Engineers of Dreams: Great Bridge Builders and the Spanning of America

by Henry Petroski

  The question of the aesthetic qualities of cantilever bridges in general and the Forth Bridge in particular was a much-discussed subject in the late 1880s. In a letter to Engineering News, F. J. Amweg, a civil engineer from Philadelphia, admitted that “the ‘cantilever fever’ is prevalent at the present time.” However, he defended the design of his city’s Market Street Bridge, in response to an article by the Pittsburgh engineer Gustav Lindenthal, who alluded to the bridge as just “another ugly looking cantilever bridge.” Such an aversion to the form foreshadowed the career of Lindenthal, who was to play a dominant role in late-nineteenth- and early-twentieth-century American bridge building, but that is a topic for the next chapter.

  In another of his many lectures on the Forth Bridge—this one in 1889, before the Library Institution of Edinburgh—Baker contended that the beauty or ugliness of the bridge had to be considered in the context of its size. He was reported to have said that “it was useless to criticise the design on paper, because the mental emotion arising from its enormous size was absent.” His remarks were in response, at least in part, to growing criticism from the likes of William Morris, who had recently said that “there would never be an architecture in iron, every improvement in machinery being uglier and uglier until they reach that supremest specimen of ugliness—the Forth Bridge.” Baker argued in his lecture that it was not possible to judge an object’s beauty without knowing the functions it was designed to satisfy. Though the marble columns of the Parthenon were beautiful in their place, for example, they would cease to be so if bored through and used as funnels for an Atlantic steamship; he admitted, however, that, “of course, Mr. Morris may think otherwise.” Baker stated that he and Fowler had indeed considered aesthetics in their design.

  Other critics of the Forth Bridge faulted it for its structural excess and its consequent great cost. Not the least among these critics was Theodore Cooper, who shortly after the great Scottish bridge was finished would make comments upon it that years later would come back to haunt him:

  You all know about the Firth of Forth bridge, the clumsiest structure ever designed by man, the most awkward piece of engineering in my opinion that was ever constructed, from the American point of view. An American would have taken that bridge with the amount of money that was appropriated and would have turned back 50 [percent] to the owners, instead of collecting when the bridge was done, nearly 40 [percent] in excess of the estimate.

  The Forth Bridge has continued to have its aesthetic and structural champions and detractors, which is not surprising considering the scale and unusualness of the structure. To give a sense of the bridge’s scale, Engineering published in 1889 a diagram that compared the size of the Forth Bridge with that of the Eiffel Tower, then recently completed in Paris. Two Eiffel Towers were drawn on their side, balancing each other foot to foot, with two equally large half-spans of the Forth Bridge superimposed. Not incidentally, this comparison with the tower widely known to have been designed to resist principally the force of the wind, could do little but reinforce the impression of a sturdy and safe bridge. However, neither this exhibition of scale or stance nor the human model demonstrating the structural principles involved, represented the bridge in full. Whereas the Eiffel Tower diagram showed only one pier and two half-spans, and whereas the human model showed two piers and their respective four half-spans, plus a suspended span, the actual Forth Bridge has three piers, six cantilever arms, and two suspended spans. Were the three towering column structures not such a dominant visual feature of the bridge, it might have been criticized as having an even number of full spans, something many bridge engineers believe makes for an inferior visual composition. However, in the Forth Bridge, the light suspended spans are not so much the focus of attention as are the heavy, straddling towers and their symmetrically cantilevered arms, so the bridge appears to bound across the firth in an odd number of leaps that works both visually and structurally.

  The completed bridge was tested in January 1890, during a severe gale, with two trains, each comprising three seventy-three-ton locomotives pulling fifty cars full of coal, with another three locomotives bringing up the rear. The total weight on the bridge was eighteen hundred tons, under which the ends of the cantilevers deflected only about seven inches. Though the trains ran side by side on the two-track bridge, they were not allowed to cross it completely—that first full crossing would not occur until January 24, when it was achieved by a special passenger train carrying the chairmen of the railroad companies. The formal opening of the Forth Bridge occurred on March 4, when the Prince of Wales, accompanied by his son, Prince George, and the dukes of Edinburgh and Fife, rode across in a train and declared the bridge opened. In a ceremony on the occasion, the prince announced that knighthood had been conferred by Queen Victoria upon Benjamin Baker and William Arrol—Sir John Fowler had already been so honored—for their work on the Tay and Forth bridges, which signaled the progress that was being made in bridge building worldwide.

  The cover of the souvenir program showed a North British Railway locomotive labeled “Progress” pulling a through passenger carriage labeled “Aberdeen to New York, via Tay Bridge, Forth Bridge, Channel Tunnel, and Alaska.” Dreams are always in advance of reality, however, and a Channel tunnel, the boring of which had in fact begun years before, would not be completed for more than a hundred years. Prior to the 1990 centennial of the Forth Bridge, it was given a thorough inspection and declared to be in fine shape, a result of its having been conscientiously maintained throughout its first hundred years; and it was predicted that “given reasonable care and maintenance [the bridge] will last for another 100 years.” Thus the Aberdeen–to-New York through train might still someday be achieved, for a trans-Siberian railway was officially completed in 1991, and a bridge connecting Siberia and Alaska across the Bering Strait was for some time the dream of such engineers as Joseph Strauss and Tung-Yen Lin. Though none of these would realize their dreams, that is not to say that their successors will not take the project up; but it will be one for another century.

  3

  While the Forth Bridge was being built in Scotland, American bridge construction was also continuing to advance. Octave Chanute had overseen the building of the two-thousand-foot-long, three-hundred-foot high Kinzua Viaduct, carrying the Erie Railroad’s new division into the mining region of northwestern Pennsylvania. Originally constructed out of wrought iron in 1882 by the Phoenix Bridge Company of Phoenixville, Pennsylvania, the structure would be replaced with a stronger steel design in 1900. (It now stands unused in Kinzua Bridge State Park.)

  The Brooklyn Bridge was completed in 1883, under the supervision of Washington Roebling and his wife, Emily Warren. With a clear span of 1,595 feet between stone towers with Gothic arches, it was famous for being the longest suspension bridge in the world, but it was soon so congested with traffic that talk of other bridges across the East River began almost immediately upon its opening. Furthermore, the bridge, designed for the lighter trolley and road traffic of almost twenty years earlier, was not structurally able to carry heavy locomotives, and for years elaborate schemes of switching cable cars at the New York and Brooklyn terminals would be an almost constant topic of discussion in the pages of Engineering News.

  The Kinzua Viaduct, as it looked in the late nineteenth century, in northwestern Pennsylvania (photo credit 3.13)

  Expanding cities and railroads needed more and more bridges, and the steadily increasing volume and weight of traffic they were being required to carry made bridge design an ever-challenging endeavor, whose practitioners welcomed new and better ways of determining the loads that various kinds of vehicles and trains imposed on a structure. There is seldom a single method that is all things to all engineers, and different individuals at a given time often have different views of how best to design a bridge. In the late nineteenth century, a technical paper published in the Transactions of the American Society of Civil Engineers, which was a principal forum for such discussions, was often follow
ed by at least as many pages of discussion by a dozen or more of the author’s prominent contemporaries (as had Schneider’s 1885 paper on the Niagara Falls cantilever). The pages of Engineering News and its successor, Engineering News-Record, also often contained exchanges between engineers, like those that had appeared in Engineering between Eads and Roebling over the design of caissons, but these tended to be more one-on-one serial debates and could be less dignified professionally.

  The bulk of what the likes of James Eads and the Roeblings wrote about bridge design appeared in the form of reports to those from whom they wanted initial or continuing financing. They published relatively little about bridge building in professional journals; therefore, with a few notable exceptions, their theoretical outlook tended not to be so influential as their practical achievements in construction. Those engineers who wrote more constantly and openly about bridge design and construction tended to have a much more immediate influence on the nature of design as practiced within the profession. Having technical papers noticed and discussed in the Transactions of the American Society of Civil Engineers was a guarantee for gaining recognition among engineers, but an alternative way to disseminate one’s ideas and methods was book publication.

  Theodore Cooper was a frequent contributor of papers to meetings and journals of the American Society of Civil Engineers, and he twice received that society’s prestigeous Norman Medal for his contributions. His 1880 medal-winning paper, “The Use of Steel for Railway Bridges,” showed him to be innovative, for no major bridge had yet been built entirely of the relatively new material. In the mid-1880s, he first published his book, General Specifications for Iron Railroad Bridges and Viaducts, which has been described as comprising “the first authoritative specifications on bridge construction that had been published and circulated.” The title soon expanded to include steel bridges. Cooper’s book was issued in its seventh edition in 1906.

  Among the things that made Cooper’s name most well known among engineers was his system of accounting in the design process for the loads of railroad trains on a bridge structure. He represented the heaviest locomotive then known by means of the forces it exerted through its driving wheels, and represented the train it pulled as a single, uniformly distributed load related to those forces. This made it convenient to modify earlier bridge designs as locomotives and the cars they pulled became heavier, which they seemed constantly to do. Cooper’s system was widely adopted, and by the early twentieth century had become “the almost universal standard for railway bridge design in America.” As convenient as his method was for dealing with trains of increasing weight, Cooper also strongly advocated the more accurate method of representing the load of a train on a bridge by the individual loads transmitted through each wheel. He published tables that made it possible for design engineers to carry out such analyses rapidly and conveniently. All such refinements in calculation meant, of course, that railroad bridges could be designed more accurately, and therefore more economically. No undue iron or steel needed to be added because of uncertainties as to how the bridge might be loaded by a heavy train moving across it.

  Cooper’s 1889 paper on American railroad bridges constituted a concise history, beginning with seventeenth-century wooden bridges and concluding with a section on the failure of bridges, then a matter of increasing concern to the railroads and their passengers. What the disaster of Tay Bridge was to Britain, the collapse of the Ashtabula Bridge almost three years to the day earlier, on December 29, 1876, had been, even more immediately, to American bridge building. The bridge had been erected in the mid-186os across a deep gorge at Ashtabula, Ohio, about fifty miles east of Cleveland on Lake Erie, on the Lake Shore & Michigan Southern Railroad. It was basically what is known as a Howe-type truss design, but with additional diagonals, and with cast iron replacing wooden members. The design had been modified by Amasa Stone. Born in Charlton, Massachusetts, in 1818, Stone began his career as a carpenter and, with his brother-in-law, William Howe, the inventor of the truss type, acquired a contract to build the first railroad bridge over the Connecticut River, in 1840, and with a partner, acquired patent rights to the Howe truss in 1842. He went on to a remarkable railroad career, which included being president of the Cleveland, Painesville & Ashtabula Railroad, which he had built, and which had merged with the Lake Shore & Michigan before the bridge collapsed. The 165-foot span was being crossed by a train of two locomotives pulling eleven railroad cars westbound at fifteen miles per hour, when the driver of the first locomotive felt the bridge sinking. He opened his throttle wide and got safely across, but the other locomotive and all the cars went down with the bridge. At the time, it was still snowing after a recent blizzard; eighty people died in the icy wreck.

  According to some accounts, Stone had been warned by engineers that the structure would be unsafe, and it was generally assumed that the bridge was cheaply built. However, after the accident it was found that, at $75,000, the bridge had actually been relatively costly to build. The engineer responsible for inspecting the structure became so disturbed by the blame heaped upon him that he committed suicide. Although Stone and his design were eventually condemned by the American Society of Civil Engineers, the exact cause of the failure is still not known with absolute certainty. One possible explanation is that a car or cars became derailed, perhaps when the driver lurched his locomotive forward because he thought the bridge was falling, and the truss was so damaged by the impact of the wheels that it could no longer support the unusual load. Whatever the cause, increased attention was devoted to bridge building in America after 1876.

  Cooper wrote, more than ten years after the Ashtabula accident, that it “not only alarmed the general public, but also shook the blind confidence of the railroad companies in their existing bridges,” and he outlined how the situation had changed in the meantime. Subsequent inspections of bridges and procedures uncovered weaknesses, not only in existing bridges but also in existing ways of designing and building them. Cooper was especially critical of designs that did not balance careful analysis with judgment about constructability. He documented how the loads that railroad trains exerted on longer bridges had about doubled from 1873 to 1889, and he called for more precise methods of analysis. In particular, Cooper advocated the adoption of his method based on locomotive wheel loadings, which he acknowledged had been “worked out independently but simultaneously by Mr. Robert Escobar, C.E., of the Union Bridge Company,” in 1880.

  Cooper called attention to an 1878 publication of his own as being “the first paper on bridge construction in which that relic of ignorance, the ‘factor of safety,’ was entirely omitted.” He recognized that bridges should be built with a considerable measure of strength beyond that calculated, but he believed that safety should be based upon rational calculation coupled with careful material testing and inspection, rather than upon an arbitrary numerical multiplier. He opposed “stringent rules,” what today would be called codes of practice, that placed mandatory numerical requirements on such things as factors of safety. Good bridge building, he maintained, could not be based “merely upon a theory of stresses.” Rather, it “must provide for all those requirements of stability and solidity which are instinctively recognized by the practical engineer, and which cannot be complied with by merely using a large factor of safety.” Cooper could not have known at the time how ironic his own words would become.

  Toward the end of his authoritative paper, Cooper summarized the benefits that had accrued from “the American system of competitive bridge construction,” which, according to him, involved aspects of evolution of truss designs, economy of details, and advances in theory absent from British practice. But Cooper also acknowledged some limitations of the competitive system, including “the American idea of building cheap railroads far in advance of the immediate demands of the regions through which they run, to settle those districts and build up a future paying traffic,” which “compelled the use of cheap and light bridge structures,” thus “favoring th
e lowest bidder.” He insisted, however, that bridges used for the traffic originally designed were safe, and that it was as irresponsible for engineers to build excessively strong bridges as excessively weak ones. He approvingly quoted a Professor Unwin:

  If an engineer builds a structure which breaks, that is a mischief, but one of a limited and isolated kind, and the accident itself forces him to avoid a repetition of the blunder. But an engineer who from deficiency of scientific knowledge builds structures which don’t break down, but which stand, and in which the material is clumsily wasted, commits blunders of a most insidious kind.

  In other words, Unwin, like Cooper, believed that bridges should be made strong enough to perform their function but not so strong that they are heavier and more expensive than they have to be. This view has always been and always will be shared by the best of engineers, for they recognize that in the final analysis engineering is part of a much larger social enterprise, and money spent unnecessarily for civil-engineering structures becomes unavailable for other civic endeavors, or for initiatives of a humanitarian kind. The line between too little and too much safety is not always very clear and distinct, however, and that is what makes the best engineering also the most difficult. It is also what can lead to the worst engineering.