Methods of defining rurality

In the field of rural criminology, analysis of domestic violence has expanded well beyond its initial American research base, now extending globally (Buzawa and Buzawa Reference Buzawa and Buzawa2017; Carrington et al. Reference Carrington, Hogg, Scott, Sozzo and Walters2018; DeKeseredy Reference DeKeseredy2021). Yet surprisingly few works use clear-cut criteria to differentiate locations (DeKeseredy et al. Reference DeKeseredy, Hall-Sanchez, Dragiewicz, Rennison and Donnermeyer2016). When Weisheit, one of the field’s founders, conducted a review of literature on rural crime in 2016, he found the most common approach was “not to define rural at all” (Weisheit Reference Weisheit2016: 10). Australian rural criminologists have observed that rurality is a “considerably more unstable, diverse and fragmented phenomenon than is commonly perceived. There is no absolute and definitive distinction between either the conceptual or geographic boundaries of urban and rural areas, or metropolitan and non-metropolitan” (Barclay et al. Reference Barclay, Donnermeyer, Scott and Hogg2007: 2). Nevertheless, leaders in this field acknowledge the need to draw a line somewhere, starting with jurisdictionally specific definitions. In Australia, criminologists currently use the concept of “degrees of remoteness” to plot the distribution of IP homicide in inner and outer regional, versus remote and very remote areas of Australia (NSW Domestic Death Review Team 2020). However, these categories are not transhistorical since the geographical boundaries and meanings of remoteness change over time. To respond to this issue, our study turned to historically specific definitions of the “character of settlement,” derived from Australian historical demographic sources (Australian Bureau of Statistics 1911).

The analysis of rural versus urban crime in Australia has relied on state capitals to stand for “urban” settlements (Nagy Reference Nagy2021; Nagy and Piper Reference Nagy and Piper2019; Piper and Finnane Reference Piper and Finnane2017a). Ironically, this approach highlights urbanity, since it describes the balance of the state (in NSW, over 300,000 square miles) as “rural” territory. A related problem is the homogenization of communities outside the capitals, which makes it difficult to assess the impact of diverse social organization forms on the outcome of lethal domestic violence prosecutions. In NSW, Sydney was consistently the state’s metropolis, but across the early twentieth century, the state’s population was distributed between smaller cities, large regional towns, plus villages, farms, missions, and outback stations. Previous studies have also used the location of trials, rather than offenses, to categorize cases geographically. This tactic further clouds the significance of rurality. For example, by applying the trial location method, the Cobcroft case would be coded urban, since the trial was held in Sydney. Categorizing cases according to the locales where IP homicides took place is more labor intensive, since it requires close analysis of police gazettes and news coverage, but it provides a sharper tool to evaluate the significance of crime location on the prosecution of IP murder.

Rural and urban New South Wales in the early twentieth century

The foundation of New South Wales through the transportation of criminals skewed the development of its population in the nineteenth century, and the White Australia policy, implemented after Federation, put a distinct racial stamp on the state’s demographic development in the twentieth century. The 1901 Immigration Restriction Act, in effect until 1958, ensured that the Commonwealth’s tiny proportion of Asian, Pacific Islanders, and other non-whites would dwindle even further, and policymakers widely assumed that “full-blood” Aboriginals would ultimately become extinct (Jupp Reference Jupp2002; Kingston Reference Kingston2006). At Federation, the population of NSW (1.3 million) was by far the largest of the Australian states, and its residents comprised over one quarter of the nation’s population (Australian Bureau of Statistics 1911). In 1901, Sydneysiders comprised more than one-third of the state’s population, and 10 years later, they outstripped Melbourne’s population, the next-largest city (Spearritt Reference Spearritt2000). As Sydney continued to grow, reaching close to two million by 1955, so did the nonmetropolitan population of NSW, which reached 1.5 million.

Over the first five decades of the century, the proportion of people living in Sydney declined to 20 percent of the state’s population, while several smaller cities grew. The largest ones (Newcastle, Wollongong, and Broken Hill) were based on mining and later steel production (see Figure 1). In the more fertile regions of the state, the dairy industry, fruit growing, and mixed farming sustained the development of several large towns of between 15 and 20 thousand residents. The leading contributor to the state’s primary production economy was the pastoral industry, with its small workforce scattered on stations (Kingston Reference Kingston2006). The population density of Sydney was almost 18,000 persons per square mile by the 1950s, whereas there was only one person per square mile in the state’s vast “Western Division,” where small towns and hamlets, plus a sprinkling of missions and outback stations retained a foothold (Carver Reference Carver1955: 221). Thus, the demographic composition of NSW in the first half of the twentieth century can be divided broadly between Sydney and the remainder of the state for rough comparisons, as earlier studies have done. However, the nonmetropolitan population comprised a complex patchwork of settlements and social settings, which must be appraised to analyze the impact of rurality on prosecutions of IP murders.

Figure 1. Map of NSW – showing major cities, plus locations of women’s rural IP homicides and trials, 1901–1955.

The criminal court system in NSW reflected the state’s uneven development and settlement, and this led to most trials taking place in the state’s largest population centers. The Supreme Court had exclusive jurisdiction over capital cases, and all criminal trials held before the Court were heard by a jury of 12 men (Woods Reference Woods2018). The Central Criminal Court (where Margaret Cobcroft was tried) was located in Sydney’s CBD, but the state’s justices also traveled on circuit to preside over trials, usually in major towns and small cities. That changed in 1912, when an act was passed to allow the Supreme Court to hear cases in “any town or place,” as proclaimed by the state Governor (New South Wales, Supreme Court, and Circuit Courts (Amendment) Act 1912, no 9 sec. 5 (1)). This law made it possible for persons from sparsely settled places, unlike the metropolis, to be called before coronial inquests and to be tried by jurors who may have been familiar with the principals, their relations, and the district in which a homicide took place. But no matter where defendants were tried, one crucial factor was consistent in NSW prior to the 1970s: the exclusively masculine cast of decision-makers who determined the outcome of capital cases.

Sources and methods: Identifying IP homicides in NSW

Systematic historical analysis of criminal prosecutions past the convict period was difficult in Australia prior to the public release of data gathered by Mark Finnane’s Prosecution Project (PP) (Finnane Reference Finnane2016). This searchable resource, which compiles data from undigitized court, police, and coronial records, has greatly facilitated criminal justice research on twentieth-century NSW, where the archival record for the prosecutions is patchy, and no body of transcripts or postconviction processing exists. For each prosecution, the database includes the offender’s name, sex, offense, plea, trial location, date of trial, verdict, judge, and outcome. With access granted to the full PP database for NSW, we were able to record every murder prosecution and its outcome from 1901 to 1955.

Because the PP database was not set up to provide information on victims, we required further sources to distinguish between different types of victim–offender relationships. To accomplish this, we linked cases from the original database to digitized historical coverage of inquests, trials, and police gazettes, accessed through the National Library Australia’s TROVE research tool (Oats and Bushnell Reference Oats and Bushnell2017). This method allowed us to populate the raw database with supplementary data: the name, gender, ethnicity, and age of every offender and victim, in addition to the type of weapons or methods used to kill, and the location of the offense. To categorize the nature of the relationships between the offenders and victims, we relied primarily on news accounts of inquests and trials. Following the standard set in recent criminological work on IP violence (Sutton and Dawson Reference Sutton and Dawson2021), we included every case for which we found evidence that the victim was a married or de facto partner, a former partner, or a known sexual or romantic contact.

To select illustrative cases for qualitative analysis, we analyzed news coverage of each IP murder prosecution of a woman through TROVE’s resources. This method complements the more common approach in histories of gender and homicide, based on highly politicized murders of the past. Instead, we used the patterns revealed through our quantitative analysis to ensure that our chosen cases illustrated key features identified in the majority of IP murder prosecutions. This approach also revealed that unreported cases, and those that did not register politically nevertheless stirred considerable public interest at the time of trial. For example, the Cobcroft case generated 65 news stories, despite the trial opening and closing in less than a day. TROVE provides access to scores of local papers, which often reprinted accounts of murders tried elsewhere, using wire services, but the descriptive headlines editors concocted (such as “Fatal Family Feud”) expressed contemporary characterizations of defendants, victims, and crimes. The reporters who sat in court produced unofficial transcripts, written in shorthand, but they also commented on the nature of the Crown’s and defense’s questioning, the demeanor of the judge, and the size and composition of the public drawn to the case. Australia’s lowbrow papers (notably Truth) were drawn to domestic tragedies, and they reported any hint of sexual impropriety in salacious detail. This qualitative evidence is essential to remind us that Margaret and Henry Cobcroft were not data points – they were real people, entangled through the law’s disposition of lethal acts of domestic violence. Yet, analyzing quantitative data is equally essential, to reveal that his death and her acquittal were components of a larger array of crimes and outcomes in early twentieth century NSW.

Patterns of rural and urban IP prosecution outcomes

To analyze our supplemented database, we began by seeking out patterns in IP murder prosecution outcomes. To code crime locations, we devised three subcategories. We defined Sydney and the state’s biggest cities of 20,000 or greater (Newcastle, Wollongong, and Broken Hill) as urban; settlements of 1,000–19,999 were defined as towns, and we categorized hamlets, stations, and missions with populations less than 1000 as rural. We also referred to historic population figures and settlement designations to determine the crime locations, drawing on the NSW Yearbooks (1904–1955) and the Commonwealth federal census population tables. Importantly, this approach defines locations historically, which entails using the date of crime to code for location, rather than relying on current-day designations of settlements. For instance, some of Sydney’s present-day outer suburbs (such as Wilberforce) were hamlets a century ago (Australian Bureau of Statistics 1911). The coverage of trials in historic newspapers provided further evidence of the character of homicide locations, which we used to supplement census definitions. The ways in which Australians distinguished between rural and regional areas and “the bush” were expressions of local culture, which shifted over time. As Hogg and Carrington caution, rural criminologists must take their cues from the “discursive construction and effects of crime and violence in rural spaces” (Hogg and Carrington Reference Hogg and Carrington2003: 300).

Using descriptive statistics and cross-tabulations, we determined patterns and trends by evaluating the role of rurality and the gender of accused offenders on the outcome of IP murder trials. Due to the modest number of cases in our dataset, it was unfeasible to conduct complex statistical tests. However, our descriptive tabulations of the baseline patterns of IP murder prosecutions provided the best data to inform our selection of illustrative cases.

Table 1 provides the distribution of IP murder prosecutions in NSW between 1901 and 1955 by crime location. Overall, there were 221 cases of IP murder prosecuted during this period: 165 male and 56 female defendants. For both males and females, just under half of their offenses occurred in nonurban locations (defined here as towns, hamlets, or stations with populations less than 20,000). This distribution of IP murders in towns and remote areas was disproportionate to the distribution of the state’s population. Prosecutions for lethal domestic violence were more common in nonurban locations, which supports recent understandings of the greater prevalence of IP homicide in rural and remote regions. Nearly half (104 of 221) of the IP murders in this period occurred outside of major cities, while on average only 30 percent of the NSW population lived in nonurban areas (Australian Bureau of Statistics 2016).

Table 1. Intimate partner murders by gender and crime location, NSW, 1901–1955

Table 2 reveals distinctions in the outcome of women’s prosecutions by size of locality. In the smallest and most remote regions of NSW, over 70 percent of women prosecuted for IP murder were acquitted (including Margaret Cobcroft). The acquittal rate was substantially lower in major cities (59 percent) and towns (45 percent). Just one nonurban woman was found not guilty by reason of insanity, a finding that differs from most studies of women and homicide in the twentieth century (Noh et al. Reference Noh, Lee and Feltey2010; Russell Reference Russell2010). Indeed, our evidence indicates that jurors did not “pathologize” women charged with IP murder; rather, they delivered not-guilty verdicts, or found them guilty of the lesser offense of manslaughter.

Table 2. Prosecution Outcomes for intimate partner murders by gender and crime location, NSW, 1901-1955

There was less variation by location in the outcomes of men’s prosecutions, but males were more than four times more likely than females to be convicted of IP murder across all locales. However, these gender distinctions in IP trial outcomes were starkest in the profile of persons found guilty and executed. No woman convicted of murder was hanged, but seven men were executed for murdering females between 1901 and 1955, and four of them were hanged for the murder of IPs in rural NSW.

The methods used to commit IP homicide also differed by gender and location. Table 3 shows that poison and firearms offered women means of inflicting lethal violence that did not rely upon physical strength; however, rural women were more likely to turn to firearms. In urban areas, it was possible to conceal a handgun, but on farms and stations, firearms were necessities, kept at hand to ward off pests and hunt for food. Some rural women (such as Myra Crowther, discussed below) acquired their own firearms for self-protection. As far as poison was concerned, women who tended house had access to lethal poisons in cities as well as country NSW. In Sydney, where rats were a major problem throughout the period under analysis, the open sale of a formulation of tasteless and odorless rodenticide led to a spate of prosecutions of women for poisoning IPs in the early 1950s. Two of those women were capitally convicted (Scrine Reference Scrine2002). Although the state executive ultimately commuted their death sentences, these poison-murder convictions indicate that the longstanding condemnation of women who capitalized on intimacy to carry out their deeds persisted (Helfield Reference Helfield1990). By contrast, the defense was better able to cast rural women, especially those who used firearms to defend themselves against violent and threatening husbands, in a less condemnatory light. Yet, jurors were prepared to go further, by acquitting rural female defendants who shot and killed male partners who did not present an imminent threat.

Table 3. Methods used for intimate partner murders by gender and size of locality, New South Wales 1901–1955

Table 4 shows that the age and marital status of offenders were additional demographic differences between men and women tried for IP murder. Although the distribution of prosecuted males’ ages was roughly similar across major cities, towns, and hamlets or stations, women charged with IP murder in rural areas tended to be older than their urban counterparts. In towns and hamlets, a full one-third of female-committed homicides were perpetrated by women over age 40, in comparison with only 13 percent in major cities. One of the oldest women in the dataset (Caroline Buckman, whose case is discussed below) was 55 when tried for the murder of her husband. Although the defense’s evidence of violence, insults, and threats also appeared in trials of younger women, she had endured her spouse’s abuse over decades. The locational breakdown of defendants suggests that Buckman and other rural women may have suffered domestic violence over longer periods in comparison with their urban counterparts; however, this evidence may also have persuaded jurors to consider women’s lethal violence unworthy of punishment as murder.

Table 4. Age of offender for intimate partner murders by gender and size of locality, New South Wales 1901–1955

The outcomes of IP murder prosecutions were also associated with the civil state of defendants. Table 5 shows substantial differences in the outcome of IP murder prosecutions, both between men and women, and across the rural–urban continuum. In major cities, towns, and hamlets or stations, men’s victims were predominantly, although not overwhelmingly, their wives. However, among women, the proportion of husbands as victims was greater as rurality increased – from about half in major metropolitan areas, to approximately 70 percent in towns, to over 90 percent in hamlets or stations. This evidence suggests that rural women were more likely to remain in deeply discordant marital relationships, further supporting the findings of rural criminologists. In Ellen Hoey’s case (discussed below) she had tried to separate from her husband before she shot him. Hoey was like other women, whose threats to leave abusive husbands triggered lethal violence, as remains the tendency today (DeKeseredy Reference DeKeseredy2021). When such cases were prosecuted as murder, jurors were more like to appraise the male victims more harshly than the female defendants.

Table 5. Civil relationship to victim for intimate partner murders by gender and size of locality, New South Wales 1901–1955

Finally, we investigated whether prosecution outcomes differed between men and women when judicial discretion determined the sentence for defendants convicted of manslaughter. The Crimes Act (NSW) defined murder as homicide committed “in reckless indifference to human life” or with “intent to kill,” and murder and attempted murder were both mandatory capital offenses (New South Wales (1900), Crimes Act (NSW), 40, ss 17-25). After a murder verdict, the only sway a judge had over the sentence was to recommend mercy: the state’s executive, not judges, had the discretionary power to determine whether to commute death sentences (Woods Reference Woods2018). However, if a jury found evidence of provocation or lack of intent, they could return a guilty verdict for manslaughter, which carried a maximum sentence of life, but allowed judges to impose as little a sanction as release on the convicted person’s own recognizance, as the Meagher case (discussed below) illustrates.

Table 6 provides the results of our analysis of sentence lengths in trials that ended in manslaughter convictions. In our dataset of 221 IP murder prosecutions, juries found 47 men and 14 women guilty of manslaughter. Given the small number of women’s cases, disaggregating these results according to our locational breakdown produced very small cell sizes. However, this table does confirm that judges sentenced women to substantially shorter prison sentences, as compared with men. Almost half the men (40 percent) convicted of manslaughter were sentenced to 10 years or longer terms imprisonment, but only one quarter of women faced such lengthy terms. This evidence of judicial lenience, combined with women’s overall higher rates of acquittal, indicates that jurors and judges generally agreed that women’s homicides of IPs were less worthy than men’s of punishment.

Table 6. Sentence length for those convicted of manslaughter or attempted murder* of intimate partner by gender, New South Wales 1901–1955

Overall, our analysis shows that despite the greater frequency of women experiencing domestic violence in rural areas in the lead-up to lethal violence, they were more likely than their urban counterparts to be acquitted, and far less likely than males to be punished severely. Indeed, jurors may have treated rural women more leniently because they believed defense evidence of severe abuse women endured prior to killing their male partners, overwhelmingly their husbands. The rarity of verdicts of not guilty by reason of insanity was striking, and this finding suggests that male jurors in NSW considered the mental state of women who killed their partners reasonable. Attributing lenience to judicial chivalry cannot account for the scarcity of severe outcomes, as Nagy’s study of neighboring Victoria (Nagy Reference Nagy2021) has shown, because trial outcomes varied between rural and urban cases. There is no reason why jurors in small towns or rural locales would be more inclined than their urban counterparts to chivalrous sentiments. Rather, male jurors and judges found evidence of provocation by abusive men credible, and they were prepared to believe women’s accounts, especially if they claimed they killed to defend themselves and protect their children. Even in cases where women killed defenseless partners, male jurors and judges declined to treat them harshly, as the following four cases illustrate.

One of us had to die

Harparary NSW was a tiny hamlet in 1910, located approximately halfway between Sydney and Brisbane on the north-west slopes of the Great Dividing Range. By the mid-nineteenth century, white squatters had seized the traditional lands of the Gamilaraay people. Although the area’s rivers and soil supported mixed farming and the raising of stock, the post-Federation government targeted the region for closer settlement, and it granted selections to alter the pattern of vast pastoralist holdings. James Joseph Buckman was one of those selectors, and he and his family lived in a small farm dwelling. Married in 1890, he and his wife, Caroline, had six children, but the union was marred by violence. In 1901, when they lived on a small property outside of Koorawatha, reports by neighbors of the couple’s rows led a local constable to look in on the family. A charge of assault causing grievous bodily harm was laid after Caroline Buckman, bleeding and bruised, told the constable, “my husband beat me and kicked me…he threatened to kill me.” She agreed to testify against him: “I don’t want to be killed.” When the case went to trial, the charge was reduced to common assault, but Buckman paid a hefty fine – twenty-two pounds, seven shillings. The couple then separated, but they were living together again in Harparary in early 1910. By that point in their conflictual marriage, Caroline Buckman decided “one of us had to die,” and she shot him dead with a revolver. In her encounter with the criminal justice system, she was the defendant, facing the death sentence if convicted of murder.

It might have taken days for authorities to discover that James Buckman was dead had Mrs. Buckman not informed the police promptly. The morning after the incident, she rode a sulky (a horse-drawn one-person cart) 30 miles to Narrabri, the nearest town with a police station, where she handed in her revolver (with one spent chamber) and confessed she had killed her husband. “Bush Tragedy” was the headline several papers used to describe the coronial inquest. Three of the couple’s teenaged children corroborated their mother’s claim that their father had frequently abused her verbally and beaten her. They added that he had compelled their mother to do hard labor around the property and refused to support his family. The sub-inspector of police in Narrabri knew the couple was “unhappy,” and he told the coroner that Mrs. Buckman had recently preferred a charge of lunacy against her husband in an effort to remove him from the house, without success. When she gave her own statement at the inquest, Caroline Buckman described the shooting as an act of self-defense. On the morning of 20 January, she fell while riding a horse in the scrub, and her husband had followed her, not to assist, but to club her with a branch. “‘I came back on purpose to murder you',” he threatened. “'For the love of my children, don’t murder me’,” she replied. Later, she reflected: “I thought it was certain death for me. I drew the revolver and fired at him.” Yet, other witnesses at the inquest told the corner Mrs. Buckman may have planned to murder her husband: she possessed a revolver and kept it well-oiled and loaded; she had tried but failed to put him away; and, according to the victim’s uncle, she had vowed: “‘I have a good mind to shoot him’.” The coroner decided the evidence against Caroline Buckman was sufficiently strong for her to be prosecuted for the “felonious and malicious” murder of her husband (Broken Hill Barrier Miner 1910 February 14).

Caroline Buckman and her children, unsupported by their father, were so poor that they had survived for months on flour and rabbits; consequently, she could ill afford to hire a solicitor or a barrister to represent her. Australian historical criminological studies have confirmed the deleterious impact of poverty on the quality of legal defenses (Piper and Finnane Reference Piper and Finnane2017a; Reference Piper and Finnane2017b). However, one of the state’s top lawyers from an illustrious legal family, Richard Windeyer, volunteered to represent the accused pro bono, and he ensured that the woman was released on bail (possibly guaranteed with his own funds). A specialist in criminal law and divorce law, Windeyer also represented Mrs. Buckman at trial, when he skilfully attacked the claims of Crown witnesses that the shooting had been planned (Bennett Reference Bennett1990). However, like most defenders of women who killed their husbands, he based the defense on the accused’s testimony that the victim had made her life, and that of her children, a misery.

The trial took place in the northern NSW town of Armidale (population 6,500), before veteran Supreme Court Justice Emmanuel Cohen, who had presided over several cases in which women were tried for the murder of husbands, including one who was convicted in 1904 of murder by poison (Strange and Heatherington Reference Strange and Heatherington2020). In an unusual move, Windeyer called the defendant to testify under oath (which exposed her to cross-examination), but the risk paid off. The jury barely had time to walk from the courtroom and return, since they reportedly delivered their verdict in 5 minutes. There is no record of the judge’s charge, in which he would have reviewed the evidence and instructed the jury on possible verdicts (guilty of murder or manslaughter or not guilty), but the newspapers reported the foreman’s rider to the verdict of not guilty: they considered the killing of James Buckman a “justifiable homicide.” According to the law of NSW, this interpretation was baseless. The Crimes Act (NSW) recognized the common law defense of self-defense, but it did not recognize any form of violence as “justified.” Despite the jury’s reason for acquitting the defendant, Justice Cohen did not admonish them. Instead, he released Caroline Buckman as a free woman. One year later, the widow remarried, becoming Mrs. Thomas Day, and the couple moved to inner Sydney. As far as Mrs. Day was concerned, after enduring two decades of violence and abuse from the father of her children, the case of the “Outback Tragedy” was closed (Armidale Chronicle 1910 April 30).

He treated me more like a beast than a woman

In 1927, evidence of a history of wife abuse and violent assaults also preceded the trial of a woman tried for the murder of her shearing contractor husband. The accused, Ellen Hoey, had tried to use the civil and criminal court system to shield her from the abuse inflicted by her husband, Norman. However, Mrs. Hoey ended up in court after she shot the man dead. She, too, turned herself in, and told the police, “No one knows what I have had to put up with.” In fact, her domestic troubles were well known to her neighbors and local authorities. The fatal incident occurred on the main street of Inverell (population 2,500), a town considerably larger than the hamlet where the Buckmans had lived unhappily, but small enough for locals to know each other’s business. In the spring of 1927, Norman Hoey was shearing on a large station northwest of town, while his wife, who had sought a judicial separation from him on the ground of cruelty in 1924, lived in Inverell with their two children. Ten days prior to the fatal encounter, Mrs. Hoey laid a complaint of assault against her estranged husband, but he had failed to appear to answer to the charge. However, he did drive up to the family home, after he broke his promise to take the children to the pictures. According to Mrs. Hoey’s testimony at the inquest, she approached his car holding a revolver, which she kept on her person because she was “terrified of her husband.” As the two argued over the cancelled children’s treat, she claimed a “flash” went off unintentionally. Once she realized she had shot her husband in the temple, she sent her daughter for the police and waited for them to arrive. Was this a planned or accidental killing? The coroner’s inquest answered that question by finding there was sufficient evidence for Ellen Hoey to be tried for her husband’s murder (Armidale Chronicle 1927 November 1).

Like Caroline Buckman, this accused husband murderer could not fund her defense, but prior to Hoey’s trial, community members banded together to secure her bail and to enlist a top-notch barrister. At the inquest, a local business owner told the coroner he considered Mrs. Hoey “a decent, honourable woman;” by contrast, Mr. Hoey’s reputation was “very bad” in regard to “women and loose living.” As soon as the widow was charged with murder, her female supporters set up a defense fund, and the first donation of twenty pounds came from the Inverell branch of the Country Women’s Association. Far from a feminist organization (Jones Reference Jones and Jones2017; Teather Reference Teather1994), the club’s members put up 1,000 pounds in sureties to secure the woman’s release on bail, and it sent circulars through the press and its branches to drum up the fees to hire a high-profile barrister to defend “their sister” (Mudgee Guardian and North-Western Representative 1927 November 1). The man who answered the call was former NSW Premier W.A. Holman, K.C.

In Ellen Hoey’s murder trial, convened in the town of Tamworth (population 7,200), Holman’s job was made easier by the prosecutor in the case. L. J. McKeon, representing the Crown, declined to cross-examine defense witnesses, including the couple’s 13-year-old daughter, who recounted many acts of her father’s physical and verbal abuse, plus attacks on the accused herself. The Crown also closed with no address to the jury. By contrast, Holman challenged the jury to consider whether the defendant would have shot her husband on the town’s main street if she had murder on her mind. In his closing address, he asked: “what woman, who had never fired a revolver before in her life, would shoot when her darling boy was between her and his father?.” When the jury returned a verdict of not guilty after an hour’s recess, Justice James accepted their finding. The closest he came to objecting to the acquittal was to caution others against sorting their differences with revolvers. “Let this be a warning to all people not to carry firearms,” the judge intoned, and further advised the “accused has run a very great risk by doing so” (Lismore Northern Star 1928 March 2). Yet, the jury appeared to appreciate that Ellen Hoey’s husband had presented a risk to her, and the relief she had sought through civil and criminal remedies had failed to punish him or stem his violent threats. The acquittal sent a stronger message than the judge’s admonition: honorable women who killed dishonorable men should not be punished.

Hell on Earth

On a remote station called Kangarooby, John Meagher, husband and father, met his end when his wife shot him on 20 August 1934. As the manager of the station, the 42-year-old victim was well known to locals, but district authorities knew him as a trouble-making drunkard, whom the police had charged in the previous few months with bad language, disorderliness, and resisting arrest (Armidale Express and New England Advertiser 1934 October 17). At the coroner’s inquest, Eileen Meagher, aged 35, stated her husband had come home drunk late on the night she shot him. During the quarrel that ensued, he threatened to kill her and one of their children. After he went to bed, she entered the bedroom and found him clutching a pea rifle. In her statement to police, she said she knew the rifle was cocked when she wrested it from her husband. Then she fired it without taking aim: “I did not mean to shoot him; it was an accident,” she told the officer in the immediate aftermath. She also described the deceased as “a good husband to me and it was drink that was the trouble” (Forbes Advocate 1934 August 24). On the basis of Mrs. Meagher’s initial statement and testimony, the coroner could have determined the death was accidental, committed in the heat of a struggle. However, the ballistic evidence suggested cooler criminal intent: John Meagher had been shot not once, but twice – in the head, and at close range.

All the women charged with IP murder in NSW in the first half of the twentieth century were too poor to hire their own counsel, and in Eileen Meagher’s case, some were also illiterate. This left her particularly reliant on the barrister who responded to the usual court request for a man to represent the indigent, a factor that typically put defendants at a disadvantage (Piper and Finnane Reference Piper and Finnane2017b). P.D. Shortland, a barrister with 12 years’ experience, based his practice in the small alpine town of Cooma (New South Wales Law Almanac 1931). He lacked the stature of Holman or the experience of Windeyer, but as soon as he arrived on scene in Bathurst (population 117,000), he came up with a plausible defense that hinged on casting the victim as a villain. The “good husband,” described in the accused’s police statement, became a dissolute brute, and he tried to make the accused into a dutiful wife and mother, pushed to defend herself. Under the Crimes Act, defendants could make an unsworn statement from the dock, rather than testify under oath on the witness stand, and Shortland used this provision to present an account of the accused’s marital troubles. In a hushed voice, Eileen Meagher told the court her husband had made her life “a veritable Hell on earth.” On several occasions, he had threatened her with a gun and a razor, and his “frenzied fits of passion” left her with “the horror and fear” that she and her four children would be killed (Cowra Free Press 1934 October 22). Because her husband had squandered his wages on drink, the family lived in a barely furnished hut without any comforts, and whenever she pleaded for money to clothe or feed the children, her husband responded with blows and insults. The defense called four male witnesses to corroborate the woman’s claims, and they all testified they had seen the victim drunk, abusive, and threatening toward his wife. This chorus of men confirmed that the victim was an unworthy, discreditable husband and father.

In the defense’s closing address, Shortland tried to persuade the jury to absolve the defendant of criminal responsibility, not through sympathy, he urged, “but on legal grounds.” “What this woman put up with during the three months prior to the tragedy possibly only God knew. Think of the mental anxiety she must have felt,” he asked them. To explain the woman’s second shot was trickier, so he tried to convince the jury to consider the two shots as “part of one act,” committed under great strain, not a loss of sanity. After deliberating for 2 hours, the jury announced its decision. Rather than convicting the defendant of murder or finding her not guilty by reason of insanity, they found her guilty of manslaughter, with a “strong recommendation of mercy” (The Bathurst National Advocate 1934 October 18). The manslaughter verdict fell short of the defense’s hoped-for acquittal, but it transferred Eileen Meagher’s fate from the jury’s hands into the judge’s.

By the time Justice Halse Rogers presided over the Meagher case, he had served 6 years on the state Supreme Court bench. As usual, he charged the jury on the prospect that the defendant might not be criminally liable on the ground of insanity, but he accepted the verdict. He then departed from convention by asking the foreman to explain the “grounds of the recommendation.” Was it “provocation” or the “general circumstances of the crime?,” the judge queried. To both questions, the foreman responded, “yes”. Before he pronounced Eileen Meagher’s sentence, Rogers further speculated, “the jury probably took the view [the defendant] was half distraught when she committed the deed.” What sentence should he now impose to acknowledge the gravity of the crime considering the “peculiar circumstances of the case?” Justice Rogers’ bench side ruminations presaged a lenient sentence – 5 years hard labor – which he further softened with a recommendation that the convicted woman be released immediately. As the judge explained, he wished to allow the convicted killer to reunite with her children, so long as a relative or other responsible person was prepared to enter into a bond to care for her. Within 2 weeks, the NSW Minister of Justice announced the government’s approval, adding that “satisfactory arrangements” had been made for the “future welfare of Mrs. Meagher” (The Bathurst National Advocate 1934 October 26).

I decided it would be better to kill him

Whether women committed IP lethal violence in the heat of the moment or by accident, female defendants who claimed they had killed abusive husbands for the sake of their children, not just their own safety, narrated accounts that conformed to the image of maternal honor. The acquittal of women charged with IP murder was less common for scheming poisoners, as the last woman executed in NSW in 1889, Louisa Collins, was alleged to have been (Kukulies-Smith and Priest Reference Kukulies-Smith and Priest2011). But when counsel defended accused female poisoners, they usually tried to raise doubt about the cause of death (Strange and Heatherington Reference Strange and Heatherington2020). By contrast, barristers who defended women who shot their husbands dead at close range had few options, other than to argue that women committed homicide because they felt they had no other option after years of brutality. In the case of Myra Crowther, she told the police: “I decided it would be better to kill him” (Sydney Truth 1945 December 16). Despite her admission – that she had shot her husband from behind with a rifle while he was lying down – the Crown, not the defense, went out of his way at her murder trial in 1945 to persuade the jury to acquit the woman.

Forty-five-year-old Myra Crowther lived with her husband and six children in a modest four-room weatherboard house in Wiradjuri country, on the outskirts of Gilgandra (population 1500). Not long after the family moved to the district, Edward Crowther’s criminal behavior and his abusive treatment of his wife and children became an open secret. The coroner’s inquest, held after Mrs. Crowther turned herself in to the police and indicated where her son had buried the body, heard evidence that the deceased subjected his children to beatings with a buckled belt and that he had threatened to kill them or send them to state homes. One of their daughters, who had since married and left home, testified that her father had made “improper suggestions” to her. Mrs. Crowther told the coroner that on the night before the shooting she had seen her husband leave their bed, pick up a razor, and walk toward the bedroom where one of the younger girls slept. The next morning, she decided it would be better to “spend my whole life behind walls than let him kill my family” (Sydney Truth 1945 November 4). As her husband lay reading, she picked up a rifle, walked behind him, aimed, and fired. Could Ellen Crowther’s lethal act of domestic violence be any crime but murder?

Although the couple’s 18-year-old son was originally charged as an accessory to murder for agreeing to bury his father’s body in a makeshift grave, he did not face trial, just his mother. Veteran reporters who covered Sydney’s Central Criminal Court beat described Myra Crowther’s prosecution as “one of the most unusual murder trials in years.” The reason for their observation was the Crown’s support of the defense’s case. The prosecutor, Charles Vincent Rooney, K.C., was a member of the state’s elite squad of Metropolitan Crown Prosecutors, a “stern accuser of the guilty,” who had built a reputation for forensic flamboyance (Tedeschi Reference Tedeschi2006: 12-13). Yet, in this case, he made it clear he disapproved of the victim’s character. In one reporter’s description of the proceedings, he indulged in alliteration to describe Rooney’s performance: “Pathos turns a Prosecutor pleader” (Sydney Sun 1945 December 16). In this IP murder case, the Crown set up the female defendant’s acquittal.

Myra Crowther’s trial opened and closed on 10 December 1945, hastened by the absence of witnesses and the lack of a case for the defense. Crown Attorney Rooney boldly declared that the “man this woman killed deserved death” (Sydney Sun 1945 December 16). Without testing any of the witness testimony from the inquest or the medical evidence as to the cause of death, Rooney told the jury there was “not one disputed fact in the whole case.” After he acknowledged it was “not a pleasant thing to speak ill of the dead,” he proceeded to describe Edward Crowther as a scoundrel: “for years he was a drunkard, gambler, thief and bully who terrorised his wife and family.” By contrast, Rooney admired the defendant’s confession as “the most frank and sincere he had ever read,” and he paid tribute to the way she had hidden her husband’s corpse, because it gave her the time she needed to drive over 130 miles to drop off her children with relatives before she returned to Gilgandra to turn herself in. Far from a criminal, this woman was “amazing from the viewpoint of her courage and steadfastness.” For Rooney, the key element of the case was the allegation of attempted incest: the “husband was depraved” (Forbes Advocate 1945 December 14; Canberra Times 1945 December 11). At Rooney’s suggestion, the judge allowed the case to go to the jury after the close of the Crown’s statement, and the jury delivered a verdict of not guilty without hearing a defense.

For a seasoned Crown Attorney and Supreme Court judge to allow this unorthodox murder trial to proceed as it did suggests that both lawmen were intent on condemning a brutal man, a danger to his wife and children. According to the law of the time, the Attorney General could have discharged the defendant, rather than proceed with an indictment and trial that was unlikely to lead to conviction, but there was sufficient evidence to try the confessed killer for murder (New South Wales, Crimes Act, 1900, Act 40, Third Schedule, Form no. 1, s. 18 (b)). Once the trial proceeded, the 12 men empanelled as jurors would normally have heard evidence and listened to the judge’s charge before reaching a verdict. In this instance, the jury likely felt obliged to find the woman (who had, after all, confessed to the planned murder of her husband) not guilty. On top of that, the jury added that this was a “justifiable homicide,” a defense still unrecognized by the NSW Crimes Act in 1945. The jurors in the trials of Caroline Buckman and Ellen Hoey also based their not-guilty verdicts on the justifiability of women’s IP homicides, which suggests a widely held belief, more pronounced in rural cases, that women who killed men who mistreated them and their children should not be held legally or morally accountable. Cruel and depraved men living in small towns and remote settlements may have had greater scope to terrorize their wives and children than their urban counterparts, but when rural women’s homicides of abusive IPs exposed men’s wrongdoings, these victims were more likely to induce disgust than sympathy.