4
$\begingroup$

I am wondering if the blow-up of $\mathbb{P}^5$ along three disjoints $\mathbb{P}^1$ (say in generic position) can be understood as a projective bundle over some nice (Fano?) variety.

If one considers instead the blow-up of $\mathbb{P}^5$ along the union of three $\mathbb{P}^3$ in generic position, then it is easily proved that it is the projectivization of $\mathcal{O}(-1,0,0) \oplus \mathcal{O}(0,-1,0) \oplus \mathcal{O}(0,0,-1)$ over $\mathbb{P}^1 \times \mathbb{P}^1 \times \mathbb{P}^1$.

So I thought that something similar could be true if the union of the three $\mathbb{P}^3$ was replaced by the union of three lines. I thought it could also be a $\mathbb{P}^2$-bundle over $\mathbb{P}^1 \times \mathbb{P}^1 \times \mathbb{P}^1$.

At least the Picard rank ($4$) and the rank of the K-theory ($18$) are the same for a $\mathbb{P}^2$-bundle over $\mathbb{P}^1 \times \mathbb{P}^1 \times \mathbb{P}^1$ and for the blow-up of $\mathbb{P}^5$ along $3$ disjoint lines.

Many thanks in advance.

EDIT : Following the suggestion of Borisov, I ask a weaker a question : Is this blow-up of $\mathbb{P}^5$ along three disjoint lines (in generic position) $K$-equivalent to a $\mathbb{P}^2$-bundle over $\mathbb{P}^1 \times \mathbb{P}^1 \times \mathbb{P}^1$?

$\endgroup$
2
  • $\begingroup$ I don't see any obvious maps to a $\mathbb P^1$ here. It would be more reasonable to ask whether this variety is a $K$-equivalent to a $\mathbb P^2$-bundle over $(\mathbb P^1)^3$. $\endgroup$ Commented Apr 21, 2014 at 2:23
  • $\begingroup$ Not birational, K-equivalent! All these varieties are rational, hence birational to each other. $\endgroup$
    – abx
    Commented Apr 21, 2014 at 10:15

2 Answers 2

5
$\begingroup$

Consider the six points $p_0=(1:0:0:0:0:0)$, $p_1=(0:1:0:0:0:0)$, and similarly $p_2,\dotsc,p_5$. Then one can take the 3 lines $\overline{p_0p_1}$, $\overline{p_2p_3}$, and $\overline{p_4p_5}$, which are in general position. These lines are toric subvarieties of $\mathbb{P}^5$ with the standard toric structure, so the blow-up $X\to\mathbb{P}^5$ along the three lines is a toric variety. Using toric geometry, one can compute the cone $\text{NE}(X)$ of effective curves, its extremal rays and the associated contractions. It turns out that $X$ is weak Fano, i.e. $-K_X$ is nef and big, and $\text{NE}(X)$ has 6 extremal rays: 3 correspond to the 3 blow-ups, while the other 3 are small and have intersection zero with $K_X$. In particular, $X$ has no projective bundle structure. The loci $E_i$ ($i=1,2,3$) of the 3 small rays are the transforms of the 3-dimensional linear subspaces of $\mathbb{P}^5$ containing 2 of the 3 lines, and they are pairwise disjoint. Each $E_i$ is isomorphic to the blow-up of $\mathbb{P}^3$ along 2 lines, and it is a $\mathbb{P}^1$-bundle over $\mathbb{P}^1\times \mathbb{P}^1$. The small contraction restricts on $E_i$ to the projection to $\mathbb{P}^1\times \mathbb{P}^1$. One can flop these 3 small rays, and get a new smooth weak Fano variety $X'$ (still toric). This $X'$ is a $\mathbb{P}^2$-bundle over $\mathbb{P}^1\times\mathbb{P}^1\times \mathbb{P}^1$, more precisely: $$X'=\mathbb{P}_{\mathbb{P}^1\times\mathbb{P}^1\times\mathbb{P}^1}\left(\mathcal{O}(1,1,0)\oplus\mathcal{O}(1,0,1)\oplus\mathcal{O}(0,1,1)\right).$$ (Attention: I am using Grothendieck's convention for Proj, while it seems to me that Libli uses the opposite one.) The $\mathbb{P}^2$-bundle has 3 special sections $S_i$ ($i=1,2,3$) with normal bundle $\mathcal{O}(-1,-1,0)\oplus\mathcal{O}(-1,0,-1)$ and so on. Each $S_i$ is the locus of a small extremal ray of $X'$, having intersection zero with $K_{X'}$, corresponding to one of the 3 flops.

$\endgroup$
3
  • $\begingroup$ That's a great answer! Thanks so much! $\endgroup$
    – Libli
    Commented Apr 25, 2014 at 8:45
  • $\begingroup$ May I ask you another question? On $X'$ there is the line bundle $p^*\mathcal{O}(-1,-1,-1)$, where $p : X' \rightarrow \mathbb{P}^1 \times \mathbb{P}^1 \times \mathbb{P}^1$ is the canonical projection. Do you have an idea what is its transform when we flop it to $X$? $\endgroup$
    – Libli
    Commented Apr 25, 2014 at 17:50
  • $\begingroup$ It is easy to see what happens to toric divisors, but it is more difficult to study a general divisor in the linear system $|p^*\mathcal{O}(1,1,1)|$. If $H'_1\subset X'$ is the pull-back of $\{pt\}\times\mathbb{P}^1\times\mathbb{P}^1$, the transform $H_1$ of $H_1'$ in $X$ is the transform of a hyperplane in $\mathbb{P}^5$ containing 2 of the blown-up lines. Similarly you can consider in $X$ the transforms $H_2$ and $H_3$ of hyperplanes in $\mathbb{P}^5$ containing the 2 other pairs of blown-up lines. Then $\mathcal{O}_X(-H_1-H_2-H_3)$ corresponds to $p^*\mathcal{O}(-1,-1,-1)$. $\endgroup$ Commented Apr 26, 2014 at 5:48
1
$\begingroup$

Edit. It turns out that what follows is simply the realization of a blowing up of $\mathbb{P}^5$ as a $\mathbb{P}^2$-bundle over $\mathbb{P}^1\times \mathbb{P}^1\times \mathbb{P}^1$. This is the blowing up that the OP already knows about, not the blowing up that the OP is asking about.

For three copies of $\mathbb{P}^1$, say $\mathbb{P}(A)$, $\mathbb{P}(B)$ and $\mathbb{P}(C)$, with respective universal invertible quotients, $$q_A:A\otimes_k\mathcal{O}_{\mathbb{P}(A)}\to \mathcal{O}_{\mathbb{P}(A)}(1), $$ $$q_B:B\otimes_k\mathcal{O}_{\mathbb{P}(B)}\to \mathcal{O}_{\mathbb{P}(B)}(1), $$ $$q_C:C\otimes_k\mathcal{O}_{\mathbb{P}(C)}\to \mathcal{O}_{\mathbb{P}(C)}(1), $$ on $X = \mathbb{P}(A)\times \mathbb{P}(B)\times \mathbb{P}(C)$, form the rank 3, locally free sheaf $$\mathcal{E} = \text{pr}_{\mathbb{P}(A)}^*\mathcal{O}_{\mathbb{P}(A)}(1) \oplus \text{pr}_{\mathbb{P}(A)}^*\mathcal{O}_{\mathbb{P}(B)}(1)\oplus \text{pr}_{\mathbb{P}(A)}^*\mathcal{O}_{\mathbb{P}(C)}(1), $$ with the associated quotient, $$ q: (A\oplus B \oplus C)\otimes_k \mathcal{O}_X \to \mathcal{E}.$$ This quotient defines an induced morphism, $$ f :\mathbb{P}_X(\mathcal{E}) \to \mathbb{P}(A\oplus B\oplus C).$$ The three obvious quotients, $$ r_A : \mathcal{E} \to \text{pr}_{\mathbb{P}(A)}^*\mathcal{O}_{\mathbb{P}(A)}(1), \ \ r_B : \mathcal{E} \to \text{pr}_{\mathbb{P}(B)}^*\mathcal{O}_{\mathbb{P}(B)}(1), \ \ r_C : \mathcal{E} \to \text{pr}_{\mathbb{P}(C)}^*\mathcal{O}_{\mathbb{P}(C)}(1), $$ which in turn defines three sections, $$ s_A: X\to \mathbb{P}_X(\mathcal{E}), \ \ s_B : X\to \mathbb{P}_X(\mathcal{E}),\ \ s_C:X\to \mathbb{P}_X(\mathcal{E}), $$ of the projection to $X$. The pairwise spans of these three sections are sub-$\mathbb{P}^1$-bundles, $$ \mathbb{P}_X(\mathcal{E}_{B,C}), \mathbb{P}_X(\mathcal{E}_{A,C}), \mathbb{P}_X(\mathcal{E}_{A,B}). $$ The images of these three subbundles are contracted under $f$ to the three $\mathbb{P}^3$s, $$\mathbb{P}(B\oplus C), \ \ \mathbb{P}(A\oplus C), \ \ \mathbb{P}(A\oplus B).$$ Thus $f$ is a birational, projective morphism that is an isomorphism over the complement of $$\mathbb{P}(B\oplus C) \cup \mathbb{P}(A\oplus C) \cup \mathbb{P}(A\oplus B).$$

$\endgroup$

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.