Skip to content

Unsung Hero: Yoshikatsu Tsuboi

Post date:

Yoshikatsu Tsuboi

Yoshikatsu Tsuboi (1908 – 1980)

Many of these images in this post were taken from the book Kenzō TangeArchitecture for the World, Edited by Seng Kuan and Yukio Lippit, in cooperation with Harvard University Graduate School of Design. For me, the most interesting chapter was the one written by Mamoru Kawaguchi, project leader for Yoshikatsu Tsuboi, Chief Engineer for the Yoyogi National Olympic Stadiums.

The collaboration between Kenzo Tange and Yoshikatsu Tsuboi began in 1951 with the Hiroshima Children’s Library which was the first reinforced concrete shell structure to be built in Japan after the war.

Hiroshima Children's Library, Hiroshima, Japan.

Tsuboi rationalised the structure to a cylinder, a partial toriod and a cone.

Their next collaboaration was Sunpu Kaikan, in Shizuoka City, 1957.


Untitled 5

Their most significant collaboration however, was for the Yoyogi National Stadiums, built for the 1964 Tokyo Summer Olympics. The large stadium was a 15,000 person stadium for the swimming events and the smaller stadium was a 4,000 stadium for basketball.


This post is about what is possible when architects and engineers collaborate, and not about structure per-se. Here’s a structural account Kanzo Tange wrote shortly after the completion of the stadiums and which was published in Domus in 1965. The images in that article are themselves reprinted from a Kenchiku Bunka article of January 1965.

Although steel had been in short supply after the war, an increase in labour costs was making concrete shell structures prohibitively expensive. Architects had been studying the possibilities of suspension structures since Mattherew Nowicki architects and engineers’ Dorton Arena (1952).

The next building to use this technology was Eero Saarinen’s Ingalls Ice Hockey Rink that opened in 1958.

Saarinen's Ingalls Ice Hockey Rink

Tange asked Tsuboi to explore the possibility of using such a structure. Tange’s intuition was that such a structure would be more economical, reduce the internal volume and thus the heating load and also simplify the acoustic treatments. A true architect, his main reason was to realise “open spaces” in which occupants can move without physical obstruction or psychological oppression.


Here’s the basic structural system for the large stadium.

Untitled 6

Kawaguchi recalls that the finite analysis calculations were performed using a hand-held calculator. If so, it must have been one of the first prototypes. The numbers in the diagram below are from a more recent computer verification and are in general agreement.

Untitled 7

A major challenge was how to deal with a structure that moves and flexes whilst being constructed. This led to the development of the “Saturn” ring joint.

The roof is also fitted with oil dampers to protect the roof integrity during gale-force winds. Tange insisted they be visible and painted red.


Untitled 12Untitled 13It’s been half a century since. We don’t know who thought of exactly what, how, when or why. But sketches like this next one for the smaller gymnasium don’t just happen by accident. This one shows a complete structural system, one having a different roof structure from the large stadium, and that requires no bracing.


Untitled 15

We will never know whether it was Tange or Tsuboi who thought it might be a good idea to hang curved steel members hanging from spiralling cable in tension. All we know is that Tsuboi thought it worth investigating. The largest problem was the huge bending moment the column and its ground beam have to resist. This brings to mind the statement that Toshikata Sano of Tokyo Imperial University made circa 1910 at a symposium titled What Should Be The Style of Our Country’s Future Architecture?

the fundamental meaning of architectural beauty is no more than support of weight and clear expression of kinetic forces.

It’s easy to forget that the SHARP Compet CS10A, the world’s first electronic transistor calculator was not released until December, 1964. It looked like this. It had a memory and could perform the four basic arithmetic functions.


In 1964, IBM announced the IBM System/360. It looked like this. (Photos IBM)


It had 8 KB of internal memory that could be expanded up to 8 MB. I had 256 KB of main storage that could be expanded to 8 MB but 512 KB, 768 KB or 1024 KB were more usual.


Finally, here’s some good words that Tange had to say about their collaboration.


Untitled 6