[Chicago Skyscrapers, 1934-1986, published by University of Illinois Press, is out now–available on Bookshop.org and Amazon.com, among other outlets. This week’s entry follows up on last week’s Skyscraper Museum webinar conversation with SOM Principal Emeritus Bill Baker about that firm’s history of concrete high rises and tube structures].
“The right means to the right ends must be found; i.e., the means must be in scale with the ends, and a philosophic base must be used to judge the relationship of structure, scale, end architecture.”
–Myron Goldsmith, “The Effects of Scale,” 1953-1987.[i]
Myron Goldsmith’s 1953 IIT Master’s thesis on tall building structures argued that scale was critical to structural systems. He began with Galileo’s physiological example: if a bone is enlarged proportionally, its cross-sectional area—and thus bearing capacity—increases exponentially by a factor of two but its volume—determinant of the bone’s self-weight and thus of the loads it carries—increases by a factor of three. A structure can only be scaled up so far, Galileo realized, before it fails under its own weight—or, as Goldsmith quoted biologist D’Arcy Wentworth Thompson: “Elephant and hippopotamus have grown clumsy as well as big.”
Goldsmith recognized this principle in building structures, particularly masonry ones, where bearing walls’ self-weight produced structurally clumsy equivalents—the massive walls of the Monadnock’s lower stories, for example. Similar limits exist for all systems and materials. In high rises, lateral resistance to wind limits simple frames to around 25 stories—lower-story columns become ungainly in buildings taller than this as shear forces build up, and shear walls or trusses are needed to maintain reasonable column sizes. Taller buildings, Goldsmith argued, require structural changes that are not simply quantitative—more cross-section or stronger materials. Instead, they require qualitative changes in the configuration of the structure itself. New systems came with architectural potential in addition to optimization. “A new structural system,” he wrote, “gives the possibility of a new aesthetic expression.”
“Engineering for efficiency is not the last and only determinant; it is possible to make a choice from several efficient schemes because of architectural, aesthetic, and environmental reasons. The human needs must give the directions.”[ii]
Goldsmith expanded “The Effects of Scale” in his 1987 monograph, Buildings and Concepts, to include urban and environmental impacts. The efficient scaling up of oil tankers in the 1970s, for instance, made more optimal vessels but brought greater risk of catastrophe, and denser and larger cities required more complex systems—circulatory, infrastructural, and economic, among others. Goldsmith’s thesis, done with Mies as advisor, formed the prototype for IIT’s M.S. program during the 1960s, much of which he and Fazlur Khan supervised along with David Sharpe, a graduate of Tuskegee’s undergraduate program who joined SOM and the IIT faculty. IIT’s Master’s program became a fertile think-tank for SOM, producing case studies of structural and architectural integration that often found their way to drawing boards downtown and vice versa.[iii]
Chestnut-DeWitt (SOM, 1962-65)
Goldsmith’s thesis saw striking application in the 43-story Chestnut-DeWitt Apartments, designed in parallel with the 37-story Brunswick. Brunswick’s wind-resisting synthesis of closely spaced perimeter columns, stiff spandrel beams, core shear walls, and linking floor slabs spurred Graham, Goldsmith and Khan’s growing interest in the ‘rebirth of the bearing wall.’ But Chestnut-DeWitt’s unique site circumstances, coupled with differences between commercial and residential programs, made for a subtle reconsideration of the Brunswick’s principle into a new structural type—the tube.
Chestnut-DeWitt’s site was an L-shaped Streeterville lot adjacent to Mies’ 860-880 Lake Shore Drive. Metropolitan Structures, re-organized after Herbert Greenwald’s death in 1959, owned the site and asked SOM to design a third pair of towers in their Lake Shore Drive cluster. Graham recalled being concerned about views to and from the iconic 860-880 and realizing that another pair of towers would crowd them.[iv] He suggested stacking Metropolitan’s program into one taller tower, leaving half the site for a low garage pavilion. This would leave breathing space around Mies’ towers, but Graham’s massing brought structural issues. Commercial floor plates like the Brunswick’s could accommodate deep shear walls or trusses in their large cores to resist wind. Alternatively, they could rely on stiff joints throughout deep, repetitive rigid frames. In apartments, however, the desire—and, in Chicago, code requirement—for natural light and ventilation in living areas produced longer, shallower building masses that were weak across their short axes. Shear walls deployed between units or at building ends could contribute lateral resistance, but apartment buildings lacked the multiple elevator shafts and condensed plumbing cores convenient for effective shear walls or wind trusses in commercial towers. Worse, residential buildings did not have the interstitial mechanical requirements that made deep, moment-resisting girders viable. Instead, concrete slab construction, thinned by the economic advantages of reducing floor-to-floor height, left little sectional area for the deep moment connections that could brace commercial towers.
Graham planned apartment layouts around short, hammer-headed corridors, wrapping bedrooms and living rooms around tightly planned bathrooms, closets, and entries into floor plates of 122’ x 78’. Multiplying this to fill Metropolitan’s program required a tower more than 40 stories tall, slender enough to require significant bracing in both directions. Exterior shear walls were out of the question given the lake and city views in all directions. Hal Iyengar and Khan instead tried to develop a central core out of fire stairs and elevator shafts, at one point suggesting twin shear walls containing these elements parallel to the building’s longer axis. Stair and elevator openings, however, frustrated these attempts.
The structural design for Brunswick, meanwhile, was developing a few months ahead of Dewitt-Chestnut’s. Its rigid but porous external bearing walls showed that a shear wall’s rigidity could be distilled into a network of moment joints around large window apertures. Brunswick’s large central core made it only partly reliant on this external frame–the tall, open lobby and large transfer girder made its core shear walls critical to the building’s stability. Khan wondered whether, with more robust connections, skyscrapers’ exterior walls themselves could provide such a slender structure’s lateral stability. Doing so required a compromise between exterior member sizes and desirable views. But the Brunswick’s upturned perimeter beams held a clue. The one place in an apartment where structural elements could intrude into the expected 8’ floor-to-ceiling height was at the exterior, where sills and air conditioner cabinets reduced window apertures anyway. Upstand beams here could provide the deep column connections necessary to create stiff moment joints in the exterior wall. Khan also realized that exterior columns could be spaced more closely together in a residential program, performing double duty as structural mullions, and forming more, narrower windows. Doubling or tripling perimeter columns meant more connections and, thus, greater overall stiffness. At some point, the distinction between a skeleton of columns and beams blurred into structurally solid walls pierced with window openings that could work as a giant, tubular cantilever beam sticking out of the ground. The resulting shape was an imperfect beam (with two webs instead of the I-beam’s one), an imperfect shear wall (perforated with dozens of window openings), and an imperfect architectural solution (window walls interrupted by columns larger than mullions)—but taken together these individual elements formed an efficient overall structure.
Thinking about the entire building as a cantilever was a paradigm shift. Hand calculations were limited to tracing loads through a structure, looking at individual elements’ capacities to resist loads and deflection. Such an elemental approach, engineers knew, provided conservative results—studies on the 55-story 1000 Lake Shore Plaza showed that its shear wall and column structure deflected only 37% as far its designers had calculated due to wind.[v] This may have been reassuring, but it was a waste of materials. Khan’s sense of the building structure as a holistic—almost organic—system marked a new approach. Understanding the flow of forces through a monolithic network required more computing power than hand calculation could provide. But the redundancies that made such structures difficult to calculate also made them efficient— ‘hyperstatic,’ dispersing forces throughout building frames in multiple, simultaneous load paths, in this case through a “shear shell” or “tube.”[vi]
Concentrating the tower’s structure on its exterior allowed more efficient unit layouts, too. Seventeen interior columns, taking gravity loads only and located based on apartment layouts rather than a structural grid, reduced spans, taking advantage of flat plate systems’ adaptability to irregular column placement. One important problem developed as Khan and Iyengar began using SOM’s new mainframe computer to analyze the structure. As the tube walls collected wind loads on their faces they would flex, lacking the backup of the Brunswick’s shear walls. As they did so, they would transfer loads to the side walls—the ‘webs’ of the cantilevered beam—only gradually, meaning that the wind-facing wall’s center would deflect farther than its ends, a phenomenon Khan called “shear lag.” The end walls would, essentially, be dragged along, causing unanticipated stresses in the corner columns. In conventional frame construction, corner columns were the least loaded since they carried only ¼ of the floor area of an interior column. But for tube structures, the team now recognized, corner columns became highly stressed elements, demanding larger sections—validating, at least in this case, the classical rules championed by Mies that doubled columns up when turning a corner.
Graham adapted DeWitt-Chestnut’s exteriors to Khan’s structural scheme. Its perimeter columns are collected by a transfer girder at the second floor, as at Brunswick, but here only at every other column, leaving 11’-0” openings at ground level for a colonnade. Like Brunswick, Graham selected travertine to clad the raw structural frame, although here it was actually installed along with a layer of rigid foam, to forestall thermal expansion and contraction. At the corners, a re-entrant detail allowed greater column depth in both directions, accommodating the shear lag stresses while providing visual emphasis. In a subtle expression of its wind bracing theory, the tower’s structural elements all become thinner as the building rises; lateral shear and bending increase toward the base, allowing the structure to be far more flexible toward the top. As the columns and edge beams thin, from 1’-11” at the base to 1’-2” at the roof, DeWitt Chestnut’s windows grow, from 3’-7” to 4’-4”.
Metropolitan Structures secured an $8 million mortgage for this innovative structure from Aetna Life through Draper and Kramer. The project received FHA support even though it was aimed at the upper middle-class market. Federal funding meant that DeWitt-Chestnut was, along with Marina City, Sandburg Village, and Outer Drive East, open to any qualified applicant regardless of race,—still unusual enough that these projects were lauded by the Chicago Commission on Human Relations.[vii] Metropolitan’s construction subsidiary began work on site in August, 1963 and the building opened to tenants in February, 1965, with rents ranging from $140 for studios up to $410 for three-bedroom units ($1200 to $3475), 30% higher than Marina City but comparable to Outer Drive East, in keeping with Metropolitan’s professional, rather than executive, market.[viii]
DeWitt-Chestnut proved a deferential contrasting backdrop to the Mies buildings, but for engineers and critics who understood this elegant, quiet block’s structural innovation and nuanced expression, it was a qualitative leap in performance, based on principles first explored by Komendant and Pei and furthered by SOM at Brunswick, but honed into a distinct, new species of skyscraper structure. Komendant, Khan and New York’s Leslie Robertson all made steps toward the pure tube structure, but credit for SOM’s team in developing the first pure tube skyscraper here—one that relied entirely on its exterior for its lateral stability—was justified. Khan’s systemic approach turned engineers from calculators to designers. His work with Graham over the next decade fused static, programmatic, and architectural form, setting height records along with high standards for integrated engineering and structural aesthetics. DeWitt-Chestnut, on that point, was more than a technical success. Architectural Record called it “one of the most sophisticated and disciplined of SOM’s sophisticated and disciplined designs.”[ix] And, if it was a deliberate visual contrast to Mies’ incomparable towers to the east it was also, according to Iyengar, a link between Khan’s structural philosophy and 860-880’s principles:
“Mies’ buildings were still framed buildings. He was still mostly concerned with expressing the frame. He didn’t get beyond that. But, his principle though, his notion of the structure having a prominence in architecture could be seen all the way through…. As long as the structures play a dominant role, creates the essence of architecture, then it becomes Miesian.”[x]
[i] Myron Goldsmith, “The Effects of Scale” in Myron Goldsmith-Buildings and Concepts. (New York: Rizzoli, 1987). 8-23.
[ii] Goldsmith, “The Effects of Scale,” op. cit. 22.
[iii] On David Sharpe’s career, see Dahna M Chandler, “Scaling the Heights of Architectural Academe.” Black Issues in Higher Education, vol. 16, no. 23, Jan. 6, 2000. 24 and Robert Lau, “The Legacy of David C. Sharpe.” CTBUH Journal, 3. 2010. 40-43. See, too, Lizondo-Sevilla, L., S.-F. José, G.-R. Zaida. “Mies and His Teaching Venues: The Triumph of Architecture over Function”. ACE: Architecture, City and Environment, Vol. 15, no. 45, Feb., 2021.
[iv] The following relies heavily on the excellent descriptions of the building’s design process in Yasmin Sabina Khan, Engineering Architecture: The Vision of Fazlur R. Khan (New York: Norton, 2004). 84-103 and Mir M. Ali, Art of the Skyscraper: The Genius of Fazlur Khan. (New York: Rizzoli, 2001). 43-44, 86.
[v] “Winds Post Challenge for Skyscraper Builders.” Chicago Tribune, June 23, 1968. D1.
[vi] An authoritative overview of tube principles is Fazlur R. Khan, Ph.D., “Tubular Structures for Tall Buildings” in Mark Fintel, ed., Handbook of Concrete Engineering (New York; Van Nostrand Reinhold, 1974). 345-354. See, too, the excellent overview of the tube concept in Robert E. Fischer, “Optimizing the Structure of the Skyscraper.” Architectural Record, Vol. 152, no. 4. October 1972. 97-104.
[vii] “Open Housing Increasing on Near North Side.” Chicago Tribune, Dec. 29, 1963. N2.
[viii] Display Ad, The Dewitt Apartments. Chicago Tribune, Feb. 14, 1965. C9.
[ix] “Sheer Tower in Chicago. [DeWitt-Chestnut]” Architectural Record, Vol. 139, no. 1. January, 1966. 161.
[x] Betty J. Blum, Oral History of Srinivasa (Hal) Iyengar. (Chicago: Art Institute of Chicago, 2008). 27.
architecturefarm 