No label defined (Q3973290)

[For the entire collection see Zbl 0722.00006.] The Hartshorne-Rao module of a curve \(C\subset\mathbb{P}^ 3_ k\) is defined as \(M_ C=\sum_{n\in\mathbb{Z}}H^ 1({\mathcal I}_ C(n)))\), where \({\mathcal I}_ C\) denotes the ideal sheaf of \(C\). Let \(C_ i\), \(i=1,2\), be two curves. Is there a construction of a curve \(C\) (in terms of \(C_ i)\) such that \(M_ C\) is isomorphic to some shift of \(M_{C_ 1}\oplus M_{C_ 2}\)? This is solved by \textit{P. Schwartau} in his unpublished thesis ``Liaison addition and monomial ideals'' [Ph. D. thesis, Brandeis University 1982), see also \textit{J. Stückrad} and \textit{W. Vogel} [``Buchsbaum rings and applications. An interaction between algebra, geometry, and topology'', VEB Deutscher Verlag der Wissenschaft (Berlin 1986; Zbl 0606.13017); published simultaneous by Springer Verlag)] for a reproduction of his arguments. In this paper the authors generalize Schwartau's result in several senses: (1) There is a liaison addition in any codimension in \(\mathbb{P}^ n_ k\), \(n\geq 3\), of schemes of mixed codimensions. --- (2) There is an addition of any number of schemes. --- (3) There is no need that the schemes are attached via a complete intersection (as in Schwartau's result). The basic idea is the construction of a scheme \(Z\) consisting set- theoretically of the union of certain given schemes such that the cohomology of \(Z\) is related to those of the given schemes in terms of a long exact sequence. This far reaching generalization yields, in the case the underlying scheme is arithmetically Cohen-Macaulay, that the cohomology of \(Z\) is the shift of the direct sum of the involved schemes. There are a number of applications of this clever technique to the construction of arithmetically Buchsbaum schemes, the double linkage, Hilbert functions, etc.