The excavation and surface surveying of the quarry area and the three hill settlements have provided abundant information on the axe production process and its organisation. The test trenches placed in the upper and lower parts of the dolerite outcrop show that the quarrying took advantage of the intense fracturation of the deposit by systematic joints. This enabled the extraction of small blocks or slabs with relatively little effort, and clearly represents an important economic advantage in comparison to the other dykes. (Foote (1916, 82), a geologist who travelled widely through southern India, mentions that the occurrence of such weathering dolerite outcrops was a very unusual phenomenon for the trap-dykes that occur in such great numbers on the Deccan plateau.) Percussion stones of dolerite, gabbro and granite, as well as chisels of antelope horn, found within the quarry debris, were used to break up the deposit. A spherical hammerstone, weighing 1767g, made out of gabbro from the northern dyke, is the heaviest artefact of this type recorded so far at Sanganakallu. The main operation involved obtaining 150-300mm long blocks, slabs or nodules of angular, sub-angular or sub-rounded shape. Their outer surface has a yellow-orange colour, resulting from oxidation of the rock minerals by weathering processes in the systematic joints. This surface alteration has proved to be a helpful trait in reconstructing how the natural slabs were worked. Traces of this 'skin' remains, to a larger or smaller extent, on c. 50% of the unfinished products as well as on some used axes. In 8% of the blanks, surface alteration appears on the dorsal as well as ventral sides of the blank, indicating the thinness of some of the selected slabs.
In general, three reduction strategies have been identified in the Sanganakallu lithic assemblage. The methods are defined mostly on the basis of the initial dolerite stone packages utilised (Brumm et al. 2007). These include:
In all cases, the dominant procedure was to work the blocks, slabs or flakes bifacially around the perimeter of the stone piece (Fig. 4). Yet, considerable variation in the knapping procedures can be observed within all three strategies, suggesting that persons with very different skills and habits were operating simultaneously in the area (Brumm et al. 2007).
Figure 4: Stages in axe production: 1. natural dolerite block, 2. initial flaking of lateral edges of the block, 3. flaked blank, 4. blank with slight pecking traces, 5. blank with more intensive pecking, 6. polished axe (note different scales; photography by José Antonio Soldevilla)
[View static image]
The main target that needs to be achieved in the knapping process is the reduction in thickness of the rough-outs, as thinner axe blanks will require less pecking and grinding. Unskilful knapping is visible in the form of thick blanks with no further modification. These resulted in discard in the quarry and settlement areas. On occasion these thick blanks were transformed into percussion tools. It is much easier and quicker to dress thin slabs or flakes (methods 2 and 3) than to work larger blocks and slabs, which requires a higher level of technical control (method 1) (Brumm et al. 2007). In this case, the rocks were reduced on one face through lateral flaking, while knapping on the opposite face was much less invasive, thus taking advantage of at least one of the flat surfaces of the slabs or blocks. Again, the underlying principle seems to be to spare effort and time during the knapping, as well as during the later pecking and grinding stages.
One possible way of determining the technical competence or variation of knapping is to consider the relation between the length and the weight of the blanks and finished axes. High technical control results in thin artefacts, which require shorter polishing processes and/or provide sharper edges. Depending on the type of natural rocks used, it also enables the maximisation of raw materials. Low flaking skills produce thick and, consequently, heavier artefacts. Therefore, the relation between length and weight provides a useful analytical test to evaluate the degree of skill manifest in the flaking process. (This agrees with observations made by Stout (2002) among the Langda adze makers of Irian Jaya. The more skilled knappers produce longer but not wider or thicker blanks. Consequently, high technical competence results in proportionally lighter artefacts.) If we correlate these two variables measured on a sample of artefacts from the Sanganakallu axe collection, it can be observed that variation in production is high (Fig. 5). On average, unskilled examples are 2.8 times heavier than the well-flaked blanks of the same size. For example, the weight of a typical 100mm long axe varies between 100 and 240g. As expected, most of the finished and used axes fall among the lighter examples. It seems reasonable to assume that many of the artefacts found during survey and excavation were in fact discarded because of the low-quality knapping and the workload that finishing them would have created. The technical organisation of production was not very rigid, and persons with different levels of skill seem to have been at work simultaneously.
Figure 5: Relation between length and weight of axe blanks, finished axes and finished axes with signs of use-wear from Hiregudda (Areas A and J) and Sannarachammagudda (Trench 10). Lines mark the relation between length and weight and are therefore indicative of higher (1/1) or lower (1/4) technical competence
About 25-30% of a test sample of finished axes and blanks from Hiregudda and Sannarachammagudda showed traces of pecking, visible as small pits (Fig. 4). This process seems to have been carried out with the same type of dolerite, gabbro, quartz and granite percussion stones as flaking, and mainly served to reduce the highest points and crests of the preforms. The pecking traces rarely cover more than 20% of the surfaces. It appears that this working stage was only applied occasionally, in order to reduce irregularities or blank thickness resulting from knapping errors.
The final stage of stone axe production consists of grinding and, eventually, polishing. As in the case of pecking, this labour-intensive process rarely affects the whole surface of the artefact, leaving traces of earlier work stages and even parts of the surface skin of the raw material (see also Allchin 1957). In many cases only the cutting edge is polished. Again, the technical organisation of the production process seems to have been guided by utilitarian and quantitative criteria, rather than by the finishing and aesthetic aspect of the products. As indicated below, it is possible that a considerable portion of the blanks or preforms were not ground in the Sanganakallu area, but in settlements further away. It is likely that only some of the artefacts were polished locally (see also Boivin et al. 2005, 77).
One of the longest known archaeological features of the three hill-settlements are the grinding or polishing grooves and hollows found in the granitic bedrock or across much of the surface of large boulders (Subbarao 1948; Boivin et al. 2007). Use-wear analysis, as developed for the study of macro-lithic artefacts (Adams 2002; Risch 2002), has been carried out on the working surfaces of these features, in order to confirm their participation in the axe-grinding process. This is the first such analysis of the features, and indeed of any grinding hollows in India. The mesoscopic observation (5-20X) of the working surfaces has shown that not all of these features served the same purpose and that many were used for different tasks. Mainly the smaller and oval hollows or 'slicks' present evidence of intensive abrasion and use-wear traces related to the grinding of stone (Fig. 6). The larger rounded hollows, in contrast, show traces related preferentially to grain processing. Experimental tests confirmed that the deeper oval features allow a nearly mechanical grinding of the blank's edges by placing the whole artefact in the hollow, pressing it with the flat hand and carrying out a rocking movement with it. Such an operation is much more comfortable than holding the blank in the hand and grinding only the frontal cutting edge against the rock surface. In fact, many artefacts show polishing traces around the periphery, especially on the bevelled edge and the opposite butt, while the flat surfaces of the dorsal and ventral faces remain largely unchanged because of a lack of contact with the convex grinding hollows.
Figure 6: Polishing hollows from Choudammagudda at an initial, middle and final stage of development (Photograph by J.A. Soldevilla)
A different form of grinding must have been carried out in the V-shaped polishing grooves, which are much less frequent in the settlements. Such features are known from ethnographic and experimental examples and allow the polishing of lateral edges (e.g. Kennedy 1962; Delage 2004, 40). In addition to the highly levelled grain surfaces, the intensity of the activity carried out in these grooves and the hardness of the worked material is also revealed by scratches and deep striations on the groove surfaces.
© Internet Archaeology/Author(s) URL: http://intarch.ac.uk/journal/issue26/26/3.html
Last updated: Wed Jul 29 2009