On the degree and genus of smooth curves in a projective space (Q2277044)

A long-standing problem in projective geometry is the determination of all possible values for the genus \(g\) of smooth curves of given degree \(d\) lying inside some projective space \({\mathbb{P}}^ n\). The classical Castelnuovo bound provides a sharp limit to the range in which g may vary, however inside it some gaps occur, expecially around the maximal value. - Complete solutions to the problem are known for \(n=3\) (Gruson- Peskine) and \(n=4, 5\) (Rathmann); in both cases, the existence of curves with all the expected genera, is established on rational surfaces of low degree. This paper is concerned with the case \(n\geq 6\). In the first section, the author provides two functions a(d,n) and b(d,n) such that, for large d, all numbers g sitting inside a(d,n), b(d,n) are allowed as genus of a smooth curve of degree d in \({\mathbb{P}}^ n\); such curves are determined through the study of the Picard group of rational surfaces with hyperelliptic hyperplane sections. The result improves previous works of Rathmann and Ciliberto-Sernesi and widens the range where the existence of curves of fixed genus is known. The upper bound is in fact sharp: as soon as \(g>b(d,n)\), gaps in the existence of curves of g occur. The second section is devoted to the cases \(n=6, 7\). The author proves, for these 2 values of n, a conjecture of J. Harris meaning that curves of `high' genus must lie on surfaces of `small' degree. All the results together, with the aid of a computer program for the examination of few scattered cases, allow the author to give a complete solution for the problem of the genus of smooth curves in \({\mathbb{P}}^ 6\). - In \({\mathbb{P}}^ 7\), only a finite number of cases for \(39<d<64\) remains unsolved.