Natural Population Change: Horsley-Nailsworth, 1775-1851

The Industrial Revolution was accompanied by self-sustaining population growth, the causes of which historical demographers have debated extensively.^[25] Attempts to explain this change have focused on the issue of whether the crucial cause was a rising birth rate or a falling burial rate. Those who have argued for the primacy of the birth rate have emphasized the impact of economic growth on fertility,^[26] while those who have advocated the primacy of the burial rate have regarded economic growth as a by-product of population increase.^[27] Nonconformity, as already noted, has been treated in this debate solely as a factor of underestimation in the Anglican record.^[28]

This section describes local trends in baptisms, births, and burials for the period 1775-1838,^[29] using aggregative analyses of Horsley's Anglican registers,^[30] as well as the birth and burial registers of the Dissenters's chapels at Nailsworth.^[31] With compound-interest growth rates, it is possible to assess the relative importance of births and burials to the overall pattern of population growth and, at the same time, compare the contributions of Anglicans and Dissenters to the composition of the general growth rate in births. Using multiple regression in time series between 1804 and 1838, it is also possible to determine the effects of economic and religious-cultural changes on short-term fluctuations in births. Before reporting these findings, discussion of the sources on which they are based is necessary, since the combined use of Anglican and Nonconformist registers within a local setting is comparatively unusual.^[32]

The Anglican registers of Horsley were chosen because of their consistency and because Horsley itself was more representative of the industrial population than was Avening. Together with Nonconformist registration of births and burials at Shortwood and Forest Green, they represent the best set of demographic records for the industrial population, consonant with the constraints of a local study.^[33] A partial exception was made with respect to marriages. Since the law required Dissenters to marry in the parish church, the Avening registration figures were added to those of

Horsley in order to compensate for the possibility of under-registration.^[34]

Horsley's baptismal registers have been supplemented by baptismal registration at the Nailsworth Episcopal Chapel. Nails-worth baptisms between 1794 and 1812 represented the incidence of underregistration among Anglicans caused by the locality's relative isolation. Using a ratio of Anglican baptisms at Nailsworth to Horsley parochial baptisms, adjustments were made in the figures for the earlier period, 1775-1793.^[35] After 1812, the number of Nailsworth Episcopal baptisms fell dramatically. This change coincided with the introduction of standardized forms to replace the old registration books. The cause was undoubtedly the virtual ending of Nailsworth's isolation, however, as residents had grown accustomed to regular intercourse with the surrounding hinterland. Although the Episcopal Chapel's baptisms have been added to the parochial figures for the 1813-1838 period, the latter may be regarded as substantially representative of all recorded Anglican baptism for the Horsley-Nailsworth region.^[36]

Apart from the problem of deliberate underregistration, a difficulty that normally arises with the use of Anglican registers is the extent to which baptisms are truly representative of births.^[37] For Horsley, baptismal dates appearing in the parochial registers closely approximate the actual date of birth. Between 1786 and 1812, indeed, the incidence of the number of children over one year of age presented for baptism became more evident. However, Horsley's vicar made note of the ages, a practice that emphasized their exceptional character. Moreover, when infants were presented for baptism beyond the normal two-week period from the date of birth, the vicar recorded their actual birth dates. Again, adjustments of yearly totals were made when the baptismal and birth years did not coincide. The intervals between births and baptisms ranged from 2.5 to 41 weeks, and the median interval was 18.5 weeks. The average number of child baptisms was four per year; they represented a very small proportion of mean Anglican baptisms for the period, which stood at fifty-eight.

Nevertheless, the practice of delayed baptism could have served as a source of underregistration in Anglican baptismal and burial records. Wrigley and Schofield have shown, for instance, that

baptism delayed by one month could have seriously affected the number of infant births registered, since infant mortality was often high. They have suggested that between 1775 and 1824 the number of Anglican baptisms at the national level could be inflated by 1.074 for every recorded baptism, and between 1825 and 1875, at the ratio of 1.090; burials could be inflated correspondingly at the ratios of 1.045 and 1.070, respectively.^[38] Accordingly, I have applied these inflators to the data collected in appendixes E and F.

From 1813, following the introduction of standardized registration forms, references to the age at baptism or to the birth date ceased, although there was nothing in principle to prevent the continuation of the practice. If the numbers of delayed baptism were small in the pre-1813 period, they probably dwindled to insignificance thereafter. The remoteness of Horsley's outlying districts was responsible for this practice and, as noted previously, ceased to be a factor in the later period.

Nor had a Baptist influence necessarily contributed to delayed baptism.^[39] It is true that many of the Anglican laity attended the Shortwood Baptist Church as hearers, but the Baptists drew a careful distinction between the status of member and hearer. They made believers' baptism the condition for membershi and prescribed a minimum age of sixteen. There were very few instances of Anglican baptism having occurred so late. Indeed, Baptists were more likely to have influenced Congregationalists, since Dissenters were especially prone to attend each other's services. However, when the Forest Green minister announced his intention to delay the baptism of members' children because of a sudden scruple regarding infant baptism, the church meeting dismissed him soon thereafter.^[40] The members preferred infant baptism, despite the presence of a strong Baptist church, with whose followers they had had regular intercourse. The Anglican laity could hardly have been influenced more greatly.

The Forest Green Congregationalists and the Shortwood Baptists each maintained birth records that provide considerable additions to the Anglican totals. The Congregationalists recorded births and baptisms for the entire period from 1775 to 1836. Baptist births, drawn from a marriage register of zoo Baptist families, were recorded only for the period 1800-1836; therefore, earlier Baptist births were estimated from the church membership roll,

using annual conversions and annual membership losses as independent variables in a regression model:^[41]

Since both registers recorded actual births, the problem of "representativeness," which arose in the case of Anglican baptisms, is not an issue in quite the same way.

Congregationalist registers provided the most continuous series. Not only do they cover a longer period, but both birth and baptism dates were frequently recorded; the residence of the parents and the maidenme of the mother are sometimes included as well. A government official, responsible for collecting them, exaggerated their degree of incompleteness when he noted that "many births and baptisms have never been entered, others but partially."^[42] There was undoubtedly some omission, but his remarks apparently referred only to the one register in which he inscribed them. When all Congregationalist registers are counted, they yield in combination a continuous series in which the only notable defect is the occassional omission of the date of baptism; the date of birth is usually recorded instead. Where the reverse was true, estimates of the birth date were extrapolated from the mean interval between births and baptisms during the decade of the 1820s.^[43] For the entire period, 675 Congregationalist births were recorded, equaling 12.3 percent of all births: that "Dissenters viewed baptism strictly as a religious ordinance," as the government official maintained, is an argument that rather favors more complete registration, as this statistic suggests.

The problem of assessing the degree of underregistration in these sources is difficult, however, because we are dealing with a much smaller population than the Anglican, and one subject to an independent structure of growth. The distinction between member and hearer, already noted, was critical. Hearers usually fluctuated between Church and Chapel,^[44] and were likely to have had their children baptized in the former. Births or baptisms recorded in Dissenters's registers were only those of members' children. Special factors, therefore, could have influenced Nonconformist birth

registration: the age-sex composition of the membership and its potentially irregular growth pattern. Even with a large membership, the number of births might have remained small if the proportion of married members of childbearing age was insufficient. However, membership growth was sufficiently high among Baptists after 1800 to suggest the existence of a normal population, and as we shall see, this growth strongly affected a rapidly rising growth rate in Baptist births.

The Shortwood church roll, from which membership figures have been derived, is especially complete; it records dates of baptism and death, dates and places of emigration, and exclusions. Migration was the most troubling factor since members, prior to their departure, did not always obtain a formal letter of dismissal to a new Baptist church and would therefore remain on the register until the minister sorted out the record. Fortunately, the minister made periodic adjustments so that the document on which analysis has been based appears trustworthy.

As a group, Dissenters' registers offer a reasonably complete picture of Nonconformist demographic trends. In general, Baptist births had the greatest impact; from 1800 to 1836 alone, they numbered 746, or 42 percent of all Nonconformist births during the entire period under study. Congregationalist births, however, increased dramatically from 1820 as a result of the schism that occurred at the Forest Green Church in that year.^[45] Quaker births, which numbered only forty for the period under study, were also added to the Nonconformist totals. If we include the estimates of Baptist births for the 1775-1799 period, the total number of Nonconformist births reaches 1,799, or 32.8 percent of all births recorded between 1775 and 1836. The magnitude of this figure suggests that recorded Nonconformist births, including estimates, may be treated as representative of all such births.

It is possible, then, to compare the growth rates in Nonconformist and Anglican births to determine their importance with respect to the growth rates in all births from 1775-1838. Such a comparison also requires analysis of the varying responses of Anglican, Baptist, and Congregationalist births to economic fluctuations. The dichotomy between sect and denomination is especially relevant in this respect. All religious groups, except the Quakers, became more denominational during the period under study, a change that undermined their cohesiveness as communi-

Fig. 7.
Time-series of All Births, 1775-1838.
Key: Solid Line = All Births/Baptisms
Small Broken Line = Anglican Baptisms
Large Broken Line = Dissenter Births

ties and their abilities to resist exogenous shocks emanating from a changing economy. Yet there were differences in degree among them. Because of a stable competent ministry, the Baptists achieved an optimal balance between the large membership growth issuing from the evangelical revival and the persistence of sectarian practices. Their stronger sense of community immunized

Fig. 8.
Time-series of Nonconformist Births 1775-1838.
Key: Solid Line = All Dissenter Births
Small Broken Line = Baptist Births
Large Broken Line = Congregationalist Births

them more effectively (although never completely) against the negative consequences of economic change.

If we compare the growth rates of Anglican and Nonconformist births, we can see how their importance to the general growth rate in births shifted during the 1775-1838 interval (see figs. 7 and 8 and table 27). Births are plotted in time series in figures 8 and 9,

Fig. 9.
Time-series of Births, Burials and Marriages 1775-1838.
Key: Solid Line = All Births/Baptisms
Small Broken Line = Marriages
Large Broken Line = Burials

and their growth rates appear in table 27. The subperiods, 1775-1799 and 1800-1838, corresponded broadly to the phases of protoindustrialization and industrial takeoff, respectively. The second phase also marked the intensification of the evangelical revival. In the first phase Anglican births grew faster than Dissenters' births but tapered off significantly in the second phase, when Dissenters' births rose rapidly. The rising general growth rate in births over the entire period was due mainly to the growth in Nonconformist births after 1800. The volume of Nonconformist births was perhaps more sensitive to religious revival than to economic flux, while the reverse seems to have been true for Anglicans.

This analysis becomes complicated, however, by the fact that the second phase was a period both of prosperity and contraction, as well as one of structural transformation of the economy. The shift to the factory system and depressions in the cloth trade induced anomie and could be expected to have had a damping effect on population.^[46] Economic growth, although the product of structural transformation, could also bring intervals of prosperity that would stimulate reproduction. Membership growth in Nonconformist churches might have benefited in either case; some joined out of despair and others, from feelings of optimism. We need to examine the effects of short-term fluctuations in economic indices on Anglican and Nonconformist births to obtain a better picture of the causal relations (see table 28).

Multiple-regression analyses accomplished this task by using the residuals of average annual wheat prices^[47] and the original observations of cloth manufacturing output^[48] as one-year, lagged independent variables for the 1804-1838 period. The unstandardized coefficients of both independent variables, as revealed in table 28, proved strong predictors of Anglican baptisms. The negative direction of the coefficient for "Rswheat" (residuals of wheat prices) meant that a decline in price in any given year stimulated reproduction, while a corresponding rise in cloth manufacturing output had a similar effect, and vice versa. By contrast, only cloth manufacturing output significantly affected reproduction among Dissenters, suggesting that as a group they were more proletarian, since the level of output necessarily serves as a surrogate for employment in the industry; the volume of their births, in other words, remained unaffected by agricultural rhythms (see table 29).

TABLE 28. Regression Coefficients: Wheat Prices (Residuals) and Cloth Production on All Baptisms and Births, 1804-1838
							Applies to the whole model
D-var	I-var	df	b	t	Prob > t	B	F	Prob > F		DWa	Ar
Anglican baptism							9.634	0.0006	0.3435	1.41	0.27
	Rswheat	1	-2.65	-2.93	0.0062	-0.41
	Cloth	1	0.000033	3.29	0.0025	0.46
Dissenter births							5.330	0.01	0.2079	1.45	0.24
	Rswheat	1	-0.99	-0.86	0.397	-0.13
	Cloth	1	0.000041	3.16	0.003	0.49
Total Baptist births							10.040	0.0004	0.3540	1.4	0.26
	Rswheat	1	-3.65	-2.17	0.038	-0.30
	Cloth	1	0.000075	3.95	0.0004	0.55
Key : "Rswheat" = residuals of wheat prices; "Cloth" = annual cloth output (in tables 29, 35, and 36 also); R 2 ; b = parameter estimate, or slope; t = t -ratio; Prob >t = probability of obtaining a value greater than t ; B = standardized b ; F = F -ratio; Prob > F = probability of obtaining a value greater than F ; Ar = first-order autocorrelation coefficient.
a The term DW refers to the Durbin-Watson statistic; values over 1.7 indicate no problem with autocorrelation; values between 0.82 and 1.7 lay within an ambiguous zone, and their degrees of significance depend on the size of the autocorrelation, reported next to DW . Values below 0.82 clearly have the problem of autocorrelation. See Wonnacott and Wonnacott, Econometrics , pp. 142 and 147.

TABLE 29. Regression Coefficients: Wheat Prices (Residuals) and Cloth Production on Nonconformist Births, 1804-1838
							Applies to the whole model
D-var	I-var	df	b	t	Prob > t	B	F	Prob > F		DWa	Ar
Baptist births							6.875	0.0034	0.2626	1.76	0.12
	Rswheat	1	-0.242	-0.453	0.6534	-0.06
	Cloth	1	0.000022	3.68	0.0009	0.55
Congregrationalist births							2.006	0.151	0.05	1.4	0.26
	Rswheat	1	-0.63	-0.924	0.362	-0.156
	Cloth	1	0.0000138	1.787	0.08a	0.302
Dissenter births							5.558	0.0087	0.216	1.5	0.27
	Rswheat	1	-0.994	-0.939	0.355	-0.14
	Cloth	1	0.000038	3.21	0.0031	0.49
Note : This analysis is based on the uninflated values for Baptist, Congregationalist, and the total of Dissenters' births, appearing in appendix E. The total of Dissenters' births in previous analyses was inflated to compensate for the possible effects of delayed baptism on infant mortality, principally m order to facilitate comparisons with Anglican baptisms. This inflation factor had no effect, however, on the regressions of the original observations of Baptist and Congregationalist births and, in fact, produced only very marginal increases in their totals. For these reasons, I chose to dispense with the inflation factor in making intra-Nonconformist comparisons.
a Although falling slightly below the 0.05 level, I find a probability of 92.0 percent to be acceptably significant.

Nevertheless, Baptists had a more proletarian following than did Congregationalists, as table 29 illustrates. Both responded positively and significantly only to cloth-manufacturing output, but this tendency was more clearly marked among the Baptists.

In general, these findings suggest that levels of outmigration among Anglicans and Dissenters differed, with the remaining Anglican population becoming less industrial. Dissenters, because of the greater cohesiveness of their community life, retained more of an industrial working-class following, although important differences were still discernible among them. Congregationalist births, as noted earlier, increased rapidly from about 1820, and this trend coincided in that same year with the schismatic formation of a new Congregationalist church at Lower Forest Green. Those who remained at Upper Forest Green did so because of an intense attachment to that church's history and traditions; those who moved to Lower Forest Green followed the lead of wealthier members of the original congregation who sought a larger following by closer proximity to Nailsworth, which had grown into a small town.

The schism reflected the unavoidable tension between sectarian and denominational tendencies. The church at Upper Forest Green retained greater élan as a community, while the church at Lower Forest Green, although benefiting initially from its new locale, appeared more vulnerable to subsequent trade depressions, which by 1845 severely damaged its membership stability.^[49] Still, when combined in time series, the births of both Congregationalist churches appear especially high after 1820. The Shortwood Baptists bridged the sect/denomination divide more successfully; despite their complaints about membership turnover,^[50] they maintained greater cohesiveness throughout the entire period and retained, as we have seen, more of a working-class following.

Evangelical religion, combined with a more effective ministry, made possible this greater cohesiveness and in doing so contributed to the growth in Baptist births. A multiple regression of cloth output, annual male and female conversions, and total annual membership loss shows only the first and last variables explaining variance in Baptist births.^[51] When cloth output and membership loss increased, Baptist births rose, despite the absence of a correlation between the independent variables (see table 30).

TABLE 30 . Regression Coefficients: Cloth Production, Conversions, and Members Lost on Baptist Births, 1804-1838
D-var	I-var	df	b	t	Prob > t	B
Baptist births
	Cloth	1	0.0000497	3.262	0.0028	0.403305
	Male conversions	1	0.290144	1.530	0.1369	0.222142
	Female conversions	1	-0.045533	-0.342	0.7350	-0.051169
	Total membership loss	1	0.576898	4.837	0.0001	0.575423
Note: F = 12.054, df = 4, N = 33, Prob > F 0.0001, ; DW = 1.423, first-order autocorrelation coefficient = 0.281.

Baptist births, in other words, rose in response to a more productive industry, however much this higher productivity issued from a declining number of firms. A delining number of firms meant higher average unemployment and increased migration, which accounted for a meaningful proportion of Baptist membership loss. Such loss should theoretically have had a damping effect on reproduction, but instead we find an increase in the number of births.

The intensity of Baptist evangelicalism explains this apparent anomaly. The conversion rate tended to rise during periods of trade depression because of the pervasive anomie induced by such crises, although a high level of membership loss, attributable to either rising mortality or higher outmigration, accompanied and modified these gains. New converts, both male and female, contributed little to a rise in Baptist births; however, among old members who stayed behind, reproduction increased. Baptist births rose at times when members displayed an especially strong attachment to their community and thereby offset the constraining effects of industrial involution. Evangelicalism, in other words, gave hope where despair and falling births might otherwise have prevailed. This effect of religious culture on the growth in Baptist births can be further illustrated by adding another membership

TABLE 31. Regression Coefficients: Cloth Production, Members Lost, and Total Membership on Baptist Births, 1804-1838
D-var	I-var	df	b	t	Prob > t	B
Baptist births
	Cloth	1	0.00000389	0.716	0.4795	0.096756
	Total membership loss	1	0.452854	4.237	0.0002	0.451696
	Total membership	1	0.035732	3.494	0.0015	0.490351
Note: F = 24.043, df = 3, N = 33, Prob > F 0.0001, ; DW = 1.716, first-order autocorrelation (auto-R ) = 0.101.

variable to the same regression model, the effect of which appears in table 31.

Total membership (the annual base membership after adjustments for losses and conversions) reflected the changing strength of the Baptist church as a community and appears in time series in chapter 9 (fig. 21). Total membership loss remained highly significant, but fluctuations in cloth output ceased to have any effect. Total membership proved to be highly significant as well; since total membership was also correlated with cloth output, it absorbed a great deal of the variance this last variable had once explained,^[52] thereby revealing how much the cohesiveness of Baptist community life could shield members from exogenous shocks.^[53]

Despite the effects of membership loss, in other words, the sharp growth in total membership until 1850 shows that the communal life of the Baptist church retained a high tensile strength. This durability enabled its members to weather the anomie of the Industrial Revolution and long-term economic decline that paradoxically accompanied it; a high growth rate in Baptist births was the result.

A combination of economic and cultural conditions, therefore, determined the growth rate in all births. Demographers have sometimes underestimated cultural considerations, but these findings show that the concept of community, informed by a religious bond, vindicates their importance to the demographic

revolution. The contribution of a rising growth rate in births to overall population growth remains to be considered. To do so, we must compare the growth in all births to the movement of burials.

A combination of economic and cultural conditions also affected the growth rate in burials. Those who have argued for the primacy of its fall in explaining population advance have stressed the centrality of eighteenth-century medical progress.^[54] Defenders of the primacy of a rising birth rate have treated the falling burial rate as only a by-product of a shift in the age structure of the population, itself the result of economic factors, and have dismissed the significance of medical advances. They have suggested, moreover, that the progress of industrialization in the early nineteenth century caused the burial rate to rise.^[55] Before comparing births and burials at Horsley-Nailsworth in light of this controversy, we must construct the time series for burials.

The growth in burials for the Horlsey-Nailsworth area, like the growth rate in births, was calculated from a composite of values drawn from Anglican and Dissenter registers. The Shortwood Baptists maintained a burial register from 1808 that doubles the number of burials recorded in the Horsley parish register (Baptist burials for the 1775-1807 period were estimated).^[56] Indeed, most of those buried at Shortwood were not members of the Baptist church. Shortwood burials, therefore, added significantly to the total number of burials for the region. Nonconformist burials also included the burial registration figures for the Quakers and Congregationalists, each of whom had maintained separate grounds. The number of Quaker burials was predictably small, but Congregationalist registration was simply incomplete. Nevertheless, their registers provided sufficient detail to make estimates for missing years.^[57] If burials in Congregationalist grounds seemed few, however, this was due to the Congregationalist practice of using the Baptist burial ground: On May 31, 1827, Mary Heskins, the wife of Francis Heskins, carpenter, had been buried, and "both [were] inter'd by Mr. Edkins [sic ]," the Congregationalist minister.^[58] By combining Quaker, Congregational, Baptist, and Anglican registration figures, together with the inflators employed by Wrigley and Schofield,^[59] an accurate growth rate in burials for the region can be established.

The findings are summarized in tables 32 to 34, and the

observations are plotted in time series in figure 9. The findings support the neo-Malthusian view, which defends the primacy of rising births. A rising growth rate in births of 1.7 percent per annum in the late eighteenth century clearly outstripped a falling growth rate in burials of -0.35 percent per annum. If burials had moved faster, they might have qualified as an independent variable, the decline of which resulted from advances in medical science. Since burials moved more slowly, their general decline more likely resulted from a previous fall in the age at first marriage, which neo-Malthusians have identified as the primary cause of population growth. The growth rate in burials, furthermore, had begun to rise during the early nineteenth century, although moderately.

When expressed as crude vital rates, we can see a similar pattern at work (see table 33). The burial rate remained low and flat between 1781 and 1801, rose noticeably between 1801 and 1811, and fell precipitously in the following two decades, only to rise moderately during the 1830s. From 1791 the birth rate remained noticeably higher than the burial rate, although fluctuating in a similar way.^[60] Nationally, as table 34 reveals, the death rate fluctuated more erratically between 1781 and 1811; it fell noticeably thereafter and to levels not much higher than those at Horsley-Nailsworth. Indeed, national death rates throughout this period did not greatly differ from those at Horsley-Nailsworth, except in 1811. National birth rates, however, were significantly higher in only 1781, 1791, and 1821; this difference simply reflected local variation.^[61] In 1801, 1811, 1831, and 1841, national birth rates closely approximated those of the locality and sometimes fell below them.

The rising growth rate in burials at Horsley-Nailsworth during the early Industrial Revolution implies a general deterioration of living standards. However, the fall in the crude burial rate after 1811 and its relative flatness thereafter suggests otherwise for most of the period under study. Indeed, an analysis of mortality by age-sex cohort shows industrialization to have had a less catastrophic effect than anticipated.