Australia: The Land Where Time Began

A biography of the Australian continent 

Tulas - Are They Linked to ENSO in Australia?

The tula is an endemic hafted tool used to work hardwoods by Aboriginal people throughout central and western Australia, totaling about 2/3 of the continent. Tulas also had other functions such as butchery and plant processing. Tulas appear to have spread rapidly across their range and to date no antecedent tool of this form is known of in Australia. According to the authors1 tulas first appear in the archaeological record at the same time as the onset of the ENSO climatic conditions.

In this paper the auhors1 propose that the appearance of this new, specialised tool at about 3,700 BP was likely to be a response to intensification of the ENSO, though the data is not yet sufficient to establish unequivocally a causal link between this tool and the ENSO intensification. As a result of this intensification aridity and climatic variability increased and lasted for nearly 2,000 years. The authors1 propose the appearance of the tula was an element of the toolkit they adopted as part of their minimisation of risk, and was a part of the wider social and economic strategy adopted to cope with the increasingly difficult climatic uncertainty. This possibility has implications for the innovation process diversity operating in Australia in the Holocene. According to the authors1 their discussion in this paper is a platform for future studies which they believe are necessary.

It has been noted that tulas used by Australian Aboriginal people were of a unique nature (Holdaway & Stern, 2004, 253-256). These distinctive tools that were used across much of the arid zone of Australia, were comprised of a discoidal flake that had been retouched and hafted on the ends of wooden shafts. They were seen to be used for scraping wood, incising and adzing in the historical period. It is a composite tool which often has the stone implement fixed to a wooden shaft, that was constructed by regular and distinctive production technology, and they have been seen to be an endemic adaptation, adopted by Aboriginal people throughout 2/3 of the Australian continent, its use being geographically focused on central and western Australia. Tulas are a technologically distinct form of tool which is useful in analysis and the authors1 have suggested that they may have had emic significance to the Aboriginal people who use them. The authors1 suggest they had adaptive significance, regardless of the perceptions of the knappers of these tools. The properties of tulas compared with those of unmodified stones varies between stratified sites that have been dated (e.g. data in Gould, 1977 vs Smith, 2006), and within palimpsest open sites (Veth, 1993), there is no suggestion of an emergence, and subsequent proliferation phase, at an earlier time, as there are for other tools such as backed artefacts. The authors1 suggest that relatively high levels of residential mobility are indicated by a number of factors, such as the often 'exotic' provenance of tulas, the high numbers and ubiquity of tulas that are found in many sites in the arid zone. The careful, repeated production and selection of robust flakes that were broad platform and deeply convex, that were required for these tools, that were high-impact hafted tools, is unique, though the method of resharpening the tula flakes was not necessarily different from that used for retouching 'scrapers' that were in use prior to the appearance of tulas in the archaeological record, and there are no known antecedents of the tula (Hiscock & Veth, 1991).

The authors1 say their primary aim in this paper is to review the evidence of the appearance of the tula and using this evidence to discuss plausible innovation processes in the Holocene in Australia. Studies have shown that tulas began appearing in archaeological sites dated to 4,000-3,000 BP (Hiscock & Veth, 1991), though it was previously believed that tulas had been in use throughout the Holocene (e.g. Gould, 1977). Models of innovation of other tools used in Aboriginal Australia have been discussed extensively (e.g. Attenbrow, 2008,2009), questions of the adaptive context in which tulas appeared and use of tulas over a wide areas of Australia have not been addressed specifically. The authors1 suggest that the process of the emergence of the tula may have differed from that of other standardised stone tool types, though the adaptive context for the appearance and spread of the tula is similar to that of other standardised stone tools from the mid to Late Holocene. A study of the existing models of technological change in Australia clarifies the nature of these differences.

Percussion or a combination of percussion and pressure is used to produce bifacial points with strait or convex convergent retouched margins, which are mostly restricted to the northwest region of mainland Australia. Backed artefacts, found across much of southern and eastern Australia, have a margin that is retouched steeply, often with a bipolar technique, and in appearance it is similar to microliths from the Old World. Evidence has been found indicating that there is a long history of these implements in Australia dating back to the early Holocene, at least, and possibly as far as the terminal Pleistocene, which largely resolved the uncertainty of their chronology. For a number of millennia they were produced at low rates, the production rate increasing in the Late Holocene (e.g. Clarkson, 2007; Hiscock, 1993, 1994, 2002,02006; Hiscock & Attenbrow, 1998, 2004; Slack et al., 2004). Proliferation events, the name used for periods in which implements were produced at much high rates, one of which occurred between 4,000-2,000 BP for bifacial points in northwest Australia and for backed artefacts in eastern Australia (Hiscock, 2006). The authors1 suggest proliferation events represent an adaptive response, and as they occur long after the invention and use of the tools, they involved the emphasising of a tool form that was pre-existing to new adaptive contexts (Hiscock, 2002; Hiscock & O'Connor, 2006).

It has been difficult to define the use of these tools and the circumstance that led to their regular production 4,000-2,000 BP has been achieved only for backed artefacts in eastern Australia. It has been found that backed artefacts were not often used for projectile armatures (points or barbs) as had been the initial belief, rather, combined residue and use-wear studies have shown they were used in a wide range of craft-related processing activities, such as wood-working, skin-working, plant-working and feather-working, and many specimens show evidence of multiple and different uses (Robertson et al., 2009). It has been suggested (Attenbrow et al., 2009), based on the lack of any evidence that indicates any shift over time in the uses and tasks of backed artefacts, that backed artefacts were especially prolific in the period 4,000-2,000 BP because the tools in which they were hafted had multiple uses. The authors1 suggest this evidence is consistent with a proposal that foragers in Australia emphasised composite tools that contained backed artefacts because of their readiness and multifunctionality, as they used them for nearly all tasks (Hiscock, 1994, 2002, 2006, 2,008). It has been proposed (Hiscock, 2002, 2006) that tools that were standardised, ready and multifunctional, that had backed artefacts attached, were emphasised at the time of the proliferation event as a result of the increased level of environmental variability associated with ENSO conditions that were intensified at this time, and as there was a need for mobile hunter-gathering at these times. Such tools reduced the risks associated with foraging, in the context of increased resource variability, by providing tools that were reliable and easily repaired, and that increased resource acquisition and manipulation on most occasions. The conclusion that they had long use lives is consistent with observations that production rates of backed artefacts increased during the period 4,000-2,0000 BP, and more extensive maintenance was carried out of other retouched flakes (Hiscock, 2008; Hiscock & Attenbrow, 2005). According to the authors1 a similar proposal has been put forward for the articulation of northern bifacial point proliferation as the climatic variability increased with ENSO intensification (Hiscock, 2002, 2006), though the nature of tool use has been studied less for points.

An image of technological change during the Holocene has been provided by these inferences, the high production rate of some tools representing an adaptive response to greater variability of the climate, at least a significant factor, that involves an emphasis on tool forms that were preexisting, as a solution to economic contests that were new. According to the authors1 they argue in this paper that the tula, the 3rd tool type to be widely distributed, appears over a time span that they describe as 'archaeologically instantaneous', correlating with the intensification of the ENSO that occurred between 4,000-2,000 BP.

Australian tula

According to the authors1 a common description for tulas is as semi-discoidal retouched flakes with a pronounced ventral bulb and a convex cutting edge when seen in plan view (Hiscock & Veth, 1991; Holdaway & Stern, 2004: 253-256; Spencer & Gillen, 1904. These stone tools were historically observed to be hafted at the end of wooden shafts and spear throwers, primarily by using Triodia-based resin (Sheridan, 1979). Central and western Australia is the centre of focus of the geographic distribution of tulas, though examples of tulas or tula-like objects have been found along the eastern seaboard central regions (McNiven, 1993; Hiscock & Attenbrow, 2005).

The authors1 suggest composite tools that contained tulas probably had a variety of functions, one of which, that has been discussed, being as an adz for use with woods that were very dense, such as mulga, Acacia aneura. Tulas could be used to scrape, grave, saw and chisel, and their construction often made them suitable for adzing hard timber (e.g. Sheridan, 1979). The authors1 suggest when tulas were being used for wood-working scraping was likely to have been the dominant use over adzing, and they were probably used for a range of other tasks such as butchery, vegetable processing, and possibly others (after Gould, 1977). They also suggest that as a category tulas are multifunctional, though in other contexts were mainly used for wood-working. The distal working edge of the tula bluntness over a short time when it was being used for work on hardwoods that would require the distal edge to be sharpened by frequent retouching to extend its working life, the authors1 suggesting that this retouching would allow the tula to be used for more than an hour (Hayden, 1977). As the distal margin is being constantly retouched the cutting edge would gradually migrate towards the proximal end of the tula and in the process changing the plan shape of the retouched edge that changes from convex to straight and ultimately to concave (e.g. Cooper, 1954; Gould, 1977; Hiscock & Veth, 1991). The changes have been used as an exemplar of morphological changes accompanying maintenance of the edge (e.g. Dibble, 1995), the changes having been documented in assemblages that have been excavated (e.g. Hiscock, 1988). The tula 'slug' is either discarded of recycled as the tool becomes less firmly fixed to the resin haft, and becomes less able to be retouched and is of less use for heavy scraping or adzing. The ease with which it can be resharpened, and the likelihood of its discarding, are affected by the steepness and character of the step terminations on the retouched edge. Tulas have been found in different stages of maintenance and use in archaeological sites, and tulas that have been completed that have remained unused have occasionally been found in archaeological sites (Veth, 1993), and caches of unused tulas have been found on rare occasions (Hiscock, 1988). At many sites slug stage specimens that have been used and retouched are frequently found. Known archaeological instances of tula and tula slug specimens in archaeological sites in Australia have been used to construct a chronology of this tool type.

Tula Technology - Chronology

As a result of sampling and taphonomic issues (Gorecki et al., 1997; Hiscock, 2001; Hiscock & Attenbrow, 1998) the construction of reliable chronologies for implements in Australia during the Holocene has been difficult. The authors1 suggest it may be difficult to be sure whether a few, or possibly no implements were made in a particular time period when the rate of implement manufacture was low, or archaeological assemblages are small, the probability of finding the rarer implements, such as tulas, is also low, in such circumstances large volume excavations and large assemblages are valuable. Another problem for determining the chronology of tools is the occurrence in some sites in which sediment accumulates rapidly, as these are often the most resilient sites in which vertical displacement of artefacts occurs after deposition. Another possible problem is the accuracy of the identification of an artefact as a tula, and in some cases this requires re-examination by an expert. The authors1 suggest these points must be remembered when reading the following account of the chronology of tulas.

The basis of the inference of a 10,000 BP Early Holocene age for tulas is the tool found at Puntutjarpa (Gould, 1977). The Puntutjarpa Rock Shelter is located near the Warburton Ranges in Western Australia that was excavated in the 1960s2. The claim of the age made by Gould is based on small flakes that had been retouched dorsally, believed by Gould to be small tulas that he called micro-adzes. The only artefacts recovered from the lower levels, dated to the Early Holocene, at this site were the 'micro-adzes', and the proposal that this tool category dated to the lower Holocene is dependent upon the correct identification of these tools as tulas. The 'micro-adzes' have been demonstrated to be technologically and functionally different from the tula, lacking small platforms, convex ventral surface and slug stage of reduction (Hiscock & Veth, 1991). It has been found that definite tulas and tula slugs, that had proportionally large platforms, are only present in the upper levels of the deposit, that probably dates to the middle or late Holocene (Hiscock, 2008: 213-214). Any further statement of age of tulas at this site has been made difficult by the imprecision of the dating of this site, the authors1 suggesting it is more profitable to look at other sequences and regions.

Tulas are known only from the most recent deposits at archaeological sites in Australia, suggesting an age of first appearance of no earlier than the late Holocene, in spite of extensive excavations being carried out at many archaeological sites in the arid zone, many of which have been dated to more than 30,000 BP (Gould, 1977; Hiscock, 2008; McNiven, 1993; Smith, 2006; Veth, 1993, 2005).

In the Cleland Hills, the deposits at Puritjarra Rock Shelter, the first appearance of tulas is at about 3,500 BP (smith, 2006: 378). According to the authors1 of the 34 tulas recovered all except 2 were distinctive slugs and 88 % had been made from chert and chalcedony, and both materials hold their edge well (Smith, 2006: 393-395). Sites such as the Devon Downs Rock Shelter on the Lower Murray, 3,500 BP; the Kwerlpe Rockshelter, Northern Territory, 3,635 BP (Gould, 1978; Smith, 1988); and the Mt Newman Rock Shelter on the Hamersley Plateau (Marwick, 2009), were other sites that contained tulas with early dates. These tulas were recovered from sites across much of arid and central Australia with consistent ages of about 3,700-3,500 BP.

Evidence of younger dates for the initial appearance of tulas comes from the east coast of Australia, though it is difficult to interpret this evidence. The maximum age that has been found in the northeast is at Platypus Rockshelter, less than 2,700 BP, and Brooyar Rockshelter, also less than 2,700 BP; Broadbeach Aboriginal Burial Ground, about 1,300 BP; Gatton Rockshelter, about 1,000 BP (McNiven, 1993). At Capertee 3 Rockshelter, on the central part of the eastern seaboard, a tula has been found that dates to the last 1,700 years (Hiscock & Attenbrow, 2005:110). The maximum age of tulas found along the east coast in uncertain as the number found in the east coast sites is small, indicating that to date there are no known tulas dating to earlier than the Late Holocene.

It has been revealed by this review by the authors1 that there is a consistent chronological pattern, in spite of the current state of uncertainty, that the earliest tulas date to 3,700-3,500 BP across much of arid and central Australia, and in eastern Australia, from about 2,700-1,500 BP. The authors1 suggest this antiquity is consistent with the changes that occurred in the climate between 3,700-2,000 BP.

Implications for arid Australia of the ENSO intensification

Tulas - ENSO and technological change

Tulas suddenly appeared and proliferated between 4,000-2,000 BP in the archaeological records in central and western Australia, coinciding with the the intensification of the ENSO, that also occurred between 4,000-2,000 BP. According to the authors1 the chronological coincidence between the appearance of tulas and the ENSO intensification raises questions regarding the causes and processes of technological change. The authors1 propose that the appearance and dispersal of tulas that occurred over much of the continent between 4,000-2,000 BP has a connection with functional and economical changes that affected desert societies at this time that was triggered by an increase in aridity generally, as well as greater climatic variability. They pose the questions below that may advance this proposition by further research.

It has been concluded by a number of archaeological studies that forager settlement patterns were changed after 4,000 BP with residential mobility of groups  increasing (Smith, 1988; Thorley, 1998; Veth, 1993). The authors1 suggest the increased mobility at this time may have been the result of reduced carrying capacity, greater patchiness of resources, and lower predictability, all resulting from intensification of the ENSO. Tulas are an example of curated tool (see Sheridan, 1979) and typically curated toolkits are associated with an increase of mobility. Alterations to toolkits would be expected to result in the context of increased residential mobility, and resources that are less reliable. It has been suggested that the tula,  as a curated tool that has multiple purposes, was part of a technology developed for a risk minimising strategy linked to increased environmental stochasticity (Hiscock, 1994, 2006).

One of the technological and economic responses that could be predicted by modelling of risk response (Hiscock, 2008:159-161) Veth, 2005) is precisely the adoption by arid zone foragers of stone tools such as tulas. Technological shifts and innovations as part of a new behavioural suite that has incorporated new ways of using landscapes, and presumably conceptualising them, were included in prehistoric responses to foraging risk in Australia. The authors1 suggest there was probably a broad base for the responses to heightened risk that was present in the late Holocene, that included the technological evolution, as well as changes in the sizes of groups, their residential mobility, the area of their territory, their trade and exchange relations, and between the social and political dynamics within and between groups of hunter-gatherers.

Tulas represent a process of technological innovation that is different from that of other forms of implement that were abundant during the Holocene, especially backed artefacts and bifacial points. The authors1 say that tulas appear to have had no precursors in Australian archaeological assemblages prior to their appearance at about 3,700 BP, so they don't represent an emphasis on, or elaboration of, any forms that were pre-existing, that appear to underpin the backed artefact and point proliferations. Tulas therefore appear to have been a novel invention that represented a response to changed conditions of foraging in which the production and maintenance of wooden implements that were 'energy-extractive', and were of technological and evolutionary significance.


The first appearance in the archaeological record of the tula in the late Holocene, based on current evidence, that can be dated to between 3,700-2,000 BP, as indicated by its sudden appearance in great numbers in a wide range of sites across central and western Australia. It is unquestionable that these timeframes coincide with the extreme ENSO conditions that were characterised by aridification across vast areas and the increased variability of climate across Australia.

Tulas appeared 'suddenly' in archaeological terms, and as there is no known evidence of precursors, the authors1 suggesting the making of tulas may have been a strategic innovation in order to deal with changes in altered social and economic circumstances, a likely triggering possibly being intensification of ENSO conditions, that would have had their most extreme impact in central Australia. The authors1 postulate that, in the context of Australian environments in the Late Holocene, the toolkits assembled to assist foragers in the new, more difficult conditions, probably contained tools that were easily repaired and deployed for many tasks, some of which may have been unforeseen. Tulas are versatile stone tools that were used to make and repair a wide range of energy extractive wooden implements such as spears, spear-throwers, wooden bowls and digging sticks, and they remained serviceable for extended periods. Among people who are highly mobile nomadic foragers, for whom territoriality and residential mobility is at a premium, such wooden implements were very useful (Hiscock, 2008, 159; Yu, 2006).

The authors1 argue that the closeness of ENSO shifts of climate and the appearance in the archaeological record of tulas suggests a link that may hint at the process of innovation. The authors1 suggest that the evolutionary process applying to the class of tools called tulas is determined  by the reason for the innovation of this tool compared to the evolution of other tools in the Holocene of Australia. If they were developed as a response to the onset of a climate change it would suggest their evolutionary process differed from that of other examples of technological change that occurred in the Holocene of Australia, such as backed artefacts and bifacial points, which display a proliferation of pre-existing element of the tool-kit. According to the authors1 this is at the core of assumptions which are the basis of selectional approaches in Archaeology. According to the authors1 these 'posit essentially that proliferation (or reward) of 'fit' behaviours and pre-existing technologies will occur, even if these are not known to be advantageous to the people at the time'1. A selection process and its associated assumptions are not confirmed to by the sudden appearance of tulas in the archaeological record. The invention of an entirely new tool led to the appearance of tulas in the archaeological record, rather than the expected selection and proliferation of an extant tool.

The authors1 suggest the implications of these conclusions need to be addressed in future research. According to the authors1 a number of questions have been raised but not answered by this study. One question is the reason different mechanisms of change occurred for different classes of tool? Do the differences inferred between selectionist (bifacial points, backed artefacts) and non-selectionist (tula) processes indicate differences in the nature of social and physical contexts or in contingent qualities of different groups or technologies? Another question arising from this study was the reason for the spread across the arid zone of tula use, as well as beyond the arid zone. The reason for the chronological difference between the earliest known tulas in arid central Australia and eastern Australia is not known. A question asked by the authors1 is whether this should be interpreted as indicating the speed of dispersal of the innovation, and should the similarity of the ages of the earliest known tulas found across central Australia be interpreted as evidence of rapid transmission by cultural means of the innovation. It is also not known if dispersion of tulas was because of its economic or social benefits, or was it associated with a dispersal of the population? The authors1 also ask if the function of the tula can be refined to provide insight into organic tool production, and such information could clarify the context of tula dispersal and innovation? They also raise the question of what the foraging risk could have been in the Late Holocene as to why it would trigger this kind of technological change?

It has been argued in the past that there may have been a significant effect on the Aboriginal populations in the arid zone that was caused by the ENSO. A time series analysis was carried out on a large database from radiocarbon dates sourced from the archaeological records of the Australian drylands. In the study it was postulated that changes in the sum probability curves in the arid interior, Pilbara region and the Murchison region were related to ENSO shifts. The relationship between ENSO and the Australian drylands was investigated when a study was subsequently carried out (Williams et al., 2008). It has been argued, using a similar 'dates as data approach' (Williams et al., 2008) that ENSO initiation in the Late Holocene resulted in a series of changes in the distribution of populations of Aboriginal people between 3,700-2,000 BP, including responses that were inferred as regional abandonment and a population that declined.

Sources & Further reading

  1. Veth, Peter, Hiscock, Peter & Williams, Allan, 2011, Are tulas and ENSO linked in Australia, Australian Archaeology, No. 72, June.
  2. Puntutjarpa - Research Data




Author: M. H. Monroe
Last updated: 24/03/2013
Journey Back Through Time
Experience Australia
Aboriginal Australia
National Parks
Photo Galleries
Site Map
                                                                                           Author: M.H.Monroe  Email:     Sources & Further reading