A lithic scatter near Aberdeen under forensic examination
Nethermills Farm lies west of Aberdeen, in the settlement of Crathes on the north bank of the River Dee. Ancient stone tools have been collected from the cultivated fields here since the mid-20th century. They can be found over some 2km, and indicate a long history of activity from the Palaeolithic to the Bronze Age; the total collection numbers over 40,000 pieces. Excavation by James Kenworthy 40 years ago focused on a significant Mesolithic site, but investigated only a tiny portion of the area of archaeological interest.
The different collecting episodes have all been separately – but relatively uniformly – catalogued, making it possible to examine the development of our understanding over several decades. While previous research has often considered farming, and how finds moved about in the ploughsoil, little attention has been given to archaeologists: what differences are there between material found in an initial sweep across a field, and later collections? Many scatter sites are fieldwalked only once, with the perception that the sample recovered is representative. Such sites are an important archaeological resource, and understanding how we create them matters. I and my colleagues – listed in the endnote – wondered if Nethermills Farm could help?
Decades of collecting
Lithics were first identified in the fields at Nethermills Farm in the spring of 1973 and there have been six main periods of collection on the eastern fields (table opposite). As might be expected, analysis and reporting of this material has varied. Roger Daly’s collection, around 500 pieces found in the 1990s, came to light only when we were test-pitting in 2019, and we have not yet been able to catalogue it.
Starting in 1973 John Grieve collected across two adjoining fields but focused on three prolific hotspots. He lacked formal training, but compiled a basic catalogue, carefully studying relevant literature (“Mercer’s
What archaeologists call lithics – stone tools and manufacturing debris – lie scattered across the UK wherever people lived and worked in the more ancient past. Collections from cultivated land are a critical source of information. But how reliable are they? As Caroline WickhamJones explains, a long-running project near Aberdeen is helping to evaluate fieldwalking
paper influenced, but not very impressed”, he wrote of one report). His collection was archived by Aberdeen City Museums and never published.
Kenworthy excavated an area of roughly 110m² under Grieve’s easternmost concentration. He analysed his finds extensively, but never completed the work. Torben Bjarke Ballin (a professional lithic specialist) resumed study in 2012, but as everything had been individually wrapped and bagged – around 30,000 pieces – it was impractical for him to examine more than a small portion. Kenworthy identified 2,750 “special” pieces, which we published as part of the excavation report in 2017.
ofars – the Over Fifties Archaeological Research Society – covered both fields using 2m transects. Their collection has been catalogued in detail and published online by Heather Sabnis (trained by Kenworthy). Mesolithic Deeside also used 2m transects and focused on the easternmost field, where test-pitting took place in 2019; Ann Clarke (a professional lithic specialist) has analysed all that material.
Kenworthy set out to investigate a Mesolithic site indicated by Grieve’s fieldwalking, finding stratified remains, including postholes and fragments of timber with bark remaining. In his final post-excavation work in 2016 he was unable to verify the structures’ existence, and sadly test-pitting in 2019 indicated that continuing cultivation has destroyed this evidence. Nevertheless, it is clear that Mesolithic people repeatedly visited the area over a considerable period of time. Most of the excavated material related to the Late Mesolithic (8000–4000bc), but there was also some Neolithic and Bronze Age material (4000–1000bc), together with seven items of apparent Late Upper Palaeolithic date (13,000– 9500bc). Radiocarbon determinations confirmed the principal activity dates from the sixth millennium bc to the late second millennium bc.
Blades, flakes & cores
Grieve’s collection was catalogued and rebagged in 2018 to provide secure storage, and standardise recording with more recent work from the field. This allowed us to scrutinise an early collection (an oft-neglected resource), while also thinking about how site information accumulates across the decades. Reflexive work such as this, often overlooked, is an important part of fully understanding lithic scatter sites.
The most obvious result of Ballin’s work on the Grieve Collection was to provide an accurate count. The collection contains material from several sites, and Aberdeen City Museum had previously registered a total of 6,736 pieces, while an earlier study by Daly noted 8,775. There are 5,701 pieces in the collection today. Discrepancies are likely to be due partly to problems with the original bags, and partly to the vagaries of previous research. In one or two cases whole categories of artefact are missing, and it is possible that material has been removed for study or exhibition and not returned.
At Nethermills Farm Ballin recorded more pieces in Grieve’s collection (3,244), than originally listed (3,151), though the difference is small (3%). Nevertheless, a number of unquantifiable biases have a bearing on the study, including the specific interests of individual collectors, the weather and the differing expertise of particular analysts. While scrutiny suggests that analysis has been reasonably consistent, some
differences in definition and recognition are inevitable. In addition, over 90% of the excavated material has not been catalogued, and what has been analysed was selected because it was “interesting” in some way. We have ignored the rest of the excavated material here because the principal focus lies on how knowledge about the site has accrued.
At first glance, there is some difference between the absolute totals of material from different collecting episodes. ofars amassed the largest fieldwalked assemblage (4,593 pieces), while both Grieve and Mesolithic Deeside collected around 3,000 pieces. Grieve recovered mainly flakes, but also many blades (long narrow flakes) and cores, and though he was interested in Mesolithic material, there were few microliths. In general, Grieve recovered larger, chunkier pieces, though it is interesting how little debitage – the roughest material – he collected. Ballin found Grieve’s blades to be wider (9–13mm) than those from the excavated sample (5–8mm). He also noted that Grieve collected more pieces with evidence of hard-hammer percussion than is common among local assemblages. Grieve did, however, recover Neolithic and Bronze Age arrowheads – three leaf-shaped points and eight barbed-and-tanged points. On balance, the lack of debitage suggests that Grieve may have been cherry-picking things that most interested him.
ofars were particularly good at recognising blades and flakes. While they recovered more debitage than Grieve, their collection contained fewer cores, but they were good at spotting microliths. ofars recovered both broad and narrow blades (a width range of 4–17mm), and their collection contains one large, unusual blade that may be Palaeolithic, as well as three leaf-shaped points. Mesolithic Deeside also collected a sizeable assemblage and were good at recovering debitage as well as blades and flakes. They collected many cores and, in general, they recovered many smaller pieces. Analysis by Mesolithic Deeside indicates that, while the spacing of fieldwalked transects influences the size of a collection, it has little impact on overall interpretation.
Not surprisingly, given that over 90% remains unanalysed, the excavated assemblage is different. While it contains fewer cores than the Grieve or Mesolithic Deeside collections, it has the greatest number of microliths. Whereas Grieve recovered nine broad-blade microliths and only one narrow-blade, most of the excavated microliths were narrowblade. Both excavation and testpitting used sieving to recover lithics, finding the smallest pieces.
There is some difference between the relative contents of individual collections. The lack of debitage in Grieve perhaps means that flakes are over-represented; the proportion of irregular debitage increases with time and may reflect the growing importance archaeologists give to knapping techniques. It is also possible that recognisable blades and flakes were removed by earlier projects, leaving the bulk of the debitage for later. It’s difficult to know whether the lack of material is due to unconscious bias or actual absence, and it should be remembered that Grieve was targeting “flint-rich” parts of the sites, so that other factors may have affected his collection. Research elsewhere indicates that different artefacts within ploughsoil tend to move differently over time, smaller pieces falling downwards and larger migrating towards the surface. Curiously, given the variation in other types, the proportion of blades is very consistent through collecting time. Grieve recovered a higher percentage of cores; not surprisingly, the excavated portion yielded a disproportionate amount of microliths and other retouched pieces.
Given these gross variations, it is interesting to see how understanding of the site has changed with each episode of work. Curiously, despite the gross differences between individual collections, general understanding of the area changes little through time, as each new episode of fieldwalking adds information to the previous view. Mesolithic activity was visible from the start. However, whereas Grieve’s work suggested that the principal focus of activity dated to the earlier (broad-blade) Mesolithic, excavation showed there to have been a greater later (narrow-blade) component. This
was balanced by ofars who also recovered broad-blade microliths, suggesting that the excavation area had its particular history.
Both Grieve and ofars have Neolithic or Bronze Age points, while excavation produced a wider range of later material. Again, artefacts relating to the Late Upper Palaeolithic were recovered from the start, and later confirmed among the excavated material and the ofars assemblage. However, these pieces would probably not have been identified as such by earlier analysis, as at the time a Scottish Palaeolithic had not been recognised. Recovery of a possible Hamburgian shouldered point from one of the test pits in 2019 provided nice confirmation of the significance of the location as a Late Upper Palaeolithic site (see feature Nov/Dec 2017/157).
Lessons for archaeology
Analysis of the field collections at Nethermills Farm over the past 50 years provides some important lessons for the way in which we research and interpret lithic scatter sites.
1 Repeated visits are necessary for a fieldwalked assemblage to be representative of a site. Even after over 3,000 pieces had been removed, it was still possible to collect several thousand more from the surface in subsequent decades. The recovery of over 30,000 lithics from excavation mostly below the ploughsoil in one small part of the site suggests why material continues to appear, at least in places: lithics may still be seen on the surface at Nethermills Farm.
2 Both Grieve and ofars recorded the importance of walking after rainfall in order to maximise the visibility of flint, but other logistical details such as the spacing of walked transects seem to matter less. Every project has its own factors, including the experience and number of walkers, time available and possible social benefits. While assemblage size will vary, interpretation remains remarkably uniform.
3 The role of community archaeology groups in fieldwalking is highlighted. The importance of repeated visits to a site over years makes fieldwalking less viable for professional teams such as commercial units with restricted finance and time. Community groups are on the spot, understand the locality, and are
enthusiastic and experienced. Nevertheless, the significance of specialist advice, in particular for lithic analysis, is also emphasised in order to ensure prompt reporting and publication of results. The weekly production of lithic data allowed Mesolithic Deeside to plan an up-todate fieldwork strategy and maintain team enthusiasm.
4 While basic archaeological interpretation of the site was available from the outset, subsequent fieldwork added significant detail. Later collections impacted on chronological interpretation, broadening the understanding of Mesolithic activity to focus on the later Mesolithic while reinforcing the presence of both earlier material (Late Upper Palaeolithic) and later (Neolithic and Bronze Age). Similarly, the importance of soft hammer percussion became more apparent.
5 Other subtle shifts perhaps reflect unconscious biases of the time or individual collectors. Irregular debitage is commoner than first recognised, and retouched pieces are more numerous and more varied. Flakes fall in relative significance, while more diversity in core type is seen. None of these affects the broad understanding of the site, but all are significant to detailed interpretation.
Old collections matter. The Grieve Collection includes material from eight sites besides Nethermills Farm, together providing information relating to prehistoric activity along the River Dee. This is supplemented by data from more recent fieldwalking projects and excavations along the length of the Dee, all of which together form an important record of the human past, adding considerably to archaeological understanding. Many similar assemblages housed in museums around the uk offer a, largely untested, archaeological resource and merit careful curation.
7 Though fieldwalking at Nethermills Farm has ceased, well-publicised interest in the site awakened memories of other material such as the Daly Collection.
Lithic scatters are an important archaeological resource. In many countries their contribution to archaeology is acknowledged, but this is not always so in the uk. The inclusion of lithic scatter sites within a protective scheduling programme is not always easy. Yet they play a valuable part in constructing the narrative of prehistory. Given the fragility of the archaeological resource and limitations of our assets, we should be taking them more seriously.
Torben Bjarke Ballin, Alison Cameron, Ann Clarke, Diane Collinson, Sheila Duthie, Gordon Noble, Irvine Ross, Heather Sabnis and Caroline WickhamJones contributed to this project as well as many volunteers. Funding has been gratefully received from the Society of Antiquaries of London, the Society of Antiquaries of Scotland, the Marc Fitch Fund, Aberdeenshire Council, Historic Environment Scotland and the National Lottery Heritage Fund. Aberdeen City Museum facilitated the loan of the Grieve Collection. See “Archaeological excavations at Nethermills Farm, Deeside, 1978–81” by C Wickham-Jones et al, Proceedings of the Society of Antiquaries of Scotland 146 (2016), doi.org/10.9750/ psas. 146.198535. Caroline Wickham-Jones is a consultant archaeologist specialising in the archaeology of Scotland’s earliest inhabitants