Definition. Let $M$ and $N$ be smooth manifolds, and let $F\colon M\to N$ be any map. We say that $F$ is a smooth map if for every $p\in M$, there exists smooth chart $\big(U,\varphi\big)$ containing $p$ and $\big(V,\psi\big)$ containing $F(p)$ such that $F(U)\subseteq V$ and the composite map $\psi\circ F\circ \varphi^{-1}\colon \varphi(U)\to\psi(V)$ is smooth.

I have to show the following proposition, but many points are not clear to me.

Proposition. Suppose $M$ and $N$ are smooth manifolds with or without boundary, and $F\colon M\to N$ is a map. Then $F$ is smooth if and only if either of the following condition is satisfied:

$(a)$ For every $p\in M$, there exist smooth charts $(U,\varphi)$ containing $p$ and $(V,\psi)$ containing $F(p)$ such that $U\cap F^{-1}(V)$ is open in $M$ and the composite map $\psi\circ F\circ\varphi^{-1}$ is smooth from $\varphi(U\cap F^{-1}(V))$ to $\psi\big(V\big)$.

$(b)$ $F$ is continuous and there exists amooth atlases $\{(U_\alpha,\varphi_\alpha)\}$ and $\{(V_\beta,\psi_\beta)\}$ for $M$ and $N$ respectevely such that for each $\alpha$ and $\beta$, $\psi_\beta\circ F\circ \varphi_\alpha^{-1}$ is a smooth map from $\varphi_\alpha(U_\alpha\cap F^{-1}(V_\beta))$ to $\psi(V_\beta)$.

Proof. I would like to proceed, hoping that it is the simplest way as follows: first prove that $F$ is smooth iff $(a)$ holds, and after showing that $(a)$ holds iff $(b)$ holds.

F is smooth iff $(a)$ holds

($\Rightarrow$) If $F$ is smooth we have that $F(U)\subseteq V$, then $U\subseteq F^{-1}(V)$. Therefore $U\cap F^{-1}(V)=U$ that is open.

$(\Rightarrow)$ If $(a)$ holds for every $p\in M$, there exist smooth charts $(U,\varphi)$ containing $p$ and $(V,\psi)$ containing $F(p)$ such that $U\cap F^{-1}(V)$ is open in $M$ and the composite map $\psi\circ F\circ\varphi^{-1}$ is smooth from $\varphi(U\cap F^{-1}(V))$ to $\psi\big(V\big)$. It only remains to show that $F(U)\subseteq V$, now if $\tilde{p}\in U$ then there exists a smooth chart $(\tilde{V},\tilde{\psi})$ containing $F(\tilde{p})$

Question 1. How can I prove that $F\big(\tilde{q}\big)\in V$?


$(b)\Rightarrow (a)$ it seems obvious.

Question 2. I have no idea how to show that $(a)\Rightarrow (b)$. Could you give me a hints?


  • $\begingroup$ For Q1, you should first replace $U$ by $U\cap F^{-1}V$ and restricting $\varphi$, then everything follows. For Q2, prove smooth implies (b), which amounts to just patching the charts. $\endgroup$ May 12, 2019 at 10:25

1 Answer 1


Question 1: $F(U)\subset V$ is not always true. But you can take a new chart of the form $(U^\prime,\varphi)$ which has this property (take $U^\prime=U\cap F^{-1}(V)$).

Question 2: If you want to do it without "$a\Leftrightarrow F \text{ smooth}$", which I guess you do:

For $p\in M$ take $(U_p,\varphi_p)$ and $(V_p,\psi_p)$ some charts around $p$ and $F(p)$ respectively. If $U_p^\prime=U_p\cap F^{-1}(V_p)$ then $\{ U_p^\prime\}_p$ is an open cover for $M$ such that $F_{\vert U_p^\prime}$ is continuous, so $F$ is continuous.

Take the atlas $\{(U_p^\prime,\varphi_p)\}_p$ for $M$. Unfortunately $\{(V_p,\psi_p)\}$ may not be an atlas for $N$ because the domain of the charts may not cover $N$ is $F$ is not surjective. You can take the original atlas of $N$ instead, which we write $\{(V_\alpha,\psi_\alpha)\}$. Then for $p$ and $\alpha$ the map

$$\psi_\alpha\circ F\circ \varphi_p^{-1}:\varphi_p(U_p^\prime\cap F^{-1}( V_\alpha))\longrightarrow\psi(V_\alpha)$$

has domain $\varphi_p(U_p^\prime\cap F^{-1}( V_\alpha))=\varphi_p(U_p\cap F^{-1}(V_p\cap V_\alpha))$ so it is just the composition

$$\underbrace{\psi_\alpha\circ\psi_p^{-1}}_{\substack{\text{smooth because of the} \\ \text{coherence of the charts}}} \circ\underbrace{\psi_p\circ F\circ \varphi_p^{-1}}_{\substack{\text{smooth by assumption} }}.$$

  • $\begingroup$ @AdamThanks for your ansewer. Perhaps it is more direct to prove that if $F$ is smooth then $(b)$ holds and viceversaa. What do you say? $\endgroup$
    – Jack J.
    May 12, 2019 at 14:02
  • 1
    $\begingroup$ @JackJ. Yes the best way might be to prove that "$F$ smooth" implies $(b)$ then $(b)\Rightarrow (a)$ and finally $(a)$ implies "$F$ smooth". But if you want to learn how to manipulate coordinate charts it's a good exercise to prove every directions! $\endgroup$ May 12, 2019 at 14:23
  • $\begingroup$ @AdamI have another problem! If I start from $F$ smooth, then $(a)$ holds, because we have shown equivalence. From this point how do I get to $(b)$? $\endgroup$
    – Jack J.
    May 12, 2019 at 14:33
  • $\begingroup$ @JackJ. you mean you want to prove "$F$ smooth" implies $(b)$ ? $\endgroup$ May 12, 2019 at 14:37
  • 1
    $\begingroup$ @JackJ. this is because $F_{\vert U_p^\prime}$ composed at the beginning and at the end by homeomorphisms (namely $\varphi_p$ and $\psi_p$) is continuous (by assumption it is even differentiable in the usual sense). (Ps: don't forget to accept the answer if it satisfies you) $\endgroup$ May 12, 2019 at 15:54

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.