arXiv · math/0312323
On the multiplicity of the hyperelliptic integrals
Abstract
Let $I(t)= \oint_{δ(t)} ω$ be an Abelian integral, where $H=y^2-x^{n+1}+P(x)$ is a hyperelliptic polynomial of Morse type, $δ(t)$ a horizontal family of cycles in the curves $\{H=t\}$, and $ω$ a polynomial 1-form in the variables $x$ and $y$. We provide an upper bound on the multiplicity of $I(t)$, away from the critical values of $H$. Namely: $ord\ I(t) \leq n-1+\frac{n(n-1)}{2}$ if $°ω<°H=n+1$. The reasoning goes as follows: we consider the analytic curve parameterized by the integrals along $δ(t)$ of the $n$ ``Petrov'' forms of $H$ (polynomial 1-forms that freely generate the module of relative cohomology of $H$), and interpret the multiplicity of $I(t)$ as the order of contact of $γ(t)$ and a linear hyperplane of $\textbf C^ n$. Using the Picard-Fuchs system satisfied by $γ(t)$, we establish an algebraic identity involving the wronskian determinant of the integrals of the original form $ω$ along a basis of the homology of the generic fiber of $H$. The latter wronskian is analyzed through this identity, which yields the estimate on the multiplicity of $I(t)$. Still, in some cases, related to the geometry at infinity of the curves $\{H=t\} \subseteq \textbf C^2$, the wronskian occurs to be zero identically. In this alternative we show how to adapt the argument to a system of smaller rank, and get a nontrivial wronskian. For a form $ω$ of arbitrary degree, we are led to estimating the order of contact between $γ(t)$ and a suitable algebraic hypersurface in $\textbf C^{n+1}$. We observe that $ord I(t)$ grows like an affine function with respect to $°ω$.
Explore related subjects
Keep this discovery
Explore connections, maps & timelines
Claire Moura. 2003-12-17. On the multiplicity of the hyperelliptic integrals. https://doi.org/10.1088/0951-7715%2F17%2F6%2F004
Cite the original work for its findings. Save a collection to share your selection of sources.