¹ The antebellum period is considered in detail in Temin, Peter, “A History of the American Iron and Steel Industry from 1869 to 1900” (unpublished M. I. T. Ph.D. dissertationxs, 1964), where fuller documentation for the post-Civil War period than is given here may also be found.Google Scholar

² See the discussion in the American Institute of Mining Engineers, Memorial of Alexander Lyman Holley (New York, 1884), pp. 171–81Google Scholar, and the references given there.

³ The actual extent of this substitution is unknown, but it was small compared to the substitution between wrought iron and steel. The use of cast iron for construction is described in Giedion, Sigfried, Space, Time and Architecture (4th ed.; Cambridge: Harvard University Press, 1962), pp. 165–206Google Scholar. One use of cast-iron pipe is described in Blake, Nelson Manfred, Water for the Cities (Syracuse: Syracuse University Press, 1956).Google Scholar

⁴ Table 2, following: the last column of the second part.

⁵ The AISA data are contained in a series of annual reports issued for 1867 and 1871 to 1911, mostly under the title of Statistics of the American and Foreign Iron Trades. The product definitions given here are adapted from those in Osborne, A. K., An Encyclopedia of the Iron and Steel Industry (New York: Philosophical Library, Inc., 1956).Google Scholar

⁶ The Census of 1869 listed the quantity of the individual products made, but not the value.

⁷ See Koopmans, Tjalling C., “Identification Problems in Economic Model Construction,” in Hood, Wm. C. and Koopmans, Tjalling C., eds., Studies in Econometric Method, Cowles Commission Monograph No. 14 (New York: John Wiley and Sons, 1953).Google Scholar

⁸ See Temin, A History, ch. v.

⁹ See ibid., ch. vi.

¹⁰ Holley is given this credit in the American Institute of Mining Engineers, Memorial.

¹¹ Data on steel rails were taken from AISA, Statistics, 1904, p. 105. The volume of all rolled-steel products can be computed from Tables 1 and 2, below, for the years after 1889: it is shown in Table 1 for the earlier years. The price of steel rails was deflated by the Warren and Pearson price index in U. S. Bureau of the Census, Historical Statistics of the United States, Colonial Times to 1957 (Washington, 1960), p. 115.Google Scholar

¹² The volumes of production of wrought iron in 1899 and in 1909 as calculated from the data in Tables 1 and 2, are about 2,100 and 1,700 thousand gross tons respectively. The price of wrought-iron bars fell from $82 in 1869 to $43 in 1889. This is a fall of 46 per cent (as opposed to the 78 per cent fall in the price of steel rails), and it is exactly balanced by a fall of the same amount in the Warren and Pearson price index. The price of wrought-iron bars, in other words, did not fall relative to the general level of prices in the economy. The price of wrought-iron bars was taken from AISA, Statistics, 1904, p. 127.

¹³ Clark, Victor S., History of Manufactures in the United States (New York: McGraw-Hill, 1929), II, 260.Google Scholar

¹⁴ ibid., II, 261–63, 274–76; III, 84–85, 118–22.

¹⁵ The prices of steel products may be seen in any of the AISA, Statistics, for this period.

¹⁶ See Temin, A History, ch. viii. The AISA, Directory to the Iron and Steel Works of the United States, published every other year during this period, shows the large number of entrants to the industry in the 1880's. It must be remembered that the scale of firms and the tariff enabled the steel manufacturers to keep the price of steel relatively high. An elastic supply curve does not indicate zero profits for existing large-scale producers.

¹⁷ All the quantitative results used from here on are taken from Table 2.

¹⁸ See U. S. Bureau of the Census, Historical Statistics, pp. 427–29.

¹⁹ See Clark, History of Manufactures, II, 348–55; III, 122–27.

²⁰ ibid., II, 347.

²¹ The note in the Census that said this was appended only to the data for iron, but the discussion in the text referred to steel. In addition, the magnitude of “otherproduct” production (close to 30 per cent of the total) shown by the uncorrected Census data was unrealistic.

²² The share of “other products” is a residual in all cases.

²³ The data on rails came from AISA, Statistics, 1904, p. 105, and the data on wire rods were taken from the same source, p. 98.

²⁴ AISA, Statistics, 1887, p. 33; 1896, p. 72; 1904, p. 60; 1905, p. 68; 1906, p. 69; 1907, p. 75; 1909, p. 89; 1910, p. 87; 1911, II, 16.

²⁵ Total rolled products were reported for 1899 (AISA, Statistics, 1904, p. 98), but they were not broken down by material. An estimate of rolled-steel products was constructed by regressing the quantity of rolled-steel products on the quantity of steel ingots and castings produced for those years (1886–90, 1904–11) for which rolled steel was reported. The data on rolled steel came from the sources listed in the previous footnote, and the data on ingots and castings came from AISA, Statistics, 1911, II, 61. The regression result was:

where S is rolled steel, I is ingots and castings, and the standard errors are shown below the estimates. (R² for the regression was .99.) Rolled steel was subtracted from total rolled products to get rolled iron, and the proportion of the total represented by rolled iron was calculated.

²⁶ The formulas are as follows, where P_I is the quantity of the product under discussion that was made of iron; P_s, the quantity of the product made of steel; T_I, the total amount of rolled iron; and T_s, the total amount of rolled steel.