In 1584, fearing death, Eleanor Cumpayne, a single woman in the parish of Halesowen (Worcestershire), commended her soul to almighty God and her body to the earth, divided her sheep and money amongst her godchildren, relatives and friends, and placed particular trust in her cousin, Margery Cumpayne, who, amongst other duties, was to bestow the sum of 20s at her funeral. Unlike most testators, Eleanor Cumpayne survived this fearful illness, dying in the summer eight years later. Eleanor is also unusual because the detailed provisions establishing her as a single woman that were made 25 years before her death also survive. In his will her father, George, bequeathed to her for her natural life a chamber next to the fire in his house in Hill, a garden, an acre of land, 3s 4d annually, a cow called Fillpayle, a mare, bedding, the best pot and pan, six sheep which were to be kept with the rest of the family's sheep and also, at the death of her mother, a pasture. The experience of Eleanor Cumpayne, living semi-independently as a single woman with the assistance of her family, suggests that the idea that early modern society expected all women to marry needs closer examination. Such a view can indeed be found in prescriptive literature. The author of The law's resolutions of women's rights (1632) stated without hesitation that all women ‘are understood either married or to be married’. For him women were defined by their marital status, and marriage was the natural expectation for all women.

Contemporary authors based their ideas of the necessity of marriage for women on two main arguments: first, that marriage was natural and demanded by scripture, particularly as a consequence of the Fall, which imposed on women the twin obligations of childbearing and subjection to the authority of a husband; and second, that marriage was an economic necessity, a notion encapsulated in the well-known adage that ‘To thrive one must wive.’ The strength of these arguments in preventing remaining unmarried from seeming to be a viable and attractive option for early modern women is the subject of this article. In particular, I shall concentrate on the second argument, the economic dimension, since it was in this respect, as the experience of Eleanor Cumpayne suggests, that social attitudes did most to encourage the possibility of women remaining unmarried.

On November 11–14 1993, Indiana University hosted a conference on the ‘History of Registration of Causes of Death’, with funding from the US National Institute on Aging and the National Institute of Child Health and Human Development. The conference brought together historians of medicine and historically-oriented demographers and epidemiologists to discuss the origins of the recording of causes of death and the possible uses of these documents in demographic and epidemiological research. Demographers and epidemiologists would like to use long-run series of causes of death to examine the effects of social and economic conditions, the availability of health care, and specific risk factors on mortality. Many important questions (such as the effects of early health experiences on old-age morbidity and mortality) are best studied with data on changes over long periods of time. However, it is very difficult to construct a consistent series of deaths by cause over time because advances in medical theory and practice have led to significant changes in the classification of diseases. For example, it is unclear whether the prevalence of heart disease was increasing, decreasing, or constant before 1940, because heart disease was often classified under other categories.

The essays in this special number of Continuity and Change offer a range of insights on the historical circumstances in which cause-of-death registration emerged. They help us to see the ways in which medical theory, medical practitioners, and their increasingly influential professional organizations shaped the conceptualization of reporting of causes of death. Günter Risse's ‘Causes of death as a historical problem’ serves as an overview of the problems that social historians of medicine find underlying any continuous history of mortality experience. Above all, he argues, medical historians react as historians, wary of Whiggish confidence in state records without attention to the ideologies governing their creation.

During the night of 24 May 1524 a fire broke out in the city of Troyes in the French province of Champagne, a fire which was to rage for two days and nights, destroying up to a quarter of the town (including at least 300 houses) and seriously damaging three of the city's churches. A number of fatalities was also reported. In the aftermath of the fire, as an assessment of the damage was made, the city council undertook a review of municipal security and ordered the expulsion of non-natives of Troyes, the safe storage of inflammable materials and the provision of water outside each house. Emergency grain supplies were also organized, as many stores had been destroyed in what had already been a poor yield that year. The extent of the fire reveals the vulnerability of urban communities at a time when fire-fighting methods were still extremely primitive and fire insurance non-existent, but the sources also demonstrate the capacity of these communities to recover from such a setback. For fire was a common enough occurrence in cities in this period, an everyday hazard in narrow streets where houses were huddled together and the primary building material and source of fuel, for both domestic use and manufacturing, was wood. Indeed, there is now a substantial literature on urban fires of the sixteenth and seventeenth centuries, which concentrates largely on the physical damage inflicted, the sources of relief available to the victims and the precautions to be taken to prevent a recurrence of such misfortune.

PROBLEM NUMBER 1: THE SHIFTING ECOLOGY OF DISEASE

We all know that human disease is a constant and natural expression of conflicting forces and agents. As biological phenomena, diseases are never static. Influenced by physical and social agents, sickness shifts constantly in specific ecological settings. In infectious diseases, for example, constant readjustments within changing physical environments occur between pathogenic agents such as micro-organisms and viruses and their human hosts. As do other complex organisms, humans display a series of adaptive mechanisms which can both prevent or create illness; a lifetime of such encounters categorized as ‘wear and tear’ phenomena leads to distinctions between chronological and physiological age that are useful in assessing susceptibility to disease and life expectancy.

To understand the broad contours of human disease and periodic epidemiological changes in various parts of the globe – what physicians have called the ‘ebb and flow’ of disease – many historians therefore subscribe to the idea of a shifting ecology of disease. This dynamic concept presupposes complex interactions between both biological and non-biological factors which are ultimately responsible for different and changing patterns of sickness in time and space, exemplified, at least superficially, by McNeill's popular book Plagues and peoples. By highlighting the interconnectedness of various possible factors, including those which could ultimately cause death, the ecological model is quite useful to both historians and demographers.

There has recently been a tremendous new awareness of male violence against women, an awareness which spans the academic world, the services that come in contact with it, and popular initiatives. Edinburgh District Council's Zero Tolerance campaign (which arose out of a report on adolescents' knowledge about, and attitude to, domestic violence) began in 1992 and was a landmark in bringing this issue out into the open in Britain. Hundreds of articles have appeared in American medical journals in the last four or five years, and March 1995 saw the launch of a new interdisciplinary journal, entitled simply Violence Against Women. And of all the women who suffer violence, the greatest number are wives whose husbands abuse them.

In contemporary societies, status as a wage worker is a fundamental source of social and political identity. Wage labour is also the main source of income for individuals, conditioning – to a large extent – access to property, patterns of consumption, and receipt of public benefits. In Jane Lewis's words, despite an unequal wage structure between the sexes, the ‘best way of avoiding poverty risks, for both men and women, according to statistics, is being in the labour market’.

Labour markets have almost always had gender disparities, with women in a secondary position in terms of wage levels and promotion opportunities. In present-day Europe, public policies are the primary mechanism for counterbalancing women's secondary position in the labour market and also within the family. But this has not always been the case. Among the profound changes in the character of European states during the nineteenth century was their transformation from being one of the mechanisms reinforcing and organizing labour markets along gender lines to becoming an institution that has promoted equal-opportunity policies in the twentieth century and attempted to improve women's disadvantaged position.

In early-nineteenth-century medical literature, one finds an elegant symmetry between causes of disease and causes of death: both were sufficient causes of particular events. However, as I will argue, by the end of the century physicians no longer sought sufficient causes of individual disease episodes – instead almost all of medical research was organized around the quest for necessary causes that were shared by all the episodes of each particular disease. Such causes carried great practical and theoretical advantages: they enabled physicians to control and to explain disease phenomena.

One might wonder why there has been no parallel change in our thinking about causes of death; to this very day, causes of death are sufficient causes of particular events. In principle there is no apparent reason why we could not identify necessary causes for classes of deaths – indeed, we sometimes do so. But, in the case of death, such causes hold little interest. Because of how they are used, sufficient causes for individual deaths are more interesting and more important to us than are necessary causes of deaths. Thus, the change in thinking about causes of disease – the change that destroyed the symmetry between causes of disease and causes of death – may not reflect simply progress within a fixed system of medical goals and values, but a profound change in the social role of physicians.

As controversy concerning the future of Britain's railways continues to be voiced, their influence in shaping social structures in the past perhaps remains unexplained. Yet the nineteenth-century revolution in transport had a most profound effect on the physical conditions and mentalities of contemporaries. Landscapes were altered, old towns irrevocably changed and new ones created; topographical spellings and pronunciation were standardized and time itself was robbed of local variation; a new vocabulary was devised and new occupations were invented. In the construction of the railways a vast labour force, around 200,000 labourers at mid century, drew workmen from their traditional employment and often very far from their original homes. In the operation of the railways, new possibilities of travel, for business and pleasure, changed fundamentally the patterns and perceptions of mobility. Given such revolutionary developments, it is highly improbable that patterns of crime and criminality and the measures taken to counteract them would be unaffected. This article attempts to assess and to illustrate the extent and nature of the changes in these areas caused by the coming of the railway.

The lunatic asylum remains one of the most remarkable institutional monuments of the modern world, dominating the social landscape of Victorian Britain and exercising a powerful attraction for social historians of medicine, an attraction almost as great as the spectre of the madhouse for contemporary novelists. Our image of the Victorian asylum is still pervaded to a surprising degree by the gloomy spectacle of the total institution presented by Michel Foucault, though it has been modified by a whole range of institutional and philosophical accounts undertaken in the past three decades. Pioneering studies by researchers such as Andrew Scull have illuminated not only the power exercised by the new asylum superintendents, armed with medical discourses of moral treatment and the early promise of curability, but also the continuing dominance of the ‘mad doctors’ in the sombre years of neo-Darwinian pessimism and eugenics doctrines. More recent contributions to the now enormous literature on the social history of insanity have shifted the focus of attention from earlier concerns with charting the rise of the asylum and the elaboration of medical discourses under the psychiatric gaze of physicians to a detailed reconstruction of the social environment of the asylum and especially to the interplay between familial circumstances and the way institutions responded to the insane. Such concerns were also clearly evident in important earlier studies by Walton, Scull, Digby and others, which drew on fundamental work by Anderson on the changing role of the family during industrialization. These scholars drew attention to the importance of family and kinship relations in the negotiation of a lunatic's passage to the Victorian asylum, as well as the role of wider forces of economic change, population growth and migration in shaping the environment in which decisions about the care of the mad were made.

The origins of joint family households in the Balkans remain an issue of interest to many scholars of family history. This article focuses on the joint family household pattern recorded in the 1905 Ottoman census in two communities in Epirus. Related issues will also be discussed, such as the variations in household-formation behaviour within a community regarded as ‘ethnically’ homogeneous, the extent to which joint family households should be viewed as an intrinsic trait of the culture and economy of the Vlachs, and the problems that arise from relying exclusively on quantitative data for understanding household and family life.

The 1905 census, the last one undertaken by the Ottoman authorities, was intended to provide information on the Empire's ethnic composition, in a context of increasing unrest and tension caused by rising nationalist movements. The 1905 population listings of the two villages discussed in this article, the Vlach village of Syrrako and the Greek one of Aristi (or Artsista, the older version of the village's name, used before the trend to ‘Hellenicize’ Slav-sounding place names), are among the few such documents relating to Epirus to have been preserved. A relatively large body of literature on Epirus communities, produced by travellers, local historians, and geographers, furnishes additional information.

The present ageing of the population in western countries, and the related problems this development poses for contemporary social policy, provide especially compelling reasons for historians to think about old age in the past. A fundamental question is whether solitariness, or residential isolation, in this phase of the life-cycle is an inherent product of industrialization and urbanization or whether past societies confronted similar problems.

Despite the growth in studies of the history of the family over recent years – studies which have considered family structure, family demography, household strategies, kinship networks, and the like – a great gap in the literature remains regarding the role and status of family members in different phases of the life-cycle in Spain, particularly with respect to age and sex. Therefore, it is not surprising that there is an absence of information on old age and our elders' life-styles in the past, or on the extent to which the aged led lives of poverty and marginalization.

This article attempts to tackle some of the issues relating to the elderly in a particular place, Bilbao, and in a historical period, the nineteenth and early twentieth centuries.

In nineteenth-century Canada, a Census was conducted every ten years, beginning in 1851. Historians have tried to use the information in the local, provincial, and national Censuses to determine the rate of mortality and causes of death. The mortality statistics are thought to be reliable after 1921, but practical as well as conceptual problems interfere with attempts to establish retrospective mortality patterns using the nineteenth-century Census. In this paper, I will review the problems with this source on mortality. I will then compare the Census data with that derived from a study of a physician's daybooks, focusing on four specific areas of discrepancy and the possible reasons for the differences. Finally, I will suggest that the combined use of both types of sources may challenge some durable assumptions and could help to establish more reliable estimates of mortality patterns.

The progress of industrialization throughout the nineteenth century had profound effects on health and mortality. One relationship that contemporaries found particularly alarming was the potentially damaging effect that the employment of women in industry outside the home could have on their families' health. The fear expressed by public health officials was that women's employment would lead to neglect of their family duties and subsequently put their children's health at risk. Women were also regarded as more susceptible to industrial hazards, particularly to lead. Hence, women's work was seen to have important consequences for maternal and infant health and was considered incompatible with their traditional duties within the home. The introduction of legislation to restrict women's employment opportunities during the latter half of the nineteenth century was partly in recognition of this conflict between work and home. At the turn of the nineteenth century, the importance of good mothering was increasingly cited by public health reformers, mainly male and middle-class, alarmed at the persistent high level of infant mortality in urban areas. In industrial areas where there were opportunities for women's employment, the extent to which it contributed to high levels of mortality became a prominent public health issue associated with growing concern over standards of motherhood. In response to increasing agitation over the employment of mothers, the 1891 Factory Act prohibited employers from knowingly hiring women within a month of childbirth.

Even after the unification of Germany in 1871, political power continued to be fragmented. The Bismarckian constitution was superimposed on a collection of previously independent states: it acknowledged their continued existence as historical regions, granted them their own constitutions, state parliaments, and extensive legislative and executive powers. At a regional level, different perceptions of the appropriate role of the state continued to exist, with Prussian centralism contrasting with the laissez-faire amateurism of Bremen and Hamburg. The creation of centralized Reich (German Empire) agencies, such as the Imperial Health Office, failed to guarantee an effective implementation of German Empire decrees, and the administrative structure of individual states remained diverse. Indeed despite increasing pressure to create a uniform nation state and standardized administrative procedures, there were still 25 separate states in 1914. In relation to the structure of medical services and the formulation of medical policy, a significant degree of regional variation persisted into the twentieth century.

In a recent article Frederik Pedersen used the records of matrimonial litigation from the York consistory, the principal Church court of the province, during the fourteenth century to make a number of observations concerning the relationship of these records to the society from which they were generated. He argued that ‘the medieval court documents do not present a random sample illustrating trends in the surrounding society’ and that litigants tended to be disproportionately drawn on the one hand from the upper echelons of society and on the other hand from locations close to York itself. He has further suggested that the age structures of male and female deponents found within the surviving cause papers do not fit the same model life tables and that this raises ‘further doubts about the representativeness of court documents as evidence of changing patterns of lay behaviour’. In this article, I shall show that his analysis is based upon a flawed methodology, is marred by error, and is ultimately mistaken. Pedersen's essential point in ‘Demography in the archives’, as outlined in the Abstract to the article, is that the people who appear in the court records are unrepresentative of society as a whole, and hence ‘that the court records tell us more about the people who used the courts than about trends in the society in which litigation arose’. My argument is that it is the very unrepresentativeness of the people and their cases that provides us with a window into the society from which the cases arise. I shall suggest ways in which the York cause paper evidence can indeed be used to illuminate broader social trends, but also suggest caveats as to the reading of individual causes.

I. THE CREATION OF MODERN DEMOGRAPHIC STATISTICS

The existence of a specialized administrative organization depends on the political, economic, and social development of a country, and is indispensable for the availability of modern statistics. As more than one author has pointed out, quantitative assessments of its population reflect a government's capacity to rule. The first Spanish attempts to obtain such quantitative knowledge are found in the so-called Relaciones ordered by Philip II in the second half of the sixteenth century; they described both European and American territories of the Crown. These early attempts, however, lacked continuity.

When I wrote ‘Demography in the archives’, it was my hope to open up a methodological discussion of geographical, social, and demographic aspects of legal sources that I believed had not received sufficient attention in the past. The matter was urgent, I felt, because I had found that there was a distinctive pattern to the reception of my work on either side of the Atlantic. When I presented papers in North America, my basic premise – that it was difficult, if not impossible, to use legal sources like the York cause papers as the basis of an analysis of social change in the century after the Black Death – was accepted with little or no comment. However, when I presented the same papers in Britain there seemed to be unease about this conclusion and one of the most frequent questions I was asked was why I did not analyse temporal variations in the material. From this I perceived the need to explain how I arrived at such a position, and after a long period of gestation the eventual outcome was ‘Demography in the archives’. I am delighted that Dr Goldberg has taken the time to go over my arguments in this meticulous fashion (see ‘Fiction in the archives: the York cause papers is a source for later medieval social history’, above in this number). He has found many errors that escaped me during the article's proofreading, and for this work I thank him. In what follows I shall concentrate on what I see as the main points of his article.

It is clear that we differ profoundly in our estimation of the value of these sources to the social historian. However, there is some common ground between us. We agree that an investigation of the categories suggested in ‘Demography in the archives’ will yield important insights into the nature of the medieval litigation material. We are, in other words, agreed that factors such as litigants' social status, their geographical distance to the courts, and the age-composition of witnesses must be examined before we can conclude anything about the usefulness of these sources. Where we differ is in our interpretation of the evidence and in the tools that we suggest are most suited for this analysis.

Barbara Evans Clements, Barbara Alpern Engel and Christine D. Worobec eds., Russia's women: accommodation, resistance, transformation. (Berkeley and Los Angeles, CA, and Oxford, England: University of California Press, 1991.) Pages xi+300. £35.00 (hardback); £13.50 (paperback). (hereafter RW)

Beatrice Farnsworth and Lynne Viola eds., Russian peasant women. (New York and Oxford: Oxford University Press, 1992.) Pages 304. £13.50. (hereafter RPW)

Barbara Alpern Engel, Between the fields and the city: women, work and family in Russia 1861–1914 (Cambridge and New York: Cambridge University Press, 1994.) Pages xi+254. £35.00. (hereafter Between)

In 1995 a Russian scholar lamented that ‘Women's history as an autonomous sub-discipline of historical science does not yet exist in Russia’, and that it had been suppressed in the Soviet Union. In contrast, western historians have produced a great deal of work on Russian women over the last 20–30 years. Much of the early work concentrated on women radicals and the feminist movement. More recently, historians have turned their attention to the mass of the female population who lived in rural Russia. The books under review here are very good examples of both the range and the quality of the work on Russian women and of the new focus on peasant women.