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

by Henry Petroski

  Cooper’s reputation was so solid in 1894 that he was appointed by President Grover Cleveland to a commission of five expert engineers that was to recommend the length of span which would be safe and practical for a bridge across the Hudson River at New York City. Two competing bridge companies had proposed two competing bridge types, one a cantilever and one a suspension bridge. The cantilever would have had a span greater by three hundred feet than the recently completed Forth Bridge, and the suspension bridge, with a thirty-one-hundred-foot main span, would have been almost twice as long as the Brooklyn Bridge. The commission’s rejection of a cantilever in favor of a suspension bridge for this dramatic site touched off a debate that was to last for years, and would include considerations relating to the interference with shipping, the length of the maximum possible suspension bridge, and the alternative of tunnels under the Hudson. That saga will take the entire next chapter to relate.

  Engineering News, in part because it was based in New York, took considerable interest in the issue of bridging the Hudson, and the extended discussion of cantilever bridges in its pages prompted one reader, perhaps unfamiliar with the discussion of the word that had taken place years earlier, to inquire as to why that spelling was used, rather than “cantaliver,” as recommended by the Century Dictionary. Engineering News admitted that, whereas it followed that “eminent authority” in such Americanized spellings as “center” and “gage,” it preferred the “cantilever” spelling “as being more euphonious and as least upsetting long-established and well-nigh universal usage.” The editors went on to reopen the question of etymology, admitting that “the origin of the word is uncertain,” an opinion to be echoed by the ponderous and authoritative Oxford English Dictionary, in which “no satisfactory suggestion could be offered” for the word that had become ever more familiar to many engineers and laypersons alike. According to Century, the editors reported, the “probable original … was ‘quanta libra,’ of what weight or balance,” and this led to the dictionary’s spelling preference. Nevertheless, the Universal Dictionary, copyrighted in 1897, was to relegate the spelling “cantaliver” to “unusual” status. Under the definition of the word relating to bridge building, this dictionary credited, perhaps because of the popularity of Benjamin Baker’s lecture and its human model, the Japanese with “the earliest known application of the principle,” and described quite precisely the already “celebrated Forth Bridge” as comprising a “double cantilever (of 1,360 ft. length) … connected by girders 350 ft. long,” thus adding up to the remarkable 1,710-foot spans that impressed Cooper and so many of his contemporaries.

  The Forth Bridge had clearly become the cantilever of the world, to the virtual exclusion of its progenitors. Under the entry “cantilever bridge,” for example, the Universal again was quite technically correct in describing such a structure as one “constructed on the cantilever system, the two sides being pushed out towards the centre and supported by a greater weight on land, until they meet and are joined at the centre.” Though the dictionary acknowledged that “numerous important bridges” were built on the principle, none were named in the entry. The temporary superstructure of the Eads Bridge in particular was no more mentioned in this American dictionary than it was in Baker’s lecture. Yet it was Eads who, almost three decades earlier, in his argument for the advantages of the arch over the truss, had so convincingly employed the bent, angled, or canted lever to make his point for his design for a bridge across the Mississippi. Indeed, according to Cooper, who could take personal pride in its construction, the St. Louis bridge was “the first practical solution of the cantilever principle on a large scale.” He went on to observe that “the erection of two balanced cantilevers, each over 250 feet, with ease, safety and economy, made clear to the mind of engineers that the cantilever was the economic method of erecting long spans over deep gorges or rivers, where ordinary methods of scaffolding would be too expensive, or subject to great risks, or where navigation forbids the obstruction of the waterway.” In fact, to the casual observer, photographs of the Eads Bridge under construction can be easily mistaken for those of the Forth Bridge. I once assembled a series of slides for a lecture by squinting at them before a light bulb and was later embarrassed to find that the Eads Bridge under construction had been projected on the screen while I was describing the Forth.

  An early proposal for a cantilever bridge across the St. Lawrence River at Quebec (photo credit 3.14)

  Cooper would not have made such a mistake. He did, however, near the end of his career, supervise the construction of a great cantilever from such a distance as to lose sight of some of the details that earlier he had written about so authoritatively. The history of his bringing to reality the idea for a bridge across the St. Lawrence River, near the city of Quebec, is, like the history of virtually all great bridges, long and arduous. A span was proposed as early as 1852 by Edward Wellman Serrell, who had by then designed and built the Lewiston & Queenston Suspension Bridge over the Niagara River between New York and Canada. That bridge was advertised as the “Largest in the WORLD!!!” because its stone towers atop the bluffs were 1,040 feet apart, but its deck was only 849 feet long because of the manner in which it joined the shores below the bluffs. Guy wires, mimicking those on Roebling’s nearby Niagara Gorge Suspension Bridge, were added after a storm in 1855, but they were detached in early 1864, when they were threatened by an ice jam that had formed in the river. The guys were still unfastened when a gale struck on the first day of February of that year, and the bridge was destroyed. Serrell’s proposal for a railway and highway bridge at Quebec was not acted upon at the time, but the site he identified was to become the location for another ill-fated bridge more than forty years later.

  4

  The Quebec Bridge Company was incorporated in Canada in 1887, with the authority to issue bonds and the right to build a railway bridge that might also serve for pedestrians and vehicles to cross the St. Lawrence River. Construction was to begin within three years, but three extensions were granted by Parliament, with the last due to expire in 1905. Though other legislation changed the name of the organization to the Quebec Bridge & Railway Company, it continued to be known by its shorter name. Legislation also declared the bridge to be for the general advantage of Canada, and so subsidies were granted to allow the work to begin in earnest.

  In 1897, E. A. Hoare, chief engineer of the bridge company, wrote to the president of the Phoenix Bridge Company in Pennsylvania and asked that any of its engineers planning to attend the annual meeting of the American Society of Civil Engineers in Quebec that June stop in to see him regarding a bridge project. Among those attending the meeting was John Sterling Deans, chief engineer of Phoenix, and he and many other visiting engineers were taken to the bridge site. Also in the group was Theodore Cooper, and within about a week Deans had written to Hoare that Cooper would be happy to lend his experience to the project, which called for a bridge with piers sixteen hundred feet apart. When invitations to bid were issued in 1898, specifications for a cantilever bridge were included; bidders proposing any other kind of bridge would have to provide their own specifications. Among the tenders received were four cantilever and three suspension-bridge designs, including a cantilever from the Keystone Bridge Company and both a cantilever and a suspended span from the Phoenix Bridge Company. Since any type of bridge at Quebec would necessarily be on a massive scale, the advice of an experienced consulting engineer was sought, and Theodore Cooper agreed to review all the plans and tenders. In spite of his prior relations with Phoenix, Cooper’s integrity as an engineer was believed to be above favoring their design for any but sound technical and economic reasons.

  Cooper preferred the cantilever designs, because he believed them to be realizable at a lower cost than suspension-bridge alternatives. In mid-1899, he reported to the Quebec Bridge Company his conclusion that the Phoenix cantilever design was indeed the “best and cheapest” overall. (In fact, Keystone’s design was considerably less expensive per
ton of steel, but it was more costly overall, since the bridge itself would have been much heavier.) Along with his recommendation, Cooper called for further exploration of the riverbed, in order to establish the final position of the bridge’s foundations and thus set a final determination of its length. Deeper foundations, for example, could require more time and money to construct than would a longer span. In early 1900, Cooper had the additional information he had requested; after studying the situation for three months, he concluded that the original pier locations, sixteen hundred feet apart, would take a year longer and be accompanied by more “real and imaginary contingencies” than shallower piers farther apart, and he recommended increasing the main bridge span to eighteen hundred feet—thus proposing a cantilever bridge with a span longer than any in the world.

  In the meantime, negotiations were going back and forth between the Quebec and Phoenix companies, with the latter concerned about the financial status of the former. Detailed design work did not progress very quickly under such circumstances, for Phoenix was not assured of being paid for its services. It was not until 1903, when plans for a National Transcontinental Railway project were revealed, that a bridge at Quebec became such a necessity that government backing was assured. By then the government had also become more interested because of planning for the Quebec Tercentenary in 1908, and it was intimated that the bridge should be ready for the celebration. Thus the pace of design work was suddenly accelerated in 1903—with consequences that were only to be realized years later.

  Cooper’s changes relevant to an eighteen-hundred-foot span were sent to Phoenix, where Peter L. Szlapka, the company’s design engineer, raised some questions about the degree to which some of the steel was stressed. In the final analysis, however, the exceptional magnitude of the structure was invoked as justification for the exceptional loading of its parts. The new specifications were submitted to Collingwood Schreiber, chief engineer of the Department of Railways and Canals, for the required government approval, and Schreiber proposed to his superior that the department “employ a competent bridge engineer to examine from time to time the detailed drawings of each part of the bridge as prepared, and to approve of or correct them,” and submit them to Schreiber for final approval. When Cooper learned of this, he wrote to Hoare at Phoenix: “This puts me in the position of a subordinate, which I cannot accept.… I have written to Mr. Schreiber that I do not see how such an engineer could facilitate the progress of the work or allow me to take any responsible steps independently of his consent.” In other words, Cooper wanted to have the last word on, and the full credit for, the design of the longest bridge in the world. A couple of weeks later, Cooper went to Ottawa to meet personally with Schreiber, after which the minister of railways and canals was advised that, “provided the efficiency of the structure be fully maintained up to that defined in the original specifications attached to the company’s contract, the new loadings proposed by their consulting engineer be accepted.” Though all plans were to continue to be submitted to Schreiber, Cooper was, for all practical purposes, to be the final authority.

  Theodore Cooper, like Benjamin Baker, never married; their bridges were their children. As the Forth Bridge was Baker’s magnum opus, so the Quebec was to be Cooper’s. Cooper had few equals in America, but at the time of his involvement in the Quebec Bridge he was an elderly man in poor health, which pretty much kept him confined to New York. Though he had visited the Quebec site on several occasions while the piers were being constructed, he never once went to Quebec during the erection of the steel superstructure. Cooper may have been “de facto, chief engineer,” and thus ultimately responsible for checking every aspect of the bridge design, but he had no staff in New York to assist him. The design of the bridge, as far as selecting the sizes of members and checking that they were not overloaded, was done to Cooper’s specifications and modifications at the Phoenix Bridge Company by Szlapka, a German-trained engineer who over the course of twenty-odd years with Phoenix had worked on many major projects. Szlapka was, however, a desk engineer, without experience in the erection phase of bridge building, and so was not necessarily in a position to judge the structure itself on his own visits to the construction site. Yet Cooper, who was known for his hypercritical disposition, had full confidence in Szlapka’s work, accepting it on faith when Cooper could not study it thoroughly himself; he had little concern in 1907 that the bridge was progressing in any but a normal fashion.

  The south arm of the Quebec Bridge had been cantilevered out about six hundred feet over the St. Lawrence River by early August 1907, when it was discovered that the ends of pieces of steel which had been joined together were bent. Cooper was notified, by letter, by Norman R. McLure, a 1904 Princeton graduate who was “a technical man” in charge of inspecting the bridge work as it proceeded, who suggested some corrective measures. Cooper sent back a telegram rejecting the proposed procedure and asking how the bends had occurred. Over the next three weeks, in a series of letters back and forth among Cooper, chief engineer Deans, and McLure, Cooper repeatedly sought to understand how the steel had gotten bent, and rejected explanation after explanation put forth by his colleagues. Cooper alone seems to have been seriously concerned about the matter until the morning of August 27, when McLure reported that he had become aware of additional bending of other chords in the trusswork and, since “it looked like a serious matter,” had the bends measured; he explained that erection of additional steel had been suspended until Cooper and the bridge company could evaluate the situation.

  Yet, even as McLure went to New York to discuss the matter with Cooper, Hoare, as chief engineer of the Quebec Bridge Company, had authorized resumption of work on the great cantilever. As soon as McLure and Cooper had discussed the bent chords, Cooper wired Phoenixville: “Add no more load to bridge till after due consideration of facts.” McLure had reported that work had already been suspended, and so contacting Quebec more directly was not believed to be urgent, but when McLure went on to Phoenixville, he found that the construction had in fact been resumed. Some conflicting reports followed, thanks in part to a telegraph strike then in progress, as to whether Cooper’s telegram was delivered and read in time for Phoenixville to alert Quebec.

  In any event, the crucial telegram lay either undelivered or unread as the whistle blew to end the day’s work at 5:30 p.m. on August 29, 1907. According to one report, ninety-two men were on the cantilever arm at that time, and when “a grinding sound” was heard, they turned to see what was happening. “The bridge is falling,” came the cry, and the workmen rushed shoreward amid the sound of “snapping girders and cables booming like a crash of artillery.” Only a few men reached safety; about seventy-five were crushed, trapped, or drowned in the water, surrounded by twisted steel. The death toll might also have included those on the steamer Glenmont, had it not just cleared the bridge when the first steel fell. Boats were lowered at once from the Glenmont to look for survivors, but there were none to be found in the water. Because of the depth of the river at the site, which allowed ocean liners to pass, and which had demanded so ambitious a bridge in the first place, the debris sank out of sight, and “a few floating timbers and the broken strands of the bridge toward the … shore were the only signs that anything unusual had happened.” The crash of the uncompleted bridge “was plainly heard in Quebec,” and the event literally “shook the whole countryside so that the inhabitants rushed out of their houses, thinking that an earthquake had occurred.” In the dark that evening, the groans of a few men trapped under the shoreward steel could be heard, but little could be done to help them until daylight. The sounds of the bridge falling and the moans of the lives it claimed would echo around the world for days, weeks, and years to come.

  The south arm of the Quebec Bridge, as it appeared just before its collapse on August 29, 1907 (photo credit 3.15)

  Within days, the story of the unread telegram was reported as part of the tragedy. According to one version, Cooper filed it in New York befor
e noon and, though delayed by the strike, the wire did reach Quebec in the middle of the afternoon—in time to save the men, if not the bridge. Cooper was first reported to have said that the message included an admonition to get off the bridge at once, but in fact it called only for halting the addition of any more steel to the structure. The less-than-urgent-sounding message thus lay on Deans’s desk until he returned to his site office at about 5 p.m., shortly before McLure himself returned from New York. Even if it had not been too late to clear the structure, no warning might have been given, for the halt asked by Cooper was in fact to buy time to analyze the anomalies that had developed in the structure. It is not clear that Cooper or anyone else believed that collapse was imminent.

  Shortly after the accident, amid speculations as to its causes and as to whether the men might have been saved, Cooper was reported to have reproached himself “for not having visited the work in two years,” and confessed that he had “tried to obtain his release from the responsibility of serving as the consulting engineer” on the Quebec Bridge because of his poor health, but “the builders would not listen to that.” With regard to the ill-fated telegram, Cooper qualified earlier reports regarding its message and pointed out that, serving officially only as consulting engineer to the project, he had no authority to order the workmen off the structure.

  A royal commission was formed immediately to inquire into the cause of the collapse of the Quebec Bridge. The commission comprised: Henry Holgate, a civil engineer from Montreal; John G. G. Kerry, a civil engineer from Campbellford, Ontario; and John Galbraith, dean of the Faculty of Applied Science and Engineering at the University of Toronto. The site was visited the day after the accident, and the taking of evidence commenced in Quebec within two weeks. Cooper himself, who remained in New York, was interviewed by the commission there for a week in mid-October. After the confusing and conflicting newspaper reports concerning what he had said about his telegram in the days immediately following the accident, Cooper had remained silent on the matter, until the commission visited “the Nestor of American bridge designers,” as Engineering News identified him in its report on the visit, thus raising his reputation to mythic proportions. Although he was actually sixty-eight at the time, he was described as “now 70 years of age” and as having been “in poor health for several years.” The trade journal’s sympathetic portrait of the engineer was no less deferential than the commission’s treatment of him: