I am looking for a reference (or an easy explanation) for the openness of the stable locus of a holomorphic family of (holomorphic) vector bundles on a compact Riemann surface parametrized by a (compact) complex manifold. For me, a holomorphic family of vector bundles on a compact Riemann surface $X$ parametrized by a (compact) complex manifold is just a holomorphic vector $E$ bundle over $T \times X$.

There is a proof in Narasimhan and Seshadri's paper "Stable and Unitary Vector Bundles on a Compact Riemann Surface" but the proof there depends on their proof of the unrelated theorem that describes stable vector bundles on a compact riemann surface via certain unitary representations of suitably defined fuchsian groups.

If there is a simpler proof in the case that all parametrized bundles are of degree zero I would like to know it.

EDIT: Note that I do not want to assume that $T$ and the vector bundle $E$ over $T \times X$ are algebraic. So e.g. Huybrechts-Lehn or Le Potier's "Lectures on Vector Bundles" aren't of any help to me, I think.

  • $\begingroup$ Retagged reference-request, hope you don't mind $\endgroup$ Jun 29, 2011 at 18:34

1 Answer 1


I think this question is discussed in Kobayashi's "Differential geometry of complex vector bundles", at least the result follows implicitly.

One way to see is as follows: For simplicity, consider holomorphic rank 2 bundles $V$ of degree 0. They are not stable if there exists a holomorphic $f\colon L\to V$ of a holomorphic line bundle $L$ of degree $0.$ If you think of a holomorphic bundle as given by a holomorphic structure $\bar\partial,$ then the result follows from the observation, that, for a family of Fredholm operators (like the $\bar\partial$ on $L^*\otimes V,$ where you vary the holomorphic structures on $L$ and $V$) the minimal kernel dimension is attained on an open subset.

  • $\begingroup$ Hi, Sebastian, sorry for answering only now. What happens for rank greater than two? $\endgroup$
    – bavajee
    Oct 6, 2011 at 14:22

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

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