UNSW

l1111111111111111 1111111111 111111111111111111111111111111 11111111

>013694880
Building:
-•
Rt and made many suggestions f
ack down muc obscure material.

UNSW

LIBRARY
 6 introduction
9 chapter 1 Building and Engineering in Ancient Times: 1000 B.C. - A.D. 500
61 chapter 2 The Medieval Era: 500 - 1400
113 chapter 3 The Renaissance: 1400 - 1630
175 chapter 4 Global Trading and the Age of Reason and Enlightenment: 1630 - 1750
233 chapter 5 Engineering Becomes a Profession: 1750 - 1800
289 chapter 6 The Harmony of Theory and Practice: 1800 - 1860
365 chapter 7 The Birth of the Modern Building: 1860 - 1920
447 chapter 8 Architectural Engineering: 1920 - 1960
535 chapter 9 The Computer Age and the Greening of Construction: 1960 - present
61 O afterword
611 glossary
613 appendix 1 "Aims and Means;' by Ove Arup
615 appendix 2 Iron and its Alloys
615 appendix 3 Concrete and Reinforced Concrete
620 endnotes
622 bibliography
627 further reading
633 index
639 illustration credits

~ -~ . '. . . . ... .;.:-__ ;,;- :· :-.·'!"""·

f; \jt-/ ?". ':.:t: ·.U'

)
; '?Jfr ~~l~ i\ ir -
1879-1962 Eugene Freyssinet 1890-1969 Owen Williams 1899-1961 Eduardo Torroja
1879-1959 Walter Bauersfeld 1891-1979 Gio Ponti 1899-1990 Fred Severud
1880-1968 Arturo Danusso 1891-1979 Pier Luigi Nervi 1900-55 Bernard Lafaille
1883-1963 Karl von Terzaghi 1893-1963 A. A. Griffith 1901-85 John Baker
1885-1959 Hardy Cross 1895-1983 R. Buckminster Fuller 1901-84 Jean Prouve
1886-1969 Ludwig Mies van der Rohe 1895-1988 Ove Arup 1903-98 Alfred Pugsley
1887-1965 Le Corbusier 1897-1976 Lydik Jacobsen 1929-19305 The Great Depression

19105-19205 Scientific management and Gantt charts 19305 Understanding of ductile behavior
of metals
19205 Understanding of brittle fracture From c.1925
Welding of structural steel 19305-19505 Understanding of plastic
19205 Concept of "effective stress"
behavior of steel structures

1919 Newcomen Society for the history 1930 First congress of International
engineering and technology formed {UK Association of Bridge and
Structural Engineering {IABSE)
1921 Building Research Station (UK) op ned
19305 First learned societies for histor
of engineering and technology

19205 Model tests used by Dischinger for design of concrete shells 19305 Model tests of wide-span concrete roof
structure by Eduardo Torroja and Pier Luigi Nervi
1923 Zeiss-Dywidag system patented
1927 Boundary layer wind tunnel 1930 Wind-tunnel testing of model of
testing of wind loads on buildings Empire State Building

1930 Model tests of soil loaded in a

centrifuge conducted by Karl Terzhagi

19305 Scale model tests for room acoustics

(using microphones)

19305 Acoustic and electrical resistance strain

gauge developed

1930-31 Model testing of dynamic building

response to earthquakes

1931 Heliodon devised to demonstrate effects

of sunlight

c.1920 Relaxation method of calculatio developed for airship structures

19205 Spirit duplication/ Hectograph ethod of copying drawings
1922 "Comfort zone " chart developed mid-19305 Moment distribution metho
introduced by Hardy Cross
c. 1920-19705 Four-function electro-me , hanical calculators
From 1920s Ray tracing used to help d sign concert hall acoustics

1921-23 Concrete shell airship 1924-25 Bauhaus building, 1930-32 Boots factory "Wets"
hangars at Orly, France Dessau, Germany Building, Beeston , England
1922 Church of Le Notre Dame, 1935 Front6n Recoletos, Madrid, Spain
Le Rainey, France
1924-25 Schroder House,
Utrecht, Netherlands 1931 Empire State Building, New York, USA
1923 Prototype Zeiss dome,
1936 Aircraft hangar, Orvieto, Italy
Jena, Germany
1927 Salle Pleyel, Paris, France 1936 Simpsons, Picadilly, London, UK
1924 "Gesolei" exhibition buildings,
Dusseldorf, Germany 1933 First hyperbolic paraboloid shell
 1902-59 Felix Samuely
1902-89 Nicolas Esquillan
1903-79 Hubert Rusch
1903-88 Max Mengeringhausen
1909-2002 John Blume
1910-97 Felix Candela

b. 1925 Frei Otto

1939-45 World War II

1950s Major structural use of aluminum

1940s Glued, laminated timber ("Glulam") developed 1952 Plastic design of steel frames
mid-1940s Ferro-cement developed by Nervi 1952 Use of glass for
structural elements
1945 Development of glass-reinforced polymers

1939 First edition of Principles of Moder

Building by the Department of Scientific and
1958 First issue of Technology and Cul re
Industrial Research (UK)
published by the Society for the Histor
of Technology (USA)

1930s-1940s MERO system developed by Max Mengeringhausen

1950s Photoelastic stress analysis of models

Mid-1950s Scale models used in

acoustic design

1950 Curta Calculator II, 10-figure hand! eld mechanical calculator

From mld·1950s Mainframe computer

used for structural analysis

1949-54 Hunstanton School , Hunstanton, England

1951 Skylon , Festival of Britain, London (prestressed cable structure)
1952 Raleigh Livestock Arena , Raleigh, 1955 Music pavilion, Kassel, Germany
North Carolina (cable net roof) (doubly-<:urved membrane)
1956-58 CNIT Building, Paris, France
1952 Lever House, New York, USA 1955-57 Inland Steel Building,

. . . . . . ..
• r. Chicago, USA

1940 1945 1950 1955 1960

Architectural Engineering
1920-1960

The advances in high-rise building during the 1880s and constructing steel-frame buildings in the United States .
1890s in the United States were made possible by cer- And it was as a direct result of this close collaboration
tain developers recognizing that their real estate dreams with engineers that a few architects of genius would , in-
would be realized not by architects working alone, but by deed, find ways to forge their identity on steel-frame
architects working in collaboration with engineers. buildings-for example, Louis Sullivan 's Guaranty
Compared to the traditional, European Beaux Arts idea of Building in Buffalo , New York (1895), and in New York
the architect 's role, this generally meant the architect City, Cass Gilbert's Woolworth Building (1913), William
had a smaller part to play. Indeed, it was difficult for ar- Van Allen 's Chrysler Building (1930), and Mies van der
chitects to find ways of expressing their individuality in Rohe 's Seagram Building (1958). In Europe , the first
multistory iron- or steel-frame commercia l office build- building that truly celebrated steel in its architecture-
ings or hotels. In the extreme, the role of the architect the Centre Pompidou in Paris, designed by architects
could be little more than providing a decorative facade Ren zo Piano and Richard Rogers in collaboration with
and lavish interiors in the foyer and other public areas. the engineering firm of Ove Arup-was not completed
until 1977.
Engineers were prominent in many of the early building
design firms in the United States . In New York, George Two remarkable features of steel-frame construction
B. Post had trained as a civil engineer before turning to were how quickly the height of buildings grew and how
building design , and in Chicago William Le Baron Jenney quickly the technology spread through North America and
was an engineer who employed both architects and en- to many large cities of South America , South Africa ,
gineers in his firm. Other practices soon learned the Australia, and the Far East. By the second decade of the
benefits of collaboration between architect and engi- twentieth century. twenty-story steel-frame buildings with
neer, and several of the most successful had one or heating and mechanical ventilation had become part of
more partners who had trained as an engineer-Adler of everyday construction practice. Built in 1931, the Empire
Adler & Sullivan , Root of Burnham & Root , and both State Building, with its 102 stories, would remain the
partners of Holabird and Roche. It was these firms' en- world 's tallest building for nearly four decades. How was sJa, 540
gineering innovations that were so attractive to develop- it that such progress could be achieved in such a short
ers in helping reduce the cost and increase the speed of span of time? Surprisingly, perhaps , the answer has little
 [Link]'l ADY WORKS .)IA 'III~ E

PART FRA:\fES

PUR ORD j SKETCH j

PAT . OR CARO OR. No.
I- - - -
OPER A TION
:l========I=====,==
Rt: c 'o
I PLANED
I l, OTTEU DrtlLl.'E ll A , Elt'n
---
T O BE BEGUN I
- -- -
TO B E FINIS H ED
I I I
----
NUM BER WANTED 15 15 15 15 15

NUMB\~~;INISIE: I DA :Y ITO: AL DA:Y ITO:AL DAILY I TOT AL DAILY I TOTAL DAILY :TO TA

- - - - - 121 I 2 I 4 I I I I I ,--1- - 1 - -
1 22 I I I 2 I 4 I I I I I I
- - - ---'-1- 2a · I 1 I 5 I - --1_ 1 _ 7 _ 1 - 1 - -
1 1-
12-1 1 2 1 7 1 1 1 51 3 1 3 1 I I I
I 26 I 4 I 11 I 2 I 7 I I I 4 I I I I
1 ~1 I I I I 1 -- 1 -- 1 --1 - - 1 -
I 28 I 1 I 12 I ., I 9 I 3 I 7 I 2 I 2 I I
I 29 I 2 I 14 I 1 I 10 I t I 8 I 1 I 3 l - - 1 -
,- -- - ----'l'---,3-0---'c-l- 1----;l----,-Vi I I I 11 I 2 I 10 I 1 I 4 1- - I =
I ::I I I I I cl I 1 1 I I I 11 I 1 I 5 I I
FEB I 2 I I I I I 1 I 12 I 1 I ti I 2 I 2
I 3 I I I 1 I 15 1 1 I 13 I 3 I o I 1 I :1
I 4 I I I I I I I - I 11 I ,, :;
I b I I I I I I I •J I 1:3 I I I Ii
1--0 1-- I __ I I I 2 I 15 I I I ,) I ~
- I 7 1--1 - I --1 I I I ., I 15 I 1 \)
---I 9 1--1--1 - I I I I 2 I 11
53 8 I I
I 10 I I I I~
I 11 I -,- - - - I 1:1
I 12 I I , - -1-,
-------
- I H I I I
-_ _ _-clc--::1-::-1--:-_-_
-_ =-c I 1.;

,_
_ _ _ _ _ _
I rn I
- I 17 I
1_
I

1- -
,,. •· •••·l
!
j-
HU'1l1
I" :
I
-1--1---- ll 1:to 1:1J .,- .\tTLu.

Frc. 20n.
539

540

538 Empire State Building, New York, 1931 . Chief engineer: Andrew J. Eken of Starret Bros. and Eken Inc. Structural
engineer: H. G. Balcom and Associates ; architect: Shreve, Lamb and Harmon. Under construction in July 1930. 539 Chart
by Henry Gantt, first developed c.1918, to help plan the effective allocation of manufacturing operations to the various
machines in a workshop. 540 Empire State Building.
 to do with building engineering. A seemingly offhand but World War I, served as visual tools for comparing the tar-
extraordinary statement by the American building histo- geted and the actual progress of projects . After the war,
rian Carl Condit highlights a remarkable feature of engi- the charts were soon adopted to help the organization
neering in general: and management of any large project, and were used
For all its height, the design of the steel framin g [for the with particular success in the construction of many dams
Empire State building] was a straightforward Job of deter- in the United States during the 1920s.
mining the sizes of members adequate to meet buckling
and wind loads . ... 1 Applying such tools in the management of large construc-
tion projects enabled both the construction time and the
Following certain breakthroughs, enormous progress can cost of such projects to be dramatically reduced. This , in
be made in a relatively short time simply by exploiting the turn, enabled real estate developers to propose ever
new opportunities. Neither the Empire State Building nor larger buildings in the secure expectation of higher prof-
dozens of other buildings constructed in the first half of its . The construction statistics for many of the 1930s
the twentieth century were particularly innovative . They skyscrapers in America are truly astonishing. The steel
were designed and built largely using construction tech- frame of the Empire State Building was built at an aver-
niques, design methods , and engineering science that age speed of four and a half stories per week, for a total
had been established by the 1870s. Most of the develop- of just 410 days for the entire building.
ments during the following sixty years or so were in the
field of what would be called, in manufacturing industries , Taller buildings were, however, made possible not only by
" production engineering "-improving the quality of the the application of principles of scientific management,
materials used, the methods for their manufacture and but also as the result of a breakthrough in the under-
assembly, and, not least, managing the whole process . standing of the behavior of soil under load . Taller build-
ings were heavier and needed more substantial
American engineer Frederick Taylor (1856-1915) pub- foundations. Engineering scientists began studying soil
lished his paper "The Principles of Scientific Manage- as a material, in the same way they studied the behavior
ment " in 1911 while working for the Bethlehem Steel of iron or timber. During the early decades of the twenti-
company on the construction of warships for the U.S. eth century a wider transformation was underway that
government. His rational analysis of the activities touched the theoretical and experimental study of all ma-
needed to complete manufacturing tasks revolutionized terials . Following the progress made in physics in under-
the process of organizing production. Taylor 's principles standing the atomic structure of matter, scientists
were rapidly adopted in many countries , in part studying all types of material began asking a more funda-
because of the rush to manufacture armaments during mental question: Why did soil (and other materials) have
World War I. They were championed in the construction the strength and stiffness that it did? Seeking and finding
industry by the building contractor Frank Gilbreth explanations for the properties of materials became a
(1868- 1924) , who, to some extent, gave them a bad new discipline known as "materials science ," and led to
name through the overenthusiastic use of stopwatches a revolution in the design and manufacture or construe-
to time the activities of construction site workers down tion of all man-made artifacts.
to the last second. By the mid-1920s American trade
unions had already ensured that such extremes were Materials Science, Soil Mechanics
on the decline . The load-bearing capacity of soils is, both metaphorically
and literally, the foundation of all building. The science of
One of the most useful tools to come out of these early soil mechanics had first been developed in the eigh-
539 days of modern management was the "Gantt Chart. " teenth century, especially by Charles Augustin de
Henry Gantt (1861-1919) was a mechanical engineer Coulomb, and continued in the nineteenth century, espe-
and management consultant who had worked for many cially by William Rankine. Their main concern had been
years under Frederick Taylor in the American defense in- the stability of retaining walls and the faces or slopes of
dustry. Gantt charts , which were first used at the end of soil newly created or exposed when excavating the
 ground to build foundations for buildings and bridge Karl Terzaghi (1883- 1963) was the engineering giant
piers , digging cuttings , and building embankments for who found the solution to this problem. Born in Prague,
fortifications , dams , canals , and railways. This work, he studied mechanical engineering at the Technische
however, had relatively little impact on the design of ac- Hochschule in Graz , Austria, before joining the army.
tual building foundations , which , even into the twentieth During this time he spent two years in Saint Petersburg
century, had not changed significantly since Roman and designing reinforced concrete structures; he went on to
medieval times. write his doctoral thesis on reinforced concrete. In 1914
he began a two-year visit to the United States, where he
Foundations were of two types: a footing made of timber, worked on the design and construction of a number of
stone , concrete , or steel and concrete that distributed large dams , and it was during this time that he began to
the load over a sufficiently large area; or, in soft soils, a turn his professional attention to his childhood passion
series of piles driven so deeply that they either rested on for geology. Terzaghi was struck by the fact that he was
rock beneath the soil or were prevented from sinking by not able to account for the behavior of engineering struc-
541 , 542 friction between the pile and the soil. The design proce- tures and their foundations in terms of the geology of the
543 dures for both these types were empirical , and took little soils themselves. Thus he began developing the principle
account of any scientific material properties of the soil. that would guide his work on soil mechanics for the next
"According to some of the best authorities ," said the six- forty years: The key to understanding the behavior of soil
teenth edition of Rankine's classic Manual of Civil is the careful and meticulous observation of its behavior,
Engineering (1887), "the test of a pile 's having been suf- both in the field and in laboratory experiments . On his re-
ficiently driven is that it shall not be driven more than one- turn to Austria in 1916 the Austrian government asked
fifth of an inch by thirty blows of a ram weighing 800 him to accept a teaching post at the Imperial School of
pounds and falling 5 feet at each blow " (italics in the orig- Engineers in Constantinople (now Istanbul ). During his
inal ). 2 The same textbook advised that footings on firm nine years in that post he undertook the theoretical and
ground , including sand , gravel, and hard clay, should be experimental work that established soil mechanics as a
made large enough to ensure that the pressure on the truly scientific discipline and culminated , in 1925 , with
earth was between 2 ,500 and 3 ,500 pounds per square th e publication of Erdbaumechanik auf bodenphysikalis-
foot (125- 175 kN / m 2 ) . Footings on soft earth should be cher Grund/age (The Mechanics of Earth Construction
built on sand or concrete to a depth that took account of Based on the Physical Properties of Soil ). This book laid
the angle of repose of the earth , which reflected the abil- the foundations for the modern subjects of soil mechan-
ity of the soil to resist shear forces . This was a minor ics and geotechnical engineering. After 1925 he spent
adaptation of the well-established design procedures for four years at the Massachusetts Institute of Technology
earth embankments. in Cambridge , Massachusetts , where he introduced and
developed the science of soil mechanics . In 1929 he re-
The main preoccupation of foundation engineers was not turned to Austria as a professor at the Technical
to avoid settlement, since this was considered inevitable, University of Vienna , which soon became the world center
but to ensure that each separate foundation settled at for engineers interested in geotechnical engineering. In
the same rate and , eventually, by the same amount. This 1936 the International Society of Soil Mechanics and
was an intelligent aim and could theoretically be achieved Foundation Engineering was founded, and Terzaghi was
by adjusting the area of each footing in proportion to the elected as its first president , a position he held for more
load it must carry, thus creating the same pressure on the than two decades , until the age of seventy-four.
soil everywhere beneath the building. However, this aim
depended on the soil having identical properties across Terzaghi 's insight was , in hindsight, simple: He recog-
the entire building footprint , an ideal situation that is sel- nized the vital role played by the water content of soil in
dom found in practice. Measurements taken in 1892 on a the soil 's ability to carry loads. Yet this insight brought
number of buildings in Chicago , including the Auditorium about a revolution in understanding the behavior of soil.
Building , showed that differential settlements of up to When water fills the voids between soil particles , the
450 mm were not uncommon. total stress at any point comprises two components: the
pressure in the water, that is, the " pore-water pressure ," graphical construction of an orthogonal grid of flow lines
and the stresses in the matrix of particles , the "effective and lines of equal pore-water pressure , it is possible to
stress. " The presence of water reduces the forces be- calculate the pressure gradient in the soil and, hence ,
tween individual soil particles and , hence , reduces the the rate at which water flows through soil. Thus any po-
force needed to cause the particles to shear past one tential adverse consequences of water flowing beneath
another. Expressed in the language of Coulomb's con- buildings or dams can be predicted , and suitable means
cept of internal friction (see chapter 6 , p. 244) , this devised for controlling the water flow and avoiding the 544

means that the presence of water in soil reduces the in- problems . Terzaghi was always conscious of the risk to a
ternal friction and , hence , its shear strength and its abil- project, and hence to the client's purse , that resulted
ity to carry loads. While this conclusion was the most from the incomplete knowledge of the nature of the soils
radical and far-reaching consequence of Terzaghi 's work, beneath a construction site . Such a lack of knowledge al-
his observational approach to the behavior of soils was ways compromises the engineer's ability to determine an
hardly less significant. economical and safe design solution before construction
has started. In order to avoid the overdesign of struc-
For virtually the first time in any building engineering dis- tures in the ground , Terzaghi insisted on a thorough site
cipline , an experimental approach to the subject was es- investigation and soil analysis before construction
tablished that put it on a par with the great scientific work began , and continuous observation of the behavior of
of German and British physicists and chemists of the the soil as construction progressed , to enable the de-
late nineteenth century, in whose tradition Terzaghi had signs to be modified to reflect the true soil conditions as
been steeped as a young engineer. He was driven by the they were revealed . Thus he was able to save the client
goal of devising a set of material properties that would money if the initial assumptions about the soil proved to
explain the curious , and often apparently inexplicable, be too conservative , and to avoid disaster if the assump-
behavior of soils when subject to loads . He felt it should tions were found to have overestimated strength or stiff-
be possible to predict the behavior of soils as reliably as ness of the soil.
that of other materials. Thus in addition to inventing the
concept of effective stress , he devised or formalized all Terzaghi ' s worldwide fame was a result of the rigor,
the tests that are today undertaken in site investigations thoroughness , and reliability of his approach, which
to establish the properties of a soil that allow its load- began with understanding the geology of a site , fol-
bearing behavio r to be predicted with confidence. lowed by analysis of geotechnical conditions , structural
Because the presence or absence of water is so impor- considerations , and , often , hydrological issu es . His ad-
tant , many of these tests relate to the propensity of a vice was sought on every type of construction work . The
soil to allow water to flow through it, that is , its perme- technical reports he wrote for his clients were highly re-
ability, and the factors that might affect permeability. garded both for their quality and their forthright nature ,
These include the degree of "consolidation " of the soil- and for many decades they served as models for other
the amount by which the sizes of voids between soil par- geotechnical engineers .
ticles have been reduced , for example , by the action of
water when the soil was formed, or increased , as a result Aircraft Design and Structural Engineering
of disruption by landslides , earthquakes, or excavations. During the early decades of the twentieth century, struc-
He thus demonstrated the need to place great emphasis tural engineering benefited considerably from research
on the history of a soil , both over geological time and in undertaken in the aircraft industry, especially the search
its more recent past. for lightweight structures for airships and planes. Much
of this research involved the study of metals and how it
To model the movement of water underground, Terzaghi might be possible to improve their strength-to-weight ra-
promoted the use of "flow nets ," which had been devised tio . This , in turn, was made possible by a number of re-
in about 1900 by an Austrian hydraulic engineer, Philipp cent developments in physics. In about 1910 the
Forchheimer (1852-1933) to visualize in two and three properties of materials were at the forefront of research
dimensions how water flows through soils . Through the in physics. In about 1910 the German physicist Max von
 542

Impermeable

' -~-f--+--~--;..--- '

,' :' : : : : : I; ; : \ \ \ •.,
d»r//7/ , /;/)Jl?JJJ);;;;7);;;;;171T//J);;;';;;;);;);;;;;;;7;?77//?JJl?l};TJ?T/?T//?77/
lmperviovs slrcdum
544

543 545

541 Victoria Station, London, 1900. Timber piles used for extension works. 542 Victoria Station. Mass concrete founda-
tions used for extension works. 543 Early twentieth-century foundations. (left) Typical Hennebique pad foundation ,
c. 1907; (right) pile made using the Considere system, c. 1908. 544 Flow net showing flow of water in soil beneath a
sheet pile (top) and a dam (bottom). 545 Lines of stress cracks with different root radii; the smaller radius causes
greater stress concentration.
Laue (1879-1960) had discovered that X-rays (identified culated using the known strength of bonds between indi-
by Wilhelm Rontgen in 1895) are diffracted (scattered) by vidual atoms or molecules . Second , for very fine thick-
the atoms of a crystal in ways that depend on the struc- nesses , below a tenth of a millimeter or so, the strength
ture of the crystal. A few years later, the British scientist of the fibers increased markedly as the diameter de-
William Henry Bragg (1862- 1942) and his son William creased. Finally, the strength of thin fibers decreased
Lawrence used the knowledge of X-ray diffraction to de- with the age of the fiber. The explanation for these phe-
duce the precise nature of crystal structures , and devel- nomena led to perhaps the most profound change of the
oped the technique known as "X-ray crystallography." twentieth century in our understanding of fracture and
Laue and both Braggs were awarded Nobel Prizes for their the strength of materials .
work, which, for the first time, enabled scientists to look
at the properties of materials at the atomic scale . A host For some decades it had been recognized that stresses in
of scientists worldwide was soon engaged in exploring materials under load are raised around holes or at the
the chemical and physical consequences of the new end of cracks in the material. Such areas are called
theories and discoveries about the structure of matter. "stress concentrations. " In 1913 C.E. Inglis , an engineer- 545

One line of inquiry was of particular concern to all design ing professor at Cambridge University, had demonstrated
engineers : analysis of the physical properties of mat- that the stress concentration factor increases signifi-
ter-including strength , stiffness, hardness , brittleness, cantly as the radius of curvature of the hole, or crack tip,
ductility, metal fatigue , and creep-in terms of atomic decreases . A practical consequence of this phenomenon
structure. It proved to be an arduous task , one that was is that the strength of a structure is effectively reduced if
completed only in the second half of the twentieth cen- cracks causing high stress concentrations are present.
tury, but its significance cannot be overstated . Today the Griffith extended the work of Inglis by considering not only
explanation of materials properties in terms of atomic the elastic energy stored in a strained material , but also
structure forms the core of university courses in nearly the energy needed to create new surfaces in a crack. He
every engineering discipline . showed that if less strain energy is released from an ex-
tension to a crack than would be required to create the
Two scientists whose contribution was of particular sig- new surfaces , then the crack does not grow or spread .
nificance for structural engineering were Alan Arnold However, if the strain energy released is greater than
Griffith (1893- 1963), a materials scientist at the Royal needed to form new surfaces , then the crack does grow
Aircraft Establishment (RAE ) at Farnborough in England , and spread- not just gradually, but even explosively.
and Geoffrey Ingram Taylor (1886- 1975), a scientist Since this explanation does not involve the radius of the
who worked for much of his life at Cambridge University crack tip , it directly relates the stress that a material can
in England . bear to the length (actually the square root of the length )
of the longest crack in the material , and also to two mate-
The Work of A. A. Griffith and G. I. Taylor rial properties- the "modulus of elasticity " and the "free
Much of the research on the strength of materials under- surface energy." Thus , for a given stress , a material will
taken after World War I at the RAE was done on "ordinary " fracture if the longest crack is equal to or greater than a
materials such as metals and timber. However, these are critical crack length, now universally known as the
complex, nonhomogeneous materials. A. A. Griffith de- "Griffith crack length. "
cided to use glass as a "model " material-a material on
which to conduct experiments and obtain results that This understanding of the conditions necessary for a
might then be applied to the " real " materials such as tim- crack to propagate enabled Griffith to explain the curious
ber and metals . The idea of using small-scale model results of his tests on glass fibers. When the fibers are
tests was familiar to aeronautical engineers for aerody- freshly drawn, their surfaces are undamaged and there
namic tests in wind tunnels . Using thin fibers of glass, are no cracks that could begin to propagate. Older fibers
Griffith noted three curious and unexplained facts. First , invariably become scratched, and these scratches act
the tensile strength of a single fiber of glass was about as microscopic cracks . If the energy ba lance at the tip
one hundred times less than the theoretical strength cal- of such a microcrack is such that the microcrack length
exceeds the Griffith crack length, then the crack will prop- The most significant impact of Griffith 's work for building
agate and the fiber will fracture. engineers was that it alerted them to the potential dan-
gers of any stress concentration that might become the
"Griffith crack theory, " as it is known , quickly became starting point for a fatigue crack and might eventually
invaluable to forensic engineers seeking the causes of lengthen to the critical Griffith crack length , resulting in
engineering disasters later in the twentieth century, disastrous brittle fracture . This new insight into the be-
many of which were found to involve the brittle fracture havior of materials led , in the 1920s , to the reassess-
of metals. Furthermore, these metals are often ductile ment of the detailed design of all steel structures. This
at normal temperatures and become brittle at low tern- reassessment focused on all holes cut in steel , espe-
peratures (typically below about -10°C). The most cially the bolted connections between columns and
famous of these forensic inquiries concerned the beams in steel-frame buildings , which required large
world 's first jet-powered airliner, the Comet , which suf- numbers of holes set close together.
fered a series of unexplained crashes in the early
1950s , apparently arising from midair explosions. It was The 1920s also saw the introduction of both gas and
finally determined that the stresses in the aircraft 's electric arc welding as a means of joining steel members.
fuselage were such that the diagonals of the near-rec- At first , welding seemed an ideal way to create a fully rigid
tangular windows were fractionally short of the Griffith connection between steel members while avoiding the
crack length . Also , the fuselage material, an aluminum large numbers of holes that created stress concentra-
alloy, was ductile at ordinary temperatures and brittle at tions. Welding, however, brought its own difficulties . One
the air temperature at the cruising altitude . When was that the heating process often changed the crys-
fatigue cracks just a few millimete rs long formed at the talline structure of steel from one that allowed ductile de-
corners of some of the windows , as tends to occur in all formation to one that is very brittle. In addition , the
materials subject to repeated changes of stress , the welding process could artificially introduce discontinu-
aperture increased to the critical Griffith crack length . ities inside the steel , and these could behave exactly like
The conditions were then such that the crack propa- cracks at sites of stress concentrations and ideal start-
gated with the explosive speed and violence of a balloon ing points for Griffith cracks .
bursting. (This analogy is apt. Griffith 's theory explains
why a small hole in a partially inflated balloon merely These difficulties led to important research programs in
causes it to deflate , while a hole of the same si ze in a Germany and the United States , where electric arc weld-
fully inflated and highly stressed balloon causes it to ing was actively developed in the 1920s in both the ship-
burst .) Since that time , aircraft windows have been building and building industries . Not only was it
made significantly smaller than th e Griffith crack length . necessary to improve the welding process to eliminate
They are also more oval-shaped to reduce the level of cracks , it was also necessary to develop methods of
stress concentration where a fatigue crack would tend nondestructive testing to verify both that welded con-
to form. nections penetrated the steel sections fully, providing
a full-strength connection , and that there were no discon-
The implications of Griffith 's work for engineers have been tinuities or cracks that might be the starting points for
enormous. Perhaps most significant has been the devel- brittle fracture.
opment of a whole family of new materials known as fiber-
reinforced composite materials, or "composites. " These One of the first welded steel structures used in a building
use very high-strength fibers of glass , carbon, aramids , was a 12-meter roof truss erected at the factory of the
and other materials embedded in a matrix usually made Electric Welding Company of America, in Brooklyn , New
of a polymer resin . The hulls of most small boats are today York , in 1920. Welding was also used to fix to the steel
made of fiberglass , and tennis rackets are often made columns the large brackets that supported the rails of an
from carbon-fiber composites. Composites are used in the overhead traveling crane in the factory. All the welded
building industry, too , in the construction of cladding connections were fully load-tested to prove their
panels, for example. strength. For several years the use of electric arc welding
546

547

546 Westinghouse Electric and Manufacturing Company building, Sharon, Pennsylvania, 1926. 547 International Silver
Company factory, Meriden, Connecticut, 1937. Exposed welded steel frame .
 grew and spread, but only as a substitute for some of the cerned with why different materials had the strength they
riveted connections; it did not result in changes to the did . Taylor 's particular interest was to explain why it be-
overall design of the structures. Only in 1926 was the comes progressively more difficult for metals to "yield "
first fully welded frame structure constructed: a five-story as they are bent back and forth. This phenomenon, which
building at the Westinghouse Electric and Manufacturing can be observed by bending a steel paper clip, is known
546 Company in Sharon, Pennsylvania. In this building the as "work hardening. " Indeed , how is it that a crystalline
beams and columns, themselves built up from flat plates material deforms in a ductile or plastic way at all? The ex-
by welding, were welded to form a continuous, homoge- planation Taylor proposed in 1934 was that the ductile
neous frame. The structure achieved the necessary rigid- behavior of metals is caused by imperfections, which he
ity to resist gravity and wind loads using about 12 percent called "dislocations ," in an otherwise regular crystal
less steel than a conventional riveted frame- 790 rather structure. These dislocations could move relatively freely 549

than 900 tons. Before the building was erected, full-size by the minor rearrangement of bonds between atoms,
prototypes of the key connection were built and tested to without the bonds having to be broken.
destruction to verify their stiffness and strength. Over the
next four years the Westinghouse company constructed Taken together, the work of Griffith and Taylor, and of
twenty-four welded frame buildings including, in 1930, D'Arcy Thompson (see chapter 7, pp. 443 , 446), trans-
the eleven-story Central Engineering Laboratory Building formed engineers ' understanding of the strength of
at its Pittsburgh factory. In 1931 the tallest welded frame materials and made it possible to raise the art of struc-
of the time-the nineteen-story Dallas Power and Light tural design to an altogether higher scientific plane. In
Building-was completed. During the 1930s welded the case of buildings, however, this new knowledge was
steel frames were widely used for industrial buildings , pri- not always applied. Aircraft and aircraft engines had
marily because they required less steel , but designers always been designed to the limit of a material ' s
soon learned that welded connection s brought to steel strength, but building components were and are
structures an elegance that could never be achieved with designed to such limits only when weight is especially
547 riveted or bolted connections. important, for example, in a long-span roof. Taylor ' s
work did , however, have an immediate impact on the
Despite the understanding of brittle fracture developed design of structural steelwork through the work of John
in the 1920s and 1930s, many structural failures were Baker and the Steel Structures Research Committee in
ultimately attributed to the formation of Griffith cracks the 1930s (see pp. 4 70, 4 72) .
because it can be very difficult to detect them or prevent
them from forming, due to metal fatigue. In recent years Minimum-Weight Structures
an understanding of Griffith crack theory has been par- As we saw in chapter 5 , the conscious search for a mini-
ticularly valuable to engineers seeking to work confi- mum-weight structure began in the 1790s with the need
548 dently with brittle materials such as glass and to to minimize the quantity of iron used to make cast iron
refurbish and reuse nineteenth-century buildings with beams and columns. In the 1830s, Eaton Hodgkinson
beams and columns of cast iron . The use of highly and William Fairbairn established the most economical
stressed tensile membranes such as those pioneered by form for a beam, and a decade later conducted experi-
Frei Otto in the mid-1950s (see p. 506 and chapter 9) ments to establish the best structure for tubular girders.
has also benefited from Griffith's insight. Understanding After the middle of the nineteenth century the quest for
the mechanism by which a hole or tear might propagate minimum-weight structures was taken up by the design-
in such highly stressed membranes has enabled design ers of long-span roof structures. To this end design engi-
engineers to keep stresses below the critical level and neers devised the " pin-jointed " truss, which enabled
avoid potentially catastrophic consequences. them to control the way the structure worked: The pin
joints prevent the transfer of bending moments from one
While designing aircraft and propellers during the World member to the next, and thus ensure that all the mem-
War I, Griffith had worked closely with the British physi- bers carry only tension or compression forces and no
cist and engineer G. I. Taylor. Taylor, like Griffith , was con- bending moments. Such structures are easily amenable
548

549

548 Exhibition Pavilion for the firm Glasbau Hahn, Frankfurt am Main, c. 1952. One of the first buildings to use glass as a
structural material for beams and columns. 549 Diagrams showing how edge (top) and screw (bottom) dislocations can
move through a crystal structure, allowing plastic deformation of an apparently rigid and brittle material.
550

551

550 The rigid airship Schutte-Lanz 1 under construction in Germany, 1911. The helical girders were of timber.
551 A selection of the many methods used to create the rigid transverse frames and bulkheads for rigid airships.
(A) Graf Zeppelin (German, 1928); (B) R101 (English, 1930) (C) Macon (USA, 1933) (D) Hindenburg (German, 1935).
 to precise determination of forces and deflections using in braced frameworks such as airships and airplanes.
either analytical or graphical techniques. This iterative method for finding a set of solutions to the
large number of equations involved was not mathemati-
In the twentieth century, engineering design benefited cally exact, but it was accurate enough for practical pur-
considerably from the development, during World War I, poses . First, a simple analysis was used to calculate a
of two entirely new types of structure that made severe set of approximate forces in all the struts and ties in the
demands on structural engineers: the rigid airship and frame. This set of data was then used as the input for a
the airplane. The rigid airship was developed first in new calculation, which reduced the inconsistencies
Germany in 1900. The British were quick to follow the within the first set of data and generated a second set
German lead, and the 1920s saw a race between engi- that was closer to the "true " solution . This, in turn , was
neers in the two countries to design larger and larger used as the input for yet another set of calculations , and
550 machines. Not only did these structures have to be suf- so on , each successive calculation cycle being closer to
ficiently strong and stiff to withstand loads that were the true solution . Finally, the changes from one calcula-
still poorly understood , they also had to be sufficiently tion to the next were judged to be insignificant, typically
light to fly. after six or seven cycles. Nevertheless, the process was
a lengthy one. Novelist Neville Shute, a designer of air-
The structure of airships generally consisted of a series ship frames in the 1920s, describes in his autobiography
of disks, made rigid like a bicycle wheel using wires that Slide Rule the elation he and his colleagues felt when
551 could be tightened to prestress the compression ele- reaching the end of such a process , often after two
ments . Parallel disks were linked by perpendicular longi- weeks or more .
tudinal members , and the whole frame was made
sufficiently rigid using additional wires to provide tensile While relaxation methods were more sophisticated than
cross bracing. Such a structure was far from statically de- the building industry needed at the time , their develop-
terminate, and calculating the forces that the struts and ment was significant for two reasons. First, they were
wire ties needed to carry presented structural engineers ideally suited for use in computer programs developed
with an enormous challenge. In order to reach the neces- during the 1950s and 1960s to analyze frameworks and
sary strength-to-weight efficiency, airships and airplanes tensile cable-net and membrane structures . Most impor-
had to be designed much more precisely than buildings. tantly, however, relaxation methods provided, in the
Aircraft designers could not make the crude approxima- 1920s, the ability to check the accuracy of more approxi-
tions made by building engineers to simplify their design mate methods. This had previously been impossible, and
calculations, nor could they give reassurance to the re- engineers using approximate methods of analysis were
sults of such approximations by using safety factors of usually in the dark as to just how "approximate " they
three or four. Aircraft designed like this would be too really were .
heavy to fly. There was also the matter of size and com-
plexity: Although it was not difficult to set up the many The ability to make more accurate calculations , and to
equations that defined the state of equilibrium in the compare these both with the results of experiments
frame of an airship, there were so many that it was impos- and with approximate calculations, was especially use-
sible to solve them all in the pre-computer age. ful for research in structural engineering. In particular,
it enabled engineers to undertake a more careful study
Research on how best to design such structures was car- of the safety of structures and thus to address safety
ried out in many establishments throughout the world, issues when designing structures. In the 1850s
but it was work done at the RAE in England that would Rankine had shown engineers how to use the factor of
later have the most direct influence on the design of safety to make allowance for the inevitable inaccura-
structures in the building industry. In the late 1910s and cies that result from using simple mathematical mod-
early 1920s the engineer Richard Southwell (1888- els of materials, loads, and structures to predict the
1970) developed what later became known as the "relax- actual behavior of built structures. In the 1910s the
ation method" for analyzing the forces and deformations Swiss engineer Robert Maillart (1872-1940) argued
that, since the self-weight of a structure is more reliably stresses and deflections even in a modest steel or con-
known than the live loads that acted on a structure , the crete-frame building would be far too complex, and far
former should be subject to a lower factor of safety (see too time-consuming, to be carried out by most structural
chapter 7 , pp. 431, 436-38). A number of engineers engineers. The point is that the highest precision is not
working at the RAE in the 1920s and 1930s, notably necessary: The art of structural engineering design lies
Alfred Pugsley (1903-98), began to take this approach in being only as precise as is necessary: it is the art of
further, making an assessment of the different contri- approximation and the intelligent use of Rankine's great
butions to the overall safety of a structure and the invention, the factor of safety. The design procedures de-
degree of uncertainty in each of the mathematical mod- veloped for frame buildings as tall as the Empire State
els used to build the full model of an entire structural are testimony to the fact that a little science can go a
system . They also looked carefully at the conse- long way. The next major impact on the design of frame
quences of the failure of individual members; some, for buildings by advances in engineering science did not
example, might cause a catastrophic failure, whereas occur until the 1970s and 1980s . Only with the aid of
others might have a less important effect on an air- computers could the complexities of the buckling and tor-
craft 's performance and so could be designed with a sional behavior of structural elements and of the dy-
lower margin of safety. By this means designers were namic loads from wind, earthquakes , and fire be
able to reduce significantly the weight of materials satisfactorily mode led .
used in airframe structures. After World War II , Pugsley
brought his expertise directly into the building industry Although design methods for both steel- and concrete-
when he became Professor of Civil Engineering at the frame structures were already widely understood and in
University of Bristol in England. The approach he and use by about 1910, they were mainly to be found in text-
his colleagues had developed in the 1930s became the books. New methods proliferated rapidly, and practicing
basis of modern reliability analysis using what are now engineers often had to choose between alternative meth-
called " partial safety factors, " not only in the structural ods and approaches without sufficient understanding to
design of aircraft and buildings but also in the design of establish which was the best to use. Most countries soon
bridges, ships, cars, and other structures. recognized that there was a need to standardize design
procedures for both steel-frame and reinforced concrete
The Design of Building Structures structures. Three types of information were collected into
In the nineteenth century iron- and steel-frame construc- national standards:
tion had been innovative , requiring an understanding of Properties of materials, including the quality of their
the most sophisticated engineering science of the day. manufacture
By the early twentieth century it had become main- The various loads that building stru ctures should be
stream, and by 1930, the job of the structural engineers designed to carry
designing the Empire State Building had become a Codes of design practice that provided suitab le meth-
straightforward task of determining the sizes of mem- ods for designing the various structura l elements of
bers (see pp. 451- 52) . buildings - columns, beams, floors and shea r walls -
and th e connections between th em.
The ability to transform the " innovative" into the
" straightforward" was testimony to the ingenious design The information in each of these areas could be com-
methods that structural engineers developed . The behav- bined to create a mathematical model of the behavior
ior of building materials-steel, concrete, and brick- of an entire structure and ensure that the proposed
work-and, especially, the soil beneath a building is far design fell within acceptable limits . Generally speaking,
from straightforward. The way they behave when acting the most important limits were in two areas: servicea-
together is even more complex. The actual loads that will bility, especially a maximum deflection for beams and
impinge on buildings are highly variable and seldom floors , and strength, for example, a maximum or " per-
known with certainty. Despite the power and speed of missible " level of stress derived by calculation . The
today's computers, calculating precisely the " actual " overall stability of structures-their ability to resist hor-
izontal (wind) loads-was often presumed to be satis- always been perceived as improvements, especially by
factory without doing any specific calculations, just as busy practicing engineers, who routinely observe that us-
it was in the 1880s before tall buildings were provided ing the previous codes very seldom led to failures or col-
with wind bracing . lapses. This sequence of events has been replicated for
other construction materials and in other countries.
It was widely agreed that the aim of standardizing design
practices was a sensible one, but it was soon found not A key function of the codes of practice was to assist reg-
to be so straightforward. The minimum requirement was , ulatory authorities in assessing the safety of proposed
of course, that acceptable safety standards be met. structures . However, in discharging this function, they
Different people had different opinions as to the best or had to err on the side of caution ; the result was that they
most suitable ways of meeting that requirement. The con- often found themselves at odds with engineers in the
sensus as to what was considered acceptable by the sci- field, as the following two examples illustrate.
entific and design communities- the "engineering
climatology" as Alfred Pugsley called it-varied with both In 1904, the Ritz Hotel in London 's Piccadilly became 552

time and place. Thus , what might have been acceptable the first building in London for which a fully steel-framed
as a design method in 1900 might not be acceptable in structure was proposed. However, the Swedish engineer
1930 or 1960, both because of the experience and un- Sven Bylander (1876-1943), who worked for an
derstanding gained in the intervening period and be- American contractor, was prevented from using a steel
cause of society's changing expectations. frame of the type already common in the United States
because it was not permitted by the London building reg-
In Britain , for example , the London County Council pub- ulations . These required a building to have external ma-
lished its Reinforced Concrete Regulations in 1915 , fol - sonry walls that carried the loads from each floor to the
lowed in 1933 by a "Code of Practice for the Use of foundations; internal columns were permitted only to
Reinforced Concrete in Buildings " drafted by a subcom- carry floor beams. Bylander's compromise , made simply
mittee of the Building Research Board of the Department to comply with the regulations , was to construct the
of Scientific and Industrial Research , the Reinforced steel frame he wanted , stopping just a meter or so short
Concrete Structures Committee. A different code was is- of the external wall. Short beams spanned the remain-
sued in 1939 by the Building Industries National Council , ing distance to the substantial external masonry wall ,
but once World War II began , this was rarely used. Both of whi ch had very little load from the floors to carry. By the
these codes , as their predecessors at the turn of the cen- time Bylander started to design his next building, the
tury had been, were " permissible stress " codes. The first Selfridges Department Store on Oxford Street, London ,
British Standard code of practice for the design of rein- the regulations had been changed to allow columns in
forced concrete , CP 114, which dealt essentially with in the external wall , as is immediately apparent in th e
situ concrete work, appeared in 1948. It largely followed building elevation. 553

the prewar codes , but incorporated guidance on the ratio

of water to cement and on the minimum cover to be pro- In 1935-36, London 's Piccadilly was the location of an-
vided to reinforcement where conditions were corrosive. other conflict between building designers and the regula-
It encouraged , but did not require , inspections every tory authorities. A new shop was being designed for
three to five years to identify cracking and corrosion of re- Simpsons clothing store by architect Joseph Emberton 554

inforcement, which was , by then, recognized as a prob- (1889- 1956) working with engineer Felix Samuely
lem. Codes of practice for prestressed concrete (CP 115) (1902-59). They wanted a dramatic, entirely column-free
and for precast concrete (CP 116) followed in 1959 and elevation at street level. Samuely proposed a steel
1965 respectively. The code-updating process has con- Vierendeel frame at the first-floor level spanning the full
tinued to the present day: CP110 was issued in 1972 , width of 24.4 meters , which would carry the loads from
BS8110 in 1985, and the European design code EC2 in all the upper-story columns to the columns at either side
the mid-1990s. Although each was intended to be an im- of the shop front. However, this did not comply with reg- 555

provement over its predecessor, the new codes have not ulations that required floor loads to be conveyed by the
553 554

552 Ritz Hotel, Piccadilly, London, 1904-06. Engineer: Sven Bylander. Under construction . 553 Selfridges Department
Store, Oxford Street, London, 1906-09. Engineer: Sven Bylander. 554 Simpsons Store, Piccadilly, London, 1936.
Engineer: Felix Samuely; architect: Joseph Emberton. 555 Simpsons Store. Steel frame structure as designed by Felix
Samuely (left); as modified to meet London Building Regulations (right).
 II II 1.1_ 1~
i j
I I I I IA111111111111111111~I
I I l_ l_ l_ j I--mTIIIIIIl)]lJJIIIIIIIIIIIIIIIIIIIITIIIIDum::n,,.I
I I
I_ I_ I_
1= 1= = ~ I I.-m:x:rrIII]T]TllllllllllllllllilO!IIIIUIPIDIInn--I
I I
-f
I I
l~llllIIIlililIII~I

-~--*=:-]
PR0P0.5 E..0 AF\RANC f_Mf.NT M001F'" ICA Tl0N5 BY L.C.C.
T01L ~MAOLD AllllLA~ llllL ... AL:,LwT T"ML &LNOIN C. MO...,LNT~ IN T,-4L Mlt.M6Lll\3 AN O
TH L ,._t. QVl llllL.0 _,TLLL ~LG'TIONS AII\I. f'I\.OPOllllT!ON.-.L TO THI!.~~ Atl'\LA~

COMPARISON OF LOAD ING SYSTEMS

COMPARISON OF BENDING SYSTEMS
555

most direct route to the nearest column resting on a (and are) generally represented on these committees:
foundation: they could not be carried downward a few materials producers, practicing designers, and the engi-
floors, and then sideways through a transfer girder to a neering research community. Typical ly, when such com-
column on a foundation. Samuely and the architects mittees met to discuss changes to the codes , the
argued their case against the London County Council engineering scientists felt it was most important to in-
until the last possible moment, but they lost. Full-width clude the latest engineering research. Practicing engi-
transfer girders had to be introduced at every story, at neers, on the other hand , were usually in favor of simpler,
556 considerable unnecessary cost and weight. quicker-and, above all, safer-design methods , even if
they were, strictly speaking, less accurate. Even as late
The Vierendeel girder was fabricated by welding to as the 1920s the design of structural frames was under-
achieve the necessary strength at the connection be- taken as a series of independent columns and beams . It
tween the vertical and horizontal portions of the girder. was widely perceived that this was both unrealistic and
This was the first substantial use of welding on a building unduly pessimistic , and that it led to more steel being
project in England-a technology that Samuely had used than was actually necessary. In practice , however,
brought from his previous work in Germany and Russia , there was no alternative . A frame structure that relies on
where welding had been developed in the sh ipbuilding in- the rigidity of the connections between beams and
dustry during the early 1920s. To satisfy the authorities, columns cannot be determined by statics alone because
the welded frame had to be cut in situ to form two large it is statically indeterminate . In principle , this problem
independent beams at first- and second-floor levels. Even had been solved in the 1850s by the French engineer
after this surgery their size was impressive, more in Emile Clapeyron (1799-1864), who devised what was
keeping with the scale of bridge engineering than the called the "theorem of three moments ," which enabled
scale of building structures. The coda to this story is that engineers to ca lculate the unknown bending moments.
Samuely's battle, like Bylander's before him, did force a However, this required solving as many simultaneous
rethinking of the regulations, and they were subsequently equations as there were unknowns-an impractical
changed to allow any arrangement of columns and method for a whole building.
beams as long as their safety could be documented.
In the 1910s the German engineer Axel Bendixen refor-
As design codes for the structures of buildings were artic- mulated Clapeyron 's method in terms of the rotation of
ulated, there followed the inevitable establishment of the ends of each structural member of a frame.
committees that reviewed codes ' content to ensure that Bendixen's "slope-deflection method " was widely used
they incorporated the latest needs and developments in for simple structures, but it still required solving the
the industry. Three groups of interested parties were same number of equations as there were unknown
556

UNSYMMETRICAL BENT As a Continuous Girder Corrected for Sidelurch

~201 Assuming fhe girder rigid the column shears
B I,m;=+52.9 m;=-106 will be proportional fo fhe values of'
l(=(j) I( . I 2· I 2 17
I!; fhal 1s, fo and
4 9+ = 4 9 36
9

1
... ~= 17 x!0.0=5.29 m;=5.29x/0=+52.9
H2 = 4.71 m; =47/X/5=-70.6
J ¾ ·Vi
+52.9 +440 :::::=>-< -50.5 -70.6
+2S.2 -22.0
~=5.29 -·n o :::::=>-< +1s. 1
-7.8 +/0.5
+6.S :::::=>-< - Z5
+3.7 -3.2
-3.l -------- +2.3
-I.! ---------... +1.5
+0.9 :=:>< - J.J
0.5 -0.4
0 ,'. -0.4 +0.3
m af Top
-5
+2.7
.sl
+47.4 -54.4
V +162 Llm, af Top
AJ -8.J Am, af Base
-50.2 +62.5' 0. m at Base
- - - - - -- -
~ = 9;; =4.88 H2 =
1
lfo9 =: J.90
Since lhese momenls correspond fo a folal sheaf'
of 8.18, !he moments for a shear of /0.0 will
-be_j_Q_ of fhese values :-
8.78 k/
~ = -57.0
M8 =+54.0
Mc=-61.8
M0 =+ 71.3

557

556 Simpsons Store, Picadilly, London , 1936. Engineer: Felix Samuely; architect: Joseph Emberton. Building under con-
struction, with huge welded steel girders, as modified to London building regulations. 557 Description of the moment
distribution method, developed by Hardy Cross, c. 1930.
 bending moments, and calculating the sizes of individual column and beam. This meant that an alternative mathe-
structural members of a multi story building would involve matical model could be imagined which considered the
many hundreds of such equations. In the days of manual frame as made up not from a series of beams and
calculation, this was either unrealistic or, for engineers columns , but from a number of cross-shaped elements
with limited mathematical skills , simply impossible. connected at pin joints. This model of the frame was very
During the early 1930s two engineers found different much easier to calculate , and has become the standard
ways out of this dilemma: Hardy Cross (1885-1959) in way of thinking about how such frame structures can be
the United States, and John Fleetwood Baker, later Lord analyzed. Most importantly, the act of drawing the bend-
Baker (1901-85) , in England. ing moment diagram forced the engineer to think about
how the structure was actually working. From the bending
As Southwell had done , Hardy Cross devised an iterative moment diagram, it was but a small step to sketching the
means of solving an almost unlimited number of equa- shear force diagram and the deflected shape of the
tions. It was a brilliant and elegant simplification of the beams and columns.
calculation procedures that had to be undertaken to ana-
lyze the bending moments in rigid-jointed frame buildings. The moment distribution method was quickly adopted by
In a world that seemed ever more complex to most engi- design engineers throughout the world because of the
neers, it came as a ray of sunlight. Hardy Cross ' s " mo- ease with which it could provide answers that were accu-
ment distribution method " brought the design of rigid rate enough , even for large and complex structures. It
frames within reach of the engineer with ordinary calcula- also became popular with university lecturers, for at the
ss1 tion skills, armed only with a slide rule. It was conceived same time that it generated numerical results, it helped
in terms of the bending moments carried in the mem- instil! into students an understanding of how the struc-
bers-an approach more in keeping with the way engi- ture was working. In this respect the moment distribution
neers think about how frames actually work than method stands alongside the methods of graphical stat-
Bendixen 's method , which considered the angles of rota- ics that revolutionized structural design in the latter half
tion at the ends of individual members. The moment dis- of the nineteenth century and remained current for
tribution method required , first , imagining all the nearly a century. Both are still regularly taught today by
connections fixed and carrying out the relatively simple those who seek to develop in students a deep feeling for
calculations to find the moments at the ends of each structural behavior.
member. Then each connection would, in the mind of the
engineer, be " released ," and any out-of-balance mo- The Plastic Behavior of Steel
ments distributed among the other members meeting at At the same time that Hardy Cross ' s method was making
the connection , in proportion to the stiffness of the mem- the structural engineer' s work in analyzing and designing
bers . Repeating this process just four or five times gener- steel structures easier, some fundamental doubts were
ated successive approximations that rapidly converged beginning to appear as to the validity of treating steel as
on the equilibrium solution. Cross 's method was also just an ideal elastic material. In addition, it began to be rec-
as easy to use for the calculation of forces and bending ognized that designing a steel-frame building as a series
moments resulting from wind loads as for those resulting of independent columns and beams was both unrealistic
from gravity loads-a significant benefit for the designers and pessimistic. In the 1920s it was common practice to
of high-rise buildings. design steel members so that the stress never exceeded
the so-called "elastic limit, " with a certain margin of
An important consequence of using Cross's moment dis- safety. However, there were a number of well-known facts
tribution method was that it required engineers to draw about the use of steel in real structures that contra-
bending moment diagrams for the frames they were de- dicted this ideal picture of steel. It was known that rolled-
signing. Not only did this help them visualize how the steel components used in building structures already
structure was carrying the loads , it also drew attention to contained significant internal stresses resulting from
the many places where there was no bending moment in the manufacturing and assembly processes, and that
the structural members-at or near the center of every these "residual stresses " could be almost as high as
the elastic limit for steel. It was also well known that elastic behavior of the steel, they would be based upon
parts of a steel structure, for instance the area around more rational assumptions, with the goal of achieving an
the rivet holes in connections, could and did deform per- improved correspondence between predicted and actual
manently, or plastically, without serious cause for con- forces and stresses in structures.
cern. Many designers of steel structures were also
aware that some of the basic assumptions they made in A series of service trials of the new design procedures
their design calculations were unnecessarily conserva- was carried out in the mid-1930s, and it was generally
tive or simply unrealistic. For example, steel has a con- agreed that they did , indeed , achieve their intended aim .
siderable reserve of strength beyond the so-called point However, despite requiring the design engineer to use
of "failure " that is usually defined as a certain, safe (but more complex engineering science and mathematics, the
rather arbitrary) proportion of its ultimate strength or its new procedures did not produce steel structures that
yield strength (the limit of its elasticity). There was also were either significantly safer or cheaper, nor did they ad-
the matter of differential settlements between the vari- dress the apparently safe plastic behavior of steel above
ous foundations supporting the columns of a building. its yield point or the remaining doubts about elastic de-
Such movements are inevitable in all buildings, and yet sign methods in general . Simplified versions of the new
they were ignored when designing steel frames . Had they procedures were developed , but these, too , were finally
been taken into account, the calculations would have rejected by the profession.
shown that many of the columns and beams would "fail, "
in the sense that the elastic stresses would exceed the Baker and his research engineers became increasingly
maximum values laid down in the design codes. concerned at this outcome. Despite their best efforts ,
they had failed to resolve the serious doubts about the
These observations about the behavior of actual steel validity of elastic design methods. Indeed , it seemed that
structures led to two conclusions: Steel structures were they were incapable of further development or improve-
usually overdesigned , and significant savings cou ld be ment to regain the confidence of engineers that is funda-
achieved through a more efficient use of steel; and, more mental to all structural design methods. A state of crisis
fundamentally, the whole basis of elastic design methods had been reached.
for steel was irrational and needed to be reassessed. In
1929 the British Steelwork Association set up the Steel A breakthrough became possible only with the realization
Structures Research Committee , under the chairman- that the problems with the existing elastic design proce-
ship of John Baker (1901-85). to review the situation and dures could be addressed only by means of a radically dif-
make recommendations for new design procedures. ferent approach . The key to the solution came in the form
of a new idea of what constituted failure. The idea that a
The work of the Steel Structures Research Committee failure was the point at which a certain limiting stress
was initially directed at investigating just how far the was reached (the limit of elasticity) was replaced by the
usual assumptions upon which steel frame design was notion that the limit should be the collapse of the struc-
based deviated from the actual conditions found in sim- ture. A certain amount of plastic deformation could there-
ple frames. A thorough program of testing model and full- fore be tolerated as long as sufficient parts of the
scale structures using various techniques of stress structure remained below their yield point, thus prevent-
analysis, including the newly developed vibrating-wire ing the structure as a whole from becoming a "mecha-
and electrical-resistance strain gauges (see pp . nism," defined as a structure lacking sufficient bracing or
4 75-76). demonstrated that the deviations were often rigid joints to prevent its collapse. This realization forced
considerable. A second stage of the committee 's work a total reexamination of structural design procedures.
was directed toward developing new design procedures. No longer was the level of stress in every part of the
Although the new procedures would still consider only the structure of prime interest; the main concern became the

558 Hunstanton School, Norfolk, England, 1949-54. Engineer: Ron Jenkins of Ove Arup & Partners; architects: Alison and
Peter Smithson. 559 Morrison air raid shelter developed by John Baker, 1937-38.
 558

1/e" ci,o nut and

t>olt ¾ "1ong

- 1/e~.
IHdlO washer t4g thic k

~ :fy ~ t di": I~~~ and

6 laths , .. ,. 5 ' 9 \t° x 18g

6 ' x f;,' x3/e"'
12 lat hs I"' x 3 ' 4 1,;' -,. 18g
Corner angles

2 •, ,; x 2 112" x 11 "'
4 ang les

3 1
Bolt s / 8" <ho • 1 // long
bg hook a nd eye:
fas t ening t o ell corner s

559 M ORR ISON SHELTER • FINAL DESIGN

 location and relative magnitude of local maximum bend- ing structures , for example , single-story steel-frame
ing moments as the applied loads were all increased , in warehouses and industrial units , plastic design proce-
proportion , up to the point at which sufficient plastic dures are now widely used. They are much simpler to use
hinges had formed to transform the structure into a than elastic procedures, and can lead to substantial
mechanism that might collapse . The safety of the struc- economies of material-one of Baker' s original goals.
ture under working loads could be achieved by ensuring However, there has been less impact on the design of
that the applied loads were less than the collapse load by larger steel-frame buildings , where the sizes of steel
a margin which came to be known as the " load factor. " beams and columns are governed not by the need to pre-
vent collapse but by the need to limit elastic deflections
The new approach required new theoretical concepts , that lead , if too large , to the cracking of the windows and
such as " plastic hinge ," "full plastic moment," " shape plaster finishes inside buildings. Nevertheless , the most
factor," and " incremental collapse ," as well as new defi- recent design codes for both steel and reinforced con-
nitions of old concepts , such as "failure " and "collapse ." crete buildings and bridges allow the use of plastic de-
The new plastic design procedures , as they came to be sign procedures.
known , were not readily accepted by many members of
the established community of designers. Indeed , even Experimental Engineering Science
half a century after this design revolution , they are still Not only did the growth of experimental physics in the late
not universally accepted. nineteenth century lead to results that had a direct im-
pact on building engineering , for example , in the work of
The first structure that Baker designed using his plastic Griffith and Taylor, it also led to the widespread growth of
design methods , immediately before the outbreak of experimentation as a means of solving practical engi-
World War II , was a lightweight bomb shelter for use by neering problems . The idea of experimentation was cer-
559 the civilian population . Named the " Morrison Shelter " af- tainly not new, but it wasn 't until the late nineteenth and
ter the then Home Secretary Herbert Morrison , it made early twentieth centuries that the benefits of setting up
full use of the ability of steel to absorb huge amounts of teams of research scientists to deal with a subject sys-
energy when deforming plastically beyond its elastic tematically were recognized. Many universities that had
limit. Its light steel frame was designed to absorb the im- previously concentrated on theoretical studies aug-
pact energy of falling masonry by deforming plastically mented by a few experiments began to refocus their ef-
without total collapse . Tens of thousands of the shelters forts on experimental work. Some government agencies ,
were made , and many thousands of deaths and severe like the Federal Materials Testing Station in Switzerland ,
injuries were prevented . The same principle is at work to- founded in 1880, and the National Physical Laboratory in
day in the thousands of miles of crash barriers alongside Britain, which opened in 1902 , were set up specifically to
our highways . It is also the principle underlying the con- undertake research into materials science and provide
struction of many buildings designed to withstand earth- authoritative data about materials.
quakes without collapse. In such buildings , the plastic
hinges in steel members are the means by which the en- By the early 1920s the benefits to be gained from an ex-
ergy of the earthquake is absorbed . perimental approach to research were becoming clear in
many fields of building engineering . A growing number of
As with many developments in building engineering, the research facilities and institutes were established that
consequences of the plastic design revolution are virtu- carried out studies in the increasing range of fields in sci-
ally invisible to those outside the engineer 's world. A ence and engineering that had an impact on building de-
number of test buildings were constructed , including one sign and construction . Many of these were established in
at the Cambridge University Engineering Department, universities; others were independent, such as the
where John Baker was a professor in the 1950s, but the Hermann-Rietschel-lnstitut for Heating and Ventilation
first architectural use of plastic design methods was for Technology established in Berlin in 1887 (see chapter 7 ,
a modest steel structure at Hunstanton School in p. 408). In the 1890s, testing stations were set up in a
558 Norfolk , England (1949-54). For certain classes of build- number of countries in continental Europe and the United
 States to study the performance of construction materi- surface and recording the test on film. For the case of a
als in fires. In 1920 a national Building Research Station large aircraft hangar which had one side open to the wind ,
was established in Britain to study all subjects related to these tests revealed the creation of a large vortex that
construction . The Earthquake Research Institute was es- created unexpected positive pressures on the lee side of
tablished at the University of Tokyo in 1924. Individuals the building.
not affiliated with a university or research organization
were also able to use experiments and careful measure- One of the earliest uses of a wind tunnel to measure
ments to make their impact on building engineering. wind pressures on tall buildings was for the Empire 562

State Building in the early 1930s. The air pressure was

A particular benefit of the various research institutes was measured at various places around the circumference
that they provided links to research in fields of inquiry of the model building and at several different heights.
outside but related to the building industry such as geol- This enabled the designers to estimate the wind loads
ogy, materials science, and mechanical and electrical en- on the entire building and on individual windows . On the
gineering. At the beginning of the twentieth century, one lee side of a tall building, for example , the air pressure
of the most fruitful areas of research, one that would is significantly lower than it is inside the building, thus
have important impacts on the design of buildings , was tending to suck the windows out. At this time only static
the relatively new field of aerodynamics. loads were investigated. The role of vortices in causing
periodic loads on structures was first recognized during
Wind is of particular concern , of course , in the design of the investigation of the spectacular collapse of the mile-
tall buildings. The first wind tunnel had been built in 1871 long Tacoma Narrows Bridge in Washington State in
by Frank H. Wenham (1824- 1908), a member of the 1940. Wind tunnel tests revealed how vortices first built
Aeronautical Society of Great Britain , to measure lift up on one side of the bridge deck and were then shed
forces on aerofoils . Gustave Eiffel built a wind tunnel in into the lee of the structure. The next vortex then built
1912, and in 1916 Ludwig Prandtl (1875- 1953) de- up on the other side of the deck , and so on. The vor-
signed and built the first closed-circuit wind tunnel , at the tices thus exerted alternating loads on the bridge deck
University of Gottingen in Germany, which was the proto- that set it in oscillating motion. The role of vortices in
type for all modern tunnels . Prandtl discovered that the causing periodic loads on tall buildings became sign ifi-
air in contact with a surface is stationary, and its speed cant when engineers began designing very slender tow-
increases to the ambient wind speed across a thin layer ers in the late 1950s.
of air he called th e " boundary layer." He found th at the
thickness of the boundary layer was highly sensitive to Experimental Stress Analysis
surface roughness . This was especially significant for Underlying the design procedures for all load-bearing ele-
building engineers , as buildings are all located in the ments of buildings is the idea that the stress in the mate-
boundary layer and themselves contribute to the rough- rial should be less than a certain prescribed level . The
ness of the earth 's su rfa ce. design engineer thus needs to know the magnitude of
stresses inside the materials of which the structural ele-
In 1927 Otto Flachsbart (1898- 1957), one of Prandtl 's ments are made. This approach had been used through-
co-workers , undertook the first wind tunnel experiments out the nineteenth century by the designers of many
on buildings , to study the effect of the thickness of the types of structure that carry loads-bridges , locomo-
560 boundary layer on wind pressures on buildings. He meas- tives , ships, high-pressure boilers, machine tools , and
ured air pressures on the walls and roofs of various types many more . Unfortunately, it is not possible to measure
of build ings , including small apartment blocks , and com- internal stresses directly. It is , however, possible to
pared the results for buildings standing in isolation with measure the loads applied to a structure and the partic-
those for groups of buildings in close proximity. He also ular load that causes it to collapse . It is also possible to
used experimental equipment in which water was used to measure the deflection of various parts of a structure, or
561 simulate the flow of air. The movement of the water was the change in length of one of its members , and the de-
made visible by sprinkling aluminum powder on the water gree to which the surface of a structural component has
 V

Schema fiirModel/
2b

560

560 Diagram showing the results of Otto Flachsbart's wind tunnel tests on a model of an aircraft hangar, Gottingen,
Germany, 1927. 561 Otto Flachsbart's wind tunnel tests. Selected frames from film recording the movement of water
(representing air) flowing past a model of the aircraft hangar. 562 Wind tunnel model of the Empire State Building, New
York, used to measure w ind pressures on the facade, c. 1931 .
561 562

been strained in tension or compression . It is then neces- The Strain Gauge

sary to construct mathematical models of the structure For simple tension and compression elements , the inter-
and the materials, based on the appropriate engineering nal stress is calculated as the force in the element di-
science , in order to establish the relationship between vided by the area of cross section . The stresses inside
the measurable external variables and the internal elements subject to bending are calculated from meas-
stresses. Following the lead taken in experimental urements of the load, the deflection and the geometry of
physics during the late nineteenth century, engineering the cross section. This allows the second moment of
scientists gradual ly began to appreciate the benefits that area, I, to be calculated , and the interna l stress calcu-
could accrue from a rigorous and methodical approach to lated , using the elastic theories of bending that were de-
their subject. By the early twentieth century the study of veloped during the eighteenth and nineteenth centuries.
the microbehavior of materials under load was flourish-
ing, and it soon became known as "experimental stress The main limitation of these approaches was that they fo-
analysis. " During the next half century experimental cused on the overall behavior of a structure rather than
stress analysis followed two distinct lines of develop- the behavior of the material itself. In the early nineteenth
ment. One was based on the strain gauge, the other on century, scientists began measuring the strain (deforma-
photoelastic stress analysis . tion) on the surface of a piece of material under load.
 They did this by measuring the change in the distance be- changes due to relative movement of the two ends. 564

tween two gauge marks on the surface when a load was Vibrating-wire strain gauges , which are bulkier than elec-
applied to the structure. Once the stiffness of the mate- trical resistance strain gauges , are used mainly for study-
rial (Young's modulus) was known, the surface stress ing larger building elements and the geotechnical
cou ld be calculated. However, the process was extremely investigation of rock movements in building foundations ,
tedious. It required the use of a traveling microscope to which need to be monitored over long periods of time.
obtain results that were sufficiently accurate , and so was Two problems with all strain gauges are their instability
of no use for components with inaccessible surfaces or over time and their sensitivity to temperature changes .
those subject to dynamic loads. For the most careful measurements, therefore, an electri-
cal thermometer is essential, and the gauges need to be
All this changed in the 1930s with the invention of the constantly recalibrated to compensate for creep in the
strain gauge. The first and still most widely used strain mounting of the gauge on the component being tested.
gauge is the electrical resistance strain gauge , the de-
sign of which is based upon the fact that the cross sec- Since their invention , strain gauges have been the essen-
tion of an electrical conductor is reduced when stretched tial tool of all research into structural behavior, and their
and this, in turn, increases its electrical resistance. impact on structural engineering has been enormous.
Electrical resistance strain gauges were first used by the They have enabled countless design procedures and
American engineer Charles Kearns in the early 1930s for mathematical models to be tested for their accuracy and
studying the stresses in aircraft propellers . He ground reliability in predicting structural behavior and , hence , im-
flat a conventional carbon-composite electrical resistor proved. Today, the readings from strain gauges are fed di-
and mounted it onto an insulating strip which he ce- rectly into computers , and stress levels and patterns can
mented to the blade. He made the electrical connection be calculated and displayed in real time if desired .
to the stationary measuring equipment by means of
brushes and rings similar to those used in electric mo- Photoelasticity, Photoelastic Stress Analysis
tors. In the late 1930s Arthur Ruge and Edward Simmons Photoelastic stress analysis , like strain gauge technol -
in the United States both used arrays of fine wires to ogy, was first developed in the aerospace industry.
achieve the same effect. In 1952 the English engineer Because of its relatively high cost , it has been used in
Peter Jackson , working with Saunders-Roe , an aircraft the construction industry mainly for large structures
manufacturer on the Isle of Wight in Britain , used a tech- such as bridges and dams and only occasionally in build-
nology that had been developed for producing printed cir- ing engineering. However, it has had an enormous im-
cuit boards to make foil strain gauges that were much pact on the understanding of stress and strain inside
smaller, easier to mount, and more reliable than wire materials under load , and thus warrants mention here .
strain gauges. Foil gauges still form the basis of experi- An optical technique that uses models made of a trans-
mental stress analysis, both singly and in the form of a parent material, photoelastic stress analysis allows engi-
rosette with three gauges at 120 degrees to enable the neers to see directly what is going on inside a component
563 principal stress directions to be established. Although under load . Although the stresses themselves are not
strain gauges were developed in the aircraft industry, made visible , the technique allows the visualization and
their use for measuring stresses in both full-size and measurement of the next best thing-the magnitude of
model structures and components quickly spread to all the difference between two orthogonal stresses at any
other industries. point. A relatively simple process allows calculation of
the stresses themselves.
Another strain gauge technology, developed by the
Maihak company in Germany in 1936, was based on the Photoelastic stress analysis depends upon the property
well-known relation between the pitch at which a string vi- of certain transparent materials to affect polarized light
brates and the tension in the string. A wire, perhaps 200 as it passes through the material, and in proportion to
to 400 mm long, is fixed at both ends , and its pitch meas- the stresses inside it. This phenomenon, called "birefrin-
ured electronically as its length, and hence its tension, gence, " or "double refraction ," was first observed in
 .-~-

Vi
,--
\ ;>( /
I
" '
Felt_>,'--~

563

564 565

567

563 Diagrams of electrical resistance strain gauge rosettes that measure the strain on the surface of a test piece in three
directions. 564 Acoustic or vibrating wire strain gauge. Typically 100-150 mm long. 565 Photograph of photoelastic
model. The model shows the distribution of stresses caused by a load on a surface. 566 Generation of the lines of princi -
pal stress in a model subject to bending. First the lines of equal inclination of the stress directions are plotted (top), then
the lines of principal stress. The lines showing tension (bottom of the beam) show the ideal position for steel reinforce-
ment in a reinforced concrete beam. 567 Photoelastic model showing a model under bending load, and stress concentra -
tion at the root of a crack.
 glass in 1815 by the Scottish physicist David Brewster placed by computer models . Photoelastic stress analysis
(1781- 1868), and given comprehensive theoretical can be used to great benefit, for example, in developing
treatment by the German scientist Franz Neumann an understanding of reinforced concrete by indicating
(1798-1895). Using this analysis, internal stresses can where steel reinforcing bars in a beam should be best lo-
be computed from the observed degree to which the cated, since they should, ideally, follow the lines of the
plane of polarization of the light is rotated . In fact, the de- tension stresses inside the beam.
gree of rotation depends not upon the magnitude of the
stress, but upon the difference between the magnitudes Experimental Analysis Using Scale Models
of the " principal stresses ," and this enables the precise Photoelastic stress analysis is a sophisticated example
direction of the internal stresses to be computed. 3 of using a test model to simulate , in an experimental lab-
oratory, the behavior of an actual , full-size structure. The
The early development of the technique of photoelastic making and testing of model structures is as old as build-
stress analysis was undertaken in the 1910s at ing itself. The earliest models would have been used to
University College , London , using test models made of establish the stability of a structure such as an arch-its
glass . Although glass is widely available, it is not highly very ability to work as a structure-or the load that can be
birefringent; it is also difficult to cut to shape , and is vul- carried by an individual beam or an assembly of struc-
nerable to fracture. Photoelastic stress analysis became tural elements making up a bridge or a roof truss, for ex-
more accessible with the development of transparent ample. The technique only really becomes useful in
epoxy resins, such as Araldite , in the 1940s. The tech- structural engineering, however, when the results of test-
nique reached its peak in the 1960s in the aerospace in- ing a scale model of a structure can be used to predict ac-
dustry, for example , at Rolls-Royce in England , where two- curately the behavior of a geometrically similar structure
and three-dimensional models of aero-engine compo- built at a larger scale.
nents were analyzed in order to find ways of reducing
stress concentrations, which are usually the points at As we know from observing Roman arches and medieval
which potentially catastrophic fatigue cracks originate . cathedrals, the stability of masonry structures is inde-
pendent of their scale. If a model arch, flying buttress ,
Even dynamic stresses can be established by loading a or dome is made to stand, then a similar structure , with
model in an oven and cooling it while still under load, lit- all dimensions increased proportionately, will also
erally freezing the stresses into the Araldite. Thin slices stand. The stability of a masonry structure depends only
are then cut from the model for analysis. The model is on its geometry (the plan , section and sizes of its corn-
then viewed in a polariscope to reveal the varying degree ponents) ; it is virtually independent of the strength ,
of rotation of the incident polarized light as a pattern of stiffness , or density of the masonry. The designers of
565 highly colorful interference fringes. For quantitative work, masonry buildings could, therefore, scale up the results
monochromatic light (usually sodium) is used to produce obtained from a model building in direct (linear) propor-
sharper fringes and measure the magnitude of the inter- tion , and design a satisfactory full -size building. This
567 nal stresses to an accuracy of a few percentage points. process cannot be used for structural components that
The same fringes can be used in qualitative analysis to are subject to bending or that may fail due to buckling,
establish the directions of the internal principal stresses since these phenomena do not vary linearly with scale.
(tension and compression) and to generate a diagram The fact that no engineering science is needed to use
that shows the flow of forces through an object or, as it is scale models for the design of masonry structures
566 usually called , the "load path. " helps explain why masonry structures developed to
such a high degree of sophistication so early in building
Photoelastic models provided a direct visual representa- history. Scale models could not be used to help design
tion of the internal effects of external loads on a struc- bending structures, for example, until the engineering
ture. This unique ability ensured their continued survival sc ience necessary to scale up the results of model
as a teaching tool long after the quantitative use of pho- tests for such structures was developed in the mid-eigh-
toelastic stress analysis in research laboratories was re- teenth century.
The shape of simple tensile structures is also independ- points of support and stiffening of the unsupported
ent of size ; fo r example, the form of a hanging chain of edges. The task of trying to create a scale model using
weights depends only on the relative sizes of the weights , real reinforced concrete, with its sand , aggregate , and
not their absolute weights. This fact was known to Robert hundreds of steel bars of different sizes , shapes , and de-
Hooke in the 1670s. Hooke also knew that the results ob- grees of restraint within the concrete, was also tricky. With
tained using a hanging (tensile) model could be used (in- so many variables to consider, it must have seemed most
verted) to establish the form of compression structures unlikely, in the early days of designing shells , that design-
such as arches. As we saw in chapter 4 , this technique ers would be able to predict with confidence the behavior
was used by engineers in 17 42 to analyze the stability of of a concrete shell spanning 50 meters or more from tests
the dome of Saint Peter's Basilica in Rome. In the 1840s, on a model just a meter or two across . Yet , by the 1950s,
models of suspension bridges were built and tested to this seemingly impossible goal had , indeed , been
establish how they would deflect as a load passed over achieved, though it had not been reached in a single step.
them. In the 1890s Antonio GaudT scaled up the geomet-
rical forms established using three-dimensional hanging The significant breakthrough came from the application
models to design the forms of many masonry vaults in of an idea developed for an entirely different engineering
his buildings (see chapter 7, pp. 441 , 443 ). If the self- problem-the search for the most efficient shape for a
weight of the structure is ignored, the axial forces in the ship 's hull. This became an issue for ship designers in
members of a triangulated , statically determinate , pin- the 1860s , when steam engines were first used to power
jointed truss are also independent of scale. From the ships. In order to determine the size of engine that a new
1850s on, models of such truss bridges were sometimes ship would require , before the ship was built, it was nec-
tested to compare the forces in each member with the essary to know the resistance the ship would encounter
values calculated using mathematical models. when moving through water. Initially, it was assumed that
the resistance comes only from friction between the hull
The situation is very different, however, for a structural of the ship and the water. Calculations based on this as-
element subject to bending or loaded in compression to sumption, however, were found to be seriously wrong. At
the point at which it may fail by buckling. The strength and first it had been thought that testing a model of a ship 's
the stiffness of a beam depend on the properties of the hull might enable the resistance of the full-size ship to be
material ; the self-weight increases with the cube of the in- determined. The results of such tests by the British
crease in scale ; and , if self-weight is ignored , the stiff- Admiralty, however, proved to be "embarrassing and ex-
ness of the structure increases with the cube of the scale pensive mistakes ," as one official report concluded .4
and its strength with the square of the scale . Rea ching
this understanding of the effect of scale had been the In the 1850s, the designers of the largest steamship
goal of scientists since they first began to look at struc- built to date , the Great Eastern (launched in 1859), were
tural behavior, from Galileo and Hooke. However, it was faced with similar problems of scale . This ship was to be
only with the work of Eaton Hodgkinson and William the largest the world had ever seen: At over 200 meters
Fairbairn , who found the most effective shape for a cast long and 26 meters wide, she would be nearly twice as
iron beam around 1830, and the most efficient cross long and have five times the displacement of the next
section for the tubular wrought iron girders of the railway largest ship . In around 1853 , William Froude (1810- 79 ),
bridges over the River Conwy and the Menai Strait in an engineer working for Brunel's Great Western Railway
North Wales in the mid-1840s, that the scale effects in company constructing the line from London to the west
bending structures were fully understood. The case of a of England, joined Brunel and John Scott Russell on the
beam loaded in bending, however, is much simpler than design team for the Great Eastern . Froude , who had stud-
the case of a thin reinforced concrete shell curved in one ied mathematics and physics at Oxford University,
or two dimensions. An analysis of the structural behavior worked on two vital issues for the new ship-how it
of such a shell requires consideration of overall bending, would roll in high seas, and how much power and fuel
the tension and compression in the plane of the shell, would be needed to take her across the Atlantic. His
and the effects of local strengthening needed at the analysis of the rolling of ships was so successful that it
was still in use in the second half of the twentieth cen- fire , th e intera ction between foundations and soils ,
tur y. His work on the resistance of a ship 's hull to mov- earthquake engineering , the effects of wind on build-
ing through the sea soon led him to the conclusion that ings , and many more .
friction is not the only cause ; he realized that the surface
wave created by a ship also impedes its movement. This Indeed , in the ea rly days of any new branch of building
insight led him to believe , contrary to prevailing opinion , engineering, full -size and scale-model tests are still the
that models of ships ' hulls might indeed be used to pre- only means by which mathematical models can be tested
diet the resistance to movement of full-size ships. In and the accuracy of their predictions established prior to
1867 Froude tested models of two different hull shapes their use in engineering design .
in three sizes-3 , 6 , and 12 feet long-and found that
they generated similar wave patterns when towed at The Development of Concrete Shells
speeds proportional to the square of the length of the In building engineering , the first widespread application
models. He concluded that the wave resistance of the of nondimensional constants to model testing was in the
model and a full-size ship would be proportional if the development of thin , curved concrete shells in the 1920s
ratio V/ V(g D) for the model and prototype had the same and 1930s. These structures, more than any other,
value . This ratio , known as the " Froude number," is di- would exploit the three essential attributes of reinforced
mensionless , which means, effectively, that it is inde- concrete that set it apart form all other construction ma-
pendent of the scale of the model .5 terials-it can be made into a thin , load-bearing sheet of
un constrained or theoretically unlimited extent; it can be
Froude 's model tests were conducted in the open water formed into any three-dimensional shape ; and , by judi-
of the river Dart in South Devon , hardly ideal for precise cious placement of the reinforcement , it can be made as
measurements. Nevertheless, their success persuaded strong and stiff as necessary for a particular application.
the British Admiralty to grant Froude £2000 to build the No such material had previously been known . Here was a
world 's first towing tank near his home in Torquay, concept that had a unique appeal to engineers , who had
where , in 1871 , he conducted model tests that influ- always had the intellectual ideal of creating structures
enced the design of the HMS Greyhound . The sea trials th at used material only where it was absolutely neces-
a year later showed that Froude 's design method was a sa ry, and in minimum quantities . A designer of a re in-
complete success . forced concrete structure was , effectively, designing the
properties of the material itself.
Froude 's work formed the basis for one of the most pow-
erful tools available to the engin eer. As long as a de- Since curved concrete sh ells were id eally suited to
signer could find a suitable dimensionless constant, it fo rm the roofs of buildings , their design also necessar-
was possible to use a scale model to simulate the behav- ily inspired the search for ways to maximize the span of
ior of a full-size prototype. Froude 's method was used to roofs while minimizing their thickness. Maximizing the
great effect in wind tunnels in the early years of the twen- efficiency with which materials are used was virtually
tieth century, and for many other engineering problems the same goal that had been achieved in the evolution
involving fluid flow. of structural forms in the plant and animal kingdom
that had been identified by D'Arcy Thompson. In princi-
The development of experimental engineering science ple , thin shells of reinforced concrete could emulate
in general, and the use of scale models in particular, the shell of an egg, a flower petal , a seashell , or the
transformed nearly every aspect of building engineering exoskeleton of a crab . Crucially, seeking such a goal
design during the first half of the twentieth century. In would also appeal to inquiring minds in ways that steel-
addition to predicting the static and dynamic behavior of and concrete-frame structures did not. It would demand
structures , scale-model tests have played a major part an unprecedented understanding of both the behavior
in the development of design methods for acoustics , of structures and the properties of materials-an
the flow of air and smoke inside buildings , both at nor- understanding that Karl Terzaghi had brought to the
mal temperatures and at elevated temperatures in a analysis of soils and A. A. Griffith had brought to the
 .=-- --G15 _ _
568

fracture of materials. It would require the development marked that since it was not possible at that time to
of new mathematical models for structures and materi- make precise calculations for such a structure, the
als, and this, in turn , required experimentation to bending moments in the asymmetric and highly
establish the accuracy and effectiveness of those stressed shell had to be calculated using a series of it-
mathematical models. erative calculations to match the outward deformation
of the shell and the resistance of the tension ring .6 At
In 1920, however, it was far from clear how the goals of the center of the span , where the compressive forces
maximum efficiency might be achieved. Few concrete were highest, ten thin ribs were added to the top of the
contractors had the necessary knowledge of theoretical shell to prevent buckling.
engineering science ; universities and other research in-
stitutes, where the theoretical understanding was to be In the early 1920s, the development of thin concrete
found, were not equipped to test concrete structures , shells benefited the fortuitous collaboration of certain in-
and building and testing a full -size roof structure would dividuals and firms who had been brought together by the
have been as costly as it was impractical. Nevertheless, need to build a planetarium in Jena , Germany. One firm
some firms and individuals continued to make progress was Dyckerhoff & Widmann ; the other was Zeiss , which
as they used their experience to develop more effective made scientific optical equipment and had a strong tradi-
design and calculation methods. The contracting firm of tion of basing its product development on scientific re-
Dyckerhoff & Widmann (see chapter 7) was not the only search . The chief design engineer at Zeiss , Walter
company building concrete shells, but from the first Bauersfeld (1879- 1959), was a mechanical engineer
decade of the century on , it had gradually developed its who had developed several of the optical projection de-
expertise in designing and building some of the largest vices for which Zeiss was well known. In 1919 Bauersfeld
and most complex shell roofs then extant, using ribs to had a practical problem to solve: He wanted to design a
provide adequate stiffness for the thin shells when nec- new type of planetarium, in which the stars and planets
essary. By measuring the behavior of the shells it con- would be projected onto the interior surface of a perfectly
structed, the firm developed more accurate and reliable hemispherical dome from an array of projectors that
design methods and was able to build bolder and bolder could move to simulate the rotation of the earth . For
roofs. The roof of the Union Theater in Saarbrucken, nearly five years Bauersfeld and his team worked on the
Germany, for example, completed in 1923, spanned twin problems of creating the projection system and the
26.15 by 23.5 meters , with a rise of just 2.8 meters- hemispherical screen. His first task was to find a plane-
ssa an unprecedented span-to-rise ratio of 8.4 to 1. The faced polyhedron that approximated the shape of a hemi-
structural engineer for the dome, Professor Georg Ruth sphere to create a screen for each projector. The dome
(1880-1945) of the Technical University at Dresden, re- had to be very light, as it was to be constructed on the
 roof of one of the buildings at the Zeiss works. He set form . The thickness of concrete was chosen to achieve 511

about devising a workable structure without any of the the same span-to-thickness ratio as an eggshell, about
traditions and preconceptions of the building industry, 1:130. The combination of the framework and concrete
approaching it as an inventor and research engineer, covering was patented in 1923 as the Zeiss-Dywidag sys-
from first principles . tern , taking letters from the firm Dyckerhoff & Widmann
AG , for whom Dischinger worked. In 1924 Bauersfeld
Devising a regular grid of triangles to form a hemisphere and Dischinger designed and built a shell roof over a
was not, however, straightforward. By good fortune workshop, using the Zeiss-Dywidag system , for another
Bauersfeld knew Ernst Haeckel, the professor of botany firm in Jena; this one spanned 40 meters and had a
at the University at Jena who , in 1887 , had published thickness of just 60 mm . Their third collaboration on a 572

remarkable illustrations of many microorganisms . major structure was another planetarium at the Zeiss
Among these was the radiolarian Aulonia hexagona, the works , spanning 25 meters.
structure of which was a nearly perfect sphere made
from hexagons and pentagons (see chapter 7, p. 443). From his first project with Zeiss , Dischinger had recog-
Bauersfeld found that he could create a regular version nized the potential of this new structural form in the
of such a structure from an icosahedron with the vertices building industry; however, he also realized there would
cut to form twelve pentagons and twenty hexagons of be limited opportunities to use the circular plan. His
nearly equal area. This form not only provided the basis first thought was to build on an elliptical base , creating
for the screens , but also defined the arrangement of a roof with double curvature . The calculations of the
bars needed to construct a framework that could support forces in the shell would be extremely complex, and the
the screens. Each pentagon and hexagon was triangu- regular geometry of the geodesic Zeiss framework could
lated to give them the necessary rigidity and, in essence , not be used . This led him to recognize the benefits of a
came to resemble the triangulated domes that Johann cylindrical shell , curved in only one direction and able to
Schwedler had developed in the 1860s and August Foppl span as a long, curved beam in the other direction . Both
had developed in the 1890s, both of whose work the formwork and the calculations would be simpler, and
Bauersfeld had encountered during his engineering stud- the shell could be adapted to a rectangular building
ies . The skeleton of the prototype 16-meter-diameter plan. Bauersfeld and Dischinger constructed their first
dome comprised 3,480 steel rods each about 600 mm cylindrical barrel vault over a workshop at the Zeiss
long, and each manufactured to an accuracy of less than works in Jena in 1924; they were assisted in the analy-
one-hundredth of a millimeter- a degree of precision sis of this shell by the young engineer Ulrich
more characteristic of an optical instrument than a con- Finsterwalder (1897- 1988), who had joined Dyckerhoff
ventional roof structure . The elegant simplicity of the & Widmann a year earlier. During the following decade
final solution- which was , in essence, the first geodesic Dischinger and Finsterwalder, joined in 1926 by Hubert
dome- gives little indication of the extreme rigor with Rusch (1903- 79) , went on to develop methods for de-
which Bauersfeld had tackled the challenge. To form the signing all the main forms of shell- shells with single
surface on which to project the images of the stars , and double curvature built on circular, rectangular, and
569, 510 Bauersfeld had first considered using gypsum, but this polygonal plans .
proved difficult to waterproof. Only after the skeleton had
been constructed did he seek advice about how he might The process developed by Dischinger was essentially
best use concrete as a more water-resistant alternative that of the research scientist, for none of these early
to gypsum. He made contact with the engineer Franz shells had a useful precedent. First, a theoretical analy-
Dischinger (1887-1953), who had worked on a number sis was made considering only forces in the plane of the
of buildings at the Zeiss factory. The solution he pro- shell (membrane theory), combined with simple bending
posed was to use the newly developed Torkret process theory. To this was added an analysis to ensure that the
(named after the company that developed it), in which compressive forces would not cause the thin shell to
the concrete was sprayed onto a mesh of wire and sup- buckle. However, this was hardly a comprehensive model
ported from beneath by a lightweight movable timber of the system , and the next stage was to build and test a
569 570

571

568 (page 481) Union Theater, Saarbrucken, Germany, 1922-23. Engineer: Dyckerhoff and Widmann . Elevation.
569 Zeiss dome under construction, Zeiss glass works, Jena, Germany, 1923. Engineers: Walter Bauersfeld and Franz
Dischinger. 570 Completed Zeiss dome. 571 A geodesic structure under construction using the Zeiss-Dywidag system,
Jena, Germany, 1924. Engineers: Walter Bauersfeld and Franz Dischinger. Torkret (gunite) is sprayed over moveable
formwork.
 model of the shell . Not only the structure, but also the signed a circular shell roof for a factory that manufac-
material and the loads had to be modeled at a reduced tu red electrical machinery. Supported on columns at the
scale. The various parameters, including load, bending corners of an octagon , the shell , which was a portion of
moment, and stiffness , varied differently with the scale a sphere, was 40 mm thick, spanned 26 meters, and
factor, some linearly and others with the square or the had a rise of just 3.3 meters . The following three years
cube of the scale. To deal with this complexity, Dischinger saw the first use of Zeiss-Dywidag shells in public build-
and Finsterwalder were able to benefit from Froude 's use ings-several spectacular market halls at Frankfurt ,
of nondimensional ratios to ensure that the scale factors Leipzig , Budapest, and Basel. The twin octagonal domes 575, 577

were appropriate for representing the specific behaviors of the market at Leipzig were 65m in diameter at the
of the full -scale system being studied-the overall stress base of the 100mm-thick shell, with a clear span of
distribution or the buckling of the shell , for example. From about 80m at ground level . To raise his confidence in un-
careful measurements of the displacement of points on derstanding how the full-size structure would perform,
the surface of the model of the shell, and the displace- as well as undertaking a mathematical analysis of the
ment at its edges , the designers were able to establish behavior of the structure, Dischinger also built a 1:60
the forces in the shell and the degree of bending it would scale model of the dome and measured its deformation
be subject to across its span. By this means they were under a variety of loads. Dischinger's theoretical studies
able to reconcile the various calculated values of stress of polygonal shell vaults, and the experience he gained
and deflection derived from the different mathematical designing and constructing the Zeiss-Dywidag roofs at
models; they could then make minor changes to the pre- the Leipzig and Basel markets , formed the basis of his
dieted values , and the results formed the basis for spec- doctoral thesis , and he was awarded his doctorate in
ifying the thickness of the concrete and the size and 1929. Following this he developed a means for calculat-
disposition of the steel reinforcement. Finally, prototype ing doubly curved shells on a rectangular base . An early
sections of some of the shells were made to ensure that prototype using this form of structure was a modestly
the predictions of the structure 's behavior made from sized roof at a factory in Wiesbaden built in 1931. The
mathematical models and small-scale models were roof was built on a 7.3-by-7.3-meter square plan , and the
573 borne out at full size . Following this process it was possi- shell was just 15 mm thick at the crown. In 1930
ble to reduce the thickness of the shell to the absolute Finsterwalder was awarded a doctorate for his work on
minimum . And the thickness (or, rather, thinness ) of the the theory of cylindrical shells using the Zeiss-Dywidag
shells was astonishing: A shell roof spanning several system as exemplified in the market hall at Budapest.
tens of meters could be just a few tens of millimeters During the 1930s the firm of Dyckerhoff & Widmann was
thick. The economy in the use of materials that this made responsible for the construction of shell roofs all over
possible was truly remarkable: For example , the dome of the world , not only from its own offices- Hubert Rusch
St. Peter 's Basilica in Rome weighs about 10,000 tons, ran the firm 's headquarters in Buenos Aires from 1931
and the ribbed arch roof of the Breslau Jahrhunderthalle to 1934- but also selling licenses to use the Dywidag
weighs some 6 ,340 tons ; the roof of the market hall at system to firms in other countries. For example, the
Leipzig, which covers a larger area than both , has a rein- Austrian engineer Anton Tedesko worked for two years
forced concrete shell that is only 100 mm thick and with Dyckerhoff & Widmann in Germany before being
weighs just 2,160 tons. asked to establish the firm's presence in the United
States through collaboration with the Chicago firm of
In 1924 Dischinger and Finsterwalder designed their Roberts and Schaefer. The first Dywidag shells in the
first major cylindrical shell roof for "Gesolei "-the United States were the Hayden Planetarium in New York,
"Gesundheit, Sozialefursorge und Leibesubungen " completed in 1935, and an exhibition building at the
(Health, Social Welfare and Physical Exercise) exhibition Chicago World 's Fair of 1933. Most spectacular, per-
574 buildings in Dusseldorf in 1925. A year later they de- haps, was the ice rink built for the Hershey Chocolate

572 A dome under construction using the Zeiss-Dywidag system, Jena, Germany, 1924. Engineers: Walter Bauersfeld and
Franz Dischinger.
 574

576

577

573 Load testing a prototype barrel vault prior to erection of the Gesolei (Gesundheit, Sozialefursorge und [Link])
exhibition building, Dusseldorf, 1925. Engineer: Franz Dischinger. 574 Gesolei exhibition building, Dusseldorf, 1925.
575 Market hall, Leipzig, Germany, 1928-29. Rendering of interior. Engineers : Franz Dischinger and Hubert Rusch ;
architect: Hubert Ritter. 576 Hershey ice rink, Hershey, Pennsylvania, 1936. Engineer: Anton Tedesko. 577 Market hall,
Leipzig. Interior.
578

579

578 Schwarzwaldhalle, Karlsruhe, Germany, 1953. Engineer: Ulrich Finsterwalder. 579 Schwartzwaldhalle.
 Company in Hershey, Pennsylvania , in 1936 , which cov- concrete shells were seldom adopted and exploited by
576 ered an area 73 meters by 104 meters. architects. The reasons for this are complex, but a
major factor was probably the considerable technical ex-
Dyckerhoff & Widmann pioneered nearly every applica- pertise needed to design them. Other reasons may be
tion of shell roofs , offering building designers unprece- that their scale is too large , and that they tend to have
dented column-free spaces beneath the roofs of market little impact on the elevation of a building. Another may
buildings , exhibition halls , aircraft hangars , bus depots , be that the concrete shell is opaque ; architects in the
railway stations , railway platforms , sports halls , and , nineteenth century embraced transparent iron and
most numerous of all, factory buildings. One of the glass roofs with enthusiasm , and in the 1980s and
largest of the factory roofs was built for Ferdinand 1990s much use has been made of translucent , tensile,
Porsche 's Volkswagen works in Wolfsburg, Germany, membrane structures. What seems to have been the ar-
which covered 166 ,000 square meters. chitect 's loss , however, was the engineer's gain.
Concrete shells-both in their simplest form and when
In addition to their work on shells , Dischinger, used in combination with stiffening ribs and other struc-
Finsterwalder, and Hubert Rusch were also pioneers in tural elements , whether cast in situ or constructed from
prestressed concrete bridge construction in the late precast un its or exploiting the advantages of prestress-
1930s and 1940s. In 1953 Ulrich Finsterwalder ing- have inspired what was surely the most creative
designed one of the first prestressed concrete roofs for period in the history of structural engineering since the
the Schwarzwaldhalle , in Karlsruhe , Germany, an audi- late Middle Ages , the era of the great cathedrals .
torium with seating for 4 ,500 spectators and a stage
that could accommodate 1 ,500 performers. It had a During the years between about 1930 and the 1960s, a
suspended saddle-shaped roof that spanned 73.5 by dozen or so structural engineers became known well be-
578, 579 48.6 meters . yond the world of engineering for their mastery of rein-
forced concrete and brickwork and the extraordinary
The contributions made by Dischinger, Finsterwalder, and architectural forms they created: Pier Luigi Nervi (1891-
Rusch to the art of designing reinforced concrete struc- 1979), Eduardo Torroja (1899- 1961), Felix Candela
tures can hardly be overestimated . These engineers de- (1910- 97 ), Eladia Dieste (1917-2000), Riccardo
veloped an extraordinary understanding of material and Morandi (1902- 89 ), Anton Tedesko (1903-94) , Eugene
its performance through a combination of increasingly Freyssinet (1879- 1962), Oscar Faber (1886-1956) ,
complex theoretical analyses of shells and careful obser- Owen Williams (1890-1969), Ove Arup (1895-1988),
vations of the behavior of real structures , both scale mod- Bernard Lataille (1900- 55), Nicolas Esquillan (1902- 89),
els and at full size . A series of seminal publications Rene Sarger (1917- 88) , and Heinz Isler (b. 1926).
describing the methods they had developed for designing Although all these engineers collaborated with archi-
concrete shells was distributed throughout the world , and tects , they also worked independently, virtually creating
their work was also embraced by designers who wanted to their own architectural vocabulary, a phenomenon un-
use reinforced concrete for applications other than shell heard of at any other time in the history of building engi-
roofs . Dischinger and Rusch 's influence continued neering. They were able to build upon the work of
through their educational and research work as profes- Dischinger and Finsterwalder through the technical pa-
sors of concrete construction at the Technische pers and textbooks those men had written, and, most im-
Hochschule in Berlin , from 1933, and the Technische portantly, through adopting and developing the
Hochschule in Munich, from 1948, respectively. experimental approach that Dischinger had pioneered.
The use of both model testing and prototype testing ,
By the late 1920s reinforced concrete was no longer together with the increasingly sophisticated techniques
just a material that provided a fireproof alternative to of experimental stress analysis as part of the design
the steel frame. It allowed the construction of an en- process, allowed innovative structural engineers to take
tirely new family of structural forms whose visual and ar- their designs beyond the range of familiar experience
chitectural impact could not be ignored. Nevertheless, with confidence.
 en
CIO
~

DRAWINGS AND COPYING

From the 1920s on , the reproduction of drawings by the blueprint process began to be replaced by the diazo , or
dyeline. Developed in the 1890s, the diazo method was also a photographic process but produced a positive copy,
rather than a negative, of an original on translucent paper or linen. Depending on which chemicals were used to
impregnate the copy paper, the lines produced were magenta, blue, black, or brown . The Diazotypes or 0zalids , the
two best-known trade names, were much quicker to produce than blueprints and did not require a washing bath.
A further advantage over blueprints was that diazo copies could be made of lines drawn with pencil, which allowed
draftsmen to correct their drawings more easily since pencil is erasable but ink must be scraped off the paper.
There were also some disadvantages. The diazo developing process involved the use of ammonia , which lingered
on the copies for some time and created an unpleasant atmosphere in the print rooms of drawing offices . Also, the
copies remained sensitive to light , which meant they had to be stored either in plan chests or rolled up. The use of
tracing cloth declined in the 1930s because of its high cost, but tracing paper remained in use until the 1950s,
when polyester films such as Mylar were introduced .

580 581

580 Engineering design office of R. Travers Morgan & Partners, London, c. 1968. 581 Detail of engineering drawing on
woven linen cloth, c. 1930.
 The Spanish engineer Eduardo Torroja, who began using load and simulated wind and snow load , and compared
concrete shells in the mid-1930s, created some of the the results with those predicted by calculation. They were
most innovative structures of their kind . For the canopy found to be close enough to convince Torroja that his un-
at the Madrid hippodrome (1935), for example, he used derstanding of the behavior of the structure was suffi-
a concrete shell resembling a gently curved and folded cient to go ahead and build it. When the roof was
582 sheet of paper. The canopy cantilevers 13 meters over complete , he measured its actual deformations and com-
the seating, and yet is just 50 mm thick. To help estab- pared them with the predictions made from calculations ,
lish where best to place the tensile steel reinforcement, from the model tests , and found they agreed to within
Torroja augmented his general understanding of struc- 100 mm. Not only did this give Torroja final confirmation 591

tural behavior by studying the behavior of a simple of the soundness of his design for this roof, it also raised
model made of thin card , to sketch the approximate pat- his confidence , and that of all other engineers engaged in
583 tern of the internal stresses. This gave him a basis for similar projects , in the power of his design procedure us-
calculating the thickness and spacing of the steel rein- ing model tests combined with relatively simple mathe-
584 forcing bars. To verify Torroja 's understanding of this matical calculations.
unprecedented structural form , and the reliability of the
design calculations, the contractor built and tested a At the Pont de Suert church (1952) in the Spanish
585 prototype unit. The test also served as a trial of the Pyrenees, Torroja devised an innovative way of making a
reusable formwork devised for the construction of the complex , curved structural shell that could be built with-
canopy. Undertaken at a time when steel was in short out using the timber formwork that was then standard
supply, the contractor broke up the prototype shell after practice when making shells of reinforced concrete . \He
the test and reused the reinforcement to make one unit combined the traditional technique of the Catalan brick
of the final roof. vault (see chapter 7 , p. 386 ) with steel reinforcement to
form the five pairs of shells that make up the nave . The
Possibly the boldest structure Torroja designed was the shells , each comprising three layers of hollow brick tiles
roof over the Front6n Recoletos sports hall in Madrid 30 mm thick , we re built without formwork , using guides
(1935). He proposed a cylindrical shell in two lobes span- to define th e curved shape of the vault and qu ick-setting
ning 32 meters wide by 55 meters long, with a concrete mortar for the inner layer. The steel re inforcement was 589, 590

shell that was 80 mm thi ck everywhere but at the inter- laid on the outer surface and covered by the outer layer of
586 section of the two lobes , where it was 150 mm. As with cement mortar facing . As a vault is not a fun icular shape
the Hippodrome roof, Torroja first had to understand in in cross section (as masonry vaults must be ), it needs to
qualitative terms how the structure worked . A simple resist some bending. The curved profile of the lobes en-
model made of card , fixed to model end walls made of ables the thin shell, just 200 mm thick, to resist th is
wood , confirmed the overall stiffness of the curved form bending, just as a seashell does . 588

and the need to provide firm support to the longitudinal

edges of the curved shells , especially where the two Like Torroja , the Italian engineer Pier Luigi Nervi was a pi-
lobes intersected. Because the necessary calculations oneer in the use of scale models during the development
for this asymmetrical structure form were so complex , of many of his structures, both to develop a better under-
the mathematical model had to be greatly simplified . This standing of how they worked and to predict the deforma-
made the calculations of stresses and deflections partic- tions that the full-size structure would experience under
ularly uncertain, especially in the perforated parts of the load. Over many years Nervi collaborated with Professor
shell containing the windows. To inspire sufficient confi- Arturo Danusso (1880- 1968) at the Polytechnic Institute
dence in his innovative design Torroja built and tested a of Milan , who developed many techniques for building
587 1:25 scale model. He measured deflections under dead and load ing scale models of structures and measuring

582 Hippodrome, Madrid, 1935. Engineer: Eduardo Torroja. 583 Hippodrome, Madrid. Diagram of the principal stresses in
the canopy. 584 Hippodrome, Madrid. Prototype of the canopy, showing tensile steel reinforcement. 585 Testing the
Hippodrome canopy prototype with sandbags.
 deflections and strains. Nervi frequently required model probably the cheapest because Nervi was a contractor as
tests to complement theoretical calculations simply be- well as a designer, and conceived all his remarkable
cause he so often designed structures that were break- structures with their method of construction foremost in
ing new ground. One example was an aircraft hangar he his mind .) Model tests were essential for such an innova-
designed and built for the Italian Air Force in 1936. It cov- tive structural form, one whose asymmetry and high de-
ered an area 111.5 by 44.8 meters , and one of the long gree of statical indeterminancy meant that considerable
sides had to provide access through doors that were 50 simplifications and approximations had to be made to
592 meters wide . Despite its apparent complexity, Nervi 's de- yield a practical mathematical model. The model tests 593a -c

sign was selected because it was the cheapest. (It was confirmed Nervi 's prediction that the many structural el-

586 Front6n Recoletos sports hall, Madrid, 1935. Engineer: Eduardo Torroja. 587 Front6n Recoletos. Model, 1935.
588 Pont de Suert church, Lleida, Spain, 1952. Engineer: Eduardo Torroja. Plaster model of one bay of the nave. 589 Pont
de Suert church. Nave, constructed of steel-reinforced brickwork. 590 Pont de Suert church. Cross section of reinforced
brickwork shell. 591 Front6n Recoletos. Illustration showing comparative deformations of the roof, 1935.
 Exterior

First layer of
hollow bricks
(with gypsum)

Second and third

layers (with
cement mortar)

Cement morter
facing

Gypsum and

- - - - - -- - - - -
Theoret ical deformation Actual deformation
0 Deformation of model
600 millimeters
Actual deformation
592

ements would work much better in combination than he Probably Nervi 's greatest innovation was a new material
was able to justify using the highly simplified theoretical he called "ferro-cemento," which was concrete made
calculations. They also demonstrated that the dimen- using fine aggregate (sa nd) reinforced with a mesh of
sions he had chosen in the preliminary design would wire just one or two millimeters in diameter. He used
largely be satisfactory, and only a few of the diagonal ribs this material to make very thin precast elements that
of the lamella roof had to be increased in thickness . could be assembled on formwork and stitched together
Measurements of the deformation of the actual structure using conventional steel reinforcement and concrete
taken during and after construction showed that the re- placed in situ . This was the innovation of a practical
sults of the model tests had been sufficiently accurate. builder, and he used many versions of this technique to
They also enabled Nervi to monitor the plastic and vis- create some of the most remarkable and beautiful struc-
cous deformations of the structures caused by the ten- tures ever built. One of these was the large sports
dency of concrete to continue to deform under prolonged palace built for the Rome Olympics in 1960. The corru- 594

loading. Further deformation ceased after about five gated dome, which spans just over 100 meters, is made
years. Toward the end of his life, Nervi still considered from V-shaped ferro-cement units around 1 meter deep
the technical studies he undertook for this hangar among and 4.5 meters long, with walls just 25 mm thick. 596

the most complex of his career.

592 Aircraft hangar, Orvieto, Italy, 1936. Engineer: Pier Luigi Nervi. 593a-c Aircraft hangar, Orvieto. Model used to
measure deformations and calculate stresses in the grid shell. Combinations of gravity and wind loads were
applied using weights, and deformations measured using clock gauges.
 - ~ -::-' ~_.J--:1 _J __J _j ~~
~~- - ~~~=-:==~~:·
593a

593b 593c

Eugene Freyssinet was one of several French engineers needed to create the curved shape in concrete can be 597, 59a

who began experimenting with using concrete in the built using straight and flat timber planks . 599, 600

form of a thin sheet , for example , for the airship

hangars at 0rly (1921- 23 ), which were some 200 Hyperbolic paraboloids caught the imagination of many
meters long and 60 meters high , spanning over 85 architects in the 1940s and 1950s , both because of
595 meters . The deeply corrugated reinforced-concrete their form and because they provided large column-free
arches , acting structurally as a folded plate , were just spaces needing only a few supports around the perime-
90 millimeters thick. Sadly, they did not survive ter. The acknowledged master of the form was the
American bombs in 1944. Spanish engineer Felix Candela , who constructed many
hundreds of buildings with roofs made using hyperbolic
The French engineer Bernard Lafaille built many types of paraboloids of reinforced concrete , mainly in his
concrete shell in the 1920s before beginning to experi- adopted home , Mexico. Some were used for mundane
ment, in about 1933, with a new structural form for industrial buildings, but many others were used in
shell structures , the hyperbolic paraboloid. This family churches of considerable architectural merit, and
of anticlastic curved surfaces can be generated by mov- their appearance was always striking. Candela 's tour
ing a straight line through space, and is unique in the de force was his restaurant Los Manantiales at 601

history of structures as being the only form that was Xochimilco near Mexico City (1958) , constructed using
identified using mathematics and engineering science his own versionof ferro-cement. It spans 42.5 meters
before it had been conceived and constructed by an en- and yet is just 42 millimeters thick, except at the free
gineer. It has the practical advantage that the formwork edges which are slightly thicker. Its form is generated
594

595 596

594 Palazzo dello Sport, Rome, 1960. Engineer: Pier Luigi Nervi. 595 Airship hangars, Orly, France, 1921-23. Engineer:
Eugene Freyssinet. 596 Palazzo dello Sport. Diagram of typical precast unit made from ferro -cement. 597 First experi·
mental hyperbolic paraboloid shell built by Bernard Lataille, 1935. 598 Hyperbolic paraboloid shell roof covering storage
sheds at Alfa-Romeo works, Milan, 1937. Engineer: Giorgio Baroni. 599 Sketches of proposed applications of hyperbolic
paraboloid shells by Fran~ois Aimand, c. 1934. 600 Diagram of the first experimental hyperbolic paraboloid shell built by
Bernard Lataille, 1935. 601 Iglesia de San Jose Obrero, Monterrey, Mexico, 1959. Engineer: Felix Candela.
602, 603 by four intersecting hyperbolic paraboloids . and " Earthquake Milne ." Milne arrived in Tokyo in 1875
to teach mechanics, and gradually became interested in
The Uruguayan engineer Eladio Dieste also built many the nature of earthquakes . Following an earthquake in
shell structures in the 1940s and 1950s , mainly using 1880 he invited two other British scientists to join him ,
reinforced brickwork rather than reinforced concrete. and together they developed various ways of measuring
This has the advantage that the external finish was less and recording the actual movements that make up an
monotonous , and does not require the expensive form- earthquake . What they discovered over the next two
604 work needed to make concrete shells. The largest con- years was contrary to then-current theories: An earth-
crete shell roof ever constructed is the double-skinned quake was not a single big bang whose effects then died
shell of the Centre des Nouvelles Industries et down , nor was it a regular periodic movement. Rather, the
Technologies (CNIT) exhibition hall , built in Paris between motions were of long duration and irregular in frequency,
1956 and 1958. The span between the corners of its tri- and occurred in all three dimensions . After his two col-
angular plan is 208 meters. The enormous span was leagues returned to Britain , Milne stayed in Japan and
made possible by the ingenious design of the French en- studied the movements of several two-story buildings
gineer Nicolas Esqu illan , who devised a lightweight during earth tremors. He found that timber buildings
scheme using two shells about 3 meters apart and only moved more than those built of brick , and that the move-
65 mm thick . These are joined by a series of diaphragm ments on the second floor were greater than those on the
walls, creating what is effectively a series of thin-walled first (ground) floor. He also found that a building built on
tubes. The top and bottom surfaces of the tubes are what was locally considered to be "soft ground " moved
curved , resembling a straw or the fibers in a tree trunk, to more than a similar one built on "hard ground. "
increase the capacity of the walls to carry compression
605, 606 forces without buckling. In 1880 Milne founded the Se ismological Institute of
Japan , and Fusakichi Omori (1868- 1923) became its di-
By the 1950s , in addition to their use for large and spec- rector. In the 1890s Milne and Omori conducted the first
tacular roofs , concrete shells were being used through- "shaking table " experiments to simulate the movement
out the world , for many tens of thousands of roofs of of an earthquake on model buildings . The table , re-
moderate span , in buildings such as markets , school strained by springs and mounted on wheels to allow it to
halls , factories , sports and exhibition halls , railway and move horizontally, was pulled sideways by ropes and sud-
bus stations , gas stations, garages , and warehouses. denly released to simulate the movements of an earth-
Having once required the efforts of the world 's foremost quake . Omori went on to study the behavior of bri ck
engineers , concrete shells had become mainstream columns in earthquakes . Apart from Omori little local in-
building products supplied off the shelf, so to speak , by terest in seismology developed , and by 1892 the
specialist contractors. Seismological Institute had ceased to exist. Milne re-
turned to Britain in 1895 after his laboratory was de-
The Dynamic Behavior of Buildings stroyed by fire , and set up the Shide Seismological
The idea that buildings could be designed and con- Observatory on the Isle of Wight. For nearly twenty years
structed to res ist damage by severe earthquakes devel- the Shide Observatory was , effectively, the world head-
oped in Japan and the West Coast of the United States quarters for earthquake seismology.
during the 1920s. The key issue was understanding ex-
actly what forces a building had to resist . This question Worldwide interest in earthquakes was revived after the
was first studied in depth in Japan in the 1880s, not by great earthquake in San Francisco , California of 1906,
the Japanese , but by three British engineering scientists and in that year F. J. Rogers , a Stanford University physics
who had taken jobs at the Imperial College of Engineering professor, bu ilt what he called a "shaking machine ,"
in Tokyo. The leader of this team was John Milne driven by an electric motor and crank , to study the effect
(1850-1913) , later known as "the father of seismology" of simulated earthquake movements on soil. In 1915

602 Los Manantiales Restaurant, Xochimilco, Mexico, 1958. Engineer: Felix Candela. 603 Los Manantiales Restaurant.
605 606

604 Chiesa parrocchiale, Atlantida , Uruguay, 1955-60. Architect and engineer: Eladio Dieste. 605 Centre des Nouvelles
industries et technologies (CN IT), Paris, 1956-58. Engineer: Nicolas Esquillan. Under construction. 606 CN IT.
 Riki Sano (1880-1956), a practicing engineer who was continued working in the field of earthquake engineering
also a professor at the University of Tokyo , wrote the as a consulting engineer in private practice. In 1964 he
first textbook on the subject, Earthquake Resistant returned to Stanford to continue his research after a gap
Building. More elaborate shaking tables were con- of over thirty years, and earned his Ph.D. degree in 1967
structed in Tokyo 's Earthquake Research Institute , at the age of fifty-nine . A few years later Stanford re-
founded in 1924, and at the California Institute of sponded to his suggestion for establishing a dedicated
Technology (Caltech ) in Pasadena after the 1925 Santa research center, and , when it opened in 197 4, the univer-
Barbara earthquake. The first systematic research into sity honored its former student by naming it the John A.
earthquake-resistant timber housing began in 1929, Blume Earthquake Engineering Center.
both at the Tokyo Earthquake Research Institute and by
Professor Lydik Jacobsen (1897-1976), the director of In 1956 Pier Luigi Nervi and the Italian architect Gio Ponti
the Vibration Laboratory at Stanford . Rather than assum- (1891-1979) faced the challenge of designing a thirty-
ing a building to be a static structure subject to a dynamic two-story building in reinforced concrete for the Pirelli
load , Jacobsen recognized the need to consider also the Company in Milan . High-rise steel-frame buildings were
dynamic response of the building. In 1930-31 he and his standard products in the 1950s, especially in the United
research students undertook the first shaking-table tests States, but concrete-frame buildings were not. Concrete
on a model of a proposed multistory building-the has the advantage that it does not vibrate as readily as
Olympic Club in San Francisco. This work used a mono- steel , and oscillations induced by wind die down more
lithic model to study the different modes of vibration of rapidly. This was an important design consideration for
the building, and highlighted the need to make a more so- the 33-story, 127-meter high Pirelli Tower, which was, at
phisticated model, one that would take into account the the time, the only high-rise building in Milan. The building 6oa
independent movement of each story of a proposed is about 70 by 18 meters in plan , and has four main
building. Jacobsen 's work inspired the next generation of columns and four secondary columns, which are com-
research students, including John Blume (1909-2002), bined with the stairwell and shear walls at each end of
who devised and built a 1:40 scale model of the pro- the building. From ground level , the cross section of the
607 posed fifteen-story Alexander Building in San Francisco columns diminishes as the load they have to support re-
in 1932-33. The model was a masterpiece of mechani- duces . Slender floor beams span up to 20 meters be-
cal engineering design , and allowed each story to move tween the columns and are cast integrally to create a rigid
in five independent ways: two horizontal , one vertical , frame that provides resistance to wind loads . Nervi used 609, 610

and two rotational about horizontal axes . A mass of alu- his experience of model testing during the previous 61 1

minum and steel plates and tubes, steel springs , and twenty years to augment and confirm his design calcu la-
steel ball bearings allowed each floor to roll. Movements tions for wind loads . In addition to measuring static 612

were displayed by pointers at every story. The analysis deflections , he studied the response of the building to dy-
undertaken by Blume and other research students namic loads to ensure that there was no possibility of the
demonstrated how modes of vibration could be calcu- wind causing potentially disastrous vibrations near the
lated from construction drawings of a building. It showed building 's natural frequencies. A 12-meter-tall model at a
the effect of floors and nonstructural elements such as scale of 1:10 was built and tested at the Experimental
partition walls on the stiffness of a building, and the Institute of Bergamo using materials that modeled the
damping effect they contributed. It also demonstrated strength and vibration characteristics of concrete. Wire
that buildings deflect more markedly by bending as a can- strain gauges provided electrical signals that could be
tilever fixed in the ground than by the horizontal shear be- displayed on the screen of an oscilloscope to identify the
tween one story and the next. This work by Blume and his various natural vibration frequencies of the structure and
researchers finally raised the understanding of the ef- to study how the vibrations decayed.
fects of earthquakes on high-rise buildings to a level that
allowed structural engineers to consider loads due to Structures in the Third Dimension, Spatial Structures
earthquakes when designing their buildings, so providing The German mathematician August Mbbius (1790-
structures that would safely resist these loads. Blume 1868) had shown in his Lehrbuch der Statik (Textbook of
sm wa

Statics ) (1837) that, for a two-dimensional framework to trusses were already being widely used for roof struc-
be fully triangulated and statically determinate , there is tures in the 1860s, and Johann Schwedler had designed
a simple relationship between the number of nodes (n) many domes using a triangulated , statically determinate
and the number of bars (k) working in either tension or framework , also in the 1860s (see chapter 7, p. 439) . In
compression , namely, k = (2n - 3 ). He also showed that the first decades of the twentieth century, three-dimen-
for a three-dimensional framework , the relationship is k sional statically determinate and statically indetermi-
= (3n - 6 ). Unfortunately, his work remained unknown to nate frameworks were widely used in the young airship
structural engineers , and both of these simp le relation- and aircraft industries , where minimum weight was the
ships were rediscovered independently by engineers overriding design constraint. In buildings , however, the
later in the century: the rule for two dimensions by Otto only significant uses of a three-dimensional framework
Mohr in 187 4 and the rule for three dimensions by during this time were the geodesic domes and barrel
August Fbppl in 1892. Meanwhile , two-d imensional vaults devised by Bauersfeld and Dischinger for their
 ~~---
....._...- ]«:"i~T \
_________'___ , __ ,,_,,,,_, _____ _

·- --------.---======- ---~.:: ___ _

-----~ - - - - - - - - - - - - ------ -,

_:·:!i~~~~:::::f.: ) {~~:,_: :::::::::::::::::::1

~:~::_ ~ ~-=--=--=--=--=--=--=--=--=--=--=--=--=--=--=--=--=-j

610

611 612

607 Scale model (1 :40) of the Alexander Building in San Francisco, California, used to measure dynamic deformations
induced by earthquakes, 1932-33. 608 Pirelli Tower, Milan, 1958. Engineer: Pier Luigi Nervi; architect: Gio Ponti.
609 Pirelli Tower. Plan showing large cross section of main columns near the foot of the tower. 610 Pirelli Tower. Interior.
611 Pirelli Tower. Section. 612 Pirelli Tower. Scale model (1 :10) to assess dynamic behavior arising from wind loads.
 616

61 5

613 Max Mengeringhausen displaying nodes of his MERO system. 614 Railway Station, Adamstown , Dublin , Ireland , 2006.
Eng ineering design and construction of roof by MERO (UK) ; architect: Building Design Partnership. 615 Union Tank Car
Company, near Baton Rouge, Louisiana, 1958. Engineer: R. Buckminster Fuller. Geodesic dome spanning 118 meters.
616 MERO system. Detail of steel node connecting multiple members of framework.
 planetariums and many Zeiss-Dywidag shell roofs , where systems were available in aluminum or plastic ; these had
the framework was embedded in concrete , thus effec- not been available in 1920. By the mid-1980s more than
tively becoming part of the reinforcement. 300,000 geodesic domes had been built in sizes rang-
ing from a few meters in diameter to over 100 meters. 615

The potential of three-dimensional frameworks to span

and enclose large spaces remained unexplored until Prestressing
the 1930s , when the German engineer Max The idea underlying prestressed structures is very old .
613 Mengeringhausen (1903-88) began developing ways of Masonry structures derive their stability from the weight
assembling buildings quickly using a standard kit of bearing down on the lower parts of the building, and me-
parts, a method that was already common for mass dieval cathedral builders often added pinnacles to but-
production of cars and other items . His investigations tresses to increase their stability. Outside the building
into the many ways in which identical shapes can be industry many structures, from ancient sailing ships to
stacked or arranged took him into three-dimensional umbrellas and bicycle wheels , derive much of their stiff-
geometry, crystallography, and the study of organic ness from prestressing using ropes or other tensile mate-
forms through the work of D'Arcy Thompson. For the rials. Prestressed concrete began to be used for bridges
assembly of frameworks, he quickly realized that the in the 1930s. The appeal of prestressing to the structural
key to a construction system lay in the method of con- engineer is that it effectively enables a cable to be loaded
necting the bars. The elegant solution he eventually in compression and, hence , allows the designer to avoid
devised , like most good engineering, seems incredibly bulky struts. One of the most elegant examples of pre-
simple and obvious . The steel node has threaded holes stressing was the Skylon , built in 1951 for the Festival of
that allow up to eighteen members from different direc- Britain in London . The Skylon was a 76-meter tall column 617, 61a

616 tions to be connected at a single point. The bars were made from aluminum, suspended 12 meters above the
made in a series of standard lengths to create various ground by just three cables at the base and stabilized by
types of framework , in different cross sections, to three more cables at its midpoint. These supports were
accommodate the forces they needed to carry. barely visible , even when seen close-up, and the column
Mengeringhausen used his MERO system for various appeared to float above the ground.
temporary structures during the 1940s before using it
as a lightweight roofing system in the 1950s, including In 1952, the roof of the new Raleigh Livestock Arena in
614 as a support for roofs of glass or transparent plastics. North Carolina (now known as the Dorton Arena) became
The large numbers of members and nodes made the the first substantial prestressed cable-net structure
full structural analysis of such space trusses rather ever built. The Norwegian-born engineer Fred Severud 619,620
tedious , and engineers usually managed by using sim- (1899-1990)-later known for his work on the Ingalls
plified calculations. The simple and repetitive geome- Hockey Rink at Yale University and the Gateway Arch in
try, however, made such structures ideal material for St. Louis (see pages 520-21)- was called in to complete
the engineers who were writing the first computer pro- the structure that had been designed by Polish architect
grams for structural analysis in the late 1950s. Matthew Nowicki (1910-50). Killed in a plane crash in
Egypt , Nowicki did not live to see his unique design come
Like Mengeringhausen, the American inventor Richard to fruition . The doubly curved arena roof consisted of two
Buckminster Fuller (1895-1983) saw the potential of virtually orthogonal sets of steel cables stretched be-
three-dimensional frameworks made of standard parts tween inclined concrete arches, and it was clad with steel
as the key to rapid construction. Structurally, the geo- sheeting. The shape of this roof was simple enough, and
desic dome patented by Fuller in 1954 was no different the steel cables were few and regular enough for the roof
from that developed in 1920 by Bauersfeld for his plane- to be calculated as an elastic prestressed structure.
tarium . Fuller, however, saw its potential as a lightweight Such a calculation procedure was already familiar to en-
roof structure that could be transported easily and as- gineers with experience designing prestressed concrete
sembled rapidly. The key difference lay in the building en- structures. The Raleigh Arena proved to be a tremendous
velope: By the 1950s a variety of lightweight cladding inspiration to many architects, including Eero Saarinen
 I ,

617 618

and Kenzo Tange, who wanted to exploit the architectural hanging structures , for which engineers were able to
possibilities of tensile structures. What they were able to make the necessary calculations.
design conceptually, however, was severely limited by the
very restricted range of simple geometrical shapes for In 1951 Fred Severud 's design office in New York was
which engineers were then able to undertake calcula- working on the plans for the Raleigh Stadium when they
tions: The calculations could only be done for surfaces were visited by a German student of architecture, Frei
that could be uniquely described using mathematical Otto (b. 1925). He , too, was impressed by the possibil-
equations. And from a practical point of view, in the days ities that doubly curved prestressed surfaces would
of hand and slide rule calculations , there was a low limit seem to offer architects. Returning to Stuttgart in
to the number of simultaneous equations that could be Germany, he began making models with a variety of
solved. The large structures of the 1950s and early materials-string, chains, nets, soap bubbles, and
1960s thus still belonged within the tradition of freely elastic sheets-to generate a range of curved shapes

617 The Skylon at the Festival of Britain, London, 1951. Engineer: Felix Samuely; architects: Philip Powell and Hidalgo
Moya. 618 The Skylon. Approximate prestressing forces: (A) 50 tons (tension); (B) 100 tons (t); (C) 150 tons (compres-
sion); (D) 150 tons (t). 619 Raleigh Livestock Arena (now Dorton Arena), Raleigh, North Carolina, 1952. Engineer: Fred
Severud; architect: Matthew Nowicki. 620 Raleigh Livestock Arena.
 619 620

and surfaces. Working with the long-established maker The high specific strength and stiffness of timber has al-
of circus tents, Strohmeyer, Otto soon devised his first ways made it especially suitable for large-span roofs. In
full size doubly curved prestressed tensile structure in the United States and Canada there was a long tradition
1955 , a modest canopy under which musicians could of building timber-truss railway bridges , partly because
621 perform outdoors. It spanned 18 meters across the timber was so readily available and also because it was
diagonals, which was twice the maximum span found in easy to transport over long distances to locations with no
Strohmeyer 's traditional circus tents. locally grown timber. For the same reasons, and because
the skills needed to erect timber structures were readily
The fact that the form was simple and the span rela- available , American engineers used timber for many
tively small , by today 's standards, is deceptive. This large roofs in the 1930s, 1940s, and 1950s when engi-
was not merely a sheet stretched between a couple of neers in Europe were more inclined to use reinforced
tent poles . It was a fully engineered roof structure. The concrete or steel.
stress in the fabric was 1 .6 tons per meter width , and
Strohmeyer had to weave custom-made fabric from The high specific stiffness and strength of wood seemed
high-quality cotton that was equally strong and stiff in to make it an ideal material for aircraft, but in an age
both directions , unlike normal tent fabrics . Steel cables when adhesives were weak and water-soluble, this ad-
16 mm in diameter were sewn into the edges of the vantage was offset by the weight of nails or screws
membrane, and their length and the precise shape of needed to assemble the wooden components. Suitable
the curved boundaries of the fabric were established adhesives were developed in the 1940s for the manufac-
from tests on scale models. The cutting pattern for the ture of plywood and other built-up wooden elements used
individual strips of fabric also had to be carefully in aircraft-for example, the laminates that formed the
shaped and measured using physical models. monocoque plywood fuselage of World War ll's Mosquito
aircraft. The research to create a strong and waterproof
New Structural Materials adhesive led to the development of epoxy resin glues,
New structural materials do not come along very often: and these opened up an entirely new range of timber
cast iron in the 1790s, reinforced concrete in the 1890s. products for use in the building industry, including both
The decade immediately after World War II saw the arrival structural materials, such as Glulam, laminated veneer
of two more , both of which owe their development to the lumber (LVL), and plywood, and nonstructural products,
wartime aircraft industries: glued, laminated timber such as chipboard and fiberboard. This benefited the
(Glulam), and aluminum. building industry after the war, especially in the United
 States and Canada, where timber was plentiful and steel In one sense , "total design " represents the culmination
623 , 625 in short supply. of Arup 's own experience and career- from his first work
designing buildings with architects in the 1930s, to
Aluminum was also in abundance immediately after founding his consulting engineering practice Ove Arup
World War II , when it was no longer needed for building and Partners in 1946 , and , finally in 1963, the formation
aircraft. The Dome of Discove ry built for the Festival of of the daughter firm Arup Associates , with design teams
Britain in 1951 was probably the largest aluminum comprising engineers of different disciplines and archi-
structure ever constructed ; aluminum was chosen not tects delivering a fully integrated building design. In an-
only for its lightness but also because it captured the other sense , it was the conclusion of a much longer story,
spirit of the postwar years , which inspired many new im- indeed , arguably, the history of building design itself.
622, 624 ages , ideas , and designs. A particular advantage of alu- After the si xteenth century, as more and more design is-
minum is that , unlike any other material, it can be easily sues became too technical for generalists to master, the
manufactured in many shapes using three different field of building design as a whole splintered into various
processes- rolling , casting , and extrusion. This ap- specialties . "Total design " represented a return to the
pealed to the sculptor, architect , and engineer Jean ideal of a process whereby many different specialties
Prouve (1901-84), and he exploited the characteristics could be considered together when designing a build-
of all three processes when he designed , in 1954, the ing-not by a single individual , but by a team.
pavilion for an exhibition in Paris celebrating the centen-
626 nial of the discovery of aluminum. Collaboration between architect , engineer, and builder
was not a new idea . Since the early nineteenth century
It is worth noting how easily these new materials were in- growing numbers of architects and engineers had worked
corporated into the structural engineer 's repertoire . By together to create designs that were , so to speak , greater
the 1950s engineers had learned to analyze materials than the sum of their individual parts : Saint George 's Hall
from the point of view of scientists rather than craftsmen . in Liverpool and London 's Crystal Palace were two promi-
Engineers now knew that materials had a series of nent examples. And the high-rise buildings in late nine-
unique generic properties , such as strength , stiffness , teenth-century America were made possible onl y by an
ductility, hardness , creep , fatigue behavior, and so on . unprecedented degree of collaboration between engi-
Knowing these properties meant that it was relatively neer, architect , and building contractor. But the total de-
easy for engineers to incorporate a new material in their sign concept- the conscious aim to achieve better
designs- a far cry from the early days of cast iron or rein- design by forming an integrated team of designers- took
forced concrete, the use of which had required that the idea of collaboration to a new level. It had its philo-
hands-on experience be built up over many years. sophical origins in the Deutsche Werkbund and the
Bauhaus group of designers in Germany in the 1910s.
Toward "Total Design"
In the summer of 1969 the engineer Ove Arup (1895- Throughout the nineteenth century, architects still clung
1988), who was then seventy-four years old , responded to to traditional construction materials despite the fact that
a request by the partners of the firm he had founded in buildings in iron and steel were being erected all around
1946 by writing a paper outlining the aims of the firm them. As building materials , iron and steel presented two
(reprinted in Appendix 1). There has seldom , if ever, been problems that limited architects ' opportunities to ex-
a better summary of what modern building engineering press themselves in the ways they had been used to with
is . Central to the firm 's aims was what Arup called "total masonry. First of all , they were inherently skeletal materi-
design " or, in later papers , "total architecture " ; the term als. In a multistory building they provided a frame on
echoed the title of Walter Gropius's 1956 book which to hang the elements that gave body and solid form
The Scope of Total Architecture. To Arup it meant that "all to a building- the floors , facade and windows , roof,
relevant design decisions have been considered air-conditioning, and elevators . Secondly, iron and steel
together and have been integrated into a whole by a well- in multistory buildings were subject to stringent fireproof-
organized team empowered to fix priorities. " 7 ing requirements. Thus architects had little opportunity,
621

624

625

621 Music pavilion, Kassel , Germany, 1955. Designed by Frei Otto; constructed by tentmaker Peter Strohmeyer.
622 Dome of Discovery at the Festival of Britain, London, 1951 . Engineer: Freeman Fox and Partners ; architect: Ralph
Tubbs. 623 Glue-laminated timber dome at Field House (Brick Breeden Fieldhouse), Montana State College, Bozeman,
Montana, 1956. Structure: Timber Structures, Inc. Diameter: 94.5 meters; rise: 15.2 meters. 624 Dome of Discovery.
Interior, showing the triangulated lattice roof trusses. 625 Field House dome. Key : (1) Radial ribs of laminated timber;
(2) Compression ring; (3) Beams of laminated timber; (4) Steel cross -bracing ; (5) Timber rafters ; (6) and (7) Tension ring.
626

in most buildings, to put iron and steel on show. The ex- dilemma was how to maintain their Beaux Arts heritage
ceptions-where designers could express the nature , or while embracing the new methods of construction. In
"soul ," of iron or steel in their architecture-were mainly Britain, where architects hung on to tradition especially
single-story buildings , which were not subject to the grimly, the situation was reflected in the architectural
codes that regulated fire protection for taller buildings. press. In 1896 a new weekly periodical-The Builders '
They included thousands of manufacturing works and hun- Journal and Architectural Record- was launched, and
dreds of market halls, as well as railway stations, exhibi- its pages were full of articles on traditional architecture
tion buildings, and glass houses in botanical gardens. in masonry and timber. The Arts and Crafts Movement
was in full swing; the domestic buildings of Charles
In the United States, from the 1880s on, commercial Voysey (1857-1941) represented the pinnacle of archi-
pressures to increase net usable floor areas had fa- tectural merit. For four years the only mention of steel or
vored steel-frame buildings. These required increased reinforced concrete in the journal 's pages was in a few
input from engineers in designing the foundations, advertisements .
structure , and building services; the architect 's role
was correspondingly diminished. In Europe, however, Finally, in 1900, The Builders ' Journal and Architectural
such a diminution of the architect's role did not appeal Record published its first article on steel-frame build-
to the architectural establishment, which had developed ings. It was written by a Mr. Homan of Homan and
over many centuries during which masonry was the dom- Rogers, Engineers, whose advertisements had ap-
inant construction medium. European architects hung peared regularly in the journal offering " Homan's fire-
on to tradition, as if blind to the world growing about proof floors " and "constructional steel and ironwork,
them . They faced a crisis of identity. Without the sculp- roofs, piers, bridges, joists and girders and concrete
tural opportunities that masonry offered, and with so fireproof floors (used in nearly 2000 buildings)" and,
much of the design of steel-frame buildings in the hands after 1897, "constructional steel skeleton buildings
of engineers, what role was there for architects? Their (American System)." Two years later there appeared a

626 Pavilion du centenaire de !'aluminium, Paris, 1954. Architect: Jean Prouve.

lengthy report on ferro-c oncrete that discussed the Architectural Concrete
Hennebique system as used in Britain by the Mouchel Reinforced concrete construction had developed in the
company. In 1904 the new monthly Fire Supplement to late nineteenth and early twentieth centuries in response
the journal strongly advocated an engineering approach to two goals: to establish more effective fireproof con-
to building design. In 1904 and 1905 the journal re- struction methods, and to find cheaper alternatives to
ported at length on the construction of the Ritz Hotel, the expensive steel frame. These goals were especially
the first major steel building in London, built by important for industrial buildings , such as warehouses
American contractors with German steel. By this time, and factories, and for commercial office buildings. The
however, Britain was some twenty years behind the catalogues of firms such as Hennebique and Wayss and
United States and at least a decade behind France and Freytag were full of such buildings , which were , funda-
Germany in their use of steel-frame construction. mentally, almost identical to their steel-frame equiva-
Finally, in May of 1906 , the journal changed its name lents: columns were located on a regular grid, between
to The Builders ' Journal and Architectural Engineer, which spanned beams that supported the floors.
and a new monthly supplement, Concrete and Steel Architects ' roles in designing such concrete-frame build-
Supplement, proclaimed: ings were similar to the roles they played in designing
[W]ith the advent of [the steel-frame and reinforced con- steel-frame buildings.
crete] the architect is gradually finding that a far greater
amount of engineering knowledge is required of him than Gradually, after about 1910, a number of pioneering ar-
was previously the case. Some architects overcome this by chitects began to realize that reinforced concrete offered
employing engineering assistants or by having consultative new sculptural opportunities that arose from the very na-
advice on the engineering aspect of the problems put be- ture of the material and its method of manufacture. It
fore them . But, no matter what assistance be obtained, he takes the shape of the mold into which the liquid con-
would be a poor architect who did not carefully study and crete is poured , and so , to a large extent, the structural
thoroughly inform himself of the primary principles of the engineer is able to introduce steel reinforcement to pro-
construction and calculation that are involved. 8 vide the necessary internal strength without necessarily
affecting the outer form. Reinforced concrete thus offers
During the next few years the various titles of the journal architects three entirely new opportunities that structural
continued to reflect the architects ' search for a new steel does not. It can generate solid, three-dimensional
identity. In 1910 it was renamed The Architects ' and forms . It can easily form curved lines and surfaces. It can
Builders ' Journal . In 1919 it became The Architects ' be used to create thin "sheets " of structure with the abil-
Journal , subtitled "For architects, surveyors, builders ity to carry loads; thus it can be used to make load-bear-
and civil engineers," and a year later it became The ing walls , floors that can cantilever out beyond their
Architects ' Journal and Architectural Engineer. Only in supports, and, most spectacularly, concrete-shell roofs.
1926 was the name shortened to The Architects ' Together these formed the basis of a new architecture,
Journal, the name it retains today. one not limited to elements in one dimension, as with a
steel column or beam , but that can work in two dimen-
While engineers' increased contribution to building de- sions, as with a slab , and in three dimensions, as with a
sign was as inevitable in Europe as it had been in 1890s solid sculptural form or a curved shell such as a vault,
America , the outcome was different. Generally speaking, dome, or doubly curved hyperbolic paraboloid. These op-
architects in Europe managed to retain their leading role portunities led to achievements that would match and
in building design , and there developed two common even surpass the achievements in masonry construction
types of building construction. Where steel-frame con- in the temples of ancient Greece , the baths of Imperial
struction did become well established , especially in Rome, the cathedrals of medieval Europe , and the great
Germany and Britain, it comprised a structural frame hid- domes of the Renaissance.
den by a substantial facade of masonry with a traditional
appearance. The alternative was to adopt reinforced con- The architect Le Corbusier (Charles Edouard Jeanneret)
crete as the material of construction. (1887-1965) conceived his " Dom-ino " skeleton in 1914. 629
627 628

1920P

629

627 Unidentified reinforced concrete factory, Paris, 1906. Engineer: Franc;:ois Hennebique. 628 (top and bottom) Grain silos,
from Le Corbusier, Vers une architecture, 1923. 629 Le Corbusier, Maison Dom -lno, 1914.
 Here , for the first time and with characteristic economy, As architects became more familiar with the new mate-
was the architectural expression of the flat slab , albeit rial , they realized they could only use it with confidence
supported by architecturally superfluous columns. through collaboration with engineers and contractors .
Throughout the 1920s architects gradually began to
By the early 1920s Le Corbusier had begun to realize tackle larger projects in concrete; buildings with a new
some of his dreams for the new material , and he cele- and distinct appearance began to emerge , especially in
brated them in his book Vers une Architecture (Towards a continental Europe . These were a far cry from the indus- 631 , 632
New Architecture) , first published in Paris in 1923 and trial , engineer-led buildings of the 1910s. 633,634
available in English in 1927 in a translation , by Frederick
Etchells, of the thirteenth French edition. Uncharacteris- Reinforced concrete offered the unique opportunity for
tically among the writings of architects , it was a paean to the architect and engineer to develop jointly the form of
engineers and builder: he eulogized the "engineer 's aes- the building elements-the mass and the surface, in Le
thetic " and the need to create the "mass-production spirit " Corbusier's terms; indeed , this was almost a require-
of conceiving, constructing, and living in mass-produced ment. An architect had to listen to the engineer 's advice ,
houses. His inspiration was the engineering of cars , ocean at least about what was possible and what not. A recep-
liners , and aircraft. While he extolled the merits of the tive architect could also listen to an engineer's proposals
buildings of ancient Greece and Rome, and the Renais- about what the most rational forms might be both for
sance , he also rejoiced in the modern factory buildings structural elements and the entire structures. Here was
621 that were being built by the hundreds throughout Europe . the same engineer's aesthetic that Le Corbusier had rec-
ognized in the design of aircraft, ships , and cars. In fact ,
Le Corbusier shunned the architect 's traditional preoccu- had he looked closely at the individual components of
pation with style- " what a feather is on a woman 's head these machines, he could have found even more inspira-
. . . sometimes pretty, though not always , and never any- tion in their scientifically crafted shapes and ingenious
thing more. " Architecture , he said , has "graver ends, " methods of assembly.
and set down "three reminders to architects " that their
art is manifested in mass and surface which , in turn, are Among the first designers to articulate the rational ap-
generated from the plan . To illustrate this point , he uses proach to the detail of building structures was the
not medieval churches or Renaissance palaces , but pho- Welshman Owen Williams, one of very few 9uilding de-
tographs of eight massive reinforced concrete grain silos signers who excelled as both a structural engineer and an
from the wheat-growing areas of Canada and the United architect. In a number of outstanding buildings in the
States- "the magnificent FIRST-FRUITS of the new age ," 1920s and 1930s , he consciously used the structural
628 as he described them .9 and sculptural characteristics of concrete to express the
structural function or duty of building elements. This had 635, 636
After about 1918 Le CorbU6ier's cause was bolstered by long been the unconsci~us outcome of engineers striv-
the fact that most countries had ceased to grant patents ing to achieve minimum weights for structures such as
for supposedly new means of reinforcing concrete with bridges , the cast iron beams and columns of early nine-
steel. Reinforced concrete had finally become a generic teenth-century mills , and the many types of roofs of large
construction material like masonry, steel, or timber. The buildings. The idea had been captured by Louis Sullivan
way was now open for architects to work directly with en- in 1896 when he observed that it is a law of the natural
gineers independently of the contractors who had owned world that "form ever follows function "; in about 1918,
the patent rights to the reinforcing systems . Initially, the the Bauhaus group of designers adopted this as their
new material was exploited by architects on a rather mod- guiding principle . When expressing structural function in
est scale-often , literally, at the domestic scale , as the reinforced concrete, however, there was a certain decep-
more adventurous private clients patronized the revolu- tion at work , since the outward appearance of reinforced
tionary thinkers. There can be no more durable image of concrete gives no indication of where the reinforcement
the flat slab than the Schroeder House by Gerrit Rietveld is located, or how much is used. More recently, when
630 (1888-1964), built in Utrecht in 1924-25. structural expression again became fashionable in the
630 631

632 633

630 Schroeder House, Utrecht, 1924-25. Architect: Gerrit Rietveld. 631 Notre Dame, Le Rainey, France, 1922. Architect:
Auguste Perret. 632 Bauhaus building, Dessau, Germany, 1926. Architect: Walter Gropius. 633 Schocken store, Stuttgart,
Germany, 1926. Architect: Erich Mendelsohn. 634 Notre Dame, Le Rainey.
 635 636

1980s, the idea that structural function should be ex- ment process for buildings, enshrined in codes of profes-
pressed in architectural form was nicely captured in the sional practice, mitigated against it. This required archi-
term "archi-structure ," a word coined by Derek Sugden, tects to complete a design and then submit the design to
an engineer with Arups for nearly forty years and a found- a number of contractors. By this means, it was presumed
ing member of Arup Associates . that the client would be able to get the cheapest , or best-
value, building. This process did not seem to allow the
During the 1950s, the Italian engineer Pier Luigi Nervi possibility of an architect desig"ning a building that was
also delighted in expressing the structural function of inherently expensive though inappropriate choice of ma-
many elements of his buildings through their forms, terials , selection of structurally inefficient spans be-
although in many instances the forms are not as tween columns, or poor or nonexistent integration of
637, 638 strictly necessary as might appear-classic examples of building services. The Crystal Palace was an early, rare
639 archi-structure. exception , with the engineer William Barlow involved at
the very start of the design, which demonstrated how ef-
Design for Construction fective a closely-integrated design and construction team
The final aspect of building construction included in could be (see chapter 6, pp. 355-59). The designers and
Arup's total design ideal was consideration, at the de- builders of American high-rise buildings achieved similar
sign stage, of how the actual construction of the bu i Id ing efficiencies when driven by strong commercial pressures
should be facilitated. Such an idea was familiar to indus- from developers.
tries where mass production was well established, such
as machine tool works in the nineteenth century and Le Corbusier was perhaps the first architect to embrace
Henry Ford's automobile factories in the twentieth, but it the manufacturing process in his philosophy of build-
was not common in the building trades. The very procure- ing. He was prophetic not only in his articulation of the

635 Daily Express Building, London, 1 930-32. Engineer: Owen Williams; architects: Ellis and Clarke with Owen Williams.
Section showing columns in the rigid portal frame. 636 Boots factory "Wets" building, Beeston, England, 1930-32.
Engineer and architect: Owen Williams. 637 Palace of Labor, Turin, Italy, 1960-61 . Engineer: Pier Luigi Nervi. 638 Palace
of Labor. Plan of floor slab. 639 Palace of Labor. Drawing for the reinforcement of floor slab.
637

Artnd!Vrct.J svpenve _ , ..
, .,,,.,,_ .. ~ - '•"w/

..
. . .

\y
- c-:---==.: - - - --· -~ -- -

638 639
 building as a machine , like a car or airplane , but in his un- up against the building regulations at Highpoint. In ac-
derstanding of the significance of the manufacturing cordance with the current regulations Lubetkin had pro-
methods that had developed in the automobile industry. posed a standard concrete frame building. Not only did
Indeed , Le Corbusier was inspired as much by the meth- this form of construction lim it the flexibility of internal
ods of mass production as by the idea of reinforced con- layouts , it also often led to cracks in concrete or plaster
crete as a material. One of his earliest proposals for finishes as the heavily loaded columns compressed and
residential buildings was that they should be bu ilt not of moved relative to the unloaded wall panels . Arup sug-
steel-reinforced concrete , but cement reinforced with as- gested replacing the columns by thin load-bearing rein-
bestos fibers , which had been developed recently as a forced-concrete walls , which would place fewer
cheap roofing material and as a way to provide fireproof- constraints on the layout of the interior and would avoid
ing for structural steelwork . He saw the use of asbestos- the problem of cracking , as well as being more econom- 642

reinforced cement as the most suitable means of ical to build . At the time , however, the London County
640 providing mass-produced buildings. Council building regulations , which were often used be-
yond London 's boundaries , did not permit internal walls
While Le Corbusier had the vision to see the potential to be load-bearing. Permission to build was granted only
benefits for architecture in considering construction is- because the building lay a few hundred meters outside
sues at the design stage of buildings , he was not a con - the London County Council boundary, and the local dis-
tractor. Ove Arup was. He worked for the Danish trict surveyor was able to take an independent view.
contracting fi rm Christiani & Nielsen , which was expand- Nevertheless , Arup was forced to compromise by
ing its activities in Brita in, and later as chief engineer adding some additional and wholly unnecessary rein-
with another Danish contractor, Kier. Arup saw how his forcem ent in order to get his idea accepted.
knowledge of construction could help arch itects , for ex-
ample , by adapting for use in buildings many techniques Immediately after World War II , Arup set up the consult-
of concrete construction that had been developed in ing engineering firm that today still bears his name . Free
civil engineering and industrial construction projects , now from the constraints of working for a firm of contrac-
such as docks , warehouses , and grain silos . One exam- tors , Arup was able to begin developing his vision of
ple , which today seems rather obvious , was to speed total design , in which architects and engineers collabo-
up concrete construction by choosing shapes and pro- rate on equal terms in developing building designs from
files that could be cast using simple formwork , erected the concept stage through to the completion of detailed
and dismantled quickly, and reused many times. Arup design. His firm soon attracted the attention of more
realized his dream of bring his engineering expertise to and more architects who wanted to work together to ere- 643, 644
benefit the design of buildings from their very concep- ate buildings that combined engineering prowess with 645, 646
tion to their completion in the work he undertook with architectural flair. 647a- d

the Russian architect Berthold Lubetkin (1901- 90) and

Tecton , the architectural practice he formed in 1932 . Engineering the Internal Environment
Arup and Tecton collaborated first on a number of build- The growing influence of engineers on the design of
ings at the London Zoo , then, in 1935, on the first of two buildings was not limited to the structure of buildings . In
Highpoint apartment buildings in north London , now the United States during the 1920s , for example , air-con-
641 considered an icon of engineer-architect collaboration. ditioning was becoming a standard feature of factories ,
The simple lines and forms of the building were a strik- hotels , theaters , and especially cinemas . Three main
ing contrast to many concrete buildings , which seemed factors contributed to this growth . The first was demand .
to emulate masonry in their use of intricate detail. Like For hotels and places of entertainment , air-conditioning
other innovative engineers before and since , Arup came became an essential feature to attract customers.

640 Maison Manoi, France, 1919. Architect: le Corbusier. 641 Highpoint I, North London, 1935. Engineer: Ove Arup;
architects : Berthold lubetkin and Tecton. 642 Highpoint I. (left) Floor plan of standard concrete frame solution originally
proposed ; (right) Arup's plan for using load-bearing concrete walls.
 ,~~

\1( ~/
'~

(
(

\ ~ r, -
' ~-'-
I .•_

•
/4

ttS if -,
~,-
' ''-..T

~
640

I ! I'

I I
1 ,,
I 1! ; "
:= :-+:~~; t. ~---,
i _- , :1 I

l __ ~ I __ __ 1 -- -· \ ::"
I

643 644

643 Brynmawr Rubber Factory, South Wales, 1946-51 . Engineer: Ove Arup & Partners; architect: Architects Co-Partnership.
644 Coventry Cathedral, Coventry, England, 1958-62. Engineer: Ove Arup & Partners; architect: Sir Basil Spence.
646

645 David S. Ingalls Hockey Rink, Yale University, New Haven, Connecticut, 1956-59. Engineer: Fred Severud ; architect: Eero
Saarinen. 646 Ingalls Hockey Rink. Sections. 647a-d, St. Louis Gateway Arch (Jefferson National Expansion Memorial), St.
Louis, Missouri, 1947-65. Engineer: Fred Severud; architect: Eero Saarinen. Completed arch and construction details.
647a

647c 647d

Indeed , one factor in the growth of the film industry was which were usually still powered by reciprocating steam
the fact that movie theaters offered a pleasant way of es- engines. Manufacturers of refrigeration plants for the
caping the unbearable heat and humidity in many cold storage industries, especially in France , Germany,
American cities. and Switzerland , had continued to develop their prod-
ucts, improving their effectiveness , making them smaller,
Second, the entrepreneurial efforts of a number of men, and increasing their capacity. This was achieved by three
most notably Willis Carrier (see chapter 7, p. 411- 12), means: by devising the mechanical means for accom-
ensured that the performance of air-conditioning equip- plishing new thermodynamic cycles in the refrigeration
ment continued to improve to meet the growing demand. process (the unique combination of compression , ex-
The capacity of early air-conditioning systems had been pansion, heating, and cooling to which the refrigerant is
limited by the slow speed of reciprocating compressors , subjected); by experimenting with different refrigerants ,
of which many hundreds were tried , and by developing culated , partially conditioned , and downward supply to-
various types of rotary compressors to replace the slow, gether in one building . It set new standards of comfort 648,649
limited-capacity reciprocating machines . New refriger- for theaters and cinemas.
ants and compressors had to be developed together,
since the geometry of the blades in the compressor de- Finally, air-conditioning was more readily introduced into
pended on the density of the refrigerant and the speed buildings as the number of engineers and architects who
of the motor. In addition , the various refrigerants , many were familiar with its requirements grew. Clearly, a large
of which were unpleasant, toxic, or corrosive chemicals, air-conditioning system could not be installed as an after-
had to have new types of seal developed to contain the thought. Yet inexperienced contractors might well not
vapor, and suitable lubricants found that did not react have been prepared to , or in fact invited to , work closely
with them. One means of avoiding some of these difficul- with the design team for a new hotel or theater. The situ-
ties was the further refinement of an absorption refriger- ation was similar to what had occurred in 1870s Chicago
ation process which had been devised in France by Carre and New York when it was becoming clear that structural
and Tellier in about 1860 and which avoided the need for engineers had to be involved early in the design process
a compressor. The key feature of the thermodynamic of large iron-framed buildings. One consequence was the
cycle of an absorption chiller is that the energy that emergence of guidance provided by air-conditioning con-
drives the system is heat-a counterintuitive notion that tractors eager to influence both clients and design teams
has puzzled many an owner of a gas-powered domestic and to sell the equipment they supplied. Another was the 650

refrigerator using what became known as the " Electrolux gradual emergence during the 1920s of consulting engi-
cycle " after the firm that first developed it. neers who specialized in the design and specification of
ventilation and air-conditioning systems. These special-
Carrier 's prominent place in the air-conditioning industry ists were engaged by the client and / or design team to
between 1910 and the 1930s was ensured by his dili- provide advice specific to the needs of a particula r build-
gent search for ways to adapt the various new technical ing, and were independent of contractors whose advice
developments in the refrigeration industry, mainly in might not have been impartial since their profit was mad e
Europe , for use in air-conditioning systems. The improve- from installing the equipment they supplied .
ments in refrigeration technology took place at the same
time that the steam engine was being replaced by the Although the spread of air-conditioning appears , in retro-
electric motor as the power source for both compressors spect , to have been a smoothly progressing technical tri-
and fans. This resulted in a huge increase in the capacity umph , the road was not always obstacle-free . From th e
of air-conditioning systems- enough to service the very very beginning there had been dissenting views , espe-
largest cinemas on the hottest and most humid days . A cially on medical grounds , concerning both the benefits
major problem in such large buildings, however, was con- of air-conditioning and the quantity of fresh or condi-
trolling the temperature and humidity sufficiently pre- tioned air that each person needed. During the first two
cisely and rapidly. The engineer Leo Lewis of the Carrier decades of the twentieth century a debate had raged as
Corporation devised an innovative way of achieving the to the nature of " right air " that should be provided for
desired control. His technique was not entirely new, but, children in schools , in particular how much fresh air each
as is often the case in engineering innovation, was a com- child needed for good health . Since the 1890s , a figure
bination of several ideas that had previously been used of 0.85 cubic meters per minute (30 cubic feet per
separately. In his system , three different streams of air minute , cfm) per person had been promoted by ventila-
were used. One stream was passed through the condi- tion engineers , with the backing of doctors. Such a quan-
tioning plant; another comprised some of the air that had tity could only be provided using mechanical ventilation,
been extracted from the auditorium and recirculated; the and equipment suppliers were quick to attack proposals
third was unconditioned fresh air. The resulting pre-mixed to allow a reduced quantity of 0 .28 cubic meters per
air stream was introduced into the auditorium at ceiling minute per person (10 cfm) , which could be achieved by
level . Lewis's innovation in 1922 at Grauman 's opening windows and providing natural ventilation.
Metropolitan Theater in Los Angeles was to use the recir- Further confusion arose when airborne dust and humid-
 ; I I I\

, I I/ I I \

I I ~ .,,.,.,

I - ,

T"

.,, ~ ;i ,~\ '

648
649

Don'ts For Theatre Ventilation

1. Don't use the mushroom system of supply for
cold air.
2. Don't pass all air through the cooler.
3. Don't omit complete mechanical exhaust with
refrig·erating systems.
4 . Don't omit automatic [Link]-e control with
refrigerating systems.
6. Don't supply cold air at low points aud expect
to pull it up with the exhaust.
6. Don't supply warm air at high points and ex-
pect to pull it down with [Link].
7. Don't expect to pull air any place. You can
push, but you cannot pull.
8. Don't conceive of a theatre as a. tight box. It
never is.
9. Don't introduce air into a. theatre auditorium
from the rear unless you know exactly where it is
going and can accurately control its temperature and
velocity.
10. Don't expect a thermostat on the main floor to
maintain conditions of comfort in the balcony, or
vice-versa.
11. Don't supply air to the main door and b1Ll-
cony, or to the main floor and dressing rooms with
the same fan .
12. Don't expect air currents to follow trained
arrows on the plans , unless you are sure the arrows
are thoroughly and properly trained.
13. Don't expect a Rolls-Royce ventilating sys-
tem at the cost of a Ford.

650
651

648 Grauman's Metropolitan Theater, Los Angeles, 1922. Cross section showing the mixing of recirculated, fresh , and
conditioned air. 649 Rivoli Theater, New York, opening day, May 1925. The Rivoli was one of the first air-conditioned the·
aters in the United States. A sign at rear advertises "cooled by refrigeration ." 650 Design rules for theater ventilation,
published in Heating and Ventilation magazine, March 1925. 651 Cover illustration of The Aerologist magazine, December
1931 , representing the dilemmas facing American air-conditioning engineers.
 ~
llJ
§~ H==:-t--f-'I.-~ --'k--\A~~~~*~ ~ >('.1--c~~~~ ~~~
K
~
~ SECTIO N
t
~ Reading 6 %- 0 · 2 %= 5 ·8 %

Ill
~

eo 70 ~o ,o
ORY f'JULf'J TEMPERRTURE °F

653

654

652 The Comfort Zone Chart, published by the American Society of Heating and Ventilation Engineers, 1922.
653 Sky factor protractors, devised by the Building Research Station, England . 654 Daylight factor slide rule, devised by
the Building Research Station, England.
 ity were brought into the debate. Matters were resolved Cinemas , theaters, some hotels , and prestige buildings
in 1922 , temporarily at least, when the American Society for large corporations followed in the 1920s and 1930s.
of Heating & Ventilation Engineers (ASH&VE) published In the early 1930s , a number of firms began breaking into
the results of research it had undertaken to establish a the domestic market with small standalone room air-con-
scientific definition of comfort in terms of wet- and dry- ditioners-the predecessors of the now-ubiquitous win-
bulb air temperatures. The results were plotted as a dow air units-that provided an easy way to retrofit
652 "comfort zone " on the psychrometric chart that Carrier air-conditioning in a building without a central system. For
had published in 1908 (see chapter 7, p. 411 ). This not commercial landlords the window unit system meant that
only provided the means for specifying the cooling level each tenant renting office space could provide air-condi-
that an air-conditioning engineer should strive to tioning, but at its own expense. Air-conditioning became
achieve , it also served to suggest that the engineers ' ar- a standard feature of commercial offices in the United
guments had a firmer scientific basis than the argu- States only during the building boom that followed World
ments put forward by advocates of natural ventilation. War II. A similar pattern occurred in each major city: Once
Research continued throughout the 1920s at the the proportion of office buildings with air-conditioning
ASH&VE laboratories to establish the ranges of tempera- reached about 20 percent , the rest had to follow quickly
ture and humidity that different people judged to be to prevent their tenants from moving to better premises.
"more " or " less " comfortable. New York and Philadelphia reached this point in 1953 , af-
ter which virtually all new office buildings were designed
Although it was recognized that an air-conditioning sys- and built with air-conditioning.
tern could deliver a controlled internal atmosphere , there
was disagreement about which environment to create. From today 's standpoint, when energy is a global con-
Should an engineer seek to reproduce the atmosphere of cern in terms of both its financial and environmental
the seaside beach or the mountain resort? It was widely costs , it seems astonishing that , before the 1960s , so
believed that the action of waves on a beach creates little effort was devoted to reducing the energy used to
ozone , and that this is what makes the sea air so healthy run buildings . The exception was the exploitation of the
and invigorating. By the end of the 1920s the members of heat pump. This device , first conceived by the British
ASH&VE had voted to ban the term "fresh air" from their physicist William Thomson (later Lord Kelvin ) in the
literature , since it was seen as favoring the idea of 1850s , is essentially a reversible refrigeration plant: It
natural ventilation ; the term "outside air " was to be used takes heat from a low-temperature source such as a river
instead . Matters were further confused by some un- and pumps it into the building at a high temperature. In
scrupulous contractors who installed ventilation systems summer it takes heat from a building 's interior and deliv-
that provided chilled air with no humidity control , claiming ers it into the relatively low-temperature water in the
that it was full air-conditioning. To compensate for the river. The benefit to the building owner is that th e energy
lack of humidity control , air speeds were increased to needed to pump this heat is less than half the quantity of
raise the wind chill factor, resulting in complaints from heat energy delivered to or removed from the building-
members of cinema audiences who suffered stiff necks a rare example of a device that effectively operates at
and chilled feet. The front cover of the air-conditioning pe- more than 100 percent effic iency. Possibly the first work-
riodical The Aerologist in December 1931 caught the ing heat-pump installation was by Thomas Graeme
mood of the various arguments that were raging concern- Nelson Haldane (1899-1981), a consulting engineer
ing the choice between air-conditioning and natural venti - who was to become president of the British Institution of
lation and the various ways of achieving them-a debate Electrical Engineers in 1948. In 1928, Haldane used an
651 that continues to the present day. experimental machine to heat his Scottish home and
London office. By the 1930s it was reported that over
Despite the controvers ies, air-conditioning, or "man- fifty domestic heat pumps were in operation in the United
made weather " as many called it, did maintain a foothold , States. In 1931 the Southern California Edison company
and continued to spread throughout the country. During installed an 800-horsepower vapor-compression refriger-
the 1910s and 1920s it was widely installed in factories. ation system in its Los Angeles office that also worked
as a heat pump in the winter. In Switzerland in the Station in England developed custom-made protractors 654
1940s, heat pumps were used for a number of district and nomograms that provided the means to estimate
heating schemes , and Zurich Town Hall was the first daylight factors on the basis of architects ' drawings .
major building to use a heat pump to combine winter They also produced a dedicated slide ru le. 653
heating with summer cooling. A similar system was in-
stalled in London 's Royal Festival Hall in 1951 , taking The effects of direct sunlight on a building-both the
heat from the river Thames in winter and using the river shadows cast by adjacent buildings and the sun 's en-
as the heat sink for summer cooling. Despite their highly ergy entering a window-are as important as the level of
efficient technical performance , heat pumps never be- daylight inside a room . To assess the effect of direct
came popular, partly because they were difficult to main- sunlight falling on a building and entering a room
tain but mostly because of their additiona l capital cost through the windows , the Building Resea rch Station de-
and the low price of fuel. veloped a simple and ingenious device called a " he-
liodon. " This modeled the shadows cast by the sun at
lighting any time of day, and could be set up to investigate a
The quality of the internal air in build ings was the object building at any position on the earth 's surface . It en-
of considerable attention during the interwar period in the abled designers to see the full range of shading condi -
United States , and occasionally in other countries , both tions that a building would experience, and quickly to
because air quality had a direct, perceivable influence on establish the periods of the day when shadows might
people 's comfort and because the growing heating and be a nuisance-calculations in th ree-dimensional geom-
ventilating industry had an enormous financial interest in etry that would otherwise have been far too tedious to
providing temperature-controlled building environments . undertake in the precomputer age.
Developments in two other engineering fields concerned
with the internal environment- lighting and acoustics- The Building Envelope
did not proceed with the same intensity because the The growing awareness of the importance of providing
need was less acute and there was less money to made good daylight inside buildings also had its effect on the
by contractors . construction of buildings. Concrete , being opaque , did
not endear itself to many architects. This limitation was
The design of windows to provide adequate inside day- overcome by the development in Germany in 1908 of a
lighting had been an empirical process until th e end of means of embedding glass blocks into a reinforced con -
the nineteenth century, partl y because there was no con- crete slab or sh ell roof to make what became known as
venient means of measuring lighting intensity and also "glass concrete ." Under various trade names , it became
because there were no agreed-upon data on how much popular for a while during the 1930s. Glass technology
light people needed . This situation changed around the improved markedly in the early twentieth century, result-
turn of the century, when scientists in Germany, Austria , ing in improved quality, larger sizes , and lower prices.
Britain, and the United States began to suspect that The glass-making process of J. H. Lubbers, which was
there was a link between vision problems in schoolchild- developed in about 1905 in the United States , made it
ren and lighting levels . By 1920 many countries had set possible to manufacture single sheets of glass up to 8
minimum lighting levels for different types of room that by 1.6 meters. At about the same time the Belgian engi-
had to be achieved using either daylight or artificial light. neer Emile Fourcault (1862- 1919) successfully devel-
The "daylight factor " had been defined in 1896 by the oped a continuous process for making drawn glass on
British electrical engineer Alexander Trotter (1857- an industrial scale. A band or ribbon of glass was lifted
194 7) as a means of assessing what proportion of the from a vat of molten glass ; once solidified , the band
maximum available daylight reached the interior of a could be cut into individual panes without pausing the
room. Calculating this required complex three-dimen- drawing process. The modern float-glass process , which
sional geometry and the analysis of light reflected off revolutionized glass-making by further increasing the
walls and ceilings , and these were not easily done. To size and quality of glass and reducing its cost , was de-
help designers, scientists at the Building Research veloped by the English firm Pilkington during the 1950s.
655 Boots factory "Wets" bu ilding, Beeston, England, 1930-32. Engineer and architect: Owen Williams. Glass-concrete roof.
656

Cost and qual ity were not the only factors that revolution- developments encouraged engineers and architects to
ized the use of glass . Normal sheet glass from the fac- make building envelopes ever more transparent. One of
tory is relatively weak and brittle. To reduce its the earliest examples of an all-glass facade in Britain
susceptibility to breakage , glass used in large building was Owen Williams 's building for the Boots company in
facades is either laminated or toughened , or both . The Nottingham (1930- 32 ). 655

first laminated "safety glass ," consisting of a film of

transparent Celluloid sandwi ched between two sheets of The means used to support the glass in a transparent
glass , was developed for use in automobile windshields building facade is as important as the glass itself. The
in 1910 by the French chemist Edouard Benedictus, and specialist discipline now known as "facade engineering "
was soon being used in buildings . Multiple-sheet lami- began to develop in the 1930s at a time when many ar-
nated glass screens were first produced in the 1930s, chitects , engineers , and builders , for different reasons ,
and the laminating material polyvinyl butyral (PVB ), which were striving to streamline building manufacturing
is still used today, was developed in 1940. Toughened processes. For some architects , the very process of pre-
glass was developed around 1930 for automobile wind- fabrication and off-site production was an aesthetic in it-
shields by the French firm Saint-Gobain , and it was self, very much in the spirit of Le Corbusier's idea of the
quickly being used in windows and building facades. This building as a machine. The French designer Jean Prouve
glass is prestressed by quickly cooling the outer surface was particularly influential. He had trained as an appren-
of hot glass sheets using air jets. As the entire glass sec- tice in the art of metalworking, and throughout his life
tion cools , the outer surface is drawn into compression as a designer retained a passionate concern for how
by the inner core acting in tension, which helps prevent things were made as well as how they appeared. He
the spread of micro-cracks in the surface of the glass . In began his work in buildings designing and making archi-
the mid-1950s the German firm Glasbau Hahn devel- tectural ironwork for staircases , doors , elevators , lights ,
oped glass cement, which enables glass elements to be and a host of other interior fittings . He later turned his
bonded to form larger structural units such as beams or attention to windows and the building envelope , using
fins for stiffening large sheets of plate glass . All these his knowledge of metalworking to devise systems of
657

660

656 Maison du peuple, Clichy, France, 1937-39. Glazed facade by Jean Prouve. 657 Maison du peuple. Detail of mullions
for the facade. 658 Lever House, New York, 1952. Structural engineer: Weiskopf and Pickworth ; architect: Skidmore,
Owings & Merrill ; designer: Gordon Bunshaft. 659 Lever House. Diagram showing juncture of the horizontal mullion and
vertical rail that guides the window cleaning gondola. 660 Apartments, Square Mozart, Paris, 1953. Aluminium facade by
Jean Prouve.
 mull ions and glazing bars for supporting glass that were erwise the acoustics was addressed largely as an empir-
656, 657 both easy to manufacture and to install in buildings. ical process. Good common sense guidance had been
Since the early 1950s, both steel and aluminum have available in the books of David Boswell Reid , W. S.
been used in curtain walling systems , which have devel- Inman, and others since the 1830s. This had served de-
aped along two lines : The "strong-back system " consists signers moderately well , and was usually adequate as
of rigid panels that are suspended directly from the build- long as concert halls were built in traditional ways. In the
ing frame; in the "stick system, " a secondary structure is 1920s, however, architects began experimenting with
constructed in situ to receive the glazing. One of the first new shapes of auditoriums to which the familiar
large buildings to use the latter type of curtain wall was acoustics design rules did not apply. It became neces-
658, 659 Lever House in New York. sary to think about the acoustics of such spaces from
first principles , considering the different paths by which
660 Like many designers after World War 11, Jean Prouve was sound might reach the listener in the auditorium and the
keen on exploiting certain specific characteristics of alu- correct balance between direct sound and sound re-
minum-its lightness , its resistance to corrosion , and its flected off the walls or ceiling. The French acoustician
ability to be pressed into panels , cast into complex three- Gustave Lyon (1857- 1936) was one of the pioneers of
dimensional shapes , and extruded to form an infinite va- this approach, which he used at the Salle Pleyel in Paris, 662

riety of complex cross sections. It was this ability of completed in 1927 .

aluminum to form extruded sections that made it so suit-
661 able for window mullions and facade systems. It was well known that sound travels in straight lines
and is reflected like light from hard surfaces , and that
Acoustics this phenomenon can lead to a nonuniform distribution
Today we think of double glazing as a means of providing of sound throughout an auditorium. Some designers
effective thermal insulation; indeed , this was its purpose went to great trouble to try to predict sound paths using 663 , 66 4

when the Romans use,d it in a some public baths and pri- what we now call "ray tracing ." Before the advent of
vate palaces in the first century. The first use of sealed computers this meant laboriously drawing lines and
double-glazing, however, was in the 1930s in a number of adjusting the pos ition and curvature of the walls and
nondomestic buildings, including hotels , apartment ceiling to achieve a number of suitable sound paths , in
blocks , and commercial offices , where it was used as a this way avoiding a nonuniform sound field . When
means of providing acoustic insulation against the grow- designing the Salle Henry-le-Boeuf in Brussels in 1928,
ing noi se from traffic on city streets . Indeed , at this time , the architect Vi ctor Horta took a great interest in its
the science of building acoustics consisted largely of acoustics , including the use of ray tracing as a design
seeking ways of solving the problem of noise transmis- tool. Despite his client's wishes , Horta (1861- 1947 )
sion between rooms. Although this was simply a nui- refused to consult Gustave Lyon , then considered to be
sance for people in apartments and hotel rooms , it was the leader of the French school of acoustical design ,
essential to eliminate such noise when designing the because he considered that Lyon 's purely scientific
growing number of sound studios for radio stations and approach paid insufficient attention to how music actu-
recording studios. Especially challenging was the need to ally sounded in concert halls . In making his design deci-
achieve acoustical separation of the control cubicle of a sions on the hall 's volume and -shape and the form of
studio while maintaining visual contact between the mu- the ceiling, Horta visited a great many halls in Europe
sic or program producer and the performers. and the United States ; he also used ray tracing . In the
early 1930s some German acousticians used rays of
When designing recording studios, auditoriums , and con- light in three-dimensional models of theaters to study
cert halls in the 1920s and 1930s , the reverberation the path taken by sound waves, but such models were
time could be predicted using Sabine's method, but oth- of no help in considering different sound frequencies ,

661 Republic National Bank, Dallas, Texas, 1955. Architects: Wallace Harrison and Max Abramovitz. Full-story pressed
aluminum panels form the facade of the thirty-six-story bu ild ing.
 l I
1111 , ,
I
II I I I 11
I
II I I I 11
I
II I I I 11
66 4 66 5

reverberation times, or the dif ferent times a sound stu dies were undertaken in a number of other university
takes to reach a listener by different sound paths . physics department s demonstrating the possibility of us-
ing mode ls during the design of an aud itorium.
Scale models were f irst used to ana lyze the acoustical
performance of auditoriums by the German physicist The science of acoustics received an unexpected boost
Friedrich Spandock in the early 1930s. Us ing simple di- during World War II , when sound was the primary means
mensiona l analysis to find suitab le dimensionless con- for locating the sites of enemy guns and, before the de-
stants, he showed that the acoustical behavior of a room velopment of radar, for detecting the approach of enemy
varies inversely with the model scale; for a one-fifth sca le aircraft. Because sound is transmitted for longer dis-
model, for example, the sound frequencies used for test- tances in water than in air, underwater microphones
ing need to be five times higher than normal frequencies . were used to detect the engines of ships and sub-
He also recognized the need to ensure that the tempera- marines. Most sophisticated of all was the development
ture , pressure , and density of the air in the model room of sonar, an underwater version of radar that detects and
were identica l to the air in a real room, since these all af- locates underwater objects such as mines and vesse ls
fect the speed at which different frequencies of sound by means of sound-wave reflection , and can also be used
travel . Spandock ' s first concern was to demonstrate that to build up a three-dimensional map of the seafloor. After
tests on a model were indeed a re liab le way to represent the war, many of the experts on underwater sound waves
the acoustical behavior of a full-size room. He then stud- became consu ltants advising building designers on
ied the effect on both the decay of sound and its distribu- acoustics. Wartime developments in the science of
tion throughout an auditorium of raked seating (which acoustics did not, however, transform building design
improved t he acou sti cs) and a sem icircula r wa ll behind overn ight. Wh en t he Royal Festiva l Hall was being de-
th e stage (whic h made t he aco ustics worse). Using a mi- signed in 1949-50 , t he primary consu ltant on acoustics
cro phone, he displayed the decay of sound in the model based his advice on over thirty years of ex perience in
auditorium on t he screen of an oscil loscope, and recorded bu ilding design and discussions with musicians, con-
the results photographical ly. In the late 1930s similar ductors , music critics, music lovers , and others with

662 Salle Pleyel , Paris, 1927. Acoustic design by Gustave Lyon. Section . 663 Salle Henry-le-Boeuf. Palais des Beaux -Arts,
Brussels, 1929. Architect and acoustic design : Victor Horta. Manual ray tracing. 664 Salle Henry-le-Boeuf. 665 Royal
Festival Hall, London , 1949-51 . Acoustic consultant: Ph ilip Hope Bagenal ; architect: Lesl ie Martin for the London County
Council Architects.
 practical experience rather than on the latest develop- out to raise a designer's level of confidence sufficiently to
ments in the science of acoustics. Some thirty years allow the proposed design to be built.
after Sabine's work, the only measurable and (relatively)
predictable design parameter was the reverberation In 1959 the Building Research Station in Britain pub-
time . Yet consultants differed in their recommendations lished a new edition of The Principles of Modern Building.
for the most desirable reverberation-time value for the First published in 1939 , this book had become a classic
Festival Hall: One said 2.2 seconds and another said 1. 7 of its kind, seeking to encourage building designers to
seconds , a very large difference. Designers settled on base all their thinking about buildings on scientific princi-
1.7 seconds, but, as it turned out , the reverberation time ples, as was already standard practice in, for example,
of the finished hall was just 1.5 seconds , which most the design of aircraft or long-span bridges. Nevertheless,
people agreed was too short. Acoustics was still far from even in 1959 the design of many aspects of buildings
being a precise science, and assessment was subjec- still involved empirical design rules which lacked an un-
tive ; some musicians and conductors liked the acoustics derlying basis of engineering science or in which the un-
665 of the hall , others did not. In its early days a great many derlying science was not apparent. The book encouraged
studies were undertaken to search for ways of lengthen- the use of basic physics and engineering science to
ing reverberation time , but with little success. Following achieve water-tight buildings , to design thermal and
further studies in the early 1960s when the hall was acoustical insulation, to achieve adequate levels of illu-
being refurbished (it was said that the Royal Festival Hall mination from daylight and artificial light, to avoid con-
was the "most studied " hall in the history of acoustics) , densation problems by controlling the change of
a number of microphones and loudspeakers were in- temperature and humidity across the building envelope ,
stalled to provide " assisted resonance. " This increased and so on. The design issues were the same as they had
the reverberation time to the desired value of 1. 7 sec- been throughout history, and, surprisingly perhaps,
onds , and the acoustics of the refurbished hall met with much of the current design guidance was little different
widespread acclaim. Nevertheless, taste in the world of from what had existed a hundred or more years earlier.
music changes. After many discussions and a number of
trials, it was finally agreed , in 1995 , to stop using as- Of particular interest is what the book did not cover.
sisted resonance , yet the acoustics of the hall are still There was little space devoted to building services , or
widely regarded as excellent. the principles underlying the provision of comfort. There
was no mention of energy. A section headed "The eco-
The period between about 1920 and 1960 saw the build- nomics of heat insulation" began : " In some heated build-
ing industry in every major country transformed. The ex- ings the heat input may be restricted by the capacity of
perimental sciences had had an impact on most aspects the heating plant or by the willingness of the owner or oc-
of building design, and by the 1950s most buildings were cupier to supply heat. Under such conditions , improved
being designed and built much as they are today. When insulation may contribute partly to reducing fuel con-
deemed necessary, engineering calculations could be sumption, and partly to providing a higher internal tem-
made for the structure and foundations, thermal perform- perature; its value must take both into account." lO There
ance, ventilation, acoustics , lighting design, and fire was no mention of how to design the acoustics of a room
safety. Only the balance between empirical and scientific in a building, although there was a section on how to re-
design calculations has changed since then. New con- duce problems of noise transmission by means of sound
struction materials such as aluminum could be integrated insulation. In these branches of building engineering, de-
into building design almost immediately because their en- signers still tended to rely on post-construction remedies
gineering properties and behavior were fully understood rather than design a building with the aid of engineering
down to the atomic level-a far cry from the early days of science to avoid the problems in the first place. Fire pro-
cast iron and reinforced concrete. It was easy to recog- tection fared rather better, and there was a section on
nize when a building design lay outside the normal range designing buildings to reduce the spread of fire. There
of experience that could be handled by standard design was, however, no mention of the term "fire engineering";
procedures. In such cases, model tests could be carried this did not appear until the 1980s .