I know that it is possible to construct an approximation of ellipses using 4 arcs, but are there any ellipses that can be constructed from arcs with 100 percent accuracy?

Usually, the construction would be slightly off: approximation vs real

Is it possible the there exists an ellipse where this is not the case? And it can be perfectly represented using arcs?

I imagine this is impossible. But in my research, I never managed to find a definitive answer.

  • 5
    $\begingroup$ Yes, they are called circles. :-) $\endgroup$ Aug 15, 2021 at 23:14
  • $\begingroup$ ...indeed ONLY circles. $\endgroup$ Aug 15, 2021 at 23:20
  • $\begingroup$ The orthogonal lines to the tangents of a circle all intersect at a single point. This is not true for an arc of an ellipse. $\endgroup$ Aug 15, 2021 at 23:27

1 Answer 1


We claim that an arc on an ellipse is the arc of a circle if and only if that ellipse is a circle. Indeed, look at the curvature: An arc of a circle has constant curvature. By direct calculation, though, you can check (see how to calculate the curvature of an ellipse) that in a non-circular ellipse, the curvature varies in such a way that it is never locally constant. (To see this, look at the linked answer's formula. Specifically, focus on $\sqrt{a^2\sin^2 t + b^2\cos^2 t}$. Calculus tells us this radical expression has derivative $$(a^2-b^2)\frac{\sin(2t)}{2\sqrt{a^2\sin^2 t + b^2\cos^2 t}}.$$ Note that as long as $a \neq b,$ this will only be zero at isolated points, so there can't be any small arc of an ellipse on which the curvature is constant.)


You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .