Imprisonment of Female Urban and Rural Offenders in Victoria, 1860-1920

This paper examines imprisonment data from Victoria between 1860 and 1920 to gather insights into the variations in incidence of women being convicted by rural versus urban courts, including close focus on the difference in types of offences being committed in urban and rural locations. This paper also details women’s mobility between both communities as well as change in their offending profiles based on their geographic locations. Our findings suggest that while the authorities were broadly most concerned with removing disorderly and vagrant women from both urban and rural streets, rural offending had its own characteristics that differentiate it from urban offending. Therefore, this demonstrates that when examining female offending, geographic location of an offender and offence must be taken into consideration.


Introduction
Criminologists have traditionally focused their attention on crime in the urban environment, not only in Australian literature but also internationally (Carrington 2007;DeKeseredy 2015). This is not without validity, as a large proportion of crime is recorded within dense urban populations. Concern with the urban environment and its perceived criminogenic properties have been the focus of social scientists since the major move from agrarian to more industrialised societies. The consequent population shift from rural to urban environments was problematised and located as the cause for changes in suicide rates, interpersonal conflict, property damage and general social malaise (Durkheim 1895(Durkheim /2007Tönnies 1957;Wirth 1938).
In the last three decades, there has been a growth in scholarship examining crime in rural communities, including criminogenic risk factors for offending and victimisation. Recent exploration of crime and policing in rural Australia highlights the very different issues that these communities in Australia face in comparison to their international counterparts. The literature has also argued that the bucolic image of regional and rural Australia as a place of peace, tranquillity and safety is not based on reality, with rural communities often facing similar crime rates and issues as urban environments (Carcach 2000;Carrington 2007;Hogg and Carrington 2006;Jobes et al. 2000). Since the 1980s, attempts have been made to deliver historical, longterm quantitative data of both urban and rural convictions by researchers but with little analysis of gender differences, nor contextualisation of the statistical data (Mukherjee, Walker and Jacobsen 1986). Research into historical crimes in rural Australia has also tended to focus on bushranging (with particular focus on Ned Kelly and the Kelly Gang), policing and unrest on the goldfields or issues with squatters (Disher 1981;McQuilton 1987;Thurgood 1988). This focus has arguably wrought a distinctive male focus in the research so far. As such, this article begins to redress this imbalance in contemporary and historical criminology knowledge by examining urban and rural women's imprisonment in Victoria between 1860 and 1920.
This article examines the offending patterns of 6,042 women imprisoned for the first time in the Australian state of Victoria between 1860 and 1920; of this group geographic data is noted for 6,027 prisoners. This dataset represents the first longitudinal study of women's criminality in Australia and one of the largest studies of historic female offending to date. It enables insights into the variations in incidence of women being convicted by rural versus urban courts, including close focus on the difference in types of offences being committed in urban and rural locations. Further, it details women's mobility between both communities as well as change in their offending profiles based on their geographic locations. Our findings suggest that while the authorities were broadly most concerned with removing disorderly and vagrant women from both urban and rural streets, rural offending had its own characteristics that differentiate it from urban offending. Therefore, this demonstrates that when examining female offending, geographic location of an offender and offence must be taken into consideration.
Defining what is considered 'rural' or 'regional' is difficult. This is due in part to definitions of rurality changing based on the discipline of investigators, as well as traditional Australian notions of regional and rural as being anything that is outside the major coastal capital cities (DeKeseredy 2015;Hogg and Carrington 2006). For the purposes of this study, Melbourne is the urban location against which 'rural' data are being compared. Melbourne grew exponentially during the nineteenth century and towns that were once on the periphery and not part of greater Melbourne in the 1860s became suburbs by 1920. For example, Dandenong was not considered part of Melbourne until after the postwar period but was an important rural township during the latter part of the nineteenth century when it acted as the gateway between the Gippsland region and the city of Melbourne. For this reason and simplicity's sake, this article will use metropolitan Melbourne boundaries from 1900 to delineate the urban region. Larger settlements outside of Melbourne-those that could be classified as regional centres-will also be classed as rural rather IJCJ&SD 102 Online version via www.crimejusticejournal.com © 2019 8(1) than urban because despite having slightly denser populations than other rural townships, their demographics and socio-legal methods were more akin to rural rather than urban environments.
Unsurprisingly, crime has been, both in the past and present, associated predominantly with men as offenders, with feminist criminology itself tending to focus on the dynamic of crime with men as offenders and women as victims (Carrington 2013;Covington and Bloom 2003;Mazerolle 2008). Contemporary criminological research into gender, crime and rural Australia has identified violence and criminal offending as driven in part by narrow definitions of masculinity in rural locations (Carrington 2007). Research into historical instances of rural criminality and violence in Australia has overwhelmingly focused on individual cases, favouring emphasis on men's offending over women's (Highland 1994;Phillips 1994). Yet, historically, in Australia and internationally, women and girls were often involved with antisocial and criminal behaviour in both urban and rural environments at much higher rates than those recorded today (D'Cruze 2000;Nagy 2015). The contemporary oversight of women's and girls' historical involvement with the criminal justice system as offenders has led to a skewed understanding as to where they in Australia fit in the criminal justice system, as well as a lack of awareness of the potential sources of twentieth-and early twenty-first centuries' women's increased offending rates.
Towards the end of the nineteenth century more women were imprisoned in Victorian jails on a day-to-day basis than any other Australian colony (Ross and Skondreas 2000). Women were routinely imprisoned in colonial Victoria for a range of offences ranging from petty (e.g., indecent language in public) to serious crimes, such as murder. Historical research into female offending in Victoria has typically concentrated on those crimes perceived to be most heavily gendered, particularly sex work and reproduction-related offences (Finch and Stratton 1988;Frances 2007;Goc 2013;Laster 1989;McConville 1980;Rychner 2017). The former has been described as being located most heavily within Melbourne or on the goldfields, while the latter-that is, abortionhas typically been depicted as a mostly urban crime, with rural women travelling to the city to procure the service. Likewise, infanticide, though recognised as not being restricted to the city, has been described as particularly associated with the urban context (Swain and Howe 1995: 91). Another focus in historical scholarship has been female murderers subjected to capital punishment (Cannon 1994;Laster 1994;Overington 2014), with little focus on spatial location. Research into more general patterns of female offending in Victoria during the late nineteenth and early twentieth centuries has only recently started to emerge, with analyses on risk factors for women's imprisonment, as well as offence variation in women's criminal careers Nagy 2017, 2018). Therefore, this paper represents an important contribution to the historical and criminological literature, which have seldom explored the relationship between female offending and geographic location.

Urban and rural women in historical Victoria
As with other parts of the Western world, there has always been a dichotomy between the urban and rural environments in Australia. Life outside the city was often considered to be, if not exactly idyllic, at least more morally uplifting than the polluted slums of the urban landscape. Moral welfare societies in Victoria mirrored their overseas counterparts in advocating the removal of fallen women and delinquent children to rural situations and institutions where they might have the best chance to reform (Finnane 1997: 76-77). At the same time, the bush was popularly imagined as a harsh and dangerous landscape, especially for women-although, one that could also offer greater freedoms (James 1989;Murphy 2010). Australian art, literature and culture have imagined the female bush dweller to be more akin to her rural male counterpart (rugged, egalitarian and independent; e.g., as portrayed in Henry Lawson's The Drover's Wife) than to urban women. While the reality of such depictions remains debatable, colonial women living in rural Victoria were subject to far different conditions and concerns than their counterparts in Melbourne. For a large portion of the nineteenth century, men outnumbered women in Victoria. In the 1861 census, there were 153.6 men for 100 women in a colony that had a total population of 540,322 individuals. However, Melbourne did not suffer from this disparity, with many suburbs or municipalities having a higher female than male population. Geelong, Victoria's second most populous area, also had more women than men listed in the census, but not by a significant margin (11,834 women to 11,125 men). Instead, it was in the mining districts (such as Ballarat and Bendigo)-where 42% of the colony's population resided-that contributed to the sex ratio difference, with 154,692 male inhabitants to only 73,489 women. By the 1881 census there was more parity between the number of women and men being recorded around Victoria, with nearly 91 women to 100 men, which was on par with the other Australasian colonies. Women outnumbered men in Melbourne and its urban municipalities (220,843 women to 213,624 men), and were beginning to dominate not only in Geelong but also in the central area of Ballarat  In addition to being distributed differently across the urban and rural environments, women in Victoria faced different socio-economic conditions dependent on their location; although, some factors were shared. For example, the 1871 census notes that only 20% of women in Victoria had a form of employment that was not domestic duties associated with being a wife or daughter, nor were they under the care of the state. Women as a group had fewer legitimate means of earning a wage than men, but those in rural areas faced even more limited employment prospects, with scarcer opportunities in the manufacturing and retail industries than in the city, as well as limited availability of labour in the hospitality industry around the goldfields (Kingston 1975). Factors affecting the number of such positions being offered-such as the economic depression of the 1890s or the absence of male workers on military service during World War I-may have had a particular effect on the lives of urban working-class women (Frances 1993).
Marriage trends were also likely to affect women's socio-economic position and resulting criminality. In the 1881 census, a trend was noted throughout the colony where there were more wives than husbands. Such disparity would indicate that the missing husbands were either residing outside the colony, claiming they were unmarried or were otherwise absent from the lives of many Victorian women. Wife desertion was certainly common and was lamented by authorities as a cause of female crime and 'immorality' (Twomey 2002). Conversely, the high marriage rates among women in rural areas that resulted from the sex imbalance in many of those regions was liable to have had a reducing effect on female crime within those environments, given the historical correlation between female offending and lack of a male provider (Piper 2015

Method
The dataset used in this analysis is drawn from the Central Register of Female Prisoners, a series of records created by Victoria's penal department to register prisoners' names, personal details and convictions. Upon a woman's first entry to prison a record was created for her that would be added onto any subsequent returns (minor or prior convictions that had not resulted in imprisonment were also usually noted). The format of the record-keeping system remained consistent across several decades and included the following details: biographical information such as birthplace, year of birth, year of arrival in Victoria (if a migrant), religion, occupation, literacy and marital status; conviction details including trial date, offence, court and sentence; descriptions of women's appearances including height, weight, tattoos and physical conditions, as well as mugshots, which began appearing in the late nineteenth century; and occasionally other comments about a woman's family history, their behaviour inside prison and details of their release or transfers to other institutions, such as charity homes or 'lunatic' asylums.
This register provides a sample of 6,042 individual women who first entered the central prison system between 1860 and 1920. It does not include the women incarcerated in Victoria's prisons during this period who had been imprisoned prior to 1860, as their records exist in earlier registers. However, three women in the sample had convictions that occurred prior to 1860, the earliest occurring in 1854 but as they were not imprisoned at that time they did not have prison records. There were also 124 women in the dataset who continued offending after 1920, the last known conviction occurring in 1947. While the information contained in this data is rich, its limitations should also be noted: biographic information such as occupation and marital status were recorded upon women's first imprisonment but were seldom updated on subsequent returns, and women may have purposively misled officials when asked to supply details about their background. This latter point manifested in aliases, and court registrars and prison officials were generally very aware of the propensity for this and could effectively record and link aliases to existing prison records. Recidivism levels may also be somewhat underestimated due to officials' occasionally lax approach to noting details of minor convictions that had not resulted in imprisonment.
Information about geographic location has been obtained from the prisoner records through the listed courts where women were tried and convicted. There were three levels of court one could find themselves: Court of Petty Sessions, Court of General Sessions and the Supreme Court. The Court of Petty Sessions (sometimes also referred to as Police Courts) was the lowest court and initially dealt almost solely with minor criminal matters such as drunkenness, vagrancy and the most trivial class of larcenies and assaults. These small courts were located all over Victoria, meaning that the Petty Sessions Court hearing a matter was almost always located very close to where the crime itself occurred. The Court of General Sessions (established in 1852) was a midpoint between the Court of Petty Sessions and the Supreme Court, empowered to try cases of non-capital crimes before juries. The Court of General Sessions was initially only established in Melbourne and Geelong, but was later introduced to other areas such as Ballarat. Finally, the highest court in Victoria was the Supreme Court, which heard matters not only in Melbourne but also from beyond the city. As with the system in England, Supreme Court judges would attend Circuit Courts in rural locations such as Ballarat, Bendigo and Geelong, as well as Beechworth, Ararat, Castlemaine, Maryborough and Sale to hear cases of serious crimes. Women charged with more serious offences would generally be committed to the General Sessions or Circuit Court Sessions located closest to where their crime took place, but this could be some distance if they were living remotely to any of the major centres. However, the vast bulk of female offending was heard by the more proximate Petty Sessions, with more than 80% of female prisoners only ever tried in this jurisdiction over the course of their criminal careers (Piper and Nagy 2017

Offence profiles of female prisoners in urban and rural Victoria
Overall, the offence profiles of female prisoners in Victoria between 1860 and 1920 were similar to results from other research in Australia and abroad for roughly the same period (Allen 1990;Zedner 1991). The bulk of female offending was weighted towards minor public order offences such as vagrancy and disorderly conduct, while theft offences greatly predominated over violent crimes when it came to more serious criminal activity (Piper and Nagy 2017). This pattern is broadly evident across women imprisoned by both urban and rural courts. However, when more specific offence types are examined some interesting divergences become apparent.
The prison registers identify 6,042 unique female offenders in the period of 1860 to 1920, of whom 6,027 have location data recorded for the court that had first sentenced them to imprisonment. The data suggest that for around three-quarters of female prisoners this was also their first conviction overall, an indication of a lack of sentencing alternatives in this period beyond the use of fines that some women found themselves unable to pay due to poverty (Finnane 1997: 33-35). Of these 6,027 women, 73.3% were sentenced from a court located in the greater Melbourne area (see Table 2). Even allowing for the disproportionate concentration of Victoria's female population in the capital during the colonial period, this figure suggests that urban women were over-represented in the prisoner population. Although Melbourne's female population accounted for only 44.8% of the female population of Victoria in 1881, Melbourne courts contributed over 78% of female prisoners. This supports contemporary research about the criminogenic properties of cities and the effect that urban disadvantage and poor living conditions have on the creation of illicit economies and behaviours (Schwartz and Gertseva 2010). It also supports characterisations of Melbourne as a cesspool of crime by nineteenth-century commentators (Davison and Dunstan 1985).
Many of the offences where urban women were particularly over-represented seem to confirm Melbourne's possession of an underworld economy (see Table 2). While urban courts accounted for 73.3% of first-time female prisoners overall, this proportion rose higher when it came to many theft offences including robbery (85.3%), larceny from the person (82.9%), burglary (77.3%) and receiving stolen goods (76.7%). Urban rather than rural women would also find themselves imprisoned at higher rates for aiding and abetting or intent to commit a felony (81.8% to the rural 18.2%). This aligns with contemporary findings that show it is property crime that most 'closely approximates the conventional wisdom that crime rates tend to increase with population size and concentration' (Carrington 2007: 33). The colonial city probably provided greater criminal opportunities for certain types of theft than rural areas. For example, robbery and larceny from the person by female offenders was predominantly committed within the context of sex work (Piper 2018), which, while not confined to the capital, was particularly prevalent there (Frances 2007). It is also possible that, as Hogg and Carrington (2006: 120-121) discovered in the modern context, there was a lower detection (and consequently imprisonment) rate for property offences in rural locations, especially for crimes such as break-ins, as police were often reliant in this period on eyewitness accounts by neighbours or passing foot traffic. Closer proximity to neighbours in urban areas may have also led to higher reporting levels of other types of crimes such as threatening life or to cause harm, with 84.7% of women imprisoned for this offence tried by Melbourne courts. Large differences in rates between urban and rural offenders for some other offences, such as suicide (threatened or attempted), might also have been due to closer proximity to neighbours who were in a position to intervene in and report attempts at suicide in Melbourne versus in rural communities.
The public nature of working-class life in colonial Melbourne and its possession of a street subculture may have likewise influenced the slightly elevated proportions of urban women first imprisoned for vagrancy (77.2%) or disorderly conduct offences (77%). These offences were routinely used to police female sex workers (Frances 2007) and appear to have been deployed in this capacity far more often than specific sex work related offences (such as soliciting or brothelkeeping), for which urban women were less strongly represented in the dataset. Additionally, homelessness, as today, was also a genuine problem, with colonial commentators observing that Melbourne was inhabited by high numbers of female beggars, most of which were concentrated among the very young and very old but could be found resident in the city year after year (Freeman 1888: 133). Contrariwise, Julie Kimber (2010) has suggested that in rural towns in Australia vagrancy charges were historically used as a means of moving 'problem' women (notably drunkards or sex workers) on to other areas.
Some of these offences speak to the particular criminal opportunities of rural environments. For example, the crime with the highest proportion of rural offenders (selling illegal liquor) was an understandable occurrence in remote areas with fewer established hotels. Perceptions of hospitality and food preparation as being feminine qualities and pursuits meant many women in colonial Victoria were able to find employment in either legitimate hotel-keeping (Wright 2003), or, like bushranger Ned Kelly's mother Ellen, in more illicit sly-grog selling (Lake 1985). Sly-grog shanties were particularly common on the goldfields, with diarist Samuel Curtis Candler (1848-19: 126) recounting an anecdote in the late 1860s about a particularly colourful character who used to go around the diggings with a two-gallon tin rum tank disguised under her dress as a pregnancy belly.
Likewise, while many forms of theft were linked to the urban centre, it makes sense that stock offences would be concentrated more in rural locations where there were more cattle, sheep, horses and other animals to steal-although, the low number of women convicted overall of this predominantly male-perpetrated offence means that from a frequency rather than a proportional perspective there was little difference between urban and rural offending rates among women. The increased proportion of rural women among those imprisoned for fraud offences presents more of a puzzle. This was possibly connected to a greater willingness among rural business owners to cash cheques or provide goods on credit, inevitably creating opportunities for some individuals to present forgeries or obtain items under false pretences. Conversely, the less anonymous nature of smaller towns perhaps meant people who practised such deceptions simply found it less easy to evade detection.
Lack of anonymity in rural areas also likely increased both the reputational threat posed by deviant behaviour and the likelihood that such behaviour would ultimately be uncovered. This would explain why rural women accounted for the majority of those imprisoned for concealment of birth, a charge usually brought against women who had failed to register the births of illegitimate children (Goc 2013: 3). Indeed, this often acted as an alternative charge to infanticide when women alleged a child had been stillborn. Other forms of female deviance, such as drunkenness and obscene language, may have also been less tolerated in close-knit regional centres, which were structured around institutions and symbols of civil and moral authority in an effort to curb the 'loose' behaviour associated with the rough-and-tumble settlements on the colonial margins. Women within rural communities may have also been less able to pay the fines usually offered as alternatives to imprisonment in relation to such offences, either because their support networks were not as extensive as urban women or because there were less independent economic opportunities.
The need to enforce order onto the chaos of the frontier also potentially influenced the policing of interpersonal violence, at least around non-fatal assaults. Police in rural Victoria were different to their urban counterparts, with a more paramilitary model being employed outside Melbourne; the law was also upheld in a very selective fashion by rural Victorian police (McQuilton 1987). This did not translate into an increased imprisonment of non-white women in the way that it sometimes did for non-white men. Only four female prisoners in the sample were identified as Indigenous; however, some others may have also had Aboriginal heritage (Piper and Nagy 2018). As Grant (2014) has explored, few Indigenous women were jailed during this period unlike Indigenous men.
However, it is noticeable that the imprisonment rate of rural women for both assaults and homicide-a crime unlikely to be subject to selective policing-was higher than the general imprisonment of rural women (but still under-represented compared to the overall female population). More contemporary studies have found that women's violence is strongly driven by rates of male violence in rural environments (Parker and Reckdenwald 2008). Some of the factors suggested by recent scholarship as contributing to violence in modern rural areas-such as isolation, heavy-drinking cultures and more limited access to support services-are likely to have also been relevant historically (Hogg and Carrington 2006: 65;Jobes et al. 2001). These would presumably act as additional risk factors for other types of crimes with stronger associations with rural areas in the sample, such as suicide, arson and property damage.

Offending patterns and spatial locations over time
While the broad patterns of female offending were fairly similar across urban and rural locations, an examination of the specific offences that resulted in women's entry into the prison system reveals considerable variation produced by environmental conditions. Beyond this, location also influenced women's offending patterns across time.
Over the course of one's criminal careers, a significant proportion of female offenders would commit different types of offences (Piper and Nagy 2017). Overall, the majority of women's offending, considering all convictions, still comprised of public order offences, followed by theft offences, then violent and other types of crimes ( Table 2). As with first offence data, factoring in location produced some significant trends. For example, rural women still showed a particular likelihood to be convicted of crimes that fell in the 'other' offences category, such as arson, attempted suicide and concealment of birth. Conversely, urban offenders make a greater showing among violent offenders when looking across criminal careers; although, this may be the case because this broad category includes more minor forms of violence, such as threatening behaviour.
In addition to engaging in different criminal activities, across time female prisoners were also liable to commit offences in different locations (see Table 3). All prisoners 6,027 70.6 24.9 4.5 * Offence categories are not mutually exclusive, as a prisoner could commit more than one type of offence over her criminal career. a  2 (2) = 67.83, p < 0.001 b  2 (2) = 23.59, p < 0.001 c  2 (2) = 7.18, p < 0.05 d  2 (2) = 6.07, p < 0.05 In all, 4.5% of the 6,027 female prisoners in the sample amassed convictions in courts both inside and outside of Melbourne. These mobile offenders made a particularly strong showing among those who had at least one conviction for theft at some point (6.4%), but were under-represented among those with a conviction for violence (3.7%). Perhaps, given that female violence has been routinely linked to the context of close interpersonal relations (Schwartz and Gertseva 2010), it was more likely to take place among those embedded in permanent, ongoing networks. Meanwhile, the risks associated with identification in theft cases perhaps in itself acted as an inducement to mobility among such offenders, encouraging known criminals to move to new areas where their face and modus operandi were less known. In discussion of the history of property crime in London, William Meier (2011: 41-66) notes that the increased mobility IJCJ&SD 109 Online version via www.crimejusticejournal.com © 2019 8(1) enabled by new systems of transportation meant that the 'traveling thief' was one of the major challenges faced by law enforcement across the late nineteenth and early twentieth centuries.
It was not only the type of offending that differed between urban, rural and mobile offenders, but also one's overall level of offending. The average mean time between women's first-and lastknown convictions was shorter for rural offenders (M = 1,018 days, SD = 2,084) than their urban counterparts (M = 1,305 days, SD = 2,343); however, both were far exceeded by the average criminal career of mobile offenders (M = 4,638 days, SD = 3,852). There was a similar trend in terms of number of convictions. While metropolitan offenders had a mean average of four convictions (SD = 7) to rural offenders' mean of three (SD = 5), the prolific recidivism of mobile offenders resulted in an average of 12 convictions (SD = 14).
These findings align with research strongly associating mobility with higher levels of offending (Barnett and Mencken 2002;Steffensmeier and Haynie 2000). Similarly, Clinard (1942) in their study of youth mobility between rural and urban environments discovered that young populations who were exposed to urban values and returned to rural communities exhibited criminal offending typologies more often associated with urban offenders. Brown (2011) also found such patterns when examining the conviction histories of a sample of 427 men imprisoned at Dartmoor in England in 1932. Brown's sample reported a much higher rate of mobility overall than that found in the Victorian female sample; however, this is probably due to differences in definitions and sampling techniques. Our designation of mobile offenders only takes into account mobility between urban and rural contexts (but potentially hides mobility across different rural contexts) or between a variety of Melbourne suburbs. Further, while our sample looks at prisoners across the whole of Victoria, Brown (2011) examined a prison that housed mostly serial offenders; thus, if recidivist offenders were more mobile, it would make sense that such a sample showed a high rate of mobility. The gender differences between the samples are also likely important, given that men historically were prone to live more mobile lives than women as a result of movement connected to the pursuit of employment (Lake 1986). The differences perhaps also point to the possibility that levels of mobility shifted across sociohistoric context, with Brown (2011: 561) underlining the significance to her study of motor-car theft in the interwar period, a factor not present in our 1860-1920 sample.
Changing conditions over time also likely resulted in changing numbers of rural, urban and mobile offenders within the 1860-1920 sample itself. Analysis of the number of women entering the prison system each year against the location of the courts in which they were convicted across their criminal careers reveals no single, consistent trend, but rather a number of significant fluctuations in the proportion of rural, urban and mobile (urban and rural convictions) women ( Figure 1).
For example, whereas almost four-fifths of women imprisoned in the 1860s would only ever be convicted in Melbourne courts, this rate fell considerably to around two-thirds of women entering the system in the 1870s and 1880s. The successive decades-which saw a decline in the overall number of women being imprisoned-then brought a progressive rise in the proportion of those women whose convictions were urban based. The proportion of mobile offenders also rose during the early 1900s. Perhaps, then, rural women disproportionately benefitted from changes in sentencing practices-such as the release of first-time offenders on probation-that saw fewer women entering the prison system overall (White 1979).

Offender profiles of female prisoners across spatial locations
Some of the variations in offending patterns outlined above may also be linked to differences evident in the personal characteristics of urban, rural and mobile offenders. Statistically, concentration of most criminal offending is in the adolescent and early adult periods in Australia (ABS 2017); however, this trend shifts in the historic female prisoners' dataset. Whereas modern research suggests the teens and early twenties as peak offending periods, with a median age of 28 years, for Victoria's female prisoners in this era it was their twenties and thirties, with a median age of first-time entry into prison being 32 years. However, age profiles shifted significantly when compared against the locations where women were convicted (Table 4).
Urban offenders tended to be younger (MED = 30 years, SD = 10) than rural (MED = 35 years, SD = 13) or mobile offenders (MED = 32 years, SD = 13) at first conviction. This means that early onset of offending may, in turn, have led to increased levels of offending persisting over one's life, which is most evident among urban offenders when compared to their rural counterparts. Consequently, this may explain why urban offenders were particularly likely to be listed as having never been married on their initial entry to the prison system (Table 4).
Conversely, rural women accounted for an increased proportion of those first convicted after 30 years of age and whose last conviction did not take place until they were over 50 years old; this includes women who were widowed on first entry to gaol (see Table 4). This may indicate a particular 'problem' group of older women in rural or regional communities who in the absence of strong familial bonds-and perhaps with declining employment prospects-found themselves cast on the support of the justice system. The elevated proportions of rural women imprisoned for drunkenness and indecent language may indicate that these were the primary ways such women were policed, or that such behaviours became more common among rural women with age and the removal of social controls exerted by patriarchal figures.  occupation as servant. However, almost all those whose listed occupation indicated that they were employed in domestic duties within their own home came from rural areas, perhaps indicating the higher levels of employment of women outside the home in urban areas. Likewise, most of those engaged in more middle-class occupations-such as nursing, teaching, journalism, acting or shop-keeping-were urban offenders.
Surprisingly, one socio-economic indicator that did not vary much by location was literacy level. Despite concerns expressed during the late nineteenth century about education levels in some parts of regional Victoria (Barcan 1980), there was no significant decline in literacy among rural offenders. Historically, those born in Australia enjoyed far higher literacy rates than those born overseas due to the early introduction of free and compulsory schooling (Lyons 2001). Migration patterns did vary across rural and urban locations, with urban offenders less likely to have migrated from Great Britain and more likely to have been born in Australia. Interestingly, mobile offenders comprised an elevated proportion of those who had migrated from places other than Great Britain, indicating that coming from a non-English speaking background may have contributed to some mobile offenders' struggle to avoid the cycle of imprisonment, as well as showing that mobility was already a significant part of their histories. More contemporary research has found migration in conjunction with a non-English speaking background, social isolation and addiction have all contributed to an exponential rise of Vietnamese women, in particular, in Victoria's prisons (Francis 2014).

Conclusion
While for a large portion of the nineteenth and early twentieth centuries there were more women inhabiting rural communities in Victoria, offending was overwhelmingly located in Melbourne. Although across all spatial locations women were most likely to find themselves imprisoned for public order or theft offences, there were differences between the demographics of urban, rural and mobile offenders. Social control, or rather a lack thereof, appears to have contributed to greater rates of urban and mobile women's offending; however, rural women could often find themselves imprisoned especially in their later years. Across all groups, offending, or rather imprisonment, was declining right through to 1920, albeit at different rates and undoubtedly for various social reasons. Urban offending consistently was the greatest contributor to women's imprisonment rates but this should not preclude attention being paid to rural female criminality. Instead, attention could most especially be paid to not only why crime in the metropolitan environment continued to be higher than the rural one, but also to what protective factors may have contributed to rural women's decreased risk of criminal behaviour at a time when rural regions of Victoria were known for their crime and violence. Differences between the three groups of female offenders certainly highlight the importance of not only investigating the types of offences that women may have found themselves imprisoned for, but also the geographical place from which women entered the penal system, and from where re-entry would occur in cases of recidivism. Overall, the results from this study highlight the necessity for further research and spatial analysis of women's historic offending and imprisonment in Victoria, including, but not limited to, women's mobility across rural communities and movement within urban environments.