Take the 2-minute tour ×
Mathematics Stack Exchange is a question and answer site for people studying math at any level and professionals in related fields. It's 100% free, no registration required.

Hopf's Umlaufsatz (better known as?) says:

Let $\gamma$ be a simple closed di fferentiable positively oriented curve in the plane. Then for its curvature $\kappa$ it holds:

$$\int_{\gamma}\kappa\ \text{d}s = 2\pi$$

I wonder if (and cannot see why not) the inversion holds, too:

Let $\kappa$ be a continuous function $\kappa: [0,1] \rightarrow \mathbb{R}$ with $\kappa(0) = \kappa(1)$ and $\int_0^{1}\kappa\ \text{d}s = 2\pi$. Then there is a simple closed di fferentiable positively oriented curve $\gamma$ (of length 1) with curvature $\kappa$.

(If this holds, $\gamma$ would be unique upto congruency via Euclidean motions.)

share|improve this question
1  
Sometimes this is called the 2-dimensional Gauss-Bonnet theorem, as in this article. // Yes, one can reconstruct a curve from its curvature, as on second page here. –  user53153 Jan 11 '13 at 20:49
1  
@5PM: A planar curve can be reconstructed from it's curvature, but the question is if the curve is closed or not. –  Jason DeVito Jan 16 '13 at 20:51
add comment

1 Answer

There is another necessary condition on $\kappa$. The so-called Four vertex theorem:

Theorem: Suppose $\kappa$ is the curvature of a simple closed curve. Then, either $\kappa$ is constant or has at least 2 minimums and 2 maximums.

The ellipse shows that there is no reason to expect more than 2 minimums or maximums of curvature.

It turns out that together with your integral condition and this condition together are necessary and sufficient for $\kappa$ to correspond to a closed simple curve. A write up, due to DeTurck, Gluck, Pomerleano, and Vela-Vick, (who say the result is due to Dahlberg) can be found here. See also a more informal (but similar) write up by Vela-Vick here.

share|improve this answer
    
Thanks, now I remember reading about the theorem long time ago... Now I also see that @Hans wanted length $1$, which is not guaranteed by the construction as far as I can tell... Or maybe Hans even wanted an arclength parametrization? I hope he'll clarify. –  user53153 Jan 16 '13 at 21:15
    
@5PM: I missed the unit length part. In that case, I'm not sure what more needs to be added. –  Jason DeVito Jan 16 '13 at 23:22
add comment

Your Answer

 
discard

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.