The image processing techniques described above demonstrated that alterations
to an unused flint surface could be reliable detected. These techniques
have accordingly been used to investigate hafting traces on drill bits (see
Unger-Hamilton et al. 1987). In
this case the classification had the limited aim of deciding whether any
traces were significantly different from unused flint, and of detecting
their origins if possible.
The drill bits from the original experiments (made in conjunction with a
microwear analysis of drill bits from Abu Salibihk, Iraq) were re-examined
microscopically and two cases of some alteration of the flint surface were
observed on the hafted areas, but these traces covered such a small area
and were so indistinct that no conclusions as to their origin could be reached.
Therefore it was decided to carry out a further experimental programme with
the specific aim of detecting hafting traces.
The experimental program is illustrated in Figure 26.
The loaded resin haft consisted of resin mixed with ochre and sandy grit
in which the drill bit was held entirely by the mastic (Figure 27a). This
was achieved by heating the mastic, inserting the drill bit and allowing
the mastic to cool and harden. Wedge hafting was achieved by making a mortise
in the end of the drill shaft in which the bit was held in place by the
pressure exerted between the drill shaft and the contact material (Figure
27b). This technique is similar to that described by Keeley as jam or wedged
hafts, and was chosen because such hafts, "characteristically allow
movement of the tool in the haft" (Keeley
1982,799), and therefore are more likely to produce hafting traces.
A refinement of this technique was to insert small chocks of wood between
the drill bit and the shaft to reduce movement in the haft, and allow the
drill bit to be centered along the axis of the shaft more accurately (Figure
27c). The forcing of the slivers of wood into contact with the drill bit
would perhaps produce characteristic hafting traces. Hafting with sinew
was carried out by cutting a deep slot into the drill haft into which the
drill bit was inserted and held in place by wrapping the sinew around the
haft and the protruding edges of the drill bit (Figure 27d). The edges were
allowed to protrude so that they would be in contact with the sinew, therefore
any movement of the bit in the haft would produce friction with the sinew.
This was to test whether any traces would be produced that were attributable
to the sinew. Sinew polish from hafting has been reported by Buller
(1983).
The worked materials were selected as being relevant to the archaeological
specimens and the previous experimental program (see
Unger-Hamilton et al. 1987).
To attempt to isolate the origin of any traces, micrographs at 200 magnifications
were taken of the tools at each technological stage. Photographs were taken
of areas of the blanks as soon as they were removed from the core and again
after being retouched to form drill bits. The drill bits were then photographed
after being hafted and used in the bow drill. Outline drawings of the blanks
and the drill bits were made so that the position of each photograph could
be recorded. It was attempted to photograph the same area on the blanks
and the drill bits, both before and after use, to produce exactly comparable
photographs. This was not always possible as some blanks required extensive
retouch to reduce them to drill bits and in some cases removing the photographed
area of the blank. All areas showing any visible alteration of the flint
surface at 200 magnifications were photographed, after the drill bits were
retouched, and again after they had been hafted and used. A 100 photographs
were taken to obtain as complete a visual record as possible, so that any
wear traces observed could be assigned to the technological stage at which
they appeared (see Figure 28).
A number of these photographs were selected for digital scanning and
subsequent texture analysis, the results of which are illustrated in Figure
29. The values of the statistics are expressed as differences from unused
flint so that the nearer a point is to the origin the closer the similarity
of that texture sample is to unused flint as described above.
As can be seen from Figure 29 most of the samples cluster around the
origin indicating their similarity to unused flint, and thus exhibiting
little surface alteration. This group includes textures of blanks, retouched
drill bits and used drill bits from photographs of the same area of the
same tool. This illustrates that in most cases no surface alteration occurred
either from retouch or hafting. Included in the diagram (Figure 29) are
five examples of photographs taken of use wear produced at the tip of the
drill bits. The area marked around the origin of Figure 29 represents 15%
of the total range of variation from unused flint. Allowing for the natural
variation of flint textures, points within this area are not considered
significantly different from unused flint. This level of significance is
calculated from the results presented in Figure
20. Six points representing areas on the hafted end of the drill bits
remain outside this area.
Point a is from a photograph taken of the bulbar scar of a blank,
and the slight surface alteration may be due to the pressure exerted between
the core and the blank when it was struck. Points b, c and d
are from photographs taken of retouched drill bits before they were hafted.
In all three cases these areas were on dorsal ridges of the drill bits.
A possible explanation is that as the tools were pressure retouched on a
hide pad, small chips of flint were removed that congregated beneath the
tool, so that when pressure was applied during retouch the drill bits were
pressed down onto the flint chips causing flint on flint contact. The dorsal
ridges would provide a particularly vulnerable area for surface alteration
under these circumstances (see Plate
6 a) and it is interesting to note that traces on dorsal ridges have
been interpreted as evidence for hafting by Keeley
(1982 Fig.1 and 5, 805), and Frison
(1968,152).
The incidence of such retouch traces prior to hafting on the drill bits
could be considered more probable because these tools are retouched around
most of the periphery, but the retouching of the non-utilised end of hafted
tools of other types is more extensive than with unhafted tools. The shaping
of endscrapers to fit the haft, or the backing of bladelets for hafting
as knives or sickle blades are examples of this.
Points e and f are from used drill bits. Point f is
also from a dorsal ridge (see Plate
6 b) of the hafted end of
drill bit number 2; e is from the central part of the ventral surface
of drill bit number 9. Neither of these examples has a polish that is sufficiently
distinctive to identify any contact material that may have caused it, and
therefore cannot be matched to any particular type of haft. Number 2 was
wedge hafted and number 9 was wedge hafted with the addition of chocks,
so that if these surface alterations were due to hafting they should exhibit
characteristics that are attributable to contact with wood. Not only do
they not do so, but the two surface alterations are quite different.
From 9 drill bits hafted by various techniques, only two small and indistinctive
areas of surface alterations were observed that occurred after hafting.
A far higher incidence of positive hafting traces would be expected to appear
on 9 drill bits if hafting traces occur with any frequency, especially considering
the intensity of use and the amount of friction between the tool and haft
involved in using drill bits in a bow drill. The drilling action involves
considerable vertical pressure applied on the capstone as well as the vigorous
rotational movement of the drill. The rotational action of the drill takes
the form of alternative clockwise and counter-clockwise movement causing
lateral vibration as the inertia of the previous movement is overcome.
The fact that surface alteration occurred on the retouched drill bits before
hafting indicates that there may be other origins for traces that may have
been attributed to hafting. Wear traces from the manufacture of tools has
been reported before by Keeley (1980)
and Moss (1983). Antler traces from
the soft hammer rubbing over the surface of the tool during retouch have
been observed (Moss 1983 see Figure
6.8d, 103). But when the manufacturing traces are indistinct and not identifiable
to a known process, perhaps there has been a tendency to ascribe these traces
to hafting because they appear on the assumed hafted area of the tool. In
this study surface alteration occurred on dorsal ridges of tools during
retouch, although those areas are well away from the edge where such traces
as hammer stone stripe (Keeley 1980)
might be expected.
The results of these experiments show that wear traces do occur during retouch
and possible minor surface alterations can occur when the blank is removed
from the core. Haft wear has been defined as wear traces "which make
little sense as traces of utilisation, but does conform to what is known
or expected of wear traces from minor movements of a tool against its haft"
(Cahen et al. 1979,681). The surface
alterations from the manufacture of these tools fall within this definition
and could easily be confused with genuine haft wear.
With prehistoric material the tools can only be observed after all the stages
of manufacture, hafting and use have taken place, and so it is impossible
to be precise in assigning the traces to a particular stage in the process.
By careful observation of this process it has been demonstrated that wear
traces on the non-operational end of a tool are not necessarily due to hafting
and it is suggested that extreme caution should be applied before attributing
any traces to hafting.