Are there ways to know the curvature of a curve $\gamma$ that lives on a surface $\mathcal{S}$starting from the gaussian curvature of $\mathcal{S}$? In general, is it possible bound the curvature of this curve with a function that depends from the gaussian curvature?


No way the tangent curvature can be fixed .For example you can draw on a sphere a geodesic, any small circle which can be any displaced arbitrary parallel, all living on the sphere.

You can have them constant when deforming together isometrically.In bending both $k_g, K$ (geodesic and Gauss curvatures) remain constant.

But you cannot derive one from the other, they are independently defined or formed from the first fundamental form coefficients.

If a product is 12,factors can be (6,2) or (4,3) or (8, 3/2) etc. Anyway you can physically while embedding in 3-space broaden, narrow up or squeeze out a flexible surface isometrically in bending.

You can also, using a "concentric" geodesic polar coordinate system draw parallel circles of changing geodesic curvature. Even a constant geodesic curvature line (that would roll out to a circle if surface were a developable) ... can be drawn on a surface of of arbitrarily changing K.

But you cannot define or extract all lines $\gamma$ on a surface from given $K$.

To find geodesic curvature we need to further differentiate on arc the inclination of line to the lines of curvature through the second fundamental form.

For surfaces of revolution $k_g$ by Liouville's formula is:

$$ k_g = \psi ^{\prime} + \sin \psi \sin \phi / r $$

and $K$ is:

$$ K = \frac{d\phi}{ds \cos \psi}\cdot \frac {\cos \phi }{r} $$

which are independent.


No, and no.

Consider a torus with "core" radius $r$ and "tube radius" $s$, lying on a table like a donut. A point i=$P$ in contact with the table surface lies on a circle of radius $r$ and an orthogonal circle of radius $s$, both geodesic, so the Gaussian curvature at $P$ is $rs$.

Now fix $k$ and pick $r = k/s$. Then $rs = k$, so for any such $s$, we have the same gaussian curvature. But the "around the tube" curves through $P$ (for various values of $s$) will have curvature $1/s$, which is unbounded as $s$ grows small.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

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