Norton Starr, Archimedean Spiral, 1973.
For the underlying curve I used an Archimedean spiral, and located the center of each network on this curve. I also imposed a slight gap between adjacent networks, rather than having them touch. Satisfying these requirements, together with some other more recondite constraints, involved a considerable amount of analysis and experimentation, almost all of it purely computational. During the design stage, there were no more than two or three trial drawings because in that era it was very time consuming to produce such pictures. In contrast, I carried out repeated numerical experimentation in order to obtain locations for the centers of the networks and radii for them which were likely to yield the image I had in mind.
This design was drawn in 1973 by fountain pen on a Calcomp drum plotter, driven by an IBM 360/75 computer. The result was then photo-reversed and reduced by fifty percent before being lithographed in black ink on white paper.