47 So High You Can’t Get Over It: Neo-classicism, Modernism and Colonial Practice in the Forming of a Twentieth-century Architectural Landmark Michael Findlay Amyas Connell (1901-80) was a New Zealand-born architect and a leading figure in British modernism. His first commission, High and Over (1929-31), for the archaeologist and classical scholar Bernard Ashmole, is among the earliest modernist houses built in England and one of the few buildings by a New Zealand architect that commands a significant place in international architec- tural history. This paper examines High and Over from a New Zealand perspec- tive. Connell’s critical profile has been partially shaped by the soubriquet ‘Wild Colonial Boys’, used to frame the work of Connell and others by the English architectural historian Dennis Sharp. Such an interpretation overlooks the depth of Connell’s experience prior to High and Over and arguably overplays his anti- theoretical and outsider status. This paper explores the design sources for High and Over. It was an unusual fusion of classical geometry, Beaux-Arts planning, Corbusian aesthetic, and Parisian décor and landscape. Connell’s interest in the villas of Hadrianic Rome was evident in High and Over’s centralized plan, which reveals a regard for clas- sical geometry that transcends the usual explanations for such practices in the Beaux-Arts teaching systems adopted in British architectural education. This rich combination of elements has polarized opinion on its merits as a piece of mod- ernist architecture. The High and Over project also included a number of related structures set in a landscape plan not usually included in analysis. The land- scape plan is considered here, with reference to French modernist garden design of the 1920s, including the designs of the Armenian architect Gabriel Guévrékian at the Paris Exposition and the Villa Noailles at Hyères (1927). High and Over occupies a place where the traditions of classicism and the emer- gent features of modernism intersect. Connell may have produced a distinctively British form of classical modernism, but it is also an artefact of New Zealand origin, with its roots in rural Taranaki. *** Amyas Douglas Connell (Fig. 1) was born in the Taranaki dairy town of Eltham in 1901, where his father was a professional photographer and landscape artist. Amyas also wanted to be an artist but was unsure of his future direction, and after his final year at school, laboured on a dairy factory construction site over INTERSTICES 09 49 viewed Le Corbusier’s Pavillion de l’Esprit Nouveau. In his recollections from the 1960s, Ward was dismissive of the Exposition apart from Le Corbusier’s contri- bution. This is a retrospective view from his position as theorist and historian of the practice, expressed at a time when Art Deco was still regarded as an inferior manifestation of modernism (Ward, 1967: 74). There was much at the exposition to excite young colonial architects. Amongst the most talked about sights was a cubist garden, Le Jardin d’eau et de luminiére, by the Armenian architect Gabriel Guévrékian (Imbert, 1993). This densely organized triangular garden used a revolving faceted glass sphere, tiered triangular concrete pools, horizontal water jets, coloured electric lights and the optical vibration of complementary colour planes to explore notions of space and time. As with other French modernist garden designs, this was nothing to do with horticulture or the picturesque. It was an ‘instant garden’, constructed from metal, glass and hard-edged concrete. Avant-garde French garden design had a close relationship to the contemporary visual and plastic arts. Exhibited elsewhere were Parisian architect Rob Mallet- Stevens’ sculptures that resembled Cubist trees made from slabs of concrete. Connell and Ward chose to study at the Architectural Atelier associated with the Bartlett School of University College, London. The Atelier has been described as an add-on to the Bartlett, surviving from the aspirations of Reginald Blomfield (1856-1942) to apply the architectural teaching systems of the French École des Beaux-Arts to British architectural education (Campbell, 1989: 133). Blomfield has often been simplistically characterized as a bombastic reactionary. While ad- amantly opposed to modernism, he was also an admirer of progressive French architecture and its teaching methods and used his influence to reform British architectural education along those lines. Studying at the Atelier did not require formal enrolment or lead to any qualifications, but it was considered a useful place to network and establish connections as well as for preparing the large and complex drawings required for competitions. The Atelier was led by French-born and Beaux-Arts trained Hector Othon Corfiato, whom both Connell and Ward regarded highly (Ward, 1967: 78). Corfiato built relatively little in his lifetime but his teaching was influential over the generation of students that came together at the Bartlett. Corfiato was cosmopolitan in outlook and ensured that the focus was on France and Northern Europe as opposed to the concentration on Ameri- can models at the competing Liverpool School of Architecture under C. H. Reilly. Connell’s New Zealand experience seems to have measured up well against his British contemporaries, as it became his intention during 1925 to compete for the prestigious Rome Scholarship in Architecture (Ward, 1967: 78). Connell was informed that he had been awarded the Rome scholarship in June 1926 (Shaw, 1926). Ward had also entered the competition and gained second place but had neglected to join the RIBA and was therefore ineligible for the Henry Jarvis Travelling Studentship. He was awarded a special studentship from the Jarvis Fund instead (Personal, 1926: 10). The Rome scholarship provid- ed an annual stipend, board and lodging – financial support that was crucial to Connell’s extended stay in Europe as he had no private means. Connell and Ward celebrated by taking three months leave of absence and returning to New Zealand where civic receptions were held to mark their wins (Shaw, 1927). They arrived back in London for Christmas 1926, working their passage once again, this time on the SS Ruapehu (Connell, 1926). Despite being “the highest tree in the orchard”, personal rifts in its management and declining standards in its teach- ing programmes had shaken the status of the British School at Rome. As Louise the summer of 1918 (J. Connell, 1985). The site is unknown but Taranaki was in the midst of a construction boom led by Eltham architects J. W. Rough and J. Duffill, who designed over 100 dairy buildings in this region alone (Bartle, 2006). Many of these were built in steel reinforced concrete. Pease’s Building in Bridge Street, Eltham (1910), was designed with a fully suspended concrete floor; and the slightly later Wilkinson’s Building (1913) was a row of five two-storey shops with a total frontage of 41 metres, using monolithic construction with beams cast into the floor slabs (South Taranaki District Council, 2000). This ferro- cement building with its classical façade incorporated the two main threads of Connell’s early practical experience in New Zealand: advanced construction and neo-classical formality. His brother Jock recalled that this first job ended follow- ing a disagreement with the foreman over the correct way to lay foundations, but the experience focused Amyas’ interest in architecture. In 1919, he travelled to Wellington to be articled to the firm of Young & Fearn, where he was offered a four-year term and a relatively generous salary of £2.3.4 per month, rising to £5.8.4 at the fourth year (Item 54, 1916-40: 8). It was common for articled pupils to receive no payment at all under these arrangements and Connell’s remuneration suggests that Stanley Fearn expected to get good work from him. British-born Fearn was the first recipient of the New Zealand Institute of Architects (NZIA) Gold Medal in 1927, awarded for the William Booth Memorial Training College in Wellington, a project actually completed in 1913 (Walker, 2005: 30). Connell received an up-to-date training with the firm and at night classes organized by the Wellington Branch of the NZIA. This Atelier training followed the model established by the Royal Institute of British Architects (RIBA) and the university-based schools of architecture. Connell’s first-year course of study included the history of architecture, practical mathematics, structural mathematics, building construction, freehand drawing and an oral examination. Solutions to design and construction problems were also submitted in the form of a ‘testimony of study’ (Item 57, 1916-40: 65). Connell won a competition organ- ized by the Wellington Branch of the NZIA for a war memorial in 1921 (NZIA Wellington Branch, 1921: 12-13), followed by a third place in an NZIA-sponsored national competition behind F. Gordon Wilson (1900-59), the future Government Architect (Notes, 1922: 75). Fearn was a strong advocate of young architects travel- ling abroad, and advised the students of the Atelier to, “Let it be the aim and object of everyone to eventually get to the Beaux Arts, Paris, even if only for six months, for that is the fountain head of the world where architectural knowl- edge is concerned” (Untitled, 1921: 86). Fearn was acknowledging that the Atelier method of training had its limitations and that talented students would benefit from exposure to British architectural teaching. The outward flow of New Zealand architects to Britain was increased after the First World War when places in the schools of architecture were offered to demobilized ex-servicemen and an expatriate colonial presence began to form in London and Liverpool. Connell met his future professional partner Basil Robert Ward (1902-76) through the Wellington Architectural Students Association. With Fearn’s encouragement, both decided to seek further opportunities in London and it added to their myth in later years that they worked their passage as coal trimmers on the SS Karamea. Connell rented a room at Clements Inn Passage in central London and worked for Arts and Crafts architects William and Edward Hunt (News and Topics, 1926: 872). Although money was tight, Connell and Ward visited the 1925 L’Exposition Internationale des Arts Décoratifs et Industriels Modernes in Paris, where they Fig. 1: Amyas Connell and sister Beatrix, 1926. Eltham Historical Society. Fig. 1 INTERSTICES 09 51 attention to the Villa’s complex hilltop setting. The largest of Tiberius’ numerous structures on the island, the Villa originally stood five levels high and was a demanding structure to represent. Even at 1/16”:1’0” scale, the plan sheet measured 7 x 6 feet, and was accompanied by elevations and sections almost as large. These draw- ings were left incomplete when Connell resigned his scholarship in January 1929 to return to London and commence his first architectural commission. Connell’s first client was the retiring director of the British School at Rome, the classical scholar Bernard Ashmole (1894-1988). Ashmole was young, energetic and liberal – an unusual combination of attributes in a director of the Rome school – but he had been appointed in 1925 in order to rescue the school’s failing reputation. He regarded Connell as the star of the Rome school and later wrote that Connell’s study of the Capitol in Rome revealed “their subtleties perhaps more thoroughly than anyone before, except perhaps their creator” (Ashmole, 1994: 51). Connell returned to London and re-established contact with a fellow colonial student from the Atelier, Stewart Lloyd Thomson (1902-90). Connell and Thomson set up at 11 Mecklenburg Place in the heart of Bloomsbury, conveni- ently close to University College and Ashmole’s office as the newly appointed Yates Professor of Archaeology. While the first existing drawings for High and Over, dated June 1929, are signed Connell & Thomson (Connell & Thomson, 1929),1 neither mentions the other in any known source and their working relation- ship is therefore conjectural. Apart from High and Over, the outlook was bleak. Jock Connell recalled that his brother had failed to hold on to a major potential client in Philip Sassoon, who had commissioned him to design a garden at his restored mansion house, Trent Park (Sharp, 1994: 62). Thomson only records on his ARIBA candidate’s statement that he began professional practice in 1929 and made no claims for collaboration on High and Over (Thomson, 1939). His first British commission, the house St Raphael in Hornchurch, London, did not even- tuate until 1932, by which time Connell had moved on, rejoined by Basil Ward. Even so, Thomson can be counted as another significant colonial practitioner of the flat-roofed, thin-walled, reinforced concrete modern house in Britain. Thomson has disappeared from the published account of Connell’s early career, but their brief association is further evidence of the colonial nexus that formed at the British schools of architecture after the First World War and had such great in- fluence over the shape of British modernism. Aside from Thomson and Ward, Connell’s milieu formed around a network of mainly colonial and foreign modernist architects. He was a friend of the Australian Raymond McGrath and 1. See also a perspective drawing in Sharp (1994: 26) Campbell has noted: “Within a decade of its creation (in 1912), the faculty began to use the Rome scholarship not simply to encourage systematic working methods, clarity of planning and good draughtsmanship but actually to discourage what it termed ‘modern tendencies’” (1989: 131). As well as being chronically under- funded, there was no consistent educational policy and the instruction tended to reinforce the divide between architecture as a profession and an academic discipline. The course was criticized for its lack of relevance to practice and a large number of Rome and Jarvis prize-winners went back to teach at schools of architecture. In 1924, the term of the scholarship was shortened to two years, with a third year available on application, in an effort to meet the requirements of students keen to enter practice. Despite its flaws, the Rome prize seemed to be created for individuals like Connell and Ward. The scheme was established to foster talent in young architects who found it difficult to pay their way through articled training and university-based education, and was open to any British subject under the age of 30. Connell and Ward were not the first New Zealanders to achieve success in this field. Edward Armstrong (1896-1992) was awarded the Jarvis Scholarship in 1921 but resigned in five months and returned to New Zealand for the Robert McDougall Art Gallery commission (Campbell, 1989: 151). The same honour went the following year to Akaroa-born George Checkley (1893-1960) who, like Armstrong, entered the Liverpool School of Architecture on an armed forces serviceman’s grant (Sharples et al, 1996: 168). Both Armstrong and Checkley went on to contribute to the British modern movement, Checkley through a pair of early houses at Cambridge and Armstrong as the designer of sensitively de- tailed blocks of working-class flats in and around London. Checkley, although a retiring individual, was an important link for the colonials of the late 1920s. He was part of the influential architectural set at Cambridge that included the Australian Raymond McGrath (1903-77) and his client Mansfield Forbes, whose remodelled Georgian house Finella (1929) was an early rallying point for British modernists. The main focus for architecture students at the British School at Rome was the restoration through drawings of ruined ancient sites. This was carried out in an interdisciplinary manner alongside students from art and archaeology departments. In his first year, Connell completed drawings of the Campodoglio including its Roman remnants, and researched the archaeological bibliography of Capri in preparation for his restoration of the Villa of Tiberius (Fig. 2). Sum- marizing his year’s work, Connell wrote of his admiration for “the great villas of Imperial Rome”, due to their interesting layouts and “purity of style” (Connell, 1927). The Rome scholarship included funding for travel and Connell reported visiting Naples, Salerno, Amalfi, Ravello, Genoa, Florence, Lucia, Pisa, Siena, Venice, Ravenna, Bologna, Padua, Turin and Milan. Additionally he spent ten days in Paris at the École des Beaux-Arts studying drawings, and managed to visit Dijon, Lyons, Marseilles and the Riviera (Connell, 1928). Under the terms of the Rome prize, Connell was obliged to apply for a renewal of his scholarship for a third year. London-based secretary of the school, Evelyn Shaw (1882-1974), wrote that, “Connell’s case appears to be so strong that I am quite prepared to do my best to get it through even if it is necessary to ask Sassoon or someone else for the money” (Shaw, 1928). Shaw was referring to a private donor of the school, Sir Philip Sassoon (1888-1939), an MP and cousin of the poet Seigfried Sassoon. Connell worked on his major reconstruction during 1928, with particular Fig. 2 Fig. 2: Connell’s drawing of the Villa of Tiberius, Capri, prepared as part of his final year of study at the Rome School, 1928. British School at Rome Archive. INTERSTICES 09 53 the internal space. Each wing housed a separate function, and circulation for the Ashmole family was via the direct route of the central hall and stair while servants were obliged to take the long way around through a complex arrange- ment of corridors and secondary stairs. Despite the understandable tempta- tion to make the three-winged arrangement completely symmetrical, Connell could not contain the kitchen in the same wing as the dining room and this was pushed out into a single-storey projection that adds an interesting imbalance to the plan. Connell’s manipulation of the hexagon into a plan answering the Ash- moles’ functional needs was both sophisticated and true to pragmatic Roman practice (Fig. 4). While Connell’s post-Rome school landscape commission for Sir Philip Sassoon was constrained to formal English design tendencies by the client’s desire to remodel his house, Trent Park, in neo-Georgian style, Connell was highly con- scious of the natural landscape setting for High and Over and how features that related to the hexagonal geometry of the house might be added to it. Terraces were planned on two levels to manage the slope, lending a second level of geo- metric complexity to both house and garden. The terraces were to be linked to a pair of triangular flowerbeds and to other garden structures including pools and a pergola (Fig. 5). In this solid rooting of the house to the site and the elaboration of garden structures, Connell’s work differs markedly from that of Le Corbusier, who was more concerned with the standardization of his houses so that they could rest on piloti, with minimal contact with the ground and on any type of site. Connell was more focused on Mediterranean villas on their stony hillsides, where walled and terraced gardens extend architectural space into the environment. Connell’s interest in French modernist garden design is reflected in a 1934 draw- ing that presented a complete garden plan for the extensive grounds (Connell & Ward, 1934). This extends an earlier plan prepared with the design of the house and which was well under way by the time High and Over was publicized in 1931. The garden close to the house was based on two equilateral triangles divided into terraced beds, one on the gentle slope rising to the left of the entrance and the other filling the steeper grade down to the base of the chalk pit. These structures strongly recalled Guévrékian’s garden at the Paris Exposition as well as his better known courtyard at the Villa Noailles at Hyères (1927), one of the few surviving modernist gardens and thus well recognized today (Brown, the urbane Russian Serge Chermaeff (1900-96)mwhose romantic design ten- dencies mostly placed them outside the emergent branch of heavily theo- rized modernism. Architect and industrial designer Christian Barman (1898-1980) was also among Connell’s circle. The site chosen by the Ashmoles was part of the Shardeloes Estate in Amer- sham, a Buckinghamshire town at the northern end of London’s Metropolitan Line. The 12-acre site was elevated but bare of trees or other significant features apart from a hollow marking the position of a disused chalk pit. The decision was made to build on the level area above the pit and Connell began work late in 1928. The initial design for High and Over was later described by Ashmole as “an Elizabethan plan, E-shaped, the two projecting wings forming a court- yard which would be open to the south” (Ashmole, 1994: 51). This first proposal was turned down as it did not make effective use of the terrain and exposure to sun. No drawings survive of this scheme, but it would seem that the basic ele- ments were in place. The combination of hovering roof canopies, punched out slot windows and plain white walls epitomized the new architecture of the 1920s and it seems probable that these formal elements would have been present in the discarded first proposal. Ashmole’s account of the planning process has the Y-plan as his suggestion, taken up by Connell who understood the potential of the 120-degree arrangement of the main parts of the building around a centralized hexagonal hub (Ashmole, 1994: 52). Reference has been made to the Edwardian free plan and British houses built with modified Y-plans (Thistlewood & Heeley, 1997: 86). It is likely that Connell was aware of these but he was more inclined to draw directly from classical Roman sources, particularly the novel geometries of the Hadrianic period, a subject that he had mastered while in Rome. An appealing notion to the modernist planner in Connell was that Roman geometry was used when it brought tangible practical advantages. A hexagon has the useful feature of containing the greatest area within the shortest perimeter. It may also be multiplied and composed into a regular pattern or subdivided into equilateral triangles. Hexagonal room shapes were incorporated into Roman baths and one particular example of hexagonal planning, the Sala Trilo- bata, adjacent to the Baths of Trajan (Fig. 3), might be considered as a model for the planning of High and Over (Jones, 2000: 92). In this instance, rooms coming off the central space were constrained in size by the length of the straight-line segment forming the perimeter of the hexagon. This set of proportions would not be effective for a plan that demanded an appropriate scale relationship be- tween the entrance hall of a modern house and its major rooms. Connell instead expanded the hexagon to define the outer walls and drew a double square start- ing from the future fountain spigot, so that the spaces leading off were propor- tioned according to their function. The fountain has posed a critical dilemma for those striving to fit High and Over into the conventional narrative of modernism. Robert Esau surmises it “may be a deliberate evocation of some Roman or Italian prototype known only to Ashmole and Connell” (Esau, 1994: 61) and it is clear that water plays a particular role in the planning of the house and the garden. The hall floor was laid with black marble cut into triangles, and an illuminated circular glass fountain was set in the centre of a six-armed star. The larger hexagon was extended into an equilateral triangle that delineated the canopy over the front entrance and the shape of the stair tower. A circle exactly half the width of the hall was cut from the ceiling and allowed light from the stair tower to flood Fig. 3 Fig. 3: The geometry of Roman villas was a major influence on Connell’s radial plan for High and Over. The Sala Trilobata adjacent to the Baths of Trajan (AD 109) is an example of Roman planning based on the hexagon (Jones, 2000: 92). Fig. 4 Fig. 5 Fig. 4: The geometry of High and Over is based on a double-square rectangle rotated 120-degrees from the plan centre. Connel & Thompson (1929); regulating lines added by Michael Findlay. Fig. 5: Connell’s design for High and Over included a complex garden plan which was largely carried out by client Ashmole and garderner Marlow. The terraced rose garden was to be matched with another descending into the chalk pit but was left uncompleted. Architect and Building News, p. 431. INTERSTICES 09 55 minor figure in French modernism but his design philosophies on architecture and gardens were carefully studied in Britain during the late 1930s (Imbert, 1993: 185). In his 1929 publication Terrasses et Jardins, he argued against living on the damp and unhealthy earth and, instead, building towards the sky where the benefits of light and sun could be enjoyed from the roof terrace (Imbert, 1993: 188). The strongly socialist Lurçat believed that even the most modest house was incomplete without a garden and the landscape design of High and Over can be seen as Connell’s rejection of Corbusier’s tabula rasa. The height that is such a striking feature of High and Over accentuated the impression that the garden was essentially composed to be seen from above, from the top of the water tower or with the roof terrace of the main house as a horizontal façade, rid of the clut- ter of complex roof lines, chimneys and the other accoutrements of the country house (Fig. 7). Also constructed by 1931 were the lodge (lived in by George and May Marlow; Fig. 8), the water tower (Fig. 9) and a generator house. The lodge, unpublished as a plan but photographed for an extensive article in the Architect and Building News in 1931 (Robertson, 1931: 433), was a two-storey cubic design with corner slot windows and a striking angled wall that sheltered a private courtyard from the roadway. The generator house continued the theme with an engaged wall projecting towards the front of the site. This small cubic form was cut through with ventilation slots, and its abstract geometry effectively introduced the un- folding drama of the main structures, made more striking by the bareness of the hill on which they stood. High above the road stood an astounding object which resolved into a cylindrical water tank on a slender concrete pole combined with a pump house and racquet court. This highly irregular structure caused the local authorities to baulk but was forced on Ashmole due to the failure of the council to supply a water pipe sufficient to service the swimming pool (Ashmole, 1994: 55). Water emerges as crucial element in Connell’s planning for High and Over, with many parts of the complex scheme being able to be read against Guévré- kian’s garden. Connell’s landscape plan for High and Over also reveals his interest in the disciplined but modern designs of Paul Vera (1888-1971), who was concerned to bridge the gap between the modern world and the classical French paysager, with its geometrical beds and shaped plantings. According to Dorothée Imbert, Vera prescribed a “‘vernacular modernism’ that praised labour” (1993: 73), a propo- sition that certainly describes Connell’s plan and Ashmole’s physical efforts alongside George Marlow to reshape the surroundings of High and Over. Vera’s new formulation of the jardin régulier acknowledged labour, reflected in Connell’s raising of the water tower to its prominence in the landscape plan and the positioning of the vegetable garden between the formal axes of the pergola and the broad walk. Tradition was not abandoned. These various garden rooms were served by the pergola and lawns edged by low hedges of box, controlled shrubs and tall clipped green walls, as were the neo-classical gardens of the past. Concrete was used as a modern substitute for stone masonry, and Ashmole recounted that, “He [Marlow] and I together cast several thousand square feet of deep concrete edges for the beds, and several thousand square feet of cement paving for the paths” (Ashmole, 1994: 59). The extension of the interior plan of High and Over into this heavily constructed synthetic environment, occurring at various levels of abstraction, was intensely architectural in terms of space-time. Movement and physical activity of all types were celebrated. The fives court at 2000: 38). Connell was on the Riviera in 1928 when the house, designed by Rob Mallet-Stevens (1886-1945), was newly completed. The similar hilltop position and framing of the garden from the roof terrace make comparison between the two projects highly tempting. Both gardens are heavily delineated, using white concrete walls and borders to create strong contrasts between ground and beds. The original plantings at Hyères were similar to Connell’s plan for High and Over and used groupings of specimen plants rather than the pebbles that domi- nate the restored garden today. Connell’s perspective drawing shows clipped shrubs adding a regular spherical feature that punctuates the descent, a device also used by Guévrékian. Three flights of steps were to lead down to a small circular pool sheltered by a stand of trees. A long pergola extended the axis of the library wing while a maze filled in another triangular space above the rose garden. These features were linked to a rectangular lawn that ran on a shal- low 15-degree angle off the main axis towards the water tower at the top of the site, where a large rectangular pool was constructed. Between the lawn and the pergola was an extensive vegetable garden divided into rectangular beds. An arrow-shaped lawn intersected the main drive to the house where a fruit orchard dropped steeply into the chalk pit and was in turn bisected by a path laid out with rectangular sections of box hedge. Plants were chosen for both their practical and architectural form and the fruit trees and juvenile Irish yews planted by Ashmole are visible in photographs taken of the site in 1931. These elements were connected by a serpentine path that wound through the site and finished at the far end of the broad walk that was protected by a hedge. Connell placed these features sparingly on the large site and they were separated by large areas of roughly mown grass. Much of this ambitious scheme was achieved. A fascinat- ing aerial photograph taken in the mid-1930s shows Ashmole and the indefati- gable gardener George Marlow removing rocks and developing Connell’s ambi- tious landscape plan (Sharp, 1994: 26). In the end, Ashmole changed the position of the swimming pool to the foot of the stairs descending the chalk pit and did not proceed with the main terraced garden or the maze (Fig. 6). Connell’s use of roof terraces, classical vistas from the three-storey house itself and retaining walls to create exterior promenades is related to the landscape theory of French modernist Andre Lurçat (1894-1970), whose Maison Guggen- buhl at 14 rue Nansouty, Paris (1927), had influenced other details at High and Over, including the roof canopies and bay window. Lurçat is now regarded as a Fig. 6 Fig. 7 Fig. 6: Connell’s landscape plan for High and Over was developed in 1929. This plan from 1934 shows the intended landscape elements with four speculative houses on the High and Over Estate. RIBA Library. Fig. 7: The view from the High and Over roof terrace, including the Lodge, 1931. Architect and Building News, p. 429. INTERSTICES 09 57 been viewed as something marginal to the work he would produce in England, the impression being that he somehow had to forget his training and relearn the principles of modernist design. Connell’s five years of practical experience in New Zealand took place under a particular set of circumstances that smoothed the transition from Edwardian design values to Corbusian modernism. Unlike many of his English peers, Connell was informed about and experienced in reinforced concrete construction techniques, having worked on sites as a labour- er and as part of Fearn’s busy practice. Visiting present-day Eltham and view- ing the existing architectural fabric of the town leaves little doubt that Connell moved towards early maturity as a designer with a comprehensive understan- ding of the properties of concrete as a construction medium. When Connell saw Corbusier’s Domino system (1914), with its point-load foundations and cast-in floor beams, it cannot have seemed altogether foreign to him. It must have been galling for Connell that no firm in Britain could provide a satisfactory price for monolithic concrete construction on High and Over (Ashmole, 1994: 55) when he had seen buildings of this type successfully and economically built in rural New Zealand. It is also clear that Connell’s training and emergent approach to design favoured the axial symmetry of neo-classical planning that shaped other early British for- ays into modernism. Connell’s conception remained firmly rooted both in the ground and in classical architectural tradition. Connell was attempting modera- tion between the European avant-garde and the neo-classical English country house, with its cultural connections passing back through Andrea Palladio to the villas of ancient Rome. This dualism has engendered two misunderstandings that are now deeply imbedded in the interpretation of High and Over – firstly, that Connell’s effort to bridge the gap between modernism and classicism in the late 1920s was an ill-judged compromise, and secondly, that the design was, rela- tively speaking, a failure owing to its conservatism. Connell’s design certainly contains distinct modernist and conservative modes, but the piecemeal reading of the whole project in terms of disparate parts offers an incomplete and flawed understanding of Connell’s intention. Similarly, the segmentation of Connell’s career into what has been characterized as a juvenile period of experimentation progressing later into a more ‘respectable’ modernism misreads High and Over and the richness of its conception as a bridge between classicism and modern- ism. This reflects a rather well-worn cultural belief that colonials achieve success by ignoring the rules, a personal narrative which Connell and Ward later took up themselves. Connell’s status in architectural history is therefore marked by the simultaneous presence and absence of the expatriate from both cultural spaces. His formative experiences in rural Taranaki and Wellington were both repressive and rich with possibilities, ensuring that when he left the orbit of New Zealand, he did not re- turn. New Zealand was an ideal training ground for architects who would find their skills in demand in Britain, a fact borne out by the presence of so many ex- patriates during the 1920s and beyond. While most went simply to improve their prospects, some, like Connell, found the timing right to take the lead. the base of the water tower, regarded by some as a bizarre novelty, makes per- fect sense in this context. The climb to the brow of the hill 40 metres above the road was rewarded by the sight of the field, called the ‘velvet lawn’, across the Misbourne valley (Ashmole, 1994: 59). The most exciting vista was that from the top of the water tower, accessible by an internal ladder. This sight line took in the large circular pool that Ashmole and Marlow excavated at the foot of the garden steps where the family swam and exercised. This vantage point also reinforced the two-dimensional nature of the modern garden, designed to be appreciated from the roof terrace as a geometrical pattern on the land. The plan reveals both Connell’s observance of French trends in modernist garden design, the Beaux- Arts tradition of geometric planning and enfilade, and the English landscape tradition within the fragmented gestalt of early modernism, where both foreign and nationalistic influences were deployed together in the rush to experiment with the new architecture. In an effort to recoup some of the considerable expense of developing the estate, Ashmole approached Connell to develop a planned subdivision on the lower part of the site. This was a risky venture as the worst effects of the great depression were only just beginning to lift and the British house-buying public was scepti- cal about the new architecture. With Ashmole’s brother-in-law Charles de Peyer investing in the project, four houses were built on the slope leading up to the lodge (Connell & Ward, 1934). With the Sun houses complete in 1935, work con- tinued on the development of Connell’s landscape plan until Ashmole’s efforts were slowed by his other commitments. Marlow was injured in an accident on the property that left him an amputee and this unfortunate event also weighed heavily on Ashmole (Ashmole, 1994: 61). The maintenance demands of the house and garden became oppressive. After the Ashmoles sold the property, High and Over deteriorated under a succession of owners. The house was divided into two flats in the early 1960s. The garden was built over, although the circular swim- ming pool remains a feature of the remnant landscape. The question remains: why would Ashmole trust an untried architect with such an ambitious first project when he had the choice of any practising architect in Britain? Connell was somewhat reticent himself in pinpointing the reason for his client’s faith, but the combination of Connell’s deep understanding of classical architecture alongside the emergent themes of modernism as seen in France and Germany in the mid-1920s was surely compelling. Connell’s colonial pragma- tism also appealed to Ashmole, who was used to hard physical work and was no doubt reassured about his architect’s practical experience with an unconven- tional construction medium. Connell’s formative experience of architecture has Fig. 8: The Lodge, built for George and May Marlow, 1931. Architect and Building News, p. 435. Fig. 9: The water tower stood above the main house at the top of the rise. It housed a squash court and an elevated platform from where the garden was viewed. Architect and Building News, p. 435. Fig. 8 Fig. 9 INTERSTICES 09 59 Shaw, E. (1928). Letter to B. Ashmole (July 15). A. Connell File, BSR Archive. Shaw, E. (1929). Letter to A. Hamilton Smith (January 15). A. Connell File, BSR Archive. South Taranaki District Council (2000). Wilkinson’s Building, 150 High St, 1913-1914 (Item H1). Eltham Heritage Inventory. Hawera: s.p. Thomson, S. L. (1939). Form of Application for Admission as a Fellow. RIBA Library, Ref. 3673. Thistlewood, D., & Heeley, E. (1997). Connell, Ward and Lucas: Towards a Complex Critique. Journal of Architecture, 2(1), 83-102. Untitled (1921), New Zealand Building Progress, XVII(4), 86. Vitou, E. (2006). Guévrékian, Gabriel. Grove Art Online. Oxford University Press. Retrieved November, 23, from http://www.groveart.com/ Walker, C. (Ed.). (2005). Exquisite Apart: 100 Years of Architecture in New Zealand. Auckland: Balasoglou Books. Ward, B. (1967). Connell, Ward and Lucas. In D. Sharp (Ed.), Planning and Architecture: Essays Presented to Arthur Korn by the Architectural Association (pp. 73-86). London: Barry & Rocliffe. References Ashmole, B. (1994). Bernard Ashmole, 1894-1988: An Autobiography. Malibu: Getty Museum. Ashmole, P. (2005). E-mail to Michael Findlay (November 27). Bartle, R. (2006). The House that Jack Built: How One Designer Changed the Face of Hawera and the Dairy Industry Forever. Taranaki Stories: Business and Industry. Retrieved November 5, from http://pukeariki.com/en/stories/businessandindustry/default/asp Benton, C. (1999). Review of Jean-Louis Cohen, L’architecte André Lurçat. Journal of the Society of Architectural Historians, 58(1), 98-100. Brown, J. (2000). The Modern Garden. London: Thames & Hudson. Campbell, L. (1989). A Call To Order: The Rome Prize and Early Twentieth Century British Architecture. Architectural History, (32), 131-51. Connell, A. (1926). Telegram to Evelyn Shaw (November 14). Copy on B. R. Ward File, The British School at Rome (BSR) Archive, Rome. Connell, A. (1927). Application for Renewal of Scholarship at The British School at Rome. A. Connell File, BSR Archive. Connell, A. (1928). Application for Renewal of Scholarship at The British School at Rome. A. Connell File, BSR Archive. Connell, A. (1929). Letter to A. Hamilton Smith (December 29). A. Connell File, BSR Archive. Connell & Thomson (1929). Proposed House at Amersham, Buckinghamshire. Plans and elevations. RIBA Library, London, Ref. ON3121. Connell & Ward (1934). High and Over Estate, Station Road, Amersham, Bucks. Proposed layout of first section. Plan. RIBA Library, Ref. ON3122. Connell, J. (1985). Pers. comm. to M. Findlay (June). Esau, R. (1994). Connell, Ward and Lucas and the Emergence of the British Modern Move- ment in Architecture. Unpublished Dissertation for the Degree of Doctor of Philosophy, Bryn Mawr College, Philadelphia. Giovenco, A. (2005). Pers. comm. to M. Findlay (November 12). Imbert, D. (1993). The Modernist Garden in France. New Haven: Yale University Press. Item 54, NZIA Minute Book Articles 1916-1940, p. 8. NZIA Archives, J. C. Beaglehole Room, Victoria University of Wellington. Item 57, NZIA Minute Book Articles 1916-40, p. 65. NZIA Archives. Jones, M. W. (2000). Principles of Roman Architecture. New Haven: Yale University Press. New Zealand Institute of Architects Wellington Branch Competition for War Memorial (1921). New Zealand Building Progress, XVII(1), 12-13. News and Topics (1926). Architects’ Journal, LXIII(June 30), 872. Notes (1922). New Zealand Institute of Architects Journal, 1(3), 75. Personal (1926, November 18). Otago Daily Times, p. 10. Robertson, H. (1931). Amoenitas: ‘High and Over’ at Amersham. Architect A. D. Connell. Architect and Building News, CXXVI(June 26), 428-35. Sharp, D. (Ed.) (1994). Connell, Ward and Lucas. London: Book Art. Sharples, J., Powers, A., & Shippobottom, M. (Eds.). (1996). Charles Reilly & The Liverpool School of Architecture, 1904-1933. Liverpool: Liverpool University Press. Shaw, E. (1925). Letter to A. Connell. A. Connell File, BSR Archive. Shaw, E. (1926). Letter to A Connell (June 29). A. Connell File, BSR Archive. Shaw, E. (1927). Letter to unknown recipient (September 19). A. Connell File, BSR Archive, Ref. H27.