Let $X$ be a nice variety resp. a manifold over the complex numbers. One defines a connection on a vector bundle $V$ on over $X$ as a $\mathbb C-$linear sheaf homomorphism

$\nabla : V\rightarrow V\otimes \Omega^1$

which satisfies the Leibniz rule.

I have read that this is equivalent to giving for each local vector field $Y\in Der_{\mathbb C}(\mathcal O_X)$ a $\mathbb C-$ linear sheaf homomorphism

$\nabla_Y : V \rightarrow V$


(1) Leibniz rule

(2) $\nabla_{fY+gZ}=f\nabla_Y+g\nabla_Z$ for $f,g \in \mathcal O_X$ and $Y,Z$ local vector fields.

I can prove that each connection in the first sense implies a connection in the second sense. But I don't see how you get from the datum of thhe $\nabla_Y$ a connection in the first sense.

Remark: By a vector field I understand a linear derivation of the structure sheaf into itself.


Using the fact that $\Omega^1$ is $\mathcal{O}_X$-locally free and dual to the sheaf of vector fields you should be able to prove $Hom_{\mathcal{O}_X}(\Theta_X,\mathcal{E}nd_\mathbb{C}(V)) = Hom_\mathbb{C}(V,V\otimes_{\mathcal{O}_X} \Omega^1)$. In local complex coordinates $(x_1,\ldots,x_n)$, $\{\nabla_Y\}$ is mapped to $\nabla: V \to V\otimes \Omega^1$ defined by $$ \nabla(v) = \sum_{i=1}^n \nabla_{\partial_i}(v) \otimes dx_i $$ (This the analogue of $df = \sum_i \frac{\partial f}{\partial x_i} dx_i$). Condition (2) ensures that this formula does not depend on the choice of coordinates.

  • $\begingroup$ Hi YBL, thanks for your enlightening answer. But I think you mean "this formula does not depend of..."? $\endgroup$ – Veen Oct 23 '11 at 14:07
  • $\begingroup$ Of course. I edited it. $\endgroup$ – AFK Oct 23 '11 at 15:44

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.