3
$\begingroup$

Let me start with concrete examples. Let $X$ be a smooth special Gushel-Mukai threefold and $\mathcal{C}(X)$ be its honest Fano surface of conics, it has two irreducible components $\mathcal{C}(X)=\overline{\mathcal{F}}\cup\pi^*\Sigma(Y)$, where $\overline{\mathcal{F}}$ is the closure of double cover of family conics which are double tangent to the branch locus $\mathcal{B}$ on $Y_5$ and $\Sigma(Y)$ is the Fano surface of lines on $Y_5$, the map $\pi:X\rightarrow Y_5$ is a double cover with branch locus $\mathcal{B}$. Note that the component $\pi^*\Sigma(Y)$ is a $\mathbb{P}^2$. Also we have the intersection of two components is $\pi^*(\rho)\cup\pi^*\{L\subset\mathcal{B}\}$, where $\rho$ is the curve of $(-1,1)$-lines on $Y$ and $L$ is line in branch locus $\mathcal{B}$.

Let $\mathcal{C}_m(X)$ be the contraction of $\mathcal{C}(X)$ along the component $\pi^*\Sigma(Y)\cong\mathbb{P}^2$ to a point. Then $\mathcal{C}_m(X)$ becomes irreducible.

First, I assume that $X$ is general in the sense that $\mathcal{B}$ is general, i.e. does not contain any line or conic. Then one can show that $\mathcal{C}_m(X)$ has the unique singular point, denoted by $q$, I call this $q$ a type $I$ singularity. Now I assume $X$ is not general but with branch locus only contains one line $L$, then $\mathcal{C}_m(X)$ still has a unique singularity $q'$ because $\pi^*L$ is still in the component $\pi^*\Sigma(Y)$ and they contract to one point, but I think in this case the singularity type of $q'$ should be different with the previous case, I call it type $II$. Now, I assume that the branch locus $\mathcal{B}$ could also contain conic. Then $\mathcal{C}_m(X)$ would have extra isolated singular point $q''$ given by those conics, since they are fixed by the geometric involution $\tau$(coming from the double cover), I call those singularity Type $III$.

My first question is are these three type singularity are of the same type or not? For example, how to tell what kind of singularity type do they have, say $A_1, A_2$ etc? I would guess that type I and type II are different but for type I and type III, I am not so sure.

Consider the similar story on ordinary Gushel-Mukai threefold $X'$, the honest Fano surface of conics $\mathcal{C}(X')$, this is always irreducible surface. If $X'$ is general, then $\mathcal{C}(X')$ is smooth and if $X'$ is non-general, then there are singular points given by $\tau$-conic $C$ with normal bundle $\mathcal{N}_{C|X'}=\mathcal{O}_C(2)\oplus\mathcal{O}_C(-2)$(or equivalently, such conic is fixed by some involution $\tau$, but this involution appears in the blow up of $\mathcal{C}(X')$ or blow down of $\mathcal{C}(X')$). It is known that there is a unique exceptional curve $E\subset\mathcal{C}(X')$ consists of $\sigma$-conics and if we contract this exceptional curve we get the minimal surface of general type $\mathcal{C}_m(X')$. If $X'$ is general, then it is smooth, if $X'$ is not general, then it has singularity $q'''$ given by the $\tau$-conic I mentioned above, in this case, I call singularity type of $q'''$ type $IV$.

My second question: Is this type $IV$ singularity type the same as type $I, III, III$? My third question: how to study these singularities in the moduli space? I tried to google, but end up with nothing useful.

$\endgroup$
3
  • $\begingroup$ I did not read the entire post. Are you trying to find sources that prove results about singularities of parameter spaces of rational curves on Fano manifolds? $\endgroup$ Commented Jul 22, 2021 at 10:40
  • $\begingroup$ @JasonStarr, thanks for reply, I am trying to find sources to tell how to study the singularities of moduli space of rational curves on Fano manifold, besides some basic property, say the number of singularties(finitely many or infinitely many). I would like to see something like tell whether a singularity is a nodal, cuspidal or some other type something like this. $\endgroup$
    – user41650
    Commented Jul 22, 2021 at 12:02
  • $\begingroup$ "I would like to see something like tell whether a singularity is a nodal, cuspidal or some other type something like this." The dimension of the parameter space of rational curves on Fano manifolds grows, roughly, like the anticanonical degree of the curve. For a parameter space of large dimension, what is your definition of "nodal" and "cuspidal"? $\endgroup$ Commented Jul 22, 2021 at 13:07

0

You must log in to answer this question.