A LITTLE FROM COLUMN A, A LITTLE FROM COLUMN B:
MERGING HIGH-RESOLUTION RESOURCE MODELLING
WITH ARCHAEOLOGICAL DATA TO CONSTRUCT AN
INDIGENOUS OCCUPATION HISTORY FOR INLAND
SOUTHWESTERN AUSTRALIA.
Alana M. Rossi
BA (Hons), MA.
Submitted in fulfillment of the requirements for the Doctor of Philosophy
School of Arts and Sciences
Fremantle
November, 2020
Declaration
To the best of the candidate’s knowledge, this thesis contains no material previously
published by another person, except where due acknowledgement has been made.
This thesis is the candidate’s own work and contains no material which has been
accepted for the award of any other degree or diploma in any institution.
Signature:
Name:
Alana M. Rossi
Date:
1 November 2020
ii
ABSTRACT
In Noongar country, southwestern Australia, most archaeological research has
focussed on the coastal margin. Only one inland site has been excavated and dated
– Mulka's Cave – but interpretations were hampered by the lack of comparative data
(Rossi 2010, 2014a). This research aims to address the knowledge gap for the
inland southwest by creating an occupation history for the area within 50 km of
Mulka's Cave, by merging high-resolution resource modelling with archaeological
data.
The study area was divided into Landscape Divisions based on landform and
vegetation characteristics; monthly availability of food and water were quantified for
each, using published sources, agricultural tools (Department of Primary Industries
and Regional Development 2016a) and a specially devised model. Optimal Foraging
Theory was then used to define the optimal seasonal route for the study area
occupants. Archaeological investigations focussed on three sites: Anderson Rocks,
Gibb Rock and Mulka's Cave. Stone artefact attributes were interpreted using Kuhn's
(1995) provisioning systems, whereby different levels of mobility favour contrasting
technological strategies.
The archaeological data and occupation model identified a system of predictable and
unpredictable residential moves; predictability and duration were dictated by the
availability of water on-site and elsewhere in the landscape. Occupation was longer
and more predictable in the northern and central study area, where groups were
tethered to more reliable water sources during summer and autumn. More frequent,
unpredictable moves occurred in the southern study area, as winter rains allowed
people to disperse throughout the landscape.
The archaeological data demonstrate that the study area was visited as early as
9650 cal. BP, but regular occupation occurred from 7775 cal. BP. Considering the
reliance on ephemeral water sources, it is unlikely that the inland southwest could
have sustained regular occupation during the Pleistocene, when conditions were
drier. This study highlights the value in combining theoretical occupation models with
archaeological data to maximise interpretive value.
iii
ACKNOWLEDGEMENTS
First and foremost, I would like to thank my primary supervisor Dr Shane Burke for
always being approachable, encouraging, and for addressing my [very] lengthy
emails and chapter drafts in a timely fashion. I am also extremely grateful for the
financial assistance, which helped cover unanticipated field costs.
I would also like to thank my secondary supervisor, R. Esmee Webb for her
assistance, particularly during the fieldwork at Mulka's Cave during my MA research.
Her support and advice were especially appreciated when meeting with Indigenous
land councils and navigating the application processes associated with the
Aboriginal Heritage Act (1972).
The following individuals and institutions are thanked for their support or assistance
in particular areas:
o
The Australian Government, for providing fees offset under the Research
Training Program (RTP) scheme;
o
Graduate Women Western Australia, whose $5000 scholarship for
radiocarbon dating was vital to this research;
o
Fisher Rare and Special Collections, University of Sydney, for providing a
high-resolution scan of Curr's (1887) map;
o
South-West Aboriginal Land and Sea Council (SWALSC), particularly the
Ballardong Traditional Owners who accompanied me to Twine Reserve:
Clarence Barron, Donna Hill and Marques Ugle. The Ballardong's lawyer,
Peter Nettleton, is also thanked for facilitating meetings and providing
assistance and advice as required;
o
The Research Office, University of Notre Dame, Fremantle, for financial
and administrative assistance;
o
Staff from the School of Arts and Sciences and Student Administration and
Fees, University of Notre Dame, Fremantle for ongoing administrative
assistance and other support;
o
Professor Chris Clarkson, Professor Peter Hiscock and Professor Simon
Holdaway for examining this thesis, and providing valuable feedback.
iv
The penultimate thanks must go to my family and friends, for their support
throughout this incredibly long journey. I am particularly grateful for their efforts in
suppressing the "WHAT ARE YOU GOING TO DO AFTERWARDS?!?" question that
everyone wants to ask/yell but represses for fear it might initiate some sort of
breakdown. My immediate family were especially supportive during the last 6 weeks
prior to draft submission, when I felt like it was all becoming too much. Simple things
like a group chat with Simpsons references, terrible jokes, photos or updates from
home always came at a time that I really needed them and reminded me that there
was life outside the PhD bubble.
Finally, I want to thank my future wife Jess Godfrey for her moral support and
encouragement when I felt like giving up. She also read everything I wrote and
provided feedback from a different perspective to my supervisors, which was greatly
appreciated. I am particularly grateful for her patience when I was frustrated that, as
a geologist, she couldn't identify raw material samples without smashing them into a
million pieces; she also successfully hid her disappointment when I wouldn't let her
do so.
v
SUMMARY TABLE OF CONTENTS
Abstract .......................................................................................................................... iii
Acknowledgements ......................................................................................................... iv
Summary Table of Contents ............................................................................................ vi
Table of Contents .......................................................................................................... vii
List of Tables ................................................................................................................. xiii
List of Figures ................................................................................................................ xix
SECTION ONE: BACKGROUND TO THE RESEARCH
Chapter 1
Introduction: Addressing the Scarcity of Archaeological Data from
Inland Southwestern Australia .................................................................... 1
Theoretical Framework: Mobility, Subsistence and Technology ............... 16
Natural Environment and Landscape of the Study Area ........................... 36
Cultural Context ........................................................................................ 59
Chapter 2
Chapter 3
Chapter 4
SECTION TWO: RESOURCE AND OCCUPATION MODEL
Chapter 5
Chapter 6
Chapter 7
Chapter 8
Constructing the Resource Model ............................................................ 76
Plant and Animal Foods .......................................................................... 101
Water Availability .................................................................................... 125
Occupation Model for the Study Area ...................................................... 151
SECTION THREE: ARCHAEOLOGICAL INVESTIGATIONS
Chapter 9
Chapter 10
Chapter 11
Chapter 12
Chapter 13
Archaeological Methods ......................................................................... 173
Anderson Rocks ..................................................................................... 202
Gibb Rock ................................................................................................ 225
Mulka’s Cave ........................................................................................... 234
Identifying Technological Provisioning Systems ..................................... 275
SECTION FOUR: MERGING THEORY WITH ARCHAEOLOGY
Chapter 14 Mobility, Site Use and Occupation Intensity ........................................... 297
Chapter 15 Conclusion: Occupation History of the Study Area ................................. 313
References ................................................................................................................... 343
APPENDICES
Appendix A
Appendix B
Appendix C
Appendix D
Appendix E
Appendix F
Appendix G
Flora Species Present in the Study Area ............................................... 376
Distribution of Plant Foods in the Study Area ......................................... 399
Distribution of Animal Foods in the Study Area ...................................... 418
Measuring Artefact Attributes ................................................................. 425
Anderson Rocks Stone Artefact Assemblage ........................................ 431
Gibb Rock Stone Artefact Assemblage ................................................... 445
Mulka’s Cave Stone Artefact Assemblage ............................................. 450
vi
TABLE OF CONTENTS
Abstract .......................................................................................................................... iii
Acknowledgements ......................................................................................................... iv
Summary Table of Contents ............................................................................................ vi
Table of Contents .......................................................................................................... vii
List of Tables ................................................................................................................. xiii
List of Figures ................................................................................................................ xix
SECTION ONE: BACKGROUND TO THE RESEARCH
Chapter 1
Introduction: Addressing the Scarcity of Archaeological Data from
Inland Southwestern Australia ....................................................... 1
1.1 Introduction ............................................................................ 1
1.2 Archaeological Research Bias in the Southwest ................... 1
1.3 Study Area, Aims and Research Questions ........................... 5
1.4 Method of Addressing Research Questions .......................... 7
1.4.1 Resource and Occupation Model ............................... 7
1.4.2 Archaeological Investigations ..................................... 10
1.5 Significance of this Research ................................................. 12
1.6 Organisation of Thesis ........................................................... 14
Chapter 2
Theoretical Framework: Mobility, Subsistence and Technology .. 16
2.1 Introduction .......................................................................... 16
2.2 Forager Theory .................................................................... 16
2.3 Technological Organisation ................................................. 20
2.3.1 Mobility ..................................................................... 21
2.3.2 Raw Material Availability........................................... 23
2.3.3 Risk........................................................................... 26
2.3.4 Summary .................................................................. 28
2.4 Technological Provisioning Systems ................................... 29
2.4.1 Individual Provisioning ............................................. 30
2.4.2 Place Provisioning .................................................... 31
2.4.3 Summary .................................................................. 34
2.5 Conclusion ........................................................................... 35
Chapter 3
Natural Environment and landscape of the Study Area ............... 36
3.1 Introduction .......................................................................... 36
3.2 Geology and Landform ........................................................ 36
3.3 Hydrology ............................................................................. 40
3.4 Climate ................................................................................. 44
3.4.1 Palaeoclimate of Southwestern Australia ................. 46
3.4.2 Modern Climate of the Study Area ........................... 50
3.5 Vegetation ............................................................................ 52
3.5.1 Botanical Provinces of Western Australia ................ 52
3.5.2 Vegetation Communities in the Study Area .............. 54
3.6 Fauna .................................................................................. 57
3.7 Conclusion ........................................................................... 58
vii
Chapter 4
Cultural Context ........................................................................... 59
4.1 Introduction .......................................................................... 59
4.2 Archaeological Background ................................................. 59
4.2.1 Archaeological Sites in the Study Area .................... 60
4.2.2 Other Relevant Research ......................................... 65
4.3 Ethnohistorical Background ................................................. 70
4.3.1 Aboriginal Social Boundaries ................................... 71
4.3.2 Ethnohistory of the Study Area ................................ 73
4.4 Conclusion ........................................................................... 75
SECTION TWO: RESOURCE AND OCCUPATION MODEL
Chapter 5
Constructing the Resource Model ................................................ 76
5.1 Introduction .......................................................................... 76
5.2 Landscape Divisions ............................................................ 76
5.3 Quantifying Plant and Animal Foods .................................... 80
5.3.1 Compiling Floral and Faunal Lists ............................ 81
5.3.2 Identifying and Classifying Food Species ................. 82
5.3.3 Determining Seasonal Availability ............................ 83
5.3.4 Benefits and Limitations of the Approach ................. 85
5.4 Quantifying Water Availability .............................................. 87
5.4.1 Representing Rainfall Variability............................... 87
5.4.2 Soil Water ................................................................. 88
5.4.3 Rock Structures ........................................................ 94
5.5 Formulating the Occupation Model ...................................... 98
5.6 Conclusion ......................................................................... 100
Chapter 6
Plant and Animal Foods ............................................................. 101
6.1 Introduction ........................................................................ 101
6.2 Plant Food .......................................................................... 101
6.2.1 Gum ........................................................................ 101
6.2.2 Lerps/Manna........................................................... 106
6.3.3 Storage Organs ...................................................... 106
6.3.4 Fruit ........................................................................ 108
6.3.5 Flowers ................................................................... 109
6.3.6 Leaves/Roots/Stems .............................................. 110
6.3.7 Seeds ..................................................................... 111
6.3.8 Summary ................................................................ 111
6.3 Animal Foods ..................................................................... 114
6.3.1 Birds and their Eggs ............................................... 114
6.3.2 Reptiles and their Eggs .......................................... 119
6.3.3 Mammals ................................................................ 121
6.3.4 Amphibians ............................................................. 121
6.3.5 Summary ................................................................ 122
6.4 Conclusion ......................................................................... 124
viii
Chapter 7
Water Availability ........................................................................ 125
7.1 Introduction ........................................................................ 125
7.2 Soil Water .......................................................................... 125
7.2.1 Low Winter Rainfall – 2010, 2012........................... 125
7.2.2 Average Winter Rainfall – 2009, 2011 .................... 128
7.2.3 High Winter Rainfall – 1992, 1998 .......................... 131
7.2.4 Summary ................................................................ 134
7.3 Rock Structures.................................................................. 136
7.3.1 Low Winter Rainfall – 2010, 2012........................... 136
7.3.2 Average Winter Rainfall – 2009, 2011 .................... 138
7.3.3 High Winter Rainfall – 1992, 1998 .......................... 142
7.3.4 Summary ................................................................ 144
7.4 Summary of Water Availability in the Study Area .............. 146
7.5 Conclusion ......................................................................... 150
Chapter 8
Occupation Model for the Study Area ........................................ 151
8.1 Introduction ........................................................................ 151
8.2 Viability of Year-Round Occupation ................................... 151
8.3 Seasonal Route ................................................................. 154
8.3.1 Regular Seasonal Route ........................................ 154
8.3.2 Impact of Rainfall Variation .................................... 163
8.4 Inter-Group Aggregation .................................................... 164
8.5 Archaeological and Technological Signatures ................... 167
8.5.1 Northern and Central Study Area ........................... 168
8.5.2 Southern Study Area (Palaeovalley) ..................... 169
8.5 Conclusion ......................................................................... 171
SECTION THREE: ARCHAEOLOGICAL INVESTIGATIONS
Chapter 9
Archaeological Methods ............................................................. 173
9.1 Introduction ........................................................................ 173
9.2 Field Methods ................................................................... 173
9.2.1 Site Selection ......................................................... 173
9.2.2 Survey and Excavation Methods ............................ 176
9.3 Lithic Analysis .................................................................... 177
9.3.1 ‘Horrid Little Bits of Stone’ – the Problem with
Quartz Artefact Analysis ......................................... 178
9.3.2 Stone Artefact Classification................................... 180
9.3.3 Attribute Analysis .................................................... 184
9.4 Other Analysed Materials ................................................... 188
9.4.1 Charcoal ................................................................. 188
9.4.2 Sediments............................................................... 188
9.4.3 Ochre ...................................................................... 189
9.4.3 Glass ...................................................................... 189
9.5 Data Analysis ..................................................................... 192
9.5.1 Quantifying Site-Specific Water Sources................ 192
9.5.2 Constructing a Temporal Framework ..................... 194
9.5.3 Identifying Technological Strategies ....................... 195
9.5.4 Quantifying Occupation Intensity ............................ 199
9.6 Conclusion ......................................................................... 200
ix
Chapter 10 Anderson Rocks ......................................................................... 202
10.1 Introduction ....................................................................... 202
10.2 Site Description ................................................................ 202
10.2.1 Water Availability .................................................. 204
10.3 Survey and Excavation Methods ..................................... 205
10.4 Organics and Dating ........................................................ 209
10.5 Stone Artefact Assemblage ............................................. 211
10.5.1 Surface Collections ............................................... 212
10.5.2 AR1 ....................................................................... 214
10.5.3 AR2-3 ................................................................... 216
10.6 Other Cultural Material ..................................................... 223
10.7 Conclusion ....................................................................... 223
Chapter 11 Gibb Rock ................................................................................... 225
11.1 Introduction ....................................................................... 225
11.2 Site Description ................................................................ 225
11.2.1 Water Availability .................................................. 225
11.3 Survey and Excavation Methods ..................................... 226
11.4 Organics and Dating ........................................................ 229
11.5 Stone Artefact Assemblage ............................................. 230
11.5.1 Surface Collections ............................................... 230
11.5.2 GR1 ...................................................................... 232
11.6 Conclusion ....................................................................... 232
Chapter 12 Mulka’s Cave .............................................................................. 234
12.1 Introduction ....................................................................... 234
12.2 Site Description ................................................................ 234
12.2.1 Water Availability .................................................. 236
12.3 Previous Research ........................................................... 239
12.4 Survey and Excavation Methods ..................................... 241
12.4.1 Main Cave Area .................................................... 241
12.4.2 The Humps ........................................................... 243
12.4.3 Camping Area ....................................................... 244
12.5 Organics and Dating ........................................................ 245
12.6 Stone Artefact Assemblage ............................................. 251
12.6.1 Entrance Platforms ............................................... 251
12.6.2 Column Sample .................................................... 252
12.6.3 MC1 ...................................................................... 253
12.6.4 The Humps Scatter ............................................... 255
12.6.5 TH1 ....................................................................... 256
12.6.6 Camping Area Scatter .......................................... 258
12.6.7 CA1 ....................................................................... 262
12.6.8 CA2 ....................................................................... 267
12.7 Other Cultural Material ..................................................... 272
12.8 Conclusion ....................................................................... 273
x
Chapter 13 Identifying Technological Provisioning Systems......................... 275
13.1 Introduction ....................................................................... 275
13.2 Gibb Rock ........................................................................ 275
13.3 Anderson Rocks ............................................................... 277
13.3.1 Surface Collections ............................................... 277
13.3.2 AR1 ....................................................................... 278
13.3.3 AR2-3 ................................................................... 280
13.4 Mulka’s Cave .................................................................... 286
13.4.1 Main Cave Area .................................................... 286
13.4.2 The Humps ........................................................... 288
13.4.3 Camping Area ....................................................... 290
13.5 Conclusion ....................................................................... 295
SECTION FOUR: MERGING THEORY WITH ARCHAEOLOGY
Chapter 14 Mobility, Site Use and Occupation Intensity ............................... 297
14.1 Introduction ....................................................................... 297
14.2 Gibb Rock ........................................................................ 297
14.3 Anderson Rocks ............................................................... 298
14.3.1 Mobility and Site Use ............................................ 298
14.3.2 Occupation Intensity Over Time ........................... 302
14.4 Mulka’s Cave .................................................................... 304
14.4.1 Mobility and Site Use ............................................ 304
14.4.2 Occupation Intensity Over Time ........................... 310
14.5 Conclusion ....................................................................... 312
Chapter 15 Conclusion: Occupation History of the Study Area..................... 313
15.1 Introduction ....................................................................... 313
15.2 Addressing the Research Questions ............................... 313
15.2.1 Theoretical Resource and Occupation Model ...... 314
15.2.2 Timescale of Aboriginal Occupation ..................... 321
15.2.3 Mobility and Occupation Patterns .......................... 324
15.2.4 Occupation Intensity and Water Availability ......... 331
15.3 Future Research Directions ............................................. 337
15.4 Conclusion ....................................................................... 339
References ................................................................................................... 343
xi
APPENDICES
Appendix A Flora Species Present in the Study Area ................................... 376
Appendix B Distribution of Plant Foods in the Study Area ............................ 399
B.1
Monthly Availability of All Plant Foods .................. 401
B.2
Monthly Availability of Plant Foods in Granite ....... 405
B.3
Monthly Availability of Plant Foods in Heath ......... 408
B.4
Monthly Availability of Plant Foods in Mallee ........ 410
B.5
Monthly Availability of Plant Foods in Saline ......... 412
B.6
Monthly Availability of Plant Foods in Thicket ....... 414
B.7
Monthly Availability of Plant Foods in Woodland .. 416
Appendix C Distribution of Animal Foods in the Study Area ......................... 418
Appendix D Measuring Artefact Attributes .................................................... 425
D.1
Metric Attributes .................................................... 425
D.2
Retouch ................................................................. 426
D.3
Breakage ............................................................... 427
D.4
Flake Scar Orientation ........................................... 428
D.5
Bipolar Reduction .................................................. 429
Appendix E Anderson Rocks Stone Artefact Assemblage ............................ 431
E.1
Cores ..................................................................... 432
E.2
Small Flaked Fragments ....................................... 433
E.3
Large Flaked Fragments ....................................... 434
E.4
Small Flakes .......................................................... 435
E.5
Large Flakes ......................................................... 437
E.6
Retouched Flakes ................................................. 443
Appendix F Gibb Rock Stone Artefact Assemblage ...................................... 445
F.1
Cores ..................................................................... 446
F.2
Small Flaked Fragments ....................................... 446
F.3
Large Flaked Fragments ....................................... 447
F.4
Small Flakes .......................................................... 447
F.5
Large Flakes .......................................................... 448
Appendix G Mulka’s Cave Stone Artefact Assemblage ................................ 450
G.1
Cores .................................................................... 451
G.2
Small Flaked Fragments ....................................... 452
G.3
Large Flaked Fragments ....................................... 453
G.4
Small Flakes ......................................................... 456
G.5
Large Flakes ......................................................... 458
G.6
Retouched Flakes ................................................. 469
xii
LIST OF TABLES
Table 2.1
Forager models and theories for achieving optimisation,
based on the references listed and summary data from
Bettinger et al. (2015), Clarkson (2007), Kelly (2013) and
Winterhalder (1981). References denote the first iteration
of a model or significant subsequent contribution.
18
Table 2.2
Behavioural characteristics arising from different resource
structures. Re-tabulated from Ambrose and Lorenz
(1990:10).
19
Table 3.1
Summary of main palaeoenvironmental data from
southwestern Australia – research focussing on the
recent past has been omitted (e.g. Cullen and Grierson
2009; Treble et al. 2003). Temperature data from
Gouramanis et al. (2012) as it does not span the entire
record. Dates are cited in the format provided by the
author. * = Perth metro area, not marked on Figure 3.7.
48
Table 3.2
Vegetation communities found in the study area.
Community names, descriptions and classifications are
based on Beard (1972, 1975, 1976, 1979, 1980). Note
that these descriptions apply only to the study area (i.e.
the Hyden, Muntadgin and Skeleton Rock systems), and
may not reflect community characteristics at a broader
scale.
55
Table 4.1
Archaeological features in the study area, excluding those
analysed for this research. ID, name and classification
taken from DPLH’s Aboriginal Heritage Inquiry System
(AHIS), which separates Registered sites (RS) from
‘Other Heritage Places’ (OHP). Unless noted otherwise,
information derives solely from DPLH site files. The map
code indicates the position of a particular feature or set of
features on Figure 4.1—features are numbered from west
to east.
62
Table 4.2
Archaeological data from Smith’s (1993:271) interior
zone, showing the number of sites in different topographic
units, and the number of sites falling within different
assemblage size categories.
70
Table 5.1
Landscape Divisions and their associated vegetation
communities (see descriptions in Table 3.2 and
distribution in Figure 3.14) and landform. Vegetation
communities are defined following Beard (1972, 1975,
1976, 1979, 1980).
77
xiii
Table 5.2
Factors governing seasonal availability for plant and
animal foods, and generalisations used to determine
periods of availability. Where no governing factors were
present the items would be available year-round.
Generalisations were employed where data were lacking.
For details, see Appendices B–C.
85
Table 5.3
Years chosen to represent rainfall variability, showing
annual and winter rainfall totals (mm) and classification,
and the number of isolated falls > 15 mm, occurring
outside winter (from Hyden station, BoM ID 010568).
88
Table 5.4
Soil Water Tool (DPIRD 2016a) data fields and inputs
used. A soil profile was selected to represent each
Landscape Division (provided in parentheses after profile)
based on generalised soil characteristics (Beard 1979,
1990; DPIRD 2016b; Murphy-White 2007; Schoknecht
and Pathan 2013). Model years were chosen to represent
a variety of rainfall quantities and distributions, as per
Chapter 5.4.1. The 1992 and 1998 model periods were
abridged due to unusually high Plant Available Water
(PAW) balances occurring at germination, resulting from
model mechanics whereby plants begin to draw on a
store of water accumulated under preceding fallow
conditions. This does not accurately represent land under
native vegetation, as plants would be drawing water yearround. The April break-season rule was voided to ensure
that germination occurs during winter, so model periods
were more directly comparable. A single crop coefficient
was used throughout, to simulate mature native
vegetation.
91
Table 5.5
Ranked plant food categories, based on return rates
(cal/hr) determined by handling time (hr/kg) and
nutritional value (cal/kg). Handling time and nutritional
content are given only where return rates were calculated
manually. Asterisks indicate inferred handling times.
99
Table 6.1
Total number of plant food products found in the study
area. The number of plant species providing a particular
food item is listed, so the overall total indicates the total
number of plant food items in the study area rather than
the number of food-bearing species. A species may yield
more than one edible product. Plant uses are based on
Bindon (1996), Low (1991), Meagher (1974) and Webb
(2000); not all uses correspond to each food item and/or
species. Food items are ranked from highest to lowest, as
per Table 5.5.
102
xiv
Table 6.2
Spatio-temporal availability of all plant food categories
within the study area, showing the number of species
providing specific products during particular months.
Foods are ordered by rank, from highest to lowest.
103
Table 6.3
Spatio-temporal availability of all animal food categories
within the study area, showing the number of species
providing a certain product during each month, across all
Landscape Divisions.
116
Table 7.1
Longevity of PAW (days) in modelled years, across all six
Landscape Divisions.
134
Table 7.2
Winter and annual rainfall totals (mm) for each of the
modelled years, and the longevity of water (days) in each
rock structure. Years have been arranged by annual
rainfall total (low to high). Note that runoff gnammas
frequently held water from the end of one year and into
the next – this is not reflected in the longevity figures.
145
Table 7.3
Longevity of water (days) in various soils and rock
structures, as a result of differently sized isolated falls.
Durations are calculated assuming that falls occurred in
December and were received by a dry structure or profile.
Daily loss for soils is based on average December figures
(4.4 mm for Granite, 4.8 mm for Thicket), and recharge
amounts as per Chapter 7.2.2. Rainfall data sourced from
Bureau of Meteorology, for Hyden (010568) for 1929–
2017 (no data for 1947). Asterisks indicate that a
structure or profile has reached its storage capacity,
limiting longevity.
149
Table 8.1
The volume of water held by various capped rock
structures and the longevity of water supply (people
days). People days = number of people x number of days,
so ten people days may be one person consuming water
for ten days, or ten people consuming water for a single
day. Daily consumption was estimated at 1.5 L per
person.
152
Table 8.2
Temporal distribution of ranked plant food items (following
Table 5.5), and unranked animal foods in each
Landscape Division.
157
Table 9.1
Artefact type definitions and the attributes recorded for
each. Flaked fragments were separated based on their
maximum dimension, and flakes by their length.
182
xv
Table 9.2
Stone artefact attributes – further information regarding
methods can be found in Appendix D. Negative flake
scars < 5 mm in maximum/percussion length have been
omitted as they more likely result from retouch or use,
and single scars can be very difficult to separate (and
therefore quantify) on all but very high-quality quartz. The
references provided point to a definition of a particular
attribute, and its method of measurement.
183
Table 9.3
Features used to identify modified and unmodified glass,
and the attributes recorded for each. To be classified as
modified glass, the piece needed to meet at least one of
the criteria listed – all other pieces were classified as
unmodified. Criteria are based on Allen and Jones (1980),
Goward (2011:97), Harrison (2000), Holdaway and Stern
(2004:30), Paterson (1999), Runnels (1976), Veth and
O'Connor (2005) and Williamson (2004). Maximum
dimensions were recorded via the same method used for
stone artefacts (see Table 9.2 and Appendix D). Edge
damage is only considered evidence of modification only
where the site is relatively undisturbed, i.e. receives little
foot or vehicular traffic.
191
Table 9.4
Artefact attribute indicators for dominant and
supplementary technological provisioning systems, using
the pairing system established in Figure 9.5. The
indicators are only of value within paired systems
(Individual Provisioning and Gearing Up, Place
Provisioning and Short-term Expedient), since they
describe the relative position of each strategy on a
continuum, whose range of values will be different to that
of the alternate pair.
198
Table 10.1
Dimensions, capacity, shape, evaporation coefficient (EC)
and modelled runoff conditions for the three gnammas at
Anderson Rocks. Depth and radius were measured for
each structure and used to calculate capacity following
the standard mathematical formula for the volume of a
cone (see Figure 9.4).
205
Table 10.2
Soil pH values for selected XUs within AR1 and AR2-3.
209
Table 10.3
Charcoal samples and their returned radiocarbon ages,
from pit AR2-3. Dates are arranged by XU depth. Asterisk
beside sample weight indicates in-situ sample – all others
derived from sieve residue. Note that no 13C
measurements were provided with results.
211
Table 10.4
Number of stone artefacts from Anderson Rocks pits and
surface collections.
212
xvi
Table 10.5
Artefact types in surface collections at Anderson Rocks.
Figures in parentheses refer to high-quality materials.
Coll. = collection. Sc. = Scatter. OSM = other surface
material, from outside defined scatters.
212
Table 10.6
Stone artefact types by XU in AR1. Figures in
parentheses refer to high-quality raw materials.
215
Table 10.7
Stone artefact types by XU in AR2-3. Figures in
parentheses refer to high-quality raw materials.
217
Table 10.8
Location, weight and colour of ochre samples from AR23.
223
Table 11.1
Number of stone artefacts from the Gibb Rock pit and
surface collections.
230
Table 11.2
Frequency of artefact types found at Scatter 1 and
Scatter 2.
230
Table 11.3
Frequency of artefact types by XU in GR1.
232
Table 12.1
Surface dimensions, capacity, shape, evaporation
coefficient (EC) and modelled runoff conditions for the six
gnammas at Mulka's Cave. Dimensions of five traditional
gnammas after Webb and Rossi (2008), cleft rockhole
from Gunn (2004, 2006). Volumetric capacities were
calculated using standard mathematical formulae
provided in Figure 9.4.
237
Table 12.2
Radiocarbon dates from inside Mulka's Cave. Asterisk
denotes AMS date. CS = Column Sample (cut in 2006),
P88 = Bowdler et al. (1989) pit, excavated in 1988. Dates
have been arranged by depth below the 2006 ground
surface using Figure 12.8. Note that Bowdler et al. (1989)
excavated by stratigraphic unit, so the thickness of each
unit varied in different parts of the pit. The maximum
depths have been cited below. All CS samples came from
sieve residue, while the location of P88 samples are
unknown.
249
Table 12.3
Radiocarbon dates for the Humps and Camping Area
pits. Asterisk denotes AMS date. All samples were
collected from sieve residue.
250
Table 12.4
Number of stone artefacts from Mulka’s Cave pits and
surface collections. Assemblages have been grouped by
site area.
251
xvii
Table 12.5
Stone artefact types by XU from the Column Sample
inside Mulka’s Cave. Figures in parentheses refer to highquality quartz.
252
Table 12.6
Stone artefact types by XU in MC1. Figures in
parentheses refer to high-quality quartz.
254
Table 12.7
Stone artefact types by XU in TH1. Figures in
parentheses refer to high-quality quartz.
256
Table 12.8
Stone artefact types in the Camping Area Scatter. Figures
in parentheses refer to high-quality materials. Note that
‘other CA’ refers to the low-density material found outside
the clusters, but within the Camping Area Scatter.
258
Table 12.9
Stone artefact types by XU in CA1. Figures in
parentheses refer to high-quality materials.
262
Table 12.10
Stone artefact types by XU in CA2. Figures in
parentheses refer to high-quality materials.
267
Table 12.11
Location, weight and colour of ochre samples from
Mulka's Cave.
272
xviii
LIST OF FIGURES
Figure 1.1
Map of dated sites in Australia, after Williams et al. (2014:
Figure 1). The Holocene sites pre-date 10,000 BP. Black
box encases all sites in Noongar country; the specific
boundary of the Single Noongar Native Title Claim is
shown in Figure 1.2.
3
Figure 1.2
Left: Location of the Single Noongar Native Title Claim
(grey) and the study area boundary (blue). Right:
Enlarged view of the study area, showing the location of
archaeological sites mentioned in the text. The eastern
boundary follows the State Barrier Fence. Base imagery
from Google earth.
6
Figure 2.1
Optimal and suboptimal strategies for different resource
structures (Ambrose and Lorenz 1990:15).
19
Figure 2.2
The relationship between quality and abundance of local
stone, and the use of formal and informal tool production
(Andrefsky 1994:30).
26
Figure 2.3
Impact of mobility on provisioning strategies: A (left):
frequency of moves and occupation duration; B (right):
length of logistical forays. (Graf 2010, after Kuhn 1995).
30
Figure 3.1
Simplified geology of Western Australia, showing extent
of the Yilgarn Craton, also known as the Yilgarn Block
(Beard 1990:41).
37
Figure 3.2
Major regions within the Yilgarn Craton (Anand and Butt
2010:1018).
38
Figure 3.3
Geology of characteristic Wheatbelt valleys (Commander
et al. 2001, reproduced from Department of Environment
2005b).
39
Figure 3.4
Geology of selected parts of the study area (A) and
beyond (B), showing the occurrence and typical
distribution of potential tool-stone. Geological imagery
(including legend) from Geological Survey of Western
Australia (1984). Note that geological maps depict the
surface and near-surface geology, so raw material
sources may not have all been accessible at groundlevel.
41
Figure 3.5
Groundwater salinity across the Yilgarn Craton (Anand
and Butt 2010:1078).
43
xix
Figure 3.6
Recently dead vegetation near Lake Gounter, Hyden, in
October 2016 (Photograph: A M Rossi). Note the
replacement of large woody shrubs by much smaller
plants with a greater salt tolerance.
44
Figure 3.7
Palaeoenvironmental sites mentioned in the text. Those
in and around Perth are not marked.
47
Figure 3.8
Major climatic trends identified in southwestern Australian
palaeoenvironmental studies. Periods were classified in
relation to the preceding period; blue =wetter, red = drier,
black = no change, grey = uncertain. All dates have been
plotted as cited, although Churchill’s (1968) dates have
been converted from BC/AD to cal. BP and plotted on the
secondary axis with the data obtained by Gouramanis et
al. (2012).
47
Figure 3.9
Harrison’s (1993:219) reconstruction of lake status at
1000-year intervals, from 6000–1000 BP.
49
Figure 3.10
Bioclimatic zones of Western Australia (Beard 1990:39).
51
Figure 3.11
Annual rainfall data from Hyden weather station (no.
10568), 1929–2017 in relation to average rainfall (blue
dashed line). Data are not available for 1947 or 1974.
51
Figure 3.12
A: Natural regions of Western Australia (Beard 1990:34);
B: IRBA 7 regions of Western Australia (Australian
Government Department of Environment and Energy
2012). Abbreviations largely reflect Beard’s (1990)
subregion names except: AVW = Avon Wheatbelt; COO =
Coolgardie, GES = Geraldton Sandplains, JAF = Jarrah
Forest, SWA = Swan Coastal Plain, VIB = Victoria
Bonaparte, YAL = Yalgoo.
53
Figure 3.13
Catenary sequence for the Mallee Region (top) and Avon
Wheatbelt Region (bottom), showing the relationship
between topography and vegetation communities (Beard
1990:116, 129).
54
Figure 3.14
Pre-European distribution of vegetation communities in
the study area. Spatial data extracted from NRInfo
(DPIRD 2016b), based on Beard’s (1972, 1979, 1980)
data. Continuous black line separates the two IBRA
regions that meet in the study area: the Avon Wheatbelt
(north) and Mallee (south). Dashed line represents the
boundary of the study area.
56
xx
Figure 4.1
Registered sites and Other Heritage Places in the study
area. White squares denote approximate site positions,
enlarged in inset maps to show their immediate
surroundings. Locational data were retrieved from
Aboriginal Heritage Inquiry System (AHIS), maintained by
the Department of Planning, Lands and Heritage. The
central point of sites/place marked on insets. Site details
can be found in Table 4.1.
61
Figure 4.2
Locations mentioned in the text.
65
Figure 4.3
Area covered by South West Native Title Settlement,
showing subdivisions (Source: National Native Title
Tribunal). ILUA = Indigenous Land Use Agreement.
72
Figure 4.4
Aboriginal social boundaries in southwestern Australia
according to various sources (redrawn from original maps
by A.M. Rossi). A: Curr (1886,1887) separated groups
into the non-circumcising Western Division (green) and
the circumcising Central Division (beige). Group names
are given in italics, where known. Red crosses indicate
locations where circumcision was practised, but recorded
locations are often vague and their placement on the map
may be somewhat arbitrary. E.g. the cross within Kokor
territory is noted in text as ‘east of York’. Original image
provided courtesy of Fisher Rare and Special Collections,
University of Sydney. B: Bates (White 1985) placed the
division for her non-circumcising group (the Bibbulmum)
further inland. Her eastern boundary locations are shown
in blue. At the time of her research, Karlgarin was the
most inland town established in the vicinity of the study
area, so was the only real point of reference. The western
boundary for the circumcising group (Karatjibbin –
specific locations shown in red) is considerably further
east. C: Davidson (1938). D: Tindale (1974). Groups east
of his circumcision/subincision boundary are not shown.
73
xxi
Figure 5.1
Vegetation and landform characteristic of each
Landscape Division – all photographs and species
information from Beard (1990:119, 121, 123, 131–132,
134, 136). A–B: Saline, dominated by short, salt-tolerant
shrubs (Atriplex spp., Tecticornia spp.). Note the taller
Eucalyptus sp. trees on slightly higher ground in B. C:
Eucalyptus spp. Woodland, note the generally open
nature of the canopy and the relatively sparse
understorey. D: Mallee, note the multi-stemmed Eucalypts
and the thicker understorey than in C. E–F: Heath,
dominated by a diverse range of low shrubs including
Grevillea spp. and Verticordia spp. Note the Mallee in the
background of E, indicating the mosaic nature of
Landscape Divisions. G: Thicket of Acacia, Allocasuarina
and Melaleuca species – note the closely spaced shrubs.
H: Granite. Note the thicker vegetation at the base of the
slope, due to water runoff. The rock is mostly bare except
where pockets of soil support shrub and small tree
growth.
78
Figure 5.2
Distribution of Landscape Divisions within the study area
– note that the Saline Landscape Division comprises both
saline valleys and salt lakes. Spatial data extracted from
NRInfo (DPIRD 2016b). Examples of each Landscape
Division are shown in Figure 5.1.
79
Figure 5.3
The relative water content of various soil types, showing
available and unavailable water (Moore 2001:82).
Available water can be extracted by plants, referred to
here as Plant Available Water (PAW). Note that a soil
may be holding water, even if none is available for plant
use (unavailable water); coarser soils (those with larger
particle sizes) have lower water-storage capacity than
finer soils but require less input before plants are able to
extract water.
89
Figure 5.4
Example output from the Soil Water Tool (DPIRD 2016a),
showing soil water (PAW) under fallow (black line) and
cropped conditions (green line), plus daily rainfall (blue
bar). Fallow and cropped values will be in-sync until
germination date ('break'); thereafter, modelled crops
draw on stored water accumulated under preceding fallow
conditions. An increase in the PAW balance is registered
only when daily input (rainfall) exceeds daily loss (through
evaporation/evapotranspiration). Daily data values (box)
are displayed when the mouse is hovered over the
relevant point on the chart output (black filled circle).
92
xxii
Figure 6.1
Examples of plant foods found in the study area –
photographs and information sources as cited. A: Acacia
acuminata, 3–5 m tall, bearing edible seeds and gum
(Bindon 1996:251). B: Edible gum exuded from Acacia
spp. (Low 1991:152). C: Typical Acacia seed pod
(Sweedman and Merritt 2006:176). D: Woody fruits (2–4
cm long) containing edible seeds of Allocasuarina humilis
(Erickson et al. 1979:21). E–F: Eucalyptus redunca and
its flowers; tree to 6 m tall (Bindon 1996:131). G–H: Lerps
on Eucalyptus spp. leaves (Low 1991:153). I:
Chamelaucium megalopetalum, edible flowers 10–15 mm
wide (Erickson et al. 1979:88). J: Enchylaena tomentosa
fruits, approximately 5 mm wide (Low 1991:167). K:
Thysanotus patersonii tubers, generally 10–30 mm long
and 5 mm wide (Bindon 1996:251). L: Carpobrotus
virescens, fruits and leaves (up to 65 mm long) are both
eaten (Daw et al. 1997:4–5). M: Atriplex vescaria shrub (≤
1 m high) which bears edible leaves (Bindon 1996:46). N:
Edible fruits of Santalum acuminatum, tree grows to 5 m,
fruit 20–40 mm wide (Daw et al. 1997:52–53).
105
Figure 6.2
Plant food categories demonstrating strong temporal
variation in availability, showing monthly species diversity
within each Landscape Division. The same colour key
applies to each chart.
107
Figure 6.3
Total number of different plant foods within each
Landscape Division.
112
Figure 6.4
Total number of plant food items (top) and rank (bottom)
available in each month across the entire study area.
113
Figure 6.5
Examples of birds found in the study area. Measurements
refer to the body length (from crown to tip of the tail)
unless noted otherwise. Common names follow species
name. Photographs and measurements from Nevill
(2014:52–55,
78–81,
90–91,
106–111).
A–C:
Passeriformes (perching birds): Phylidonyris niger, whitecheeked honeyeater, 18 cm (A); Pomatostomus
superciliosus, white-browed babbler, 21 cm (B); Petroica
goodenovii, red-capped robin, 12 cm (C). D–G:
Waterbirds: Himantopus himantopus, black-winged stilt,
37 cm (D); Tadorna tadornoides, Australian shelduck, 56–
72 cm (E); Charadrius ruficapillus, red-capped plover, 15
cm (F); Vanellus tricolor, banded lapwing, 27 cm (G). H:
Dromaius novaehollandiae, emu, 160–190 cm tall. I:
Leipoa ocellata, malleefowl, 55–60 cm. J: Barnardius
zonarius, Australian ringneck, 36 cm.
115
xxiii
Figure 6.6
Monthly availability of birds (top) and birds’ eggs (bottom)
in each Landscape Division. The same colour key applies
to each chart.
118
Figure 6.7
Example of native non-avian fauna present in the study
area. Common names are provided after species name,
measurements cited are maximum total length (TL –
including tail) or body length (BL – excluding tail)
Photographs and measurements from Nevill (2014:124–
129, 136–137, 142–143, 151, 156–157, 162–163, 184–
185). A–D: Reptiles: Ctenophorus cristatus, bicycle
dragon, TL 370 mm = (A); Liopholis multiscutata, bull
skink, TL 250 mm (B); Underwoodisaurus milii, barking
gecko, TL 165 mm (C); Varanus gouldii, sand monitor, TL
1.6 m (D). E–K: Mammals: Sminthopsis crassicaudata
(family Dasyuridae), fat-tailed dunnart, BL 75 mm (E);
Dasyurus geoffroii (family Dasyuridae), chuditch, BL 310–
360 mm (F); Macropus fuliginosus, western grey
kangaroo, BL 0.95–2.2 m (G); Chalinolobus gouldii,
Gould’s wattled bat, BL 65–75 mm (H); Isoodon obesulus,
southern brown bandicoot, BL 340 mm (I); Trichosurus
vulpecula, common brushtail possum, BL 380 mm (J);
Tachyglossus aculeatus, short-beaked echidna, BL 400
mm (K). L: Myobatrachus gouldii, turtle frog, BL 50 mm.
120
Figure 6.8
Total number of animal food items available in each
month across the entire study area.
123
Figure 6.9
Number of different animal food items within each
Landscape Division. Note that Heath and Thicket both
preserve a single amphibian species, not visible at the
scale of this figure.
123
Figure 7.1
Daily moisture balance of soils in each Landscape
Division (mm of plant available water – PAW – coloured
lines) and rainfall (black bars) for years with low winter
rainfall. Note that the same colour key applies to both.
127
Figure 7.2
Daily moisture balance of soils in each Landscape
Division (mm of plant available water – PAW – coloured
lines) and rainfall (black bars) for years with average
winter rainfall. Note that the same colour key applies to
both.
129
Figure 7.3
Daily moisture balance of soils in each Landscape
Division (mm of plant available water – PAW – coloured
lines) and rainfall (black bars) for years with high winter
rainfall. Note that the same colour key applies to both.
133
xxiv
Figure 7.4
Daily water balance (mm) in rock structures (coloured
lines) and rainfall (black bars) for years with low winter
rainfall. Note that the same colour key applies to both.
137
Figure 7.5
Daily water balance (mm) in rock structures (coloured
lines) and rainfall (black bars) for years with average
winter rainfall. Note that the same colour key applies to
both.
141
Figure 7.6
Daily water balance (mm) in rock structures (coloured
lines) and rainfall (black bars) for years with high winter
rainfall. Note that the same colour key applies to both.
143
Figure 7.7
Frequency of different rainfall scenarios in the study area
from 1929–2017. A (left): number of low, average and
high winter rainfall years. B (right): frequency of variously
sized isolated falls. Based on Bureau of Meteorology data
for the Hyden station (010568); note no data were
recorded in 1947.
147
Figure 7.8
Generalised water availability in soils and rock structures
in each Landscape Division, in average (top), low
(middle) and high (bottom) winter rainfall years. Note that
Heath, Mallee, Saline, Thicket and Woodland areas only
provide soil water. Solid lines indicate the generally
consistent presence of water over the specified period,
i.e. being dry for no more than a few days at a time.
Dashed lines indicate when water may be available – its
presence/absence depends on the timing of rainfall
events. Unusually large pre- or post-winter falls have
been voided (see Chapter 5.4.3), as they mask the
normal longevity of water from winter rain. Water-holding
periods for average (2009, 2011) and high (1992, 1996)
winter rainfall years are based on the mean of both years.
For low winter rainfall years (2010, 2012), only 2012 data
was used as they were deemed more broadly
representative of low rainfall conditions. In winter 2010 a
few large falls were interrupted by three weeks of limited
rainfall, causing soils to dry out when they normally would
not have. Mallee soils are not pictured for low rainfall
conditions as they held water for just a single day.
148
Figure 8.1
Occupation model for the study area under average
rainfall conditions, showing the primary water sources
(blue), primary foraging (green) and residential areas
(yellow) and the optimal portion of the study area
(orange) that should be occupied at particular times of the
year. For division location see Figure 8.2.
155
xxv
Figure 8.2
Location of the northern, central, southern parts of the
study area and the optimal months to occupy each. Note
that these divisions are not firm boundaries, but merely
indicative of the broad areas referred to in the text and
Figure 8.1.
156
Figure 9.1
Location of archaeological sites excavated by the author.
174
Figure 9.2
Procedural key used to classify flaked stone via the
materialist approach developed by Hiscock (2007).
Unflaked stone preserves no evidence of flake removals
or having been detached from a larger piece. Note that
ventral surfaces were identified via any of the normal
landmarks, such as bulb of percussion, ripple marks,
termination etc, following Holdaway and Stern
(2004:108). Retouch scars must be initiated from or
modify the ventral surface – see Table 9.1 and Appendix
D.2 for more information. Note the amendments for
quartz-dominated assemblages, including the use of
impact points or shatter facets to identify flaked quartz, as
well as flaked fragments incorporating those pieces
where breakages or poor-quality material make it difficult
to confirm the presence/absence of a ventral surface.
Similarly, poor-quality material likely masked retouch in
several instances – where such uncertainty was
encountered, pieces were classified as in the lower
category, as flakes.
181
Figure 9.3
Variable quality of quartz within the study area. Quality
increases from left to right. Top row: very poor. Second
row: poor–moderate. Third row: moderate–good. Bottom
row: high quality. White bars are 1 cm.
186
Figure 9.4
Shapes used to represent morphological variation
between individual gnammas, and the mathematical
formulae used to determine their total capacity and the
volume of water with depth.
193
Figure 9.5
Major differences in technological organisation between
the four provisioning systems: Place Provisioning (PP),
Individual Provisioning (IP), Gearing Up (GU), and Shortterm Expedient (EXP).
196
xxvi
Figure 10.1
Features seen at Anderson Rocks – all photographs by
A.M. Rossi. Individual pictured is 1.64 m tall. A–B: Northfacing view of quartz vein striking north-south up the
outcrop. C: Large pan (10.9 m wide) on top of outcrop,
facing east. Note the outflow channel leading downslope.
D: Deep channels created by water flow – these link
many pans together in an interconnected system. E–F:
Pans linked by channels. Note the higher elevation of
some pans that, when full, would overflow into the
channels directing water to lower lying structures. G:
South-east facing view of large pan that overflows
downslope to gnammas (behind small circular patch of
shrubs in top right of photo). This pan receives runoff
from upper slopes and channel systems. H: North-west
facing view of the two gnammas that receive runoff from
upslope pan and channel system (which continues to flow
after rain has ceased), as well as the surrounding slope.
Note that the well-defined flow of water does not reach
the third gnamma, which lies to the left of those pictured.
I: South-east facing view of all three gnammas – the black
staining indicating inflow channels to the left and central
gnamma. The smallest gnamma (right) receives runoff
only from its immediate surrounds. Largest gnamma (left)
has diameter of 1.43 m.
203
Figure 10.2
Total volume of water stored within Anderson Rocks
gnammas during high (top), average (middle) and low
(bottom) winter rainfall years.
206
Figure 10.3
A: Aerial photograph of Anderson Rocks. White square
indicates the area magnified in image B. B: Main
surveyed area of Anderson Rocks, showing the location
of gnammas, artefact scatters and pits. Note the thicker
vegetation separating the eastern and western
exposures. Features too small to outline at the scale of
this figure have been indicated by arrows. Base imagery
from Google Earth.
207
Figure 10.4
A: Vegetated area near Scatter 1 and pit AR1, facing
north-east. Note that the scatter and pit were immediately
downslope of the thick vegetation pictured, where only
low ground cover was present. B: Scatter 2, located
within a small pan bisected by a quartz vein, facing north.
Pan measures 1.63 m long by 0.91 m wide (maximum
dimensions). C: close-up of the quartz vein in Scatter 2.
D: Location of Scatter 3 and pit AR2-3 within open
vegetated area southeast of the gnammas, facing southwest. Note the granite slab marks the position of the
backfilled pit. All photographs by A.M. Rossi.
208
xxvii
Figure 10.5
Charcoal weight with depth in AR1 (left) and AR2-3
(right).
210
Figure 10.6
Depth-age curve for AR2-3. Solid line joins radiocarbon
age determinations, while linear trendline (dotted line)
was used to estimate the age of the lowest cultural
material, from XU32 (-825 mm).
211
Figure 10.7
Artefact discard rates in AR2-3. The rate has been
adjusted to artefacts/0.5 m2/100 years to allow for the
difference in excavated area in the lower deposits, as
only pit AR3 continued below 600 mm.
218
Figure 10.8
Proportion of small flakes on high-quality material (left),
and breakage rate for all small flakes (right) from AR2-3.
220
Figure 10.9
Selected attributes recorded on large quartz flakes, by
XU, in AR2-3. A-B (top): Average negative flake scar
(blue bars) and platform counts (orange bars) for highquality materials (A, left) and lower-quality materials (B,
right). C-D: average mm/scar values (C, left) and
proportion of broken flakes (D, right) for high-quality
materials (black bars) and lower-quality materials (grey
bars).
221
Figure 11.1
Shallow pans on eastern side of Gibb Rock; B: Fenced off
wall to capture runoff from the outcrop, facing west. Both
photographs by A.M. Rossi.
226
Figure 11.2
A: Aerial photograph of Gibb Rock. White rectangle
indicates area magnified in image B. Note the extensive
wall constructed to filter runoff to the large circular tank,
near the car park on the eastern side of the outcrop. B:
Location of Scatter 1 and Scatter 2 (within white
boundaries), and pit GR1 (red arrow). Note the thicker
vegetation north of the car park area. Base imagery from
Google Earth.
227
Figure 11.3
A: Area around Scatter 1, facing south. B: Area around
Scatter 2, facing east. C: quartz weathering from parent
granite, in Scatter 1. All photographs by A.M. Rossi.
228
Figure 11.4
Weight of charcoal with depth in pit GR1.
229
xxviii
Figure 12.1
Cultural and natural features present at The Humps. All
photographs by A.M. Rossi. A: the main entrance to
Mulka's Cave, facing west. Note the raised platforms
either side of the entrance, created when erosion
removed the sediments below. B: Two paintings inside
the entrance to Mulka's Cave, note the yellow handstencil
inside the circular motif. Lefthand painting measures 127
x 36 cm, righthand painting 83 cm wide (Gunn 2004:26).
C: Raised walkway outside Mulka's Cave, facing south.
D: Red and white pigmented handstencils and handprints
inside Mulka's Cave. E: View of raised walkway from the
top of Mulka's Cave. F: Gnamma 1, measuring 1.4 m long
and 0.85 m wide. Note the shadows cast by the gnamma
walls, protecting the water from evaporation.
235
Figure 12.2
A: Aerial photograph of The Humps. White rectangle
indicates the area magnified in image B. B: Location of
natural and cultural features around the Humps, including
Mulka's Cave itself, the Humps Scatter, Camping Area
Scatter, lizard trap and the main water sources. Features
too small to outline at the scale of this figure are indicated
by arrows. Pit locations for Mulka's Cave and the
Camping Area are illustrated in Figure 12.4 and Figure
12.6. Locations for cleft rockhole and Gnamma 5 are
approximations, based on Figure 3 from Gunn (2006:22).
Base imagery from Google Earth.
236
Figure 12.3
Total volume of water available at Mulka's Cave from the
gnammas and cleft rockhole (solid line) and the gnammas
only (dashed line), during high (top), average (middle)
and low (bottom) winter rainfall years.
238
Figure 12.4
Main Cave Area, facing west, showing the location of the
Entrance Platforms (flat areas indicated by white arrows),
Column Sample and pit MC1 (red arrows). Photograph:
A.M. Rossi.
243
Figure 12.5
A: View of The Humps outcrop taken from the area
around Gnammas 1 and 2, facing south west. Note the
sparsely vegetated shoulder on which the Humps Scatter
is located. B: The Humps Scatter, showing pit TH1, facing
north east. Gnammas 1 and 2 occur on the flat granite
exposure visible in the top right of the image. C: A portion
of the Camping Area, facing west. Note the relatively
clear area surrounded by thicker border vegetation – no
artefacts were present in the latter area.
244
xxix
Figure 12.6
Camping Area artefact scatter, showing the extent of
surface artefacts (light grey shading), the location of the
high-density surface clusters (Clusters A–F, dashed
circles) and pits CA1 and CA2. Note the scatter is
surrounded by thick border vegetation (dark grey
shading).
245
Figure 12.7
Charcoal weight by XU within pits CA1 (top), CA2
(middle) and TH1 (bottom).
246
Figure 12.8
Stratigraphic section (Bowdler et al. 1989: Figure 4.2a)
superimposed with their XU levels (blue), sediment
column (orange) and the approximate position of the
2006 Column Sample (pink). XUs numbered by each
author. Rossi (2014a) after Bowdler et al. (1989).
248
Figure 12.9
Depth-age curves for inside Mulka’s Cave (left) and other
parts of the site (right). Solid lines join radiocarbon age
determinations, while linear trendlines (dotted lines) were
used to estimate the age of the lowest cultural material,
represented by the end point of those lines. Inside
Mulka’s Cave, Wk-27113 (P88 XU15) and Wk-29188
(P88 XU16) were statistically indistinguishable. The lower
sample was selected as it was derived from a larger
charcoal sample.
249
Figure 12.10
Small core from MC1 XU4. Note the presence of four
original crystal facets (indicated by blue arrows).
254
Figure 12.11
Artefact discard rates for pit TH1.
256
Figure 12.12
Selected attributes recorded on large quartz flakes within
the Camping Area Scatter. Cl. = Cluster. A–B (top):
Average negative flake scar (blue bars) and platform
counts (orange bars) for high-quality materials (A – left)
and lower quality materials (B – right). C–D: average
mm/scar values (C – left) and proportion of broken flakes
(D – right) for high-quality materials (black bars) and
lower quality materials (grey bars).
260
Figure 12.13
Artefact discard rates for pit CA1.
262
Figure 12.14
Proportion of small flakes on high-quality material (left),
and breakage rate for all small flakes (right) from pit CA1.
264
xxx
Figure 12.15
Selected attributes recorded on large quartz flakes, by
XU, in CA1. A–B (top): Average negative flake scar (blue
bars) and platform counts (orange bars) for high-quality
materials (A – left) and lower quality materials (B – right).
C–D: average mm/scar values (C – left) and proportion of
broken flakes (D – right) for high-quality materials (black
bars) and lower quality materials (grey bars).
265
Figure 12.16
Dorsal surface of three quartz flakes from CA1 XU2. Note
the brown crust that covers most of the dorsal surface
(including negative flake scars) and has been removed in
other areas; this crust could not be removed by washing
or rubbing. Crust was absent on ventral surfaces. White
bar = 1 cm.
266
Figure 12.17
Artefact discard rate in pit CA2.
267
Figure 12.18
Proportion of small flakes on high-quality material (left),
and breakage rate for all small flakes (right) from pit CA2.
269
Figure 12.19
Selected attributes recorded on large quartz flakes, by
XU, in CA2. A–B (top): Average negative flake scar (blue
bars) and platform counts (orange bars) for high-quality
materials (A – left) and lower quality materials (B – right).
C–D: average mm/scar values (C – left) and proportion of
broken flakes (D – right) for high-quality materials (black
bars) and lower quality materials (grey bars).
270
Figure 12.20
A (left): quartz flake from CA2 XU4. Lateral margins
preserve notches formed by previous flake removals.
Black bar = 1 cm. B (right): an idealised core that could
produce notched flakes. Note the long, parallel flake
scars, and the scalloped edge evident in cross section. If
the platform or base of the core was removed, flakes
would exhibit similar scalloping.
271
Figure 12.21
Flaked glass from Entrance Platforms at Mulka's Cave.
Note the small negative flake scars concentrated along
the right distal margin, while the other remains sharp and
intact. Stippling represents unmodified surface, black bar
= 1 cm.
273
Figure 14.1
Technological provisioning systems (top) and occupation
intensity (bottom) at Anderson Rocks, based on dated
archaeological content from AR2-3 and surface material;
the latter was presumed to date from the last 50 years or
so. Where provisioning systems could not be confidently
identified, these have been omitted.
301
xxxi
Figure 14.2
Technological provisioning systems (top) and occupation
intensity (bottom) at Mulka’s Cave. Provisioning systems
are shown for all dated pits and the combined surface
assemblages, which were presumed to date from the last
50 years. CS = Column Sample. Where provisioning
systems could not be confidently identified, these have
been omitted. Occupation intensity is based on
assemblages from the Camping Area.
308
Figure 15.1
Occupation
periods
represented
by
subsurface
archaeological deposits at Anderson Rocks (dark blue),
Gibb Rock (red) and Mulka's Cave (light blue). CS =
Column Sample, P88 = 1988 pit excavated by Bowdler et
al. (1989). GR1 was not dated but based on the short
depth of deposit the assemblage is probably no more
than a few hundred years old. Note the presence of
cultural material throughout the occupation periods,
indicating no significant hiatuses in site visitation.
322
Figure 15.2
Occupation intensity (top) for each site and effective
rainfall conditions for dated sites (middle) and the study
area as a whole (bottom). Note the Gibb Rock
assemblages were presumed to date to the last few
hundred years, and all surface assemblages to the last 50
years.
334
xxxii
CHAPTER 1
INTRODUCTION: ADDRESSING THE SCARCITY OF ARCHAEOLOGICAL DATA
FROM INLAND SOUTHWESTERN AUSTRALIA
1.1 INTRODUCTION
This chapter provides the background to the author's research into Aboriginal
occupation of inland southwestern Australia. To begin, the archaeology of the
Noongar country, southwestern Australia, is considered, with specific reference to
the research bias that has focussed on coastal areas and neglected the inland
environs. The archaeological and ethnohistorical data sources are then summarised
for the coastal southwest and the Western Desert. These provide a stark contrast to
the limited information available for the inland southwest but cannot be applied due
to considerably different environmental conditions. The study area is then defined,
with reference to Tindale's (1974) boundary that separates the Noongar from the
desert groups, as well as a distinct natural boundary that it replicates. The major
research aim is outlined, and four specific research questions are devised to address
it. The methodology forms a two-pronged approach, whereby a theoretical
occupation model based on resource distribution is integrated with archaeological
data. The latter includes that collected by the author and the fragmentary remains
documented in the wider area. The methods associated with each approach are
outlined, and the limitations discussed. The significance of this research is then
evaluated, both in the results it aims to provide, and the novel approach employed.
Finally, the organisation of the thesis is outlined, with reference to the four distinct
sections; each address a particular topic throughout several chapters.
1.2 ARCHAEOLOGICAL RESEARCH BIAS IN THE SOUTHWEST
When compared to other parts of the continent, Western Australia is not particularly
well represented in the national database of dated archaeological sites. However,
Noongar country – defined here as the area covered by the Single Noongar Native
Title Claim – preserves a particularly problematic record. While sites are present
throughout southwestern Australia, most excavated and dated sites are located
within 100 km of the present coastline (Figure 1.1). The area yields some of the
1
oldest sites in Australia – Devil's Lair and Upper Swan both preserve occupation
sequences dating to around 45 ka, and the former was arguably visited from 50 ka
(Allen and O'Connell 2014; Balme 2014; Pearce and Barbetti 1981; Turney et al.
2001). All other sites post-date 22,000 cal. BP, and many show occupation
throughout the Last Glacial Maximum, 21 ± 3 ka (Dortch 1996; Reeves et al. 2013;
Williams and Smith 2013). Various site types are represented, including rockshelters,
quarries, open sites, rock art sites, and shell middens. While the specifics of these
sites are not relevant here, the diversity of this archaeological sample is noteworthy.
The ethnohistorical record is similarly varied, unsurprising since early European
activity was focussed on the coast, particularly around Perth and Albany. There is an
extensive collection of letters, diaries and field notebooks that contain detailed
observations of Aboriginal people in the 19th century (e.g. Grey 1841; Hammond
1933; Hassell 1975; Hercock 2014; Moore 1884 – see list in Haebich and Tilbrook
1981). This substantial body of archaeological and ethnohistorical data has provided
a wealth of knowledge on various facets of Aboriginal life, including occupation
patterns (e.g. Anderson 1984; Bird 1985; Smith 1993), land and resource
management practices (e.g. Hallam 1975, 1989; Lullfitz et al. 2017), and food
sources (e.g. Meagher 1974).
The inland southwest presents a stark contrast to the rich data sources described
above. Only one archaeological site has been dated – Mulka's Cave, a decorated
rockshelter that was the focus of the author's MA research (Rossi 2010, 2014a).
While Rossi (2014a) demonstrated that initial occupation occurred around 8000 cal.
BP, the stone artefact assemblage was of little interpretive value due to the lack of
comparative archaeological data. Ethnohistorical data was also scarce, so the author
was unable to form any strong conclusions about how the site was used, or how it fit
into the broader context of southwestern Australia.
2
Figure 1.1 Map of dated sites in Australia, after Williams et al. (2014: Figure 1). The Holocene sites
pre-date 10,000 BP. Black box encases all sites in Noongar country; the specific boundary of the
Single Noongar Native Title Claim is shown in Figure 1.2.
Several archaeological locations are known in the area surrounding Mulka's Cave,
but most were identified during cultural heritage surveys for development works, so
limited data were recorded and made available. Bird (1985) and Smith (1993) both
considered some inland areas in their large-scale analyses, but the data sources
were optimised for their individual research questions, so were of little comparative
value. While several ethnohistorical sources referred to the Aboriginal group that
lived around Hyden (e.g. Davidson 1938, Tindale 1974), other observations were
limited to a few brief statements by Roe (1836, 1852), Landor and Lefroy (1843) and
Hammond (1933), from which little information could be drawn. This problem persists
today and typifies the scarcity of archaeological and ethnohistorical data for the
entire inland southwest.
This same research bias is not evident elsewhere in the state – there is an extensive
suite of sites in the Western Desert, which also stretches into the Northern Territory
and South Australia. As in the coastal southwest, site types are varied, and range
from rockshelters and rock art sites to ochre mines and stone arrangements
3
(Williams et al. 2008). Several Pleistocene sites are present, including Karnatukul
(formerly Serpent's Glen), which was occupied from 47,830 cal. BP (McDonald et al.
2018); several other sites date to around 35,000 cal. BP (Smith 2013:79). The
ethnohistorical record is also fairly rich, but for a different reason to the coastal
southwest. Due to the remote location, European settlement occurred much later,
and the process of cultural dislocation proceeded much more slowly (Haebich and
Tilbrook 1981:2). Therefore, by the time Tonkinson and Gould began their fieldwork
in the 1960s, they were able to observe many facets of traditional life that had been
lost elsewhere. Both published various accounts and addressed several topics,
including mobility and occupation patterns, subsistence behaviour, and aggregation
events (e.g. Gould 1968, 1969a, 1969b, 1971a, 1971b, 1984, 1991; Tonkinson
1978).
While both the coastal southwest and the Western Desert yield a relative abundance
of archaeological and ethnohistorical data, it is of limited relevance to the inland
southwest. The coastal regions receive more rainfall, and preserve freshwater lakes,
rivers and potable groundwater; rainfall is more limited in the inland southwest, and
more variable year-to-year (Bureau of Meteorology data). Other differences in the
climate, hydrology and geology of each area has created distinct vegetation
communities, which harbour different faunal assemblages. Therefore, there is no
reason to expect a similar antiquity of occupation, nor that demographic or mobility
patterns would be comparable between the coastal and inland southwest. Similarly,
the Western Desert data cannot be applied. All Pleistocene sites were associated
with a reliable water source of some variety, while none are found in the inland
southwest (Bird et al. 2016). Further, in the Western Desert, rainfall is erratic, both
within and between years, and can be extremely localised (Gould 1991; Tonkinson
1978:7). As a result, mobility patterns would have been vastly different than in the
inland southwest where rainfall is highly seasonal. In contrast to the coastal
southwest and Western Desert, then, little is known about the antiquity or nature of
Aboriginal occupation of the inland southwest. It is this extensive knowledge gap that
the author wishes to address with the current research.
4
1.3 STUDY AREA, AIMS AND RESEARCH QUESTIONS
While the research gap identified above affects most of inland southwestern
Australia, it was necessary to define a more manageable-sized area for this research
to focus on. The position of the study area was primarily dictated by Mulka's Cave,
the only excavated and dated site in the area. A targeted sample of granite outcrops
identified three other sites with archaeological potential. Therefore, the study area
was defined as a square measuring approx. 100 x 100 km, centred on the point
between the sites (Figure 1.2); the area east of the State Barrier Fence (formerly No.
1 Rabbit Proof Fence) was excluded, as it broadly coincides with a significant
cultural and natural boundary. Near the Hyden township, the fence-line closely
resembles
the
position
of
Tindale's
(1974)
frontier
that
divides
circumcising/subincising desert groups from the Noongar, who did neither; this
boundary defines the eastern extent of the Single Noongar Native Title Claim. While
the circumcision boundary was not solid, nor necessarily static (Gibbs and Veth
2002), it is reasonable to assume that, as Noongar, the study area occupants may
not have had regular access to the land further east.
The State Barrier Fence also separates the Eastern Goldfields region from other
parts of southwestern Australia. At a broad scale, the geology, drainage and climate
of the study area has more in common with the Eastern Goldfields than the areas to
the west (Anand and Butt 2010 – see Chapter 3.2). At a finer scale, however,
significant differences emerge. Greenstone belts (bands of metavolcanic and
metasedimentary rock) are common in the Eastern Goldfields, while in the study
area basement rock is almost exclusively granite. These geological variations create
distinct soil profiles, each of which have different water retention properties; this
strongly influences vegetation formations, which in turn affects faunal assemblages.
These differences were significant enough that the Eastern Goldfields was separated
from southwestern Australia for the purposes of biodiversity surveys (e.g. Biological
Surveys Committee 1984). Considering the importance of soil, flora and fauna in the
resource and occupation model (see Chapter 1.4), the author chose to exclude the
area east of the State Barrier Fence.
5
Figure 1.2 Left: Location of the Single Noongar Native Title Claim (grey) and the study area boundary
(blue). Right: Enlarged view of the study area, showing the location of archaeological sites mentioned
in the text. The eastern boundary follows the State Barrier Fence. Base imagery from Google earth.
The main aim of this research is to address the knowledge gap for inland
southwestern Australia by producing an occupation history for the study area.
Specifically, the main aim is to answer one particular question: when and how did
Aboriginal people use the study area, and did the nature and/or intensity change
over time? This broader aim will be achieved by considering four specific research
questions:
1. How were food and water distributed across the study area and, following
Optimal Foraging Theory, how would seasonal and interannual resource
availability influence human occupation patterns?
2. What is the timescale of Aboriginal occupation in the study area, and how
does this compare to other parts of Australia?
3. What do the archaeological remains reveal about mobility and occupation
patterns within the study area, and how do these results fit within the
theoretical model?
4. How did occupation intensity vary over time, and can this variation be
explained with reference to long-term shifts in water availability?
6
The methods of answering these questions, and their limitations, are discussed
below.
1.4 METHOD OF ADDRESSING RESEARCH QUESTIONS
While no human society should be viewed through a framework of environmental
determinism, there is no doubt that the natural environment and landscape places
severe limitations on how humans can utilise an area. In the context of Indigenous
occupation of Australia, the most important factors were the availability of water, food
plants and animals. Therefore, this research employs a two-pronged approach. The
first requires construction of a theoretical occupation model based on critical
resource distribution, while the second involves archaeological investigations. These
complimentary data sources will be merged to provide an occupation history for the
study area. The methods for each stage of the approach, and the associated
limitations, are discussed briefly below; more detailed information is available in
Chapters 5 and 9.
1.4.1 Resource and Occupation Model
Before the occupation model can be constructed, it is necessary to quantify food and
water distribution throughout study area, allowing for variation over space and time.
To address spatial variation, several Landscape Divisions will be defined, based on
dominant landform and vegetation associations in the study area as recorded by
Beard (1972, 1979, 1980); food and water resources will be quantified for each
Division. Lists of native flora and fauna will be compiled from the biodiversity surveys
conducted on various nature reserves in the study area; vegetation and landform
descriptions will be used to identify the relevant Landscape Division/s represented
therein. Plant foods will be identified using published lists (e.g. Bindon 1996; Bird
and Beeck 1988; Meagher 1974), allowing for unlisted specimens that are known to
be edible; all animals will be considered edible, and any associated eggs. Monthly
availability of each food item will then be defined using the relevant biological data –
this will provide a monthly record of every plant and animal food item available in
each Landscape Division.
7
The study area has saline groundwater and no freshwater surface drainage, so the
only potable water is rainwater, temporarily preserved in soils and rock structures. As
a result, it is important to consider temporal variation not only within a single year,
but between years. Therefore, water availability will be considered during low,
average and high winter rainfall years, defined with reference to modern rainfall data
from Hyden (BoM station no. 010568). Soil water will be evaluated using the Soil
Water Tool (Department of Primary Industries and Regional Development 2016a),
which was designed to model the water-holding period of various agricultural soils
under different conditions, using local rainfall data. A representative soil profile will
be chosen for each Landscape Division, based on generalised soil characteristics
outlined in the literature (Beard 1979, 1990; Department of Primary Industries and
Regional Development 2016b; Murphy-White 2007; Schoknecht and Pathan 2013).
Rock structures (pans, gnammas) will be modelled manually, using local rainfall data
in conjunction with evaporation and runoff coefficients; this will produce a daily water
balance for different structures under various rainfall conditions.
The distribution of critical resources undoubtedly places constraints on when and
how intensively people can occupy certain parts of the landscape. Considering this,
Optimal Foraging Theory (OFT) provides a number of rules that quantify the optimal
response to varying resource distributions. The OFT family of models considers
which resources should be targeted first, when diets should diversify, and which
resources can be profitably transported longer distances (Macarthur and Pianka
1966; Orians and Pearson 1979). Similarly, they consider which foraging patches
(represented here by Landscape Divisions) should be searched first, and when
groups should move to a new patch (Charnov 1976; Macarthur and Pianka 1966). To
use these rules, plant foods will be ranked by their return rates (cal./hr), based on
published data. Animal foods will not be ranked as the calculations are more
complex for mobile prey, and the data are unavailable. Indeed, fixed resources are
more likely to dictate occupation patterns (Bird 1985:131, 140). The spatio-temporal
availability of differently ranked resources, as well as unranked resources and water,
will be considered in light of the OFT tenets, to determine the optimal seasonal
occupation pattern for the study area.
8
Method-specific limitations are evaluated in Chapter 5, but it is important to identify
the limitations and assumptions that impact the resource and occupation model at a
broader scale. These are listed below:
1. Landscape Divisions are presumed to have consistent water-holding
properties and plant/animal foods resources throughout the study area.
However, it is conceivable that this may vary at a smaller scale, and that
each iteration of a particular Division may differ in their resource base;
2. The effect of annual rainfall on plant and animal foods cannot be
considered, as the relevant data are not available. However, foods were
unlikely to have been equally abundant under all rainfall conditions;
3. The model will consider the distribution of critical resources as the sole
factor dictating occupation patterns. Shelter requirements are not
considered, since thick vegetation, rockshelters or hollowed boulders
provided natural protection from the elements. Further, shelters could be
constructed from bark, branches and other plant matter – Roe (1836) noted
the remnants of many such structures in the inland southwest;
4. Human landscape management/modification is not considered. The most
widespread technique was firing, which can be used to increase access to
prey and to promote plant growth (Bird et al. 2005, 2008, 2013; Gould
1971b; Hallam 1975, 1989; Jones 1969). However, the inland southwest
was fired less frequently than coastal regions, and the vegetation
communities characteristic of the study area were infrequently burnt when
encountered in other parts of the southwest with more intensive firing
regimes (Abbott 2003; Lullfitz et al. 2017; Prober et al. 2016). Therefore,
fire regimes are unlikely to have significantly impacted the density or
distribution of resources in the study area;
5. Population density and territory size will not be considered, as the data are
lacking. The study area is assumed to represent a distinct territory
occupied by a single group, and the optimal seasonal route is defined with
reference to this area only. It is somewhat misleading to consider an area
in isolation, since the group may have had access to some Noongar land
outside the boundary. Similarly, the study area may have supported more
than one group, each with distinct portions of land.
9
1.4.2 Archaeological Investigations
This study is distinct from a regional analysis (e.g. Attenbrow 2004; McBryde 1974)
as it must necessarily omit a defining feature of these: extensive site survey of the
study area or a representative sample of it (McBryde 1986; White 2011). This cannot
be achieved for the study area, since most of the land has been cleared for
agriculture and is privately owned. Known archaeological site locations are therefore
skewed towards those areas of the landscape that are less affected by such factors
– granite outcrops and salt lakes. Nevertheless, it is still important to consider the
archaeological record from a number of sites to construct a meaningful occupation
history, as:
Any sequence based on the evidence of one site, however rich the site,
may be unreliable, due to local specialized features of that site not readily
recognized on first investigation. Variations may be found from site to site,
not only in the quantity of material left behind by their prehistoric
occupants, but also in the types of implements represented.
Assemblages from sites of approximately the same geographical area
may differ in their components...It is only on the basis of excavated,
stratified, assemblages from a number of sites within the same region
that a sound reconstruction of the cultural sequence for that region’s
prehistory may be made’ (McBryde 1974:16–17).
Therefore, this research will consider archaeological remains from several sites.
Those from Anderson Rocks, Gibb Rock and Mulka's Cave will be considered in
detail, as these granite outcrops were the focus of the author's excavations and
surface collections (Figure 1.2). Another outcrop, Twine Reserve, will be considered
more briefly, for reasons outlined in Chapter 9.2.1; discussion will be limited to
observations made by the author during a site visit, and their relevance within in the
context of the occupation model. Where sufficient details are available, other
archaeological remains from the study area will be considered in conjunction with the
model and the author's data.
Stone artefacts from Anderson Rocks, Gibb Rock and Mulka's Cave are classified
using a materialist approach, based on Hiscock (2007). In contrast to typological
systems that rely on artefact form, and often inferred function, a materialist approach
is less subjective as it classifies artefacts based on their position within the reduction
sequence. This places more importance on artefact attributes than the frequency of
different artefact types. To address questions of mobility and occupation patterns,
10
assemblage data will be considered with references to Kuhn's (1995) provisioning
systems. The strength of this framework lies in the integration of mobility, raw
material availability and risk into a single model – one that also follows the tenets of
ecological/forager theory. Kuhn (1995:20) accepts that people move to access
critical resources (food and water) but, following Bamforth (1986), acknowledges that
this changes spatial relationship between people and stone sources, which are rarely
truly ubiquitous.
Kuhn (1995) proposed two contrasting solutions to ensure that technological needs
were consistently met: Individual Provisioning and Place Provisioning. Individual
Provisioning arises under increased residential mobility, when unpredictable access
to replacement stone is managed by the manufacture of a portable, multifunctional
toolkit. This kit is heavily maintained and recycled; artefacts are only discarded when
they cannot be repaired or recycled. In contrast, Place Provisioning involves a
particular site being furnished with toolmaking potential, in the form of raw material
stockpiles. Artefacts are produced in a more expedient manner, and discarded after
use; when another task arises, new flakes are struck. This strategy is only optimal
when residential mobility is decreased, and the duration and predictability of
occupation increases. Each strategy produces distinct assemblages due to
differences in raw material requirements, discard behaviour and reduction
sequences. Occupation intensity is considered with reference to these provisioning
systems, as well as other commonly employed proxies such as reduction intensity,
artefact discard rates, and post-depositional breakage.
The main limitations in the archaeological investigations relate to the sample,
specifically the number of sites investigated, the size and distribution of pits, and how
representative these samples are at a broader scale; the issue is not unique to this
research, as all archaeological excavation involves some level of sampling. Three
sites were targeted for more detailed investigation. While an occupation history
would ideally derive from a larger sample, this is partially addressed by investigating
different areas of each site. This increases the number of comparative assemblages
available and allows the antiquity of occupation to be defined based on several
different dated deposits. However, the author's pits are small, 500 x 500 mm, to
permit different site areas to be investigated while still remaining within the one
11
square metre excavation limit imposed by Section 16 of the Aboriginal Heritage Act
(1972). Small pits are more susceptible to sampling biases, particularly in terms of
less frequently encountered artefact types (e.g. retouched pieces). It is therefore
necessary to consider the possibility that such pieces were not truly absent, but
simply outside the pit boundary. Nevertheless, Hiscock's (2007) and Kuhn's (1995)
approaches both rely more heavily on artefact attributes than types, so small pit size
should have a minor impact. It is probably more influential in terms of yielding
smaller overall assemblages, in which certain attribute patterns may be harder to
identify.
1.5 SIGNIFICANCE OF THIS RESEARCH
This study is significant as it aims to address the knowledge gap for inland
southwestern Australia, typified by the lack of excavated and dated sites compared
to the other parts of the country (Williams et al. 2014). At present, there are
insufficient archaeological or ethnohistorical data to indicate when and how the study
area was used by Aboriginal people; these questions will be directly addressed by
this research. The study area may be of additional interest due to its location in
something of an intermediate area climatically, environmentally and culturally – it lies
between the well-watered coastal southwest and the arid interior, and adjoins the
boundary separating southwestern Australia from the Eastern Goldfields, and the
Noongar from desert groups. As a result, the study area should preserve its own
unique occupation history, unlike those from the coastal southwest or Western
Desert, where data are more abundant. The occupation history will allow the Mulka's
Cave assemblage to be interpreted, so the site can be placed in its regional context.
This is particularly important since it is the most profusely decorated Noongar rock
art site, and receives more than 80,000 visitors annually, partly due to its vicinity to
Wave Rock (Gunn 2004, 2006). The theoretical occupation model will allow
fragmentary archaeological remains or particular site types (e.g. granite outcrops) to
be interpreted, and predict the type of occupation sequences that might be found in
different parts of the study area. These results may not be applicable to the entire
inland southwest, but will certainly be of some broader relevance beyond all but the
eastern boundary of the study area.
12
While the importance of increasing the body of archaeological data for the inland
southwest cannot be overstated, the significance of this research is also in the
approach itself, particularly with regard to quantifying resource distribution. This is
not in itself a novel idea, but the scale and integration of analysis is noteworthy. Here,
the theoretical occupation model and archaeological assemblages are treated as
complementary data sources, with neither taking precedence over the other. Too
often, resource data are considered supplementary to archaeological findings, and
provide background or contextual information rather than being fully integrated into
analyses (e.g. Bird 1985; Smith 1993). Plant and animal food lists are generally brief,
incomplete, and compiled for large regions; as a result, these underestimate
resource diversity and spatial variability. In contrast, the author aims to compile an
extensive list of all plant and animal foods likely present within each Landscape
Division, and to define their availability on a monthly basis. Further, the focus is on
organism-specific data that dictates when an edible product was available, rather
than ethnohistorical records detailing the season that particular food sources were
used. The latter may be unreliable where items were available for longer periods, or
year-round, since season of use will be partially dictated by the relative abundance
of other food products. This method will provide a more complete, less subjective
record of food resources in the study area, while also addressing spatio-temporal
variation at a finer scale than previously attempted in southwestern Australia.
As with food resources, water availability is often considered more broadly, by simply
noting the types of sources that may have been used during certain seasons (e.g.
Bird 1985:176). Ephemeral water sources are rarely considered in detail, probably
because they were less integral to human survival, as most areas provide potable
groundwater, or freshwater lakes or swamps. In the study area, however, this is not
the case. As a result, new methods are devised to quantify the longevity of rainwater
stored in soil profiles and rock structures under low, average and high rainfall
conditions. These data will provide a high-resolution representation of how water
availability varied over space and time, and indicate how this would have constrained
human occupation. The author is not aware of any previous attempts to quantify
ephemeral water sources at all, let alone in this manner; it may be of use in other
areas where water is similarly limited.
13
1.6 ORGANISATION OF THESIS
Section One (Chapters 1–4) provides the background and contextual information of
relevance to this research. Chapter 1 has outlined the research bias in the Noongar
country, southwestern Australia, which has resulted in little knowledge on when or
how the inland environs were used. It detailed the research questions formulated,
and the methods devised to address this knowledge gap, and the significance of
doing so. Chapter 2 details the theoretical background, focussing on forager theory
and
technological
organisation,
which
are
united
through
Kuhn's
(1995)
technological provisioning systems. Chapter 3 focusses on the natural environment
of the study area. The geology and hydrology are described, noting the impact of
clearing on hydrological conditions. The Quaternary and modern climate is evaluated,
with particular focus on the conflicting nature of the Holocene palaeoenvironmental
records for southwestern Australia. Finally, the vegetation systems and fauna
distribution are described. Chapter 4 describes the cultural context of the study area
and beyond. The earliest archaeological evidence in Australia is briefly evaluated,
followed by a discussion of the archaeological locations in the study area, and the
few relevant studies from further afield. The fragmentary ethnohistorical record is
then evaluated.
Section Two (Chapters 5–8) describes the various facets of the resource and
occupation model. Chapter 5 details the methods used to quantify resource
availability and formulate the model. Six Landscape Divisions are defined to address
spatial variation within the study area. Discussion then turns methods of quantifying
food and water availability in each – while the former relies on biodiversity surveys
and published data, the latter involves some more unique methods, including the use
of an agricultural model. The means of converting these data into an occupation
model are then outlined. Chapter 6–7 provide the descriptive results; the former
outlines the spatio-temporal availability of plant and animal foods, while the latter
details the longevity of ephemeral water sources under low, average and high rainfall
conditions. Chapter 8 presents the occupation model, based on these data. It defines
the optimal seasonal route that the study area occupants may have followed, and
notes which technological provisioning systems should characterise certain sites or
occupation periods. It also considers whether aggregation events could have
occurred in the study area, and where and when these might have been held.
14
Section Three (Chapters 9–13) focusses on the archaeological investigations.
Chapter 9 outlines the methodology, beginning with field methods, before outlining
the materialist approach used to analyse the quartz-dominated assemblage, and the
reasons for its selection. Artefact attributes are then discussed, and archaeological
indicators formulated for each technological provisioning system, as well as different
levels of occupation intensity. Chapters 10–12 focus on each of the excavated sites
– Anderson Rocks (Chapter 10), Gibb Rock (Chapter 11) and Mulka's Cave (Chapter
12). The sites are described, and their specific water resources quantified. The
organics and any radiocarbon dates are discussed, before the stone artefact
assemblages, and the attributes collected from them, are detailed. Other cultural
material, such as ochre or flaked glass, are then briefly considered. Chapter 13
considers the assemblage from each site in Kuhn's (1995) framework, to identify the
technological provisioning systems represented within each pit or surface collection,
allowing for temporal shifts within the former.
Section Four (Chapters 14–15) unites the theoretical occupation model with the
archaeological data. Chapter 14 considers how Anderson Rocks, Gibb Rock and
Mulka's Cave were used, with reference to site function and overall mobility patterns.
Where possible, changing intensity over time is then considered, and linked to shifts
in water availability. Finally, Chapter 15 summarises the research to provide an
occupation history for the study area. Each of the four main research questions is
answered and the broader implications of those results considered. Future research
directions are then identified, before the concluding remarks are made.
15
CHAPTER 2
THEORETICAL FRAMEWORK: MOBILITY, SUBSISTENCE AND TECHNOLOGY
2.1 INTRODUCTION
This chapter discusses the theoretical framework of this study, outlining how
occupation patterns are influenced by ecological conditions, and how such patterns
may be detected within the archaeological record. Forager models are briefly
discussed, as they provide solid predictions about how people should move around
the landscape and alter their subsistence strategies and settlement patterns to
optimise returns under varying resource distribution. Technological organisation is
then evaluated, focussing on how behavioural and physical attributes of technology
can be affected by mobility, raw material availability, and risk. Finally, Kuhn’s (1995)
provisioning systems are described; this concept unites forager theory with
technology by predicting the different ways that people can ensure their
technological needs are met under various conditions.
2.2 FORAGER THEORY
Forager Theory falls under the umbrella of Human Behavioural Ecology (HBE), a
field that explains human behavioural diversity using models grounded in
evolutionary ecology (Winterhalder and Smith 2000). While not the sole determining
factor, the environment is a major element constraining human occupation – HBE
considers how human behaviour can be affected by various environmental
conditions (Nettle et al. 2013). A major subset of HBE is Optimal Foraging Theory
(OFT), a family of models that relate the resource base to certain behaviours or
choices, with the overall goal of optimisation (utility increase), i.e. to maximise
returns (Kelly 2013:33–34; Mithen 1989). Foraging efficiency is often considered to
be a proxy for reproductive fitness so, following evolutionary theory, optimal foraging
should be the goal, whether conscious or not (Kelly 2013:34; Smith 1979). There is a
vast body of research dedicated to HBE and forager theory, so only relevant
contributions are discussed further; more comprehensive treatments can be found in
Bettinger et al. (2015), Kelly (2013), Nettle et al. (2013) and Winterhalder (1981).
16
Individual OFT models may identify which available resources should be pursued,
the best places for foraging, when resources should be processed before transport,
and when groups should move to new foraging grounds. HBE and OFT have their
roots in ecological science, so many models were initially developed for animals (e.g.
Charnov 1976; Emlen 1966; Horn 1968; MacArthur & Pianka 1966; Orians and
Pearson 1979), and subsequently modified for, or applied to, hunter-gatherer
societies (e.g. Dyson-Hudson and Smith 1978; Harpending and Davis 1977; Hawkes
& O'Connell 1992; Wilmsen 1973). Individual models are often complex
mathematical computations (Kelly 2013:46), involving the following:
o
Goal: generally, to maximise caloric energy return for time spent foraging
(including search, handling, transport and processing);
o
Currency: calories, or energy;
o
Constraints: e.g. the amount of time that can be dedicated to foraging;
o
Options: various potential food sources or foraging patches.
Applying a foraging model requires intimate knowledge of the landscape and the
various resources found therein – their nutritional value, handling time, transport
costs and distribution. That such detailed knowledge is required to generate arguably
simple models is a common criticism of forager theory (Bettinger et al. 2015:133;
Grøn 2012; Smith 1983). Models also do not allow for choices unrelated to energy
requirements or costs, for example items selected or omitted due to preference or
taboo, but as Kelly (2013:76) noted, these models are not intended to replicate
reality, but to model expected outcomes if certain conditions are met.
While detailed knowledge is required to apply individual forager models, their
underlying rules (Table 2.1) are robust enough to be successfully applied in isolation,
or incorporated into broader analyses (e.g. Clarkson 2004, 2007). Ambrose and
Lorenz (1990) used OFT models as the foundation of their resource structure
scheme. They characterised resources according to their predictability and
abundance, creating four extremes of the resource-structure continuum: predictable
and dense (e.g. well-watered woodlands), predictable and scarce (e.g. semi-arid
woodland, scrub), unpredictable and dense (e.g. grasslands with seasonal herds),
unpredictable and scarce (e.g. deserts). Using several OFT models, they
characterised residential mobility, settlement patterns, and social organisation for
17
groups inhabiting each distinct resource zone (Figure 2.1; Table 2.2). While Ambrose
and Lorenz (1990) merely summarised findings from existing forager models, their
data provide a useful synthesis while also permitting OFT models to be more widely
applied by relating predictions from individual models to specific resource structures.
Table 2.1 Forager models and theories for achieving optimisation, based on the references listed and
summary data from Bettinger et al. (2015), Clarkson (2007), Kelly (2013) and Winterhalder (1981).
References denote the first iteration of a model or a significant subsequent contribution.
Model
name
/subject
Method to achieve optimisation
Reference
Dietbreadth
model
Foods are ranked according to returns (energy value for time spent
procuring and processing). Higher ranked foods are pursued first,
and lower ranked resources are utilised when higher ranked items
are unavailable or used to supplement the diet when higher ranked
items are less abundant.
Macarthur and
Pianka 1966
Patch
choice
model
Patches (areas of land that contain resources, which are distributed
unequally across landscape) are ranked according to return rates.
Patches with higher returns are searched first, moving to lower
ranked patches as higher ones are depleted.
Macarthur and
Pianka 1966
Group size
and
dispersion
Where resources are stable and evenly distributed, and variance in
returns is low, foragers should be dispersed in small units; where
resources are mobile and clumped (e.g. greater spatial variation in
returns), people should be aggregated in a larger group. Residential
mobility increases as resources become scarcer and less
predictable. Moves will be opportunistic where resources are
unpredictable and scheduled where they are predictable.
Harpending and
Davis 1977;
Horn 1968;
Wilmsen 1973
Marginal
Value
Theorem
Patches are not exploited to exhaustion. People will move to a fresh
patch when the return rate in the current patch reaches the average
for all patches, allowing for travel time. Patches are abandoned
sooner when the environment is of better quality overall. This model
may not hold true when occupation is tethered to certain resources,
e.g. a stable water source.
Charnov 1976
Territory
size / range
Range (e.g. territory size) should be greater in poorer environments,
where resources are scarcer or patchily distributed. Territories will
be smaller where resources are dense and more evenly distributed.
Harpending and
Davis 1977
Economic
Where resources are dense and predictable, territories are small, so
Defensibility the cost of defending territory is exceeded by benefits. Elsewhere
Model
costs are too great, leading to passive territories (e.g. undefended,
semi-permeable) or shifting territories/high mobility.
Central
Place
Foraging
Model
Dyson-Hudson &
Smith 1978
Allows for food being transported to a central place rather than being Orians and
consumed at the point of extraction. Unprocessed resources should Pearson 1979
be transported to camp until the return rate (including travel time) is
greater for processed vs unprocessed foods – processing will
become more profitable as transport distance increases. Distance to
camp determines what resources are worth collecting, e.g. lowranked foods can't be transported far as costs outweigh returns, so
the range of viable resources narrows with distance from camp.
18
Figure 2.1 Optimal and suboptimal strategies for different resource structures (Ambrose and Lorenz
1990:15).
Table 2.2 Behavioural characteristics arising from different resource structures. Re-tabulated from
Ambrose and Lorenz (1990:10).
Resource
structure
Territorial
strategy
Information
exchange
Residential
mobility
Group
size
Population
density
Diet
breadth
A: predictable
and dense
Territorial
defence
Low
Low,
scheduled
Small
High
Moderate
B: predictable
and scarce
Home
range, semipermeable
Undefended,
very
permeable
Undefended,
very
permeable
Medium
Medium,
scheduled
Small
Medium
High
Very high
High,
opportunistic
Large
Medium
Very low
High
Very high,
opportunistic
Very
small
Very low
Very high
C: unpredictable
and dense
D: unpredictable
and scarce
19
2.3 TECHNOLOGICAL ORGANISATION
Technological organisation is 'the study of the selection and integration of strategies
for making, using, transporting, and discarding tools and the materials needed for
their manufacture and maintenance' (Nelson 1991:57). Previously, technological
variation had been attributed to stylistic or functional concerns (see discussion in
Rolland and Dibble 1990), but it was later recognised that technology had the power
to reveal elements of human behaviour, as technology responds to social and
economic factors. Hence, the ‘need to acquire resources in different locations, to
move around the landscape, to remain settled at a place, to transport different kinds
of resources and material needs, and many other variables condition the
technological strategies employed at a particular time and place' (Nelson 1991:88)
Technology, then, can be regarded as a problem-solving strategy and a physical
manifestation of how people interact with the environment (Kuhn 1995:19). As such,
it is vital to integrate technology into theoretical concerns, not only because stone
artefacts dominate archaeological assemblages, but to generate a more complete
picture of past human behaviour (Torrence 1983, 1989).
Several factors influence technological organisation, including specific functional
requirements (e.g. Ferris 2015), manufacture costs and resource returns/tool
efficiency (e.g. Clarkson et al. 2015; Garvey 2015; Kuhn and Miller 2015), raw
material availability (e.g. Andrefsky 1994; Bamforth 1986; Manninen and Knutsson
2014), demographic changes (e.g. Henrich 2004; Shennan 2001), mobility (e.g.
Hiscock 1996; Parry and Kelly 1987; Shott 1986) and risk (e.g. Collard et al. 2005,
2013; Hiscock 1994; Torrence 1983, 1989, 2001). The resulting variation can be
behavioural – relating to the timing or location of stone-working activities – or
physical, relating to the design of individual artefacts, the composition of
assemblages, or the reduction techniques used to produce them (e.g. Binford 1979;
Hiscock 1994; Nelson 1991; Parry and Kelly 1987; Shott 1986; Torrence 1983, 1989,
2001). A thorough discussion of these contributing factors is well outside the scope
of this thesis, so the focus here is centred on three major factors: mobility, raw
material availability, and risk. Arguments predicated on assemblage diversity or
implement design (e.g. Bleed 1986; Collard et al. 2005, 2013; Shott 1986; Torrence
1983, 1989, 2001) are not considered below, as the principles are difficult to apply to
amorphous, debris-dominated assemblages, such as those frequently encountered
20
in Australia (Hiscock and Clarkson 2000; Kuhn 1995:19–20). Instead, the discussion
is limited to those case studies yielding results relevant within an Australian context.
2.3.1 Mobility
Generally, two types of mobility are recognised: residential and logistical. Residential
mobility occurs when an entire group moves to a new location to exploit particular
resources, while logistical mobility involves small task groups leaving a central
location to procure resources from further afield, before transporting them back for
the entire group to consume (Binford 1980; Kelly 2013:78–79). As discussed above,
the magnitude of residential mobility is dictated by the abundance and predictability
of resources, while increased logistical mobility can arise when residential mobility is
impeded, often due to the incongruent or patchy distribution of critical resources – for
example where permanent water is present but food is scarce (Binford 1980).
Neither residential nor logistical mobility, however, should be treated as a single
variable, as each comprises several parameters, including frequency of moves,
distance per move and annual distance travelled (Shott 1986). Discussion here
focusses on the relationship between residential mobility and stone reduction
techniques – specifically standardised versus unstandardised techniques, and the
use of bipolar flaking, as these features are easily identified in Australian stone
artefact assemblages.
Parry and Kelly (1987) relied heavily on Binford's (1973, 1977) concept of curated
and expedient technology to define their standardised and unstandardised reduction
techniques, but directly relate the use of these techniques to the level of residential
mobility. They note that unstandardised (expedient) technologies require minimal
time, effort or skill to execute, as no attempt is made to control flake form. Instead,
the most suitable piece for a particular task is chosen from the flaking products;
pieces are not reworked, conserved or retained after initial use. Unstandardised
techniques consume a lot of stone, so raw material selection is dictated by what is
locally available in sufficient quantity, though bipolar flaking may be used to conserve
material where it is less abundant (Parry and Kelly 1987). In contrast, standardised
(curated) technologies require considerable time, energy and skill to execute, as well
as good quality raw material that is used in a much more efficient manner. Tools are
portable, formal and multifunctional, and are repaired and reused many times.
21
However, standardised technologies are much costlier to manufacture, maintain and
even use, since generalised tools are less efficient than specialised tools and
retouched edges are blunter than fresh ones (Parry and Kelly 1987; Torrence 1983).
In their archaeological example, Parry and Kelly (1987) identified a shift to
standardised reduction that could not be related to local conditions – including the
availability of raw materials – but instead coincided with the occurrence of large,
permanent villages, indicating a shift to a sedentary lifestyle (Parry and Kelly 1987).
They argued that for highly mobile groups, the greater cost of standardised
technologies is outweighed by the benefits conferred through the portability of items,
conservation of raw material (since access may be limited) and the likelihood of most
tasks being achievable with a multifunctional toolkit. Where mobility is reduced, the
cost of standardised technologies is too high and many of the benefits evident under
greater mobility are not required (e.g. portability) – therefore, expedient technologies
are produced (Parry and Kelly 1987). These expedient technologies may also be
associated with greater mobility where raw material is ubiquitous and abundant,
negating the transport of tools between residences (Bamforth 1986; Parry and Kelly
1987).
Hiscock (1996) also evaluated the influence of residential mobility on technological
organisation, adopting a narrower focus than Parry and Kelly (1987), by linking the
level of bipolar reduction to residential mobility. Parry and Kelly (1987) noted that
bipolar reduction was evident in their expedient assemblages – which reflected more
sedentary groups – but only linked it to situations of reduced raw material availability.
However, Hiscock (1996) considered that conservation may arise independent of
raw material availability and increase in concert with occupation duration. He studied
several assemblages from the South Alligator River area of the Northern Territory,
focussing on sites in woodland areas and comparing them to those on the floodplain
margin. At woodland sites, bipolar cores became more common as the distance to
raw material sources increased, indicating the type of raw material conservation
noted elsewhere (Bamforth 1986 – see Chapter 2.3.2). However, all floodplain sites
were at, or extremely close to, raw material sources, yet bipolar cores were more
than twice as frequent there than at any woodland sites, even those furthest from
raw material sources (Hiscock 1996: Figure 1). Therefore, variation in bipolar
frequency could not be explained with reference to source location alone. Instead,
22
Hiscock (1996) noted that bipolar knapping often occurs near the end of a reduction
sequence, when a core would otherwise be abandoned due to low inertia; it would
be beneficial for less mobile groups to extend core use-lives, and thereby reduce raw
material procurement and transport costs. Therefore, he concluded that the
floodplain sites, characterised by the greater frequency of bipolar cores, represented
longer occupations (Hiscock 1996).
It may seem unnecessary for groups to conserve stone when sources are nearby,
but the benefits are evident when considered in the context of utility increase. Since
foragers seek to maximise returns while minimising time and energy expenditure,
there would be little benefit to incurring the costs associated with stone procurement
– minimal as they may be – when existing supplies could be further reduced using a
different technique, one that requires little time or skill to execute and produces more
cutting edge per core than freehand flaking (Pargeter and de la Pena 2017; Pargeter
and Elen 2017; Parry and Kelly 1987).
2.3.2 Raw Material Availability
Binford's (1979) concept of embedded procurement arose from his fieldwork in
Alaska and Central Australia, where he observed that people rarely made trips
specifically to procure stone. Instead, raw material collection was embedded within
subsistence tasks – people brought home stone when foraging or hunting returns
were low to maximise transport potential. Binford (1979) argued that embedded
procurement could drastically reduce, and even entirely void, direct procurement
costs. As stone collection occurred during regular food gathering activities, Binford
(1979) posited that raw materials used for tool manufacture are, therefore, a direct
reflection of the foraging range accessed from a particular site. Garvey (2015) noted
that the widespread acceptance of Binford's (1979) concept places more importance
on the means of conserving stone, once procured, than on wholly evaluating the
mechanism of procurement itself. Regardless of whether material is procured directly
or via an embedded strategy, access to raw material can impact technological
organisation independent of other factors. Here, the focus is on the distribution of
raw material sources and the physical properties of the stone itself.
23
The distribution of raw materials can influence whether they are included within the
technological system and, if they are, how intensively they are reduced. Bamforth
(1986) found that where a single raw material source was utilised, periods of greater
residential mobility encouraged higher levels of recycling and maintenance. He
argued, however, that this was not because mobility inherently required the use of a
more formal toolkit, but because changing mobility restricted access to the stone
source, and therefore encouraged conservation; conversely, when access improved,
there was less need to recycle and maintain tools. Where multiple sources existed –
at varying distances from an occupation site – Bamforth (1986) found that local,
lower quality materials were retouched less and were frequently discarded before
they were broken, while higher quality material from more distant sources was
subjected to increased levels of maintenance and recycling. In addition, pieces were
more likely to be broken before being discarded. If this high-quality material was
collected during subsistence activities (embedded procurement) it may reflect a
portion of territory that was visited long ago, or where the next visitation date is
uncertain (Clarkson 2007:25). However, reduction intensity may also be related to
transport and procurement costs if the material was targeted directly. More distant
materials incur greater procurement and transport costs, so the resultant artefacts
must be used for longer periods (and therefore retouched, maintained and recycled)
to offset costs and confer some benefit over using lower quality, more local materials
(Garvey 2015; Kuhn and Miller 2015). Where shifts in mobility bring people closer to
a higher quality source, however, procurement costs are reduced. This may make it
profitable to use the higher quality material or, where it was already in use, permit
less intensive reduction and more frequent artefact replacement (Garvey 2015).
It is also important to consider the physical properties of various raw materials,
particularly those that are local and abundant, as they are more likely to dictate
technological decisions. Not all raw materials are equally suited to artefact
manufacture – higher quality materials are generally more homogenous and isotropic,
so they not only flake more predictably than lower quality stone but are more suited
to reworking and retouching, and may retain their sharp edges for longer (Bleed
1986; Cotterell and Kamminga 1979; Whittaker 1994:65–66).
24
Andrefsky (1994) devised a framework for predicting how the abundance and quality
of local stone impacts technological organisation, specifically the use of formal or
informal tool production (Figure 2.2). He found that where local stone was abundant
it dominated assemblages, regardless of the quality of the material; however, the
technological system adopted varied in response to the quality of the stone. Where it
was abundant, high-quality material was reduced both formally and informally,
though low-quality stone was always associated with informal reduction, often
supplemented with small quantities of formal tools made of higher quality material
(Andrefsky 1994). These results were echoed by Manninen and Knutsson (2014),
who identified an extreme case whereby local stone was so abundant but of such a
poor quality that the entire technological system was simplified to incorporate it,
rather than rationing high-quality stone as expected (Andrefsky 1994; Bamforth
1986). Andrefsky (1994) found similar levels of formal and informal reduction in both
short- and long-term occupations (i.e. higher and lower residential mobility), where
the raw material source was constant. Elsewhere, he found evidence for informal
technology associated with higher levels of residential mobility since, as others had
argued, this strategy could be adopted where raw material availability permitted
(Bamforth 1986; Parry and Kelly 1987). Similarly, Manninen and Knutsson (2014)
identified that their observed switch to informal technology was associated with
increased mobility, which took people further from the source of good-quality raw
material. These results appear to address Bamforth's (1986) question of why people
would make and transport formal tools if material was abundant by answering, ‘they
didn’t’.
25
Figure 2.2 The relationship between quality and abundance of local stone, and the use of formal and
informal tool production (Andrefsky 1994:30).
2.3.3 Risk
In the context of subsistence, risk relates to the probability of failing to procure
resources, or not meeting dietary requirements (Hiscock 1994; Torrence 1989). Risk
can vary in timing and severity (the cost of failure), which can be relatively minor,
whereby additional time or energy costs are incurred, or severe enough that it
endangers the lives of individuals. Risk can be caused by several factors, including
uncertainty of resource distribution (when people move into unfamiliar areas),
climatic shifts, resource depletion/scarcity/mobility, and the limited spatio-temporal
availability of prey (Hiscock 1994; Torrence 1989). Torrence (1989, 2001) argued
that risk is greatest where diet breadth is lowest, especially where people rely on a
few species of large game that are highly mobile and only available during a certain
season. While this particular scenario is not relevant to Australian research, other
risk factors would certainly have been influential. To a certain extent, risk can be
minimised by utility increase, i.e. altering diet-breadth, mobility and aggregation
patterns to emulate the optimal solution for local conditions (see Chapter 2.2;
Clarkson 2007:12–14; Gould 1991; Hiscock 1994). However, technology can also
26
reduce risk through both behavioural and physical characteristics – technological
activities can be scheduled to reduce time/energy costs and ensure that tools are
ready for use when required, and risk can also be mitigated by the physical attributes
of artefacts themselves; these strategies are discussed in more detail below.
Initially an independent concept, Torrence's (1983) time-budgeting model was later
subsumed into her broader risk reduction concept (Torrence 1989, 2001). Her
central idea was that, to reduce risk, people must schedule technological tasks so
they do not interfere (either through time or energy consumption) with more
important subsistence tasks, while simultaneously ensuring that tools are ready for
use. These technological tasks can incorporate raw material procurement as well as
artefact manufacture, maintenance and repair. She assumed that raw materials were
collected via embedded procurement, since Binford (1979) demonstrated that this
was the lowest cost method of procurement, but this requires that usable stone is
found within the foraging range. Manufacture and maintenance tasks should be
performed during down-time, either after dark or during other activities, for example
while waiting for game (Binford 1977, 1979). Where risk is more frequent or the cost
of failure is greater, manufacture and maintenance should be anticipatory to ensure
that tools are ready before they are required; when risk is more limited, tools can be
made or repaired as they are needed (Torrence 1989, 2001). As Clarkson (2007:15)
summarised, the archaeological implications for risk reduction should be that raw
materials reflect foraging range, and that more complex assemblages – representing
manufacture, maintenance and discard – would occur at residential sites, where
down-time is predictable and abundant.
As noted earlier, design elements can be difficult to observe in Australian lithic
assemblages that are often dominated by debris and other amorphous material.
However, Hiscock (1994, 2002a) related the mid-Holocene introduction of the
Australian Small Tool Tradition (ASTT) – identified by its small, formal, highly
standardised implements – to risk reduction required by resource uncertainty that
arose due to climatic shifts, possibly related to El Niño Southern Oscillation (ENSO),
and/or people moving into unfamiliar parts of the landscape. Despite having
generalised knowledge of the wider area, for example the types of plant and animal
foods and stone that may be available, people would be unfamiliar with local
27
resource distribution, so would have trouble predicting when artefacts may be
required and when/where raw materials may be encountered (Hiscock 1994).
Torrence (1989, 2001) noted that well-organised technology is required to
successfully reduce risk. ASTT implements certainly conform to this definition and
provide clear advantages to people wishing to reduce risk. The small, standardised
tool forms such as points, backed artefacts and tulas – each of which has a distinct
geographical distribution (Hiscock 1994: Figure 2) – have long been argued to be the
only surviving components of composite tools used for a variety of tasks (Fullagar et
al. 2009; Kamminga 1980; Robertson 2009; Robertson et al. 2009). These tools
were portable, multifunctional, reliable, maintainable, and had redundant features
built in to guard against component failure, making them advantageous under risky
conditions that may promote increased mobility (Bleed 1986; Hiscock 1994, 2002a;
Shott 1986; Torrence 1983, 1989, 2001). Standardisation reduces repair time,
conserves raw material and normalises performance, desirable characteristics where
risk is heightened (Clarkson 2007:18–19; Hiscock 2006). Some regions exhibit a
decline in these formal tools during the late Holocene – Hiscock (1994, 2002a)
attributed this to less risky conditions, as effective precipitation had possibly
increased, and resource structures were well-known. Groups were possibly less
mobile, so portable, multifunctional toolkits were no longer required, although some
formal elements were retained where they were functionally necessary (Hiscock
1994).
2.3.4 Summary
The discussion above indicates that several factors influence technological
organisation, and that their effects can be manifested in various ways. Further, the
same technological strategy can arise under different conditions. For example,
expedient technologies may arise as a result of reduced mobility, abundant raw
material, or low-risk conditions. Similarly, each of these factors is interrelated –
mobility shifts may occur as a response to risk and subsequently alter access to raw
material sources. In each case study, the relevant author identified the factor that
they believed exerted the strongest influence over technological organisation when
related to their own dataset, but it is important to acknowledge that the dominant
influence may differ according to the unique characteristics of a site or area, i.e.
28
there is no single factor – whether it be mobility, raw material availability, or risk –
that will override all others in every scenario. It is therefore necessary to consider the
local conditions, including raw material distribution, mobility and subsistence
strategies, as well as risks that may arise through environmental change or altered
settlement patterns. All of these factors have the potential to influence technological
organisation, but the dominant mechanisms, as well as expressions of variation, will
differ over space and time.
2.4 TECHNOLOGICAL PROVISIONING SYSTEMS
Kuhn's (1995) provisioning model successfully united the two previously discussed
bodies of theory: forager models and technological organisation, incorporating the
major themes of mobility, raw material availability, and risk. Following ecological
theory, Kuhn accepted that mobility is primarily dictated by the abundance and
predictability of critical resources: food and water. Movement alters the location of
people relative to raw material sources – which are rarely truly ubiquitous – so
people may find themselves in an area with an abundance of food, but no usable
stone (Kuhn 1995:20). Therefore, Kuhn saw two ways for people to ensure that they
have access to the technology they need, when they need it: Individual Provisioning
involves people carrying their technological requirements with them in the form of
portable, multi-functional toolkits; in contrast, Place Provisioning involves furnishing a
specific location with toolmaking potential (raw material) or finished implements.
Following Binford (1977, 1979) and Torrence (1983), Kuhn (1995:25) believed that
manufacture and maintenance tasks occur during down-time, which is most
predictable and abundant at residential locations. As a result, the provisioning
strategy employed should be evident from the archaeological debris generated at
these sites. This in turn yields information on mobility, since the scale, frequency and
predictability of residential moves dictates the optimal technological provisioning
system (Figure 2.3a). The contrasting provisioning systems are discussed in more
detail below.
29
Figure 2.3 Impact of mobility on provisioning strategies: A (left): frequency of moves and occupation
duration; B (right): length of logistical forays. (Graf 2010, after Kuhn 1995).
2.4.1 Individual Provisioning
In situations of high residential mobility, often required where resources are scarce
or unpredictable (Ambrose and Lorenz 1990), the opportunity to reprovision with
stone may be poorly known or difficult to predict. In those cases, Kuhn (1995) argued
that individuals should provision themselves with their technological needs – in the
form of finished tools, blanks and small standardised cores – to guard against the
risk of being unprepared. However, this preparedness must be weighed against the
cost of constantly carrying around one's technological requirements. Transport costs
are arguably the overriding factor constraining technology for highly mobile groups,
so the toolkit must be portable but also versatile and multifunctional, since
specialised tools can rarely be accommodated (Gould 1969a:76; Kuhn 1995; Nelson
1991; Shott 1986). Since future access to raw material sources is uncertain, and
mobile toolkits leave little room for spares, the use-life of individual implements
should be extended as far as possible (Kuhn 1995). Items should be designed for
durability and be reworked and recycled when they become blunt or damaged, and
discarded only when they cannot be modified for continued use (Bleed 1986; Binford
1979; Kuhn 1995). Replacements must be manufactured from blanks, or flakes
detached from transported cores. The greater use-life of such tools favours the use
of a higher quality material, so stone types should be carefully selected, prioritising
those materials that flake predictably, retain sharp edges for longer, are more
resistant to breakage, and can be easily reworked (Bleed 1986; Garvey 2015;
Goodyear 1989; Graf 2010; Mackay 2005; Torrence 1989).
30
Invariably, unanticipated tasks will arise that cannot be completed with existing tools.
In these cases, tools must be made by modifying components of the toolkit or, where
possible, manufacturing more expedient tools from locally available raw material
(Kuhn 1995; Mackay 2005). Clarkson (2007:16) noted that this type of strategy will
be especially important under increased residential mobility, since frequent moves
constrain the diversity of transported toolkits (Shott 1986).
The above data permit a prediction regarding what sort of assemblages should be
generated at residential sites where Individual Provisioning was the dominant system.
Elements of the mobile toolkit have extended use-lives so should be infrequently
encountered in the archaeological record and – where they are present – show
evidence of extensive modification through retouch and reworking (Kuhn 1995;
Mackay 2005). While tools may be infrequently discarded, sites should still preserve
some traces of heavily modified pieces through debris generated by retouch,
reworking and converting blanks into implements (Kuhn 1995:26). This evidence
should comprise small flakes of good quality raw material. Some level of primary
reduction may also occur by the use of small, portable cores, also of higher quality
material – this flaking should be standardised and show attempts at raw material
conservation, through the use of blade cores or attempts at rejuvenation by
overhang removal (Graf 2010; Kuhn 1995; Mackay 2005; Mitchell 2016:159; Parry
and Kelly 1987). As with elements of mobile toolkits, cores themselves may be rare
within archaeological deposits (Mackay 2005), but evidence of their use should be
preserved in flaking debris. Lower quality materials would be used more expediently
and generate a similar signature to Place Provisioning, i.e. the artefacts would show
no evidence of standardisation, retouch, or raw material conservation, and pieces
would be discarded after light use (Kuhn 1995; Mackay 2005). Presumably, a more
limited quantity of debris would accumulate via the use of expedient technology as a
short-term risk reduction strategy, when compared to its widespread use within Place
Provisioning.
2.4.2 Place Provisioning
As residential mobility decreases, the duration and predictability of occupation
increases (Figure 2.3a; Kuhn 1995). It is therefore possible for people to identify
areas in the landscape where lengthy visits can occur at regular intervals in the
31
future, and to provision these places by stockpiling raw material. Stockpiling material
negates the ongoing transport costs associated with mobile toolkits, while
simultaneously avoiding direct costs of raw material transport by embedding
collection within foraging tasks – raw material selection therefore reflects the
foraging range (Bamforth 1986; Binford 1979; Kuhn 1995). The abundance of stone
means that people can adopt a less formal, unstandardised, expedient technology,
whereby tools are made as required and discarded once they are no longer
functional or after a task has been completed (Binford 1977; Kuhn 1995). Stone
consumption is much higher than that associated with formalised reduction
techniques, so expedient technology can only be widely adopted if stone is abundant,
either by stockpiling or through natural abundance (Bamforth 1986; Kuhn 1995;
Nelson 1991; Parry and Kelly 1987). Nevertheless, informal, expedient technologies
have several important benefits over the more standardised, formal reduction
techniques that characterise Individual Provisioning:
o
Unmodified raw material can be converted into whatever type of tool is
required, without being constrained by the size and form of blanks or small
transported cores (Kuhn 1995:24);
o
Unstandardised reduction is faster and easier than more formal reduction
techniques (Parry and Kelly 1987);
o
Fewer costs are entailed by discarding a blunt or broken tool and making a
new one, when compared to reworking or recycling an existing tool
(Bamforth 1986);
o
A fresh edge (i.e. newly detached flake) is always sharper than a retouched
edge (Parry and Kelly 1987);
o
A specialised tool (suited to a narrower range of tasks) will always be more
efficient than a generalised, multifunctional tool (Torrence 1983);
o
Raw material quality is less of a constraint – lower quality material can be
incorporated since tools are discarded after use or breakage and do not
need to be reworked (Graf 2010; Manninen and Knutsson 2014; Mitchell
2016:152).
Increased stone consumption means that expedient strategies will create much more
archaeological debris than Individual Provisioning and may overwhelm the latter
32
where both systems were in use at a single location (Binford 1977; Kuhn 1995:26).
As the focus is on primary reduction, cores should be a frequent component of the
stone artefact assemblage, and both cores and flakes will be much larger than under
Individual Provisioning (Mackay 2005; Speal 2009; Sullivan and Rozen 1985).
Flakes should be unretouched and many will show little or no macroscopic evidence
of use, while broken or worn tools will be discarded without further modification – no
efforts were made to extend the use-life of individual artefacts (Kuhn 1995; Mackay
2005); the lack of retouch should limit the number of very small flakes. There should
be little evidence of standardised flaking, manifested by a wide array of artefact sizes
and shapes. Raw material choices will be indicative of the foraging range, but may
trend towards the best material available locally – it should show little pre-processing
before transport, evidenced by larger fragment sizes and the presence of cortex
(Binford 1979; Clarkson 2007:24–25).
Despite the abundance of raw material, lengthy visits and the time since last raw
material procurement may encourage some level of raw material conservation and
economisation (Clarkson 2007:25), i.e. by using bipolar reduction that allows small
cores to be reduced further than would be possible using other techniques. Hiscock
(1996) found that even where raw material was abundant, the incidence of bipolar
flaking increased with occupation duration, while Parry and Kelly (1987) found that
bipolar artefacts were common among expedient assemblages. Therefore, it seems
likely that bipolar flaking would be characteristic of assemblages generated by Place
Provisioning. While it may seem contradictory to employ a conservation strategy
within a more wasteful, expedient technology (one characterised by the short-term
use of artefacts and lack of reworking/recycling), it is possibly due to costs relative to
benefits. Since it only alters the mechanism of primary reduction, the bipolar
technique represents a simple, low-cost means of conserving raw material while
retaining the benefits of an expedient technology.
Reduced residential mobility may, in some cases, be accompanied by increased
logistical mobility, especially where environments are patchy and returns are variable
(Binford 1980; Clarkson 2007:24; Kuhn 1995:27). Under these conditions, people
travel greater distances from the central residential location in search of resources, a
characteristic of Binford's (1980) collectors. Specific tools would be required for long
33
logistical forays, beyond those expediently manufactured items used for tasks
occurring at the residential site. People would therefore 'gear up' at the residential
location by manufacturing the required tools (Binford 1979). This is a form of
Individual Provisioning, but it occurs in concert with, rather than in opposition to,
Place Provisioning (Figure 2.3). Longer logistical forays generally target specific
resources, since these must be highly ranked enough to offset transport and fieldprocessing costs (Binford 1980; Orians and Pearson 1979). A narrower and more
predictable range of tasks would be conducted, so tools could be more specialised –
and therefore more efficient – than the generalised toolkit required under conditions
of high residential mobility. Portability and durability would still be integral
components, however (Clarkson 2007:24; Torrence 1983). Therefore, while
expedient technologies can accommodate lower quality materials, more careful
selection of stone would presumably characterise tools required for long logistical
forays. Evidence will be scarcer than the debris arising from expedient tool
manufacture but – if present – it should be distinguishable from traditional Individual
Provisioning. During the process of 'gearing up', tools are made from cached raw
materials, rather than the transported blanks or small cores used under traditional
Individual Provisioning. Therefore, evidence of primary reduction should dominate,
and debris should be more abundant than under traditional Individual Provisioning
and also show a greater variety of sizes, including larger flakes and cores; smaller
flakes arising from retouch may occur where a tool’s function required a retouched
edge. Standardised reduction and raw material conservation should be far less
evident than under Individual Provisioning.
2.4.3 Summary
Perhaps the greatest success of Kuhn's (1995) scheme is its reliance on flaking
methodology and retouch indices to identify signatures characteristic of Individual
and Place Provisioning. In contrast to models that use assemblage diversity or
design criteria, Kuhn's can be applied to fairly amorphous lithic collections dominated
by debris with few formal tools and little typological variation. This perhaps explains
why his model is commonly invoked when interpreting Australian lithic assemblages
(e.g. Clarkson 2004, 2007; Mackay 2005; Mitchell 2016) that are well-known for their
amorphous nature (Hiscock and Clarkson 2000). While Individual and Place
Provisioning have distinct technological signatures, Kuhn (1995:25–26) noted that
34
these strategies are not mutually exclusive. All mobile groups employ some form of
Individual Provisioning, in that there are some items that people always transport
with them, and some level of Place Provisioning. The dominant system simply varies
according to the circumstances at a given time or place, as a group's mobility
fluctuates in response to resource availability. Shifts in provisioning systems will be
most evident, then, from broad, inter-assemblage analyses rather than considering
lithics from an isolated site (Kuhn 1995:29).
2.5 CONCLUSION
Forager models demonstrate that optimisation can be achieved under varying
resource distributions by altering several components of hunter-gatherer life,
including which resources should be targeted, when one foraging patch should be
rejected for another, when resources should be processed before transport, whether
people should aggregate in a large group or disperse into small units, and how
frequently they should move their residential base. These findings are robust enough
to use without reference to the underlying mathematical models and are frequently
incorporated into archaeological analyses. Technological organisation can be
influenced by several factors, including mobility, raw material availability, and risk,
which are interrelated, to a certain degree. These factors dictate where and when
technological activities occur, which raw materials are used and to what extent they
are reduced, whether formal or informal technologies are adopted, and the reduction
techniques used to produce them. However, technological organisation must be
considered via a multifaceted approach geared towards local conditions, as no single
factor is universally dominant. Significantly, Kuhn's (1995) provisioning systems unite
forager theory with technological organisation and define optimal technological
solutions under different levels of residential mobility; more importantly, his scheme
can be applied to amorphous Australian assemblages. Using his predictions in
concert with local resource distribution, it should be possible to identify those parts of
the landscape where more frequent and less predictable moves would favour
Individual Provisioning, in contrast to more scheduled, predictable moves that permit
Place Provisioning. This provides a valuable context for analysing and interpreting
fragmentary archaeological remains.
35
CHAPTER 3
NATURAL ENVIRONMENT AND LANDSCAPE OF THE STUDY AREA
3.1 INTRODUCTION
This chapter outlines the natural environment of the study area, focussing on those
aspects that would dictate the availability of resources for its human occupants. The
geology, landform, and hydrology are described, as well as the impact of vegetation
clearing on hydrological conditions. The Quaternary climate of the continent is
evaluated, focussing on those periods relevant to human occupation in Australia –
the late Pleistocene and the Holocene. The Holocene palaeoenvironmental records
of southwestern Australia are summarised, with particular focus on the few records
from inland areas; the modern climate of the study area is also quantified, as it forms
a framework upon which to superimpose past climatic changes. Finally, the preEuropean distribution of flora and fauna is assessed after considering the impact
caused by clearing native vegetation for agriculture.
3.2 GEOLOGY AND LANDFORM
The study area, like much of southwestern Australia, is underlain by the Yilgarn
Craton (Figure 3.1). At 657,000 km2, this is the largest of the continent’s three
cratons, each of which represents a fragment of the Earth's crust that has avoided
significant deformation through tectonic processes for more than a billion years
(Anand and Butt 2010; Blewett et al. 2012; Johnson 2009:59). The rocks that make
up the Yilgarn – mainly granitoids and gneiss with north-northwest oriented bands of
greenstone (heterogeneous deposits of metavolcanic and metasedimentary rock,
more common in the eastern portion of the craton) – were formed during the
Archaean period, mostly between 3000 and 2600 million years ago (Ma), though
some parts date to 3700 Ma (Anand and Butt 2010; Anand and Paine 2002; Blewett
et al. 2012). There is considerable variation in the geochronology, geological history,
drainage and soil types across the craton, so it should not be considered a
homogenous unit.
36
Figure 3.1 Simplified geology of Western Australia, showing extent of the Yilgarn Craton, also known
as the Yilgarn Block (Beard 1990:41).
Anand and Butt (2010) divided the Yilgarn craton into three regions: Southwestern,
Southern and Northern (Figure 3.2), based on a combination of geological, drainage,
climatic, soil and groundwater characteristics. The Southern region incorporates the
study area and is characterised by low relief and gently undulating terrain, while the
adjacent Southwestern region comprises a plateau with high relief and a much more
dissected landscape (Anand and Butt 2010); these topographic differences result
from drainage histories that diverged around 16 million years ago. Large
palaeorivers had once snaked across much of southwestern Australia, but regular
flow ceased after the onset of aridity in the mid-Miocene between 16 and 11 Ma
(Martin 2006). In the Eocene, uplift created the Darling Range and rejuvenated
37
drainage west of the Meckering line (Commander et al. 2001). New river systems
were created in the Southwestern region, but east of the Meckering line (the Zone of
Ancient Drainage, which broadly corresponds to the Wheatbelt region) drainage
remains uncoordinated, and connected flow occurs only after periods of
exceptionally heavy rainfall (Anand and Butt 2010; Beard 1999; Bettenay 1962;
Hatton et al. 2003).
Figure 3.2 Major regions within the Yilgarn Craton (Anand and Butt 2010:1018).
As a result of the inactive drainage and tectonic stability noted above, the
topography of the Wheatbelt is, as a whole, flat and subdued. The lowest points in
the landscape are the broad, flat valley floors up to 15 km wide – relicts of
palaeodrainage systems at least 45 million years old (Commander et al. 2001). The
oldest of these deposits are the narrow palaeochannel sands and clays (Figure 3.3)
that are concealed by younger, Quaternary deposits that infill the valley. Within these
valley floors are a characteristic feature of the inland southwest – chains of salt lakes
and playas. Tertiary laterite formations are common in small pockets, while the
remnant sandplains derived from this cover a large portion of the landscape (Figure
3.4; Murphy-White and Leoni 2006). Archaean granite outcrops exist throughout the
landscape – these are surface expressions of more resilient portions of the
38
underlying bedrock that have resisted the weathering to which the surrounding rock
has succumbed (Campbell 1997). Gneisses are largely confined to the east, outside
the study area (Figure 3.4).
As a whole, the topography is unimposing, and the slopes gradual; only the
occasional granite outcrop breaks the horizon. Such outcrops were valued by
Aboriginal people as they capture and trap rainfall in crevices and depressions,
provide a range of plant and animal foods, and act as natural lookouts (Bayly 1999;
Bindon 1997; Smith 1993:20).
Figure 3.3 Geology of characteristic Wheatbelt valleys (Commander et al. 2001, reproduced from
Department of Environment 2005b).
The study area yields three raw materials that are suitable for artefact manufacture:
quartz, silcrete and dolerite. Silcrete (Figure 3.4: Czb) occurs within lateritic profiles,
generally above deeply weathered bedrock (Czo). It becomes exposed as the
overlying duricrust (Czl) erodes or where the silcrete layer is otherwise exposed, as
in breakaways (Chin et al. 1984). Silcrete occurs in small patches, predominantly
south of the main palaeodrainage channel, but is much more common in the
greenstone belt east of the study area (Figure 3.4a). Dolerite is available in the large
discontinuous dolerite-gabbro dykes (Figure 3.4: Pd) striking ENE along the main
palaeochannel; smaller dykes are fairly common throughout the study area. Quartz
occurs as veins in granite bedrock, formed by hydrothermal processes whereby fluid
is forced into a void and subsequently cools into a solid – the oldest, coarsest quartz
39
is found in the core of the vein, surrounded by younger, finer textured material
(Abeysinghe 2003:6); a single source can therefore yield quartz of varying quality (de
Lombera Hermida 2009). Quartz veins are especially common in the Southwestern
region and can outcrop at the surface where the surrounding rock has weathered
away (Abeysinghe 2003:172, Figure 111). One vein is large enough to appear on
geological maps of the study area (Figure 3.4a), but smaller veins could be
encountered in any area of exposed bedrock.
Pegmatite bodies can yield high-quality quartz from their cores (often 10–15 m wide),
as crystals have the space to form unimpeded, so can reach considerable sizes
(Abeysinghe 2003:174; Fetherston et al. 2017:91). A single pegmatite vein occurs in
the study area (Figure 3.4a) but the cores are generally encased by several metres
of pegmatite, so the quartz may be inaccessible unless the core has become
exposed (Abeysinghe 2003:164). Quartzite (Alm) and banded chert (Aic) are present
in the greenstone belt outside the study area immediately east of the vermin-proof
fence (Figure 3.4b) – both are well-suited to artefact manufacture (Chin et al. 1984;
Glover 1984).
3.3 HYDROLOGY
The study area lies within the Lockhart River catchment, one of four sub-catchments
in the Avon Basin. While the Avon is one of Australia's major river systems, most of
the 120,000 km2 basin (including the Lockhart catchment) drains internally, meaning
connected flows are rare except after exceptionally wet periods (Bettenay 1962;
Department of Water 2009; Galloway 2004). This lack of flow means that weathering
products and, more importantly, salts, are not removed from the system and instead
accumulate over long periods (Anand and Butt 2010). Salt has been naturally
accumulating in the soils for many thousands of years, transported by westerly winds
from the Indian Ocean, but it is also present in rainfall and in the regolith (Blewett et
al. 2012; Clarke et al. 1998; De Dekker 1983; Hatton et al. 2003; Salama 1994).
These salts are concentrated in areas where water pools and subsequently
evaporates (salt lakes and playas), leaving behind ever-increasing quantities of salt
that active drainage would otherwise remove. Salinity is, therefore, a natural and
prominent part of the landscape (Bettenay et al. 1964; Department of Water 2009).
40
Figure 3.4 Geology of selected parts of the study area (A) and beyond (B), showing the occurrence and typical distribution of potential tool-stone. Geological
imagery (including legend) from Geological Survey of Western Australia (1984). Note that geological maps depict the surface and near-surface geology, so
raw material sources may not have all been accessible at ground-level.
41
This natural salinity defines the landscape of the study area, which is characterised
by saline groundwater, saline surface features, and a lack of freshwater lakes or
rivers. Saline groundwater results from the interplay of several factors including
increasing aridity, long residence times, low hydraulic gradients, and evaporative
concentration of salts around playas and discharge features (Anand and Butt 2010;
Bettenay et al. 1964; George 1992; Magee 2009).
Aside from on the coastal plain, most groundwater is above the safe salinity level for
human consumption, around 2100–1600 mg/L (Figure 3.5; NHMRC, NRMMC
2011:973–974). The salinity levels are even greater in the study area, commonly >
20,000 mg/L, and up to 55,000 mg/L in palaeodrainage channels (Bowen and
Benison 2009; Ghauri 2004; Hollick 2004). The saline and hypersaline groundwater
combined with low rainfall accounts for the lack of freshwater surface features in the
study area. In contrast, saline surface features have been part of the landscape for
millions of years, acting as discharge points for naturally occurring saline
groundwater (Bettenay et al. 1964; Bowler 1986; Department of Water 2009; George
and Coleman 2001; Salama 1994). Nevertheless, salt lakes and playas are not static
features, but instead respond to climate change. George and Coleman (2001)
argued that current salt lakes in the Wheatbelt are presently half the size they were
20,000 years ago, during the last dune-building phase. Salt lakes are also dynamic
across shorter time-scales, as water-level and salinity vary throughout the year in
response to rainfall (Bowler 1986; Halse et al. 2003).
Much of the landscape has also been affected by secondary (dryland) salinity, an
unfortunate by-product of agricultural development, specifically the clearing of native
vegetation for introduced crops. In the study area, less than 13% of native vegetation
remains (Department of Water 2009). Native plants are deep-rooted, perennial, and
will effectively use all water available to them over the course of a year (Mitchell et
al. 2009; Wildy et al. 2004). In contrast, introduced crops are shallow-rooted annual
plants that are inactive for several months of the year, and therefore have lower
water requirements. Hatton et al. (2003:344) concluded that before the Wheatbelt
was cleared, 'very little rainfall was discharged in liquid form'. Instead, most rain
infiltrated the soil where it fell and was available for plant use or evaporated from the
ground on which it landed; runoff was limited, as was recharge to the regional aquifer
42
system (Farrington and Salama 1996; Hatton et al. 2003). Rainwater that would
previously have been intercepted and utilised by native vegetation is now surplus to
crop needs and is redistributed as runoff to lower lying areas (e.g. valley floors)
where it infiltrates the soil and accumulates. This increased recharge can
significantly raise the water table in just decades (Farrington and Salama 1996;
George et al. 1997; Hatton et al. 2003).
Figure 3.5 Groundwater salinity across the Yilgarn Craton (Anand and Butt 2010:1078).
As the water table rises, it dissolves and mobilises the salt present in the deeper
parts of the soil profile below the major rooting zone for vegetation and brings it
closer to the surface, where it is detrimental to plant growth (Clarke et al. 2002;
Hatton et al. 2003). This process creates newly salinised land and increases the salt
concentration in naturally saline land. In many cases, Melaleuca shrubs (and
associated species) that inhabit naturally saline areas cannot adapt to increased salt
loads, so these plants die and are replaced by those with a greater salt tolerance,
such as samphires (Department of Water 2009). Stands of dead vegetation are
common in these areas, indicating the very short timeframe across which these
processes operate (Figure 3.6). Salt lakes have also been drastically affected –
under native vegetation conditions, they would have held water for a few months
43
after winter rains and been dry by summer, except where significant summer rains
were generated by cyclones that occasionally arrived from the North West
(Department of Water 2009; Halse et al. 2003). The rising water table, in conjunction
with additional input from surface runoff, means that these lakes now hold water for
far longer and, in many cases, have lost their natural wetting and drying cycles to the
detriment of the local plants and animals (Department of Environment 2005a;
Department of Water 2009; George et al. 2008; Halse et al. 2003). The water is
becoming saltier, and salinity levels now frequently exceed 200,000 mg/L, although
salinity and water chemistry vary between water bodies (Benison et al. 2007; Bowen
and Benison 2008; Cale et al. 2004).
Figure 3.6 Recently dead vegetation near Lake Gounter, Hyden, in October 2016 (Photograph: A M
Rossi). Note the replacement of large woody shrubs by much smaller plants with a greater salt
tolerance.
3.4 CLIMATE
Today, Australia has several climatic zones – from the temperate south, the arid
deserts of the red centre, to the tropical north. This variation is unsurprising
considering the continent spans 33 degrees of latitude and 40 degrees of longitude;
past climates would have varied similarly, but conditions were very different from
those at present. Climatic variation is driven by the interplay of many factors, such as
palaeogeography, volcanic activity, insolation levels, as well as shifts in the Earth’s
orbit and rotational axis (Hays et al. 1976; Laskar et al. 2004; Wanner et al. 2008).
The characteristics of this past climate can be determined through a variety of
methods, including the analysis of ice- and sediment-cores, pollen, alluvial/lacustrine
deposits, and speleothems (e.g. Churchill 1968; Galloway 1965; Harrison 1993;
Moros et al. 2009; Pickett et al. 2004; Quigley et al. 2010). Most techniques involve
44
the use proxy data such as lake levels, for example, which respond to the amount of
effective precipitation received during a specific period. Effective precipitation is
dictated by the combination of total precipitation, temperature, evaporation, wind,
and vegetation cover (Reeves et al. 2013). Hence, it can be very difficult to quantify
specific climatic conditions, so periods are often characterised only in a relative
sense (i.e. wetter vs. drier).
The late Quaternary, which incorporates the Pleistocene and Holocene epochs, is
the only period relevant to human occupation of Australia. Changes in the Earth’s
orbit and rotation (the Milankovitch cycle), are largely responsible for the
characteristic feature of the late Quaternary global climate, where periods of aridity –
glacial periods – are interspersed with interglacials, when conditions were wetter and
more favourable to humans (Bowler et al. 1976; Galloway 1965; Kershaw et al.
2003; Martin 2006). The changing sea-level, which resulted from ice sheet
expansion and retraction, would have intensified the variations caused by
astronomical cycles. During the Last Glacial Maximum (21 ± 3 ka: Reeves et al.
2013)
continent-wide
climatic
data
were
most
synchronous:
sea-surface
temperatures were 3–6 °C colder than present; sea-levels were approximately 100–
130 m lower than today, exposing around 3 million sq. km of continental shelf;
aeolian/dune-building processes were at their height; and drier conditions are
evident from many terrestrial records (Barrows et al. 2007; Brooke et al. 2017;
Johnson 2009:180; Lewis et al. 2013; Nanson et al. 1992; Reeves et al. 2013; Zheng
et al. 2002). However, effective precipitation may have reached its minimum a few
thousand years later, between 15 and 11 ka, but the timing varied significantly
across the continent (Harrison 1993; Kershaw and Nanson 1993; Reeves et al.
2013; Turney et al. 2006; Williams et al. 2009).
Worldwide, Holocene climatic variation was far less severe than during the
Pleistocene (Mayewski et al. 2004; Wanner et al. 2008); this pattern is echoed in
Holocene Australian records (Bowler et al. 1976; Chappell 1991; Dodson 2001).
Broadly, the early Holocene (12–8 ka) was characterised by warmer and wetter
conditions across the continent, but the timing of onset varied regionally (Reeves et
al. 2013). Seas reached their current levels between 8 and 7.5 ka (Lewis et al.
2013). The late Holocene (5–0 ka) climate was more variable, a characteristic
45
attributed to the El Niño Southern Oscillation (ENSO) in the north and east of the
continent, and a redistribution of westerly wind belts in the south (Reeves et al.
2013). The varying distribution and influence of these systems resulted in another
characteristic of the late Holocene – distinct spatial variation in climate – which
means that it is important to consider southwestern Australia separately.
3.4.1 Palaeoclimate of Southwestern Australia
It
is
well
known
that
southwestern
Australia
is
poorly
represented
in
palaeoenvironmental records (Gouramanis et al. 2012; Hesse et al. 2004; Petherick
et al. 2013; Reeves et al. 2013); long-term Holocene data are available from just a
few sites (Figure 3.7). While the specifics of most studies are irrelevant here, the
discordance between their results is noteworthy (Figure 3.8; Table 3.1). Opinion is
divided about whether the current climate is wetter (Churchill 1968; Dodson and Lu
2000; Kendrick 1977; Semeniuk 1986; Yassini and Kendrick 1988), drier
(Gouramanis et al. 2012; Harrison 1993; Zheng et al. 2002), or the same (Newsome
and Pickett 1993) as that in the mid- to late Holocene – scholarly opinion is equally
divided concerning the timing and length of perceived wet/dry periods (Figure 3.8).
Hence, the discordance between the palaeoenvironmental studies ‘raises the
question of how much reliance should be placed on the results’ (Newsome and
Pickett 1993:246). Due to the location of the study area, inland records (Harrison
1993; Zheng et al. 2002) are most relevant, and are discussed below. These data
are more coherent, but this may be a product of an extremely small sample size.
Harrison’s
(1993)
continental
lake-level
analysis
includes
four
sites
from
southwestern Australia: Lake Grace, Lake King, Myalup Swamp and Storeys Lake,
but the latter only preserves an isolated record at 17,000 BP. The remaining three
sites provide a record from 6000 BP to the present, but data are never available for
all three sites for any specific millennium (Figure 3.9). Lake-levels were inferred from
lithology or stratigraphy, and multiple radiocarbon (14C) dates were available for
each site except Myalup Swamp, for which only one age determination was
obtained. The data are extremely limited, but the overall pattern indicates higher
lake-levels at 6000 and 5000 BP, while low-lake states dominate thereafter until the
present (Harrison 1993).
46
Figure 3.7 Palaeoenvironmental sites mentioned in the text. Those in and around Perth are not
marked.
Figure 3.8 Major climatic trends identified in southwestern Australian palaeoenvironmental studies.
Periods were classified in relation to the preceding period; blue = wetter, red = drier, black = no
change, grey = uncertain. All dates have been plotted as cited, although Churchill’s (1968) dates have
been converted from BC/AD to cal. BP and plotted on the secondary axis with the data obtained by
Gouramanis et al. (2012).
47
Table 3.1 Summary of main palaeoenvironmental data from southwestern Australia – research
focussing on the recent past has been omitted (e.g. Cullen and Grierson 2009; Treble et al. 2003).
Temperature data from Gouramanis et al. (2012) as it does not span the entire record. Dates are cited
in the format provided by the author. * = Perth metro area, not marked on Figure 3.7.
Author
Sites/study area
Proxy
Major palaeoenvironmental results
Churchill
(1968)
1. Boggy Lake
2. Flinders Bay
Swamp
3. Fremantle*
4. Myalup
Swamp
5. Perth*
6. Rottnest
Island
7. Scott River
Swamp
8. Weld Swamp
9. West Lake
Muir
Pollen
Several shifts from wet to dry conditions, as
follows:
o 5000–2600 BC: wetter
o 2600–500 BC: drier (maximum at 1200
BC)
o 500 BC – AD 400: wetter
o AD 400–1400: drier
o AD 1400–present: wetter
Kendrick
(1977)
Swan River near
Guildford*
Molluscs
Drier conditions from before 6700 BP to around
4500 BP. Wetter conditions thereafter.
Semeniuk
(1986)
1.
2.
3.
4.
Lancelin
Whitfords*
Rockingham
Leschenault
Peninsula
Calcrete
7000–3500 BP mostly semi-arid, small band
along west coast with subhumid-humid. After
2800 BP regional trend to increased humidity.
Yassini and
Kendrick
(1988)
Point Waylen*
Ostracods
Drier conditions 6000– 4100 BP. Wetter
conditions thereafter.
Harrison
(1993)
1. Lake Grace
2. Lake King
3. Myalup
Swamp
Lake-levels
Higher lake levels 6000–5000 BP. Low-lake
states almost exclusive thereafter.
Newsome
and Pickett
(1993)
1. Loch McNess Pollen
2. Boggy Lake
Largely stable climate 9000 BP to present;
possibility of shorter arid periods throughout.
Dodson and
Lu (2000)
Byenup Lagoon
Pollen
Record begins around 4800 BP, indicating
regional increase in effective precipitation.
Zheng et al.
(2002)
1. Lake Cowan
2. Lake Lefroy
Aeolian and
lacustrine
sediments
High lake event 10,000– 4000 BP. Reactivation of
aeolian activity indicating increased aridity 4000–
2000 BP, possibly into the present.
Gouramanis
et al. (2012)
3. Barker
Swamp
Pollen,
charcoal,
gastropods,
ostracods
Record begins 7.4ka, indicating regional increase
in effective precipitation. Several smaller shifts:
o 7.4–7.2 ka: wetter
o 7.2–6.7 ka: drier
o 6.7–6.2 ka: wetter
o 6.2–5.8 ka: drier
o 5.8–4.5 ka: wetter
o 4.2–2.3 ka: drier
o 2.3–1.4 ka: wetter
o 1.4 ka – present: dry/climate as today.
48
Figure 3.9 Harrison’s (1993:219) reconstruction of lake status at 1000-year intervals, from 6000–1000
BP.
49
Finer grained data are available from the Lake Cowan/Lake Lefroy system more than
600 km east of the present coastline. Based on optically-stimulated luminescence
(OSL) and radiocarbon dates from aeolian and lacustrine sediments, Zheng et al.
(2002) proposed that the Holocene climate transitioned through two major phases:
1. Early to mid-Holocene: High lake event 10,000–4000 BP;
2. Late Holocene: Reactivation of aeolian activity 4000–2000 BP, possibly
continuing into the present.
If aeolian activity and lake-levels are indeed accurate proxies for effective
precipitation, then conditions at Lakes Cowan and Lefroy were wetter from 10,000 to
4000 BP and became drier thereafter.
3.4.2 Modern Climate of the Study Area
To understand the variation in effective precipitation discussed above, it is necessary
to consider the baseline environmental conditions upon which these variations are
superimposed. The study area is within Beard’s (1981, 1990) Extra Dry
Mediterranean climatic zone (Figure 3.10), where maximum rainfall occurs in winter
and there are seven or eight ‘dry’ months (where evaporation exceeds precipitation)
each year. In Hyden, maximum summer temperatures average 34° C but may reach
48° C. Winter minima average 5° C, but may fall as low as -6° C at night (Bureau of
Meteorology data). Average annual rainfall is approximately 340 mm (based on
1929–2017 data), but in many years this average is not reached; every five to ten
years falls are significantly above average (Figure 3.11); this highly variable rainfall
pattern is characteristic of Mediterranean climates (Dodson 2001). If the limited
palaeoenvironmental data for inland southwestern Australia (Harrison 1993; Zheng
et al. 2002) are applicable, then this climatic pattern has been in place for at least
4000 years.
50
Figure 3.10 Bioclimatic zones of Western Australia (Beard 1990:39).
Figure 3.11 Annual rainfall data from Hyden weather station (no. 10568), 1929–2017 in relation to
average rainfall (blue dashed line). Data are not available for 1947 or 1974.
51
3.5 VEGETATION
Palaeovegetation data for southwestern Australia are rare, but those available data
indicate little shift in vegetation patterns during the Holocene (Dodson 2001; Dodson
and Lu 2000; Itzstein-Davey 2004; Newsome and Pickett 1993; Pickett 1997; Pickett
et al. 2004). The one site with a longer record spanning the last 31,000 years (Bibra
Lake, in Perth), indicates that the biome around Perth at 18,000 BP was the same as
that of the present-day (temperate sclerophyll woodland and shrubland) but different
species may have dominated (Pickett 1997; Pickett et al. 2004). Regardless of
whether floral stability indicates a largely unchanging climate or the resistance of
vegetation communities to climate change, the vegetation patterns evident before
agricultural development had probably been in place for millennia.
3.5.1 Botanical Provinces of Western Australia
Beard (1990) divided Western Australia into three main provinces (Figure 3.12a):
South-West (SWP), Eremaean (EP) and Northern Province (NP). The boundaries of
these major provinces are broadly congruent with the climatic zones shown in Figure
3.10, indicating that, as Beard (1990:50) stated, the vegetation of Australia has
adapted to thrive in the nutrient-poor conditions and is instead constrained only by
the amount of moisture in the soil. The study area lies within Beard’s (1990) SouthWest Province, whose eastern boundary roughly separates the Extra-Dry and SemiDesert Mediterranean zones (Figure 3.10). Here, deeply weathered soil profiles
allow plant roots access to the pallid zone (weathered rock) where moisture
accumulates, enabling trees to survive through the dry season. This unique feature
of the landscape allows much more densely wooded areas, such as those in the
lower southwest, to be found within Mediterranean-type climates (Beard 1990:50,
55). The SWP is a biodiversity hot-spot, known for its high proportion of endemic
species, around 3000 of the 5710 identified (Beard et al. 2000). Clearing has had a
disastrous effect on the natural environment of the study area, but the floral species
themselves endure. In the Wheatbelt (approx. 75% of the SWP), almost 350 species
are now considered rare or endangered, but by the 1980s only 24 had been
rendered extinct (Hobbs 2003; Hopper et al. 1990; Leigh et al. 1984).
Beard’s (1990) provinces have been separated into a number of regions (Figure
3.12a), most of which have been retained in IBRA 7 (Interim Biogeographic
52
Regionalisation for Australia, Version 7 – Figure 3.12b). Most of the study area falls
within the Mallee region, while the northern extremity lies in the Avon Wheatbelt
region. The Mallee Region is so-named for the dominance of the mallee vegetation
structure, comprising small multi-stemmed eucalypts. Mallee originally covered
around 75% of the region, while woodlands, scrub-heath and thicket were
comparatively rare (Beard 1990:130 & 135). In contrast, the Avon Wheatbelt only
preserves small isolated patches of mallee, and is instead dominated by woodlands
and thicket (Beard 1990:118). Beard (1990) related the distribution of these different
vegetation units to topography, and therefore soil type and depth (Figure 3.13), but
there is some argument concerning exactly how topography and associated
variables influence vegetation formations on a more general level (Dirnböck et al.
2002). The broad catenary sequence is similar in both regions, with hypersaline
areas inhabited by halophytes, heavy lowland soils by woodlands, duplex slopes by
mallee, and sandplains by thicket and/or scrub-heath (Figure 3.13).
Figure 3.12 A: Natural regions of Western Australia (Beard 1990:34); B: IRBA 7 regions of Western
Australia (Australian Government Department of Environment and Energy 2012). Abbreviations
largely reflect Beard’s (1990) subregion names except: AVW = Avon Wheatbelt; COO = Coolgardie,
GES = Geraldton Sandplains, JAF = Jarrah Forest, SWA = Swan Coastal Plain, VIB = Victoria
Bonaparte, YAL = Yalgoo.
53
Figure 3.13 Catenary sequence for the Mallee Region (top) and Avon Wheatbelt Region (bottom),
showing the relationship between topography and vegetation communities (Beard 1990:116, 129).
3.5.2 Vegetation Communities in the Study Area
Beard (1975) mapped the entire South West Botanical Province at 1:250 000 scale,
the highest resolution vegetation data available for the entire region. He aimed to
reconstruct the distribution of native vegetation before European arrival, using
remnant vegetation, aerial photography and soil-vegetation catenae. Vegetation was
classified based on the nature, size and density (canopy/foliage coverage) of the
dominant stratum, plus any diagnostic plant species. The spatial data are preserved
NRInfo (DPIRD 2016b), as well as published maps and accompanying memos
(Beard 1972, 1976, 1979, 1980). Where particular plant communities recurred
following the same catenary sequence, they were grouped into vegetation systems –
effectively subdivisions of a botanical province (Beard 1976).
Three vegetation systems are present in the study area: the Hyden, Muntadgin and
Skeleton Rock systems. The same catenary sequences are evident in each,
although they preserve different proportions of vegetation communities due to local
54
variation in landform and soils. These communities are summarised in Table 3.2 and
their distribution mapped in Figure 3.14.
Table 3.2 Vegetation communities found in the study area. Community names, descriptions and
classifications are based on Beard (1972, 1975, 1976, 1979, 1980). Note that these descriptions
apply only to the study area (i.e. the Hyden, Muntadgin and Skeleton Rock systems), and may not
reflect community characteristics at a broader scale.
Vegetation
community
Description
Broombush
Thicket
Very dense one- or two-layered shrub community. Dense upper layer dominated
by Allocasuarina, Acacia and Melaleuca > 1 m, generally 2–2.5 m tall, 30–70%
canopy coverage. Lower layer may be absent or suppressed, comprising sparse
layer of shrubs < 1 m tall.
Granite rocks
Outcrop bare or lichen-clad; small trees and shrubs may occur where sufficient soil
has accumulated in clefts and depressions. Fringe vegetation diverse and varied,
frequent thickets and woodlands.
Mallee
Open Eucalypt shrub assemblage, > 1 m but < 10 m tall – canopy coverage
generally 10–30%, but up to 70%. Eucalyptus eremophila almost always present,
associated with several other eucalypts. Variable low shrub understorey, often
dominated by Melaleuca.
Melaleuca
thicket
On swampy (saline) ground. Closed (30–70%) community of Melaleuca shrubs >
1m tall.
Scrub-heath
Mixed, stratified, partly open shrub assemblage. Lower closed layer of shrubs < 1
m tall, 30–70% canopy coverage. Open (10–30%) upper layer of shrubs > 1m tall.
No dominant species.
Succulent
steppe
On saline soils. Samphire communities comprising low succulent shrubs. May form
ground layer only or be associated with teatree shrubs and/or Eucalypt woodland.
Teatree
scrub
On saline soil. Irregular open community of shrubs 1–6 m tall, commonly 2.5–3.5
m. Single-layered formation, canopy coverage 10–30%. Several species, all
Melaleuca, with occasional admixture of small eucalypts. Ground vegetation
absent, or scattered grass and samphire.
Woodland
On the heaviest soils, lower in the landscape. Woodland comprising Eucalyptus
trees 10–25 m tall, with 10–30% canopy coverage; dominated by E.
salmonophloia, often in association with E. longicornis. Understorey dominated by
Melaleuca pauperiflora, irregularly scattered and reaching up to 3.5 m tall, plus
various eucalypts. Lower shrub layer varies depending on proximity to salt.
55
Figure 3.14 Pre-European distribution of vegetation communities in the study area. Spatial data
extracted from NRInfo (DPIRD 2016b), based on Beard’s (1972, 1979, 1980) data. Continuous black
line separates the two IBRA regions that meet in the study area: the Avon Wheatbelt (north) and
Mallee (south). Dashed line represents the boundary of the study area.
Mallee is dominant in the south of the study area, reflecting its local prevalence
within the Mallee region; it has a much more restricted distribution further north,
within the Avon Wheatbelt region, but there are sizeable areas of mallee with
patches of woodland. True woodlands are much more limited in extent, but
substantial areas occur around larger salt lakes in the south-eastern portion of the
study area where soils are heavier. While structurally different, similar species are
often found in woodland and mallee (Beard 1972, 1979). Thicket and scrub-heath
have a mutually exclusive distribution: the former is prevalent in the north of the
study area and the latter restricted to the south. The structure of these formations is
also inverted; thickets comprise a dense upper layer of taller shrubs, sometimes
accompanied by a sparse understorey of small shrubs, while heath exhibits a dense
lower shrub layer and a sparse upper layer (Table 3.2). Similar species may be
present in both formations, though thickets often have a less diverse assemblage
where the understorey is absent (Beard 1972, 1979). Teatree scrub and Melaleuca
56
thicket are similar, both found on saline soils in palaeochannels, but the shrubs are
more densely packed in thickets. Succulent steppe is much more limited in
distribution, confined to the southwestern portion of the study area in the main
palaeochannel. Finally, granite rocks outcrop throughout the study area. These
preserve a multitude of microhabitats, ranging from eternally dry soils to permanently
shaded and moist cave deposits, identifying granite outcrops as local biodiversity
hotspots (Hopper et al. 1997). Outcrops channel water, nutrients and soil to the base
of the rock, permitting denser vegetation than would otherwise be possible while also
providing valuable water sources for humans (Bayly 1999; Bindon 1997; Schut et al.
2014). Each of the vegetation communities described above would offer a unique
suite of plant foods for human consumption, from eucalypt-derived lerps and manna,
to Acacia and other tree seeds, as well as various fruits, roots and flowers (see
Chapters 5–6).
3.6 FAUNA
The native fauna of the Wheatbelt has been significantly impacted by extensive landclearing. Of the 43 native ground-dwelling mammal species that originally inhabited
the Wheatbelt, 13 have disappeared since European arrival, and just 12 species are
considered abundant or common (Kitchener et al. 1980). A total of 195 bird species
have been identified, of which two are now extinct and 95 have suffered a decline in
range or abundance (Saunders and Ingram 1995). At least 22 species have become
more widespread and numerous following agricultural development (Saunders
1989); predatory birds are now more common due to introduced prey including pests
(mice, rabbits) and stock animals, while other birds have been attracted by new plant
foods such as cereal crops (Storr 1991). A total of 1000 wetland invertebrates, 665
ground-dwelling spiders, 86 reptile and 21 frog species have been identified in the
Wheatbelt (McKenzie et al. 2004; Pinder et al. 2004), but the impact of clearing on
these populations is unknown. Kitchener et al. (1980) argue that few lizards have
become extinct since populations persist in small patches where their habitat is
preserved, such as granite outcrops.
The species mentioned above would not all be present in the study area, as the
Wheatbelt incorporates an area of almost 25 million hectares (Hobbs 2003). While
57
smaller scale faunal data are available, they are limited to surveys of isolated
reserves. For example, McKenzie et al. (1973) studied the faunal assemblage at
Dragon Rocks, 30 km southeast of Hyden, and identified just 15 native mammals
(including macropods, possums, and the echidna – Tachyglossus aculeatus), 59
birds (including the emu – Dromaius novaehollandiae), 19 reptiles (including two
dragons) and four amphibian species. This list likely underestimates pre-clearing
faunal diversity but indicates the variety of animal foods available to Aboriginal
people, who are known to have eaten almost any animal they could catch (Hassell
1975:10).
3.7 CONCLUSION
Long periods of tectonic stability and inactive drainage have created a
characteristically subdued landscape in the study area, dominated by gentle slopes
and low relief; salt lakes are common in the relict palaeodrainage systems. Clearing
of native vegetation has altered the landscape by raising the naturally saline water
table and creating salinised and waterlogged soils. Salt lakes have been a natural
part of the landscape for thousands of years, but before clearing they probably only
held water for a few months each year. While the pre-clearing landscape can be
quantified there is, at present, no coherent palaeoenvironmental reconstruction;
opinion is divided about whether the current climate is wetter or drier than that of the
preceding period.
The limited data from inland southwestern Australia indicate that drier conditions
have been in place for the last 4000 years. Prior to clearing, mallee covered most of
the study area, with scrub-heath in the south transitioning to thicket in the north;
woodlands and salt-tolerant vegetation was also present. The fauna of the study
area suffered devastating species loss following the introduction of agriculture, but
several bird species have benefitted from the introduced food sources and more
abundant water. Hence, modern faunal surveys do not necessarily accurately
represent pre-European distribution.
58
CHAPTER 4
CULTURAL CONTEXT
4.1 INTRODUCTION
This chapter provides the cultural context, both archaeological and ethnohistorical,
for the study area. The archaeological background is evaluated, including the earliest
evidence for Aboriginal people in Australia, in marginal environments, and in the
southwest. The few southwestern studies with some inland component are then
discussed. Finally, the known archaeological locations within the study area are
identified – they are few in number, and none have been investigated in detail or
radiometrically dated. The ethnohistorical sources are then analysed, though these
are few, fragmentary, and fraught with biases. Ethnographic information collected
during recent cultural heritage surveys provides a valuable supplement to this limited
dataset.
4.2 ARCHAEOLOGICAL BACKGROUND
It is generally accepted that Aboriginal people first moved from island South East
Asia to Sahul (the landmass comprising mainland Australia, Tasmania and New
Guinea, which were joined at times of lower sea level), sometime around 60,000
years ago. Two sites with occupation evidence from 60–50 ka, Nauwalabila and
Madjedbebe/Malakunanja II, were deemed problematic by some due the questioned
association between artefacts and dated materials (see contrasting debates in
Hiscock 2013 and Allen and O'Connell 2014). However, Clarkson et al. (2017) reexcavated and re-dated Madjedbebe and demonstrated compelling evidence in the
form of diverse stone artefact assemblages, ochre, and evidence of plant processing
for occupation by 65,000 cal. BP. Nevertheless, the main body of archaeological
evidence in Australia post-dates 47,000 cal. BP (e.g. David et al. 2013; Morse et al.
2014; Slack et al. 2004; see list in Allen and O’Connell 2014). Following colonisation,
cross-continent movement was fairly rapid, and most parts of the continent, including
the present arid zone, have sites pre-dating 40,000 cal. BP (e.g. Allen and O’Connell
2014; Smith 2013:79–80).
59
Aboriginal people were in southwestern Australia relatively soon after initial
colonisation – while 50,000 cal. BP dates for Devil's Lair (Turney et al. 2001) are
disputed, the 45.44 ± 2.57 ka date is widely accepted (Allen and O'Connell 2014;
Balme 2014); Upper Swan dates to around the same time, 44.77 ± 3.92 ka (Pearce
and Barbetti 1981, calibrated in Allen and O'Connell 2014). Despite the more coastal
location of these sites, there is no reason to presume that inland areas could not
have been occupied this early – it would simply have depended on the local
landscape, specifically the distribution of critical resources, particularly water. Early
sites in the arid zone elsewhere in Australia are all associated with a reliable water
source of some kind (Bird et al. 2016; Smith 2013:78–91).
4.2.1 Archaeological Sites in the Study Area
While several archaeological sites have been identified around the study area and in
the wider region, little is known about them, and none have been radiometrically
dated. A total of 16 registered sites and ‘other heritage places’ (as defined by the
Department of Planning, Lands and Heritage – DPLH) are within the study area,
excluding those analysed as part of this research (Figure 4.1, Table 4.1); expanding
the area would yield more sites but add little comprehensive data. Sites are mostly
found on granite outcrops and valley floors around salt lakes, echoing Bird’s (1985)
and Smith’s (1993) results (see Chapter 4.2.2). Corsini (1993) noted that he visited
several unnamed granite outcrops in the area and found artefacts in most places
where water pools after rain. Nevertheless, the number and distribution of known
archaeological locations should not be presumed representative of the area as a
whole – preservation, survey and reporting biases must be considered. Sites are
more likely preserved in landforms unaffected by agriculture, such as granite
outcrops and salt lakes. Site locations skew towards areas surveyed for
development
(roads,
pipelines,
infrastructure)
or
places
reported
by
locals/landowners, rather than from the results of systematic research surveys. This
likely explains the relatively high number of rock art sites, as these are easily
identified by people with no archaeological training.
60
Figure 4.1 Registered sites and Other Heritage Places in the study area. White squares denote approximate site positions, enlarged in inset maps to show
their immediate surroundings. Locational data were retrieved from Aboriginal Heritage Inquiry System (AHIS), maintained by the Department of Planning,
Lands and Heritage. The central point of sites/place marked on insets. Site details can be found in Table 4.1.
61
Table 4.1 Archaeological features in the study area, excluding those analysed for this research. ID,
name and classification taken from DPLH’s Aboriginal Heritage Inquiry System (AHIS), which
separates Registered sites (RS) from ‘Other Heritage Places’ (OHP). Unless noted otherwise,
information derives solely from DPLH site files. The map code indicates the position of a particular
feature or set of features on Figure 4.1— features are numbered from west to east.
ID
Name
DPLH
class
Map
code
Archaeological
content
Reference
16787
Kondinin Pipeline 1
OHP
1
Artefacts
Staub and Taylor 1999
18508
Corrigin-Hyden Road
OHP
2
Artefacts
16791
Kondinin Pipeline 5
RS
3
Artefacts
Staub and Taylor 1999
16788
Kondinin Pipeline 2
RS
4
Artefacts
Staub and Taylor 1999
16789
Kondinin Pipeline 3
OHP
5
Artefacts
Staub and Taylor 1999
5839
Camel Peaks
RS
6
Rock art
Davidson 1952
4438
Wave Rock Scarred
Tree
RS
7
Modified tree,
artefacts
Corsini 1993
4661
Hippo’s Yawn
OHP
7
Rock art
5840
Hyden Rock
RS
7
Rock art, artefacts
Davidson 1952
Glendenning 2004
Peck 1982
21384
Wave Rock Scarred
Tree
OHP
7
Modified tree
Glendenning 2004
21385
Wave Rock Rockholes OHP
7
Gnammas
Glendenning 2004
21386
Wave Rock Isolated
Find
OHP
7
Artefact
Glendenning 2004
21387
Wave Rock
RS
7
5844
Hyden
OHP
8
Gnamma, artefacts
5610
Graham Rock
OHP
9
Gnammas
5611
Hyden-Hyfield
RS
10
Rock art
Glendenning 2004
The sites listed in Table 4.1 are described below; unless cited, information derives
from DPLH site files. The sites clustered on or around Wave Rock are discussed as
one unit since they potentially represent a single site complex. No effort was made to
visit the listed sites (some no longer exist), rather information is reported as it
appears in the site files or heritage survey reports.
Kondinin Pipeline 1: A low-density artefact scatter on and around a salt lakebed,
with two higher-density clusters, the first measuring 551 m2 yielded 300 artefacts and
the second approximately 315 m2 had 90 artefacts. A sample of 20 artefacts
62
comprised 18 quartz – predominantly debris (n=13), with four complete flakes and
one broken flake – along with a quartzite flake and flake fragment. Most artefacts
were small, < 10 mm in maximum dimension; no pieces measured > 30 mm (Staub
and Taylor 1999).
Corrigin-Hyden Road: An artefact scatter adjacent to a granite exposure. Six brown
bottle glass artefacts (five scrapers and one retouched piece) were found in a 10 x
15 m area. Isolated finds along the same road alignment comprised two quartz
flaked pieces and a quartz flake on salt pans/flats, a quartz flaked piece on a creekline, and a worked brown glass piece on the granite itself.
Kondinin Pipeline 5: A low-density artefact scatter measuring 85 x 36 m on a
lunette, adjacent to a salt lake. Quartz debris measuring 15–30 mm dominated the
assemblage. A studied sample comprised eight quartzite artefacts (three flakes, two
pieces of debris, plus a single flake fragment, core and broken flake) and 156 quartz
artefacts (92 pieces of debris, 23 broken flakes, 22 flakes, 14 flake fragments and
five cores); a disturbed area west of the scatter also yielded nine artefacts (Staub
and Taylor 1999).
Kondinin Pipeline 2: A low-density artefact scatter on and around the bed of a saltlake. Two higher density clusters were found on the western shore: one 25 m2 with
50 artefacts, and a two square-metre area yielding 20 artefacts. Quartz debris
dominated both assemblages, mostly ≤ 20 mm in maximum dimension. A sample
comprised three quartzite artefacts (a broken flake, a flake and flake fragment) and
27 quartz artefacts (20 pieces of debris, four flake fragments, two flakes and a
broken flake) (Staub and Taylor 1999). Nearby, Quartermaine (2000) found three
quartz flaked pieces, a quartz flake and a worked glass piece.
Kondinin Pipeline 3: An artefact scatter on and around a salt lake. Some artefacts
were found on the lakebed, but most were concentrated in a band 2 m wide around
the lake edge, approximately 1–2 artefacts/m2. Artefacts were mostly quartz debris,
≤ 20 mm in maximum dimension. A sample of 27 artefacts (all quartz) was classified
as nine broken flakes, nine pieces of debris, six flakes, plus a flake fragment, core,
and a flake subsequently used as a core (Staub and Taylor 1999).
63
Camel Peaks: A rockshelter on granite outcrop, housing red handstencils and a
painted rake-like motif with five tines (Davidson 1952).
Wave Rock: Site 21387 does not refer to specific archaeological content, rather
encompasses the entire Wave Rock area, which Glendenning’s (2004) informants
identified as an important ethnographic site that was used as a camp, hunting place,
and water source. Two rock art sites, Hyden Rock (5840) and Hippo’s Yawn (4661),
occur on the eastern edge of the Wave Rock outcrop. Hyden Rock comprises two
small rockshelters housing red handstencils and a white circle (Davidson 1952); both
shelters have since been damaged by fire and/or graffiti, but motifs are still visible
(Glendenning 2004; Peck 1982). A few white quartz flakes and a quartzite grindstone
were also found nearby. Hippo’s Yawn originally preserved at least three
handstencils, but they were inadvertently removed with overlying graffiti; no artefacts
are present. Other archaeological features around Wave Rock include an isolated
broken quartz flake (site 21386), a series of gnammas (site 21385) and two
scarred/modified trees (sites 4438 and 21384), the first of which was associated with
13 quartz pieces, four of which had clearly been flaked. Corsini (1993) surveyed the
Wave Rock area and located five artefact scatters and 23 isolated artefacts, totalling
60 quartz pieces (23 of which had definitely been flaked) and a possible
hammerstone made on basalt. Only one registered site (4438) could be identified
from his limited information. The remaining cultural material he noted may be
incorporated within or additional to the sites described above.
Hyden: Comprises a deep gnamma (> 1.5 m deep) and some quartz fragments on a
granite exposure.
Graham Rock: Two gnammas were identified on the granite outcrop, but no
artefacts were present on Graham Rock or the neighbouring outcrop, even in
rockshelters where sediment had accumulated.
Hyden-Hyfield: Paintings (no stencils) located in a breakaway rockshelter, on
granite.
64
4.2.2 Other Relevant Research
So little archaeological data exists for the study area and its surrounds that it is
necessary to go further afield for comparative material. Significant research projects
have been conducted in southwestern Australia (e.g. C. Dortch 2000; J. Dortch
2000, 2004; Ferguson 1985; Hallam 1972, 1975), but they are focussed on the
coastal plain or the heavily-forested southwestern corner. Due to the vastly different
environment of the study area – specifically the comparative scarcity of water –
occupation intensity and mobility patterns would not be comparable, and hence a
different archaeological signature should be expected. Therefore, the only studies of
relevance to this research are those that include an inland component, in areas with
similar climate, geology and/or hydrology (e.g. Anderson 1984; Bird 1985; Smith
1993).
Figure 4.2 Locations mentioned in the text.
65
Swan Coastal Plain–Darling Plateau
Anderson (1984) investigated how Aboriginal occupation varied between three
distinct landscapes: the Swan Coastal Plain, the jarrah forests immediately east of
the Darling Scarp, and the Darling Plateau, up to 180 km from the present coastline,
near Yealering (Figure 4.2). She analysed data from almost 450 sites and found that
the size of sites (based on the number of artefacts found, excluding amorphous
debris) and artefact densities varied significantly between the three environmental
zones. On the coastal plain, ‘major’ sites (with > 500 artefacts) were more common
and sites generally had more artefacts than those in the other zones. Assemblages
included a wider range of artefact and raw material types, as well as grinding stones
and greater amounts of debitage (Anderson 1984:34). In contrast, no major sites
were found in the jarrah forests, where minor (< 50 artefacts) sites instead
dominated, but a few intermediate sites (50–500 artefacts) were also present. Inland
sites (all near water sources) fell between these two extremes – major sites were
rare but present, and artefact numbers were consistently lower than at coastal sites.
Artefact assemblages contained a more restricted range of raw materials, and were
quartz-dominated, but dolerite was also fairly common (Anderson 1984:19); more
detailed artefact data were not provided for inland sites.
Anderson (1984:34, 37) related differential site distribution to resource availability
and human mobility and concluded that the Swan Coastal Plain and the inland
Darling Plateau were both inhabited by distinct groups, while the forested area acted
as a transitional zone between them. She argued that the people who lived on the
coastal plain had a regular seasonal round. They aggregated on the coast, or around
larger waterbodies during summer and autumn, to exploit the wealth of aquatic
resources. Food was scarcer in winter, so people broke into smaller groups and
some moved into the forested area – explaining the predominance of small sites
there – while the others ranged more widely among the plain; by late spring,
coastward movement recommenced. In contrast, the more limited resource base
meant that groups living on the Darling Plateau had a less defined seasonal round.
These groups were more mobile, so ranged over longer distances, but also utilised
the forest area more heavily as its swamps held water for longer periods. Due to the
general lack of dated sites, Anderson (1984) did not discuss how long these
occupation patterns had been in place.
66
Bremer Bay–Lake Grace
Bird (1985) studied a transect running from Bremer Bay to Lake Grace (Figure 4.2),
incorporating 180 sites across four environmental zones: coast (n=37), coastal plain
(n=32), intermediate (n=58) and inland salt lake (n=53). Her research aim focussed
on lithic utilisation strategies, specifically the reduction intensity of chert, which was
of limited availability compared to the more ubiquitous quartz. Like Bamforth (1986),
she demonstrated that chert was conserved as distance from source increased,
while quartz reduction was largely unchanged. While many of Bird’s conclusions are
largely irrelevant here, her site distribution data – specifically the sizeable inland
sample – warrants discussion. Bird (1985:96, 150) classified surface artefact scatters
based on artefact density, allowing for the size of the area over which artefacts were
distributed. Her study area yielded 23 major sites (> 1 artefact/m2), 90 minor sites (<
1 artefact/m2), 12 diffuse scatters (< 1 artefact/m2 but over 100,000 m2 in area), 33
isolated artefacts and 22 other sites (burials, quarries, stone arrangements and shell
scatters). She argued that her study area was populated at a density of four people
per 100 square kilometres (Bird 1985:129). Based on those data, and the
assumption that water availability constrained patterns of aggregation and dispersal,
Bird (1985:176) proposed the following seasonal route for her study area:
o
Feb–May: People would cluster in a few large camps, based around
permanent, reliable water sources (gnammas, soaks, coastal dunes);
o
June–September: People would disperse into smaller groups and use more
ephemeral water sources (rock pools, areas with standing water after winter
rains);
o
October–January: Camps varied in size, with people using intermediate
water sources (soaks, small swamps, etc.).
Bird (1985:150) identified 53 sites in the Inland Salt Lake zone: 30 minor, nine major,
six diffuse scatters, seven isolated artefacts, and one quarry/stone arrangement.
Based on geomorphological evidence, she considered that surface artefact scatters
probably post-dated 5000 BP, but none were directly dated (Bird 1985:115). Detailed
artefact data are unavailable, but the assemblages were heavily quartz-dominated,
with small quantities of chert and silcrete (Bird 1985:340–344). Most sites were in
open Eucalypt woodland on valley floors near salt lakes – just one granite outcrop
67
site was represented (Bird 1985:161, 163). Nevertheless, she cautioned that this
does not necessarily represent occupation patterns but rather sampling bias, as she
did not record a representative sample of sites, but instead initially targeted those
areas with known or potential water sources, or where artefacts had been reported
previously (Bird 1995:96).
Sites were more common in the southern part of the Inland Salt Lake zone, where
freshwater lakes and swamps occur (Bird 1985:131, 162). Nevertheless, Bird
(1985:161) noted that there was no strong association between site locations and
water sources. Indeed, based on Bird’s (1985:340–344) data, the average distance
between a site and a known freshwater source was 2.4 km. People generally camp
within 100 m of freshwater (e.g. Gould 1968), but only 13 of Bird’s 53 inland sites
were this close to known water sources. It is possible, then, that the bulk of the sites
represent occupation when surface water was abundant (during winter), as people
dispersed into smaller groups at this time; this may explain the predominance of
minor sites. Sites closer to water sources may have been used between October and
May, with the timing and length of occupation depending on the reliability of
individual water sources. Four major sites were within 100 m of known freshwater
sources. These may represent people clustering there at times when other water
was scarce, but this idea cannot be evaluated using Bird’s (1985) data.
Southern coast–Mount Ridley
Smith (1993, 2011) conducted a regional analysis along a transect of the southern
coast between Esperance and Israelite Bay, stretching inland as far as Mount Ridley
and Mount Ragged (Figure 4.2). The area incorporated three environmental zones:
offshore islands, coast and the interior. She surveyed approximately 10% of her
study area and recorded 300 sites and analysed the data from 217 of them, of which
215 yielded artefacts (Smith 1993:223; Smith 2011). Her systematic survey
methodology and deliberate focus on surveying all landscape types equally meant
that site distribution could be used as a proxy for settlement patterns (Smith
1993:211). Across all three zones, sites were most common on granite outcrops,
where artefact densities were correspondingly higher (Smith 1993:228). She related
the popularity of granite outcrops to their water catchment properties, the availability
of shelter, diversity and predictability of plant foods, and use of high-points as
68
lookouts to monitor weather patterns as well as the movement of people and game
(Smith 1978, 1993:237).
Throughout her study area, most sites were 500–800 m from water sources (known
or potential) and, unsurprisingly, artefact scatters were generally found within 100 m
of gnammas (Smith 2011). It is important to note that while Smith (1993:229)
assumed all salt lakes would have access to soaks of drinkable water, this is not
always the case, as Roe (1836, 1852) found. Of the 193 artefact scatters she
recorded, 86% (n=167) had < 350 artefacts, including debris; only four scatters, all
on the coast, yielded > 1000 artefacts (Smith 2011). Based on the consistently
limited spatial extent of her sites, their regular distribution, and low artefact counts,
Smith (1993:287–288) argued that her study area was occupied by small, highly
mobile groups, and that their mobility was likely a response to resource scarcity and
unpredictability. The few sites with more artefacts may represent areas where people
congregated, possibly with people from other cultural groups.
The inland zone sample comprised 50 artefact sites and two quarries, an average of
one site per 6.25 km2 (Smith 1993:270–272). Nearly 60% of sites were on/around
granite exposures, but salt lakes, rivers/swamps and sandplains were also often
associated with archaeological material (Table 4.2). Most sites were adjacent to thick
vegetation, in which people could take shelter if required (Smith 1993:281). No
detailed artefact data are available for the interior sites, but most (n=46, 88%)
yielded < 350 artefacts and 21 had < 100 artefacts, echoing the small-site pattern
evident across her entire study area. Larger assemblages were restricted to granite
and salt lakes, with the former yielding the only sites with 501–1000 artefacts (Table
4.2). Overall, sites on granite accounted for 75% of artefacts discarded in the interior,
followed by salt lakes (15%) and watercourses (7%) (Smith 1993:271–272).
One interior site, Mount Ridley, also has rock art, including hand stencils, bird tracks,
and complex linear designs; most of the nearly 1000 artefacts were discarded well
away from the rock art. Smith (1993:276) said her informants indicated that Mount
Ridley was a meeting place, where people from Esperance gathered with their
neighbours from the north. Mitchell’s (2016) analysis supported this idea – he
identified silcrete sourced from Ngadju country, to the north, and evidence of intrasite spatial organisation that Conkey (1980) considered an indicator of aggregation.
69
Mitchell (2016:172) also argued that evidence of both Place and Individual
Provisioning indicated repeated use of Mount Ridley over time combined with longrange travel, both of which would be expected at an aggregation site.
Table 4.2 Archaeological data from Smith’s (1993:271) interior zone, showing the number of sites in
different topographic units, and the number of sites falling within different assemblage size categories.
Topography
Number of artefacts
TOTAL
0–30
31–99
100–350
351–500
501–1000
Granite
4
4
18
2
2
30
Salt lake
2
2
1
2
0
7
Rivers/swamp
2
2
2
0
0
6
Sandplain
4
1
2
0
0
7
Dune/gravel
1
0
0
0
0
1
Chert ridge
0
1
0
0
0
1
TOTAL
13
10
23
4
2
52
4.3 ETHNOHISTORICAL BACKGROUND
In 1826 the British founded Albany, and the Swan River colony (now Perth) followed
in 1829 (Stannage 1981). He was not the first to head inland in search of agricultural
land, but Surveyor General J. S. Roe (1852) appears to be the first European who
ventured near the study area – he passed through Karlgarin (Figure 4.2) and just
south of Hyden during an extensive trip southeast of Perth in 1848–1849. In the
1860s, pastoral stations were established at Narembeen and Kellerberrin (Leake
1950). Holland's Track was cut in 1893, and runs from Broomehill to Coolgardie,
passing through King Rocks approximately 30 km north-east of Hyden. Europeans
were in the study area soon after: the words ‘M. Cronin 1894’ were painted inside
Mulka’s Cave, but were subsequently removed (Randolph 1973). While the
surrounding area was farmed from the 1920s, the township of Hyden was not
gazetted until 1932 (Landgate 2015).
Ethnographic and historical documents may provide a wealth of information but, as
McBryde (1979) and Trigger (1982) noted, these sources can be problematic due to
biases held by observers and/or authors, as well as their methods of collecting data.
Several early documents mention the Indigenous inhabitants of southwestern
Australia, but generally focus on information useful to future settlers/explorers, such
as Aboriginal names for specific water sources, or ‘exotic’ subjects like initiation
70
ceremonies, burial rites, kinship systems, spiritual beliefs, and tribal warfare. These
do not present a balanced picture of Aboriginal life (Green 2011; McBryde 1979).
Furthermore, sources make no allowance for the impact of introduced disease on
traditional life (Briscoe 2003; Campbell 2002; Cleland 1928; Dowling 1997; Green
1981) or the disruption of traditional occupation patterns by colonial settlement and
agriculture (Host and Owen 2009; Neville 1948; Radcliffe-Brown 1930). The sources
must therefore be treated with caution.
More recent ethnographic information is also available, predominantly from cultural
heritage surveys. Under Section 18 of the Western Australian Aboriginal Heritage
Act (1972), any development must be preceded by a cultural heritage survey,
involving Aboriginal people with a demonstrated connection to the area. O’Connor et
al. (1989:54) argued that appropriate spokespeople for specific areas should meet at
least one of the following criteria:
1. Long term (normally childhood) association with one or more members of
the ‘pivotal generation of culture transmitters’ – individuals who had, as
children themselves, been exposed to people who passed on their
traditional knowledge;
2. Demonstrated knowledge of an area’s natural resources, including food
plant and animal locations, hunting grounds, camping areas and water
sources;
3. Politically active individuals, who have worked hard to amass traditional
cultural knowledge relevant to a specific area.
The information disclosed by such people can also be biased and problematic, as it
represents only that portion of traditional knowledge that has survived the cultural
dislocation that followed European settlement. Nevertheless, it is a welcome
supplement to the fragmentary documentary record.
4.3.1 Aboriginal Social Boundaries
Much of southwestern Australia is covered by the South West Native Title Settlement
(Figure 4.3); the Traditional Owners of this area all identify as Noongar, but
subgroups also exist. Whereas the current native title boundary is rigid for legal and
administrative reasons, boundaries in the past were likely more fluid, both spatially
71
and temporally. It is generally accepted that, before European arrival, the southwest
was inhabited by a group of people with similar languages and customs. They
shared a common initiation rite, scarification, and did not practice circumcision or
sub-incision, distinguishing them from the eastern groups who did (Bates 1938; Curr
1886, 1887; Tindale 1974; White 1985). The boundary that separated circumcising
from non-circumcising groups was arguably not a static border but rather a westward
moving frontier before its impetus was arrested by European settlement (Gibbs and
Veth 2002). In all but Curr’s (1886, 1887) system, the study area lies within the noncircumcising group (Figure 4.4), but local group names and territories vary. The
occupants of the study area have been identified as Kikkar (Davidson 1938), EastOf-Perth tribe (Hammond 1933 – though his geographical boundaries are vague),
and Njakinjaki (Tindale 1974). Hyden appears to be on the boundary for Curr’s
(1886, 1887) Kokor group, but he noted no adjacent groups. While the inhabitants of
the study area were evidently Noongar, proximity to the (moving) circumcision
boundary means that desert influences were likely active. Social gatherings and
visits were common both within and between tribes (Hammond 1933), and groups
were permitted into other people’s lands provided correct protocols were observed
(Myers 1982). Mechanisms for cultural exchange were certainly in place, so each
tribe/group should not be viewed as isolated entities.
Figure 4.3 Area covered by South West Native Title Settlement, showing subdivisions (Source:
National Native Title Tribunal). ILUA = Indigenous Land Use Agreement.
72
A
B
C
D
Figure 4.4 Aboriginal social boundaries in southwestern Australia according to various sources
(redrawn from original maps by A.M. Rossi). A: Curr (1886,1887) separated groups into the noncircumcising Western Division (green) and the circumcising Central Division (beige). Group names
are given in italics, where known. Red crosses indicate locations where circumcision was practised,
but recorded locations are often vague and their placement on the map may be somewhat arbitrary.
E.g. the cross within Kokor territory is noted in text as ‘east of York’. Original image provided courtesy
of Fisher Rare and Special Collections, University of Sydney. B: Bates (White 1985) placed the
division for her non-circumcising group (the Bibbulmum) further inland. Her eastern boundary
locations are shown in blue. At the time of her research, Karlgarin was the most inland town
established in the vicinity of the study area, so was the only real point of reference. The western
boundary for the circumcising group (Karatjibbin – specific locations shown in red) is considerably
further east. C: Davidson (1938). D: Tindale (1974). Groups east of his circumcision/subincision
boundary are not shown.
4.3.2 Ethnohistory of the Study Area
Ethnohistorical information for the study area is limited, as historical records of inland
travels barely mention Aboriginal people. For example, while Roe (1836)
encountered Aboriginal people around Narembeen and Mt Hampton (Figure 4.2), he
gave few details of the encounter. Most accounts are dominated by the search for
water, and in what locations and quantities it was eventually encountered –
73
unsurprising since several exploration parties nearly came to disastrous ends for
want of water (e.g. Roe 1852). When Landor and Lefroy (1843) were travelling
between Dumbleyung and Lake Grace, their guides refused to venture further east
(closer to the study area), ostensibly due to lack of food and water. This perceived
harshness of the landscape may explain why early settlers believed that Aboriginal
people rarely, if at all, visited the study area, and merely passed through on their way
to better lands (Meeking 1979; Mouritz 1986). However, Landor’s and Lefroy’s
(1843) experience may actually reflect their guides’ unwillingness to travel beyond
their own lands to an area where they may be less familiar with the location of food
and water sources, creating significant risk (Hiscock 1994). Indeed, Landor and
Lefroy turned northwards and shortly found the land far more hospitable, indicating
the variability of inland environs.
Hammond (1933:21, 23) stated that people from the ‘eastern regions’ dealt with
resource scarcity and variability by splitting into small groups and moving around
frequently; similar strategies are predicted under Optimal Foraging Theory (see
Chapter 2.2). During his inland journey, Roe (1836) only encountered Aboriginal
people travelling in small groups of up to ten individuals or noted the remains of a
single hut; around present-day Narembeen he identified the remains of five huts
clustered around a large water hole. The more limited visibility associated with small,
highly mobile groups may also have contributed to early settlers’ belief that
Indigenous people did not systematically occupy the study area.
There is some evidence that ceremonial and social events may have occurred in the
study area, including at Wave Rock:
We call Wave Rock ‘Kutter Kich’. This place is very important to us. This
is on the boundary where a lot of people used to meet for ceremonies.
Our old people walked the trail from the nanuk (neck) of the river at
Guildford to Kutter Kich for thousands and thousands of years. (Winmar
1996:28).
This suggests that the study area may have functioned as a place where occupants
could meet with their western neighbours, and perhaps also with desert groups. A
member of the Kalaako group, the Njaki Njaki’s closest neighbour from across the
circumcision boundary, indicated that his people had ceremonial and trade
relationships with their western neighbours, but that this contact was infrequent
74
(Macintyre et al. 1992). He did not state in whose territory these meetings occurred,
but similar inter-group aggregations have been identified in the archaeological record
(Mitchell 2016). If these meetings also included groups from nearer to the coast, the
study area would have provided a convenient intermediate meeting place between
the west coast and the inland desert areas. The study area also apparently
contained sites of mythological significance, which may have been taboo areas for
women and children; these mythological sites were on or near granite outcrops
(Macintyre et al. 1992). One prominent informant noted that Mulka’s Cave at The
Humps was likely a significant ceremonial or mythological site (Macintyre et al.
1992); another individual noted that when he set up camp near Mulka’s Cave, elders
told him to move further away as he should not stop there (Goode 2011). These sites
may have been used in conjunction with, or separately from, the social gatherings
noted above.
4.4 CONCLUSION
While ethnohistorical sources are limited, they indicate that the study area people
were part of the larger southwestern group, the Noongar, who marked initiation by
scarification, not circumcision. People living in the study area had contact with their
neighbours from as far away as the west coast, as well as those from the
circumcising desert groups (Macintrye et al. 1992; Winmar 1996). Gatherings with
their neighbours possibly occurred at Wave Rock, and other granite outcrops in the
area may have had ceremonial or mythological significance; neither claim can be
addressed with the current suite of archaeological data. Due to the harsh
environment of the study area, its occupants were probably highly mobile and
travelled in small groups, moving across the landscape according to local resource
availability. This suggestion is supported by the archaeological data from the study
area and its immediate surrounds, as well as other parts of inland southwestern
Australia that is dominated by small sites focussed on and around granite outcrops
or salt lakes. While no sites in the area have been dated radiometrically,
Quartermaine (2000) found flaked bottle glass, indicating that Aboriginal people were
using the area after Europeans arrived in Western Australia, contrary to the beliefs of
early settlers (Meeking 1979; Mouritz 1986).
75
CHAPTER 5
CONSTRUCTING THE RESOURCE MODEL
5.1 INTRODUCTION
This chapter details the methods used to construct the resource model for the study
area, beginning with the means of expressing spatial variability in resource
distribution by creating of Landscape Divisions defined by vegetation and landform
characteristics. The process for compiling floral and faunal lists is then outlined,
including the identification of food plants and animals, and the means of determining
their seasonal availability and distribution across the landscape. The procedure for
quantifying the longevity of rainwater preserved within soil profiles and rock
structures is then discussed, allowing for the highly variable rainfall patterns evident
in the climatic data. Finally, the methods of constructing the occupation model are
explained, focussing on the ranking system used to classify plant foods.
5.2 LANDSCAPE DIVISIONS
The study area is small, but there is no reason to assume that food and water
resources would be uniformly distributed across its landscape. It was therefore
necessary to devise some means of representing variation at a manageable level.
This is commonly achieved by dividing a landscape into different units; this approach
is employed in archaeological research (e.g. Bird 1985; Smith 1993) and agricultural
land management (e.g. Murphy-White 2007; Murphy-White and Hauck 1999). The
precise divisions formulated depend on the size of the study area, as well as the
aims of specific studies. For this research, the study area has been divided into
distinct landscape types – referred to hereafter as Landscape Divisions – based on
landform and vegetation characteristics recorded by Beard (1972, 1979, 1980). His
data were selected since they represent the pre-European distribution of native
vegetation, and extensive spatial data are easily accessible via the NRInfo web
application (DPIRD 2016b). Six Landscape Divisions were recognised: Granite,
Heath, Mallee, Saline, Thicket and Woodland – these names are capitalised to
distinguish them from generic vegetation communities or landform characteristics.
Most incorporate a single vegetation community, but the Saline division comprises
76
several that occur on naturally saline soil, as well as salt-lakes themselves (Table
5.1). Mosaic units were classified based on the dominant vegetation form or, where
neither was dominant (e.g. scrub-heath/thicket), the more diverse assemblage. The
distribution and character of each Landscape Division are shown in Figures 5.1–5.2.
Vegetation communities are dictated by soil, topography and hydrology, so the
Landscape Division unit should capture and express variation in all pertinent
resources including food plants, vertebrate fauna, and water; sites within the same
Landscape Division are therefore presumed to share identical resource bases. This
assumption of homogeneity is an oversimplification, albeit commonly employed (e.g.
Smith 1993:203), and it is important to note that variation will occur at a much
smaller scale than can be depicted in Figure 5.2. Indeed, most vegetation types
grade into one another so while firm boundaries are delineated, they are somewhat
artificial. Nevertheless, each Landscape Division should accurately characterise the
dominant vegetation and landform of a given area, which facilitates analysis at a
broad, landscape level (determining how the entire study area could have been used)
as well as permitting site-specific analysis of resource abundance. The latter is
especially significant, as it allows all archaeological sites in the study area – even
those for which scant data are available (e.g. those merely listed on the Aboriginal
Heritage register) – to be better contextualised by an understanding of the local
resource base and the impact it has on occupation patterns.
Table 5.1 Landscape Divisions and their associated vegetation communities (see descriptions in
Table 3.2 and distribution in Figure 3.14) and landform. Vegetation communities are defined following
Beard (1972, 1975, 1976, 1979, 1980).
Landscape
Division
Vegetation community /
landform
Woodland
Woodland
Mallee
Mallee
Heath
Scrub-Heath
Thicket
Broombush Thicket
Granite
Granite outcrops
Saline
Melaleuca Thicket
Succulent steppe
Tea-tree scrub
Salt lakes
77
A
B
Figure removed due to copyright restrictions
C
D
E
F
G
H
Figure 5.1 Vegetation and landform characteristic of each Landscape Division – all photographs and
species information from Beard (1990:119, 121, 123, 131–132, 134, 136). A–B: Saline, dominated by
short, salt-tolerant shrubs (Atriplex spp., Tecticornia spp.). Note the taller Eucalyptus sp. trees on
slightly higher ground in B. C: Eucalyptus spp. Woodland, note the generally open nature of the
canopy and the relatively sparse understorey. D: Mallee, note the multi-stemmed Eucalypts and the
thicker understorey than in C. E–F: Heath, dominated by a diverse range of low shrubs including
Grevillea spp. and Verticordia spp. Note the Mallee in the background of E, indicating the mosaic
nature of Landscape Divisions. G: Thicket of Acacia, Allocasuarina and Melaleuca species – note the
closely spaced shrubs. H: Granite. Note the thicker vegetation at the base of the slope, due to water
runoff. The rock is mostly bare except where pockets of soil support shrub and small tree growth.
78
Figure 5.2 Distribution of Landscape Divisions within the study area – note that the Saline Landscape Division comprises both saline valleys and salt lakes.
Spatial data extracted from NRInfo (DPIRD 2016b). Examples of each Landscape Division are shown in Figure 5.1.
79
5.3 QUANTIFYING PLANT AND ANIMAL FOODS
The resource base undoubtedly affects when and how humans can occupy a given
area. It is well-known that Aboriginal people generally had a highly mobile lifestyle,
and frequently moved around in response to shifts in resource abundance (e.g.
Clarke 2009). It is therefore important not to simply analyse a list of foods available
in a given area, but to consider their seasonal availability and varying distribution
across the landscape. This is not a new concept; various studies have outlined the
seasonal food availability in southern Western Australia. Meagher (1974) compiled
the first comprehensive list of food sources – both plant and animal – exploited by
Aboriginal people in southwestern Australia by conducting an extensive review of
historical sources. Aside from identifying the genera or species described in various
sources, she also noted the month or season in which the document was written, or
the observation made, as an indication of the season during which a particular food
was exploited. Others have attempted similar seasonal analyses for southern
Western Australia (e.g. Bird 1985; Smith 1993; Webb 2000), but Meagher's (1974)
study remains the most detailed. In contrast, few have considered the spatial
variation resource distribution, either neglecting this factor entirely, or considering it
in a more superficial manner. Bird (1985) simply noted those food types that would
be restricted to one or more of her environmental zones – coast, coastal plain,
intermediate and inland salt lake – without providing a comprehensive list of all
available foods. She noted the preferred habitat of some plant food species (Bird
1985: Appendix A – Table 6), but this information was not integrated into her
analysis. Similarly, while Smith (1993) specified the vegetation communities near
each of her sites – and assumed that each represented a distinct resource zone –
she did not quantify the resource beyond identifying those communities in which
some targeted food species occurred (Smith 1993:101).
While the aim of this study is not unique – to quantify the spatio-temporal availability
of plant and animal foods – the approach differs from those described above in its
focus and scale. The intention here was not only to identify the edible plant and
animal species present in the study area, but to produce a finer grained analysis by
defining the distribution of every exploited species across each of the six Landscape
Divisions, considering the seasonal availability of all associated food items.
Therefore, this research sought to quantify the full extent of spatial variation, rather
80
than merely assuming its presence. Similarly, previous approaches focussed on
records of the season in which different food products were used by Aboriginal
people, while the method used here relied on organism-specific information to
determine when each plant or animal food was available. While there is some
overlap between these measures, the period of use does not necessarily represent
the entire window over which a food is available. Indeed, the period of use or nonuse may vary between areas, depending on the relative abundance or dearth of
other food items available at a particular time. For the same reason, taboo foods
were not considered, and indeed there are no relevant data available for the study
area. The methodology employed, outlined in more detail below, was optimised to
produce high-resolution data illustrating spatio-temporal variation in food resource
availability, since this would undoubtedly have placed constraints on human
occupation of the study area.
5.3.1 Compiling Floral and Faunal Lists
Before food sources could be analysed, a list of the plant and animal species found
within each Landscape Division was compiled; invertebrate fauna were omitted as
data were lacking. Much of the study area has been cleared for agriculture, but the
various surveys documenting the remnant native flora and vertebrate fauna in
uncleared areas provided a good source of data. A total of 33 sites (often very large
nature reserves) provided usable data, and most incorporated several vegetation
communities. Beard’s (1972, 1979, 1980) larger scale studies were also of value
since they described individual formations within the study area. The relevant
Landscape Division was identified by comparing vegetation and landscape
descriptions to those in Tables 3.2 and 5.1. Figure 5.2 was not used as a
classification tool since it merely indicates the dominant vegetation and landscape
type and so does not represent variation at a smaller scale. Formations in recently
burnt or otherwise disturbed areas were excluded from analysis.
Community descriptions were then used to compile species lists for each Landscape
Division; floral data were abundant, so vegetation lists were compiled by marking
each species against the Landscape Division/s in which it had been observed.
Faunal data were more limited, and many reports provided general faunal lists rather
than noting which animals were encountered in each vegetation community.
81
Therefore, as well as noting in which Landscape Divisions an animal was observed,
they were also matched against records of animals’ habitat preference or targeted
food sources (e.g. some birds are attracted to flowers of certain plants). This
reduced bias towards those Landscape Divisions overrepresented in more detailed
faunal surveys (e.g. McKenzie et al. 1973) and allowed for faunal mobility, as
animals may have been observed in just one of their characteristic areas (e.g. Halse
et al. 2004). The lists were checked against published and online sources (see list in
Appendices A–C) to reflect taxonomic changes and to remove introduced species or
native animals observed outside their natural range. Genus-level identifications (e.g.
Macropus sp; Eucalyptus spp.) were removed unless no exact species were
identified in a particular Landscape Division, or the named species did not represent
the diversity of species within a Division.
5.3.2 Identifying and Classifying Food Species
To identify potential food species, the floral and faunal lists were compared to
published accounts of plants and animals eaten by Noongar people (e.g. Bindon
1996; Bird and Beeck 1988; Meagher 1974 – see complete list in Appendix B).
These lists are often based on records of explorers and early settlers, particularly
Grey's (1841) diaries, but most only include species that had been named, or those
identified by their descriptions, meaning it is almost impossible to compile a list of all
plant species utilised (Clarke 2003). This does not affect the faunal assemblage, as
Hassell (1975:10) stated that Aboriginal people did not discriminate between animal
species and that 'everything was for food'. A list of animals available in an area, then,
probably closely resembles those that were eaten. The faunal assemblage was
divided into four categories: amphibians, birds, mammals and reptiles, with a subcategory for bird and reptile eggs. Eggs were presumed to occur within the same
Landscape Division as the animals themselves. The echidna (Tachyglossus
aculeatus) also lays eggs, but they are retained in a pouch, so cannot be
encountered independent of the animal; these eggs were not listed separately.
For plants, however, there were other factors to consider – primarily whether the
plant is edible, inedible or, in certain cases, toxic. Given Clarke's (2003) comment
that plant food lists are often incomplete, plants that are known or likely to be edible,
such as members of the Orchidaceae family (Pate and Dixon 1982:222), have been
82
included even where there are no recorded instances of them being eaten.
Considering Aboriginal people could render some toxic items edible, such as
Macrozamia kernels (Beaton 1982; Beck 1992), it is reasonable to assume that any
edible plant would have been eaten. Plant species are listed where they have been
identified in the literature and grouped by genus where the identification was at
genus or family level. Species-level data were prioritised over more general
information. Genus-level data (e.g. Acacia gum is edible) was only applied to species
that were not individually identified in food lists, under the assumption that if an
identified species also provided that food, it would be noted in the literature. Plant
foods were divided into seven categories, defined below:
o
Flowers: flower structure and/or nectar;
o
Fruit: fleshy fruits and the kernels/seeds;
o
Gum: resinous substance produced by trees under stress, generally
through heat and lack of moisture (Nussinovitch 2010:8) and exuded
through injury points;
o
Leaves/roots/stems: main body of the plant, including stems, leaves and/or
fibrous roots (that transport water and nutrients rather than storing them);
o
Lerps/manna: secretions produced by insects sucking sap containing
organic nutrients, produced by photosynthesis (Atkins and Ruan 2010–
2016);
o
Seeds: all seeds from trees, shrubs and grasses except large kernels
contained within fleshy fruits;
o
Storage organs: Stem or root structures (tubers, bulbs, corms, etc.) used to
store nutrients for prolonged periods, usually below the ground surface.
Some species have perennial storage organs, while others consume the
organ to reshoot following a summer dormancy, and spend the growing
season refilling and replenishing it (Brown et al. 2008:7; Pate and Dixon
1982). This affects the period over which certain storage organs would be
available.
5.3.3 Determining Seasonal Availability
The seasonal availability of plant foods and animals is dictated by climate, or the
biology, life cycle and habits of an organism (Table 5.2). Seasonal availability, (rather
83
than season of use), was defined using published sources in the first instance,
prioritising data relating to the study area or southwestern Australia over more
generalised information. Terrestrial animals were present year-round, while
migratory birds were seasonal; eggs were available during the relevant bird and
reptile breeding seasons. Plant food availability was more difficult to determine, as
while the plant itself may be present year-round, the edible product may not.
Availability was determined from species-level information and, when that was
lacking, data were sourced from the same genus. Occasionally, generalisations had
to be made, relating the timing of a particular item’s availability to the flowering
period, as the latter data were abundant (e.g. Paczkowska and Chapman 2000;
Western Australian Herbarium 1998-). The factors governing the availability of all
food items are shown in Table 5.2.
Availability was analysed in monthly intervals. Foods were available if present for
most of a given month. To avoid misrepresentative inflation, two caveats were
imposed. First, only the main period of availability was noted. For example, plants
may have a secondary flowering period that occurs after heavy summer rains, and
animals may have a similarly irregular breeding cycle under unusual climatic
conditions. These periods were omitted in favour of the regular annual cycle. Second,
some plant foods may be available in small quantities throughout the year, but in
significantly greater abundance during a particular season, especially when
availability is dictated by climatic variables. In those cases, the period of availability
has been defined as the period of increased production. This may underrepresent
the diversity of food sources at certain times of year, but the small amounts available
were unlikely to influence occupation patterns in the same way that a seasonal
bounty would. Thus, the period of availability may define the only period during which
a food was available, or the period when abundance greatly exceeded that at any
other time of year. Periods of heightened value (e.g. tubers with high water content
in summer, macropods in better condition through winter) have been identified as
they may affect the timing of resource exploitation.
84
Table 5.2 Factors governing seasonal availability for plant and animal foods, and generalisations
used to determine periods of availability. Where no governing factors were present the items would be
available year-round. Generalisations were employed where data were lacking. For details, see
Appendices B–C.
Category
Plants
Animals
Availability
Flowers
o As per individual genus/species.
Fruit
o As per individual genus/species and/or:
• Unripe: two months after flowering;
• Ripe: three months after flowering;
• Naturally dried (e.g. on tree or ground): one
month after fruiting period ends.
Gum
o Periods of high temperature and limited moisture.
Leaves/roots/stems
o As per individual genus/species.
Lerps/manna
o Periods of increased photosynthesis (higher
temperature, more sunlight hours).
Seeds
o As per individual genus/species and/or:
• Unripe seeds (tree/large woody shrubs): one
month before ripe seeds or two months after
flowering;
• Ripe seeds (small shrubs, herbs and grasses):
one month after flowering.
Storage organs
o Perennial organs: year-round;
o Annual organs: dormant period plus a month
either side, when near maximum size.
Amphibians
o Year-round.
Birds
o Non-migratory: year-round;
o Migratory: as per individual migration patterns.
Bird eggs
o Non-migratory: breeding season;
o Migratory: breeding season, breeding location
preferences and timing of presence in study area.
Mammals
o Year-round.
Reptiles
o Year-round.
Reptile eggs
o Breeding season.
5.3.4 Benefits and Limitations of the Approach
The floral and faunal lists (Appendices A–C) were integral to evaluating the spatiotemporal availability of plant and animal foods, but their limitations must be
acknowledged. Beard (1972, 1979) attempted to quantify the pre-European
distribution of vegetation communities – data that formed the backbone for the
Landscape Divisions used herein – while all floral and faunal surveys recorded the
modern distribution of plant and animal species. This approach therefore applied
85
modern species data to pre-European spatial data, but it is well-known that
European farming and introduced species have been detrimental to the native flora
and fauna (Abbott 2002, 2008; Benson 1991; Bickford and Gell 2005; Dickman 1996;
Witt et al. 2006; Woinarski et al. 2015 – see also Chapter 3.6). Some effects can be
ameliorated: obviously alien species, or those outside their natural range, were
omitted from the lists here, but some may remain. Similarly, extinct species were not
represented; even if the relevant species could be identified, data were too scarce to
assign each to the relevant Landscape Division/s. These omissions will largely affect
the faunal assemblage, as floral species are more likely to endure (see Chapter
3.5.1 and 3.6). Finally, while all reasonable measures were taken to ensure that
floral and faunal species could be confidently assigned to a Landscape Division,
there is no reason to assume that all the species listed will be represented in every
iteration of a particular Landscape Division. The lists are merely indicative of species
that may be encountered in a particular area. In many cases, animals should not be
considered restricted to any particular Landscape Division – they may merely be
more commonly found in certain areas. However, mobile resources are less likely to
impact site patterning (Bird 1985:131), so any introduced bias should not have
serious implications.
Despite the limitations noted above, this approach has several important benefits: it
makes use of the abundant data created as a by-product of agriculture and
maximises its application. While a number of locations contributed floral and faunal
data, they were extended by considering each vegetation community separately,
capitalising on the fact that vegetation is heavily mosaicked at a smaller scale; each
site therefore yielded data that could be applied to multiple Landscape Divisions.
This increased sample size meant that local data were sufficient and should address
an issue identified by Gibson et al. (2004), whereby Beard's structural vegetation
units do not correlate with species composition at a larger scale. It also permitted the
compilation of more comprehensive floral and faunal lists for each Landscape
Division, facilitating a higher resolution analysis that quantified the full extent of
spatial variation in food resource distribution, in contrast to previous studies (e.g.
Bird 1985; Smith 1993). This finer grained analysis allows a more thorough treatment
of factors that may affect or dictate occupation patterns.
86
5.4 QUANTIFYING WATER AVAILABILITY
Water availability is paramount to human occupation of any area – it cannot be
transported as readily as food. As noted in Chapter 3.3, there are no freshwater
lakes or rivers in the study area, and the regional groundwater is saline–hypersaline.
Small pockets of less saline groundwater can be found around granite outcrops, but
these are at depths > 20 m, so wells must be drilled to exploit them (Bettenay and
Hingston 1963; Davidson 1977); these sources were inaccessible to the Indigenous
occupants of the study area. Some plants yield water but are mostly used while
travelling or during extreme drought (Clarke 2011: 80–81); these are not considered
further, as they are unlikely to impact occupation patterns. The main source of
potable water, therefore, was preserved rainfall, herein separated into what falls or
runs onto soil, and that captured by rock structures. This distinction is important, as
water will be lost at different rates according to the storage mechanism. This
research seeks to quantify the longevity of water in various soil profiles and rock
structures, under different rainfall conditions, as the spatio-temporal availability of
these sources probably patterned human occupation.
5.4.1 Representing Rainfall Variability
All potable water sources in the study area are forms of preserved rainfall, so their
availability cannot be evaluated without considering the inter-annual rainfall
variability noted previously (see Chapter 3.4.2). Rainfall variation was generalised as
low, average or high, to characterise winter rainfall and annual totals from the Hyden
meteorological station (BoM ID 010568), for 1929–2017. Category ranges were
calculated assuming that values within one standard deviation of the mean
represented a fairly average winter (97–177 mm) or annual (256–424 mm) total;
values above or below those ranges were classified as high or low, respectively.
Rain falling outside the winter period was considered by quantifying isolated falls of
15 mm and above; they can increase the annual total considerably. Falls were
classified in 10 mm increments, from 15 mm to ≥ 105 mm. Cumulative falls were not
considered as they could not be generalised in the same manner. Indeed, larger
cumulative falls often incorporated one or more isolated falls, so were captured
elsewhere. Six years were selected, primarily to represent variation in winter rainfall
total – as rain is most predictable and reliable during winter – while also taking care
87
to represent some variation in annual rainfall total and the number of isolated falls
(Table 5.3). More recent years were prioritised for data quality and completeness,
because the Hyden station does not record the post-clearing drop in winter rainfall
that occurred in other parts in the Wheatbelt (e.g. Andrich and Imberger 2013). The
Hyden station does not record evaporation data and historical evaporation data are
not stored in the same way as rainfall data. Therefore, average daily evaporation
values were derived from the monthly totals for Kondinin, 60 km west of Hyden (Luke
et al. 1987).
Table 5.3 Years chosen to represent rainfall variability, showing annual and winter rainfall totals (mm)
and classification, and the number of isolated falls > 15 mm, occurring outside winter (from Hyden
station, BoM ID 010568).
Year
Winter rainfall
Annual rainfall
Total
Classification
Total
Classification
Isolated
falls ≥ 15
mm
1992
217.1
High
514.1
High
5
1998
178
High
324.6
Average
1
2009
150.1
Average
295.9
Average
2
2010
69.5
Low
142.1
Low
0
2011
126.4
Average
472.4
High
2
2012
84.2
Low
280
Average
2
5.4.2 Soil Water
Soil water is defined here as that stored temporarily within the soil profile, rather than
that which has entered the groundwater system. While soaks and wells are often
mentioned in the archaeological literature, they are generally near-surface
expressions of fresh groundwater, or subsurface remnants of seasonal watercourses
fed by runoff in sparsely vegetated areas (Brodie et al. 2012; Gould 1968, 1971a,
1984). There are scant references to soil water in the historical literature, but Roe
(1836, 1852) recounts several, often unsuccessful, attempts at digging for water in
inland southwestern Australia. Nevertheless, the archaeological literature has never
attempted to quantify soil water availability, which probably reflects its lower
importance due to higher/more reliable rainfall and/or the presence of other
significant freshwater sources, such as lakes, rivers, or potable groundwater. The
88
lack of such resources in the study area heightens the importance of soil water, so in
order to accurately reconstruct spatio-temporal occupation patterns, the longevity of
this source must be quantified.
Modelling soil water content
The water-holding properties of soils vary due to the interplay of several factors,
including soil texture, profile depth, the presence/absence of permeability barriers
(e.g. less permeable subsoils, as in duplex soils), and salt content (Hillel 2004;
Houser 2003; Moore 2001; Schoknecht and Pathan 2013). These variables influence
permeability and hydraulic conductivity – how quickly soils transport water from the
ground surface into the profile, and the rate at which this water subsequently moves
through the profile – as well as the amount of water a profile can store and the depth
at which it is stored. Therefore, modelling soil water generally requires complex
mathematical or computer models (see review by Ranatunga et al. 2008), as well as
the expertise to correctly apply them. The Soil Water Tool (DPIRD 2016a) differs
because, while it is underlain by extensive mathematical theory devised by Ritchie
(1972), it was designed to be easily accessible. Using a few simple inputs and local
rainfall data, it models the daily Plant Available Water (PAW – Figure 5.3); this tool
was used to quantify the availability of soil water in the study area.
Figure 5.3 The relative water content of various soil types, showing available and unavailable water
(Moore 2001:82). Available water can be extracted by plants, referred to here as Plant Available
Water (PAW). Note that a soil may be holding water, even if none is available for plant use
(unavailable water); coarser soils (those with larger particle sizes) have lower water-storage capacity
than finer soils but require less input before plants are able to extract water.
89
Table 5.4 outlines the specific inputs used in the Soil Water Tool, the most important
of which is the crop coefficient, which defines the rate of crop evapotranspiration
relative to evaporative potential, as dictated by climatic conditions (Allen et al. 1998).
This coefficient varies according to a plant’s developmental stage, but herein a single
coefficient was used throughout to simulate mature vegetation. Accurate coefficients
are readily available for crops (e.g. Allen et al. 1998) but not for other vegetation,
partly due to the greater number of variables such as plant health, density and water
availability (Pereira et al. 2015). The crop coefficient results from the interplay of
various climate- and organism-specific variables, but a major contributor is the Leaf
Area Index (LAI), that quantifies the proportion of ground surface shaded by foliage.
LAI is more influential than overall plant size, as small crops with greater LAI values
(e.g. lettuce, cabbage) have higher crop coefficients than larger tree-crops (e.g.
citrus), where ground-coverage is not as dense (Allen et al. 1998 – Table 12). LAI is
generally lower for natural vegetation, especially in arid or semi-arid areas (Allen et
al. 1998). Therefore, a conservative coefficient of 0.5 was selected, indicating that
the vegetation lost water at around half the rate of atmospheric demand. This
coefficient was applied to all the soil profiles (and the Landscape Divisions they
represent) since plant size is not the defining factor in rate of water loss.
Data from the Soil Water Tool cannot be exported and is only output as a graph.
However, daily values are displayed when the mouse pointer is hovered over the
appropriate part of the chart (Figure 5.4). These data had to be manually entered
into an Excel spreadsheet for analysis; data-entry accuracy was verified by spotchecks and visual comparison of Excel charts and the original output. Analytical
methods were simple – quantifying the longevity of soil-water under various rainfall
conditions, using each modelled soil profile as a direct representation of a particular
Landscape Division. The quantity of Plant Available Water (PAW) was assumed to
broadly represented the amount that could be removed from the soil by digging. For
the Granite Landscape Division, however, there are significant differences between
the modelled sand profile and real-world conditions. In the model, soils receive input
only from rain falling directly onto the ground surface. This represents the real-world
scenario in most Landscape Divisions, but Granite areas receive additional input
from runoff – when water is shed from outcrops onto the surrounding soils. This extra
input heightens the importance of accurately determining profile depth, since deeper
90
profiles have greater water storage capacity. The modelled sand profile is 2.5 m
deep (Oliver and Robertson 2009), but soils around Granite outcrops are generally
only 0.8–1.5 m deep (DPIRD 2016b; Schoknecht & Pathan 2013). Differences likely
exist between other model profiles and the Landscape Divisions they represent, but
under normal conditions storage capacity invariably exceeds input, rendering them
immaterial.
Table 5.4 Soil Water Tool (DPIRD 2016a) data fields and inputs used. A soil profile was selected to
represent each Landscape Division (provided in parentheses after profile) based on generalised soil
characteristics (Beard 1979, 1990; DPIRD 2016b; Murphy-White 2007; Schoknecht and Pathan 2013).
Model years were chosen to represent a variety of rainfall quantities and distributions, as per Chapter
5.4.1. The 1992 and 1998 model periods were abridged due to unusually high Plant Available Water
(PAW) balances occurring at germination, resulting from model mechanics whereby plants begin to
draw on a store of water accumulated under preceding fallow conditions. This does not accurately
represent land under native vegetation, as plants would be drawing water year-round. The April
break-season rule was voided to ensure that germination occurs during winter, so model periods were
more directly comparable. A single crop coefficient was used throughout, to simulate mature native
vegetation.
Field
Description
Input
Weather
station
Selected from a list of BoM and DPIRD
facilities in the Wheatbelt.
o
Hyden BoM Station (010568)
Soil type
Selected from a list of ten idealised soil
profiles defined by Oliver and Robertson
(2009).
o
o
o
o
o
o
Sand (Granite)
Sandy Earth (Heath)
Deep Sandy Duplex (Mallee)
Shallow Loamy Duplex
(Saline)
Gravel (Thicket)
Loamy Earth (Woodland)
Date range
Maximum 12-month period must be
selected if cropped simulations are
required.
o
o
o
o
o
o
20/3/1992 – 31/12/1992
20/4/1998 – 31/12/1998
1/1/2009 – 31/12/2009
1/1/2010 – 31/12/2010
1/1/2011 – 31/12/2011
1/1/2012 – 31/12/2012
Break
season date
Quantity of rainfall required to prompt
germination, specified as a value over
three days after 25 April and/or after 5
June.
o
5 mm of rain over three days
after 5 June
Crop
coefficient
Rate of evapotranspiration relative to
atmospheric demand. Separate values
can be provided for each of four
recognised growth stages: initialisation,
development, mid-stage, late-stage.
o
0.5 (throughout)
91
Figure 5.4 Example output from the Soil Water Tool (DPIRD 2016a), showing soil water (PAW) under
fallow (black line) and cropped conditions (green line), plus daily rainfall (blue bar). Fallow and
cropped values will be in-sync until germination date ('break'); thereafter, modelled crops draw on
stored water accumulated under preceding fallow conditions. An increase in the PAW balance is
registered only when daily input (rainfall) exceeds daily loss (through evaporation/evapotranspiration).
Daily data values (box) are displayed when the mouse is hovered over the relevant point on the chart
output (black filled circle).
Special methods had to be devised to model PAW in Granite soils due to the
additional input and shallower profile depth, as noted above. The precise amount of
additional input varies due to the individual morphology of each outcrop, as well as
rainfall characteristics (Fernie 1930; Laing and Hauck 1997). For consistency, it was
assumed that Granite soils received double the input that other areas did, comprising
direct input (rainfall) plus the same amount again through runoff. On average, sands
can store 90 mm of water per vertical metre of deposit. A Granite profile 1.15 m in
depth (the mid-point between the 0.8–1.5 m depth range for Granite soils) could hold
approximately 100 mm of water. To calculate the daily PAW balance of Granite soils,
it was necessary to determine the starting balance and the subsequent rate of water
loss. These measures are outlined below:
92
o
Starting balance: the quantity of water held in a Granite profile at
germination, or after the profile re-wets. This was calculated by adding the
PAW balance of the sand profile – which incorporates the daily rainfall
minus the quantity of water required to recharge the profile and the daily
loss – to the rainfall amount again, to simulate additional input through
runoff;
o
Daily losses: the amount of water lost from a Granite profile on a daily
basis, through evapotranspiration. Daily losses were calculated for each
day that the sand profile held water (since the losses will be identical for
Granite profiles) across all modelled years and used to calculate an
average daily loss value for each month. Sand profiles never held water in
October or November, so these values had to be interpolated based on
those from other months.
Beginning with the Granite starting balance, subsequent daily PAW balances were
determined by doubling daily rainfall and applying the observed daily loss (from the
sand profile) or, where that value was unavailable, the relevant average daily loss
value. PAW balances in the Granite profile were not permitted to exceed 100 mm, to
reflect the limited profile depth.
Due to the methods employed, a starting balance could only be calculated when the
sand profile responded to a rainfall event, but Granite would likely respond when
sand did not due to the impact of runoff. These responses could not be modelled, but
could be estimated based on the rainfall amount, drying time, and the quantity of
water required to recharge a profile devoid of water (see Chapter 7.2.2).
Benefits and limitations of the approach
It is important to acknowledge the limitations of this approach, both those inherent in
the Soil Water Tool, and those arising from the application of model results to
address the research aims. The Soil Water Tool does not allow for certain barriers to
water infiltration (water-repellence), sodicity (salt content), or other characteristics
that may impact water-storage capacity. However, these variables are minor
compared to those encompassed within the model, so the omissions are unlikely to
alter the results appreciably. The main limitations concern the interpretation of the
93
results, in terms of how applicable specific values are to the real-world scenarios this
research aimed to replicate. For example, it was assumed that any water that can be
extracted by plants (PAW) would also be accessible to humans by digging, but
plants may be more efficient at removing water from the soil; there are no data
permitting this idea to be evaluated. The crop coefficient applied (0.5) may also
introduce errors – at best, it is an accurate indication of the longevity of water under
native vegetation and at worst, merely a measure of how long various soils provide
water to a crop transpiring at half the rate of atmospheric demand. In either case, the
patterns are broadly representative (i.e. the relative longevity of PAW in different
soils in different years) but specific values such as dates, quantities and duration
should be accepted with caution, particularly for Granite areas, where model data
was not used in its original form. Finally, there is no reason to expect soil and
vegetation characteristics to be consistent throughout a single Landscape Division.
Pereira et al. (2015) noted that natural vegetation coefficients rarely represent even
a minority of scenarios, due to the variables noted above. Nevertheless, the
assumption of intra-Division homogeneity is inherent to the use of Landscape
Divisions as analytical units. Where considerable intra-Division variation was
conceivable, this was factored into interpretations.
The benefits of using the Soil Water Tool considerably outweigh the limitations of the
approach used here. The Tool is underlain by extensive theory and mathematical
models, is tailored to Western Australian soil and rainfall conditions, and provides
predictions and calculations that would otherwise be inaccessible to those unfamiliar
with soil physics. It permits a far more robust analysis of the spatio-temporal
availability of soil water than would otherwise be possible by analysing basic soil
characteristics. The data permit the establishment of firm timeframes for the
longevity of soil water in different Landscape Divisions, even if some specifics should
be treated with caution.
5.4.3 Rock Structures
Herein, rock structures are defined as depressions that form naturally through the
erosion of basement rock and permit the temporary storage of rainwater – these are
assumed to be present primarily within the Granite Landscape Division. Two main
forms are recognised: pans are wide, shallow depressions, generally < 100 mm
94
deep; and gnammas, narrow cavities that can reach several metres in depth,
especially if they intersect natural fractures (Campbell 1997). While other forms are
recognised in the geological literature (e.g. Campbell 1997), these types are often
defined based on their mode of formation, while the focus here is on physical
characteristics that would affect the rate of water-loss. Rock structures (particularly
gnammas) were often discussed in the scientific literature, but more often with
reference to their overall capacity (e.g. Gunn 2006; Rossi 2010; Timms 2013; Webb
and Rossi 2008) without regard for how they receive and lose water. In low rainfall
zones such as the study area, capacity alone can be an inaccurate means of
characterising rock structures since, like soils, their capacity may greatly exceed
input.
It is well-known that Aboriginal people used the water stored in rock structures (e.g.
Bayly 1999, 2015; Roe 1836, 1852) – therefore, an attempt was made to quantify the
volume of water captured by gnammas and pans under various input conditions, and
to determine the longevity of the water supplies. However, claims that Aboriginal
people intentionally enlarged depressions (e.g. Bayly 1999; Lullfitz et al. 2017) are
not addressed, since these do not affect the rate of water capture or loss in the
generalised structures considered here.
Modelling the longevity of water in rock structures
Rock structures are closed systems, as they have no interaction with soil or the
groundwater table. Therefore, quantifying input and water-loss is simple using rainfall
and evaporation data, but the latter cannot be used in its original form. While there is
some inconsistency and inaccuracy associated with the Evaporation Pan instrument
(Lowe et al. 2009), the main concern is how accurately the Pan Evaporation rate
represents evaporation in natural bodies of water. Evaporation Pans are raised metal
constructions, with sides that are exposed to direct sunlight – they cannot store heat
in the way that natural bodies of water can (Dingman 1994; Tanny et al. 2008). While
there are many complex methods of estimating evaporation relative to the Pan
Evaporation rate (e.g. Deardorff 1977, 1978; McMahon et al. 2013), the most
common is to formulate a Pan Coefficient, a proportion of the Pan Evaporation rate
that can be applied to various water bodies. Coefficients are not necessarily
consistent year-round, and a unique value may be required for each month or
95
season due to the thermal properties of larger bodies of water (Lowe et al. 2009).
Lakes and reservoirs tend to have higher water temperatures than Evaporation Pans
in winter, and lower in summer, due to this insulating effect (Winter 1981). Globally,
there has been considerable research into developing coefficients for reservoirs and
lakes (e.g. Tanny et al. 2008), to quantify water-loss in agricultural or urban settings.
Australian research is comparatively limited, and an average pan coefficient of 0.79
is generally applied (Lowe et al. 2009). This coefficient cannot be applied to rock
structures, but it forms a valuable baseline measure from which to develop
coefficients for pans and gnammas.
Domingues-Villar et al. (2008) found that shallow rock structures (pans) lost water at
a rate similar to Pan Evaporation, but a precise rate was not specified. Evaporation
pans have exposed sides, and metal transfers more heat to stored water than rock
will – therefore, a coefficient of 0.9 was applied to rock pans herein, as water would
be lost far more quickly than from reservoirs. Unlike reservoirs, gnammas store
water further below the ground surface and have a small surface area, providing
more protection from the wind and sun, two of the main drivers of evaporation (Shaw
2014:64–65). Hillel (2004:421) noted that areas sheltered from wind had evaporation
rates around 20% lower than unsheltered areas. To reflect this, a coefficient of 0.4
was applied to gnammas. In reality, evaporation would likely vary with depth, with the
coefficient being closer to the rock pan measure when water was stored close to the
surface (i.e. when the gnamma was nearly full) but lower where water was stored at
the base of a very deep gnamma. Nevertheless, a coefficient of 0.4 should represent
most scenarios. Seasonal coefficient variation was not considered here, since small
bodies of water will not store heat in the same way that larger ones do (Dingman
1994).
Rock structure dimensions vary greatly (Timms 2013), but morphological
generalisations were required in order to apply the methods used in this research. A
maximum water-storage capacity of 100 mm was imposed on pans due to their
shallow depth, but no limit was placed on gnamma depth. Structures were assumed
to be straight-sided, so dimensions were identical at the top and bottom of the
structure. A sub-category was created to represent structures receiving additional
input: runoff gnammas and runoff pans. These subcategories were necessary since
96
not all structures will receive runoff. As with Granite soils (see Chapter 5.4.2), runoff
structures were assumed to receive twice the daily rainfall value, while regular pans
and gnammas received direct input only. Water-loss was assumed to occur only
through evaporation – animal use was discounted as it cannot easily be quantified.
Similarly, water-loss or conservation by human actions (drinking, using lids to reduce
evaporation) was not considered, as the aim was to quantify the longevity of water
under natural conditions; consumption and conservation are considered elsewhere
(see Chapters 8–12).
Using the coefficients derived above, in conjunction with the relevant rainfall and
evaporation data, it was possible to determine the daily water-balance of each of the
four rock structure types (gnammas, runoff gnammas, pans, runoff pans) across the
six modelled years, using an Excel spreadsheet. Evaluating longevity was simple,
but the use of a spreadsheet permitted more flexibility and allowed conditions to be
altered to better replicate the desired scenarios. For example, the longevity of water
was often affected by large falls that occurred outside winter. These falls could be
voided to investigate the longevity of water accumulating via winter rains. Similarly,
isolated falls of various sizes could be simulated when structures were dry, and the
longevity of water evaluated. The results of these analyses are provided in Chapter 7.
Limitations
The main limitations with this method concern generalisations, and thus how
representative the results may be. Rock structures are undoubtedly more common in
the Granite Landscape Division, but they may occur elsewhere on exposures too
small to be visible at the scale represented in Figure 5.2. Similarly, not all Granite
areas will exhibit all types of rock structures – pans are ubiquitous, but gnammas are
less common. These issues cannot be resolved at a Landscape Division level, but
the presence/absence of different rock structures will be considered when evaluating
the resource base of specific archaeological sites and localities. Similarly, the
generalised dimensions and runoff characteristics of rock structures will not
represent all scenarios. Some structures may receive runoff from a larger area or
may be more protected from evaporation by vegetation or overhanging rock – these
structures would then receive and lose water at different rates. Moreover, an outcrop
will shed varying proportions of individual rainfall events, based on the size and
97
timing of the fall (Laing and Hauck 1997). Nevertheless, generalisations permit broad
trends to be evaluated, and the baseline data can be amended to better represent
the specific morphological and runoff characteristics of rock structures at particular
sites (see Chapters 9–12). Finally, the coefficients used may introduce error, as with
the crop coefficient used when quantifying soil water (see Chapter 5.4.2). While
precise quantities and longevity estimates may therefore be flawed, the patterns
should be representative of the relative longevity of water in various structures,
under various rainfall conditions.
5.5 FORMULATING THE OCCUPATION MODEL
An accurate occupation model cannot solely rely on the diversity of food resources
available in each Landscape Division, but must also consider the nutritional content
and procurement costs (time taken to collect and process particular resource using
traditional technology and methods) associated with different resource types. These
data are used to calculate return rates, generally expressed as the number of
calories that can be obtained per handling hour (Table 5.5). By ranking foods by
return rates, it is possible to determine which may exert more control over
occupation patterns, since higher ranked items should be pursued first, and patches
yielding them should be prioritised over lower ranked patches (Macarthur and Pianka
1966; see Chapter 2.2). For plant foods, the required data were fairly easily
accessible. Published return rates were used where available, i.e. where the
genus/species used in the calculation was representative of, or similar to, those
occurring in the study area (Table 5.5). Otherwise, return rates were calculated
manually, using published handling times for plant food categories in conjunction
with representative nutritional data, either for a genus/species abundant in the study
area or a published average value. Handling times were unavailable for several plant
food categories, so values were inferred by determining whether one kilogram of a
certain food would take longer to collect/process than another for which handling
time is known. In reality, handling time and return rates vary according to the
individual undertaking the task (Cane 1987; O'Connell and Hawkes 1981) and the
species being collected, but this variation could not be incorporated into the model.
98
Table 5.5 Ranked plant food categories, based on return rates (cal./hr) determined by handling time
(hr/kg) and nutritional value (cal./kg). Handling time and nutritional content are given only where
return rates were calculated manually. Asterisks indicate inferred handling times.
Moderate
High
Rank
Resource
category
Handling Nutrition Return
(hr/kg)
(cal./kg) (cal./hr)
Gum
0.25*
2370
9480
o Nutrition – Brand-Miller and Holt 1998
(average for gum).
Lerps/
manna
0.5*
2870
5740
o Nutrition – Brand-Miller et al. 2014:234
(Psylla sp. scale lerps, manna).
Storage
organs
0.5
1290
2580
o Handling time – O'Connell et al. 1983
(average of rates provided).
o Nutrition – Brand-Miller and Holt 1998
(based on averages for bulbs and tubers).
Fruit
0.5
950
1900
o Handling time – O'Connell and Hawkes
1981; O'Connell et al. 1983.
o Nutritional – Brand-Miller and Holt 1998
(average for fruit).
1462
o Bird et al. 2009 (Grevillea eriostachya).
1013
o Nutrition – Brand-Miller and Holt 1998
(based on averages for leaves and
stalks/buds).
538
o O'Connell and Hawkes 1981 (Acacia spp.).
Flowers
Low
Reference
Leaves/
roots /stems
Seeds
0.75*
760
Handling time is even more complex for animal foods due to varying methodologies
– some researchers cite post-encounter returns, while others incorporate tracking
time along with success rates, which vary seasonally (see Bird et al. 2009, 2012).
Published return rates (using a consistent definition of handling time) were not
available for all faunal categories, and could not be calculated manually; furthermore,
prey size is not an accurate indicator of rank (Bird et al. 2009). Therefore, the
distribution of animal types was incorporated into the model but in a less formal
manner. Indeed, fixed resources (plants) are more likely to dictate occupation
patterns than mobile prey (Bird 1985:131, 140).
Based on these data, optimal foraging locations were identified in Landscape
Divisions yielding the greatest diversity of higher ranked resources at a given time,
but allowance was made for instances where diversity was not representative of
abundance, for example, if a dominant genus or species provided a particular food
product. While fauna could not be ranked, their distribution must be considered in
order to accurately characterise the total resource base accessible at a particular
time and place. Faunal distribution was often used to identify optimal patches where
99
plant foods could not, i.e. where two Landscape Divisions had similar suites of plant
food but different faunal assemblages. Animal foods were assumed to be more
influential in seasons when plant foods were less abundant and seasonal
abundances (e.g. waterbirds arriving in winter) were assumed to exert a stronger
control than fauna that were available year-round. Residential sites were presumed
to occur within/near optimal foraging locations, in areas where potable water could
be found.
5.6 CONCLUSION
This chapter has demonstrated how the Landscape Divisions, based on vegetation
and landform characteristics, can represent spatial variation at a manageable level.
Plant and animal foods were identified within larger floral and faunal lists compiled
for each Landscape Division, and the temporal availability of each food item
quantified using published species- or genus-level data and a series of
generalisations, where required. Freshwater drainage is lacking, and groundwater is
saline, so water availability is strongly linked to rainfall, which varies considerably
between years. Inter-annual rainfall variability was represented by selecting several
model years, based on the quantity of winter rain and the distribution of other falls.
These model years were used to evaluate the longevity of water in several different
soil profiles – each representing a distinct Landscape Division – using the Soil Water
Tool (DPIRD 2016a). These model years were also used in a specially devised
method to evaluate longevity of water in rock structures. These methods provide a
useful framework to quantify resource availability across all Landscape Divisions,
under varying climatic conditions. The availability of low, moderate and high-ranked
plant foods indicates optimal foraging locations that, when combined with the
presence of potable water, can indicate which parts of the landscape were used at
different times of the year.
100
CHAPTER 6
PLANT AND ANIMAL FOODS
6.1 INTRODUCTION
This chapter focuses on quantifying the spatio-temporal availability of plant and
animal resources that were identified following the methods outlined in Chapter 5.
Plants and animals are considered separately, as different factors govern their
distribution over space and time. The seasonal availability of each plant food
category is considered, as well as the diversity of species in each Landscape
Division, and how species composition affects the period over which each particular
product was available. These data are then summarised to indicate when plant foods
are most abundant, and where particular types are most commonly found. Animal
foods are evaluated similarly, both the animals themselves and their eggs, where
relevant, while also considering when certain species would be in peak condition and
when others would be more easily encountered and captured. The temporal
distribution of animal foods is then evaluated, before a final discussion outlining how
faunal diversity varies across the study area.
6.2 PLANT FOODS
A total of 467 plant species found in the study area furnished 674 different edible
plant food items (Appendix B.1). Their frequency is shown in Table 6.1, along with a
brief description of their recorded uses for each category. These data indicate that a
diverse range of plant food products were present in the study area, a selection of
which are pictured in Figure 6.1. They are unevenly distributed both seasonally and
spatially, so the availability of each must be considered in the context of the relevant
season and Landscape Division.
6.2.1 Gum
A total of 101 species produce edible gum, comprising 69 Acacia, 30 Hakea, as well
as Eucalyptus occidentalis and Pittosporum phillyreoides – the Acacias are some of
the most prolific gum-producers world-wide (Cribb and Cribb 1975:184). Gum is only
available during summer (Table 6.2), as it is produced when trees undergo stress,
101
normally due to heat and/or drought (Nussinovitch 2010:8). Gum is the highest
ranked plant food due to its high caloric content combined with low handling time, as
it can be exuded in great quantity and requires no processing before consumption
(Figure 6.1b). The greatest diversity of gum-bearing species occurs in Mallee (n=60),
followed by Heath (n=52), Woodland (n=38), Thicket (n=30) and Granite (n=27); only
eight species are present in Saline areas. Acacia dominates in all Landscape
Divisions, but Hakea is also a strong contributor in Mallee and Heath areas.
Table 6.1 Total number of plant food products found in the study area. The number of plant species
providing a particular food item is listed, so the overall total indicates the total number of plant food
items in the study area rather than the number of food-bearing species. A species may yield more
than one edible product. Plant uses are based on Bindon (1996), Low (1991), Meagher (1974) and
Webb (2000); not all uses correspond to each food item and/or species. Food items are ranked from
highest to lowest, as per Table 5.5.
Plant food item
No.
species
o Eaten in natural state;
o Blended with other ingredients to make drinks;
o Gum made into balls or cakes for future use.
101
Lerps/manna
o Eaten in natural state;
o Blended with other ingredients to make drinks.
46
Storage organs
o Eaten raw or roasted;
o Species with high water content eaten to quench thirst.
68
Fruit
o Flesh eaten when fruits ripe;
o Unripe fruits roasted and eaten;
o Flesh eaten or stored once dried (intentionally or
naturally);
o Large seed or kernel inside fruit occasionally eaten,
raw or roasted.
50
Flowers
o Occasionally eaten whole or with other parts of plant;
o Nectar sucked out of flowers;
o Flowers soaked to make sweet drink.
252
Leaves/roots/stems
o
o
o
o
Leaves/stems eaten raw, roasted or steamed;
Tender leaf base sole edible part for some species;
Roots cooked;
Root bark pounded into flour and mixed with other
ingredients.
38
Seeds
o Ripe seeds eaten raw;
o Ripe seeds ground into flour, mixed with water and
formed into damper or cake that was then baked in
ashes;
o Certain unripe seeds roasted to render them edible.
119
High
Gum
Moderate
Low
Recorded uses
TOTAL
674
102
Table 6.2 Spatio-temporal availability of all plant food categories within the study area, showing the number of species providing specific products during
particular months. Foods are ordered by rank, from highest to lowest.
HEATH
GRANITE
Summer
Winter
Spring
Dec
Jan
Feb
Mar
Apr
May
Jun
Jul
Aug
Sep
Oct
Nov
Gum
27
27
27
0
0
0
0
0
0
0
0
0
Lerps/manna
3
3
3
0
0
0
0
0
0
0
0
0
Storage organs
49
49
49
49
10
10
10
10
10
10
10
49
Fruit
14
14
9
8
5
6
9
10
12
12
12
14
Flowers
37
17
17
3
5
5
11
16
33
61
61
49
Leaves/roots/stems
14
14
14
15
15
15
15
15
15
15
15
15
Seeds
33
32
14
13
13
14
14
14
29
35
35
35
TOTAL
177
156
133
88
48
50
59
65
99
133
133
162
Gum
52
52
52
0
0
0
0
0
0
0
0
0
Lerps/manna
15
15
15
0
0
0
0
0
0
0
0
0
Storage organs
14
14
14
14
6
6
6
6
6
6
6
14
Fruit
22
27
16
15
8
8
16
16
17
17
17
18
Flowers
111
57
57
15
19
22
36
47
86
156
156
131
7
7
7
8
8
8
8
8
8
8
8
8
Leaves/roots/stems
MALLEE
Autumn
Seeds
60
60
32
32
32
33
33
33
58
61
61
61
TOTAL
281
232
193
84
73
77
99
110
175
248
248
232
Gum
60
60
60
0
0
0
0
0
0
0
0
0
Lerps/manna
32
32
32
0
0
0
0
0
0
0
0
0
Storage organs
9
9
9
9
5
5
5
5
5
5
5
9
Fruit
17
20
11
11
7
7
15
15
16
16
16
17
Flowers
117
76
76
11
14
15
29
37
69
153
152
130
Leaves/roots/stems
9
9
9
10
10
10
10
10
10
10
10
10
Seeds
67
67
26
26
26
27
27
27
65
68
68
68
TOTAL
311
273
223
67
62
64
86
94
165
252
251
234
103
Table 6.2 Continued
WOODLAND
THICKET
SALINE
Summer
Autumn
Winter
Spring
Dec
Jan
Feb
Mar
Apr
May
Jun
Jul
Aug
Sep
Oct
Nov
Gum
8
8
8
0
0
0
0
0
0
0
0
0
Lerps/manna
17
17
17
0
0
0
0
0
0
0
0
0
Storage organs
11
11
11
11
6
6
6
6
6
6
6
11
Fruit
11
10
7
3
1
3
4
5
7
8
9
10
Flowers
49
41
36
0
0
0
3
6
16
54
53
51
Leaves/roots/stems
29
29
29
29
29
29
29
29
29
29
29
29
Seeds
12
11
4
4
4
5
6
6
14
14
14
13
TOTAL
137
127
112
47
40
43
48
52
72
111
111
114
Gum
30
30
30
0
0
0
0
0
0
0
0
0
Lerps/manna
16
16
16
0
0
0
0
0
0
0
0
0
Storage organs
7
7
7
7
3
3
3
3
3
3
3
7
Fruit
13
16
12
11
4
4
8
9
9
9
9
10
Flowers
84
49
49
13
15
18
30
39
61
112
112
97
Leaves/roots/stems
6
6
6
7
7
7
7
7
7
7
7
7
Seeds
34
34
16
16
16
16
16
16
32
34
34
34
TOTAL
190
158
136
54
45
48
64
74
112
165
165
155
Gum
38
38
38
0
0
0
0
0
0
0
0
0
Lerps/manna
25
25
25
0
0
0
0
0
0
0
0
0
Storage organs
15
15
15
15
3
3
3
3
3
3
3
15
Fruit
15
15
11
10
6
7
9
10
13
14
14
16
Flowers
68
48
48
4
5
5
12
19
35
84
84
75
Leaves/roots/stems
11
11
11
12
12
12
12
12
12
12
12
12
Seeds
40
40
14
13
14
15
15
15
40
42
42
42
TOTAL
212
183
153
54
40
42
51
59
103
155
155
160
104
B
D
C
A
I
Figure removed due to copyright restrictions
G
F
E
K
M
H
J
L
N
Figure 6.1 Examples of plant foods found in the study area – photographs and information sources as
cited. A: Acacia acuminata, 3–5 m tall, bearing edible seeds and gum (Bindon 1996:251). B: Edible
gum exuded from Acacia spp. (Low 1991:152). C: Typical Acacia seed pod (Sweedman and Merritt
2006:176). D: Woody fruits (2–4 cm long) containing edible seeds of Allocasuarina humilis (Erickson
et al. 1979:21). E–F: Eucalyptus redunca and its flowers; tree to 6 m tall (Bindon 1996:131). G–H:
Lerps on Eucalyptus spp. leaves (Low 1991:153). I: Chamelaucium megalopetalum, edible flowers
10–15 mm wide (Erickson et al. 1979:88). J: Enchylaena tomentosa fruits, approximately 5 mm wide
(Low 1991:167). K: Thysanotus patersonii tubers, generally 10–30 mm long and 5 mm wide (Bindon
1996:251). L: Carpobrotus virescens, fruits and leaves (up to 65 mm long) are both eaten (Daw et al.
1997:4–5). M: Atriplex vescaria shrub (≤ 1 m high) which bears edible leaves (Bindon 1996:46). N:
Edible fruits of Santalum acuminatum, tree grows to 5 m, fruit 20–40 mm wide (Daw et al. 1997:52–
53).
105
6.2.2 Lerps/Manna
Lerp refers to tiny deposits on Eucalyptus leaves, comprising starchy and sugary
excretions of sap-sucking psyllids (Figure 6.1g–h). Other insects cause trees to
exude sugary liquids that harden into flakes, called manna (Low 1991:153). Both
substances occur only on Eucalyptus and are produced as a result of insects feeding
on sap, specifically phloem sap, which carries organic nutrients within a liquid
substrate. Phloem sap is produced in leaves and shoots as a result of
photosynthesis and transported to non-photosynthetic parts of the plant (Atkins and
Ruan 2010–2016). Photosynthesis requires sunlight, and while it occurs year-round
in evergreen trees, rates are significantly increased where temperatures are higher
and there are more sunlight hours per day (Bell and Williams 1997); both of these
measures peak in summer. Therefore, the availability of lerps/manna is restricted to
the summer months (Table 6.2) following the methodology employed herein (see
Chapter 5.3.3).
According to Low (1991:153), lerps and manna were important foods to Aboriginal
people, and could be gathered in considerable quantity. These foods are ranked
highly due to their high nutritional value and low handling time (Table 5.5). A total of
46 Eucalyptus species in the study area produce lerps/manna – the remaining three
species may also have provided these foods, but no data were available. The
greatest diversity of lerps/manna-bearing species is found in Mallee (n=32) and
Woodland (n=25) areas. Heath, Thicket and Saline areas all preserve 15–17 species,
while only three are present in Granite.
6.2.3 Storage Organs
A total of 68 plants supplied edible storage organs, dominated by Caladenia (n=15),
Drosera (n=9), Platysace (n=7) and Thelymitra (n=6). Storage organs provide the
best returns of all moderately ranked plant foods, as their good nutritional content is
combined with short handling time (Table 5.5). A two-phase availability period is
evident as some species have perennial storage organs, while others exhaust their
organ to reshoot following a period of summer dormancy (Figure 6.2). Their organs
are then replenished over the subsequent growing season, but this replacement
effectively renders them unavailable for that period.
106
60
Storage organs
Granite
Mallee
Thicket
Heath
Saline
Woodland
50
no. species
40
30
20
10
0
Dec
Jan
Feb
Mar
Apr
May
Jun
Jul
Aug
Sep
Oct
Nov
Jun
Jul
Aug
Sep
Oct
Nov
Jun
Jul
Aug
Sep
Oct
Nov
Jun
Jul
Aug
Sep
Oct
Nov
30
Fruit
25
no. species
20
15
10
5
0
Dec
Jan
Feb
Mar
Apr
May
180
Flowers
150
no. species
120
90
60
30
0
Dec
Jan
Feb
Mar
Apr
May
90
Seeds
75
no. species
60
45
30
15
0
Dec
Jan
Feb
Mar
Apr
May
Figure 6.2 Plant food categories demonstrating strong temporal variation in availability, showing
monthly species diversity within each Landscape Division. The same colour key applies to each
chart.
107
During the first phase, April–October, only the 18 species with perennial storage
organs are available. Diversity is greatest in Granite areas (n=10), followed by Saline
and Heath (n=6 each), Mallee (n=5), Woodland and Thicket (n=3 each). From
November–March, an additional 50 species are available – those that spent the
previous months refilling their storage organs in preparation for the impending
dormant period. In this second period, Granite areas still yield the most diverse
assemblage (n=49), and no other Landscape Division preserves < 25% of the
species found in the study area. However, plants with perennial storage organs now
form but a small proportion of the overall assemblage so, elsewhere, the spatial
pattern of diversity shifts. Woodland areas now preserve the second most diverse
assemblage (n=15), closely followed by Heath (n=14), while Saline, Mallee and
Thicket all yield less diverse assemblages (n=7–11). Five species have high watercontent during summer; 3–4 of these can be found in most Landscape Divisions, but
only two are present in Woodland and Saline areas (Appendix B).
6.2.4 Fruit
Edible fruits are provided by 50 species representing 18 genera, the most dominant
being Leucopogon (n=11), Persoonia (n=8) and Cassytha (n=6). Most fruits are
small, but the quandong (Santalum acuminatum) can measure 20–40 mm in
diameter (Figure 6.1n). A total of 18 species, including Leucopogon spp. belong to
the Eriaceae family that produce very small fruits less < 10 mm in diameter, with
hard stones surrounded by sweet pulp (Low 1991:128). Fruit rank moderately among
plant foods, since their small size increases handling time. Overall diversity is
greatest in Heath areas (n=32), followed by Mallee (n=26), Woodland (n=21),
Granite and Thicket (n=19 each). Saline areas supply the fewest species, just 14.
This pattern holds true for most of the year, but some seasonal variation is evident.
Fruits diversity peaks during December–January, when 44 of the 50 fruit species are
available, 10–27 of which are found in each Landscape Division (Figure 6.2). The
assemblage is most diverse in Heath areas, unsurprising since Eriaceae, 16 of which
produce fruit during summer, are common in heath formations. Diversity decreases
in February and March, as Leucopogon no longer produces fruit – at this time,
Thicket areas yield the second most diverse assemblage after Heath. Fruit
availability reaches its minimum in April, when only 12 species are producing fruit,
108
and no more than eight can be found in a single Landscape Division. The most
diverse assemblages are present in Heath, followed by Mallee, Woodland, Granite,
Thicket and Saline areas. This pattern persists thereafter, but fruits become more
abundant from June, when Leucopogon are again available. At this time, diversity
values are more clustered – Heath and Mallee yield similarly-sized assemblages
(15–16 species), as do Woodland, Granite and Thicket (8–9 species). This clustering
becomes less pronounced by late winter and values are more evenly spread
throughout spring, as the availability of fruit increases. The increase is particularly
pronounced in Granite and Woodland areas, whose late spring diversity values
matched or exceeded those of summer, when diversity peaked in all other
Landscape Divisions.
6.2.5 Flowers
Flowers comprise the most abundant form of plant food, with 252 species
representing 31 genera. The assemblage is heavily dominated by Eucalyptus (n=47),
Melaleuca (n=47), Hakea (n=30), Grevillea (n=25) and Verticordia (n=22); Grevillea,
Melaleuca and Verticordia yield no edible products aside from their blooms. Flowers
were occasionally eaten whole, or with other parts of the plant, but most were
exploited purely for their sweet nectar. Flowers are a moderately ranked resource
because most species only produce a small amount of nectar in each flower,
increasing handling time and limiting returns (Table 5.5). The most diverse
assemblages are found in Heath (n=158) and Mallee (n=155) areas, followed by
Thicket (n=114), Woodland (n=86), Granite (n=62) and Saline (n=54). Temporal
variation is pronounced (Figure 6.2) – many species flower in spring and summer,
but several produce flowers year-round, or at other times of the year. The distinct
flowering window exhibited by each genus or species, combined with the varying
representation of genera create a unique pattern of seasonal flower availability within
each Landscape Division.
A total of 189 plants flowered at some point during summer, including every species
from four of the dominant genera; only Hakea were not flowering. Unsurprisingly,
then, summer diversity was greatest in those areas with the highest number of total
species, namely Mallee and Heath, followed by Thicket, Woodland, Saline and
Granite. In all cases, diversity was greatest in December, when 37–117 species
109
flowered in each Landscape Division. Diversity then decreases in January, when
Verticordia and most Grevillea are no longer flowering, and reaches its minimum in
March, at which point Eucalyptus and Melaleuca are no longer in bloom. Just 26
species produced flowers at some point during autumn, including ten Banksia that
flower throughout the year. There is a slight increase in diversity as autumn
progresses, but diversity is greatest where the Banksia genus is well-represented,
specifically Heath (n=15–22), Thicket (n=13–18) and Mallee (n=11–15); Woodland
and Granite provided no more than five species per month, and no plants in Saline
areas flowered in autumn.
From June–August diversity increased, as Grevillea and Hakea begin to flower at
some point during winter. This increase is especially pronounced in Heath, Mallee
and Thicket areas (61–86 species), where those genera are well-represented. In
contrast, other Landscape Divisions yielded more reduced assemblages, between
16 and 35 species. By September, overall diversity reaches its peak, as several
dominant genera begin to flower, including Eucalyptus (n=46) and all Melaleuca
(n=47) and Verticordia (n=22). Diversity is greatest in Heath and Mallee areas,
where over 150 species are flowering, and particularly limited in Granite and Saline,
where fewer than 65 are. Species diversity declines slightly throughout spring, as
Hakea (n=30) do not flower after October; the decline is most pronounced in areas
where that genus is well-represented, but the inter-Division diversity rankings remain
largely unchanged.
6.2.6 Leaves/Roots/Stems
Leaves/roots/stems is the least diverse plant food category in the study area – just
38 species are present, including six Atriplex and five Frankenia. Despite the limited
diversity, leaves/roots/stems are available throughout the year (Table 6.2). Only one
species – Thysanotus patersonii – exhibits seasonal variation due to summer
dormancy, at which point the above-ground component of the plant dies off (Daw et
al. 1997). Leaves/roots/stems are a low-ranked resource as their low nutritional
content means that large quantities must be collected to provide sufficient calories,
increasing handling time (Table 5.5). This resource is most diverse in Saline areas,
which provided a total of 29 species, 16 of which were not present in any other
Landscape Division. Diversity and exclusivity are pronounced here, because many
110
plants within the leaves/roots/stems category are adapted to survive in salt-affected
areas. Less than half the study area assemblage was present elsewhere – just 15
species were found in Granite areas, 12 in Woodland, 10 in Mallee, eight in Heath
and seven in Thicket areas.
6.2.7 Seeds
After flowers, seeds are the most diverse plant food category found in the study area.
A total of 119 species from 11 genera produce edible seeds. The assemblage is
heavily dominated by Acacia (n=70; Figure 6.1c) and Hakea (n=30). While seeds
have good caloric value, particularly Acacia, the nutritional benefits are offset by the
increased handling time required for collection and processing (Table 5.5). The
seed-production periods of the dominant genera – one of which lasts for six months
while the other occurs year-round – dictate the temporal variation in this food group
and creates two distinct six-month availability periods (Figure 6.2).
During the first period, February–July, seeds are comparatively rare. The bulk of the
assemblage comprises Hakea (n=30) and Allocasuarina (n=9); their seeds are
available year-round as they are encased in woody fruits that persist on the plant for
a year or more (Wildlife Society of Western Australia 2013). Just three species have
shorter seed-bearing windows centred around autumn and/or winter. Overall
diversity was therefore skewed towards areas where the Hakea and Allocasuarina
were well-represented, primarily Heath (n=33), followed by Mallee (n=27), Thicket
(n=16), Woodland (n=15), Granite (n=15) and Saline (n=6). Seeds are abundant
from August–January, when up to 116 different species produce seeds, including
most of the Acacia genus. The peak occurs from September to November, where
each Landscape Division yields 14–68 different types of seeds. The more
heterogenous assemblage alters the spatial pattern of diversity, with the greatest
assemblage found in Mallee (n=68) and Heath (n=61) areas, followed by Woodland
(n=42), Granite (n=35), Thicket (n=34) and Saline (n=14) areas.
6.2.8 Summary
Overall, plant foods are most diverse in Mallee and Heath areas (360 and 341 items,
respectively) and scarcest in the Saline Landscape Division, which yields < 150 plant
foods (Figure 6.3). However, there is strong temporal variation in plant food
111
availability (Figure 6.2), so rather than focussing on overall diversity it is important to
consider what items are available in different seasons, how these items are ranked,
and where the greatest diversity and/or quantity can be found. Plant foods are most
abundant in summer, peaking at 585 different items in December. More importantly,
all seven plant food categories are available, including the highly ranked gum and
lerps/manna that are unavailable for the rest of the year (Figure 6.4). Lerps/manna is
derived from Eucalyptus, so would be most abundant in Woodland and Mallee areas
where that genus dominates. Both of these Divisions also preserve a diverse range
of gum-bearing plants, but these are restricted to the understorey, so would be
smaller and less common. Despite the limited species diversity (n=30), Thicket areas
would yield the greatest quantity of gum, since Acacia is one of the three dominant
genera and a prolific gum-producer (Table 3.2; Cribb and Cribb 1975:184). A more
varied assemblage is present in Heath areas, but the quantity of gum would be
reduced due to a more diverse floral assemblage combined with smaller plant sizes.
Moderate-ranked summer foods comprise storage organs, fruits and flowers. The
latter two categories are abundant in Mallee and Heath, while Granite provides three
times as many plants with storage organs – which have the highest returns of the
moderately ranked plant foods (Table 5.5) – than all other Landscape Divisions. Five
species also yield high water content in summer, and Mallee, Heath and Granite
areas each yield three or four of these (Appendix B).
400
350
Flowers
Fruit
Gum
Leaves/roots/stems
Lerp/manna
Seeds
Storage organs
300
no. items
250
200
150
100
50
0
Granite
Heath
Mallee
Saline
Thicket
Woodland
Figure 6.3 Total number of different plant foods within each Landscape Division.
112
All plant foods
700
600
Flowers
Fruit
Gum
Leaves/roots/stems
Lerp/manna
Seeds
Storage organs
500
no. items
400
300
200
100
0
Dec
Jan
Feb
Mar
Apr
May
Jun
Jul
Aug
Sep
Oct
Nov
Plant foods by rank
700
Low
Moderate
600
High
no. items
500
400
300
200
100
0
Dec
Jan
Feb
Mar
Apr
May
Jun
Jul
Aug
Sep
Oct
Nov
Figure 6.4 Total number of plant food items (top) and rank (bottom) available in each month across
the entire study area.
By autumn, plant foods are much scarcer – highly ranked items are no longer
available and most other categories exhibit a marked decline in species diversity.
From April, moderately ranked foods are restricted to plants that retain their
subsurface storage organs year-round (n=18) as well as 23 species with edible
flowers, and 13 with fruit. As in summer, these foods are most diverse in Heath,
Mallee and Granite areas, though the latter mostly yields storage organs, which are
also fairly diverse in Saline areas. At this time, both low-ranked items
(leaves/roots/stems and seeds) comprise one-third of the available plant foods; most
are
available
year-round,
rather
than
coming
into
season
in
autumn.
Leaves/roots/stems are most abundant and diverse in Saline areas, while a range of
seed-producers can be found in Heath and Mallee. Thicket would provide a more
homogenous assemblage, but seeds could be collected in great quantity as one of
the three dominant genera (Allocasuarina) retains seeds year-round. As winter
approaches plant foods are still comparatively scarce, but as the season progresses
more fruits and flowers become available. The latter are abundant in Thicket areas
113
and the former in Heath and Mallee, but Woodland and Granite also provide a good
range of fruits. By spring, moderately ranked foods are present in good supply, but
the increase is primarily due to the flower category, which is well-represented in
Heath, Mallee and Thicket areas. Lower-ranked seeds are also increase in
abundance in those same Landscape Divisions. By November, storage organs are
again widespread and diverse as many species have replenished their subsurface
energy stores over the growing season, in preparation for summer dormancy.
Storage organs would once again be an important food source, concentrated in
Granite areas.
6.3 ANIMAL FOODS
A total of 142 vertebrate fauna species are present in the study area, comprising 89
birds, 28 reptiles, 20 mammals and five amphibians (Appendix C). As well as the
meat from these animals, eggs were available from 85 bird and 23 reptile species.
Each of these animal types, and the resultant food products, are discussed
separately below, in order of abundance.
6.3.1 Birds and their Eggs
A total of 89 bird species are present, representing 15 orders (Appendix C); 85 of
these birds breed in the study area. Passeriformes (perching birds) dominate the
assemblage, but waterbirds (n=9) and cockatoos/parrots (n=8) are fairly common;
emu and malleefowl are also present (Figure 6.5). Most species are non-migratory
(n=68), and inhabit the study area for the entire year, while the remainder (n=21) are
rarely present for more than six months of the year. It is these migratory patterns that
create the seasonal variation in the avian assemblage, but temporal variation in each
Landscape Division is more subdued (Table 6.3; Figure 6.6). Woodland and Mallee
always yield the most diverse assemblages, while Heath, Granite and Thicket are
closely grouped, yielding around 15–20 fewer species than Woodland. The least
diverse assemblages are generally found in Saline areas, but waterbirds (n=9) are
restricted to this Division – the three sedentary species are likely confined to the
more permanent saline waters in the eastern portion of the study area. Emu
(Dromaius novaehollandiae) are ubiquitous, while malleefowl (Leipoa ocellata) are
restricted to Mallee, Heath and Thicket areas.
114
A
B
C
F
G
Figure removed due to copyright restrictions
D
H
E
I
J
Figure 6.5 Examples of birds found in the study area. Measurements refer to the body length (from
crown to tip of the tail) unless noted otherwise. Common names follow species name. Photographs
and measurements from Nevill (2014:52–55, 78–81, 90–91, 106–111). A–C: Passeriformes (perching
birds): Phylidonyris niger, white-cheeked honeyeater, 18 cm (A); Pomatostomus superciliosus, whitebrowed babbler, 21 cm (B); Petroica goodenovii, red-capped robin, 12 cm (C). D–G: Waterbirds:
Himantopus himantopus, black-winged stilt, 37 cm (D); Tadorna tadornoides, Australian shelduck, 56–72 cm (E); Charadrius ruficapillus, red-capped plover, 15 cm (F); Vanellus tricolor, banded lapwing,
27 cm (G). H: Dromaius novaehollandiae, emu, 160–190 cm tall. I: Leipoa ocellata, malleefowl, 55–60
cm. J: Barnardius zonarius, Australian ringneck, 36 cm.
115
Table 6.3 Spatio-temporal availability of all animal food categories within the study area, showing the number of species providing a certain product during
each month, across all Landscape Divisions.
MALLEE
HEATH
GRANITE
Summer
Autumn
Winter
Spring
Dec
Jan
Feb
Mar
Apr
May
Jun
Jul
Aug
Sep
Oct
Nov
Amphibian
4
4
4
4
4
4
4
4
4
4
4
4
Bird
48
47
46
46
47
47
48
49
50
50
49
49
Bird eggs
16
4
0
2
2
1
4
15
39
46
43
28
Mammal
9
9
9
9
9
9
9
9
9
9
9
9
Reptile
15
15
15
15
15
15
15
15
15
15
15
15
Reptile eggs
11
0
0
0
0
0
0
0
0
11
11
11
TOTAL
103
79
74
76
77
76
80
92
117
135
131
116
Amphibian
1
1
1
1
1
1
1
1
1
1
1
1
Bird
50
49
48
47
49
49
50
52
53
52
52
51
Bird eggs
18
5
0
2
2
1
4
15
40
48
46
29
Mammal
11
11
11
11
11
11
11
11
11
11
11
11
Reptile
21
21
21
21
21
21
21
21
21
21
21
21
Reptile eggs
16
0
0
0
0
0
0
0
0
16
16
16
TOTAL
117
87
81
82
84
83
87
100
126
149
147
129
Amphibian
0
0
0
0
0
0
0
0
0
0
0
0
Bird
61
60
59
58
60
62
63
65
67
65
65
62
Bird eggs
19
6
0
2
1
1
5
20
53
62
57
36
Mammal
14
14
14
14
14
14
14
14
14
14
14
14
Reptile
23
23
23
23
23
23
23
23
23
23
23
23
Reptile eggs
19
0
0
0
0
0
0
0
0
19
19
19
TOTAL
136
103
96
97
98
100
105
122
157
183
178
154
116
Table 6.3 Continued
Summer
WOODLAND
THICKET
SALINE
Dec
Jan
Autumn
Feb
Mar
Apr
Winter
May
Jun
Jul
Spring
Aug
Sep
Oct
Nov
Amphibian
0
0
0
0
0
0
0
0
0
0
0
0
Bird
40
39
39
38
39
40
45
46
50
49
47
44
Bird eggs
12
3
0
1
2
2
6
18
44
43
39
26
Mammal
2
2
2
2
2
2
2
2
2
2
2
2
Reptile
4
4
4
4
4
4
4
4
4
4
4
4
Reptile eggs
3
0
0
0
0
0
0
0
0
3
3
3
TOTAL
61
48
45
45
47
48
57
70
100
101
95
79
Amphibian
1
1
1
1
1
1
1
1
1
1
1
1
Bird
45
44
44
44
46
47
48
49
49
49
48
46
Bird eggs
15
4
0
2
1
1
3
15
39
47
43
23
Mammal
7
7
7
7
7
7
7
7
7
7
7
7
Reptile
8
8
8
8
8
8
8
8
8
8
8
8
Reptile eggs
6
0
0
0
0
0
0
0
0
6
6
6
TOTAL
82
64
60
62
63
64
67
80
104
118
113
91
Amphibian
0
0
0
0
0
0
0
0
0
0
0
0
Bird
66
65
63
61
62
64
65
67
69
68
69
67
Bird eggs
21
5
0
2
2
2
5
18
54
65
60
40
Mammal
17
17
17
17
17
17
17
17
17
17
17
17
Reptile
23
23
23
23
23
23
23
23
23
23
23
23
Reptile eggs
19
0
0
0
0
0
0
0
0
19
19
19
TOTAL
146
110
103
103
104
106
110
125
163
192
188
166
117
Overall, seasonal change in avian diversity is limited, with minimum and maximum
values for a single Landscape Division varying by 4–12 species. From December–
March all assemblages decline, as eight migratory species leave the area. The
decrease is most pronounced in Woodland areas, where seven of these species are
found. Diversity then slowly increases throughout autumn and winter, as migratory
birds arrive, and peaks in August, when 49–69 species are present in each
Landscape Division. The increase is particularly evident in Saline areas, due to the
arrival of migratory waterbirds that arrive when ephemeral salt lakes fill with winter
rain. As a result, Saline, Granite and Thicket areas now yield similarly diverse
assemblages (49–50 species). Diversity decreases through spring, as nine migratory
birds leave and only three arrive. This decline is most pronounced in Saline areas,
due to the departure of migratory waterbirds, so the gap in diversity between this and
other Landscape Divisions again begins to widen.
90
Birds
80
Granite
Heath
Mallee
Saline
Thicket
Woodland
70
60
no. species
50
40
30
20
10
0
Dec
Jan
Feb
Mar
Apr
May
Jun
Jul
Aug
Sep
Oct
Nov
Jun
Jul
Aug
Sep
Oct
Nov
80
Bird eggs
70
60
no. species
50
40
30
20
10
0
Dec
Jan
Feb
Mar
Apr
May
Figure 6.6 Monthly availability of birds (top) and birds’ eggs (bottom) in each Landscape Division. The
same colour key applies to each chart.
118
Avian egg-laying is seasonal (Figure 6.6) – of the 89 species of birds that visit or
inhabit the study area, < 30% produce eggs during summer (n=22). Most of these
species lay in December and, unsurprisingly, the pattern of diversity parallels that of
birds themselves, and is highest in Woodland and lowest in Saline areas. Diversity
declines through January when only six species, including the malleefowl, are laying.
By February, no eggs are available, and numbers remain low through autumn, with
no more than two species laying eggs in any Landscape Division at one time.
Significantly, emu produce their extremely large eggs in May, and could be found
throughout the study area.
By June, eggs are still scarce, with the most diverse assemblage found in Saline
areas, where migratory waterbirds breed. Diversity increases through winter, as nonmigratory birds also begin laying, and a separation starts to form between Woodland
and Mallee areas, where diversity increases at a greater rate, and the remaining
Landscape Divisions where fewer bird species are found. In Saline areas, egg
diversity peaks in August (n=44) and declines thereafter. Elsewhere, the maximum
occurs in September when Woodland and Mallee yield 62–65 species, while Heath,
Granite and Thicket preserve 46–48. Diversity then decreases sharply through
October and November, and the gaps between Landscape Divisions begin to narrow
once again.
6.3.2 Reptiles and their Eggs
The reptile assemblage comprises 25 lizards and three snakes. Most of these
produce eggs except four lizards (all skinks) and one snake that give birth to live
young. The lizard group includes 10 skinks (family Scincidae), nine geckos (suborder
Gekkota) and five dragons (Figure 6.7). Reptiles are most diverse in Woodland and
Mallee areas, which both harbour 23 species, including one – a species of gecko –
not found elsewhere; these Landscape Divisions provide all snake and most dragon
and gecko species. Slightly fewer reptile species are present in Heath areas (n=21),
and the dragon and gecko assemblages are less diverse than in Woodland and
Mallee areas. Just 15 reptiles are present in Granite areas, including two dragons,
and a sand monitor. Reptiles were comparatively rare in Thicket (n=8) and Saline
areas (n=4).
119
B
D
A
E
C
Figure removed due to copyright restrictions
H
F
G
K
I
J
L
Figure 6.7 Example of native non-avian fauna present in the study area. Common names are
provided after species name, measurements cited are maximum total length (TL – including tail) or
body length (BL – excluding tail) Photographs and measurements from Nevill (2014:124–129, 136–
137, 142–143, 151, 156–157, 162–163, 184–185). A–D: Reptiles: Ctenophorus cristatus, bicycle
dragon, TL 370 mm = (A); Liopholis multiscutata, bull skink, TL 250 mm (B); Underwoodisaurus milii,
barking gecko, TL 165 mm (C); Varanus gouldii, sand monitor, TL 1.6 m (D). E–K: Mammals:
Sminthopsis crassicaudata (family Dasyuridae), fat-tailed dunnart, BL 75 mm (E); Dasyurus geoffroii
(family Dasyuridae), chuditch, BL 310–360 mm (F); Macropus fuliginosus, western grey kangaroo, BL
0.95–2.2 m (G); Chalinolobus gouldii, Gould’s wattled bat, BL 65–75 mm (H); Isoodon obesulus,
southern brown bandicoot, BL 340 mm (I); Trichosurus vulpecula, common brushtail possum, BL 380
mm (J); Tachyglossus aculeatus, short-beaked echidna, BL 400 mm (K). L: Myobatrachus gouldii,
turtle frog, BL 50 mm.
120
Reptile species are present year-round (Table 6.3), but would be more accessible in
spring and summer as they are considerably more active, especially when seeking
mates. In winter, they are far less active, so would be especially difficult to catch
unless their burrows could be located; as a result, opportunities to hunt them would
be limited. Eggs are only available for four months of the year, September–
December. The greatest variety occur in Woodland and Mallee areas, where 19
different types of egg were available, followed by Heath (n=16) and Granite (n=11);
reptile eggs are rare in Thicket (n=6) and Saline (n=3) areas, due to their limited
reptile assemblages.
6.3.3 Mammals
Twenty mammal species were found in the study area, comprising six Dasyurids
(belonging to the family Dasyuridae – carnivorous marsupials, all small and mouselike in the study area except for a chuditch), three bats, three macropods, three
possums, two rodents, as well as a single echidna, numbat and bandicoot species
(Figure 6.7). A total of 17 species inhabit Woodland areas, including four that do not
occur elsewhere (a numbat, possum, bat and Dasyurid). Mallee areas housed 14
mammal species, while Heath had only 11, lacking all bats and many of the
Dasyurids found in Woodland and Mallee areas. Mammals were even rarer in the
remainder of the study area – just nine species were recorded at Granite areas and
seven at Thicket. Only two mammal species – one macropod and the echidna –
were present in Saline areas. Temporal variation is very subdued, as all mammal
species are present year-round (Table 6.3). However, two macropod species
(Macropus fuliginosus, Osphranter robustus) are in peak condition from June to
September due to abundant winter pasture; their condition deteriorates from October
as water became scarcer. It is possible that other herbivores would exhibit similar
seasonal variation in bodily condition.
6.3.4 Amphibians
Just five species of amphibian are found in the study area: two frogs, two toadlets
and one froglet (Figure 6.7). Amphibians are only found in three Landscape Divisions
– Granite, Heath and Thicket. They are most diverse in Granite areas, which yield
two toadlets, a frog and a froglet. The turtle frog (Myobatrachus gouldii) is confined
to Heath and Thicket. This frog can live in areas without standing water since eggs
121
are laid in the soil and the young frogs develop therein (Cogger 2014:172; Cronin
2014:38). All species are present year-round (Table 6.3), but Meagher (1974) noted
that people often preferred to eat females as they may be holding eggs at various
times of the year. The breeding period, when eggs may be present, generally lasts
for 3–4 months and is generally concentrated in autumn or winter. In Granite areas,
eggs may be found in three different species from March–June; just one species
breeds until September. The turtle frog has a different breeding cycle – it calls for
mates around July, and breeds several months later, so the young frogs are ready to
emerge with the first autumn rains (Cronin 2014:38; Western Australian Museum
n.d.). Hence, the females may contain eggs from October to December.
6.3.5 Summary
While the suite of plant foods varies seasonally, there is little temporal variation in
the faunal assemblage (Table 6.3, Figure 6.8). Seasonal changes relate primarily to
bodily
condition,
ease
of
capture
or
egg
production
rather
than
the
presence/absence of particular species. In spring and summer, reptiles are most
active, most easily caught, and are also producing eggs. In winter, macropods are in
peak condition after feeding on abundant winter pasture. Migratory waterbirds also
arrive in the area and many avian species are laying eggs. Reptiles would be difficult
to catch, however, as they typically retreat to nests or burrows to await warmer
weather. Female amphibians may contain eggs anytime from autumn to spring,
depending on the species. Variation is more pronounced at the spatial level,
reflecting the habitat preferences of different animal species, so it is necessary to
consider the overall faunal diversity represented at each Landscape Division.
Woodland areas exhibit the most diverse faunal assemblage, providing the largest
number of animal foods (n=200) and the maximum diversity in five of the six food
categories (Figure 6.9). Nearly all reptile and mammal species recorded in the study
area can be found in Woodland locations, although amphibians are notably absent.
Mallee areas supply slightly fewer animal food products overall, but reptiles (and
their eggs) are as diverse as in Woodland areas, and a good range of birds and
mammals are also present. Heath areas yield 155 animal foods, and all six
categories are present. Bird and bird egg diversity is lower than in Woodland and
Mallee, and just over half of the mammals present in the study area can be found in
122
Heath (n=11) – larger mammals would be less abundant due to the dense vegetation
(Macintyre et al. 1992). Slightly fewer reptiles and their eggs are present but, unlike
Woodland and Mallee areas, Heath harbours a single amphibian, the turtle frog,
which may hold eggs in winter (Appendix C).
250
200
Amphibian
Bird
Bird eggs
Mammal
Reptile
Reptile eggs
no. items
150
100
50
0
Dec
Jan
Feb
Mar
Apr
May
Jun
Jul
Aug
Sep
Oct
Nov
Figure 6.8 Total number of animal food items available in each month across the entire study area.
250
200
Amphibian
Bird
Bird eggs
Mammal
Reptile
Reptile eggs
Granite
Heath
no. items
150
100
50
0
Mallee
Saline
Thicket
Woodland
Figure 6.9 Number of different animal food items within each Landscape Division. Note that Heath
and Thicket both preserve a single amphibian species, not visible at the scale of this figure.
The faunal assemblage is less diverse in Granite areas, but this Landscape Division
exhibits the greatest range of amphibians, of which four species are present.
Females of three species may hold eggs in autumn, and another species in winter.
Of the reptiles, birds and mammals found within the study area, around half are
represented in the Granite Landscape Division, but both species of macropod are
123
present. Thicket areas have an even lower faunal diversity but, as in Granite and
Heath areas, all six animal food categories are represented. Bird and bird egg
diversity is similar to Granite, but mammals are slightly scarcer and larger ones
would be rare, as in Heath, due to closely-spaced vegetation. The reptile
assemblage is heavily reduced, with less than 30% of study area reptiles found in
Thicket areas. Saline areas consistently exhibit the least diverse fauna. While 108
animal food items are present, 99 of these are birds or bird eggs – only four reptile
and two mammal species are present. The avian assemblage is almost as diverse
as Heath, Granite and Thicket areas, although the Saline Landscape Division
represents the only area where waterbirds may flock in winter, when salt lakes are
holding water.
6.4 CONCLUSION
Plant foods are clearly abundant throughout the study area, but availability varies
over space and time. They are abundant in spring and summer, when all or most
categories are well represented, while autumn and winter are characterised by a
reduced suite of plant foods. Heath and Mallee areas generally supply the greatest
diversity of plant foods across all ranks, while the fewest species are associated with
Saline environments, although the relative diversity evident in each Landscape
Division can vary seasonally depending on the species composition – specifically the
proportion of plants providing their products year-round compared to those with more
restricted periods of availability. Temporal variation is less pronounced among the
fauna, as most animals are available year-round, but some migratory birds are
present for shorter periods. Bodily condition and ease of capture varies seasonally
for terrestrial animals, as does the availability of eggs, concentrated in spring and
early summer for reptiles, and August–October for birds. Woodland and Mallee
areas preserve the most diverse fauna, while Saline areas yield few non-avian
species.
124
CHAPTER 7
WATER AVAILABILITY
7.1 INTRODUCTION
This chapter seeks to quantify the longevity of water sources available to human
occupants of the study area. Due to its unique hydrological conditions, the only form
of potable water is rainwater, preserved in either the soil profile or rock structures.
Soil water availability is evaluated using the Soil Water Tool (DPIRD 2016a), and
rock structures using methods devised specifically for this research, as outlined in
Chapter 5. The availability of water within soil profiles and rock structures is
quantified throughout six modelled years that represent periods of low, average and
high winter rainfall, and also incorporate variously sized isolated falls (occurring
outside winter); the spatio-temporal variation in each of these resources is then
characterised. Finally, the soil water and rock structure data are synthesised to
define the availability of potable water within the study area and its Landscape
Divisions under different rainfall conditions, while considering the frequency at which
various conditions occur.
7.2 SOIL WATER
To quantify the availability of soil water in the study area, it is first necessary to
analyse longevity within each Landscape Division – as represented by a unique
idealised soil profile – under low (< 97 mm), average (97–177 mm) and high (> 177
mm) winter rainfall conditions. Each rainfall scenario is discussed separately, below.
7.2.1 Low Winter Rainfall – 2010, 2012
The small amount of rain received in winter 2010 (69.5 mm) meant that soil water
was a short-lived resource and only available in limited quantities (Figure 7.1). Five
of the six Landscape Divisions never had more than 11 mm available at any one
time. Even Granite areas, despite receiving additional input by runoff, exceeded 40
mm of Plant Available Water (PAW) on just one occasion. Following germination (the
point at which modelled crops begin growing and drawing on soil water – see Table
5.4) on 15 June 2010, Mallee soils held water for just a single day, while Saline and
125
Thicket areas preserved PAW for four or five days, respectively, and Heath and
Woodland for ten days; Granite soils dried after 19 days. This relatively short waterholding period resulted from the rainfall distribution – while 14 mm fell on 15–16 June,
just 2.7 mm was received over the remainder of the month. Most soils re-wetted as a
result of 27 mm of rainfall between 9 and 12 July, but Mallee soils remained dry.
Saline areas held water for three days, while Thicket, Heath and Woodland areas
retained water for 6–10 days, before all dried permanently in mid-July. Subsequent
falls were insufficient for any of these soils to re-wet, so none held water for more
than three weeks over the entire year. Granite soils dried on 4 August, after holding
water for 26 consecutive days, and 45 days in total. Granite soils may have
responded to 16.2 mm that fell over five days in mid-August, but this cannot be
evaluated with the current methodology. If PAW did accumulate, it would have been
fairly short-lived, since just 0.8 mm of rain fell over the next two weeks.
In 2012, most profiles held water more consistently than they had in 2010.
Nevertheless, in five of the six Landscape Divisions water was present only for a few
weeks in June. This is hardly surprising, considering that half of the winter rainfall
total fell during 7–30 June. As in 2010, Mallee soils only held water for one day, as
the preceding dry period meant that much larger falls would be required to recharge
these profiles. Saline and Thicket soils held water intermittently until 24 and 25 June,
respectively – Saline areas temporarily dried twice for a total of six days, while
Thicket soils dried for just two days. Heath and Woodland areas held PAW
consistently for three weeks, before drying on 28 June. Aside from Granite, only
Heath and Woodland soils held more than 10 mm of PAW at any one time. None of
the dry profiles responded to two 11 mm falls in August, and a number of larger falls
later in the year (26.6 mm on 1–2 November 2012, 60.6 mm on 24–29 November
2012) were preceded by too lengthy a dry period for these soils to again accumulate
PAW. Granite soils held water consistently for 75 days, from 7 June to 20 August
2012, nearly two months after Heath and Woodland areas dried. The early
November falls may have elicited a weak response, but PAW would have been very
short-lived. However, those in late November (totalling 60.6 mm of rain over five
days) would have resulted in PAW being present for around two and a half weeks,
when the recharge amount and average daily loss are considered (see also Chapter
7.2.2 and 7.2.4).
126
45
40
1…
1…
22-Jun-10
2…
2…
2…
29-Jun-10
3…
3…
6-Jul-10
6…
9…
13-Jul-10
1…
1…
1…
20-Jul-10
2…
2…
27-Jul-10
2…
3…
3-Aug-10
2…
5…
8…
10-Aug-10
1…
1…
17-Aug-10
1…
2…
24-Aug-10
2…
2…
2…
31-Aug-10
1…
4…
7-Sep-10
7…
1…
2010
14-Sep-10
1…
1…
1…
21-Sep-10
2…
2…
28-Sep-10
2…
1…
5-Oct-10
4…
7…
1…
12-Oct-10
1…
1…
19-Oct-10
1…
2…
26-Oct-10
2…
2…
3…
2-Nov-10
3…
6…
9-Nov-10
9…
1…
16-Nov-10
1…
1…
2…
23-Nov-10
2…
2…
30-Nov-10
3…
Granite
Woodland
Saline
Heath
6…
Mallee
3…
7-Dec-10
Thicket
27-Dec-12
35
20-Dec-12
30
13-Dec-12
25
6-Dec-12
20
29-Nov-12
15
22-Nov-12
10
15-Nov-12
5
8-Nov-12
0
1-Nov-12
20
25-Oct-12
PAW (mm)
15
18-Oct-12
10
11-Oct-12
5
4-Oct-12
0
27-Sep-12
15-Jun-10
2012
20-Sep-12
60
13-Sep-12
50
6-Sep-12
40
30-Aug-12
30
23-Aug-12
20
16-Aug-12
7…
1…
1…
1…
1…
2…
2…
2…
1…
4…
7…
1…
1…
1…
1…
2…
2…
2…
3…
3…
6…
9…
1…
1…
1…
2…
2…
2…
3…
2…
5…
8…
1…
1…
1…
2…
2…
2…
2…
2…
5…
8…
1…
1…
1…
2…
2…
2…
2…
1…
4…
7…
1…
1…
1…
1…
2…
2…
2…
1…
4…
7…
1…
1…
1…
1…
2…
2…
2…
3…
10
9-Aug-12
0
2-Aug-12
40
26-Jul-12
35
19-Jul-12
rain (mm)
PAW (mm)
30
12-Jul-12
25
5-Jul-12
20
28-Jun-12
15
21-Jun-12
5
14-Jun-12
10
0
127
Figure 7.1 Daily moisture balance of soils in each Landscape Division (mm of plant available water – PAW –
coloured lines) and rainfall (black bars) for years with low winter rainfall. Note that the same colour key applies
to both.
rain (mm)
7-Jun-12
9…
1…
14-Dec-10
1…
1…
21-Dec-10
2…
2…
28-Dec-10
2…
3…
7.2.2 Average Winter Rainfall – 2009, 2011
In 2009, Heath, Thicket and Woodland held water for similar periods, while the
remaining Landscape Divisions had different patterns of soil moisture presence
(Figure 7.2). The overall quantity of PAW was greater than in the low rainfall years,
but only Granite held > 35 mm at any one time. Heath, Thicket and Woodland areas
all accumulated PAW on 20 June 2009 and held it for 43–44 days, drying on 2–3
August, prompted by sub-2mm falls after 22 July. These soils then remained dry for
nearly two weeks, before re-wetting following 12.8 mm of rain on 15 August;
responses were likely aided by 11.8 mm that fell between 6–14 August. Heath and
Thicket areas finally dried on 24 August, while Woodland areas held water for one
more day – these areas yielded PAW for a total of 52 to 54 days. Saline and Mallee
areas held PAW for far shorter periods. Saline soils accumulated water on 20 June
but did not re-wet after drying on 2 August, despite reasonable falls in the first half of
August. While 25.2 mm of rain fell during the first eight days of the model period,
Mallee soils did not accumulate water until 30 June 2009 after a 12.2 mm rainfall
event. This water was retained for just two days before the profile dried, and
subsequently re-wetted after 21.2 mm of rain fell over two days in early July. Mallee
soils were devoid of water by 28 July, having yielded PAW for just 20 days. Despite
some moderately sized falls later in the year (14 mm on 11–12 September, 12.6 mm
on 27 October 2009, 18.2 mm on 13 November), none of these Landscape Divisions
accumulated PAW; higher evaporation rates meant that input did not sufficiently
recharge profiles.
In contrast, Granite soils held water continuously for 121 days, from 20 June until
they dried on 19 October 2009. The profile may have briefly re-wetted following the
27 October fall, but water would have been short-lived due to the increased
evaporation rate characteristic of the warmer months. Subsequent rainfall events
(18.2 mm on 13 November, 11 mm on 19 November) would have been insufficient to
recharge a completely dry Granite profile (i.e. one that was not just devoid of PAW
but had rather approached or reached its minimum volumetric water content), but
may have elicited a weak response due to the short dry period that preceded the
falls.
128
PAW (mm)
1…
1…
2…
2…
2…
3…
2…
5…
8…
1…
1…
2…
2…
2…
2…
Mallee
Granite
Woodland
Saline
Heath
9
-0
2…
ec
2…
9
-0
2…
28-Dec-11
ec
2…
9
-0
1…
21-Dec-11
ec
1…
Thicket
8…
1…
14-Dec-11
9
-0
ec
5…
7-Dec-11
9
-0
ov
30-Nov-11
9
-0
ov
2…
23-Nov-11
9
-0
ov
1…
16-Nov-11
9
-0
ov
9-Nov-11
09
ct-
2-Nov-11
09
ct-
26-Oct-11
09
ct-
19-Oct-11
09
ct-
1…
120
9…
12-Oct-11
09
ct-
6…
2009
3…
9
-0
ep
3…
5-Oct-11
100
2…
9
-0
ep
2…
28-Sep-11
2011
2…
9
-0
ep
1…
1…
21-Sep-11
80
1…
9
-0
ep
9…
14-Sep-11
60
6…
9
-0
ug
3…
7-Sep-11
40
3…
9
-0
ug
2…
2…
31-Aug-11
20
2…
9
-0
ug
1…
24-Aug-11
9
-0
ug
1…
9
-0
ug
1…
17-Aug-11
9
l-0
Ju
1…
9
l-0
Ju
4…
7…
10-Aug-11
9
l-0
Ju
1…
9
l-0
Ju
2…
3-Aug-11
9
-0
un
2…
9
-0
un
2…
27-Jul-11
0
2…
20
1…
1…
20-Jul-11
15
1…
10
8…
13-Jul-11
20-Jun-09
22-Jun-09
24-Jun-09
26-Jun-09
28-Jun-09
30-Jun-09
2-Jul-09
4-Jul-09
6-Jul-09
8-Jul-09
10-Jul-09
12-Jul-09
14-Jul-09
16-Jul-09
18-Jul-09
20-Jul-09
22-Jul-09
24-Jul-09
26-Jul-09
28-Jul-09
30-Jul-09
1-Aug-09
3-Aug-09
5-Aug-09
7-Aug-09
9-Aug-09
11-Aug-09
13-Aug-09
15-Aug-09
17-Aug-09
19-Aug-09
21-Aug-09
23-Aug-09
25-Aug-09
27-Aug-09
29-Aug-09
31-Aug-09
2-Sep-09
4-Sep-09
6-Sep-09
8-Sep-09
10-Sep-09
12-Sep-09
14-Sep-09
16-Sep-09
18-Sep-09
20-Sep-09
22-Sep-09
24-Sep-09
26-Sep-09
28-Sep-09
30-Sep-09
2-Oct-09
4-Oct-09
6-Oct-09
8-Oct-09
10-Oct-09
12-Oct-09
14-Oct-09
16-Oct-09
18-Oct-09
20-Oct-09
22-Oct-09
24-Oct-09
26-Oct-09
28-Oct-09
30-Oct-09
1-Nov-09
3-Nov-09
5-Nov-09
5
5…
0
140
120
100
2…
2…
6-Jul-11
PAW (mm)
rain (mm)
80
60
40
20
0
0
29-Jun-11
110
100
90
80
70
60
50
40
30
20
10
129
Figure 7.2 Daily moisture balance of soils in each Landscape Division (mm of plant available water – PAW –
coloured lines) and rainfall (black bars) for years with average winter rainfall. Note that the same colour key
applies to both.
rain (mm)
As in 2009, the 2011 model period was characterised by intermittent soil water
availability. Saline, Thicket, Woodland and Heath areas had all received PAW on 29
June 2011 but dried out between 2 and 5 July because the 9.5 mm of rain received
in the first three days was followed by just 0.7 mm over the subsequent nine-day
period. These soils re-wetted after 14.5 mm of rain fell on 11–12 July and held PAW
until 17–19 August. Only Woodland soils held water consistently during this second
period – all other soils dried temporarily at least once, for up to four days at a time.
As in 2009, Mallee soils did not accumulate PAW when all other soils did, but
remained dry until almost a month later, when cumulative falls (58.8 mm between 29
June and 28 July) had sufficiently recharged the profile. Mallee soils subsequently
dried on 2 August, after holding water for four days. None of these profiles
accumulated PAW as a result of several > 10 mm falls that occurred from September
to November.
Granite soils held water consistently from 20 June to 27 November but, as a result of
the rainfall distribution noted above, the profile very nearly dried out in early July. In
contrast to all other Landscape Divisions, Granite soils held water for longer after the
2011 winter than they did after high winter rainfall in 1992 and 1998. This is due to
the 2011 post-winter rainfall distribution, whereby the average winter total was
followed by a few moderately sized falls in later months (12.6 mm on 19 September,
13 mm on 23 October, 12.6 mm on 4 November). The Granite profile was still
holding water when these falls occurred and was thus able to respond strongly to
them. These falls not only arrested PAW decline, but actually increased balances
sufficiently to prolong the longevity of water. These falls would have elicited a much
weaker response had they occurred more than a few days after Granite soils had
dried, as the profile would need to recharge before PAW would accumulate.
Only Granite and Thicket soils, the two coarsest profiles, responded when 105 mm
of rain fell on 7 December 2011. The 5.4 mm received the previous day would have
been largely ineffective due to higher atmospheric demand at this time of year. It is
well-known that coarser profiles respond more readily to rainfall events as lower
quantities of water are required for recharge (Figure 5.3; Moore 2001:86), but it is
noteworthy that even such a large rainfall event was insufficient to recharge the
remaining soils. Thicket soils had been dry for 112 days, yet still accumulated 41.6
130
mm of PAW as a result. The remaining 63.4 mm represents the quantity of water
required to recharge the profile, plus the daily loss. The Thicket profile stored this
water for just 13 days but would have dried slightly sooner were it not for the 11.6
mm fall on 13 December that temporarily slowed the rate of loss. The sand profile
(from which Granite soils were modelled, after allowing for runoff – see Chapter 5.4.2)
accumulated 4 mm of PAW as a result of the same fall, meaning 101 mm was
consumed through recharge and daily loss. For the shallower Granite profile, just 46
mm would be required to recharge a completely dry profile – this could be achieved
with 23 mm of rain. The Granite profile had only been devoid of PAW for nine days
before the 105 mm fall, so even less water would be required for recharge. In any
case, the Granite profile could not accept all the water available to it, due to the
limited storage capacity of 100 mm. Granite soils stored this water until the end of
the model period (31 December 2011) and, based on the rate of loss, would have
held it for perhaps another eight days into 2012, presuming no further rainfall
occurred. This isolated fall, as well as the subsequent 11.6 mm fall noted above,
therefore resulted in the presence of PAW for a total of 33 days in Granite soils.
7.2.3 High Winter Rainfall – 1992, 1998
Apart from Granite, all of the Landscape Divisions demonstrated similar patterns of
water-retention in 1992. They held water, rarely > 20 mm at any one time, from 13
June until late September/early October (Figure 7.3). All of the five Division’s soils
temporarily dried at least three times, but Mallee soils dried on five separate
occasions and for longer periods than the others. The three major drying events
were each prompted by six- to twelve-day periods where rain was scarce, and daily
falls were < 2 mm; PAW balances were not large enough to insulate profiles against
drying under these circumstances. These profiles would likely have remained dry if
not for 20 mm of rain that fell on 20 September, followed by 7.1 mm over the next
four days. This brought the September rainfall total to 53.4 mm, while the mean is
just 28.1 mm (BoM data). This extra post-winter input permitted all soils to re-wet for
another five to eight days, before drying between 25 and 28 September. Thereafter,
Mallee, Heath and Saline soils remained dry, having held water for a total of 75, 92
and 94 days respectively. Thicket and Woodland soils briefly re-wet after 10 mm of
rain fell over two days in early October, and finally dried on 8 October 1992; they
held water for 99 and 102 days respectively. Subsequent falls were too small (< 4
131
mm) and well-spaced to permit any of these profiles to accumulate PAW. In contrast,
Granite soils held water consistently from 13 June to 3 November, due to the
additional input they received from runoff. Their larger PAW balance, which reached
the 100 mm storage capacity on several occasions, allowed them to retain PAW
through the rainless periods that prompted all other profiles to dry. Nevertheless,
even with extra input from runoff, the light falls later in 1992 were insufficient to
recharge the Granite profile.
In 1998, the same five Landscape Divisions were again closely grouped, as they
were in 1992. However, in contrast to the earlier period, soils retained water
continuously from 5 June until their final drying date. This probably resulted from
rainfall distribution – a total of 60.4 mm of rain fell on 5–11 June, and all soils stored >
40 mm of PAW. These larger balances meant that all soils retained water even
though only 16.4 mm of rain fell from 12 to 30 June 1998. Balances then increased
due to strong rainfall in the first half of July (39.6 mm from 1–15 July) before
gradually declining due to limited falls thereafter. All profiles dried within a week of
each other: Saline and Heath on 29 July, and Thicket, Woodland and Mallee – in that
order – between 2 and 4 August. Just 12 mm of rain fell over the first 15 days of
August, so soils remained dry, rendering the later falls (22.2 mm on 28 and 29
August) insufficient to recharge the dry profiles. The well-spaced nature of late winter
rainfall events meant these soils held water for a much shorter period than they had
in 1992, despite the winter rainfall totals varying by < 40 mm. As in all other modelled
periods, Granite areas held water for the longest. In this case PAW was available for
136 days, from 5 June to 18 October, over ten weeks after all other areas had dried.
Subsequent rainfall events thereafter were too small (generally < 3 mm) and widely
spaced to permit Granite soils to again accumulate PAW.
132
120
2…
2…
27-Jun-92
2…
1…
4-Jul-92
4…
7…
11-Jul-92
1…
1…
1…
18-Jul-92
1…
2…
25-Jul-92
2…
2…
1-Aug-92
3…
3…
6…
8-Aug-92
9…
1…
15-Aug-92
1…
1…
22-Aug-92
2…
2…
2…
29-Aug-92
3…
2…
5-Sep-92
5…
8…
1…
1992
12-Sep-92
1…
1…
19-Sep-92
2…
2…
26-Sep-92
2…
2…
3-Oct-92
2…
5…
8…
10-Oct-92
1…
1…
17-Oct-92
1…
2…
24-Oct-92
2…
2…
2…
31-Oct-92
1…
4…
7-Nov-92
7…
1…
14-Nov-92
1…
1…
1…
21-Nov-92
2…
2…
28-Nov-92
2…
1…
1…
Heath
7…
12-Dec-92
Saline
4…
Granite
5-Dec-92
Woodland
25-Dec-98
1…
Mallee
18-Dec-98
1…
1…
20-Jun-92
Thicket
11-Dec-98
100
4-Dec-98
80
27-Nov-98
60
20-Nov-98
40
13-Nov-98
20
6-Nov-98
0
30-Oct-98
30
23-Oct-98
25
16-Oct-98
20
9-Oct-98
15
2-Oct-98
5
25-Sep-98
10
18-Sep-98
13-Jun-92
1998
11-Sep-98
0
4-Sep-98
120
28-Aug-98
100
21-Aug-98
PAW (mm)
rain (mm)
80
14-Aug-98
60
7-Aug-98
40
31-Jul-98
5…
8…
1…
1…
1…
2…
2…
2…
2…
2…
5…
8…
1…
1…
1…
2…
2…
2…
2…
1…
4…
7…
1…
1…
1…
1…
2…
2…
2…
3…
3…
6…
9…
1…
1…
1…
2…
2…
2…
3…
3…
6…
9…
1…
1…
1…
2…
2…
2…
3…
2…
5…
8…
1…
1…
1…
2…
2…
2…
2…
2…
5…
8…
1…
1…
1…
2…
2…
2…
2…
20
24-Jul-98
0
17-Jul-98
30
10-Jul-98
25
3-Jul-98
20
26-Jun-98
15
19-Jun-98
5
5-Jun-98
12-Jun-98
10
0
133
Figure 7.3 Daily moisture balance of soils in each Landscape Division (mm of plant available water – PAW –
coloured lines) and rainfall (black bars) for years with high winter rainfall. Note that the same colour key
applies to both.
Title
rain (mm)
1…
1…
19-Dec-92
1…
2…
26-Dec-92
2…
2…
3…
7.2.4 Summary
Figures 7.1–7.3 indicate that, for most of the study area, soil water was short-lived,
and its presence centred on winter. Initial winter falls recharged soil profiles and
provided the impetus for PAW to accumulate, while the size and distribution of
subsequent falls dictated the quantity, longevity and consistency of soil water. There
is a simple relationship between the quantity of winter rainfall and the longevity of
soil water – most soils stored water for longer periods after high rainfall winters than
after low rainfall winters (Table 7.1). Nevertheless, the timing was still influential. For
example, in 1998 (a high winter rainfall year) soils were mostly dry by early August
because over half the monthly rainfall total fell over a few days in late August, when
soils had been dry for too long to respond. Similarly, most soils dried at some point
during the winter of 1992 despite this year having the highest winter rainfall total
(n=217.1 mm). Falls were light and widely spaced, so PAW balances were too low to
prevent profiles drying during short rainless periods.
Truly isolated falls, those that occur when profiles have been dry for long periods of
time, were largely ineffective, since input must be sufficient to recharge the profile
before water can be made available to plants (Figure 5.3). Granite and Thicket areas
required 46 and 63.4 mm of water to recharge their respective soil profiles. For
Granite soils, recharge could be achieved with a fall of around 25 mm, due to the
extra input received from runoff. Recharge values are unavailable for other
Landscape Divisions, but must exceed 100 mm due to the lack of response in
December 2011.
Table 7.1 Longevity of PAW (days) in modelled years, across all six Landscape Divisions.
Landscape
Division
Granite
Heath
Woodland
Mallee
Saline
Thicket
Low winter rainfall
Average winter rainfall
High winter rainfall
2010
2012
2009
2011
1992
1998
45
17
74
21
121
52
177
41
144
92
136
54
20
1
21
1
54
20
45
4
102
75
59
60
7
10
11
16
43
52
32
50
94
99
54
58
134
Overall, Mallee areas had the poorest soil water potential. They took the longest to
accumulate PAW, were generally the first to lose it, and were most likely to dry out
during short rainless periods. In periods of low or average winter rainfall, soil water
was present for just a single day or a few weeks (Table 7.1). Water was more
plentiful after high winter rainfall, when Mallee soils held water for up to 11 weeks,
but the PAW balance still rarely exceeded 20 mm at any one time (Figure 7.1–7.3).
Saline soils generally held water for longer than Mallee profiles, but also frequently
dried out during short rainless periods when superior soils retained water. Like
Mallee, these soils were also less likely to re-wet once dried. Saline areas may yield
soil water for 1–13 weeks, depending on the quantity of winter rainfall. In some parts
of the Saline Landscape Division, sodic soils/subsoils may render this stored water
unpotable – indeed when Roe (1836:299) dug for water near a salt lake he
encountered only brackish, undrinkable water.
Thicket retained water for slightly longer than Saline soils, preserving PAW for up to
14 weeks. Heath and Woodland areas out-performed Thicket soils in lower rainfall
periods, holding water for an additional week or more, but all three areas performed
similarly under higher rainfall conditions. However, Thicket soils could accumulate
PAW from large isolated falls > 65 mm. The smaller amount of water required to
recharge coarsely textured Thicket profiles allowed these soils to respond, even after
lengthy dry periods, when the finer textured Heath and Woodland soils cannot.
Granite areas represented the best prospects for soil moisture under all rainfall
conditions, primarily due to increased input through runoff. This increased the
consistency, quantity and longevity of soil moisture. All other soil profiles rarely held >
10–40 mm PAW, with the upper limit occurring during high rainfall winters. In
contrast, Granite soils stored 40–50 mm of PAW during low winter rainfall years
(2010, 2012), and achieved their upper storage limit (100 mm) several times during
all other modelled years. Even when winter rainfall was limited, Granite soils held
water for at least 6.5 weeks; under average or high rainfall conditions these soils
could hold water for up to 22 weeks. Their limited profile depth and coarse texture,
combined with extra input from runoff, means that Granite soils accumulated PAW
after isolated falls > 25 mm. They benefit more from smaller rainfall events that are
spread over a longer period, as occurred in 2011 (Figure 7.2) than closely-spaced
135
heavier falls, as the profile quickly reaches its water-storage capacity under higher
input conditions.
7.3 ROCK STRUCTURES
Four types of rock structure are considered here: gnammas, pans, runoff gnammas
and runoff pans. Each exhibits unique characteristics of water reception and
retention. The availability of water in each structure is described for periods of low,
average and high winter rainfall, using the same six model periods as above.
7.3.1 Low Winter Rainfall – 2010, 2012
Rainfall was infrequent for the first four months of 2010, so all rock structures only
held small quantities of water (< 20 mm) for short periods, no more than nine days,
centred around specific March and April rainfall events (Figure 7.4). Pans only
responded to the heaviest falls due to the lack of runoff combined with fairly high
evaporation rates. By mid-May, falls were greater and evaporation rates lower, so
structures held water for 6–16 days. In contrast, runoff gnammas received enough
input to retain water consistently through to winter and well beyond. Winter rains
commenced when 14 mm fell over two days in mid-June but falls thereafter were
fairly irregular. The season was characterised by long periods where falls are absent,
or very light (< 1 mm), punctuated by a few heavier events. Pans held water
inconsistently throughout winter for no more than 19 days at a time. Gnammas and
runoff pans accumulated large enough balances from early to mid-July falls (30.5
mm over five days) to retain water throughout subsequent dry periods, when just 3.3
mm of rain fell over four weeks. Runoff pans dried on 29 August and gnammas on 4
September, after continuously holding water for 51 and 57 days, respectively. Pans,
runoff pans and gnammas all remained dry throughout the remainder of September
and October, as falls were light and widely spaced; only 5.9 mm of rain fell during
that two-month period. In contrast, runoff gnammas accumulated a large enough
water balance to retain water throughout much of this period. They dried on 22
October after holding water continuously for more than five months. All structures
briefly re-wet in mid-November when 8.4 mm of rain fell, but none held water for
more than five days. Only runoff gnammas responded to two lighter falls in
136
2…
22-Jan-10
2…
29-Jan-12
2…
29-Jan-10
2…
5-Feb-12
5…
5-Feb-10
5-…
12-Feb-12
1…
12-Feb-10
1…
19-Feb-12
1…
19-Feb-10
1…
26-Feb-12
2…
26-Feb-10
2…
4-Mar-12
4…
5-Mar-10
5-…
11-Mar-12
1…
12-Mar-10
1…
18-Mar-12
1…
19-Mar-10
1…
25-Mar-12
2…
26-Mar-10
2…
1-Apr-12
1…
2-Apr-10
2-…
9-Apr-10
9-…
1…
16-Apr-10
1…
22-Apr-12
2…
23-Apr-10
2…
29-Apr-12
2…
30-Apr-10
3…
6-May-12
6…
7-May-10
7-…
13-May-12
1…
14-May-10
1…
20-May-12
2…
21-May-10
2…
27-May-12
2…
28-May-10
2…
3-Jun-12
3…
4-Jun-10
4-…
10-Jun-12
1…
11-Jun-10
1…
17-Jun-12
1…
18-Jun-10
1…
24-Jun-12
2…
1-Jul-12
1…
8-Jul-12
8…
15-Jul-12
1…
22-Jul-12
2…
29-Jul-12
2…
5-Aug-12
5…
12-Aug-12
1…
19-Aug-12
1…
26-Aug-12
2…
2-Sep-12
2…
9-Sep-12
16-Sep-12
23-Sep-12
30-Sep-12
7-Oct-12
14-Oct-12
21-Oct-12
28-Oct-12
4-Nov-12
11-Nov-12
18-Nov-12
9-Jul-10
9-…
16-Jul-10
1…
23-Jul-10
2…
30-Jul-10
3…
6-Aug-10
6-…
13-Aug-10
1…
20-Aug-10
2…
27-Aug-10
2…
3-Sep-10
3-…
10-Sep-10
1…
17-Sep-10
1…
1…
2…
24-Sep-10
2…
1-Oct-10
1-…
8-Oct-10
8-…
3…
7…
1…
15-Oct-10
2…
22-Oct-10
2…
29-Oct-10
4…
5-Nov-10
1…
12-Nov-10
1…
19-Nov-10
26-Nov-10
2…
3-Dec-10
9…
10-Dec-10
1…
16-Dec-12
17-Dec-10
2…
23-Dec-12
24-Dec-10
3…
30-Dec-12
31-Dec-10
2…
2…
5-…
1…
1…
2…
3-…
1…
1…
2…
3…
Gnamma
2…
1…
Pan
9-Dec-12
2-…
Runoff gnamma
2-Dec-12
2…
2-Jul-10
Runoff pan
25-Nov-12
9…
25-Jun-10
2010
8…
2012
8-Apr-12
15-Apr-12
December (3.2 mm on 21 December, 2.6 mm on 25 December), retaining water for a
1…
22-Jan-12
single day in each case, due to higher atmospheric demand.
8-…
15-Jan-10
80
8-Jan-10
1…
60
8…
40
8-Jan-12
15-Jan-12
20
0
20
15
10
5
0
1-…
100
depth of water (mm)
1-Jan-10
120
rain (mm)
180
160
140
120
100
80
60
40
1-Jan-12
1…
20
0
40
35
30
25
20
15
5
10
0
137
Figure 7.4 Daily water balance (mm) in rock structures (coloured lines) and rainfall (black bars) for
years with low winter rainfall. Note that the same colour key applies to both.
depth of water (mm)
rain (mm)
As in 2010, the beginning of 2012 was characterised by limited water availability –
each structure held water for short periods centred on rainfall events large enough to
offset daily loss through evaporation. Runoff gnammas held water for up to nine days
at a time, while all other structures dried no more than three days after rain. Pans
only responded to larger falls, and held water for just one day after 10.2 mm of rain
fell in mid-March. Pans, runoff pans and gnammas held water for 7–16 days in May,
while runoff gnammas retained water consistently until winter. Winter rainfall
distribution differed from 2010, however. A total of 26.8 mm of rain fell over a week in
early June 2012, but subsequent falls were more frequent than in 2010, and the
intervening dry periods shorter. As a result, most structures consistently held water
throughout winter (Figure 7.4). In contrast, pans dried after holding water for 23
consecutive days in June, as the water balance (< 15 mm) was insufficient to survive
the late June and early July dry period. For the remainder of winter, pans held water
for less than six days at a time, while runoff pans and gnammas retained water
consistently until 31 August and 13 September, respectively. All three structures then
re-wetted several times in September, for one or two days at a time. Subsequent
larger (29.4 mm 31 October–2 November; 60.6 mm 25–29 November) allowed them
to hold water for 4–11 days in the first instance, and 7–18 days after the later fall.
Due to the considerable balance accumulated during winter (> 110 mm), runoff
gnammas were still holding water when the October and November falls occurred.
This increased the water balance sufficiently that they still held a small quantity of
water on 31 December, 241 days after they began to fill; the remaining 22 mm would
have been lost over the first five days of January 2013. Were it not for the larger
isolated falls noted above, which are well above the November mean (19.2 mm –
BoM data), runoff gnammas would probably have dried in early to mid-November.
Due to the higher annual rainfall total, created by heavier and more frequent rain
outside the winter period, all structures held water for considerably longer than they
had in 2010.
7.3.2 Average Winter Rainfall – 2009, 2011
In early 2009, a few moderate to large falls helped all rock structures accumulate
water at least once. All structures except pans held water for up to three days in midJanuary after an 8 mm fall (Figure 7.5). On 28 January, 21 mm of rain fell, followed
138
by 10.4 mm on 1 February. Pans and runoff pans dried out between these two
events, and only held water for one to three days at a time. In contrast, gnammas
and runoff gnammas did not dry out between these closely spaced falls.
Nevertheless, they still only held water for fairly short periods, just six and 14 days,
respectively, highlighting the impact of greater atmospheric demand in the summer
months. Most rock structures held water intermittently, for up to 15 days at a time, in
May–early June, but runoff gnammas accumulated enough to remain wet until winter
rains commenced. From 19 June, rainfall events became more frequent and a few
large falls (> 12 mm) in late June and early July, combined with subsequent smaller
falls, allowed all structures to build up large stores of water. This permitted them to
retain water throughout winter, even when rainfall was scarce. Pans finally dried on 1
September, but regained water three times in September–October, for no more than
four days at a time, after daily or cumulative falls of 8–15 mm. Runoff pans dried on
10 October, and briefly re-wetted in late October, while gnammas retained water
consistently until 4 November. All rock structures briefly accumulated water as a
result of two larger November falls (18.2 mm on 13 November, 11 mm on 19
November), and held this water for up to five days. Subsequent falls were too light to
exceed the daily loss through evaporation, so pans, runoff pans and gnammas all
remained dry for the rest of the year. Runoff gnammas, however, were still holding
water on 31 December, a total of 224 consecutive days. This water would have been
lost by 11 January the following year, presuming no further input occurred. Unlike
2012, the November 2009 rains were only slightly above average (31.4 mm).
Therefore, the increased water longevity in 2009 was created not by isolated falls,
but due to higher winter rainfall and the large store of water (> 265 mm) this created
in runoff gnammas.
During first four months of 2011 water was present intermittently, for up to eight days
at a time in runoff gnammas, while other structures never yielded water for more
than four consecutive days. In the earliest months, pans only responded to falls
above 10 mm (15.6 mm on 6 January, 11.4 mm on 30 January), and held water for a
single day. They responded to smaller falls thereafter, when evaporation rates were
lower, but held this water for equally short periods. By mid-May, runoff gnammas had
accumulated enough water to prevent them from drying during short rainless periods.
All other structures still held water intermittently throughout June, due to low rainfall
139
(total 18.6 mm). From late June, gnammas and runoff pans held water consistently,
while pans filled intermittently until mid-July. Thereafter, rainfall events became
heavier and more frequent – 70.9 mm fell between 11 and 31 July. This input
allowed pans to hold water for 51 consecutive days, before they dried on 31 August.
Subsequent wet periods were short (up to six days) and centred on falls of 5–15 mm.
Runoff pans did not dry until 13 October, after holding water for 110 consecutive
days. A total of 35.5 mm of rain fell on 18–23 October, which increased the monthly
total to more than twice the average (49.1 mm, mean 21.6 mm). This permitted pans
and runoff pans to accumulate water for up to six and 18 days at a time, respectively,
before they both dried in early November. These late October falls also increased
the water balance in gnammas and runoff gnammas, which were still holding winter
water. Gnammas finally dried on 17 November, after holding water for 145
consecutive days. At this time, runoff gnammas were still holding > 250 mm of water.
On 7 December 2011, 105 mm of rain was received – this is the single largest
individual fall ever recorded by the Hyden meteorological station. This input allowed
water to accumulate in all previously dry rock structures and increased the already
considerable store of water in runoff gnammas. Nonetheless, pans dried after 12
days, despite another > 20 mm of rain that fell over the following week. Runoff pans
dried after 15 days because their limited depth meant they could not accept
additional input generated by runoff. Due to these falls and another 7.2 mm that fell
on 31 December, three of the four rock structures held water on the final day of the
model period. Runoff pans would have held this water for a day, while gnammas
would have preserved their water supply until about 6 January 2012. Runoff
gnammas held > 350 mm on 31 December, so would have yielded water for the first
few months of 2012. Had the 7 December fall not occurred, these structures would
probably have held water until early to mid-January.
140
depth of water (mm)
1…
22-Jan-11
2…
22-Jan-09
2…
29-Jan-11
2…
29-Jan-09
2…
5-Feb-11
5…
5-Feb-09
5…
12-Feb-11
1…
12-Feb-09
1…
19-Feb-11
1…
19-Feb-09
1…
26-Feb-11
2…
26-Feb-09
2…
5-Mar-11
5…
5-Mar-09
5…
12-Mar-11
1…
12-Mar-09
1…
19-Mar-11
1…
19-Mar-09
1…
26-Mar-11
2…
26-Mar-09
2…
2-Apr-11
2…
2-Apr-09
2…
9-Apr-11
9…
9-Apr-09
9…
16-Apr-11
1…
16-Apr-09
1…
23-Apr-11
2…
23-Apr-09
2…
30-Apr-11
3…
30-Apr-09
3…
7-May-11
7…
7-May-09
7…
14-May-11
1…
14-May-09
1…
21-May-11
2…
21-May-09
2…
28-May-11
2…
28-May-09
2…
4-Jun-11
4…
4-Jun-09
4…
11-Jun-11
1…
11-Jun-09
1…
18-Jun-11
1…
18-Jun-09
1…
25-Jun-11
2…
25-Jun-09
2…
2-Jul-11
2…
2-Jul-09
2…
9-Jul-11
9…
9-Jul-09
9…
16-Jul-11
1…
16-Jul-09
1…
23-Jul-11
2…
23-Jul-09
2…
30-Jul-11
3…
30-Jul-09
3…
6-Aug-11
6…
6-Aug-09
6…
13-Aug-11
1…
13-Aug-09
1…
20-Aug-11
2…
20-Aug-09
2…
27-Aug-11
2…
27-Aug-09
2…
3-Sep-11
3…
3-Sep-09
3…
10-Sep-11
1…
10-Sep-09
17-Sep-11
1…
17-Sep-09
24-Sep-11
2…
24-Sep-09
1-Oct-11
1…
1-Oct-09
8-Oct-11
8…
8-Oct-09
15-Oct-11
1…
15-Oct-09
22-Oct-11
2…
29-Oct-11
2…
5-Nov-11
5…
12-Nov-11
1…
19-Nov-11
1…
24-Dec-11
31-Dec-11
1…
2…
3…
19-Nov-09
26-Nov-09
3-Dec-09
10-Dec-09
17-Dec-09
24-Dec-09
31-Dec-09
1…
8…
1…
2…
2…
5…
1…
1…
2…
3…
1…
1…
2…
3…
Gnamma
17-Dec-11
1…
12-Nov-09
2…
Pan
10-Dec-11
3…
5-Nov-09
1…
Runoff gnamma
3-Dec-11
2…
29-Oct-09
1…
Runoff pan
26-Nov-11
22-Oct-09
2009
15-Jan-09
300
1…
250
8…
15-Jan-11
200
1…
8-Jan-09
150
1-Jan-09
8…
50
0
25
20
15
5
10
0
1…
8-Jan-11
2011
1-Jan-11
100
rain (mm)
550
500
450
400
350
300
250
200
150
100
0
50
110
100
90
80
70
60
50
40
30
20
10
0
141
Figure 7.5 Daily water balance (mm) in rock structures (coloured lines) and rainfall (black bars) for
years with average winter rainfall. Note that the same colour key applies to both.
depth of water (mm)
rain (mm)
7.3.3 High Winter Rainfall – 1992, 1998
In 1992, early water-holding events were rare – just one occurred, when 20.8 mm of
rain fell over three days in February (Figure 7.6). Pans, runoff pans and gnammas
held water for up to three days, while runoff gnammas retained water for 15 days,
aided by subsequent 3–4 mm falls that slowed the rate of loss. These falls were
insufficient to replenish supplies in the other structures. In mid-March, however,
103.7 mm of rain fell over eight days, and all structures accumulated water. These
falls were sufficient for runoff gnammas to retain moisture until winter and beyond.
Pans held water for 13 days but dried for one day between the initial 33 mm input
(13–16 March) and the subsequent 70.7 mm (17–20 March), again highlighting the
rapid rate of water-loss in this type of rock structure. Runoff pans and gnammas
dried by early April and early May, respectively. Thereafter, dried structures rewetted for short periods centred on various rainfall events, holding water for up to 15
days at a time. Early June was dry, but water was more abundant after strong midJune rainfall totalling 66.7 mm. As a result, all structures held enough water to
prevent them drying during July. August and September falls were well above
average, 110.7 and 53.4 mm respectively – this permitted all rock structures to
consistently hold water throughout these months. Pans dried on 2 October and
runoff pans on 22 October, after holding water for 113 and 133 consecutive days,
respectively. Runoff pans could not take full advantage of these falls due to their
limited storage capacity, which was reached 22 times between July and September.
Gnammas dried on 5 December, as the October and November falls barely slowed
the rate of water loss; these falls were so widely spaced that no pans re-wetted. In
contrast, runoff gnammas were holding almost 600 mm of water in mid-October.
That permitted these structures to retain water until 31 December, a total of 294
consecutive days or over 80% of the year. The remaining > 355 mm would have
been lost within the first few months of 1993, presuming average rainfall was
received.
142
depth of water (mm)
22-Jan-92
2…
29-Jan-98
2…
29-Jan-92
2…
5-Feb-98
5…
5-Feb-92
5…
12-Feb-98
1…
12-Feb-92
1…
19-Feb-98
1…
19-Feb-92
1…
26-Feb-98
2…
26-Feb-92
2…
5-Mar-98
5…
4-Mar-92
4…
12-Mar-98
1…
11-Mar-92
1…
19-Mar-98
1…
18-Mar-92
1…
26-Mar-98
2…
25-Mar-92
2…
2-Apr-98
2…
1-Apr-92
1…
9-Apr-98
9…
8-Apr-92
8…
16-Apr-98
1…
15-Apr-92
1…
23-Apr-98
2…
22-Apr-92
2…
30-Apr-98
3…
29-Apr-92
2…
7-May-98
7…
6-May-92
6…
14-May-98
1…
13-May-92
1…
21-May-98
2…
20-May-92
2…
28-May-98
2…
27-May-92
2…
4-Jun-98
4…
3-Jun-92
3…
11-Jun-98
1…
10-Jun-92
1…
18-Jun-98
1…
17-Jun-92
1…
25-Jun-98
2…
24-Jun-92
2…
2-Jul-98
2…
1-Jul-92
1…
9-Jul-98
9…
8-Jul-92
8…
16-Jul-98
1…
15-Jul-92
1…
23-Jul-98
2…
22-Jul-92
2…
30-Jul-98
3…
29-Jul-92
2…
6-Aug-98
6…
5-Aug-92
5…
13-Aug-98
1…
12-Aug-92
1…
20-Aug-98
2…
19-Aug-92
1…
27-Aug-98
2…
26-Aug-92
2…
3-Sep-98
3…
2-Sep-92
2…
10-Sep-98
1…
9-Sep-92
9…
17-Sep-98
1…
16-Sep-92
1…
24-Sep-98
2…
23-Sep-92
2…
1-Oct-98
1…
30-Sep-92
3…
8-Oct-98
8…
7-Oct-92
7…
15-Oct-98
1…
14-Oct-92
2…
21-Oct-92
2…
28-Oct-92
5…
4-Nov-92
1…
11-Nov-92
1…
18-Nov-92
2…
25-Nov-92
3…
2-Dec-92
1…
9-Dec-92
1…
16-Dec-92
2…
23-Dec-92
3…
30-Dec-92
22-Oct-98
29-Oct-98
5-Nov-98
12-Nov-98
19-Nov-98
24-Dec-98
31-Dec-98
4…
1…
1…
2…
2…
9…
1…
2…
3…
Gnamma
17-Dec-98
2…
Pan
10-Dec-98
2…
Runoff gnamma
3-Dec-98
1…
Runoff pan
26-Nov-98
1992
2…
700
1…
22-Jan-98
600
15-Jan-92
500
8…
1…
400
8-Jan-92
15-Jan-98
300
1-Jan-92
8…
200
0
35
30
25
20
15
5
10
0
1…
8-Jan-98
1998
1-Jan-98
1…
100
rain (mm)
400
350
300
250
200
150
100
50
0
35
30
25
20
15
5
10
0
143
Figure 7.6 Daily water balance (mm) in rock structures (coloured lines) and rainfall (black bars) for
years with high winter rainfall. Note that the same colour key applies to both.
depth of water (mm)
rain (mm)
No rain fell in the first two months of 1998, so the first opportunity for any rock
structure to accumulate water was when 4 mm of rain fell on 2 March, followed by
12.6 mm on 10 March. Water was only present for one to two days after the first fall,
but for up to a week after the second. Water was available for longer periods in April
as a result of 15.4 mm of rain that fell on 8–10 April, followed by 37 mm on 16–17
April. Pans held water for eight days, while runoff pans and gnammas both yielded
water for around three weeks. The strong April rainfall allowed runoff gnammas to
hold water consistently until winter, but was only present intermittently in other
structures until 60.4 mm of rain fell on 5–11 June. They retained this water for at
least three months, due to average July and August rainfall. Pans dried on 24 August
after holding water for 80 consecutive days, but re-wetted for another 11 days soon
after. The inconsistent presence of water resulted from the distribution of August
rainfall, as over half of the monthly total fell on 27–31 August. Runoff pans dried on 9
October after holding water for 126 days. Again, longevity was impeded by the 100
mm storage capacity, which was reached on 13 separate occasions before they
dried. Gnammas retained water for nearly a month longer, until 6 November, while
runoff gnammas still held approximately 100 mm of water on 31 December. This
would have been lost by 21 January 1999, presuming no additional input occurred.
7.3.4 Summary
The relationship between annual rainfall total and overall longevity of water in rock
structures is simple (Table 7.2). Rain is more effective in winter due to lower
evaporative demand, but water can accumulate fairly easily at other times of year, as
long as falls exceed the daily loss for a particular structure. Therefore, rainfall
distribution is less significant than overall quantity. Rain is generally scarce in the
beginning of each year, but structures may hold water for short periods centred on
larger rainfall events; this water is generally very short-lived due to higher
evaporation rates. Water is more consistently present from May–August when falls
become heavier and more frequent and evaporation rates are lower. As winter
rainfall increases, so too does the quantity, longevity and consistency of water in
rock structures. When winter rainfall is low, most structures will still preserve water,
albeit inconsistently, for most of the season. When winter rainfall is average or higher,
structures hold water consistently throughout winter, and frequently a month or more
beyond.
144
Post-winter longevity is dictated by the water-balance remaining at the end of winter,
as well as the size and distribution of subsequent rainfall events. Structures
frequently accumulate water in the later in the year, from isolated falls or short
periods of sustained rainfall, although that water is rarely preserved for more than a
week.
Table 7.2 Winter and annual rainfall totals (mm) for each of the modelled years, and the longevity of
water (days) in each rock structure. Years have been arranged by annual rainfall total (low to high).
Note that runoff gnammas frequently held water from the end of one year and into the next – this is
not reflected in the longevity figures.
Year
Annual
rainfall
Winter
rainfall
Gnammas
Pans
Runoff
gnammas
Runoff
pans
2010
142.1
69.5
108
43
183
95
2012
280
84.2
155
60
257
139
2009
295.9
150.1
169
92
244
142
1998
324.6
178
202
106
277
168
2011
472.4
126.4
212
99
264
181
1992
514.1
217.1
258
141
311
190
Pans were consistently the poorest performing rock structures, as they held water for
the shortest time and in the smallest quantity. They never reached their 100 mm
storage capacity due to the lack of runoff combined with a high evaporation rate that
nears atmospheric demand. Water-holding periods were closely centred on heavy or
sustained rainfall events. Pans were most likely to dry out during short rainless
periods, but could hold water consistently through winter, and up to a month beyond,
if rainfall was average or better. When winter rainfall was low, water was present for
as little as six weeks. Isolated falls were largely ineffective – even the largest (105
mm) resulted in water being present for less than two weeks. Runoff pans held water
for longer periods, and more consistently than pans that did not receive runoff. Water
was present throughout winter in all but the lowest rainfall year (2010) and runoff
pans generally preserved winter water for a month longer than pans. Runoff was
most valuable in low rainfall years, as storage capacity was quickly reached under
average or high input conditions. This limited the effectiveness of large isolated falls,
so the 105 mm fall was preserved for only 16 days. Runoff pans held water for
almost 14 weeks in 2010 and for just over half of the year under higher rainfall
conditions (1992), when the storage capacity was a stronger constraint.
145
Gnammas consistently held water for longer than pans and runoff pans, due to the
lower evaporation rate. This improved performance was especially evident under
higher rainfall conditions – when runoff pans were limited by their low storage
capacity – which permitted gnammas to retain water for nearly 37 weeks in 1992,
more than two months longer than runoff pans. Even under low rainfall conditions,
gnammas held water for 15 to 22 weeks of the year. Runoff gnammas always
yielded the greatest quantity of water, and held it for the longest period, due to the
due to the additional input from runoff combined with the lower evaporation
coefficient. Their large water balances prevented them from drying out during long
rainless periods when other structures lost their water. As a result, in five of the six
years modelled, runoff gnammas continuously held water from before winter until 31
December; they were generally dry for just two to four months a year. Runoff
gnammas could also accumulate water from small isolated falls and preserve
heavier falls for around two months.
7.4 SUMMARY OF WATER AVAILABILITY IN THE STUDY AREA
Of the 88 years for which seasonal data are available (the Hyden station collected no
weather information in 1947), 63 exhibited average winter rainfall. This equates to
roughly 72% of the sample, or seven years in every ten (Figure 7.7a). Therefore, the
water availability during those years represents the normal distribution of potable
water in the study area, if unusually large pre- or post-winter falls are discounted
(see Chapter 5.4.3). In average rainfall winters, water is available in most Landscape
Divisions, but the timing varies (Figure 7.8). At the beginning of winter rain is
sporadic, so water is only available in Granite areas with gnammas, runoff gnammas
or runoff pans, which begin to accumulate water in mid-May. When rain becomes
heavier and more frequent, generally in mid- to late June, soil water is available in
most Landscape Divisions. Mallee profiles, however, accumulate water later in the
year and hold it, intermittently, for a few weeks at best. Heath, Saline, Thicket and
Woodland soils hold water fairly consistently until mid-August, but the generally low
PAW balances mean these soils may temporarily dry out during prolonged periods of
limited rainfall. By late August, water is only available in the Granite Landscape
Division. Where present, pans will hold water until the end of winter, and runoff pans
until mid-October. All Granite soils will hold PAW until late October. Beyond this,
146
water is only preserved at Granite locations housing gnammas and runoff gnammas.
The former are dry by early November, while runoff gnammas can hold water until
January–February the following year.
180
70
160
60
140
50
no. falls
no. winters
120
40
100
80
30
60
20
40
10
20
0
0
low rainfall
average rainfall
high rainfall
15-24.9
25-34.9
35-44.9
45-54.9
55-64.9
65-74.9
75-84.9
85-94.9
95-104.9
≥ 105
fall size (mm)
Figure 7.7 Frequency of different rainfall scenarios in the study area from 1929–2017. A (left): number
of low, average and high winter rainfall years. B (right): frequency of variously sized isolated falls.
Based on Bureau of Meteorology data for the Hyden station (010568); note no data were recorded in
1947.
High and low rainfall winters are far less frequent, accounting for just 11% and 17%
of the sample, respectively, or around one or two years in every ten. Most soils and
rock structures begin accumulating water at the same time as in average winter
rainfall years, and dry out in the same order, but the longevity of their stored water is
vastly different. In low rainfall winters, Mallee areas only retain water for a day, and
Heath, Saline, Thicket and Woodland soils and profiles are dry by early July.
Depending on rainfall distribution, the timing of water-holding periods may vary, but
water will be consistently present unless rainfall events are interrupted by abnormally
long drier periods, as in 2010 (see Chapter 7.2.1). Thereafter, water is again
confined to Granite areas, but the longevity and consistency of its water retention is
reduced. Pans hold water intermittently through most of winter, and even gnammas
and runoff pans may dry temporarily if rainfall events are too widely spaced. Granite
soils are dry by mid-August, and runoff gnammas before the end of spring. In high
rainfall winters, Heath, Saline, Thicket and Woodland soils all retain soil water into
early September, as does Mallee, in contrast to low and average winter rainfall years.
Granite soils dry in late October and gnammas hold water until the end of November.
Larger water balances permit runoff gnammas to retain water until early to midFebruary. Even under high rainfall conditions, then, Granite is the only Landscape
Division that regularly preserves winter rains beyond early September (Figure 7.8).
147
Figure 7.8 Generalised water availability in soils and rock structures in each Landscape Division, in
average (top), low (middle) and high (bottom) winter rainfall years. Note that Heath, Mallee, Saline,
Thicket and Woodland areas only provide soil water. Solid lines indicate the generally consistent
presence of water over the specified period, i.e. being dry for no more than a few days at a time.
Dashed lines indicate when water may be available – its presence/absence depends on the timing of
rainfall events. Unusually large pre- or post-winter falls have been voided (see Chapter 5.4.3), as they
mask the normal longevity of water from winter rain. Water-holding periods for average (2009, 2011)
and high (1992, 1996) winter rainfall years are based on the mean of both years. For low winter
rainfall years (2010, 2012), only 2012 data was used as they were deemed more broadly
representative of low rainfall conditions. In winter 2010 a few large falls were interrupted by three
weeks of limited rainfall, causing soils to dry out when they normally would not have. Mallee soils are
not pictured for low rainfall conditions as they held water for just a single day.
148
It is also important to consider isolated falls of moderate size (≥ 15 mm) that occur
outside winter. A total of 266 such falls was recorded at Hyden between 1929 and
2017. Frequency declines as fall size increases; more than 60% of the sample
comprises falls in the smallest category (15–24.9 mm), while falls of ≥ 75 mm made
up just 2% of the sample (Figure 7.7b). In most cases, water would accumulate only
in Granite areas, but the longevity of water in an individual Granite location would
depend on fall size and the types of rock structures present. The distribution of pans
and runoff pans is somewhat irrelevant, however, as Granite soils outperform these
structures in all scenarios (Table 7.3).
If evenly distributed, the lightest isolated falls would occur around twice per year.
Soils would preserve water for a single day, gnammas for up to five days, but runoff
gnammas may hold water for 11 days (Table 7.3). Falls of 25–34.9 mm occur every
eighteen months, and water would then be present for slightly longer – up to six days
in Granite soils, eight days for gnammas, and over two weeks in runoff gnammas.
Larger falls, between 35 and 64.9 mm, occur more rarely, once every two and a half
years. Again, water would only be present in Granite areas, for 6–20 days in soils
and up to four weeks in rock structures.
Pan
1–2
2–3
3–4
9–10
≥ 10
Runoff
gnamma
6–11 11–16 16–20 20–25 25–29 29–34 34–38 38–42 42–46
≥ 46
Runoff
pan
3–5
5–7
7–9
9–11*
Soil
0–1
1–6
6–10
10–15 15–20
Soil
0
0
0
Thicket
10–12 12–14 14–17 17–19 19–21 21–24
4–5
0
5–6
11*
0
6–7
7–8
8–9
≥ 105 mm
8–10
85–94.9 mm
35–44.9 mm
5–8
75–84.9 mm
25–34.9 mm
3–5
65–74.9 mm
15–24.9 mm
Gnamma
55–64.9 mm
Storage type
Granite
45–54.9 mm
Landscape
Division
95–104.9 mm
Table 7.3 Longevity of water (days) in various soils and rock structures, as a result of differently sized
isolated falls. Durations are calculated assuming that falls occurred in December and were received
by a dry structure or profile. Daily loss for soils is based on average December figures (4.4 mm for
Granite, 4.8 mm for Thicket), and recharge amounts as per Chapter 7.2.2. Rainfall data sourced from
Bureau of Meteorology, for Hyden (010568) for 1929–2017 (no data for 1947). Asterisks indicate that
a structure or profile has reached its storage capacity, limiting longevity.
≥ 24
11*
11*
11*
11*
11*
20–
23*
23*
23*
23*
23*
1–3
3–5
5–7
7–9
≥9
149
No falls of 65–74.9 mm have been recorded at Hyden and falls of ≥ 75 mm and
above are very rare, occurring every 14.5 years. Water would be more prevalent
around Granite areas, present for up to 23–24 days in soil and gnammas, and up to
6.5 weeks in runoff gnammas. In contrast to other isolated falls, however, Thicket
soils would begin to accumulate PAW if ≥ 75 mm of rain was received. Nevertheless,
this water would be short-lived, around 7–9 days for the heaviest falls. Even the
largest isolated fall recorded (105 mm) was insufficient to recharge soils in Heath,
Mallee, Saline and Woodland Landscape Divisions, so these areas would not
accumulate water from any isolated falls unless their profiles had dried very recently.
In contrast, the strong response of rock structures to isolated falls means that water
would accumulate in Granite areas as a result of sub-15 mm falls. These falls are far
more frequent, but the resulting water would be absent from soil profiles and pans,
and short-lived in other rock structures.
7.5 CONCLUSION
The availability of potable water in the study area is limited, over both space and
time. Soil profiles are the only source of stored water for five of the six Landscape
Divisions, but this water rarely lasts through winter, let alone beyond. Of these areas,
only Thicket profiles will briefly accumulate water at other times as a result of rare
isolated falls > 65 mm that occur every 14.5 years or so. In contrast, Granite areas
regularly yield water outside the winter months. Winter rain is preserved for longer
periods due to extra input through runoff, as well as the greater protection from
evaporation that some rock structures offer. Granite soils can accumulate water after
falls of ≥ 25 mm, and rock structures easily respond to even smaller isolated falls as,
unlike soils, the length of the preceding dry period has no bearing on the amount of
water accumulated. Nevertheless, water is generally scarce by December. Only
runoff gnammas occasionally end years with a water balance large enough to persist
more than a few weeks into the following year, either as a result of rare high rainfall
winters – that occur around once ten years – or even rarer large isolated falls.
150
CHAPTER 8
OCCUPATION MODEL FOR THE STUDY AREA
8.1 INTRODUCTION
This chapter outlines an occupation model for the study area, focussing on three
main questions: whether occupation could be sustained year-round, what sort of
seasonal movements would be undertaken, and if/when inter-group aggregation
events could be held. These questions are addressed with reference to the
distribution of food and water resources as quantified in Chapters 6 and 7,
incorporating the findings from Forager Models discussed in Chapter 2. Then, the
archaeological signature of the occupation model is described, focussing on the
density and distribution of archaeological sites and other remains, and how this will
vary over time and space – not only between different Landscape Divisions but also
in different parts of the study area. Technological features are evaluated, describing
some attributes that should be present in stone artefact assemblages originating
from various sites and parts of the study area, drawing heavily on Kuhn's (1995)
concept of technological provisioning systems.
8.2 VIABILITY OF YEAR-ROUND OCCUPATION
Before seasonal routes can be evaluated, it is necessary to determine whether the
study area could sustain year-round occupation under normal rainfall conditions, or if
people could only remain in the area during certain seasons. Food is available
throughout the year, but plant foods are less abundant in autumn and winter (Table
6.2). This situation is common elsewhere in southwestern Australia (e.g. Anderson
1984; Bird 1985), so the reduced quantity of food is unlikely to hinder occupation
during those seasons. Instead, the availability of potable water was the limiting factor.
The data discussed in Chapter 7 suggested that year-round occupation was
impossible, as potable water could not be found in any Landscape Division between
mid-January and mid-May (Figure 7.8). This probably oversimplifies water availability
– particularly in gnammas – by not considering the purposeful preservation of water,
the rate of human consumption, the impact of increased runoff and the specific
volume of water preserved; each of these factors are evaluated below.
151
Individual gnammas are morphologically unique. Timms (2013) measured 80
gnammas in southwestern Australia, 15 of which are within the study area – on
average, these measured 2.02 m2 in surface area, and 720 mm in depth. A gnamma
with those dimensions would store approximately 2 litres of water for every mm of
depth, assuming that it had vertical sides. These average values can be used to
evaluate the specific quantity of water available in various types of gnammas
throughout the year, while also considering factors influencing the longevity of water,
such as the use of caps.
It is well-known that Aboriginal people capped gnammas with slabs of rock or piles of
branches to limit evaporation and keep the water clean (Bayly 1999, 2015; O'Connor
and Prober 2010). Caps preserve the water level at the time they are applied;
thereafter, human consumption is the only action reducing the water level. After
winter rains ceased, gnammas that receive no runoff held on average 78 mm of
water (Figure 7.5). Using the value derived above, this would represent around 156
litres of water. If caps were applied at this time, and each individual consumed 1.5
litres per day, a single gnamma would supply water for 104 people-days (10 people
consuming water on a single day = 10 people days; Table 8.1). It is unlikely that
capped gnammas could provide much more than a short-term or emergency source
of water, possibly during longer distance logistical forays. In contrast, runoff
gnammas achieve their maximum water balance in early spring, as they are able to
benefit from smaller, post-winter falls. On average, they held a maximum of 273 mm
of water, representing 546 litres and 364 people-days, if the structure was capped.
Table 8.1 The volume of water held by various capped rock structures and the longevity of water
supply (people days). People days = number of people x number of days, so ten people days may be
one person consuming water for ten days, or ten people consuming water for a single day. Daily
consumption was estimated at 1.5 L per person.
Rock structure
Capping date
Volume of
water (L)
No. people
days
Gnamma
End of winter
156
104
Runoff
gnamma
Early spring
546
364
High-input
gnamma
End of spring
End of summer
1440
620
960
413
152
As outlined in Chapter 5.4.3, runoff gnammas were modelled fairly conservatively by
assuming that they captured rain that fell directly within their margins and received
the same amount again through runoff – effectively doubling the rainfall. Even
though granite never sheds all the water that falls on its surface (Fernie 1930; Laing
and Hauck 1997), a modelled runoff gnamma requires a total capture area just 2.5–3
times its size. This accurately represents a gnamma below a wider, pan-like
depression that funnels water into the structure. For example, a gnamma 1 m in
diameter would need an overlying pan of around 1.73 m in diameter to capture twice
the amount of rainfall. However, considerably more runoff would be delivered into
gnammas below larger pans or those located at the bases of slopes or fed by long
intake channels; Bayly (1999) suggested that such channels were deliberately
created by Aboriginal people for this very reason. In these cases, the main limiting
factor is a gnamma’s depth, which limits its capacity to accept and store additional
runoff. If a gnamma received direct rainfall plus four times that amount again (i.e.
captured water from an area 8–10 times its size – a modestly sized slope or intake
channel), it would easily hold 720 mm of water by late winter and retain it, with minor
fluctuations in water balance, until the end of spring; thereafter, increased
evaporation and limited input depletes the water balance. If capped at the end of
spring, high-input gnammas would preserve 1440 litres of water, representing 960
people-days (Table 8.1). Even if these structures were left uncapped until the end of
summer, they would preserve more water than a standard runoff gnamma that was
capped at the optimal time (620 L compared to 546 L for runoff gnammas).
Even when capped, gnammas and runoff gnammas permit comparatively short-term
occupation. This could be extended if several gnammas were present at one location,
or if they stored a greater volume of water per millimetre of depth. Nevertheless,
year-round occupation probably relied on the use of high-input gnammas in
conjunction with other sources during the rainless period. People generally used the
most ephemeral water sources first, before moving to more permanent supplies
(Anderson 1984; Bird 1985; Gould 1968, 1969b; Tonkinson 1978:29). It follows, then,
that after exhausting surface and soil water reserves, gnammas would be used,
followed by runoff gnammas. Only then would people have moved to the most
reliable source – high-input gnammas – which would be the only viable source of
water until the winter rains began. Particular gnammas could possibly be capped at
153
the optimal time by people foraging in the local area, rather than left uncovered (and
losing water) until residential groups arrived on-site. Therefore, depending on the
distribution, size, morphology and runoff characteristics of various gnammas, careful
management of all water sources probably permitted year-round occupation of the
study area. However, this occupation would be tethered to a subset of the Granite
Landscape Division for much of the year.
8.3 SEASONAL ROUTE
Seasonal routes are heavily influenced not only by the distribution of particular
resources over space and time, but also their predictability. When food and water
resources are predictable, seasonal routes are fairly regular, such as those identified
in parts of southwestern Australia and those portions of the Western Desert where
rainfall is seasonal, albeit limited (Anderson 1984; Bird 1985; Cane 1987; Veth
1987). In other parts of the Western Desert, rainfall can be very localised and highly
variable both within and between years. There, regular seasonal routes were lacking,
and people instead followed the rains – several years often passed before a specific
area was revisited, when the rains chanced to lead them back in that direction
(Gould 1969b, 1991; Tonkinson 1978:29). Resource distribution is fairly predictable
in the study area, so a more regular seasonal route would be possible. This regular
route is first described, defined as that which occurs under average rainfall
conditions, before the impact of higher and lower winter rainfall on this annual
pattern is considered.
8.3.1 Regular Seasonal Route
Summer
In summer, occupation was tethered to those Granite areas with several large runoff
and/or high-input gnammas (Figure 8.1). It is unlikely that any single location would
yield sufficient water to support sizeable groups of people, so there could be no
large-scale congregation around reliable water such as occurred elsewhere in the
southwest during the drier months (Anderson 1984; Bird 1985). Similarly, water
would be relatively short-lived at most places with runoff gnammas, necessitating
moves to high-input gnammas at some point during summer. As well as providing
the most reliable water sources, Granite areas offered valuable protection from
154
bushfires – for this reason, fire-sensitive plants are often well-represented at granite
outcrops (Hopper 2000; Hopper et al. 1997). Of the plant foods available during
summer, gum and lerps/manna are highly ranked due to their high nutritional value
combined with limited handling time, as neither requires processing before
consumption (Table 5.5). Lerps/manna would be abundant in Eucalypt-dominated
Mallee and Woodland areas, while gum would be prolific in Thicket due to the
frequency of Acacia (Tables 6.2 and 8.2). Heath would also provide gum, though in
more limited amounts than Thicket areas.
Figure 8.1 Occupation model for the study area under average rainfall conditions, showing the
primary water sources (blue), primary foraging (green) and residential areas (yellow) and the optimal
portion of the study area (orange) that should be occupied at particular times of the year. For division
location see Figure 8.2.
155
Figure 8.2 Location of the northern, central, southern parts of the study area and the optimal months to occupy each. Note that these divisions are not firm
boundaries, but merely indicative of the broad areas referred to in the text and Figure 8.1.
156
Table 8.2 Temporal distribution of ranked plant food items (following Table 5.5), and unranked animal foods in each Landscape Division.
WOODLAND
THICKET
SALINE
MALLEE
HEATH
GRANITE
Summer
Autumn
Winter
Spring
Dec
Jan
Feb
Mar
Apr
May
Jun
Jul
Aug
Sep
Oct
Nov
High
30
30
30
0
0
0
0
0
0
0
0
0
Moderate
100
80
75
60
20
21
30
36
55
83
83
112
Low
47
46
28
28
28
29
29
29
44
50
50
50
Unranked
103
79
74
76
77
76
80
92
117
135
131
116
High
67
67
67
0
0
0
0
0
0
0
0
0
Moderate
147
98
87
44
33
36
58
69
109
179
179
163
Low
67
67
39
40
40
41
41
41
66
69
69
69
Unranked
117
87
81
82
84
83
87
100
126
149
147
129
High
92
92
92
0
0
0
0
0
0
0
0
0
Moderate
143
105
96
31
26
27
49
57
90
174
173
156
Low
76
76
35
36
36
37
37
37
75
78
78
78
Unranked
136
103
96
97
98
100
105
122
157
183
178
154
High
25
25
25
0
0
0
0
0
0
0
0
0
Moderate
71
62
54
14
7
9
13
17
29
68
68
72
Low
41
40
33
33
33
34
35
35
43
43
43
42
Unranked
61
48
45
45
47
48
57
70
100
101
95
79
High
46
46
46
0
0
0
0
0
0
0
0
0
Moderate
104
72
68
31
22
25
41
51
73
124
124
114
Low
40
40
22
23
23
23
23
23
39
41
41
41
Unranked
82
64
60
62
63
64
67
80
104
118
113
91
High
63
63
63
0
0
0
0
0
0
0
0
0
Moderate
98
78
74
29
14
15
24
32
51
101
101
106
Low
51
51
25
36
26
27
27
27
52
54
54
54
Unranked
146
110
103
97
104
106
110
125
163
192
188
166
157
Following Macarthur and Pianka’s (1966) diet-breadth model, it is safe to assume
that people would want access to both high-ranked plant food categories. However,
other plant and animal foods would also be required since residential mobility may
have been impeded by the infrequency of water sources. The most optimal summer
camp location would be in Granite areas in the northern portion of the study area
(Figure 8.2) – gum would be abundant in the extensive Thicket formations and
lerps/manna in the intervening Mallee, which also provides a good range of lower
ranked resources. Heath areas would still be accessible, and pockets of Woodland
are common in the northern Mallee formations (Figure 3.14; Table 8.2). Mallee and
Woodland areas both yield an abundance of fauna (Table 6.3), including reptiles that
could be easily caught during summer. Less fauna is present in Granite, Heath and
Thicket areas and larger game would be especially rare in dense vegetation
(MacIntyre et al. 1992). Granite areas also provide a wealth of moderately ranked
storage organs, including a considerable number that are only available for five
months centred around summer – importantly, many have high water-content at this
time (Table 6.2; Appendix B.2).
As occupation span increased, higher ranked resources would become depleted, so
lower ranked items would be pursued (Macarthur & Pianka 1966). Acacia seeds
would be an ideal choice due to their high calorie, fat and protein content but, like
most seeds, they require extensive processing so are ranked poorly compared to
other plant foods (Brand-Miller and Holt 1998; O'Connell and Hawkes 1981; Table
5.5). Again, the northern portion of the study area provides the optimal solution,
because Acacia seeds are abundant in Thicket areas – this resource could not be
intensively exploited from further afield, since low-ranked items can only be profitably
transported short distances (Orians and Pearson 1979). People may have stored
some of these Acacia seeds and gum-cakes for autumn, when plant foods are
scarcest, but this would only be profitable around high-input gnammas, where
occupation may continue beyond summer.
The northern area therefore provides a diverse diet through Mallee and Heath
Divisions, while permitting easy access to numerous Acacia-derived resources found
in Thickets, some of which would be especially significant as the stays lengthened;
high-ranked gum and lerps/manna would also be abundant. In contrast, the southern
158
part of the study area would be suboptimal for human occupation at this time.
Woodland and Mallee areas provide various types of game as well as lerps/manna,
but Heath areas are smaller and Thicket is inaccessible, restricting the abundance
and diversity of high-ranked resources and limiting the exploitability of low-ranked
seeds that cannot be profitably transported.
Autumn
In autumn, the availability of potable water is still restricted to certain Granite areas –
runoff gnammas were likely exhausted except where very large structures may have
formed; occupation was most likely centred around high-input gnammas (Figure 8.1).
High-ranked plant foods were unavailable after summer, heightening the importance
of other resources. Moderate-ranked resources (fruit, flowers and storage organs)
are most diverse in Heath, Mallee and Thicket areas, but flowers dominated Thickets
(Tables 6.2 and 8.2). Granite areas provide an abundance of plants with storage
organs, but only a fraction would be available after March. Low-ranked resources are
readily available in all of these Landscape Divisions, and Thicket would again
provide an abundance of seeds, now derived from Allocasuarina – another of the
three dominant genera – that preserve their seeds year-round inside woody fruits.
Animals would be a more important food source during autumn and the most diverse
are found in Woodland and Mallee areas. Heath and Granite also provide
reasonable faunal assemblages, and the latter supports three amphibian species
that may contain eggs in autumn.
Plant foods are scarce in autumn and high-ranked items are entirely absent. It would
therefore be vital to have access to the greatest overall diversity of resources. The
southern portion of the study area is again suboptimal. Thicket areas are
inaccessible and Heath pockets are small and widely spaced. Saline areas are
accessible, but yield the fewest moderate-ranked resources (although storage
organs are relatively diverse), and fauna are scarce. The optimal solution can
instead be found in the central study area (Figure 8.2), where Heath and Thicket
meet, specifically at Granite areas surrounded by Heath and Mallee. Patches of
Woodland occur within Mallee, and Thicket areas are also easily accessible (Figure
3.14). Depending on the distribution and capacity of high-input gnammas, it is
possible that certain occupation sites could be used throughout summer and into
159
autumn. While the residential location may be fixed, there would likely be a shift in
the foraging range, with more intensive use of Heath areas instead of Thicket (Figure
8.1). Longer stays would deplete resources, possibly necessitating an increased
foraging range, likely incorporating long logistical forays to target plant or animal
foods that could be profitably transported; capped gnammas or well-timed isolated
falls may provide valuable water sources for such journeys. From mid-May, flexibility
was increased somewhat – potable water was still restricted to Granite areas but
may be intermittently present at outcrops with runoff pans and any gnammas. These
water sources could be used as ‘stepping stones’ to allow people to begin moving
away from summer and autumn camps towards winter foraging areas in preparation
for rain.
Winter
At the beginning of winter, rain is generally infrequent and light, so reliable water is
restricted to Granite areas (Figure 8.1). From late June, however, potable water can
be found in all Landscape Divisions. This newly available water opens up parts of the
landscape that were previously inaccessible, either through the lack of water or
being too far from summer and autumn camps. Moderate-ranked plant foods are
most diverse at Heath, Mallee and, to a lesser extent, Thicket areas (Table 8.2).
Heath and Mallee have similar assemblages of fruits and storage organs, but flowers
are more diverse in Heath; all supply a variety of low-ranked items. In contrast,
Thickets yield few storage organs, fruits or low-ranked resources, but flowers are
diverse. Mallee and Woodland again yield the most diverse faunal assemblages, but
winter rains fill the ephemeral salt lakes, prompting the arrival of migratory waterbirds
in Saline areas – a great number of these and other birds lay eggs from July. Their
eggs are most diverse in Woodland, Mallee and Saline areas, and waterbirds are
unsurprisingly restricted to salt lakes within the latter Landscape Division. Macropods,
and potentially other herbivores, are at peak condition in winter and concentrated in
the more open Woodland and Mallee areas (Appendix C). Reptiles would be difficult
to catch after retreating to their winter burrows.
Aside from Mallee, in winter there is a disconnect between the Landscape Divisions
yielding the most diverse plant foods and those providing more abundant animal
foods. The optimal location could therefore be the central study area where Mallee
160
(with some pockets of Woodland) is interspersed with Thicket and Heath, or the
southern study area, where large expanses of Mallee occur along with Woodland,
Saline and some Heath areas (Figure 8.2). The southern palaeovalley may have
been preferable for several reasons – first, it provides access to comparatively large
expanses of four of the five Landscape Divisions yielding diverse plant or animal
assemblages, while the central area has only three. Secondly, the central area was
occupied through autumn so would likely be suffering from some level of resource
depletion and would benefit from people moving away to allow resources to
regenerate. Finally, Aboriginal people are well-known for exploiting seasonal
bounties, and characterise their seasons not only by weather conditions but by the
plant and animal foods that become available at various times, and the activities
associated with procuring them (e.g. Clarke 2009, 2011:54; Collard and Harben
2010; O'Conner and Prober 2010). The abundance of birds and their eggs, available
primarily in Saline, Mallee and Woodland areas, would invariably be targeted.
Similarly, while most moderate-ranked resources are rare in Saline areas, this
Landscape Division yields the most diverse assemblage of plants with perennial
storage organs, after Granite (Table 6.2).
In winter, then, people would likely have spread out into the newly accessible
southern part of the study area – particularly around the palaeodrainage network –
making use of ephemeral surface and soil water. Occupation may have been more
intensive in the eastern area, where larger salt lakes may attract a greater number of
waterbirds. Water is only intermittently available in Mallee soils (Figures 7.2 and 7.8),
so occupation sites should be primarily in Saline or Woodland areas, including where
Woodland occurs within the Mallee Landscape Division. Granite areas may be too
wet for comfortable habitation and Heath areas too dense – while access could be
created via burning, heath vegetation was infrequently fired elsewhere in
southwestern Australia (Prober et al. 2016). Considering the ephemerality of most
water sources, residential mobility would be greater during winter. Frequent moves
would also allow access to new resource patches, which may reduce reliance on
low-ranked items such as seeds, which require considerable processing. Longer
logistical forays should be unnecessary, since residential mobility is unimpeded and
water is universally available, albeit in somewhat limited supply. The reduced
resource base and ephemeral water would favour small groups, probably smaller
161
than those occupying the most reliable water sources in summer and autumn, so
some level of social fragmentation likely occurred, as it did elsewhere in the
southwest (Anderson 1984; Bird 1985).
Spring
From mid- to late winter, and certainly by spring, the choice of residential locations
narrowed, as water is only available in the Granite Landscape Division (Figure 7.8).
Plant and animal foods are mostly distributed as during autumn, but are becoming
more abundant; reptiles are active and easily caught, and many are also laying eggs
(Table 6.3). Plant foods are becoming somewhat more abundant, and moderateranked resources are again concentrated in Heath and Mallee areas (Table 8.2).
While Saline areas still provide waterbirds and some eggs, Woodland and Mallee
areas offer considerably more bird eggs, combined with far more diverse reptile
assemblages. Other animal foods are also well-distributed in Woodland and Mallee
areas, but the condition of large herbivores begins to deteriorate from October.
Occupation was likely concentrated in the southern part of the study area, as in
winter, but residential sites would have shifted towards Granite areas, where soil
water was available. These reserves would be exhausted by mid- to late October, so
groups would be forced to relocate to gnammas (Figure 8.1). Standard gnammas
would only support short-term occupation, unless several very large structures were
grouped together, but the presence of runoff or high-input gnammas would permit
considerably longer stays.
By November, runoff gnammas and high-input gnammas would be the main sources
of water (Figures 7.8 and 8.1). At this time, most migratory waterbirds would have
left, and ephemeral salt lakes would be dry; some waterbirds would still be laying,
probably confined to the larger, permanent salt lakes in the eastern portion of the
study area. Nevertheless, Saline areas would yield fewer eggs than most other
Landscape Divisions in spring (Table 6.3). There would be fewer benefits associated
with Saline valleys, so groups possibly began moving northwards towards the central
study area, where extensive Mallee and Heath formations occur with pockets of
Woodland. Heath and Mallee areas provide a diversity of moderate-ranked food
resources, with Woodland areas yielding a similar diversity of fruit and storage
organs but fewer flowers. An abundance of storage organs would be newly available
162
in Granite areas, specifically those that refilled their subsurface stores over the
growing season in preparation for their summer dormancy. The central study area
would have had nearly five months to recover from autumn occupation, so resources
should again be fairly abundant. This shift placed groups in an ideal location before
summer, when the northern part of the study area again represented the optimal
settlement location.
8.3.2 Impact of Rainfall Variation
While the quantity and distribution of rainfall influences the quality and abundance of
plant and animal foods, this impact is difficult to measure and negligible compared to
its effect on water availability. The latter should manifest itself primarily in the length
of occupation at particular sites. Under average rainfall conditions, water is almost
universally available in the study area for around two months, allowing maximum
flexibility in settlement patterns; at all other times, occupation is tethered to the
Granite Landscape Division (Figure 8.1). Under low rainfall conditions, however,
there is perhaps one month during which people could spread out and utilise
ephemeral water sources (Figure 7.8). Soil water, gnammas and runoff gnammas all
hold water for shorter periods than under average rainfall conditions, so those areas
could support correspondingly shorter occupation periods – only high-input
gnammas would permit extended stays. Occupation would be centred around these
both in the southern palaeovalley before the northward move in November, and in
the central and northern study area from late spring through autumn. This would be
especially problematic in the north and central study area as occupation is tethered
to these sites for long periods even under average rainfall conditions. Any extension
would heighten resource depletion, requiring more intensive use of lower ranked
plant resources as well as longer logistical forays to target more distant resources.
It is difficult to determine whether water supplies in these areas would last through to
the following winter without site-specific data on the distribution and capacity of highinput gnammas. Movement into adjacent areas may be possible, but this would
impose additional strain on their resource base, especially as the hydrological
situation is unlikely to be improved nearby. Instead, it may have been necessary for
some groups to remain in the palaeovalley in summer and autumn, around high-
163
input gnammas, and to forage more widely to offset the reduced diversity of plant
foods and the effects of resource depletion from the winter and spring occupation.
In contrast, high rainfall conditions offer increased flexibility as soil water supplies
allow occupation of non-Granite areas for three months of the year. In contrast to
average and low rainfall years, Mallee areas accumulate soil water fairly consistently
when rains are strong, so people would be free to camp in the extensive Mallee
formations at this time. From early spring, occupation would again be tethered to
Granite areas, but all types of gnammas would permit longer occupation than under
average rainfall conditions. However, these supplies may not have been heavily
utilised, as Granite soils provided water until the end of October (Figure 7.8), and
these permitted greater freedom of movement around the area. The increased water
availability would not necessarily translate to longer occupation of the winter foraging
grounds in the southern palaeovalley, but this may be optimal if higher rainfall
encouraged more waterbirds to travel to the study area and to remain, and breed, for
a longer period. In this case, gnammas would provide a viable water source after
Granite soils had dried, until the northward movement had commenced. Once
occupation shifted to the central and northern study area, runoff gnammas would
provide water for considerably longer than under average rainfall conditions –
occupation would therefore be tethered to high-input gnammas for a correspondingly
shorter period. Importantly, this would reduce resource strain around them,
permitting fewer logistical forays and less intensive use of low-ranked resources.
8.4 INTER-GROUP AGGREGATION
The spatio-temporal availability of resources dictates regular seasonal movements
but also influences the timing and location of inter-group aggregations for social,
ritual or ceremonial purposes. As Gould (1969a:121) noted 'Aboriginal ceremonial
life is not cut off from the practicalities of subsistence and daily living.' Aggregation
patterns depend on local ecological conditions and people may aggregate due to
resource abundance or, conversely, scarcity, such as when food is abundant but
reliable water sources are rare (Conkey 1980; Stanner 1965). Some argued that
food storage was an important component of larger aggregations (Hallam 1975:162;
Gammage 2011:296; Pascoe 2013:106–107), while others claim that events were
164
instead timed to coincide with natural periods of abundance (Brown 1918; Clarke
2009; Gould 1969b, 1984, 1991; Hassell 1975). In either case, considerable
quantities of food and water would be required.
As noted in Chapter 4.2.2, Noongar people gathered into larger residential groups
during summer, when food was abundant, in areas where reliable water was
available, often on the coast (Anderson 1984; Bird 1985). While neither Anderson
(1984) nor Bird (1985) specifically mention social or ceremonial gatherings, it is likely
that if they occurred within their respective study areas, their timing would coincide
with these regular periods of intra-group aggregation; large groups could not be
supported at other times. Similarly, O'Connor and Prober (2010) noted that in the
warmer months, abundant food and reliable water meant that permanent coastal
residents of the Great Western Woodlands (east of the Wheatbelt) could be joined
by people from further inland, so meetings and social gatherings could be held.
Similar patterns existed in parts of the Western Desert where rainfall is predictable –
people dispersed during the rainy period (in this case, summer) when food was less
abundant, and congregated around more permanent water in the cooler, drier
months when abundant food allowed ceremonies and meetings to occur (Cane 1987;
Veth 1987). Elsewhere in the Western Desert, rainfall is much less predictable and
seasonal rounds were less well-defined (Gould 1968, 1969b, 1991; Tonkinson
1978:29). Food could not be stockpiled for ceremonial events, as storage is only
beneficial where there is a guarantee of returning to a particular location in the
foreseeable future (Gould 1969b, 1991). Therefore, aggregations could only occur
during the periods of natural abundance that occur after good rains, when plant
foods, water and game were all simultaneously plentiful – consequently, these
aggregations were often many years apart (Gould 1969b, 1984, 1991).
Aggregations require either a synchronous abundance of food and water or, more
commonly, abundant food combined with a water source capable of sustaining a
large group of people for a short time. It is unlikely that food could be stored in
sufficient quantity to permit winter gatherings, when water was more plentiful, so
aggregation events in the study area would need to coincide with periods when food
was naturally abundant. Summer would be ideal, due to the availability of highranked plant foods (gum and lerps/manna) and December may have been
165
particularly suitable as good quantities of bird and reptile eggs were still available
(Table 6.3). However, the asynchronous distribution of critical resources means that
this summer food abundance was accompanied by limited water availability. While
larger isolated falls (≥ 35 mm) provide water that will withstand evaporation for a
week or more (Table 7.3), it is unlikely these sources would support the rate of
consumption associated with aggregations. The Hyden meteorological station has
only recorded three ³ 35 mm falls in December since record keeping began in 1928.
Therefore, aggregations could only occur by relying on winter rainfall stored in
gnammas. Under average rainfall conditions, runoff or high-input gnammas would
preserve sufficient water if capped at their peak; following high winter rainfall, all
gnammas would probably be capable of supporting aggregations, especially if more
than one was present at a particular site. Aggregating after high winter rainfall may
be advantageous for several reasons: it places fewer restrictions on the location of
aggregation sites, so they can be selected qualities other than the presence of a
specific type of gnamma; water will be more widely available following heavy winter
rains, which may facilitate travel to and from the study area; the increased longevity
of stored water supplies mean that reserving a source for aggregation would not
impact regular occupation by decreasing the amount of water available for
consumption at other times; and plant and animal foods may be more abundant or of
higher quality after good winter rainfall (e.g. macropods may stay at peak condition
through October and beyond). High winter rainfall occurs roughly once per decade
(BoM data; Chapter 7.4) so aggregation events may have been infrequent, and held
many years apart, as they were in the Western Desert (Gould 1969b, 1984, 1991).
It is also important to consider the parts of the study area that would best
accommodate aggregations, since large groups could significantly deplete the
resource base around a particular site, placing future occupation at risk. It seems
reasonable to assume that aggregations were held in areas that were otherwise
unused in summer, such as the southern palaeovalley (Figure 8.2). The central study
area may have been within the summer foraging range, so would be less ideal. In
the south, high-ranked lerps/manna would be abundant in the Mallee and Woodland
areas and gum could be procured, albeit in smaller quantities, from most accessible
Landscape Divisions. Moderate- and low-ranked plant resources, fauna and eggs
166
are fairly prevalent in Woodland and Mallee areas. Similarly, salt lakes may still be
holding water after strong winter rainfall, so waterbirds may still be in the area.
Granite provides an abundance of moderate-ranked storage organs throughout
summer. If required, low-ranked resources such as Acacia seeds could be utilised.
While these are most abundant in Thicket areas, they can be found in smaller
quantities in all Landscape Divisions, so their proximity to the aggregation site should
mean that transport costs are not prohibitive. Similarly, the timing and location of an
aggregation is highly predictable as it is dictated by the quantity of winter rainfall and
the presence of suitable rock structures and other site characteristics. Therefore, it
may have been possible to stockpile foods at the site in advance during the regular
palaeovalley occupation period. Gathering low-ranked items, such as Acacia seeds,
could be embedded in regular foraging, reducing costs in the same manner as raw
material procurement.
8.5 ARCHAEOLOGICAL AND TECHNOLOGICAL SIGNATURES
As demonstrated above, residential mobility varies considerably when the north and
central part of the study area are compared to the southern palaeovalley;
archaeological and technological signatures should vary accordingly. While quartz is
readily available at some Granite areas (see Chapter 3.2) it is not as ubiquitous so
the widespread adoption of an informal technology could probably not be supported.
As a result, evidence of Individual and Place Provisioning should vary according to
the degree of residential mobility. Furthermore, temporal change may be identified in
the archaeological record, since the length of occupation at particular sites is linked
to the quantity of winter rainfall. While the previous discussion centred on individual
years where winter rainfall is higher or lower than average, it is unlikely that such
short-term shifts would be identifiable in the archaeological record. Therefore,
changes will only be reflected across a larger time-scale, but the same patterns
should be in place. For example, the archaeological signature expected during a
high rainfall year would hold true over a longer period of increased rainfall. Analysis
of changing occupation patterns may thereby provide valuable palaeoenvironmental
data by indicating shifts in localised water availability as a proxy for effective
precipitation (Rossi 2014b).
167
8.5.1 Northern and Central Study Area
It was argued above that occupation was concentrated in the central and northern
study area from late spring through autumn, but the scarcity of water limited
residential mobility, tethering groups to runoff or high-input gnammas (Figure 8.1).
This tethering meant that the timing and location of occupation was highly
predictable, favouring a Place Provisioning strategy dominated by more expedient
stone reduction technologies with limited recycling, retouch and maintenance of
items (Kuhn 1995). Raw material should be predominantly quartz from the local area,
as embedded procurement can occur during foraging, aided by the widespread
distribution of the material. Longer term occupations may encourage some raw
material conservation – this will be reflected by bipolar flaking techniques, which
conserve stone while maintaining the benefits of a less formal technology (Hiscock
1996; Parry and Kelly 1987; Chapter 2.3.1). However, long occupations create a
higher risk of resource depletion, so grindstones and/or grinding patches may
indicate more intensive use of seeds – a low-ranked resource. Similarly, longer
logistical forays were likely required to target distant food resources that could be
profitably transported. In these cases, there should be some level of 'gearing up', a
specialised form of Individual Provisioning (Binford 1979; Kuhn 1995). There may be
some evidence, then, of higher quality raw materials along with retouch and more
controlled flaking, but this debris will be numerically overwhelmed by evidence of
Place Provisioning (Kuhn 1995:29). Higher quality raw materials would probably
comprise finer textured quartz with fewer internal fracture planes, since few other
types are available within the study area. Dolerite is heavy and coarse-grained, so
more suited to ground artefacts (Glover 1984) and less likely to feature within a
portable toolkit.
Residential sites in the central and northern study area should have strong
archaeological visibility for two reasons. First, a greater quantity of archaeological
debris is generated by Place Provisioning when compared to Individual Provisioning
strategies (Kuhn 1995:29). Furthermore, occupation is tethered to small areas near
reliable water, as people rarely situate themselves more than 100m from water
during drier months – the subsequent reoccupation of the same location results in
the continued accumulation of archaeological material (Binford 1980; Gould 1968).
As these sites will be palimpsests of several occupation events, they may preserve
168
evidence of changing occupation intensity over time, representing shifts in the
quantity of water that could be procured. Under low rainfall conditions, runoff
gnammas can support shorter occupations. In contrast, sites at high-input gnammas
should demonstrate correspondingly longer stays, as no other viable water sources
exist in the study area during the relevant seasons. The reverse is true under high
rainfall conditions, where increased water availability at runoff gnammas permits
shorter occupations at high-input gnammas. These shifts may be expressed as
variations in the amount of debris deposited (discard rates), length of reduction
sequences, and the incidence of bipolar reduction (Hiscock 1996). Longer
occupations at high-input gnammas may also produce more evidence of ‘gearing up’,
due to the increased importance of logistical forays.
While residential sites should be located primarily around reliable water sources,
there may be evidence of extractive sites or very short-term occupation elsewhere,
reflecting logistical forays to procure higher ranked food resources. These sites will
be more difficult to detect archaeologically, as the unpredictability of visitation would
discourage Place Provisioning and limit the formation of stratified archaeological
deposits. However, if local stone (most likely quartz) was available nearby, expedient
tools may have been produced and discarded on-site. Bipolar flaking should be
absent, due to short occupation phases, but there may be some evidence of toolkit
maintenance, in the form of debris comprising high-quality raw materials.
8.5.2 Southern Study Area (Palaeovalley)
The southern palaeovalley was the focus of occupation throughout winter and most
of spring (Figure 8.1). During winter, water is more widely available, permitting a
degree of flexibility in occupation patterns impossible at other times of the year.
While the optimal Landscape Divisions can be predicted – primarily Saline and
Woodland areas (including those occurring as pockets within Mallee) – there was
little need to reoccupy the same precise location on subsequent occasions.
Residential sites in Mallee areas lacking pockets of Woodland can only be the result
of high winter rainfall, as intermittent soil water means they cannot be occupied at
other times (see Chapter 8.3.2). Therefore, Individual Provisioning should have been
be the dominant strategy, but if raw material was available nearby there could be
some level of expedient artefact manufacture. Residential moves would be fairly
169
frequent due to the ephemerality of water resources, as well as the reduced suite of
plant foods, so occupation debris will be scarcer than in the northern and central
study area and should demonstrate little or no evidence of bipolar flaking. The
reduced chance of reoccupation will also limit the accumulation of archaeological
material in specific locations.
After winter rains cease, only Granite areas can support occupation, primarily due to
the presence of soil water. Nevertheless, precise residential locations are largely
unpredictable, since people would be free to choose from any one of the numerous
Granite zones occurring in the area. Similarly, campsites could be selected from
innumerable locations at any Granite outcrop, since water is fairly ubiquitous. Hence,
Individual Provisioning should dominate, likely supplemented by the local quartz that
is accessible at many Granite outcrops. Archaeological visibility will be lower than in
the northern and central study area, but higher than in Woodland and Saline areas,
as Granite acts as a tether, albeit weaker than elsewhere in the study area. At
gnammas, the occupation location is predictable, but duration is dictated by the
number and size of rock structures present, the amount of runoff received, and the
availability of water elsewhere in the area. Gnammas and runoff gnammas should be
characterised by Individual Provisioning, as visits will invariably be short, either due
to limited water storage (low winter rainfall years), the timing of regular seasonal
movements (average winter rainfall) or greater freedom in selecting residential sites
(high winter rainfall – see Chapter 8.3). High-input gnammas permit longer and more
predictable visits, even under low rainfall conditions, so Place Provisioning may
dominate at these locations. Further, temporal change should be evident – low
rainfall should be accompanied by increased occupation intensity, while they would
be occupied less intensively under high rainfall conditions. These shifts should be
primarily reflected in artefact discard rates and the length of reduction sequences.
Aggregation sites should only be found in the southern palaeovalley area. The
archaeological signature at these sites will differ from the regular occupation sites
described above. Location should be semi-predictable: a Granite area with several
large gnammas, ideally capturing some degree of runoff, that can be capped.
Aggregation events may require more intensive use of low-ranked resources, so
there may be some evidence of seed processing through grindstones and/or grinding
170
patches. While Place Provisioning is generally not beneficial for shorter term
occupation, the intensity of occupation – in this case dictated by group size – may
warrant stockpiling some raw material, but the short occupation span would
discourage the use of bipolar techniques. This Place Provisioning should be
accompanied by Individual Provisioning, reflecting groups that have travelled longer
distances to attend. Exotic materials may be present, though this depends on the
types of stone available in the visitors’ home territory. Conkey (1980) and Mitchell
(2016) noted that aggregation sites often demonstrate a level of segregation, with
different parts of the site used for specific activities. This may create distinct
assemblages in each area, or at least distinct clusters of archaeological material.
While aggregations were probably infrequent, their archaeological visibility should be
reasonable due to large group sizes, potential spatial organisation and some use of
Place Provisioning. Rainfall patterns dictate the frequency with which these events
can be held, so discard rates may be linked to longer term precipitation records. For
example, higher discard rates would indicate more frequent events, which could only
occur when high rainfall years were common, presuming that group size was
relatively consistent over time and that events could not be sustained under average
rainfall conditions.
8.6 CONCLUSION
Despite the lack of freshwater lakes and rivers, or potable groundwater, the study
area could be occupied year-round. However, for much of the year groups would be
restricted to a subset of the Granite Landscape Division, where high-input gnammas
capture runoff from larger areas. From late spring to early autumn groups would
occupy the central and northern part of the study area, with residential sites tethered
to the few reliable water sources found therein. For the remainder of the year, people
would reside in smaller groups in the southern palaeovalley, frequently moving their
residential bases between different parts of the Saline and Woodland Landscape
Divisions to make use of ephemeral soil water resources. Following the cessation of
winter rain, occupation would be tethered to Granite. Inter-group aggregations could
be held about once a decade in the southern palaeovalley in summers after heavy
winter rain. The archaeological and technological signatures should also exhibit a
pronounced north-south dichotomy. In the central and northern study area,
171
occupation was lengthy and predictable, favouring a Place Provisioning strategy. In
the southern palaeovalley, residential mobility was increased, and occupation
location was far less predictable, so Individual Provisioning should dominate,
although some Place Provisioning may occur at aggregation sites and high-input
gnammas. Throughout the study area, occupation intensity at high-input gnammas
should demonstrate an inverse relationship with rainfall, reflecting tethering when
water availability is limited, and greater freedom under higher rainfall conditions.
172
CHAPTER 9
ARCHAEOLOGICAL METHODS
9.1 INTRODUCTION
This chapter outlines the methodology applied during collection and analysis of
archaeological material. First, field methods are described, beginning with the siteselection process before considering the techniques used during archaeological
excavation and the collection of surface artefacts. Lithic analysis procedures are
then detailed, with particular focus on the problematic nature of quartz, before
describing the classification system adopted and the attributes measured. The
analysis of other collected materials – charcoal, sediment samples, ochre and glass
– is then outlined. Finally, the data analysis methods are described, focussing on
quantifying site-specific water sources and using artefact data to identify
technological provisioning systems and varying levels of occupation intensity.
9.2 FIELD METHODS
This section outlines the main field methods employed throughout this research.
Derivations from these general methods, as well as more specific details (i.e. the
location of individual test-pits), can be found in Chapters 10–12.
9.2.1 Site Selection
As noted earlier (see Chapter 1), Mulka's Cave was the initial focus of this research
because the extensive suite of rock art made it a potentially significant site, but the
author’s previous interpretations were hampered by a lack of archaeological context
(Rossi 2010, 2014a). To address this problem, an entirely new lithic analysis was
required, combined with comparative data from the surrounding area. The
Department of Planning, Lands and Heritage’s online database was consulted to
identify registered sites and 'other heritage places' within 50 km of Mulka's Cave.
The relevant reports and site files were then evaluated to determine if any of these
locations had potential for undisturbed sub-surface archaeological content. Locations
were discounted if on private property (for reasons of access), if they preserved
limited or no archaeological content (i.e. lost through development, or if literature
173
noted a lack of artefacts), and where there was evidence of disturbance or erosion.
Only a few open artefact scatters (ID 16787, 16791, 18508 – see Table 4.1) as well
as two water sources (5120 Anderson Rocks, 5121 Anderson Rocks 2 aka Twine
Reserve) met these criteria. In the case of 5120 and 5121, the presence/absence of
artefacts could not be confirmed as there was no associated literature. Anderson
Rocks and Twine Reserve (Figure 9.1) were prioritised because if these sites had
been occupied, the tethering effect of water sources would have encouraged
occupation within a more restricted area, thereby increasing the chance of
encountering stratified archaeological deposits. Both sites were visited, and each
exhibited a series of gnammas as well as obvious (Twine Reserve) or potential
artefacts (Anderson Rocks), warranting small-scale excavation of both sites.
Figure 9.1 Location of archaeological sites excavated by the author.
Unfortunately, investigation at Twine Reserve had to be abandoned (see below). To
offset this loss of data a more intensive investigation was planned at Anderson
Rocks. Nevertheless, another site was still required to maximise comparative data. It
was cost-prohibitive to consider another location that met Section 5 of the Aboriginal
Heritage Act (1972), so the decision was made to search other areas for potential
174
sites. The prevalence of agriculture creates a distinct preservation bias in the study
area, so the search was limited to areas of state-owned land – primarily large granite
outcrops or salt lakes, and a few small wooded reserves. Several areas yielded no
cultural material, or even potential cultural material, with the exception of Gibb Rock,
a granite outcrop that lacked gnammas but preserved potentially flaked stone. Gibb
Rock provided a mix of similarities and contrasts to the other two sites: all are in the
Granite Landscape Division but have vastly different water sources. The three sites
could then be used to address the newly emerged question regarding the use of
occupation history at certain areas as a proxy for effective rainfall (Rossi 2014b).
Twine Reserve (aka Anderson Rocks 2)
It is necessary to include a brief discussion of the Twine Reserve investigation as it
initially formed part of the site sample and some fieldwork was undertaken.
Permission to test-pit was sought, and subsequently granted, under Section 16 of
the Aboriginal Heritage Act (1972). As a condition of their support, the Ballardong
people (Traditional Owners for inland southwestern Australia – see Figure 4.3)
required three monitors to be on-site during the excavations, as well as the
implementation of an Intellectual Property (IP) agreement before reporting of any
results deriving from the excavations. The monitoring process required each
Traditional Owner (TO) to be paid $500 per day for their time, in addition to travel
allowances plus food and accommodation costs. On-site monitoring was not required
during earlier research at Mulka's Cave so, perhaps naively, the author assumed
that this would not be required for the current research, since all sites are
administered by Ballardong. Monitoring costs had therefore not been allowed for
during the planning and budgeting stage for this research, which was primarily selffunded. It should be noted that the amounts requested by the Ballardong are in-line
with those specified by the Noongar Standard Heritage Agreements that were
implemented in June 2015, as a part of the yet to be finalised South West Native
Title Settlement.
Available funds could only support a three-day field trip to Twine Reserve, with the
designated Indigenous monitors. While such a short visit would permit a cursory
investigation, at best, the archaeological content was deemed significant enough to
warrant the expenditure. The trip was scheduled for November 2015 with the aim to
175
excavate one or two small (500 x 500 mm) test-pits, as shallow deposits were
anticipated; unfortunately, various delays meant that little was actually achieved
during the limited timeframe. The first pit was very time-consuming to excavate due
to several pockets of heavily cemented soil (likely due to an ants’ nest encountered
lower in the pit) and was eventually abandoned due to limited archaeological
content. Before another pit could be initiated one of the monitors fell ill, and it was
clear that it was in her best interests to return to Perth immediately – therefore,
excavation could not continue under the conditions set down by the Ballardong
people. Due to financial constraints, further excavations at Twine Reserve were
impossible. For this reason, as well as the limited archaeological content within the
single excavated test-pit, no attempt was made to develop the IP agreement;
instead, the results of this excavation will not be discussed further. However, this site
features in the broader conclusions made herein, since the data did not derive from
the excavation, but rather from the author’s observations made at a time outside of
the November 2015 visit.
9.2.2 Survey and Excavation Methods
Before any fieldwork was undertaken on registered Aboriginal sites, permission was
sought under Section 16 of the Aboriginal Heritage Act (1972). While this specifically
relates to excavation, in practice it also applies to the removal of cultural objects from
the ground surface (i.e. surface artefacts should not be collected without obtaining
permission). No excavation or removal of cultural material occurred at a registered
site before the issuing of a Section 16 permit – these permits generally allow the
excavation of one square metre across a site. These same guidelines were adhered
at other locations, even where permissions were not required, as there was a strong
chance these would be registered as sites in the future, based on the results of this
research.
All sites were evaluated on foot, but systematic grid searches were impossible due to
large areas and limited access. Instead, investigation focussed on accessible areas
where archaeological content was likely, i.e. at the base of outcrops, areas where
sediment had accumulated on the rock, and around any water sources. Noteworthy
features (e.g. gnammas) were positioned via GPS and photographed; gnammas
were also measured, including their intake areas, where present. Surface artefacts
176
or potential artefacts were treated differently depending on whether these occurred
in identifiable concentrations (surface artefact scatters) or were more sparsely
distributed across the site. GPS coordinates were used to define the extent of
artefact concentrations; small and/or lower density scatters were often collected as a
single unit, while larger and/or denser scatters were subdivided depending on the
distribution
of
material
(see
Chapters
10–12).
Elsewhere,
individual
artefacts/potential artefacts were positioned via a GPS reading and collected along
with any other potential cultural material visible within a certain radius, generally less
than three metres. These methods permitted a reasonable level of spatial control
within the available timeframes.
At all sites, 500 x 500 mm pits were excavated. This permitted greater intra-site
spatial analysis while remaining within the one square metre limit. Pits were
positioned in areas offering the best chance of deep, undisturbed and stratified
archaeological content, generally in areas where surface artefacts (or potential
artefacts) were densest and there were no obvious signs of disturbance such as
drainage channels or animal burrows. Pit locations were recorded via GPS, and the
pit aligned on cardinal compass points where possible. Excavation was conducted
by the author and her secondary supervisor, R.E. Webb. Due to the lack of visible
stratigraphy the deposit was removed in arbitrary 25 mm excavation units (XUs), but
XU depth often varied according to the consistency of the sediment, excavator error,
or the presence of intrusive elements. Sediment samples were only taken from
selected pits and XUs, as no soil changes were apparent. Excavated sediment was
weighed, sieved through nested 6 mm and 3 mm mesh, and the retained material
bagged. This matter was subsequently sorted, separating natural debris from
charcoal, potential artefactual stone and other items of interest (e.g. glass, ochre).
Natural debris was retained but not studied further, while the remaining material was
subject to the additional analyses outlined below.
9.3 LITHIC ANALYSIS
Before outlining the lithic analysis methods, it is necessary to evaluate the dominant
raw material in study area – quartz. The problems associated with this material,
including difficulties identifying intentional flaking, heavily influence subsequent
177
methodology in terms of both the classification system adopted and the specific
attributes selected for analysis.
9.3.1 ‘Horrid Little Bits of Stone’ – the Problem with Quartz Artefact Analysis
Quartz (SiO2) is a mineral and has a crystalline structure; the size of crystals and the
way in which they aggregate can vary widely. Macrocrystalline quartz, the type to
which analysts refer when lamenting the difficulty of quartz artefact analysis,
comprises those varieties with a crystal structure that is identifiable at a macroscopic
level. Specimens may be aggregations of smaller irregularly formed (anhedral)
crystals (often visible only as grains), creating a texture ranging from fine- to coarsegrained or, if space and environmental conditions permit, each individual quartz
crystal can grow to such a size that its crystal structure is regular (euhedral) and
easily visible to the naked eye (Rafferty 2012:231). Crystal size forms a continuum,
and both large euhedral crystals and aggregations of anhedral crystals (of various
sizes) may be found within the same vein, since crystal size is related to the rate at
which the liquid SiO2 cools and the space available for crystal formation (see also
Chapter 3.2). While Hallam’s (1977:169) characterisation of some quartz as ‘horrid
little bits of stone’ was made in jest, this in fact reflects the attitudes of many analysts
to macrocrystalline quartz due to the heterogeneity of the material and its
unpredictable fracture patterns. This complicates typological analysis of flaked quartz
as well as the subclassification of quartz materials.
In isotropic materials (e.g. obsidian, single large quartz crystals), physical properties
are identical in every direction. These materials flake predictably because the
fracture path is dictated by core morphology and knapping technique (Cotterell and
Kamminga 1987). However, for aggregations of quartz crystals, fracture follows an
easier, predefined path along crystal boundaries – the larger the crystal facets, the
more influence these have over the shape of a flake. Aggregations of small crystals
may flake fairly well, as the fracture can largely follow its own path, while
aggregations of larger crystals will flake less predicably and produce blocky, chunky
flakes. Many varieties of quartz can complicate archaeological analyses, then, for the
following reasons:
178
1. Flakes may lack regular flake landmarks, where detached along internal
flaw planes or large crystal facets (Holdaway and Stern 2004:116–118;
Witter 1990:43);
2. Quartz has a higher fragmentation rate than other materials, so flakes are
frequently split as they are removed from the core; no single knapping
variable controls fragmentation (Driscoll 2011a; Tallavaara et al. 2010).
Broken flakes may be misclassified as flaked pieces, especially when
material is of poor-quality, which will affect artefact counts;
3. It may be difficult to identify retouch or other features, especially on
coarser grained material.
These issues are at the core of the ongoing argument of whether quartz-dominated
assemblages are best represented by an artefact typology designed specifically for
quartz and accounting for its irregular flaking properties, or whether existing
typologies designed for flint-like materials can be successfully applied (see summary
in Driscoll 2010:78–80). Driscoll (2011b) found that even where experienced
analysts were provided with a typology and description of quartz artefact types, and
then asked to classify an experimentally produced assemblage accordingly, their
classifications varied widely; clearly a universally adopted typology would not
alleviate all the difficulties associated with quartz artefact analysis.
The heterogeneity of macrocrystalline quartz makes subclassification difficult, as
texture, colour and opacity all form a continuum rather than a series of discrete
categories. In archaeological literature, quartz is generally divided into two
categories based on colour and opacity, using some variation of the crystal
quartz/clear quartz/rock crystal vs. milky quartz/vein quartz system. This may be a
reflection of the gemmological classification system, which relies mostly on the
colour of the material to subdivide quartz (e.g. rose quartz, smoky quartz), or a
consequence of Dickson’s (1977:102) argument that translucency is the strongest
indicator of how well quartz will flake. As noted above, however, it is crystal size and
the presence of internal flaw planes that actually determines how well quartz flakes.
Single quartz crystals may be transparent, translucent, or opaque – the traditional
archaeological colour-based classification scheme does therefore not address how
flakeable individual types are. Martinez Cortizas and Llana (1996) proposed four
179
morphostructural groups of quartz based on crystal size and the presence of internal
flaw planes; these have been demonstrated to more accurately characterise flaking
predictability (de Lombera-Hermida 2008, 2009). In practice, however, it is difficult to
determine quartz formation processes at an artefactual level – indeed, Driscoll
(2011a) demonstrated that it is often impossible to accurately distinguish grain size
without microscopic analysis, perhaps explaining why the morphostructural
classification system has not been widely adopted.
9.3.2 Stone Artefact Classification
Classification systems for flaked stone artefacts can generally be characterised as
either typological or materialist/technological. In the former system, artefacts are
classified into different types based on morphological characteristics (e.g. Bird 1985),
with a distinct emphasis placed on the quantity of retouched pieces or ‘formal tools’
(e.g. geometric microliths, backed blades, points). In a materialist system, pieces are
classified into mutually exclusive categories based on the position within the stone
reduction sequence, thereby examining the form of the artefact ‘in terms of the
mechanisms by which it was created, rather than the presumed purpose for which it
was created’ (Hiscock 2007:202).
The main criticism of typological analyses is the underlying assumption that each
morphological type represents a distinct and meaningful category to those who made
and used the tools. In fact, it has been demonstrated that one particular
morphological type – the geometric microlith – did not have one distinct function but
was rather used for a multitude of everyday tasks (e.g. Robertson 2009; Robertson
et al. 2009). Other backed artefacts have been shown to have similarly varied
functions (Fullagar et al. 2009; McDonald et al. 2007; Robertson 2005; Robertson
and Attenbrow 2008). Thus, despite the geometric microlith being a more
aesthetically pleasing form (and one that requires a greater investment of time to
manufacture), this does not necessitate its use solely for ‘higher’ functions. Similarly,
Taipale et al. (2014) and Knutsson et al. (2015) demonstrated that quartz pieces
previously classified as flaking debris had in fact been used as tools, without further
modification, again questioning the inferred link between form and function. In
contrast, a materialist approach classifies artefacts without reference to how (or
indeed if) they were used, merely how they were created (Hiscock 2007).
180
A materialist approach was employed herein as these systems accurately separate
different stages in the reduction sequence, so can be easily employed in concert with
Kuhn's (1995) provisioning systems that rely on technological attributes rather than
the presence of distinct artefact forms (see Chapter 2.4). The classification system
was modelled on Hiscock's (2007), with some minor adjustments to better reflect the
quartz-dominated assemblages found in the study area. The procedural key below
was used to evaluate potentially flaked stone, separating unflaked stone from flaked
fragments, cores, flakes and retouched flakes (Figure 9.2). Some unflaked stone
was non-local – where these pieces were unlikely to have arrived on-site via natural
processes, they were classified as manuports (Table 9.1).
Figure 9.2 Procedural key used to classify flaked stone via the materialist approach developed by
Hiscock (2007). Unflaked stone preserves no evidence of flake removals or having been detached
from a larger piece. Note that ventral surfaces were identified via any of the normal landmarks, such
as bulb of percussion, ripple marks, termination etc, following Holdaway and Stern (2004:108).
Retouch scars must be initiated from or modify the ventral surface – see Table 9.1 and Appendix D.2
for more information. Note the amendments for quartz-dominated assemblages, including the use of
impact points or shatter facets to identify flaked quartz, as well as flaked fragments incorporating
those pieces where breakages or poor-quality material make it difficult to confirm the
presence/absence of a ventral surface. Similarly, poor-quality material likely masked retouch in
several instances – where such uncertainty was encountered, pieces were classified as in the lower
category, as flakes.
181
Table 9.1 Artefact type definitions and the attributes recorded for each. Flaked fragments were
separated based on their maximum dimension, and flakes by their length.
Artefact
Attributes to record
Core
Any piece that exhibits a negative flake scar
and/or shatter facets, but no ventral surface.
Flakes have been struck from the core, but it
preserves no evidence of having been
detached from a larger piece.
1.
2.
3.
4.
5.
6.
7.
8.
9.
Flaked
fragment
Any piece preserving negative flake scars
and/or shatter facets but cannot be confidently
assigned as either a core or a flake – the
presence/absence of a ventral surface is
uncertain. This may incorporate broken pieces,
or those on such poor-quality raw material that
features are difficult to identify. Flaked
fragments have clearly been modified by
humans but cannot be classified further.
Small fragments (< 10 mm):
1. Raw material;
2. Combined weight and no;
Flake
1
Definition
Any piece that has been detached from a core,
and not retouched (though may exhibit other
edge damage, breakage). In contrast to most
definitions of a flake (e.g. Holdaway and Stern
2004), even those devised specifically for
quartz, flakes do not need to preserve evidence
of conchoidal fracture (e.g. bulb of percussion)
or fracture plane propagation (impact point,
zone of impact shatter), simply a ventral
surface indicating a previous attachment to a
larger piece. These may demonstrate edge
battering or breakage and may or may not have
been used.
Retouched
flake
A flake that has had further flakes struck from
it, i.e. where flakes have been initiated from
and/or modify the ventral surface or the
interface between the dorsal and ventral
surface (Hiscock 2007). Scars must be
complete, to distinguish retouch from previous
flake removals. Pieces exhibiting only
microscars or edge battering were not
considered retouched, since these result from
use or post-depositional processes.
Manuport
Pieces of non-local unmodified stone
= for complete/nominally broken pieces only
2
1.
2.
3.
4.
5.
6.
7.
Raw material;
Individual weight;
Maximum dimensions;
No. negative flake scars;
Grams/scar;
No. platforms used;
Bipolar;
Unmodified surface;
Other.
Large fragments (≥ 10 mm):
Raw material;
Individual weight;
Maximum dimensions;
No. negative flake scars;
Grams/scar;
No. platforms used;
Other.
Small flakes (< 10 mm):
1. Raw material;
2. Combined weight and no;
3. Breakage frequency.
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
Large flakes (≥ 10 mm):
Raw material;
Individual weight;
Oriented/max. dimensions;
No. negative flake scars1;
Mm/scar2;
Flake scar orientation1;
Unmodified surface1;
Bipolar;
Breakage;
Edge damage;
Other.
1.
2.
3.
4.
5.
6.
7.
Raw material;
Individual weight;
Oriented/max. dimensions;
Type of retouch;
Edge damage;
Breakage;
Other.
1. Raw material;
2. Individual weight.
= for orientable flakes only
182
Table 9.2 Stone artefact attributes – further information regarding methods can be found in Appendix
D. Negative flake scars < 5 mm in maximum/percussion length have been omitted as they more likely
result from retouch or use, and single scars can be very difficult to separate (and therefore quantify)
on all but very high-quality quartz. The references provided point to a definition of a particular
attribute, and its method of measurement.
Attribute
Summary method and/or variables
Raw material
Type: quartz (and quality), BIF, chalcedony, silcrete.
Weight
Individual or combined, depending on artefact type.
Reference
Oriented
Percussion Length of fracture plane (i.e. following the flaking axis) Holdaway and
dimensions length
from platform to termination in a straight line.
Stern
2004:138–140
Width at
At halfway point of percussion length line, at right
midpoint
angle.
Thickness Taken at the point where percussion length and width
at midpoint at midpoint intersect.
Maximum
Maximum
dimensions length
The longest line that can be drawn between two
points on the piece.
Maximum
width
Longest line that can be drawn at right angle to
maximum length.
Maximum
thickness
Longest line that can be drawn at right angle to
maximum thickness.
Breakage
For small flakes, the total number of broken pieces
was noted. For large flakes and retouched artefacts,
breakage type (transverse/longitudinal/marginal) and
the type of fragment remaining were noted.
Number of negative
flake scars
Complete and partial scars ≥ 5 mm percussion/max.
length, excluding retouch/damage scars.
Weight/length per scar
No. flake scars in relation to artefact size, classified
by percussion length (mm/scar) or weight (g/scar).
Negative flake scar
orientation
Quadrats (1–4) from which negative flake scars ≥ 5
mm have been initiated.
Number of platforms
No. platforms used to create flake scars.
Proportion of
unmodified surface
The proportion of unmodified surface evident on a
core or a dorsal flake surface: < 25%, 25–50%, > 50–
75%, > 75%.
Bipolar
Presence/absence/uncertain.
Edge damage
Presence of edge damage noted, and general
description (location, extent). Does not refer to
retouch or recent/fresh breaks (sieving etc).
Retouch type
Backing, unifacial retouch.
Other
Other notable features, e.g. presence of hafting resin,
original crystal facets, parallel flake scars.
Holdaway and
Stern
2004:138–140,
189
Hiscock 2002b
Holdaway and
Stern
2004:146–147.
Holdaway and
Stern 2004:144,
193
Holdaway and
Stern
2004:158–160.
183
The benefits of the materialist approach are clear – it removes the subjectivity from
flaked artefact classification and voids the argument about whether quartz artefacts
are best classified with a flint- or quartz-based typology. Nevertheless, poor-quality
quartz limited the visibility of certain features, likely artificially enhancing the number
of flaked fragments (i.e. flaked artefacts that cannot be classified further) and
reducing the number of retouched items where coarse textures mask retouch scars.
While not exclusive to quartz, bipolar flaking can sometimes split cobbles so that
positive and negative surfaces cannot be identified, i.e. both positive and negative
scars are lacking (Holdaway and Stern 2004:196). These items would likely be
classified as unflaked stone, unless other features were present. Nevertheless,
misclassification is less problematic than within typological systems as it is the
distribution of attributes that is the primary focus under a materialist approach, rather
than the frequency of certain artefact types.
9.3.3 Attribute Analysis
A number of metric and non-metric attributes were recorded for stone artefacts,
depending on both the classification and size of each piece (Tables 9.1–9.2; see
detailed methodological information in Appendix D). The primary aim of the attribute
analysis was to quantify reduction intensity (i.e. how heavily materials have been
worked and/or reworked) since this accurately separates Individual and Place
Provisioning strategies (Kuhn 1995; see Chapter 2.4). It was also important to
consider other features of relevance to Kuhn's (1995) technological provisioning
systems, as well as occupation intensity and assemblage composition. Specific
attributes were selected for their relevance to the aims above, as well as their ease
of observation on (often poor-quality) quartz, which dominated assemblages.
Most raw material types were fairly easily classified, since they represent those
commonly found in southwestern Australia. Unfamiliar types were identified following
standard non-invasive geological techniques, based on colour, texture, weight and
structure, as well as comparison to type specimens or general texts (e.g. Blatt and
Tracy 1996; Hefferan and O'Brien 2010; Perkins 2011). Four raw material types
were recognised, as below:
184
o
Quartz: This material varied widely from glassy smooth pieces to coarsetextured fragments with visible fracture planes, crystal aggregations and/or
flaws (Figure 9.3);
o
Silcrete: A metasedimentary rock comprised of reworked grains cemented
within a silica matric (Langford-Smith 1978:3). Grain size and angularity
varied considerably. Specimens with larger, more angular grains would be
more accurately classified as silicified granite but have been grouped with
silcrete as they derive from the same weathering profile, at a greater depth
(Taylor and Eggleton 2017);
o
BIF: Banded Iron Formation, chert and/or ironstone. These categories were
combined because they cannot be visually distinguished from small
samples, and may have derived from the same source rock, probably in the
greenstone belt east of the study area (Figure 3.4);
o
Chalcedony: Chemically identical to quartz but forms under different
conditions; it is differentiated from quartz by a waxy lustre. It possibly
originates from chalcedonic silica caprock outside the study area (Figure
3.4: Czj).
Raw material quality is a major factor that separates expedient technologies from
transported toolkits, with higher quality materials selected for the latter. It was
therefore necessary to identify these materials, since expedient strategies create
considerably more archaeological debris that may mask signs of alternate
provisioning systems. BIF and chalcedony were always high-quality, while silcrete
never was; quartz, however, was incredibly variable. High-quality quartz was defined
as the finer grained (or entirely smooth/glassy) material that lacked visible
flaws/fracture planes/crystal aggregations (Figure 9.3). Considering the quartzdominated assemblages, the quality classification of quartz was integral. However, it
must be treated with caution since, as noted above (see Chapter 9.3.1), quartz can
be highly variable within a single vein/specimen. Therefore, a high-quality quartz
piece may not always be indicative of the qualities of its parent core. Furthermore, it
is not always possible to accurately distinguish quartz quality at a macroscopic level
(Driscoll 2011a).
185
Figure 9.3 Variable quality of quartz within the study area. Quality increases from left to right. Top
row: very poor. Second row: poor–moderate. Third row: moderate–good. Bottom row: high quality.
White bars are 1 cm.
Reduction intensity was measured using the standard indicators: negative flake scar
and platform counts, proportion of unmodified surface (as quartz rarely displays a
cortex) and frequency of bipolar reduction (Table 9.2). Flakes/cores with more
unmodified surface represent earlier stages in the reduction sequence, while flake
scar and platform counts increase as the sequence lengthens as does the incidence
of bipolar reduction.
Retouch intensity was not investigated – while several measures accurately
characterise the loss of mass through retouch (e.g. Clarkson 2002; Kuhn 1990) the
link to reduction intensity only exists where retouch was performed for edge renewal
or artefact repair. In the study area, most retouched artefacts were backed, so the
186
retouch formed part of the initial manufacture sequence, and was not linked to
ongoing maintenance to extend artefact use-life. Similarly, standard indicators of
reduction intensity were not applied to retouched artefacts, as they had been heavily
modified from their original form.
Artefact size is often linked to reduction intensity, with larger pieces argued to
represent earlier stages in the reduction sequence since as a core reduces in size,
only smaller flakes may be detached from it (Clarkson and O'Connor 2014).
However, Lin et al. (2016) demonstrated that neither core size nor reduction intensity
patterned the size distribution of artefacts in a measurable way, because the
products are so overwhelmingly small that any increase in large flakes has a
negligible effect at an assemblage level. Therefore, artefact dimensions were only
linked to reduction intensity in certain instances, i.e. where a core was so small that it
could not be reduced further. Elsewhere, artefact size was accommodated in a
different way. Early stage analysis indicated that flake scar counts were heavily
biased by artefact size – larger artefacts (particularly flakes) were more likely to
exhibit higher scar counts than smaller artefacts, since they had larger surfaces on
which to preserve flake scar evidence. Furthermore, the same flake scar count may
have different implications for reduction intensity according to the artefact size. For
example, a core with six flake scars may indicate higher reduction intensity if it
weighs only a few grams, but far less intensive reduction if it weighs 50 grams. For
flakes, scar count was considered against percussion length, by dividing the total
length by the number of scars, to derive a mm/scar value. The same technique was
used for cores and large fragments, substituting weight for length since it better
represents size for more complex three-dimensional pieces. By allowing for artefact
size, these measures more accurately represent reduction intensity, and also reduce
the bias introduced by comparing larger artefacts to smaller ones.
Where present, a number of qualitative measures were also noted. Hafting resin –
normally a dark, solid encrustation on one or more margins – is an important feature
since it may indicate the use of a mobile, multifunctional toolkit incorporating
redundant features (Hiscock 1994; Torrence 1983). The presence of original crystal
facets indicates that a piece of quartz came from a large natural crystal. These
crystals would be ideal as transported cores, since the material therein is often
187
homogenous and of higher quality than when it comprises an aggregation of smaller
crystals. The presence of elongated, parallel flake scars was noted, as it may
indicate more formal, standardised reduction techniques that permit raw material
conservation. It was also important to quantify artefact breakage since it may be
indicative of discard behaviour. Further, transverse snaps are commonly associated
with trampling, which can be linked to greater occupation intensity (Hall and Love
1985; Hiscock 1988:22; Holdaway and Stern 2004:114). Severe edge damage was
noted, as it may have prompted artefact discard, or resulted from trampling.
9.4 OTHER ANALYSED MATERIALS
While lithics were the dominant form of cultural material found at the sampled sites,
charcoal was present, as were small quantities of European glass and ochre.
Sediment samples were also collected for contextual information, as soil conditions
can impact the preservation of organics. Each of these material types, and the
methods of analysis, are treated separately below.
9.4.1 Charcoal
Charcoal was bagged by XU and weighed. Certain samples were selected for
radiocarbon dating according to the concentration of cultural material within the
same XU (i.e. to date peaks in discard rate, or the earliest/latest cultural material),
the depth of the sample compared to the overall depth of deposit, and the quantity of
charcoal preserved. Where sample size permitted, in situ samples were chosen over
fragments recovered from sieve residue, as very small fragments can be blown into
test-pits during excavation or moved downwards through the deposit by insect
activity. Samples were sent to the Radiocarbon Dating Laboratory at the University
of Waikato, New Zealand – most were sufficiently small that AMS (Accelerator Mass
Spectrometry) dating was required. The samples selected and the returned age
determinations are provided in Chapters 10–12.
9.4.2 Sediments
Small sediment samples were collected from certain pits/XUs primarily for pH
measurements, as a soil that is acid (pH < 7), neutral (pH 7) or alkaline (pH > 7) can
influence the preservation of organics such as bone or charcoal (Barham and
188
Huckleberry 2014). It must be noted, however, that pH is just one of many variables
that influence the preservation of organics. For example, bone preservation is
dictated by intrinsic factors (those of the bone itself, i.e. density, age, porosity,
whether it had been cooked etc.) as well as extrinsic or environmental factors, such
as soil chemistry, pH, moisture content and temperature (Beisaw 1996:7–8; Gordon
and Buikstra 1981). Nevertheless, pH provides a broad indication of how well
organics may be preserved had they originally been present in archaeological
deposits. While bone is better preserved in more alkaline environments (pH > 7),
these same environments are less ideal for charcoal preservation (Barham and
Huckleberry 2014; Braadbaart et al. 2009; Gordon and Buikstra 1981).
Soil pH was tested with an over the counter colourimetric test kit. While such kits are
less accurate than pH readings obtained via electrometric methods, they are
commonly employed in archaeological analyses due to their accessibility (Barham
and Huckleberry 2014). Sediment samples were tested after the completion of
fieldwork, once materials had been transported to Perth. Specific samples and the
results of these are provided in Chapters 10–12. It is important to note, however, that
soil pH is a dynamic variable that results from the interplay of chemical and physical
properties of the soil and can therefore change fairly rapidly – current levels may not
be indicative of those in the past (Barham and Huckleberry 2014).
9.4.3 Ochre
Ochre is a natural pigment whose colour is derived from iron-oxides, iron-sulphides,
and other minerals suspended within a matrix, generally clay (Wallis et al. 2016). It is
fairly easily identified by vibrant colours, greasy texture and small particle-size, and
should produce a colour-streak when abraded against a surface (Wallis et al. 2016;
Watts 2010). Ochre was sorted by colour family (i.e. reds, yellows), and fragments
counted and weighed within each XU. No further analysis was undertaken, as
sourcing requires specialist equipment and techniques, as well as reference samples
(e.g. Green and Watling 2007; Scadding et al. 2015; Wallis et al. 2016).
9.4.4 Glass
Glass was roughly categorised as modern and historical – obviously modern glass
(i.e. Coca Cola bottles, beer bottles) was not collected unless it occurred below the
189
ground surface, where it may be indicative of post-depositional disturbance.
Historical glass requires evaluation to determine if it had been intentionally flaked, as
there is well-documented, continent-wide archaeological and historical evidence for
Aboriginal people flaking glass, primarily bottle glass (e.g. Allen 1969; Goward 2011;
Harrison 2000; Horne and Aiston 1928:11; Moore 1884:28–29; Paterson 1999;
Tindale 1941; Williamson 2004). Glass may have been desirable due to its
isotropism, meaning it flakes more predictably than most types of stone (Cotterell
and Kamminga 1987). Where flaked glass is present, this indicates that Aboriginal
people were visiting sites after European arrival and, where it can be determined,
after the date the vessel was manufactured. Within the context of this research,
worked glass has significant implications for site-use history, and is a valuable
method of indirectly dating deposits and assemblages.
Several factors impede the identification of intentionally flaked glass. First, the low
fracture toughness means that glass breaks more easily than stone and is more
susceptible to post-depositional damage, the results of which often mimic retouch
(Allen and Jones 1980; Beaumont 1961; Conte and Romero 2008; Cotterell and
Kamminga 1987). Furthermore, the inherent differences between glass bottles and
stone cores mean that flaking methodology can be varied. Thick bottle bases can be
used as cores (base core reduction) and flaked like stone, producing cores and
flakes with the standard landmarks (Goward 2011:16; Harrison 2000). The alternate
strategy, known as worked fragment reduction, involves simply smashing a bottle
and choosing the fragments best suited to a particular task, often without modifying
them further (Goward 2011:16; Harrison 2000). Unless they have been retouched
into a formal tool type, the products of such a strategy can be incredibly difficult to
identify – various studies have confirmed the use of expedient fragments through
microwear and residue analysis, even where these pieces lack macroscopic edge
damage (Goward 2011:22; Martindale and Jurakic 2006; Wolski and Loy 1999).
Various attempts have been made to define criteria against which to evaluate
potentially flaked glass (e.g. Allen and Jones 1980; Cooper and Bowdler 1988;
Goward 2011; Harrison 2000; Paterson 1999; Runnells 1976; Veth and O'Connor
2005; Williamson 2004). The bulk of these studies are concerned with the base core
reduction technique, particularly distinguishing retouch from post-depositional
190
damage. Despite decades of research, no set of universally applicable criteria has
been formulated and context is often considered the most important criterion of all
(e.g. Allen and Jones 1980; Beaumont 1961). Effectively, this means that the same
piece might be considered artefactual if it was discovered in an isolated, out of the
way location, while it would be evaluated more critically were it found in a high-traffic
(foot or vehicular) area.
Herein, glass was divided into modified and unmodified – further classification was
unnecessary as the focus was primarily on flaked glass as a means of dating
deposits or assemblages derived from Aboriginal occupation. Table 9.3 lists the
criteria used to evaluate glass fragments – the criteria selected were those relevant
to the small, fragmentary glass assemblage that was encountered. It is important to
note, however, that the absence of microwear and residue analyses means that
expediently used fragments (classified as modified glass) will be underrepresented in
the assemblage. Glass assemblages were too small to provide meaningful data on
reduction intensity – therefore, only basic descriptive attributes were recorded (Table
9.3). No attempt was made to date individual glass fragments.
Table 9.3 Features used to identify modified and unmodified glass, and the attributes recorded for
each. To be classified as modified glass, the piece needed to meet at least one of the criteria listed –
all other pieces were classified as unmodified. Criteria are based on Allen and Jones (1980), Goward
(2011:97), Harrison (2000), Holdaway and Stern (2004:30), Paterson (1999), Runnels (1976), Veth
and O'Connor (2005) and Williamson (2004). Maximum dimensions were recorded via the same
method used for stone artefacts (see Table 9.2 and Appendix D). Edge damage is only considered
evidence of modification only where the site is relatively undisturbed, i.e. receives little foot or
vehicular traffic.
Type
Modified glass
Criteria
o
o
o
o
Unworked glass
o
Features of a normal conchoidal flake (e.g. bulb of
percussion, etc);
Physical resemblance to known formal-tool type (e.g.
geometric microlith);
Edge damage (undisturbed sites only);
Unmistakable/unequivocal retouch, characterised by:
• Continuous flake removals (i.e. individual
scars are not separated by unmodified
areas), which may be overlapping;
• Restricted to one or two edges;
• Regular-sized flake removals;
• Uniform direction of force (though backing
may be bidirectional);
• Consistent scar patina i.e. flake removals
produced in a single event.
None of the above.
Attributes
1.
2.
3.
4.
1.
Individual
weight;
Maximum
dimensions;
Colour;
Type of
modification.
Combined
weight.
191
9.5 DATA ANALYSIS
Several steps were required in the data analysis process to accurately merge the
theoretical model with archaeological data. First, the longevity of site-specific water
sources had to be modelled. Then, a temporal framework had to be constructed to
evaluate change over time. It was then necessary to merge artefact data with
theoretical models by identifying technological signatures that should accompany
various provisioning systems and the indicators of occupation intensity therein.
9.5.1 Quantifying Site-Specific Water Sources
While the data from Chapters 6–7 are deemed representative of food resources and
soil water available at each archaeological site, Chapter 8 demonstrated that the
longevity and volume of water varied considerably according to gnamma
morphology. Therefore, it was necessary to make some attempt to more accurately
represent the quantity and longevity of water at each studied site, since this would
directly impact occupation span under various rainfall conditions. For each gnamma,
the following were considered:
o
Depth;
o
Surface area;
o
Shape (cylinder, elliptic cylinder, cone, inverted triangular prism – Figure
9.4);
o
Impact of above on evaporation coefficient;
o
Level of runoff (none, 2x, 5x).
It was fairly straightforward to track the daily water balance for cylinders and elliptic
cylinders, as each millimetre of depth consistently represents the same volume of
water; this is easily calculated using the structure's surface area. These gnammas
were modelled following the methods outlined in Chapter 5.4.3. Cones and inverted
triangular prisms, however, narrow with depth. Therefore, a given quantity of water
(e.g. a 5 mm rainfall event) will be stored over different vertical distances depending
on the water level at the time of the fall. Similarly, 5 mm of daily evaporation will
represent a different volume of water, depending on the depth at which water is
stored. As a result, both rainfall and evaporation must be considered in terms of
water depth and volume. Similarly, evaporation coefficients had to be altered for
192
these structures, since their sloping walls offer less protection by shadow-casting,
and water is exposed to direct sunlight for a greater portion of the day.
Figure 9.4 Shapes used to represent morphological variation between individual gnammas, and the
mathematical formulae used to determine their total capacity and the volume of water with depth.
193
Using formulae shown in Figure 9.4, which assumes that structures narrow at a
consistent rate, it was possible to construct an Excel spreadsheet to calculate the
volume of water held by each structure as the depth of water changed. The daily
water balance was then calculated using the following measures, in sequence:
o
Daily rainfall volume: surface area of structure x rainfall amount (allowing
for runoff multiplier), then converted to litres;
o
Pre-evaporation daily water volume: above added to the previous day's
volume (or zero, for the first model day);
o
Pre-evaporation water level: above converted to mm;
o
Post-evaporation water level: above minus daily evaporation with
appropriate coefficient applied;
o
Post-evaporation daily water volume: above converted to litres.
The water balance was tracked for each of the modelled years used previously,
representing periods of low, average and high winter rainfall. The water-holding
periods for each gnamma are provided in Chapters 10–12, along with the specific
values used to calculate them.
9.5.2 Constructing a Temporal Framework
A major goal of this research was to investigate change over time, which inherently
requires good temporal control within test-pits. In stratigraphically homogenous
deposits, the excavation unit (XU) forms the primary analytical unit. Standard
archaeological practice dictates that all material contained within a single XU is
considered contemporaneous, but it is likely a palimpsest of several occupation
events. Where available, radiocarbon dates were treated as the basal date for their
XU of origin. Calibration provides a date range within a certain confidence interval,
but this can complicate analysis of temporal change. Therefore, the midpoint of the
date range at 95.4% confidence (as cited on the data sheets returned by the
University of Waikato) was used as an indication of the calibrated basal date for a
particular XU. For consistency, all dates sourced before 2013 were recalibrated
using OxCal 4.2.4 (Bronk Ramsey 2009) and SHCal13, southern hemisphere
atmospheric data (Hogg et al. 2013). If an XU or surface accumulation contained
historical glass, these were presumed to post-date European arrival in southwestern
194
Australia (1826), though many likely post-dated the 1860s, when European people
were in the vicinity of Hyden (see Chapter 4.3).
Due to the financial cost and sample size requirements of radiocarbon dating, it is
rarely viable to obtain more than a few age determinations for a single excavated pit.
It was therefore necessary to construct a depth-age curve that effectively fills in the
blanks between radiocarbon dates by assuming a constant sedimentation rate – this
allows a broad date to be assigned to undated XUs. A curve was constructed for
each dated pit, but the data were not applied to undated portions of the site, as there
is no reason to expect sedimentation to occur at the same rate.
9.5.3 Identifying Technological Strategies
In order to interpret attribute data and archaeological remains, it was necessary to
construct technological signatures outlining what should be found at a site
characterised by Individual or Place Provisioning strategies. While there is a wealth
of literature on the subject (e.g. Clarkson 2004, 2006, 2007; Graf 2010; Kuhn 1995;
Mackay 2005; Mitchell 2016), the focus must be on those features preserved in the
archaeological record of the study area and identifiable within the data collected
during this research. It is also important to consider the subsidiary strategies that
may occur within the larger technological system, such as the manufacture of
expedient tools to supplement the transported toolkit required by the Individual
Provisioning strategy, or ‘gearing up’ by preparing a toolkit for the long logistical
forays that may be necessary under Place Provisioning (see also Chapter 2.4). The
archaeological literature rarely addresses the signatures associated with these
supplementary strategies, so some attempt was made to define these herein by
considering the similarities and differences between Gearing Up or Short-term
Expedient strategies (as they are referred to hereafter), and the dominant Individual
or Place Provisioning strategies.
It was first necessary to consider the major differences in technological organisation
between each of the four systems (Figure 9.5), as this will impact the attributes that
serve as key indicators. Raw material quality represents a good primary indicator,
since higher quality materials are required for transported toolkits. In contrast,
expedient technologies can accommodate poorer quality raw materials, which are
195
generally more abundant in the landscape. While high-quality material may be
stockpiled under Place Provisioning, its limited distribution would discourage
reduction in a more expedient manner, particularly since lower quality materials
would be readily available. Instead, it would probably be reserved for certain
artefacts, such as those required for logistical forays, when use-life must be
extended. Therefore, it was assumed that high-quality materials derived from
Individual Provisioning or Gearing Up, while the remainder represented Place
Provisioning or Short-term Expedient strategies.
Figure 9.5 Major differences in technological organisation between the four provisioning systems:
Place Provisioning (PP), Individual Provisioning (IP), Gearing Up (GU), and Short-term Expedient
(EXP).
It was then possible to separate the paired strategies based on aspects of
technological organisation. Because Place Provisioning accompanies longer site
visits, reduction sequences are longer, reduction intensity is higher, and some raw
material conservation may be required; the opposite is true under Short-term
Expedient strategies. While Gearing Up is a form of Individual Provisioning, it differs
196
in some important ways that should be visible at an assemblage level. In contrast to
traditional Individual Provisioning, the toolkit is not required for extended, ongoing
use, but rather for a particular journey, after which people will return to the same
base camp. Therefore, the location and availability of replacement stone is known;
this should translate to lower levels of reduction intensity and raw material
conservation. The reduction sequence is longer under Individual Provisioning but,
due to the higher residential mobility, only a small part of the sequence will be
represented at any particular site.
Using the technological contrasts set up above, it was possible to formulate a list of
attributes that should characterise assemblages arising from each provisioning
system (Table 9.4). Only the most reliable indicators were selected, as sample
size/bias was often an issue, especially for assemblages on high-quality raw
materials. The focus is on attributes rather than artefact type frequency, but in some
cases the latter was indicative. For example, retouched pieces will be rare under
Place Provisioning, but could occur if they were required for a specific task. In
contrast, they should never be found within Short-term Expedient strategies, since
formal tools would have been available in the transported kit.
These signatures were evaluated against the archaeological data. When
assemblages from several XUs shared the same characteristics, these were
grouped to increase sample size and improve the visibility of provisioning systems.
Nevertheless, small sample sizes precluded the use of more complex statistical
measures. Instead, the focus was on simple expressions of central tendency along
with graphical representations of data. Indeed, as Cowgill (2015) notes, these
simpler patterns and methods are often overlooked in favour of more complex
statistical techniques, to the detriment of broader research questions.
197
Table 9.4 Artefact attribute indicators for dominant and supplementary technological provisioning
systems, using the pairing system established in Figure 9.5. The indicators are only of value within
paired systems (Individual Provisioning and Gearing Up, Place Provisioning and Short-term
Expedient), since they describe the relative position of each strategy on a continuum, whose range of
values will be different to that of the alternate pair.
Attribute
Place Provisioning
Short-term Expedient
Discard rate/assemblage size
Larger
Smaller
Negative flake scar average
Higher
Lower
Negative flake scar range
Higher
Lower
Platform average
Higher
Lower
Platform range
Higher
Lower
g/scar
Lower
Higher
mm/scar
Lower
Higher
Bipolar core frequency
Higher
Lower
Retouched pieces
Some
None
Individual Provisioning
Gearing Up
Breakage rate
Higher
Lower
Heavy edge damage rate
Higher
Lower
Standardisation rate
Higher
Lower
Discard of usable artefacts
Lower
Higher
Bipolar core frequency
Higher
Lower
Negative flake scar average
Higher
Lower
Platform average
Higher
Lower
g/scar
Lower
Higher
mm/scar
Lower
Higher
Unmodified core surface
Lower
Higher
Attribute
198
9.5.4 Quantifying Occupation Intensity
A major aim of this research was to evaluate occupation intensity over time. Intensity
shifts in response to the number of visitors and visit duration, but these factors are
difficult to separate, so intensity is best considered in terms of the number of peopledays spent at a site. The dominant provisioning systems inherently indicate different
levels of occupation intensity: once visits become longer and more predictable, Place
Provisioning replaces Individual Provisioning. However, this switch merely means
that a certain threshold has been reached, as occupation may still further intensify
within a Place Provisioning system. Considering this strategy produces the most
archaeological debris, such shifts should be readily identifiable in the stone artefact
assemblages. A number of indicators were considered:
o
Evidence of Gearing Up;
o
Proportion of bipolar cores;
o
Range of negative flake scar counts;
o
Range of platform counts;
o
Range of unmodified surface;
o
Post-depositional breakage;
o
Artefact discard rates.
Most of these attributes reflect the length of the reduction sequence, so higher
values indicate the longer sequences that occur during more intense occupation
periods. Post-depositional breakage may be higher where activity is more intense,
while Gearing Up indicates restricted residential mobility and the need for long
logistical forays to supplement the local resource base. However, none of the
indicators are reliable enough to be used in isolation. For example, it is often hard to
separate post-depositional breakage from pieces that broke during use. Similarly,
Gearing Up activities may have occurred in a different part of the site and so be
absent in particular deposits, or may represent preparation for long trips unrelated to
resource procurement (i.e. aggregation events). If at least four high-intensity
indicators were identified, this was considered a sufficiently robust measure of
occupation intensity. The lack of any specific indicator was not considered
particularly important, as small pit sizes meant that some attributes (e.g. discard
rates, bipolar cores) will be sensitive to sampling biases.
199
While the data above relates to individual assemblages, it is necessary to consider
occupation intensity at a site-wide level, rather than treating each pit or area
separately. For example, Individual Provisioning may dominate in one pit but if
contemporaneous Place Provisioning assemblages were found in another part of the
site, then the former does not represent a large-scale reduction in occupation
intensity. Instead, it probably indicates that different parts of the site were used in
different ways, or at different times of year. Considering this, it was possible to
identify three levels of occupation intensity in the archaeological record:
o
Low: dominated by Individual Provisioning and/or Short-term Expedient
strategy across all deposits representing a particular occupation period.
Represents short, unpredictable occupation;
o
Moderate: Place Provisioning, with no/few high intensity indicators.
Represents longer, more predictable occupation;
o
High: Place Provisioning with at least four high intensity indicators.
Represents longest, most intense occupation.
9.6 CONCLUSION
The main archaeological sample used in this research comprises three locations
(Anderson Rocks, Gibb Rock and Mulka's Cave), all located within the Granite
Landscape Division. A fourth site – Twine Reserve – had to be abandoned due to
logistical and financial constraints. A number of 500 x 500 mm test-pits were
excavated at each site to permit intra-site spatial analysis without exceeding the one
square metre excavation limit. Lithic analysis methods were heavily influenced by the
properties of quartz, a material that dominates the study area assemblages. Quartz
is highly variable and can fracture irregularly, often lacking regular hallmarks of
intentional flaking. The ambiguities of this material are minimised within a materialist
approach rather than the standard typological classification schemes; under a
materialist approach, attribute data is of greater importance than artefact type
frequency. Attributes were chosen to reflect reduction intensity, as this is a key factor
in separating Kuhn's (1995) Individual and Place Provisioning systems. Other
materials (e.g. sediment, glass) are mostly of value in terms of quantifying the
preservation environment or dating deposits and assemblages. The formulated
200
technological signatures, covering both dominant and subsidiary strategies, will allow
attribute data to be interpreted within the framework of Kuhn's (1995) technological
provisioning systems. Occupation intensity can be evaluated by considering the
dominant provisioning system and the presence of high-intensity indicators, primarily
based on the length of the reduction sequence. Site-specific water data will allow
these data to be interpreted within the framework of resource availability.
201
CHAPTER 10
ANDERSON ROCKS
10.1 INTRODUCTION
This chapter discusses the landscape and archaeological content of Anderson
Rocks. To begin, the natural setting and characteristics of the outcrop are described,
before consideration turns to the availability of water in pans, soils and – most
significantly – the three gnammas. Field methods are then discussed, encompassing
the surface collections and the three excavated pits. The charcoal assemblage is
quantified, with specific reference to the influence of soil conditions on charcoal
preservation and the radiocarbon age determinations derived from selected samples.
The stone artefact assemblages are then described, focussing on the frequency and
distribution of artefactual stone and the major trends within attribute data. Finally, the
small ochre collection is briefly evaluated.
10.2 SITE DESCRIPTION
Anderson Rocks is 31 km north of Hyden, within the Granite Landscape Division.
This area is bordered by Thicket to the north and elsewhere by Heath, with the
surrounding Mallee at least 2.5 km away (Figure 9.1). The outcrop itself reaches a
height of 392 m ASL and covers approximately half a square kilometre. Several long
quartz veins are visible on the outcrop, measuring from 4.5 to 16 cm wide (Figure
10.1a–b). Anderson Rocks was originally classified as an 'other heritage place
(lodged)' on the DPLH register. This label applies to locations that are potentially
Aboriginal sites but have yet to be evaluated. During the course of this research DAA
(now DPLH) completed the evaluation and stated that Anderson Rocks did not meet
the criteria of a site under Section 5 of the Aboriginal Heritage Act 1972. The area is
therefore not subject to the conditions imposed by the AHA.
202
B
A
E
C
D
F
I
G
H
Figure 10.1 Features seen at Anderson Rocks – all photographs by A.M. Rossi. Individual pictured is
1.64 m tall. A–B: North-facing view of quartz vein striking north-south up the outcrop. C: Large pan
(10.9 m wide) on top of outcrop, facing east. Note the outflow channel leading downslope. D: Deep
channels created by water flow – these link many pans together in an interconnected system. E–F:
Pans linked by channels. Note the higher elevation of some pans that, when full, would overflow into
the channels directing water to lower lying structures. G: South-east facing view of large pan that
overflows downslope to gnammas (behind small circular patch of shrubs in top right of photo). This
pan receives runoff from upper slopes and channel systems. H: North-west facing view of the two
gnammas that receive runoff from upslope pan and channel system (which continues to flow after rain
has ceased), as well as the surrounding slope. Note that the well-defined flow of water does not reach
the third gnamma, which lies to the left of those pictured. I: South-east facing view of all three
gnammas – the black staining indicating inflow channels to the left and central gnamma. The smallest
gnamma (right) receives runoff only from its immediate surrounds. Largest gnamma (left) has a
diameter of 1.43 m.
203
10.2.1 Water Availability
There is an abundance of pans at Anderson Rocks, many of which are quite
expansive and, when full, resemble small pools. These pans are interconnected
system linked by deeply scoured channels created by water flow over millennia –
these channels distribute runoff and spill-over between pans (Figure 10.1c–f).
Nevertheless, the shallow depth of these pans (mostly < 100 mm deep) limit the
longevity of water and, like the modelled runoff pans these would generally be dry by
October (Figure 7.8). The single deeper pan (160 mm) holds water for slightly longer.
As elsewhere in the Granite Landscape Division, soils may store water for a month
or so beyond winter, but the precise longevity depends on the depth of the profile
and the amount of runoff received.
Three gnammas are present at Anderson Rocks; a few others are forming elsewhere
on the rock but are currently too small to be considered as water sources. All three
gnammas receive additional input through runoff, but the situation is further
complicated for the larger two gnammas, which are part of the interconnected flow
system noted above. In this case, runoff is not simply generated by a single slope or
an overlying pan, but rather a system of slopes, pans and channels that all
eventually lead to the gnammas (Figure 10.1g–i). The gradient of channels varies,
and the storage capacity of individual pans dictates when they will overflow,
releasing water into the system further down the outcrop. It is well outside the scope
of this thesis (and the expertise of its author) to reconstruct the incredibly complex
system of inflow and runoff, but it is important to note that the closest analogue –
high-input gnammas (Table 10.1) – is an oversimplification and will overestimate the
water level at some times, and underestimate it at others. The gnammas’ conical
structures provide less protection from direct sunlight, so an evaporation coefficient
of 0.65 was applied.
When full, the Anderson Rocks gnammas hold 593 litres of water (Table 10.1). Due
to the lower volume combined with increased runoff, the gnammas approach or
reach maximum capacity during winter, under all modelled rainfall conditions.
Gnammas reach their capacity earlier as the quantity of winter rainfall increases, but
the post-winter longevity is dictated by the amount of spring rainfall received (Figure
10.2). The influence of strong post-winter falls is clearly evident from the 2009, 2011
204
and 2012 model periods, where more water was available at 1 November than after
high winter rainfall. Isolated falls had a strong, albeit temporary, effect on water
balances due to the considerable runoff received by two of the three gnammas. For
example, the 11.6 mm fall on 13 December 2011 deposited over 150 L of water
across the three structures, but this balance reduced fairly rapidly due to greater
evaporative demand (Figure 10.2).
Table 10.1 Dimensions, capacity, shape, evaporation coefficient (EC) and modelled runoff conditions
for the three gnammas at Anderson Rocks. Depth and radius were measured for each structure and
used to calculate capacity following the standard mathematical formula for the volume of a cone (see
Figure 9.4).
Structure
Radius (mm)
Depth (mm)
EC
Runoff
Shape
Capacity (L)
Gnamma 1
420
222
0.65
2x
Conical
41
Gnamma 2
605
401
0.65
5x
Conical
153.7
Gnamma 3
715
744
0.65
5x
Conical
398.3
TOTAL
593
10.3 SURVEY AND EXCAVATION METHODS
In 2016, the entire south-eastern portion of the Anderson Rocks outcrop was
searched for potential cultural material. Border vegetation was generally too thick to
penetrate, so survey of the surrounding area was restricted to less densely
vegetated pockets. The western portion of the outcrop was omitted as there was no
vehicular access and thick vegetation separated the eastern and western
outcroppings at ground level (Figure 10.3). Three low-density accumulations of
surface material were found – Scatter 1 on a vegetated shoulder of the outcrop,
Scatter 2 at the base of a small pan (that also contained a quartz vein) and Scatter 3
in an open vegetation formation at the base of the outcrop (Figures 10.3–10.4).
These scatters were each collected either as a single unit or as a series of smaller
collections that were later merged into distinct collections based on their location
(see Chapter 9.2.2). A few isolated artefacts were found at other parts of the site and
were combined into a single unit for analysis.
205
high winter rainfall
700.00
1992
1998
600.00
water volume (L)
500.00
400.00
300.00
200.00
100.00
1 Dec
1 Nov
1 Oct
1 Sep
1 Aug
1 Jul
1 Jun
1 May
1 Apr
1 Mar
1 Feb
1 Jan
0.00
average winter rainfall
700.00
2009
2011
600.00
water volume (L)
500.00
400.00
300.00
200.00
100.00
1 Dec
1 Nov
1 Oct
1 Sep
1 Aug
1 Jul
1 Jun
1 May
1 Apr
1 Mar
1 Feb
1 Jan
0.00
low winter rainfall
700.00
2010
2012
600.00
water volume (L)
500.00
400.00
300.00
200.00
100.00
1 Dec
1 Nov
1 Oct
1 Sep
1 Aug
1 Jul
1 Jun
1 May
1 Apr
1 Mar
1 Feb
1 Jan
0.00
Figure 10.2 Total volume of water stored within Anderson Rocks gnammas during high (top), average
(middle) and low (bottom) winter rainfall years.
206
A
B
Figure 10.3 A: Aerial photograph of Anderson Rocks. White square indicates the area magnified in
image B. B: Main surveyed area of Anderson Rocks, showing the location of gnammas, artefact
scatters and pits. Note the thicker vegetation separating the eastern and western exposures. Features
too small to outline at the scale of this figure have been indicated by arrows. Base imagery from
Google Earth.
Three 500 x 500 mm pits were excavated in 2016, following the standard methods
outlined earlier, including the use of arbitrary excavation units (XUs) due to the lack
of visible stratigraphy (see Chapter 9.2.2). AR1 was positioned within Scatter 1, in an
area where vegetation was limited to low groundcover. The sediment was
waterlogged due to recent rains and became muddier as excavation progressed. At
360 mm below the surface (the base of XU13) water began intruding into the pit,
making continued excavation impossible. Large rock slabs were also present,
indicating that bedrock was probably nearby, so the pit was terminated and backfilled.
AR2 and AR3 were excavated within Scatter 3, approximately 40 m southeast of the
gnammas in an open area surrounded by thicker scrub/heath vegetation (Figures
10.3, 10.4d). Small roots and insect burrows were encountered within each pit –
these intrusions may have caused some vertical movement of very small pieces, but
there was no evidence of major disturbances. AR2 was excavated to a depth of 600
mm (XU24); cultural material was still present, but time constraints forced
termination and backfill of the pit. The author returned a few months later to excavate
AR3, which was positioned as near to AR2 as possible to intercept the same cultural
207
deposits while remaining out of the backfilled disturbance zone. XU33 was culturally
sterile and time constraints were again a factor, so the pit was terminated at -850
mm and backfilled. Due to their proximity, AR2 and AR3 are analysed as a single
unit, referred to hereafter as pit AR2-3. XU depths were identical throughout AR2
and AR3, except in one instance where two AR2 XUs (18–19) had to be combined to
accommodate a 50 mm XU in AR3. XU numbering follows AR3 since that was the
deeper deposit.
A
C
B
D
Figure 10.4 A: Vegetated area near Scatter 1 and pit AR1, facing north-east. Note that the scatter
and pit were immediately downslope of the thick vegetation pictured, where only low ground cover
was present. B: Scatter 2, located within a small pan bisected by a quartz vein, facing north. Pan
measures 1.63 m long by 0.91 m wide (maximum dimensions). C: close-up of the quartz vein in
Scatter 2. D: Location of Scatter 3 and pit AR2-3 within open vegetated area southeast of the
gnammas, facing south-west. Note the granite slab marks the position of the backfilled pit. All
photographs by A.M. Rossi.
208
10.4 ORGANICS AND DATING
Charcoal was far more common in pit AR1, and this pit yielded more than three
times as much charcoal (n=87.7 g) as AR2-3, despite the latter being a larger and
deeper pit. Soil pH values indicate that this discrepancy may be due to a
preservation bias. As noted elsewhere (see Chapter 9.4.2), alkaline environments
(pH > 7) are less conducive to charcoal preservation. The sediment in AR1 averages
pH of 5 while AR2-3 is much closer to, and even within, the alkaline end of the scale
(Table 10.2). Despite those same alkaline soil conditions being more favourable for
bone preservation, none was found. Within each pit, charcoal weight shows the
normal decline with depth, indicating the more recent material had been better
preserved (Figure 10.5). In AR1, charcoal was most abundant in XU4 and XU5,
which yielded 15.65 and 20.8 g, respectively. Over half of the XUs preserved more
than 5 g of charcoal. In contrast, just six XUs in AR2-3 had more than one gram of
charcoal. Charcoal weight peaked in XU9 that yielded 2.65 g.
Table 10.2 Soil pH values for selected XUs within AR1 and AR2-3.
Pit
XU
pH
AR1
1
5
5
5
10
5
13
5
1
6
5
6
10
6
15
6.5
20
7
25
6
30
6
33
6
AR2-3
209
weight (g)
0
5
10
weight (g)
15
20
25
0
AR1 - XU
1
2
3
4
5
6
7
8
9
10
11
12
AR2-3 - XU
13
1
2
3
4
5
6
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
Figure 10.5 Charcoal weight with depth in AR1 (left) and AR2-3 (right).
Five samples from AR2-3 were dated, but all samples originated from the deeper pit,
AR3. All were AMS dated due to small sample sizes, and only two in situ charcoal
samples were large enough to permit radiometric dating (Table 10.3). The upper four
dates are in sequence, with XU3 returning an age of 164 ± 20 BP, and XU25 5928 ±
20 BP. Interestingly, XU3 and XU10 are separated by at least 150 mm of vertical
deposit, yet their ages differ by just 45 radiocarbon years. The XU3 sample was
derived from a large chunk of charcoal in the north-eastern portion of the pit – this
fragment continued into the pit wall and unexcavated deposit beyond. Charcoal was
only removed and collected where it occurred within the AR3 pit boundary, so it is
possible that a small fragment of the remainder became dislodged during excavation
and was introduced into XU10. The lowest sample, from XU32, returned an age of
2471 ± 20 BP, and is out of sequence with all other dates (Table 10.3). It is possible
that the small sample of charcoal had moved vertically down the pit, possibly due to
the root and insect activity noted above. Those dates considered more reliable (Wk44677, Wk-44679 and Wk-44680) were used to construct a depth-age curve, which
indicated that the lowest cultural material in AR2-3 dates to around 8200 cal. BP
(Figure 10.6). Sedimentation rates were highest in the top half of the pit and slowed
below XU20.
210
Table 10.3 Charcoal samples and their returned radiocarbon ages, from pit AR2-3. Dates are
arranged by XU depth. Asterisk beside sample weight indicates in-situ sample – all others derived
from sieve residue. Note that no 13C measurements were provided with results.
XU
Depth
(mm)
Sample
number
Weight
(g)
%Modern
C14 age
BP
Calibrated age BP
3
50–75
Wk-44677
0.25*
98 ± 0.2
164 ± 20
266–0 (68.2%)
275–0 (95.4%)
10
225–250
Wk-44678
0.1
97.5 ± 0.2
207 ± 20
280–149 (68.2%)
290–74 (95.4%)
20
500–525
Wk-44679
0.15
62.8 ± 0.2
3734 ± 20
4083–3982 (68.2%)
4145–3926 (95.4%)
25
625–650
Wk-44680
0.1*
47.8 ± 0.1
5928 ± 20
6735–6671 (68.2%)
6780–6645 (95.4%)
32
800–825
Wk-44681
0.1
73.5 ± 0.2
2471 ± 20
2675–2362 (68.2%)
2699–2353 (95.4%)
cal. BP
0
1000
2000
3000
4000
5000
6000
7000
8000
9000
0
50
100
150
200
250
300
depth (mm)
350
400
450
500
550
600
650
700
750
800
850
Figure 10.6 Depth-age curve for AR2-3. Solid line joins radiocarbon age determinations, while linear
trendline (dotted line) was used to estimate the age of the lowest cultural material, from XU32 (-825
mm).
10.5 STONE ARTEFACT ASSEMBLAGE
A total of 821 stone artefacts were recovered from Anderson Rocks. Most of these
came from the three pits, particularly AR2-3 which contributed over 80% of artefacts
(Table 10.4). The assemblage was heavily quartz dominated (96%), and most quartz
211
was of moderate quality (i.e. exhibiting some flaw planes or granularity in texture),
but there were some examples of extremely poor and very fine-quality material. The
non-quartz artefacts comprised seven silcrete, nine chalcedony and 16 BIF pieces.
The results of each pit and surface collection are discussed below, beginning with
the surface material. The focus is on the attributes integral to interpretations of each
pit/area; full artefact data can be found in Appendix E.
Table 10.4 Number of stone artefacts from Anderson Rocks pits and surface collections.
Pit/Collection
Quartz
Silcrete
BIF
Chalcedony
TOTAL
AR1
95
3
2
0
100
AR2-3
647
3
14
8
672
Scatter 1
9
0
0
0
9
Scatter 2
9
0
0
0
9
Scatter 3
24
1
0
1
26
Other surface material
5
0
0
0
5
789
7
16
9
821
TOTAL
10.5.1 Surface Collections
Just 49 artefacts were recovered from surface collections at Anderson Rocks, more
than half from Scatter 3. Scatters 1 and 2 each yielded nine pieces, and five were
recovered from other parts of the site. Most of the artefacts were made on quartz
(n=47), and the two non-local materials (silcrete and chalcedony) occurred within
Scatter 3. Similarly, all of the high-quality materials were present in Scatter 3, where
they outnumbered artefacts made on lower quality stone. All artefact types were
present at Scatter 3, while the remaining collections preserved more restricted
assemblages (Table 10.5).
Table 10.5 Artefact types in surface collections at Anderson Rocks. Figures in parentheses refer to
high-quality materials. Coll. = collection. Sc. = Scatter. OSM = other surface material, from outside
defined scatters.
Core
Small
fragment
Large
fragment
Small
flake
Large
flake
Retouch
Manuport
TOTAL
Sc.1
2
1
1
0
3
1 (1)
0
8 (1)
Sc.2
1
0
4
0
4
0
0
9
Sc.3
1
(2)
(2)
3 (1)
5 (9)
(2)
1
10 (16)
OSM
1
0
2
0
1
1
0
5
TOTAL
5
1 (2)
7 (2)
3 (1)
13 (9)
2 (3)
1
32 (17)
Coll.
212
Cores
Only five cores were present, all quartz of moderate–good quality. Two were found
at Scatter 1, and one each in all other surface collections (Table 10.5). The largest
core weighed 46.65 g, was made on the poorest quality material, and was found
outside the defined scatters. Elsewhere, cores weighed no more than 5.3 g, and
were of much finer textured quartz. Each core exhibited between five and nine
negative flake scars, 0.38–5.18 g/scar, and the maximum values occurred on the
largest core; all cores had less than 25% of their surface unmodified. Between two
and four platforms were evident on each specimen, with the maximum recorded on
three of the five cores. Bipolar cores were exclusively found in Scatter 1 and the
surrounding material, while Scatters 2 and 3 yielded only freehand cores (Appendix
E.1).
Flaked Fragments
Only three small fragments were present, unsurprising due to the bias against small
items being noticed on the ground surface. All were quartz, and one was found in
Scatter 1 and the remainder in Scatter 3. The Scatter 1 fragment was poorer quality
quartz (quite granular in nature) while those from Scatter 3 were on higher quality
material.
Nine large fragments were found on the surface, all quartz. Four were recovered
from Scatter 2, including three on very poor-quality raw material. All other collections
yielded one or two examples, and high-quality material was restricted to Scatter 3.
One notable example was found in the 'other surface material' – this piece was fairly
rounded compared to other fragments, and the ridges between negative flake scars
also appear to have been rounded. It possibly represents a core or fragment
subsequently ingested by a bird (as a gizzard stone), which may explain its
occurrence outside a defined scatter. As the identification was contentious, it was
classified as a fragment, following the methods described earlier (see Chapter 9.3.2).
It also preserved seven negative flake scars, while all other large fragments had no
more than three (Appendix E.3). At Scatter 3, artefacts averaged 2.5 scars, 0.15
g/scar and two platforms. Elsewhere, reduction intensity was lower, when the large
potential core is excluded (avg. 2 scars, 1.19 g/scar, 1.6 platforms).
213
Flakes
Only four small flakes were present, all quartz and all from Scatter 3; two were made
on high-quality quartz (Appendix E.4). A total of 22 large flakes were present,
comprising 21 quartz and a single chalcedony flake. Large flakes were particularly
common at Scatter 3, where 14 were found, including the nine on high-quality
materials. All other collections yielded between one and four examples, and three
from Scatter 2 were made on very poor-quality quartz (Table 10.5). Reduction
intensity was similar across higher and lower quality materials. The latter averaged
2.9 scars, 6.21 mm/scar and 1.8 platforms, while high-quality materials averaged 2.6
scars, 6.12 mm/scar and 2.2 platforms. Broken (n=5) and heavily edge damaged
flakes (n=3) were all confined to Scatter 3 (Appendix E.5).
Retouched Flakes
Five retouched flakes were found on the surface; all were backed and made on
quartz. Three high-quality quartz specimens were present, including both pieces
from Scatter 3 and another from Scatter 1. None of the retouched artefacts were
broken or had suffered heavy edge damage (Appendix E.6).
Manuports
A single manuport was present at Scatter 3. This piece was a very weathered chunk
of silcrete, weighing 27.3g, with a distinctly rounded surface. The material was quite
coarse and poor quality, so the absence of flaking seems unsurprising. It may have
been a hammerstone or grindstone, but no pitting or battering was evident. However,
the coarse-textured material would make such features difficult to identify. In any
case, the piece was certainly transported to the site by humans.
10.5.2 AR1
AR1 yielded 100 stone artefacts, with all types present except cores (Table 10.6).
Generally, artefacts became more numerous with depth until their peak in XU9
(n=18), but the decline in XU6–7 is noteworthy. The basal portion of the pit (XU10–
13) contributed comparatively few stone artefacts, with no single XU yielding more
than six pieces. Just four non-quartz artefacts were present (three silcrete, one BIF)
in AR1, between XU4 and XU9 (Table 10.6).
214
Table 10.6 Stone artefact types by XU in AR1. Figures in parentheses refer to high-quality raw
materials.
XU
Core
Small
fragment
Large
fragment
Small
flake
Large
flake
Retouch
TOTAL
1
2
3
4
5
6
7
8
9
10
11
12
13
TOTAL
0
0
0
0
0
0
0
0
0
0
0
0
0
0
3
3
3
1 (3)
7
3
7
8 (1)
8
2
4
1
1
51 (4)
0
0
1
0
0
0
0
1
4
1
0
0
2
9
0
1
4
(1)
1 (1)
1
0
1 (1)
4 (1)
(1)
(1)
(3)
0
12 (9)
1
(1)
0
3
0
0
0
(1)
1
0
0
(1)
0
5 (3)
0
0
1
1
(1)
0
0
2
0
(1)
(1)
0
0
4 (3)
4
4 (1)
9
5 (4)
8 (2)
4
7
12 (3)
17 (1)
3 (2)
4 (2)
1 (4)
3
81 (19)
Flaked Fragments
A total of 55 small fragments were present in AR1, comprising 54 quartz and one BIF.
Unsurprisingly, many were poor-quality quartz, precluding their identification as
either flakes, cores or retouched pieces; only four were made on high-quality
materials. Numerically, small fragments were most abundant in XU7–9 (n=24) but
contributed over 60% to each assemblage between XU5 and XU8 (Table 10.6).
Just nine large fragments were present in AR1, all quartz of generally poor quality.
These artefacts were most common in XU9 (n=4) but comprised a greater proportion
of assemblages between XU9 and 13 (Table 10.6). Most weighed less than 0.6g, but
one from XU3 weighed 10.45 g and was very poor-quality quartz. It may have been
tested as a core and subsequently abandoned due to poor flaking properties; the
presence/absence of a ventral surface could not be confirmed due to the material’s
poor quality. Large fragments exhibited between one and three negative flake scars
each, more commonly two or three, 0.29 g/scar (excluding the largest fragment) and
one to two platforms. However, poor-quality raw material meant that scar and
platform count could not be determined for nearly half of the specimens (Appendix
E.3).
215
Flakes
AR1 yielded 21 small flakes, which were most numerous in XU9 (n=5) and XU3
(n=4). They generally contributed between 10 and 30% to each assemblage, but a
greater proportion were present in XU3 (> 40%) and XU12 (approx. 60%). Only nine
small flakes were made on high-quality materials. These were concentrated in XU8–
12, and five were broken.
Only eight large flakes were present in AR1, and five of these were from the upper
four XUs (Table 10.6); five were on moderate quality quartz and silcrete, and three
on high-quality quartz. Of the three high-quality flakes, one was broken and another
exhibited heavy edge damage. These flakes averaged 2 negative flake scars, 6.31
mm/scar, and 1.5 platforms. The remaining flakes displayed 1–3 scars (avg. 2.2),
3.29–26.57 mm/scar (avg. 9.49 mm/scar) and 1–2 platforms (avg. 1.6) – none
exhibited heavy edge damage, and only one was broken (Appendix E.5).
Retouched Flakes
Retouched flakes were the least common artefact type in AR1 – only seven were
present, all quartz, and all backed. Three of these occurred between XU3 and XU5
and the remaining four between XU8 and XU11. Three were made on high-quality
quartz, while the remainder were on moderate–good quality material. The latter were
restricted to XU3–8, while high-quality materials were more common in the lower
XUs (Table 10.6; Appendix E.6). Four retouched flakes were broken, including all
three made on high-quality quartz.
10.5.3 AR2-3
A total of 672 stone artefacts were present in AR2-3, but the basal unit (XU33) was
culturally sterile. Most artefacts were concentrated in XU12–23 (n=518), and only 34
were found in XU24–32 (Table 10.7). However, XU thickness and pit size varied with
depth, so discard rates provide a more accurate indication of artefact concentrations
(Figure 10.7). Cultural material accumulated fairly slowly in the earlier, basal
deposits, with discard rates frequently below 5 artefacts/100 years. The highest rates
occurred in XU17–20, around 2950–4035 cal. BP, and decreased until the European
period (Figure 10.7). Artefacts were mostly quartz (n=647), but over half the
216
assemblage was made on high-quality materials (n=364) – these materials
dominated in two artefact categories: small flakes and retouched pieces. High-quality
materials outnumbered others in XU9–10, XU15–19 and XU29–32 (Table 10.7). BIF
and chalcedony artefacts were particularly common in XU13–20 (n=14/22) but were
also present elsewhere in the pit. The large artefact assemblage from AR2-3 means
that change over time can be more easily addressed; as a result, some of the data
are presented in greater detail.
Table 10.7 Stone artefact types by XU in AR2-3. Figures in parentheses refer to high-quality raw
materials.
XU
Core
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
TOTAL
0
0
0
0
0
1
0
(1)
(1)
0
0
0
1
0
(1)
2
2
2
(1)
1
(1)
0
0
0
0
0
0
0
0
0
0
0
9 (5)
Small
fragment
(1)
(1)
1
0
(1)
5 (2)
1
0
1
(3)
1
2
1 (2)
0
5
2 (1)
(2)
16 (9)
4 (1)
2
(3)
2
(3)
0
1
1
0
1
(1)
0
(1)
0
46 (31)
Large
fragment
0
1
0
0
1
1
(1)
0
0
0
1
0
0
2
0
0
(1)
3
6
1
0
0
2
0
1
0
0
0
0
0
0
0
19 (2)
Small
flake
0
1 (1)
4 (3)
(3)
3 (6)
7 (1)
9 (1)
4 (4)
(10)
(8)
(7)
7 (7)
2 (10)
10 (10)
12 (20)
8 (9)
6 (26)
16 (56)
3 (33)
10 (20)
15 (11)
5 (17)
10 (1)
1
5
1 (1)
(2)
1 (1)
(1)
(3)
(1)
(2)
140 (275)
Large
flake
0
0
1
3
(1)
5 (3)
5
1
0
0
1 (1)
4 (1)
4 (1)
4 (2)
3 (1)
1 (5)
5 (1)
14 (5)
3 (5)
13 (4)
6 (2)
0
3
3
1
0
0
0
0
2 (1)
(1)
(1)
82 (35)
Retouch
TOTAL
0
0
0
0
0
0
0
0
0
(1)
(1)
0
(1)
1
(1)
2 (1)
(2)
4 (6)
1 (1)
2 (1)
1 (1)
0
1
0
0
0
0
0
0
0
0
0
12 (16)
(1)
2 (2)
6 (3)
3 (3)
4 (8)
19 (6)
15 (2)
5 (5)
1 (11)
(12)
3 (9)
13 (8)
8 (14)
17 (12)
20 (23)
15 (16)
13 (32)
55 (76)
17 (41)
29 (25)
22 (18)
7 (17)
16 (4)
4
8
2 (1)
(2)
2 (1)
(2)
2 (4)
(3)
(3)
308 (364)
217
no. artefacts/0.5 m2/ 100 yrs
0.00
5.00
10.00
15.00
20.00
25.00
30.00
35.00
1
2
3
4
5
6
7
8
9
10
11
12
13
14
AR2-3 XU
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
Figure 10.7 Artefact discard rates in AR2-3. The rate has been adjusted to artefacts/0.5 m2/100
years to allow for the difference in excavated area in the lower deposits, as only pit AR3 continued
below 600 mm.
Cores
A total of 14 cores were present, all quartz, including five on high-quality quartz
(Table 10.7). Cores were concentrated in XU15–21 (n=10), where each XU had one
to two specimens; none were found below XU22. Size and reduction intensity varied
between the standard and high-quality quartz cores. The latter averaged 5.4 scars,
0.51 g/scar, and 3.4 platforms. Four of the five cores weighed < 2.5 g and were
bipolar – reduction technique could not be determined for the heaviest core; only one
core had more than 25% of its surface unmodified. While other quartz cores had a
higher scar average (n=5.89), other reduction intensity values were lower (avg. 0.95
g/scar, 3.1 platforms). These cores were considerably heavier, and only two weighed
218
< 2.5 g, both from XU16. Three had more than 25% of their surface unmodified, and
five were bipolar, concentrated in XU16–20 (Appendix E.1).
Flaked Fragments
A total of 77 small fragments were recovered from AR2-3, all quartz, including 31 on
high-quality quartz (Table 10.7). Small fragments were most common between XU15
and XU23, where 50 occurred. Half of these were found in XU18, which yielded five
times as many small fragments as most other XUs; only five were present in XU24–
32.
Large fragments were comparatively rare in AR2-3 – only 21 were recovered,
accounting for just 3.1% of the entire artefact assemblage. All were on quartz,
generally of moderate quality, and only two high-quality specimens were present.
Large fragments were concentrated between XU14 and XU23, where 15 of the 21
specimens occurred, and especially common in XU18–19 (n=9); none were found
below XU25. All large fragments displayed 1–4 negative flake scars, most commonly
2–3, and averaged 0.45 g/scar. Over 70% were dual-platform fragments, while the
remainder had no more than three platforms (Appendix E.3).
Small Flakes
Small flakes were the most common artefact type in AR2-3 and were present in most
XUs. Of the 415 small flakes recovered – 402 of which were quartz – 287 were found
between XU14 and XU22 (Table 10.7). Over 65% of small flakes were made on
high-quality material, so standard-quality small flakes were the minority in most XUs
(Table 10.7). Over 50% of the small flakes were broken; separate breakage data
were not available for standard- and high-quality materials, but the proportion of
broken flakes with depth follows a similar pattern to the proportion of high-quality
materials (Figure 10.8). This may indicate that those flakes were more likely to be
broken than those made on lower quality material.
219
% high-quality
20
40
60
% broken
80
100
0.00
0
4
4
8
8
12
12
16
AR2-3 XU
AR2-3 XU
0
0
20.00
40.00
60.00
80.00
100.00
16
20
20
24
24
28
28
32
32
Figure 10.8 Proportion of small flakes on high-quality material (left), and breakage rate for all small
flakes (right) from AR2-3.
Large flakes
A total of 117 large flakes were recovered from AR2-3, including eight on non-local
raw materials. Large flakes were most common between XU12 and XU21, where
over 70% of the assemblage was found (n=84). They were especially concentrated
in XU18–20, but were proportionally more significant in XU20, where they comprised
over 30% of the assemblage. The distribution of high-quality materials (n=35) largely
followed this same pattern, as they represented < 30% of the large flake assemblage
(Table 10.7). High-quality materials are discussed separately to lower quality
materials, below.
Flakes on high-quality raw material displayed 1–5 flake scars, 1–3 platforms, and
averaged 5.55 mm/scar. However, there was some variation with depth. The highest
scar averages occurred in XU14–15, but values were also fairly high in XU20 and
XU31–32 (Figure 10.9a). Platform averages also peaked in XU15, while the lowest
averages were concentrated in XU21 and XU30–31. Higher scar averages were
generally accompanied by higher platform averages, but this was not always the
case (i.e. XU14 – Figure 10.9a). Mm/scar values were generally higher (indicating
220
lower reduction intensity) in the upper deposits, but sample sizes were small in those
XU. Generally, most XUs averaged < 6 mm/scar, and values were below 4 mm/scar
in XU14–15 (Figure 10.9c). Five flakes had long, parallel narrow flake scars,
indicating they may have derived from formal, standardised cores. Two were present
in XU21, and one each in XU6, XU14 and XU18. Only three large flakes had
suffered serious edge damage, but the breakage rate was considerably higher, over
65% of those flakes where breakage/lack of breakage could be confidently identified.
In fact, all flakes were broken in nine of the 13 XU that contributed breakage data –
these maximum breakage rates were concentrated between XU11 and XU17 (Figure
10.9d).
number
0.5
1
1.5
2
2.5
number
3
3.5
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
4
0
4.5
No. NFS
No. platforms
AR2-3 XU
AR2-3 XU
0
4
6
8
10
1.5
2
2.5
3
3.5
4
4.5
No. NFS
No. platforms
% broken
12
14
16
0
18
high-quality
other
AR2-3 XU
AR2-3 XU
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
2
1
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
mm/scar
0
0.5
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
10
20
30
40
50
60
70
80
90
100
high-quality
other
Figure 10.9 Selected attributes recorded on large quartz flakes, by XU, in AR2-3. A–B (top): Average
negative flake scar (blue bars) and platform counts (orange bars) for high-quality materials (A – left)
and lower quality materials (B – right). C–D: average mm/scar values (C – left) and proportion of
broken flakes (D – right) for high-quality materials (black bars) and lower quality materials (grey bars).
221
Among the lower quality materials, individual values were more varied. Each
exhibited 0–7 scars and 1–4 platforms, but flakes were generally larger (avg. 6.73
mm/scar). Flake scar and platform averages showed similar patterns – both peaked
in XU8 and remained fairly high in XU12–15 (Figure 10.9b). Ranges were highest in
XU18 and XU20, which yielded the largest flake assemblages. Mm/scar values were
consistently higher than in the high-quality assemblage, and averages reached 16.72
mm/scar. However, most values were < 8 mm/scar; they were lowest in XU8, and
peaked in XU16–17, and XU23 (Figure 10.9c). Two flakes from XU20 had a patina,
possibly formed by weathering, on their dorsal surface. This coating covered some
negative flake scars (distinguishing it from cortex or unmodified surface), while other
flake removals had exposed fresher material beneath. The parent core may have
been exposed for a considerable amount of time between flaking episodes.
All unmodified surface categories were represented, and 12 flakes had more than 25%
of their surface unmodified; six of these flakes came from XU20. A total of 32 flakes
were bipolar – the technique was especially dominant in XU7 and XU12–14, but was
also fairly common in XU18 and XU20, where sample sizes were larger. Ten flakes
had suffered severe edge damage, but breakage rates were fairly low. Of the 18 XU
that contributed data, seven yielded no broken flakes, and most had less than 50%
of their flakes broken. Breakage rates were consistently low from XU18–21, and only
whole flakes were present from XU23–25 (Figure 10.9d).
Retouched Flakes
A total of 28 retouched pieces were present in AR2-3, of which 16 were made on
high-quality materials, including three on chalcedony. All retouched artefacts were
found between XU10 and XU23, and over half (n=15) were from XU18–20. Ten were
present in XU18, with no more than three in any other XU. Proportionally, however,
retouched items were the most significant in XU16 where they comprised nearly 10%
of the assemblage. Lower quality materials were particularly common in XU16–20
(n=9; Table 10.7).
The high-quality materials were predominantly backed (n=14). The remaining two
pieces were unifacially retouched, but this retouch also served to blunt an edge; it
was not steep enough to be classified as backing. Only one of the high-quality
222
pieces had suffered severe edge damage, but seven were broken, including five
from XU18. Among the lower quality materials, eight artefacts were backed, while
four exhibited unifacial retouch. Only four of these retouched pieces were broken,
and another exhibited significant edge damage (Appendix E.6).
10.6 OTHER CULTURAL MATERIAL
Aside from stone artefacts, ochre was the only other cultural material present at
Anderson Rocks. It was restricted to AR2-3 where it was fairly scarce, with a total of
just 0.75 g recovered. Three colours were represented in varying quantity: brownishyellow (0.4 g), red (0.3 g) and the scarcest, a brownish-red (< 0.05 g). The samples
are too small for any real analysis, but the vertical distribution is noteworthy. Ochre
was primarily concentrated in XU18–22, where all but 0.15 g was recovered. Only
one other XU – XU8 – yielded any ochre (Table 10.8).
Table 10.8 Location, weight and colour of ochre samples from AR2-3.
Pit
XU
Depth (mm)
Weight (g)
Colour
AR2-3
8
175–200
0.15
red
AR2-3
18
425–475
0.05
brownish-red
AR2-3
21
525–550
< 0.05
brownish-red
AR2-3
22
550–575
0.15
red
AR2-3
22
550–575
0.4
brownish-yellow
10.7 CONCLUSION
The most long-lived water source at Anderson Rocks are the three gnammas, two of
which are also fed by the system of interconnected channels and pans that filter
runoff down the outcrop. These gnammas fill to capacity even after low winter rainfall,
and the longevity of water is determined by spring rainfall, which offsets water-loss
through evaporation. Isolated falls (> 10 mm) create a considerable, but temporary,
increase in water availability. Three pits were excavated (two of which intersect the
same cultural deposits) at Anderson Rocks and a number of surface collections
made around the site, but artefacts were most common in the excavations. Charcoal
was scarce in AR2-3 but fairly abundant in AR1, probably due to soil conditions – the
latter pit was not dated due to a comparative lack of artefacts and a shallower
223
deposit. The five radiocarbon dates from AR2-3 are largely in sequence, with the
lowest reliable sample dating XU25 to 5928 ± 20 BP, around 6780–6645 cal. BP.
The depth-age curve indicates that cultural material began to accumulate around
8200 cal. BP.
The stone artefact assemblage was heavily quartz-dominated, but only AR2-3
provided a stratified sample large enough to evaluate change over time. High-quality
quartz was more common, especially among small flakes and retouched pieces.
Small flakes and small fragments dominated the assemblage in the upper portion of
AR2-3, with few cores or retouched pieces present; all artefacts had less than 25%
of their surface unmodified. Most artefacts were discarded in the middle XUs (XU12–
23), and the bulk of non-quartz artefacts were also found here, along with most of
the cores and retouched pieces. Artefact discard rates were highest, particularly in
XU17–20. Reduction intensity often peaked in these same XU, as measured by
negative flake scar and platform counts, and weight/length per scar. Broken and
damaged artefacts were also more common than elsewhere in the pit. Unmodified
surface values were more varied and reached their maximum levels in the middle XU,
among the lower quality materials. Most of the ochre – which was confined to pit
AR2-3, was also found within these XU. In the lower portion of the pit, artefacts were
comparatively scarce – no cores were present, and few large fragments or retouched
pieces were found.
224
CHAPTER 11
GIBB ROCK
11.1 INTRODUCTION
This chapter focusses on the physical characteristics of Gibb Rock and the
archaeological material derived therein. To begin, Gibb Rock and its environmental
setting are outlined, with specific relation to the Landscape Divisions represented in
the wider area and the availability of water at the site itself. The survey, surface
collection and excavation methods are then detailed. Finally, the archaeological
material is quantified and described, beginning with organic material before turning
to the major archaeological content – the stone artefact assemblage – and the
attribute data collected.
11.2 SITE DESCRIPTION
Gibb Rock is approximately 40 km NNE of Hyden, in the central part of the study
area where Thicket and Heath meet (Figure 9.1). The rock and its immediate
surrounds fall within the Granite Landscape Division, but beyond this lies an
extensive boundary of Thicket. Vegetation shifts to Mallee (with pockets of Woodland
– see Figure 3.14) around 800 m from Gibb Rock on the western and southern sides.
The outcrop itself peaks at 416 m ASL and covers an area approximately 0.68 km2.
The granite is fairly coarse-grained, and large chunks of quartz were observed
weathering from the parent rock in several locations. Gibb Rock is currently a water
reserve and does not appear on the DPLH register.
11.2.1 Water Availability
No gnammas are present at Gibb Rock, but shallow pans are common (Figure 11.1a)
– however, stored water would be short-lived (see Chapter 7.3.4). At present, a
portion of the outcrop has been fenced off and a wall erected to capture runoff and
funnel it to storage tanks for agricultural use (Figure 11.1b). Previously, water would
have flowed down the outcrop and been stored in the surrounding soils. These would
have held water for longer than pans, but the specific water-holding period would be
225
defined by the depth of the profile. In any case, water would be available during and
immediately following winter rains but would be scarce thereafter.
A
B
Figure 11.1 A: Shallow pans on eastern side of Gibb Rock; B: Fenced off wall to capture runoff from
the outcrop, facing west. Both photographs by A.M. Rossi.
11.3 SURVEY AND EXCAVATION METHODS
Gibb Rock fieldwork was conducted in October 2016. Due to accessibility issues, it
was not possible to search the nearly 4 km perimeter of Gibb Rock. Much of the rock
itself was fenced off, as noted above, and no attempt was made to enter this area.
Instead, survey focussed on the central eastern side of the outcrop, where vehicle
tracks provided easy access (Figure 11.2). The area north of the carpark was
searched, and no potential cultural material was noted, but thick vegetation impeded
access and visibility (Figure 11.2). The southern portion was more easily accessible
as thick vegetation was interspersed with more sparsely vegetated areas. Two such
areas had potential surface artefact concentrations – the first (Scatter 1) where a
pocket of soil and vegetation had developed on the lower portion of the outcrop itself,
and the second (Scatter 2) in a sandy area at the base of the outcrop (Figures 11.2,
11.3a–b).
226
A
B
Figure 11.2 A: Aerial photograph of Gibb Rock. White rectangle indicates area magnified in image B.
Note the extensive wall constructed to filter runoff to the large circular tank, near the car park on the
eastern side of the outcrop. B: Location of Scatter 1 and Scatter 2 (within white boundaries), and pit
GR1 (red arrow). Note the thicker vegetation north of the car park area. Base imagery from Google
Earth.
At Scatter 1, there were several areas where quartz was naturally weathering from
the bedrock (Figure 11.3c). This quartz was quite variable in quality and very friable
on the weathered surface but had a finer, closer texture beneath. If it was quarried,
rather than collected after weathering from the rock, it may have been a source of
usable
material.
The
potential
artefacts
were
irregularly
distributed
and
comparatively sparse, so when a piece was encountered, a GPS reading was taken,
and the piece collected along with any other potential cultural material within a twometre radius; these small collections were subsequently grouped due to low artefact
density. Deposit was probed and proved shallow (approx. 150 mm) so there was no
real potential for excavation at Scatter 1.
227
A
B
C
Figure 11.3 A: Area around Scatter 1, facing south. B: Area around Scatter 2, facing east. C: quartz
weathering from parent granite, in Scatter 1. All photographs by A.M. Rossi.
228
At Scatter 2, no quartz exposures were present, and there was a lower density of
potential artefactual material, all concentrated within approximately 10 m of the base
of the outcrop. Surface material was collected following the same methods for
Scatter 1. Depth probing indicated deposit was at least 300 mm deep, so there was
some potential for excavation. Therefore, a single 500 x 500 mm pit (GR1) was
positioned within Scatter 2 near some of the more convincing potential artefacts.
This pit was excavated following the standard methods noted earlier (see Chapter
9.2.2). The soil was coarse and sandy, making it difficult to maintain a consistent XU
(excavation unit) depth, so XU thickness varied from 20 to 40 mm. As the pit
progressed, medium sized roots began to intrude, and few potential artefacts were
noted in situ or in the sieve. As a result, GR1 was terminated at a depth of 130 mm
below the ground surface, and subsequently backfilled. All sieve residue from GR1,
along with material obtained from surface collections, was returned to Perth for
analysis.
11.4 ORGANICS AND DATING
A total of 1.6 g of charcoal was present in GR1, with no more than 0.4 g in any single
XU. Charcoal weight declined from XU1–4, before increasing again in the basal unit
(Figure 11.4). Due to the limited depth of deposit and the dearth of archaeological
content, no attempt was made to date material from GR1. As a result, no pH
readings were taken, since the main aim of that test was to determine the impact of
soil chemistry on the preservation of organics (see Chapter 9.4.2).
weight (g)
0
0.1
0.2
0.3
0.4
0.5
1
GR1 - XU
2
3
4
5
Figure 11.4 Weight of charcoal with depth in pit GR1.
229
11.5 STONE ARTEFACT ASSEMBLAGE
A total of 88 artefacts were recovered from Gibb Rock, comprising 87 quartz and one
silcrete. None of the quartz was high quality, and most was extremely variable,
coarse-grained, and had visible crystal aggregations (Figure 9.3: top left). Most
artefacts came from surface collections that contributed nearly 70% of artefacts
(Table 11.1). The results of the pit (GR1) and surface collections are discussed
separately below but, in most cases, small sample sizes were a considerable
constraint on attribute analysis. Raw data are available in Appendix F.
Table 11.1 Number of stone artefacts from the Gibb Rock pit and surface collections.
Pit/Collection
Quartz
Silcrete
TOTAL
Scatter 1
53
0
53
Scatter 2
6
1
7
GR1
28
0
28
TOTAL
87
1
88
11.5.1 Surface Collections
Of the 60 surface artefacts found at Gibb Rock, all but seven came from Scatter 1.
However, Scatter 2 yielded the only non-quartz piece found at Gibb Rock (Table
11.1). Neither scatter preserved any retouched flakes, and Scatter 2 also had no
flaked fragments; small flakes and small fragments were rare throughout (Table
11.2).
Table 11.2 Frequency of artefact types found at Scatter 1 and Scatter 2.
Collection
Core
Small
fragment
Large
fragment
Small
flake
Large
flake
Retouch
TOTAL
Scatter 1
2
1
21
3
26
0
53
Scatter 2
4
0
0
1
2
0
7
TOTAL
6
1
21
4
28
0
60
Cores
Six cores were present, comprising five quartz and one silcrete. Despite lower
overall artefact numbers, four of the cores came from Scatter 2, where cores
comprised 57% of the artefact assemblage. Quartz cores were much heavier at
230
Scatter 1, ranging from 62.1 to 143g – at Scatter 2, however, all weighed under 12g,
averaging 9.28g. The smallest Scatter 2 core was the only bipolar specimen. At
Scatter 1, quartz cores averaged three negative flake scars, 43.5 g/scar and two
platforms, while those at Scatter 2 averaged eight scars, 1.48 g/scar and four
platforms. All quartz cores at Scatter 2 had < 25% unmodified surface, while those
from Scatter 1 exhibited much higher values (50–75%, > 75%). The single silcrete
core – from Scatter 2 – weighed 232g and had a single negative flake scar and
platform, with > 75% of its surface unmodified (Appendix F.1). This piece may be a
broken hammerstone but no pitting or battering was evident. However, the extremely
rough texture may mask such features.
Flaked Fragments
All flaked fragments came from Scatter 1 and comprised one small and 21 large
fragments (Table 11.2). The increased sample size in the latter meant that large
fragments represented the best source of reduction intensity data from the surface
collections, but only 14 provided negative flake scar and platform data. On average,
each exhibited two negative flake scars, 1.84 g/scar and 1.5 platforms. One
particular specimen, which had a smooth, fine texture compared to most Gibb Rock
quartz (but still not high quality), exhibited seven scars and three platforms; all others
had no more than three flake scars and one to two platforms (Appendix F.3).
Flakes
Small flakes were scarce on the surface, comprising three from Scatter 1 and a
single example from Scatter 2. In contrast, a total of 28 large flakes were present, all
but two from Scatter 1. None exhibited edge damage and just one broken flake was
recovered, from Scatter 1. Bipolar and freehand reduction were equally common at
Scatter 2, but bipolar flakes outnumbered freehand 2.6:1 at Scatter 1. At Scatter 1,
flakes averaged 2.1 negative flake scars, 8.46 mm/scar and 1.6 platforms, but
individual specimens displayed between zero and five scars and up to three
platforms. At Scatter 2, one flake had no negative flake scars and a single platform,
while the other was a dual platform flake with three scars (Appendix F.5). Unmodified
surface was variable – each scatter yielded pieces in both the lowest (< 25%) and
highest categories (> 75%). All four categories were represented at Scatter 1, but
over 70% fell within the lowest unmodified surface category (< 25% – n=17).
231
11.5.2 GR1
A total of 28 artefacts, all quartz, were present in GR1 – artefactual stone was most
abundant in the uppermost XU, which contributed 50% of the artefacts. There was
an overall decline with depth thereafter, and the lowest two XU contributed just three
artefacts. The extremely small sample size means that data were scarce, as the
most numerous categories (small flakes, small fragments) contribute no usable
attribute data. Few large flakes were found in GR1, and no cores or retouched
pieces were present (Table 11.3).
Table 11.3 Frequency of artefact types by XU in GR1.
0
Small
fragment
4
Large
fragment
0
Small
flake
8
Large
flake
2
0
3
0
1
1
3
0
4
0
2
4
0
1
0
5
0
1
TOTAL
0
13
XU
Core
1
2
Retouch
TOTAL
0
14
0
5
0
0
6
1
0
0
2
0
0
0
0
1
0
12
3
0
28
Small fragments were the most numerous category (n=13), and all but two were
found in the upper three XUs, where most of the cultural material was concentrated
(Table 11.3). Small flakes were also common (n=12) and were most abundant in
XU1 (n=8). Only three large flakes were found, all in XU1–2. This was the only class
of artefacts found in GR1 for which detailed attributes were recorded; sample size is
therefore very problematic and average values must be treated with caution. On
average, flakes had 2.5 flake scars, 4.58 mm/scar and originated from dual-platform
cores. There was no evidence for bipolar reduction, and unmodified surface was
consistently < 25%, but poor-quality raw material meant that each attribute could
only be measured for a single flake (Appendix F5).
11.6 CONCLUSION
Gibb Rock would have provided fairly easy access to three Landscape Divisions,
and the food resources found therein. Granite and the surrounding Thicket are the
dominant zones, but Mallee areas occur less than 1 km from the site. Water would
be preserved in Gibb Rock pans and soils during winter but would be scarce soon
232
after rains ceased. While archaeological survey was impeded by thick vegetation
and general accessibility issues, cultural material was found in two main areas – on
the rock itself (Scatter 1), and at the base of the outcrop (Scatter 2). Excavation in
the latter area provided few artefacts and limited organic content, so no attempt was
made to date the deposit. The Gibb Rock stone artefact assemblage was small, and
dominated by poor-quality quartz, possibly quarried locally from the exposures on
the outcrop. Most artefacts came from Scatter 1, which yielded all artefact types
except retouched pieces that were absent from the site as a whole. Generally,
reduction intensity was variable, but attribute values were most often concentrated
toward the lower end of the scale.
233
CHAPTER 12
MULKA’S CAVE
12.1 INTRODUCTION
This chapter focuses on Mulka's Cave by first considering the natural setting of the
site and its resource base, particularly the longevity of water in the deeper rock
structures. The extensive body of previous research is outlined, from the earliest
sources mentioning the site to later archaeological and rock art studies. Field
methods are then described for the collections made during the 2006 development
work as well as subsequent investigations permitted under Section 16 of the
Aboriginal Heritage Act 1972. The charcoal assemblage is quantified, focussing on
the resultant radiocarbon dates, including those derived from earlier research (Rossi
2010). The stone artefact assemblage is then discussed, with each of the eight
pits/surface collections treated separately. Finally, a brief analysis of ochre and
flaked glass is presented.
12.2 SITE DESCRIPTION
Mulka's Cave is located within the Humps Nature reserve, 18 km north east of
Hyden, within the southern part of the study area. The site itself falls within the
Granite Landscape Division and is surrounded by vast areas of Mallee. Heath is
accessible around 6.5 km to the north west, while Woodland and Saline areas lay
around 15 km south (Figure 9.1). The Humps outcrop reaches 420 m ASL and
covers an area approximately 0.5 km2, but other smaller granite exposures also
occur within the reserve.
Mulka's Cave is a registered site (ID 5842) that comprises surface artefact
concentrations, gnammas, and a constructed lizard habitat ('lizard trap'). However, it
is undoubtedly best known for the decorated boulder that houses 452 pigmented
motifs in a variety of colours (Figure 12.1; Gunn 2006). These cultural and natural
features all occur on the eastern side of The Humps (Figure 12.2).
234
B
A
D
C
E
F
Figure 12.1 Cultural and natural features present at The Humps. All photographs by A.M. Rossi. A:
the main entrance to Mulka's Cave, facing west. Note the raised platforms either side of the entrance,
created when erosion removed the sediments below. B: Two paintings inside the entrance to Mulka's
Cave, note the yellow handstencil inside the circular motif. Lefthand painting measures 127 x 36 cm,
righthand painting 83 cm wide (Gunn 2004:26). C: Raised walkway outside Mulka's Cave, facing
south. D: Red and white pigmented handstencils and handprints inside Mulka's Cave. E: View of
raised walkway from the top of Mulka's Cave. F: Gnamma 1, measuring 1.4 m long and 0.85 m wide.
Note the shadows cast by the gnamma walls, protecting the water from evaporation.
235
A
B
Figure 12.2 A: Aerial photograph of The Humps. White rectangle indicates the area magnified in
image B. B: Location of natural and cultural features around the Humps, including Mulka's Cave itself,
the Humps Scatter, Camping Area Scatter, lizard trap and the main water sources. Features too small
to outline at the scale of this figure are indicated by arrows. Pit locations for Mulka's Cave and the
Camping Area are illustrated in Figure 12.4 and Figure 12.6. Locations for cleft rockhole and
Gnamma 5 are approximations, based on Figure 3 from Gunn (2006:22). Base imagery from Google
Earth.
12.2.1 Water Availability
As at most granite outcrops, several shallow pans are present, and soils at the base
of the rock would store water for several weeks after winter rains ceased (Figure 7.8).
The most significant water sources, however, are the deeper rock structures. Five
gnammas are present, four on flat exposures of granite north of the decorated
boulder and the fifth mid-slope on the western side of the outcrop (Figure 12.2). The
distance from the sloping outcrop means that four gnammas receive little runoff – at
best, from an overlying pan. Some small, flat slabs of granite were found near the
gnammas, and may have been used as lids to limit water loss through evaporation.
One further rock structure exists, a cleft rockhole that has developed on a fault line
on The Humps outcrop (Figure 12.2) – this structure is fed by the considerable runoff
generated by the rock. The rockhole is interesting as its depth exceeds that of most
gnammas, yet the surface area is more than five times greater than the largest
gnamma. It also has a v-shaped profile, making its shape analogous to an inverted
triangular prism (Figure 9.4). In smaller, straight-sided gnammas, water is protected
by shadows cast by the structure’s walls except when the sun is directly overhead;
protection is even greater when the structures are not full to capacity. In contrast, the
236
cleft rockhole's large surface area and gently sloping sides means that protection is
much more limited, so an evaporation coefficient of 0.8 was applied (Table 12.1).
Table 12.1 Surface dimensions, capacity, shape, evaporation coefficient (EC) and modelled runoff
conditions for the six gnammas at Mulka's Cave. Dimensions of five traditional gnammas after Webb
and Rossi (2008), cleft rockhole from Gunn (2004, 2006). Volumetric capacities were calculated using
standard mathematical formulae provided in Figure 9.4.
Surface
dimensions (mm)
Depth
(mm)
EC
Runoff
Shape
Capacity
(L)
Gnamma 1
Radius a = 700
Radius b = 425
1100
0.4
None
elliptic
cylinder
1028.09
Gnamma 2
Radius = 1000
1000
0.4
2x
cylinder
3141.59
Gnamma 3
Radius a = 225
Radius b = 100
300
0.4
None
elliptic
cylinder
21.21
Gnamma 4
Radius = 100
200
0.4
None
cylinder
6.28
Gnamma 5
Radius = 125
350
0.4
5x
cylinder
17.18
Cleft rockhole
Length = 8500
Width = 2000
800
0.8
5x
triangular
prism
6800.00
Structure
TOTAL
11,014.35
The six rock structures have a combined capacity of just over 11,000 L, more than
half of which derives from the cleft rockhole that holds 6800 L alone (Table 12.1).
Regardless of rainfall conditions, capacity was never reached (Figure 12.3), because
only the cleft rockhole and Gnamma 5 ever completely filled. Under high winter
rainfall conditions, water volume peaked at 8000–8500 L, around 75% of total
capacity; it was slightly lower after average winter rainfall at around 7600 L. The cleft
rockhole completely filled under either rainfall condition, albeit at different rates, so
the difference in overall volume reflects the quantities of water stored in the
gnammas. It is important to distinguish between these sources since gnammas could
be capped, to limit evaporation, while the cleft rockhole could not. Gnammas held
1250–1650 L after high winter rainfall, and around 900–950 L when winter rains were
average. In all cases, water balance steadily decreased from October, unless strong
spring falls arrested the decline. By 31 December, 1500–4000 L of water was
available, of which between 165 and 740 L was stored in gnammas. Even after low
winter rainfall, water volume peaked at around 6000–6800 L; the cleft rockhole did
not completely fill, and gnammas always held less than 400 L. All structures took
longer to accumulate water and began to lose it soon after winter. By 31 December,
237
gnammas were almost dry, and 650–1400 L was available in the cleft rockhole
(Figure 12.3).
high winter rainfall
9000.00
1992
1998
8000.00
7000.00
water volume (L)
6000.00
5000.00
4000.00
3000.00
2000.00
1000.00
average winter rainfall
9000.00
1 Dec
1 Nov
1 Oct
1 Sep
1 Aug
1 Jul
1 Jun
1 May
1 Apr
1 Mar
1 Feb
1 Jan
0.00
2009
2011
8000.00
7000.00
water volume (L)
6000.00
5000.00
4000.00
3000.00
2000.00
1000.00
1 Dec
1 Nov
1 Oct
1 Sep
1 Aug
1 Jul
1 Jun
1 May
1 Apr
1 Mar
1 Feb
1 Jan
0.00
low winter rainfall
8000.00
2010
2012
7000.00
6000.00
water volume (L)
5000.00
4000.00
3000.00
2000.00
1000.00
1 Dec
1 Nov
1 Oct
1 Sep
1 Aug
1 Jul
1 Jun
1 May
1 Apr
1 Mar
1 Feb
1 Jan
0.00
Figure 12.3 Total volume of water available at Mulka's Cave from the gnammas and cleft rockhole
(solid line) and the gnammas only (dashed line), during high (top), average (middle) and low (bottom)
winter rainfall years.
238
12.3 PREVIOUS RESEARCH
The earliest published accounts of Mulka’s Cave date to the early 1950s (Davidson
1952; Day 1951; Serventy 1952; Willey 1950), but Davidson conducted his fieldwork
much earlier, in 1938–1939. Willey (1950) and Day (1951) both published short
newspaper articles that briefly describe the decorated boulder but these documents
mainly focussed on their personal experiences at the site. These records are
interesting from a historical perspective but are not relevant to this research.
Serventy (1952) and Davidson (1952) both attempted more serious publications,
mostly concerning the rock art, but both have been superseded by the later, far more
comprehensive rock art recording undertaken by Gunn (2004, 2006). These early
sources are not discussed further as summaries are available elsewhere (Rossi
2010:32–37).
The first archaeological investigation was conducted in 1988, when Bowdler et al.
(1989) excavated a 1 m2 pit inside the dripline of the main entrance to Mulka’s Cave.
A total of 210 stone artefacts were identified, the lowest from XU13a that also
yielded a radiocarbon date of 420 ± 50 BP (lab code not supplied); they believed
XU14–17 (0.7–1.1m below the surface) to be archaeologically sterile. The artefacts
were largely quartz, and comprised 139 fragments, 61 flakes, seven retouched
pieces and three cores. The upper deposits (XU1–10a) were heavily disturbed, with
glass and modern material found throughout. Only 31 artefacts were recovered from
XU11–13a, which they believed to have some stratigraphic integrity (despite the
presence of a filled drainage channel running through these strata), leading Bowdler
et al. (1989:31) to conclude that Mulka’s Cave ‘was occupied, in a brief and
intermittent fashion, by Aboriginal people since perhaps 500 years ago’.
As part of his site management plan (Gunn 2003), Gunn (2004, 2006) conducted a
thorough re-recording of the rock art at Mulka’s Cave. Echoing Davidson’s (1952)
earlier remarks, Gunn also concluded that the rock art assemblage was noteworthy
for the region, with regard to the range of colours and motif types, as well as the
overall number of motifs. However, Gunn’s analysis revealed that the degree of
discordance between this site and those in the surrounding area was even more
marked than Davidson (1952) perceived. Gunn recorded a total of 452 motifs, while
most Noongar sites yield fewer than 100 (Webb and Gunn 2004). The assemblage
239
comprised 275 handstencils, 40 sprayed areas, 23 handprints, 23 paintings, three
drawings and an object stencil, in addition to many uninterpretable fragments. The
designs were executed in red, orange, yellow, cream and white pigment. Despite
their well-protected location, nearly 80% of motifs were in poor or very-poor condition,
leading Gunn to conclude that the visible artwork began to be created between
3000–2000 years ago. Motif condition and superimposition indicated that the art was
not created in a single episode, meaning the site had been repeatedly visited over
the last few thousand years (Gunn 2004, 2006).
While Gunn’s main focus was documenting the artwork, he also conducted an onsite analysis of 130 stone artefacts from the drainage channel outside the main
entrance to Mulka’s Cave. The assemblage comprised 96% quartz, most of which
was debris or amorphous flaked material that could not be further classified (n=110),
with only 16 flakes and four retouched pieces recovered. Gunn (2004, 2006)
concluded that Mulka’s Cave was a recurrent focus for occupation over the last 3000
years and was used for both ritual and general activities. He argued that people
would primarily have utilised the space outside the main entrance to the boulder,
where the surface artefacts were found.
Rossi’s (2010, 2014a) reinvestigation of the site was prompted by the conflicting
interpretations offered by Bowdler et al. (1989) and Gunn (2004, 2006), specifically
regarding the antiquity and intensity of occupation. Most of the earlier findings are
largely irrelevant and superseded by the current research, so will not be discussed
further. There are, however, some aspects that warrant discussion. Through
comparison of photographs, plans and sketches created over several decades, it
was demonstrated that heavy visitation since the 1980s resulted in the loss of 10–15
m3 of sediment from inside Mulka’s Cave, representing one vertical metre at the
cave mouth (Rossi 2010; Rossi and Webb 2007, 2008; Webb and Rossi 2008). This
sediment was redeposited onto the slope outside the main entrance, meaning the
multitude of artefacts found in this area have no spatial or temporal control, and are
of limited analytical value. Furthermore, Rossi (2010, 2014a) reanalysed the material
excavated by Bowdler et al. (1989), including their large samples of 'natural debris'.
While the assemblage data cannot be incorporated in the current analysis (due to
the typological methodology employed at the time), Rossi (2010, 2014a) identified
240
nearly three times as many artefacts (n=604 compared to 210), including several
from XU14–17, which Bowdler et al. (1989) considered archaeologically sterile.
12.4 SURVEY AND EXCAVATION METHODS
All fieldwork at Mulka's Cave was conducted between 2006 and 2008 as part of
Rossi's (2010) MA research. Work took place in two stages, that associated with the
2006 site development, and that permitted between 2007 and 2008 for research
purposes. As per Gunn's (2003, 2004) recommendations, a raised walkway was
installed outside the main entrance to Mulka's Cave in 2006, to prevent further
degradation of the slope. Under Section 18 of the Aboriginal Heritage Act 1972, all
ground disturbance had to be monitored for archaeological content. At this time,
walkway postholes were excavated, and surface collections made in other areas of
the site that would be impacted by development. In 2007, permission to excavate
was sought and subsequently granted under S16 of the AHA, (permit no. 418). The
one square-metre allowance was executed via the excavation of four 500 x 500 mm
pits, allowing investigation in an undisturbed area around the cave mouth, as well as
other areas of interest identified during development work.
Much of the material collected during the 2006 site works were unusable, as erosion
and/or earlier site works has voided spatial and temporal controls. The out of context
material is not incorporated within the current study, but full assemblage details are
available in Rossi (2010). The remaining pits and surface collections focus on three
parts of the site, the Main Cave Area, the Camping Area, and the Humps outcrop
itself. Each of these areas, and the specific methodologies employed are discussed
below.
12.4.1 Main Cave Area
The Main Cave Area was the focus of the 2006 work but, as noted above, most of
the material is decontextualised and of no further use in the current analysis. Only
one area suffered limited impact – the raised platforms outside the cave mouth,
created when erosion scoured out the sediments below. These platforms receive no
vehicular traffic and minimal foot traffic, noteworthy considering Mulka's Cave
receives an estimated 80,000 visitors a year (Gunn 2006). Surface artefacts were
241
collected from these platforms during the 2006 development work. Artefact density
was low, so each was collected as a single unit – these have been amalgamated for
analysis (Entrance Platforms).
The platform areas noted above also extended just inside the cave mouth,
representing valuable interior deposits that had not been lost to erosion. As the 2006
development work progressed, the edges of these platforms were in danger of
collapse, so they were sampled to evaluate the 420 BP date obtained by Bowdler et
al. (1989). The author was not on-site at the time, but R.E. Webb cut a 200 x 200
mm column sample to a depth of -600 mm in 12, 50 mm XUs (Figure 12.4). Webb
sorted the sieve residue on-site, and potential artefacts and dateable charcoal
samples were passed on to the author for analysis – at present, the remaining
charcoal samples cannot be located. Cultural material was comparatively scarce
(and entirely lacking in the basal XU), probably due to the small pit size.
Nevertheless, the assemblage is sufficient considering the main aim was to redate
the interior deposits, while ensuring that archaeological content was present in dated
XUs.
After the Section 16 permit was granted, a final attempt was made to sample
deposits in the Main Cave Area. In 2008, a single 500 x 500 mm pit (MC1) was
excavated on a raised platform south of the cave entrance (Figure 12.4), following
the standard procedures outlined earlier (Chapter 9.2.2). Archaeological content was
limited, and MC1 was terminated after six XUs, at 300 mm below the surface.
242
Figure 12.4 Main Cave Area, facing west, showing the location of the Entrance Platforms (flat areas
indicated by white arrows), Column Sample and pit MC1 (red arrows). Photograph: A.M. Rossi.
12.4.2 The Humps
During the 2006 development work, most of the Humps outcrop was inspected for
signs of Aboriginal usage. Artefacts were noted on one small shoulder of the outcrop
that overlooked gnammas, where sediment and some vegetation (mostly ground
cover with scattered low shrubs) had developed (Figure 12.5a–b). A low-density
surface concentration was identified, spanning a 10 x 5 m area; in 2007, this scatter
was collected as a single unit. Immediately thereafter, the area was probed to
determine depth of deposit, which proved fairly consistent. A 500 x 500 mm pit (TH1)
was therefore located on the flattest area in the centre of the scatter; it was
terminated when bedrock was encountered at -320 mm, at the base of XU6.
243
A
B
C
Figure 12.5 A: View of The Humps outcrop taken from the area around Gnammas 1 and 2, facing
south west. Note the sparsely vegetated shoulder on which the Humps Scatter is located. B: The
Humps Scatter, showing pit TH1, facing north east. Gnammas 1 and 2 occur on the flat granite
exposure visible in the top right of the image. C: A portion of the Camping Area, facing west. Note the
relatively clear area surrounded by thicker border vegetation – no artefacts were present in the latter
area.
12.4.3 Camping Area
The Camping area was identified in 2006 when members of the Mulka's Cave
Aboriginal Steering Committee pointed out that as children they camped in a belt of
trees between the gnammas and the decorated boulder, when brought to the site by
their grandparents. The area was searched and a surface artefact scatter was
located in a clearing bordered by larger shrubs (Figure 12.5c) – the more heavily
vegetated area was also searched but yielded no artefacts. The Camping Area was
not to be affected by the development works, so no further investigation occurred
until the S16 permit was issued in 2007. At that time, the scatter was mapped, and
244
found to span an area measuring approximately 130 m long and 30–35 m wide.
Within this scatter, six higher density concentrations were noted, each covering
approximately 20 square metres (Figure 12.6: Clusters A–F). These clusters were
collected individually, while the surrounding low-density material was collected as a
single unit. Due to the extent of the scatter, two 500 x 500 mm pits were excavated
within. CA1 was positioned within Cluster B and terminated at 450 mm below the
surface (after XU9) at which point the sediment became cemented. CA2 was
positioned at the centre of the scatter, between Cluster C and D. It was terminated
after six XUs, at 300 mm below the surface, when saprolite was encountered.
Figure 12.6 Camping Area artefact scatter, showing the extent of surface artefacts (light grey
shading), the location of the high-density surface clusters (Clusters A–F, dashed circles) and pits CA1
and CA2. Note the scatter is surrounded by thick border vegetation (dark grey shading).
12.5 ORGANICS AND DATING
Bowdler et al. (1989) recovered a considerable quantity of charcoal (over 1.2 kg)
during their excavation inside Mulka's Cave. While the Column Sample total could
not be quantified due to the loss of undated samples noted above, at least 5 g of
charcoal was present, a not insubstantial quantity considering the small size of the
pit. In contrast, none was present in MC1. A total of 3.75 g was found in TH1 and
was fairly evenly distributed throughout the deposit, as four of the six XUs yielded
0.7–0.85g; the basal XU preserved just 0.2g (Figure 12.7). CA1 yielded 8.95 g of
charcoal, mostly concentrated in XU1–2 (7.2 g). Charcoal was especially scarce
towards the base of the pit – none was present in XU7, and XU8–9 preserved less
than 0.05 g each (Figure 12.7). A similar amount was present in CA2, 8.8 g, despite
245
the pit being 150 mm shallower than CA1. Again, the bulk was from the upper
deposits, with XU2 and XU3 yielding 2.8 and 3.75 g respectively. Charcoal quantity
then declined to the base of the pit (Figure 12.7).
charcoal weight (g)
0
1
2
3
4
5
1
2
CA1 - XU
3
4
5
6
7
8
charcoal weight (g)
0
0.5
1
1.5
2
2.5
3
3.5
4
1
CA2 - XU
2
3
4
5
6
charcoal weight (g)
0
0.2
0.4
0.6
0.8
1
1
TH1 - XU
2
3
4
5
6
Figure 12.7 Charcoal weight by XU within pits CA1 (top), CA2 (middle) and TH1 (bottom).
Soil samples were unavailable for further analysis – these were passed to R.E.
Webb during the author's MA research and could not be located thereafter. However,
basic tests conducted at the time indicated that pH was generally < 5 (Rossi
2010:97). As noted elsewhere (see Chapter 9.4.2), such acid conditions are
246
conducive to charcoal preservation, so the lack of charcoal in MC1 cannot be
explained in terms of soil pH.
A total of 16 radiocarbon dates were obtained from Mulka's Cave, comprising 12
from the author's own excavations and four from the pit excavated by Bowdler et al.
(1989); six were sourced as part of the author’s MA research (Rossi 2010). Dates
were obtained from the Main Cave Area, Camping Area and The Humps.
The dates from inside Mulka's Cave are complicated by the fact that they came from
two separate pits in areas where the deposit was eroding at different rates. The top
of each pit was first considered in reference to the 1952 ground level, as this
represented the earliest visual record of the cave and the closest analogue to a
depositional (vs. erosive) surface. This permitted Rossi (2014a) to correlate
excavation depths across both pits allowing direct comparison of the dates from
each (Figure 12.8; Table 12.2). All dates are in sequence except for Wk-35800 which
is indistinguishable from modern carbon; it probably represents accidental inclusion
of more recent material when the Column Sample was cut (see Rossi 2014a). The
uppermost date (Wk-35798) lies within the disturbed deposits of the 1988 pit (XU1–
10a, see 12.3 above), so may be equally affected. The author had planned to avoid
these deposits but had performed an incorrect correlation between the Column
Sample and 1988 pit at the time of sample submission (Rossi 2014a). The dates
from XU15 and XU16 of the 1988 pit, both of which pre-date 8000 cal. BP, overlap at
one standard deviation (at 68.2% probability) and are statistically indistinguishable.
When the in-sequence dates are considered, the depth-age curve indicates that
cultural material began to accumulate around 9650 cal. BP (Figure 12.9). The
earliest material derives from the 1988 pit, while that from the Column Sample postdates 6000 cal. BP.
247
Figure 12.8 Stratigraphic section (Bowdler et al. 1989: Figure 4.2a) superimposed with their XU levels (blue), sediment column (orange) and the approximate
position of the 2006 Column Sample (pink). XUs numbered by each author. Rossi (2014a) after Bowdler et al. (1989).
248
Table 12.2 Radiocarbon dates from inside Mulka's Cave. Asterisk denotes AMS date. CS = Column
Sample (cut in 2006), P88 = Bowdler et al. (1989) pit, excavated in 1988. Dates have been arranged
by depth below the 2006 ground surface using Figure 12.8. Note that Bowdler et al. (1989) excavated
by stratigraphic unit, so the thickness of each unit varied in different parts of the pit. The maximum
depths have been cited below. All CS samples came from sieve residue, while the location of P88
samples are unknown.
XU
Pit
Depth
(mm)
Sample no.
Sample
wt (g)
C14 age
BP
Calibrated age BP
Reference
CS
3
100–150 Wk-35798*
~2
200 ± 25
281–145 (68.2%)
290–0 (95.4%)
Rossi 2014a
P88
12
125–300 Wk-29186
not
available
1061 ± 31
961–914 (68.2%)
970–805 (95.4%)
Rossi 2014a
CS
7
300–350 Wk-35799*
~2
2118 ± 25
2082–2009 (68.2%) Rossi 2014a
2148–1940 (95.4%)
P88
14
350–475 Wk-29187
not
available
4285 ± 36
4858–4711 (68.2%) Rossi 2014a
4870–4626 (95.4%)
P88
15
475–575 Wk-27113*
~2
7386 ± 30
8194–8050 (68.2%) Rossi 2010,
8299–8026 (95.4%) 2014a
CS
11
500–550 Wk-35800*
~1
Modern
P88
16
575–675 Wk-29188
not
available
Modern
7284 ± 41
8156–8002 (68.2%) Rossi 2014a
8166–7973 (95.4%)
cal. BP
cal. BP
0
Rossi 2014a
0
1000 2000 3000 4000 5000 6000 7000 8000 9000 10000 11000
1000
2000
3000
4000
5000
6000
7000
8000
0
0
CA1
MC
CA2
50
100
TH1
100
200
150
300
500
depth (mm)
depth (mm)
200
400
250
300
600
350
700
400
800
450
900
500
Figure 12.9 Depth-age curves for inside Mulka’s Cave (left) and other parts of the site (right). Solid lines join
radiocarbon age determinations, while linear trendlines (dotted lines) were used to estimate the age of the
lowest cultural material, represented by the end point of those lines. Inside Mulka’s Cave, Wk-27113 (P88
XU15) and Wk-29188 (P88 XU16) were statistically indistinguishable. The lower sample was selected as it was
derived from a larger charcoal sample.
249
Two dates were sourced from the Humps pit (TH1), one basal (XU6) and the other
from XU2. The basal sample returned an age of 1236 ± 20 BP, while XU2 was more
recent, 454 ± 20 BP (Table 12.3). A total of seven dates were obtained for the
Camping Area – four from CA1 and three from CA2. In CA1, all but the lowest date
(XU8, 92 ± 30 BP) are in sequence. That date was on a single, small fragment of
charcoal that may have been introduced into the pit during excavation. The lowest
reliable sample, from XU6, dates to 1626 ± 21 BP (Table 12.3). The CA2 dates are
all in sequence, ranging from 1340 ± 94 BP for XU2, to 5093 ± 30 BP for XU5.
Depth-age curves indicate that artefacts began to accumulate in CA2 around 6700
cal. BP and from 2500 cal. BP in CA1. The vast differences in sedimentation rates is
of note, as both pits are separated by < 60 m of flat, featureless terrain. Sediment in
TH1 accumulated at a similar rate to CA1 (Figure 12.9).
Table 12.3 Radiocarbon dates for the Humps and Camping Area pits. Asterisk denotes AMS date. All
samples were collected from sieve residue.
Pit
TH1
CA1
CA2
XU
Depth
(mm)
Sample no.
2
70–120
6
2
Sample
wt (g)
C14 age
BP
Wk-38220*
0.7
454 ± 20
504–472 (68.2%)
510–343 (95.4%)
300–320 Wk-38221*
0.2
1236 ± 20
1172–1066 (68.2%)
1180–1059 (95.4%)
3.9
326 ± 69
453–290 (68.2%)
500–147 (95.4%)
Rossi 2010
Rossi 2010
40–90
Wk-24311
Calibrated age BP
5
200–250 Wk-24864*
0.3
1541 ± 30
1411–1349 (68.2%)
1470–1308 (95.4%)
6
250–300 Wk-38218*
0.25
1626 ± 21
1520–1432 (68.2%)
1533–1416 (95.4%)
8
350–400 Wk-24312*
< 0.05
92 ± 30
2
50–100
2.8
4
160–220 Wk-38219*
5
220–270 Wk-24310*
Wk-24309
Reference
239–24 (68.2%)
254–0 (95.4%)
Rossi 2010
1340 ± 94
1300–1093 (68.2%)
1364–981 (95.4%)
Rossi 2010
1.05
4308 ± 21
4855–4830 (68.2%)
4873–4711 (95.4%)
0.8
5093 ± 30
5891–5743 (68.2%)
5908 –5665 (95.4%)
Rossi 2010
250
12.6 STONE ARTEFACT ASSEMBLAGE
A total of 969 stone artefacts were recovered from Mulka's Cave, 702 of which
derived from the Camping Area. In contrast, the Main Cave Area yielded just 90
artefacts. The assemblage was heavily quartz dominated, comprising over 96% by
artefact number (Table 12.4). Most quartz was of fairly good quality, with fewer
examples of the extremely poor material that was found elsewhere. Non-quartz
materials consisted of silcrete (n=21), BIF (n=8) and chalcedony (n=3), all of which
were confined to the Camping Area. The results of each pit and surface collection
are briefly discussed below, arranged by site area and beginning with surface
material. Due to the number of collections/pits (and the generally short depth of
deposit), only selected data are tabulated or graphed – raw data are available in
Appendix G.
Table 12.4 Number of stone artefacts from Mulka’s Cave pits and surface collections. Assemblages
have been grouped by site area.
Main
Cave
Humps
Camping
Area
Quartz
Silcrete
BIF
Chalcedony
TOTAL
Entrance Platforms
8
0
0
0
8
Column Sample
63
0
0
0
63
MC1
19
0
0
0
19
Humps Scatter
21
0
0
0
21
TH1
156
0
0
0
156
CA Scatter
174
9
0
0
183
CA1
304
9
5
3
321
CA2
192
3
3
0
198
937
21
8
3
969
TOTAL
12.6.1 Entrance Platforms
Just eight stone artefacts were recovered from the Entrance Platforms, comprising
five large flakes, two large fragments and a core, all on lower quality quartz. The
core was small (1.75 g), bipolar, and exhibited four scars detached from three
platforms; < 25% of its surface was unmodified (Appendix G.1). The large fragments
varied in weight from 0.9 to 4.25 g, but each exhibited a single scar and platform
(Appendix G.3). Each large flake had 2–5 scars (avg. 3.4), 3.15–7.86 mm/scar (avg.
5.41) and 2–3 platforms (avg. 2.2). All had < 25% of their surface unmodified, and
none were broken, but two were damaged (Appendix G.5)
251
12.6.2 Column Sample
The 2006 Column Sample yielded 63 stone artefacts, all quartz, including 41 on
high-quality materials. Small flakes dominated the assemblage (n=36), while cores
and retouched pieces were rare; all cores, small fragments and retouched pieces
were made on high-quality quartz. The limited pit size and small assemblage means
that vertical trends should be treated with caution, so discard rates were not
calculated. Overall, the number of stone artefacts increased with depth until the peak
in XU9 (n=16) and declined thereafter; no artefacts were found in XU1, XU3 or XU12
(Table 12.5). While small sample size is problematic, the assemblage can be
characterised in general terms.
Table 12.5 Stone artefact types by XU from the Column Sample inside Mulka’s Cave. Figures in
parentheses refer to high-quality quartz.
0
0
0
Small
fragment
0
0
0
Large
fragment
0
0
0
Small
flake
0
1 (1)
0
Large
flake
0
0
0
0
0
0
0
(1)
0
0
0
(1)
0
(1)
0
0
0
(2)
0
(2)
(5)
0
1 (1)
1
1
0
(1)
0
0
3 (2)
3
2
4 (2)
(6)
(4)
(5)
(7)
(1)
10 (26)
0
(1)
1
3 (1)
1 (1)
4 (3)
0
0
9 (6)
XU
Core
1
2
3
4
5
6
7
8
9
10
11
TOTAL
Retouch
TOTAL
0
0
0
0
1 (1)
0
0
0
0
0
0
(1)
0
0
(1)
3
3 (3)
6 (2)
4 (7)
1 (6)
4 (12)
(7)
(3)
22 (41)
Just one core was found in the Column Sample, in XU8 (Table 12.5). It was made on
high-quality material, a smooth-textured black opaque quartz. It weighed 3.25 g and
exhibited six flake scars detached from four platforms. Less than 25% of the surface
was unmodified, and a combination of feather and axial terminations indicates that
both freehand and bipolar reduction techniques were used (Appendix G.1).
A total of five small flaked fragments were present, all made on high-quality quartz.
Four were found in XU9 and XU11, and the remaining small fragment was recovered
from XU5. Large fragments were also comparatively rare; five were found in XU5–9,
including two on high-quality quartz (Table 12.5). The high-quality material was more
252
intensively reduced, and averaged 4 flake scars, 0.18 g/scar and 2 platforms; the
lower quality material averaged 2 scars, 0.5 g/scar and 1.5 platforms (Appendix G.3).
Small flakes were the most common artefact type in the Column Sample, comprising
more than 55% of the artefact assemblage (n=36). They were especially
concentrated between XU6 and XU10, where 28 were found (Table 12.5). Lower
quality materials were only present in XU2–6, while high-quality quartz was found
throughout. None of the small flakes from XU2–6 were broken, and < 15% from
XU10 were. Elsewhere, 60–100% of small flakes had been broken (Appendix G.4).
A total of 15 large flakes were present, all between XU5 and XU9; numbers peaked
in XU7 (n=4) and XU9 (n=7 – Table 12.5). Six flakes were made on high quality
quartz, and averaged 2.3 negative flake scars, 5.05 mm/scar and 1.6 platforms.
Individual values were highest in XU5–7, and one flake in XU5 exhibited several
small partial flake scars, indicating it came from a highly reduced core. Only one
flake was broken, and another had suffered heavy edge damage – both were found
in XU9. The remaining nine flakes were made on lower quality material, and
averaged 2.1 flake scars, 5.01 mm/scar and 1.8 platforms (Appendix G.5).
Only one retouched piece – a backed artefact – was present, in XU9 (Table 12.5).
The piece was made on high-quality quartz, was unbroken and exhibited no
significant edge damage (Appendix G.6).
12.6.3 MC1
Just 19 stone artefacts were recovered from MC1, all quartz, including five on highquality quartz; all artefact types were represented except small fragments. XU1–2
were archaeologically sterile, and artefact numbers peaked in XU4 (n=7) and
declined to the base of the pit (Table 12.6). The very small sample size means that
the MC1 assemblage can only be characterised in very basic terms.
Just one core was found, in XU4, made on a smoky clear quartz with a smooth,
glassy surface (Table 12.6). This piece is noteworthy as it preserved four original
crystal facets, representing the tip of a large individual crystal. The opposing surface
was covered in negative flake scars and retained no trace of a ventral surface. This
piece therefore represents the use of a large crystal as a core, or possibly a flake
253
that had been subsequently modified to the extent that the entire ventral surface was
removed – had any trace of this surface remained, the piece would have been
classified as retouched (see Chapter 9.3.2). The core was extremely small (0.65 g)
so was presumably worked via the bipolar method. Three negative flake scars were
preserved, all initiated from a single platform (Figure 12.10; Appendix G.1).
Table 12.6 Stone artefact types by XU in MC1. Figures in parentheses refer to high-quality quartz.
XU
Core
Small
fragment
Large
fragment
Small
flake
Large
flake
Retouch
TOTAL
1
2
3
4
5
6
0
0
0
(1)
0
0
0
0
0
0
0
0
0
0
1
1
0
0
0
0
0
0
1
1
0
0
1 (1)
2 (1)
3
2
0
0
1 (1)
1 (1)
0
0
0
0
3 (2)
4 (3)
4
3
TOTAL
(1)
0
2
2
8 (2)
2 (2)
14 (5)
Figure 12.10 Small core from MC1 XU4. Note the presence of four original crystal facets (indicated
by blue arrows).
Two large fragments were present in MC1, one each in XU3 and XU4; both were on
lower quality quartz (Table 12.6). The XU3 fragment was made on a greyish quartz
of moderate texture and exhibited three scars detached from three platforms. In
contrast, the XU4 specimen was made on extremely poor-quality quartz that
comprised an aggregation of small crystals that gave it a very irregular texture. It
weighed 8 g and exhibited just one negative flake scar (Appendix G.3).
254
Only two small flakes were found, in XU5–6. Both were made on lower quality quartz
and had been broken (Appendix G.4). In contrast, large flakes were the most
numerous artefact category, and comprised 10 of the 19 artefacts present; at least
two large flakes were found in each XU (Table 12.6). Only two flakes were made on
high-quality quartz, from XU3–4. They averaged three negative flake scars, 5.12
mm/scar and two platforms; one exhibited several overlapping small scars, indicating
it derived from a highly reduced parent core. Neither flake was broken, but one had
suffered severe edge damage. Eight flakes were made on lower quality material –
reduction intensity was lower, averaging 2.1 flake scars, 12.28 mm/scar and two
platforms. Only one of these flakes was broken, from XU6 (Appendix G.5).
Only XU3 and XU4 both yielded one retouched piece made on high-quality quartz
and another on lower quality material (Table 12.6). All were backed except one
unifacially retouched piece from XU4. None of the retouched pieces were broken or
exhibited major edge damage (Appendix G.6).
12.6.4 The Humps Scatter
The Humps Scatter yielded 21 stone artefacts, all quartz, dominated by high-quality
materials (n=14). The assemblage comprised nine large flakes, four small flakes,
three cores, two large fragments, two retouched pieces and a small fragment. Two
cores were made on high-quality quartz, while the remainder was fine-textured, but
exhibited some flaws and fracture planes. All cores were small (2.2–3.05 g), bipolar,
and preserved 6–7 flake scars detached from 2–3 platforms (Appendix G.1). None
had more than 25% of their surface unmodified, and one high-quality quartz core
preserved three original crystal facets, the largest used as the main platform.
A single small fragment was found, made on high-quality quartz (Appendix G.2).
Large fragments were almost as rare, with just two present, including one on lower
quality material. The better quality piece was more intensively reduced (7 scars, 0.29
g/scar, 3 platforms) than the lower quality piece, but the latter still displayed five
scars (0.43 g/scar) detached from two platforms (Appendix G.3).
All four small flakes were unbroken; three were made on high-quality quartz while
the fourth exhibited a blockier structure. Of the nine large flakes, only three were
made on lower quality raw material. These averaged 4.7 flake scars, 4.96 mm/scar
255
and 2.3 platforms; unmodified surface was always < 25%. In contrast, the six highquality materials were less intensively reduced, averaging 2.8 scars, 6.74 mm/scar
and 2.2 platforms; one flake had 25–50% of its surface unmodified. Three highquality flakes were broken, and none exhibited severe edge damage (Appendix G.5).
Two retouched pieces were present, both exhibiting unifacial retouch in the form of
flake removals on their ventral surfaces. One was made on glassy smooth quartz,
while the other was a more granular variety. Neither was broken, and only the lower
quality piece exhibited significant edge damage (Appendix G.6).
12.6.5 TH1
A total of 156 artefacts were excavated from TH1, all quartz, including just 24 on
high-quality quartz (Table 12.7). Discard rates ranged from 7.6–21 artefacts/100
years, and values exceeded 12 artefacts/100 years in five of the six XU. The highest
values occurred in XU2 and XU4, and the lowest in XU5 (Figure 12.11).
Table 12.7 Stone artefact types by XU in TH1. Figures in parentheses refer to high-quality quartz.
0
0
0
0
0
Small
fragment
4 (4)
8 (7)
10
13
6
Large
fragment
3
3
4
6
3
Small
flake
14
14
2 (4)
6 (3)
6
Large
flake
3 (1)
3
2
1
3
0
0
3
44 (11)
0
19
2 (1)
44 (8)
0
12 (1)
XU
Core
1
2
3
4
5
6
TOTAL
Retouch
TOTAL
3
2
(3)
3
2 (1)
27 (5)
30 (7)
18 (7)
29 (3)
20 (1)
3
13 (4)
8 (1)
132 (24)
no. artefacts/100 yrs
0.00
5.00
10.00
15.00
20.00
25.00
1
2
TH1 XU
3
4
5
6
Figure 12.11 Artefact discard rates for pit TH1.
256
Small fragments were the most common artefact type in TH1, with a total of 55
present. They were most numerous between XU2 and XU4, where 38 were found;
just three were present in the basal XU, and high-quality materials were only found in
XU1–2 (Table 12.7). A total of 19 large fragments were found in TH1, all on lower
quality quartz. They were most common in XU4 (n=6), and none were present in
XU6 (Table 12.7). Most fragments had 1–3 negative flake scars, and 1–2 platforms,
while a single specimen from XU3 had five scars and three platforms. Reduction
intensity peaked in XU1 (avg. 2.3 scars, 0.18 g/scar, 1.7 platforms) and showed a
general decline with depth. Flake scar and platform averages were lowest in XU5
(avg. 1 scar, 1 platform), but those fragments were among the lightest (avg. 0.22
g/scar – Appendix G.3).
Small flakes were the second most numerous artefact type in TH1, with a total of 52
recovered. XU1 and XU2 yielded 14 each, while all other XUs had three to nine
specimens. High-quality materials were only found in XU3–4 and XU6 (Table 12.7).
Large flakes were rare – only 13 were found, seven of which occurred in the upper
two XU, including the single large flake made on high-quality material; no large
flakes were present in XU6 (Table 12.7). The high-quality quartz flake was broken
and displayed three flake scars (3.68 mm/scar) detached from a single platform. The
flake scars were long, narrow and parallel, and the flake itself was fairly long and thin,
so it may have derived from a more standardised core. The lower quality quartz
flakes averaged 2.4 scars, 6.56 mm/scar and two platforms. Three had > 25% of
their surface unmodified, including two from XU3. Reduction intensity was highest in
XU1, where the single bipolar flake also occurred, and lowest in XU3 (Appendix G.5).
A total of 17 retouched pieces were found in TH1, with two to three from each XU
(Table 12.7). All were backed except two unifacially retouched pieces, from XU1 and
XU4. The four specimens made on high-quality materials were restricted to XU3
(n=3) and XU5 (n=1). All of the high-quality quartz pieces were backed and only one,
from XU3, was unbroken. The XU5 piece preserved traces of probable hafting resin,
indicating it formed part of a composite tool. Nine of the 13 lower quality pieces were
broken, including all pieces from XU5–6 (Appendix G.6).
257
12.6.6 Camping Area Scatter
A total of 183 artefacts were recovered from the Camping Area Scatter, all quartz
(including 50 on high-quality quartz) except for nine silcrete specimens. Artefacts
were most numerous in the surrounding low-density material (n=70, including three
silcrete), while the densest cluster, Cluster B, yielded 38. Clusters E and F were the
sparsest, with just seven and six artefacts respectively. High-quality materials were
proportionately most common in Clusters B and C (Table 12.8).
Table 12.8 Stone artefact types in the Camping Area Scatter. Figures in parentheses refer to highquality materials. Note that ‘other CA’ refers to the low-density material found outside the clusters, but
within the Camping Area Scatter.
Area
Core
Small
fragment
Large
fragment
Small
flake
Large
flake
Retouch
Manuport
TOTAL
Cluster A
Cluster B
Cluster C
Cluster D
Cluster E
Cluster F
Other CA
TOTAL
0
4 (1)
0
1 (1)
0
1
1
7 (2)
0
0
(2)
0
0
0
1
1 (2)
3
2 (1)
3
2
1
1
11 (2)
23 (3)
3 (1)
(3)
(2)
1 (1)
0
(1)
3 (1)
7 (9)
10 (1)
19 (7)
10 (3)
11 (1)
3 (1)
3
36 (13)
92 (26)
(4)
(1)
0
1
1 (1)
0
(2)
2 (8)
1
0
0
0
0
0
0
1
17 (6)
25 (13)
13 (7)
16 (3)
5 (2)
5 (1)
52 (18)
133 (50)
Cores
Nine cores were present in the Camping Area Scatter, comprising seven quartz (two
high quality) and two silcrete. Cores were especially common within Cluster B, which
yielded five cores, including both silcrete specimens. No cores were present in
Clusters A or E (Table 12.8). The two high-quality quartz cores were found in
Clusters B and D; both weighed < 3g, were bipolar, and had been heavily reduced.
They displayed ten flake scars each (0.24–0.27 g/scar), detached from 4–6 platforms.
Reduction intensity was more varied among the lower quality materials. The two
silcrete cores, both from Cluster B, were the largest, and weighed 14.15–60.2 g. The
heavier was a single platform core with four negative flake scars and > 50–75% of its
surface unmodified; the smaller displayed ten scars detached from four platforms
and had < 25% unmodified surface. The latter was made on a finer grained silcrete,
so the increased reduction intensity is not surprising. Lower quality quartz cores
preserved 3–11 scars, 0.14–5.52 g/scar and 1–5 platforms; the lowest values were
258
all found in Cluster F, where the single core had > 75% of its surface unmodified.
Three cores were bipolar, one each from Clusters B, D and F. The latter weighed
16.55 g so the flaking method was probably not related to core size, but may have
been a response to the poor quality, blocky material (Appendix G.1).
Flaked Fragments
Only three small fragments were found, comprising two high-quality quartz pieces
from Cluster C and one lower quality quartz piece from the surrounding area (Table
12.8). The 26 large fragments (all quartz) were primarily concentrated in the
surrounding area, which yielded 13; all other Clusters had no more than three each.
Only three large fragments were made on high-quality quartz, including two from the
surrounding area and one from Cluster B. These fragments preserved 2–4 scars,
0.08–0.67 g/scar and two platforms. Reduction intensity was more varied among the
lower quality quartz specimens. These exhibited 1–9 scars, 0.2–6.75 g/scar and 1–4
platforms. Reduction intensity values were greatest in Cluster E (6 scars, 3 platforms)
and lowest in Cluster F (1 scar, 1 platform, 6.75 g/scar); Clusters B–D yielded values
toward the higher end of the scale (Appendix G.3). One specimen from Cluster C
preserved long, narrow, parallel flake scars, indicating it derived from a more
standardised core.
Flakes
A total of 16 small flakes were found in the Camping Area Scatter, all quartz, with no
more than four in any single collection; none were present in Cluster E. Small flakes
on high-quality materials (n=9) outnumbered those on lower quality material (n=7),
and Clusters B, C and F yielded only high-quality materials (Table 12.8). Both small
flakes from Cluster C were broken, as well as three of four from Cluster A; unbroken
flakes dominated elsewhere (Appendix G.4).
Large flakes were the most common artefact type in the Camping Area Scatter, and
a total of 118 were recovered, including six on silcrete. The largest assemblage
(n=49) was present in the surrounding material, while large flakes were
comparatively rare in Clusters E and F, where few artefacts were found. Only 26
large flakes were made on high-quality material – these were particularly common in
Cluster B and the surrounding material (Table 12.8). These flakes exhibited 1–7
259
scars, 1–4 platforms and 1.68–11.66 mm/scar. All reduction intensity measures
peaked in Cluster A, while Clusters D and E had the lowest values across one or
more attributes (Figure 12.12a, c). Variation in individual values was greatest in
Cluster B and the surrounding material, where sample sizes were largest. One flake
from the surrounding material exhibited the same notching noted in the CA2 sample
(see Chapter 12.6.8). Broken flakes were rare, and restricted to Cluster B (< 15%)
and the surrounding material (< 17% – Figure 12.12d); just a single flake exhibited
significant edge damage, from Cluster B (Appendix G.5).
number
0
1
2
3
4
number
5
6
7
8
Cl. A
0
1
2
3
4
5
6
Cl. A
No. NFS
No. NFS
No. platforms
No. platforms
Cl. B
Cl. B
Cl. C
Cl. C
Cl. D
Cl. D
Cl. E
Cl. E
Cl. F
Cl. F
Other CA
Other CA
mm/scar
0
Cl. A
7
1
2
3
4
% broken
5
6
7
high-quality
8
other
0
Cl. A
Cl. B
Cl. B
Cl. C
Cl. C
Cl. D
Cl. D
Cl. E
Cl. E
Cl. F
Cl. F
Other CA
Other CA
10
20
30
40
50
60
70
80
high-quality
90
100
other
Figure 12.12 Selected attributes recorded on large quartz flakes within the Camping Area Scatter. Cl.
= Cluster. A–B (top): Average negative flake scar (blue bars) and platform counts (orange bars) for
high-quality materials (A – left) and lower quality materials (B – right). C–D: average mm/scar values
(C – left) and proportion of broken flakes (D – right) for high-quality materials (black bars) and lower
quality materials (grey bars).
260
Large flakes on lower quality materials displayed 1–9 scars, 2.32–12.72 mm/scar
and 1–4 platforms. However, the collection was dominated by dual platform flakes
(n=42), and those with 2–5 scars (n=80). Reduction intensity measures peaked in
Cluster E and were lowest in Cluster F, but the small sample sizes must be
considered; elsewhere, variation in average values was fairly subdued (Figure
12.12b–c). Clusters B–C and the surrounding material exhibited the greatest range
in individual reduction intensity measures, but values were fairly diverse throughout.
Broken flakes were more common than among the high-quality materials – while
Clusters D–F yielded only whole flakes, the remainder had 11–25% of theirs broken
(Figure 12.12d). Nineteen flakes exhibited significant edge damage; these were most
common in Cluster B and the surrounding material, which each yielded five
(Appendix G.5).
Retouched Flakes
A total of ten retouched flakes were found in the Camping Area Scatter. All were
quartz, and all but one was backed. High-quality materials dominated, and just a
single unifacially retouched piece and a backed artefact were made on lower quality
materials. Retouched pieces were most common in Cluster A (n=4), while none were
found in Clusters C or F (Table 12.8). The unifacially retouched piece was a large
flake (53.65 g) on fairly rough-textured quartz. Several flakes had been removed
from the ventral surface, in areas where flake morphology provided a good platform
– this probably reflects the use of a flake as a core. None of the retouched pieces
were broken, across higher or lower quality material. Three of the higher quality
pieces exhibited more significant edge damage, and were restricted to Clusters A–B
and the surrounding material (Appendix G.6).
Manuports
A single manuport was found in Cluster A, comprising a large unmodified chunk of
silcrete weighing 25.4 g. It was rounded but showed no evidence of battering but
may have been collected for future use as a core, hammerstone or grindstone.
261
12.6.7 CA1
A total of 321 artefacts were recovered from CA1, comprising 304 quartz (including
115 of high quality), nine silcrete, five BIF and three chalcedony (Tables 12.4 and
12.9). Non-quartz artefacts were concentrated in XU2–5 (n=15), and high-quality
materials contributed around 45 to > 50% of assemblages from XU2–4 (Appendix G).
Artefact discard rates were highly varied in CA1, ranging from < 2 to > 30
artefacts/100 years. Rates were fairly consistent from XU2–5, around 17–19
artefacts/100 years, and peaked in XU6. Rates then declined with depth until XU9,
which registered the lowest value (Figure 12.13).
Table 12.9 Stone artefact types by XU in CA1. Figures in parentheses refer to high-quality materials.
XU
Core
Small
fragment
Large
fragment
Small
flake
Large
flake
Retouch
TOTAL
1
0
4
0
3
(2)
(2)
7 (4)
2
3
0
2
4 (1)
7 (3)
3
0
10 (9)
16 (17)
2 (5)
7 (14)
0
1 (1)
19 (15)
33(35)
4
5
6
7
8
9
TOTAL
3
0
0
0
0
0
5
0
7
8
8 (3)
2 (2)
1
41 (9)
3
6
1
8
2 (1)
0
23 (1)
9 (20)
22 (7)
(14)
4 (7)
1 (2)
5
70 (76)
17 (7)
12 (2)
6
10
2
0
56 (30)
2 (3)
(1)
0
0
0
0
3 (7)
34 (30)
47 (10)
15 (14)
30 (10)
7 (5)
6
198 (123)
no. artefacts/100 yrs
0.00
5.00
10.00
15.00
20.00
25.00
30.00
35.00
40.00
1
2
3
CA1 XU
4
5
6
7
8
9
Figure 12.13 Artefact discard rates for pit CA1.
262
Cores
Five cores (all quartz) were present in CA1, all between XU3 and XU4 (Table 12.9).
All were quite fine-textured, but exhibited natural fracture planes and flaws, so were
not classified as high-quality quartz. Two larger cores were present (16.15 and 82.15
g), one each in XU3 and XU4, and three smaller (< 4.5 g). The larger cores had
higher flake scar (n=9–12) and platform counts (n=4–5), but reduction intensity was
lower when the size of nodules was considered (n=1.35–9.13 g/scar). In contrast,
the smaller cores were all bipolar, and averaged five scars, three platforms, and 0.5
g/scar. Two cores (the largest and one smaller), both from XU4, had 25–50% of their
surface unmodified (Appendix G.1).
Flaked Fragments
A total of 50 small quartz fragments were present in CA1, over half of which were
found between XU5 and XU7 (n=26), while XU4 yielded none. High-quality materials
were only present in XU2–3 and XU7–8 (Table 12.9, Appendix G.2). CA1 also
yielded 24 large fragments, all quartz. These artefacts were especially common in
XU5 (n=6) and XU7 (n=8), while other XUs preserved no more than three. XU8
yielded the only large fragment on high-quality material (Table 12.9) – it weighed just
0.25 g and displayed five scars (0.05 g/scar). The remaining large fragments
preserved 1–4 scars (avg. 2.1), 1–3 platforms (avg. 1.7) and 0.05–2 g/scar (avg.
0.61). Reduction intensity was highest in XU1, but XU7 preserved the greatest range
of flake scar and platform counts (Appendix G.3).
Flakes
Small flakes were the most common artefact type in CA1, with a total of 146 present.
They were particularly common in XU3–5, but numbers then declined towards the
base of the pit (Table 12.9). Small flakes were one of the few categories where
pieces made on high-quality materials (n=76, including 5 non-quartz) outnumbered
those of lower quality. High-quality materials were particularly dominant in XU3–4,
and XU6. Several examples from XU3 came from highly reduced parent cores,
based on the frequency of small, overlapping flake scars. Overall, < 20% of small
flakes were broken. Separate breakage data were not available (i.e. by raw material
quality), but rates were often higher in those XU where high-quality materials were
263
more common, which may indicate those were more frequently broken (Figure
12.14). The breakage rate was highest in XU6 (n=35.71%), where all small flakes
were made on high-quality material (Appendix G.4).
% high-quality
0.00
20.00
40.00
60.00
% broken
80.00
100.00
0.00
0
1
1
2
2
3
3
4
4
40.00
60.00
80.00
100.00
XU
CA1 XU
0
20.00
5
5
6
6
7
7
8
8
9
9
Figure 12.14 Proportion of small flakes on high-quality material (left), and breakage rate for all small
flakes (right) from pit CA1.
Large flakes were the second most common artefact type in CA1, with a total of 86
present, including 30 on high-quality materials (28 quartz, 2 BIF). Nearly 70% of
large flakes (n=56) were found in XU3–5, and no high-quality materials were present
below XU5 (Table 12.9). High-quality flakes displayed 1–7 flake scars (avg. 2.9),
2.39–11.52 mm/scar (avg. 5.08) and 1–4 platforms (avg. 1.9). Reduction intensity
values peaked in XU4 and XU1 and were lowest in XU5 (Figure 12.15a, c). XU3
displayed the greatest range of individual values, as well as fairly high averages, and
preserved the only piece with > 25% of its surface unmodified. No flakes were
broken in XU1 or XU4, where reduction intensity values were highest, while half of
those from XU5 were (Figure 12.15d). Only three flakes had suffered significant
edge damage, one each from XU2, XU3 and XU5 (Appendix G.5). Two flakes from
264
XU3 preserved original crystal facets, while one from XU4 displayed long, narrow,
parallel flake scars.
Flakes on lower quality materials preserved 1–7 scars (avg. 3.2), 1.69–15.62
mm/scar (avg. 5.27) and 1–3 platforms (avg. 2), indicating a similar reduction
intensity to high-quality materials, at a pit-wide level. Average values peaked in
XU7–8, and XU1, but variation between individual XUs was much more subdued
than in the high-quality assemblage (Figure 12.15a–c). Two flakes had 25–50% of
their surface unmodified (one each from XU4 and XU6), while values were higher in
XU5–6 (50–75%). Breakage rates were highest in XU2–3 (33–50%), while XU8
yielded no broken flakes (Figure 12.15d). Six had severe edge damage, including
two each from XU4 and XU7 (Appendix G.5).
number
0
0.5
1
1.5
number
2
2.5
3
3.5
0
4
No. NFS
1
0.5
1
1.5
2
2.5
3
3.5
1
No. platforms
2
2
3
3
4
4
CA1 XU
CA1 XU
No. platforms
5
5
6
6
7
7
8
8
9
9
mm/scar
0
1
1
2
3
% broken
4
5
6
7
0
high-quality
1
2
2
3
3
4
4
CA1 XU
CA1 XU
10
20
30
40
50
60
70
80
90
100
high-quality
other
other
5
4
No. NFS
5
6
6
7
7
8
8
9
9
Figure 12.15 Selected attributes recorded on large quartz flakes, by XU, in CA1. A–B (top): Average
negative flake scar (blue bars) and platform counts (orange bars) for high-quality materials (A – left)
and lower quality materials (B – right). C–D: average mm/scar values (C – left) and proportion of
broken flakes (D – right) for high-quality materials (black bars) and lower quality materials (grey bars).
265
In XU2, six of the seven large flakes (two on lower, four on higher quality material)
and several of the smaller flakes preserved a crust on their dorsal surfaces. This
crust overlay some negative flake scars but had been removed in other areas when
subsequent flakes were detached (Figure 12.16). No flakes displayed any crusting
on their ventral surfaces, so it was probably formed when flakes were still attached to
their parent cores. Scar distribution indicates there was some length of time between
flaking episodes, separated by the crust-forming event.
Figure 12.16 Dorsal surface of three quartz flakes from CA1 XU2. Note the brown crust that covers
most of the dorsal surface (including negative flake scars) and has been removed in other areas; this
crust could not be removed by washing or rubbing. Crust was absent on ventral surfaces. White bar =
1 cm.
Retouched Flakes
A total of 10 retouched flakes were found in CA1, comprising nine quartz and one
chalcedony. Retouched pieces were most common in XU4 (n=5), and none occurred
below XU6. The three artefacts made on lower quality material were restricted to
XU3–4 (Table 12.9); all were backed and unbroken, but one exhibited significant
edge damage. The seven high-quality pieces comprised five backed and two
unifacially retouched pieces. The latter occurred in XU1, where backed artefacts
were not found. Both unifacial pieces terminated in a point, and had flakes removed
from their ventral surfaces. Each also preserved a blunt margin of partial negative
flake scars, created by prior flake removals when the piece was attached to its
parent core – these pieces may have been used in the same way as intentionally
backed pieces. One also preserved patches of a thick, resinous material near the
266
blunted margin, indicating these pieces had probably been hafted. Neither had been
broken or heavily damaged, but two backed artefacts from XU4–5 had suffered
transverse snaps (Appendix G.6).
12.6.8 CA2
A total of 198 stone artefacts were recovered from CA2, comprising 192 quartz,
three silcrete and three BIF, across all artefact types; 83 pieces were made on highquality materials. High-quality materials were proportionally more common in XU1–2
but were present throughout the pit. They outnumbered lower quality material in just
two categories: small flakes and retouched pieces (Table 12.10). Despite the fairly
large assemblage, artefact discard rates were low, due to the extremely slow
sedimentation rate. Values peaked at < 6 artefacts/100 years in XU2 and dropped as
low as 0.99 artefacts/100 years in XU6 (Figure 12.17).
Table 12.10 Stone artefact types by XU in CA2. Figures in parentheses refer to high-quality materials.
XU
Core
Small
fragment
Large
fragment
Small
flake
Large
flake
Retouch
Manuport
TOTAL
1
2
3
4
5
6
TOTAL
0
0
0
1
0
0
1
5
6
7
10 (3)
10 (2)
1 (2)
39 (7)
(2)
1 (1)
1
2 (1)
3
5
12 (4)
2 (5)
3 (15)
11 (18)
20 (5)
8 (3)
(1)
44 (47)
2 (1)
(1)
4 (4)
5 (7)
4
0
15 (13)
(1)
1 (7)
1 (1)
1 (2)
(1)
0
3 (12)
0
0
0
1
0
0
1
9 (9)
11 (24)
24 (23)
40 (18)
25 (6)
6 (3)
115 (83)
no. artefacts/100 yrs
0.00
1.00
2.00
3.00
4.00
5.00
6.00
7.00
1
2
CA2 XU
3
4
5
6
Figure 12.17 Artefact discard rate in pit CA2.
267
Cores
Just one core was present in CA2, recovered from XU4, where artefacts were the
most numerous (Table 12.10). The core was made on fine-textured quartz, but
natural fracture planes were evident. It weighed 23.15 g and exhibited 16 negative
flake scars (1.45 g/scar) initiated from six different platforms. Less than 25% of the
surface was unmodified, and the core was not bipolar (Appendix G.1).
Flaked Fragments
A total of 46 small fragments were found in CA2, all quartz. These artefacts were
most numerous between XU4 and XU5, which together yielded over half (n=25) of
the assemblage. Only seven were made on high-quality materials – these were
restricted to XU4–6 (Table 12.10). All 16 large fragments were quartz, and were
especially common in XU4–6, where eleven were found. Just four were made on
high-quality materials, which were particularly frequent in XU1–2 (Table 12.10). On
average, the high-quality fragments preserved 1.8 flake scars, 0.33 g/scar and 1.8
platforms. Most values were similar for lower quality materials, but flake scar
averages were higher (avg. 2.4 scars, 0.34 g/scar, 1.8 platforms – Appendix G.3).
Temporal trends could not be analysed as fragments were clustered in XU5–6, and
small sample size plagued the upper deposits.
Flakes
Small flakes were the most abundant stone artefact type in CA2, comprising 88
quartz, two BIF and one silcrete; over half (n=47) were made on high-quality
materials. Small flakes were especially common in XU3–4, where nearly 60% of the
assemblage was concentrated, but most high-quality pieces came from XU2–3
(Table 12.10). Three of the high-quality flakes from XU3–4 exhibited long, narrow,
parallel flake scars, indicating they had been struck from highly reduced,
standardised parent cores. In XU1–5, breakage rates generally increased in concert
with the proportion of high-quality materials, implying that these pieces were more
often broken; XU6 was hampered by small sample size (Table 12.10, Figure 12.18).
268
% high-quality
20.00
40.00
60.00
% broken
80.00
100.00
0.00
0
1
1
2
2
3
CA2 XU
CA2 XU
0.00
0
20.00
40.00
60.00
80.00
100.00
3
4
4
5
5
6
6
Figure 12.18 Proportion of small flakes on high-quality material (left), and breakage rate for all small
flakes (right) from pit CA2.
Of the 28 large flakes found in CA2, all were quartz except for one silcrete and one
BIF flake from XU1–2; higher quality materials were almost as common as lower
quality pieces. Large flakes were concentrated in XU3–4, which yielded a total of 20,
while all other XUs preserved no more than four each (Table 12.10). Among the
high-quality materials, flakes displayed 2–5 flake scars, 2.16–7.6 mm/scar and 1–3
platforms. All three values indicated that reduction intensity increased with depth,
until its peak in XU4, but the limited samples in XU1–2 must be considered (Figure
12.19a, c). None of the flakes in XU1–2 were broken, but 50–67% of those from
XU3–4 were (Figure 12.19d).
The lower quality materials exhibited a greater range of reduction intensity values,
and displayed 1–6 scars, 2.84–12.34 mm/scar and 1–3 platforms. These values
peaked in different XUs – flake scar and platform averages were highest in XU3 and
XU5, respectively, while XU1 produced the lowest mm/scar value (Figure 12.19b–c).
Between 50 and 67% of flakes were bipolar in these same XU. XU3 also yielded the
greatest range of individual flake scar and platform values (Appendix G.5). No lower
269
quality flakes were broken, but two had suffered significant edge damage (Figure
12.19d).
number
0
0.5
1
1.5
number
2
2.5
3
0
3.5
0.5
1
1.5
2
2.5
3
4
No. NFS
No. NFS
1
3.5
1
No. platforms
3
3
CA2 XU
2
CA2 XU
2
No. platforms
4
4
5
5
6
6
mm/scar
0
1
2
3
4
% broken
5
6
7
0
8
10
20
30
40
50
60
70
80
90
100
high-quality
1
1
other
high-quality
other
2
3
3
CA2 XU
CA2 XU
2
4
4
5
5
6
6
Figure 12.19 Selected attributes recorded on large quartz flakes, by XU, in CA2. A–B (top): Average
negative flake scar (blue bars) and platform counts (orange bars) for high-quality materials (A – left)
and lower quality materials (B – right). C–D: average mm/scar values (C – left) and proportion of
broken flakes (D – right) for high-quality materials (black bars) and lower quality materials (grey bars).
Several of the high-quality quartz flakes were noteworthy for various reasons. Two
specimens from XU1 and XU3 exhibit original crystal facets, indicating the use of a
large crystal as a core. Four specimens from XU4 exhibited interesting flake scar
patterns, in the form of long, narrow, parallel scars, and/or a series of uniformly sized
and distributed partial scars that formed notches along a lateral margin. The latter
may be produced from more formal cores with a series of narrow, parallel flake scars.
If such a core is rotated and new flakes are struck – removing the old platform or
base of the core – these flakes will retain a scalloped edge where the new flake
270
intersects with the existing scars (Figure 12.20). Interestingly, while the flake scar
patterns indicated the production of long, narrow quartz flakes, only two were found,
in XU3–4. One exhibited a length/width ratio of 4.84, and the other 2.7 despite
having suffered a transverse snap (Appendix G.5).
Figure 12.20 A (left): quartz flake from CA2 XU4. Lateral margins preserve notches formed by
previous flake removals. Black bar = 1 cm. B (right): an idealised core that could produce notched
flakes. Note the long, parallel flake scars, and the scalloped edge evident in cross section. If the
platform or base of the core was removed, flakes would exhibit similar scalloping.
Retouched Flakes
A total of 15 retouched flakes were present, all but three on high-quality quartz. All
XUs yielded no more than three retouched artefacts, except XU2 which preserved
eight; none were present in XU6 (Table 12.10). Nine of the retouched pieces were
backed, while the remainder exhibited unifacial retouch. The lower quality materials
were exclusively unifacially retouched – two had scalloped retouch along one or
more margins, while the other was a larger flake (4.1 g) that had subsequently been
used as a core. None were broken or exhibited severe edge damage. The highquality assemblage comprised nine backed pieces and three that had been
unifacially retouched; one of the latter preserved traces of possible hafting resin.
Only three of the high-quality quartz pieces were broken, all backed, from XU2 (n=1)
and XU4 (n=2). No unbroken pieces had suffered significant edge damage
(Appendix G.6).
271
Manuports
A single silcrete manuport was found in CA2, in XU4. The piece weighed over 13 g
and had a very weathered, rounded surface. It may preserve weathered negative
flake scars, but the coarseness of the material made this difficult to confirm. The
piece lacked the battering associated with hammerstones, but again the poor texture
may mask these features.
12.7 OTHER CULTURAL MATERIAL
Some ochre and flaked glass was also found at Mulka's Cave. Ochre was relatively
scarce, totalling 0.6 g, and was only present in CA1 (XU7) and CA2 (XU3–4). Three
colours were present: red, brownish-red, and brownish-yellow. The latter was most
abundant, comprising 0.4 g, while red was the scarcest (0.05g – Table 12.11). The
sample size was too small for any further analysis.
Table 12.11 Location, weight and colour of ochre samples from Mulka's Cave.
Pit
XU
XU depth (mm)
Wt (g)
Colour
CA2
3
100–160
0.15
Brownish-red
CA2
4
160–220
0.05
Red
CA1
7
300–350
0.4
Brownish-yellow
Unworked glass was not recovered in any of the excavations discussed herein, so is
not considered further. Just one example of flaked glass occurred, on the Entrance
Platforms outside Mulka's Cave. Several fragments of unworked glass were found in
the same area, but the colour of these pieces (clear and amber) indicates they
derived from different source vessels to the worked piece. The flaked piece was pale
green and had flake scars along one margin, effectively blunting it, while the
opposing margin remained sharp (Figure 12.21). Such a dichotomous distribution is
unlikely to result from post depositional processes, and indeed the Entrance
Platforms receive minimal foot traffic (see Chapter 12.4.1). There were no
distinguishing features on the glass, so it can be dated only in the broadest sense,
as post European arrival in Western Australia. The piece weighed 1 g and measured
17 mm long and 13.59 mm wide. Three large (≥ 10 mm) negative flake scars were
present, each initiated from a different platform.
272
Figure 12.21 Flaked glass from Entrance Platforms at Mulka's Cave. Note the small negative flake
scars concentrated along the right distal margin, while the other remains sharp and intact. Stippling
represents unmodified surface, black bar = 1 cm.
12.8 CONCLUSION
Water was fairly abundant at Mulka’s Cave, regardless of winter rainfall quantity, due
to the properties of the cleft rockhole. It filled later and held water for shorter periods
as winter rainfall decreased, but post-winter longevity was strongly influenced by
spring falls. In contrast to the other sites investigated herein, Mulka's Cave has a
history of research beginning in the 1930s. Much of the earlier work concentrated on
the rock art while later, more detailed studies (e.g. Bowdler et al. 1989; Gunn 2004,
2006) produced conflicting results. The author focussed on three main parts of the
site: The Main Cave Area, the Humps Scatter, and the Camping Area. The former
was problematic as many of the collections are unusable because erosion of the
rockshelter sediments has redistributed archaeological material. Inside Mulka's Cave,
occupation commenced around 9650 cal. BP, while the Camping Area deposits are
slightly more recent – CA2 dated to 6700 cal. BP and CA1 to 2500 cal. BP. Artefacts
began accumulating on the Humps outcrop after 1200 cal. BP. The different dates
for CA1 and CA2 illustrate the problems with applying dates from one pit to another,
even when they are located in the same part of a site.
Over 950 stone artefacts were recovered – most came from the Camping Area, while
cultural material was scarce in the Main Cave Area. The assemblage was heavily
quartz dominated, with a small quantity of BIF, silcrete and chalcedony restricted to
273
the Camping Area, which also yielded the only ochre fragments. Artefacts on highquality material outnumbered those on lower quality material in only one area, the
Column Sample, but high-quality materials were also fairly common in the Camping
Area. Most subsurface stone artefact assemblages were dominated by small flakes,
but small fragments were also fairly common. A number of pieces from around the
site indicate the use of large quartz crystals as cores, and several other pieces,
generally on high-quality quartz, came from highly reduced or standardised parent
cores. A single flaked glass piece found on the Entrance Platforms indicates that
Mulka's Cave was visited by Aboriginal people after Europeans arrived in Western
Australia.
274
CHAPTER 13
IDENTIFYING TECHNOLOGICAL PROVISIONING SYSTEMS
13.1 INTRODUCTION
The aim of this chapter is to identify the technological provisioning systems used at
Gibb Rock, Anderson Rocks, and Mulka's Cave by considering the data presented in
Chapters 10–12 against the archaeological signatures formulated in Chapter 9.5.3.
Raw material quality and conservation, as well as the intensity and length of
reduction sequences, are used to identify the dominant provisioning systems
(Individual and Place Provisioning), as well as supplementary systems (Gearing Up
and Short-term Expedient). Gibb Rock is evaluated first, as it provides a clear
signature that forms a useful baseline against which to evaluate assemblages from
other sites. The Anderson Rocks and Mulka's Cave assemblages are then
considered. In both cases, larger sample sizes, varied pit/surface collection locations
and the availability of radiocarbon dates allows spatio-temporal variation in
provisioning systems to be evaluated.
13.2 GIBB ROCK
Given the small artefact assemblages present at Gibb Rock, one of the best sources
of data was the quality of quartz. As noted earlier (see Chapter 9.5.3), Place
Provisioning strategies require collecting the best material available within the local
foraging range (referred to hereafter as 'stockpile-quality material', i.e. moderate–
good quality), and stockpiling it at a site where future occupation is predictable. In
contrast, Individual Provisioning and Gearing Up require higher quality materials to
prolong artefact use-lives and, especially under Individual Provisioning, to minimise
waste by ensuring predictable flaking properties. The quartz used at Gibb Rock is, at
best, variable, and at worst, incredibly poor. The outer, weathered surface is coarsetextured, friable and plagued by numerous natural fracture planes and flaws (Figure
9.3: top left). The internal material is closer textured, but still heavily flawed. This
quartz would be wholly unsuitable for the manufacture of a mobile toolkit, and it is
unlikely that such material would be collected and transported to the site as part of a
Place Provisioning strategy, when the assemblages from nearby sites indicate that
275
much better quality quartz was readily available in the wider area. Instead, the
assemblage seems to reflect the use of local raw material, as all artefacts are
visually consistent with the material from on-site exposures; the use of local material
occurs under Short-term Expedient strategies (see Chapter 2.4.1).
Where the relevant data were available, they primarily indicated the lower reduction
intensity that is characteristic of Short-term Expedient strategies. At Scatter 1,
negative flake scar and platform averages were low across all artefact types (2-3
scars, 1.5–2 platforms). Individual values were especially low when considered in
conjunction with artefact size – cores averaged 43.51 g/scar, large fragments 1.84
g/scar, and large flakes 8.46 mm/scar. In addition, all cores and over 13% of large
flakes had more than 50% of their surface unmodified (Appendix F). Bipolar
reduction was more than twice as common as freehand reduction, but rather than
indicating increased reduction intensity or raw material conservation, the prevalence
was probably linked to the poor quality of the local quartz; bipolar reduction is often
helpful when flaking poor-quality quartz. Furthermore, pieces detached along natural
fracture planes (which were abundant in the local quartz), can mimic the axial
termination characteristic of quartz bipolar flakes. This is especially problematic
where crushing cannot be observed due to poor texture.
Only eight artefacts from GR1 and Scatter 2 provided reduction intensity data: large
flakes and cores. Negative flake scar and platform averages from large flakes were
similar to Scatter 1, but the flakes there were considerably larger. The quartz cores
had been more intensively reduced, averaging eight flake scars and four platforms,
on cores weighing < 12 g. These cores were made on better quality material than
most at Gibb Rock, but still not predictable or fine-textured enough to form part of a
transported toolkit. Instead, it seems visually consistent with the local quartz located
below the weathering front. The increased reduction intensity may therefore indicate
removal of the outer, weathered quartz to access the more closely textured material
within. The cores were certainly flaked on-site then, but the scarcity of large flakes
may indicate that the reduction event and core discard occurred in different parts of
the site. Overall, despite some more intensive reduction than might be expected
under a Short-term Expedient strategy, there is insufficient evidence of any other
technological provisioning system.
276
13.3 ANDERSON ROCKS
At Anderson Rocks, some assemblages were hampered by sample size issues, but
technological systems were generally identifiable throughout, particularly when the
Short-term Expedient signature from Gibb Rock was used as a basis for comparison.
The surface material is discussed first, then the two test-pits: AR1 and AR2-3. A
temporal timescale can only be defined for AR2-3; no dates were obtained for AR1
and there is no reason to assume that sedimentation proceeded at the same rate.
13.3.1 Surface Collections
Three of the four surface collections (all but Scatter 2) provided fairly clear evidence
of Place Provisioning. These assemblages were made on stockpile-quality material,
comprising moderate- to good-quality quartz (generally with some flaws and
variability in texture), as well as some silcrete. Negative flake scar and platform
averages were higher than at Gibb Rock (avg. 3.1–7 scars, 1.8–3.5 platforms,
combined across relevant artefacts from all three scatters), and artefacts were
generally smaller. Individual values were often quite varied but more evenly
distributed, rather than clustered toward the lower end of the range. Cores were
almost exclusively small and heavily reduced, and three of the four were bipolar –
one likely as a response to blocky raw material texture, and the remaining two due to
their small size. In some cases, flake scars on cores outnumbered flakes found
within the same surface collection. These cores clearly had a longer use-life and had
been flaked in various parts of the site. The available evidence points to the longer
on-site reduction sequences that characterise Place Provisioning (Figure 9.5).
Scatters 1 and 3 also preserved evidence of more formal assemblages made on
high-quality materials, primarily quartz and a small amount of chalcedony. Reduction
intensity was similar to the Place Provisioning assemblage, across large fragments
(avg. 2.5 scars, 2 platforms, 0.15 g/scar) and large flakes (avg 2.6 scars, 2.2
platforms, 6.12 mm/scar). No artefacts retained parallel flake scars, or any other
indicators of more standardised reduction techniques. Finally, three retouched
pieces were present, none of which were broken or heavily damaged. The moderate
reduction intensity combined with the lack of raw material conservation and
standardised reduction is more consistent with a Gearing Up strategy, used within
the Place Provisioning system.
277
Scatter 2 preserved a different signature. As elsewhere, the assemblage was made
on two contrasting materials but, in this case, these comprise a very poor-quality
local quartz (the scatter was located within a pan that contains an exposed quartz
vein) and a few items of stockpile-quality (i.e. moderate–good) quartz. The local
material had been lightly reduced; pieces averaged 1.7 scars and 1–2 platforms.
These values are similar those from Gibb Rock, but the Anderson Rocks artefacts
were generally smaller, probably due to the limited width of the quartz vein. Raw
material quarrying, as well as artefact manufacture, use and discard all occurred
within a metre or so of the quartz source – this is indicative of a Short-term
Expedient strategy, where artefacts were manufactured to meet an immediate need,
and discarded thereafter.
The remaining three artefacts at Scatter 2 were made on finer textured quartz that
exhibited some variability and flaws. These artefacts comprised a single core, large
fragment and large flake. All had been more intensively reduced than the local quartz
(avg. 3–7 scars, 3 platforms), indicating they had previously been flaked elsewhere
on-site, under a Place Provisioning strategy. As Scatter 2 was located within a pan, it
is possible that these artefacts washed downslope via the interconnected channel
system described earlier (see Chapter 10.2.1). If they were intentionally discarded
here, they must post-date the local quartz assemblage, as they could otherwise have
provided a source of better quality material for expedient artefact manufacture.
13.3.2 AR1
In AR1, XU1–9 preserved a different technological signature to the lower deposits,
XU10–13. In the upper deposits, stockpile-quality material was most abundant,
predominantly quartz with a small amount of silcrete. This assemblage was
dominated by small fragments (n=43), which contribute no useful information on
reduction intensity, but their prevalence does indicate a considerable amount of onsite reduction. Just six large fragments and five large flakes contributed useful data
on reduction intensity. Negative flake scar and platform values were very similar to
those at Gibb Rock (avg. 2–2.2 scars, 1.6–1.7 platforms), but the average
weight/length per scar was higher in AR1 (avg. 3.76 g/scar, 9.49 mm/scar),
indicating a lower reduction intensity. However, the range of individual values was
similar, and the small sample size at AR1 makes the average values somewhat
278
misleading as they are easily skewed by outliers. Four retouched pieces (all backed)
were present, on stockpile-quality material – only one was broken. These may have
been required for specific on-site tasks, or been part of a Gearing Up assemblage,
since shorter, predictable artefact use-life might not always require the use of highquality materials. Retouched pieces would not be manufactured under a Short-term
Expedient strategy, since similar artefacts would be available within the transported
toolkit. Therefore, these deposits probably represent Place Provisioning, with lower
reduction intensity indicating that the area was used in a different, less intensive
manner to other parts of Anderson Rocks. There may, of course, have been some
mixing of contrasting signatures within the same archaeological deposits.
The remainder of the XU1–9 assemblage was made on high-quality material and
exhibited characteristics of both Individual Provisioning and Gearing Up. Flake scar
and platform counts were similar to the Place Provisioning assemblage (as may be
expected under Gearing Up), but artefacts were slightly smaller. There was no
evidence of standardised reduction techniques that may occur under Individual
Provisioning, but partial scars on one flake indicate that it came from a highly
reduced parent core. Few artefacts were discarded, and the single retouched piece
was broken, while both large flakes were not. The inconsistent reduction intensity
and discard behaviour may indicate a mixture of both provisioning systems, or simply
be an artefact of small sample size.
The lower deposits (XU10–13) yielded much fewer artefacts, so sample size was
more problematic. Nevertheless, in contrast to the upper deposits, the assemblage
contained more even quantities of lower and higher quality materials. The lower
quality material comprised eight small and three large fragments. The latter had
higher flake scar averages than at Gibb Rock (2.5 scars, 0.17 g/scar), but showed
less core rotation. The small sample makes it impossible to discern between a Shortterm Expedient strategy (using stockpiled materials) or Place Provisioning – the
scarcity of artefacts may support the former, but sampling bias must also be
considered. The high-quality quartz comprised five small flakes, two retouched
pieces and one large flake; all were broken except two small flakes. The lack of
artefacts and the dominance of broken pieces may be indicative of the raw material
conservation that occurs under Individual Provisioning, since intentional discard (i.e.
279
not including waste products created by retouch) should be limited to broken or
unusable items that cannot be repaired.
13.3.3 AR2-3
AR2-3 displayed several shifts in provisioning systems over time. Due to the depth of
the pit, it has been divided into three sections for discussion: the upper (XU1–11),
middle (XU12–21) and lower deposits (XU22–32).
Upper deposits (XU1–11)
XU1–3 (138–0 cal. BP) yielded a fairly even mix of stockpile- and higher quality
material. The latter comprised small fragments and small flakes made on high-quality
quartz, along with a single small BIF flake. This assemblage seems more consistent
with reworking existing artefacts rather than manufacturing new ones, so is more
indicative of Individual Provisioning. The stockpile-quality material comprised a few
small flakes as well as a single large flake, small fragment and large fragment. Data
on reduction intensity were scarce, so this assemblage could not be reliably
compared to the archetypal Short-term Expedient assemblage from Gibb Rock.
However, the scarcity of artefacts may indicate a Short-term Expedient strategy
using stockpiled material.
In contrast, XU4–8 (1220–138 cal. BP) preserved evidence of more intensive
reduction of stockpiled material. Negative flake scar (3–4) and platform averages (2–
2.2) were higher across cores (n=1), large fragments (n=2) and large flakes (n=14),
all of which were generally smaller than their Gibb Rock counterparts. Large flakes
also exhibited a greater range of individual values, with up to seven scars and four
platforms evident. These features are more characteristic of Place Provisioning,
where longer reduction sequences occur.
XU5–8 (1220–354 cal. BP) also yielded some artefacts on higher quality materials,
comprising quartz and some BIF. While high-quality quartz may be stockpiled and
used in a Gearing Up strategy, high-quality exotics (such as BIF) are more likely to
be part of a toolkit where artefacts need to be preserved indefinitely. The BIF
artefacts comprised a small whole flake (just under 10 mm in percussion length – too
large to have been produced by retouch) as well as two large flakes. These must
280
have been struck from a transported core, which would invariably be part of a
transported toolkit. All large flakes were relatively long and narrow, and scars of the
same morphology were found on one large flake and the single core – this indicates
the use of a more formal reduction technique. Two of the four large flakes were
broken (and another was possibly broken), and the single core was very small (2.45
g), bipolar, and probably discarded because no further flakes could be removed from
it. This use of exotic materials, standardised reduction techniques and selective
discard are all characteristic of Individual Provisioning.
XU9–11 (1870–1220 cal. BP) was dominated by high-quality material. Only four
artefacts were made on stockpile-quality material, and only one yielded data on
reduction intensity, so that sample was too small to evaluate. The higher quality
material was primarily quartz, with one chalcedony piece also found. Artefacts
comprised a core, three small fragments, 25 small flakes (mostly broken), one large
flake, and two retouched pieces (both unbroken). Primary reduction clearly occurred
on-site, as many of the broken small flakes would have measured > 10 mm when
whole. Some level of secondary modification may be indicated by the small
fragments, and those small flakes fine enough to have been detached by retouch. A
single small flake exhibited long, narrow, parallel flake scars, but there were no other
traces of standardised reduction. Data on reduction intensity were scarce, but the
single core was very small (1.85 g), bipolar and preserved eight scars initiated from
two platforms. The discard of unbroken retouched pieces is more consistent with
Gearing Up, but Individual Provisioning is supported by the broken chalcedony flake,
heavily exhausted core, and small flake with parallel scars. However, the latter two
artefacts could be generated by Gearing Up, if there was some use of more formal,
intensive reduction techniques to conserve high-quality material. It seems more likely,
then, that the assemblage represents Gearing Up, but the small sample size makes
the signature somewhat ambiguous.
Middle deposits (XU12–21)
In contrast to XU9–11, XU12–18 (3603–1870 cal. BP) exhibited clear evidence of
Place Provisioning. There was abundant evidence of primary reduction, comprising
35 large flakes, seven cores, seven retouched pieces and five large fragments. Many
of the small fragments (n=26) and small flakes (n=61) probably originated from
281
primary reduction, but some may have been produced by secondary modification. In
addition to greater artefact numbers, reduction intensity was consistently high,
averaging 2.7–5.9 scars, 2–3.1 platforms, 0.25–0.94 g/scar and 6.62 mm/scar.
Individual values were also more varied than would be expected under a Short-term
Expedient strategy. All cores weighed less than 11 g, and while the smaller ones had
been flaked to exhaustion, larger ones retained a good amount of unmodified
surface (25–50%) and preserved usable material. Bipolar reduction was common
among cores (80%) and large flakes (45%), which may indicate some conservation
of raw material at a core level. Finally, the seven retouched pieces may have been
manufactured for a logistical foray (under Gearing Up), or they may simply have
been required for specialist tasks.
Higher quality materials were found throughout XU12–18, in varying quantities, but
XU12–17 preserved a different signature to XU18. XU12–17 (3170–1870 cal. BP)
yielded two large BIF flakes, three small BIF flakes and two small chalcedony flakes;
all small flakes were whole and weighed < 0.05 g, so probably derived from retouch
(Appendix E.4). Clearly, then, secondary modification of exotic materials occurred
on-site, along with some degree of primary reduction. These activities are consistent
with Individual Provisioning, if the presence of exotic materials can be considered a
diagnostic feature of that strategy.
The high-quality quartz assemblage from XU12–17 was much larger, and the
number and range of artefacts indicated that a fair amount of primary reduction
occurred. Reduction intensity was similar to the Place Provisioning assemblage (avg.
2.3–4 scars, 2–3 platforms), but the range of individual values was more restricted.
Four flakes were quite long and narrow, and two had scars of the same shape,
indicating some standardised reduction of high-quality quartz. These specimens had
been more intensively reduced than the remainder of the high-quality quartz flakes
(avg. 2.8 scars vs 1.8 scars), but the latter could rarely be oriented, so lacked
platform data. The breakage rate of high-quality quartz flakes was very high – seven
of nine were undoubtedly broken, compared to 39% of stockpile-quality quartz flakes.
The latter averaged 5.2 mm thick for unbroken pieces and 2.4 mm for broken
specimens, while the entire high-quality quartz assemblage averaged 2 mm thick. It
is possible, then, that the higher breakage rate does not indicate selective discard,
282
but simply that a greater proportion of the assemblage was more susceptible to postdepositional breakage. The high-quality quartz assemblage clearly shows some
evidence of Gearing Up, based on frequency of primary reduction, the lower
reduction intensity, and the discard of unbroken retouched pieces. The more highly
reduced standardised pieces could be part of the same strategy, or may support the
Individual Provisioning strategy suggested by the exotic raw materials.
In contrast, the entire high-quality assemblage from XU18 (3603–3170 cal. BP)
appears to derive from Individual Provisioning. Reduction intensity data were
somewhat limited, but values for large quartz flakes were higher (avg. 3.3 scars, 2.7
platforms) than even the more intensively reduced components from XU12–17. One
large flake was long and narrow, and retained five long, narrow scars on its dorsal
surface, evidence of standardised reduction techniques. One small flake retained an
original crystal facet; large crystals would be the perfect choice for transported cores,
since the material therein is generally homogenous and predictable, and the flat
surfaces provide ideal platforms. Finally, there was evidence of preferential discard
behaviour that was largely restricted to broken or non-functional items. Forty per cent
of the large flakes were broken, and 83% of the retouched pieces. If these were
broken post-deposition, the more numerous artefacts (flakes) would be expected to
show a higher rate of breakage than the scarcer artefacts (retouched pieces), due to
the greater likelihood of the former being encountered on the ground surface.
Furthermore, the broken high-quality quartz flakes actually had a slightly greater
average thickness (3.3 mm) than their unbroken counterparts (3.1 mm), an almost
identical value to that of unbroken stockpile-quality flakes (3.4 mm). These data
indicate that the high-quality artefacts were probably broken during use before being
discarded.
XU19 (3819–3603 cal. BP) was dominated by high-quality material (n=41), with a
smaller quantity of stockpiled material present (n=17). The stockpiled material
indicated Place Provisioning, as reduction intensity was higher than at Gibb Rock
(avg. 2.5–3 scars, 2 platforms, 5.88 mm/scar, 0.71g/scar). In addition, a single
backed piece was found; this piece would not be required under a Short-term
Expedient strategy, as the required task could be performed with the transported
toolkit. Among the higher quality material, reduction intensity was somewhat varied.
283
Values were higher for the single core, which was sufficiently small to discourage
further reduction (4 scars, 4 platforms, 2.25 g, bipolar). Values were not particularly
high for large flakes, and averaged 2.4 scars, 1.8 platforms and 6.53 mm/scar.
Several flakes displayed smaller, overlapping flake scars on their dorsal surfaces,
indicating they had been struck from heavily reduced parent cores, but there were no
signs of standardised flaking techniques. Four of the five large flakes were broken,
as was the single retouched piece. However, due to the small stockpile-quality
assemblage, it was not possible to determine whether this reflected post-depositional
breakage or preferential discard. As such, the high-quality assemblage may have
derived from Individual Provisioning, Gearing Up, or a combination of both strategies.
A Place Provisioning strategy clearly dominated in XU20–21 (4571–3819 cal. BP).
The stockpile-quality material exhibited fairly high reduction intensity (avg. 2.5–8
scars, 2.1–4 platforms), and the single core was flaked by bipolar reduction due to its
small size. Large flakes (n=19) showed extensive variation in individual flake scar
(0–6) and platform counts (1–4), and seven had more than 25% of their surface
unmodified. Several of these flakes had staining on their dorsal surface – this stain
overlay some scars but had been lost in other areas due to subsequent flake
removals. This indicates a reuse of older/stored material that is characteristic of
Place Provisioning. Finally, three retouched pieces were present (all intact),
indicating some Gearing Up, or the need for specialist tools for particular tasks.
High-quality material was also well-represented in XU20–21, but reduction intensity
data were only available from a single core and six large flakes. The core was quite
heavily reduced (7 scars, 4 platforms, 0.88 g/scar) but, based on its weight (6.15 g),
it could probably be reduced further. Large flakes averaged 3 scars and 1.6
platforms, which was fairly similar to the Place Provisioning assemblage. Two of
these flakes had parallel flake scars, indicating some level of standardisation in
previous flake removals, and most of the large flakes were themselves fairly long
and narrow. Half of the large flakes were broken; this may represent selective
discard, as these flakes were only slightly thinner, on average, than their unbroken
counterparts. Two retouched pieces were also present – neither was broken or
heavily damaged. Most of the assemblage seems more consistent with Gearing Up,
since reduction intensity was not particularly high, and discard was not limited to
284
broken items or exhausted cores. Nevertheless, some isolated visits may have
occurred under Individual Provisioning.
Lower deposits (XU22–32)
XU22–28 (7350–4571 cal. BP) yielded a smaller assemblage so the technological
signatures were less clear. The deposit was dominated by stockpile-quality material,
but small amounts of high-quality material were also present. The stockpiled material
was fairly lightly reduced – average values were similar to those at Gibb Rock,
except large fragments were considerably lighter (avg. 2–2.3 scars, 1.7–2 platforms,
8.84 mm/scar, 0.19 g/scar). This lower reduction intensity, combined with the overall
scarcity of artefacts (and low discard rates), might indicate Short-term Expedient use
of stockpiled material. A single retouched piece was present, which may indicate that
some longer visits occurred, under a Place Provisioning strategy. Higher quality
materials were represented by three small quartz fragments and 22 small flakes (19
quartz, 3 BIF). The absence of larger items, and the presence of exotic materials, is
more consistent with artefact maintenance, and possibly some primary reduction of
small transported cores, as would be the case under Individual Provisioning.
The lowest deposits, XU29–32 (8200–7350 cal. BP) were associated with the
smallest sample size. Stockpile-quality material was limited to two large flakes from
XU30, so reduction intensity data were incredibly limited. Nevertheless, flake scar
average was low (1.5) and one of the flakes had 25–50% of its surface unmodified.
This appears to reflect Short-term Expedient use of stockpiled material. Higher
quality quartz was more abundant, and comprised two small fragments, seven small
flakes and three large flakes. Again, reduction intensity data were limited and of little
value. However, there was clearly some on-site reduction of higher quality material,
and some small flakes and fragments may have derived from secondary modification.
Due to the small sample size, it is difficult to attribute this assemblage to Individual
Provisioning or Gearing Up, but the scarcity of artefacts would seem to support the
former, since reduction sequences would be shorter.
285
13.4 MULKA'S CAVE
As in Chapter 12, the Mulka's Cave assemblages have been separated by site area:
the Main Cave Area, the Humps, and the Camping Area. Sample sizes are variable
both within and between areas, but most were large enough to permit fairly confident
identification of technological systems. Radiocarbon dates were obtained from all
three areas, allowing timeframes to be placed on most archaeological deposits, and
the technological provisioning systems identified therein.
13.4.1 Main Cave Area
In the Main Cave Area, small sample sizes hampered the identification of
technological systems, particularly in MC1 and the Entrance Platforms. At the latter,
only eight artefacts were present, all on stockpile-quality (moderate–good) material.
Reduction intensity measures were generally higher than at Gibb Rock on the core
and large flakes (avg. 3.4–4 scars, 2.2–3 platforms, 0.44 g/scar, 5.41 mm/scar),
while large fragments were less intensively reduced (avg. 1 scar, 1 platform, 2.58
g/scar). All five large flakes had their entire dorsal surface covered by negative flake
scars (i.e. no unmodified surface), meaning the parent cores had been previously
flaked, elsewhere on-site. This reuse of stockpiled material is consistent with Place
Provisioning, but the small number of artefacts means that reduction intensity
measures can be heavily skewed by previous flaking events. The assemblage may
represent a short site visit, when stockpiled materials were reduced in a Short-term
Expedient manner, or simply indicate that this area of the site was used differently to
others.
Stockpile-quality material dominated MC1, but reduction intensity was fairly low,
similar to that at Gibb Rock (avg. 2–2.1 scars, 2 platforms). While the lower reduction
intensity is more consistent with a Short-term Expedient strategy, two retouched
pieces were also found; as noted elsewhere, retouched pieces would not be
produced under that strategy. The assemblage probably represents a level of Place
Provisioning, but with a shorter reduction sequence than normal; this potentially
indicates this part of the site was used less intensively or in a different manner to
other parts of Mulka's Cave. High-quality materials, all quartz, were much less
abundant. This assemblage comprised one core, two large flakes and two retouched
pieces. The core was highly reduced (0.65 g), and made on a large natural quartz
286
crystal, an ideal choice for transported cores. The large flakes had been more
intensively reduced than the stockpiled material (avg. 3 scars, 2 platforms, 5.12
mm/scar), and one preserved an abundance of small, partial scars, indicating it was
struck from a highly reduced parent core. None of the large flakes or the retouched
pieces were broken, but one large flake had suffered heavy edge damage. The
sample size was too small to accurately separate Individual Provisioning from
Gearing Up, and indeed each are supported by different features of the assemblage.
The highly reduced transported core is more consistent with Individual Provisioning,
while the discard of intact pieces is more synonymous with Gearing Up.
Column Sample
The Column Sample inside Mulka's Cave yielded the largest sample size in the Main
Cave Area, so technological provisioning systems were more readily identified. In
XU2–6 (1570–48 cal. BP), stockpile-quality material dominated over high-quality
quartz, but the entire assemblage yielded limited data on reduction intensity, since it
was dominated by small flakes (n=13), from which no relevant attributes were
collected. The stockpiled assemblage comprised ten small flakes, two large
fragments and one large flake; it may have originated from Place Provisioning or
Short-term Expedient use of stockpiled material. The higher quality assemblage was
even smaller, just three small flakes, a small fragment, large fragment and a large
flake. While data were limited, reduction intensity measures were fairly high (4–6
scars, 3 platforms), especially when the size of pieces were considered (0.3 g/scar,
3.8 mm/scar). Furthermore, the dorsal surface of the large flake exhibited several
small, partial scars, indicating it came from a highly reduced parent core – the flake
was possibly detached to rejuvenate this same core. The reduction intensity and
general scarcity of artefacts seems more consistent with Individual Provisioning than
Gearing Up.
XU7–9 (4250–1570 cal. BP) was dominated by high-quality material, but the
abundance of large flakes on lower quality material means reduction intensity data
were fairly abundant across both material types. The lower quality large fragment
had been lightly reduced (1 scar, 1 platform), while flakes had similar flake scar and
platform averages to Gibb Rock (2.1 scars, 1.8 platforms). However, by virtue of their
small size, these flakes showed a greater reduction intensity (4.99 mm/scar).
287
Similarly, they demonstrated significant variation in individual scar, platform and
unmodified surface values, considering the relatively small sample size. These
features are more consistent with Place Provisioning, as reduction sequences are
longer. The higher quality assemblage was generally consistent with Gearing Up –
reduction intensity was similar to the Place Provisioning assemblage (2 scars, 1–1.3
platforms), and most large flakes/retouched pieces were discarded without being
broken or heavily damaged. Only the single core, made on black smooth-textured
quartz, had been more heavily reduced (6 scars, 4 platforms, 0.54 g/scar). No other
artefacts were made on this same material, so the core may not have been flaked
on-site, but the very small pit size must be taken into account.
The lowest Column Sample deposits, XU10–11 (6000–4250 cal. BP), were
associated with the smallest sample size, which comprised only eight small flakes
and two small fragments, all on high-quality quartz. Reduction intensity data were not
available, but the dominance of small artefacts indicates that on-site activities
probably included artefact maintenance and/or flaking small, transported cores; this
is more consistent with Individual Provisioning.
13.4.2 The Humps
As in most other collections, the Humps Scatter yielded artefacts made on both
higher and lower quality material, the latter suitable for stockpiling (n=14 and 7,
respectively). The stockpiled material was heavily reduced (avg. 4.7–6 scars, 2–3
platforms, 0.37–0.43 g/scar, 4.96 mm/scar), so is more consistent with Place
Provisioning than a Short-term Expedient strategy. Similarly, the presence of a single
retouched piece is inconsistent with the latter. High-quality artefacts comprised six
large flakes, three small flakes, two cores, one small fragment, one large fragment
and one retouched piece. Based on the artefact types, some level of primary
reduction clearly occurred in this area, and indeed all of the small flakes are too large
to have been produced by retouch. Both cores were heavily reduced (avg. 6.5 scars,
2.5 platforms, 0.43 g/scar), bipolar, and probably discarded as they were too small to
flake further. Reduction intensity was also fairly high for large fragments (7 scars, 3
platforms, 0.29 g/scar) but less so among large flakes (avg. 2.8 scars, 2.2 platforms,
6.74 mm/scar). Three of the large flakes were broken, while the retouched piece was
intact. This assemblage seems more consistent with Gearing Up, based on the
288
discard behaviour as well as the frequency of on-site primary reduction. The higher
reduction intensity probably reflects some effort to conserve better quality material,
which would be encountered less regularly than stockpile-quality quartz.
Stockpile-quality material dominated throughout TH1 (n=132), while higher quality
material was comparatively scarce (n=24). The latter assemblage comprised 11
small fragments, eight small flakes, four retouched pieces and one large flake. Highquality material was most common between XU1–3 (601–0 cal. BP, n=14), while just
five pieces were present in the lower half of the pit (XU4–6, 1120–601 cal. BP). The
small flakes and fragments may have originated from secondary modification or
artefact maintenance, and only the single large flake provided concrete evidence of
on-site primary reduction. Three of four retouched pieces were broken, and one
retained probable hafting resin, indicating it was part of a composite tool, which
would be common in transported toolkits. The single large flake was also broken,
and its long, narrow, parallel flake scars indicate that it came from a more
standardised core. All of these features support Individual Provisioning.
The TH1 stockpiled material comprised good mix of small (n=88) and large artefacts
(n=44). Large fragments had a fairly low average scar and platform count (1.7 and
14, respectively) but showed a greater reduction intensity when their small size was
considered (avg. 0.26 g/scar); large flakes were more intensively reduced than at
Gibb Rock (avg. 2.4 scars, 2 platforms, 6.56 mm/scar). Overall, then, reduction
intensity is more consistent with Place Provisioning. A total of 13 retouched pieces
were also present – these may have been required for specific tasks under Place
Provisioning, or represent some level of Gearing Up, since shorter artefact use-life
may permit the use of slightly lower quality material. Nearly 82% of retouched pieces
were broken, compared to just 40% of large flakes, which may indicate some
selective discard. It is possible that some retouched pieces were part of Individual
Provisioning assemblages (i.e. the definition of high-quality material was too narrow),
but the Humps area may simply have been the focus of major retooling events prior
to long logistical forays.
289
13.4.3 Camping Area
The Camping Area preserved the largest stone artefact assemblages from Mulka's
Cave, so technological provisioning systems were relatively easy to identify, in most
cases. The surface material is discussed separately to each of the excavated testpits, CA1 and CA2.
Camping Area Scatter
Within the Camping Area Scatter, Clusters A–D yielded consistently similar
signatures. Reduction intensity of stockpile-quality material was higher than at Gibb
Rock across all measured artefacts, including cores (avg. 6–8.5 scars, 2–3
platforms), large fragments (avg. 2.7–5.5 scars, 1.3–2.5 platforms, 0.45–1.43 g/scar)
and large flakes (avg. 3–3.8 scars, 1.7–2.3 platforms, 4.61–6.61 mm/scar). One
large fragment from Cluster C exhibited long, parallel flake scars, indicating that
formal reduction was not always confined to high-quality material. This higher
reduction intensity, combined with greater artefact numbers, is characteristic of Place
Provisioning. In addition, individual flake scar, platform and size values were quite
varied, particularly in Cluster B, indicating longer reduction sequences. Two silcrete
cores provided concrete evidence of reuse of stockpiled material – these cores
displayed a total of 14 negative flake scars, but the entire Camping Area Scatter
yielded only six large silcrete flakes. These cores must have been flaked elsewhere
on site, or in the distant enough past that the resultant flakes were buried. This reuse
of older material can skew reduction intensity data, but the effect is negligible on
large, diverse assemblages such as those from Clusters A–D.
High-quality raw material was much scarcer in Clusters A–D. Large flakes were more
intensively reduced than the Place Provisioning assemblage (avg. 3–7 scars, 2–3
platforms, 2–6.1 mm/scar), but only one derived from a standardised core, based on
the presence of long, parallel flake scars. Both cores were bipolar and displayed ten
scars detached from 4–6 platforms, and were heavily reduced, weighing < 2.7 g.
One of the large flakes weighed considerably more than others (4.15 g), so could
have been used as a core if raw material conservation was integral, as it is under
Individual Provisioning. Furthermore, none of the five retouched pieces were broken,
and only one of the twelve large flakes was. Therefore, despite the smaller sample
size, the discard behaviour seems more consistent with Gearing Up; the increased
290
reduction intensity and some standardised flaking may be indicative of attempts to
conserve raw material at the core level.
Clusters E and F both yielded much smaller assemblages (n=7 and 6, respectively),
and indicated shorter term use of stockpiled material. In Cluster F, flake scar and
platform averages were fairly low (1–3, 1–1.5 platforms, 5.52–6.75 g/scar), and the
core had more than 75% of its surface unmodified. Values were considerably higher
in Cluster E (avg. 5.5–6 scars, 3 platforms), but the small assemblage size indicates
that much of this reduction happened elsewhere on-site, so does not accurately
characterise the level of stone flaking in this area. A single retouched piece was
present at Cluster E, a large flake (> 50 g) that had several flakes detached from its
ventral surface. This probably reflects the use of an older, larger flake as a source of
raw material, rather than retouching a flake for a specific purpose, so is not
inconsistent with the Short-term Expedient use of stockpiled material.
Higher quality material was scarce in both Clusters E and F, limited to just one small
flake at the latter, and a single large flake and retouched piece at the former. It was
difficult to make any conclusions on the basis of so few artefacts, but the large flake
retained an original crystal facet, so came from a large crystal that could have been
used as a transported core. The retouched piece was possibly broken which, in
conjunction with the scarcity of artefacts, may indicate some level of Individual
Provisioning.
The surrounding, low-density material from outside the Clusters was more difficult to
interpret. Stockpile-quality material was fairly heavily reduced (2.9–11 scars, 1.4–5
platforms, 0.82–1.3 g/scar, 5.76 mm/scar), which is more consistent with Place
Provisioning. However, considering the low-density spread of the material (70
artefacts in 3900 m2), pieces probably preserved evidence of reduction that occurred
elsewhere on site, rather than in the immediate vicinity of the discard location. Highquality material was scarcer and reduction intensity varied, but was generally slightly
higher than the stockpiled assemblage (2.5–3.9 scars, 2–2.5 platforms, 0.37 g/scar,
4.39 mm/scar). Only two of the 13 large flakes were broken, and neither of the
retouched pieces. Despite the limited data, this seems inconsistent with Individual
Provisioning, so may represent some level of Gearing Up, or simply the use of higher
quality material within a Place Provisioning strategy.
291
CA1
The sizeable assemblage of stockpile-quality material in CA1 preserved clear
evidence of Place Provisioning throughout the entire occupation period (2500 cal.
BP–recent past). Reduction intensity was consistently higher than at Gibb Rock,
across all artefact types, when all XUs were grouped (avg. 2.1–7.2 scars, 1.7–3.6
platforms, 0.61–2.4 g/scar, 5.27 mm/scar); even when XUs were considered
separately, reduction intensity always exceeded that of a Short-term Expedient
strategy. There was considerable variation among individual flake scar and platform
counts, as well as the amount of unmodified surface. Several cores had been flaked
to exhaustion, while others retained usable material. This variation, in combination
with higher average values, is indicative of the longer reduction sequences that
characterise Place Provisioning.
Higher quality material occurred in all but the basal unit, XU9 (2500–2158 cal. BP),
but the provisioning systems varied over time. Gearing Up dominated in XU1 and 3–
4, while XU2 and 5–8 assemblages were more consistent with Individual
Provisioning. In XU1 (144–0 cal. BP) and XU3–4 (1056–324 cal. BP), high-quality
material was fairly abundant, but only large flakes (n=23) provided reduction intensity
data. Average values were very similar to the stockpiled assemblage, but the flakes
themselves were slightly smaller (avg. 3 scars, 1.9 platforms, 4.98 mm/scar). Discard
behaviour showed no conservation of raw material at the tool/flake level, as
breakage rates were incredibly low – only 15% (n=3) of large flakes (n=3) and 17%
(n=1) of retouched pieces were broken. This lack of conservation is much more
consistent with Gearing Up than Individual Provisioning. Three retouched pieces on
stockpile-quality (from XU3–4) may also be part of this Gearing Up assemblage, if
lower quality material was indeed suitable when artefacts did not have to be
preserved indefinitely. Exotic raw materials were scarce, but the presence of three
pieces in XU3 may indicate that some level of Individual Provisioning occurred at this
time.
Individual Provisioning is represented in XU2 and XU5–8, but these XUs preserve
different evidence of the same strategy. In XU2 (324–144 cal. BP), the assemblage
comprised one small fragment, nine small flakes (including two exotic raw materials)
and five large flakes. A distinct crust was evident on the dorsal surface of all large
292
flakes, and many of the smaller flakes. Considering the absence of crusting on the
ventral surfaces, it was clearly formed when the flakes were attached to the parent
core. If the aim of reduction was to produce flakes for on-site use, one would expect
a mixture of flakes, some with the crust (representing those with dorsal surface
exposed during the crust-forming event), and others lacking it (those struck after the
crust was removed). It seems more likely, then, that the flakes were produced when
the core was trimmed, to remove the crusted surface and expose the unaffected
material beneath. This indicates some desire to maximise utility of material per unit
weight, by removing waste prior to transport. This is more consistent with Individual
Provisioning, where greater mobility means that transport costs must be reduced
wherever possible. Transported cores were an important part of Individual
Provisioning, representing toolmaking potential that is otherwise lacking, considering
the distance/time to next source is unknown; this reduces the risk of being without a
certain tool when required. This same risk is absent on logistical forays, since they
are shorter and target a specific range of resources before owners return to base
camp, where raw material stocks are available. It is more efficient, then, for people to
equip themselves with the tools required for the appropriate tasks, than to transport
cores.
XU5–8 (2158–1056 cal. BP) also preserved clear evidence of Individual Provisioning
but, in this case particularly through selective artefact discard, which was primarily
limited to small/unusable pieces and broken items. Large flakes were rare – only two
were present, one of which was broken; the single retouched piece was also broken.
The majority of the assemblage comprised small flakes (mostly unbroken) and
fragments < 10 mm. These may have been produced by flaking small, transported
cores, or from reworking existing artefacts. Finally, there were a greater number of
artefacts made on exotic materials (n=7). These materials are more likely to be
components of transported toolkits, since higher quality material flakes more
predictably and prolongs artefact use-life.
CA2
As in CA1, evidence of Place Provisioning was found throughout CA2. The
stockpiled material comprised moderate to good quality quartz, as well as one piece
of coarser grained silcrete. One core was recovered, which exhibited evidence of
293
extensive prior reduction (16 scars, 6 platforms), but still retained a good amount of
usable material (23.15 g). Large fragments and large flakes also showed higher
reduction intensity (2.4–3 scars, 1.8–2 platforms, 0.34 g/scar, 5.16 mm/scar) than
would be expected under Short-term Expedient strategies, and a greater range of
individual values, indicating longer reduction sequences. It is generally hard to detect
Short-term Expedient use of stockpiled material among a Place Provisioning
assemblage, since reduction intensity values are skewed by previous reduction
events, and Place Provisioning creates far more archaeological debris. However,
there was some evidence in the form of a unifacially retouched piece from XU4. This
flake was fairly large and heavy (4.1 g), and the retouch comprised a single large
flake removal from the ventral surface. If the aim was indeed to detach and use the
new flake, rather than modify the form of the existing flake, it would seem
inconsistent with Place Provisioning, as a new flake could be struck from the parent
core, or other stockpiled material. It seems more consistent with the reuse of a
previously discarded flake for a new purpose (in this case, as a core), which is more
synonymous with a Short-term Expedient strategy. Nevertheless, Place Provisioning
dominated throughout the entire occupation period, from 6700 cal. BP until the
recent past.
As in CA1, the character of high-quality assemblages varied over time. XU1–3
(2982–0 cal. BP) were dominated by evidence of Gearing Up. Large fragments and
flakes exhibited lower reduction intensity than the Place Provisioning assemblage
across most measures (avg. 1.3–2.3 scars, 1.3–1.7 platforms, 0.28 g/scar, 5.89
mm/scar), indicating shorter reduction sequences. Breakage rates were low – only
two of six large flakes were broken, and one of nine retouched pieces. This is
inconsistent with the conservation behaviour expected under Individual Provisioning.
Four non-quartz pieces were found, as well as two small flakes that preserved
evidence of more standardised reduction; these may indicate that some visits
occurred under Individual Provisioning.
In contrast to the upper deposits, XU4 (4792–2982 cal. BP) preserved clear
evidence of Individual Provisioning. Reduction intensity was consistently higher than
the Place Provisioning assemblage, across large fragments (3 scars, 3 platforms)
and large flakes (avg. 3.3 scars, 2.7 platforms, 4.44 mm/scar). Five of the seven
294
large flakes exhibited evidence of standardised reduction, with more than one
indicator present on several flakes. One retained long, narrow, parallel flake scars,
while three flakes are long and narrow themselves. Another three have notches on
their lateral margins, comprising regularly sized and spaced partial flake scars
(distinct from retouch), created from previous flake removals when the standardised
parent core was oriented differently (Figure 12.20). The fact that few large flakes
were found indicates that much of this reduction occurred elsewhere, as would be
the case under Individual Provisioning. One small flake also exhibited parallel flake
scars, and another at 90 degrees to it, indicating standardised reduction and rotation
of the parent core; the remainder of the small flakes are fine enough to have been
produced by retouch. Both retouched pieces were broken, but only three of the
seven large flakes were. However, the remainder may have been produced to
rejuvenate the core, rather than to use the resultant flake, so their discard does not
necessarily indicate a lack of raw material conservation. While greater reduction
intensity could occur under Gearing Up, the selective discard and standardised
reduction are much more characteristic of Individual Provisioning.
XU5–6 (6700–4792 cal. BP) yielded much smaller assemblages on high-quality
material; these comprised a single retouched piece, four small fragments and four
small flakes (all whole). The retouched piece was not broken or heavily damaged;
this is more consistent with Gearing Up, since functional pieces should not be
discarded under Individual Provisioning. However, considering its extremely small
size (0.05 g), it may represent artefact loss rather than intentional discard. Similarly,
there was little evidence of primary reduction, since all small flakes and fragments
could have derived from reworking existing artefacts or converting flakes into tools;
the dominance of secondary reduction is more consistent with Individual Provisioning.
Therefore, while the small sample size must be considered, the limited assemblage
from XU5–6 is more likely a product of Individual Provisioning.
13.5 CONCLUSION
Using the data provided in Chapters 10–12, it was possible to identify the various
technological provisioning systems employed at the three studied sites. At Gibb
Rock, assemblages were small and individual collections were often quite restricted.
295
A Short-term Expedient strategy dominated at a site-wide level, evident from the use
of low-quality, local quartz, and lower reduction intensity measures. Provisioning
systems were much more varied at Anderson Rocks and Mulka's Cave; different
systems dominated at particular times, or in specific site areas. At Anderson Rocks,
the first 3600 years were dominated by Individual Provisioning and Short-term
Expedient strategies, evident from limited artefact discard and less intensive
reduction of lower quality material. Around 4600 cal. BP there was a shift to Place
Provisioning, which continued for the next 2700 years, but some Individual
Provisioning and Gearing Up also occurred during this period. From 1870 cal. BP the
dominant system shifted from Place to Individual Provisioning several times. The
surface deposits were dominated by Place Provisioning, but a small assemblage on
the outcrop indicated Short-term Expedient reduction of low-quality quartz from a
local vein.
In the Main Cave Area at Mulka's Cave, only the Column Sample provided a large
enough assemblage to permit concrete identification of the technological
provisioning systems in use. Individual Provisioning dominated in the earliest (6000–
4250 cal. BP) and latest deposits (1570–48 cal. BP), while Place Provisioning, often
supplemented by Gearing Up, occurred between. On the Humps outcrop, the
subsurface deposits indicated the simultaneous use of contrasting strategies,
Individual and Place Provisioning, throughout the 1120-year occupation period. The
surface assemblage only preserved evidence of Place Provisioning and Gearing Up.
The latter was evident from extensive on-site reduction of high-quality materials, and
the discard of unbroken, undamaged artefacts made on the same material. In the
Camping Area, Place Provisioning dominated the surface and subsurface deposits
throughout the entire 6700-year occupation period. Stockpiled materials were
intensively reduced, and the variance in individual values indicated longer reduction
sequences. Supplementary and contrasting strategies were identified in various
deposits, based on assemblages made on higher quality raw materials: Individual
Provisioning was more common in the earlier deposits, while Gearing Up occurred
more frequently in the later assemblages.
296
CHAPTER 14
MOBILITY, SITE USE AND OCCUPATION INTENSITY
14.1 INTRODUCTION
This chapter describes how Anderson Rocks, Gibb Rock and Mulka's Cave were
used, and how site use changed over time. To begin, the resource and occupation
model is used as a framework to interpret the technological provisioning systems
identified in Chapter 13, considering both the dominant (Place Provisioning,
Individual Provisioning) and supplementary systems (Gearing Up, Short-term
Expedient). These data identify the season/s during which each site may have been
visited, how different parts of the site were used, and the general level of residential
mobility throughout the occupation period. Then, changes in occupation intensity
over time are considered by identifying those periods of higher or lower intensity
superimposed on the normal pattern evident within a selected pit; this can only be
performed for dated deposits that yielded sufficiently large stone artefact
assemblages. Finally, potential reasons for intensity shifts are identified, with
particular links to water availability.
14.2 GIBB ROCK
No samples from Gibb Rock were dated, but based on the limited depth of GR1, the
cultural material probably dates to the last few hundred years; older material may be
present in unsampled parts of the site. Further, the assemblages were too small to
consider change over time. While limitations in intra-site sampling must be
considered, the available evidence suggests that Gibb Rock was probably not a
residential location, even in the short-term. There was no evidence of the Place
Provisioning strategy that should accompany long stays, and no trace of toolkit
maintenance that characterises Individual Provisioning – these renewal activities
occur at residential sites where downtime is predictable and abundant (Clarkson
2007:15; Torrence 1983). The lack of Place Provisioning is not particularly surprising,
given the absence of gnammas or rockholes that store water for longer periods. The
only time the site could be visited for longer periods would be during winter, when
surface water was freely available. As noted elsewhere (see Chapter 8.3.1), the
297
optimal response to more widespread water availability would be for people to
spread out into the landscape, exploiting the areas that could not be visited during
drier months. This increased mobility is incompatible with Place Provisioning. The
lack of Individual Provisioning is more surprising, since it indicates that the site was
not used as a residential base at all, even under a system of increased residential
mobility.
Short-term Expedient assemblages, such as those found at Gibb Rock, are created
when a particular tasks could not be completed with the available tools. While this is
often associated with Individual Provisioning, due to the use of a generalised mobile
toolkit, it could also occur under Place Provisioning if an unanticipated task occurred
away from the home base. It is therefore not possible to evaluate the general
mobility of the Gibb Rock visitors based solely on the expedient assemblages
preserved. It is only possible to conclude that the site was visited for very short
periods, probably during foraging trips, and was not a residential location. If the
predictive model is correct, these trips would have occurred in spring–summer, when
people were based at a well-watered location nearby. This would explain the
absence of Individual Provisioning assemblages, since the lack of water would make
it impossible for people to camp onsite at that time.
14.3 ANDERSON ROCKS
Pit AR2-3 indicates that Anderson Rocks was visited from 8200 cal. BP until the
recent past. XU1–32 all yielded cultural material, indicating that there were no
significant hiatuses in site occupation. However, it is impossible to address continuity
of occupation at a finer scale with the available data. Nevertheless, it is possible to
address questions of site use and changing occupation intensity over time, using the
data outlined in Chapters 10 and 13.
14.3.1 Mobility and Site Use
The technological provisioning systems identified in Chapter 13 provide important
information on mobility, which can be combined with other indicators to determine
when and how Anderson Rocks was used. Occupation evidence was clustered in
two main site locations: most was in the vegetated area near the gnammas, while
298
smaller assemblages were present several hundred metres northwest, atop the
outcrop. These site areas appear to have been used in different ways, so are
discussed separately below.
Gnamma area
The depth of archaeological deposit in AR2-3 clearly indicates that this same area
was regularly reoccupied over several thousand years. Considering the proximity to
gnammas (approx. 50 m), this suggests that those structures provided the main
source of water during these visits – otherwise, there would be no reason for
occupation to be repeatedly tethered to this same location. It has been demonstrated
that, as a long-lived water source, gnammas would not be used until more
ephemeral sources (e.g. surface water, soil, pans) had been exhausted (see Chapter
8; Anderson 1984; Bird 1985; Gould 1968, 1969b; Tonkinson 1978:29). Therefore, it
is unlikely that the bulk of occupation evidence was deposited in winter, or soon after,
but instead when water was no longer widely available.
Following the model derived earlier (see Chapter 8.3), Anderson Rocks should have
been occupied in November and/or March–May, when the central study area was
the optimal residential location. The gnammas would regularly hold water in
November, but the precise quantity available depends on the amount of spring
rainfall, which preserves the water balance by partially offsetting evaporation. At this
time, occupation would be predictable, and extended stays would have been
permitted under most rainfall conditions; this would favour Place Provisioning. Some
visits could occur from March–May, but this would depend on the size and frequency
of autumn rainfall events; these sporadic visits would favour Individual Provisioning.
The evidence from AR2-3 and Scatter 3 indicated two fairly distinct occupation
periods, with the transition between the two occurring around 4571 cal. BP, at the
interface between XU21 and XU22. In the earlier deposits (8200–4571 cal. BP),
Individual Provisioning was dominant, with some Short-term Expedient reduction of
stockpiled materials. This indicates that visits were shorter and/or more sporadic,
which necessitated the maintenance of a formal, transported toolkit, supplemented
by expedient artefacts where required. While Individual Provisioning could suggest
the site was visited in autumn rather than spring, it seems an unlikely explanation for
299
the 3600-year dominance of that strategy. It is more probable that a reduction in the
normal water-holding period precluded long visits at this time, increasing residential
mobility. However, the presence of stockpiled material as low as XU30 indicates that
initial occupation may not have occurred under these conditions. Before 7775 cal.
BP, the site must have been visited regularly enough, and for long enough periods,
to warrant stockpiling material. It is therefore possible that there is additional
occupation evidence (displaying the Place Provisioning signature) in the unsampled
deposits, including those below XU33.
The later period (4571 cal. BP–recent past) was dominated by Place Provisioning,
but there were short periods where small sample size prevented the strategy being
clearly identified. Nevertheless, in the last four and a half thousand years visits to
Anderson Rocks must have occurred at predictable, regular intervals and for longer
periods. As noted above, this could probably only occur during spring, as the
gnammas' water balance declines swiftly from December. Several assemblages
were characteristic of 'Gearing Up', which involves the manufacture of specialist
tools before a logistical foray. This strategy is associated with longer site visits, when
restricted mobility has limited access to certain resources, requiring an increase in
logistical mobility. Gearing Up assemblages were particularly identifiable between
4571–3819 cal. BP (XU20–21) and 3170–1870 cal. BP (XU12–17), and in the recent
past (Scatter 3).
Despite dominance of Place Provisioning over the last 4500 years or more, there
was evidence of the contrasting strategy (Individual Provisioning) within several of
the same XUs (Figure 14.1). It was often difficult to separate the traces of Gearing
Up and Individual Provisioning, but the latter was particularly evident in XU18 (3603–
3170 cal. BP), XU5–8 (1220–354 cal. BP) and XU1–3 (138–0 cal. BP). The
intermixture of Place and Individual Provisioning could arise in one of two ways. It
may indicate a short-term shift between the two strategies across a timescale short
enough to preserve both in the same deposit – for example, a period of several
years or more where longer visits were suspended, and Place Provisioning was
replaced by Individual Provisioning. There is, however, an alternate possibility: that
these contrasting strategies were complementary to a certain degree. If Place
Provisioning reflects the normal, predictable occupation, then Individual Provisioning
300
may reflect site visits at other times of year. These out of season visits could occur
after large isolated summer/autumn falls, when gnammas would hold water for a
short period, allowing people to temporarily move further afield from their wellwatered refuges. Individual Provisioning was particularly common within the last
1200 years. This may indicate an increase in the conditions that facilitated out of
season visits, or more frequent shifts in occupation patterns at this time.
Figure 14.1 Technological provisioning systems (top) and occupation intensity (bottom) at Anderson
Rocks, based on dated archaeological content from AR2-3 and surface material; the latter was
presumed to date from the last 50 years or so. Where provisioning systems could not be confidently
identified, these have been omitted.
Outcrop
Occupation evidence was less abundant on the outcrop itself, limited to
assemblages from Scatters 1 and 2, and test-pit AR1. Scatter 2 preserved quite a
distinct assemblage compared to others from Anderson Rocks, as artefacts were
made on local raw material obtained from a narrow, exposed quartz vein. The entire
artefact use-life, from quarry to discard, occurred at the raw material source. If
people were camping on-site, at least some of these activities would probably occur
301
in other areas of the site, such as the main occupation zone. Similarly, the visit (and
associated task) was possibly short-term enough that it was not worth collecting
material from the stockpile, considering the availability of the local material.
Therefore, the assemblage probably derived from people passing through, possibly
on a foraging trip, while based at a site nearby.
Scatter 1 and AR1 showed repeated occupation of the same small area atop the
outcrop. AR1 was not dated, so temporal change can only be considered in terms of
earlier and later deposits. The lower/earlier deposits (XU10–13) preserved evidence
of Individual Provisioning, indicating shorter visits, and a higher degree of mobility at
this time. This is unlikely to represent a widespread increase in mobility, based on
the dominance of Place Provisioning in AR2-3 over the last 4500 years. Instead, it
probably represents out of season visits to Anderson Rocks. It seems unlikely these
visits would occur in winter (or any period of sustained rainfall), as runoff from the
outcrop makes the area fairly damp and unpleasant. The upper deposits (XU1–9)
exhibited less intensive reduction of stockpiled material. However, the presence of
several retouched pieces on stockpile-quality quartz indicated the assemblages did
not derive from a Short-term Expedient strategy. Instead, the area was probably
used within the larger Place Provisioning system but in a different way. While the
gnamma area was the focus of occupation, the outcrop itself may have been used
for a more casual, restricted set of stone-working activities. Indeed, Scatter 1
preserved direct evidence of Place Provisioning and Gearing Up, indicating that
longer visits had occurred on the outcrop in the recent past, at least.
14.3.2 Occupation Intensity Over Time
Change over time could only be evaluated for AR2-3, as no other deposits were
dated. However, its large assemblage size combined with the increased depth of
cultural deposit means that periods of higher and lower intensity are readily
identifiable. XU15–17, 18 and 20 all showed evidence of more intensive occupation,
but in different ways. XU15–17 preserved a strong Gearing Up signature, which
indicates that site visits were long enough to require logistical forays, for which
specific tools were made. In addition, artefact discard rates were high, and there was
a considerable degree of post-depositional flake breakage that can accompany more
intense occupation periods. Most cores were incredibly small (≤ 1.6 g) bipolar, and
302
probably could not be flaked further – this indicates some degree of raw material
conservation at the core level, which often accompanied longer visits. Artefact
discard rates were also high in XU18 and XU20, and each exhibited varied individual
flake scar, platform and unmodified surface values. These features are indicative of
extended reduction sequences that accompany longer or more intense occupation
periods. Further, all cores from XU18 were bipolar, and XU20 preserved strong
evidence of Gearing Up.
Less intense occupation is evident when Individual Provisioning or Short-term
Expedient strategies dominate all deposits from a particular period. This indicates a
shift from longer occupations (under Place Provisioning) to shorter and/or more
sporadic visits. From 8200–4571 cal. BP (XU22–32), stockpiled materials were
reduced far less intensively, and high-quality assemblages probably represented
reworking of existing pieces rather than large-scale artefact manufacture. Artefact
discard rates were also at their lowest in these deposits. As noted above, the
presence of stockpiled material from 7775 cal. BP means that earlier or
contemporaneous Place Provisioning episodes had occurred. Therefore, the short,
sporadic visits preserved in the lower deposits do not reflect initial occupation of
Anderson Rocks.
It is also important to consider the reasons for interruption of the normal occupation
patterns, specifically the less intensive occupation from 8200–4571 cal. BP and more
intense periods from 4036–3819 and 3603–2520 cal. BP (Figure 14.1). As
demonstrated above, regular use of the AR2-3 area was probably linked with the
gnammas. As a result, the water preserved within them dictated the intensity of
occupation that could be supported at any given time. Shifts in intensity would not be
expected if gnammas were always capped at their peak, since they filled to capacity
even under low rainfall conditions (Figure 10.2). Instead, intensity was probably
influenced by the natural water-holding periods, which dictated the quantity of water
available when people arrived on-site in late spring. As noted elsewhere (see
Chapter 10.2.1), longevity of water was strongly influenced by the amount of spring
rainfall. Strong spring falls frequently followed average or high winter rainfall, but
rarely occurred when winter rainfall was low (BoM data). Therefore, lower intensity
occupation may occur during low rainfall years, while higher intensity would only be
303
possible in average or high rainfall years when spring falls were strong. Any
extension of the normal occupation period would be beneficial, as it would delay the
move to the normal summer refuges; occupation was already tethered to those sites
for long periods due to the scarcity of reliable water sources in the study area.
14.4 MULKA'S CAVE
Mulka's Cave was visited from around 9650 cal. BP but most evidence post-dates
6700 cal. BP. Thereafter, the site was visited regularly until the recent past, but
different parts of the site were used in different ways, and at different times.
Changing intensity was especially evident in the Camping Area, where the pits were
well-dated and yielded the largest stone artefact assemblages.
14.4.1 Mobility and Site Use
Based on the artefact assemblages and technological provisioning systems identified
therein, site use varied across Mulka's Cave. The Humps and Camping Area
preserved similar evidence, so are considered together, while the Main Cave Area is
treated separately. Site-wide data are then considered to evaluate potential evidence
for inter-group gatherings at Mulka's Cave.
Main Cave Area
The Main Cave Area yielded the smallest assemblages, so technological signatures
were often unclear. Nevertheless, artefact discard rates and other assemblage
characteristics indicate that occupation was less intensive inside and around the
rockshelter than in other parts of the site. It is important to note, however, that most
occupation evidence from this area was out of context, so intensity may be
underestimated by considering only the undisturbed portions of the assemblage. In
any case, the rockshelter and the slope outside do not appear to have been the
focus of activity at the site, contra to Gunn (2006). It may have been visited
separately to, or in conjunction with, other parts of the site.
The earliest deposits from the Column Sample (XU10–11, 6000–4250 cal. BP) inside
Mulka's Cave preserved evidence of Individual Provisioning, indicating greater
mobility at this time, at least in terms of visits to rockshelter itself. It is uncertain
304
whether the same signature is preserved in basal portions of the 1989 test-pit, but
low artefact numbers may support this (Rossi 2014a). From 4250–1570 cal. BP
Place Provisioning dominated, with some evidence of Gearing Up, which indicates
that stays were more regular and longer, and that some visits were long enough to
prompt logistical forays. Interestingly, it is during this period that the rock art was
created, following Gunn's (2004, 2006) age estimate of 3000–2000 cal. BP. These
longer site visits may have been associated with rock art creation, and indeed Gunn
(2004, 2006) concluded that motif superimposition indicated that Mulka’s Cave had
been visited regularly over the last few thousand years. However, the inferred link
cannot be further evaluated with the available data.
From 1570 cal. BP, Column Sample assemblages demonstrated a mix of Individual
Provisioning and Short-term Expedient use of stockpiled material. At this time, then,
greater mobility and shorter and/or less predictable visits again prevailed. While MC1
was not dated, the artefact-bearing layers (XU3–6) probably date to this same period.
The assemblages indicated some Short-term Expedient artefact manufacture along
with a formal strategy that could not be reliably identified, but had some features in
common with Individual Provisioning. The uppermost deposits in the Column Sample
and MC1 were both sterile, but surface artefacts, including one flaked glass piece,
demonstrate that these areas were visited in the recent past, and that people were
still using stockpiled material for expedient artefact manufacture.
Camping Area and The Humps
All three pits (CA1, CA2, TH1) showed repeated reoccupation of the same areas, but
different time periods and activities were represented by each pit. The Camping Area
was clearly the focus of occupation at Mulka's Cave, based on the quantity and
spread of surface artefacts, as well as the stratified deposits yielding archaeological
material. This fits well with the assertions made by Traditional Owners, who stated
they used to camp there when they visited the site in their childhood. This, in turn,
indicates some continuity in site use from the distant to the more recent past. The
Camping Area was demonstrably in use during period the rock art was created but
may also have preserved some more tangible links in some of its deposits. Ochre
was found in three XUs deposited from 4792–1173 cal. BP, with most of the sample
305
post-dating 3000 cal. BP, when Gunn (2006) argued the rock art began to be created.
However, ochre had a variety of uses, not all connected to art production.
The Humps outcrop was also visited regularly, but for a far shorter period than other
parts of the site. The lower reduction intensity (compared to the Camping Area)
possibly indicates that it was primarily used as a satellite area for stone-working.
This location may have been chosen for the excellent views it affords of the
surrounding area, allowing occupants to monitor the movements of people, animals,
rain clouds, or even bushfires.
Throughout the Camping Area and The Humps deposits, Place Provisioning
dominated from the earliest assemblages (6700 cal. BP) to the surface material,
indicating that visits to the site had always been predictable and long enough to
warrant stockpiling raw material. As a result, some inferences can be made
regarding the main season of occupation. As noted elsewhere (Chapter 8.5.2),
winter water abundance would favour Individual Provisioning, as people were free to
move around the landscape and visit areas that were inaccessible during the drier
months. Therefore, it would be suboptimal to stay at Mulka's Cave for long periods
during winter. In contrast, it would be the ideal spot in September/October, when
freedom was curtailed because all ephemeral sources of water had dried up, but the
southern part of the study area was still the optimal foraging location. Similarly, while
gnammas could be capped to preserve their water supply, the largest source – the
cleft rockhole – could not. Due to its large surface area and sloping walls, it also lost
water at a greater rate than gnammas. Considering the abundance (up to 6800 L)
and ephemerality of this source, it would invariably have been used. It may have
prompted some level of congregation, whereby people came together after winter to
use this shorter lived supply before dispersing to the smaller, more reliable sources
in summer. In this case, the dominance of Place Provisioning is not surprising. The
rockhole provided a reliable source of water, even after low winter rainfall, which
permitted long, predictably timed visits under all rainfall conditions.
Several deposits also preserved evidence of Gearing Up, which results from
restricted residential mobility and the need to travel further afield to procure specific
resources. The strategy first appeared in CA2 XU3, but due to the slow rate of
deposition, the date range was fairly broad (2982–1172 cal. BP). The basal XU of
306
CA1 dated to 2500 cal. BP, but Gearing Up assemblages were only present in
deposits post-dating 1056 cal. BP. The strategy was well represented in surface
material both in the Camping Area and The Humps; its absence in TH1 probably
related to that part of the site being used in a different way, as noted above.
Therefore, these longer visits, and the logistical forays they prompted, possibly
occurred as early as around 3000 cal. BP, but became increasingly common in the
last 1000 years or so, if the excavated deposits can be considered a representative
sample.
In contrast to Gearing Up, traces of Individual Provisioning were present throughout
most of the 6700-year occupation period – only deposits from 2982–2158 cal. BP
lacked clear evidence of the strategy (Figure 14.2). However, many individual XUs
preserved no evidence of Individual Provisioning, and there were no deposits where
this strategy was not accompanied by Place Provisioning. Nevertheless, Mulka's
Cave was regularly visited by more mobile groups who relied on transported toolkits;
such visits would have been short and less predictably timed. As noted above, the
coexistence of Individual and Place Provisioning in the same archaeological deposit
can indicate short-term shifts between the two strategies, or their use at different
times of year. The former seems unlikely, considering the lengthy period over which
both strategies are present, in contrast to the shorter periods of overlap at Anderson
Rocks (Figure 14.1). Further, the rockhole would always preserve enough water for
longer stays during the normal, post-winter occupation period, even in low rainfall
years. It seems more likely, then, that Individual Provisioning represents visits
outside the normal occupation periods associated with Place Provisioning. These
visits may have occurred during summer or autumn as, if capped, the gnammas
would have provided a reasonable source of water. This may have allowed people to
temporarily move from their well-watered refuges in other parts of the study area,
taking the pressure of the critical resources found there. Traces of Individual
Provisioning were distinctly less common in the last 1000 years or so, particularly
during the last 150 years (Figure 14.2). Despite the abundance of surface artefacts,
only Cluster F preserved possible traces of the strategy. During this time, the
predictable, longer visits still occurred, but out of season occupation was rare. It is
possible that European arrival may have interrupted traditional occupation patterns,
but this idea cannot be evaluated with the available data.
307
Figure 14.2 Technological provisioning systems (top) and occupation intensity (bottom) at Mulka’s
Cave. Provisioning systems are shown for all dated pits and the combined surface assemblages,
which were presumed to date from the last 50 years. CS = Column Sample. Where provisioning
systems could not be confidently identified, these have been omitted. Occupation intensity is based
on assemblages from the Camping Area.
Evidence of aggregation at Mulka’s Cave
Many have argued that Mulka's Cave had some ritual significance, often based on
the quantity and diversity of its rock art, as well as particular motifs (Goode 2011;
Gunn 2006; Macintyre et al. 1992; Webb and Rossi 2008). However, Rossi
(2010:111–113) was unable to find any supporting evidence in the archaeological
record, as assemblages from the Camping Area and Main Cave Area were fairly
similar. While the aim of this research was not to re-address the question of
ceremony at Mulka's Cave, it is worthwhile to consider the inferences that can be
made based on the newly available data, specifically whether people gathered at the
site, for ritual or social reasons. There are two main questions to consider – first,
308
whether the resource availability could support larger groups of people and, secondly,
if the site preserves any evidence of this.
As noted elsewhere (see Chapter 8.4), gatherings would be best held in the southern
part of the study area, in December, after high winter rainfall (one year in ten). The
timing and location would provide enough food and water for larger groups without
impacting the survival of the regular occupants. Mulka's Cave fits the location
criterion, but also provides a unique abundance of water. As demonstrated earlier
(see Chapter 12.2.1), the cleft rockhole captures water easily, and stores it in great
volume for a considerable period; this source could easily support larger groups of
people. Under average or high rainfall conditions (eight years in ten), the structure
held approx. 2900–4700 L of water at 1 December. Mulka's Cave would therefore
permit greater freedom in terms of which years gatherings could occur. While large
enough supplies would probably be available in low rainfall years, this water would
be used for general occupation, due to reduced availability elsewhere (see Chapter
14.4.2).
The rock art in Mulka's Cave began to be created around 3000–2000 years ago
(Gunn 2004, 2006), but both the rockshelter and the Camping Area had been visited
for thousands of years before this. Visits occurred on a regular basis both before and
after the rock art creation, based on the evidence of Place Provisioning throughout
the Camping Area deposits. Clearly, then, the entire occupation history cannot be
linked to ceremonial events, if the rock art is considered indicative of ritual site use.
However, the site still preserves some evidence that gatherings or ceremonial events
may have occurred within the broader occupation scheme. Indeed Gunn (2006)
argued that Mulka's Cave was probably used for a mix of general and ritual purposes.
Kelleher (2002:270) claimed that ceremonial sites should preserve distinct activity
areas, with the ritual areas separated from the main domestic areas; this should be
visible in artefact density and assemblage characteristics. As demonstrated above,
the Main Cave Area was used far less intensively than The Humps or the Camping
Area. While Place Provisioning dominated elsewhere, the rockshelter assemblages
more often displayed evidence of Individual Provisioning. This may indicate that it
was used in a different way to the rest of the site, with less focus on day to day
activities. However, the limited sample is problematic, and indeed the Column
309
Sample yielded some evidence of Place Provisioning from 4250–1570 cal. BP, so
some level of general occupation did occur inside Mulka's Cave. It was also argued
above that Individual Provisioning assemblages from the Camping Area probably
indicated
out
of
season
visits.
These
visits
may
have
occurred
for
ceremonial/gathering purposes; visitors may have camped in the designated area/s
and maintained their transported toolkits as required. If gatherings did occur at
Mulka's Cave, the general paucity of non-quartz materials may indicate that contact
was primarily with other Noongar rather than desert groups from the east, who had
more regular access to exotic materials (Figure 3.4).
14.4.2 Occupation Intensity Over Time
Unlike Anderson Rocks, most deposits at Mulka's Cave were dated. Nevertheless,
sample sizes were often too small to properly evaluate changing intensity over time.
The Camping Area yielded the largest assemblages and, as the main focus of
occupation, should preserve the best evidence of intensity shifts. At Anderson Rocks,
periods of reduced occupation intensity were identified when Individual Provisioning
or Short-term Expedient strategies dominated all deposits dating to a particular
period. At Mulka's Cave, however, Place Provisioning was ubiquitous throughout the
entire occupation period represented by the author's pits (Figure 14.2). While the
lowest deposits in the Column Sample (XU10–11, 6000–4250 cal. BP) only
preserved evidence of Individual Provisioning, contemporaneous deposit from the
Camping Area indicate that Place Provisioning was in use at this time. In this case,
the contrasting strategies relate to how different parts of the site were used rather
than reflecting a large-scale shift in occupation intensity across the entire site
complex.
Only CA1 preserved evidence of increased occupation intensity, in XU3–4, that
dated to 1056–324 cal. BP (Figure 14.2). Artefact discard rates were fairly high, and
both deposits preserved clear evidence of Gearing Up. Negative flake scar and
platform counts were their most varied in these XU, and 50–67% of cores were
bipolar. All of these features indicate longer reduction sequences, which is a fairly
reliable proxy for increased occupation span. Values were consistently higher and
more varied in XU4 than XU3, and the former also preserved the largest number of
artefacts with > 25% of their surface unmodified. Based on these data, site visits
310
were longer and/or more intense between 1056–324 cal. BP, particularly from 1056–
723 cal. BP. Interestingly, this overlaps with the period represented by TH1, where
cultural material began to accumulate around 1120 cal. BP. This area may have
been used as a response to increased intensity, prompting spatial segregation of
some stone-working activities.
At any site, longer and/or more intense visits are only possible when sufficient food
and water resources are available. However, the capture properties and capacity of
the cleft rockhole meant that, even under low rainfall conditions, more than 5700 L of
water was available at the beginning of spring. As a result, the intensity of
occupation was probably limited not by the availability of water, but by food
resources, and the practical distance people can travel to forage and/or transport
items to site. At some stage, depletion of surrounding food plants and animals will
reduce foraging returns and increase transport costs – at this point, it becomes more
profitable to move to a new location (Charnov 1976). Therefore, under average or
high rainfall conditions, people would leave Mulka's Cave before the rockhole dried,
since both water and more abundant food resources could be found elsewhere.
However, under lower rainfall conditions, water would be scarcer and more patchily
distributed across the group's territory, curtailing people's freedom to leave. The
large capacity and ephemerality of the rockhole supply would instead encourage
more intense occupations during the water-holding period. People would remain onsite for as long as possible (and possibly gather in larger numbers), despite lower
foraging returns and increased transport costs. These longer stays would serve as a
risk reduction strategy, by conserving water in other, more long-lived sources, until
later in the season. These sources were already the focus of extended
summer/autumn occupation within the normal seasonal route, so their conservation
would be especially vital under low rainfall conditions. It seems plausible, then, the
period of intense occupation at Mulka's Cave from 1056–324 cal. BP reflects the
restricted availability of water in the wider area, which occurred under low rainfall
conditions.
311
14.5 CONCLUSION
The data collected earlier, in conjunction with the occupation model, allowed several
conclusions to be made about how each site was used over time. Gibb Rock was
visited in the more recent past, probably over the last few hundred years. The site
was not a focus of residential occupation but was probably visited during
spring/summer foraging trips when people were based at a location with reliable
water. At Anderson Rocks, most occupation occurred when the gnammas held water,
probably during spring. Visits were shorter in the earliest period (8200–4571 cal. BP),
probably due to limited water availability, so people were more mobile and depended
heavily on transported toolkits to meet their technological needs. For around the last
four and a half thousand years, visits were longer, indicating that water was more
abundant than during the earlier period. Occupation was particularly intense from
4036–3819 and 3603–2520 cal. BP, probably due to higher spring rainfall extending
the water-holding period of the gnammas. The outcrop itself was used for stoneworking activities, but was also visited during foraging trips, when unexpected tasks
were completed using quartz from an exposed vein.
At Mulka's Cave, Place Provisioning occurred across the entire 6700-year
occupation period represented by the author’s excavations. Visits were therefore
predictable and long enough to warrant stockpiling raw material. This is unsurprising,
as the cleft rockhole provided an abundant source of water, even under low rainfall
conditions. The site was probably occupied in September–October, when water was
becoming scarcer again, before people moved northwards to more optimal spring
foraging grounds. Individual Provisioning was used throughout most of the
occupation period, and may represent out of season visits, when capped gnammas
would provide a reasonable supply of water, or people who travelled to the site for
ritual or social gatherings. The Main Cave Area may have been the focus of those
ceremonies, if indeed they occurred, but domestic activities were concentrated in the
Camping Area. Occupation was more intense from 1056–324 cal. BP, when Mulka's
Cave may have acted as a refuge, due to limited water availability elsewhere.
312
CHAPTER 15
CONCLUSION: OCCUPATION HISTORY OF THE STUDY AREA
15.1 INTRODUCTION
This chapter ties together the different threads of this research to address the
primary aim – to determine when and how people used the study area – and places
these results in their broader context. Each of the four research questions are
addressed in turn. The first focuses on the theoretical occupation model that was
devised based on critical resource distribution and Optimal Foraging Theory, while
the second defines the occupation period represented by the three excavated
sites. The third considers mobility and occupation patterns revealed by the
archaeological record, and how these fit within the theoretical model. The final
research question quantifies temporal shifts in occupation intensity and relates them
to water availability. Discussion then turns to the future research directions, before
the main findings of the study are summarised, and its significance evaluated.
15.2 ADDRESSING THE RESEARCH QUESTIONS
This research sought to answer one broad question: 'when and how did Indigenous
people use the study area, and did the nature and/or intensity change over time?'
This aim was addressed with reference to four specific research questions:
1. How were food and water distributed across the study area and, following
Optimal Foraging Theory, how would seasonal and interannual resource
availability influence human occupation patterns?
2. What is the timescale of Aboriginal occupation in the study area, and how
does this compare to other parts of Australia?
3. What do the archaeological remains reveal about mobility and occupation
patterns within the study area, and how do these results fit within the
theoretical model?
4. How did occupation intensity vary over time, and can this variation be
explained with reference to long-term shifts in water availability?
313
Each question is addressed separately below, with reference to the methods, source
data and wider context and significance, as appropriate.
15.2.1 Theoretical Resource and Occupation Model
This section addresses the first research question: ‘how were food and water
distributed across the study area and, following Optimal Foraging Theory, how would
seasonal and interannual resource availability influence human occupation patterns?’
To begin, the spatio-temporal distribution of food and water supplies are outlined for
the study area. Then, the theoretical occupation model is discussed, focussing on
the seasonal route and patterns of aggregation that would be permitted by the
availability of critical resources. Finally, the results of the occupation model are
placed in the context of similar data from southwestern Australia and other parts of
the state.
Distribution of plant and animal foods
Plant foods were divided into seven categories, ranked by their return rates: gum
and
lerps/manna
(high);
flowers,
fruit
and
storage
organs
(moderate);
leaves/roots/stems and seeds (low). A total of 674 plant foods from 467 species
were identified. Flowers were most numerous (n=252), but seeds (n=119) and gum
(n=101) were also fairly common. Plant foods were most abundant in summer, which
was the only season when all categories were available, including the two highly
ranked ones; all storage organs were available, including those that consume their
organ to reshoot after a period of dormancy, and fruit diversity also peaked. Food
was scarcest in autumn, when highly ranked items were no longer available;
moderately ranked foods were limited to a few perennial storage organs, flowers and
fruits. Lower ranked items would have been more important at this time –
leaves/roots/stems were available year-round, as were seeds from Hakea (n=30)
and Allocasuarina (n=9). Plant foods became more abundant through winter to
spring, as more species began to produce fruit, flowers and seeds; the entire suite of
storage organs was again available by late spring.
Overall, plant foods were generally the most diverse in Heath and Mallee. Saline
yielded the most limited variety, but leaves/roots/stems were most diverse in this
Division; storage organs were concentrated at Granite. However, species diversity is
314
not always indicative of the quantity of a particular food that may be available. For
example, Thicket formations have fewer species than most other areas, but the three
dominant genera all provide food items: Acacia provide edible seeds and gum,
Allocasuarina have edible seeds that are available year-round, and Melaleuca yield
edible flowers – these foods would be available in good supply. Similarly,
lerps/manna derive from Eucalyptus, so would be abundant in Woodland and Mallee
areas, where that genus dominates.
Animal foods were categorised as follows: amphibians, birds, bird eggs, mammals,
reptiles and reptile eggs; the latter only included eggs that were laid (rather than held
inside the body), as they could be encountered independent of the animal. 142
animal species were identified, dominated by birds (n=89). Eggs were available from
85 birds and 23 reptiles. Temporal variation in animal foods was fairly subdued:
amphibians, mammals and reptiles were present year-round, but their ease of
capture and bodily condition varied. Female amphibians may have held eggs, mostly
from March–June, while macropods (and possibly other herbivores) were in peak
condition from June–September. Reptiles would have been difficult to catch in winter,
when they are dormant, but more readily encountered when active during spring and
summer; their eggs were available from September–December. Some migratory bird
species visited the study area for short periods, but most birds were non-migratory.
As a result, the minimum and maximum counts for a single Landscape rarely differed
by more than ten species. Bird eggs were mostly available from August–October,
and few were present in February–June. Overall, fauna was most diverse in
Woodland and Mallee. Saline consistently yielded the least diverse assemblage,
which was dominated by birds and their eggs. However, these would be available in
good supply during winter and early spring, when migratory waterbirds travelled to
the study area to breed. Large mammals were probably rare in Heath and Thicket
areas due to the closely spaced vegetation.
Distribution of water
Generally, soil water was short-lived, and its presence centred on periods of
sustained rainfall. Soils stored more water, for longer periods, after stronger winter
rains. However, the distribution of falls was also important. If soils were dry for too
long, the rainfall was often insufficient to recharge the profile – in those cases, Plant
315
Available Water (PAW) did not accumulate. Under low winter rainfall conditions,
PAW was present in most soils for a few days to a few weeks. In contrast, Mallee
soils never accumulated water, and Granite profiles held PAW until late winter, due
to the additional input through runoff. Under average rainfall conditions, most soils
retained PAW throughout most of winter; Mallee only held water intermittently and,
along with Saline soils, was less likely to rewet once dry. Granite soils retained water
until late October, and the 100 mm storage capacity was reached on several
occasions. When winter rainfall was high, most soils held water in greater quantities
and for a longer period, until early September; Granite soils held water until early
November. Overall, Granite areas provided the best prospects for soil water, due to
the influence of runoff shed from outcrops. The profile was frequently constrained by
the upper storage limit, so benefitted from smaller falls spread over a long period,
when capacity was less of a hindrance. Dry Granite profiles could accumulate PAW
from isolated falls > 25 mm but would hold water for less than a week unless the falls
were > 35 mm; such falls were comparatively rare (approx. 15% of all falls ≥ 15 mm).
Heath, Thicket and Woodland soils held water for similar periods, while Mallee was
the poorest performing profile – it only held PAW consistently under high rainfall
conditions.
Rock structures (pans and gnammas) were confined to the Granite Landscape
Division. As they are closed systems, rock structures only lose water only by
evaporation, so there was a simple relationship between the annual rainfall total and
the number of days the structures held water. Falls were more effective in winter,
when evaporation was lower, but water could accumulate relatively easily at other
times of year. Pans held water for the shortest period, generally only during winter or
other periods of sustained rainfall; water was only intermittently present when winter
rainfall was low. Runoff pans fared slightly better, but still dried sooner than
gnammas due to the storage limit imposed by their shallow depth. Nevertheless,
they held water fairly consistently throughout winter, regardless of rainfall quantity.
Both types of gnamma performed better than pans, especially under high rainfall
conditions, as water could be stored over a greater depth and was lost more slowly.
Runoff gnammas consistently outperformed gnammas and were only dry for 2–4
months of a year. Due to the larger water balances accumulated, they held water
during long, rainless periods when other structures dried.
316
Due to the presence of rock structures, Granite was the only Landscape Division that
regularly preserved water beyond early spring, even under high rainfall conditions.
Structures would accumulate water after small, isolated falls, but this would generally
last less than a week (Table 7.3; Figure 7.8). None of the structures held water
beyond February, but this probably reflects the model limitations. Runoff gnammas
were modelled fairly conservatively, by assuming they received twice the amount of
rainfall. However, gnammas situated at the base of a slope, or at the end of a long
inflow channel (high-input gnammas), would receive considerably more input. If
these structures had the capacity to store this input, they could hold water for much
longer periods.
Occupation model
Under OFT, higher ranked items should be pursued first, but lower ranked items will
be incorporated when those are unavailable or limited in supply (Macarthur and
Pianka 1966). Foraging patches can be similarly ranked, and those providing greater
returns will be utilised first; where possible, people will move to a new foraging patch
when resource depletion means that returns would be greater in a new patch
(Charnov 1976; Macarthur and Pianka 1966). Finally, distance from camp dictates
what resources can be profitably transported – higher ranked items can be
transported greater distances before costs become prohibitive, while lower ranked
items can only be moved over shorter distances (Orians and Pearson 1979).
Considering these rules and the resource distribution data, the seasonal route below
would be optimal under average rainfall conditions; variations under lower or higher
rainfall are noted, where relevant.
Summer: Larger groups would be tethered to Granite areas with capped
runoff gnammas and/or high-input gnammas; group size would be limited
by the volume of water available. Restricted residential mobility and
predictable occupation location would favour Place Provisioning. Highranked foods were most abundant in Woodland, Mallee and Thicket areas,
but people would also need access to a diversity of moderate- and lowranked plant resources and animals, since residential mobility was impeded.
The optimal foraging patches would be in the northern study area, which is
dominated by Thicket, and Mallee that preserves small pockets of
317
Woodland (Figure 3.14). Mallee and Woodland areas would have provided
an abundance of fauna, and storage organs were readily available at
Granite. Acacia seeds are low-ranked, despite their excellent nutritional
content, but were abundant in Thicket. These could be profitably
transported over short distances, so could be incorporated into the diet as
required.
After high winter rainfall (around one in every ten years), aggregation
events could occur, ideally in December in the southern palaeovalley. Food
would have been relatively plentiful, albeit less so than in other parts of the
study area; water would also be available if gnammas had been capped
after rains ceased. Importantly, increased consumption of critical resources
would not affect normal summer occupation, since that occurred further
north.
Autumn: Water availability was still limited, probably restricted to highinput gnammas at particular Granite areas; Place Provisioning would still
be the optimal technological system. Plant foods were far less abundant,
and no highly ranked items were available; people would therefore need to
broaden their diet and include more low-ranked resources (Macarthur and
Pianka 1966). The central study area, where Thicket and Heath meet,
would be optimal. Foraging returns would be highest in Heath and Mallee
areas, where moderately ranked foods were most diverse, and Granite also
preserved an abundance of storage organs. Fauna from Mallee and
Woodland areas would be particularly vital during autumn, due to the more
limited suite of plant foods available. If required, Allocasuarina seeds could
be collected from nearby Thicket areas, but only if campsites were within
fairly short distances. By mid-May, groups could begin to move to different
Granite outcrops, as most rock structures begin to accumulate water from
early rains.
Winter: From late June, soil water was available in all Landscape Divisions,
albeit intermittently in Mallee. Mallee was the only area that yielded diverse
plant and animal foods during winter; plant foods were abundant in Heath,
and fauna in Woodland, while migratory waterbirds were present in Saline
318
areas. The southern palaeovalley would be the optimal residential location,
as it would provide access to all of these Landscape Divisions. As a
response to widespread ephemeral water, people would disperse across
the area in smaller groups, visiting areas that were inaccessible during the
drier months. The increased mobility and unpredictable occupation location
would favour Individual Provisioning, when technological needs are met by
the use of a transported toolkit. However, by late winter, water was only
available at Granite areas, as all other soils had dried. Under low rainfall
conditions, this shift to Granite would have occurred sooner, while it could
be delayed for several weeks under high rainfall conditions.
Spring: By spring, water was only available in Granite areas, in soils and
rock structures. Individual Provisioning would dominate in most cases, but
some Place Provisioning may occur around larger rock structures where
longer, more predictable visits occurred. Plant foods were more abundant,
and reptiles were active after their winter dormancy. Heath, Mallee and
Woodland areas provided the greatest diversity of plant/and or animal
foods; many waterbirds were still present in early spring, but would
probably be more common in the larger salt lakes in the eastern portion of
the study area. Therefore, the southern palaeovalley would still represent
the optimal foraging location. By November, most migratory waterbirds
would have departed, so there would be fewer benefits associated with the
southern part of the study area. People probably moved towards the central
study area in preparation for a move further northward in summer.
Under low rainfall conditions, small groups may have remained in the
palaeovalley and spread out to utilise small supplies of water from capped
gnammas, to reduce the strain on water sources in the central and northern
study area. Conversely, if food resources were particularly abundant after
strong winter rains, it may have been optimal for some groups to delay the
northward move since, by summer, occupation would again be tethered to
those few locations with reliable water.
319
Broader context and significance
Anderson (1984) and Bird (1985) both reconstructed seasonal routes for different
parts of southwestern Australia. Anderson (1984) believed that one group occupied
the Swan Coastal Plain, and another the inland Darling Plateau. Reliable water and
abundant aquatic resources allowed the Plain group to congregate on the coast in
summer and autumn. In winter, they broke into smaller groups and dispersed along
the plain and the forested area to the east; by late spring, people began to move
back to the coast. She argued that inland groups had an irregular route and, since
resources were less abundant, that these people were more mobile, and ranged
over larger areas (Anderson 1984:34, 37). Bird (1985) studied a transect on the
southern coast of Western Australia, inland to Lake Grace. She believed that water
would dictate seasonal patterns of aggregation and dispersal. From February–May,
people would have clustered in large groups around permanent, reliable water. From
June–September, groups would fragment and spread out to use more ephemeral
water. In October–January, intermediate water sources would be used, and camp
sizes would vary according to the capacity of these sources (Bird 1985:176).
The study area seasonal route is similar in several ways, but it occurred over a much
smaller region than those considered by Anderson (1984) and Bird (1985). In each
case, people dispersed during winter, as a response to less abundant food
resources and more widely available water. There was some level of congregation in
summer, but water sources in the study area probably could not support groups as
large as those around more reliable, permanent water sources elsewhere in the
southwest. Finally, the predictability of resources permitted a regular seasonal route,
in contrast to the erratic occupation patterns that occurred in some parts of the
Western Desert (Gould 1969b, 1991; Tonkinson 1978:29).
In some parts of Western Australia, reliable water was available during seasons of
food abundance, allowing different groups to gather for ceremonies or social events
(Cane 1987; O'Connor and Prober 2010; Veth 1987). In contrast, unpredictable
rainfall in the southern Western Desert meant that groups did not have regular
seasonal routes, but instead followed the rain. Food could not be stockpiled for
events, since it could be many years before groups were able to return to a particular
location. Instead, events were held after strong rains, when water, plant and animal
320
foods were all readily available. As a result, these aggregations were held very
infrequently (Gould 1968,1969b, 1984, 1991). The same was true in the study area,
where aggregations would best be held the summer after strong winter rains, which
occur around once every ten years. Water could be preserved in capped gnammas,
to accompany the bountiful food supplies available in summer. Since summer
gatherings were predicated on winter rains, and a location having the required
structures to preserve this water, the location and timing of events was more
predictable than in the southern Western Desert. As a result, some food could be
stockpiled, including lower ranked resources that might be collected via embedded
procurement, in a manner akin to flakeable stone.
15.2.2 Timescale of Aboriginal Occupation
This section addresses the second research question: 'what is the timescale of
Aboriginal occupation in the study area, and how does this compare to other parts of
Australia?' This question is addressed in two parts, first by defining the occupation
period, by identifying the oldest and most recent cultural material, with reference to
the archaeological investigations at Anderson Rocks, Gibb Rock and Mulka's Cave,
as detailed in Chapters 10–12. The second section relates to the wider context of
these results, specifically how they compare to those from other parts of Australia,
whether they can be considered representative of the study area as a whole (i.e.
whether earlier material may be present at other sites), and why these data are
significant.
Timing of occupation
Both dated sites (Anderson Rocks and Mulka's Cave) exhibit evidence of early
Holocene occupation. Mulka's Cave preserved the earliest cultural material in the
basal unit of the pit excavated by Bowdler et al. (1989); the author subsequently
dated charcoal from various units (Rossi 2014a). The depth-age curve indicates that
artefacts began to accumulate around 9650 cal. BP (Figure 12.9), but the earliest
directly dated sample was deposited more than 1500 years later (1988 pit XU16,
Wk-29188: 8166–7933 cal. BP at 95.4% – Chapter 12.5). Most occupation evidence
was found in the Camping Area, and post-dated 6700 cal. BP (Figure 15.1).
Considering the vastly different occupation periods represented by CA1 and CA2,
321
despite similar depths of deposit, it is possible that older cultural material is present
in unsampled deposits of the Camping Area.
Anderson Rocks demonstrated a similar antiquity of human occupation. In pit AR2-3,
the lowest artefact-bearing layer (XU32) dated to around 8200 cal. BP (Figure 10.6).
This early occupation evidence showed that people were meeting their technological
needs through Individual Provisioning and, soon thereafter, Short-term Expedient
flaking of stockpiled material. Stockpiling, or Place Provisioning, was the optimal
solution when people knew they would be at a particular location regularly, and for
long periods. Clearly, then, Anderson Rocks had been visited under such conditions
before 7775 cal. BP. The site may preserve more occupation evidence below the
culturally sterile XU33 or in other parts of the site; these deposits may be
considerably older, or broadly contemporaneous to those from AR2-3 XU32.
Figure 15.1 Occupation periods represented by subsurface archaeological deposits at Anderson
Rocks (dark blue), Gibb Rock (red) and Mulka's Cave (light blue). CS = Column Sample, P88 = 1988
pit excavated by Bowdler et al. (1989). GR1 was not dated but based on the short depth of deposit
the assemblage is probably no more than a few hundred years old. Note the presence of cultural
material throughout the occupation periods, indicating no significant hiatuses in site visitation.
It was also important to place some timeframe on the most recent occupation
periods, particularly to determine whether occupation continued after European
arrival. CA1, TH1, the Column Sample, and AR2-3 all yielded direct dates from 138–
48 cal. BP for XUs with basal depths of 75–150 mm (Tables 10.3, 12.2–12.3). Based
on the deposition rates, much of the upper and surface material probably dates to
the European period. Gibb Rock was not dated, but the assemblages probably date
to within the last few hundred years – most material was found on the present
ground surface, so some overlap with the European period is likely.
322
Rossi (2010) identified ten flaked glass pieces from Mulka's Cave. These constitute
direct evidence that the rockshelter and the slope outside were used after European
arrival. Only one of these glass flakes derived from the undisturbed collections that
were reanalysed herein, so the differences in artefact typology and identification
means the presence of the remaining nine pieces must be interpreted with caution.
The same applies to flaked glass found in surface assemblages elsewhere in the
study area (e.g. Quartermaine 2000 – see Chapter 4.2.1). Nevertheless, the
evidence clearly indicates that occupation occurred after European arrival, and on a
regular basis, contrary to the statements of early settlers (e.g. Meeking 1979; Mouritz
1986).
Broader implications and significance
While the initial colonisation date, route and conditions are still hotly debated, there
is fairly reliable evidence that Aboriginal people had reached northern Australia by
65,000 cal. BP (Clarkson et al. 2017). People had certainly reached southwestern
Australia by around 45,000 cal. BP, possibly as early as 50,000 cal. BP, if the more
contentious dates from Devil's Lair are accepted (Allen and O'Connell 2014; Balme
2014; Pearce and Barbetti 1981; Turney et al. 2001). Most parts of the country –
including the present-day arid zone – have sites pre-dating 40,000 cal. BP (e.g. Allen
and O'Connell 2014; McDonald et al. 2018; Smith 2013:79–80). Clearly, then, if the
three sampled sites can be considered representative, the study area has been
occupied for a small portion of the continent's 65,000-year human history.
The late occupation of the study area possibly comes down to water availability – all >
30,000 BP sites in the arid zone are associated with some sort of reliable water
source (Bird et al. 2016; Smith 2013:78–91). In the study area, freshwater lakes and
rivers are absent, and the groundwater saline/hypersaline. The only source of
potable water was rainfall preserved temporarily in soil profiles and rock structures.
Considering the ephemerality of water under modern rainfall conditions, it is unlikely
that widespread occupation could occur during the drier conditions that characterised
the last glacial period, beginning around 32 ka and terminating at the early Holocene
(Petherick et al. 2013; Reeves et al. 2013). Therefore, while the present sample is
limited, it seems unlikely that any Pleistocene sites or cultural deposits would be
found in the study area, or the inland southwest. If present, they would probably
323
indicate short-lived incursions into the area, when climatic conditions permitted,
rather than more regular, permanent use of the area.
Despite the relative recency, the early Holocene dates from Anderson Rocks and
Mulka's Cave are significant in the context of the inland southwest. As noted earlier
(see Chapter 1.2), the area is vastly underrepresented in the national database of
dated archaeological sites compiled by Williams and Smith (2012, 2013) and
Williams et al. (2008, 2014). Prior to this research, Mulka's Cave was the only site in
the area that provided radiocarbon determinations in association with cultural
material. However, during their geomorphological analysis at Bennett Lake, Suzuki
et al. (1982:70) claimed to have found artefacts in a unit they believe was deposited
between 7500 and 4500 BP. However, when Bird (1985:111–113) visited the site,
she was unable to locate any artefacts in undisturbed deposits. Based on
geomorphological evidence, Bird (1985:115) believed that the surface scatters from
her Inland Salt Lake Zone were exposed by the erosion of mid- to late Holocene
units, so all assemblages post-dated 5000 BP. Mulka's Cave and Anderson Rocks,
then, represent the first reliable evidence for early Holocene occupation of inland
southwestern Australia, and provide clear evidence of its ongoing use until the
European period. The Mulka's Cave data certainly refute the claims by Bowdler et al.
(1989) that the site was only visited over the last few hundred years.
15.2.3 Mobility and Occupation Patterns
This section addresses the third research question: ‘what do the archaeological
remains reveal about mobility and occupation patterns within the study area, and
how do these results fit within the theoretical model?’ This discussion is divided into
three sections. The first comprises site histories for Gibb Rock, Anderson Rocks and
Mulka's Cave, focussing on how Kuhn's (1995) technological provisioning systems
indicate different levels of residential and logistical mobility. Then, the occupation
model is evaluated against the results from the three excavated sites as well as data
from other archaeological locations in the study area. The final section summarises
the general mobility and occupation patterns from the study area and addresses the
broader relevance of various conclusions.
324
Site histories
At Gibb Rock, poor-quality local quartz was reduced in a Short-term Expedient
manner; the resultant artefacts were used to supplement the transported toolkit to
complete unanticipated tasks. Reduction intensity was low, and the assemblage was
small, less complex, and lacked evidence of toolkit maintenance. This indicates that
site visits were short, and no down-time was available (Clarkson 2007:15; Torrence
1983); Gibb Rock was therefore not a residential location. Instead, it was probably
visited during foraging trips, while people were based at a nearby site. Since Gibb
Rock was not a residential site, it is not possible to consider the level of mobility of
the groups who visited.
At Anderson Rocks, some cultural material was present on the outcrop, but most
was concentrated around the gnammas that acted as a tether and led people to
reoccupy the same area over more than 8000 years. These residential visits
probably occurred during spring/early summer, when the gnammas held water that
was not widely available elsewhere. Mobility varied over time – the earliest period
began at unknown point in the past when visits became long and predictable enough
to prompt stockpiling of raw material. From 8200–4571 cal. BP, Individual
Provisioning and Short-term Expedient reduction of this stockpiled material indicates
that visits were shorter. From 4571 cal. BP, visits became longer and more
predictable, evident from more intensive reduction of stockpiled material,
characteristic of Place Provisioning; this may indicate an increase in the gnammas
water-holding period. Gearing Up assemblages indicate some increase in logistical
mobility during the later period, at 4571–3819 cal. BP, 3170–1870 cal. BP, and
during the recent past. Longer visits may have been more common at these times,
requiring logistical forays to procure more distant resources. The coexistence of
Place and Individual Provisioning assemblages, particularly over the last 1200 years,
may indicate out of season visits (probably after summer/autumn isolated falls)
and/or short-term shifts between contrasting strategies.
Some satellite stone-working activities occurred atop the outcrop at Anderson Rocks.
While these deposits were not dated, they likely accumulated over the last few
thousand years; the longer occupation span during that period may have prompted
some spatial segregation of activities. A small surface scatter indicates that the site
325
was also visited during foraging trips, when local quartz was used to manufacture
expedient artefacts. The task must have been very short and easily accomplished,
considering raw material stockpiles were available elsewhere on-site.
The earliest occupation evidence from Mulka’s Cave derives from the rockshelter,
where cultural material began accumulating around 9650 cal. BP. However, all
available assemblages from the rockshelter post-date 6000 cal. BP. These are
dominated by Individual Provisioning and Short-term Expedient strategies, despite
the presence of Place Provisioning in contemporaneous deposits in other parts of
the site. Longer visits occurred from 4250–1570 cal. BP; some may have been linked
with rock art creation, which Gunn (2004, 2006) estimated began around 3000 years
ago. The rockshelter was used less intensively than other parts of Mulka's Cave, but
it was not possible to determine if it had any ritual significance as claimed by Goode
(2011) and Macintyre et al. (1992). However, the Main Cave Area assemblages do
not support Gunn's (2006) idea that the rockshelter and slope were the main focus of
occupation, but the present analysis excluded the disturbed deposits that yielded
most of the cultural material from that area. While the 420 BP date obtained by
Bowdler et al. (1989) was clearly erroneous, their idea of intermittent visitation is
broadly representative of much of the occupation period.
The Camping Area was the focus of residential occupation, from 6700 cal. BP until
the recent past. This indicates some continuity from the distant to the more recent
past, when Traditional Owners camped in this same area when brought to site as
children. Place Provisioning occurred throughout the deposits, so visits were long
and predictable enough for raw material stockpiling to consistently represent the
optimal technological strategy. This predictability, combined with the proximity of the
Camping Area to the gnammas, indicates that visits were tied to the water-holding
period of the rock structures. The cleft rockhole provided an abundant but ephemeral
source of water under all rainfall conditions and may have been particularly valuable
from mid–late spring when water was less widely available. Individual Provisioning
occurred throughout most of the deposit – considering the reliability of the water
sources, it probably represents out of season visits. Gearing Up assemblages all
post-date 3000 cal. BP, indicating increased logistical mobility thereafter.
326
From 1120–1056 cal. BP three interesting patterns emerged, and persisted until the
recent past: Gearing Up assemblages were more common, Individual Provisioning
was used less frequently, and the Humps outcrop was first used as a satellite stoneworking area (Figure 15.1). This was probably a combined response to increased
occupation intensity from 1056–324 cal. BP. Increased intensity would require more
regular logistical forays to offset resource depletion, and out of season visits may be
discouraged by the already lengthy period of tethered occupation. Finally, more
intense visits may have required some spatial organisation, prompting an area to be
dedicated to particular stone-working activities.
Evaluating the occupation model
While the sample is somewhat limited, is possible to evaluate the model against
archaeological results from this research, and those of others. Briefly, the model
summarised above (see Chapter 15.2.1) predicted that Place Provisioning would
dominate in the central and northern study area, around reliable water sources. In
the southern area, Individual Provisioning would be more frequently used, since
residential mobility was greater in winter and part of spring, but some Place
Provisioning might occur around certain rock structures that were the focus of
occupation in mid- to late spring. Superficially, the studied sites fit this pattern.
Realistically, however, the site histories summarised above are sensical without
recourse to the model, based purely on the availability of water. Occupation was
longer and more predictable in locations with more reliable water sources (Anderson
Rocks, Mulka's Cave), while the site without water (Gibb Rock) was not used as a
residential location. However, archaeological remains from the rest of study area
permit the model to be further evaluated.
Of the 17 other archaeological locations in the study area, 16 are in the southern
palaeovalley. This may indicate widespread winter dispersal predicted by the model,
but it is important to note that most were identified during cultural heritage surveys
that accompanied development work, which is understandably more common near
the Hyden township. It is difficult to place these archaeological locations in context,
as none have been excavated or dated, and stone artefact data are very limited.
Where available, these data can be difficult to integrate due to methodological
327
differences. As a result, discussion is limited to those sites that can be evaluated in
context of the model or archaeological data collected during this research.
Winmar (1996:28) claimed that Wave Rock was a place where various Noongar
groups gathered. It is difficult to evaluate this idea based on the scant archaeological
data, but some elements from the site complex may support the notion: the scarred
trees may have denoted the special nature of the site; the single grindstone may
indicate that low-ranked seeds were used to support larger groups of people;
gnammas were present; and the site is in the southern part of the study area that the
model identified as optimal. However, in the absence of limited artefact data, no
concrete conclusions can be made.
Four sites preserved gnammas: Wave Rock Rockholes (21386), Hyden (5844),
Graham Rock (5610) and Twine Reserve (5121). The latter is in the central study
area, while the remainder are in the southern palaeovalley. No artefacts were found
at Graham Rock or the Wave Rock Rockholes, and artefacts were scarce in the
wider area surrounding the latter (see Chapter 4.2.1). A few quartz fragments were
present at Hyden (5844), which also preserved a gnamma > 1.5 m deep. While the
volume of water captured would depend on the amount of runoff received, this
source would preserve water for long periods, due to its depth. The paucity of
cultural material at the southern gnamma sites supports the idea that this part of the
study area was primarily used in winter, when surface water was more abundant. In
contrast, Twine Reserve preserved extensive surface artefact scatters made on nonlocal quartz – probably indicative of Place Provisioning – and likely retains
subsurface archaeological deposits. Timms (2013) noted seven gnammas at the site;
the largest three range from 0.8–1.95 m in depth and hold almost 10,000 L of water
between them. Based on the author's observations, these gnammas would collect
runoff from a considerable expanse of rock. Twine Reserve would have been an
ideal spot to camp during the drier months, possibly explaining the increased
archaeological visibility; this also supports the model's predictions.
Staub and Taylor (1999) investigated four salt lake sites: Kondinin Pipeline (KP)-1,
KP-2, KP-3 and KP-5. Most artefacts were found around the lake edges, indicating
these areas were primarily visited when the lakes were full, as predicted by the
model. However, the presence of some artefacts on the lakebeds indicates some
328
limited use at other times of year. Stone artefact data are limited, but certain
assemblage characteristics are of interpretive value. The assemblages were
dominated by 'debris', which are probably analogous to the author's flaked fragments.
In KP-5, the largest artefact sample studied, 59% of quartz artefacts were debris;
similarly high values were present in most other samples. In AR2-3, fragments
comprised 66.3% of the stockpile-quality assemblage, and only 33.7% of artefacts
made on higher quality material. It seems likely, then, that most of the KP fragments
were made on lower quality raw material. As a result, either Place Provisioning or a
Short-term Expedient strategy must have been in use.
Artefact totals are not available for any sites, but sample sizes indicate > 164
artefacts were found at KP-5, and > 390 at KP-1; artefact density was often greater
than in the Camping Area clusters, and total artefact numbers were higher than in
the entire Camping Area Scatter. However, it is unlikely the KP assemblages
represent Place Provisioning, due to the infrequency of cores (0–3.7% compared to
5% in AR2-3 XU12–18), their uniformly small size (< 30 mm in maximum dimension),
and the use of flakes as cores. This level of reduction and recycling would not be
necessary under Place Provisioning or Short-term Expediency, since raw material
would be readily available from a stockpiled or local source. Instead, the
assemblages may represent a more casual form of stockpiling, whereby a limited
quantity of material was accumulated during a particular visit; this material was used
during subsequent visits, and discarded flakes/cores were flaked again rather than
more material being collected. Therefore, as predicted by the model, occupations
were probably not long or predictable enough to warrant regular stockpiling of raw
material. The presence of a raw material source, in the form of previously discarded
items, may have acted as a tether for future visits, so may explain why artefacts
were clustered in particular areas along the shorelines. However, this increased
archaeological visibility may also indicate that people sometimes gathered in greater
numbers than predicted by the model, to exploit the seasonal abundance provided
by waterbirds.
Implications for mobility and occupation patterns throughout the study area
In the study area resources were predictable, but not especially abundant
(particularly water), so occupants probably lived in small groups as in other inland
329
areas (Hammond 1933:21, 23; Roe 1836; Smith 1993:287–288). It is not possible to
evaluate territory size, but it would have been larger than in more resource-rich
areas, and some boundaries may have been semi-permeable (Dyson-Hudson and
Smith 1978; Harpending and Davis 1977). Overall, mobility was dictated by the
availability of water – winter abundance allowed people to spread out, while the drier
months tethered occupation to the few more reliable water sources; the length of
occupation at individual sites shifted according to the availability of water on-site and
elsewhere in the landscape. This created a system characterised by both
scheduled/predictable and unscheduled/unpredictable moves. The former occurred
when people travelled to a known water source, while unscheduled moves may
occur during winter, or if large isolated falls permitted people to temporarily leave
their summer/autumn camps. As a result, the same sites could be visited at several
different times of year, and possibly for different reasons – while Mulka's Cave may
have had a ceremonial function, general occupation clearly occurred there also.
Certainly, the entire occupation and mobility system was more predictable than in
parts of the Western Desert, where erratic and localised rainfall precluded normal
seasonal routes (Gould 1969b, 1991; Tonkinson 1978:29). This regular, consistent
use of the study area is in stark contrast to statements by early settlers, who claimed
that people simply passed through on their way elsewhere (Meeking 1979; Mouritz
1986), or Landor’s and Lefroy's (1843) Aboriginal guides, who claimed that food and
water were not available. The latter probably reflects a risk avoidance strategy, as
the guides were not familiar with the distribution of food and water resources in the
study area.
While the bulk of mobility data derived from technological provisioning systems, raw
material types also provided some insights. High-quality exotic raw materials (BIF,
chalcedony) were only found at Mulka's Cave and Anderson Rocks; these materials
probably derived from the greenstone belt beyond the eastern boundary of the study
area, approximately 80–85 km from the sites (Figure 3.4). These materials were rare,
but were more commonly found in deposits that also preserved evidence of Gearing
Up (n=12/22 in AR2-3; n=3/8 in CA1, n=3/3 in CA2 – Appendices E and G).
Considering the distance from source, it is unlikely that these exotic materials were
collected during logistical forays to procure food resources. There are two possible
explanations: first, that tools made on high-quality exotics formed part of the
330
permanently retained toolkit, and that maintenance activities occurred before a
standard logistical foray along with the manufacture of other required artefacts on
high-quality quartz. Alternatively, some quartz Gearing Up assemblages may reflect
preparations for long travels for other reasons, perhaps social gatherings with people
from further east. Small quantities of exotic materials, or artefacts, may have been
transported back to the base camp and subsequently maintained.
Regardless of how high-quality exotics fit into the broader technological scheme, and
the extent of logistical mobility implied, the presence of these materials indicate that
study area occupants had some access to the greenstone belt itself, or the people
who lived there. The limited quantity of material means this access or contact was
infrequent. It seems unlikely, then, that the greenstone belt was part of the study
area group's territory or normal seasonal route. Contact with these people likely
increased after 4000 cal. BP, when most of the exotic artefacts were discarded. If
Tindale's (1974) cultural boundary is accurate, this may explain the general paucity
of the high-quality exotics. While Noongar and desert groups had some social
contact, it was infrequent (Macintyre et al. 1992). If this reasoning is correct, it
indicates that the boundary separating Noongar from the desert groups had been
located west of the greenstone belt, in the approximate position mapped by Tindale
(1974), for the entire occupation period represented by the Anderson Rocks and
Mulka's Cave deposits. This is in direct contrast to Gibbs and Veth's (2002) assertion
that the boundary attained its present position only after European arrival arrested its
westward progression. However, it is a very small sample of artefacts on which to
base such an argument, which also relies on the exotic material having originated
from the greenstone belt (or another area east of the study area), rather than an
alternate source within Noongar territory. Nevertheless, if this theory is correct, and
gatherings did occur within the study area, these must have been primarily with other
Noongar groups, or people from areas where quartz was the dominant raw material
used for artefact manufacture.
15.2.4 Occupation Intensity and Water Availability
This section addresses the final research question: ‘how did occupation intensity
vary over time, and can this variation be explained with reference to long-term shifts
in water availability?’ The discussion begins with a consideration of the shifts in
331
intensity evident at each site, and how they might be dictated by water availability.
While it was not the aim of this research, the collected data allowed Rossi's (2014b)
earlier idea to be evaluated, namely the use of occupation intensity as a proxy for
effective rainfall. These intensity data were used to construct an effective rainfall
record for each site, which is then compared and refined into a single record for the
study area and evaluated against the palaeoenvironmental record.
Different patterns of occupation intensity were evident at each of the three studied
sites based on the use of Individual or Place Provisioning and the presence of highintensity indicators (e.g. post-depositional flake breakage, longer reduction
sequences). Intensity was consistently low at Gibb Rock, where Short-term
Expedient strategies were ubiquitous. The deposits were not radiocarbon dated, but
the assemblages probably date to the last few hundred years. Water was only
available during winter, at which time Individual Provisioning was the optimal solution
to permit the greater mobility required for people to access areas that could not be
visited during the drier months. Therefore, it is unlikely any different pattern of
intensity would be evident at Gibb Rock, even if older occupation deposits were
found.
At Anderson Rocks, low intensity occupation occurred from 8200–4571 cal. BP,
when Individual Provisioning dominated. Thereafter, intensity was moderate, with
bursts of higher intensity occupation lasting from a few hundred (4036–3819 cal. BP)
to over 1000 years (3603–2520 cal. BP). A brief low-intensity period occurred from
138 cal. BP, but considering the abundance of Place Provisioning assemblages on
the surface, this period probably only lasted around 75 years or so. The gnammas
could only store < 600 L of water, so people would have stayed as long as was
possible to reduce the length of time spent at their normal summer refuges, which
were the focus of occupation for much longer periods. Therefore, the extreme
variation in occupation intensity was linked to the natural water-holding period of the
gnammas, which dictated the quantity of water available by the time people arrived
on-site. Considering this, occupation intensity was linked to rainfall quantity as below:
o
Low intensity = low rainfall;
o
Moderate intensity = average or high rainfall;
o
High intensity = average or high rainfall, strong spring falls.
332
At Mulka's Cave, moderate occupation intensity dominated throughout the 6700-year
occupation period represented by the author's excavations; this was probably due to
the reliability of the cleft rockhole water supply. One period of higher intensity was
identified, from 1056–324 cal. BP. The rockhole never held < 5000 L by spring, so
food resources would invariably be exhausted while it still held water. It would be
suboptimal to remain on-site thereafter, unless water was not widely available
elsewhere – in that case, Mulka's Cave could be used as a refuge. Therefore,
occupation intensity at Mulka's Cave responded to the availability of water in the
wider area. This can be linked to effective rainfall as below:
o
Moderate intensity = average or high winter rainfall;
o
High intensity = low winter rainfall.
A rainfall record was produced for each site based on the inferred response of
occupation intensity to varying water availability (Figure 15.2). When these records
are compared, several conflicting periods are evident. From 8200–4571 cal. BP and
138–50 cal. BP, low rainfall was indicated by the Anderson Rocks record, while the
Mulka's Cave data identified these as periods of average–high rainfall. Intensity at
Anderson Rocks was dictated by the availability of water on-site, while Mulka's Cave
occupation responded to how abundant it was in the wider area. Further, the
Anderson Rocks gnammas were more susceptible to drying than other structures,
due to their limited capacity and conical shape. Therefore, rainfall may have been
low enough to preclude long visits to the site, but not so low that Mulka's Cave had
to be used as a refuge. From 1056–324 cal. BP, the Mulka's Cave record indicated
low rainfall, while Anderson Rocks data implied average–high rainfall. At Anderson
Rocks, both Individual and Place Provisioning assemblages date to this period. The
latter were smaller than elsewhere in the pit, while the former were more readily
identified than in most other XU. The dominance of Individual Provisioning probably
indicates that low rainfall conditions prevailed for most of the period, while the small
Place Provisioning assemblages may represent short periods of increased rainfall,
when longer visits could occur. The latter would not be visible at Mulka's Cave, as
such evidence would be numerically overwhelmed and thereby masked by the highintensity signature.
333
Figure 15.2 Occupation intensity (top) for each site and effective rainfall conditions for dated sites
(middle) and the study area as a whole (bottom). Note the Gibb Rock assemblages were presumed to
date to the last few hundred years, and all surface assemblages to the last 50 years.
Considering the data presented above, it is possible to combine the Anderson Rocks
and Mulka's Cave records to produce a higher resolution, more refined rainfall record
for the study area. Nevertheless, due to the limitations inherent in each record, there
are periods where it is not possible to quantify effective rainfall precisely, but merely
narrow the range of possible conditions. Five rainfall conditions could be identified,
by comparing and combining the records:
334
o
Very low: high-intensity indicators at Mulka's Cave;
o
Very low–low: Individual Provisioning dominant at Anderson Rocks, no data
from Mulka's Cave;
o
Low: Individual Provisioning dominant at Anderson Rocks;
o
Low–high: unclear Anderson Rocks signature, no high-intensity signatures
at Mulka's Cave;
o
Moderate–high: moderate intensity at Anderson Rocks and/or Mulka's
Cave.
The combined rainfall record indicates that the mid-Holocene was dominated by
lower rainfall conditions (Figure 15.2). For the next 2700 years, average or high
rainfall conditions persisted, with two periods of probable higher spring falls (4036–
3819 and 3603–2520 cal. BP). During the last 1000 years, drier conditions again
dominated, but rainfall temporarily increased for short periods.
While the above pattern is interesting, it is difficult to integrate into the wider
palaeoenvironmental
record
for
southwestern
Australia,
primarily
due
to
methodological limitations. The author had previously considered how occupation
intensity might be used as a proxy for effective rainfall (Rossi 2014b), but the current
research was not designed to test that theory. At best, it was hoped that the
collected data might allow the idea to be further developed. As a result, the methods
here were not designed to test rainfall variation over time, but instead to consider
how different rainfall conditions (as represented by modern data) might have affected
water availability and human occupation at various sites. There are three particularly
significant points of difference between this research and palaeoenvironmental
studies:
1. Palaeoenvironmental reconstructions generally characterise years as
wetter or drier, while three categories were used herein: low, average and
high. There will invariably be some overlap between the author's 'average'
rainfall years, and those characterised as wetter or drier by others;
2. This study utilised modern rainfall data, as recorded by the Hyden
meteorological station (no. 010568). As such, these data are only
representative of periods of lower, average and higher rainfall that occur
during the current climatic phase. Earlier periods may have had a wetter or
335
drier baseline, so palaeoenvironmental reconstructions consider the
amount of effective rainfall compared to the modern climate. There is a
difference of scale, then, between this research and palaeoenvironmental
studies. The former will reflect smaller scale shifts in effective rainfall, while
the latter are designed to identify larger scale changes;
3. Finally, gnammas are closed systems with no groundwater input. In
contrast, most palaeoenvironmental proxies originate from waterbodies that
interact with the regional groundwater table. As a result, each will respond
to shifts in effective rainfall in different ways, and the results are not
necessarily directly comparable.
Keeping the above in mind, there are some notable similarities and differences
between the rainfall record produced here and the palaeoenvironmental record from
southwestern Australia. Interestingly, the record has little in common with those from
inland southwestern Australia (Harrison 1993; Zheng et al. 2002) that show wetter
conditions until 4000 BP (Figure 3.8); at that time Anderson Rocks record exhibited
lower occupation intensity, indicative of drier conditions. In contrast, the same broad
pattern of drier followed by wetter conditions is evident in three coastal records:
those produced by Kendrick (1977), Semeniuk (1986) and Yassini and Kendrick
(1988). The timing of their wet to dry transitions (4500–2800 BP) are not directly
comparable, as none of their dates were calibrated, but they fall within a similar
period to the 4571 cal. BP date identified here. Similarly, the drier period from 1056–
324 cal. BP fits fairly well with the record produced by Gouramanis et al. (2012), who
identified drier conditions from 1400 cal. BP (Figure 3.8). However, these similarities
should not be overstated. Considering the discordance of the southwestern
palaeoenvironmental record, it would be difficult to produce a pattern that did not in
some way resemble another. Indeed, if Pickett (1997) was correct, and the
southwest has experienced no major Holocene climatic change, then the record
produced for the study area may be capturing more minor fluctuations in effective
rainfall. While the accuracy of the effective rainfall record is inconclusive, the idea
itself shows some potential and warrants further consideration, as it represents a
novel use of occupation intensity data. Such data are often interpreted through a
palaeoenvironmental lens but, to the author's knowledge, have never been used as
a proxy for palaeoenvironmental conditions themselves.
336
15.3 FUTURE RESEARCH DIRECTIONS
This study has identified several areas that warrant future research. These fall into
three main categories – the first two relate to the study area, and inland southwest
as a whole, and involve increasing the body of archaeological data, and more
accurately quantifying water availability. The third is more closely focussed on the
assemblages collected from Anderson Rocks, Gibb Rock and Mulka's Cave, and
outlines how certain methodological changes could maximise the quantity and
reliability of available data.
The primary aim for any future research should be increasing the quantity of
archaeological data available for the study area, and the inland southwest in general.
Rossi (2010, 2014a) had difficulty interpreting the Mulka’s Cave assemblage,
primarily due to the lack of archaeological data from the surrounding area. The Gibb
Rock assemblage typified the value of comparative data – this assemblage
preserved a clear Short-term Expedient signature, and so provided quantifiable
reduction intensity values that could be compared to those from other assemblages.
This allowed the author to accurately distinguish between different provisioning
systems made on lower quality raw material. If more data were available, the same
method could be used to separate traces of Gearing Up from Individual Provisioning.
At present, it is possible to theorise that reduction intensity should be higher under
Individual Provisioning, but the smaller assemblages produced by both strategies
means that these differences cannot be quantified. If high-quality materials were
found at a site that had no reliable water, they would likely be part of an Individual
Provisioning assemblage. While the small sample size may be problematic, it would
provide some point of reference for reduction intensity values. As a result,
excavation and surface collections should focus on sites without reliable water
(probably around granite or salt lakes, due to the preservation bias associated with
agricultural land), as well as those with larger rock structures, particularly in the
central and northern study area. The latter represent the best opportunity to evaluate
the antiquity of occupation in the study area, since the water sources imposed a
strong tethering effect. Considering how the cleft rockhole shaped occupation of
Mulka's Cave, Twine Reserve would be an ideal candidate for investigation, since
the large gnammas represent a reliable water source. A more thorough survey of
337
Anderson Rocks and Gibb Rock would also be worthwhile, since access issues
limited the portions of the site the author could visit.
As demonstrated by this research, rainwater temporarily held in deeper rock
structures was vital to year-round occupation of the study area. Therefore, it is
pertinent to consider ways of more accurately quantifying these sources. Here, daily
water balance was modelled using local rainfall data and an evaporation coefficient
based on the inferred rate of water loss compared to Evaporation Pans and larger
reservoirs. The coefficient is the main source of potential error and would be best
refined by logging the ongoing water balance of a particular gnamma. Ideally, this
gnamma would be deep, and receive no additional input from runoff. The
daily/weekly water level could be measured and considered in conjunction with local
rainfall data to determine the rate of water-loss through evaporation. This would also
indicate how the evaporation rate varied with depth, since water held near the
bottom of the gnamma would evaporate more slowly due to increased protection
from the sun and wind. Ideally, different shaped structures could be studied to test
the individual coefficients applied to those with sloping sides. Similarly, it would be
beneficial to more accurately characterise site-specific water sources, with particular
reference to the level of runoff received – this is dictated by the size of the area
generating runoff, as well as the slope and texture of granite, and the amount of
rainfall received (Fernie 1930); if these data were available, they could be factored
into the model. Accurate data is vital for larger structures like the cleft rockhole at
Mulka's Cave, whose water-holding properties strongly influenced interpretations. By
more accurately characterising these water sources, particularly how they perform
under various rainfall conditions, it may be possible to further develop the idea of
using occupation intensity at particular sites as a proxy for effective rainfall.
Finally, several methodological improvements could be applied to the cultural
material analysed during this research. Ideally, more charcoal samples should be
dated, allowing deposition and discard rates to be refined, and also to establish the
age of certain deposits (e.g. pit AR1). Small flakes (< 10 mm long) dominated most
assemblages, but limited data was collected from these artefacts. If the arbitrary
boundary between small and large flakes was reduced to 5 mm, this would increase
the sample size of larger flakes, for which more detailed attributes were recorded; as
338
a result of this change all flake scars on large flakes would have to be counted,
excluding those produced by retouch or edge damage. The reduced flake length
boundary would also more accurately separate flakes produced by retouch, which
are less likely to measure > 5 mm. If small flakes were classified by raw material
quality, and breakage rates recorded, the extent of retouch could be reliably
determined based on waste products, which are more likely to enter the
archaeological record than the retouched items themselves. The increased stone
artefact data may allow more accurate identification of different provisioning systems,
and aid in distinguishing Individual Provisioning from Gearing Up. Finally, the
material excavated by Bowdler et al. (1989) warrants reanalysis, using the same
methods as in this research, allowing for the changes suggested above. This would
provide a more accurate indication of how the rockshelter at Mulka's Cave was used,
and the increased artefact sample size may allow the question of ceremonial/ritual
occupation to be addressed further.
15.4 CONCLUSION
The main aim of this research was to determine when and how the study area was
used by Aboriginal people. It was demonstrated that the study area was visited as
early as 9650 cal. BP, and that regular occupation occurred from at least 7775 cal.
BP and continued until after European arrival. Predictable resource distribution
allowed year-round use of the study area and permitted a regular seasonal route –
the intensity of occupation at particular sites depended on the water available there
and elsewhere in the landscape. Summer and autumn occupation centred on larger
gnammas in the northern and central study area, respectively. Residential mobility
was limited, due to the restricted availability of water, so Place Provisioning
dominated, and some Gearing Up occurred when stays were longer. People moved
into the southern palaeovalley in winter when more abundant water allowed them to
disperse more widely – this period of freedom was shorter under low rainfall
conditions, and longer after strong winter rainfall. The increased residential mobility
encouraged Individual Provisioning, supplemented by Short-term Expedient
reduction of lower quality materials. Place Provisioning was restricted to Mulka's
Cave, as the large ephemeral source of water held by the cleft rockhole permitted
predictable spring occupation under all rainfall conditions. The source held sufficient
339
water for Mulka's Cave to be used as a refuge after very low winter rainfall, when
water was scarce in the wider area. By late spring, groups moved northward to the
central study area, when the few reliable water sources again limited residential
mobility.
Aggregation events, for ceremonial or social reasons, could have been held in the
southern study area the summer after strong rainfall, when the required resources
would have been available. However, this idea could not be addressed with the
archaeological data. The quartz dominated assemblages indicate that, if these
aggregation events did occur, they primarily involved Noongar groups. The paucity of
high-quality exotics indicates the study area occupants had little access to the
greenstone belt to the east, or the people who resided there; despite the proximity to
desert groups, then, their influence may have been fairly limited.
The occupation history summarised above is particularly significant since little
archaeological or ethnohistorical data were available for the study area, or the inland
southwest in general. The results demonstrate some similarities with parts of the
southwest, where regular seasonal routes involved dispersal during winter, and parts
of the Western Desert, where the lack of reliable water limited the frequency of
aggregation events (Anderson 1984:34, 37; Bird 1985:176; Gould 1969b, 1984,
1991). After the author’s MA research (Rossi 2010), only six dates were available for
inland southwestern Australia, all from Mulka's Cave (Williams and Smith 2013); no
other sites yielded comparative artefact assemblages. At present, Anderson Rocks
and Mulka's Cave yield a combined total of 22 radiocarbon dates, of which 17 were
considered reliable. Artefact assemblages from Anderson Rocks and Gibb Rock both
represent valuable comparative data that allowed Mulka's Cave to be placed in a
regional context. Furthermore, Anderson Rocks had previously not met the criteria of
a registered site under Section 5 of the Aboriginal Heritage Act (1972). As a result of
this study, that ruling may need to be re-evaluated, and Gibb Rock may also be
added to the register; the author will communicate these results to the appropriate
authority. Overall, the results of this research represent a significant improvement in
the number of known, dated archaeological sites. It has demonstrated that the study
area was regularly occupied from the early to mid-Holocene transition and suggests
that Pleistocene sites are unlikely to be found anywhere in the inland southwest, as
340
drier conditions would have limited the water held by ephemeral sources, which were
vital to human occupation.
While the value of archaeological and theoretical data for the inland southwest
should not be underestimated, much of the significance of this research is in its
approach, and apparent success thereof. This approach required compilation of a
theoretical occupation model based on the distribution of critical resources, which
was then merged with archaeological data. While opportunities to test the model
were limited, its success is evident from the lack of conflict between model
predictions and archaeological data. Particularly strong support was found in the
southern study area, where most gnammas were associated with limited
archaeological content – the model predicted that this part of the study area would
be mostly used in winter and early spring, when water was more widely available.
The model is a useful comparison to archaeological data, but is also an interpretive
tool in its own right. It can be used to place specific sites in a wider context, based on
the Landscape Division and part of the study area in which they occur. It also
identifies the areas with archaeological potential for longer occupation sequences,
due to the tethering effect of more reliable water sources in the central and northern
study area. These inferences could be used to target particular locations for
excavation, to evaluate/confirm the antiquity of occupation in the study area and
provide valuable comparative stone artefact data.
This research has highlighted the importance of considering spatio-temporal
variation in the distribution of critical resources, and how this high-resolution data
can permit detailed occupation models to be formulated. It has confirmed the value
of such models, particularly in areas where comparative data are rare, and indicated
how they can be integrated into an archaeological analysis, with neither archaeology
nor theory taking precedence over the other. This study has devised novel ways to
quantify the longevity of rainwater temporarily stored in soil profiles and rock
structures – this approach could be applied to other areas where ephemeral water
sources were integral to human survival. It has also demonstrated the value of floral
and faunal surveys when compiling lists of plant and animal foods available prior to
European arrival. This approach would be particularly useful in other agricultural
areas, where extensive clearing prompts regular biodiversity audits of uncleared land.
341
Finally, this research has supported Rossi's (2014b) earlier assertions that, in some
cases, occupation intensity can be uses as a proxy for effective rainfall; this identified
periods of lower rainfall from 8200–4571 cal. BP, and throughout most of the last
1000 years. While the accuracy of the results is still in question, the method warrants
further investigation, particularly due to the conflicting palaeoenvironmental records
from southwestern Australia.
342
REFERENCES
Abeysinghe, P.B. 2003 Silica resources of Western Australia. Western Australia
Geological Survey, Mineral Resources Bulletin 21. Perth: Geological Survey of
Western Australia.
Abbott, I. 2002 Origin and spread of the cat, Felis catus, on mainland Australia, with
a discussion of the magnitude of its early impact on native fauna. Wildlife Research
29:51–74.
Abbott, I. 2003 Aboriginal fire regimes in south-west Western Australia: evidence
from historical documents. In I. Abbott and N. Burrows (eds), Fire and ecosystems of
south-west Western Australia: impacts and management, pp.119–146. Leiden:
Backhuys Publishers.
Abbott, I. 2008 Historical perspectives of the ecology of some conspicuous
vertebrate species in south-west Western Australia. Conservation Science Western
Australia 6(3):1–214.
Allen, J. 1969 Archaeology and the history of Port Essington. Unpublished PhD
thesis, School of Archaeology and Anthropology, Australian National University,
Canberra.
Allen, J. and R. Jones 1980 Oyster Cove: archaeological traces of the last
Tasmanians and notes on the criteria for the authentication of flaked glass artefacts.
Papers and Proceedings of the Royal Society of Tasmania 114:225–233.
Allen, J. and J.F. O’Connell 2014 Both half right: updating the evidence for dating
first human arrival in Sahul. Australian Archaeology 79:86–108.
Allen, R.G., L.S. Pereira, D. Raes and M. Smith 1998 Crop evapotranspiration –
guidelines for computing crop requirements. FAO Irrigation and Drainage paper 56.
Rome: Food and Agriculture Organisation of the United Nations.
Ambrose, S.H. and K.G. Lorenz 1990 Social and ecological models for the Middle
Stone Age in Southern Africa. In P. Mellars (ed.), The emergence of modern humans:
an archaeological perspective, pp.3–33. New York: Cornell University Press.
Anand, R.R. and C.R.M Butt 2010 A guide for mineral exploration through the
regolith in the Yilgarn Craton, Western Australia. Australian Journal of Earth
Sciences 57(8):1015–1114. https://dx.doi.org/10.1080/08120099.2010.522823
Anand, R.R. and M. Paine 2002 Regolith geology of the Yilgarn Craton, Western
Australia: implications for exploration. Australian Journal of Earth Sciences 49(1):3–
162.
Anderson, J. 1984 Between plateau and plain: flexible responses to varied
environments in south-western Australia. Occasional Papers in Prehistory 4.
Canberra: Research School of Pacific Studies, Australian National University.
343
Andrefsky, W. 1994 Raw-material availability and the organization of technology.
American Antiquity 59:21–34.
Andrich, M.A. and J. Imberger 2013. The effect of land clearing on rainfall and fresh
water sources in Western Australia: a multi-functional sustainability analysis.
International Journal of Sustainable Development and World Ecology 2013.
https://dx.doi.org/10.1080/13504509.2013.850752
Atkins, C. and Y-L. Ruan 2010–2016 Phloem transport. In R. Munns, S. Schmidt and
C. Beveridge (eds), Plants in Action (2nd ed.). Retrieved from <
http://plantsinaction.science.uq.edu.au/content/chapter-5-phloem-transport >
Atlas of Living Australia (n.d.) Atlas of Living Australia [Computer Software].
Retrieved from < https://www.ala.org.au/ >
Attenbrow, V. 2004 What’s Changing: Population Size or Land-use Patterns? The
Archaeology of Upper Mangrove Creek, Sydney Basin. Terra Australis 21. Canberra:
Australian National University.
Australian Biological Resources Study 2009 Australian Faunal Directory [Computer
Software]. Retrieved from < https://biodiversity.org.au/afd/home >
Australian Government Department of Environment and Energy 2012 Interim
Biogeographic Regionalisation for Australia (IBRA), Version 7. Retrieved 7
September 2018 from
< http://www.environment.gov.au/land/nrs/science/ibra/australias-bioregions-maps >
Balme, J. 2014 Devils Lair: occupation intensity and land-use. Australian
Archaeology 79:179–186.
Bamforth, D.B. 1986 Technological efficiency and tool curation. American Antiquity
51(1):38–50.
Barham, A. and G. Huckleberry 2014 Sediments. In J. Balme and A. Paterson (eds),
Archaeology in practice: a student guide to archaeological analyses (2nd ed.), pp.47–
84. West Sussex: Wiley-Blackwell.
Barrows, T.T., S. Juggins, P. De Deckker, E. Calvo and C. Pelejero 2007 Long-term
sea surface temperature and climate change in the Australian–New Zealand region.
Paleoceanography 22:PA2215. https://dx.doi.org/10.1029/2006PA001328
Bates, D. 1938 The passing of the Aborigines. London: John Murray Ltd.
Bayly, I.A.E. 1999 Review of how indigenous people managed for water in desert
regions of Australia. Journal of the Royal Society of Western Australia 82(1):17–25.
Bayly, I. 2015 The importance of gnammas for Indigenous Australians. Wildlife
Australia 52(2):23–25.
344
Beard, J.S. 1972 Vegetation survey of Western Australia. The vegetation of the
Southern Cross area, Western Australia. Map and Explanatory memoir, 1:250 000
series. Sydney: Vegmap Publications.
Beard, J.S. 1975 The vegetation survey of Western Australia. Vegetatio 30(3):179–
187.
Beard, J.S. 1976 Vegetation survey of Western Australia. The vegetation of the
Boorabbin and Lake Johnston areas, Western Australia. Map and explanatory
memoir, 1:250 000 series. Sydney: Vegmap Publications.
Beard, J.S. 1979 Vegetation survey of Western Australia. The vegetation of the
Hyden area, Western Australia. Map and explanatory memoir, 1:250 000 series.
Sydney: Vegmap Publications.
Beard, J.S. 1980 Vegetation survey of Western Australia. The vegetation of the
Corrigin area, Western Australia. Map and explanatory memoir, 1:250 000 series.
Sydney: Vegmap Publications.
Beard, J.S. 1981 Vegetation survey of Western Australia. Swan. Series 1:1,000,000.
Crawley: University of Western Australia Press.
Beard, J.S. 1990 Plant life of Western Australia. Kenthurst: Kangaroo Press.
Beard, J.S. 1999 Evolution of the river systems of the south-west drainage division,
Western Australia. Journal of the Royal Society of Western Australia 82:147–164.
Beard, J.S., A.R. Chapman and P. Gioia 2000 Species richness and endemism in
the Western Australian flora. Journal of Biogeography 27(6):1257–1268.
Beaton, J.M. 1982 Fire and water: aspects of Australian Aboriginal management of
cycads. Archaeology in Oceania 17(1):51–58.
Beaumont, P. 1961 The problem of glass implements. Bulletin of the South African
Museums Association 7(8):161–163.
Beck, W. 1992 Aboriginal preparation of Cycas seeds in Australia. Economic Botany
46(2):133–147.
Beisaw, A. 1998 Differential preservation and recovery: taphonomy of bone
preservation at the Thomas/Luckey site, Ashland, NY. Unpublished MA thesis, State
University of New York, Binghamton.
Bell, D.T. and J.E. Williams 1997 Eucalypt Physiology. In J.E. Williams and J.C.Z.
Woinarski (eds), Eucalypt ecology: Individuals to ecosystems, pp.168–196.
Cambridge: Cambridge University Press.
345
Benison, K.C., B.B. Bowen, F.E. Oboh-Ikeunobe, E.A. Jagniecki, D.A. LaClair, S.L.
Story, M.R. Mormile and B-Y. Hong 2007 Sedimentology of acid saline lakes in
southern Western Australia: newly described processes and products of an extreme
environment. Journal of Sedimentary Research 77:366–388.
Benson, J. 1991 The effect of 200 years of European settlement on the vegetation
and flora of New South Wales. Cunninghamia 2(3):343–370.
Bettenay, E. 1962 The salt lake systems and their associated aeolian features in the
semi-arid regions of Western Australia. Journal of Soil Science 13(1):10–17.
Bettenay, E., A.V. Blackmore and F.J. Hingston 1964 Aspects of the hydrologic cycle
and related salinity in the Belka Valley, Western Australia. Australian Journal of Soil
Research 1964(2):187–210.
Bettenay, E. and F.J. Hingston 1963 The quality of groundwaters in the central
wheatbelt of W.A. Journal of the Department of Agriculture 4(4):216–219.
Bettinger, R.L., R. Garvey and S. Tushingham 2015 Hunter-gatherers:
archaeological and evolutionary theory (2nd ed.). New York: Springer.
Bickford, S. and Gell, P. 2005 Holocene vegetation change, Aboriginal wetland use
and the impact of European settlement on the Fleurieu Peninsula, South Australia.
The Holocene 15(2):200–215.
Bindon, P.R. 1996. Useful bush plants. Perth: Western Australian Museum.
Bindon, P.R. 1997 Aboriginal people and granite domes. Journal of the Royal
Society of Western Australia 80:173–179.
Binford, L.R.1973 Interassemblage variability – the Mousterian and the functional
argument. In C. Renfrew (ed.), The explanations of culture change – models in
prehistory, pp.227–254. London: Duckworth.
Binford, L.R. 1977 Forty seven trips. A case study in the character of archaeological
formation processes. In R.V.S. Wright (ed.), Stone tools as cultural markers, pp.24–
36. Canberra: Australian Institute of Aboriginal Studies.
Binford, L.R. 1979 Organization and formation processes: looking at curated
technologies. Journal of Anthropological Research 35(3):255–273.
Binford, L.R. 1980 Willow smoke and dogs' tails: Hunter-gatherer settlement systems
and archaeological site formation. American Antiquity 45(1):4–20.
Biological Surveys Committee 1984 The biological survey of the Eastern Goldfields
of Western Australia, part 1: introduction and methods. Records of the Western
Australian Museum Supplement 18:1–19.
346
Bird, C.F.M. 1985. Prehistoric lithic resource utilisation: a case study from the
southwest of Western Australia. Unpublished PhD thesis, Centre for Prehistory,
University of Western Australia, Crawley.
Bird, C.F.M. and C. Beeck 1988 Traditional plant foods in the southwest of Western
Australia: the evidence from salvage ethnography. In B. Meehan and R. Jones (eds),
Archaeology with ethnography: an Australian perspective, pp.113–22. Canberra:
Research School of Pacific Studies, Australian National University.
Bird, D.W., R.B. Bird and B.F. Codding 2009 In pursuit of mobile prey: Martu hunting
strategies and archaeofaunal interpretation. American Antiquity 74(1):3–29.
Bird, R.B., D.W. Bird, B.F. Codding, C.H. Parker and J.H Jones 2008 The “fire stick
farming” hypothesis: Australian Aboriginal foraging strategies, biodiversity, and
anthropogenic fire mosaics. Proceedings of the National Academy of Sciences of the
United States of America 105(39):14,796–14,801.
Bird, D.W., R.B. Bird and C.H. Parker 2005 Aboriginal burning regimes and hunting
strategies in Australia’s Western Desert. Human Ecology 33(4):443–464.
Bird, D.W., B.F. Codding, R.B. Bird and D.W. Zeanah 2012 Risky pursuits: Martu
hunting and the effects of prey mobility: reply to Ugan and Simms. American
Antiquity 77(1):186–194.
Bird, M.I., D. O’Grady and S. Ulm 2016 Humans, water and the colonization of
Australia. Proceedings of the National Academy of Sciences 113(41):11,477–11,482.
https://dx.doi.org/10.1073/pnas.1608470113
Bird, R.B., N. Tayor, B.F. Codding and D.W. Bird 2013 Niche construction and
Dreaming logic: aboriginal patch mosaic burning and varanid lizards (Varanus gouldii)
in Australia. Proceedings of the Royal Society B 280:20132297.
https://dx.doi.org/10.1098/rspb.2013.2297
Blatt, H. and R.J. Tracy 1996 Petrology: igneous, sedimentary, and metamorphic
(2nd ed.). New York: W.H. Freeman and Company.
Bleed, P. 1986 The optimal design of hunting weapons: maintainability or reliability.
American Antiquity 51(4):737–747.
Blewett, R.S., B.L.N. Kennett and D.L. Huston 2012 Australia in time and space. In
R.S. Blewett (ed.), Shaping a nation: a geology of Australia, pp.47–120. Canberra:
Geoscience Australian and ANU E Press.
Bowdler, S., J. Harris, A. Murphy, G. Nayton and C. Pocock 1989. A test excavation
at Mulka’s Cave (Bates Cave) near Hyden, Western Australia. Unpublished report
prepared for the Department of Aboriginal Sites, Western Australian Museum, Perth.
Bowen, B.B. and K.C. Benison 2009 Geochemical characteristics of naturally acid
and alkaline saline lakes in southern Western Australia. Applied Geochemistry
24:268–284.
347
Bowler, J.M. 1986 Spatial variability and hydrologic evolution of Australian lake
basins: analogue for Pleistocene hydrologic change and evaporite formation.
Palaeogeography, Palaeoclimatology, Palaeoecology 54:21–41.
Bowler, J.M., G.S. Hope, J.N. Jennings, G. Singh and D. Walker 1976 Late
quaternary climates of Australia and New Guinea. Quaternary Research 6(3):359–
394.
Braadbaart, F., I. Poole and A.A. van Brussel 2009 Preservation potential of charcoal
in alkaline environments: an experimental approach and implications for the
archaeological record. Journal of Archaeological Science 36:1672–1679.
https://dx.doi.org/10.1016/j.jas.2009.03.006
Brand, J. 2005 WA sandalwood (Santalum spicatum) establishment guide for
farmland in the Wheatbelt, July 2005. Retrieved 27 March 2017 from
< http://avongro.com.au/Webpages/documents/sandal2005Jul.pdf >
Brand-Miller, J.C. and S.H.A. Holt 1998 Australian Aboriginal plant foods: a
consideration of their nutritional composition and health implications. Nutrition
Research Reviews 11:5–23.
Brand-Miller, J., K.W. James and P.M.A. Maggiore 2014 Tables of composition of
Australian Aboriginal foods. Canberra: Aboriginal Studies Press.
Briscoe, G. 2003 Counting, health and identity: a history of Aboriginal health and
demography in Western Australia and Queensland, 1900–1940. Canberra:
Aboriginal Studies Press.
Brodie, R.S., K.C. Lawrie and D.P. Commander 2012 Groundwater – lifeblood of the
continent. In R.S. Blewett (ed.), Shaping a nation: a geology of Australia, pp.333–
379. Canberra: Geoscience Australia and ANU E-press.
Bronk Ramsey, C. 2009 Bayesian analysis of radiocarbon dates. Radiocarbon
51(1):337–360.
Brooke, B.P., S.L. Nichol, Z. Huang and R.J. Beaman 2017 Palaeoshorelines on the
Australian continental shelf: Morphology, sea-level relationship and applications to
environmental management and archaeology. Continental Shelf Research 134:26–
38.
Brown, A., P. Dundas, K. Dixon and S. Hopper 2008 Orchids of Western Australia.
Crawley: UWA Press.
Brown, A.R. 1918 Notes on the social organization of Australian tribes. The Journal
of the Royal Anthropological Institute of Great Britain and Ireland 48:222–253.
Cale, D.J., S.A. Halse and C.D. Walker 2004 Wetland monitoring in the Wheatbelt of
south-west Western Australia: site descriptions, waterbird, aquatic invertebrate and
groundwater data. Conservation Science Western Australia 5(1):20–135.
348
Campbell, E.M. 1997 Granite landforms. Journal of the Royal Society of Western
Australia 80(3):101–112.
Campbell, J. 2002 Invisible invaders: smallpox and other diseases in Aboriginal
Australia 1780–1880. Melbourne: Melbourne University Press.
Cane, S. 1987 Australian Aboriginal subsistence in the Western Desert. Human
Ecology 15(4):391–434.
Chapman, A. and J. Dell 1977 Reptiles and frogs of Bendering and West Bendering
nature reserves. Records of the Western Australian Museum Supplement 5:47–58.
Chappell, J. 1991 Late Quaternary environmental changes in eastern and central
Australia, and their climatic interpretation. Quaternary Science Reviews 10(5):377–
390.
Charnov, E. 1976 Optimal foraging: the marginal value theorem. Theoretical
Population Biology 9(2):129–136.
Chin, R.J., A.H. Hickman and R. Thom 1984 Hyden SI 50-4 Australia 1:250 000
geological series – Explanatory notes. Perth: Geological Survey of Western Australia.
Churchill, D.M. 1968 The distribution and prehistory of Eucalyptus diversicolor F.
Muell., E. marginata Donn ex Sm., and E. calophylla R. Br. in relation to rainfall.
Australian Journal of Botany 16(1):125–151.
Clarke, C.J., R.J. George, R.W. Bell and T.J. Hatton 2002 Dryland salinity in southwestern Australia: its origins, remedies and future research directions. Australian
Journal of Soil Research 40(1):93–113.
Clarke, C.J., R.J. George, R.W. Bell and R.J. Hobbs 1998 Major faults and the
development of dryland salinity in the western wheatbelt of Western Australia.
Hydrology and Earth System Sciences 2(1):77–91.
Clarke, P.A. 2003 Australian ethnobotany: an overview. Australian Aboriginal Studies
2003/2:21–38.
Clarke, P.A. 2009 Australian Aboriginal ethnometeorology and seasonal calendars.
History and Anthropology 20(2):79–106.
https://dx.doi.org/10.1080/02757200902867677
Clarke, P.A. 2011 Aboriginal people and their plants. Dural: Rosenberg Publishing.
Clarkson, C. 2002 An index of invasiveness for the measurement of unifacial and
bifacial retouch: a theoretical, experimental and archaeological verification. Journal
of Archaeological Science 29:65–75.
Clarkson, C. 2004 Technological provisioning and assemblage variation in the
Eastern Victoria River region, Northern Australia: A Darwinian approach.
Unpublished PhD thesis, The Australian National University, Canberra.
349
Clarkson, C. 2006 Interpreting surface assemblage variation in Wardaman country,
Northern Territory: an ecological approach. In S. Ulm and I. Lilley (eds), An
archaeological life: papers in honour of Jay Hall, pp.177–190. Aboriginal and Torres
Strait Islander Studies Unit Research Report Series, 7. Brisbane: University of
Queensland.
Clarkson, C. 2007 Lithics on the land of the lightning brothers: the archaeology of
Wardaman country, Northern Territory. Terra Australis 25. Canberra: ANU e-press.
Clarkson, C., M. Haslam and C. Harris 2015 When to retouch, haft or discard?
Modelling optimal use/maintenance schedules in lithic tool use. In N. Goodale and W.
Andrefsky (eds), Lithic technological systems and evolutionary theory, pp.117–138.
New York: Cambridge University Press.
Clarkson, C., Z. Jacobs, B. Marwick, R. Fullagar, L. Wallis, M. Smith, R.G. Roberts,
E. Hayes, K. Lowe, X. Carah, S.A. Florin, J. McNeil, D. Cox, L.J. Arnold, Q. Hua, J.
Huntley, H.E.A. Brand, T. Manne, A. Fairbairn, J. Shulmeister, l. Lyle, M Salinas, M.
Page, K. Connell, G. Park, K. Norman, T. Murphy and C. Pardoe 2017 Human
occupation of northern Australia by 65,000 years ago. Nature 547:306–310.
https://dx.doi.org/10.1038/nature22968
Clarkson, C. and S. O’Connor 2014 An introduction to stone artefact analysis. In J.
Balme and A. Paterson (eds), Archaeology in practice: a student guide to
archaeological analyses (2nd ed.), pp.151–206. West Sussex: Wiley-Blackwell.
Cleland, J.B. 1928 Diseases amongst the Australian Aborigines. London: Bale, Sons
and Danielsson Ltd.
Coates, A. 1992 Flora and vegetation survey of Dragon Rocks nature reserve (no. A
36128). Unpublished report prepared for the Department of Conservation and Land
Management, Western Australia.
Cogger, H. 2014 Reptiles and amphibians of Australia. Collingwood: CSIRO
Publishing.
Collard, M., B. Buchanan, M.J. O’Brien and J. Scholnick 2013 Risk, mobility or
population size? Drivers of technological richness among contact-period western
North American hunter-gatherers. Philosophical Transactions of the Royal Society B
368:20120412. https://dx.doi.org/10.1098/rstb.2012.0412
Collard, L. and S. Harben 2010 Nartj Katitj Bidi Ngulluckiny Koorl? (Which
knowledge path will we travel?). Studies in Western Australian History 26:75–95.
Collard, M., M. Kemery and S. Banks 2005 Causes of toolkit variation among huntergatherers: a test of four competing hypotheses. Canadian Journal of Archaeology
29:1–19.
350
Commander, D.P., N. Schoknecht, W. Verboom and P. Caccetta 2001 The geology,
physiography and soils of Wheatbelt valleys. In Dealing with salinity in Wheatbelt
valleys: processes, prospects and practical options: proceedings of a conference
held at Merredin, Western Australia – 30th July – 1st August 2001. Perth: Viv Read
and Associates.
Conkey, M.W. 1980 The identification of prehistoric hunter-gatherer aggregation
sites: the case of Altamira. Current Anthropology 21(5):609–630.
Conte, I.C. and F.G. Romero 2008 Microwear analysis of retouched glass fragments
from Fortlet Miñana, Azul, Argentina, 1860–1863.International Journal of Historical
Archaeology 12(3):2480262.
Cooper, Z. and S. Bowdler 1988 Flaked glass tools from the Andaman Islands and
Australia. Asian Perspectives 37(1):74–83.
Corsini, S. 1993 Investigation of Aboriginal sites in the vicinity of Wave Rock, Hyden.
Unpublished report prepared for the Mallee Aboriginal Corporation and the Western
Australian Museum.
Cotterell, B., and J. Kamminga 1979 The mechanics of flaking. In B. Hayden (ed.),
Lithic usewear analysis, pp.97–112. New York: Academic Press.
Cotterell, B. and J. Kamminga 1987 The formation of flakes. American Antiquity
52(4):675–708.
Cowgill, G.L. 2015 Some things I hope you will find useful even if statistics isn’t your
thing. Annual Review of Anthropology 44:1–14.
Cribb, A.B. and J.W. Cribb 1975 Wild food in Australia. Sydney: William Collins
Publisher.
Cronin, L. 2008 Cronin’s key guide: Australian mammals. Crows Nest: Jacana Books.
Cronin, L. 2014 Cronin’s key guide: Australian reptiles and frogs. Crows Nest:
Jacana Books.
Cullen, L.E. and P.F. Grierson 2009 Multi-decadal scale variability in autumn-winter
rainfall in south-western Australia since 1655 AD as reconstructed from tree rings of
Callitris columellaris. Climate Dynamics 33(2–3):433–444.
Curr, E.M. 1886 The Australian race: its origin, languages, customs, place of landing
in Australia and the routes by which it spread itself over that continent, Volume 1.
London: Trübner and Co.
Curr, E.M. 1887 Continental Australia showing the routes by which the Aboriginal
race spread itself throughout the continent [map]. Victoria: Lands and Surveys
Department.
351
David, B., B. Barker, F. Petchey, J-J. Delannoy, J-M. Geneste, C. Rowe, M.
Eccleston, L. Lamb and R. Whear 2013 A 28,000-year-old excavated painted rock
from Nawarla Gabarnmang, northern Australia. Journal of Archaeological Science
40:2493–2501.
Davidson, D.S. 1938 An ethnic map of Australia. Proceedings of the American
Philosophical Society 79:649–680.
Davidson, D.S. 1952. Notes on the pictographs and petroglyphs of Western Australia
and a discussion of their affinities with appearances elsewhere on the continent.
Proceedings of the American Philosophical Society 96:76–117.
Davidson, W.A. 1977 Hydrogeology and drilling results of the 1969–1970 drought
relief program. Geological Survey of Western Australia Record no. 1977/3. Perth:
Geological Survey of Western Australia.
Daw, B., T. Walley and G. Keighery 1997 Bush tucker plants of the south-west.
Como: Department of Conservation and Land Management.
Day, R.B. 1951 The Humps beyond Hyden. The West Australian Saturday 27
January 1951 p. 18.
Deardorff, J.W. 1977 A parameterization of ground-surface moisture content for use
in atmospheric prediction models. Journal of Applied Meteorology 16:1182–1185.
Deardorff, J.W. 1978 Efficient prediction of ground surface temperature and
moisture, with inclusion of a layer of vegetation. Journal of Geophysical Research
83:1889–1903.
De Dekker, P. 1983 Australian salt lakes: their history, chemistry, and biota – a
review. Hydrobiologia 105(1):231–244.
De La Pena, P. 2015 A qualitative guide to recognize bipolar knapping for flint and
quartz. Lithic Technology 40(4):1–16.
Dell, J. 1977 Birds of Bendering and West Bendering nature reserves. Records of
the Western Australian Museum Supplement 5:31–46.
de Lombera Hermida, A. 2008 Quartz morphostructural groups and their mechanical
implications. Annali dell’Università degli Studi di Ferrara 2008:101–104.
de Lombera Hermida, A. 2009 The scar identification of lithic quartz industries. In F.
Sternke, L. Eigeland and L.J. Costa (eds), Non-flint raw material use in prehistory,
pp.5–11. Oxford: Archaeopress.
Department of Environment 2005a The ecology of wheatbelt lakes. Water notes 33.
Perth: Department of Environment
Department of Environment 2005b The Wheatbelt's ancient rivers. Water notes 34.
Perth: Department of Environment.
352
Department of Primary Industries and Regional Development (DPIRD) 2016a Soil
Water Tool [Computer Software]. Retrieved from
< https://www.agric.wa.gov.au/climate-weather/soil-water-tool >
Department of Primary Industries and Regional Development (DPIRD) 2016b NRInfo
[Computer Software]. Retrieved from < https://maps.agric.wa.gov.au/nrinfo/ >
Department of Water 2009 Waterway assessment of the Camm River: Lockhart
River confluence to Hyden. Water resource management series report number WRM
57. Perth: Department of Water.
Dickman, C.R. 1996 Impact of exotic generalist predators on the native fauna of
Australia. Wildlife Biology 2(3):185–195.
Dickson, F.P. 1977 Quartz flaking. In R.V.S. Wright (ed.) Stone tools as cultural
markers: change, evolution and complexity, pp.97–103. Canberra: Australian
Institute of Aboriginal Studies.
Diez-Martin, F., P.S. Yustos, M. Dominguez-Rodrigo and M.E. Prendergast 2011 An
experimental study of bipolar and freehand knapping of Naibor Soit quartz from
Olduvai Gorge (Tanzania). American Antiquity 76(4):690–708.
Dingman, S.L. 1994 Physical hydrology. New York: Macmillan.
Dirnböck, T., R.J. Hobbs, R.J. Lambeck and P.A. Caccetta 2002 Vegetation
distribution in relation to topographically driven processes in southwestern Australia.
Applied Vegetation Science 5(1):147–158.
Dodson, J.R. 2001 Holocene vegetation change in the Mediterranean-type regions of
Australia. The Holocene 11(6):673–680.
Dodson, J.R. and J.J. Lu 2000 A Late Holocene vegetation and environment record
from Byenup Lagoon, south-western Australia. Australian Geographer 31(1):41–54.
Dominguez-Villar, D., C. Arteaga, R. Garcia-Gimenez, E.A. Smith and J. Pedraza
2008 Diurnal and seasonal water variations of temperature, pH, redox potential and
conductivity in gnammas (weathering pits): implications for chemical weathering.
Catena 72:37–48.
Dortch, C.E. 2000 Past Aboriginal hunter-gatherer economy and territorial
organisation in coastal districts of Western Australia's lower Southwest. Unpublished
PhD thesis, Centre for Archaeology, Department of Anthropology, University of
Western Australia, Crawley.
Dortch, J. 1996 Late Pleistocene and recent Aboriginal occupation of Tunnel Cave
and Witchcliffe rock shelter, Southwestern Australia. Australian Aboriginal Studies
2:51–60.
353
Dortch, J. 2000 Palaeo-environmental change and the persistence of human
occupation in south-western Australian forests. Unpublished PhD thesis, Centre for
Archaeology, Department of Anthropology, University of Western Australia, Crawley.
Dortch, J. 2004 Palaeoenvironmental change and the persistence of human
occupation in south-western Australian forests. BAR International Series 1288.
Oxford: Archaeopress.
Dowling, P. 1997 “A great deal of sickness”: introduced diseases among the
Aboriginal people of colonial southeast Australia, 1788–1900. Unpublished PhD
thesis, Australian National University, Canberra.
Driscoll, K. 2010 Understanding quartz technology in early prehistoric Ireland.
Unpublished PhD thesis, College of Art and Celtic Studies, School of Archaeology,
University College Dublin, Belfield.
Driscoll, K. 2011a Vein quartz in lithic traditions: an analysis based on experimental
archaeology. Journal of Archaeological Science 38:734–745.
Driscoll, K. 2011b Identifying and classifying vein quartz artefacts: an experiment
conducted at the World Archaeological Congress, 2008. Archaeometry 53(6):1280–
1296. https://dx.doi.org/10.1111/j.1475-4754.2011.00600.x
Dyson-Hudson, R. and E.A. Smith 1978 Human territoriality: an ecological
reassessment. American Anthropologist 80:21–41.
Emlen, J.M. 1966 The role of time and energy in food preference. The American
Naturalist 100(916):611–617.
Erickson, R., A.S. George, N.G. Marchant and M.K. Morcombe 1979 Flowers and
plants of Western Australia (revised ed.). Sydney: Reed.
Farrington, P. and R.B. Salama 1996 Controlling dryland salinity by planting trees in
the best hydrogeological setting. Land Degradation and Development 7:183–204.
Ferguson, W.C. 1985 A mid-Holocene depopulation of the Australian Southwest.
Unpublished PhD thesis, Australian National University, Canberra.
Fernie, N. 1930 Water supplies from rock catchments in the Western Australian
wheat belt. The Journal of the Institution of Engineers Australia 2(6):198–208.
Ferris, J.M. 2015 Procurement costs and tool performance requirements:
determining constraints on lithic toolstone selection in Baja Calafornia Sur. In N.
Goodale and W. Andrefsky (eds), Lithic technological systems and evolutionary
theory, pp.139–155. New York: Cambridge University Press.
Fetherston, J.M., S.M. Stocklmayer and V.C. Stocklmayer 2017 Gemstones of
Western Australia second edition. Geological Survey of Western Australia, Mineral
Resources Bulletin 25. Perth: Geological Survey of Western Australia.
354
Florabank (n.d.) Fact Sheets. Retrieved 27 March 2017 from
< http://www.florabank.org.au/lucid/key/species%20navigator/media/html/index.htm >
Fullagar, R., J. McDonald, J. Field and D. Donlon 2009 Deadly weapons: backed
microliths from Narrabeen, New South Wales. In M. Haslam, G. Robertson, A.
Crowther, S. Nugent and L. Kirkwood (eds), Archaeological science under a
microscope: Studies in residue and ancient DNA analysis in honour of Thomas H.
Loy, pp.258–270. Terra Australis 30. Canberra: ANU E Press.
Galloway, P.D. 2004 Agricultural sub-regions of the Avon River basin. Resource
Management Technical Report 284. Western Australia: Department of Agriculture.
Galloway, R.W. 1965 Late Quaternary climates in Australia. The Journal of Geology
73(4):603–618.
Gammage, B. 2011 The biggest estate on earth: how Aborigines made Australia.
Crow's Nest: Allen and Unwin.
Garvey, R. 2015 A model of lithic raw material procurement. In N. Goodale and W.
Andrefsky (eds), Lithic technological systems and evolutionary theory, pp.156–171.
New York: Cambridge University Press.
Geological Survey of Western Australia 1984 Hyden SI 50-4 [map] Australia 1:250
000 geological series. Perth: Geological Survey of Western Australia.
George, R.J. 1992 Hydraulic properties of groundwater systems in the saprolite and
sediments of the wheatbelt, Western Australia. Journal of Hydrology 130:251–278.
George, R., J. Clarke and P. English 2008 Modern and palaeogeographic trends in
the salinisation of the Western Australian wheatbelt: a review. Australian Journal of
Soil Research 46:751–767.
George, R.J. and M. Coleman 2001 Hidden menace or opportunity – groundwater
hydrology, playas and commercial options for salinity in wheatbelt valleys. In Dealing
with salinity in Wheatbelt valleys: processes, prospects and practical options:
proceedings of a conference held at Merredin, Western Australia – 30th July – 1st
August 2001. Perth: Viv Read and Associates.
George, R., D. McFarlane and B. Nulsen 1997 Salinity threatens the viability of
agriculture and ecosystems in Western Australia. Hydrogeology Journal 5(1):6–21.
Ghauri, S. 2004 Groundwater trends in the central agricultural region. Resource
Management Technical Report 269. Perth: Department of Agriculture and Food.
Gibbs, M. and P. Veth 2002 Ritual Engines and the Archaeology of Territorial
Ascendancy. In S. Ulm, C. Westcott, J. Reid, A. Ross, I. Lilley, J. Pragnell and L.
Kirkwood (eds), Barriers, Borders, Boundaries: Proceedings of the 2001 Australian
Archaeological Association Annual Conference Tempus 7, pp.11–19. Brisbane:
University of Queensland.
355
Gibson, N., G.J. Keighery, N.M. Lyons and A. Webb 2004 Terrestrial flora and
vegetation of the Western Australian wheatbelt. Records of the Western Australian
Museum Supplement 67:139–189.
Glendenning, W. 2004 A report of an Aboriginal heritage survey of the area of
proposed water treatment facilities and associated pipework upgrades at Karlgarin
and Wave Rock, near Hyden. Unpublished report prepared for the Water
Corporation.
Glover, J.E. 1984 The geological sources of stone for artefacts in the Perth Basin
and nearby areas. Australian Aboriginal Studies 1984/1:17–25.
Goode, B. 2011 An ethnographic Aboriginal heritage survey of the proposed
development of Reserve 23128 in Hyden, Western Australia. Unpublished report
prepared for South West Aboriginal Land and Sea Council (SWALSC).
Goodyear, A.C. 1989 A hypothesis for the use of crypto-crystalline raw materials
among Paleoindian groups of North America. In C.G. Ellis and J.C. Lothrop (eds),
Eastern Paleoindian lithic resource use, pp.1–9. Boulder: Westview Press.
Gordon, C.C. and J.E. Buikstra 1981 Soil pH, bone preservation, and sampling bias
at mortuary sites. American Antiquity 46(3):566–571.
Gould, R.A. 1968 Living archaeology: the Ngatatjara of Western Australia.
Southwestern Journal of Anthropology 24(2):101–122.
Gould, R.A. 1969a Yiwara: Foragers of the Australian Desert. London: Collins.
Gould, R.A. 1969b Subsistence behaviour among the Western Desert Aborigines of
Australia. Oceania 39(4):253–274.
Gould, R.A. 1971a The archaeologist as ethnographer: A case study from the
Western Desert of Australia. World Archaeology 3(2):143–177.
Gould, R.A. 1971b Uses and effects of fire among the Western Desert Aborigines of
Australia. Mankind 8:14–24.
Gould, R.A. 1984 Rock pools and desert dances. Natural History 93(3):61–71.
Gould, R.A. 1991 Arid-land foraging as seen from Australia: Adaptive models and
behavioral realities. Oceania 62(1):12–33.
Gouramanis, C., J. Dodson, D. Wilkins, P. De Deckker and B.M. Chase 2012
Holocene palaeoclimate and sea-level fluctuation recorded from the coastal Barker
Swamp, Rottnest Island, south-western Western Australia. Quaternary Science
Reviews 54:40–57.
Goward, T. 2011 Aboriginal glass artefacts of the Sydney region. Unpublished Hons
thesis, Department of Archaeology, University of Sydney, Camberdown/Darlington.
356
Graf, K.E. 2010 Hunter-gatherer dispersals in the mammoth-steppe: technological
provisioning and land-use in the Enisei River valley, south-central Siberia. Journal of
Archaeological Science 37:210–223.
Green, N. 2011 Explorers and the Aborigines. In M. Herlock and S. Milentis (eds),
Western Australian exploration 1836–1845, pp.419–437. Perth: Hesperian Press.
Green, N.J. 1981 Aborigines and white settlers in the nineteenth century. In C.T.
Stannage, C.T. (ed.), A new history of Western Australia, pp.72–123. Crawley:
University of Western Australia Press.
Green, R.L. and R.J. Watling 2007 Trace element fingerprinting of Australian ocher
using Laser Ablation Inductively Coupled Plasma-Mass Spectrometry (LA-ICP-MS)
for the provenance establishment and authentication of Indigenous art. Journal of
Forensic Sciences 52(4):851–859. https://doi.org/10.1111/j.1556-4029.2007.00488.x
Grey, G. 1841 Journals of two expeditions of discovery in north-west and western
Australia, during the years 1837, 38 and 39. London: Boone.
Grøn, O. 2012 Our grandfather sent the elk – some problems for hunter-gatherer
predictive modelling. Quartär 59:175–188.
Gunn, R.G. 2003 Mulka’s Cave Aboriginal Rock Art Site: a discussion of
management options. Unpublished report prepared for the Department of Indigenous
Affairs.
Gunn, R.G. 2004 Mulka’s Cave Aboriginal rock art site: an archaeological recording.
Unpublished report prepared for the Department of Indigenous Affairs and HydenKarlgarin Landcare Group.
Gunn, R.G. 2006 Mulka’s Cave Aboriginal rock art site: its context and content.
Records of the Western Australian Museum 23:19–41.
Gunness, A. 2002 A survey of vegetation and flora: Wogarl Reserve, in the Graball
Catchment, Shire of Narembeen. Nedlands: Wildflower Society of Western Australia.
Haebich, A. and L. Tilbrook 1981 South west Aboriginal studies bibliography: with
annotations and appendices. Mount Lawley: Mount Lawley College.
Hall, J. and W.R.F. Love 1985 Prickly Bush, a site with backed blades on the
Brisbane River: a pilot study towards the measurement of site “disturbance”.
Queensland Archaeological Research 2:71–81.
Hallam, S.J. 1972 An archaeological survey of the Perth area, Western Australia: a
progress report on art and artefacts, dates and demography. Australian Institute of
Aboriginal Studies Newsletter 3(5):11–19.
Hallam, S.J. 1975. Fire and hearth: a study of European usage and European
usurpation in southwestern Australia. Canberrra: Australian Institute of Aboriginal
Studies.
357
Hallam, S.J. 1977 Topographic archaeology and artifactual evidence. In R.V.S.
Wright (ed.), Stone tools as cultural markers: change, evolution and complexity,
pp.169–177. Canberra: Australian Institute of Aboriginal Studies.
Hallam, S.J. 1989 Plant usage and management in southwest Australian Aboriginal
societies. In D.R. Harris and G.C. Hillman (eds), Foraging and farming: the evolution
of plant exploitation, pp. 136–147. London: Unwin Hyman Ltd.
Halse, S.A., M.N. Lyons, A.M. Pinder and R.J. Shiel 2004 Biodiversity patterns and
their conservation in wetlands of the Western Australian wheatbelt. Records of the
Western Australian Museum Supplement 67:337–364.
Halse, S.A., G.B. Pearson and S. Patrick 1993 Vegetation of depth-gauged wetlands
in nature reserves of south-west Western Australia. Department of Conservation and
Land Management Technical Report 30. Como: Department of Conservation and
Land Management.
Halse, S.A., J.K. Ruprecht and A.M. Pinder 2003 Salinisation and prospects for
biodiversity in rivers and wetlands of south-west Western Australia. Australian
Journal of Botany 51(6):673–688.
Hammond, J.E. 1933. Winjan’s people. Perth: Imperial Printing Company.
Harpending, H. and H. Davis 1977 Some implications for hunter-gatherer ecology
derived from the spatial structure of resources. World Archaeology 8(3):275–286.
Harrison, R. 2000 ‘Nowadays with glass’: regional variation in Aboriginal bottle glass
artefacts from Western Australia. Archaeology in Oceania 35:34–47.
Harrison, S.P. 1993 Late Quaternary lake-level changes and climates of Australia.
Quaternary Science Reviews 12(4):211–231.
Hassell, E. 1975 My dusky friends. Dalkeith: C.W. Hassell.
Hatton, T.J., J. Ruprecht and R.J. George 2003 Preclearing hydrology of the
Western Australian wheatbelt: target for the future? Plant and Soil 257:341–356.
Hawkes, K. and J. O’Connell 1992 On optimal foraging models and subsistence
transitions. Current Anthropology 33(1):63–66.
Hays, J.D., J. Imbrie and N.J. Shackleton 1976 Variations in the earth’s orbit:
pacemaker of the Ice Ages. Science 194(4270):1121–1132.
Hefferan, K. and J. O’Brien 2010 Earth minerals. West-Sussex: Wiley-Blackwell.
Henrich, J. 2004 Demography and cultural evolution: why adaptive cultural
processes produced maladaptive losses in Tasmania. American Antiquity 69:197–
218. https://dx.doi.org/10.2307/4128416
358
Hercock, M. (ed.) 2014 The Western Australian explorations of John Septimus Roe
1829–1849. Carlisle: Hesperian Press.
Hesse, P.P., J.W. Magee and S. van der Kaars 2004 Late Quaternary climates of the
Australian arid zone: a review. Quaternary International 118:87–102.
Hillel, D. 2004 Introduction to environmental soil physics. San Diego: Academic
Press.
Hiscock, P. 1988 Prehistoric settlement patterns and artefact manufacture at Lawn
Hill, northwest Queensland. Unpublished PhD thesis, Department of Anthropology
and Sociology, University of Queensland, St Lucia.
Hiscock, P. 1993 Bondaian technology in the Hunter Valley, New South Wales.
Archaeology in Oceania 28:65–76.
Hiscock, P. 1994 Technological responses to risk in Holocene Australia. Journal of
World Prehistory 8(3):267–292.
Hiscock, P. 1996 Mobility and technology in the Kakadu coastal wetlands. IndoPacific Prehistory Association Bulletin 15:151–157.
Hiscock, P. 2002a Pattern and context in the Holocene proliferation of backed
artifacts in Australia. In R.G. Elston and S.L. Kuhn (eds), Thinking small: global
perspectives on microlithization, pp.163–177. Archaeological Papers of the American
Anthropological Association 12. Washington: American Anthropological Association.
Hiscock, P. 2002b Quantifying the size of artefact assemblages. Journal of
Archaeological Science 29(3):251–258.
Hiscock, P. 2006 Blunt and to the point: changing technological strategies in
Holocene Australia. In I. Lilley (ed.), Archaeology of Oceania: Australia and the
Pacific Islands, pp.69–95. Malden: Blackwell Publishing.
Hiscock, P. 2007 Looking the other way. A materialist/technological approach to
classifying tools and implements, cores and retouched flakes. In S. McPherron and J.
Lindley (eds), Tools or cores? The identification and study of alternative core
technology in lithic assemblages, pp.198–222. Newcastle: Cambridge Scholars
Publishing.
Hiscock, P. 2013 Occupying new lands: global migrations and cultural diversification
with particular reference to Australia. In K.E. Graf, C.V. Ketron and M.R. Waters
(eds), Paleoamerican odyssey, pp.3–11. College Station Texas: Department of
Anthropology, Texas A&M University.
Hiscock, P. 2015 Dynamics of knapping with bipolar techniques: modeling transitions
and the implications of variability. Lithic Technology 40(4):342–348.
https://doi.org/10.1179/2051618515Y.0000000021
359
Hiscock, P. and C. Clarkson 2000 Analysing Australian stone artefacts: an agenda
for the twenty first century. Australian Archaeology 50:98–108.
Hobbs, R.J. 2003 The wheatbelt of Western Australia. Pacific Conservation Biology
9(1):9–11.
Hogg, A., Hua, Q., Blackwell, P., Niu, M., Buck, C., Guilderson, T., Heaton, T.,
Palmer, J., Reimer, P., Reimer, R., Turney, C. and S. Zimmerman 2013 SHCal13
Southern Hemisphere Calibration, 0–50,000 Years cal BP. Radiocarbon 55(4):1889–
1903. https://doi.org/10.2458/azu_js_rc.55.16783
Holdaway, S. and N. Stern 2004. A record in stone: the study of Australia's flaked
stone artefacts. Canberra: Aboriginal Studies Press.
Hollick, A. 2004 South East Humps focus catchment report. Research Management
Technical Report 291. Perth: Department of Agriculture and Food.
Hopper, S.D. 2000 Creation of conservation reserves and managing fire on granite
outcrops – a case study of Chiddarcooping Nature Reserve in the Western
Australian wheatbelt. Journal of the Royal Society of Western Australia 83:173–186.
Hopper, S.D. 2001 Plant list for Hyden Rock. In C.R. Twidale and J.A. Bourne (eds),
Field Guide to Hyden Rock, including Wave Rock, pp.64–68. Hyden: Wave Rock
Management Pty Ltd.
Hopper, S.D., A.P. Brown and N.G. Marchant 1997 Plants of Western Australian
granite outcrops. Journal of the Royal Society of Western Australia 80(3):141–158.
Hopper, S.D. and H. Lambers 2014 Human relationships with and use of Kwongkan
plants and lands. In H. Lambers (ed.), Plant life on the sandplains in southwest
Australia: a global biodiversity hotspot, pp.285–306. Crawley: UWA Publishing.
Hopper, S.D., S. van Leeuwen, A.P. Brown and S.J. Patrick 1990 Western
Australia's endangered flora and other plants under consideration for declaration.
Perth: Department of Conservation and Land Management.
Horn, H.S. 1968 The adaptive significance of colonial nesting in the Brewer’s
Blackbird (Euphagus cyanocephalus). Ecology 49:682–694.
Horne, G. and G. Aiston 1928 Savage life in Central Australia. London: Macmillan.
Host, J. and C. Owen 2009 “It’s still in my heart, this is my country”: the single
Noongar claim history. Perth: University of Western Australia Press.
Houser, P.R. 2003 Infiltration and soil moisture processes. In T.D. Potter and B.R.
Colman (eds), Handbook of weather, climate and water: atmospheric, chemistry,
hydrology and societal impacts, pp.493–506. Hoboken: John Wiley and Sons.
Itzstein-Davey, F. 2004 An Early Holocene palaeoenvironmental record from Two
Mile Lake, South-Western Australia. Australian Geographer 35(3):317–332.
360
Jeske, R.J. and R. Lurie 1993 The archaeological visibility of bipolar technology: an
example from the Koster site. Midcontinental Journal of Archaeology 18(2):131–160.
Johnson, D. 2009 The geology of Australia. Melbourne: Cambridge University Press.
Jones, R. 1969 Fire-stick farming. Australian Natural History 16:224–228.
Kamminga, J. 1980 A functional investigation of Australian microliths. The Artefact
5:1–18.
Kelleher, M.H. 2002 Archaeology of sacred space: the spatial nature of religious
behaviour in the Blue Mountains National Park, Australia. Unpublished PhD thesis,
University of Sydney, Sydney.
Kelly, R.L. 2013 The lifeways of hunter-gatherers: the foraging spectrum (2nd ed.).
New York: Cambridge University Press.
Kendrick, G.W. 1977 Middle Holocene marine molluscs from near Guildford,
Western Australia, and evidence for climate change. Journal of the Royal Society of
Western Australia 59(4):97–104.
Kershaw, P., P. Moss and S. van der Kaars 2003 Causes and consequences of
long-term climatic variability on the Australian continent. Freshwater Biology
48(7):1274–1283.
Kershaw, A.P. and G.C. Nanson 1993 The last full glacial cycle in the Australian
region. Global and Planetary Change 7(1):1–9.
Kitchener, D.J. and A. Chapman 1977 Mammals of Bendering and West Bendering
nature reserves. Records of the Western Australian Museum Supplement 5:17–29.
Kitchener, D.J., A. Chapman and B.G. Muir 1980 The conservation value for
mammals of reserves in the Western Australian wheatbelt. Biological Conservation
18(3):179–207.
Knutsson, H., K. Knutsson, N. Taipale, M. Tallavaara and K. Darmark 2015 How
shattered flakes were used: micro-wear analysis of quartz flake fragments. Journal of
Archaeological Science: Reports 2:517–531.
Kuhn, S.L. 1990 A geometric index of reduction for unifacial stone tools. Journal of
Archaeological Science 17:585–593.
Kuhn, S.L. 1995 Mousterian lithic technology: an ecological perspective. New Jersey:
Princeton University Press.
Kuhn, S.L. and D.S. Miller 2015 Artifacts as patches: the marginal value theorem
and stone tool life histories. In N. Goodale and W. Andrefsky (eds), Lithic
technological systems and evolutionary theory, pp.172–197. New York: Cambridge
University Press.
361
Laing, I.A.F. and E.J. Hauck 1997 Water harvesting from granite outcrops in Western
Australia. Journal of the Royal Society of Western Australia 80(3):181–184.
Landgate 2015 Historical town names. Retrieved 20 February 2016 from
< http://www0.landgate.wa.gov.au/maps-and-imagery/wa-geographic-names/namehistory/historical-town-names#H >
Landor, H. and H.M. Lefroy 1843 Landor and Lefroy’s expedition to the south-east of
York, January 1843 (transcript). In M. Herlock and S. Milentis (eds), Western
Australian exploration 1836–1845, pp.327–333. Perth: Hesperian Press.
Langford-Smith, T. 1978 A select review of silcrete research in Australia. In T.
Langford-Smith (ed.), Silcrete in Australia, pp.1–11. Armidale: University of New
England.
Laskar, J., P. Robutel, F. Joutel, M. Gastineau, A.C.M. Correia and B. Levard 2004 A
long term numerical solution for the insolation quantities of the Earth. Astronomy and
Astrophysics 428(1):261–285.
Leake, B.W. 1950. Countryman. The Western Mail Supplement, 17 August 1950.
Leigh, J., R. Boden and J. Briggs 1984 Extinct and endangered plants of Australia.
Melbourne: MacMillan.
Lewis, S.E., C.R. Sloss, C.V. Murray-Wallace, C.D. Woodroffe and S.G. Smithers
2013 Post-glacial sea-level changes around the Australian margin: a review.
Quaternary Science Reviews 74:115–138.
Lin, S.C., C.M. Pop, H.L. Dibble, W. Archer, D. Desta, M. Weiss and S.P. McPherron
2016 A core reduction experiment finds no effect of original stone size and reduction
intensity on flake debris size distribution. American Antiquity 81(3):562–575.
Low, T. 1991 Wild food plants of Australia. Sydney: HarperCollins.
Lowe, L.D., J.A. Webb, R.J. Nathan, T. Etchells, H.M. Malano 2009 Evaporation
from water supply reservoirs: an assessment of uncertainty. Journal of Hydrology
376:261–274.
Luke, G.J., K.L. Burke and T.M. O’Brien 1987 Evaporation data for Western
Australia. Resource Management Technical Report 65. Perth: Department of
Agriculture and Food.
Lullfitz, A., J. Dortch, S.D. Hopper, C. Pettersen, R. Reynolds and D. Guilfoyle 2017
Human niche construction: Noongar evidence in pre-colonial southwestern Australia.
Conservation and Society 15(2):201–216.
MacArthur, R.H. and E.R. Pianka 1966 On optimal use of a patchy environment. The
American Naturalist 100(916):603–609.
362
Macintyre, K., B. Dobson, J. Harris and G. Quartermaine 1992 Report on an
ethnographic and archaeological survey of the proposed KondininBounty/Forrestania mines power line. Unpublished report prepared for the State
Energy Commission of Western Australia (SECWA).
Mackay, A. 2005 Informal movements: changing mobility patterns at Ngarrabullgan,
Cape York Australia. In C. Clarkson and L. Lamb (eds), Lithics' down under':
Australian perspectives on lithic reduction, use and classification, pp.95–107. BAR
International Series 1408. Oxford: Archaeopress.
Magee, J.W. 2009 Palaeovalley groundwater resources in arid and semi-arid
Australia: a literature review. Geoscience Australia Record 2009/03. Canberra:
National Water Commission and Geoscience Australia.
Manninen, M.A. and K. Knutsson 2014 Lithic raw material diversification as an
adaptive strategy – technology, mobility, and site structure in Late Mesolithic
northernmost Europe. Journal of Anthropological Archaeology 33:84–98.
Martin, H.A. 2006 Cenozoic climatic change and the development of the arid
vegetation in Australia. Journal of Arid Environments 66(3):533–563.
Martindale, A. and I. Jurakic 2006 Identifying expedient glass tools from a postcontact Tsimshian village using low power (10–100×) magnification. Journal of
Archaeological Science 33(3):414–427.
Martinez Cortizas, A. and C. Llana 1996 Morphostructural variables and the analysis
of their effects on quartz blank characteristics. In N. Moloney, L. Raposo and M.
Santonja (eds), Non-flint stone tools and the Palaeolithic occupation of the Iberian
Peninsula, pp.49–53. BAR International Series 649. Oxford: Archaeopress.
Mayewski, P.A., E.E. Rohling, J.C. Stager, W. Karlen, K.A. Maasch, L.D. Meeker,
E.A. Meyerson, F. Gasse, S. van Kreveld, K. Holmgren, J. Lee-Thorp, G. Rosqvist,
F. Rack, M. Staubwasser, R.R. Schneider and E.J. Steig 2004. Holocene climate
variability. Quaternary Research 62:243–255.
McBryde, I. 1974. Aboriginal Prehistory in New England: an archaeological survey of
Northeastern New South Wales. Sydney: Sydney University Press.
McBryde, I. 1979 Ethnohistory in an Australian context: independent discipline or
convenient data quarry? Aboriginal History 3(1/2):128–151.
McBryde, I. 1986 Australia’s Once and Future Archaeology. Archaeology in Oceania
21(1):13–28.
McDonald, J.L., D. Donlon, J.H. Field, R.L.K. Fullagar, J. Brenner Coltrain, P.
Mitchell and M. Rawson 2007 The first archaeological evidence for death by
spearing in Australia. Antiquity 81:877–885.
363
McDonald, J., W. Renyen, F. Petchey, K. Ditchfield, C. Byrne. D. Vannieuwenhuyse,
M. Leopold and P. Veth 2018 Karnatukul (Serpent’s Glen): a new chronology for the
oldest site in Australia’s Western Desert. PLOS One 13(9):e0202511.
https://doi.org/10.1371/journal.pone.0202511
McKenzie, N.L., A.A. Burbidge and N.G. Marchant 1973 Results of a biological
survey of a proposed wildlife sanctuary at Dragon Rocks near Hyden, Western
Australia. Department of Fisheries and Fauna Western Australia Report no. 12.
Perth: Department of Fisheries and Fauna.
McKenzie, N.L., N. Gibson, G.J. Keighery and J.K. Rolfe 2004 Patterns in the
biodiversity of terrestrial environments in the Western Australian wheatbelt. Records
of the Western Australian Museum Supplement 67:293–335.
McMahon, T.A., M.C. Peel, L. Lowe, R. Srikanthan and T.R. McVicar 2013
Estimating actual, potential, reference crop and pan evaporation using standard
meteorological data: a pragmatic synthesis. Hydrology and Earth System Sciences
17:1331–1363.
Meagher, S.J. 1974 The food resources of the Aborigines of the south-west of
Western Australia. Records of the Western Australian Museum 3:14–65.
Meeking, J. 1979. The history of Hyden. Northam: Mid-West Print.
Mitchell, M.B. 2016 The Esperance Nyungars, at the Frontier: an archaeological
investigation of mobility, aggregation and identity in late-Holocene Aboriginal society,
Western Australia. Unpublished PhD thesis, The Australian National University,
Canberra.
Mitchell, P.J., E. Veneklaas, H. Lambers, S.S.O. Burgess 2009 Partitioning of
evapotranspiration in a semi-arid eucalupt woodland in south-western Australia.
Agricultural and Forest Meteorology 149:25–37.
Mithen, S.J. 1989 Modeling hunter-gatherer decision making: complementing
Optimal Foraging Theory. Human Ecology 17(1):59–83.
Moore, G. (ed.) 2001 Soil guide: a handbook for understanding and managing
agricultural soils. Agriculture Western Australia Bulletin No. 4343. Perth: National
Landcare and Agriculture Western Australia.
Moore, G.F. 1884 Diary of ten years eventful life of an early settler in Western
Australia, and also a descriptive vocabulary of the language of the Aborigines.
London: Walbrook.
Moros, M., P. De Dekker, E. Jansen, K. Perner and R.J. Telford 2009 Holocene
climate variability in the Southern Ocean recorded in a deep-sea sediment core off
South Australia. Quaternary Science Reviews 28(19):1932–1940.
Morse, K., R. Cameron and W. Reynen 2014 A tale of three caves: new dates for
Pleistocene occupation in the inland Pilbara. Australian Archaeology 79:167–178.
364
Mouritz, M. 1986. Only the wind was free: 60 years in Wave Rock country. Canberra:
National Library of Australia.
Muir, B.G. 1977 Biological survey of the Western Australian wheatbelt, part 2:
vegetation and habitat of Bendering Reserve. Records of the Western Australian
Museum Supplement 3:1–142.
Muir, B.G. 1978 Some nature reserves of the Western Australian wheatbelt, part 4:
Narembeen Shire. Unpublished report prepared for the Department of Fisheries and
Wildlife.
Muir, B.G. 1979 Some nature reserves of the Western Australian wheatbelt, part 20:
Kondinin Shire. Unpublished report prepared for the Department of Fisheries and
Wildlife.
Murphy-White, S. 2007 Farm water for the north-eastern and eastern wheatbelt of
Western Australia (zones 2 and 5). Resource Management Technical Report 318.
Perth: Department of Agriculture and Food.
Murphy-White, S. and E. Hauck 1999 Agricultural water supply areas of the northeastern and eastern wheatbelt, Western Australia. In Water 99 joint congress –
Brisbane, Australia 6–8 July 1999. Barton: Institute of Engineers, Australia.
Murphy-White, S. and P. Leoni 2006 Lockhart catchment appraisal 2005. Resource
Management Technical Report 310. Perth: Department of Agriculture and Food.
Myers, F.R. 1982 Always ask: resource use and land ownership among Pintupi
Aborigines of the Australian Western Desert. In N.M. Williams and E.S. Hunn (eds),
Resource managers, pp.173–196. Canberra: Australian Institute of Aboriginal
Studies.
Nanson, G.C., D.M. Price and S.A. Short 1992 Wetting and drying of Australia over
the past 300 ka. Geology 20(9):791–794.
Native Grass Resources Group (n.d.) Native grass propagation. Retrieved 27 March
2017 from < https://nativegrassresourcesgroup.files.wordpress.com/2016/08/nativegrass-propagation.pdf >
Nelson, M.C. 1991 The study of technological organization. Archaeological Method
and Theory 3:57–100.
Nettle, D., M.A. Gibson, D.W. Lawson and R. Sear 2013 Human behavioural ecology:
current research and future prospects. Behavioral Ecology 24(5):1031–1040.
https://doi.org/10.1093/beheco/ars222
Nevill, S. 2014 Guide to the wildlife of Perth and the south west. Perth: Simon Nevill
Publications.
Neville, A.O. 1948 Contributory causes of Aboriginal depopulation in Western
Australia. Mankind 4:3–13.
365
Newsome, J.C. and E.J. Pickett 1993 Palynology and palaeoclimatic implications of
two Holocene sequences from southwestern Australia. Palaeogeography,
Palaeoclimatology, Palaeoecology 101(3):245–261.
NHMRC, NRMMC 2011 Australian drinking water guidelines paper 6 national water
quality management strategy. Canberra: National Health and Medical Research
Council and National Resource Management Ministerial Council.
Nussinovitich, A. 2010 Plant gum exudates of the world: sources, distribution,
properties, and applications. Boca Raton: CRC Press.
O’Connell, J.F. and K. Hawkes 1981 Alyawara plant use and optimal foraging theory.
In B. Winterhalder and E.A. Smith (eds), Hunter-gatherer foraging strategies:
ethnographic and archaeological analyses, pp.99–125. Chicago: University of
Chicago Press.
O’Connell, J.F., P.K. Latz and P. Barnett 1983 Traditional and modern plant use
among the Alyawara of Central Australia. Economic Botany 37(1):80–109.
O’Connor, M.H. and S.M. Prober, S.M. 2010 A calendar of Ngadju seasonal
knowledge. Floreat: CSIRO Sustainable Ecosystems.
O'Connor, R., G. Quartermaine and C. Bodney 1989 Report on an investigation into
the Aboriginal significance of wetlands and rivers in the Perth-Bunbury region. Perth:
Western Australian Water Resources Council.
Oliver, Y.M. and M.J. Robertson 2009 Quantifying the benefits of accounting for yield
potential in spatially and seasonally responsive nutrient management in a
Mediterranean climate. Australian Journal of Soil Research 47:114–126.
Orians, G.H. and N.E. Pearson 1979 On the theory of central place foraging. In D.J.
Horn, R.D. Mitchell, and G.R. Stairs (eds), Analysis of ecological systems, pp.154–
177. Columbus: Ohio State University.
Paczkowska, G. and A.R. Chapman 2000 The Western Australian flora: a descriptive
catalogue. Perth: Wildflower Society of Western Australia, Western Australian
Herbarium, CALM and the Botanic Gardens and Parks Authority.
Pargeter, J. and P. de la Pena 2017 Milky quartz bipolar reduction and lithic
miniaturization: experimental results and archaeological implications. Journal of Field
Archaeology 42(6):551–565. https://doi.org/10.1080/00934690.2017.1391649
Pargeter, J. and M.I. Eren 2017 Quantifying and comparing bipolar versus freehand
flake morphologies, production currencies, and reduction energetics during lithic
miniaturization. Lithic Technology 2017.
https://doi.org/10.1080/01977261.2017.1345442
Parry, W. and R.J. Kelly 1987 Expedient core technology and sedentism. In J.K.
Johnson and C.A. Morrow (eds), The organisation of core technology, pp.285–304.
Boulder: Westview Press.
366
Pascoe, B. 2013 Dark emu, black seeds: agriculture or accident? Broome: Magabala
Books.
Pate, J.S. and K.W. Dixon 1982 Tuberous, cormous and bulbous plants: biology of
an adaptive strategy in Western Australia. Crawley: University of Western Australia
Press.
Paterson, A. 1999 Confronting the sources: the archaeology of culture-contact in the
south-western Lake Eyre Basin, Central Australia. Unpublished PhD Thesis,
Department of Archaeology, University of Sydney, Camberdown/Darlington.
Pearce, R.H. and M. Barbetti 1981 A 38,000-year-old archaeological site at Upper
Swan, Western Australia. Archaeology in Oceania 16:173–178.
Peck, C. 1982 Monitoring and conservation work on sites in south-west – September
1982. Unpublished report prepared for the Western Australian Museum.
Pereira, L.S., R.G. Allen, M. Smith and D. Raes 2015 Crop evapotranspiration
estimation with FAO56: Past and future. Agricultural Water Management 147:4–20.
Perkins, D. 2011 Mineralogy (3rd ed.). Essex: Pearson.
Petherick, L., H. Bostock, T.J. Cohen, K. Fitzsimmons, J. Tibby, M.-S. Fletcher, P.
Moss, J. Reeves, S. Mooney, T. Barrows, J. Kemp, J. Jansen, G. Nanson, A.
Dosseto 2013 Climatic records over the past 30 ka from temperate Australia – a
synthesis from the Oz-INTIMATE workgroup. Quaternary Science Reviews 74:58–
77.
Pickett, E.J. 1997 The late Pleistocene and Holocene vegetation history of three
lacustrine sequences from the Swan Coastal Plain, southwestern Australia.
Unpublished PhD thesis, Department of Geography, University of Western Australia,
Crawley.
Pickett, E.J., S.P. Harrison, G. Hope, K. Harle, J.R. Dodson, A.P. Kershaw, I.C.
Prentice, J. Backhouse, E.A. Colhoun, D. D’Costa, J. Flenley, J. Grindrod, S.
Haberle, C. Hassell, C. Kenyon, M. Macphail, H. Martin, A.H. Martin, M. McKenzie,
J.C. Newsome, D. Penny, J. Powell, J.I. Raine, W. Southern, J. Stevenson, J-P.
Sutra, I. Thomas, S. van der Kaars, J. Ward 2004 Pollen-based reconstructions of
biome distributions for Australia, Southeast Asia and the Pacific (SEAPAC region) at
0, 6000 and 18,000 14C yr BP. Journal of Biogeography 31(9):1381–1444.
Pinder. A.M., S.A. Halse, J.M. McRae and R.J. Shiel 2004 Aquatic invertebrate
assemblages of wetlands and rivers in the Wheatbelt region of Western Australia.
Records of the Western Australian Museum Supplement 67:7–37.
Prober, S.M., E. Yuen, M.H. O'Connor and L. Schultz 2016 Ngadju kala: Australian
Aboriginal fire knowledge in the Great Western Woodlands. Austral Ecology 41:716–
732.
367
Quartermaine, G. 2000 Report on an archaeological investigation of Aboriginal sites
Corrigin to Hyden section of the Brookton Highway. Unpublished report prepared for
Halpern Glick Maunsell.
Quigley, M.C., T. Horton, J.C. Hellstrom, M.L. Cupper and M. Sandiford 2010
Holocene climate change in arid Australian from speleothem and alluvial records.
The Holocene 20(7):1093–1104.
Radcliffe-Brown, A.R. 1930 The social organization of Australian Tribes (part 1).
Oceania 1(1):34–63.
Rafferty. J.P. (ed.) 2012 Minerals. New York: Brittanica Educational Publishing.
Ranatunga, K., E.R. Nation and D.G. Barratt 2008 Review of soil water models and
their applications in Australia. Environmental Modelling and Software 23(9):1182–
1206.
Randolph, P. 1973. Bates Cave, Hyden. Unpublished report prepared for the
Department of Aboriginal Sites, Western Australian Museum, Perth.
Reeves, J.M., T.D. Barrows, T.J. Cohen, A.S. Kiem, H.C. Bostock, K.E.
Fitzsimmons, J.D. Jansen, J. Kemp, C. Krause, L. Petherick, S.J. Phipps 2013
Climate variability over the last 35,000 years recorded in marine and terrestrial
archives in the Australian region: an OZ-INTIMATE compilation. Quaternary Science
Reviews 74:21–34.
Rick, A. 2011 Survey and analysis of plant communities growing on gypsum in the
Western Australian wheatbelt. Unpublished report prepared for the Wheatbelt NRM
Region and the Department of Environment and Conservation, Western Australia.
Ritchie, J.T. 1972 Model for predicting evaporation from a row crop with incomplete
cover. Water Resources Research 8(5):1204–1213.
Robertson, G. 2005. Backed artefact use in eastern Australia: a residue and usewear analysis. Unpublished PhD thesis, School of Social Science, University of
Queensland, St Lucia.
Robertson, G. 2009 Aboriginal craft and subsistence activities at Native Well I and
Native Well II, Central Western Highlands, Queensland: results of a residue and usewear analysis of backed artefacts. In M. Haslam, G. Robertson, A. Crowther, S.
Nugent and L. Kirkwood (eds), Archaeological science under a microscope: studies
in residue and ancient DNA analysis in honour of Thomas H. Loy, pp.239–257. Terra
Australis 30. Canberra: ANU E Press.
Robertson, G. and V. Attenbrow 2008 Skin-working at Emu Tracks 2, New South
Wales, Australia: an integrated residue and use-wear analysis of backed artefacts.
Lithic Technology 33:31–49. https://doi.org/10.1080/01977261.2008.11721059
Robertson, G., V. Attenbrow and P. Hiscock 2009 Multiple uses for Australian
backed artefacts. Antiquity 83:296–308.
368
Roe, J.S. 1836. Explorations east of York to Wongan Hills, October to November
1836, field books (transcript). In M. Hercock (ed.), The Western Australian
explorations of John Septiumus Roe 1829–1849, pp.236–297. Perth: Hesperian
Press.
Roe, J.S. 1852 Report of an expedition under the Surveyor-General, Mr J.S. Roe, to
the south-eastward of Perth, in Western Australia, between the months of
September, 1848, and February, 1849, to the Hon. The Colonial Secretary. Journal
of the Royal Geographical Society of London 22:1–57.
Rolland, N. and H.L. Dibble 1990 A new synthesis of Middle Palaeolithic variability.
American Antiquity 55(3):480–499.
Rossi, A.M. 2010 An archaeological re-investigation of the Mulka’s Cave Aboriginal
rock art site, near Hyden, southwestern Australia. Unpublished MA thesis, School of
Arts, Edith Cowan University, Joondalup.
Rossi, A.M. 2014a Re-evaluating the antiquity of Aboriginal occupation at Mulka’s
Cave. Australian Archaeology 78:39–46.
Rossi, A.M. 2014b Unconventional proxies: using archaeological measures of
occupation intensity to infer Holocene climatic variability in inland southwestern
Australia. Paper presented at the Australian Archaeological Association conference,
Cairns.
Rossi, A.M. and R.E. Webb 2007 The consequences of allowing unfettered tourist
access at an Aboriginal site in a fragile environment: the erosive effect of trampling.
Conservation and Management of Archaeological Sites 9:219–236.
Rossi, A.M. and Webb, R.E. 2008 The erosive effect of tourism at an Aboriginal rock
art site on the western edge of the arid zone in south-western Australia. Antiquity 82
Project Gallery. Retrieved 10 January 2014 from
< http://www.antiquity.ac.uk/projgall/rossi315/ >
Runnels, C. 1976 More on glass implements from Greece. Newsletter of Lithic
Technology 5(3):27–31.
Salama, R.B. 1994 The evolution of saline lakes in the relict drainage of the Yilgarn
River, Western Australia. In R.W. Renaut and W.M. Last (eds), Sedimentology and
geochemistry of modern and ancient saline lakes, pp.189–199. SEPM Special
Publication no. 50. Tulsa: Society for Sedimentary Geology.
Saunders, D.A. 1989 Changes in the avifauna of a region, district and remnant as a
result of fragmentation of native vegetation: the wheatbelt of Western Australia. A
case study. Biological Conservation 50:99–135.
Saunders, D.A. and J.A. Ingram 1995 Birds of southwestern Australia: an atlas of
changes in distribution and abundance of wheatbelt avifauna. Chipping Norton:
Surrey Beatty and Sons.
369
Scadding, R., V. Winton and V. Brown 2015 An LA-ICP-MS trace element
classification of ochres in the Weld Range environ, Mid West region, Western
Australia.
Journal
of
Archaeological
Science
54:300–312.
https://doi.org/10.1016/j.jas.2014.11.017
Schoknecht, N.R. and S. Pathan 2013 Soil groups of Western Australia: a simple
guide to the main soils of Western Australia (4th ed.). Resource Management
Technical Report 380. Perth: Department of Agriculture and Food.
Schut, A.G.T., G.W. Wardell-Johnson, C.J. Yates, G. Keppel. I. Barab, S.E. Franklin,
S.D. Hopper, K.P. Van Niel, L. Mucina and M. Byrne 2014 Rapid characterisation of
vegetation structure to predict refugia and climate change impacts across a global
biodiversity hotspot. PLoS ONE 9(1):e82778.
https://dx.doi.org/10.1371/journal.pone.0082778
Semeniuk, V. 1986 Holocene climate history of coastal southwestern Australia using
calcrete as an indicator. Palaeogeography, Palaeoclimatology, Palaeoecology
53(2):289–308.
Serventy, V.L. 1952 Cave paintings near York and Hyden. The Western Australian
Naturalist 3:121–131.
Shaw, E. 2014 Hydrology in Practice (3rd ed.). London: CRC Press.
Shennan, S.J. 2001 Demography and cultural innovation: a model and some
implications for the emergence of modern human culture. Cambridge Archaeological
Journal 11:5–16.
Shott, M. 1986 Technological organization and settlement mobility: an ethnographic
examination. Journal of Anthropological Research 42(1):15–51.
Simpson, K. and N. Day 2004 Field guide to the birds of Australia (7th ed.).
Campberwell: Viking.
Slack, M.J., R.L.K. Fullagar, J.M. Field and A. Border 2004 New Pleistocene ages for
backed artefact technology in Australia. Archaeology in Oceania 39:131–137.
Smith, E.A. 1979 Human adaptation and energetic efficiency. Human Ecology
7(1):53–74.
Smith, E.A. 1983 Anthropological applications of optimal foraging theory: a critical
review (including comments and reply). Current Anthropology 24(5):625–651.
Smith, M.A. 2013 The archaeology of Australia's deserts. Cambridge: Cambridge
University Press.
Smith, M.V. 1993 Recherche à l’Espérance: a prehistory of the Esperance region of
South-Western Australia. Unpublished PhD thesis, Department of Anthropology,
University of Western Australia, Crawley.
370
Smith, M.V. 2011 Moving on: an archaeological record of mobility in the Esperance
area of south-western Australia. Records of the Western Australian Museum
Supplement 79:16–29.
Speal, C.S. 2009 The economic geography of chert lithic production in the southern
Maya lowlands: a comparative examination of early-stage reduction debris. Latin
American Antiquity 20(1):91–119.
Stannage, C.T. (ed.) 1981. A new history of Western Australia. Crawley: University
of Western Australia Press.
Stanner, W.E.H. 1965 Aboriginal territorial organization: estate, range, domain and
regime. Oceania 36(1):1–26.
Staub, A. and J. Taylor 1999 Report on an Aboriginal heritage survey: proposed
Hyden to Kondinin water pipeline. Unpublished report prepared for WML consultants.
Storr, G.M. 1991 Birds of the south-west division of Western Australia. Records of
the Western Australian Museum Supplement 35:1–150.
Sullivan, A.P. and K.C. Rozen 1985. Debitage analysis and archaeological
interpretation. American Antiquity 50(4):755–779.
Suzuki, H., Y. Uesugi, K. Endo, H. Ohmori, K. Takeuchi and K. Twasaki 1982.
Studies in the Holocene and recent climatic fluctuations in Australia and New
Zealand. Tokyo: University of Tokyo Press.
Sweedman, L. and D. Merritt 2006 Australian seeds: a guide to their collection,
identification and biology. Collingwood: CSIRO Publishing.
Taipale, N., K. Knutsson and H. Knutsson 2014 Unmodified quartz flake fragments
as cognitive tool categories: testing the wear preservation, previous low
magnification use-wear results and criteria for tool blank selection in two Late
Mesolithic quartz assemblages from Finland. In J. Marreiros, N. Bicho and J.F.
Gibaja (eds), International conference on use-wear analysis: use-wear 2012,
pp.352–361. Newcastle upon Tyne: Cambridge Scholars Publishing.
Tallavaara, M., M.A. Manninen, E. Hertell and T. Rankama 2010 How flakes shatter:
a critical evaluation of quartz fracture analysis. Journal of Archaeological Science
37:2442–2448.
Tanny, J., S. Cohen, S. Assouline, F. Lange, A. Grava, D. Berger, BB. Teltch and
M.B. Parlange 2008 Evaporation from a small water reservoir: direct measurements
and estimates. Journal of Hydrology 351:218–229.
Taylor, G. and R.A. Eggleton 2017 Silcrete: an Australian perspective. Australian
Journal of Earth Sciences 2017. https://doi.org/10.1080/08120099.2017.1318167
Timms, B.V. 2013 Geomorphology of pit gnammas in southwestern Australia.
Journal of the Royal Society of Western Australia 96:7–16.
371
Tindale, N. 1941 A Tasmanian stone implement made from bottle glass. Papers and
Proceedings of the Royal Society of Tasmania 1941:1–2.
Tindale, N.B. 1974 Aboriginal tribes of Australia: their terrain, environmental controls,
distribution, limits and proper names. Berkeley: University of California Press.
Tonkinson, R. 1978 The Mardudjara Aborigines: living the dream in Australia's
Western Desert. New York: Holt, Rinehart and Winston.
Torrence, R. 1983 Time budgeting and hunter-gatherer technology. In G. Bailey
(ed.), Hunter-gatherer economy in prehistory, pp.11–22. Cambridge: Cambridge
University Press.
Torrence, R. 1989 Retooling: towards a behavioral theory of stone tools. In R.
Torrence (ed.), Time, energy and stone tools, pp.57–66. Cambridge: Cambridge
University Press.
Torrence, R. 2001 Hunter-gatherer technology: macro- and microscale approaches.
In C. Panter-Brick, R.H. Layton and P. Rowley-Conwy (eds), Hunter-gatherers: an
interdisciplinary perspective, pp.73–98. Cambridge: Cambridge University Press.
Treble, P.C., J.M.G. Shelley and J. Chappell 2003 Comparison of high-resolution
sub-annual records of trace elements in a modern (1911–1992) speleothem with
instrumental climate data from southwest Australia. Earth and Planetary Science
Letters 216:141–153.
Trigger, B.G. 1982 Ethnohistory: problems and prospects. Ethnohistory 29(1):1–19.
Turney, C.S.M., M. Bird, L.K. Fifield, R.G. Roberts, M. Smith, C.E. Dortch, R. Grun,
E. Lawson, L.K. Ayliffe, G.H. Miller, J. Dortch and R.G. Cresswell 2001 Early human
occupation at Devils Lair, southwestern Australia, 50,000 years ago. Quaternary
Research 55:3–13.
Turney, C.S.M., S. Haberle, D. Fink, A.P. Kershaw, M. Barbetti, T.T. Barrows, M.
Black, T.J. Cohen, T. Corrège, P.P. Hesse, Q. Hua, R. Johnston, V. Morgan, P.
Moss, G. Nanson, T. Van Ommen, S. Rule, N.J. Williams, J-X. Zhao, D. D’Costa, YX. Feng, M. Gagan, S. Mooney and Q. Xia 2006 Integration of ice-core, marine and
terrestrial records for the Australian Last Glacial Maximum and Termination: a
contribution from the OZ INTIMATE group. Journal of Quaternary Science
21(7):751–761.
Veth, P. 1987 Martutjarra prehistory: variation in arid zone adaptations. Australian
Archaeology 25:102–111.
Veth, P. and S. O’Connor 2005 Archaeology, claimant connection to sites, and
Native Title: employment of successful categories of data with specific comments on
glass artifacts. Australian Aboriginal Studies 2005(1):2–15.
372
Wallis, L., J. Huntley, M. Marsh, A. Watchman and A. Ewen 2016 PXRF analysis of a
yellow ochre quarry and rock art motifs in the Central Pilbara. Journal of the
Anthropological Society of South Australia 40:134–155.
Wanner, H., J. Beer, J. Bütikofer, T.J. Crowley, U. Cubasch, J. Flückiger, H. Goosse,
M. Grosjean, F. Joos, J.O. Kaplan, M. Küttel, S.A. Müller, I.C. Prentice, O. Solomina,
T.F. Stocker, P. Tarasov, M. Wagner and M. Widmann 2008 Mid- to Late Holocene
climate change: an overview. Quaternary Science Reviews 27(19):1,791–1,828.
Watts, I. 2010 The pigments from Pinnacle Point Cave 13B, Western Cape, South
Africa. Journal of Human Evolution 59:392–411.
https://dx.doi.org/10.1016/j.jhevol.2010.07.006
Webb, R.E. 2000 Reconstructing the ethnobotany of a depleted flora: food plant
availability in the Murchison Basin, Western Australia, prior to European arrival. In
R.A. Nicholson and T.P. O’Connor (eds), People as an agent of environmental
change, pp.92–116. Oxford: Oxbow Books.
Webb, R.E. and R.G. Gunn 2004 Re-recording culturally significant Aboriginal sites
in south-western Australia as a guide to Noongar usage of the region in the past.
Rock Art Research 21:93–97.
Webb, R.E. and A.M. Rossi 2008 How was Mulka’s Cave, an Aboriginal rock art site
near Hyden, in south-central Western Australia, used by the people who decorated
its walls, when the present entrance was much smaller? Records of the Western
Australian Museum 24:307–318.
Western Australian Herbarium 1998- Florabase – the Western Australian Flora
[Computer Software]. Retrieved from < https://florabase.dpaw.wa.gov.au/ >
Western Australian Museum n.d. Frogwatch – Frogs – Retrieved 27 March 2017
from < http://museum.wa.gov.au/explore/frogwatch/frogs/ >
White, I. (ed.) 1985 Daisy Bates: the native tribes of Western Australia. Canberra:
National Library of Australia.
White, P. 2011 Changing perspectives in Australian archaeology, part I. Regional
archaeology in Australia. Technical Reports of the Australian Museum Online
23(1):3–5.
Whittaker, J. C. 1994 Flintknapping: making and understanding stone tools. Austin:
University of Texas Press.
Wildlife Society of Western Australia 2013 Seed notes for Western Australia.
Retrieved 27 March 2017 from < https://www.dpaw.wa.gov.au/about-us/science-andresearch/publications-resources/114-seed-notes-for-western-australia >
Wildy, D.T., J.S. Pate and J.R. Bartle 2004 Budgets of water use by Eucalyptus
kochii tree belts in the semi-arid wheatbelt of Western Australia. Plant and Soil
262:129–149.
373
Willey, O. 1950 In search of the wheatbelt cave of Mulka the terrible. The West
Australian Saturday 28 October 1950 p. 18.
Williams, A.N. and M.A. Smith 2012 AustArch 2: A database of 14C and
luminescence ages from archaeological sites in the Top End. Australian Archaeology
74:146.
Williams, A.N. and M.A. Smith 2013 AustArch 3: A database of 14C and
luminescence ages from archaeological sites in southern Australia. Australian
Archaeology 76:102.
Williams, A.N., M.A. Smith, C.S.M Turney and M. Cupper 2008 AustArch1: A
database of 14C and luminescence ages from archaeological sites in the Australian
arid zone. Australian Archaeology 66:99.
Williams, A.N., S. Ulm, M. Smith and J. Reid 2014 AustArch: A database of 14C and
non-14C ages from archaeological sites in Australia – composition, compilation and
review (data paper). Internet Archaeology 36. https://dx.doi.org/10.11141/ia.36.6
Williams, M., E. Cook, S. van der Kaars, T. Barrows, J. Shulmeister and P. Kershaw
2009 Glacial and deglacial climatic patterns in Australia and surrounding regions
from 35,000 to 10,000 years ago reconstructed from terrestrial and near-shore proxy
data. Quaternary Science Reviews 28(23–24):2398–2419.
Williamson, C. 2004 Finding meaning in the patterns: the analysis of material culture
from a contact site in Tasmania. In R. Harrison and C. Williamson (eds), After
Captain Cook: the archaeology of the recent Indigenous past in Australia, pp.75–102.
Sydney: Archaeological Computing Laboratory.
Wilmsen, E.N. 1973 Interaction, spacing behavior, and the organization of hunting
bands. Journal of Anthropological Research 29(1):1–31.
Winmar, R. 1996 Walwalinj: the hill that cries. Nyungar language and culture. Perth:
Dorothy Winmar.
Winter, T.C. 1981 Uncertainties in estimating the water balance of lakes. Journal of
the
American
Water
Resources
Association
17(1):82–115.
https://dx.doi.org/10.1111/j.1752-1688.1981.tb02593.x
Winterhalder, B. 1981 Optimal foraging strategies and hunter-gatherer research in
anthropology: theory and models. In B. Winterhalder and E.A. Smith (eds), Huntergatherer foraging strategies: ethnographic and archaeological analyses, pp.13–35.
Chicago: University Press.
Winterhalder, B. and E.A. Smith 2000 Analyzing adaptive strategies: human
behavioural ecology at twenty-five. Evolutionary Anthropology 9(2):51–72.
Witt, G.B., J. Luly and R.J. Fairfax 2006 How the west was once: vegetation change
in southwest Queensland from 1930–1995. Journal of Biogeography 33(9):1,585–
1,596.
374
Witter, D. 1990. Regions and resources. Unpublished PhD thesis, Prehistory
Department, Research School of Pacific Studies, Australian National University,
Canberra.
Woinarski, J.C.Z., A.A. Burbidge and P.L. Harrison 2015 Ongoing unravelling of a
continental fauna: decline and extinction of Australian mammals since European
settlement. Proceedings of the National Academy of Sciences 112:4531–4540.
https://dx.doi.org/10.1073/pnas.1417301112
Wolski, N. and T.H. Loy 1999 On the invisibility of contact: residue analyses on
Aboriginal glass artefacts from western Victoria. The Artefact 22:65–73.
Yassini, I. and G.W. Kendrick 1988 Middle Holocene ostracodes, foraminifers and
environments of beds at Point Waylen, Swan River Estuary, southwestern Australia.
Alcheringa 12(2):107–121.
Zheng, H., C.M. Powell and H. Zhao 2002 Eolian and lacustrine evidence of late
Quaternary palaeoenvironmental changes in southwestern Australia. Global and
Planetary Change 35(1):75–92.
375
APPENDIX A
FLORAL SPECIES PRESENT IN THE STUDY AREA
The floral list was compiled using literature derived from surveys and analyses
performed within the study area (Beard 1972, 1979; Coates 1992; Department of
Water 2009; Gunness 2002; Halse et al. 1993; Hopper 2001; Muir 1977, 1978, 1979;
Rick 2011). Species are listed alphabetically, with taxonomy and naming conventions
as per Florabase (Western Australian Herbarium (1998-). Any non-current names
encountered in the references above have been updated; taxonomic changes made
after 1 April 2017 will not be reflected. A cross in the relevant column indicates that a
plant species was found in that particular Landscape Division. Genus-level
identifications (e.g. Acacia sp., Eucalyptus spp. – assumed to represent two distinct
species) have generally been omitted, based on the assumption that the relevant
species was likely identified in another source; they were retained or amended in
certain circumstances, as below:
o
Entries were retained where the genus-level identification/s was the only
representation of a genus in a particular Landscape Division;
o
Spp. was converted to sp. where the floral list included just one named
species of the same genus in a particular Landscape Division, i.e. the named
species could not possibly incorporate all of the genus-level identifications.
Landscape Divisions have been abbreviated, as below:
o
GR = Granite
o
HEA = Heath
o
MAL = Mallee
o
SAL = Saline
o
TH = Thicket
o
WL = Woodland
376
Species
Acacia acanthoclada
GR
HEA
MAL
SAL
X
X
X
X
Acacia acuaria
X
Acacia acutata
X
X
Acacia ancistrophylla
X
Acacia anfractuosa
Acacia assimilis
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
Acacia bidentata
X
X
Acacia brachyclada
X
Acacia brachyphylla
X
Acacia celastrifolia
X
X
X
Acacia chrysocephala
X
Acacia colletioides
X
Acacia consanguinea
X
X
Acacia coolgardiensis
X
Acacia cupularis
X
Acacia densiflora
Acacia dentifera
X
X
Acacia dermatophylla
X
Acacia desertorum
X
X
Acacia dielsii
X
X
Acacia ephedroides
X
Acacia eremophila
X
X
Acacia erinacea
X
X
X
X
X
Acacia filifolia
X
Acacia fragilis
Acacia hemiteles
X
X
Acacia ericifolia
Acacia fauntleroyi
X
X
X
X
X
X
X
Acacia inceana
X
X
Acacia lasiocarpa
X
X
X
X
X
Acacia leptopetala
X
Acacia leptospermoides
X
Acacia lineolata
X
Acacia lirellata
X
Acacia mackeyana
X
Acacia merinthophora
X
Acacia merrallii
Acacia microbotrya
X
X
Acacia intricata
Acacia lasiocalyx
X
X
X
Acacia beauverdiana
Acacia chrysella
WL
X
Acacia acuminata
Acacia andrewsii
TH
X
X
X
X
X
X
X
377
Species
Acacia mimica
GR
HEA
MAL
X
X
X
Acacia moirii
X
Acacia multispicata
X
SAL
TH
X
X
X
Acacia myrtifolia
X
Acacia neurophylla
X
X
Acacia newbeyi
X
Acacia pulchella
X
Acacia pycnocephala
X
Acacia rigens
X
Acacia rostellata
X
Acacia saligna
X
X
Acacia sessilispica
X
X
X
Acacia shuttleworthii
X
Acacia sphacelata
Acacia spinosissima
X
X
X
X
X
X
X
X
X
Acacia tetanophylla
X
Acacia tratmaniana
X
X
X
X
Acacia uncinella
X
X
Acacia verriculum
X
X
X
X
Acacia viscifolia
X
Acacia yorkrakinensis
X
Acrotriche patula
X
X
X
X
Actinobole uliginosum
X
Actinotus superbus
X
Adenanthos argyreus
Adenanthos flavidiflorus
X
X
X
Allocasuarina acutivalvis
X
X
X
Allocasuarina campestris
X
X
X
X
X
Allocasuarina corniculata
Allocasuarina huegeliana
X
Allocasuarina humilis
Allocasuarina microstachya
X
X
X
X
X
Allocasuarina pinaster
X
Allocasuarina spinosissima
X
Allocasuarina thuyoides
X
Alyxia buxifolia
X
Amphibromus nervosus
X
Amphipogon caricinus
Amphipogon debilis
X
X
Acacia signata
Acacia trigonophylla
X
X
Acacia sedifolia
X
X
X
Acacia sclerophylla
Acacia sulcata
WL
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
378
Species
GR
HEA
MAL
Amphipogon strictus
SAL
TH
X
Amyema miquelii
X
Amyema miraculosa
X
Andersonia lehmanniana
X
Andersonia parvifolia
X
Andersonia sp.
X
X
Angianthus tomentosus
X
Anigozanthos humilis
X
Anthotium rubriflorum
X
Aphelia brizula
X
Aphelia nutans
X
X
X
X
Aristida contorta
X
Astroloma epacridis
X
X
X
Astroloma serratifolium
X
X
X
Astroloma sp. Dumbleyung
X
X
X
X
Atriplex bunburyana
X
Atriplex hymenotheca
X
Atriplex paludosa
X
Atriplex semibaccata
X
X
X
X
X
X
X
Austrostipa drummondii
X
X
Austrostipa hemipogon
X
X
X
X
X
X
Austrostipa juncifolia
X
Austrostipa pycnostachya
X
X
X
Austrostipa trichophylla
X
Austrostipa variabilis
X
X
Baeckea grandibracteata
X
X
Baeckea muricata
X
Baeckea sp.
Baeckea spp.
X
X
Atriplex vesicaria
Austrostipa elegantissima
X
X
Atriplex amnicola
Austrostipa compressa
X
X
Argyroglottis turbinata
Atriplex sp.
WL
X
X
X
X
X
X
Baeckea sp. Koonadgin
X
Banksia audax
X
X
X
X
X
X
X
X
X
X
Banksia cirsioides
X
Banksia densa
X
Banksia drummondii
X
Banksia elderiana
X
Banksia erythrocephala
X
Banksia laevigata
X
X
X
X
X
379
Species
GR
TH
WL
X
X
X
X
X
X
Banksia violacea
X
X
X
Beaufortia bracteosa
X
X
X
Beaufortia micrantha
X
X
X
Beaufortia orbifolia
X
X
X
Beaufortia puberula
X
X
X
Beaufortia schaueri
X
X
X
X
Bertya sp.
X
X
X
X
Beyeria brevifolia
X
X
X
X
X
X
X
Banksia rufa
Banksia sphaerocarpa
X
HEA
MAL
X
Beyeria lechenaultii
Billardiera lehmanniana
X
Blennospora drummondii
X
X
X
Blennospora phlegmatocarpa
X
X
Boronia coerulescens
X
X
X
Boronia crassifolia
X
Boronia subsessilis
X
Boronia ternata
X
Borya constricta
X
Borya laciniata
X
Borya nitida
X
X
Borya sphaerocephala
X
X
Bossiaea concinna
X
X
X
X
X
X
X
X
X
X
X
X
Bossiaea walkeri
X
X
Brachyloma geissoloma
X
Brachyloma mogin
X
Brachyscome bellidioides
X
Brachyscome iberidifolia
X
Brachyscome perpusilla
X
Bromus arenarius
X
Bulbine semibarbata
X
Caesia micrantha
X
Caladenia denticulata
X
Caladenia dimidia
X
Caladenia doutchiae
X
Caladenia falcata
X
Caladenia flava
X
Caladenia hirta
X
Caladenia hoffmanii
X
Caladenia longicauda
X
Caladenia mesocera
X
Caladenia paradoxa
X
X
Billardiera variifolia
Boronia capitata
SAL
X
X
X
X
X
X
X
X
380
Species
GR
HEA
MAL
SAL
TH
Caladenia pendens
X
Caladenia pulchra
X
Caladenia reptans
X
Caladenia roei
X
Caladenia vulgata
X
Calandrinia calyptrata
X
Calandrinia eremaea
X
Calandrinia granulifera
X
X
X
Calandrinia polyandra
Calectasia grandiflora
X
X
X
Callistemon phoeniceus
X
X
X
X
Callitris arenaria
X
Callitris canescens
X
Callitris columellaris
Callitris preissii
WL
X
X
X
X
X
X
X
X
Callitris roei
X
X
X
X
Callitris verrucosa
X
X
X
Calothamnus gilesii
X
X
X
Calothamnus huegelii
X
X
Calothamnus lateralis
X
Calothamnus quadrifidus
X
X
X
X
Calotis hispidula
X
Calytrix sapphirina
X
X
X
X
X
X
X
X
X
X
X
Calytrix simplex
X
X
Calytrix strigosa
X
Calytrix violacea
X
X
X
Carpobrotus edulis
X
Carpobrotus modestus
X
Carpobrotus virescens
X
Cassytha aurea
Cassytha glabella
X
X
Calytrix fraseri
Calytrix leschenaultii
X
X
Calytrix breviseta
Calytrix desolata
X
X
X
X
X
X
Cassytha melantha
X
Cassytha nodiflora
X
Cassytha pomiformis
X
Cassytha racemosa
X
X
X
X
Caustis dioica
X
Centrolepis aristata
X
Centrolepis drummondiana
X
Centrolepis pilosa
X
Centrolepis polygyna
X
X
X
X
381
Species
Chamaescilla corymbosa
GR
HEA
MAL
X
SAL
TH
X
Chamaescilla spiralis
X
Chamaexeros fimbriata
X
Chamelaucium ciliatum
X
X
Chamelaucium megalopetalum
X
X
Chamelaucium pauciflorum
X
X
X
X
Chamelaucium sp. Merredin
Cheilanthes austrotenuifolia
X
Cheilanthes distans
X
Cheilanthes sieberi
X
Cheilanthes tenuifolia
X
Cheiranthera filifolia
X
Choretrum glomeratum
X
X
Choretrum pritzelii
X
X
Chorizema aciculare
X
X
X
X
X
X
Clematis delicata
Clematis spp.
X
X
Chloris truncata
Chthonocephalus pseudevax
WL
X
X
Coleanthera myrtoides
X
Comesperma drummondii
X
X
X
Comesperma integerrimum
X
Comesperma scoparium
X
Comesperma volubile
X
X
X
X
X
Conospermum amoenum
X
Conospermum bracteosum
X
Conospermum brownii
X
Conospermum croniniae
X
X
Conospermum filifolium
X
X
X
X
Conospermum stoechadis
X
Conostephium preissii
X
X
X
Conostylis aculeata
X
Conostylis albescens
X
Conostylis argentea
X
Conostylis petrophiloides
X
Coopernookia strophiolata
X
X
X
Corynotheca micrantha
X
Crassula colorata
X
Crassula decumbens
X
Crassula exserta
X
X
X
Cratystylis conocephala
X
X
X
X
X
Cryptandra apetala
X
Cryptandra dielsii
X
382
Species
GR
Cryptandra leucopogon
HEA
MAL
X
X
Cryptandra minutifolia
X
Cryptandra nutans
X
X
X
X
X
Cryptandra pungens
Cryptandra sp.
TH
X
Cryptandra myriantha
Cryptandra polyclada
SAL
WL
X
X
X
X
X
X
X
X
X
Cyanicula amplexans
X
Cyanicula aperta
X
Cyanicula ashbyae
X
Cyanostegia angustifolia
X
Cyanostegia lanceolata
X
Cyanostegia microphylla
X
Cyathostemon heterantherus
X
X
Cyphanthera microphylla
X
Dampiera angulata
X
X
Dampiera glabrescens
Dampiera haematotricha
X
X
X
Dampiera heteroptera
X
Dampiera juncea
X
Dampiera lavandulacea
X
X
X
X
Dampiera luteiflora
X
Dampiera oligophylla
X
Dampiera sacculata
X
X
X
Dampiera stenostachya
X
Dampiera wellsiana
X
Darwinia diosmoides
Darwinia sp. Karonie
X
X
X
X
Dasymalla axillaris
X
X
Dasymalla terminalis
X
X
Daucus glochidiatus
X
Daviesia abnormis
X
X
Daviesia audax
Daviesia benthamii
X
X
X
X
X
X
Daviesia cardiophylla
X
Daviesia daphnoides
X
Daviesia hakeoides
X
X
X
X
X
Daviesia incrassata
X
Daviesia inflata
X
X
X
X
Daviesia intricata
X
Daviesia lancifolia
X
Daviesia nudiflora
X
Daviesia preissii
X
X
Daviesia rhombifolia
X
X
X
X
383
Species
GR
HEA
MAL
Daviesia spiralis
X
X
Daviesia teretifolia
X
Daviesia uncinata
X
Daviesia uniflora
X
Desmocladus asper
Desmocladus lateriflorus
X
TH
WL
X
X
X
X
SAL
X
X
X
X
Desmocladus myriocladus
X
Desmocladus parthenicus
X
Dianella revoluta
X
Dichopogon fimbriatus
X
Dicrastylis parvifolia
X
X
X
X
Didymanthus roei
X
Disphyma crassifolium
X
Diuris corymbosa
X
Diuris picta
X
Diuris setacea
X
X
X
Diuris sp.
X
Dodonaea amblyophylla
X
Dodonaea bursariifolia
Dodonaea caespitosa
X
X
X
X
Dodonaea inaequifolia
X
Dodonaea microzyga
X
Dodonaea ptarmicaefolia
X
X
X
Dodonaea stenozyga
X
X
Dodonaea viscosa
X
Drosera andersoniana
X
Drosera androsacea
X
Drosera bulbosa
X
Drosera glanduligera
X
Drosera macrantha
X
X
X
X
X
X
Drosera macrophylla
X
X
X
X
Drosera sp.
X
Drosera stricticaulis
X
Drosera subhirtella
X
X
Drummondita hassellii
X
X
Ecdeiocolea monostachya
X
X
Einadia nutans
X
X
X
X
X
X
Enekbatus stowardii
X
X
Enneapogon caerulescens
Eragrostis dielsii
X
X
Enchylaena lanata
Enchylaena tomentosa
X
X
Drosera pycnoblasta
Elythranthera brunonis
X
X
X
384
Species
GR
HEA
Eremaea beaufortioides
X
Eremaea pauciflora
X
MAL
SAL
X
TH
WL
X
X
Eremophila brevifolia
X
Eremophila decipiens
Eremophila drummondii
X
X
X
X
X
Eremophila lehmanniana
X
Eremophila scoparia
X
X
X
Erempohila glabra
Eriachne ovata
X
X
X
Erichsenia uncinata
X
Ericksonella saccharata
X
X
Ericomyrtus serpyllifolia
X
X
Eriochilus dilatatus
X
Erodium cygnorum
X
X
X
X
X
Erymophyllum tenellum
X
X
Eucalyptus albida
X
X
Eucalyptus alipes
X
X
X
Eucalyptus annulata
X
Eucalyptus astringens
X
Eucalyptus burracoppinensis
X
X
Eucalyptus calycogona
X
X
Eucalyptus capillosa
X
Eucalyptus celastroides
X
Eucalyptus comitae-vallis
X
Eucalyptus cylindriflora
X
X
X
X
X
X
X
Eucalyptus dissimulata
X
X
X
Eucalyptus falcata
X
X
X
X
X
X
X
X
X
X
X
Eucalyptus grossa
X
Eucalyptus incrassata
X
X
Eucalyptus kondininensis
X
X
Eucalyptus leptopoda
X
X
X
X
X
X
X
Eucalyptus longicornis
X
X
X
X
X
X
Eucalyptus moderata
Eucalyptus occidentalis
X
X
Eucalyptus horistes
Eucalyptus myriadena
X
X
Eucalyptus gracilis
Eucalyptus merrickiae
X
X
Eucalyptus gardneri
Eucalyptus loxophleba
X
X
Eucalyptus flocktoniae
Eucalyptus foecunda
X
X
Eucalyptus densa
Eucalyptus eremophila
X
X
X
X
X
385
Species
GR
HEA
Eucalyptus oleosa
MAL
SAL
X
TH
WL
X
Eucalyptus ornata
X
Eucalyptus ovularis
X
Eucalyptus phaenophylla
X
Eucalyptus phenax
X
Eucalyptus pileata
X
Eucalyptus platycorys
X
Eucalyptus redunca
X
X
X
X
Eucalyptus rigidula
X
Eucalyptus salicola
X
X
X
X
X
X
X
Eucalyptus salmonophloia
X
X
X
X
Eucalyptus salubris
X
X
X
X
Eucalyptus sargentii
X
Eucalyptus sheathiana
X
Eucalyptus spathulata
Eucalyptus sporadica
X
X
Eucalyptus subangusta
X
X
X
X
X
X
Eucalyptus transcontinentalis
X
X
X
X
X
X
X
X
Eucalyptus wandoo
X
Eucalyptus yilgarnensis
X
X
Euryomyrtus leptospermoides
X
Euryomyrtus maidenii
X
Eutaxia neurocalyx
X
Exocarpos aphyllus
X
X
X
Exocarpos sparteus
X
X
X
X
X
Frankenia desertorum
X
Frankenia drummondii
X
Frankenia punctata
X
Frankenia tetrapetala
X
Gahnia ancistrophylla
X
X
X
Gahnia drummondii
X
X
X
Gahnia trifida
X
X
Gastrolobium densifolium
X
Gastrolobium hookeri
X
Gastrolobium parviflorum
X
Gastrolobium punctatum
X
X
X
X
Gastrolobium crassifolium
X
X
Frankenia cinerea
X
X
X
X
X
X
X
Gastrolobium reticulatum
Gastrolobium spinosum
X
X
X
X
Gastrolobium trilobum
X
X
Glischrocaryon aureum
X
Glischrocaryon flavescens
X
Glischrocaryon roei
X
X
X
X
X
X
386
Species
GR
HEA
MAL
Gnephosis brevifolia
SAL
X
Gnephosis tridens
X
Gompholobium hendersonii
X
X
Gonocarpus confertifolius
X
X
X
Goodenia coerulea
X
Goodenia convexa
X
Goodenia pinifolia
X
X
Goodenia salina
X
X
X
X
X
Goodenia scapigera
X
Grevillea acacioides
X
Grevillea acuaria
X
X
Grevillea anethifolia
X
X
X
Grevillea asteriscosa
X
Grevillea cagiana
X
X
Grevillea didymobotrya
X
X
X
Grevillea eriostachya
X
X
X
Grevillea eryngioides
X
X
X
Grevillea excelsior
X
Grevillea hookeriana
X
X
Grevillea huegelii
X
X
X
X
X
X
X
Grevillea integrifolia
X
X
Grevillea involucrata
X
Grevillea oligantha
X
X
Grevillea paradoxa
Grevillea petrophiloides
X
X
X
X
Grevillea pilosa
X
X
X
X
X
Grevillea shuttleworthiana
Grevillea teretifolia
X
Grevillea umbellulata
X
X
X
X
X
Grevillea wittweri
X
Grevillea yorkrakinensis
X
X
X
X
Gunniopsis septifraga
X
Hakea adnata
X
X
Hakea ambigua
X
X
Hakea baxteri
X
Hakea corymbosa
X
X
X
Hakea cygna
X
X
X
Hakea falcata
X
X
X
Grevillea prostrata
Hakea erecta
X
X
X
Grevillea incrassata
Grevillea paniculata
WL
X
Gnephosis tenuissima
Gonocarpus nodulosus
TH
X
X
X
X
X
X
X
387
Species
GR
HEA
MAL
Hakea francisiana
X
X
Hakea gilbertii
X
X
Hakea horrida
X
X
Hakea incrassata
X
X
Hakea kippistiana
SAL
X
X
X
Hakea lissocarpha
X
X
Hakea marginata
X
X
X
Hakea meisneriana
X
X
X
X
Hakea minyma
X
X
X
X
X
Hakea obliqua
X
X
Hakea pandanicarpa
X
X
X
X
Hakea platysperma
X
X
Hakea preissii
X
Hakea prostrata
X
X
X
Hakea scoparia
X
X
X
Hakea subsulcata
X
X
X
X
X
X
X
X
X
X
Hakea trifurcata
X
X
X
Halgania lavandulacea
X
Halgania sp.
Halgania spp.
X
X
Helichrysum leucopsideum
X
Hemigenia dielsii
X
Hemigenia diplanthera
X
Hemiphora uncinata
X
Hibbertia eatoniae
Hibbertia exasperata
X
X
X
X
X
X
Hibbertia glomerosa
X
X
Hibbertia gracilipes
X
Hibbertia hemignosta
X
Hibbertia pungens
X
X
X
X
Hibbertia recurvifolia
X
X
X
Hibbertia stowardii
X
Hibbertia uncinata
X
Hibbertia verrucosa
X
X
X
Homalocalyx thryptomenoides
Hyalochlamys globifera
X
X
Hibbertia lineata
Hibbertia rupicola
X
X
Hakea oldfieldii
Hakea sulcata
X
X
Hakea newbeyana
Hakea petiolaris
WL
X
Hakea lehmanniana
Hakea multilineata
TH
X
X
X
388
Species
GR
Hyalosperma demissum
X
Hyalosperma glutinosum
X
Hybanthus floribundus
X
Hydrocotyle alata
X
Hydrocotyle callicarpa
X
Hydrocotyle diantha
X
HEA
SAL
TH
X
X
X
X
X
X
Hypocalymma puniceum
X
Hypolaena pubescens
X
Isoetes australis
X
Isolepis stellata
X
Isopogon attenuatus
X
Isopogon buxifolius
X
Isopogon divergens
X
X
X
X
X
X
Isopogon polycephalus
X
X
X
Isopogon scabriusculus
X
X
X
Isopogon teretifolius
X
X
X
Isopogon villosus
X
X
Isotoma petraea
X
Jacksonia condensata
X
Jacksonia nematoclada
X
Jacksonia racemosa
X
X
Juncus sp.
X
Kennedia prostrata
X
Kippistia suaedifolia
X
X
Kunzea jucunda
X
X
Kunzea preissiana
Kunzea pulchella
X
X
Kennedia prorepens
Kunzea micromera
WL
X
Hydrocotyle rugulosa
Hypocalymma angustifolium
MAL
X
X
X
X
Labichea stellata
X
Lasiopetalum microcardium
X
Lawrencella davenportii
X
Lawrencella rosea
X
X
Lawrencia squamata
X
X
Laxmannia grandiflora
X
Laxmannia paleacea
X
Laxmannia squarrosa
X
X
Lepidium rotundum
X
Lepidobolus chaetocephalus
X
X
Lepidobolus preissianus
X
X
Lepidosperma angustatum
X
X
X
Lepidosperma drummondii
X
X
X
X
X
X
X
X
X
X
389
Species
GR
Lepidosperma effusum
Lepidosperma gracile
HEA
MAL
X
X
X
X
X
X
X
X
X
X
X
X
Lepidosperma sanguinolentum
X
X
X
Lepidosperma scabrum
X
Lepidosperma tenue
X
X
X
Lepidosperma tuberculatum
X
X
X
Leptomeria pauciflora
X
Leptomeria preissiana
X
Leptosema daviesioides
X
Leptospermum erubescens
X
X
Leptospermum incanum
X
X
Leptospermum inelegans
X
Leptospermum nitens
X
X
X
X
X
X
X
X
X
X
X
Leptospermum oligandrum
X
Leptospermum spinescens
X
X
Leucopogon blepharolepis
X
X
Leucopogon conostephioides
X
X
Leucopogon crassifolius
X
X
X
X
Leucopogon cuneifolius
X
X
X
X
Leucopogon dielsianus
X
X
X
X
X
Leucopogon fimbriatus
X
Leucopogon hamulosus
X
X
Leucopogon obtusatus
X
X
X
Leucopogon sp. Forrestania
X
Leucopogon tamminensis
X
X
X
X
Lobelia gibbosa
X
Lobelia heterophylla
X
Lobelia tenuior
X
Lomandra collina
X
Lomandra effusa
X
X
Lomandra mucronata
X
Lomandra sp.
X
Loxocarya cinerea
X
Loxocarya sp.
Lycium australe
Lyginia barbata
X
X
Leucopogon sp.
Levenhookia dubia
X
X
Lepidosperma rigidulum
Leucopogon oxycedrus
WL
X
Lepidosperma pubisquameum
Lepidosperma resinosum
TH
X
Lepidosperma leptostachyum
Lepidosperma longitudinale
SAL
X
X
X
X
X
X
X
X
390
Species
GR
Lysinema ciliatum
HEA
MAL
X
X
SAL
X
Maireana brevifolia
X
Maireana carnosa
X
Maireana erioclada
X
Maireana oppositifolia
X
X
Malleostemon tuberculatus
X
Melaleuca acuminata
X
X
X
Melaleuca adnata
X
X
X
Melaleuca atroviridis
X
Melaleuca brevifolia
X
Melaleuca calyptroides
X
Melaleuca cardiophylla
X
X
X
X
X
Melaleuca carrii
X
Melaleuca condylosa
Melaleuca cordata
X
X
X
X
X
X
X
X
Melaleuca cuticularis
X
Melaleuca densa
X
Melaleuca depauperata
X
Melaleuca eleuterostachya
X
X
X
Melaleuca elliptica
X
X
Melaleuca fulgens
X
X
X
X
X
X
Melaleuca grieveana
X
Melaleuca halmaturorum
X
Melaleuca hamata
X
X
X
Melaleuca hamulosa
X
X
Melaleuca haplantha
X
X
Melaleuca lanceolata
X
X
X
X
X
Melaleuca lateriflora
X
X
Melaleuca lateritia
X
X
Melaleuca laxiflora
X
X
Melaleuca lecanantha
X
Melaleuca leptospermoides
X
X
X
X
Melaleuca macronychia
WL
X
Maireana amoena
Maireana triptera
TH
X
X
X
X
X
X
X
X
X
Melaleuca marginata
X
Melaleuca nesophila
X
X
Melaleuca oldfieldii
X
Melaleuca pauperiflora
X
X
X
Melaleuca pentagona
X
X
X
Melaleuca platycalyx
X
X
X
X
X
X
X
Melaleuca pungens
Melaleuca radula
X
Melaleuca scabra
X
X
X
X
X
X
X
391
Species
GR
HEA
MAL
Melaleuca scalena
SAL
TH
X
Melaleuca seriata
X
Melaleuca spathulata
X
X
X
X
Melaleuca spicigera
X
X
X
X
Melaleuca subtrigona
X
X
X
X
Melaleuca thyoides
Melaleuca uncinata
X
X
X
X
Melaleuca urceolaris
X
Melaleuca villosisepala
X
Melaleuca viminea
X
X
X
X
X
X
Microcorys ericifolia
X
Microcorys exserta
X
Microcorys sp. stellate
X
Microcorys subcanescens
X
X
Microcybe ambigua
X
Microcybe multiflora
X
X
X
Micromyrtus obovata
X
X
X
X
X
Microtis eremaea
X
Microtis media
X
Millotia myosotidifolia
X
Millotia tenuifolia
X
X
X
X
X
Micromyrtus racemosa
X
X
X
X
Mirbelia dilatata
Mirbelia floribunda
X
X
Mirbelia ramulosa
X
X
Mirbelia spinosa
X
X
Mirbelia trichocalyx
X
X
X
X
X
X
Monotaxis grandiflora
X
X
Muehlenbeckia adpressa
X
Neurachne alopecuroidea
X
Nicotiana cavicola
X
Olax benthamiana
X
X
X
X
X
Olearia axillaris
X
Olearia homolepis
X
Olearia incondita
X
Olearia muelleri
X
Olearia muricata
Olearia revoluta
X
X
Mesomelaena stygia
Mirbelia spp.
X
X
X
Mesomelaena preissii
Micromyrtus imbricata
X
X
Melaleuca undulata
Menkea spp.
WL
X
X
X
X
X
X
X
392
Species
GR
Olearia sp. Eremicola
HEA
MAL
X
Opercularia vaginata
X
Ophioglossum lusitanicum
X
X
Orianthera flaviflora
X
Orianthera nuda
X
X
Orianthera tortuosa
X
X
Oxalis perennans
X
X
Patersonia juncea
X
X
X
Pauridia occidentalis
Persoonia coriacea
X
X
X
Persoonia hakeiformis
X
X
X
X
X
Persoonia inconspicua
Persoonia quinquenervis
WL
X
Opercularia hispidula
Pauridia glabella
TH
X
Olearia sp. Kennedy Range
Parietaria debilis
SAL
X
X
X
X
X
Persoonia saundersiana
X
X
Persoonia striata
X
Persoonia teretifolia
X
Persoonia trinervis
X
X
Petrophile circinata
X
X
Petrophile conifera
X
Petrophile divaricata
X
X
X
X
X
X
X
X
X
X
X
X
X
Petrophile squamata
X
X
X
X
Petrophile striata
X
Phebalium filifolium
X
X
X
X
Phebalium lepidotum
X
X
X
Phebalium megaphyllum
X
X
X
X
X
X
Petrophile ericifolia
X
Petrophile scabriuscula
X
Petrophile seminuda
X
Phebalium tuberculosum
X
Pheladenia deformis
X
Philotheca gardneri
X
Philotheca tomentella
Phyllangium paradoxum
X
Phyllanthus calycinus
X
Phylloglossum drummondii
X
Physopsis lachnostachya
X
X
X
X
X
X
X
Pimelea angustifolia
X
Pimelea argentea
X
Pimelea graniticola
X
X
X
X
393
Species
GR
Pimelea imbricata
HEA
MAL
X
X
Pimelea suaveolens
X
Pimelea sulphurea
X
Pimelea sylvestris
X
TH
X
X
X
X
X
Plantago debilis
X
Platysace commutata
X
X
Platysace deflexa
X
Platysace effusa
X
X
X
X
X
X
X
X
Platysace juncea
X
Platysace maxwellii
X
Platysace trachymenioides
X
Podolepis capillaris
X
X
X
X
Podolepis lessonii
X
X
X
X
Podolepis tepperi
X
X
X
Podotheca angustifolia
X
Podotheca gnaphalioides
X
Poranthera microphylla
X
Prasophyllum gracile
X
Prasophyllum triangulare
X
Psammomoya choretroides
X
X
X
X
Pterostylis barbata
X
Pterostylis pyramidalis
X
Pterostylis recurva
X
X
Pterostylis sanguinea
X
X
X
X
X
X
Pterostylis scabra
X
Pterostylis sp.
X
Ptilotus fasciculatus
X
X
Ptilotus gaudichaudii
X
X
Ptilotus holosericeus
X
Ptilotus manglesii
X
Ptilotus spathulatus
X
X
X
X
X
X
X
X
X
Puccinellia stricta
X
Pultenaea sp.
X
Pultenaea verruculosa
Regelia inops
Rhagodia drummondii
X
X
Pterochaeta paniculata
Quinetia urvillei
X
X
Pleurosorus rutifolius
Ptilotus polystachyus
X
X
Pityrodia lepidota
Platysace compressa
WL
X
X
Pittosporum angustifolium
Pittosporum phillyreoides
SAL
X
X
X
X
X
X
X
394
Species
GR
HEA
Rhagodia preissii
Rhagodia spp.
X
Rhodanthe citrina
X
Rhodanthe laevis
Rhodanthe manglesii
MAL
SAL
X
X
X
X
X
X
X
X
X
X
X
Rhodanthe pygmaea
Ricinocarpos glaucus
TH
X
X
Rinzia carnosa
X
X
Rinzia crassifolia
X
X
Rinzia polystemonea
X
Roycea pycnophylloides
X
Roycea spinescens
X
Ruppia maritima
X
Rytidosperma caespitosum
X
Rytidosperma setaceum
X
Rytidosperma sp. Goomalling
X
Santalum acuminatum
X
Santalum murrayanum
Santalum spicatum
X
X
X
X
X
X
X
Sarcocornia globosa
X
Scaevola restiacea
X
X
X
X
Schoenus calcatus
X
Schoenus globifer
X
Schoenus hexandrus
X
X
X
Schoenus sp.
X
X
X
X
X
Scholtzia sp.
X
Sclerolaena costata
X
Sclerolaena diacantha
X
Sebaea ovata
X
Senecio glossanthus
X
Senecio sp.
X
Senna artemisioides
X
Senna cardiosperma
X
Senna charlesiana
X
Siloxerus multiflorus
X
Spartochloa scirpoidea
X
Spergularia marina
Spiculaea ciliata
X
X
Schoenus armeria
Scholtzia parviflora
X
X
X
Schoenus nanus
X
X
Sarcocornia blackiana
Scaevola spinescens
WL
X
X
X
X
X
X
X
X
X
395
Species
GR
HEA
MAL
Spyridium subochreatum
X
Stachystemon polyandrus
X
Stackhousia huegelii
X
X
Stackhousia monogyna
X
X
Stackhousia pubescens
X
Stirlingia simplex
X
Stylidium breviscapum
X
Stylidium bulbiferum
X
Stylidium calcaratum
X
X
Stylidium dichotomum
X
Stylidium dielsianum
X
X
X
X
X
X
X
Stylidium involucratum
X
Stylidium leptophyllum
X
Stylidium neglectum
X
X
Stylidium nungarinense
X
Stylidium petiolare
X
Stylidium piliferum
X
X
X
Stylidium repens
X
X
X
Stylidium sacculatum
X
X
X
X
X
Stypandra glauca
X
Stypandra jamesii
X
X
Styphelia tenuiflora
X
Synaphea interioris
X
Synaphea petiolaris
X
X
X
X
X
Tecticornia doliiformis
X
Tecticornia halocnemoides
X
Tecticornia indica
X
Tecticornia lepidosperma
X
Tecticornia leptoclada
X
Tecticornia loriae
X
Tecticornia lylei
X
Tecticornia moniliformis
X
Tecticornia pergranulata
X
Tecticornia pterygosperma
X
Tecticornia undulata
X
Templetonia sulcata
X
X
Stylobasium australe
Templetonia aculeata
X
X
Stylidium yilgarnense
Synaphea spinulosa
WL
X
Stenanthemum stipulosum
Stylidium squamellosum
TH
X
Stackhousia scoparia
Stylidium ecorne
SAL
X
X
X
X
396
Species
Tetrapora preissiana
GR
HEA
MAL
X
X
X
Tetrapora tenuiramea
SAL
X
Thelymitra antennifera
X
Thelymitra flexuosa
X
TH
X
Thelymitra latiloba
Thelymitra macrophylla
X
X
Thelymitra sargentii
X
X
Thelymitra spiralis
X
Thomasia spp.
X
Threlkeldia diffusa
Thryptomene australis
X
X
X
Thryptomene cuspidata
X
Thryptomene kochii
X
Thryptomene racemulosa
X
X
X
X
X
Thysanotus manglesianus
X
X
X
X
Thysanotus sparteus
X
X
X
Thysanotus thyrsoideus
X
Trachymene cyanopetala
X
Trachymene ornata
X
X
X
Trachymene spp.
X
X
Tricoryne tenella
X
X
Tricostularia compressa
X
X
Triglochin calcitrapa
X
Triglochin centrocarpa
X
Triglochin sp.
X
X
Triodia scariosa
X
Trioidia rigidissima
X
Tripterococcus brunonis
X
Trymalium daphnifolium
X
Urodon dasyphyllus
X
X
X
X
Verticordia auriculata
X
Verticordia brownii
X
Verticordia chrysantha
X
X
Verticordia chrysanthella
X
X
Verticordia densiflora
X
Verticordia drummondii
X
Verticordia endlicheriana
X
Verticordia eriocephala
X
X
Velleia cycnopotamica
Verticordia acerosa
X
X
Trachymene sp.
Utricularia tenella
X
X
Thryptomene urceolaris
Thysanotus patersonii
WL
X
X
X
X
X
X
X
X
X
X
X
X
X
X
397
Species
GR
Verticordia gracilis
HEA
MAL
X
X
Verticordia habrantha
TH
WL
X
X
Verticordia inclusa
X
Verticordia insignis
X
Verticordia integra
X
Verticordia mitodes
X
Verticordia multiflora
SAL
X
X
X
X
Verticordia pennigera
X
Verticordia picta
X
X
Verticordia plumosa
X
X
Verticordia roei
X
X
X
X
Verticordia serrata
X
X
X
X
Verticordia tumida
X
X
X
Vittadinia australasica
X
Vittadinia cuneata
X
X
X
X
X
X
Wahlenbergia preissii
X
X
X
X
Waitzia acuminata
X
X
X
X
X
X
X
X
X
X
X
Westringia cephalantha
Westringia rigida
Westringia sp.
X
X
Wilsonia humilis
X
Wurmbea graniticola
X
Wurmbea sinora
X
Xanthorrhoea nana
X
X
X
X
398
APPENDIX B
DISTRIBUTION OF PLANT FOODS IN THE STUDY AREA
The identification of food plants and the seasonal availability of their exploited parts
have been based largely on databases and publications, including Atkins and Ruan
(2010-2016), Bell and Williams (1997), Bindon (1996), Bird (1985), Bird and Beeck
(1988), Brand (2005), Brown et al. (2008), Cribb and Cribb (1975), Daw et al. (1997),
Dortch (2004), Florabank (n.d.), Hopper and Lambers (2014), Low (1991), Meagher
(1974), Native Grass Resources Group (n.d.), Nussinovitch (2010), Paczkowska and
Chapman (2000), Pate and Dixon (1982), Sweedman and Merritt (2006), Webb (2000),
Western Australian Herbarium (1998-) and Wildlife Society of Western Australia
(2013); only those species present in the study area (Appendix A) are included below.
Where species- or genera-specific information was unavailable, periods of availability
were determined using the assumptions and generalisations outlined in Chapter 5.3.3.
Species are listed alphabetically (within food categories for Appendix B.2–B.7), with
taxonomy and naming conventions as per Florabase (Western Australian Herbarium
(1998-). Any non-current names encountered in the references above have been
updated; taxonomic changes made after 1 April 2017 will not be reflected. Where an
individual plant species has been identified as a food source, it has a separate entry.
Where a genus was identified, all members have been incorporated within a single
entry; the figure in parentheses indicates the number of species present in the study
area, excluding those that have a separate entry. Genus-level identifications were
treated as per Appendix A and contributed to the totals for the study area (Appendix
B.1) only where they were the sole representation of a particular genus, or the named
species did not reflect the extent of diversity. Again, spp. entries were assumed to
reflect two distinct species, and totals were calculated as such for each Landscape
Division (Appendix B.2–B.7).
A cross within a month column denotes that the particular food item was available
during that month; blank cells indicate that the item was unavailable at that time. For
genus entries, the period of availability has been defined as that which applies to the
majority of species – in most cases there was minimal variation between members of
the same genus. The only exception is Banksia spp. (n=9), as each species flowered
399
at many different times throughout the year, and no dominant period could be identified
either within or between species; these flowers have been marked as available yearround. Blue filled cells represent periods of high water-content in storage organs.
Plant food categories have been abbreviated as below:
o
FL = Flowers
o
FR = Fruit
o
GM = Gum
o
L/M = Lerp/manna
o
L/R/S = Leaves/roots/stems
o
SE = Seeds
o
SO = Storage organs
400
B.1 MONTHLY AVAILABILITY OF ALL PLANT FOODS
Oct
Nov
X
SE
X
X
X
X
X
Acacia beauverdiana
SE
X
X
X
X
X
Acacia microbotrya
GM
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
SE
Acacia saligna
Acacia spp. (66)
X
GM
X
X
X
SE
X
X
X
GM
X
X
X
SE
X
X
Acrotriche sp.
FR
Allocasuarina spp. (9)
SE
X
X
Amyema spp. (2)
FR
X
X
Aristida sp.
SE
Astroloma serratifolium
FR
X
X
Astroloma spp. (2)
FR
X
X
Atriplex spp. (5)
L/R/S
X
Atriplex vesicaria
L/R/S
X
Aug
X
Jul
X
Jun
GM
May
Acacia acuminata
Apr
Part eaten
(category)
Mar
Species
Sep
Spring
Feb
Winter
Jan
Autumn
Dec
Summer
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
SE
Baeckea spp. (3)
FL
Banksia sphaerocarpa
FL
Banksia spp. (9)
FL
X
Beaufortia spp. (5)
FL
X
Billardiera lehmanniana
FR
X
Billardiera variifolia
FR
Brachyloma spp. (2)
FR
X
Caesia micrantha
SO
X
X
X
X
Caladenia spp. (15)
SO
X
X
X
X
Calandrinia polyandra
L/R/S
X
X
X
X
X
X
X
X
X
X
X
X
SO
X
X
X
X
X
X
X
X
X
X
X
X
FL
X
X
X
X
X
L/R/S
X
X
X
X
X
X
X
X
X
X
X
X
SE
X
X
X
X
X
SO
X
X
X
X
X
Callistemon sp.
FL
X
X
X
X
Calothamnus spp. (4)
FL
X
X
X
X
X
Calytrix spp. (8)
FL
X
X
X
X
Carpobrotus spp. (2)
FR
X
X
X
L/R/S
X
X
X
X
X
X
X
Calandrinia spp. (3)
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
401
Species
Part eaten
(category)
Mar
Apr
May
Jun
Jul
Aug
Sep
Oct
Nov
Spring
Feb
Winter
Jan
Autumn
Dec
Summer
Carpobrotus virescens
FR
X
X
X
L/R/S
X
X
X
X
X
X
X
X
X
X
X
X
Cassytha glabella
FR
X
X
X
X
X
X
X
X
X
X
X
X
Cassytha racemosa
FR
X
X
X
X
X
X
X
X
X
X
X
X
Cassytha spp. (4)
FR
X
X
X
X
X
X
Chamelaucium spp. (4)
FL
X
X
X
X
X
X
X
X
X
X
X
X
Cyanicula spp. (3)
SO
X
X
X
X
Cyathostemon sp.
FL
Darwinia spp. (2)
FL
X
Dianella revoluta
FR
X
X
X
X
L/R/S
X
X
X
X
X
X
X
X
X
X
X
X
Dichopogon fimbriatus
SO
X
X
X
X
X
X
X
X
X
X
X
X
Disphyma crassifolium
L/R/S
X
X
X
X
X
X
X
X
X
X
X
X
Diuris spp. (3)
SO
X
X
X
X
Dodonaea viscosa
FR
Drosera spp. (9)
SO
X
X
X
X
Einadia nutans
L/R/S
X
X
X
X
Elythranthera sp.
SO
X
X
X
X
Enchylaena tomentosa
FR
X
L/R/S
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
Enekbatus sp.
FL
Enneapogon sp.
SE
Eragrostis dielsii
SE
Eremaea spp. (2)
FL
X
X
X
X
Eriachne sp.
SE
X
X
X
X
Ericomyrtus sp.
FL
X
X
X
Erodium cygnorum
L/R/S
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
SE
Eucalyptus loxophleba
FL
X
X
X
L/R/S
X
X
X
Eucalyptus occidentalis
GM
X
X
X
Eucalyptus oleosa
L/M
X
X
X
Eucalyptus spp. (45)
FL
X
X
X
L/M
X
X
X
X
X
X
Eucalyptus wandoo
FL
Exocarpos aphyllus
FR
X
X
Exocarpos sparteus
FR
X
X
Frankenia spp. (5)
FL
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
L/R/S
X
X
X
X
X
X
X
X
X
X
X
X
Gahnia spp. (3)
L/R/S
X
X
X
X
X
X
X
X
X
X
X
X
Grevillea eriostachya
FL
X
X
X
X
X
X
X
X
X
X
X
X
402
GM
X
X
X
SE
X
X
X
X
X
X
X
X
X
Hypocalymma spp. (2)
FL
Kennedia prostrata
FL
Kunzea spp. (4)
FL
Lepidium sp.
L/R/S
X
X
SE
X
Leptospermum spp. (6)
FL
X
Leucopogon spp. (11)
FR
X
Lomandra spp. (3)
FL
Lycium australe
X
X
X
X
X
X
X
X
L/R/S
X
X
FR
X
X
Jul
Aug
Sep
Oct
Nov
X
X
X
X
X
X
X
L/R/S
X
X
X
Melaleuca spp. (47)
FL
X
X
X
Micromyrtus spp. (3)
FL
Microtis spp. (2)
SO
X
X
X
X
Oxalis sp.
L/R/S
X
X
X
X
Persoonia spp. (8)
FR
X
X
X
Pittosporum phillyreoides
GM
X
X
X
Platysace deflexa
SO
X
X
X
Platysace effusa
SO
X
X
Platysace maxwellii
SO
X
Platysace spp. (4)
SO
Prasophyllum spp. (2)
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
Spring
Jun
FL
Winter
May
X
Apr
FL
Hakea spp. (30)
Autumn
Mar
Grevillea spp. (24)
Feb
Part eaten
(category)
Jan
Species
Dec
Summer
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
SO
X
X
X
X
X
Pterostylis recurva
SO
X
X
X
X
X
Pterostylis spp. (4)
SO
X
X
X
X
X
Regelia sp.
FL
X
Rinzia spp. (3)
FL
Santalum acuminatum
FR
Santalum murrayanum
FR
L/R/S
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
Santalum spicatum
FR
Scaevola spinescens
FR
Scholtzia sp.
FL
Spiculaea sp.
SO
Styphelia tenuiflora
FR
Tecticornia indica
L/R/S
X
X
X
X
Thelymitra spp. (6)
SO
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
403
Aug
Sep
Oct
Nov
X
Jul
X
Jun
X
X
Spring
May
L/R/S
X
Winter
Apr
FR
Autumn
Mar
Threlkeldia diffusa
Feb
Part eaten
(category)
Jan
Species
Dec
Summer
X
X
X
X
X
X
X
X
X
Thryptomene spp. (5)
FL
Thysanotus manglesianus
L/R/S
X
X
X
X
X
X
X
X
X
X
X
X
SO
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
Thysanotus patersonii
FL
L/R/S
X
X
X
X
X
X
X
X
X
X
X
X
X
SO
X
X
X
X
Thysanotus sp.
SO
X
X
X
X
X
X
X
X
X
X
X
X
Thysanotus thyrsoideus
SO
X
X
X
X
X
X
X
X
X
X
X
X
Triglochin spp. (2)
SO
X
X
X
X
X
X
X
X
X
X
X
X
Verticordia spp. (22)
FL
X
X
X
X
Wurmbea spp. (2)
SO
X
Xanthorrhoea sp.
FL
X
X
X
X
X
X
X
X
404
B.2 MONTHLY AVAILABILITY OF PLANT FOODS IN GRANITE
Banksia sphaerocarpa
FL
Banksia sp.
FL
X
Calandrinia spp. (3)
FL
X
Calothamnus sp.
FL
X
Calytrix spp. (2)
FL
Cyathostemon sp.
X
X
X
X
X
X
X
X
X
X
X
X
X
X
FL
X
X
X
Ericomyrtus sp.
FL
X
X
X
Eucalyptus loxophleba
FL
X
X
X
X
X
X
Eucalyptus spp. (3)
FL
X
X
X
Grevillea spp. (5)
FL
X
Hakea spp. (9)
FL
Hypocalymma sp.
FL
Kennedia prostrata
FL
Kunzea spp. (2)
FL
Leptospermum spp. (2)
FL
Lomandra spp. (2)
FL
Melaleuca spp. (10)
FL
Micromyrtus sp.
FL
Scholtzia sp.
FL
Thryptomene sp.
FL
Thysanotus patersonii
FL
Verticordia spp. (9)
FL
Xanthorrhoea sp.
FL
Astroloma serratifolium
FR
X
X
X
X
X
X
X
Astroloma sp.
FR
X
X
X
X
X
X
Billardiera variifolia
FR
X
X
X
Carpobrotus virescens
FR
X
X
X
Cassytha glabella
FR
X
X
X
X
X
X
Dianella revoluta
FR
X
X
X
X
Dodonaea viscosa
FR
Enchylaena tomentosa
FR
X
Exocarpos aphyllus
FR
X
X
Exocarpos sparteus
FR
X
X
Leucopogon spp. (3)
FR
X
X
Lycium australe
FR
X
X
Persoonia spp. (2)
FR
Santalum acuminatum
FR
Santalum spicatum
FR
Scaevola spinescens
FR
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
Jul
X
Jun
X
Apr
X
Mar
X
Feb
X
Jan
Nov
X
Oct
FL
Spring
Sep
Baeckea spp. (2)
Winter
X
X
Aug
Part eaten
(category)
Autumn
May
Species
Dec
Summer
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
405
Species
Part eaten
(category)
Mar
Apr
May
Jun
Jul
Aug
Sep
Oct
Nov
Spring
Feb
Winter
Jan
Autumn
Dec
Summer
Acacia acuminata
GM
X
X
X
Acacia microbotrya
GM
X
X
X
Acacia saligna
GM
X
X
X
Acacia spp. (14)
GM
X
X
X
Hakea spp. (9)
GM
X
X
X
Pittosporum phillyreoides
GM
X
X
X
Atriplex sp.
L/R/S
X
X
X
X
X
X
X
X
X
X
X
X
Calandrinia spp. (3)
L/R/S
X
X
X
X
X
X
X
X
X
X
X
X
Carpobrotus virescens
L/R/S
X
X
X
X
X
X
X
X
X
X
X
X
Dianella revoluta
L/R/S
X
X
X
X
X
X
X
X
X
X
X
X
Enchylaena tomentosa
L/R/S
X
X
X
X
X
X
X
X
X
X
X
X
Erodium cygnorum
L/R/S
X
X
X
X
X
X
X
X
X
X
X
X
Eucalyptus loxophleba
L/R/S
X
X
X
X
X
X
X
X
X
X
X
X
Gahnia spp. (2)
L/R/S
X
X
X
X
X
X
X
X
X
X
X
X
Lomandra spp. (2)
L/R/S
X
X
X
X
X
X
X
X
X
X
X
X
Lycium australe
L/R/S
X
X
X
X
X
X
X
X
X
X
X
X
Thysanotus patersonii
L/R/S
X
X
X
X
X
X
X
X
X
Eucalyptus spp. (3)
L/M
X
X
Acacia acuminata
SE
X
X
X
X
X
Acacia microbotrya
SE
X
X
X
Acacia saligna
SE
X
X
Acacia spp. (14)
SE
X
X
Allocasuarina spp. (4)
SE
X
X
Calandrinia spp. (3)
SE
X
X
Eriachne sp.
SE
X
Erodium cygnorum
SE
Hakea spp. (9)
SE
X
X
X
X
X
X
X
Caesia micrantha
SO
X
X
X
X
X
X
Caladenia spp. (11)
SO
X
X
X
X
Calandrinia spp. (3)
SO
X
X
X
X
X
X
Cyanicula sp.
SO
X
X
X
X
Dichopogon fimbriatus
SO
X
X
X
X
Diuris spp. (3)
SO
X
X
X
X
X
Drosera spp. (7)
SO
X
X
X
X
X
Elythranthera sp.
SO
X
X
X
X
X
Microtis spp. (2)
SO
X
X
X
X
X
Platysace maxwellii
SO
X
X
X
X
X
X
X
X
X
X
X
X
Platysace spp. (2)
SO
X
X
X
X
X
X
X
X
X
X
X
X
Prasophyllum spp. (2)
SO
X
X
X
X
X
Pterostylis recurva
SO
X
X
X
X
X
Pterostylis spp. (3)
SO
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
406
X
Thelymitra spp. (4)
SO
X
X
X
X
X
Thysanotus patersonii
SO
X
X
X
X
X
Triglochin spp. (2)
SO
X
X
X
X
Wurmbea spp. (2)
SO
X
X
X
X
X
X
X
X
X
X
X
Nov
X
Oct
X
Sep
X
Aug
X
Jul
SO
Jun
Spiculaea sp.
May
Part eaten
(category)
Apr
Species
Mar
Spring
Feb
Winter
Jan
Autumn
Dec
Summer
X
X
407
B.3 MONTHLY AVAILABILITY OF PLANT FOODS IN HEATH
X
Callistemon sp.
FL
X
Calothamnus spp. (4)
FL
X
Calytrix spp. (6)
FL
Chamelaucium spp. (4)
FL
X
Eremaea spp. (2)
FL
X
Ericomyrtus sp.
FL
Eucalyptus loxophleba
FL
X
X
X
Eucalyptus spp. (15)
FL
X
X
X
Grevillea eriostachya
FL
X
X
X
Grevillea spp. (14)
FL
X
Hakea spp. (23)
Nov
FL
Oct
Beaufortia spp. (5)
Sep
X
Spring
Aug
FL
Jul
Banksia spp. (9)
Winter
Jun
FL
May
Banksia sphaerocarpa
Apr
Baeckea spp. (3)
Part eaten
(category)
FL
Species
Autumn
Mar
Feb
Jan
Dec
Summer
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
FL
X
X
X
Kunzea spp. (3)
FL
X
X
X
X
Leptospermum spp. (5)
FL
X
X
X
X
Lomandra spp. (2)
FL
X
X
X
Melaleuca spp. (26)
FL
X
X
X
Micromyrtus spp. (3)
FL
X
X
X
Regelia sp.
FL
X
X
Rinzia sp.
FL
X
X
Scholtzia sp.
FL
X
X
X
Thryptomene spp. (4)
FL
X
X
X
X
X
Thysanotus patersonii
FL
X
X
X
X
X
Verticordia spp. (20)
FL
X
X
X
Xanthorrhoea sp.
FL
X
X
X
Acrotriche sp.
FR
X
X
X
X
Astroloma serratifolium
FR
X
X
X
X
X
X
X
X
X
X
X
X
Astroloma sp.
FR
X
X
X
X
X
X
X
X
X
X
X
X
Brachyloma sp.
FR
X
X
X
X
X
X
X
Cassytha glabella
FR
X
X
X
X
X
X
X
X
X
X
X
X
Cassytha spp. (4)
FR
X
X
X
X
X
X
Dianella revoluta
FR
X
X
X
X
Exocarpos aphyllus
FR
X
X
X
Exocarpos sparteus
FR
X
X
Leucopogon spp. (10)
FR
X
X
Persoonia spp. (7)
FR
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
408
GM
X
X
X
Acacia microbotrya
GM
X
X
X
Acacia saligna
GM
X
X
X
Acacia spp. (26)
GM
X
X
X
Hakea spp. (23)
GM
X
X
X
Dianella revoluta
L/R/S
X
X
X
X
X
Eucalyptus loxophleba
L/R/S
X
X
X
X
Gahnia spp. (2)
L/R/S
X
X
X
Lomandra spp. (2)
L/R/S
X
X
Santalum murrayanum
L/R/S
X
X
Thysanotus patersonii
L/R/S
Eucalyptus spp. (15)
L/M
X
X
Acacia acuminata
SE
X
Acacia beauverdiana
SE
X
Acacia microbotrya
SE
Acacia saligna
SE
X
X
Acacia spp. (26)
SE
X
X
Allocasuarina spp. (8)
SE
X
X
Aristida sp.
SE
Hakea spp. (23)
SE
X
X
Caladenia spp. (2)
SO
X
Drosera spp. (2)
SO
Platysace deflexa
X
Nov
Acacia acuminata
X
Oct
FR
X
Sep
Styphelia tenuiflora
Spring
Aug
FR
Jul
Santalum murrayanum
Winter
Jun
Santalum acuminatum
Part eaten
(category)
FR
Species
May
Apr
Autumn
Mar
Feb
Jan
Dec
Summer
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
SO
X
X
X
X
X
X
X
X
X
X
X
X
Platysace effusa
SO
X
X
X
X
X
X
X
X
X
X
X
X
Platysace maxwellii
SO
X
X
X
X
X
X
X
X
X
X
X
X
Platysace sp.
SO
X
X
X
X
X
X
X
X
X
X
X
X
Pterostylis recurva
SO
X
X
X
X
X
Pterostylis sp.
SO
X
X
X
X
X
Thelymitra sp.
SO
X
X
X
X
X
Thysanotus patersonii
SO
X
X
X
X
X
Thysanotus sp.
SO
X
X
X
X
X
X
X
X
X
X
X
X
Thysanotus thyrsoideus
SO
X
X
X
X
X
X
X
X
X
X
X
X
409
B.4 MONTHLY AVAILABILITY OF PLANT FOODS IN MALLEE
Sep
Oct
Nov
Spring
Aug
Jun
Winter
May
Apr
Autumn
Mar
Feb
Jan
Dec
Summer
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
Part eaten
(category)
FL
Banksia sphaerocarpa
FL
Banksia spp. (6)
FL
X
Beaufortia spp. (5)
FL
X
Calandrinia sp.
FL
X
Callistemon sp.
FL
X
Calothamnus spp. (3)
FL
X
Calytrix spp. (4)
FL
Chamelaucium spp. (2)
FL
Cyathostemon sp.
FL
Eremaea sp.
FL
Ericomyrtus sp.
FL
Eucalyptus loxophleba
FL
X
X
X
Eucalyptus spp. (31)
FL
X
X
X
Grevillea eriostachya
FL
X
X
X
Grevillea spp. (13)
FL
X
Hakea spp. (20)
FL
Hypocalymma sp.
FL
Leptospermum spp. (4)
FL
Lomandra spp. (2)
FL
Melaleuca spp. (33)
FL
Micromyrtus spp. (2)
FL
Regelia sp.
FL
Rinzia spp. (2)
FL
Thryptomene spp. (2)
FL
Thysanotus patersonii
FL
Verticordia spp. (12)
FL
Xanthorrhoea sp.
FL
X
Acrotriche sp.
FR
Astroloma serratifolium
FR
X
X
X
X
X
X
X
Astroloma spp. (2)
FR
X
X
X
X
X
X
Billardiera variifolia
FR
X
X
Cassytha racemosa
FR
X
X
X
X
Cassytha sp.
FR
X
X
X
X
Dianella revoluta
FR
X
X
X
X
Dodonaea viscosa
FR
Exocarpos aphyllus
FR
X
X
Exocarpos sparteus
FR
X
X
Leucopogon spp. (8)
FR
X
X
Persoonia spp. (4)
FR
Jul
Species
Baeckea spp. (2)
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
410
Sep
Oct
Nov
Spring
Aug
Jun
Winter
May
Apr
Autumn
Mar
Feb
Jan
Dec
Summer
X
X
X
Part eaten
(category)
FR
Santalum murrayanum
FR
Styphelia tenuiflora
FR
Acacia acuminata
GM
X
X
X
Acacia microbotrya
GM
X
X
X
Acacia spp. (38)
GM
X
X
X
Hakea spp. (20)
GM
X
X
X
Atriplex sp.
L/R/S
X
X
X
X
X
Calandrinia sp.
L/R/S
X
X
X
X
Dianella revoluta
L/R/S
X
X
X
Eucalyptus loxophleba
L/R/S
X
X
Gahnia spp. (2)
L/R/S
X
X
Lomandra spp. (2)
L/R/S
X
Santalum murrayanum
L/R/S
X
Thysanotus patersonii
L/R/S
Eucalyptus oleosa
L/M
X
X
X
Eucalyptus spp. (31)
L/M
X
X
X
Acacia acuminata
SE
X
X
Acacia beauverdiana
SE
X
X
Acacia microbotrya
SE
Acacia spp. (38)
SE
X
X
Allocasuarina spp. (6)
SE
X
X
Calandrinia sp.
SE
X
X
Hakea spp. (20)
SE
X
X
X
X
X
X
X
X
Calandrinia sp.
SO
X
X
X
X
X
X
X
X
Drosera spp. (2)
SO
X
X
X
X
Platysace effusa
SO
X
X
X
X
X
X
X
X
X
X
X
X
Platysace maxwellii
SO
X
X
X
X
X
X
X
X
X
X
X
X
Platysace sp.
SO
X
X
X
X
X
X
X
X
X
X
X
X
Pterostylis sp.
SO
X
X
X
X
X
Thysanotus patersonii
SO
X
X
X
X
X
Thysanotus sp.
SO
X
X
X
X
Jul
Species
Santalum acuminatum
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
411
B.5 MONTHLY AVAILABILITY OF PLANT FOODS IN SALINE
X
X
Eucalyptus spp. (17)
FL
X
X
X
Frankenia spp. (5)
FL
X
X
Grevillea spp. (3)
FL
X
Hakea spp. (2)
FL
Leptospermum sp.
FL
Lomandra sp.
FL
Melaleuca spp. (18)
FL
Rinzia sp.
FL
Verticordia sp.
FL
X
Amyema sp.
FR
X
X
Billardiera lehmanniana
FR
X
X
X
Carpobrotus spp. (2)
FR
X
X
X
Dianella revoluta
FR
X
X
X
Dodonaea viscosa
FR
Enchylaena tomentosa
FR
X
Exocarpos aphyllus
FR
X
X
Leucopogon sp.
FR
X
X
Lycium australe
FR
X
X
Santalum acuminatum
FR
X
Santalum spicatum
FR
Scaevola spinescens
FR
Threlkeldia diffusa
FR
Acacia spp. (6)
GM
Hakea spp. (2)
GM
Atriplex spp. (5)
Nov
X
Oct
X
FL
Sep
FL
Eucalyptus loxophleba
Aug
Darwinia sp.
Spring
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
Jul
FL
Jun
Calytrix sp.
Winter
May
X
Apr
FL
Autumn
Mar
Calandrinia sp.
Feb
Part eaten
(category)
FL
Jan
Species
Baeckea sp.
Dec
Summer
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
L/R/S
X
X
Atriplex vesicaria
L/R/S
X
Calandrinia polyandra
L/R/S
Calandrinia sp.
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
L/R/S
X
X
X
X
X
X
X
X
X
X
X
X
Carpobrotus spp. (2)
L/R/S
X
X
X
X
X
X
X
X
X
X
X
X
Dianella revoluta
L/R/S
X
X
X
X
X
X
X
X
X
X
X
X
Disphyma crassifolium
L/R/S
X
X
X
X
X
X
X
X
X
X
X
X
Einadia nutans
L/R/S
X
X
X
X
X
X
X
X
X
X
X
X
Enchylaena tomentosa
L/R/S
X
X
X
X
X
X
X
X
X
X
X
X
Erodium cygnorum
L/R/S
X
X
X
X
X
X
X
X
X
X
X
X
412
Species
Eucalyptus loxophleba
Part eaten
(category)
L/R/S
Mar
Apr
May
Jun
Jul
Aug
Sep
Oct
Nov
Spring
Feb
Winter
Jan
Autumn
Dec
Summer
X
X
X
X
X
X
X
X
X
X
X
X
Frankenia spp. (5)
L/R/S
X
X
X
X
X
X
X
X
X
X
X
X
Gahnia spp. (2)
L/R/S
X
X
X
X
X
X
X
X
X
X
X
X
Lepidium sp.
L/R/S
X
X
X
X
X
X
X
X
X
X
X
X
Lomandra sp.
L/R/S
X
X
X
X
X
X
X
X
X
X
X
X
Lycium australe
L/R/S
X
X
X
X
X
X
X
X
X
X
X
X
Tecticornia indica
L/R/S
X
X
X
X
X
X
X
X
X
X
X
X
Threlkeldia diffusa
Thysanotus
manglesianus
Eucalyptus spp. (17)
L/R/S
X
X
X
X
X
X
X
X
X
X
X
X
L/R/S
X
X
X
X
X
X
X
X
X
X
X
X
L/M
X
X
X
Acacia spp. (6)
SE
X
X
X
X
X
X
Allocasuarina spp. (2)
SE
X
X
X
X
X
Atriplex vesicaria
SE
Calandrinia sp.
SE
X
X
X
X
X
Eragrostis dielsii
SE
X
X
X
Erodium cygnorum
SE
X
X
X
X
Hakea spp. (2)
SE
X
X
X
X
X
Lepidium sp.
SE
X
X
X
X
X
Caladenia spp. (3)
SO
X
X
X
X
Calandrinia polyandra
SO
X
X
X
X
X
X
X
X
X
X
X
X
Calandrinia sp.
SO
X
X
X
X
X
X
X
X
X
X
X
X
Drosera spp. (2)
SO
X
X
X
X
Platysace effusa
SO
X
X
X
X
X
X
X
X
X
X
X
X
Platysace maxwellii
Thysanotus
manglesianus
Triglochin sp.
SO
X
X
X
X
X
X
X
X
X
X
X
X
SO
X
X
X
X
X
X
X
X
X
X
X
X
SO
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
413
B.6 MONTHLY AVAILABILITY OF PLANT FOODS IN THICKET
Sep
Oct
Nov
Spring
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
Part eaten
(category)
FL
Banksia sphaerocarpa
FL
Banksia spp. (8)
FL
X
Beaufortia spp. (5)
FL
X
Calothamnus spp. (2)
FL
X
Calytrix spp. (4)
FL
Chamelaucium spp. (3)
FL
X
Darwinia sp.
FL
X
Enekbatus sp.
FL
Eremaea sp.
FL
Ericomyrtus sp.
FL
Eucalyptus loxophleba
FL
X
X
X
Eucalyptus spp. (15)
FL
X
X
X
Grevillea eriostachya
FL
X
X
X
Grevillea spp. (12)
FL
X
Hakea spp. (13)
FL
Leptospermum spp. (2)
FL
Lomandra sp.
FL
Melaleuca spp. (20)
FL
Micromyrtus spp. (3)
FL
Scholtzia sp.
FL
Thryptomene spp. (3)
FL
Thysanotus patersonii
FL
Verticordia spp. (11)
FL
X
Astroloma serratifolium
FR
X
Brachyloma spp. (2)
FR
X
Cassytha glabella
FR
X
X
X
X
X
X
Cassytha spp. (2)
FR
X
X
X
X
X
X
Dianella revoluta
FR
X
X
X
X
Exocarpos aphyllus
FR
X
X
X
Leucopogon spp. (4)
FR
X
X
Persoonia spp. (6)
FR
Santalum acuminatum
FR
X
Acacia acuminata
GM
X
X
X
Acacia spp. (16)
GM
X
X
X
Hakea spp. (13)
GM
X
X
X
Atriplex sp.
L/R/S
X
X
X
X
X
X
X
X
Dianella revoluta
L/R/S
X
X
X
X
X
X
X
X
Eucalyptus loxophleba
L/R/S
X
X
X
X
X
X
X
X
Jul
Species
Baeckea spp. (2)
Aug
Jun
Winter
May
Apr
Autumn
Mar
Feb
Jan
Dec
Summer
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
414
Species
Gahnia sp.
Part eaten
(category)
L/R/S
Mar
Apr
May
Jun
Jul
Aug
Sep
Oct
Nov
Spring
Feb
Winter
Jan
Autumn
Dec
Summer
X
X
X
X
X
X
X
X
X
X
X
X
Lomandra sp.
L/R/S
X
X
X
X
X
X
X
X
X
X
X
X
Oxalis sp.
L/R/S
X
X
X
X
X
X
X
X
X
X
X
X
Thysanotus patersonii
L/R/S
X
X
X
X
X
X
X
X
X
Eucalyptus oleosa
L/M
X
X
X
Eucalyptus spp. (15)
L/M
X
X
X
Acacia acuminata
SE
X
X
X
X
X
Acacia beauverdiana
SE
X
X
X
X
X
Acacia spp. (16)
SE
X
X
X
X
X
X
Allocasuarina spp. (3)
SE
X
X
X
X
X
X
X
X
X
X
X
X
Hakea spp. (13)
SE
X
X
X
X
X
X
X
X
X
X
X
X
Drosera spp. (2)
SO
X
X
X
X
Platysace effusa
SO
X
X
X
X
X
X
X
X
X
X
X
X
Platysace maxwellii
SO
X
X
X
X
X
X
X
X
X
X
X
X
Pterostylis sp.
SO
X
X
X
X
X
Thysanotus patersonii
SO
X
X
X
X
X
Thysanotus sp.
SO
X
X
X
X
X
X
X
X
X
X
X
X
X
415
B.7 MONTHLY AVAILABILITY OF PLANT FOOD IN WOODLAND
Sep
Oct
Nov
Spring
Aug
Winter
May
Autumn
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
FL
X
Calandrinia sp.
FL
X
Calothamnus sp.
FL
X
Calytrix spp. (2)
FL
Cyathostemon sp.
FL
Darwinia sp.
FL
X
Eremaea sp.
FL
X
Eucalyptus loxophleba
FL
X
X
X
Eucalyptus spp. (25)
FL
X
X
X
Eucalyptus wandoo
FL
X
X
Grevillea spp. (8)
FL
Hakea spp. (9)
FL
Kunzea sp.
FL
Leptospermum spp. (3)
FL
Lomandra sp.
FL
Melaleuca spp. (18)
FL
X
Regelia sp.
FL
X
Rinzia sp.
FL
X
X
Thryptomene sp.
FL
X
X
Thysanotus patersonii
FL
X
X
Verticordia spp. (2)
FL
Acrotriche sp.
FR
Amyema sp.
FR
X
X
Astroloma serratifolium
FR
X
X
X
X
Astroloma spp. (2)
FR
X
X
X
Brachyloma sp.
FR
X
Cassytha racemosa
FR
X
X
Cassytha sp.
FR
X
Dianella revoluta
FR
X
Dodonaea viscosa
FR
Enchylaena tomentosa
FR
X
Exocarpos aphyllus
FR
X
X
Exocarpos sparteus
FR
X
X
Leucopogon spp. (2)
FR
X
X
Persoonia spp. (3)
FR
Santalum acuminatum
FR
Santalum spicatum
FR
Jul
Beaufortia spp. (2)
Jun
X
Apr
FL
Mar
Banksia spp. (3)
Feb
Part eaten
(category)
FL
Jan
Species
Baeckea sp.
Dec
Summer
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
416
Jun
Jul
Aug
Sep
Oct
Nov
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
L/R/S
X
X
X
X
X
X
X
X
X
X
X
X
Enchylaena tomentosa
L/R/S
X
X
X
X
X
X
X
X
X
X
X
X
Erodium cygnorum
L/R/S
X
X
X
X
X
X
X
X
X
X
X
X
Eucalyptus loxophleba
L/R/S
X
X
X
X
X
X
X
X
X
X
X
X
Gahnia sp.
L/R/S
X
X
X
X
X
X
X
X
X
X
X
X
Lomandra sp.
L/R/S
X
X
X
X
X
X
X
X
X
X
X
X
Thysanotus patersonii
L/R/S
X
X
X
X
X
X
X
X
X
Eucalyptus spp. (25)
L/M
X
X
Acacia acuminata
SE
X
X
X
X
X
Acacia microbotrya
SE
X
X
X
Acacia saligna
SE
X
X
X
X
X
Acacia spp. (24)
SE
X
X
X
X
X
X
Allocasuarina spp. (4)
SE
X
X
X
X
X
X
Calandrinia sp.
SE
X
X
X
X
X
Enneapogon sp.
SE
Erodium cygnorum
SE
X
X
X
X
Hakea spp. (9)
SE
X
X
X
X
X
Caladenia spp. (2)
SO
X
X
X
X
Calandrinia sp.
SO
X
X
X
X
Cyanicula spp. (2)
SO
X
X
X
X
X
Diuris sp.
SO
X
X
X
X
X
Drosera spp. (2)
SO
X
X
X
X
X
Platysace maxwellii
SO
X
X
X
X
Pterostylis spp. (2)
SO
X
X
X
X
X
Thelymitra spp. (2)
SO
X
X
X
X
X
Thysanotus patersonii
SO
X
X
X
X
Triglochin sp.
SO
X
X
X
X
Part eaten
(category)
FR
Jan
X
Species
Scaevola spinescens
Dec
May
Spring
Apr
Winter
Mar
Autumn
Feb
Summer
X
X
Acacia acuminata
GM
X
X
X
Acacia microbotrya
GM
X
X
X
Acacia saligna
GM
X
X
X
Acacia spp. (24)
GM
X
X
X
Hakea spp. (9)
GM
X
X
X
Eucalyptus occidentalis
Pittosporum
phillyreoides
Atriplex spp. (3)
GM
X
X
X
GM
X
X
X
L/R/S
X
X
Calandrinia sp.
L/R/S
X
Dianella revoluta
L/R/S
Einadia nutans
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
417
APPENDIX C
DISTRIBUTION OF ANIMAL FOODS IN THE STUDY AREA
The table below indicates fauna present in the study area – all of which are considered
food sources (see Chapter 5.3.2). The list was compiled based on surveys conducted
in the study area (Chapman and Dell 1977; Dell 1977; Kitchener and Chapman 1977;
McKenzie et al. 1973; Muir 1978, 1979), with habitat and behavioural information from
other databases and publications (Atlas of Living Australia, n.d.; Cogger 2014; Cronin
2008, 2014; Nevill 2014; Simpson and Day 2004; Storr 1991; Western Australian
Museum n.d.). For ease of reference, species are listed alphabetically within animal
type, with taxonomy and naming conventions as per the Australian Faunal Directory
(Australian Biological Resources Study 2009); taxonomic changes made after August
2018 will not be reflected.
The first series of columns indicates the month/s during which a particular species was
present in the study area – a cross indicates that the animal was present in the study
area during the majority of the specific month. Yellow filled cells indicate the months
during which a particular species or bird or reptile may be laying eggs; grey filled cells
indicate the period during which female amphibians may contain eggs. Green filled
cells indicate when certain macropods are in peak condition.
The second series of columns (to the right of the thick black line) refers to the presence
of each species within the six Landscape Divisions. A cross indicates that the species
had been observed in a particular area, while an asterisk indicates that it likely
inhabited this area, based on generalised habitat information sourced from the
references listed above. A blank cell indicates that the animal was not found in this
particular area. The seasonal availability of animals and associated products will be
identical for each Landscape Division in which the animal can be found.
Landscape Divisions have been abbreviated, as below:
o
GR = Granite
o
SAL = Saline
o
HEA = Heath
o
TH = Thicket
o
MAL = Mallee
o
WL = Woodland
418
Jul
Aug
Sep
Oct
Nov
GR
X
X
X
X
X
X
X
X
X
X
X
X
Heleioporus albopunctatus
Amphibian
X
X
X
X
X
X
X
X
X
X
X
X
X
Myobatrachus gouldii
Amphibian
X
X
X
X
X
X
X
X
X
X
X
X
Pseudophryne guentheri
Amphibian
X
X
X
X
X
X
X
X
X
X
X
X
X
Pseudophryne occidentalis
Amphibian
X
X
X
X
X
X
X
X
X
X
X
X
X
Acanthagenys rufogularis
Bird
X
X
X
X
X
X
X
X
X
X
X
X
*
*
*
X
*
*
Acanthiza apicalis
Bird
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
Acanthiza chrysorrhoa
Bird
X
X
X
X
X
X
X
X
X
X
X
X
X
*
X
X
Acanthiza pusilla
Bird
X
X
X
X
X
X
X
X
X
X
X
X
*
*
*
Acanthiza uropygialis
Bird
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
Aegotheles cristatus
Bird
X
X
X
X
X
X
X
X
X
X
X
X
Anas gracilis
Bird
X
X
X
X
X
X
Anas superciliosa
Bird
X
Anthochaera carunculata
Bird
Anthrochaera chrystoptera
Bird
X
X
X
Anthus novaeseelandiae
Bird
X
X
Aquila audax
Bird
X
X
Ardea pacifica
Bird
Artamus cyanopterus
Bird
X
X
Artamus personatus
Barnardius zonarius
semitorquatus
Cacomantis flabelliformis
Bird
X
X
Bird
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
*
WL
Jun
X
TH
May
SAL
Animal
type
Amphibian
Apr
MAL
Species
Crinia pseudinsignifera
Mar
HEA
Spring
Feb
Winter
Jan
Autumn
Dec
Summer
*
X
*
*
*
*
X
X
X
X
*
*
X
*
*
*
X
*
X
*
*
X
*
X
X
X
*
*
X
X
X
X
X
X
*
*
Bird
X
X
X
X
X
X
*
Cacomantis pallidus
Bird
X
X
X
X
X
X
*
Calamanthus fuliginosus
Bird
X
X
X
X
X
X
X
X
X
X
X
X
Charadrius ruficapillus
Bird
X
X
X
X
X
X
X
X
X
X
X
X
X
*
*
*
*
*
*
*
X
*
*
*
X
419
SAL
TH
WL
*
*
*
*
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
Bird
X
X
X
X
X
X
X
X
X
X
X
X
Cygnus atratus
Bird
X
X
X
X
X
X
X
X
X
X
X
X
Dicaeum hirundinaceum
Bird
X
X
X
X
X
X
X
X
X
X
X
X
Dromaius novaehollandiae
Bird
X
X
X
X
X
X
X
X
X
X
X
X
Drymodes brunneopygia
Bird
X
X
X
X
X
X
X
X
X
X
X
X
Egretta novaehollandiae
Bird
X
X
X
X
X
X
Eopsaltria australis
Bird
X
X
X
X
X
X
X
X
X
X
X
X
*
*
X
Epthianura albifrons
Bird
X
X
X
X
X
X
X
X
X
X
X
X
*
X
X
Eurostopodus argus
Bird
X
X
X
X
X
X
*
X
Falco berigora
Bird
X
X
X
X
X
X
X
X
X
X
X
X
Falco cenchroides
Bird
X
X
X
X
X
X
X
X
X
X
X
X
*
*
*
X
*
*
Gavicalis virescens
Bird
X
X
X
X
X
X
X
X
X
X
X
X
*
X
X
X
*
X
Gerygone fusca
Bird
X
X
X
X
X
X
X
X
*
*
Gliciphila melanops
Bird
X
X
X
X
X
X
X
X
X
Grallina cyanoleuca
Bird
X
X
X
X
X
X
X
X
X
Himantopus himantopus
Bird
X
X
X
X
X
Cinclosoma castanotum
Bird
X
X
X
X
X
X
Climacteris rufa
Colluricincla harmonica
rufiventris
Coracina novaehollandiae
Bird
X
X
X
X
X
Bird
X
X
X
X
Corvus coronoides
Bird
X
X
Cracticus nigrogularis
Bird
X
Cracticus tibicen dorsalis
Bird
Cracticus torquatus
Bird
X
X
X
X
Apr
X
Mar
Bird
Feb
Cincloramphus cruralis
Jan
Animal
type
Bird
Jun
X
Species
Chrysococcyx basalis
Dec
Nov
X
Oct
*
Sep
*
Aug
*
Jul
MAL
Spring
HEA
Winter
GR
Autumn
May
Summer
*
*
X
X
*
X
X
*
*
*
*
X
X
X
*
X
X
*
X
X
X
X
*
X
X
X
X
X
*
X
X
*
*
*
*
X
X
*
*
*
*
*
X
X
*
*
X
X
X
X
X
X
*
X
X
X
*
*
*
*
X
420
X
X
X
X
X
X
X
X
X
X
Hylacola cauta
Bird
X
X
X
X
X
X
X
X
X
X
X
X
Lalage suerii
Bird
X
X
X
X
X
X
X
Leipoa ocellata
Bird
X
X
X
X
X
X
X
X
X
X
X
X
Lichenostomus cratitius
Bird
X
X
X
X
X
X
X
X
X
X
X
X
Lichmera indistincta
Bird
X
X
X
X
X
X
X
Lophoictinia isura
Bird
X
X
X
X
X
Malurus lamberti
Bird
X
X
X
X
X
X
X
X
X
X
X
X
*
*
*
*
Malurus pulcherrimus
Bird
X
X
X
X
X
X
X
X
X
X
X
X
*
X
X
X
*
Manorina flavigula
Bird
X
X
X
X
X
X
X
X
X
X
X
X
*
*
*
*
*
Melithreptus brevirostris
Bird
X
X
X
X
X
X
X
X
X
X
X
X
*
*
X
*
X
Merops ornatus
Bird
X
X
X
X
X
Microeca fascinans
Bird
X
X
X
X
X
X
X
X
X
X
X
X
*
Myiagra inquieta
Bird
X
X
X
X
X
X
X
X
X
X
X
X
*
*
Neophema elegans
Bird
X
X
X
X
X
X
X
X
X
X
X
X
*
*
X
X
Nesoptilotis leucotis
Bird
X
X
X
X
X
X
X
X
X
X
X
X
*
*
X
Ninox novaeseelandiae
Bird
X
X
X
X
X
X
X
X
X
X
X
X
*
Nymphicus hollandicus
Bird
X
X
X
X
X
X
X
X
X
X
X
X
*
Oreoica gutturalis
Bird
X
X
X
X
X
X
X
X
X
X
X
X
*
X
X
Pachycephala pectoralis
Bird
X
X
X
X
X
X
X
X
X
X
X
X
*
*
X
Pachycephala rufiventris
Bird
X
X
X
X
X
X
X
X
X
X
X
X
*
*
Pardalotus striatus
Bird
X
X
X
X
X
X
X
X
X
X
X
X
Parvipsitta porphyrocephala
Bird
X
X
X
X
X
X
X
X
X
X
X
X
*
Petrochelidon nigricans
Bird
X
X
X
X
X
X
X
*
X
Petroica goodenovii
Bird
X
X
X
X
X
X
X
*
X
X
X
X
X
X
*
X
*
X
*
*
X
X
X
X
WL
X
TH
Nov
SAL
Oct
MAL
Sep
X
GR
Aug
X
Jul
*
Jun
X
May
*
Animal
type
Bird
Apr
*
Species
Hirundo neoxena
Mar
HEA
Spring
Feb
Winter
Jan
Autumn
Dec
Summer
*
*
*
*
*
*
*
*
X
X
X
*
*
*
*
*
*
X
*
*
*
X
X
X
X
X
X
*
*
X
*
X
X
*
X
X
*
*
*
421
Species
Phaps chalcoptera
Animal
type
Bird
Mar
Apr
May
Jun
Jul
Aug
Sep
Oct
Nov
GR
HEA
MAL
SAL
TH
WL
Spring
Feb
Winter
Jan
Autumn
Dec
Summer
X
X
X
X
X
X
X
X
X
X
X
X
*
*
*
X
X
*
Phylidonyris niger
Bird
X
X
X
X
X
X
X
X
X
X
X
X
*
*
*
Platycercus icterotis
Bird
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
Platycercus varius
Bird
X
X
X
X
X
X
X
X
X
X
X
X
*
*
*
X
*
*
Platycercus zonarius
Bird
X
X
X
X
X
X
X
X
X
X
X
X
*
X
X
X
X
X
Podargus strigoides
Bird
X
X
X
X
X
X
X
X
X
X
X
X
*
Polytelis anthopeplus
Bird
X
X
X
X
X
X
X
X
X
X
X
X
*
*
*
X
*
*
Pomatostomus superciliosus
Bird
X
X
X
X
X
X
X
X
X
X
X
X
*
*
*
X
X
*
Ptilotula ornata
Bird
X
X
X
X
X
X
X
X
X
X
X
X
*
*
*
*
*
Pyrrholaemus brunneus
Bird
X
X
X
X
X
X
X
X
X
X
X
X
X
*
X
Rhipidura leucophrys
Bird
X
X
X
X
X
X
X
X
X
X
X
X
X
Sericornis frontalis
Bird
X
X
X
X
X
X
X
X
X
X
X
X
*
Smicrornis brevirostris
Bird
X
X
X
X
X
X
X
X
X
X
X
X
*
Stipiturus malachurus
Bird
X
X
X
X
X
X
X
X
X
X
X
X
Strepera versicolor
Bird
X
X
X
X
X
X
X
X
X
X
X
X
Tadorna tadornoides
Bird
X
X
X
X
Turnix varius
Bird
X
X
X
X
X
X
X
X
X
X
X
X
Vanellus tricolor
Bird
X
X
X
X
X
X
X
X
X
X
X
X
Zosterops lateralis
Bird
X
X
X
X
X
X
X
X
X
X
X
X
Antechinomys laniger spenceri
Mammal
X
X
X
X
X
X
X
X
X
X
X
Cercartetus concinnus
Mammal
X
X
X
X
X
X
X
X
X
X
Chalinolobus gouldii
Mammal
X
X
X
X
X
X
X
X
X
X
Dasyurus geoffroii
Mammal
X
X
X
X
X
X
X
X
X
Isoodon obesulus
Mammal
X
X
X
X
X
X
X
X
Macropus fuliginosus
Mammal
X
X
X
X
X
X
X
X
*
X
*
*
X
X
*
X
X
X
X
X
X
X
X
X
*
X
X
X
*
X
X
X
*
*
*
*
X
*
*
*
*
*
*
X
X
*
*
X
X
*
X
X
X
X
*
X
X
X
X
X
*
X
X
X
X
*
*
*
X
X
X
X
X
X
X
X
*
*
X
X
422
Jul
Aug
Sep
Oct
Nov
X
X
X
X
X
X
X
X
X
X
X
Notamacropus irma
Mammal
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
Notomys mitchellii
Mammal
X
X
X
X
X
X
X
X
X
X
X
X
X
*
X
X
X
Nyctophilus geoffroyi
Mammal
X
X
X
X
X
X
X
X
X
X
X
X
X
Osphranter robustus
Mammal
X
X
X
X
X
X
X
X
X
X
X
X
*
Phascogale calura
Mammal
X
X
X
X
X
X
X
X
X
X
X
X
X
Pseudomys occidentalis
Mammal
X
X
X
X
X
X
X
X
X
X
X
X
X
Sminthopsis crassicaudata
Mammal
X
X
X
X
X
X
X
X
X
X
X
X
Sminthopsis granulipes
Mammal
X
X
X
X
X
X
X
X
X
X
X
X
*
X
Sminthopsis murina
Mammal
X
X
X
X
X
X
X
X
X
X
X
X
X
*
Tachyglossus aculeatus
Mammal
X
X
X
X
X
X
X
X
X
X
X
X
*
*
Tarsipes rostratus
Mammal
X
X
X
X
X
X
X
X
X
X
X
X
Trichosurus vulpecula
Mammal
X
X
X
X
X
X
X
X
X
X
X
X
X
Vespadelus pumilus
Mammal
X
X
X
X
X
X
X
X
X
X
X
X
X
Anilios australis
Reptile
X
X
X
X
X
X
X
X
X
X
X
X
*
Crenadactylus ocellatus
Cryptoblepharus
plagiocephalus
Ctenophorus cristatus
Reptile
X
X
X
X
X
X
X
X
X
X
X
X
*
Reptile
X
X
X
X
X
X
X
X
X
X
X
X
X
Reptile
X
X
X
X
X
X
X
X
X
X
X
X
*
Ctenophorus maculatus
Reptile
X
X
X
X
X
X
X
X
X
X
X
X
Ctenophorus ornatus
Reptile
X
X
X
X
X
X
X
X
X
X
X
X
X
Ctenophorus salinarum
Reptile
X
X
X
X
X
X
X
X
X
X
X
X
X
Ctenotus impar
Reptile
X
X
X
X
X
X
X
X
X
X
X
Ctenotus pantherinus
Reptile
X
X
X
X
X
X
X
X
X
X
Ctenotus schomburgkii
Reptile
X
X
X
X
X
X
X
X
X
Delma australis
Reptile
X
X
X
X
X
X
X
X
X
WL
Jun
X
TH
May
SAL
Apr
MAL
Animal
type
Mammal
GR
Species
Myrmecobius fasciatus
Mar
HEA
Spring
Feb
Winter
Jan
Autumn
Dec
Summer
*
*
*
*
X
X
X
X
X
X
X
*
*
*
X
*
X
X
*
*
*
*
X
*
X
*
X
*
X
X
X
*
X
*
X
X
*
X
*
X
X
X
*
X
X
X
X
*
*
*
X
*
*
423
May
Jun
Jul
Aug
Sep
Oct
Nov
GR
HEA
X
X
X
X
X
X
X
X
X
X
X
*
*
Demansia sp.
Reptile
X
X
X
X
X
X
X
X
X
X
X
X
Egernia kingii
Reptile
X
X
X
X
X
X
X
X
X
X
X
X
*
Gehyra variegata
Reptile
X
X
X
X
X
X
X
X
X
X
X
X
X
Hesperoedura reticulata
Reptile
X
X
X
X
X
X
X
X
X
X
X
X
Lerista distinguenda
Reptile
X
X
X
X
X
X
X
X
X
X
X
X
*
*
Lialis burtonis
Reptile
X
X
X
X
X
X
X
X
X
X
X
X
*
*
*
Liopholis multiscutata
Reptile
X
X
X
X
X
X
X
X
X
X
X
X
*
X
X
Lucasium maini
Reptile
X
X
X
X
X
X
X
X
X
X
X
X
*
X
Morethia obscura
Reptile
X
X
X
X
X
X
X
X
X
X
X
X
X
*
*
Parasuta gouldii
Reptile
X
X
X
X
X
X
X
X
X
X
X
X
*
*
*
*
*
Pogona minor
Reptile
X
X
X
X
X
X
X
X
X
X
X
X
X
*
X
X
Strophurus spinigerus
Reptile
X
X
X
X
X
X
X
X
X
X
X
X
*
X
*
Tiliqua occipitalis
Reptile
X
X
X
X
X
X
X
X
X
X
X
X
*
*
X
Tiliqua rugosa
Reptile
X
X
X
X
X
X
X
X
X
X
X
X
*
*
X
Underwoodisaurus milii
Reptile
X
X
X
X
X
X
X
X
X
X
X
X
*
*
*
Varanus gouldii
Reptile
X
X
X
X
X
X
X
X
X
X
X
X
*
*
*
WL
Apr
X
TH
Animal
type
Reptile
SAL
Species
Delma fraseri
Mar
MAL
Spring
Feb
Winter
Jan
Autumn
Dec
Summer
*
X
X
X
*
*
*
*
*
*
*
*
X
X
*
X
*
X
*
*
X
424
APPENDIX D
MEASURING ARTEFACT ATTRIBUTES
Artefactual material was analysed by XU within each test-pit. Relevant attributes were
recorded in a database of Microsoft Excel spreadsheets. Due to the issues associated
with often poor-quality quartz, some attributes could not be observed on particular
artefacts. Selected attributes are discussed in more detail below, where earlier
methodological information (see Chapter 9.3.3) was insufficient.
D.1 METRIC ATTRIBUTES
Artefacts were weighed on digital scales that measured weight in 0.05 g increments.
All artefacts were weighed individually except for flaked fragments and flakes
measuring less than 10 mm in percussion/maximum length – these were weighed as
a group containing other artefacts of the same raw material and type, from the same
pit and XU. Length, width and thickness measurements were taken using digital
calipers that measured in 0.01 mm increments.
Maximum dimensions were recorded for flaked fragments and cores as the former
cannot be oriented, and there is no orientation system that applies to all varieties of
core (Holdaway and Stern 2004:186). Oriented dimensions were recorded for flakes
and retouched flakes where the flaking axis could be identified (i.e. platform to
termination). Where this was not possible (often due to breakage or extensive retouch)
maximum dimensions were recorded, but in a slightly different manner than for cores
and flaked fragments due to the presence of distinct dorsal and ventral surfaces.
These dimensions were measured as below:
Maximum dimensions:
o
Maximum length: a line drawn between the two points separated by the
greatest distance on the same surface (flakes, retouched flakes) or any
surface (cores, flaked fragments);
o
Maximum width: the longest line that can be drawn at 90 degrees to
maximum length;
425
o
Maximum thickness: the longest line that can be drawn at 90 degrees to
maximum width (cores, flaked fragments) or the maximum distance
between the dorsal and ventral surfaces (flakes, retouched flakes).
Oriented dimensions:
o
Percussion length: a line measured from the platform (bisecting bulb of
percussion/impact point, where present) to termination, following the flaking
axis (i.e. the direction that the fracture has progressed from impact until
exiting the core);
o
Width at midpoint: measured at the midpoint of percussion length, at right
angles;
o
Thickness at midpoint: distance between the dorsal and ventral surfaces,
measured at the point where percussion length intersects with width at
midpoint.
D.2 RETOUCH
As discussed earlier (see Chapter 9.3.2), retouch was identified where additional flake
removals occurred after the ventral flake surface was created – negative flake scars
should therefore occur on the ventral surface itself or be initiated from this newly
created surface (Hiscock 2007). In the latter case, at least some complete scars
should remain (i.e. the latest in the sequence, if scar superimposition is evident);
micro-scars resulting from edge damage are not considered retouch. If only partial
negative scars are preserved on flake margins, these instead represent flake removals
that occurred when the artefact was still attached to the parent core – scars that are
subsequently split when the new flake is created – not retouch (Hiscock 2007;
Holdaway and Stern 2004:161). While it should therefore be relatively easy to separate
retouch from edge damage and previous flake removals, in practice, poor-quality raw
material hampered some efforts; where retouch was ambiguous, pieces were
classified as flakes. Two types of retouch were encountered, defined below:
o
Backing: steep uni- or bidirectional retouch that unites the dorsal and ventral
surfaces by creating a blunt margin located opposite the working edge –
retouch can be created via the bipolar technique, so crushing may be
evident (Hiscock 1993; Holdaway and Stern 2004:159).
426
o
Unifacial retouch: retouch occurring on only one surface of the flake, either
dorsal or ventral, modifying the inferred working edge (Holdaway and Stern
2004:158–159).
D.3 BREAKAGE
The presence/absence of breakages were noted for flakes and retouched flakes. The
type of break was quantified, where possible, but in many cases poor quality raw
material prevented this. Breakage types and preserved portions were classified
following Hiscock (2002b), as illustrated in Figure D.1.
Figure D.1 Breakage types and the resultant flake fragments produced (Hiscock 2002b:253).
427
D.4 FLAKE SCAR ORIENTATION
Negative flake scar orientation was recorded following Holdaway and Stern
(2004:146–147), via the use of a quadrat system – this method could only be applied
to artefacts that could be oriented along the flaking axis, where flake scar directionality
was not obscured by poor-quality raw material. An 'x' is superimposed over the artefact,
with the central point positioned in the midpoint of the dorsal surface, and the end of
the upper arms joining each edge of the platform; where the platform was missing, the
piece was oriented following the flaking axis and the arms positioned at the 'corners'
of the flake, i.e. where the left and right lateral margins intersected with the break.
Quadrats are numbered in a clockwise direction, beginning with the quadrat that
contained/would have contained the platform (Figure D.2).
Figure D.2 Method of recording flake scar orientation on flakes and retouched flakes (Holdaway and
Stern 2004:147).
Each quadrat was evaluated to determine whether flakes had been initiated therein,
based on negative bulbs of percussion, ripple mark orientation, and/or general scar
topography. It was important to specify the precise quadrats containing initiations –
rather than simply counting how many quadrats contain them – as the number of core
rotations represented will depend on whether the platform quadrat (Q1) preserved
evidence of prior flake removals. For example, a flake exhibiting flake removals from
just one quadrat may indicate no core rotation (when Q1 contains the initiations, thus
428
previous flakes were detached from the same platform as the current flake) or a single
rotation (Q2–Q4, when the quadrat exhibiting initiations does not contain the platform).
Therefore, the presence/absence of flake initiations was recorded for quadrats one to
four on each suitable flake.
D.5 BIPOLAR REDUCTION
Bipolar flaking involves three separate components: a core, a hammerstone and an
anvil, but the arrangement of these elements can vary. The bipolar technique is often
separated into axial and non-axial reduction, depending on the spatial relationship
between the impact point (where the hammerstone strikes the core) and the area
where the core contacts the anvil. In axial reduction, the blow is directed towards the
anvil, compressing the core between the hammerstone and anvil, and detaching a
flake (Diez-Martin et al. 2011; Hiscock 2015). In non-axial or oblique reduction, core
placement and/or direction of blow means that force exits the core without reaching
the anvil – the only thing that differentiates this technique from freehand knapping is
the use of an anvil to stabilise the core. These major methodological differences affect
the form of flaked products, and indeed the lack of compressive forces in non-axial
techniques makes those products difficult to differentiate from those created via
freehand flaking (Diez-Martin et al. 2011; Hiscock 2015). Therefore, only axial bipolar
is considered here; Hiscock (2015) considers this a 'strong' definition for bipolar flaking.
There is no single feature that signifies the use of the axial bipolar technique. Instead,
raw material type and quality dictate bipolar and freehand signatures, so diagnostic
features must be appropriate to a particular raw material type (De La Pena 2015;
Jeske and Lurie 1993, Pargeter and De La Pena 2017). For quartz, the features that
most accurately separate the products of bipolar from freehand knapping are crushing
and termination type (Figure D.3): bipolar terminations are generally axial or splintered
and flakes often preserve platform segments at the distal end, along with extensive
crushing at the platform and distal margin; freehand flakes and cores rarely exhibit
crushing, instead preserving feather, step, hinge or overshot (aka plunging)
terminations (Diez-Martin et al. 2011; Pargeter and De La Pena 2017). The bipolar
technique was identified by the presence of one or more of the diagnostic features
noted above. Those lacking any such features (i.e. exhibiting the freehand counterpart)
429
were considered to be the product of freehand reduction. As with other attributes, the
presence/absence of bipolar reduction could not always be observed on poor-quality
raw material.
A
B
Figure D.3 Flake terminations. A: side views of different terminations showing the path the fracture
takes through the core (Cotterell and Kamminga 1987:684). B: dorsal surface of a flake with a splintered
termination – distal end is at the bottom of the photograph (Pargeter and De La Pena 2017:10).
430
APPENDIX E
ANDERSON ROCKS STONE ARTEFACT ASSEMBLAGE
This appendix contains the raw artefact data for the Anderson Rocks stone artefact
assemblages. Weight (grams) and dimensions (millimetres) have been rounded to
two decimal places. Blank cells denote unusable or missing data. Unusable data
was that deemed not representative of a particular attribute, i.e. mm/scar for a flake
that could not be oriented. Where data were absent, this indicated that a particular
attribute could not be confidently measured on a certain specimen, often due to
poor-quality raw material. Artefact types have been ordered by pit/scatter
(beginning with surface material), XU depth and raw material. See Tables 9.1–9.2
and Appendix D for details on attributes recorded for each artefact type.
Abbreviations are as follows:
o
Bck. = backed
o
Ret. = retouch type
o
BIF = Banded Iron Formation
o
RM = raw material
o
Bk = breakage
o
Sc. = scatter
o
Chal. = chalcedony
o
Sc1 = Scatter 1
o
Comb wt. = combined weight
o
Sc2 = Scatter 2
o
Dims = dimensions
o
Sc3 = Scatter 3
o
ED = edge damage
o
Sil. = silcrete
o
HQ = high-quality raw material
o
T = thickness
o
L = length
o
Trans. = transitional
o
Lng. = longitudinal
o
Trv. = transverse
o
Max. = maximum dimensions
o
Uni = unifacial
o
Mrg. = marginal
o
%Unmod = proportion of
o
#NFS = no. of negative flake scars
o
Ort. = oriented dimensions
o
W = width
o
OSM = other surface material
o
Wt = weight
o
#Plat = no. of platforms
o
XU = excavation unit
o
Qtz = quartz
unmodified surface
431
Pit/Sc.
XU
RM
HQ
Wt (g)
L (mm)
W (mm)
T (mm)
#NFS
g/scar
#Plat
%Unmod
Bipolar
E.1 CORES
Sc1
-
Qtz
N
4.9
23.62
16.39
15.13
5
0.98
2
< 25
Y
Sc1
-
Qtz
N
5.3
24.43
20.82
15.02
8
0.66
4
< 25
Y - trans
Sc2
-
Qtz
N
2.65
20
15.22
8.99
7
0.38
3
< 25
N
Sc3
OSM
-
Qtz
Qtz
N
N
5.3
46.65
25.59
41.9
21.53
35.04
10.51
32.39
6
9
0.88
5.18
4
4
< 25
< 25
N
Y - trans
AR2-3
6
Qtz
N
6.2
25.05
19.72
10.86
4
1.55
2
AR2-3
8
Qtz
Y
2.45
22.67
12.97
7.47
4
0.61
4
< 25
Y
AR2-3
AR2-3
9
13
Qtz
Qtz
Y
N
1.85
3.4
15.7
22.62
13.06
14.94
9.57
10.54
8
8
0.23
0.43
2
3
< 25
< 25
Y
AR2-3
15
Qtz
Y
1.15
14.56
13.2
6.7
4
0.29
3
25-50
Y
AR2-3
16
Qtz
N
1.45
15.69
11.08
9.56
9
0.16
4
< 25
Y
AR2-3
16
Qtz
N
1.6
17.9
14.11
6.55
3
0.53
3
25-50
Y
AR2-3
17
Qtz
N
5.5
22.95
21.65
12.91
5
1.10
3
AR2-3
17
Qtz
N
10.65
31.53
25.74
14.21
5
2.13
3
25-50
N
AR2-3
18
Qtz
N
5.3
22.93
22.55
12.51
4
1.33
2
25-50
Y
AR2-3
AR2-3
18
19
Qtz
Qtz
N
Y
6.25
2.25
26.93
22.68
19.69
17.85
10.35
7.64
7
4
0.89
0.56
4
4
< 25
< 25
Y
Y
AR2-3
20
Qtz
N
3.35
23.55
18.77
8.69
8
0.42
4
< 25
Y
AR2-3
21
Qtz
Y
6.15
30.33
17.28
10.78
7
0.88
4
< 25
N
432
E.2 SMALL FLAKED FRAGMENTS
XU
RM
No.
Comb.
Wt (g)
No. HQ
Sc1
-
Qtz
1
0.5
0
Sc3
-
Qtz
2
0.35
2
AR1
1
Qtz
3
0.25
0
AR1
2
Qtz
3
0.15
0
AR1
3
Qtz
3
0.2
0
AR1
4
Qtz
4
0.45
3
AR1
5
Qtz
7
0.7
0
AR1
6
Qtz
3
0.35
0
AR1
7
Qtz
7
0.3
0
AR1
8
BIF
1
0.05
1
AR1
8
Qtz
8
0.45
0
AR1
AR1
9
10
Qtz
Qtz
8
2
0.8
0.15
0
0
AR1
11
Qtz
4
0.7
0
AR1
12
Qtz
1
0.05
0
AR1
AR2-3
13
1
Qtz
Qtz
1
1
0.2
< 0.05
0
1
AR2-3
2
Qtz
1
0.1
1
AR2-3
3
Qtz
1
0.05
0
AR2-3
AR2-3
5
6
Qtz
Qtz
1
7
< 0.05
1.45
1
2
AR2-3
7
Qtz
1
0.15
0
AR2-3
9
Qtz
1
0.1
0
AR2-3
AR2-3
10
11
Qtz
Qtz
3
1
0.1
0.05
3
0
AR2-3
12
Qtz
2
0.3
0
AR2-3
13
Qtz
3
0.35
2
AR2-3
AR2-3
15
16
Qtz
Qtz
5
3
0.3
0.15
0
1
AR2-3
17
Qtz
2
0.05
2
AR2-3
18
Qtz
25
2
9
AR2-3
AR2-3
19
20
Qtz
Qtz
5
2
0.3
0.2
1
0
AR2-3
21
Qtz
3
0.2
3
AR2-3
22
Qtz
2
0.35
0
AR2-3
23
Qtz
3
0.2
3
AR2-3
25
Qtz
1
0.25
0
AR2-3
26
Qtz
1
< 0.05
0
AR2-3
28
Qtz
1
< 0.05
0
AR2-3
AR2-3
29
31
Qtz
Qtz
1
1
< 0.05
0.05
1
1
Pit/Sc.
433
Pit/Sc.
XU
RM
HQ
Wt (g)
L (mm)
W (mm)
T (mm)
#NFS
g/scar
#Plat
E.3 LARGE FLAKED FRAGMENTS
Sc1
-
Qtz
N
4.6
21.56
17.05
11.04
2
2.30
2
Sc2
Sc2
-
Qtz
Qtz
N
N
0.7
1.75
12.9
20.57
9.5
12.99
6.2
6.8
1
2
0.70
0.88
1
1
Sc2
-
Qtz
N
2.35
17.44
12.7
9.3
3
0.78
3
Sc2
-
Qtz
N
2.55
22.71
12.91
8.48
2
1.28
1
Sc3
Sc3
-
Qtz
Qtz
Y
Y
0.35
0.35
11.85
11.26
7.34
6.93
4.54
5.04
2
3
0.18
0.12
2
2
OSM
-
Qtz
N
6.25
23.57
22.05
11.64
7
0.89
2
OSM
-
Qtz
N
6.75
30.68
20.27
10.56
AR1
AR1
3
8
Qtz
Qtz
N
N
10.45
0.4
37.51
11.62
22.96
7.16
10.81
4.49
1
10.45
1
AR1
9
Qtz
N
0.15
10.16
4.72
3.77
AR1
9
Qtz
N
0.2
10.19
5.18
4.16
2
0.10
2
AR1
9
Qtz
N
0.45
11.8
10.86
3.73
AR1
9
Qtz
N
2.15
23.02
11.22
7.43
3
0.72
2
AR1
10
Qtz
N
0.4
11.15
7.18
5.16
3
0.13
1
AR1
13
Qtz
N
0.4
12.58
6.07
4.32
2
0.20
1
AR1
13
Qtz
N
0.6
11.46
5.7
5.19
AR2-3
2
Qtz
N
0.2
12.71
4.41
2.85
3
0.07
1
AR2-3
5
Qtz
N
0.4
11.99
6.45
4.8
3
0.13
2
AR2-3
6
Qtz
N
6.3
27.66
20.23
12.1
4
1.58
2
AR2-3
AR2-3
7
11
Qtz
Qtz
Y
N
0.5
0.25
14.09
12.56
6.09
5.54
4.95
4.81
1
2
0.50
0.13
1
2
AR2-3
14
Qtz
N
0.65
14.7
8.53
5.53
4
0.16
2
AR2-3
14
Qtz
N
1.1
15.03
12.54
6.5
2
0.55
2
AR2-3
AR2-3
17
18
Qtz
Qtz
Y
N
0.2
0.2
15.01
11.11
4.43
5.46
3.52
3.77
3
1
0.07
0.20
2
1
AR2-3
18
Qtz
N
0.35
13.01
7.79
4.83
3
0.12
2
AR2-3
18
Qtz
N
0.85
13.44
10.15
6.8
4
0.21
3
AR2-3
AR2-3
19
19
Qtz
Qtz
N
N
0.15
0.2
10.38
10.3
6.95
4.65
2.97
3.61
2
3
0.08
0.07
2
2
AR2-3
19
Qtz
N
0.6
12.14
10.16
4.97
2
0.30
2
AR2-3
19
Qtz
N
0.65
11.84
9.68
5.18
3
0.22
2
AR2-3
AR2-3
19
19
Qtz
Qtz
N
N
3.15
5.15
19.56
26.73
14.15
18.44
9.98
10.33
3
2
1.05
2.58
2
2
AR2-3
20
Qtz
N
2.25
22.9
15.56
5.79
4
0.56
3
AR2-3
23
Qtz
N
0.25
12.53
6.3
3.83
2
0.13
2
AR2-3
23
Qtz
N
0.5
10.01
6.63
8.24
2
0.25
2
AR2-3
25
Qtz
N
2.15
27.43
11.92
7.17
434
E.4 SMALL FLAKES
XU
-
RM
Qtz
No.
4
Comb.
Wt (g)
0.55
No. HQ
1
No.
broken
0
AR1
2
Qtz
1
< 0.05
0
0
AR1
3
Qtz
4
0.15
0
3
AR1
AR1
4
5
Qtz
Qtz
1
2
0.05
0.35
1
1
1
1
AR1
6
Sil.
1
< 0.05
0
0
AR1
8
Qtz
2
0.2
1
1
AR1
AR1
9
9
Qtz
Sil.
4
1
0.45
0.05
1
0
1
0
AR1
10
Qtz
1
< 0.05
1
1
AR1
11
Qtz
1
0.05
1
1
AR1
AR2-3
12
2
Qtz
Qtz
3
2
0.1
0.2
3
2
1
1
AR2-3
3
BIF
1
< 0.05
1
0
AR2-3
3
Qtz
6
0.35
2
1
AR2-3
AR2-3
4
5
Qtz
Qtz
3
9
0.05
0.2
3
6
3
5
AR2-3
6
BIF
1
0.1
1
0
AR2-3
6
Qtz
7
0.5
0
1
AR2-3
AR2-3
7
8
Qtz
Qtz
10
8
0.5
0.35
1
4
5
6
AR2-3
9
Qtz
10
0.55
10
8
AR2-3
10
Qtz
8
0.35
8
7
AR2-3
AR2-3
11
12
Qtz
Qtz
7
14
0.15
0.7
7
7
6
9
AR2-3
13
Chal.
2
0.05
2
0
AR2-3
13
Qtz
10
0.75
8
8
AR2-3
AR2-3
14
15
Qtz
BIF
20
2
1.8
< 0.05
10
2
11
0
AR2-3
15
Qtz
30
1.5
18
14
AR2-3
16
Qtz
17
0.85
9
5
AR2-3
AR2-3
17
17
BIF
Qtz
1
31
< 0.05
1.35
1
25
0
12
AR2-3
18
BIF
1
0.35
1
0
AR2-3
18
Qtz
71
4.7
55
41
AR2-3
AR2-3
19
20
Qtz
Chal.
36
1
2.55
0.1
33
1
25
1
AR2-3
20
Qtz
29
2
19
18
AR2-3
21
Qtz
26
1.8
11
19
AR2-3
22
Qtz
21
0.95
17
12
AR2-3
22
Sil.
1
< 0.05
0
0
Pit/Sc.
Sc3
435
Pit/Sc.
AR2-3
XU
23
RM
BIF
No.
1
Comb.
Wt (g)
< 0.05
No. HQ
1
No.
broken
0
AR2-3
23
Qtz
10
1
0
3
AR2-3
24
Qtz
1
< 0.05
0
0
AR2-3
AR2-3
25
26
Qtz
BIF
5
1
0.7
< 0.05
0
1
2
0
AR2-3
26
Qtz
1
< 0.05
0
0
AR2-3
27
Qtz
2
0.75
2
0
AR2-3
AR2-3
28
28
BIF
Qtz
1
1
0.05
0.15
1
0
0
0
AR2-3
29
Qtz
1
< 0.05
1
1
AR2-3
30
Qtz
3
0.1
3
2
AR2-3
AR2-3
31
32
Qtz
Qtz
1
2
< 0.05
0.1
1
2
0
2
436
HQ
Wt (g)
Dims.
L (mm)
W (mm)
T (mm)
Bk
#NFS
mm/scar
#Plat
%Unmod
Bipolar
ED
-
Qtz.
N
0.4
Ort.
13.02
5.96
3.52
N
3
4.34
1
< 25
N
N
Sc1
Sc1
-
Qtz.
Qtz.
N
N
0.55
1.5
Ort.
Ort.
10.42
10.15
8.36
13.4
4.83
6.36
N
N
3
4
3.47
2.54
2
2
< 25
< 25
Y
Y
N
N
Sc2
-
Qtz.
N
0.8
Ort.
14.54
12.66
3.55
N
2
7.27
N
N
Sc2
-
Qtz.
N
1.25
Ort.
14.83
11.35
6.2
N
1
14.83
1
Y
N
Sc2
Sc2
-
Qtz.
Qtz.
N
N
1.65
2.3
Ort.
Ort.
15.11
30.3
11.56
11.78
5.47
5.14
N
4
2
3.78
15.15
2
3
< 25
Y
Y
N
N
Sc3
-
Chal.
Y
0.25
Ort.
10.17
10.48
1.83
Y
1
10.17
< 25
N
N
Sc3
-
Qtz.
Y
0.15
Max.
12.38
6.06
3.98
Y
4
< 25
N
Sc3
Sc3
-
Qtz.
Qtz.
Y
Y
0.15
0.15
Max.
Ort.
12.14
10.02
7.97
7.86
1.58
1.42
Y
N
1
1
10.02
1
< 25
< 25
N
N
Y
Sc3
-
Qtz.
N
0.25
Ort.
11.9
10.97
2.21
N
4
2.98
3
< 25
N
Y
Sc3
-
Qtz.
Y
0.25
Ort.
12.38
7.7
2.25
N
3
4.13
2
< 25
Y
Y
Sc3
Sc3
-
Qtz.
Qtz.
N
Y
0.25
0.35
Ort.
Ort.
12.73
13.44
5.6
5.98
2.91
3.34
N
N
2
3
6.37
4.48
1
3
< 25
< 25
N
N
N
N
Sc3
-
Qtz.
Y
0.4
Ort.
11.92
10.66
2.48
N
4
2.98
2
< 25
N
N
Sc3
-
Qtz.
N
0.5
Ort.
13.39
8.59
2.54
N
2
6.70
1
< 25
Y
N
Sc3
Sc3
-
Qtz.
Qtz.
N
Y
0.5
0.65
Max.
Ort.
19.53
10.82
12.2
12.21
1.99
4.13
Y
Y
1
3
3.61
2
< 25
< 25
N
N
N
Sc3
-
Qtz.
Y
1.1
Ort.
22.36
8.71
4.34
N
3
7.45
3
< 25
Y
N
Sc3
-
Qtz.
N
2.4
Ort.
20.02
18.08
5.35
N
4
5.01
2
< 25
Y
N
OSM
AR1
1
Qtz.
Qtz.
N
N
0.4
0.25
Ort.
Ort.
10.54
10.87
11.64
11.3
2.32
2.33
N
N
5
3
2.11
3.62
2
2
< 25
< 25
N
N
N
Bk type
RM
Sc1
Pit/Sc.
XU
E.5 LARGE FLAKES
Trv.
Trv.
437
HQ
Wt (g)
Dims.
L (mm)
W (mm)
T (mm)
Bk
#NFS
mm/scar
#Plat
%Unmod
Bipolar
ED
Y
0.55
Ort.
12.8
7.68
3.89
N
2
6.40
2
< 25
Y
N
AR1
4
Qtz.
N
0.5
Ort.
11.98
7.72
3.88
N
2
5.99
2
< 25
N
N
AR1
AR1
4
4
Qtz.
Sil.
N
N
4.7
0.2
Ort.
Ort.
26.57
16
18.6
8.98
6.36
1.09
N
1
2
26.57
8.00
1
1
< 25
< 25
Y
N
N
N
AR1
8
Qtz.
Y
0.3
Ort.
12.44
8.04
1.88
N
2
6.22
1
< 25
N
Y
AR1
9
Qtz.
N
0.15
Ort.
9.86
6.2
2.21
Y
3
3.29
2
< 25
N
N
AR1
AR2-3
12
3
Qtz.
Qtz.
Y
N
0.2
1.55
Max.
Ort.
11.43
20.11
6.83
15.9
2.85
4.31
Y
N
AR2-3
4
Qtz.
N
0.1
Ort.
10.87
3.57
1.37
Y
AR2-3
4
Qtz.
N
0.2
Ort.
13.26
4.8
2.33
N
AR2-3
4
Qtz.
N
0.85
Ort.
13.44
12.74
6.45
AR2-3
5
BIF
Y
0.15
Max.
10.62
5.7
2.36
Y
2
AR2-3
6
BIF
Y
0.15
Ort.
11.71
6.32
1.82
N
2
5.86
AR2-3
6
Qtz.
Y
0.05
Ort.
10.23
2.95
1.15
1
AR2-3
AR2-3
6
6
Qtz.
Qtz.
N
Y
0.3
0.3
Ort.
Ort.
17.22
13.05
7.8
4.01
2.39
5.95
Y
AR2-3
6
Qtz.
N
0.35
Ort.
13.81
8.29
2.94
N
AR2-3
6
Qtz.
N
0.4
Ort.
11.85
7.72
4.39
Y
AR2-3
AR2-3
6
6
Qtz.
Qtz.
N
N
0.7
0.9
Ort.
Ort.
12.89
11.03
8.56
11.87
5.04
5.16
AR2-3
7
Qtz.
N
0.15
Max.
10.62
7.87
2.82
1
AR2-3
7
Qtz.
N
0.35
Ort.
13.34
5.33
3.12
3
4.45
1
< 25
AR2-3
AR2-3
7
7
Qtz.
Qtz.
N
N
1.3
1.55
Ort.
Ort.
20.69
17.02
10.72
10.27
5.67
5.41
N
N
3
1
6.90
17.02
2
1
< 25
Y
Y
N
N
AR2-3
7
Qtz.
N
1.75
Ort.
18.04
11.65
7.58
N
6
3.01
4
< 25
Y
N
N
Bk type
RM
Qtz.
XU
2
Pit/Sc.
AR1
Trv.
3
6.70
2
< 25
Y
N
Y
Y
N
2
6.63
3
< 25
N
N
2
6.72
2
< 25
Lng.
Trv.
Lng.
N
< 25
N
N
1
< 25
N
N
10.23
2
< 25
N
N
2
3
8.61
4.35
1
1
< 25
< 25
N
Y
N
3
4.60
2
N
N
2
5.93
2
< 25
N
N
3
7
4.30
1.58
2
3
< 25
< 25
Y
N
N
N
Y
N
438
T (mm)
5.88
AR2-3
11
Chal.
Y
0.15
Ort.
11.36
5.84
0.85
Y
AR2-3
AR2-3
11
12
Qtz.
Qtz.
N
Y
0.15
0.15
Max.
Max.
11.07
11
5.22
9
2.62
1.42
Y
Y
AR2-3
12
Qtz.
N
0.15
Ort.
11.91
4.46
1.83
AR2-3
12
Qtz.
N
0.6
Ort.
13.74
8.4
3.57
AR2-3
AR2-3
12
12
Qtz.
Qtz.
N
N
1.05
2.5
Ort.
Ort.
14.71
19.76
11.75
15.88
4.47
6.68
AR2-3
13
Qtz.
Y
0.15
Max.
10.83
6.92
2.41
Y
AR2-3
13
Qtz.
N
0.6
Ort.
19.73
6.71
5.31
N
AR2-3
13
Qtz.
N
1.5
Ort.
16.16
16.02
4.76
AR2-3
13
Qtz.
N
1.65
Ort.
19.55
12.72
5.53
AR2-3
13
Qtz.
N
1.9
Ort.
12.03
12.37
AR2-3
14
BIF
Y
0.45
Ort.
15
AR2-3
AR2-3
14
14
Qtz.
Qtz.
Y
N
0.05
0.15
Ort.
Max.
AR2-3
14
Qtz.
N
0.3
AR2-3
14
Qtz.
N
AR2-3
AR2-3
14
15
Qtz.
Qtz.
AR2-3
15
AR2-3
Lng.
ED
W (mm)
11.02
Bipolar
L (mm)
13.02
%Unmod
Dims.
Ort.
#Plat
Wt (g)
0.75
mm/scar
HQ
N
#NFS
RM
Qtz.
Bk type
XU
8
Bk
Pit/Sc.
AR2-3
4
3.26
3
< 25
N
N
1
11.36
2
< 25
N
N
< 25
< 25
N
N
N
1
1
1
11.91
N
5
2.75
2
< 25
Y
N
N
N
4
4
3.68
4.94
3
4
< 25
25-50
Y
Y
N
Y
Trv.
2
N
2
< 25
N
N
2
9.87
2
< 25
Y
N
2
8.08
3
N
4
4.89
2
6.63
N
5
2.41
2
12.58
2.92
N
3
5.00
1
11.64
10.9
2.59
9.74
1.76
2
Y
Y
Trv.
4
2
2.91
Ort.
10.39
7.2
2.73
Y
Trv.
2
0.8
Ort.
15.54
8.49
4.78
N
N
Y
3.35
0.15
Ort.
Ort.
18.92
13.24
16.12
6.62
7.83
1.09
N
Y
Qtz.
N
0.2
Ort.
15.3
7.41
2.52
Y
15
Qtz.
N
0.25
Ort.
10.06
6.46
2.06
Y
AR2-3
AR2-3
15
16
Qtz.
BIF
N
Y
1.1
0.25
Ort.
Ort.
14.01
10.01
12
14.74
5.62
1.44
N
AR2-3
16
Qtz.
Y
0.1
Ort.
10.47
4.93
1.68
Y
< 25
N
N
Y
N
< 25
N
N
2
< 25
< 25
N
N
N
N
5.20
2
< 25
N
N
4
3.89
3
< 25
Y
N
Trv.
4
4
4.73
3.31
3
3
< 25
< 25
Y
N
N
Y
Mrg.
4
3.83
2
< 25
N
N
Trv.
3
3.35
3
< 25
N
N
3
3
4.67
3.34
2
2
< 25
< 25
N
N
N
N
2
5.24
1
< 25
N
439
Wt (g)
Dims.
L (mm)
W (mm)
T (mm)
Bk
%Unmod
Bipolar
ED
0.15
Max.
12.73
4.68
2.17
Y
1
< 25
N
N
AR2-3
16
Qtz.
Y
0.15
Max.
12.62
4.26
2.09
Y
2
< 25
N
N
AR2-3
AR2-3
16
16
Qtz.
Qtz.
Y
N
0.25
0.9
Ort.
Ort.
13.09
16.72
5.71
16.09
2.06
2.33
< 25
N
Y
N
N
AR2-3
17
Qtz.
N
0.1
Max.
10.9
7.42
2.54
Y
1
< 25
N
N
AR2-3
17
Qtz.
Y
0.25
Max.
11.38
9.2
3.2
Y
3
< 25
N
N
AR2-3
AR2-3
17
17
Qtz.
Qtz.
N
N
0.3
0.35
Ort.
Ort.
13.1
11.23
6.07
6.26
2.57
3.11
Y
< 25
< 25
N
N
N
AR2-3
17
Qtz.
N
1.45
Ort.
31.19
8.48
5.18
AR2-3
17
Sil.
N
0.3
Ort.
12.64
6.7
2.1
AR2-3
18
BIF
Y
0.2
Ort.
10.79
10.32
AR2-3
18
Qtz.
N
0.1
Max.
13.11
AR2-3
18
Qtz.
N
0.15
Ort.
AR2-3
18
Qtz.
N
0.15
AR2-3
AR2-3
18
18
Qtz.
Qtz.
N
N
AR2-3
18
Qtz.
AR2-3
18
AR2-3
AR2-3
Trv.
Trv.
2
1
6.55
16.72
#Plat
HQ
Y
mm/scar
RM
Qtz.
#NFS
XU
16
Bk type
Pit/Sc.
AR2-3
2
2
3
4
4.37
2.81
2
3
N
1
31.19
2
Y
N
N
2
6.32
1
< 25
N
N
1.47
N
1
10.79
1
< 25
N
N
4.9
2.04
Y
3
< 25
N
N
10.4
6.69
2.18
N
2
5.20
2
< 25
N
N
Ort.
10.32
7.24
1.57
N
3
3.44
3
< 25
N
N
0.2
0.25
Ort.
Max.
13.22
13.06
5.49
6.7
2.23
2.4
Y
3
2
4.41
2
< 25
N
N
N
Y
0.25
Ort.
11.08
7.66
2.3
N
4
2.77
3
< 25
N
N
Qtz.
N
0.25
Ort.
14.18
5.22
2.69
N
0
1
> 75
N
N
18
18
Qtz.
Qtz.
N
N
0.3
0.4
Ort.
Ort.
13.79
12.46
5.97
10.38
2.68
2.46
N
N
3
1
4.60
12.46
2
1
< 25
25-50
Y
Y
N
N
AR2-3
18
Qtz.
N
0.5
Ort.
14.8
7.7
4.65
N
2
7.40
2
< 25
N
N
AR2-3
18
Qtz.
Y
0.55
Max.
15.24
12.23
3.57
Y
2
AR2-3
AR2-3
18
18
Qtz.
Qtz.
N
Y
0.65
0.65
Ort.
Ort.
11.5
16.24
10.73
12.11
3.56
3.05
Y
3
2
3.83
8.12
AR2-3
18
Qtz.
N
0.7
Ort.
17.99
6.49
5.11
N
3
6.00
Trv.
Trv.
< 25
Y
2
3
< 25
N
N
N
2
< 25
Y
N
440
mm/scar
#Plat
%Unmod
4.44
3.92
N
N
2
5
7.11
4.96
1
2
< 25
< 25
N
Y
N
N
10.7
3.87
N
3
4.68
2
< 25
N
N
10.27
4.91
1.87
Y
3
< 25
N
Y
Ort.
Ort.
10.1
10.6
5.34
10.73
0.85
2.8
Y
N
Lng.
1
2
10.10
5.30
2
2
< 25
< 25
N
N
N
Y
0.4
Ort.
20.18
6.56
3.06
Y
Trv.
4
5.05
2
< 25
N
Y
0.55
Ort.
10.14
10.04
2.94
Y
Trv.
3
3.38
2
< 25
N
N
0.7
Ort.
15.9
10.23
2.95
Y
Trv.
2
7.95
1
< 25
N
1.2
Ort.
21.86
14.17
3.1
N
3
7.29
2
< 25
Y
N
Qtz.
N
0.1
Ort.
12
5.35
1.25
N
2
6.00
1
< 25
N
Y
20
Qtz.
N
0.15
Max.
11.81
4.91
3.06
Y
3
AR2-3
AR2-3
20
20
Qtz.
Qtz.
N
Y
0.15
0.15
Ort.
Ort.
14.31
10.24
4.32
3.99
1.3
2.04
N
Y
AR2-3
20
Qtz.
N
0.15
Max.
11.64
8.08
1.48
Y
1
AR2-3
20
Qtz.
N
0.2
Ort.
11.61
7.01
1.93
N
0
AR2-3
AR2-3
20
20
Qtz.
Qtz.
Y
N
0.2
0.2
Ort.
Ort.
13.78
13.02
6.23
6.18
1.73
1.91
N
Y
2
3
AR2-3
20
Qtz.
N
0.25
Ort.
13.13
7.8
2.47
N
AR2-3
20
Qtz.
Y
0.35
Ort.
12.83
7.14
3.28
AR2-3
AR2-3
20
20
Qtz.
Qtz.
N
Y
0.4
0.45
Ort.
Ort.
10.75
12.12
10.42
12.76
AR2-3
20
Qtz.
N
0.7
Ort.
12.63
10.01
Qtz.
N
0.75
Ort.
16.17
11.58
4.31
Y
AR2-3
18
Qtz.
N
1.05
Ort.
15.26
9.52
4.6
AR2-3
AR2-3
18
18
Qtz.
Qtz.
N
Y
1.15
1.55
Ort.
Ort.
14.22
24.82
10.14
12.59
AR2-3
19
Chal.
Y
1.2
Ort.
14.04
AR2-3
19
Qtz.
Y
0.05
Max.
AR2-3
AR2-3
19
19
Qtz.
Qtz.
Y
N
0.1
0.3
AR2-3
19
Qtz.
N
AR2-3
19
Qtz.
Y
AR2-3
19
Qtz.
Y
AR2-3
19
Qtz.
AR2-3
20
AR2-3
N
N
N
< 25
< 25
N
N
N
N
25-50
N
N
1
> 75
N
N
6.89
4.34
1
2
< 25
< 25
N
N
N
N
1
13.13
1
25-50
N
Y
N
4
3.21
2
< 25
N
N
2.8
2.87
N
Y
3
4
3.58
3.03
3
2
25-50
< 25
Y
N
N
N
4.23
N
2
6.32
3
< 25
Y
N
Trv.
Trv.
Trv.
2
3
ED
#NFS
N
18
Bipolar
Bk type
Y
Bk
< 25
T (mm)
2
W (mm)
3.05
L (mm)
5
Dims.
N
Wt (g)
< 25
HQ
2
RM
5.39
XU
3
Pit/Sc.
Mrg.
AR2-3
7.16
3.41
1
2
441
XU
RM
HQ
Wt (g)
Dims.
L (mm)
W (mm)
T (mm)
Bk
#NFS
mm/scar
#Plat
%Unmod
Bipolar
ED
20
Qtz.
N
1.8
Ort.
17.05
13.3
5.45
N
2
8.53
3
< 25
N
N
AR2-3
20
Qtz.
N
1.9
Ort.
20.6
12.13
5.41
N
2
10.30
2
25-50
Y
AR2-3
AR2-3
20
20
Qtz.
Qtz.
N
N
2.85
9.45
Ort.
Ort.
28.47
29.97
12.47
24.31
6.67
8.75
N
N
3
6
9.49
5.00
3
4
25-50
< 25
N
Y
N
N
AR2-3
21
Qtz.
Y
0.05
Ort.
11.37
5.97
1.61
Y
2
5.69
< 25
N
N
AR2-3
21
Qtz.
N
0.15
Ort.
14.19
6.71
1.88
N
2
7.10
2
25-50
N
N
AR2-3
AR2-3
21
21
Qtz.
Qtz.
Y
N
0.2
0.25
Ort.
Ort.
10.88
10.18
7.52
8.88
2.53
1.59
Y
3
2
3.63
5.09
1
3
< 25
< 25
Y
N
N
N
AR2-3
21
Qtz.
N
0.3
Ort.
12.41
5.34
3.82
N
4
3.10
2
< 25
Y
N
AR2-3
21
Qtz.
N
0.5
Ort.
11.16
12.33
2.63
N
2
5.58
1
< 25
N
N
AR2-3
21
Qtz.
N
1.4
Ort.
14.07
14.83
4.22
N
3
4.69
2
< 25
Y
N
AR2-3
21
Qtz.
N
2.35
Ort.
18.5
17.82
4.47
N
4
4.63
2
< 25
Y
N
AR2-3
23
Qtz.
N
0.5
Ort.
11.33
7.81
4.28
N
1
11.33
2
< 25
Y
Y
AR2-3
23
Qtz.
N
0.75
Ort.
19.14
9.36
3.78
N
2
9.57
1
Y
Y
AR2-3
AR2-3
23
24
Qtz.
Qtz.
N
N
8.8
0.25
Ort.
Ort.
30.26
12.63
20.97
8.45
12.1
2.12
N
N
2
3
15.13
4.21
2
> 50-75
< 25
N
N
N
N
AR2-3
24
Qtz.
N
1.65
Ort.
17.2
18.79
4.31
N
3
5.73
2
< 25
N
N
AR2-3
24
Qtz.
N
1.8
Ort.
20.52
11.85
4.34
N
3
6.84
2
< 25
Y
N
AR2-3
AR2-3
25
30
Qtz.
Qtz.
N
Y
0.65
0.2
Ort.
Ort.
18.08
10.12
8.18
10.69
4.4
1.49
N
Y
2
2
9.04
5.06
1
1
< 25
Y
N
N
N
AR2-3
30
Qtz.
N
0.25
Max.
18.52
8.03
1.73
Y
1
N
N
AR2-3
30
Qtz.
N
1.05
Ort.
19.37
11.63
3.37
N
2
9.69
2
25-50
Y
N
AR2-3
AR2-3
31
32
Qtz.
Qtz.
Y
Y
0.1
0.15
Ort.
Ort.
10.76
13.2
4.87
4.01
1.34
1.73
N
N
3
3
3.59
4.40
1
2
< 25
< 25
N
Y
N
N
Bk type
Pit/Sc.
AR2-3
Trv.
Trv.
Trv.
442
Wt (g)
Dims.
L (mm)
W (mm)
T (mm)
Ret.
Bk
Qtz.
Y
0.2
Ort.
8.73
7.43
1.98
Bck.
N
N
Sc1
Sc3
-
Qtz.
Qtz.
N
Y
0.85
0.15
Ort.
Ort.
14.48
10.09
10.95
5.15
3.9
2.15
Bck.
Bck.
N
N
N
Sc3
-
Qtz.
Y
0.5
Ort.
11.39
10.46
3.28
Bck.
N
N
OSM
-
Qtz.
N
1.35
Ort.
21.87
12.45
3.66
Bck.
N
N
AR1
AR1
3
4
Qtz.
Qtz.
N
N
0.85
0.65
Ort.
Ort.
12.91
16.55
9.91
9.3
5.15
3.36
Bck.
Bck.
Y
N
Trv.
N
Y
AR1
5
Qtz.
Y
0.1
Ort.
7.5
4.77
1.37
Bck.
Y
Lng., Trv.
N
AR1
8
Qtz.
N
0.05
Ort.
7.84
4.47
1.68
Bck.
N
N
AR1
AR1
8
10
Qtz.
Qtz.
N
Y
0.15
< 0.05
Ort.
Ort.
8.91
6.54
6.99
4.99
1.81
0.83
Bck.
Bck.
N
Y
Lng., Trv.
N
N
AR1
11
Qtz.
Y
0.15
Ort.
7.22
4.73
2.74
Bck.
Y
Trv.
N
AR2-3
10
Qtz.
Y
0.15
Ort.
9.92
7.68
1.77
Bck.
N
N
AR2-3
AR2-3
11
13
Qtz.
Qtz.
Y
Y
0.3
0.2
Ort.
Ort.
17.38
14.27
5.14
4.6
2.44
2.18
Bck.
Bck.
N
N
N
N
AR2-3
14
Qtz.
N
1.85
Ort.
15.62
12.63
4.81
Uni.
N
N
AR2-3
15
Qtz.
Y
0.15
Ort.
12.49
5.71
2.11
Uni.
Y
AR2-3
AR2-3
16
16
Qtz.
Qtz.
N
Y
0.3
0.45
Ort.
Ort.
13.14
14.72
8.25
7.21
2.72
3.38
Bck.
Bck.
N
N
Y
N
AR2-3
16
Qtz.
N
0.65
Max.
15.78
9.6
4.62
Uni.
AR2-3
17
Chal.
Y
0.5
Ort.
17.81
8.62
2.86
Bck.
N
N
AR2-3
AR2-3
17
18
Qtz.
Chal.
Y
Y
0.3
0.2
Ort.
Ort.
9.28
14.43
6.88
5.09
2.36
1.7
Bck.
Bck.
N
Y
N
N
Trv.
Trv.
ED
HQ
-
Bk type
RM
Sc1
Pit/Sc.
XU
E.6 RETOUCHED FLAKES
N
443
Pit/Sc.
XU
RM
HQ
Wt (g)
Dims.
L (mm)
W (mm)
T (mm)
Ret.
Bk
Bk type
ED
AR2-3
18
Qtz.
Y
0.25
Ort.
11.68
7.44
1.51
Bck.
Y
Mrg.
N
AR2-3
18
Qtz.
Y
0.3
Ort.
8.22
7.68
3.96
Bck.
Y
Trv.
N
AR2-3
AR2-3
18
18
Qtz.
Qtz.
N
N
0.35
0.4
Ort.
Ort.
14.49
17.53
6.18
7.9
2.56
2.35
Bck.
Bck.
Y
Y
Trv.
Trv.
N
N
AR2-3
18
Qtz.
Y
0.4
Ort.
12.95
7.22
3.99
Bck.
Y
Trv.
N
AR2-3
18
Qtz.
Y
0.5
Ort.
15.7
10.04
2.48
Bck.
Y
Trv.
Y
AR2-3
AR2-3
18
18
Qtz.
Qtz.
Y
N
0.5
0.9
Ort.
Ort.
11.02
18.8
10.42
9
3.74
3.23
Bck.
Uni.
N
Y
Trv.
N
N
AR2-3
18
Qtz.
N
2.7
Ort.
18.95
15.43
5.34
Bck.
N
AR2-3
19
Chal.
Y
< 0.05
Ort.
4.71
4.89
0.89
Bck.
Y
Trv.
N
AR2-3
19
Qtz.
N
0.1
Ort.
6.11
5.57
2.33
Bck.
Y
Trv.
N
AR2-3
20
Qtz.
N
0.2
Ort.
12.58
6.25
1.41
Bck.
N
N
AR2-3
20
Qtz.
Y
0.9
Ort.
20.37
9.61
3.47
Bck.
N
N
AR2-3
20
Qtz.
N
1.8
Ort.
22.67
14.5
4.41
Bck.
N
N
AR2-3
AR2-3
21
21
Qtz.
Sil.
Y
N
0.55
4.9
Ort.
Ort.
12.1
43.05
13.29
16.93
2.39
6.52
Uni.
Uni.
N
N
N
AR2-3
23
Qtz.
N
0.35
Ort.
9.61
7.96
2.63
Bck.
N
N
N
444
APPENDIX F
GIBB ROCK STONE ARTEFACT ASSEMBLAGE
This appendix contains the raw artefact data for the Gibb Rock stone artefact
assemblages. Weight (grams) and dimensions (millimetres) have been rounded to
two decimal places. Blank cells denote unusable or missing data. Unusable data
was that deemed not representative of a particular attribute, i.e. mm/scar for a flake
that could not be oriented. Where data were absent, this indicated that a particular
attribute could not be confidently measured on a certain specimen, often due to
poor-quality raw material. Artefact types have been ordered by pit/scatter
(beginning with surface material), XU depth and raw material. See Tables 9.1–9.2
and Appendix D for details on attributes recorded for each artefact type.
Abbreviations are as follows:
o
Bk = breakage
o
Sc. = scatter
o
Comb wt. = combined weight
o
Sc1 = Scatter 1
o
Dims = dimensions
o
Sc2 = Scatter 2
o
ED = edge damage
o
Sil. = silcrete
o
HQ = high-quality raw material
o
T = thickness
o
L = length
o
Trans. = transitional
o
Max. = maximum dimensions
o
%Unmod = proportion of
o
#NFS = no. of negative flake scars
o
Ort. = oriented dimensions
o
W = width
o
#Plat = no. of platforms
o
Wt = weight
o
Qtz = quartz
o
XU = excavation unit
o
RM = raw material
unmodified surface
445
Pit/Sc.
XU
RM
HQ
Wt (g)
L (mm)
W (mm)
T (mm)
#NFS
g/scar
#Plat
%Unmod
Bipolar
F.1 CORES
Sc1
Sc1
-
Qtz
Qtz
N
N
62.1
143
63.83
86.29
36.02
39.44
19.61
38.88
4
2
15.53
71.50
2
2
> 50-75
> 75
N
Sc2
-
Qtz
N
4.9
20.17
17.74
10.58
9
0.54
4
< 25
Y - trans
Sc2
-
Qtz
N
11.35
28.69
23.37
18.14
4
2.84
Sc2
Sc2
-
Qtz
Sil.
N
N
11.6
232
28.7
69.59
23.15
54.69
16.19
59
11
1
1.05
232.00
4
1
< 25
> 75
N
N
F.2 SMALL FLAKED FRAGMENTS
XU
RM
No.
Comb.
Wt (g)
No. HQ
Sc1
GR1
1
Qtz
Qtz
1
4
0.2
0.35
0
0
GR1
2
Qtz
3
0.15
0
GR1
3
Qtz
4
0.4
0
GR1
GR1
4
5
Qtz
Qtz
1
1
0.1
0.25
0
0
Pit/Sc.
446
RM
HQ
Wt (g)
L (mm)
W (mm)
T (mm)
#NFS
g/scar
#Plat
Sc1
-
Qtz
N
0.3
12.58
5.72
3.78
3
0.10
2
Sc1
-
Qtz
N
0.4
10.18
8.09
5.92
1
0.40
1
Sc1
Sc1
-
Qtz
Qtz
N
N
0.5
0.55
14.13
14.56
5.84
9.53
5.42
4.59
2
0.28
2
Sc1
-
Qtz
N
0.7
13.11
9.71
6.8
2
0.35
2
Sc1
-
Qtz
N
0.7
13.31
9.04
7.24
Sc1
Sc1
-
Qtz
Qtz
N
N
0.7
1
15.8
16.81
7.94
11.85
5.44
5.49
1
2
0.70
0.50
1
1
Sc1
-
Qtz
N
1.25
15.59
10.96
7.15
2
0.63
2
Sc1
-
Qtz
N
1.3
22.83
8.85
5.27
Sc1
-
Qtz
N
1.35
20.27
7.8
6.38
Sc1
-
Qtz
N
1.45
16.48
13.06
5.59
1
1.45
1
Sc1
-
Qtz
N
1.5
18.16
12.42
8.19
Sc1
-
Qtz
N
1.7
18.28
9.66
7.05
1
1.70
1
Sc1
Sc1
-
Qtz
Qtz
N
N
2.85
2.9
18.23
22.43
11.55
17.32
8.55
10.04
2
1
1.43
2.90
1
1
Sc1
-
Qtz
N
4.45
25.43
13.1
9.3
2
2.23
2
Sc1
-
Qtz
N
10.9
39.28
17.31
14.21
1
10.90
1
Sc1
Sc1
-
Qtz
Qtz
N
N
15.2
15.65
37.94
39.77
29.41
22.65
13.09
17.04
7
2.24
3
Sc1
-
Qtz
N
45.75
45.71
34.32
24.13
Pit/Sc.
XU
F.3 LARGE FLAKED FRAGMENTS
F.4 SMALL FLAKES
XU
RM
No.
Comb.
Wt (g)
No. HQ
Sc1
-
Qtz
3
1.55
0
Sc2
GR1
1
Qtz
Qtz
1
8
0.25
0.6
0
0
0
GR1
2
Qtz
1
< 0.05
0
1
GR1
3
Qtz
2
0.15
0
GR1
4
Qtz
1
0.05
0
Pit/Sc.
No.
broken
0
447
#Plat
%Unmod
Bipolar
ED
#NFS
-
Qtz
N
0.2
Ort.
12.43
6.46
1.96
N
2
Sc1
Sc1
-
Qtz
Qtz
N
N
0.2
0.35
Ort.
Ort.
12.24
13.34
8.84
5.72
2.52
2.98
Y
N
1
2
12.24
6.67
Sc1
-
Qtz
N
0.45
Ort.
10.04
7
4.23
N
0
Sc1
-
Qtz
N
0.45
Max.
10.87
9.37
4.87
1
Sc1
Sc1
-
Qtz
Qtz
N
N
0.45
0.45
Ort.
Ort.
10.95
12.13
14.08
5.89
2.68
3.16
N
2
2
5.48
6.07
3
1
Sc1
-
Qtz
N
0.5
Ort.
14.82
7.68
3.55
N
1
14.82
1
Sc1
-
Qtz
N
0.5
Ort.
13.08
9
3.1
N
2
6.54
2
< 25
Sc1
Sc1
-
Qtz
Qtz
N
N
0.55
0.55
Ort.
Ort.
11.91
13
8.19
8.97
4.12
4.53
N
1
11.91
2
> 50-75
Sc1
-
Qtz
N
0.55
Ort.
12.57
8.2
4.16
N
2
6.29
1
< 25
N
N
Sc1
-
Qtz
N
1
Ort.
12.19
10.57
4.21
N
3
4.06
1
< 25
Y
N
Sc1
Sc1
-
Qtz
Qtz
N
N
1.1
1.15
Ort.
Ort.
18.17
16.99
10.7
9.64
3.08
3.99
N
N
3
3
6.06
5.66
2
2
< 25
< 25
Y
N
N
Sc1
-
Qtz
N
1.4
Ort.
16.49
13.62
3.79
N
1
16.49
1
25-50
N
N
Sc1
-
Qtz
N
1.65
Ort.
18.23
10.72
6.96
N
3
6.08
2
< 25
Sc1
Sc1
-
Qtz
Qtz
N
N
1.7
1.7
Ort.
Ort.
23.28
17.24
10.26
9.78
4.77
6.43
N
N
3
3
7.76
5.75
2
1
< 25
< 25
Y
Y
N
N
Sc1
-
Qtz
N
1.8
Ort.
13.63
18.73
4.58
N
3
4.54
3
< 25
Y
N
Sc1
-
Qtz
N
2.55
Ort.
16.22
16.13
7.1
N
5
3.24
3
< 25
Y
N
Sc1
Sc1
-
Qtz
Qtz
N
N
2.6
3
Ort.
Ort.
20.91
18.68
16.3
19.57
5.57
5.3
N
N
4
2
5.23
9.34
2
1
< 25
< 25
N
Y
N
N
Bk type
Sc1
Pit/Sc.
Bk
N
> 75
T (mm)
Y
1
W (mm)
1
< 25
< 25
L (mm)
N
Dims.
Y
Wt (g)
< 25
HQ
2
RM
6.22
XU
mm/scar
F.5 LARGE FLAKES
N
N
< 25
< 25
N
Y
N
N
N
N
Y
N
N
N
N
448
RM
HQ
Wt (g)
Dims.
L (mm)
W (mm)
T (mm)
Bk
#NFS
mm/scar
#Plat
%Unmod
Bipolar
ED
Qtz
N
3.85
Ort.
25.77
17.32
7.42
N
2
12.89
1
25-50
Y
N
Sc1
Sc1
-
Qtz
Qtz
N
N
3.9
4.35
Ort.
Max.
22.84
30.37
18.7
22.44
6.77
6.66
N
N
1
1
22.84
1
25-50
> 50-75
Y
N
N
Sc2
-
Qtz
N
0.95
Ort.
11.16
13
5.06
N
3
3.72
2
< 25
N
N
Sc2
-
Qtz
N
3.6
Ort.
27.08
17.93
5.44
N
0
1
> 75
Y
N
GR1
GR1
1
1
Qtz
Qtz
N
N
0.2
0.35
Ort.
Ort.
10.48
11.78
7.63
6.82
2
4.64
N
N
2
3
N
N
N
GR1
2
Qtz
N
0.35
Ort.
11.33
9.62
2.33
N
Bk type
XU
-
Pit/Sc.
Sc1
5.24
3.93
2
2
< 25
N
449
APPENDIX G
MULKA’S CAVE STONE ARTEFACT ASSEMBLAGE
This appendix contains the raw artefact data for the Mulka’s Cave stone artefact
assemblages. Weight (grams) and dimensions (millimetres) have been rounded to
two decimal places. Blank cells denote unusable or missing data. Unusable data
was that deemed not representative of a particular attribute, i.e. mm/scar for a flake
that could not be oriented. Where data were absent, this indicated that a particular
attribute could not be confidently measured on a certain specimen, often due to
poor-quality raw material. Artefact types have been ordered by site area, pit/scatter
(beginning with surface material), XU depth and raw material. See Tables 9.1–9.2
and Appendix D for details on attributes recorded for each artefact type.
Abbreviations are as follows:
o
Bck. = backed
o
Ort. = oriented dimensions
o
BIF = Banded Iron Formation
o
Oth. = other surface artefacts
o
Bk = breakage
o
#Plat = no. of platforms
o
CAS = Camping Area Scatter
o
Qtz = quartz
o
Chal. = chalcedony
o
Ret. = retouch type
o
Cl. = cluster
o
RM = raw material
o
Comb wt. = combined weight
o
Sc. = scatter
o
CS = Column Sample
o
Sil. = silcrete
o
Dims = dimensions
o
T = thickness
o
ED = edge damage
o
Trans. = transitional
o
EP = Entrance Platforms
o
Trv. = transverse
o
HQ = high-quality raw material
o
Uni = unifacial
o
HS = Humps Scatter
o
%Unmod = proportion of
o
L = length
o
Max. = maximum dimensions
o
W = width
o
Mrg. = marginal
o
Wt = weight
o
#NFS = no. of negative flake scars
o
XU = excavation unit
unmodified surface
450
Pit/Sc.
XU/Cl.
RM
HQ
Wt (g)
L (mm)
W (mm)
T (mm)
#NFS
g/scar
#Plat
%Unmod
Bipolar
G.1 CORES
EP
CS
MC1
HS
HS
8
4
-
Qtz
Qtz
Qtz
Qtz
Qtz
N
Y
Y
N
Y
1.75
3.25
0.65
2.2
2.55
16.28
24.11
13.29
18.72
19.89
11.89
13.12
11.65
16.37
17.39
10.51
8.49
4.15
10.39
9.33
4
6
3
6
6
0.44
0.54
0.22
0.37
0.43
3
4
1
3
2
< 25
< 25
25-50
< 25
< 25
Y
Y - trans
Y
Y
Y - trans
HS
CAS
B
Qtz
Qtz
Y
Y
3.05
2.35
18.36
16.58
14.14
16.54
10.16
11.57
7
10
0.44
0.24
3
4
< 25
< 25
Y - trans
Y
CAS
CAS
CAS
CAS
CAS
CAS
CAS
CAS
B
B
B
B
D
D
F
Oth.
Qtz
Qtz
Sil.
Sil.
Qtz
Qtz
Qtz
Qtz
N
N
N
N
N
Y
N
N
3.05
6.05
14.15
60.2
0.85
2.65
16.55
14.3
19.09
21.96
35.43
59.53
10.93
17.69
33.78
31.02
18.47
19.77
30.18
45.51
11.53
14.76
21.92
27.55
10.28
16.37
15.59
27.63
8.58
10.33
16.65
19.17
11
9
10
4
6
10
3
11
0.28
0.67
1.42
15.05
0.14
0.27
5.52
1.30
4
3
4
1
2
6
1
5
< 25
< 25
< 25
> 50-75
< 25
< 25
> 75
< 25
Y - trans
N
N
N
Y
Y
Y
N
CA1
CA1
CA1
CA1
CA1
CA2
3
3
4
4
4
4
Qtz
Qtz
Qtz
Qtz
Qtz
Qtz
N
N
N
N
N
N
2.5
16.15
0.95
4.3
82.15
23.15
18.23
43.29
14.81
25.38
58.5
34.74
16.28
21.41
9.37
15.44
52.4
27.93
8.27
19.77
9.14
13.86
40.23
20.68
6
12
4
5
9
16
0.42
1.35
0.24
0.86
9.13
1.45
2
5
4
3
4
6
< 25
< 25
< 25
25-50
25-50
< 25
Y
N
Y
Y - trans
N
N
451
G.2 SMALL FLAKED FRAGMENTS
XU/Cl.
5
9
RM
Qtz
Qtz
No.
1
2
Comb.
Wt (g)
< 0.05
0.15
No. HQ
1
2
11
-
Qtz
Qtz
2
1
0.15
0.1
2
1
TH1
TH1
TH1
TH1
TH1
TH1
CAS
CAS
CA1
CA1
CA1
CA1
CA1
CA1
CA1
1
2
3
4
5
6
C
Oth.
1
2
3
5
6
7
8
Qtz
Qtz
Qtz
Qtz
Qtz
Qtz
Qtz
Qtz
Qtz
Qtz
Qtz
Qtz
Qtz
Qtz
Qtz
8
15
10
13
6
3
2
1
4
5
10
7
8
11
4
0.6
1.25
1.25
1.4
0.4
0.3
0.2
0.35
0.45
0.85
1.35
0.7
0.95
1.5
0.4
4
7
0
0
0
0
2
0
0
1
3
0
0
3
2
CA1
CA2
CA2
CA2
CA2
CA2
CA2
9
1
2
3
4
5
6
Qtz
Qtz
Qtz
Qtz
Qtz
Qtz
Qtz
1
5
6
7
13
12
3
0.2
0.6
0.6
0.9
2.35
2
0.45
0
0
0
0
3
2
2
Pit/Sc.
CS
CS
CS
HS
452
XU/Cl.
RM
HQ
Wt (g)
L (mm)
W (mm)
T (mm)
#NFS
g/scar
EP
EP
CS
CS
CS
CS
CS
MC1
5
5
6
7
9
3
Qtz
Qtz
Qtz
Qtz
Qtz
Qtz
Qtz
Qtz
N
N
N
Y
N
N
Y
N
0.9
4.25
0.4
1.8
0.25
1.25
0.1
2.25
15.27
25.32
11.81
16.63
11.9
17.73
10.38
18.56
9.31
18.14
6.93
15.79
5.05
9.82
5.73
13.89
7.43
10.56
5.08
7.88
4.22
6.01
2.75
7.67
1
1
3
6
2
1
2
3
0.90
4.25
0.13
0.30
0.13
1.25
0.05
0.75
1
1
2
3
MC1
HS
4
-
Qtz
Qtz
N
Y
8
2
32.7
21.26
20.11
12.93
14
7.16
1
7
8.00
0.29
1
3
HS
TH1
TH1
TH1
TH1
TH1
TH1
TH1
TH1
TH1
TH1
TH1
TH1
TH1
TH1
TH1
TH1
TH1
TH1
TH1
CAS
CAS
CAS
CAS
CAS
CAS
CAS
CAS
CAS
CAS
CAS
1
1
1
2
2
2
3
3
3
3
4
4
4
4
4
4
5
5
5
A
A
A
B
B
B
C
C
C
D
D
Qtz
Qtz
Qtz
Qtz
Qtz
Qtz
Qtz
Qtz
Qtz
Qtz
Qtz
Qtz
Qtz
Qtz
Qtz
Qtz
Qtz
Qtz
Qtz
Qtz
Qtz
Qtz
Qtz
Qtz
Qtz
Qtz
Qtz
Qtz
Qtz
Qtz
Qtz
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
Y
N
N
N
N
N
N
N
2.15
0.25
0.45
0.55
0.25
0.3
0.5
0.3
0.6
0.65
0.7
0.15
0.25
0.3
0.3
0.5
0.55
0.15
0.2
0.3
0.6
0.9
1.7
1.55
4.25
6.6
0.8
2.65
3.95
1.55
7.05
19.23
10.76
13.03
15.63
11.18
10.5
15.71
11.65
13.6
12.33
14.55
12.62
12.7
12.02
11.14
13.49
12.98
11.39
10.7
8.78
13.91
16.8
19.45
22.58
24.58
29.19
17.37
21.39
24.77
17.32
27.57
17.37
5.47
9.42
7.26
5.86
6.31
6.24
6.07
9.05
8.25
9.19
4.45
5.04
6.54
6.63
6.88
6.83
5.05
6.89
6.36
6.67
8.95
10.71
10.49
18.87
20.57
7.19
14.83
18.74
16.68
19.93
6.58
4.01
3.68
5.39
4.2
3.5
4.18
3.93
6.49
6.94
4.59
3.06
4.6
4.16
3.49
3.67
5.01
2.86
3.24
3.67
6.2
5.9
9.84
7.85
12.01
13.19
6.41
9.64
11.3
6.62
13.54
5
2
2
3
1
2
1
1
5
1
2
1
2
2
1
1
0.43
0.13
0.23
0.18
0.25
0.15
0.50
0.30
0.12
0.65
0.35
0.15
0.13
0.15
0.30
0.50
2
2
1
2
1
2
1
1
3
1
2
1
2
2
1
1
1
1
1
3
3
2
4
2
9
3
2
7
4
5
0.15
0.20
0.30
0.20
0.30
0.85
0.39
2.13
0.73
0.27
1.33
0.56
0.39
1.41
1
1
1
1
2
1
2
1
4
1
2
2
2
2
#Plat
Pit/Sc.
G.3 LARGE FLAKED FRAGMENTS
1
1
3
453
RM
HQ
Wt (g)
L (mm)
W (mm)
T (mm)
#NFS
g/scar
Qtz
Qtz
Qtz
Qtz
Qtz
Qtz
Qtz
Qtz
N
N
Y
N
N
N
N
N
4.75
6.75
0.15
0.4
0.6
0.65
0.8
0.85
22.04
27.2
10.99
10.26
12.08
12.12
16.71
14.84
18.02
26.45
5.92
9.34
6.98
11.72
7.71
9.28
11.81
11.26
2.71
4.96
5
6.21
6.29
5.56
6
1
2
2
2
3
4
4
0.79
6.75
0.08
0.20
0.30
0.22
0.20
0.21
3
1
2
1
CAS
CAS
Oth.
Oth.
Qtz
Qtz
N
Y
1.55
2
17.77
20.34
9.67
12.53
9.31
7.95
3
3
0.52
0.67
1
CAS
CAS
CAS
CAS
Oth.
Oth.
Oth.
Oth.
Qtz
Qtz
Qtz
Qtz
N
N
N
N
2.15
2.45
2.85
3.15
23.03
16.36
20.25
22.62
12.29
15.02
12.37
16.61
6.47
11.9
11.34
9.94
3
3
4
3
0.72
0.82
0.71
1.05
1
2
2
1
CAS
CA1
CA1
CA1
CA1
CA1
CA1
CA1
CA1
CA1
CA1
CA1
CA1
Oth.
2
2
2
4
4
4
5
5
5
5
5
5
Qtz
Qtz
Qtz
Qtz
Qtz
Qtz
Qtz
Qtz
Qtz
Qtz
Qtz
Qtz
Qtz
N
N
N
N
N
N
N
N
N
N
N
N
N
4.05
0.8
1.1
1.6
0.1
0.15
1.75
0.15
0.3
0.65
1.15
1.6
2
32.5
14.4
21.81
18.48
10.15
11.6
21.28
11.83
10.84
18.21
17.97
17.65
20.01
13.66
7.31
7.2
13.03
5.48
6.26
8.71
3.77
6.21
7.68
14.56
12.82
14.61
11.02
7.23
7.07
7.94
2.32
5.31
7.48
3.24
4.16
4.77
4.89
8.67
9.27
1
2
4.05
0.40
1
2
3
0.53
2
3
1
0.05
1.75
3
1
2
0.15
2
1
3
1
1.15
0.53
2.00
1
2
1
CA1
CA1
CA1
CA1
CA1
CA1
CA1
CA1
CA1
CA1
CA1
CA1
CA2
CA2
CA2
6
7
7
7
7
7
7
7
7
8
8
8
1
1
2
Qtz
Qtz
Qtz
Qtz
Qtz
Qtz
Qtz
Qtz
Qtz
Qtz
Qtz
Qtz
Qtz
Qtz
Qtz
N
N
N
N
N
N
N
N
N
Y
N
N
Y
Y
Y
0.9
0.3
0.45
0.45
0.55
0.55
0.6
0.85
0.9
0.25
0.35
1.65
0.25
0.5
0.2
11.39
13.69
10.69
10.81
12.02
12.63
13.93
12.72
16.99
10.31
11.42
15.29
12.91
12.69
13.39
10.94
6.35
7.37
7.83
6.32
6.25
7.11
10.02
6.77
5.27
5.42
11.09
4.01
7.78
4.36
7.87
3.67
5.97
5.3
6.08
6.07
6.16
5.81
6.3
3.97
4.55
8.89
3.41
5.03
4.19
1
1
2
1
4
0.90
0.30
0.23
0.45
0.14
1
1
3
2
0.20
0.43
5
0.05
3
2
1
1
0.55
0.13
0.50
0.20
#Plat
XU/Cl.
E
F
Oth.
Oth.
Oth.
Oth.
Oth.
Oth.
Pit/Sc.
CAS
CAS
CAS
CAS
CAS
CAS
CAS
CAS
2
1
3
2
1
1
454
Pit/Sc.
XU/Cl.
RM
HQ
Wt (g)
L (mm)
W (mm)
T (mm)
#NFS
g/scar
#Plat
CA2
CA2
CA2
CA2
CA2
CA2
CA2
CA2
2
3
4
4
4
5
5
5
Qtz
Qtz
Qtz
Qtz
Qtz
Qtz
Qtz
Qtz
N
N
N
N
Y
N
N
N
1.1
0.15
0.2
0.35
1.5
0.4
0.5
0.6
18.24
11.14
10.34
10.66
20.62
12.58
11.65
11.45
7.94
5.16
5.73
7.04
10.1
5.47
11.17
9.32
7.66
3.06
3.64
5.55
8.65
4.26
4.67
5.42
3
1
2
3
3
1
2
3
0.37
0.15
0.10
0.12
0.50
0.40
0.25
0.20
2
1
2
3
3
1
2
2
CA2
CA2
6
6
Qtz
Qtz
N
N
0.45
0.55
10.62
13.42
6.55
6.24
5.1
5.61
3
2
0.15
0.28
CA2
CA2
CA2
6
6
6
Qtz
Qtz
Qtz
N
N
N
0.9
2
2.5
13.62
15.43
19.65
9.17
11.61
14.29
6.26
11.07
8.1
4
2
3
0.23
1.00
0.83
1
2
455
G.4 SMALL FLAKES
XU/CL.
RM
No.
Comb.
Wt (g)
No. HQ
No.
broken
CS
CS
CS
CS
CS
CS
CS
CS
CS
2
4
5
6
7
8
9
10
11
Qtz
Qtz
Qtz
Qtz
Qtz
Qtz
Qtz
Qtz
Qtz
2
3
2
6
6
4
5
7
1
0.05
0.15
0.15
0.45
0.5
0.3
0.5
0.3
0.3
1
0
0
2
6
4
5
7
1
0
0
0
0
4
3
3
1
1
MC1
MC1
5
6
Qtz
Qtz
1
1
0.4
0.35
0
0
1
1
HS
TH1
TH1
TH1
TH1
TH1
TH1
CAS
CAS
CAS
CAS
CAS
CAS
CA1
CA1
1
2
3
4
5
6
A
B
C
D
F
Oth.
1
2
Qtz
Qtz
Qtz
Qtz
Qtz
Qtz
Qtz
Qtz
Qtz
Qtz
Qtz
Qtz
Qtz
Qtz
Chal.
4
14
14
6
9
6
3
4
3
2
2
1
4
3
1
1.65
1.2
2.25
0.45
1.2
0.55
0.2
1.2
1.1
0.3
0.5
0.05
0.85
0.3
0.1
3
0
0
4
3
0
1
1
3
2
1
1
1
0
1
0
CA1
CA1
2
2
Qtz
Sil.
15
3
1.3
0.1
8
0
3
1
CA1
CA1
CA1
CA1
CA1
CA1
CA1
CA1
CA1
CA1
CA1
CA1
CA2
CA2
3
3
3
4
5
5
5
6
6
7
8
9
1
1
BIF
Qtz
Sil.
Qtz
BIF
Qtz
Sil.
Chal.
Qtz
Qtz
Qtz
Qtz
BIF
Qtz
1
31
1
29
2
24
3
1
13
11
3
5
1
6
0.1
4.25
0.15
2.55
0.15
2.5
0.25
0.1
1.1
1.9
0.35
0.45
0.1
0.95
1
16
0
20
2
5
0
1
13
7
2
0
1
4
1
6
0
6
0
3
2
0
5
1
0
0
0
1
Pit/Sc.
3
0
2
0
0
1
0
1
456
Pit/Sc.
CA2
CA2
CA2
CA2
CA2
CA2
CA2
XU/CL.
2
2
3
3
4
5
6
RM
BIF
Qtz
Qtz
Sil.
Qtz
Qtz
Qtz
No.
1
17
28
1
25
11
1
Comb.
Wt (g)
0.05
1.05
3.45
0.15
4.2
1.45
0.05
No. HQ
1
14
18
0
5
3
1
No.
broken
1
6
9
1
3
2
0
457
5.88
5.18
7
9.84
8.59
5.32
1.29
0.93
2.19
3.8
4.5
3.53
3.81
N
N
N
N
N
N
N
N
CS
CS
CS
CS
CS
CS
CS
MC1
MC1
MC1
MC1
9
9
9
9
9
9
9
3
3
4
4
Qtz
Qtz
Qtz
Qtz
Qtz
Qtz
Qtz
Qtz
Qtz
Qtz
Qtz
N
N
Y
Y
N
Y
N
Y
N
Y
N
0.2
0.2
0.25
0.25
0.3
0.3
1
0.65
1.45
0.75
0.75
Ort.
Ort.
Ort.
Max.
Ort.
Ort.
Ort.
Ort.
Ort.
Ort.
Ort.
12.8
10.45
10.28
12.08
10.01
12.71
14.77
14.65
15.37
16.04
15.33
4.71
5.26
9.28
7.61
6.68
7.72
9.63
10.13
16.9
10.19
8.01
3.3
2.73
1.89
3.81
3.64
2.04
6.19
3.39
3.02
3.24
3.78
N
N
Y
N
N
Y
N
N
N
N
N
N
3
3
4
5
2
4
2
3
2
4
1
2
2
5.72
5.31
5.03
3.15
7.86
3.80
5.06
3.93
3
2
2
2
2
3
N
N
N
Y
Y
N
N
N
N
3
< 25
< 25
< 25
< 25
< 25
< 25
< 25
< 25
< 25
< 25
4.65
6.03
1
< 25
25-50
N
N
2
2
2
1
0
2
4
3
6.40
5.23
5.14
1
2
1
N
6.36
3.69
4.88
1
1
2
3
< 25
25-50
< 25
< 25
> 75
< 25
< 25
< 25
3
2
5.35
7.67
1
3
< 25
< 25
2
N
N
N
N
N
N
N
ED
12.21
14.4
15.41
12.82
14.41
7.35
6.67
4.06
7.31
9.23
11.94
5.9
9.54
Bipolar
T (mm)
17.16
15.94
20.11
15.77
15.71
15.21
10.12
11.78
12.81
18.6
18.38
11.2
12.05
%Unmod
W (mm)
Ort.
Ort.
Ort.
Ort.
Ort.
Ort.
Ort.
Ort.
Max.
Ort.
Max.
Max.
Ort.
#Plat
L (mm)
1.5
1.55
2.45
2.6
2.65
0.6
0.1
0.05
0.2
0.85
0.95
0.1
0.5
mm/scar
Dims.
N
N
N
N
N
Y
N
Y
N
N
N
N
Y
#NFS
Wt (g)
Qtz
Qtz
Qtz
Qtz
Qtz
Qtz
Qtz
Qtz
Qtz
Qtz
Qtz
Qtz
Qtz
Bk type
HQ
5
6
7
7
7
7
8
8
Bk
RM
EP
EP
EP
EP
EP
CS
CS
CS
CS
CS
CS
CS
CS
Pit/Sc.
XU/Cl.
G.5 LARGE FLAKES
Y
Y
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
Y
N
Y
N
N
N
458
W (mm)
T (mm)
Bk
21.66
12.01
12.74
16.91
6.76
15.22
9.49
9.5
4.88
3.94
4.58
5.69
3.28
5.65
2.06
3.3
N
N
N
N
Y
N
Y
Y
HS
HS
-
Qtz
Qtz
Y
Y
0.85
1.25
Ort.
Ort.
17.04
15.65
8.66
11.85
2.98
4.91
N
Y
HS
HS
HS
HS
HS
TH1
TH1
TH1
TH1
TH1
TH1
TH1
1
1
1
1
2
2
2
Qtz
Qtz
Qtz
Qtz
Qtz
Qtz
Qtz
Qtz
Qtz
Qtz
Qtz
Qtz
Y
Y
N
N
N
Y
N
N
N
N
N
N
1.3
1.45
2
2.9
2.9
0.2
0.55
0.6
3.7
0.15
0.45
2.05
Max.
Ort.
Ort.
Ort.
Ort.
Ort.
Ort.
Ort.
Ort.
Ort.
Ort.
Ort.
20.56
14.59
28.5
22.47
16.33
11.04
10.29
10.19
22.95
10
13.58
23.53
12.35
10.05
11.88
15.47
11.24
6.44
13.4
12.26
15.03
5.18
5.6
15.5
4.74
5.71
5.32
5.8
8.98
2.13
4.24
4.22
10.84
2
4.06
5.11
N
N
N
N
Y
Y
N
N
N
N
N
TH1
TH1
TH1
TH1
3
3
4
5
Qtz
Qtz
Qtz
Qtz
N
N
N
N
0.15
0.25
1.1
0.05
Ort.
Ort.
Ort.
Ort.
10.81
11.99
19.88
10.11
6.43
6.22
12.22
3.62
2.07
2.84
4.46
1.13
27.30
5.13
9.55
19.19
2
2
2
1
Trv.
Trv.
1
3
2
1
2
4
4
2
4.83
3.57
6.49
Trv.
2
2
Trv.
Trv.
N
Y
Trv.
Y
Trv.
ED
L (mm)
27.3
15.39
19.1
19.19
11.54
19.33
14.28
12.97
Bipolar
Dims.
Ort.
Ort.
Ort.
Ort.
Max.
Ort.
Ort.
Ort.
%Unmod
Wt (g)
4
0.85
1.4
2.45
0.25
2.5
0.4
0.5
#Plat
HQ
N
N
N
N
N
N
Y
Y
mm/scar
RM
Qtz
Qtz
Qtz
Qtz
Qtz
Qtz
Qtz
Qtz
#NFS
XU/Cl.
4
5
5
5
6
6
-
Bk type
Pit/Sc.
MC1
MC1
MC1
MC1
MC1
MC1
HS
HS
2
2
2
< 25
< 25
< 25
< 25
< 25
< 25
< 25
N
N
N
N
N
N
N
N
N
N
8.52
7.83
3
2
< 25
25-50
N
N
N
N
5
2
4
5
5
3
3
2
4
1
2
4
7.30
7.13
4.49
3.27
3.68
3.43
5.10
5.74
10.00
6.79
5.88
2
3
2
2
1
3
< 25
< 25
< 25
< 25
< 25
< 25
< 25
25-50
< 25
Y
N
Y
Y
N
N
Y
3
< 25
N
1
1
3
2
10.81
11.99
6.63
5.06
2
1
2
2
25-50
25-50
< 25
< 25
N
N
N
N
2
1
Y
Y
Y
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
459
#NFS
mm/scar
#Plat
%Unmod
Bipolar
ED
N
N
N
N
5.3
9.2
5.75
2.52
2.53
2.83
4.5
2.8
3.33
3.01
4.29
3.73
N
N
Y
N
N
N
N
N
Y
N
N
7
4
2
3
3
2
3
4
3
4
4
2.00
6.88
6.68
3.94
5.06
6.66
5.51
3.35
4.96
3.69
4.64
3
3
1
3
2
1
2
2
2
3
2
< 25
< 25
< 25
< 25
< 25
< 25
< 25
< 25
< 25
< 25
< 25
Y
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
Y
Y
Y
N
N
4.06
3.63
3.38
4.06
Y
N
N
N
3
2
7
4
5.64
5.24
1.68
4.38
1
1
4
2
< 25
< 25
< 25
< 25
N
N
N
N
N
N
Qtz
Qtz
Qtz
Qtz
Qtz
Qtz
Qtz
Qtz
N
N
N
N
N
N
N
N
0.2
0.25
0.3
0.3
0.4
0.4
0.6
0.7
Ort.
Ort.
Ort.
Ort.
Ort.
Ort.
Ort.
Ort.
11.21
10.67
11.36
11.69
14.38
11.67
12.98
13.16
6.77
8.38
7.9
6.07
8.12
6.92
10.78
12.89
2.58
1.38
2.28
3.55
2.01
3.27
3.72
3.57
CAS
CAS
A
A
Qtz
Qtz
N
N
0.85
1.15
Ort.
Ort.
20.47
11.26
9.25
12.6
4.71
5.02
CAS
CAS
CAS
CAS
CAS
CAS
CAS
CAS
CAS
CAS
CAS
CAS
A
A
A
B
B
B
B
B
B
B
B
B
Qtz
Qtz
Qtz
Qtz
Qtz
Qtz
Qtz
Qtz
Qtz
Qtz
Qtz
Qtz
N
Y
N
Y
N
N
N
N
Y
N
N
Y
1.5
1.85
2.2
0.35
0.5
0.5
0.6
0.6
0.6
0.65
0.7
0.75
Ort.
Ort.
Ort.
Ort.
Ort.
Ort.
Ort.
Ort.
Ort.
Ort.
Ort.
Ort.
16.35
13.99
27.53
13.36
11.82
15.17
13.31
16.54
13.4
14.87
14.76
18.56
13.85
9.91
20.11
8.13
11.85
10.72
7.24
12.77
11.11
11.84
12.86
7.97
CAS
CAS
CAS
CAS
B
B
B
B
Qtz
Qtz
Qtz
Qtz
N
Y
Y
N
0.8
0.8
0.85
0.9
Ort.
Ort.
Ort.
Ort.
16.91
10.47
11.75
17.51
8.54
11.96
16.05
10.8
Bk type
< 25
< 25
5
5
A
A
A
A
A
A
Bk
2
2
T (mm)
6.82
5.63
W (mm)
3
2
L (mm)
N
N
Dims.
N
Y
Y
N
Y
N
N
N
Wt (g)
N
N
N
N
N
N
N
N
HQ
< 25
< 25
< 25
< 25
< 25
< 25
< 25
< 25
RM
2
2
2
2
3
2
1
2
XU/Cl.
3.74
3.56
5.68
2.92
3.60
3.89
4.33
6.58
Pit/Sc.
3
3
2
4
4
3
3
2
TH1
TH1
CAS
CAS
CAS
CAS
CAS
CAS
Y
N
Trv.
Y
N
N
Trv.
Trv.
Trv.
Trv.
460
T (mm)
5.15
3.31
5.16
5.18
7.11
6.01
6.08
6.01
CAS
CAS
B
B
Qtz
Qtz
N
N
3
3.55
Ort.
Ort.
21.44
17.35
15.87
18.13
CAS
CAS
CAS
CAS
CAS
CAS
CAS
CAS
CAS
CAS
CAS
CAS
B
B
B
C
C
C
C
C
C
C
C
C
Qtz
Qtz
Qtz
Qtz
Qtz
Qtz
Qtz
Qtz
Qtz
Qtz
Qtz
Qtz
N
N
N
N
N
Y
N
Y
N
N
N
Y
4.25
4.3
5.15
0.2
0.4
0.65
1.2
1.2
2.3
2.6
3.3
4.15
Ort.
Ort.
Ort.
Ort.
Ort.
Ort.
Ort.
Ort.
Ort.
Ort.
Max.
Ort.
20.24
22.08
20.04
10.22
10.14
12.26
14.72
16.66
17.43
25.43
22.91
23.32
CAS
CAS
CAS
CAS
C
C
C
C
Qtz
Qtz
Sil.
Sil.
N
N
N
N
4.25
4.9
0.4
10.75
Max.
Ort.
Ort.
Ort.
24.09
26.74
11.47
35.6
ED
W (mm)
15.76
9.62
15.21
10.6
12.38
15.18
15.56
17.83
Bipolar
L (mm)
16.21
19.33
10.74
18.27
15.86
11.91
18.69
19.4
%Unmod
Dims.
Max.
Ort.
Ort.
Ort.
Ort.
Ort.
Ort.
Ort.
#Plat
Wt (g)
1
1.05
1.2
1.4
1.55
1.8
2.3
2.35
2
2
3
3
4
1
4
< 25
< 25
< 25
< 25
< 25
< 25
< 25
< 25
N
N
N
N
N
N
N
N
Y
N
N
N
N
N
N
N
5.36
2.48
2
3
< 25
< 25
N
N
Y
Y
5
3
5
2
4
4
4.05
7.36
4.01
5.11
2.54
3.07
3
2
2
1
2
2
< 25
< 25
< 25
< 25
< 25
< 25
Y
N
6
5
2
2.78
3.49
12.72
3
2
1
< 25
< 25
< 25
N
Y
N
N
2
11.66
2
25-50
N
N
N
N
Y
N
N
N
N
N
Y
Y
N
N
N
Y
N
3
7
1
5
3
1
2
< 25
< 25
< 25
< 25
N
3.82
11.47
7.12
mm/scar
HQ
N
Y
N
Y
N
N
N
N
#NFS
RM
Qtz
Qtz
Qtz
Qtz
Qtz
Qtz
Qtz
Qtz
Bk type
XU/Cl.
B
B
B
B
B
B
B
B
Bk
Pit/Sc.
CAS
CAS
CAS
CAS
CAS
CAS
CAS
CAS
N
N
N
N
N
N
N
4
3
3
7
5
4
3
5
6.44
3.58
2.61
3.17
2.98
6.23
3.88
8.87
7.51
N
N
4
7
16.35
17.41
19.33
6.92
6.85
11.29
13.44
12.32
6.81
15.63
13.82
18.08
8.43
8.2
7.91
1.79
3.46
4.64
4.54
4.44
13.32
3.82
7.97
7.37
N
N
N
N
N
N
N
N
Y
19.9
10.49
13.27
18.74
8.45
13.53
3.15
10.81
Trv.
Mrg.
N
N
N
Y
N
N
N
N
461
Dims.
L (mm)
W (mm)
T (mm)
Bk
#NFS
mm/scar
0.2
0.45
0.55
0.95
1.7
2.2
2.35
2.7
Ort.
Ort.
Ort.
Ort.
Ort.
Ort.
Ort.
Ort.
10.41
14.53
12.67
13.45
14.15
16.49
22.48
18.31
7.03
10.61
13.39
12.79
11.75
10.53
12.01
18.37
2.38
2.23
1.92
3.51
7.1
7.92
3.6
6.43
N
N
N
N
N
N
N
N
2
2
2
2
5.21
7.27
6.34
6.73
1
1
1
< 25
< 25
< 25
< 25
5
3
3
3.30
7.49
6.10
2
2
2
CAS
CAS
D
D
Qtz
Qtz
N
N
2.75
4.5
Ort.
Ort.
20.08
33.57
15.54
12.36
7.7
10.12
N
N
5
3
4.02
11.19
4
2
CAS
CAS
CAS
CAS
CAS
CAS
CAS
CAS
CAS
CAS
CAS
CAS
D
D
E
E
E
E
F
F
F
Oth.
Oth.
Oth.
Qtz
Sil.
Qtz
Qtz
Qtz
Qtz
Qtz
Qtz
Qtz
Qtz
Qtz
Qtz
N
N
N
Y
N
N
N
N
N
N
Y
N
6.1
5.65
1.55
1.8
3.95
6.25
0.75
0.75
1.85
0.2
0.2
0.25
Max.
Ort.
Ort.
Ort.
Ort.
Ort.
Ort.
Ort.
Ort.
Max.
Ort.
Ort.
34.15
24.84
20.66
25.5
13.89
21.84
14.08
17.34
19.11
10.09
10.92
13.4
16.91
29.3
13.53
8.25
20.11
25.66
14.16
14.45
13.58
6.13
6.71
7.31
10.21
6.46
4.78
6.33
9.13
8.52
3.56
2.12
6.78
2.94
2.26
2.43
N
N
N
N
N
N
N
N
N
Y
Y
N
4
4
3
4
6
9
2
6.21
6.89
6.38
2.32
2.43
7.04
2
2
2
4
4
1
6.37
2
5.46
6.70
CAS
CAS
CAS
CAS
Oth.
Oth.
Oth.
Oth.
Qtz
Qtz
Qtz
Qtz
Y
N
N
N
0.3
0.3
0.3
0.35
Max.
Max.
Ort.
Ort.
11.46
10.89
13.15
10.91
6.1
8.52
6.5
11.27
4.27
3.54
2.54
2.55
Y
N
N
Trv.
3
3
2
2
6
2
3
3
4.38
3.64
ED
Wt (g)
N
N
N
N
N
N
N
Y
Bipolar
HQ
%Unmod
RM
Qtz
Qtz
Qtz
Qtz
Qtz
Qtz
Qtz
Qtz
#Plat
XU/Cl.
D
D
D
D
D
D
D
D
Bk type
Pit/Sc.
CAS
CAS
CAS
CAS
CAS
CAS
CAS
CAS
< 25
< 25
< 25
N
N
N
N
N
N
Y
Y
N
Y
Y
N
Y
N
N
N
< 25
N
N
N
< 25
< 25
< 25
25-50
< 25
< 25
< 25
N
1
1
25-50
< 25
< 25
< 25
N
N
Y
N
N
N
N
N
N
N
N
Y
N
N
N
N
N
N
N
N
Y
2
2
< 25
< 25
< 25
< 25
N
N
N
N
N
N
N
462
T (mm)
Bk
3.49
3.64
3.39
3.27
3.06
3.49
3.77
3.27
N
Y
Y
N
N
N
N
N
CAS
CAS
Oth.
Oth.
Qtz
Qtz
N
N
0.6
0.6
Max.
Ort.
16.63
12.67
7.69
8.55
4.82
5.63
CAS
CAS
CAS
CAS
CAS
CAS
CAS
CAS
CAS
CAS
CAS
CAS
Oth.
Oth.
Oth.
Oth.
Oth.
Oth.
Oth.
Oth.
Oth.
Oth.
Oth.
Oth.
Qtz
Qtz
Qtz
Qtz
Qtz
Qtz
Qtz
Qtz
Qtz
Qtz
Qtz
Qtz
N
N
N
N
N
Y
N
N
N
N
N
N
0.6
0.65
0.7
0.75
0.8
0.9
0.9
0.9
0.9
0.95
0.95
0.95
Ort.
Ort.
Ort.
Ort.
Ort.
Ort.
Ort.
Ort.
Ort.
Ort.
Ort.
Ort.
15.8
11.39
12.63
11.85
24.16
13.58
15.26
11.36
11.68
17.06
16.45
15.1
7.73
8.93
14.45
12.35
15.51
12.17
14.34
12.2
10.67
10.26
10.9
11.37
CAS
CAS
CAS
CAS
Oth.
Oth.
Oth.
Oth.
Qtz
Qtz
Qtz
Qtz
N
N
N
N
1
1.1
1.1
1.1
Ort.
Ort.
Ort.
Ort.
12.99
13.11
21.02
16.24
12.31
9.35
13.12
13.7
Trv.
3
1
4
2
2
3
6
5
4.06
2
3.38
5.84
8.00
5.23
2.40
2.79
2
2
2
2
4
3
< 25
< 25
< 25
> 50-75
< 25
< 25
< 25
< 25
N
N
ED
W (mm)
6.83
8.1
7.06
8.27
8.95
8.04
6.98
8.67
Bipolar
L (mm)
12.19
14.32
13.51
11.68
16
15.7
14.41
13.93
%Unmod
Dims.
Ort.
Max.
Ort.
Ort.
Ort.
Ort.
Ort.
Ort.
#Plat
Wt (g)
0.35
0.4
0.45
0.5
0.5
0.5
0.55
0.55
mm/scar
HQ
Y
Y
N
N
N
Y
Y
Y
#NFS
RM
Qtz
Qtz
Qtz
Qtz
Qtz
Qtz
Qtz
Qtz
Bk type
XU/Cl.
Oth.
Oth.
Oth.
Oth.
Oth.
Oth.
Oth.
Oth.
Pit/Sc.
CAS
CAS
CAS
CAS
CAS
CAS
CAS
CAS
N
N
N
N
N
N
N
N
Y
N
N
N
N
N
N
N
N
1
3
4.22
1
> 50-75
< 25
4.02
4.58
4.24
4.45
3.24
3.1
3.14
5.15
6.04
3.86
3.86
5.41
N
N
N
N
N
N
N
N
N
Y
N
N
2
3
4
3
3
4
3
2
3
3
3
3
7.90
3.80
3.16
3.95
8.05
3.40
5.09
5.68
3.89
5.69
5.48
5.03
1
2
3
2
2
2
4
1
2
2
3
2
< 25
< 25
< 25
< 25
< 25
< 25
< 25
< 25
< 25
< 25
< 25
< 25
N
N
N
N
N
N
N
Y
Y
N
N
Y
N
N
Y
N
3.7
6.5
3.57
2.03
N
N
N
N
2
3
4
3
6.50
4.37
5.26
5.41
2
1
3
2
< 25
< 25
< 25
< 25
N
N
N
N
N
N
Y
N
Trv.
N
N
N
N
N
N
N
463
Bipolar
1.3
1.4
1.45
1.7
1.75
1.9
2.25
2.4
Ort.
Ort.
Ort.
Ort.
Ort.
Ort.
Ort.
Max.
16.4
23.81
19.48
18.24
22.7
16.59
23.42
20.5
13.3
10.83
10.59
16.93
12.82
13.99
16.97
17.01
4.51
5.08
5.98
4.32
6.62
6.32
5.08
7.3
N
N
N
N
N
N
N
N
4
3
3
3
3
4
3
4
4.10
7.94
6.49
6.08
7.57
4.15
7.81
2
2
3
2
2
3
2
CAS
CAS
Oth.
Oth.
Qtz
Qtz
Y
Y
2.4
2.65
Ort.
Ort.
20.04
24.62
12.13
10.45
5.14
5.73
N
N
5
3
4.01
8.21
CAS
CAS
CAS
CAS
CAS
CAS
CA1
CA1
CA1
CA1
CA1
CA1
Oth.
Oth.
Oth.
Oth.
Oth.
Oth.
1
1
2
2
2
2
Qtz
Qtz
Qtz
Sil.
Sil.
Sil.
Qtz
Qtz
Qtz
Qtz
Qtz
Qtz
Y
N
N
N
N
N
Y
Y
Y
N
Y
Y
3.8
7.05
12.55
0.85
3.25
5.05
0.2
0.4
0.15
0.15
0.25
0.45
Ort.
Ort.
Ort.
Ort.
Ort.
Ort.
Max.
Ort.
Ort.
Max.
Max.
Max.
22.52
26.2
33.6
14.29
19.77
30.99
10.6
14.69
10.28
10.16
10.73
11.8
25.26
30.6
24.38
14.3
23.85
20.74
7.83
7.79
9.7
7.18
7.91
11.77
5.79
7.55
14.96
3.5
5.39
7.55
2.42
2.69
1.75
2.7
3.88
3.8
N
N
N
N
Y
N
5
5
3
2
6
5
2
4
CA1
CA1
CA1
CA1
2
2
2
3
Qtz
Qtz
Qtz
BIF
Y
Y
N
Y
1.1
1.35
1.9
0.35
Ort.
Ort.
Ort.
Ort.
14.71
15.04
16.79
12.7
13.05
11.99
13.72
6.95
4.27
4.72
6.23
2.57
N
N
N
N
N
N
Y
Trv.
#NFS
Y
N
N
N
N
Y
N
N
Bk type
Qtz
Qtz
Qtz
Qtz
Qtz
Qtz
Qtz
Qtz
Pit/Sc.
Oth.
Oth.
Oth.
Oth.
Oth.
Oth.
Oth.
Oth.
N
2
3
< 25
< 25
N
< 25
< 25
< 25
N
Y
Y
N
Y
3
5
3
5.01
3.36
4.23
3
1
ED
%Unmod
2
N
N
N
N
N
N
N
N
Y
Y
N
N
#Plat
N
N
N
N
N
N
3.67
< 25
< 25
< 25
< 25
< 25
< 25
< 25
< 25
mm/scar
3
3
2
1
2
4
Bk
4.50
5.24
11.20
7.15
3.30
6.20
T (mm)
N
N
W (mm)
N
N
L (mm)
< 25
< 25
Dims.
3
2
Wt (g)
N
N
N
N
N
HQ
< 25
< 25
< 25
< 25
25-50
N
N
N
N
N
N
Y
N
RM
N
N
XU/Cl.
< 25
< 25
CAS
CAS
CAS
CAS
CAS
CAS
CAS
CAS
N
N
N
N
464
5.52
2
N
N
N
N
4
3
3
3
2.70
4.07
4.16
5.02
2
2
2
2
Ort.
Ort.
Ort.
Max.
Max.
Max.
Ort.
Ort.
11.19
10.28
10.17
10.44
11.9
10.63
11.52
10.94
4.02
4.6
4.86
4.63
4.56
6.44
4.52
7.26
0.96
2.59
1.61
2.2
2.52
1.36
2.44
2.05
N
N
Y
Y
Y
Y
N
N
CA1
CA1
3
3
Qtz
Qtz
N
Y
0.3
0.35
Max.
Ort.
11.24
14.29
7.74
5.86
4.42
3.79
Y
N
3
3
CA1
CA1
CA1
CA1
CA1
CA1
CA1
CA1
CA1
CA1
CA1
CA1
3
3
3
3
3
3
3
3
3
3
4
4
Qtz
Qtz
Qtz
Qtz
Qtz
Qtz
Qtz
Qtz
Qtz
Qtz
Qtz
Qtz
N
Y
Y
N
Y
N
Y
Y
Y
N
Y
Y
0.35
0.5
0.6
0.75
1
1.3
1.35
1.85
2.55
4.05
0.1
0.15
Max.
Ort.
Ort.
Ort.
Ort.
Ort.
Ort.
Ort.
Ort.
Ort.
Max.
Ort.
10.93
12.48
12.18
16.99
19.02
11.97
13.68
20.97
16.72
25.79
11.96
11.04
9.84
10.02
9.51
11.52
10.85
13.37
13.86
10.5
15.77
18.01
4.67
6.52
4.09
3.71
3.93
2.71
3.09
5.12
5.07
4.08
6.55
8.41
2.48
2.47
N
N
N
N
N
N
N
N
N
CA1
CA1
CA1
CA1
4
4
4
4
Qtz
Qtz
Qtz
Qtz
N
N
N
N
0.2
0.2
0.25
0.3
Ort.
Ort.
Ort.
Ort.
10.78
12.22
12.48
15.05
8.58
7.96
7.39
5.02
2.44
1.85
2.22
3.57
Trv.
#NFS
0.05
0.1
0.1
0.1
0.1
0.1
0.15
0.15
Bk type
Y
N
Y
N
Y
Y
Y
Y
2
3
2
3
4
1
1
2
N
N
N
N
N
ED
2
Qtz
Qtz
Qtz
Qtz
Qtz
Qtz
Qtz
Qtz
Bipolar
#Plat
N
3
3
3
3
3
3
3
3
Pit/Sc.
%Unmod
mm/scar
3
1
2
2
< 25
< 25
< 25
< 25
< 25
< 25
< 25
< 25
< 25
< 25
Bk
2
3
1
2
T (mm)
6.24
3.05
5.66
6.34
3.99
4.56
6.99
2.39
8.60
W (mm)
5
2
4
3
3
3
3
3
7
3
L (mm)
2
Dims.
4.76
< 25
< 25
Wt (g)
2
2
HQ
11.52
5.47
< 25
< 25
< 25
< 25
< 25
< 25
25-50
< 25
RM
1
2
1
XU/Cl.
5.60
3.43
5.09
CA1
CA1
CA1
CA1
CA1
CA1
CA1
CA1
N
N
N
N
N
N
N
N
N
Y
N
Y
N
N
N
N
N
N
N
N
< 25
N
N
N
N
N
N
N
N
N
Y
N
N
< 25
< 25
< 25
< 25
N
N
N
N
N
Y
N
N
N
N
N
N
465
T (mm)
Bk
2.26
2.36
3.82
3.8
2.27
3.45
1.77
3.24
Y
N
N
N
Y
N
N
N
CA1
CA1
4
4
Qtz
Qtz
Y
N
0.75
0.85
Max.
Ort.
15.81
16.92
12.15
12.26
3.52
3.22
CA1
CA1
CA1
CA1
CA1
CA1
CA1
CA1
CA1
CA1
CA1
CA1
4
4
4
4
4
4
4
4
5
5
5
5
Qtz
Qtz
Qtz
Qtz
Qtz
Qtz
Qtz
Sil.
BIF
Qtz
Qtz
Qtz
N
Y
N
N
N
N
N
N
Y
Y
N
N
1
1.1
1.15
1.75
2.75
3.8
4.45
0.3
0.45
0.1
0.15
0.35
Ort.
Ort.
Ort.
Ort.
Ort.
Ort.
Ort.
Ort.
Ort.
Ort.
Ort.
Ort.
11.84
15.12
14.81
18.04
18.28
18.87
18.25
10.61
13.55
10.58
10.39
10.3
10.93
10.79
15.08
18.25
15.98
14.53
19.38
10.8
8.29
5.38
5.38
8.4
CA1
CA1
CA1
CA1
5
5
5
5
Qtz
Qtz
Qtz
Qtz
N
N
N
N
0.35
0.4
0.45
0.45
Ort.
Ort.
Ort.
Ort.
11.09
13.57
13.99
15.06
7.09
7.89
9.39
8.95
Mrg.
Trv.
2
2
4
3
2
4
1
3
5.24
10.81
2.73
3.78
6.20
2.86
15.62
5.58
2
1
2
2
2
2
1
1
< 25
< 25
< 25
< 25
< 25
< 25
< 25
< 25
N
2
2
8.46
1
5.39
5.71
4.76
5.58
6.5
9.83
9.28
2.55
2.15
1.27
1.75
3.26
N
N
N
N
N
N
N
N
N
Y
N
N
7
5
3
3
6
4
4
2
2
2
2
1
1.69
3.02
4.94
6.01
3.05
4.72
4.56
5.31
6.78
5.29
5.20
10.30
2
4
3
3
3
2
2
3
3.26
4.06
3.13
2.61
N
N
N
N
5
2
4
2
2.22
6.79
3.50
7.53
3
1
2
2
Trv.
1
1
1
ED
W (mm)
8.1
7.04
7.46
8.41
8.04
8.55
9.43
7.89
Bipolar
L (mm)
10.47
21.62
10.9
11.35
12.4
11.42
15.62
16.75
%Unmod
Dims.
Ort.
Ort.
Ort.
Ort.
Ort.
Ort.
Ort.
Ort.
#Plat
Wt (g)
0.3
0.35
0.35
0.35
0.35
0.4
0.4
0.55
mm/scar
HQ
N
N
Y
N
N
Y
N
Y
#NFS
RM
Qtz
Qtz
Qtz
Qtz
Qtz
Qtz
Qtz
Qtz
Bk type
XU/Cl.
4
4
4
4
4
4
4
4
Pit/Sc.
CA1
CA1
CA1
CA1
CA1
CA1
CA1
CA1
N
N
< 25
< 25
N
N
N
< 25
< 25
< 25
< 25
< 25
< 25
25-50
< 25
< 25
< 25
< 25
Y
N
N
Y
Y
Y
N
N
N
N
N
N
N
Y
N
N
N
N
N
N
Y
N
N
Y
N
Y
N
N
N
N
N
< 25
< 25
< 25
< 25
N
N
N
N
N
N
N
N
N
N
N
466
N
N
N
Y
N
N
N
N
N
N
N
N
2
6
2
4
3
4
2
4
3
6
4
5
8.45
3.32
5.08
3.53
3.79
2.97
5.72
3.39
3.95
1.97
3.42
5.46
2
25-50
< 25
2
1
< 25
< 25
1
3
3
2
2
3
< 25
< 25
< 25
< 25
N
Y
N
N
N
N
N
N
N
Y
N
N
N
N
N
Y
N
N
N
Y
N
N
N
N
5.82
6.29
2.3
2.34
N
N
3
4
2
2
4.32
4.65
2
2
< 25
< 25
< 25
7.60
2
Y
Y
N
N
N
N
N
N
Ort.
Ort.
Ort.
Ort.
Ort.
Ort.
Ort.
Max.
11.82
12.4
16.6
19.29
18.9
11.15
10.81
11.02
6.04
10.57
9.71
9.53
16.02
8.86
4.41
6.42
4.06
2.89
3.75
6.12
7.5
1.3
1.35
3.05
N
Y
N
N
N
N
N
Y
CA1
CA1
6
6
Qtz
Qtz
N
N
0.3
0.35
Ort.
Ort.
10.52
11.64
6.49
7.8
2.83
2.65
CA1
CA1
CA1
CA1
CA1
CA1
CA1
CA1
CA1
CA1
CA1
CA1
6
6
7
7
7
7
7
7
7
7
7
7
Qtz
Qtz
Qtz
Qtz
Qtz
Qtz
Qtz
Qtz
Qtz
Qtz
Qtz
Qtz
N
N
N
N
N
N
N
N
N
N
N
N
0.75
6.45
0.2
0.25
0.3
0.45
0.55
0.65
0.75
1.1
1.4
3.4
Ort.
Ort.
Ort.
Ort.
Ort.
Ort.
Ort.
Ort.
Ort.
Ort.
Ort.
Ort.
16.89
19.94
10.15
14.1
11.36
11.89
11.44
13.55
11.85
11.8
13.69
27.32
7.82
17.12
4.92
6.72
7.05
7.95
8.72
8.45
12.52
9.37
9.15
14.08
CA1
CA1
CA2
CA2
8
8
1
1
Qtz
Qtz
Qtz
Qtz
N
N
N
Y
1.3
1.55
0.2
0.35
Ort.
Ort.
Max.
Ort.
12.95
18.6
10.54
15.2
11.37
8.26
7.6
8.48
N
Trv.
Trv.
< 25
ED
4.34
10.05
2.48
3.01
2.65
4.21
5.09
4.72
3.5
5.54
7.25
7
0.5
0.55
0.75
1.4
2.05
0.15
0.1
0.2
Bipolar
#Plat
N
N
N
N
N
N
N
N
N
N
Pit/Sc.
%Unmod
mm/scar
N
Qtz
Qtz
Qtz
Qtz
Qtz
Sil.
Qtz
Qtz
#NFS
> 50-75
< 25
5
5
5
5
5
5
6
6
Bk type
1
1
Bk
10.52
3.88
T (mm)
1
3
W (mm)
N
L (mm)
N
N
N
N
N
N
N
N
Dims.
N
N
N
Y
N
N
Y
Wt (g)
< 25
< 25
< 25
< 25
> 50-75
< 25
< 25
< 25
HQ
1
3
2
3
1
2
2
RM
5.91
4.13
5.53
3.22
9.45
5.58
3.60
XU/Cl.
2
3
3
6
2
2
3
2
CA1
CA1
CA1
CA1
CA1
CA1
CA1
CA1
467
N
N
Y
CA2
CA2
3
3
Qtz
Qtz
N
N
0.85
4
Ort.
Ort.
13.32
18.28
10.51
15.02
7.03
9.9
CA2
CA2
CA2
CA2
CA2
CA2
CA2
CA2
CA2
CA2
CA2
CA2
4
4
4
4
4
4
4
4
4
4
4
4
Qtz
Qtz
Qtz
Qtz
Qtz
Qtz
Qtz
Qtz
Qtz
Qtz
Qtz
Qtz
Y
N
Y
Y
N
N
Y
Y
N
N
Y
Y
0.15
0.2
0.2
0.25
0.3
0.3
0.4
0.5
0.6
0.65
4.75
< 0.05
Max.
Ort.
Max.
Max.
Ort.
Ort.
Ort.
Ort.
Ort.
Ort.
Ort.
Ort.
10.22
11.28
12.36
11.11
12.34
10.61
10.78
11.72
19.11
10.84
24.56
11.09
5.83
6.51
11.85
6.94
6.22
12.5
9.81
9.49
7.83
13.06
22.28
2.29
2.48
2.4
1.67
3.5
2.68
2.42
3.86
3.15
2.9
4.02
7.36
0.82
CA2
CA2
CA2
CA2
5
5
5
5
Qtz
Qtz
Qtz
Qtz
N
N
N
N
0.2
0.25
0.3
1.85
Max.
Ort.
Ort.
Ort.
10.99
11.36
13.8
18.57
4.98
7.07
5.62
13.04
3.34
4.14
4.43
5.42
N
Y
N
N
4
2
3
2
4.09
5.48
2
< 25
< 25
< 25
Y
N
N
2
3
2
6.66
3.82
6.11
1
2
1
25-50
< 25
< 25
N
N
N
N
N
N
3
6
4.44
3.05
2
3
< 25
< 25
Y
Y
Y
N
Y
Y
2
2
3
4
1
3
5
3
3
2
4
2
N
N
12.34
3.54
2.16
3.91
6.37
5.42
6.14
5.55
< 25
< 25
< 25
< 25
< 25
< 25
< 25
< 25
< 25
< 25
< 25
2
4
4
4
2.84
3.45
4.64
N
N
N
N
N
N
N
N
N
Trv.
5.64
1
3
2
2
2
3
2
3
< 25
< 25
< 25
< 25
ED
T (mm)
8.91
0.98
2.1
2.07
2.31
2.67
3.22
2.41
Bipolar
W (mm)
25.91
4.21
6.59
6.96
7.47
4.99
6.88
9.81
%Unmod
L (mm)
16.34
10.96
11.2
11.93
10.93
13.31
11.46
12.22
#Plat
Dims.
Ort.
Ort.
Max.
Max.
Ort.
Ort.
Ort.
Ort.
mm/scar
Wt (g)
5.1
< 0.05
0.1
0.15
0.2
0.2
0.25
0.35
#NFS
HQ
N
Y
Y
Y
N
Y
Y
N
Bk type
RM
Sil.
BIF
Qtz
Qtz
Qtz
Qtz
Qtz
Qtz
Bk
XU/Cl.
1
2
3
3
3
3
3
3
Pit/Sc.
CA2
CA2
CA2
CA2
CA2
CA2
CA2
CA2
N
N
N
Y
Y
N
N
N
Y
N
N
N
N
N
N
N
N
N
N
N
N
N
N
Y
Y
N
N
N
N
N
N
N
N
Y
N
Y
468
Dims.
L (mm)
W (mm)
T (mm)
Ret.
Y
N
Y
N
Y
Y
N
N
N
N
N
N
Y
0.75
0.3
1.2
0.25
1.3
0.95
2.5
0.1
0.1
0.3
0.1
0.2
0.05
Max.
Ort.
Ort.
Ort.
Ort.
Ort.
Ort.
Ort.
Ort.
Ort.
Ort.
Ort.
Ort.
16.42
11.64
12.32
12.2
26.37
12.73
18.17
7.28
5.68
11.78
7.37
8.6
6.13
9.24
7.78
12.57
6.78
12.99
12.93
17.22
3.42
4.69
5.88
4.86
6.28
3.48
4.47
3.98
5.89
2.32
2.52
5.78
5.98
1.67
2.7
4.63
2.35
2.81
1.85
Bck.
Bck.
Bck.
Bck.
Uni.
Uni.
Uni.
Bck.
Bck.
Uni.
Bck.
Bck.
Bck.
N
N
N
TH1
TH1
TH1
TH1
TH1
TH1
TH1
TH1
TH1
TH1
TH1
3
3
4
4
4
5
5
5
6
6
6
Qtz
Qtz
Qtz
Qtz
Qtz
Qtz
Qtz
Qtz
Qtz
Qtz
Qtz
Y
Y
N
N
N
Y
N
N
N
N
N
0.1
< 0.05
0.1
0.15
0.2
0.05
0.1
< 0.05
0.1
0.1
0.2
Ort.
Max.
Max.
Ort.
Ort.
Ort.
7.46
5.12
6.62
9.55
6.37
5.51
5.6
3.64
4.83
5.52
10.44
4.16
1.89
1.84
2.87
2.28
3.08
2.2
N
Y
Y
Ort.
Ort.
Ort.
Ort.
4.95
7.24
5.96
8.44
3.99
5.96
4.5
5.45
2.03
1.9
2.5
2.71
Bck.
Bck.
Bck.
Bck.
Uni.
Bck.
Bck.
Bck.
Bck.
Bck.
Bck.
N
N
N
Y
Y
Y
N
Y
N
Y
Y
Y
Y
Y
Y
ED
Wt (g)
Qtz
Qtz
Qtz
Qtz
Qtz
Qtz
Qtz
Qtz
Qtz
Qtz
Qtz
Qtz
Qtz
Bk type
HQ
9
3
3
4
4
1
1
1
2
2
3
Bk
RM
CS
MC1
MC1
MC1
MC1
HS
HS
TH1
TH1
TH1
TH1
TH1
TH1
Pit/Sc.
XU/Cl.
G.6 RETOUCHED FLAKES
N
N
Trv.
Trv.
Trv.
Trv.
Trv.
Trv.
Trv.
Trv.
Trv.
N
N
N
Y
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
469
Wt (g)
Dims.
L (mm)
W (mm)
T (mm)
Ret.
0.2
0.35
0.6
0.7
1.1
0.5
0.25
53.65
Ort.
Ort.
Ort.
Ort.
Ort.
Ort.
Max.
Ort.
8.77
13.36
20.13
17.93
16.04
13.47
12.96
53.7
7.94
9.48
7.56
8.77
12.93
9.99
9.42
43.42
2.94
2.35
3.09
3.29
3.57
2.21
2.74
16.33
Bck.
Bck.
Bck.
Bck.
Bck.
Bck.
Bck.
Uni.
N
N
N
N
N
N
N
N
Y
CAS
CAS
Oth.
Oth.
Qtz
Qtz
Y
Y
0.15
0.45
Ort.
Ort.
12.3
12.84
5.31
8.04
1.95
2.61
Bck.
Bck.
N
N
N
Y
CA1
CA1
CA1
CA1
CA1
CA1
CA1
CA1
CA1
CA1
CA2
CA2
1
1
3
3
4
4
4
4
4
5
1
2
Chal.
Qtz
Qtz
Qtz
Qtz
Qtz
Qtz
Qtz
Qtz
Qtz
Qtz
Qtz
Y
Y
N
Y
Y
Y
N
Y
N
Y
Y
Y
0.1
0.1
0.6
0.9
0.05
0.1
0.15
0.4
0.5
0.1
0.3
0.1
Ort.
Ort.
Ort.
Ort.
Ort.
Ort.
Ort.
Ort.
Ort.
Ort.
Ort.
Ort.
7.95
11.55
17.04
24.95
11
10.03
15.11
14.66
12.47
9.09
10.8
12.82
6.35
4.76
8.91
7.52
3.36
6.91
4.74
6.96
9.83
4.95
10.35
3.09
1.91
2.05
2.93
4.86
1.91
1.14
1.43
2.32
2.94
2.14
2.23
2.66
Uni.
Uni.
Bck.
Bck.
Bck.
Bck.
Bck.
Bck.
Bck.
Bck.
Uni.
Bck.
N
N
N
N
Y
N
N
N
N
Y
N
N
N
N
N
N
N
Y
N
N
N
N
N
CA2
CA2
CA2
CA2
2
2
2
2
Qtz
Qtz
Qtz
Qtz
Y
Y
N
Y
0.15
0.15
0.15
0.2
Ort.
Ort.
Ort.
Ort.
12.83
8.93
10.86
15.61
4.55
5.91
7.15
4.11
3.39
1.83
1.76
1.83
Bck.
Bck.
Uni.
Bck.
N
N
N
N
N
N
N
N
ED
HQ
Y
Y
Y
Y
Y
N
Y
N
Bk type
RM
Qtz
Qtz
Qtz
Qtz
Qtz
Qtz
Qtz
Qtz
Bk
XU/Cl.
A
A
A
A
B
D
E
E
Pit/Sc.
CAS
CAS
CAS
CAS
CAS
CAS
CAS
CAS
N
N
N
Y
Y
Trv.
Trv.
470
Wt (g)
Dims.
L (mm)
W (mm)
T (mm)
Ret.
0.4
0.5
0.85
0.65
0.95
0.1
0.3
4.1
Ort.
Ort.
Ort.
Ort.
Ort.
Max.
Max.
Ort.
15.13
12.36
19.31
14.73
13.02
7.09
12.26
22.34
7.36
10.76
9.39
8.79
12.61
5.72
6.45
16.63
2.48
2.66
4.54
3.6
4.75
2.08
3.91
8.53
Bck.
Uni.
Bck.
Uni.
Uni.
Bck.
Bck.
Uni.
N
CA2
5
Qtz
Y
0.05
Ort.
9.75
3.31
1.7
Bck.
N
Y
N
N
Y
Y
N
Trv.
ED
HQ
Y
Y
Y
N
Y
Y
Y
N
Bk type
RM
Qtz
Qtz
Qtz
Qtz
Qtz
Qtz
Qtz
Qtz
Bk
XU/Cl.
2
2
2
3
3
4
4
4
Pit/Sc.
CA2
CA2
CA2
CA2
CA2
CA2
CA2
CA2
N
N
N
N
N
N
Y
N
N
471