Determination of flint implement function. Wear analysis of scrapers from Sarup

Abstract

Determination of Flint lmplement Function. Microwear Analysis of Scrapers from Sarup

Based on Keeley's studies of microwear on flint implements (9) Helle Juel Jensen, Peter Rasmussen and the author have carried out a number of experiments at Moesgaard with a view to analysing microwear in prehistoric flint implements from Denmark.

The experiments were carried out by working with various types of flint tools on a number of materials which should represent most of those present in Denmark in prehistoric times: wood in the form of oak and Scotch pine as representing hard- and softwoods respectively, red-deer antler, fresh tubular bones from cattle and pigs, raw meat in the form of pork, hide from red deer and wild pig, plaice and cod fish and plant materiale comprising fresh and dry grass, ripe corn (wheat) and withered reeds.

After use, the test implements were cleaned in sodium hydroxide, or hydrochloric acid (implements used on bone or antler), and then examined under the microscope at 250X magnification. The microwear traces observed correspond closely to Keeley's descriptions (9), and we were able to separate traces of wear deriving from work in the following materials: wood, bone/antler, hide, meat and plant material.

Using the test implements, we caried out a series of blind tests to gain experience of microwear analysis and to ascertain how much confidence might be attached to it in prehistoric flint. These blind tests were performed in the course of about two months and at the end of this period we could identify 80-90 % of a series of 10-12 tools correctly.

On the basis of our experiments, I have applied microwear analysis to Middle Neolithic flint scrapers from the Sarup site in south-west Funen (12). The investigation covers 161 scrapers, recovered from a layer dated to late MN II (C-14 date 2390±90 BC) (14), and I succeeded in identifying the wear marks in 118 pieces (73 % of the material).

In most cases, the wear marks were, as expected, found in connection with the retouched part of the tool, but in 5 implements this was not so. These were all more or less blade-like flakes with a retouched distal end, exhibiting wear along an unretouched side. These implements must on this account be considered knives, where the retouched end is merely a finger rest.

In 40 scrapers (26 % of the 156 classified as such on the basis of wear), striations were observed in connection with wear running parallel to the edge. An example of this is seen in fig. 1b (the arrow, as in the following examples, showing where the picture was taken). This unusual orientation of the striations, which is not in accordance with ordinary scraper movements, may naturally be fortuitous. But since it has been found on so many of the implements, it may manifest a deliberate function in which the working movements have run parallel to the edge. An explanation may possibly be found in ethnographical observations. In connection with a description of the use of stone tools among New Guinea highlanders, M. Strathern has remarked that tools fashioned from flakes often have different functions, and that the same tool may therefore be used both for scraping and for cutting (18). This information on a varied use of tools provides a possible explanation of the situations parallel with the edge found on Neolithic scrapers, the scrapers possibly constituting a rather unspecialized implement group sometimes used as knives or saws, for example.

The result of the wear analysis was that I found among scrapers evidence only of wood and hide working. Of the 5 implements which on account of the disposition of wear must be regarded as knives, 4 had worked in wood and 1 could not be identified. In connection with the description of wood and hide use-wear respectively, 2 examples of test implements with this type of wear and 2 examples of prehistoric implements with corresponding use­wear are presented.

Wood polish. Keeley describes i.a. wood polishing as bright and smooth and gently curved over or "domed" on the high points of the microtopography, in time forming a reticular pattern if the flint surface is coarse (19). The characteristic wood polishing with this distribution is seen clearly at the arrows in fig. 2 (test flint knife which has cut fresh oak 1000 times). In fig. 3 (implements which have scraped dry oak 1000 times) the polishing is more strongly developed and thus covers a greater part of the surface, but here too, the characteristic domed polishing is seen.

Fig. 4a-b, artefact no. X565: One of the implements already mentioned which must be regarded as a knife due to the location of the wear along a sharp lateral edge (indicated with a bracket in fig. 4a). In fig. 4b, scattered examples of the domed wood polishing are seen along the edge. Since the implement is made of very pale flint, the wear marks are not particularly distinct in the photograph. Fig. 5a-b, artefact no X1372: The wood polishing is seen as pale areas along the edge in the centre of the picture. Due to their location at several microtopographic levels, it has been possible during photography to focus on only one point (arrow).

Hide polish. In his description of this type of wear, Keeley writes among other things: "Two microwear features link all the hide polishes: (1) relatively severe attrition of the working edge (that is, removal of flint material from the edge by means other than breakage) which results in a markedly rounded edge, and (2) diffuse, shallow linear features running in the direction of use" (20). The "relatively severe attrition of the working edge" is without doubt the most characteristic sign of hide working tools. In the test hide scrapers, this wear appeared very rapidly, and after about 30 minutes' use on dry hide, the rounding of the edge was in some cases so pronounced that it could be seen with the naked eye. It was also seen on blades used for cutting hide. In fig. 6, a section of the edge of a test hide scraper used on dried red deer hide for 30 minutes is seen. As in the following examples of hide scrapers, the picture was taken at 60X magnification. This rather low magnification in relation to the pictures of wood polish (240X) has been chosen because it is sufficient to reveal the characteristic strong rounding of the edge, but at the same time gives, with the greater depth of field obtained, a better impression of this type of wear. Besides the rounding of the edge, fig. 6 shows the "diffuse, shallow linear features, running in the direction of use", which Keeley mentions. To illustrate the strong wear on the edge of hide scrapers, an example is shown of a tool before and after use. In fig. 7a, a section of the edge of a scraper before use is shown, and in fig. 7b the same area after the tool has scraped dried red deer hide for 30 minutes. Fig. 8a-b artefact no. 1684: Double scraper in which both edges have been used for scraping hide. In fig. 8b, the rounded edge characteristic of hide scrapers is seen. Fig. 9a-b, artefact no. X906: Hide scraper with the same characteristic marks of wear as fig. 8b.

By counting implements with a double edge as 2, the number of implements with identified wear reaches 122, distributed as follows:

Wood . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 103

Hide scrapers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  17

Double function (hide & wood) . . . . . . . . . . . . . . . . . . . . . . 2

Wood scrapers obviously form the largest group. It is also important to note that in only 2 implements was wear from working in more than one material found. I think this indicates that the implements were in most cases made for a single purpose and discarded immediately after use. This interpretation is also supported by the fact that Neolithic settlements as a rule yield large numbers of scrapers.

The studied material contained 11 of the 18 subtypes into which N. H. Andersen has divided the Sarup material. No correlation between these sub-types and function (use­wear marks) could be demonstrated. While working on the use-wear analysis, I noticed that the hide scrapers generally had very thin edge in relation to the other scrapers. In order to investigate whether there is any connection between edge thickness and function, I have measured the edge thickness of all scrapers where all of the edge is preserved. This has been done in the following manner. The length of the edge (retouched portion) is measured followed by the thickness 1/3 and 2/3 of the way along the edge, 5 mm in. As edge thickness may vary somewhat in the individual scraper, I have elected to measure thickness at two places and use the mean of these two values to express it. The reason for measuring edge thickness and not edge angle is that I find it difficult, especially on thin edges, to measure the latter. Moreover, a thin edge, on account of steep trimming, may form a rather large angle unrelated to edge thickness. As edge thickness has been measured only on scrapers where the whole edge is preserved, the investigation comprises 143 implements, grouped as follows:

Wood scrapers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 90

Hide scrapers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  15

Double function (wood & hide) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2

Undetermined . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36

Edge thickness is measured in mm to the nearest 1/2 mm. Fig. 10a shows the distribution of edge thickness in all the scrapers, and fig. 10b and 10c covers wood and hide scrapers, respectively. It is apparent that hide scrapers form a group with a very small edge thickness, the mean edge thickness for hide scrapers being 4.7 mm, whereas that for wood scrapers is 9.4 mm. By means of the non-parametrical Mann-Whitney U test (23) it has been investigated whether the apparent difference in edge thickness between wood and hide scrapers is in fact fortuitous. The test revealed a highly significant correlation between edge thickness and function. It may therefore be concluded that the selection of scrapers for a particular purpose was apparently based on edge thickness, whereas other morphological features, as found in the sub-types, have been of no consequence. A corresponding tendency is seen in K. Knutsson's investigation of the scrapers from the Late Neolithic settlement Borneck-Ost, near Hamburg (24). Fig. 15 of the publication shows the relation between wear and edge angle. As is pointed out, there is a distinct correlation between implements identified as hide scrapers and low edge angles, whereas scrapers interpreted as having been used in hard materials generally have larger edge angles. The same result has been arrived at by E. N. Wilmsen in an investigation of paleo-Indian implements, in which he finds that hide scrapers (scrapers with rounded edges) generally have low edge angles in relation to scrapers which on the basis of chipped edges are interpreted as having been used on hard materials. The supposition that the selection of scrapers for particular purposes was based on quite simple criteria, largely edge thick­ness/angle, is also supported by ethnographical observations. Thus White and Thomas have pointed out that New Guinea highlanders do not distinguish between core and disc implements, and that the selection of stone tools for particular purposes takes only size and edge type (edge angle) into account (26).

Over the years, several more or less conflicting theories as to how wear marks are formed on flint implements have been advanced. This discussion has in particular concentrated on the conspicuous "corn gloss". Several archaeologists have held that this type of gloss results from a wearing down of the flint surface, whereas others believe that it derives from the accretion of silica (28, 29, 30, 31, 32, 33). With a view to a better evaluation of the various theories, we have carried out a small investigation of a Neolithic flint sickle with a highly developed gloss. This tool has been sawn acros and the resulting section photographed under the microscope as shown in fig. 11. The picture shows the edge, and it will be seen that no layering of the surface can be observed, whereas the edge is distinctly rounded, suggesting wearing down.

Jens Jeppesen

Oversættelse: Peter Crabb