Let $X\rightarrow Y$ be a smooth morphism of schemes and consider the blowup of the fiber product along the diagonal, $W:=Bl_{\Delta}X\times_Y X$.

a.Does there exist a collection of smooth morphisms of schemes $X_i\rightarrow Y_i$ such that

1.the schemes $W_i:=Bl_{\Delta_i}X_i\times_{Y_i} X_i$ are irreducible and

2.there exist morphisms $W_i\rightarrow W$ ,such that $\{W_i\rightarrow W\}$ is an etale cover?

[Here $\Delta_i$ denotes the diagonal of $X_i\times_{Y_i} X_i$.]

b.Same question with the additional assumption that $Y$ is smooth.

In case $Y$ is singular, the answer in a. is negative, see Will Sawin's answer below.

  • $\begingroup$ What are you denoting by $W$? $\endgroup$
    – dhy
    Aug 28, 2014 at 0:05
  • $\begingroup$ Sorry, $W$ is the blowup of fiber product along the diagonal $W=Bl_{\Delta}X\times_Y X$, I'll edit the post. $\endgroup$
    – matthew
    Aug 28, 2014 at 0:07
  • $\begingroup$ Can't one take $Y_i$ an etale cover of $Y$ and $X_i = X \times_Y Y_i$? Fiber product and blow-up are both etale-local constructions, so everything will just be a base change from $Y$ to $Y_i$, hence etale. $\endgroup$
    – Will Sawin
    Aug 28, 2014 at 2:56
  • $\begingroup$ @Will Savin:Why are the resulting schemes irreducible? $\endgroup$
    – matthew
    Aug 28, 2014 at 4:51
  • $\begingroup$ What does the index $i$ mean? $\endgroup$ Aug 28, 2014 at 6:50

1 Answer 1


Let $Y$ be the nodal cubic curve and let $X$ be its finite etale double cover. Then $X \times_Y X$ is just a union of two copies of the finite etale double cover. Blowing up changes nothing.

You have a slight problem here - the finite etale double cover is union of two $\mathbb P^1$ which intersect at $0$ and $\infty$. It has no etale cover by any irreducible schemes, because each irreducible scheme must map to only one of the two $\mathbb P^1$s, but any etale neighborhood of either $0$ or $\infty$ intersects both $\mathbb P^1$s.

So I think no $W_i$ work, which means the answer is no.

  • $\begingroup$ Thank you, Will. I was originally interested in the case where $Y$ is smooth, which I didn't mention in the hypothesis. I'll edit the post to include this case, too. $\endgroup$
    – matthew
    Aug 28, 2014 at 22:26

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.