MathOverflow is a question and answer site for professional mathematicians. 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

Consider a compact surface $\mathcal{S} \subset \mathbb{R}^3$ without boundary and assume we are given the Gaussian curvature $K(x), x\in \mathcal{S}$. It is know that Gaussian curvature does not fully determine the Riemannian metric, as for example discussed in

Does the curvature determine the metric?

But for a given Gaussian curvature, does one have the freedom to choose a Riemannian metric such that the Gauss-Bonnet theorem holds?:

$$\int_{\mathcal{S}} K(x) d\mu_g(x) = 2\pi\; \chi(\mathcal{S}),$$

where $\mu_g$ is the Riemmanian volume measure induced by the metric $g$ and $\chi(\mathcal{S})$ is the Euler characteristic.

Is there even a constructive method for a Riemmanian metric $g$ under the conditions that the resulting Gaussian curvature is $K(x),x\in \mathcal{S}$ and the Gauss-Bonnet theorem is fulfilled?

(maybe this question is too general as stated above, but then might become more meaningful with assumptions on $K$.)

share|cite|improve this question
If the surface $S$ is diffeomorphic to a sphere or torus and the function $K$ is everywhere negative, clearly this is not possible. Similarly, if $S$ is diffeomorphic to a torus or higher genus surface and $K>0$ everywhere, this is not possible. – Ben McKay Apr 4 '14 at 13:11
a related MO question could be… – Ali Taghavi Apr 6 '14 at 8:18
up vote 12 down vote accepted

If $K$ is to be the Gauss curvature of some metric on a compact two-dimensional manifold $M$ with Euler characteristic $χ(M)$, then from the Gauss-Bonnet theorem one has the following conditions: (a) $K$ is positive somewhere if $χ(M)>0$, (b) $K$ changes sign or $K≡0$ if $χ(M)=0$, (c) $K$ is negative somewhere if $χ(M)<0$. Theorem: $K∈C^∞(M)$ is the Gauss curvature of some Riemannian metric on the two-dimensional, compact, connected manifold $M$ if and only if $K$ satisfies the above sign condition. MR0375153 (51 #11349) Kazdan, Jerry L.; Warner, F. W. Existence and conformal deformation of metrics with prescribed Gaussian and scalar curvatures. Ann. of Math. (2) 101 (1975), 317–331. 53C20 (58G99)

share|cite|improve this answer
Interesting result! I guess the minor question that remains then, is which Riemannian surfaces are embeddable in $\mathbb{R}^3$, since for example the flat torus is not. – Jaap Eldering Apr 4 '14 at 16:50
Not a minor question at all. Embeddability in $\mathbb{R}^3$ is a very complicated problem, even locally. There are abstract Riemannian metrics on surfaces with marked point, so that no neighborhood of the marked point admits any isometric embedding into $\mathbb{R}^3$: MR2393070 (2009k:53177) Reviewed Nadirashvili, Nikolai(F-PROV-APT); Yuan, Yu(1-WA) Improving Pogorelov's isometric embedding counterexample. (English summary) Calc. Var. Partial Differential Equations 32 (2008), no. 3, 319–323. 53C45 – Ben McKay Apr 4 '14 at 17:11

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.