Let $X_{2,3}\subset\mathbb{P}^n$, with $n\geq 5$, be a complete intersection of a quadric $X_2$ and a cubic $X_3$ containing a $2$-plane $H$. Assume $X_2$ and $X_3$ to be general among the hypersurface of the same degree containing $H$. In particular $X_2$ and $X_3$ are smooth.

Question: If $n = 5$ is $X_{2,3}$ necessarily singular? Is $X_{2,3}$ smooth for $n\geq 6$?


2 Answers 2


If $X \subset \mathbb{P}^n$ is a non-degenerate, smooth complete intersection variety of dimension at least $3$, then the restriction map $$\operatorname{Pic}(\mathbb{P}^n) \to \operatorname{Pic}(X)$$ is an isomorphism by Grothendieck-Lefschetz theorem. So, we get $\operatorname{Pic}(X) = \mathbb{Z}$, generated by the hyperplane section.

This implies that $X$ contains no linear spaces of codimension $1$; in particular, the threefold $X_{2,3} \subset \mathbb{P}^5$ is necessarily singular.


If $n=5$ then let $\mathbb P^5$ have coordinates $x_0,\ldots,x_5$ and suppose the plane is $H=\mathbb P^2_{(x_0:x_1:x_2)}$. The two equations of $X$ are necessarily of the form $$ \begin{pmatrix} A_1 & B_1 & C_1 \\ D_2 & E_2 & F_2 \end{pmatrix}\begin{pmatrix} x_3 \\ x_4 \\ x_5 \end{pmatrix} = 0 $$ for some polynomials $A_1,\ldots,F_2\in\mathbb C[x_i]$ of the indicated degree. The scheme $Z\subset \mathbb P^5$ defined by the $2\times 2$ minors of the $2\times 3$ matrix has codimension 2, and thus $Z\cap H$ is nonempty (generically given by some points). Since both equations have order of vanishing $\geq2$ along $(Z\cap H)\subset X$, the 3-fold $X$ must be singular there.

In dimension $n\geq6$ the scheme $Z$ (obtained by the analogous argument) has codimension $n-3\geq3$ so in the general case $Z\cap H=\emptyset$ and the $n$-fold $X$ is smooth (along $H$ at least). Essentially, since the matrix never drops rank we can always use the equations to eliminate two of the variables $x_3,\ldots,x_n$, showing that $X$ is smooth of dimension $n$ at every point along $H$.

  • 4
    $\begingroup$ It's not true that both equations have order of vanishing $\geq 2$ at these points but rather that some linear combination does (but this still implies a singularity). $\endgroup$
    – Will Sawin
    Jan 18 at 12:58
  • $\begingroup$ Sorry, yes good point. That is what I meant to say. $\endgroup$
    – Tom Ducat
    Jan 18 at 14:55

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.