1 An incomplete list of historical studies of age at leaving home would include Wall, R., ‘The age at leaving home’, Journal of Family History, 3 (1978)CrossRefGoogle Scholar, and ‘Leaving home and the process of household formation in pre-industrial England’, Continuity and Change, 2 (1987)Google Scholar, and Schofield, R., ‘Age-specific mobility in an eighteenth century rural English parish’, Annales de Démographie Historique (1970)Google Scholar, which describe leaving home in England; and Modell, J., Furstenberg, F. F., and Hershberg, T., ‘Social change and transitions to adulthood in historical perspectives’, Journal of Family History 1 (1975)Google Scholar, which compares the transition to adulthood in the United States in the nineteenth and twentieth centuries. For an economic perspective on age at leaving home with a somewhat different focus from the one offered here, see Galenson, D., ‘Economic determinants of the age at leaving home: evidence from the lives of nineteenth-century New England manufacturers’, Social Science History 11, 4 (1987).CrossRefGoogle Scholar

2 Hajnal, J., ‘European marriage patterns in perspective’, in Glass, D. V. and Eversley, D. E. C. eds., Population in history (London, 1965)Google Scholar, and Hajnal, J., ‘Two kinds of preindustrial household formation system’, Population and Development Review 8 1982.CrossRefGoogle Scholar

3 Major tillage crops are wheat, barley, oats and potatoes; computed from Mitchell, B. R., European historical statistics 1750–1975, 2nd edn. (New York, 1980)Google Scholar, series D. Figures on holdings from Crotty, R., Irish agricultural production: its volume and structure (Cork, 1966), 351.Google Scholar An index of the number of male agriculturalists per thousand occupied males falls from 627 to 572 between 1901 and 1911 for Ireland as a whole, but increases for the province of Connaught and many rural counties in other provinces: Fitzpatrick, D., ‘The disappearance of the Irish agricultural laborer, 1841–1912,’ Irish Economic and Social History VII (1980), table I.Google Scholar

4 Fitzpatrick's index of the number of labourers per 100 farmers falls from 180 to 146 between 1861 and 1901 (Fitzpatrick, ‘Disappearance’, table n).

5 Population and household figures from the decennial censuses of Ireland. Teitelbaum's study of fertility patterns in Ireland suggest that the fertility transition did not begin before 1921 (Teitelbaum, M. S., Fertility decline in Britain (Princeton, 1984).Google Scholar This finding rests on an uncritical acceptance of birth-registration data. More recent studies show some fertility control after the 1880s; see especially David, P. A. and Sanderson, W. C., ‘Measuring marital fertility control with CPA’, Population Index 54 (1988)CrossRefGoogle ScholarPubMed; and O'Grada, C. and Duffy, N., ‘Fertility control in Ireland and Scotland c. 1880–1930: some new findings’ (Working Paper, University College, Dublin, 1989).Google Scholar

6 Micks, W. L., An account of the constitution, administration, and dissolution of the Congested Districts Board for Ireland: from 1891 to 1923 (Dublin, 1925), 78.Google Scholar

7 Fitzpatrick, D., ‘A share of the honeycomb: education, emigration and Irishwomen,’ Continuity and Change I (1986).Google Scholar

8 These comments are not intended as a general history of Irish emigration. For an excellent overview see Fitzpatrick, D., Irish emigration 1801–1921, Studies in Irish Economic and Social History No. 1 (Dublin, 1984).Google Scholar

9 Calculated from Commission on Emigration and Other Population Problems, Reports (Dublin, 1954), table 95.Google Scholar

10 O'Grada's careful review of Irish emigration statistics shows that they seriously undercount emigrants from Connaught as well as emigrants to Britain, especially at mid-century (O'Grada, C., ‘A note on 19th-century Irish emigration statistics’, Population Studies 29 (1975).CrossRefGoogle Scholar Employing official statistics creates the false impression that emigration from Connaught increased during the last four decades of the nineteenth century.

11 Emigration from Europe was especially heavy during the 1880s. For the period 1881–1885, net emigration from Germany was 4.3 per thousand; gross emigration from Ireland ranged from 12.6 to 21.6 per thousand in the years 1881–1885 (Burgdörfer, F., ‘Migration across the frontiers of Germany’, in Willcox, W. ed., International migrations, vol. II (New York, 1931), table 120Google Scholar; W. E. Vaughan and A. J. Fitzpatrick, Irish historical statistics, table 54.)

12 ‘Overseas’ emigrants exclude those to Great Britain; figures from Commission on Emigration and other Population Problems, table 124.

13 Discussion of rural Irish households has centred on the existence or prevalence of a stem-family household system. Arensberg and Kimball asserted the prevalence of stem-family households in County Clare in the 1930s, and argued that such households reflected long practice. The controversy reflects the debate set off originally by Peter Laslett's demonstration that small, nuclear-family households were the norm at least in England in the early-modern and modern period. See Laslett's introduction to Laslett, P. and Wall, R. eds., Household and family in past time (London, 1972)CrossRefGoogle Scholar; Arensberg, C. and Kimball, S., Family and community in Ireland (Cambridge, MA, 1968)CrossRefGoogle Scholar; Gibbon, P. and Curtin, C., ‘The stem family in Ireland’, Comparative Studies in Society and History, XX (1978)Google Scholar; and Fitzpatrick, D., ‘Irish farming families before the First World War,’ Comparative Studies in Society and History, xxv (1983).Google ScholarGuinnane, T., ‘Intergenerational bargains, emigration, and the rural Irish household system’ (unpublished working paper Princeton University, 1989) argues that discussions of the stem-family conflate the distinct issues of co-residence and succession.Google Scholar

14 In addition to the references cited in the previous footnote, historical works include Connell, K. H., ‘Marriage in Ireland after the Famine: the diffusion of the match’, Journal of the Statistical and Social Inquiry Society of Ireland (1955–1956)Google Scholar, and ‘Peasant marriage in Ireland: its structure and development since the Famine’, Economic History Review, 2nd series, xiv (1962)Google Scholar; Fitzpatrick, D., ‘Marriage in post-Famine Ireland’, in Cosgrove, A. ed., Marriage in Ireland (Dublin, 1985).Google Scholar These works reply on the archives of the Irish Folklore Commission, held at University College, Dublin; Fitzpatrick supplements this record with additional accounts, such as newspapers.

15 The extent of agreements for parental support was recognized by the Local Government Board in 1910, when it ruled that parents who turned their farms over to a son had not intentionally deprived themselves of means, and so were not made ineligible for the Old Age Pension (Local Government Board (Ireland), ‘Annual report for the year ending 31 March 1909’, House of Commons Sessional Papers, 1910, vol. 54, xiii.Google Scholar) In cases where the heir's father died before transferring the farm, the deceased man's will typically stipulated support for the widow as a condition for taking the farm. For Altenteil see Sieder, R., Sozialgeschichte der Familie (Frankfurt an Main, 1987), 65–72.Google Scholar

16 The argument mentioned in the text is hinted at by Fitzpatrick, , ‘Marriage in post-Famine Ireland,’ pp. 120–3Google Scholar, and is brought up very frequently by Irish scholars. The inter-censal marriage of an heir is a very rare event. Figures refer to children of the 1901 household head only. T. Guinnane, ‘Intergenerational bargains’, compares other implications of the folklore account to the patterns found in the manuscript census sample.

17 Folklore sources mention other practices, including helping to earn a dowry by working as a servant in Ireland or Great Britain.

18 Computed from the 1901 and 1911 Census of Ireland as reported in Vaughan, W. E. and Fitzpatrick, A. J. eds., Irish historical statistics: population, 1821–1971 (Dublin, 1978), table 27.Google Scholar The comparable figure for males is 19 per cent. It is not possible to estimate the percentage of women in the manuscript sample who leave home to marry between 1901 and 1911.

19 There was also substantial migration to Belfast; the counties reported in Table 3, as well as the focus on Dublin, are intended to supplement analysis below. The published census does not report birth-place data cross-tabulated by age, and there are no other published quantitative sources on internal migration, so more precision on the role of internal migration in the life-cycle of young Irish people awaits exploitation of the manuscript census schedules for Dublin and other receiving areas. David Fitzpatrick suggested calculating the ratio of internal migrants to numbers emigrated from that county. This ingenious measure may mix cohorts between numerator and denominator; we could, in principle, recover the ages of the outmigrants, but we do not know whether the internal migrants were younger or older. Recent work suggests important changes in female employment patterns in rural areas during the last part of the nineteenth century. See Bourke, J., ‘Women and poultry in Ireland, 1891–1914’, Irish Historical Studies 99 (1987))Google Scholar, and ‘Husbandry to housewifery: Ireland 1890–1914’ (unpublished Ph.D. thesis, Australian National University, 1989).Google Scholar

20 Ages computed from the information of marital duration in the 1911 census. There are 641 wives and 642 husbands with usable data. The sample is documented below.

21 See, among the numerous studies of the Irish-born in the United States, Diner, H., Erin's daughter in America (Baltimore, 1983).Google Scholar

22 Some 30–40 per cent of males and females are servants at ages 15–19 in the Icelandic, Norwegian, and Flemish communities Hajnal cites. Over half of all 15 to 19-year-olds are servants in some rural Danish parishes. J. Hajnal, ‘Two kinds’, tables 13 and 14.

23 Breen, R., ‘Farm servanthood in Ireland, 1900–40,’ Economic History Review, 2nd series, xxxvi (1983).Google Scholar

24 The Republic of Ireland does not enforce the hundred-year seal on manuscript census schedules. The census schedules are located in the Public Record Office, the Four Courts, Dublin. The sample includes the District Electoral Divisions (DEDs) of Scarriff (County Clare); Drummin, Bundorragha, Erriff, Owennadornaun, and Kilgeever (County Mayo); Newtown, Staholmog, and Cruicetown (County Meath); and Coolattin, Coolboy, and Shillelagh (County Wicklow). See Guinnane, T., ‘Migration, marriage, and household formation: the Irish at the turn of the century’ (unpublished Ph.D. dissertation, Department of Economics, Stanford University, 1987) for more extensive description.Google Scholar

25 The most obvious difference between the sample and the Irish populace is in religion; lacking any Ulster locale, the sample is 92 per cent Roman Catholic, compared to 74 per cent of the Irish populace in 1911. Most non-Catholics in the sample are found in the Wicklow subsample, where 23 per cent are non-Catholics. Guinnane, ‘Migration, marriage, and household-formation’, table 2; Census of Ireland, 1911, General Report, table 121. To forestall confusion it should be noted that the manuscript census sample is not a random sample of individuals from all of Ireland, nor is it a random cluster sample (the locales were not chosen on the basis of any randomization process). Matching the 1901 and 1911 census schedules, and the entries in the valuation books, would pose formidable problems for an all-Ireland random sample of households or individuals, particularly given the frequent occurrence of many surnames in rural Ireland.

26 Where a son takes over his parents' household between 1901 and 1911 this is quite obviously the same household. Less clear are two instances in which a son-in-law or nephew not present in 1901 takes over, retaining an elderly widow present in 1901.

27 The Revision Books are located in the Valuation Office, Dublin.

28 See Hajnal, J., ‘Age at marriage and proportions marrying’, in Population Studies VII (1953).Google Scholar

29 For example, Modell, J., Furstenberg, F., and Hershberg, T., ‘Social change and transitions to adulthood in historical perspective’, Journal of Family History 1 (1975).Google Scholar

30 Others have noted that imbalances between in- and out-migration rates can affect SMAM and similar measures. (Hajnal refers to this problem in his original article. See ‘Age at marriage and proportions marrying’, 112.Google Scholar) The last problem discussed in the text is similar, but unrelated to migration per se. If we are interested in age at leaving home defined for subgroups for which information on the adults gives no clue as to their childhood origins, then current-status methods risk inappropriate comparison of those who have left home to those who have not. That is to say, if some farmer's sons become shopkeepers, then information from one census will yield an undercount of the number of adult males who were originally farmer's sons.

31 Census of the Irish Free State, 1926, volume v, part 1, table 22 and 23. Little is known about Irish mortality in the late-nineteenth and early-twentieth century, in part because the death registration system was seriously defective in many areas. Life-tables calculated from death-registration data without any corrections are highly questionable.

32 For example, if boys disappear earlier in life from labourers' households than from prosperous farmers' households, it may be due to higher child mortality in poor households, or to a lower age at leaving home.

33 The relatively few visitors in these households implies that few are missing from these households, unless for some reason the households in the sample were relatively undesirable places to visit.

34 That is, unobserved heterogeneity in any waiting-time process, such as leaving home, can induce a spurious, negative relationship between the duration until the event (here, the young person's age) and the probability of occurrence for that event (leaving home). Guinnane, T., ‘Household, the life-cycle, and migration: leaving home in rural Ireland, 1901–1911’ (unpublished working paper, Princeton University)Google Scholar, develops this issue in detail and estimates an econometric model of age at leaving home using this same manuscript census sample.

35 The occupation groupings pertain to the stated occupation of the principal earner in the household. In only a very few households are there individuals who report distinctly different occupations; in most cases the father (or widowed mother) is a ‘farmer’, and children call themselves ‘farmer's sons’, etc. The farmer class in Table 4 includes graziers and the labourer class includes herds. The residual category ‘Other’ is comprised of artisans, shopkeepers, and government and estate employees. This classification scheme may obscure differences between various occupation classes, but the sample is too small to permit less aggregation. Small, medium, and large farmers are those whose land and buildings are rated at less than £4, £4-£30, and more than £30, respectively.

36 The census may suffer from some undercounting of children, and this undercounting may be more pronounced for females than for males. Information presented later on the behaviour of older children does not support the view that the differences reported in Table 4 are simply an artefact of differential undercounting by sex.

37 The econometric models reported in the companion paper ‘control’ for the effect of locale, and suggest a counter-caution: some of the apparent effects of locales in Table 5 stem from differences in the composition of each locale. That is, Mayo differs from Meath both because a larger proportion of household heads in Mayo are fanners, and because of purely local effects apparently unrelated to occupational structure.

38 The term ‘head’ as used in Table 6 includes ‘co-heads,’ or wives. None of the female heads aged 25–29 in Table 6 are never-married.

39 Fitzpatrick argues that the term ‘visitor’ is best interpreted as ‘resident relative’, Fitzpatrick, , ‘Irish farming families’, 358.Google Scholar Visitors are here classified as ‘not related’.

40 The comparable figure for women is 3.5 per cent (N = 142). Both figures pertain to the 1911 census. Arensberg and Kimball note that when they did their fieldwork in 1932–1934, a male who did not yet head his own household was called a ‘boy’ (Arensberg, and Kimball, , Family and community, 55).Google Scholar

41 Statistical modelling of the probability that a household disappears between 1901 and 1911 suggests that a larger number of girls in a household does not increase the probability that the household disappears.

42 The published censuses of Ireland do not cross-classify data on servants by age-groupings small enough to be of any use for this essay. Breen used the 1901 and 1911 census schedules for his study area, but unfortunately never tabulates his data to show the proportion of a given age-sex group who are servants.

43 The poultry market is mentioned in Slater's royal national directory of Ireland, 9th edn. (London, 1894), p. 48.Google Scholar

44 The locales were chosen deliberately to be on the borders of counties, making them sensitive to county of birth, the only measure of internal migration available in the census. Most internal migrants in the Clare subsample were born in County Tipperary.

45 More than a few young people say they are the household head's ‘servant’, but then list an occupation of ‘labourer’. The terms ‘domestic servant’, ‘farm servant’, and ‘servant’ are taken to be synonyms; given the imprecision in these labels, no attempt was made to distinguish farm servants from domestic servants.

46 Since many young people left Ireland before reaching ages 15–19, the figures presented in the text pertain to the subset of the cohort remaining in Ireland at that age.

47 One often hears the comment that Irish heirs were not able to marry because their sisters ‘refused to leave home’, and no potential wife wanted a household burdened by surplus siblings. Guinnane, T., ‘Marriage substitutes and non-marriage in turn-of-the-century Ireland’ (unpublished working paper, Princeton University, 1989)Google Scholar argues that the reverse causality - men were less intent on marriage when a sister remained at home, taking care of household tasks - is more plausible, and presents evidence in support of that view.

48 Fitzpatrick, ‘A share of the honeycomb;’ and ‘The modernization of the Irish female’, in O'Flanagan, P., Ferguson, P. and Whelan, K. edsd., Rural Ireland 1600–1900: modernization and change (Cork, 1987).Google Scholar