On the connectivity of the branch and real locus of M₀,[n+1]

From MaRDI portal
Publication:2331684



Abstract: If ngeq3, then moduli space mathcalM0,[n+1], of isomorphisms classes of (n+1)-marked spheres, is a complex orbifold of dimension n−2. Its branch locus mathcalB0,[n+1] consists of the isomorphism classes of those (n+1)-marked spheres with non-trivial group of conformal automorphisms. We prove that mathcalB0,[n+1] is connected if either ngeq4 is even or if ngeq6 is divisible by 3, and that it has exactly two connected components otherwise. The orbifold mathcalM0,[n+1] also admits a natural real structure, this being induced by the complex conjugation on the Riemann sphere. The locus mathcalM0,[n+1](mathbbR) of its fixed points, the real points, consists of the isomorphism classes of those marked spheres admitting an anticonformal automorphism. Inside this locus is the real locus mathcalM0,[n+1]mathbbR, consisting of those classes of marked spheres admitting an anticonformal involution. We prove that mathcalM0,[n+1]mathbbR is connected for ngeq5 odd, and that it is disconnected for n=2r with rgeq5 is odd.


Let \(\mathcal{M}_{g,[n+1]}\) be the moduli space of isomorphism classes of \((n + 1)\)-marked Riemann surfaces of genus \(g\), where \(g \geq 0\) and \(n \geq -1\) are integers such that \(3g - 2 + n > 0\). This space is a complex orbifold of dimension \(3g - 2 + n\). Its branch locus \(\mathcal{B}_{g,[n+1]} \subset \mathcal{M}_{g,[n+1]}\) consists of the isomorphism classes of those admitting non-trivial conformal automorphisms. It is an interesting problem to study connectivity in these spaces. In the particular case of \(n=-1\), \textit{G. Bartolini} et al. proved in [Ann. Acad. Sci. Fenn. Math. 38, 245--258 (2013; Zbl 1279.14032)] that \(\mathcal{B}_{g,[0]} \subset \mathcal{M}_{g,[0]} = \mathcal{M}_g\) is connected only for \(g \in \{3, 4, 13, 17, 19, 59\}\). The complex orbifold \(\mathcal{M}_{g,[n+1]}\) admits a natural antiholomorphic automorphism of order two (a real structure) which is induced by the usual complex conjugation. Let us consider the locus \(\mathcal{M}_{g,[n+1]}(\mathbb{R}) \subset \mathcal{M}_{g,[n+1]}\) of fixed points (the real points) of such a real structure which consists of the isomorphic classes of those admitting anticonformal automorphisms and consider the sublocus \(\mathcal{M}^{\mathbb{R}}_{g,[n+1]} \subset \mathcal{M}_{g,[n+1]}(\mathbb{R})\) of those classes having a representative admitting an anticonformal involution, or equivalently, the representative being definable over the real numbers. It has been proved that \(\mathcal{M}^{\mathbb{R}}_{g,[0]} \subset \mathcal{M}_{g}\) is connected and \(\mathcal{M}_{g,[0]}(\mathbb{R}) \subset \mathcal{M}_g\) is also connected but \(\mathcal{M}_{g,[0]}(\mathbb{R})\setminus \mathcal{M}^{\mathbb{R}}_{g,[0]}\) is not in general connected. In the paper under review, the authors deal with isomorphisms classes of \((n+1)\)-marked spheres. For each \(n \geq 3\), they study the connectivity of both spaces \(\mathcal{B}_{0,[n+1]}\) and \(\mathcal{M}^{\mathbb{R}}_{0,[n+1]}\). In Theorem 1, it is obtained that the branch locus \(\mathcal{B}_{0,[n+1]}\) is connected if either (i) \(n \geq 4\) is even or (ii) \(n \geq 6\) is divisible by 3. Otherwise, it has exactly two connected components. The results on the connectivity of \(\mathcal{M}^{\mathbb{R}}_{0,[n+1]}\) are given in the statement (4) of Theorem 2 in the introduction of the paper, but they become clearer as stated in three propositions in Section 4, where they are proved. Proposition 4: If \(n \geq 5\) is odd, then \(\mathcal{M}^{\mathbb{R}}_{0,[n+1]}\) is connected. Proposition 5: If \(n = 2r\), where \(r \geq 5\) is an odd integer, then \(M^{\mathbb{R}}_{0,[n+1]}\) is not connected. Moreover, for \(r = p\), where \(p\) is a prime integer, the real locus \(\mathcal{M}^{\mathbb{R}}_{0,[2p+1]}\) has exactly \((p - 1)/2\) connected components. Proposition 6: If \(n = 4p\), where \(p \geq 2\) is a prime integer, then \(\mathcal{M}^{\mathbb{R}}_{0,[n+1]}\) is not connected if and only if \(p \geq 7\). In the case \(n = 4r\), where \(r \geq 1\) is odd, but different from a prime, the authors observe in Remark 9 that \(M^{\mathbb{R}}_{0,[n+1]}\) is connected for \(r = 1, 9, 15, 21, 27, 33\) and it is not connected for \(r = 25, 35\). Summing up, the authors explicitly give the 32 values of \(n\in \{4, \dots, 100\}\) such that the real locus is not connected. Two of these values rise doubts in the opinion of the reviewer: the number \(84 = 4\cdot 21\), which contradicts what has been previously said for \(r=21\), and the number \(88=4\cdot 22\), not covered by the above results. In Subsection 1.1, the authors also give two results about hyperelliptic Riemann surfaces and generalized Fermat curves as two easy Corollaries from the above results.



Cites work









This page was built for publication: On the connectivity of the branch and real locus of \(\mathcal{M}_{0,[n+1]}\)

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q2331684)