2
$\begingroup$

I am currently looking at stratified Mukai flop. Roughly speaking, this is a construction that, starting with a grassmannian $G$ inside a hyperkahler $X$ produces a birational manifold $X^*$ (with a dual grassmannian $G^*$ inside). The easiest example is the standard Mukai flop, namely the case in which the grassmannian $G$ is a projective space $\mathbb{P}^n$.

I was looking for a concrete example of this situation.

In the case $G=\mathbb{P}^n$, I can construct such an example starting from a rational curve $C\cong\mathbb{P}^1$ inside a $K3$ surface $S$. The Hilbert scheme of points then provides $C^{[n]}\cong\mathbb{P}^n\subset S^{[n]}$.

What about a generic grassmannian $G=G(k,n)$ with $k\neq1,n-1$?

Thank you very much for any comments, remarks and answers.

$\endgroup$
  • $\begingroup$ For a quartic surface $S$ in $\mathbb{P}_{\mathbb{C}}^3$ that contains no lines, the Grassmannian $G$ of lines $L$ in $\mathbb{P}_{\mathbb{C}}^3$ embeds in $\text{Hilb}^4_{S/\mathbb{C}}$ via the rule $[L] \mapsto [L\cap S]$. $\endgroup$ – Jason Starr Mar 17 '16 at 19:06
  • $\begingroup$ By the way, in the example above, since $\text{Hilb}^4_{S/\mathbb{C}}$ is an $8$-dimensional hyperKaehler manifold, and since $G$ is $4$-dimensional, it is a maximal isotropic subvariety. So the embedding of $G$ does not factor through some projective space inside of $\text{Hilb}^4_{S/\mathbb{C}}$ of dimension strictly larger than $4$ (isotropic varieties have dimension less than or equal to $4$). $\endgroup$ – Jason Starr Mar 17 '16 at 19:08
3
$\begingroup$

This is just the general case of the example from my comment above. Let $S\subset \mathbb{P}^g_{\mathbb{C}}$ be a K3 surface of degree $2g-2$ that contains no curve that spans a $\mathbb{P}^r$ with $r<g-1$. Any such curve would be contained in (many) hyperplane sections of $S$, hence would have degree $< 2g-2$. So if $\text{Pic}(S)$ is generated by $[\mathcal{O}_{\mathbb{P}^g}(1)|_S]$, then $S$ contains no such curve.

Now let $G$ be the Grassmannian of dimension $2g-2$ parameterizing linear subspace $\Lambda\subset \mathbb{P}^g$ of dimension $g-2$. There is a morphism $G \to \text{Hilb}^{2g-2}_{S/\mathbb{C}}$ by $[\Lambda] \mapsto [\Lambda \cap S]$. Since $\text{Hilb}^{2g-2}_{S/\mathbb{C}}$ has dimension $4g-4$, $G$ is a maximal isotropic subvariety of $\text{Hilb}^{2g-2}_{S/\mathbb{C}}$. So it does not factor through any projective spaces embedded in $\text{Hilb}^{2g-2}_{S/\mathbb{C}}$.

Edit. I originally had an incorrect comment here speculating on a modular interpretation of the Mukai flop.

$\endgroup$
  • $\begingroup$ Thank you very much. This is exactly what I was looking for! $\endgroup$ – User3773 Mar 18 '16 at 11:48

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

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