Copyright © 2002 by Christine Whitehead and Kevin Sharpe. All rights reserved. National Geographic Society Research Reports, 1976 Projects (Washington, DC: National Geographic Society, 1978), pp. 31-50.

An Analysis of Prehistoric Engravings on Boulders in Koonalda Cave, South Australia

Christine E. Sharpe

CONTENTS.

Background. 1

The Upper Chamber 2

Meanders and Engravings. 4

Grant Recipient: Christine E. Sharpe, Auckland, New Zealand.

Grant 1523: In support of an expedition to the art sanctuary in Koonalda Cave, South Australia.

An investigation of engravings on boulders in the northwest passage of Koonalda Cave, South Australia, was undertaken in January 1976 on an expedition supported by the National Geographic Society and the South Australian Museum (Sharpe, 1977; Sharpe and Sharpe, n.d.[1] and n.d.[2]).

Background

During the summer of 1972-1973, my husband, Kevin Sharpe, and I were part of a group led by Alexander Gallus to examine evidence of a prehistoric people who utilized Koonalda Cave. There, we found carefully engraved boulders on the floor of the cave’s Upper Chamber (Sharpe and Sharpe, 1976; see Figs. 1 and 2). Koonalda Cave has been known over many years for its prehistoric flint mining, stone constructions, and markings at least 20,000 years old made by human fingertips being stroked across the soft, powdery limestone surface of the cave walls (Wright, 1971).

Koonalda Cave, located 14 mi (23 km) from the coast − the Great Australian Bight − and 54 mi (87 km) from the Western Australian border (Fig. 3), is one of the largest of the 170 known caves on the Nullarbor Plain, part of a limestone karst of some 78,000 mi²(200,000 km²) (Dunkley and Wigley, 1967; Lowry and Jennings, 1974). The land around Koonalda is fairly arid, with only about 8 in. (200 mm) of rain failing per year, and an evaporation rate that always exceeds rainfall (Leonard, 1980).

In 1957, Gallus had recognized Koonalda’s archeological significance, and continued his investigations until 1976 (Gallus, 1968, 1971, 1977). Another archeologist to operate in Koonalda was Richard Wright of the University of Sydney (Wright, 1971). Robert Edwards and Leslie Maynard have described the cave’s finger markings and wall engravings in a number of publications (Edwards and Maynard, 1967, 1969; Maynard and Edwards, 1971).

The Upper Chamber

The floor of the Upper Chamber of Koonalda Cave comprises at least five rockfalls of different ages and degrees of weathering (which renders the rocks smooth and rounded) (Figs. 3, 4, and 5). The oldest collapse, rockfalls C and F, is engraved; engravings within the other areas are sparse if not absent altogether. Within rockfall C are individual engraved rocks and a number of human activity areas, some obviously intentionally constructed and flanked by engraved boulders. This part of the cave is a prehistoric ritual sanctuary of some magnitude.

The boulders in rockfall A are fairly smooth and rounded, and those in rockfall D are more rough and jagged. But the latter have been used significantly by humans in that there is an impressive series of activity areas constructed here, together with bone placements and a chalcedony flake, presumably an engraving tool. The remains of torches abound in all but rockfalls B and F, the most recent. In them is no evidence of human use, but underneath them − and under A and D as well − can be found the original smooth, rounded, and engraved boulder floor.

An important question to be answered is by what mechanism did the rockfalls become smooth and rounded. This process is active on rockfalls over already engraved smooth and rounded boulders, and difficulties obviously exist if the process is achieved by water flow as is often supposed. In fact, the most promising mechanism is that of exudation or salt weathering, in which crystals form in the surface pores of the limestone, forcing off particles. There are, however, details to be settled with this proposal (Sharpe and Sharpe, n.d.[1]).

The above perspective on the Upper Chamber was a result of the 1976 investigation. At that time, also, a thorough survey was made of the cave, especially noting the exact positions of the engraved boulders.

Meanders and Engravings

During the last few years there has been a growing interest in the subject of cave meanders and line engravings, and more and more Australian caves containing them are coming to light. Cutta Cutta and Kintore Caves in the Northern Territory, published by W. P. Walsh in 1964, contain Koonalda-type markings (Walsh, 1964). There are finger markings in the Orchestra Shell and Ross Caves north of Perth, Western Australia, and a little farther south in Morfitt’s Cave (Hallam, 1971; Dortch, 1976). In the caves around Mount Gambier in South Australia, finger markings and engravings have been discovered, and we believe they also are to be found in a number of other Nullarbor caves. In Victoria, markings can be seen in Princess Margaret Rose Cave, McEachern’s Cave, and in New Guinea 2 Cave (Walsh, 1964; Link, 1967; Stubbs, 1979).

Alexander Marshack is well known for his studies of prehistoric engravings on small portable pieces of stone and bone from Europe (Marshack, 1972a, b, c, 1975, 1977, 1979). His intention is to discover something of the thought patterns of prehistoric people. By careful microscopic observation, he has found that the tools used to incise the marks differ from mark to mark or group of marks to group of marks, which he feels proves that the marks were made over a period of time or perhaps by different people. The marks tend to form sets, each an obvious unit in itself, which Marshack describes as “mythic” rather than “arithmetic” (tally marks); that is, the line series brings to mind an idea that sparks off many associations relating to myths.

In relation to Koonalda Cave and the Australian meander tradition, Marshack provides us with specific tools to set about recording the lines in a way that may be helpful in their later interpretation. He personally suggested we investigate how the lines were made, how they were built up over time, and whether they were drawn with different tools.

Aim and Method

An aim during the 1976 expedition to Koonalda was to follow Marshack’s suggestion of trying to discover something about the internal structure of the engraved lines on the boulders. By doing this, we hoped to see if there were any similarities between the engravings on boulders of the northwest passage of Koonalda Cave and those that Marshack had studied in Europe. We were especially interested in his claim that the work was participatory in nature, and wanted to see if the line series or streams were consistent in that all the lines of which they were composed had the same cross sections (i.e., were made with the same point) and that different streams had different cross sections. A stream can be defined at first visually as being a group of parallel lines which have a common direction and a consistent interval between lines, and which appear to have the same cross section and degree of wear.

First, a suitable sample of engravings was chosen, one relatively clear and easy to “read” (some areas are so heavily engraved that the lines are virtually impossible to sort out). The work was then undertaken with a strong side light (gas), and hand-held candle which could be moved across the streams until the light was thrown directly into each line. This eliminated the shadows thrown by the side light and, with the aid of a magnifying glass, allowed a look into the lines, to establish the contour of their cross sections.

It was possible to establish which groups superimpose others by examining the junction at which they crossed. These results can be seen in Figures 6-13. An important word of caution needs to be appended. The figures drawn are replications of those made in the cave and show the overlays and the cross sections; the visual discrimination between different streams has been (almost) ignored. The point was to test the hypothesis that a visually definable stream shows consistency in the cross sections of its constituent lines, and consistency in the way its lines overlay or are overlaid by lines of other streams. As can be seen from the drawings, this hypothesis is born out in most of the line sets examined, in that lines grouped by common cross sections and common overlayings do in general correspond with the visually discernable streams.

However, this is subject to a number of circumspections. Because the above hypothesis itself and the method of drawing the lines evolved during the examination of the lines themselves, there are inaccuracies in the figures and the method of representation that cannot be rectified without further examination of the lines in situ (the hypothesis, however, still seems well established). For instance, some lines are given the same level in the overlay structure, and so are represented by the same symbol, when this is not ascertainable since they have no crossing or touching with other lines. Some lines have been given the same level of overlay when they do meet or cross one another, meaning their relative overlaying was not looked for. And thirdly, some sets of lines have no crossings or meetings at all and yet an overlay structure has been given to them. Thus, the method of representation used has its drawbacks and in further investigations should be modified.

Results

1. Two areas on one particular boulder were studied, one is the central section and the other directly above a hole in the foot of the boulder.

The central section (Fig. 6) lies to the left of a natural hole in the limestone that penetrates the boulder to the depth of 30 cm. For convenience, this grouping is described in four clusters physically separated from each other, but which contain different streams. Working from left to right we have the top cluster 1. In this, stream 1 slants from top left to bottom right, and comprises closely spaced parallel lines; there are four of these with a fifth farther to the right but still at roughly the same angle, though curved. The second line is broken. The cross section for these is a flat-bottomed U. Laying on top of these are the regularly spaced marks of stream 2. The first four of these slant to the right across the previous set, the fifth and broken seventh are vertical, and the sixth (lower down) slants back toward the left with a small line slanting from the left toward and meeting it (shallow cross section). Across the lower part of the first stream leaning to the left is a third stream, more deeply engraved than either the first or second. These have deep V cross sections.

Cluster 2 has two streams only, one overlaying the other (called 4 and 5 for convenience). Stream 4 slants from the top right to the bottom left, one long line and a group of four small lines, the first of which branches at the bottom to the left. The nine lines of stream 5 slant from the bottom right to the top left. The first of the 4 group has a shallow U cross section, and on the other three a deep V.

Cluster 3 has four distinctive levels, the lowest (stream 6) has an extremely low U cross section and arches almost horizontally upward toward the left where it intersects with cluster 2. The next layer (7) is composed of four vertical lines, overlapping 6. Stream 8 consists of three wavy lines, parallel and slanting toward the right-hand upper corner. Cutting over all of these, and deepest, is the curved line probably part of stream 5 in cluster 2.

Cluster 4 has four horizontal wavy lines (9) crossing the top; they are very faint and shallow and have a V-shaped cross section. Directly below and to the middle of these are seven parallel lines (10) which slope toward each other, the set of which seems to branch off one of two very deep lines (11). A tiny group to the right consists of three curved lines (12) overlaid by three straight vertical ones (13).

2. The lines from the lower portion of the above boulder are shown in Figure 7. They form an “arrow” pointing downward to a hole at the floor level of the boulder, reminiscent of a similar “arrow” in the “squeeze” area which points downward toward a mining hole (Gallus, 1968). It is a very simple grouping. Four lines (stream 1) slant toward the hole from above left and these are the deepest and have a squared-off U cross section; two more of these lines join the stream at its base. From the right slant four short thin lines (deep and thin U cross section); there is a gap and two more similar lines at the same angle meet the hole (stream 2). In the center of the cluster are four deep lines (stream 3) slanting in toward each other and pointing downward to the junction of the left and right streams, and also pointing to the hole.

3. The section selected to draw from another boulder shows interesting stream patterns which appear to link together natural holes in the limestone surface (Fig. 8). The lowest right-hand hole shown in the drawing has lines (stream 1) drawn toward it and then going inside it. It was rather difficult to sort out superimpositions in this example owing to the great confusion of lines. Those drawn have been selected from this mass. There is little variation in cross section although we can still see that, like the first drawing, the cross sections are fairly consistent with the streams.

4. Two drawings have again been made from one boulder, which this time faces onto an activity area. The engraved surface is badly cracked and it may soon flake off.

Superimpositioning is obvious in the first example (Fig. 9). The form is basically tent-like, with two strong streams of parallel lines leaning toward each other. It appears that the left-hand side stream (1) overlays the right-hand side (2). A smaller stream of lines (3) is between the two sides and pointing toward their apex.

These three general aspects are similar to the lower cluster on example (2), except this “arrow”-type form points upward. Variation in cross section can again be observed.

5. The second example from this boulder is a complex patterned surface immediately below the first (Fig. 10). It appears to contain two systems joined together, one large (upper) and one small (lower) with various short parallel streams scattered about.

A vague similarity exists between the two systems in that, in both, two separate streams slant from the lower-left to the upper-right and two more slant through these from the upper-left to the lower-right. Perhaps this shows some left/right symmetry.

6. This is a single system consisting of seven separate streams (Fig. 11), four running roughly vertically (1, 2, 3, 4) and three crossing the upper section of these like an arrow pointing to the right (5, 6, 7). The three vertical streams incline slightly from lower-left to top-right. The “arrow” portion consists of two sets of four parallel lines (5 and 7), and a further parallel group (6) joins these from behind like a tail.

7. The next engraving to be described was found on the “altar” stele (Fig. 12). Briefly, the engraving has two parts: a large and vertical V (apex at A) of streams 1, 7, and 8, and a small “fan” grouping of parallel-line streams at the upper-left. The large V is intersected at the top of its right arm by the base of a second and upturned V of streams 6 and 9. The “fan” consists of four streams of parallel lines, two of two (2 and 5), and two of four (3 and 4). Stream 2 is crossed by the other two-lined stream 5.

The most interesting aspect of this cluster is the intimate association of the engraved lines with the natural structure of the limestone. Streams 7 and 8 connect holes. A long hole sets the direction of the shorter arm (stream 1) of the first V mentioned, the lines being a continuation of the upper end of the hole. A natural surface crack is followed in contour by stream 6 forming the left-hand arm of the second and upside-down V. The longest line of this four-line stream meets the crack about halfway up, and the remaining three shorter lines lock into the end of the space formed like fingers into a hand.

8. The final area of lines to be described is located on the lower stone of the two at the entrance to the “cavern” (Figs. 1, 2, and 13). The lines do not form clusters as do most of the others, but are, almost without exception, vertical; some intersect at their lower ends and there are a few small side-branches. As there are few places where lines cross, it is impossible to work out any real time sequence, but three levels have been sorted out. As in the previous example, a longish hole is incorporated into the design.

Conclusions

There are many aspects to be considered in analyzing the markings found in Koonalda Cave, which may well represent in Australia the earliest human artistic expression. Nine findings of our expedition, briefly listed below, will be of assistance in interpreting these markings aged in excess of 20,000 years.

1. Cross sections seem consistent across a stream of lines.

2. Cross sections range from a sharp V to a broad, shallow U.

3. There is a direct relationship between natural features of the boulder surface and some engraved lines. For instance, small pits or holes are joined by lines, or have lines passing through or coming out of them. Long cracks are imitated by lines engraved parallel to them. And large holes are often indicated by a complex of a number of streams converging on the holes.

4. Fine-line groups tend to underlie thicker ones.

5. The majority of parallel lines are in groups of four. This is consistent with the finger markings (presumably executed with the four fingers).

6. Marshack (pers. comm.) mentions groups of parallel lines, overlapping each other and moving systematically down the rock face or wall. We noticed no structuring similar to this on Koonalda boulders.

7. The engravings on the lower boulders of the cavern are distinctly different from the others that were studied. They are mainly parallel and move uniformly up and down the face (see Figs. 1 and 13). They are also emphatically drawn and thus possibly related to the striking form of the “cavern.” In general, it appears that engraving patterns associated with outstanding “hole” formations such as the “cavern” have a distinctly different form from other boulder engravings; they are clear, vertical lines with fairly even spacing and lack any underlying mesh of lines. Perhaps this definite change of style in engraving implies that different sets of lines have different purposes or meanings.

8. In some systems of streams there is rough left/right symmetry.

9. A number of clusters contain V-shapes of streams bisected at the apex by other streams.

Marshack’s method can tell us a great deal about the (temporal) structure of the Koonalda markings. More, of course, must be done along this line. But the future task must also be to discover more of what they meant to their makers. This will require a number of techniques including anthropological, structural, psychological, and artistic approaches, each of which will tell something, perhaps, but no one of which will probably tell the whole story.

Acknowledgments

As principal investigator, I wish to express my thanks to the National Geographic Society and to the South Australian Museum for their financial and other support toward the 1976 expedition to Koonalda Cave. I am also grateful to the South Australian Aboriginal and Historic Relics Advisory Board (for permission to enter the cave), Mr. and Mrs. Cyril Gurney of Koonalda Station, Dr. Alexander Marshack, Professor Hallam Movius, Jr., and to the members of the expedition: Dr. Alexander Gallus (nominated by the South Australian Museum as field investigator), Dr. Kevin Sharpe (photographer), Messrs. Ian Lewis and Kevin Mott (surveyors), and Mr. Neil Chadwick (representing the South Australian Museum). Dr. Kevin Sharpe gave invaluable help in the preparation of this report.

References

DORTCH, C. E. 1976. Two engraved stone plaques of late Pleistocene age from Devil’s Lair, Western Australia. Arch. Phys. Anth. Oceania, vol. 11, no. 1, pp. 32-44.

DUNKLEY, J. R., and WIGLEY, T.M.L., eds. 1967. Caves of the Nullarbor, 61 pp., illus. Speleological Research Council, Sydney.

EDWARDS, R., and MAYNARD, L. 1967. Prehistoric art in Koonalda Cave. Proc. South. Aust. Br. Roy. Geogr. Soc. Australasia, vol. 68, pp. 11-17.

_________. 1969. Prehistoric art in Koonalda Cave (Australia). Bol. Cent. Cam. Stud. Preist., vol. 4, no. 4, pp. 117-130.

GALLUS, A. 1968. Parietal art in Koonalda Cave, Nullarbor Plain, South Australia. Helictite, vol. 6, no. 3, pp. 43-49.

_________. 1971. Results of the exploration of Koonalda Cave, 1956-1968. Pp. 87-133 in Archaeology of the Gallus Site, Koonalda Cave, 133 pp., illus., R. V. S. Wright, ed. Australian Institute of Aboriginal Studies, Canberra.

_________. 1977. Schematisation and symboling. Pp. 370-386 in Form in Indigenous Art, 486 pp., illus., P. J. Ucko, ed. Australian Institute of Aboriginal Studies, Canberra.

HALLAM, S. J. 1971. Roof markings in the “Orchestra Shell” Cave, Wanneroo, near Perth, Western Australia. Mankind, vol. 8, no. 2, pp. 90-103.

LEONARD, B. E. 1980. South Australian year book no. 15: 1980, 666 pp., illus. Australian Bureau of Statistics, South Australian Office, Adelaide.

LINK, A. G. 1967. Late Pleistocene-Holocene climatic fluctuations; possible solution pipe-foibe relationships; and the evolution of limestone cave morphology. Zeitschr. Geomorph., vol. 11, no. 2, pp. 117-145.

LOWRY, D. C., and JENNINGS, J. N. 1974. The Nullarbor karst, Australia. Zeitschr. Geomorph., vol. 18, no. 1, pp. 35-81.

MARSHACK, A. 1972a. Cognitive aspects of upper Paleolithic engraving. Curr. Anthr., vol. 13, nos. 3-4, pp. 445-477.

_________. 1972b. The roots of civilization, 413 pp., illus. McGraw-Hill Book Co., New York.

_________. 1972c. Upper Paleolithic notation and symbol. Science, vol. 178, no. 4063, pp. 817-828.

_________. 1975. Exploring the mind of ice-age man. Nat. Geogr., vol. 147, no. 1, pp. 62-89.

_________. 1977. The meander as a system: the analysis and recognition of iconographic units in upper Palaeolithic compositions. Pp. 286-3 17 in Form in Indigenous Art, 486 pp., illus., P. J. Ucko, ed. Australian Institute of Aboriginal Studies, Canberra.

_________. 1979. Upper Paleolithic symbol systems of the Russian Plain: cognitive and comparative analysis. Curr. Anthr., vol. 20, no. 2, pp. 271-311.

MAYNARD, L., and EDWARDS, R. 1971. Wall markings. Pp. 59-80 in Archaeology of the Gallus Site, Koonalda Cave, 133 pp., illus., R.V. S. Wright, ed. Australian Institute of Aboriginal Studies, Canberra.

SHARPE, C. E. 1977. Koonalda Cave: the beginning of artistic expression. New Quart. Cave, vol. 2, no. 3, pp. 226-234.

SHARPE, C. E., and SHARPE, K. J. 1976. A preliminary survey of engraved boulders in the art sanctuary of Koonalda Cave, South Australia. Mankind, vol. 10, no. 3, pp. 125-130.

SHARPE, K. J., and SHARPE, C. E. n.d.[1]. The Upper Chamber of Koonalda Cave, South Australia: A second report. Preprint.

_________. n.d.[2]. Koonalda in the Nullarbor: A prehistoric Australian heritage. Preprint.

STUBBS, D. 1979. Prehistoric art of Australia, abridged ed., 80 pp., illus. Sun Books, Melbourne.

WALSH, W. P. 1964. Unexplained markings in Kintore and Cutta Cutta Caves, Northern Territory, Australia. Helictite, vol. 2, pp. 83-91.

WRIGHT, R.V.S., ed. 1971. Archaeology of the Gallus Site, Koonalda Cave, 133 pp., illus. Australian Institute of Aboriginal Studies, Canberra.