I'm just wondering about my proof for the following fact. I get the feeling it is almost trivial but I am still getting a feel for geometry and so it doesn't seem 'obvious' to me just yet.

The projection $\pi$ of the fiber bundle $(E,\pi,M,F)$ is a submersion. Is the below argument a correct proof of this fact? Would it be considered 'the obvious' proof?

To show $\pi$ is a submersion, we must show that for each $p\in E$, the differential $d\pi_p:T_pE\rightarrow T_{\pi(p)}M$ is surjective.

Since, $(E,\pi,M,F)$ is a fiber bundle, there exists a neighbourhood $U\subseteq M$ containing $\pi(p)$ and diffeomorphism $\phi:\pi^{-1}(U)\rightarrow U\times F$ , such that $\pi|_{\pi^{-1}(U)}=\pi_1\circ \phi:\pi^{-1}(U)\rightarrow U$. Thus, showing $d\pi_p$ is surjective is equivalent to showing $d(\pi_1\circ \phi)_p={d\pi_1}_{\phi(p)}\circ d\phi_p$ is surjective. Since $\phi$ is a diffeomorphism, $d\phi_{p}$ is an isomorphism, so it is surjective. Furthermore, ${d\pi_1}_{\phi(p)}$ is surjective since projection onto the first factor is a submersion. Since the composition of two surjections is a surjection itself, it follows that ${d\pi_1}_{\phi(p)}\circ d\phi_p=d\pi_p$ is surjective. Thus, $\pi$ is a submersion.

  • $\begingroup$ You convinced me. $\endgroup$ – Amitai Yuval Feb 7 '15 at 17:52

Yes your proof is correct. You can formulate it a little bit shorter as follows: a map is a submersion if and only if there exists local charts such that the representation of the map in this charts is a projection (one direction is very easy, the converse follows from the inverse function theorem). Now for the projection of a fibre bundle, we just take a local trivialization $\pi^{-1}(U) \to U \times F$ and observe that the local representation of the bundle projection is the projection on the first factor. Essentially, this boils down to your argument.

  • $\begingroup$ Thanks very much for this. Do you know where I can find a proof of your statement? $\endgroup$ – beedge89 Feb 9 '15 at 23:11
  • $\begingroup$ You mean the "submersion iff locally projection"-part? I think it is contained in Marsden, Ratiu & Abraham "Manifolds, Tensor Analysis, and Applications". You can find an older version of this in the internet. $\endgroup$ – Tobias Diez Feb 9 '15 at 23:36

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.