Previous Next

Drawing in the Land

6

Database Profile

This chapter provides a broad overview of the locational distribution of rock markings within the study area and considers what these patterns may indicate in respect of occupation, land use and sociocultural context. A summary of the database is presented, and the different types of site contexts and rock markings are quantified. The overall structure of the database in terms of site context, distribution and density is outlined first. The context of each site and the nature and distribution of their rock markings are then examined to identify variability and trends in spatial distribution. These patterns will be used to inform the analyses of temporal and spatial variability in rock art, presented in later chapters.

6.1 The Database

A total of 810 sites are in the Site Database, which includes data from the Illawarra Prehistory Group Database (IPG/DB) and the Research Database (R/DB). A summary of the different site contexts, their catchment location and frequency are shown in Table 6.1. The major site context in the database is rock shelters (n = 627; 77.4%). Open contexts account for 183 (22.6%) of the total site count. The Cataract contains the highest site count; the Nepean, the lowest. The difference in the density of sites in each catchment is explored below. Site density differences may reflect varying levels of occupation and land use.

Table 6.1 Number and percentage of site contexts (open or shelter) by river catchment (compiled from IPG/DB and R/DB).

Catchment

Open

Shelter

Total per catchment

Percentage per catchment

Cataract

Area: 207 sq km

90

194

284

35.1

Cordeaux

Area: 165 sq km

37

127

164

20.3

Avon

Area: 173 sq km

32

171

203

25.1

Nepean

Area: 246 sq km

24

135

159

19.6

Total

Area: 791 sq km

183

627

810

100

Total contexts (%)

22.6

77.4

100

Source: Table reproduced from Dibden (2011).

The total study area covers c. 791 square kilometres and the average site density in the Upper Nepean catchment is 1.02 sites per square kilometre. The site density for each catchment is listed below:

  • Cataract: 1.37 sites per square kilometre
  • Cordeaux: 1 site per square kilometre
  • Avon: 1.17 sites per square kilometre
  • Nepean: 0.64 sites per square kilometre

The Cataract contains the highest site frequency and the highest overall site density. Likewise, the Nepean contains the lowest site frequency and the lowest overall site density. The Cordeaux and Avon each contain comparable sites densities, which are slightly lower than the Cataract but considerably higher than the Nepean. This pattern of decreasing site density from north to south is a continuation of the density patterns between the Georges/Woronora and Cataract rivers identified previously by Sefton (1988). Site density will be examined further in respect of specific site types and the abundance of their various traits.

6.2 Open Site Contexts

As indicated, open site contexts account for 22.6 per cent of the total site count in the database. All open sites except one (EC12) contain grinding grooves. Of the 182 sites with grinding grooves (see Plate 1.8), 11 contain engraved groove channels (described below) and five have engraved rock art. In the following sections, grinding groove sites will be examined first and then open sites with engraved groove channels and engraved rock art.

6.2.1 Grinding Groove Sites

Grinding groove sites form a significant component of the archaeological record within the Upper Nepean catchment. Here, as in Attenbrow’s Upper Mangrove Creek study (2004:148), they are regarded as indicators of Aboriginal land use. The grinding groove data are used to examine, amongst other aspects of social life and subsistence, spatial variability with regard to the nature of Aboriginal occupation and land use. Patterns in the variable distribution of grinding groove sites in respect of topography have been observed elsewhere in the Sydney Basin. Attenbrow (2004) found that these sites were generally restricted to ridge landforms. Sefton (1988:54) also found that grinding groove sites on the northern Woronora Plateau were most frequently located at or near plateau level.

Grinding groove sites have a range of different groove counts. It is expected that by exploring count variability, correlated with environmental location, information might be gleaned regarding the foci of occupation and land use and this may be broadly informative of sociocultural context. Here the distribution of these sites and their variable groove counts is examined in respect of a number of geographic and environmental variables. First, the nature of these sites, their estimated antiquity and other matters relevant to this research are discussed.

Grinding groove sites contain grooves in rock surfaces that are produced through the shaping and/or sharpening of ground-edge stone hatchet heads (Attenbrow 2004). Groove size and morphology is known to be variable in the broader Sydney Basin, which suggests that they can result from the sharpening of a variety of different tools, and the preparation of food (Attenbrow 2004:43). In the study area, groove dimensions indicate that all are hatchet-grinding grooves.

A broad temporal framework for the age of grinding groove sites can be inferred on the basis of the age of ground-edge hatchet heads found within archaeological deposits. Across Australia, there is significant variation in the timing of the introduction of ground-edge hatchet technology and, in the south-east, the earliest hatchet heads date to the fourth millennium BP (Dibden 1996:35; Attenbrow 2004:241), and no earlier than 3,500 years ago (Hiscock 2008:155). The Pleistocene age of ground-edge hatchets elsewhere in Australia is well established (Morwood & Trezise 1989:78; Clarkson et al. 2015; Hiscock et al. 2016). Kamminga (1982:103) states that their later appearance into south-east Australia can be understood as the introduction of a pre-existing technology into new areas. At the southern margins of the Sydney Basin, hatchet heads found in excavated sites date to the last 3,000 years (Bowdler 1970; Lampert 1971; Flood 1980; Boot 1994:334). Using these dates and evidence from further north, it is suggested that grinding groove sites in the study area can be no older than 3,500 years. Given that hatchets were still used at the time of European occupation, the use of some grinding groove sites may have spanned this temporal range.

Morwood and Trezise (1989:85) suggest that the introduction of new technologies, such as ground-edge tools, in southern Australia acted to increase extractive efficiency and was adopted as a response to increased demands placed on regional production systems, due to population increase. According to Attenbrow (2004:241), the introduction of ground-edge technology in the Sydney Basin in the fourth millennium BP coincides with changes in climate that saw the advance of colder and drier conditions. She suggests that ground-edge implements may have been a part of risk minimisation strategies adopted for negotiating changes in vegetation and faunal communities associated with environmental fluctuation. In the Sydney Basin, there was a further change in the second and first millennia BP when a diversification in subsistence methods using ground-edge implements occurred (Attenbrow 2004:241; McDonald 2008a). Based on the presence of hatchet heads and fragments of ground edges in dated assemblages, Attenbrow (2004:223) argues that while grinding groove sites in the Upper Mangrove Creek catchment were used during the past 4,000 years, they were perhaps more commonly produced in the last 2,000 years. It is reasonable to assume that grinding groove sites in the study area may also have been used with greater frequency in the past 2,000 years.

It is well recognised that ground-edge stone hatchets played a significant role in Aboriginal subsistence activities (Sharp 1952:69; Binns & McBryde 1972:1; Dickson 1978:19, 35; Giopoulos 1986:2). Hatchets were used for a range of utilitarian purposes and functioned as a general multi-purpose tool (Dickson 1978:35).

While grinding grooves were formed as the result of specific technological activity and can be considered functional expressions of behaviour, they also have the potential to be informative of social aspects of life. Given the central importance of ground-edge implements in the economic life of Aboriginal people during the late Holocene, it is considered here that the location of grinding groove sites, particularly those with large groove counts, is potentially informative about areas in the Upper Nepean catchment that were habitually utilised within a specific social context. Whether or not that related to an exclusively gendered domain, general domestic space or some other mode is, however, unknown. It is also considered that grinding groove sites may have been imbued with cultural meaning and significance entirely independent of their functional context. This is plausible given that hatchets feature in the rock art of the Sydney Basin, as pigment and engraved graphics, and stencils, from which can be inferred the suggestion of their sociocultural meaningfulness.

It is well documented that in historic times, some stone hatchet heads were imbued with a high culturally constituted value and significance and that they featured as prestige items in formal contexts of gift exchange in the south-east (Sharp 1952; McBryde & Harrison 1981; McBryde 1984; Dibden 1996). An ethnographic study of the Yir Yoront in northern Queensland revealed that hatchet exchange was mediated by kinship structure and thus played a social role in generalising and standardising interpersonal relationships (Sharp 1952). Sharp also argues that hatchets in that area were imbued with masculine symbology.

Both men and women used hatchets, but the literature emphasises that they were most commonly men’s tools. When not in use, hatchets were worn tucked in men’s belts (Dickson 1978:19, 30). Hatchets can be considered to have been a form of material culture physically associated with the bodily appearance of men. It is feasible to consider that hatchets may have been gendered objects and that they could potentially have encoded information relating to, amongst other things, social identity, age, status and differing degrees of ritual knowledge. An analysis of hatchets from south-eastern New South Wales, including a number from the study area and adjacent coastal plain, established that they exhibit stylistic variability, which may indeed signify that they were used in strategies relating to the negotiation of social identity (Dibden 1996).

Grinding hatchet heads on stone can result in the formation of grinding grooves, which, similar to rock art, creates indelible marks on the rock surface and land. Grinding groove sites may have become significant and meaningful locales over time given their reference to an important item of material culture and their strong material presence in the landscape. Sites containing high groove counts are now visually significant marked locales. While the original motivation that led people to choose to grind hatchet heads at a specific place is now not understood, it is possible over time and, as places became increasingly embellished with grooves, that the meaning and significance of that locale was correspondingly changed. Grinding groove sites may have provided a physical and conceptual reference to the ancestral past and activities of previous generations. Because of the enduring physicality of grinding groove sites, they may have been meaningfully constituted expressions of place and mnemonic of past events and personal and group history (cf. Peterson 1972:16). This notion is explored further in later chapters.

A total of 182 grinding groove sites are distributed across the study area and 2,015 grooves occur in these (Dibden 2011:Appendix 4). A small number of grinding groove sites contain other rock markings in the form of engraved rock art and engraved groove channels. These additional traits will be discussed further below.

The Cataract has almost half of all open grinding groove sites (n = 90; 49.5%) and more than half of the grinding grooves (n = 1,086; 53.9%). The Avon and Cordeaux catchments contain similar numbers of sites; however, the Cordeaux contains markedly more grooves than the Avon. The Nepean contains relatively low site (13%) and groove counts (11%). The difference in the density of site numbers and groove counts between each catchment is explored further below.

The open context grinding groove site density in the entire study area is 0.2 sites per square kilometre and average grinding groove density is 2.5 grooves per square kilometre. The grinding groove site and groove density in each catchment is as follows:

  • Cataract: 0.4 sites per square kilometre; grinding groove density is 5.2 per square kilometre.
  • Cordeaux: 0.2 sites per square kilometre; grinding groove density is 2.6 per square kilometre.
  • Avon: 0.18 sites per square kilometre; grinding groove density is 1.6 per square kilometre.
  • Nepean: 0.1 sites per square kilometre; grinding groove density is 0.1 per square kilometre.

Grinding groove site and groove densities diminish from north to south. The Cataract contains the highest grinding groove density and the Nepean contains the lowest overall site and groove density. The Cordeaux and Avon contain similar site densities as each other, but the Cordeaux contains the highest groove density. There is a strong correspondence between the open grinding groove site numbers and densities in each of the catchments and the overall site frequencies and densities as outlined previously in Section 6.1.

In all catchments, the minimum groove count per site is one. The maximum number of grooves per site differs, with the Avon possessing the lowest maximum count of 41 grooves per site and the Cordeaux having the highest count of 92. The average groove count per site is 11.1. The majority of sites in all catchments possess low numbers of groove counts. The most frequently occurring groove count per site (mode) for all catchments is one and the overall median is five. Sixteen sites only possess groove counts of 31 or more. Approximately 70 per cent of open context grinding groove sites have 10 or less grooves.

The sites containing higher groove counts are predominantly located in the north-east sector of the study area (Figure 6.1). Large grinding groove sites occur on, or within very close proximity to, the main watershed divides and ridge crest landforms. As suggested in Chapter 2, these landforms are potentially areas habitually used for domestic occupation and movement through country.

A total of 129 sites have 10 or less grinding grooves. Most sites contain very low grinding groove counts, and the pattern of decreasing sites counts with higher numbers of grooves is a general trend. Almost 45 per cent of sites with 10 grooves or less possess one or two grooves only. This suggests that the majority of grinding groove activity is generally ad hoc, is expedient and not spatially focused. By comparison, those with large groove counts can be considered to reflect repeated and intensive use, albeit possibly over a long period of time, and this suggests that they have been produced as a result of a strong spatially focalised motivation.

The following analyses explore the distribution of these sites in accordance with a number of environmental variables. The majority of sites are located in upper valley slope contexts (n = 152; 84%). All sites with groove counts of 21 or more are located in upper valleys.

The majority of grinding groove sites are located in areas corresponding to the main thoroughfares in the study area (n = 146; 80%): the Major Divides, Spurs off Major Divides and Minor Divide landforms. Fewer sites are located on Spurs off Minor Divides, which are generally landforms located in the interior. The larger sites are located in upper valley slopes and on landforms corresponding to the main thoroughfares.

The majority of grinding groove sites (n = 171; 93%) are located in first- and second-order stream contexts. This is not surprising, given that most of these lower order streams are located on the upper valley slopes. Far fewer sites are in third- and fourth-order streams. The sites containing high grinding groove numbers (>10) are all located in first- and second-order stream contexts.

Figure 6.1

Figure 6.1 Location of size coded (as per grinding groove counts) open context grinding groove sites.

Source: Map reproduced from Dibden (2011).

6.2.2 Engraved Groove Channels

Engraved groove channels form a very minor component of the archaeological record in the study area (Table 6.2). Their function is uncertain, and it is not known whether they served a utilitarian purpose or otherwise. Despite their rarity and uncertainty in regard to function, they are a significant material and visual component of open context sites (Plate 1.8). Engraved groove channels occur across the Sydney Basin (Sefton 1988:65). While their function is occasionally surmised, they have not been the subject of a specific study and analysis. Sefton (1988:65) suggests that they appear to be functional, acting to control water flow around potholes. However, more often than not (94 per cent of sites), grinding groove sites do not possess engraved groove channels, and so a purely utilitarian purpose is questionable. It is beyond the scope of this research to conduct a detailed analysis of engraved groove channels. However, their distribution will be examined in respect of their relationship with other rock marks, and a number of geographic and environmental variables.

Engraved groove channels are linear (usually longer than 1 metre), narrow (c. 3 centimetres wide), shallow (c. 2 centimetres deep), incised or pecked into open rock platforms. Engraved groove channels can be relatively straight or sinuously curved. Although most groove sites do not have channels, when they are found, channels are typically at groove sites rather than in isolation; they can also occur with engraved rock art. They are usually in sites that have large numbers of grinding grooves. The age of engraved groove channels in the Sydney Basin is unknown. Given their association with grinding grooves, they may be of a comparable age range and antiquity.

Table 6.2 Summary description of sites with engraved groove channels (description as per IPG Site Cards).

Site

Catchment

Stream order

Graphics

Grinding grooves (gg)

Engraved groove channels (egc)

BMS1

Cataract

1

Yes

68

1 egc

C29

Avon

1

No

17

1 ‘L’ shaped egc, which diverts water to ggs & away from pothole

C41

Avon

2

No

1

1 egc from 1 pothole to another

DCC28

Cordeaux

1

No

1

1 egc around pothole; gg below pothole

EC2

Cordeaux

1

Yes

26

1 egc keeping seepage out of a pothole

Gill51

Cataract

1

No

54

1 egc extending from a small pothole diverting water past a large pothole

Lod31

Cataract

1

No

30

1 egc with pick marks visible; 1 egc extending from gg

Lod5

Cataract

1

No

45

1 egc at top of shallow pothole

T1

Cordeaux

1

No

59

1 egc diverting water from main water flow to side of the site

UA3

Avon

1

No

41

1 egc extends from a grinding groove in an alignment with water flow

Wall15

Cataract

1

No

52

2 egcs

Source: Table reproduced from Dibden (2011).

A total of 12 engraved groove channels are present in 11 open sites in the study area. Most sites contain one engraved groove channel only. The Cataract contains the highest count (n = 5). The Avon and Cordeaux contain three and there are none in the Nepean. The majority are associated with relatively large numbers of grinding grooves. All engraved groove channels are located in upper valley slope contexts, and all but one is located within first-order streams. The majority of sites (n = 9) are situated on main thoroughfares in the study area (Figure 6.2). It is evident that engraved groove channels are patterned preferentially in regard to topographic variables. This feature of their environmental context is shared with large grinding groove sites and engraved rock art.

Figure 6.2

Figure 6.2 Location of sites with engraved groove channels.

Source: Map reproduced from Dibden (2011).

6.2.3 Engraved Rock Art

The open context engraved rock art on the Woronora Plateau consists of pecked and abraded motifs of tracks, symbols, human figures, fish and macropods (Sefton 2003a:9), and is formally consistent with the regional Sydney–Hawkesbury engraving style (cf. McMah 1965; Attenbrow 2002:146; Sefton 2003a:9; McDonald 2008a). Engraved imagery is believed to have been produced by ‘conjoined-puncturing’, whereby a series of pits (measuring up to 3 centimetres in diameter and up to 1 centimetre deep) have been punctured into the bedrock; pits can overlap to form a continuous groove or may remain unjoined (Clegg 1985). In some instances, pits have been subsequently abraded. The objects used to create engraved graphics are unknown but may have included stone, shell, bone, wood and hatchet heads (Attenbrow 2002:147).

Approximately 50 open engraved rock art sites are on the Woronora Plateau. Sefton (1988:63) describes 22 of these in her study, and notes their restricted distribution within 11 kilometres of the coast. McDonald (1994:331) analysed 717 open engraved sites possessing 7,904 motifs in the wider Sydney Basin. South of the Georges River, engraved rock art sites were found to decline in both number and assemblage size (McDonald 1994:333).

The antiquity of the Sydney–Hawkesbury engraved rock art is not known, although it is certain that some engraved imagery was produced during the early period of European occupation (Edwards 1971:361; Attenbrow 2002:150). McDonald (2008a) argues that they date from the early Bondaian because of their stylistic similarity with shelter pigment art, but that the majority may have been produced within the last 1,000 years. Sefton (2003a) also argues that on the basis of weathering processes, they cannot be very old.

Of the Upper Nepean’s six open context sites containing engraved imagery: five contain one image and the other contains four (Dibden 2011:Appendix 5). These sites are located in the Cataract and Cordeaux and in the north-eastern corner of the study area, and none are located more than 11 kilometres from the coast (Table 6.3). This pattern conforms to the one Sefton (1988) identified. The decline in site numbers and density south of the Georges River, as identified by McDonald (2008a), continues southward across the Woronora Plateau. The most southerly occurrence of the Sydney Basin open context engraved rock art (EC2 and EC12) is at the mid/eastern end of the Cordeaux/Cataract divide (Figure 6.3).

Table 6.3 Description of engraved rock art in open contexts (per IPG Site Cards).

Name

Location

Motif description and associations

EC2

Cordeaux

One engraving, described by the IPG as a female symbol: an engraved circle with a central pit. 26 grinding grooves and one engraved groove channel.

EC12

Cordeaux

One whale motif.

BC3

Cataract

One macropod motif; one grinding groove.

BC4

Cataract

One emu motif; 12 grinding grooves.

BMS1

Cataract

Four motifs described as male and female symbols; 68 grinding grooves and one engraved groove channel.

Lod47

Cataract

One female with an axe in her right hand, with headdress, a waistband, one left breast and ‘clear female symbol at fork of legs’; 10 grinding grooves.

Source: Table reproduced from Dibden (2011).

As with large grinding groove sites and engraved groove channels, engraved rock art exhibits strong patterning in relation to environmental variables. All are located on thoroughfares on major watershed or minor divide landforms. All engraved rock art is found on upper slope and crest landforms, and in first-order stream contexts. Most of the engraved imagery is associated with substantial numbers of grinding grooves. Two engraving sites also contain engraved groove channels. The clustering of open engraving sites in the north-east sector of the study area is comparable with the general pattern of large grinding groove site distribution. Given the paucity of all these site traits (large grinding groove sites, engraved groove channels and engraved rock art) in the study area, their general spatial co-occurrence and similar environmental and topographical locational patterning are notable. The significance of this pattern in respect of a cultural landscape will be explored further in Chapter 9.

Figure 6.3

Figure 6.3 Location of open context engraved rock art sites.

Source: Map reproduced from Dibden (2011).

6.2.4 Open Context Sites—Summary

A number of notable features and geographic and environmental patterns are discernible in the open context site locational and distribution data. These are summarised as follows:

  • The majority of open context sites contain grinding grooves only. A small percentage contain engraved grooves channels (n = 11; 6%) and/or engraved imagery (n = 5; 2.7%). One open site contains an engraved image only.
  • The Cataract catchment contains almost half the grinding groove sites and more than half of the groove counts. The frequency and density of grinding groove sites and numbers of grooves per catchment diminishes from north to south.
  • Engraved groove channels occur in highest frequency in the Cataract and do not occur in the Nepean catchment.
  • Open context sites with engraved imagery occur in a restricted geographic area at the north-east of the study area and within 11 kilometres of the coastline. The two sites in the Cordeaux catchment represent the southern extent of engraved Sydney–Hawkesbury rock art.
  • Grinding groove counts vary between 1–92, with a mode of one and median groove count of five. The majority of sites contain low grinding groove counts. The range in median groove count is comparably low in all catchments. This comparable structural patterning suggests that the different densities of these sites between catchments are a reflection of variable intensities of land use in the Upper Nepean.
  • The majority of grinding groove sites contain single or otherwise very low groove counts suggesting that, generally, hatchet-grinding activity was not spatially focused. However, the pattern of a considerably fewer number of sites containing high groove counts does indicate that those sites are of a different order of significance. These locales were the focus of repeated activity.
  • The majority of grinding groove sites are located in upper valley slope/first-order stream contexts and landforms that are likely to have been thoroughfares in the Upper Nepean catchment. In Chapter 2, it was proposed that these landforms are those that people were likely to occupy habitually. The presence of the larger grinding groove sites in these places cannot necessarily be seen to confirm this, although the relationship is suggestive that this may be the case. In addition, given that the maintenance of such an important tool is likely to take place in a domestic or base camp context prior to embarking on hunting, foraging or journeying, the presence of the majority of these sites in these landforms gives further weight to this argument.
  • All grinding groove sites with large groove counts, all engraved groove channels and all engraved imagery are located in upper valley slope/first-order stream contexts. Large groove counts, engraved groove channels and engraved rock art frequently co-occur. Furthermore, the majority of sites that contain these traits are located on thoroughfare landforms and are clustered in the north-east sector of the Upper Nepean.

The following discussion compares these patterns with those identified by Sefton (1988) in her study area (Georges/Woronora and Cataract catchments). Sefton’s (1988:54) area contained 3,912 grooves in 316 open sites, with an average count of 12.4 grooves per site. This mean calculation is slightly higher than in the Upper Nepean and the range of groove counts (1–209) is considerably greater. However, both study areas show remarkable agreement in structural patterning. Sefton (1988:55) also found that 70 per cent of sites contained 10 or fewer grooves.

Sefton (1988:62) examined the density of grinding grooves, and found that higher densities occurred in the Georges/Woronora River basin than the Cataract. She argues that this pattern indicates a greater Aboriginal population in the former area. The ratio between grinding grooves and rock art sites in each catchment is equivalent, and she argues that this supports her assumption that grinding grooves are indicators of settlement patterns and density of catchment usage. The patterns identified for the Upper Nepean provide further support of a trend on the Woronora Plateau, of diminishing grinding groove site and groove count density, from north to south. Sefton (1988:54) identifies that most grinding groove sites are located at or near plateau level, and in swamps or creek beds at the heads of creeks. They are rarely found in valley bottom locations. She argues that, while suitable sandstone is not distributed everywhere in the landscape, not all outcrops contain grooves where it does occur, and that sites did not appear to be overused. This is also the case in the Upper Nepean and, accordingly, grinding groove site distribution can be regarded as a factor of choice rather than as having been ultimately determined by environmental constraints. Attenbrow (2004:224) identified a similar pattern of grinding groove distribution relative to topography in the Upper Mangrove Creek catchment, where grinding grooves have a more restricted distribution than any other archaeological trait. She found that most sites with grinding grooves (74%) occur in upper elevations in periphery ridgetop zones (comparable with major watershed divide landforms as defined in this research).

The locational patterning of engraved groove channels in Sefton’s study area is also comparable with the Upper Nepean catchment, in which all but two of 21 sites are located at plateau level. The remaining two are located in creek beds in upper valley slope contexts (Sefton 1988:67). All engraved groove channels, except one, occur in association with grinding grooves. All sites with engraved groove channels (except one, which is interpreted to be unfinished) contain high numbers of grooves. While the frequency of grinding groove sites and engraved groove channels is high in the northern catchment, the proportional frequency of engraved groove channels and grinding groove sites is the same in both (Sefton 1988:67). All 22 engraved rock art sites in Sefton’s (1988:63) study are located in either ridge-top or near plateau-level locations. Sefton found that while the geographic distribution of engraved rock art (as confined to the eastern sector) and grinding groove sites do not coincide, the frequency of both diminishes from north to south, and the proportional frequency between the two is the same for both catchments.

The patterns in open context site contents and geographic and environmental location identified in this section, along with broader trends, as identified north of the Upper Nepean catchment by Sefton (1988), will be discussed again in conjunction with patterns in sheltered rock art in Chapters 8 and 9.

6.3 Shelter Site Contexts

Shelter site contexts account for 627 (77%) of the total site count in the database. Of these, 509 contain rock art, of which 110 have been subject to detailed recording for this research (Figure 6.4). As with open context sites, data relating to the remaining shelter sites have been obtained from the IPG site recordings, its various reports (Sefton 1989, 1990, 1991, 1992, 1994, 1995, 1996, 1997, 2000, 2003a) and others (Biosis 2007). The analyses of shelter contexts and their traits is undertaken on two levels. The first is based on general geographic, environmental and site contents data compiled from the IPG/DB and R/DB, and the second is based on detailed shelter morphology and rock art micro-topographical locational data from the R/DB. These latter data are based on a sample of the sites in the study area, and so the analyses of shelter morphology and micro-topographic location are conducted to indicate only the diversity and range of situational contexts that exist in relation to sheltered rock art. In subsequent chapters, analyses will be undertaken in respect of these variables to investigate their relationship with rock art in more depth, and whether or not there is a temporal signature to this diversity.

Figure 6.4

Figure 6.4 Location of rock art rock shelters in the Research Database.

Source: Map reproduced from Dibden (2011).

The IPG have not recorded rock shelters that do not contain observable ‘direct’ evidence of use, although many have been encountered during their surveys (Sefton 1988:70). Sefton (1988:71) refers to Attenbrow’s (1987) work in the Upper Mangrove Creek catchment where, upon excavation, 60 per cent of such sites were found to contain deposit. The implication is that shelters with archaeological deposit are likely to be underrepresented in the IPG/DB. This underrepresentation is uniform across the entire Upper Nepean and, hence, is unlikely to influence site density calculations. This data limitation is of minor importance to the main thrust of this research.

6.3.1 Shelter Contents—General

The majority of shelters (81%) contain rock art in the form of graphic and gestural (stencils and prints) marks (Table 6.4; Dibden 2011:Appendix 6). Other archaeological traits may also be present. Grinding grooves are present in 21 shelters. Stone artefacts have been recorded in 278 shelters, 118 of which do not contain rock art. Small numbers of other features, such as the caching of stone tools and ‘quarrying’ of quartz pebbles, have been recorded.

Table 6.4 Summary of archaeological traits per shelter in the Upper Nepean catchment (IPG/DB and R/DB).

Catchment

Stone artefacts

Grinding grooves

% with ggs

Rock art

% with art

Art counts

% art count

Total shelters

Cataract

85

12

57.1

160

31.4

1,424

27.8

194

Cordeaux

46

1

4.8

114

22.4

1,494

29.2

127

Avon

72

4

19

140

27.5

1,562

30.5

171

Nepean

75

4

19

95

18.7

640

12.5

135

Total

278

21

509

5,120

627

Source: Table reproduced from Dibden (2011).

The density of shelters with rock art, across the entire study area, is 0.6 sites per square kilometre. A total of 5,120 graphic and gestural marks are distributed at an average density of 6.4 per square kilometre. The site and rock art mark density in each catchment is as follows:

  • Cataract: Sheltered rock art site density is 0.8 sites per square kilometre; rock art mark density is 6.8 per square kilometre.
  • Cordeaux: Sheltered rock art site density is 0.7 sites per square kilometre; rock art mark density is 9.1 per square kilometre.
  • Avon: Sheltered rock art site density is 0.8 sites per square kilometre; rock art mark density is nine per square kilometre.
  • Nepean: Sheltered rock art site density is 0.4 sites per square kilometre; rock art mark density is three per square kilometre.

Similarities and differences are evident in patterns of sheltered rock art site density compared with open context grinding groove data. The Nepean catchment contains the lowest overall site and rock art mark density. In contrast to open sites where the Cataract catchment contains significantly higher site and groove densities, the Cordeaux and Avon catchments contain sheltered rock art site densities that are comparable with that found in the Cataract. The Cataract contains a lower density of rock art marks compared with the Cordeaux and Avon. This pattern is unexpected given trends in various other lines of evidence (Section 6.1.1 and 6.2), which reveal a pattern of site and trait decline from north to south. It is also notable that the Cordeaux and Avon both possess higher frequencies of rock art motifs in fewer shelters, compared with the Cataract and Nepean. The question of what may have influenced this different geographic pattern in sheltered rock art distribution is analysed further in Chapter 9.

Over half (57%) of the rock shelters with grinding grooves are in the Cataract catchment. This distribution is comparable with open grinding groove sites in which 49.5 per cent are in the Cataract. Also, comparable with open contexts, the Cataract has the highest frequency (68.1%) of grooves in shelters. This correspondence of grinding groove frequency between open and shelter contexts in the Cataract suggests that the pattern relating to the geographic distribution of sheltered rock art density is different, and potentially informative of different behavioural expression relating to rock art.

The sheltered grinding groove site density in the entire study area is 0.02 sites per square kilometre; average grinding groove density is 0.09 grooves per square kilometre. The sheltered grinding groove site and groove density in each catchment is as follows:

  • Cataract: 0.057 sites per square kilometre; grinding groove density is 0.236 per square kilometre.
  • Cordeaux: 0.006 sites per square kilometre; grinding groove density is 0.006 per square kilometre.
  • Avon: 0.023 sites per square kilometre; grinding groove density is 0.064 per square kilometre.
  • Nepean: 0.016 sites per square kilometre; grinding groove density is 0.045 per square kilometre.

Sheltered grinding groove site and groove densities diminish significantly between the Cataract and the southern rivers. The Cordeaux, however, contains the lowest site and groove density, and this contrasts with the open site density pattern. The Avon contains higher sheltered grinding groove density than the Nepean, which is comparable with the open site pattern.

The descriptive statistics for shelter context rock art marks reveal a number of additional similarities and differences between the catchments (Table 6.5). The maximum number of rock art marks per site differs between catchments. The Nepean possesses the lowest maximum count of 41 rock art marks per site and the Cordeaux has the highest count of 206. The average rock art mark count per site is 10. The median count is four for each of the Cataract, Cordeaux and Avon catchments.

Table 6.5 Summary statistics relating to shelter contexts with rock art marks per catchment (IPG/DB and R/DB).

Catchment

Number of sites

Total image count

Range of counts

Average count

Median

Standard deviation

Cataract

160

1,424

1–146

8.8

4

15.7

Cordeaux

114

1,494

1–206

13.1

4

26.8

Avon

140

1,562

1–108

11.1

4

18.3

Nepean

95

640

1–41

6.7

3

8.4

Total

509

5,120

1–206

10

4

18.6

Source: Table reproduced from Dibden (2011).

Very few sites have large rock art mark counts. Nearly 80 per cent of shelter context rock art sites have 10 or less rock art marks. Less than 5 per cent of sites have rock art mark counts of 31 or more. This pattern of distribution is similar to that encountered in the exploration of grinding groove counts in open contexts. The production of rock art also appears to conform to a pattern of spatially focalised activity represented by sites with high rock art counts on the one hand and, on the other, a much more situationally diffuse activity characterised by limited and often one-off events.

The locations of size-coded sheltered rock art sites are shown in Figure 6.5. The most obvious spatial pattern is that the Nepean catchment does not contain shelters with large rock art counts. The remaining catchments all contain large sheltered art sites, but each has different locational patterns. The largest site in the Cataract is located in the valley near the dam wall and in the centre of the catchment. This contrasts with the Cordeaux and Avon, where large sites are located on or within close proximity to watershed divides or ridge crests of minor divides. There is some tendency for these large sites to be relatively close to the Illawarra Escarpment, rather than within the interior of the plateau.

As the majority of shelter contexts contain 10 or fewer rock art marks, the distribution in these sites is explored further. Rock art counts of one or two occur in 34.8 per cent of the total number of shelters in the study area. Shelters with one (n = 89) or two (n = 88) rock art marks account for 44.3 per cent of this group of shelters.

While shelters with low rock art counts are possibly representative of marking activity undertaken within discrete time frames or temporal phases, those with large rock art counts may represent accumulations of imagery over significantly longer periods of time and in successive temporal phases. An analysis of the spatial distribution of sheltered rock art sites will be undertaken in Chapter 9, which explores locational patterns further and examines variability between different temporal rock art phases.

The following discussion explores the distribution of sheltered rock art sites in accordance with the same suite of environmental variables used to examine grinding groove sites. Given that temporal variability may be a significant determinant in regard to rock art location, the purpose of these analyses is to reveal the range of environmental variability only. Approximately half of shelters with rock art (55.6%) are located in upper valley slope contexts. The remainder are distributed almost equally between mid and lower slope contexts. This distribution of rock art differs significantly from the pattern in open grinding groove sites, although there are slightly more sites with rock art counts of 60 or more in upper valley slope rather than mid or low contexts.

The Cataract catchment contains rock art shelters in comparable densities in all slope contexts. However, the remainder of the catchments, generally, contain more sites on upper slopes, particularly in the Nepean. The physical differences between each catchment were described in Chapter 2, and it was noted that occupation and travel within the Cataract valleys was likely to be less constrained than in other catchments. The pattern in different site locations between each catchment most likely reflects the differences in topography between them, and a general preference to occupy all valley locations in less treacherous terrain, or the converse.

The majority of rock art sites (71%) are located in areas that correspond to the main thoroughfares in the study area. This pattern of preferential location near major thoroughfares is comparable with that relating to open grinding groove sites, although to a lesser extent (see Dibden 2011:Table 6.20). The different terrain of the Cataract catchment is likely to be influencing this distribution to some degree.

The location of sheltered rock art sites in relation to stream order also diverges when compared with that identified for grinding grooves sites. While the majority are located in lower order stream contexts, they are also commonly located in higher order contexts, and this is commensurate with the greater numbers of sites located on mid and lower slopes. Significantly, sites with high rock art counts are preferentially located in lower stream order contexts.

Figure 6.5

Figure 6.5 Location of size coded (as per rock art counts) sheltered rock art sites.

Source: Map reproduced from Dibden (2011).

6.3.2 Shelter Rock Art—Marks

In this section, shelters with rock art and their variability in respect of different categories and counts are examined briefly. Pigment graphics occur in 461 of the 509 rock art shelters. Stencil imagery occurs in 130 shelters. Prints are relatively uncommon (five sites). Gestural applications of pigment marks occur in 23 sites, and gestural subtractive marks such as pitting occur in 21. The latter categories are likely to be underrepresented because, except for large red pigment smears, these mark types have not been recorded by the IPG. Engraved graphics occur in two shelters.

Rock art occurs in shelters in three main groupings: pigment graphics only, stencils only and graphics and stencils (Table 6.6). The majority of shelters (74.5%) possess graphics only. Forty-seven sites (9.2%) contain stencils only, and 83 (16.3%) have both graphics and stencils.

Table 6.6 Data on number and percentage of the main rock art categories in each catchment (IPG/DB and R/DB).

Catchment

Graphics only

% only graphics

Stencils only

% only stencils

Graphics & stencils

% graphics & stencils

Cataract

104

65%

16

10%

40

25%

Cordeaux

90

79%

8

7%

16

14%

Avon

108

77.1%

15

10.8%

17

12.1%

Nepean

77

81.1%

8

8.4%

10

10.5%

Total

379

74.5%

47

9.2%

83

16.3%

Source: Table reproduced from Dibden (2011).

The relative percentage frequencies of these groupings vary between catchments. The Cataract catchment has more shelters with stencils and fewer with graphics only, compared with the southern catchments. This difference in the relative abundance of a specific trait between the Cataract and other catchments has been identified previously in regard to other data. The greater frequency of stencilling in the Cataract is explored at a deeper analytical level in Chapters 8 and 9.

6.3.3 Sheltered Rock Art—Shelters

Rock art occurs in a wide range of shelter morphological types and sizes. Shelter variability is explored here with data from the Research Database (see Dibden 2011:Appendix 7:Shelter Database). In all aspects relating to rock art shelter size, living space and art panels, shelters are highly variable. While the minimum length, height, width and volume calculations are very low, the largest shelters are of enormous dimensions. Given the mode of zero in respect of living space categories, it is common that shelters provide space that is neither suitable for use as a living/work area, nor as a sleeping space. The range in size of rock art surfaces is also significant and is between 1 and 103 square metres.

In Table 6.7, shelter variability is described further in respect of shape, the nature of floors and dampness, location of rock art in respect of optimal living space, and location within elevated rock faces. These shelter categories will be examined in later chapters in relation to the nature of the rock art and temporal ordering. Of note here is the predominance of highly uneven rock-floors in half of all rock art shelters, and this accounts for the low number of shelters with Living or Optimal Living Space. It is also notable that eight shelters are perched in elevated contexts within vertical rock faces.

Table 6.7 Numbers of morphological types and other shelter characteristics (n = 110: R/DB).

Category

Shelter counts

Shape

dome

15

open

20

scallop

14

square

25

wedge

36

Floor

Primarily rock

56

Primarily sediment

38

Half/half

16

Relationship to ground surfaces

Shelter perched in elevated rock faces

8

Location of art to relationship to Optimal Living Space

Not Applicable (i.e. no Living Space)

41

Away

4

Adjacent ceiling

2

Adjacent wall

52

Adjacent wall and ceiling

11

Dampness

Dry

95

Mixed

9

Wet

6

Source: Table reproduced from Dibden (2011).

The variability that the rock shelters possess suggests that marking the land with rock art is likely to have been motivated by a range of purposes and meanings. The different variables relating to shelters provide opportunities and constraints to human beings as embodied individuals.

6.3.4 Sheltered Rock Art—Micro-topography

Rock art occurs in a range of micro-topographic contexts within shelters (Table 6.8). Variability in rock art location is explored in this section with data from the Research Database (see Dibden 2011:Appendix 8:Mark Database). A total of 2,565 rock marks have been identified in the 110 rock shelters recorded for this research. The majority of rock marks have been produced on walls and in open visibility contexts. Eighty-three marks have been identified to be in a relationship of isomorphic congruence with natural features on the rock art panels. The lowest mark is 5 centimetres above a shelter floor, and the highest is c. 3 metres. The average height is 106 centimetres, and the median and mode are both 100 centimetres. Plates 6.1, 6.2 and 6.3 illustrate variable visibility contexts in which rock marks can occur.

Table 6.8 Numbers of locations and visibility characteristics of rock marks (R/DB).

Category

Rock mark counts

General location

Wall

2,034

Ceiling

269

Concavity

262

Visibility

Open

2,134

Moderate

261

Hidden

170

Isomorphic congruence

In a relationship of isomorphic congruence

83

Source: Table reproduced from Dibden (2011).

Plate 6.1

Plate 6.1 Location of a ‘hidden’ motif on the roof (50 centimetres above the rock floor) of a small concavity in the rear wall of a shelter (rock shelter SCR14; mark ID 2459).

What is particularly interesting about this shelter is that the rear wall contains abundant suitable rock art panels that are devoid of art, other than the small hidden motif that cannot be seen unless one kneels down and looks up into the roof of the concavity.

Source: Photograph by Julie Dibden, 2011.

Plate 6.2

Plate 6.2 Location of a highly visible motif situated c. 3 metres above the ground (as measured from base of image to the ground) on an outside face of a shelter (rock shelter DCC12; mark ID 1534).

Source: Photograph by Julie Dibden, 2011.

Plate 6.3

Plate 6.3 Close-up of mark ID 1534.

Source: Photograph by Julie Dibden, 2011.

6.3.5 Shelter Context Sites—Summary

The notable features and geographic and environmental patterns discernible in the shelter context site data are summarised as follows:

  • Shelters with rock art are distributed in different densities between the catchments in a pattern that contrasts with open context grinding groove site distribution, except in the Nepean.
  • The Avon and Cordeaux catchments contain comparable site densities with the Cataract, but higher densities of motifs.
  • Grinding grooves in shelters have a comparable density distribution with that found in open grinding groove sites.
  • The Avon and Cordeaux catchments contain more motifs in fewer shelters than the other catchments.
  • Rock art counts per shelter vary between 1–206, with a mode of one and median count of four. The majority of shelters contain low rock art counts.
  • Generally, rock art activity was neither intensive nor spatially focused. The fewer number of shelters that contain high rock art counts suggests that more intensively marked sites were utilised within a different sociocultural context to smaller sites.
  • Rock art shelters are distributed relatively evenly between landforms, but larger sites are almost exclusively located on landforms in which domestic occupation and movement through country is inferred. In the Cataract catchment, large sites occur in valley bottoms. Elsewhere, they are located most frequently on watershed divides or the crests of ridges.
  • Shelters contain predominantly graphic rock art with lesser frequencies of gestural marks. The Cataract catchment has a greater proportional frequency of stencils.
  • Rock art shelters are extremely variable in their morphology and dimensions, and the majority have highly uneven rock floors. The majority of shelters do not possess attributes that would have allowed for their use as domestic habitation sites. The sociocultural context of rock art production and perception may not have been equivalent in all shelters.
  • Rock art is predominantly on shelter walls in contexts of open visibility. However, rock art occurs in a range of micro-topographic contexts within shelters, suggesting that the sociocultural context of rock art production and perception was highly variable.

6.4 Summary

In this chapter, an overview of the database has been presented and the broad categories of archaeological material and the contexts in which these occur have been quantified. General trends in the variability of site and trait frequency, and their spatial density across the study area have been identified. The Cataract catchment contains the greatest numbers of sites and the highest overall site density. Site density in the Cordeaux and Avon catchments is comparable, but lower than the Cataract and higher than the Nepean. This pattern has been mirrored, if not amplified, by the grinding groove site data. The structural comparability between each catchment in the Upper Nepean, and with areas to the north (Sefton 1988), suggests that the pattern of variable site density is real and not a reflection of sampling bias. The spatial trend of site frequency and density decreasing from north to south is possibly an indicator of differential levels of Aboriginal land use across space in the period since c. 3,500 BP, when it is assumed that ground-edge stone hatchets began to be used in the region. This spatial pattern, in regard to site density, is only behaviourally meaningful if an equitable opportunity exists for rock platform and shelter use in all catchments of the study area. It has been beyond the limits and scope of this project, and possibly any research, to quantify the availability of suitable stone exposures across the study area in order to test for this. Based on personal observation and the IPG records and documents, it is the case that stone platforms and exposures are abundant across all catchments. The geology and geomorphology are relatively uniform. Accordingly, the site-density patterning in the Upper Nepean is believed to be a result of human choice, rather than being environmentally determined.

It has also been identified that open context rock art and engraved groove channels occur in greater numbers in the Cataract catchment than elsewhere, and not at all in the Nepean. The southerly limit of open context engraved rock art in the Cordeaux catchment, and engraved groove channels in the central area of the Avon, marks the cessation of practices that were clearly of decreasing and less importance in the area of the southern Woronora Plateau. In subsequent chapters, the question as to what these patterns in the Upper Nepean may indicate will be addressed.

An unexpected pattern, identified in this chapter, is that sheltered rock art abundance in the Cordeaux and Avon catchments does not mirror the decreasing density trends found in other site patterns. The Cataract catchment contains densities of sheltered rock art sites comparable with that found in the Cordeaux and Avon, but has lower densities of rock art marks. This suggests that the Avon and Cordeaux catchments were the focus of more intensive shelter context rock marking than when compared with the Cataract.

Open grinding groove and sheltered rock art sites possess a similar diversity in the number of marks they contain, and predominantly have low counts of grinding grooves or rock art. On the other hand, small numbers of both site types contain a contrastingly high number of marks. This variability may be inferred to reflect differences in respect of the use of land, which may be the result of their production within different sociocultural contexts.

The variability in rock art shelters, including their environmental and topographic location, morphology, the nature of the rock marks they contain, and the micro-topographic location of marks, has been defined in this chapter. The topographic location of rock art shelters has been found to be more variable than open context grinding groove sites. This cannot be accounted for by the availability of suitable sandstone shelters and exposures.

The examination of diversity in the physical location of rock marks, whether on a macro- or micro-topographic scale, focuses attention on questions of sociocultural context, occupation and meaning. Further analyses will be undertaken in following chapters to address the questions that have arisen in this chapter.


Previous Next