MathOverflow is a question and answer site for professional mathematicians. Join them; it only takes a minute:

Sign up
Here's how it works:
  1. Anybody can ask a question
  2. Anybody can answer
  3. The best answers are voted up and rise to the top

Let $X$ a variety. Say $X=\operatorname{Spec} A$.

Consider two ideals of $A$, say $I$ and $J$, with equal radical ; and consider the blow-ups of X with centre $I$ and $J$, say $Y_I$ and $Y_J$. How can I decide if the blow-up map $Y_I \to X$ factors through $Y_J$ ?

For example, if $X = \operatorname{Spec}k[x,y]$ is the affine plane. Then we can consider the blow-ups with centre $I_1 = (x,y)$, $I_2 = (x,y^2)$, $I_3=(x^2,y^2)$ and $I_4=(x^2,xy,y^2)=I_1^2$. We have $I_4\subset I_3\subset I_2\subset I_1$. Let $Y_i$ denote the corresponding blow-ups.

It is well-known that $Y_1$ and $Y_4$ are isomorphic. However $Y_2$ and $Y_3$ are two other different varieties. Using the universal property of the blow-up we see that $Y_1\to X$ factors through $Y_3$ but not through $Y_2$.

What kind of property should I look at to predict this kind of factorization, without computing the actual blow-up ?

share|cite|improve this question
up vote 2 down vote accepted

There are two conditions I can think of.

  1. Perhaps $I = J \cdot J'$ for some other ideal $J'$. Indeed, blowing up a product of ideals is the same as blowing up one ideal (say $J$) and then blowing up the total transform of the other (say $J'$).

  2. The other is requiring that $I = \overline{J}$, here this is integral closure of ideals (more generally $I$ and $J$ have the same integral closure and $I \subseteq J$). Indeed, two ideals have the same integral closure if and only if they have the same normalized blow-up and if their total transforms on the common normalized blow-up agree.

Let me give an alternate explanation of 2., suppose for example that $I = \overline{J}$. Then the blow-up of $I$ and the blow-up of $J$ have the same normalization and $Y_I$ is just a partial normalization of $Y_J$.

So this explains some of your examples. Indeed, the integral closure of $I_3$ is just $I_4$. On the other hand, $I_4 = I_1^2$. So by 1., $I_1$ and $I_4$ have the same blowup but by 2., the blow up of $I_4$ factors through the blowup of $I_3$. However, if you consider $I_1 \cdot I_2$ or $I_3 \cdot I_2$, you will find that their blowups factor through $I_2$.

For smooth surfaces, there is a theory of factorization of complete (integrally closed) ideals due to Zariski. This should allow you to answer these sorts of questions even more precisely.

share|cite|improve this answer
I think this will help me a lot! The point 2 is really great, I was not aware of it. – Lierre Feb 14 '12 at 8:19

The universal property of the blowup $\pi:Bl_I X\to X$ of an ideal $I$ in $X$ is that the sheaf of ideals $\pi^{-1}I\cdot O_{Bl_I X}$ is invertible. So, if you want to find out whether $Bl_I$ factors through $Bl_J$ you just have to check whether $\pi^{-1}J\cdot O_{Bl_I X}$ is invertible.

share|cite|improve this answer

Your Answer


By posting your answer, you agree to the privacy policy and terms of service.

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