4
Given the lack of rock shelters, the main archaeological site types on An-barra land are all open. As mentioned in Chapter 2, there are three main types: shell middens, shell mounds and earth mounds (gun-gapula). They are typical of other areas on the coastal plains across the Top End of northern Australia (Brockwell et al. 2011). The sites date from approximately 3,700 years ago until recently. The establishment of sites follows the progradation of the landscape, similar to what occurs in Blue Mud Bay (Faulkner 2013), but not for shell mounds in the Weipa area where coastal progradation slowed during the late Holocene (Shiner et al. 2013) and there does not seem to be a link between site age and distance to the coast (Holdaway et al. 2017; Morrison 2015). On An-barra territory, as described earlier, older sites, like Mu-garnbal, are located inland, while more recent sites, like Guna-jengga and Jurnaka, are located on the current coast (Map 3.1).
As well as following the evolution of the landscape, the faunal assemblages of these sites reflect change over time. As shellfish inhabit specific ecosystems, they are useful indicators of past environmental conditions and subsistence strategies. A change in species indicates a change in environment. For example, a shift from marine to estuarine species reflects a change in exploitation strategies focused on the sea to the mangroves and estuary of An-gartcha Wana.
The oldest site on An-barra land investigated so far is Mu-garnbal, a shell midden dated to c. 3,700 years BP (Table 3.1). It is in Martay territory on the eastern side of An-gartcha Wana, approximately 7 km from the coast (Map 3.1). Rhys compared Mu-garnbal with one of the modern-day coastal sites: “This is an inland outstation for the wet season and is in real contrast to the coastal strand loopers of Lalarrgadjirripa. Fascinating to hear that they are camped on a gun-gapila [mound]. I must go and see” (RJ FN 1973, Bk 14, 73).
Without knowing their age, Betty commented on the differences between the shell assemblages at Mu-garnbal and those on the coast:
A sample taken from this site reflected its surroundings and contrasted with the… coastal middens… the main species being: Anadara [now Tegillarca] granosa (79%), Cerithidea anticipata (8%), Telescopium telescopium (6%), Terebralia palustris (4%), rest (6%)… The presence of mangrove species rarely used by the Anbarra is an important feature of these percentages. The large contribution of Anadara [Tegillarca] granosa also contrasted with Anbarra predation. Tapes [Marcia] hiantina was not recorded in the sample (Meehan 1982a, 167).
The dominance of T. granosa, a marine species, at Mu-garnbal (3,724–3,345 cal BP) is consistent with the site being close to the former coastline prior to progradation. This is also consistent with other T. granosa dominated sites across northern Australia at this time from northern Western Australia across the Top End to far north Queensland (see discussion on change over time below).
The other middens analysed for this study are younger. They date between 1,392 cal BP and recent, in descending chronological order:
These middens were dominated by species different from the ones Betty observed the An-barra collecting. In all cases, Dosinia was the dominant shellfish, except for the youngest site, Ngarli ji-bama, which was dominated by Mactra (Tables 3.8 and 3.19). These middens reflect a pattern of foraging from an environment different, though still marine based, from that observed by Betty in the 1970s, when the dominant shellfish collected was Marcia hiantina (see discussion below) and Modiolus modulaides from the home base at Ngarli ji-bama. There is a similar diversity of shellfish species in both shell mounds and middens.
The shell mounds are dated between 1,356 cal BP and recently. In descending chronological order, they are Muyu a-jirrapa (1,356–867 cal BP), Gulukula (875–465 cal BP) and Yuluk a-jirrapa (636–165 cal BP) (Table 3.1). As stated in the last chapter, the contents of the Muyu a-jirrapa Shell Mound excavation were never analysed as the assemblage was lost in the Canberra bushfires of 2003. However, Rhys’ observation from his 1974 samples was that the mound was dominated by Dosinia (RJ 1974, FN Bk 1, 90). Gulukula was also dominated by Dosinia (Table 3.4) and Yuluk a-jirrapa by Coecella horsfieldii (Table 3.20). Again, the assemblages of these shell mounds reflect shell foraging strategies from a variety of marine environments different from those that prevail at the sites today. The date of 1,356–867 cal BP for Muya A-jirripa Shell Mound makes it contemporary with Muya A-jirripa Coastal Shell Midden (1,392–987 cal BP) (Table 3.1). Therefore, the fact that Dosinia is the dominant shellfish in both Muya A-jirripa Shell Mound and Coastal Shell Midden suggests that occupants of both sites were foraging from the same resource base (see previous chapter).
Although there are numerous earth mounds located on the An-gartcha Wana floodplains, Ji-bena (1,284–790 cal BP) was the only one that was excavated, and an auger sample was taken from Anamanba (796–423 cal BP). The Ji-bena assemblage reflects a change in environmental conditions and exploitation strategies on the An-gartcha Wana floodplains. As Ji-bena is well dated archaeologically, these changes can be related to environmental shifts indicated by the floodplain’s geomorphology, which have also been dated (see Chapter 2). Ji-bena also had the widest range of archaeological remains of all the sites excavated.
The assemblage was dominated by both marine and estuarine shellfish. There were also other faunal material and some stone artefacts (Table 3.7). Over time, the faunal remains indicate that a variety of different environments were being targeted from Ji-bena (Figure 3.14). Shellfish record the foraging in open beaches and mangrove environments. Chelodina rugosa (known as “long-necked turtle” in the Top End), northern brushtail possum (Trichosurus arnhemensis) and northern brown bandicoot (Isoodon macrourus), mark the exploitation of both freshwater swamps and open woodlands (Goodfellow 1993, 48).
Ji-bena now lies 8 km inland and the presence of marine shellfish in its assemblage suggests that the Ji-bena started life as a shell mound, perhaps when the coast was closer. Dosinia juvenilis dominated the shellfish assemblage from c. 1,300 years BP, immediately post the establishment of the mound until 988–609 cal BP (spit 8), when exploitation of this species declined and subsequently disappeared (Figure 3.12). Although they are far fewer in number, other marine species follow much the same pattern but did not disappear altogether. The foraging focus subsequently switched to mangrove shellfish, predominantly Telescopium. While less frequent, they are, with a couple of exceptions, present throughout the history of the mound, peaking after 988–609 cal BP, when the marine species decline dramatically (Figure 3.12). However, they too decline post this peak, perhaps indicating a general reorganisation of foraging strategies away from Ji-bena onto the prograding coastline. Betty notes that Telescopium is a gastropod and, in general, the An-barra did not seem to like them – they mainly collected them when bivalves were scarce or hard to get. There followed a period of abandonment, apart from some sporadic occupation, until 710–333 cal BP (spit 5) when it was reoccupied as an earth mound. At that time, there was an increase of freshwater turtle in the upper levels of Ji-bena. This new foraging strategy coincides with the emergence of freshwater conditions in Balpilja close to the site, where the focus was now on freshwater produce from the swamp (Table 3.10, Figure 3.14).
The sample of Geloina coaxans (mangrove mud whelk) augered from 2.3 m below the surface of nearby Anamanba earth mound (Map 3.1) was dated to 796–423 cal BP (Table 3.1). This date overlaps with the decline of marine shellfish and the increase in mangrove species at Ji-bena, suggesting that Anamanba was occupied originally when the foraging focus had already shifted from marine to mangrove following environmental change associated with coastal progradation.
There is little stone suitable for knapping available on the coastal plains of An-gartcha Wana. The only stone quarry in Gu-jingarliya country is Lawuk a-jirrapa (stone- spear-point-place), a low rocky ridge on the western edge of the An-barra home base of Ngarli ji-bama (Figure 2.1, Map 3.1; Meehan 1982a, 26–7; Meehan & Jones 2005). Ji-bena is the only archaeological site that contains stone artefacts, mainly in its upper levels (Table 3.11, Figure 3.15). Apart from the quartzite artefacts, which were probably sourced locally from Lawuk a-jirrapa, the ground artefacts made from ochre, haematite and sandstone must have been traded in, possibly from sources in Gun-nartpa country to the south (Table 3.12). The lack of stone is notable as it emphasises An-barra reliance on ephemeral raw materials (bone, shell and wood) for their material culture, which was elaborate and multi-faceted (Meehan & Jones 2005). Betty recorded and collected many such items on her visits. This ethnographic collection is now housed in the MAGNT in Darwin.
The An-barra make a distinction between shell middens and shell mounds. Betty commented:
While walking through the bush Anbarra people often commented about a specific site, describing its place in the annual cycle and naming the people who had camped there as well as what they would have eaten. Thus, some sites are designated ‘cold weather camps’, others ‘rain-time camps’, and so on. When people cannot name the people who inhabited a particular site, they usually attribute it to ‘dead men’, meaning that it was inhabited by Gidjingali ancestors whose names they cannot, or do not want, to remember (Meehan 1982a, 166).
In this context, Rhys commented about Guna-jengga as a remembered rrawa, that is, recognised as a camp site not a Dreaming (RJ FN 1973, Bk 23, 3–4). Betty also explained: “The structure of all midden deposits attributed to ‘dead men’ are consistent with contemporary Anbarra customs. In fact, the important home base sites of Ngalidjibama and Kopanga are located on areas containing extensive non-continuous midden deposits” (Meehan 1982a, 166–7).
About the shell mounds, Betty said: “These have an obvious human origin, but the Anbarra say they are not the work of man but of the ‘Dreaming’. The mounds do not occur on the coast but inland at least 1 km on a series of fossil dunes. They are found on both sides of the river in Anbarra, Gulala and Matai territory” (Meehan 1982a, 167). Here, this especially relates to the Gulukula and Yuluk a-jirrapa shell mounds, both of which are Dreaming places.
The relationship between contemporary Anbarra customs, ‘dead men’ sites and ‘Dreaming’ is one of the intriguing problems in the prehistory of the Blyth River area. In the contemporary diet Tapes hiantina [Marcia hiantina] is the dominant species. In the Anbarra ‘dead men’ sites the contribution of this species [is] below 14%. While in the Kula Kula mound it formed but 2% of the shell sample. Conversely, Dosinia juvenilis, which was relatively unimportant in 1972–73 (0.4%), comprised 91% of the shells in this last sample (Meehan1982a, 168).
Again, these observations are supported by the archaeological analysis in the previous chapter.
In terms of structure, the 1972–73 An-barra and ‘dead men’ camps extending along the beach front dunes and leaving extensive intermittent deposits are basically the same, but both differ markedly from the ‘Dreaming’ mounds which imply isolated, confined camping areas. From their geomorphological positions and the fact that they are no longer remembered as the camps of men, we believe that the ‘Dreaming’ mounds are considerably older than the ‘dead’ men sites and are related to the exploitation of a coastline that is different from the present one. Shell mounds similar in size and shape are situated both to the west at Maningrida and to the east at Milingimbi [Map 1.1] and on the Ngan-galala plains, to say nothing of the Weipa and other mounds further afield in Cape York, Queensland. The Gidjingali ‘Dreaming’ mounds may thus belong to this broader class of sites whose ecological and social contexts are not yet fully understood (Meehan 1982a, 168).
Despite Betty and Rhys’ thoughts about the older age of the shell mounds, the radiocarbon chronology established that they are contemporary with the shell middens, the shell midden at Mu-garnbal being by far the oldest site on An-barra land. So, age does not determine whether the An-barra classify sites as belonging to “dead men” or the “Dreaming”; rather, it is the foraging behaviour that created the sites in the first place. The An-barra recognise middens as belonging to people because they still create such sites themselves today, while the shell mounds are the result of foraging strategies from the past that are no longer familiar, hence they were created by mythical beings from the Dreaming (for further discussion, see Brockwell 2013).
The archaeology on An-barra land reflects the changing landscape over the last 3,700 years. The establishment of the sites follows the progradation of the landscape, with the older sites, like Mu-garnbal located inland, to more recent sites, like Guna-jengga and Jinawunya located on the current coast. The predominance of Tegillarca (a marine species) at Mu-garnbal reflects the time when the site was located adjacent to the seashore before the floodplains of An-gartcha Wana silted up, as discussed previously. Betty said of the change in diet over time,
though I do not have detailed figures for their [Matai] shellfish consumption [at Maganbal] during 1972–73 I know that they at least ate shellfish of all the species represented with the addition of Tapes [Marcia] hiantina from Lalarr-gadjirripa and various species of Mytilidae from Moganarra. As for the Anbarra, Tapes [Marcia] hiantina have become more important for the Matai in recent times (1982, 166–7).
The cheniers adjacent to the coast at Jinawunya (1,264–887 cal BP) and Gulukula (1,150–750 cal BP) are the most recent. The archaeological sites of the same names are dated to 875–465 cal BP (Gulukula) and 740–430 cal BP (Jinawunya). The next chenier inland, Lorrkon a-jirrapa, has been dated to 1,839–1,355 cal BP, and its associated archaeological site to 1,433–1,046 cal BP (Table 3.1).
With increasing rainfall in the last 1,500 years, freshwater ponded behind the cheniers and large mangrove swamps became freshwater wetlands. Dating of faunal assemblages from Ji-bena shows that Balpilja Swamp was freshwater from at least 700 years ago and probably as early as 1,000 years BP (Figure 3.14; Brockwell 2013; Chappell & Jones 1999; Thurtell et al. 1999). These dates emphasise the recent nature of the current An-gartcha Wana landscape.
Despite these changes, Betty and Rhys constantly emphasise the continuity in the traditions between the archaeological sites and present-day camps: “In no prehistoric site did I notice any species that were not eaten at some time during 1972–73, but the relative importance of some species was markedly different in these old sites” (Meehan 1982a, 168).
The ‘dead men’ sites represent both continuity with, and contrast between, contemporary Gidjingali practices. The structures of the sites are consistent with modern Anbarra behaviour, the species of shellfish eaten also being the same. The contrast is in the relative importance of these species. In 1972–73 Tapes [Marcia] hiantina were collected far more frequently than any other species contributing 61% of the gross weight of all shellfish collected during the year. For people living at the Kopanga [Aningarra’s Camp] and Gunadjang-ga ‘dead men’ sites this species contributes merely 14% and 3% respectively. Dosinia juvenilis taking its place at both sites and contributing 33% and 60% respectively (Meehan 1982a, 167).
This observation is supported by the analysis of the shell middens in the previous chapter.
Rhys reflected on the Guna-jengga archaeological site:
The interesting thing is that the ‘dinner time camp’ associated with the fish trap – i.e. the one under the 2 bushes on the dune – is in line with the middle and is stratigraphically just slightly above it – or the same level. Thus we have a ‘recent’ but ‘prehistoric’ midden – and one from the ethnographic present being found in the same place and for the same purpose – the old traditions and camp site use continued to the observed present – e.g. the little hearth and midden – deposited a week or two ago (RJ FN 1973, Bk 21, 27).
Both Betty and Rhys note that a modern hearth overlies the existing midden at Muyu a-jirrapa: “About 4.30 pm we went to some work on the eroding dune (coastal midden) at Miadjirripa. It is about 12" thick and comes almost to the present-day surface. Right on top of this midden there was a ‘recent’ hearth with crab, fish and angalidjauwurigia remains visible in it” (BM FN 1974 Bk 1, 90). Rhys commented: “Thus, the modern activity is located directly above the older midden, and its debris (of modern activity) forming its uppermost component” (RJ 1974 FN, Bk 1, 84).
Rhys reflected:
I felt here, walking over the middens, that even if I could not get permission to dig at one, it did not really matter – the unique and important factor here, was that this was one of the few places in Australia (or the world for that matter) where hunters are still intimately tied to their landscape and to that of their distant forebears. Must get a date for the top of the midden – to give a value to the time it takes for H/G culture to forget an old camping ground. F.G. also said that when people want to camp on the middens – the Kula Kula tends to press them away… ‘Kula Kula shift ‘im’ (RJ 1972 FN, Bk 9, 19–20).
The onset of the Holocene climatic optimum (9,000–6,000 years BP) coincided with rapid expansion of populations and establishment of new sites across much of Australia. Williams et al. (2015) argue that more favourable environmental conditions at the end of this period allowed longer periods of residence, resulting in a shift to more sedentary lifestyles in some parts of the continent.
Williams et al. (2015) argue that, despite highly variable climates, over the last 6,000 years in Australia, the increased number of radiocarbon dates from archaeological sites in this period suggests increasing population. After sea-level stabilisation c. 6,000 years ago, sedimentation from rivers filled coastal embayments creating ideal shallow-water conditions for shellfish beds. During that time, there was increased exploitation of coastal resources along the northern and eastern coastlines with sites being mainly shell mounds or middens. Building of conical shell mounds (mainly of Tegillarca but also of other species like Dosinia and Marcia) occurred across northern Australia from c. 4,000 years BP, including Queensland and Western Australia (Clune & Harrison 2009; Cochrane 2014; Harrison 2009; Holdaway et al. 2017; O’Connor 1999; Shiner et al. 2013; Veitch 1999). In the Top End of Australia, the construction of shell mounds ceased abruptly c. 500 years ago (Brockwell et al. 2009). However, it seems they may have continued into more recent times on the Weipa Peninsula in western Cape York, where Morrison (2013, 2014, 2015) argues shell mound building persisted until contact, and on the Abydos Plain in Western Australia (Clune & Harrison 2009; Harrison 2009). Researchers suggest that changes in Tegillarca habitats around 800 to 500 years ago contributed to a notable ecological shift from open beach and mudflats to the mangrove-rich environments seen along much of the coastline today (Bourke 2012; Faulkner 2013; Hiscock 1997). Evidence from the earth mound at Ji-bena suggests similar environmental transformations occurred along the Blyth River during this period (Brockwell et al. 2005). Furthermore, there is widespread documentation of social and economic adaptations across northern Australia at that time. For example, Brockwell (2009) highlights changes in faunal and stone tool assemblages in earth mounds along the floodplains of the lower Adelaide River 600 years ago. In the Torres Strait, significant transformations in demographics, mobility, rituals, seascape construction, social ties and exchange practices are noted between 800 and 600 years ago, as well as between 500 and 400 years ago (Barham et al. 2004, 37; David & Badulgal 2006; McNiven 2006, 9–10). In north Queensland, David and Wilson (1999) describe Ngarrabullgan, a mountain now regarded as ritually dangerous and avoided by traditional custodians, which bears evidence of substantial occupation before 600 years cal BP (Hiscock 2008, 274–5; Hiscock & Faulkner 2006, 219–20).
Bourke et al. (2007) suggest that major climatic shifts between 800 and 500 years ago caused environmental change that led to these shifts in economic and cultural practices around Darwin. Isotope analysis of shell samples taken from Darwin Harbour, An-gartcha Wana, and Blue Mud Bay in eastern Arnhem Land show widespread shifts in sea surface temperatures throughout the region at this time (Brockwell et al. 2013).
Evidence of a possible causal factor for economic and social change in the late Holocene in northern Australia comes from Kuwae in Vanuatu, where there was a major volcanic eruption in the 1450s CE. It was one of the largest volcanic eruptions recorded, profoundly influencing not just the immediate area in Vanuatu but also extending its effects worldwide. Locally, the eruption devastated the former island of Kuwae, leaving behind the islands of Tongoa and Epi, which significantly reshaped the political, linguistic and ecological landscapes of central Vanuatu (Ballard et al. 2023).
On a global scale, the Kuwae eruption is often linked to the onset of the Little Ice Age (c. 1,450–1,530 CE) in the Northern Hemisphere that led to economic hardships across various regions. Current research being conducted at ANU by Stuart Bedford and Chris Ballard seeks to verify both the timing and magnitude of the eruption at its source, located on the islands of Epi and Tongoa (Culture, History and Language 2024). This data may shed light on the chronology of late Holocene environmental change and changes in Aboriginal economic and social strategies in northern Australia.
In 1988, Betty Meehan and Rhys Jones published an edited volume, Archaeology with Ethnography: An Australian Perspective, based on papers presented at the 1983 AAA conference on the same theme. Their main motivation for this publication was their reaction against the processualists who at that time argued a purist view that “archaeology, to be a respectable and autonomous discipline, must be capable of carrying out its entire intellectual processes within its own terms and with no derivation of core ideas from cognate disciplines” (Meehan & Jones 1988, viii).
There were some seminal papers presented in this volume, which are still widely quoted. Meehan (1982a) based her information on her long periods of fieldwork with the An-barra people of the An-gartcha Wana in the 1970s. She defined “dinner time camps” as “small campsites used during the day while people are engaged in hunting trips away from their home bases where people cook and eat food that has been procured up to that time” (Meehan 1982a, 26; 1988b). She argued that dinner time camps could be distinguished archaeologically as mainly being small in size and containing monospecific faunal remains, whereas base camps are larger and much more varied in their assemblages.
More recently, there has been renewed interest in and debate about archaeology using ethnography as an interpretive tool, with an increasing commitment to community archaeology and Indigenous-driven research projects (e.g. Allen & Rowe 2014; Clarke 2002; David et al. 2021, 2024; Guilfoyle et al. 2013; Hiscock 2013; Hiscock & Faulkner 2006; Holdaway & Allen 2011; MacFarlane et al. 2005; McNiven 2016; Molle et al. 2023; Morrison 2014; Ross et al. 2013; Thomas et al. 2023; Urwin et al. 2024; Wright et al. 2021).
Several authors propose that the “ethnographic present” began at least 500 years ago in coastal northern Australia (e.g. Brockwell 2013; David & Badulgal 2006; David & Weisler 2006; David & Wilson 1999; Hiscock 2013; Hiscock & Faulkner 2006; McNiven 2006), with major changes to previous environments, and Indigenous economies from that time onwards similar to those recorded in the Contact period. The importance of this argument is that ethnographic analogy then becomes a valid tool in archaeological interpretation of the past over the last 500 years.
Hiscock and Faulkner (2006) explore the formation of shell mounds along the northern coastal plains, drawing on ethnographic analogy and mythology. Researchers (e.g. Bourke 2005; Clune & Harrison 2009; Cribb 1996; Morrison 2003; Veitch 1999) have relied on ethnographic insights to decode the origins as well as the ritualistic and social aspects of these mounds. However, Hiscock and Faulkner assert: “Attempts to impose historic ideologies and cosmologies on earlier times fail to acknowledge the magnitude and rate of economic and ideological change on the tropical coastline of Australia” (2006, 209). Their argument suggests that between 3,000 and 500 years ago, before shell mound construction halted, foragers operated in a landscape that differed significantly from what we see today. Moreover, the fact that many existing shell mounds now sit in barren, unproductive areas has led many Aboriginal communities to perceive them not as human-made structures, but rather as natural formations or as creations of their “Dreaming” ancestors. Consequently, Hiscock (2008, 272–5) and Hiscock and Faulkner (2006, 216) contend that employing modern ethnographic comparisons is inadequate for understanding shell mounds. They suggest that
historically recorded understandings of the mounds probably emerged only after the termination of the economic and environmental system that created them, as these relic structures were conceptualised by people with transformed economies and views about the land. During the last 800–600 years, the myths about and uses of these sites came to reflect the concerns, perceptions and ideology of historic and proto-historic Aboriginal people (Hiscock & Faulkner, 2006, 217).
The shell mound “Dreamings”, Yuluk a-jirrapa and Gulukula, both on An-barra territory, demonstrate that such is the case. However, the shell middens there also represent relics from a transformed landscape but they are recognised by the An-barra as being humanly constructed “dead men” sites, even though Mu-garnbal is dated to 3,500 years ago. This implies that the situation is more complex than Hiscock and Faulkner have argued as the An-barra today have insight into settlement patterns of previous landscapes. This indicates the value of ethnographic analogy and Indigenous explanations to interpret local archaeological sites dated to the late Holocene.
Margaret Carew (pers. comm. 18 July 2024), the linguist who assisted in the translation of Betty’s text “For the An-barra” into Gu-jingarliya (see following chapter) noted that:
The group name An-barra has long been translated into English as “people of the rivermouth” (cf. Gurrmanamana et al. 2002). This name is built on the body part word barra. It sounds the same but has a different meaning from barra “monsoon wind” and is a different part of speech in grammatical terms.
There is a class of words in Burarra that are “shape classifiers” (e.g. barra “base part”, jawa “throat”, bama “head”, gochila “belly” etc.). These words are often derived into other words, where the semantics of shape and part/whole relations are relevant. For example, jin-gochila is a word built on gochila “belly”, and means “mother” or “mother’s clan group” (derived by adding the female prefix jin- onto the base word). An-barra is a word built on barra “bottom, underneath part”. We get this by adding the masculine prefix an- onto the base word barra (an- can also agree with a group of people). It literally means “people associated with something that is underneath”.
As a body part or shape classifier word, barra refers to the bottom of something and also the underneath or submerged part of something. For example, barra gu-jirra “the depths of a billabong”, and gu-barra “in the deep water” are some of the constructions with this term where the “underwater” meaning is evident.
It was curious why barra is recruited for “rivermouth”, given that another body part jawa “throat” is more commonly used for this part of a river. From reading Betty and Rhys’ work, as well as that of anthropologist Geoff Bagshaw, it was realised that An-barra, as a group name, refers to the fact that these people own country that is submerged by the sea and includes sites in the rivermouth of An-gartcha Wana.
As indicated previously, the An-barra have an intimate knowledge of their landscape and are acutely aware of changes over time. The meaning of An-barra demonstrates that knowledge.
In eastern Arnhem Land, a multi-disciplinary study using geomorphology, archaeology, anthropology and linguistics determined that toponyms from Yolngu country around Blue Mud Bay (Map 1.1) were established up to 3,000 years ago and that the current social system may have considerable time depth (Morphy et al. 2020). This may also be the case with An-barra territory.
White and Parsons (1973, 1976) carried out population genetics research in Arnhem Land in the 1970s that is relevant to An-barra history in that area. White provided a summary of his relevant findings below:
… using fingerprints and some other single-gene characters, the four small language groups around Maningrida, including the Anbarra or Burarra, are distinctive and can be distinguished from other surrounding populations. As they are genetically distinguishable, so too are they by language. Linguists have shown that these groups belong to four language families, the highest level of language differentiation in Australia. The fact that there is little genetic differentiation within this small cluster indicates that there was intermarriage across the language family boundaries but not inland. The fingerprint characters, particularly total ridge count, help resolve the relationship of the Yolngu of northeast Arnhem Land to the Burarra. The latter have been considered by some to belong to the Yolngu complex. However, it is clear from [this] study that there is marked genetic differentiation between the two, when the Yolgnu are considered as a whole, suggesting reproductive isolation over a significant period of time, or different origins. The genetic results also suggest that the Anbarra with their immediate neighbours have differentiated genetically either through reproductive isolation, that is, highly localised marriage distances over many generations with evidence of very limited marriages away from the coast, or demographically unstable small populations, that is numbers have fluctuated markedly through time, although it is possible that both these mechanisms have operated across many generations. Certainly, the very favourable subsistence base, with access to both marine and terrestrial (plants and animals) resources, provides opportunity for population stability through sufficient time to generate both the observed linguistic and population genetic difference in this part of coastal Arnhem Land (pers. comm. 2023).
The An-barra Archaeological Project aimed to explore pre-colonial settlement and subsistence patterns in central Arnhem Land, considering the significant environmental changes on northern Australia’s coastal plains during the mid to late Holocene. Additionally, the archaeological findings were analysed alongside ethnographic insights from Betty’s book, Shell Bed to Shell Midden (Meehan 1982a), highlighting both connections and disparities while seeking to understand the underlying reasons.
The site types on An-barra territory – shell mounds, shell middens and earth mounds – are typical of other areas on the coastal plains across the Top End of northern Australia. The earliest site so far recorded, Mu-garnbal, dated to 3,700 years, is located inland close to the palaeo-estuary of An-gartcha Wana. Establishment of sites generally over the next few thousand years followed the evolution and progradation of the landscape with the younger ones closest to the present-day coast. Like the archaeological record from other regions in late Holocene northern Australia, Ji-bena records major changes in An-barra subsistence patterns c. 800–500 years ago suggesting major environmental change at that time. A possible causal factor is the massive volcanic eruption in Kuwae, Vanuatu, c. 1,450 CE, which would have disrupted climate and sea surface temperatures, creating new environments. Change and continuity are both themes of the research. Foraging patterns recorded by Betty and Rhys in the 1970s differ from those found archaeologically, in terms of shellfish species harvested. Settlement patterns differ in that large conical shell mounds are no longer created and low, diffuse shell middens are the more recent archaeological expression. Despite changes in foraging behaviour, the shellfish species eaten today are the same as those eaten in the past, just the proportions differ, and “dead men” sites are similar to shell middens today and recognised by the An-barra as such.
The ethnographic and linguistic evidence indicates that the An-barra are well aware of changes over time to their landscape. Despite its recent nature, it is covered in place names emphasising An-barra adaptability and the continuity of their social traditions in the face of change. Continuity is also emphasised by the evidence of genetic differentiation from neighbouring groups that suggests highly localised marriage over many generations. In conclusion, ethnography and Aboriginal world views can add depth and detail to previous lifeways that archaeology cannot, and both complement and enrich archaeological interpretation.