Numerical invariants of surfaces in \({\mathbb P}^4\) lying on small degree hypersurfaces (Q1736078)

By a famous theorem of \textit{G. Ellingsrud} et al. [Invent. Math. 95, 1--11 (1989; Zbl 0676.14009)] for all \(\alpha <6\) there exists a constant \(d_1(\alpha)\) such that all smooth surfaces in \({\mathbb P}^4\) with \(K^2 \leq \alpha \chi \left( {\mathcal O}\right)\) have degree smaller than \(d_1(\alpha)\). One of the ingredients of the proof is that if \(K^2 < 6 \chi \left( {\mathcal O}\right)\) then the surface is contained in a quintic threefold. The main result of this paper is a generalization of this result to higher values of \(\alpha\) under the assumption that such a quintic threefold do exist. Theorem. Assume \(\alpha >6\). Then there is a number \(d(\alpha,5)\) such that every smooth surface which satisfies \(K^2 \leq \alpha \chi \left( {\mathcal O}\right)\) and lies on a quintic hypersurface of \({\mathbb P}^4\) has degree at most \(d(\alpha,5)\). The notation \(d(\alpha,5)\) comes from a question of the author, asking if the result generalizes to surfaces contained in a hypersurface of degree \(m\) of \({\mathbb P}^4\), in which case he would call the constant \(d(\alpha,m)\). The author consider a smooth surface \(X\subset {\mathbb P}^4\) and set \(m\) for the minimal degree of a hypersurface containing \(X\). Then he constructs a nontrivial extension class \[ \xi \in \mathrm{Ext}^1 (\Omega_X,{\mathcal O}_X(-K_X+(m-5)H)) \] This allows the author to obtain some numerical constraints on the invariants of a surface contained in a quartic threefold or a quintic threefold of \({\mathbb P}^4\) that are of independent interest. The proof of the theorem uses these constraints and Ellinsgrud-Peskine techniques.