This question is about the diffeomorphism of $\mathbb{C}P^1$ and $S^2$. At the end of youler's answer, we read

"the general fact that two manifolds are diffeomorphic when you can give them each a coordinate atlas with the same transition maps."

I am not sure why this is true. I have not even a geometric intuition.

In my lecture notes, we were given the same example, up to showing that the two manifolds have the same transition maps. I think we are expected to understand that the transition maps imply that the two manifolds must be diffeomorphic to each other, but I don't see why.

If possible, I would like an intuitive reasoning, followed by a rigorous answer.


2 Answers 2


Suppose that $X, Y$ are manifolds with atlases $\{\phi_\alpha: \alpha\in A\}$ and $\{\psi_\alpha: \alpha\in A\}$ such that for any two pairs of indices $\alpha, \beta\in A$, $$ \phi_\beta\circ \phi_\alpha^{-1}= \psi_\beta\circ \psi_\alpha^{-1}. $$
Define the map $f: X\to Y$ by the formula $$ f(x)=y $$ whenever $\phi_\alpha(x)= \psi_\alpha(y)$ for any pair of charts with common index whose domains contains $x$ and $y$ respectively. The fact that the transition maps of the two atlases are the same implies that this map is well-defined (i.e. are independent of the choice of the above charts). To see that this map is a diffeomorphism, one needs to check smoothness of this map (and of its inverse) in local coordinates, i.e. smoothness of compositions $$ \psi_\alpha \circ f \circ \phi_\alpha^{-1} $$ and $$ \phi_\alpha \circ f^{-1} \circ \psi_\alpha^{-1}. $$ But these compositions are equal the identity maps for every $\alpha\in A$. Hence, $f$ and its inverse are diffeomorphisms. Thus, $f: X\to Y$ is a diffeomorphism.

The converse statement holds as well: Suppose that there is a diffeomorphism $f: X\to Y$. Then $X, Y$ can be equipped with atlases that have the same transition maps. Indeed, given an atlas $\{\psi_\alpha: \alpha\in A\}$ on $Y$, define charts on $X$ by the formula $$ \phi_\alpha= \psi_\alpha\circ f. $$ Direct computations shows that transition maps for the two atlases are the same.


Consider two manifolds $M$ and $N$, with atlases $\{(U_{\alpha}, \phi_{\alpha})\}$ and $\{(V_{\alpha}, \psi_{\alpha})\}$ respectively. Let $\tau = \phi_{\beta} \circ\phi_{\alpha}^{-1}:\phi_{\alpha}(U_{\alpha}\cap U_{\beta})\rightarrow\phi_{\beta}(U_{\alpha}\cap U_{\beta})$ and $\xi=\psi_{\beta} \circ\psi_{\alpha}^{-1}:\psi_{\alpha}(V_1\cap V_{\beta})\rightarrow\psi_{\beta}(V_{\alpha}\cap V_{\beta})$ be the transition maps associated with each atlas.

If $\tau = \phi_{\beta} \circ\phi_{\alpha}^{-1} = \psi_{\beta} \circ\psi_{\alpha}^{-1} = \xi$, we then have $$\phi_{\beta}(U_{\alpha}\cap U_{\beta})=\psi_{\beta}(V_{\alpha}\cap V_{\beta}) \implies M \supset U_{\alpha}\cap U_{\beta}=V_{\alpha}\cap V_{\beta} \subset N,$$ since both $\phi_{\beta}$ and $\psi_{\beta}$ are homeomorphisms.

Let's denote this open set with $S=U_{\alpha}\cap U_{\beta}=V_{\alpha}\cap V_{\beta}$.

As stated in youler's answer, consider the map $\Phi:M \rightarrow N$, defined by $\Phi(\phi_{\alpha}^{-1}(S))=\psi_{\alpha}^{-1}(S)$. Its inverse $\Phi^{-1}:N \rightarrow M$, also exists and is defined by $\Phi(\psi_{\alpha}^{-1}(S))=\phi_{\alpha}^{-1}(S)$. If we can somehow show that $\Phi$ and $\Phi^{-1}$ are both smooth... then indeed the two manifolds are diffeomorphic.


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.