Fallingwater Case Study: Site, Cantilevers and Plans (Frank Lloyd Wright)
A Fallingwater case study for students: the brief, the site over Bear Run, the plan around the boulder hearth, the concrete trays, why the cantilevers sagged and how they were repaired.
Published on 11/11/202614 min read
Fallingwater is a weekend house Frank Lloyd Wright designed in 1935 for the Kaufmann family of Pittsburgh, built from 1936 to 1937 over a waterfall on Bear Run, Pennsylvania. It teaches three things: reading a site, planning around one fixed point, and what happens when a daring structure meets ordinary construction. Below: brief, site, plan, section, materials, climate, and the 2001 to 2002 cantilever repair.
Fallingwater at a glance
Architect: Frank Lloyd Wright (1867 to 1959), with his engineers William Wesley Peters and Mendel Glickman. The client also hired engineers, Metzger-Richardson Co.
Client: Edgar J. and Liliane Kaufmann, owners of Kaufmann's Department Store, Pittsburgh.
Location: Bear Run, Mill Run, Fayette County, Pennsylvania, USA, about 113 km (70 miles) southeast of Pittsburgh.
Dates: drawn September 1935; built April 1936 to December 1937; guest house 1939.
Type: weekend house, now a museum: a living room, three bedrooms and service rooms.
Size: no source we opened gives a floor area. The Conservancy says terrace area almost equals indoor area.
Status: given to the Western Pennsylvania Conservancy in 1963 (the 1976 landmark nomination says October 1969), open since 1964, US National Historic Landmark, UNESCO World Heritage since 2019 within "The 20th-Century Architecture of Frank Lloyd Wright".
Kaufmann's employees' association bought 1,600 acres (about 650 ha) at Bear Run in 1926, and the family took over the core tract in 1933 (timeline; nomination). Their son, Edgar Kaufmann Jr., joined Wright's Taliesin Fellowship in autumn 1934. His parents met Wright there that November and asked for a weekend house near a waterfall on their land (The Kaufmann Family). The family loved the falls, and Wright put the house over the most dramatic of them rather than across from them (Designing Fallingwater).
Wright first saw the site on 18 December 1934. The apprentice Edgar Tafel's account, summarised in Theodore Ceraldi's 1999 ACSA paper, has Wright drawing the house on the site plot plan in September 1935, in the roughly two and a half hours before Kaufmann arrived at Taliesin. Working drawings followed in January 1936.
Fallingwater site analysis
The facts that drive the design, from the 1976 nomination unless noted:
Position: on the north bank of Bear Run, above the upper waterfall, in mature woodland.
Orientation: facing about 150 degrees, roughly south-southeast, over the stream.
Back edge: a rock cliff about 4.9 m (16 ft) high. The driveway runs between cliff and house, and you enter from this north side through a loggia.
Rock: a 1938 Museum of Modern Art press release says the bedrock is the foundation, an anchor for beams and the hearth of the fireplace.
Water: the stream forms a pool partly under the living-room terrace and leaves as a waterfall at the other end (MoMA).
Instead of a house across the stream looking at the falls, Wright pinned it to the boulders at their edge and pushed the floors out over the water. After his first visit he wrote to Kaufmann that the house was taking shape in his mind "to the music of the stream" (December 1934 letter, quoted by Ceraldi). For setting up a survey and analysis sheets like this, see our guide to documenting a building for a case study.
The view from below the falls shows the idea at once: horizontal concrete terraces held out from a vertical stone core over the stream. Photo: Carol M. Highsmith, Public domain, via Wikimedia Commons
Concept and design ideas
A stone core, free concrete planes. A vertical mass of local sandstone holds the chimney and stairs; concrete floors and terraces reach out from it at different lengths.
Inside and outside as one surface. Edgar Kaufmann Jr., quoted in the nomination, said Wright insisted on floors at one level in the same stone on both sides of the glass, and continuous ceilings. The terraces are the rooms continued.
The site as structure. The boulder is the hearth and the bedrock carries the piers. The Conservancy presents the house as the clearest example of Wright's organic architecture, the union of art and nature (World Heritage page).
Fallingwater floor plan and zoning
Edgar Kaufmann Jr. explained the plan in one sentence, quoted in the nomination: Wright set the chimney mass on a large boulder whose top became the hearth, and swung the spaces around it. The living room goes south over the falls, the entrance spaces east, the kitchen west, and a narrow strip for stairs and dining north. Four low stone piers mark the centre bay of the living room.
Wright placed the hearth on the boulder and arranged the rooms around it, with the living room pushed south over the falls. Diagram: MOA
Living level: one large living room with dining, a glazed hatch with steps down to the stream, the kitchen and a small entry; steel and glass doors to east and west terraces.
Second level: master bedroom and terrace over the living room, Kaufmann's dressing room over the kitchen with a west terrace, and a guest bedroom.
Third level: a bedroom (later a library) and a gallery opening onto a terrace.
Guest wing, 1939: bedroom, lounge and pool on the hill above, servants' rooms over a carport, reached by a curved walkway under a concrete canopy.
For your sheet: show the zoning, the hearth as the fixed point and the terrace-to-room ratio. Our Farnsworth House case study, coming in this series, makes a good pair.
Where to see the real plans: Wright's originals are at Columbia University's Avery Library, viewable there or through JSTOR, and the Library of Congress holds free as-built drawings from the Historic American Buildings Survey. The Conservancy's architectural drawings page explains both; the archive is under the Frank Lloyd Wright Foundation's copyright. Take any dimension you draw from those, never from a photograph.
Fallingwater section and the cantilevers
People read the house as stone walls with concrete sticking out. Edgar Kaufmann Jr. called that wrong: Wright conceived a series of horizontal concrete trays with upturned stiffening edges, spaced apart by short stone masses. Each tray is structural on its underside; above it an air space carries the stone-flagged floor (nomination). The upturned edges, the parapets you see from outside, act as beams.
The trays carry the house; when their edges sagged, cables tightened along the main beams in 2001 to 2002 stopped the movement. Diagram: MOA
The living-room tray sits on bolsters on the bedrock and cantilevers south on four main beams. Ceraldi, summarising Robert Silman Associates' analysis, explains that the master terrace above was never designed as a true cantilever on its own: it leans partly on steel T-sections in the living-room window, which pass load down to those four beams. The two levels work as a pair. The house rises more than about 9 m (30 ft) above the falls (Conservancy). Wright himself wrote in the January 1938 Architectural Forum (quoted by MoMA) that next time "parapets will carry the floors", or the two would work as one.
What went wrong in 1936
From Ceraldi's account, built on Silman's analysis and the archive letters:
Two sets of engineers. In August 1936 Kaufmann's engineers had twice the number of 25 mm (1 in) square bars placed in the living-room cantilever beams. Wright protested in angry letters before the two made up.
No camber. Formwork was set dead level, with no upward crown to offset the first deflection.
Heavier than drawn. The master terrace slab was cast thicker than designed.
Early loading. The contractor, Walter Hall, built a storage shanty at the tip of the master terrace before its concrete reached 28-day strength.
Metzger-Richardson reported structural cracks around 1 December 1936 in cantilever beams and the master terrace parapets (Ceraldi), and John Matteo's 2025 summary dates the start of monitoring to that month. Kaufmann had a surveyor measure the terraces from 1941 until his death in 1955 (Western Pennsylvania Conservancy, 2013).
Fallingwater restoration: the repair of 2001 to 2002
The sag kept growing: about 100 mm (4 in) by 1955, and nearly 180 mm (7 in) in a University of Virginia thesis study, dated 1994 in the Conservancy's preservation history and 1995 in its 2013 article. Robert Silman Associates started work in 1995. In Scientific American (September 2000) Silman reported the west terrace edge down by up to 146 mm, the east terrace edge by up to 184 mm and the master terrace's south end by about 114 mm, with monitoring over a year and a half showing the cracks still growing. (We read these figures on the article page reproduced in Matteo's slides.)
1997: temporary steel shoring went in under the living-room cantilever, down to the stream bed (Ceraldi; Matteo).
2001 to 2002: 557 floor flagstones were numbered and lifted. Steel cables were placed alongside three of the four main living-room beams, anchored in concrete blocks and the piers, and jacked tight. The east and west terrace joists were post-tensioned, and carbon fibre strengthened the master terrace and Kaufmann Sr.'s terrace (preservation history; Conservancy, 2013; Bob Silman in memoriam).
Result: an April 2002 Associated Press report (via the Lawrence Journal-World) said the jacks lifted the structure about 13 mm (half an inch) off the shoring and the cracks closed. The house was left slightly tilted, because forcing it level could have broken windows. A decade on, the Conservancy measured movement of about 0.25 mm (1/100 in).
Where sources disagree: the Conservancy timeline dates the cables to 2002, Matteo to 2001. Matteo lists west terrace carbon fibre in 2004, not 2001 to 2002. Sag figures vary by date, place and who measured, from about 100 mm in 1955 to about 150 mm (6 in) in the AP report and 184 mm in Silman's survey.
Cracks reopened in the piers in 2011 and the terraces in 2012, leading to a year of monitoring (Conservancy, 2013). From 2023 to March 2026 the World Heritage Preserved project replaced roofs, re-waterproofed terraces by lifting and relaying the flags, and injected nearly 12 tons of grout into hollow voids in the main house's stone walls, through which water had long reached the rooms.
The structural story starts within months of the first pour and has never fully stopped: the house needs engineering care as well as housekeeping. Diagram: MOA
Fallingwater materials and construction
Stone: walls, piers and chimney of native sandstone (nomination), locally quarried Pottsville sandstone (Conservancy), quarried on site from 1935 (Matteo). Floors inside and out are 50 mm (2 in) sandstone flags.
Concrete: cantilevered reinforced concrete slabs, painted the light ochre Wright specified in 1937 and repainted at least six times by 1959 (preservation history).
Steel and glass: steel window and door sash in the red the Conservancy calls Cherokee red; at the corners, glass meets glass with no post.
Timber and cork: black walnut plywood millwork and mostly built-in furniture designed by Wright (nomination); cork-lined bathrooms (MoMA).
Light, climate and services
Glass under deep overhangs: MoMA noted in 1938 that the living room is almost all glass on three sides, yet the thin concrete overhangs keep out glare.
Low ceilings and a big hearth: the Conservancy describes the low ceilings as part of the house's sense of shelter.
Built-in light: the nomination notes fluorescent lighting built into the ceilings, used to imitate daylight or as indirect light.
No air conditioning, even today, according to the visit page.
The weak side is water. The Conservancy's notes list failing roof and terrace waterproofing, ponding, water drawn into stone joints, freeze-thaw damage and cork stained where leaks persist.
Reception: praise, cracks and leaks
MoMA opened a small exhibition of twenty enlarged photographs and plans on 25 January 1938. The 1976 nomination records Wayne Andrews calling it the world's most famous modern house, and Vincent Scully placing it among the finest works of twentieth-century art. In 1964, its first public year, it drew 29,792 visitors (timeline). For its owners it has meant decades of surveys and repairs.
What designers in India can take from it
Our reading, for practice in Indian conditions:
A cantilever is decided on site as much as on paper. Level formwork, a thicker slab and a shed on green concrete hurt Fallingwater as much as the design did. IS 456:2000, clause 11.3, says forms should not come off until concrete reaches at least twice the stress it will then carry. For ordinary Portland cement at 15°C or more it suggests 14 days for props under slabs spanning over 4.5 m and 21 days under beams over 6 m, and the props left must carry the full dead load plus construction loads. Agree camber and stripping sequence in writing with your structural engineer, and keep material stacks off fresh slabs.
One engineer of record. In 1936 extra steel went in without the designer agreeing. Keep one structural engineer responsible and put every change on revised drawings.
Plan the water path. Indian flat RCC terraces, stone-clad walls and hill houses meet the same problems in the monsoon: falls to drains, a membrane under stone floors, drips at edges, and solid, well-filled walls. See our guides to waterproofing by climate and building in Himachal and Uttarakhand hill towns.
Build on the rock you have. Stone quarried on site and a hearth on an existing boulder tie the house to its ground. On rocky Indian plots, map outcrops and boulders in the survey before planning, and have a geotechnical engineer confirm any rock you build on.
The Conservancy runs Fallingwater at 1491 Mill Run Road, Mill Run, Pennsylvania. Its visit page lists a guided house tour, a longer in-depth tour that allows personal photography, and a grounds-only pass. In 2026 house tours run from 14 March to 31 December, booked ahead, and the site closes on Wednesdays. Children under 6 cannot take the guided house tour. Expect a 400 m (quarter mile) gravel walk, more than 100 steps inside and no public transport. Check the official page before you plan.
Wright designed it in 1935 after his first visit in December 1934. Construction began in April 1936 and the Kaufmanns moved in in December 1937. The guest house followed in 1939.
Who designed Fallingwater, and for whom?
Frank Lloyd Wright, with his engineers William Wesley Peters and Mendel Glickman, for Edgar J. and Liliane Kaufmann of Pittsburgh. Their son Edgar Kaufmann Jr. gave it to the Western Pennsylvania Conservancy in 1963.
What materials is Fallingwater made of?
Local sandstone for walls, piers and chimney; reinforced concrete for the cantilevered floors and terraces; steel sash; sandstone flag floors; black walnut plywood built-ins; and cork in the bathrooms.
Why did Fallingwater's cantilevers sag, and is it safe now?
Engineers point to overstressed, under-reinforced beams, formwork without camber, a slab thicker than drawn and early loading during construction. In 2001 to 2002 cables were post-tensioned along the main living-room beams and other terraces were strengthened. The house was not levelled; movement since has been very small, and it is monitored regularly.
Where can I find Fallingwater's floor plan and section drawings?
Wright's originals are at Columbia University's Avery Library, in person or through JSTOR. As-built drawings from the Historic American Buildings Survey are free at the Library of Congress. The Conservancy's architectural drawings page explains how to reach both.
Sources
Fallingwater (Western Pennsylvania Conservancy): Timeline (accessed 3 Nov 2026)