The Quebec Bridge Company was formed in 1887 and for the next
thirteen years, very little was done. In 1900, Thomas Cooper, a consultant
in civil engineering specializing in bridge building was
appointed the company’s consulting engineer for the duration of the
work. The Sixth Street bridge in Pittsburgh, the Seekonk bridge in
Providence, and the Second Avenue bridge in New York were
already part of his portfolio.
Conscious of the precarious financial situation of the Quebec
Bridge Company, Cooper recommended the cantilever superstructure
proposed by the Phoenix Bridge Company, of Phoenixville,
Pennsylvania as the best and cheapest of the proposals. He also recommended
that the span of the bridge be increased from 1600 feet
to 1800 feet to minimize the cost of constructing the piers supporting
the bridge. The specifications were also modified to allow for
greater unit stresses in the structure. The Quebec Bridge was to be
the longest cantilever bridge in the world.
For the next three years, the assumptions underlying the modified
design of the most technically ambitious bridge in the world went
unchallenged. After the Canadian government guaranteed a bond
issue in 1903, the construction shifted into high gear. In the rush to
minimize delays, the assumed weight of the revised bridge was not
recalculated. Instead, work continued with the estimated weight the
Phoenix Company had provided with the original proposal. Cooper
was personally offended when the Canadian Department of Railways
and Canals requested that the plans be independentlyreviewed and approved. With full confidence in his own design and
expertise, Cooper managed to turn the independent review into a
rubber stamp approval by unqualified individuals.
Subsequent warnings were also summarily ignored. In 1906, the
Phoenix Company’s inspector of material reported that the actual
weight of steel put into the bridge had already exceeded the original
estimated weight. The final weight of the bridge was estimated to
be eleven million pounds higher than originally thought. The alternative
being to start building the bridge all over again, Cooper concluded
that the increase in stresses was acceptable.
In early August 1907, the lower horizontal pieces running the
length of the bridge began to show signs of buckling. The Phoenix
Company insisted that they were already bent when they left the
shop in Phoenixville and work continued. They made no effort to
explain why a deflection had increased by an inch and a half in the
past week. On August 29th, the south arm of the bridge collapsed
under its own weight, killing 73 workers.
The bridge was eventually redesigned and rebuilt, weighing two
and a half times more than its predecessor. Ironically, it suffered a
tragedy of its own in 1916 when the central span fell into the river
while it was being hoisted into place. The bridge was finally completed
in 19181. It is still in use today and it is still the longest cantilever
bridge in the world.
The parallel with today’s micro-electronic designs is obvious. The
next design is always more challenging than the previous one and it
takes the designers into previously uncharted waters. A design cannot
go unchallenged simply because it worked in the last implementation.
Changes in the specifications or new functionality cannot be
assumed to work simply because they were designed by the best
engineers. Errors will be made. It is important that any design be
independently verified to ensure that it is indeed functionally correct.
The alternative is to manufacture a non-functional design.
Hopefully, no one will get killed. But many could lose their jobs.