Discontinuous dreaming networks: analyses of variability in Australian pre-historic petroglyphs.
Abstract. This paper analyses the spatial variation within a corpus of pre-Historic petroglyphs, the Panaramitee, which has been represented in the literature as a homogeneous pan-Australian 'style'. Correspondence analysis and group average cluster analysis found that there were five regional groups
KEYWORDS: Panaramitee--Petroglyph--Regionality--Dreaming Tracks--Australia
Introduction
This paper explores the degree of spatial variation within a particular corpus of Australian pre-Historic petroglyphs. Dominated by tracks and simple nonfigurative motifs, this corpus of rock art has been labelled the 'Panaramitee style' (Maynard 1976, 1979). It has been represented as homogeneous at a continental level in terms of technique, form and motif proportions. By contrast, two other defined sets of Australian rock art 'styles', the 'simple figurative' and 'complex figurative', are more variable and spatially restricted, and have also been argued to be more recent than the Panaramitee (Maynard 1976, 1979).
These 'styles' were defined by Maynard (1976, 1979) in a simple model that she proposed for the development of Australian rock art. This scheme has been widely used as a basis for further research (e.g. Clegg 1981b, 1984a, 1984b, 1986, 1987, 1992, n.d., in prep.; David 1988; David and Chant 1995; Flood 1997; Franklin 1984, 1986, 1989, 1991, 1992; McDonald 1982, 1983, 1988; Morwood 1979, 1984, 1988; Nobbs 1984; Smith 1992), as well as being cited in syntheses of Australian archaeology (e.g. Mulvaney and Kamminga 1999; White 1984; White and Habgood 1985). It is also a useful baseline for overviewing geographical and chronological change in Australian rock art (Morwood 2002). In contrast to other proposed Australian rock art sequences, Maynard's synthesis appears to explain much of the pronounced spatial and chronological variation (Flood 1997; Morwood 2002).
Following standard archaeological practice, Maynard named the first of the 'styles' in her sequence after a type-site in South Australia, Panaramitee, which contained an extensive and typical example of this form of rock art. The Panaramitee consists of pecked petroglyphs of 'macropod' and 'bird tracks' (1), 'human footprints', circles, dots, crescents, spirals, radiate designs and only a small proportion of figurative motifs other than tracks (Fig. 1). The forms of the motifs are bands, thick outlines and solid figures. These petroglyphs reveal a pan-continental distribution, being found in South Australia, central Australia, western New South Wales, at Ingaladdi (now known as Yingalarri) in the Northern Territory, in the Laura area of Cape York Peninsula, the Mt Isa region of north-western Queensland, and in Tasmania (Maynard 1979: Fig. 4-1) and have been identified more recently in Western Australia at sites such as Scott River (Clarke 1983), and also in the Northern Territory in Kakadu National Park (Chaloupka 1993; Flood 1997; Sullivan 1988) and at Helen Springs (Mulvaney 2001).
[FIGURE 1 OMITTED]
According to Maynard (1976, 1979) and Edwards (1966a, 1971) before her, the most significant feature of the sites in South Australia, central Australia and western New South Wales (the 'classic sites', Maynard 1979: Fig. 4-1) is a constancy in the relative proportions of motifs, while petroglyphs of this nature found in other regions of Australia (e.g. Laura and Tasmania) do not have this feature. Maynard observed that this was because the smaller number of petroglyphs at the latter sites meant that not enough figures were accumulated to result in the motif proportions found at the widely distributed 'classic' sites, or because the sites were not sufficiently well described in the literature to determine motif proportions. However, since these peripheral sites displayed identical techniques, forms and range of motifs, they too were included in the Panaramitee Maynard 1976, 1979).
By contrast with the Panaramitee, the succeeding simple figurative and complex figurative styles' in Maynard's sequence each comprise several distinct, regionally varied styles of petroglyphs and pictograms united by a series of common characteristics, in particular:
* The predominance of figurative motifs in the form of simplified silhouettes of human and animal models.
* A considerably decreased frequency of tracks compared to their dominance in the Panaramitee.
* The strong standardisation of portrayals, with human beings depicted frontally, animals and birds in profile, snakes and lizards in plan view.
* The inclusion of decorative details, which consist of stripes, bars, dots, and the like as infill, and different colours for the outlining of painted solid figures.
Simple figurative styles occur around the northwestern, northern and eastern edges of Australia, as well as at sites in western New South Wales, which constitute the furthest inland examples (Maynard Fig. 4-1).
The main distinction between the simple figurative and the complex figurative 'styles' is that the latter is, 'in some respect, more sophisticated than crudely naturalistic' (Maynard 1979: 100). The 'complex figurative' is a corpus of extremely diverse rock art 'styles', found exclusively in coastal regions of the north and north-western corner of the continent. Examples are the Mimi and x-ray paintings of the Arnhem Land escarpment (e.g. Brandl 1982), the Gwion Gwion (Bradshaw) and Wandjina figures of the Kimberley region (Crawford 1968, 1977), and the 'Gurangara' petroglyphs (now known as Woodstock figures, Bednarik 2002; McNickle 1985; Morwood 2002) found at certain sites in the Pilbara.
A variety of archaeological evidence was adduced by Maynard in support of the sequence, including absolute dates where available (e.g. for petroglyphs at Ingaladdi/Yingalarri), other evidence regarding the age of rock art, distribution, quantitative analysis, and some superimpositions (Maynard 1976, 1979). More recently, it has been proposed that the Panaramitee may have persisted over a long period of time (Franklin 2004), as indicated by the minimum age of 13 000 years obtained for buried petroglyphs at Early Man in the Laura region (Rosenfeld 1981a), a maximum age of about 13 500 years for pecked weathered circles on a boulder at Puritjarra, central Australia (Smith 1996; Rosenfeld and Smith 2002) and the continued use of Panaramitee-type motifs in recent Aboriginal artistic systems (see e.g. Anderson and Dussart 1988).
Maynard's (1976, 1979) model identified an interesting pattern within Australian rock art, which saw a change from an early, continent-wide homogeneous 'style', to a series of separate, regionally varied 'styles'. Maynard suggested that this pattern parallelled that within Australian stone artefacts (see also Flood 1997), but otherwise did not explain the archaeological significance of her model.
This paper explores the degree of spatial variation within the first of Maynard's 'styles', the Panaramitee, in an attempt to:
* First, assess the claims for a homogeneous Australia-wide corpus of petroglyphs in greater detail and with a more extensive sample of sites than was available to her.
* Secondly, further examine the significance of the claimed differential variation between the Panaramitee on the one hand, and the simple figurative and complex figurative 'styles' on the other. Previous analyses of the variation within the simple figurative 'styles' (Franklin 1984, 1989) indicate that many of the claims concerning this corpus of art can be upheld. Similarly, two separate multivariate analyses confirmed that the Panaramitee can be distinguished as an entity from the simple figurative 'style' (Franklin 1991, 1992, 2004).
The analyses
The following major question is addressed in the analyses presented in this paper:
Do the petroglyph sites which have been referred to the Panaramitee represent a homogeneous body of rock art, or can regionality be identified within them?
The Panaramitee has been represented in the archaeological literature as a homogeneous corpus of petroglyphs (e.g. White 1984). Although Maynard (1976, 1979), drawing on the work of Edwards (1964, 1965a and b, 1966a and b, 1968a and b, 1971), claimed that this was so in a general sense, they noted that there was also variation in terms of particular motifs which were unique to different sites (e.g. 'owl' motifs occurred at Red Gorge [Mountford and Edwards 1964], human figures with radiating headdresses at N'Dhala Gorge [Edwards 1971; Forbes 1983]) and in the proportions of motifs, which were only constant at the large sites in the 'arid zone' (see Mulvaney and Kamminga 1999; Smith 1989 and Veth 1989 for the location of this area and fluctuations in its extent over time), and not in the small sites at the periphery of the continent (Maynard 1976, 1979). However, this variation has often been overlooked in the archaeological literature. Furthermore, the variation within the simple figurative and complex figurative 'styles' was claimed by Maynard to be much greater than that within the Panaramitee, occurring in the first two categories of 'styles' in terms of technique, form and the qualities that constitute 'character' (see Maynard 1977: Table 5).
The analyses presented in this paper represent a test of the first part of Maynard's sequence, the Panaramitee, by examining whether or not this corpus of rock art is internally consistent. The analyses therefore explore the variation between petroglyph sites that Maynard and others since have referred to the Panaramitee (see below). Another test of Maynard's sequence examined whether or not the Panaramitee could be distinguished as an entity from the simple figurative 'style' (Franklin 1991, 1992, 2004; see above).
Methods
Spatial variation within the Panaramitee is explored Australia-wide using correspondence analysis and cluster analysis of motif proportions between sites. Analyses of presence/absence and diversity of motifs are presented in Franklin (1992, 2004).
The data used in the analyses were in the form of counts of motifs from each of the sites. A typology of motifs was therefore required to enable this counting procedure to be as consistent and standardised as possible (see Franklin 2004).
The master typology used in the analyses by necessity had to incorporate motif typologies used by previous researchers, as it was not possible to visit all the sites used in the analyses to obtain counts of motif types. The recordings and counts of petroglyphs produced by other researchers had to be used. The master classification for motifs throughout the continent that was established (Table 1) extended the localised typologies which were previously only useful for individual sites or regions.
[TABLE 1 OMITTED]
Counts of motifs were obtained from both published and unpublished sources, from either tables of data, or if these were not available, from line drawings or tracings of the petroglyphs (Tables 2 and 3). In the latter case, motifs were counted from the illustrations according to the pre-established typology. New motif counts were also obtained during fieldwork, which was undertaken in the Laura and Mt Isa regions of north-east and north-west Queensland and in southwestern Australia at petroglyph sites near Scott River and on Edah Station.
Abraded grooves, which have the appearance of cut marks in the rock, occur at many petroglyph sites, either interspersed with other engraved motifs, or on separate surfaces not associated with other petroglyphs. They may be so numerous that it is often extremely difficult to obtain an accurate count for them, and available figures probably represent an underestimate of their true number. Consequently, abraded grooves were excluded from the motif counts and not considered in any of the analyses. Similarly, although amorphous and indistinct motifs were included in the counts for some sites, these 'types' were also excluded from consideration because they are not meaningful in this context.
Although it is recognised that there are probably chronological variations through time, each site has been examined as a whole (see also Franklin 1992), rather than separated into purported chronological groupings, as this was the only way that analysis could follow (see also Ross 1997, who faced a similar dilemma in her analysis of pictograms in northwest central Queensland). Although Flood (1997) has criticised this approach, her argument is rather circular, as she stated:
An additional problem is that most sites are treated as a whole, even if their art spans 40,000 years! Only occasionally can the art of a site be divided into different periods which are analysed separately ... This probably sounds an insuperable obstacle, but in the case of the Panaramitee tradition in fact it is not, since at classic sites such as Karolta and Panaramitee North virtually no change in style or subject matter is discernible throughout the site's history (Flood 1997: 195)
If this is the case, it is not necessary to divide the sites into hypothetical periods and analyse each one. Indeed, in most cases there was no basis upon which to undertake such separate analyses, as the various researchers provided data insufficient to this task. Flood's argument is also partly based upon her acceptance of the cation-ratio dates for petroglyph sites in South Australia (Dorn and Nobbs 1992; Dorn et al. 1988; Nobbs and Dorn 1988, 1993). These dates are no longer accepted, even by the researcher involved in the dating program (Dorn 1996; see Franklin 2003, 2004 for a discussion of this issue). Finally, a basic premise upon which the analyses in this paper are based is that archaeologists cannot at present deal very easily with time in rock art, but we can deal with space.
Analytical techniques
Each petroglyph site was described in terms of the incidence (frequency) of motif types in the master typology. The sites were then compared with each other in terms of these motif counts using the multivariate techniques correspondence analysis (e.g. Bolviken et al. 1982; Hill 1974) and group average cluster analysis (e.g. Doran and Hodson 1975; Hodson 1970).
The analyses used:
* A correspondence analysis program published by Wright (1985) for the direct factoring of a data matrix and written for use on portable microcomputers.
* A series of programs, EUCLID, HIERCL and DENDRO, for the group average cluster analysis (R. V. S. Wright pers. comm.). A standardisation procedure of 'normalisation by object' removed the effect of absolute abundance of motifs at large sites.
The data
The specific sample of sites (see Fig. 2 and Table 2) was selected for analysis on the following bases:
[FIGURE 2 OMITTED]
(i) The sites were claimed to belong to the Panaramitee (i.e. petroglyphs in the Olary Province and the Laura region, Sturt's Meadows, N'Dhala Gorge, Bundaberg, Gatton, Einasleigh, Carbine Creek in the Mt Isa region, and Mt Cameron West [Preminghana], Maynard [1976, 1979]).
(ii) Other workers referred the petroglyphs to, or compared them with, the Panaramitee (e.g. Scott River and Edah, Clarke [1983]; petroglyph sites in central western Queensland, Morwood [1979, 1984]; petroglyphs on the Koolburra Plateau, Flood [1987]).
(iii) The sites appear to show similarities with petroglyph sites attributed to the Panaramitee (e.g. other petroglyphs in the Mt Isa region; Sundown Point in Tasmania; a site on the McArthur River in the Northern Territory).
(iv) The status of the sites as representatives of the Panaramitee is disputed (i.e. the Early Man petroglyphs in the Laura region by Rosenfeld [1981b]; N'Dhala Gorge by Forbes [1982, 1983]).
This selection process resulted in a sample of some 51 petroglyph sites from throughout Australia (Table 2, Fig. 2). In some cases regional data were available for analysis (e.g. the Olary Province, central western Queensland), whereas in other cases only isolated sites were available for particular regions (e.g. N'Dhala Gorge for central Australia, Fig. 2). The sample included 15 sites from South Australia, one site from New South Wales, two from the Northern Territory, 28 from Queensland, three from Tasmania and two from Western Australia (Table 2).
Results
The first three axes from the correspondence analysis, accounting for a total of 78.84% of the variation, show that there is inter-regional variation in terms of differing emphases on motifs or combinations of motifs (Franklin 2004: Figs. 3:12--3:14). This inter-regional variation is shown most clearly in the plot of the second and third axes (Fig. 3), which accounts for 36.07% of the total variation. Five major groups are apparent on the scatter diagram, resulting in the identification of five regions--central, Cape York, Carpentaria, Tasmania and central western Queensland.
[FIGURE 3 OMITTED]
At the bottom left of the scatter diagram (Fig. 3), two of the three Mt Isa sites (CC, FC), plus Einasleigh (E) and McArthur River (MR), form a group that might be labelled Carpentaria. This grouping is due to their large percentages of complex circles (v.10, see Table 3), and is consistent with a group identified in the cluster analysis (Fig. 4). The remaining Mt Isa site, Browns Creek (BC), was separated from the others, as was also the case in the cluster analysis (Fig. 4).
[FIGURE 4 OMITTED]
Except for the Laura River Crossing (LR), all of the Laura sites have grouped together at the bottom right of the scatter diagram on the basis of their proportions of dots/pits (v.5), and in some cases relatively larger percentages of complex lines (v.8), meandering lines/mazes (v.6) and radiating lines (v.12, Table 3). N'Dhala Gorge (NG), Browns Creek (BC) and Gatton (G) have also grouped with these sites on the same bases. In the cluster analysis (Fig. 4), Laura River (LR) grouped with the Rockholes (R), Wharton Hill (WH) and Morven (M), a cluster not inconsistent with the scatter diagram (Fig. 3).
The groups towards the top middle and right of the scatter diagram comprise predominantly central western Queensland and Olary sites. However, three of the sites in the latter region, Yunta (Y), Winnininnie 2 (Wi2) and Karolta 2 (K2), have been separated slightly from the others due to their similar percentages of crescents (v.11, Table 3). Oulnina (O) has also been separated, grouping with the Flinders Ranges site (FR) and Greenes Creek (GC) because of a predominance of simple circles (v.9). Both of these groupings are consistent with clusters identified on the dendrogram (Fig. 4). Despite these separations, there is still a relatively large, clear group of Olary sites in the top middle of the scatter diagram (Fig. 3).
The central western Queensland sites form an overall group towards the top right of the scatter diagram, while the group at the middle left consists of two of the Tasmanian sites and Edah (Ed) from Western Australia. These sites have emphases on both simple and complex circles (v.9 and v.10, Table 3), but with a greater percentage of the former. On the dendrogram (Fig. 4), Mt Cameron West (Preminghana, MCW) and Sundown Point (SP) clustered together, whereas Edah (Ed) clustered with two of the Mt Isa sites, Einasleigh (E) and the McArthur River (MR).
The group at the top left of the scatter diagram (Fig. 3) comprises Oulnina (O), the Flinders Ranges site (FR) and Greenes Creek (GC), while Bundaberg (B) is placed on its own. The first three sites contain simple circles (v.9) almost to the exclusion of all other motifs (Table 3). Bundaberg (B) also has a marked concentration on simple circles and small percentages of all other motifs (Table 3). On the dendrogram (Fig. 4), these four sites also form a cluster.
The regional groups on the scatter diagram (Fig. 3) are not quite as clear-cut in the cluster analysis (Fig. 4), where they are further subdivided. For example, the central western Queensland sites are divided into three major clusters on the dendrogram (Fig. 4). However, sites from a particular region still group initially with other sites from the same region (e.g. at the 0.56 level). If we take the clusters at an even broader level, say 1.12, then on the top arm are positioned all of the Tasmanian and Carpentarian sites, all except two of the Olary sites, and half of Cape York, while on the bottom two arms we have all of the central western Queensland sites, two Olary sites and the remainder of Cape York. A relatively high level of regionality is therefore still apparent.
The first three axes from the correspondence analysis, accounting for a total of 78.84% of the within-group variation (Fig 3; see also Franklin 2004: Figs. 3:12 and 3: 14), thus separated petroglyph sites from throughout the continent into regional groupings (see further discussion below). The cluster analysis, which considers 100% of the total within-group variation, separated the sites into clusters that reflect variation within regions to a greater extent, although still indicating a relatively high degree of regionality.
The results of the analyses presented in this paper and in Franklin (1992, 2004) can be summarised as follows:
1) The correspondence analysis identified regional groupings in terms of the emphasis on different motifs or combinations of motifs (Fig. 3).
2) The cluster analysis separated the sites into groups that reflect to a greater extent variation within regions (Fig. 4), although a relatively high degree of regionality is still apparent.
3) Sites that are identical in terms of the presence/ absence of motifs may be very different in terms of proportions of motifs (Franklin 1992, 2004: Tables 3:3, 3:10 and 3:11).
4) The only sites to contain the whole range of Panaramitee-type motifs occur in the centre of the continent, and elsewhere only at one other site (Bundaberg) (Franklin 1992, 2004).
5) Sites in the centre of the continent tend to be more heterogeneous than those at the periphery, and have a greater range of variation in terms of coefficients of diversity (Franklin 1992, 2004: Tables 3:12 and 3:13, Fig. 3:16).
Discussion
The most important finding from the analyses was that there are separate regional groupings of sites. These regions have an emphasis on different motifs or combinations of motifs within the overall range of 13 types. The motif emphases plus, in some cases, significant presences or absences of motifs can be used to define different regional groups, which have been labelled central, Cape York, Carpentaria, Tasmania, and central western Queensland (Table 4. See also Franklin 1992, 2004; Flood 1997). Each of the regional groups can be given a general description (Table 4), even though there is a certain amount of variation within them (see Fig. 4). A discussion of each regional group follows.
[TABLE 4 OMITTED]
Central region
The sites that make up this regional group (Table 4) correspond with the sites that were labelled 'classic' Panaramitee (Maynard 1976, 1979), as well as others in the Manunda-Yunta drainage basin (Nobbs 1981, 1982a--f, 1983a--g, 1984, pers. comm.). Following Edwards (1966a, 1971), Maynard claimed that there was a constancy of motif proportions at these sites, which she stated all came from the 'arid zone' (see above). The claim for a constancy of motif proportions at these sites can no longer be upheld when a larger Australia-wide site sample is analysed, as in this paper. There is a range of emphases on different motifs at different sites in the Olary Province (e.g. Karolta 2 [K2] has an emphasis on dots, the Rockholes [R] on 'macropod tracks', Table 3), so that motif proportions do not remain constant between sites, as is apparent from the quantitative data presented by other researchers (David 1988; Morwood 1985; Nobbs 1984). This range of emphases was reflected in the within-region variation identified in the cluster analysis (Fig. 4). Therefore, with the benefit of a larger sample, there is more variation between sites in the central regional group than was acknowledged in the original definition of the Panaramitee (Maynard 1976, 1979).
However, even with this internal variation, the central regional group can be distinguished as a whole from the others. In contrast to other regional groups, which tend to have a marked emphasis on one or two motifs, the central regional group tends to have smaller proportions of a range of different motifs. That is, motifs do not dominate sites to such an extent. The characteristic motifs are mostly those used to define the Panaramitee--'macropod' and 'bird tracks', 'human footprints', circles, dots, crescents, spirals (which were counted as complex circles in the analyses) and radiating lines.
Cape York Peninsula
In this regional group, 'bird tracks' are quite common at some sites, while 'macropod tracks', 'human footprints' and 'other tracks' are infrequent. Simple and complex circles are also not common (Table 4). Other features that allow Cape York to be distinguished as a regional group (Table 4) are the predominance of pits, and the greater relative frequency (compared to other areas) of mazes, simple and complex lines, and radiating lines at some sites.
The only site in Cape York Peninsula that does not fit the pattern identified for the Cape York regional group is the Laura River site (LR), which was consistently differentiated in the analyses from the other Cape York petroglyphs (Figs 3 and 4, Franklin 2004: Figs 3:12 and 3:14), and was therefore not included in this group (Table 4). Due to its larger proportion of figurative motifs (Table 3), Laura River tended to group closer to central western Queensland rather than with Cape York (Figures 3 and 4). However, 'macropod and bird tracks' and 'human footprints' are also well represented at the Laura River site, which indicates greater similarities with the central regional group than the other groups (Figures 3 and 4).
Although not located in Cape York Peninsula, the Gatton rock art site (G) linked with the Cape York region (Figs 3 and 4, Table 4) probably due to the predominance of dots/pits. This grouping demonstrates the widely dispersed similarities between the petroglyph sites included in the analyses, and the fact that there may be parallels between petroglyph sites across the continent.
Carpentaria
This regional group was found to have, in particular, a marked emphasis on complex circles and, to a lesser extent, simple circles (Table 4). Radiating lines and mazes are more common than at sites in the other regional groups with the exception of Cape York. 'Tracks' are not common, and of them, 'bird tracks' are the most frequent. The result that the Mt Isa sites form part of a distinct regional group confirmed claims made by Morwood (1979), and confirmed by Franklin (1996a, 2004), that the rock art of the Mt Isa region has distinctive features, in particular high proportions of complex circles.
As was the case with the Cape York regional group, there is one site located geographically close to the others that does not fit the pattern identified for its regional group. This site, Browns Creek (BC), was consistently differentiated in the analyses from the other petroglyphs (Figs 3 and 4, Franklin 2004: Figs 3:12 and 3:14), and was therefore also excluded from this regional group. Browns Creek grouped closer to the Cape York petroglyphs (Figs 3 and 4, Franklin 2004: Figs 3:12 and 3:14) and some of the sites of the central regional group (Fig. 4) than the Carpentaria sites due to its larger proportion of dots/ pits (Table 3).
Tasmania
This regional group is dominated by circles of all kinds, although simple circles are more frequent than complex circles (Table 4). Even though Greenes Creek (GC) was separated from Sundown Point (SP) and Mt Cameron West (MCW) on axis 3 in the correspondence analysis (Fig. 3), it can still be included in this regional group as it was part of the same overall cluster on the dendrogram (Fig. 4) and it contains simple circles almost to the exclusion of all other motifs. Greenes Creek is just more extreme on a theme that is common to the other Tasmanian sites.
Central western Queensland
There is considerable variation between sites in this regional group, although it was separated as a whole from the other regional groups on the correspondence analysis (Franklin 2004: Figs 3:12 and 3:14 especially). The sites in central western Queensland range from those with an emphasis on 'tracks' of all kinds and/or crescents, to those having a more marked emphasis on figurative 'vulva' motifs (Table 4). Some sites have even higher proportions of 'tracks' than some of the sites in the central regional group, especially 'human footprints' and 'other tracks' (in this case, unidentified tracks and hands).
Sites that clustered with more than one regional group
Some sites (Laura River Crossing, Browns Creek and Scott River) show similarities with more than one of the regional groups isolated in the analyses, and have therefore been excluded from Table 4. For example, Scott River (SR) showed similarities with sites in the Cape York regional group and central western Queensland (Fig. 3). This result again confirms the widespread similarity of the petroglyph sites analysed in this paper.
Discussion
Panaramitee petroglyphs have generally been represented in the literature as a relatively homogeneous corpus of art, although some variation was also claimed in terms of distinctive motifs ('fingerprints') that were unique to one or two sites (Edwards 1966a, 1971; Maynard 1976, 1979; see above).
The analyses presented in this paper have indicated that there are separate regional groups within the body of petroglyphs that can be referred to the Panaramitee, although the regional groups are not necessarily mutually exclusive. Therefore, the question arises as to whether or not the term 'Panaramitee style' is of any utility as a concept. The analyses have clearly shown that some of the claims made concerning this corpus of petroglyphs (Maynard 1976, 1979; Edwards 1966a, 1971) can no longer be supported. The claim that there is a constancy of motif proportions between sites in the Olary Province and central Australia can no longer be upheld when a larger Australia-wide sample of sites is analysed, an observation also supported previously by Nobbs' (1984) quantification of motifs in the Olary Province (and see also David 1988; Morwood 1985). The claim for a constancy of motif proportions also led to the argument for a homogeneous, pan-continental petroglyph 'style', the Panaramitee (Maynard 1976, 1979). This 'style' was said to be more homogeneous in South Australia, central Australia and western New South Wales (at the 'classic' sites) than at the periphery of the continent, where there were either not enough engraved motifs accumulated to result in the motif proportions found at the 'classic' sites, or they were not adequately described in the literature for motif proportions to be determined. The peripheral sites were included in the 'style' due to similarities in technique, motif range, and the forms of the petroglyphs.
According to these previous claims, the sites in the centre of the continent should be the most homogeneous. However, an analysis of diversity of motifs (Franklin 1992, 2004) showed that this is not the case. The most heterogeneous sites are found in the centre of the continent, and the greatest range of variation in coefficients of diversity also tends to occur here. Sites at the periphery of the continent (e.g. Laura, Tasmania and Western Australia) tend to be more homogeneous, showing marked concentrations on particular motif types. Although some of the peripheral sites have a smaller number of motifs, as Maynard claimed (e.g. Mt Cameron West, MCW, see Table 2), others, such as Bundaberg (B) have many more, but still tend to be homogeneous (Franklin 2004: Table 3:12).
Since it has been shown that there are regional variations in the emphasis on particular motifs, and that there is variation in the regions that were previously claimed to be the most homogeneous, are there any consistent features that draw together the sites analysed in this paper? The sites are all similar in terms of technique--pecking, and pecking and abrading in the case of some of the central western Queensland sites. The range of motifs is also, in general, similar across the continent, comprising 'animal tracks', 'non-figurative' motifs, and a few different types of figurative motif at some sites, such as the 'vulvas' that are found in large percentages at some of the central western Queensland sites. The occurrence of a generally similar motif range at the sites in the Australia-wide sample is not to say that all sites contain the full (though still relatively restricted) range of 13 motif types, as an analysis of presence/absence of motifs indicated (Franklin 1992, 2004). There are significant absences of motifs from some regional groups (e.g. 'macropod' and 'other tracks' from Tasmania), as well as different regional emphases on motifs. Generally, though, the motif emphasis in this corpus of art can be distinguished from that found in, for example, the simple figurative styles, where the emphasis is on (other) figurative motifs, rather than tracks.
Therefore, in these very general terms, there are similarities which hold this corpus of art together, and which might be used to define the 'style'. However, except for motif range and the widespread distribution of the sites throughout the continent, these are not the features used to define the 'Panaramitee style'. The term should therefore only be used in a very broad sense. It should no longer encompass the concept of homogeneity over a wide area of Australia, as has been the case in much of the literature (e.g. Maynard 1976, 1979; White 1984; White and Habgood 1985; White and O'Connell 1982).
The concept of the Panaramitee as a 'style' has been debated (e,g. Bednarik 1988, 1994; 1995; Maynard 1988; Rosenfeld 1991). Flood (1997) prefers to use the term 'tradition' to describe the Panaramitee, as it '... implies culture contact and continuity but not necessarily close similarity or uniformity'. (p. 179). Use of this term is appropriate, as it covers '... longterm continuity in either individual technologies or attributes' and 'a temporally ordered series of archaeological phases or cultures that show cultural similarities to each other' (Bahn ed. 2001: 452), and appears to fit with the results obtained in the analyses presented above. This concept of 'tradition' also parallels the use of the term to describe Australian stone tool technologies (e.g the 'core tool and scraper tradition'). 'Tradition' is more appropriate than the term 'style', as the latter term refers to '... a highly specific and characteristic manner of doing something ...' which is '... always peculiar to a specific time and place ...' (Sackett 1977: 370), and in the case of the petroglyph sites analysed in this paper, we are probably dealing with chronological as well as spatial variation.
Like Flood (1997), I also prefer to retain the term 'Panaramitee' for this tradition of rock art, as many Australian rock art researchers and archaeologists have found it a useful term (e.g. Clegg 1988, 1992; David 1988; David and Chant 1995; Franklin 1984, 1986, 1989, 1991, 1992; McDonald 1982, 1983, 1988; Morwood 1979, 1984, 1988, 2002; Smith 1992), and it follows the standard archaeological practice of labelling a tradition or culture after a type-site which contains an extensive and typical example of the material in question (e.g. the Mousterian after the type-site of Le Moustier).
However, it should be made clear that the label 'Panaramitee' only refers to the manifestation of the tradition within Australia (see also Flood 1997). As Bednarik (1995) has shown, some of the rock art of other continents also resembles the Panaramitee. This observation is made clear by the results of the blind test undertaken by Bednarik (1995), whereby eight Australian rock art specialists were sent copies of unlabelled drawings of petroglyphs and asked to identify where they came from. Although the petroglyphs were all from North and South America, Africa and Europe, and the score of wrong answers was 98%, there was nevertheless internal consistency in the answers provided by the specialists, who identified the petroglyphs as being from particular sites in Australia. The results of Bednarik's test therefore tend to confirm the widespread similarities of this form of rock art within Australia, as indicated by the results of the multivariate analyses presented in this paper, and in no way negate the use of the term 'Panaramitee tradition' within Australia, even though there may be world-wide resemblances. As Bednarik (1995: 4) has asked 'why is there such stylistic consistency in pre-iconic petroglyphs the world over?', an interesting question which is beyond the scope of this paper.
Discontinuous Dreaming networks: a model for the explanation of patterns within Australian pre-Historic petroglyphs
The following functional model is proposed as one possible framework to account for the regional differences in motif choice observed in the multivariate analyses. It has yet to be thoroughly tested and is presented as a hypothesis.
Dreaming tracks
The model suggests that the regionality found within the Panaramitee reflects the depiction in petroglyphs of the motifs relevant to the particular parts of Dreaming tracks on which people reside or to which they own the ceremonial rights. It is a predictive model that can be assessed and tested as new petroglyph sites are rediscovered and Dreaming tracks mapped in detail.
The significance of 'Dreamings' and 'Dreaming tracks' or 'song lines' has been stressed in many studies of Aboriginal cosmology and land ownership (e.g. Chatwin 1987; David 2002; Elkin 1934; Gunn 1997, 2003; Layton 1992; Morphy 1983; Moyle 1983; Munn 1973; Spencer and Gillen 1938; Strehlow 1978; Sutton 1988a--c, 1990). Dreaming tracks reflect the activities of the Dreamtime ancestors during the creative era as they emerged from the earth and travelled across the country along lengthy tracks or circled within more narrowly defined regions. During their travels, the ancestors created various geographical features still present in the landscape before finally re-entering the earth.
Dreaming tracks sometimes covered considerable distances, a thousand kilometres or more, and in one example spanned the entire continent from the Roper River to Port Augusta, going through a number of transformations in the process (Sutton 1990). Therefore, as the Dreaming tracks show regional variation, the motifs used to depict them will also vary regionally. The greater diversity of motifs at petroglyph sites in the centre compared with the periphery of the continent concurs with the intersection of numerous Dreaming tracks in central Australia. Similarly, the result that the total corpus of engraved motifs occurs mainly in the centre may be due to the often considerable spatial extent of its Dreaming tracks, compared with the more spatially restricted tracks at the periphery of the continent, and is also probably combined with a greater use of the motifs over time (potentially greater than 10 000 years, as indicated by the dating results from Puritjarra, Smith 1996; Rosenfeld and Smith 2002). Both of these results also concur with a possibly greater use of the central Australian petroglyphs as aggregation sites compared to those at the periphery (see Conkey 1980).
There are differing interpretations for the antiquity of Dreaming tracks. Some tracks may be of considerable antiquity, others more recent. Possible evidence for the former appears in the following statement:
... many of the myths are centrally concerned with underpinning the rights of particular human groups to specific areas of land, and they often symbolize relationships of alliance or political disjunction between territorial groups, especially in the case of travelling myths. Even these myths to some extent lead a life of their own, not quite matching with current land tenure or the current state of intergroup politics. Sutton (1988b: 18; my emphasis)
The observation that the travelling myths, expressed on the ground by Dreaming tracks, do not match exactly with current land tenure indicates that the Dreaming tracks may refer to a past situation. In other words, while land tenure may change, it seems that Dreaming tracks may not, and may therefore possibly be long-standing. Consequently, their comparison with petroglyphs, whose age is also unknown but which may also be of considerable antiquity (Rosenfeld 1981a, 1981b; Rosenfeld and Smith 2002), is at least worthy of further consideration.
Similarly, Mulvaney and Kamminga (1999) implied that Dreaming tracks may have some antiquity when they pointed out that the direction from which the Dreaming beings travelled in northern Australia, from across the sea to Arnhem Land, is revealing of the sorts of ancient links overseas that archaeologists have claimed. They further argued that in general, rock art is '... the most valuable type of material evidence for reconstructing the prehistory of Dreaming beliefs and practices ...' (p. 359), citing by way of example the beginnings of Rainbow Serpent imagery in Kakadu rock paintings some 6000 years ago as evidence of continuous religious beliefs from that distant time (Tacon et al. 1996). Other researchers have pointed to the continuous occupation of Australia for a vast period of time and asked: 'Can one doubt the corresponding antiquity of the Dreamtime beliefs, considering how universal were the basics of Aboriginal cosmology?' (Moore 1988: 32).
Given that the Dreaming track hypothesis proposed here accounts for the variations observed within the Panaramitee-tradition petroglyphs analysed in this paper, and that there is evidence the petroglyphs date back some 14 000 years at Early Man and Puritjarra, the associated Dreaming tracks can also be interpreted as having similar antiquity.
On the other hand, some Dreaming tracks are interpreted as being more recent, for example in the Victoria River District and at a few sites in central Australia. In the former region, attempts were made to trace back archaeologically the history of rock paintings, which express the current belief system relating to Dreaming tracks (David 2002; David et al. 1994). Excavations at various sites throughout the region demonstrated that painting activity dated mostly to the last 1400 years or so, implying a recent date for the associated Dreaming tracks. Similarly, at Therreyererte in central Australia, an abrupt increase in the number of stone artefacts about 500-600 years ago was interpreted as marking the commencement of inter-tribal ceremonies at this major site on a native cat Dreaming track, and possibly the origin of the Dreaming mythology (Smith 1988).
Further excavations will be necessary at other sites whose recent history and associated mythology are already known before we can attempt to establish the antiquity of Dreaming tracks (and at present, only a few sites have been excavated, all from central Australia, David 2002). However, it will still be incumbent upon excavators to establish links between the excavated material, artistic activities and purported Dreaming tracks, and firm demonstrations of this nature may not always be possible.
Dreaming tracks frequently extended across group boundaries, facilitating meetings between local groups and travelling groups for gift exchange and rituals associated with the myth relevant to the track. The use of similar motifs across vast areas at sites related to particular Dreamings is well documented (Ross 1997), and suggested that the shared understanding of motif forms provided a means for negotiating the rights and obligations of travellers along the tracks. The repetition of motifs between regions implied a shared knowledge that assured travellers of their right to move through the territory and which established an affilial relationship between the owners and the travellers. It is also therefore not surprising that Dreaming tracks frequently correlated with the trade routes documented in recent times (Ross 1997), suggesting that both formed a means for the interaction of people across the landscape, and for the diffusion of similar motifs across vast areas.
The emphasis on Dreaming tracks seems to have been particularly pronounced in the central desert regions of Australia, although they were also important in Arnhem Land and the Kimberleys, but less so in Cape York (Peterson 1986). The significance placed upon Dreaming tracks in arid regions meant that boundaries between groups tended to be loosely defined, also a function of the uncertainty of resources in these regions. So, for example, for the Kurintji living on the margins of the arid zone, handover points were located on each Dreaming track where one estate group's primary interest gave way to another's. Since each group normally 'ran' more than one track across its estate, ownership of estates could be defined by the directions in which the collection of sections of tracks ran (Peterson 1986). As many Dreaming tracks may be more than 1000 km long, many sites will have associations with other sites well outside the particular 'tribal' territory, providing a non-genealogical link between different individuals, clans and sites many kilometres apart. The landscapes of different groups were therefore linked into an even broader cultural landscape encompassing much of arid and semi-arid Australia (Gunn 1995, 1997, 2000, 2003).
Dreamings rarely travelled in clear, straight lines from one place to the next, and often their paths were very convoluted. In many cases, they travelled from major waterholes to minor ones, clearly avoiding the next major waterhole, which is invariably related to another Dreaming being travelling in another direction (Gunn 2003), thus complicating both the mapping of Dreaming tracks and their interpretation.
Art motifs were indissolubly linked with Dreaming tracks and ownership of particular parcels of land, as the following examples show. Most recent artistic productions of Aboriginal people relate to depictions of Dreamtime myths relating the journeys of the ancestral beings, and often constitute maps of the country through which the ancestors travelled. The art depicts specific topographical features on the Dreaming track that relates to the artist. This depiction of country applies to artistic productions in all non-rock art media: the engraved and painted sacred objects, body paintings, ground paintings, sand drawings (Munn 1973), bark paintings (Morphy 1977b, 1983), contemporary Western Desert acrylic paintings (Anderson and Dussart 1988), and even the toas of Lake Eyre (Morphy 1977a; Sutton 1988a). It may equally well apply to pre-Historic rock paintings and petroglyphs (see Franklin 2004), particularly when the similarities between the motifs in the different media are taken into consideration.
Indeed, Gunn (2003) noted the co-occurrence of rock art and Dreaming tracks, where in Arrernte territory in central Australia, all rock art recorded to date occurred at sites associated with a specific Dreaming. However, not all Dreaming sites along Dreaming tracks necessarily contained rock art and Gunn (2000, 2003) could find no consistent graphic associations between motif type, colour or technique at rock art sites along the path taken by a specific Dreaming being. For instance, although a distinctive shield motif was confined broadly to a north-south distribution at sites to the south of Alice Springs, the five site complexes at which it occurs all have different Dreamings, and this distribution is cut at each complex by Dreamings travelling eastwest, southeast-northwest or southwest-northeast. Although these observations might appear to contradict the general tenets of the model outlined above, Gunn stressed that at present his conclusions only applied to Arrernte rock art, so that the general explanation provided by the model still stands.
Elkin (1949) suggested that at least part of the extensive trade network that extended between north-west central Queensland and Lake Eyre was the track of the Mura Mura Dreaming beings, which were associated with the dog cult-totem. The engraved motifs associated with this myth are concentric arcs, lines and circles, all of which are found in the petroglyphs and paintings of northwest Queensland (Ross 1997), and are common motifs in the Panaramitee tradition. The frequent coincidence between trade routes, Dreaming tracks and of necessity waterholes in Aboriginal Australia was emphasised by Ross (1997).
In western New South Wales, Clegg (1993a) observed similarities between engraved motifs at the Panaramitee site of Gap Hills and a red kangarooeuro myth cycle that also showed links with stories from the Flinders Ranges. However, differences with the petroglyphs in the nearby sites of Sturt's Meadows, Mootwingee and Euriowie suggested that their significances probably related to other Dreamings (see also Clegg 1993b), a case where sites in close proximity to each other need not necessarily lie on the same Dreaming track.
Morwood (1979, 2002) noted the association between a particular painted rock art motif and a Dreaming track in the central Queensland highlands. He noted that the distribution of a distinctive paired tortoise motif, found at only four sites associated with water sources on the upper Warrego and upper Nogoa Rivers, ran parallel to the paths taken by two Dreaming beings, a snake and a goanna. The tortoise sites were also found to occur on a line some 75 km long that marked a territorial boundary between Aboriginal groups in these regions (see Morwood 2002: Fig. 8.24). Morwood argued that the similarities between known Dreaming tracks and the linear distribution of a distinctive rock art motif suggested that the rock art sites marked points of significance along the track.
In north-east Arnhem Land, the situation regarding the artistic depiction of Dreaming tracks appears to be less complex than it is in the desert. For example, Morphy (1983) observed that among the Yolngu of north-east Arnhem Land, each area of land is associated mainly with the activities of one set of ancestral beings. Although Dreaming tracks frequently intersect, only one of the basic design types found in Yolngu bark paintings is associated with any given area of land. Similarly, in the western Kimberleys, each clan had a unique ancestral hero whose Dreaming track ended in that clan's estate, so that the tracks in this region tended to be relatively short (Layton 1985). If similar situations pertained in the past, the result would be restricted spatial distributions of particular motifs in different regions, as observed for the peripheral petroglyph sites in the analyses. However, the occurrence of an overall basic shared repertoire of motifs at most petroglyph sites might ultimately reflect their links, via the Dreaming tracks and trade routes, with petroglyph sites in the centre of the continent.
Dreaming tracks, boundary maintenance and information exchange
The links forged by Dreaming tracks in central Australia are also consistent with the function of style in boundary maintenance and information exchange, where open social networks are required in the harsh, unpredictable environments of the arid zone, where Panaramitee-tradition petroglyphs tend to occur. The greater degree of homogeneity and the widespread distribution of the Panaramitee tradition, compared to the 'simple figurative' and 'complex figurative', corresponded with the possible function of the art to cement social networks between widely dispersed groups, enabling access to the resources of other groups during periods of shortage. By contrast, the greater degree of variation between the different styles of the 'simple figurative' and 'complex figurative' and their spatially more restricted distribution suggest that they may have acted as territorial markers in the resource-rich areas in which they are found (see also Franklin 2004).
Studies of territorial organisation in recent Aboriginal society (e.g. Peterson 1986) support this interpretation. In the arid zone where resources can be unpredictable and precarious, Aboriginal groups tended to be small, and social and territorial networks were extended over a large area facilitating easy movement across fluid boundaries. These adaptations were crucial to survival so that during periods of shortage in particular regions, access to the resources of other groups was ensured. People in these regions were related to everyone else through a moiety or section system (Peterson 1986; see also Yellen and Harpending 1977).
By contrast, in resource-rich areas such as coastal Arnhem Land, the tropical woodlands and southeastern Australia, the environment can support a significantly larger population density than is possible in arid regions. Consequently, groups tended to be larger, territories smaller and more clearly defined (although territories in central Australia may also be well defined), a greater degree of boundary maintenance was required with elaborate procedures for crossing from one to another, and the groups were corporate in nature. People in resource-rich areas were not related to each other through a moiety or section system, but maintained ties mainly with the clans of father, mother, wife and so on (Peterson 1986).
The differing functions for style, as indicative of open and closed social networks respectively, were exploited in Gamble's (1982, 1983, 1986, 1991) explanation for the similarities between Venus figurines across a large area of Europe (see also Barton et al. 1994; Davidson 1997a), by Jochim (1983) in his model for the restricted distribution of cave art in the resource-rich areas of the Dordogne and Cantabria during a period of climatic deterioration, and by Brandt and Carder (1987) to explain an increase in the diversity and distribution of pastoral rock art in the Horn of Africa during a period of changing climatic conditions.
They have also been usefully applied to explain the differing degrees of variation in Australian rock art, both on a continent-wide and a within-region basis (Cole and David 1992; David 1991, 1992, 2002; David and Chant 1995; David and Cole 1990; David and Lourandos 1997; Davidson 1997b; Flood 1987; Gunn 2002; Lewis 1988, 1997; McDonald 1992, 1993, 1994, 1998, 1999; Morwood 1979, 1980, 1981, 1984, 1992, 2002; Morwood and Hobbs 1995a, 1995b; Ross 1997; Smith 1989, 1996; Veth and McDonald 2002).
These models have been complemented by studies that attempt to map the movements of materials across the landscape in order to pinpoint the antiquity of trade and other social networks. For example, the widespread exchange of pearl and baler shell from the north-west coast of Australia into the Great Sandy Desert may indicate an antiquity of no more than a few thousand years for such networks, as all specimens of shell so far found in desert regions are no older than this (Veth 2000; Smith and Veth 2004).
Similarly, geochemical 'fingerprinting' of pigment from geological samples can enable the determination of whether ochre recovered from excavations, or samples from particular paintings, are from specific sources, allowing inferences of the extent of past trade routes and alliance networks to be drawn from the geographical distribution of pigments (David et al. 1993; David et al. 1995; Goodall et al 1996; Smith et al. 1998). Furthermore, in situ analysis enables ochre from rock slabs that may have fallen from shelter walls, as well as that from different art panels, to be matched with ochre pieces in excavated deposits, and incorporated with research like that cited above which provenances ochre to particular sources (O'Connor and Fankhauser 2001).
It is clear from these studies that sourcing of pigments and in situ analysis of ochre have great potential for monitoring past exchange systems and alliance networks, and the way in which these may have changed over time.
Continuous and discontinuous symbolic systems
The use of non-figurative motifs in symbolic systems also helps to explain the persistence of the overall pattern of similarity in the Panaramitee tradition over possibly a prolonged period of time (see above). The potential for individual non-figurative motifs to have a range of different meanings suggests it is possible that the meanings of motifs might have changed over time and distance according to context, while the morphology of the motifs might not. This possibility is allowed for by the extreme simplicity of motifs in the Panaramitee tradition, compared to the greater complexity of motifs in the 'simple figurative' and 'complex figurative', where meanings for motifs tend to refer to only one class of phenomena and to be more specific than is the case for non-figurative motifs (Munn 1966).
Some examples will serve to highlight these points (see Tables 5 and 6). Among the Warlpiri people of central Australia, motifs '... cover highly general categories, each of which includes a variety of different classes of phenomena' (Munn 1966: 940). For instance, the circle can be used to specify a waterhole, fire, fruits of different kinds, and so on (Table 5). In ordinary usage, only one of these meanings is relevant at a time, so that people using the visual system select in each instance one specific meaning from the range of possible meanings. The visual categories comprehended by each motif in the Warlpiri system are so broad that an indefinite number of phenomena can be specified. The result is that new meanings can be readily incorporated into the system without the need to derive new motifs. Meaning ranges of this type are labelled discontinuous (Munn 1966; Table 5).
[TABLES 5-6 OMITTED]
By contrast, in Yirkalla bark paintings of northeast Arnhem Land, the meaning ranges of motifs are labelled continuous because they only refer to one phenomenon and not to several different classes of phenomena, as among the Warlpiri. Therefore, snake and tree motifs (Table 6) are probably used for all varieties of snakes and trees recognised by the people. However, even though different species of tree may be specified by the 'tree' motif, the same motif cannot also be used for waterholes, fires, and so on, as is the case with the circle motif among the Warlpiri.
This is not to say that motifs with continuous meaning ranges do not occur among the Warlpiri, or that discontinuous categories do not occur in the Yirkalla system. For example, most tracks (which are figurative motifs) have continuous ranges in Warlpiri sand storytelling, in that they carry a standard signification: they convey the species, number and direction of individuals moving across the country. Nevertheless, the use of highly general visual categories with discontinuous meaning ranges is particularly emphasised in central Australia, while motifs with continuous meaning ranges receive greater emphasis in Yirkalla bark paintings (Munn 1966, 1973).
Using motifs that have discontinuous meaning ranges, as in the Warlpiri system, means that it is possible to operate with a much smaller number of elements than in a system that relies largely on continuous categories, because the number of classes of meaning in the former can be expanded without increasing the repertoire of elements. However, in a visual system that emphasises continuous meaning ranges, new motifs must be added to incorporate new meanings (Munn 1966). These observations go some way towards explaining the greater degree of variation within the 'simple figurative' than within the Panaramitee tradition. With the simple and complex figurative genres, we could be dealing with motifs which have continuous meaning ranges, while motifs at Panaramitee tradition petroglyph sites may well have discontinuous meaning ranges. It will not be possible, however, to determine the particular meanings subsumed by each motif at pre-Historic art sites.
Conclusions
This paper has explored the degree of variation within a corpus of Australian pre-Historic petroglyphs, the Panaramitee, which has been represented in the literature as a homogeneous pan-Australian 'style'. It was found that there was inter-regional variation in terms of differing emphases on motifs within an overall restricted range across the continent. Five major regional groups were defined--central, Cape York, Carpentaria, Tasmania and central western Queensland. A greater diversity of motifs was identified in the central regional group compared to more marked concentrations on particular motifs in the other regional groups, such as simple circles in Tasmania. Petroglyph sites in the centre of the continent were therefore more diverse than those located at the periphery.
Following Flood (1997), it is suggested that the term 'Panaramitee style' should be replaced in the literature by the term 'Panaramitee tradition', as 'tradition' covers the long-term continuity in motif types that occurs in the petroglyph sites analysed in this paper. The regionality found within the Panaramitee tradition is interpreted as reflecting the depiction in petroglyphs of the motifs relevant to particular parts of Dreaming tracks, which sometimes covered considerable distances across the continent. Dreaming tracks frequently extended across group boundaries and also correlated with trade routes, suggesting that both formed a means for the interaction of people across the landscape, and for the diffusion of similar motifs across vast areas.
These observations are also consistent with the open social networks required in the harsh, unpredictable environments of the arid zone, where Panaramiteetradition petroglyphs tend to occur. The greater degree of homogeneity and the widespread distribution of the Panaramitee tradition, compared to the 'simple figurative' and 'complex figurative', corresponded with the possible function of the art to cement social networks between widely dispersed groups, enabling access to the resources of other groups during periods of shortage.
The use of non-figurative motifs in symbolic systems also helps to explain the persistence of the overall pattern of similarity in the Panaramitee tradition over possibly a prolonged period of time. The potential for non-figurative motifs to have a range of different discontinuous meanings suggests it is possible that the meanings of motifs might have changed diachronically according to context, while the morphology of the motifs might not.
The hypothesis for the variation within the Panaramitee tradition that has been proposed in this paper can be referred to as the discontinuous Dreaming network model. In deriving this model, it has been shown that analyses of spatial variation in Australian rock art can yield profitable results that can be used in constructing models of Australian pre-History.
Note: The views expressed in this paper do not represent those of the Queensland Environmental Protection Agency.
Final MS received 5 December 2006.
REFERENCES
Anderson, C. and F. Dussart 1988. Dreamings in acrylic: Western Desert art. In P. Sutton (ed.), Dreamings: the art of Aboriginal Australia, pp. 89-142. Penguin, Melbourne.
Bahn, P. G. (ed.). The penguin archaeology guide. Penguin Books, Ringwood, Victoria.
Barton, C. M., G. A. Clark and A. E. Cohen 1994. Art as information: explaining Upper Palaeolithic art in western Europe. World Archaeology 26: 185-207.
Bednarik, R. G. 1988. Comment on F. D. McCarthy, 'Rock art sequences: a matter of clarification'. Rock Art Research 5: 22-24.
Bednarik, R. G. 1994. Rock art as a cultural determinant. AURA Newsletter 11(2): 9-12.
Bednarik, R. G. 1995. Taking the style out of the Panaramitee style. AURA Newsletter 12(1): 1-5.
Bednarik, R. G. 2002. First dating of Pilbara petroglyphs. Records of the Western Australian Museum 20: 415-429.
Bolviken, E., E. Helskog, K. Helskog, I. M. Holm-Olsen, I. Solheim and R. Bertelsen 1982. Correspondence analysis: an alternative to principal components. World Archaeology 14: 41-60.
Brandl, E. J. 1982. Australian Aboriginal paintings in western and central Arnhem Land: temporal sequences and elements of style in Cadell River and Deaf Adder Creek. Australian Institute of Aboriginal Studies, Canberra.
Brandt, S. A. and N. Carder 1987. Pastoral rock art in the Horn of Africa: making sense of udder chaos. World Archaeology 19: 194-213.
Chaloupka, G. 1993. Journey in time: the world's longest continuing art tradition. The 50,000-year story of the Australian Aboriginal rock art of Arnhem Land. Reed, Chatswood, New South Wales.
Chatwin, B. 1987. The songlines. Picador, London.
Clarke, J. 1983. An Aboriginal engraving site in the south-west of Western Australia. Records of the Western Australian Museum 11: 63-67.
Clegg, J. 1977. A Sausserian model of prehistoric art. The Artefact 2: 151-160.
Clegg, J. 1981a. Notes towards mathesis art. Clegg Calendars, Sydney.
Clegg, J. 1981b. Report on preliminary work at Eight-Mile Creek, western New South Wales. Privately published paper.
Clegg, J. 1984a. Enquetes preliminaires a Sturt's Meadows: un gisement Australien de petroglyphes du style Panaramitee. L'Anthropologie 88: 595-610.
Clegg, J. 1984b. Comment on M. F. Nobbs, 'Rock art in Olary province, South Australia'. Rock Art Research 1: 111-117.
Clegg, J. 1986. Prehistoric pictures as evidence about religion. World Archaeological Congress, Southampton, Cultural attitudes to animals including birds, fish and invertebrates, Volume 3, precirculated papers. Unwin Hyman, London.
Clegg, J. 1987. Style and tradition at Sturt's Meadows. World Archaeology 19: 236-255.
Clegg, J. 1992. Rules of similarity in Panaramitee engraving sites. In J. McDonald and I. P. Haskovec (eds), State of the art: regional rock art studies in Australia and Melanesia. Proceedings of Symposium C, 'Rock art studies in Australia and Oceania' and Symposium D, 'The rock art of northern Australia' of the First AURA Congress, held in Darwin in 1988, pp. 32-38. Occasional AURA Publication 6. Australian Rock Art Research Association, Melbourne.
Clegg, J. 1993a. Gap Hills: a new simple figurative style engraving site in western New South Wales. Rock Art Research 10: 55-58.
Clegg, J. 1993b. Style at Sturts Meadows and Gap Hills. In M. Lorblanchet and P. G. Bahn (eds), Rock art studies: the post-stylistic era or where do we go from here?, pp. 115-126. Oxbow Monograph 35, Oxbow Books, Oxford.
Clegg, J. n.d. Preliminary investigations at Sturt's Meadows: an Australian petroglyph site in the Panaramitee style. Privately published paper.
Clegg, J. in prep. Sturt's Meadows: ancient Aboriginal engravings in arid Australia. British Archaeological Reports International Series.
Cole, N. and B. David 1992. 'Curious drawings' at Cape York Peninsula. An account of the rock art of the Cape York Peninsula region of north-eastern Australia and an overview of some regional characteristics. Rock Art Research 9: 27-36.
Conkey, M. W. 1980. The identification of prehistoric hunter-gatherer aggregation sites: The case of Altamira. Current Anthropology 21: 609-620.
Crawford, I. M. 1968. The art of the Wandjina. Oxford University Press, Melbourne.
Crawford, I. M. 1977. The relationship of Bradshaw and Wandjina art in north-west Kimberley. In P. J. Ucko (ed.), Form in indigenous art: schematisation in the art of Aboriginal Australia and prehistoric Europe, pp. 357-369. Australian Institute of Aboriginal Studies, Canberra.
David, B. 1988. Comment on F. D. McCarthy, 'Rock art sequences: a matter of clarification'. Rock Art Research 5: 22-24.
David, B. 1991. Fern Cave, rock art and social formations: rock art regionalisation and demographic models in southeastern Cape York Peninsula. Archaeology in Oceania 26: 41-57.
David, B. 1992. Cave paintings of the Mitchell-Palmer Limestone Belt: Preliminary results. In J. McDonald and I. P. Haskovec (eds), State of the art: regional rock art studies in Australia and Melanesia. Proceedings of Symposium C, 'Rock art studies in Australia and Oceania' and Symposium D, 'The rock art of northern Australia' of the First AURA Congress, held in Darwin in 1988, pp. 76-82. Occasional AURA Publication 6. Australian Rock Art Research Association, Melbourne.
David, B. 2002. Landscapes, rock-art and the Dreaming: an archaeology of preunderstanding. Leicester University Press, London.
David, B. and D. Chant 1995. Rock art and regionalisation in north Queensland prehistory. Memoirs of the Queensland Museum 37(2).
David, B. and N. Cole 1990. Rock art and inter-regional interaction in northeast Australian prehistory. Antiquity 64: 788-806.
David, B. and H. Lourandos 1997. 37,000 years and more in tropical Australia: investigating long-term archaeological trends in Cape York Peninsula. Proceedings of the Prehistoric Society 63: 1-23.
David, B., E. Clayton and A. Watchman 1993. Initial results of PIXE analysis on northern Australian ochres. Australian Archaeology 36: 50-57.
David, B., A. Watchman, R. Goodall and E. Clayton 1995. The Maytown ochre source. Memoirs of the Queensland Museum 38: 441-445.
David, B., I. McNiven, V. Attenbrow, J. Flood and J. Collins 1994. Of Lightning Brothers and White Cockatoos: dating the antiquity of signifying systems in the Northern Territory, Australia. Antiquity 68: 241-251.
Davidson, I. 1997a. The power of pictures. In M. W. Conkey, O. Soffer, D. Stratmann and N. G. Jablonski (eds), Beyond art: Pleistocene image and symbol, pp. 125-159. Memoirs of the California Academy of Sciences 23, California Academy of Sciences, San Francisco.
Davidson, I. 1997b. Review of B. David and D. Chant, 'Rock art and regionalisation in north Queensland prehistory'. Archaeology in Oceania 32: 216-217.
Doran, J. E. and F. R. Hodson 1975. Mathematics and computers in archaeology. Edinburgh University Press, Edinburgh.
Dorn, R. I. 1996. Uncertainties in 14C ages for petroglyphs from the Olary province, South Australia. Archaeology in Oceania 32: 214-215.
Dorn, R. I. and M. Nobbs 1992. Further support for the antiquity of South Australian rock engravings. Australian Aboriginal Studies 1992(1): 56-60.
Dorn, R. I., M. Nobbs and T. A. Cahill 1988. Cation-ratio dating of rock-engravings from the Olary Province of arid South Australia. Antiquity 62: 681-689.
Edwards, Robert 1964. Rock engravings and stone implements of Pitcairn Station, north-eastern South Australia. Records of the South Australian Museum 14: 643-662.
Edwards, Robert 1965a. Rock engravings and incised stones: Tiverton Station, north-east South Australia. Mankind 6: 223-231.
Edwards, Robert 1965b. Rock engravings and Aboriginal occupation at Nackara Springs in the north-east of South Australia. Records of the South Australian Museum 15: 9-27.
Edwards, Robert 1966a. Comparative study of rock engravings in South and central Australia. Transactions of the Royal Society of South Australia 90: 33-38.
Edwards, Robert 1966b. Field work in the Northern Territory. Friends of the South Australian Museum 3: 11-15.
Edwards, Robert 1968a. Prehistoric rock engravings at Thomas Reservoir, Cleland Hills, western central Australia. Records of the South Australian Museum 15: 647-670.
Edwards, Robert 1968b. Unique rock art in central Australia. Hemisphere 12: 2-7.
Edwards, Robert 1971. Art and Aboriginal prehistory. In D. J. Mulvaney and J. Golson (eds), Aboriginal man and environment in Australia, pp. 356-367. Australian National University Press, Canberra.
Edwards, Ron 1967. Cave art in the Newcastle Ranges. Northern Folk 18: 6-7.
Elkin, A. P. 1934. Cult-totemism and mythology in northern South Australia. Oceania 5: 171-192.
Elkin, A. P. 1949. The origin and interpretation of petroglyphs in south-east Australia. Oceania 20: 119-157.
Flood, J. M. 1987. Rock art of the Koolburra Plateau, north Queensland. Rock Art Research 4: 91-126.
Flood, J. M. 1997. Rock art of the Dreamtime: images of ancient Australia. HarperCollins, Sydney.
Forbes, S. 1982. Aboriginal rock engravings at N'Dhala Gorge: an analysis of a central Australian 'Panaramitee style' rock art site. B. Lett. thesis, Department of Prehistory and Anthropology, Australian National University.
Forbes, S. 1983. Aboriginal rock engravings at N'Dhala Gorge, Northern Territory. In M. Smith (ed.), Archaeology at ANZAAS 1983, pp. 199-213. Western Australian Museum, Perth.
Franklin, N. R. 1984. Of !macropods and !men: an analysis of the simple figurative styles. B.A. (hons) thesis, Department of Anthropology, University of Sydney.
Franklin, N. R. 1986. Stochastic vs. emblemic: an archaeologically useful method for the analysis of style in Australian rock art. Rock Art Research 3: 121-140.
Franklin, N. R. 1989. Research with style: a case study from Australian rock art. In S. J. Shennan (ed.), Archaeological approaches to cultural identity, pp. 278-290. Unwin Hyman, London.
Franklin, N. R. 1991. Explorations of the Panaramitee style. In P. Bahn and A. Rosenfeld (eds.), Rock art and prehistory: papers presented to Symposium G of the AURA Congress, Darwin 1988, pp. 120-135. Oxbow Books, Oxford.
Franklin, N. R. 1992. Explorations of variability in Australian prehistoric rock engravings. PhD thesis, Department of Archaeology, La Trobe University.
Franklin, N. R. 1996a. An analysis of rock engravings in the Mt Isa region, northwest Queensland. In P. Veth and P. Hiscock (eds), Archaeology of northern Australia, pp. 137-149. Tempus 4, Anthropology Museum, University of Queensland, St Lucia.
Franklin, N. R. 1996b. Rock art in Australia and New Guinea 1990-1994: research, management and conservation. In P. G. Bahn and A. Fossati (eds), Rock art studies news of the world 1: recent developments in rock art research (Acts of Symposium 14D at the NEWS95 World Rock Art Congress, Turin and Pinerolo, Italy), pp. 145-161. Oxbow Monograph 72, Oxbow Books, Oxford.
Franklin, N. R. 2003. Rock art in Australia and New Guinea 1995-1999: research, management and conservation. In P. G. Bahn and A. Fossati (eds), Rock art studies news of the world 2: developments in rock art research 1995-1999, pp. 133-146. Oxbow Books, Oxford.
Franklin, N. R. 2004. Explorations of variability in Australian prehistoric rock engravings. BAR International Series 1318, British Archaeological Reports, Oxford.
Gamble, C. 1982. Interaction and alliance in Palaeolithic society. Man n.s. 17: 92-107.
Gamble, C. 1983. Culture and society in the Upper Palaeolithic of Europe. In G. N. Bailey (ed.), Hunter-gatherer economy in prehistory: a European perspective, pp. 201-211. Cambridge University Press, Cambridge.
Gamble, C. 1986. The Palaeolithic settlement of Europe. Cambridge University Press, Cambridge.
Gamble, C. 1991. The social context for European Palaeolithic art. Proceedings of the Prehistoric Society 57: 3-15.
Goodall, R., B. David and J. Bartley 1996. Non-destructive techniques for the analysis and characterisation of pigments from archaeological sites: the case of Fern Cave. In S. Ulm, I. Lilley and A. Ross (eds), Australian archaeology '95: proceedings of the 1995 Australian Archaeological Association Annual Conference, pp. 183-187. Tempus 6, Anthropology Museum, University of Queensland, St Lucia.
Gunn, R. G. 1981. The petroglyphs of Sundown Point, northwest Tasmania: some observations and recommendations. Unpubl. report, Tasmanian National Parks and Wildlife Service, Hobart.
Gunn, R. G. 1995. Regional patterning in the Aboriginal rock art of central Australia: a preliminary report. Rock Art Research 12: 117-128.
Gunn, R. G. 1997. Rock art, occupation and myth: the correspondence of symbolic and archaeological sites within Arrernte rock art complexes of central Australia. Rock Art Research 14: 125-136.
Gunn, R. G. 2000. Central Australian rock art: a second report. Rock Art Research 17: 111-126.
Gunn, R. G. 2002. Our country, their country: preliminary comparisons of the pigment rock-art across the Arrernte/Western Desert borderlands. In S. Ulm, C. Westcott , J. Reid, A. Ross, I. Lilley, J. Prangnell and L. Kirkwood (eds), Barriers, borders, boundaries: proceedings of the 2001 Australian Archaeological Association Annual Conference, pp. 109-119. Tempus 7, Anthropology Museum, University of Queensland, St Lucia.
Gunn, R. G. 2003. Arrernte rock-art: interpreting physical permanence in a changing social landscape. Australian Aboriginal Studies 2003(1): 52-73.
Haglund, L. 1975. Survey of sites significant to Aborigines, McArthur River region, Northern Territory. Unpubl. report, Dames and Moore for Mimets Development Pty. Ltd.
Hill, M. O. 1974. Correspondence analysis: a neglected multivariate method. Applied Statistics 25: 340-354.
Hodson, F. R. 1970. Cluster analysis and archaeology: some new developments and applications. World Archaeology 1: 299-320.
Jochim, M. A. 1983. Palaeolithic cave art in ecological perspective. In G. Bailey (ed.), Hunter gatherer economy in prehistory, pp. 212-219. Cambridge University Press, Cambridge.
Layton, R. 1985. The cultural context of hunter-gatherer rock art. Man n.s. 20: 434-453.
Layton, R. 1991. Trends in the hunter-gatherer rock art of western Europe and Australia. Proceedings of the Prehistoric Society 57: 163-174.
Layton, R. 1992. Australian rock art: a new synthesis. Cambridge University Press, Cambridge and New York.
Lewis, D. J. 1988. The rock paintings of Arnhem Land, Australia: social, ecological and material culture change in the post-glacial period. BAR International Series 415, Oxford.
Lewis, D. J. 1997. Bradshaws: The view from Arnhem Land. Australian Archaeology 44: 1-16.
Maynard, L. 1976. An archaeological approach to the study of Australian rock art. M.A. (hons) thesis, Department of Anthropology, University of Sydney.
Maynard, L. 1977. Classification and terminology in Australian rock art. In P. J. Ucko (ed.), Form in indigenous art: schematisation in the art of Aboriginal Australia and prehistoric Europe, pp. 387-402. Australian Institute of Aboriginal Studies, Canberra.
Maynard, L. 1979. The archaeology of Australian Aboriginal art. In S. M. Mead (ed.), Exploring the visual art of Oceania, pp. 83-110. University Press of Hawaii, Honolulu.
Maynard, L. 1988. Comment on F. D. McCarthy, 'Rock art sequences: a matter of clarification'. Rock Art Research 5: 30.
McCarthy, F. D. 1969. The rock engravings of Mt Cameron West, north-west Tasmania. Unpubl. report, Australian Institute of Aboriginal Studies, Canberra, and Tasmanian Museum and Art Gallery, Hobart.
McDonald, J. 1982. On the write track. B.A. (hons) thesis, Department of Anthropology, University of Sydney.
McDonald, J. 1983. The identification of species in a Panaramitee style engraving site. In M. Smith (ed.), Archaeology at ANZAAS 1983, pp. 236-272. Western Australian Museum, Perth.
McDonald, J. 1988. Comment on F. D. McCarthy, 'Rock Art sequences: a matter of clarification'. Rock Art Research 5: 28-30.
McDonald, J. 1992. Rock art in the Sydney region: synchronic and functional variation in a dual-media, regional art body. In J. McDonald and I. P. Haskovec (eds), State of the art: regional rock art studies in Australia and Melanesia. Proceedings of Symposium C, 'Rock art studies in Australia and Oceania' and Symposium D, 'The rock art of northern Australia' of the First AURA Congress, held in Darwin in 1988, pp. 15-24. Occasional AURA publication 6, Australian Rock Art Research Association, Melbourne.
McDonald, J. 1993. On a clear day, you can see Mount Yengo or: investigating the archaeological manifestations of culturally significant foci in the prehistoric landscape. In J. Steinbring, A. Watchman, P. Faulstich and P. S. C. Tacon (eds), Time and space: dating and spatial considerations in rock art research. Proceedings of Symposia F, 'The dating of rock art' and Symposium E, 'Spatial considerations in rock art', Second AURA Congress, Cairns 1992, pp. 84-91. Occasional AURA publication 8, Australian Rock Art Research Association, Melbourne.
McDonald, J. 1994. 'Everywhere you go you know he's been there before'. The influences of Fred McCarthy on rock art research in the Sydney Basin. In M. Sullivan, S. Brockwell and A. Webb (eds.), Archaeology in the north. Proceedings of the 1993 Australian Archaeological Association Conference, pp. 40-53. Northern Australian Research Unit, Darwin.
McDonald, J. 1998. Shelter rock-art in the Sydney Basin--a space-time continuum: exploring different influences on stylistic change. In C. Chippindale and P. S. C. Tacon (eds), The archaeology of rock-art, pp. 319-335. Cambridge University Press, Cambridge.
McDonald, J. 1999. Bedrock notions and isochrestic choice: evidence for localised stylistic patterning in the engravings of the Sydney region. Archaeology in Oceania 34: 145-160.
McNickle, H. P. 1985. An introduction to the Spear Hill rock art complex, northwestern Australia. Rock Art Research 2: 48-59.
Moore, D. R. 1988. Comment on F. D. McCarthy, 'Rock art sequences: a matter of clarification'. Rock Art Research 5: 30-32.
Morphy, H. 1977a. Too many meanings: an analysis of the artistic system of the Yolngu of northeast Arnhem Land. PhD. thesis, Australian National University.
Morphy, H. 1977b. Schematisation, meaning and communication in toas. In P. J. Ucko (ed.), Form in indigenous art: schematisation in the art of Aboriginal Australia and prehistoric Europe, pp. 77-89. Australian Institute of Aboriginal Studies, Canberra.
Morphy, H. 1983. 'Now you understand'--an analysis of the way Yolngu have used sacred knowledge to retain their autonomy. In N. Peterson and M. Langton (eds), Aborigines, land and land rights, pp. 110-133. Australian Institute of Aboriginal Studies, Canberra.
Morwood, M. J. 1979. Art and stone: towards a prehistory of central western Queensland. Ph.D. thesis, Department of Prehistory and Anthropology, Australian National University.
Morwood, M. J. 1980. Time, space and prehistoric art: a principal components analysis. Archaeology and Physical Anthropology in Oceania 15: 98-109.
Morwood, M. J. 1981. Archaeology of the central Queensland Highlands: the stone component. Archaeology in Oceania 16: 1-52.
Morwood, M. J. 1984. The prehistory of the central Queensland Highlands. In F. Wendorf and A. E. Close (eds), Advances in world archaeology, pp. 325-379. Academic Press, New York.
Morwood, M. J. 1985. Facts and figures: notes on rock art in the Mt Isa area, northwestern Queensland. Rock Art Research 2: 140-145.
Morwood, M. J. 1988. Comment on F. D. McCarthy, 'Rock art sequences: a matter of clarification'. Rock Art Research 5: 32-33.
Morwood, M. J. 1992. Changing art in a changing landscape: A case study from the upper Flinders region of the north Queensland Highland. In J. McDonald and I. P. Haskovec (eds), State of the art: regional rock art studies in Australia and Melanesia. Proceedings of Symposium C, 'Rock art studies in Australia and Oceania' and Symposium D, 'The rock art of northern Australia' of the First AURA Congress, held in Darwin in 1988, pp. 60-70. Occasional AURA Publication 6, Australian Rock Art Research Association, Melbourne.
Morwood, M. J. 2002. Visions from the past: the archaeology of Australian Aboriginal art. Allen and Unwin, Sydney.
Morwood, M. J. and D. R. Hobbs 1995a. Conclusions. In M. J.
Morwood and D. R. Hobbs (eds.), Quinkan prehistory: the archaeology of Aboriginal art in S.E. Cape York Peninsula, Australia, pp. 178-185. Tempus 3, Anthropology Museum, University of Queensland, St. Lucia.
Morwood, M. J. and D. R. Hobbs 1995b. Themes in the prehistory of tropical Australia. Antiquity 69: 747-768.
Mountford, C. P. and R. Edwards 1964. Rock engravings in the Red Gorge, Deception Creek, northern South Australia. Anthropos 59: 849-859.
Moyle, R. 1983. Songs, ceremonies and sites: the Agharringa case. In N. Peterson and M. Langton (eds), Aborigines, land and land rights, pp. 66-93. Australian Institute of Aboriginal Studies, Canberra.
Mulvaney, J. and J. Kamminga 1999. Prehistory of Australia. Allen and Unwin, Sydney.
Mulvaney, K. 2001. Cultural images: the petroglyphs of a sandstone quarry, Helen Springs, Northern Territory, Australia. Rock Art Research 18: 40-54.
Munn, N. D. 1966. Visual categories: an approach to the study of representational systems. American Anthropologist 68: 936-950.
Munn, N. D. 1973. Walbiri iconography: graphic representation and cultural symbolism in a central Australian society. Cornell University Press, London.
Nobbs, M. F. 1981. Mannahill Rock Engraving Site (Karolta). Unpubl. report, Australian Institute of Aboriginal Studies.
Nobbs, M. F. 1982a. Karolta 2 Rock Engraving Site. Unpubl. report, South Australian Heritage Branch, Department of Planning and Environment.
Nobbs, M. F. 1982b. Oulnina Rock Engraving Site. Unpubl. report, Australian Institute of Aboriginal Studies.
Nobbs, M. F. 1982c. Aboriginal rock art in Olary province, South Australia, on Bimbowrie, Outalpa, Karolta. Unpubl. report, Australian Heritage Commission.
Nobbs, M. F. 1982d. Tattawuppa Rock Engraving Site. Unpubl. report, Australian Institute of Aboriginal Studies.
Nobbs, M. F. 1982e. Rowes Dam Rock Engraving Site. Unpubl. report, Australian Institute of Aboriginal Studies.
Nobbs, M. F. 1982f. Yunta Rock Engraving Site. Unpubl. report, Australian Institute of Aboriginal Studies.
Nobbs, M. F. 1983a. Yunta Springs Rock Engraving Site. Unpubl. report, Australian Institute of Aboriginal Studies.
Nobbs, M. F. 1983b. Florina Rock Engraving Site. Unpubl. report, Australian Institute of Aboriginal Studies.
Nobbs, M. F. 1983c. Morialpa Rock Engraving Site. Unpubl. report, Australian Institute of Aboriginal Studies.
Nobbs, M. F. 1983d. Mount Victor Rock Engraving Site. Unpubl. report, Australian Institute of Aboriginal Studies.
Nobbs, M. F. 1983e. Winnininnie Rock Engraving Site. Unpubl. report, Australian Institute of Aboriginal Studies.
Nobbs, M. F. 1983f. Winnininnie 2 Rock Engraving Site. Unpubl. report, Australian Institute of Aboriginal Studies.
Nobbs, M. F. 1983g. Aboriginal rock art in Olary province on Teetulpa, Winnininnie, Florina and Karolta. Unpubl. report, Australian Institute of Aboriginal Studies.
Nobbs, M. F. 1984. Rock art in Olary province, South Australia. Rock Art Research 1: 91-118.
Nobbs, M. F. and R. I. Dorn 1988. Pleistocene age determinations for rock varnish formation within petroglyphs: cation-ratio dating of 24 motifs from the Olary region, South Australia. Rock Art Research 5: 108-144.
Nobbs, M. F. and R. I. Dorn 1993. New surface exposure ages for petroglyphs from the Olary Province, South Australia. Archaeology in Oceania 28: 18-39.
O'Connor, S. and B. Fankhauser 2001. Art at 40,000 BP? One step closer: an ochre covered rock from Carpenter's Gap Shelter 1, Kimberley region, Western Australia. In A. Anderson, I. Lilley and S. O'Connor (eds), Histories of old ages: essays in honour of Rhys Jones, pp. 287-300. Pandanus Books, The Research School of Pacific and Asian Studies, The Australian National University, Canberra.
Peterson, N. 1986. Australian territorial organisation: a band perspective. Oceania Monograph 30, University of Sydney, Sydney.
Preiss, K. A. 1962. Aboriginal rock carvings and paintings at 'The Canyon', Flinders Ranges, South Australia. The South Australian Naturalist 39: 5-12.
Quinnell, M. C. 1972. Aboriginal rock engravings near Rocky Scrub Creek, Junction View, southeast
Queensland: Tryon's Pigeon Creek site re-recorded. Memoirs of the Queensland Museum 16: 215-221.
Rola-Wojciechowski, C. 1983. 'A bit of bundy': the 'Bundaberg' engraving site. B.A. (hons) thesis, Department of Archaeology and Palaeoanthropology, University of New England.
Rosenfeld, A. 1981a. Excavations at the Early Man Shelter. In A. Rosenfeld, D. Horton and J. Winter (eds), Early Man in north Queensland: art and archaeology in the Laura area, pp. 5-34. Terra Australis Series 6, Australian National University, Canberra.
Rosenfeld, A. 1981b. Rock engravings in the Laura area. In A. Rosenfeld, D. Horton and J. Winter (eds), Early Man in north Queensland: art and archaeology in the Laura area, pp. 50-89. Terra Australis Series 6, Australian National University, Canberra.
Rosenfeld, A. 1991. Panaramitee: dead or alive? In P. Bahn and A. Rosenfeld (eds), Rock art and prehistory: papers presented to Symposium G of the AURA Congress, Darwin 1988, pp. 136-144. Oxbow Books. Oxford.
Rosenfeld, A. and M. A. Smith 2002. Rock-art and the history of Puritjarra Rock Shelter, Cleland Hills, central Australia. Proceedings of the Prehistoric Society 68: 103-124.
Ross, J. 1997. Painted relationships: an archaeological analysis of a distinctive anthropomorphic rock art motif in northwest central Queensland. BA (hons) thesis, Department of Archaeology and Palaeoanthropology, University of New England, Armidale.
Sackett, J. R. 1977. The meaning of style in archaeology: a general model. American Antiquity 42: 369-380.
Smith, C. 1992. Colonising with style: reviewing the nexus between rock art, territoriality and the colonisation and occupation of Sahul. Australian Archaeology 34: 34-42.
Smith, M. A. 1988. The pattern and timing of prehistoric settlement in central Australia. PhD thesis, Department of Archaeology and Palaeoanthropology, University of New England.
Smith, M. A. 1989. The case for a resident human population in the Central Australian Ranges during full glacial aridity. Archaeology in Oceania 24: 93-105.
Smith, M. A. 1996. Prehistory and human ecology in central Australia: an archaeological perspective. In S. R. Morton and D. J. Mulvaney (eds), Exploring central Australia: society, the environment and the 1894 Horn expedition, pp. 61-73. Surrey Beatty and Sons, Chipping Norton.
Smith, M. A. and P. M. Veth 2004. Radiocarbon dates for baler shell in the Great Sandy Desert. Australian Archaeology 58: 37-38.
Smith, M. A., B. Fankhauser and M. Jercher 1998. The changing provenance of red ochre at Puritjarra rock shelter, central Australia: Late Pleistocene to present. Proceedings of the Prehistoric Society 64: 275-292.
Spencer, B. and F. J. Gillen 1938. The native tribes of central Australia. Macmillan and Co., London.
Stockton, J. 1977. Site report: Greenes Creek Aboriginal engraving site. Tasmanian National Parks and Wildlife Service Occasional Paper 1, Hobart.
Strehlow, T. G. H. 1978. Central Australian religion: personal monototemism in a polytotemic community. Australian Association for the Study of Religions, Adelaide.
Sullivan, H. 1988. Rock engravings in Kakadu National Park. Paper presented in Symposium C, First Australian Rock Art Research Association Congress, Darwin.
Sutton, P. 1988a. The morphology of feeling. In P. Sutton (ed.), Dreamings: the art of Aboriginal Australia (exhibition catalogue), pp. 59-88. Penguin, Melbourne.
Sutton, P. 1988b. Dreamings. In P. Sutton (ed.), Dreamings: the art of Aboriginal Australia (exhibition catalogue), pp. 13-32. Penguin, Melbourne.
Sutton, P. 1988c. Responding to Aboriginal art. In P. Sutton (ed.), Dreamings: the art of Aboriginal Australia (exhibition catalogue), pp. 33-58. Penguin, Melbourne.
Sutton, P. 1990. The pulsating heart: large scale cultural and demographic processes in Aboriginal Australia. In B. Meehan and N. White (eds), Hunter-gatherer demography: past and present, pp. 71-80. Oceania Monograph 39, University of Sydney, Sydney.
Tacon, P. S. C., M. Wilson and C. Chippindale 1996. Birth of the Rainbow Serpent in Arnhem Land rock art and oral history. Archaeology in Oceania 31: 103-124.
Veth, P. 1989. Islands in the interior: a model for the colonization of Australia's arid zone. Archaeology in Oceania 24: 81-92.
Veth, P. 2000. Origins of the Western Desert language: convergence in linguistic and archaeological space and time models. Archaeology in Oceania 35: 11-19.
Veth, P. and J. McDonald 2002. Can archaeology be used to address the principle of exclusive possession in native title. In R. Harrison and C. Williamson (eds), After Captain Cook: the archaeology of the recent indigenous past in Australia, pp. 122-132. Sydney University Archaeological Methods Series 8, Archaeological Computing Laboratory, University of Sydney, Sydney.
White, J. P. 1984. Australian prehistory: time for a reassessment? The Quarterly Review of Archaeology 5: 11.
White, J. P. and P. J. Habgood 1985. La prehistoire de l'Australie. La Recherche 167: 730-737.
White, J. P. and J. F. O'Connell 1982. A prehistory of Australia, New Guinea and Sahul. Academic Press, Sydney.
Wright, R. V. S. 1985. Detecting pattern in tabled archaeological data by principal components and correspondence analysis: programs in basic for portable microcomputers. Science and Archaeology No. 27: 35-38.
Yellen, J. E. and H. Harpending 1972. Hunter-gatherer populations and archaeological inference. World Archaeology 4: 244-253.
RAR 24-819
Dr Natalie R. Franklin
Cultural Heritage Branch
Environmental Protection Agency
P.O. Box 15155
City East, QLD 4002
Australia
E-mail: natalie.franklin@epa.qld.gov.au
(1.) In this paper, petroglyph motifs are given names, which refer to their appearance to the observer, and not labels, which refer to the meaning or interpretation of the motifs. For example, use of the term 'kangaroo track' or 'macropod track' as a motif name implies only that the figure reminds the observer of a kangaroo or macropod track, not that it is assumed the artist's original intention was to depict such tracks. Clegg (1977, 1981a) introduced the notation '!' as a prefix for motifs, thus specifically excluding the attribution of meaning or intent. However, I do not intend to use this notation for the naming of motifs, although the usage of motif names in this paper is the same as that proposed by both Clegg (1977, 1981a) and Maynard (1977). Therefore, a motif is named according to the shape that it reminds me of, and is not intended to imply that the motif is necessarily a depiction of that shape.
Table 2. The sites included in the analyses, their abbreviations in
Figures 3 and 4, and the source of the data. Unless otherwise stated,
data are derived from tables of motif counts from each source.
Site Abbreviation Source/form of data
South Australia
Flinders Ranges FR Preiss (1962)
Florina F Nobbs (1984)
Karolta 1 K1
Karolta 2 K2
Mt Victor MV
Oulnina O
Rockholes R Gale (pers. comm.)
Rowes Dam RD Nobbs (1984)
Salt Creek SC Gale (pers. comm.)
Tatt awuppa T Nobbs (1984)
Wharton Hill WH Gale (pers. comm.)
Winnininnie 1 Wi1 Nobbs (1984)
Winnininnie 2 Wi2
Yunta Y
Yunta Springs YS
Western New South Wales
Sturt's Meadows SM Clegg (in prep., pers. comm.)
Northern Territory
N'Dhala Gorge NG Forbes (pers. comm.)
McArthur River MR Haglund (1975, pers. comm.)
--drawing of engraved panel
from photographic montage
Tasmania
Greenes Creek GC Stockton (1977)--tracings of
petroglyphs, checked during
personal fieldwork
Mt Cameron West MCW McCarthy (1969)--tracings of
petroglyphs, list of motifs
Sundown Point SP Gunn (1981)--list of motifs,
table of counts, sketch
plan of site
Queensland--Cape York
Early Man EM Rosenfeld (1981b)--tracings
of engraved panels
Early Man Shelter EMC Rosenfeld (1981b)
C
Early Man Shelter EMH
H
Emu Dreaming ED Trezise unpublished scaled
Death Adder DA illustrations, Australian
Gallery Institute of Aboriginal and
Laura River LR Torres Strait Islander
Possum Gallery PG Studies, Canberra--all
recordings checked during
personal fieldwork
Green Ant GA Flood (pers. comm.)
Echidna Shelter ES
Queensland--Mt Isa region
Browns Creek BC Personal fieldwork
Carbine Creek CC Morwood (1985)
Einasleigh E Ron Edwards (1967)--drawing
of engraved panel
Frank Creek FC Personal fieldwork
South-east Queensland
Bundaberg B Rola-Wojciechowski (1983)
Gatt on G Quinnell (1972)--tracing of
engraved panel
Central western Queensland
Buckland Creek 1 BC1 Morwood (1979)
Bull Hole BH
Dooloogarah 2 D2
Goat Rock 1 GR1
Goat Rock 2 GR2 Morwood (1979)
Morven M
Native Well 1 NW1
Native Well 2 NW2
Ochre Site 1 O1
Paddy's Cave PC
Plateau 1 P1
Twelve Mile 12MC
Crossing
Weir 1 W1
Western Australia
Edah Ed Personal fieldwork
Scott River SR
Table 3. Percentages of motifs at sites in the Australia-wide sample
(rounded to whole numbers). For site abbreviations, see Table 2. NB
complex circles could not be determined for the central western
Queensland sites (BC1 - W1 inclusive).
Motifs
'Macropod 'Bird 'Human Other Dots/
Site tracks' tracks' footprints' 'tracks' pits
FR 1 12 0 0 0
F 16 10 3 3 12
K1 15 16 2 0 19
K2 12 10 0 0 42
MV 41 6 0 0 0
O 0 15 4 0 0
R 41 17 4 1 9
RD 31 13 4 1 10
SC 33 17 2 1 7
T 22 9 3 0 6
WH 40 24 4 1 6
Wi1 20 16 1 2 6
Wi2 22 6 0 0 37
Y 15 14 1 1 29
YS 28 13 4 1 21
SM 28 27 2 3 32
NG 17 13 1 0 28
GC 0 0 0 0 0
MCW 0 2 0 0 16
SP 0 0 0 0 32
SR 21 47 0 0 11
Ed 6 4 1 0 6
BC 27 7 0 0 51
CC 2 13 0 2 7
E 0 11 0 0 8
FC 2 3 0 0 9
MR 0 3 0 0 12
B 0 2 0 1 13
G 0 6 0 0 69
BC1 22 13 6 20 1
BH 5 41 9 5 4
D2 1 3 0 6 0
GR1 13 15 10 8 0
GR2 13 58 7 3 2
M 65 22 0 0 0
NW1 1 6 0 1 1
NW2 5 1 0 0 0
O1 25 35 2 3 0
PC 5 29 3 3 0
P1 9 24 0 2 2
12MC 19 17 17 9 6
W1 10 21 12 14 6
DA 0 19 0 0 58
EM 3 30 0 0 20
EMC 2 31 0 0 21
EMH 5 67 0 0 5
ES 9 17 1 1 64
ED 0 14 0 0 79
GA 9 3 3 2 50
LR 36 17 22 6 1
PG 2 10 0 0 71
Motifs
Meandering Simple Complex Simple Complex
Site lines/mazes lines lines circles circles
FR 1 0 0 73 1
F 3 4 0 40 0
K1 0 5 0 38 2
K2 0 0 0 20 5
MV 0 0 0 35 0
O 0 0 0 81 0
R 0 1 1 17 4
RD 0 1 0 35 1
SC 1 0 0 26 6
T 0 1 0 43 14
WH 0 2 1 14 4
Wi1 1 9 0 40 1
Wi2 1 12 0 12 1
Y 1 2 0 22 4
YS 0 9 1 20 0
SM 1 1 1 2 0
NG 8 0 2 7 12
GC 0 0 1 96 3
MCW 2 9 1 44 22
SP 0 13 2 34 17
SR 0 4 12 3 0
Ed 0 1 0 33 22
BC 1 3 2 2 3
CC 11 8 2 13 28
E 2 8 0 28 30
FC 2 7 2 22 39
MR 8 24 6 6 38
B 2 7 3 62 8
G 1 13 2 2 2
BC1 3 3 0 2 0
BH 3 2 0 15 0
D2 0 0 0 0 0
GR1 1 3 0 0 0
GR2 0 1 1 0 0
M 0 3 1 3 0
NW1 0 7 0 1 0
NW2 0 12 0 0 0
O1 0 10 3 1 0
PC 0 19 3 0 0
P1 1 1 0 2 0
12MC 5 1 0 9 0
W1 2 2 0 2 0
DA 5 11 3 0 1
EM 20 14 2 2 2
EMC 5 18 8 0 2
EMH 14 0 9 0 0
ES 0 0 0 0 0
ED 1 1 1 1 0
GA 6 3 1 13 8
LR 0 0 0 1 0
PG 4 6 2 0 3
Motifs
Radiating Other Total
Site Crescents lines figuratives No.
FR 12 0 0 127
F 8 0 1 497
K1 3 0 0 1579
K2 10 0 1 94
MV 18 0 0 17
O 0 0 0 26
R 4 0 1 5857
RD 4 0 0 442
SC 6 0 1 2471
T 2 0 0 236
WH 4 0 0 2246
Wi1 4 0 0 925
Wi2 9 0 0 147
Y 11 0 0 396
YS 3 0 0 1069
SM 3 0 0 13 901
NG 4 6 2 5462
GC 0 0 0 66