Hostname: page-component-cd9895bd7-fscjk Total loading time: 0 Render date: 2024-12-25T17:06:41.223Z Has data issue: false hasContentIssue false
  • English
  • Français

Illegitimacy and its implications in mid-eighteenth-century London: the evidence of the Foundling Hospital

Published online by Cambridge University Press:  29 January 2009

Adrian Wilson
Get access Rights & Permissions [Opens in a new window]

Abstract

An abstract is not available for this content so a preview has been provided. Please use the Get access link above for information on how to access this content.
Image of the first page of this content. For PDF version, please use the ‘Save PDF’ preceeding this image.'
Type
Articles
Information
Continuity and Change , Volume 4 , Issue 1 , May 1989 , pp. 103 - 164
Copyright
Copyright © Cambridge University Press 1989

Access options

Get access to the full version of this content by using one of the access options below. (Log in options will check for institutional or personal access. Content may require purchase if you do not have access.)

References

ENDNOTES

1 For the increasing importance of marriage, see Wrigley, E. A. and Schofield, R. S., The population history of England 1541–1871: a reconstruction (London, 1981)Google Scholar; Outhwaite, R. B., ed., Marriage and Society: studies in the social history of marriage (London, 1981)Google Scholar; Smith, R. M., ‘Marriage processes in the English past: some continuities’, in Bonfield, L., Smith, R. M. and Wrightson, K. eds., The World we have Gained: histories of population and social structure (Oxford, 1986), 43–99.Google Scholar

2 See Wrigley, E. A., ‘Marriage, fertility and population growth in eighteenth-century England’Google Scholar, in Outhwaite, ed., Marriage and Society, 137–85.Google Scholar

3 Laslett, Peter, Oosterveen, Karla and Smith, Richard M. eds., Bastardy and its Comparative History (London, 1980)Google Scholar, contributions by Macfarlane; Oosterveen et al.; Newman; Levine and Wrightson; Wrightson; Laslett.

4 Wrigley, E. A., ‘A simple model of London's importance in changing English society and economy 1650–1750’ Past and Present 37 (1970) 44–70.CrossRefGoogle Scholar

5 A useful review of such studies is Meyer, Jean, ‘Illegitimates and foundlings in preindustrial France’Google Scholar, in Laslett, et al. eds., Bastardy, 249–63Google Scholar. Note that foundlings are specifically excluded from counts of illegitimate births in studies of English illegitimacy: Laslett, Peter, Family life and illicit love in earlier generations (Cambridge, 1977) 154–5.CrossRefGoogle Scholar

6 This claim arises chiefly from the Registrar-General's figures for the period 1842–1902. See Peter Laslett, ‘Introduction: comparing illegitimacy over time and between cultures’, in Laslett, et al. eds., Bastardy, 1–64Google Scholar. For the suggestion that this was ‘a continuation of earlier trends’, see Laslett, , Family Life, 146Google Scholar. Some support for the latter view comes from Finlay, Roger, Population and metropolis: the demography of London 1580–1650 (Cambridge, 1981) 149CrossRefGoogle Scholar. Here it emerges that even with foundlings counted as bastards, an eight-parish sample of some 25,000 births yielded an illegitimacy ratio of only 2.0 per cent in the early seventeenth century, compared with a contemporary national ratio of 2.3 per cent. However, several qualifications attach to this comparison. First, as Finlay has elsewhere pointed out, foundling children were ‘deposited …in the richer parishes’ (‘The population of London, 1580–1650’, unpublished Ph.D. thesis, University of Cambridge, 1976, 185)Google Scholar. To compensate for this, it might well be necessary to bring into view a still larger number of parishes, particularly since the single poor parish of St Botolph, Bishopsgate, contributed over 40 per cent of the births in the sample and generated the suspiciously low illegitimacy ratio of 0.6 per cent. Second, there is some evidence that the chronology of change in London illegitimacy between 1580 and 1650 was different from the national pattern (ibid., 186). Third, it might be more appropriate to compare London illegitimacy levels with those of the immediate hinterland. As Laslett has shown, the region ‘East’, which included Middlesex and Essex, had an overall illegitimacy ratio of 1.2–1.4 per cent in the period 1581–1640 (Family Life, 142; compare Laslett, et al. eds., Bastardy, 31–2)Google Scholar. Using this regional comparison, the London ratio of 2.0 per cent could be made to look high rather than low.

7 The contemporary national bastardy ratio was rising from 3.143 per cent (1750–1754), through 3.550 per cent (1755–1759), to 4.032 per cent (1760–1764): Laslett, , ‘Introduction’Google Scholar, in Laslett, et al. eds., Bastardy, 14Google Scholar. For the numbers of London births at this time, see note 22 below; for the numbers of foundlings, Section III below. (As will emerge in Sections VIII and X, the numbers fluctuated considerably during the General Reception. The figure of 10 per cent is a rough average.)

8 McClure, Ruth K., Coram's children: the London Foundling Hospital in the eighteenth century (New Haven, 1981)CrossRefGoogle Scholar; Nichols, R. H. and Wray, F. A., The history of the Foundling Hospital (London, 1935)Google Scholar. See also the works cited in note 11 below.

9 Notably in the conception of the General Reception as ‘indiscriminate admission’: Nichols, and Wray, , Foundling Hospital, 47–63Google Scholar; George, M. Dorothy, London life in the eighteenth century (London, 1927) 57Google Scholar. See also, in similar vein, Stone, Lawrence, The family, sex and marriage in England 1500–1800 (London, 1977) 475–6.Google Scholar

10 It began innocuously enough in May 1760 with the suggestion that the Hospital might take in the orphans of soldiers. For a study of nineteenth-century London illegitimacy based on successful petitions to the Foundling Hospital, see Gillis, John R., ‘Servants, sexual relations, and the risks of illegitimacy in London, 1801–1900’, Feminist Studies 5 (1979) 142–73.CrossRefGoogle Scholar

11 Such is the burden of the writings on the subject of the Hospital's former child and midnineteenth-century secretary John Brownlow. See his Chronicles of the Foundling Hospital (London, 1847)Google Scholar and History of the Foundling Hospital (London, 1848).Google Scholar

12 See Greater London Record Office, London (hereafter GLRO), F. H./Daily Committee Minutes, 25 March 1741; Nichols, and Wray, , Foundling Hospital, 38–9Google Scholar; McClure, , Coram's Children, 49–51.Google Scholar

13 See Shorter, Edward, The making of the modern family (New York, 1985)Google Scholar, paperback edition (Glasgow, 1977) 170–6, and Stone, , The Family, 476Google Scholar. For a critique of the logic of such inferences, see Wilson, Adrian, ‘Inferring attitudes from behaviour’, Historical Methods 14 (1981) 143–4.CrossRefGoogle Scholar

14 As was noted on 18 June 1755, with an order to the physicians and surgeons to ‘take particular care’ in the future: Thomas Coram Foundation for Children, London (hereafter TCFC), General Committee Minutes, vol. 5.

15 Usually the number of children to be taken in was twenty. In such a case, a bag was filled with 20 white balls, 5 red balls (possibly 4: General Committee Minutes, loc. cit.) and enough black balls to make up a number equal to the number of children who had been brought in during the two hours for which the doors were open in advance. The women then drew one ball each. Those drawing black balls were sent away; those with red waited on the side; and those drawing white balls had their children inspected, one by one, for age and health. Rejected children from this latter group were to be replaced by others from the red-ball group. It will be noted that the latter could not, in turn, be replaced; yet on every occasion but one the full 20 children were taken in. Thus some children above the age-limit may well have been admitted, as I shall subsequently suggest – and still more could of course have been brought. For the official description of this procedure, and for a contemporary engraving depicting it, see McClure, , Coram's Children, 76–8.Google Scholar

16 Ibid, 76–136.

17 ‘To the day’, except for those 11 days (3–13 September) lost in 1752 with the shift from the Old Style (Julian) to the New Style (Gregorian) calendar. Cheney, C. R. ed., Handbook of dates for students of English history (London, 1978) 10–11 and 154–5.Google Scholar

18 See Section VI below (children aged over two months) and Appendix 3 (the effective cutoff of three months).

19 There were 22 takings-in with ‘spacing’ (interval since the previous taking-in) of seven weeks or less. Amongst these, the median spacing was 5 weeks, and the median number of children brought was 58. This implies a rate of 603 children per year. Using a more complex method (see Appendix 3), we arrive at an estimate of 597 per year. (As far as can be ascertained, children were seldom or never brought to more than one taking-in: see note 40 below.)

20 Some 76 children were brought to the final taking-in, of whom 20 were admitted and the remaining 56 were rejected. Let us imagine that a further x mothers were ‘deterred’ from bringing children to this taking-in. Then some (56 + x) ‘potential foundlings’ existed at this point. After 25 days the General Reception started. Many of these (56 + x) children will have been brought within the first week or so of the General Reception. So too will foundlings who had ‘accumulated’ during the intervening 25 days. Hence a large-scale initial ‘rush’ is to be expected at the start of the General Reception. Thus it is not surprising to find (from the billets – see Appendix 1) that in the first week of the General Reception an estimated 184 London foundlings were brought to the Foundling Hospital – far more than had ever been brought to any taking-in. (The number of 184 has been inflated, from 147 London billets, to take account of the fact that 45/223 billets for this week were missing or had missing or ambiguous information.) In the next three weeks, as this initial rush declined, the numbers of London foundlings were 57, 44 and 39. If we assume, for the sake of argument, that (a) all the (56 + x) children were in fact brought during the first four weeks of the General Reception and (b) the rate at which foundlings were ‘accumulating’ was constant between the penultimate taking-in and the end of the fourth week of the General Reception, we can estimate the value of x. A more precise estimate is obtained by relaxing assumption (b), i.e. by taking into account seasonal and secular change (using the findings of section VII and VIII below). This estimate turns out to be 49. That is, a taking-in which saw 76 children brought also saw a further 49 mothers ‘deterred’, or roughly two mothers deterred for every three who did bring children. Expressed as an index, the implied rate is 64 deterred for every 100 brought. This will be a conservative estimate, since assumption (a) above prevents us from taking account of the fact that some of the (56 + x) children will have been ‘dropped’ on church porches, or will have died, between the final taking-in and the start of the General Reception. A largely independent estimate, not subject to this limitation, is given in Appendix 3: there the number deterred becomes 82 for every 100 brought. Given that we expect the estimate of 64 to be conservative, the two estimates are in excellent agreement.

21 (a) See Section V below, (b) This emerges from the baptisms-deficit in the London Bills of Mortality: see Figure 5 and note 57 below, (c) See Sections VIII and X below.

22 The annual number of births in London can be estimated at 19,945 for the 1750s, and 21,609 for the 1760s. See Wrigley, and Schofield, , Population history of England, 79Google Scholar (corrected London baptisms), 101 (correction for nonconformity and late baptism), 138 (correction for nonobservance). For the 1760s nonconformity/late baptism correction I have used the mean of the factors for the 1750s and 1770s; the non-observance corrections for 1756 and 1766 were assumed to apply to the 1750s and 1760s respectively. The figure for the 1750s should be inflated, by about 1.02, to take account of the baptisms-deficit in the London Bills of Mortality due to the General Reception (Section VIII below). This produces a corrected figure of 20,344.

23 The full Charter is reproduced in Nichols, and Wray, , Foundling Hospital, 329–336.Google Scholar

24 George, , London Life, 215.Google Scholar

25 Shorter, , Modern Family, 175Google Scholar, citing Peyronnet's study of Limoges. Compare also Meyer, , ‘Illegitimates and foundlings’, 253–6Google Scholar, citing works on Rennes, Lyons and St Malo as indicating significant numbers of legitimate children amongst foundlings in these towns. For the suggestion that London foundlings were ‘legitimate as well as illegitimate’, see Laslett, , Family life, 110Google Scholar, note 7. Stone, , The family, 476Google Scholar, goes further, claiming (without supporting evidence) that ‘a majority [of London foundlings] seem to have been legitimate children of couples who were financially unable to support them.’

26 These are rough estimates derived from the extent of adult-mortality in the London Bills of Mortality at this time. Indirect confirmation is provided by An account of the rise, progress and stale of the General Lying-in Hospital (London, 1768), 3Google Scholar. Here, out of 8,768 women delivered since 1752, 1,069 were described as widows, and 1,943 as ‘objects who were seduced’, that is, as bastard-bearers. Thus amongst the women relieved by the General Lying-in Hospital, there was apparently one who had been widowed in pregnancy for every two bastard-bearers. Postnatal orphaning will have been still more common, since two parents were at risk, over twelve months rather than nine months.

27 On the billets, see Appendix 1. Numbers of foundlings brought to takings-in are given in TCFC, General Committee Minutes, vols. 3–5, passim. Numbers admitted during the General Reception over time can be found in GLRO, F. H./Secretary: Children/Nursery Books, vols. 1–4.

28 Wrigley, and Schofield, , Population history of England, 292Google Scholar note 14, citing studies from Belgium, Sweden and France in periods ranging from the seventeenth to the twentieth century. See also Wrigley, , ‘Marriage, fertility and population growth’, 163–5Google Scholar, outlining results for England 1661–1820, based on family-reconstitution studies.

29 London bread-prices (discussed in note 45 below) show considerable year-to-year change (Figure 5) and a very weak seasonal cycle. Even in a period when secular change was unusually slight (1749–1751), an analysis of variance shows that seasonal effects account for only 21 per cent of the variance, and generate an F ratio of less than unity.

30 On the London Bills of Mortality see Laxton, Paul, ‘A guide to the London Bills of Mortality’ (Cambridge, forthcoming)Google Scholar; Wrigley, and Schofield, , Population history of England, 77–83Google Scholar; Christie, James, Some account of parish clerks (London, 1893)Google Scholar; Ditchfield, P. H., The parish clerk (London, 1907)Google Scholar; George, , London life, 318Google Scholar note 1 and the sources there cited; Finlay, Population and metropolis; and Marshall's compilation, cited in note 49 below. The collection I have used is that in the Guildhall Library, London.

31 It can be seen from Figure 1 that in the week's admissions sampled from the eighth 4-week interval, the number of children from the distant counties dropped to zero. This was doubtless the result of severe weather at this time (5–11 January 1757). In these four weeks as a whole, 206 children were brought: in the sampled week, only 20. Thus the weather caused a freak drop in the bringing of children even from London; it is intelligible that this would have hit still harder at the distant counties. Hence this particular week in the sample is unrepresentative.

32 Of the 327 billets I have sampled from 1759 (see Appendix 1), 90 carried written tokens which yielded geographical information. The farthest places mentioned (mostly on one token each) were Dorset, Monmouthshire, Herefordshire, Wales, Cheshire, Shropshire (three tokens) and Durham. With the sole exception of Wales, all these places had been ‘recruited’ by August 1757. (For the record, 19 of these 90 children came from London. As is explained in note 38 below, this proportion was certainly unrepresentative.)

33 Chi-square = 4.4, df = l, p < 0.05. For guides to such statistical tests, see Siegel, Sidney, Nonparametric statistics for the behavioural sciences (Tokyo, 1956)Google Scholar, and Blalock, Hubert M. Jr., Social statistics (New York, 1960)Google Scholar. On the possible discrepancy between the Out-Parishes and the rest of London, see also the discussion of the baptisms-deficit at the end of Appendix 2.

34 Based on one-week-in-four samples (Appendix 1). For this purpose a replacement week was sampled for the 8th interval (cf. note 31 above). The estimates were derived in two different ways: by taking as representative of any given 4 weeks (a) the numbers of London children, and (b) the proportions of children coming from London, in the sampled week. The figure of 2,018 for the first 52 weeks is the mean of two such estimates (1,988 and 2,048). The figures of 1,792 and 1,902 represent the extremes from all the 4×2 = 8 estimates for 52-week periods starting 4, 8, 12 or 16 weeks after the beginning of the General Reception.

35 Cf. note 20 above. It should be added that, as we shall see in Sections VIII and X below, the General Reception numbers were abnormally high, chiefly because of the massive price-rise which began shortly before the General Reception.

36 See Wrigley, E. A., ‘Births and baptisms: the use of Anglican baptism registers as a source of information about the numbers of births in England before the beginning of civil registration’, Population Studies 31 (1977) 281–312CrossRefGoogle ScholarPubMed. Here the highest infant mortality mentioned (Liverpool 1875) was 234 per 1,000 (308). Even if we take the highest mentioned exogenous infant mortality of 211 (Liverpool 1875), and add the highest endogenous infant mortality of 122 (Banbury 1700–1749: see 286), we arrive at a total of 333 per year. That is, on the most pessimistic possible assumptions, two-thirds of infants survived to the end of the first year of life. On these assumptions, over 40 per cent of postnatal orphans will have been aged over 6 months. Yet, as is explained in the text, only three children in the sample of 327 had age-estimates of over 6 months (all three being described as ‘nine months’), and in the one case which can be checked, the true age was only 6 months.

37 See Wilson, Adrian, ‘Participant or patient? Seventeenth-century childbirth from the mother's point of view’, in Porter, Roy ed., Patients andpractioners: lay perceptions of medicine in pre-industrial society (Cambridge, 1985) 129–44Google Scholar; Notes and Queries, 9th series, 3, (1899) 212Google Scholar: The Oxford English Dictionary (12 vols., Oxford, 1933)Google Scholar, under ‘month’ (3.f) and ‘uprising’ (2.c).

38 (a) In these combined samples, of the 181 foundlings with estimated ages under 30 days, only 37 (20 per cent) were supplied with place-information. By contrast, of the 146 foundlings with age-estimates of 30 days or more, 53 (36 per cent) had placeinformation. This difference is statistically significant at the 0.01 level (chi-square = 10.0, df = 1). Thus younger children are under-represented amongst those with placeinformation, (b) It is also certain that London children were under-represented. Only 19 of the 90 children supplied with place-information were described as having been born in London parishes. If this proportion were representative, then in 1759 only about 850 foundlings per year would havè come from London; but in fact, as will emerge in Section X below, the number was certainly greater than this and probably about twice as high.

39 See note 15 above.

40 A by-product of this result is to suggest that it was very unusual for a child to be brought to two successive takings-in. At any given taking-in, about two-thirds of the children were balloted out and sent away (notes 15 and 19 above). In principle, the mothers of these children could have brought them again to the next taking-in, to try a second time. Had this been common, then there would have been many children aged over two months even at takings-in spaced five weeks apart – since after an elapse of five weeks, most of the children brought to the previous taking-in would now be aged over two months. As we have seen, this did not happen. Thus it would appear that after being balloted out of a taking-in, mothers did not return a second time. Did they ‘drop’ their children on the church porch? We do not know.

41 The only child in these samples known to be aged over 60 days – its actual age being 74 days – was also the only child of known age coming from the last five children at a long-spaced taking-in. This child's age was, of course, described as ‘two months’ on its billet. Only one other child of known age was described as ‘two months’; this child's true age was 55 days, and it came at a short-spaced taking-in.

42 The reason for this flattening of the cycle is twofold. The long period over which paternal death is possible (9 months) has the effect that the mortality-component of pregnancy-orphaning is relatively weakly cycled. And the synchrony of the two component cycles means that combining them in this way – i.e., effectively with a five-month lag – is always ‘disadvantageous’ to the strength of the resultant cycle.

43 See Wrigley, and Schofield, , Population history of England, 293–6Google Scholar. In addition, we should consider the special case of maternal mortality, that is, the death of the mother either in childbirth oras a result of childbirth. Such deaths were recorded as ‘childbed’ in the London Bills of Mortality: for the period 1749–1756 they totalled 1,508, a rate of 188.5 per year, implying a maternal mortality rate of around 9 per thousand (if referred to total births: note 22 above) or 13 per thousand (if referred to baptisms in the Bills of Mortality). This figure is of the expected order of magnitude: see Roger Schofield, ‘Did the mothers really die? Three centuries of maternal mortality in “the world we have lost”’; in Bonfield, , Smith, and Wrightson, eds., The world we have gained 251–60Google Scholar; Loudon, Irvine, ‘Deaths in childbed from the eighteenth century to 1935’, Medical History 30 (1986), 1–41CrossRefGoogle ScholarPubMed, at 17–18. Although it might well be expected that maternal death would occasion the sending of the child to the Foundling Hospital, several considerations rule this out as a major source of foundlings. (1) A disproportionate number of maternal deaths will have been associated with stillbirths: Schofield, , op. cit.Google Scholar, and Eccles, Audrey, Obstetrics and gynaecology in Tudor and Stuart England (London, 1982) 125–30Google Scholar. (2) Even if all deaths ‘in childbed’ led to the child's being sent to the Foundling Hospital, this will account for less than one in five of the foundlings during the period of takings-in, and a still lower proportion during the General Reception. (3) The seasonal cycle of ‘childbed’ burials in the Bills of Mortality 1749–1756, though roughly similar to that of adult burials in general, showed two distinctive features: a sharp drop in December, and an even sharper rise in late July and early August. Neither of these features is reflected in the seasonal cycle of numbers of foundlings (Figure 4).

44 The baptisms-deficit was estimated by assuming a linear growth in numbers of baptisms between 1755 and 1761. This generated ‘expected’ numbers of baptisms in the Bills of Mortality throughout the General Reception. A 52-week moving window, moving in steps of 4 weeks, was used. For each of the 38 positions of this window, the actual Bills of Mortality baptisms were subtracted from the ‘expected’ numbers, yielding the deficit plotted in Figure 5.

45 The derivation of seasonally adjusted numbers of children brought to takings-in, used in Figure 5, is explained at the end of Appendix 3. Bread prices and adult burials are taken from the weekly Bills of Mortality. These figures require some explanation, (a) Burials were given in age-ranges of 0–2, 2–5, 5–10, 10–20, 20–30, and so on up to 90–100. The numbers of burials in the various 10-year age-ranges between 20 and 60 moved up and down (both seasonally and from year to year) in close synchrony. Thus, while the decision to use the full range 20–60 to represent adult burials for Figure 5 is an arbitrary one, it does not affect the result, (b) Bread prices were specified in ‘Assizes of Bread’ held (irregularly) about seven times a year. The prices were given in two different forms: the prices (shillings, pence, farthings) of loaves of three standard weights, and the weights (pounds, ounces, drams) of loaves of three standard prices. For each of these six loaf-sizes, two varieties of flour (three in the case of the penny-loaf) were specified, with different prices. Thus there were 13 different price-measures in all. Usually, but not invariably, all these 13 price-measures moved up or down simultaneously. The measure I have used is the weight of the household flour penny-loaf, the cheapest loaf specified in the Bills. This has been tabulated in intervals of 4 weeks, and then converted to 52-week moving averages. Thus each point on the price-curve in Figure 5 represents the average weight of 52 household-flour penny-loaves. The scale of prices in Figure 5 is linear for weights, rather than for prices, as this simplified the calculations.

46 For Figure 6, as for Figure 5, the weight of the household-flour penny-loaf was the price-measure selected. The 4-week price-tabulations were made in such a way that the taking-in itself fell during the first week of the 4-week interval designated as ‘lag 0’ in Figure 6. Thus takings-in can be considered to have taken place between ‘lag 1’ and ‘lag O’.

47 See Hair, P. E. H., ‘Bridal pregnancy in rural England in earlier centuries’, Population Studies 20 (1966) 233–43CrossRefGoogle ScholarPubMed, and ‘Bridal pregnancy in earlier rural England further examined’, ibid. 24 (1970) 59–70; Laslett, , Family Life, 128–30Google Scholar, and ‘Introduction’ in Laslett, et al. eds., Bastardy, 7–12Google Scholar (and several other contributions in the latter volume); Wrigley, and Schofield, , Population history of England, 254Google Scholar; Wrigley, , ‘Marriage, fertility and population growth’, 158–63.Google Scholar

48 Wrigley, and Schofield, , Population History of England, 368–9, 374–6Google Scholar (econometric analyses by Ronald Lee) 327–8, 348–50, 417–30, and chapter 11 (454–84) passim; Houlbrooke, Ralph A., The English family 1450–1700 (London, 1984), 67.Google Scholar

49 These figures are taken from Marshall, J., Mortality in the metropolis (London, 1832) 67.Google Scholar Wheat prices are for Oxford, further up the Thames, but it is likely that London price-movements were very similar.

50 See Wrigley, and Schofield, , Population history of England, 30.Google Scholar

51 Shorter, , Modern family, 171.Google Scholar

52 Billet 2, 148 (18 August 1756). Much more common, in fact, was for the written token simply to state the child's name.

53 Wrightson, Keith, ‘The nadir of English illegitimacy in the seventeenth century’Google Scholar, in Laslett et al. eds., Bastardy, 176–91Google Scholar, gives an example (179) of a bastard-bearer's father maintaining her and her child until forced by the manorial jury's fines to eject them. Such pressure from local sources of collective authority is a recurring theme of Wrightson's chapter. This suggests that the apparent difference between the treatment of bastard-bearers and that of pregnant widows could well have originated outside the family, even if it was mediated through the family. For a similarly telling example, see Chaytor, Miranda, ‘Household and kinship: Ryton in the 16th and 17th centuries', History Workshop 10 (1980) 25–60, at 48.CrossRefGoogle Scholar

54 Elliott, Vivien Brodsky, ‘Single women in the London marriage market: age, status and mobility, 1598–1619’Google Scholar, in Outhwaite ed., Marriage and Society, 81–100Google Scholar (see 87–8). There are some indications that remarriage was easier for widows than for widowers in seventeenth-century London. See the interesting discussion by Carlton, Charles, The court of orphans (Leicester, 1974) 67–73Google Scholar, particularly 69–70. Nationally, it is known that the numbers of remarriages were very high in the seventeenth century, and fell considerably in the course of the eighteenth century (Wrigley, and Schofield, , Population history of England, 258–9.Google Scholar Remarriage was clearly an attractive prospect for widows and widowers who held even small amounts of real property. See Laslett, Peter, The world we have lost (London, 1965Google Scholar; 1976 edition, 103–5); Spufford, Margaret, Contrasting communities (Cambridge, 1974) 116–7CrossRefGoogle Scholar; and Chaytor, , ‘Household and kinship’, 37–8.Google Scholar Chaytor's paper has been extensively criticised by Houston, Rab and Smith, Richard, ‘A new approach to family history?’ History Workshop 14 (1982), 120–31.CrossRefGoogle Scholar In connection with remarriage, Houston and Smith point out the importance of the widow's landholding and the state of the general land market as further determinants of the propensity to remarry. Despite this and other technical criticisms, they observe that Chaytor has ‘every justification’ for ‘arguing a case for the distinctive circumstances of women’ (124) and accept that ‘temporary privations were often part of the widow's life cycle’ (125).Google Scholar

55 Wrigley, , ‘Marriage, fertility and population growth’Google Scholar, passim, particularly 146, 155–63, 167–71.

56 See ibid., 161–2. As Wrigley observes (161, note 36), even at the height of ‘repetitive’ bastard-bearing in the late eighteenth century, it turns out that over 80 per cent of all bastard births were first births. Using this figure, a bastardy ratio of 4 per cent implies that 12.8 per cent of first births were illegitimate. The true level will be somewhere between this conservative estimate and 16 per cent of first births. (Cf. also Wrigley, and Schofield, , Population history of EnglandGoogle Scholar, 233 note 66).

57 The baptisms-deficit peaked at about 1,200 per year (Figure 5). Of the 108 London billets I have examined from Phase 1 of the General Reception (Appendix 1), 66 described the child as having been christened already, and 42 as not christened. These reports were probably accurate. (The expected numbers christened and not christened are 64 and 44, based on a comparison between the baptisms-deficit and the estimated total numbers of London foundlings in the first year of the General Reception.) Assuming that these proportions remained constant (i.e., that the age-profile of foundlings did not change), we can use these proportions to inflate the baptisms-deficit so as to yield an estimate of the total number of London foundlings. Inflating 1,200 by 108/42 yields an estimate of just over 3,000. The lower 0.025 confidence limit of this estimate is 2,484. For the numbers of London births at this time, see note 22 above.

58 Brown, Robert Lee, ‘The rise and fall of the Fleet marriages’Google Scholar, in Outhwaite ed., Marriage and society, 117–36.Google Scholar Brown's Table I (123) shows 5,551 recorded Fleet marriages in 1740, and an estimated number of 6,609 for the same year. The total number of marriages in London was of the same order some two decades later (see notes to Figure 7 above).

59 This claim was made at the time by Joseph Massie and other critics of the General Reception. See McClure, , Coram's children, 108–9.Google Scholar

60 The final level of the baptisms-deficit was about 600 per year (Figure 5). Using the same procedure as in note 57 above, this yields an estimate of 1,543 for the total numbers of London foundlings at this time.

61 Lee, Ronald, ‘Short-term variation: vital rates, prices, and weather’Google Scholar, in Wrigley, and Schofield, , Population history of England, 456–401Google Scholar, at 375. The elasticity is the percentage change in the dependent variable induced by a one per cent change in the explanatory variable. This incorporates change in both directions, but can most readily be grasped by considering the case of a one per cent increase in prices. The response of nuptiality to price-increases tends to take the form of a depression for up to four years (including the year of the price-change), followed by a smaller brief compensatory upswing. Concentrating on the former effect, and taking the mean of Lee's elasticities for 1641–1745 and 1746–1834, we find a total elasticity of –0.2325, of which just over one half (–0.118) occurs after a lag of a year or more. See also Wrigley and Schofield in ibid. 327–8; and compare note 72 below.

62 Levine, David and Wrightson, Keith, ‘The social context of illegitimacy in early modern England’Google Scholar, in Laslett et al., eds., Bastardy, 158–75Google Scholar, at 171; Wrightson, , ‘The nadir of English illegitimacy’, 189Google Scholar; Wrigley, , ‘Marriage, fertility and population growth’, 158–60.Google Scholar See also Smith, , ‘Marriage processes’, 90.Google Scholar

63 Wrightson, , ‘The nadir of English illegitimacy’, 188–91Google Scholar, citing the experience of Burnham-on-Crouch, Essex, in the early 1630s. Compare also Levine, and Wrightson, , ‘The social context of illegitimacy’, 163, 165Google Scholar (Figure 5.1) and 171–2, dealing with an illegitimacy crisis some three decades earlier in Terling, Essex. In the latter case, a protracted economic crisis lasting five years (1594–8) halved the numbers of marriages and produced a subsequent illegitimacy peak which lasted for eight years (1598–1605).

However, the dearth of 1630–1631 had no effect on illegitimacy in Terling, (172)Google Scholar, in sharp contrast to events in nearby Burnham. To interpret this contrast, we would need to know (a) what happened in Burnham in the 1590s, and (b) the numbers of marriages in Terling during and after the dearth of 1630–1631.

64 Wrigley, , ‘Marriage, fertility and population growth’, 158–61Google Scholar; Smith, , ‘Marriage processes’, 85–8.Google Scholar

65 Elliott, , ‘Single women’, 86–7Google Scholar; Finlay, Population and metropolis, 137–9. The mean female ages at marriage given here are, respectively, 20.5 (for 496 London-born women marrying by licence, 1598–1619) and 23.1 (reconstructed mean for 70 London-born women identified from four parish registers, c. 1580–1650). By comparison, the mean of the female mean ages-at-marriage in 12 non-London family reconstitution studies was 26.0 in the period 1600–1649 (Wrigley and Schofield, Population history of England, 248, 255).Google Scholar See also Laslett, , ‘Introduction’Google Scholar, in Laslett et al. eds., Bastardy, 21.Google Scholar

66 Such a role for London was classically suggested by Wrigley, , ‘A simple model’.Google Scholar

67 See note 6 above, and Smith, , ‘Marriage processes’, 98 note 169.Google Scholar

68 This estimate may be conservative, in that it takes no account of the fact that the Out-Parishes were probably under-represented amongst children brought to the Foundling Hospital. (For this point see the end of section V above, and the end of Appendix 2.)

69 See note 7 above.

70 The financial interests of orphans of freemen in London and certain provincial boroughs were handled by local ‘courts of orphans’. See Carlton, , The court of orphansGoogle Scholar, cited in note 54 above. The records of these courts may serve as a starting-point for such an investigation.

71 Levine, David, Family formation in an age of nascent capitalism (New York and London, 1977).Google Scholar Levine finds that in Terling, 1760–1784, the level of illegitimacy was much higher in ‘hungry years’ than in ‘normal harvest years’ (132); that the adoption of the Speenhamland system abolished this effect (133); and that Shepshed, Lancashire, 1758–1850, shows no such effect at all, despite manifesting years of crisis illegitimacy (139–41). Clearly, therefore, the impact of prices on illegitimacy was heavily dependent on the local economic and social context.

72 Wrigley and Schofield's rural data, as analysed by Lee, yields the following picture: In the period 1641–1745, the negative effect of prices on nuptiality (cf. note 61 above) was spread roughly equally between annual lags of 0, 1 and 2. In the period 1746–1834, by contrast, two-thirds of the effect occurred at lag 0, and the remaining one-third at lag 1, while lag 2 now showed a powerful positive compensatory effect. Thus in the earlier periods, prices acted over a long interval (36 months), centring on a point lagged by about 12 months; in the later period, prices acted over a shorter interval (24 months), centring on a point lagged by perhaps 6 months. In view of this dramatic shift, it would be interesting to see how the corresponding monthly pattern, which is available from 1691, behaved over different subperiods within the period 1691–1834. This might help to resolve the discrepancy which Lee notes between the monthly and annual data. See Lee, , ‘Short-term variation’, 369, 375.Google Scholar The pattern shown by the monthly data (369) seems consistent with the ‘London’ courtship model, whereas the annual data, even for 1746–1834, imply a slower or ‘rural’ process. See also Wrigley, and Schofield, , Population history of England, 349–50.Google Scholar

73 GLRO, F. H./Billet Books (here the billets are mostly in correct numbered sequence): TCFC, vols. 18, 35 and 133 (these billets had probably been extracted from the main series partly in cases of reclaiming, and partly as curiosities). For illustrated accounts of the billets, and especially of the more colourful of the ‘tokens’ attached to them, see McClure, , Coram's children, 81–6Google Scholar, and Nichols, and Wray, , Foundling Hospital, 119–26.Google Scholar

74 Wrigley, and Schofield, , Population history of England, 286–93.Google Scholar This pattern specifically excludes London (286 note 3).

75 Dyer, Alan, ‘Seasonally of baptisms: an urban approach’, Local Population Studies 27 (1981), 26–34.Google Scholar

76 Boulton, Jeremy, Neighbourhood and society: A London suburb in the seventeenth century (Cambridge, 1987), 50–53CrossRefGoogle Scholar; Finlay, ‘Population of London’, 181–3.

77 See note 28 above.

78 Baker, R. A. and Neider, J. A., The GLIM System Release 3 (1978).Google Scholar GLIM is far easier to use, as well as more flexible and powerful, compared with certain other statistical packages.

79 See Edwards, A. W. F., Likelihood: an account of the statistical concept of likelihood and its application to scientific inference (Cambridge, 1972).Google Scholar