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

by Henry Petroski

  Among the many pithy phrases in architect Aymar Embury’s conciliatory works, he wrote that “it is always easier to remember than to invent.” It was here that he and the engineers did part ways, though inadvertently. As much as Ammann and his contemporaries referred to early-nineteenth-century suspension bridges, like Telford’s Menai, as aesthetic models, they did not find it easy (or think it relevant) to remember how the light decks were tossed about in the wind. This professional amnesia was endemic by the mid-1930s, and the effects of it began to manifest themselves in a big way. As construction neared completion on the Bronx-Whitestone and preparations were being made for attaching the floor, the deck began to swing when the wind blew at certain angles, and it moved back and forth lengthwise between the towers. Though this “pendulum action” was unexpected, it does not seem to have excessively alarmed Ammann and his associates, who simply installed short guy cables at the bridge’s midspan and braking devices at the towers. There was “marked improvement” in the behavior of the bridge, but the first cables had to be replaced by heavier ones. These slight modifications to control the motion of the new structure were apparently taken in stride, and the Bronx-Whitestone opened on April 30, 1939, in time for the World’s Fair.

  Contemporary suspension bridges had been designed to the same structural aesthetic as the George Washington and Bronx-Whitestone, and some of those other bridges were also beginning to exhibit excessive motion. Even the Golden Gate Bridge, which at forty-two hundred feet had of course surpassed the George Washington as the longest suspended span in the world, was more flexible than anticipated. In heavy winds, the Golden Gate moved sideways by as much as fourteen feet, but engineers calculated that this movement stressed the bridge less than did the expected variation of temperature. Though the deck of the Golden Gate was stiffened by a conventional deep truss, it was extremely slender relative to its great length, and this resulted in great flexibility. Two much shorter suspension bridges, designed by David Steinman and stiffened like the Bronx-Whitestone with plate girders, also exhibited considerable motion in their decks. The eight-hundred-foot Thousand Islands International Bridge, opened in 1938 over the St. Lawrence River between New York and Ontario, and the 1,080-foot Deer Isle Bridge, opened in 1939 over Penobscot Bay in Maine, behaved similarly to the Bronx-Whitestone and were likewise fitted with restraining devices, albeit of a different nature. Although considerably stiffened since their initial opening, these bridges remained flexible. In 1978, for example, hundreds of people, including Joan Mondale, the wife of the then vice-president, were stranded on Deer Isle for several hours when the deck of the two-lane bridge began “swelling.” A local resident characterized the bridge as normally having “a certain amount of play in it,” but on that day the sway in the light breeze was far greater than during the seventy-mile-per-hour winds of the previous winter.

  The Bronx-Whitestone Bridge after stiffening trusses were added in the mid-1940s (photo credit 5.19)

  Such vagaries of behavior of their bridges in the wind led to considerable rethinking about calculations and testing of models by engineers in the late 1930s and into 1940, but most must have thought as Ammann did: “We have had to deal with very small movements, and would have felt no concern about them had they not tended to produce discomfort in some persons under unfavorable conditions.” The movements were in most cases “very small” relative to the size of the structure (on the order of one foot in a thousand, for example) and were akin to the swaying of a skyscraper today. Although such motions are not thought to be life-threatening to the structure or its inhabitants, they can be psychologically distracting and can have adverse economic implications if people do not want to occupy a swaying skyscraper or cross a swinging bridge. In 1940, however, few seemed to be excessively worried about such things.

  Among the official consulting engineers on the George Washington Bridge project were two who, with Ammann, were responsible for the very flexible bridges that followed from the aesthetic imperative. These were, of course, Joseph Strauss and Leon Moisseiff. Just as Strauss was a consultant for the George Washington Bridge, so Ammann fulfilled a similar role for the Golden Gate. Such interlocking relationships, common among the engineering elite, explain to a considerable degree how the state of the art can advance almost in lockstep, so that many different structures, in this case suspension bridges, can share the same aesthetic characteristics—and the same behavioral flaws.

  It was Moisseiff’s development of the deflection theory that enabled all the slender and flexible bridges to be designed in the first place. Leon Solomon Moisseiff was born in Latvia in 1872, when it was part of the Russian empire. He attended the Baltic Polytechnic Institute in Riga for two years, but his student political activities reportedly led his family to emigrate to the United States in 1891. They settled in New York City, where Leon worked as a draftsman before enrolling in Columbia University, from which he graduated in 1895 with a degree in civil engineering. He worked as a draftsman for the New York Rapid Transit Railroad Commission, as a design engineer with the Dutton Pneumatic Lock and Engineering Company (doing work on drydocks, gates, and lock improvements proposed for the Erie Canal), and as a draftsman with the Bronx Department of Street Improvements, before joining the New York Department of Bridges in 1898 as chief draftsman and assistant designer. It was in this position that he worked on the Williamsburg, Queensboro, and Manhattan bridges and met Gustav Lindenthal. In 1910, Moisseiff became engineer of design for the Bridge Department, and in 1915 he struck out on his own as a consulting engineer. In 1920, he was appointed chief designer of the Delaware River Bridge, which remained his office’s major project until the record span was completed in 1926. Thereafter, through 1940, he consulted on virtually every major suspension-bridge project in the United States, including many an engineer’s dream that did not materialize. Even if biographical sketches and memoirs have to be taken with a grain of salt, since they often rely so heavily on the word of friends and relatives, sometimes they do contain a grain of truth: “Although he did not always receive formal credit, Moisseiff was the principal designer of the George Washington, Bronx-Whitestone, Tacoma Narrows, and Mackinac bridges.”

  Leon Moisseiff (photo credit 5.20)

  II

  The twenty-eight-hundred-foot main span of the Tacoma Narrows Bridge made it the third-longest suspension bridge when it was completed in 1940. In keeping with the engineering aesthetic and economic thinking of the times, the bridge deck was stiffened with plate girders. However, although the Tacoma Narrows’ main span was five hundred feet longer than that of the Bronx-Whitestone, only eight-foot-deep girders were used to stiffen the roadway, because it was much narrower than that of the New York span. The combination of a longer span with shallower depth and narrower width made the Tacoma Narrows Bridge more flexible than any other. Nevertheless, the two-lane crossing of the Narrows, about thirty miles south of downtown Seattle, provided a reasonable highway alternative to taking ferries between Seattle and the Olympic Peninsula, across Puget Sound. As soon as the Tacoma Narrows opened in July, drivers noticed how flexible it was, the wave motion of its roadway bringing cars that were ahead of a driver on the bridge alternately into and out of view as the pavement rose and fell. Rather than scaring toll-paying customers away, however, the bridge became affectionately known as Galloping Gertie and attracted even more traffic as an unintended amusement ride. Although some of the riders were reported to have become seasick, traffic over the bridge in its first two weeks of operation was twice what engineers had expected.

  A bridge across the Narrows had been proposed as early as 1933 by the Tacoma Narrows Bridge Company, which had obtained a franchise and was then seeking capital. However, the growing sentiment for publicly owned bridges and utilities led to a competing application by Pierce County, on the peninsula. In 1937, no doubt fueled by the success of bridges like the George Washington and the almost complete Golden Gate, the State Legislature created the Washington Toll Bridge Authority, which
took over the Pierce County initiative and its application for a construction grant from the federal Public Works Administration. As was typical, various conceptual designs had been considered, including a cantilever bridge and a multiple-span suspension bridge such as had been recently completed across San Francisco Bay. By mid-1938, the State Highway Department had made a preliminary design for a suspension bridge with a single main span of twenty-six hundred feet, and two side spans of thirteen hundred feet each, all resting on a stiffening truss twenty-two feet deep. The structure was to carry a twenty-six-foot-wide roadway and two four-foot sidewalks. The total width of the bridge, including the stiffening truss, was to be only thirty-nine feet—a remarkably narrow deck relative to the length of the bridge.

  The Tacoma Narrows Bridge, when it was opened in July 1940 (photo credit 5.21)

  As consulting engineer, Moisseiff was asked to study the Highway Department’s design, and he submitted his initial report in July 1938. His first criticism addressed the unequal tower heights. Though these had been chosen to accommodate the unequal elevations of the two ends of the bridge, they meant that the entire roadway of the preliminary design had an upward incline toward the higher, Tacoma shore, and the consulting engineer criticized it in no uncertain terms:

  Unless there are very valid reasons which compel the making of the towers of unequal heights the towers should be of identical design and fabrication. Economic fabrication and good appearance demand it. The symmetry of the structure should be adhered to.

  Moisseiff’s solution was to “raise the west end of the bridge by 19.5 ft.” With regard to the twenty-two-foot stiffening truss of the Highway Department design, the consultant found that it could not “effectively stiffen the bridge except at great cost.” He proposed eight-feet-deep plate girders, which would not only “result in a neat and pleasing appearance” but also “be about one cent per lb. less than for a truss.” And he reported that his studies showed it “best to attain rigidity by shortening the side spans and by a reduction of sag ratio” in the cables. This was “not only a better solution but also the cheapest,” he concluded. In a second part of his report, Moisseiff recommended further that the spacing of the suspenders supporting the roadway from the main cables be increased from thirty to fifty feet, not only to achieve a more pleasing appearance but to effect a further savings of about $35,000 out of the total estimated cost of $6 million. Ironically, he also argued for keeping the “height of the towers to a minimum due to the relatively great effect of transverse wind pressure.”

  In September 1938, an application to secure a federal loan for the bridge project was submitted by the Washington Toll Bridge Authority to the Reconstruction Finance Corporation, which funded numerous projects of the Public Works Administration. As was standard procedure, the application was referred to the Legal, Finance, and Engineering divisions of the Administration, but it was a review for the Reconstruction Finance Corporation that raised the strongest concerns about the soundness of the project. Theodore L. Condron, advisory engineer to the bond purchaser, was a septuagenarian consulting engineer best known for designing the steel structure for the seventy-two-bell carillon in the Rockefeller Memorial Chapel of the University of Chicago. In his report on the application Condron identified the board of consulting engineers as consisting of Charles E. Andrew, bridge engineer of the San Francisco-Oakland Bay Bridge, and chairman of the board; Luther E. Gregory, a retired navy rear admiral and resident of Olympia, Washington; and R. B. McMinn, a bridge engineer with the U.S. Bureau of Roads in Portland, Oregon. The consulting engineer Moisseiff, and his associate Frederick Lienhard, were identified as “New York engineers” associated with the design of the San Francisco-Oakland Bay and Golden Gate bridges.

  The report of the board of consulting engineers on the Moisseiff-modified plans had found them to be “in satisfactory shape for receipt of bids,” even though the board did not examine the project in detail. Time did not permit a checking of stresses in the cables or stiffening system, for example, but the board had “full confidence in Mr. Moisseiff,” considering him “to be among the highest authorities in suspension bridge design.” With this endorsement, Condron may have thought that the approval of the project before him would have been a routine matter; the more he looked at the plans, however, the more doubts he seemed to have. In particular, Condron had serious reservations because of the extremely narrow width of the proposed bridge relative to its main-span length. When he compared this ratio with that of recently completed suspension bridges, the Tacoma Narrows was definitely more slender than any of them, and thus Condron could not see it as just a routine application of bridge-building experience. Even the Golden Gate Bridge, then the longest suspension span in the world, was not nearly so slender as the Tacoma Narrows design, as Condron’s tabulation showed:

  Advisory engineer Condron may well have known of the surprising flexibility of the Golden Gate Bridge, and he had heard that “certain tests had been made on models of suspension bridge spans” at the University of California. When Condron could find no published reports on those tests, he went to Berkeley to confer with Professor R. E. Davis about concerns over the deflection of the very slender Tacoma design, whose construction loan was awaiting approval. Condron reported that Davis “felt reasonably confident that the lateral deflections of the Tacoma Narrows Bridge as designed and determined by Mr. Moisseiff would be in no way objectionable to users of the bridge.” As if to document as best he could the authority of Moisseiff and the deflection theory, Condron quoted from a 1933 report on the accuracy of calculation that the theory permitted: “Moisseiff and Lienhard have presented a method which is closely accurate for determining lateral deflections of truss and cable stresses in the truss due to lateral forces.” Whereas Condron had gone to Berkeley with questions about vertical as well as lateral deflections, he appears to have been reassured only about the latter, however, and he seems to have gone to lengths in his report to make that point clear. The problems with the bridge would not, of course, be with the lateral deflections.

  Condron continued to have doubts about the design, and even a letter from Moisseiff to him could not put them to rest. When Moisseiff wrote that, considering the slenderness of the bridge, its stiffness was “rather satisfactory,” Condron pointed out that “there seems to be some question even in his mind as to whether the obtained stiffness is other than rather satisfactory.” In the end, however, the consulting engineer to the Reconstruction Finance Corporation acceded to authority and expertise:

  In view of Mr. Moisseiff’s recognized ability and reputation, and the many expressions of approval and comment of his methods of analyses of stresses and deflections in the designs of long span suspension bridges, particularly as expressed by the engineers who participated in the discussion of the paper presented before the American Society of Civil Engineers by Messrs. Moisseiff and Lienhard entitled “Expansion [sic] Bridges under the Action of Lateral Force,” I feel we may rely upon his own determination of stresses and deflections.

  The Freudian slipping of “expansion” for “suspension” into the title of the paper by Moisseiff and Lienhard may have indicated Condron’s fundamental unwillingness to concede that the Tacoma Narrows Bridge was stiff enough. Nevertheless, the weight of evidence presented by experts in the discussion of the key theoretical paper was too much for the lone advisory engineer to refute. In that discussion, the University of California models were repeatedly referred to, and Dean Charles Derleth found that their confirmation of the theory was “gratifying.” He pointed out that the paper of Moisseiff and Lienhard “had its inception in the early debates on the Golden Gate design,” with the authors “seeking a convincing argument to justify shallow stiffening trusses and slender wind-bracing for a 4000-ft. span.” Though “engineers of considerable accomplishment” had argued that deck widths approaching two hundred feet might be necessary, Moisseiff and Lienhard’s method of analysis had justified a ninety-foot roadway and was considered sufficient to
“silence all arguments for unnecessary floor widths.” Derleth also reached beyond mathematical analysis, to a “poetic rather than a mechanical figure of speech,” to emphasize how important the cables were relative to the deck of very long suspension bridges, saying he liked to “describe the theory of Messrs. Moisseiff and Lienhard as assigning to the floor system the nature of a kite in the wind, with the cable and suspenders acting as the restraining strings and tail.” Few but the likes of Condron seem to have worried more about the kite than the strings of the “different species in a genus of suspension bridges” that had been evolving toward the Tacoma Narrows Bridge in the wake of Moisseiff and Lienhard’s theory.

  For all the expertise that was assembled against him and to which he felt obligated to defer, however, Condron could not bring himself to unqualified approval in the conclusion to his report: