MathOverflow is a question and answer site for professional mathematicians. It's 100% free, no registration required.

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

This is a follow up to my previous question, and I have lowered my demands to a situation as follows:

Let $X$ be an algebraic variety, $\mathcal{I}$ a coherent sheaf of ideals and $\mathcal{L}$ a line bundle on $X$. Let $h_1,\ldots,h_k\in\mathcal{L}(X)$ be certain global sections with the property that under any local trivialization $\psi:\mathcal{L}|_U\xrightarrow{\ \sim\ }\mathcal{O}_U$, we have $\psi_U(h_i|_U)\in\mathcal{I}(U)$. In other words, $ Z(\mathcal{I})\subseteq Z(h_i)$ for all $i$.

Let $\beta:Y:=\mathrm{Bl}_\mathcal{I}(X)\to X$ be the blow-up of $X$ along $\mathcal{I}$. Now, I am looking for a line bundle $\mathcal{M}$ on $Y$ and global sections $g_0,\ldots,g_k\in\mathcal{M}(Y)$ such that $Z(g_i)$ is the strict transform of $Z(h_i)$.

From Artie's comment in my previous question, I assumed that the bundle $\beta^\ast\mathcal{L}\otimes \mathcal{E}^\vee$ should satisfy my demands, where $\mathcal{E}=\bigoplus_{d\ge 0} \mathcal{I}^{d+1}$ is the exceptional ideal sheaf of the blow-up and $\mathcal{E}^\vee$ its dual.

My question: Is this the $\mathcal{M}$ I am looking for, and if yes, why? I do no see what the corresponding global sections would be.

share|cite|improve this question
The condition in the second paragraph can be rewritten as $h_i \in \Gamma(X,{\mathcal I}\otimes{\mathcal L})$, the latter being considered as a subspace of $\Gamma(X,{\mathcal L})$. – Sasha Jun 21 '11 at 18:38
That's neat indeed! – Jesko Hüttenhain Jun 21 '11 at 21:07
up vote 3 down vote accepted

Jesko, you still need more. The condition $Z(\mathscr I)\subseteq Z(h_i)$ ensures that the strict transform of these sections will be different from their pull-back, but you also need a condition that guarantees that they will be linearly equivalent.

Here is the point more concretely: Let $D_i=Z(h_i)$ and $\sum E_j=\beta^{-1}(Z(\mathscr I))$. Now $$\beta^*D_i \sim \widetilde D_i+\sum a_{ij}E_j$$ where $\widetilde D_i$ is the strict transform of the divisor $D_i$ and the $a_{ij}$ are the appropriate numbers. In order for $\widetilde D_i\sim \widetilde D_{i'}$ to be true you need that $a_{ij}=a_{i'j}$ for all $j$. In other words, you need some condition that guarantees that the $D_i$ (in some complicated sense) have the same multiplicity along $Z(\mathscr I)$.

share|cite|improve this answer
I should first tell you that I am very grateful for the great deal of patience and help that you're providing. So, thanks a lot! I also understand your point now - but is there any way to control the $a_{ij}$? For instance, if $X$ is projective and $\mathcal{L}=\mathcal{O}_X(d)$, do I have a chance? At least for $d=1$, this should be possible. – Jesko Hüttenhain Jun 21 '11 at 21:07
Jesko, I would expect that it's more about $\mathscr I$. If that is relatively simple, say it defines a smooth irreducible subvariety and $X$ is smooth along $Z(\mathscr I)$, then there is only one $E_j$ and the $a_{ij}$ is the multiplicity of $D_i$ along $Z(\mathscr I)$. This is well-defined, so you could assume that they are equal. If $Z(\mathscr I)\subseteq X$ is more complicated it is difficult to tell the $a_{ij}$ from initial data. Then again, this construction is a way to define the $a_{ij}$ and they do mean some kind of multiplicity, but I suppose you would like something else – Sándor Kovács Jun 21 '11 at 21:42
I should have said that upfront: Everything I am dealing with may safely be assumed to be smooth (both $Z(\mathcal{I})$ and $Z(h_i)$ and also $X$ itself), so I'll probably get what I want, yet. Thanks again! – Jesko Hüttenhain Jun 21 '11 at 22:55

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.