Hostname: page-component-cd9895bd7-jn8rn Total loading time: 0 Render date: 2024-12-18T20:25:47.190Z Has data issue: false hasContentIssue false
  • English
  • Français

Market Structure, Industrial Research, and Consumers of Innovation: Forging Backward Linkages to Research in the Turn-of-the-Century U.S. Steel Industry

Published online by Cambridge University Press:  13 December 2011

Janet T. Knoedler
Janet T. Knoedler
Affiliation:
Janet T. Knoedler is assistant professor of economics atBucknell University.
Get access Rights & Permissions [Opens in a new window]

Abstract

Although the U.S. steel industry's concentrated market structure and well-established production technology curbed active research by most steel firms, between 1880 and 1910 vertical research arrangements between steel producers and steel consumers, notably the Pennsylvania Railroad, became a key factor in promoting both increased innovation in basic steel products and increased innovative effort by steel producers, albeit slowly and gradually. Thus, research into steel was initiated not by steel producers but by steel consumers, who established in-house industrial research laboratories and interfirm cooperative research arrangements as a means to solve their technical problems with steel products. They also began to work toward creating an institution—the American Society for Testing Materials—that would allow for effective interaction with other consuming firms and, eventually, with producing firms to exchange information and build consensus.

Type
Articles
Information
Business History Review , Volume 67 , Issue 1 , Spring 1993 , pp. 98 - 139
Copyright
Copyright © The President and Fellows of Harvard College 1993

Access options

Get access to the full version of this content by using one of the access options below. (Log in options will check for institutional or personal access. Content may require purchase if you do not have access.)

References

1 See, for example, Adams, Walter and Mueller, Hans, “The Steel Industry,” in The Structure of American Industry, ed. Adams, Walter, 8th ed. (New York, 1990), 72–100Google Scholar; Berglund, Abraham, “The United States Steel Corporation and Price Stabilization,” Quarterly Journal of Economics 38 (Nov. 1923): 1–29CrossRefGoogle Scholar; Chandler, Alfred D. Jr., Scale and Scope: The Dynamics of Industrial Capitalism (Cambridge, Mass., 1990)Google Scholar; Hogan, William T., Economic History of the Iron and Steel Industry in the United States (Lexington, Mass., 1971)Google Scholar; Parsons, Donald and Ray, Edward John, “The United States Steel Consolidation: The Creation of Market Control,” Journal of Law and Economics 18 (April 1975): 181–219CrossRefGoogle Scholar; Scherer, Frederic and Ross, David, Industrial Market Structure and Economic Performance, 3d ed. (1970; Boston, Mass., 1990), 236, 252, 293, 669Google Scholar; Schroeder, Gertrude, The Growth of Major Steel Companies, 1900–1950 (Baltimore, Md., 1953)Google Scholar; and Tiffany, Paul, “The American Steel Industry in the Postwar Era,” in Changing Patterns of International Rivalry: Some Lessons from the Steel Industry, ed. Abe, Etsuo and Suzuki, Yoshitaka (Tokyo, 1991), 245–65.Google Scholar

2 The basic oxygen furnace brought about significant reductions in both production time and costs by permitting production of a “heat” of steel in about forty minutes compared to six to eight hours under the open-hearth method used by most U.S. plants. Yet most U.S. steel producers were very slow in adopting this technology; a Bethlehem executive was quoted in the 1950s: “We don't move until we are sure that we can get the best possible yield out of a new technique or facility.” See Tiffany, Paul A., The Decline of American Steel (New York, 1988), 133Google Scholar; see also Acs, Zoltan, The Changing Structure of the U.S. Economy: Lessons from the Steel Industry (New York, 1984), 91Google Scholar; and Adams, Walter and Dirlam, Joel, “Big Steel, Invention, and Innovation,” Quarterly Journal of Economics 80 (May 1966): 167–89.CrossRefGoogle Scholar Continuous casting provided better quality steel and a further reduction in production time; steel could be taken directly from the refining process in a continuous production line and formed into slabs, billets, or blooms without going through the intermediate stage of forming the refined steel first into ingots and then cooling it and reheating it before producing slabs, billets, or blooms. See Tetsuji Okazaki, “Import Substitution and Competitiveness in the Prewar Japanese Iron and Steel Industry,” in Abe and Suzuki, eds., Changing Patterns of International Rivalry, and also Tiffany, “The American Steel Industry in the Postwar Era,” 251-52.

3 For the complete story of the history of the British steel industry, see, for example, Dillard, Dudley, Economic Development of the North Atlantic Community (Englewood Cliffs, N.J., 1967)Google Scholar; and Landes, David, The Unbound Prometheus (New York, 1969).Google Scholar See also Chandler, Scale and Scope, 283-84; Elbaum, Bernard, “The Steel Industry before World War I,” in The Decline of the British Economy, ed. Elbaum, Bernard and Lazonick, William (New York, 1986)Google Scholar; and Steven Tolliday, “Competition and Maturity in the British Steel Industry, 1870–1914,” in Abe and Suzuki, eds., Changing Patterns of International Rivalry, 39-51.

4 Tweedale, Geoffrey, “Science, Innovation, and the ‘Rule of Thumb’: The Development of British Metallurgy to 1945,” in The Challenge of New Technology, ed. Liebenau, Jonathan (Aldershot, England, 1988), 63–70Google Scholar; Tolliday, “Competition and Maturity in the British Steel Industry,” 57.

5 Landes, The Unbound Prometheus, 263; Chandler, Scale and Scope, 492; and Rainer Fremdling, “The German Iron and Steel Industry,” in Abe and Suzuki, eds., Changing Patterns of International Rivalry, 113-36. The Thomas process was the basic version of the Bessemer process.

6 Landes, The Unbound Prometheus, 267.

7 Fremdling, “The German Iron and Steel Industry,” 120. As is often noted, the conscious application of science to industry in the German industrial revolution of the late nineteenth century was instrumental in the country's rapid emergence in many key industries of the day, including steel. See also Dillard, Economic Development, 308-16; Landes, The Unbound Prometheus, 346-52; Lee, Robert, “The Paradigm of German Industrialisation,” in German Industry and German Industrialisation, ed. Lee, W. R. (London, 1991)Google Scholar; and Marshall, Alfred, Industry and Trade (1919; New York, 1970), 121–39Google Scholar.

8 Tweedale, Geoffrey, Sheffield Steel and America (Cambridge, England, 1987), 186–87.Google Scholar

9 See Allen, Robert C., “International Competition in Iron and Steel, 1850–1913,” Journal of Economic History 39 (Dec. 1979): 911–14CrossRefGoogle Scholar; and Steven B. Webb, “Tariffs, Cartels, Technology and Growth in the German Steel Industry, 1879–1914,” ibid. 40 (June 1980): 309-11, 328.

10 Allen, “International Competition in Iron and Steel,” 913; and Tiffany, Decline of American Steel, 8-9, 11. See also Taussig, F. W., A Tariff History of the United States, 8th ed. (New York, 1931), 361–408.Google Scholar In U.S. Steel's earliest sorties into international markets, the firm attempted to organize the global market along the lines of its domestic organization, but such efforts met with opposition from the U.S. government and little enthusiasm from foreign producers.

11 As Adams and Dirlam observed in 1966, “In innovation as in invention, the giants of the steel industry have lagged, not led.” See Adams and Dirlam, “Big Steel, Invention, and Innovation,” 175. See also Chandler, Scale and Scope, 136; Galambos, Louis, “The American Economy and the Reorganization of the Sources of Knowledge,” in The Organization of Knowledge in Modern America, ed. Oleson, Alexandra and Voss, John (Baltimore, Md., 1979)Google Scholar; and Schroeder, The Growth of Major Steel Companies.

12 For a sampling of the standard literature, see Kamien, Morton and Schwartz, Nancy, “Market Structure and Innovation: A Survey,” Journal of Economic Literature 13 (1975): 1–37Google Scholar; Mansfield, Edwin, The Economics of Technological Change (New York, 1968)Google Scholar; Nelson, Richard R., “The Economics of Invention: A Survey of the Literature,” Journal of Business 32 (April 1959): 101–27CrossRefGoogle Scholar; Nelson, , Technology, Economic Growth, and Public Policy (Washington, D.C., 1967), 29-30, 35–42Google Scholar; and Schmookler, Jacob, “Economic Sources of Inventive Activity,” in The Economics of Technological Change, ed. Rosenberg, Nathan (New York, 1971), 135.Google Scholar The threshold concept is developed by Markham, Jesse, “Market Structure, Business Conduct, and Innovation,” American Economic Review 55 (May 1965): 323–32.Google Scholar One of the preeminent economists to look at the economics of innovation is Edwin Mansfield, who attempted an empirical analysis of the technological superiority of large firms in his “Size of Firms, Market Structure, and Innovation,” Journal of Political Economy 71 (1963): 556–76CrossRefGoogle Scholar, and concluded that large firms in the three industries of his study did not always account for the largest share of innovations; thus, contrary to the views attributed to John Kenneth Galbraith and Joseph Schumpeter, disproportionately large firms in concentrated industries did not seem to be a prerequisite for innovation, nor was there a distinct relationship between the degree of concentration of an industry and the industry's rate of technological change. In a later study, he argued that small firms played important roles in the early stages of R&D (presumed to be the less expensive stages), and that “a slight amount of concentration may promote more rapid invention and innovation"; see Mansfield, et al. , The Production and Application of New Industrial Technology (New York, 1977), 16.Google Scholar A good survey of both sides of the theoretical and empirical economics literature is found in Baldwin's, William Market Power, Competition and Antitrust Policy (Homewood, Ill., 1987)Google Scholar and Scherer's Industrial Market Structure and Economic Performance.

13 For discussion of some of the historical and theoretical background for interfirm and user-based research and early examples of user-based research in the U.S. automobile industry, see Knoedler, Janet T., “Early Examples of User-Based Industrial Research,” Business and Economic History 22 (Fall 1993): 285–94.Google Scholar

14 Chandler, Scale and Scope, 128-37; Chandler, , The Visible Hand: The Managerial Revolution in American Business (Cambridge, Mass., 1977), 266–69Google Scholar; Clark, Victor S., History of Manufactures in the United States (New York, 1929), 2: 77-79, 205, 232-33, 333Google Scholar; and Temin, Peter, Iron and Steel in Nineteenth-Century America: An Economic Inquiry (Cambridge, Mass., 1964), 163.Google Scholar The trend in the use of larger and larger furnaces became a competition for a time—the famous Lucy furnace, built at Pittsburgh in 1872, was the largest furnace to date, 75 feet high by 20 feet in diameter, and was built to ensure a supply of pig iron for the Pittsburgh steel works. It produced the record output to that date—475 tons of Bessemer steel per week. When Carnegie's firm, the Edgar Thomson Steel Company, began to produce Bessemer steel in 1873, Carnegie built the famous Isabella furnace to rival the Lucy furnace in output levels. In 1873 Cambria also built two furnaces of the same design. Other notable and similar innovations included: increasing the number of heats; using larger and more converters in a single plant; using removable bottoms for Bessemer converters; using better refractory lining in the furnaces; using hydraulic crane systems to move ingots; using two-high reversing mills; rolling rails directly from the ingot without shearing or reheating the bloom; using coke for the smelting fuel; using by-product coke and charcoal ovens; and moving to water cooling and dry-air blasts. Chandler notes that as a result of the increased throughput, Carnegie's costs fell dramatically and thus subsequently dropped rail prices from $67.50 per ton in 1880 to $17.63 per ton. And even with the fall in prices, the firm's profits rose throughout. See Scale and Scope, 129; and The Visible Hand, 269.

15 See, for example, Knoedler, “The Transition from Rule-of-Thumb to Science in Industry,” unpub. MS, and “Backward Linkages to Industrial Research in Steel, 1870–1930” (Ph.D. diss., University of Tennessee, 1991)Google Scholar; Mowery, David and Rosenberg, Nathan, Technology and the Pursuit of Economic Growth (New York, 1989), 29CrossRefGoogle Scholar; Rosenberg, Nathan, “The Commercial Exploitation of Science by American Industry,” in The Uneasy Alliance: Managing the Productivity-Technology Dilemma, ed. Clark, Kim B., Hayes, Robert H., and Lorenz, Christopher (Boston, Mass., 1985), 22Google Scholar; Birr, Kendall, “Science in American Industry,” Science and Society in the United States, ed. Van Tassel, David D. and Hall, Michael G. (Homewood, Ill., 1966), 65.Google Scholar

16 Howard Bartlett acknowledged that many of those among the ranks of corporate management held suspicions toward science's “impractical dreamer [who] … belonged in the university where he would not upset the methods that had worked for many years.” See Bartlett, Howard R., “The Development of Industrial Research in the United States,” in National Research Council, Research—A National Resource, vol. 2 (Washington, D.C., 1940), 24.Google Scholar John J. Beer and W. D. Lewis observed that the tradition of trade secrecy in industry conflicted with scientists' orientation toward sharing and disseminating the results of their work, and that industrial scientists often represented a “potentially antagonizing influence with regard to the working force.” See “Aspects of the Professionalization of Science,” Daedalus, Fall 1963, 765.

17 Lewis, W. D., “Industrial Research and Development,” in Technology in Western Civilization, ed. Kranzberg, Melvin and Pursell, Carroll (New York, 1967), 621Google Scholar; Bartlett, “The Development of Industrial Research in the United States,” 25. As John Rae has observed, for the most part, the early industrial scientists in the U.S. assumed “ancillary roles in what was still a largely empirical technology"; see Rae, “The Application of Science to Industry,” in Oleson and Voss, eds., The Organization of Knowledge in Modern America, 252. See also Birr, “Science in American Industry,” 41. Although the number of industrial scientists increased in the late nineteenth century, in 1900 there were still only 276 chemists employed on a full-time basis in the chemicals industry. See also Galambos, “The American Economy and the Reorganization of the Sources of Knowledge,” 269, 272. According to J. S. Unger of Carnegie Steel's industrial research lab, as of 1900 industrial research was not performed by individuals specially hired for the work but by a variety of people—the chemist, the engineer, or the departmental superintendent. See Unger, J. S., “Some Practical Views of Research,” in American Society for Testing Materials, Topical Discussion on Cooperation in Industrial Research (Philadelphia, Pa., 1918), 58Google Scholar.

18 See Temin, Iron and Steel in Nineteenth-Century America, 156; and Rosenberg, “The Commercial Exploitation of Science by American Industry,” 27.

19 Mowery, David, “The Emergence and Growth of Industrial Research in American Manufacturing, 1899–1945” (Ph.D. diss., Stanford University, 1981)Google Scholar.

20 Clark, , History of Manufactures in the United States, 2: 78Google Scholar.

21 The Iron Age, 10 Jan. 1910, 72.

22 This data comes from the National Research Council's 1920 and 1921 editions of Industrial Research Laboratories of the United States (Washington, D.C.). The Carnegie story is often cited as evidence of the firm's progressiveness; yet most accounts also note that the firm placed a higher value on the chemist's commercial value than on the technical progressivity that he represented. Carnegie is quoted as gloating about the fact that the chemist allowed the firm to purchase its raw materials on the basis of chemical analysis rather than on the basis of the mine owner's reputation. See Bartlett, “The Development of Industrial Research in the United States,” 27-28; and Chandler, The Visible Hand, 268-69. The chemistry lab's activities were relegated to the Carnegie firm's overall goal of minimizing costs; and Carnegie's comments also indicate that the lab had few early emulators. Moreover, as Chandler notes, the chemist hired by Carnegie was “a learned German.”

23 On industry rankings see Schroeder, The Growth of Major Steel Companies, 38, 43; and Parsons and Ray, “The United States Steel Consolidation.” However, these firms are cited as large firms because they both appeared on Navin's list of the top 500 firms in the United States as of 1917—Inland as number 87 and Taylor-Wharton as number 474. See Navin, Thomas R., “The 500 Largest American Industrials in 1917,” Business History Review 44 (Autumn 1970): 360–86.CrossRefGoogle Scholar

24 See National Research Council, Industrial Research Laboratories of the United States (1921, 1946).

25 Birr, “Science in American Industry,” 41; Rosenberg, “The Commercial Exploitation of Science by American Industry,” 27. Among the products of early steel innovations were three of great importance: better blowing engines for blast furnaces, improved rolling mills, and development of mass production technology to produce barbed wire, fences, nails, and springs. See Frank Sisco, a steel industry representative, “Research in the Iron and Steel Industry,” in National Research Council, Research—A National Resource, vol. 2 (Washington, D.C., 1940), 158Google Scholar.

26 Jewett, Frank B., “Industrial Research,” National Research Council, Reprint and Circular Series of the National Research Council, no. 4 (1919): 15Google Scholar (paper read before the Royal Canadian Institute, Toronto, Canada, 8 Feb. 1919).

27 Sisco, “Research in the Iron and Steel Industry,” 162.

28 Steel producers alternated between using pools to curtail output and increase price and engaging in vicious price wars. See, for example, Clark, History of Manufactures in the United States, 2: 223; and Temin, Iron and Steel in Nineteenth-Century America, 208.

29 See, for example, Yamawaki, Hideki, “Dominant Firm Pricing and Fringe Expansion: The Case of the U.S. Iron and Steel Industry,” Review of Economics and Statistics 67 (Aug. 1985): 429CrossRefGoogle Scholar; Lamoreaux, Naomi, The Great Merger Movement in American Business, 1895–1904 (New York, 1985), 118-26, 135–58CrossRefGoogle Scholar; and Scherer, F. M., Industrial Market Structure and Economic Performance, 2d ed. (Chicago, Ill., 1980), 178–80Google Scholar.

30 In 1901 U.S. Steel dominated production of the majority of the finished steel products produced in the U.S.: it produced 75 percent of steel billets and T rails, 90 percent of street railroad rails, 60 percent of all rails, 85 percent of structural steel, 80 percent of tin plate, 85 percent of wire, 85 percent of wire and wire products, and “substantially all of the sheet steel, tubes, and tin plating capacities.” As well, it controlled a substantial portion of steel inputs: 46 percent of all iron ore mined in the United States, 45 percent of pig iron, 72 percent of Bessemer steel, 60 percent of open-hearth steel, and 67 percent of total ingots. See Schroeder, The Growth of Major Steel Companies, 38, 43; and Parsons and Ray, “The United States Steel Consolidation,” 211.

31 Adams points out that “[t]he corporation remained sufficiently big, however, to keep its competitors ‘in line’ without threats and without displays of force”—a form of “friendly competition.” See “The Steel Industry,” 152. U.S. Steel maintained that stable prices were good and believed that dissemination of information regarding steel prices to other steel producers was important: this was “quite apparent in testimony by major figures before Congress during the early post-merger years.” See Parsons and Ray, “The United States Steel Consolidation,” 208. Price stabilization was also an explicit goal of the new corporation in Berglund's view: “[f]or the development of a large and well-established iron and steel business, price stabilization would therefore be a natural and farsighted policy. … a great combination is organized primarily to regulate or restrict competition and, therefore, production. The object of such a consolidation is not so much a better adaptation to market conditions as control of those conditions. … [i]ts attitude, furthermore, toward independents has never been characterized by any attempt at what is called destructive competition.” See Berglund, “The United States Steel Corporation and Price Stabilization,” 5, 29.

32 Rae, “The Application of Science to Industry”; Galambos, “The American Economy and the Reorganization of the Sources of Knowledge”; Mowery and Rosenberg, Technology and the Pursuit of Economic Growth.

33 Sisco, “Research in the Iron and Steel Industry,” 162.

34 The American Rolling Mill lab is often described as the first true industrial research laboratory in the steel industry, or the first lab recognized as such. This may result from the bias against testing as industrial research. See Bartlett, “The Development of Industrial Research in the United States,” 58; Sisco, “Research in the Iron and Steel Industry,” 162; and Borth, Christy, True Steel: The Story of George Matthew Verity and His Associates (Indianapolis, Ind., 1941)Google Scholar; see National Research Council, Industrial Research Laboratories of the United States (1921, 1946).

35 This represented consolidation of varied research activities that had been carried out for years by the corporation's various subsidiaries, under a central unit. Bartlett observed that at the formation of U. S. Steel, all of the constituent companies had labs in which “more or less systematic investigation had been carried on for some years.” See Bartlett, “The Development of Industrial Research in the United States,” 59; and also The Iron Age, 14 April 1927, 1096.

36 In 1910 U.S. Steel acquired the Heroult patents for electric steel making and refining from the French steel producers who owned the patents. See The Iron Age, 17 Nov. 1910, 1110. Galambos has observed that it was not surprising that U.S. Steel lagged in moving into organized industrial research:

This awesome holding company brought together most of the large corporations in an industry that had already undergone a long period of horizontal and vertical integration. Initially, the central office did little more than collect standardized statistical reports and prevent member firms from competing with each other. Stabilization of prices was not an automatic consequence of concentration…. It was some years, in fact, before the firm even began to achieve thoroughgoing centralization. One can understand, then, why the U. S. Steel Corporation was not a particularly aggressive firm, at least during this phase of its development. Its managers were content to stabilize the industry even though that meant gradually losing part of their control of the market to smaller and more aggressive rivals. U. S. Steel did not invest heavily in industrial research; it was not an innovative company. It represented the type of industrial combine that was more concerned with centralization and consolidation than with innovations leading to improved productivity in the transitional years through 1919.

See Galambos, “The American Economy and the Reorganization of the Sources of Knowledge,” 276.

37 The Iron Age, 14 April 1927, 1096.

38 Ibid., 2 Aug. 1900, 19; Iron Age acclaimed this innovation in 1906: “[t]he Bethlehem special structural shapes will be an innovation in steel construction, greatly extending the range of application of rolled steel shapes with a simplification of detail and an improvement in structural design.” See The Iron Age, 1 Nov. 1906, 1146. Although the Grey mill had been used successfully in Germany for several years, Bethlehem was the first American company to put the mill into practice. See The Iron Age, 26 Sept. 1907, 831-38.

39 National Research Council, Industrial Research Laboratories of the United States (1946).

40 See ibid. (1921). Note that the listing of research laboratories for this year for U.S. Steel referred the reader to the Carnegie entry; the corporation as a whole had not yet integrated industrial research into its overall operations.

41 The Iron Age, 1 Sept. 1927, 567.

42 Ibid. Iron Age noted that the number of employees for the Carnegie Steel Co., a subsidiary of U.S. Steel, was not included in its total.

43 Ibid., 15 May 1930, 1459.

44 Ibid., 1460. However, G. B. Waterhouse of MIT, in response, stated that “[t]he steel industry is so vast that certain lines of research are not fully appreciated, but rapid advances have been and are being made in tool steels, alloy steel and alloy cast iron, heat treatment and the physical metallurgy of steel…. The main contribution of science to the steel industry has been the cultivation of scientific methods in laboratories and plants, and today technical men occupy many of the authoritative positions. … a great deal of careful scientific work has been carried on and applied and much literature is being published as the result.”

45 See, for example, Chandler, Scale and Scope, 139; Schroeder, The Growth of Major Steel Companies, 112, 211; Brozen, Yale, “R&D Differences among Industries,” in Economics of Research and Development, ed. Tybout, Richard (Columbus, Ohio, 1965), 85–86Google Scholar; Mansfield, “Size of Firms, Market Structure, and Innovation,” 566; Markham, “Market Structure, Business Conduct, and Innovation,” 327; Perazich, George and Field, P. M., Industrial Research and Changing Technology (Philadelphia, Pa., 1940)Google Scholar; and Terleckyj, Nestor, Research and Development: Its Growth and Composition (New York, 1963), 109Google Scholar. See also Adams, Walter, “The Steel Industry,” in The Structure of American Industry: Some Case Studies, ed. Adams, Walter (New York, 1954), 152Google Scholar; and Adams and Mueller, “The Steel Industry” (1990), 73.

46 For the quote, see Adams and Dirlam, “Big Steel, Invention, and Innovation.” For discussion of innovation in technologically advancing industries, see, for example, David Hounshell and John Kenly Smith, Jr., who describe industrial research in the chemicals industry, and in the specific case of Du Pont, in Science and Corporate Strategy: Du Pont R&D, 1902–1980 (New York, 1988)Google Scholar, and Leonard Reich, who describes GE's aggressive move into industrial research in The Making of American Industrial Research: Science and Business at GE and Bell, 1876–1926 (New York, 1985)Google Scholar.

47 See American Society for Testing Materials, Yearbook (Philadelphia, Pa., various years).

48 See National Research Council, Industrial Research Laboratories of the United States (1921). The role of the automobile industry in promoting greater industrial research in steel is discussed in Knoedler, “Early Examples of User-Based Industrial Research.”

49 Mowery and Rosenberg, Technology and the Pursuit of Economic Growth, 30, 40; Rosenberg, “The Commercial Exploitation of Science by American Industry,” 28, 31. See also Bartlett, “The Development of Industrial Research in the United States,” 27-29.

50 See Knoedler, “Backward Linkages to Industrial Research in Steel, 1870–1930,” chaps. 3 and 4.

51 Quotation from “Departments of Physical and Chemical Tests,” n.d. (c. 1913), box 661, Pennsylvania Railroad Papers, Accession 1810, Hagley Museum and Library, Wilmington, Delaware (hereafter PRR Papers); see Knoedler, “Backward Linkages to Industrial Research in Steel, 1870–1930,” chap. 3.

52 See Sellew, William, Steel Rails (New York, 1913), 4Google Scholar.

53 C. D. Young, “The New Physical and Chemical Laboratory of the Pennsylvania Railroad Company,” ASTM, Proceedings (1915), part 2, 349-50.

54 Quotation from “Departments of Chemical and Physical Tests”; Young, “The New Physical and Chemical Laboratory of the Pennsylvania Railroad Company,” 351.

55 See Knoedler, “Backward Linkages to Industrial Research in Steel, 1870-1930,” chap. 3.

56 See Rosenberg, Nathan, Inside the Black Box (New York, 1982), 157Google Scholar, and “The Commercial Exploitation of Science by American Industry,” 30. See also Knoedler, “The American Railway Engineering Association and Improvement in Steel Rails,” unpub. MS.

57 Proceedings of the American Railway Engineering and Maintenance of Way Association 13 (1912): 851–52Google Scholar.

58 Ibid., 852.

59 Ibid., 854.

60 See Pennsylvania Railroad Specifications No. 4-A, for Billet or Bloom Steel for Forgings, 15 Aug. 1893; Pennsylvania Railroad Specifications No. 6, for Crank Pin Steel, 8 Nov. 1897; and Pennsylvania Railroad Specifications No. 5-A, for Carbon Steel Driving Axles, 21 Oct. 1903; box 659, PRR Papers.

61 Only one other railroad reported the establishment of a testing facility before 1900; two others established their labs in the decade following 1900. See National Research Council, Industrial Research Laboratories of the United States (1921). However, a 1911 PRR in-house survey examined the laboratory facilities of other railroads. Comments from these surveys indicate that there was considerable industrial research ongoing at many railroads at the beginning of the second decade of the twentieth century, although the PRR had the largest and most sophisticated lab. See J. D. Meyer to C. D. Young, 16 Nov. 1911; H. B. MacFarland to C. D. Young, 14 Nov. 1911; James H. Gibboney to C. D. Young, 9 Nov. 1911; and “Railroad Chemists and Engineers of Tests”; all box 714, PRR Papers.

62 As the Penn stated in its 1908 specifications, the crucial aspects of improving quality—the manufacturing details that could make better steel—were left to the manufacturer: “Most important of the features of the new specifications is placing more upon the manufacturer the responsibility for the character of the rail produced. The company recognizes that it is merely a purchaser, not a manufacturer. Considerable latitude is, therefore, to be allowed in the methods of manufacture utilized, so long as the result is a sound rail.” See “New Pennsylvania Railroad Specification for Rails” (editorial), The Iron Ape, 13 Feb. 1908.

63 Knoedler, “Backward Linkages to Industrial Research in Steel, 1870–1930,” chap. 3.

64 E. F. Kenney, “The Latest Results with Steel Rails,” The Iron Age, 2 July 1908, 43.

65 See Knoedler, “Backward Linkages to Industrial Research in Steel, 1870–1930,” chap. 2.

66 The Iron Age, 12 Dec. 1907, 1696. Carnegie Steel was at this time part of the U.S. Steel Corporation, but most of the constituent firms in the corporation continued to operate as separate entities. See, for example, Galambos, “The American Economy and the Reorganization of the Sources of Knowledge.” The Iron Age, 30 April 1903, 28. In 1903 the AASM included: American Iron & Steel Manufacturing; American Steel & Wire; American Steel Hoop; Bethlehem Steel; Cambria Steel; Carbon Steel; Carnegie Steel; Central Iron and Steel; Colorado Fuel & Iron; Crucible Steel; Diamond State Steel; Glasgow Iron; Illinois Steel; Inland Steel; Jones & Laughlin Steel; Lackawanna Steel; Lorain Steel; Lukens Iron & Steel; Maryland Steel; National Steel; National Tube; Otis Steel; Passaic Steel; Pennsylvania Steel; Pittsburgh Forge & Iron; Reading Iron; Republic Iron & Steel; A. & P. Roberts; Shelby Steel Tube; Standard Steel Works; Tennessee Coal, Iron & Railroad; Tidewater Steel; and Worth Brothers.

67 The Iron Age, 16 Aug. 1900, 19.

68 Ibid., 15 Jan. 1902, 8.

69 Ibid., 26 June 1913, 1570.

70 Ibid.

71 See Knoedler, “Backward Linkages to Industrial Research in Steel, 1870–1930,” chaps. 3 and 4; Association of American Steel Manufacturers (AASM), Manufacturers' Steel Specifications, Rails (Pittsburgh, Pa., 1906), 2–3Google Scholar. As Charles Dudley stated in 1907, “Multiple tests are pernicious and should be abandoned. Retests, including the sampling, should never be made unless there is reasonable evidence to think that there is an error in the first test…. Shipments of material once fairly rejected, should never be accepted and used if the material is of such a kind that safety or risk to human life is involved.” See ASTM, Proceedings (1907), 35.

72 The Iron Age, 30 April 1903, 28; see Knoedler, “Backward Linkages to Industrial Research in Steel, 1870–1930”; and “Recent Advances in the Standardization of Steel Specifications,” The Iron Age, 30 April 1903, 28-29.

73 “American Research Work on Rails,” The Iron Age, 12 Sept. 1912, 614.

74 The development of a science of metallurgy or “materials science” was a slow process, “by no means completed by the outbreak of the First World War.” See Mowery and Rosenberg, Technology and the Pursuit of Economic Growth, 31.

75 See “Records of the President's Committee on Rail Failures,” box 664, PRR Papers.

76 See The Iron Age, 6 March 1930, 720.

77 See Young, C. D., “Condensed Information Concerning the American Society for Testing Materials” (Philadelphia, Pa., 1916), 1Google Scholar.

78 See The Iron Age, 17 May 1900, 12; and Knoedler, “Backward Linkages to Industrial Research in Steel, 1870–1930,” chap. 4; The Iron Age, 10 May 1900, 12.

79 See, for example, The Iron Age, 25 July 1907, 241; and ASTM, Proceedings (1906).

80 The Iron Age, 16 June 1902, 8; and see Knoedler, “Backward Linkages to Industrial Research in Steel, 1870–1930,” chap. 4.

81 ASTM, Proceedings (1903); and Knoedler, “Backward Linkages to Industrial Research in Steel, 1870-1930.”

82 Of 168 members in 1901, 37 represented steel-producing firms; 13 represented railroads; 39 represented independent testing firms; and 34 were university professors. See American Section, IATM, Proceedings (1902).

83 Membership increased steadily over the next three decades, from 175 in 1902, to 600 in 1905, to 2,100 by 1916, 2,500 by 1919, and 4,200 by 1928. See ASTM, Proceedings (1904), 572; Young, “Condensed Information Concerning the American Society for Testing Materials”; The Iron Age, 26 June 1919, 1748, and 5 July 1928, 28.

84 The Iron Age, 6 July 1905, 13.

85 See ibid., 16 June 1902, 8.

86 See Knoedler, “Backward Linkages to Industrial Research in Steel, 1870–1930.”

87 ASTM, Proceedings (1904), 572. Specifications issued included those for structural steel for bridges, ships, and buildings; for open-hearth steel boiler plate and rivet steel; for steel rails; for steel splice bars; for steel axles and steel tires; for steel forgings and castings; for wrought and foundry pig iron; for cast iron pipe and car wheel; for locomotive cylinders; and for malleable and gray iron castings. See ibid. (1906), 23.

88 ASTM, Yearbook (Philadelphia, Pa., 1901), 235, 245Google Scholar.

89 ASTM, Proceedings (1917), 41-43.

90 The Iron Age, 12 Sept. 1912, 614-15.

91 The proposal was to drop test every fifth lot of steel rather than every lot, thus reducing the probability of rejection based on the drop test. The language had previously stated that test pieces should “preferably” be taken from the top. As steel-consuming firms believed that most of the defects collected at the top of the ingot, they thought that this provision would provide a more certain way of detecting defects. See The Iron Age, 6 July 1905, 17.

92 ibid.

93 ASTM, Proceedings (1905), 163.

94 The Iron Age, 6 July 1905, 17.

95 Ibid., 5 July 1906, 17.

96 ASTM, Proceedings (1907), 39.

97 The Iron Age, 27 June 1907, 1948.

98 ASTM, Proceedings (1907), 113.

99 Ibid., quotations from 89 and 118.

100 Ibid., 94. “… the skilled employees of a rail mill are paid by the ton. … therefore the danger exists that sometime the full running of the mill may be under what can be designated as a strained condition.” See ibid., 115.

101 Ibid., 55.

102 Ibid., quotation, 55; 74.

103 Ibid., 111.

104 Ibid., 39. The members of this subcommittee included Bostwick and Carhart of Carnegie Steel, Kenney and Thackray of Cambria Steel, Dudley of the PRR, Marburg of the University of Pennsylvania, and Webster, an independent consulting engineer.

105 The Iron Age, 27 June 1907, 1955.

106 ASTM, Proceedings (1908), 19-20, 22; quotation, 27.

107 Ibid., 28; “Bessemer metal is, many times, cast in the ingot mold, before the reactions are complete.”

108 Ibid., 35-39, quotations at 34 and 27.

109 Ibid., 74-93, 94-98.

110 Ibid., 99, 106, 108.

111 The Iron Age, 9 July 1908, 116.

112 The technical detail was the presence of manganese sulphide, believed to be responsible for a large number of defects in rails. See The Iron Age, 8 July 1909, 100; quotation, 96.

113 ASTM, Proceedings (1909), 109.

114 Ibid., 77-89. Fay and Wint's specific recommendations included: lower sulphur content, changes in pouring the molten metal into the ingot mold by producers of steel, and giving more attention to electric refining. Job and Dudley, ibid., 90-105; Fay quotation, 106. In defense of the steel producers, Henry Howe, professor at Columbia and long affiliated with Taylor-Wharton, argued that “the practical steel manufacturers may have made more examinations than Mr. Fay is aware of.”

115 The Iron Age, 8 July 1909, 96.

116 Ibid., 4 April 1912, 846.

117 Ibid.

118 ASTM, Proceedings (1912), 27.

119 Ibid. (1913), 582; The Iron Age, 27 Feb. 1913, 539; ibid., 12 March 1914, 680.

120 Ibid., 27 Feb. 1913, 590. However, he made the significant observation that “failures are confined to what may be called defective rails, and … most rails stand up to the service.”

121 ASTM, Proceedings (1914), 89.

122 Ibid. (1915), 34 (quotation); 35. For example, neither AREA nor ASTM specifications addressed the considerable differences between the chemistry of the finished steel in the rails and the chemistry of the ladle analysis. Yet PRR analyses had shown substantial discrepancies in the results of these two forms of chemical analysis, and for that reason, PRR specifications had always detailed not only the requisite chemical composition but also the methods to be used to determine these chemical tests.

123 Ibid., 37. As Gibbs noted, “any deviation from the beaten path is immediately met by an increase in the price asked, the reasonableness of which the purchaser has no means of determining. In at least one case of which I have knowledge, the actual expense of every kind … due to rail breakage on a certain system was less than one-fifth of the annual increase in the cost of rail due to the introduction of a proposed new specification designed to secure the quality of rail desired.”

124 Ibid., 35; quotation, 40.

125 Ibid. (1916), 94.

126 Chemical analysis could be used to identify the extent of segregation, although “the facilities which have been found necessary by the large railroad (the PRR) which has used this method are probably beyond the reach of many smaller railroads.” This was the most reliable method of detection found to date. A second proposed method—estimation of the extent of segregation using Brinell hardness tests—had not yet shown practical results. See The Iron Age, 6 July 1916, 25.

127 ASTM, Proceedings (1917), 113; The Iron Age, 5 July 1917, 11.

128 ASTM, Proceedings (1917), 111-12. This was a method, however, that was not yet perfected. Howard of the Bureau of Standards reported on the bureau's efforts and stated the “necessity of acquiring more information upon the successive stages through which the metal in a rail passes from the time of fabrication until rupture is reached.”

129 Ibid., 112. This report was prepared by P. H. Dudley.

130 Knoedler, “Backward Linkages to Industrial Research in Steel, 1870–1930,” chap. 4.

131 ASTM, Proceedings (1904), 29.

132 The Iron Age, 8 July 1909.

133 Ibid., 5 July 1928, 28; see also ibid., 18 Jan. 1923, 237.

134 See ibid., 6 March 1930, 720.

135 Editorial, The Iron Age, 24 July 1919, 250; ibid., 6 March 1930, 720.

136 Knoedler, “Backward Linkages to Industrial Research in Steel, 1870–1930,” chap. 5.

137 The Iron Age, 22 May 1930, 1536.

138 See Adams, “The Steel Industry,” 152.

139 Some of the more recent literature in the economics of innovation has begun to recognize the importance of users to the innovation process. See for example Jorde, Thomas and Teece, David, eds., Antitrust, Innovation, and Competitiveness (New York, 1992)Google Scholar; Lundvall, Bengt-Ake, Product Innovation and User-Producer Interaction (Aalborg, 1984)Google Scholar; and Eric Von Hippel, “The Dominant Role of Users in the Scientific Instrument Process,” Research Policy, 1976, 212-39.

140 David Mowery has suggested that cooperative research organizations were more numerous in the United States in the years before the Second World War, but that during and after the war, “dramatic changes in the structure of the U.S. research system” caused U.S. firms to rely almost exclusively on intrafirm innovation. See “The Development of Industrial Research in U.S. Manufacturing,” American Economic Review 80 (May 1990): 348. See also Knoedler, “Backward Linkages to Industrial Research in Steel, 1870–1930”; and Knoedler, “Early Examples of User-Based Research.”