Holomorphic functions on bundles over annuli (Q930539)

This paper in several complex variables makes a contribution to the so-called Serre problem which asks whether the total space of a holomorphic fiber bundle is a Stein manifold if both its base and fibers are. There are instances when the answer is Yes and others when it is No. The main theorem of this article is the following. Main theorem. The special total space \(E_m(D,M)\) is Stein if and only if \(m\log\varrho(M)\leq2\pi^2\). Here the base is the annulus \(1<| W| <e^m\) of modulus \(m>0\) in the \(W\)-plane, \(S_m\) is the strip \(| \text{Im}\,w| <m/(4\pi)\) in the \(w\)-plane, that is the universal covering of the above annulus under the map \(W=e^{2\pi i\,w-m/2}\), \(M\) is an invertible \(d\times d\)-matrix, \(d\geq2\), such that the entries of \(M\) and \(M^{-1}\) are integers, \(\varrho(M)\) is the maximum of the spectral radii of \(M\) and \(M^{-1}\), \(D\) is a Stein Reinhardt domain in the space of \((z_1,\dots,z_d)\in({\mathbb C}\setminus\{0\})^d\) that is invariant under the action \(z\mapsto\big(\prod_{j=1}^dz_j^{M_{ij}}\big)_{i=1}^d=z^M\) of the matrix \(M\) on \(z\), and \(E_m(D,M)\) is the quotient space of \(S_m\times D\) under the cyclic group generated by the map \((w,z)\mapsto(w+1,z^M)\). The proof expands invariant holomorphic functions \(f{:}\,S_m\times D\to{\mathbb C}\) into Hartogs Laurent series \(f(w,z)=\sum_k g_k(w)z^k\), where the \(g_k\in{\mathcal O}(S_m)\) are holomorphic functions on the strip \(S_m\) indexed by row vectors \(k\in{\mathbb Z}^d\), which due to the invariance of \(f\) satisfy that \(g_{kM^j}(w)=g_k(w+j)\). It also analyzes the orbits of the matrix \(M\) on the space \({\mathbb Z}^d\) of indices \(k\) using linear algebra such as eigenvalues and eigenvectors and Jordan normal forms. Making use of a Phragmén-Lindelöf type of maximum principle the author concludes that some \(g_k\) must vanish identically under some circumstances, which then makes \(E_m(D,M)\) not Stein. The paper is well written and pleasant to read (except perhaps for a profusion of bold italics in mathematical formulas that call for especially sharp eyes).