¹ Copeland, Melvin T., The Cotton Manufacturing Industry of the United States (Cambridge: Harvard University Press, 1923), p. 84Google Scholar.

² Draper, George O., “The Present Development of the Northrop Loom,” Transactions of the New England Cotton Manufacturers' Association (hereafter cited as Transactions), No. 50 (1895)Google Scholar and Navin, Thomas R., The Whitin Machine Works Since 1831 (Cambridge: Harvard University Press, 1950), p. 273CrossRefGoogle Scholar, estimate the average run before stopping at five minutes.

³ Draper put the weaving cost in one pound of print cloth at 2.8 cents out of a total labor cost of 5.728 cents. The estimate is for 1887 and for an unidentified New England mill. Draper, George, Labor-Saving Looms (3d ed; Hopedale, Mass.: Draper Company, 1907), p. 185Google Scholar. This figure is corroborated by a report of the United States Commissioner of Labor cited by Smith, T. R. in his The Cotton Textile Industry of Fall River, Massachusetts (New York: Kings Crown Press, 1944), p. 101Google Scholar. According to this report, which is for 1888–89, labor costs per yard of print cloth amounted to 12.377 cents of which 5.712 cents, or 46 per cent, was for weaving. Another supporting estimate is given in a report of M. Campbell, general manager of the Woonsocket Machine & Press Company, Woonsocket, Rhode Island, to the New England Cotton Manufacturers' Association annual meeting of 1897. Weaving costs of 3.303 cents per pound of cloth constituted 46 per cent of total labor costs per pound; Textile World, XIV (Jan. 1899), 60.

⁴ Navin, pp. 181–87.

⁵ For accounts of the development of the Draper loom, see Draper, Transactions; Draper, William F., “Continued Development of the Northrop Loom,” Transactions, No. 74 (1903), pp. 158–76Google Scholar; Navin, pp. 181–87; and Chase, William H., Five Generations of Loom Builders (Cambridge: Harvard University Press, 1950)Google Scholar. Technical descriptions of the Draper loom are provided by G. Draper, Transactions, and Townsend, I. U., “Northrop Loom,” Scientific America, Supplement to Vol. LIV (Sept. 1902)Google Scholar.

⁶ In addition to the citations immediately following, see Harris, Seymour, Economics of New England (Cambridge: Harvard University Press, 1952), p. 292CrossRefGoogle Scholar. Wolfbein, Seymour. The Decline of a Cotton Textile City (New York: Columbia University Press, 1944), p. 69Google Scholar; and Stelzer, I. M., “The Cotton Textile Industry,” in The Structure of American Industry (3d edition, edited by Adams, Walter; New York: Macmillan, 1961), p. 59Google Scholar.

⁷ Thomas, Edward W., “Comparative Life of Cotton Machinery and Valuations,” Transactions, No. 63 (1897), pp. 384–85Google Scholar.

⁸ Textile World, XIV (Jan. 1898), 24Google Scholar.

⁹ Smith, p. 119.

¹⁰ The Textile Industries (Philadelphia: Winston, 1938), pp. 74–75Google Scholar. Harris and Wolfbein (note 6) cite Michl as the source of their data on New England's adoption of the Draper loom.

¹¹ Chase, pp. 22–23.

¹² The figures Michl cites are for the change in looms in place, not sales of the Draper loom. The number of looms in place in the South and North increased between 1899 and 1914 by 153,000 and 81,000 respectively—the net increase in looms in place in the United States in this period being reduced by the decline in looms in place in other regions from 41,000 in 1899 to 28,894 in 1914. These figures are of importance in explaining the sales of Draper looms to the two major regions, the rate of growth of the industry obviously having an impact on the number of units of the new technique that are purchased. But it is clear that Draper looms could be and were purchased as replacements for existing equipment—that is, that some firms found it profitable to scrap existing equipment and purchase Draper looms in order to maintain a given level of capacity. That such replacement did occur is reflected in the fact that sales of the loom to the North of 150,464 greatly exceeded the number of looms added in New England, and, as pointed out later, sales to the North were almost exclusively to New England. Also, since the fifteenyear interval covered in the data is less than the normal twenty-year life over which these looms were depreciated, the excess of sales of Draper looms over looms added probably reflects only the replacement of existing looms by Draper looms and not that of Draper looms by newer Drapers.

¹³ Letter from Mr. J. H. Richardson, Treasurer, Draper Corporation, to Irwin Feller January 23, 1963.

¹⁴ Textile World, XXVII (Jan. 1911)Google Scholar; italics are in original.

¹⁵ The normal life of looms was put at twenty years, with depreciation charges put at 5 per cent per year; Uttley, T. W., Cotton Spinning and Manufacturing in the United States (Manchester: University Press, 1905), p. 26Google Scholar. Actual mill practice must have varied quite widely. A Tariff Commission study of 1912 which covered 40 New England and 27 southern mills noted that 25 per cent of the 127,572 looms in the study were over 20 years old. A few looms (532) dated back to before the Civil War. The report does not give any breakdown of age of looms by region, but it is probable, given the relative size of the industry in each region prior to 1890, that most of the looms over 20 years of age were located in New England. U. S. Congress, House, United States Tariff Commission, Cotton Manufactures, 62d Cong., 1st sess., 1912, House Doc. 643, p. 472.

¹⁶ Navin, p. 605.

¹⁷ A most important defect is that neither portion of the ratio may have any meaningful relationship to the relevant economic conditions. For example, in a period where expansion of plant and equipment is called for, if lax or inept management decides to add only a fraction of what would be profitable to capacity and yet provides for this increase with the new technique, the measure of diffusion would be 1. Yet this is certainly no proof of the economic rationality of the entrepreneurs concerned. Important as this defect is to the proxy, the same criticism can be leveled against all measures of diffusion that state how fast a technique was adopted without providing a norm by which one can say whether the observed rate of adoption was the correct one.

¹⁸ Textile World, XIV (1898), 24–25Google Scholar.

¹⁹ Smith, pp. 108–9.

²⁰ Fabricant, Solomon, The Output of Manufacturing Industry, 1899–1937 (New York: NBER, 1940), pp. 174–75Google Scholar.

²¹ Copeland, Melvin T., “Progress of the Automatic Loom: Note,” Quarterly Journal of Economics, XXV (Aug. 1911), 747Google Scholar.

²² As stated by G. Draper in 1895, the Company felt that with such a large field it could “only advance by degrees. We took an easy problem at the outset and made plain prints with bobbin filling. On cop fillings we have run single looms successfully, but wish to experiment on a larger scale before delivering looms for that purpose”; Draper, Transactions, p. 98.

²³ Frankel, Marvin, “Obsolescence and Technological Change in a Maturing Economy,” American Economic Review, XLV (June 1955), pp. 296–319Google Scholar.

²⁴ Tariff Commission (cited in n. 15), p. 498.

²⁵ U. S. Census Bureau, Biennial Census of Manufacturers, 1927, p. 258, Table 12.

²⁶ U. S. Census Bureau, Thirteenth Census of the United States, 1910, Vol. X, Manufactures, p. 57Google Scholar.

²⁷ G. Draper, Labor-Saving Looms, p. 65; the report of the Manufacturers Record is cited in Smith's, Robert S.Mill on the Dani (Durham, N. C. : Duke University Press, 1960), p. 72Google Scholar.

²⁸ G. Draper, Labor-Saving Looms, preface.

²⁹ Ibid., p. 215.

³⁰ Stern, Boris, “Mechanical Changes in the Cotton Textile Industry,” Monthly Labor Review, LX (Aug. 1937), 316–41Google Scholar.

³¹ Pepperell Manufacturing Company, Records, V.DD-1 (Baker Library, Harvard University), p. 35. There is no mention in the Pepperell records or in those of the other mill companies of any credit on old looms being offered by any loom manufacturer besides Draper.

³² Lyman Mills, Records, V.MFC-1; V.PO-12 (Baker Library), p. 143.

³³ Laconia Mills, Records, V.FB-7 (Baker Library), p. 22.

³⁴ Ibid., p. 23.

³⁵ Mills, Lyman, Records, V.MFC-2, March 30, 1911Google Scholar; December 3, 1912.

³⁶ Uttley, pp. 66–67.

³⁷ Navin, p. 534.

³⁸ Tariff Commission, p. 464.

³⁹ Ibid., p. 463.

⁴⁰ G. Draper, Labor-Saving Looms, p. 47.

⁴¹ Tariff Commission, p. 11.

⁴² Ibid., p. 12.

⁴³ Taussig, Frank W., Some Aspects of the Tariff Question (Cambridge: Harvard University Press, 1934), pp. 274–75Google Scholar.

⁴⁴ Uttley, p. 6.

⁴⁵ Ibid., p. 24.

⁴⁶ ibid., p. 26.

⁴⁷ Ibid.

⁴⁸ Ibid.

⁴⁹ Uttley's error is that he has computed depreciation charges according to two different principles. He has made one estimate on the basis of physical deterioration and the other on the basis of economic obsolescence. The twenty-year life given for the plain loom is a physical one. Economic obsolescence, however, refers to a diminution in the income stream to be derived from using a technique because of the appearance of a rival technique. The shorter life estimated for the Draper loom is based upon the assumption that as newer, more efficient, Draper looms made their appearance and as their price fell because of the expiration of patent protection, the revenue stream to be derived from the older Draper looms in use would be reduced. The estimated fourteen-year economic life of the Draper loom might not be the actual life obtained; what is clear, though, is that if the economic life of the Draper loom is estimated at less than its physical life, then similar deductions must be made against the plain loom. In that automatic and plain looms are functional rivals, technical improvements and a fall in price in the former have the effect of reducing costs of production on new Draper looms to the disadvantage of all other firms, including those using the plain looms, thereby reducing the income stream and consequently the economic life of these looms. Thus in understating depreciation charges against the plain loom, Uttley has understated the savings to be derived from adopting the Draper loom.

⁵⁰ Ibid., p. 26.

⁵¹ Ibid., p. 51.

⁵² Ibid., p. 57. The price per cut on plain looms was 31.45 cents in the North and 23.30 cents in the South.

⁵³ Draper, G., Transactions (1895), p. 95Google Scholar.

⁵⁴ Ibid., p. 108.

⁵⁵ After General William.F. Draper had argued that his company had always acted “on the principle of giving the purchaser the lion's share of the benefit,” Kent, W. J. remarked, “General Draper has just told you when we purchase anything from the firm of George Draper & Sons the purchaser gets the lion's share. I was not aware of that fact before, and if such is the case, my sympathy goes out to the lamb everytime”; Transactions, No. 60 (1896), pp. 134–35.Google Scholar

⁵⁶ This approach would be in accord with present organization theory, which holds that problem-solving search activity will initially focus on alternatives close to existing practices before exploring substantially different alternatives. See Cyert, Richard and March, James G., A Behavioral Theory of the Firm (Englewood Cliffs, N. J.: Prentice-Hall, 1964)Google Scholar.

⁵⁷ Smith, p. 115.

⁵⁸ Mills, Lyman, Records, V.MPC (Dec. 18, 1900)Google Scholar.

⁵⁹ Knowlton, Evelyn H., Pepperell's Progress (Cambridge: Harvard University Press, 1948), p. 139CrossRefGoogle Scholar.

⁶⁰ G. Draper, Labor Saving Looms, p. 38.