Kawada 2010

aded from ascelibrary.org by New York University on 05/10/15. Copyright ASCE.
For personal use only; all rights r

Downloaded from ascelibrary.org by New York University on 05/10/15. Copyright ASCE. For personal use only; all rights reserved.
Suspension Bridge
History of the Modern
Other Titles of Interest
Beyond Failure: Forensic Case Studies for Civil Engineers, by Norbert J.

Delatte Jr. (ASCE Press, 2009). Narrates the circumstances of important
failures that have had wide-reaching impacts on civil engineering practice.
(ISBN 978-0-7844-0973-2)
Bridginess: More of the Civil Engineering Life, by Brian Brenner. (ASCE

Press, 2009). Offers an entertaining appraisal of the mindset, practice, and
profession of civil engineers, with a focus on the technical, social, and sym-
bolic lives of bridges. (ISBN 978-0-7844-1040-0)
Engineering Legends: Great American Civil Engineers, by Richard G. Wein-

gardt. (ASCE Press, 2005). Sketches the lives and achievements of 32 great U.S.
civil engineers, from the 1700s to the present. (ISBN 978-0-7844-0801-8)
In the Wake of Tacoma: Suspension Bridges and the Quest for Aerodynamic
Stability, by Richard Scott. (ASCE Press, 2001). Comprehensively describes
the changes imposed on the design of suspension bridges as a result of the 1940
collapse of the first Tacoma Narrows Bridge. (ISBN 978-0-7844-0542-0)
Landmark American Bridges, by Eric DeLony. (ASCE Press, 1993). Presents

photographs and descriptions of more than 990 historic bridges, tracing the
evolution of U.S. bridge building. (ISBN 978-0-87262-857-1)
Washington Roebling’s Father: A Memoir of John A. Roebling, edited by

Donald Sayenga. (ASCE Press, 2009). Portrays the life and achievements of
a legendary engineer through the eyes of his equally accomplished son, with
annotations for hundreds of people, places, events, and technologies. (ISBN
978-0-7844-0948-0)
History of the Modern

Suspension Bridge
Solving the Dilemma between
Economy and Stiffness
Tadaki Kawada
Translated by Harukazu Ohashi
Edited by Richard Scott
Library of Congress Cataloging-in-Publication Data
Kawada, Tadaki, 1935-

[Kindai tsuribashi no rekishi. English]
History of the modern suspension bridge : solving the dilemma between economy and stiffness /
Tadaki Kawada ; translated by Harukazu Ohashi ; edited by Richard Scott.

p. cm.
Includes bibliographical references and index.
ISBN 978-0-7844-1018-9
1. Suspension bridges—Design and construction—History. I. Scott, Richard, 1956
Feb. 10- II. Title.
TG400.K34213 2011
624.2’309—dc22
2009049073
Published by American Society of Civil Engineers

1801 Alexander Bell Drive
Reston, Virginia 20191
www.pubs.asce.org
Any statements expressed in these materials are those of the individual authors and do not necessar-
ily represent the views of ASCE, which takes no responsibility for any statement made herein. No
reference made in this publication to any specific method, product, process, or service constitutes
or implies an endorsement, recommendation, or warranty thereof by ASCE. The materials are for
general information only and do not represent a standard of ASCE, nor are they intended as a refer-
ence in purchase specifications, contracts, regulations, statutes, or any other legal document.
ASCE makes no representation or warranty of any kind, whether express or implied, concern-
ing the accuracy, completeness, suitability, or utility of any information, apparatus, product, or
process discussed in this publication, and assumes no liability therefor. This information should
not be used without first securing competent advice with respect to its suitability for any general
or specific application. Anyone utilizing this information assumes all liability arising from such
use, including but not limited to infringement of any patent or patents.
ASCE and American Society of Civil Engineers—Registered in U.S. Patent and Trademark Office.
Photocopies and reprints. You can obtain instant permission to photocopy ASCE publica-
tions by using ASCE’s online permission service (http://pubs.asce.org/permissions/requests/).
Requests for 100 copies or more should be submitted to the Reprints Department, Publications
Division, ASCE (address above); e-mail: permissions@asce.org. A reprint order form can be
found at http://pubs.asce.org/support/reprints/.
Cover photographs courtesy of Jason Hsu /Wikimedia Commons (Akashi Kaikyo Bridge) and
Mary K. Baird /MorgueFile (Spider web).
Copyright © 2010 by the American Society of Civil Engineers.

All Rights Reserved.
ISBN 978-0-7844-1018-9
Text originally published in Japanese by Kensetsutosho.
© 2002 Tadaki Kawada
ISBN 4-87459-216-3
Manufactured in the United States of America.
18 17 16 15 14 13 12 11 10 12345
Contents
preface to the english edition . . . . . . . . . . . . . . . . ix

preface. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xiii
acknowledgments . . . . . . . . . . . . . . . . . . . . . . . . . . . xv
1 The Emergence of Modern

Suspension Bridges . . . . . . . . . . . . . . . . . . . . . . . 1
The Universality of Primitive Suspension
Bridges and Their Structure 1
Suspension Bridges Were Unknown to Western
Civilization until the Medieval Period 8
The Birth of the Modern Suspension Bridge 18
Finley’s Glory and Setback 21
2 Trials and Errors in Europe . . . . . . . . . . . . . . . . . 27

Introduction to the United Kingdom 27
Accidents in the United Kingdom and
Problems Relating to Stiffness 36
Introduction to France 46
French-Style Suspension Bridges 52
Early Suspension Bridge Theory and
Its Limitations 60
v
vi history of the modern suspension bridge
3 North American Engineers Develop the

Rigid Suspension Bridge . . . . . . . . . . . . . . . . . . 71
Homecoming to America 71
The Niagara Railway Suspension Bridge 77

John Roebling 80
The Brooklyn Bridge 86
4 Emergence of the Deflection Theory

and Its Development . . . . . . . . . . . . . . . . . . . . . 97
Theory of the Stiffened Suspension Bridge 97
The Emergence of the Deflection Theory 102
The Development of Suspension Bridges in
the Early Twentieth Century 106
The Trend toward Longer Spans 110
Two Large Projects in San Francisco 117
5 Ambush: The Dynamics of Wind . . . . . . . . . . . . 129

Disaster of the Tacoma Narrows Bridge 129
The Dynamics of Wind 135
Inevitable Accident or Human Error? 140
Aftermath in America following the Tacoma
Narrows Bridge Disaster 146
6 New Developments beyond North America . . . . 159

From North America to Europe: British
Challenge for Innovation 159
Aerofoil Revolution 164
Glory and Tragedy of the Severn Bridge 172
Mass: Engineers’ Forgotten Asset 182
7 Two Record-Breaking Suspension Bridges . . . . . 193

Japan’s Akashi Kaikyo Bridge 193
Denmark’s Storebælt Bridge 201
The End of the Century: The Millennium
Bridge 214
contents vii
8 Epilogue: No End to the Dilemma . . . . . . . . . . . 225

The Development of Modern Suspension
Bridges and Stiffness 225
Prospects for the Twenty-First Century 227

Lessons from History: Concluding Remarks 233
index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 237
about the author, translator, and editor . . . . . 245
This page intentionally left blank

Preface to the English Edition
Japan, from humble forays into the field in the

1950s, today stands as a leading builder of suspension
bridges. Its engineers have taken tremendous strides over
the past four decades, brilliantly adapting contemporary
practice emerging out of the ashes of the 1940 collapse of
the Tacoma Narrows Bridge, and devising new methods
of analysis, design, construction, and maintenance. For
more than a decade the longest suspension bridge in the
world—the Akashi Kaikyo Bridge—has stood as dramatic
testimony to this remarkable progress, as have many other
large suspension bridges built in this island nation.
No small figure in this transformation has been
Dr. Tadaki Kawada. Dr. Kawada’s long association with
suspension bridges—as an academic, practicing engineer,
and head of a large bridge construction firm—has brought
him into intimate contact with the theory, history, design,
and building of large suspension bridges. His association
with the massive Honshu-Shikoku project, which alone
involved 10 suspension bridges, has occupied much of his
working life. This experience has brought profound knowl-
edge, deep understanding, and a unique perspective to the
story that unfolds over the pages that follow—a story of
suspension bridges, the largest, most dramatic, and grace-
ful of bridge forms.
ix
x history of the modern suspension bridge
Dr. Kawada’s narrative is enriched by two other qualities. His passion

for the historical aspects of suspension bridge development has been mani-
fested in his authorship of several Japanese-language books on the subject.
These books have contributed much to that nation’s knowledge and appre-
ciation of the international nature of suspension bridge history. His exten-
sive contributions to Japanese developments in the field have also infused
this book, providing a Japanese perspective on a history long penned by
North American and European writers.
Now, an English-language audience has the opportunity to benefit
from Dr. Kawada’s experience and knowledge. Dr. Kawada uses an impor-
tant theme, that of tension between the economy of construction and the
need for stiffness, to recount the trials and tribulations of the fascinating
and sometimes turbulent history of the modern suspension bridge. This ten-
sion (perhaps an appropriate metaphor, given the subject) has consistently
shaped suspension bridge design since the inception of the modern form by
American Judge Finley in the early 1800s. Through this lens, Dr. Kawada
vividly emphasizes that the lessons of yesteryear—even those of two centu-
ries past—remain relevant to practicing engineers today and to the public
that crosses the bridges engineers design, build, and maintain.
The lessons learned in the aftermath of the Tacoma Narrows Bridge
collapse generated new knowledge to build safe, efficient, and stable suspen-
sion bridges. But these structures—even in today’s world of vast engineering
knowledge, sophisticated modeling, and computing power—can still pro-
duce unexpected and unpleasant surprises, as attested by the vexing (but
eventually solved) vibration of London’s ultramodern Millennium pedes-
trian bridge across the Thames in 2000. Dr. Kawada’s theme of balancing
economy and stiffness provides an absorbing framework by which to explore
the range of design challenges—from static and aerodynamic stability all the
way to mass psychology—that have faced engineers from the earliest days
of the suspension bridge to the structures of today. Moreover, this approach
has enduring utility in assessing where the future development of suspension
bridges may lead, as Dr. Kawada explores at the book’s conclusion.
In preparing this translation of Dr. Kawada’s book, we would like to
express our thanks to a number of people who made this project possible.
First, our sincere appreciation goes to Dr. Tadaki Kawada for embracing this
project so fully and for assisting us at various points during its preparation,
including furnishing of many of his own photographs and illustrations. We
would also like to recognize the untiring efforts of Dr. Shigeru Echigo, who
performed yeoman service in obtaining the many publishing permissions for
photographs, illustrations, and charts which made it possible for this book
to faithfully mirror the original Japanese edition. Without Dr. Echigo’s per-
preface to english edition xi
sistence, patience, and organization, this book would not be what you see
before you. Our sincere thanks also go to Betsy Kulamer of ASCE Press for
shepherding the book through the myriad processes that lead to publication,
and for her patience, humor, and always-positive and enthusiastic outlook in
the face of what may have at times seemed insurmountable obstacles. Matt
Boyle, ASCE books production manager, provided a friendly and encourag-
ing hand as the book moved through production. Also at ASCE, Corinne
Addison shouldered the task of putting all of the illustration permissions in
order. And finally, our thanks to Anda Divine, whose deft touch at copyediting
is so evident in the final product.
It is our hope that this translation captures the eloquence of the origi-
nal Japanese-language edition of Dr. Kawada’s 2002 book, and in so doing
communicates to future generations of engineers the continuing relevance
of the themes of economy and stiffness for spans that may someday dwarf
even the Akashi Kaikyo Bridge.
Harukazu Ohashi, translator, Tokyo, Japan

Richard Scott, editor, Peterborough, Canada
September 2009
This page intentionally left blank

Preface
This book is intended to contribute to the future

development of suspension bridges by tracing how modern
suspension bridges have evolved from their early form as
they appeared in Western civilization only 200 years ago.
The book explores how suspension bridges have been
developed to make such gigantic structures as the Akashi
Kaikyo Bridge in Japan and the Storebælt Bridge in Den-
mark feasible at the end of the twentieth century. The
developments, trials, and errors experienced during the
course of suspension bridge development are described
and thoroughly illustrated.
“When he was thinking about how to build a bridge
across the River Tweed, Sir Samuel Brown stopped while
observing a spider’s web. Right at this time he discovered
the suspension bridge” (Bender 1872). This anecdote, in
which the concept of the modern suspension bridge was
born through the observation of a spider’s web, is similar to
how Newton gleaned a hint of universal gravitation from
a falling apple.
Compared with masonry and timber bridges, which
dominated bridge construction in the nineteenth century,
a suspension bridge—made feasible by virtue of lower
labor and material requirements and a shorter construc-
tion period—was extremely advantageous economically.
As it became known that a catalyst for this development
was the observation of a spider’s web, modern suspension
xiii
xiv history of the modern suspension bridge
bridges also became flexible and undulating, that is, lacking stiffness. How
to produce the required stiffness in modern suspension bridges without sac-
rificing the superior characteristic of economy is a problem that has since
preoccupied engineers.
The dual problem of stiffness and economy in suspension bridges
introduces conflicting concepts. A heavy structure, while meeting high
stiffness requirements or the need for a robust suspended structure, can be
uneconomical. In contrast, when economy is overly emphasized, the bridge
can lack stiffness, and examples of problems caused by this approach are
too numerous to count. Over the past 200 years of advances in materi-
als and methods, the introduction of new design theories, and the use of
wind tunnels to refine design concepts, modern suspension bridges have
been built with constantly improving technology. With the wisdom of and
unceasing efforts by engineers, the conflict between stiffness and economy
in modern suspension bridges has been nearly conquered, as evidenced by
the enormous Akashi Kaikyo Bridge (nearly 2,000 m in span), completed
in 1998. The modern suspension bridge at the end of the twentieth century
has attained the following status: “The basic concept is that of a suspension
bridge . . . [using] . . . tried-and-tested techniques” (Binney 2000).
However, an unexpected event involved London’s pedestrian-only Mil-
lennium Bridge over the Thames River, which was built in 2000 to celebrate
the two millenia since the birth of Christ. This bridge made full use of “estab-
lished technology,” yet had to be closed and was out of service just 3 days
after its opening due to vibration caused by a lack of stiffness. The ghost of
inadequate stiffness—once thought to have been expunged by engineers after
200 years—was still alive and asserted itself in this state-of-the-art bridge.
The steps taken since the emergence of modern suspension bridges
very much reflect a history of the struggle to conquer the conflict between
stiffness and economy. And, I should say, that this struggle has not yet finally
come to an end at the beginning of the twenty-first century.
Tadaki Kawada
Tokyo, Japan
November, 2002
References
Bender, C. (1872). “Historical Sketch of Improvements in Suspension Bridges to
the Present Time.” Transactions of the American Society of Civil Engineers,
ASCE, New York.
Binney, M. (2000). “Revolutionary design liable to give off bad vibration.” The
Times (London), June 12.
Acknowledgments
This book was started as a doctoral thesis and was

completed under the guidance of Professor Yozo Fujino of
the Graduate School for Engineering Research at the Uni-
versity of Tokyo. I offer my special thanks for his support.
Professor Emeritus Manabu Ito, Professor Osamu Shino-
hara, Professor Masahiko Kunijima, and Associate Profes-
sor Masato Abe of the Graduate School for Engineering
Research of the University of Tokyo, Professor Hiromichi
Higashihara of the Earthquake Research Institute of the
University of Tokyo, and finally Dr. Kunikatsu Nomura
of Kawada Industries, Inc.—all these authorities kindly
reviewed the manuscript before its completion and pro-
vided me with invaluable advice. I deeply thank them all.
For obtaining permissions from many people and
sources who gladly approved the use of photographs, I
must also extend my thanks to Dr. Shigeru Echigo and
his staff at the Research Division of Kawada Industries,
Inc., and Mr. Sean J. Johnston and Ms. Yoko Ohmiya of
Kawada Industries, Inc., who also generously devoted their
time to the tedious work of proofreading and to develop-
ing the index and the figures in this book.
Only upon completion of the book did I fully become
aware of how many people had helped me in its creation. I
sincerely appreciate their efforts.
xv
1
The Emergence of
Modern Suspension
Bridges
Nowadays, there is no doubt that the suspension
bridge constitutes the main solution for long-span struc-
tures. The 1,991-m span Akashi Kaikyo Bridge in Japan
and the 1,624-m span Storebælt Bridge in Denmark are the
longest spans that humanity has built up to the twenty-first
century, and both are suspension bridges.
A more recent structural type is the cable-stayed bridge,
which has recently demonstrated significant advancements.
Yet even the largest of these bridges—the 890-m span Tatara
Bridge in Japan and the 856-m span Normandy Bridge in
France—are not yet comparable in length to suspension
bridge spans. The suspension bridge is best suited for extend-
ing span lengths even farther, and for nearly 100 years it has
established its superiority in long spans, only 200 years since
it emerged in its modern form. Considering that a masonry
arch bridge in ancient Rome has survived for more than
2,000 years, the suspension bridge is a relatively new struc-
tural form.
The Universality of Primitive

Suspension Bridges and
Their Structure
Primitive suspension bridges were probably used by people
even before the arch bridge in ancient Rome mentioned
1
2 history of the modern suspension bridge
above. Even to this day indigenous peoples not exposed to modern civ-
ilization and without writing and metals continue to build unexpectedly
adequate suspension bridges. For example, in 1959 the anthropologist
Edward Weyer described a suspension bridge built by an African Pygmy

tribe, as illustrated in the photo in Fig. 1-1: “Suspension bridges like this,
built entirely of vines,” he wrote, “have earned the Pygmies a reputation as
human spiders. The Pygmies and the Bushmen are the only two groups in
Africa that live without any agriculture or herding” (Weyer 1959).
Weyer reported separately on a suspension bridge built by a tribe living
in primitive conditions in New Guinea. Two decades later, Japanese explor-
ers led by Katsuichi Honda obtained the photograph shown in Fig. 1-2
and the tribal map shown in Fig. 1-3, where a suspension bridge was built
by a tribe in Papua New Guinea (Honda 1976). Their tools consisted of
stone knives and wooden digging sticks but no containers. They were a truly
primitive people whose living conditions are reminiscent of the Stone Age.
As seen in unexpectedly high-quality suspension bridges built by unsophisti-
cated cultures that still reflect a Stone Age existence, it is understood that the
origin of this type of structure is quite old, perhaps even prehistoric.
I once undertook research concerning when the structure of the sus-
pension bridge first appeared in the human historical record (Kawada
1972). The first such mention comes from the beginning of the fifth century
in China. Buddhist priest Faxian Sanzan of East Jin traveled to India seeking
the Holy Sutra, and a description of the journey, published after his return,
contains the following:
Keeping to the range, the party journeyed on in a south-westerly direc-

tion for fifteen days, over a difficult, precipitous, and dangerous road,
the side of the mountain being like a stone wall, ten thousand feet in
height. On nearing the edge, the eye becomes confused; and wishing
to advance, the foot finds no resting place. Below is a river, named the
Indus. The men of former times had cut away rock to make a way down,
and had placed ladders on the side of the rock. There are seven hundred
steps in all; and when these and the ladders have been negotiated, the
river is crossed by a suspension bridge of ropes. The two banks of the
river are somewhat less than eighty passes apart. (Nagasawa 1971)
The route that Sanzan traveled is considered similar to the route

that British adventurer Sir Francis Younghusband tried to explore in the
late nineteenth century. Therefore, these two people crossed a similar type
of suspension bridge at the same location although they lived at differ-
ent times. According to Younghusband, “The bridge consisted of three
Source: Weyer (1959).

Figure 1-1. Suspension bridge of a Pygmy tribe in Africa.
the emergence of modern suspension bridges
3
Figure 1-2. A suspension bridge built by primitive people in New Guinea.

Source: Honda (1976), with permission from Suzusawa Publishing Co., Ltd.
thick ropes made by the weaving of twisting twigs from birch trees. When
crossing, one has to step on one rope and hold two ropes on each side”
(Younghusband 1896).
Along with this explanation, Younghusband furnished a photograph
of a suspension bridge built in Ascori Village, China, shown in Fig. 1-4.
Since this photo unfortunately provides only a long-distance view, Fig. 1-5
is also provided, which shows a bridge situated upstream along the Indus
River in western Tibet. It should undoubtedly be considered similar to the
type of suspension bridge depicted in Fig. 1-4.
the emergence of modern suspension bridges 5
Figure 1-3. Uginba Village and a suspension bridge.

Source: Honda (1976), with permission from Suzusawa Publishing Co., Ltd.
We are surprised to observe the similarities between ancient bridge

structures and those built by such disparate recent-day indigenous peoples
as a Pygmy tribe in Africa and a Papua tribe in New Guinea. The main cable
where people walk also forms the bridge deck, and the two or more cables
that hang at both sides as the handrails form a V-section. This is a common
configuration of these suspension bridges. The shortcoming of the V-section
Figure 1-4. Suspension bridge near Askori Village, China.

Source: Younghusband (1896).
Figure 1-5. Primitive suspension bridge upstream along the Indus River in western
Tibet.
Source: Steinman (1918).
suspension bridge is that the loading point greatly deflects and pulls the
upper two cables toward one another. For this reason, it is said in New
Guinea that only one person at a time can cross such suspension bridges.
In Tibet, helpers in front and back push the upper cables apart with their
back and feet when crossing with heavy loads (Iwahori 1956). It is natural
to assume that this cross section would have been transformed from a V- to
a U-section by adding multiple cables to the deck.
Primitive suspension bridges with a U-section have existed in various
places in the world. For example, one such span in Peru was reported in the
mid-1800s by early explorers (Prescott 1847). Since the emergence of modern
suspension bridges, however, it appears that these spans are no longer of much
use today, although some of them are rebuilt every two years or so for the pur-
pose of transferring engineering and cultural knowledge. Several examples of
this type of suspension bridge were built in Japan. There are not many such
suspension bridges remaining today because the natural vines used for the
cables are easily damaged. However, one exception is the Kazura Bridge (Fig.
1-6) in Tokushima Prefecture, which exists to this day thanks to its rebuilding
every two years (Tokushima Prefecture Education Board 1955).
Figure 1-6. The Kazura Bridge in Iya Village, Japan.

Kazura is a type of vine.
Source: Courtesy of Nishi-Iya Village Tourist Association.
Figures 1-7 and 1-8 are invaluable images taken in the Andes Moun-
tains in the 1980s by a Japanese photographer, Jun Takano; they tell us how
this type of bridge is erected (Takano 1988).
Suspension Bridges Were Unknown to

Western Civilization until the Medieval Period
Although primitive suspension bridges are found in various places through-
out the world, from Africa to New Guinea, Japan, and South America, it is
Figure 1-7. Villagers gather with ropes to replace a suspension bridge in Peru.
Source: Takano (1988), with permission from Yama-Kei Publishers Co., Ltd.
Figure 1-8. A villager adds ropes to the sides of a Peruvian suspension bridge.
Source: Takano (1988), with permission from Yama-Kei Publishers Co., Ltd.
thought that “rope dancing” or “rope crossing” spans consisting of just a

single cable predate even these early spans. This is evident from reports of
surprising discoveries by explorers in the hinterlands of China and South
America. For example, academic explorer Jiro Kawakita reported the fol-
lowing in northwest Nepal in 1955:
Incidentally, when we arrived at the village there was something that

surprised us. It was that there was no bridge to cross the Haruken
[meaning “the other side”]. . . . Instead of a bridge, there was a thick
rope stretched from the Turuken [meaning “this side”] to the Haruken.
It was made from cutting yak’s skin into strips and weaving it together
to make a thick rope. It was larger than the size of a grip, and deflected
under its own weight.

One week after arriving at the village, I visited the Haruken. . . .
There was no other means to cross other than rope dancing. There is a
sliding timber called a Garbo. It looked like a half section of pipe, the
inside of the hardwood was hollowed, and its thickness was one foot.
By placing it over the thick rope, sliding became easy. A thin rope could
go through the outside of the Garbo. The thin rope was made of yak
fur, was strongly hand-woven, and is long enough to wrap a body two
or three times. By using this, the weight of one’s body is supported from
the hanging rope.
The procedures of crossing are like this. First, put the Garbo over
the thick rope. Wrap the thin rope attached to the Garbo around your
body several times and securely fasten. Hold the upper portion of the
Garbo by your hands and hang from the thick rope. Then, with the help
of the weight of your body, you can automatically slide down to about
the midpoint of the river, where the rope deflects most. From there, turn
over on your back, hold the rope with both hands, and twist around
using your feet. With your arms’ strength and the momentum developed
by swinging your body right and left as both legs twist upward, your
body is pulled up. (Kawakita 1960)
From the photo in Fig. 1-9 showing rope dancing, it is also obvious that
a great effort is required to cross the remaining portion of the span after
reaching the lowest point deflected by the person’s weight. Pulling from the
other side by another rope was subsequently invented, and this marked the
emergence of “rope crossings,” shown in one of its simplest forms in Fig.
1-10 (Heim 1967).
Rope crossings are found worldwide in mountainous areas, not only
in Kashmir and Nepal but also in China, Japan, and South America. From
Fig. 1-11 depicting the tarabita type of rope crossing in South America
around 1750, it is known that large animals such as horses also crossed
these spans (Juan and Ulloa 1750).
The Japanese “basket crossing” is similar to the rope crossing. Figure
1-12 features the “Basket Crossing in Hida,” a ukiyo-e painting by Hiro-
shige Ando in Pictures of Scenic Beauty in Sixty Four Provinces. A similar
type was used until roughly the Meiji period (1868–1912) at Gokayama in
Toyama Prefecture in Japan (Ishibashi 1972). It still remains as an example
of a yaen (a Japanese term, “monkey run”) over the Uwayu River in the
Figure 1-9. Rope dancing.

Source: Kawakita (1960).
Figure 1-10. Rope crossing.

Source: Heim (1967), with permission from Akane Shobo.
12
Source: Juan and Ulloa (1750).

Figure 1-11. Rope crossing in South America.
history of the modern suspension bridge
Figure 1-12. Basket crossing as seen in a ukiyo-e painting.

Source: Narasaki (1971), with permission from Shin’ichi Ikenaga.
Totsu Gorge, Nara Prefecture (Ueda 1984). (It is said that the term “monkey
run” originates from the appearance of a person on a basket moving by tug-
ging on a rope, thus resembling a monkey running along a vine.)
As we have seen, primitive suspension bridges began first with a mono-
cable in rope dancing, rope crossing, and basket crossing spans. They then
evolved to a V-section consisting of three cables, and then to a U-section
with multiple cables. It is questionable as to whether two-cable spans ever
existed, although historian Joseph Gies claimed that a two-cable suspension
bridge existed in which a walking deck was fixed on two tightly stretched
cables, and that a three-cable suspension bridge was subsequently developed
(Gies 1963).
I once accepted the explanation shown in Fig. 1-13, but I have come
to question this view. I originally believed that it was natural for suspen-
sion bridges to evolve from one, to two, and then to three cables because
of the difficulty in procuring material for the main cable, which forms the
essential part of the structure. However, I am now inclined to think that the
number of cables increased from one to three for reasons of constructabil-
ity and safety.
This perspective is also suggested from photographs of bridge replace-
ment work in Peru (Figs. 1-7 and 1-8). And, by way of careful observation,
a suspension bridge in Guatemala (Fig. 1-14) cited by Gies as an example
of a two-cable suspension bridge appears to be relatively new since it looks
as if it uses wire rope. Therefore, I am now tempted to believe that the
two-cable suspension bridge emerged as a simplified form of the U-section
suspension bridge.
Figure 1-15 shows “Suspension Bridge at the Border of Hietsu,”
painted by Hokusai Katsushika, which is often cited as a typical exam-
Figure 1-13. Evolution of primitive suspension bridges.

Source: Kawada (1981).
Figure 1-14. An old-type suspension bridge.

Source: Gies (1963).
Figure 1-15. “Suspension Bridge at the Border of Hietsu,” painted by Hokusai

Katsushika.
Source: Narasaki (1971), with permission from Shin’ichi Ikenaga.
ple of primitive suspension bridges in Japan. It is a two-cable suspension

bridge. However, the structure is technically problematical and it is dif-
ficult to believe that it existed as shown. Hokusai probably never visited
the bridge, and drew it by developing a concept only from second-hand

knowledge.
In any case, whether they included one, two, or multiple cables as
previously described, primitive suspension bridges were distributed widely
over continents from central Asia, Africa, and China to South America,
and in islands such as Japan and New Guinea. It is also believed that they
were universally used to cross ravines in mountainous areas before recorded
human history and before civilization.
It is curious that no evidence of such a primitive suspension bridge
existed in Western civilization. For some reason, the suspension bridge type
of structure did not exist in the Western world until the Renaissance. In
fact, only relatively recently did the Western world learn about suspension
bridges, probably after the beginning of the sixteenth century. The first
recorded evidence is a rope suspension bridge built over the Padus River
by the military alliance of Germany and Switzerland in 1515. This bridge
was built to transport cannons across the river. Then in 1569, French
Admiral Coligny built a rope suspension bridge over the Clain River for
capturing the city of Poitiers, France (Noirfontaine 1832). These suspen-
sion bridges served as emergency crossings for military use, and iron was
not yet used at that time.
However, the book Machinae Novae (“New Machines”), thought
to be published around 1595 to 1617 by the Italian engineer Verantius,
presented the Pons Unius Funis (“Unique Rope Bridge”), a basket crossing
shown in Fig. 1-16, and the Pons Ferreus (“Iron Bridge”), a span of eyebar
chains shown in Fig. 1-17 (Verantius ca. 1599–1617).
The suspension bridge was still a rare structural form in European civ-
ilization during the Renaissance, and it was indeed a “new machine.” The
first bridge to use an iron chain was built across the Oder River near Glo-
rywitz in Prussia in 1734 (Bender 1868). Later, in 1741, the Winch Bridge
(Fig. 1-18) was built in Britain (Schaechterle and Leonhardt 1940). William
Hutchison, in The History of Antiquities of the County Platine of Durham,
described it thus:
About two miles [3.2 km] above Middleton, where the river Tees falls
in repeated cascades, a bridge, suspended on iron chains, is stretched
from rock to rock, over a chasm near 60 ft [18 m] deep, for the passage
of travelers, but particularly for miners: the bridge is 70 ft [21 m] in
length, and a little more than 2 ft [60 cm] broad, with a hand rail on
one side, and planked in such a manner that the traveler experiences all
Figure 1-16. Basket crossing, as shown by Verantius.

Source: Verantius (ca. 1595–1617).
Figure 1-17. A suspension bridge, as shown by Verantius.

Source: Verantius (ca. 1595–1617).
Figure 1-18. The Winch Bridge in Britain (1741).
the tremulous motion of the chain, and sees himself suspended over a
roaring gulph on an agitated, restless gangway to which few strangers
dare trust themselves. (Hutchison 1794)
As stated above, the structure was nothing more than a primitive sus-
pension bridge, except that the cables were made of iron. Even so, it was
used for 60 years until 1802 (Tyrrell 1911). The superiority of iron as a raw
material in suspension bridges was clearly demonstrated over the use of
plant fibers, which required replacement every few years.
The Birth of the Modern Suspension Bridge

The use of iron chains greatly extended the service life of suspension bridges,
and yet the Winch Bridge was not structurally equivalent to a modern
suspension bridge, primarily because the cable itself constituted the deck.
Whether the cable is made of plant fiber or iron, as long as the cables func-
tion as the deck and people must walk directly on them along their curve, it
is nothing more than a primitive suspension bridge.
Before the suspension bridge was developed to support horse-drawn
carriages and, later, cars in modern traffic, a great leap beyond the primi-
tive suspension bridge was required. First, the deck should not follow the
catenary of the main cables, but instead should form more of a flat roadway.
Second, to avoid excessive deflection from horse-drawn carriages and cars,
a certain degree of stiffness was essential.
The first suspension bridge that conquered these problems inherent

to primitive suspension bridges was built in 1801 in Pennsylvania by James
Finley (1762–1828). Not much is known about his background, but he was
born in 1756 in Ireland and emigrated to America at a young age. In 1770

he was admitted to the University of Pennsylvania (possibly as a prepara-
tory student). He later became a judge in Fayette, Pennsylvania and, at the
same time, became an inventor of suspension bridges and in their early days
a distributor of them as well (Cordier 1820). As shown in Fig. 1-19, the
deck of his Jacob’s Creek Bridge, the world’s first modern suspension bridge,
could support people and horse-drawn carriages, and the main cables were
separated from the deck. The era of the modern suspension bridge started
with this bridge, which Finley described thus:
In 1801, I erected the first bridge on this construction over Jacob’s Creek
in accordance with a contract between Fayette and Westmoreland coun-
ties . . . 70 feet [21-m] span, 12½ feet [3.8 m] width, warranted for
50 years all but the flooring for . . . $600. . . . I wonder if there exists
another method that is structurally light, robust, long-enduring, easy to
build, and easy to repair and replace members. (Finley 1810)
This description is cited from a letter Finley addressed to the U.S.

Patent Agency in 1810. The content of the letter is interesting and from it
one can easily discern both his confidence and talent. Finley listed five fea-
tures of his invention:
1. Installed anchorages and towers.
2. Separation of the main cables from the deck.
3. Identical angle of the main cables as they diverge from both sides
of the tower tops.
4. Suspended hangers extending from the main cables.
5. Attachment of the deck to the suspenders.
Figure 1-19. Jacob’s Creek Bridge in Pennsylvania, the first modern suspension
bridge.
Source: Finley (1810).
Finley stressed the separation of the deck used by people and horse-
drawn carriages from the main cables that support the loads. He claimed
that the distribution of loads to many hangers would eliminate excessive
deformations in the main cables. Technically, Finley understood the effect

of the stiffening girder in suspension bridges, but offered the following cau-
tion as to the design of the deck structure: “. . . four or more joists will be
necessary for the upper tier to extend from end to end of the bridge. Each
will consist of more than one piece; the pieces had best pass each other side
by side so that the ends may rest on different joists on the lower tier” (Finley
1810). This measure was not alone sufficient, but Finley clearly recognized
the benefits of stiffening in suspension bridges.
By the early 1800s, suspension bridges had been known to Western
civilization for some time but they were considered dangerous, primitive
structures used only by technically unadvanced societies and as emergency
bridges for the military. However, with the suspension bridge’s new emer-
gence as a permanent structure that could carry horse-drawn carriages, Fin-
ley’s name became widely known. With his simple, robust, and economical
method that claimed to be “light, robust, long-enduring, easy to build, and
easy to repair and replace members,” the suspension bridge system became
widely accepted in frontier America and led to a building boom. In 1807, the
Potomac Bridge (38.6-m span with a 4.5-m width) was built in Washington,
D.C., and in 1809 two of the first modern multiple-span suspension bridges
were built—the Schuylkill Falls Bridge (a two-span suspension bridge with
equal 46-m spans) and the Neshammy Creek Bridge (a two-span suspension
bridge with equal 30-m spans) in Bucks County, Pennsylvania.
In 1810, three relatively large bridges were built: the Brandywine
Creek Bridge near Wilmington, Delaware, a 43.5-m single-span suspension
bridge with a 9-m width accommodating a two-lane roadway; the Ken-
tucky River Bridge in Franklin County, Kentucky, a suspension bridge of
two 50.1-m spans; and the Newburyport Bridge built over the Merrimac
River in Essex County, Massachusetts, also known as the Essex-Merrimac
Bridge. The latter was the longest single span—expanding the span length
to 72 m—and it accommodated two-way traffic separated by a median. It
was a truly splendid structure.
The Essex-Merrimac bridge became the most famous of Finley’s sus-
pension bridges thanks to its inclusion in a bridge engineering book written
by Thomas Pope (1811). Public expectations of this bridge type were enor-
mous, as the report in a local newspaper reveals:
The chain bridge, recently thrown over the Merrimac, three miles
[4.8 km] from this town, is now open for the accommodation of travel-
ers. . . . It consists of a single arch of 240 ft [73 m] span and 40 ft [12 m] at

its greatest elevation from the water. The abutments, 49 ⫻ 25 ft [14.7 ⫻
7.5 m] at the base, 45 ⫻ 20 ft [13.5 ⫻ 6.1 m] at the top, and 37 ft [11.3m]
high, are constructed entirely of rough and split stone. . . .

On the abutments at each side of the river, framed piers or uprights
are erected, of a sufficient height, and capped with stout timber to sup-
port the chains, from which the flooring of the bridge is supported. The
chains, ten in number, were passed over the caps of the uprights, and
extended or slackened so as to pass under the two middle cross joists
of the flooring, and describe a sinking or curve, the entire distance of
which, from a right line, is one-seventh part of the span; these are tripled
where they bend over and rest on the caps; are each 516 ft [157.2 m]
in length and 256 ft [78 m] along the curve line between the uprights.
The ends descend from the uprights (with an angle of the same inclina-
tion that the curve makes inwards), to a space below the surface of the
ground, and there secured sufficiently to counterbalance the bridge and
any possible weight that may be brought thereon.
The whole quantity of iron used in chains, securities, etc.,
is 22 tons. The two middle cross joists of the flooring rest upon the
chains, and other joists are attached to them in nearly a horizontal line
by suspenders from the curve. On the joists the string pieces are laid
lengthwise, which receive the plank. The whole flooring admits two
passages, 15 ft [4.6 m] wide, rises 3 ft [0.9 m] in the center, is firmly
connected and strengthened by a stout railing, and has a very light and
agreeable appearance. Horses with carriages may pass upon a full trot
with very little perceptible motion of the bridge. The whole weight of
flooring, chains, suspenders, etc., commonly supported, does not exceed
100 tons, transient weight, the third part of which, it is presumed, can
never be brought upon the bridge at the same time. The total cost is
about $25,000. John Templeman, Esq., of the territory of Columbia,
the contractor. . . .” (Newburyport Herald, 14 December 1827)
Finley’s Glory and Setback

The 1810s proved to be a very successful decade for Finley. He could not
manage to build all these chain suspension bridges by himself, but about
40 bridges were built in nearly 10 years under license using his patent.
The major bridges built in that time period include the Northampton
Bridge in Pennsylvania (about a 30-m span) in 1811 and the Lehigh Chain
Bridge (also known as the Allentown Bridge), also in Pennsylvania, in 1814
(a four-span continuous suspension bridge with a total length of 142.5 m).

This latter span accommodated a two-lane roadway and a 1.8-m-wide
pedestrian way on each side (Niles’ Weekly Register, 4 February 1845).
Other bridges were built during this period but their completion dates are
unknown. They include, for example, the Cumberland Bridge in Maryland
(39-m span), two identically named Brownsville bridges (a 36-m span with
a 5.4-m width and a 39-m span with a 4.5-m width), and others.
Suspension bridges boomed in these early days but the glory did not
last long because of a series of unintended but continuing accidents that
beset Finley’s suspension bridges. The first suspension bridge that Finley
provided with a “50-year guarantee,” the Jacob’s Creek Bridge, lasted only
half that long, and others collapsed in an even shorter time. The first accident
occurred in 1810 when, just 3 years after completion, the Potomac Bridge
was destroyed by flood. The next loss occurred in 1811 when, only 2 years
following completion, the Schuylkill Falls Bridge collapsed under a herd
of cattle crossing the span. A new suspension bridge, rebuilt immediately
after the incident, also collapsed within 5 years, in January 1816, this time
under the weight of accumulated snow. In fact, many suspension bridges
fell under the weight of winter snows during this period. For example, one
of the Brownsville bridges collapsed when a fully loaded horse-drawn car-
riage passed over the snow-covered bridge in the winter of 1820. A typical
example of a similar accident occurred at the Newburyport Bridge in 1827,
as a local newspaper reported:
On Tuesday morning last, the Essex-Merrimac Bridge [a name by which

it was also known] gave way in the center, from the parting of the chains
which supported it. On the bridge at the moment, was a loaded team, drawn
by six oxen and two horses, driven by two men, Messrs. Carleton and Jack-
man, all of whom were precipitated 40 ft [12 m] into the river beneath.
Five of the ten chains, which supported the bridge, were snapped
in different places, and now remain upholding the broken and shattered
timber, altogether as sad a wreck as we ever witnessed. At the instant of
the crash, the light evolved by the friction of the chains resembled the
vivid streaming of a meteor.
Various causes are assigned for the accident, and none, with more
probability, than the united effect of the incumbent pressure of the
immense body of snow lying upon the bridge, and the frost which had
contracted the particles of iron. These produced tenseness in the chains
which was incapable of resisting the additional pressure of the loaded
team, and the whole gave way. (Newburyport Herald, 9 February 1827)
After the accident, it was decided to immediately rebuild the bridge.
The bridge will be built in two entirely distinct parts, without any con-
nection whatever between the two, so that, on any apprehension of

danger from passing with a heavy load, that part may be taken which is
not laden. The bridge will be supported by two or three chains in addi-
tion to the former number. (Newburyport Herald, 27 February 1827)
The photographs in Figs. 1-20 and 1-21 show that suspension bridge
after strengthening. It is interesting to understand the structure of this type
of suspension bridge of that era. This reinforced suspension bridge with-
stood continued use for about 80 years, until 1909 (Miles 1911).
As suggested by the fact that the Newburyport Bridge lasted for
80 years after being strengthened through the addition of only 20% more
cable, the series of accidents afflicting Finley’s suspension bridges never
implied that his suspension bridges suffered from structurally fatal faults.
Rather, collapses were caused by a lack of consideration of snow loading,
thermal stress, problems associated with the raw material used for link chain
cables, and so forth. In other words, these faults were attributable to the
primitive levels of design, materials, construction, and other features, and in
the level of engineering as a whole. Nevertheless, with continuing accidents,
Figure 1-20. Side view of the Newburyport Bridge in Massachusetts.

Source: Miles (1911).
Figure 1-21. Front view of the Newburyport Bridge.

Source: Miles (1911).
Finley’s fame declined and the Lehigh Gap Bridge (48-m main span, 24-m
side spans, and 4.8-m width) built in 1826 would be the last of his modern
suspension bridges (Boyer and Jerry 1937). Finley’s patented suspension
bridge using iron chain cables was never again built in North America.
It is said that Finley died at the age of 72 in 1828. Three years before
his death, the Jacob’s Creek Bridge collapsed when a six-horse drawn car-
riage passed over it, just halfway into the span’s 50-year guarantee.
References
Bender, C. (1868). Historical Sketch of Improvements in Suspension Bridges to the
Present Time, ASCE, New York.
Boyer, W. H., and Jerry, I. A. (1937). “An early American suspension bridge.” Civil
Engineering, 7(5).
Cordier, J. (1820). Histoire de la Navigation Interieure, Vol. 12, Firmin Didot,
Paris.
Finley, J. (1810). “A description of the patent chain bridge.” Portofolio, 3(6).
Gies, J. (1963). Bridges and Men, Doubleday, New York.
Heim, G. (1967). Places of Gods [in Japanese], Kenji Ozaki, trans., Akane Shobo,
Japan.
Honda, K. (1976). Highlander in New Guinea [in Japanese], Suzusawa Publishing,
Japan.
Hutchison, W. (1794). The History of Antiquities of the County Platine of Durham,

F. Jollie, Carlisle.
Ishibashi, N. (1972). A Land of Mystery: Mount Etchu-Goka [in Japanese],
Hokkoku, Japan.
Iwahori, Y., ed. (1956). Karakorum [in Japanese], Bungei-Shunju Co., Japan.
Juan, J., and Ulloa, A. de. (1750). Relacion historia del viaje a la America Meridional,
Vol. 1.
Kawada, T. (1972). “In search of the origin of the suspension bridge.” [in Japanese].
Koukyou, 6, Association of Steel Bridge Construction.
Kawada, T. (1981). Cultural History of Suspension Bridges [in Japanese], Gihodo,
Japan.
Kawakita, J. (1960). A Country of Birds’ Graves: Description of Exploration at
Mysterious Land in Himalaya [in Japanese], Kobunsha, Japan.
Miles, A. (1911). “The old Essex-Merrimac chain suspension bridge at Newburyport,
Mass, and tests of its wrought-iron links after 100 years’ service.” Engineering
News, 66(5).
Narasaki, M. (1971). Famous Bridges in Various Countries [in Japanese], Kodansha
Co., Ltd., Japan.
Nagasawa, K., trans. (1971). Story of Hogen and Journey of Sohun, Heibonsya [in
Japanese], Toyo Bunko, Tokyo, 194.
Noirfontaine, A. B. de. (1832). “Memoir sur les Ponts de Cordages.” Annales des
Ponts et Chaussées, Vol. 4.
Pope, T. (1811). A Treatise on Bridge Architecture, A. Niven, New York.
Prescott, W. H. (1847). History of the Conquest of Peru, Harper & Brothers,
New York.
Schaechterle, K. W., and Leonhardt, F. (1940). “Hängebrücken I.” Die Bautechnick,
33 (August 2).
Steinman, D. B. (1918). “Bridges.” In Encyclopedia Americana, Americana Corp.,
New York and Chicago.
Takano, J. (1988). Mother Earth of Andes [in Japanese], Yamato Keikoku Co.,
Japan.
Tokushima Prefecture Education Board. (1955). Iya and the Kazura Bridge [in Japanese],
Village of Nishi-Iya, Japan.
Tyrrell, H. G. (1911). Bridge Engineering, Tyrrell, Chicago.
Ueda, A. (1984). Bridges and Japanese [in Japanese], Iwanami Sinsho, Japan.
Verantius, F. (ca. 1595–1617). Machinae Novae.
Weyer, E., Jr. (1959). Primitive Peoples Today, Doubleday, New York.
Younghusband, F. (1896). The Heart of the Continent, John Marry, London.
2
Trials and Errors

in Europe
Introduction to the United Kingdom
Although many of James Finley’s suspension bridges
had unfortunate endings in the North America, his unique
concept was introduced by Thomas Pope in A Treatise on
Bridge Architecture (1811) and became widely known in
Europe. Britain was the first European country where the
Finley-type suspension bridge was built. In fact, the pioneer
spirit in that country had already been exhibited by wire
cable bridges such as the Winch Bridge (1741, actually a
primitive—not a modern—suspension bridge), as described
in Chapter 1. The wire cable bridge appeared to be similar
to a modern suspension bridge at first glance, though it
featured a totally different structure. As Fig. 2-1 shows,
the major difference is that the cables were not continu-
ous between the towers. From our present-day knowledge,
it should rather be classified as a cable-stayed bridge, and
was perhaps influenced by Verantius’s seventeenth-century
“New Machine.”
The first of this kind of bridge was the Galashiels
Bridge (33.6-m span) built in 1816, which was followed
by the King’s Meadows Bridge (33.5-m span) and the
Dryburgh Abbey Bridge (79.2-m span), both completed
in 1817. These bridges were pedestrian walkways with
a width of only 1.2 m, and they oscillated easily. In fact,
there were two Dryburgh Abbey bridges, an old and a new.
Construction of the first span began on April 13, 1817, and
was completed 4 months later. But this bridge, like many
27
Figure 2-1. King’s Meadow Bridge.

Source: Stevenson (1821).
early wire cable bridges, was destroyed in a strong wind on January 15,
1818. It did not survive more than half a year.
On the same site, a new bridge was built that was in appearance simi-
lar to the old bridge, but there was a remarkable change in its basic concept.
The new design eschewed the cable-stayed form and introduced a modern
suspension bridge of the Finley type, with chain cables strung between the
tower tops and hangers suspended from them. (Strangely, the concept of
stringing cables between main towers did not yet exist in Great Britain). The
Scottish engineer Robert Stevenson described the improved design:
. . . after a better design . . . and in less than three months it was again
opened to the public. This bridge is now constructed upon the catenar-
ian principle, agreeably to [Fig. 2-2], the roadway being suspended by
perpendicular rods of iron, from main or catenarian chains. . . . the
roadway has also been strengthened by a strongly trussed wooden rail,
which also answers the purpose of a parapet, on each side of the bridge,
the good effects of which were particularly exemplified, while the bridge
was building. (Stevenson 1821)
For the reconstruction of the Dryburgh Abbey Bridge, Captain Samuel

Brown (1779–1852) introduced the design concept of Finley’s modern sus-
pension bridge and opened a new era in Britain. Brown was involved in the
Figure 2-2. Reconstructed Dryburgh Abbey Bridge.

trials and errors in europe 29
manufacture and development of chains for ships during his service in the
navy, and around 1808 invented a bar chain made of round or flat bars with
holes at both ends to supplement defects in the strength of traditional link
chains made from iron rings (Kemp 1972). Figure 2-3 shows the structure
of the flat eyebar chain for which he took out a patent. However, the flat
eyebar chain was not used in the reconstruction of the Dryburgh Abbey
Bridge, and 4-mm-diameter round eyebar chains were adopted. As is appar-
ent from Fig. 2-2, the diagonal cables remained and the bridge retained a
strong transitional appearance.
Brown built the first modern suspension bridge in Europe, the Union
Suspension Bridge, which spans the River Tweed along the border between
Scotland and England. The bridge’s span of 140 m and width of 5.5 m made
Eyebar Link Bolt
Assembled Link
Hanger
Figure 2-3. Brown’s structure of the flat eyebar chain.

Source: Kemp (1972), with permission from Emory L. Kemp.
it the world’s largest suspension bridge when completed in 1820, and its
carrying capacity for loaded horse-drawn carriages changed the prevailing
view that only pedestrians were able to cross suspension bridges (Figs. 2-4
and 2-5). Of course, eyebar chains were used for the cables and various
concepts for connection methods of the vertical hangers can be seen in the
photograph and in Fig. 2-3.
The bridge’s cable sag of 9.15 m produced a sag ratio of about 1⬊15,
which is relatively flat from our present-day perspective. This bridge has
undergone several rehabilitations and reinforcements but it retains essen-
tially its original appearance and is still in use under a 2-tonne load restric-
tion (PCEI 1997, 50).
The modern suspension bridge and Brown’s name, already known to
the British public through his reconstruction of the Dryburgh Abbey Bridge,
became even more popular through the success of the Union Suspension
Bridge. Brown was 45 years old in 1821, retired from the navy and in busi-
ness as a manufacturer of chain cables. He had taken up the vocation of sus-
pension bridge engineer. His first bridge after establishing his independence
was the Trinity Chain Pier, completed in 1821 (Brown 1822). As shown in
Fig. 2-6, Brown designed a 1.2-m-wide, three-span suspension bridge for the
213-m pier, built in Newhaven at the mouth of the Forth River. Subsequently,
in November 1823, Brown completed the even larger Brighton Chain Pier,
with a total length of 346.5 m and a 3.8 m width (Fig. 2-7). Here, he built a
suspension bridge with four spans of 77.8 m and a cable sag of 5.5 m.
The eyebar chain cable developed by Brown gained acceptance in Brit-
ain because it ensured greater strength than linked chain cables. The Menai
Bridge in Wales (177-m span) in 1826, the Széchenyi Bridge in Hungary
(203-m span) in 1849, and the Clifton Bridge in Bristol (214-m span) in
1864 continued the construction of famous bridges that remain to this day.
The Menai Suspension Bridge was the first bridge to exceed the record
length span of the Union Bridge. Built by Thomas Telford over a 7-year
period between 1819 and 1826, this bridge featured a 168-m main span
Figure 2-4. Sketch of Brown’s Union Bridge.

Figure 2-5. Photograph of Brown’s Union Bridge.

Source: Hopkins (1970), with permission from David C. Hopkins.
Figure 2-6. Trinity Chain Pier.

Source: Brown (1822).
(177 m between the tower tops) and a 3.66-m width to accommodate one traf-
fic lane in each direction and a central 1.22-m wide pedestrian walkway. The
deck was suspended 30.5 m above the water and its masonry towers reached
a height of 46.6 m. Three to four 16-m span masonry arches were used in the
side spans. The structure’s 2,187 tonnes of iron and its 520-m length elevated
the bridge to an incomparable scale for its era (Figs. 2-8 and 2-9).
Figure 2-7. Brighton Chain Pier.

“The pier on its last legs.”
Photograph by Thomas Donovan, 1896.
Figure 2-8. Menai Bridge upon completion, viewed from the side.
One interesting aspect of this bridge is that prior to construction, an

experiment was performed to stretch a 180-m-long chain cable and sus-
pend 57 hangers at designated locations over a valley near the construction
site. In this way, cable geometry and cable tension were confirmed, and the
23.5-tonne weight of the chain cable and the required force of 39.5 tonnes
for lifting it to the tower tops were obtained experimentally (Smiles 1904).
These results were subsequently employed in the actual construction. For
example, as shown in Fig. 2-10, a 140-m-long and 1.8-m-wide raft was
used to erect the chain cables, on which the assembled chain cables were
transported to the site and lifted directly to the tower tops.
The Széchenyi Bridge, built in Budapest and popularly called the
Lánchíd Chaine Bridge, was completed by two British engineers, Tierney
Clark for the design and Adam Clark for the construction. It was really a
British suspension bridge; the cable was naturally built from eyebar chains
imported from Britain. Tierney and Adam had been thought to be brothers
(Shirley-Smith 1964, 68; Kawada 1981), but Tierney came from England
and Adam from Scotland; having been born in different places, they had no
blood relationship (Medobedo 1999, 50; PCEI 1997, 337).
Figure 2-9. Menai Bridge upon completion, viewed from one end. Note the trans-
verse struts between the cables.
This suspension bridge consisted of a 203-m main span and 91-m

side spans, with a 14-m width. It was a heavy bridge, with strong stiffen-
ing trusses providing extremely high rigidity. Two days before the span’s
completion, Hungarian troops with 270 horse-drawn cannons crossed the
bridge, and the bridge withstood this unexpected loading test (Civil Engi-
neer and Architect’s Journal 1849). The bridge was eventually destroyed in
the Second World War but was subsequently restored. Its beauty, looming
Figure 2-10. Cable erection plan for the Menai Bridge.

Source: Smiles (1904).
over the waters of the Danube River, still makes the bridge a Budapest land-
mark (Fig. 2-11).
To end the era of British eyebar chain cable suspension bridges, the
Clifton Bridge was completed in 1864 as the largest span of its type. This
project started with an open design competition for a bridge to be erected
over the abyss of Avon River in Clifton, near Bristol, England. The first
competition in 1829 was a failure, with no feasible design proposals. There-
fore, a second competition was initiated toward the end of 1830. This time
the prizewinner was a young Isambard Brunel, only 24 years old, whose
design was selected over those of such famous engineers as Samuel Brown
and Thomas Telford (Body 1976). Upon learning that his first proposal of a
more than 200-m span suspension bridge was not accepted because of anxi-
ety about the wind, Brunel presented a 194-m span bridge for the second
competition, moving one tower base considerably forward, as shown in
Fig. 2-12. Thus, he succeeded in keeping the span length less than 200 m
and finally won the competition (though the tower-to-tower distance of the
bridge remained 214 m, exceeding the limit of 200 m).
The construction of the Clifton Bridge began in 1831 under Brunel’s
supervision, but was affected by the turbulent social and economic condi-
tions of the time and took 33 years to complete. During that period, Brunel
died in 1853 at the age of 53 and thus did not see the bridge through to
completion on December 8, 1864. The Clifton Bridge is well preserved and
Figure 2-11. Széchenyi Bridge in Hungary.

Source: New Civil Engineer, 25 July, 1996, with permission from New Civil Engineer.
Figure 2-12. Proposal for the Clifton Suspension Bridge (Brunel’s final plan).
Source: Watercolor by Samuel Jackson, K4077, with permission from Bristol’s City Museum
& Art Gallery.
remains in service, though with load restrictions. It retains essentially the

same appearance as it did long ago (Figs. 2-13 and 2-14). The tower shape
differed from his proposal, having been changed by Brunel during construc-
tion on account of budget difficulties.
Accidents in the United Kingdom and

Problems Relating to Stiffness
Although British suspension bridges appeared to steadily evolve and intro-
duced the change from link chain to eyebar chain, they could not escape the
accidents that plagued Finley’s suspension bridges in America. For example,
in 1836, Brown’s 1823 Brighton Chain Pier collapsed spectacularly after
about 13 years of use. In a report to the Royal Scottish Society of Arts, J. S.
Russell assessed the calamity:
The same span of the Brighton Chain Pier (the third from the shore),
has now twice given way in a storm. The first time it happened in a
dark night, and the storm was accompanied by much thunder and light-
ning: the general opinion of those who do not inquire into the causes
of such matters was that it was destroyed by lightning; but the persons
Figure 2-13. Current-day view of the Clifton Suspension Bridge, side view.
Source: Courtesy of Hiroshi Isohata.
employed about the pier, and whose business it was to repair it, were
satisfied that the first fracture was neither caused by lightning nor by the
waters, but by the wind.
The fracture this year was similar to the former, and the cause
evidently the same. This time, it gave way half an hour after midday,
on the 30th of November 1836, and a great number of people were
therefore enabled to see it.
The upper one of the two sketches [Fig. 2-15] annexed, shows the
greatest degree of undulation it arrived at before the roadway broke;
and the under one shows its state after it broke; but the great chains
from which the road is suspended remained entire.
. . . but on the 29th of November, 1836, the wind had almost the
same violence as a tropical hurricane, since it unroofed houses and threw
down trees. To those who were at Brighton at the time, the effect of such
a storm on the chain pier was a matter of interest and great curiosity. For
a considerable time, the undulations of all the spans seemed nearly equal.
The gale became a storm about eleven o’clock in the forenoon, and by
noon it blew very hard. Up to this period many persons from curiosity
38
Source: Author.
Figure 2-14. View of the Clifton Suspension Bridge from one end.
Figure 2-15. Failure of the Brighton Chain Pier.

Source: Russell (1839), with permission from Royal Scottish Society of Arts.
went across the first span, and a few were seen at the further end; but
soon after midday the lateral oscillations of the third span increased to a
degree to make it doubtful whether the work could withstand the storm;
and soon afterwards the oscillating motion across the roadway, seemed
to the eye to be lost in the undulating one. . . .
At last the railing on the east side was seen to be breaking away,
falling into the sea . . . (Russell 1839)
It was not clear whether torsional vibration was observed, but it reminds
us of the disaster of the Tacoma Narrows Bridge in Washington State more
than a century later.
Some suspension bridges of that period failed from causes other than
wind. For example, the Montrose Suspension Bridge, built by Brown to
span the South Esk River in Scotland, was opened in December, 1829 but
collapsed only 3 months later. The reason for the disaster is related to the
fact that this suspension bridge provided a spectacular spot for people to
observe boat racing. On that day, a crowd of some 700 gathered on the
bridge and when boats passed underneath, people moved simultaneously
from one side to the other. The cables snapped under the eccentric loading
and the impact of the sudden crowd movement (Rendel 1841).
It is said that the worst error in the suspension bridges built by Brown
was the Tees Railway Suspension Bridge in North Yorkshire, England. This
86-m span suspension bridge was originally designed for roadway use, but
suddenly the need for the railway became so great that a reckless trial run
was made. E. W. Serrell wrote in the American Railway Journal:
The only bridge known to the public, built upon the suspension principle,
that has had locomotive trains upon it, was on the line of the Stockton and
Darlington Ry. in England. This bridge had been built for common road
travel, but being in a position to be used by the railway company, the track
was laid on it, and a trial made. Mr. Stephenson stated that the platform
rose up 3 ft before the locomotive at ordinary speed; and that the entire
work was nearly destroyed by the passage of the train. (Serrell 1853)
With respect to loads, horse-drawn carriages and railway trains are com-
pletely different and this trial run was criticized as an “inappropriate use of
a suspension bridge.” It should be said that the need for stiffness in suspen-
sion bridges was not yet fully recognized at that time (Syngleman 1832).
Neither Brown nor Telford had sufficient understanding of the neces-
sary stiffness of suspension bridges. This can be understood by studying a
photograph of the Menai Suspension Bridge following its completion (Fig.
2-9). Its lattice structures on both sides of the deck were too flexible to act as
a stiffening truss. As such, Telford’s bridge was nothing but an unstiffened
suspension bridge. The deck at that time was made of timber. The Menai
Suspension Bridge had no storm cables and, though it was designed to be
relatively heavy with a weight of 435 kg/m2, it could not escape becoming a
victim of wind. W. A. Provis reported on the incidents of major wind-induced
damage to this bridge:
December 12, 1825: The suspension bridge frequently oscillated when

it was nearly completed. The cables of the center span moved separately,
and the transverse struts connecting them as a countermeasure proved
to be effective [refer to Fig. 2-9].
February 6, 1826: One week after the bridge’s opening on Janu-
ary 30, the suspension bridge underwent violent oscillation during a
heavy gale. As a result, the wooden transverse beams of the deck failed
and some hangers were broken. As an immediate repair, thick oak wood
plates were installed beneath the deck between the transverse beams
and firmly attached with iron stirrups. These were very effective and no
damage occurred for almost 10 years.
January 23, 1836: A severe gale caused undulations in the suspen-
sion bridge which at the span’s quarter point reached a double ampli-
tude of 4.9 m. Wearing of the oak planking under repetitive loading for
a period of 10 years as well as shrinkage attributable to drying was felt
to have reduced the span’s stiffness. A partial repair was made, but no
strengthening was undertaken at this time.
January 6 to 7, 1839: The weakened suspension bridge swayed
violently during which torsional vibration was induced. About one-
third of the total number of 444 hangers were damaged, and only the
pedestrian walkway located along the center of the roadway survived.

The roadways along both sides of the span were heavily damaged, and
in some places sagged for a distance of about 50 m and were blown
about by wind. The transverse struts connecting the cables failed but
fortunately the cables were not damaged. (Provis 1842)
As described above, the Menai Bridge was struck by strong winds sev-
eral times and was repaired each time. The timber deck was replaced with a
steel structure in 1892. During 1938 and 1939 rehabilitations, the wrought
iron chains were replaced with steel chains and the center two cables were
removed, leaving two rows along both sides of the bridge. A steel stiffening
truss was also installed at this time (ENR 1939). Having undergone several
major rehabilitations, the Menai Bridge continues to be used, as shown in
Figs. 2-16 and 2-17, but with load restrictions.
In fact, the design competitions for the Clifton Suspension Bridge held
in 1829 and 1830 occurred after the Menai Bridge suffered vibrations and
damage from wind. Telford, the judge in the first competition, therefore
rejected the suspension bridge plans exceeding 200 m in span submitted by
Brunel and others as reckless attempts from those with little regard for wind.
In contrast to Telford, Robert Stephenson (1803–1859) showed great
interest in the stiffness of suspension bridges and was more cautious. He
was the only son of George Stephenson (1781–1848), an inventor of the
railroad, and from birth was also destined to be a railroad engineer. Telford
had built the Menai Suspension Bridge for road use over the Menai Straits
Figure 2-16. View of the Menai Bridge from the side.

Source: Courtesy of Mikio Matsui.
42
Source: Courtesy of Tadaki Kawada.

Figure 2-17. View of the Menai Bridge from one end.
along the line of a major trunk route connecting England and Ireland.
Robert Stephenson, on the other hand, built a railroad bridge across the
Menai Straits near Telford’s bridge.
With the success of the Menai Bridge as a road bridge, a suspension

bridge was at first considered for the Britannia railroad bridge. However,
due to the lack of stiffness in traditional suspension bridges and because
deflections caused by running trains were excessive, an effort was made to
supplement stiffness and limit deflections to an allowable level by adopt-
ing a box section for the steel girder. Based on the suspension bridge plan,
the towers to support the main cables were actually built as shown in Fig.
2-18. At this time (1848) the suspension bridge was obviously intended for
construction (Beckett 1969). However, when determining the dimensions of
the stiffening girder, the ever-prudent Stephenson requested the assistance of
two authorities in British engineering, William Fairbairn (1789–1879) and
Eaton Hodgkinson (1789–1861). After repeated experiments and theoreti-
cal studies, the box girder became so large that it was concluded that cables
were not necessary (Timoshenko 1953).
As seen in Fig. 2-19, a huge box with about a 9-m depth and a 4-m
interior width was designed to span the distance of 146 m. As a result, it was
decided that trains would cross the bridge inside the box. Since completion
of this bridge in 1850, it has been known as the Britannia Tubular Bridge
Figure 2-18. Britannia Bridge during construction in 1848.

Source: Beckett (1969), with permission from Octopus Publishing Group.
Figure 2-19. Cross section of the Britannia Bridge.

Source: Timoshenko (1953), with permission from Dover Publications.
(Dempsey 1864). It goes without saying that a tube is a form of pipe, but
it also describes the subway lines in London. The term “tubular bridge” is
thus somewhat of an oxymoron because one thinks of a train crossing the
bridge when it is actually entering a tunnel.1 In any event, it may be because
of the severe loads imposed by running trains that British engineers aban-
doned the concept of truss-stiffened suspension bridges.
In the period following the introduction of Finley’s concept into Great
Britain, several suspension bridges were built, all of them unstiffened. After
being dismissed in the design of the Britannia Bridge, the concept of the stiff-
ened suspension bridge so common today was not to take root in Britain.2
Figure 2-20 shows the Britannia Bridge as built, without the suspension
cables. This bridge was significant in its introducing a new structural form:
the box girder. At the same time, it was a somewhat ironic monument that
demonstrated the limitations of British engineers of the day in their inability

to open the door to the modern stiffened suspension bridge. Nonetheless,
they came close to the concept.
In fact, the enthusiasm for suspension bridges rapidly died down in
the United Kingdom around this time because of a continuing spate of acci-
dents. The 1864 completion of the Clifton Suspension Bridge, delayed more
than 30 years following the initiation of construction, marked the end of the
era of the British-style eyebar chain suspension bridge initiated by Samuel
Brown. Since then, this form of suspension bridge has never reappeared in
Great Britain.
Figure 2-20. Britannia Tubular Bridge.

This photo shows the bridge in its original form. Only the masonry towers of the original
bridge remained when the bridge was reinforced by its transformation to a steel arch bridge
after a fire in the 1970s (PCEI 1997).
Source: Beckett (1969), with permission from Octopus Publishing Group.
Introduction to France
Knowledge of the modern suspension bridge concept born in North America
was first transferred to Britain, and then to the Old Continent—Europe—and

in particular to France, where it was further developed. Having suffered the
major disturbances of the French Revolution that began in 1789 and then
the Napoleonic Wars (1799–1815), France’s progress in accepting the modern
suspension bridge was delayed and lagged behind that of Britain. As a result,
the nation as a whole was eager to embrace suspension bridges, and the gov-
ernment led the introduction of this new technology. During this period, Brit-
ain was in the midst of the Industrial Revolution. With respect to suspension
bridges, Brown obtained a patent for the eyebar chain, Telford began construc-
tion of the Menai Suspension Bridge in 1818, and the Union Bridge was com-
pleted in 1820 as a genuine road bridge able to carry horse-drawn carriages.
Stimulated by suspension bridge developments in neighboring Great
Britain, the French government in 1821 and 1823 dispatched an energetic
scholar, Louis Henri Navier (1785–1836), to study them. Navier, who
taught structural mechanics at the École des Ponts et Chaussées, compiled
his findings in the Investigation and Research Report on Suspension Bridges,
published in 1823 (Navier 1823). Figure 2-21 shows the front page of his
official report, which translates to “Report Submitted to Mr. Bequey, Direc-
tor General of the French Ministry of Bridges, Roads, and Mines and Inves-
tigation and Research Paper on Suspension Bridges.”
Dear Director General, Your Excellency assigned me to investigate sus-

pension bridges by visiting Britain and report the merits and shortcom-
ings of this new technology. I have put my heart and soul to answer to
your directions dated August 13, 1821.
During this period, I visited Britain first from September to Novem-
ber in 1821, and followed by from March to April in 1823, and visited
most of the major chain cable suspension bridges by visiting the sites, and
made an effort to solve the problems of this kind of structure. This is sin-
cerely to report on what I have obtained and results of investigation and
research by attached investigation and research report. . . . (Navier 1823)
Navier’s report contained the following three chapters, and the report
was delivered to major departments within the French government to help
deepen knowledge of suspension bridges:
Chapter I: History of Suspension Bridges and Results of Investi-

gation in Britain
Figure 2-21. Title page of Navier’s report on British suspension bridges.

Source: Navier (1823).
Chapter II: Theoretical Study of Suspension Bridges

Chapter III: Design Examples (a road bridge and an aqueduct
bridge designed by Navier)
This rising interest in suspension bridges was manifested in plans for

the Pont des Invalides in Paris, as shown in a design example in Chapter III
of Navier’s report. It featured a 155-m span and an 8.8-m width, as shown
in Fig. 2-22. Its span was about 20 m shorter than the Menai Bridge, then
under construction, but it was about 20 m longer than the Union Bridge and
Figure 2-22. Pont des Invalides, Paris.

Source: Navier (1823).
would be the world’s second-largest. For the cable, a genuine forged chain
cable was adopted similar to the British style. Construction of this suspen-
sion bridge began under Navier’s supervision in July, 1824, but regrettably
proved to be a “phantom bridge” that was never completed.
The construction of the bridge was postponed when the anchorage was
observed to have undergone a small amount of sliding. This sliding
occurred in the night of September 6 and 7, 1826 and was caused by
the fracture of a 32-cm diameter water pipe on the Champs Élysées side
of the riverbank in Paris. . . . The construction of the Pont des Invalides
was not resumed. Instead, the already-completed portion of the bridge
was demolished and this great and beautiful structure disappeared with-
out a trace. (de Brony 1837)
The accident was caused by the loss of bearing capacity from flood-
ing and a lack of earth pressure at the front of the anchorage. In any event,
following the failure of the Pont des Invalides, Navier never again built a
suspension bridge.
Instead, French engineer Marc Seguin and his brother Camille began
to promote suspension bridges in France. It was largely through their efforts
that the wire cable was adopted, replacing the British-style eyebar chain cable,
thus generating a major new development in modern suspension bridges.

An opportunity for the Seguins appeared in the December 8, 1821, issue of
Le Moniteur, which contained an article on American suspension bridges
(Cordier 1821). Marc Seguin, who distrusted iron manufacturing technol-

ogy, immediately thought of adopting iron wire:
The adoption of wire is easier and also safer. The mechanical testing
needed for a chain cable is no longer required in a wire cable, because a
defect in material or manufacturing in a single member of chain would
jeopardize the cable as a whole, whereas wire is twice as strong as bar
and in addition the required testing is completed during its cold forming
process. And when a bundle of wires is spanned between towers and
because the binding does not allow the slippage of wire, even the break-
age of a single wire results in a loss of just one wire in the strength of
the cable. (Seguin 1824)
Subsequently, in 1822 Marc and Camille Seguin began research exper-

iments on wire. For example, it is said they invented the technique of dip-
ping wire in boiled linseed oil to provide effective corrosion protection. At
the same time, they observed that wire processed in this way underwent
a slight reduction in strength. In 1823, Marc Sequin built an 18-m-span,
50-cm-wide model of a suspension bridge on his property. On January 26,
1824, after compiling the research results, the Seguin brothers submitted a
paper to Director General Becquey of the Ministry of Bridges, Roads, and
Mines (Seguin 1824). On the inside cover page of the paper was the detailed
title “Desciption Metrée et Estimation d’Un Pont en Fil de Fer, Construit sur
la Rivière [Calculation and Estimate of Wire Bridge Built over the Galore
River in Saint-Vallier].” Featured were a design and estimates for a 30-m
span of 1.65-m width for the Pont de Saint-Vallier, as shown in Fig. 2-23.
The response from Becquey to Marc Seguin’s overture was immediate.
He distributed a copy of Seguin’s paper to prefectural governors, appending
the following letter dated February 18, 1824:
I have also thought of the advantages of using wire in suspension

bridges. . . . It is no doubt that the bar chain was successful in extend-
ing span length and supporting heavy loads. We have no worry in this
British-developed method in construction and calculation as Professor
Navier proved. However, the wire cable to be attempted by the Seguin
brothers for use in construction over the Rhône River will clearly reduce
the construction period, ease the construction, and reduce costs. More-
over, the stiffness of the suspension bridge is not reduced as a result.
Figure 2-23. Pont de Saint-Vallier, in France.

Source: Seguin (1824).
I would like to ask Sir Governors to order your civil staff mem-
bers to investigate the river conditions and disseminate how suspension
bridges would be advantageous by comparing them with the traditional
bridge types. (Reverdy 1982)
The first full-scale bridge to implement Seguin’s concept was the Pont
Saint-Antoine, completed in Geneva, Switzerland on August 1, 1823. This
suspension bridge spanned two moats of a castle with 33.5-m and 23-m
spans, and a width of 1.85 m. Guillaume-Henri Dufour (1787–1875), a
technical officer in the Swiss Army, built the bridge with assistance from
Marc Seguin. Dufour had been educated at the École des Ponts et Chaussées
and had joined the French Army in the Napoleonic Wars. Although the
Pont de Saint-Antoine was built in Switzerland, in reality it can be said that
it is a genuine French-style suspension bridge (Peters 1987). It was the first
modern suspension bridge built for actual use in Europe. (Prior to that,
Seguin had built two experimental bridges, but they were not actually used
and were soon demolished.)
It was the Seguin brothers in Annonay who reminded me of adopting a

wire suspension bridge . . . the Seguin’s elder brother was kind enough
to come over to Geneva for this new technology and visited the site with
me. We laid the first foundation stone for the construction together.
(Dufour 1824)
On May 28, 1824, Marc Seguin began construction of his first full-
scale suspension bridge, the Pont de Tournon over the Rhône River. As
Figure 2-24. Marc Seguin’s design for the Pont de Tournon.

Source: Seguin (1824).
shown in Fig. 2-24, this bridge featured a center tower with two 85-m
spans and a width of 4 m. Each side of the bridge was supported from a
cable strand consisting of 112 wires of 3-mm diameter. Suspended vertical
hangers supported transverse beams of oak, each comprising a 5-m long,
60-cm ⫻ 30-cm rectangular section. This bridge was completed in 1 year
and 3 months, but the span was later raised (in 1842) by government edict
because it hindered the navigation of steam ships that began to increase in
numbers on the Rhône. The original cables remained in use for 140 years
until the bridge was demolished in 1965. Figure 2-25 shows the Pont de
Tournon after having been raised; it is interesting to note the traces of struc-
tural innovations found in the connection details between the main cable
and vertical hangers, and in the main cable connection to the central tower.
Figure 2-25. Pont de Tournon after its height was raised in 1842 to accommodate
increasing river traffic.
The Pont de Tournon was a significant achievement, the first bridge

built in more than 500 years over the Rhône River along the 200-km dis-
tance between Lyon and Orange (the previous one was the Pont Saint Esprit,
a span of 25 masonry arches built in 1307). The suspension bridge caused

a sensation. It leapt across the river in a span that could never have been
conceived for a masonry arch bridge, and, at one-third to one-fifth of the
cost, the span reflected significant economy.
When the Pont de Tournon’s success became certain, the Minister
of Commerce and Public Works, Comte d’Argout, delivered an economic
stimulative directive dated August 24, 1831, to prefectural governors: “In
order to provide jobs for workers . . . the civil engineers of departments
are requested to investigate ferry locations over the rivers and study the
feasibility of bridging” (Marrey 1990, 125). The effect of d’Argout’s notice
was a rapid implementation of suspension bridges in France, built under
the leadership of Director General Becquey. The boom was such that the
Seguin brothers alone built more than 100 bridges. Bridge historian H. J.
Hopkins (1970) depicted the situation effectively, as shown in Fig. 2-26.
Based on A. A. Jakkula’s comprehensive “History of Suspension Bridges in
Bibliographical Form” (1941), this diagram at first glance reveals a peak in
suspension bridge construction in France generally between 1830 and 1850.
Considering that more than 50 bridges with unknown dates of completion
were not included, the diagram suggests that more than 200 suspension
bridges were built in this 20-year period in France alone.
The fact that in the same period about 20 bridges were built in Great
Britain, and not more than 20 bridges in the United States and other Euro-
pean countries, reflects the popularity of suspension bridges in France. The
promoters behind this suspension bridge boom were such influential figures
as Director General Becquey, his technocrats, and the Seguin brothers.
French-Style Suspension Bridges

Initially, France attempted to learn the technology of the modern suspension
bridge from Great Britain, but seems to have changed its attitude following
the failure of Navier’s Pont des Invalides. France later developed its own
methods by adopting wire for cables to create distinctive French-style sus-
pension bridges. Fortunately, knowledge of these early French-style bridges
can be gleaned from an 1831 report submitted to the government by Louis-
Joseph Vicat (1786–1861), a French engineer and the inventor of artificial
cement. In this report, eight suspension bridges were listed as having been
built across the Rhône River. By 1830, only 13 or 14 suspension bridges had
been built in the whole of France, and more than half of these were built
Figure 2-26. Number of suspension bridges completed per country per year.
downstream of Lyon over the Rhône River. The eight suspension bridges are
listed below in order from upstream:
• Vienne Bridge: two spans of 86 m each, two lanes, completed in

1829
• Tain Bridge (also known as the Tournon Bridge): two spans of 85 m
each, one lane with a widening at the center of the span, completed
in 1825
• Vallance Bridge: two spans of 117 m each, two lanes, completed in

1827
• Serrières Bridge: two spans of 101 m each, one lane with a widening
at the center of the span, completed in 1829

• Andance Bridge: two spans of 90.4 m each, one lane, completed in
1827
• St.-Andeol Bridge: three spans of 85 m each, one lane over both side
spans and two lanes over the center span, completed in 1830
• Beaucaire Tarascon Bridge: four spans with two 120-m center spans
and two 93.6-m side spans, two lanes, completed in 1828
• Fourgues Bridge: two spans of 69.7 m each, one lane with a widen-
ing at the center, completed in 1830.
Those suspension bridges built in the 5 years following 1825 embody

differences in span length and in the number of spans, but the arrangement
of wire cables, the method of their anchorage, and the framing of the bridge
deck were virtually identical. Their structure can be discerned from Fig. 2-27,
a diagram adapted from Vicat, and from the photograph in Fig. 2-28. These
eight bridges were all multiple-span suspension bridges with a center tower
situated in the river; no bridge crossed the river with a single central span. The
Figure 2-27. Early French-style suspension bridge elements.

Source: Vicat (1831).
Figure 2-28. Pont de Saint-Andeol in Provence, France.

riverbed was excavated to a depth of 2 m or so at the foundation of the center

tower, and sheet piling was followed by the casting of concrete. Finally, the
foundation had to be protected by stones for scour, as shown in Fig. 2-28.
A distinguishing feature of these eight suspension bridges was, how-
ever, their cables and their anchorage systems. These spans lacked the anchor
blocks characteristic of modern suspension bridges; instead, the cables were
fixed to the tower foundations. Vicat observed:
The composition of the cable in terms of its arrangement was the same
as in any suspension bridge regardless of the number of strands; they
are laid on the same plane over the tower top at a spacing of 15 cm to
20 cm. Between the towers, their sags do not describe exactly the same
curve, but they do form a cable acting as a single unit. At the middle
tower, the strands from both sides crisscross and, at the tower base [as
shown in Fig. 2-27 and the photograph in Fig. 2-28], each hang down
the opposite face of the tower and are anchored to the tower base.
The hangers suspending the deck are attached to different strands

in sequential order. This means that depending on the number of strands
(e.g., strands 3, 5 and 6), hangers at every 3rd, 5th, and 6th points along
the deck are attached to the corresponding strand. . . . (Vicat 1831)
For strands made of iron wires, Vicat explained the procedure this way:
Wire [finished with corrosion protection] was hung over a reel and
wound successively to a flanged drum. This drum can carry about 30 kg
to 40 kg of wire, and when a bundle of wire is finished, then the end of
the wire is connected with the pulling end of the next wire, and wound
up until the drum is filled.
The forming of a strand is very simple. Two timber piles were
driven into the earth for the required strand length, and supports
installed in between at appropriate distances to finish the preparations.
A saddle for the strand [croupière in French, as shown in Fig. 2-29]
is attached to a strong nail driven into the pile. When this preparation
is completed, a drum wound with wire was carried via a hand cart back
and forth between the piles. The drum rotates as the cart moves forward
and the wire is drawn. When it reaches the pile, a waiting worker pulls
the wire with a timber vice to the strand saddle. (Vicat 1831)
The real achievement of Vicat, who greatly influenced the structure

of early French-style suspension bridges and left an interesting legacy of
knowledge in several areas of suspension bridge design, is the invention of
cable spinning. Vicat observed the variation of tension between individual
wires in parallel-wire strands bundled in the Seguin style. The difference in
tension generally does not vary by more than 1.5% in a fabricated parallel-
wire strand bound by a thin, annealed iron wire over a certain distance,
although with longer distances this is not necessarily so. When the strand is
Figure 2-29. Strand saddle.

bent at an angle of 120 degrees over the tower top, and if the strand is fab-
ricated in such a way as to not permit any slippage between wires, then the
difference in tension can be up to 30%. Slack wires were actually observed
in the Pont de Vienne and the Pont d’Andance.

Because the Seguin brothers had noticed differences in tension between
individual wires, they did not allow the manufacture of parallel wire strands
on a flat plane. By setting up two towers as in an actual bridge, wire was
stretched at the predetermined sag to build up a strand. Furthermore, appro-
priate loads were applied at the hanger locations. The strand was then coated
for corrosion protection. Complete strands were later removed from the tem-
porary towers, transported to the site, and erected. Because of the nature of
these procedures, the size of the parallel-wire strand (i.e., the size of the cable)
was determined from the capacity of the erection and handling equipment.
Consequently, the strands were subject to size restrictions.
Vicat questioned this method of manufacturing parallel-wire strands.
Why were separate temporary towers necessary when towers were already
being built for the suspension bridge itself? And if the cable carrier for erec-
tion has the capacity to pull a strand, then there should be no problem to
build up the strand wire-by-wire. From this idea, Vicat invented the concept
of aerial spinning, in which wire is erected directly at the site:
When the piers, towers and anchorages of a suspension bridge are com-
pleted and the anchorages are ready to fix the cables, the preparation
will be complete by winding two set wires of equal lengths into reels.
The wire is then stretched by using puller equipment comprising a loop
system connecting one tower top to the other tower top. This equip-
ment acts as a chain pump and rather faster than a system using people
pulling on the ground. . . . (the number of wires to be pulled could be 2,
4, or 6 . . .). (Vicat 1831)
For erecting the cable, a guide wire is stretched as a gauge in advance.

The tension of each succeeding wire is matched by adjusting its sag to the
guide wire. Even accounting for temperature change, unless the strands
become large enough to produce significant temperature differences the
length is automatically adjusted. And the benefits are not only limited to
construction and economy:
The shortcoming of existing suspension bridges was the separate use

of 10 to 20 strands as a cable. Although the strand arrangements on
the tower top were rather simplified, it lacked balance over the entire
structure. . . .
And, in the traditional method, construction becomes more dif-

ficult as the span becomes longer, requiring larger machines for cable
erection such as winches and cranes. . . . (Vicat 1831)
Such problems were resolved without difficulty by adopting the aerial

spinning method and bundling the wires into a cable. Therefore, Vicat insisted
that:
We have set limits on ourselves for the strand length and sizes so long
as we made use of prefabricated strands, and this means that we missed
the biggest advantage of wire cable. . . . The author’s proposed method
[aerial spinning] has made unnecessary such extra costs otherwise
required, and the same machine is usable for even a 500-m span as in a
50-m span. I have firm confidence on this point, and rather wonder why
it has not been adopted. . . . (Vicat 1831)
Although the aerial spinning method was superior, it took some time
before it became established among engineers, and prefabricated strands
continued to be utilized in France. Only in exceptions where deep valleys
made it difficult to build a middle tower—requiring relatively long spans—
was the aerial spinning method adopted. Examples included the Pont de
Roche-Bernard and the Pont de Tours, both completed in 1840.
The Pont de Roche-Bernard had a span of 198 m. Its designer, Nicolas Le
Blanc, stretched a catwalk between the tower tops, on which wire was manu-
ally pulled via a pulley as shown in Fig. 2-30 (Leclerc 1850). Regrettably, this
suspension bridge, similar in appearance to the Menai Suspension Bridge, had
a short life and collapsed in a strong wind in 1852 (Fig. 2-31). The Pont de
Saint-Christophe, completed in 1847, is believed to have been built using the
aerial spinning method. This bridge extended the span using this method to
183.6 m, and was characterized by the anchorage of the main cables as shown
in Fig. 2-32. The two upstream and downstream cables were connected at the
anchorage as one cable (Leclerc 1850). With few such exceptions, the majority
of suspension bridges built in France at this time employed a middle tower to
shorten the span lengths and were built using prefabricated strand cables.
Figure 2-30. Aerial spinning method by Le Blanc.

Source: Leclerc (1850).
Source: Leclerc (1850).

Figure 2-31. Pont de Roche-Bernard.
Figure 2-32. Pont de Saint-Christophe.

trials and errors in europe
59
One remains curious as to why aerial spinning was not adopted for
the Pont Suspendu de Fribourg in Switzerland, built by M. J. Chaley and
opened in 1834. As shown in Fig. 2-33, a catwalk descended to the riverbed
during construction. It should have been possible to extend this catwalk

to the opposite riverbank and adopt the aerial spinning method. Chaley
was a French-born engineer and, because he mentioned Vicat as a source
in his paper on the bridge, it is difficult to believe that he was not aware of
the aerial spinning method (Chaley 1835). In any case, the Fribourg Bridge
exceeded the span of the Menai Suspension Bridge by nearly 90 m. When
completed in 1834, this 266-m span suspension bridge was the world’s lon-
gest. It served for 90 years and became a Fribourg landmark (Fig. 2-34) until
it was replaced by a concrete arch bridge in 1924.
Early Suspension Bridge Theory

and Its Limitations
The modern suspension bridge was invented by James Finley in the early
nineteenth century and then spread throughout the Western world, but it
was built without the benefit of theory, in both America and Britain. In a
sense, the absence of theory was one cause of repeated accidents in these
early suspension bridges.
The geometry of a stretched cable had long been of interest. As early
as the seventeenth century, Galileo Galilei had speculated its shape to be a
parabola. Robert Hook also studied its shape. At the end of that century,
Philippe de La Hire (1640–1718) wrote a book that dealt with the tension in
cables in comparison to the axial forces acting on arches, and presented the
Figure 2-33. General plan of Pont Suspendu de Fribourg.

Source: Chaley (1835).
Figure 2-34. Pont Suspendu de Fribourg.
force polygon (La Hire 1695). It is said that Jacob Bernoulli (1654–1705)
was the first person to describe cable geometry by way of an exact mathe-
matical equation. Living in Basel, Switzerland as part of the Bernoulli family
of famous mathematicians, Bernoulli studied the following question: What
will be the curve of a uniform string when hung from the same height? He
proved that the curve is a catenary, not a parabola as Galileo had assumed.
Another Swiss, Leonhard Euler (1707–1783), studied under Jacob’s
brother Johann Bernoulli (1667–1748). In his book Methodus Inveniendi
Lineas Curvas [. . .] published in 1744, Euler deduced that a cable under a
uniform weight w distributed along the cable formed a catenary by consid-
ering a small section mn and integrating 冮 wsds, either by the direct method
or by the energy method (Fig. 2-35). Euler had been invited to Russia at
the age of 20 and taught at the Science Academy in St. Petersburg, and he
invited another Swiss pupil of Daniel Bernoulli (Johann’s son), the mathe-
matician Nicolas Fuss (1755–1826), to teach there. Fuss attempted to prove
that a cable carrying a uniform load along the length of a span is shaped as
a parabola. In connection with a plan to build a suspension bridge across
the Neva River, Fuss had studied and clarified the geometry of a cable under
uniform loading—as in a suspension bridge cable—as a parabola.
Such was the development of theoretical studies prior to Thomas Tel-
ford. They all dealt simply with cable geometry; the “theory” of suspension
Figure 2-35. Euler’s sketch of the equilibrium of cable (“methodus inveniendi lineas
curvas . . .”).
bridges had not yet been established. But Telford was unsatisfied with this
trend and he attempted additional theoretical study. In connection with
investigations for the unrealized Runcorn Bridge (300-m main span and
150-m side spans), Telford entrusted the strength testing of eyebar chains
made from malleable iron to mathematician Peter Barlow (1776–1862)
(Smiles 1904). Barlow used the catenary equation to determine cable ten-
sion, and Telford asked British mathematician Davies Gilbert (1767–1839)
to create a table of the tedious catenary calculations to make this informa-
tion easier to use. It was published as Tables for Computing All the Circum-
stances of Strains, Stress of Suspension Bridges in 1829.
Telford himself was rather skeptical of the theory and considered it
more reliable to confirm calculations through experiments with a model. To
this end, Telford build an apparatus sufficient for a 300-m span, as shown in
Fig. 2-36 (Timoshenko 1953). By using this apparatus, Telford acquired the
necessary data for the design and erection of the Menai Suspension Bridge,
as already mentioned. He also found that the elongation of iron members
suddenly increased at about one-half of their breaking (or yield) strength.
From these results, Telford decided to use less than one-third of the breaking
strength for actual service loads.
Henri Navier, who had compiled the studies on early suspension
bridges, is credited as the most influential figure behind their theoretical
advancement. At that time, institutions for civil engineering education only
existed in France, such as the École des Ponts et Chaussées, founded in
Figure 2-36. Telford’s experimental apparatus.

1747. Navier, who had taught at this school since 1819, was the actual
founder of modern structural mechanics. As H. Shutraub (1976) wrote:
A surprisingly great number of methods, still in daily use, of calculating

the dimensions of structures go back to Navier, or have at least been
re-formulated by him. The fact that we are able, today, to construct
safely and economically, is mainly [because of] the methods of struc-
tural analysis, that particular branch of mechanics which is based on the
actual working conditions of a structure. These methods were created,
within little more than a decade, by a single man, Navier.
Navier compiled the results of his several trips to Great Britain and his
investigations, and presented them in his Investigation and Research Report
on Suspension Bridges, as described earlier. In the report’s second chapter, he
performed a detailed study of the theory of suspension bridges. He first con-
sidered the cable geometry of a suspension bridge as a parabola, as it carries
a longitudinally uniform load; he rejected the catenary. As a result, the cable’s
characteristics became clear. The horizontal tension (H) is proportional to
the longitudinal unit loading (w) and the second power of span length (l), but
is inversely proportional to the cable sag (f). It is expressed in this equation:
From this, Navier concluded: “Therefore, it is no use to worry about

increased deflections as the span becomes longer. In contrast, a suspension
bridge becomes less prone to deflection. The reason is that the cable in a
longer span becomes bigger and heavier, generating an increase in longitu-
dinal unit weight.”
Navier studied not only the effects of dead loads on suspension

bridges, but also the deformation and vibration of suspension bridges under
live loads such as vehicles. As a result, by demonstrating that the effects of
impacts and vibration caused by live loads is in relative terms reduced as the
scale of a suspension bridge increases, he somewhat hastily concluded: “In
suspension bridges, as the span increases, the structural difficulty is reduced,
and as the scale gets bigger and bolder, the success becomes more certain”
(Navier 1823).3 In other words, he said that suspension bridges are safe
when their scale increases, but this is quite a misleading and dangerous con-
clusion. However, Navier’s reputation was too great for engineers at the
time to acknowledge his errors. If Navier could have known that the Menai
Bridge would be damaged by wind vibration (1825–1839), and of the inci-
dent in 1831 at Britain’s Broughton Bridge (50-m span), which collapsed
from resonance caused by the marching of soldiers in step (Suttcliff and Sut-
tcliff 1965), it is very likely that his conclusions would have been different.
In any event, it was unfortunate that Navier at that time was utterly
unaware of resonance in suspension bridges and the threat of wind. In a
sense, he completed the early theoretical work in (unstiffened) suspension
bridges and he performed diligent analysis by using the most advanced
mathematics of that time. But for Navier the cable was still the only element
of the suspension bridge worthy of consideration, and he never conceived
the idea of stiffening with girders. This perspective is very obvious when one
considers Navier’s thoughts on thermal change. After studying the effects
of thermal change on cable geometry and tension, Navier proposed the use
of rollers at the tower tops. Then, after pointing out that a similar thermal
effect occurred in the bridge deck, Navier expressed the following view-
point: “When the deck is designed to be continuous for the entire span, the
joints of longitudinal members comprising the deck would be better if not
fixed, because it is desirable to allow a little sliding against each other when
the longitudinal girders expand and contract.”
Again, recall Finley’s advice concerning deck assembly. He claimed that
the deck should be as unified as possible and as rigid as possible. By contrast,
Navier proposed a loose structure and, consequently, a flexible deck.
As long as one did not expect to obtain stiffness from the deck and
railings, it was natural for Navier to conclude that one should supplement
stiffness by increasing cable tension. For this reason Navier suggested a shal-
lower cable sag. Generally, the cable’s sag ratio is expressed by n ⫽ f/l, as
shown in Fig. 2-37. Finley adopted a ratio of about 1/7, whereas Navier
considered an appropriate value to be about one-half of this ratio, that is,
about 1/12 to 1/15. From our present-day knowledge, Finley, who tried to
restrain excessive displacement in suspension bridges with stiffness in the
Figure 2-37. Comparison of Finley’s cable and Navier’s cable (n ⫽ f/l).

Source: Kawada (1981), with permission of Tadaki Kawada.
railings and deck, was far closer to the essence of the suspension bridge,
whereas Navier believed that suspension bridges depended on weight and flat
cables, and that they would become inherently safe as their scale increased.
In fact, the subsequent development of modern suspension bridges proves
that Finley was correct, but there were enough engineers who were dazzled
by Navier’s fame and his theory that sometimes wrong paths were taken.
In contrast to Navier—who introduced modern suspension bridges as
a scholar and studied them theoretically, but who had never built a single
suspension bridge himself—Marc Seguin and his brother promoted the
French-type wire suspension bridge and became leading figures in the field.
Perhaps because they learned Finley’s invention directly from America, they
better understood the role of railings and decks in enhancing the stiffness of
suspension bridges than did Navier (Bender 1868). Marc Seguin observed:
The weakest point of suspension bridges is the vibration induced by

dynamic loading. For its prevention, the only way is to increase the stiff-
ness by any means, and the best method to my knowledge is to make
the railings robust.
It is better to use thick timber for the upper and lower chords
of railings, and assemble the truss from diagonal members crossing at
right angles as in the St. Andrews Cross. And for their connections, use
iron bolts firmly fastened to obtain the necessary strength and stiffness.
(Seguin 1825)
For a better understanding of what Seguin means, refer to the drawings of

truss assemblies in the Fourgues Bridge in Fig. 2-27 and the Saint-Chris-
tophe Bridge in Fig. 2-32.
Although some French suspension bridges were built with atten-

tion to deck and railing stiffness, this alone could not conquer the prob-
lem completely. In fact, the once-booming era of French suspension bridge
building was brought to an end in the 1850s because of a continuing series
of accidents (refer to Fig. 2-26). Suspension bridges at this time were charac-
terized by low stiffness and light weight, and were prone to deflections and
oscillation. A typical accident affecting these suspension bridges involved
resonance set off by marching troops. Posted at the entrance to European
suspension bridges of that era were warnings such as: “Do not walk in step
on the suspension bridge.”
Mathematician and engineer J. V. Poncelet (1788–1867) first explained
the cause of failure of suspension bridges from the phenomenon of reso-
nance. He demonstrated such resonance in a beam when a periodic force
is applied with a frequency close to the natural frequency of the beam, and
verified the reason why a dangerous situation arose when soldiers marched
in step (Timoshenko 1953; Poncelet 1829). The most famous accident
ascribed to this resonance in a suspension bridge was the collapse of the
Basse-Chaîne Bridge in Angers, France, in 1850 (Fig. 2-38). This 102-m
span suspension bridge had been in service since 1839 but failed under
resonance from marching troops, resulting in a tragic 226 casualties (Dupit
and Hougan 1850).
Suspension bridges in this period had not yet become safe structures.
Figure 2-38. Tragic incident at the Pont Suspendu de Basse-Chaîne.

Notes
1. The subway is officially called the Underground, and the operating entity
of subways in London is London Underground Limited (LUL). Therefore,

the “Tube” is similar to the “Métro” subway (abbreviation of Chemin de fer
Métropolitan) in Paris. The London subway began operations in 1863, and
therefore the completion of the Britannia Bridge was 13 years earlier.
2. J. M. Rendel was an exception among Britain’s engineers. Under Telford, Rendel
dealt with the 1830s accident of the Montrose Bridge and succeeded Telford
after his death in 1834. Therefore, he was the principal person undertaking the
repair of the span’s hurricane damage, and obtained an effective result by which
to approach the essence of stiffening in suspension bridges. Regrettably, Rendel
only worked on the Montrose Bridge’s repair and did no other work of his own.
For more detail, refer to Rendel (1841) or Kawada (1981, 135–137).
3. The original is as follows: “La difficulté de ces constructions diminue quand
l’étandue des arches deviant plus considérable, et la succès est d’autant mieux
assuré que l’entreprise est plus grande et semble plus hardie.”
References
Beckett, D. (1969). Bridges: Great Buildings of the World, Paul Hamlyn, London.
Bender, C. (1868). “Historical Sketch of Improvements in Suspension Bridges to the
Present Time.” Trans. ASCE, ASCE, New York.
Body, G. (1976). Clifton Suspension Bridge: An Illustrated History, Moonraker
Press, Bradford-on-Avon, U.K.
Brown, S. (1822). “Description of the Trinity Pier of suspension at Newhaven near
Edinburgh.” Edinburgh Philosophical J., 6.
Chaley, M. J. (1835). “Pont Suspendu de Fribourg.” Annales des Ponts et Chaussées,
Suisse.
Civil Engineer and Architect’s Journal (CEAJ). (1849). “Pesth suspension bridge.”
Civil Engineer and Architect’s Journal, 12(144).
Cordier, J. (1821). “L’histoire de la navigation interieur.” Le Moniteur, Decembre 8.
de Brony, M. (1837). “Notice biographique sur H. Navier, membre de l’Institut Ryale
de France (Acadmie des Sciences). Officier de Légion d’Honneur, Inspectour
des Ponts et Chaussées.” Annales des Ponts et Chaussées.
Dempsey, G. D. (1864). Tubular Bridges, Virtue Brothers, London. Reprinted in
1970 by Redwood Press, San Jose, Calif.
Dufour, G. H. (1824). Description du Pont Suspendu en Fil de fer Construit â
Genève, Chez J. J. Paschoud, Paris.
Dupit, M. de C., and Hougan, R. (1850). “Rapport de la commission d’enquète
nommée par arrête de M. le préfet de Maine-et-Loire, en date du 20 avril
1850, pour rechercher les causes et les circonstances qui ont amené la chute de
pont suspendu de Basse-Chaîne.” Annales des Ponts et Chaussée, III.
Engineering News-Record (ENR). (1939). “New cables, wider roadway for
famous Menai suspension bridge in Wales.” Engineering News-Record,
123(5).
Hayden, M. (1976). The Book of Bridges, Marshall Cavendish, London.
Hopkins, H. J. (1970). A Span of Bridges: An Illustrated History, David &

Charles, Newton Abbot, U.K.
Jakkula, A. A. (1941). “A history of suspension bridges in bibliographical
form.” Bull. Agricult. Mech. College Texas, 12(7). This 10-year work
involved considerable effort to compile a list of all recorded suspension
bridges to that time, including source information.
Kawada, T. (1981). A Cultural History of Suspension Bridges [in Japanese],
Gihodo, Japan, 140.
Kemp, E. L. (1972). “Samuel Brown: Britain’s pioneer bridge builder.” In
History of Technology, Mansell, London.
La Hire, P. de. (1695). Traité de Mecanique : Ou l’on Explique tout ce qui est
Nécessaire dans la Pratique the Arts, & les Propriétés des Corps Pesants
lesquelles ont un plus Grand Usage dans la Physique, De l’Imprimerie
Royale, et se vend chez Jean Anisson, Paris.
Leclerc, M. M. (1850). “Notice sur la construction du Pont Suspendu de
Saint-Christophe (sur le Scorff), près Lorient.” Annales des Ponts et
Chaussées, 20.
Marrey, B. (1990). Les Ponts Modernes en France 18c-19c Siecles, Picard,
Paris.
Medobedo, G. (1999). Story of World’s Bridges [in Japanese], translated by
Naruse Teruo, Sankaido, Japan.
Navier, H. (1823). Rapport et Mémoire sur les Ponts Suspendus, Imprimérie
Royale, Paris.
Publication Committee on European Infrastructure (PCEI). (1997). European
Infrastructure, Japan Society of Civil Engineers, Tokyo.
Peters, T. F. (1987). Transitions in Engineering: Guillaume Henri Dufour and
the Early 19th Century Cable Suspension Bridges, Birkhäuser Verlag,
Basel.
Poncelet, J.-V. (1829). Introduction à la Méchanique Industrielle.
Pope, T. (1811). A Treatise on Bridge Architecture, A. Niven, New York.
Provis, W. A. (1842). “Observations on the effects predicted by wind on the
suspension bridge over Menai Straits, more especially as relates to the
injuries sustained by the roadways during the storm of January, 1839,
together with brief notices of various suggestions for repairing the
structure.” Trans., III, Institution of Civil Engineers, London.
Rendel, J. M. (1841). “Memoir of the Montrose suspension bridge.” Min.,
Proc., 1, Institution of Civil Engineers, London.
Reverdy, G. (1982). “Les ponts suspendus au XIXc siècle.” In Ponts de France,
G. de Grattesat, ed., Collection Tradition Technique, Presse de l’École
Nationale des Ponts et Chaussées.
Russell, J. S. (1839). “On the vibration of suspension bridges and other structures,
and the means of preventing injury from this cause.” Trans., Royal Scottish
Soc. Arts, 1( January 16).
Seguin, M. (1824). Desciption d’Un Pont en Fil de Fer, Chez Bachelier, Paris.
Serrell, E. W. (1853). “The applicability of suspension bridges to railways.” Amer.

Railway J., 26, March 26.
Shirley-Smith, H. (1964). The World’s Great Bridges, Phoenix House, London.
Shutraub, H. (1976). The Technical History of Construction: Developments in
Engineering Construction Technology [in Japanese], translated by Ichiro
Fujimoto, Kajima, Japan.
Smiles, S. (1904). “Lives of engineers: Medcalfe and Telford.” In History of Roads,
John Murray, London.
Stevenson, R. (1821). “Description of bridges of suspension.” Edinburgh Philosophical
J., 5(10).
Suttcliff, A., and Suttcliff, P. D. (1965). Stories from Science, Cambridge University
Press, Cambridge, U.K.
Syngleman, J. (1832). “Letter to editor: misapplication of suspension bridges.”
Mechanics’ Magazine (London), 18(490).
Timoshenko, S. P. (1953). History of Strength of Materials, McGraw-Hill, New
York; reissued 1983, Dover Publications, New York.
Vicat, L. J. (1831). “Pont suspendu en fil de fer sur le Rhône. Rapport au Conseil
d’État, Directeur Général des Ponts et Chaussées.” Annales des Ponts et
Chaussées, 1.
3
North American
Engineers Develop the
Rigid Suspension Bridge
Homecoming to America
In the aftermath of the setbacks for inventor James
Finley, North America underwent a dry spell from the 1820s
until the 1840s, when modern suspension bridges re-emerged.
These new spans were cloaked in avant garde developments
from the advanced country of France, thanks to promotion
by Charles Ellet, Jr. (1810–1862) (Fig. 3-1). Ellet was born
on a farm in Pennsylvania and, although he obtained perhaps
only a rudimentary education, he possessed a strong desire
to learn. He worked as an assistant surveyor during his late
teens, learned mathematics and French independently, and
saved money for further education. In 1830, at the age of
20, Ellet set sail for Paris for further studies.
In Paris, Ellet attended lectures at the École Polytech-
nique where he learned civil engineering. What interested
him most were modern suspension bridges, as advanced
technology was then enabling their boom in France. In 1831
Ellet visited several European countries to see suspension
bridges, including those in France. Of particular note was
the Pont Suspendu de Fribourg just starting construction in
Switzerland. By exceeding the 177-m-span Menai Bridge,
this new 266-m span would become the world’s longest.
Because work on the bridge had just begun, Ellet could wit-
ness its construction at an early stage.
71
Figure 3-1. Portrait of Charles Ellet, Jr.

Source: Shank (1974), with permission from American Canal and Transportation Center.
When Ellet returned to America in early 1832 after 2 years of study in

France, he had become a complete adherent of the French-style wire suspen-
sion bridge. Coincidentally, he learned of a competition to design a bridge
crossing the Potomac River, and he immediately proposed an approximately
300-m span suspension bridge. He was only 22 years old. He did not win that
competition, so it took a little more time for Ellet to actually build a suspen-
sion bridge; the 109-m span Schuylkill River Bridge, built in 1841 through
1842 in Fairmont, Pennsylvania, was his first (Kemp 1979). Of course, this
suspension bridge was strongly influenced by Ellet’s experiences in France,
and was indeed a French-style suspension bridge. It served for more than
30 years until it was replaced with a truss bridge in 1879 to accommodate
increased traffic.
The success of the Schuylkill River Bridge, the first wire cable sus-
pension bridge in America, elevated both the status of modern suspension
bridges and Ellet’s fame in America in a manner similar to that experienced
by Samuel Brown with his Union Bridge in Britain. Because of this early
success, Ellet obtained several major construction contracts in quick suc-
cession. Most notable was a contract for the Wheeling Bridge (308 m, to
north american engineers develop the rigid suspension bridge 73
span the Ohio River at Wheeling, West Virginia), which would exceed the
Fribourg Bridge as the world’s longest span. In 1847, a contract was initi-
ated for the Railway and Carriage Bridge over the Niagara River between
Canada and the United States (244-m span), the world’s first suspension
bridge to carry standard rail traffic.
The Wheeling Bridge was unquestionably the world’s longest suspen-
sion bridge, exceeding the Menai Bridge (177-m span) and the Fribourg
Bridge (266-m span) by a large margin. The bridge deck cleared the water
by 29.6 m and its two six-strand cables contained a total of 6,600 wires.
Its form resembled that of M. J. Chaley’s Fribourg Bridge; in the construc-
tion of the cables, Ellet adopted the same method of erecting prefabricated
parallel-wire strands. Three years were required for construction and the
bridge was completed safely in 1849 to the acclaim of local townsfolk
(Fig. 3-2).
The Niagara Railroad Bridge, which began construction in 1848,
one year after the Wheeling Bridge, was a different situation. The towns of
Wheeling, West Virginia, and Niagara, New York, were 500 km apart and,
considering the poor road situation at that time, it was difficult for Ellet to
serve as chief engineer for both projects. Nevertheless, Ellet dared to take
on all that responsibility, considering it an opportunity to realize a dream
that would never come again. So he tackled both spans by himself, without
Figure 3-2. Wheeling Bridge (West Virginia) upon completion in 1849.

Source: Shapiro (1983), with permission from Dover Publications.
giving the slightest consideration to an offer of assistance from John Roe-

bling, another excellent engineer.
In the spring of 1848, construction began on the Niagara Bridge with
the building of a temporary 235-m-span, 2.7-m-wide footbridge 3 km down-

stream from Niagara Falls at the construction site. Ellet was also quite a
showman and was adept at dealing with the press, so he held a kite-flying
contest for children before the inauguration of construction. Whoever’s kite
reached the opposite side of the gorge for the first time would be awarded
a prize of $5. By hauling successively thicker kite strings, a cableway was
eventually completed. When that was finally in place, Ellet immediately sus-
pended a gondola from it, boarded it, and performed a daring aerial traverse
100 m above the Niagara River to the applause of bystanders. A temporary
footwalk was completed in the summer of that year, and before the railing
was installed Ellet performed a second crossing of the footbridge, this time
on horseback.
But the two contracting railroad companies, the Niagara Falls Inter-
national Bridge Company (United States) and the Niagara Falls Suspension
Bridge Company (Canada), began to criticize Ellet’s performance. Ellet was
not often present at the construction site, but when he did appear he usually
did something flamboyant. Not surprisingly, his behavior was frowned upon
by discerning people. Discord between the companies and Ellet put the proj-
ect in jeopardy when, in the summer of that year and without the compa-
nies’ permission, Ellet allowed people to cross the footbridge and collected
tolls. A total of $5,000 was collected in the first 6 months alone, compared
to the $30,000 cost of the footbridge, and a dispute erupted between Ellet
and the companies over who should keep the revenue. Finally, in December
1848, Ellet was dismissed (Steinman and Watson 1957).
As I will explain below, the construction of the Niagara Railway Sus-
pension Bridge was suspended for 3 years until John Roebling assumed
responsibility for it. The temporary footbridge remained until Roebling
completed the new bridge, and was used as a connecting bridge during con-
struction, as originally planned. It was then removed in 1854, marking the
end of Ellet’s career as a suspension bridge engineer.
On May 17, 1854, Ellet’s suspension bridge at Wheeling collapsed in
a storm, just 5 years following its completion. The collapse was reported in
the local newspaper as follows:
With feelings of unutterable sorrow, we announce that the noble and

world-renowned structure, the Wheeling Suspension Bridge, has been
swept from its strongholds by a terrific storm, and now lies a mass of
ruins. Yesterday morning thousands beheld this stupendous structure, a
mighty pathway spanning the beautiful waters of the Ohio, and looked
upon it as one of the proudest monuments of the enterprise of our citi-
zens. Now, nothing remains of it but the dismantled towers looming
above the sorrowful wreck that lies beneath them.

About 3 o’clock yesterday we walked toward the Suspension
Bridge and went upon it, as we have frequently done, enjoying the cool
breeze and the undulating motion of the bridge. . . . We had been off
the flooring only two minutes, and were on Main street when we saw
persons running toward the river bank; we followed just in time to see
the whole structure heaving and dashing with tremendous force.
For a few moments we watched it with breathless anxiety, lung-
ing like a ship in a storm; at one time it rose to nearly the height of
the tower, then fell, and twisted and writhed, and was dashed almost
bottom upward. At last there seemed to be a determined twist along the
entire span, about one half of the flooring being nearly reversed, and
down went the immense structure from its dizzy height to the stream
below, with an appalling crash and roar.
For a mechanical solution of the unexpected fall of this stupen-
dous structure, we must await further developments. We witnessed
the terrific scene. The great body of the flooring and the suspend-
ers, forming something like a basket swung between the towers, was
swayed to and fro like the motion of a pendulum . . . each vibration
giving it increased momentum, the cables, which sustained the whole
structure, were unable to resist a force operating on them in so any
different directions, and were literally twisted and wrenched from
their fastenings. . . .
We believe the enterprise and public spirit of our citizens will
repair the loss as speedily as any community could possibly do. It is a
source of gratulation that no lives were lost by the disaster. (Wheeling
Intelligencer, May 18, 1854)
It was indeed a nightmarish incident of sufficient impact to deal a

severe blow to Ellet’s reputation and honor all at once. The collapse was
induced by asymmetric torsional vibration, and the reporter’s eye caught the
phenomenon shrewdly as “Each vibration giving it increased momentum.”
The collapse was attributed to the structural weakness and low stiffness
of the French-style suspension bridge that Ellet had imported to America.
There were no engineers except one who could understand the essence of
the event. Four days later The New York Times reported the events in full,
but an engineering journal simply reported that a suspension bridge had
been once again wrecked in a storm.
The Wheeling Bridge was rebuilt by John Roebling (1806–1869), an

exceptional engineer who understood the significance of the span’s 1854
failure.1 One can easily see that the reconstructed bridge (Fig. 3-3) featured
increased structural stiffness by way of a stiffening truss and stay cables

radiating from the towers, as compared with Ellet’s original bridge in Fig.
3-2. Although it is difficult to discern in Fig. 3-2, Ellet’s original bridge
featured six horizontally aligned cables in the French style, tightened by flat
steel plates to which vertical hangers for suspending the deck were attached.
By contrast, Roebling’s bridge cables were squeezed into a single round
one, wrapped with a thin wire for corrosion protection. Cable clamps were
installed from which hangers were draped to support the deck.
Clearly, Roebling’s method is superior in terms of corrosion protec-
tion, variation in stresses, cable strength, and method of connection for
the vertical hangers. It thus became a standard specification for subsequent
suspension bridges. The superiority of Roebling’s rehabilitation is so obvi-
ous that this suspension bridge remains in use today, more than 140 years
later.2
Figure 3-3. Present-day Wheeling Bridge, which was rebuilt by John Roebling after
its collapse in 1854.
Source: Courtesy of Michael Cuddy and TranSystems, Inc.
The Niagara Railway Suspension Bridge

As mentioned previously, the construction of the Niagara Railway Suspen-
sion Bridge was interrupted for three years until an engineer was appointed
to succeed Ellet, who had built only a temporary footbridge and was then
dismissed. At that time in Britain, Samuel Brown’s chain cable suspen-
sion bridges failed one after another, and in France, wire cable suspension
bridges were also plagued by accidents. Most horrific was the collapse of
the Basse-Chaîne Bridge in Angers when soldiers marched across it, setting
off large undulations that caused the bridge to fail, resulting in the tragic
loss of 226 lives. These accidents revealed the fact that suspension bridges
easily oscillated because of their low stiffness. This defect of low stiffness in
suspension bridges became critical, especially in railway bridges. Therefore,
in Britain the 146-m span Britannia Bridge, completed in 1850 and initially
planned as a suspension bridge, was built instead as a huge box girder bridge.
Robert Stephenson, the leading engineer of British railroads at that time and
probably the highest authority in civil engineering in the world, concluded
that a suspension bridge would not be suitable for railways because of its
inadequate stiffness.
Against this background, when John Roebling was named chief engi-
neer of the Niagara Railway Suspension Bridge in 1851, engineers worldwide
reacted with surprise to the apparent rashness of this project. In particular,
Stephenson immediately wrote a letter to Roebling advising withdrawal
from such recklessness. “If your bridge succeeds then mine have been mag-
nificent blunders” (Steinman 1945, 167). But Roebling had good prospects
for success. After he substituted his design for Ellet’s, Roebling accepted the
position of chief engineer. Soon after, the Niagara Railway Bridge resumed
construction in 1851. Contrary to most people’s anxieties, the world’s first
railroad suspension bridge was completed after 4 years of construction.
On March 16, 1855, the first train crossed the bridge, marking the first
time in human history that a full-size train crossed a wire-cable suspension
bridge.
The bridge marked the launch of a new era that altered the general
perception of suspension bridges. The prevailing opinion at that time was
that it was inconceivable that a suspension bridge could withstand the pas-
sage of heavy trains, but this new structure (Fig. 3-4) allayed such skepti-
cism. The Niagara Railway Suspension Bridge was duly designed for the
combined use of railroad and roadway. In fact, this bridge served well for
the following 42 years, and in the end it endured loads as heavy as two and
a half times the design load. It was surely a great success.
Figure 3-4. John Roebling’s Niagara Railway Suspension Bridge.

Source: Steinman (1945), with permission from Houghton Mifflin Harcourt.
Roebling explained the reason for his success very clearly in a report
submitted following the bridge’s completion:
The success of this work may now be considered an established fact.

The trains of the New York Central, and the Great Western Railroad in
Canada, have been crossing regularly since the 18th of March, averag-
ing over thirty trips per day.
One single observation of the passage of a train over the Niagara
Bridge will convince the most skeptical, that the practicability of Sus-
pended Railway Bridges, so much doubted heretofore, has been success-
fully demonstrated.
The work which you did me the honor to entrust to my charge,
has cost less than $400,000. This same object accomplished in Europe
would have cost four millions without serving a better purpose, or
insuring greater safety. . . .
Professional and public opinion having been adverse to Suspended
Railway Bridges, the question now turns up: What means have been
used in the Niagara Bridge, to make it answer for Railway traffic? The
means employed are Weight, Girders, Trusses and Stays. With these any
degree of stiffness can be insured, to resist either the action of trains,
or the violence of storms, or even hurricanes. And I will here observe,

that no Suspension Bridge is safe without some of these appliances. The
catalogue of disastrous failures is now large enough to warn against
light fabrics, suspended to be blown down, as it were, in defiance of the

elements. A number of such fairy creations are still hovering about the
country, only waiting for a rough blow to be demolished.
Without adding much to the weight of the structure, a surprising
degree of stiffness has been obtained by the united action of the gird-
ers and trusses. They have fully realized my expectations. The pressure
of an Engine and of a whole train of cars is much distributed, that
the depression caused is not readily observed. . . . The next means of
stiffness I have applied are wire-rope stays, above as well as below the
floors. . . . (Roebling 1855)
In order to understand the meaning of combining “girder” and “truss”

(a concept now unfamiliar to us), refer to Fig. 3-5, which is a photo of the
Niagara Railway Suspension Bridge taken from another angle. Roebling
clearly separated the upper girder (for railroad) and the lower girder (for
roadway). The main cables are also separated. Between the upper and lower
timber girders, timber posts and steel rod tension members were rigidly
assembled to make a truss. The use of a separate timber truss for a railing
Figure 3-5. John Roebling’s Niagara Railway Suspension Bridge (combined road-
way and railroad bridge).
further increased stiffness. With this combination of girder and truss, a more
rigid stiffening system was created.
Roebling’s understanding of suspension bridges was remarkable. He well
grasped the cause behind the failure of the Wheeling Bridge, and his words
appeal to us with persuasiveness. Before Roebling, there was no one who so
deeply understood the meaning of “stiffening” in suspension bridges.
Interestingly, two other engineers also built suspension bridges over
the Niagara Gorge during this period. One was Edward W. Serrell, who
built the Queenston-Lewiston Suspension Bridge, a 320-m span completed
in 1851 which linked Queenston, Ontario and Lewiston, New York. The
bridge frequently oscillated following its completion, so hold-down stays
were therefore immediately installed. However, when these stays were frozen
by drift ice they were temporarily removed to avoid the risk of damage. On
this occasion the suspension bridge easily collapsed under the force of a
strong wind (N.Y. Engineering News 1894).
Another engineer was Samuel Keefer, who built the Clifton Suspension
Bridge at Niagara Falls in 1869, forerunner to the present Rainbow Bridge
(an arch bridge). The suspension bridge’s 386-m span was the world’s longest
at that time, but its width was just 3 m. The deck was originally built from
timber, as were the stiffening trusses and towers. Many people crossed the
bridge because of its famous location, and within 2 years the timber deck
was replaced with steel. Finally, in 1888 the width was increased, transform-
ing the span’s appearance. But after about 7 months the span was struck by
a storm; it collapsed at midnight, after which virtually nothing remained
(Steinman 1945, 190–192).
Charles Ellet did not possess a deep knowledge of suspension bridges,
and neither did Keefer and Serrell. Therefore, Roebling’s prophecy came
true concerning the “suspicious bridges” they built.
John Roebling
John Roebling understood the meaning of “stiffness” in modern suspension
bridges, and it was he who successfully undertook the construction of a rail-
road suspension bridge that all the world’s authorities said was impossible.
Ellet had embraced both the French spirit and the French-style suspension
bridge, but Roebling, with the spirit of German pragmatism, sought the
“essence” of the suspension bridge and established the modern stiffened
suspension bridge.
Roebling was born and raised in Germany, and later studied at the
Berlin Royal Science and Engineering School (later Berlin University), the
leading educational institute in science and engineering in Germany. As a

student, Roebling’s interest in suspension bridges was sparked by a lecture
on bridge engineering given by J.F.W. Dieteleyn. When he learned that
Germany’s first suspension bridge was being built in Banberg, Roebling

went there to observe its construction, and eventually selected suspension
bridges for his graduation thesis (Kawada 1974, 20–21). After graduating
in 1836, Roebling entered employment with the Prussian government and
worked for 3 or 4 years, but soon became tired of life as a civil servant.
In 1831, at the age of 25, Roebling emigrated to New World America. In
the beginning, he dreamed of large-scale farming, as did many immigrants,
but the results were not favorable. Around 1837, he asked his wife to
take care of the house and farm and sought a career in canal and railroad
building.
Having once again returned to construction, Roebling first noticed
the difficulties posed by the handling and weakness of hemp rope. Time
and again hemp ropes were the cause of major accidents. Roebling
searched for a stronger rope and devised a wire rope formed by twisting
iron wires. With a primitive second-hand machine for making hemp rope
and trials from scratch with the aid of nearby farmers, Roebling invested
in and successfully fabricated wire ropes. The year was 1841 and he was
just 35 years of age. The next year Roebling obtained a patent for the
“Manufacturing Method of Wire Rope” and his manufacturing business
became successful. This marked the birth of the wire rope industry, and
the nascent Roebling company would become the leading manufacturer
in the United States—and, as such, in the world—by the time his children
matured.
During the process of developing wire rope, Roebling deepened the
knowledge of iron wire. He eventually concluded that iron wire would
provide the best solution for the economical construction of large bridges
requiring a major leap in span, the biggest challenge for bridge technology.
In 1841, Roebling then went one step further by transforming the aerial
spinning concept for parallel-wire cable devised by Vicat into a practical,
modern method. He described it thus:
The nature of my invention consists in the combination of any number

of wires, laid parallel to each other, so that they form a round cylinder,
and occupy the same position respectively for their entire length; all the
wires to be uniformly strained in thus collecting them, and the whole
to be wrapped with wire, either at intervals or throughout the whole
length. . . . The wrapping may be made of annealed wire, so as to form
a close cover.
There are two different methods by which the uniform strain of

all the wires composing the rope can be attained. . . . When all the wires
are suspended so that they deflect equally, they will be equally strained
by their own weight. A change of temperature cannot affect the pro-

cess injuriously, as all the wires will rise or sink in proportion as they
are contracting or expanding. . . . All the wires being placed parallel to
each other, uniformly strained throughout and not twisted, the greatest
strength will be obtained by the least quantity of wire. (Roebling 1841)
The opportunity to demonstrate Roebling’s concepts of parallel-wire

cables and aerial spinning came unexpectedly early. One of the aqueduct
bridges crossing the Allegheny River in Pittsburgh was irreparably damaged
by drifting ice in the winter of 1843–1844 and required immediate replace-
ment. The Canal Company launched a public competition for the design,
and Roebling, who submitted the only suspension bridge among the total of
43 proposals, was awarded the contract. His bridge consisted of seven spans
of 49.4 m each that supported a 2.6-m-deep aqueduct, 5 m wide across the
top and 4.3 m wide across the bottom. The bridge would have to bear the
weight of water exceeding 2,000 tonnes. The cable actually consisted of two
separate cables per side, each made of 1,900 wires bundled into a cable of
17.8-cm diameter. The aqueduct was filled with water on May 22, 1845.
Some people made disquieting prophecies such as “The bridge will fall if
water flows,” but the aqueduct did not move at all and went on to be a
great success.
With this success, Roebling built four similar aqueduct bridges in the
period up to 1850 (Steinman 1945, 105; Kawada 1974, 44). One of them, the
164-m-long aqueduct consisting of 4 spans of 41 m built in Lackawaxen,
Pennsylvania, exists to this day; although no water flows across the bridge, it
is still used as a pedestrian and one-lane road bridge (Fig. 3-6). It is unques-
tionable that his aqueduct bridges are structurally identical to the suspen-
sion bridge. Wire cables endured the weight of thousands of tonnes of canal
water in the same manner as from the loading of locomotives and trains.
With the accumulation of this valuable experience, Roebling soon
tackled the Niagara Railway Suspension Bridge. Prior to this bridge, Roe-
bling had built his first vehicular suspension bridge, the Monongahela
Bridge at Pittsburgh, between May 1845 and February 1846. This span
replaced a covered bridge built by Louis Wernwag that had been destroyed
by fire, and was constructed by reusing the piers. It was an eight-span struc-
ture close to 500 m in total length. Roebling’s sketch of this bridge is shown
in Fig. 3-7.
Figure 3-6. John Roebling’s aqueduct bridge at Lackawaxen, Pennsylvania.

Roebling’s role in building the Niagara Railway Suspension Bridge from

1851 to 1855 has already been described. With this success, the feasibility
of supplying necessary stiffness to modern suspension bridges—heretofore
having experienced repeated trials and errors since the days of Finley—was
demonstrated to the world. Henceforth, the suspension bridge was posi-
tioned to become the champion of truly long-span bridges.
The first long-span bridge that Roebling built was the Ohio River
Bridge (also known as the Cincinnati-Covington Bridge) at Cincinnati. This
bridge leapt the river in a world-record span of 322 m. Initially, work began
with a 3- to 4-year construction schedule, but the financial panic which
began in 1857, combined with the disruption of the Civil War between 1861
and 1865, interrupted the work. Finally, construction finished at the end of
1866 after 10 years of hardship, and the bridge was opened on New Year’s
Day, 1867. Figure 3-8 is a Harper’s Weekly illustration of the Ohio River
Bridge immediately after its completion, showing the deck accommodating
two tracks for horse-drawn trolley cars, stiffening trusses, and Roebling-
style tower stays stretching from the towers.
For more than 130 years, this suspension bridge built by Roebling has
served as an important means for transportation across the Ohio River and
continues to be a Cincinnati landmark. A bronze statue of Roebling holding
a triangular scale stands by the bridge (Fig. 3-9), and the bridge is affection-
ately called “Roebling’s Bridge” by the local citizenry (Fig. 3-10).
nloaded from ascelibrary.org by New York University on 05/10/15. Copyright ASCE. For personal use only; all rights rese
84
Figure 3-7. Monongahela Bridge (a sketch by John Roebling).
Figure 3-8. Ohio River Bridge at Cincinnati upon completion.

Figure 3-9. Bronze statue of John Roebling beside the Ohio River Bridge.
Source: Author.
The Brooklyn Bridge

The completion of the Ohio River Bridge was well publicized in Europe.
Furthermore, Americans were now planning to build the gigantic Brooklyn
Bridge in New York City. A British technical journal reported:
Having completed this work, Mr. Roebling is about to undertake one

of far greater magnitude, that for connecting the city of New York with
Brooklyn, across the East River. It is necessary, here, that a headway of
one hundred and thirty feet [40 m] be maintained in midchannel, and
Figure 3-10. The Ohio River Bridge between Cincinnati, Ohio, and Covington,
Kentucky, also known as “Roebling’s Bridge.”
Source: Author.
the great suspension span will be one thousand six hundred feet [488 m]
in the clear, or much more than twice that at Clifton, nearly three times
that of the Menai suspension bridge. . . . This great opening will be
approached, on either side, by a succession of arches, the whole length
being between one and a half and two miles [2.4 to 3.2 km], and the
total estimated expense is six million dollars, or, at the present value of
American currency, nearly one million sterling. (Engineering 1867)
It was not simply that the span was long; the bridge was also planned
with an unprecedented deck width of 26 m accommodating two roadway
lanes in each direction, in addition to two transit tracks, an elevated pedes-
trian way in the center and, underneath, space for various pipes and util-
ity lines as shown in Fig. 3-11(top). The bridge was designed to support a
modern transportation system, and thus changed the entire appearance of
existing suspension bridges (Brooklyn Museum 1983, 173). Figure 3-11 also
shows later transitions of this suspension bridge to cope with evolving trans-
portation demands. The change in 1898 involved the installation of trolley
tracks as shown in (middle) of the same figure. With the advancement of the
motor age, however, all the tracks were removed in 1952 to provide three
lanes in each direction, thus creating a six-lane roadway; the structure of the
Figure 3-11. Evolution of Brooklyn Bridge deck cross section.

(top) Upon completion in 1883; (middle) With two trolley tracks installed in 1898; (bottom)
Structure altered to present-day road and pedestrian configuration in 1952 by removal of trol-
ley tracks (design by D. B. Steinman).
Source: Brooklyn Museum (1983), courtesy of New York City Department of Transportation.
stiffening truss was also altered to its present appearance, as shown in Fig.
3-11(bottom). In any event, while suspension bridges were believed to have
the advantage of economical construction and were considered suitable for
low-traffic regions such as mountainous areas, Roebling claimed that they
could carry modern urban traffic, including railways. Their later develop-
ment would endorse the correctness of his claim.
As a prelude to construction, the New York Bridge Company was estab-
lished on April 16, 1867, to build the Brooklyn Bridge, and on May 23 the
company appointed Roebling its chief engineer. At this point, it appears that
Roebling had completed all necessary investigations and designs, for on Sep-
tember 1, 1867, only 3 months following his appointment as chief engineer, he
submitted a report describing the bridge concept (Roebling 1867). Although
it was called a preliminary investigation report, the document described the
bridge location and provided design details that proved to be almost identical
to the bridge as it was actually built; quantities and costs were also precisely
estimated. A single, complete vision of the majestic Brooklyn Bridge crossing
the East River was already in Roebling’s mind (Fig. 3-12).

Roebling never saw the completed Brooklyn Bridge. On June 6, 1869,
while surveying prior to construction, he suffered an accident. From the
injury, fatal tetanus set in; on July 22, just before the inauguration of con-
struction, John Roebling died at the age of 63. Not until August did the
Authority officially appoint Washington Augustus Roebling (1834–1926),
John Roebling’s son, as chief engineer to succeed him. At this time, Washing-
ton was just 32 years old, a young engineer who had majored in civil engineer-
ing at the Rensselaer Polytechnic Institute in Troy, New York. No one other
than Washington was considered as a suitable successor because he learned
directly from his father and had always assisted him in the investigation and
the design of the Brooklyn Bridge (Fig. 3-13).
On January 3, 1870, the construction of the nineteenth century’s great-
est bridge began with work on the tower foundations. The famous drama
of its construction started; Washington would realize his father’s dream
after enduring 13 years of hardship (McCullough 1972). This was prior to
Thomas Edison’s inventing the incandescent light bulb, so oil lamps and gas
lights were commonly used for illumination. The pneumatic caissons for the
foundations were made of wood and their joints were filled with the rags of
old fishing nets for airtightness.
Figure 3-14 illustrates a cross section of the timber caisson for the Brook-
lyn tower foundation, and Fig. 3-15 shows the working situation inside the
Figure 3-12. The dream of John Roebling (original design).

Source: Steinman (1945), with permission from Houghton Mifflin Harcourt.
Figure 3-13. John A. Roebling (left) and his son, Washington Roebling.
Sources: (John Roebling) Steinman (1945), with permission from Houghton Mifflin Harcourt;
(Washington Roebling) Shapiro (1983), with permission from Dover Publications.
caisson. Work continued under severe conditions unimaginable today; fires

occurred frequently, interrupting construction.
Even more ominous were attacks of the bends (“decompression sick-
ness”). At that time, without knowing what caused this deadly condition
or how to cure it, underwater workers died one after another as the New
Figure 3-14. Construction of the Brooklyn tower caisson.

Source: Harper’s Weekly, December 17, 1870, with permission from the Brooklyn Museum
Library Collection.
Figure 3-15. Work inside the caisson.

Source: Courtesy of the New York Public Library.
York tower caisson sank deeper. Finally, chief engineer Washington Roe-
bling himself collapsed inside the caisson, a victim of the bends. He was
just 35 years of age. Miraculously, he survived, but his lower body was
paralyzed and he lost physical strength. A vocal cord ailment impaired his
ability to speak and he also had difficulty in hearing. It is said that after this
Figure 3-16. Washington Roebling directing Brooklyn Bridge construction from

his bed.
Source: Brooklyn Daily Eagle, 1948.
accident Roebling never visited the bridge construction site again. Figure
3-16 depicts the ailing and haggard Roebling observing the construction
site through his nearby apartment window and giving directions. His wife
Emily became his hands and feet, delivering his instructions to the workers
at the construction site.
Ever since human beings began using suspension bridges, and even fol-
lowing the transformation of these bridges into modern form, they contin-
ued to be susceptible to deflections and oscillations, and remained a poten-
tial danger if exposed to strong wind. It was John Roebling who had added
the necessary stiffness to the suspension bridge and demonstrated that it
could be a modern bridge form suitable for use in long spans. The Brooklyn
Bridge was the pinnacle of Roebling’s engineering knowledge, but was com-
pleted only through the devotion of his son Washington and his wife.
Figure 3-17 is a lithograph of the Brooklyn Bridge upon completion
in 1883. It appears to have been widely distributed as a poster, because
one was brought home by a Japanese citizen who visited New York at that
time. This poster was dedicated to a famous shrine in Kotohira, Sanuki
Prefecture, Japan, as a votive picture tablet (Kawada 1985, 215). The story
goes that this lithograph sparked the first proposal for the future Honshu-
Shikoku system of bridges connecting the islands of Honshu and Shikoku
across the Inland Sea of Japan (Murakami 1998, 94).
Figure 3-17. Brooklyn Bridge upon completion in 1883.

Source: Currier and Ives lithograph, 1883.
Once suspension bridges acquired the necessary stiffness, their position

and fame in the field of long-span bridges in the United States was confirmed.
However, anxiety over their use had not been completely eliminated in Europe.
An illustrative example was the Forth Railway Bridge, built near Edinburgh,
Scotland, between 1882 and 1889 (Paxton 1990). This bridge was origi-
nally designed by Thomas Bouch as a two-span suspension bridge, a unique
configuration as shown in Fig. 3-18(A). In spite of more than a year having
passed following the actual start of foundation construction, the design was
abruptly changed to a cantilever truss bridge as shown in Fig. 3-18(B). The
reason for this change was the unfortunate 1879 collapse of the Tay Bridge,
which spanned the Firth of Tay in Scotland. It had been designed by the same
Thomas Bouch, and this disaster resulted in his resignation as chief engineer
for the Forth Railway Bridge (Kawada 1987). But the real reason was that
British engineers had not yet learned how to stiffen suspension bridges and
were nervous about accommodating railroads on these types of structures.
This brings to mind Bouch’s use of stays extending from the tower juncture
at roadway level up to the cable (i.e., cable stays). Compared with Roebling’s
tower stays, which stretched from the tower tops to the deck, cable stays (in
the nineteenth-century sense of the term, not today’s commonly used term)
are considered to have quite different effects in the restraint of deformations
of the bridge deck.
Figure 3-18. Original plan (A) and final plan (B) for the Forth Railway Bridge near
Edinburgh, Scotland.
Source: Paxton (1990), with permission from Thomas Telford Publishing.
Notes
1. Emory Kemp claimed a different view from what is described in this book, stating
that the Wheeling Bridge was reconstructed by Ellet (Kemp 1979). Kemp (1999)
reiterated his view in a later paper, but since the reconstructed suspension bridge
embodies the Roebling style, I have reflected Steinman’s opinion (Steinman and
Watson 1957, 208).
2. No Finley-built suspension bridge remains extant. Therefore, the Wheeling suspen-
sion bridge is the oldest existing modern suspension bridge in North America.
References
Brooklyn Museum. (1983). The Great East River Bridge, 1883–1983, Abrams, New
York.
Engineering. [Editorial], London, Vol. 4, July, 1867.
Kawada, T. (1974). A Story of Brooklyn [in Japanese], Kagaku Shokan, Japan.
Kawada, T. (1985). Bridges and Romance in History [in Japanese], Gihodo, Japan.
Kawada, T. (1987). Egg of Beaumont: Truth of Failure of the Tay Bridge [in Japa-
nese], Kensetsu Tosho, Japan.
Kemp, E. L. (1979). “Links in a chain: the development of suspension bridges
1801–70.” Structural Engineer, August.
Kemp, E. L. (1999). “Charles Ellet, Jr. and the Wheeling suspension bridge.” Proc.,
Int. Conf. on Historic Bridges, West Virginia University Press, Morgantown,
W. Va.
McCullough, D. G. (1972). The Great Bridge, Simon and Schuster, New York.
Murakami, K. (1998). Spanning the 21st Century: A Story of the Akashi Kaikyo
Bridge, Maruzen, Japan.
N.Y. Engineering News. (1894). “The old and new suspension bridges over the
Niagara River at Lewiston.” N.Y. Engineering News, 42(2).
Paxton, R. (1990). 100 Years of the Forth Bridge, Thomas Telford, London. Trans-
lated into Japanese by the Study Committee on Civil Engineering History, Japan
Society of Civil Engineers, 1992.
Roebling, J. A. (1841). “Specifications.” Patent application for “New and Improved
Mode of Manufacturing Wire Ropes,” March 27. Rutgers University, New
Brunswick, N.J., Alexander Library, Roebling Family Archives.
Roebling, J. A. (1855). Final Report of John A. Roebling, Civil Engineer, to the Pres-
ident and Directors of the Niagara Falls Suspension and Niagara Falls Interna-
tional Bridge Companies, Lee, Man & Co., Rochester, N.Y.
Roebling, J. A. (1867). “To the President and Directors of the New York Bridge
Company on the Proposed East River Bridge,” September 1, 1867.
Shank, W. H. (1974). Historic Bridges of Pennsylvania, American Canal and Trans-
portation Center, York, Pa.
Shapiro, M. J. (1983). A Picture History of the Brooklyn Bridge, Dover, New York.
Steinman, D. B. (1945). The Builders of the Bridge: The Story of John Roebling and
His Son, Harcourt, Brace, New York.
Steinman, D. B., and Watson, S. R. (1957). Bridges and Their Builders, Dover Pub-
lications, New York.

4
Emergence of the
Deflection Theory
and Its Development
Theory of the Stiffened

Suspension Bridge
The early studies concerning the theoretical expla-
nation of suspension bridges included the force polygon,
catenary, and parabola. Louis Henri Navier compiled and
published his work in his Investigation and Study Report
on Suspension Bridges which, as described earlier, became
a bible for engineers for almost half a century. However,
Navier dealt only with the theory of unstiffened suspen-
sion bridges, in which the stiffening girder does not exist in
terms of our contemporary understanding (Navier 1823).
The theoretical treatment of suspension bridges stiff-
ened either by girder or truss was probably first attempted
by William Rankine in his 1858 Manual of Applied Mechan-
ics. Rankine presented Peter Barlow’s opinion that “From
the results of model testing, I found that the stiffness of the
suspension bridge could be increased by way of an unex-
pectedly lightweight girder” (Rankine 1858). Rankine sub-
sequently obtained an approximate solution to determine
the stress in a stiffening girder by assuming its deflected
curve, and concluded that “[T]he transverse strength of the
stiffening girder should be four twenty-sevenths of a simple
girder of the same span to bear a uniform load of the same
intensity.” Whether or not this is true, it seems that this
97
investigation was the first theoretical analysis of a stiffening truss (Timosh-

enko 1953, 200–201). Subsequently, in 1877, Wilhelm Ritter of the Riga
Institute of Technology (in today’s Republic of Latvia) attempted a theoreti-
cal treatise of the stiffened suspension bridge.

The methods developed by Rankine and Ritter were not widely accepted
because these solutions assumed an extremely high degree of rigidity in the
stiffening girder, while stating “[S]uspension bridges can be stiffened by unex-
pectedly lightweight girders.” If one accepts the assumptions of Rankine and
Ritter alike, stiffening girders do not deform from their initial configurations
even after live loads are applied, and thus act as a rigid body.
When John Roebling began the design of the Brooklyn Bridge in 1867,
existing theories of the suspension bridge involved either this rigid-body
theory or the theory of Navier-style unstiffened suspension bridges. Roe-
bling was dissatisfied with both approaches, and as a result he designed the
bridge independently from them. As Gustav Lindenthal later pointed out,
Roebling experimentally determined the necessary dimensions without the
benefit of theory in designing his series of suspension bridges, including the
Brooklyn Bridge:
The first theory for the stiffening of a suspended span by the English
engineer and mathematician, Rankine (1855), was not used in the
United States as far as is known, because it gave heavier and more
expensive trusses than the empirical Roebling System—low stiffening
trusses with stays up to the towers, successfully used for road bridges
in this country. The deflection and stresses in the Brooklyn Bridge were
studied with the aid of a model which indicated for the long span a
sufficiently stiff structure under the exigencies of highway traffic. . . .
(Lindenthal 1934)
Even so, the Brooklyn Bridge generated rapidly increasing interest in

the theory of stiffened suspension bridges. Several books and papers were
published during its construction that attempted to predict the actual behav-
ior of the stiffening girder. For example, in 1880 Celeste Clericetti published
The Theory of Modern American Suspension Bridges, in which he presented
an approximate solution for a Brooklyn-type suspension bridge with cables,
hangers, tower stays, stiffening trusses, and other features. In the following
year Charles Bender published Suspension Bridges of Any Desired Degree
of Stiffness, in which he described a suspension bridge with a combination
of elastic cables and elastic girders. Also in 1881, Müller-Breslau presented
the Theory of Chain Stiffened by Beam, which really took its cue from the
Elastic Theory (discussed in detail below).
emergence of the deflection theory and its development 99
On the other hand, the engineering theory of the arch bridge was
remarkably well developed at this time, and Alberto Castigliano published the
Theory of Equilibrium of Elastic Structures and its Application in 1879. In
1886, Maurice Léby published Calculation Method for Stiffened Suspension

Bridges and 2 years later Joseph Mélan’s Theory of Steel Arches and Suspen-
sion Bridges was published, which could be considered the definitive version
of the Elastic Theory.
Nowadays, in discussions of the early theories of stiffened suspen-
sion bridges, only Mélan is popular, but in fact many other scholars and
engineers explored the subject, as just described. Mélan’s first presentation
of the Elastic Theory in 1888 contained an error in the way thermal stress
was treated. He also devoted his efforts only to an analysis of three-hinged
suspension bridges with a hinge at center span. However, he continued to
refine his ideas in subsequent editions, and in the 1906 edition Mélan finally
presented a generalized theory for the suspension bridge with two-hinged
stiffening girders. During this period, M. T. Godard published a solution
for stiffened suspension bridges using influence lines (Godard 1894). The
historical transition of the theory of stiffened suspension bridges was later
described by Leon Moisseiff:
With the advent of suspended stiffening trusses Rankine in England and

Ritter in Switzerland, at about the middle of the last century, developed
the theory of the rigid stiffening truss. According to the assumptions of
this theory, the suspended truss remains absolutely rigid under the live
load, and the original curve remains undisturbed. This assumption by
itself tacitly eliminates the effect of the dead load of the bridge. As spans
were comparatively low and trusses were easily made very stiff, the error
so committed while substantial, was not serious. In later years, with the
growth of the knowledge of elastic structures, the stiffening truss began
to be treated as an elastic member. Müller-Breslau and Mélan developed
to a higher degree the elastic theory of suspension bridges. Perhaps,
because few and comparatively small suspension bridges were built in
Europe at that time, the effect of deformation in these bridges was not
given due attention. At any rate, the original assumption contained in
the Rankine theory, that the cable geometry remains undisturbed, was
tacitly taken over by the elastic theory. (Moisseiff 1925)
The theory that ignores cable deformations from live loads is called the
Elastic Theory of suspension bridges. By contrast, a more accurate theory
that includes cable deformation is generally called the Deflection Theory.
Here I will try to distinguish the differences between these two theories.
Figure 4-1. The relationship between the cable and the stiffening girder in suspen-
sion bridges.
Source: Author.
Figure 4-1 displays the relationship between the cables and the stiffen-
ing girder in suspension bridges. Since the Elastic Theory assumes cables
that do not deform under live loads, the basic equation of the Elastic Theory
is as follows:
M ⫽ M0 ⫺ Hy (4-1)
where
M ⫽ bending moment at any section x along the stiffening girder
M0 ⫽ bending moment as a simple beam
H ⫽ additional horizontal tension in cable due to live load
y ⫽ cable ordinate.
The term on the right side of this equation, Hy, is a product of the
increment of the cable’s horizontal tension due to live load and the cable’s
vertical distance as a reduction of bending moment from that of a simply
supported beam. This is an effect of being suspended from a cable, which is
an advantage of the suspension bridge.
The term y in this case is shown as the coordinate system in Fig.
4-1. When the dead load is considered as a horizontally uniform load, this
becomes a parabola as expressed below:
4f
y⫽ (l ⋅ x ⫺ x2 ) (4-2)
l2
where
l ⫽ span length
f ⫽ cable sag.
On the other hand, in the Deflection Theory, the deformation, d, to be

produced in the cable (and therefore a deformation of the stiffening girder)
is considered. As a result, the bending moment in the Elastic Theory is fur-
ther reduced by the effect of d, (Hw ⫹ H)d, that is:
M ⫽ (M0 ⫺ Hy) ⫺ (Hw ⫹ H)d (4-3)
where Hw is the cable’s horizontal tension due to dead load, w.

It is important to note that this basic equation of the Deflection Theory
as shown in Eq. 4-3 contains the term of the cable’s horizontal tension Hw
due to dead load and, by this, the “effect of dead load in suspension bridge”
pointed out by Moisseiff is considered.
Differentiating Eq. 4-3 for a second time, we have:
d 2 M ⎡ d 2 M0 d2y ⎤ d2η
⫽ ⎢ ⫺ H ⎥ ⫺ (H w ⫹ H )
dx2 ⎣ dx
2
dx2 ⎦ dx2
where the second derivative of the bending moment generally gives the
load Px and the fourth derivative of deflection d equals the second-order
derivative of the bending moment. The following equation will become
Eq. 4-4, which is the other form of the basic equation of the Deflection
Theory derived by Mélan:
⎡ d4η d2y ⎤ d2η

Px ⫽ ⎢ EI 4 ⫺ H 2 ⎥ ⫺ (Hw ⫹ H) 2
⎣ dx dx ⎦ dx
or
d4η d2η d2y

EI ⫺ (H w ⫹ H ) ⫽ Px ⫹ H (4-4)
dx 4 dx2 dx2
where
Px ⫽ live load
E ⫽ Young’s modulus of the stiffening girder
I ⫽ moment of inertia of the stiffening girder.
The second term on the right-hand side of the equation includes y,

which is not relevant to d. Differentiating Eq. 4-2 for a second time, we
obtain d2y/dx2 ⫽ ⫺(8f /l2): const. Therefore Eq. 4-4 can be expressed as
follows:
d4η d2η 8f
EI ⫺ (H w ⫹ H ) ⫽ Px ⫺ H 2 (4-4v)
dx 4 dx2 l
When the right-hand term Px ⫺ H(8f/l 2) in Eq. 4-4v takes on a constant

value, it becomes a differential equation with a fourth-order derivative of d,
and it is now feasible to obtain a solution. However, it is tedious to solve this
equation in practice, and Mélan himself thought it impractical. He therefore
considered the Elastic Theory to be sufficient for general use in suspension
bridges. As a point of interest, Mélan always called it the “More Exact
Theory” (Shimada 1964), and the term Deflection Theory was adopted by
Moisseiff and his colleagues. The term appeared early in the twentieth cen-
tury only in the United States.
The Emergence of the Deflection Theory

In 1801, the modern suspension bridge emerged in North America with
James Finley’s Jacob’s Creek Bridge. It was then embraced by engineers in
Britain, then France, and subsequently returned to North America. Through-
out the nineteenth century, cable materials gradually evolved from chain, to
eyebar, and then to wire, thus extending possible span lengths.
The pinnacle of nineteenth-century suspension bridges is the Brooklyn
Bridge. This bridge extended the span range up to 486 m and solved the
stiffness problem, the weakest element in suspension bridges, and thus con-
stituted a fully developed form of the modern long-span suspension bridge.
Figure 4-2 displays a genealogical diagram of suspension bridge development
compiled by Emory L. Kemp, with the Brooklyn Bridge at its pinnacle.
The twentieth century started with an accident on the Brooklyn
Bridge on July 24, 1901 that required the sudden closure of the bridge
for 36 hours. New Yorkers were temporarily deprived of transportation
service across the East River. Engineering News reported the situation: “At
near the midspan of the north main cable in New York’s Brooklyn Bridge,
an accident of successive failure of nine suspenders occurred. Therefore the
north lane was closed for 36 hours during the replacement” (Engineering
News, August 1901).
Figure 4-2. Genealogical development of modern suspension bridges in the nine-

teenth century.
Source: Kemp (1979), with permission from Emory L. Kemp.
In fact, this accident triggered the new development of suspension

bridges in the twentieth century and the era of long-span suspension bridges,
for it was through this accident that Leon Moisseiff, a tragic engineering
genius, appeared on the stage:
In the summer of 1901, due to a combination of causes, a number of

suspension rods in the main span of the Brooklyn Bridge failed. At that
time this was the only crossing over the East River and carried most of
the passenger traffic from Brooklyn, N.Y. to New York City. The Wil-
liamsburg Bridge was then in the process of construction. The uninter-
rupted travel over the Brooklyn Bridge was then of vital importance to
the inhabitants of New York. Great public interest was aroused by the
incident. . . .
During the controversy on the behavior and safety of the Brook-
lyn Bridge which followed the publication of the engineering report, the
writer worked on an analysis of the stresses and deformations of that
bridge. He found that while there were no official measurements of the
deflection of the bridge under load, such deflections as could be observed
daily were of much less magnitude than would result from computations
based on the elastic theory. Further considerations showed that, neglect-

ing altogether the effect of the stiffening trusses on the deflection of the
bridge and going back to the simple computations of the equilibrium
polygons under dead load and under dead and live load, much closer
agreement with the actual behavior of the bridge would result. It became
apparent that the tacit neglect by the Elastic Theory of the effect of the
dead load on the deformation of the bridge under live load was a serious
error for large bridges. This is explained by the fact that the Rankine-
Ritter theory was based on a rigid stiffening truss and thus logically
introduced the error. It was later taken over without further scrutiny by
the “fathers” of the Elastic Theory. Because of this, the latter theory is
grossly in error for suspension bridges. (Moisseiff 1935, 1205–1206)
In this way, the twentieth century began with a denial of the Elastic
Theory; ironically, this accident provided the opportunity to do so. Instead
of the now-discredited Elastic Theory, attention was focused on the “More
Exact Theory” once again, the one that Mélan had treated less seriously
because he considered it to be impractical.
Leon S. Moisseiff, then a young and spirited engineer still under 30 years
of age, rediscovered Mélan’s theory and correctly appreciated its significance
even more than Mélan himself. Moisseiff established the theory as ortho-
doxy in the design of twentieth-century long-span suspension bridges. Mois-
seiff described the development of the theory as follows:
While working on the development of a more correct analysis, the

writer directed his attention to the short treatment given by Mélan to
what he called “The More Exact Theory” [reference to Mélan (1888)].
Strangely, Mélan did not put much emphasis on the practical results of
the theory. The writer found that Mélan’s analysis of a one-span sus-
pension bridge, taking into account the effect of the dead load on the
deformation of the bridge under live load, gave solutions which closely
approached actual conditions.
. . . Realizing that there can be no theory which, applied to a
practical structure, will be absolutely exact, the writer subsequently
denoted the developed theory by the name of “The Deflection Theory,”
having in mind that it is based on the deflections of the cable and of the
truss. Thus, the Manhattan Bridge was designed in 1904 by the Deflec-
tion Theory, and may be considered the first modern suspension bridge.
(Moisseiff 1935, 1206)
The great importance of the Deflection Theory is not a math-
ematical one. While resulting in greater accuracy, its chief merit in com-
parison with the Elastic Theory is not that of greater accuracy but that
of very substantial economy. (Moisseiff 1925)
This was how the “More Exact Theory of Suspension Bridges”—abandoned

by Mélan—was rediscovered by New York engineer Moisseiff. Newly renamed
the “Deflection Theory of Suspension Bridges,” it underwent a reconsidera-
tion that made it suitable for designing long-span suspension bridges.
The first suspension bridge to be calculated using the Deflection Theory
was the Manhattan Bridge, as Moisseiff pointed out. Built from 1904 through
1909, this 450-m span structure spanned the East River near the Brooklyn
Bridge (Fig. 4-3). In 1907, during its construction, an accident occurred at
the 549-m span Québec Bridge in Canada, which upon completion was to
have been the world’s longest span. During erection of the truss by cantile-
vering, the cantilever arm buckled and collapsed, killing 74 people (Hopkins
1970, 160–164; Tanaka 2000). As a result of this tragedy, the stability of
large-scale bridges came under immediate scrutiny.
The engineering verification of the Manhattan Bridge was entrusted
to Ralph Modjeski (1861–1940), an authority on bridges. Modjeski had
graduated at the top of his class from the École des Ponts et Chaussées
Figure 4-3. The Manhattan Bridge (center) and the Brooklyn Bridge (bottom right).
Source: Courtesy of Bojidar Yanev.
in Paris, then emigrated to America and entered bridge engineering prac-

tice. Acknowledged as the leading authority in bridge engineering, he was
appointed to the Committee of Reconstruction for the Québec Bridge. In
this capacity, Modjeski completely reexamined the Deflection Theory, and

the results were reported in his 1909 report on the Manhattan Bridge. In this
report, Modjeski officially verified the accuracy of the Deflection Theory, as
had been claimed by Moisseiff.1 Modjeski later went on to design what was
then the world’s longest suspension bridge, the 533-m span Delaware River
Bridge which linked Camden, New Jersey, and Philadelphia, Pennsylvania.
For this bridge, without hesitation, Modjeski appointed Moisseiff as engi-
neer of design.
This is how the Deflection Theory became established as the leading
design theory for long-span suspension bridges. Subsequent long-span bridges
in America were all designed applying this theory.
The Development of Suspension Bridges

in the Early Twentieth Century
As the twentieth century commenced, modern suspension bridges in North
America began to make great strides toward longer spans. The first such
effort was the Williamsburg Bridge (489.5-m span, connecting Manhattan’s
Lower East Side and Brooklyn), which in surpassing the Brooklyn Bridge
became the world’s longest suspension bridge and ended the former span’s
20-year reign.
The completion of this bridge in 1903 made it technically a twen-
tieth-century bridge, but it was actually designed toward the end of the
1800s. Therefore, its design was not based on the Deflection Theory, but
instead on the old Elastic Theory. As stated earlier, it is not correct to say
that the Brooklyn Bridge was designed using the Elastic Theory—John
Roebling had determined the sections experimentally with the combina-
tion of tower stays. His was a unique concept (Rankine 1858, Chapter 1;
Moisseiff 1935). Therefore, the Williamsburg Bridge was the first example
in which the Elastic Theory was applied to a suspension bridge of nearly
500-m span. Its engineer was Leffert L. Buck (1837–1907). Born in New
York, Buck entered the Civil War while a sophomore student at St. Law-
rence College. After retiring from the army, Buck studied civil engineering
at Rensselaer Polytechnic Institute. Following graduation, he was involved
in a wide range of bridge construction in North and South America, includ-
ing the rehabilitation of Roebling’s Niagara Railroad Suspension Bridge.
Buck eventually demolished this suspension bridge and replaced it with an

arch span that remains in service to this day.
In 1895, Buck was appointed chief engineer of the Bridge Department
of New York City and became the engineer responsible for technical mat-
ters associated with the Williamsburg Bridge. At that time, railroads and the
horse-drawn carriage constituted the primary transportation modes. The
advisability of running railroad tracks on suspension bridges was still ques-
tioned in Europe, and while John Roebling had just proven it could be done,
even his Niagara Railroad Bridge lasted at best only 40 years and had to be
replaced by a more rigid arch bridge.
The Williamsburg Bridge (Fig. 4-4) was designed using the Elastic
Theory, and as such reflected the load and deflection criteria of a railroad
bridge; it was consequently stiffened by a truss of twice the depth of the
Brooklyn Bridge. J. K. Finch, a professor of civil engineering at Columbia
University, commented:
With the construction of the Williamsburg Bridge in 1903, the greatly

increased loads of suburban cars and the new theory of stiffening truss
action led to the building of the heaviest, clumsiest suspension which
ever has been or probably ever will be erected. Recalling the grace-
ful lines of the Brooklyn Bridge, the clumsy towers, the huge stiffen-
ing truss, 40 ft [12 m] deep, and the use of vertical suspenders only,
make Williamsburg look like an elephant in comparison with a giraffe.
(Finch 1941)
As Finch points out, with a 12-m truss depth and a 1⬊40 depth-to-span
ratio, the Williamsburg Bridge is proportionally the most massive suspension
Figure 4-4. Cross section of the Williamsburg Bridge.

Source: Courtesy of Mary Evans Picture Library.
bridge ever built. By way of comparison, the Brooklyn Bridge features a

5.3-m-deep stiffening truss and a depth-to-span ratio of 1⬊90, and the Man-
hattan Bridge has a 7.3-m-deep stiffening truss with a depth-to-span ratio
of 1⬊60.
From the photos in Figs. 4-5 through 4-7, taken from nearly the same
angle for three suspension bridges—the Brooklyn, the Williamsburg, and
the Manhattan—the heavy appearance of the Williamsburg Bridge is easily
noticed at first glance.
Buck adopted a new type of tower, using steel instead of the traditional
masonry, yet the towers were notorious for their clumsiness and “bandy-
legged” appearance. One of the critics of the Williamsburg Bridge was Gustav
Lindenthal, Commissioner of Bridges for New York City from 1902 to 1903.
At that time, Lindenthal was not yet familiar with the Deflection Theory in
suspension bridges. Because of his understanding that a suspension bridge is
nothing more than an inverted arch according to the Elastic Theory, Linden-
thal proposed an eyebar chain-cable suspension bridge for the Manhattan
Bridge, as shown in Fig. 4-8. Figure 4-9 is his illustration of the suspension
bridge as an inverted arch (Engineering News-Record, October 1, 1903).
As observed from these figures, Lindenthal sought to increase stiffness by
Figure 4-5. Brooklyn Bridge.

Source: Author.
Figure 4-6. Williamsburg Bridge.

Source: Author.
Figure 4-7. Manhattan Bridge.

Source: Courtesy of Bojidar Yanev.
Figure 4-8. Lindenthal’s unbuilt eyebar chain design for the Manhattan Bridge.
Source: Engineering News-Record, October 1, 1903.
bracing the cables instead of the stiffening truss. This probably reflects his
attachment to the idea of an inverted arch.
As described earlier, Moisseiff continued to develop the Deflection
Theory during this period, and in the end Lindenthal’s stiffened chain-cable
suspension bridge was not accepted. Thus, the Manhattan Bridge became
the first suspension bridge to be designed using the Deflection Theory, and
its construction began in 1904 according to Moisseiff’s design.
The Trend toward Longer Spans

The Manhattan Bridge proved to be a great success because of its design
adopting the new Deflection Theory. The Roebling-style designs as reflected
in the Brooklyn Bridge incorporated many experience-based factors. For
example, Roebling’s tower stays were determined empirically without cal-
culations because the complexity of their analysis surpassed the ability of
contemporary engineers; instead, Roebling relied solely on his experience-
based intuition. Therefore, tower stays were not favored by many engineers
and were rarely used in later suspension bridges (Moisseif 1935).
Figure 4-9. A suspension bridge as an inverted arch.

Source: Engineering News-Record, October 1, 1903.
The Elastic Theory, as used in the Williamsburg Bridge, lacked theo-

retical relevance for increasing span lengths; it was clear that the stiffen-
ing truss became too large, resulting in an aesthetically and economically
undesirable structure. Therefore, the Deflection Theory was a new tool that
could effectively solve this contradiction, and its power and utility were
demonstrated by the success of the Manhattan Bridge.
For the next half-century, steady progress was made in the span length
of suspension bridges in the United States (Table 4-1). The span record
established by the Williamsburg Bridge lasted for about 20 years, until it
was broken in 1924 by the Bear Mountain Bridge. Designed by Howard C.
Baird, this suspension bridge crosses the Hudson River about 55 km north
of the center of New York City.
Then, only 2 years later, this record was exceeded by the 533-m span
Delaware River Bridge. This bridge realized a century-long dream of the
citizens of Philadelphia and Camden, and because of the geographical simi-
larity of the site to Manhattan and Brooklyn, the bridge was called the
Second Brooklyn Bridge (Black 1936, 108).2 The design team for this bridge
consisted of Ralph Modjeski as chairman of the Board of Engineers, Leon
Moisseiff as engineer of design, Daniel E. Moran as consultant for the foun-
dations, and Paul P. Crét of the University of Pennsylvania as architect. This
type of specialist staff team subsequently became the standard approach for
long-span bridge projects.3
Table 4-1. Progress in Record-Span Suspension Bridges

Year Completed Bridge Name Span Length (m)
1883 Brooklyn 486.0
1903 Williamsburg 487.5
1909 Manhattana 450.0
1924 Bear Mountain 497.0
1926 Delaware River 533.0
1929 Ambassador 564.0
1931 George Washington 1,067.0
1936 San Francisco-Oakland Baya 704.0
1937 Golden Gate 1,280.0
1939 Bronx-Whitestonea 701.0
1940 Old Tacoma Narrowsa 853.0
1957 Mackinaca 1,158.0
1964 Verrazano-Narrows 1,298.0
aBridges so marked did not establish record span lengths.
This bridge, on which construction began in December 1922, was the

first suspension bridge to exceed 500 m in span. Its width was a generous
39 m, the weight of steel totaled 61,700 tonnes, and the cables alone weighed
7,000 tonnes. A distinguishing feature of this bridge was its main cables.
Preceding spans such as the Brooklyn Bridge, the Williamsburg Bridge, and
even the Manhattan Bridge featured two cables on each side, for a total of
four cables. However, in the Delaware River Bridge, only one cable per side
(thus, a total of two cables) was used. As a result, the cable diameter became
an unprecedented 76.2 cm (Figs. 4-10 and 4-11). Construction progressed
smoothly without interruption and the bridge was completed on schedule.
Figure 4-10. Main cable of the Delaware River Bridge.

Source: American Cable Company (1926).
Figure 4-11. Delaware River Bridge.

Source: American Cable Company (1926).
It was opened on July 1, 1926, 3 days before the 150th anniversary of Amer-
ican Independence. It was reported that construction accidents resulted in
13 casualties, including one young member of the engineering staff.
The Delaware River Bridge’s record span was exceeded only 3 years
later by the 564-m span Ambassador Bridge in 1929. Because this bridge
was built to connect border cities in two nations, Detroit in the United States
and Windsor in Canada, an American/Canadian joint-venture company was
formed to undertake the project. The design was handled by the construc-
tion company and Jonathan Jones was named as lead engineer. However,
Jones hired several consultants, including Moisseiff for the superstructure,
Daniel Moran for the substructure, and architect Cass Gilbert (1859–1934)
for aesthetic treatments. The bridge owner also retained Modjeski as its
consultant. One of the characteristics of this bridge is that both side spans
were not suspended, as shown in Fig. 4-12, and naturally this made the
bridge more economical.
The trend toward longer-span suspension bridges in the United States
reached its pinnacle in the 1930s when the 1,067-m-span George Washing-
ton Bridge doubled the existing span record, and leapt beyond the 1,000-m
Figure 4-12. Plan view of the Ambassador Bridge linking Detroit, Michigan, and
Windsor, Canada.
Source: Watson (1937).
span for the first time in human history. The building of a bridge across
the Hudson River as a western approach to Manhattan, the center of New
York, had been longed for since the city’s early days, and therefore several
plans had been proposed over the years. Some of them are shown in Fig.
4-13, including an arch bridge, a cantilever truss, and Lindenthal’s proposal
for a suspension bridge stiffened by braced chain cables. All these designs
reflect the engineers’ best efforts of the era. Finally, a plan for the suspension
bridge proposed by Othmar Ammann (1879–1965) was accepted (Ammann
1933), and its feasibility was in part established by the success of the Man-
hattan Bridge—the suspension bridge that had demonstrated the veracity of
the new Deflection Theory.
In 1921, the Port Authority of New York and New Jersey was founded
by the states of New York and New Jersey, with the crossing of the Hudson
River as a primary objective. A comparative study of tunnels and bridges
was performed at the outset, but a suspension bridge proposed by Ammann
became the favored plan. In 1927, the Authority appointed Ammann as its
bridge engineer, and on September 21 of that year it officially inaugurated
the construction of the George Washington Bridge. Listed as technical advi-
sors for what was perhaps the most significant bridge of the twentieth cen-
tury were well-regarded engineers such as Moisseiff for the superstructure,
Moran for the substructure, and architect Gilbert, together with Ammann’s
mentor Lindenthal.
One of the most significant characteristics to be noted concerning
the George Washington Bridge was that engineers Ammann and Moisseiff
rendered stiffening trusses unnecessary in the world’s longest suspension
bridge. Ammann explained the rationale as follows:
Extensive studies convinced the writer that for a long-span suspension

bridge a rigid system was not necessary. He was also familiar with the
fact that by the application of the correct or so-called deflection theory,
as distinguished from the ‘elastic theory,’ to a more or less flexible system,
Figure 4-13. Examples of bridge proposals to span the Hudson River, entering New
York City from the west.
Source: Ammann (1933).
material economies can be effected. This is inherently due to the stiffen-

ing effect of the dead load, which effect is ignored in the so-called elastic
theory. . . .
As a result of lengthy theoretical investigations, supplemented by
observations on mechanical models, made in an endeavor to find the
appropriate degree of rigidity of the stiffening trusses for the George
Washington Bridge, the writer came to the conclusion that the arrange-
ment of nearly flexible trusses in the finished bridge, and the omission of
trusses in the initial stage of a single highway deck, were perfectly per-
missible and would secure a degree of rigidity at least equivalent to that

of any of the aforementioned large modern bridges. (Ammann 1933)
The George Washington Bridge was initially designed as a double-

deck bridge, as shown in Fig. 4-14. With an 8.8-m truss depth and a corre-
sponding depth-to-span ratio of 1⬊120, the truss was much shallower than
that of the Williamsburg Bridge (a bridge of approximately half the span
with a 12.2-m deep truss and a depth-to-span ratio of 1⬊40). Therefore, it
is not as stiff, and this is why the George Washington Bridge is described as
“a semi-flexible truss.”
But Ammann and Moisseiff reached the conclusion that even this was
unnecessary, and only the upper deck to carry the initial stages of traffic was
built as shown in Fig. 4-15 and in the photo in Fig. 4-16. It stood as nothing
more than an unstiffened suspension bridge. The only attribute that made
this feasible was the stiffening effect of the dead load; as weight increased,
the stiffness of long-span suspension bridges was enhanced, thus rendering
a truss unnecessary, as Ammann and Moisseiff claimed.
Figure 4-14. Ammann’s and Moisseiff’s original design for the George Washington
Bridge truss.
Source: Ammann (1933).
Figure 4-15. As-built suspended structure of the George Washington Bridge in 1931.
Source: Dana et al. (1933).
As a matter of interest, a total of about 95,000 tonnes of steel were used

in the George Washington Bridge, and four cables 91.4 cm in diameter—
larger than those of the Delaware River Bridge—were used; they weighed
30,000 tonnes.
The views of Ammann and others were in this sense correct. Until the
George Washington Bridge was transformed into a double-deck structure in
1962 as per the original plans, the unstiffened suspension bridge had been
in service essentially problem-free for 30 years.
Two Large Projects in San Francisco

With the completion of the George Washington Bridge in 1931, engineers had
succeeded in building a 1,000-m span without any intermediate supports. The
suspension bridge as a structural type had made this achievement possible.
Figure 4-16. George Washington Bridge upon completion in 1931.

Source: Courtesy of the Port Authority of New York & New Jersey.
The success of the George Washington and Delaware River bridges on the
East coast proved a decisive influence over two major projects in Califor-
nia that had been pending for many years, the Golden Gate Bridge and the
San Francisco-Oakland Bay Bridge. Two huge projects, similar in scale to the
Akashi Kaikyo Bridge and the Seto Ohashi Bridge in Japan’s Honshu-Shikoku
Bridge project, were begun simultaneously in 1933 in the San Francisco Bay
area (Fig. 4-17) (Kawada 1990).
The first span was the exceptionally famous Golden Gate Bridge, to
which Joseph Baermann Strauss (1870–1938) as chief engineer devoted his
life. Since 1921, Strauss had planned a hybrid structure consisting of a can-
Figure 4-17. San Francisco Bay and the location of the Golden Gate Bridge and the
San Francisco-Oakland Bay Bridge.
Source: U.S. Steel (1936), with permission of American Bridge Co.
tilever truss and suspension bridge spanning 1,200 m, as shown in Fig. 4-18.
He obtained a patent for the concept and energetically campaigned for its
construction (Petroski 1995, 277). However, when it became apparent that
the George Washington Bridge would be completed successfully, Strauss

abandoned his patented concept and designed the crossing as an orthodox
suspension bridge. Figure 4-19 shows the new design, in which Ammann,
Moisseiff, and Derleth were deeply involved as the three members of the
Advisory Board of Engineers established by the Bridge District’s Board of
Directors. A peculiarity was its art deco-style tower design developed by
consulting architect Irving F. Morrow.
Construction of the Golden Gate Bridge began on January 5, 1933. The
most difficult and time-consuming task was the construction of the marine
foundation for the San Francisco tower pier. It was planned to build this foun-
dation 20 m under the water and about 340 m offshore, where the strait meets
the open waters of the Pacific Ocean. Fortunately, an extended sheet of rock
was relatively shallow, and therefore it was planned to install a trestle to the
pier site. However, because the actual rock surface had been polished by the
action of strong tides, it was extremely difficult to install the trestle on the
exposed rock. Then the original trestle was immediately destroyed by wave
action, but was rebuilt to a much more robust design.
Construction of the San Francisco tower foundation itself proved even
more difficult. The magnitude of the challenge was amply demonstrated by
Figure 4-18. Strauss’s original plan for the Golden Gate Bridge, a hybrid structure
consisting of a cantilever truss and suspension bridge.
Source: Derleth Collection, Water Resources Center Archives, University of California, Berkeley.
Reproduced with permission.
Figure 4-19. Strauss’s final design of the Golden Gate Bridge, which was strongly
influenced by Ammann and Moisseiff.
Source: Ito et al. (1999), courtesy of Tadaki Kawada.
the fact that its design underwent significant change four times. The original
plan was to float a steel caisson into a concrete fender that would serve as
a breakwater and provide protection from ship collision; the pier would
then be built after excavating to a depth of 33 m (Golden Gate Bridge and
Highway District 1970).
Construction progressed to the point that the 10,000-tonne steel cais-
son was towed to the site and pulled inside the fender. However, the mooring
ropes broke immediately because of direct exposure to surging waves from
the open sea that entered the fender. The caisson then swayed and struck
the fender, exposing it to danger as well. In the end, the enormous caisson
had to be pulled out of the fender and was scuttled in the open sea. Work to
close the fender proceeded so as to eliminate damage from surging waves,
and the design was subsequently changed to transform the fender itself into
a foundation. With several design changes to the fender, the bearing area
of the tower foundation increased to 2.4 times the size of the original plans
(van der Zee 1986, 210–213).
The construction of the San Francisco tower foundation and tower
was the most difficult challenge of this project. They required 2 years and
4 months to complete, which encompassed more than half of the span’s over-
all construction period of 4 years and 3 months. The remaining work was
completed in just under 2 years, including construction of the cables, stiffen-
ing trusses, and other tasks. On May 27, 1937, the 1,280-m span Golden
Gate Bridge was opened, and so began its reign of about a quarter-century as
the world’s longest span bridge (Fig. 4-20).
Figure 4-20. Golden Gate Bridge and bronze statue of Joseph B. Strauss.
Source: Author.
Although in span not equaling the Golden Gate Bridge, the San Fran-
cisco-Oakland Bay Bridge was unquestionably the world’s largest bridge
to date in terms of the scale of its construction. As shown in Fig. 4-21,
the bridge’s double-deck structure initially accommodated two interurban
tracks and nine road lanes. It now accommodates 10 road lanes. In addi-
tion to twin 704-m-span suspension bridges, this major bridge complex
includes a 164-m-long tunnel on Yerba Buena Island and a 427-m-span
cantilever truss. Its total length of 13.2 km, including the 8.1-km over-water
section and approaches, is no less impressive than Japan’s Seto Ohashi route
between Kojima and Sakaide in the Honshu-Shikoku Bridge project. The lat-
ter’s combined highway-railroad bridge features a 9.4-km-long over-water
section and a total length of 13.1 km when approaches are included. More-
over, the San Francisco-Oakland Bay Bridge was built before the Second
World War, a half-century earlier than the Japanese bridge.
For this bridge, Modjeski, Moisseiff, and Moran again served on the
Board of Consulting Engineers. The largest problem in this project was how
to connect the 3-km section between San Francisco and Yerba Buena Island.
With the state of engineering at that time, this distance was too great to
Figure 4-21. San Francisco-Oakland Bay Bridge, original plan and cross section.
envisage spanning with a single suspension bridge. However, the resulting

concept—to align two suspension bridges end-to-end—generated the new
problem of constructing a huge artificial island for a common anchorage
(W4 in Fig. 4-21). This artificial island was of an extraordinary scale and
had to be embedded in rock at the unprecedented depth of 65 m to resist the
several thousand tonnes of horizontal pull from the cables, which entered
the anchorage at a height of 84 m above sea level.
It was Daniel Moran who resolved this problem brilliantly. As a founda-
tion consultant, Moran had demonstrated his abilities in the George Wash-
ington and Delaware River bridges. From the standpoint of workability, the
pneumatic caisson method at that time was limited to depths of only 38 m. For
the San Francisco-Oakland Bay Bridge, Moran developed the concept of the
multiple-domed caisson that integrated the best features of both dredging and
pneumatic caissons. With this concept, global tilting could be controlled by
adjusting the buoyancy of each dome by means of compressed air (Fig. 4-22).
An assembly of 55 steel tube caissons, each 4.5 m in diameter and

assembled in a 5 ⫻ 11 arrangement to form a gigantic 27.6-m-long ⫻ 59.1-m-
wide block, was floated under its own buoyancy and towed to the site. After
positioning at the site, concrete was cast in openings between the steel tubes.
This gradually lowered the caisson, expelling air from the steel tubes in the
process (Fig. 4-23). As the large caisson sank, each steel tube was extended
upwards through the addition of more sections. Before the caisson reached
deep bedrock at a depth of 65 m, the caps of the steel tube caissons were
removed, allowing excavation to continue through layers of sedimentary
rock and mixed boulder-clay soils, which were dredged by grab buckets.
Frequently the caisson cutting edge struck boulders, causing the caisson to
tilt, and this required divers to enter the caisson to remove them by under-
water blasting.
Moran’s idea was successful. The cutting edge passed through the 30-m-
thick sedimentary layer and reached bedrock in June 1934, about 1 year after
Figure 4-22. Mock-up, multiple-dome caisson conceptualized by Daniel Moran.

Source: State of California (1934).
Figure 4-23. Construction procedure, multiple-dome caisson.

sinking had begun (State of California 1934). Yet even after the caisson had
settled on bedrock, other difficulties arose, such as leveling the rock and com-
pacting the tremie concrete. Eventually, however, the gigantic middle anchor-
age, with a rectangular cross section measuring 59.1 m longitudinally and
27.6 m transversely, and standing 65 m deep and 84.3 m above water level,
was completed (Fig. 4-24) (U.S. Steel 1936). This artificial island was created
using the largest caisson built to that time, and was proudly reported by engi-
neers to be larger than the pyramid for King Khufu, with more concrete than
was used in the Empire State Building.
The construction of the middle anchorage W4 was the most challeng-
ing aspect of building the San Francisco-Oakland Bay Bridge. Subsequent
work steadily progressed, and all construction was completed and the bridge
opened to traffic in November, 1936. With an amazingly short construction
period of only 3 years and 4 months, this immense project was now finished
(Fig. 4-25).
Figure 4-24. Middle anchorage W4 of San Francisco-Oakland Bay Bridge.

Figure 4-25. San Francisco-Oakland Bay Bridge, completed in late 1936.

Notes
1. It was F. F. Turneaure who actually undertook the analysis under Modjeski.
Turneaure (1911) contains his published results.

2. This bridge is also known as the Philadelphia-Camden Bridge, or the Camden
Bridge, but its official name (since 1955) is the Benjamin Franklin Bridge.
3. Both the Brooklyn Bridge and the Williamsburg Bridge, including their aesthetic
features, were designed by the chief engineer without the benefit of architectural
advice. Most critics hold that this explains why the Williamsburg Bridge was such
an unattractive design.
References
American Cable Company. (1926). The World’s Greatest Suspension Bridge—Phila-
delphia to Camden, The American Cable Company, Inc., Philadelphia.
Ammann, O. H. (1933). “George Washington Bridge: general conception and devel-
opment of design.” Trans. ASCE, Vol. 97, Paper No. 1818.
Bender, C. B. (1881). “Suspension bridges of any desired degree of stiffness.” Van
Nostrand’s Engineering Magazine, D. Van Nostrand, New York.
Black, A. (1936). The Story of Bridges, Whittlesey House, McGraw-Hill, New York
Castigliano, A. (1879). Théorie de l’Equilibre des Systèmes Élastiques et ses Applica-
tions, Turin.
Clericetti, C. (1880). “The theory of modern American suspension bridges.” Proc.,
Institute of Civil Engineers, London.
Dana, A., Andersen, A., and Rapp, G. M. (1933). “Design of superstructure.” Trans.
ASCE, Vol. 97, Paper No. 1820.
Finch, J. K. (1941). “Wind failures of suspension bridges or evolution and decay of
the stiffening truss.” Engineering News-Record, March 13.
Godard, M. T. (1894). “Recherche sur le calcul de la résistance des ponts suspen-
dus.” Annales des Ponts et Chaussées, Vol. 8.
Golden Gate Bridge and Highway District. (1970). The Golden Gate Bridge: Report
of the Chief Engineer to the Board of Directors of the Golden Gate Bridge and
Highway District, California, September 1937, Highway and Transportation
District, San Francisco.
Hopkins, H. J. (1970). A Span of Bridges: An Illustrated History, David & Charles,
Newton Abbot, U.K.
Ito, M., Kawada, T., et al. (1999). Opening the Age of Super Long-Span Suspension
Bridges: Further Challenge of Engineers [in Japanese], Kensetsu Tosho, Japan.
Kawada, T. (1969). Theory and Calculations of Long-Span Suspension Bridges [in
Japanese], Bridge Publication Committee, Japan.
Kawada, T. (1990). “Spanning over the San Francisco Bay.” Bridge and Foundation,
August.
Kemp, E. L. (1979). “Links in a chain: the development of suspension bridges

1801–70.” Structural Engineer, 57A(8).
Léby, M. (1886). “Mémoir sur le calcul des ponts suspendus rigides.” Annales des
Ponts et Chaussées.
Mélan, J. (1888). “Theorie der eisernen Bogenbrücken und der Hängebücken.” In

Handbuch der Ingenieuwissenschaten, 2nd ed., Wilhelm Engelmann, Leipzig,
Germany.
Moisseiff, L. S. (1925). “The towers, cables and stiffening trusses of the bridge
over the Delaware River between Philadelphia and Camden.” J. Franklin Inst.,
October.
Moisseiff, L. S. (1935). “Note to ‘A Generalized Deflection Theory for Suspension
Bridges,’ by D. B. Steinman.” Trans. ASCE, Vol. 100, 1205–1209.
Müller-Breslau, H. (1881). “Théorie der durch einen Balken versteiften Kette [Theory
of chains stiffened beams].” Zeitschrift d. A. I. zu Hanover.
Navier, L. H. (1823). Rapport et Mémoire sur les Ponts Suspendus, Imprimérie Royale,
Paris.
Petroski, H. (1995). Engineers of Dreams: Great Bridge Builders and the Spanning
of America, Vintage, New York, p. 277.
Rankine, W. J. M. (1858). Manual of Applied Mechanics, R. Griffin, London.
Ritter, W. (1877). Versteifungsfachewerke bei Bogen und Hängebrücken [Stiffened
Arch Bridges and Stiffened Suspension Bridges],” Zeitschrift für Bauwesen.
Shimada, S. (1964). Applied Mathematics for Civil Engineering [in Japanese], Kyoritsu
Shuppan, Japan
State of California. (1934). First Annual Progress Report San Francisco-Oakland
Bay Bridge, State of California, July 1.
Tanaka, H. (2000). “Collapse of the Quebec Bridge and its reconstruction.” Bridge
and Foundation, October.
Timoshenko, S. P. (1953). History of Strength of Materials, McGraw-Hill, New York.
Turneaure, F. F. (1911). “Exact methods of calculation.” In Theory and Practice of
Modern Framed Structures, 9th ed., Wiley, New York.
U.S. Steel. (1936). San Francisco–Oakland Bay Bridge, United States Steel Corp.,
Pittsburgh.
Watson, W. J. (1937). A Decade of Bridges: 1926–1936, J. H. Jansen, Cleveland, Ohio.
van der Zee, J. (1986). The Gate: The True Story of the Design and Construction of
the Golden Gate Bridge, Simon & Schuster, New York.
5
Ambush: The
Dynamics of Wind
Disaster of the Tacoma
Narrows Bridge
We have seen that the success of the George Washington
Bridge, a span that breached the 1,000-m span limit for the
first time ever, helped pave the way for an even longer-span
suspension bridge, the Golden Gate Bridge. However, the
contributions of the George Washington Bridge were not
limited to extended span length alone. This bridge also ver-
ified that stiffening trusses could be rendered unnecessary
in long-span suspension bridges because of their inherently
large self-weight.
With this result, for the 701-m-span Bronx-Whites-
tone Bridge, Othmar Ammann decided to adopt a plate
girder that had previously been used only in medium- and
small-span suspension bridges. Ammann describes this sus-
pension bridge, on which planning was inaugurated in 1935
and construction completed in April 1939, as follows:
It is now well established that long-span suspension

bridges for modern highway traffic may have a rela-
tively flexible stiffening system, and that the degree
of flexibility has a material effect upon the economy
of the design. In this respect the Bronx-Whitestone
Bridge marks another radical departure from past the-
ories and practice. Its stiffening girders have greater
flexibility in relation to span length than any other
suspension bridge built in recent years, except the
George Washington Bridge in its present state with
129
only a single unstiffened highway deck. The latter bridge, however, with
a present roadway capacity equivalent to that of the Bronx-Whitestone
Bridge, has a suspended dead weight per lineal foot of 2.5 times greater,
a center span 56 percent longer, and side spans somewhat shorter, all of
which factors contribute to greater rigidity of the unstiffened cables.
It was the aim . . . on esthetic as well as structural and economical
grounds, to restrict the height of the floor structure to a minimum, to
avoid trusses, and to keep the top at such an elevation above the floor as
not to obstruct the view of the landscape from passing vehicles. A depth
of 11 ft [3.35 m] for the stiffening girders was found to be sufficient and
fit best into the floor structure. This is only 1/210 of the length of the
center span and only 1/70 of the side span. (Ammann 1939)
Thus, in the Bronx-Whitestone Bridge, the trusses that James Finley

introduced to the world and which became a requisite for modern suspen-
sion bridges, disappeared. Instead, a plate girder was adopted as the stiffen-
ing structure. As the photo in Fig. 5-1 reveals, a suspension bridge stiffened
by plate girders is economical and also very light and graceful in appear-
ance. The new form was admired as the pinnacle of modern technology.
Figure 5-1. Bronx-Whitestone Bridge upon completion in 1939.

Source: Leonhardt (1982), with permission from Leonhardt, Andrä, und Partner.
ambush: the dynamics of wind 131
Following the introduction of the Deflection Theory, American engi-

neers extended the span length of suspension bridges and continued to
reduce the depth-to-span and width-to-span ratios of their stiffening trusses
and girders. Table 5-1 provides a summary of this progress.

With this trend, Leon Moisseiff, without hesitation, adopted a plate
girder for the Tacoma Narrows Bridge when he undertook to design what
would be the world’s third longest suspension bridge. Its span would be
about 20% longer than that of the Bronx-Whitestone Bridge but, because
of expected low traffic volumes, two lanes were considered sufficient and
the span’s width (the distance between the cables) was set at 11.9 m. As a
result, the design evolved to a suspension bridge of unprecedented slender-
ness, with the suspended structure featuring a depth-to-span ratio of 1:350
and a width-to-span ratio of 1:72. Moisseiff probably never imagined that
this would lead to the specter of future failure.
Construction of the Tacoma Narrows Bridge began in October 1938
and the bridge was completed in July 1940. Only 4 months after opening,
in November of the same year, the bridge was tossed about in a wind of
approximately 19 m/s (42 mph) and collapsed as the wind tore it to pieces
(see Fig. 5-2). A week following the accident, Engineering News-Record
reported on the scene, including vivid photographs:
Wind wrecked the 2,800-ft [853-m] main span of the Tacoma Narrows
suspension bridge shortly before noon Thursday, November 7. Almost the
entire suspended structure between the towers was ripped away and fell
into the waters below, but the 1,100-ft [346-m] side spans remained intact.
Cables and towers survived and held up the weight of the side spans,
though the latter sagged about 30 ft [10 m] as the towers, which are fixed
at the base by steel anchors deeply embedded in the concrete piers, were
bent back sharply by the unbalanced pull of the side-span cables.
Table 5-1 Depth- and Width-to-Span Ratios for Major Suspension

Bridges
Truss (Girder)
Year Span Length Depth-to-Span Width-to-
Completed Bridge Name (m) Ratio Span Ratio
1903 Williamsburg 487.5 1⬊40 1⬊14
1909 Manhattan 450 1⬊60 1⬊13
1931 George Washington 1,067 (0) 1⬊33
1937 Golden Gate 1,280 1⬊164 1⬊47
1939 Bronx-Whitestone 701 (1⬊209) 1⬊31
1940 Tacoma Narrows 853 (1⬊350) 1⬊72
(1) (5)
(2) (6)
(3) (7)
(4) (8)
Figure 5-2. Collapse of the Tacoma Narrows Bridge in November 1940.
Source: University of Washington Libraries, Special Collections: (1) negative UW21427; (2) nega-
tive UW2142c; (5) and (7) negative UW 26371; and (8) negative FAR026, with permission.
(3), (4), and (6) Tacoma Narrows Bridge film.
A wind reported as 42 mph [19 m/s] was blowing on the morning

of the accident. Higher winds had been experienced previously without
damage, but this wind caused vertical wave motion that developed a lag
or phase difference between opposite sides of the bridge, giving the deck
a cumulative rocking or side-to-side rolling motion.
Failure appeared to begin at mid-span with buckling of the stiff-
ening girders, although lateral bracing may have gone first. Suspenders
snapped and their ends jerked high in the air above the main cables,
while sections of the floor system several hundred feet in length fell
out successively, breaking up the roadway toward the towers until only
stubs remained. (ENR 1940a).
The Engineering News-Record further reported:
It was by far the most slender suspension bridge ever built, both later-
ally and vertically, the width of the roadway and spacing of cables being
only 39 ft (1⬊72 span ratio). And the depth of the stiffening girders was
8 ft (1⬊350 span ratio).
In general conception it followed the model set by the Whitestone
Bridge [2,300 ft or 701-m span], but was longer in span and much nar-
rower. To suit the expected traffic it was built for only two traffic lanes
against six on Whitestone and Golden Gate and an ultimate of twelve
on the George Washington Bridge.
The destructive wave movements that wrecked the bridge were
the sequels to a history of disturbing oscillation that had been under
observation and study since the bridge was built. Similar but much
smaller oscillation occurs on the Whitestone Bridge, but while there it
amounts to only a few inches, the vertical waves in the Tacoma Nar-
rows were very much greater, having reached as much as 50 in. [1.25-m]
amplitude. . . . (ENR 1940a)
The oscillation of the Tacoma Narrows Bridge seemed to be well

known following the bridge’s completion, and local people soon bestowed
the bridge with the honorable name of “Galloping Gertie” (Gertie being a
nickname for Gertrude). The actual vibration was carefully observed for the
4-month period of the bridge’s existence by F. B. Farquharson, an engineer-
ing professor at the University of Washington. The results of these observa-
tions are shown in Fig. 5-3. This figure shows all vertical bending modes,
and in fact the motion actually observed over this period was limited to the
vertical mode. The torsional mode appeared suddenly just before the bridge
failure (Farquharson 1949).

Figure 5-3. Vertical bending modes observed in the Tacoma Narrows Bridge.
Source: University of Washington Libraries, Special Collections, with permission.
The Dynamics of Wind

As the vibration became problematic following the completion of the
Tacoma Narrows Bridge, several studies continued in the 4 months before

the span’s failure, to determine its cause and to devise possible counter-
measures. One of these studies involved a series of wind tunnel tests per-
formed by Farquharson at the University of Washington. The results clearly
indicated that the cross section of the Tacoma Narrows Bridge was very
unstable with respect to wind.
The top left part of Fig. 5-4(A) depicts the experimental results of a
modeled cross section of the Tacoma Narrows Bridge. It reveals that the coef-
ficient of lift, where the angle of attack changes from ⫺10 degrees to ⫹10
Figure 5-4. Results of wind tunnel tests performed by F. B. Farquharson in the

summer of 1940 to assess vibration in the Tacoma Narrows Bridge.
The angle of attack, _, is positive for an upward wind against the horizontal deck and negative
for a downward wind.
degrees, varied between the positive and the negative as in a sinusoidal curve.
Here, the rate of change in the angle of attack at around 0 degrees is suspi-
cious and reveals an aerodynamic instability. This was the main culprit behind
the continued vibration in the Tacoma Narrows Bridge over its 4-month exis-
tence. It was evident that the energy of the vibration was attributable to vorti-
ces emanating from the edge of the girder. This phenomenon is known as the
Kármán vortex. In other words, the Tacoma Narrows Bridge was so flexible
that it suffered oscillations under the action of Kármán vortices.
In terms of countermeasures, one approach was to increase the stiff-
ness of the suspension bridge. However, this approach proved difficult to
implement, so other measures were considered. These included reducing the
generation of vortices and weakening the forces they created by making
the vortices smaller. The methods are listed below, and Fig. 5-4 shows the
results of wind tunnel testing (Bowers 1940).
1. Install a fairing to cover the outside of the girder to smooth out the
flow of air [Fig. 5-4(B)].
2. Install a vane discontinuously along the outside of the girder to
reduce the scale of generated vortices [Fig. 5-4(C,D)].
3. Punch holes in the girder web to smooth out the flow of air, thus
making the girder more open, like a truss [Fig. 5-4(E,F)].
It became clear that all the methods were effective. However, even
with holes punched through the girder web, some instability remained when
the openness ratio was as low as 15% (E). It was found that two rows of
holes in the girder with a 25% openness ratio, as denoted in (F), proved
effective. In contrast, one row of larger holes with the same openness ratio
(25%) did not improve stability.
The measure that was eventually planned for installation in the Tacoma
Narrows Bridge is shown in Fig. 5-4(D), and was chosen for reasons of aes-
thetics and economy. The vane was made from wood and looked like an aile-
ron. It was about 48 cm high and was rounded with about a 32-cm radius.
Engineers planned to attach the fairing 1.9 m from the stiffening girder
using steel brackets. After all preparatory work had been completed, work
was scheduled to start on November 7, the day the bridge collapsed. The
bridge had failed due to violent torsional vibrations never before observed
in it, and because no one could have foreseen the emergence of torsional
vibration in the bridge, engineers had no way to identify this mechanism of
failure in wind tunnel testing using sectional models.
Therefore, the world’s first full-span model wind tunnel facility was
hurriedly built at the University of Washington. The research results from
this wind tunnel clarified the vibration of the suspension bridge by wind and
the nature of wind dynamics. In a report, Farquharson indicated two types
of vibration in suspension bridges, with the results shown in Figs. 5-5 and
5-6 (Farquharson 1952).
Limited Vibration
In this mode, vibration increases in amplitude with increasing wind
speeds. However, when the amplitude reaches a certain magnitude, it dimin-
ishes and finally disappears. Vibration modes 0-NV to 4-NV shown in Fig.
5-5 are forms of vertical vibration. Moreover, all vibration modes shown in
Fig. 5-3 are also vertical in nature and are therefore self-limiting.
Catastrophic Vibration
As wind speed increases, in this mode of vibration—once it occurs—
the amplitude rapidly diverges without convergence and finally develops
Figure 5-5. Wind speeds vs. vibration in wind tunnel testing of the full-span model.
0-NV, vertical bending vibration with no node; 1-NT, 1-node torsional vibration; 1-NV, 1-node
vertical bending vibration; 2-NV, 2-node vertical bending vibration; 3-NV, 3-node vertical
bending vibration; 4-NV, 4-node vertical bending vibration.
Figure 5-6. Role of the center diagonal stays.

Source: Courtesy of Tadaki Kawada.
to the point of failure. The torsional vibration noted as 1-NT(2ND) in Fig.

5-5 is the only one of this kind of catastrophic vibration. The asymmet-
ric torsional mode that appeared in the last moments of the Tacoma Nar-
rows Bridge and caused its failure was this catastrophic 1-NT(2ND) mode.
According to these wind tunnel test results, a catastrophic vibration noted
as 1-NT(2ND) occurred at a reduced wind speed of 0.91 m/s (2 mph). This
mode of vibration had not appeared during lower wind speeds. This wind
speed for the Tacoma Narrow Bridge, which corresponds to 19 m/s (42
mph) in the prototype Tacoma Narrows Bridge, proved to be a dangerous
point of departure and is therefore called the critical wind velocity.
Clearly, there is no emergence of catastrophic vibration below the
critical wind velocity. By contrast, in some cases catastrophic vibration may
not appear if structural restraint is provided even if wind speeds reach the
critical wind velocity. Instead, severe vertical vibrations of other modes with
similar frequencies may appear. The Engineering News-Record stated that
in the Tacoma Narrows Bridge, “Even during strong winds there was no
report of (torsional vibration)” (ENR 1940a).
The Tacoma Narrows Bridge had experienced winds exceeding the
critical wind velocity, but because a catastrophic vibration had not emerged
for a period of 4 months, it was felt that the span exhibited some type of
structural restraint. It appeared that this restraint was probably provided by
the center diagonal stays that tied the cables to the stiffening girder at the
center span (Fig. 5-6). The evidence revealed that the most frequent vibra-
tions observed in the Tacoma Narrows Bridge involved symmetric modes, as
noted by types 0 and 2 in Fig. 5-3. The general asymmetric vibration observed
in the Bronx-Whitestone and elsewhere—spans that did not provide center
diagonal stays—did not appear in the Tacoma Narrows Bridge (before the
center diagonal stays failed). It is therefore obvious that the center diagonal
stays had an effect in restraining vibration. These center diagonal stays were
subjected to repetitive reversible stresses for a period of 4 months. When
observed at around 9 o’clock in the morning on the day of the collapse,
and verified in a 16-mm film recording the situation, they were found to be
alternately slackened and tightened because of the bridge’s repeated upward
Figure 5-7. Center diagonal stays just before the collapse.

Source: University of Washington Libraries, Special Collections, negative UW26371, with
permission.
and downward motion. Figure 5-7 is a still frame from the film and Fig. 5-8
shows the failed center diagonal stays following the bridge’s collapse.
The repetitive stresses over this 4-month period must have caused a
fatigue failure of the center diagonal stays. As a result, the restraining force
that had prevented the onset of torsional vibration over the 4 months dis-
appeared. The catastrophic vibration emerged immediately following the
failure of the center diagonal stays (Kawada 1975, 55–60).
Figure 5-8. Center diagonal stays after the collapse.

Inevitable Accident or Human Error?

Following the failure of the Tacoma Narrows Bridge, the federal govern-
ment immediately created a board of inquiry to investigate the accident.
This board consisted of three eminent engineers headed by Ammann, who
was considered the leading bridge engineer of the day, particularly in the
field of suspension bridge engineering. The other two members were Theo-
dore von Kármán, a professor of aerodynamics at the California Institute
of Technology and an authority in fluid mechanics, and whose name was

associated with the Kármán vortex, and Glenn B. Woodruff, the designer of
the San Francisco-Oakland Bay Bridge superstructure.
The investigation board’s report was quickly compiled in the 4 months
following the collapse, but was voluminous and comprehensive (Ammann
et al. 1941). The report was adeptly written, providing first a summary of
conclusions and a preface, followed by the chapters as follows:
Chapter 1: History, organization, design, construction and behav-

ior of the bridge
Chapter 2: Review of the design of the Tacoma Narrows Bridge
Chapter 3: Comparison of the Tacoma Narrows Bridge with other
suspension bridges
Chapter 4: Aerodynamic forces acting on suspension bridges and
the oscillations resulting therefrom
Chapter 5: General conclusions
Appendices: (about 140 pages in total)
The Appendices covered several interesting topics, including designer

Moisseiff’s letter that introduced the design change to the plate-girder from
the truss for stiffening; testimony by witnesses to the accident; a theoretical
analysis of the vibration of suspension bridges; and results of wind tunnel
testing. The first two of the board’s 16 conclusions were:
1. The Tacoma Narrows Bridge was well designed and built to safely
resist all static forces, including wind, usually considered in the design
of similar structures. Its failure resulted from excessive oscillations
caused by wind action.
2. The excessive vertical and torsional oscillations were made possible
by the extraordinary degree of flexibility of the structure and by its
relatively small capacity to absorb dynamic force. It was not realized
that the aerodynamic forces which had proven disastrous in the past
to much lighter and shorter flexible suspension bridges would affect a
structure of such magnitude as the Tacoma Narrows Bridge, although
its flexibility was significantly greater than any other long suspension
bridge. (Ammann et al. 1941)
These first two conclusions were in a sense the most meaningful of

all that were presented in the report. They echoed an editorial published
in Engineering News-Record 2 weeks after the accident, which influenced

public opinion at that time and finally served to absolve Moisseiff of respon-
sibility for the design error:
Incompetence or neglect cannot be blamed for the Tacoma Narrows

collapse, the greatest bridge failure in history. Acknowledged leaders
in bridge engineering had built the structure, and both design and con-
struction were carried out with conscientious care. The collapse resulted
from actions and forces previously ignored or known to be unimportant
that became dominant in the new range of magnitude and proportions
of this exceptional bridge. (ENR 1940b)
This is how Moisseiff was held to be blameless. However, he was deeply

hurt by the collapse of the Tacoma Narrows Bridge, and within 3 years he
died. Ammann’s obituary marking his death stated:
The one great disappointment in Mr. Moisseiff’s career was the failure
of the Tacoma Narrows Bridge, the design of which he had originated
and guided. It would be improper for his fellow professionals to put the
blame for that failure entirely upon Mr. Moisseiff’s shoulders, for he
followed a trend in long span suspension bridge design which appeared
justified at the time. (Amman 1943)
Here Ammann’s true feelings were revealed. Ammann practiced in

the same era as Moisseiff and he, too, followed the trends in the theory
of long-span suspension bridges prevailing at the time. Therefore, when
I reread the report of the investigative committee into the failure I am
tempted to think that it seemed designed to establish Moisseiff’s innocence.
For example, a negative comment as described below was made regarding
the center diagonal stays installed as an emergency measure following the
bridge’s completion, but it was again accompanied by a weak excuse that
exonerated Moisseiff:
On the Tacoma Narrows Bridge it was apparently less marked. . . .

From these experiences the conclusion may be drawn that central stay
ropes, when properly designed, exercise a limited damping effect on
vertical oscillations, but they are largely ineffective when loose and may
be a source of danger when the motions of the cables become large.
(Ammann et al. 1941, 55–57)
On December 5, 1940, while Ammann and his colleagues were taking

great pains in their report to vindicate Moisseiff, the Engineering News-
Record published an article on two other suspension bridges stabilized by

cable stays:
Occurrence of disturbing vertical movement or undulations in suspen-

sion bridges of slender proportions antedated the recent experiences at
Whitestone and Tacoma by more than a year. At both the Thousand
Islands Bridges over the St. Lawrence River in New York, and the Deer
Isle Bridge in Maine, marked undulation was observed when the bridges
neared completion. The engineers for the Thousand Islands Bridge, Rob-
inson and Steinman, developed a successful control system consisting of
diagonal rope stays from roadway level at the towers to the cables of the
main span, and a year later, in June, 1939 applied the same system to
the Deer Isle Bridge. In both cases the undesirable undulation has been
brought under complete control. (ENR 1940c)
The article continued as follows and provided the diagrams shown in

Figs. 5-9 and 5-10 that highlight the configuration of the stay cables:
Two suspension spans form the principal elements of the two-lane high-
way bridge across the St. Lawrence. The American crossing has a main
span of 800 ft [244 m] and two 350 ft [107-m] side spans with deck
width and cable spacing of 30 ft 6 in. [9.3 m]. . . .
Undulatory movement was first noticed in June 1938, when the
framework of the suspended structure was complete and the I-Beam-
Lok grid of the floor slab, with attached steel form plates, was being put
down. It was then thought that the addition of the floor concrete would
steady the bridge, but this proved not to be the case. . . .
Figure 5-9. Thousand Islands Bridge (American crossing) over the Saint Lawrence
River, connecting northern New York in the United States with southeastern Ontario
in Canada.
Source: Ohashi (2003), with permission from Harukazu Ohashi.
Figure 5-10. Deer Isle Bridge spanning the Eggemoggin Reach in Maine.
Source: Ohashi (2003), with permission from Harukazu Ohashi.
The first effort at control consisted of installing temporary wire-

rope diagonal guys between the central cable band and the adjacent
suspender connections at the top of the stiffening girder. These ties elim-
inated the longitudinal swing of the span and reduced the undulation,
changing it into a motion with node at center—one quarter point of the
span going up while the other quarter point went down. . . .
For further control various methods of diagonal bracing by stay
ropes were studied with a conclusion that a set of stays anchored to
the ends of the stiffening girders and fanning out to connect with cable
bands on the main span cables was best. . . . In consequence a tempo-
rary stay system of this type was installed, two days before the formal
opening and dedication of the structure (Aug. 18, 1938). . . .
The stays comprised two diagonals at each corner of the main
span, each consisting of two parts of 3⁄4-in. [19 mm] steel hoist rope with
steamboat ratchet for adjustment. They ran to the fourth and sixth cable
bands, and were pulled up to about 5,000 lb. [2.2 tonnes] tension. . . .
After the opening of the bridge certain winds still caused motion,
the amplitude now being about 6 in. [15 cm]. The permanent installa-
tion was therefore designed with three stays instead of two, connected to
the third, fifth and seventh cable ends, each placed under 10,000 lb. [4.5
tonnes] tension. Each stay was made of a 11/8 in. [28.6 mm] prestressed
galvanized strand pinned to the stiffening girder and a cable band. . . .
Since this installation (Jan. 24, 1939) there has been no report of
undulation, and it is believed that if any oscillatory movements remain
they are imperceptible. (ENR 1940c)
As can be seen in Fig. 5-11, the Thousand Islands Bridge featured

slender proportions as did the Tacoma Narrows Bridge, and it looks as if
Figure 5-11. Thousand Islands Bridge (American crossing).

Source: Ratigan (1959), with permission from William B. Eerdmans.
it would soon oscillate. It is said that the vibration was adroitly controlled
through the use of the center tie and cable stays. On the other hand, the
Deer Isle Bridge proved more problematic:
While the Thousand Islands problem was being worked out in the field,
construction of another suspension bridge of slender proportions was
under way at Deer Isle, Maine, under the same engineers. Since it had
longer span, narrower deck and greater camber than the former struc-
ture, it was expected to develop oscillation. . . . (ENR 1940c)
In fact, the Deer Isle Bridge had begun to oscillate during construction. At
this time, stays were installed over both the main and side spans. By tension-
ing each one to 4.5 tonnes, the vibration could be completely controlled.
Clearly, the article in Engineering News-Record brought Ammann’s
efforts to vindicate Moisseiff almost to naught. Ammann had tried to prove
that the collapse of the Tacoma Narrows Bridge was inevitable by stating
that, given the state of knowledge of bridge engineers at that time, wind
dynamics had never been considered. Yet the ENR article clearly stated—as
if to antagonize Ammann—that the vibration in suspension bridges had
been known for a year before the accident and, furthermore, could have
been prevented. I wonder how Ammann rationalized his thinking, but he
dismissed the article without hesitation in the passage below, which clearly
captures his thinking:
The oscillations of the Thousand Islands and Deer Isle bridges and the
corrective installations are described in an article in Engineering News-
Record of December 5, 1940, entitled Two Suspension Bridges Stabi-
lized by Cable Stays. . . .
These bridges are more nearly comparable in size to some of
the early flexible suspension bridges. They give no clue to the possible
behavior of a suspension bridge 3½ times longer and 6 times heavier.
(Ammann et al. 1941, 73)
Aftermath in America following the

Tacoma Narrows Bridge Disaster
The failure of the Tacoma Narrows Bridge influenced subsequent bridge
engineering worldwide and, with this, the science of bridge aerodynamics
was born (Hirai 1956, 371–415; Kawada 2000, 25–32).
The first and most direct result of this new emphasis on aerodynamics
was found in the newly reconstructed Tacoma Narrows Bridge, completed in
1950. The foundations of the original Tacoma Narrows Bridge were reused
after strengthening. Its 853-m main span was the same as the former bridge,
although the suspended structure was completely changed. As shown in Fig.
5-12, the old bridge’s stiffening girder was changed to a stiffening truss with
a higher degree of torsional stiffness. The deck featured gaps of open grat-
ing, and three sets of center diagonal stays were provided.
The changes can be seen in Fig. 5-12(top) and (middle), both drawn to
the same scale. The depth-to-span ratio was increased to 1⬊85 as compared
to 1⬊350 for the old bridge, and the width-to-span ratio was also improved to
1⬊47 for the new bridge against 1⬊72 for the old bridge. Naturally, the span’s
weight was increased; the cables’ carrying weight was 18,160 tonnes for the
new bridge versus 11,250 tonnes for the old bridge, a 1.6-fold increase. The
weight of the cables for the new bridge was increased to 5,441 tonnes from the
3,817 tonnes in the old bridge, making them about 1.4 times heavier. Thus, in
total, the new bridge was about 50% heavier than the former span (Gotchy
1990). The reconstructed Tacoma Narrows Bridge has been in use for more
than half a century and no wind-related damage has ever been reported. This
is evidence that the aerodynamic measures have functioned well.
Figure 5-12. Aerodynamic measures implemented on the new Tacoma Narrows

Bridge.
(top) Cross section of old Tacoma Narrows Bridge; (middle) Cross section of new Tacoma Nar-
rows Bridge; (bottom) Stiffening truss at midspan of new Tacoma Narrows Bridge.
The next span deserving recognition for its attention to aerodynam-

ics is the 1,158-m span Mackinac Bridge, built between 1954 and 1957
to connect the noncontiguous upper and lower peninsulas of the state of
Michigan. Moisseiff actually undertook an early design for this gigantic sus-
pension bridge while the Tacoma Narrows Bridge was under construction.
For this design Moisseiff adopted a shallow plate-girder suspension bridge
similar to the Tacoma Narrows Bridge (Gies 1963).
Fortunate or not, the collapse of the Tacoma Narrows Bridge and
American involvement in the Second World War had the effect of postpon-
ing construction of this bridge until the end of the war. In 1950, the proj-
ect was reinitiated. Having been “baptized” by the failure of the Tacoma
Narrows Bridge, Moisseiff’s design had to be completely reevaluated. The
selection of committee members to examine the bridge design was entrusted
to Ivan C. Crawford, dean of the University of Michigan School of Engi-
neering. The leading authority in suspension bridge design at that time was
unquestionably Ammann, who, having built the first suspension bridge to
exceed 1,000 m in span and having led the investigation into the Tacoma
Narrows Bridge failure, stood at the pinnacle of the field. On the other
hand, David B. (“D. B.”) Steinman, who had gained a solid reputation by
publishing a paper entitled “Rigidity and Aerodynamic Stability of Suspen-
sion Bridges” (Steinman 1943), was strongly grounded in aerodynamic
theory. Steinman was rather more clearly insistent concerning his theories of
wind and vibration in suspension bridges. After struggling with the decision,
Crawford finally selected both Ammann and Steinman as committee mem-
bers. Another member was Glenn Woodruff, the superstructure designer
of the San Francisco-Oakland Bay Bridge and an influential engineer who
had worked cooperatively with Ammann on the committee investigating the
Tacoma Narrows Bridge failure.
The committee now had two giant, influential figures, each taking
different positions. According to rumors, the committee soon became a
forum in which Ammann and Steinman argued against each other openly,
with Woodruff acting as a referee. Ammann soon had enough and quit
the committee; as a result, the resolution of all engineering problems was
handed to Steinman. In January 1953, he was officially entrusted with
the design of the Mackinac Bridge. Woodruff later joined Steinman as an
associate.
Given this background, Steinman’s Mackinac Bridge—in Fig. 5-13,
showing comparisons with other suspension bridges built to that time—did
not approach the Golden Gate Bridge in terms of center span length. How-
ever, it did become the world’s second longest suspension bridge, exceed-
ing the George Washington Bridge. And if the entire suspended structure
between anchorages is compared, its 2,545-m length surpassed the Golden
Figure 5-13. Comparison of suspension bridges in terms of cable-supported span

length.
Source: Steinman and Nevill (1957), with permission from William B. Eerdmans.
Gate Bridge’s 1,966-m distance, making the Mackinac Bridge unequivocally

the world’s longest suspension bridge (Steinman and Nevil 1957).
Steinman explained the wind considerations actualized in this design
and the results as follows:
By utilizing all of the new knowledge of suspension bridge aerodynam-

ics, particularly my own mathematical and scientific discoveries and
inventions, I have made the Mackinac Bridge the most stable suspension
bridge, aerodynamically, that has ever been designed.
This result has been achieved not by spending millions of dollars
to build up the structure in weight and stiffness to resist the effects,
but by designing the cross section of the span to eliminate the cause of
aerodynamic instability. The vertical and torsional aerodynamic forces
which tend to produce oscillations have simply been eliminated.
The basic feature of this high degree of aerodynamic stability is
the provision of wide open spaces [through 76.2-cm stringers] between
the stiffening trusses and the outer edges of the roadway [refer to Fig.
5-14(top)]. The trusses are spaced 68 ft [20.7 m] apart, but the roadway
is only 48 ft [16.5 m] wide. This leaves open spaces 10 ft [3 m] wide on
each side, for the full length of the three suspension spans.
When you drive over the three suspension spans in the central
portion of the bridge, you can see the stiffening trusses beyond the edges
of the roadway, with the open spaces between them. These open areas
constitute the scientific design. They eliminate the closed corners in
which pressure concentrations are producible by wind, and they also
eliminate the solid areas on which such pressure differences would oth-
erwise act to produce oscillations of the span.
By this feature alone (the open spaces between roadway and stiff-
ening trusses), the critical wind velocity (the wind velocity at which
potential oscillations can start) was increased from 40 miles per hour
[18 m/s] to 632 miles per hour [282 m/s]!
But I was not satisfied with raising the critical velocity to this fabu-
lous figure. For still further perfection of the aerodynamic stability, I
provided the equivalent of a wide opening in the middle of the roadway
on the suspension spans. The two outer lanes each 12 ft [3.66 m] wide, are
made solid, and the two inner lanes and the central mall, 24 ft [7.1 m]
wide together, are made of open-grid construction of the safest, most
improved type. By this additional feature of aerodynamic design—add-
ing the central opening to the lateral openings previously described—I
achieved a further increase in aerodynamic stability and raised the criti-
cal wind velocity from 632 miles per hour [282 m/s] to a critical wind
velocity of infinity! (Steinman 1958, 15–19)
The Mackinac Bridge also featured a robust center tie, as shown in Fig. 5-14
(bottom), which fully incorporated the latest findings in wind engineering.
A comparison between the George Washington and Mackinac bridges
clearly demonstrates their differences. Figures 5-15 and 5-16 show the scaled
plans for the stiffening trusses and the towers. The four main cables (two
on each side), the framing of the stiffening trusses, and the placement of the
roadway in proximity to the laterals—all these features reflect clear differ-
ences from those of the Mackinac Bridge (Kawada 1999). The towers of the
George Washington Bridge and the Verrazano-Narrows Bridge (the latter is
discussed in detail below) are essentially the same, whether truss construc-
tion or not, and also resembles those of the Bronx-Whitestone Bridge.
The differences between the conservatively designed suspension
bridges and those designed with up-to-date aerodynamic measures are
shown in Table 5-2. Most notable is the huge difference in economy. It may
not be reasonable to simply compare these bridges because of the differ-
Figure 5-14. Measures taken to improve aerodynamic performance of the Mackinac

Bridge.
(top) cross section; (bottom) robust center tie.
Sources: (top) Steinman (1959); (bottom) Joyce (1959).
Figure 5-15. Comparison of stiffening truss cross sections in the George Washington
Bridge, the Verrazano-Narrows Bridge, and the Mackinac Bridge.

Figure 5-16. Comparison of towers of the George Washington Bridge, the Verrazano-Narrows Bridge, and the Mackinac
Bridge.
Table 5-2. Comparison of Major Suspension Bridges Built in the

United States
George Verrazano-
Washington Golden Gate Mackinac Narrows

Year completed 1931 1937 1957 1964
Main span length (m) 1,067 1,280 1,158 1,298
Total span length (m) 1,402 1,823 2,274 1,966
Deck width (cable 32.3 27.4 20.7 31.4
spacing) (m)
Number of lanes 14 6 4 12
Height of tower (m) 181.4 227.4 168.2 210.3
Cable diameter (mm) 914 ⫻ 4 914 ⫻ 2 622 ⫻ 2 904 ⫻ 4
and number of cables
Cable weight (tonnes) 30,200 22,100 11,300 37,800
Total structural steel 94,600 83,600 38,100 132,100
weight (tonnes)
ence in the number of lanes. Even so, in terms of both the total steel weight
and cable weight, the Mackinac Bridge is less than one-third the weight of
the Verrazano-Narrows Bridge. Conversely, the Verrazano-Narrows Bridge
consumed three times more structural steel than the Mackinac Bridge (Figs.
5-17 and 5-18). Another interesting disparity to be pointed out here is found
in the center ties or center diagonal stays. These were installed in the recon-
structed Tacoma Narrows Bridge and the Mackinac Bridge but not in the
Verrazano-Narrows Bridge.
It is interesting to learn how actual bridge designs were affected by
aerodynamics. The Verrazano-Narrows Bridge offers a good example. For
many years Steinman had proposed a suspension bridge across the mouth
of New York Bay. From the mid-1920s Steinman had named this cross-
ing the Liberty Bridge; in campaigning to secure the design commission, he
described this span as his lifework (Kawada 1998).
However, in the mid-1950s, when momentum for building the bridge
acquired tangible form, the work was entrusted to Ammann. Steinman found
it difficult to convince the client, the Triborough Bridge and Tunnel Author-
ity, otherwise. Ammann had been the Authority’s chief engineer between
1934 and 1939 for that Authority and had designed the Bronx-Whitestone
Bridge. However, for some reason, in designing the world’s longest suspen-
Figure 5-17. Mackinac Bridge.

Source: Author.
sion bridge Ammann ignored the latest findings in bridge aerodynamics.

Where Steinman attempted to apply the lessons of the failure of the Tacoma
Narrows Bridge through the creation of aerodynamically favored cross sec-
tions and the use of center ties to improve aerodynamic stability, Ammann
insisted (or could not help insisting) that mass is everything and therefore
simply extrapolated his experience with the George Washington Bridge. He
designed the Verrazano-Narrows Bridge as another version of his George
Washington Bridge, built more than 30 years earlier (Ammann 1966).
In any event, construction of this record-breaking suspension bridge
with a 1,298-m center span began soon after the completion of the Macki-
nac Bridge, and the gigantic Verrazano-Narrows Bridge was completed in
1964. This bridge reigned as the world’s longest-span suspension bridge for
about a quarter-century and was the last in the United States to hold this
position. It was not only the world’s longest suspension at that time, but
also remains the heaviest suspension bridge ever built. This record as the
heaviest span will never be surpassed, for bridge engineers thereafter used
aerodynamics as a tool and, with the demand for increasingly economical
structures, would no longer rely solely on weight, as did Ammann.
Figure 5-18. Verrazano-Narrows Bridge.

Source: Author.
References
Ammann, O. H. (1939). “Planning and design of Bronx-Whitestone Bridge.” Civil
Engineering, April.
Ammann, O. H., von Kármán, T., and Woodruff, G. B. (1941). The Failure of the
Tacoma Narrows Bridge: A Report to the Honorable John M. Carmody,
Administrator, Federal Works Agency, Washington D.C., Report by the Board
of Engineers, March 28.
Ammann, O. H. (1943). “Obituary: Leon S. Mossieff.” Engineering News-Record,
September 9.
Ammann, O. H. (1966). “Verrazano-Narrows Bridge: conception of design and
construction procedure.” Proc., ASCE, March.
Bowers, N. A. (1940). “Model tests showed aerodynamic instability of Tacoma
Narrows Bridge.” Engineering News-Record, November 21.
Engineering News-Record (ENR). (1940a). “Tacoma Narrows Bridge wrecked by
wind.” Engineering News-Record, November 14.
ENR. (1940b). “Dynamic wind destruction.” [Editorial] Engineering News-Record,
November 21.
ENR. (1940c). “Two recent bridges stabilized by cable stays.” Engineering News-
Record, December 5.
Farquharson, F. B. (1949). Aerodynamic Stability of Suspension Bridges with

Special Reference to the Tacoma Narrows Bridge. Part I: Investigation Prior
to October, 1949, University of Washington, Engineering Experiment Station,
Bulletin No. 116, Part I.
Farquharson, F. B. (1952). Aerodynamic Stability of Suspension Bridges with Special

Reference to the Tacoma Narrows Bridge. Part III: The Investigation of the
Original Tacoma Narrows Bridge under the Action of Wind, University of
Washington, Engineering Experiment Station, Bulletin No. 116, Part III,
June.
Gies, J. (1963). Bridges and Men, Doubleday, New York.
Gotchy, J. (1990). Bridging the Narrows, Gig Harbor Peninsula Historical Society,
Gig Harbor, Wash.
Hirai, A. (1956). “Failure of the Tacoma Narrows Bridge and its lesson.” [in
Japanese] In Steel Bridge III, Chapter 12, Gihodo, Japan.
Joyce, W. E. (1959). “Mackinac Bridge: longest bridge cables ever constructed.”
ASCE Civil Eng., 29(1), 54–57.
Kawada, T. (1975). Who Wrecked Galloping Gertie? The Mystery of the Tacoma
Narrows Bridge Disaster [in Japanese], Kensetsu Tosho, Tokyo, Japan.
Kawada, T. (1987). Modern Suspension Bridges, Riko Tosho, Japan.
Kawada, T. (1998). “People contributed to the development of steel structures (18):
Ammann and Steinman (1).” Japan Society for Steel Construction, No. 30.
Kawada, T. (1999). “People contributed to the development of steel structures (19):
Ammann and Steinman (2).” Japan Society for Steel Construction, No. 31.
Kawada, T. (2000). “Failure of the Tacoma Narrows Bridge: aerodynamics in its
infancy.” [in Japanese] Assoc. Japan Wind Eng., No. 83.
Leonhardt, F. (1982). Brücken: Ästhetic und Gestaltung, Leonhardt, Andrä, und
Partner, Berlin.
Ohashi, H. (2003). “Bridges in America: Thousand Island Bridge.” [in Japanese]
Civil Engineering Journal, Japan, 44(2), February.
Ratigan, W. (1959). Highways over Broad Waters: Life and Times of David Steinman,
Bridgebuilder, Eerdmans Publishing, Grand Rapids, Mich.
Steinman, D. B. (1943). “Rigidity and aerodynamic stability of suspension bridges.”
Proc., ASCE, November.
Steinman, D. B. (1958). “The design of the Mackinac Bridge.” In Mighty Mac: The
Official Picture History of the Mackinac Bridge, L. A. Rubin, ed., Wayne State
University Press, Detroit, Mich.
Steinman, D. B., and Gronquist, C. H. (1959). “Mackinac Bridge: superstructure
design and construction.” ASCE Civil Eng., 29(1), 48–53.
Steinman, D. B., and Nevill, J. T. (1957). Miracle Bridge at Mackinac, Eerdmans
Publishing, Grand Rapids, Mich.
6
New Developments
beyond North America
From North America to Europe:
British Challenge for Innovation
Since the completion of Charles Ellet’s 308-m span
Wheeling Bridge in 1849, the United States had for more
than a century continuously held the record for the world’s
longest suspension spans. In particular, America dominated
the field of long-span suspension bridges exceeding 1,000
m in span; no others existed anywhere in the world. At the
summit stood the 1,158-m span Mackinac Bridge designed
by D. B. Steinman and 1,298-m span Verrazano-Narrows
Bridge designed by Othmar Ammann.
However, the American domination of long-span
suspension bridges came to an end in the 1960s. Several
suspension bridges in the 1,000-m span range were con-
structed in rapid succession in the United Kingdom and
Portugal, as listed below:
• September 1964: Forth Road Bridge, United King-

dom, 1,006-m span
• August 1966: Tagus River Bridge, Portugal, 1,013-m
span
• September 1966: Severn Bridge, United Kingdom,
988-m span
The Forth Road Bridge was built across the Firth

of Forth next to the Forth Railway Bridge, a twin 521-m
span cantilever bridge that held the honor of being the world’s
longest span toward the end of the nineteenth century. This
159
bridge was the first long-span suspension bridge built outside North Amer-
ica, so when the United Kingdom started its construction in 1958, British
industry organized a nationwide consortium, the A.C.D. Bridge Co.:
1. Two design firms joint venture (JV):

• Mott, Hay and Andersen
• Freeman Fox and Partners
2. Three construction firms JV:
• Sir William Arrol and Co., Ltd.
• Cleveland Bridge and Engineering Co., Ltd.
• Dorman Long Bridge and Engineering, Ltd.
The British had, as we have seen, first learned the technology of early
suspension bridges from America, and again began to absorb American
long-span technology by requesting engineering services for construction of
the cables from the John A. Roebling’s Sons Co. However, with a desire to
rationalize and economize their designs, British engineers introduced many
innovations, such as the replacement of the heavy American-style concrete
deck with an orthotropic steel deck (Fig. 6-1). The towers of the Forth Road
Bridge, especially, had a sleek appearance. But one result of this design was
Figure 6-1. General plan of the Forth Road Bridge in east central Scotland.
new developments beyond north america 161
that it was too rational and light. As a result, the towers began to oscillate
in winds, making bridge workers seasick (Studio G. Hutton 1964).
To describe the economy of this bridge in the context of its era, one
can compare it with the Verrazano-Narrows Bridge, for example, which

was about 25% longer in span. The steel weight of the Forth stiffening truss
is 1/2.8 of the corresponding weight in the Verrazano-Narrows Bridge. The
weight of the cables is 1/4.7 and that of the towers is only 1/8.0 of the cor-
responding weight in the Verrazano-Narrows Bridge. One cannot generalize
too much because of the difference in the number of lanes and therefore
in the magnitude of live loads, but the economy of the Forth’s design is
impressive. The Forth Road Bridge breached the 1,000-m span barrier out-
side America and was officially opened to traffic on September 4, 1964, by
Queen Elizabeth II (Fig. 6-2).
A similarly well-known bridge exceeding 1,000 m in span outside
North America is the Tagus River Bridge connecting the city of Lisbon to
the municipality of Almada on the left bank of the Tagus River.1 Yet the
approach taken by the Portuguese government was completely different
from that taken in the United Kingdom. For this bridge, a widely adver-
tised international design competition was held. From the four teams that
Figure 6-2. Forth Road Bridge, which officially opened in 1964.

submitted proposals—one Portuguese, two European conglomerates, and

a conglomerate from the United States—an American team was selected
because it had the most experience. The U.S. design team was led by the
consulting firm of Steinman, Boynton, Gronquist & Birdsall, with Tudor

Engineering Co., the American Bridge Division of U.S. Steel, Morrison
Knudsen Co. Ltd., and others for construction. As a result, the Tagus River
Bridge was a direct import of American technologies. This is obvious from
the general plan shown in Fig. 6-3. Here, open grating was adopted in the
deck, which is supported by laterals through the stringers. Such features as
these facilitated air flow and are based on the same design principles used in
the Mackinac Bridge (Office of Tagus River Bridge 1966).
However, this bridge was required to accommodate two rail tracks in
the future and therefore several additional measures were introduced, such
as widening the distance between the tower legs and placing the line of the
main cables beyond the outer edge of the truss. As a result, it was planned
that the stiffening truss would pass through the towers, enabling it to be
Figure 6-3. General plan of the Tagus River Bridge in Lisbon, Portugal.
continuous. Future cable strengthening in the form of stay cables was taken
into consideration.
In 1997, 30 years after the bridge’s completion, authorities decided
to add the rail tracks, but over this period traffic conditions had changed
greatly. A standard-weight train track was now adopted in place of the light-
rail plan originally contemplated, and an additional road lane was to be
added to provide a total of three lanes in each direction. For these reasons,
the original cable-stayed concept was abandoned and instead the bridge was
strengthened through the addition of two cables, each 35.5 cm in diameter.
The new cross section after rehabilitation is shown in Fig. 6-4, and a photo
taken during the aerial spinning of the secondary cables is shown in Fig. 6-5.
It was reported that the work was performed under average traffic volumes
of 140,000 vehicles per day (Parker 1997; Reina 1997).
Figure 6-4. Cross section after strengthening the Tagus River Bridge.
Before retrofitting (left, “Existing”) and after retrofitting (right, “New”).
Source: Reina (1997), with permission from Estradas de Portugal, S.A. Centro de Documen-
tação e Informação.
Figure 6-5. Secondary cable installation (aerial spinning) on the Tagus River Bridge.
Source: Reina (1997), with permission from Estradas de Portugal, S.A. Centro de Documen-
tação e Informação.
Aerofoil Revolution
With the lessons learned from the failure of the Tacoma Narrows Bridge
in 1940, engineers in the United States conquered the problems of aero-
dynamic stability and further extended span lengths. These efforts generated
successful, gigantic suspension bridges such as the Mackinac Bridge and the
Verrazano-Narrows Bridge. Two mainstream approaches were used to fur-
ther enhance aerodynamic performance. One adopted a stiffening truss and
open grating decks, thus eliminating the generation of vortices (or at least
reducing their size). The other approach increased stiffness by increasing the
mass (or weight) of the suspension bridge. Center stays and cable stays were
sometimes used, but these were only considered measures of a supplemen-
tary or emergency nature. As a result, mainstream American aerodynamic
measures involved the use of trusses and mass.
However, European engineers began to develop a completely different
approach to that of the Americans by adopting an aerofoil section that could
reduce wind forces and suppress the emergence of vortices. This method could
be inferred by extrapolating the results of a series of wind tunnel tests performed
on the cross section of the original Tacoma Narrows Bridge by F. B. Farquhar-

son prior to that bridge failure. His tests had actually indicated an improvement
in the stability of the cross section through the addition of a fairing (refer to
Chapter 5 for a complete discussion) (Farquharson 1949; Kawada 1975).

Suspension bridge design did not develop in this direction in the United
States, however, and the idea was finally presented in the international com-
petition for the Tagus River Bridge held in 1960. The idea emerged with
a proposal from Dr. Ing. Fritz Leonhardt, a West German structural engi-
neer and university professor. As shown in Fig. 6-6, Leonhardt’s design was
Figure 6-6. Leonhardt’s proposal for the Tagus River Bridge.

The proposal featured an aerofoil-like deck, a monocable considered to be effective against
torsional vibration, and diagonal hangers.
innovative and bold, and featured a truly aerofoil-like deck, a monocable

considered to be effective against torsional vibration, and even diagonal
hangers (Office of the Tagus River Bridge 1966). It is possible that this idea
was not adopted because it was considered too innovative (the U.S. design
won the competition), but this new concept of an aerofoil-like stiffening
girder brought about an aerofoil revolution in subsequent world-ranked
long-span suspension bridges.
The bridge that would embody this revolution opened just one month
after the Tagus, but in concept it was a generation apart. Following the
Second World War a truss-stiffened suspension bridge across the Severn
Estuary linking England and Wales was to have been built, but changing
national priorities saw the design adapted to the Forth Road Bridge. The
intervening years saw a shallower truss adopted for the Severn Bridge, but
even this design would appear outdated in comparison to the streamlined
box girder design ultimately used. According to Leonhardt, the Severn
Bridge’s streamlined box section was derived from his work as follows:
During my wind tunnel tests at the NPL [National Physical Labora-

tory] in Teddington [U.K.] with my new cross sections, there were tests
for the Severn Bridge going on with models of the truss type. After the
favorable results of my tests became known, the design for the Severn
Bridge was changed. A cross section with a flat hollow box girder and
with thin fin-shaped cantilevers for walkways at the upper third of the
depth were [sic] developed which became characteristic for modern
English suspension bridges. (Leonhardt 1982)
Some have accepted that the Severn’s evolution actually occurred as

described above. In any event, as shown in Fig. 6-7, the cross section of the
Severn Bridge gradually changed as wind tunnel testing continued, and what
finally emerged was an unprecedented, revolutionary aerofoil section box
girder (Walshe and Rayner 1962; Kawada 1987).
With the change from the truss to an aerofoil section, the suspension
bridge became significantly more lightweight and economical, as described
by British bridge designer Sir Gilbert Roberts:
The advantages of the plated box form of construction, if it were practi-

cable, had been apparent to the author for some years before the oppor-
tunity occurred to make a thorough study of it for the Severn Bridge.
The more practical advantages are obvious. The full width of the top
surface of the box uninterrupted by longitudinal gaps is available as
useful road area and the surface requiring external painting is very much
Figure 6-7. Evolution of Severn Bridge cross section.

(top) Preliminary cross section (adopted for Forth Road Bridge); (middle) First revised cross
section; (bottom) Final cross section.
reduced. The interior of the box can be sealed off and requires no main-
tenance. The torsional stiffness of the deck of a suspension bridge is a
very useful property in dealing with aerodynamic oscillations, and the
box section has a much greater torsional stiffness, weight for weight,
than any arrangement of trusses and cross-girders.
Further there is the prospect of considerable saving in weight
of steel as compared with a truss girder bridge with separate non-
participating deck, since the steel in the box section is capable of resist-
ing stresses in several directions at the same time—local and general
bending, lateral bending, shear and torsion.
The reduction in suspended weight is reflected in further savings
in cable, towers, and anchorages, so that even if the box section costs the
same as the trusses there would be an overall saving in money. Another
advantage of the box section is the reduction in wind drag on the deck
to one third of that for the trusses; this is particularly significant in the
design of the towers which sustain 70 percent of the wind force on the
deck as a lateral reaction at the top. . . . (Roberts 1966)
In this way, the trend toward the American type of suspension bridge fol-
lowing the failure of the Tacoma Narrows Bridge was reversed again, to one of
more lightweight and economical design. Also, in contrast to existing bridges
using vertical hangers, a new approach was taken to enhance structural damp-
ing, described by Roberts: “Attention was therefore directed to the provision
of added structural damping, and the author had the idea of making use of the
energy absorbing properties, or hysteresis, of the wire ropes or strands forming
hangers. . . . For this purpose the two parts of each hanger are separated and
inclined to the vertical to form a triangular system. . . .” (Roberts 1966). From
this observation, the Severn Bridge, shown in Fig. 6-8, came to incorporate the
epoch-making diagonal hangers for the first time in the world.
A revolution had now occurred in the development of suspension
bridges. Having been stiffened by a light aerofoil or streamlined box and by
triangular, truss-like diagonal hangers, the Severn Bridge changed suspen-
sion bridge design worldwide upon completion in 1966 with its extremely
economical and innovative design (Fig. 6-9).
By comparing steel weights among major suspension bridges built
from the 1960s to the 1980s, as shown in Table 6-1, one can easily under-
stand how innovative the adoption of an aerofoil box section was for sus-
pension bridges. (Note that the George Washington Bridge does not belong
to this period, that the side spans of the First Bosporus Bridge are not sus-
pended, and that the Humber Bridge employed concrete towers.)
Compared with American suspension bridges, those built in Europe
after the 1960s reflected significantly reduced steel weights, and this trend
Figure 6-8. Diagonal hangers in the Severn Bridge.

was most pronounced in the Severn, Bosporus, and Humber bridges. The
Tagus River Bridge in Portugal was a Steinman-style design that fully incor-
porated all aerodynamic considerations. Even so, the differences between
the truss and aerofoil sections are immediately apparent. Significantly large
differences are found in the towers; the steel weight of the Severn Bridge
towers was only one-fifth (and in the First Bosporus Bridge only about one-
third) that of towers in the Tagus River Bridge.
For example, Fig. 6-10 shows the general outline of the Severn Bridge
tower. In contrast to the cellular structure used in traditional American
suspension bridges, the tower leg is a rectangular box consisting of stiff-
ened external shells and internal diaphragms. The connections were bolted
between flanges and only fastened internally. As a result, no bolt heads are
found on the outside of the towers (Associated Bridge Builders 1966).
Earlier, I described the emergence of the Severn Bridge as an aero-
foil revolution. Indeed, it really can be considered a revolutionary develop-
ment. The Severn Bridge ended the predominance of long-span suspension
bridges by the United States. Instead, the United Kingdom produced the
Severn Bridge, and in the glow of worldwide attention stood Sir Gilbert
Roberts, who led the design consultant Freeman Fox and Partners. This firm
then went on to win the design competition for the First Bosporus Bridge
in Turkey. When this 1,074-m span bridge was completed in 1973, Free-
man Fox and Partners turned its sights to the design of the Humber Bridge

Table 6-1. Comparison of Steel Weights of Major Suspension Bridges from the 1960s to the 1980s
Weight of Steel (tonnes)
Year Suspended
Bridge Name Country Completed Main Span (m) Total Structure Cables Towers
Verrazano-Narrows USA 1964 1,298 132,000 45,200 37,800 49,100
Forth Road UK 1964 1,006 30,500 16,300 8,100 6,100
Tagus River Portugal 1966 1,013 38,000 16,300 7,700 14,000
Severn UK 1966 988 18,800 11,400 4,700 2,700
First Bosporus Turkey 1973 1,074 19,000a 8,700 5,500 4,800
Humber UK 1981 1,410 27,000a 16,000 11,000 Concrete
a
First Bosporus has no suspended side spans; the Humber Bridge towers are concrete.
Figure 6-9. Upon completion in 1966, the Severn Bridge changed suspension bridge
design worldwide.
Source: Author.
Figure 6-10. Severn Bridge tower legs are rectangular boxes consisting of stiffened
external shells and internal diaphragms.
Source: Associated Bridge Builders (1966).
near Kingston upon Hull in England. With a 1,410-m span, it would be

the world’s longest suspension bridge. When completed in 1981, this span
ended America’s hegemony in the genre.
Figure 6-11 shows the general plan of the First Bosporus Bridge; that
of the Humber Bridge is shown in Fig. 6-12. It is obvious at first glance
that these bridges use essentially the same structural concepts as the Severn
Bridge, except that the side spans of the Bosporus Bridge are not suspended
and concrete towers are used on the Humber Bridge (British Bridge Builders
1981). Figures 6-13 and 6-14 show photos of each bridge, and both are very
slender, reminding us of the old Tacoma Narrows Bridge. As a result, these
suspension bridges reflected significant economies in comparison to the
American type. For example, if one compares the steel weight of suspended
structures, that of the Humber Bridge (110 m longer than the Verrazano-
Narrows Bridge) is only 35% of the latter span’s and is similar to that found
in the Forth Road and the Tagus River bridges.
Glory and Tragedy of the Severn Bridge

Another Severn-style suspension bridge—the momentous result of the aero-
foil revolution—enabled the United Kingdom to finally exceed America’s span
record with the Humber Bridge, the world’s longest suspension bridge in the
Figure 6-11. General plan of the First Bosporus Bridge in Turkey.

Figure 6-12. General plan of the Humber Bridge in England.

Figure 6-13. First Bosporus Bridge.

Source: Author.
Figure 6-14. Humber Bridge.

Source: Author.
1980s. This bridge also brought with it significant economies in comparison

to American-style suspension bridges. On July 17, 1981, Queen Elizabeth II
herself was in attendance to announce its opening. It was a glorious moment
for British bridge technology, which had now surpassed America about a cen-
tury after the completion of the 521-m span Forth Railway Bridge in 1889.
Ironically, following the opening of the Humber Bridge, structural
problems began to emerge in the Severn Bridge, the prototype for the
Humber. These problems began with the wind-induced vibration of its
diagonal hangers, a phenomenon that had never been observed in verti-
cal hangers. Damping devices were immediately installed as a suppression
measure (Fig. 6-15), although such superficial remedies could not resolve
the situation. Beginning in the spring of 1982, articles with headlines such
as “English Bridge Cables Fail” began to frequently appear (Engineering
News-Record, April 8, 1982). And Flint & Neill Partnership, the firm that
undertook an investigation of the Severn Bridge, submitted a report indicat-
ing that replacement of the hangers alone would cost £5 million.
Dr. Ing. Hellmut Homberg, a German consulting engineer, clearly
explained the problems with the diagonal hangers. He analyzed the diago-
nal hanger as a truss diagonal member that is subjected to reversing stresses,
and concluded: “Diagonal hangers suffer high stress amplitudes, vary-
ing from zero to levels of tension exceeding the allowable design limit of
Figure 6-15. Damping devices installed at the Severn Bridge hangers.

Source: Author.
600 N/mm2 by more than 50 percent, thereby accelerating fatigue failure

which is directly proportional to amplitude and maximum working stress”
(New Civil Engineer, April 15, 1982).
Freeman Fox and Partners immediately rebutted such criticism. In the
same journal, two weeks after publication of Dr. Homberg’s comments, Ber-
nard Wex (one of the firm’s partners) wrote that the problems arose mainly
because of significantly increased traffic volumes (in other words, the live
load condition) as well as the synchronized effects of many finely tuned
bridge attributes (New Civil Engineer, April 29, 1982).
As the debate continued, the Severn Bridge’s already dire condition

worsened. First, it became evident that current traffic volumes were now
double the original design estimate, and it was further concluded that the
problems in the suspension bridge were not limited to the hangers but were
in fact more extensive. Here is a sampling of articles:
“Live Loads Soar on U.K. Bridge” (Engineering News-Record, July 1,

1982)
“Severn Report Lists Wide-Ranging Weaknesses” (New Civil Engineer,
Nov. 18, 1982)
“Severn Bridge Woes Mount” (Engineering News-Record, Dec. 2, 1982)
“Report Puts $54-Million Tag on U.K. Bridge Upgrading” (Engineering
News-Record, May 26, 1983)
“Severn Fears Grow with ‘Sudden Collapse’ Warning” (New Civil
Engineer, Nov. 3, 1983)
From the titles of articles published over a period of a year or so, the
seriousness of the Severn Bridge’s condition became the subject of much
speculation. Figure 6-16 shows the hangers for which repair was deemed
essential at that time, and also shows the fatigue cracks in welds found on
the interior of the stiffening girder [from the report by Flint & Neill Part-
nership (1982)]. The photographs in Fig. 6-17 and 6-18 show the actual
retrofitting work. These photographs formed part of a paper I co-authored
which was published in Bridge and Foundation Engineering, Japan after my
return from a visit to the bridge site. The photos were taken with Atsushi
Hirai, then Professor Emeritus at the University of Tokyo. The repair work
on the Severn Bridge continued under tight traffic restrictions, with two of
its four lanes closed.
The most severe problems in the Severn Bridge were found in the
towers. As described in the preceding section, the towers were extremely light
and fragile, and could no longer endure the current conditions of increased
live loads. As a result, they required extensive reinforcement. Fortunately,
the tower legs in the Severn Bridge consisted of a box made up of four plates
(of course, stiffened by diaphragms), and were not of multicellular design.
Here, engineers planned to insert steel pipes inside each of the four cor-
ners of each tower leg—in essence, building another tower inside the tower
(New Civil Engineer Jan. 3, 1985). The last operation in this tower repair
work involved the load transfer to the new towers, performed in January
through March of 1990. The work was reported in a British technical jour-
nal with an explanatory drawing (Fig. 6-19). According to this paper, the
four steel pipes installed within the four corners of the tower column were

Figure 6-16. Locations of hangers and defects in Severn stiffening girders.
Source: Flint and Neil Partnership (1982), with permission.
(A)
(B)
Figure 6-17. Remediation of hangers on the Severn Bridge.

(A and B) Remediation of hangers; (C) A. Hirai, professor emeritus, University of Tokyo.
Source: Hirai and Kawada (1984), courtesy of Tadaki Kawada.
(C)
Figure 6-17. (Continued)
fixed at 6-m intervals in consideration of buckling; sliding bearings allow

for vertical movement. Three jacks were set at the base through triangular
jacking collars. At the top of the tower, the four pipes were assembled into a
single unit to support the strengthened cable saddle. A jacking force of 2,000
tonnes (with a total of 25 mm per lifting stroke) was required, correspond-
ing to one-third of the dead load reactions at the tower top. The first jacking
operation was cautiously performed with a 600-tonne load, and stresses in
the members were checked. The work was performed at a time of minimum
Figure 6-18. Repair of fatigue cracks in welds on the Severn Bridge.

Figure 6-19. Diagram of jacking the Severn Bridge tower during repairs.
Source: New Civil Engineer, Jan. 18, 1990.
traffic between 11 p.m. and 3 a.m., such that traffic was not halted during
this time (New Civil Engineer, Jan. 18, 1990).
As described above, the program of repair work encompassed the
hanger cables and their connections to the stiffening box girder, the repair
of the girder itself, and tower strengthening. The total cost of the repair pro-
gram reached £33 million. The strengthening and retrofitting cost 2.5 times
the Severn’s initial construction cost of £13 million, although there was a
time lapse of 16 to 17 years between the two [Financial Times (London),
Jan. 5, 1984].
Mass: Engineers’ Forgotten Asset

The Severn Bridge, whose ingenuity was once regarded as the dominant idea
of the time, began suffering from the fatal disease of structural fatigue after
about 10 years of service. This was the most shocking and controversial
topic for bridge engineers since the 1940 disaster of the Tacoma Narrows
Bridge. Through repeated wind tunnel tests and theoretical analyses, the
aerodynamic dangers for suspension bridges that had preyed on the Tacoma
Narrows Bridge decades earlier were believed to have been conquered and
eliminated by now. Moreover, the stiffening girder of the Severn Bridge used
the ingenious invention of an aerofoil box section that overcame the inher-
ent weakness of plate girders “since the steel in the box section is capable
of resisting stresses in several directions at the same time” (Roberts (1966).
Accordingly, the Severn Bridge had been praised both for its higher torsional
stiffness and for its remarkable cost savings. In reality, however, after 16 to
17 years of service the Severn Bridge had deteriorated so much that it was
feared it would soon come to the end of its life. Therefore, the bridge had to
be repaired at enormous expense.
Something was odd with this suspension bridge. Yes, something must
have been wrong. The Severn Bridge once predominated the long-span sus-
pension bridges in the world and was praised as a masterpiece of modern
bridge engineering that made the most use of the fruits of recent wind tunnel
technologies. But the reality of the bridge was this: There must be some
essential oversight; some fatal blind spot must have been inherent in this
masterpiece of a suspension bridge. But what was this blind spot?
Reinvestigating the wind tunnel tests for the Severn Bridge myself,
I could not help but conclude that, from the beginning, the testing lacked
the factor of mass and that it had been too preoccupied with the shape of
the stiffening girder cross section. Previous post-Tacoma suspension bridges
in America were all stiffened by trusses accompanied with concrete decks,
and therefore their mass was inherently large. Hence, limited vibrations
such as vortex-induced oscillation or gust response would become neither a
problem nor structurally fatal (Fig. 6-20) (Ito 1985).2 In the Severn Bridge,
the significance of mass in suspension bridges was evidently forgotten. The
span was too light, and from this emerged a structure extremely sensitive to
the effects of wind and vehicular traffic. Although a catastrophic divergent
vibration did not occur during wind tunnel testing, the suspension bridge
was constantly in motion under the action of real winds and vehicles.
How sensitive was the Severn Bridge? Could its sensitive properties
be improved by stiffening with a truss or increasing the dead load (mass)?
Figures 6-21 and 6-22 show the results of trial calculations by Hirai and
Figure 6-20. Typical aerodynamic oscillations in suspension bridges.

Source: Ito (1985), with permission from Manabu Ito.
Kawada (1985) based on the criteria shown in Table 6-2. As clearly shown
in Fig. 6-21(top), a Severn-type stiffening girder (Sev.) deflects much more
easily than a Forth Road Bridge-type (For.), the difference reaching about
a 20% greater magnitude. However, if mass is increased in the Severn-type
girder [Sev.(M)], the resulting increase of bending stiffness reduces the mag-
nitude of deflection by about 40% in comparison to the original Severn-type
and 20% below the Forth Road Bridge-type. In contrast, Fig. 6-21(top)
clearly indicates that a light suspension bridge deflects easily, and therefore
tends to vibrate and becomes sensitive to live loading. The mass added to
Sev.(M) was assumed to be exactly the weight of a reinforced concrete deck
of 2.5 tonne/m2, which corresponds to the 20-cm-deep and 14.9-m-wide
deck used in the side span of the Forth Road Bridge.
Figure 6-21(bottom) is the bending moment diagram. It is obvious
that the bending moment of the truss For. is the largest among the other sus-
pended structures, and that the bending moment in a box section is reduced
by 40% by adding mass, thus proving the superiority of adding mass.
Long-span suspension bridges may oscillate under the action of wind
rather than from live loading. Figure 6-22 displays the results of an investi-
gation by Hirai and Kawada (1985) into the lateral deflectibility of stiffen-
ing girders (deflections for main spans are shown). Figure 6-22(top) reveals
that the truss section For. experiences the most deflection, and that a stream-
lined section such as that found in the Severn Bridge is significantly supe-
rior. Again, the effectiveness of adding mass is demonstrated in the reduc-
Figure 6-21. Comparison of conditions of stiffening by live loading.

(top) maximum vertical displacement diagram; (bottom) maximum vertical bending moment
diagram. Sev., Severn-type stiffening girder; Sev.(M), Severn-type girder with increased mass;
For., Forth-type girder.
Figure 6-22. Comparison of conditions of stiffening by wind loading for the center
span.
(top) Maximum lateral displacement diagram; (bottom) maximum lateral bending moment
diagram. Sev., Severn-type stiffening girder; Sev.(M), Severn-type girder with increased mass;
For., Forth Road Bridge-type girder.
Table 6-2. Dimensions and Stiffening Conditions used in the Trial

Calculations for the Severn and Forth Bridges
Severn Fortha Severn (M)
Dead Load
Cable and hanger (tonnes/m/bridge) 2.76 3.90 (4.0) 4.80
Suspended structure (tonnes/m/ 9.55 11.50 (17.0) 17.00
bridge)
Total (tonnes/m/bridge) 12.31 15.40 (21.0) 21.80
Wind Load
Wind load acting on cable, Wc 0.217 0.256 0.284
(tonnes/m/bridge)
Wind load acting on girder, Wg 0.238 1.080 0.238
(tonnes/m/bridge)
Stiffness
Area of cable (m2/bridge) 0.324 0.452 0.556
Moment of inertia in vertical 1.126 4.7 (3.5) 1.126
direction, Ix (m4/bridge)
Moment of inertia in lateral 48.070 38.0 (27.0) 48.070
direction, Iy (m4/bridge)
Torsional stiffness factor of girder, 2.898 1.3 (1.3) 2.898
J (m4/bridge)
Severn indicates Severn Bridge model, prototype of the Severn bridge stiffened by lightweight
box girder. Forth indicates Forth Bridge model stiffened by truss with orthotropic steel deck.
Severn (M) indicates Severn Bridge model stiffened by lightweight box girder with addition of
mass proposed by Hirai and Kawada (1985). The weight in this model is the reinforced con-
crete deck used in the Forth Bridge.
aParameters in ( ) are for the side spans when they differ from the main span.
Source: Hirai and Kawada (1985).

tion of deflection in the streamlined section by 23%. In the same way, the
maximum lateral bending moments shown in Fig. 6-22(bottom) indicate the
superiority of adding mass to the box section [Sev.(M)].
As seen so far, the addition of mass improves resistance against static

wind loads, but this mass addition also influences a span’s dynamic proper-
ties. Although the mass added in this study did not influence frequencies,
as shown in Fig. 6-23, their amplitudes in terms of vortex-induced vibra-
tion and oscillation from buffeting—as well as critical wind speeds—were
significantly reduced, thus exhibiting an improvement (Hirai and Kawada
1985; Nomura et al. 1985).
The shortcomings in the Severn Bridge as observed so far, and the
valuable lessons obtained from this bridge, were reflected in the construc-
tion of subsequent suspension bridges. The difference is seen in the illus-
trative examples of two spans designed by Freeman Fox and Partners, the
First Bosporus Bridge (completed in 1973) and the Second Bosporus Bridge
(completed in 1988) in Turkey.3 Both bridges were built across the Bos-
porus Straits, not far from one other (Fig. 6-24), and feature very similar
proportions with almost identical main spans and unsuspended side spans.
In addition, both employed a Severn-style aerofoil section for their stiffening
girders. However, although both bridges appear at first glance to be twins,
there is an essential difference in design principle. First, the diagonal hang-
Figure 6-23. Improvement in dynamic properties of suspension bridges by the addi-

tion of mass.
(A) tendency in vortex-induced oscillation; (B) tendency in buffeting (gust response); and
(C) critical wind speed.
Figure 6-24. Comparison of the First and Second Bosporus bridges.

(top) First Bosporus Bridge; (bottom) Second Bosporus Bridge.
ers adopted in the Severn Bridge and then in the First Bosporus Bridge were
abandoned in the Second Bosporus Bridge. Second, by abandoning diagonal
hangers, mass (weight) was increased to supplement the reduced stiffness.
Figure 6-25 shows the results of my trial calculations to verify whether
stiffness reductions could be supplemented only by mass addition, as theory
might suggest. The span of the model was identical to that of the Second
Bosporus Bridge. The First Bosporus Bridge’s dead load of D1 (570 kgf/m2),
including the cables’ weight, was applied to the model with diagonal hang-
ers. A total dead load of D2 (670 kgf/m2) was applied to a Second Bosporus
Bridge model with vertical hangers, where the ratio of D2/D1 is nearly 1.18.
The results shown in Fig. 6-25 indicate that the 18% increase in dead load that
was actually adopted for the Second Bosporus Bridge could increase stiffness
by a similar amount as that provided by diagonal hangers alone. Even with
vertical hangers, the bending moments and displacements were significantly
reduced. This result demonstrates that the addition of mass could make the
problematic diagonal hangers unnecessary (Kawada 1989; Ito et al. 1999).
With the Verrazano-Narrows Bridge being the last of a series of heavy
American-style spans, the worldwide trend had shifted toward suspension
bridges of lighter weight. Now the trend started to again shift in another
direction: retaining the Severn-style aerofoil section but abandoning diag-
onal hangers and supplementing the loss in stiffness by way of increased
dead load. Figure 6-26 compares unit steel weights in (kgf/m2) in ordinate
and span lengths (m) in the abscissa for major aerofoil-section suspension
bridges built toward the end of the twentieth century since (and including)

Figure 6-25. Influence of configuration of hangers and mass.
the Severn Bridge. In this figure, the only suspension bridges using diagonal
hangers are the Severn, Humber, and the First Bosporus bridges. These three
suspension bridges are early spans of the genre, whereas all subsequently
built spans adopted vertical hangers.

The steel weight per unit of deck area (m2) of the Second Bosporus
Bridge is 33% heavier than that in the First Bosporus Bridge, and further-
more is 9% heavier than even that of the Humber Bridge, a bridge that is
30% longer in span. Consequently, the unit-steel weights of later aerofoil-
section suspension bridges diverge from the average trend of the earlier
Severn and Humber bridges built in the United Kingdom. That is, the prob-
lems raised by the Severn Bridge were conquered by two methods: abandon-
ing diagonal hangers, and adding mass. These measures are exactly what my
colleagues and I have advocated since 1984:
Since the time, however, when the stream-lined box section for the stiff-
ening girder was adopted for the Severn Bridge, it is regrettable that
engineers seemed to have forgotten the role and the significance of the
weight or mass in a suspension bridge. Instead, they invented inclined
hangers for the purpose of stabilizing a light suspension bridge, and on
the contrary made matters worse.
Figure 6-26. Steel weights in streamlined box suspension bridges.

B1, First Bosporus; B2, Second Bosporus; K1, Kurushima #1; K2, Kurushima #2; K3, Kurush-
ima #3; HA, Hakucho; HU, Humber; N, Nankai; O, Ohshima; SB, Storebælt; SE, Severn.
To tell the truth, if only we had remembered correctly what John

Roebling said, and if we had added mass to the suspended box girder, the
inclined hangers would be of no use and would not have been contrived.
In our opinion, the stream-lined box section with additional

mass should have prevailed as the best solution for both structural and
economical reasons, not the solution with inclined hangers. (Hirai and
Kawada 1985)
Notes
1. The Tagus River Bridge has three popular names. Before completion it was
known as the Tagus River Bridge, after the name of the river to be crossed. After
completion the bridge was named the Salazar Bridge, after the dictator of Portugal
at that time. However, 8 years later, when the dictatorship was overthrown on
April 25, 1974, the date of the liberation became the bridge’s name. Therefore,
the present official name is Ponte de 25 April.
2. When the natural frequency of a suspension bridge matches the frequency of
wind, then theoretically resonance occurs and it diverges (dotted line in Fig.
6-20). However, it is difficult to believe that a wind with the same frequency
can be of long duration, and therefore in reality the gust response is a limited,
nondivergent oscillation.
3. Neither the First Bosporus Bridge nor the Second Bosporus Bridge is the official
name of these spans. The first bridge is the Bosporus Bridge and the second bridge
is the Fatih Sultan Mehmet Bridge. However, the popular names of the First and
Second Bosporus bridges are used in this book.
References
Associated Bridge Builders. (1966). The Severn Bridge Superstructure, Associated
Bridge Builders, Ltd., U.K.
British Bridge Builders. (1981). Bridging the Humber, Cerialis Press, York, U.K.
Farquharson, F. B. (1949). Aerodynamic Stability of Suspension Bridges with
Special Reference to the Tacoma Narrows Bridge. Part I: Investigation Prior
to October 1949. University of Washington, Engineering Experiment Station,
Bulletin No. 116, Part I.
Flint & Neill Partnership. (1982). M4 Severn Crossing: Structural Feasibility Study,
Interim Report, October, Flint & Neill Partnership, London.
Hirai, A., and Kawada, T. (1984). “What is going on in suspension bridges: Severn,
Bosporus, and Golden Gate.” Bridge & Foundation Engineering, Japan,
October.
Hirai, A., and Kawada, T. (1985). “Proposal for suspension bridge with additional
mass: a solution for Severn Bridge problem.” Kawada Technical Report, 4
(January), 14–20.
Ito, M. (1985). “Cable supported bridges and winds.” Kawada Technical Report, 4
(January), 5–9.
Bridges: Further Challenge of Engineers [in Japanese], Kensetsu Tosho,
Japan.
Kawada, T. (1975). Who wrecked Galloping Gertie? The Mystery of the Tacoma
Narrows Bridge Disaster [in Japanese], Kensetsu Tosho, Tokyo, Japan.
Kawada, T. (1987). Modern Suspension Bridges, Riko Tosho, Japan.
Kawada, T. (1989). “Long-span suspension bridge again changed the trend.”
Kawada Technical Report, 8 (January).
Leonhardt, F. (1982). Brücken: Ästhetic und Gestaltung, Leonhardt, Andrä, und
Partner, Berlin.
Nomura, K., Oka, K., Maeda, K., and Yoneda, M. (1985). “Study on the
characteristics of a mass added suspension bridge.” Kawada Technical Report,
4 (January), 21–35.
Office of the Tagus River Bridge. (1966). The Salazar Bridge, Gabinete de Ponte
Sabre O Tejo, Lisbon, Portugal.
Parker, D. (1997). “High wire live action.” New Civil Engineer, August 21/28.
Reina, P. (1997). “A crossing revisited.” Engineering News-Record, September 1.
Roberts, G. (1966). “The Severn Bridge design: a new principle of design.” Proc.,
Intl. Symp. on Suspension Bridges, Lisbon, Portugal.
Studio G. Hutton. (1964). Forth Road Bridge Superstructure, Rudby Yarm,
Yorkshire, U.K.
Walshe, E., and Rayner, D. V. (1962). “A further aerodynamic investigation for the
proposed Severn Suspension Bridge.” Aero Rpt. No. 1010, National Physical
Laboratory (NPL), Teddington, U.K.
7
Two Record-Breaking
Suspension Bridges
Japan’s Akashi Kaikyo Bridge

With the completion of the Humber Bridge in 1981,
Britain took the honor of having built the world’s longest
span and brought the long-standing American reign in this
field to an end. But Britain’s hold on the long-span throne
did not last even 20 years; she had to relinquish her posi-
tion before the end of the twentieth century because in 1988
in Japan, and in 1991 in Denmark, construction began on
two gigantic suspension bridges that were scheduled for
completion before 2000.
Japan’s Akashi Kaikyo Bridge was planned as a
1,990-m main span with two 960-m side spans for a total
suspended length of 3,910 m, and the Storebælt Bridge in
Denmark features a 1,624-m main span with 536-m side
spans for a total suspended length of 2,712 m. Both vastly
exceed the Humber’s 1,410-m main span. These two enor-
mous suspension bridges were completed soon after one
another but they sharply contrasted in terms of their design
concepts. This reminds one of the conceptual contrast
between Othmar Ammann and D. B. Steinman in the Ver-
razano-Narrows and Mackinac bridges; in these new struc-
tures, Japan took a traditional stance, as had Ammann, but
it was Denmark that boldly introduced new concepts, as
had Steinman.
The Japanese engineers’ traditional approach was
not surprising, considering that until the end of the Second
World War not one Japanese bridge could be described as
a worthy long-span bridge; Japanese bridge engineering
193
had not yet developed far enough. Even so, with Japan’s postwar economic
recovery, its nationwide development of highways was initiated and its
desire to build long-span bridges started to gain strength.
Briefly looking back at the steps taken by Japan in the development of

long-span bridges, its first bridge to exceed 200 m in span was an arch bridge,
the Saikai Bridge (216-m span) in 1955. Span lengths steadily increased, and
with the Akashi Kaikyo Bridge completed in 1998, Japan finally came to
possess the world’s largest and longest span bridge.
• 1962: Wakato Bridge, 367-m span
• 1973: Kanmon Bridge, 712-m span
• 1983: Innoshima Bridge, 770-m span
• 1985: Ohnaruto Bridge, 876-m span
• 1988: Shimotsui-seto Bridge, 940-m span
• 1988: Kita Bisan-seto Bridge, 990-m span
• 1988: Minami Bisan-seto Bridge, 1,100-m span
For Japanese engineers, who began promoting long-span suspension
bridges but had essentially no background in this field, America was the
supreme front-runner and mentor. Therefore, from the Wakato Bridge to
the Akashi Kaikyo Bridge, Japan learned from America, absorbed American
technologies, and eventually surpassed America—which had ended its era of
large suspension bridge construction in the 1960s—as a dominant force in
suspension bridge engineering (Gimsing 1999).
Examples of representative suspension bridges built in Japan include
the Innoshima Bridge in Fig. 7-1, the Minami and Kita Bisan-seto bridges
Figure 7-1. Innoshima Bridge.

Source: Author.
two record-breaking suspension bridges 195
in Fig. 7-2, and the Akashi Kaikyo Bridge in Fig. 7-3. All featured a consis-
tently traditional form reflecting American mainstream design, with diago-
nally braced steel towers and stiffening trusses, without the influence of
the revolutionary aerofoil section. As the Akashi Kaikyo’s cross section in

Fig. 7-4 shows, the deck is supported on transverse trusses by way of string-
ers, and the open grating providing open gaps reflects the results of the
wind tunnel testing and aerodynamic design that followed the 1940 Tacoma
Narrows Bridge failure (Kurino 1998). Of course, the results of engineer-
ing research and the developments of this period (for example, the vertical
stabilizer) were incorporated (Yasuda et al. 1978). Other features not shown
here include a light-weight orthotropic deck instead of a heavy concrete
deck, and welded, instead of bolted, connections, all of which contributed
to reduction in overall weight.
Major technical problems not prevalent in America that had to be
overcome in the Akashi Kaikyo Bridge were the threats of earthquakes and
typhoons. In fact, the Hanshin earthquake struck on January 17, 1995 at the
time when the towers and cables were completed. The sliding of active faults
feared during the planning stage became a reality near the bridge’s center
span, and caused the displacements shown in Fig. 7-5. The plan had to there-
fore be hastily changed because the earthquake lengthened the main span by
Figure 7-2. Minami (foreground) and Kita Bisan-seto (background) bridges.

Source: Author.
Figure 7-3. Akashi Kaikyo Bridge.

Source: Author.
Figure 7-4. Suspended structure of the Akashi Kaikyo Bridge.

Source: Kurino (1998), with permission from Honshu-Shikoku Bridge Expressway Corporation.
Figure 7-5. Effects of the Hanshin earthquake (January 1995) on the Akashi Kaikyo
Bridge.
0.8 m—to 1990.8 m from 1990.0 m—and increased the Awaji Island side span
by 0.3 m, from 960.0 m to 960.3 m. Moreover, the profile at center span was
raised to 97 m, 1 meter above the original elevation of 96 m (Kurino 1998).
The Akashi Kaikyo’s approximately 300-m-tall steel towers also pre-
sented problems that had never been experienced in previous suspension
bridges. It was well known that some towers of long-span suspension bridges
vibrated in winds during construction and that workers became seasick. All
these previously reported vibrations occurred only during the time of tower
erection, and once the cables were in place they acted like stays to quell the
vibrations (Shirley-Smith 1967; Miura and Fujiwara 1996).
However this situation was not applicable to the 287-m-tall Akashi
Kaikyo towers. As the following list of steel tower heights (above the foun-
dations) for existing major suspension bridges reveals, the Akashi Kaikyo
towers are exceedingly tall. (Because concrete towers are excluded from this
list, the 156-m-tall concrete towers of the Humber Bridge are not shown.)
• Mackinac Bridge, 168 m

• George Washington Bridge, 181 m
• Verrazano-Narrows Bridge, 210 m

• Golden Gate Bridge, 227 m
• Akashi Kaikyo Bridge, 287 m
The great height of the Akashi Kaikyo towers resulted in an excessively

flexible structure, and engineers remained concerned that vibration would
not be suppressed even after the main cables were completed and when
all the dead loads were placed on the towers. It was consequently decided
to take the unprecedented step of installing damping devices as permanent
measures, whereas they had previously been employed only during the erec-
tion stages. Figure 7-6 shows the mechanism of the tuned mass damper and
Figure 7-6. Damping devices for the Akashi Kaikyo Bridge.

the locations where they were installed. For each tower, eight units weighing
10.5 tonnes and 12 units weighing 9.5 tonnes, respectively, were installed
as permanent devices comprising about 0.9% of the tower’s total weight
(Kurino 1998).
The first operation after the erection of the towers was the hauling of
a pilot rope. In the Akashi Kaikyo Bridge, a helicopter was used successfully
for the first time in long-span suspension bridge construction. A 10-mm-
diameter pilot rope made of light and strong aramid fiber was carried across
the approximately 2,000-m-wide navigation channel by a helicopter, with-
out interfering with navigation, through the use of a tension-controlled
unreeler developed for this operation. The procedure is shown in Fig. 7-7
(Yasuda and Takeno 1994; Yokoyama et al. 1995).
Since the erection of cables in long-span suspension bridges usually
takes a year or more, there is a high probability of being struck by typhoons,
which in Japan are very common. Instead of aerial spinning, which involves
the sequential laying out of individual parallel wires, the prefabricated
strand method was adopted in this project. Each prefabricated strand com-
prises dozens to more than 100 bundled wires. This method of erecting
bundled wires instead of individual ones had actually been invented in early
French suspension bridges long ago (Vicat 1831; Kawada 1981, 178). How-
ever, once John Roebling established the foundations of the modern aerial
spinning method, bundled wires became a remnant technology, although
its advantages were once again recognized in the Newport [Rhode Island]
Bridge built in the late 1960s (ENR 1968). Because no long-span suspen-
sion bridges had been built in North America for almost a half-century,
Japan improved upon this method. After it was first used in Japan on the
Kanmon Bridge (completed in 1973), the prefabricated strand method was
used for almost all of the Honshu-Shikoku bridges (excluding the Shimot-
sui-seto Bridge, the only exception where aerial spinning was used). It has
thus become the common erection method in Japan.
Figure 7-8 shows a cross section of the Akashi Kaikyo’s main cable
and its strand composition, and Fig. 7-9 shows the cables during erection.
It is clear that the prefabricated strand method is more efficient and effec-
tive against strong winds in comparison to the aerial spinning method.
This superiority is evidenced by comparisons of the output of erected cable
wires per day—23 tonnes in the Forth Bridge, 30 tonnes in the Severn
Bridge, and 53 tonnes in the Humber Bridge. With prefabricated strands,
85 tonnes were achieved daily in the Kanmon Bridge in its first application,
170 to 180 tonnes in the Kita and Minami Bisan-seto bridges, and a record
358 tonnes per day in the Akashi Kaikyo Bridge (Tarumi Construction
Office 1995).

Figure 7-7. Akashi Kaikyo Bridge: pilot rope crossing using a helicopter. PR, pilot rope.
Source: Yokoyama et al. (1995), with permission from Masanori Yokoyama.
Figure 7-8. Main cable and prefabricated strand of the Akashi Kaikyo Bridge.
For corrosion protection of the completed cables, a unique method of

dehumidification was adopted for the Akashi Kaikyo Bridge. In contrast to
existing methods of protection against water penetration using a covering of
coated wrapping wire, a unique method was developed that removes water
in the cable by forcing dried, dehumidified, and desalinized air into the inte-
rior spaces of the cables (Fig. 7-10) (Saeki and Furuya 1997; Shimomura
et al. 1998).
The Akashi Kaikyo Bridge was opened to traffic on April 5, 1998, its
1,991-m span the longest in the world. The bridge stands as a proud testa-
ment to decades of extensive research and development by Japanese engineers
since the opening of the 367-m-span Wakato suspension bridge in 1962.
Denmark’s Storebælt Bridge

Construction of the Storebælt Bridge in Denmark was inaugurated about
4 years after that of the Akashi Kaikyo Bridge in Japan. This bridge was to
be built over an international navigation route, the only route connecting
the Baltic Sea to the North Sea. To ensure navigational safety, the bridge
became a suspension bridge with a 1,624-m span, exceeding by 15% the
main span of the Humber Bridge, which until that time had enjoyed the title
of the world’s longest suspension bridge.1
Figure 7-9. Erection of cable by prefabricated strand method for the Akashi Kaikyo
Bridge.
Source: Nippon Steel (1996), with permission of Nippon Steel.
Figure 7-10. Cable corrosion protection system by dehumidification, Akashi Kaikyo

Bridge.
Source: Saeki and Furuya (1997), with permission from Honshu-Shikoku Bridge Expressway
Corporation.
Although the Akashi Kaikyo Bridge had already begun construction in

Japan and further developed traditional American suspension bridge tech-
nologies, the Storebælt Bridge followed the stream of aerofoil revolution
originating in Europe in the 1960s and first realized in the Severn Bridge
in Great Britain. Its contrasting concept can be easily understood from the
general plan shown in Fig. 7-11 (see also Fig. 7-12). Most notable is the
Storebælt Bridge’s adoption of a shallow, streamlined box girder instead of
traditional stiffening trusses.
As is already well known, the fall of the Tacoma Narrows Bridge in
1940 brought about the consideration of girder depth (h) as an important
characteristic of stiffening systems. Massive stiffening trusses replaced the
slender stiffening girder. Expressed as a ratio of (h) to span length (l), the h/l
of 1/209 in the Bronx-Whitestone Bridge and 1/350 in the Tacoma Narrows
Bridge reflected the last spans of their era with such slender proportions. As
a result of the Tacoma Narrows Bridge collapse, no truss-stiffened suspen-
sion bridges have since been built with an h/l ratio of less than 1/200. For
Figure 7-11. Storebælt suspension bridge: general plan. Construction began in 1991.
example, this ratio is 1/85 in the reconstructed Tacoma Narrows Bridge (1950),
1/100 in the Mackinac Bridge (1957), 1/178 in the Verrazano-Narrows
Bridge (1964), and 1/142 in Japan’s Akashi Kaikyo Bridge (1998). However,
the aerofoil revolution also significantly changed the trend in depth-to-span
ratios. The ratios are 1/324 for the Severn Bridge (1966) and 1/313 for
the Humber Bridge (1981). This adoption of slender stiffening box girders
produced ratios reminiscent of the era that spawned the first Tacoma Nar-
rows Bridge; the slenderness trend in h/l ratios finally peaked at 1/406 in the
Storebælt Bridge (1998).
The difference in depth-to-span ratios greatly influences the weight
of the stiffening girder. A comparison is shown in Table 7-1. (Note that a
simple comparison of total steel weight is not entirely relevant because the
suspended span lengths and lane numbers are different.) As a result of the
adoption of a streamlined box, the steel weight of the stiffening girder in
the Storebælt Bridge is just 56% of that in the Akashi Kaikyo Bridge. The
Storebælt’s stiffening girder is a welded closed box section. Its interior was
not painted because it is dehumidified by dry air injection (Gimsing 1998).
The diagonal hangers used in the Severn Bridge were not adopted; it is thus
believed that added mass can supplement the stiffness otherwise provided by
Figure 7-12. Suspension portion of the Storebælt Bridge.

Source: Author.
diagonal hangers (Kawada 1989; Ito et al. 1999). Other notable features of
the Storebælt Bridge that differ from traditional suspension bridges include
its use of concrete towers (adopted after the Humber Bridge), in addition
to a three-span continuous stiffening girder that had never been used in a
suspension bridge designed exclusively for road traffic.
Niels J. Gimsing, civil engineer and Professor Emeritus in the Depart-
ment of Structural Engineering and Materials at the Technical University
of Denmark, points out that the Storebælt Bridge’s approximately 2,700-m
Table 7-1. Comparisons of Steel Weights

Total Steel Weight for
Steel Weight per Deck Area Stiffening Girder
(kgf/m2) (tonnes)
Storebælt Bridge 424 29,680
Akashi Kaikyo Bridge 760 89,300
long stiffening girder is like a huge swinging log, and over its entire length
is suspended by hangers:
The highest concrete pylons are the 254 m high pylons of the Store-
bælt Bridge. Here the arrangement of struts is unique as only two struts
are found between the pier and the pylon top with the intermediate
strut positioned at mid height and not beneath the girder. So the girder
actually floats through the pylons without vertical support. This clearly
emphasizes the fact that the girder is actually hanging in the cable
system and does not have any vertical support directly on to the pylon.
(Gimsing 1999)
As the photographs of the towers of both bridges taken during construction

in Fig. 7-13 show, their contrast is remarkable.2 Employing a continuous
stiffening girder, the bridge could avoid the extreme stress increases in the
girder at the towers in a different manner from those road-railroad suspen-
sion bridges supported by the bearings at the towers, such as the Tagus
River Bridge and the Minami and Kita Bisan-seto bridges.
Details of the stiffening girder at the Storebælt Bridge tower are shown
in Fig. 7-14.
1. Side bearings that constrain transverse movement of the girder

2. Hydraulic buffers that restrain torsional vibration of the stiffening
girder, but not other transverse displacements. For this reason, the
hydraulic system is cross-coupled as shown in the figure.3
The anchorage design fixes the main cables at a point 60 m above the
sea, and is indeed a rational and smart design (Fig. 7-15). The portion below
sea level is composed of a gigantic 55-m-wide, 122-m-long, 16-m-high con-
crete caisson with a total weight of 50,000 tonnes that was built in a dry dock
and towed to the bridge site (A/S Storebæltforbindelsen 1995; Ito et al. 1999).
The anchorages form the only supports for the 2,700-m-long stiffening girder,
and here hydraulic buffers were installed as shown in Fig. 7-16. Spherical
bearings were also used to permit rotation at the end supports.
At the anchor blocks, a hydraulic damper arrangement allows for slow

horizontal movement up to ⫾1 m and free rotation of the girder. For
faster movements caused by asymmetrical and moving traffic loads, the
buffers provide a strong damping effect for the longitudinal movement
of the bridge girder. (Ostenfeld 1996)
(A)
Figure 7-13. Comparison of towers.

(A) Akashi Kaikyo Bridge; (B) Storebælt Bridge.
Source: Author.
(B)
Figure 7-13. (Continued)
Figure 7-14. Connection details between tower and stiffening girder, Storebælt Bridge.
Source: Gimsing (1998), with permission from A/S Storebæltsforbindelsen.
Figure 7-15. Storebælt Bridge anchorage.

Source: Gimsing (1998), with permission from A/S Storebæltsforbindelsen.
The main cables of the Storebælt Bridge were erected by the tradi-
tional aerial spinning method. However, instead of traditional sag adjust-
ments, an ingenious “controlled tension method” was adopted in which
the sag is determined under an applied constant tension of about 80%
that of the free-hanging wires. As a result, 149 tonnes of wire were erected
daily, thus establishing a record for aerial spinning (Gimsing 1999). In fact,
and as the late Blair Birdsall acknowledged, the world’s first use of the
controlled tension method was by Japanese engineers, and the method was
subsequently employed in the Shimotsui-seto Bridge (940-m span) (Bird-
sall and Celis 1997). With this new method, the sag adjustments that were
previously required twice—once during wire spinning and again during
strand forming—were now only required once, during strand forming, thus
greatly increasing efficiency.
Figure 7-17 shows the erection scheme of this method as adopted
in the Shimotsui-seto Bridge (Okukawa and Hirahara 1988). Only a few
suspension bridges have employed aerial spinning in Japan, and the Shim-
otsui-seto Bridge is the only one in the series of Honshu-Shikoku bridges
to do so. From such a rare example, a marvelous technical innovation
emerged, and in this case it contributed to the success of Denmark’s Store-
bælt Bridge.
Finally, let me describe wind problems associated with the Storebælt
Bridge. Its towers—the second tallest in a suspension bridge, whose 254-m
height is only exceeded by those of the Akashi Kaikyo Bridge—are made of
concrete, and therefore were thought to feature a high degree of stiffness.
On the contrary, vibration was observed in the fall of 1995 during strong
winds that struck the towers just after their completion. Vortex-induced
vibration occurred in winds with an average speed of 17 m/s (38 mph) and
the tower tops oscillated with an amplitude of approximately 200 mm. It
(A)
(B)
Figure 7-16. Details between anchorage and stiffening girder, Storebælt Bridge.
(A) plan view; (B) longitudinal section.
Source: Ostenfeld (1996), with permission from A/S Storebæltsforbindelsen.
Figure 7-17. Erection plan for the controlled tension method used on the Shimotsui-
seto Bridge in Japan.
Source: Okukawa and Hirahara (1988), with permission from Honshu-Shikoku Bridge
Expressway Corporation.
imposed no undue stresses and no corrective measures were taken because it

was expected that the vibration would not occur once the cables were fixed
in place (Gimsing 1999).
The most serious concern arose during the study of the early erection
stages of the stiffening girder. Erection was planned to start from the center
span with the lifting of large segments. However, it was feared that aero-
dynamic instability would occur during these early stages if the girders were
erected using conventional procedures. Figure 7-18 shows critical wind
velocities revealed by wind tunnel tests at several stages of girder erection:
10% to 11%, 18%, and 31% with corresponding eccentricity parameters.
The figure clearly indicates lower critical wind velocities during the early
Figure 7-18. Critical wind speeds and eccentricity of girder erection on the Höga
Kusten Bridge in Sweden.
The Höga Kusten Bridge is similar to the Storebælt Bridge and was built at about the same
time. The “0” in the abscissa means that the eccentricity of erection is zero and that erection
proceeds symmetrically.
Source: Tanaka et al. (1996), with permission from Hiroshi Tanaka.
stages of girder erection. Interestingly, by shifting the center of the erected

girders from the span center longitudinally by an amount 6, the critical wind
velocities were significantly increased (Tanaka et al. 1996; Tanaka 1998).
To be precise, Fig. 7-18 does not show the results for the Storebælt
Bridge but presents an experimental result from the Höga Kusten Bridge in
Sweden, a span completed in 1997 just before the Storebælt Bridge. As seen
from the plan in Fig. 7-19, it is obvious that the Höga Kusten is designed
with essentially the same concept as the Storebælt, except that its span length
and width differ slightly. (Its towers, like those of the Storebælt, were also
cast in concrete.) If the erection of the girder proceeded symmetrically from
the center of the span, aerodynamic stability was quite low until the point of
30% erection, as seen in the lower curve in Fig. 7-20. In contrast, if erection
proceeded asymmetrically as shown by A, B, C, D, . . . in Fig. 7-20, aerody-
Figure 7-19. General plan, Höga Kusten Bridge, Sweden.

Figure 7-20. Aerodynamic stability and asymmetric erection.

These data suggest that asymmetrical girder erection is superior to that of erection started con-
ventionally from the center of the main span. This method was employed on both the Storebælt
and Höga Kusten bridges.
namic stability was somewhat low up to a 10% deck erection stage, as the
upper curve in this figure indicates. Moreover, the degree of instability was
much less and the critical design wind velocity reached 60 m/s (134 mph) as
soon as an approximately 15% deck erection stage was reached. From the
above results, it became evident that the asymmetrical girder erection shown
in Fig. 7-20 was superior to that of erection started conventionally from the
center of the main span, and this was employed in both the Höga Kusten
and Storebælt bridges.
From June 4 to 7, 1998, with Denmark’s Queen Margrethe II in atten-
dance, the completion of the Storebælt Bridge was celebrated nationwide.
The bridge was opened on June 14, just 2 months after the opening of the
Akashi Kaikyo Bridge in Japan.
The End of the Century:

The Millennium Bridge
As an event to celebrate the 2,000 years since the birth of Christ, the opening
ceremony of the Millennium Bridge was held on June 10, 2000, in London.
This small pedestrian bridge is about 300 m long, with three spans of 80 m,
144 m, and 100 m, and is just 4 m wide. The design was selected from
among 229 entries in a widely advertised international competition held by
the British government. The final award was made to a team consisting of
world-renowned architect Lord Norman Foster, sculptor Sir Anthony Caro,
and Ove Arup and Partners, a structural design firm of prominence not
only in the United Kingdom but throughout the world. Foster described
the concept of this bridge as follows: “While the basic concept is that of a
suspension bridge, the objective was to push tried-tested techniques as far
as possible to create a uniquely thin bridge profile forming a slender blade
across the Thames” (Binney 2000).
The design was indeed light and innovative, as seen in Figs. 7-21 and
7-22. This impression was emphasized by the silver-white color that matched
the stainless steel railings; the designers proud description of the bridge as a
“blade of light” was indeed correct.
However, the Millennium Bridge suddenly began to oscillate follow-
ing the opening ceremonies, as people began to cross the span. The bridge
had never oscillated when inspectors crossed the span, nor had it done so
during relatively strong winds. The bridge’s behavior unnerved its users and
it had to be completely closed only 3 days after opening. According to the
Figure 7-21. Millennium Bridge, London, 2000.

Source: Author.
subsequent investigation, the oscillation at this time was a first asymmetric

lateral vibration with a node at the midpoint of the center span. The fre-
quency was about 1 Hz (Reina and Cho 2000). An example of this is shown
in Fig. 7-24—a pure lateral sinusoidal mode with a frequency of 0.9 Hz.
The first suspect in the suspension bridge’s vibration was wind, but
the bridge had never oscillated in wind. According to the executive direc-
tor of the Millennium Bridge Foundation, (the client for the bridge), “The
design was tested phenomenally carefully in three separate wind tunnel
tests, including a particularly elaborate one in Canada” (Binney 2000). His
comments acknowledged that because of the span’s innovative structure,
extreme attention had been paid to aerodynamic stability so that the bridge
was designed to never be prone to wind-induced vibration. However, the
bridge did in fact vibrate. But why did a bridge with no susceptibility to
aerodynamic instability vibrate?
The cause of the vibration in the Millennium Bridge was not wind, but
pedestrians—a factor that had been completely overlooked by the structural
engineers. As such, the bridge had swayed under a crowd of pedestrians
that rushed onto it just after the opening ceremony. According to Dr. John
Figure 7-22. Millennium Bridge: cable and bridge deck.

Source: Author.
Parker, an expert in vibration whose comments were published in New Civil

Engineer about 10 days following the closure of the Millennium Bridge:
A pedestrian’s center of gravity is about 1 m above his feet and as he

walks normally he puts his feet alternately about 100 mm each side of
a center line. Thus each pedestrian exerts a cyclic lateral force of about
8% of body weight at a frequency of around 1 Hz.
If a large number of people walk—not in step—across a bridge,
the resultant mean lateral force [H] would be the force from one pedes-
trian [h] multiplied by the square root of the total number [N] [that is,
H ⫽ h N ].
This would probably be enough to start the bridge swaying at the
critical frequency. At this point more and more pedestrians would find
it more comfortable to walk in phase with the movement, feeding more
and more lateral energy into the structure. (Parker 2000a; bracketed
material inserted by author)
It had never been considered that the thrust caused by pedestrians could
cause vibration in a bridge. James Fitzgerald, CEO of Ove Arup and Part-
ners, immediately responded to this observation by stating “There is nobody

who has ever measured such a force” (Parker 2000b).
However, there were examples of such pedestrian-caused vibrations in
bridges. Of particular note was a Japanese pedestrian bridge that in 1992

encountered vibrations similar to those of the Millennium Bridge on its
opening day. Yozo Fujino of the University of Tokyo investigated this vibra-
tion in detail, and he reported the results in the English-language journal
Earthquake Engineering and Structural Dynamics in 1993 (Fujino et al.
1993; Pacheco and Fujino 1993). The structure is a cable-stayed pedestrian
bridge as shown in plan view in Fig. 7-23. The bridge is situated between a
Figure 7-23. Bridge subject to lateral vibration from pedestrian movement.

(top) Side plan; (bottom) section of steel box girder.
Source: Fujino et al. (1993), with permission from Yozo Fujino.
motorboat racing park and a bus terminal, and once an event is finished the
bridge quickly becomes extremely crowded with people rushing to leave.
The vibration observed in the bridge was a first-mode lateral vibration with
a frequency of 0.9 Hz, as shown in Fig. 7-24. The source of the vibration
was a horizontal thrust periodically applied by the walking motion of pedes-
trians, as shown in Fig. 7-25.
When a person walks, periodic vertical forces are naturally generated, but
horizontal thrusts that are also applied at a frequency of about 1 Hz become
the driving force behind resonance when this frequency is close to the natural
frequency of the bridge. The conclusion of Fujino’s paper was extremely clear
concerning the cause of vibration in crowded pedestrian bridges:
First a small lateral motion is induced by the random lateral human

walking forces, and walking of some pedestrians is synchronized to the
girder motion. Then resonant force acts on the girder, consequently the
girder motion is increased. Walking of more pedestrians is synchro-
nized, increasing the lateral girder motion. In this sense, this vibration
has a self-excited nature. Of course, because of the adaptive nature of
human beings, the girder amplitude will not go to infinity and will reach
a steady-state. (Fujino et al. 1993)
Eventually, the cause of vibration in the Millennium Bridge became

clear. With the cause established, engineers worked to establish a strategy
Figure 7-24. Dominant lateral mode.

Figure 7-25. Pedestrian and mechanism of bridge vibration.

for retrofitting the span. One of the important priorities was to preserve
the slender aesthetics of the span devised by the design team. To counteract
vibration, the engineers decided to employ combinations of the following
measures:
• Install chevron bracing, and viscous and tuned mass dampers

beneath the deck (Fig. 7-26) (Russell 2001)
• Install diagonal bracing with viscous joints at the piers (Fig. 7-27)
(Parker 2000c)
• Install tie-downs (fixed to concrete blocks) with viscous dampers on
the side span at the south riverbank (Fig. 7-28) (Parker 2000d)
Again looking at the photos in Figs. 7-21 and 7-22, the structural pecu-
liarities of the span become more apparent. First, the sags of the four cables
stretched out along each side of the span are only 2.3 m at the center span.
Figure 7-26. Damping devices installed beneath the deck.

Source: Author.
Figure 7-27. Diagonal bracing.

Source: Parker (2000c), with permission from New Civil Engineer.
Figure 7-28. Viscous dampers on the south side span.

Source: Parker (2000d), with permission from New Civil Engineer.
With this, the Millennium Bridge’s sag ratio is a surprisingly low 1⬊62.6, in
contrast to about 1⬊8 to 1⬊15 in conventional suspension bridges. Further-
more, the cables usually support the stiffening girder via hangers (i.e., as
tension members) but in this bridge the transverse beams—acting as bend-
ing members—support the stiffening girder. Moreover, in the structures con-
necting the stiffening girders to the transverse beams, there are no diagonal
members for bracing.
In conclusion, the bridge was light and extremely low in resistance
against lateral bending. This state of affairs must entirely have been the
result of thinking that if only the wind problem was solved, oscillations
induced by lateral forces would not occur and the design would be suf-
ficient. The rehabilitation work against vibration in the Millennium Bridge
began in 2001 and the bridge was reopened on February 22, 2002. The
rehabilitated Millennium Bridge is shown in Fig. 7-29.
Arup’s project engineer, Roger Ridsdill-Smith, explained the effects of
the rehabilitation as follows:
On opening day the worst effects occurred when there were around 2,000
people on the bridge—that’s around 1.3 pedestrians per square meter.
At the quarter points of the main span the bridge was swaying
about 70 mm each way with alarming lateral accelerations of nearly
0.4g. After the retrofit we calculate that even with as many as two pedes-
trians per square meter the biggest deflection will be less than 20 mm
and the maximum lateral acceleration 0.04g. (Reina 2001)
Figure 7-29. Millennium Bridge with damper installed after rehabilitation.

Source: Author.
Notes
1. This span, called the East Bridge, is actually part of a much larger project
called the Great Belt (or Storebælt) Fixed Link. The 6,790-m bridge containing
the suspension span connects Zealand with the small island of Sprogø, and is
paralleled by an 8,000-m-long immersed rail tunnel; to the west, between Sprogø
and Funen, is a 6,611-m-long combined rail and road bridge.
2. A patent was taken out earlier in Japan that cited the advantages of a “swinging
log-type” continuous suspension bridge without intermediate support: Patent
No. 1096733, Continuous Suspension Bridge, Application Date: May 10, 1973;
Date of Notice to Public: May 28, 1981; Inventor: Kuranishi, Shigeru (then a
professor at Tohoku University).
3. As a matter of fact, the hydraulic buffers or dampers with a cross-coupled
system have never been installed in the Storebælt East Bridge. In April 2008,
Ernst Laursen, operations manager of A/S Storebælt, wrote to me as follows:
“. . . the hydraulic dampers were not installed at the pylons because the final wind
tunnel tests with a full bridge model revealed that the dampers would only give
a marginal improvement in the torsional restraint of the bridge girder. However,

the pylons and the bridge girder were prepared for future installation of hydraulic
pendulum if necessary, but now after 10 years in service we have seen no need for
pendulum installations.”
References
A/S Storebæltforbindelsen. (1995). Storebælt [in English], A/S Storebæltforbindelsen,
Copenhagen.
Binney, M. (2000). “Revolutionary design liable to give off bad vibration.” The
Times (London), June 12.
Birdsall, B., and Celis, J. (1997). “A bridge for the 21st century.” ASCE Civil Eng.,
October.
Engineering News-Record (ENR). (1968). “Shop-fabricated cable strands make
bridge history at Newport.” Engineering News-Record, June 13.
Fujino, Y., et al. (1993). “Synchronization of human walking observed during lateral
vibration of a congested pedestrian bridge.” Earthquake Eng. Struct. Dyn.,
22, 741–758.
Gimsing, N. J., ed. (1998). East Bridge, A/S Storebaeltsforbindelsen, Copenhagen.
Gimsing, N. J. (1999). “The Akashi Kaikyo Bridge and Storebaelt East Bridge [in
Japanese].” In Long Span Bridges and Aerodynamics, T. Miyata et al., eds.,
Springer-Verlag, Tokyo.
Bridges: Further Challenge of Engineers [in Japanese], Kensetsu Tosho, Japan.
Kawada, T. (1981). Cultural History of Suspension Bridges [in Japanese], Gihodo,
Japan.
Kawada, T. (1989). “Long-span suspension bridges again changed the trend.” [in
Japanese] Kawada Technical Report, Vol. 8.
Kurino, S. (1998). “Akashi Kaikyo Bridge.” Bridge and Foundation Engineering,
Japan, 32(8), August.
Miura, K., and Fujiwara, Y. (1996). “Suspension Bridges: Technology and Its Transition
[in Japanese].” Japan Society of Civil Engineers, pp. 113–114, December.
Nippon Steel (1996). Cable Construction for the Akashi Kaikyo Bridge [in Japanese],
Report to Honshu-Shikoku Bridge Authority, Nippon Steel and Kobe Steel.
Okukawa, A., and Hirahara, N. (1988). “Erection of cables for the Shimotsui-seto
Bridge.” [in Japanese] Honshi Technical Report, 12(45).
Ostenfeld, K. H. (1996). “Comparison between different structural solutions.
The Great Belt East.” Proc., 15th Congress IABSE, Structural Engineering
in Consideration of Economy, Environment and Energy, International
Association for Bridge and Structural Engineering, Zurich.
Pacheco, B., and Fujino, Y. (1993). “Keeping cables calm.” Civil Engineering, October.
Parker, D. (2000a). “Marching pedestrians blamed for bridge sway.” New Civil
Engineer, June 22.
Parker, D. (2000b). “Universities to investigate pedestrian forces on bridge.” New

Civil Engineer, June 29.
Parker, D. (2000c). “Arup picks double dampers to stop Millennium Bridge sway.”
New Civil Engineer, November 16.
Parker, D. (2000d). “Marching to a new tune.” New Civil Engineer, November 23,
12–13.
Reina, P. M. (2001). “Work begins to halt swaying of London’s pedestrian crossing.”
Engineering News-Record, March 5.
Reina, P., and Cho, A. (2000). “Spans sway underfoot in Europe.” Engineering
News-Record, July 10.
Russell, H. (2001). “End in sight for Millennium Bridge problems: final solution.”
Bridge Design & Eng., First Quarter, 9.
Saeki, S., and Furuya, K. (1997). “Corrosion protection for cables [in Japanese].”
Doboku Sekou, 38(7).
Shimomura, M., Sugiyama, T., and Hanai, T. (1998). “Corrosion protection system
for the Akashi Kaikyo Bridge’s cables.” [in Japanese] Honshi Technical Report,
22(86).
Shirley-Smith, H. (1967). “Supply and erection of the main superstructure.” In Forth
Road Bridge, Institution of Civil Engineers, London.
Tanaka, H., et al. (1996). “Aerodynamic stability of suspension bridge with partially
constructed bridge deck.” Proc., 15th Congress IABSE, Structural Engineering
in Consideration of Economy, Environment and Energy, International Association
for Bridge and Structural Engineering, Zurich.
Tanaka, H. (1998). “Aeroelastic stability of suspension bridge during erection.”
Struct. Eng. Intl., February.
Tarumi Construction Office. (1995). Report on Cable Construction for the Akashi
Kaikyo Bridge, First Construction Bureau, Honshu-Shikoku Bridge Authority,
Nippon Steel–Kobe Steel Joint Venture, December.
Vicat, L. J. (1831). “Pont suspendu en fil de fer sur le Rhône. Rapport au Conseil
d’État, Directeur Général des Ponts et Chaussées.” Annales des Ponts et
Chaussées, Vol. 1.
Yasuda, K., and Takeno, H. (1994). “Spanning pilot rope by helicopter for the
Akashi Kaikyo Bridge.” Bridge and Foundation Engineering, Japan, 28(6).
Yasuda, M., Fujii, Y., and Ueda, T. (1978). “Measure for aerodynamic stability for
a truss stiffened suspension bridge.” Proc., 33rd Ann. Conv., Japan Society of
Civil Engineers, Japan, 1–114.
Yokoyama, T., Shimizu, A., and Yamanashi, N. (1995). “Spanning pilot rope work by
helicopter for the Akashi Kaikyo Bridge.” Kawada Technical Report, Vol. 14.
8
Epilogue: No End
to the Dilemma
Development of Modern
Suspension Bridges and Stiffness
The 200-year history of the modern suspension bridge
has primarily been a struggle to harmonize the needs of
economy and stiffness and conquer the conflict between the
two. Even in primitive suspension bridges, ideas evolved
for stiffening that improved the walking comfort of pedes-
trians before the emergence of modern suspension struc-
tures. Examples are seen in the cable bracing of a suspen-
sion bridge built by African Pygmies as shown in Fig. 1-1
in Chapter 1, and in the vine “spider’s web” stay cables of
the Iya-Kazura Bridge in Tokushima Prefecture, Japan (Fig.
1-6). Clearly, these measures increased the vertical stiffness
of these suspension bridges and worked to resist torsion.
The Jacob’s Creek Bridge built by James Finley in
1801 was considered the first modern suspension bridge
with stiffening trusses that significantly increased vertical
stiffness (Fig. 1-19). For some reason, in the design concepts
of the modern suspension bridges that later spread to Eng-
land and France, an understanding of the role of the stiff-
ening truss regrettably became ambiguous, and massive,
unstiffened suspension bridges continued to be built. One
reason for this trend was that a working theory pertaining
to stiffened suspension bridges had not yet been developed.
As a result, cable theories based on catenaries and parabolas
predominated and unstiffened suspension bridges prevailed.
225
This was true with Louis Henri Navier in France as well as Thomas Telford
and Isambard Brunel in England—they all built the same types of structures.
For them, the object of consideration was vertical stiffness alone, and the
required attributes for this were mass and cable tension.

It is not surprising that early suspension bridges with low torsional
stiffness collapsed as the victims of wind-generated aerodynamic forces.
More than a few spans were destroyed by wind, such as the Brighton Chain
Pier (a four-span continuous suspension bridge), the Menai Bridge, and the
Wheeling Bridge built by Charles Ellet after his return to America. They
exhibited vibrations as prototypes of the later Tacoma Narrows Bridge, and
collapsed.
It was John Roebling who fully understood the significance of stiff-
ness in these early modern suspension bridges and demonstrated several
possible ways to enhance it. In his suspension bridge spanning the Niagara
River gorge he successfully proved that a train could be carried safely—an
idea that had been abandoned in the United Kingdom because of the per-
ceived unsuitability of suspension bridges for railroad use. Later, Roebling
designed the 486-m-span Brooklyn Bridge, the longest suspension bridge
of the nineteenth century. Roebling’s objective in this period was chiefly to
provide for bending stiffness, but it is not clear whether he had any idea of
torsional stiffness.
Roebling ensured a required level of stiffness with a combination
of three components: stiffening trusses, mass, and stay cables—but other
bridge engineers of those days disagreed on how to provide stiffness in sus-
pension bridges. For example, Gustav Lindenthal, a mentor to both Othmar
Ammann and D. B. Steinman, felt that cables should be stiffened as shown
in Figs. 4-8 and 4-13 (third from top) because he believed a suspension
bridge was simply an inverted arch. His idea was correct in the context of
the Elastic Theory as applied to the design of suspension bridges, but when
the Deflection Theory was developed by Leon Moisseiff and colleagues, the
Elastic Theory became outdated.
The Deflection Theory accelerated the lengthening of suspension
bridges, and asserted that the essence of the vertical stiffness in suspension
bridges lay in the mass of the cables and suspended structure. This theory
concluded that a stiffening truss is not necessary if there is enough mass,
which was not wrong so long as this concept was concerned with verti-
cal stiffness alone. Unfortunately, engineers such as Ammann and Moisseiff
forgot the damage to previous suspension bridges caused by wind, and were
not aware of torsional stiffness and torsional vibration. That ignorance is the
ultimate reason behind the disaster at the Tacoma Narrows in 1940. After
epilogue: no end to the dilemma 227
investigating the failure of the Tacoma Narrows Bridge, bridge engineers

better understood the need for torsional stiffness in suspension bridges. Sub-
sequently, plate girders of low torsional stiffness were avoided and the truss
became widely used in America as a structure with high torsional stiffness.

Beginning in the 1960s, Europe began to adopt the streamlined box girder.
In contrast to vertical and torsional stiffness, lateral stiffness had
been considered a minor issue because the truss and the deck framing were
thought to provide sufficient stiffness to resist lateral wind loads. However,
London’s Millennium Bridge, designed without any consideration of lateral
stiffness, suffered as a painful reminder of the retaliation of forces unac-
counted for.
Prospects for the Twenty-First Century

To conclude, I will endeavor to present the prospects for how the modern
suspension bridge will continue to evolve, in what ways it will be improved,
and what sorts of possibilities remain in the newly arrived twenty-first cen-
tury. Is it possible to further extend the approximately 2,000-m maximum
span of modern suspension bridges? If so, what will the new limit be?
One approach to answer this question is shown in Fig. 8-1. Here, the
limits of span length for three typical sag ratios, f/l ⫽ 1/12, 1/10, and 1/8,
are presented, with the weight of the cable shown as wc. The sum of the
dead load of the suspended structure to be supported by the main cables
and the live load is shown as ws, where the ratio of ws divided by wc is set
in abscissa; span length is represented on the ordinate.
As clearly shown in Fig. 8-1, when span length increases, the ratio
of ws/wc rapidly decreases, that is to say, the weight of the main cables
increases and the load of the suspended structure comparatively decreases.
When ws/wc is zero, the load of the suspended structure is nil, meaning a
condition whereby the main cables are free-hanging. This diagram indicates
that although the span limit varies with sag ratio, the limit of span is roughly
between 5,000 m and 7,000 m.
On the other hand, Fig. 8-1 clearly suggests that the load of the sus-
pended structure should be reduced as the span increases. However, since the
failure of the Tacoma Narrows Bridge, it has been recognized that torsional
stiffness is required in suspended structures to suppress torsional flutter. Iron-
ically, this torsional stiffness is reduced proportionally to the second order of
span length. As a result, according to J. R. Richardson (1981), steel weight
for the suspended structure should rapidly increase, as shown in Fig. 8-2.
Figure 8-1. Span limits in suspension bridges.

The span limit varies with the sag ratio. Sag ratios, f/l ⫽ 1/12, 1/10, and 1/8; wc, cable weight;
ws, sum of the dead load of the suspended structure to be supported by the main cables and
the live load.
Figure 8-2. Spans versus steel weights of stiffening girder.

b/c, deck width/deck depth ratio; UD, divergence speed.
Source: Adapted from Richardson (1981).
Figures 8-1 and 8-2 are difficult to reconcile since the former indicates
that a lighter suspended structure is preferred as span length increases, while
the latter reveals that the weight of the suspended structure increases with
span length. This contradiction has been a challenge for engineers who have
sought to further extend spans in long-span suspension bridges. A good
example is the Messina Strait Bridge in Italy, in which engineers have put
a 3,000-m span in sight for the first time. Figures 8-3 and 8-4 present the
prize-winning proposals selected in the 1969 international design competi-
tion. All plans reflected a difficult struggle to avoid increasing the weight of
the suspended structure, yet it is doubtful that these designs would have pro-
duced stable structures as expected. In fact, I once tried another approach
for a 3,000 m span suspension bridge, which was feasible for either a truss
or a streamlined box girder so far as their torsional stiffness was increased
to a certain degree.
Such bridges would reach a critical wind velocity only in major wind-
storms such as typhoons, and the arrival of such windstorms could be
accurately predicted with recent developments in weather forecasting tech-
nology. Figure 8-5 shows the critical wind velocity of a suspension bridge
Figure 8-3. Competition proposal for the Messina Strait Bridge by the A. Musmeci
Group.
Figure 8-4. Competition proposal for the Messina Strait Bridge by the A. Nervi
Group.
Figure 8-5. Effect of mass on critical wind speeds.

The line indicating 82 kgf/mm2 (120 ksi) is the allowable working stress adopted for the Akashi
Kaikyo Bridge. The vertical line indicating 107 kgf/mm2 (157 ksi) (on the right-hand side of the
figure) indicates the allowable working stress during bridge erection.
designed with hypothetical spans of 2,000 m, 2,500 m, and 3,000 m. The

abscissa is cable stress, and the vertical line indicating 82 kgf/mm2 (120 ksi)
is the allowable working stress adopted for the Akashi Kaikyo Bridge in
Japan. The vertical line indicating 107 kgf/mm2 (157 ksi) (on the right-hand
side of the figure) indicates the allowable working stress during bridge erec-
tion. This difference could be used for additional mass using water, as an
example, to raise the critical wind velocity by 17% to 18%. Conversely, the
amount of weight could be reduced in the initial design, thus making it eco-
nomical. The trial calculation revealed the reduction in dead load accounted
for 18% to 20% of the suspension bridge as a whole (Kawada 1995).
A twin deck suspension bridge proposal developed by J. R. Richard-
son through investigation and research for the Messina Strait crossing might
be the closest solution for the conflict between the further extension of span
length and reducing the weight of suspended structures (Richardson 1984).
A stiffened transverse beam separates the independent decks, and each deck
is suspended from cables as shown in conceptual form in Fig. 8-6. The result-
ing aerodynamic damping moment as shown in Fig. 8-7 would prevent the
emergence of flutter (Nakazaki 1998). The effect of ensuring aerodynamic
stability without a weight penalty is tremendous. Figure 8-8 compares the
construction cost-benefits with a conventional suspension bridge.
The Messina Strait Bridge will most probably be the first suspension
bridge that surpasses the span record of 3,000 m in the twenty-first cen-
tury. For that record-breaking span, the latest twin box suspension bridge
is proposed as the results of the recent developments of bridge engineering.
The cross section in Fig. 8-9 and the perspective view in Fig. 8-10 show the
proposed Messina Strait Bridge with its 3,300-m center span.
Figure 8-6. Twin box suspension bridge concept developed for the Messina Strait
Bridge.
Source: Richardson (1984), with permission from IABSE.
Figure 8-7. Aerodynamic damping moment of twin flat plates.

Source: Nakazaki (1998), with permission from Shunzo Nakazaki.
Figure 8-8. Economy of twin box suspension bridges.

Source: Brown (1999), with permission from Springer-Verlag GmbH.
Figure 8-9. Cross section of the Messina Strait Bridge.

Source: Stretto di Messina S.P.A. (1993), with permission from Kazuhiko Kawashima.
Lessons from History: Concluding Remarks

In this book I have described certain high points in the history of modern sus-
pension bridges. One motive behind this book is my belief that learning from
past history and the trials and errors of predecessors will in fact lead to new
developments in technology for the future. David Billington, director of the
Figure 8-10. Perspective view of the proposed Messina Strait Bridge.

Source: Brown (1999), with permission from Springer-Verlag GmbH.
Program in Architecture and Engineering at Princeton University, once stated:

“History, for structural engineers, is of an importance equal to science” (Bil-
lington 1977).
I have approached the history of modern suspension bridges in the

same manner. Etymologically, in Greek, history is historia, meaning “quest”
and “investigation.” Learning from predecessors’ mistakes in engineering is
very meaningful, but long time lags often render such experiences irretriev-
able. However, unless we utilize the invaluable lessons obtained from the
past, mistakes can be repeated. According to Henry Petroski, a civil engi-
neering professor at Duke University who specializes in failure analysis,
The more meaningful design information in case histories of success-

ful projects is: what failure criteria were employed, and what failure
avoidance strategies were incorporated. Nineteenth-century designers
like John Roebling were often quite explicit about such aspects of fail-
ure, and structurally daring suspension bridges like Niagara Gorge and
Brooklyn are testaments to the values of proactive failure analysis in
design. (Petroski 1994)
Petroski believes that the success of engineer John Roebling is attrib-

utable to the fact that Roebling learned from the history of past mistakes.
Roebling developed the suspension bridge as a modern, reliable structure
capable of safely carrying modern transportation modes, and initiated the
trend toward ever-lengthening of spans in the United States. Furthermore,
Petroski astutely points out some examples of mistakes that did not benefit
from history:
The apparent ignorance, displacement, or casualness of such a failure-

based principle, as exemplified by the apparent preoccupation of design-
ers a half-century after Roebling with such criteria as aesthetics and
economy, created a design climate that culminated in the collapse of the
Tacoma Narrows Bridge. (Petroski 1994)
Dr. Yotaro Hatamura is a Professor Emeritus at the University of

Tokyo and the founder, in 2002, of the Association for the Study of Fail-
ure. In his book Recommendations for Error Studies, which explains the
importance of learning from mistakes (including historical events), he states:
“Human beings learn from mistakes and deepen their thought” (Hatamura
2000). One of the important aspects of learning from history is clearly learn-
ing from mistakes. This does not simply mean to learn passively to avoid
making the same mistakes. For example, lessons from the Tacoma Narrows
Bridge spawned the new field of aerodynamics that has, without a doubt,
profoundly influenced subsequent modern suspension bridges. It has also
significantly contributed to other engineering disciplines.
Thus, the study of history does not only involve the examination of
mistakes. History, whether a mistake or success, is insufficient if it exists
only as knowledge. Each work or example is a product of its era, social
background, and demand. How one understands the way the result influ-
ences subsequent events is to grasp the large tide of dialectic development
and the “stream” of history. This is to say, in engineering as well as virtually
all other fields, the greatest significance of learning history is to establish
an historical vision. Without an understanding of the path of history, the
absence of an historical vision is akin to sailing the oceans without a navi-
gational chart. Therefore, great engineers and scholars in the past all learned
from history.
For example, Andrea Palladio (1508–1580), the greatest engineer
of the medieval period, studied Roman technology, and his knowledge
is encapsulated in his 1570 Quattro Libri dell’Architettura; Stephen P.
Timoshenko, a master of structural mechanics, wrote an historical master-
piece, History of Strength of Materials (1953); and David B. Steinman, one
of the great American bridge engineers, co-wrote a comprehensive history
entitled Bridges and Their Builders (1957). Many other engineers, such as
Fritz Leonhardt and David Billington, studied history and wrote enlighten-
ing history books.
The driving force propelling the development of suspension bridges has
been—and continues to be—the conquering of the conflict between “eco-
nomic efficiency,” which was inherited from primitive suspension bridges,
and the dearth of “stiffness,” which continues to threaten the safety of sus-
pension bridges even today.
References
Billington, D. R. (1977). “History and esthetics in suspension bridges.” J. Struct.
Div. ASCE, 103(8), 1655–1672.
Brown, W. C. (1999). “Long span projects: a personal view.” In Long-Span Bridges
and Aerodynamics, Toshiro Miyata et al., eds., Springer-Verlag, Heidelberg.
Hatamura, Y. (2000). Recommendations for Error Studies [in Japanese], Kodansha,
Japan.
Hirai, A., trans. (1972). Crossing over Strait of Messina [in Japanese], Honshu Shikoku
Bridge Authority, Japan.
Kawada, T., ed. (1987). Modern Suspension Bridges [in Japanese], Riko Tosho,
Japan.
Kawada, T. (1995). “Prospect of super long-span suspension bridge and a proposal

of suspension bridge with added mass in strong winds.” Kawada Technical
Report, 14, 1–5.
Nakazaki, S. (1998). “Why a twin-box suspension bridge is superior in aero-
dynamics.” Kyoryo, 34(12).

Petroski, H. (1994). Design Paradigms: Case History of Error and Judgment in
Engineering, Cambridge University Press, Cambridge, U.K.
Richardson, J. R. (1981). The Development of the Concept of the Twin Suspension
Bridge. National Maritime Institute, Feltham, U.K.
Richardson, J. R. (1984). “Aerodynamic stability of twin suspension bridge
concept.” Structural Engineering Today and Tomorrow, Proc., IABSE 12th
Cong., International Association of Bridge and Structural Engineers, Zurich.
Steinman, D. B., and Watson, S. R. (1957). Bridges and Their Builders, Dover Publ-
ications, New York.
Stretto di Messina S.P.A. (1993). “The design of the bridge over the strait of
Messina.”
Timoshenko, S. P. (1953). History of Strength of Materials, McGraw-Hill, New York;
reissued 1983, Dover Publications, New York.
Index
accidents/collapses, 226–27 Bronx-Whitestone Bridge, 129–30

construction casualties, 90–91, 113 George Washington Bridge, 114–16
error studies, 234–35 Golden Gate Bridge, 119–20
Finley’s bridges, 22–24 Mackinac Bridge, 154–55
French bridges, 58, 66, 77 reputation, 148, 155
government investigations, 140–46 Verrazano Narrows Bridge, 150–55,
Tacoma Narrows Bridge, ix, x, 39, 159, 193
131–34, 138–39 Andance Bridge, 54
U.K. bridges, 36–45, 64 aqueduct bridges, 83
Wheeling Bridge, 74–76 arch bridges, 1
wind vibration, 40–41, 64, 129–55, Argout, Compte d’, 52
226 artificial islands, 123–25
aerial spinning, 57–60, 81–82, 199,
210 Baird, Howard C., 111
aerodynamic stability, 146–55, 235 bar chain cable, 29, 30, 35, 46,
aerofoil box sections, 164–72, 195, 62, 77
203–4 Barlow, Peter, 62, 97
open grating decks, 164, 195 basket crossing, 10–14, 16–17
stiffening girder erection, 211–14, “Basket Crossing in Hida” (Hiroshige),
222n2 10, 13
twin box bridges, 231–33 Basse-Chaîne Bridge, 66, 77
wind tunnel testing, xiv, 135–37, Bear Mountain Bridge, 111
164–66 Beaucaire Tarascon Bridge, 54
aerofoil box sections, 164–72, 195, Becquey, Director General, 46–47, 49,
203–4 52
Akashi Kaikyo Bridge, ix, xiii–xiv, 1, Bender, Charles, 98
118, 193–201 the bends (decompression sickness),
allowable working stress, 230–31 90–91
cross-sections, 195–96 Berlin Royal Science and Engineering
steel weight, 205 School, 80–81
tower vibration, 197–99, 207 Bernoulli, Jacob, 61
Allegheny River Bridge, 82 Bernoulli, John, 61
Ambassador Bridge, 111, 113–14 Billington, David, 233–34, 235
Ammann, Othmar, 226–27 Birdsall, Blair, 209
accident investigations, 140–46 bolt connections, 195
237
Bosporus bridges. See First Bosporus coefficient of lift, 135–36

Bridge; Second Bosporus Bridge Coligny, Gaspard II de, 16
Bouch, Thomas, 92–94 concrete towers, 168, 197, 205,
box girder bridges, 43–45, 77 211–14
Brandywine Creek Bridge, 20 controlled tension method, 210–11

Bridges and Their Builders (Steinman Crawford, Ivan C., 148
and Watson), 235 Crét, Paul P., 111
Brighton Chain Pier, 30, 32, 36–39, critical wind velocity, 139, 229–31
226 Cumberland Bridge, 22
Brittania Tubular Bridge, 43–45, 77
Bronx-Whitestone Bridge, 111, damping, 168, 198–99, 206–9,
129–30, 143, 155, 203 218–22, 222nn2–3, 231–32
Brooklyn Bridge, 86–93, 98, 102–3, Danish bridges. See Storebælt Bridge
106, 108, 110–11, 127n3, 226 decks
caisson work, 90–91 aerofoil box sections, 164–72, 195,
stiffness, 91–93 203–4
Brown, Samuel, 28–31, 35, 39–40, 46, double-deck structures, 122
72, 77 open grating decks, 164, 195
Brownsville Bridges, 22 orthotropic decks, 195
Brunel, Isambard, 35, 226 primitive bridges, 5–7
Buck, Leffert L., 106–10 roadway form, 18–21
bundled wires, 199 stiffness, 18, 20
thermal change, 64
cable diameter, 112 two-cable bridges, 14–15
cable geometry, 60–65 See also railway bridges
cable sag, 64–65 Deer Island Bridge, 143–46
cable-stayed bridges, 1, 27–28, 94, 164 Deflection Theory, 97–106, 110–27,
aerodynamic stability, 146–55 131, 226
cable diameter, 112 Delaware River Bridge, 106, 111–13,
center diagonal stay failure, 138–40 127n2
vertical oscillations, 140–46 depth-to-span ratios, 131, 204
caisson work, 90–91, 121, 123–24 Derleth, Charles, 119
Calculation Method for Stiffened diagonal hangers, 168, 174–82,
Suspension Bridges (Léby), 99 190–92
cantilever truss bridges, 93, 122 Dieteleyn, J.F.W., 81
Caro, Anthony, 214 Dryburgh Abbey Bridges, 27–30
Castigliano, Alberto, 99 Dufour, Guillaume-Henri, 50
catastrophic vibration, 137–40
center diagonal stay failure, 138–40 early modern bridges
Chaley, M. J., 60, 73 French wire bridges, 46–60, 71
Cincinnati-Covington Bridge, 83, number completed per year, 53
85–87 theoretical aspects, 60–66
Clark, Adam, 33 U.K. bridges, 27–45
Clark, Tierney, 33 U.S. bridges, 18–24
Clericetti, Celeste, 98 earthquakes, 195–97
Clifton Bridge, 30, 35–38, 41, 80 École des Ponts et Chaussées, 62–63
Index 239
economic efficiency, xiii–xiv, 225, 235 girders, 64, 78–80, 97

Elastic Theory, 98–104, 106–10, 226 box girders, 43–45, 77
Ellet, Charles, Jr., 71–80, 94n1, 226 Deflection Theory, 97–98
error studies, 234–35 plate girders, 129–30, 141
Essex-Merrimac Bridge, 20–24 stiffening girders, 206–9, 222nn2–3

Euler, Leonard, 61–62 wind vibration, 136–37
eyebar chain cable, 29, 30, 46, 62, 77 Godard, M. T., 99
Golden Gate Bridge, 111, 117–21,
Fairbairn, William, 43 129, 153, 198
fairings, 135–36, 165 government investigations, 140–46
Farquharson, F. B., 133, 135–37, 165
Fatih Sultan Mehmet Bridge, 187–91, Hanshin earthquake, 195–97
191n3 Hatamura, Yotaro, 234–35
Faxian Sanzan, 2–4 helicopters, 198–200
Finch, J. K., 107–8 Hirai, Atsushi, 178, 183–91
Finley, James, x, 19–24, 64–65, 94n2, Hiroshige, Ando, 10, 13
130, 225 historical overview, 1–24
First Bosporus Bridge, 168–70, early modern bridges in France,
172–73, 187–90, 191n3 46–60
Fitzgerald, James, 217 early modern bridges in the UK,
flat eyebar chain cable, 29, 30, 46, 62, 27–45
77 early modern bridges in the US,
Forth Railway Bridge, 93, 174 18–24
Forth Road Bridge, 159–61, 166, 170, primitive bridges, 1–8
199 Renaissance-era bridges, 16–18
Foster, Norman, 214 rigid suspension bridges, 71–94
Fourgues Bridge, 54, 66 twentieth-century bridges, 106–27
Freeman Fox and Partners, 169, 175, The History of Antiquities of the
187 County Platine of Durham
French bridges (Hutchinson), 16–17
early modern bridges, 46–55 History of Strength of Materials
École des Ponts et Chaussées, 62–63 (Timoshenko), 235
wire cable and anchorage systems, “History of Suspension Bridges in
54–60, 71, 75 Bibliographical Form” (Jakkula),
Fribourg Bridge, 60–61, 71, 73 52
Fujino, Yozo, 217 Hodgkinson, Eaton, 43
Fuss, Nicolas, 61 Höga Kusten Bridge, 212–13
Hokusai, Katsushika, 14–16
Galashields Bridge, 27 Homberg, Hellmut, 174–75
Galileo Galilei, 60–61 Honda, Katsuichi, 2
George Washington Bridge, 111, Honshu-Shikoku project, ix, 92, 118,
113–17, 129, 150–53, 197 122, 199
Gies, Joseph, 13–14 Hook, Robert, 60
Gilbert, Cass, 113–14 Hopkins, H. J., 52
Gilbert, Davies, 62 Humber Bridge, 168–74, 190–91, 193,
Gimsing, Niels J., 205–6 199, 204
concrete towers, 168, 197, 205 Lehigh Chain Bridge, 21–22, 24

steel weight, 170 Lehigh Gap Bridge, 24
Hutchinson, William, 16–17 length (span length), 1, 64, 227–33
hydraulic buffers, 206–9, 222n3 aerodynamic stability, 146–55
Deflection Theory, 110–27, 131,

Innoshima Bridge, 194 226–27
inverted arches, 108–10 Leonhardt, Fritz, 165–66, 235
Investigation and Research Report lift, 135–36
of Suspension Bridges (Navier), limited vibration, 137
46–48, 63–65, 97 Lindenthal, Gustav, 98, 108–9, 114,
iron, 18–24, 62 226
iron wire. See wire cable
Iya-Kazura Bridge, 7, 225 Machinae Novae (Verantius), 16–17,
27
Jacob’s Creek Bridge, 19–20, 22, 24, Mackinac Bridge, 111, 148–54, 159,
102, 225 193, 197
Jakkula, A. A., 52 Manhattan Bridge, 104–5, 108–12
Japan’s bridges, ix, 193–95 Manual of Applied Mechanics
Akashi Kaikyo Bridge, ix, xiii–xiv, (Rankine), 97
118, 193–201 “Manufacturing Method of Wire
earthquakes and typhoons, 195–97, Rope,” 81–82
199 masonry arch bridges, 1
Honshu-Shikoku project, ix, 92, mass, 182–91
118, 122, 199 materials, xiv
pedestrian bridges, 217 iron, 18–24, 62
primitive bridges, 7, 225 iron wire, 49, 65–66
Jones, Jonathan, 113 primitive bridges, 8–9, 14, 18
steel, 108–9, 170
Kanmon Bridge, 194, 199 Mélan, Joseph, 99, 101–2, 104–5
Kármán vortex, 135–37, 141 Menai Bridge, 30–34, 40–43, 46, 62,
Kawada, Takaki, ix–x, 183–91 64, 226
Kawakita, Jiro, 9–10 Messina Strait Bridge, 229–33
Kazura Bridge, 7, 225 Methodus Inveniendi Lineas Curvas
Keefer, Samuel, 80 (Euler), 61
Kemp, Emory, 94n1, 102–3 Millennium Bridge, x, xiv, 214–22,
Kentucky River Bridge, 20 227
King’s Meadows Bridge, 27–28 Minami Bisan-seto Bridge, 194–95,
Kita Bisan-seto Bridge, 194–95, 199, 199, 206
206 Modjeski, Ralph, 105–6, 111, 113,
123
Lackawaxen Bridge, 83 Moisseiff, Leon S.
La Hire, Philippe de, 60–61 Deflection Theory, 102–6, 110,
Lánchíd Chaine Bridge, 33–35 226
Lausen, Ernst, 222n3 Delaware River Bridge, 111, 113
Le Blanc, Nicolas, 58–59 George Washington Bridge, 114–16
Léby, Maurice, 99 Golden Gate Bridge, 119–20
Index 241
Mackinac Bridge, 148 Potomac Bridge, 20, 22, 72

Tacoma Narrows Bridge, 131, prefabricated strand cables, 58–59,
141–42, 145–46 199–203
Monongahela Bridge, 82, 84 primitive bridges, 1–8, 225
Montrose Bridge, 39, 67n2 single-cable bridges, 9–14

Moran, Daniel E., 111, 113–14, two-cable bridges, 13–15
122–23 U-section bridges, 7
More Exact Theory, 101–2, 104–5 vine bridges, 2–4, 7
Müller-Breslau, H., 98–99 V-section bridges, 5–7
multiple-domed caisson, 123–24 Provis, W. A., 40–41
Navier, Louis Henri, 46–48, 52, Quattro Libri dell’Architettura

62–65, 97, 226 (Palladio), 235
Neshammy Creek Bridge, 20 Québec Bridge, 105–6
Newburyport Bridge, 20–24 Queenston-Lewiston Bridge, 80
Newport Bridge, 199
New York Bridge Company, 88 railway bridges, 39–40
Niagara Railroad Bridge, 73–74, box girder bridges, 43–45, 77
77–80, 106, 226 Niagara Railroad Bridge, 73–74,
Normandy Bridge, 1 77–80
Northampton Bridge, 21 Tagus River Bridge, 162–63
Williamsburg Bridge, 106–12
Ohio River Bridge, 83, 85–87 Rainbow Bridge, 80
Ohnaruto Bridge, 194 Rankine, William, 97–98, 99
Old Tacoma Narrows Bridge, 111 Recommendations for Error Studies
open grating decks, 164, 195 (Hatamura), 234–35
orthotropic decks, 195 Rendel, J. M., 67n2
Ove Arup and Partners, 214 resonance, 191n2
Richardson, J. R., 227–28, 231
Palladio, Andrea, 235 Ridsdill-Smith, Roger, 221–22
parallel-wire strands, 56–57, 81–82 “Rigidity and Aerodynamic Stability of
Parker, John, 216–17 Suspension Bridges” (Steinman),
pedestrian vibration, x, xiv, 214–22, 148
227 rigid suspension bridges, 71–94
Petroski, Henry, 234 Brooklyn Bridge, 86–93, 102–3,
plate girders, 129–30, 141 110–11
pneumatic caisson method, 123–24 Ellet’s bridges, 71–76, 94n1
Poncelet, J. V., 66 Niagara Railroad Bridge, 73–74,
Pons Ferreus, 16–17 77–80, 106
Pons Unius Funis, 16–17 Roebling’s bridges, 76–93
Pont des Invalides, 47–48, 52 Wheeling Bridge, 72–76, 94nn1–2
Ponte de 25 April. See Tagus River Ritter, Wilhelm, 98
Bridge Roberts, Gilbert, 166–72
Pope, Thomas, 20–21, 27 Robinson, Holton, 143
Port Authority of New York and New Roche-Bernard Bridge, 58–59
Jersey, 114 Roebling, Emily, 90–91
Roebling, John, 74, 80–86, 234 span length. See length

Brooklyn Bridge, 86–90, 98, 102–3, spider webs, xiii–xiv
106, 108, 110–11 St. Andiol Bridge, 54–55
Niagara Railroad Bridge, 74, 77–80, stay cables, 76, 78–79, 94, 164
106 steel tower heights, 197–99

statue, 83, 86 steel weights, 170, 205–6, 227–29
stiffening methods, 98, 226 Steinman, David B., 143, 226, 235
Wheeling Bridge, 76 Mackinac Bridge, 148–51, 154, 159,
wire rope, 81–82, 199 193
Roebling, Washington A., 89–93 Tagus River Bridge, 162, 169
Roman bridges, 1 Stephenson, Robert, 28, 41–43, 77
rope bridges, 4, 8–14 stiffness, xiv, 18, 40–45, 91–93, 203–6,
rope crossing, 10–14 225, 235
rope dancing, 9–14 aerofoil box sections, 165–72, 195,
Runcorn Bridge, 62 203–4
Russell, J. S., 36–39 Deflection Theory, 102–6, 110–27,
131, 226
sag adjustments, 210 Elastic Theory, 98–104, 106–10, 226
See also wire cable mass, 182–91
sag ratios, 227–28 pedestrian vibration, x, xiv, 214–22,
Saikai Bridge, 194 227
Saint-Antoine Bridge, 50 railings and decks, 65–66
Saint-Christophe Bridge, 58–59, 66 railroad bridges, 43–45, 78–79
Saint-Vallier Bridge, 50 torsional stiffness, 227–28
Salazar Bridge. See Tagus River Bridge tower stays, 94, 106, 110
San Francisco-Oakland Bay Bridge, See also accidents/collapses; girders;
111, 118–19, 122–27, 141 trusses
Schuylkill Falls Bridge, 20, 22 Storebælt Bridge, xiii, 1, 193, 201–14,
Schuylkill River Bridge, 72 222n1
Second Bosporus Bridge, 187–90, steel weight, 205
191n3 stiffening girders, 206–9, 222nn2–3
Seguin, Camille, 48–50, 65–66 Strauss, Joseph Baermann, 118–21
Seguin, Marc, 48–52, 65–66 stringers, 195
Serrell, Edward W., 39–40, 80 structural damping, 168
Serrières Bridge, 54 “Suspension Bridge at the Border of
Seto Ohashi Bridge, 118, 122 Hietsu” (Hokusai), 14–16
Severn Bridge, 159, 166–72, 199, 204 Suspension Bridges of Any Desired De-
diagonal hanger failures, 174–78, gree of Stiffness (Bender), 98
190–91 Széchenyi Bridge, 30, 33–35
repair costs, 182
steel weight, 170 Tables for Computing All the Cir-
tower weakening, 178–82 cumstances of Strains, Stress of
Shimotsui-seto Bridge, 194, 199, Suspension Bridges (Gilbert), 62
209–211 Tacoma Narrows Bridge, ix, x, 39,
Shutraub, H., 63 131–34, 226–27
single-cable bridges, 9–14 center diagonal stay failure,
Index 243
138–40 twin box bridges, 231–33

investigation, 140–42 two-cable bridges, 13–15
reconstruction, 146–48 typhoons, 195, 199
stiffening trusses, 203–4
Tagus River Bridge, 159, 161–66, U.K. bridges

169–70, 191n1, 206 accidents/collapses, 36–45
Tain Bridge, 53 aerodynamic advances, 159–91
Takano, Jun, 8 early modern bridges, 27–35
Tatara Bridge, 1 Millennium Bridge, x, xiv, 214–22,
Tay Bridge, 93 227
Tees Railway Bridge, 39–40 Union Bridge, 29–31, 46, 72
Telford, Thomas, 30–33, 35, 40–43, U.S. bridges
46, 61–63, 226 accidents/collapses, 22–24, 131–46
tension/balance of economy and stiff- early modern bridges, 18–22
ness, x rigid suspension bridges, 71–94
Theory of Chain Stiffened by Beam twentieth-century bridges, 106–27
(Müller-Breslau), 98 U-section bridges, 7
Theory of Equilibrium of Elastic
Structures and its Application Vallance Bridge, 54
(Castigliano), 99 Verantius, 16–17, 27
The Theory of Modern American Sus- Verrazano-Narrows Bridge, 111, 159,
pension Bridges (Clericetti), 98 193
Theory of Steel Arches and Suspension aerodynamic stability, 150–55, 190
Bridges (Mélan), 99 steel weight, 170
thermal change, 64 tower height, 198
Thousand Island Bridges, 143–46 vertical oscillations, 140–46
Timoshenko, Stephen P., 235 vibration. See pedestrian vibration;
torsional stiffness, 227–28 wind vibration
Tournon Bridge, 50–52, 53 Vicat, Louis-Joseph, 52–60
towers Vienne Bridge, 53
concrete, 168, 197, 205, 211–14 vine bridges, 2–4, 7
damping, 168, 198–99, 206–9 viscous damping, 218–22
steel tower heights, 197–98 von Kármán, Theodore, 141
stiffness, 94, 106, 110, 206–9 V-section bridges, 5–7
weakening, 178–82
wind vibration, 197–99 Wakato Bridge, 194
tower stays, 94, 106, 110 weight. See steel weights
transverse struts, 34 welded connections, 195
A Treatise on Bridge Architecture Wernwag, Louis, 82
(Pope), 27 Wex, Bernard, 175
Trinity Chain Pier, 30, 32 Weyer, Edward, 2
trusses, 64, 76, 78–80, 88, 97 Wheeling Bridge, 72–76, 94nn1–2, 226
Deflection Theory, 97–98 Whitestone Bridge. See Bronx-Whites-
stiffening trusses, 164, 203–6 tone Bridge
wind vibration, 136 width-span ratios, 131
tubular bridges, 43–45 Williamsburg Bridge, 106–12, 127n3
Winch Bridge, 16–18, 27 torsional flutter, 227–29

wind tunnel testing, xiv, 135–37, tower vibration, 197–99, 211–14
164–66 verticle bending modes, 133–34
wind vibration, 40–41, 64, 129–55, wire cable, 48–49, 52, 65–66
226 aerial spinning, 57–60, 81–82, 199,

aerodynamic stability, 146–55 210
catastrophic vibration, 137–40 bundled wires, 199
critical wind velocity, 139, 229–31 controlled tension method, 210–11
diagonal hanger failure, 174–82, corrosion protection, 201, 203
190–91 French-style bridges, 54–60, 71, 75
Kármán vortex, 135–37 prefabricated strand cables, 58–59,
limited vibration, 137 199–203
mass, 182–91 wire rope, 81–82
steel weights, 170, 205–6, 227–29 Woodruff, Glenn B., 141, 148
Storebælt Bridge, 211–14
Tacoma Narrows Bridge, 131–34, Younghusband, Francis, 2–4
138–39
About the Author,

Translator, and Editor
TADAKI KAWADA, PH.D., (author) is currently on the board

of directors of Kawada Industries, headquartered in Tokyo,
Japan; previously he served as president (1977–1996) and
chief executive officer (1996–2005). During his career at
Kawada Industries, he was involved in many noteworthy
bridge projects, including the Honshu-Shikoku suspen-
sion bridge, the Tatara cable-stayed bridge, and the third
Tacoma Narrows suspension bridge. He earned his under-
graduate degree in French from the Tokyo University of
Foreign Languages in 1958, studied as a research student
at the Bridge Research Laboratory of the University of
Tokyo for 4 years, and received a doctorate in engineering
from the University of Tokyo in 2002. He is the author of
many articles and books, including Who Wrecked Gallop-
ing Gertie? The Mystery of the Tacoma Narrows Bridge
Disaster (Kensetsu Tosho, 1975) and A Cultural History of
Suspension Bridges (Gihodo, 1985), both in Japanese.
HARUKAZU OHASHI, PH.D., (translator) is currently execu-

tive director of Nippon Engineering Consultants Co., Ltd.,
Tokyo. Previously he held positions with the Honshu-
Shikoku Bridge Authority in Japan and with Parsons Cor-
poration (into which the well-known bridge engineering
firm Steinman, Boynton, Gronquist & Birdsall was absorbed)
in New York City. His career as a structural engineer has
245
focused on the design, construction, and maintenance of long-span bridges,

including the world’s longest single span, the Akashi Kaikyo Bridge; the
world’s only triple suspension bridge, the Kurushima Bridge; and the Mes-
sina Strait Bridge in Italy, which will supersede the Akashi Kaikyo Bridge as
the longest single span. He earned a bachelor’s degree in 1980 and a mas-
ter’s degree in 1981, both in civil engineering from the University of Michi-
gan. He earned his doctorate from the Tokyo Institute of Technology, where
his research focused on fatigue issues in orthotropic steel deck suspension
bridges. He has published technical papers on long-span bridge design and
fatigue and translated several books on bridges and structural engineering
into Japanese and English.
RICHARD SCOTT, M.E.S., (editor) is currently a waterway heritage planner

with Parks Canada, where he is responsible for planning along the Trent-
Severn Waterway, site of the world’s tallest hydraulic lift lock. Previously,
he was senior regional planner with the National Capital Commission in
Ottawa, Canada, where the 1996 Greenbelt Master Plan—for which he
was project manager—garnered national awards from both planning and
landscape architecture communities. He earned a bachelor’s in land use
planning from Ryerson University and a master’s in environmental studies
from York University, both in Toronto, and he has written several papers on
planning. He also holds a long-standing passion for bridges and authored
In the Wake of Tacoma: Suspension Bridges and the Quest for Aerody-
namic Stability (ASCE Press, 2001), the Japanese-language translation of
which was awarded the Japan Association for Wind Engineering 2005 out-
standing publication prize.

Kawada 2010

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Kawada 2010

Uploaded by

Copyright:

Available Formats

aded from ascelibrary.org by New York University on 05/10/15. Copyright ASCE.

For personal use only; all rights r

Other Titles of Interest

Beyond Failure: Forensic Case Studies for Civil Engineers, by Norbert J.

Bridginess: More of the Civil Engineering Life, by Brian Brenner. (ASCE

Engineering Legends: Great American Civil Engineers, by Richard G. Wein-

Landmark American Bridges, by Eric DeLony. (ASCE Press, 1993). Presents

Washington Roebling’s Father: A Memoir of John A. Roebling, edited by

History of the Modern

Kawada, Tadaki, 1935-

Tadaki Kawada ; translated by Harukazu Ohashi ; edited by Richard Scott.

Published by American Society of Civil Engineers

Copyright © 2010 by the American Society of Civil Engineers.

preface to the english edition . . . . . . . . . . . . . . . . ix

1 The Emergence of Modern

2 Trials and Errors in Europe . . . . . . . . . . . . . . . . . 27

3 North American Engineers Develop the

The Niagara Railway Suspension Bridge 77

4 Emergence of the Deﬂection Theory

5 Ambush: The Dynamics of Wind . . . . . . . . . . . . 129

6 New Developments beyond North America . . . . 159

7 Two Record-Breaking Suspension Bridges . . . . . 193

8 Epilogue: No End to the Dilemma . . . . . . . . . . . 225

Prospects for the Twenty-First Century 227

This page intentionally left blank

Preface to the English Edition

Japan, from humble forays into the field in the

Dr. Kawada’s narrative is enriched by two other qualities. His passion

Harukazu Ohashi, translator, Tokyo, Japan

This page intentionally left blank

This book is intended to contribute to the future

This book was started as a doctoral thesis and was

The Universality of Primitive

Edward Weyer described a suspension bridge built by an African Pygmy

Keeping to the range, the party journeyed on in a south-westerly direc-

The route that Sanzan traveled is considered similar to the route

Source: Weyer (1959).

Figure 1-2. A suspension bridge built by primitive people in New Guinea.

Figure 1-3. Uginba Village and a suspension bridge.

We are surprised to observe the similarities between ancient bridge

Figure 1-4. Suspension bridge near Askori Village, China.

Figure 1-6. The Kazura Bridge in Iya Village, Japan.

Suspension Bridges Were Unknown to

thought that “rope dancing” or “rope crossing” spans consisting of just a

Incidentally, when we arrived at the village there was something that

under its own weight.

Figure 1-9. Rope dancing.

Figure 1-10. Rope crossing.

Source: Juan and Ulloa (1750).

Figure 1-12. Basket crossing as seen in a ukiyo-e painting.

Figure 1-13. Evolution of primitive suspension bridges.

Figure 1-14. An old-type suspension bridge.

Figure 1-15. “Suspension Bridge at the Border of Hietsu,” painted by Hokusai

ple of primitive suspension bridges in Japan. It is a two-cable suspension

the bridge, and drew it by developing a concept only from second-hand

Figure 1-16. Basket crossing, as shown by Verantius.

Figure 1-17. A suspension bridge, as shown by Verantius.

Figure 1-18. The Winch Bridge in Britain (1741).

The Birth of the Modern Suspension Bridge

The ﬁrst suspension bridge that conquered these problems inherent

born in 1756 in Ireland and emigrated to America at a young age. In 1770

This description is cited from a letter Finley addressed to the U.S.

deformations in the main cables. Technically, Finley understood the effect

ers. . . . It consists of a single arch of 240 ft [73 m] span and 40 ft [12 m] at