3
$\begingroup$

Let $X$ be a smooth projective variety over $\mathbb{C}$. Let $D$ be a Cartier divisor on $X$. If $C$ is a curve on $X$ such that the intersection $C\cdot D <0$, we have $C \subseteq \mathbf{B}(D)$, where $$\mathbf{B}(D):= \bigcap_{m \in \mathbb{N}, F\in |mD|}F$$ is the stable base locus of $D$.

I want to know if there is any result related to the inverse direct. To be precise, I want to know if the followings are true or false:

(1)For a general point of $\mathbf{B}(D)$, is there a curve $C$ passing through that point with $C\cdot D<0$?

(2)The union of all curves with negative intersections with D has the same dimension as $\mathbf{B}(D)$, that is $$\dim \mathbf{B}(D) = \dim \bigcup_{C, C\cdot D < 0}{C}\qquad?$$

$\endgroup$

1 Answer 1

7
$\begingroup$

Both are false. Let $X$ be the blow-up of $\mathbb P^3$ at $8$ very general points, and let $D = 2H - \sum E_i$ be the linear system of quadrics through the $8$ points. There is a pencil of such quadrics. The stable base locus of $D$ is a genus $1$, degree $4$ curve (obtained as the intersection of any two quadrics). But $D$ is nef.

$\endgroup$
3
  • 2
    $\begingroup$ You could a little closer if you instead considered the diminished base locus $\mathbf B_-(D) = \bigcup_{\text{$A$ ample}} \mathbf B(D+A)$, which is a subset of $\mathbf B(D)$. Both are probably still false, but it's at least true that if $\mathbf B_-(D)$ is non-empty, $D$ is not nef. $\endgroup$
    – user47305
    Jan 4, 2015 at 16:20
  • $\begingroup$ Thank you for your answer! But I did not quite follow your example: In your example, $X$ is a toric variety, hence any nef divisor is base point free (in particular $D$ is basepoint free if it is nef). Where did I understand wrong? $\endgroup$
    – Li Yutong
    Jan 5, 2015 at 6:12
  • 2
    $\begingroup$ $X$ is not toric! You can blow up $4$ general points in $\mathbb P^3$ and get a toric variety, but no more ([1,0,0,0] etc are torus-invariant points, but that's it). However, it's still a Mori dream space for up to $7$ points, and so nef divisors are semiample. One you hit $8$, everything goes bad; it's a lot like $\mathbb P^2$ blown up at $9$ points versus $8$. $\endgroup$
    – user47305
    Jan 5, 2015 at 6:58

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.