Australia: The Land Where Time Began

A biography of the Australian continent 

Ground-Edge Axe – Oldest in the World Coincides With Human Colonisation of Australia

In this paper Hiscock et al. report evidence of the earliest ground-edge axe in the world, at 44,000-49,000 BP. The age of this axe coincides with the time when the Aborigines were still in the process of colonising the Australian continent. The discovery of ground/polished axes exemplifies the diversification of technological practices that occurred as modern humans dispersed out of Africa, as this type of axe has not been associated with the dispersal of Homo sapiens across Eurasia. Ground-edge axes have now been found in 2 colonised lands at the time humans arrived and Hiscock et al. therefore argue that these technological strategies are associated with the adapting of economies and social practices to new environmental contexts.

The evidence of ground-edge axes was uncovered in northern Australia dating to 44,000-49,000 BP. This age makes it the earliest evidence of a ground-edge axe to be reported in the world to date, and such evidence has implications for the dispersion of modern humans form Africa, as well as the nature of the first occupation by humans in Australia.

Australian stone lithic industries from the Pleistocene have, according to Hiscock et al., been persistently chacterised as being extremely and uniformly simple, being tools that were unstandardized and expedient, which means that the discovery of ground-edge axes in Australia is challenging.

Ground-edge axes have been reported across much of northern Australia that dated to the terminal Pleistocene, at Widgingarri 1 and Carpenter’s Gap I and 3 in the Kimberley, Western Australia (O’Connor, 1999; O’Connor et al., 2014) at Malanangerr, Nauwalabila 1, Nawamoyn, and Nawarla Gabarnmang, in western Arnhem Land (Geneste et al., 2010; Jones, 1985; Schrire, 1982; White, 1967), and on Cape York, Sandy Creek (Morwood and Trezise, 1989). Hiscock et al. say these ground-edge axes were invented locally. In the islands to the north of Australia such ground-edge axes first appear in the Neolithic, also, there is no evidence that this technology was introduced to the Australian continent. The unanswered questions until now are when axes were invented, and in what manner did that invention relate to the colonisation process? In this paper Hiscock et al. present evidence for the production of axes close to the time of colonisation of Australia.

There are not large numbers of axes in assemblages as they are long lived. Therefore the early axe chronology cannot be based solely on recovery of whole axes, which are rare, and the discovery of the presence of such axes from well-dated excavations depends on flakes that have been removed from ground bevels on axes as they were being resharpened or repair of damaged and worn edges. The identification of axes that have been dated to 35,000 BP in Australia (Geneste et al., 2010) has been based on such flakes with parts of a bevelled edge on them. There have been suggestions of even earlier axes from Madjedbebe (Malakunanja II) based on the presence of small flakes of volcanic material in sediments that were dated to more than 40,000 BP (Clarkson et al., 2015). Hiscock et al. are cautious of this interpretation as the flakes lack diagnostic ground bevels (Clarkson et al., 2105: 173). The repair of a bevel often involved removing a number of flakes before the edge was reground, and such a repair cycle may be repeated several times, the process making an order of magnitude more flakes than axes deposited in the archaeological record. It is, therefore, the polished bevel that defines specimens as ground-edge axes, and reshaping the flakes that remove the bevel as identifiable as the complete axe. The angle of ground bevels ranges from 60o to 100o (Dickson, 1981: 104), and the characteristics of the edge can vary significantly over the life of the axe as multiple uses and repairs are carried out on it (Kononenko et al., in press). As this range of angles overlaps with the range of angles that are produced by core reduction they are not reliable as a sole diagnostic trait. The ground surface that is highly polished is the only morphological feature that is unique to axes. Extensive abrasion with another rock is used to make smooth surfaces which cannot be produced incidentally by other knapping actions such as the preparation of platforms. It has been shown by experiments that grinding basalt to a polished level takes from 1.5 to 5.0 hours depending on the abrasive agent that is used (Dickson, 1980). Hundreds of forceful strokes are needed to produce a smooth bevel, even under optimal conditions. This proposition is confirmed experimentally and it is indicated that though the smoothness of ground surfaces vary, they are always smoother than fracture surfaces. Hiscock et al. say the key indicator of axes they use is convergent bevels with high surface smoothness which is achieved by extensive abrasion, and this is applied to the identification of small flakes from axes that are produced by bevel reshaping.

Recent discovery of such flakes have demonstrated that axes were made in Australia at least 30,000-35,000 BP (Geneste et al., 2010; Jones, 1985; Morwood & Trezise, 1989; O’Connor, 1999; O’Connor et al., 2014). Evidence that has been presented in this paper from excavations at Carpenter’s Gap Shelter 1 demonstrates that ground-edge axes were being made in Northern Australia more than 10,000 years earlier. This is the earliest evidence to be reported in the world to date of ground-edge axes, and it reveals that the first Australians were technological innovators who developed grinding and abrading as techniques to be used for shaping a range of new implements that included hafted ground-edge axes. Hiscock et al. argue that the evidence from Carpenter’s Gap Shelter 1 shows that this kind of innovation arose as humans dispersed from Africa as they invented regional traditions as part of their adaptations to new landscapes.

Carpenter’s Gap Shelter 1

Carpenter’s Gap 1 (CG1) is among the oldest known habitation sites in Australia that have been dated by the radiocarbon technique. The first excavations at this site were carried out in 1992 and 1993 during which 5 1 m square test pits were dug to bedrock (Frawley & O’Connor, 2010; O’Connor, 1995). Square A2, which is close to the large rockfall that trapped the deposit within the upper part of the shelter, to produce the artefactual material that is discussed in this paper. Units average a depth of 2 cm in this site excavation but were dug within depositional units. A hearth 10 cm deep, e.g., would be removed separately from other sediments, to be treated as 1 stratigraphic context, though it would be divided into excavation units of 2 cm depth to enhance assessments of provenance.

The sediments have primarily accumulated in the shelter, dating from the Upper Holocene, as a result of in situ weathering of layers of softer sedimentary rocks that are embedded in the limestone reef the cave formed in, as well as a component of aeolian deposits (Vannieuwenhuyse et al., in press), overlie sediments of Pleistocene age that have been dated to approximately 49,000 cal. BP through to approximately 18,000 BP). Assemblages of Holocene age were found to be restricted to layers 1-4, and most of the deposit accumulated prior to the LGM. Cultural material was deposited throughout the site, beginning in excavation unit 61, which was significantly below deposits dating to 44,000-49,000 BP. The axe fragment that was of the earliest age was recovered from excavation unit 52, near the base of the cultural sequence.

The specimen referred to in this paper as Carpenter’s Gap Axe Flake 1 was recovered from Square A2 unit 52 was designated cg1/a2/52/1. A charcoal sample was also found with it and it was dated to 48,875-43941 cal. BP (WK-37976). Hiscock et al., argue that the axe fragment and the charcoal sample are associated stratigraphically and therefore constitute evidence that the axe grinding technology that was employed in the manufacture of the axe at or immediately after the arrival of the humans in Australia.

Some have questioned the chronological integrity of the early Australian assemblages (e.g. Allen & O’Connell, 2003, 2014; O’Connell & Allen, 2014), their argument being that for older specimens their post-depositional relocation placed them in a false association resulting in radiometric estimates of an early age. The reply of Hiscock et al. is that though their critique is overdrawn (Hiscock, 2013), the possibility of movement should be examined for each deposit. The assemblages from the Pleistocene at CG1were evaluated to determine if they had been affected by vertical displacement by looking for size-sorting of artefacts within the lower deposit. This test of postdepositional movement of materials within archaeological deposits, as there are a variety of processes that act to lower small specimens and/or raise larger ones (Bocek, 1986; Cahen & Moeyersons, 1977; Hofman, 1986; McBrearty, 1990; Schiffer, 1987; Stockton, 1973; Wood & Johnson, 1978). Therefore, Hiscock et al. predicted that there would be smaller specimens in unit 52 and adjacent levels than in ones that were immediately higher, if there had been significant movement involving displacement of specimens into unit 52 from higher levels in the deposit. With this in mind Hiscock et al. examined the relationship between the size of artefacts and their depth  for specimens in excavation units 45-60, which represents MIS 3 – the period prior to the Last Glacial Maximum (LGM). Using univariate GLM (General Linear Model) and non-parametric regression statistical tests established that there was no significant relationship between depth and artefact mass

(F = 0.043, d.f.= 15, p = 0.975, rs = 0.011, p = 0.914, N = 100),

maximum artefact dimension

(F = 0.882, d.f. = 15, p = 0.586; rs = 0.079, p = 0.433, N = 100) or

Flake percussion length

F = 0.998, d.f. = 12, p = 0.477; rs = -0.141, p = 0.384, N = 40). 

Hiscock et al. view the failure to find size-sorting as refuting the hypothesis that vertical movement of artefacts had occurred within the oldest levels of the deposit. Other lines of evidence are consistent with this conclusion. For example, excavation units 51-53 commonly contain basalt flakes, though they are rarer in high levels, 42-50, which indicates there is minimal ‘reservoir’ of similar specimens from which the axe flake, cg1/a2/52/1, could possibly have derived. Also, small and large artefacts, as well as the limestone plaque that was recovered from the base of the deposit, that was covered with ochre (O’Connor & Frankhauser, 2011), was discovered which was lying horizontally. Regular displacement of material is not suggested by these observations. As a consequence, Hiscock et al. were confident that this specimen is associated, stratigraphically and temporally, with the radiocarbon sample in that excavation unit, having an antiquity of 44,000-49,000 ca. BP.

Axe production demonstration

The interpretation that is offered by Hiscock et al. relies on reliable data as well as a clear identification of the technological character of the specimen in question.

Archaeological comparison

The comparative sample used by Hiscock et al. was comprised of artefacts from 3 categories comprising a total of 50 artefacts;

(1)   The specimen that was discussed in this paper, Carpenter’s Gap Axe Flake 1,

(2)   11 axes and axe fragments recovered from sites in the Kimberley and adjacent regions, and

(3)   38 basal flakes that were not ground from levels 48-52 of CG1.

Included in the 3rd category were flakes which had surfaces that were weathered and slightly patinated.

Mean values for Ra ratios and Rzjis ratios differ significantly for basalt ground-edge axes and flakes with no grinding

(Ra ratios t = -3.810, d.f. = 10, p = 0.003; Rzjis ratios: t = -3.089, d.f. = 10,  p = 0.011).

Contrasting with this the Ra and Rzjis ratios of ground-edge axes do not differ significantly from those of unground axes, and the specimen from excavation 52 of CGI that is reported in this paper

(Ra ratios t = -0.541, d.f. = 10,  p = 0.601; Rzjis ratios: t = -0.542, d.f. = 10, p = 0.600).

According to Hiscock et al. these results are consistent with the proposition that the smoothing of the platform and dorsal surface is unlike the surface of basalt at Carpenter’s Gap 1 that has been flaked or weathered and is indistinguishable from the ground faces of axes. Given that the morphology of the platform and its junction with the dorsal face of cg1/a2/52/1 is the same as is seen on the typical bevels of axes, and that abrasion, that has been extensive and laborious, has smoothed the basalt to the same extent as is observed on axes, the conclusion of Hiscock et al. is that Carpenter’s Gap Axe Flake 1 (cg1/a2/52/1) must be a flake that has been removed from the polished edge of a ground edge axe.

Technological novelty and the colonisation of Australia

Hiscock et al. suggest that the date of the production of ground-edge axes is close to, and possibly immediately after, the age that has generally been accepted for the colonisation of Sahul (Hiscock et al., 2008). It is now clear that ground-edge axes first appear in the archaeological record shortly after the landfall of the earliest colonisers. Therefore, there is now evidence of substantial technological innovation in the context of the process of colonisation of Sahul.

There is also a remarkable parallel, the first appearance in the Japan archipelago of ground-edge tools that coincided with the arrival of Homo sapiens in Japan about 38,000 BP (Takashi, 2012). The axes from the Pleistocene in Australia and Japan that are known of are distinctly differ distinctly in size and shape from each other, and represent separate technological innovations, and Hiscock et al. speculate that they both possibly built on grinding applications that were pre-existing such as the grinding of haematite for pigment or the production of bone tools. It is suggested that the timing of these innovations in 2 separate lands at the point of colonisation, that dispersing humans often innovated as they entered new territories, and not maintaining technologies that had been previously employed. Adjustments to provisioning and production systems that local materials were suited to, and the availability/costs of materials, as well as new economic and social systems in the new landscapes that were serviced by these novel technologies. It is suggested that something of the magnitude and structure of technological experimentation and innovation in Australia can be illuminated by describing the growth of regional diversity in the production of waisted axes and ground-edge axes.

Technological diversity and regional traditions

According to Hiscock et al. geographic variation and regional traditions of behaviour are evident in the technology of the modern humans who were colonising Sahul. The use of hafted ground-edge axes in northern Australia and flaked, waisted unground axes in Papua New Guinea, though the complete lack of axes in the southern 2/3 of the Australian continent, epitomise this (Balme & O’Connor, 2014; Geneste et al., 2010; O’Connor, 1999; 18, Summerhayes et al., 2010). Around the time of colonisation these divisions originated and persisted through to the Holocene, by which time axes began to appear in the assemblages recovered from archaeological sites in most parts of southern mainland Australia and in New Guinea polished adzes are found in deposits. Hiscock et al. say these regional distinctions persisted for 40,000 years, which was presumably bolstered by distinctions in language and social views.

A new image of the dispersion of modern humans out-of-Africa is offered by the findings of the study that is reported in this paper. Flexible and novel adaptations were displayed by cultural groups who occupied new lands such as Sahul and Japan, which is revealed archaeologically in the invention of new technological strategies, such as hafted, ground-edge axes. The construction of cultural differences between groups in different regions and the cultural distinctions formed at colonisation in some instances, were extremely long lasting. It was concluded by Hiscock et al. that with the dispersal of H. sapiens dynamic adaptive modification of cultural systems occurred in conjunction with the dispersal, which played a significant role in the successful expansion of modern humans around the world, as well as leading to a long-lasting differentiation of human societies.


The antiquity of the production of ground-edge axes in Australia has been progressively pushed back, which is a reflection of the increasingly sensitive dating techniques, as well as the gradual increase of archaeological sample sizes. Hiscock et al. conclude, based on their discovery of the specimen at Carpenter’s Gap 1, which dated to 44,000-49,000 BP, that the production of ground-edge axes is broadly coincident with the colonisation of Australia by modern humans. It is suggested by Hiscock et al. that axe production was probably invented within Australia a short time after the first humans arrived in the continent, and they have noted 2 implications of this inference.

(1)   The emergence of novelty during the human global dispersal. With the continuing spread of humans, technology was not only losing the diversity it had evolved in Africa; it was also being transformed by invention of tools that were of entirely novel diversity.

(2)   The early invention of ground bevelled edges on axes from Northern Australia is a marker for regional behavioural distinctions that are long-lasting, demonstrating spatial differentiation in traditions and adaptive patterns beginning in the earliest period of exploration and settlement.

The technological elements of these regional distinctions persisted for 40,000 years, which indicates that these differences in technology were part of deep social and linguistic distinctions within Sahul.

Sources & Further reading

  1. Hiscock, P., S. O’Connor, J. Balme and T. Maloney (2016). "World’s earliest ground-edge axe production coincides with human colonisation of Australia." Australian Archaeology 82(1): 2-11.


Author: M. H. Monroe
Last updated: 18/05/2016
Journey Back Through Time
Experience Australia
Aboriginal Australia
National Parks
Photo Galleries
Site Map
                                                                                           Author: M.H.Monroe  Email:     Sources & Further reading