Let $P_1, P_2, Q$ denote the Hilbert scheme of a plane conic in $\mathbb{P}^3$, a quartic and a degree $d$ surface in $\mathbb{P}^3$. Then there is a natural inclusion map $i$ from Hilbert flag scheme $\mathrm{Hilb}_{P_1,Q}$ to $\mathrm{Hilb}_{P_2,Q}$ under the map, $(C,X) \mapsto (2C,X)$. Then is the image under the composition of the maps, $i$ with the natural projection map $\mathrm{pr}_2$,

$\mathrm{Hilb}_{P_1,Q} \to\mathrm{Hilb}_{P_2,Q}\to \mathrm{Hilb}_{Q}$

an irreducible component of the image of $\mathrm{pr}_2$? If so can this result be generalized to the case when we can replace plane conic and quartic by curves $C_1, C_2$ such that $rC_1$ has the same Hilbert polynomial as $C_2$?

Note:Hilbert flag scheme $\mathrm{Hilb}_{P_i,Q}$ parametrize pairs of the form $C \subset X$ where $P_i$ is the Hilbert polynomial of $C$ and $X$ is a degree $d$ surface in $\mathbb{P}^3$.

  • 3
    $\begingroup$ I disagree that you have such a morphism. You have a rational transformation which is regular on the open subset parameterizing pairs $(C,X)$ such that $C$ is a Cartier divisor in $X$. Also, when $d$ equals $4$, clearly the image of $\text{Hilb}_{P_2,Q}$ equals all of $\text{Hilb}_Q$ (consider hyperplane sections), whereas the image of $\text{Hilb}_{P_1,Q}$ will be a proper closed subset. $\endgroup$ Commented Jun 22, 2012 at 17:18
  • $\begingroup$ Is it obvious that not every quartic surface contains a plane conic? $\endgroup$
    – Will Sawin
    Commented Jun 22, 2012 at 17:33
  • 3
    $\begingroup$ @Will -- Yes, it is obvious. It is 9 linear conditions on quartic surfaces to contain a given smooth plane conic, yet there is only an 8-parameter family of plane conics in $\mathbb{P}^3$. $\endgroup$ Commented Jun 22, 2012 at 17:36
  • $\begingroup$ I don't understand the map on Jason's open subset either. I assume $Hilb_{P_2}$ is the Hilbert scheme of plane quartics. If $C \subset X$ is planar, that does not mean that $2C \subset X$ is also. You need to adjust $P_2$ to be the Hilbert polynomial of the first infinitesimal neighborhood of $C \subset X$. You can compute this by determining the degree of the normal bundle of $C$ (via adjunction), but the answer will depend on the degree $d$ of $X$. $\endgroup$ Commented Jun 23, 2012 at 9:10
  • 1
    $\begingroup$ @Arend: The OP says "quartics", not "plane quartics". Indeed if you interpret the question with "plane quartics", it is absurd. $\endgroup$ Commented Jun 24, 2012 at 12:10

1 Answer 1


This question has been answered in the comments by Jason Starr. The morphism $i$ is in general not defined and the composition will in general not map to a component of the target. I am reposting this as a CW answer; if it gets upvoted, this question will not reappear on the front page.


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.