14

I am reading Hartshorne's Algebraic Geometry, II.6 about Cartier Divisor. It is defined to be the global section of the sheaf $K^*/O^*$. Then it said: " thinking of the properties of the quotient sheaves, we see that a Cartier divisor can be described by an open cover $U_i$ of $X$, and for each $i$ an element $f_i \in \Gamma(U_i, K^*)$ such that for $i,j$, $f_i/f_j \in \Gamma(U_i \cap U_j, O^*)$. I couldn't make sense of this part. Can someone explain more for me please? So, we need to take the presheaf $U \rightarrow K^*(U)/O^*(U)$, and make it a sheaf? I am always uncertain about how to deal with sheafification.

Long
  • 1,630

1 Answers1

12

This is a nice example of describing a global section of the sheafification of the (quotient) presheaf by local data that glues. In other words, the sheaf quotient $\mathscr K^\ast / \mathscr O^\ast$ is the sheafification of the presheaf quotient $U\mapsto \mathscr K^\ast(U) / \mathscr O^\ast(U)$, and by the properties of sheaves and definition of sheafification, we know that a global section of $\mathscr K^\ast / \mathscr O^\ast$ is determined by a collection of local sections of the presheaf quotient that glue on overlaps.

Now a local section of the presheaf quotient is just an element $f_i\pmod{ \mathscr O^\ast(U_i)}$, and asking that a collection $\{ f_i \}$ glues is the same as requiring that $f_i = f_j \pmod{\mathscr O^\ast(U_{ij})}$, which is the same as $f_i/f_j \in\mathscr O^\ast(U_{ij})$.

Andrew
  • 11,179
  • 2
    So, is it the fact that, if I have a presheaf F, then the global section of $F^$ will be the collection of local sections of F that gluable ? Is this true for $F^(U)$ for U any open set? I was thinking the global section of $F^*$ should be more complicated than that. – Long Aug 03 '13 at 20:55
  • Dear @LongMai, the sheafification is the smallest sheaf that has all the local sections of the presheaf, so yes, a section of $F^\ast$ on $U$ equals a collection of local sections of the presheaf on an open covering of $U$, which is basically what Hartshorne's direct construction of $F^\ast$ says on p. 64. – Andrew Aug 03 '13 at 21:14
  • @Hugo Your statement holds only when the presheaf is separated. – QU Binggang Jun 17 '20 at 02:52