Definition 1

A surface of dimension k (or k-dimensional surface or k-dimensional manifold) in $\mathbb R^n$ is a subset $S\subset \mathbb R^n$ each point of which has a neighborhood in $S$ homeomorphic to $\mathbb R^k$.

Definition 2

A set $A(S) := \{\phi : I_k \to U_i, i \in\mathbb N\} $of local charts of a surface $S $ whose domains of action together cover the entire surface (that is, $S = \bigcup U_i$), is called an atlas of the surface $ S$.

Since $A(S)$ is countable, but I wanna know if every surface $S$ has an atlas? If not, could you give a counter example?

  • $\begingroup$ This is not a right definition of an atlas. What book are you reading? $\endgroup$ Jun 30 at 4:25
  • $\begingroup$ Mathematical Analysis by Vladimir A.Zorich $\endgroup$
    – LEY
    Jun 30 at 4:26
  • 1
    $\begingroup$ What is $I_k$? What are the $U_i$? What definition of a surface do you have in mind? A 2-dimensional manifold admits an atlas by definition… $\endgroup$
    – PrudiiArca
    Jun 30 at 4:27
  • $\begingroup$ Then take another look at Zorich's definition. $\endgroup$ Jun 30 at 4:28

The trick is that any subspace of $\mathbb{R}^n$ is second-countable and hence Lindelöf: any open cover has a countable subcover. So, since your $S$ according to Definition 1 has an open cover by sets homeomorphic to $\mathbb{R}^k$, there is a countable subcover which gives an atlas according to Definition 2.


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.