5.4 Lithic Identification and Analysis

While the actual attributes to be recorded in any one analysis depend on both the assemblage and on the questions to be asked, lithic analysis typically includes measuring a number of fields including the length, width and thickness of an artefact as well as sometimes recording its weight (metrical analysis, although the value of recording and analysing weight is debatable), in order to produce a series of statistics that describe any given piece (see the 422kb .pdf ScARF download: Lithic identification and analysis). On unbroken pieces the length-width ratio, for example, gives an indication of how long a flake is in relation to its width. Flakes with a length-width ratio of 2:1 or greater and where there are other indications of the use of blade technology, such as the presence of blade cores, are usually assumed to be blades and if it is less than 2:1 then they are categorized as flakes; however, these proportions should be used only as a guide as blades can be more squat with a lower length-width ratio while flakes can be blade-like. If an assemblage is dominated by flakes with a length-width ratio of 2:1 or more it would be classified as belonging to a blade-based industry. However, blades do not necessarily have to be twice as long as they are wide as it is the intentional flaking of a linear detachment, often with parallel sides, that creates a blade form.

The classification of flints can be hampered by the widespread plough damage that is common on lithic artefacts recovered from arable fields. Identification can be obscured by parts of the piece being missing or by edge damage appearing similar to intentional retouch, and statistical characterisation hampered by high incidence of breakage (whether ancient or modern) so it is important to be able to distinguish between breaks resulting from modern plough damage and intentional chipping (see Mallouf 1982). Useful explanations of how to get started in measuring, describing, analysing and classifying lithic artefacts can be found in Andrefsky (1998), Saville (1980), and also Watson (1968).

The illustration of flint artefacts requires a thorough understanding of how flints have been struck and how such characteristics are then portrayed using appropriate conventions. The key manuals that help unlock the door to lithic illustration are those by Martingell and Saville (1988), Addington (1986), and the more general work by Griffiths et al. (1990). A number of universities provide classes within Continuing Education Departments in archaeological illustration, while membership of the Institute for Archaeologists (Archaeological Illustrators Section) brings the benefits of seminars, conferences, publications, and the opportunity to seek advice from professional illustrators.

Presenting lithic scatter data from fieldwalking and test pit projects can usefully be made more compatible so projects can be compared.

See (Table 1 in the 422kb .pdf ScARF download: Lithic identification and analysis) for an example of summarising information from fields walked as part of one project as a single chart and used to produce lithic density per hectare counts.

Strategies for Lithic Artefact Analysis and Classification

The excavation of Mesolithic sites in Scotland often yields substantial lithic assemblages, potentially numbering in excess of 2000 pieces per metre square (including micro-debitage), and yielding over 250,000 pieces as excavated assemblages (Mithen 2000, 302). Even smaller scale sites or investigations can produce total assemblages in the region of 1500 pieces (Wickham-Jones and Dalland 1998). The recovery of diagnostic microliths, cores, or debitage pieces from deposits of more recent date also requires due consideration and specialist evaluation. The presence of such large quantities of stone-working debris has consequences for excavation and field-recovery methodologies especially the impact of topsoil striping and the use of wet sieving (see Wickham-Jones 1990 1990, 103 for discussion of the effect of recovery method on assemblage profile, and section 5.1.3 above). The large volume of material frequently generated from Mesolithic excavations also represents a substantial post-excavation commitment in terms of both time and resources and raises longer term storage, curation, and access issues.

The excavations at Kinloch, Rum introduced the first rapid, basic computerised classification system for large lithic assemblages in Scotland (Wickham-Jones 1990 1990). Prior to this, individual researchers often used their own idiosyncratic typologies (e.g. Mercer 1971) or applied imported schema (Cormack 1970). The Rum system provided basic categorisation by raw material (e.g. flint, quartz), blank type (e.g. flake, core, chunk), presence of cortex (using a tripartite division), regularity (at least 10mm of acute edge), presence or absence of retouch and/or edge damage, metric dimensions, and condition (e.g. fresh, abraded, corticated) for all pieces greater than 10mm in size. While this schema has since undergone development. Lithic analysts working on Scottish assemblages often base their categorisation upon it, or on that subsequently elaborated from it by the SHMP (Southern Hebrides Mesolithic Project; Finlayson et al. 1996; 2000). The utility of the routine recording of some of these categories has been questioned (Ballin 2000; Saville 2002; Saville et al. 2007; Saville and Ballin 2009) and there is still divergence in approaches to small size debitage (<10mm max. size) and the treatment of cores and retouched pieces (e.g. MacGregor and Donnelly 2001; Ballin and Johnson 2005). In the main, however, a broad consensus in approaches to the treatment and classification of Mesolithic lithic assemblages is emerging between many analysts. This offers considerable potential for cross-comparison and to elucidate genuine temporal and regional differences in stone-working traditions.

The use of a rapid basic classification method in combination with sub-sampling to provide a detailed technological and typological profile of the assemblage has proven particularly effective when dealing with large 'palimpsest', potentially multi-period, assemblages. A more biographical approach to cores in conjunction with detailed debitage studies has also proved informative in defining the dynamic process of lithic reduction. This has enabled some parameters of lithic skill to be defined and led to the identification of novice knappers - most likely children - in at least one assemblage (Finlay 2008; Mithen and Finlay 2000a). While there has been a traditional bias towards retouched pieces such as microliths, it is the mainstay of the unmodified component of the assemblage that often yields detailed understandings of stone-working techniques and can be used to address issues of tradition, mobility, materiality, and identity.

The adoption of detailed attribute based schema for microliths and other retouched pieces initiated by the SHMP (Finlayson et al. 2000) as well as schemes to classify microlith fragments (Finlay 2009) seeks to address the problem of idiosyncratic and unsystematic classification schemes and to foster greater consistency in classification, particularly between different analysts. These schemes still use basic microlith type categories (e.g. scalene triangle) but attempt to explore the nature of variation. Such approaches have proved particularly fruitful in combination with usewear studies (see below), for example at some of the sites on Islay usewear was found to be more likely to occur on those microliths with more angular forms (Finlayson and Mithen 1997, 2000). The use of an integrated chaine-opèratoire approach has also elucidated different routines of production and subtle manufacturing differences in microlith creation (Finlay 2000; 2003).

The recognition of Palaeolithic and Mesolithic material requires a broad and detailed knowledge of British and European assemblages and highlights the importance of the specialist evaluation of excavated assemblages, including older collections as well as a detailed understanding of local raw material to resolve issues of provenance. The identification of the Upper Palaeolithic artefacts in the Late Hamburgian (Havelte) assemblage at Howburn, South Lanarkshire (Ballin et al. 2010) and the Federmessergruppen assemblage at Kilmelfort Cave, Argyll (Saville and Ballin 2009) are based solely on typological association, in the absence of in situ deposits and any organic material for radiocarbon dating. Collaboration with colleagues in mainland Europe and especially those in northern Europe is helping to fully appreciate the nuances of shared and disparate traditions of stone-working during this period and enhance approaches to, and understandings of, local variation in stone tool use as well as raw material constraints.

The diversity of lithic raw materials used in prehistoric Scotland creates challenges for the analyst and studies have augmented understanding of the Mesolithic use of beach pebble flint (Mithen 2000), quartz (Ballin 2009) chert (Ballin and Johnson 2005; Wright nd), bloodstone (Wickham-Jones 1990) and baked mudstone and chalcedonic silica (Saville et al. in press). An important aspect is the local availability of resources which can be highly variable (Wickham-Jones and Collins 1978; Wickham-Jones 1986) but there may be other selection criteria involved, for example colour, texture, and aesthetic preferences to consider.  The variability of conchoidal fracture properties in some of these raw materials also necessitates greater flexibility in terms of analysis and impacts on recovery and site comparison. Condition is also a factor, for heavy surface patination (cortication) and burning often restricts identification even to raw material type, although condition itself can reveal much about artefact post-depositional histories, site formation processes, and identify the reworking of earlier artefacts. More analytical and experimental research needs to be undertaken to fully understand fully the particular properties of the suite of raw materials exploited in Scotland.

Integral to understanding Mesolithic lithic technology in Scotland has been the use of experimental replication (e.g. Mithen et al. 2000; Finlay 2003; 2008). This has proved helpful in understanding raw material constraints and technological signatures which have mainly been identified by using technological attribute analysis rather than refitting. The presence of so many large palimpsest sites has often precluded refitting as a routine strategy. Refitting has been more successful in relation to particular raw materials (Mithen and Finlay 2000b) and at smaller more discrete sites, such as Kilellan Farm, Islay (Saville 2005), although studies that link pieces between different sites in the landscape (e.g. Conneller 2005; Schaller-Åhrberg 1990) have yet to be attempted in Scotland.

A suite of microwear studies has been undertaken on Scottish microliths and other artefact and debitage classes using both low-power (Bradley 1985; Finlayson 1990; Finlayson and Mithen 1997, Finlayson and Mithen 2000) and high-power microscopy techniques (e.g. Hardy 2004). These studies reveal the complexity of microlith function and the flexibility of composite implements and the role of unmodified pieces and as such offer strong caveats against uniform interpretations of artefact function.

Coarse stone and pebble tools, such as the distinctive so-called limpet hammers and limpet scoops were among the earliest forms identified from shell midden sites (Anderson 1898). There is considerable variety in the types and forms of coarse stone implements (hammerstones, anvils, pebble tools, etc.) recovered from sites and these have suffered from lack of field recognition and unsystematic analysis (Clarke 1990; 2009; section 4.1.1 above). A number of waisted pebbles found as surface finds in the Tweed Valley and interpreted as net sinkers have also traditionally been attributed to the period (Saville 2004c; Warren 2005). The question of whether pebble stone axeheads are an element of pre-Neolithic assemblages in Scotland is still debated (Saville 1994c; 2009). Other distinctive forms of coarse stone implements such as hollowed stone palettes have been recognised at a couple of sites (Finlay et al. 2003) and limited residue analysis has also been undertaken with mixed results (Finlay and Whitehead 2000). No decorated or incised pieces are as yet known from Scotland although these have been found elsewhere in the British Isles, for example at Rhuddlan, north Wales (Berridge with Roberts 1994) or the incised flint pebbles at Hengistbury Head, Dorset (Barton 1992).

Though the study of Mesolithic coarse stone tools has not flourished in the same way as that of flaked lithic assemblages it holds considerable potential, highlighted by the work of Clarke (2009). There has been no overall consistent methodological approach, nor has there been consensus regarding terminology, or approach in the recording of wear traces. The consequence is that stone tool catalogues which have been produced by various researchers over decades cannot be easily interrogated for intra-site comparisons.

Experimental approaches to pebble tools and associated bone and antler forms have also formed an important part of archaeological investigations, from the experimental use of concrete limpet hammers to knock limpets off rocks (Clark 1956) to more recent studies involving hide working (Griffitths and Bonsall 2001; Barlow and Mithen 2000; Birch 2009).

Other categories of stone finds include fire-cracked rocks and pieces of pumice. While not deliberately modified, fire-cracked rocks have tended to be overlooked and more systematic recovery and analysis of these pieces would be likely to yield enhanced understandings of Mesolithic pyrotechnologies. Pieces of modified and unmodified pumice are also infrequent finds (mostly on coastal west coast sites). Several of those that have been analysed are dated to particular volcanic eruptions which provide a useful dating control (Newton 1999).