¹ The Young Steam Engineer's Guide (Philadelphia [1826]), p. 1Google Scholar. This book was a reissue of Evans' Abortion of the Young Steam Engineer's Guide, first published in 1805.

² See Roe, Joseph Wickham, English and American Toolbuilders (New York: McGraw-Hill, 1926), chs. i, iiGoogle Scholar.

³ W[inram], H[enry] Dickinson, A Short History of the Steam Engine (Cambridge, Engl.: The University Press, 1939), pp. 95–97Google Scholar. Evans was called by some “The Watt of America.” Thurston, Robert H., A History of the Growth of the Steam-Engine (Ithaca: Cornell University Press, 1939), p. 159Google Scholar.

⁴ Dickinson, pp. 95–97.

⁵ Journal of the Statistical Society of London (later, Journal of the Royal Statistical Society), II (Jan. 1840), 440Google Scholar; U. S. Congress, House, Report on the Steam Engines in the United States, H. Doc. No. 21, 25th Cong., 3d sess., 1838. Referred to hereafter as Report on Steam Engines. Only 65 out of 305 stationary steam engines in Birmingham were high pressure, while only 63 out of over 1,200 stationary steam engines in the United States were low pressure in 1838.

⁶ There were at least 10 steam engines built in America before Evans introduced his, and his engine was pirated extensively in the years before 1825. In fact, the opposition to Evans was probably a reaction to his attempted patent monopoly and his litigious personality, rather than to his invention: a reflection of the demand for the high-pressure engine, not of its lack. See Bishop, J. Leander, A History of American Manufactures from 1608 to 1860 (3 vols.; Philadelphia, 1866), I, 502, 510, 534, 547, 577Google Scholar; Victor Clark, S., History of Manufactures in the United States (3 vols.; New York: McGraw-Hill, 1929), I, 408–9Google Scholar; Greville, and Bathe, Dorothy, Oliver Evans: A Chronicle of Early American Engineering (Philadelphia: Historical Society of Pa., 1935), pp. 207, 265Google Scholar. Evans' patent expired in 1825; ibid., pp. 220–21.

⁷ Clark, I, 409; U. S. Congress, House, Documents Relative to the Manufactures in the United States, collected by the Secretary of the Treasury, Louis McLane; Exec. Doc. No. 308, 22d Cong., 1st sess., 1833; Report on Steam Engines.

⁸ Allen H. Fenichel, “Growth and Diffusion of Power in Manufacturing, 1838–1919” (paper delivered to the Conference on Research in Income and Wealth, Chapel Hill, 1963), uses the Report on Steam Engines for an estimate of steam power in manufacturing with no additions.

⁹ It is possible that some of them were sold and that the same engine sometimes appears in both reports under different ownership. This is occasionally recorded in the Report on Steam Engines (e.g., p. 294), but it does not appear to have been common.

¹⁰ Bathe, pp. 207, 278–79; Report on Steam Engines, pp. 192, 215. We do not know that Rush & Muhlenberg were making engines in the early 1820's, our information on the firm being derived from the 1838 data. But it would be surprising if the firm maintained its existence through the 1820's while building only a single steam engine in its first six years. It made three steamboat engines in 1826–27. The Report on Steam Engines (p. 165) lists a second engine built by Oliver Evans, but as the date of its construction is given as 1830 the attribution is uncertain.

¹¹ Dickinson, p. 106. It should be noted that the average age of the low-pressure steam engines shown in Table 2 is higher than that of U. S. engines as a whole.

¹² These data were calculated from the Report on Steam Engines.

¹³ See North, Douglass C., The Economic Growth of the United States, 1790–1860 (Englewood Cliffs, N.J.: Prentice-Hall, 1961), pp. 166–68Google Scholar, and, for machinery, Roe, p. 109.

¹⁴ The precise numbers in Table 1 depend on the definitions of regions, but the conclusions do not. The regional breakdown used here differs from that used elsewhere primarily in its division of several states into two regions. This results from the use of the Allegheny Mountains as the division between the Middle Atlantic and the western regions and the allocation of what is now West Virginia to the West rather than to the South. As the mountains were a far greater barrier to trade than state boundaries, this division is preferable to the more usual one. (See the notes to Table 1 for the possibility that the difference between the South and the other regions was exaggerated.)

¹⁵ North, pp. 159–62; Rosenberg, Nathan, “Technological Change in the Machine Tool Industry, 1840–1910,” Journal of Economic History, XXIII (Dec. 1963), 418–20Google Scholar.

¹⁶ Eighth Census of the United States, III, Manufactures of the United States in 1860 (Washington, 1865), 738Google Scholar.

¹⁷ The South is only a possible exception, as it may have been more profitable to use imported steam engines in the South than to build them. The pattern of trade shown in Table 1 would then have been the result of southern exploitation of their comparative advantage, not a reflection of their lack of knowledge.

¹⁸ Report on Steam Engines. There were nine American builders who made more than one low-pressure engine and fourteen who made only one. There were also, of course, some English builders making high-pressure engines; 65 such engines existed in Birmingham in 1838. Journal of the Statistical Society of London, II (1840), 440Google Scholar.

¹⁹ Evans, The Young Steam Engineer's Guide, pp. 5–11.

²⁰ Lardner, Dionysius, The Steam Engine Familiarly Explained and Illustrated (3d American ed., from the 5th London ed.; Philadelphia, 1836), pp. 279–80Google Scholar; Renwick, James, Treatise on the Steam Engine (2d ed.; New York, 1839), pp. 161–63Google Scholar.

²¹ See Kuhn, Thomas S., “Energy Conservation as an Example of Simultaneous Discovery,” in Clagett, Marshall, Critical Problems in the History of Science (Madison: University of Wisconsin Press, 1959), pp. 321–56Google Scholar.

²² For the range of efficiency as measured by coal consumption see the discussion in Part II below. The Philadelphia Water Works had a Boulton and Watt engine and an Evans engine, which they compared in the period 1815–22. The two engines operated the same diameter pump and used roughly the same amount of fuel per gallon of water pumped. The relative capital costs, etc., are not known; Bathe, p. 227. Clark, I, 409, asserts that high-pressure engines were more wasteful of fuel than low-pressure engines, and that they were used for that reason in the United States. This seems to be an inference from the “loss” of the heat of vaporization in the high-pressure steam engine; see also Hunter, Louis C., Steamboats on the Western Rivers (Cambridge: Harvard University Press, 1949), pp. 132–33Google Scholar.

The problem of comparison is complicated by the fact that the distinction between high- and low-pressure engines was not as sharp as has been implied here. The distinguishing characteristic of a low-pressure engine was the possession of a condenser, but a wide variety of engines had condensers. The Cornish engines in Britain used condensers and high-pressure steam (Dickinson, p. 104).

In addition, high-pressure engines were smaller and cheaper than low-pressure engines (Hunter, pp. 129–30). This may have been partially offset by more rapid depreciation of high-pressure engines, as our discussion has suggested, and its significance is not clear.

²³ Report on Steam Engines; Hunter, pp. 130–33. Hunter also says high-pressure engines used more fuel.

²⁴ The classification of builders who had made five or more engines of a particular type as “important” builders of that type is an arbitrary classification designed to eliminate the case of major builders of one type of engine who had also made one or two engines of another type.

²⁵ Dickinson, p. 91; Cardwell, D. S. L., The Organisation of Science in England: A Retrospect (London: Heinemann, 1957), pp. 80–81Google Scholar; Hunter, pp. 132–33, 289–304.

²⁶ Clark, I, 410, asserted that steam power was five times as expensive as water-power about 1840. He did not take into account the differences between the two sources of power, however, and his data cannot be taken as evidence of a disequilibrium in this market.

²⁷ The cost of transportation shown in Table 4 was divided approximately equally between the cost of transporting raw cotton and the finished textiles and the cost of bringing coal for heat. The costs of transporting the cotton seem to have been understated by James, as a result of an understatement of the cotton used per spindle.

This is offset, however, by a possible overstatement of the cost of heat. Users of steam power could use the steam for heat if they used high-pressure engines, but there is no evidence that this was widespread. James' mills did not do so, and the extra cost for heat incurred by using waterpower may have been a cost in theory only. See Montgomery, p. 157; Merriam, J. C., “Steam” in Eighty Years' Progress of the United States (Hartford, 1869), p. 253Google Scholar; De Bow's Review, VII (1849), pp. 128–34Google Scholar.

²⁸ The annual cost figure for coal consumption in Table 4 implies an hourly coal consumption of 3–4 pounds per horsepower (at the following rates: $6.50 per ton of coal; 2,000 pounds per ton; 300 working days per year; 10–11 working hours per day). The range of coal consumption in England and America at this time was from about 2 to about 10 pounds per horsepower per hour, reflecting the diversity of engines and practice about 1840. Evans, p. 60: P[aul] Hodge, R., The Steam Engine, Its Origin and Gradual Improvement (New York, 1840), pp. 119–20Google Scholar; Journal of the Franklin Institute, n.s., XXV (1840), p. 342; James (see Table 4, Sources), pp. 66–68.

The practice assumed by James was good, but not the best; it appears to be the rate his own mills were getting. The James Steam Mills were using about 4 pounds in the early 1850's-it is possible that they were using almost as little ten years previously. They used 5.5 pounds of coal per horsepower per hour for both heat and power. As a condensing engine was used, a deduction for the coal used for heating must be made to discover the coal consumption for power alone. Assuming the ratio of fuel for power and for heat shown in Tame 4 held in the 1850's, the 4-pound rate can be derived. The engine in question was replaced in 1855 by a Corliss engine using about half as much fuel (Merriam, p. 253).

²⁹ The allowance for depreciation includes both the direct charge for depreciation and a reduction of the capital costs to take account of the declining balance of invested funds. If both steam engines and waterwheels depreciated at a linear rate of 10 per cent per year, the annual capital costs of the two sources at a 6 per cent interest rate would fall from 60 per cent of total costs for water and 20 per cent for steam to 40 and 10 per cent respectively.

³⁰ See Fenichel, Table B-8, for data on total power usage by industry in 1869.

³¹ Report on Steam Engines, p. 379; Robert Gallman, “Gross National Product in the United States, 1834–1909” (paper delivered to the Conference on Research in Income and Wealth, Chapel Hill, 1963).

³² This result was obtained by dividing the ratio of horsepower used per dollar of value added for the industry as a whole by the ratio to steam-powered cotton mills alone. The use of power in steam-powered cotton mills was derived as the product of the power used per spindle and the value added per spindle (Montgomery, pp. 124–25, 214–16).

Woolen textiles represented a significant proportion of the textile industry, and their presence renders the above calculation even more speculative than it would be for the cotton industry alone. If woolen mills did not use any steam power, the conclusion in the text is unaffected. If they used some steam power, the figure in the text is too low (high) if the horsepower used in woolen milk per dollar of value added was lower (higher) than in cotton mills. In addition, water-powered mills used more power per spindle than steam-powered mills (due to the use of a different type of spindle), and the proportion of power used in the textile industry in 1838 that was generated by steam was only about 10 per cent. See the note to Table 4.

³³ The average coal consumption per horsepower declined steadily in the first half of the nineteenth century, while the technology of waterpower utilization remained relatively stable up to the introduction of the water turbine after 1840. In addition, waterpower in New England was beginning to become scarce after about 1830. See Hodge, p. 120; Clark, I, 404.

³⁴ The ratio of value added per employee was about one third (36 per cent) more for the iron industry than for industry as a whole in 1859, and this may be projected back to 1838 (Taylor, George Rogers, The Transportation Revolution, 1815–1860 [New York: Holt, Rinehart and Winston, 1951], p. 243Google Scholar; Lebergott, Stanley, Manpower in Economic Growth [New York: McGraw-Hill, 1964], p. 510)Google Scholar. Multiplying this ratio by the ratio in the last column of Table 5 gives the ratio of steam power per employee in the iron industry to the steam power per employee in industry as a whole. The amount of steam power used in 1838 amounted to. 072 horsepower per employee for manufacturing as a whole, implying that the ratio was about. 12 for the iron industry (Report on Steam Engines, p. 379; Lebergott, p. 510). In a steam-powered rolling mill, about. 85 horsepower per employee was used; in a steam-power blast furnace, about. 37 (horsepower data from the Pennsylvania returns in the Report on Steam Engines; employment data from Temin, Peter, Iron and Steel in Nineteenth-Century America [Cambridge: M.I.T. Press, 1964], pp. 86–87, 108)Google Scholar. I assume that the amount of power used per employee was independent of the nature of the power used. Then, if the iron industry had been composed entirely of rolling mills, steam power would have been about 15 per cent of the total power used; if the industry had been only blast furnaces, about 30 per cent. As employment was probably somewhat lower in rolling mills than in blast furnaces, the ratio for the industry as a whole was near 25 per cent. (Less wrought iron than pig iron was produced at this time, and output per employee was similar in the two branches of the industry. See Temin, pp. 25–28, 86–87, 108.)

As the amount of power used per employee varied in the cotton industry according to the type of power used, there was a divergence between the proportion of power used that was generated from steam (about 10 per cent) and the proportion of the industry's value added that was produced in steam-powered plants (about 15 per cent); The 25 per cent figure for the iron industry is the proportion of employees who worked in steam-powered plants; I have assumed this was equal to the proportion of the industry's power derived from steam.

³⁵ See Temin, Part I. By 1869, the iron industry used a far higher proportion of steam power than the cotton industry (Fenichel, Table B-13).

³⁶ The west-south-central region used power primarily for sugar mills in 1838. This region accounted for 7,800 horsepower, while trie horsepower allocated to the foodproducts sector as a whole was only 9,800. The estimate in the text is thus very conservative (Fenichel, tables A-1, A-2).

³⁷ De Bow's Review, I (1846), 53–55Google Scholar; Silliman, B., Manual on the Cultivation of the Sugar Cane (Washington, 1833), pp. 31, 46Google Scholar.

³⁸ Porter, George R., The Nature and Properties of Sugar Cane (London, 1830), pp. 172–73Google Scholar; Wik, Reynold M., Steam Power on the American Farm (Philadelphia: University of Pennsylvania Press, 1953), p. 6CrossRefGoogle Scholar.

³⁹ Hollerith, Herman, “The Statistics of Power Used in Manufactures” in Tenth Census of the United States, II, Manufactures (Washington, 1883), 496Google Scholar.

⁴⁰ Bathe, pp. 159–67; Evans, Oliver, Young Millwright and Miller's Guide (Philadelphia, 1860 edition)Google Scholar; Kuhknann, Charles Byron, The Development of the Flour-Milling Industry in the United States (Boston: Houghton Mifflin, 1929), pp. 80–81, 96Google Scholar. The parallel between the Pittsburgh Steam Mill and the Albion Steam Corn Mill, built in London by Boulton and Watt in 1786, is quite good. See Lord, John, Capital and Steampower, 1750–1800 (London: P. S. King & Son, 1923), pp. 131, 162–66Google Scholar.

⁴¹ This includes Fenichel's large estimated use of steam power for lumber products in this region.

⁴² The average size of steam engines in this group was under 10 horsepower, compared to an average of about 20 for all manufacturing and over 100 for iron rolling mills. The industries in this group included—in order of the amount of steam power they used in 1838—foundries and machine shops, leather and tanning, paper and publishing, grinding white lead, etc. (Report on Steam Engines).

⁴³ “Coal and steam, therefore, did not make the Industrial Revolution; but they permitted its extraordinary development and diffusion.” Landes, David S., “Technical Change and Development in Western Europe, 1750–1914,” in The Cambridge Economic History of Europe, VI, Part I (Cambridge, Engl.: The University Press, 1965) 329Google Scholar.

⁴⁴ For a recent statement, see Habakkuk, H. J., American and British Technology in the Nineteenth Century (Cambridge, Engl.: The University Press, 1962)Google Scholar. For critical discussion, see Peter Temin, “Labor Scarcity and the Problem of American Industrial Efficiency in the 1850's,” forthcoming.

⁴⁵ Homer, Sidney, A History of Interest Rates (New Brunswick, N.J.: Rutgers University Press, 1963), pp. 195–96, 286–87Google Scholar. Interest rates in the eastern United States were higher than those in Britain, and interest rates in the western United States were higher than those in the East.