M.I.T. MONOGRAPHS IN ECONOMICS
An Economic Inquiry
THE M.I.T. PRESS
Massachusetts Institute of Technology
Copyright 1964 by The Massachusetts Institute of Technology
THE NEW IRON INDUSTRY,
Before The New Era
The first attempt by Europeans to produce iron in the United States was initiated in 1619. Unfortunately, this enterprise was destroyed by Indians in 1622, probably before it made any iron. A more successful attempt took place in the Massachusetts Bay Colony in 1645, and it is from this event that we may date the American iron industry.1
By 1700 the American colonies were producing about 1,500 tons of iron annually, or something under 2 per cent of the probable world production and perhaps about 10 per cent of the probable British production. The British industry was based on imported iron from Sweden and elsewhere, even though imports of iron were taxed heavily. The first regulation to differentiate between American iron and other imports was an act of 1750, which (as amended in 1757) provided that American iron could be admitted to England free of duty. As this act was an attempt to increase the supplies of raw material for the English industry without destroying the colonial market for its products, it carried with it the restriction that no new ironworks producing finished products could be built in the colonies after June 24, 1750. The prohibited classes of ironworks included slitting mills, which made nails, plating mills, which made hammered sheets and tin‑plate iron, and steel furnaces, which made blister steel.
Despite several changes in the details of the laws about the colonial iron industry, their intent remained faithful to the act of 1750. Nevertheless, the British regulatory policy was only moderately successful. While imports from the colonies rose, they never attained a level of high significance. At their peak in 1771, American exports of bar and pig iron were about 2,000 and 5,000 tons, respectively, compared with British imports of 43,000 tons of bar iron from other sources (but no other imports of the far less valuable pig iron). In addition, American production of iron rose from about 10,000 to about 30,000 tons in the same period. This iron no doubt found its way into consumption within the colonies by way of ironworks prohibited by colonial law. The inability of the English legislators to determine the direction of the colonial industry can be attributed to their internal conflicts of interest, but it also derived from their inadequate knowledge of the colonies and lack of appreciation for the rapidly growing colonial demand. In any case, the production of iron in America on the eve of the Revolution may have been as high as 15 per cent of world production, a proportion that was not again equaled until well into the nineteenth century. And while this high production was largely for domestic consumption, the colonies were net exporters of unfinished iron.2
In the last quarter of the eighteenth century the British iron industry went through a technological revolution that abruptly altered its relation to the world and to the American iron industry. This revolution started with the introduction of coke as a blast‑furnace fuel, replacing the time‑honored fuel of charcoal. The use of coke may have been proven a technological possibility as early as 1709 by Abraham Darby, but its general commercial feasibility had to wait upon further developments of technique, taking another half‑century. Once the skill of making iron with coke had been advanced to the proper level, however, the transformation of the British iron industry was rapid. In 1788 only 24 out of 77 pig iron furnaces in Britain used charcoal; in 1806, only 11 out of 173.3 The new technology not only removed England from dependence on its depleted forests and enabled it to exploit its abundant coal, it also opened the way for further improvements in making iron. Britain changed from a high‑cost producer to a low‑cost producer and, as a result, from a net importer to a net exporter of iron.
The United States, on the other hand, continued to employ the old technology. It became a high‑cost producer and, despite the imposition of tariffs, a net importer of iron. Nevertheless, the iron industry in this country grew until in 1830 the production of pig iron was close to 200,000 tons. This figure, the product of a free trade convention in Philadelphia and a protectionist convention in New York, represents actual data for about 45 per cent of the furnaces "reported" and imputed production for the rest. But as there was no census of manufactures in 1830, it is the best we have.4
It would be unwise to make a strict statement about the rate of growth in the years before and immediately after 1830. The extant data are too fragmentary and come at too infrequent dates for a cited growth rate to have much meaning. But it may not be too misleading to suggest that before the 1840's, for which more exact data have survived, the iron industry of the United States was expanding only slightly more rapidly than the population. If the census figure of 54,000 tons of pig iron produced in 1810 can be believed, the per capita production of iron in the United States did not rise between the Revolution and that date.5 In any case, the iron industry, like the economy, was only making its first tentative steps toward expansion. It was still using the old techniques and the traditional fuel, charcoal, inherited from previous centuries, much as the economy was largely concentrated in its familiar location. The new techniques that were being brought over from England had all the hazards of a new frontier, and as Leland Jenks says of the frontier in 1830, "It was a speculation, not a going concern.6
Before considering the "speculation" in new ironmaking techniques that occupied the American iron industry after 1830, we must ask how the industry arrived in the situation in which it found itself at that time. The American iron industry in 1830 was still operating almost exclusively on the basis of a traditional technology, despite the very successful exploitation of a newer technology in England, an exploitation that threatened to ruin the American industry if unrestricted trade were permitted.
Why was the American industry trying to protect itself against the foreign methods of making iron in preference to importing them?7 There is not sufficient information on costs to answer this question fully, but an examination of the technology in question provides a plausible explanation.
Most of the iron used at this time was in the form of wrought iron, often called bar iron after the shape in which it was most commonly sold. Wrought iron could be forged into different shapes due to the relatively pure state of the iron of which it was composed. It was made, therefore, by removing all elements but iron from iron ore. The primary offending element was oxygen, and it had been discovered long before the nineteenth century that wrought iron could be made most profitably in a two‑stage process. The direct process of making wrought iron from ore did not die out until well into the nineteenth century, but the products of "bloomary forges" which made iron in this way were already only of minor importance in 1830.8
The first step in the indirect manufacture of wrought iron utilized a blast furnace, a container in which iron ore and fuel were burned. The oxygen was banished from the iron ore, but its place was taken by part of the carbon from the fuel. The resultant alloy of iron and carbon was not malleable, although it had a relatively low melting point and could be cast into different shapes. This alloy was known as pig iron when it was used as an intermediate product and as cast iron when it was used in iron castings.
The second step in this indirect process was to expel the carbon from pig iron to make wrought iron. This process can be accomplished by the use of great heat, which causes the desired chemical reactions to take place, but the technology of the early nineteenth century did not have the capability of containing large quantities of metal at this heat or the knowledge of its effects. Mechanical labor can be substituted for heat in the refining of pig iron into wrought iron, and the method of refining that the nineteenth century inherited depended upon this substitute. The iron alternately was placed in a refinery fire and beaten with a hammer on an anvil. The forges in which this process was carried out could not process large quantities of iron, and the process required a large quantity of labor to compensate for the small amount of heat that the iron received in its frequent trips to and from the fire.
In the course of the eighteenth century, inventors in England changed both steps in the indirect process of making wrought iron.9 The traditional fuel for the blast furnace –charcoal - was replaced by coke, and the traditional method of refining pig iron by hammering was replaced by a method known as puddling. These innovations helped the English iron ‑industry; the question before us is why they were neglected by American ironmasters. The question of fuel for the blast furnace will be postponed until the detailed discussion of Chapter 3. The point of interest here is that the form in which the new fuel was introduced in England was not suitable for use in America; auxiliary innovations were needed, of which the most notable was the hot blast. Differences between the known resources in the two countries resulted in differences in relative prices, and auxiliary innovations were needed to make the English innovations profitable at American prices.
The process of puddling involved the use of a reverberatory furnace. In this type of furnace, the iron was separated from the fuel by a low wall, and the flames from the burning fuel were conducted over the iron. The lack of contact between the fuel and iron meant that a more impure fuel than the traditional charcoal could be used without contaminating the metal. This fuel was coke: too impure for use in refining by hammering, but not for use in puddling.10 And the separation of the iron and the fire enabled the iron to be manipulated without removing it from the heat. The puddler operated upon the iron through ports in the wall of the furnace. His labor was commuted from lifting and hammering to a sort of stirring activity: puddling.
The increase in heat obtained by use of a reverberatory furnace was enough to dispense with the labor of hammering, and it was discovered that the use of hammers to shape the resultant wrought iron could be dispensed with also. The replacement for hammers was grooved rolls, which could be used to make many shapes of uniform cross section. The use of grooved rolls was closely associated with that of puddling, and the two innovations were usually adopted together in the United States.11
But although the increase in heat obtained by leaving the iron in the furnace continuously was enough to reduce the labor required to refine it, it was not enough to keep the iron in a molten state. The melting point of wrought iron is higher than that of cast iron, and as the transformation to the former took place, the iron changed from a liquid into a pasty substance. Because puddled iron could not be kept molten, it could not be rid of the bits and pieces of nonmetallic slag that it picked up in the course of puddling. Much of the slag was expelled by “squeezing" the iron after it was removed from the furnace, but a non-homogeneous composition remained a hallmark of wrought iron.
Puddling and rolling saved labor and fuel costs. The saving was enough to induce their general adoption in England, but not in America where costs were different. The problem was that puddling had a few offsetting defects that had to be balanced against its cost-saving attributes. One of these difficulties was the large wastage in the original process, equal to about one‑half of the iron put into the furnace. This was partly due to the use of a sand bottom in the puddling furnace, following the practice of casting iron in sand. In 1818, the practice of using an iron bottom in the puddling furnace was introduced, which greatly reduced the amount of waste in puddling (to about seven hundredweights per ton of bar iron produced), increased the output of the furnace, and even improved the quality of the resultant iron. 12 This innovation permitted the puddling process to be used to advantage in America, both by reducing costs of puddling and by increasing its ability to use inferior irons. It took a few years for news of the innovation to spread to America, but by 1830 the stage was set for an improvement in the competitive position of the American iron industry.
I James M. Swank, History of the Manufacture of Iron in All Ages (second edition; Philadelphia: American Iron and Steel Association, 1892), pp. 103‑112.
2 The preceding discussion has been adapted from the exposition in Arthur Cecil Bining, British Regulation of the Colonial Iron Industry (Philadelphia: University of Pennsylvania Press, 1933).
3 H. R. Schubert, History of the British Iron and Steel Industry from c. 450 B.C. to A.D. 1775 (London: Routledge and Kegan Paul, 1957), p. 335; Thomas S. Ashton, Iron and Steel in the Industrial Revolution (Manchester: At the University Press, 1924), p. 99; John Gloag and Derek Bridgwater, A History of Cast Iron in Architecture (London: George Allen and Unwin, 1948), p. 42.
4 For a complete discussion of rhese figures, see Appendix A.
5 Appendix C, Table C. 1; Appendix A.
6 Leland Hamilton Jenks, The Migration ofBritish Capital to 1875 (New York: Alfred A. Knopf, 1927), p. 73.
7 This is the opposite question to that considered by H. J. Habakkuk in American and British Technology in the Nineteenth Century (Cambridge, England: At the University Press, 1962). Habakkuk tries to explain why American technology in manufacturing industries was ahead of British technology in the first half of the nineteenth century. The iron industry, however, was not ahead of Britain in technological development at this time, and the discussion here is based on that fact. The discrepancy between these two treatments appears to have arisen because Habakkuk took his cue from the light manufacturing industries of New England which differed in their technological development from the heavier industries of the Middle Atlantic states.
8 See the discussion of wrought‑iron production later in this chapter.
9 See Ashton for a detailed description of the innovations and their invention.
10 Louis C. Hunter, "The Influence of the Market upon Technique in the Iron Industry in Western Pennsylvania Up to 1860," Journal of Economic and Business History, 1 (February, 1929), 241‑281.
11 This refers to the general adoption of puddling and rolling, not their initial introduction. See Chapter 5.
12 Harry Scrivenor, History of the Iron Trade (London, 1854), p. 252; J. C. Carr and W. Taplin, History of the British Steel Industry (Cambridge: Harvard University Press, 1962), p. 98n. The use of iron replaced an "acid" lining by a neutral one; this permitted phosphorus, a troublesome trace element, to be drawn off in the slag. See the discussion of linings in Chapter 7.
13 See Ashton, pp. 54‑59, for a discussion of the introduction of crucible steelmaking in eighteenth‑century England.
Excerpt – Pages 57 - 62
We may now turn from our discussion of previous explanations of the delay in adopting coke to the problem itself. We shall attempt to derive a direct estimate of costs in order to see if relative prices before about 1860 made the production of iron with coke unprofitable. If so, we may ask why this price structure existed and whether it was changes on the supply side or on the demand side that altered the profitability of making coke pig iron in the 1850's. If it was not unprofitable to use coke, we must reexamine our reasoning in the hope of discovering why coke was not adopted before it was. One more preliminary task remains, however, before we can pass on to the consideration of costs and profits; we must understand the technology in question before we can understand the structure of costs.
The typical blast furnace of 1830, nearly indistinguishable from its counterpart of fifty years earlier, was built of masonry. It was about 30 feet high, 20 feet along each side of its square base, and it sloped inward slightly as it rose. The interior of the furnace measured only about nine feet across at its widest point (the "bosh") and was lined with sandstone or soapstone, the forerunners of refractory brick. Iron ore, charcoal, and limestone were thrown into the open top of the furnace by men who had carried baskets of these materials across a bridge built from a rise in the ground to the top of the furnace. Air was pumped into the furnace near the bottom by means of an opening (the "tuyere") made of cast iron or brass and not yet protected by water-cooling. The air was supplied by bellows or wooden tubs powered by a water wheel. This blowing apparatus was shortly to be superseded by machines made of cast iron, often employing steam engines for power. The pressure produced by the blowing engines was only about one or one-and-one-half pounds per square inch; stronger pressure was not desired for fear of "blowing the charcoal to pieces" 10
The first major change in American blast‑furnace practice was the introduction of the hot blast, first introduced in England in 1828. This modification of blast-furnace technique was based on a very simple idea. The furnace was operated by blowing a blast of air into a column of ore and fuel in order to produce combustion. The materials had to be heated to burn properly, and part of the energy in the fuel was being used to heat the blast-furnace charge and the incoming air to the point where the desired chemical reactions could take place. If the blast was already heated before it entered the furnace, it would heat the charge and lessen the amount of fuel necessary to smelt the iron.
This innovation could be attached to any existing blast furnace. In its simplest form, it was just a set of pipes over a fire in which the blast was heated on its way to the furnace. A more economical version, because it did not use any extra fuel, was to place the pipes through which the blast passed at the top of the blast furnace where they could be heated by the combustion of the waste gases from the furnace itself. It was in this form, and more efficient later modifications, that the innovation was generally adopted in America. Once the usefulness of waste furnace gases was seen, they were employed for making steam to provide power for a variety of machines, including the blowing engines.
News of the hot blast was available in the New World with little delay. Articles appeared on savings in England and on the necessary machinery, and in 1834 a New Jersey blast furnace adopted the first practical hot blast in the United States. As modified in the following year to use waste gases to heat the blast, the innovation saved the furnace 40 per cent on its fuel costs." But despite the impression created by savings such as these and by the articles claiming savings up to 75 per cent for the Clyde Iron Works in Britain, the hot blast was not a universal benefit. Overman summarized the incidence of its effects:
The economical advantages arising from the application of hot blast, casting aside those cases in which cold blast will not work at all, are immense. The amount of fuel saved, in anthracite and coke furnaces varies from thirty to sixty per cent. In addition to this, hot blast enables us to obtain nearly twice the quantity of iron within a given time than we should realize by cold blast. These advantages are far more striking with respect to anthracite coal than in relation to coke, or bituminous coal. By using hard charcoal, we can save twenty per cent of fuel, and augment the product fifty per cent. From soft charcoal we shall derive but little benefit, at least where it is necessary to take the quality of iron into consideration.12
The primary interest of the hot blast is in those areas where its effect was the greatest, that is, in connection with mineral fuel. In fact, before the application of the hot blast to the smelting of iron with anthracite, it was not practical to use that fuel at all. In such a case, the effects of the hot blast are difficult to separate from the effects of the use of mineral fuel, and the two will be treated together. But the hot blast alone is of some interest.
The convention of 1849, which collected so much useful information, classified the charcoal furnaces listed by the nature of the blast. From this, two things emerge. There is no difference in the percentage of old and new firms which adopted the hot blast, demonstrating that the hot-blast apparatus could be added to old furnaces without much trouble. But there is a sizable difference in the extent to which the hot blast was adopted in different regions of Pennsylvania. If the state is split into three - the East, the Juniata Valley, a region in the mountains which produced high‑quality iron for sale in. both Philadelphia and Pittsburgh, and the remainder of the West. We find that in the first two regions over half of the charcoal furnaces used the hot blast in 1849, but that in the last region only one‑fifth did so. 13 We may ask why this differential exists.
The cost of adopting the innovation was not high, as was argued above and as can be seen from a direct examination of the convention's data. The increased capital costs were less than 10 per cent of total capital costs, and were more than compensated for by the increased output resulting from the hot blast.14 The extant statistics are not complete enough to allow a detailed cost calculation in light of the variable nature of this innovation, but we may isolate two possibilities why the hot blast was not used in the West as much as in the East, paralleling the two possibilities that will emerge from our discussion of coke. The raw materials in the West may not have been suitable for the hot blast, that is, there may have been the wrong kind of wood to make charcoal, or there may have been a difference in demand that deterred the use of this innovation in the West.
Quality differences existed between some hot‑blast and cold-blast iron, and continuing discussion of quality problems supports the contention these differences persisted over time.15 The extent to which they were reflected in prices, however, remains obscure. When the process is specified, which is far from universal, only a slight differential is detectable.16 In addition, as the passage from Overman suggests, this quality differential was related to the quality of the charcoal used. It was also, not surprisingly, related to the quality of the iron ore used. These differences are all sufficiently small that, at this level of investigation, it is not possible to separate their effects and to isolate differences in demand or differences in raw materials as the cause of the observed inter‑regional difference. While such a situation is lamentable, the hot blast derived its major importance from its permissive effect on the use of mineral fuel, and we hope to have better luck in our investigation of the larger question.
In the course of the 1830's a Welshman and an American discovered independently that the hot blast enabled anthracite to be used in the blast furnace. Although coke was well‑known as a blast‑furnace fuel by then, the difficulty of igniting anthracite had prevented its similar use. The American inventor died before he could test the commercial feasibility of his invention, but the Welsh inventor, David Thomas, was brought to the United States, where he built the first anthracite furnace that was both a commercial and a technical success in 1839. Its distinguishing features were its large size and powerful blast: the latter was six pounds per square inch, according to one report. "With the erection of this furnace commenced the era of higher and larger furnaces and better blast machinery, with consequent improvements in yield and quality of iron produced."17
The construction of this furnace, and specifically its important blowing engines, was quite difficult. Much of the machinery was brought from England, and arranging to have the rest built in this country posed many problems. The cylinders for the blowing engine are a case in point. Thomas wanted cast‑iron cylinders of five‑foot diameter and six‑foot stroke. He arranged to have them brought from England before he left for the United States, but they were too large to go through the hatches of the ship hired and were not brought over immediately. Thomas tried to have the cylinders made in this country, but there were no boring‑mills in the United States that were big enough for the job. The Southwick foundry of Philadelphia finally undertook the job, enlarging their boring‑mill for the purpose. The price of the cylinders was $0.12 a pound, or $280 a ton.18
After these difficulties and his famous commercial success, Thomas continued his leadership of the anthracite region, blowing in a total of five furnaces for his employers and forming the Thomas Iron Company in 1854 with his sons.19 His example was widely followed and the amount of iron made with anthracite rose rapidly. The contrast of this movement with the continuing neglect of the new technology in the West is striking and should be reflected in the price and cost structures of the two regions.
10 Arthur Cecil Bining, The Pennsylvania Iron Manufacture in the Eighteenth Century (Harrisburg: Pennsylvania Historical Commission, 1938); Bulletin, 19 (October 7, 1885), 266.
11 JFI, New Series, 5 (1830), 215; 9 (1832), 339; 10 (1832), 130; 18 (1836), 127; Swank, p. 453. The full titles of frequently cited periodicals can be found in the Bibliography.
12 0verman, P. 442.
13 Convention of iron Masters, Documents Relating to the Manufacture of Iron, Published on Behalf of the Convention of Iron Masters which Met in Philadelphia on the 20th of December, 1849 (Philadelphia, 1850), Tables. The differences among regions were statistically significant at the 5 per cent level, while intertemporal differences were not.
15 JFI, N. S., 24 (1839), 334‑345; Overman, for example, pp. 436‑440; Iron Age (March 11, 1875), p. 23.
16 Louis C. Hunter, "A Study of the Iron Industry of Pittsburgh Before 1860," unpublished doctoral dissertation, Harvard University, 1928, pp. 392‑433.
17 Swank, pp. 354‑361; William Firmstone, "Sketch of Early Anthracite Furnaces, 11 AIME, 3 (October, 1874), 152‑156, esp. 155; Iron Age, 43 (March 7, 1889), 348.
18 SamueI Thomas, "Rerniniscences of the Early Anthracite‑Iron Industry," AIME, 29 (September, 1899), 901‑928. The normal price for castings was about $70 a ton, or one‑fourth the price of the cylinders. (Berry, p. 276.)
19 Swank, pp. 360‑361; Bulletin, 16 (June 28, 1882), 173.
The discovery of the Bessemer process at the middle of the nineteenth century was a critical event in the development of the iron and steel industry and of the railroads as well. This chapter describes the introduction of the Bessemer process and the other methods and improvements in the manufacture of steel that followed it. The following two chapters show the effects of these methods on the firms that used them; the final two chapters of this Part place the making of steel into the context of the iron and steel industry as a whole.
Before Henry Bessemer's famous discovery, steel was made from pig iron in two steps. Carbon was removed from pig iron by puddling to make wrought iron, and then carbon was put back into the iron by reheating wrought iron with charcoal in a steelmaking furnace. This made blister steel, which could be converted into shear steel, a relatively uniform product, by welding several bars together, or into crucible steel, a completely uniform product, by melting pieces in small crucibles. Bessemer set out to make an improved wrought iron for military use. He found that blowing air through molten pig iron removed carbon as cheaply as existing methods and also kept the iron molten. The melting point of iron rises as the carbon is removed, and the difficulties of keeping wrought iron molten had led to the production of non-homogeneous wrought iron and steel. Bessemer created a uniform product, and he discovered in addition that he could stop the process partway and produce steel. Unfortunately the timing of the stop was critical, and the product was of very uneven quality. Robert Mushet solved this problem by allowing all the carbon to be removed, as in making wrought iron, and then adding some back as in the older processes of steelmaking. All this was done while the metal was molten, and the product had the uniform consistency of crucible steel, although it lacked the high quality of that product.1
A similar discovery was made in this country about the same time by an ironmaster named William Kelly. He did not develop the process to the point of commercial usefulness, but he developed it far enough to be able to convince the patent commissioners that he should get a patent in preference to Bessemer for the process.2 The parallel discovery of this process by two people working independently may indicate that there was some kind of "pressure" for the discovery of a new process in the iron and steel industry, but the main importance of the parallel discovery for us is its creation of conflicting patents in the United States. Kelly had the process patent, Bessemer had patents for the machinery, and Mushet had a patent for the replacement of the carbon. All these patents were in existence by 1857, and although the English rights to Mushet's patent were soon lost by a tax default, the American rights continued in force.
The first public account of Bessemer's invention was in the paper he delivered to the British Association for the Advancement of Science in August, 1856. Although many people were skeptical of Bessemer's claim, the importance of the paper, if true, was widely recognized, and it was widely reported in America.3 Abram Hewitt was among the first to react to the paper, and he built an experimental Bessemer converter at Trenton soon after he learned of Bessemer's paper. The converter, however, was never used; the people at Trenton heard about the "failure", of the Bessemer process and abandoned their project.4
There were two failures of the initial Bessemer process. The first was the inability to yield a uniform carbon content, which was solved by Mushet's innovation. The second was that this process, unlike puddling, did not remove phosphorus, and that therefore only a restricted class of iron could be used. It took time to discover this fact and to find a suitable class of iron. The knowledge of chemistry was sufficiently rudimentary at that time, in addition, that the only way of discovering if an iron was suitable was to try it. As late as 1866, the owners of the American patent rights could only repeat Bessemer's claim that iron containing more than 0.1 per cent sulphur, 0.075 per cent phosphorus, or 1.75 per cent silicum was not "well adapted" to the Bessemer process, and state: "We are prepared to test irons at seventy‑five dollars per ton of pig, the iron to be delivered, and the product to be removed at the owner's cost. This test includes sufficient hammering, rolling, and cold bending to determine the various practical qualities of the product.5
The date of this quote, 1866, is usually taken as the start of the commercial manufacture of Bessemer steel in America. The decade between the publication of Bessemer’s paper and the adoption of his innovation in America was a time when many problems were solved, or at least made manageable. The problems may be grouped into three classes: legal, technical, and financial. The legal problems were the result of the conflicting patents, about which, much unrecorded bargaining took place. The second group of problems includes the chemical problems just mentioned and other defects of the early Bessemer manufacture. We know the technical solutions adopted, but little about the search for them. The third class of problems are those that derived from the difficulty of convincing people to use the new process. Without an adequate knowledge of the technical and legal problems, it is hard to say much about the financial problems of these years.
I Sir Henry Bessemer, An Autobiography (London: Offices of "Engineering," 1905).
2 John Newton Boucher, William Kelly: A True History of the So‑Called Bessemer Process (Greensburg, Pennsylvania: Published by the Author, 1924) ,James M. Swank, History of the Manufacture of Iron in All Ages (second edition; Philadelphia, 1892), pp. 396‑400.
3 Hunt's, 35 (1856), 499‑500; ARJ, 29 (1856), 595; JFI, 3rd series, 32 (1856), 267.
4 W. F. Durfee, "The Manufacture of Steel," Popular Science Monthly, 39 (October, 1891), 729‑749.
5 Bulletin, I (December 5, 1866), 97.
Excerpt – Pages 132 - 142
The Bessemer Process In Use
The owners of the combined patents lost no time in trying to attract people to use their patents. They placed advertisements for licensees in the AISA Bulletin on a continuing basis starting at the beginning of 1867. They stated that the patents for "the Pneumatic or Bessemer Process" had been consolidated under a trusteeship, and named Z. S. Durfee as the agent to whom one could apply for a license 16 They then issued a pamphlet giving more information on the costs of the new process.17 The cost of a plant with 2 three‑ton converters was given as $80,000; of a "five‑ton plant" with steam power, $125,000; and of first‑class apparatus with fireproof buildings and duplicate machinery making 50 tons of ingots in twenty‑four hours, $200,000. It cost only two‑thirds as much for a Bessemer plant as for a crucible steel, charcoal bloom, or puddled bar plant of the same capacity, the trustees asserted, and it took only 30 men to run a five‑ton Bessemer plant.
Royalties for the new process were set in sterling to harmonize with Bessemer's charges in England. The base charge was one pound sterling for each gross ton of iron used to make ingots for rails, and higher rates were charged for ingots for other purposes. Making an allowance for the waste in conversion, the charge was about $5 per ton for the use of the Bessemer process. It was asserted that the agreement of 1866 greatly lowered the royalties charged,18 but evidence is lacking to support this contention, and it may be only a myth of the steel industry.
In addition there was an initial cost to a licensee of $5,000, in return for which Winslow, Griswold, and Morrell furnished plans of a plant and information on the processes involved. The accounts of the licensee firm had to be kept open to Winslow, Griswold, and Morrell in order to provide for checks on royalty payments. The works and processes had to be open to Winslow, Griswold, and Morrell also, and Winslow, Griswold, and Holley opened the Troy works to licensees in return. They would employ two people at a time from a licensee firm to Work at Troy in the first two years of the license, although they would not pay their wages. In contrast, however, with this free communication among licensees and licenser, none of the information supplied by Winslow, Griswold, and Morrell was to be communicated to anyone not a licensee.19
These provisions for communication are the most interesting feature of the licenses, and they raise the question of whether the institutional form adopted by an industry affects its technological development. The communication facilities embodied in the licenses were widely used, and new ones were added. Almost all the early Bessemer works were built according to plans drawn by Holley, presumably acting as part of the licensing firm. Of the eleven plants in operation in 1880, Holley designed six, consulted on the construction of three more, and was the inspiration for the remaining two which were copied after one of the first six. The technical personnel of these firms were involved in a continuing game of musical chairs, and the managers of the newer works usually had been trained at one of the earlier ones.20 Frequent meetings were held of the five or six top engineers of the industry to discuss common problems. The principal participants of these meetings were the following: Holley; John Fritz, who was then at the Bethlehem Iron Company where the meetings were often held; George Fritz, John's brother, who had taken over for him at Cambria; Captain R. W. Hunt, the author of the industry history just cited and at that time the manager of the Troy works; and Captain William Jones, the brilliant manager of the Edgar Thomson Steel Works, Carnegie I s plant.21
Finally, Holley wrote a series of confidential reports on technical subjects which were distributed among this circle. On their title page they bore the following legend: "These papers are printed, not as a publication, but for the convenience of my clients, and for their exclusive use." The reports were only a few pages in length, and they each treated a specific subject: a noteworthy feature of a plant, a new process, a new machine. They were published in two series in the years 1874‑1877, and there were fifteen to twenty of them.22 After the formation of the Bessemer Association (that is, the Bessemer Steel Company) in 1877, they were for the exclusive use of its members. Holley is said to have regarded them as his best work, and his clients found them extremely useful.23 The uniform size of the Bessemer steelworks in existence in 1880 has been attributed to the newness of the process and the "immaturity of firms,"24 but it would seem that the extensive communication among firms, the all‑pervading influence of Holley, and the almost exclusive production of one product‑ rails ‑would account for this phenomenon.25
The communication network established in the early Bessemer steel industry appears to have been designed in large part to make use of the extraordinary talents possessed by Holley. The success of these arrangements is evident in the great influence of Holley and the extent to which his innovations helped the American industry. R. W. Hunt stated in the memorial to Holley, "An imperfect knowledge of the chemical requirements of the [Bessemer] process, an utter absence of tested and approved refractory materials, and, above all, imperfect machinery, were the conditions of the problem which, in 1864, Holley set himself to solve."26 He worked in all these areas, and his reports and inventions reflect his wide range of interests.27 But his primary importance was probably in the last‑named category, the improvements in machinery. Two major innovations may be noted: the "American" or Holley floor plan, and the Holley bottom.
The original British floor plan for a Bessemer plant had two converters facing each other across a deep pit which contained molds to receive the molten metal from the converters. Two converters were used because the operations of a converter were discontinuous, and the output of a single converter would occupy the auxiliary equipment only a short part of the time. Heavy auxiliary equipment was needed to power the Bessemer machinery itself, and also to move and process the large amounts of metal the Bessemer converter used. The indivisibility of this machinery meant that large plants were necessary for the efficient utilization of the Bessemer process and that large outputs were needed for the success of these plants. The converters only worked a small percentage of the time because of the technical characteristics of the production, the difficulty of moving the metal from place to place, and the time needed to rebuild the converter lining consumed in the extraordinarily high heat of the process. The bottom of the converter would wear out after one to three heats, the converter would have to be cooled, and a man would climb inside and repair the lining of the converter.28 Having two converters kept the machines employed more of the time, but the original plan was not satisfactory.
Holley set himself to improve the speed with which the converters could be used by two means. He arranged the converters to facilitate the movement of the metal, and he shortened considerably the time required to repair the converter lining. The first was done by placing the two converters side by side instead of facing, raised high off the ground to let them discharge their contents at ground level rather than in a pit. This arrangement opened up the working space around the converters and permitted further innovations to speed the handling of materials.29 But the converter lining still wore out with great regularity and slowed operations behind the speed of the auxiliary equipment. Holley introduced his new bottom in 1869‑1870, and received a patent on it in 1872.30 The previous practice had been to repair the lining inside the converter shell, which meant cooling it sufficiently for a man to get inside. Holley's innovation was to make the bottom of the shell removable. A worn‑out bottom could be taken off the converter and a new one put on without cooling the converter itself; the saving in time is obvious. Holley claimed for himself the use of refractory brick, the preformed bottom, and the pre-drying of the bottom. The Holley plan gave the converters a "front" where the metal could be handled with facility; it also gave them a "back" where the bottoms could be handled equally expeditiously.
These innovations were far from the only ones introduced at this time, but they were among the most important, and they characterize the spirit of the innovations. Holley and Fritz are the major names in this process, but there were many active engineers whose influence cannot be accurately estimated from this distance. A race among steel men developed as each tried to increase the speed of his plant. Holley's innovations started and permitted this race, and many subsequent improvements continued it. The record of increasing speed could be described in detail, but without an explicit idea of the cost structure of steel mills, the proliferation of numbers would serve little function.31 In 1878 the twenty British Bessemer works in operation produced 800,000 tons of Steel.32 In the same year, the ten active American firms made 650,000 tons,33 about half again as much apiece as their British counterparts. By 1880, the American firms had individual capacities of over 100,000 tons, and production was not far under capacity.34
We would like to know if the ever‑increasing size of the American steelworks was economical. The great emphasis placed upon it indicates that it was, but it does not provide a test. The similarity of American steelworks likewise precludes a firm test. Information has survived about a few of the unsuccessful aspirants to membership in the Bessemer industry, but usually without enough information to know why they failed. The National Iron Armor Company of Chester, Pennsylvania, is typical of these firms. It obtained a license for the Bessemer process before 1868 and built a plant with five‑ton converters. Holley commented that the works were well‑located and well constructed. But we hear no more of them, and the causes of their failure are unknown.35
The Freedom Iron and Steel Works of Lewistown, Pennsylvania, is the only early and unsuccessful entrant to the industry about which information has survived. It made its first "blow" on May 1, 1868, and failed in 1869. The Bessemer works were dismantled and taken to Joliet, Illinois, where they were used by the Joliet Steel Company.36 The causes for this failure were apparently manifold, for all of the obvious things that could be wrong with a steelworks were ascribed to the Freedom Works. First, it was the only American steel mill to use the English floor plan in preference to Holley's 37 In addition, the iron used contained too much phosphorus, and the product was of bad quality.38 Finally, the Freedom Company was probably undercapitalized due to the failure of its president to appreciate the large scale of the new process .39 Each of these factors would have been sufficient to cause failure, and the effects of each of them alone cannot be estimated.
The similarity of Bessemer steel plants at 1880 and our ignorance of the precise causes of the failures in the Bessemer steel industry before that date, then, preclude direct testing of the assertion that high speeds were necessary for these plants. Nevertheless, the indivisibility of Bessemer machinery must have been a potent force leading to faster operations, and the participation of all Bessemer steel plants in the developments leading to greater speed implies that they offered benefits by reducing costs.
Alexander Holley died in 1882. His passing marked, and may have caused in part, a shift in the development of the steel industry and a decline in the uniformity of steel plants. Perhaps the most important of the changes in the supply of steel after 1880 was the rise in the importance of the open‑hearth process. This competitor to the Bessemer process was invented a decade after Bessemer's discovery, but it became important in America only after an additional delay of about two decades .40 We must understand why the Bessemer process was more suitable for use in the 1870's and 1880's and why the open‑hearth process was used in increasing proportion after those decades before inquiring further into other changes in steelmaking. To do this, we must describe the open‑hearth process and an important innovation of the 1880's; the basic process.
The open‑hearth process may be seen, in retrospect,. as a logical development of the puddling process. The puddling process reduced the amount of labor necessary to remove the carbon from pig iron by increasing the heat to which the iron was subjected. The open‑hearth process increased the heat still further and, by exceeding the melting point of wrought iron, was able to dispense entirely with the need for manipulation of the iron. The iron was placed on a hearth open to the flames from the fuel being burnt, much as in a puddling furnace. The difference was that the fuel gas, was burnt in the regenerative stoves introduced by William Siemens. In these stoves, the gas was alternately burned in one of two compartments of firebrick checker-work while the exhaust from the hearth was drawn through the other. The combustion chamber was thus preheated by the exhaust fumes. Heat from the burning gas was added to this existing reservoir of heat, and new heights of temperature were attained.
The open‑hearth furnace, in other words, was a puddling furnace to which greater heat could be supplied and, consequently, in which there was no need for a puddler. The contrast between this new process, first introduced in 1866, and the Bessemer process is apparent. The latter process was based on a new source of energy, the impurities in the iron itself, and it required for its operation a set of machinery unlike any that had existed before. Both processes, however, employed high heat to eliminate labor, and they both produced a uniform product that was immediately labeled steel .41
The first people to try the open‑hearth process were, of course, Cooper and Hewitt. They built an open‑hearth furnace at Trenton in 1868, but it was not successful. Present as an assistant at this experiment, however, was Samuel T. Wellman, who played a role in the introduction of the open‑hearth process similar to that played by John Fritz in the rail mill and Holley in the Bessemer steel mill. He was present at two of the four open‑hearth furnaces built in the following five years, and he was able to correct the errors in design that had led to failure at Trenton.42
Wellman was not as towering a figure as Holley or Fritz, either because of his native ability or because they were still active. Holley saw the possibilities of the open‑hearth process while he was still engaged in improving the Bessemer manufacture. His reports to his clients talk about the newer process, he worked with Wellman at some of the early open‑hearth furnaces, and he was connected with the introduction of the Pernot revolving open‑hearth furnace. This was one of the major innovations of the period, although it is unclear how much effect it had .43
The major problem of the early open‑hearth furnaces was maintenance of the furnace in the face of the high heat attained.44 This was similar to the problem of the Bessemer converter lining, but the solution adopted was not clear. It would appear that the development of better refractory materials was the only major change in this area 45 Another difficulty of the process was the cost of labor required to charge the furnace. The open‑hearth furnace, in contrast with the Bessemer converter, did not start out using molten inputs and could not be charged in the easy way that Bessemer's machinery allowed. Wellman introduced a machine that charged open‑hearth furnaces, saving not only labor but the capital costs implicit in cooling the furnace for hand charging. They were universally adopted and became a standard feature of open‑hearth works.46
The open‑hearth process was not widely used until after 1885, due to the high cost of steel made this way before that time, but there were reasons for a few people to adopt it. One reason, of course, was the quality of the product. This may have been real or fancied at this early date, but some people thought the quality of open‑hearth steel was more controlled.47
Bessemer steelworks also had the advantage of plentiful scrap, and some mills adopted this process to use the scrap.48 Non-Bessemer plants had the inducement, conversely, that the rights to use the Bessemer process were not available for part of this period. The small size of the open‑hearth plant, present for all users, was an additional factor inviting people to use it. The minimum size of an efficient open‑hearth plant was far smaller than that of a Bessemer plant, the average sizes (by capacity) of the two types of plant differing by a factor of ten in 1880, although innovations such as machine charging were narrowing the gap after that date .49 The difference followed from the relative time needed to transform iron into steel in a Bessemer converter (about twenty minutes) and in an open hearth furnace (six hours or more) and the cost of capital per unit of output implied by this at different outputs. The more "divisible" process was adopted in the interstices of the economy in the 1870's, there to wait until the discovery of the basic process which would lower its costs.
Both the Bessemer and open‑hearth processes were initially introduced with an acid lining in the refining vessel, the socalled acid process. This variant was the one in use in 1880, and it is the costs of the two acid processes that were referred to in the immediately preceding pages. In 1879, two Englishmen, Thomas and Gilchrist, introduced a practical way for using a basic lining in making steel and introduced the "basic" process.50
The problem with the acid process was that it did not remove phosphorus in the course of refining the iron; the phosphorus in the iron would not combine with the acid lining. As phosphorus was injurious to the metal being produced, the acid process could only use iron made from ore that did not contain much phosphorus. A whole class of iron ores was therefore excluded from use with these processes, as has been mentioned above. The basic process substituted a basic lining for an acid one, allowing the phosphorus in the iron to combine with the lining and be carried off in the slag. The use of a basic lining changed the cost structure of the steel industry, and it was its invention, more than any other single factor, which shifted the balance between Bessemer and open‑hearth steel.
The costs of producing steel in the nineteenth century, however, are hard to discover. From the moment the Bessemer process was adopted to any large extent, the steel men came under fire for the supposed effects of their restrictive policies and the protective tariff they worked so hard to get .5 1 They retaliated by telling folk tales about what was happening in the industry, leaving the investigator to guess about their costs and profits. The best cost comparison that we have, therefore, is not as good as could be desired, but its explicitness gives us as much information as could be hoped.
The costs are the fruits of an investigation of the steel industry undertaken by the United States Commissioner of Corporations after 1900. The investigation was designed to show the effects of the formation of the United States Steel Corporation, but if the data can be trusted, they may communicate other information as well .52 The costs given are all for the period after the turn of the century, and they therefore show the structure of costs after the steel industry had shifted its direction toward open‑hearth steel. We may examine these costs and attempt to infer from them what the structure of costs was before the decision was made. Some of the relevant data are given in Table 6.1.
The costs of the two processes were very close, and the decision to use one rather than the other would appear not to have been dominated by cost savings. Nevertheless, there was nothing to be lost by using the basic open‑hearth process at this point, and even a small something to be gained. This structure of costs made steel men receptive to arguments in favor of this process, and we ask how it came about.
AVERAGE COSTS OF STEEL INGOTS
IN THE PITTSBURGH DISTRICT,
($ per gross ton)
(Acid) Basic Bessemer Open‑Hearth Difference
(1) (2) (2)‑ (1)
Other works costs 1.34 2.13 +0.79
Total works cost 16.54 16.25 ‑0.29
General and miscellaneous expenses 0.83 0.69 ‑0.14
Total book cost 17.37 16.94 ‑0.43
Source: U.S. Commissioner of Corporations, 111, 166. The Bessemer billet ingot figures were used in preference to the rail ingot figures.
The last column of Table 6.1 shows that the small difference in the works costs of the two methods was the result of two offsetting influences. The works costs were the costs connected with the actual steelworks, as opposed to those costs that were part of the entire firm and could not be allocated directly to any single plant. Works costs can be broken down, as is done in the table, into material and nonmaterial costs. The former consisted primarily of the cost of pig iron and scrap, less than $.50 being charged to manganese and limestone. The latter consisted of a variety of small expenses of which the most easily recognizable are perhaps labor, fuel, steam, supplies and tools, and materials in repairs and maintenance. The table shows that the nonmaterial costs for the basic open‑hearth process were higher than those of the acid Bessemer, while the material costs were lower.
These two offsetting influences may be attributed to different causes. The nonmaterial costs of the acid open‑hearth process were as high or higher than those for the basic openhearth process, not only for this late date, but also earlier.53 As far as we know, also, they were both higher than the equivalent costs for the Bessemer process all through this period, and we may say that the higher nonmaterial costs of the newer process were due to the fact that it used an open‑hearth and not that it used a basic lining.
Material costs are a different story. The acid open‑hearth and the acid Bessemer process used the same kind of pig iron,
16 Bulletin, I (February 27, 1867), 212.
17 John F. Winslow, John A. Griswold, and Dan'l J. Morrell, Trustees, and Z. S. Durfee, General Agent, The Pneumatic or Bessemer Process of Making Iron and Steel (Philadelphia, 1868).
18 Iron Age (May 9, 1878), pp. 15‑16.
19 License for Kelly, Bessemer and Mushet patents (n.p., n.d.). The license also provided that the licensee would commence work within a year and that the agreement applied to all future patents of the three men named owned by the licensing parties.
20 Robert W. Hunt, "A History of the Bessemer Manufacture in America," AIME, 5 (June, 1876), 201‑216.
21 John Fritz, The Autobiography of John Fritz (New York: John Wiley and Sons, Inc., 1912), p. 160; E. B. Coxe, another engineer and author of the industry, often attended these meetings also. George Fritz died in 1873 [E & MJ, 16 (1873), 153].
22 The following reports are in the possession of the Library of Congress:
First Series (1874‑1875) and Supplements (1875‑1876)
3 The Siemens steel manufacture, from pig and ore, at Landore and at thesteel works of Scotland.
4 West Cumberland iron and steel works.
5 The Pernot revolving hearth furnace and practice, for Siemens‑Martinsteel and for puddling.
6 Brown, Bayley and Dixon's steel works.
7 Barrow haernetite iron and steel works.
8 The steel manufacture in France and Belgium.
Suppl. to 2 The value of manganese. The practice in Germany.
2nd Suppl. to 2 The effect of manganese, carbon, phosphorus, and other in gredients on iron and steel.
Second Series (1877)
2 The direct use of the blast‑furnace metal in the Bessemer process.
3 Price's retort furnace for reheating and puddling.
4 Report on the phosphorus steel manufacture, with and without the Sherman process.
5 Rolling mill improvements. Rolling double‑length rails direct from the ingot, blooming‑tables for reversing trains, etc.
6 The hot‑blast cupola and utilizing converter flame for heating cupola blast.
7 Solid steel castings for ordnance, structures, and general machinery, by the Terrenoire process.
23 American Institute of Mining Engineers (AIME), Memorial of Alexander Lyman Holley (New York, 1884),1). 136.
24 R. N. Grosse, "Determinants of the Size of Iron and Steel firms in the United States, 1820‑1880," unpublished doctoral dissertation, Harvard University, 1948, p. 214.
25 For the proportion of Bessemer steel used to make rails, see Appendix C, Table C.11.
26 AIME, Memorial p. 31.
27 For a list of Holley's inventions, see ibid., pp. 69‑70; Bulletin, 16 (November 15 and 22, 1882), 309.
28 Fritz, pp. 155‑156.
29 The new layout was patented as "a combination of crane, converter, and chimney." AIME, Memorial..., 1). 69. Pictures of a Bessemer plant around 1890, with the Holley floor plan, are given in Carnegie Brothers and Company, Ltd., The Edgar Thomson Steel Works and Blast Furnaces (Pittsburgh, 1890).
30 U.S. Patent 133,938, granted December 17, 1872.
31 Clark documents part of this race: Victor S. Clark, History of Manufactures in the United States (3 vols.; New York: McGraw‑Hill Book Company, Inc., 1929), Vol. 11, pp. 264‑265.
32 J. C. Carr and W. Taplin, History of the British Steel Industry (Cambridge: Harvard University Press, 1962), pp. 97, 108.
33 Appendix C, Table CA.
34 Grosse, pp. 281‑282; Appendix C, Table CA.
35 Holley, The Bessemer Process . . ., pp. 12, 37; Winslow, Griswold, and Morrell, p. 30. Other obscure aspirants were listed in National Association of Iron Manufacturers, Statistical Report for 1872 (Philadelphia, 1873), p. 132.
36 Swank, p. 411; National Association of Iron Manufacturers, p. 21.
37 Holley, The Bessemer Process . . ., p. 35.
38 Durfee, "The Manufacture of Steel, continued," p. 23; Iron Age (April 18, 1878), p. 5.
39 Andrew Carnegie, Autobiography of Andrew Carnegie (Boston: Houghton Mifflin, 1920), p. 178. Carnegie may have been connected with the Freedom Iron Company, the parent company of the Freedom Iron and Steel Works, in which case he would be in a position to know about capital supplies. In his statement of income for 1863, he recorded $250 from the Freedom Iron Company. Carnegie MSS, Manuscript Division, Library of Congress, Vol. 111.
40 Appendix C, Table C.5.
41 See AIME, Memorial . pp. 171‑181, and references given there.
42 Joseph G. Butler, Jr., Fifty Years of Iron and Steel (fourth edition; Cleveland: The Penton Press Co., 1923), pp. 68, 180; The Otis Steel Company Pioneer, Cleveland, Ohio (Cambridge, Massachusetts, privately printed, 1929), no page numbers.
43 AIME, Memorial . . ., p. 32; Bulletin, 13 (June 4, 1879), 138; Harry Huse Campbell, The Manufacture and Properties of Iron and Steel, (second edition; New York: The Engineering and Mining journal, 1903), p. 206.
44 Bulletin, I I (December 26, 1877), 337.
45 Note the existence of a furnace "rebuilding fund" for the open‑hearth process in U.S. Commissioner of Corporations, Report on the Steel Industry (Washington, 1913), Vol. 111, p. 166.
46 Frank Popplewell, Some Modern Conditions and Recent Developments in Iron and Steel Production in America (Manchester: The University Press, 1906), pp. 95‑96.
47 Iron Age (April 27, 1876), p. 11; Clark, Vol. 11, p. 267. Clark (pp. 270‑275) appears to accept the Bessemer steel men's reasoning that there was no important difference between the two types of steel at any time.
48 Iron Age (April 13, 1876), p. 1; American Iron and Steel Association, Directory to the Iron and Steel Works of the United States (Philadelphia, 18761900), 1876‑1880.
49 Grosse, pp. 281‑282; U.S. Census of Manufactures, 1905, Vol. IV, Special Reports on Selected Industries (Washington, 1908), 14.
50 Carr and Taplin, pp. 98‑100; Lillian Gilchrist Thompson, Sidney Gilchrist Thomas: An Invention and Its Consequences (London: Faber and Faber, 1940).
51 See Chapter 8.
52 U.S. Commissioner of Corporations, Report on the Steel Industry (Washington, 1913).