Integrable systems and algebraic surfaces (Q1918369)

The paper is devoted to thorough analysis of algebraically integrable Hamiltonian systems (i.e., to complex integrable systems for which the compactified and desingularized joint level sets of the Hamiltonians are Abelian varieties with the linear structures given by the Hamiltonian [cf. \textit{P. van Moerbeke}, Proc. Symp. Pure Math. 49, Pt. 1, 107-131 (1989; Zbl 0688.70012)]), especially to the particular class of examples determined by some natural systems on the coadjoint orbits of a loop algebra of polynomials in one variable \(\lambda\) with values in a finite-dimensional semisimple Lie algebra [cf., e.g., \textit{M. Adler} and \textit{P. van Moerbeke}, Adv. Math. 38, 267-317 (1980; Zbl 0455.58017) and ibid., 318-379 (1980; Zbl 0455.58010)]. The reduced coadjoint orbit has a description as an open set of the union in the family of Jacobians, corresponding to a family of curves in the \((z,\lambda)\)-plane, and its symplectic geometry is related to that of the plane. The author deals with the following problems: Which integrable systems fit into the loop algebra framework, if not, are there other algebraic surfaces which can be invariantly associated to them, and classification of these surfaces. We cannot refer the deep results here. In rough terms, the vanishing of a certain invariant of the integrable system ensures a surface \(Q\) generalising the \((z,\lambda)\)-plane, in the loop case this \(Q\) is rational, and the classification is made according to whether certain curves \(S_h\) in \(Q\) (determined by the Lagrangian fibration of the integrable system) are hyperelliptic or not. Several examples related to recent work by M. R. Adams, J. Harnad, the author (coadjoint orbits), N.J. Hitchin (moduli spaces of stable \({\mathfrak {gl}}(r)\) pairs), E. K. Sklyanin (unusual brackets), S. Makai (\(K-3\) surfaces), and the genus two case are discussed. [The erratum concerns Proposition 3.5 of the paper].