English Population History from Family Reconstruction: 1580-1837.

E.A. Wrigley, R.S. Davies, J.E. Oeppen and R.S. Schofield (New York: Cambridge University Press, 1997. xxii plus 657pp. $85.00).

This tome tabulates thousands of family histories from Anglican parish registers reconstituted under the auspices of the Cambridge Group for the History of Population and Social Structure. If that enterprise has flagged of late, contrary to the authors' hopes, this publication may reinvigorate the field, as much by what it does not do as by what it does.

This equivocal judgement reflects the book's character as a flagship. Guaranteed to be an essential reference for European social historians, it eschews the excitement of speculation to convey more authority. The authors limit themselves to the twenty-six parishes with the best records, and for the years after 1789 to just eight. In interpreting those selected data, they look to statistics or physiology, rather than reaching out to a full range of life. The book takes for granted that economics explains most demographic behaviour. Yet the laborious task of linking families, name by name, to occupations or other social circumstances is left to future "textured local studies."

Comparison with analogous projects highlights national stereotypes. The central Institut National d'Etudes Demographiques directed a statistically rigorous geometric sampling. German investigations draw on local genealogies, celebrating the particularism of each Heimat. In England, disparate volunteer labours reveal a homogeneous countryside, in which individuals moved easily within a market formed by shared social norms.

The original characteristics of English rural history are not novel discoveries, but they bear repeating. Marriage was the motor of this quarter-millennium. By the 1830s, first-time brides averaged 23, three years younger than before 1730. Abetted by an increase in conceptions out of wedlock and a slight shrinking of intervals between births, falling ages at marriage generated a population boom: England grew four times faster than its continental rivals from 1550 to 1820.

That boom took place within a low-pressure demographic regime, in which both death and birth rates were low. The key was extended breastfeeding. Infants almost always continued nursing past their first birthdays and sometimes past their second, for a mean duration of nineteen months. As in southwest France and Frisia, breastfeeding kept children alive while delaying the arrival of siblings. These English customs seem independent of region, era, and occupational structure.

One neglected shift did disrupt chronological homogeneity. Partisans of a general crisis of the seventeenth century will take heart to learn that within their period (say 1630-1730), ages at marriage were higher, births out of wedlock were rarer, brides were less often pregnant, parity progression ratios were lower, and births came more seldom to wives in their forties. When it comes to periodization, the authors prefer a long eighteenth century from 1680 to 1820. However, the empirical results break in the middle, with the "very odd" seventeenth century confuting standard life tables until 1750.

One brief observation suggests a profitable recasting of debates over protoindustrialization. Parishes where larger percentages of men engaged in manufacturing by 1831 were characterized both by the highest ages at marriage in all eras and by the greatest fall in those ages over time.

The defences of the completeness and reliability of the sources are convincing, with one exception. The registers record sharply lower infant mortality after 1750. The drop, from 19% to 16%, is concentrated in the endogenous half of that mortality, deaths in the first month associated with genetic weaknesses. Such a plummet is redolent of the omission of perinatal deaths. As the delay before christening increased, lives measured in minutes might appear less frequently in baptism and even burial registers.

Reviewing reasons for regarding this decline in infant mortality as genuine illustrates the range of argument family reconstitution makes possible. The authors point to a coincident reduction in maternal mortality, from 12 to 9 and then after 1800 to just 6 deaths per thousand confinements. Yet this is much what one would expect if the very existence of the child that killed the mother went unrecorded. There seems no corresponding advance in wives' life expectancies; if fatal deliveries were under-reported, deaths in childbed may go unrecognized.

1750 also marks a drop in the sex ratio among endogenous deaths. This is compatible with decaying registration, as males were over-represented among the children provoking difficult labour and succumbing during it. A fall in the frequency of twins at much the same time arouses further suspicion. The observed frequency of twins is sensitive to endogenous mortality, since loss of even one of the pair disguises the fact that twins were conceived at all. From this point of view, the correlations between parishes with low maternal mortality and low twin mortality and low infant mortality generally are all too predictable.

If neonate mortality did not actually decline, it would be in keeping with death rates above one month and one year, and even in adulthood. At these ages, the Cambridge Group now sees the quantity of life increasing only after 1840. Apart from half-hearted suggestions that obstetrics may have improved, they offer no reason to suppose the youngest of the young escaped this trend.

Surprisingly, the first issue raised by English data remains unsettled. For Louis Henry, the main point of family reconstitution was to uncover the contours of natural fertility, the distribution of births in the absence of the stopping pattern known as family limitation. Anthony Wrigley's pioneering study of Colyton in Devon suggested that the years 1647-1719 (that general crisis again?) were noteworthy, not only for the constriction of marital fertility from without, but also for the deliberate halting of childbearing within marriage. If true, it was of theoretical interest that a population might take up family limitation for an extended period and then abandon it.

Here the authors once more steer a cautious course. Their words insist that the English were indeed subject to natural fertility. Yet the Coale-Trussell indices they calculate report statistically significant levels of family limitation, especially between 1650 and 1699. Half that ostensible family limitation is due to large numbers of pregnant brides (making their fertility seem unduly low at higher ages and longer durations of marriage, when it did not enjoy such a boost; so much the worse for tests of statistical significance).

The question remains: what about Colyton? A footnote keeps open the possibility that it was home to some genuine limiters, but the book never comes to grips with the question. Only close cross-referencing of ages at last birth with ages at marriage and gaps between spouses' ages could establish whether the greater diminution in fertility in longer-established marriages between 1650 and 1749 marks family limitation. Until such intricate work is linked to sources beyond the parish registers, the world we have lost will not be reclaimed in full.

Ernest Benz Smith College