|
Article Excerpt
Abstract
Analysis of the stone artefact assemblage excavated from a stratified midden in a sandstone rockshelter at Balmoral Beach in Middle Harbour, Sydney, has revealed various strategies were adopted in manufacturing asymmetric backed artefacts, also known as Bondi points. The irregular morphology of many Bondi points and the small size of most in this assemblage suggest a need to economize and improvise, which we propose was due to the relatively limited availability of suitable stone materials in coastal Sydney. It was not only a question of distance to source and access but of abundance--there are few sources of suitable stone close to Balmoral Beach. The documented technological organization shows that people at Balmoral Beach were creatively backing flakes in several different ways to produce Bondi points with standardized width and thickness--length of the complete Bondi points was seemingly not as important. This observation, combined with the evidence that some broken Bondi points were re-worked, is discussed in the context of reliability and maintainability, design principles that may have been adopted for their manufacture.
Keywords: Balmoral Beach, Sydney, hacked artefacts, manufacturing, reliability, maintainability
**********
Excavations in an Aboriginal shell midden at Balmoral Beach in Middle Harbour, Sydney, revealed an abundant flaked stone assemblage in which backed artefacts were a dominant tool type. Much archaeological debate in Australia has focused on backed artefacts--particularly where and when they were first made, their naming, characteristics and distribution (e.g. Wieneke and White 1973; Pearce 1977; Hiscock and Attenbrow 1998, 2004; Mulvaney and Kamminga 1999:234-35; Hiscock 2002). Other studies have described general manufacturing technology (e.g. Flenniken and White 1985) or more detailed strategies in stone rich areas such as the Hunter Valley (Hiscock 1986, 1993; Moore 2000). In addition, the concept that the late Holocene proliferation of backed artefacts was associated with reducing 'risk'--when colonizing new environments or at times of climatic/environmental change--has been examined (Hiscock 1994; Attenbrow 2004:238-41).
Our study investigates whether the manufacture of backed artefacts at Balmoral Beach was influenced by the choice of particular design principles. The study was prompted by the initial examination of the backed artefacts here which indicated that many had irregular morphologies and that on average they were relatively small in size compared to other assemblages of backed artefacts in the Sydney region.
The geographical distribution, quality, accessibility and abundance of stone is said to have influenced the way people organized themselves in a particular environment (Bamforth 1992:131-33), and there was thus a fundamental link between the availability of stone and the way it was worked. If so, this should be apparent in stone artefact assemblages in coastal Sydney, which could be described as a relatively stone-poor environment (see below). (In this study, we define the coastal zone as that part of the country that includes the ocean and estuarine shorelines; this includes land that extends back from the coastline for a distance of ~30 km.) The influence of raw material availability may be evident in the strategies and techniques employed, as well as the adaptability and creativity seen in the use of the stone--in whatever form it was available--to make backed artefacts, in particular Bondi points. In this paper, we explore the technological strategies behind the manufacture of backed artefacts, as well as the design principles (also referred to as design variables and design criteria) that may have influenced their morphology, and the known sources of raw materials used for making flaked stone tools at Balmoral Beach 2. In the analysis, we focus on establishing how cores were worked and how backed artefacts were manufactured. Before doing so, however, we describe the site and the flaked stone assemblage.
Background to the site and stone artefact assemblage
The sandstone rockshelter at the southern end of Balmoral Beach measures 6.8 m by 3.5 m with a height of 3 m, faces north-east--directly towards the beach and to the mouth of Port Jackson (Fig. 1). In the 1960s, deposit to a depth of about one metre was removed from the top of the midden, both inside and outside the rockshelter, during roadworks and landscaping. Despite this action, initial excavations for site management purposes in 1992 revealed the existence of undisturbed rich deposits extending below the current ground level to a depth of some 2 m (Attenbrow 1992a, 1992b). It was therefore decided to continue excavating the site as part of the Port Jackson Archaeological Project, during which spatial and temporal variations in land and resource use in the Port Jackson catchment were being investigated (Attenbrow 1991).
[FIGURE 1 OMITTED]
In 1992 and 1993, several squares totalling 4.5 [m.sup.2] were excavated in different parts of the midden both inside and outside the rockshelter (Fig. 2) (Attenbrow 1993, 1994:6-20; Hashimoto 1993). The well-stratified deposits were excavated within a framework of 50 x 50 cm squares, which were named according to an alpha-numeric grid system. Only Squares M5 and M6, inside the shelter, were excavated until sandstone bedrock was reached at a depth of almost 2 m below present ground level (Fig. 3). The stone artefact assemblage discussed in this paper comes from one of these squares, M6. This square was selected as it is the only one excavated to bedrock that has had the excavated materials fully sorted. The deposits in this area of the site were divided into four layers: Layer 1 (12-18 cm deep) disturbed re-deposited sediments, overlaying the undisturbed midden; Layer 2, a thick shelly layer extending to a maximum depth of 65 cm below surface; Layer 3, consisting of well-stratified, shell-free, compacted, stone artefact rich sediments, extends to a depth of ~130 cm bs; and Layer 4, sandy mottled sediments, to bedrock at ~200 cm below present ground level. The backed artefacts discussed in this paper came principally from Layers 2 and 3, with only four coming from Layer 1.
Seven radiocarbon determinations are available from this area of the site. These radiocarbon ages, particularly Beta 55984 and Beta 60308 on shell from the top of the in situ deposit with conventional radiocarbon ages of 3000 [+ or -] 80 bp and 2960 [+ or -] 60 bp respectively, and Beta 58864 and Beta 60307 on charcoal from the basal layer of the deposit with ages of 3780 [+ or -] 140 bp and 3530 [+ or -] 150 bp, indicate that the undisturbed deposits of Layers 2, 3 and 4 in squares M4 to M8 accumulated between 3390-4500 cal bp and 2450-3050 cal bp. (Ages were calculated using Calib 5.0.1: shell samples [DELTA]R=3 [+ or -] 69, Ulm 2006; charcoal samples using atmospheric Southern Hemisphere option; calibration datasets: Hughen et al. 2004; McCormac et al. 2004). The 1960s roadworks bad removed some 2500-3000 years of accumulated cultural deposits from this part of the shelter.
[FIGURE 2 OMITTED]
[FIGURE 3 OMITTED]
Characteristics of the flaked stone assemblage
Square BB2/M6 yielded 2747 flaked artefacts and seven artefacts with ground surfaces. There were also many heat-shattered pieces, which included materials from which the flaked stone artefacts were made as well as the local sandstone. In addition, there were ambiguous 'fractured pieces', which may have been flaked or been naturally fractured, but were of non-local materials. During the analysis a series of nominal and ordinal characteristics were recorded for the flaked stone assemblage--the cores, retouched flakes and flaking debris--and it is the results of analysing these characteristics that are discussed below.
Silcrete predominates with 66% of the 2754 artefacts made from this material (Table 1). The next most abundant materials are IMTC with 13% and FGS 9% (Table 1 caption explains abbreviations). All other materials form only very small proportions of the assemblage, each being <5%. By weight (i.e. reflecting the amount of material brought into the site), silcrete still predominates but is a much lower percentage of the assemblage (48%), principally because there are a large number of very small flakes of this material. The igneous artefacts represent a much higher percentage by weight (21%) than by number (4%) as, although they are relatively few in number, they are larger on average than artefacts of other materials.
The flaked stone artefacts include 21 cores, 1311 complete flakes, 1257 broken flakes, and 158 undiagnostic flaked pieces (Table 2). A large proportion of the assemblage consists of unmodified complete flakes (46%, n=1261). The modified flaked component, i.e. artefacts with retouch and/or use-wear, is relatively small--125 artefacts (4.6%), which includes 114 retouched flakes, as well as ten unretouched flakes and one core with usewear (Tables 2 and 3). The 114 retouched flakes include 75 backed artefacts, 37 scrapers, and four burins (including one elouera and one scraper with burin scars) (Table 3). The cores and retouched flakes, in particular the backed artefacts, are the focus of this paper.
The 1257 broken flakes were divided into four classes (Table 2). Only a small percentage of the broken flakes in the assemblage had retouch or use-wear (n=74, 5.9%), yet they represent more than half the modified flakes (59.7%). This feature of the assemblage suggests that although some broken flakes were selected for retouching or use as tools, flakes were also broken, for example, (a) during detachment from the core; (b) while being retouched; (c) during use; (d) deliberately prior to or after retouch; (e) broken through treadage; (f) a combination of the above. Some of these scenarios are explored below by examining how cores were reduced and the characteristics of retouched artefacts. However, further examination of the broken margins and conjoin analysis is required to identify more clearly the mechanisms involved in flake breakage.
Cores and core reduction
Twenty-one cores, representing only 0.8% of the assemblage, were identified; 14 were worked using a bipolar technique (Tables 4 and 5). Three were core fragments. The bipolar cores had distinctive crushing and step-flaking at one or both ends, the result of being struck at 90[degrees] by a hammerstone when resting on an anvil. One bipolar core had use-wear along a section of one lateral margin. The six non-bipolar complete cores are single-platform (n=1) and multi-platform (n=5, including one bi-directional). All cores had at least one complete negative flake scar. On average, the non-bipolar cores are all larger than the bipolar cores (Table 4), though there are overlaps in all dimensions. The bipolar cores are all less than 25 mm long, except for one silcrete core at 46.7 mm. The non-bipolar cores are all greater than 26 mm long.
Cores are made from almost all identified...
|
