8

I'm reading John Lee's Introduction to Smooth manifolds 2nd edition. Consider the problem 10-1:

Define an equivalence relation on $\mathbb{R}^2$ by $(x,y) \sim (x',y')$ if and only if $(x',y')=(x+n,(-1)^n y)$ for some $n \in \mathbb{Z}$. Let $E=\mathbb{R}^2 / \sim$ and $q:\mathbb{R}^2 \rightarrow E$ be the quotient map. Let $\pi_1:\mathbb{R}^2 \rightarrow \mathbb{R}$ be the projection onto the first factor and $\epsilon:\mathbb{R} \rightarrow S^1,x \mapsto e^{2\pi i x}$ be the smooth covering map. Then there exists a map $\pi:E \rightarrow S^1$ such that $\pi \circ q=\epsilon \circ \pi_1$ since $\epsilon \circ \pi_1$ is constant on each equivalence class.

(a) Show that $E$ has a unique smooth structure such that the quotient map $q:\mathbb{R}^2 \rightarrow E$ is a smooth covering map.

(b) Show that $\pi:E \rightarrow S^1$ is a smooth rank-1 vector bundle.

(c) Show that it is not trivial.

My attempted solution:

proposition 4.40 in the book says

Suppose $M$ is a connected smooth $n$-manifold and $f:E \rightarrow M$ is a topological covering map. Then $E$ is a topological $n$-manifold and has a unique smooth structure such that $f$ is a smooth covering map.

This proposition is similar to part(a), however, the map $q$ is from $\mathbb{R^2}$ to $E$ not the other way around. How should I manipulate $q$ to satisfy this proposition?

For part(b), the only thing I know is that part (a) gives $E$ is smooth but I have no clue about the rest.

For part(c), Lee defines a trivial bundle to be a bundle with a global trivialization. Suppose $E$ has a global trivialization $\Phi:\pi^{-1}(S^1) \rightarrow S^1 \times \mathbb{R}$.

If I can show $\pi^{-1}(S^1)=E$, then since $\Phi$ is a homeomorphism we have $E \cong S^1 \times \mathbb{R}$. Then if I can show they can't be homeomorphic to each other, the result follows.

But is it true that $\pi^{-1}(S^1)=E$ and mobius strip can't be homeomorphic to the unit cylinder?

bbw
  • 1,273

1 Answers1

1

For (a), $q: \mathbb{R}^2 \rightarrow E$ in the problem corresponds to $f: E \rightarrow M$ in Proposition 4.40. So there really is no problem.

For (b), for $p$ in $S$, any interval $U$ containing of $p$ that is of length, e.g. 0.5, will do. The local trivialization should be fairly obvious.

For (c), See this.