, Woodman (ibid.) has commented on the nature of Mesolithic archaeology: “do we always have to seek the Mesolithic version of the ‘Little House on the Prairie’ i.e. sites where we expect all of life’s dramas were played out”. Little (2005, 82) has appositely discussed the debate:
“ultimately, while both sides have succeeded in identifying weaknesses in the other’s proposed settlement models for the Late Mesolithic, neither model challenges the accepted (economically determined) norms of Mesolithic settlement, or attempts to address the variety of social factors active in structuring settlement patterns across the country”.
What the general impression of Later Mesolithic sites has presented researchers is locations of short term stays, and specialised sites. Woodman and Anderson (1990) have suggested that the presumed paucity and ephemeral nature of Later Mesolithic evidence, in that often a single flake is found, relates to the dynamics of their society – they suggest that a single flake may represent more extensive use of that area, for instance, the maintenance of traps, weirs and so forth. Moreover, they (1990, 386) argue that if lithics were being curated – following Binford’s model – this will not leave evidence of activity; they note sites that have returned Mesolithic dates, but have not produced any lithics, such as Mitchelstowndown East, Co. Limerick and Valencia Island, Co. Kerry. To these two can be added the Mesolithic trackway at Lullymore Bog, Co. Kildare (see below, p. 113).
Woodman suggests that where a plethora of lithics have been found, such as at Lough Gara and Newferry in the Bann Valley, this could be signs of a multitude of repeated, short visits to the same place over long periods of time, as opposed to evidence for sustained, long term settlement (Woodman 2003, 15). The excavations of a knapping floor at Bay Farm on the Antrim Coast produced no evidence for settlement per se – instead it was argued that the evidence suggested repeated visits to a lithic production site (Woodman and Johnson 1996). The Newferry assemblage contained substantially less cores and substantially more retouched pieces than the Bay Farm assemblage, and the Newferry flint was of a higher quality that is to be found on the coast, and the communities avoided the lower quality material available in the Bann Valley itself (Woodman and Anderson 1990, 382-3).
Newferry lies to the north of Lough Beg, which in turn is a northern extension of Lough Neagh. A series of excavations were carried out there, beginning with the Harvard Archaeological Mission in the 1930’s (Woodman 1977, 155). Woodman began a research excavation in 1971, which became a rescue excavation the following season due to impending farm improvement works – the site had initially been noted by Smith after diatomite cutting had been carried out there (ibid.). While the earliest evidence on the site is Early Mesolithic, the main activity is Later Mesolithic with Woodman commenting that the Later Mesolithic material is evidence of a “fully evolved industry” (ibid., 194). The excavations dated material beginning at 6225 ± 145 bc (organic mud, zone 9 ) with the main activity from 5680 ± 195 bc (wood, zone 8) to 3465 ± 90 bc (charcoal, zone 3), with most of the material presumed as not being in situ (ibid., 156-77). The upper zone – zone 2 – (not dated) contained Mesolithic and Neolithic material, as well as what was presumed to be eroded Mesolithic material from the two lower zones (ibid., 179-81). The zones were devised by the geology as opposed to the archaeology of the site (ibid., 160).
Woodman noted various changes in the lithic assemblage through the zones, especially with more evidence of lithic production – more cores and waste – in zone 7 than in later zones: these later zones consisted of lithics brought to the site, with some zones with no cores identified (ibid., 185). Woodman argued that over time a broader flake became more dominant; tanged flakes were more common in the earlier zones, with these being superseded by the leaf shaped butt trimmed flakes in the upper zones – Woodman suggested that this change may be related to a difference in hafting as opposed to a differing implement type; points are not apparent after zone 4, with the bar form apparent from the upper part of zone 5; the raw materials for the axes changed from a predominant use of schist to mudstone – these two rock types accounted for 90% of the axes; backed knives were in all zones with little apparent change through time (ibid., 187-9). Other implements apparent were picks, borers, spokeshaves, polishing stones, elongated pebbles and bone points, and Woodman noted that over time the range of tools became more restricted (ibid., 192).
Woodman argued that this was primarily a temporary fishing site; the range of lithics were only those non-organic tools that survived – the lithics probably represented tools for the maintenance of weirs and traps, as well as general purpose tools. There was little evidence of faunal remains due to the acidic nature of the soil. He argued that as the area changed from a series of river channels to marsh the site was gradually abandoned (ibid., 193).
As noted in the previous chapter, it was argued by Woodman and Johnson (1996, 228) that there was the possibility of the Bay Farm production site’s material being moved to the Bann Valley and Lough Neagh. The pits at Bay Farm were stated as enigmatic by the excavators. Similarly to the Early Mesolithic at Mt. Sandel and the upper levels of Newferry, chert appears at Bay Farm – this time in the form of a single large flake in one of the pits. Aside from the chert flake, the rest of the flint found in this pit was treated as ‘industrial waste’ (Woodman and Johnson 1996, 155-6), but the placement of a chert flake amongst the flint suggests something more involved than the disposal of industrial waste. This can arguably be seen as a purposeful deposition of a different kind of stone amongst the more pedestrian flint – if we can call it that – that is abundant in the area.
The coastal focus of many Later Mesolithic sites compared to the Early Mesolithic has been taken as evidence not for a change in patterning but the fact of rising sea levels over the periods disguising earlier evidence. Excavations of middens on the east coast at Rockmarshall, Co. Louth (Mitchell 1949), at Sutton, Co. Dublin (Mitchell 1956; 1972) and on Dalkey Island, Co. Dublin (Liversage 1968) showed evidence of the coastal resources used, and these excavations complemented the idea of the Mesolithic peoples as beachcombers mentioned in chapter two (Movius 1942; Macalister 1949).
The first Mesolithic site to be excavated on the west coast was Ferriter’s Cove, which produced evidence of significantly smaller middens – small dumps of shells – than the east coast examples. Here was evidence for repeated short visits to a coastal site during the summer and autumn months, with the dating of the site spread over about a thousand years – mainly from 5000 to 4000 cal. BC, with the main activity around 4500 cal. BC (Woodman et al. 1999, 114). While there was some evidence for linear patterning of stake-holes and one arc of stake-holes, these were not considered evidence for huts – some were possibly evidence for fish-drying/smoking (ibid., 130; 154). Possible roasting pits were discovered, and burnt stone platforms were interpreted as cooking places; the evidence of numerous burnt stones was suggested as implying the use of them for roasting or boiling of food (ibid., 126).
The faunal remains were dominated by marine fish, which were represented by 16 identified species; 10 species of shellfish were identified, of which dog-whelk predominated. The minor role played by birds in the diet was considered a factor of seasonality of the site, and the mammal remains were dominated by pig with some hare; the remaining mammal bones were of indeterminate species (ibid., 87-91). The discovery of sheep and cow remains shall be dealt with in the following section.
The lithic assemblage was predominantly of locally sourced stone: greenstone, tuff, rhyolite, and flint, with minimal amounts of siltstone, quartz and chert. Beach pebbles were also used, and 13 polished stone axes were also recovered, 5 of which had been deposited together (ibid., 153). The authors commented that the usual Later Mesolithic-type lithics – the butt-trimmed flakes and related forms – were proportionally underrepresented at Ferriter’s Cove in comparison to other sites, and that, of those present, “few of them could be considered classic examples”; they argued “it is important to emphasise diversity within the Later Mesolithic assemblages rather than always attempting to fit them within one template” (ibid., 76).
As can be seen from Fig. 3-1, the evidence for Mesolithic communities at Ferriter’s Cove is in almost complete isolation in the landscape, apart from the Mesolithic date returned from the site on Valencia Island. The Ferriter’s Cove site has described as a place that was visited recurrently for short periods of time, highlighting the fact that numerous other sites, both coastal and inland must be in the area. The model of communities moving seasonally between coast and inland was presumed to be the case in Ireland for both periods of the Mesolithic (Woodman 1978). However, in light of stable isotope analysis – which measures the protein component from bone collagen in diets (Schulting 1998, 206) – it is now argued that similarly dated Later Mesolithic human remains had a differing food intake. The analysis of a human femur from the midden at Rockmarshall gave a result of -18.1‰, which does not indicate a predominantly marine diet (Woodman et al. 1997, 143); this would support the notion of communities moving between coast and inland. While the Ferriter’s Cove human remains indicated a marine based diet (-14‰ & -14.1‰), the Killuragh Cave remains indicated a terrestrial diet (-21.96‰ & -21.3‰) (Woodman et al. 1999, 143). A terrestrial diet is taken to mean “C-3 plants, the flesh of C-3 herbivores and freshwater fish” (Lanting and Van Der Plicht 1998, 162). Schulting has noted that salmon register a marine isotopic signature not a freshwater fish signature (Schulting 1998, 207). These stable isotope results for the Killuragh Cave human remains led Woodman et al. (1999, 143) to ask “was it actually possible for some Mesolithic communities to subsist exclusively in the interior of Ireland?”
Little (2005) has taken the traditional coastal Mesolithic research focus to task, arguing that insufficient time and energy has been given towards the inland areas of Ireland, where there has been for years a multitude of evidence of Mesolithic activity. Commenting on her present PhD research on the Mesolithic in the midlands, she suggests that the comparative ease of working in coastal areas is one reason for the neglect of the midlands, where millennia of peat formation hampers any straightforward research agenda (ibid., 82). Little comments that the spatial and temporal relationship between the use of the coast and inland by the Mesolithic inhabitants is poorly understood, with the failure to follow up on the initial discovery of Mesolithic material in the midlands, the exception being Mitchell’s excavation at Clonava, Lough Derravarragh, Co. Westmeath (ibid.).
Recently, Fredengren undertook the excavation of another old findspot upriver from Clonava, at Lough Kinale, which was noted as a surface lithic scatter after drainage works lowered the lake level (Fredengren 2004, 29). This site, with a strong Neolithic presence as well, has shown to be an artificially heightened natural island, constructed with layers of stone, peat, timber and brushwood. The brushwood seems to have been pegged in to place, and it is argued that the site witnessed repeated visits (ibid.). A rectangular structure was noted, but it is unclear as of yet whether this is Mesolithic or Neolithic in date. Due to waterlogging there has been excellent preservation of organic material, with worked wood and bone surviving. One interesting aspect of the site is the lack of significant amounts of fish bone – the usual interpretation of such a site, a platform on a lake, would be to suggest that this was a fishing spot, however this does not seem to be a main activity (Fredengren, pers. comm.). The osteologist noted that this site produced good evidence for mature wild pigs – usually it has been younger pigs remains found – and in comparison to previous samples from Ireland, the pigs were quite massive, much larger than had been previously thought (McCarthy, pers. comm.).
Along with the site at Lough Kinale, there is a growing body of evidence for the construction of various types of ‘sites’ in the landscape. Fredengren’s (2002) work on the crannogs of Lough Gara has again raised the question as to whether some of these may be of Mesolithic date. She investigated a series of brushwood and timber pilings on the shore of Inch Island – this area had turned up lithic in the 1950’s, and Woodman (1978, 322) had argued, contra Raftery’s and Cross’s arguments for a Stone Age date for many of these crannogs, that the relation between the lithics and crannogs was spurious and that the lithics probably arrived there from erosion from a higher spot on the shoreline (for further discussion of the Lough Gara lithics see below, Section 5.5.10). During Fredengren’s work a Later Mesolithic date (4230-3970 BC) was returned from one of the timbers on Inch island, and also an Early Mesolithic date (7330-7050 BC) from a piece of the brushwood (Fredengren 2002, 120). While she maintains that these, and the possibly related traces of a stone causeway, cannot be specifically described as crannogs, she argues that the “new results show that there is more to the Lough Gara material than eroded deposits from earlier shorelines” (ibid., 121). Fredengren has noted that the difficulty in assessing these numerous crannogs that were noted by Raftery and Cross is that they are now covered over by grass (ibid., 132). However she argues that the stone platforms may well be of Mesolithic date.
The excavations of Moynagh Lough by Bradley (1991) again showed evidence of the construction of island platforms in the Mesolithic. Here, two knolls in the lake were heightened with a layer of white marl, stones, and brushwood. Bradley noted that the mud was “almost completely sterile and its presence is difficult to explain” (ibid., 7). The majority of the lithics, both chert and flint, were found in and on top of this layer which also contained charcoal which dated to 5270+-60 BP (ibid., 9).
While the above locales produced evidence for lithics, at Mitchelstowndown East, Co. Limerick and Valencia Island, Co. Kerry Mesolithic dates were returned with no evidence of lithics. The former site (6585±30 BP) comprised a platform of oak beside a stream, and the latter (6560±120 BP) was a stone platform which had a baulk of fossil oak dated to 8910±150 BP (Woodman and Anderson 1990, 386). Brindley and Lanting (1998) undertook a dating programme on a selection of Irish trackways and produced evidence for the construction of a pine trackway at Lullymore Bog, Co. Kildare in the Mesolithic, at a site where no lithics turned up. This date was so early (7140±70 BP) that other samples were tested and these proved the dates to be accurate as Mesolithic – 7145±35 BP for the pine track, 6120±45 BP for the overlying peat, and 6120±30 BP for pine from a pine ‘surface’ (ibid., 47). The pine trackway was constructed of transversely laid, radially split pine, with a width of 1.8m; the finder suggested that “the timber was worked and formed a linear structure which had all the appearances of a trackway” (ibid.).
Recent monitoring of development works along the River Liffey in Dublin City discovered a series of fishing traps, stake rows and a brushwood platform which have dated to the Later Mesolithic, again with no lithics apparent, apart from a single burnt blade found away from the structures (McQuade forthcoming). The earliest date reported (6090-5890 cal. BC) was for a wattle fish trap consisting of two parallel wattle fences constructed from hazel sails, rods and rushes – most of the sails were pared towards the base, and some had side branches pared as well; some had been torn rather than cut; the rods and sails were an average of 13mm in diameter. One of the fences survived to a length of 5.12m long, while the other only survived for 1m. A horizontal wattle panel lay between the two fences; this was 1.3m wide and extended for 2.1m and consisted of tightly woven hazel, with an average diameter of 25mm (ibid.).
A contemporary c-shaped fish trap was located 31m to the southeast of the wattle trap – this was also predominantly made from hazel, with one ash, and a stake was dated to 6080-5840 cal. BC (2 sigma calibration). 8m northwest of this were a stake row and brushwood platform – these were also contemporary and constructed predominantly of hazel; the hazel from the platform was up to 7cm in diameter and dated to 6080-5870 cal. BC. A row of stakes dated to 5910-5710 cal. BC ran for 25.1m; a similarly dated basket trap was located 10m away and was considered to be related to this row of stakes. Further stake rows and clusters of stakes were identified, as well as three oak planks which were deposited as drift wood along the shoreline (ibid.).
This site was discovered under metres of silt, and the lowest levels were at minus 5.8m O.D – in total a strip of 16m by 60m was excavated (ibid.). The dates suggest the use of this stretch of the river over a period of a few hundred years, and this excavation represents the first definitive proof of fishing traps, weirs and so forth in the Irish Mesolithic. The site also suggests the use of coppicing of wood (Warren, pers. comm.). One of the difficulties with ascertaining this is that hazel is very proficient as a naturally self-coppicing tree (Mabey 1997, 88), but the large quantities needed for making the lengths of traps possibly suggests human help in the coppicing. This is of considerable importance, as this is an early example of such practices – in his recent article on the Neolithic transition, Rowley-Conwy (2004, 96) has suggested that coppicing “implies sedentism” and commitment to place that can be seen as a hallmark of the Neolithic, and that coppicing “cannot be reconciled with a nomadic settlement pattern”. If the material is indeed evidence for coppicing in the Irish Mesolithic, Rowley-Conwy’s remarks are either questionable, or else this implies that current understandings on sedentism in the Later Mesolithic are questionable.
Another interesting aspect of this site is that it highlights a further role for hazel in the Mesolithic. Hazel generally gets a hard time in the literature – the ubiquity of hazelnuts on sites due to their excellent preservational qualities often means that they are either taken for granted as just a prehistoric foodstuff by archaeologists, or dismissed as ubiquitous because of their survival rate. Here, however, hazel is highlighted as the overwhelmingly dominant raw material used by the communities in their constructions – as the surrounding landscape has been surmised as “zones of salt marsh, reed swamp and fenn carr” (McQuade forthcoming), other wood such as willow, dog wood, birch, and alder would have also been commonplace, but there was a clear bias towards the use of hazel. This is not surprising as hazel is an aptly suitable timber for working in wattling, but may suggest a more involved relationship between Mesolithic communities and the hazel tree. Other trees used at the locality were dogwood, ash, birch, and oak (ibid.).
As mentioned in the previous chapter, the Bally Lough Project looked at the Stone Age settlement around the River Barrow’s estuary and then around the middle reaches of the river from Carlow to Monasterevin (Zvelebil et al. 1996). This region was for the most part devoid of research and early prehistoric evidence, apart from the known monuments. Interestingly, considering they subtitled one of their articles “the ‘Riverford Culture’ revisited”, the authors assign ground and polished axes as post-Mesolithic artefacts (Green and Zvelebil 1990, 65; Zvelebil et al. 1996, 33). The initial work on the coast and estuary produced evidence for a rhyolite quarry. The authors noted that while the excavations of the quarry showed evidence of sustained (Mesolithic to Bronze Age) and substantial use (over 30,000 lithics excavated from the site), lithics made from rhyolite were found almost exclusively in a very local area around the quarry (Zvelebil et al. 1987, 18). This apparently highly localised nature of the rhyolite use led the authors to four questions:
“(1) How intensive was, in fact, rhyolite production, bearing in mind the high waste to finished product ratio of its reduction process? (2) Who was in control of the rhyolite source – perhaps a group not resident in the region? (3) Was rhyolite obtained for export, and if so, what was its destination? (4) Why did local the local people ignore the advantages of this raw material [over the local small pebble flint] for making large and medium-sized tools?” (Green and Zvelebil 1990, 70).
The predominant raw material was flint, which came from small beach pebbles, with a small amount of chert, basalt, and quartz (ibid.). The authors suggested that the coastal and estuarine areas were favoured in the Mesolithic compared to inland, and the lack of riverine evidence in this area may be due to peat formation and alluvial events (ibid.).
The survey then continued in following years, where they moved some 70 km upriver from the estuary. Here the (prehistoric, not specifically Mesolithic) lithic raw material changed, with a preponderance of chert, basalts, andesites and rhyolites – flint accounted for 20% of the lithics, with the authors suggesting that the larger flint tools were imported from outside the region, as the available flint nodule size was too small for such tool-making. 5% of the lithics were quartz or quartzite (Zvelebil et al. 1996, 21-3). The density of lithics was smaller in the basin in comparison to the estuary, leading the authors to suggest this may have been due to alluvial deposition, the use of the coastal area as primary reduction sites, hence creating more artefacts, or due to the ease with which worked flint in the estuary was spotted by the researchers, compared to the basalt/rhyolite material which is not as straightforward to identify as worked (ibid., 31). The authors argue that a clear pattern of the Mesolithic material is for it to lie in the alluvial/till boundary, with the authors suggesting that much Mesolithic activity which would have occurred along the river is buried below alluvium (ibid., 36).
The Bally Lough Project is of clear importance to Irish archaeology, tackling a for the most part devoid archaeological landscape and producing evidence from the Early Mesolithic onwards. However, the various papers are scattered widely over differing journals and book sections, and the project would seem to have run out of steam in a sense. There was no real integration of the coastal and upriver sections tackled, and no move to interpret the findings in a greater or more lucid fashion – beyond a substantial focus on taphonomic issues – such as has characterised Zvelebil’s writings in other works (e.g. 1997; 2003).Top of Page
Over the last few years there has been a dramatic increase in the discovery of Mesolithic human remains. These have been recovered from a shell midden, and more recently from two caves, Ferriter’s Cove, and from a number of cremation burials located on the banks of the Shannon. This section will discuss the various lines of evidence and discuss the interpretations of the burials.
The key discovery in the 1970’s of an Early Mesolithic presence on the Shannon system in the Midlands altered the picture of the Irish Mesolithic as it was realised that areas where there had been no previous research or collecting, could produce evidence. A quarter of a century later a series of Early Mesolithic cremations were excavated on the banks of the Shannon, at Hermitage, Co. Limerick (Collins and Coyne 2003). This excavation will fundamentally alter perceptions of the Irish Mesolithic.
The original date returned for the first cremation (pit A) was 7550-7290 cal. BC, and 7030-6630 cal. BC and 6610-6370 cal. BC for the other two cremations (a Bronze Age date was returned for another cremation at the site) (ibid., 25-6). These dates were from charcoal samples, and at a conference in May 2006 Collins (2006) announced dates for the cremated bones which are similar to the original dates: 7530-7320 cal. BC for pit A and 7090-7030 cal. BC for the cremation from area B.
The cremated bones from pit A were taken as representing one adult individual, possibly male – due to the small size of bones, no more detailed osteological determinations could be made, and the determination of male was questionable (Lynch 2001, 112-3). The weight of bone recovered was 1979g – which is above average from archaeological contexts – suggesting that the whole body was cremated, and great care taken in collecting and depositing the remains into the pit – the reddened clay with imbedded bone from the pit may be from the pyre, suggesting the shovelling of material from the pyre (ibid., 115). Lynch commented that the cremation was “expertly carried out”, and “the process of cremation would have been a considerable effort, requiring significant planning and a major input of time and effort” – the author argued that the condition of the bones in the pit suggests that the people who undertook it were familiar with the processes of cremation (ibid., 114).
Pit A had a post in it which was suggested as being a grave marker; against this was placed a polished stone axe: “the cremated remains were then deposited around the axe and post, in a semi circular arrangement” (Collins and Hayes 2001, 80). Along with the axe, 17 lithics came from the same context of the pit, with one of these a diagnostic microlith, and another a possible microlith (Woodman 2001, 121). Three of the lithics (including the microlith and possible microlith) and the axe from the pit showed evidence of burning (Collins and Coyne 2003, 25; Woodman 2001, 121). For the lithics in general from the site (chert, flint, siltstone, mudstone, and possibly shale and limestone), Woodman suggested that it was an Early Mesolithic assemblage, with diagnostic Later Mesolithic material as well (ibid., 126-7).
The cremated bones from area B were taken as representing one adult individual, possibly male – again, due to the small size of bones, no more detailed osteological determinations could be made, and again the determination of male was questionable: along with the 179g of bone, a tiny amount of cremated fish bone was also present (ibid.). The human remains well less well burnt in the cremation compared to pit A, and while pit A was suggestive as having been an entire body, the area B cremation may have been only selective parts, as the skull was considerably overrepresented while torso and limbs were underrepresented – Lynch noted that the amount of undiagnostic bone may be creating this bias (ibid., 116). Eleven chert flakes and flake fragments were from the came context, along with nine naturally fractured pieces of chert and one piece of flint debris (Woodman 2001, 123).
Due to a lack of identified warping of the bones from both cremations, it was argued that this may mean that the bodies were placed on top or in the middle of the pyre (ibid., 115). Further, the fragments of bones in both cremations were very small, and this was taken as possible evidence that some process such as pounding or grinding of the bones was undertaken after the cremation was completed, as part of the burial ritual (ibid., 118). While a third cremation returned a Mesolithic date, only one bone is registered in the finds (Collins and Hayes 2001, 97), and this is not included in the osteoarchaeological analysis – six lithics also came from the fill (Woodman 2001, 124).
These are the earliest dates for cremations in Europe, and the earliest dates for any type of burial in Ireland. Given the early dates returned for the bone, the comments by the osteologist relating to the cremation practices – made before the material was dated – are of considerable importance. These comments highlight that this was not a haphazard disposal of a body, but a carefully executed cremation. It is a sobering thought that the analysis suggests that the cremation undertaken was of a developed, sophisticated nature: a cremation practice that presumably was of some antiquity at a time of its undertaking. The cremations at Hermitage highlight a persistent place (sensu Pollard 2000) in the landscape– the reuse of a locale for burial rites over a period of many centuries, with evidence also for continuity of place and rite in the Bronze Age. While no Neolithic cremations were uncovered, this development-led excavation only examined a 10m wide strip, therefore further evidence for Mesolithic and later periods’ burials is certainly a possibility.
Human remains were also discovered during excavations at Killuragh Cave, Co. Limerick; the cave overlooks the Mulkear River, a tributary of the Shannon and lies some 15km as the crow flies south east of Hermitage. The excavations produced, along with lithics, Early Mesolithic human remains dated to 7194-6658 cal. BC and 7000-6546 cal. BC; Later Mesolithic human remains gave dates of 4730-4460 cal. BC and 4460-4160 cal. BC; Neolithic human remains gave a date of 3624-3348 cal. BC, as well as a dog bone dated to 3940-3630 cal. BC (Woodman and O’Shaughnessey 2003). Again, this highlights the use of persistent places in the landscape over considerable periods. Woodman and O’Shaughnessey (2003) argue that while the human remains are related to the cave, they do not represent burials inside the cave as such, but possibly “either introduced or washed into the cave from outside” (ibid.).
Whether the continuity of this cave as a locale for the deposition of human remains is a signature of a continuity of the same communities’ from the Mesolithic to the Neolithic is a moot point. The difficulty is in interpreting whether a persistent place such as this represented: either 1) an ancestral burial place to the Neolithic inhabitants, 2) an explicit appropriation of renowned Mesolithic locales by intrusive Neolithic communities, or 3) an incidental reuse of a landmark in the landscape. Or indeed, whether the supposed Neolithic dates are in fact late Later Mesolithic! The orthodox presumption is that all post-4000 cal. BC dates are Neolithic. However, a lack of Mesolithic sites after 4000 cal. BC simply means a lack of excavated dated sites, rather than a past reality that the Mesolithic communities and their way of life were vanquished on the convenient tidy date of 4000 cal. BC. After all, the Mesolithic and Neolithic are our analytical constructs as opposed to something historically meaningful.
The same argument can be made for the Stoney Island bog body from Co. Galway. While O’Floinn (1995, 139) has suggested this skeleton retrieved from a bog as the earliest Neolithic example, the calibrated dating by Brindley and Lanting in the same book (1995, 134) gives four calibrated dates: 4226-4196 cal. BC; 4152-4056 cal. BC; 4052-3958 cal. BC; and 3840-3826 cal. BC (the authors suggest that the earliest date is a contaminated sample). So we have a body which either: 1) straddles the transitional period, 2) falls up to 150 years before the traditional transition date, or 3) falls over 150 years after the traditional date – consequently there is over a 300 year period in which this person may have lived and died, or in other words some 15 generations. Woodman (2000, 40) has suggested that the 13C results of 21‰ for the Stoney Island bog body “would suggest a land-based diet which is more likely to indicate a farming economy”. But this is forgetting the evidence for the ‘land-based diet’ results (-21.9‰ and -21.3‰) for the Later Mesolithic remains of Killuragh Cave mentioned previously; the Early Mesolithic remains at the cave also returned similar 13C results: -19.95‰ and -20.86‰ (Woodman and O’Shaughnessey 2003).
Recently, human bones which had been found by cavers in a cave in Srahmore, Co. Leitrim have been dated to the Later Mesolithic or Early Neolithic (4217 (4030) 3967 cal. BC (two sigma)) (Dowd, pers. comm.), with the dates being remarkably similar to the Stoney Island bog body. This cave, which lies on the north face of Srahmore Mountain close to the Sligo-Leitrim border, has not been excavated so the fuller picture of this site remains to be revealed.
At Ferriter’s Cove several pieces of human bone and teeth were found, one dating to 4225-3950 cal. BC, and the other to 4250-3980 cal. BC (Schulting 1999, 219). At Rockmarshall Co. Louth, middens overlooked a former coastal lagoon in Dundalk Bay, and a human femur was dated to 4774-4366 BC (O’Sullivan 2002, 11). O’Sullivan has highlighted some current ideas in Mesolithic research, touching on the notion of the middens as visible features in the landscape, recurring places for visitation; he describes the routines of such visits – sitting, chatting, eating, and looking out over the landscape (ibid.). He suggests that what the bones were doing in the middens can be read in a number of ways – from Scottish evidence, he relays that the middens may have been used as resting places for bodies while they rotted, and once they had rotted the bones were removed, with some accidentally left behind; alternatively, the bones may have been purposively deposited in the middens, as a place of ancestry (ibid.).
O’Sullivan’s paper has placed some of the known Mesolithic burials in Ireland in context of wider European discoveries. However, he has overstated the Irish case for midden burials, as the bones and teeth from Ferriter’s Cove were not found in the context of the middens – as previously mentioned, the deposits of shells at Ferriter’s Cove were more discrete dumps than substantial middens: these would not have stood out as visible cultural markers in the landscape. So the only human remains from a midden in Ireland is from Rockmarshall. A perplexing omission in the article is any discussion of the Killuragh Cave remains, as this site clearly adds to the variety of locales used for mortuary practices. While at the two coastal sites the interpretation can be made for the bones of ancestors remaining close to daily activity, the cave depositions may represent a liminal area, removed from daily routines. While the other sites were out in the open, in plain sight for all to see, the cave is more secluded, indeed hidden. So, rather than one tradition, there was a variety at play contemporaneously.Top of Page
In this section I will turn my attention to the Mesolithic-Neolithic transition, and outline the evidence that has been put forward by researchers in Ireland for the introduction of agriculture. In comparison to the Irish Mesolithic period, the Neolithic has seen a strong focus of theorising and research over many years, with the highly visible megalithic monuments a magnet for researchers. There has been an abundance of publications on the Neolithic over the years, and also on the question of the arrival of farming practices onto the island. Therefore, to keep it succinct, I will only touch on aspects of the Neolithic pertaining to the question of the transition period itself as much as possible, and will not dwell on the megalithic research to any great degree which has a bountiful reading list on its own. I will start with the Ballyglasss Project’s work on the passage tombs in Sligo and Burenhult’s interpretation of the economy and society in the Neolithic period, and then outline Woodman’s concerns with the use of radiocarbon dating in assessing the transition period. I then turn to the Bally Lough Project’s work on the introduction of farming in the South-east, and Kimball’s work in Donegal. This will be followed by Sheridan’s, Tresset’s, and Woodman’s disagreement with the forager-centric model as put forward by Burenhult and Zvelebil et al. The Neolithic house is seen as an arrival of an architectural tradition and seen in strong contrast to the Mesolithic societal structure, so I will dwell awhile on this topic, touching on the megalith phenomena as I do. Finally, I will discuss the stable isotope models that have been put forward and finish with the palynological evidence for farming.
As I have outlined in the previous chapter, the Neolithic as an entity has evolved since its inception in archaeological circles, with it now being firmly linked with the practice of agriculture. The earlier theorising by Knowles and his contemporaries did not make this link. Indeed, they reasoned that the first Neolithic inhabitants did not practice the arts of agriculture at all (Knowles 1896, 663). A few decades later, the Childean revolutionary vision placed farming at the forefront of the Neolithic (1927, 1). It has remained there since. However, what has changed is the emphasis placed on agriculture itself as an entity inside the Neolithic – baldly, from a social evolutionary perspective of people progressing to the stage of an agricultural economy (Childe 1927; Macalister 1949), to an economic perspective (with hang-ups from the previous perspective) (Burenhult 1984; Case 1969; Herity 1974; Woodman 2000), to the more current social transformation model whereby the social realities and ideas of farming are as vital as the economic determinations (Cooney 2000; Thomas 2001; Tilley 1994; Whittle 1996). Along with agriculture, new types of material culture such as lithic types and pottery and especially the megalithic phenomenon is tied in with the Neolithic, and therefore tied in with farming. Case’s (1969) thesis that the megalith building came after a period of initial farming, once a sustainable population and economy was in place, was argued against by those who thought of the megaliths as artefacts of pioneering agriculturalists staking their hold on the land (Herity 1974).
Burenhult’s Ballyglasss project was designed specifically to test the theory as to whether the indigenous Mesolithic populations were responsible for the construction of the monuments, and to examine the cultural ecology and economy of these builders, and how this changed diachronically (Burenhult 1984). Here, the builders of the megaliths were seen as being Mesolithic in terms of their “socio-economical” character – a key part of the evidence for this was the “offerings of unopened sea-shells in the excavated monuments” (Burenhult 1980, 5). Burenhult argued that cattle did not play a part in their economy, but rather it was a “complex hunter strategy” that typified the economy, which led to an increase in population density, and social complexity – these factors led to the construction of the megaliths (Burenhult 1984, 138-9). The key evidence to support this were the middens and the hut sites, which were taken as suggesting a seasonal round of settlement. Later, ‘Neolithic’ elements such as cattle and cereals were introduced through contact with farming communities of Eastern Ireland (ibid.).
For Burenhult, the Mesolithic-Neolithic transition in Sligo was mediated on an indigenous population developing to a stage of complexity which eventually led to the introduction of farming practices such as cattle rearing and cereal growing, but that the megaliths themselves were built by Mesolithic people. The key factor missing from this interpretation was that the usual Later Mesolithic assemblage seen throughout Ireland, the macrolithic butt-trimmed forms and so forth were absent from all sites investigated by Burenhult’s team. Therefore, no site investigated could be considered Later Mesolithic, leading to the conclusion that these sites were from a later period – in other words Neolithic in date.
Woodman (2000) has set out the dilemmas of the discussion of the transition in terms of the radiocarbon dating evidence used. He argues that there is a gap from Later Mesolithic sites dating up to 5500 BP, to Neolithic sites dated to after 5000 BP, with no reliable dates in between (ibid., 242). Moreover, he cites Pilcher who maintained that to attempt to use conventional radiocarbon dating to interpret real ages closer than 500 years were not valid (ibid., 226). This would suggest that conventional dating cannot be used to ascertain historical events such as complex transitions. Added to this are the fluctuations in the 14C levels in the atmosphere, which are interpreted as being “considerable” around 4200-4000 BC (ibid., 240).
Woodman suggests that supposed early dates for Neolithic houses, such as Ballynagilly, Co. Antrim and Ballygalley, Co. Antrim tend to be from wood or charcoal, which may create a false sense of antiquity to the buildings (ibid.). McSparron (2003) produced a list of radiocarbon dates from nine Early Neolithic houses which come from material not susceptible to the old-wood effect (four of these houses are from the group of seven recorded at Corbally, Co. Kildare). This list highlights how the early dates given are put into perspective by using non wood samples, and that – for these houses at least – a start point of c. 3800 cal. BC seems apparent, which falls 200 years after the start of the Early Neolithic (sensu Cooney 2000).
Earlier I discussed the Bally Lough Project in terms of the Mesolithic evidence. A key objective of their project was also the investigation of the agricultural transition in the region “in order to explore the roles of indigenous and external influences” (Green and Zvelebil 1990, 59). They argued that the ensuing Neolithic period saw an increase in the diversity in the lithic assemblages, and an adoption of pressure flaking “and polishing as techniques to extend edge life”. As mentioned they considered chipped and ground stone axes as Neolithic (ibid., 64). Peterson (1990, 93-4) considered the use of bipolar working as a continuation from the Mesolithic, however Woodman (pers. comm.) has disagreed, arguing that there is no evidence of the use of the bipolar method of working in the Mesolithic. They argued that the evidence from the coast and estuary showed a geographical continuity over the periods, and that this was matched in the middle Barrow Valley, with an extension in the Neolithic into upland areas (Zvelebil et al. 1996, 36). Rather than seeing foragers being displaced totally by arriving farmers, they argued that one must consider an overlap, a zone of interaction in which both the foragers and farmers played respective parts (Green and Zvelebil 1990, 58). They raised the possibility that the initial arrival of farming may have come from the continent and from Ireland spread to Britain, with the suggestion that this may have occurred through contact but not necessarily through a substantial population movement (ibid., 86).
Kimball’s Lough Swilly Survey was designed to investigate the Neolithic transition in Donegal, and to compare his results to that of the Bally Lough Project. He argued that whereas The Bally Lough results showed continuity in geographical locations over the transition, his results showed that Mesolithic evidence was located solely in aquatic locations, and post-Mesolithic was more widespread (Kimball 2000A, 77). Further, he argues that discontinuity is apparent in the lithic material use and procurement strategies, with Later Mesolithic characterised by the use of non-local material, and the post-Mesolithic characterised by local, apart from the axes (ibid.).
In a number of papers Sheridan (2003A; 2003C 2004) has also been highly critical of the idea that indigenous foragers were responsible for the introduction of farming to Britain and Ireland, with little population movement involved; she has been particularly scathing in her attack towards Julian Thomas, calling this forager-centric thesis the “Thomasian orthodoxy”. However, he clearly was not the first or last to suggest these thoughts. She argues to forget the latest theoretically-driven models and to look at the clear empirical evidence. Pottery styles (ibid, 7-9) and the fully formed techniques of ceramic production (Sheridan 2004, 12), point to the arrival of culture groups in toto from the continent. Along with the pottery are the early tomb types which she also argues highlight a French connection (Sheridan 2003C, 12).
Tresset (2003, 25) concurs with Sheridan and argues that: “it would appear wholly far-fetched to posit that local Mesolithic groups sailed to the continent and brought back domesticated animals”. But where is the material culture of the continental farmers that brought the cattle? “One could imagine that the new incomers failed in their attempt to settle in Ireland. [Or] erosion, silting-up and Holocene sea level changes…may have washed away any such traces of littoral settlement” (ibid., 26). She argues that the early dates for cattle, such as Ferriter’s Cove, may represent a continental connection which failed, and the later “Early Neolithic” cattle evidence in the form of the settlements at Tankardstown, Co. Limerick and Cloghers, Co. Kerry dating to after 4000 BC “could be linked to some connection with Britain, as suggested by the association with the Carinated Bowl tradition” (ibid., 27).
Woodman concurs with this view for the early cattle remains and argues that, as well as the pottery, the lithic assemblage of the Irish Early Neolithic (large leaf-shaped arrowheads, convex end scrapers, plano-convex knives (Woodman 1993)) resembles that of Britain and significantly differs from that of the lithic technology of the Irish Later Mesolithic (Woodman et al. 1999, 149). He has strongly disagreed with the thesis put forward by Zvelebil et al., arguing that simply because people from the two periods were using the same area and resources, it does not follow that they were the same people, or that the Mesolithic communities had an input into the arriving Neolithic way of life – it simply shows that they were using the same landscape, and that the coast and so forth were important in the Neolithic as well.
The phenomena of the Neolithic house has become a key feature in the debate in the Neolithic transition, as the sheer numbers found, over 90 according to Grogan’s (2002) latest publication on them, stands in stark contrast to the known Mesolithic dwellings, and Cooney (2000, 14) argues that these buildings are evidence of a defined architectural tradition. Added to this, Cooney (ibid., 54) has remarked for the Neolithic in general that the house evidence can help researchers move beyond megaliths as the research point of contact with the Neolithic, and will balance out the idea that Britain and Ireland were typified by a lack of settlements in this period. Moreover, the positioning of these houses can do away with the concept of tombs as proxies for settlement (sensu Cooney 1983).
The earliest examples of the Neolithic houses are suggested as being for the most part larger than those of the Middle Neolithic (Grogan 2002, 521) and this may well be a significant factor in the meaning and use of these buildings. The earliest types also tend to be rectangular, but Cooney (2000, 66-7) noted that this does not strictly play out for every site, and he suggests that the individual site history’s should be examined, rather than an all-Ireland type framework. Grogan (2002), and especially Cooney (2000), have discussed Neolithic houses at length, so here I will briefly mention four different sites that have been published to some degree, which, while showing the commonality of the template, highlight the peculiarities of them.
The first is Cloghers, Co. Kerry, mentioned by Tresset above. This house is the closest known Neolithic activity to Ferriter’s Cove, which lies some 50km out on the Dingle Peninsula. While the excavation of Ferriter’s Cove was initiated on the discovery of a plano-convex knife, little further Neolithic evidence was forthcoming (Woodman et al. 1999). Within a few years, development around the town of Tralee has transformed the known Neolithic activity in the area. A few kilometres from the Cloghers house, other excavations have uncovered a passage tomb and hilltop enclosure on a limestone reef, 50m south of the River Lee at Ballycarty (Connolly 1999), and a preliminary report for another nearby site stated that this was a complex of pits, stake holes and post holes that produced “Early Neolithic” pottery along with lithics (Dunne 2000).
The rectangular house at Cloghers was situated on the north slope of the same limestone reef overlooking the river which met the sea a short distance away (Kiely 2003, 182). While the external measurements of the house gives it an area of c. 100m² – making it the largest Neolithic building recorded in Ireland – Grogan (2002, 519) notes that as the building was divided into three compartments, he gives the area at 72.45 m². This building was both plank built and of posts and the builders cut into the bedrock in places to make the foundation; the lithics of greenstone, quartz, chert, and flint as well as ceramics were found in the foundation trenches, including a quartz core that was placed at the base of one of the posts (Kiely 2003, 184). Kiely (ibid., 187) suggested that a focus of deposition was at the entrance of the building. A large number of stake-holes, pits, and what was taken as a fence line, were also excavated outside of the house, and there was evidence for the making of ceramics on site (ibid.). The house was argued as having been destroyed by fire (ibid.). The dating of two hazelnut shells gave dates of 3695-3630 and 3575-3535 cal. BC for the first and 3765-3640 cal. BC for the second. This dating suggests, contra Tresset and Grogan, that this house was occupied towards the end of the Early Neolithic.
A group of seven houses were excavated in Corbally, Co. Kildare. Purcell (2002) excavated three of these, and noted in her article that four others had since been discovered. For the three that she excavated, she was unsure of the contemporaneity of the houses, suggesting that they were occupied within a “very short time-span” (ibid. 70). All three were post and plank built, and the lithics and ceramics from the three houses were similar, but with more retouched tools recovered from house 1 compared to the other two (ibid.). House 1 had a living area of c. 53 m², while house 3 had a living area of c. 37 m² (ibid., 34; 59).
Only one animal bone was recorded – this was from an internal posthole in house 3 (ibid., 72). One of the houses returned dates for a single grain of Triticum dioccum of 3775 -3631 cal. BC, 3581-3567 cal. BC, 3561-3537 cal. BC; a fragment Corylus avellana of 3891-3881 cal. BC, 3799-3641 cal. BC; and charcoal of 4240-3925 cal. BC, 3875- 3810 cal. BC (Purcell 2002, 46). The wood and charcoal dates are significantly earlier than the grain of wheat, which brings this house closer to the Middle Neolithic sensu Cooney (2000, 14), not Early Neolithic as stated by Ó Drisceoil (2004, 180) and implied by Grogan (2002).
While Purcell (2002) noted that there are no known megaliths in the Co. Kildare, another recent excavation of a Neolithic structure, this time in Drummenny Lower, Co. Donegal, saw the structure on the crest of a ridge, overlooking a court tomb 200 m away which sat beside a river in the valley below (Dunne 2004, 165). In discussing the excavation, Dunne (ibid.; 170) commented that she was avoiding the term house as the function of this rectangular, c. 60m² building was uncertain, for one because “the occurrence of a dramatic incline in the floor surface would have made occupation somewhat uncomfortable”. The floor declined in height by 1m over the width of the building, which was c. 6.3m wide. It was argued that the building was destroyed by fire, and that a possible cremation pyre was also noted 30m away (ibid., 170).
While the Drummenny Lower building overlooked a court tomb, at Ballyglasss, Co. Mayo a court tomb was built on top of the foundations of a dismantled rectangular post-built house (Ó Nualláin 1972). So we can see that these buildings and megalithic structures were both intimately related, and, in the example of Kildare, not related. Cooney has commented on the placement of monuments, suggesting that a common theme is the use of places already renowned. The biography of the place was elaborated through the construction of monuments there, with the monuments, like the houses, having biographies (Cooney 2000, 125).
In terms of the identification of Early Neolithic monuments and their part in the Neolithic transition, the dating is extremely difficult. In her article on the chronology of megalithic tombs, Sheridan (2003B) regards passage tombs as being early, but in terms of the court and portal tombs, she regards it as difficult to state. She suggests that the idea of a transformation from a wooden mortuary feature to a court tomb is plausible for the Ballymacaldrack, Co. Antrim megalith and in terms of the portal tombs, she sees them as being contemporary with the court tomb (ibid., 70). Cooney (ibid., 14) argues court and passage tombs can be considered early, while Whittle (2004, 81) suggests portal tombs may be early types as well.
Caulfield’s work has revealed extensive field boundaries of stone and earthen banks along the coast of North Mayo, especially around the area of the Ceide Fields, which also have evidence for round and rectangular houses and megaliths, such as at Ballyglasss. O’Connell and Molloy (2001) have reviewed the evidence for farming in the Early Neolithic, and in the case of the Ceide Fields, there are two pollen profiles, taken 16 km apart, one by the coast at Glenura, and the other inland at Garrynagran. Both are associated with pre-bog walls.
The Glenura profile showed a localised picture of a c. 500m radius; up to c. 3900 cal. BC this showed a pine dominated woodland, with farming initiated before the elm decline of c. 3900 cal. BC (ibid., 103-4). They suggested that the intensive farming period was from c. 3800 cal. BC and argued that the field boundaries may have been built then. They noted an increase in tree growth, especially hazel from c. 3450 cal. BC.
The Garrynagran profile – which showed a regional picture – the pre-elm decline woodland was dominated by hazel, elm, and pine, and lesser amounts of oak, and unlike the Glenura there was no evidence for pre-elm decline forest clearance (ibid., 108;110). As mentioned in the previous chapter another study on the area has shown that an area outside the known distribution of field walls produced evidence for extensive farming outside of the known distribution of the pre-bog walls. Again it is unclear whether this could signify a chronological marker or differing farming practices in the region.
Down the coast at Lough Sheeauns, Co. Galway, another profile was taken in an area surrounded by megaliths, but not associated with Neolithic field boundaries (ibid., 110). Here, the pre-elm decline forest was dominated by oak, with hazel, alder, and birch present, and elm and pine in lesser amounts; an opening up of the forest (possibly regionally) was noted at c. 4430 cal. BC with the authors suggesting natural or possibly human causes (ibid., 113). After the elm decline there is a “dramatic landnam phase”, and by c. 3600 cal. BC woodland regeneration.
As mentioned previously, stable isotope analysis has been used to understand dietary practices, and this has become an integral part in discussions of the Neolithic transition. The use of stable isotope analysis to indicate either a marine or terrestrial source of protein in the diet is hotly debated, with polarised views as to whether the results can be clear cut evidence of differing dietary patterns. On the one hand it is argued that a (more or less) straightforward case can be made (Schulting and Richards 2000) – the following table (Table 3-1) from Schulting (1998, 206) presents a schematic view of the differing stable isotope results. On the other hand it is suggested that a myriad of factors can weigh against such interpretations (Milner et al. 2003).
Table 3-1 Schematic view of the differing stable isotope results; table from Schulting (1998, 206)
The interpretations of the stable isotope values from Britain, France, and Denmark suggest that there was a sudden shift away from a marine diet beginning in the Neolithic (Schulting 1998, passim). Milner et al. (2003, passim) argued that the suddenness of this shift was not so sudden if the dates are compared, and suggested that the sampling size is meagre. Further, they argue that the variable of “intra-population dietary variation related to gender, age or status” is not accounted for adequately in the interpretations (ibid., 15). They also suggest that “consumption of marine foods does not always result in a marine stable isotope value”, as these values can vary, for instance with the salinity of the water, and also that the introduction of cereal agriculture may have an effect of masking a marine component to the diet (ibid., 16-7).
In discussing his stable isotope analysis, Schulting (2004, 23-5) argues that “cows and fish don’t get along” commenting that “there is little in common between fisher folk and pastoralists”. Citing ethnographic work from Africa and North America he argues pastoralists culturally disinclined towards fish; physiologically, switching seasonally from fish to meat “can cause severe illness” (ibid., 25). However, it is arguable that you do not have to leave Ireland for ethnographic work of the relationship between pastoralists and fishers and the compatibility of the differing protein intakes (see Pl. 3-1).
Clearly it is not an either/or choice, or a matter of only eating one or the other in a separate season. Indeed, Caulfield (1983, 213) made this point in his retort to Burenhult’s thesis: “it can be argued that the persistent fisher/farmer communities which one finds along the Connaught coast today is a more appropriate model for settlement when the megalithic tombs were being constructed”.
These dates are about 1000 years older than traditionally cited (Woodman 1985; Waddell 2000), as recalibration has pushed them back significantly (Woodman 2003); a recent reanalysis of the calibrated dates published in his 2003 paper, has brought the dates published there forward slightly (Woodman, pers. comm.).[return].
This has now changed as there is evidence of Later Mesolithic structures from the Toome bypass excavations (Prudame 2004); (This site is unpublished).[return].
Attempts to build such sod structures suggest that it could have been made entirely of sods (Woodman 2006C).[return].
As mentioned, a recent reanalysis of the calibration of these dates has brought them slightly forward in time, starting from c. 7900 cal. BC (Woodman, pers. comm.).[return].
Some axes in the National Museum collection have recently been interpreted as showing signs of use as ardshares, due to the scratch marks (MNI 1987:184, Dangan Lower, Co. Galway); this may be representative of a more common use of axes as digging tools.[return].
Recently, Sternke (pers. comm.) has looked at the Lough Gara assemblage in order to include it in the new Irish Mesolithic database - she did not find any cores that she thought could be Early Mesolithic.[return].
The author does not state whether this date is calibrated or not.[return].
Woodman (2006B) has commented that there would seem to be pre-midden activity as well.[return].
It is not stated whether this is a calibrated date.[return].
It is not stated whether this is a calibrated date.[return].
It is not stated whether this is a calibrated date.[return].
In terms of the loose use of dating, Fredengren (2002, 127), also citing Brindley and Lanting, states the dating of this person as being "calibrated to 4230-3940 BC". It is unclear how she arrived at these dates. It would seem she roughly bundled the first three dates together, and ignored the latest date - however this would still not tally. If she has grouped these dates, it is a perilous use of dating to present three separate dates in such a manner. Interestingly this dating of "4230-3940" is also given by Sheridan (2003B, 70) for this burial in her paper on chronologies of megaliths.[return].
As noted, Woodman does not regard the Mesolithic human remains as burials within the cave itself (Woodman and O'Shaunessey 2003).[return].
A similar slope was evident in the Gortaroe II Neolithic house: "the house was built on a slope with a gradient of 1:8 across its short axis (i.e. the northeast of the house was almost 1m higher than the southwest side)", which was 6.8m wide (Gillespie forthcoming, 104; 127).[return].