Counting conics in complete intersections (Q977136): Difference between revisions

Let \(X\) be a smooth projective variety containing a quasi-line \(l\). It is true that \(e_0(X,l)\leq e(X,l)\) where \(e(X,l)\) is the number of quasi-line which are deformations of \(l\) and pass through two given general points of \(X\) and \(e_0(X,l)\) is the number of quasi-lines which are deformations of \(l\) and pass through a general point \(x\) of \(X\) with a given general tangent direction at \(x\). In general the inequality may be strict. The main goal of the paper is to give an elementary proof of the following result. If \(X \subseteq {\mathbb{P}}^{n+r}\) is a general smooth \(n\)-dimensional complete intersection of multi-degree \((d_1,\dots,d_r)\) with \(d_1+\dots+d_r=\frac{n+1}{2} + r\), then (i) the family of conics contained in \(X\) is a nonempty, smooth and irreducible component of the Chow scheme \(\mathcal{C}(X)\); (ii) a general conic \(C\) contained in \(X\) is a quasi-line of \(X\) and \(e_0(X,C)=e(X,C)=\frac{1}{2} \prod_{i=1}^r(d_i - 1)!d_i!\). The formula \(e(X,l)=\frac{1}{2} \prod_{i=1}^r(d_i - 1)!d_i!\) was obtained also by \textit{A. Beauville} [Mat. Fiz. Anal. Geom. 2, No. 3-4, 384-398 (1995; Zbl 0863.14029)] in a different way, as a conseguence of his computation of the quantum cohomology algebra \(H^*(X,{\mathbb{Q}})\) of a complete intersection.