I'm a little confused about fiber bundles. I have a specific example and would appreciate someone clarifying this for me. Let $f:M\times U\rightarrow TM$ be a map from a product space $M\times U\rightarrow$ to the tangent bundle TM of M. Is ($M\times U$,f,TM) a fiber bundle with $M\times U\equiv E$ the total space and TM the base space? If so, which is the fiber space? Or should the base be just M? For example, let $f(x,u)=\dot x=g(x)h(u), x \in M, u \in U$, for some g,h. Is f the projection map of the a fiber bundle? Should it be TM or $T_{x} M$, the tangent space of M at x? Thanks in advance!

  • 2
    $\begingroup$ You seem more than a little confused. $M \times U$ is a trivial fibre bundle over $M$, with fibre $U$. The projection map is the map $M \times U \to M$ such that $(m, u) \mapsto m$. $TM$ is another fibre bundle over $M$, with projection map $TM \to M$ such that $v \mapsto x$ for all $v \in T_x M$. $TM$ is generally not a trivial bundle over $M$. $\endgroup$
    – Zhen Lin
    Dec 16, 2011 at 3:27
  • $\begingroup$ What's $\dot x$ here? $\endgroup$ Dec 16, 2011 at 3:30
  • $\begingroup$ Thanks Zhen but i've already understood what you're saying. My main problem is the map f:M×U→TM. Can this be a bundle? $\endgroup$
    – Jorge
    Dec 16, 2011 at 3:34
  • 1
    $\begingroup$ @Jorge But $x \in M$, isn't it? I don't really see what $f, g, h$ are supposed to be. It would be good to write down what it means to be a fibre bundle over $TM$, and see whether you've satisfied the requirements. $\endgroup$ Dec 16, 2011 at 3:36
  • 2
    $\begingroup$ Please provide more information. From the definition you need at the very least $f:M\times U\to TM$ to be surjective. For a general map this is obviously false (for example, let $f = \iota\circ\mathrm{pr}_1$ where $\mathrm{pr}_1$ is the projection from $M\times U$ to $M$ and $\iota$ is the inclusion of $M$ into $TM$ as the 0 section). So please be more precise and edit into the question what you know about the map $f$. $\endgroup$ Dec 16, 2011 at 15:48

1 Answer 1


In the definition of a fiber bundle, the map from the total space $E$ to the base space $M$ is the canonical projection, meaning locally $f:M\times F\to M$ takes $f(x,p)=x$ for fiber $F$. In your example, I guess $f:M\times U\to TM$ could be the projection of a fiber bundle if $TM$ $was$ a trivial bundle, say $M=\mathbb{R}^4$. But also note that most people would write $TM$ as the the entire bundle, meaning $TM=M\times \mathbb{R}^n$ locally, not just $TM=\mathbb{R}^n$.

EDIT: Note that I keep saying "locally". For any fiber bundle $f:E\to M$ with fiber $F$, you can write $f_i:U_i\times F\to M$ for a local covering $\{U_i\}$ of $M$. I won't write down all the details of that, but you should be familiar with it as well.


You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .