Let $F\rightarrow \mathbb{P}^{1}$ be a vector bundle, and let $\mathcal{F}$ be its sheaf of sections. I have read that the sheaf $\mathcal{F}(m)/\mathcal{F}$ is a skyscraper sheaf, but I do not know how to prove it.

I do not know if it is of any help, but I am going to show what I have thought until now. We have $$ \mathcal{F}(m)=\mathcal{F}\otimes\mathcal{O}_{\mathbb{P}^{1}}(m). $$ Since $H^{0}(\mathbb{P}^{1},\mathcal{O}_{\mathbb{P}^{1}}(m))\simeq K[X_{0},X_{1}]_{m}$, I guess that we may consider $(X_{1})$ as the distinguished point of $\mathbb{P}^{1}$ with respect to $\mathcal{F}(m)/\mathcal{F}$. Nevertheless, I do not know how to continue.

Any help would be appreciated.

  • $\begingroup$ To speak of $F(m)/F$ you should first chose a map $F \to F(m)$. What map do you have in mind? $\endgroup$ – Sasha Oct 23 '16 at 10:11
  • $\begingroup$ @Sasha I am not sure, but it is induced in some way by the map $\mathcal{O}_{\mathbb{P}^{1}}\rightarrow \mathcal{O}_{\mathbb{P}^{1}}(m)$ induced by $K[X_{0},X_{1}]_{0}\rightarrow K[X_{0},X_{1}]_{m}, k\mapsto X_{1}^{m}\cdot k$ with our interpretation. $\endgroup$ – H. Jackson Oct 23 '16 at 12:38

This works for any curve $C$. If $p$ a point on $C$ there is a short exact sequence $$ 0 \to \mathcal{O}(-p) \to \mathcal{O} \to k(p) \to 0. $$ Take a tensor product of this exact sequence with $\mathcal{F}(p)$. It stays exact because $F$ is a vector bundle. So, you get a short exact sequence $$ 0 \to \mathcal{F} \to \mathcal{F}(p) \to \mathcal{F}(p)\otimes k(p) \to 0. $$

Sheaf $\mathcal{F}(p)\otimes k(p) \cong \mathcal{F} \otimes k(p)$ is a skyscraper sheaf supported on $p$.

On $\mathbb{P}^1$ we have $\mathcal{O}(1) = \mathcal{O}(p)$ and $\mathcal{O}(m) = \mathcal{O}(mp)$, where $m \in \mathbb{Z}$. For positive $M$ we generalize previous argument by starting with short exact sequence $$ 0 \to \mathcal{O}(-mp) \to \mathcal{O} \to k_m(p) \to 0. $$ where $k_m(p)$ is a skyscraper sheaf supported on $x$ of length $m$. Then we tensor this short exact sequence with vector bundle $\mathcal{F}(mp)$.

| cite | improve this answer | |
  • $\begingroup$ Maybe it is not clear in the statement of my question, but $m$ is supposed to be an integer (furthermore, I want it to be positive). Let $p$ be a point of $C$. Could we generalize your idea for the divisor $m\cdot p$? Thank you! $\endgroup$ – H. Jackson Nov 3 '16 at 19:20
  • $\begingroup$ Sorry, I confused twisting by a point (as divisor) and twisting by number (i.e. by hyperplane divisor), but on $\mathbb{P}^1$ this is the same thing! And this argument also can be generalized to the case you interested in. $\endgroup$ – Alex Nov 3 '16 at 21:04

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.