Needles in Haystacks: Why Finding Massacres in the Australian Archaeological Record is So Challenging

By Lynley Wallis

The publication of a map showing frontier massacre sites across Australia by eminent historian Professor Lyndall Ryan has generated enormous public interest in the past 12 months (Figure 1). The map represents a wonderful resource with substantive research behind it. However, users should be aware that the “massacre locations” shown on it (and most other similar maps), are mostly “best guess estimates” of where such events took place. Most of the locations have not been “ground-truthed”, a term which archaeologists frequently use to indicate that someone has been out to the site, ideally with knowledge holders (be they, for example, Aboriginal traditional owners or land holders) and confirmed through either oral testimony or other means, such as archaeological or geophysical survey, exactly where the event took place. Ground-truthing is the best way to know precisely where to locate any place on a map.

Figure 1 Massacre map by Lyndall Ryan.

Oral histories are handed down through repeated retelling and, in the repetitious retelling of the story to ensure its fundamental components are retained, details are often lost, such as exactly where or exactly when something happened. Elements of the story are almost always discarded with the passage of time, and rarely were detailed maps drawn or geographic details written down to accompany the verbal handing down of knowledge (see an earlier post in which we touch on this here). The dispossession of land that followed European invasion meant that many Aboriginal people have had limited opportunities to visit the places recounted in stories told them by their Elders, and thus many stories have been handed down about places when the people were not at that particular place. As such, the visual cues that would usually help a story teller remember certain details of the story are sometimes missing.

As you might well imagine, when the record indicates that a massacre event took place on “ABC Creek”, and ABC Creek is 100 km long, placing a single “dot on a map” to indicate where the massacre occurred involves a level of guesswork and intuition. This is not to denigrate the efforts of Professor Ryan — far from it — her online map serves as a valuable resource for people wanting to learn more about Australian history*. But it does have some limitations.

Beyond the inherent challenges associated with mapping such events based on historical and oral sources, other factors are at play in mapping massacres, including whether they might be preserved in the archaeological record at all. If they aren’t, then no amount of ground-truthing will ever be able to locate them.

Hence it’s useful forpeople to learn something of the complexities of archaeological site formation to appreciate just why physically locating colonial massacre sites in the landscape is such a difficult exercise, even with the benefit of historic or contemporary massacre maps, and why physical evidence of such events is not routinely likely to survive to the present.

Various factors will influence whether physical evidence of a massacre will be preserved in the archaeological record, such as:

  • Where the event took place;
  • The relative number of victims and the spatial nature of the event(s);
  • Who the perpetrators were and the mode of killing they adopted; and,
  • How the victims’ bodies were treated after death

Drawing heavily on an academic paper that my colleague Mirani Litster and I wrote in 2011 (a copy of which can be found here), and another by Bryce Barker (a copy of which can be found here) I explore some of these issues in the rest of this blog post.

Where massacres occurred

Unsurprisingly, massacres tended to be carried out in the more isolated parts of the frontier rather than close to larger European settlements. As townships grew in size and local Aboriginal populations declined (through deliberate killings, death by disease or their tactical retreat to safer country beyond the frontier), the possibility of hostile interactions occurring in close proximity to townships became slimmer. This meant there was less chance there would be disapproving witnesses around to report (either in verbal or written form) what they had seen. Consequently, deaths were more likely to occur on the less regulated frontiers of newly established pastoral and mining districts, away from large European population centres.

After the Myall Creek massacre in NSW in 1838, ‘European’ offenders became far more circumspect about talking or writing openly about killings in which they had been involved. This particular massacre involved a group of stockmen who rode into a pastoral station and killed most of the Aboriginal inhabitants. Word of the killings soon became common knowledge and the event resulted in a trial in which 7 of the 11 stockmen were found guilty of murder, and hanged for their crimes on 18 December 1838. This was the first time that the killing of Aboriginal people resulted in the offenders being punished, and ultimately led to the emergence of more clandestine — rather than overt — acts of violence against Aboriginal people.

The remote location of many massacres not only reduced the chance of offenders being witnessed or caught, but also increased the time that might elapse between the event itself and the possible discovery of the crime scene, either in the immediate aftermath or by modern researchers. Much of Queensland, particularly those parts where pastoralism is the norm, remains isolated and remote even today. Huge swaths of country are typically rarely visited by more than a handful of people in any one year, and thus finding specific locations within them, at which massacres occurred decades ago, is especially challenging (Barker 2007:9).

Nevertheless, some exceptions to this are likely to exist in specific cases where there was either (a) some form of documented investigation of the event soon after its occurrence**, or (b) where strong oral testimony by people (often, but not always, Indigenous) with an intricate awareness of the landscape, has survived.

The relative number of victims and the spatial distribution of events

While we will never know how many people were killed on the frontier***,  knowing something of the relative number of victims and the spatial nature of killings is important for understanding the chances of material evidence of such events surviving in the archaeological record.

As with most archaeological site types, evidence that is clustered in a discrete location, as opposed to being thinly spread over a greater expanse, and evidence that survives in a greater amount, means a greater likelihood of its preservation, survival, detectability and recovery.

Massacre events with high victim counts and/or large amounts of ordnance being expended will theoretically have the most potential for archaeological discovery. However, many massacres were in the form of punitive expeditions, which involved killing small numbers of people, sometimes over large distances [in what Barker (2007:10) called opportunistic ‘hit-and-run attacks’]; hence they have very slim chances of being incorporated into the archaeological record. Further, the archaeology of such small-scale, short-lived events is also strongly affected by the treatment of victims after death; this is considered in more detail below.

Perpetrators of massacres and specific modes of killing adopted

Massacres were enacted by pastoralists, farmers, travellers, miners and other civilians —these all being people (generally men) whose livelihoods required them to reside at the furthest extents of European expansion. Massacres were also, of course, carried out by the Native Mounted Police (NMP) force raised by the colonial Government explicitly to protect those living and working on the frontier.

The environment of fear in which colonists lived cannot be underestimated: it would have largely been an extremely foreign and threatening environment (Heather Burke has written about this before on our blog here). Visual depictions of violent confrontations between settlers and Indigenous people by colonial artists (such as shown in Figure 2) fed the anxiety that settlers felt about the frontier, contributing to the general sense that settlement at the colonial boundaries quite literally meant putting one’s ‘life on the line’.

Figure 2 Conflict between settlers and Indigenous people, possibly by James Bonwick. NLA, Rex Nan Kivell Collection: NK1300.

Who the perpetrators were, their level of fear, and their concerns about being caught, has implications for the methods they used to carry out killings, and therefore for the type of evidence we could expect to find preserved archaeologically. The main killing methods included firearm injury, poisoning, and to a lesser degree stabbing (or bayoneting) and blunt force trauma. With killings of non-Aboriginal people, it was usually spearing or the blow of a waddy or boomerang that was the ultimate cause of death.

Potential evidence resulting from these types of killings likely to remain at the scene are either (1) consumable components of the instruments utilised (e.g. bullets or bullet casings), as European weapons themselves are unlikely to have been discarded (though typically spears remained in victims’ bodies or nearby according to many written accounts of non-Aboriginal deaths), or (2) the victims’ skeletal remains.

Even though Tony Pagels has written previously about the ordnance and weapons issued to the NMP (e.g. here and here), there will be many instances whereby specific ordnance supplies wouldn’t have been used, thereby making it difficult to unequivocally comment on who was present at a site based solely on the discovery of a particular type of bullet or casing. Further, as much of the ordnance issued to the NMP was also available to the civilian population, evidence for its presence in a particular location may do little more than merely indicate that guns were involved.

Despite certain television programs suggesting otherwise, the cause of death can be difficult to determine from the analysis of bones, since death often involves trauma to soft rather than hard tissues. Thus, if a victim was shot, stabbed or suffered blunt force trauma, the impact would need to have intercepted bone to be apparent in their skeletal remains.

Even if bones with impact trauma are recovered, tying the death to a specific perpetrator can be challenging. One exception to this might be the presence of Aboriginal skeletal remains showing evidence of gunshot trauma. Such a finding would point strongly to either non- Indigenous perpetrators or NMP troopers, since few Indigenous people had access to such weaponry. Archaeologist Mike Rowland (2004) convincingly described a case related to the massacre of 7 or 8 Woppaburra men from the Keppel Islands in which one skull had entry and exit wounds from a low-velocity bullet. This information, when combined with that available from historic accounts, presented a compelling case for these men being victims of a massacre, built on escalating lines of evidence.

The opposite to this is the finding of European or Chinese (the latter of whom were prolific on many of the goldfields of Qld) remains with portions of a spear embedded in them. While it is not unknown for Aboriginal remains from the pre-colonial period to be located with such evidence (such as Narabeen man), thus far we don’t know of any such non-Aboriginal remains, though there are many written accounts of such deaths (as well as those caused by blunt force trauma from blows by weapons) during the colonial period.

Another form of frontier killing, particularly after the Myall Creek massacre, was poisoning, perhaps the most well-known instance of which was the poisoning of dozens of Aboriginal people at Kilcoy station in 1842. Pastoralists, shepherds, miners and surveyors routinely faced the prospect of their provisions being stolen when they were absent from their huts. One response was to lace flour with arsenic or strychnine in the hope that the death of the thieves would not only prevent them from re-offending, but that others would “learn a lesson” (Foster et al. 2001:82), though this sometimes resulted in retaliation by survivors:

MAYTOWN, June 10. The body of a Chinese miner, working about three-quarters of a mile from his hut on the left-hand branch of the Palmer River, was found on Thursday last a few yards from where he had been working. The body was covered up with stones, the head battered in, and presenting a most ghastly sight. It seems the unfortunate man had been missing several days, his mates reported the matter to the police, and stated they believed the blacks had killed him. From marks on the body and head the weapon used appears to have been a tomahawk. Yu Kee is the murdered man’s name, and he was well-known to Europeans. The motive for the crime is supposed to have been revenge. It is said that deceased or some of his mates had poisoned a quantity of flour and left it so that the blacks could easily obtain possession. As a matter of fact one died and many became very sick. (Brisbane Courier, 21 June 1895, p2)

However, the likelihood of any evidence for death by poisoning appearing in the archaeological record is minimal. This is because soft tissues of victims generally have to be well preserved for the evidence of the poison to survive in them. In Australia, human remains are almost exclusively “skeletonised”, with soft tissues generally never surviving owing to the highly seasonal weather patterns we experience. Evidence for poisoning as the particular cause of death of victims is therefore unlikely to be found in any human remains located through archaeological investigation.

Treatment of victims after death

A further factor affecting the potential archaeological record of massacres is the manner in which victims’ remains were treated after death: these are what archaeologists refer to as “natural and cultural site formation processes” (Schiffer 1987).

If victims aren’t buried soon after death, natural transformation processes, including scavenging, cause their bodies to be disarticulated and scattered across the landscape. You only have to drive down a highway in outback Australia to see how quickly roadkill disappears through the efforts of scavengers and weather.

Cultural transformation processes, such as deliberate burial (either in mass or singular graves), attempts at destruction of evidence (such as through burning the victims’ bodies), and traditional ceremonial funerary practices, would also have played a role in affecting whether skeletal material will survive.

Burial of victims was one obvious option, sometimes in mass graves:

… I found a grave into which about 20 [Indigenous people] must have been thrown. A settler taking up a new country is obliged to act towards them in this manner or abandon it. (Black as cited in Kiddle 1962:121)

The larger the grave, the greater the possibility of it being detected today using such techniques as remote sensing, geophysical prospection or geochemistry. While the latter methods are technologically possible, they remain almost impossible in reality in Australian contexts without the existence of other forms of evidence to indicate initially where such a grave might exist (such as oral testimony or a historical map).

Even if graves are found, it needs to be ascertained whether they were of victims of a massacre since, as discussed above, the mode of killing might not be immediately apparent. However, even without necessarily locating ordnance or other evidence of violent trauma on the bones, the discovery of mass graves of Aboriginal people itself in remote locations would be highly suspicious. The number and patterning of the skeletal remains within such graves could also tell us something about their origin. For example, victims of violence might be expected to be interred in a haphazard fashion, with little care taken.

Another element of the treatment of victims after death is whether Indigenous people had the chance to inter victims traditionally after death. If this did in fact happen, not only will the evidence of any massacre be limited, because it will no longer be in situ at the location of the murder, but the bodies will appear to be traditionally interred; unless evidence for non-Indigenous trauma can be found on the bones or strong oral testimony exists, identifying them as massacre victims will be near impossible. However, it is highly likely that survivors, after witnessing such violence, may not have been able to recover the bodies of victims through fear of encountering a similar fate. And in some places there simply may not have been enough survivors to provide the appropriate death rites.

While burial was one option available to perpetrators to minimise their risk of exposure, they may have been disinclined to exert the necessary physical labour required for such an undertaking and, in very remote locations where outsiders were unlikely to stumble across the evidence, burial might well not have occurred. Certainly in some instances squatters complained that bodies of NMP victims were not buried, such as the brothers John and Alexander Mortimer of Manumbar Station on the Burnett River who complained to ‘the Officer in command of the Party of Native Police, about the “blacks you left dead on our run” , entreating him to either “bag or bury” the victims as “it is far from pleasant for us to have the decomposing remains of four or five blackfellows lying unburied” (Sydney Morning Herald, 25 March 1861, p2). Furthermore, the racist views of some settlers with regard to the ‘non-humanity’ of their victims meant they may have not felt it was ‘appropriate’ or ‘necessary’ to afford Aboriginal people the same treatment in death as they would their Christian counterparts.

In contrast to burial, the burning of bodies was a less labour-intensive though more callous practice (i.e. when being undertaken to hide a crime, rather than as part of an established respectful cultural tradition) and appears from various historical accounts to have been common. Such an approach to destroying the evidence has several archaeological implications. Human bones are typically not fully destroyed by burning and, ironically, cremated bone survives better in buried contexts than unburnt bone. Hence, while burning reduces the quantity of skeletal evidence, it also increases the possibility of any remaining bone surviving archaeologically. However, despite this, the ability to determine the origin (human or non-human) of burnt bone is very difficult, since it is often highly fragmented and individual types of bones are hard to identify. Examples of this are known from contemporary accounts on the frontier, such as:

William Bales, of Rosebank station was murdered by two blacks about 20 miles miles from Brighton Downs. The date of the crime is not known. The murder was witnessed by two gins, who state that the body was burnt The scalp and teeth only have been found. (Darling Downs Gazette, 16 June 1886, p3)

Because of the factors I have discussed in this post, evidence of massacre events in the archaeological record is unlikely to be found through random chance. And archaeological evidence alone is unlikely to provide a line of evidence sufficiently robust to silence critics who doubt the ubiquitous occurrence of such violence on the frontier. Litster and Wallis (2001:114) suggested that for this reason it is critical to adopt “a hierarchy of documentary sources, oral histories, material cultural evidence and skeletal remains” to allow for massacre locations to be determined at a ‘possible’, ‘probable’ or ‘highly likely’ level.

In closing it’s worth remembering the following saying: an absence of evidence isn’t evidence of absence, in this case with respect to whether massacres occurred or not. Just because we may not be able to physically pinpoint the exact location of a colonial-era massacre today does not mean that such events did not occur. In the end it should not matter whether we can identify exactly where a frontier killing took place; instead, the importance should lie in the acknowledgement that such killings did indeed happen, and contemporary truth telling is a practical means of healing some of the past wrongs and moving forward together.


*And that we hope will be extended by the database we have built on this topic that will be launched in December 2019.

**The Irvinebank massacre is an excellent example of the first, and we will explore that in a future blog post.

*** Though see estimates made by researchers such as Broome (2003). This is not at all to invalidate the emotions associated with the death of even a single individual.


Barker, B. 2007 Massacre, frontier conflict and Australian archaeologyAustralian Archaeology 64:9-14.

Broome, R. 2003 The statistics of frontier conflict. In B. Attwood and S.G. Foster (eds), Frontier Conflict: The Australian Experience, pp.88-98. Canberra: National Museum of Australia.

Davidson, I., H. Burke, L.A. Wallis, B. Barker, N. Cole and E. Hatte 2018 Connecting Myall Creek and the Wonomo. In J. Lydon and L. Ryan (eds), Remembering the Myall Creek Massacre, pp. 100-111. Sydney: NewSouth Publishing.

Foster, R., Hosking, R. and Nettelbeck, A. 2001 Fatal Collisions: The South Australian Frontier and the Violence of Memory. Wakefield Press: Adelaide.

Kiddle, M. 1962 Men of Yesterday: A Social History of the Western District of Victoria. Melbourne University Press: Melbourne.

Litster, M. and L.A. Wallis 2011 Looking for the proverbial needle? The archaeology of Australian colonial frontier massacres. Archaeology in Oceania46:105-117.

Rowland, M.J. 2004 Myths and non-myths: Frontier ‘massacres’ in Australian history – The Woppaburra of the Keppel Islands. Journal of Australian Studies81:1-16.

Schiffer, M.B. 1987 Formation Processes of the Archaeological Record. Alberquerque: University of New Mexico Press.

4 thoughts on “Needles in Haystacks: Why Finding Massacres in the Australian Archaeological Record is So Challenging

  1. Thanks again for another interesting article . there are a couple of massacre sites around mt Isa that have the potential to be located by finding the spent shell casings as the documentated locations are at specific waterholes . for example , paroo waterhole , the waterhole above the junction of cabbage tree creek and the Leichhardt river and the only waterhole on slaughter creek .

Leave a Reply