Auguste Perret and the Concrete Column
from Concrete: The Vision of a New Architecture
Peter Collins
In examining Perret’s doctrine in detail, it will be convenient to begin with his attitude towards the design of columns, the element in his vocabulary to which he attached the most importance. Columns have traditionally been regarded as the very symbol of architecture, since they express more forcibly than any other element the peculiar dignity of the architect’s achievement, and it is only within recent years, mainly as a result of their abuse in the nineteenth century, that their use on façades and in interiors has been looked upon with mistrust. As we have seen, Perret’s substitution of reinforced concrete columns for load-bearing partition walls at 25b rue Franklin was interpreted by the younger generation as an invitation to do away with visible supports altogether, and this attitude became more and more pronounced as the new concepts of spatial integration spread, and the desire to exploit the slenderness of new structural materials increased. The traditional idea of a colonnade, with its rhythm of vertical supports forming imposing perspectives throughout interiors and across façades, was discarded as both structurally unnecessary and inappropriate to modern planning, so that the columns themselves, which at the beginning of the century were already being disguised by Art Nouveau decorators as naturalistic stems, were now reduced to their minimum frequency and dimension so as to be as unobtrusive as possible, if not completely unseen.
Auguste Perret liked columns, and proliferated them with all the enthusiasm of the architects of the past. It was not only the inherent structural dignity of the column itself which captured his imagination but the powerful emotional effect created by receding ranks of columns, and the optical function fulfilled by such rhythmic sequences in creating an awareness of scale beyond the effective bounds of stereoscopic sight. However much the Renaissance principles of perspective might be outmoded as a means of pictorial representation, the abiding reality of its laws as a means of apprehending spatial relationships seemed to him incontestable, and instead of diminishing or camouflaging his structural supports, he sought every means at his disposal to isolate them in space, and make their rhythm provide the dominant unifying element of his designs. This did not mean that he limited himself on principle to spans only possible in older types of construction, or multiplied columns unnecessarily where they were not structurally required; but it did mean that instead of feeling morally bound always to use the maximum spans obtainable by civil engineering, he considered himself free to use intermediate supports whenever desirable, and saw no objection to dividing a large hall with interior colonnades, provided that these were not in any way detrimental to its use.
In order that full aesthetic advantage could be taken of these structural columns, it was necessary to ensure that each would be a pleasing object in itself. Pride in his new material, and a growing contempt for veneers and renderings, made him seek beauties inherent in concrete, but unlike most of the early theorists, he was not rash enough to assume that the resultant shapes must be of a kind never seen before. On the contrary, he based his researches on those inalienable characteristics, derived from structural and optical laws, which all vertical supports must have in common, and being a building contractor, he perceived that it would be the methods of casting the material, rather than the nature of the material itself or any presupposed theoretical analyses and calculations, which would be the cause of such modifications as might distinguish concrete from worked stone.
The engineers of the period, who had had no reason to abandon the conventional theoretical assumptions of steel frame construction, designed each stanchion as one continuous support rising from floor to floor, and considered the most distinctive characteristic of reinforced concrete frames to be the ‘haunches’ of the beams, namely the flared ends of the beams which increased the area of concrete under compression near the vertical supports. It soon became apparent to Perret, however, that despite the effective continuity of such columns from the point of view of theoretical calculations, they were not in fact constructed continuously, but were cast independently in sections one storey high. Similarly, though the beam might be conceded theoretically as requiring an increased amount of concrete at its extremities, it was in practice more efficient to make the form-work completely regular, and hence horizontal at its lower edge. In consequence, Perret reversed the normal conception of a reinforced concrete frame. The columns now butted into the beams, and the latter formed a simple grid pattern on a level plane, so that henceforth continuity was horizontal and hence also visible, and each internal column appeared as a distinct element, expressing the way it was cast. It was as if the standard Hennebique diagram had been turned on its side.
The next stage was to consider the character which the form-work should take. In Hennebique’s earliest industrial buildings, the columns had been of simple rectangular section, which was logical enough, since this was the cheapest way of assembling timber planks, and Perret himself had used a similar method in the garage in the rue de Ponthieu, where the only major modification was the introduction of diminution and entasis in the columns of the façade. He did not have any occasion to experiment with the shape of vertical supports in monumental buildings during the ensuing twenty years, for the columns of the Théâtre des Champs-Elysées had had to be veneered with marble or plaster, and his other private commissions were mainly for the most utilitarian kind of industrial building, where such refinements would have been extravagant, if not out of place. In 1922, however, an unexpectedly favourable opportunity presented itself when he was commissioned to build the church of Notre Dame du Raincy, for the propaganda of Viollet-le-Duc had disposed the public to accept the idea of ecclesiastical architecture as structure unadorned, and it was thus possible to concentrate here on the refinement of bold and original structural methods without the risk of outraging popular taste, or inviting unfavourable comparisons with the processes of factory design.
The result was undoubtedly the most revolutionary building constructed in the first quarter of the present century, and it justly occupied the place of honour in the first issue of the first magazine specifically devoted to the New Architecture: L’Architecture Vivante. The design comprised four rows of free-standing columns 37 ft. high, spaced 33 ft. apart along the length of the nave, and diminished from 17 in. at the foot to 14 in. at the summit. Being free-standing, and thus unaffected by the normal need to receive the abutment of intermediate beams or partitions, there was no practical obligation to make the columns rectangular in section, and therefore, despite the increased cost of the form-work, Perret made them round. This shape was preferred for two reasons; firstly, because it was most economical (in the structural sense of the word), in that it provided constant rigidity from every angle, like a tree trunk, and was, as he himself pointed out, ‘best adapted for a member subjected to compression’1; secondly, because it was more satisfactory optically as a result of the gradations of shadow and constancy of silhouette. The architects of the eighteenth century, like those of antiquity, had studiously avoided free-standing square columns (or ‘pilasters’ to give them their correct, though now misapplied name), because apart from their clumsiness, their apparent width varied according to whether they were seen diagonally or from the front. As geometric projections or drawings, a series of pilasters might well appear similar in appearance to columns, but when viewed in perspective, the width of each pilaster would seem to increase as the angle of vision became more oblique. It was thus not merely in obedience to Platonic ideals (whereby the circle was regarded as the configuration of perfection), or in obedience to the belief that the most ‘natural’ architectural forms were those found in natural organisms such as trees, that the Greeks used cylindrical columns; it was also in obedience to visual experience and optical laws, which demonstrated that rhythms discernible on plan did not necessarily produce the same rhythmic effects when projected into three-dimensional space.
To give greater elegance to the columns of Notre Dame du Raincy, and also (since this was his first attempt to give entasis to round columns) to make provision for any inaccuracies in casting, Perret modelled the surface of the form-work with a series of flutings. These were not regular segmental grooves, as in antique columns, but an alternation of curved projections and angular fillets more suggestive of Gothic composite piers. Similarly, by applying diminution and entasis to columns twenty-three diameters high, Perret showed that the forms were in no sense dictated by a subservience to either Greek or Gothic prototypes, but were on the contrary an attempt to extract the most rational elements from both. Nevertheless, it is clear that although he scrupulously observed the structural characteristics of the new material as he then understood them, Perret was still, like Labrouste before him, unable to free his mind entirely from standard historical conventions, and although he must have known, from reading Choisy, that Mycaenean columns tapered in the opposite direction,* he as yet saw no cause to modify the traditional assumption that a column must always derive its stability from being narrower at the summit than at the base.
It is not known precisely when he first began to question this assumption, but his first radical modification of the traditional lithic shape seems to have been introduced when designing the new Palais de Chaillot in 1934 for the 1937 Paris Exhibition. The authorities had decided to demolish the old Trocadéro Palace, built by Davioud as a temporary centre-piece for the 1878 exhibition, and to substitute a new monumental group of permanent buildings on the same site (the Mont de Chaillot) which would fulfill the same function of housing exhibition and concert halls, and would also terminate effectively the splendid axial vista extending through the Eiffel tower from Gabriel’s École Militaire. Perret, who had been entrusted with the commission, abandoned the earlier Baroque composition, with its dominating central structure and curved concatenated wings, and substituted two symmetrical groups of museum buildings linked by an open colonnade. The resultant composition was, in character, reminiscent of Jules Hardouin-Mansart’s Grand Trianon at Versailles, but the difficulty here was to design a colonnade appropriate both to the new material and the vastly increased scale. Perret finally decided to arrange giant columns, 83 ft. high, in four rows, making a hundred and four in all, across the whole width of the Place du Trocadéro, and roof them at the same height as the four-storey museum blocks on each side.
The task of designing an open colonnade 17 ft. higher than that of the Madeleine presented aesthetic problems of the greatest delicacy, since the whole visual effect of the ensemble depended on the success of this focal element of the design. The columns could conceivably have been designed as simple cylinders, like those of the nearby Musée d’Art Moderne to be built contemporaneously as part of a similar though much smaller composition, but the harshness of such a rigid silhouette woul...