Quaternary Science Reviews 21 (2002) 1023–1037 A Late Pleistocene vegetation history from the Australian semi-arid zone Judith H. Fielda,*, John R. Dodsonb, Ian P. Prosserc a Archaeology A14, University of Sydney, NSW, Sydney 2006, Australia Department of Geography, University of Western Australia, Perth 6907, WA, Australia c Division of Water Research, CSIRO, GPO Box 1666, Canberra 2601, ACT, Australia b Abstract Cuddie Springs is an ephemeral lake in central northern New South Wales, Australia. The upper 3 m of sediment consist of lacustrine clays containing a Late Pleistocene sequence of extinct and extant fauna, and in the upper 1.7 m, an associated archaeological record. Changes observed in the pollen sequence include: (i) a peak in charcoal values corresponding to a dramatic decline in Casuarina woodland to chenopod shrubland at 2.5 m, respresenting a climatic shift to more arid conditions; (ii) chenopod shrubland moved into decline with the spread of grasslands around 1.7 m, and the amelioration in climatic conditions persisted until approximately 28,000 BP. A regime emerged which resulted in extended lake dry periods and peak aridity by approximately 19,000 BP and (iii) at 1 m depth, around 19,000 BP a shift to peak arid conditions is observed with a return of Chenopodiaceae and a decline in grasses. The lake entered an ephemeral phase that has persisted until the present day. The broad palaeoenvironmental framework of lake history, climate and vegetation change spans the archaeological and faunal records from Cuddie Springs. The direct association enables a closer examination of causation in faunal extinctions and human subsistence activities in the Australian arid zone. r 2002 Elsevier Science Ltd. All rights reserved. 1. Introduction Towards the close of the Pleistocene a number of dramatic environmental changes are recorded in the fossil record for the Australian continent (as elsewhere). The culmination of the glacial period at the Last Glacial Maximum (LGM), the extinction of a suite of large animals (the megafauna) and the arrival of people have all provided a complex record for the Australian arid zone. In Australia, the range of sites where pollen records for the Late Pleistocene are found is restricted to the more temperate margins of the continent (Kershaw, 1985; Singh and Geissler, 1985; Colhoun and van de Geer, 1988; Pickett, 1997). The aridity of the Australian interior has precluded widespread formation of organicrich deposits with pollen. Data comes mainly from salt lakes, springs and anomalous peat deposits that generally have only shallow time-depth records (e.g. Bell et al., 1989; Boyd, 1990). The exceptions to these examples are studies from north-west Australia, the *Corresponding author. Tel.: +61-2-9351-7412; fax: +61-2-93515712. E-mail address: j.field@chem.usyd.edu.au (J.H. Field). Nullarbor Plain, Lake Frome, Ulungra Springs and Cuddie Springs (Martin, 1973; Wyrwoll et al., 1986; Dodson and Wright 1989; Singh and Luly, 1991; Dodson et al., 1993). Of these studies only Ulungra Springs and Cuddie Springs provide information for the pre-LGM. Tracking the course of people and fauna through time in the arid zone has been difficult also because there are few preserved stratified sites. Open sites are rarely found intact leading most investigations to be focused on rockshelters and caves, locations often devoid of palaeoenvironmental data. As a result, researchers have turned to charcoal, phytolith and faunal studies in an attempt to reconstruct vegetation histories (Vrba, 1981; Dortch, 1984; Smith et al., 1995; Bowdery, 1998). While providing some clues to the nature of the local environment, faunal deposits can be of limited value considering the ‘‘disharmonious’’ assemblages that are known from the Pleistocene with no modern analogues (e.g. Graham and Lundelius, 1984). The open archaeological site of Cuddie Springs is interesting because of the associated archaeological, palaeoenvironmental and faunal records (Field and Dodson, 1999). The latter two extend well beyond the 0277-3791/02/$ - see front matter r 2002 Elsevier Science Ltd. All rights reserved. PII: S 0 2 7 7 - 3 7 9 1 ( 0 1 ) 0 0 0 5 7 - 9 1024 J.H. Field et al. / Quaternary Science Reviews 21 (2002) 1023–1037 limits of the radiocarbon technique and possibly beyond known human occupation of the southern part of the continent (see O’Connell and Allen, 1998). The aim of this paper is to report a Late Pleistocene environmental history of the semi-arid zone around Cuddie Springs. This study provides a record of vegetation and environment covering a pre-LGM time period not previously reported for arid zone pollen deposits. 2. Cuddie Springs Cuddie Springs (Fig. 1) was known as a fossil fauna site for over a century (Abbott, 1881; Wilkinson, 1885) but only recently has it also been established as containing important archaeological and palaeoenvironmental records. The results to date include: (i) establishing the presence of people in this part of the continent at B35,000 years ago; (ii) the co-existence of humans with a suite of extinct animals (megafauna); and (iii) the presence of seed-grinding implements in a preLGM context (Fullagar and Field, 1997; Field and Fullagar, 1998; Field and Boles, 1998; Field and Dodson, 1999). The preservation of palaeoenvironmental information was established in earlier studies (Dodson et al., 1993) and revealed the potential for Cuddie Springs to provide a substantial contribution to our knowledge concerning arid zone vegetation histories for the Late Pleistocene. 2.1. Environmental setting The Cuddie Springs lakebed is located on the northwest plains of New South Wales between the Macquarie River and Marra Creek channels and is not part of a currently active river system. The site is on the southern boundary of the summer rainfall belt in the semi-arid zone, with a mean rainfall of approximately 400 mm/ year. The average temperatures for the winter months range from 5.31C to 18.51C, with summer maxima often exceeding 401C. The lake bed is approximately 2 km in diameter and lies in a landscape of o20 m relief. The fossil and archaeological records are found in the centre of a claypan (approximately 200 m in diameter) on the lake floor (Furby, 1995, Fig. 2). The area around Cuddie Springs supports vegetation that is characteristic of semi-arid environments (Fig. 3). The lake is characterised by grey alluvial soils which support Eucalyptus largiflorens and Eucalyptus microtheca interspersed with Acacia stenophylla. Atriplex species and Muehlenbeckia occur with Rhagodia spinescens, Goodenia sp. and some unidentified grasses. The mistletoe, Lysiana exocarpi sp. tenius is found on many of the trees in the area. The surrounding red soil plains support vegetation that includes Callitris columellaris, Casuarina luehmanii, Eucalyptus populnea, Geigera parviflora, Flindersia maculosa and a range of shrubs including Atriplex, Eremophila, Santalum, Senna, Lepidium and Piemelea species. Early last century Cuddie Springs was reported as the only source of water between the Barwon and the Macquarie River during dry seasons (Anderson and Fletcher, 1934). The claypan fills after local storms from rainwater and local surface run-off, and the greater lake floor may fill after exceptional rainfall taking up to 12 months to dry (see Field and Dodson, 1999) (Fig. 4). When the claypan floods (B1 in 3 years), large numbers of waterbirds are attracted to the Fig. 1. Location of Cuddie Springs in south-eastern Australia and in relation to the current river systems of the north-western riverine plains of New South Wales. (Illustration: Fiona Roberts). J.H. Field et al. / Quaternary Science Reviews 21 (2002) 1023–1037 1025 Fig. 2. The Cuddie Springs claypan. Excavations at the centre of the claypan (arrow) have revealed a record of fossil fauna, environment and archaeology. (Photo: J. Field). Fig. 3. Vegetation on the lake floor at Cuddie Springs, characterised by Eucalyptus species that tolerate periodic inundation. This photo was taken during a wet year, in dry years there is no understorey vegetation visible. (Photo: J. Field). area and many aquatic plant species emerge. The claypan coincidentally dried after the sinking of a bore on the adjacent property of Moranding in 1902, and it was proposed then that this was responsible for the drying of Cuddie Springs. The site has been described in the literature as both a spring and a 1026 J.H. Field et al. / Quaternary Science Reviews 21 (2002) 1023–1037 Fig. 4. View of the claypan (background) from the southern margins showing local runoff held back from the claypan by a low levee (foreground). The claypan fills after local rainfall and may take several months to dry. (Photo: J. Field). mound spring with interpretations of its formation ranging from meteorite impacts to deflation hollows and eroded structural domes (Habermahl, 1980; Meakin, 1991). On the north-eastern margin of the lake there is a remnant source bordering dune approximately 30 cm high, comprised of coarse and fine sand (see Fig. 5). Numerous artefacts, including backed pieces (Hiscock and Attenbrow, 1996) and grinding stones have been observed eroding out of the dune. Cuddie Springs is located on Quarternary alluvium and aeolian deposits, consisting of gravels, sand, silt and clay. These deposits are derived from fluvial sources to the west (Marra Creek), the north (Barwon River) and the east (Macquarie River). These fluvial systems no longer contribute sediments to the Cuddie Springs lake. The dark features on the aerial photos that appear to be palaeo-channels to the north and south of the lake floor (Fig. 5) are of a wavelength and width that are many times larger than the present rivers in the region. The palaeo-channels and the feature called Cuddie Springs are a boundary (a low ‘‘back swamp’’) between the Marra Creek floodplain and the slightly higher red soils to the east. 2.2. Stratigraphy of the claypan sediments The stratigraphy of the Cuddie Springs claypan was documented by the Geological Survey of New South Wales in 1990 after they recovered a 54 m core from the claypan periphery (Meakin, 1991). Following archaeological excavations undertaken by Field and associates in 1991, the stratigraphy in the claypan centre was also recorded (Dodson et al., 1993; Furby et al., 1993; Field and Dodson, 1999; Fig. 6). 2.2.1. Geological survey of NSW DDH: claypan periphery (Profile 1, Fig. 6) From the base of the DDH to approximately 38 m depth, the core consisted of gravels and medium to coarse sands indicative of a fluvial sequence. Between 39 and 38 m depth, carbonaceous fragments and pieces of pyrite were found and these have been interpreted by Meakin (1991) as indicating the presence of vegetation, possibly on a swampy sandbank. The Geological Survey of NSW also obtained one TL determination of 467,0007150,000 at a depth of 40–45 m in the unconsolidated sand. This would appear to be a minimum date considering the presence of Cretaceous sediments at 54 m depth (Meakin, 1991). J.H. Field et al. / Quaternary Science Reviews 21 (2002) 1023–1037 Fig. 5. Aerial photograph of the Cuddie Springs lake and surrounds. The lake floor is approximately 2 km in diameter and the claypan (arrow) is where the fossil records are found. (Photo courtesy of the NSW Government Department of Lands). Above 38 m depth to approximately 12.6 m from the surface, the sediments were described by Meakin as lacustrine silts deposited in a low energy depositional environment. The upper section of this unit is interpreted by Meakin to represent a shallow lake phase with a higher energy regime than in previous periods as suggested by the ‘‘angled cross beds and fine sandy laminae’’ (Meakin, 1991). From 12.6 to 1.5 m depth, there are fine to medium grained, well sorted sands, which are overlain by lacustrine clays. These clays, comprising the upper 1.5 m of sediments, are interpreted as intermittent flood deposits (Meakin, 1991). 2.2.2. Archaeological excavations: claypan centre (Profile 2, Fig. 6) In 1991, a pit at the claypan centre was excavated to a depth of 3 m (Dodson et al., 1993). The sediments in the claypan centre are at the lowest point on the lake floor and are constantly damp below about 1 m depth. Sampling of the deposits below 3 m was undertaken using a sand auger. Particle size analysis of the sediments was compiled to 5 m depth for the 1991 1027 excavations (Dodson et al., 1993; Fig. 7). Further sampling was conducted in March 1994 and ceased at 10 m depth due to waterlogging and slumping of the deposit. The results showed that the Cuddie Springs profile consists of a number of lithologically distinct units as described below. From approximately 10 to 7 m depth at the claypan centre, the sediments consist of gravels and coarse sand with high concentrations of bone, both complete and fragmented. In situ deposition of the bone is suggested by the absence of any significant abrasion. Some of the bone recovered had teeth marks on the surface and all of the bone exhibited various degrees of mineralisation. Nodules of waterlogged clays (gley) were also found but they contained no pollen. These deposits are interpreted as a fluvial stage prior to the formation of the lake. From 7 m to 3 m depth there are clean, well sorted medium to fine grained sands of possible aeolian origin. An aeolian source is inferred from the particle size and the degree of sorting that is common for wind blown sands as compared with fluvially sorted sands (e.g. Watson, 1989). At 5 m depth the sediments contain little silt and clay (o15%) most probably introduced by root channels from the overlying deposits. This horizon appears to correspond to the sand horizon identified by Meakin (1991) from 12.6 to 1.5 m in the Geological Survey drill hole (Profile 1, Geol. Survey DDH, Fig. 6). The variable depth of the sand suggests a dune. At the top of the aeolian sand unit there is a sharp boundary where the lacustrine clays begin and this is also marked by an apparent increase in the concentration of fossil bone. The lacustrine facies consist of a sequence of sandy clays and silts that extend to the surface (Fig. 6). These have been interpreted to represent pan and ephemeral conditions and are mostly grey to greyish brown in colour (Munsell colour 5Y 7/1), are massive in structure and highly cohesive. A series of auger samples has shown that the upper 3 m of lacustrine clays are continuous to the edge of the lake. Two layers of bone identified as occurring below the archaeological record sediments have sharp upper boundaries, possibly representing erosion surfaces. In both layers (at B1.8 and B2 m depth), bones are found in a matrix of quartz gravel and coarse sands, interpreted as lake shore (beach lag) depositional environments. Overlying the bone horizons are lake deposits comprised principally of clay with a minor component of sand and numerous small root channels. The clay deposits are interpreted as a shallow lake or swamp, with decreasing clay content corresponding to ephemeral lake conditions. Between B1.7 and 1.35 m depth there is sharp increase in the clay and silt content of the sediments. From 1.35 m to approximately 1.1 m depth the particle size analysis indicates a return to the ephemeral conditions seen in the lower levels. At approximately 1028 J.H. Field et al. / Quaternary Science Reviews 21 (2002) 1023–1037 comprises clay and silt similar to that observed between 1 and 0.47 m depth and is consistent with deposition from intermittent flooding such as the ephemeral conditions observed in the present day. Bone has been recovered from all stratigraphic levels at the claypan centre and the concentration of bone appears to increase in the coarse gravel deposits at around 7 m depth. The colour of the bone was highly variable in most levels ranging from light brown friable bone to very dense black material. ICP-AES analysis of the bone has shown that manganese is the primary mineralising element throughout the profile and the colour of the bone directly correlates to the manganese content (K. Privat and J. Field, unpublished results, 2000). At the lowest levels of sampling with the auger, bone comprised about 80% of the auger bucket contents and was found in varying degrees of fragmentation. At the base of the 1991 excavations, in the surface of the aeolian sands at 3 m depth, a number of bone elements, including a Diprotodon sp. mandibular ramus were found in a para-conformal orientation. Above 1.7 m depth the bone is unabraded indicating an in situ deposition in the claypan. Fig. 6. Stratigraphy of the Geological Survey DDH core (Profile 1) and the archaeological excavations at the centre of the Cuddie Springs claypan (Profile 2). The DDH core was recovered approximately 60 m from the excavations. No bone was reported nor was pollen preserved in the core sediments. DS: Deflation Surface; FS: Ferruginised Sands, which are overlain by a concreted beach lag deposit. (Illustration: Megan Mebberson). 1.1–1 m depth there is a concentration of stone, bone and charcoal interpreted as a deflation pavement that dips to the south-west. At 0.47–0.45m there is a layer of finely laminated silts, divided in places into two layers indicating deposition under standing water. From 0.45 m to approximately 0.25 m depth there is a band of well sorted orange gravel and quartz sand within which are found numerous exfoliated teeth, mineralised bone fragments and artefacts. This unit is discontinuous over a wider area, is similar to the sands recovered at depth (>5 m), and appears to represent spoil from the well digging in the 1950s. The upper 25 cm of deposit 2.2.3. Chronology The upper lacustrine deposits have been dated using conventional and AMS radiocarbon techniques (Field and Dodson, 1999). The fifteen radiocarbon dates indicate that the unit with the human/megafauna overlap (B1.7–1 m depth) was deposited sometime between approximately 34,000 and 28,000 BP (Field and Dodson, 1999). The disconformity at 1 m represents a time-compressed unit of up to 10,000 years. While the upper metre of sediment is disturbed, the earliest date from this level is approximately 19,000 years. One OSL age has been obtained at 1.55 m of 35,40075800 (ANUOD118a.) (Field and Dodson, 1999). Additional radiocarbon determinations for the AMS technique were prepared using the ABOX-SC procedure (Bird et al., 1999; Fifield et al., in press) and returned ages that fall within the range of radiocarbon ages obtained for archaeological levels 1–4 using conventional and standard AMS methods (see Table 1). 3. Sampling and pollen preparation Samples for pollen analysis (3 cm2) were collected at 5 cm intervals from 3.05 m depth to the surface. They were placed into glass vials in the field with a spatula that was cleaned after each use and then transported to the laboratory. These samples supplement a previous pollen analysis (Dodson et al., 1993) in which samples were collected at 5 cm intervals to a depth of 1.8 m. J.H. Field et al. / Quaternary Science Reviews 21 (2002) 1023–1037 1029 swamps and back waters, often blanketing large areas by the massing of innumerable plants’’ (Aston, 1973; see also Cunningham et al., 1992: 34). Foster and Harris (1981: 202) observe that the ideal growing conditions for Azolla are those ‘‘where the effects of turbulence and periodic flooding will not fragment the colonies’’. The Tertiary age beds that were the subject of the study by Foster and Harris (1981) were interpreted as shallow, permanent lakes, based on the Azolla content and the fine grained enclosing sediment. The presence of Azolla in the fossil pollen record is therefore interpreted as an indicator of freshwater conditions at Cuddie Springs. The pollen diagram was prepared using the TILIA program (Grimm, 1991). The taxa in the diagram have also been grouped as trees, shrubs, herbs, aquatics and fern taxa. The sequence has also been divided into a number of zones based on observations of sediment changes and major pollen changes through the profile. The full age of the fossil pollen deposits is unknown since the lower layers extend beyond the limits of radiocarbon. However, the age determinations for the beginning of the human/megafauna overlap, at about 1.7 m depth, are around B35,000 years (Field and Dodson, 1999). 4. Pollen Analysis 4.1. Zone 1.3.10–2.50 m; age unknown Fig. 7. Particle size analysis of the Cuddie Springs claypan deposit (Profile 2, Fig. 6) (after Furby, 1991; Dodson et al., 1993). The relative percentages of clay, silt, fine and coarse sand and gravel with depth to 5 m are shown. 3.1. Chemical preparation of pollen samples The pollen was separated from the high clay content of the sediments by heavy liquid separation using ZnBr2 in bromoform (Faegri and Iversen, 1975). The organic fraction was collected onto an acetate based filter paper, which was dissolved in acetone and removed before treatment with HF and acetolysis (Moore et al., 1991). The samples were dehydrated through an alcohol series and mounted in silicone oil (4000 cs viscosity). Terrestrial and aquatic taxa were counted until a minimum of 200 grains of dryland taxa had been identified for each level. These counts formed the pollen sum used in the pollen diagram (Fig. 8). The charcoal was analysed using the point count method of Clark (1982). Azolla spores were counted separately and are represented in the diagram as raw counts. Spores and massulae of the small freshwater fern Azolla have been used in this study as an indicator of relatively high-level lake conditions. Azolla today is generally found ‘‘floating on still, fresh waters of stock tanks, lagoons, Zone 1 is defined by high levels of Casuarinaceae with relatively high levels of Poaceae and Asteraceae throughout this zone. Eucalyptus and Acacia are also represented, although at lower frequencies compared with Casuarinaceae. Some herbaceous taxa (e.g. Malvaceae) are also present. Towards the upper boundary of Zone 1, there is a sharp increase in charcoal values corresponding to a sharp decrease in the relative abundance of Casuarinaceae, Malvaceae and herbaceous taxa. Chenopodiaceae, Dodonaea and Asteraceae also decrease at the top of Zone 1. Changing freshwater lake levels are indicated by the fluctuating counts of Azolla spores. 4.2. Zone 2.2.5–1.9 m; age unknown The lower boundary of Zone 2 is defined by a sharp decrease in Casuarinaceae, associated with a corresponding increase in Chenopodiaceae values. Poaceae and Asteraceae values maintain low levels, but increase towards the top of the zone. The aquatic taxa Cyperaceae, Typha and Myriohphyllum are represented throughout this zone. Azolla levels are high at the base of Zone 2, a continuation of the high levels in Zone 1. Eucalyptus appears to be represented by consistently low numbers throughout the sequence. The semi-arid 1030 Table 1 Summary of pollen, geomorphology, archaeology and faunal data for the Cuddie Springs lacustrine facies Pollen zone Arch. levela Lithostratigraphy Ageb Climate Environs Archaeology/Fossil finds Surface 8 6 Clay, silt Late Holocene (1470770 Beta 81,385) Semi-arid Open forest with chenopod shrublands, ephemeral lake conditions Surface scatters of flaked and ground stone artefacts, scarred trees 0.5–0.05 m B1–0.5 m 7 6 5 5 Clay and silt Clay, silt and fine sand Arid Arid B1.2–1 m 5 2–4 Clay, silt, some fine sand with a small component of coarse sand Pleistocene/Holocene 14,820770 Beta 81,376 5590760 Beta 81,375 19,2707320 Beta 44,374 28,7707300 Beta 81,377 32,9007510 Beta 81,378 28,7407340 ANUA 10011c 28,5907480 ANUA 13012c 32,0007550 ANUA 10319c B1.7–B1.2 m 4 1 Predominantly clay and silt, formation of peds, fine roots throughout. High Azolla counts plus aquatic taxa Semi-arid Chenopod shrublands and grasslands, scattered trees. Shallow marshy freshwater lake. Conditions becoming moister during this period B1.9–B1.7 m 3 Pre-human Semi-arid Chenopod shrubland/scattered trees, high energy depositional environment No archaeology detected, extinct and extant faunal species represented 2.5–B1.9 m 2 Pre-human Unknown Semi-arid 3.1–2.5 m 1 Pre-human Clay, silt, fine and coarse sands and gravel. Cemented lag deposit of bone and stone Mainly silt, clay and fine sands Fine sands decreasing with clay and silt increasing 33,3007530 Beta 81,380 30,9907360 Beta 81,381 32,5807510 Beta 81,382 29,5707280 Beta 46,171 35,40072800 ANUOD118a 28,7807350 ANUA 10012c 31,34071000 ANUA 12309c 32,4207460 Beta 81,383 29,1707360 Beta 81,383 Unknown Unknown Semi-arid Chenopod shrubland, extended high lake levels Casuarina forest, grass and herbaceous understorey, fluctuating lake levels No archaeology detected, extinct and extant faunal species represented No archaeology detected, extinct and extant faunal species represented a After Field and Dodson (1999) Dates listed in stratigraphic order (see Field and Dodson, 1999 for detail). c Samples prepared using the ABOX-SC technique (Fifield et al., in press). b Semi-arid Chenopod shrubland, extended dry lake conditions. Grasslands with scattered trees. Ephemeral lake conditions As for 1–0.5 m Disturbed deposits, flaked and ground stone artefacts, bone from modern and extinct faunal species Flaked and ground stone artefacts from all stages of manufacture. Ochre, charcoal pieces > 1 cm. Wide range of activites (e.g. wood and plant working, butchering). Fragmented bones of megafauna and extant animal species in situ, cf. campsite. Flaked stone artefacts in early stages of manufacture only, principally butchering implements. Complete and broken bones of megafauna in situ, some extant species present. Diprotodon, Genyornis, Sthenurus and Macropus titan represented J.H. Field et al. / Quaternary Science Reviews 21 (2002) 1023–1037 Depth from surface J.H. Field et al. / Quaternary Science Reviews 21 (2002) 1023–1037 Fig. 8. Pollen and charcoal diagram showing relative percentages of pollen versus depth for the Cuddie Springs deposits. The radiocarbon dates are listed on the left-hand side of the diagram. The asterisk on the right-hand side of the diagram indicates the level where stone artefacts first appear in the sequence. Where pollen values are too low to register on the diagram, a cross marks the levels where these taxa are present. (Illustration: Fiona Roberts). 1031 1032 J.H. Field et al. / Quaternary Science Reviews 21 (2002) 1023–1037 conditions, as indicated by the increase in Chenopodiaceae, persist throughout Zone 2. However, long term flooding of the lake was common. A slight decrease in Chenopodiaceae and an increase in Poaceae, Asteraceae and Cyperaceae at the top of Zone 2 suggests a change to less arid local conditions. 4.3. Zone 3.1.9–1.7 m; (?) >60,000–B35,000 BP Zone 3 represents the period just prior to the appearance of the first stone artefacts. Stratigraphically, it comprises a layer of coarse gravel overlain by a well sorted and concreted layer of bone and stone interpreted as a lag deposit and may represent hiatus in sediment accumulation. Zone 3 is likely to be a period of low deposition or erosion due either to deflation or fluvial activity. Some of the bone is slightly rounded and cemented into the silt and gravels indicating a hiatus is likely. Arboreal pollen types are rare, although values for Chenopodiaceae and Asteraceae are high. The ephemeral lake conditions persist with Cyperaceae and Myriophyllum still present, however Azolla is now virtually absent. The depositional environment is uncertain in Zone 3, however, the high percentage of gravel suggests that the deposit may have been part of an erosion channel infill. Throughout Zone 3, the vegetation around Cuddie Springs appears to have been a chenopod shrubland with scattered trees. Some herbaceous species are present and a number of aquatic taxa indicate a positive water balance. It is possible that there was deflation of some of the fine-grained sediments before Zone 3 was sealed and this zone is now capped by a band of ferruginised sands. There is weak iron staining with no evidence of an ironpan or lateritic crust. 4.4. Zone 4.B1.7–B1.2 m;B35,000 BP A gradual decrease in levels of Chenopodiaceae is seen through Zone 4. Poaceae and Asteraceae numbers increase with a broader range of other shrub and herb taxa present (i.e. Zygophyllaceae, Solanaceae, Malvaceae, Heterodendron, Brassicaceae and Apiaceae). Fewer trees are present and their representation in the pollen record may be masked by a greater local pollen influx. Azolla is found in greater numbers through Zone 4 indicating lake full conditions. This interpretation is also supported by the particle size analysis showing increases in silt and clay, with very little sand and gravel (Dodson et al., 1993). The deposit has numerous fine roots and plant matter and the clay breaks up into peds. Around 35,000 years ago, Cuddie Springs was a shallow and marshy freshwater lake in an arid environment, the catchment supported both chenopod shrublands and grasslands. The lower values of Casuarinaceae may be a function of a distant source. The presence of Acacia polyads indicates a local source (as opposed to monads which are possibly from more distance sources). Aquatic taxa are constantly represented and towards the top of the zone an increase in charcoal is observed. A gradual decrease in Chenopodiaceae and increase in grass, herbs and aquatic taxa suggest that conditions became slightly moister during this period. 4.5. Zone 5.B1.20–1.00 m;B31,000–19,000BP The period leading up to the LGM is represented by Zone 5 which is sealed at the upper boundary by a deflated surface comprising bone, stone and charcoal. The pollen evidence indicates that around 28,000 BP, there was a wetter period and the highest levels of aquatic taxa were found during this time. The ephemeral conditions returned following the more permanent nature of the previous zone. Casuarinaceae levels increased and Poaceae and Asteraceae and other herbs were present. Chenopodiaceae did not dominate the vegetation as seen following 28,000 BP and in the older deposits. The deflation pavement, which forms the upper boundary of Zone 5, represents a period of approximately 10,000 years. The concentration of stone and bone in the deflation surface suggests a cessation of sedimentation and/or sediment deflation (possibly by increased windiness). 4.6. Zone 6. 1–0.5 m; 19,000–(?) 10,000 BP Zone 6 covers a time period that includes the LGM. Recent excavations have revealed extensive disturbance through this horizon and inverted radiocarbon ages at the base suggest bioturbation or reworking of sediments may have taken place. The pollen record for this zone is therefore regarded as unreliable. However, in broad terms, the absence of Azolla massulae through this zone is an indication of the extended dry periods and along with the general Chenopodiaceae representation, is consistent with the LGM as described in Dodson and Wright (1989) and Dodson (1989). The Cuddie Springs site was well within the arid zone during the Last Glacial period with a dramatic change in the vegetation as an arid climatic regime dominated. A Holocene record is not thought to be represented in this deposit. 4.7. Zone 7. 0.5–0.05 m; the historic period No pollen preserved. J.H. Field et al. / Quaternary Science Reviews 21 (2002) 1023–1037 4.8. Zone 8. Surface: present The pollen spectra from the surface sample reflect the present day environment at Cuddie Springs, i.e., the immediate environs on the lake floor and possibly a component of the vegetation from the surrounding red soil plains. Some species, identified on the red soil plains, in particular Callitris sp. have not been identified in any of the sediment samples analysed. The Eucalyptus component of the vegetation is higher than at any other time in the pollen record with Casuarina levels similar to those observed around 30,000 BP (at approximately 1.1 m depth). The Azolla levels are very low when compared to pre-LGM levels, though the recent flooding of the lake in January 1995 resulted in an extended period of lake full conditions which may see a significant rise in the Azolla input to the pollen record. The relative concentrations of pollen represented in the surface sample from Cuddie Springs indicate that the source of pollen is predominantly from the immediate environs of the lake floor. The regional pollen taxa in the present day context are species such as Callitris sp. some species of Eucalyptus and Casuarina, which are found up to several kilometres from the claypan (see Dodson, 1983). 5. Fire history Three distinct peaks can be observed in the charcoal record at 2.5, 1.15 and 1 m depth. The lowest peak (at 2.5 m) occurs prior to the arrival of people and beyond the limits of radiocarbon dating. High charcoal levels correlate with an abrupt change in the local vegetation, i.e. a decrease in Casuarinaceae values and an increase in Chenopodiaceae. The peak in charcoal at this level is most likely to be the result of local fires (e.g. as in Clark, 1982) although a human cause cannot be eliminated until these levels have been investigated by archaeological excavation. 6. Discussion 6.1. Vegetation history The vegetation history at Cuddie Springs shows distinct changes and variation starting from the commencement of sedimentation of the lacustrine facies through to the LGM. There are four distinct phases in the pollen record. The first two phases correspond to the period prior to the arrival of people at Cuddie Springs and the onset of the final two phases begin when an archaeological record is present (see Table 1). 1033 An open woodland of Casuarina dominates the record in the opening stages. The understorey is composed mainly of Poaceae and Asteraceae species. Eucalyptus species are a minor component of the record with contributions from Acacia. Azolla counts indicate varying lake levels with intermittent drying of the lake floor. The high levels of Casuarina may be fringing vegetation and thus an indicator that Cuddie Springs was a floodplain depression. Callitris, often viewed as an indicator of dryness in vegetation histories, is noticeably absent from the Cuddie Springs record, possibly because of distance to source and its susceptibility to destruction (Dodson, 1979). The decline of Casuarina at the top of Zone 1 is coincident with a large peak in charcoal values. Casuarina has been described as a fire sensitive species and as such may have suffered a decline through burning rather than climatic change. The second distinct vegetation phase at Cuddie Springs (>35 ka) occurs prior to the arrival of people and is marked by the high levels of Chenopodiaceae, a gradual decline in Casuarina, constant levels of Poaceae and varying levels of Asteraceae. Some aquatic taxa are recorded in low frequencies through these levels with an increase in Cyperaceae following the decline of Azolla. The charcoal concentrations through Zone 2 are constant but very low, suggesting that no major fire events occurred during this time. The disconformity at 1.7 m forming the upper limit of Zone 2 is characterised as a high-energy depositional environment, possibly a stream channel with rapid deposition of sediments. The third broad phase of the Cuddie Springs vegetation history (B35–B28 ka) begins at about the same time as evidence of people are first detected at the site. The local vegetation suggests a semi-arid climatic regime at this time, signalled by a shift to open grasslands with scattered Eucalyptus, Acacia and Casuarina trees. Azolla values rise and are maintained through this zone by permanent inundation of the claypan, and possibly the larger lake floor on intermittent occasions as seen in the present day (Furby, 1995). Particle size analysis for the third phase reveals a marked increase in fine silts and clays with the absence of gravels and coarse sands. Sediments indicate marshy conditions with the formation of peds and the abundance of fine roots in the profile. The formation of peds may mean seasonal or occasional drying out of the claypan. The presence of Nitella oogonia, Azolla massulae, and Chydorid Cladocera carapace, head shields and fragments combine to support an interpretation of still, shallow freshwater conditions and incomplete drying of the lake (Ralph Ogden, ANU, pers. comm., 1994). Rapid deposition of sediment may indicate frequent storm events and/or possibly continuous filling of the lake from a sparsely vegetated 1034 J.H. Field et al. / Quaternary Science Reviews 21 (2002) 1023–1037 catchment. Fossil remains of Grus rubicundus (brolga) have been also identified from these levels (Furby, 1995). Brolga are now only seen when the claypan has been flooded, as are many other species of water birds (e.g. pelicans, ducks and spoonbills). The compacted stone pavement interpreted as a deflation surface marks the upper limit of this phase. The eventual drying of the lake and development of high wind activity, also seen in the aeolian dune activity in central Australia (Wasson, 1989), may have led to the loss of sediments on the Cuddie lake floor and the construction of the source bordering dune which is found on the north-east margin. The period represented by the deflation surface does not necessarily indicate a cessation of sedimentation but extended dry periods and high wind activity or strong variability between high and low water conditions that resulted in a time averaging of the record. Beach sands contributed to the lunette at lake full conditions plus the deflation of sediments off a contracted lake floor at low lake levels. The fourth phase (B19–B6ka) can only be described in the very broadest terms because of the extensive disturbance which has been identified through archaeological excavation and evidence in the apparently homogeneous pollen spectra. Sometime around 19,000 BP, sedimentation recommenced with periodic inundation of the claypan. The homogeneous lake muds deposited since 19,000 BP are in contrast to the stratigraphic profile developed after the ephemeral flooding events recorded in the earlier zones. The high levels of Chenopodiaceae are consistent with extended drying of the claypan surface. The absence of Azolla spores supports this conclusion. There is no intact Holocene record at the centre of the claypan that may be a function of the depositional environment during the Holocene with warmer climates leading to greater evaporation and wetting and drying of the upper sediments (precluding pollen preservation). The present day Cuddie Springs is marked by the irregular filling and drying of the claypan. There are also periodic droughts and relatively high winds that may have prevented a constant sedimentation regime following the LGM. The Cuddie Springs record shows a shift from Casuarina forest to Chenopodiaceae in the early phases of the record, with a shift to grasslands and then marked aridity around the time of the LGM. Modern pollen spectra show a shift to semi-arid vegetation, probably developing sometime during the Holocene. The absence of a Holocene record makes the timing of this change impossible to define. In the present day setting, Cuddie Springs is an ephemeral lake that irregularly fills with water. A recent flooding event of January/February, 1995 is the first historic recorded filling of the ancient lake floor (see Field and Dodson, 1999). Initially, Cuddie Springs was part of a riverine floodplain, probably joined to the Darling River system and later followed by a number of short-lived lacustrine phases. 6.2. Stratigraphic associations for the human/megafauna overlap The combined records for pollen, charcoal, lake history and archaeology, provide a picture of sequentially deposited horizons within a sealed stratigraphic unit (Zones 4 and 5), albeit formed in a relatively short time frame (ca.7000 years) (Field and Dodson, 1999). This interpretation is supported by the composition of stone tool and fossil bone assemblages that exhibit distinct changes through time corresponding to phases in lake hydrology (see Fullagar and Field, 1997; Field and Dodson, 1999; Table 1). The observed changes in vegetation through Zone 5 directly correlate to climatic and vegetation changes in time frames established for other sites from across the continent (Bowler et al., 1976; Dodson, 1989; Dodson and Wright, 1989; Kershaw et al., 1991). The age of the Cuddie Springs deposit below 1.7 m depth is beyond the limits of radiocarbon. The first archaeological evidence of human occupation at Cuddie Springs is not accompanied by any significant increases in microscopic charcoal and it is not until higher in the sequence that peaks of charcoal are observed. These peaks coincide with an apparent intense period of occupation of the lake floor as supported by the density of the archaeological evidence. The peak in microscopic charcoal at 1.15 m in Zone 5 is associated with an increase in large pieces of charcoalFrecovered during archaeological excavation, >1 cm in maximum dimension, angular and shows little abrasion or rounding. The peak in concentration of charcoal in Zone 5 (equivalent to archaeological levels 2, 3 and 4) is followed by a marked decrease at its upper boundary. The charcoal increases in Zone 5 correlate with a decrease in Poaceae (grasses), Casuarinaceae and aquatic taxa. The development of chenopod shrublands at the expense of grasses and some tree taxa may have been accelerated by local burning during a period of developing aridity. The persistence of charcoal through Zone 6 during maximum dry conditions, in conjunction with the presence of a range of stone artefacts, suggests that Cuddie Springs may not have been abandoned during the LGM, as has been observed in other arid zone sites (e.g. O’Connor et al., 1993). 6.3. The implications for the archaeological and megafaunal records The two main explanatory models for the Pleistocene faunal extinctions revolve around the effects of climatic J.H. Field et al. / Quaternary Science Reviews 21 (2002) 1023–1037 change and the timing of the arrival, and subsequent activities of people (Dodson et al., 1988; Flannery, 1990; Horton, 2000). For the arid zone, the combination of factors may have been very different from those acting in the more temperate coastal zones. Climate may well have been the driving force behind the arid zone extinctions. As Horton (1984) argues, the loss of free water as a result of rainfall depression during the lead up to the LGM may have had irreversible effects on the large animals such as the Diprotodon. Not only could it not migrate to betterwatered areas, but the shift from shrublands to grasslands at Cuddie Springs during this period would have selected against it in favour of grazers such as the red kangaroo (Macropus rufus) (see Dodson, 1989). The disappearance of the large bird Genyornis newtoni from the arid zone around 50,000 BP (Miller et al., 1999) may be another example of an animal ill-equipped to deal with climate change and reconfiguration of habitats. The Emu (Dromaius novaehollandiae) persisted in these habitats through the extinction period documented in this study. Elements of Diprotodon and Genyornis make up a significant proportion of the bone assemblage in the early archaeological levels (Field and Dodson, 1999; Zone 4). The animals appear to have died at a waterhole: by predation from humans, perishing during a local drought or a combination of both. Local conditions were certainly favourable for browsers during this time. However, around this time there is a shift from chenopod shrublands to grasslands (Zones 4–5), and in the archaeological record the introduction of seed-grinding stones (Fullagar and Field, 1997; Field and Fullagar, 1998). The appearance of these specialised technologies implies an environment of uncertainty developed. Seed-grinding stones are only found in arid/semi-arid zone contexts and the exploitation of seeds is viewed as a resource choice when people wished to maintain a presence in a resource depleted environment (Tindale, 1977). By the time the deflation pavement formed at the upper limit of the megafauna/human overlap, the megafauna had all but disappeared. The environment was more arid and the lake had entered an extended dry period. Fire does not seem to have played a major environmental role in the archaeological levels at Cuddie Springs, and through this period peaks in charcoal are correlated to camp-fires with people camping around a possible soak, or digging for water. The apparent broad subsistence base of the first inhabitants and the increasing aridity associated with the lead up to the LGM would see the latter as the driving force in the localised events at Cuddie Springs around 30,000 BP. 1035 The Cuddie Springs vegetation history has confirmed the trends in general aridity and temperature for latitudes of 321S in eastern Australia observed for the Ulungra Springs record further east (Dodson and Wright, 1989). It also demonstrates that fire histories are not necessarily accurate indicators of a human presence on the landscape and finally, it shows that climate change is certainly a significant factor in arid zone faunal extinctions during the Late Pleistocene. Acknowledgements Funding for the project was provided by the University of New South Wales, Australian Geographic, National Geographic and an ARC Project Grant. Field was the recipient of an Australian Postgraduate Award. We wish to thank Chris Myers, the Johnstone family, Jan Currey, Doug Green, the Brewarrina Local Aboriginal Land Council, the Walgett Shire Council and numerous volunteers for help with the project. Megan Mebberson and Fiona Roberts are thanked for producing the line diagrams. Thanks are also due to Eric Colhoun, Jim Rose and an anonymous referee for constructive comments. References Abbott, W.E., 1881. Notes of a journey on the Darling. Proceedings of the Royal Society of NSW 15, 41–70. Anderson, C., Fletcher, H.O., 1934. The Cuddie Springs bone bed. The Australian Museum Magazine 5, 152–158. Aston, H.I., 1973. Aquatic Plants of Australia. Melbourne University Press, Melbourne. Bell, C.J.E., Finlayson, B.L., Kershaw, A.P., 1989. Pollen analysis and dynamics of a peat deposit in Carnarvon National Park, central Queensland. Australian Journal of Ecology 14, 449–456. Bird, M.I., Ayliffe, L.K., Fifield, K., Cresswell, R., Turney, C., 1999. Radiocarbon dating of ‘old’ charcoal using a wet oxidationF stepped combustion procedure. Radiocarbon 41, 127–140. Bowdery, D., 1998. Phytolith analysis applied to Pleistocene– Holocence archaeological sites in the Australian arid zone BAR International Series 695. Bowler, J.M., Hope, G.S., Jennings, J.N., Singh, G., Walker, D., 1976. Late Quaternary climates of Australia and New Guinea. Quaternary Research 6, 359–394. Boyd, W.E., 1990. Quaternary pollen analysis in the arid zone of Australia: Dalhousie Springs, Central Australia. Review of Palaeobotany and Palynology 64, 331–341. Clark, R.L., 1982. Point count estimation of charcoal in pollen preparations and thin sections of sediment. Pollen et Spores 24, 523–535. Colhoun, E.A., van de Geer, G., 1988. Darwin Crater, the King and Linda Valleys. In: Cainozoic Vegetation of Tasmania, Handbook prepared for the Seventh International Palynological Congress, 28th August–3rd September, 1988. Department of Geography, University of Newcastle, pp. 30–71. Cunningham, G.M., Mulham, W.E., Milthorpe, P.L., Leigh, J.H., 1992. Plants of Western New South Wales. Inkata Press, Sydney. 1036 J.H. Field et al. / Quaternary Science Reviews 21 (2002) 1023–1037 Dodson, J.R., 1979. Late Pleistocene vegetation and environments near Lake Bullenmerri, Western Victoria. Australian Journal of Ecology 4, 419–427. Dodson, J.R., 1983. Modern pollen rain in south-eastern New South Wales, Australia. Review of Palaeobotany and Palynology 38, 249–268. Dodson, J., 1989. Late Pleistocene vegetation and environmental shifts in Australia and their bearing on faunal extinctions. Journal of Archaeological Science 16, 207–217. Dodson, J.R., Wright, R.V.S., 1989. Humid to arid to subhumid vegetation shift on Pilliga sandstone, Ulungra Springs, New South Wales. Quaternary Research 32, 182–192. Dodson, J.R., Enright, N.J., McLean, R.F., 1988. A late Quaternary vegetation history for far northern New Zealand. Journal of Biogeography 15, 647–656. Dodson, J.R., Fullagar, R., Furby, J., Jones, R., Prosser, I., 1993. Humans and megafauna in a late Pleistocene environment from Cuddie Springs, north western new south Wales. Archaeology in Oceania 28, 94–99. Dortch, C., 1984. Devil’s Lair: a study in prehistory. Western Australian Museum, Perth. Faegri, K., Iversen, J., 1975. Textbook of Pollen Analysis, 3rd Revised Edition. Blackwell Scientific, Oxford. Field, J., Boles, W., 1998. Genyornis newtoni and Dromaius novaehollandiae at 30,000 B.P. in central northern New South Wales. Alcheringa 22, 177–188. Field, J., Dodson, J., 1999. Late Pleistocene megafauna and archaeology from Cuddie Springs, southeastern Australia. Proceedings of the Prehistoric Society 65, 275–301. Field, J., Fullagar, R., 1998. Grinding and pounding stones from Cuddie Springs and Jinmium. In: Fullagar, R. (Ed.), A Closer Look: Recent Australian Studies of stone Tools. Sydney University Archaeological Methods Series 6. ISBN 1 86451, pp. 365–369. Fifield, L.K., Bird, M.I., Turney, C.S.M., Hausladen, P.A., Santos, G.M., di Tada, M.L., in press. Radiocarbon dating of the human occupation of Australia prior to 40 ka BPFsuccesses and pitfalls. Radiocarbon, in press. Flannery, T., 1990. Pleistocene faunal loss: implications of the aftershock for Australia’s past and future. Archaeology in Oceania 25, 45–67. Foster, C.B., Harris, W.K., 1981. Azolla capricornica sp. nov. First Tertiary record of Azolla Lamarck (Salviniaceae) in Australia. Transactions of the Royal Society of South Australia 105, 195–204. Fullagar, R., Field, J., 1997. Pleistocene seed-grinding implements from the Australian arid zone. Antiquity 71, 300–307. Furby, J.H., 1991. A preliminary study of Late Pleistocene megafauna, humans and environment at Cuddie Springs, north-western New South Wales. Unpublished B.A. (Hons) Thesis, School of Geography, UNSW. Furby, J.H., 1995. Megafauna under the microscope, archaeology and palaeoenvironment at Cuddie Springs. Unpublished Ph.D. Thesis, School of Geography, UNSW. Furby, J.H., Fullagar, R., Dodson, J.R., Prosser, I., 1993. The Cuddie Springs bone bed revisited, 1991. In: Smith, M.A., Spriggs, M., Frankhauser, B. (Eds.), Sahul in Review: Pleistocene Archaeology in Australia, New Guinea and Island Melanesia. Department of Prehistory, Research School of Pacific Studies. ANU, Canberra, pp. 204–210. Graham, R.W., Lundelius Jr., E.L., 1984. Coevolutionary disequilibrium and Pleistocene extinctions. In: Martin P.S., Klein R.G. (Eds.), Quaternary Extinctions: A Prehistoric Revolution. The University of Arizona Press, Tucson, Arizona, pp. 223–249. Grimm, E.C., 1991. Tilia Graph, Version 1.19. Research and Collection Section, Illinois State Museum, Illinois. Habermahl, M.A., 1980. The Great Artesian Basin, Australia. BMR Journal of Australian Geology and Geophysics 5, 9–38. Hiscock, P., Attenbrow, V., 1996. Backed into a corner. Australian Archaeology 42, 64–65. Horton, D.R., 1984. Red Kangaroos: last of the Australian megafauna. In: Martin, P.S., Klein, R.G. (Eds.), Quaternary Extinctions: A Prehistoric Revolution. The University of Arizona Press, Tucson, pp. 639–680. Horton, D.R., 2000. The pure state of nature. Sacred cows, destructive myths and the environment. Allen & Unwin, Sydney. Kershaw, A.K., 1985. An extended late Quaternary vegetation record from north-eastern Queensland during the last two glacial/interglacial cycles. Proceedings of the Ecological Society of Australia 13, 179–189. Kershaw, A.P., D’Costa, D.M., McEwen Mason, J.C.R., Wagstaff, B.E., 1991. Palynological evidence for Quaternary vegetation and environments of mainland south-eastern Australia. Quaternary Science Reviews 10, 391–404. Martin, H.A., 1973. Palynology and historical ecology of some cave excavations in the Australian Nullarbor. Australian Journal of Botany 21, 283–316. Meakin, S., 1991. The Cuddie Springs Pleistocene megafauna locality: site description and drilling results. Unpublished Report, The Geological Survey of New South Wales Department of Minerals and Energy. GS Report No.: GS1991/034. Miller, G., Magee, J.W., Johnson, B.J., Fogel, M.L., Spooner, N.A., McCulloch, M.T., Ayliffe, L.K., 1999. Pleistocene extinction of Genyornis newtoni: Human impact on Australian megafauna. Science 283, 205–208. Moore, P.D., Webb, J.A., Collinson, M.E., 1991. Pollen Analysis, Second Edition. Blackwell Scientfic, Oxford. O’Connell, J.F., Allen, J., 1998. When did humans first arrive in Greater Australia and why is it important to know? Evolutionary Anthropology 6, 132–146. O’Connor, S., Veth, P., Hubbard, N., 1993. Changing interpretation of postglacial human subsistence and demography in Sahul. In: Smith, M.A., Spriggs, M., Fankhauser, B. (Eds.), Sahul in Review: Pleistocene Archaeology in Australia, New Guinea and Island Melanesia. Department of Prehistory, Research School of Pacific Studies. ANU, Canberra, pp. 95–105. Pickett, E.J., 1997. The Late Pleistocene and Holocene vegetation history of three lacustrine sequences from the Swan Coastal Plain, south-western Australia. Unpublished Ph.D. Thesis, University of Western Australia. Singh, G., Geissler, E.A., 1985. Late Cainozoic history of vegetation, fire, lake levels and climate at Lake George, New South Wales. Philosophical Transactions of the Royal Society of London, Series B 311, 379–447. Singh, G., Luly, J., 1991. Changes in vegetation and seasonal climate since the last full glacial at Lake Frome, South Australia. Palaeogeography, Palaeoclimatology, Palaeoecology 84, 75–86. Smith, M.A., Vellen, L., Pask, J., 1995. Vegetation history from archaeological charcoals in central Australia: the late Quaternary record from Puritjarra rock shelter. Vegetation History and Archaeobotany 4, 171–177. Tindale, N.B., 1977. Adaptive significance of the Panara or grass seed culture of Australia. In: Wright, R.V.S. (Ed.), Stone Tools as Cultural Markers: Change, Evolution and Complexity. Humanities Press, New Jersey, pp. 345–349. Vrba, E.S., 1981. The significance of bovid remains as indicators of environment and predation patterns. In: Behrensmeyer, A.K., Hill, A.P. (Eds.), Fossils in the Making: Vertebrate Taphonomy and Paleoecology. The University of Chicago Press, Chicago, pp. 247–271. Wasson, R.J., 1989. Desert dune building, dust raising and palaeoclimate in the Southern Hemisphere during the last 280,000 years. J.H. Field et al. / Quaternary Science Reviews 21 (2002) 1023–1037 In: Donnelly, T.H., Wasson, R.J. (Eds.), CLIMANZ 3, CSIRO Division of Water Resources, Canberra, pp. 123–137. Watson, A., 1989. Windflow characteristics and aeolian entrainment. In: Thomas, D.S.G. (Ed.), Arid Zone Geomorphology. Belhaven Press, London, pp. 209–231. 1037 Wilkinson, C.S., 1885. President’s Address, Annual General Meeting. Proceedings of the Linnean Society of NSW 9, 1207–1241. Wyrwoll, K.H., McKenzie, N.L., Pederson, B.J., Tapley, I.J., 1986. The Great Sandy Desert of north-western Australia: the last 7000 years. Search 17, 208–210.