Geographical variation in Australian backed artefacts: trialling a new index of symmetry

NUMBER 79 | DECEMBER 2014

Australian Archaeology, the official publication of the Australian Archaeological Association Inc., is a refereed journal published since 1974. It accepts original articles in all fields of archaeology and other subjects relevant to archaeological research and practice in Australia and nearby areas. Contributions are accepted in eight sections: Articles (5000–8000 words), Short Reports (1000–3000), Obituaries (500–2000), Thesis Abstracts (200–500), Book Reviews (500–2000), Forum (5000), Comment (1000) and Backfill (which includes letters, conference details, announcements and other material of interest to members). Australian Archaeology is published twice a year, in June and December. Notes to Contributors are available at: <www.australianarchaeologicalassociation.com.au>.

Australian Archaeology is indexed in the Arts and Humanities, Social and Behavioural Sciences, and Social Sciences Citation Indices of the Thomson Reuters Web of Knowledge, SCOPUS, Australian Public Affairs Information Service (APAIS), and Anthropological Literature and Anthropological Index Online.

Australian Archaology is ranked as a tier A journal by the European Reference Index for the Humanities and French Agence d’Evaluation de la Recherche et de l’Enseignement Supérieur.

Subscriptions are available to individuals through membership of the Australian Archaeological Association Inc. or to organisations through institutional subscription. Subscription application/renewal forms are available at <www.australianarchaeologicalassociation.com.au>. Australian Archaeology is available through Informit and JSTOR.

Design and Print: Openbook Howden

Front Cover: Cailey Maclaurin and Samantha Aird examining a fish trip on Bentinck Island in the Gulf of Carpentaria (Annette Oertle, entered in the AAA2013 Photography Competition).

All correspondence and submissions should be addressed to:

Australian Archaeology

PO Box 10, Flinders University LPO

Flinders University SA 5048

Email: [email protected]

<http://www.australianarchaeologicalassociation.com.au>

The views expressed in this journal are not necessarily those of the Australian Archaeological Association Inc. or the Editors.

© Australian Archaeological Association Inc., 2014

ISSN 0312-2417

Editors

Heather Burke Flinders UniversityLynley Wallis Wallis Heritage Consulting

Editorial Advisory Board

Brit Asmussen Queensland MuseumVal Attenbrow Australian MuseumHuw Barton Leicester UniversityNoelene Cole James Cook UniversityPenny Crook La Trobe UniversityInes Domingo Sanz University of BarcelonaJudith Field University of New South WalesJoe Flatman University College LondonRichard Fullagar University of WollongongTracy Ireland University of CanberraJudith Littleton University of AucklandMarlize Lombard University of JohannesburgAlex Mackay University of Wollongong Scott L’Oste-Brown Central Queensland Cultural Heritage ManagementJo McDonald The University of Western AustraliaPatrick Moss The University of QueenslandTim Murray La Trobe UniversityJim O’Connell University of UtahSven Ouzman The University of Western AustraliaFiona Petchey University of WaikatoAmy Roberts Flinders UniversityKatherine Szabo University of WollongongNancy Tayles University of OtagoRobin Torrence Australian MuseumPeter Veth The University of Western AustraliaAlan Watchman Flinders UniversityDavid Whitley ASM Affiliates Inc.Nathan Woolford Nathan Woolford Consultants

Short Report Editor

Sean Winter The University of Western Australia

Book Review Editors

Alice Gorman Flinders UniversityClaire St George Ochre Imprints

Thesis Abstract Editor

Tiina Manne The University of Queensland

Editorial Assistant

Susan Arthure Flinders University

Commissioned Bloggers

Jacqueline Matthews The University of Western AustraliaCarly Monks The University of Western AustraliaMichelle Langley The Australian National UniversityJordan Ralph Wallis Heritage Consulting

45

December 2014, Volume 79

Editorial | Heather Burke and Lynley A. Wallis iii

Articles

Chronological trends in late Holocene shell mound construction across northern Australia: Insights from Albatross Bay, Cape York Peninsula | Michael Morrison 1

Earthenware of Anuru Bay: A reassessment of potsherds from a Macassan trepang processing site, Arnhem Land, Australia, and implications for Macassan trade and the trepang industry | Daryl Wesley, Tristen Jones, Sue O’Connor, Jack Fenner and William R. Dickinson 14

Transforming the inedible to the edible: An analysis of the nutritional returns from Aboriginal nut processing in Queensland’s Wet Tropics | Anna Tuechler, Åsa Ferrier and Richard Cosgrove 26

The central lowlands of the Hunter Valley, NSW: Why so few early sites have been found in this archaeologically-rich landscape | Philip Hughes, Nigel Spooner and Daniele Questiaux 34

A Kaurna burial, Salisbury, South Australia: Further evidence for complex late Holocene Aboriginal social systems in the Adelaide region | Timothy D. Owen and F. Donald Pate 45

The making of a radical archaeologist: The early years of Vere Gordon Childe | Robin Derricourt 54

Monitoring change at Aboriginal rock art sites | Natalie Franklin 65

Putting WA archaeology on the map: The inestimable contribution of Charlie Dortch

Guest edited by Sandra Bowdler, Jane Balme and Joe Dortch 77

Charlie Dortch: History and archaeology across three continents | Joe Dortch, Jane Balme and Sandra Bowdler, with Peter Randolph 78

Charlie Dortch | Wayne Webb 81

And a suggestion from one of our readers: A personal note | Sandra Bowdler 81

Charles E. Dortch Publication List 83

Both half right: Updating the evidence for dating first human arrivals in Sahul | Jim Allen and James F. O’Connell 86

Aboriginal landscape burning and its impact on the summer monsoon of northern Australia | Karl-Heinz Wyrwoll and Michael Notaro 109

What to make of the ‘Murchison Cement’? A re-examination of a megafaunal fossil site in the Mid West, Western Australia | Ashleigh Murszewski, Ingrid Ward and Matthias Leopold 116

Geographical variation in Australian backed artefacts: Trialling a new index of symmetry | Peter Hiscock 124

A Norfolk Island basalt adze from coastal New South Wales | Peter White, Christian Reepmeyer and Geoffrey Clark 131

Observations on edge-ground stone hatchets with hafting modifications in Western Australia | Kim Akerman 137

Backed points in the Kimberley: Revisiting the north-south division for backed artefact production in Australia | Tim Maloney and Sue O’Connor 146

Maritime deserts of the Australian northwest | Peter Veth, Kane Ditchfield and Fiona Hook 156

Table of Contents

65

01

26

14

i

December 2014, Volume 79

A tale of three caves: New dates for Pleistocene occupation in the inland Pilbara | Kate Morse, Richard Cameron and Wendy Reynen 167

Devils Lair: Occupation intensity and land-use | Jane Balme 179

Intergenerational archaeology: Exploring niche construction in southwest Australian zooarchaeology | Joe Dortch, Carly Monks, Wayne Webb and Jane Balme 187

Malimup: A Tasmanian Aboriginal Hoabinhian site in the southwest of Western Australia? | Sandra Bowdler 194

A mid- to late Holocene sequence from Weld Range, Mid West, Western Australia, in local, regional and inter-regional context | Vicky Winton, Viviene Brown, Jamie Twaddle, Ingrid Ward and Nicholas Taylor 203

Referees 216

Thesis Abstracts - Available online

The Occupation of Bakers Flat: A Study of Irishness and Power in Nineteenth Century South Australia | Susan Arthure

Vanished Value | William Doring

Health, Diet and Migration Prior to the Establishment of the Pre-Angkorian Civilisation of Southeast Asia | Jennifer Newton

World War II Conflict Aviation Archaeology: Managing World War II Aviation Sites in Australia and the Marshall Islands | Fiona Shanahan

The Evidence of the Dutch Occupation of the Western Australian Coast Following the Vergulde Draeck (1656) Shipwreck | Robert (Bob) Sheppard

Gimme Shelter: Archaeology and the Social History of Structural Defence in Adelaide, 1941–1943 | Martin Wimmer

Book Reviews - Available online

Love’s Obsession: The Lives and Archaeology of Jim and Eve Stewart, by Judy Powell | Andrew Sneddon

Excavations, Surveys and Heritage Management in Victoria Volume 1, edited by Ilya Berelov, Mark Eccleston and David Frankel | Pamela Ricardi

Backfill - Available online

Obituary: Stephen Mark Free (23 March 1966–9 May 2014)

Obituary: Herman Mandui (1969–2014)

156

131

203

167

116

ii

PUTTING WA ARCHAEOLOGY ON THE MAP:THE INESTIMABLE CONTRIBUTION OF CHARLIE DORTCH

THEMED SECTIONGUEST EDITORS: SANDRA BOWDLER, JANE BALME AND JOE DORTCH

Image: Charlie Dortch at Devils Lair, southwest Australia.

December 2014, Volume 79:124–130

TH

EM

ED

SE

CT

ION

Geographical variation in Australian backed artefacts:Trialling a new index of symmetry

Peter HiscockDepartment of Archaeology, School of Philosophical and Historical Inquiry, Faculty of Arts and Social Sciences, The University of Sydney,

Sydney NSW 2006, Australia <[email protected]>

Abstract

Geographical variation in backed artefact size and morphology has long been recognised in Australia. This paper evaluates a novel measure of symmetry that can quantify regional and continental-scale geographic patterns in backed artefact forms. The result indicates that we can construct new depictions of regional differences in Australian backed artefacts, and that evolutionary explorations of those differences are worthwhile.

In the century since the discovery of backed artefacts in Australia, archaeologists have explored their patterning through space and time. The vigorously debated question of backed artefact chronology appears to have finally been resolved in favour of a model of terminal Pleistocene emergence of distinct backed artefacts followed by a proliferation in their production during the late Holocene (Hiscock 2008; Hiscock and Attenbrow 1998, 2004; Slack et al. 2004). In contrast, the extent of regional variation in the timing of these trends, and the causes of temporal changes in production rates, continue to be investigated and disputed (e.g. Attenbrow et al. 2009; Bowdler 2011; Hiscock 1994, 2002; Hiscock et al. 2011; Kuhn 2011; McNiven 2011; Shott 2011; White 2011a, 2011b). Much has recently been revealed about the operation of backed artefacts as tools, with the demonstration that, along the east coast, they did not normally function as spears or public symbols without other uses; they typically functioned as craft tools for producing organic artefacts (Robertson 2002, 2005; Robertson et al. 2009). Redescriptions of prehistoric patterns of production and use, through both discoveries of new material and re-evaluation of previously published assemblages, have fundamentally redirected discussions about how and why backed artefact technology evolved in Australia. Models of warfare-intensive periods and migration/diffusion in the mid-Holocene have given way to models of adaptations in response to physical, economic and social factors.

One of the missing elements in these discussions is a depiction of the geographical variation in the morphology, function and antiquity of backed artefacts. Over the last couple of decades Australian archaeologists have focused on establishing chronological frameworks, but earlier studies had also been concerned with questions about the geographical distribution of backed artefacts. For instance, the geographical limits of backed artefacts and particularly the existence and location of a perceived northern limit to their distribution was extensively discussed (e.g. Flood

1995; Hiscock 2001; Hiscock and Hughes 1980; Mulvaney 1969, 1985; Pearce 1974; Smith and Cundy 1985). Another geographic question that was much discussed in the twentieth century emerged from early typological perceptions of morphological differences in the size, shape and symmetry of backed artefacts across the country. For instance, one widely discussed proposition was that asymmetrical and symmetrical varieties could be distinguished and that they show, disparate geographic tendencies in the archaeological record (e.g. Mitchell 1949:49). One of the key commentators, Mulvaney (1969:124), described the Lake Eyre Basin and southeastern Australia as the distributional focus of symmetrical backed artefacts, while the east coast was a focus of asymmetrical forms, saying that ‘While Bondaian types also occur in South Australia [SA] and Victoria, there was an extraordinary concentration on coastal sites in southern New South Wales [NSW]’. Mulvaney (1985:212) developed a model in which asymmetrical backed artefacts dominated collections on the west and east coasts, while symmetrical forms dominated collections in Victoria and SA. Any such continental-scale variation would demand explanation and would evoke a discussion as elaborate as that which developed in the search for an explanation of chronological change in backed artefacts. Although Mulvaney’s model has been virtually ignored in discussions over the past two decades, it remains a potent, albeit untested, image of the configuration of Australian assemblages. This paper tests the model, through the use of an original quantitative measure of symmetry, and comments on the possibility of developing evolutionary models of Australian backed artefacts.

Typological Measurements of Backed Artefact Symmetry

Australian typologists have long recognised that the size and shape of backed artefacts may have varied geographically across the continent. Measurement of this variation was one reason for the proliferation of classes in early typological

124

December 2014, Volume 79:124–130

TH

EM

ED

SE

CT

ION

Peter Hiscock

schemes, although archaeologists progressively abandoned finer typological distinctions in a search for the broad patterns of backed artefact distribution. For several decades the differentiation of only two classes of backed artefact (or blade) has been common, as summarised by Pearce (1974:300):

Most Australian workers recognize [sic] two major typological categories of backed blades, geometric and Bondi. The geometric type covers various angular shapes such as triangle, trapeze and roughly symmetrical segments and crescents, often trimmed near the bulbar end as well as the distal end. The Bondi type is trimmed to a point at one end, and is asymmetrical about a plane normal to the longest dimension of the artefact.

This widely adopted convention of classifying backed artefacts into geometric and Bondi types conflates multiple morphological characteristics. As the above quote reveals, these types imply a correspondence of plan shape and symmetry: the ‘geometric’ form is said to comprise plan shapes, such as triangles, trapezes and crescents, all of which are implied to be symmetrical, while the unipointed Bondi is implied to be asymmetrical. Constructing these types from multiple characteristics was a common procedure employed by Mitchell (1949:49), McCarthy et al. (1946) and many subsequent researchers from the early twentieth century.

In the 1960s relative specimen length was explored by Glover (1967) as a quantitative expression of those morphological differences, based on measurements of length and width. He proposed a two-fold division of backed artefacts into classes labelled Groups A and B. Group A consisted of specimens with a length:width ratio less than 2:1, and Glover (1967:419) commented that such backed artefacts would ‘all fall into what is often called a class of geometric microliths and they are variously shaped crescentic, triangular and trapezoid’. Group B consisted of specimens with a length:width ratio more than 2:1, and included what Glover described as ‘asymmetrical points’. Although Glover (1967) expressed doubts about the value of distinguishing geometric from other forms, it is clear from his discussion that he saw a broad equivalence between Group A and geometric backed artefacts, and Group B and the Bondi forms. This conjunction is also clear in the works of other researchers who employed terms interchangeably (e.g. Mulvaney 1985:212).

These typological approaches are based on presumptions that, in Australian assemblages, particular combinations of characters exist, and that there were limited and repeated permutations of trait combinations. Using those conventions researchers have often discussed the category labelled geometric as being relatively short, bilaterally symmetrical around the chord, and with a plan shape that is triangular, trapezoidal or crescentic; while the Bondi class is described as having none of those features, being elongated, pointed at one end and asymmetrical. Those particular combinations seem simple and convenient, but there is no logical reason that these characteristics should be combined in only these ways. Figure 1 illustrates symmetrical and asymmetrical variants of each of the plan shapes often discussed in the Australian literature; there is no theoretical or mechanical reason for shapes such as triangles or crescents to be symmetrical. In a similar way there is no theoretical reason for symmetrical specimens, including triangles and crescents, to be shorter than all asymmetrical specimens, as implied in Glover’s A/B distinction. If such patterns exist they need to be empirically established, but suitably framed analyses have not yet been undertaken.

Presumed co-variation of traits, such as symmetry, plan shape, elongation and thickness, have underpinned divisions between symmetrical and asymmetrical backed artefacts, but those classifications assume rather than test the nature of trait combinations. When constructing categories such as ‘geometrics’ or ‘Bondis’ archaeologists have selected different traits or given different weights to each trait in classificatory process, leading to highly variable practices in classifying backed artefact symmetry.

In the absence of empirical evidence that shape, elongation and symmetry are highly correlated in Australian backed artefacts, and in view of the high level of interobserver variability demonstrable in subjective, unquantified, classifications of backed artefact symmetry, it is more productive to measure these traits independently of each other. This approach will help to establish the association of these traits, while giving unambiguous distributional data on morphological variation in backed artefacts.

Backed Artefact Symmetry Index (BASI)

In developing a quantitative measure of backed artefact symmetry it is not necessary to seek to measure a comprehensive record of plan shape. Modern technologies of 3D scanning are capable of making an overall record of size and morphology, but in fieldwork conditions and for testing a specific question, such as what level of symmetry is found on specimens, it is likely that dedicated, simple measures will be more cost-efficient and productive. The index presented here is designed to express only the degree of symmetry in the position of the widest point of a backed artefact, irrespective of other characteristics of specimen size, shape or elongation and the index therefore represents a very specific feature of symmetry. Consequently the ‘Backed Artefact Symmetry Index’ (BASI) measures one element of bilateral symmetry, namely how closely the widest part of the specimen corresponds to the mid-point of the chord. This index is designed to be calculated with a few simple variables, and involves using calipers to make only the three measurements illustrated in Figure 2 and defined as follows:

• CHORD length = straight line distance between the two end points of the chord, representing the location

Figure 1 Illustration of both symmetrical and asymmetrical variants for backed artefact shape categories.

125

December 2014, Volume 79:124–130

TH

EM

ED

SE

CT

ION

Geographical variation in Australian backed artefacts: Trialling a new index of symmetry

at which the chord meets with a distinct retouched or non-retouched margin. On some specimens the chord is curved and in these instances the chord length may not lie precisely parallel to the unretouched chord margin;

• MAXIMUM width = the greatest dimension from the chord to the opposite margin, at right angles to the chord length. Note the measurement is orientated relative to the chord length, but is measured from the chord itself rather than the line defined as chord length; and,

• LENGTH to maximum width (LMW) = distance from one end of the chord, along the line of the chord, to the point where maximum width was taken (or where the line defined by maximum width would intersect with the chord length). On any specimen two such measurements can be made: one from either end of the chord. The measurement that represents the length to maximum width is the smaller of the two possible measurements, irrespective of whether that measurement represents the distal or proximal portion of the specimen.

From these three variables BASI is calculated as:

• BASI = Length to maximum width / (0.5 * Chord length)

This index yields values that range between 0 and 1. BASI values of 0 occur when the maximum width is at the end of the chord, and values close to 0 indicate that the maximum width of the specimen is located close to one end of the chord. In contrast, BASI values of 1 reveal that the maximum width of the specimen is exactly mid-way along the chord, and values close to 1 indicate that the maximum width is located near the centre of the chord length. Consequently asymmetrical specimens will have lower BASI values, while symmetrical specimens will have higher values.

Accuracy and Error in BASI Measurements

The accuracy of BASI measurements is influenced by at least two classificatory activities that lithic analysts are obliged to undertake. The first is the systematic and correct identification of backed artefacts. In part this is an issue of the consistency with which analysts define and differentiate implement types, but the dominant concern is not the definition of backed

artefacts as a class of implement, it is the regularity with which typologists in the past have mistaken unretouched fragments for backed artefacts. In museum collections across Australia error rates of experienced researchers in identifying backed artefacts are commonly in the range of 10–40% of specimens. Identification errors typically consist of over-identification, as unretouched specimens are mistakenly classed as backed artefacts. There are many reasons that unretouched flakes in particular are misclassified as backed artefacts, including the confusion of dorsal platforms, platform facetting, overhang removal and heat shattering, with backing retouch. This observation will be explored further on another occasion, but a case study of such misidentifications, at an error rate of 33%, has previously been presented for the Sandy Hollow site in NSW (Hiscock 1993:71–72). In this context the implication of high identification error rates is the potentially significant contaminating effects on the average BASI values calculated for an assemblage. Obviously the reliability of BASI values is in part an expression of the accuracy of classifications created by analysts.

This principle of stringent selection of samples extends to the treatment of broken specimens. While BASI can be measured accurately and repeatedly on complete specimens, it will generally be unreliable if applied to broken specimens. This is because backed artefacts frequently break transversely, truncating chord length and therefore affecting the calculated BASI value. Transverse breakage typically transforms asymmetrical backed artefacts into more symmetrically shaped specimens, while broken symmetrical specimens can appear distinctly asymmetrical. It appears that discussions of backed artefact symmetry have often proceeded without explicitly considering the completeness of specimens, or the danger that, when dealing with fragments, the perceived level of symmetry in an assemblage may have been measuring the level of fragmentation rather than the pattern of backed artefact symmetry. BASI should be calculated only on complete specimens to avoid this hazard. However, amongst the complexity that emerges from excluding broken specimens in lithic analysis there is always the possibility that broken specimens are systematically different to complete ones, because the difference somehow increases the likelihood of breakage. In this particular case it might be that a variable more associated with symmetrical specimens, such as thickness or cross-sectional shape, alters their breakage rate relative to asymmetrical specimens. This possibility can be evaluated by comparing traits, such as thickness at maximum width, on both complete and broken specimens where the measurements can be reliably made.

One intrinsic complexity of the BASI measure should also be explained. Because the BASI measurement, by definition, can be taken at only one point on each specimen, it may be that, on some specimens, the measured point appears atypical of the overall plan shape. This will probably occur on a few specimens within a large assemblage of backed artefacts, but is more likely to happen on asymmetrical than symmetrical specimens. On asymmetrical specimens it is possible to have the widest portion of the artefact centrally located, while the two ends of the specimen have very different plan shapes, and for the specimen to therefore be perceived as asymmetrical while having a high BASI value. The distally retouched form traditionally labelled as a ‘Woakwine’ is an obvious example of a specimen in which this combination of features might be found. Atypical BASI values of this kind are less likely to be observed on complete specimens conventionally classified

Figure 2 The three measurements required to calculate BASI: length, width and length to maximum width (LMW). A illustrates an asymmetrical specimen, while B illustrates a symmetrical one.

126

December 2014, Volume 79:124–130

TH

EM

ED

SE

CT

ION

Peter Hiscock

as symmetrical, for which a maximum width close to one end of the object, giving a low BASI value, is definitionally improbable. As a consequence it can be predicted that specimens conventionally classified as symmetrical will probably display a smaller dispersion of BASI values than those classified as asymmetrical. This difference in the operation of the index for the two typological categories does not constitute a problem for the use of the index; it simply reminds us that BASI is specific in measuring one aspect of backed artefact symmetry and should not be thought of as measuring all dimensions of symmetry. Since complete, finished, asymmetrical specimens with high BASI values are likely to be rare, I expect they will normally have a minimal effect on robust assemblage statistics.

Obviously the effects of atypical values might be diminished by procedures such as measuring the second widest point, in the hope this would be more typical of the overall symmetry. Such a ‘Secondary BASI’ may offer a smaller range of values in asymmetrical specimens but it is operationally far more difficult to apply, requiring elaborate judgments and multiple measurements. With maximum width it is simple and quick to estimate the widest point place the calipers at that point and then without adjusting the instrument move it left and right along the length to confirm that no other place is further from the chord and hence that this is the widest point. Measuring the second widest point (as opposed to the third or fourth widest) is more time consuming, and in any case would not ensure that an atypical value was still not measured. Anyone using a secondary width procedure should specify the practice, but I have chosen not to do this and to measure only the maximum width to illustrate the analytical value of BASI.

In any case, the positive qualities of an index such as this should be stressed. Note that percentages of broad typological categories, such as symmetrical or asymmetrical backed artefacts, are relatively imprecise measures of symmetry. The level of interobserver difference in the classification of Australian backed artefacts has not been adequately quantified, but, in light of error rates in identification, typological systems have been highly variable in their implementation. Furthermore, dichotomous categories such as symmetrical fail to discriminate between specimens that are strongly symmetrical, moderately symmetrical or weakly symmetrical. Consequently the difference between a collection of backed artefacts that are dominated by weakly asymmetrical specimens and one dominated by weakly symmetrical specimens may be too subtle to be adequately characterised by percentages of classes in conventional classifications. It is in contexts where a sensitive and consistent measure of symmetry is desired that a continuous metrical index such as BASI will be valuable. In addition, the BASI should reduce interobserver variation because its components are metrical and specifically defined. Hence, the BASI should be a measure of symmetry that is both sensitive and repeatable.

Employing BASI to Measure Spatial Patterning in Australia

Equipped with the BASI it is possible to return to the question of geographical variation in backed artefact symmetry with which I began the paper. Figure 3 presents a map of geographic changes in mean BASI values for assemblages

of backed artefacts across Australia. This map has been constructed as follows:

• Only data from complete backed artefacts have been used, for reasons explained above. Definitions of backed artefacts as retouched flakes with steep, bipolar or anvil-rested flaking were strictly maintained as per my previous usage (Hiscock 1993, 2002, 2006; Hiscock and Attenbrow 1996).

• Only sites with assemblages of 10 or more complete specimens were incorporated in the analysis to reduce the impact of atypical specimens.

• Assemblages were only employed when I was confident the collection was either complete, representative or minimally subject to collector bias. This meant that many of the large amateur collections held in museums were rejected for the purposes of this analysis.

• For this initial publication I measured only those specimens that would have been classed as backed artefacts by several generations of archaeologists. In particular this meant excluding the ‘Juan knives’ meticulously recorded by Lamb (1996, 2005, 2011) on the northeast coastline, even though I am convinced they qualify as backed artefacts, simply to avoid unnecessary critiques. Nonetheless, Lamb’s large backed artefacts are mostly asymmetrical and are unlikely to alter northeastern isopleths radically. For information the maximum backed artefact length in these data is less than 60 mm and the thickest specimens are less than 14 mm. Furthermore, I have applied no other value judgments as to which individual specimens should be measured. For instance, I measured unbroken Woakwine points from the deserts, even though they typically have low BASI values and we have reason to think that most, if not all, are unfinished triangular-shaped symmetrical backed artefacts. But to exclude them as unfinished would be injudicious and unnecessary. The main effect of including Woakwine specimens would be to reduce the Lake Eyre BASI values, which are already very high; removing all Woakwine points would exaggerate rather than diminish the published pattern and would add a selection bias.

Figure 3 Preliminary map of the geographical variation in backed artefact symmetry in mainland Australia, displayed as isopleths of average BASI as per text.

127

December 2014, Volume 79:124–130

TH

EM

ED

SE

CT

ION

Geographical variation in Australian backed artefacts: Trialling a new index of symmetry

• Mean BASI values were calculated for each site, without regard for specimen raw material or the chronological differences between specimens within the site. I acknowledge that, by disregarding these variables, a partial analysis is obtained, but as this is a first approximation and the research is ongoing it is deemed appropriate. For readers troubled by this an ANOVA of BASI accounting for raw material produced a non-significant relationship (F= 0.954, d.f. = 8, p = 0.471). These variables will be explored in further studies.

• Isopleths of mean BASI values, at 0.05 intervals, were constructed using a simple linear estimation method.

• Isopleths are drawn only for those portions of the continent with adequate data (see above).

The map constructed in this fashion should be considered a preliminary record of BASI variation for several reasons. This sample was collected to allow proof of concept, not as a final batch for defining geographical variation, and as a consequence there may be some concerns about its size. The dataset currently comprises approximately 2000 specimens, complete and fragmented, a sample that will no doubt polarise readers. To put this sample in perspective this sample is not small: it is the same order of magnitude as that used for continental-scale studies of lithic variation elsewhere in the world (e.g. Buchanan et al. 2014; O’Brien et al. 2014). For that reason it is a reasonable sample with which to explore the opportunities offered by quantitative analysis of backed artefact form. However, the sample nonetheless comprises a limited number of observations with which to study fine-grained patterns at a continental scale (Table 1). The sites are not evenly dispersed and it is likely that the collection of further data from some areas, particularly from Victoria and Western Australia (WA), may alter the locations of some isopleths discussed below. Furthermore, the map assembled in this fashion is atemporal, and does not reveal any shifts in backed artefact BASI values through time.

Samples included in this study are drawn from many of the well-known sites that have been published and used in previous studies of backed artefact variation (including Bondi Beach, Kurnell, Curracurrang 1, Capertee 3, Kenniff Cave, James Range East, Santa Teresa, Puritjarra, Puntutjarpa, South Bullsbrook, Walyunga and Northcliffe 2), as well as from many other sites available in Australian museums and private collections (Table 1). Those collections were supplemented by field measurements taken by myself over the last two decades, such as at Lawn Hill, in desert SA, the Moreton region and the Sydney Basin. With these cautions explicit I argue that this evidence yields a preliminary, yet useful, expression of the geographic variation in backed artefact symmetry.

The resulting map yields a remarkably clear image of geographical variation in this one trait of backed artefact form. Perhaps the most noteworthy point about the map is the highly patterned nature of assemblage variation that is revealed. The variation across the landscape appears to be structured coherently and without obvious exceptions. This confirms the observations of both amateur and professional typologists and collectors: that variation in the composition of backed artefact assemblages is geographically structured at a broad scale.

The BASI map demonstrates a focus on production of very symmetrical backed artefacts in the interior, and especially in

SiteMean+Standard

DeviationNumber

(All Specimens)

6 Mile Waterhole 0.763+0.194 16

8 Mile Creek 0.875+0.087 13

Altona Bay 0.773+0.173 10

Arumvale 0.760+0.093 10

Barrack Point 0.398+0.137 10

Bellambi 0.557+0.135 11

BGW97 0.611+0.202 17

Bondi Beach 0.551+0.199 57

Cape le Grand 0.737+0.171 61

Capertee 3 0.536+0.219 207

Cheetup region 0.541+0.118 10

Curracurrang 1 0.632+0.212 231

Finniss Springs 0.785+0.079 10

Fromms Landing 0.823+0.207 15

HD3 0.841+0.140 135

HV1 0.555+0.226 35

James Range East 0.856+0.154 49

Kenniff Cave 0.834+0.145 21

Kurnell 0.511+0.147 49

Lake Dumbleyung 0.637+0.214 10

Lake Jasper 0.855+0.085 10

Maidstone 0.901+0.046 10

MG Open 0.784+0.170 59

Monger’s Lake 0.829+0.114 10

Mucklandama Creek 0.663+0.199 10

North Pool 0.898+0.031 10

Northcliffe 0.785+0.140 66

Ooramina WH 0.826+0.253 11

Phillip Pond 0.874+0.094 106

Portion 5-6 0.858+0.139 47

Puntutjarpa 0.797+0.143 45

Puritjarra 0.737+0.204 25

Roxby Downs Station 0.873+0.125 76

Santa Teresa 0.861+0.180 80

South Bullsbrook 0.628+0.195 95

Split Rock 0.609+0.147 10

The Pines 0.864+0.099 62

Walyunga 0.718+0.181 103

Wiluna 0.861+0.083 10

Windang Beach 0.698+0.232 10

Woodside 0.708+0.225 55

Yarrimie 0.640+0.272 19

Table 1 Mean BASI values for complete specimens from key sites in the analysis.

128

December 2014, Volume 79:124–130

TH

EM

ED

SE

CT

ION

Peter Hiscock

dry, desert areas. The most symmetrical assemblages occur across the Lake Eyre Basin and adjacent lands. Production of more asymmetrical backed artefacts was frequent closer to the coast, with assemblages of highly asymmetrical specimens on the eastern seaboard. This pattern reveals the value of the BASI measurement and encourages Australian analysts to consider, once again, the spatial patterns in such artefact classes.

Discussion

In the remarkable first edition of his synthesis of Australian prehistory Mulvaney (1969:123–126) developed a number of related statements about geographic variation in backed artefacts. His assertion that there was a northerly limit to their distribution has since been robustly demonstrated by Smith and Cundy (1985), although sampling issues have meant that the precise distribution remains elusive (Hiscock 2001). Mulvaney’s (1969:124, 1985:212) conclusion that geometric forms were concentrated in the Lake Eyre drainage basin and inland southeast, while the Bondi form was concentrated in coastal areas of southern NSW, is also partly borne out by the evidence provided in Figure 3, in so far as the BASI index reveals patterns that can be mapped on to a century old typological division. This preliminary map of geographical variation in backed artefact symmetry suggests there may be broadly coherent structure to the continental patterning.

Given this outcome it will be worth encouraging future analyses to explore geographical variation further. To date, late Pleistocene and early Holocene backed artefacts are most well documented in the east and northeast of Australia, and this category of retouched flake appears to be substantially younger in many interior arid and semi-arid landscapes. If backed artefacts in different regions have a common origin, these chronological differences most likely hint at dispersal of backed artefact production systems and/or associated composite technology, presumably with accompanying systems for transmitting these operations and of conceptualising them in varied social contexts. In this situation I hypothesise that at least some of the geographical variation in backed artefact morphology is related to that dispersal history, as modification of the morphological variation in backed artefacts arose from a range of mechanisms that might have included transmission error, co-opting and modification of backed artefact form as composite tools or tool use varied, novelty induced by the adjustment of backing processes to different blank production strategies, and divergence in the backed forms to facilitate social signalling in learning/production contexts (see Hiscock 2014). In North America cladistics analyses are increasingly being used to search for phylogenetic relationships and examine the divergence of forms from common ancestors, and such approaches are worth applying in the Australian context to spatial variation in backed artefacts (see Hiscock and O’Brien 2014).

Conclusion

Despite more than a century of study the evolutionary trends in prehistoric Australian lithic technologies remain little known and this is true of even well studied technological activities, such as the manufacture of backed artefacts. Once depicted as a simple progression through three distinct, yet geographically uniform, stages, we now recognise that

Australian technologies display substantial variation in response to multiple factors. The recognition of diversity and non-stadial change has perhaps drawn our attention to the complexity of local patterns and adaptations and away from the continental stage across which evolutionary dramas have been played out. In this paper I have trialled a novel yet simple measure to quantify continental-scale geographic patterns in the symmetry of backed artefacts. The BASI provides an additional characteristic with which to explore geographical variation in Australian backed artefacts and the historical contingency underlying technological evolution throughout the terminal Pleistocene and early Holocene.

Acknowledgements

Research was partially funded by grants from the Northern Territory University and The Australian National University. I thank the following institutions for allowing me to examine collections: Australian Museum, Queensland Museum, Northern Territory Museum, SA Museum and the WA Museum. I especially thank the following individuals for providing access to material in their care: Val Attenbrow, Peter Bindon, Leanne Brass, Stan Florek, Andrew Hughes, Boone Law, Fiona Mowat, Colin Pardoe, Michael Quinnell, Richard Robins, Moya Smith, Peter Thorley and Peter Veth. A special thanks to Charlie Dortch, who was invaluable during the period in which WA specimens were measured, and who always made my trips to Perth more enjoyable and productive. Charlie has been a significant force in Australian archaeology and a researcher of great vision.

References

Attenbrow, V., G. Robertson and P. Hiscock 2009 The changing abundance of backed artefacts in southeastern Australia: A response to Holocene climate change? Journal of Archaeological Science 36:2765–2770.

Bowdler, S. 1981 Hunters in the Highlands: Aboriginal adaptations in the eastern Australian uplands. Archaeology in Oceania 16:99–111.

Bowdler, S. 2011 The dead hand of Jeremy Bentham. Australian Archaeology 72:69–70.

Buchanan, B., M.J. O’Brien and M. Collard 2014 Continent-wide or region-specific? A geometric morphometrics-based assessment of variation in Clovis point shape. Archaeological and Anthropological Sciences 6:145–162.

Flood, J. 1995 Archaeology of the Dreamtime. Sydney: Angus and Robertson.

Glover, I.C. 1967 Stone implements from Millstream Station, Western Australia: Newall’s Collection re-analysed. Mankind 6:415–425.

Hiscock, P. 1993 Bondaian technology in the Hunter Valley, New South Wales. Archaeology in Oceania 28:64–75.

Hiscock, P. 1994 Technological responses to risk in Holocene Australia. Journal of World Prehistory 8:267–292.

Hiscock, P. 2001 Sizing up prehistory: Sample size and composition of artefact assemblages. Australian Aboriginal Studies 2001(1):48–62.

Hiscock, P. 2002 Pattern and context in the Holocene proliferation of backed artefacts in Australia. In S. Kuhn and R. Elston (eds), Thinking Small: Global Perspectives on Microlithisation, pp.163–177. Anthropological Papers of the American Anthropological Association (AP3A) 12.

Hiscock, P. 2006 Blunt and to the point: Changing technological strategies in Holocene Australia. In I. Lilley (ed.), Archaeology in Oceania: Australia and the Pacific Islands, pp.69–95. Malden: Blackwell.

129

December 2014, Volume 79:124–130

TH

EM

ED

SE

CT

ION

Geographical variation in Australian backed artefacts: Trialling a new index of symmetry

Hiscock, P. 2008 Archaeology of Ancient Australia. New York: Routledge.

Hiscock, P. 2014 Learning in lithic landscapes: A reconsideration of the hominid ‘toolmaking’ niche. Biological Theory 9:27–41.

Hiscock, P. and M.J. O’Brien 2014 Evolutionary images of Australian microlithic industries. Unpublished paper presented at the XVII World UISPP Congress, Burgos.

Hiscock, P. and P.J. Hughes 1980 Backed blades in northern Australia: Evidence from northwest Queensland. Australian Archaeology 10:86–95.

Hiscock, P. and V. Attenbrow 1996 Backed into a corner. Australian Archaeology 42:64–65.

Hiscock, P. and V. Attenbrow 1998 Early Holocene backed artefacts from Australia. Archaeology in Oceania 33:49–63.

Hiscock, P. and V. Attenbrow 2004 A revised sequence of backed artefact production at Capertee 3, New South Wales. Archaeology in Oceania 39:94–99.

Hiscock, P. and V. Attenbrow 2005 Australia’s Eastern Regional Sequence Revisited: Technology and Change at Capertee 3. BAR International Series 1397. Oxford: British Archaeopress.

Hiscock, P., C. Clarkson and A. Mackay 2011 Big debates over little tools: Ongoing disputes over microliths on three continents. World Archaeology 43:653–664.

Kuhn, S.L. 2011 Are social and operational explanations incompatible? Australian Archaeology 72:70–71.

Lamb, L. 1996 A methodology for the analysis of backed artefact production on the South Molle Island Quarry, Whitsunday Islands. In S. Ulm, I. Lilley and A. Ross (eds), Australian Archaeology ‘95: Proceedings of the 1995 Australian Archaeological Association Annual Conference, pp.151–159. Tempus 6. St Lucia: Anthropology Museum, The University of Queensland.

Lamb, L. 2005 Rock of Ages: Use of the South Molle Island Quarry, Whitsunday Islands, and the Implications for Holocene Technological Change in Australia. Unpublished PhD thesis, Department of Archaeology and Anthropology, The Australian National University, Canberra.

Lamb, L 2011 Rock of Ages. South Molle Island Quarry, Whitsunday Islands: Use and Distribution of Stone through Space and Time. Oxford: Archaeopress.

McCarthy, F.D., E. Brammell and H.V.V. Noone 1946 The stone implements of Australia. Memoirs of the Australian Museum 9:1–94.

McNiven, I.J. 2011 Backed artefacts as a material dimension of social inclusiveness. Australian Archaeology 72:71–72.

Mitchell, S.R. 1949 Stone Age Craftsmen: Stone Tools and Camping Places of the Australian Aborigines. Melbourne: Tait Book Company.

Mulvaney, D.J. 1969 The Prehistory of Australia. London: Thames and Hudson.

Mulvaney, D.J. 1985 Australian backed blade industries in perspective. In V.N. Misra and P.Bellwood (eds), Recent Advances in Indo-Pacific Prehistory, pp.211–217. New Delhi: Oxford and IBH Publishing.

Mulvaney, J. and J. Kamminga 1999 Prehistory of Australia. Sydney: Allen and Unwin.

O’Brien, M.J., M.T. Boulanger, B. Buchanan, M. Collard, R.L. Lyman and J. Darwent 2014 Innovation and cultural transmission in the American Paleolithic: Phylogenetic analysis of eastern Paleoindian projectile-point classes. Journal of Anthropological Archaeology 34:100–119.

Pearce, R.H. 1974 Spatial and temporal distribution of Australian Backed blades. Mankind 9:300–309.

Robertson, G. 2002 Birds of a feather stick: Microscopic feather residues on stone artefacts from Deep Creek Shelter, New South Wales. Tempus 7:175–182.

Robertson, G. 2005 Backed Artefact Use in Eastern Australia: A Residue and Use-Wear Analysis. Unpublished PhD thesis, School of Social Science, The University of Queensland, St Lucia.

Robertson, G., V. Attenbrow and P. Hiscock 2009 Multiple uses for Australian backed artefacts. Antiquity 83:296–308.

Shott, M.J. 2011 On social and functional process in backed tools. Australian Archaeology 72:72.

Slack, M., R. Fullagar, J. Field and A. Border 2004 New Pleistocene ages for backed artefact technology in Australia. Archaeology in Oceania 39:131–137.

Smith, M.A. and B.J. Cundy 1985 Distribution maps for flaked stone points and backed blades in the Northern Territory. Australian Aboriginal Studies 1985(2):32–37.

White, J.P. 2011a Backed artefacts: Useful socially and operationally. Australian Archaeology 72:67–69.

White, J.P. 2011b Backing(:) The future. Australian Archaeology 72:73–75.

130