British shipbuilders had something new to boast about at the Chicago World Fair in 1893. That year, William Denny and Brothers of Dumbarton placed an advertisement in one of the Fair’s catalogues highlighting a recent engineering feat: “light draught work for Eastern Rivers.” The advertisement revealed an indisputable fact: Denny, the reputable Clyde shipbuilding firm, was a leader in this new specialization of the industry’s expertise. But even more interesting is what the advertisement didn’t mention. Far from the Clyde, these new techniques of shipbuilding came from engineers and businessmen half a globe away. Denny’s advertisement may have been placed in the catalogue for the British Section of the World Fair, but by the end of the nineteenth century British shipbuilding was anything but simply “British.”
The new expertise that Denny’s ad highlighted—ship design and construction for far away, foreign rivers—was a response to renewed imperial demands of the late nineteenth century. Increasingly, local governments and private companies operating in British colonies and protectorates drew on the powers of steam technology to chart and control freshly conquered territories. But this “steamship imperialism” was not merely a case of Clyde technology re-locating to the edges of empire.  Indeed, Denny’s expertise was founded on a rich, global data collection enterprise—the testing of steamers on foreign waters, and the meticulous recording and compilation of test results—and Denny’s earned its bragging rights by coordinating this data among client companies operating on Eastern rivers.
The results of these Eastern river trials can be found in the National Maritime Museum, amidst the 1000+ box collection of Denny and Brothers’ technical records. There, six large, belt-buckled volumes of Reports on Ganges and Irrawaddy Steamers reveal the way in which foreign river considerations permeated into ship design, and how experimental methods in the shipbuilding industry were reconfigured in the far reaches of British imperial waters.
Accumulating and sharing trial data was part of William Denny’s larger project to “apply the teachings of science to everyday commercial practice.”  His firm earned a scientific reputation for pioneering the commercial use of experimental tank trials in 1884, and for inaugurating more rigorous vessel testing with progressive measured mile trials a decade earlier. But just conducting these trials was not enough to ensure better ship construction. Even when trials were conducted by experts, Denny and his colleagues believed that many factors still “militate seriously against reliable data being procured.”  Faced with these problems, Denny set to work amassing a larger data set of trial results. This data set became a vital part of the experimental process, allowing “the designer of new ships […] to estimate with greater assurance of attaining satisfactory results.” 
Data sharing was meant to increase the reliability of trial results but did much more, ushering in qualitative changes to the shipbuilding industry. Denny and Brothers compiled unique data that its client companies generated, and today we can find these compilations in the Reports. For example, the Irrawaddy Flotilla Company (IFC) was one of Denny’s biggest clients, and it sent back notes from its operations on the easternmost rivers of British India. After receiving Denny-built steamers that were shipped overseas in pieces, the IFC would reassemble the vessels in its Burmese dockyards and put them to trial on the Rangoon River. IFC fleet engineers typed up trial results, often stressing crucial local considerations such as the priority of smaller drafts over higher speeds. In this way, fleet engineer Robert Barr offered his assessment of two Denny-built paddle steamers, the Kawlin and the Kentung. The IFC’s “practical staff” had watched the vessels race, and although the Kentung won the race, the Kawlin was undoubtedly superior. Weighed down with the same amount of tonnage, the Kawlin’s mean draft came in 4 inches below the Kentung’s. Barr noted this advantage, considering the viability of both vessels if they were put to work on the Mandalay-Myingyan Ferry route. Steamers on that route typically “leave Mandalay downwards daily with heavy Rice cargoes, so the ‘Kawlin’ will be able to carry more than former Ferry steamers, without risk of grounding.” 
Denny and Brothers sought out this practically-minded feedback, encouraging clients “to send us useful criticisms.”  The shipbuilding industry had not always been attentive to foreign river contexts, and beginning in the 1880s these global trial results facilitated a shift toward the construction of shallow-draft vessels.  Further design changes were to follow, including experimentation with new building materials that would offer a steamer more protection from the unique environment of Eastern rivers. Denny and Brothers began innovating with steel in the late 1870s, building its first steel steamer, the Taeping, for the IFC. The motivation didn’t come from Dumbarton but from the Eastern waters themselves: “They built this steamer of steel because they had suffered so severely from many of their steamers being snagged and sunk by trees and stones lying in the bed of the river.” Part of putting steel to the test was not just seeing how the Taeping fared on the Irrawaddy River, but determining whether the disassembled vessel’s steel plates could make the initial journey from Dumbarton to Rangoon. “This, although a practical test, was a very severe one.” The plates ultimately survived the trip without any breakage, and the IFC experience with the Taeping gave Denny cautious optimism for the use of steel in ship construction.
Fuel consumption was another feature of ship design that was negotiated through global trial data. The IFC stressed its low priority when reporting on the failed trials of the vane wheel steamer Fano. Although savings in fuel consumption were valuable, IFC manager R. Sinclair stressed that they were not worth pursuing if the trade off was speed. Compared to the Gozo, the Fano consumed 20% less fuel. But “the call everywhere throughout the Delta is for speed to enable passengers to reach their landing place as early in the afternoon as possible so that those who have far to go from the river bank may reach their homes before dark after which they are afraid to travel. This is a point of much importance.”  The Gozo was speedy and easily maneuvered, and perceptions of safety and security in British Burma made this a much better ship design.
Reconfiguring experimental methods out in empire
Not only did the objects of testing (the vessels) undergo revision as a result of this data collection effort, but the very testing process itself was subject to reconsideration. This testing required that steamers make multiple runs on a pre-set course (the designated “measured mile”) at varying speeds to gauge more precisely the relationship between engine power and vessel speed. In order to conduct these trials in British Burma, the IFC had to replicate the measured mile course on the Irrawaddy. But Denny’s Reports reveal how difficult this replication was. Ensuring the appropriate conditions at the testing site was easily frustrated. Poor weather conditions often obscured test results, and river traffic could also bring trials to a halt. Most unpredictable, however, were tidal conditions: quickly changing tides dealt a serious blow to the accuracy of speed readings.  Although engineers’ experience persuaded them “that it should be true, that similar ships tried under similar conditions do give similar results,” the hopes of achieving comparable results were dim. “The difficulty, of course, is to ensure similarity of conditions, which is much greater than the difficulty of producing similar ships.” 
Yet the Reports also reveal the experimental practices that client companies innovated in an effort to overcome flaws in the test design. Clients shared these testing practices with Denny and Brothers. After extensive testing of its creek steamer the Nikko, the IFC found a way to overcome rapid tidal fluctuations. It sent these results to Denny, showing how it could generate reliable speed readings by conducting four full-speed single runs with only ten minutes lapsing between the starts of consecutive runs, and averaging the results.  Denny and Brothers was then in a position to transmit these ideas to other clients experiencing similar testing problems. Thus when one of its other clients, the India General Navigation and Railway Company (IGN&RC), reported shaky trial results of its steamer, the Devana, Denny suspected that tidal fluctuations were responsible for erroneous speed readings. Denny recommended that the IGN&RC conduct revised trials following the model of the IFC. Sharing the IFC’s trial methods, Denny suggested that “in order to obtain accurate results and comparisons […] in waters that are rapidly changing in speed of flow,” the IGN&RC should institute a similar protocol of performing at least three single runs at any given speed. 
We can see from the Reports on Ganges and Irrawaddy Steamers how shipbuilding expertise in the late nineteenth and early twentieth century was premised on much more than the traditional know-how concentrated on the Clyde. In Denny’s case, Eastern river expertise was a result of widely-circulating “practical” and scientific knowledge, and both steamer-construction and experimental methods were improvised through trials far away from Dumbarton.
 Royal Commission for the Chicago Exhibition, Official Catalogue of the British Section (London: William Clowes & Sons, 1893).
 For more on the role of steamships as “pioneers in imperialism,” see Daniel Headrick, “The tools of imperialism: technology and the expansion of European colonial empires in the nineteenth century,” The Journal of Modern History 51 (1979), and Daniel Headrick, The Tools of Empire: Technology and European Imperialism in the Nineteenth Century (New York: Oxford University Press, 1981).
 “Memoir of Mr. William Denny, F. R. S. E.,” Transactions of the Institution of Naval Architects 28 (1887), 457.
 See discussion following: William Denny, “On the difficulties of speed calculation,” Transactions of the Institution of Engineers and Shipbuilders in Scotland 18 (1874-75), 219.
 David Pollock, Modern Shipbuilding and the Men Engaged in It: A Review of Recent Progress in Steamship Design and Construction, Together With Descriptions of Notable Shipyards, and Statistics of Work Done in the Principal Building Districts (London E. & F. N. Spon, 1884), 118.
 Letter from J. A. Folsom to John A. Galbraith on the “Kawlin,” c. 04 May 1914 (Reports on Ganges and Irrawaddy Steamers, Vol. II, DBAB0958, William Denny and Brothers, Ltd. Collection, National Maritime Museum).
 Archibald Denny, “On the practical application of stability calculations,” Transactions of the Institution of Naval Architects 28 (1887), 375.
 For more discussion of changes in the shipbuilding industry due to increased attention to client demand, see especially: Robert V. Kubicek, “The design of shall-draft steamers for the British Empire, 1868-1906,” Technology and Culture 31 (1990).
 See discussion following: B. Martell, “On steel for shipbuilding,” Transactions of the Institute of Naval Architects 19(1878), 25.
 Letter from R. Sinclair to the Secretary of the Irrawaddy Flotilla Company, 11 July 1924 (Reports on Ganges and Irrawaddy Steamers, Vol. IV, DBAB0959, William Denny and Brothers, Ltd. Collection, National Maritime Museum).
 In a paper presented at the 23rd session of the Institution of Naval Architects, one member described how the results of progressive measured mile trials “are only approximately reliable where the tide is either very slow or very uniform, and would be quite unreliable where the tide is fast and variable.” J. Harvard Biles, “On progressive trials,” Transactions of the Institution of Naval Architects, 23 (1882), 191.
 Letter from William Denny and Brothers, Ltd. to Messrs. Duncan Macneill & Co., 16 August 1930 (Reports on Ganges and Irrawaddy Steamers, Vol. V, DBAB0956, William Denny and Brothers, Ltd. Collection, National Maritime Museum).
 Notes by William Denny and Brothers on the “T.S.S. Devana,” 05 February 1918, volume 2, 1918 (Reports on Ganges and Irrawaddy Steamers, Vol. II, DBAB0958, William Denny and Brothers, Ltd. Collection, National Maritime Museum).
 For more on the localization of experimental methods in late-nineteenth century shipbuilding, see especially Leggett’s work on the replication of test tanks: Don Leggett, “Replication, re-placing, and naval science in comparative context, c. 1868-1904,” The British Journal for the History of Science 46 (2013).