This discussion leads me to wonder that if a finite number of circular arcs in R^3 are fitted together to make a C^1 closed curve that's knotted, what is the smallest number of distinct arcs that can do this? --Dan ________________________________________________________________________________________ It goes without saying that .