Cone theorems and cyclic covers (Q4026027)

From MaRDI portal

!

This is the item page for this Wikibase entity, intended for internal use and editing purposes. Please use the normal view instead:

scientific article; zbMATH DE number 124403
Language Label Description Also known as
default for all languages
No label defined
    English
    Cone theorems and cyclic covers
    scientific article; zbMATH DE number 124403

      Statements

      0 references
      21 February 1993
      0 references
      cone theorem
      0 references
      formal completion
      0 references
      cyclic covering
      0 references
      normal threefolds
      0 references
      deformations of curves
      0 references
      log-terminal singularities
      0 references
      Cone theorems and cyclic covers (English)
      0 references
      [For the entire collection see Zbl 0744.00034.] This note is a continuation of the author's paper in J. Algebr. Geom. 1, No. 2, 293-323 (1992; see the preceding review). The aim is again to prove various forms of the cone theorem using methods similar to the original geometric arguments of \textit{S. Mori} [cf. Ann. Math., II. Ser. 110, 593-606 (1979; Zbl 0423.14006) and 116, 133-176 (1982; Zbl 0557.14021)]. The basic idea is the same as in the author's paper cited above. Instead of deforming a curve \(C \subset X\) directly, we construct a covering \(Y \to X\) and deform another morphism \(D \to Y\) where \(D\) is a suitable covering of \(C\). In the author's mentioned paper, \(Y\) was a bug- eyed cover of \(X\), therefore a nonseparated algebraic space. The method of the present article is to replace \(X\) by its formal completion along \(C\) and then find a suitable cyclic covering \(Y\) of this formal scheme. The new method has both disadvantages and advantages. It can handle only quotients by cyclic groups and not by arbitrary groups. The method however is still strong enough to recover the most general cone theorem for normal threefolds. The main advantage is that it can be used to investigate deformations of curves that are contained in the singular locus of \(X\). In fact this method shows that the deformation problem of \(C \to X\) where the canonical class of \(X\) is \(\mathbb{Q}\)-Cartier can be reduced to another deformation problem \(D\to Y\) where the canonical class of \(Y\) is Cartier. Theorem. Let \(X\) be a projective threefold over a field of characteristic zero. Assume that \(K_ X\) is \(\mathbb{Q}\)-Cartier and that \(X\) has only log- terminal singularities with the exception of finitely many points. Then (1) The set of extremal rays is locally finite in the open subset \(\{z \in N_ 1 (X) | z \cdot K_ X < 0\}\). (2) For every extremal ray \(R\) the locus of \(R\) is covered by rational curves \(L_ R\) such that \([L_ R] \in R\) and \(-6 \leq L_ R \cdot K_ X < 0\). The locus of a ray \(R\) is defined to be the set theoretic union of the closed points of all curves \(D \subset X\) such that \([D] \in R\). Presumably the 6 in the statement can be replaced by 4.
      0 references

      Identifiers

      0 references
      0 references
      0 references
      0 references