Once Wright had decided the location of the house, he had the obvious problem of building it there. The location of the north bank of Bear Run was not large enough to provide a foundation for a typically built Wright house.
Beyond this issue, there were also the clients’ needs that had to be met. The Kaufmanns planned to entertain large groups of people, so the house would need to be larger than the plot allowed. Also, Mr. and Mrs. Kaufmann requested separate bedrooms as well as a bedroom for their adult son and an additional guest room.
Wright’s solution to the problem of space came when he decided on a cantilevered structure. The structural design for Fallingwater was undertaken by Wright in association with staff engineers Mendel Glickman and William Wesley Peters, who had been responsible for the columns featured in Wright’s revolutionary design for the Johnson Wax Headquarters.
Preliminary plans were issued to Kaufmann for approval on October 15, 1935, after which Wright made a further visit to the site and provided a cost estimate for the job. In December 1935 an old rock quarry was reopened to the west of the site to provide the stones needed for the house’s walls. Wright only made periodic visits during construction, instead assigning his apprentice Robert Mosher as his permanent on-site representative. The final working drawings were issued by Wright in March 1936 with work beginning on the bridge and main house in April 1936.
The strong horizontal and vertical lines are a distinctive feature of Fallingwater
The construction was plagued by conflicts between Wright, Kaufmann, and the construction contractor. Uncomfortable with what he saw as Wright’s insufficient experience using reinforced concrete, Kaufmann had the architect’s daring cantilever design reviewed by a firm of consulting engineers. Upon receiving their report, Wright took offense and immediately requested Kaufmann to return his drawings and indicated he was withdrawing from the project. Kaufmann relented to Wright’s gambit and the engineer’s report was subsequently buried within a stone wall of the house.
After a visit to the site in June 1936, Wright rejected the stonemasonry of the bridge, which had to be rebuilt.
For the cantilevered floors, Wright and his team used upside down T-shaped beams integrated into a monolithic concrete slab which both formed the ceiling of the space below and provided resistance against compression. The contractor, Walter Hall, also an engineer, produced independent computations and argued for increasing the reinforcing steel in the first floor’s slab. Wright refused the suggestion. While some sources state that it was the contractor who quietly doubled the amount of reinforcement,according to others, it was at Kaufmann’s request that his consulting engineers redrew Wright’s reinforcing drawings and doubled the amount of steel specified by Wright.
In addition, the contractor did not build in a slight upward incline in the formwork for the cantilever to compensate for the settling and deflection of the cantilever. Once the concrete formwork was removed, the cantilever developed a noticeable sag. Upon learning of the steel addition without his approval, Wright recalled Mosher.
With Kaufmann’s approval, the consulting engineers arranged for the contractor to install a supporting wall under the main supporting beam for the west terrace. When Wright discovered it on a site visit he had Mosher discreetly remove the top course of stones. When Kaufmann later confessed to what had been done, Wright showed him what Mosher had done and pointed out that the cantilever had held up for the past month under test loads without the wall’s support.
In October 1937, the main house was completed.