MathOverflow is a question and answer site for professional mathematicians. It's 100% free, no registration required.

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 complete $C^\infty$ Riemannian metric on $\mathbb R^2$ of positive sectional curvature.

  1. Is the metric embeddable as the boundary of a convex subset of $\mathbb R^3$?

  2. Is the embedding unique?

  3. Are there generalizations of 1-2 to complete noncompact surfaces of nonnegative sectional curvature?

  4. What are good references for these matters?


$\bullet\ $ after doing some reading on the subject I found that the assertion 1 is true in the sense that the surface is isometric, as a metric space, to the boundary of a convex body in $\mathbb R^3$ (as proved by Alexandrov back in 1942). The matter of uniqueness is well-understood.

$\bullet\ $ However, one should not expect the boundary to be smooth, e.g. there are examples of $C^\infty$ metrics of nonnegative curvature on $S^2$ which cannot be isometrically $C^3$-embedded into $\mathbb R^3$.

$\bullet\ $ If the curvature is positive, then smoothness can be achieved as proved by Pogorelov and Nirenberg (independently in the 1950s).

$\bullet\ $ Local smooth isometric embedding for nonnegatively curved surfaces was established by Lin in 1985.

$\bullet\ $ A more recent reference for these matters is the book by Burago and Zalgaller, Geometry III, Encyclopedia of Mathematical Sciences.

share|cite|improve this question
Why is this titled "On Alexandrov embedding theorem"? What is the Alexandrov embedding theorem? (I am probably supposed to know this. In my defense, it is possible that I once knew it years ago, but have forgotten.) – Deane Yang Apr 21 '10 at 22:34
Alexandrov embedding theorem says that any positively curved metric on $S^2$ is isometric to the boundary of a convex body (in fact, very little regularity is assumed about the metric: it should be inner and have positive curvature in comparison sense). I know there is a version of this theorem for positively curved metrics on the plane, possibly with some condition on total curvature. Anyway, I think those who are able to answer know the relevant background; I do not. – Igor Belegradek Apr 21 '10 at 22:52
I googled it, and found this as the top hit:… :) – Nate Eldredge Apr 21 '10 at 22:54
@Igor: Thanks! I knew only the theorem for smooth metrics, which is due to Pogorelov and Nirenberg. I imagine that Alexandrov proved it under weaker regularity assumptions. But did he do this before, simultaneously, or after Pogorelov and Nirenberg? – Deane Yang Apr 22 '10 at 2:18
@Igor: Thanks for the update on your question. – Deane Yang Apr 29 '10 at 2:20
up vote 7 down vote accepted

Is the metric embeddable as the boundary of a convex subset of 3?

YES, it is a limit case of standard Alexandrov's theorem. Moreover one can choose any embedding of cone at infinity and construct the embedding. This is a theorem of Olovyanishnikov --- one of three students of Alexandrov who died in the war.

Is the embedding unique?

NO, but I suspect it is unique once you fixed the convex embedding of the cone at infinity. It might follow from the proof of Pogorelov's theorem but I was not able to check his proof.

Are there generalizations of 1-2 to complete noncompact surfaces of nonnegative sectional curvature?

I'm not sure what you mean --- if it has strictly positive curvature at one point then it is automatically $\mathbb R^2$. If it is $\mathbb R^2$ then it is all the same.

share|cite|improve this answer
Just to confirm, is it true that any complete metric of nonnegative sectional curvature on $\mathbb R^2$ is isometric to the boundary of a convex subset of $\mathbb R^3$? I have never seen this stated in the literature. – Igor Belegradek Apr 22 '10 at 3:16
@Igor, it is in the paper linked to theorem of Olovyanishnikov. – Anton Petrunin Apr 22 '10 at 3:34
@Anton, I cannot find the relevant statement in Olovyanishnikov's paper but on the same site I found an earlier paper of Alexandrov: "Existence of a convex polyhedron and of a convex surface with a given metric. (Russian. English summary) Rec. Math. [Mat. Sbornik] N.S. 11(53), (1942). 15--65.", which in Theorem 3 proves what I asked for for nonnegatively curved metrics on the plane with positive curvature at one point. – Igor Belegradek Apr 22 '10 at 4:33
@Igor, see §3 "Основная теорема" – Anton Petrunin Apr 22 '10 at 14:54
Thanks, Anton! By the way, this site for Mat.Sbornik is amazing. I never expected to find PDF's of articles from 1940s. And they even have English summary; unbelievable. – Igor Belegradek Apr 22 '10 at 17:40

I am not sure about some details, but there is a standard reference to most results of this type: "Extrinsic geometry of convex surfaces" by A.V. Pogorelov which is about 700 pp. in either English or Russian. As I later discovered, this book is essentially a union of 3-4 previous books that Pogorelov wrote on different topics. He even copied entire chapters from older books, including a chapter dealing with extensions of the Alexandrov embedding theorem to various functionals of curvatures and some global parameters. Anyway, this book is hard to read, but it a great source of material and references.

share|cite|improve this answer
I was going to get Pogorelov's book but I was hoping that some expert could give me answers before I read the book. – Igor Belegradek Apr 21 '10 at 23:06
I'm pretty sure that very few people besides Pogorelov understand the contents of this book, but I second Igor's recommendation. I haven't looked at the book in a long time and never really understood it, but there is a lot in there that I think deserves to be better understood and known. – Deane Yang Apr 22 '10 at 2:15

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.