1 See the recent influential pamphlet by Michael, Anderson, Approaches to the history of the Western family, (1980).Google Scholar

2 Wall, , ‘Woman alone in English society’, Annales de démographic historique (1981).CrossRefGoogle Scholar

3 This categorisation scheme, first outlined by Laslett, in Household and family in past time (1972), 31Google Scholar, takes account only of the kin structure of the household. The basic unit is that of the simple family (married couple alone or married couple or lone parent with unmarried children). When other relatives are also present not forming a family in their own right, the household is termed ‘extended’ or multiple if two or more related families are present. The scheme also allows for two types of no family household: solitaries and related and/or unrelated persons living together not forming families of their own.

4 See my ‘Work welfare and the family: an illustration of the adaptive family economy‘, in Bonfield, Smith and Wrightson, , The world we have gained (1986).Google Scholar

5 The fullest account of life-cycle service is to be found in Laslett, Family life and illicit love in earlier generations (1977), 29–65.Google Scholar For the term ‘life-cycle servant’ see ibid. 34.

6 Hajnal,‘Two kinds of household formation system’, in Wall, Robin and Laslett, , Family forms in historic Europe (1983), esp. 68–71, 85–8.Google Scholar

7 Although see my‘The age at leaving home’, Journal of Family History 3 (2) (1978).Google Scholar This relies on a different methodology from that applied in the present study in that calculations were made of the number of children not living with a parent who still had at least one parent alive. The exit rate is naturally higher than that.

8 For rejection by a daughter of her right to inherit see Le, Play, Ouvriers des deux mondes (1857). The case comes from the lower Pyrenees in the mid-nineteenth century.Google Scholar

9 A guide to lists of inhabitants, copies of which are held by the Cambridge Group, is being published in the Journal, Local Population Studies. In 1980, there were 568 lists in the collection, if enumerations from the censuses of 1801, 1811, 1821 and 1831 are included, and since 1980 a further 96 lists have been added. Of the 568 lists, 36 per cent named individuals and 24 per cent (not always the same lists) gave details on relatives.

10 The problems of identifying the households in past populations and the variety of their functions are explored in greater depth in my introduction to Wall, Robin and Laslett, , Family forms in historic Europe (1983), 7–13.Google Scholar

11 Enumeration Abstract 1841 England and Wales, 1, 4; Enumeration England and Wales 1801 lists the questions on a preliminary page prior to summarising the population of Great Britain. The emphasis in the text is my own.

12 On the increase in the number of relatives in the household, see Laslett, , Family life and illicit love (1977), 24Google Scholar; Anderson, ‘Household structure and the industrial revolution’, in Laslett, and Wall, , Household and family in past time (1972), 220.Google Scholar Anderson also comments on the increase in the number of lodgers, 220, 225, but note Laslett's revised estimate of the number of lodgers in pre-industrial England, Household and family in past time, 134.Google Scholar

13 To avoid possible confusion over the term ‘child’, children and step-children of the head have been termed ‘offspring’ in the tables and figures in this paper.

14 There is always the possibility that some servants were the (unacknowledged) relatives of the head. References to the employment of relatives as servants occur in the seventeenth-century diaries of Pepys and Josselin. Where servants can be identified as relatives, for example through a shared surname, they have been counted as relatives and not as servants in the following tables. The correction, however, can only be a minimal one, see Laslett, , Household and family in past time (1972), 57.Google Scholar For Josselin, see Macfarlane, , The family life of Ralph Josselin (1970), 148.Google Scholar

15 These arguments are set out at greater length in my ‘The household: demographic and economic change in England, 1650–1970’, in Wall, Robin and Laslett, , Family forms in historic Europe (1983).Google Scholar For the demography of pre-industrial England, see Wrigley, and Schofield, , The population history of England (1981).Google Scholar

16 A considerable body of literature now exists on what scholars intend (and intended) by the use of the term ‘proto-industrial’. As used in the present paper, it betokens home or small workshop based manufacture of goods for a non-local market.

17 The research of Valerie Fildes (see Breasts, bottles and babies (1986) and ‘The English wet nurse and her role in infant care, 1538–1800’, Medical History (forthcoming)) and Gillian Clark (forthcoming in Local Population Studies) has revealed that nurse children were sent from London into a necklace of surrounding counties and at times even further afield where they might stay with the same nurse for varying periods, up to several years in some cases. Unfortunately, the central question of how many parents actually put their children out to nurse is still unanswered as the children are only recorded on burial or on rare cases, as at Baling, with their nurses. An argument could perhaps be put forward for excluding both nurse children and schoolboys from the resident population as outsiders with a permanent residence elsewhere but, on the other hand, there is no way of ascertaining now to what extent the resident population ought to be augmented for temporary absentees.Google Scholar

18 Smith, R. M., ‘Population and its geography in England, 1300–1730’, in Butlin, R. and Dodgshon, R., An historical geography of England and Wales (1978), 217.Google Scholar

19 Wrigley, and Schofield, , The population history of England (1981), 333–5, 528.Google Scholar

20 It is customary to see the husbandman as intermediate in status between the yeoman and the labourer. Laslett, , The world we have lost, 3rd edn (1984), 39, 44–5.Google Scholar The precise meaning of the term in a community like Ealing where there was only one individual identified as a labourer or at Ardleigh where there was none is another matter, and amongst the husbandmen, particularly at Ardleigh, there may have been a number, even a majority of individuals who might have been termed labourers.

21 The payments are documented in a remarkable book by F. F. Erith, published in 1978, Ardleigh in 1796: its farms, families and local government. A thorough analysis of the relationship between household forms and poor-law policy has been carried out by Thomas Sokoll, a research student at the Cambridge Group.

22 See my ‘Work welfare and the family: an illustration of the adaptive economy’ in Bonfield, Smith and Wrightson, , The world we have gained (1986).Google Scholar See also the profile of the life cycle in rural West Flanders in my ‘Does owning real property influence the form of the household?’, in Wall, Robin and Laslett, , Family forms in historic Europe (1983), 393.Google Scholar

23 The household of another clergyman containing 7 females and 37 males, 36 of them gentlemen boarders, has been included in the total population but is excluded from all further calculations here as no ages were specified.

24 This at least would appear to be the position of Medick as enumerated in his now classic article, ‘The proto-industrial family economy’, Social History 3 (1) (1976).Google Scholar

25 The households of the two clergyman employed at most 20 of Winwick's total of 55 servants.

26 The Winwick list, too, fails the strict criteria set out for inclusion in Table 1 in that spouses and offspring are not specifically identified.

27 Persons without a specified relationship to the head and sharing the surname of the head of the household have been classed as spouse if female, listed second in the household, and within 20 years of the head's age, otherwise as offspring. If the surname is different and they were not styled servant, it has effectively been assumed that they were unrelated to anyone in the household and they appear in the category, ‘others’.

28 Offspring are under-estimated through the inability to identify step-children. Relatives of ever-married women cannot be identified as their maiden names are never recorded.

29 This result is derived from pooling all the reconstitution studies available to the Cambridge Group 1982. See Wrigley, and Schofield, , ‘English population history from family reconstitution’, Population Studies 37 (1983), 166.Google ScholarPubMed

30 My thanks are due to David Wall for the information on the age gap between spouses in Chilvers Colon in 1781.

31 A much larger number of enumerations are available for the first three decades of the nineteenth century. A few only have been included in Tables 3–6 for the purposes of comparison. A broader definition of offspring was adopted for these tables than for Figure 1, in that all unmarried children resident with at least one parent are counted as offspring whereas in Figure 1 offspring include only the children of the household head.

32 Macfarlane would appear to have been the first to have used this technique on a historical population. Macfarlane, , The family life of Ralph Josselin (1979), 209.Google Scholar

33 A study of the number of children living apart from their parents was completed by me in 1978, in part based on the rare lists that specified the number of children ever born and surviving to the date of the list, in part based on linking families in the census to entries for these families in the parish register to identify absent children. ‘The age at leaving home’, Journal of Family History 3 (2) (1978).Google Scholar

34 An estimate suggested by Wrigley, and Schofield, , ‘English population history from family reconstitution’, Population Studies 37 (2) (1983), 179.Google ScholarPubMed

35 Anderson, , Family structure in nineteenth-century Lancashire (1971), 126.Google Scholar

36 Smith, R. M., ‘Population and its geography in England 1500–1730’, in Butlin, R. and Dodgshon, R., An historical geography of England and Wales (1978), 210–11.Google Scholar

37 It has been suggested to me by Tom Doig, Curator of the Cambridge Folk Museum, who is involved in an intensive study of the population of Barkway, that the exodus is to be explained by the education local girls received, which made them much in demand as domestic servants. At Stoke Poges it is clear that the younger absent girls were indeed employed as domestic servants as the list gives details of absent children. Whether these opportunities in domestic service could have caused such a marked exodus of daughters is less certain, as the demand for domestic servants must have been a fairly general one.

38 Kussmaul, , Servants in husbandry (1981), 120 f.Google Scholar, and on servants in general see my ‘Regional and temporal variations in English household structure from 1650’, in Hobcraft, and Rees, , Regional demographic development (1979), 97–8.Google Scholar

39 The decision to exclude them was taken because it was felt that the offspring of widows might well be concentrated in the older age groups and thereby perhaps conceal that, other factors being equal, relatively more sons than daughters would be present in the older age groups.

40 Enumeration Abstract, 1 (1831), viii.Google Scholar