Introduction

In a recent paper in Antiquity, The tula adze: manufacture and purpose (Reference MooreMoore 2004), Mark Moore suggests that the production of tula adze-flakes is inefficient, with siret fracture being a common cause for failure. In this, otherwise well informed, paper Moore notes that siret failure was 'approaching' 70 per cent in his own replication experiments, a high figure under any circumstances, but one that needs to be treated with caution.

Knapping tula adze flakes - some technological considerations

Lithic replicative studies are of great use in understanding the techniques and technology of stone knapping - with one proviso: if an experimental knapper has a high failure rate in an experimental situation, it cannot be automatically inferred that the aboriginal technique being examined has not been fully comprehended.

While acknowledging that siret fracture is a major cause of failure in the production of gull-winged flakes it is not necessarily a failure rate of high order. From my own knapping experience I would suggest a less than 10 per cent rate of failure is a more realistic figure to consider.

I have been manufacturing and using tula adzes for more than thirty years and in this time have examined tula adze-flakes across the continent. My observations suggest that Moore has failed to notice a number of subtle clues present in assemblages of tula flakes that indicate the manner in which indigenous knappers minimised the probability of failure. While I do not have statistical data to present, I can say that the features described are in my experience more commonly found on tula flakes than on any other flaked stone artefact.

The first clue is that tula flakes may show one or more unresolved ring cracks present on the striking platform. The presence of these ring cracks suggests that the knapper is using very controlled force to remove the flake. When the force is not sufficient to effect flake removal an unresolved ring crack is created. The knapper then uses a second, slightly more forceful stroke to remove the flake. It may also indicate that the knapper takes at least one preliminary stroke at the core, in the manner of a carpenter driving a nail, before removing the flake with a second, equally controlled but more forceful blow.

The second feature that reinforces this suggestion is that in any collection of tula flakes or adzes, a number with double cones of percussion can be expected. Double cones occur when the blow that removes the flake initiates a ring crack adjacent to the ring crack created by the first 'setting up' stroke. The removal of the flake clears the rear section of the unresolved ring crack, leaving it revealed on the ventral flake surface as a secondary bulb of percussion.

The presence of either the unresolved ring cracks or the secondary bulbs of percussion suggest that indigenous knappers were ensuring that both the placement and velocity of the knapping stroke were precisely calculated. When siret fractures have occurred in knapping tulas experimentally it is clear that the platform is too thin, in relation to its width and excessive force has been used to remove the flake.

Unresolved ring cracks that are virtually circular in form; and the usually prominent sub-conic form of the bulb of percussion immediately below the ring crack on the ventral surface of tula flakes and the rapidly expanding bulb of percussion, indicate that a high velocity force was applied at an angle nearer to 90° to the platform rather than the usual 60° - 80° used for general flake removal. The dimensions of the ring cracks and PFAs suggest that hard hammerstones were used to remove the flakes and contact with the core was concentrated and minimal.

The sum of these clues suggests therefore; that tula-flake knappers used precise, high velocity forces to regularly create short, wide flakes with pronounced and rapidly expanding bulbs of percussion in a controlled manner with a far lower rate of failure than suggested by Moore.

It should also be noted that large tula adze flakes, that is with platform widths greater than 35mm were regularly produced on the Barkly Tableland and in the Eyre Basin, Elsewhere within their range tula adze flakes are generally smaller. Less force is required to detach such flakes and consequently the problem of failure through siret fractures may not be of the magnitude suggested by Moore.

Further points in relation to the efficiency of the technique must certainly be the lack of complexity of the knapping process and the time involved to produce tula flakes. Unlike the production of Folsom points, which involves a complex sequence of discrete events that require possibly an hour or more to accomplish, a tula flake can be produced in a matter of seconds. As most tula flake production appears to have been undertaken at the quarry or other stone source, failures of either 50 per cent or more would not preclude a knapper producing an annual supply of suitable adze flakes in less than an hour.

The function of tula adzes

In relation to the function of tulas, it is commonly held that tulas were created as a response to the working of several varieties of hardwood commonly utilised by indigenous Australians of the arid regions.

However, having examined the form and distribution of Australian grindstones, receptacles and tula adzes over the past few decades, I would suggest that tula adzes, usually presented as a response to the need to work arid zone hardwoods, reflect much more in terms of the general economy and the environment. After all Aboriginals living in other areas of Australia also regularly worked many hardwoods without tulas. Similarly, an examination of ethnographic adzes from the western desert regions of Australia and the literature relating to the material culture of this region clearly show that tulas were neither made nor used in historic times. Archaeologically however tula adzes and slugs are found throughout the Western Desert.

Tula adzes are, on the other hand, found both archaeologically and ethnographically in the hot, but not so arid Pilbara region. They also occurred ethnographically in the Kimberley, south of the Fitzroy River, in the Tanami Desert and central Australia, across the Barkly Tablelands and though the riverine areas of western Queensland into the Eyre Basin.

In all these areas extremely well made grindstones, often of very large sizes were also used in recent times and there is ample evidence that grains and other seeds were regularly stored in very large wooden receptacles. Ashwin (1932: 64) describes finding a large Aboriginal camp near Newcastle Waters in the Northern Territory, which included a well constructed shelter, 7' (2+m) high and 16' (5m) in circumference. As well as containing large bundles of spears, the shelter contained 17 large wooden dishes, each over a metre in length, containing an estimated ton of seed. Ashwin believed the seed to be derived from a wild rice that grew in a vast 80km stretch of country that lay to the north of Newcastle Waters. Throughout the regions where tulas were used ethnographically, large carrying vessels were consistently made of the softwood Erythrina verspitilio. In these areas shields were also generally made from this timber or, in the case of the Pilbara region, also of the softwood Brachychiton gregorii and medium density timber Hakea loria. It is necessary therefore to consider that tulas were created to work softwoods as much as they were a response to the working of harder timbers.

I would contend that, tula adzes were intimately associated with the production of soft- and hard-wood vessels for large scale storing of seeds and grains, softwood shields and possibly a more sedentary existence afforded by the storage of grains; rather than the production of the hardwood implements that are found in the more mobile hunter/gatherer economy as practised in recent times by the people of the Western Deserts. If this is the case, then the archaeological presence of tulas in the Western Desert suggests that, in the past, there existed an economy based on a higher reliance of seeds and grains and the storing of these foodstuffs, than occurred at the time of contact.

The decline in the practice of storage of seeds and grains, and the lessening for the need for greater numbers of storage vessels led to the abandonment of manufacture and use of the tula adze in the Western desert, even though they were still used by neighbouring groups in the Pilbara, south Kimberley, Tanami and Central Australia. An examination of a number of factors, including changes in vegetation patterns in the Western Desert over time, may provide greater data on the reasons why tulas dropped out as an element of the Western Desert suite of stone artefacts.

In conclusion, the production of tula adze flakes is not the high risk strategy as suggested by Moore, but rather one by which, in a manner of minutes a knapper can produce a supply of adze flakes for either personal use or exchange, that would last many months before requiring replacement.