Let $M$ be a smooth manifold. For each coordinate chart $(U,\varphi)$ one can form the basis $\partial_1,...,\partial_n$ which are local vector fields defined on $U$. For fixed $x$ this forms a basis of $T_xM$ and $\partial_i|_x$ is obtained as the image of $e_i=(0,...,0,1,0,...,0)$ by the mapping which establishes the isomorphism beetween $\mathbb{R}^n$ and $T_xM$. More concretely, $\partial_i|_x$ is the equivalence class of the curve $\gamma$ where $\gamma(0)=\varphi^{-1}(\varphi(x)+te_i)$. So far we haven't used the Riemannian structure. Les us now suppose that $M$ is Riemannian manifold. I have some general questions about the orthogonality of these canonical basis and the domain over which it may hold:
1. In general $\partial_1,...,\partial_n$ need not to be orthogonal. However one can perform Gramm-Schmidt orthogonalization procedure to obtain local orthonormal frame $E_1,...,E_n$. The domain of this frame would be the same as for $\partial_i$'s. Therefore one can consider local orthonormal frames on the domain of each coordinate chart.

When is it possible to choose $E_i$'s in such a wat that the commutator $[E_i,E_j]$ vanish?

And from the other side

  1. Can we assume without the loss of generality that in a given point $x \in M$ the local frame $\partial_1|_x,...,\partial_n|_x$ is orthogonal?
  1. By the Frobenius theorem, if $[E_i,E_j]$ vanishes for every $i,j$, then there is a parametrization $\varphi$ whose induced $\partial_i$'s are the $E_i$'s. It then follows that $\varphi$ is an isometry, and the curvature must vanish all over the range of $\varphi$.

  2. Yes, just take any parametrization and compose it with an appropriate linear transformation of $\mathbb{R}^n$.

| cite | improve this answer | |
  • $\begingroup$ Thank you! Could mi also give the reference for results which you invoked in 1? I only know the Frobenius theorem which relates the notion of integrability of a bundle with the notion of foliation. $\endgroup$ – truebaran Sep 26 '15 at 7:40
  • $\begingroup$ @truebaran I guess you can find all the necessary details in Introduction to Smooth Manifolds by Lee. $\endgroup$ – Amitai Yuval Sep 26 '15 at 8:58

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.