# Are there any surfaces that contain both positive and negative Gaussian curvature?

Would a torus shape be considered to have both positive and negative Gaussian curvature?

-

## migrated from physics.stackexchange.comSep 16 '13 at 7:55

This question came from our site for active researchers, academics and students of physics.

Yes, there are many of those surfaces. In general saddle points will result in negative Gaussian curvature because the two principle radii of curvature are opposite in sign whereas peaks and holes will result in positive Gaussian curvature because their principle radii of curvature have the same sign (either both negative or both positive).

A few examples of surfaces with both positive and negative Gaussian curvature can be readily found in nature: a pear (usually), a peanut shell, a baseball bat (some negative Gaussian curvature at the handle).

A torus indeed has both positive and negative Gaussian curvature, because it has saddle points, a whole ring of them at the inside of the torus, and it has `peaks' at the outside. So it depends whether you are inside (in the hole) or outside the torus whether the Gaussian curvature is negative respectively positive (see e.g. this module on curvature).

-

Consider a torus $T$ in $\Bbb R^{3}$. By the following question, $T$ has a point of positive curvature.

Any compact embedded $2$-dimensional hypersurface in $\mathbb R^3$ has a point of positive Gaussian curvature

By the Gauss-Bonet theorem, $\int_T K \ dA = \chi(T)=2-2g=0$, where $g=1$ is the genus of the torus. Thus, since $T$ has a point (and therefore a neighborhood) of positive curvature, it must have a neighborhood of negative curvature in order to make the above integral vanish.

This might be overkill, but I like it.

-

It is a theorem of Hilbert that any closed smooth surface without boundary in 3 space must have a point of positive Gauss curvature. So any surface that also has a point of negative Gauss curvature will have both.

The integral of the Gauss curvature over a surface is 2pi times its Euler characteristic. The torus has Euler characteristic zero so it must always have regions of negative Gauss curvature to cancel the regions of positive curvature guaranteed in Hilbert's theorem. There is no way to warp it so that its Gauss curvature has only one sign. Similarly surfaces with more than one handle all have negative Euler characteristic and thus must always have both positive and negative Gauss curvature.

This is not true for the torus in 4 space. Here the torus can be given zero Gauss curvature everywhere.

-