Ahead of the Curve: Dublin Airport and the Duval Plan

From the IAR Archive:

Frederick O’Dwyer follows the evidence to unearth the design sources for Dublin Airport’s original terminal at Collinstown

In Ireland the subject of airport architecture has always been associated in the public imagination with the original Dublin Airport terminal at Collinstown (1937-42), an exciting International Style exemplar, harmonizing aesthetics and functionality with its curved plan, tiered viewing terraces and external spiral staircases. It won the RIAI gold medal for the young Desmond FitzGerald (Fig 7), airports architect of the Office of Public Works, in 1943. While its construction was due to the enlightened policy of Seán Lemass, minister for industry and commerce, the design was kept under wraps until after the end of the war, creating a mystery about its origins and indeed doubts that FitzGerald, who never designed another building of such quality, could have done it unaided. Sean Rothery, who had worked with FitzGerald as a student in the post-war era, found him reluctant to discuss his inspiration for the airport or indeed his young assistants, when he interviewed him in 1985.1 In his later book, Rothery stated that many contemporaries believed that he was ‘not the main author of the design but that it was largely the work of the young team under him’ though he conceded that this view was not conclusive.2 It was certainly very different from an RIAI prize-winning 1934 airport FitzGerald designed as a student project, a conventional brick-faced rectilinear structure placed tangentially to a circular airfield of the type then popular in Germany.3 In contrast, Hugh Pearman has written of the Collinstown team: ‘Starting with a blank sheet of paper and clearly aware of the writings of Le Corbusier, they set to work. The modern world was best epitomized by the latest generation of wonderful aeroplanes, and so the building would respond aesthetically to the design of the planes themselves’.4 Two questions arise. Firstly, was the paper really blank; would one design a hospital for instance, without undertaking planning research and studying precedents, and secondly, is FitzGerald entitled to be regarded as the ‘main author’? No account of the building was published before the end of the war in Europe, the first substantive piece appearing in the Irish Builder and Engineer on 28 July 1945. It stated: ‘this curved plan … is now largely accepted as the best solution to the problem of providing the smallest possible plan to serve the maximum number of aircraft, by supplying a long periphery to accommodate them. It is believed that Collinstown was the first airport to incorporate this principle, which has since been widely adopted, a notable example being the Washington, USA airport.’ The tradition that Dublin Airport was sui generis seems to be largely grounded on this anonymous statement. In fact the new National Airport in Washington, by Cheney, Goodman and Chandler, was completed in 1941, a year before Collinstown, while a similar plan already had been used for Delano and Aldrich’s main terminal at New York Municipal Airport (La Guardia), built 1937-39. Might it not be more likely that these buildings were derived from a common prototype? The British journal Flight published a full description of Washington on 27 November 1941, just ten days before America entered the war, which describes a convex layout very similar to Dublin, permitting an optimal location of runways unobstructed by buildings, as the ‘Duval plan’.

Albert-Bernard Duval (b.1891), an aviator and a lieutenant in the French naval reserve, is now an almost completely forgotten figure. He first described his plan or système, apparently unpatented, in the journal L’Aéronautique in 1921. It was later featured in the popular journal L’Illustration in August 1929 in which he illustrated examples of projects being developed at Lyon-Bron, by the architects Chomel and Verrier (where the terminal is clearly an angular prototype for Collinstown), and at Middle River, Baltimore. Not described however, was a smaller but significant Duval-inspired American project, Burbank, California, by the Austin Company of Cleveland (Fig 6). Opened in May 1930, this was to prove the most influential of all though by then Duval’s interest had shifted to aerial navigation, four editions of his treatise on the subject, written with Louis Hébrard, appearing between 1922 and 1950.

In 1929 the Royal Institute of British Architects set up an aerodromes committee, with the objective of studying airport design. Its members soon spotted the emergence and indeed advantages of the Duval plan though in the four years of its existence, they seemingly never mentioned him by name, using the generic term ‘V-type layout’ in their reports. The committee’s secretary John Dower visited Lyon-Bron while Graham Dawbarn (1893-1976), who received an RIBA bursary in 1931 to visit airports in the United States, made a particular study of Burbank during a three-week tour in a small plane piloted by his associate Nigel Norman. Norman (1897-1943), an aeronautical engineer and squadron-leader in the auxiliary air force, was the chairman of Airwork, an aviation conglomerate which had built Heston aerodrome in 1929 (designed with L.M. Austin). This had a shallow V-plan whereas at Burbank the hangars were set at ninety degrees with a curved terminal at the apex (Fig 8); identical to the plan subsequently used at Collinstown (Fig 9), albeit that the Burbank elevations were in the Spanish mission style. The virtues of Burbank (and similar ‘runway airports’ at San Francisco, New Orleans, Buffalo and Pittsburg) were subsequently extolled over European grass airstrips by Norman and Dawbarn in a number of lectures and articles on airport design, though Duval was never mentioned. It was only after the RIBA committee was subsumed by the air ministry into a multi-disciplinary advisory board in 1933 that civil-engineer members like Sir Leopold Savile gave the Frenchman his due credit.5

In 1933 Norman set up an aviation consultancy with Dawbarn and Alan Muntz. To promote the venture they cultivated the Marquess of Londonderry (air minister 1931-35), for whom they built an airfield at Newtownards and through whose influence they obtained the commission to design a major new airport at Lydda in Palestine (1934), on the empire route to India. While not curved, Lydda (Fig 5) looked very like Collinstown in profile, ascending in a series of terraces while the airfield was laid out on the Duval pattern. For their next big project, a municipal airport terminal at Elmdon, commissioned by Birmingham Corporation in 1935, Norman and Dawbarn used a coat-hanger shaped plan, curved on the airside (Fg 4). This was rejected by the client in mid-1936 on cost grounds,6 exacerbated by excessive circulation areas and the provision of surplus unfitted-out space for future expansion. Notwithstanding these problems, the architects continued to publicise it for some months; its successor, a smaller very different design with striking concrete canopies, was approved in July 1937.

The Irish government had long-term plans for the development of the old RFC station at Collinstown as a civil airport; this was approved in principle on 6 December 1935 though nothing much happened. It was envisaged that the two Dublin local authorities would contribute substantially to the capital and running costs, the government having committed itself the same month, at a conference in Ottawa, to prioritising the building of a transatlantic airport on the western seaboard as well as setting up a joint operating company with Imperial Airways and Pan Am for services to North America. Rineanna, on the river Shannon, had been selected for development as a combined landplane/flying-boat facility with a temporary base for the latter at Foynes. The Irish authorities seemed convinced that flying boats would prevail for some years on the transatlantic and proceeded on that basis, projecting a longer timeline, in excess of two years, for the preparation of the Rineanna airfield. However, the British air ministry was already proceeding with the construction of a large landplane base at Hattie’s Camp (later renamed Gander) in Newfoundland. Although the airlines realized that flying boats could not land in Newfoundland in winter, they showed no particular interest in purchasing landplanes; the flying boats would take an indirect route via the Azores between October and April. All this changed on 19 November 1936, when a Pan Am delegation, with Charles Lindbergh as technical advisor, arrived in Dublin to meet Irish ministers. In private discussions Lindbergh dropped a bombshell: Pan Am and Imperial Airways wanted to commence landplane proving flights in September 1937 when the Newfoundland airport would be ready. There being no prospect of developing Rineanna in that time, the cabinet was forced to switch its focus to Collinstown where the existing grass runways could be upgraded. Officials were dispatched to London to obtain the advice of the air ministry while the OPW assigned the task of preparing designs to its principal architect, T.J. Byrne (1876-1939), who on 11 January 1937 took on FitzGerald as an assistant architect. Advertisements were also placed for a senior (project) architect, but Byrne never filled the post. The secretary of the department of industry and commerce, John Leydon, had initially thought that temporary buildings might suffice but then decided to proceed with a proper terminal, a ‘first-class airport’, even though no commitment had been received from the local authorities (who subsequently dropped out). It emerged that the initial proving flights would be carried out by a new type of aircraft, the de Havilland Albatross, which the air ministry was ordering directly from the manufacturer, and that a suitable hangar would have to be built at Collinstown. Byrne now joined an inter-departmental committee which had been sitting since September; he reported regularly on progress, but did not bring FitzGerald along until 21 April. Byrne later produced a chronicle which reveals that the ‘sketch plan stage’ ran from 15 February to 3 May 1937; interim sketch designs (not located) being produced on 11 and 24 March.7 The 3 May scheme was subsequently modified by shortening the wings and adding a fourth storey. A partial set of the June 1937 general arrangement drawings survives in the OPW, revised in June 1939, with FitzGerald’s name at the foot and initialled by him, D.O’T. and C.G.H. The latter, unidentified, may be a technician8 while the former is Dermot O’Toole (1910-70), who had qualified before FitzGerald and had worked, inter alia, for Scott and Good. Four other assistant architects named by Rothery appear to have been taken on later, during the working drawing and construction phases. FitzGerald’s own coloured but uncaptioned prints (in the Irish Architectural Archive) appear to be from the June 1937 set, and differ only in minor details from the drawings revised in 1939.

The Collinstown design, on the Duval system, may be seen as a refined version of Norman and Dawbarn’s Elmdon 1, with a more efficient use of space, notably tighter circulation and provision for future expansion by way of a largely open-framed second floor rather than the objectionable unfitted-out shell. The spiral staircases at each end are a notable common feature but are exposed rather than enclosed. While the landside elevation of Collinstown (Fig 3) has much in common with Lydda, the airside is much more progressive than either it or Elmdon, being International Style rather than art deco. The influences here are German, the glass banding (Fig 1) recalling Erich Mendelsohn’s Schocken department store in Chemnitz (1930) and Columbushaus in Berlin (1931). Both FitzGerald and O’Toole were familiar with Mendelsohn’s work.9 If, as appears to be the case, the Birmingham design was available as a template, how might this have arisen? The most obvious connection is that Dawbarn was employed to design the Collinstown and Rineanna hangars, which were to be based on one he had built for Nigel Norman at Heston in 1935. While Byrne formally suggested retaining Dawbarn’s services for these on 19 March 1937,10 this was probably first raised by the air ministry at its meeting with Irish officials in early December 1936.11

In his chronicle, Byrne states that the only research undertaken for Collinstown was a visit to the RIBA exhibition ‘Airports and Airways’, in late February 1937, accompanied by the OPW’s chief engineer J.P. Candy (but not FitzGerald); an inspection of the new (circular) Gatwick terminal and a small aerodrome at Gravesend the same week; and the perusal of some periodicals. This was in keeping with an instruction from Leydon that only one or two airports need be inspected. Byrne had advised the committee that his attendance at the exhibition would facilitate the preparation of sketch plans on his return.12 A perusal of the catalogue reveals little of real relevance to airport planning except the curiosity that a model of Norman and Dawbarn’s unexecuted Elmdon terminal was withdrawn before the exhibition opened, even though the City of Birmingham was thanked in the credits. Was a deal struck with regard to it? A connection may be found in a special airports issue of L’Architecture d’Aujourd’hui (September 1936), formerly owned by FitzGerald and now in the IAA.13 Much of the content extols the virtues of the Duval system but there are also floor plans of Elmdon alongside one of which is a pencil sketch for the plan of Dublin Airport (Fig 2). Could it be that Dawbarn gave the magazine to Byrne or FitzGerald and made the sketch, pointing out how a reversion to the Burbank double curve would solve the Elmdon problems? Such a link may explain why Byrne was defensive about the Collinstown design, imposing an embargo on publication from the start, resisting requests from the press and even Lemass and, as late as October 1938, expressing the opinion ‘that the building plans should not be published until the structures are approaching completion’.14

It may be argued that many of the doubts surrounding FitzGerald’s authorship arise from the sui generis hypothesis which is plainly incorrect – if Collinstown is viewed instead as a derivative but progressive design, superior to Dawbarn’s Elmdon, then it is easier to visualize him as the ‘main author’, albeit with the supports of senior and junior staff. While Byrne drove the project forward (up until his sudden death from a heart attack in January 1939) and provided the link with Dawbarn, the aesthetic freshness of Collinstown accords more with the work of young enthusiasts like FitzGerald and O’Toole rather than that of a sixty-year old of the Edwardian school. FitzGerald’s name is on the drawings and, from the evidence of the surviving written records, he worked prodigiously. The runways at Collinstown were ready by late 1937, but the Albatrosses were not; the proving flights were postponed to the autumn of 1938 and ultimately cancelled. Pan Am did not fly landplanes across the Atlantic until after the war.

Frederick O’Dwyer is an architect, architectural historian and town planner.

From the IAR Archive
First published in the Irish Arts Review Vol 29, No 2, 2013