AR23.
23 June 2008
Copyright © 2004 by Kevin Sharpe. All rights reserved.
Unpublished.
THE UPPER CHAMBER OF KOONALDA CAVE, SOUTH AUSTRALIA:
its Rockfalls, Their Weathering and Use
by
Kevin Sharpe
The Graduate College, Union Institute and University, Cincinnati, Ohio
Harris Manchester College, Oxford University
10 Shirelake Close,
Oxford OX1 1SN,
United Kingdom
kevin.sharpe@tui.edu
www.ksharpe.com
·
Ask Robert re publishing: send him finished ms
plus other comments, given that he’s already seen it.
·
Need to discuss AR16: Gallus’s paper on the Upper Chamber.
·
Dates from Gallus (pers. comm. 7
November 1978):
I have two new C14
dates.
A.
ANU-1201.
‘Radiocarbon age’ larger than 20,450
B.P. Sample size was only 12%
of the requirements, thus this [is] a minimum
age. Personal comment by H. A. Polach: ‘Low C14 activity of your sample combined with small sample size does not allow us to define an
absolute age ± error in conventional terms. However the detected C14 activity places the radiometric age within the minimum age
given and 52,000 B.P., with 95% probability with an indicated age of 29,000
+ 11,600 – 4,600.
B.
Pieces of wood and wood-pulp. Collected from surface
dust, North West Passage (‘Sanctuary’) and surface of Crevice 1c. ‘Radiocarbon age’: 18,200
± 300 B.P.
I would say that this sample marks the end of activity in the North West
Passage. The original old floor on which the polished boulders stand must be
somewhat older, and the wall-markings somewhat older as the original floor from
where they have been made has already collapsed in several areas.’
·
What is Gallus’s 1986 paper that Bednarik thinks is the
last word?
·
Print. Leslie read. Revise.
·
Plates and
figures.
·
Print. Send out.
KEYWORDS.
Australian archaeology – finger flutings – Koonalda Cave –
line markings – Nullarbor Plain – salt weathering
ABSTRACT.
This paper describes the Upper Chamber of Koonalda Cave in
which, perhaps 19,000 years
ago or more, people engraved the walls, marked the walls with their fingers,
engraved and arranged the rocks on the floor, and deposited animal parts, all under light of torches some of whose
remains still sit where left on the tops of boulders. The visitors may have
used the whole of the Upper Chamber in this way, though rockfalls more recent
cover the majority of evidence for this. The Upper Chamber contains at least five
rockfalls of different ages and degrees of weathering. The boulders of the
oldest collapse show prehistoric human use and line markings, with sparse or no
markings on other areas. Under the most recent rockfalls, however, lies the
original smooth, rounded, and marked boulder floor. The boulders on the floor
possibly weather by a process called salt weathering to become smooth and
rounded.
CONTENTS.
Introduction. 2
Part One:
Description of the Upper Chamber 2
Rockfall
C. The Smooth Rounded Boulders. 2
Rockfall
A.. 2
Rockfall
B.. 2
Rockfall
D.. 2
Rockfall
E.. 2
Rockfall
F. 2
Stratigraphy. 2
Part Two: The
Smoothing and Rounding of the Boulders. 2
Claws
and Guano of Bats. 2
Water
Action. 2
Water
Dripping. 2
Dissolving
in a Pool 2
Air
Movement 2
Insolation. 2
Salt
Weathering. 2
Summary and
Conclusions. 2
Acknowledgements. 2
Note. 2
References. 2
Referees’
Comments. 2
FIGURES.
1.
Koonalda
Cave – location and floor
plan (after J. B. Hinwood, 1960).
2.
Northwest passage of Koonalda Cave (after J. B. Hinwood, 1960).
3.
Floor G5-I
(after Chadwick and Smith). Apart from those mentioned, the edge of the floor
is defined by what appear to be built-up walls of rocks and stones.
4.
The expected ‘stratigraphy’ of boulders as a result of
differential rounding of variously aged rock falls (after Jennings), and the
expected stages in the smoothing and rounding of boulders by salt weathering.
PLATES.
1.
The Upper Chamber of Koonalda Cave, looking from the
‘directional stele’ toward the ‘ramparts’ over Rockfall C (photograph by Kevin
Mott).
2.
Rock F6-b,
floor F6-I. Note the
engravings and the stones leaning against them. Scale in centimeters.
3.
‘Cache’ in floor E4-I showing the set of vertebrae in the dust. Scale
in inches.
4.
The two incised rocks (upper one, D2-a, and lower one, D2-b) at the entrance to the
‘cavern.’ D2-b
appears as a ‘snake head.’ Scale in centimeters.
5.
Floor G5-I
viewed from the northwest. Note the apparently placed rocks around the
circumference of the floor and the two round and smooth boulders G4-a and G5-b on the upper right. The skull found on
the floor was on the stone to the left of the ruler. Scale in centimeters.
6.
The ‘bird’-shaped smooth and rounded boulder that is
incised and partially under the rock fall on the slope down to the ‘squeeze.’
Scale in centimeters.
Many scientists have investigated Koonalda Cave
in the Nullarbor Plain of South Australia, starting with Hunt and Wells in 1904 (Wells and Hunt 1919), and perhaps McCullough in 1892 (McCullough 1979). Gallus first recognized
the archaeological significance of the cave in 1956 and continued his work there during
the following two decades (Gallus 1964,
1966, 1968a-d, 1971, 1977; summarized in Flood 1997). The Australian Institute of
Aboriginal Studies also mounted a multidisciplinary expedition in the 1960s to research the cave
(Wright 1971a-d;
for other archaeological articles on Koonalda Cave, see Cox 1960; Pretty 1960; Pretty and Gallus 1967; Wigley 1966).
The floor of Koonalda
Cave comprises three
passages in the limestone and that end in lakes (see Figure 1 and Thomson 1949: 49-50).
Gallus undertook most of his excavations in the Gallus Site, the lower level of
the Northwest Passage, just in from the cave
entrance. This paper focuses on the next level of this passage, higher, now
dry, and called the Upper Chamber (see Figure 2).
FIGURE 1. Koonalda Cave
– location and floor plan (after J. B. Hinwood, 1960).
FIGURE 2. Northwest
passage of Koonalda
Cave (after J. B.
Hinwood, 1960).
The 30-meter
ascent from the Gallus Site to the Upper Chamber reaches the top at a series of
large rocks that resemble ‘ramparts’ and overlook the Gallus Site. The
remainder of the trail to the end of the Upper Chamber clambers over and around
boulders. It descends a little from the ‘ramparts’ and then ascends to the
highest point of the path at what was previously called the ‘directional stele’
(Sharpe and Sharpe 1976:
127). It then
passes nearly level and then descends a little to a 10-meter drop before the ‘squeeze,’ a 30-centimeter-high, five-meter-long
channel. On the other side of the ‘squeeze’ juts a small
platform 18 meters above a lake, the
end of the Northwest Passage and the terminus
of the West Passage of the cave.
Work proceeded during 1973 in
the portion of the Upper Chamber where fine lines incise some of the boulders
on the floor. Whitehead and Sharpe plotted a large number of these marked
boulders, some with torch stubs and bones sitting on top of or beneath them.
Their preliminary mapping of the engraved boulders led them to visualize the
Upper Chamber as a series of engraving-flanked pathways (Sharpe and Sharpe 1973: Figure 2). The chamber is now interpreted
more broadly. The investigations continued in 1976 with the methodology of Marshack (Marshack 1972a, 1972b, 1972c, 1975,
1976a, 1976b, 1977, 1979). Sharpe noticed at this time that the whole
floor of the Upper Chamber comprises boulders smoothed and rounded to varying
degrees and of human use, even if more recent rockfalls cover the majority of
these boulders and artifacts. That is, the floor of the Upper Chamber results
from a number of different rockfalls of various ages, the oldest of which is
now smoothed and rounded. Prehistoric peoples used this floor and often altered
it. But most of this rockfall and its human evidence are now hidden under more
recent rock fall. This report attempts to establish this point, to describe the
floor of the Upper Chamber and its artifacts from this perspective, and to
suggest how the boulders might become smooth and rounded.
The paper comprises two parts. The first describes the
various rockfalls in the Upper Chamber, mentioning the artifact finds and how
people changed the floor. The sequence used to name the rockfalls starts at the
‘ramparts’ and ends at the ‘squeeze’ (see Plate 1 for photographs of the Upper Chamber and Figure 2 for the layout of the Upper
Chamber divided into rockfalls). Rockfalls C and F are probably the oldest,
followed by A and D, and then, most recently, B and E. Rockfalls C and F
underlie the other rockfalls and consist of smooth and rounded boulders with
line incisions on their surfaces. (This paper does not discuss the line
markings in any detail, but other reports cover the subject (Sharpe 1982, 2000; Sharpe and Sharpe 1980).)
The second part of the paper discusses the weathering of the
boulders in the chamber: How do they become smooth and rounded?
The 1973
expedition to Koonalda most extensively investigated what is now called
Rockfall C, where the majority of boulder line markings appear. Irregularly
shaped (see Figure 2),
Rockfall C sometimes merges into its neighboring rockfalls on the southeast
because they too are smooth. Rockfall C runs from the top of the climb to the
Upper Chamber almost as far back as the ‘directional stele’ and from one wall
to the other. Its boulders vary in shape and size, but all are smooth and
rounded at least on their upper sides. A fingernail – not a touch – can mark
their generally firm surfaces. Parallel lines incise many of them, but a thick
dust often covers the boulders and the markings.
This area of smooth and rounded boulders is the most humanly
used portion of the present surface of the Upper Chamber. In their report on
the Upper Chamber – after their first visit to Koonalda – Whitehead and Sharpe
(Sharpe and Sharpe 1976)
chiefly see it as a series of pathways that engraved boulders define and
delineate. Its use is now seen as far more complex, comprising the construction
and use of floors, the engraving of rocks and walls, the deposition of bones,
the placement of torches for illumination, the following of pathways, and so
on.
This paper adopts a particular terminology to describe the
human use of the Upper Chamber. Terms such as ‘area’ and ‘floor’ are applied in
a casual geographical way. The ordering A through F of the various rockfalls of
the Upper Chamber follows their geographical sequence from the ‘ramparts’ to
the ‘squeeze’; the lettering says nothing about their relative historical ages
(Figure 2). Terms
such as ‘floor’ are archaeological and suggest the use for some activity or
other of that section of the cave floor, albeit perhaps some millimeters below
the present surface, by prehistoric Aborigines. Many geographical features and
floors are given human terms, such as ‘cavern’ and ‘Chadwick’s Hole.’ A grid
reference system for the Upper Chamber is also adopted to number rocks and
floors; in the grid square E4,
for instance, are rocks E4-a
and E4-b (the last
letters are applied with no particular sequence in mind) and floor E4-I (letter postscripts are applied
to rocks and Roman numerals to floors).
Floor F6-I.
In their report, Whitehead and Sharpe mention the ‘elephant head’ rock in an
area they designate as P2
(Sharpe and Sharpe 1976:
128-29) between two pathways 15-8 and 7-8. This rock lies in what is now
called floor F6-I.
Floor F6-I,
approximately 2.5
by 1.5 meters,
looks past a few small and partially
incised stones on its southeastern end to a geographical area called the
‘amphitheater’ some three meters below. A large incised boulder (whose markings
have rather deteriorated) sits on its northeastern side and, behind this, the
floor of the cave drops sharply into a sort of pit some four meters deep. In
this pit are lines pictured in Whitehead and Sharpe’s report (Gallus 1977: Plate 2; Sharpe and Sharpe 1976: 128). The well lined ‘elephant head’ rock,
F6-b, shaped as its
name suggests and with its ‘trunk’ reaching onto the periphery of the floor,
half encloses the northwestern side of the floor.
PLATE 2. Rock F6-b, floor F6-I. Note the engravings and the stones leaning
against them. Scale in centimeters.
Between F6-b
and E6-a, a medium
sized and incised rock on the southwestern edge of the floor, lie several
stones. Some of the stones appear loose and rough (see Plate 2). Some are small
and rounded and their crystalline surfaces do not point downwards, in contrast
with those of almost all other rocks in the Upper Chamber. They may, therefore,
not sit in their original positions. Moreover, as Plate 2 shows, they cover a well-defined series of
line markings on rock E6-a,
to the point of touching the incisions. These two factors suggest the
intentional construction of the floor by prehistoric people.
When this construction occurred requires investigation.
Perhaps people created the various constructions noticed in the Upper Chamber
(including the series of floors in Rockfall D) during a particular period. Or
perhaps construction occurred throughout the time that people visited the cave
(see Gallus 1971:
112-23).
Floor G6-I.
The second floor to mention lies near what Whitehead and Sharpe call point 8, at the foot of Rockfall D. Its
smooth rounded boulders back onto and, in some instances, sit under the rougher
Rockfall D (see Figure 2).
Two piles of stones help define the floor (one on the southeastern edge and one
on the northwestern edge); perhaps people created the piles, especially the
southeastern one, in the process of clearing and forming the floor. On this
pile were found pieces of wood and charcoal, the latter also scattered
abundantly over the floor.
Floor E5-I.
Three pieces of bone were found in 1973
at the point where path 3-4 leaves path 7-8.
The longest of the bones, presumably part of a large marsupial shoulder blade,
is about 15
centimeters long and was half buried in the dust. The revised perspective of
the Upper Chamber places these pieces of bone near the southeastern entrance of
floor E5-I.
Floor E5-I
measures approximately 2
by 2.5 meters and
occupies two levels, a small hump
separating its smaller and hollowed
southwestern section from its northeastern half. The rocks that form the
southwestern boundary are of two types. First, two rounded and smoothed
boulders; second, smaller smoothed
rocks with their jagged sides showing (possibly their original undersides).
This suggests intentional placement. Behind these rocks lies floor E4-I that is discussed below and that
centers on a ‘cache’ containing several vertebrae and engraved plaques.
Prehistoric visitors to the cave may have also placed the
stones on the northwestern end of floor E5-I.
Further, perhaps the stone halfway down the slope from this end of the floor
has recently slipped from the top (or someone has knocked it off) because it
shows no sign of settling into the floor it rests on and a space exists for it
with the line of other stones. This row of rocks separates floors E5-I from E5-II. A large rounded boulder has recently
suffered vandalism; someone has cut (presumably with a handpick) a pair of deep
lines into the stone and, further up it, someone has hacked a hole for what
looks like a candleholder.
Floor E4-I.
As reported previously (Sharpe and Sharpe 1976: 127),
a curved opening enters the boulder floor of the cave on the southeastern side
of rocks E4-a, E4-b, and E4-c (that Sharpe and Sharpe 1976 calls 1M, 1K,
and 1L
respectively). This opening contains a number of vertebrae and a flat stone
covers it (which is ill-fitting and wobbles when stood on; hence the find) (see
Plate 3).
PLATE 3. ‘Cache’ in floor E4-I showing the set of vertebrae in
the dust. Scale in inches.
Whitehead and Sharpe call this a ‘cache,’ and, on the 1976 visit, Chadwick investigated
it further. (The following derives from Chadwick’s notes; such remarks will be
prefixed in the future with ‘after Chadwick.’) The four vertebrae sit in dust
(see Plate 3) and
rest on a rough piece of rock, the removal of which exposed smaller pieces of rough pebbles. Three centimeters of
earth later lies a piece of incised limestone, which Chadwick drew and
replaced. It measures 6
by 9 centimeters and
four pairs of parallel lines incise one side. Subsequently, another piece of
incised rock 6 by 9 centimeters appears (it may be
useful to compare these two lined plaques with others found in Koonalda and the
Devil’s Lair Cave – see Bednarik 1991:
27, 1998; Dortch 1976; Gallus 1971: Plate 1X-6).
After removing dust and loose pebbles, Chadwick drew the
cave floor around the ‘cache’ and more thoroughly investigated for
stratigraphy. Red dust covers the floor itself. He describes the surface rocks
as flat and wedged together in sinter, making them difficult to remove. Twigs
and carbon appear under some of the looser of these rocks in places where they
could not at present have slipped from the surface. The stratigraphic sequence
appears as follows: a surface of animal
droppings (bats, etc.); gray dust and carbon; wood; rosa
dust; red dust; rock. Under two rocks in the floor, Chadwick found this stratigraphy:
rock covering; twigs, bone, and animal
droppings; gray dust intermingled with animal
droppings; fine red dust; and flat base rocks.
These minor investigations suggest a little about the human
use of the floor. The engraved plaques may have been placed in the ‘cache,’
covered with layers of pebbles and dust, the vertebrae deposited and, finally,
the ‘cache’ capped with a flat stone. The floor around the ‘cache’ may have
been intentionally constructed to make it reasonably flat and around (or before)
a large engraved smooth and rounded boulder.
A more extensive excavation between E4-b and the ‘cache’ occurred in 1975 and provides more
information on the local stratigraphy (after Michael Smith). The excavation
descended at least 40
centimeters (in some places rocks stopped it), and consisted of a ten
centimeters square test trench and a main excavation of about one-third of a
square meter. The following summarizes the stratigraphic sequence:
·
Level 1
comprises one to two centimeters of loose surface material in which are patches
of unconsolidated fine grayish-pinkish soil. In it are also bat bones and
carbon (mainly from the gray unit).
·
Level 2
comprises consolidated red soil containing a lens of gray material. In this
level (three centimeters deep in the test trench) is a large piece of carbon,
perhaps burned torch, with smaller
pieces of carbon surrounding it.
·
Level 3
comprises several different lenses and layers and descends five centimeters in
the test trench. It occurs sooner on the E4-b side of the trench than on the E4-c side. A pale sand, very compact
and perhaps laminated, occurs even closer to the E4-b side. The level in general comprises a
light red-brown speckled soil with thin bands of a dark chocolate brown, a very
fine material (almost like an ochre; perhaps layers of bat guano) that appears
mottled and seems to cling to the rocks.
·
Level 4
comprises a fine, unconsolidated, light-pink soil. Fine crystalline droplets
lie in the lower part of the level (plus gravel and small
pebbles up to a centimeter in diameter, and fine dark-red lenses), and a number
of the rocks occur throughout it. It was not excavated to its full depth in the
main trench and seems to continue down; in the test trench, it descends ten
centimeters.
·
Level 5
comprises a fine, unconsolidated, white soil and appears only in the test
trench where it descends ten centimeters. Carbon occurs in all levels except
the lower ones.
The ‘Amphitheater.’ Little can be added to the 1973 description of path 3-4 (Sharpe and Sharpe 1976: 127), except that a number of paths descend the
steep slope and that this section of the Upper Chamber looks like an
‘amphitheater’ overlooking a large rock D6-e
that resembles an ‘altar’ or ‘monument’ stele. These subjective names may help
in visualizing the area.
The 1973
path from 3 to 4 moves from the northwestern top of
the slope down it and toward the southeast to about rocks D6-a and D6-b or to the side of D6-c and D6-d, and then in the other direction,
completing the descent to the large ‘altar’ stele D6-e. Pieces of wood lie on the slope. The
path passes by rock D6-f
that features the series of interconnected, smoothed tunnels the previous
report mentions (Sharpe and Sharpe 1976:
128).
Large crevices perforate the floor around D6-e and a deposit of red ochre
emerges from the wall beside it. Its top bears a pocket containing charcoal
ends (see Gallus 1977:
Plate 9; someone’s
boot destroyed them between the 1973
and 1976 visits).
The previous report mentions these last two features (Sharpe and Sharpe 1976: 127).
On the northwestern side of D6-e sits the huge slab of rock on which lie the
floors including F6-I.
The slab forms one side of a narrow chamber some four meters long that was
described as ‘a sort of pit’ in the above discussion of floor F6-I. On the D6-e side of the slab appear a number of
examples of the boulder line marks (see Plate II in Sharpe and Sharpe 1976 for a section of this). Two
lines emerge from a natural hole on the slab wall of the chamber and run down
the face of the rock for 20
centimeters (see the earlier report, Sharpe and Sharpe 1976: 128; see also Gallus 1977: Plate 1). A series of deep and rough-looking line
markings exist on the wall side of this chamber (perhaps the lines that Edwards
and Maynard describe as a ‘short set of very deep parallel finger
marking…deeper than any others observed in the cave…[and which] seem more
eroded than usual’ (Maynard and Edwards 1971: 67)).
Someone has recently dug at the base of these lines.
The earlier report also mentions a flat-topped rock that
looks like a ‘work bench’ and that stands between paths 3-4
and 5-6. This is now called D5-a. It sits at the top of the slope
overlooking the ‘amphitheater.’ The flakes of flint that the previous report
mentions on this ‘bench’ are amorphous, but prehistoric visitors may have used
them to engrave lines (a flake from the floor of floor G5-I – and that may have been an engraving
tool – is discussed below).
The ‘Terrace.’ Whitehead and Sharpe call the gentle ‘terraced’
slope from the centerline of the Upper Chamber – along which traverses the main
path – and down to the wall, path 5-6. A similar slope of incised
rounded boulders (the ‘amphitheater’) bounds it on the northwest, but the
‘terrace’ generally comprises smaller
boulders or just dirt. The foot of the slope consists mainly of relatively new
rock collapse (Rockfall B; see Figure 2)
and this too helps separate the ‘terrace’ from the ‘amphitheater.’ In
prehistoric times before the collapse, however, this whole area may have been
one, bound on the northwest by the huge slab on which lie the floors such as F6-I, and on the southeast by the
‘ramparts,’ the beginning of the Upper Chamber. Another collapse (Rockfall A)
at present bounds it and some of its rocks lie on the smooth rounded boulders
on the ‘terrace’s’ southeastern side (for instance, they lie on C6-a and C6-b). The slope looks as though it is
terraced because it comprises a number of rocks holding back the dirt, a very
red dust. Someone recently has apparently dislodged one of the rocks on the
slope (downhill from C5-a;
not all the rocks are numbered) because beside it is a hole in which it would
fit, and what probably was its crystalline underside now forms a side.
Floor C5-I.
Two parts of the ‘terrace’ are singled out as floors though a number of other
of its sections might be similarly labeled. Floor C5-I begs for the name ‘floor’ because here
was found made one of the most impressive artifacts of the 1973 visit to the cave: a large (15 centimeters) torch, charred on
one end, thickly coated with dust, and sitting on a boulder (see Gallus 1977: Plate 8; Sharpe and Sharpe 1976: 127). A considerable amount of charcoal
sits on the floor around the boulder.
Floor D5-I.
This floor lies further up the slope than C5-I – beside the main path through the Upper Chamber
– is narrow (about 1
by 2.5 meters),
and overlooks the slope down to the wall. A significant amount of wood rests on
this floor.
Crevice C4-a.
This feature goes into the cave floor beside the rock of the same number. The
earlier report also mentions it (Sharpe and Sharpe 1976: 128). This may have marked an entrance to an area of
Rockfall C-type smooth and rounded boulders, but Rockfall A now covers it. It
is discussed further when considering the latter rockfall.
Hole Beside Rock C4-b.
A horizontal hole enters under the rocks beside rock C4-b. This hole first drew investigators’
attention because the remains of an insect sit on a rock that forms part of the
floor of the hole (for a thorough discussion on Nullarbor Cave
insects, see Richards 1977).
A series of lines mark the rock at the end of this tight hole, a difficult
place for a human to reach. These lines are discussed in more detail when
Rockfall A is considered.
The ‘Cavern.’ Whitehead and Sharpe’s 1973 path 9-10
travels from the center of the Upper Chamber to what is called the ‘cavern’
(see Sharpe and Sharpe 1976:
128), a deep
closet-type of hole into the cave floor near the wall. A deposit of red ochre
emerges from the wall at this point and the surface of the cave floor comprises
white dust. Incisions mark the slope immediately before its entrance (rocks D2-c and D2-d). The ‘snake head’ rock (D2-b) and its mate (D2-a) at the entrance also exhibit
lines (see Plate 4;
Sharpe and Sharpe 1976:
Plate III), some of the most striking line markings in the Upper Chamber.
PLATE 4. The two incised rocks (upper one,
D2-a, and lower one,
D2-b) at the
entrance to the ‘cavern.’ D2-b
appears as a ‘snake head.’ Scale in centimeters.
The ‘cavern’ opens into the floor of the cave like a wedge
and extends about four meters. Its entrance at the large end of the wedge
(where sit the two incised rocks D2-a
and D2-b) measures
about one meter square. The entrance to the ‘cavern’ is space between D2-c and D2-e, the ‘cavern’s’ tail or thin edge
reaching under D2-e
and C2-a. Wood,
charcoal, animal droppings, a small amount of bone, and the remains of various
insects sit on its floor. More line markings appear about halfway down it.
The floor of the ‘cavern’ offers the second most intriguing
thing about it after the markings at its entrance. It comprises an infill of
loose and small rubble, not sintered
or crystalline. Was this rubble placed here or did it fall or roll in
naturally? No place appears to exist for it to fall into the ‘cavern’ given its
present roof. If it somehow fell in from the front entrance area, a hill or
mound would be expected up to the fall-in point and no evidence is seen of
this. How did the rubble even itself off along the four or so meters of the
‘cavern’ to create a more-or-less flat floor? Perhaps prehistoric people found
the rubble here and placed the massive and incised rocks at the entrance (a
super-human effort). Perhaps a rockfall created the infill and the roof of the
‘cavern’ fell into place later (unlikely, because the ‘cavern’s’ ceiling
boulder D2-a
probably belongs to the original smooth and rounded boulder floor). Or, perhaps
the prehistoric visitors placed the rubble here. This matter needs investigation.
The prehistoric users of the Upper Chamber may have filled the ‘cavern’ to
cover something, as they may have done with the ‘cache’ mentioned above.
The ‘Fireplace.’ A ‘fireplace’ was found to the
northeast of D2-f.
Chadwick excavated and unearthed (after Chadwick):
·
Pieces of rock with a white dust covering sit on
the surface. Bones (some burned) of small
animals and birds lie between the
pieces of rock. Because the small
bones are not under the rocks, they probably fell there.
·
A tooth, claw, and leg bone of an animal, probably a wallaby (the claw may have served as
an engraving tool) lie under the rocks.
·
Five centimeters of rosa
dust interspersed with bones large and small,
and large pieces of charcoal and wood.
·
Five centimeters of impacted large pebbles and
gravel-like mixture interspersed with a smaller
amount of charcoal and bones.
·
Ten centimeters of large rocks in rosa dust.
·
Twenty-one centimeters of larger rocks.
·
Eight centimeters of the floor of rosa dust.
‘Chadwick’s Hole.’ Because, as will be mentioned
below, a hole on the northeast wall is named after Edwards who (with Maynard)
describes it in papers on the cave, it seems apt to name a hole on the other
side of the cave after its present-day finder, Chadwick. ‘Chadwick’s Hole,’
like ‘Edwards Hole,’ lies in the original floor of smooth rounded boulders and
goes down as a continuation of the wall of the cave. It descends behind the
rocks against the wall and directly opposite rock D2-f, extending some five meters in three
levels (after Chadwick). The small
entrance leads down to the first level, a slope where investigators can
nestle and contort their bodies through its rough rocks. The second and main
level opens out to a relatively large chamber, approximately 3 by 2 by 2.5
meters, and horizontal so that a person can stand in it. An incised smooth and
rounded boulder sits on its floor. The third level opens off a hole in
the floor of the second level and goes down some 2.5 meters. Deep and eroded lines incise
the smooth and soft wall above it and lines also exist around its edge. Rosa dust has collected throughout the hole, with white
dust predominating near the walls.
Chadwick crawled above and over the jagged rocks, up through
the top of the hole toward its roof, and found incisions on the smooth surfaces
of the fallen rocks. Only the surfaces that are smooth and turned outwards
appear to display the marks, not the jagged surfaces.
Other Rocks, Wall Sections, and Floors. A large
number of the incised boulders within Rockfall C have not been mentioned in
association with other features. One much photographed is rock F6-c, which sits on the
northwestern-wall side of floors G6-I
and II. The wall around this point shows many line markings (the sections of
the wall that Edwards and Maynard call ‘N. E. Wall Sections 2 and 3’; see Maynard and Edwards 1971: 67). Whitehead and Sharpe’s report mentions two
other features in this part of Rockfall C: a floor hole (near 8 in G6), and rock F6-d
that lies on the surface near the wall and resembles a torso (see Sharpe and
Sharpe 1976: 128-29). A number of other floors also exist in
the rockfall, for instance floor F5-I
on which sit three vertebrae and what looks like a shoulder blade.
A red dust often sits on the rocks that comprise Rockfall A
(see Figure 2) and
they appear smooth, but without the extensive rounding of their companions in
Rockfall C. The previous report defines the northwestern segment of Rockfall A
as path 13-14 (Sharpe and Sharpe 1976: Figure 2) but, because various paths move
across the area, this nomenclature may confuse. No lines appear on its rocks
despite the many suitable places for incising. Humans may have used it, but
apparently nowhere near as much as Rockfall C of smooth and rounded boulders.
Two possible floors, B2-I
and C2-I, lie on a
plateau of the rockfall and may have had human use. The dust slope up beside C2-a and C2-b, and around C2-b, provides access to them from the
‘cavern.’ Two amorphous flint flakes and a number of twigs lie under two loose
floor rocks on floor C2-I.
The Gallus-Site sides of the two rocks B2-a
and C2-a – two of
the ‘ramparts’ – show faint line markings (it is unclear whether or not these
rocks are old smooth and rounded boulders). A collection of dead insects sits
between C3-a, C3-b, and D3-a. When standing between D3-a, D3-c, and C3-b
on the one side, and D4-a,
D4-b, and D4-c on the other, the difference
between the two sets of rocks becomes apparent. The ones on the northeast are
large, rounded, and incised; the others are usually small,
more jagged, and not incised.
The northeastern side of Rockfall A looks like an old
rockfall that is weathering. It lies between the central pathway and the
northeastern wall, and the ‘terrace’ and the ‘ramparts.’ Some deteriorated line
marks incise rocks in this section of the fall near the wall.
Rockfall A covers the older area of incised smooth and
rounded boulders because such rocks with incisions lie under it. Four examples
of this were found accidentally. Two features were mentioned above:
·
Crevice C4-a.
·
The hole beside rock C4-b. This warrants further description. It
is a horizontal hole with two rocks forming its sides, another its back, and a
fourth its ceiling. Looking from its northeastern end through the few
centimeters between the ceiling and the upper side of a floor boulder, one can
see lines incised on the floor rock. This arrangement of rocks, the top ones
(roughish non-incised boulders) having fallen into their present position on
top of incised smooth and rounded boulders, looks even more convincing from the
southwestern side.
·
In addition to the above two examples, line
markings show on smooth boulders protruding about two thirds of the way up the
slope to the Upper Chamber from the Gallus Site. These rocks may represent the
forward remains of the incised smooth and rounded boulder floor.
·
On the last day of the 1976 visit to Koonalda, a roll of exposed
film fell between two rocks from the large boulder slab on the top of the
‘ramparts’ B4-a. It
could be seen down among the rocks, but was out of reach. Small stones were removed and Chadwick squeezed in to
retrieve the film. He also found a large vertical face of line markings about
four or so meters down.
A smaller and
obviously more recent rockfall, Rockfall B, lies against the wall a little
further up the chamber than Rockfall A. The only incised lines in this collapse
appear on slabs that fell from the wall, and probably incised when the slabs
formed part of the wall. Some markings here are quite dissimilar from others in
Koonalda Cave (apart from a set on rough boulders
near the ‘squeeze’). A large floor hole beside the wall here is called
‘Edwards’ Hole’ because it is probably the one he and Maynard describe (Maynard
and Edwards 1971:
64) in this section
of the Upper Chamber.
Rockfall D contains the greater part of what was previously
called path 15-8 (Sharpe and Sharpe 1976: 127-28) and the ‘directional stele.’ It is possibly the
second most recent collapse (see Figure 2)
that prehistoric humans used, though few lines mark its rocks (one example
appears in floor G6-I,
on a rock whose other side helps define floor G5-II). Many bones have been found here. It may well
be the last rockfall that Aborigines used before the more massive Rockfall E
and the cave’s abandonment. It is obvious, by looking at the edges of this
area, that it comprises rough rubble fallen onto the smooth and rounded
boulders.
FIGURE 3. Floor G5-I (after Chadwick and Smith). Apart from those mentioned,
the edge of the floor is defined by what appear to be built-up walls of rocks
and stones.
Floor G5-I
(see Plate 5, a
photograph into the southeastern half of floor G5-I, and Figure 3). As Plate 5
shows, the rockfall has been mostly cleared from this floor and piled up around
its edge. Its ‘entrance’ lies between two smooth rounded boulders, G4-a and G5-a, on the right side of the photograph and
which lie at the side of the main path, 1-2, through the Upper Chamber. The
floor is ‘exited’ by an opening in the stone wall around it, about the
center-left of the photograph. This leads onto other floors in Rockfall D.
Lines sparsely incise the southwestern side of rocks G4-a and G5-a, but no lines appear on their floor G5-I side apart from two pecked
into G5-a. As the
photograph also shows, half of the area, the southeastern half, lies at a lower
level, perhaps scooped out. A flat piece of rock balancing on top of two side
rocks some 40
centimeters high sits at the floor’s extreme west, a little below the base of
the photograph. A few pieces of wood sit inside this ‘cache.’
A 25-centimeter
square on the southeastern edge of the floor was excavated in 1975 and taken down five
centimeters to the limestone rock base (after Smith).
·
Level 1,
down about one centimeter from the surface, comprises fine gray soil with
carbon and bones.
·
Level 2
of 1-2 centimeters is of consolidated
pink soil, similar to but not quite as consolidated as level 2 in the trench by rock E4-b (see above). It contains several
large pieces of charcoal, wood, burnt wood, pebbles, and a rodent jaw.
·
Level 3
is of a very dark chocolate-red loam similar to the chocolate-brown bands in
level 3 in the
trench by rock E4-b.
It varies in thickness from a few millimeters above the rock floor to solid
pockets five centimeters deep in crevices in the rocks, and it contains carbon.
·
Level 4
lies deep in the crevices of the limestone floor and comprises a white
unconsolidated matrix. It also contains carbon.
·
The base floor comprises packed and decomposing
limestone rocks that are flat, giving the appearance of being laid like
cobbles.
The stratigraphy resembles that found in floor E4-I and in the trench beside rock E4-b.
Chadwick extended this investigation in 1976. He found charcoal in fine
flakes embedded in the compact dust and scattered all over the floor. The
stratigraphy Chadwick reports is: 0.5
centimeters of gray sterile dust; one centimeter of rosa
dust; 0.5
centimeters of compacted dark-red, clay-like dust; unsintered rock and loose
rubble with no evidence of crystallization and lying in coarse rosa dust. The two reports agree.
Chadwick also found a flint flake in the compacted red dust,
perhaps an engraving implement. Gallus suggests (pers. comm.) that it resembles
a variety of fine burins he found in a level in trench II of the Gallus Site,
under what he calls ‘rough red’ (see Gallus 1971).
A number of bones have also been found on the floor. The
skull of a kangaroo without its mandible sits on a rock not far off the floor.
A small pocket of bones and a 10-centimeter long cut or broken
piece of bone lie on the northwestern section of the floor. Prehistoric users
of the cave may have brought these bones, or the pieces of animals of which these bones are the remains, into the
Upper Chamber. (Bones of small
marsupials, birds, and bats regularly occur in the Upper Chamber; these
presumably have a non-human source.)
Floors G5-II,
III, IV, G6-III.
These four floors we not investigated as thoroughly as G5-I, but they show similar characteristics:
the placing of stones in a wall around their edges (especially G5-III, and between G5-II and floor G6-I).
Of note are the well-incised smooth rounded boulder on the
right and part of floor G6-I.
Of note also are the stones apparently placed on the left of this rock and that
form part of the shared boundary. (The rough, recently fallen rocks clearly
overlie the smooth rounded boulders here.) Other characteristics these floors
share include the apparent ‘scooping-out’ of floors (especially G5-III). Both also contain the
remains of pieces of large marsupials: two vertebrae on G5-IV, and the smashed remains of bones
including what looks like a large kangaroo skull on G5-II, probably fallen off a nearby rock.
Virtually no evidence of line incising was found on the inside faces of
boulders around these floors; however, hardly any surfaces suitable for
engraving exist here either, and the smooth rocks usually are in the floor and
walked over.
The ‘Directional Stele’ H5-a (see Figure 2). This thin, vertical boulder marks the high point
of the path from the ‘ramparts’ to the ‘squeeze’ at the rear of the Upper
Chamber, and assists a person walking this path by standing out as a point to
which to proceed. The previous report, therefore, called it the ‘directional
stele’ (Sharpe and Sharpe 1976:
127). One side of
this rock, that facing the usual path taken through the Upper Chamber, is
smooth – and sparsely incised – and the other side is rather jagged. It
probably formed part of the roof, therefore, before it fell and tilted. It also
lies on smooth rounded boulders and appears to have lain there longer than the
rocks in Rockfall E immediately to the northeast of it; further, a white dust
covers Rockfall E while red dust also adheres to the ‘directional stele.’
Smooth rounded boulders lie to its southwest for a few meters before further
fresh rockfall. It is probably the northern-most tip of the rockfall in which
lie floors G5-I to
IV, G6-III, at least
the northern-most tip visible as the present surface. A large wooden torch lies
beside it.
The presumably most recent rockfall, Rockfall E, lies to the
northwest of Rockfall D (see Figure 2).
It comprises jagged white rocks and shows little sign of smoothing. There is no
red dust on it. Also, within it exist no signs of human occupation: no floors,
bones, or torches. No line markings, either. The most extensive search for the
floor of incised smooth and rounded boulders, which possibly sits under it,
occurred here.
Hole Number 1.
The search for line markings on smooth rounded boulders underneath Rockfall E
started with a small hole (about 80 centimeters high and one meter
deep) northeast of the ‘directional stele’ (see Figure 2). Rubble separates two neighboring rounded
boulders and what was probably a large piece of roof (the underside is smooth
and not incised, and the upper side is rough) sits on their tops. If a piece of
the rubble up against the smooth rounded boulders covers lines, some lines
might cease abruptly at the edge of the piece of rock and, when the rock is lifted
away, the lines should continue under where the rock covered. Some incisions on
the smooth boulders did cease at a piece of rough rubble leaning up against one
of the boulders. This was tried. The strong line markings that went up to the
edge of the piece of rubble did continue straight on over the section of the
smooth rounded boulder that the piece of rubble covered. These pieces of rubble
had almost sintered together and were tightly attached to their surrounds.
Little crystallizing had formed underneath them. After the first piece of
rubble was removed and the line markings exposed, a second piece was removed,
exposing still more markings.
What might have formed the original floor was sought because
it might have had wood and/or charcoal on it. The base floor is a long way down
beneath the rubble, however, and, anyway, wood could fall through to it. The
rock stood on to see the line incisions (which now forms part of the floor of
the cave), has worn and well-spaced incisions on its upper face as well as
sharply incised and closer-together markings; it too is a smooth and rounded
boulder.
Hole Number 2.
About halfway down the gentle slope from the ‘directional stele’ to the sudden
descent to the ‘squeeze,’ against the northeastern wall, a hole enters between
the rocks that comprise the rubble floor (see Figure 2). Down this hole about four meters, a
chamber under the rubble floor of the Upper Chamber opens up, about five meters
long and two meters wide, and in most places high enough for a person to stand
up in. It has a flat rubble floor and one side is the smooth, non-incised wall
of the cave. The inside ceiling is also smooth and not incised and was probably
a piece of the cave’s ceiling. This hole may, therefore, not be the one that
Maynard and Edwards describe in what they call section NE5 of the Upper Chamber’s wall (see Maynard
and Edwards 1971:
67). Rather, it
appears to lie in their section NE4
of the cave wall, where they found ‘virtually no markings’ (Maynard and Edwards
1971: 67).
Incised boulders do exist here, however. A small gap opens at the southeastern end of the hole
and wall and smooth rounded boulders supporting
the rubble of the rockfall on their tops can be seen some four meters further
down. Some of this rubble was cleared away and Chadwick descended down to the
rounded boulders. Inside this lower hole sits a large incised boulder, the top
of which the rockfall has smashed (after Chadwick). It was incised before the
rockfall because some of the lines would have continued under it and because
the knocked off pieces lying beside the boulder also carry incisions. These
were the only line markings found down this hole. Four large pieces of carbon
sit on a sloping shelf adjacent to the incised boulder. All the rubble here
appears loose and unsintered. The dust is rosa
in color, whereas the dust on the floor of the cave outside the hole is white.
Hole Number 3.
More holes were investigated under the rockfall to the northwest of the
‘cavern,’ around the ‘fireplace’ described above. Only one of these holes was
suitable to photograph, ‘hole number 3.’
Rubble had to be removed from the entrance to the hole’s chamber that contains
a rounded and incised boulder. Wood and charcoal also lie inside the hole, the
charcoal in rosa dust, with white dust
covering both (after Chadwick).
Other places in the Upper Chamber show the rounded and
smooth boulder floor under the angular rockfall. This includes an incised
rounded boulder about 30
meters past the ‘directional stele’ toward the ‘squeeze,’ near the northeastern
wall, and which protrudes out of the rubble (see Figure 2).
The hypothesis being proposed is that, at one stage, the
Upper Chamber of Koonalda Cave consisted wholly of smooth rounded boulders
(Rockfall C) that animals, humans,
or both incised with lines and in conjunction with which people performed
activities. Various rock collapses (Rockfalls A, B, D, and E) fell onto this
floor at different times, some of which prehistoric Aborigines also used (for
instance, Rockfall D). The alternative explanation is that the line makers came
across the smooth rounded boulders underneath the angular rockfall and incised
them, sometimes sealing up the chambers or covering over the rocks on
completing the marks. Such a proposition would, if amplified along these lines,
become impossible to disprove, because insistence that Aborigines or animals created the state in which the incisions now
exist would counter any evidence for the thesis presented here. To argue this
way wears very thin, however. In addition, Maynard and Edwards (1971: 68) also invoke the suggestion that
collapse occurred after Aborigines marked the walls. Another paper addresses
whether the human wall markings relate to the boulder incisions (Sharpe 2000).
Rockfall F is the original floor of smooth and rounded
boulders – perhaps part of Rockfall C – after Rockfall E ceases (see Figure 2) and where the ceiling of the cave
steps down. It starts at the slope down to the ‘squeeze’ and goes through the
‘squeeze’ itself. Its start, however, is unclear because Rockfall E has rolled
down the slope and mixed with the smooth rounded boulders.
Rockfall F forms the most famous part of Koonalda Cave for
in it are the masses of wall markings, including the impressive finger flutings
and the equally impressive wall engravings above the entrance to the ‘squeeze.’
Incised boulders lie here too. One example sits on the slope
down to the ‘squeeze’ and looks like a bird (see Figure 2 and Plate 6). It ‘perches’ in a gap going in under the fall,
the original ceiling of the cave roofing the gap. It may have dislodged and
fallen to its present position, or may reside more or less where it sat before
the rockfall. A stream of lines incises its right face.
PLATE 6. The ‘bird’-shaped smooth and rounded
boulder that is incised and partially under the rock fall on the slope down to
the ‘squeeze.’ Scale in centimeters.
A smooth rounded and incised boulder sits in the flattish
part of Rockfall F, in the midst of the finger flutings (see Figure 2). No fallen rubble occurs with it,
so it probably lies in its original position. This rock is rather soft and a
vivid white in color, and its line markings incise its wall side (the
southwestern wall of the cave). The finger flutings on the walls and the incisions
on the walls and boulders may have been contemporary and related acts.
Other boulders in Rockfall F also show incisions, but no
thorough investigation of them has been made.
Gallus excavated in this area and discovered mining trenches
(which, he suggests, a set of wall engravings point to), dating to around 20,000
years B.P. (Gallus 1971:
127-28 summarizes his results).
The following table summarizes the stratigraphic findings
(below the surface) from the five small excavations so far reported from the
Upper Chamber:
The first two excavations (E4-b and E4-I)
lie close to each other and about midway between the second pair (G5-I Smith and G5-I Chadwick) and the last
(‘fireplace’). The first pair and the last are in Rockfall C. The second pair
may also lie in Rockfall C because floor G5-I is partly in it and partly in Rockfall D. The
stratigraphies roughly parallel each other apart from the ‘fireplace’; however,
it is a disturbed area and next to a wall, which means that more dust may have
settled in it.
Two courses might help establish a unified stratigraphic
sequence for the whole of the Upper Chamber but, with so little deposit
anywhere, stratigraphies are hard to establish, especially after people have
walked over the thin dust layer.
The first possible course would compare the ‘twiggy layer’
deposit (comprising mainly twigs and bat droppings) with another deposit
comprising thicker pieces of wood and charcoal (see Symon 1971: 18-20).
Perhaps the first deposit originates from the surface of the Nullarbor
Plain and the second from humans. This suggests that water washed
in the twigs. But how could water flow account for this (this point is further
discussed below)? Another hypothesis suggests that the ‘twiggy layer’
represents the remains of torches, not made of thick roots or branches, but of
twigs lashed together and perhaps soaked in an animal
fat (a testable proposition). Aborigines could have carried bundles of twigs to
the Upper Chamber and made torches up there, leaving broken and unburned twigs
scattered around. Or, pieces of twigs could fall from torches carried up to the
Upper Chamber and burned. A problem with the torch origin of the twigs raises
itself, though: the ‘twiggy layer’ appears mostly in inaccessible places and
under rocks. One way to answer this suggests that the rocks fell onto the
twigs, that humans placed them there, or that the twigs washed under the rocks.
More likely, the twiggy layer existed over most of the floor and dust covered
it before moderns trampled the twigs to powder. It would be hard then to
separate the two deposits: the ‘twiggy layer’ and that of torches of wood and
charcoal. It is also hard to separate them because the torches also appear
under rocks and lie with the twigs, and because some of the twigs have charred
ends; in other words, they may both have a human origin.
A second potential way to create a relative stratigraphy
would consider the color and texture of the dust. A coarse and white, sometimes
gray dust forms the surface layer over many parts of the chamber. Rosa and red
dust deposits also exist, the latter especially in the area of the smooth
rounded boulders. Do the different dusts relate to any particular time or
climactic period (Gallus 1971:
92)? The floor
around the smooth rounded boulders appears older than the rest of the floor of
the Upper Chamber, and yet the ceiling above it rises higher than any other
part of the chamber, especially a portion of it between what look like joint
lines. The portion of the ceiling between the joint lines is red in color,
perhaps a clay infill between beds of limestone (Frank 1971b: 41 suggests this for the origin of clay deposits in
the Gallus Site). This may account for the color of the dust over the area of
smooth rounded boulders. That limestone forms the ceiling for the rest of the
Upper Chamber (see Figure 2)
could explain the dust color there too. (The redder ceiling covers floors E4-I and G5-I, but not the ‘fireplace.’)
Stratigraphic sequences will probably vary from place to
place in the Upper Chamber, then, depending on the age of the surface, the
deposit that forms the ceiling of the cave above it, its degree of human use,
the proximity of ochre deposits, and so on. To establish a universal
stratigraphic sequence for the Upper Chamber may prove too difficult.
The floor of the Upper Chamber appears to result from a
sequence rockfalls distinguished by their different degrees of weathering.
Rockfalls C and F appear the oldest, the most smoothed and rounded and in
places under other rockfalls, followed by A, D, and then, most recently, B and
E. Lines mark the rocks in Rockfalls C and F and they show evidence of human
use. Subsequent rockfalls have obscured much of this rockfall and its human
evidence. The origin – animal or
human – of the line markings remains unclear (Bednarik 1995; Flood 1997; Sharpe 2000).
A reasonable explanation for the rounding and smoothing
process is needed to help judge the appropriateness of distinguishing between
rockfalls on the basis of their different degrees of weathering. Further, the
process may require conditions, such as the running of water, that affect the
underlying line markings. It also affects the explanation for the origin of the
masses of small twigs on the floor
of the cave – from natural sources (in water flowing from the surface, for
instance), or from human sources (twigs brought in to use as torches).
The floor of the cave comprises boulders mostly fallen from
the roof of the cave. The roof had weathered to become smooth and so, on
falling from the roof, the majority of rocks would have a jagged face upward
and a smooth face downward. How do the boulders then naturally become smooth
and rounded on their upper faces?
Any of a number of mechanisms may explain this process. The
often-quoted cave weathering mechanisms and those considered especially
pertinent include:
1. By
erosion from the claws or guano of bats.
2. By
the action of water running over the rocks, as in a stream.
3. By
the action of water dripping onto the rocks.
4. By
the rocks lying in a pool or a lake and dissolving.
5. By
the action of air rapidly blowing over them.
6. By
insolation weathering, the rocks expanding and contracting with temperature
changes.
7. By
salt weathering, crystals forming in the surfaces pores of the rock and forcing
off particles of limestone.
The Upper Chamber contains insufficient bat guano for the
first proposed mechanism to apply (Hooper 1958; Jennings 1971: 38;
King-Webster and Kenny 1958;
Ollier 1969: 47).
The process of water action (2) is sometimes cited as the smoothing and rounding
mechanism for the Upper Chamber boulders (Jennings 1963: 54). It might at first seem the most reasonable and
predominant form of erosion. After all, the appearance of the rocks resembles
those found in a river, a smooth and hard outer crust covers them (Jennings 1971: 40-41),
and smoothed tubes more than ten centimeters in diameter run through them.
Several factors counter this suggestion. The undersides of
the boulders are not necessarily smooth and rounded, and are often heavily
crystalline. Water flow would have worn all surfaces. The major difficulty with
the process of water action arises from the direction of water flow. In the
Upper Chamber, water must flow uphill. Jennings
(1963: 54) notes that the flow marks in
the squeeze area of the cave suggest a southward flow. The water would have
flowed through the squeeze and then over the boulders. However, the rounded
boulders – which a ‘rapid and turbulent flow’ appears to have rounded – lie
somewhat higher than the squeeze. Jennings
explains this by positing an ‘uphill flow under hydrostatic pressure.’ He later
considers (Jennings 1978),
though, ‘that some of the larger rounded boulders are so large as to make the
currents required to round them too great to be likely in the cave.’
A water-worn proponent might suggest the existence of a
similar cave just above the present Upper Chamber. A rapid flow of water in
this cave smoothed the boulders, which then collapsed into the present chamber.
The present smooth and rounded boulders thus derive from the floor of the
higher chamber. The shape of the Upper Chamber’s ceiling and the even greater
flow problem this hypothesis implies, however, negate it as a suitable
explanation. Additionally, the lack of evidence for the remains of an opening
into such a chamber and the need to explain why the smooth surfaces of the
boulders sit uppermost – when we might expect a more random distribution –
further abrogate this suggestion.
Beneath Rockfall A sit incised smooth and rounded boulders,
and any water-dissolving process acting on the rocks above would remove the
markings. The process cannot explain the smoothing and rounding of non-incised
boulders, other than for the oldest rockfalls.
What about the two features of the rocks that suggest the
water-warn hypothesis? The smoothed tubes more than ten centimeters in diameter
and that run through the rocks may have formed when the boulders were part of
the roof during the phreatic preparation phase in the evolution of the cave (Jennings 1978). Second, the smooth and hard
outer crust that covers the rocks may have formed through a process such as
oxidization or the formation of surface salt crystals.
To cite water flow as one of the mechanisms for the
smoothing and rounding of the boulders is, therefore, not as simple as might
first appear.
The third possible eroding force – water dripping on to the
boulders – is irrelevant because of the absence of carbonate speleothems in the
Upper Chamber (Frank 1971b:
32; Jennings 1967: 26; Lawler 1953: 339-40,
345; Lowry and
Jennings 1974: 72-73). Moreover, a drip would round a boulder only by
deposition; solutional erosion by a drip would only sculpt channels (Jennings 1978).
The process of dissolving can also be discounted (Jennings 1967: 24; Lowry and Jennings 1974: 69). Quite angular and rough boulders sit
in Nullarbor Cave lakes and show no signs of rounding
because the water is evidently too saturated. (Limestone does dissolve,
however, in exceptional unsaturated portions of some lakes in Nullarbor caves.)
Dissolving as the boulder weathering process also encounters the two problems
raised above: those of water flow and the rounding of boulders over line
markings.
Wind action as the weathering process for the cave boulders
also proves unsatisfactory. Anderson
(1964: 129) and Wigley, Wood, and Smith (1966) suggest air movement as the
reason for tafoni and other erosional forms in the Mullamullang Cave.
Jennings (1967: 25) thinks this does not offer an adequate
explanation. The erosion lacks, he writes, ‘the streamlined forms which surface
wind erosion produces,’ and the wind velocities are not ‘high enough over
adequate duration to be responsible for cave features.’ Any streamlined
features, he further suggests (1978),
‘would be aligned in a common direction unless disturbed since fashioned by the
wind.’ No aerofoil-shaped rocks appear in the Upper Chamber of Koonalda Cave
and no air currents of any great magnitude pass through it. (However, some of
the caves and blowholes of the Nullarbor do have large air movements through
them (Anderson 1964:
128, 130-31; Lawler 1953: 342-43;
Wigley 1967;
Wigley and Brown 1976:
333-34; Wigley and Wood 1967: 32-33;
Wigley, Wood, and Smith 1966).
Further, the breathing phenomena (Flood 1997) reported for these chambers suggests an air
flow reversal within them, making any wind-eroded rocks more round in shape
than aerofoil in one direction.)
Neither do the temperature changes recorded in the Nullarbor
caves seem adequate to cause insolation weathering of the boulders (Lowry 1968: Plate 22).
This leaves process (7),
that of salt weathering (or exudation, crystal wedging, salzsprengung,
salt frittering, flaking, or granulation). Many authors refer to it as the
cause of small-scale weathering in
Nullarbor caves and blowholes (Frank 1971a:
101, 1971b: 41; Hunt 1970: 17-18; Jennings 1967: 25, 1968:
50, 1971: 38; Lowry 1964: 16,
1968: 42, 1970: 31; Lowry and Jennings 1974: 59, 71;
Wigley and Hill 1966)
and some consider it an important geomorphological process in many environments
(Goudie 1986;
Mustoe 1982). A
well-recognized spatial association exists between weathering and the presence
of soluble salts (Young 1987:
962).
The salt weathering classification contains three potential
mechanisms (Cooke and Smalley 1968):
1. The
heating of salts within confined spaces in rock surfaces can exert pressure on
the rock and cause flaking and granulation. Even with a relatively limited
temperature range, salt weathering can produce rapid splitting and granular
disintegration of rock (Goudie 1986).
2. The
expansion of anhydrated salts passing to a hydrated state within confined
spaces in a rock surface can cause sufficient stress to force off particles.
Alternatively, salts passing from one hydrated state to a higher one under
temperature and humidity changes can cause a similar effect (Goudie and
Wilkinson 1977: 20). Nocita (1987) suggests that this process mostly
operates in hot, arid environments.
3. Salt
solutions within a few millimeters of the surface of rocks can, on evaporation,
precipitate salts that ‘wedge off the surface grains of limestone’ (Lowry 1964: 16). This process can wedge off surface
grains, flake and scale off various sized rock fragments, as well as split
rocks (Goudie 1986:
284; Jutson 1934: 255; Mustoe 1982: 111; Thornbury 1954: 38-39).
Experiments by Goudie in
1974 show that
the salt heating process (1)
is relatively ineffective. Further experiments published in 1986 show that temperature plays
a significant role, however, but the temperatures used simulated those on
desert surfaces rather than cave interiors. Koonalda Cave
does not have the necessary temperature fluctuations to promote this
granulation process (Cooke and Warren 1973:
66-67; Young 1987: 963).
Evans (1969-70: 154-55)
reviewed salt weathering as a cause of weathering (see also Jutson 1918 and 1934: 254-56
for an early Australian use of this explanation; and Twidale 1968: 140-42 in relation to Jutson’s use). In fact, it is
often cited as the cause of decay in building stone (reviewed again in Evans 1969-70: 156-57; see also Goudie 1986; Reed 1947; Schaffer 1967; Winkler and Wilhelm 1970). In-depth interest and
experimentation have looked at the mechanisms particularly promoting salt
weathering (Buckley 1951:
468-79; Cooke and Warren 1973: 66-71;
Evans 1969-70; Goudie 1974,
1986; Goudie,
Cooke, and Evans 1970; Mabbutt 1977: 27-29).
They show that salt weathering mechanisms (2) and (3)
would both work effectively in the limestones of Koonalda Cave
(Goudie 1974;
Goudie, Cooke, and Evans 1970: 45). Further, Koonalda contains more than
ample deposits of limestone dust covering the boulders and the floor; this may
include ‘rock meal’ or ‘rock flour,’ a diagnostic indicator of salt weathering
(Goudie 1986;
Higgins 1990: 296; Pohl and White 1965: 1464; Wellman and Wilson 1965: 1098). In Goudie’s experiments (1986) using Lower Carboniferous
siliceous sandstone, the process also produced substantial quantities of fine
sediment. Another important sign, crystallization, appears in the Upper Chamber
both under the rocks and in the scaling and slicing off sections of the wall
(Frank 1971a: 96, 101, 1971b: 32,
41; Lowry 1967; Maynard and Edwards 1971: 64; Wigley and Hill 1966: 38). Jennings
(1967: 25, 1971: 38) accepts salt weathering as an explanation of
some cave formations (see also White 1976:
308-9). He (1978) agrees that it probably caused the
frittering of the roof in Mullamullang
Cave to produce the
‘Dune,’ a ten-meter-high pile of dust.
Salt weathering, in fact, seems the most likely cause of the
smoothing and rounding of the rockfalls in the Upper Chamber of Koonalda Cave.
Several significant points need answering, however:
1. Several
experts tentatively discuss the conditions necessary for salt weathering
(Higgins 1990: 296; Lowry 1964: 16; Lowry and Jennings 1974: 59, 71).
Others then readily accept salt weathering as the explanation of breakdown in
Nullarbor caves (Hunt 1970:
17-18). A more cautious approach
should be adopted. Lowry (1968:
42-43) cites the conditions necessary
for salt weathering to take place – in particular, for salt weathering to cause
the dome development in the shallow Nullarbor caves – as at least three: ‘the
rock must be sufficiently porous; the interstitial fluid saline; and . . . the
air is not saturated with water.’ These conditions probably occur in these
caves. Whether they constitute the necessary conditions for salt weathering to
occur requires further research. Whether they are satisfied for the boulders in
Koonalda remains unknown.
2. Ollier
(1969: 12-14) notes a common criticism of the salt
precipitating process: ‘It is rather hard to explain why crystal growth should
continue against the pressure of the confining rock.’ Evans
(1969-70: 167-73) cites experimental evidence for the process
under certain conditions: high super-saturation and rapid evaporation (Rice 1977: 120). Thus, many authors suggest it as the
mechanism explaining weathering in arid areas such as deserts (Nocita 1987) and Antarctica,
and where wetting and drying is common – as in coastal regions (Cooke and
Warren 1973: 68; Evans
1969-70: 159-64; Mustoe 1982;
Ollier 1974: 20; Young 1987). Writes Mabbutt (1977: 27-28):
‘Super-saturation is favored under desert conditions where surface heating and
drying winds cause excessive evaporation, and crystallization forces are
therefore likely to reinforce thermal
stresses. The forces may be cumulative with repeated solution and
re-crystallization under high super-saturation.’ In Koonalda, even if a salt or
salts highly super-saturate the moisture in the boulders, is the cave
atmosphere conducive to rapid evaporation? Perhaps the cave’s past
speleoclimates need taking into account.
3. The
conditions necessary for the hydration of the salts offer another challenge
(Cooke and Warren 1973:
67-68; Higgins 1990; Mabbutt 1977: 28-29).
Crystallization in the anhydrous form usually occurs under high temperatures
and low humidity, and moisture is later absorbed, particularly after a wet
period. The process repeats itself. Drops in humidity cause dehydration and anhydrous
salts fill the resultant spaces. Re-hydration follows this, and the process
continues. This alternative wetting and drying creates the most effective
disruption to the surface of the rock. Do these conditions occur in Koonalda Cave?
4. The
origin of moisture presents another problem when looking at salt weathering as
the smoothing and rounding mechanism. Salt solutions supposedly evaporate close
to the surface of the boulders and deposit crystals whose growth or hydration
exert pressure and force off surface grains of limestone. But, where does the
salt solution come from? The situation differs from the disintegration of
building stones by salt weathering. Here, rain falls onto the stone wall and
percolates down through the stones to accumulate in the lower portion of the
wall. Most of the damage occurs here. The situation also differs from salt
weathering on the walls and roof of the cave. Here, water percolates down from
the surface of the plain and, when it reaches the cave walls, it evaporates.
Two potential sources may provide the moisture in the
boulders: the atmosphere and dampness in the floor.
- The
Atmosphere. If the atmosphere in Koonalda Cave usually remains
constant humidity-wise and temperature-wise, as it seems to, the rock may
not absorb moisture that then evaporates back into the atmosphere.
Seasonal changes may alter the inside climate, however; the temperature in
the cave, for instance, may vary from one season to another. Rain outside
the cave may cause large changes in humidity inside.
- Dampness in the
Floor. It makes more sense, though, for the moisture to rise in the
boulders by capillary action from the floor of the cave, especially after
rain (Jutson 1934:
347; Winkler
and Wilhelm 1970:
568). ‘The
wick effect’ that Goudie (1986)
propounds applies here. Numerous field observations record the upward
migration of saline solutions into rock and the subsequent damage of the
rock. Buildings with foundations in the capillary fringe or saturated zone
(such as the ruins of Mohenjo-Daro
on the alluvial plain of the Indus)
suffer from this process. Boulders transported onto moist playa surfaces
from alluvial fans decay rapidly as salt migrates into them. Capillary
action can draw up salts from deep zones of saturation (Baker 1990: 237).
5.
If so, salt weathering would affect the lower portions
of the boulders more (and thus perhaps render them more cavernous) than the
upper portions. The upper portions are, on the other hand, exposed to a more
evaporation-inducing atmosphere. This would result in more smoothing and
rounding on the upper surfaces. Large crystals grow on the undersides of many
of the otherwise smooth and rounded boulders. The weathering of a rockfall of
several boulders depth would, in a similar way, attack the outer boulders
rather than those deeper down – but the lower boulders would be more subject to
migrating salts than those on top. The situation could be complex and warrants
a thorough investigation. Jennings
(1978) also
points out that weathering of the rocks in a pile will lead to some rotation
with lower surfaces becoming upper surfaces and therefore subject to
weathering. So, he suggests, a stratigraphy of rounding should exist (see
Figure 4).
FIGURE 4. The
expected ‘stratigraphy’ of boulders as a result of differential rounding of
variously aged rock falls (after Jennings), and the expected stages in the
smoothing and rounding of boulders by salt weathering.
Ollier (1969: 186), in his discussion of the flaking by salt
growth in granite, says, ‘Individual blocks that weather by flaking become
rounded because the process attacks corners and edges more than faces. When a
boulder is quite rounded it shrinks, and the radius of curvature becomes smaller.’ He notes also that flaking does not extend
below ground surface and that, when it does occur on concave surfaces, it tends
to exaggerate the curvature. Evans (1969-70: 152, see also 157-58)
writes: ‘Rounding by granular disintegration is the commonest effect of salt
crystallization.’
Salt weathering presumably weathers jagged surfaces to
smooth and rounded surfaces; that is, jagged parts wear down to the same level
as the valleys between them. This makes sense (but needs further research)
because a jagged portion has two faces to force off particles from, whereas a
hollow has less of a tendency to become deeper (see Mowat 1962). The process should,
therefore, develop in stages (see Figure 4):
·
Stage I represents a jagged, relatively recently
fallen rock.
·
It becomes smooth and humped (stage II) in the
process of salt weathering and the curvatures of both concave and convex
surfaces become exaggerated (Ollier and Tuddenham 1961: 264).
·
The humps more or less disappear in the third
stage as the weathering removes convexities at twice the rate of concavities,
and the boulder becomes smooth and rounded.
If this
scenario is correct, the boulders of Rockfalls B, D, and E collapsed the most
recently and are in stage I of weathering. The rocks of Rockfall A follow in
stage II, and those of Rockfalls C and F in the final stage III and constitute
the oldest collapse.
6.
Boulders become less and less susceptible to salt
weathering as they continue to experience it because their surface area
decreases and becomes a crust. This crust becomes more and more crystalline,
harder, and impervious to the flow of moisture (Lowry 1970: 31; Mabbutt 1977: 30;
Ollier 1969: 80). This process also needs
further investigation. Evans (1969-70: 157) describes skins of crust peeling away from
limestone; portions of skins with line markings are peeling off boulders in the
Upper Chamber. The formation of the crust means that lines incised on rocks in
stage III of weathering will last longer than lines on rocks at a lower stage.
This may explain why few line markings appear in Rockfall A and only faint ones
in Rockfall D.
The ceilings and walls of Nullarbor caves show the results
of salt weathering, including hollows and tafoni (the convex and rounded
projections left between the hollows) (Howard and Kochel 1988: 28; Jutson 1934:
255; Lowry 1970: 31; Lowry and Jennings 1974: 71). The suggestion that salt weathering
causes smoothing and rounding of boulders, then, extends beyond what scholars
and other observers have seen as its work in Nullarbor caves. As the way for
the rounding and smoothing of the Upper Chamber boulders, salt weathering
offers a promising geomorphological hypothesis; many details need clarification
and verification.
This clarification is necessary from an archaeological point
of view as well. Not only would it help judge the appropriateness of
distinguishing between rockfalls from their different degrees of weathering,
but it may – as mentioned above – also
require conditions that affect line markings on rocks below the current floor
of the cave, and affect the understanding of other potential human artifacts.
The archaeological remains and wall markings in the Upper
Chamber of Koonalda Cave on the South Australian portion of the Nullarbor Plain
have been known for many years. Further line markings on the floor boulders of
this chamber were discovered by 1973,
together with torches, wood, and bone remains, possibly dating to at least 19,000
years B.P. This paper reports some of the findings of a second investigation of
the Upper Chamber in 1976.
The floor of the Upper Chamber comprises at least five
rockfalls of different ages and degrees of weathering (which renders the rocks
smooth and rounded). This part of the cave, its deposits and features, is
described in detail. Incised lines appear on rocks of the oldest collapse,
Rockfalls C and F, with few or no markings appearing on rocks in the other
collapses. Individual marked rocks and a number of apparently cleared areas
exist within Rockfall C, some areas obviously intentionally constructed with
marked boulders flanking them. The rocks in Rockfall A are fairly smooth and
rounded. Those in Rockfall D are more rough and jagged but have been used
significantly by humans in that a series of apparently cleared areas occur
here, together with bone deposits and a flint flake, perhaps an engraving tool.
The most recent rockfalls are B and E. No evidence of human use exists in them,
but underneath them and the others lays the original smooth, rounded, and
incised boulder floor. A few small
excavations have also been carried out and define comparable stratigraphic
sequences.
Further study of the rockfall sequence in relation to line
marking and mining activities might prove valuable, including as a means to
date the activities.
What mechanism smoothes and rounds the rockfalls? This process,
whatever it is, operates on rocks on top of already incised smooth and rounded
boulders. This counts against water flow as the process, despite what many
writers suppose. After examining this and a number of other mechanisms, this
paper favors salt weathering or salt weathering, in which crystals form in the
surface pores of the limestone and force off particles. Several details need
confirming with this as the weathering mechanism before proposing it with
certainty.
Two further words of caution are necessary. The first
results from the fact that the deposit lies mostly on the surface (if not under
massive rock falls) and so easily suffers damage, and has been recently damaged
by people who did not know what is there and have walked over artifacts, crushing
bones, torches, and erasing engravings, even poking holes into engraved walls
and rocks. The crushing of artifacts under foot happens easily because often a
thin layer of dust covers them and walkers find it hard to see in the dark.
Only authorized people should now enter the cave.
The second matter apologizes to those who seek a complete
archaeological history of the Upper Chamber. This paper reports from the
perspective just described on work so far carried out, and does not thoroughly
document these 2000
square meters of human use. Careful classification and analysis remains to be
carried out on the finds as well. The understanding of the Upper Chamber
continues to increase and investigators should acquire as full a picture as
possible before further irreversible archaeological activity begins.
Many people were involved in the 1976 expedition to Koonalda and the
preparation of this paper:
·
For financial support,
encouragement and advice: the National Geographic Society (especially Mary
Griswold Smith and Victor Boswell).
·
For supplying gear, sponsorship, and other support: the South Australian
Museum (especially Graeme
Pretty).
·
For permission to enter Koonalda Cave:
the South Australian Aboriginal and Historical Relics Advisory Board.
·
For accompanying Sharpe on this visit: Neil
Chadwick, Sandor Gallus (nominated by the South Australian
Museum as the field
investigator), Ian Lewis, Kevin Mott, and Christine Sharpe (now Christine
Whitehead).
·
For hospitality while at Koonalda: Mr. and Mrs.
Cyril Gurney.
·
For stimulus and support:
Alexander Marshack and Hallam Movius Jr.
·
For information and comments in the preparation
of the report: Robert Bednarik, A.
Cooper, Sandor Gallus (including his permission to use the notes of Neil
Chadwick, and those of 1975
by Smith, Cooper, and Ross on the Upper Chamber), Joe Jennings, Mary Lacombe, C. Merewether, Cliff Ollier, Graeme
Pretty, Betty Ross, and Mike Smith.
·
For assistance in the preparation of the
manuscript itself: Alf Armstrong, Sandra Myer, Helen Fawbert, and the
Universities of Auckland and Otago.
·
Most of all, the author expresses his great debt
to the late Sandor Gallus.
Further figures and plates that help elucidate this paper
can be found at http://www.ksharpe.com/www/articles/ articledetail.cfm?article_id=306.
Anderson, E. G.
1964. Nullarbor
expedition 1963-4. Helictite 2(4): 121-34.
Baker, V. R. 1990.
Spring sapping and valley network development. In C. G.
Higgins and D. R. Coates (eds.), Groundwater Geomorphology: The Role of
Subsurface Water in Earth Surface Processes and Landforms. Special Paper 252. Boulder, Colorado:
The Geological Society of America,
pp. 235-66.
Bednarik, R. G.
1991. On natural
cave markings. Helictite 29 (2): 27-41.
_________ 1995. The discrimination of rock
markings. Rock Art Research 11
(1): 23-44.
_________ 1998.
Microscopic analysis of ‘engraved plaques’ and other objects from Devil’s Lair.
Journal of the Royal Society of Western
Australia 81:
23-33.
Buckley, H. E.
1951. Crystal Growth.
New York:
John Wiley and Sons.
Cooke, R. U.
and I. J. Smalley
1968. Salt
weathering in deserts. Nature 220
(5173): 1226-27.
Cooke, R. U.
and A. Warren 1973. Geomorphology in Deserts.
London: B. T.
Batsford.
COX, H. 1960.
Australia’s
ice-age men. People (13
April): 17-19.
DORTCH. C. E. 1976.
Two engraved stone plaques of late Pleistocene age from Devil’s Lair, Western Australia. Archaeology
and Physical Anthropology in Oceania 11 (1): 34-44.
EDWARDS, R., and MAYNARD, L. 1967. Prehistoric art in Koonalda Cave. Royal Geographical Society of Australia,
South Australian Branch Proceedings 68:
11-17.
Evans, I. S. 1969-70. Salt crystallization and rock
weathering: a review. Revue Geomorphologie Dynamique 19 (4): 153-77.
Frank, R. 1971a. Cave sediments as palaeoenvironmental
indicators, and the sedimentary sequence in Koonalda Cave.
In D. J. Mulvaney and J. Golson (eds.), Aboriginal Man and Environment in Australia. Canberra: Australian National University
Press, pp. 94-104.
_________ 1971b. The sediments.
In R. V. S. Wright (ed.), The Archaeology of the Gallus Site, Koonalda Cave. Canberra: Institute of Australian
Studies, pp. 31-44.
GALLUS, A. 1964.
Two Australian stratigraphic sequences. Current Anthropology 5 (2): 127
_________ 1966.
Comment on Pradel, ‘Transition from Mousterian to Perigordian.’ Current
Anthropology 7 (1): 39-40.
_________ 1968a.
Archaeological excavations at Koonalda. Nullarbor Plain,
1957-1967. Anthropological Society
of South Australia
Journal 6 (7): 4-8.
_________ 1968b.
Excavations at Koonalda Cave,
South Australia. Current
Anthropology 9 (4): 324-25.
_________ 1968c
Parietal art in Koonalda
Cave, Nullarbor Plain, South Australia.
Helictite 6 (3): 43-49.
_________ 1968d.
Results of the 1968
Koonalda Cave (S. A.) expedition of the
Archaeological Society of Victoria. Artefact, No. 10: 1-4.
_________ 1971.
Results of the exploration of Koonalda Cave, 1956-1968.
In Archaeology of the Gallus Site, Koonalda Cave
(R. V. S. Wright, ed.), pp. 87-133. Australian Institute of Aboriginal
Studies, Canberra.
_________ 1977.
Schematization and symboling. In Form in Indigenous Art (P. J. Ucko,
ed.), pp. 370-86. Australian Institute of
Aboriginal Studies, Canberra.
Goudie, A. 1974. Further experimental
investigation of rock weathering by salt and other mechanical processes. Zeitschrift
für Geomorphologie Supplement 21
(3): 1-12.
_________ 1986. Laboratory simulation of ‘the wick
effect’ in salt weathering of rock. Earth Surface Processes and Landforms
11: 275-85.
_________, R. Cooke,
and I. Evans 1970. Experimental investigation of rock
weathering by salts. Area 2
(4): 42-8.
_________ and J. Wilkinson 1977. The Warm Desert
Environment. Cambridge:
Cambridge University Press.
Higgins,
C. G. 1990. Seepage-induced cliff recession and regional
denudation. In C. G. Higgins and D.
R. Coates (eds.), Groundwater Geomorphology: The Role of Subsurface Water in
Earth Surface Processes and Landforms. Special Paper 252. Boulder, Colorado:
The Geological Society of America,
pp. 291-318.
Hooper, J. H. D. 1958. Bat erosion as a factor in
cave formation. Nature 182
(4647): 1464.
Howard, A. D. and
R. C. Kochel 1988. Introduction to cuesta
landforms and sapping processes on the Colorado Plateau. In A. D. Howard, R. C.
Kochel and H. E. Holt (eds.), Sapping Features of the Colorado Plateau. National
Aeronautics and Space Administration Special Publication SP-491, pp. 6-56.
Hunt, G. S. 1970.
The origin and development of Mullamullang Cave N37, Nullarbor
Plain, Western Australia.
Helictite 8 (1): 3-26.
Jennings, J. N. 1958. The limestone geomorphology of the
Nullarbor Plains (Australia).
Actes Deuxième Congres Internationale Spéologie, Bari. 1: 371-86.
_________ 1963.
Some geomorphological problems of the Nullarbor Plain.
Royal Society of South Australia
Transactions 87:
41-62.
_________ 1967. The surface and
underground geomorphology. In J. R. Dunkley and T. M. L. Wigley (eds.), Caves
of the Nullarbor. Sydney:
Speleological Research Council, pp. 13-31.
_________ 1971. Karst. Canberra: Australian National
University.
_________ 1978. Pers. Comm.
Jutson, J. T. 1918. The influence of salts in
rock weathering in sub-arid Western
Australia. Royal Society of Victoria Proceedings 30 (2): 165-72.
_________ 1934. The
physiography (geomorphology) of Western
Australia. 2nd
edn. revised. Geological Survey of Western
Australia, no. 95.
King-Webster, W. A.
and J. S. Kenny 1958. Bat erosion as a factor in
cave formation. Nature 181
(4626): 1813.
Lawler, D. 1953. Wonderland under the
Nullarbor. Wildlife and Outdoors 18
(4): 336-45.
Lowry, D. C. 1964. The development of Cocklebiddy Cave, Eucla Basin, Western Australia.
Helictite 3 (1): 15-19.
_________ 1967.
Halite speleothems from the Nullarbor Plain, Western Australia. Helictite
6 (1): 14-20.
_________ 1968. The origin of blowholes
and the development of domes, by exudation in caves of the Nullarbor
Plain. Geological Survey of Western Australia Annual Report for 1967, pp. 40-44.
_________ 1970. Geology of the
Western Australian part of the Eucla
Basin. Western
Australian Geological Survey Bulletin, no. 122.
_________ and J. N.
Jennings 1974.
The Nullarbor karst, Australia.
Zeitschrift für Geomorphologie 18
(1): 35-81.
Mabbutt, J. A. 1977.
Desert Landforms. Canberra:
Australian National University.
MARSHACK, A. 1972a
Cognitive aspects of Upper Paleolithic engraving. Current Anthropology 13 (3-4):
445-77.
_________ 1972b.
The Roots of Civilization. McGraw
Hill, New York.
_________ 1972c.
Upper Paleolithic notation and symbol. Science 178 (4063): 817-28.
_________ 1975.
Exploring the mind of ice-age man. National Geographic 147 (1): 62-89.
_________ 1976a.
Some implications of the Paleolithic symbolic evidence for the origin of
language. Current Anthropology 17
(2): 274-82.
_________ 1976b.
Use versus style in the analysis and interpretation of Upper Paleolithic image
and symbol. Presented to the Ninth International Scientific Congress on
Prehistory and Protohistory, Nice.
_________ 1977.
The meander as a system: the analysis and recognition of iconographic units in
Upper Paleolithic compositions. In Form in Indigenous Art (P. J. Ucko,
ed.), pp. 286-317. Australian Institute of
Aboriginal Studies, Canberra.
_________ 1979.
Upper Paleolithic symbol systems of the Russian Plain: cognitive and
comparative analysis. Current Anthropology 20 (2):
271-311.
MAYNARD, L. 1974.
The archaeology of Australian Aboriginal art. Presented to the Symposium: The
Art of Oceania.
_________ and EDWARDS, R. 1971. Wall markings. In The Archaeology of the
Gallus Site, Koonalda
Cave (R. V. S.
Wright, ed.), pp. 61-80. Australian Institute of
Aboriginal Studies, Canberra.
Mowat, G. D. 1962.
Progressive changes of shapes by solution in the laboratory. Cave Notes 4 (6): 45-49.
Mustoe, G. E. 1982.
The origin of honeycomb weathering. Geological Society of America Bulletin
93: 108-115.
Nocita, B. W. 1987. Thermal expansion of salts as an effective physical
weathering process in hot desert environments. Geological Society of America
Abstracts with Programs 19:
790.
Ollier, C. D. 1969. Weathering. Edinburgh: Oliver and
Boyd.
_________ 1974. Weathering
and Landforms. Basingstoke,
England:
Macmillan Education.
_________ and W. G.
Tuddenham 1961. Inselbergs of Central
Australia. Zeitschrift für Geomorphologie 5 (4): 257-76.
Pohl, E. R.
and W. B. White 1965. Sulfate minerals: their
origin in the central Kentucky
karst. American Mineralogist 50
(9): 1461-65.
PRETTY, G. L. 1960. The CEGSA Nullarbor
expedition 1959/60, archaeological notes. Sydney University Speleological Society
Journal 6 (1): 2-7.
_________ and Gallus, S. A. 1967. The anthropology and archaeology of
the Nullarbor Plain. In Caves of the
Nullarbor (J. R. Dunkley and T. M. L.Wigley, eds.), pp. 47-49. Speleological Research Council, Sydney.
Reed, J. J. 1947. Deterioration of stone
(Iguimbrite) on a Wellington
building. New Zealand
Journal of Science and Technology 28(B)
(4): 249-52.
Rice, R. J. 1977. Foundations of
Geomorphology. London:
Longman.
Schaffer, R.
J. 1967. Causes
of the deterioration of building materials. 1. Chemical and physical causes. Chemistry and
Industry, No. 38:
1584-6.
SHARPE, C. E. 1977.
Koonalda Cave in the beginning of artistic
expression. New Quarterly
Cave 2 (3): 226-34.
_________ and SHARPE, K. J. 1976. A preliminary report of engraved
boulders in the Art Sanctuary of Koonalda
Cave, South Australia.
Mankind 10 (3): 125-32.
SHARPE, k. j. 1982. A Working Paper on
the Meanders in Koonalda Cave,
South Australia. Preprint.
Http://www.ksharpe.com/www/articles/articledetail.cfm?article_id=215.
_________ 2000.
Incised Linear Markings: Animal or
Human Origin? Preprint.
Http://www.ksharpe.com/www/articles/articledetail.cfm?article_id=120.
_________ and SHARPE, C. E. 1980. The Upper Chamber of Koonalda Cave, South
Australia: A Second Report. Preprint. Http://www.ksharpe.com/www/articles/articledetail.cfm?article_id=300.
SYMON, D. E. 1971.
The vegetation surrounding Koonalda
Cave. In The
Archaeology of the Gallus Site, Koonalda
Cave (R. V. S.
Wright, ed.), pp. 17-20. Australian Institute of
Aboriginal Studies, Canberra.
THOMPSON, J. M. 1949.
The Nullarbor caves system. Royal Society of South Australia Transactions 73 (1): 48-51.
Thornbury, W.
D. 1954. Principles
of Geomorphology. New York:
John Wiley and Sons.
Twidale, C. R. 1968. Geomorphology: With Special Reference to Australia. Melbourne: Australia.
Wellman, H. W.
and A. T. Wilson 1965. Salt weathering, a
neglected geological erosive agent in coastal and arid environments. Nature
205 (4976): 1097-98.
WELLS, L. A., and HUNT, G.
W. 1919. Nullarbor Plain. Royal Geographical Society of Australia,
South Australian Branch Proceedings 19:
137-40.
White, W. B. 1976. Cave minerals and
speleothems. In T. D. Ford and C. H. D. Cullingford (eds.), The Science of
Speleology. London:
Academic Press, pp. 267-327.
WIGLEY, T. M. L. 1966.
An anthropological note. In Mullamullang
Cave Expeditions, 1966 (A. L. Hill, ed), p. 33. Cave Exploration Group (South Australia), Adelaide.
_________ 1967.
Non-steady flow through a porous medium and cave breathing. Journal of
Geophysical Research 72
(12): 3199-205.
_________ and M. C. Brown
1976. The
physics of caves. In J. D. Ford and C. H. D. Cullingford, The Science of
Speleology. London:
Academic Press, pp. 329-58.
_________ and A. L. Hill
1966. Cave
decoration. In A. L. Hill (ed.), Mullamullang
Cave Expeditions, 1966. Adelaide: Cave Exploration Group (South Australia), pp. 37-39.
_________ and I.
D. Wood 1967. Meterology of the Nullarbor
Plain caves. In J. R. Dunkley and T. M. L. Wigley (eds.), Caves
of the Nullarbor. Sydney:
Speleological Research Council, pp. 32-34.
_________, I. D. Wood, and M. Smith 1966.
Meteorological aspects. In A. L. Hill (ed.), Mullamullang Cave
Expeditions, 1966.
Adelaide: Cave
Exploration Group (South Australia),
pp. 40-46.
Winkler, E. M.
and E. J. Wilhelm 1970. Salt burst by hydration
pressures in architectural stone in urban atmosphere. Geological Society of America
Bulletin 81 (2): 567-72.
WRIGHT, R. V. S. (ed.), 1971a. Archaeology of the Gallus Site, Koonalda Cave. Australian Institute of
Aboriginal Studies, Canberra.
_________ 1971b.
The archaeology of Koonalda
Cave. In Aboriginal
Man and Environment in Australia (D. J. Mulvaney and J. Golson, eds.), pp.105-13. Australian National
University Press, Canberra.
_________ 1971c.
The cave. In Archaeology of the Gallus Site, Koonalda Cave
(R. V. S. Wright, ed.), pp. 22-28. Australian Institute of
Aboriginal Studies, Canberra.
_________ 1971d.
The flints. In Archaeology of the Gallus Site, Koonalda Cave
(R. V. S. Wright, ed.), pp. 48-58. Australian Institute of
Aboriginal Studies, Canberra.
Young, A. R.
M. 1987. Salt as
an agent in the development of cavernous weathering. Geology 15: 962-66.
A file on this site, AR66.htm, contains referees’ comments
on the original version of this paper, many of which may still apply the paper
above.
