On Sat, 17 Mar 2001, Patrick Powers wrote: > The sort of approximation you consider for the > circumference is behind Ramanujan's approximation too. Only he uses a more > elaborate method than a simple mean and as a result his formula is exact > for the case where the ellipse is virtually a circle and only a few percent > out when it is nearly a straight line... > > Patrick
Only about half a percent off at the limiting case! But it bothers me that it's consistently *low*... (As if I could do better than Ramanujan!) Dave
