On the homology of the moduli space of stable curves (Q1064368)

Stable curves were introduced by \textit{P. Deligne} and \textit{D. Mumford} in Publ. Math., Inst. Haut. Étud. Sci. 36 (1969), 75-109 (1970; Zbl 0181.488). Their moduli space \(\bar {\mathcal M}_ g\) is a compactification of the classical moduli space \({\mathcal M}_ g\) of Riemann surfaces. Both spaces have dimension 3g-3 and the locus \({\mathcal D}=\bar {\mathcal M}_ g- {\mathcal M}_ g\) is the sum of \(1+[g/2]\) divisors of \(\bar {\mathcal M}_ g\). These define homology classes in \(H_{6g-8}(\bar {\mathcal M}_ g;{\mathbb{Q}})\); the Weil-Peterson Kähler form \(\omega\) on \({\mathcal M}_ g\) extends to a closed form on \(\bar {\mathcal M}_ g\) and so defines, by Poincaré duality, another class in \(H_{6g-8}(\bar {\mathcal M}_ g;{\mathbb{Q}})\). It is proved in this paper that these \(2+[g/2]\) cycles essentially form a basis of \(H_{6g-8}(\bar {\mathcal M}_ g;{\mathbb{Q}})\), if \(g>2\) (theorem 5.1). From this result, the author deduces interesting consequences about the form \(\omega\). The main technique is the construction of \(2+[g/2]\) analytic 2-cycles in \(\bar {\mathcal M}_ g\) and the computation of their intersection pairing with the above \(2+[g/2]\) classes in \(H_{6g-8}({\mathcal M}_ g;{\mathbb{R}})\), which turns out to be a non-singular pairing. This technique for constructing analytic cycles can also be used to construct higher dimensional homology classes.