Legacy archaeology: Aboriginal subsistence response to Holocene environmental changes using faunal evidence from archaeological sites on the Lower Murray,South Australia

Abstract

Ngaut Ngaut (Devon Downs) and Tungawa (Fromm’s Landing) 2 and 6 are located in the Gorge Section of the Lower Murray River. They were excavated more than 60 years ago. Unusually, they preserved fauna over the 6000 or 7000 years of occupation. Assessing this record, it is concluded that Aboriginal agents were responsible for the middens in these rockshelters. Following this, Ngaut Ngaut and the Tungawa sites are compared in terms of their dating, stratigraphy and changes in the fauna through time. While the majority of species are present throughout at all three sites, there are shifts in the number of animals in concert with Holocene environmental changes. After 3000 BP, the trend is to increased attention being given to resources from the riparian and river zones and away from the dryland Murray Plains. An increase in shellfish and the presence of crayfish gastroliths support this contention. Nearby Tartanga Island provides a record of Holocene sedimentary changes in the Murray River associated with altered sea level and flood regimes, particularly the deposition of the Monoman and Coonambidgal formations. The latter creating a landscape of highly productive swamps and backwaters. The information from these legacy excavations supports the conclusion that a shift in the locus of Aboriginal hunting and gathering activities accompanied mid- and late-Holocene environmental changes on the Lower Murray River.

Keywords

aboriginal rockshelters hunting and gathering legacy faunal collections Lower Murray River mid-Holocene changes zooarchaeology

Introduction

Ngaut Ngaut (Devon Downs) and Tartanga Island were excavated 90 years ago and Tungawa (Fromm’s Landing 2 and 6) 60 years ago (Hale and Tindale, 1930; Mulvaney, 1960; Mulvaney et al., 1964). Over that time, excavation and post-excavation methods have been transformed. This raises the question as to whether reports from these excavations contain any useful archaeological evidence. Here we consider the archaeozoological data from these sites in terms of Aboriginal subsistence activities and whether these respond to mid-Holocene environmental changes. To do this, the reliability of the excavations and faunal information will be appraised followed by a summary of the Holocene environmental evidence available. The faunal data, as indicative of Aboriginal activities, will then be assessed against environmental changes taking place through the mid- to late-Holocene. This work is part of Judith Littleton’s Royal Society of New Zealand Marsden Fund Grant (14-UOA-19) to study the human remains from the Roonka site and is being undertaken in collaboration with and permission of the River Murray and Mallee Aboriginal Corporation Inc (RMAC) and the Ngaiawang, the Traditional Owners of Roonka (Littleton et al., 2017).

The relevance of questions asked here is that elsewhere changes in the archaeological record have been linked with Holocene climatic shifts (Smith et al., 2008; Williams et al., 2008, 2015). Williams et al. (2015) argue that Aboriginal arid zone populations expanded in the mid-Holocene (5–4 kya cal BP) followed by a contraction associated with the onset of drier, more variable conditions. This was followed by a further expansion in the past 2000 years. In addition, they suggest that, with the onset of variable conditions, Aboriginal communities responded by adopting more efficient technologies allowing them to forage in less favourable patches and collect lower ranked foods. Similarly, Asmussen and McInnes (2013) claim that Aboriginal foragers, faced with a mid-Holocene climatic deterioration adopted new technologies to exploit previously unused environments and made greater use of foods requiring costly preparation. Hill et al. (2016) studying limb bones from Roonka suggest that females expanded their foraging radii to compensate for ENSO affected climate changes (increased aridity and greater climatic fluctuations) post 4000 BP. Finally, and in contrast, Dortch et al. (2012), analysing faunal remains from cave deposits in a Mediterranean-type environment (southwestern Western Australia) found little change in Aboriginal hunting of high ranked prey species through the Holocene.

Using legacy archaeological data to test these questions has its limitations. Re-analyses of the faunal material from these sites in the collections of the South Australian Museum, or new excavations, are clearly warranted. Until that time, however, synthesising the faunal information from these sites is a significant first step, generating new questions and reiterating the scientific importance of these collections, an importance that sits alongside their cultural significance to people of the River Murray and Mallee region.

Tartanga Island, Ngaut Ngaut and Tungawa 2 and 6 are located in the Lower Murray Gorge, a 280 km stretch of the Murray River between Overland Corner and Mannum (Figure 1). The Gorge formed when Pleistocene sea levels were lower and the river cut through limestones to a depth of c 100 m, (Twidale et al., 1978). As sea levels rose during the early Holocene, the Murray filled this trench and it now flows on its floodplain within a valley 3–4 km wide and c 30 m below the surrounding plains. The plains surrounding this part of the Lower Murray are semi-arid sandy dunes covered with Mallee woodland and an understory of shrubs and grasses.

Figure 1.

Location map of the Lower Murray River, South Australia, showing archaeological sites and places mentioned in the text.

Ngaut Ngaut and Tungawa 2 and 6 are rockshelters in cliffs on the east and western side of the river respectively. Tartanga is a stratified open site on a channel island, immediately to the north of Ngaut Ngaut. Apart from the original excavation reports, there have been a number of more recent studies of these sites. Fusco (2014) made use of the data from Tungawa 2 in her survey of the terrestrial mammal faunas of the Murray Mallee. In addition, Smith (1978, 1982) reanalysed Ngaut Ngaut and Roberts (1998) reanalysed Tungawa 2, the latter focussing on stone artefacts and the potential for isotope study (Bland et al., 2012; Roberts et al., 1999). Ngaut Ngaut is the subject of collaboration between Mannum Aboriginal Community Association Inc (MACAI) and Flinders University (Bland et al., 2012; Roberts, 2012). Within the Gorge, there have also been excavations at McBean Pound and at Roonka, together with reanalysis of the skeletal material and stratigraphy from Roonka using collections archived at the South Australian Museum (Littleton et al., 2017; Paton, 1983; Pretty, 1977; Walshe, 2009). In addition, there have been archaeological surveys and excavations of Aboriginal middens and cooking mounds to the north of the Gorge and of middens to its south (Jones et al., 2017, 2022; Thredgold and Roberts, 2017; Westell and Wood, 2014; Westell et al., 2020; Wilson, 2017; Wilson et al., 2012).

Assessing faunal information from Gorge archaeological sites

Tartanga Island is a stratified open site on a channel island, originally it was thought to have been part of an eroded western bank of the river. The site consists of a lower (A–E) and an upper series of Beds (F–I). A radiocarbon date from Bed C (Broecker et al., 1956, L-271E 6030 ± 120, see Table 1 for cal. BP) places this level of the site at the transition between the Monoman and Coonambidgal Formations (see discussion below, Hale and Tindale, 1930: 149–158; Twidale et al., 1978: 40). Materials from the Lower Beds were carbonate-coated and stained in contrast to those from the Upper Beds which appeared fresh. Faunal and artefactual material was surface collected, sieved from sands (sieve size not stated) or came from a narrow 40 m trench across the site (Hale and Tindale, 1930: 154).

Table 1.

C14 dates and cal BP dates for Tungawa 2 and 6, Ngaut Ngaut and Tartanga Island from Mulvaney (1960), Mulvaney et al. (1964), Smith (1982), Broecker et al. (1956). Calibrations https://c14.arch.ox.ac.uk/oxcal.html.

Tungawa 2
Layer	Lab no	C14 years BP	Cal. BP 95.4%
Layer 0–3	None submitted	Unknown
Layer 4	NZ-365, R 456/2	3240 ± 80	3255–3293 (2.7)
			3327–3644 (91.6)
			3666–3684 (1.1)
Layer 6	P308	3756 ± 85	4407 (95.4)
Layer 8	P309	3881 ± 85	4032 (1.1)
Layer 8	P309	3881 ± 85	4527 (94.3)
Layer 9	P311	4055 ± 85	4299–4326 (1.6)
			4353–4369 (0.9)
			4386–4830 (92.9)
Layer 10	NZ-364, R 456/1	4850 ± 100	5321–5420 (10.9)
			5483–5759 (80.2)
			5822–5885 (4.2)
Tungawa 6
Layers 1–7	None submitted	Unknown
Layer 8	NPL 28	2950 ± 91	3354 (95.4)
Layer 11	NPL 29	3170 ± 94	3158–3614 (95.2)
Layer 11	NPL 29	3170 ± 94	3626–3630 (0.2)
Layer 16	NPL 63	3450 ± 90	3480–3538 (4.4)
			3544–3927 (90.3)
			3949–3961 (0.7)
Ngaut Ngaut
Layers I–IV	None submitted	Unknown
Layer V	GAK 1021	2980 ± 90	2892–2904 (0.8)
Layer V	GAK 1021	2980 ± 90	2923–3375 (94.6)
Layer VII	GAK 1022	3460 ± 100	3476–3980 (95.4)
Layer VIII	GAK 1023	4360 ± 100	4651–4670 (0.8)
			4702–4758 (3.1)
			4808 to −5306 (91.5)
Layer IX	L 217G	4290 ± 140	4451–4453 (0.5)
Layer IX	L 217G	4290 ± 140	4519–5300 (94.9)
Layer XI	GAK 1024	5180 ± 100	5664–5674 (0.5)
Layer XI	GAK 1024	5180 ± 100	5711–6208 (94.9)
Tartanga Island
Bed C	L-271E	6030 ± 120	7208–7246 (1.6)
			6626–7170 (93.4)
			6569–6581 (0.4)

At Ngaut Ngaut, Tindale excavated a narrow trench in spits, followed by a square c 3 m × 3.5 m excavated through 12 layers to a depth of 6.2 m. Layers consisted of sand, ash, hearths, charcoal and shells. These were continuous across the site, except where interrupted by pits or rockfall. Material was passed through 4 mm sieves, some wet sieved (Hale and Tindale, 1930: 173–177). Smith (1982) estimated that less than 10% of the shelter within the dripline was excavated. Assessing Tindale’s excavation at Ngaut Ngaut, Smith (2000) found a high standard of excavation, one, he argued, that would stand against contemporary measures

A trench 10 m long and between 2 and 4 m wide was excavated across Tungawa 2 to a depth of 6 m. The uppermost metre, disturbed by rabbit burrows, was excavated in 10 cm spits and the remainder in terms of discernible stratigraphic differences. Material was passed through a bank of sieves (4 and 6 mm) and wet sieved where possible (Mulvaney, 1960: 59). Stratigraphic layers were continuous across the site sloping towards the shelter entrance. An estimated 20% of the site was excavated. Layer 3 at Tungawa 2 consisted of a thick band of yellow sand with little archaeological material, which Mulvaney (1960: 65) suggests showed less frequent and less intense occupation than any other period in the site’s history.

Tungawa 6 was excavated in two seasons ending with a trench 8.5 m long, 3 m wide and 3.5 m deep (6 m deep in terms of slope). Eighteen layers were distinguished. The lower sediments consisted of rockfall and fine to medium yellow sands, the upper series comprised mostly occupational debris, hearths and shell. Mulvaney et al. (1964: 483–485) note that Layers 6, 10, 13, 16, 17 and 18 were mostly fine loose yellow sand, interpreted as periods when there was low or no occupation, either at this part of the site, or at the site itself. Mulvaney et al. (1964: 481) noted that all material was sieved. It is presumed that methods used at Tungawa 2 were followed at Tungawa 6.

In addition to humans, bone accumulations may be the product of other predators or result from pitfalls. Fusco (2014: 82–83) compared sites, including Tungawa 2 (FL2), with different agents of accumulation: raptors, pitfalls and humans, concluding the higher species richness and bodyweight measures at FL2 were the result of human predation. Comparing Tungawa 2 with the Mypolonga 2 site (located between Mannum and Murray Bridge), Fusco (2014: 71) noted a relative absence of smaller animals at Tungawa 2, which she put down to a combination of Aboriginal hunting preference and the use of large sieve sizes.

Tasmanian devils and dingoes are potential agents of bone accumulation. Remains from these carnivores are highly fragmented, to the extent that few identifiable bones, beyond tooth fragments, are left (Marshall and Cosgrove, 1990: 110; Walshe, 2000, 2014: 277–278). The faunal remains from Ngaut Ngaut and Tungawa 2 and 6 are relatively well preserved and consist of identifiable bones from a diverse range of species excavated from within layers of shell with distinct hearths and stone and bone artefacts. On this basis, Smith (1982: 112), who reanalysed Ngaut Ngaut, concluded the majority of animals present at Tartanga Island, Ngaut Ngaut and the Tungawa sites were brought there by human agents.

As well as the factor of differential preservation, faunal remains recovered from an archaeological investigation are a sample of what was available, what people chose to catch, what they brought back and what has survived the processes of butchery, cooking and disposal. In addition, processes other than, or additional to, food acquisition might be operating (Russell, 2011).

Aboriginal hunters might preferentially pursue large game, or large quantities of small game, in order to make food gifts to in-laws or seniors or to gain prestige, a form of costly signalling (Altman and Peterson, 1988; Bliege Bird et al., 2001; Conolly, 2017; Sackett, 1979). Between 1841 and 1845, Edward Eyre was resident at Moorundie 45 km to the north of Ngaut Ngaut. He documented variations in Aboriginal seasonal hunting and gathering activities, individual and cooperative work groups and activities organised in terms of gender and age together with multiple food restrictions in terms of gender, age and context, with larger animals being shared in a manner that satisfied social claims (Eyre, 1845: 252–259, 283–288, 293–294). Wilson et al. (2022) discusses faunal material from these archaeological sites from an Indigenous Ngati perspective, animals that are in a special relationship with individuals or clans. While some of the processes discussed here might be visible in the record, others will not. The measures of consistency and plausibility, allied with the identification of similarities and differences within and between sites, should allow some insights into the processes operating.

Specialists identified the fauna (Hale and Tindale, 1930: 217; Mulvaney, 1960; Mulvaney et al., 1964). Fusco (2014: 33) accepted identifications for Tungawa 2, except where there was revised nomenclature (e.g. Macropus canguru = Macropus fuliginosis). As noted, Ngaut Ngaut was re-examined by Smith (1982) with identifications to species, genus or to general groups, for example ‘Small wallaby’, where fauna could not be identified further. As far as can be ascertained, mammalian identifications were on mandibles/teeth, except for a few specimens at Tungawa 6, where limb bones were used (Wakefield in Mulvaney et al., 1964: 494–496). Counts, where made, were in terms of minimum number of individuals (MNI). Information about the number of identified specimens (NISP) and the identification of post-cranial skeletal material is not available for these sites (cf. Mein and Manne, 2021).

To have full confidence in the faunal materials from Ngaut Ngaut and Tungawa 2 and 6 requires that they were excavated and analysed to modern standards. While not reaching this level of certainty, the excavation, recovery and analysis of these sites were sufficient to allow us to generate hypotheses capable of being tested if and when further information becomes available. Tartanga Island, an open stratified site, was not excavated with the same precision as the shelter sites. As a result, Tartanga Island will be discussed separately to the other sites.

Holocene environmental changes in south eastern Australia

A general consensus regarding Holocene climatic variations in south-eastern Australia has emerged over the past 20 years based on the study of proxy measures; windblown dust, river alluvium, river flows, lake levels, dune mobility, spelaeothems and pollen studies which taken together indicate changes in temperature and effective precipitation through the Holocene (see Walker et al., 2018). The early Holocene (c. 11.7–8.2 kya BP) was a period when the climate was recovering from the Last Glacial Maximum with rising temperatures, more effective precipitation, and, in southern Australia, enhanced westerly wind flows (Gliganic et al., 2014; Quigley et al., 2010: 1101; Shulmeister, 1999; Shulmeister et al., 2016: 1442). Modifying this picture, Fletcher and Moreno (2012) suggest multimillennial scale zonal shifts in SWW-driven wetness which created differences between eastern and western SE Australia in the early to mid- Holocene. Barr et al. (2019), Kemp et al. (2012: 71) and Quigley et al. (2010: 1102) argue for a peak in effective precipitation and warm temperatures 8.2–4.2 kya, (the mid-Holocene climatic optimum) with a decline beginning about 5 kya, the onset of an ENSO-dominated climate. The period from c 4.2 ka BP was one of increased variability, sometimes occurring as decadal and centennial length droughts (Gingele et al., 2007; Gliganic et al., 2014; Shulmeister, 1999; Wilkins et al., 2013: 294). Barr et al. (2019: 5) describe the period from the mid-Holocene as a transition from predominantly high precipitation and low frequency variability to a drier climate with enhanced variability.

The last 1500 years in Southeastern Australia show a decline in ENSO frequency and a return to conditions similar to those of today. This involved increased winter rainfalls, rising lake levels in western and southwestern Victoria and the initiation of peat formation in the Mount Lofty Ranges (Buckman et al., 2009: 1027; Kemp et al., 2012: 72; Marx et al., 2011: 311; Wilkins et al., 2013: 793; Williams et al., 2008: 256). Analysing stable carbon and nitrogen isotopes from faunal material from Tungawa 2, Roberts et al. (1999: 49) concluded there had been an arid period on the Lower Murray between 4050 and 3080 BP, with the period after 3080 BP being consistent with a climate similar to that of the present.

At the Last Glacial Maximum (LGM) sea levels around the Australian coast were approximately 120 m below present mean sea level (Yokoyama et al., 2001). As noted previously, the lower sea level caused the Murray River to cut a Gorge through limestones to a depth of about 65 m below the present valley floor, stripping its fill (Twidale et al., 1978: 39). Sea levels rose rapidly from c. 17,000 BP reaching present mean sea level at c.7000 BP (Belperio et al., 2002; Lewis et al., 2013: 123–125). The rise in sea levels caused the Murray to aggrade depositing coarse-grained quartz sands (Monoman Formation), dated near its top to c 7200 BP (Sloss et al., 2007; Twidale et al., 1978: 40).

There is considerable debate concerning sea levels and the status of the Lower Murray River through the Holocene (e.g. Helfensdorfer et al., 2020). Tibby et al. (2021) and Bourman et al. (2022), however, argue that when neotectonism and glacio-hydroisostatic adjustments are taken into account, the mid-Holocene sea level at the Murray mouth did not exceed 1 m AHD and that the Lower Murray remained fresh to brackish throughout. Between c. 7000 and 3000 BP, the river deposited fluvial clays, silts and fine grained sands of the Coonambidgal Formation over the top of Monoman Formation (Twidale et al., 1978: 40), with the upper layers of the Coonambidgal Formation being described as the ‘Mannum muds’ (Hubble et al., 2014). Sea levels fell slightly between c 6000 and 3500 BP (Job et al., 2021: 61) causing the Murray to incise its floodplain to its current level, creating the Lower Murray River Valley as we know it today.

Dating of the archaeological collections

Radiocarbon determinations for Tungawa 2, 6, Ngaut Ngaut and Tartanga Island, along with calibrations, are shown on Table 1. Working from the calibrated dates and site stratigraphy, Table 2 shows the approximate correspondence between dates and layers at the four sites. All sites are middle to late-Holocene in age. Tartanga Island Bed C (6.5–7.2 kya cal BP) provides a date for the upper part of the Monoman deposition event and for the mid-Holocene highstand. The lowest layers at Ngaut Ngaut (Layers XI and XII) and Tungawa 2 (Layer 11) show occupation began at c 5–6 kya, with Tungawa 6 covering only the upper part of the sequence (i.e. the last 4 kya). The original excavators were interested in establishing the antiquity of these sites and as a result, Ngaut Ngaut (Layers I–IV), Tungawa 2 (Layers 1–3) and Tungawa 6 (Layers 1–7) are undated with securely dated layers beginning about 3000 BP (Insert Table 2).

Table 2.

Approximate correspondence between cal BP dates and layers at Tungawa 2, 6, Ngaut Ngaut and Tartanga Island.

Dating	0–3 kya	3–4 kya	4–5 kya	5–6 kya	6–7 kya
Tungawa 2	Layers 0–3	4–5	6–8	9–11
Tungawa 6	Layers 1–7	8 to 18B
Ngaut Ngaut	Layers I–IV	V–VII	VIII–IX	X–XI	XII
Tartanga Isld	Beds I–F			Beds E–A

Faunal data – mammals

The number (MNI) of mammals identified at Ngaut Ngaut, Tungawa 6 and 2 is shown on Tables 3 –5 Following Frankel (1991: 61), Tables 3 –5 are organised on the assumption that the rate of deposition was constant through time. This spreads the contents of these layers evenly in these tables creating, as much as possible, comparable samples and allowing comparisons between sites to be made.

Table 3.

Mammalian fauna species, MNI identified at Ngaut Ngaut (after Smith, 1982: 112) and estimated dating (after Frankel, 1991: 61).

Dating	0–1 kya	1–2 kya	2–3 kya	3–4 kya	4–5 kya	5–7 kya	Total	%
Layer	I	II–III	IV	V–VI	VII–IX	X–XII
Antechinus sp.	0	1	0	0	0	0	1	0.7
Sarcophilus harrisii	0	0	0	1	1	0	2	0.4
Dasyurus maculatus	0	1	1	1	3	1	7	5
Thylacine sp.	0	0	0	2	1	0	3	2.1
Perameles notina (prev. bougainville)	0	1	0	2	2	0	5	3.6
cf. Isodon obesulus	1	0	1	0	1	1	4	2.9
Pseudocheirus peregrinus	0	1	1	0	0	0	2	1.4
Trichosurus vulpecula	1	1	1	1	3	2	9	6.4
cf. Lasiorhinus latifrons	0	0	0	2	1	0	3	2.1
Bettongs/Potorous	1	1	1	1	2	0	6	4.2
Wallaby small	2	9	8	10	18	3	50	35.7
Macropus fuliginosus	1	1	1	3	3	2	11	7.9
Hydromys chrysogaster	0	1	1	1	0	2	5	3.4
Rattus sp.	2	7	7	9	3	2	30	21.4
Leporillus sp.	0	0	0	0	1	0	1	0.7
Canis familiaris dingo	0	1	0	0	0	0	1	0.7
Total	8	25	22	33	39	13	140	99.8

Table 4.

Mammalian species, MNI and estimated dating for Tungawa 6 (after Wakefield in Mulvaney et al., 1964: 495).

Dating	0–1 kya	1–2 kya	2–3 kya	3–3.5 kya	3.5–4 kya	Total	%age
Layer	1–2	3–4	5–7	8–12	13–18B
Myrmecobius fasciatus	0	0	0	0	1	1	1
Perameles notina	2	1	3	0	0	6	6.2
Chaeropus ecaudatus	0	0	0	1	0	1	1
cf. Isoodon obesulus	0	0	0	0	1	1	1
Trichosurus vulpecula	2	1	1	4	2	10	10.3
Bettongia penicillata	1	1	2	4	6	14	14.4
Lagorchestes leporides	1	0	2	2	0	5	5.2
Lagostrophus fasciatus	0	0	0	0	1	1	1
Notamacropus eugenii	1	1	1	2	0	5	5.2
Macropus fujliginosus	0	0	1	0	4	5	5.2
Hydromys chrysogaster	0	0	2	1	2	5	5.2
Rattus lutreolus	10	8	8	9	3	38	39
Rattus fuscipes	0	0	0	2	2	4	4.1
Canis familiaris dingo	0	0	0	1	0	1	1
Total	17	12	20	26	22	97	99.8

Table 5.

Mammalian species, MNI and estimated dating for Tungawa 2 (after Wakefield in Mulvaney et al., 1964: 497).

Dating	<1 kya	1–2 kya	2–3 kya	3–4 kya	4–5 kya	5–6 kya	Total	%
Layer	0	1–2	3–4	5–6	7–8	9–11
Antechinus flavipes	5	0	1	0	1	0	7	1.2
Antechinus sp.	0	1	0	0	0	0	1	0.2
Dasycercus cristicauda	0	0	0	0	1	0	1	0.2
Sminthopsis cf. murina	0	1	1	0	0	0	2	0.3
Dasyurus geoffroyi	0	0	0	2	0	0	2	0.3
Dasyurus maculatus	0	0	1	1	0	1	3	0.5
Sarcophilus harrisii	0	0	0	1	1	0	2	0.3
Myrmecobius fasciatus	0	0	0	0	1	0	1	0.2
Thylacine sp.	0	0	0	0	1	0	1	0.2
Perameles notina	5	4	2	3	5	2	21	3.6
Chaeropus ecaudatus	2	1	0	0	0	0	3	0.5
cf. Isoodon obesulus	2	1	0	1	1	0	5	0.9
Pseudocheirus peregrinus	1	3	0	0	0	0	4	0.7
Trichosurus vulpecula	10	4	2	0	0	0	16	2.7
cf. Lasiorhinus latifrons	1	0	0	1	2	1	5	0.9
Bettongia penicillata	14	4	4	26	20	3	71	12.1
Bettongia lesueur	0	1	0	1	0	0	2	0.3
Potorus platyops	0	0	0	2	2	0	4	0.7
Lagorchestes leporides	9	2	4	14	8	1	38	6.5
Lagostrophus fasciatus	8	2	4	24	14	2	54	9.2
Onychogalea lunata	1	0	0	0	0	0	1	0.2
Notamacropus eugenii	3	3	2	4	4	0	16	2.7
Macropus fuliginosus	2	0	4	7	4	3	20	3.4
Hydromys chrysogaster	2	1	0	3	1	0	7	1.2
Rattus lutreolus	138	33	34	40	24	9	278	47.3
Rattus fuscipes	6	1	5	1	0	0	13	2.2
Pseudomys auritus	1	0	1	0	0	1	3	0.5
cf. Pseudomys/Notoymys	0	1	2	0	0	0	3	0.5
Conilurus albipes	1	1	0	1	0	0	3	0.5
Canis familiaris dingo	1	0	0	0	0	0	1	0.2
Total	212	64	67	132	90	23	588	100.2

Adding identifications from Finlayson (in Hale and Tindale, 1930) and Wakefield (in Mulvaney et al., 1964: 496–498) to those of Smith (1982) allows us to include Lagorstrophus fasciatus, Lagorchestes cf., leporides, Notamacropus cf. eugenii and Onychogalea lunata in the small wallaby category and Potorous platyops in the Bettong/Potoroo category. These further identifications give us confidence that species at Ngaut Ngaut are similar to those at the Tungawa sites, supporting Wakefield’s conclusion that ‘The mammals of [these] shelters represent an essentially modern fauna, [showing] no major variation in composition during the time of accumulation- approximately 5000 years’ (in Mulvaney et al., 1964: 498).

There are differences in the number of taxa identified at Ngaut Ngaut (n = 16), Tungawa 6 (n = 14) and Tungawa 2 (n = 30). The differences in taxa mostly concern smaller mammals, with more of these being identified (present?) at Tungawa 2. In addition, the MNI of mammals at Tungawa 2 (n = 588) is four times the size as that from either Tungawa 6 (n = 97) or Ngaut Ngaut (n = 140) correlating with the larger volume of Tungawa 2 excavated. Of the 588 animals and 30 species at Tungawa 2, however, 11 species make up 92% (n = 541) of the MNI. The remaining 19 species (8% of the total) consist of five or fewer individuals and removing those species (i.e. those with an individual score of less than 1% of total MNI) makes the data from Tungawa 2 less unwieldy and more comparable with that from the other two sites (compare Tables 3 –6).

Table 6.

MNI of major species identified at Tungawa 2.

Dating	<1 kya	1–2 kya	2–3 kya	3–4 kya	4–5 kya	5–6 kya	Total	%
Layer	0	1–2	3–4	5–6	7–8	9–11
Antechinus flavipes	5	0	1	0	1	0	7	1.3
Perameles notina	5	4	2	3	5	2	21	3.9
Trichosurus vulpecula	10	4	2	0	0	0	16	3.0
Bettongia penicillata	14	4	4	26	20	3	71	13.1
Lagorchestes leporides	9	2	4	14	8	1	38	7.0
Lagostrophus fasciatus	8	2	4	24	14	2	54	10.0
Macropus eugenii	3	3	2	4	4	0	16	3.0
Macropus fuliginosus	2	0	4	7	4	3	20	3.7
Hydromys chrysogaster	2	1	0	3	1	0	7	1.3
Rattus lutreolus	138	33	34	40	24	9	278	51.4
Rattus fuscipes	6	1	5	1	0	0	13	2.4
Total	202	54	62	122	81	20	541	100.00%

A number of mammalian species occur in most levels at all three sites suggesting that these are targetted species. These include small wallabies (Lagorchestes leporides, Lagostrophous fasciatus, Notamacropus eugenii), large kangaroos (Macropus fuliginosus), bandicoots (Perameles notina, prev. bougainville), brushtail possums (Trichosurus vulpecula) and water and bush rats (Hydromys chrysogaster, Rattus lutreolus and Rattus fuscipes). To this list should be added Bettongia penicillata, Brush-tailed bettong, which is plentiful at Tungawa 2 and 6. Bettongs/Potoroos are recorded in low numbers in most layers at Ngaut Ngaut.

Non-mammalian fauna

Smith (1982) provides identifications for non-mammalian species (shellfish, birds, reptiles, fish and crustacea) at Ngaut Ngaut, where the deposit largely consisted of layers of freshwater shellfish interspersed with ash and sand. Freshwater mussels, Velesunio ambiguous and Alathyria jacksoni were the predominate species in all layers with Corbicula australis, Notopala hanleyi, Thiara balonnensis and Bullinus tenuistriata present as minor shellfish species. Shellfish bulk sample weights and the MNI of non-mammalian species are shown on Table 7. Apart from shellfish, a few species are targetted (lizards, turtles, fish) and these are present in all layers. Emu eggshell was present at Ngaut Ngaut in most layers. Overall, there is little change in the non-mammalian fauna over the life of the site. There are two exceptions. First, while acknowledging that estimating shellfish quantities from bulk samples has difficulties (Table 7), Smith (1982: 111–112) and Frankel (1991: 65) suggest an increase in the quantity of shellfish being gathered in the final 2000 years of site use, together with yabbies (Cherax destructor) which are present and increase in numbers only post c. 3000 BP.

Table 7.

Shellfish bulk samples, other non-mammalian fauna species and MNI for Ngaut Ngaut (Frankel, 1991: 63, after Smith, 1982: 112–113).

Dating	<1 kya	1–2 kya	2–3 kya	3–4 kya	4–5 kya	5–7 kya	Total	%
Layer	I	II–III	IV	V–VI	VII–IX	X–XII
Shellfish wt (g)	1023	446	205	516	984	n/a	3174
Waterbirds
Anas cf. castanea (Chestnut teal)					1		1	0.4
Biziura lobata (Musk duck)		1					1	0.4
Large duck unident.						1	1	0.4
Phalacrocorax carbo (Great Cormorant)						1	1	0.4
Microcarbo melanoleucos (Little shag)						1	1	0.4
Phalacrocorax varius (Pied shag)		1					1	0.4
Grus rubicunda (Brolga)						1	1	0.4
Other birds
Aquila audax (Wedge-tail eagle)		1					1	0.4
Dromaius novaehollandiae (Emu)						1	1	0.4
Unident. Bird, small, medium		1	1				2	0.8
Reptiles
Pogona vitticeps (Bearded Dragon)	2	2	1	2	7		14	5.4
Varanus gouldii (Goanna)	1	1	1				3	1.2
Tiliqua scincoides (Bluetongue)			1		5		6	2.4
Trachydosaurus rugosa (Shingleback)	2	2	1	4	1		10	4
Morelia spilota (Carpet python)	1	1					2	0.8
Notechis scutatus (Tiger snake)					1		1	0.4
Serpentes unident.				1			1	0.4
Amphibians (Turtles)
Emydura macquarii (Murray R. turtle)			1	1	1	2	5	2
Chelodina longicollis (Long necked turtle)	1	1	1	2	3	1	9	3.5
Chelodina expansa (Snake necked turtle)						1	1	0.4
Fish
Maccullochella peelii (Murray cod)	4	7	2	5	7	3	28	10.8
Macquaria ambigua (Golden perch)		5	1		1		7	2.7
Tandanus tandanus (Catfish)					1		1	0.4
Crustacea
Cherax destructor (Yabbie)	60	72	25	2			159	61
Euastacus armatus (Murray crayfish)	1	1					2	0.8
Total (excluding shellfish)	72	96	35	17	28	12	260	100.6

n/a: not available.

Information regarding the non-mammalian faunal materials from Tungawa 2 and 6 was presented in qualitative rather than numerical terms. Mulvaney (1960): 81 and Mulvaney et al. (1964: 494) notes the same major shellfish species as found at Ngaut Ngaut and the list of non-mammalian fauna at Tungawa 2 and 6 (Table 8) parallels that at Ngaut Ngaut (Table 8). As at Ngaut Ngaut, Cherax destructor (yabbies, freshwater crayfish) only occur at Tungawa 2 and 6 from layers 3 and 14 respectively (post c 3000 BP).

Table 8.

Non-mammalian fauna recorded from Tungawa 2 and 6 (after Mulvaney, 1960: 62–73, Mulvaney et al., 1964: 494).

Species/category	Tungawa 2	Tungawa 6
Birds
Unidentified birds	Present in all layers	Present in most layers
Dromaius novaehollandiae eggshell	Present in all layers	Present in most layers
Possible Mallee fowl eggshell	Not recorded	Layers 11–18
Reptiles
Pogona sp. (bearded dragons)	Present in layers 4–8	Not recorded
Tiliqua scincoides and/or T. rugosa	Present in all layers except layer 11	Present in all layers
Varanus sp	Layers 0, 4–8	Layers 1–10
Serpentes	Layers 0, 9–11	Not recorded
Amphibians
Chelodina sp.	Present in most layers	Present in most layers
Fish
Unidentified fish	Present in most layers	Present in most layers
Maccullochella peelii	Present in several layers	Layers 16–18B
Catfish sp.	Not recorded	Layer 11
Crustacea
Cherax destructor	Present in layers 0, 1–3	Present in layers 1–14

The Gorge shelter sites (Ngaut Ngaut, Tungawa 2 and 6) are situated at the interface between the xeric environment of the plains and the mesic environment of the river and river valley. The faunal remains from these sites indicate that Aboriginal hunter gatherers utilised both zones throughout. However, the plains and the river have different Holocene histories. The Murray Plains responded to climatic and environmental changes affecting southern Australia as a whole. While these also impacted the Murray River, the catchment of the Murray covers an area from the Eastern Highlands to south-eastern Queensland and its flow and sediment load responded to changes outside the immediate region. In addition, the base level of the river was affected by the fall and rise in sea levels creating phases of aggradation and deposition. As a result, it is useful to consider the faunas from the plains separately to those from the river valley. These cannot be entirely segregated as animals, like humans, can move between zones and some species have wide tolerances, so this separation is about preferred habitats.

Murray plains fauna

The Murray Plains are covered with woodland, tall open shrubland and dunes with Mallee, saltbush and porcupine grass (Robinson et al., 2009). Its arid and semi-arid nature is exacerbated by sandy soils and a limestone substrate. Summers are hot, with a mean maximum between November and March of 30°C, winters are cooler with average monthly maximum of less than 20°C and minimums of 8°C. While rainfall is fairly even across the year, rainfall in the coolest months is more effective because of reduced evaporation (Australian Bureau of Meterology, nd). As a result, autumn, winter and early spring are productive in terms of grass and shrub growth, the fruiting and flowering of shrubs and trees and water availability (Specht, 1984).

In terms of the species represented in the archaeological sites, typical species from this habitat include mammals: Perameles notina, Bettongia penicillata, Lagorchestes leporides, Lagostrophous fasciatus, Notamacropus eugenii and Macropus fuliginosus; birds: emu, emu and Mallee fowl eggs; reptiles: Pogona sp. (bearded dragons), Tiliqua sp. (shingle-back lizards) and Varanus sp. (goannas). These species are present throughout at Ngaut Ngaut, and Tungawa 2 and 6 (Tables 3 –8). The sample from the earliest layers at Ngaut Ngaut and Tungawa 2 is small, with numbers of major mammalian species and reptiles (lizards) being greatest in the period c. 3000–5000 years BP and drop thereafter. Only Tungawa 2 shows an increase in the number of mammals in the last 1000 years. Comparing Tables 4 –9 allows us to see changes in detail that have taken place over the 6000–7000 years of site use.

The earliest/lowest levels at Ngaut Ngaut and Tungawa 2 are depauperate in terms of the number of species and the number of individuals, showing only the most common species.

At Ngaut Ngaut, Tasmanian devils (Sarcophilus harrisii) and Thylacines (Thylacinus cynocephalus) do not occur above Layer VI (dated by Layer V, 2892–3375 cal BP, Table 1). Similarly, two Tasmanian devils are recorded at Tungawa 2, Layers 5 and 7 (Layer 4, 3255–3684 cal BP) with a single Thylacine tooth at Layer 7. These dates are in accord with White et al.’s (2018) assessment that both Thylacines and Tasmanian devils became extinct on the Australian mainland between 3179 and 3227 cal. BP.

The highest numbers of the major mammalian species at all three sites occur between c. 3000 and 5000 BP (4 kya at Tungawa 6) followed, with some exceptions, by a fall in numbers.

Exceptions mostly concern Tungawa 2, where large kangaroos and small wallabies show an increase in numbers in the uppermost layer (last 1000 years? of use of this site).

Within the suite of mammals obtained from the plains, large kangaroos (M. fuliginosus, weighing between c 20 and 70 kg), wombats (cf. Lasiorhinus latifrons, 19–32 kg) and devils (Sarcophilus harisii, 5–12 kg), would represent significant windfall captures. Many of the animals coming from the plains, however, consisted of wallabies, rat kangaroos, bandicoots and reptiles. These are small package items difficult to capture in the Murray Plains environment, an almost endless mosaic of coarse-grained patches, where the pursuit and capture of small mammals and lizards would be a matter of chance, knowledge and skill. Eyre (1845: 276–277, 282–283) and Clarke (2009: 151–2) note the use of spears, clubs and cooperative hunting drives using fire, nets and brushwork fences on the Lower Murray to capture kangaroos, wallabies and emu (see also Berndt, 1947: 75, Finlayson, 1952: 45–46 for descriptions of hunting hare wallabies, Lagorchestes sp. with clubs, nets and fire drives). Lizards were caught when disturbed by hunting parties or when the bush was fired. In addition to the use of fire in active hunting, Bliege Bird et al. (2008: 14799) and Bolton and Latz (1978) observe that many small mammals and reptiles benefit from vegetation patches created by burning and, as a result, burning maximises habitat diversity and foraging returns.

The plants adjacent to these shelters, and the herbivores dependent on them, are made up of fire and drought-adapted vegetation (mallee-form Eucalypts, heaths and chenopods on sandy and calcareous soils), well able to cope with drier circumstances (Mashmedt, 2009: 50; Robinson et al., 2009: 216). The response to drier more variable conditions in this environment is likely to be reduction in the number of plants and animals present rather than species change (cf. Dodson, 2001). While the major mammalian species brought back to Ngaut Ngaut and Tungawa 2 and 6 are present throughout, there is a reduction in numbers of plains animals at these sites between 3000 and 1000 BP, most marked in the numbers of small wallabies and Bettong/Potoroos, at Tungawa 2 (Table 9).

Table 9.

Comparing MNI major species/categories present at Ngaut Ngaut, Tungawa 2 and 6.

Dating	0–1 kya	1–2 kya	2–3 kya	3–4 kya	4–5 kya	5–6 kya*	Total N
Large K/roo Ngaut Ngaut*	1	1	1	3	3	2	11
Large K/roo Tungawa 2	2	0	4	7	4	3	20
Large K/roo Tugawa 6	0	0	1	4	n/a	n/a	5
Wallaby small Ngaut Ngaut	2	9	8	10	18	3	50
Wallaby Small Tungawa 2	21	7	10	42	26	3	109
Wallaby small Tungawa 6	2	1	3	5	n/a	n/a	11
Bettong/Potoroo Ngaut Ngaut	1	1	1	1	2	0	6
Bettong/Potoroo Tungawa 2	14	5	4	29	22	3	77
Bettong/ Potoroo Tungawa 6	1	1	2	10	n/a	n/a	14
Trichosurus vulpecula Ngaut Ngaut	1	1	1	1	3	2	9
Trichosurus vulpecula Tungawa 2	10	4	2	0	0	0	16
Trichosurus vulpecula Tungawa 6	2	1	1	6	n/a	n/a	10
Perameles notina Ngaut Ngaut	0	1	0	2	2	0	5
Perameles notina Tungawa 2	5	4	2	3	5	2	21
Perameles notina Tungawa 6	2	1	3	0	n/a	n/a	6
Rattus sp. Ngaut Ngaut	2	7	7	9	3	2	30
Rattus sp. Tungawa 2	144	34	39	41	24	9	291
Rattus sp. Tungawa 6	10	8	8	16	n/a	n/a	42
Total	220	86	97	189	112	29	733

n/a (not applicable) for Tungawa 6.

Ngaut Ngaut Layers X–XII are 5–7 kya BP, data from Tables 3 –5 above.

The increase in numbers in Level 0 at Tungawa 2 might be a reflection of the return to less variable conditions over the last 1000 years (Marx et al., 2011: 331), a response to changed local conditions or to differences in the age of occupation of the three sites. As noted above, the uppermost layers at Ngaut Ngaut, Tungawa 2 and 6 are undated limiting our ability to compare the three sites in the final period of use.

Murray River and the Gorge valley

The Murray River and the gorge contain a complex set of features; the river channel, anabranches, islands, billabongs, swamps and terraces, with alternating steep or eroded cliffs on its margins. River productivity is related to flow, particularly when the river approaches bank full stage and overtops its banks (Ellis et al., 2016: 15; Gawne et al., 2007: 1070–1071, Humphries et al., 2014; Sheldon, 1994: 74) When floodwaters enters this complex environment, there is a burst of nutrients which stimulates the breeding and growth of aquatic organisms. This is highest at peak and falling flood, when nutrients flow from and back into the main channel.

Prior to its regulation in 1922 through locks and weirs, the Murray River in South Australia had two major components a falling and low phase during the months of December to April, and a rising and flood phase from May to November. The river was lowest between February and June and highest between September and December (Goode and Harvey, 2009; Maheshwari et al., 1995: 24). To a degree, productivity on the Murray Plains and the river were inversely correlated, with the Murray Plains being most productive during the cooler autumn and winter months, while the river saw peaks in productivity in spring and summer during flood rising and falling stages. Three resource zones are relevant when considering fauna obtained from the river and its margins.

a. Riparian forest and woodlands. These are divided into a red gum zone on the river banks and low-lying areas, with Eucalyptus camaldulensis and a grass and spike-rush understory, and a black box zone on the higher, outer margin of the floodplain with E. largiflorens, and a Melaleuca sp., lignum and chenopod understory (Margules and Partners Pty Ltd, 1990: 13). Species in this habitat included Antechinus flavipes, Dasyurus maculatus, Bettongia penicillata, Pseudocheirus peregrinus, Trichosurus vulpecula, Rattus fuscipes along with Tiliqua sp. (blue-tongue lizards) and birds.

b. Littoral zone and swamps. Vegetation here consists of grasses, sedges and rushes, including Typha sp., Phragmites australis and Cyperus gymnocaulus (Walker et al., 1992: 285). Hydromys chrysogaster, Rattus lutreolus, shellfish, freshwater turtles, yabbies and waterbirds are available within this zone.

c. River channels, anabranches, backwaters and billabongs were important for shellfish, Murray cod, catfish, freshwater turtles, yabbies and birds.

Knowledge of non-mammalian faunas at Tungawa 2 and 6 is limited especially in terms of numbers of fish and freshwater turtles. For the most part, the animals identified as coming from the riparian, littoral and river zones are present throughout at all three sites suggesting constant exploitation of them (Tables 3 –9). Exceptions, however, are as follows;

There is a jump in the numbers of B. penicillata in the uppermost layer of Tungawa 2.

Possums (T. vulpecula and P. peregrinus) are present at Tungawa 2 and P. peregrinus at Ngaut Ngaut only in the last 3000 years. This suggests either a change in the local riparian forests or increased hunting close to the river.

While the numbers of Rattus lutreolus at Tungawa 2 are large, especially in the uppermost layer, the small size of these animals (50–200 g) suggests they are not a major source of food and may have been collected as part of other activities in swamps or at the water’s edge.

Crayfish (especially yabbies, C. destructor) are present only in the last 3000 years at all three sites and at Ngaut Ngaut increase in numbers through time (Tables 8 and 9).

Smith (1982: 114) and Frankel (1991: 60, 64–65) reviewing the data for Ngaut Ngaut concluded there was an increase in the proportion of shellfish being gathered after 2000 BP (Table 8).

The evidence for crayfish at these sites is in the form of gastroliths created prior to moulting. Crayfish gastroliths and emu eggshell are indicators of the autumn, winter and early spring periods (Martin, 1999: vi; Walker et al., 2009: 296–297). Smith (1982: 112–113) took their presence at Ngaut Ngaut and the Tungawa sites to indicate Aboriginal occupation during the cooler months, while Eyre (1845: 255, 303) associated the autumn and winter with hardship noting people resided in shelters along the cliffs at this time. This was a time of clear, cold water in the river, when spearfishing drives took place, and large and small crayfish (Euastacus armatus and Cherax destructor respectively) were collected (Eyre 1845: 262). Eyre (1845: 252, 267–268) provides a detailed description,

‘U-kod-ko, the crayfish of the small kind is obtained by the women wading into the water in a long close line, stooping downward walking backwards whilst they grope with their hands for the crayfish. . .or by the men, wading and using a large low net called a wharro which is dragged along by 2 or 3 of them’.

B. penicillata, the brush-tailed bettong, is associated with lignum (Duma florulenta) thickets on the Murray riverflats, it is a mycophagous (fungi-eating) mammal and both it and lignum benefit from frequent fires (Krefft, 1862: 20–21; Johnson, 1995).

Factors affecting the information

Earlier reference was made to processes other than food selection which might bias results. That the hunting of mammals, in this case wallabies, might be fulfilling roles in addition to their economic function is demonstrated by a burial at the Roonka site (Burial 108). Around the head of this adult male (20–35 years old) was a double stranded chaplet of notched Notamacropus eugenii incisors, from an estimated 40+ individual animals. This burial dates to 6020–6660 cal BP, and the chaplet is an unusual mark of prestige (Littleton et al., 2017: 102; Pretty, 1977: 312–314). N. eugenii were relatively common at both Tungawa 2 and 6 (Tables 4 –6) and present at Ngaut Ngaut. Assuming processes such as costly signalling were active, it can be concluded that, as far as the plains fauna was concerned, these were present at all three sites throughout the period of occupation.

The extinction of Thylacines (Thylacinus cynocephalus) and the presence of dingoes (Canis familiaris), coincident at c 3000 BP, may have affected the abundance of mammals on the Murray Plains (Letnic et al., 2014). Koungoulos (2017: 38) presents evidence that dingoes prey on larger mammals (>10 kg) and consequently their arrival might have created conditions where small to medium-sized mammals (5–10 kg) increased in number. Fillios et al. (2010) set out strategies for investigating the impact of the introduction of the dingo on faunal assemblages in terms of changes in species abundance and bone fragmentation. That it is the smaller mammals which suffer the greatest decline post 3000 BP at Ngaut Ngaut and the Tungawa sites (Table 9), however, suggests that dingoes may have had some, but not a great, impact on these results.

The presence of crayfish gastroliths and emu eggshell indicate occupation of the shelter sites during the cooler months of the year. There is negative evidence that the fauna in the shelters might not represent summer and early spring occupation. Eyre was unspecific about the timing of some activities, for example, shellfishing, or the hunting of kangaroos with spears or throwing sticks, using pit falls, brush fences, fire drives and/or nets. However, he relates other activities to their environmental and seasonal circumstances. He described food as being most abundant during spring and summer and drew particular attention to the mass capture of waterbirds and fish using nets, describing large quantities of fish being taken in weirs across creeks as flood waters fell (Eyre, 1845: 218, 284, 252–253, 268). Waterbirds, fish and freshwater turtles are present in most layers at all three shelter sites, but where we have MNI information, that is, at Ngaut Ngaut, the numbers are relatively small and do not represent large scale capture. While this is negative evidence, it has to be placed alongside the evidence for winter/ early spring occupation in the form of emu eggshell, present in almost every layer at these sites.

There are traces of quondong (Santalum acuminatum) at Tungawa 2 and 6 and Scirpus (now Bolboschoenus) fluviatilis fibre at Tungawa 6, however, no other plant remains were noted. A number of plant species have been recorded as being economically important for Aboriginal people living along the Murray River, including roots such as Microseris scapigera, fruit such as Nitraria schoberi and Santalum acuminatum (available on the Murray Plains), and water plants, Triglochin procera (now Cycnogeton procerum), and Typha sp. in the swamps and on the river’s edge (Clarke, 1988, 2009; Cleland, 1966; Eyre, 1845). Clarke (1988) states that Microseris, Typha, Triglochin and Oxalis (all root crops) were important as they were available all year round, citing evidence from Angas and Eyre that Typha sp., where it occurred, was a staple food. Similarly, belilllah, identified by Gott (1982) as club rush (Bolboschoenus fluviatilis), another important root food, grew luxuriously on low lying areas once the floods had receded (Eyre, 1845: 269).

Tartanga Island

Tartanga Island, while close to Ngaut Ngaut, is located in a different ecological zone being on the floodplain and consisting of deposits laid down initially as an old shoreline. Beds A–E (Monoman Formation) and F– I (Coonambidgal Formation) reflect aggradation in the river in response to sea level changes. The faunal information (from Hale and Tindale, 1930: 154–159) is shown in Table 10. Tartanga is dated by the C14 date from Bed C and by the Monoman and Coonambidgal formations.

Table 10.

Faunal information for Tartanga Island (Hale and Tindale, 1930: 154–159).

Bed I	Sandy, broken mussel shells and other signs of Aboriginal occupation, a few hand mills.
Bed H	Sandy clay, Unionid mussels, burnt stones and ash, some in thick stratified layers, a few fragments of tortoise bone.
Bed G	Sterile, grey clay.
Bed F	Sterile, blue-black clay.
Break?
Bed E	Unionid mussels, Notopala hanleyii, Bulinus pectorosus (?), very large vertebrae Maccullochella peelii, Emydura macquarii, burnt stones.
Bed D	Unionid mussels, Notopala hanleyii, several Macropus fuliginopus, Maccullochella peelii, Emydura macquarii, Chelodina longicollis, other mammal fauna similar to Bed C, burnt stones.
Bed C	Unionid mussels, Notopala hanleyii, Notamacropus cf. eugenii, Trichosurus vulpecula, large no. of vertebrae and jaws of Maccullochella peelii, Tandanus sp, unidentified smaller fish, Emydura macquarii, Chelodina longicollis, burnt stones (Dated 6569–7246 cal BP)
Bed B	Unionid mussels, Macropus sp., vertebrae and jaws of Maccullochella, Chelodina, burnt stones
Bed A	Unionid mussel shells, burnt stones (hearthstones)

Tartanga Island is one of the oldest sites discussed here and appears to have been occupied earlier than Ngaut Ngaut or Tungawa 2. The stratigraphic column shows active midden deposited c 5500 to >7000 BP, with a faunal collection consisting of shellfish, fish, turtles and small and large mammals. At this point in the life of the site (Beds A to E), the collection of both mammalian and non-mammalian fauna is similar to that of Ngaut Ngaut and the Tungawa sites, though the numbers of Murray cod (Maccullochella peelii) and other fish may be greater at Tartanga. There appears to be a break, followed by the deposition of Coonambidgal clays in Beds F and G. The uppermost beds at Tartanga (Beds H and I) show renewed Aboriginal occupation consisting of shell midden and a few turtle bones. These latter beds show a somewhat different hunting pattern to those at the shelter sites, possibly reflecting changed conditions on the river.

Conclusion

The Murray River flows through its Riverina, Mallee and Gorge sections for a distance of 1500 km (Littleton et al., 2021). Outside the Gorge, the river over this distance is characterised by multiple channels, anabranches and cut-off meanders. This complex landscape is a response to the combined effects of changes in the river’s flow regime, the post Pleistocene rise in sea level and the deposition of fine sediments of the upper Coonambidgal Formation during the mid-Holocene (c. 6000–3000 BP) (Firman, 1966). Coincident in time and space with the formation of this landscape are Aboriginal cooking mounds, associated with the harvesting and processing of wetland plant foods such as Typha, Triglochin and Bolboschoenus sp., mounds which appear of the Murray Riverine Plain as early as 4300 BP (Martin, 2011: 171). Similarly, Bourman et al. (2022: 4–6) note the emergence of numerous cooking mounds at Calperum floodplain, north of Renmark, at c. 3500 cal BP (Map 1). They suggest these mounds are evidence for the utilisation of aquatic plants, an economic response to the mid-Holocene ENSO-influenced climate. However, they are also a response to changes in river morphology associated with the deposition of Coonambidgal sediments, changes which created highly productive swamp environments on the floodplains where aquatic plants could thrive. Within this scenario, changes in river flow, sea level and sediments improved the productivity of the river zone at the same time that climatically-induced changes had a negative effect on the adjacent Murray Plains.

Ngaut Ngaut, Tungawa 2 and 6 and Tartanga Island were first occupied at the time that post-Pleistocene sea levels reached their highest point. After initial occupation, shifts in Aboriginal activities take place in concert with environmental changes posited for this region (Table 11). The faunal evidence from the Gorge archaeological sites is consistent with a fall in the productivity of the surrounding plains, post-3000 BP, alongside a rise in the productivity of the river and swamp environment, with Aboriginal people adjusting their activities in response. A reduction in the number of animals being taken from the Murray Plains, allied with greater attention being given to shellfish, crayfish and probably plant foods, supports Williams et al. (2015) and Asmussen and McInnes (2013) contention that mid-Holocene environmental changes could stimulate a shift towards previously underutilised environments and lower ranked, higher cost foods.

Table 11.

Holocene environmental changes charted against faunal information from Gorge archaeological sites.

Time Period	Conditions	Response?
8–4 kya BP	Peak in temperatures and effective precipitation, sea-level rise to mid-Holocene high stand, deposition of Monoman sands.	TI (Beds A to E), NN and T2 occupied with high numbers of plains mammals and reptiles exploited together with mammals, fish, shellfish and turtles from the river zone..
c.5 kya BP	Onset of ENSO dominated climate, deposition of Coonambidgal clays and fine silts.
4.2–1 kya BP	Increased variability, long duration drier periods interspersed with unpredictable wet phases. Murray River has modern physiognomy.	NN and T2 and 6, fall in numbers of plains mammals and reptiles, increased activity in riparian, swamp and river edge zones, crayfish and increased shellfish, aquatic plants? TI (Beds H and I) shellmidden and turtles.
1 kya BP to Present	Conditions on the plains and river valley similar to those of today.	NN, T2, 6 upper layers undated, Increased numbers of plains, riparian and swamp animals at T2 Layer 0 .

TI: Tartanga Island; NN: Ngaut Ngaut; T2 and 6: Tungawa 2 and 6.

The faunal information and the changes documented at Ngaut Ngaut and Tungawa 2 and 6 appears to reflect an autumn, winter and early spring pattern. Ethnographic evidence suggests we are missing part of the archaeological record for this area, namely open midden sites containing evidence for spring and summer occupation. Some of this record may have been lost to development. Prior to the uptake of river frontage properties for pasture and irrigation farming, substantial mounds of mussel shells (Alathyria jacksoni and Velesunio ambiguous) were common along the river,

‘. . .the country around the Nor’West bend. . .. was a paradise of fish, fowl, and game. . .Even today immense heaps of shell testify to the. . .feasts which took place along the banks of the Murray. . .’ (The Adelaide Chronicle July 7, 1932: 41–44).

The faunal collections from Ngaut Ngaut, Tungawa 2 and 6 and Tartanga Island, curated in the South Australia Museum, are highly significant. They provide a comparative perspective on Aboriginal- animal interactions over the past 7000 years, a record that is unusual in the Australian archaeological context. Furthermore, many of the mammals, common and hunted over the preceding 6000 years, are now extinct, locally extinct or on endangered lists. This fauna from these sites documents an ecological and cultural world that no longer exists (Fusco et al., 2016; Menkhorst, 2009; Wilson et al., 2022).

New excavations, allied with modern approaches to dating and midden analysis, would allow greater certainty about the trends discussed here, particularly for the post 3000 BP period. Nonetheless, this exercise has allowed us to ask significant questions about the past and it has better defined the archaeological issues pertaining to this section of the Murray River.

Footnotes

Acknowledgements

This paper has benefitted from comments by Keryn Walshe, Peter Menkhorst and Peter Hiscock. We acknowledge the support and collaboration of the River Murray and Mallee Aboriginal Corporation (RMAC) in this research. Seline McNamee, School of Social Science, University of Auckland, drew the Map. Vale Mike Smith.

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This work is part of a larger project, Royal Society of New Zealand Marsden Fund Grant (14-UOA-19), with Judith Littleton as PI.

ORCID iD

Harry Allen

References

Altman

Peterson

(1988) Rights to game and rights to cash among contemporary Australian hunter-gatherers. In: Ingold

Riches

Woodburn

(eds) Hunters and Gatherers 2: Property, Power and Ideology. Oxford: Berg, pp.73–149.

Asmussen

McInnes

(2013) Assessing the impact of mid-to-late holocene ENSO-driven climate change on toxic macrozamia seed use: A 5000 year record from eastern Australia. Journal of Archaeological Science 40: 471–480.

Australian Bureau of Meterology (n.d.) Climate statistics for Australian locations. Available at: http://www.bom.gov.au/climate/averages/tables/cw_024018.shtml (accessed March 2019).

Barr

Tibby

Leng

, et al. (2019) Holocene El Niño-Southern Oscillation variability reflected in subtropical Australian precipitation. Scientific Reports 9: 1627.

Belperio

Harvey

Bourman

(2002) Spatial and temporal variability in the holocene sea-level record of the South Australian coastline. Sedimentary Geology 150: 153–169.

Berndt

(1947) Wuradjeri magic and “Clever Men”. Oceania 18: 60–86.

Bland

Roberts

Campbell

; The Mannum Aboriginal Community Association Inc (MACAI) (2012) New interpretations for the stone artefact assemblage from Ngaut Ngaut (Devon Downs), South Australia. Journal of the Anthropological Society of South Australia 36: 46–65.

Bliege Bird

Bird

Codding

, et al. (2008) The “fire stick farming” hypothesis: Australian Aboriginal foraging strategies, biodiversity, and anthropogenic fire mosaics. Proceedings of the National Academy of Science 105(39): 14796–14801.

Bliege Bird

Smith

Bird

(2001) The hunting handicap: Costly signaling in human foraging strategies. Behavioral Ecology and Sociobiology 50: 9–19.

10.

Bolton

Latz

(1978) The western hare-Wallaby Lagorchestes hirsutus (Gould) (Macropodidae), in the Tanami Desert. Australian Wildlife Research 5: 285–293.

11.

Bourman

Murray-Wallace

Wilson

, et al. (2022) Holocene freshwater history of the Lower River Murray and its terminal lakes, Alexandrina and Albert, South Australia, and its relevance to contemporary environmental management. Australian Journal of Earth Sciences 69: 605–629.

12.

Broecker

Kulp

Tucek

(1956) Lamont natural radiocarbon measurements III. Science 124: 154–165.

13.

Buckman

Brownlie

Bourman

, et al. (2009) Holocene palaeofire records in a high-level, proximal valley-fill (Wilson Bog), mount lofty ranges, South australia. The Holocene 19(7): 1017–1029.

14.

Clarke

(1988) Aboriginal use of subterranean plant parts in southern South Australia. Records of the South Australian Museum 22(63): 76.

15.

Clarke

(2009) Aboriginal culture and the riverine environment. In: Jennings

(ed.) Natural History of the Riverland and Murraylands. Adelaide: Royal Society of South Australia Inc, pp.142–161.

16.

Cleland

(1966) The ecology of Aboriginal man in South and Central Australia. In: Cotton

(ed.) Aboriginal Man in South and Central Australia. Adelaide: Government Printer, pp.111–158.

17.

Conolly

(2017) Costly signalling in archaeology: Origins, relevance, challenges and prospects. World Archaeology 49: 435–445.

18.

Dodson

(2001) Holocene vegetation change in the Mediterranean-type climate regions of Australia. The Holocene 11(6): 673–680.

19.

Dortch

Balme

Ogilvie

(2012) Aboriginal responses to Late Quaternary environmental change in a Mediterranean-type region: Zooarchaeological evidence from south-western Australia. Quaternary International 264: 121–134.

20.

Ellis

Cheshire

Townsend

, et al. (2016) Fish and Flows in the Murray River Catchment - A Review of Environmental Water Requirements for Native Fish in the Murray River Catchment. Queanbeyan, NSW: NSW Department of Primary Industries.

21.

Eyre

(1845) Journals of Expeditions of Discovery Into Central Australia, and Overland From Adelaide to King George’s Sound, in the Years 1840-1841, Facsimile Reprint 1964, vol. II. Adelaide: Libraries Board of South Australia.

22.

Fillios

Gordon

Koch

, et al. (2010) The effect of a top predator on kangaroo abundance in arid Australia and its implications for archaeological faunal assemblages. Journal of Archaeological Science 37: 986–993.

23.

Finlayson

(1952) The Red Centre: Men and Beast in the Heart of Australia. Sydney: Angus and Robertson.

24.

Firman

J B

(1966) Stratigraphy of the Chowilla area in the Murray Basin. Quarterly Geological Notes 20: 3–7.

25.

Fletcher

M-S

Moreno

(2012) Have the southern westerlies changed in a zonally symmetric manner over the last 14,000 years? A hemisphere-wide take on a controversial problem. Quaternary International 253: 32–46.

26.

Frankel

(1991) Remains to be Seen: Archaeological Insights into Australian Prehistory. Melbourne: Longman Cheshire.

27.

Fusco

(2014) The Biodiversity of the Holocene Terrestrial Mammal Faunas of the Murray Mallee, South Australia. Adelaide: BSc Biodiversity and Conservation Programme, Flinders University of South Australia.

28.

Fusco

McDowell

Prideaux

(2016) Late-Holocene mammal fauna from southern Australia reveals rapid species declines post-European settlement: Implications for conservation biology. The Holocene 26(5): 699–708.

29.

Gawne

Merrick

Williams

, et al. (2007) Patterns of primary and heterotrophic productivity in an arid lowland river. River Research and Applications 23: 1070–1087.

30.

Gingele

De Deckker

Norman

(2007) Late Pleistocene and holocene climate of SE Australia reconstructed from dust and river loads deposited offshore the River Murray Mouth. Earth and Planetary Science Letters 255: 257–272.

31.

Gliganic

Cohen

May

J-H

, et al. (2014) Late-Holocene climatic variability indicated by three natural archives in arid southern Australia. The Holocene 24(1): 104–117.

32.

Goode

Harvey

(2009) Water resources. In: Jennings

(ed.) Natural History of the Riverland and Murraylands. Adelaide: Royal Society of South Australia Inc, pp.162–177.

33.

Gott

(1982) Ecology of root use by the Aborigines of Southern Australia. Archaeology in Oceania 17(1): 59–67.

34.

Hale

Tindale

(1930) Notes on some human remains in the lower Murray Valley, South Australia. Records of the South Australian Museum 4: 145–218.

35.

Helfensdorfer

Power

Hubble

TCT

(2020) Atypical responses of a large catchment river to the holocene sea-level highstand: The Murray River, Australia. Scientific Reports 10: 7503.

36.

Hill

Durband

Walshe

(2016) Risk minimization and a late holocene increase in mobility at Roonka Flat, South Australia: An analysis of lower limb bone diaphyseal shape. American Journal of Physical Anthropology 161: 94–103.

37.

Hubble

De Carli

Clarke

, et al. (2014) The Lower Murray River’s Mannum Muds: A Holocene Age Lacustrine Deposit in a Bedrock Gorge. American Geophysical Union, Fall Meeting 2014, abstract id. EP13A-3496.

38.

Humphries

Keckeis

Finlayson

(2014) The river wave concept: Integrating river ecosystem models. Bioscience 64(10): 870–882.

39.

Job

Penny

Morgan

, et al. (2021) Multi-stage holocene evolution of the River Murray Estuary, South Australia. The Holocene 31(1): 50–65.

40.

Johnson

(1995) Interactions between fire, mycophagous mammals, and dispersal of ectromycorrhizal fungi in Eucalyptus forests. Oecologia 104: 467–475.

41.

Jones

Roberts

Westell

, et al. (2022) Aboriginal earth mounds of the Calperum Floodplain (Murray Darling Basin, South Australia): New radiocarbon dates, sediment analyses and syntheses, and implications for behavioural change. The Holocene 32: 816–834.

42.

Jones

Morrison

Roberts

, et al. (2017) An analysis of indigenous earth mounds on the Calperum floodplain, Riverland, South Australia. Journal of the South Australian Anthropological Society of South Australia 41: 18–62.

43.

Kemp

Radke

Olley

, et al. (2012) Holocene lake salinity changes in the Wimmera, southeastern Australia, provide evidence for millennial-scale climate variability. Quaternary Research 77: 65–76.

44.

Koungoulos

(2017) Canis dingo and the Australian smaller-fauna trend: A new explanatory model integrating ecological data. Journal of Archaeological Science Reports 14: 38–45.

45.

Krefft

JLG

(1862) On the vertebrated animals of the lower Murray and Darling, their habits, economy and geographical distribution. Transactions of the Philosophical Society of New South Wales 1862–1865: 1–33.

46.

Letnic

Fillios

Crowther

M S

(2014) The arrival and impact of the dingo. In: Glen

Dickman

(eds) Carnivores of Australia. Melbourne: CSIRO Publishing, pp.55–69.

47.

Lewis

Sloss

Murray-Wallace

, et al. (2013) Post-glacial sea-level changes around the Australian margin: A review. Quaternary Science Reviews 74: 115–138.

48.

Littleton

Karstens

Allen

(2021) Murray River Societies in Australia through the lens of bioarchaeology. In: McNiven

David

(eds) The Oxford Handbook of the Archaeology of Indigenous Australia and New Guinea. Oxford: Oxford Handbooks on Line. In press.

49.

Littleton

Petchey

Walshe

, et al. (2017) A preliminary redating of the holocene Roonka burials, south-eastern Australia. Archaeology in Oceania 52(2): 98–107.

50.

Maheshwari

Walker

McMahon

(1995) Effects of regulation on the flow regime of the River Murray, Australia. Regulated Rivers Research & Management 10(1): 15–38.

51.

Margules and Partners Pty Ltd (1990) River Murray riparian vegetation study. Prepared by Margules and Partners Pty Ltd, P and J Smith and Department of Conservation Forests and Lands. Report to the Murray-Darling Basin Commission, Canberra.

52.

Marshall

Cosgrove

(1990) Tasmanian devil (Sarcophilus harrisii) scat-bone: Signature criteria and archaeological implications. Archaeology in Oceania 25(3): 102–113.

53.

Martin

(1999) Overcoming the constraints of emu egg laying season. RIRDC report 99/23. Barton, ACT: Rural Industries Research and Development Corporation.

54.

Martin

(2011) Palaeoecological evidence associated with earth mounds of the Murray Riverine Plain, south-eastern Australia. Environmental Archaeology 16: 162–172.

55.

Marx

Kamber

McGowan

, et al. (2011) Holocene dust deposition rates in Australia’s Murray-Darling Basin record the interplay between aridity and the position of the mid-latitude westerlies. Quaternary Science Reviews 30: 3290–3305.

56.

Mashmedt

(2009) Soils. In: Jennings

(ed.) Natural History of the Riverland and Murraylands. Adelaide: Royal Society of South Australia Inc, pp.50–64.

57.

Mein

Manne

(2021) Identifying marsupials from Australian archaeological sites; current methodological challenges and opportunities in zooarchaeological practice. Archaeology in Oceania 56: 133–141.

58.

Menkhorst

(2009) Blandowski’s mammals: Clues to a lost world. Proceedings and Transactions of the Royal Society of Victoria 121(1): 61–89.

59.

Mulvaney

(1960) Archaeological excavations at Fromm’s landing on the Lower Murray River, South Australia. Proceedings of the Royal Society of Victoria 72: 53–86.

60.

Mulvaney

Lawton

Twidale

(1964) Archaeological excavations of rock shelter No. 6 Fromm’s Landing, South Australia. Proceedings of the Royal Society of Victoria 77: 479–516.

61.

Paton

(1983) An Analysis of Aboriginal Subsistence in the Lower Murray District, South Australia. BA Hons Thesis. Canberra: Anthropology, Australian National University.

62.

Pretty

(1977) The cultural chronology of the Roonka Flat: A preliminary consideration. In: Wright

RVS

(ed.) Stone Tools as Cultural Markers: Change, Evolution and Complexity. Canberra: Australian Institute of Aboriginal Studies, pp.288–331.

63.

Quigley

Horton

Hellstrom

, et al. (2010) Holocene climate change in arid Australia from speleothem and alluvial records. The Holocene 20(7): 1093–1104.

64.

Roberts

(1998) Digging deeper: A reanalysis of the archaeology of Fromm’s Landing, South Australia. Unpublished Honours Thesis, Archaeology Department, Flinders University.

65.

Roberts

; The Mannum Aboriginal Community Association Inc (2012) Ngaut Ngaut: An Interpretive Guide. Intellectual Property Issues in Cultural Heritage (IPinCH) Project, Archaeology Department, Simon Fraser University, Burnaby, British Columbia, Canada and the Mannum Aboriginal Community Association Inc. (MACAI), Nildottie, South Australia.

66.

Roberts

Donald Pate

Hunter

(1999) Late Holocene climatic changes recorded in macropod bone collagen stable carbon and nitrogen isotopes at fromms landing, South Australia. Australian Archaeology 49(1): 48–49.

67.

Robinson

Lang

Croft

, et al. (2009) Flora and vegetation. In: Jennings

(ed.) Natural History of the Riverland and Murraylands. Adelaide: Royal Society of South Australia Inc, pp.206–282.

68.

Russell

(2011) Social Zooarchaeology : Humans and Animals in Prehistory. Cambridge: Cambridge University Press, pp.1–10.

69.

Sackett

(1979) The pursuit of prominence: Hunting in an Australian Aboriginal community. Anthropologica 21: 223–246.

70.

Sheldon

(1994) Littoral ecology of a regulated Dryland River (River Murray, South Australia) with reference to the Gastropoda. PhD Thesis, University of Adelaide, Adelaide.

71.

Shulmeister

(1999) Australasian evidence for mid-holocene climate change implies precessional control of Walker circulation in the Pacific. Quaternary International 57–58: 81–91.

72.

Shulmeister

Kemp

Fitzsimmons

, et al. (2016) Constant wind regimes during the last glacial maximum and early holocene: Evidence from Little Llangothlin Lagoon, New England Tablelands, eastern Australia. Climate of the Past 12: 1435–1444.

73.

Sloss

Murray-Wallace

Jones

(2007) Holocene sea-level change on the southeast coast of Australia: A review. The Holocene 17(7): 999–1014.

74.

Smith

(1978) A note on the fauna from Devon Downs Shelter. Australian Archaeology 8: 19–21.

75.

Smith

(1982) Devon downs reconsidered. Oceania 17: 109–116.

76.

Smith

(2000) ‘The opening chapter of the Romance of excavation in Australia’: Reflections on Norman Tindale’s archaeology. Historical Records of Australian Science 13(2): 151–160.

77.

Smith

Williams

Turney

, et al. (2008) Human—environment interactions in Australian drylands: Exploratory time-series analysis of archaeological records. The Holocene 18(3): 389–401.

78.

Specht

(1984) The Mediterranean climate of Australia and its impact on the structure and distribution of vegetation. Bulletin de la Société Botanique de France. Actualités Botaniques 131(2-4): 225–234.

79.

Thredgold

Roberts

; The River Murray and Mallee Aboriginal Corporation (2017) Stone artefacts and earth mounds at Calperum Station, South Australia, together with a consideration of comparative Murray Darling Basin Data. Journal of the Anthropological Society of South Australia 41: 93–122.

80.

Tibby

Bourman

Wilson

, et al. (2021) A large mid-holocene estuary was not present in the lower River Murray, Australia. Scientific Reports 11: 12082.

81.

Twidale

Lindsay

Bourne

(1978) Age and origin of the Murray River and gorge in South Australia. Proceedings of the Royal Society of Victoria 90: 27–42.

82.

Walker

Thoms

Sheldon

(1992) Effects of weirs on the littoral vegetation of the River Murray, South Australia. In: Boon

Calow

Petts

(eds) River Conservation and Management. New York, NY: Wiley, pp.271–292.

83.

Walker

Madden

Williams

, et al. (2009) Freshwater invertebrates. In: Jennings

(ed.) Natural History of the Riverland and Murraylands. Adelaide: Royal Society of South Australia Inc, pp.283–305.

84.

Walker

Head

Berkelhammer

, et al. (2018) Formal ratification of the subdivision of the holocene Series/Epoch (Quaternary System/period): Two new global boundary stratotype sections and points (GSSPs) and three new stages/subseries. Episodes 41(4): 213–223.

85.

Walshe

(2000) Carnivores, taphonomy and dietary stress at puntutjarpa, Serpent's Glen and intitjikula. Archaeology in Oceania 35: 74–81.

86.

Walshe

(ed) (2009) Roonka: Fugitive Traces and Climatic Mischief. Adelaide: Government of South Australia, Hyde Park Press.

87.

Walshe

(2014) A Taphonomic Analysis of the Vertebrate Material From Allen's Cave: Implications for Australian Arid Zone Archaeology. A Thesis Submitted for the Degree of Doctor of Philosophy. The Australian National University, Canberra.

88.

Westell

Roberts

Morrison

, et al. (2020) Initial results and observations on a radiocarbon dating program in the Riverland region of South Australia. Australian Archaeology 86(2): 160–175.

89.

Westell

Wood

(2014) An introduction to earthen mounds in South Australia. Journal of the Anthropological Society of South Australia 38: 30–65.

90.

White

Saltré

Bradshaw

CJA

, et al.(2018) High-quality fossil dates support a synchronous, Late-Holocene extinction of devils and thylacines in mainland Australia. Biological Letters 14: 20170642.

91.

Wilkins

Gouramanis

De Deckker

, et al. (2013) Holocene lake-level fluctuations in Lakes Keilambete and Gnotuk, southwestern Victoria, Australia. The Holocene 23(6): 784–795.

92.

Williams

Santoro

Smith

, et al. (2008) The impact of enso in the Atacama Desert and Australian Arid Zone: Exploratory time-series analysis of archaeological records. Revista de Anthropología Chilena, Special Publication 40: 245–259.

93.

Williams

Ulm

Turney

, et al. (2015) Holocene demographic changes and the emergence of complex societies in prehistoric Australia. PLoS One 10(6): e0128661.

94.

Wilson

(2017) Holocene archaeology and Ngarrindjeri Ruwe/Ruwar (land, body, spirit): A critical indigenous approach to understanding the Lower Murray River, South Australia. PhD Thesis, Department of Archaeology, School of Humanities, Faculty of Education, Humanities and Law at Flinders University, Adelaide.

95.

Wilson

Fallon

Trevorrow

(2012) New radiocarbon ages for the lower Murray river, South Australia. Archaeology in Oceania 47(3): 157–160.

96.

Wilson

Roberts

Fusco

(2022) New data and syntheses for the zooarchaeological record from the lower Murray River Gorge, South Australia: Applying a ngatji lens. Australian Archaeology 88: 200–214.

97.

Yokoyama

Purcell

Lambeck

, et al. (2001) Shore-line reconstruction around Australia during the Last Glacial Maximum and Late Glacial Stage. Quaternary International 83–85: 9–18.