Longitudinal Analysis of Australian Filicide Perpetration Trends: Filicide in Victoria, 1860–1920

The historical examination of filicide in Australia is limited and often focuses on case studies of maternal filicides. Longitudinal trends of Australian filicide offending have focused almost exclusively on the late twentieth and early twenty-first centuries. Our study aims to fill a gap in Australian criminological knowledge about filicide. Utilising prison and Supreme Court records from 1860 and 1920, we plot the extent of filicide offending by men and women in Victoria to create a more comprehensive picture of filicide perpetration. This study also tests whether identified motives and risk factors for filicide today can be applied to historical data, to make these data accessible to criminologists studying filicide in the twenty-first century.


Introduction
Mary Carnie was 33 years old, pregnant with her sixth child and living with her husband and five children in Naringaringalook, about 200 kilometres north of Melbourne, Australia. On the morning of 6 September 1900, Mary poisoned herself and her two youngest children (Gertrude, aged four years old, and James Percy, aged one) with strychnine inside a locked room of the Carnie residence. Her husband, Thomas, was a dairy farmer, and upon returning home he found the door to the room locked; he broke it open and found the children, dead, with Mary, barely lucid, beside them. Mary had written a suicide note in which she stated that ongoing physical pain in her head, back and legs, and the impending birth of her sixth child were too painful to contemplate, hence, her attempting suicide. Because she did not wish to leave her children motherless, she had poisoned them as well. During the murder trial, it was established that Mary was known by her husband, maid, and her sister as 'irritable', which they had attributed to neuralgia (of which the witnesses were very understanding) and her grief from the loss of her sister-in-law just a year earlier. Thomas was certain that his wife was a good mother; she loved her children, and he stated with no irony that 'when she threatened to beat the children it was only in the ordinary maternal way' (VPRS30/P0/001229). The jury found Mary not guilty on the grounds of insanity, and she was detained by the Governor's pleasure for two years in a lunatic asylum.
Filicide in Australia, especially in a historical context, is under-researched. Longitudinal trends of filicide offending have focused on the late twentieth and early twenty-first centuries (1980-2010s), and historical research into filicide has concentrated almost exclusively on maternal infanticide (Allen JA 1990;Burton 1986;Goc 2016;Laster 1989;Rychner 2017;Swain and Howe 1995). There has only been one study published about longitudinal, historical paternal child homicide in Australia (Kaladelfos 2013). Considering that filicide is primarily perpetrated by fathers and stepfathers in Australia today (Brown et al., 2019), it is important to expand upon knowledge of filicides. Our aim in this paper is to include paternal filicide, filicide-suicides and familicides, along with maternal filicides, to create a more comprehensive picture of filicide perpetration. In doing so, we are also testing whether motives and risk factors associated with filicides today could be applied to historical data. Our argument is twofold. We argue that the narrow focus on maternal infanticides in historical research has obscured knowledge of filicides and made it difficult for criminologists to utilise findings from historical data. We also argue that Australian criminology has not connected with historical research in a sustained manner, and so the shifts in offending and victimisation for filicide in Australia over an extended period are unclear.

Understanding Filicide
Filicide refers to the killing of a child by a parent or parent equivalent (e.g., a stepfather). The crime is a distinctive type of homicide and is often understood to include the sub-category of neonaticide, where infants are killed within the first 24 hours of life (Resnick 1969(Resnick , 1970. International research has sought to understand infanticide to inform policy and legal prevention (Brown, Tyson and Arias 2018), and such research has traditionally looked at perpetrator motivation. The United Nations estimates that 95,000 children are killed annually worldwide, and the majority (over 56%) are killed by their parents. Of these filicides, 64.2% occurred in high-income countries (Stöckl et al., 2017). The crime remains a significant issue in Australia, where one child is killed by a parent each fortnight on average (Goldsworthy 2019). Filicides accounted for 15% of domestic homicide incidents recorded in 2012-2014 (Bryant and Bricknell 2017). The crime is increasingly incorporated into national conversations about family violence in the wake of high-profile, often public filicides, such as those perpetrated by Rowan Baxter (2019), Arthur Freeman (2009) andRobert Farquharson (2005).
Phillip J Resnick's landmark criminological study of filicide, conducted in Pennsylvania, devised five categories for offender motive: 'altruistic', 'acutely psychotic', 'unwanted child', 'accidental' and 'spouse revenge' (Resnick 1969). Altruistic filicides, identified as the most common motive, sought to relieve or prevent children's suffering, sometimes followed by parental suicide. Acutely psychotic filicide referred to cases where parents killed while suffering from a mental disorder, hallucination, epilepsy or delusion. Accidental filicides were classified according to a lack of intent to kill and were usually an unforeseen outcome of extreme discipline or punishment of the child. The spouse revenge category referred to deliberate killings to inflict suffering on the other parent (Resnick 1969). D'Orbán (1979) devised six similar categories in a study of maternal filicides in an American prison, including motives for where children were unwanted, abusive mothers, neonaticides, women who killed in retaliation in order to inflict suffering on the other parent, where mothers were mentally unwell or enacting what they considered as 'mercy killings'. While studying filicide cases in England, Wilczynski (1997) developed a classification system with 10 types of motive, moving beyond D'Orbán and Resnick's classifications to include altruistic, killing through discipline, no intent to kill or injure, psychosis in the parent, Munchausen Syndrome by Proxy, jealousy or rejection of victim, unwanted child, retaliatory, killings secondary to sexual abuse of the victim or another person, and cases where the motive is unknown. Additional contributing factors to filicides have been identified. Scott (1973) andD'Orbán (1979) identified alcohol and drug use, while Daly and Wilson (1994) recognised the greater risk posed by stepfathers compared with biological fathers in the Canadian and British contexts. The 'motivational' approach to filicide has received criticism, with Silva and colleagues (1998) arguing that social and cultural environments should be taken into account, and that understanding motive alone leads to ineffective evaluations of the crime. Nevertheless, motive continues to inform filicide research through five key categories: altruistic, fatal abuse (sometimes referred to as accidental), retaliatory, neonaticide and mental illness (Brown and Tyson 2012).
The 1990s saw a shift from motive categorisation to the identification of risk factors contributing to filicide (Meyer et al., 2001;Overpeck et al., 1998;Wilczynski 1997b). Wilczynski (1997a) proposed risk factors that should be taken into account when studying filicides: social factors (such as childcare responsibilities or parental income), demographics (age of the perpetrator, socio-economic class or ethnicity), demographics of the victim, psychiatric factors (including substance abuse), situational factors at the time of the act, family history (such as a history of child sexual abuse), prior family violence and prior contact with agencies such as Child Protective Services. Korbin (1987) emphasised poor social support as a significant risk factor for mothers to commit filicide.
Filicides are anomalous as a type of homicide in that women are as likely as men to be perpetrators. Early studies of maternal filicide addressed the scarcity of literature examining women as perpetrators (Alder and Baker 1997;Alder and Polk 1996;Allen H 1987;Bourget and Bradford 1990;Brownstein et al., 1994). Scholars from around the world have sought to break away from the simplistic, medicalised account of maternal filicide by showing that determinants such as housing problems, lack of social support, poverty and the mothers' backgrounds as victims of child sexual abuse or domestic violence contribute to filicides (Alder and Baker 1997;Wilczynski 1991). Alder and Baker (1997) contested the popular explanation of women's acts being a psychological 'outburst' resulting from stress. Rather, they argued that filicide cases revealed the 'dark side of the burden of motherhood as it is socially and culturally constructed', and described female perpetrators who could not easily be classed as 'passive victims' of circumstance (Alder and Baker 1997, p. 36).
Dominant or marginal 'masculinities', and the social and historical contextualisation of men's behaviour as a theoretical framework in the 1990s, gave rise to insightful scholarship on paternal filicides (Connell 1993). Alder and Polk (1996) identified the prevalent scenarios in which filicides occurred, comparing male parent and male non-parent offenders. Their study of child homicides perpetrated in Victoria between 1985 and 1994 found that the most common category of paternal filicides was 'fatal nonaccidental injuries', where the intent was to discipline or punish the child rather than to kill. This was followed by the categories of altruism and spousal revenge (Alder and Polk 1996). Family violence, particularly in the context of marital separation or child custody disputes, is now well established as a risk factor for filicides and familicides (Johnson 2006;Kirkwood, McKenzie and Tyson 2013;Liem, Hengeveld and Koenraadt 2009). Similar to maternal filicide research, Alder and Polk (1996) demonstrated that filicides occurred in a range of situations that defied any 'universal' explanation for masculine behaviour that had been submitted in previous studies. Comparisons between maternal and paternal filicides are fruitful in understanding the gendered 'pathways' to filicide (Dawson 2015;Eriksson et al., 2016). However, paternal filicides remain relatively understudied and require more attention, particularly as Australia has a higher rate of paternal filicide compared with other countries (Mouzos and Rushworth 2003).

Filicide: Historical Backgrounds
Much of the historical literature in Australia has focused on maternal infanticide due to the gendered politics surrounding the crime in the late nineteenth and early twentieth centuries (Burton 1986). From the 1870s, governments showed increasing concern over the number of abandoned or deceased infants found in drains, ponds, trains and dredged up in the Yarra River (Swain 2010). Women, particularly young, unmarried domestic servants, faced economic adversity and social stigma when they gave birth to illegitimate children. Contraception was unaffordable or unattainable for most, and abortions were illegal and highly dangerous (Allen JA 1982;Swain and Howe 1995). Infanticide remained extremely difficult to prosecute, due to the 'dark figure' of unknown infant deaths and the difficulty of proving the infant had lived prior to the murder. Medical ambiguities often saw women convicted of the lesser offence of concealment of birth (Swain and Howe 1995).
Some have argued that the medicalisation of infanticide in these trials sought to remove women's agency in their capacity to perpetrate wilful, premeditated violence (Ainsley 2000;Jones 1982). English and Victorian legislation would codify this medicalised understanding of maternal infanticide in the twentieth century, separating infanticide from murder, with the former considered a crime where the mother suffered a disturbance of mind as a result of pregnancy (Infanticide Act 1922 (UK); Crimes Act 1949 (Vic)). This offence carried a reduced sentence equivalent to manslaughter (Crimes Act 1949 (Vic), s4, 1).
Outside maternal infanticides and the social and legislative responses to them, there have been few historical studies focused on paternal infanticides or filicides more broadly. Analyses of historical child abuse provide insightful statistics regarding the contexts in which children were killed (Thearle and Gregory 1988). Lake (1986) was among the first Australian scholars to centre masculinity as a historical consideration, arguing that just as women were socialised into 'femininity', men were socialised into 'masculinity', a shifting category that was informed by nation, class and race. Historians demonstrated that men who killed children did so in different circumstances to women, and social responses to their crimes also differed (Arnot and Usborne 2002; Kaladelfos 2013). Kaladelfos (2013) found that men were more likely to kill older children, kill multiple children at once or perpetrate familicides. Motives for filicides, as explained in perpetrator confessions or in court, primarily centred on financial pressures to cater for the family. Other cases were perpetrated in revenge of unfaithful wives or in response to a 'slighted paternal authority' due to the sexual immorality of daughters (Kaladelfos 2013). Moving away from simplistic explanations of a rough, masculine culture that rejected the domestic, these cases reveal understandings of nineteenth-century fatherhood, family responsibility and male shame (Kaladelfos 2013). The end of the century saw the ascent of 'conjugal masculinity' and a reshaping of fatherhood that advocated for fathers to play an emotional, not merely financial, role in their children's lives (Hogan 1999). Kaladelfos (2013) demonstrates that, where familial bonds seemed strong, paternal filicides were more likely to be interpreted in terms of mental instability. In England, men who perpetrated infanticide received more leniency before the courts than men who committed other crimes (Shepherd 2013), a claim that warrants further analysis in the Australian context.

Methodology: Filicide in Victoria
For this research, we have applied contemporary definitions of filicide to our analysis. We are not concerned with the specific offence of concealment of birth, where it could not be established whether the mother had intended to kill her newborn, or if she had inappropriately disposed of a stillborn child. We have classified neonaticide as a crime where children have been killed within 24 hours of being born (Porter and Gavin 2010).
For our research sample, we include the deaths of all children (biological or step) from birth until the age of 18 years. We are using filicide as both an umbrella term and crime sub-category. As an umbrella term, filicide is being used to signify the deaths of children killed by their parents (biological or step) from birth to 18 years. However, as a sub-category, it is also applied to children who are between one and 18 years old when killed. We chose this definition because there is no clear term for children aged above infancy who are killed by their parents (West 2007). Legal definitions of infanticide can vary based on jurisdiction-we classified children who were killed up to the age of one year as infanticides in line with common research definitions (Porter and Gavin 2010;West 2007).
This pilot study details the historical filicide trends in Victoria between 1860 and 1920. This study has tested internationally accepted filicide motive categories to investigate whether they can be applied to historical filicide cases, or if these historical cases exhibit overlooked motivator categories that can be applied to contemporary offending. We have also examined whether currently accepted risk factors for filicide offending are found in historical cases. Therefore, this study provides a long-term criminological perspective of filicide perpetration in Australia that has been missing in filicide research, allowing a closer examination of how the criminal justice system has responded to filicide over time.
We utilise historical criminology methods and understanding to further our knowledge about filicide offending in this paper. Historical criminology is not clearly identified as a branch of criminology, and there is little dedicated scholarship to serve as a foundation for inquiry. This article takes the position that historical criminology is best defined as criminology done in a historical mode (Churchill 2019) because historical criminology is 'the only methodologically robust means by which to evaluate the potential for the past to help explain the present' (Lawrence 2019: 493). Using archival sources to create richer and clearer continuities and dislocations between the past and the present has been successfully applied to topics of penal history (Cox and Godfrey 2020;Guiney 2018;Nagy andPiper 2020, 2019;Nagy 2017, 2018), policing (Churchill 2018;Finnane 1990), immigration (Kaladelfos and Finnane 2018), and crime prevention (Churchill 2016). Therefore, the focus of this article is on creating a historical work of criminology, rather than writing the history of crime-this distinction is an important one. Our research keeps the contemporary issues as consequential and equal to the concerns of investigating the past.
It is important to acknowledge that criminology is built upon theories of offending and deviance, whereas history is not typically theory-driven. Applying criminological theories to the analysis of historical filicide offending means that social science theories can be applied to the past. These techniques are not typically employed by historians, so there is an opportunity for criminologists to draw upon a rich source of historical material in their current explorations of infanticide, filicide and familicide. Historical criminology has not been utilised to the same extent or manner in Australia as in the UK or Europe (Nagy 2021) and certainly has not been applied to criminological studies of filicide. This article intends to adopt historical criminology to rectify this situation.
The database used in this research draws upon multiple sources. Data for convicted women were taken from the Victorian Central Register of Female Prisoners, a series of records created by prison authorities in Victoria. A record would be created for a woman upon her first entry into prison, in which future imprisonment details were entered when necessary. Prison administrators generally ensured that women's records were not doubled-up if a woman was re-convicted, and that other prisoner's details were not entered against a woman's name. The Central Register of Female Prisoners contains the data of 6,042 women who first entered prison between 1860 and 1920. These data were entered into an SPSS database for statistical analyses.
The information about male offenders was sourced from the Victorian Supreme Court Register. Identification of capital cases was the primary endeavour, and from this selection, filicide cases were filtered out. As with the female prisoner data, attention was on the years between 1860-1920 for this initial study.
Additional information for both male and female offenders was sourced from the Capital Cases Files from the Public Records Office of Victoria (PROV). These files contain prosecution trial briefs, depositions, inquest records, physical evidence (for instance, maps), medical records and other information sourced during police investigations to be presented at the accused's trial. Where information was unclear or unable to be sourced from PROV, the National Library of Australia's Trove website was used to source newspaper records.
Prison records contain the details of men and women who were found not guilty by reason of insanity (NGRI), who served part of their terms in prison while awaiting entry into asylums. Although they were found to be not guilty by a jury, they were often treated in a similar fashion to those found guilty by prison authorities and in administrative data prepared for State officials. Juries would sometimes return an NGRI verdict where they were convinced that the crime had taken place as presented by the prosecution but did not wish to see the man or woman sentenced to death (Nagy 2021;Rychner 2017).
In total, 50 cases of filicide where an adult was convicted of the death of a child under 18 years were identified during the 60 years under investigation.

Characteristics of Filicide: Offender and Offence
Violent offending and imprisonment for all forms of offences were decreasing in Victoria between 1860 and 1920 Nagy 2017, 2018). Homicide rates for men and women were at 0.5 per 100,000 by 1920; these had remained stable for women and dropped from 3.5 per 100,000 for men from 1860 (Nagy 2021). In contrast, the rates of convictions for filicide were increasing during the same period. As Figure 1 demonstrates, convictions for filicide increased dramatically during the financial boom in the 1880s. The financial depression of the 1890s through to the first decade of the twentieth century correlated with a drop in convictions, before they began rising again from 1910. However, when broken down on a yearly basis, we see that higher rates of convictions were recorded in some years, with 1888 and 1900 recording three each (see Figure 2). Only 23 years had no recorded convictions for a filicide offence. Unsurprisingly, the majority of offenders were women (see Figure 3), and most victims of both men and women were infants (children under the age of 12 months). Cases involving older children were more likely to have multiple victims, whereas neonaticide and infanticide almost exclusively had one victim per offender. Three of the female perpetrators had accomplices: Annie Bell's mother was accused of helping her daughter kill a female infant, but charges were soon dropped; Elizabeth Cooke's husband, George Cooke, was accused of killing their 13-year-old son but charges were again dropped; and Bridget Elizabeth Magee was found guilty of murdering her infant, and her lover Peter Kane was a co-accused. In one case, the mother, Janet Dibben, was charged with manslaughter for the death of her infant son although the child had died in the care of Ellen Gardiner, a baby farmer whom Dibben had paid to care for the child. Although many of the victims were newborns and infants, a number of victims were older children (see Table 1). The oldest child to be killed by a parent was the 13-year-old son of Elizabeth and George Cooke. Regardless of whether the type of filicide was for an infant or an older child, the adult responsible was more likely to be convicted of murder rather than manslaughter (Table 1). Even so, executions for filicide were rare (see Table 2); only two offenders, one man and one woman, were hanged for their actions. James Johnston was 35 years old when he was convicted of murder for his familicide in Ballarat in 1891. Johnston shot and killed his wife, smothered his four children and attempted to kill himself by ingesting poison. In 1895, Emma Williams was executed for the murder of her infant son, John, after his body had been weighed down with a stone and thrown into the Port Melbourne lagoon. Both executions occurred in the latter part of the nineteenth century. However, the low execution rate suggests that responses to filicide were not as harsh as what might be expected, and instead demonstrates a willingness of juries to consider the motivations of the accused. Juries during the nineteenth century could recommend mercy in murder convictions. That is, they could ask the judge to recommend to the government of the day that the death sentence be commuted. IJCJSD 10(2) 2021 When considering the age of offenders, filicide was overwhelmingly committed by those aged between 20 and 39 years of age (70%) (see Table 3). Filicide was committed marginally more often by those aged 20 to 29 years (36%) than any other age group. This mirrors the findings from research today in Victoria and Australia (Brown et al., 2019), and internationally (West 2007). Although the youngest offenders from the historical sample were only a year older than the youngest known offender from contemporary data (18 years in comparison to 17 years for the twenty-first century data), the oldest offender today was 75 years old in comparison to the historical data where the oldest offender was 62 years old. Although female offenders were found across all age groups (and predominantly in the under 20, 20-29 and 30-39 categories), male offenders were older, all aged above 30. Neonaticide and infanticide were associated with younger, female offenders. Five of the six neonaticides (83%) were committed by women aged under 29 years, and 17 of the 29 infanticide offenders (59%) were under 29. These offences were linked to a lack of support (financial or social) or fears of unemployment. All neonaticides were carried out by a single woman who could not afford to leave their employment to raise the child. In half of the neonaticide cases, the women gave birth and killed their child while at their place of work.
The causes of death in the historical filicide sample were varied (see Table 4).The neglect of a child was the most common cause of death (22% of cases)-this included malnutrition, accidentally turning onto a sleeping child when the mother was drunk or pushing a child who then fell and hit their head. Suffocation or strangulation accounted for 20% of deaths, drowning for 16% and poisoning and stabbing or slashing both accounted for 12% of deaths. Although Australian research from the twenty-first century indicates that the majority of filicide deaths are caused by beatings (25%) (Brown et al., 2019), blunt force trauma accounted for only 10% of deaths in the historical data. Strangulation or suffocation, which is the second most common cause of death in filicide cases today (14%) (Brown et al., 2019), was marginally higher in the historical sample (20%). These historical data, while only marginally different from contemporary Australian data, are far more different to contemporary international research (Ward 2007).

Motives and Risk Factors
We identified motives for filicide using the six main categories utilised in filicide research today: altruistic, accidental (including fatal abuse and neglect), neonaticide, psychiatric, retaliatory and unwanted child (see Table 5). We then applied Wilczynski's (1997aWilczynski's ( , 1997b proposed risk factors to our sample. Detailed biographic histories were precluded because defendants were predominantly working-class men and women who left little behind in the way of a historical record. In many cases, numerous possible motives could be gleaned from case depositions, leading us to record an apparent primary motive and any possible secondary or tertiary motives. The motive could be garnered from defendant confessions to police, suicide notes, witness testimonies or medical depositions where the accused was examined to see if they were fit to stand trial. We then evaluated the risk factors apparent in each case. IJCJSD 10(2) 2021 Perhaps unsurprisingly, unwanted children was the most common motive for filicide as either a primary or secondary motive (48%), making this factor the most typical of the nineteenth and early twentieth century. All 'unwanted child' killings were infanticides, most commonly perpetrated within the first two months of birth. These killings were typically perpetrated by women aged between 18 and 24 years of age who were employed in some form of domestic service or hospitality. In many cases, the women were living at their place of employment.
Almost all children who were victims of 'unwanted child' deaths were illegitimate. A common pattern appeared with women hiding the pregnancy from their family and employers, giving birth at a lying-in hospital, and then attempting to dispose of the infant secretly some weeks later. We have designated neonaticide as a separate motive category because the actions of the mother and the death of the child were distinct to other motive categories (including 'unwanted'). In the case of neonaticides, women often gave birth alone, muffled the child or tied something around the mouth so it would not cry out, before strangling or drowning the infant and placing the body somewhere it would not be found. Common risk factors included social isolation (many women had relatives in other states), no support from the father of the child (due to desertion, or the mother and father were not in an ongoing relationship), shame, and fear of being exposed and losing employment. This fear was so great that in some cases, women committed filicide and then immediately returned to work-in 1892, Mary Fitzgerald gave birth to her child, drowned it in a water hole in a nearby paddock, and immediately proceeded back to work to ensure she kept her job.
Two men perpetrated infanticide on unwanted children. Here, the risk factors were similar: illegitimacy, possible shame and lack of social support. There is evidence that one of the men, Charles Glanfield, was unemployed. However, there was also evidence that Glanfield wanted to force a miscarriage while his partner was pregnant. After birth, he expressed wanting to rid himself of the child and ostensibly did so against his partner's wishes. In the second case, James Greenhalgh poisoned his infant after the mother sued him for refusal to contribute maintenance payments to the child.
Altruism was a common primary motive in infanticide and filicide cases, particularly those perpetrated in the 1890s and at the turn of the century. Economic depression in the 1890s contributed to financial hardship in colonial working-class families. Perpetrators framed their actions in altruistic terms, wishing to spare their children from a slow death by starvation. Where altruism was a primary motive, the cases were accompanied by suicide attempts where parents could not bear to leave their children to suffer once they were gone. Emma Williams, a perpetrator vilified by the press and executed for the murder of her two-year-old son in 1895, was living in poverty, with no family support or support of the father. Williams had endeavoured several times to adopt out her son or place him in the care of family interstate, to no avail. Williams was employed as a prostitute, and as was the reality with many working women in the late nineteenth century, children hindered any chance of future work.
Mental illness was the most difficult risk factor to establish. From the 1870s, the insanity defence became increasingly popular in trials for filicide (Rychner 2017). There were few practising psychiatrists in Victoria during the late nineteenth century, and little in the way of consistency among diagnoses. Few offenders had medical records regarding mental health prior to their crime. While mental illness was alleged as a risk factor for 44% of perpetrators in the sample, this figure should be treated with caution.
Accidental killings occurred evenly across infanticide and filicide cases. All perpetrators of accidental killings were women, except for one married couple who were both charged. Poverty, alcohol abuse and low socio-economic status were the most prevalent risk factors for accidental killings occurring from both fatal abuse and neglect. In two fatal abuse cases, there was evidence of prior attempts to place infants or children into care. In one instance, the child died while in care and the mother was convicted for manslaughter along with the carer.
Fatal abuse filicides were often the endpoint of a history of physical abuse enacted by mothers. In the case of Clara Grumley, police had filed a report regarding her mistreatment and repetitive beating of her fiveyear-old step-son a year earlier. Prior convictions or allegations of abuse surfaced as a major risk factor in these cases. Alcohol arose as a factor where women were living in poverty, had no stable housing or employment, and in one case, had a prior conviction for vagrancy (Bridget O'Day). Catherine Kelly's infant was found dead in her arms when police arrested her for drunkenness in 1901. In cases such as these, defendants often stated that they had done all they could for their children, spending what money they had on food and clothing for them. Their infants often died from exposure or malnourishment.
Only five cases of the sample could be classed as retaliatory among either primary or secondary motives. Retaliatory killings today are more commonly found among male perpetrators (Brown et al., 2019), yet the majority of perpetrators from the historical sample were women. Mental illness was a predominant risk factor in these killings. All perpetrators in retaliatory killings were married, with the crime often committed as the culmination of sustained marital grievance. For example, Sarah Williams was a mother to seven children, ostensibly isolated in terms of social support (her relatives lived in South Australia) and was constantly trying to feed and clothe her children. Though he was employed, her husband spent his wages on drinking and gambling, often returning home late, if at all. The infanticide occurred in the course of a public confrontation between husband and wife. Williams, having located her husband in the city, demanded money to buy medicine for the youngest child who had been ill for some time. When her husband refused, she said, 'then see what I'll do with it' and threw the infant off the Falls Bridge (The Argus, 17 Dec.1885, 11).
In the one case of a male-perpetrated retaliatory killing, the crime was committed on the same day that his partner left the relationship. The perpetrator had a history of possessive behaviour leading up to the crime. At the committal hearing, he stated that 'this is all caused by my wife drinking and going with other men' (VPRS264/P0001/6). This mirrors contemporary offenders' motives for killing their families (Brown and Tyson 2012;Debowska, Boduszek and Dhingra 2015). Domestic abuse figured in two cases. Both Mary Fitzgerald (1892) and Ellen McNab (1898) had been married to husbands who had thrown them out of home. Both women mentioned violence on the part of their husbands while married, and McNab's parents saw first-hand the injuries sustained during her marriage. Keeping in mind the difficulties of locating known instances of domestic violence in the historical record, physical, sexual and verbal abuse was common in marital households in the late nineteenth century (Allen JA 1990). This violence occurred within a culture of strict non-interference, with many women staying silent or approaching religious sisters rather than neighbours or police. Regarding homicides, children were not being used as retaliatory victims historically as much as they are now. Between 1890 and 1900, nine men were convicted for the murder of an intimate partner in Victoria; undoubtedly this number is conservative as more recently, between 2010 and 2014, there were 26 convictions for intimate partner homicide in Victoria (Australian Domestic and Family Violence Death Review Network 2018). However, recent data suggest that the victims of men's violence have changed since the historical data were collected. From this study, we can see that men were unlikely to kill their children, and there was no case of a man killing a stepchild. The historical victims of men's violence in cases of domestic dispute were almost wholly women, while women were responsible for most child deaths. Between 2000 and 2012, men were offenders of filicide in 53% cases, and in 14% of cases, it was a stepfather who killed children in his care (Brown et al., 2019). There has been a dramatic shift between 1920 and 2000 in men's filicide offending, which warrants closer examination for future research.

Filicide-Suicides
Our research supports the findings established by Shackelford, Weekes- Beasley (2005, 2008) who first analysed filicide-suicide cases in Chicago from 1965 to 1994 and then from 1870 to 1930. Their findings suggested that filicides with multiple, older children as victims would more likely end with the attempted or completed suicide of the mother. They also suggest that older parents were more likely to attempt or complete suicide after killing their children. Although our research did not investigate cases where suicide was completed by the perpetrator, the suicide attempts by parents within our dataset fit this pattern.
By considering the Victorian dataset, it appears that the majority of historical filicides were infanticides committed on unwanted children. Yet the filicide data reveals a strong pattern of filicide-suicides that has evaded historical attention in Australia. In our sample, 20% of filicides were followed by a suicide attempt. Most filicide-suicides were perpetrated by married women in the home, with an attempt to kill more than one child. All perpetrators were aged between 30 and 41. These women wished to commit suicide but feared for the fate of their children once they were gone. In this sense, the killings could be considered altruistic. Although much was made of Mary Carnie's physical pains as a defence for her behaviour, she felt that she was a burden on her husband. She attempted to kill her children and take her own life so that her husband could 'marry the woman he wished to marry'. In the cases of Ellen Scarcebrook (1891) and Ellen Youngman (1899), the perpetrators were afraid they were 'mad', that they would end up in the 'madhouse' and abandon their children. They were also convinced they had passed on a hereditary, maternal mental illness to the next generation.
Common risk factors across all suicide-filicides included mental illness, isolation, fear of financial insecurity and anxieties about providing for children during life or after death. Significantly, seven out of 10 filicide-suicides were perpetrated in the 1890s, the decade of the financial crisis that peaked in 1893 and left families feeling the effects for years. The one case involving a male perpetrator was a familicide. James Johnston killed his wife and four children in 1891 before attempting suicide himself due to significant financial debt and the perceived inability to provide for his family.

Conclusion
This pilot study examined the longitudinal trends of filicide offences in Victoria between 1860 and 1920.
The aim was to see if current criminological frameworks for filicide, with recently established motives and risk factors, could be applied to a historical sample. We have argued that the narrow focus on maternal infanticides in historical research has made it difficult for criminologists to utilise findings from historical data.
The focus on maternal infanticides and case study analysis, rather than a longitudinal analysis of filicide cases, has limited the collection of research applicable to criminological studies into filicide in Australia today. This article goes some way in rectifying that problem in Victoria, and it is clear that similar work needs to be done across the country. Historical research has the potential to meaningfully contribute to contemporary criminological concerns, as demonstrated by studies in other countries. It is necessary for Australian criminology to connect with Australian history and develop its own historical criminology research to ensure a solid foundation for theoretical explorations of crime, violence, victimisation and its criminal justice systems. Relying upon European and US historical or longitudinal data for exploration of contemporary Australian concerns will mislead scholars and policymakers in the long term. This article has demonstrated that contemporary theory can be applied in historical filicide cases. It also demonstrates that merging historical and criminological research can uncover a greater pool of records for criminologists (court and inquest case records as well as prison ones, among others).
The current characteristics of filicide offending in Australia, such as risk factors and motives, are somewhat borne out by our analysis of historical cases. For younger mothers, the risk factors of financial instability and lack of family or community networks continue to contribute to infanticide and neonaticide perpetration. Comparison of this data with that conducted by Shackelford, Weekes- Beasley (2005, 2008) in Chicago suggests that there is a continuity in the risk of older parents killing multiple children and attempting suicide. Categories of motive are applicable to the historical sample and unearth historically specific contexts such as a heavy colonial drinking culture, the mobility of family networks between colonies, the isolation of family life in rural towns, the inadequate access to contraception and the very real consequences of unintentional pregnancy on women's financial stability and means of survival.
However, the profile of the offenders differs markedly between our research findings and what is characteristic of offenders today. As our research into historical data demonstrates, fathers were in the minority when it came to filicide offending. In contrast, today in Victoria, male filicide offenders account for 62% of offenders, and men comprise 52% of filicide offenders in Australia more widely (Brown et al., 2019). Family violence, though difficult to discern in the historical record, has emerged as a major risk factor in filicide perpetration. Also, those who committed retaliatory killings on children were historically female, in contrast to the recent trend of high-profile male retaliatory killings in Australia. Although 14% of filicides today are committed by a stepfather (and stepmothers were not responsible for any filicide cases between 2000 and 2012 in Australia) there were no stepfathers in our research sample (and one stepmother). This indicates that between 1920 and 2000 there has been a marked shift in who perpetrates filicide in Victoria, and Australia more broadly. The reasons for this shift are unclear at the moment but are certainly worth investigating. It is clear that longitudinal trends between historical data and the present-day need to be examined. Our research demonstrates that historical criminology is a powerful research tool and has the potential to inform prevention strategies and responses to filicide offending and victimisation in present-day Australia.