I have a two-dimensional vector space ${\mathbb C}^2$ with basis $e_m, f_1$ and action of ${\mathbb C}^*$ by $t \cdot e_m = t^m e_m$ and $t \cdot f_1 = f_1$ and I have the projective line ${\mathbb P}^1$ which is the set of lines through the origin in this ${\mathbb C}^2$. I have the tautological line bundle $\Lambda$ on this ${\mathbb P}^1$ whose fiber over the line $L$ in ${\mathbb C}^2$ is the vector space $L$. I also have the trivial line bundle ${\mathcal O}$ on ${\mathbb P}^1$. Finally, I have the skyscraper sheaf ${\mathbb C}_{e_m}$ on ${\mathbb P}^1$ supported on the point of ${\mathbb P}^1$ which corresponds to the line spanned by $e_m$ in ${\mathbb C}^2$.

Note that all of the sheaves $\Lambda$, $\mathcal O$ and ${\mathbb C}_{e_m}$ are ${\mathbb C}^*$-equivariant.

Question: how to write a short exact sequence of ${\mathbb C}^*$-equivariant sheaves on ${\mathbb P}^1$ which involves $\Lambda$, ${\mathcal O}$ and ${\mathbb C}_{e_m}$ possibly with twists of grading?

Namely, non-equivariantly we have a short exact sequence $0 \to \Lambda \to {\mathcal O} \to {\mathbb C}_{e_m} \to 0$, is this an exact sequence of equivariant sheaves, or do we need to add twists of grading somewhere to make this exact sequence equivariant?


1 Answer 1


This is equivariant. The map $\Lambda \to \mathcal O$ explicitly on sections sends a section $(e_m, f_1) $ valued in $ L \subseteq \mathbb C^2$ to the section $f_1$ of the trivial line bundle. This map is an isomorphism everywhere but the point corresponding to the line spanned by $e_m$, and vanishes to order $1$ there. So the cokernel is the map to $\mathbb C_{e_m}$ given by evaluating at that point.

Both the projection and evaluation maps are manifestly $\mathbb G_m$-equivariant.


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.