I got up from the breakfast table to go to the keyboard and my eyes went to a white mass-market Corelle platter with three ripening peaches on it. The outline of the platter seemed pleasing, and as I looked I realized it was because its oval shape was not just two semicircles joined by straight lines. The sides had some curvature of their own. That curvature seemed “right” to me. There is probably a perfectly ordinary reason for this; familiar living things such as trees, people, and fish tend not to embody severe changes of shape, such as from straight to curve. Curvature tends to be smoothly continuous.

Then I remembered that the first "oval" pistons made for Honda's attempt to make a four-stroke viable in 500cc Grand Prix racing had straight sides. The 19,000-rpm NR500 race engine had piston-ring sealing problems, and when engine design became the responsibility of Suguru Kanazawa, one of the changes he made was to add curvature to the previously flat sides of its pistons. Why did he try this? Why did it work?

One simple possibility comes to mind, that applying compression to the straight parts of the early piston rings could make them bulge either outward or inward. Such compression might arise as the ring expanded at operating temperature, pressing its semicircular ends against matching parts of the cylinder wall. This would compress the less-curved parts of the ring, driving them only outward against the cylinder wall.

Honda NR 500
On a mid-'80s trip to Japan, Cycle World editors saw a Honda NR500 displayed at HRC HQ. From the November 1984, issue: "We also learned that Honda engineers have solved the sealing problems that bothered the Dykes-type rings used on the NR500 pistons."Cycle World

That is just speculation, but what’s sure is that the best designers are guided by both their formal knowledge and by a personal aesthetic—an accumulated sense of the rightness of things. It is said that Yoichi Oguma, who led Honda racing in the time of Freddie Spencer, insisted that junior engineers accompany him to museums of design and industry to discuss the variety of forms and ideas presented there. Formal knowledge is essential—the ability to determine the vibratory modes of crankshafts, for example—but that knowledge has to be about something in the real world.

Elliot Morris, who produced pioneering mag wheels for racing motorcycles beginning in 1973, seemed slightly embarrassed to reveal how he’d decided to give his wheels seven spokes. “I’d noticed that a lot of things in nature have odd numbers, so that’s what I chose.”

V-twin and carbon structures designer John Britten once speculated, “I think if only we could know enough, we’d find that the separate aesthetics of mathematics, of machine parts, of music are all related—all part of a single aesthetic.”

Have a look at cast-iron pulleys and other spoked machine elements from 19th-century American factories and you will see both curved spokes and odd numbers of spokes. Those were forms found least likely to crack as they solidified in their sand molds.

No knowledge is irrelevant.