Mathematics Stack Exchange is a question and answer site for people studying math at any level and professionals in related fields. Join them; it only takes a minute:

Sign up
Here's how it works:
  1. Anybody can ask a question
  2. Anybody can answer
  3. The best answers are voted up and rise to the top

Let $\gamma$ be a smooth curve in Euclidean space of length $2\pi$ whose curvature function satisfies $-1 < k(t) < 1$. Can $\gamma$ be closed?

This seems like it should be an easy exercise, at least in the plane, but I'm stuck...

share|cite|improve this question
up vote 2 down vote accepted

do Carmo's Differential Geometry of Curves and Surfaces mentions Fenchel's theorem on page 399:

The total curvature of a simple closed curve is at least $2\pi$

(plus something about when equality holds; this statements differs from the wikipedia page for Fenchel's theorem.)

In your case, you have that the total curvature $\int |k(s)| ds$ is $< 2\pi$. So the curve cannot be closed (if it is simple).

There is a generalization to piecewise regular closed curves in $\mathbb R^n$, known as the Fenchel-Borsuk theorem, though I can't find a handy reference right now.

share|cite|improve this answer
How popular is do Carmo's book? Because I see it mentioned all the time. Should I buy it? – Lays Oct 9 '13 at 5:07

If the curve was closed, then the Gauss-Bonnet formula for curves would apply, that is, $\int_0^{2\pi} k(t) dt= ±2\pi$ (where the curve is parameterized by arc length), but by the inequality $|k(t)|<1$ we get $\int_0^{2\pi} k(t) dt<\int_0^{2\pi}1dt = 2\pi$ and $\int_0^{2\pi} k(t) dt> -2\pi$.

edit : We might need to add the hypothesis that the curve is simple for this method to work, so this might not be a good solution.

edit 2 : I think that if the curve is not simple, then $\int_0^{2\pi} k(t) dt= 2k\pi$ where k is the winding number of the curve, which would fix this solution. See this link

share|cite|improve this answer
This is almost a perfect answer. But the winding number is defined relative to a point; it's not an absolute quantity. So there's just a bit more to be said. – Bruno Joyal Jul 19 '11 at 22:24
Right, I meant the winding number of curve traced by the unit tangent vector. – Vhailor Jul 19 '11 at 22:32

Take the mid-point and assume that the unit tangent vector at that point is looking in the direction of the $x$-axis. Now estimate the $x$-projection of the unit tangent vector at the distance $t$ from the midpoint. It is greater than $\cos t$ because the angle of rotation is less than $t$ due to the curvature restriction. But $\int_{-\pi}^\pi\cos t\,dt=0$, so you finish to the right of where you started. This works in all dimensions.

share|cite|improve this answer
This is too hand-wavy for me. – Eric Naslund Jul 19 '11 at 21:50

Your Answer


By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.