Proceedings of the First International Congress on Construction History, Madrid, 20th-24th January 2003, ed. S. Huerta, Madrid: I. Juan de Herrera, SEdHC, ETSAM, A. E. Benvenuto, COAM, F. Dragados, 2003.

Inventing a history for structural engineering design

Bill Addis

It is often convenient to as sume that history is a series commercial, contractual, management, social, of facts waiting to be discovered. But there are so political, etc.).' many facts. The facts that we do discover are, to a large extent, auto-selected according to who we are Despite much searching, however, it became clear and why we are looking for them. This influences that very few people had attempted to answer the where we look for facts and what we do with them author's most burning question -how were the when we have found them. The history that is so structures designed? To a large extent, the challenge generated can be very personal, even subjective, and became to create or invent the idea of a history for the great challenge is to persuade others of the value engineering designo Rather than using someone else' s or reward to be gained from adopting a certain map to navigate the territory, the challenge was to approach to history. draw a new map. And an important aspect of this As a practising design engineer, the author's own challenge was to distil the very essence of journey into the past began with questions about how engineering design, and, especially, to distinguish it people in former times might have designed things. fram general ideas of «technology» and «engineering The search uncovered a wonderful range of material science».2 in various different strands of structural engineering Engineers themselves have seldom been effective history: is promoting the nature of their art -they have generally been too busy getting on with projects- - the development of different types of structure and this is nearly as true today as five hundred or a (e. g. suspension bridges, timber roof trusses, thousand years ago. Professor Fritz Leonhardt, from dams, masonry domes); Stuttgart, was one of the most eminent design engineers of the 20th century. In a book about what - developments and inventions in engineering technology (especially materials, plant and engineers do he wrote: machinery, and methods of manufacture or construction); Der Beruf des Bauingenieurs ist in der Offentlichkeit fast nicht bekannt. Wenn vam bauen die Rede is, dan n denkt - biographies of engineers; man die meisten an den Architekten. Wenn ein the stories of individual projects; - Bauingenieur Bauten entwirft, dann wird er in den - developments in engineering science and Zeitungcn als Architekt besprochen. So ging es dem mathematics «

In essence, to design an arte1'act is to plan what you were able to convince themselves that the resulting are intending to do be1'ore you start, and to devise a structure could, indeed, be built, and then that they practical and economic way 01' constructing it. With were able to persuade the clients and even the public characteristic simplicity, Oye Arup once wrote that: that they would be possible. And furthermore, they were so success1'ul in putting their case that they Design is nothing el se than indicating a sensible way of were able to raise the considerable sums of money building. lt includes al! drawings, specifications, necessary to undertake the projects. descriptions and detailed instructions about what should be bui!t and how it should be built (Arup 1985). 11'the engineer's job is to be summed up in one phrase, then, it can be to create the confidence lo start Even de1'initions such as this, however, overlook a huilding. This provides a focus for historical study. most important aspect 01' engineering design -the How, at different times and places, have engineers process 01' achieving the necessary confidence to found the necessary confidence to build? begin building. It is one thing to sketch a magni1'icent 11',at any time in history, the project engineer did bridge or roo1' for a building, quite another to be not feel confident that a certain structure would work, con1'ident that it can be constructed and will sa1'ely he would have to decide what he needed to do in carry the loads to which it may be subjected. This order to gain that confidence. This is the art of the aspect distinguishes 1'undamentally the work 01'those engineer. Seeing and understanding how something engaged in civil and structural engineering from those similar has already been done is probably the most engaged in the design of small machines or consumer important help an engineer can gel. For artefacts, who never need to address questions as to unprecedented or especially challenging problems, whether their creation will work 1'irst time, or be safe, scientists and mathematicians (01'ten called or be possible to make. «philosophers» in ancient times) have al so been an Engineering design is a process with two key imponant source of expertise that can raise the outputs: engineer' s con1'idence. This confidence must overcome any doubts about - a description of what has to be manu1'actured success the engineer may harbour, as well, of course, and built: materials, speci1'ications, dimensions, as any doubts by the client whose money and construction method. This may be a verbal reputation are also at risk. A history of engineering instruction, but, especially when the building is design, then, must address these di1'ferent aspects 01' large or complex, would usually be written the design process: down or in the form oí' a drawing or a model: and - what was the final design for the anefact or - a just(fication or explanation of the design structure? proposals that have been made. This may be - how was the final design arrived at or devised? achieved on the strength of precedent, 1'ull- - what means did the design engineer use to scale tests, experiments on small physical increase his confidence, hefore commenÓng models, various calculations, or the use of construction, that the proposed structure would mathematical models, nowadays based on work and could be constructed, sufficiently 1'or engineering science. him to begin building.

A study 01' what engineers do can make the A history of confidence among engineering process seem all very mechanicaJ, with resulting designers would be a challenging subject to undertake artefact an almost inevitable resulto But this is not since con1'idence depends very much on the how it seems to the design engineer, him or herself. experience and the cultural and intellectual It is important to consider the challenge and the background oí' the engineer as well as the context in unknown aspects of a project that the design team which the engineer is working. There are also faces. From an engineer' s point of view, the principal important questions about how di1'ferent engineers in achievement in constructing the Pantheon, or a large the past might have perceived the gap between what cathedral or the Ei1'fel Tower, was that some people they knew had already been done and what they ]nventing a history for structural engineering desing 1]5 might have believed to be possible, though not yet and stmcture, as well as what we now eall the internal know by what means. And the confidence is needed environment (Jighting, heating, ventiJation and at even the most fundamental leve]s such as the acoustics) in a single, holistic design process. The reliability and sophistication of mathematieal visual appearance, sense of space and function (the calculations and the accuracy of dimensions «architecture» in the modern, narrow sense) became a generated by geometric techniques of producing distinct concern during the fifteenth and sixteenth drawings of complex shapes or eonnections in a centuries. About a century later, designers first began buiJding. A history of engineering design needs to to think about the load-bearing aspects ofbuildings in address these aspects of the skilJ. This was a terms of loads (weight), materials and structure. challenge that faced the author when working in the Thinking separately about materials and stmctures team devising an exhibition about the work of grew during the late-seventeenth and eighteenth engineers some years ag04 -had do you put on centuries, folIowing Galileo's work, and aspects of display? On that occasion the answer to this question the internal environment carne to be considered included a collection of the engineer' s aids to independently of the fabric of the building during the calculation -tables of Jogarithms, slide mIes and Jate-nineteenth and early twentieth centuries. mechanical and electrical calculators. There were also It is al so worth noting that structural understanding some design engineers' notebooks with their ]ittJe is neither a new phenomenon, nor one that requires a sketches and the calculations they made when know]edge of statics and elasticity. It is a commonly- exploring a new idea. These things really do capture he]d misconception that new types of structure and something of what engineers do, but their apparent stmcturaJ forms were devised first by mathematicians insignificance may stilI give only a hint at how a final or scientists and later taken up by builders, engineers grand scheme may have come about. or other designers of stmctures. In fact, the opposite There are, of course, prablems in using modern is the case, with perhaps just one exception -the words like «engineer», «desigll» and «designef» to hyperbolic paraboloid, whose stmctural properties refer to the construction of a Gothic cathedral or, were discovered in the 1930s. Many children and indeed, any edifice completed before the twentieth sculptors display a remarkable understanding of century when these words acquired their modern materials and stmctures without a knowledge of meanings. The author has not found it ditIicult to use engineering statics, and many stmctures fram long such words to describe the past, without implying the ago show their creators understood the essenee of alI professiona] demarcations, knowledge and working the basic stmctural actions better, perhaps, than many methods of the twentieth century. At the risk of being engineers today. accused of tautology, it is possibJe to use them to A history of stmctural design would need to embrace the work that must have be en undertaken to address how all the stmctural elements of buildings complete a cathedral, without specifying precisely were designed. A few examples do exist -for what it was or who did it (since we do not know). It example, beams (Yeomans 1987; Skempton 1956, is possible to discuss the proeess of designing a Sutherland 19), foundations (Peck 1948), retaining cathedral, for instance, without becoming entangled walIs (Kerisel 1993), and stability design (Mainstone in questions as to whether it was designed by an 1988).5 architect or an engineer, or whether its stmcture was Two examples from the history of engineering designed separately fram the architecture. These design will ilIustrate some of the points made above questions are, Jiterally, meaningless because to use -the design of Gothic cathedrals and the design of the very words «engineef» or «structure» is beams for use in buildings. anachronistic. What happened, over a period of many centuries, was that the activity of designing a large building, THE DESIGN OF GOTHIC CATHEDRALS whoever did it, was gradually broken down into more and more distinct issues. Thus the methods of In some ways, designing a cathedral is not as difficult designing a Greek temple or a cathedral embraced as it may sound. Most importantly, we already know visual appearance, sense of space, function, materials that it can be done -and this was true in 1200 too, for 116 B. Addis many large cathedrals had been built before the maximum area 01' window in a wall increased 1'rom period we know as Gothic, especially in France. perhaps 30% to around 80%, the horizontal thrusts of Nowadays we also know that the masonry is loaded the vaults (and wind loads on the roof) were carried only in compression and that the stresses in the out over covered aisles by means of highly efficient materials are very small compared to the strength of t1ying buttresses. The world had seen no similar the stone or brick (Iess than about 10%). The strength transition in building engineering since the 2ndcentury of the masonry is, therefore, virtually irrelevant to the AD and would have to wait 1'or700 years 1'or a similar success of the structure as a whole. period 01' dramatic structural development in early The structure must, however, be stable. Arches and Victorian England. During the 400 or so years of the vaults must not thrust sideways so much that they Gothic period there was some steady development and push over the walls or columns; they must also be improvement in various elements 01' the cathedral, curved such that lines of thrust remain within the especiaJly the slenderness of columns and fiying thickness of the masonry. Also, the loaded elements buttresses. However, these changes were relatively must not be so slender that they fail by buckling. The smaJl compared to the very sudden developments 01' stability of a masonry structure depends, then, on the the mid-12th century. relative size and disposition of the individual pieces. The beginning 01' what we know as the Gothic lt does not, generally, depend on the absolute size of period can be pinpointed quite precisely as about the building. Hence it is easily possible to build a 1134: in or around that year new building commenced model masonry structure and scale it up to whatever on the cathedrals of Sto Denis and Chartres. Within a size you want -the equilibrium conditions 1'or the few years several others were begun, all in the region model and the full-size structure are the same. around Paris called the Ile de France and by the year Nowadays we can demonstrate this using statics; 1300, some 60 Gothic cathedrals in France and 40 in engineers also knew this in 1200, based on their own England, were complete or under construction. The direct experience and the evidence 01' thousands 01' style continued to spread throughout Germany, Italy, success1'ul masonry buiJdings. Spain and central Europe until the late 16thcentury. The exception to these generalities about scaling By and large, we know how the cathedrals were up 1'rom models is the question of wind loads. A built and we also know quite a lot about how they masonry structure must not be overturned by winds were designed. We know the names 01' many from any direction. We now know that the force hundreds 01' the individuals who were involved with exerted by the wind on an object does not vary in their design and construction. Un1'ortunately we have simple proportion to the wind speed. We also know no explicit design procedures dating from the early that the speed 01' real winds increases as you rise Gothic periodo This is hardly surprising considering 1'urther above the ground. It is quite Jikely that these the secrecy surrounding the skiJls 01'trades such as the non-linear effects were the cause of some coJlapses of mason -it was forbidden to divulge any information early large masonry structures, especiaJly during outside the masons' lodge, either to other masons or construction. The main causes of damage to non-masons. We also have direct evidence of what cathedrals were, and still are, in fact, the settlement or was designed -the buildings themselves. Many movement 01'the 1'oundations and the burning 01'roo1' documents relating directly to the designs 01' trusses after lightning strikes. cathedrals have survived including several Something dramatic happened to cathedral and sketchbooks 01' details and plans of cathedrals, the church design in the J2th century. After several most famous of which is by Villard de Honnecourt. hundred years of gradual development since the end Many more-substantial working drawings survive 01' the Roman empire, the designs 1'or large churches and we know that some 01' these 1'ormed part 01' the took a sudden and sharp change of direction. Round building contracts. arches were replaced by pointed arches, long barrel From such evidence it is possible to work out how vaults were replaced by several discrete structural some of the plans and shapes have been constructed. bays formed by quadripartite vaults (intersecting They seem to be based mainly on two geometric pointed vaults), the groin vault was repJaced by the techniques or manipulations - the combining of ribbed vault, columns were reduced in thickness, the various circular ares and the «rotating 01' the square» Inventing a history for structural engineering desing 117

-a procedure by which a square with an area of half number and geometry. With numbers there were of another can be created by joining the centres of ratios, squares, multiples, series and so on; in each side. These were used to generate an enormous geometry there were circles, triangles, squares, range of plans and elevations for cathedrals and their spheres, cubes and so on, and the properties components, and several notebooks from the late associated with these shapes Gothic period include details of such geometric This use of geometry. harmonics and number was design procedures. We do have some examples of also incorporated into Christian philosophy. Saint such methods, although from rather later, contained Augustine (354-430) and Boethius (480-525) (both in some design from the 14th and 15th centuries were philosophers and hermeneuts) provided the (Shelby 1979; Sanabria 1982; Addis 1990). means. Their works on the science of music, At first sight, such geometric procedures would mathematics and architecture sought to demonstrate, seem to be of little relevance to what we would now using harmony and geometry, the underlying caJl the structural design of the building. Certainly principies of the world as created by God. Further they were not based on statics. But it aJl depends on manifestations of these principIes were, of course, to how we look at the matter of designo If we are looking be found in the Bible. Augustine took the Biblical at the history of engineering design at a time when it passage Omnia in mensura et numero et pondere is meaningless to talk of statics, any design disposuisti (Thou hast ordered aJl things in measure, procedures must be of interest, especiaJly in the light number and weight) and applied Pythagorean and of the discussion about the nature of design earlier in neo-Platonic methods to the interpretation of the this paper. It is also worth lookjng more closely at Christian universe, its creation and its order. what was meant by «geometry» in the 12th and 13th Conceming the design of buildings, the Bible also centuries (von Simson 1952, 1956). provided dimensional details of a number of Most importan ti y we know that, somehow, significant structures -the Ark, Moses' Tabernacle, building designers were inspired with a new Solomon's Temple and the Celestial Temple revealed confidence to push back the boundaries of what was to Ezekiel in a vision. A weJl-known 14th century possible and to be much more economical in their use masonic poem even claims that Solomon actuaJly of materials. How did people come to believe that it «taught» architecture in a manner «but little different would be possible to build cathedrals taJler, wider, from that used today» and that this science was longer and more daring than ever before? What directly transmitted to France. The writings of happened, or might have happened to give rise to this Augustine and Boethius dominated the middle ages, confidence? Par! of the answer seems to be bound up and the cosmic applicability of the laws of harmony with geometry and, in particular, with Euclid. figures boldly in writings about both music and building throughout the Gothic periodo It was into this phiJosophical tradition, in northem The role of geometry in the mediaeval world mediaeval France, that the books on geometry by the Greek mathematician Euclid were suddenly Since Classica] Greek times (and probably earlier), introduced. AJI copies in Greek or Latin had been 10st people had sought explanations for natural or destroyed and the books had survived only in phenomena in order to understand the world they Arabic translation. While the English scholar Adelard lived in -rainbows, the musical notes made by of Bath was a student of Thierry at Chartres. he vibrating strings, the orbits of the stars and planets, visited Domenicus Gundissalinus. the Archdeacon of the trajectory of missiles. and so on. In every field of Segovia, which had recently been recaptured from the what we now caJl natural science- optics, acoustics. Moors. There he carne across a copy of Euclid's astronomy, music, mechanics, botany -philosophers Elements 01 Geometry in Arabic which he translated had explained phenomena in terms of geometry, into Latin and brought back to his community of harmonics and number which were seen as earthly feJlow scholars at Chartres in the mid 1120s (some manifestations ofthe principies the god(s) had used to saya few years later). The appearance of Euclid create the universe. These explanations were seems to have had a profound effect on their work - generally expressed in terms of the absolute truths of it has been said that, under this influence from Euclid 118 B. Addis

«the school of Chartres attempted to transform to use in every conceivable way and created, quite theology into geometry». literally, a new type of geometry -geometria Geometry had, of course, survived as a practical art theorica. throughout the Middle Ages and the appearance of The distinction between geometria theorica et Euclid improved the level of geometrical knowledge practica was probably first made by Hugh of Saint which could be learnt. Improved geometry facilitated Victor sometime between 1125 and 1141. In doing so the more accurate description of proposed building he was looking back to the philosophies of Plato and designs and was of great practical use in the Aristotle in distinguishing the practical skilJs from the construction process -setting out building s and theoretical (contemplative) skills which had been enabling the accuracy of the finished parts and their made possible by the appearance of Euclid. Hugh put relative disposition to be checked to greater accuracy. the theoretical tool to good use in helping to explain Such an improvement alone would have enabled and justify information given in the Scriptures. For builders to contemplate larger and taller buildings. instance, he calculated that the reported size of the However, it was in the capacity to provide Ark (40.000 inches) would indeed have been large justification of designs that geometry probably had a enough to accommodate all the animals of the world more profound effect. Euclid introduced a crucial and their food (see Victor 1979,3 and 32). new ingred:~nt -the notion of the geometrical proof. There soon followed a number of geometry text This provided the perfect tool for the philosophers at books, some purely practical, and others, such as one Chartres to argue their views logically and to j ustify written around 1140 by Gundissalinus which also decisions made in a wide variety of contexts dealt with theoretical matters. In this work he including, probably, the design of buildings. Just as distinguished the two geometries according to their occurred after the invention of calculus some 600 respective purposes and duties (Table 1). years later, philosophers put the new theoretical tool

Geometria theorica Geometria practica

Finis (purpose) to teach something to do something Officium (duty) to give reasons and dispel doubt to give measurements or limits which the work should not surpass

Table I The several functions of mediaeval geometry, according to Gundissalinus cited in (Victor 1979: 9)

Other 12th century geometry treatises appeared. princip1e of falsification whereby a hypothesis shou1d The prologue to one of these: be rejected if it 1eads to conclusions found to be at variance with experience. Grosseteste typified the almost makes the practical side of geometry seem contemporary attitude to geometry in declaring that: subservient or secondary to the theoretical . . . The use of theoretical methods in practical geometry seems to have without geometry it is impossihle to ul1derstal1d l1ature, increased between the twelfth and the fourteenth sil1ce all forms of l1atural hodies are il1 essel1ce centuries. At first their roJe was ancillary to the purposes geometrical al1d can be reduced to Unes, al1gles al1d of practical geometry. Once proofs had found a place in regular figures. (Victor 1979) practical geometry, their role increased and changed. Theoretical proof became the goal even of practical In brief, we find the ro1e of geometry in the geometry. (Victor 1979) mediaeva1 period was rather like the ro1e that modern science -physics, mechanics, materia1s science, We also find, in the work ofthe philosopher Robert etc.- p1ays in the world today. It provided Grosseteste (c. 1175-1253), just a few decades later, exp1anations of why the world and the heavens were perhaps the first mention of what is now called the as they were, and how they worked." hypothetico-deductive method in science, and the Inventing a history for structural engineering desing 119

The gothic design revolution Beams have been used in buildings 1'orthousands 01' years for floors and roo1's. In classical Greece the choice To a mediaeval cathedral designer, then, a was generalJy between timber and stone. Both 01'these geometricaJ model of the building could serve not materials occur in nature in larger sizes than they are only as a mathematical model to describe the needed in buildings so extra work is needed to reduce proposed construction, but also, being based on the size down to what would be the minimum possib]e. geometry, it would serve in some manner to justify jts lt was there1'ore normal to make a compromise and use adequacy, in conjunction with other engineering a 1'orm 1'or beams that was easy to manufacture and knowledge, of course. While we do not know exactly large enough to carry the loads. Most beams were, how they used geometry to «give reasons and dispel there1'ore, 01' rectangular cross section and constant doubts», the expJosion of interest in geometry during shape along their length. It was welJ-known that the the 1120s and I l30s did occur at the very beginning strength of a beam increased in direct proportion to its of what we now know as the Gothic era. The breadth and more than in direct proportion with its cathedrals of St Oenis, Sens and Chartres were al! depth (with the square of the depth, in 1'act). designed in the 1130s; another dozen folJowed during To conceive most buildings, there1'ore, the choice the next 25 years. of size was very limited and easy to learn and pass on Any plausible account of building design in the to young bui]ders. It would depend on whether: Gothic era must take account of the significance of geometry in the mind of mediaeval mano Even the - the beam was in the floor or roo1' (i. e. on the scant evidence presented here would indicate that load) geometry would indeed have helped provide the earJy - the beam was of wood or stone cathedral designers with the confidence they needed to - the type of wood (or stone) propose bolder and bolder designs for the cathedra]s. - the span between the supports The changes that occurred in the design of cathedrals during the J2th and 13th century were not The types of material and the dimensions needed technologica] developments in the conventional sense were wel! known and would have hardly needed to be -no new material s or structural devices were written down. Nevertheless, as our modern scientific invented. The changes were largely human changes- approach to the worJd developed during the 18th they went on in the minds 01' the designers and century manuals did start to appear, especially for builders. For this reason it is appropriate to refer to timber. These gave tables and simple 1'ormu]ae for this sort of change as a revolution -a design establishing the scantlings (dimensions) 01'the timbet revolution.7 components 01' floor and roof structures (Yeomans These changes were largely complete by the 1987). There was ]ittle or no attempt to justi1'y these middle of the l3th century and comprised the first 01' dimensions, 1'or they were so well-established. two design revolutions during the Gothic era. The In fact, there are some very early examples of stone other began a bout a century later when the first signs beams 1'rom the 5th-3cd century BC in which the began to appear of taking jnto account the weights of cross section of stone has be en reduced to ]ess than a materials and the loads these imposed on parts of rectangle, in order to improve ¡he structuraI efficiency buildings (Sanabria 1982; Addis 1990). (Coulton 1977). For the very Iargest spans- the one in Figure 3, 1'rom 400 BC, was 6,2 metres in Iength - the depth 01'the beam could be reduced away from the THE DESIGN OF BEAMS centre of the beam without reducing the strength too mucho In modern ]anguage, more stone is used where The second story in the history of structural the bending moment is larger and to improve the engineering took place over many centuries -the ration of cross-sectional area to second moment 01' development of design methods for beams used in area. In this case, however, the increased efficiency of structural trames. One key moment in this story was the beam was at considerable expense in terms of the development of the I-beam and its introduction labour and would only be worth considering for into buildings in the 1830s. unusually large spans. 120 B. Addis

Section BB 28Qmm fact, the search did not take very long as much of the A theoretical and practical work had already been

coffered ceiling slab completed by many scientists and mathematicians 910mm earlier in the 18thcentury. 410mm ¡ The early application of simple bending theory did not, however, fully ret1ect the asymmetric properties 560mm of cast iron which is much weaker in tension than in compression. By the 1820s another moment of crisis in design procedures had arisen. This led William Fairbairn, with Eaton Hodgkinson, an engineering scientist at Manchester University, to search for new design procedures that would generate the optimal 6150mm shape for a cast iron beam for use in the many high- Figure 1 rise factory buildings that Fairbairn's firm was Stone beam building. In this way he hoped to out do the competitors! And he did. The result of the work was the I-beam (Hodgkinson 1831). However, this was As builder' s mathematical skills improved towards not the symmetrical I-beam we now know, since cast the end of the 18th century, so me books contained iron is six times stronger in compression than in simple formulae which effectively summarised the tension. The lower (tension) flange of the beam must data previously presented in tables. These were therefore ha ve an area six times that of the upper empirical formulae and were not engineering science (compression) t1ange. The resulting section was first as we know it -they dealt with only one material and used in 1834 in Orrells Mill in Manchester and the rectangular cross-sections, constant along the length earliest surviving example is at the famous Saltaire of the beam. The enabled someone to calculate the Mill by William Fairbaim (Fitzgerald 1988).9 relative strength of a beam-relative, that is to a This paper has reached only the middle of the 19th similar beam but of smaller or larger dimensions. The century in this extract from the story of designing formulae contained empirical constants with no beams, but it continues through the next 150 years scientific significance at all. The relative strength was right up to the presento10 Now the mathematical distinguished from the absolute strength which was models of beams used by engineers to justify their the concern of scientists seeking to explain how design decisions are much more sophisticated and can strong a beam would be. even represent the behaviour of beams under .tire This situation changed with the introduction of cast loads as well as gravity loads, but that is another iron in buildings, as a fire-proof material for columns story. and floor beams at the end of the 18th century (Skempton 1956; Sutherland 1984).8 Unlike with timber and stone, when you make a beam from iron, CONCLUSION it was important to use as little material as possible- the more material you use the more it will cost and This paper has sought to demonstrate that the you had to keep the structure as light as possible, historical development of engineering design is a since iron is three times more dense than stone and subject both different and separate from other themes seven times denser than timber. There was, therefore, in the historical study in construction. for the first time in history, benefit in using the At present the written history of engineering design minimum amount of material possible -the search is uneven1y covered, even though the source material for «minimum-weight» structures had begun and the may existo As yet, for instance, no-one has traced the established deign procedures were unable to help history of how timber and, later, wrought-iron roof designers in their task- a period of crisis had arrived. trusses were designed (as opposed to constructed). This started the urgent search for the most efficient Rather more serious is the absence of due attention cross-section of beam and shape, from end to end. In from historians to probably the two most important Inventing a history for structura] engineering desing 121

] subjects in the history of engineering design-the Addis, W. 997. Free will and determinism in the development and use by designers of graphicaJ statics conception of structures. Joumal of the Intemational and the factor of safety. These are themes in need of Association for Shell and Spatial Structures, 38, n° 2: research. 83-89. Addis, W. (Ed.) 1999. Structural and Civil Engineering Unfortunately practising engineers often have little Design. Studies in the History of Civil Engineering, 12. time or inclination to foJlow these directions, so it A]dershot: Ashgate (Variorum). must be hoped that some professional researchers Addis, W. 2001. Creativity and Innovation: the structural have been stimulated by this paper to take up the engineer's contribution to designo Oxford: Architectura] chaJlenge. Press. Arup, OYe. 1985. The prob]em of producing quality in building. The Arup Joumal (Ove Arup's 90'h Birthday ] Notes Issue), 20, n° (Spring): 23-25. Bow]ey, Marion. 1966. The British Building Industry: Four ]. See, for instance, Bowley 1966 and Nisbet ]997. Studies in Response and Resistance to Change, ] 2. For further discussion of these issues, see Addis 990, Cambridge: University Press. ] 1994, ]997 and 2001. Coulton, J. J. 977. Ancient Greek Architects at Work: 3. The profession of building engineer is hard]y known by problems of structure and design. Ithaca, New York: Come]] University Press. the public. When building is mentioned, peop]e usually think of the architect; when a building engineer designs Fitzgerald, R. 1988. The Development of the Cast Iron something, the press refer to him as an architect. This Frame in Texti]e Mills to 1850. Industrial Archaeology has been my experience during my who]e life. 1 am of Review, 10, n° 2 (Spring): 127-145. ] the opinion that the building engineer himself is ]argely Hodgkinson, E. 831. Theoretical and Experimenta] to b]ame for this state of affairs. Researches to ascertain the Strength and Best Form of 4. L' art de l' ingenieur-constructeur, entrepreneur, Iron Beams. Memoirs of the Literary and Philosophical inventeur at the Centre Pompidou, Paris, 1998. (Picon Society of Manchester, 2",1series, 5: 407-544. ]997). Kerise], J. 1993. History of retaining wall designo Retaining 5. These and other examples are reprinted in Addis 1999. Structures (Proceedings of conference: Ed. C. R. 1. 6. These issues are discussed in depth in Addis 1990. This Clayton), Thomas Te]ford: ]-16. book is now out of print, but a few copies are availab]e Leonhardt, Fritz, 1981. Der Bauingenieur und seine from the author. Aufgaben. Stuttgart: Deutsche VerJags-Anstalt. 7. This phrase is drawn, by analogy, from the idea of the Mainstone, R. M. 1988. Stability concepts from Renaissance scientific revolution that the American philosopher to today. In Stable-unstable. Centre for Conservation of Thomas Kuhn deve]oped some thirty years ago and Historie Bui]dings, Leuven University Press, Be]gium: which has transformed our understanding of the 65-78. development and history of science (Kuhn 1970, Addis Nisbet, James. 1997. A Proper Price: Quantity Surveying in 1990). London 1650-/940. London: Stoke PubJications. ] 8. These two papers are included in Addis ]999. Peck, R. B. 948. History of buildings foundations in 9. This paper is included in Sutherland ]997, along with Chicago. University of Illinois, Engineering Experiment Sutherland ]990. Station, Bulletin Series 373. 10. A further episode in this story, the design of steel beams Picon, A. (Ed.). 1997. L'art de l'ingénieur-constructeur, using plastic theory, is given in the author's other paper entrepreneur. inventeur. Paris: Éditions du Centre in this conference: The nature of Progress in structural Pompidou. ] engineering. Sanabria, S. L. 982. The mechanisation of design in the 16'h century: the structural formulae of Rodrigo Gil de Hontañón. Joumal of the Society (Jl Architectural Historians, 41: 281-293. Skempton, A. W. 1956. The origin of iron beams. Actes du REFERENCE LlST VIII' congres intemational d'histoire des sciences. Fiorence: 1029-1039. Addis, W. ]990. Structural Engineering, the Nature ol Shelby, L. R. and Mark, R. (1979). Late gothic structural Theory and Design. Chichester: ElIis Horwood. design in the Instructions of Lorenz Lechler. Architectura Addis, W. 1994. The Art of the Structural Engineer, (Journal of the History of Architecture, Munich), 9: London: Artemis. 113-131. 122 B. Addis

Strike, James. 1991. Construction into design: the influence Victor, S. K. (1979). Practical geometry in the High Middle of new methods of construction on architectural designo Ages, Memoirs of the American Phi/osophical Society, 1690-1990. Oxford: Architectural Press. 134. Sutherland. R. J. M. 1984. The birth 01' stress: an historie al von Simson, O. G. 1952 .The gothic cathedral: design and review, contained in The art and practice of structural meaning, JOllrnal 01' the Society 01' Architectural design, London: IstructE: 5-15. Historians, 11, n° 3: 6-16. Sutherland, R. J. M. 1990. The age of cast iron: w/w sized von Simson, O. G. (1956) The gothic cathedral. Pantheon, the beams? In Thorne, R. 1990. The iron revolllÚon: New York. architects, engineers and structllral innovation Yeomans. D. T. 1987. Designing the beam: from rules 01' 1780-1880. London: RIBA: 25-34. thumb to calculations, JOllrnal of the lnstitllte 01' Wood Sutherland, R. J. M. (Ed.) 1997. Structural iron, 1750-1850. Science, 11. nOl, Issue 61 (June): 43--49. Vol. 9 of series Studies in the History 01' Civil Engineering. Aldershot: Ashgate (Variorum).