Form Follows Forces
Architectoid Research Series · Structure & Space
The Engineers Who Broke the Box: Nervi, Candela, and the Structure Behind Free Architecture
Pier Luigi Nervi and Félix Candela solved the same problem from opposite directions: how to make thin concrete span vast space at a price anyone could build.
 |
| Detail of the Section of the Olympic Stadium for the Kuwait Sports Centre submitted by Studio Nervi, 1969 |
Every builder who has tried to enclose a genuinely free interior, one continuous space that curves where it wants to curve and opens where it wants to open, runs into the same enemy sooner or later. It is not the client, and it is not the budget, though both do their share of damage. It is the post-and-beam frame. The orthogonal grid of columns and beams that has carried Western building for centuries is magnificently efficient at one thing: producing boxes. For most of history, escaping the box meant heroic expense. The dome of the Pantheon, the Gothic nave, the great train sheds: each was a triumph, and each was a one-off that only an empire, a church, or a railroad could afford.
In the middle decades of the twentieth century, two structural engineers working an ocean apart changed the terms of that bargain. They proved, in built concrete, that a roof did not have to be a lid on a box. It could be a shell: thin, curved, continuous, spanning enormous distances without a forest of columns, and shaped by the flow of forces rather than the geometry of the lumberyard. More importantly, they proved it could be cheap. Pier Luigi Nervi in Italy and Félix Candela in Mexico arrived at that proof by nearly opposite methods, and the difference between their methods is one of the most instructive stories in twentieth-century construction. This post walks through how each of them actually built, project by project, using the construction photographs as the primary text. If you are a student of architecture, this is a story worth learning slowly.
01 · The Idea Underneath
Space Is the Subject. Structure Is the Price of Admission.
Start with a claim that still sounds radical if you take it seriously: the real subject of architecture is not the façade, not the mass, not the plan on paper. It is the interior space itself, the void a building encloses and the experience of moving through it. Everything else is packaging. By that measure, the history of architecture is really the history of humanity's growing technical ability to liberate space, and every great leap in spatial freedom was paid for, in advance, by a leap in structural technology. The Roman arch bought the basilica. The flying buttress bought the Gothic nave. The steel frame bought the open floor.
What no one had before midcentury was a way to make continuous curved enclosure economical: a buildable, repeatable, affordable structural vocabulary for free space, beyond the heroic one-off. That is the specific thing Nervi and Candela supplied. They did not invent the idea of spatial freedom. They invented its economics. And because both men ran their own construction companies, the invention happened where it had to happen, on the job site, one formwork decision at a time.
The Palazzetto dello Sport, Rome (built 1956–57 by Nervi & Bartoli), with architect Annibale Vitellozzi; dome engineered by Nervi. 1,620 prefabricated ferrocemento pieces, erected in forty days, forming a structure that is simultaneously the frame, the ceiling, and the ornament — there is nothing to add and nothing to hide.
02 · Rome
Nervi: Form as the Shortest Path to the Ground
Pier Luigi Nervi (1891–1979) was an engineer's engineer, trained at Bologna, and he built his reputation on a conviction he defended for fifty years: structural form should never be imposed on a building from outside. It should be discovered, by following the actual path that loads take on their way to the ground and giving that path physical shape. Where the path of forces curved, the concrete curved. Where forces gathered, material gathered. Where they thinned out, so did the structure. Beauty, in Nervi's account, was not a goal you pursued directly; it was what an honestly solved structure looked like when you were finished.
The problem Nervi faced was scarcity. In 1939 the autarchy regime restricted steel-bar reinforced concrete, and the shortage pushed him into a systematic cost analysis of everything in the construction process. It yielded two conclusions that governed the rest of his career: minimize the concrete, and eliminate the wooden formwork, which he identified as the single largest hidden cost in concrete construction. His trick, patented in 1943, was ferrocemento: layers of fine woven steel mesh, the wires a millimeter or less in diameter on a grid of about a centimeter, packed together and saturated by hand with a rich cement-sand mortar. Because the stacked mesh holds its own shape, the mortar is plastered directly onto the steel. No formwork at all. The extreme subdivision of the reinforcement transforms the material's character: instead of discrete bars buried in brittle mass, something close to homogeneous, ductile, almost sheet-metal-like, workable down to three centimeters thick. Nervi proved it the way a builder proves things. He put up a wave-roofed experimental warehouse at his own yard in 1945, and, the same year, launched a 165-ton motor yacht with a ferrocemento hull thirty-five millimeters thick, lighter than wood and forty percent cheaper.
His second invention was a way of seeing. In a loaded slab, bending forces flow along curved paths, trajectories that swirl and concentrate near the columns and run in long arcs across the open field, the way grain flows around a knot in wood, or the way bone lays down its fibers along the stresses in a femur. Engineers call these paths isostatic lines, and in 1949 Nervi's firm patented a system, conceived in his office by the engineer Aldo Arcangeli, for running the ribs of a floor slab along them. A rib aligned with the force path works purely, with nothing wasted; the thin slab between only spans locally. What to look for in the photographs below: the ribs of the Gatti Wool Factory ceiling are not a decorative pattern. Follow any rib with your eye. Near each column the ribs bunch and swirl, because that is where the forces crowd on their way down; out in the open field they relax into long, calm arcs, because there the loads travel farther with less urgency. You are looking at a stress diagram, cast at full scale and left in plain sight.
Gatti Wool Factory, Rome (1951) · 5 images · swipe or use arrows. The floor slab's ribs run along the isostatic lines — the actual curved paths bending forces take through a loaded slab on their way to the columns. The coffered ceiling is not a pattern applied to structure; it is a picture of the forces themselves, made permanent.
The third invention is the one that made the Palazzetto possible, and it deserves to be walked through slowly, because it is the heart of the Nervi System. The pieces are called tavelloni. Nervi borrowed the word from everyday Italian construction, where a tavellone is an ordinary large hollow clay tile used to span between floor joists. The choice of word tells you how he thought about them: not heroic components, just tiles, humble repeatable units of a larger surface. His tavelloni were thin ferrocemento pans, a few centimeters thick, cast face-down in molds. For the Palazzetto dome, 1,620 of them were produced, in the diamond and triangular shapes you can see in the ceiling photographs above, and here is the number a student should hold onto: those 1,620 pieces came from only about thirteen distinct shapes. The dome's geometry was worked out so that the same few molds could be reused hundreds of times each. The expensive thing in shell construction had always been the one-off curved timber centering. Nervi replaced nearly all of it with molds amortized over hundreds of castings.
The construction sequence has three beats. First, cast on the ground. Every piece was produced at grade, at bench height, by crews working in good light and good conditions, with quality control that overhead casting can never match. Second, place in the air. The cured pieces went up onto light, movable scaffolding and were arranged into the dome pattern. Each tavellone was made with upturned edges, so that wherever two pieces sat side by side, their raised edges formed a channel, a trough running along what would become a rib. Steel reinforcement was laid into this channel network. Third, knit together with one pour. A crane working through the oculus at the crown poured concrete into the channels, forming the ribs, and capped the whole assembly with a thin slab, about four centimeters, over the top. When it cured, pieces, ribs, and cap became one monolithic dome, six hundred tons of roof spanning sixty meters, erected in about forty days. The dome's outward thrust runs down the inclined perimeter trestles into a prestressed concrete ring buried in the foundation, which is where the forces finally come to rest.
Then the punchline, and it is the whole Nervi ethos in a single construction detail: the formwork was never stripped. The tavelloni stayed in place permanently, and their under-surface is the finished ceiling. The famous coffered pattern of the Palazzetto is not decoration applied to the dome, and it is not even structure laid bare. It is the mold, promoted to finish. Every line you see in that ceiling is either the joint between prefabricated pieces or the underside of a rib channel doing structural work. There is nothing to add and nothing to hide, because the means of construction and the final architecture are the same object.
Nervi gave a structural engineer's proof to an old ethical claim: that expressive form is a consequence of function honestly pursued, not a layer applied afterward.
Two more things about Nervi matter, because they cut against the caricature of the engineer as a human calculator. The first: as a professor in Rome from 1946 to 1961, he insisted to his students that mathematics was necessary but never sufficient, and that trained intuition, built up over a working lifetime of watching how concrete actually behaves under load, carried at least equal weight in the design of a structure. The second is easy to miss and explains almost everything: Nervi was the contractor. His firm, Nervi & Bartoli, was a construction company, and he won his greatest commissions not by argument but by tender, by being the cheapest bid in the room. He priced what he drew and built what he priced. The forms were disciplined by building economics from the first sketch, which is precisely why they could be both radical and affordable. Keep both thoughts. They reappear, in a different accent, in Mexico City.
03 · Mexico City
Candela: Geometry as a Found Form
Félix Candela (1910–1997) came to the same destination from the opposite direction. Trained as an architect in Madrid, where as a student he watched Eduardo Torroja's thin vaulted shells go up over the Zarzuela racecourse grandstand, Candela fled Spain after the Civil War and landed in Mexico in 1939. Within a decade he had stopped practicing architecture in the conventional sense and reinvented himself as a designer-builder of thin-shell concrete structures, eventually completing several hundred of them through his own construction company, Cubiertas Ala.
The problem is the one every beam has. A beam or a flat slab resists load by bending: compression along its top, tension along its bottom. That is why it needs depth, and depth is where the material, the weight, and the money go. Candela's trick was less an invention than a recognition, and it is worth being precise about where the magic lives, because it is not in the material. His shells are ordinary reinforced concrete: standard mesh and bar, standard mix, no better than a sidewalk's. What is extraordinary is what the geometry does to the physics. A doubly curved surface changes the rules. The curvature lets the load travel as compression, tension, and shear within the surface itself, in what engineers call membrane action, with almost no bending anywhere. No bending means no need for depth. What to look for in the photograph below: the roof of the Cosmic Rays Pavilion is as little as five-eighths of an inch thick, about the depth of a dinner plate, and it was a working laboratory roof, not a sculpture. Its thickness was set not by strength, which was never in question, but by the practical minimums of covering the steel and placing the mix.
Cosmic Rays Pavilion, UNAM, Mexico City (1951), with architect Jorge González Reyna; shell by Candela. A working laboratory roof reduced to five-eighths of an inch of concrete — the building that announced what the hypar could do.
The surface behind nearly all of Candela's mature work is the hyperbolic paraboloid, the hypar, and the property that makes it buildable is the one a student should memorize: although the surface is curved in two directions, it can be generated entirely from straight lines. Sweep a straight line through space along two tilted rails and the saddle appears. At Los Manantiales in Xochimilco, four hypars intersect to form an eight-lobed groined vault, a shell a few centimeters thick with no interior column anywhere. What to look for below: the design drawing shows the straight-line generators laid across the curved surface, the geometric skeleton of the whole enterprise, and the construction photograph shows what that skeleton meant in practice: a forest of simple wooden shoring carrying flat, straight boards that together approximate a surface no single board ever bends to follow.
Los Manantiales, Xochimilco (1958), with architect Joaquín Álvarez Ordóñez. Four intersecting hyperbolic paraboloids forming an eight-lobed groined vault, a shell a few centimeters thick, and no interior column anywhere — Candela's clearest statement that geometry itself can be the structure.
If Los Manantiales is Candela's geometry at its most lyrical, the open chapel at Lomas de Cuernavaca (1958, with architects Guillermo Rosell and Manuel Larrosa) is the same geometry pushed to its structural frontier: a single saddle-shaped hypar, self-supporting, spanning thirty meters and rising to twenty-one at its open mouth, the largest free-edge hyperbolic paraboloid built to that date. The slider below is a complete construction sequence, and it repays a slow read. In the first frames, look at the formwork: every board is straight, ordinary flat lumber laid along the surface's two families of straight-line generators by ordinary carpenters. No steam-bending, no curved centering, no exotic labor. In the middle frames, the standard mesh and bar go down directly over the boards, and the concrete follows, just four centimeters thick, an inch and a half, over nearly the whole shell, thickening only where the surface gathers its forces and delivers them to the foundation. In the final frames, the formwork has been struck, and the striations of the straight boards remain legible on the underside of the curve. The construction method is recorded permanently in the finished surface. The drawing and the building are one and the same.
Chapel of Lomas de Cuernavaca, Morelos (1958), with architects Guillermo Rosell and Manuel Larrosa; shell by Candela · 8 images · swipe or use arrows. Construction sequence: straight-timber formwork following the hypar's rulings, reinforcement placed, the shell poured to an inch and a half, and the formwork struck — the striations of the flat boards still legible in the finished concrete.
The economics deserve their own paragraph, because they are the reason Candela built hundreds of shells while his European admirers built few. His crews were small teams of Mexico City workers using hand tools. Concrete went up in buckets and was placed by hand, troweled onto the sloping formwork in a continuous, carefully choreographed sequence so that no cold joints formed in a shell only four centimeters thick. The formwork lumber, being straight and undamaged by the work, was struck, stacked, and reused on the next project, which is a large part of how Cubiertas Ala could routinely underbid conventional steel trusses. And because membrane action leaves so little bending to account for, the mathematics collapsed too: for a hypar under uniform load the entire analysis reduces to a single constant, a uniform shear throughout the shell, one line of arithmetic where the European academies demanded eighth-order differential equations. Candela dismissed their elaborate elastic theory as false precision applied to a cracked, creeping material that never obeys its assumptions anyway. Where real bending does appear, at the edges and supports, he answered with geometry rather than mass: parabolic edges left free and thin, stiffening hidden inside the groins.
Candela liked to say that his structural analysis was “a hobby,” a deliberately disarming way of stating the position the numbers above make plain. The form came first, found through an intuitive, exploratory process closer to a craftsman's than a calculator's; the mathematics served the shape, verifying rather than generating it. He insisted that any structural form worth building had to satisfy the aesthetic test and the structural test together, refusing to rank one above the other. Set him beside Nervi and the contrast sharpens both men. Nervi derived form from the flow of forces with the discipline of a proof. Candela discovered form inside geometry with the confidence of a builder who had watched hundreds of his own shells stand up. Opposite temperaments, opposite methods, and nearly identical results: thin, continuous, curving concrete enclosing large, free, uninterrupted space at a buildable price. One left his formwork in the building forever; the other left only its fingerprint in the concrete. Both eliminated the cost, and neither hid the evidence.
04 · Los Angeles
The Arrival: Free Space in the Nature of the Material
Why does any of this matter to a blog about organic architecture? Because the shell was the answer to a question the organic tradition had been asking for half a century. Wright broke the box decades before thin shells existed: dissolved its corners, opened its plan, cantilevered its edges into the landscape, and demanded that every solution be found in the nature of the material rather than applied against it. What he never had was an economical structure for continuous free-spanning enclosure. John Lautner, who carried Wright's project further than anyone, spent his career searching for exactly that: not a style to borrow but a structural solution to arrive at, unique to each site and each space, in the nature of the material. The work of Nervi and Candela was proof that the search had a destination.
Lautner went and met the proof in person. In the late 1950s, he and his client-collaborator Kenneth Reiner, the inventor-industrialist behind Silvertop, traveled to Mexico, where through the circle around the artist Mathias Goeritz they met Candela himself and toured several of his built shells. In the spring of 1960 the two men made a second research trip across Europe, ending in Rome, where Lautner studied and personally photographed Nervi's concrete structures rising for that summer's Olympic Games. Neither trip was tourism, and neither was a search for a concept to borrow. Lautner had already committed to the shell before he traveled. What he went looking for was the state of the art at full scale, built evidence from the two men who had carried the same material furthest from the engineering side, while he was in the middle of exactly the kind of structural problem they had spent their careers solving.
At Silvertop, Lautner had begun with a spatial concept, a soaring room bracketed between two curved walls under one gigantic arcing roof, and had worked through several structural schemes, including a conventional wood roof, before committing to what the space actually demanded: a vast concrete shell. His arrival at the solution came through a third technology, one neither Nervi nor Candela relied on: post-tensioning. Working with T.Y. Lin, the era's foremost authority on the method, Lautner realized a five-inch post-tensioned concrete roof spanning eighty feet, its strength drawn from steel tendons stressed within the slab after the pour, suspended from curved beams and poured in one continuous crane-and-bucket operation that remains one of the great construction stories in Los Angeles residential building. It was not Nervi's statics-first derivation and not Candela's geometry-first discovery. It was an architect with genuine working knowledge of structure, running back to his Taliesin construction years, directing engineering toward a space he had already committed to, and finding the answer where the organic tradition said it had to be found: in the nature of the material itself.
Silvertop, Silver Lake (begun 1956). A five-inch post-tensioned concrete shell spanning eighty feet — Lautner's spatial concept, refined into buildable structure with engineer T.Y. Lin.
05 · What It Adds Up To
Two Methods, One Proof
Set the two engineers side by side and the middle decades of the twentieth century come into focus as the moment the box finally became optional. Nervi derived free form from the discipline of statics and proved that an honestly solved structure needs no ornament because it is the ornament; his tavelloni made the mold and the finish one object. Candela discovered free form inside geometry and proved that it could be built by ordinary crews, with ordinary concrete, over straight ordinary lumber, at ordinary cost. Between them they turned spatial freedom from a heroic exception into a line item a contractor could price. That an architect in Los Angeles then arrived at his own solution, by his own route, was not a coincidence. It was the proof doing its work.
The revolution did not end with them. Candela's line runs directly into Santiago Calatrava, who has acknowledged the debt openly and whose bridges, stations, and museums carry the thin-shell sensibility, structure as sculpture and geometry as strength, into the present century. Late in life, Candela himself joined the design of the Oceanogràfic complex in Valencia, where the lobed restaurant pavilion is, in essence, the Xochimilco flower redrawn forty years on. Nervi's prefabrication methods became the quiet grammar of long-span construction worldwide, and both men remain foundational references wherever engineers study how geometry can substitute for mass. Their buildings are sixty years old and still cited as the state of the art.
The lesson underneath it all is single and unfashionable: spatial freedom is never free. Someone pays for it in structural intelligence, in following forces honestly, in finding the geometry that wants to stand, in refusing the cheaper frame when the space demands the shell. The engineers who broke the box did not do it for architecture's sake. They did it because it was true. That the truth turned out to be beautiful is exactly what the organic tradition had been claiming all along.
ARCHITECTOID RESEARCH SERIES · STRUCTURE & SPACE · SULLIVAN → WRIGHT → LAUTNER
Comments
Post a Comment