Introduction

1In the Western world, fertility generally declined during the second half of the nineteenth century. However, the onset and intensity of this change varied between countries – and sometimes even between regions in the same country (Coale & Watkins, 1986). In many countries, including Italy (Kertzer & Hogan, 1989), Switzerland (Praz, 2005), and England (Szreter, 1996), researchers have identified several fertility transitions. This was also the case with Canada in general (Joubert, 2013; Gossage & Gauvreau, 2007; McInnis, 2000) and Quebec in particular (Gauvreau et al., 2007).

2At the turn of the twentieth century, Quebec’s population was characterized by the presence of multiple demographic regimes associated with the provinces’ main ethno-religious groups, which were divided along religious (Catholic / Protestant) and linguistic (English / French) lines. D. Gauvreau, P. Gossage, and D. Gervais (2007) describe a variety of fertility regimes shaped by the material transformations associated with the establishment and consolidation of industrial capitalism, cultural changes and migrations. They suggest that, in comparison to other groups, French Canadians were slow to adopt “Malthusian” reproductive behaviour. The fertility transition by Protestants was evidently beginning in the late nineteenth century but by contrast a significant decline in French Canadian fertility, first observed in urban areas, did not occur until the start of the 1920s. Prior to this date, lower fertility rates could only be noticed among few subgroups of Quebec’s French Canadian population (Gauvreau et al., 2007). Consequently, the province of Quebec, which is the only Canadian province where the majority (roughly 80 %) of the population has French and Catholic origins, experienced a period of “over marital fertility” at the country scale in the late nineteenth century (Bouchard & Lalou, 1993; Gauvreau et al., 2007).

3Contrasts in fertility levels are also observed between Quebec’s urban areas. Quebec City’s fertility rates remained higher than those of Montreal, the largest city in the province, as well as those of smaller industrial cities like St. Hyacinthe and Sherbrooke (Marcoux, 2002; Gauvreau & Gossage, 2000). Eventhough a large number of people coming from surrounding countryside settled in Quebec city at the turn of the century, R. Marcoux and M. St-Hilaire (2003) point out that women born in rural areas did not have higher fertility rates than those born in the city. Modernization perspective cannot be considered ipso facto as an explanation in this case. These authors explore more deeply the relationship between family economy dynamics and reproductive behaviours among Quebec City three most important trades (carpenters, carters and shoemakers). They discuss the particular case of the shoemakers whose fertility rates are the highest in the context of footwear manufacturing in Quebec City, where cottage craft production, relying on the input of all members of the family, persisted well into the early twentieth century industrialization context (Marcoux et al., 2008). In a comparative study of Quebec City and Manchester, New Hampshire – home to a large French Canadian community at the turn of the twentieth century – M.-E. Harton (2017) shows how reproductive behaviors varied according to the structure of employment and occupation based on the dominant industries in each city, where in Quebec City, the largest socio-professional category is made up of skilled workers, compared to Manchester where French Canadians are mostly semi-skilled or unskilled workers. In the former city, married women rarely report paid employment, and children, boys and girls, work both informally and formally, while in the latter women and children experienced factory work in greater proportions. The uneven distribution of occupations of family members and of activities performed by each of them according to gender and intergenerational relations in both cities are thus associated with different reproductive outcomes but the diversity of such patterns also suggests that social networks played additionally an important role.

4This article takes a step forward in the study of the interconnections between modes of production and reproduction strategies. It focuses on the families of Quebec City leather workers, who had an above-average number of children despite the ongoing industrial transformation and global fertility transition, and provides a micro-scale analysis of family reproduction through the study of spatial distribution of differentials of fertility. Recognizing that factors behind reproductive behaviour are not strictly material, we adopt a more complex analytical model that highlights the gender and generational perspective on the households of Quebec City leather workers. The family wage economy was rooted in people’s experience of working and living together. Men, women, and children performed different activities within the family economy and in domestic life. In turn, these differentiated roles shaped perceptions of children, the care they needed, and their usefulness over the short, medium, and long terms. It reflected norms and ways of seeing that were based on the lived experiences of those individuals who built industrial cities during the nineteenth and twentieth centuries. This article sustains the idea that the diversity of reproductive patterns uncovered using the family-wage economy model suggests that kinship and social networks played an important role as well.

5Two theoretical frameworks are mobilised here. First, we draw upon the concept of “communication communities” developed by S. Szreter (1996, 2011). Communication communities comprise the social networks through which social and gender identities are acquired, reproduced and negotiated (Szreter, 1996). Within such communities, individuals share similar commonsense expectations of the roles that men, women and children would perform. They sustain reproductive strategies related to distinct gender relations and perceived relative costs and benefits of childrearing. As a result, the “communication communities” theoretical framework emphasizes the complexity of social structures, as well as the historical character of the main determinants of fertility, which are shaped by the temporal and geographic contexts within which demographic regimes operate as also suggested by S. Greenhalgh (1996) and D. I. Kertzer (1991).

6Second, we rely upon the idea that intergenerational solidarities between married adults and their parents, and therefore those between grandparents and grandchildren, were present in the past despite higher mortality rates and that the main transformation that has affected them over the past centuries relates to their meaning, as illustrated by Vincent Gourdon (2008). Especially, this author argues that the urbanization at the turn of the twentieth-century, far from dislocating kin networks, has favored them while transforming them. Thus, cohabitation has certainly declined but residential proximity, allowing for daily connections, was common in many urban areas, particularly in nineteenth and early twentieth century working class families. The study of kin propinquity must therefore enlighten family reproduction, and more specifically shed new light on fertility behaviours. According to the “helpful grandmother” hypothesis [1], the assistance that a grandmother provides to her daughter may help maximize a family’s reproductive potential. However, few sources allow us to explore these relationships from several angles. Following Bengston et al. (2012), kinship ties are based on six different forms of solidarity: affectional, consensual, functional, associational, normative, and structural. We here concentrate on three aspects of the structural dimension of these “married children and parents” relationships, namely their quantity, their types, and their spatial proximity.

7By a careful examination of the spatial distribution of leather workers’ households within Quebec City and kin propinquity among this group, we show differences relevant to fertility rates that might otherwise be hidden by cultural and economic similarities. This paper considers proximity and kinship networks as a vehicle for the social and gender identities that could shape reproductive behaviour. More specifically, it looks at how fertility behaviour varied across Quebec City and it describes the spatial concentration of specific subgroups according to various fertility rates. Using a dataset that links geocoded 1911 census microdata to Quebec marriage records, we undertake a household-level analysis of the social space inhabited by these families. Our microsociological approach therefore provides an ideal and direct vantage point on the gender and generational dynamics at play within the leather workers group through evidence of the neighbourhood interactions.

8The article draws upon a three-fold analysis of social space and family reproduction among Quebec City leather workers. After a short presentation of the context, data and methods, and after a brief overview of the structure of leather workers families, it begins with a micro-scale examination of the general spatial distribution of reproductive outcomes in Quebec City. Then, it examines the spatial concentration of specific leather workers’ subgroups based on fertility levels. Finally, it looks at the impact of kin propinquity, namely both maternal and paternal grandparents, on reproductive behaviour to explore more deeply social and family reproduction.

Quebec City at the turn of the twentieth century

9Founded in 1608, Quebec City was the province’s second-largest urban centre in the late nineteenth century. It lagged behind Montreal in terms of both population growth and industrial development (Hare et al., 1987). Between 1881 and 1911, Quebec City’s population grew 26 %, from 62,445 to 78,618. By contrast, Montreal, which had more than 500,000 inhabitants by 1911, saw its population increase by 257 % over the same period. But despite relatively modest population growth, Quebec City’s cultural composition changed dramatically as the nineteenth century drew to a close. M. St-Hilaire and R. Marcoux (2001) describe how the French Canadian share of the population grew thanks to the arrival of a large number of migrants from the rural areas surrounding the city. Meanwhile, other cultural groups, especially Irish Catholics and Anglo-Protestants, left Quebec City in large numbers following a major decline in the transportation sector. As a result, the percentage of French Canadians rose from just 56.6% in 1851 to 85.2 % in 1911.

10Quebec City’s economy has been shaped by its location on the north shore of the St. Lawrence River. In the mid-nineteenth century, it primarily served as a transit port, especially for timber exports, and as a shipbuilding centre. Over the course of the nineteenth century, the economy shifted from heavy industry (production assets) to light industry (consumer goods). The manufacturing sector, especially textiles and footwear, emerged as the largest employer in both Montreal (Burgess, 1981) and Quebec City (Courville, 2001).

11J. Burgess’s work on Montreal leather workers describes a slow transition from cottage craft production to industrial production (Burgess, 1986). Without minimizing the impact on workers, she suggests that this shift was quite complex and that it occurred in several stages. In Quebec City, artisanal and industrial production were able to coexist in both shoemaking and leather working, largely thanks to the implementation of the sweating system (low-paid piecework) (Thivierge, 1981b; Bluteau et al. 1980). Industrial rationalization and the commercialization of shoes gradually changed production processes (division of tasks, advent of machine tools, increased production, etc.). The number of people working in the sector, whether in factories or at home, increased considerably during the period. Meanwhile, the required labour became increasingly less skilled as production processes became more fragmented and retail markets continued to expand.

12In 1851, Quebec City had no factories. Footwear was produced entirely by hand. But during the second half of the nineteenth century, fewer and fewer leather workers were employed by small workshops, and a majority soon worked in factories with between 50 and 250 employees. The division of labour and the use of machines made these factories between 60 and 80 times more productive than traditional workshops. This larger production capacity helped meet demand from a retail market that was rapidly expanding at the turn of the twentieth century (Courville, 2001). By 1911, the footwear sector was the largest source of employment in Quebec City.

13Each shoe factory was divided into five or six departments (soles, assembly, lasting, etc.), which were connected by a fleet of small wagons. Each department was responsible for a series of steps in the manufacturing process, which involved over 150 separate operations in 1910. Tanneries were smaller operations than shoe factories. Some traditional tanneries remained open at the turn of the twentieth century and craft production persisted. Compared to footwear production, work in the tanneries was slower-paced, less specialized and less labour-intensive (Bluteau, 1981).

14Existing ethnographic research on leather workers in Quebec City (Dupont & Mathieu, 1981) stresses the slow pace of modernization, compared to developments in Montreal and Ontario. For instance, footwear producers in Quebec City were relatively slow to adopt the sewing machine. Shoemaking involved sewing together pieces of leather and this work could be performed manually by women or children at home, rather than by machine in a factory. The same held true for other forms of piecework, as well as a wide range of household tasks. Therefore, women and children constituted invaluable assets in leather workers families.

Data and methods

15Microdata [2] from the 1911 census make it possible to study household characteristics, residential patterns and micro-scale economic activities from a cross-sectional perspective. They provide a general overview of the population and its structure, as well as details on every member of each individual household. These details include a person’s age, sex, marital status, relationship to the head of household, place of birth [3] (of the individual and both parents), origin, religion, language spoken most often, and school attendance. In terms of information relevant to our case study of leather workers and their families, the census asked whether an individual had a profession or occupation and, if so, whether they were an employer, an employee, or self-employed. They were also asked for certain details about their workplace. All this information was collected for each individual in a household – including men, women and children.

16In addition to the full-count household-level microdata from the 1911 census compiled and geocoded by the “Population et histoire sociale de la ville de Québec” (PHSVQ) project, the “Integrated infrastructure of the Quebec population historical microdata” (IMPQ) project has linked census microdata with vital records [4]. Linked data provide a multi-generational perspective on family reproduction and kinship networks.

17Moreover, the ability to link census information to Quebec City’s built environment has allowed us to undertake an even more detailed analysis of the spatial distribution and reproductive behaviour of the leather-working community at the turn of the twentieth century [5]. In addition to precisely mapping the distribution of leather workers in the city, this makes it possible to examine spatial and socioeconomic patterns involving specific subgroups, based on the location of individual dwellings.

18To estimate fertility levels, we use an indirect measure: age-standardized own children estimates [6]. These estimates were calculated based on the number of children under the age of five belonging to women who were between the ages of 15 and 49 and who were married to a household head. This allows us to control for the age structure of the population under study. The total number of children in each maternal age category was then divided by the corresponding number of mothers. Since they only take surviving children into account, own-children estimates reflect reproductive outcomes or “effective” fertility.

Leather workers and their families in 1911

19In 1911, in Quebec City, 1633 men declared working in shoemaking, footwear manufacturing, or tanning. That represents 7.9% of all men over the age of 15 who were active in the workforce. This popular and well-established employment sector was also defined ethno-culturally: 97.5% of Quebec City leather workers were French Canadian [7]. In total, the city was home to 1592 French Canadian leather workers. Of these, 1245 worked in footwear production and 347 worked in tanneries. Two thirds of the French Canadian leather workers in Quebec City were household heads. Of these 1062 individuals, 842 (79%) were married to a woman between 15 and 49 years old. These women constitute our sample for the following analysis of reproductive behaviour.

20With an average size of 5.6 individuals, leather workers’ households were larger than other households in Quebec City, including those headed French Canadians in other lines of work. They also reflected a stronger influence of family and kinship ties, as well as important socio-economic differences (table 1).

Tabl. 1

Household structures, Quebec City, 1911. Based on regular family households only (< 30 individuals)

Household structures, Quebec City, 1911. Based on regular family households only (< 30 individuals)

21Nine out of ten French Canadian leather workers were employees, meaning that a large majority of leather workers were wage earners. Therefore, only a handful of individuals operated their own businesses. In 1912, the city’s five largest factories employed three out of four leather workers (Courville, 2001). Nevertheless, about 20% of leather workers who were employees worked either at home or in a small workshop [8]. By contrast, only 76.7% of the entire Quebec City workforce was made up of employees, 9.1% of whom worked at home or in a workshop.

22A careful examination of the occupations declared by leather workers’ wives and children also suggests that this economic sector relied on particular means of production that combined both the cottage craft tradition and industrial processes. Only 1.9% of leather workers’ wives reported being employed, compared to an average of 2.5% for married women in the city as a whole. Of course, this does not mean that women married to leather workers were not contributing to the family economy. Rather, their contributions were more likely to involve work that, from the perspective of the census, was considered “housework in [their] own home [done] without salary or wages” (Instructions to Officers, Commissioners and Enumerators, 1911: 32). The same was probably true of girls growing up in these households, since they were less likely to attend school and more likely to either report paid employment or no daily occupation. The latter strongly suggests that they were actively involved in domestic work that was tied to the family economy.

Neighbourhoods and socio-economic spatial distribution

23The same topographical features that attracted Quebec City’s founders to the site – the St. Lawrence River, the St. Charles River, and Cap Diamant – have continued to shape its development over the last four centuries. As its population has increased, the city steadily expanded westward (figure 1). At the turn of the twentieth century, it was divided in three main areas: the Upper Town (St. Jean Baptiste, Montcalm, Du Palais, and St. Louis neighbourhoods), the Lower Town (St. Pierre and Champlain neighbourhoods) and the St. Charles Valley (St. Vallier, St. Roch, St. Sauveur, and Jacques Cartier neighbourhoods).

Fig. 1

Neighbourhoods, Quebec City, 1911

Neighbourhoods, Quebec City, 1911

24Military, administrative and religious institutions dominated the Upper Town. Inside the city’s fortified walls, neighbourhoods extending down the eastern slopes of the promontory (St. Louis and Du Palais) were home to bourgeois and middle-class families, as well as more prominent residents. During the second half of the nineteenth century, limited space in Upper Town and population growth in Lower Town helped drive industry into what would become downtown Quebec City. And although shipbuilding soon began to decline, recently developed sections of St. Jean and St. Roch proved fertile ground for the city’s emerging footwear sector. Shoe factories were built alongside tanneries and workshops that had been in the area for several decades. In the years that followed, St. Roch also became the centre of the local textile industry and home to the city’s main railway station (Courville, 2001).

25The socio-economic structure of Quebec City remained highly stable from the onset of industrialization through its consolidation at the turn of the twentieth century. The late-nineteenth-century spatial differentiation of socio-economic groups described by N. Lanouette (2006) therefore remained largely unchanged in 1911. Figure 2 shows how skilled manual workers (including most leather workers) were highly over-represented in the St. Sauveur neighbourhood, as well as in St. Vallier, Stadacona, and Limoilou. In St. Roch and Jacques Cartier, manual workers were overrepresented in western areas and underrepresented in eastern areas.

Fig. 2

Location quotients for men aged 15-64 who reported being skilled manual workers, Quebec City, 1911

Location quotients for men aged 15-64 who reported being skilled manual workers, Quebec City, 1911

Location of leather workers’ households

26Leather workers’ households were concentrated in areas along the St. Charles River. This was especially true of the densely populated St. Sauveur neighbourhood, although large numbers of leather workers also lived in St. Vallier, St. Roch, and Jacques Cartier. Smaller concentrations could be found in the St. Jean neighbourhood (the only part of the Upper Town that was home to a significant number of leather workers) and along the north shore of the St. Charles River (figure 3).

Fig. 3

Spatial distribution of leather workers, Quebec City, 1911

Spatial distribution of leather workers, Quebec City, 1911

Note: This map shows leather workers’ dwellings but this study draws upon all leather worker’s household.s

27Our analysis of the structure and spatial distribution of industrial activity in Quebec City takes all production facilities – and not just factories with more than five employees – into account. Leather manufacturing facilities – including shoe factories, tanneries, and currieries – were concentrated in the Jacques Cartier neighbourhood, near the junction of St. Vallier, Langelier, and St. Joseph streets. The proximity of major railway lines and docks helped facilitate market access, while the large number of leather workers living nearby ensured the availability of an adequate labour force. The one out of five leather workers who reported working as employees from home or in a small workshop were more tightly concentrated in the area than those who reported working in another type of facility (such as a factory), or those who reported being an employer or self-employed (results not shown). These latter categories of leather workers were more likely to live in St. Sauveur or St. Vallier, neighbourhoods located in the western reaches of the area shown in figure 3.

28To sum up, Quebec City leather workers were almost all engaged in wage labour. They formed an homogenous group and their households were concentrated in one area of the city. It thus suggests that these neighbourhoods relied on close interactions and favored closely-knit community. What remains to be seen is whether the social space inhabited by leather workers and their families – defined in terms of kinship ties and geographic distribution of specific subgroups – was associated with fertility differentials.

Leather workers’ fertility differentials at the city scale

29Harton’s (2017) comparative study of French Canadian family reproduction gives for Quebec City higher own-children estimates for skilled workers than for other groups, including the semi or unskilled workers. She argues that this probably reflects the persistence of the family wage economy, deeply rooted in the leather-working trades, even in the midst of industrialization. Our estimates specifically calculated for leather workers point in this direction. Wives of workers in the footwear sector and those of tanners had an average of 1.16 and 1.20 young children, which is quite higher than the overall figure of 1.06 children under the age of five per married woman between the ages of 15 and 49.

30Figure 4 shows the geographic distribution of these own-children estimates. Each hexagon corresponds to an area with a radius of 75 metres. The highest fertility rates were in the Lower Town workingclass neighbourhoods of St. Vallier and St. Sauveur, as well as in Stadacona (on the opposite side of the St. Charles River) and in Limoilou (to the north-east). By contrast, neighbourhoods in the Upper Town had the lowest fertility rates. St. Roch, Jacques Cartier, and St. Jean were all somewhere in between.

Fig. 4

Own-children estimates by hexagon for leather workers’ families, Quebec City, 1911

Own-children estimates by hexagon for leather workers’ families, Quebec City, 1911

31These differences in fertility rates were probably more significant than the figures suggest, given the higher rates of infant mortality in the Lower Town (Gagné, 2004). Working-class families in St. Sauveur, St. Roch, and Jacques Cartier also experienced higher rates of infant mortality than residents of nearby St. Vallier. The situation likely remained the same in 1911, given the ongoing growth of working-class neighbourhoods marked by widespread economic insecurity and overcrowding. Leather workers and their families therefore faced generally poor living conditions (Thivierge, 1981b) that made it less likely for their children to survive infancy. This was particularly true of the conditions in the Jacques Cartier neighbourhood, which was crossed by the sewers leading from the Upper Town to the St. Charles River.

Communication communities and reproductive outcomes

32In their study of footwear manufacturing in 1901, Marcoux et al. (2008) demonstrate that wage-earning families had more children than those of employers and self-employed individuals working in the same sector. Furthermore, those whose head of household declared being an employee and worked at home or in a workshop had higher fertility rates than those of factory workers, likely reflecting reproductive strategies that placed a higher value on the work performed by women and children.

33So far, we have addressed three aspects of the spatial distribution of Quebec City leather workers in 1911. First, skilled manual workers were overrepresented in neighbourhoods located along the St. Charles River, especially further west. Second, there was a high concentration of leather workers in the same area, as well as in the St. Jean neighbourhood. Finally, effective fertility rates were higher in neighbourhoods with large numbers of leather workers, especially St. Sauveur and St. Vallier. We now turn our attention to whether the spatial distribution of leather workers’ households can be associated with various levels of reproductive outcomes. We hypothesize that the existence of communication communities connected to various reproductive strategies could be approached by the examination of spatial clustering of different groups. Did those Quebec City leather workers who have the greatest number of young children tend to live nearby?

34In this case, we use individually geo-referenced households as our unit of analysis, rather than hexagons representing aggregated data (see above). We also follow Wong’s (1999) suggestion of using standard deviational ellipses to illustrate and compare the spatial characteristics of population groups [9]. Such ellipses express the location and orientation of all data points in a distribution in relation to the mean centre. The ones presented below represent the first standard deviation from the central point of the distribution. The individual data points correspond to women between the ages of 15 and 49 who were married to the household head of a family active in the leather-working sector (figure 3). Each map contains four ellipses, which correspond to women with different numbers of children under the age of five.

35The ellipses in figure 5, top, show the ellipses for women aged 15 to 49. These overlap almost perfectly illustrating that the overall spatial distribution of leather workers’ families therefore did not vary significantly according to fertility rates. Nevertheless, a small difference can be seen between women who had three or more young children and those who had no children under the age of five. Although the two groups were dispersed over a similar area, their spatial distribution was centred at different points along the east-west axis. This is confirmed by the S index – based on the ratio of the intersection and union of two ellipses – which provides a statistical measure of spatial correlation (or differentiation). A value of 0 indicates perfect correlation and a value of 1 indicates perfect differentiation. The ellipse representing women who had three or more young children has a lower correlation with the one representing women who had no young children (0.151) than with the ones representing women who had one or two young children (0.020 and 0.000, respectively). At the same time, the ellipse representing women who had no young children has a lower correlation with the one representing women who had three or more young children (0.166) than with the ones representing women who had one or two young children (0.051 and 0.070, respectively; see appendix 1 for all calculations).

Fig. 5

Standard deviational ellipses for women married to a leather worker who was a household head, by number of children under five, Quebec City, 1911. Top: women aged 15-49. Middle: women aged 30-44. Bottom: women aged 35-39

Standard deviational ellipses for women married to a leather worker who was a household head, by number of children under five, Quebec City, 1911. Top: women aged 15-49. Middle: women aged 30-44. Bottom: women aged 35-39

36Could these differences in spatial distribution be related to age structure? The ellipses in figure 5, middle, are this time drawn exclusively for women between the ages of 30 to 44, that is to say those most likely either to have recently had a child or to have stopped childbearing. On the one hand, many women under the age of 30 would have been married for less than five years and therefore less likely to have higher numbers of young children. On the other hand, reproductive outcomes of women between the ages of 45 and 49 years might be reduced substantially for biological reasons.

37There is a slightly larger difference in the size of the two ellipses corresponding to the women who had three or more young children and those who had none (figure 5, middle) in comparison with the previous ellipses observed (figure 5, top). The S index calculations show that the ellipse representing women who had three or more young children has a higher correlation with the one representing women who had no young children (0.143 in figure 5, middle, compared to 0.151 in figure 5, top). Likewise, the latter ellipse has a higher correlation with the one representing women who had three or more (0.138 in figure 5, middle, compared to 0.166 in figure 5, top; see appendix 1 for all calculations). However, it is important to note women between the ages of 30 and 44 who had either one or two young children tended to be more spatially dispersed and cover a similar area, with women who had two young children somewhat more concentrated in the western part of the city.

38The ellipses in figure 5, bottom, are based on data for women between the ages of 35 and 39 and show clear differences in spatial distribution, potentially indicating the prevalence or absence of stopping behaviour. Such behaviour normally produces a distinct pattern of social and sometimes spatial distribution among women in their late thirties. On the one hand, women who had two children under the age of five were concentrated in a smaller area within St. Sauveur and St. Vallier, the two neighbourhoods that were home to the majority of home-based leather workers. On the other hand, women with one or no young children were more spatially dispersed. Although they were more likely to live in St. Sauveur, they were also spread across St. Roch and Jacques Cartier, placing the central point of the distribution in the manufacturing district. Granted, the distribution of women who had three or more children under five was also more dispersed. However, there are very few women in this group, and the fact that a few of them lived on the opposite side of the St. Charles River creates an ellipse that is strangely skewed toward the north-east.

39These results can be interpreted in various ways, since the corresponding cases likely reflected a range of circumstances related to reproductive behaviour. Some of these women may not have wanted any children, but ended up having some. Some may have wanted a large family but only managed to have one or two children. And some may have had a larger number of children, of whom only one or two survived. Granted, the distribution of women who had three or more children may also reflect some of these situations. But we would argue that, in the bulk of cases, these larger families may be partly due to the result of a different kind of reproductive strategy. Consequently, the fact that women displaying higher reproduction outcomes were concentrated in the west is consistent with previous results. St. Sauveur and St. Vallier had the highest overall effective fertility rates in the city and these neighbourhoods were also home to large numbers of leather workers who were heads of household and who worked from home or in a workshop. This would have made it easy for wives to assist with various tasks, especially sewing, and for daughters to lighten their mother’s load in assisting with food preparation or washing, for instance. Of course, factors related to the built environment, such as the presence of houses with attached workshops, may also have helped attract leather workers with larger families to these areas. J. Burgess (1986) shows the predominant role of kin and neighbourhood in building the Montreal closely-knit craft leather workers’ community in the first decades of the nineteenth century. Here, we suggest that the resulting geographic proximity may also have fostered social interactions related to the purchase of raw material and supplies, the production of goods and the selling of them that supported the creation of various communication communities. Our results therefore illustrate that large families were more closely tied to group belonging.

Kin propinquity and social reproduction

40As in many societies, kinship ties have played a key role in structuring social networks in Quebec (Fortin, 1987). Through the early twentieth century, French Canadian family connections were at the core of relationships established among co-workers (Bluteau, 1981; Hareven, 1982), neighbours (Olson & Thornton, 2011) and migrants (Mathieu, 1987; Frenette, 1998; Mimeault, 2013). In particular, such ties had been fundamental to the social networks of leather workers since the seventeenth century (Thivierge, 1981a; Burgess, 1986). Although tanning and shoemaking had both experienced major transformations, they continued to be passed down as family trades.

41Studies that explore the intergenerational transmission of fertility from mother to daughter have so far produced conflicting results (Brunet & Vézina, 2015; Gagnon & Heyer, 2001). The same holds true for research on the ways that grandmothers possibly support their daughters’ fertility or reduce infant mortality (Sear & Coall, 2011). But although the role of the paternal grandmother remains unclear, the presence of the maternal grandmother is generally associated with increased fertility and lower rates of infant mortality, leading to better reproductive outcomes (Mace & Sear, 2005; Voland & Beise, 2005).

42Since leather workers households relied on a complex family economy, we consider the presence of grandparents from both sides of the family, that is to say the parents of the leather workers who were household heads and those of their wives. Thanks to the data linking undertaken by the IMPQ, we are able to identify grandparents living in Quebec City at the time of the 1911 census, whether they lived or not within their daughter’s or within their daughter-in-law’s household.

43Of the 2132 household of leather worker heads and wives in our sample, 305 (14.3%) individuals had a least one parent living in the city. The proportions are almost identical for heads (14.8%) and their wives (14.0%). In both cases, they were more likely to have a mother (84 mothers of household heads and 81 mothers of wives) than a father (77 and 65) living in the city. Most mothers lived in a separate household, with only 19 (22.6%) living with a son active in the leather-working industry and only 25 (30.9%) living with a daughter married to a leather worker. In other words, a leather worker’s wife was more likely to share a home with her mother than with her mother-in-law. And even when the wife’s mother lived outside the household, she still tended to live nearby. On average, wives lived closer to their mothers than to their mothers-in-law, fathers-in-law, or fathers. Meanwhile, median distances and standard deviations show that while half of mothers-in-law lived very close to their daughters-in-law (within 147 metres), the other half tended to live farther away than the women’s mothers. Fathers, especially the fathers of leather workers’ wives, generally lived farther away than mothers (results not shown).

44The vast majority of wives of leather workers who had at least one parent or parent-in-law living nearby were under the age of 50 (table 2). We can therefore assess the impact of the presence of a maternal or paternal grandmother on reproductive outcomes. Maternal grandmother is defined as the leather worker wife’s mother and the paternal grandmother as the head’s mother. Our results show that families with a maternal grandmother living in the city had more young children (1.36) than those with a paternal grandmother (1.09) or no grandmother living in the city (1.17). This case study therefore sustains the “helpful grandmother” hypothesis on the maternal side. This is in line with previous studies showing that the presence of a maternal grandmother had a positive impact on child survival in the St. Lawrence Valley during the seventeenth and eighteenth centuries (Beise, 2005), as well as in Quebec’s Saguenay region during nineteenth and early twentieth centuries (Dillon et al., 2017).

Tabl. 2

Own-children estimates (ages 0-4) for women aged 15-49 and married to a leather worker who was a household head, Quebec City, 1911

Own-children estimates (ages 0-4) for women aged 15-49 and married to a leather worker who was a household head, Quebec City, 1911

45But living arrangements matter. Whether the grandmother cohabit, live nearby or at a greater distance affected their daily relationships and the general process of family reproduction. Grandmothers living closely could help with childcare and domestic chores and they could provide emotional support, for instance. But they could also be needing-care adults as aging individuals and there might have existed tense relationship between family members too. In this case study, the positive impact proves stronger in cases where each of the maternal or the paternal grandmother lived beyond the household than when she lived with the family. In both cases, her presence within the household is associated with lower own-children estimates than those for families with grandmothers living elsewhere in Quebec City (1.44 vs 1.23 and 1.06 vs 0.84).

46Did the distance between a grandmother’s home and that of her daughter or daughter-in-law affect fertility differentials among leather workers’ families? To find an answer, we divide both maternal and paternal grandmothers living outside the household into three groups of equal size, based on how closely they lived to their daughter or daughter-in-law. The correlation with own-children estimates is positive and linear on the maternal side. In other words, a shorter distance between a maternal grandmothers’ home and that of her daughter produces a higher own-children ratio. In the case of paternal grandmothers, the highest own-children estimates are associated with the middle group – those who lived neither especially close to nor especially far from their daughter-in-law’s home. These findings reflect the results of studies on the proximity and coresidence of maternal grandmothers in other historical contexts (Hacker & Roberts, 2017). This suggests that better reproductive outcomes depended on the availability of care and who cares who.

47As mentioned above, we are also interested in whether the transmission of reproductive strategies was tied to social reproduction. For instance, did specific contexts of socialization cause leather workers’ wives to adopt various reproductive strategies? Do intergenerational ties influence family reproduction? Were the children of leather workers more likely to have larger families? To begin with, our results reveal a very high rate of social reproduction rates among Quebec City leather workers. In the majority of cases, the fathers of leather workers (64 out of 75) and of leather workers’ wives (49 out of 59) were leather workers themselves. In fact, these figures may understate the degree of social reproduction, since some older fathers may not have reported an occupation to the census, even though they had been leather workers. In those few cases where fathers reported an occupation unrelated to the leather-working industry, the relationship to family size was different for leather workers who were household heads compared to their wives. Own-children estimates for the sons of leather workers are comparable to those of individuals whose fathers were not leather workers. By contrast, the daughters of leather workers had more young children than women whose fathers reported a different type of occupation.

48The results of our multivariate analysis provide an even stronger indication that reproductive behaviour was transmitted, at least partially, matrilineally among Quebec City leather workers (see appendix 2 for all calculations). Logistic regression analysis makes it possible to isolate the effect that each of the factors discussed above had on shaping leather workers’ families and on determining the likelihood of their having at least one child under the age of five in 1911 (compared to the likelihood of having no young children). The statistical models take into account the age of leather workers’ wives; the presence of a maternal or paternal grandmother within or beyond the household; the neighbourhood where they lived; whether leather workers or their wives had a father who was also a leather worker; and whether leather workers who were household heads worked from home, in a small workshop, or in a factory (analysis conducted only for those individuals who reported being wage-earning employees). Each model reflects the net effect of each of these factors – that is to say, they keep as a constant all of the other variables introduced in the model.

49Having a maternal grandmother living in Quebec City but not living within the same household increased the chances that a leather workers’ household included at least one child under the age of five in 1911. Furthermore, the effect was greater in cases where the maternal grandfather was a leather worker, suggesting that cultural models associated with the family economy and related reproductive strategies were transmitted through female socialization networks. We consider this hypothesis all the more plausible given that this phenomenon was not visible among paternal grandparents and that we take into account differences in neighbourhood (a proxy of various infant mortality rates) as well as the type of workplace (home-based or adjacent workshop, as opposed to other places such as factory) reported by wage-earning leather workers. It therefore appears that, in Quebec City’s leather-working community at the turn of the twentieth century, the closeness of the mother-daughter relationship played a key role in articulating modes of production and reproduction.

Conclusion

50This detailed spatial analysis shows that reproductive behaviour varied across Quebec City in 1911. First of all, there was a marked difference between own-children estimates for the Upper Town, where effective fertility rates were lower, and those for the St. Charles Valley, where fertility levels were higher. Moreover, married women living in the western neighbourhoods of St. Vallier and St. Sauveur tended to have more surviving children than their neighbours in St. Roch and Jacques Cartier, where more leather workers lived.

51Second, our results suggest that the spatial distribution of leather workers’ households was partly associated with differences in reproductive strategies. The phenomenon is most noticeable among families where the wife was between the ages of 35 and 39. Households with more young children were concentrated in the western part of the St. Charles Valley, suggesting that reproductive behaviour associated to larger families was more closely tied to group belonging than “Malthusian” reproductive behaviour. In other words, the absence of stopping behaviour was more prevalent than elsewhere in Quebec City.

52Third, kinship ties and kin propinquity influenced reproductive outcomes among leather workers. The presence of a maternal grandmother in the city, especially if she did not live with her daughter, had a significant positive effect on the number of young children in a household. Moreover, the closer a mother lived to her daughter (without sharing a household), the higher the latter woman’s own-children estimate. However, the presence of a paternal grandmother did not have this positive effect on fertility outcomes. Our most original finding relates to the influence of gender and intergenerational dynamics on social reproduction in the leather-working community. Specifically, the presence in the city of a maternal grandfather who was also a leatherworker further increased the positive impact of the maternal grandmother on the likelihood that a household would include young children.

53In these multiple ways, our micro-demographic study highlights complex interconnections between modes of production and reproduction strategies, uncovering distinct patterns of fertility behaviour in Quebec City during the industrial period. In keeping with J. Burgess’s (1981) interpretation of the slow pace of change in Montreal’s footwear industry, we argue that family wage economy and kinship ties, deeply rooted in the shoemaking and leatherworking trades, persisted through not only the introduction but also the consolidation of industrial capitalism. Our results suggest that spatially circumscribed communication communities sustained rational reproductive strategies that supported different modes of production and that were transmitted through kinship networks. These networks were based on gender and generational relationships and we assume that their propinquity in turn generated identities that reflected economic and material conditions as well as shared ideas and ideologies. Therefore, our views express that “communication communities” is a suitable theoretical framework for the study of Quebec City leather workers’ community at the turn of the twentieth century.

54These are only parts of the whole and we thus recognize the existence of other types of socialization networks that could also influence reproductive behaviors. For instance, the workplace could have been the focus of important networks for both men and women. Likewise, parish life and other kinship ties (such as those between siblings or cousins) could have influenced the spatial distribution of households at the local level. As S. Szreter writes (2011, 67), “the history of human reproduction is an extraordinarily complex and multi-faceted subject,” and this was an attempt to capture a small part of the big picture.