The surface \(C\)-\(C\) on Jacobi varieties and 2nd order theta functions (Q1189612)

The main results are: Theorem 1. Let \(C\) be an irreducible smooth complete curve of genus \(g\geq 5\) over \(\mathbb{C}\), \(\theta\subset JC\) be the theta-divisor on the Jacobian \(JC\) of the curve \(C\) and \(C-C\) be the surface in \(JC\) consisting of the points \(x-y\in JC\), for all \(x,y\in C\). Then, the following equality holds: \(C-C=\{a\in JC|\;a+\text{sing} \theta\subset\theta\}\), where \(\text{sing} \theta\) is the subset of singular points in \(\theta\). Theorem 2. Let \(\hat\theta=\{\zeta_{g-1}\in\text{Pic}^{g-1}(C)|\;h^ 0(\zeta_{g-1})\geq 1\}\subset\text{Pic}^{g-1}(C)\) be the canonical model of the theta-divisor and \(\xi':=K-\xi\in\text{Pic}^{g- 1}(C)\) for all \(\xi\in\text{Pic}^{g-1}(C)\), where \(K\) is the canonical class of \(C\). Then, if \(g\geq 5\) and the curve \(C\) is not trigonal: \(C- C=\bigcap_{\xi\in W^ 1_{g-1}}(\hat\theta_ \xi+\hat\theta_{\xi'})\), where \(W^ 1_{g-1}=\{\zeta_{g- 1}\in\text{Pic}^{g-1}(C)|\;h^ 0(\zeta_{g-1})\geq 2\}\), and \(\hat\theta_{-\xi}\) denotes the translate of \(\hat\theta\) by the divisor \((-\xi)\in\text{Pic}^{g-1}(C)\). If \(C\) is trigonal, then \(C- C=(W^ 0_ 3-g^ 1_ 3)\cup(g^ 1_ 3-W^ 0_ 3)\). Corollary. If \(g\geq 5\), then \(C-C={\displaystyle{\bigcap_{{D\in| 2\theta|\atop \mu_ 0(D)\geq 4}}}}D\), where \(| 2\theta|\) denotes the complete linear system of the divisor \(2\theta\) and \(\mu_ 0(D)\) denotes the multiplicity of \(D\) at zero. Thus, the author gives an exhausting answer to the hypothesis by \textit{B. van Geemen} and \textit{G. van der Geer} [Am. J. Math. 108, 615-641 (1986; Zbl 0612.14044)]: Let \(\Gamma_{00}\) be the vector space of sections from \(H^ 0({\mathcal O}_{JC}(2\theta))\), which vanish with multiplicity \(\geq 4\) at \(0\in JC\), and \(F=\{x\in JC|\;s(x)=0\) for all \(s\in\Gamma_{00}\}\), then \(F=\{x-y|\;x,y\in C\}\) in the case \(g\geq 5\) (the cases \(g=1,2,3\) and 4 are considered in the work immediately). -- The author gives exhausting answers also to the following questions: 1. (Mumford) Let \(D\) be the class of divisors of degree 0 on \(C\) such that for all divisors \(E\) of degree \(g-1\), for which \(| E|\) is a pencil then either \(| D+E|\neq\emptyset\) or \(|- D+E|\neq\emptyset\). Is \(D\equiv a-b\) for some \(a,b\in C\) then? 2. (The author) Let \(D\) be the class of divisors of degree 0 on \(C\) such that for all divisors \(E\) of degree \(g-1\), for which \(| E|\) is a pencil, \(| D+E|\neq\emptyset\). Is \(D\equiv a-b\) for some \(a,b\in C\) then?