6013569ArticlesENGLISH COUNTY POPULATIONS IN THE LATER 18TH Blackwell Publishing Ltd.

Oxford, UK and Malden, USAEHRThe Economic History Review0013-0117Economic History Society 20062007 CENTURYE. A. WRIGLEY

Economic History Review, 60, 1 (2007), pp. 35–69

English county populations in the later eighteenth century1
When directing the first English census John Rickman was intent not only on discovering the size of the population in 1801 but also on tracing past trends both nationally and for individual counties. He returned to the latter investigation on several later occasions, notably in the 1830s. There have been many subsequent attempts to improve upon his national estimates, but his estimates of county totals have continued to be used extensively, either unchanged or slightly modified. Rickman was aware that his estimates were subject to wide margins of error. For the later eighteenth century it is possible to produce new estimates which are probably substantially more accurate, taking advantage of the fact that after Hardwicke’s Act (1753) the registration of marriages in Anglican parish registers, unlike that of baptisms and burials, was virtually complete. They show that the contrast between population growth rates in ‘industrial’ counties and those in which agriculture continued to predominate were significantly more marked than suggested by Rickman’s estimates. The same exercise that produces county estimates also yields hundredal totals, which will in future allow a more refined account of relative growth and stagnation to be made.

lthough much attention has been given in recent years to tracing the history of national population trends, regional and local growth rates have been comparatively neglected. This does not reflect disinterest, but arises either from the apparent lack of new data on which to base any revision of existing estimates, or from a failure to devise a better method of making use of existing data. This article represents an attempt to provide new and more trustworthy estimates of English county populations for the period 1761–1801.2 John Rickman, who directed the taking of the first four censuses, was interested in attempting to trace the past history of the population of Britain no less than in recording the contemporary situation. His work has long remained the starting point for those interested in this topic; indeed, it has


1 The work underlying this essay was made possible by an ESRC grant (RES-000-23-0131), entitled ‘Male occupational change and economic growth in England 1750–1851’. An earlier draft was greatly improved by the comments made by Peter Kitson and Leigh Shaw-Taylor. Later comments from the Review’s anonymous referees were much appreciated. 2 In the early censuses England was taken to include Monmouth, but in this article all data referring to England exclude Monmouth. England therefore consists of 39 counties, or 41 if the three Ridings of Yorkshire are treated as equivalent to counties, as in this article.

© Economic History Society 2006. Published by Blackwell Publishing, 9600 Garsington Road, Oxford OX4 2DQ, UK and 350 Main Street, Malden, MA 02148, USA.



frequently simply been reproduced either in its original form or with slight modification. Perhaps the most widely quoted set of estimates in recent decades has been that published by Deane and Cole almost half a century ago.3 Their county totals mirror those of Rickman, except that they argued that his estimates for Devon and Middlesex were not credible and substituted their own, and that they constrained the county totals to match national population totals that differed from those of Rickman. 4 The absolute county totals therefore differ from those of Rickman, but if a rank order of counties is drawn up reflecting the relative population growth rates between 1751 and 1801 to be found in the two series, the order of counties in the two lists is identical apart from the placing of Middlesex.5 While recognizing the weaknesses of Rickman-based estimates of county populations, most scholars have felt that there was no alternative but to make use of them. 6 Rickman based his estimates on the annual totals of baptisms, burials, and marriages secured as part of the census exercise. These returns, usually referred to as the PRA (Parish Register Abstracts) were the product of enquiries made of all Anglican incumbents. In 1801, in conducting the first census, he had asked for returns of the annual totals of baptisms, burials, and marriages, not only for the immediate past but for a scattering of years earlier in the eighteenth century. These returns settled the argument about whether the population of the country was broadly stationary as Malthus and others believed, or whether it was increasing.7 It was immediately clear not only that numbers were rising, but also that the growth was rapid. In 1802 Rickman published estimates of the scale and timing of growth of both the national population and the counties in the eighteenth century. 8 In each of the next three censuses he required similar returns for the preceding decade, producing thereby continuous annual totals for all three series from 1780 onwards. He had long been interested in attempting to reconstruct population totals for the whole period of parochial registration. He was very conscious of the fact that the 1801 returns had been defective in that many parishes had been missed, and in any case he had no returns for any period before 1700. In the 1830s, therefore, he planned a more ambitious and
Deane and Cole, British economic growth, tab. 24, p. 103 and associated text. Deane and Cole preferred Brownlee’s estimates of national totals. A discussion of the methods employed and the assumptions made by scholars such as Finlaison, Farr, Griffith, and Brownlee, all of whom produced estimates of national population growth in the eighteenth century, may be found in Wrigley and Schofield, Population history of England, app. 5. It should be noted that all these scholars depended on the returns secured by Rickman but corrected his data in different ways. 5 It might be expected that Deane and Cole’s revision of Devon would also have changed the rank order of counties, but Devon was the slowest growing country in both their county series, and therefore left the rank ordering of counties unaffected. 6 For example, Lee, in the course of a stimulating discussion of regional growth patterns, having made it clear that he thought Deane and Cole’s growth estimates unconvincing, added: ‘But they remain the only estimates of eighteenth-century regional population’. Lee, British economy since 1700, p. 127. 7 The nature and the notable duration of the controversy are discussed in Glass, Numbering the people, and well illustrated in his two companion volumes, Population controversy and Development of population statistics. 8 Rickman, Observations on the results of the Population Act.
4 3

© Economic History Society 2006 Economic History Review, 60, 1 (2007)



better grounded exercise. In 1836 he asked incumbents for annual totals of baptisms, burials, and marriages for three-year periods centring on 1570, 1600, 1630, 1670, 1700, and 1750. He hoped that this exercise would enable him to make authoritative estimates of the size of the population at these dates. As one element in the 1831 PRA returns he had asked for information about the date from which the registers survived in each parish, and he was therefore well informed about the proportion of parishes with surviving registers at each date. His method in making use of the new returns was to assume that baptism, burial, and marriage rates had been constant in each county throughout the whole pre-census period; to make separate estimates of population totals at each date based on the three types of event and on this assumption; and to derive best estimates of the population total for each county at each date by averaging the three resulting totals.9 National totals were then obtained by summing the totals for each county.10 There were many potential sources of error in Rickman’s procedures, of which he was well aware. A prolonged constancy in the baptism, burial, and marriage rates was most unlikely. Even if the rates had been constant, registration coverage was not. In particular the spread of nonconformity in the eighteenth century meant that many baptisms were not recorded in Anglican registers. Other deficiencies exaggerated the problem substantially.11 The same was true of burials, though the marked deterioration began at a later date.12 In the seventeenth and early eighteenth centuries marriage registration was at times very seriously defective.13 The mere fact that the population estimates derived from the three series often differed substantially implied that the assumptions underlying Rickman’s method were unsound. Furthermore, although estimates based on events over a threeyear period were less liable to distortion from shocks and random fluctuations than those based on a single year, a serious epidemic or a severe harvest failure might easily result in a misleading estimate even if the basic method had been sound. Yet Rickman’s county estimates, for all their deficiencies, have been very widely used in the absence of any method or body of data which might be expected to produce more reliable results. One

9 There is a fuller account of Rickman’s procedures in Wrigley and Schofield, Population history of England, pp. 572–4. 10 A summary of Rickman’s methods and results, including the text of the letter that he sent to selected incumbents in 1836, may be found in 1841 Census, Preface, pp. 34–7. Rickman himself had died in 1840. His executors passed on the documents containing his calculations to the Home Office. Edmund Phipps and Thomas Vardon, acting on behalf of the first Registrar General, Thomas Henry Lister, who had also died very recently, published the table in the census and briefly described Rickman’s method of calculation. Their text suggests some reserve about Rickman’s results. They wrote: ‘We may, at a future, period, enter into a consideration of the merits of the calculation, deeming it sufficient for the present to state that there is reason for supposing the estimate hereby arrived at to be an approximation to the truth’: ibid., p. 35. 11 Wrigley and Schofield, Population history of England, tab. 4.6, p. 101, and accompanying discussion. 12 Ibid., pp. 89–96. 13 Ibid., app. 4.

© Economic History Society 2006 Economic History Review, 60, 1 (2007)

It is artificially simple. and this article is intended to help set the scene in that connection. In this article their value will be illustrated by considering their significance in the context of the changing occupational structure of England in the later eighteenth century. WRIGLEY reason. 70 per cent of the population depend on agriculture. 1 (2007) . An imaginary population is divided between those who make a living from agriculture and those who depend on other sources of income. 30 per cent on other activities. such as the later eighteenth century. while the comparable figures in the ‘non-rural’ counties are 20 and 80 per cent respectively. 60. See esp. but its relative importance was declining steadily. Agriculture remained by far the largest single industry. while many secondary and tertiary occupations expanded rapidly. indeed. 11. for their continued use is that his method produced county totals as well as national totals. ‘Men on the land’. 332. In ‘rural’ counties.12. Table 1 illustrates the possible scale of change as a result of differential rates of growth. p. At the beginning of the period. especially in a period of unusually rapid population growth. it is possible for there to be major changes in the national occupational structure even though in each component area it remained unchanged. I The type of economic growth taking place in England in the later eighteenth century implies that its occupational structure was changing in sympathy. 70 per cent of the population live in ‘rural’ counties and only 30 per cent in ‘non-rural’ counties. Adult male employment in agriculture grew only slightly in the first half of the nineteenth century and there is little reason to suppose that matters were greatly different in the preceding halfcentury. county population growth rates might be expected to differ substantially as a result. which implies that in 14 There is a discussion of trends in adult male agricultural employment in the early nineteenth century in Wrigley. It is possible at one extreme to imagine a situation in which occupational structure changed everywhere in a roughly similar fashion so that differences in population growth rates in different parts of the country would make little difference to the national picture.38 E. A. Change in the national occupational structure will arise from some combination of structural change in given areas and the rates of growth of population in those localities compared with other areas. whereas most later research has concentrated simply on the estimation of national totals.14 Since many of the occupations in which numbers were rising rapidly were highly concentrated in limited areas. provided that the occupational structure in the component areas differed and that they experienced markedly different population growth rates. The relative importance of these two factors remains to be established for England in the later eighteenth century. tab. At the other extreme. Reliable county population totals are valuable for many purposes. © Economic History Society 2006 Economic History Review.

say. with only 40 per cent of the population making a living from agriculture and 60 per cent from secondary and tertiary occupations. but.5 24 96 120 45 52 60 At time x ‘Rural’ counties 70 ‘Non-rural’ counties 30 Total 100 At time x + 50 ‘Rural’ counties 75 ‘Non-rural’ counties 60 Total 135 At time x + 100 ‘Rural’ counties 80 ‘Non-rural’ counties 120 Total 200 Percentage distribution: agricultural/other Time x Time x + 50 Time x + 100 70/30 20/80 70/30 20/80 the country as a whole 55 per cent of population depends on agriculture. Differential growth rates alone account for the changing balance between agriculture and other occupations. Anglican baptism registers sometimes consistently record the occupation of the father of a child at baptism. Where this is the case the relative size of different occupations can be specified. For example.5 48 70. while the remaining 45 per cent are outside the agricultural sector. while the ‘other’ figure has risen to 52 per cent. An illustration of the possible effect of differential growth rates on overall occupational structure Total population Occupational split: agricultural/other 70/30 20/80 Agricultural population 49 6 55 52. As a result. it makes it clear that the issue of differential regional population growth rates is important if effective use is to be made of many of the available sources that throw light on occupational structure in the past because they often enable the percentage distribution of occupations to be assessed but not the absolute numbers involved. while that in the ‘nonrural’ counties has doubled from 30 to 60. While table 1 fails.5 56 24 80 55 48 40 Other population 21 24 45 22.1780 rising to 4 per cent © Economic History Society 2006 Economic History Review. but the population in the ‘rural’ counties has risen only from 70 to 75.ENGLISH COUNTY POPULATIONS IN THE LATER 18TH CENTURY 39 Table 1. in the country as a whole the share of the agricultural population has fallen to 48 per cent of the total. 1 (2007) . A further 50 years elapses: the ‘rural’ population again increases modestly from 75 to 80. 2 per cent of men were shopkeepers or grocers c. though it may be clear that. and the overall pattern changes further. 60. Fifty years later the population has risen by 35 per cent overall. of course.5 12 64. to do justice to the complexity of change in any historical economy. the ‘non-rural’ population again doubles from 60 to 120. There has been no change in the occupational structure of the population either in the ‘rural’ or the ‘non-rural’ areas.

and for all years from 1780 to 1800. in stark contrast with baptism and burial registration. large city. but for the first half any estimates of population size and trends must remain largely dependent on Rickman unless new estimates are made. From 1754 onwards marriage registration therefore was in principle fully reliable. metropolis. but not the size of the occupied population. © Economic History Society 2006 Economic History Review. and every subsequent tenth year to 1770. For the second half of the period the census provides information about population size and therefore growth trends in different parts of the country. A.17 The only groups exempted from the provisions of the Act were Jews. ‘Effects of population redistribution’. and therefore the totals involved. and coverage deteriorated markedly in the later decades of the century. 16 and much of the available source material is drawn from sources that enable the relative importance of different occupations in a given period to be assessed. 1 (2007) . but Woods showed that mortality might improve significantly in each type of settlement—rural. 16 See note 1 above for details of the grant. knowing that from 1754 marriage registration was virtually complete.40 E. 33. may be unknown. Woods. To the degree that there was a markedly different pattern of population growth between those English counties which were. small town. Hardwicke’s Act of 1753 made it impossible to contract a valid marriage in England unless it was celebrated in an Anglican parish church according to the Prayer Book and recorded in the parish register in a prescribed form. and none for earlier years.—but because there was much more rapid growth of population in the least-healthy settlement types. 60. and members of the royal family. primarily agricultural and those counties in which the bulk of the population from an early date were dependent on manufacturing or service employment. The opportunity to generate improved county estimates arises from a fact with which Rickman himself was familiar. WRIGLEY 50 years later. little change in local occupational patterns may mask important changes in the national picture. 15 An example of the importance of this consideration may be found in Woods’s intriguing discussion of mortality trends in nineteenth-century England. whereas earlier. For much of the century expectation of life at birth changed very little in England as a whole.15 II It was opportune to reconsider county population trends in the later eighteenth century because the Cambridge Group for the History of Population and Social Structure is undertaking an ESRC-funded research project into the changing occupational structure of England in the period 1750–1850. Quakers. and remained. but required marriage totals only for each year from 1754 onwards. local improvement would not be mirrored nationally. the size of the adult male population at the two dates. 1710. etc. 17 26 George II. cap. it was seriously defective. and thus resolves the problem. and that is indeed reflected in his instructions to incumbents at the time of the first census. In 1801 he asked for totals of baptisms and burials for the years 1700. The number of births and deaths substantially exceeded the number of baptisms and burials recorded in the Anglican registers.

Several of the problems stem from the nature of the 1801 census and its defects. further. therefore. is the best of evidence on the subject. This possibility was not overlooked in the past. is not as straightforward as might appear at first sight. it were safe to assume that the marriage rate had remained essentially unchanged in the half-century preceding the 1801 census. They noted that the estimate based on marriage data alone ‘represents the closest approximation to the truth’. p. 60. and of the increase or diminution thereof. because the register of marriages may be deemed perfectly correct. one of which was based on 11-year averages of marriage totals centring on 1801 and 1781. 1 (2007) . it would be a straightforward matter to calculate county population totals covering the whole period following the inception of Hardwicke’s Act. If. British economic growth. therefore. or for a longer period. by basing estimates of earlier population totals on the average frequency of marriages over a block of years centring on a convenient ‘census’ date. and especially by the price of provisions.20 Yet the census (both the Enumeration Abstract and the PRA) was ordered to be printed on 21 December of the same year. the annual total of marriages were known with precision for each county. so that no safe inference concerning the increase or diminution of population can be drawn from the comparison of any single years with each other. Rickman wrote: A great variation in the annual amounts of marriages is caused by the circumstances of the times. cap. Again. Deane and Cole. The whole work of assembling and collating the returns. It is little wonder that the census Rickman. 15.18 Rickman repeated this observation in the prefaces to each of the subsequent censuses which he directed. In his observations on the results of the first census. or at least greatly reduced. 4.19 The use of marriage totals from the PRA as a basis for inferring population trends. Deane and Cole produced estimates of county totals for 1781 using two different methods. 102. for example. however. Observations on the results of the Population Act.ENGLISH COUNTY POPULATIONS IN THE LATER 18TH CENTURY 41 In certain circumstances. The difficulty that there were substantial fluctuations in marriage totals from year to year could be overcome. was accomplished in no more than eight months. but the average amount of the marriages for five years together. 20 41 George III. An act for the taking an account of the population of Great Britain. The Act that made provision for the census specified that Justices of the Peace should require the overseers of the poor within their jurisdictions to appear before them no sooner than 10 April nor later than 30 April 1801 in order to present returns made under the Act for the parishes for which they were responsible. and if. p. 19 18 © Economic History Society 2006 Economic History Review. the use of marriage data may offer a more reliable and exact guide to population trends than estimates based on all three types of events recorded in parish registers in the manner of Rickman. though there was no previous experience to call upon.

to describe how they were attempted. is 1. which is not capable of demonstration in the absence of censuses before 1801. indeed in most years there was exact agreement. county totals do not always equal the sum of the hundred totals and. broken by a few major errors. Suppose that the crude marriage rate was falling in the later eighteenth century. The printed total for Devon in 1792. But credible estimates cannot be produced so simply. there were discrepancies in six years between 1754 and 1800 (1754. A. for example. County population totals are to be estimated for the ‘census’ years 1761. It is essential to attempt to identify the parishes that were missed and to compensate for them.000. 1781. In Derbyshire. In some counties there were only occasional inconsistencies. Disagreement between the two figures was uncommon before 1766 (other than in 1761). except for 1780 (2. Thus if. for example.288). In the PRAs printed in subsequent censuses the agreement was always very close. WRIGLEY volumes are neither internally consistent nor complete.619. In 1790 the printed total exceeded the sum of the county totals by 4. and 1798). for example. The population of Wiltshire. the population total in 1801 were known to be 50. it is convenient next to describe the strategy adopted in generating the new estimates to be presented. the marriages for 1755–67 providing an average for 1761.42 E. These parishes were listed as an addendum to the main return for the county. In order to make clear the necessity for such corrections.000 and the ratio were 0. while in 1797 it fell short of the latter by 2.000.22 This imposes a significant burden of comparison and revision before the best use can be made of the marriage data. 1770. 1771. There was then a further period of generally good agreement. 22 The accuracy of the printed county totals compared with the sum of the constituent hundreds varied greatly.839 persons were living in parishes for which the returns arrived chiefly in January 1802. In Buckinghamshire. probably attributable to poor proofing.000. 1759. 1768. The ratio of the average centred on 1761 to the average for 1801 will then establish the relative size of the population of the county in question. then the 21 In some counties returns arrived only after the main body of material had been sent to the printer. but of this total 20. From 1767 to 1781 in each year the national total fell short of the sum of the county totals by a figure in the range between 1. and 1791.619. © Economic History Society 2006 Economic History Review.000 and 1. 21 In the case of the PRA. III In principle the method employed might appear simple.107. the sum of the hundredal totals failed to match the printed county total in every year except 1755–63 and 1771. There are comparable problems when the national total is compared with the sum of the printed county totals. 1767. They can only be produced if several initial problems are recognized and solved.7 the implied population in 1761 would be 35. 60. for example. in contrast. 1 (2007) . similarly. while the sum of the hundredal totals is 2. and so on).500. and to deal with other sources of potential error. A first difficulty with the method lies in its implicit assumption that the crude marriage rate (CMR) did not change over time. was returned as 185. the printed national totals sometimes differ substantially from the sum of the county figures. the discrepancy was very large indeed. A more serious problem is posed by the fact that returns for several hundred parishes are missing. as an aggrieved note in the census text makes clear. In order to produce these totals. 13-year averages of marriage totals are calculated centring on each of the ‘census’ years and for 1801 (for example. usually because of a misprint. Occasionally.

Moreover. At its nearest point it was 50 miles from the rest of Durham. © Economic History Society 2006 Economic History Review. p. it is convenient to include them within Northumberland. Their stability relative to each other. The rates for 1801.ENGLISH COUNTY POPULATIONS IN THE LATER 18TH CENTURY 43 population in 1761 would be overstated using this method since a given population would have contracted more marriages per head at the earlier date than at the later one.1. tab. however. The national population totals at each date were taken to be those obtained by inverse projection. There is an additional The population totals in question may be found in Wrigley et al. 1 (2007) 23 . This difficulty. Islandshire. The early censuses undercounted children and omitted men serving in the army and navy. These two hundreds continued to be treated as part of Durham in the first five censuses. the opposite would be the case.24 County totals at each census date were increased in the same ratio as the national total to preserve internal consistency. Durham and Northumberland have been ‘modernized’. 1811. If the crude marriage rate were rising. and 1821 are set out in table 2.. Assuming a constant rate would exaggerate the true population total and cause the population growth rate to be underestimated. A6. It should be noted that in table 2 and all later tables. Although a general change in the level of the crude marriage rate can be controlled by using inverse projection totals. to preserve comparability over a longer time period. 24 Wrigley and Schofield. the scale of the undercount varied between censuses. tab. indeed. Population history of England. Rates based upon them are therefore too high. 60. however. thus counteracting any distortion produced by assuming constant crude marriage rates. but. changes in the relative level of the rate between different counties might still present a problem. of course. The issue of the stability of the relative levels of county crude marriage rates over time cannot be addressed directly in the later eighteenth century in the absence of county population totals not themselves derived from marriage data (which would involve circularity of argument). can be tested during the early decades of the nineteenth century since the successive censuses provide both county population totals and marriage totals from the PRA. 614–5. A9. English population history. They are higher than the census totals. 595.7.23 The sum of the county population totals for any date before 1801 can therefore be constrained to match those produced by inverse projection. The county of Durham included the hundreds of Islandshire and Norhamshire. is not insuperable since the equivalent of census population totals exist as a result of the use of the technique of inverse projection in conjunction with the estimates of annual totals of births and deaths produced by an earlier research project of the Cambridge Group (marriage data were not used in the inverse projection exercise). pp. They are based on the average number of marriages over 13-year periods centred on the dates in question. was the northernmost area within Northumberland immediately south of Berwick-onTweed. which were enclaves of territory within Northumberland.

44 7.73 8.43 8.23 CMR 1821 8.37 6.71 9. 60.89 7.11 7.45 8.85 8.01 8. 1821.23 8.14 6.99 7.11 7.64 10.06 9.37 11.82 9.02 7.67 8.29 7.93 7.08 8.09 8.21 5.72 7.29 9.05 7.23 7.90 7.14 6.82 7.70 6.31 8.83 8. 1811.24 7.11 7.55 8.53 8.84 9. 33–4 below.26 8.72 7.47 9.05 7.62 9. and 1831.92 7.50 8. The national totals were taken to be the sum of the totals for the individual counties.20 7.16 7.01 10.19 7.02 6.08 7.77 7.58 8.49 7.86 8.59 Sources: Marriage totals were taken from Parish Register Abstracts of the censuses of 1801.95 7.42 8. E.80 6.83 7.48 8.87 7.05 8. into the following century.64 7.44 8.60 7. WR England 8.13 7. advantage to this decision in that marriage totals in a wide district along the Scottish border were misleadingly low throughout much of the later eighteenth century and.42 7.44 Table 2. A.59 6.42 6. indeed.37 8.90 8.17 7. 1 (2007) .21 7.95 7.73 8.14 7. For derivation of population totals see accompanying text.68 7.96 7.37 8.79 7. This circumstance is discussed below.86 6.55 6.59 9.19 7.29 7.45 9. 1811.05 6.66 8.29 7.06 8.82 7.99 6.83 7.72 6. ER Yorkshire.19 7. and 1821 CMR 1801 CMR 1811 8.29 8.76 8.52 8.38 7.11 8.12 7.49 7. NR Yorkshire.36 7. © Economic History Society 2006 Economic History Review.57 10.49 7.20 Bedfordshire Berkshire Buckinghamshire Cambridgeshire Cheshire Cornwall Cumberland Derbyshire Devon Dorset Durham Essex Gloucestershire Hampshire Herefordshire Hertfordshire Huntingdonshire Kent Lancashire Leicestershire Lincolnshire Middlesex Norfolk Northamptonshire Northumberland Nottinghamshire Oxfordshire Rutland Shropshire Somerset Staffordshire Suffolk Surrey Sussex Warwickshire Westmorland Wiltshire Worcestershire Yorkshire.88 8.79 6.85 7.25 Confining the problem to one county rather 25 See pp.67 8.86 7.98 7.29 6.90 8.57 9.99 8.04 7.29 7.47 5.28 7. WRIGLEY County crude marriage rates in 1801.11 6.23 7.

therefore. The same conclusion is suggested by a different statistic. and the East Riding of Yorkshire. In each case it is demonstrable that the aberrantly high © Economic History Society 2006 Economic History Review. 13. the rank changes in the two decades were in opposite directions (or in one or both decades there was no change of rank). Hampshire.0. The simplest way in which to test the stability of the relative level of the county marriage rates is to produce a rank ordering of the counties for each of the three dates and then test the extent of change from one date to the next. Kent.6. the latter was distorted by the special circumstances of four counties. Westmorland was ranked ninth in 1801 (ordering from the lowest rate to the highest). where the rates for 1801 and 1821 were respectively: 9. In the 1831 census the Liberty of St Peter of York disappeared and its constituent parishes were distributed to hundreds in all three Ridings.000.23 and 8.9 per 1. It therefore moved two places between 1801 and 1811 and one place during the next decade. and Kent. The population totals for the three Ridings in 1821 and 1831 are therefore not directly comparable without adjustment. At the other extreme. These were Devon. For simplicity. in 11 cases the ranking in 1821 was the same as in 1801 or had changed by only one place. were for the most part limited on the evidence afforded by the rates in the early nineteenth century. In three cases this was clearly a ‘wartime’ effect. 60. In only four counties were there big changes in same direction in both decades. though not trivial. In 20 of the 37 counties (excluding the four ‘rogue’ counties). A higher figure in the second decade is to be expected. the rankings of Essex. If these four are excluded from the calculation.1 and 7. because rates changed less in this decade than in its predecessor so that a small absolute change could result in a relatively marked change in ranking. 9. In six counties there was one small and one large change in the same direction.000 respectively) though the rate in 1801 was somewhat higher at 8. the average falls to 3. and 10 over the two decades. Changes in the relative positions of the counties. 1 (2007) . Worcestershire. Consider next the four ‘rogue’ counties. In the remaining 17 counties both changes were in the same direction. Gloucestershire. and sixth in 1821. 11. These averages suggest a relative stability in ranking between counties.73.ENGLISH COUNTY POPULATIONS IN THE LATER 18TH CENTURY 45 than two is a further benefit of the rearrangement. in the next decade it was 4. For example. Devon. The average change in rank number among the 41 counties in the decade 1801–11 was 2.59. seventh in 1811. The national rates in 1811 and 1821 were almost identical (8.2 and 7.8 and 8. therefore. A similar point should be noted in relation to the three Yorkshire Ridings. In each case the crude marriage rate in 1801 was far higher than in 1821. and to facilitate future comparisons. and Sussex changed by 13. but for reasons discussed in the next paragraph.20 per 1. In seven of these cases the changes were both small (where small is taken to be three places or fewer). and 9.81.59. Hampshire. In these four cases the overall change between 1801 and 1821 was marked. I have attempted to provide for the three Ridings rates and totals as the three were constituted after the abolition of the Liberty of St Peter of York.

Almost one half of the rise in the total of marriages over the same period in Hampshire occurred in Portsmouth.46 E. for example. such as Deptford.38 after adjustment. Woolwich. though Hull also had a high marriage rate. since assuming a high marriage rate must depress the size of the population to which it refers. and Kent should probably be lower in 1761 and higher in 1771. For example. Thus of the rise for the county as a whole (580) over three-quarters (438) was accounted for by a wartime boom in Plymouth. This procedure is not an ideal solution since there were earlier wars with comparable. population rose in the remaining 33 counties).772.000.226. WRIGLEY rate at the earlier date was linked to exceptional activity in those ports which were major naval bases. It is plain that assuming that the marriage rate found in 1801 in these four counties also held good for the preceding 40 years would result in a severe overstatement of the rate of growth over the period. but it reduces the distortions which would otherwise arise. the number of marriages in the eight-year period 1756–63 was 3. Broadly similar considerations apply in the other two cases. The earlier wars also caused distortions in the estimated populations in Devon and Kent because other naval bases were affected in the same way as Portsmouth.000 in Portsmouth.38 = 1. The East Riding of Yorkshire displays a similar pattern to that found in the ‘naval’ counties. The ‘true’ totals for Devon.000 in 1801.1683). including Rutland and Huntingdon. These are exceptionally high figures. and Sheerness in the case of Kent.0 per 1. the two smallest.9 per 1. 5 in all three counties the population in 1771 is smaller than in 1761.59 compared with 8.8 per cent higher than if the original 1801 rate had been used (9. if smaller. The 1821 rate was increased in each case by 0. the Hampshire marriage rate in 1821 was 7. Portsmouth in the case of Hampshire.000.000 in Plymouth. In order to avoid this it was assumed that the prevailing rate over the period prior to 1801 should be based on the rate relating to 1821 and therefore on marriage frequencies over the period 1815–27.20). In all four counties the crude marriage rate after the end of the war period was far below its wartime level (see table 2). the amount by which the national rate in 1801 exceeded the rate in 1821 (8. Hampshire. The effects of earlier wars are clearly visible. In Plymouth the average annual number of marriages rose from 461 in the 13-year period centring on 1791 to 869 in the period centring on 1801. the reason is less clear in this case. 27 26 © Economic History Society 2006 Economic History Review.27 The CMR in Hull was 12. The comparable totals for Devon as a whole were 2.79) means that the population of Hampshire at each census date from 1761 to 1791 is 16.9 per l.646 and 3. In the eight following years this total fell to 2.99 per 1.79/8. Using this rate rather than the 1801 rate (9.39 per 1. and 18. 60.26 In all four cases comparison of their rates in 1801 with the rates in neighbouring counties underlines the improbability that the high rates were other than a temporary phenomenon. A. The crude marriage rate for the period centring on 1801 was 19. 1 (2007) . effects on naval bases and marriage totals. In Portsmouth. and Plymouth in the case of Devon. which becomes 8.063. Chatham. It is significant that in tab. a pattern rarely found in other counties (there were five other counties in which population fell in this decade.

60. If the population of a county in 1801 is known and its size at earlier dates is to be estimated from marriage totals for the relevant periods. © Economic History Society 2006 Economic History Review. Just as it is appropriate to base the calculation of the populations of the four counties at earlier ‘censuses’ on a lowered crude marriage rate. say. In a county where there had previously been a substantial number of missing parishes. the ratio of 500 to 250 is the same as that between 400 and 200. 1801 and 1761) the estimated population in 1761 would be exactly the same as if 20% of the population was not covered at each date (yielding totals of 400 and 200). Assessing the necessary corrections is a laborious process. Account must therefore be taken of ‘missing’ parishes in the 1801 PRA.72 per cent. There were few if any parishes missed in the PRA returns of the second census. Two further preliminary operations need to be described. If 100% of the population is covered (yielding totals of events of 500 and 250 respectively for. it might be thought that it is only necessary that the proportion of the population covered should not change. 1 (2007) . cannot be treated in isolation. At first blush it is not clear that this is necessary. Hence the growth in population from 1791 to 1801 would be overstated. it is also appropriate to raise the marriage rates in the other 37 counties marginally when estimating their populations over the four preceding decades. it was essential to try to identify the scale of omissions from the 1801 PRA for each county. since the second group would include a period of fuller coverage.28 But it must not be forgotten that the process involves comparing marriage totals for two 13-year periods. This implies that the rates in the other 37 counties should be increased sufficiently to reduce their populations at each ‘census’ from 1761 to 1791 by 1. keyed to the 1801 population total.ENGLISH COUNTY POPULATIONS IN THE LATER 18TH CENTURY 47 The four counties. thereby offsetting the increase in the populations of the four counties at these dates brought about by the substantial reductions in their crude marriage rates just described. of which one will be 1795–1807. It is reasonable to suppose that the excess of marriages occurring in these counties at the turn of the century would have taken place elsewhere in England but for the distortions produced by wartime conditions. however. This period includes data taken from the 1811 PRA for the years 1801–7 inclusive. The scale of the adjustment should be such as to produce the same national population totals as would have resulted if the 1801 rates had been used for every county. calculate the size of the 28 As an example. adding uncorrected totals for the years 1795–1800 to the totals taken from the 1811 census would cause the increase in marriages between the 1785–97 period and the 1795–1807 period to be overestimated. The 1801 PRA listed the parishes in each hundred from which a return had been obtained and which had been incorporated in the hundred totals at the time that the text was sent to the printer. and not that coverage should be complete because the relative change would be identical in the two cases. This suggests a straightforward solution—to compare the list of parishes from which a return had been obtained with a list of the parishes in the hundred in question. First.

In the remaining 232 hundreds there was no reason to alter the published totals Provided. that in some cases the published marriage totals over the period 1754–1800 needed to be decreased rather than increased as in the illustration given above. a somewhat higher figure than in the earlier survey. In the present exercise a total of 695 missing parishes was found. See Wrigley and Schofield. Population history of England. p.33 This total is lower than that obtained in the earlier exercise. discussed in the Population history of England. 30 29 © Economic History Society 2006 Economic History Review. ‘Rickman’s parish register returns of 1801 and 1841’. Thus. as when a township is listed within a parish which kept no registers. and the situation was worsened rather than improved when Rickman made an initial attempt to correct matters by a follow-up exercise. The net total of missing parishes is therefore 571. A7. It is reasonable to conclude that his final total of missed parishes in England was 713. but in some instances confusion over the correct naming of a parish complicates judgement. In the event. 1 (2007) .48 E. overstates the extent of the overall difference between the number of parishes and chapelries in which marriage registers were kept and the number of returns made by incumbents. because there were also 124 cases in which a registering unit is named in the list of parishes for a given hundred. 621. The range and severity of the problems involved. 31 Ibid. but probably agrees with it fairly closely since there were a significant number of chapelries that maintained baptism and burial registers. if in a given hundred. 32 Ceteris paribus it might be expected that the total would be lower rather than higher because of the restriction of the present exercise to parishes in which marriage registers were kept in the later eighteenth century. A. who was keenly interested in trying to establish how many places had been missed in the 1801 PRA returns and the follow-up exercise. 33 Rickman. Population history of England. were listed which kept no registers themselves.70 per cent (100/92 × 100 = 108. of course. containing 8 per cent of the total population of the hundred in 1801. however. WRIGLEY missing population as a fraction of the total for the hundred.and post-1800 totals were to be on a comparable basis.30 It was then estimated that a total of 632 parishes were missed. Wrigley and Schofield. 7.32 This total. but which belonged to a different hundred. and inflate the PRA marriage totals appropriately. made further enquiries in 1811. but where no marriages were celebrated. if the pre. app. There were many chapelries that kept baptism and burial registers but where no marriages were celebrated. Precision is beyond reach. p. The listing of parishes from which a return had been secured proved frequently at fault.31 Any such estimate can only be approximate. need not be rehearsed again here. the total of marriages for the whole period 1754–1800 would need to be increased by 8.10. hundredal totals were increased in a total of 307 cases but decreased in a further 63 cases. 60. of course.70). that all the parishes in question maintained marriage registers. 601. for example. but where the parish as a whole may have been intended.29 Unfortunately no such simple solution is feasible because of the shortcomings of the 1801 census. tab. Some places.. The misplacement of parishes in the ‘wrong’ hundred means. five parishes were missing from the returns.

For example. 6. Combining a crude marriage rate with a county marriage total yields an estimated population total for each ‘census’ date. the implied total in 1761.035. is 51. and 100. 35 34 © Economic History Society 2006 Economic History Review. and use the cumulation of these county totals to produce national totals. It is worth noting that this conclusion is strongly underwritten by earlier work carried out at the Cambridge Group. The principle followed in this exercise was to start with the published totals for individual hundreds. the total number of marriages in Bedfordshire.58. 7. or an annual average total of 450.623. after adjustment to offset the special treatment of the four counties whose marriage rates were aberrantly high in 1801.944. build county totals from them.847.517.678. and 7. A lower figure for 1761 is to be expected. in others there is room for doubt.358. summarized in table 3.310. 450 × (1.145.9828 = 51.224.845. In the great majority of cases the dates of the period of defective registration show that the problem was confined to the early years of registration. The Bedfordshire total for 1761 of 51. The second preliminary operation can be described very briefly.34 In some instances the reasons for an inconsistency may be clear. It is very probably because. Totals built up in this way were privileged over published county and national totals unless there was a compelling reason to do otherwise.000 (table 2).310. See p.338. 100. The very close agreement between the two sets of national totals listed in the last paragraph implies.38.703. of course.55 per 1.36 Final county population totals were obtained using the ratios between the two sets of national totals.703 derived from marriage data. that the crude marriage rate varied surprisingly little in England over the 40-year period in question. The comparable totals from inverse projection were: 6. for example.145. The substantial inconsistencies between the published national totals of marriages and the totals obtained by cumulating the county totals is mirrored by similar inconsistencies between the published county totals and the sum of the totals for their component hundreds. coverage was defective in a significant number of parishes in the early years of the new régime.25). 36 The footnotes refer to a total of more than 230 parishes in which returns were defective.078 to produce a final figure of 53. as the footnotes published in the PRA for 1801 make clear. The national totals for 1761 to 1791 produced in this fashion were as follows: 6. except in 1761 (the percentages are 97.338/6. 100.093. after correction for missing parishes.139.865.ENGLISH COUNTY POPULATIONS IN THE LATER 18TH CENTURY 49 since all the parishes in the hundred had made returns and there were no ‘surplus’ parishes. and 7.000 /8.55) × 0. 60. 7.703. Expressing the former series as a percentage of the latter demonstrates very close agreement.078.26. 6. centring on 1761. is increased in the ratio 6. in 1755–67 was 5. Marriage totals were See note 22 above.206. Since the crude marriage rate in the county in 1801 (based on marriages from 1795–1807) was 8. 13 above for explanation of the final adjustment figure.662.35 Similar calculations for each county enable a national total to be built up. 1 (2007) .

3 to 8.015 0.310. abstracted from a total of 404 English parishes.794 67.000 (see tab.50 Table 3. WRIGLEY Crude marriage rates based on totals taken from a sample of 404 English parishes Average annual total of marriages Population 6. For comment see associated text.3. Table 4 shows marriage totals for the successive 13-year periods centring on ‘census’ dates for each of the hundreds within Bedfordshire.246 73. A. it is a straightforward matter to produce estimated national rates based on events recorded over the same 13-year periods as were used when making PRAbased calculations.58 8.358 7.59a 1.50 Ratio of rate in col. 2). Willey’s original totals were raised by 24 per cent.002 1. Stodden.991 1755–67 1765–77 1775–87 1785–97 1795–1807 54. 496–502. 1 (2007) 37 . pp. but the totals for Barford. The second panel shows the totals after making adjustments to reflect the proportion of the population of the hundred that lived in parishes for which no PRA return was included in the 1801 census. The rates are remarkably stable and.71 8.697 Sources: Marriages: Wrigley and Schofield.999 0. 60.. Population totals: Wrigley et al.246 57. The top panel gives the ‘raw’ totals. this comparison tends to confirm the accuracy both of the PRA-based totals and those derived from the 404 parish sample.623. and then inflated to produce ‘national’ totals. in the case of Willey substantially. 614–5. Willey.57 8. Since estimates were made for each year. they are uncannily close to the national PRAbased rate for 1801. © Economic History Society 2006 Economic History Review.999 0.37 IV It may be helpful to give an illustration of the successive steps involved in generating county marriage data and estimating county and hundredal population totals. as will be clear from the ratios in the final column.845.669 61.139 7. Population history of England.671.59 per 1. tab.206. Note: a The national rate in 1801 was 8. tab. E. reflecting the proportion of the population of the hundred living in parishes from which no return was The comparison also provides new and important evidence that the re-weighting of the parishes in the 404 parish sample on which the reconstruction of national population trends in the Population history of England was based captures the national pattern accurately. and Bedford are increased in the second panel.60 8.678 8. A9. English population history. The totals were re-weighted to offset the fact that they were not a random sample. Since it is highly improbable that such close agreement between the two series of totals and rates could have arisen by chance.439 Crude marriage rate per 1.000 8. pp.1. In most cases no change was necessary because all the parishes made returns.338 6. A2.

704 3.421 11.013 63.532 6.979 56.107 6.449 7.576 6.532 8.246 6.050 3.178 4.332 6.704 3.006 5.449 3.535 10.351 6.286 6.622 6.256 5.093 54.128 4.834 5.730 9.493 3.798 7.433 4.513 10.730 11.758 6.680 9.264 8.056 5.466 62.423 3.774 6.542 66.926 4.152 3.767 7.079 1.929 6.768 7.718 4.916 791 852 819 907 1112 1.516 7.695 3.298 52.704 3.461 59.615 14.847 6. totals constrained by inverse projection 349 370 393 454 436 572 560 616 590 617 382 405 430 496 455 572 560 616 590 617 4.067 6.587 3.138 1.238 6.532 Source: The Parish Register Abstract and the Enumeration Abstract of the 1801 Census.593 4.017 6. Notes: Because of the effects of rounding.509 8.916 4.916 4.038 4.925 4.044 7.534 4.152 6.598 3.035 345 367 410 466 421 536 576 582 598 713 440 454 445 498 486 408 451 503 563 562 5.618 7.502 5.851 9.298 Biggleswade Clifton Flitt Manshead Stodden Willey Bedford Redbornestoke Wixamtree Bedfordshire Co.836 57.129 3.839 5.080 6.338 1.485 3.648 4.128 7.138 1.338 1.257 3.632 9.258 3.449 3.451 4.839 10.298 53.460 4.097 3.704 5.548 8.461 4.229 1.587 6.519 6.353 3. 60.069 5.372 66.152 3. An example of the derivation of a series of county population totals for the later eighteenth century: Bedfordshire © Economic History Society 2006 Economic History Review. 1 (2007) Barford 261 317 307 305 317 261 317 307 305 317 3.746 4.557 825 932 982 1.485 3.079 1.035 361 384 429 487 431 665 715 722 742 781 440 454 445 498 486 436 482 538 602 581 4.394 7.532 ENGLISH COUNTY POPULATIONS IN THE LATER 18TH CENTURY 53.335 14.384 4.415 6.129 791 852 819 907 1112 1.298 Panel 1 1755–67 1765–77 1775–87 1785–97 1795–1807 Panel 2 1755–67 1765–77 1775–87 1785–97 1795–1807 Panel 3 1801 totals 1761 1771 1781 1791 1801 Panel 4 1761 1771 1781 1791 1801 3.115 1.836 57.103 3.129 66.965 9.767 6.680 7.Table 4.159 3.152 6.466 62.093 54.115 1.129 3.767 3. 51 .822 4.680 10.284 12. the county total may not always equal the sum of the hundredal totals.229 1.818 4.128 6.273 4.588 5.839 14.252 7.588 12.129 66.589 3.522 6.325 3.557 825 932 982 1.493 3.642 6.056 3.564 3.163 3.

consistent. The fourth panel completes the process. The county totals themselves are set out in table 5. for example.847. which have been constrained to make their sum equal the inverse projection national totals. tab. As an example. 41 3.152. 39 38 © Economic History Society 2006 Economic History Review. WRIGLEY obtained. quoted above p.39 The national population total for 1801 published in the census was 8. Hundredal totals are cumulated to produce the county totals shown in the final column of the third panel. 40 On a later occasion I hope to publish a revised. are shown in bold. depends on a process that moves up from hundredal marriage totals to the derivation of county totals consistent with the national totals produced by inverse As. The subsequent inverse projection exercise produced a gratifying similar total for 1801. The population of Bedfordshire in 1801 is recorded in the census as 63. or from 8.439 (Wrigley et al. It is essential that the final hundredal totals should sum to the appropriate county total to ensure consistency throughout the system. but this total needs to be slightly corrected for several errors. It suggested that the published national total in the 1801 census should be increased by 4. independent of each other.836/56.589.285. Therefore the Barford total in panel 4 is 3. tab. The two exercises were.. It is of interest to note that when the work which gave rise to the Population history of England was in train. 38 In the third panel the process is taken a step further.461 = 3. of course. A6.485. Therefore the census total for Bedfordshire was increased by the ratio between the latter total and the former. in 1801 455.152 = 3.671. 5. 595). 40 whereas the inverse projection figure was 8. 15.461. pp. 60. therefore.49%.265 (Wrigley and Schofield. In the final column the county totals. Population history of England. the total for 1755–67.41 The sequence of steps which yields an estimate of the population of a county for each ‘census’ date from 1761 to 1801. the population in 1761 is taken as 382/455 × 4.1. English population history.852 to 8.7.695 × 54.52 E.836 to 56. resulting in a revised total of 8. A9. The county marriage totals in panel 2 were those which were used in estimating the population of the county at each ‘census’ date. and integrated series of population data taken from the first six censuses. This total is unsatisfactory because of the undercount of young children and the exclusion of men in the armed forces. while the population in 1801 was 4.285. the panel 3 total for Barford in 1771 is 3. and the ratio between the panel county total and the final county total for that year is 54.298 in 1801. 1 (2007) . This mirrors the assumption made when calculating county totals that the marriage rate in any given hundred is unlikely to have varied greatly over the period.671. 614–5). A. even though in some instances there clearly were substantial differences in the marriage rate between hundreds in the same county. resulting in a revised total of 66. Therefore. The top line of panel 3 lists the population of each hundred in 1801.589.695.531. 8. They in turn provide the basis for constraining the hundredal totals in the third panel to conform to the county total. The size of each hundred at each earlier ‘census’ date is taken to be captured by the ratio of the marriage total at that date relative to the total in 1801.658. Thus the annual average marriage total for Barford in 1761 was 382. p.439.852.291.393. an estimate of the impact of these factors was made.

Hundred-based descriptions of both demographic and occupational change will be published in due course. It is important to produce hundredal totals in parallel with the county totals because within counties there were always smaller areas whose economic and demographic histories varied greatly. expressed as percentages.671. it is convenient to analyse the contrasting experience of different counties by considering them in three primary groups. engineering. and the complex of metal working. moving down from national and county totals to provide estimates of hundredal populations for each county. seven in all. Growth was often vigorous in the towns within a county even though it was sluggish in its rural hundreds. 5. the West Riding. © Economic History Society 2006 Economic History Review. which is paralleled by figure 1. the framework knitting and lace manufacturing areas in Nottinghamshire. the Staffordshire potteries. At the other extreme there was a small group of counties in which growth exceeded 50 per cent. but which also reverses direction. units with greater internal homogeneity can readily be produced. for example. included towns or boroughs that are treated as separate units in the PRA and whose population history can be reconstructed in the same way. while 22 counties grew by less than 25 per cent. and these in turn are shown sorted in ascending order. see tab.4 per cent. Details are set out in table 6. Table 5 shows the county populations for each ‘census’ date from 1761 to 1801. Over this period the national population increased by 37.ENGLISH COUNTY POPULATIONS IN THE LATER 18TH CENTURY 53 projection. The first group consists of three counties in which growth was rapid because of the dominant influence of the expansion of London. In order to bring out some of the implications of this new set of population estimates. V The focus of attention can now be switched from the logic by which county and hundredal totals were calculated to considering the results of the exercise. and smallwares manufacture in and around Birmingham. comprising 24 counties in all. and Derbyshire.338 to 8. By using hundreds as the building blocks. It also shows the ratio of the latter to the former. and ‘agricultural’ groups had been identified. 1 (2007) . 60. with Wiltshire the slowest and Lancashire the fastest growing county. Cheshire. or combine elements from two or more counties. presenting the data in map form. The third group consists of a massive wedge 42 From 6.310. Many counties. which caused rapid growth in Warwickshire. ‘industrial’. but are outside the scope of this article. plus a fourth residual group consisting of those that remained when the ‘London’. 42 Only 12 counties matched or exceeded this level of increase. Such units may either represent a subdivision of a county.439. The second group includes the bulk of the areas in which manufacturing industry was growing fastest: the textile districts and other centres of industrial expansion in Lancashire.

709 101.325 66.895 122.686 129.4 147.9 106.4 119.718 703.497 158.882 95.367 93.113 124.703 133.440 200.1 111.1 120.0 120.926 214.618 429.262 119.289 321.0 120.537 196.360 301.7 233.717 120.041 218.4 108.3 133.5 121.402 118.5 118.874 104.268 107.888 38.9 117.850 110.974 204.812 262.301 229.427 243.8 142.864 104.389 34.017 200.54 Table 5.298 115.071 112.695 326.129 109.342 83.383 160.720 93.2 130.362 113.5 113.836 101.862 122.2 124.038 142.179 96. WRIGLEY Bedfordshire Berkshire Buckinghamshire Cambridgeshire Cheshire Cornwall Cumberland Derbyshire Devon Dorset Durham Essex Gloucestershire Hampshire Herefordshire Hertfordshire Huntingdonshire Kent Lancashire Leicestershire Lincolnshire 53.9 118.9 113.3 122.703 176.603 157.6 117.0 120.779 202.687 148.0 137.652 97.267 81.048 39.554 233.855 99.9 111.4 111.705 294.814 54.845 214.0 121.525 358.936 80.8 119.716 236.3 113.466 106.646 140.755 209.0 Wiltshire Hertfordshire Rutland Northamptonshire Huntingdonshire Norfolk Berkshire Herefordshire Buckinghamshire Essex Westmorland Cambridgeshire Oxfordshire Worcestershire Yorkshire.829 37.9 115.3 109.884 62.868 35. .949 78.028 181.601 168.662 182.422 95.845 298.9 118.347 95.4 119.488 210. A.467 91.944 534.4 127.648 84. 60.079 103.407 107. 1 (2007) Sources: see text.255 215.926 89.9 106.807 133.619 259.571 177.593 279.6 120.1 136.115 97.576 166.3 113.783 188.8 © Economic History Society 2006 Economic History Review.210 328.902 169.902 90.657 116.2 148.555 136.093 101.402 127.243 193.370 235.3 115.159 89. 1771 1781 1791 1801 Ratio 1801/1761 Counties in rank order Population growth in the English counties 1761–1801 Ratio 1801/1761 County 1761 E.9 122.848 140. NR Lincolnshire Devon Northumberland Dorset Gloucestershire Suffolk 105.278 102.042 107.145 229.201 353.108 57.939 95.588 96.967 205.

609 163.676 36.342 182.6 136.4 Bedfordshire Somerset Shropshire Leicestershire Hampshire Durham Kent Cumberland Cheshire Yorkshire.183 8.1 161.316 167.108 262.942 7.166 274.667 6.766 154.292 146.745 128.219 154.475 591.055 15.368 238.2 157.588 165.897 137.4 144.0 156.345 155.216 99. WR Surrey Lancashire 1761 1771 1781 1791 1801 Ratio 1801/1761 Counties in rank order Ratio 1801/1761 124.172 199.427 104. WR England 574.421 130.073 281.702 194.113 180.416 263.798 148.106 134.766 115.993 41.544 7.354 15.845.5 119.610 217.2 165.433 256.053 178.524 194.426 16.800 141.531 152. ER Yorkshire.050 38.9 119.439 148.5 109.0 160.169 141.646 108.7 124.606 109.757 106.540 201.2 142.0 118.0 148.412 123.232 96.3 137.139 776.623.3 127.984 189.1 130.729 43.010 97.475 397.615 285.794 174. 55 .689 139.9 120.282 133. 60.610 524.4 144.232 148.303 234.178 456.143 146.9 119.105 178.319 192.3 126.3 126.5 105.075 183. 1 (2007) County 652.406 103.275 15.105 194.358 692.711 286.7 137.293 250.9 156.1 233.111 220.174 159.178 272. ER Cornwall Derbyshire Middlesex Staffordshire Warwickshire Nottinghamshire Sussex Yorkshire.347 156.727 183.915 126.859 141.0 148.163 102.002 6.817 131.5 147.791 253.206.4 108.0 126.148 98.268 37.467 125.9 161.8 168.671.310.123 147.976 138.6 165.503 134.781 114.179 175.643 17.883 218.348 357.544 227.7 157.9 111.064 158.5 160.3 133.4 Middlesex Norfolk Northamptonshire Northumberland Nottinghamshire Oxfordshire Rutland Shropshire Somerset Staffordshire Suffolk Surrey Sussex Warwickshire Westmorland Wiltshire Worcestershire Yorkshire.633 91.382 144.563 127.338 ENGLISH COUNTY POPULATIONS IN THE LATER 18TH CENTURY Sources: see text.© Economic History Society 2006 Economic History Review.370 166.386 138.578 176.678 855.3 168.101 215.9 126. NR Yorkshire.

7 0.9 37.533 79.338 321.179 3.7 11.212 53.3 0.412 18.412 1.9 6.9 26.224 8.9 0. A.6 2.9 18. WRIGLEY Kent Middlesex Surrey London group Cheshire Derbyshire Lancashire Nottinghamshire Staffordshire Warwickshire Yorkshire.9 4.5 2.093 101.2 0.071 112.3 1.0 3.9 15.467 125.420 66.1 20.0 11.4 60.5 9.6 1.4 2.9 56.5 0.278.2 3.2 16.9 2.3 42.7 9.3 24.9 19.981.3 1.5 2.671.9 8.209 13.4 11.8 27.676 357.406 977.303.316 183.6 0.5 2.667 1.367 93.537 168.8 1.8 5.2 1.182 113.8 0.757 10.4 48.3 9.406 648.8 26.5 2.048 39.439 86.101 36.503 59.310.4 8.2 22.525 703.3 0.949 78.232 159.2 2.289 218.5 3.407 91.204 13.3 17.8 20.111 217.0 65.483 15.8 24.2 5. 5 .8 17.5 2.1 1.370 1.0 2.7 15.360 574.113 285.807 113.2 1.1 2.7 9.299 29.382 96.8 4.555 146.8 2.418 14.5 100.370 181.964 481.3 1.814 256.0 133.926 95.0 3.105 220.055 15.8 37.7 0.886 6. 1 (2007) Source: see tab.781 250.1 2.868 35.588 96.298 115.200 140.0 © Economic History Society 2006 Economic History Review.332.578 138.5 1.588 1.646 200.729 591.7 48.073 194.1 1.053 233.6 1.0 1.3 1.5 10.1 2.2 1.3 0.794 114.5 0.433 167.4 2.183 2.718 855.0 27.8 1.8 3.5 1. Percentage of English total Population 1801 Increase 1761–1801 Percentage increase Population 1761 Population 1801 Differential growth rates in three county groups Population increase 1761–1801 Population 1761 E.56 Table 6.092 2.440 236.897 137.8 2.429 12.7 2.919 36.4 3.800 176.5 34.5 6.615 281.953.697.703 200.3 15.241 6.9 68.1 49.6 0.458.681 402.9 41.306 43.3 74.5 0.143 1.516 975.845 301.983 2.7 20.794 35.0 100.8 1.731 255.3 0.6 1.9 3.549 90.7 1.812 102.730 54.361.643 17.7 57.9 13.4 3.686 2. 60.9 4.4 100. WR Industrial group Bedfordshire Berkshire Buckinghamshire Cambridgeshire Essex Hertfordshire Huntingdonshire Lincolnshire Norfolk Northamptonshire Oxfordshire Rutland Suffolk Wiltshire Agricultural group Rest of England England 235.5 0.357 281.148 55.2 0.3 2.

60.9 per cent.9 per cent. 2 = growth in the range 10–19. 14. and London Note: The number shown for each county identifies the scale of population growth over the period 1761–1801: 1 = growth in the range 0–9.ENGLISH COUNTY POPULATIONS IN THE LATER 18TH CENTURY 57 Agricultural group Industrial group 0 20 40 80 km 120 160 London group Figure 1.9 per cent. 1 (2007) . © Economic History Society 2006 Economic History Review. indicates growth in the range 130–139. industrial. Population growth 1761–1801 in three county groups: agricultural. and so on. Thus the figure for Lancashire.

A growth of 15. At the end of the period the comparable figures were 16. therefore. in the industrial group 74.8 per cent. but were also modest when compared with many agricultural areas on the continent at the same period.5 per cent of the national population lived in the London group. Indeed.0 per cent).3 and 22. an illustration of the difference between a market-orientated.8 per cent. and when English agriculture was characterized both by high yields per acre and high output per man. and in the dominantly agricultural group 15. the West Riding of Yorkshire (9. there is a sense in which their share is overstated.8.0 per cent over a 40-year period implies an annual growth rate of only 0.9 per cent). It is worth noting that these rates were not only low by comparison with much of the rest of the country. 38. 60. in the London group of counties the comparable figure was 49. This was. perhaps. More than three-fifths of the total national increase over the period. WRIGLEY of land stretching from Lincolnshire and Essex in the east to Wiltshire in the west. a dramatic exemplification of the extent to which growth was concentrated in a comparatively small group of counties in which agriculture did not figure prominently.7 per cent in the industrial group. their growth rate falls from 15.8 per cent of the total national increase took place in just three counties: Lancashire (17.4 per cent between 1761 and 1801. and 26. capitalist agriculture and peasant economies where subsistence sometimes remained an important objective and sons © Economic History Society 2006 Economic History Review.7 per cent). a feature which is brought out more vividly by considering the shares of each group in the overall growth of population in England as a whole. 26.58 E. There were therefore marked changes in the relative size of the three groups. when some parts of the country were experiencing a strikingly rapid expansion based on secondary and tertiary employments.0 per cent to 13.5 per cent. the size of the population supported by agricultural employment changed little (within the agricultural counties employment outside agriculture was rising faster than employment on the land.4 per cent of the total national increase. so that the purely agricultural growth was modest indeed).9 per cent in the agricultural group. The London group accounted for 20. and Middlesex (11. for the agricultural group 10. At a time when the population of England as a whole was rising faster than that of most other European countries. If the populations of the towns and boroughs for which the PRA provided separate returns in 1801 are subtracted from the populations of this group. 1 (2007) . Indeed it was also concentrated within these two groups.9 per cent). A. 20. Whereas population in England as a whole grew by 37. took place in the first two groups (61. Each group presents a marked contrast with the national average.6 per cent).35 per cent per annum. including much of the best agricultural land in the country and covering 29 per cent of its surface area.0 per cent (in the towns and boroughs within these counties the increase was 31.0 per cent.3 per cent. for the industrial group the comparable figure was 41.3 per cent. At the beginning of the period 15. It is equally striking that the belt of agricultural counties claimed such a modest portion of the total increase.

rates of natural increase were frequently higher in rural than in urban or industrial parishes. Comparisons across space raise as many questions as similar comparisons over time. ‘Nouvelle distribution des populations’. and rural populations did not make their livings exclusively from agriculture.43 The contrast in growth rates between the three groups is far too great for any but a small part of it to be attributable to differences in rates of natural increase. tab. pp. 182–94 (nuptiality). Mortality differentials were greater. all four dioceses were growing briskly in the second half of the eighteenth century. they cannot be used to generate estimated county populations for 1751. but since. and at a time when the urban proportion was falling rather than rising. ‘Population estimates of Belgium’. This was probably true of much of western France. 15. There is an initial difficulty in that. © Economic History Society 2006 Economic History Review. There were nevertheless substantial tracts in which growth was slight or absent. The latter is shown in two versions. How great a contrast is there between the new estimates and earlier calculations? This issue may be approached by comparing the growth patterns arising from the work of Deane and Cole with those produced by the present exercise. The available evidence suggests only modest differences in nuptiality and marital fertility in different kinds of parishes. both in the heartland of Europe and towards its periphery. for example. Ratios representing the population increases over the half-century preceding the first census are shown in table 7 together with those based on the table published by Deane and Cole. in a rough-and-ready way. p. Klep. p. In some cases. ‘La peuplade’. 1. Such examples could be multiplied. English population history. Population estimates are usually subject to wide error margins. as in urban or industrial settings. 164–5. Dupâquier. the dates for which estimates are available seldom coincide. 15. and pp.. Population growth accelerated generally in Europe in the second half of the eighteenth century. tab. even though total populations were rising rapidly. fig. in general. 44 Evidence bearing on this issue may be found in Wrigley et al. however. app. pp. But there were many agricultural areas in which growth was more rapid than in the group of 14 agricultural counties in central southern England. and about the extent of the contrast between England and other countries. which shows how widespread was the fall in the urban percentage throughout continental Europe in the second half of the eighteenth century. In Norway. Population and society in Norway. rapid growth came from a low base because of previous devastation by war. Drake. by assuming that growth in each county in the decade 1751–61 took place at the same pace as in 1761–71 and then constraining the resulting county totals to sum to the 1751 national population total obtained by inverse projection. but would repay further examination. 80.44 The bulk of the difference in growth rates between different areas must be largely attributable to net migration flows.ENGLISH COUNTY POPULATIONS IN THE LATER 18TH CENTURY 59 might be retained on a family holding even though the marginal family member produced less than he consumed. The problem can be overcome.. 1 (2007) . The Belgian provinces afford several examples of this. 60. pp. since the marriage returns are available only from 1754. 505. mortality was greater where population densities were high. See also Bairoch. 268–79 (mortality). the opening date in the tables above. 501–7 (fertility). but Deane and Cole provided no estimates for 1761. In column 4 the ratios shown are taken directly from 43 This issue is far too complex to be pursued here. This ensures that the overall rise in numbers is plausible while preserving the differential growth rates of the counties. a country with only a tiny urban population. Yet a systematic investigation of the topic might reveal much of interest about the relative strength of the influences leading to the release of labour from the land or its retention.

3 175.0 115.2 127.5 130.9 127. 103 © Economic History Society 2006 Economic History Review.9 124.4 125.0 132.7 134.8 172. 60.3 116.5 121.7 131.0 170. A.3 130.3 133.5 Deane and Cole (5) Growth ratio 1801/1751(adjusted) 114.0 132.4 182.1 119.7 135. ER Derbyshire Cheshire Middlesex Sussex Staffordshire Warwickshire Nottinghamshire Surrey Yorkshire.1 110. WR Lancashire (2) Growth ratio 1801/1751 105.9 125.7 125.3 Sources: Tab.3 115.9 123.1 218.5 114.3 128.4 143.5 208.8 160.5 147.7 177. tab.6 183. p.5 169.8 137.7 131.4 125.5 116.8 188.8 119.8 165.2 132.7 117.1 137.8 126.9 143.8 120.2 189. 5.7 129.4 115.9 124.6 118.7 126.2 128. ER Yorkshire.0 163.3 176. WRIGLEY Table 7.4 184.7 (3) County Devon Northumberland Bedfordshire Westmorland Wiltshire Northamptonshire Oxfordshire Cambridgeshire Somerset Durham Gloucestershire Norfolk Huntingdonshire Essex Herefordshire Berkshire Hertfordshire Dorset Buckinghamshire Suffolk Shropshire Lincolnshire Middlesex Rutland Leicestershire Cornwall Cumberland Worcestershire Yorkshire.5 107.8 119.5 127. British economic growth.8 185.4 162.3 137.2 150.4 132.4 130.3 139.8 192.9 126.3 142.7 124.8 116.4 129.5 121.6 145.2 174.7 186.1 131. 1 (2007) .6 168.2 120.1 125.0 126.2 140.1 129.9 133.4 146. 24.1 175.0 159.9 278.4 172.60 E.9 166.3 173.1 151. Deane and Cole.9 125.6 126. NR Derbyshire Hampshire Nottinghamshire Warwickshire Sussex Staffordshire Cheshire Kent Yorkshire.6 162.3 134. WR Surrey Lancashire (4) Growth ratio 1801/1751(original) 115.7 146.7 136.6 164.7 209.9 138. NR Bedfordshire Suffolk Somerset Shropshire Kent Durham Cumberland Cornwall Yorkshire.3 181.1 143.8 124.5 125.1 151.9 119.5 159.3 149.9 175.8 116.2 110.5 128.3 199.1 169. The new estimates of growth ratios 1751–1801 compared with those of Deane and Cole (growth ratios measured by expressing the population of each county in 1801 as a percentage of its population in 1751) New estimates (1) County Rutland Wiltshire Hertfordshire Huntingdonshire Devon Northamptonshire Berkshire Norfolk Buckinghamshire Herefordshire Essex Dorset Cambridgeshire Gloucestershire Hampshire Northumberland Oxfordshire Westmorland Lincolnshire Worcestershire Leicestershire Yorkshire.7 140.3 147.9 127.

Or again. This rehearsal of contrasts between the two lists could be considerably extended.83 in Deane and Cole’s list but 2. and similarly those where agriculture was comparatively unimportant in 1831. In making this calculation to preserve comparability between the two sets of estimates Wales and Monmouth were deleted from the original list of Deane and Cole. 103. Counties that were heavily agricultural early on remained so well into the new century. This can be measured with confidence for the first time in the 1831 census.3 per cent in column 5. as in the ‘industrial’ and ‘London’ groups of counties. 46 45 © Economic History Society 2006 Economic History Review. it seemed sensible to reduce the percentage increases by 464/502. There is a general difference between the two lists. 24. The 10 counties with the lowest growth rates between 1751 and 1801 in the new estimates (table 7) had an average of 50. there are more counties experiencing very limited growth during the half-century in the new list. Hampshire achieves a growth close to two-thirds in Deane and Cole. but by only 29 per cent in the new list. 47 See pp. therefore. Lancashire’s growth. 60. which helps to explain some of the individual contrasts.5 per cent for Devon shown in column 4 is reduced to 14. basing estimates on baptism and burial as well as marriage totals). where the present exercise suggests 6 per cent. sometimes. Thus the increase of 15.45 In the next column they have been slightly reduced.6 per cent of Deane and Cole. rises still further.64 in the new list. and at the other end of the distribution. it is likely that their rank ordering remained fairly stable. it would appear.4 per cent. Perhaps because of the averaging effect produced by the method of estimation employed by Rickman (that is. British economic growth. spectacular even in Deane and Cole’s estimation. Deane and Cole’s estimates imply a rise in national population of 50. inexplicably. because of the assumption made about the prevailing crude marriage rate before 1801 described above: the same point explains the contrast in the position of Kent on the two lists). a marked contrast with the 26 per cent in the new list (chiefly. p. A clue to the nature of the inaccuracies or inconsistencies in the estimates of Deane and Cole may be found in a consideration of the percentages of the adult male population engaged in agriculture. Deane and Cole have Rutland increasing by 40 per cent. of course. Table 8 sets out data from the 1831 census.2 per cent between 1751 and 1801. 1 (2007) . for example.47 Worcestershire rises by 46 per cent on the Deane and Cole list. which enables variations in the prominence of agricultural employment within the male labour force to be appreciated. 11–12 above.46 The comparable rise in the inverse projection estimates is 46. were not heavily agricultural 70 years earlier. Although the absolute county percentages in agriculture declined between the later eighteenth century and 1831. The ratio of the fastest to the slowest growing county is 1. tab. approaching a tripling in the half-century period. The rank order of the two series in table 7 differs substantially. To make the ratio increases taken from the two series directly comparable.ENGLISH COUNTY POPULATIONS IN THE LATER 18TH CENTURY 61 their table.

356 967 1.8.020 58.267 37.096 139. in the latter there is a substantial rise between the first and second figures.2 45. the comparable percentages in the Deane and Cole list were 43.6 48.617 3.1 50.698 15.5 per cent.441 11.873 3.2 39.6 23.624 6.995 17.343 45.781 4.8 per cent.518 857 4.4 55.050 5.498 43.3 45.731 3.374 4.213 14.799 15.229 3.683 26.178 116.239 14.113 20.510 2.8 35.658 2.392 16.556 38.320 9.643 2.6 19.502 739.590 13.5 36.9 54.614 19.930 42.9 47.229 4.139 3.486 19.636 93.679 399 397 2. the next 10 40.121 727 1.644 3.901 1.998 1.675 2.152 2.1.0 50.5 29.2 16.582 39.711 2.683 16. © Economic History Society 2006 Economic History Review.385 23.870 24.376 119.330 1.234 1.221 42.812 79.812 Agriculture: occupiers not employing labour 474 458 453 1.655 313.1 31.714 2.911 14. E.661 Percentage in agriculture 59. 37.078 16.916 45.084 35.700 5.907 14.9 11.WR England 22.2 41.3 45.076 83.023 94.3 29.568 37.656 6.397 16.5 54.1 38.203 56.099 49.687 19.971 19.8 54.5.2 30.0 26.234 74.930 69.160 2.361 30.466 17. and the final 11 24.015 2.361 6.243 2.117 1.059 3.565 67.170 47.832 6.7 Total in agriculture 13. NR Yorkshire.239 1.329 43.967 36.124 15.8 47.9 30.914 4. ER Yorkshire. 1 (2007) .927 24.474 24. and 28.613 2.376 2. the 10 next lowest in growth rates had an average of 43.261 2.793 53.311 14.62 Table 8.642 32.094 16.940 56.4 55.812 33.321 15.334 18.328 2.5 18.348 19.474 95.296 28.761 26. Enumeration Abstract adult males engaged in agriculture in 1831.632 71.376 37.045 6.243 9.626 37.715 78.001 115.5 per cent.839 4.914 47.334 10.5 35.032 3.1 38.505 1.556 101.796 49.039 52.260 1.107 16.448 28.675 12.910 13.649 1.414 458 424 2.743 15. 49.342 34.464 15.167 11.711 29.2 55.6 50.718 1.142 1. WRIGLEY Males aged 20 and over engaged in agriculture in England in 1831 Males twenty years of age Agriculture: occupiers employing labour 1.266 11.949 10.6 47.671 4.636 3.234 20.9 41.737 40. A.774 2.504 35.861 59.544 888 1.8 56.9 3. The former set shows a regular decline.413 21.646 24.396 231.561 3.421 4.910 17.6 55.330 1.496 23.608 3.614 58.387 2.685 1.204 490 2.633 Labourers employed in agriculture 11.234 973.763 23.466 18.242 38.266 4.588 14.010 10.3 28.594 29.617 7.950 7.045 79.802 16.050 Source: 1831 Census.040 16.018 4.527 24.521 93.526 1.152 9.106 Bedfordshire Berkshire Buckinghamshire Cambridgeshire Cheshire Cornwall Cumberland Derbyshire Devon Dorset Durham Essex Gloucestershire Hampshire Herefordshire Hertfordshire Huntingdonshire Kent Lancashire Leicestershire Lincolnshire Middlesex Norfolk Northamptonshire Northumberland Nottinghamshire Oxfordshire Rutland Shropshire Somerset Staffordshire Suffolk Surrey Sussex Warwickshire Westmorland Wiltshire Worcestershire Yorkshire.593 35. 60.435 3.5.272 12.535 358.0 39.846 1.856 18.838 1.056 7.257 3.775 10.054 429 3.173.188 37.666 3.6 23.708 14.

occupational change produced by differential population growth rates in counties with contrasting occupational structures could be substantial. and in the agricultural group 65.ENGLISH COUNTY POPULATIONS IN THE LATER 18TH CENTURY 63 The first represents what might be expected on general grounds. In the London group 12. of course. It is of interest to note. © Economic History Society 2006 Economic History Review. the percentage of the adult male labour force engaged in agriculture would have fallen by 2. To complete the picture. The second does not. while in the rest of England the comparable percentage was 37.4 per cent between 1761 and 1801.0 to 60.5. It alone might well reduce the percentage employed in agriculture by. Table 9 mirrors the grouping of county data in table 6.0 to 39.7 per cent. Over a period as long as a century.0. It is improbable that the counties comprising the second group in the Deane and Cole set were both substantially more heavily agricultural in occupational structure in 1831 and yet grew faster than the first group in their set where agriculture was less dominant. therefore. Assume that in this residual group of counties 45 per cent were in agriculture. The new estimates of county growth rates enable the same issues to be examined in the context of late eighteenth-century England.48 VI This brief review of some of the salient characteristics of population change in the later eighteenth century clears the decks for a reconsideration of the interaction between differential population growth rates and change in occupational structure which was discussed in general terms above and illustrated in table 1.6 per cent of adult males were engaged in agriculture. or. and therefore also to affect totals derived from his work. 5 per cent with a matching rise in the percentage in secondary and tertiary occupations over a hundredyear period. as an illustration of possibilities. that throughout the period 1761 to 1801 the percentage in agriculture in the London group had been 17.6 per cent. All the percentages would in all probability have been higher in the later eighteenth century. since it would produce the same result.4 per cent. from 42. in the industrial group 25. in the agricultural group 52. In marked contrast. and all this without any changes in the corresponding percentages in individual counties. that the national agricultural percentage in 1801 in table 10 closely resembles that Much more could be said about the reasons why Rickman’s method of estimating county totals was likely to produce inaccurate results. and it continued well into the nineteenth century. and in the industrial group 19. 1 (2007) 48 . On the assumptions just described.0. account must be taken of the balance of the population.0. that these were the average percentages during this period.2 per cent of adult males were so employed. while those engaged in non-agricultural occupations would have risen similarly from 58. but it would be out of place to engage in a lengthy discussion in this article. say. incidentally. as may be seen in table 10. 60. England counties exhibited this contrast in the later eighteenth century. Suppose. alternatively.

655 358.910 13.242 19.741 78.2 50.7 59.413 21.007.0 45.221 45. 1 (2007) Percentage of total population in 1761 15.510 2.632 83.0 25.715 79.8 31.170 37.856 24.348 19.907 18.2 17. 3) 2.6 14.023 34.974 13.8 23.971 19.6 17.903 23.4 55.568 37.9 48.2 12.0 Percentage of national adult male labour force in agriculture in 1801 (col.9 3.940 71.6 55.716 973.5 16.626 12.763 38.527 18.3 22.6 16.979 3.9 23.234 181.504 35.018 4.617 7.8 54.326 22. WR Industrial group Bedfordshire Berkshire Buckinghamshire Cambridgeshire Essex Hertfordshire Huntingdonshire Lincolnshire Norfolk Northamptonshire Oxfordshire Rutland Suffolk Wiltshire Agricultural group Rest of England England Source: see tab.5 52.503 377.376 58.683 16.3 55.498 43.099 56.7 26.582 101.004 1. 1 × col.5 54.666 923.2 11.9 54.793 39.916 19.624 42.2 19.050 Table 10.64 Table 9.5 30.361 74.392 16.385 43. 6 © Economic History Society 2006 Economic History Review.334 338.6 5.687 29. Illustration of the possible effect of differential growth rates on the share of agriculture in the occupational structure of England in the later eighteenth century Assumed percentage in agriculture in later eighteenth century London group Industrial group Agricultural group Rest of England England Source: Tab.8 26. E.3 45.4 Percentage of national adult male labour force in agriculture in 1761 (col.321 16.7 42.930 58.272 45.5 34. 60.0 47.535 93. 1 × col.6 29.0 Percentage of total population in 1801 16.6 50.039 648.2 37.6 18.5 56.001 79.565 593.9 29.5 39.521 119. A.7 Kent Middlesex Surrey London group Cheshire Derbyshire Lancashire Nottinghamshire Staffordshire Warwickshire Yorkshire.9 6. WRIGLEY The percentages of the adult male labour force in agriculture in England in 1831 (males aged 20 and over) Total Engaged in agriculture 42.3 47.239 231.5 20.106 Per cent in agriculture 36.0 .0 65.173.178 313.9 37. 8 115. 2) 2.6 15.084 35.

2 per cent after a century. 5. 332. 11. Poverty. For example. and population.2 per cent after 50 years and 20. it is to be expected that many occupations employed a broadly similar percentage of the workforce and expanded at similar rates in all parts of the country. textile employment percentages varied greatly from county to county. the proportion of adult males practising as barristers nationally might still have risen significantly. so that differential local population growth rates would have had little influence on national trends.9. consider an illustrative exercise such as that shown in table 1. assume no change in the percentages employed in textile manufacture in either group of counties. No other industry approached the size of agriculture within the labour force which made agriculture an obvious subject for an illustrative exercise of the sort just undertaken. pp. 51 To take a rather trivial illustration. will make it possible to establish the relative importance of compositional change and local change in altering the national occupational structure. tab. ‘Men on the land’. so that contrasting population growth rates might have a powerful influence on national changes. On these assumptions the national percentage in textile employment would rise from 11 per cent at the beginning of the period to 15. when combined with information about changes in the occupational structure of individual counties. Within tertiary employments. while 20 per cent lived in counties in which 35 per cent of the labour force was similarly employed. and any differences in their occupational percentages by region were muted.12. Differential regional growth rates would make little difference to national patterns in these circumstances. almost a doubling from the initial level. In the textile counties population doubles every 50 years whereas in the other counties growth is much slower.51 Or again. 60. though there was no change in textile occupational percentage in any county. with a 5 per cent growth in each half century. Wrigley. for example. 50 49 © Economic History Society 2006 Economic History Review. The potential importance of compositional change varied greatly depending on whether occupational structure was or was not uniform between regions. but because London was growing more rapidly than the national population as a whole (and barristers were heavily concentrated in the capital). 174–5.ENGLISH COUNTY POPULATIONS IN THE LATER 18TH CENTURY 65 inferred from the data in the early censuses.50 though the same was not true for all kinds of tertiary employment. In contrast. Assume that at the beginning of a period lasting a century 80 per cent of the population lived in counties in which 5 per cent of the labour force worked in textiles. This was true of a wide range of retail trades. tab. occupations such as carpenter or tailor were very widespread throughout the country.49 Information about differential population growth rates. p. As in the earlier exercise. but mutatis mutandis comparable exercises related to other industries would show that substantial change might arise from differential population growth rates alone. The mid-nineteenth century censuses provide abundant evidence of the absence of regional variation in the percentage share of total employment that characterized a wide range of tertiary employments: Wrigley. 1 (2007) . in the secondary sector. for example. it is conceivable that the proportion of men who made a living as barristers did not change significantly or rose only modestly in London. progress.

60. The extent of the contrast between the swiftest and slowest growing parts of the country appears much more vivid when using hundredal data than when using county data. Middlesex. that of the hundreds in the rising industrial areas of England.801 to 49. for example.196 between 1761 and 1801. or by 214 per cent. WRIGLEY VII There were both changes in occupational structure and differences in population growth rates in all localities so that the contributions that each sub-division within the country made to the national whole was affected by both factors.298 between 1761 and 1801. by 97 per cent. Edmonton.1 per cent. in 1761 Lancashire housed only 4. The star performer was the hundred of Salford (here taken to include the town of Manchester). A. Its population grew from 93. the fastest growing after Salford was not be found elsewhere in Lancashire. it is possible to assemble groups of contiguous agricultural hundreds covering substantial areas in which the population was falling rather than rising over the same 40-year period. was a small county in area which housed the bulk of the population of London yet it included four hundreds. and Spelthorne. At the other extreme. To take the most extreme example.8 per cent of population growth down to 1801.The rest of the county also grew rapidly. Where contrasts in local population growth rates were as marked as was the case in England in the later eighteenth century. still more striking contrasts can easily be produced by using hundreds as the units of analysis rather than counties. insights that cannot be secured from county data. or by only 3. however. they are distinct. Analytically. the possibility that in some industries they may have accounted for the bulk of any change in their relative importance in the national occupational structure clearly exists. This article makes little use of the hundredal population estimates that were generated in parallel with the county estimates. Lack of space prevents any extensive demonstration of this point. both generally and in individual industries. because the county was seldom a uniform unit.9 per cent of the national population in 1761. © Economic History Society 2006 Economic History Review. Or again. for example. Nor should it be overlooked that. with 18 per cent of the population of the 14 agricultural counties in 1761 enjoyed a rise in numbers 57 per cent greater than that in the agricultural block. however. They offer. Almost as striking is the fact that the agricultural county group housed 26.8 per cent of the population of England. but not equally quickly. Their combined population rose from 47.767 to 294. yet accounted for only 10. Elthorne. It is intriguing to note. but it is readily illustrated. will be an important element in the current research project designed to identify changes in the occupational structure of England in the period 1750–1850. the population in all parts of Lancashire was growing quickly in this period. Gore.0 per cent of the national growth in population took place in this single county (table 6). yet over the next 40 years 17. Lancashire. whose populations were almost stationary in the later eighteenth century. equivalent to an annual growth rate of almost 3 per cent. but by comparison with Salford its progress was almost staid.66 E. and establishing their relative importance. 1 (2007) .

In the PRA volume of the 1831 census there are rough maps of hundredal boundaries within each county. there is good reason to treat them as part of Northumberland). Both description and analyses benefit markedly.ENGLISH COUNTY POPULATIONS IN THE LATER 18TH CENTURY 67 nor in the West Riding. 60. but in Cheshire. as well as Macclesfield itself. Neighbouring hundreds along the border also appeared to suffer from sharply declining populations.6. It was. Nottinghamshire. Take Islandshire as an example (it will be recalled that.52 Using a marriage-based method of calculation. The hundreds of Glendale and Bambrough. and Berwickon-Tweed. Between 1761 and 1801 the hundred of Macclesfield. 53 See above pp. Using the hundred as the unit of description makes it feasible to construct regional units with a far greater degree of internal homogeneity than is possible when using the county as the unit. I hope. in effect. That the apparent falls are spurious becomes plain when the crude marriage rates for the hundreds are considered. Bambrough. and Durham carries the following text printed along the border: ‘The inhabitants of the border usually resort to Scotland to be married’. especially in Northumberland. The custom of moving across the border to marry was well known at the time. The map covering Cumberland. but the rapidity of change in this part of east Cheshire is cloaked within the larger county unit. and the town of Berwick-on-Tweed resembled Islandshire though their population declines were less extreme (in 1801 their populations. Glendale. Staffordshire. I hope in the future to make much fuller use of the hundredal estimates. a southward extension of Salford hundred. integrating the later eighteenth-century data with those derived from the early censuses. although Islandshire and Norhamshire were hundreds within Durham in this period. come to be regarded as an improvement on their predecessors. Other parts of the county were very slow growing during this period. using the same method of estimation. which included such towns as Stockport. In one area it is certain that estimates based on marriage totals must be wrong: the hundreds bordering Scotland. calculated from the average number of marriages in 1795–1807 and related to the populations in 1801 were 1. 9–11. Dukinfield. Islandshire would appear to have experienced very severe population decline in the later eighteenth century.3. and 2. The rates for Islandshire.8 per 1. Further work may identify good reason to modify some of the estimates.000 respectively. 60.7. VIII The new estimates of county population totals shown in table 5 will. The proximity of Scotland encouraged an increasing number of couples to move north of the border to contract their marriages. Westmorland. 2. and Hyde. Broxton and Wirral. to the point that its population in 1801 was only about 36 per cent of the 1761 total. or Warwickshire. 3. Northumberland. 1 (2007) . but each total should be regarded at best as lying towards the centre of the range of plausible possibilities rather than exact. The two westernmost hundreds. were 51. grew by 140 per cent. and 56 per cent respectively of their 1761 levels). 53 52 © Economic History Society 2006 Economic History Review. showed no growth.

and Tindale both had rather low crude marriage rates (5. 60. Since the purpose of this article was to present the new estimates of county populations in the later eighteenth century and to illustrate one of their uses for economic history. covering the whole period conventionally equated with the industrial revolution. The census provides population data both for the county and the hundred down to 1841.1111/j.0 per 1.55 Long runs of data always offer possibilities that are more restricted when runs are short. their impact on the related county total is limited. they will allow the question of the impact of differential county growth rates on occupational structure to be pursued more effectively. Eskdale hundred had a land border with Scotland. there was marriage ‘leakage’ across the border in these hundreds it appears to have been present throughout the period 1761 to 1801 since the marriage totals suggest population growth in both hundreds. and it might prove necessary to collect new data for substantial periods of time from large numbers of parishes in order to estimate county totals. lower than in hundreds such as Morpeth and Castle. It is. Fortunately. such aberrations appear to have been rare.00355. which had no common border with Scotland.9. If. There is scope for much further work in this vein. even so. © Economic History Society 2006 Economic History Review. while Cumberland hundred faced Scotland across the Solway Firth. however.2006. but escape across the border appears to have been a constant rather than an increasing phenomenon.68 E. The published population totals at all levels from the country to the hundred stand in need of adjustment to offset the effects of under-registration and service in the army and navy. of course. Correction for underregistration would be problematic. the Northumberland estimates should be regarded with some reserve. 1 (2007) . but Cumberland hundred resembles the border hundreds in Northumberland. however.000 in the period centred on 1801). 55 It would. it would involve vastly more effort than the present exercise.1468-0289. Coquetdale. substantial in Tindale. and an attempt was made to limit any distorting impact on the county population totals by treating the most severely affected hundred as paralleling other Northumberland hundreds situated away from the border which they resembled. In Northumberland. I have not moved outside the time frame which they set. In the county of Cumberland.3 per 1. The marriage rate there was 4. A. if feasible at all.000 in 1801 when in the rest of the county the rate was 7.x 54 Other hundreds in Northumberland and Cumberland also bordered on Scotland. In Eskdale the marriage rate suggests that relatively few couples moved north to celebrate their weddings. Coverage was less good in all three series (baptisms. burials. modest in Coquetdale. since population growth was quite brisk between 1761 and 1801 in the new estimates. but. WRIGLEY The populations of these hundreds in 1761 are greatly exaggerated when using the present method for calculating late eighteenth-century population totals.54 Because the populations involved were relatively small. but consistent series covering the whole period 1761 to 1841 are feasible. Cambridge Group for the History of Population and Social Structure Date submitted Revised version submitted Accepted 24 March 2005 28 September 2005 5 December 2005 DOI: 10. be equally. and marriages).9 and 6. an artificial restriction. or perhaps even more valuable to extend the county series further backward in time. In particular.

Smith. in M. R. H. R. VI). VII). P.ENGLISH COUNTY POPULATIONS IN THE LATER 18TH CENTURY 69 Footnote references Bairoch. A. XI). 1811 Census.P. Woods. R.. Enumeration abstract (P. 1541–1871: a reconstruction (1981). I. XV). De la Renaissance à 1789 (Paris. 1 (2007) . 1991).. Enumeration abstract (P. Wrigley. J. S. II. The development of population statistics). The eighteenth century population controversy and the development of census and vital statistics in Britain (Farnborough. English population history from family reconstitution 1580–1837 (Cambridge. Glass. Enumeration abstract. XXXVII. Rickman. ed. M. 193–229. R. W. Histoire de la population française. V. 1986). and XXXVIII).. ‘Population estimates of Belgium by province (1375–1831)’. II.P. Parish register abstract (P.P. Drake.. and Schofield..P. A. and Schofield. Wrigley. Journal of Economic History. Poverty. 2004).. in J. progress. Lee. A. Parish register abstract (P. pp.. P. La révolution démographique 1750–1914 (Paris. The development of population statistics (Farnborough. and Parish register abstract. 1812. pp. E. 60. Histoire des populations de l’Europe. 485–507. supplement to the 1801 Census Reports (reprinted in Glass. 2.P. V. Enumeration abstract (P. 68–81. Klep. E.P. 1833. Dupâquier. Numbering the people. III (P. J. 1962). ed... III (1802). 1843. XV). Bardet and J. eds. ed. Oeppen. C. 1822. Wrigley. Official publications 1801 Census. R. Davies. 645–51. 1802. in L. The population controversy (Farnborough. M. XI). 1831 Census. Poulain et al. J. 1988). Glass. 1822. 1998). The British economy since 1700: a macroeconomic perspective (Cambridge.. XXII). S. pp... P.. Bonfield. Wrigley. 1986).P. vols. 1973). The world we have gained: histories of population and social structure (Oxford. Deane.P. in J-P. M. ‘The effects of population redistribution on the level of mortality in nineteenth-century England and Wales’. Historiens et populations: Liber Amicorum Etienne Hélin (Louvain-la-Neuve. British economic growth 1688–1959: trends and structure (Cambridge. eds. A.. 1821 Census. Observations on the results of the Population Act. ‘Une nouvelle distribution des populations: villes et campagnes’. Dupâquier.. ‘Men on the land and men in the countryside: employment in agriculture in early nineteenth-century England’.. M. and population (Cambridge. and K. The population history of England. pp. XXXVI.. pp. 45 (1985). S. D.. 1969).. A. Population and society in Norway 1735–1865 (Cambridge. 1973). and Cole. D. D. Parish register abstract (P... Dupâquier. © Economic History Society 2006 Economic History Review. V... E. Wrightson. 1973). eds. 41 Geo. 295–336.. Enumeration abstract (P. ‘La peuplade’. 1997). 1802. E. 1841 Census.. 1812. Glass.

Sign up to vote on this title
UsefulNot useful