Mathematics Stack Exchange is a question and answer site for people studying math at any level and professionals in related fields. Join them; it only takes a minute:

Sign up
Here's how it works:
  1. Anybody can ask a question
  2. Anybody can answer
  3. The best answers are voted up and rise to the top

Suppose I have two Euclidean metrics $\mu_1, \mu_2$ on a given vector bundle $\xi$. Does anyone know of necessary and/or sufficient conditions to ensure that there is a homeomorphism $\phi: E(\xi) \to E(\xi)$ between fibers such that $\mu_1 \circ \phi = \mu_2$?

I vaguely recall results relating to this question, but it's been a while since I studied algebraic topology.

share|cite|improve this question

It turns out that there is no additional requirement: a Euclidean structure on a vector bundle $\xi$ is unique up to isometric isomorphism of Euclidean vector bundles. This is Exercise 2-E of Milnor and Stasheff's Characteristic Classes.

In fact, we can say more.

Theorem. Let $\xi_1 = (E_1, p_1, B)$ and $\xi_2 = (E_2, p_2, B)$ be two vector bundles over a common base space $B$, and let $\mu_1$ be a Euclidean metric on $\xi_1$ and $\mu_2$ a Euclidean metric on $\xi_2$. If $\phi: \xi_1 \longrightarrow \xi_2$ is a vector bundle isomorphism, then $\phi$ is homotopic through vector bundle isomorphisms to an isometric isomorphism $\tilde{\phi}: (\xi_1,\mu_1) \longrightarrow (\xi_2,\mu_2)$.

Proof. First, suppose $\xi_1$ and $\xi_2$ are trivial bundles, so that the isomorphism $$\phi: B \times \mathbb{R}^n \longrightarrow B \times \mathbb{R}^n$$ is given by $$\phi(x,v) = (x,g(x)v)$$ for some map $$g: B \longrightarrow \mathrm{GL}(n).$$ Now $\mathrm{O}(n)$ is a deformation retract of $\mathrm{GL}(n)$, so $g$ is homotopic to some map $$h: B \longrightarrow \mathrm{O}(n).$$ Write $g_t$ for this homotopy, so that $g_0 = g$ and $g_1 = h$. Define $$\phi_t: B \times \mathbb{R}^n \longrightarrow B \times \mathbb{R}^n$$ by $$\phi_t(x,v) = (x,g_t(x)v).$$ Then $\phi_t$ is a homotopy of vector bundle isomorphisms from $\phi$ to $\tilde{\phi} = \phi_1$, and clearly $\tilde{\phi}$ is an isometric isomorphism of vector bundles.

For the general case where $\xi_1$ and $\xi_2$ are not trivial bundles, apply the above to a locally finite cover of $B$ that simultaneously trivializes $\xi_1$ and $\xi_2$. $~\Box$

share|cite|improve this answer
I don't see the last part of your argument. If you choose a locally finite cover $\{U_\alpha\}$ and apply the above you get that every $\phi_\alpha:=\phi|_{U_\alpha}$ is homotopic through isomorphisms to an isometry $\tilde \phi_\alpha$ on $U_\alpha$. But for $U_\alpha \cap U_\beta \neq \emptyset$ why would the isometries $\tilde \phi_\alpha$ and $\tilde \phi_\beta$ agree on the overlap? Of course you could paste the various $\tilde \phi_\alpha$ together with a partition of unity. This would yield a well-defined map, but one might lose the isometry property on the overlaps. – Meneldur May 6 '12 at 10:36

Your Answer


By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.