2
$\begingroup$

Let the n-dimensional sphere be defined as the set: $$ \mathbb{S}^{n-1} := \{(x_1, x_2,..., x_n)\in \mathbb{R}^{n-1}: x_1^2 + x_2^2+...+x_n^2 = 1 \}$$

If we cover $\mathbb{S}^{n-1}$ with this Atlas of 2 charts $$ \mathscr{A}= \{ (U_N, \phi_N), (U_S, \phi_S)\}$$ using the stereographic projections:

$$\phi_N: U_N:=\mathbb{S}^{n-1}\setminus \{(0,0,...,0,1)\}\rightarrow \mathbb{R}^{n-1}, \\ \phi_N(x_1, x_2,..., x_n) = (\frac{x_1}{1-x_n}, \frac{x_2}{1-x_n},..., \frac{x_{n-1}}{1-x_n}) $$ and $$\phi_S: U_S:=\mathbb{S}^{n-1}\setminus \{(0,0,...,0,-1)\}\rightarrow \mathbb{R}^{n-1}, \\ \phi_S(x_1, x_2,..., x_n) = (\frac{x_1}{1+x_n}, \frac{x_2}{1+x_n},..., \frac{x_{n-1}}{1+x_n}) $$ How can you proof that the topology derived from the differential structure provided by $\mathscr{A}$ is, in fact, the usual topology of $\mathbb{R}^{n-1}$ restricted to $\mathbb{S}^{n-1}$? Moreover, how can you proof that this topology is Hausdorff and also second countable?

$\endgroup$
2
  • $\begingroup$ There is a theorem of differential topology that if $U \subset \mathbb R^k$ is open, if $f : U \to \mathbb R^n$ is a smooth injection, and if $p \in U$, then there exists $r > 0$ such that $B(p,r) \subset U$ and $f \mid B(p,r)$ is a homeomorphism between the sets $B(p,r)$ and $f(B(p,r))$, each with the subspace topology. Is this kind of thing familiar to you? $\endgroup$
    – Lee Mosher
    Feb 26, 2019 at 23:45
  • $\begingroup$ It does make sense for me, and I think I know how to finish the exercise knowing that theorem, but actually, no, our teacher hasn't given us that theorem yet. Thank you anyways;; $\endgroup$
    – sasha
    Feb 27, 2019 at 19:29

1 Answer 1

1
$\begingroup$

To see that you get the usual topology, observe that chart mappings are continuous for the subspace topology on $U_N$ and $U_S$. Since you can easly write down continuous inverses, you conclude that the charts are homeomorphisms from $U_N$ and $U_S$ (endowed with the topology inherited from the subspace topology on $S^{n-1}$). This readily implies that the topologies coincide.

The Hausdorff and second countability properties are also easly verified directly. Given two points such that at least one of them is different from $N$ and $S$, there is one chart containing both points, and you can separate them in that chart. For $N$ and $S$ you can take the unit balls in the two charts as disjoint open neighborhoods. For second countability, you simply proof that the preimages of balls with rational centers and radii in the two charts form a basis for the the induced topology.

$\endgroup$

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.