¹ A more detailed discussion is provided in an appendix to “Long Term Changes in the National Product of the United States since 1870,” a paper presented at the 1951 meeting of the International Association for Research in Income and Wealth and to be published in 1952 by Bowes and Bowes, Cambridge, England, in Income and Wealth. Series II: Income and Wealth of the United States: Trends and Structure.

² All these figures are taken or derived from Martin, Table 1, p. 6, and are the variant adjusted for price changes by a cost-of-living index. The results for totals adjusted by an index of the general price level are roughly the same.

The concept used by Martin closely approximates that commonly referred to as “national income at factor cost.” Martin does not include undistributed corporate profits, a relatively negligible item during the period under consideration.

³ Ibid., pp. 9–12.

⁴ Ibid., p. 8. The reference is to the period 1799–1849, but it applies equally to the four decades under discussion.

⁵ There may be a slight difference in the dating of population (and gainfully occupied) figures and those for income originating: the latter are designated in Martin's book 1799, 1809, etc., whereas the former are for the census dates of the following year. However, as Martin indicates, his estimates, prior to 1899, “apply to no specific year but to a twelve months' period, beginning and ending within the two years beginning on January 1 of the year indicated.” p. 134.

⁶ Martin assigns the value of home or family manufactures reported in the Census of Agriculture to manufacturing (see p. 137). For the earliest census for which I could find these data (1839), the reported value is $29 million (See “7th Census of the United States, 1850,” Statistical View of U.S. Compendium, pp. 179 ff.). This compares with Martin's estimates of total income from agriculture (net) of $548 million. However, further allowance should be made for income from work by farm residents in cities and for their receipts of property income from sources other than agriculture.

⁷ Calculated on the assumptions stated in the text, and setting the ratio in column 4 at 0.5, product per worker outside of agriculture, expressed as a ratio to country-wide product per worker, amounted in 1799 to 2.3.4 = ([100] - [72.8xo.5]) ÷ 27.2. In 1839, the product of workers in agriculture was (68.6x0.5), and that of nonagricultural workers (31.4 x 2.2.34). The sum of these two products for 1839 is 107.7, compared with 100.0 in 1799. The increase in product per worker (member of the labor force) is thus 7.7 per cent.

⁸ The figures are from Historical Statistics, Series D-47–61, p. 64.

⁹ Adams, T. M., Prices Paid by Vermont Farmers etc. (Burlington, Vermont): Vermont Agricultural Experiment Station, Bulletin No. 507 (February 1944)Google Scholar.

¹⁰ Ibid., Table 47, p. 97.

¹¹ If one traces Martin's procedure to see how he secured such unacceptable results, the main reason is found in the use of Mulhall's estimates of agricultural capital (to which a large weight is assigned) in calculating the index used to extrapolate gross farm output to the beginning of the century (see Martin, pp. 135–36). It is difficult to understand why an item subject to changing valuation, and whose accuracy and relevance to estimating physical output of agricultural production are most doubtful, was used at all.

¹² Historical Statistics for the United States, Series G-13 and G-14, p. 142.

¹³ Rather similar results were found in comparisons of movements in prices received by farmers and in prices paid by them for manufactured products available for Vermont back to 1797. See Adams, Prices Paid by Vermont Farmers.

¹⁴ New York: Macmillan Co., 1915.