On nonprimitive Weierstrass points (Q1755543)

The paper by establishes utilities of numerical semigroups in studying stratifications of \(\mathcal{M}_{g, 1}\). For a genus \(g\) numerical semigroup \(S\), by relating the effective weight of \(S\) and the codimension of some strata of \(\mathcal{M}_{g, 1}\), the paper reports strong results in concern with components of \(\mathcal{M}^S_{g, 1}\). In section \(2\), Pflueger studies the cases where \(\mathrm{codim} \mathcal{M}^S_{g,1}\) is known, with insistence on clear proofs in the case of numerical semigroup with two generators. After reviewing the works of Eisenbud-Harris, he relates, through an inequality, the effective weight of a numerical semigroup of genus \(g\) to the Deligne bound. He, moreover, determines the possibilities for the numerical semigroups \(S\) where this inequality turns to be an equality. If \(S\) is a numerical semigroup with two number of generators or if \(S=N_{d, \delta}\), where \(S=N_{d, \delta}\) is an specific type of numerical semigroups, then it is proved in Propositions \((2.14)\) and \((2.18)\) that \(\mathcal{M}^S_{g, 1}\) is irreducible of codimension equal to the effective weight of \(S\). He finishes section two by a nice example, and indeed the smallest genus, of a numerical semigroup for which \(\mathrm{codim} \mathcal{M}^S_{g, 1}\) differs from the effective weight of \(S\). The main result of section \(3\) is a characteristic free proof of the Eisenbud-Harris' ``regeneration theorem''. In order to do this, he distinguishes an open subset of the scheme of limit linear series consisting of the so called ``refined limit linear series''. In section \(4\), the author refines his study once more by introducing the effective subsequences of a numerical semigroup of genus \(g\). If \(T\) is an effective subsequence of a numerical semigroup \(S\) of genus \(g\), then he obtains an equivalent statement for existence of \textit{effectively proper} points for \(\mathcal{M}^S_{g, 1}\) in terms of the existence of \textit{dimensionally proper} points for \(\tilde{\mathcal{G}}^r_{g, d}(d-T)\). Using the machinery produced in section \((4)\), the author proves that, for \(S\) a numerical semigroup of genus \(g\), if \(\mathcal{M}^S_{g,1}\) is nonempty and if \(X\) is an irreducible component of it, then \[ \dim X \geq \dim \mathcal{M}_{g, 1}-\mathrm{ewt}(S), \] where \(\mathrm{ewt}(S)\) denotes the effective weight of \(S\). This is the first main result of this paper. The second main result of the paper states that if \(S\) is a genus \(g\) numerical semigroup with \(\mathrm{ewt}(S)\geq g-2\), then \(\mathcal{M}^S_{g, 1}\) has an effectively proper component; furthermore, if \(\mathrm{char} k=0\), then the assertion holds for all numerical semigroups with \(\mathrm{ewt}(S)\geq g-1\). This is proved in section \(6\) by using the machinery produced in previous sections alongside an analogue of Eisenbud-Harris' Proposition \(5.2\) in