Maximal rank of space curves in the range A (Q1999423)

A projective space curve \(C\subset\mathbb P^3\) has \textit{maximal rank} if all the natural restriction maps \(H^0(\mathcal O_{\mathbb P^3}(m))\to H^0(\mathcal O_C(m))\) are either injective or surjective. A long-standing conjecture predicts the existence of components of the Hilbert scheme whose general element corresponds to a smooth, irreducible, maximal rank space curve of genus \(g\) and degree \(d\), when \(g\) is not too big with respect to \(d\). Here the notion of \textit{``not too big''} is related with the fact that if \(d,g\) grow and \(g\sim O(d^\alpha)\) with \(\alpha>(3/2)\), then the conjecture cannot hold asymptotically as soon as \(h^1(N_C)=0\), i.e. as soon as the Hilbert scheme is smooth of the expected dimension around \(C\). Thus, the authors concentrate in looking for a constant \(K\) such that the conjecture holds in the range \(g\leq Kd^{3/2}\). Indeed, they find constants \(K, \epsilon\) such that the conjecture holds (and moreover the curves satisfy \(h^1(N_C(-1))=0\)) in the range \(g\leq Kd^{3/2}-6\epsilon d\). The proof is obtained by induction, by smoothing reducible curves whose components satisfy suitable cohomological properties. The procedure is particularly devoted to solve the case \(d< g+3\), which was missing in the previous literature, the delicate point being that for \(d< g+3\) the Hilbert scheme of smooth curves often turns out to be reducible.