Let $M$ and $N$ manifold oriented with $M$ connected. Show that $F\colon M\rightarrow N$ is local diffeomorphism if and only if $F$ preserves or reverses orientation.

My initial idea is to use orientation forms but I do not know how.


Let $F$ be a local diffeomorphism and let $\omega$ be orientation-determining non-vanishing top form on $N$. Then we have that $F^*\omega$ is a non-vanishing form on $M$ and it determines an orientation on $M$. Now because $M$ is connected the orientation determined by $F^*\omega$ and the original one on $M$ must either match at all points or at none. In the first case the both orientations are consistent, while in the latter case the postively oriented bases wrt to $F^*\omega$ are negatively oriented wrt the initial orientation. WLOG we can assume that the first case is true in our calculations.

Now let $p \in M$ and let $f:U \to F(U)$ be a diffeomorphism on some open neighbourhood around $p$. Also $p$ is the domain of some local frame, as $M$ is an orientable manifold. WLOG we can take it to be $U$. (Indeed if the domain is $V$ we can consider the intersection $U \cap V$ and the restriction of $f$ on it).

Now let $\{E_1,\dots,E_m\}$ be a positively oriented local frame on $U$. As $f$ is diffeomorphism we have that $\{df(E_1),\dots,df(E_m)\}$ is a frame on $f(U)$. Then it's positively oriented if $\phi(df(E_1),\dots,df(E_m)) > 0$. This is true as:

$$\omega(df(E_1),\dots,df(E_m)) = f^*\omega(E_1,\dots,E_m) = F^*\omega(E_1,\dots,E_m)>0$$

As this is true for every $p$ in $M$ we have that $F$ preserves the orientation.

First of all for the other direction we must have that $dF$ is non-singular in order for the orientation preserving (reversing) to be defined. Then by the Inverse Function Theorem we have that $F$ is a local diffemoprhism.

  • $\begingroup$ Where did you use that $M$ is connected? $\endgroup$ – Mancala Jun 8 '18 at 23:30
  • $\begingroup$ @Mancala In the first paragraph. The fact that $M$ is connected means that $F^*\omega$ determines the same (or opposite) orientation on $M$ wrt to the original. $\endgroup$ – Stefan4024 Jun 8 '18 at 23:37
  • $\begingroup$ @Mancala If $M$ has, let's say $2$ parts, then it's possible for $F^*\omega$ to match the original orientation on one component, while not on the other. $\endgroup$ – Stefan4024 Jun 8 '18 at 23:40
  • $\begingroup$ @Mancala I've elaborated my answer a bit more. Hopefully it will help you. $\endgroup$ – Stefan4024 Jun 8 '18 at 23:48
  • $\begingroup$ @ Stefan4024 thanks! $\endgroup$ – Mancala Jun 9 '18 at 20:37

$(\Rightarrow)$ Let $\omega$ an oriteted positively (or negatively) form of $N$. We will show that $F^*\omega$ is a form of positively or negatively oriented orientation of $M$. Since $M$ is orientable, then there exists a $\eta$ form of orientation of $M$.

Since $M$ is connected, then $M$ has only two orientations, namely $\eta$ or $- \eta $. Like this,

$$F^*\omega=f\eta \text{ or } F^*\omega=f(-\eta) \text{, with }f>0$$

Therefore, $F$ preserves or reverses the orientation.

Is correct?

  • $\begingroup$ Assuming that $F^*\omega =\eta \text{ or } F^*\omega=-\eta$ is wrong, as the forms can determine the same orientation, but they are not necessarily the same. But you can say that the orientation determined by $F^*\omega$ is same as the one determined by $\eta$ or $-\eta$. $\endgroup$ – Stefan4024 Jun 10 '18 at 10:34
  • $\begingroup$ @Stefan4024 Yes, I forgot to put the correction factor $f>0$! $\endgroup$ – Mancala Jun 10 '18 at 13:08

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.