This research has benefited from the helpful comments of Robert Ankli, Lou Cain, Stanley Engerman, Duncan McDougall, Donald Schaefer and Joseph Swanson. I also wish to thank Paul Taylor for assistance in carrying out many of the calculations underlying the estimates.

¹ The term workforce is used here to represent a labor measure more accurately termed “gainful workers.”

² The service sector is defined as the following industries: trade, transportation, utilities, finance, professional and personal services, education, and government.

³ See my earlier paper “Urbanization and the Growth of the Service Workforce,” Explorations in Economic History, VIII (Spring 1971)Google Scholar for a discussion of this issue.

⁴ See for example McKelvey, B., The Urbanization of America (New Brunswick, New Jersey: Rutgers University Press, 1963)Google Scholar or Pred, A., The Spatial Dynamics of Urban Industrial Growth, (Cambridge, Massachusetts: M.I.T. Press, 1966)Google Scholar.

⁵ See Clark, C., “The Economic Functions of a City in Relation to Its Size,” Econometrica, XIII (April 1945), 97–113CrossRefGoogle Scholar or Fisher, A.G.B., The Clash of Progress and Security, (London: McMillan and Co., 1935), pp. 100–127Google Scholar. See also Easterlin's, R. comments on “The Service Industries in the Nineteenth Century,” Production and Productivity in the Service Industry, Fuchs, V., ed. (New York: NBER, 1969), pp. 357–359Google Scholar. Easterlin did not state a view of urban structural change, but the most reasonable interpretation of his position implies that the service share of the urban workforce was rising throughout the nineteenth century.

⁶ Swanson, J. and Williamson, J., “The Growth of Cities in the American Northeast, 1820–1870,” Explorations in Entrepreneurial History, IV (Fall 1966)Google Scholar, Supplement.

⁷ The evidence relating to rural development is even more sparse. S. Kuznets suggested that services available in the countryside were minimal at best. Population Redistribution and Economic Growth, U.S. 1870–1950, III, (Philadelphia: American Philosophical Society, 1964), p. xxviGoogle Scholar.

⁸ See my earlier paper.

⁹ Swanson and Williamson, “The Growth of Cities …,” p. 33.

¹⁰ Ibid., especially pages 40–41 for a discussion of the role of the rural market in explaining urban growth.

¹¹ U.S. Bureau of the Census, 9th Census of the U.S., 1870, I (Washington, D.C.: G.P.O.), Table XXXII; 10th Census of the U.S., 1880, Vol. I, Table XXXVI; Eleventh Census of the U.S., 1890, Vol. 2, Tables 116–118; Twelfth Census of the U.S., 1900, Vol. II, Table 94; Thirteenth Census of the U.S., 1910, Vol. 4, Tables III, IV.

¹² These aggregate figures were obtained from Carson, Daniel, “Changes in the Industrial Composition of Manpower since the Civil War,” Studies in Income and Wealth, XI (New York: NBER, 1949),Google Scholar and Gallman, R. and Weiss, T., “The Service Industries in the Nineteenth Century,” Studies in Income and Wealth, XXXIV (New York: NBER, 1969)Google Scholar. These latter figures are only minor revisions of Carson's estimates of gainful workers in the various service industries.

¹³ For services, the U.S. totals were obtained from Gallman and Weiss, “Service Industries …,” Table A-12. For all other industries the totals were obtained from Carson, “Changes …,” Table 1, with a minor adjustment made for differences between his estimates of laborers, n.o.s. in services and those underlying the Gallman-Weiss figures.

¹⁴ The estimates are discussed further in the Appendix, and the data are presented there in tabular form. It may be useful to know that the urban share of the total labor force was 28 percent in 1870, 32 percent in 1880, 41 percent in 1890, 45 percent in 1900, and 53 percent in 1910. The residual rural shares were 72, 68, 59, 55 and 47 percent.

¹⁵ My estimating procedure may be responsible for some of the stability. The workforce structure for cities with populations less than 25,000 was assumed to be identical to that for cities with populations between 25,000 and 100,000. If the smaller cities had different industrial structures, then my estimate of the urban distribution will be different from the actual. There would then be some difference with regard to changeover time because the smaller cities became less important in the aggregate. If the smaller city workforce distribution changed differently than did that in larger cities, then the pattern of stability might be changed further.

The data for 1870 and 1890 indicate that there was virtually no variation in workforce structure by city size in those cities for which data are available. The variation across regions was of greater importance and should outweigh any effects the smaller cities may have had. For this reason I estimated the U.S. figures on a region by region basis.

¹⁸ Some writers apparently minimized the rise in the labor force participation rate. By assuming a constant participation rate, and focusing on only one industry or industrial sector in which the workforce increased more rapidly than urban population, one would conclude that the industry increased relative to the urban workforce. If the participation rate was increasing, then the industry or industrial sector under investigation may have been increasing only at the same rate as the rest of the workforce, and there would have been no change in relative importance. The evidence in Table 4 indicates a substantial rise in the urban participation rate.

¹⁷ Gallman and Weiss, “Service Industries …,” Table 8, p. 302.

¹⁸ Ibid., Table A-1, p. 306. Value added by the hand trades and shelter is excluded from the sectoral total.

¹⁹ The classification of public service employees is one of the more suspect of the census groups. See the discussion in D. Carson, “Changes ….”

²⁰ Frank Lewis, University of Rochester, is now engaged in such a study. “The Relative Movement of Labor from Agriculture to Manufacturing in the U.S.: 1869–1899,” (Mimeo), was presented at the 1971 annual meeting of the Cliometrics Society, Madison, Wisconsin.

²¹ It has been asserted that intermediate services were entirely absent from the countryside. (See S. Kuznets Population Redistribution …, p. xxvi).

²² The manufacturing workforce declined slightly after 1900.

²³ J. Ermisch and T. Weiss, “The Impact of the Rural Market on the Growth of the Urban Workforce, United States, 1870–1910,” May 1972 (Mimeo).

²⁴ Several assumptions were necessarily made in order to conduct this preliminary analysis, and these assumptions imply this explanation. It is quite possible that alternative explanations are more tenable. In particular, if worker productivity growth in rural non-agricultural industries was slower than we assumed, then the rural economy would have been unable to supply as large a share of the increased demand, and urban goods and services would have substituted for rural ones.

²⁵ The trade and transport share rose from 38 percent in 1870 to 42 percent in 1880 and 1890, and declined to 36 percent in 1900 and 30 percent in 1910. The professional and personal share declined from 50 percent in 1870 to 39 percent in 1890 and then increased to 47 percent by 1910. (Shares can be derived from data in Appendix Table II).

²⁶ The population living in towns and villages with populations less than 2,500 is excluded from the urban population count, and with some exceptions would be the locus of the rural non-agricultural workforce. The evidence shows that the village share of the rural population rose from 12 percent in 1890 to 14 percent in 1900 and to 16 percent in 1910. See Historical Statistics of the United States (Washington, D. C.: G.P.O., 1960)Google Scholar, Series A-207, 208. It was only after 1890 that the rural non-agricultural workforce increased in importance (Table 3). The growth of the village population, then, is consistent with the growth of the rural workforce. The case would be strengthened even more if one could show that the village share remained constant in earlier years. However, evidence on village populations is unavailable for earlier years.

²⁷ D. Carson, “Changes …,” Table 1.

²⁸ Lebergott, S., “Labor Force and Employment, 1800–1860,” Studies in Income and Wealth, XXX (New York: NBER, 1966), Table 1Google Scholar.

²⁹ If the urban and rural population shares are held constant at the mean values for the period (urban share 35.5 percent, rural 64.5) and used to weight the urban and rural participation rates of 1870 and 1910, the resulting weighted U.S. participation rate would have changed by.045 percentage points. If the urban and rural participation rates were held constant and equal to the mean for the period (urban rate,.405; and rural.33) and used to weight the urban and rural population shares of 1870 and 1910, the result would be a rise of only.016 percentage points. In other words, the changing participation rates account for almost seventy-five percent of the calculated.061 percentage point rise in the U.S. participation rate. Using weights other than the means alters the results somewhat, but the basic conclusion is unchanged.

³⁰ S. Lebergott makes the point that the census counts since 1890 indicate that reported unemployment in the farm population is small and so shows less variation than non-farm series (Manpower in Economic Growth, New York: McGraw-Hill, 1964), pp. 187–88Google Scholar.

³¹ See Easterlin, R., Population, Labor Force and Long Swings in Economic Growth (New York: NBER, 1968)Google Scholar, Ch. 7 for evidence on this phenomenon in the twentieth century.

³² Lebergott, Manpower …, Table A-15, p. 522.

³³ The peak unemployment rate was higher during the 1890's than 1870's, which may explain some of the difference in the behavior of the urban participation rate in the two decades. (Lebergott, Manpower …, Table 4–3, p. 187).

³⁴ Weiss, “Urbanization …,” 255–57.

³⁵ The urban share of the U.S. workforce rose from 28 percent in 1870 to 53 percent in 1910. This reflects the rise in urban population and in the urban participation rate. Even if the participation rate had not changed, the urban share of the U.S. workforce would have risen to almost 52 percent, solely on the strength of the rise in urban population.