Case Study 9: Lithic Refitting: A Case Study from Star Carr, North Yorkshire
Chantal Conneller (Manchester University)
Refitting is a technique that was employed by the pioneers of prehistoric research (Spurrell 1880; Smith 1894) and has become a common tool in understanding lithic material (Cziesla et al. 1990; Hofman and Enloe 1992). Though many classes of artefacts can be refitted, the technique is more frequently employed on lithics, because as a ‘subtractive technology’ the entire manufacturing process can be reconstructed and its spatial dimensions understood.
The uses that refitting can be put to are numerous: it is an important tool for understanding site taphonomy, for unpicking palimpsests, for understanding techniques and both the social and spatial aspects of lithic technology and its organisation in the landscape. Above all, it can provide a temporal narrative of the occupation of a site, particularly useful on sites where other artefact classes are lacking.
Refitting is a time-consuming technique that demands a certain amount of space to lay out material. It works best if it is undertaken in short bursts of around an hour or so a day and is much quicker if undertaken by someone with a good knowledge of lithic technology. While it can be employed on any excavated lithic scatter, certain variables make refitting more labour-intensive on some sites.
Refitting takes relatively little time on a small, single-occupation Upper Palaeolithic site, particularly one with heterogeneous raw material. Large sites, diminutive raw material, taphonomic issues and homogeneous or patinated raw materials all make refitting more difficult. On such sites it is worth thinking carefully about how refitting should be employed; what questions might be answered given the nature of the site. On very large sites, for instance on the scale of Stainton West (see Case Study 6), refitting on a large scale is clearly unfeasible, but targeting rare raw material types might be useful in understanding how particular materials were transported to site (eg as blades or tools, or unworked nodules) and how products were moved around site.
Refitting also aids considerably in pattern recognition, so even a small-scale refitting project focusing on one particular area of a complex site will aid the analyst in understanding the parameters of the technology and patterns of transportation of lithics across a site. Star Carr, the main case study of this contribution, is a large, complex, repeatedly reoccupied site and here targeted refitting was used to understand the complex palimpsest on the dryland part of the site and discrete activity episodes on the lake edge.
Refitting is a vital taphonomic tool. Very few sites that predate the Loch Lomond Stadial have survived unaffected by landscape processes and refitting can help assess the evidential value of a particular site. It can provide an understanding of the extent of vertical movement through the soil profile, but also of more complex issues of site taphonomy, such as the effects of colluviation, winnowing or truncation.
At Hengistbury Head, two distinct artefact horizons were present, leading Campbell (1977) to suggest two occupations – a lower horizon dating to the Upper Palaeolithic and an upper, Mesolithic scatter, produced by size-sorting. However, refits between the two layers indicated a single Late Upper Palaeolithic assemblage (Collcutt et al. 1990). More sophisticated use of refitting to understand site taphonomy depends on its success – or lack of it.
Work on two Final Palaeolithic scatters at La Sagesse Convent (Conneller et al. 2007) on the alluvial plain of the river Test, revealed that one was intact (indicated by coherent refit sequences) while the other had been truncated by a flooding event (indicated by poor refitting success and gaps in the sequence that did not make sense in terms of human activity).
Early prehistoric excavations often reveal a number of discrete or overlapping scatters with uncertain relationships. Refitting has often been used as a way of determining contemporaneity between them (eg Cahen et al. 1979), but caution and judgement do need to be used.
De Bie and Caspar (2000) found a number of refits between the different scatters of the Belgian site of Rekem, but they were relatively few and in one direction only (eg towards a particular scatter), suggesting scavenging of flint from an earlier occupation. In this case, a judgement needs to be made about the number of refits, their direction and whether they might make sense in terms of different activity areas from a single occupation.
Technology and skill
Refitting has long been used to reconstruct the knapping sequence and thus offers a window into prehistoric technologies and decision-making. It can also provide a nuanced understanding of the technological process, and the balance between learnt ways of proceeding and individual variation in knapping style. It is an important way of assessing individual skill, the extent to which an individual is able to anticipate and correct knapping problems.
Schlanger (1996) has used refitting to argue that early Neanderthals at Maastricht-Belvedere were able to respond flexibly to both problems and potentials in the raw material, anticipating the problems that poor raw material might bring and ably exploiting larger nodules.
Assessments of skill have also been used to elucidate both the presence of specialists and unskilled children. At the Paris Basin Magdalenian site of Etiolles, Pigeot (1991) used refitting to discern a specialist knapper in close proximity to a hearth; this specialist produced large blades that were distributed across the site. The specialist was surrounded by part-skilled apprentices and beyond these by children, learning to knap flint by imitating the gestures of the specialist.
Refitting provides important information on movement of lithic material – both around the site, where tools produced in one area may be used in other areas, and in the broader landscape, where refitting permits an understanding of the form in which raw material entered a site (as nodules, cores, blanks or tools) and what was removed for future use elsewhere.
Refitting (and a lack of it) has revealed different raw material transportation strategies in the Vale of Pickering from the Final Palaeolithic to the Mesolithic. Final Palaeolithic lithic scatters at Seamer K demonstrate the presence of finished tools in speckled grey flint, contrasting with knapping sequences in white Wolds flint. This suggests that raw material was carried mainly in the form of finished tools and retooling occurred at the site, using local flint.
By contrast, refitting long blade sequences show that people were bringing prepared exotic nodules of flint to Seamer C, along with a stock of blades that do not refit to knapping sequences (see Fig 9.1).
Finally, in the Early Mesolithic, people brought complete, tested, or partially knapped beach pebbles to a number of different sites in the area (Conneller 2007).
Results: lithic refitting at Star Carr
At first glance, Star Carr does not seem a suitable site for a programme of refitting. It is a large site with nearly 25,000 lithic artefacts recovered; it was also repeatedly occupied, with Bayesian modelling suggesting repeated visits over a period of around 800 years (Milner et al. 2018). As a result, refitting was targeted in certain areas, either where it was anticipated that refitting would be easier, or where refitting would be able to address particular problems of site interpretation.
Star Carr is made up of three different zones: first, the area that during occupation would have been open lake water; secondly, the peaty lake-edge area that over time developed into an area of fen carr; and finally the densely and repeatedly occupied area of the dryland.
Each of these zones has a very distinctive type of archaeology. The area of the lake water was a zone of deposition where knapping did not occur. The assemblage from this area consists of blades and tools from a vast number of sources. Refitting was not attempted in this area though distinctive raw material types were checked in case they could be integrated into dryland sequences and thus inform on the date of a particular dryland lithic scatter (since these could not be fitted into the Bayesian model which was focused on the wetland).
The second zone of the site, the lake edge and fen carr, is characterised by the build-up of peat over several centuries. Here people came on occasion to carry out particular tasks. These short episodes are characterised by small, discrete, high-resolution scatters and were thus prime candidates for successful refitting. A scatter of flint around a hearth was refitted and found to represent the reworking and resharpening of axes.
As well as providing an understanding of the technology of this process (longitudinal removals along one face, followed by a tranchet blow on the other), the refitting showed that several axes were resharpened, some were discarded within the scatter, while others were removed for future use (see Fig 9.2).
Refitting was undertaken more extensively across the dryland where repeated occupation and poor faunal preservation rendered the area less readily readable. The dryland was characterised by a series of structures and middens (often associated with large quantities of lithic material), knapping scatters, and activity areas.
One large midden in the west of the site consisted of a high frequency of burnt material which would make refitting very difficult so none was attempted here, but the remainder of the lithics from the dryland were laid out in trays by area.
While there was a high level of success in a couple of areas (see Fig 9.3), in general even fairly discrete-looking scatters did not refit well. This seems to be because material was being scavenged during later occupations of the site and also because some scatters represent midden materials cleared from elsewhere; it is likely that the scatters that refitted well were relatively late in the sequence of occupation.
Refitting particularly focused on one of the structures (the eastern structure) to understand how it was used: whether it had been cleared out or different occupations could be discerned. Refits occurred over some vertical distance, suggesting movement through the profile or disturbance by people when the house was cleared out. Long refit sequences were not present, indicating that extensive knapping of cores was not undertaken within the structure; this was a task for workshops outside. Instead, the people would remove a couple of blades or flakes from a core for immediate use.
Also present in the structure were personal tools brought in for repair – a small axe with a refitting resharpening sequence, for example (Fig 9.4). Material also appeared to have been cleared out of the structure to midden areas to the north and north-east, as indicated by refits (Fig 9.5).
Refitting at Star Carr enhanced the understanding of complex areas of the site, particularly the use of structures, and provided a detailed picture of discrete activities in the peaty lake-edge zone. While refitting is an open-ended task through which more information will always be acquired (albeit with gradually diminishing returns), targeting particular areas in an effort to answer particular questions, or focusing on areas where the flint is likely to be more readily refitted, is a useful means of extracting maximum information using this technique without huge time investments.
Cahen, D, Keely, L H and Van Noten, F, 1980 ‘Stone tools, toolkits and human behaviour in prehistory’. Current Anthropology 20, 661–83
Campbell, J B, 1977 'The Upper Palaeolithic of Britain'. Oxford, Clarendon Press
Collcutt, S N, Barton, R N E and Bergman, C A, 1990 ‘Refitting in context: a taphonomic study from a Late Upper Palaeolithic site in sands on Hengistbury Head, Dorset, Great Britain’, in Czeisla, E Eickhoff, S, Arts, N and Winter, D (eds) 'The Big Puzzle: international symposium on refitting stone artefacts' (Studies in Modern Archaeology I). Berlin, Holos Verlag, 219–36
Conneller, C, 2007 ‘Inhabiting new landscapes: settlement and mobility in Britain after the last glacial maximum’. Oxford Journal of Archaeology 26(3), 215–37
Conneller, C, Ellis, C, Allen, M J, Macphail, R and Scaife, R, 2007 ‘A Final Upper Palaeolithic site at La Sagesse Convent, Romsey, Hampshire’. Proceedings of the Prehistoric Society 73, 191–227
Cziesla, E, Eickoff, S, Arts, N and Winter, D (eds) 1990 'The Big Puzzle: international symposium on refitting flint artefacts' (Studies in Modern Archaeology I). Berlin, Holos Verlag
De Bie, M and Caspar, J P, 2000 'Rekem: a ‘Federmesser’Camp on the Meuse River Bank'. Leuven and Asse‐Zellik, Leuven University Press and Instituut voor het Archeologisch Patrimonium
Hofman, J L and Enloe, J G, 1992 'Piecing Together the Past: applications of refitting studies in archaeology'. Oxford, BAR International Series 578
Milner, N, Conneller, C and Taylor, B, 2018 'Star Carr: Volume 1. A persistent place in a changing world'. York, White Rose Press
Pigeot, N, 1990 ‘Technical and social actors: flintknapping specialists and apprentices at Magdelenian Etoilles’. Archaeological Review from Cambridge 9(1), 126–41
Schlanger, N, 1996 ‘Understanding Levallois: lithic technology and cognitive archaeology’. Cambridge Archaeological Journal 6(2), 231–54
Smith, W G, 1894 'Man the Primeval Savage'. London, Edward Stanford
Spurrell, F C J, 1880 ‘On the discovery of the place where Palaeolithic implements were made at Crayford’. Quarterly Journal of the Geological Society 36, 544–9