Officers of the US Army Venture into Research, 1840–60

Notes

• See, for example, Todd Shallet, ‘Building Waterways, 1802–1861, Science and the United States Army in Early Public Works’, Technology & Culture, 31 (1990), 18–50 and Structures in the Stream: Water, Science and the Rise of the US Army Corps of Engineers (Austin: University of Texas Press, 1994).
   Google Scholar
• The three volumes are: Officers of the Ordnance Department, Reports on the Strength and other Properties of Metals for Cannon (Philadelphia, 1856); T. J. Rodman, Reports of Experiments on the Properties of Metals for Cannon (Boston, 1861); A. A. Humphreys and H. L. Abbot, ‘Report upon the Physics and Hydraulics of the Mississippi River’, Professional Papers of the Corps of Topographical Engineers, United States Army (Philadelphia, 1861).
   Google Scholar
• The Third System forts typically had multiple tiers of guns placed behind thick masonry or ashlar walls. A number of them, such as Fort Knox in Maine, survive. The cannon failures are described by Louis F. Gorr, ‘The Foxall Columbian Foundry: An Early Defence Contractor in Georgetown’, Records of the Columbia Historical Society, 71/72 (1971–72), 34–59.
   Google Scholar
• Lee M. Pearson, ‘The ‘Princeton’ and the ‘Peacemaker’: A Study in Nineteenth Century Naval Research and Development’, Technology & Culture, 7·2 (1966), 163–83; Edward L. Beach, United States Navy: 200 Years (New York: Holt, 1986).
   Google Scholar
• Alexander L. Holley, A Treatise on Arms and Armour (New York: Van Nostrand, 1865).
   Google Scholar
• Daniel Treadwell, Short Account of Improved Cannon (Cambridge, Mass.: Metcalf, 1845); On the Construction of Improved Ordnance (Cambridge, Mass.: Welch, 1862). His work has been described as ‘the truly revolutionary invention of the nineteenth century for artillery’ by Bernard Brodie, Sea Power in the Machine Age (Princeton: Princeton University Press, 1941).
   Google Scholar
• Charles Ellet, Mississippi and Ohio Rivers: Containing Plans for the Protection of the Delta from Inundation (Philadelphia, 1853).
   Google Scholar
• Humphreys’s background, the influence of his somewhat difficult personality and the subsequent influence of the report on river management are explained in Martin Reuss, ‘Andrew A. Humphreys and the Development of Hydraulic Engineering: Practice and Technology in the Army Corps of Engineers, 1850–1950’, Technology & Culture, 26·1 (1985), 1–33; and in John M. Barry, Rising Tide: the Great Mississippi Flood of 1927 (New York: Simon & Schuster, 1998), ch. 2.
   Google Scholar
• Lists of textbooks used are found in Travels through North America during the Years 1825 and 1826 by Bernhard, Duke of Saxe-Weimar and Eisenach, 1 (Philadelphia, 1828), 110–18; and S. L. Falk, ‘Soldier Technologist: Alfred Mordecai and the Beginnings of Science in the United States Army’ (PhD dissertation, Georgetown University, 1959). See also J. Morrison, Jr, ‘The Best School in the World’, West Point, the Pre-Civil War Years, 1833–1866 (Kent Ohio, Kent State University Press, 1986).
   Google Scholar
• Dennis Mahan, An Elementary Course of Civil Engineering, for use of Cadets of the United States Military Academy (New York, 1837) and subsequent editions.
   Google Scholar
• These limitations of the West Point curriculum are explained in Eda Kranakis, Constructing a Bridge (Cambridge, MIT Press, 1997), p. 242.
   Google Scholar
• T. S. Reynolds, ‘The Education of Engineers in America before the Morrill Act of 1862’, History of Education Quarterly, 32·4 (1992), 459–82.
   Google Scholar
• The Fort Pitt Foundry built its cannon works in 1814 and from 1821 regularly contracted to cast cannon. It failed in the panic of 1837, was revived by W. J. Totten and Charles Knapp and upon Totten’s death in 1850 was operated as Knapp, Wade & Co. The West Point Foundry cast its first cannon in 1820. In 1836 Captain Robert P. Parrott of the Ordnance Department resigned his commission to become its superintendent; see Steven A. Walton, ‘The West Point Foundry in Larger Perspective’, IA, Journal of the Society for Industrial Archaeology, 35 (2009), 9–14. Cyrus Alger’s South Boston Iron Co., started in the 1830s, had cast no cannon for the government before 1841, see Gorr. Alger established a reputation for numerous inventions including use of malleable cast iron for plows and for rolling-mill rolls before undertaking cannon. Edward C. Ezell, ‘The Development of Artillery for the United States Land Service before 1861; With Emphasis on the Rodman Gun’ (MA thesis, University of Delaware, 1963), p. 111; Charles B. Norton, American Inventions and Improvements in Breech-Loading Small Arms (Springfield, 1880), p. 362. The Columbian Foundry was un-modernized and in terminal decline by 1840.
   Google Scholar
• Lt. Col. Anne Louis de Toussard, appointed in 1798, was the army’s inspector of cannon until the office was abolished in 1802. Thereafter little was done about cannon quality until the Ordnance Department was created on the eve of the War of 1812 under Commissary-General of Ordnance, Decius Wadsworth. By 1815 Wadsworth had established standard cannon designs. See Ezell, pp. 79, 94. The 1821 reorganization of the army abolished the Ordnance Department but it was then re-established in 1832.
   Google Scholar
• The Ordnance Department also wanted the parts of muskets and rifles uniform, i.e. interchangeable, and devoted much effort in creating gauging systems for use at the national and private armouries to attain this end. See Robert Gordon, ‘The Armoury Gauging System and Interchangeable Manufacture’, Arms & Armour, 7·1 (2010), 40–52.
   Google Scholar
• In 1840 the Ordnance Department sent Rufus Baker, Benjamin Huger, Alfred Mordecai and William Wade to Europe to study cannon casting technique. They visited foundries in Britain, France and Sweden and bought samples of cannon from each for trial back home. See Alfred Mordecai, Military Commission to Europe in 1855 and 1856 (Washington, 1861). The department began a programme of experimentation at the commercial foundries under the direction of William Wade (1789–1875), who had been commissioned as a lieutenant of artillery in 1813 but was a civilian employee of the Ordnance Department. In 1842 the department obtained authority to appoint attending agents to supervise the work of the commercial foundries.
   Google Scholar
• Rodman, p. 93; Ezell, p. 140.
   Google Scholar
• Ibid., p. 93 and reports for 26 October 1849; Ezell, pp. 153, 156.
   Google Scholar
• John A. B. Dahlgren (1809–70) made a similar innovation for the US Navy at this time. The navy’s research on cast-iron cannon is not documented in reports and books comparable to the army’s, but appears to have followed the same lines.
   Google Scholar
• Peter Barlow, A Treatise on the Strength of Timber (London, 1837). Barlow was instructor in mathematics at the Royal Military Academy at Woolwich. For a critique of Barlow’s work, see Isaac Todhunter, History of the Elasticity and Strength of Materials, v. i (Cambridge, 1886). The development of the requisite theory is traced in Stephen P. Timoshenko, History of the Strength of Materials (New York: McGraw-Hill, 1958), chs 4 and 5.
   Google Scholar
• Rodman, p. 93. First he used ratios and some algebra to show that stress in the tube walls decreased as the square of the radius. This was a restatement of Barlow’s rule. Next he computed the overall strength of a gun tube by integrating Barlow’s solution for the stress distribution across the thickness of the tube. He calculated ∫a/x2 dx across the thickness of the tube, where a is the strength of the metal and x is the radial distance from the axis. He did the integration correctly and found that the strength of a tube one calibre thick was 2/3 the strength of the metal. He could have had the same result without the calculus. Next the reader encounters the first of Rodman’s convoluted, obscure explanations of his results: Hence the effective resistance of a gun one caliber thick is 2/3rds of that which half a caliber thickness would offer if the strains were all equal to the tensile strain. He was referring to an imaginary force on a sectioned tube of half-calibre thickness. He asserts after further discussion that the pressure in the bore of a gun one calibre thick develops ‘nine times the strain on the interior that it does on the exterior, independent of previous strain’, which also follows directly from Barlow’s formula without calculation. These calculations ignore the additional strength from the closed end of the gun tube.
   Google Scholar
• The correct solution to the thick tube problem appears in Gabriel Lamé’s textbook Leçons sur la Théorie Mathématique de l’Élasticté des Corps Solides (Paris, 1852). The Krupp calculation is reproduced in Holley, p. 94.
   Google Scholar
• Rodman, p. 192. Treadwell’s critique is in his 1862 pamphlet. His criticism of Rodman’s pressure gauge in this publication is faulty.
   Google Scholar
• W. Johnson, ‘T. J. Rodman: Mid Nineteenth Century Gun Barrel Research and Design for the US Army’, International Journal of Impact Engineering, 9·1 (1990), 127–59.
   Google Scholar
• C. H. Gibbons, Materials Testing Machines (Pittsburgh, 1935); R. B. Gordon, ‘Strength and Structure of Wrought Iron’, Archeomaterials, 2 (1988), 109–37. William Wade also introduced measurement of mass density for assessing the soundness of gun castings and the properties of cast iron. As early as 1759 Marc Rene had suggested measuring the density of the metal in cannon as a non-destructive test that would detect porosity. See Ezell, p. 84. Measurement of the mass density of the cannon would reveal porosity problems since the metal in the cannon should have very nearly the mass density of solid cast iron. This was a quantitative test that could be made with good accuracy and was adopted by the Ordnance Department in the course of Wade’s work. Low density was clearly bad since it indicated large porosity. The problem was that high density was not in itself proof of high quality since a small amount of porosity at critical places would be undetected and could lead to failure.
   Google Scholar
• The rifle-making machinery supplied to the British Enfield Armoury is a conspicuous example. See David Williams, The Birmingham Gun Trade (Stroud: Tempus, 2004), pp. 55–72.
   Google Scholar
• G. E. Dieter, Mechanical Metallurgy (New York: McGraw-Hill, 1986), pp. 213–14. For an evaluation of Wade’s work, see Todhunter, v. II Part 2, 688–96. The reduction of ductility is sensitive to the microstructure of the individual specimens tested.
   Google Scholar
• Rodman, reports of July 1851, August 1852 and July 1855. That their results were reliable is shown by the difference in silicon content they found between cold and hot blast iron, an average of 1·9% for hot blast vs. 0·6% for cold blast that accords with modern analyses on historic specimens.
   Google Scholar
• C. J. B. Karsten, Handbuch der Eisenhütten- kunde (Berlin, 1841).
   Google Scholar
• Booth and Morfit’s chemical analyses do show us the problem that ordnance officers faced in trying to get control of the iron they used for casting cannon. A convenient way of examining their composition data is to compute the carbon equivalent, CE = %C + (%P + %Si)/3. See William Rostoker, ‘Troubles with Cast Iron Cannon’, Archeomaterials, 1 (1986), 69–90. The CE shows how close the iron is to the 4·5% eutectic composition and is an indicator of the strength of the iron when cast. For maximum strength in a grey iron casting the CE should be about 3·5. The CE values among the thirty-two analyses reported by Booth and Morfit range from 3·06 to 5·56, showing that little control had been attained in the iron used in the cannon tested. High CE iron is likely to have the large graphite flakes in its structure that are a source of weakness. The two cannon listed in the table of analyses as failed had high CE and a large proportion of graphic carbon. There are significant differences in the average CE among the foundries; at 3·86 Fort Pitt was closest to the ideal, West Point was 4·47 and Columbian and Bellona, combined, 4·39. Holding iron molten, or melting and recasting are ways of reducing the CE. However, the only evidence on the form and distribution of these constituents was what could be gleaned from the appearance of the fracture surface of a broken bar or the surface of a casting. Only after 1880 were the techniques of metallography available to reveal the internal structure of cast iron.
   Google Scholar
• Ordnance officers believed that if they could just find the right kind of iron their cannon would be strong. They shared the then prevalent view that iron alone or in a blend from specific ores or furnaces was particularly suitable for certain products. They ran innumerable trials with pig purchased from different furnaces, alone or in blends. Experience had shown practising foundry men that holding molten iron in the melting furnace could improve the quality of castings. Wade made numerous trials such as the one described in his 16 March 1844 report, of casting a series of cannon with one batch of iron used immediately upon melting or held molten for one, two and three hours before casting. Rodman in a trial of the cooled-core casting technique first selected a mixture of pig from the Greenwood, Springfield and Salisbury furnaces, some already melted and recast, for his cannon. In another experiment, in 1848, anticipating Whitworth’s 1870 compressed steel process, Wade tried applying hydrostatic pressure to the iron in a cannon mould as it solidified in the hope it would produce sounder metal. The ordnance reports describe dozens of such experiments. Each of the cannon made by these variations in procedure was tested by firing to destruction. At least some European innovators appear to have dealt with principles and practice more successfully. The work of Robert Mallet (1810–81) offers a contrast to the combination of blind experimentation and theory used by the Ordnance Department. Mallet, an 1830 graduate of Trinity College Dublin, where he had studied mathematics, was manager of a major iron foundry. He combined his business practice with a lively interest in engineering and natural science (his work in seismology remains highly regarded). He kept abreast of the literature and had read the US ordnance reports. He had studied Karsten’s treatise on metallurgy. Mallett undertook construction of heavy cast-iron mortars reinforced with tensioned iron bands. When he examined failed hydraulic cylinders he observed the crystalline grain structure formed during cooling and noted how sections of weakness resulted where growing grains intersected. He correctly deduced that the grain structure was controlled by the direction of heat flow in the mould and so could be controlled. These observations are presented in his 1856 book, where he also described faulty structures within the iron itself and the possibility of deleterious residual stresses that arose during solidification of the metal.
   Google Scholar
• By 1820 Karsten had shown that the composition and through it the structure, of iron could be controlled in the melting furnace. This appears to have not been known to Wade and others in the Ordnance Department.
   Google Scholar
• The explosion of Stockton’s poorly fabricated wrought iron cannon on the USS Princeton in 1844, coupled with lobbying by US foundries holding cannon contracts, were partly to blame for the Ordnance Department’s reluctance to adopt novel cannon designs. The fault here was with poor execution of a good idea: the Franklin Institute’s examination of the Princeton cannon showed that some of the iron bars had areas that had never been welded.
   Google Scholar
• Marshall J. Bastable, ‘From Breech Loading to Monster Guns: Sir William Armstrong and the Invention of Modern Artillery’, Technology & Culture, 33 (1992), 213–47.
   Google Scholar
• G. Galer, F. Kemmish and R. Gordon, Connecticut’s Ames Ironworks, Family, Community, Nature and Innovation in an Enterprise of the Early American Republic (New Haven: Connecticut Academy of Arts and Sciences, 1998).
   Google Scholar
• Engineers had a great deal of experience running level lines along the Mississippi River and its tributaries and in their investigations of possible railroad routes to the west.
   Google Scholar
• In modern terminology a rating curve, the relation between water height and discharge was determined for each section studied.
   Google Scholar
• The thoroughness of the survey’s work is illustrated by its demonstration that the effect of wind over the Gulf of Mexico has a greater influence on the river flow than do the tides in the Gulf.
   Google Scholar
• The bottom samples showed that the river bed is older, non-alluvial sediment. At that time the implications of J. L. R. Agassiz’s ice age theory proposed in 1840 remained unaccepted in America, so the geological history needed for interpreting the evolution of the Mississippi channel was not yet at hand to help Humphreys with his interpretation. See Robert V. Bruce, The Launching of Modern American Science, 1846–76 (Ithaca: Cornell University Press, 1988), p. 344.
   Google Scholar
• R. L. Meade and J. A. Moody, ‘Cause for the Decline of Suspended-sediment Discharge in the Mississippi River System, 1940–2007’, Hydrological Processes, 24 (2010), 35–49. R. H. Kesel, ‘The Decline in the Suspended Load in the Lower Mississippi and its Influence on Adjacent Wetlands’, Environmental Geology and Water Science, 11·3 (1988), 271–81. Since 1950 the sediment load has decreased by as much as 70% due to closure of dams on the Missouri River.
   Google Scholar
• With a combination of observations and reasoning based on physics Chézy concluded that the flow in two similar channels could be compared by the proportionality V2P/AH = v2p/ah, where v is the water speed, p the wetted perimeter of the channel, a the area of the cross section of the flow and h the slope of the channel. Capital letters apply to the second channel. Hunter Rouse and Simon Ince, History of Hydraulics (Ames: Iowa State University Press, 1957), pp. 117, 141, 163, 173.
   Google Scholar
• In 1839 John B. Jarvis used the open channel flow equations developed by Prony, Eytelwein and Robison to demonstrate that the Croton Aqueduct water supply to New York City would deliver the volume of water required. John B. Jarvis, Reminiscences (Syracuse: Syracuse University Press, 1971), p. 130.
   Google Scholar
• Humphreys and Abbot, p. 29.
   Google Scholar
• Ibid., p. 331.
   Google Scholar
• Ibid., p. 402.
   Google Scholar
• In fact, he had made at least a 100% error. His description of this flood back in chapter two gave the highest river flow as 4·16×104 m3/s (see p. 177). Subsequent experience showed that a flood control system would have to deal with river flows at least as large 8·5×104 m3/s, see H. Miller and E. Silberhorn, ‘Managing Water Resources in the New Orleans Area’, Journal of the Water Pollution Control Federation, 56·9 (1984), 995–1002.
   Google Scholar
• Humphreys and Abbot, p. 30.
   Google Scholar
• A typical levee designed according to the Humphreys criterion had to be built with a volume of soil and therefore mass, twenty times greater by 1920 than in 1859. This large mass placed on the river bank increased subsidence and thereby reduced the effective levee height. Careful levelling surveys showed how the height of the flood plain decreased away from the river. Humphreys drew the correct conclusion that this was because coarser sediment was deposited close to the river during inundation of the flood plain. However, he failed to recognize the implication this had for the fate of the flood plains if his plan to suppress flooding were implemented.
   Google Scholar
• American Journal of Science, 33, 2nd series (1862), 181–90.
   Google Scholar
• Humphreys and Abbot, pp. 29–30.
   Google Scholar
• The problem of scale in fluid mechanics was addressed in the second half of the nineteenth century by the introduction of dimensionless ratios. The Reynolds number, which defines the conditions for the transition from laminar to turbulent flow, is one of the most familiar of these.
   Google Scholar
• Rodman, p. 55.
   Google Scholar
• This issue has been fully explored by Reuss and Barry.
   Google Scholar
• Walter R. Johnson and Benjamin Reeves, ‘Report of the Committee of the Franklin Institute of the State of Pennsylvania on the Explosion of Steam Boilers, of Experiments made at the Request of the Treasury Department of the United States, Part II, Containing the Report of the Subcommittee to whom was referred the Examination of the Materials Employed in the Construction of Steam Boilers’, Journal of the Franklin Institute, 19 (1837), 75–109, 157–93, 241–77, 325–61, 409–51; ibid., 20 (1837), 1–31, 73–113.
   Google Scholar
• H. J. Viola shows how the federal government States was ill-equipped to receive the collections from the Exploring Expedition with the result that much irreplaceable material was lost. H. J. Viola, ‘The Story of the US. Exploring Expedition’, in Magnificent Voyagers, ed. by H. J. Viola and C. Margolis (Washington: Smithsonian Institution, 1985). See also Bruce, pp. 208–09.
   Google Scholar
• The Ordnance Department had similar difficulty in dealing with the transition from wrought iron to steel in its rifle barrels. See Robert Gordon, ‘Issues in the Introduction of Tonnage Steel in the United States, 1867–1883’, Journal of the Historical Metallurgy Society, 39 (2011), 42–51.
   Google Scholar
• W. E. Berkhimer, Historical Sketch of the Organization, Administration, Material and Tactics of the Artillery, United States Army (1884), p. 273.
   Google Scholar
• David M. Hansen, ‘Zaliniski’s Dynamite Gun’, Technology & Culture, 25 (1894), 264–79.
   Google Scholar
• On the Monitor’s cannon, see Beach.
   Google Scholar
• This episode is described fully by Barry, ch. 5.
   Google Scholar
• Ibid., p. 91.
   Google Scholar
• See, for example, E. S. Reich, ‘US Integrity Effort hits Troubled Water’, Nature, 484 (2012).
   Google Scholar