¹ Unless these conditions are met, no measure of covariation can be made independent of seasons, harvest conditions, and long-term trends. The method has not yet been pushed back into the seventeenth century, although Chartres will reportedly be presenting a manipulation of Houghton's data for the 1690s in the forthcoming fifth volume of the Agrarian History of England and Wales.Google Scholar

² Granger, C. W. J. and Elliott, C. M., “A Fresh Look at Wheat Prices and Markets in the Eighteenth Century,” Economic History Review, n.s. 20 (08. 1967), pp. 257–65.CrossRefGoogle Scholar

³ Overton, Mark, “Agricultural Change in Norfolk and Suffolk, 1580–1740” (diss., Cambridge, 1980);Google ScholarOverton, Mark, “Estimating Crop Yields from Probate Inventories: An Example from East Anglia,” this JOURNAL, 39 (06 1979), pp. 363–78;Google ScholarYelling, J. A., “Probate Inventories and the Geography of Livestock Farming,” Transactions of the Institute of British Geographers, 51 (11. 1970);Google ScholarSkipp, V. H. T., Crisis and Development: An Ecological Case Study of the Forest of Arden, 1570–1674 (Cambridge, 1978).CrossRefGoogle Scholar

⁴ Thirsk, Joan, “The Content and Sources of English Agrarian History after 1500,” Agricultural History Review, 3 (Part 2, 1955), p. 71;Google ScholarYelling, “Probate Inventories,” p. 111.Google Scholar

⁵ Duckham, A. N., Agricultural Synthesis: The Farming Year (London, 1963).Google Scholar

⁶ See Wrigley, E. A. and Schofield, R. S., The Population History of England, 1541–1871: A Reconstruction (Cambridge, Mass., 1981), chap. 8.Google Scholar

⁷ It is somewhat churlish to boast of the availability of data when the set that forms the basis of the study fell into the author's lap, courtesy of the Economic and Social Research Council-Cambridge Group. Nonetheless, marriage registers survive for more places over more years than does any other single source for agrarian history; the larger project should see 100–200 parishes added to their sample.Google Scholar

⁸ Kussmaul, Ann, “Time and Space, Hoofs and Grain: The Seasonality of Marriage in England,” Journal of Inerdisciplinary History, 15 (Spring 1985).Google Scholar

⁹ An illustration of the first two patterns can be found in the marriage map, Figure 8.4, of Wrigley and Schofield's Population History. There, in 1601–1710, autumn marriage peaks are found in the arable east and midlands, and spring/summer peaks in the pastoral west. Their Table 8.6 showed, moreover, the close association of autumn marriage with parishes mapped by Thirsk as “mixed” farming types, and of April–June marriage with “open pasture” parishes.Google Scholar

¹⁰ Marriage peaks have been used in this study, rather than troughs. The trough in marriages in arable places, measured in terms of, say, August marriages, would be acceptable, but the workimposed trough in pastoral regions coincides too neatly with Lent, as the image of the Paschal Lamb should suggest.Google Scholar

¹¹ See Appendix for a complete exposition of method. In brief, indexes were derived from the raw totals of marriages per month, adjusting for days in months and total marriages; double-month totals for autumn and spring (halved to bring the random value back down to 100) were calculated for overlapping 40-year periods. Parishes were dropped from observation in any period when the total of marriages in that period fell below 60.Google Scholar

¹² The parishes of Figure 1 were not all drawn from a hat, because of their interesting patterns; it was tempting to include autumnal Barley, Hertfordshire, and vernal Cowfold, Sussex. Orwell and Willingham are two of the “Contrasting Communities” of Margaret Spufford's work; Shepshed and Bottesford are the parishes of David Levine's study of rural industrialization; Earsdon and Sedgley are among the reconstitution parishes of the Cambridge Group's sample. See Spufford, Margaret, Contrasting Communities: English Villagers in the Sixteenth and Seventeenth Centuries (Cambridge, 1974), pp. 94–99, 121–32;CrossRefGoogle ScholarLevine, David, Family Formation in an Age of Nascent Capitalism (New York, 1977), pp. 16–17, 88–91. Sedgley is unusually well documented, with a published parish register that includes occupations (changes in occupation closely match the pattern of its seasonality of marriages), and transcribed probate inventories. Close inspection of its marriage seasonality reveals that the period of greatest decline in agricultural seasonality must have occurred shortly after 1621. There is a gap in the occupational listings in the register from 1627 to 1677, but between 1578 and 1625, 38 percent of fathers of baptized children were given agricultural designations, while 48 percent were returned as metal workers (N = 1,463); by 1675/85, the proportion in agriculture had plummeted to 9 percent, while the proportion in metal-working had leapt to 66 percent (N = 724). See Staffordshire Parish Registers Society, Sedgley, 1558–1684 (1940–1941); and J. S. Roper, Sedgley Probate Inventories, 1614–1787.CrossRefGoogle Scholar

¹³ For all but one of the remaining examples the 129 parishes with market charters were dropped from the sample; included on the map are the “regions” employed in Figure 7 (which, incidentally, is the exercise that includes the market towns).Google Scholar

¹⁴ The triangles are largely indicative of January/February or December/January marriage peaks. There is a high likelihood that these had three very different and regionally specific causes, representing variously the early lambing of the southwest, the survival of a pattern of post-Advent, pre-Lent marriage (such as that noted in France, for example) in the northwest, and the nonagricultural seasonality of industrial places, where any months may have been the relative maxima. On the first point, see Bettey, J. H., “The Development of Water Meadows in Dorset during the Seventeenth Century,” Agricultural History Review, 25, part 1, (1977), p. 41; on the second, northern Staffordshire, Cheshire, and Derbyshire hired even their farm servants in January and February (see Ann Kussmaul, “Statute Sessions and Hiring at Fairs,” forthcoming); on the third, a communication from Marguerite Dupree concerning the December marriages of Staffordshire pottery workers has suggested that that, too, was a slack time in the work year, coinciding with the slowing of transportation of raw materials as canals froze.Google Scholar

¹⁵ See Kussmaul, “Time and Space,” for companion maps to Figures 2 and 3.Google Scholar

¹⁶ The coefficient of variation was preferred to the standard deviation (on which it is based) for demonstrating the growing dispersion of the indexes in the later seventeenth century because of its resistance to influence by varying sizes of the means; see also Kussmaul, “Time and Space,” for graphic demonstrations of the process.Google Scholar

¹⁷ Note that the increase in distance is independent of overall sample size, varying rather with the proportion of the sample that was A-type, in this definition: the mean minimum distance remains virtually unchanged at less than 20 kilometers as the sample size grows, until the midseventeenth century; it then jumps, while the sample size remains little changed.Google Scholar

¹⁸ See Appendix for a discussion of “tuning,” for the totals of parishes per type in each period, and for a demonstration of a second useful tuning.Google Scholar

¹⁹ The choice of Cramer's V is crucial here: its size is independent of the sample size, unlike the chi-square statistic on which it is based. The number of parishes in observation (60 or more marriages) increased until 1601/40, then fell, only to rise again to a maximum in 1781/1820. The paired periods for the frequency analysis were odd-to-odd and even-to-even pairs of the overlapping periods. In no instance of pair-wise comparison, it should be noted, was the sequence of types independent (by the criterion of the chi-square statistic, at the 0.0001 level of error).Google Scholar

²⁰ It is tempting to read too much into the value of the minimum Cramer's V in 1601/40 to 1641/80. The Interregnum was marked by much upset to the registralion of marriages—legal (Civil Marriage), religious, and political. Underregistration reached its highest level, and the pattern of seasonality grew suspiciously flat (see Wrigley and Schofield, The Population History, p. 25). Four of the comparisons between periods involve the years 1641/60, and the persistence of types is more likely to have been weakened to an immeasurable extent by statistical “noise” then, than in any of the other comparisons. Cheering, though, is the value of Cramer's V. 1601/40 to 1661/1700, that is, avoiding the Interregnum entirely: while higher than the original minimum of 0.33, at 0.43 it would become the new minimum of the series.Google Scholar

²¹ Chartres, J. A., Internal Trade in England, 1500–1700 (London, 1977);CrossRefGoogle ScholarChartres, J. A., “Road Carrying in England in the Seventeenth Century: Myth and Reality,” Economic History Review, 2nd ser. 30 (02. 1977), pp. 73–94.CrossRefGoogle Scholar

²² To simplify the diagram the regions have been given symmetrical production-possibility curves, and, at the unintegrated start, have been assumed to be producing identical outputs of Hoofs.Google Scholar

²³ Market towns were included in the exercise, and spurious variation at low levels of both indexes was corrected for by excluding from observation, for any one period, combinations of autumn and spring indexes located within 50 index points of the (100, 100) no-seasonality combination. Instead of using the simple ratio between spring and autumn marriages, which varies between zero and 1 on the high autumn side, and 1 and infinity on the high spring side, the ratio was transformed into the angle formed by the observation with reference to a base at the no-seasonality combination of 100, 100. Thus the corresponding ranges for the transformed ratio become c.0° to 45° and 45° to c.90° (the measures exceed 90° and fall short of 0° when the autumn or spring indexes, respectively, drop below 100. I don't think in radians, the values returned by the statistical software for the arctangent of the spring/autumn ratio, so the results have been expressed in degrees.For this exercise only, a cutoff of 48 marriages was employed (see Appendix).Google Scholar

²⁴ Figure 2 suggests three other arbitrary regions. Region 3, the southeast, is bimodal (caused in large part by strong differences in specializations between arabic East Kent and the rest of the region); its bimodality is at its strongest in periods 8 and 10. The largest region (4) includes much of the Midlands and the East, shows the extreme diversity that would be expected; the distribution begins loosely arabic, with a long pastoral tail, and becomes very bimodal by the last two periods. Region 5, the North, is represented by too few parishes to read much from its diagram except a general move in the pastoral direction.Google Scholar

²⁵ Wrigley, E. A., “Urban Growth and Agricultural Change in Early Modern England,” Journal of Interdisciplinaiy History, 15 (Spring 1985).Google Scholar

²⁶ See, for examples, McCloskey, D. N., The Applied Theory of Price (New York, 1982), pp. 427–32.Google Scholar

²⁷ The zone P₁W in Figure 7 would be the “empty land” in North American applications; such a conception is even less tenable applied to old, settled England.Google Scholar

²⁸ The local zone P₁W for sixteenth-century Kent, Sussex, and East Anglia lies under the Channel and the North Sea.Google Scholar

²⁹ Wrigley, “Urban Growth”; Chartres, J. A., “Food Supply and Internal Trade” (paper presented at the Conference on the Economic and Social History of London, 1500–1700,London,1981), p. 16.Google Scholar

³⁰ vnes, Jan de, The Economy of Europe in an Age of Crisis, 1600–1750 (Cambridge, 1976), pp. 176–209;Google ScholarThirsk, Joan, Economic Policy and Projects: The Development of a Consumer Society in Early Modern England (Oxford, 1978).Google Scholar

³¹ After three months' journey from East Anglia, the geese won by two days. Fussell, G. E. and Goodman, Constance, “Eighteenth Century Traffic in Livestock,” Economic History, 3 (1936), p.235.Google Scholar

³² It is most rewarding that the findings lend support to the concentration on the later seventeenth and early eighteenth centuries of that classic of economic history, Westerfield's, R. B.Middlemen in English Business, Especially between 1660 and 1760, Transactions of the Connecticut Academy of Arts and Sciences, vol. 19 (New Haven, 1915),Google Scholarif not to his characterization of the whole period as one of “meager local specialization” (p. 130).Google Scholar

³³ Rainfall was gauged from the modem Ordnance Survey Ten-Mile sheets, Yearly Averages, 1916–50 (London, 1967); there is no early-modem substitute, but at least modern rain is less likely to have been influenced by past English agricultural practice than are modem soil types.Google Scholar

³⁴ In a simple set of linear regressions of mean rain (in inches) on mean distance from London (in kilometers, to continue the metric mélange of the figure), the slope for the “arabic” means is 0.035 (significant at 0.0033 level of error), that for all “pastoral” observations is –0.003 (wrong sign, and not significant, 0.10 level of error), but for both “arabic” and “pastoral” the means is 0.057 (significant at 0.0001 level of error). The R-squares for all but the last equation are derisory; that for the last (all means) is 0.74.Google Scholar

³⁵ To the extent that marriage seasonality might become “customary,” a tacit rule guiding behavior, a lag might be expected between changes in the pattern of work and changes in the pattern of marriages. I do not expect to find a lag, but have not yet found parallel sources as densely and precisely dated as marriage seasonality against which to test for length of lag.Google Scholar

³⁶ The disturbance of the Interregnum (see footnote 20) to the statistical series is particularly unfortunate with respect to isolating precise causes of change, that is, in nominating causes consistent in time with the changes in seasonality (if indeed the closely complicated tangle can ever be unraveled). This relates particularly to the relation of turning points in seasonality with turning points in real wage, population, and relative price series.Google Scholar

³⁷ The organizational change of enclosure may have increased the flexibility of production choices, thus permitting greater differentiation in practice. The argument that the seventeenth century saw the greatest amount of land enclosed has recently been revived (Wordie, J. R., “The Chronology of English Enclosure, 1500–1914,” Economic History Review, 2nd ser. 36 (11. 1983), pp. 483–505). The relation of enclosure to changes in marriage seasonality, especially swings between autumn and spring seasonality, will soon be investigated, but even if the two are found to be associated, nothing is likely to be learned from the test about the primacy of causes, especially on the question of whether general movements in relative prices or the newly felt imperatives of comparative advantage were more important.CrossRefGoogle Scholar

³⁸ In 1601/40, 66 of the parishes in regions identified by Thirsk as “mixed” typed as “arable,” whereas only 43.5 were “expected” to be, had the marriage seasonality and farming types been independent (using the “tuning” of types relative to period means: see Appendix); only 14 typed as “pastoral” (21.5 expected). Forty-eight parishes in wood pasture and open pasture regions (except for those identified as dairying regions) typed as “pastoral” (32.0 expected); only 13 typed as “amble” (34.8 expected). But parishes in dairying regions were randomly spread over the seasonality types: 6 typed as “arable” (6.6 expected) and 5 as “pastoral” (6.1 expected).Google ScholarThirsk, Joan, “The Farming Regions of England,” in The Agrarian History of England and Wales, vol. IV, 1500–1640, ed. Thirsk, J. (Cambridge, 1967), 4.Google Scholar

³⁹ Thirsk, “Farming Regions,” p. 4.Google Scholar

⁴⁰ See Appendix, however, for a method of spotting tendencies towards specialization.Google Scholar

⁴¹ See, for example, Hey, David G., “A Dual Economy in South Yorkshire,” Agricultural History Review, 17, part 2 (1969), pp. 108–19.Google Scholar

⁴² E. A. Wrigley, “Urban Growth.”Google Scholar

⁴³ Witness the decline of the East Anglian woolens industry; an exception to the process would have been regions characterized by female rural industrial employments such as lace-making and straw-plaiting, strong in certain areas of highly commercial grain production (and, interestingly, resistant to identification by marriage seasons as nonagricultural, almost as if men's work determined the seasonality of marriage—but that, as they say, is another paper).Google Scholar

⁴⁴ As the larger study will show, it is also suited to be fashioned into a diagnostic tool for local history (as Figure 1 suggested) to throw major specializations and the timing of change into better focus; regions can be more cleanly delineated, with a sample larger than the present 404 parishes of the Cambridge Group; the location and volatility of rural industry will also be explored, as will the relation of industrial growth and decline to the increase in agricultural specialization.Google Scholar