Assume $\mathbb{P}^1_k$ is the projective line over an algebraically closed field. In Hartshorne, Chapter II.6, Corollary 6.17, Hartshorne claims that the generator of $\text{Pic}(\mathbb{P}^1_k)$ is isomorphic to $\mathcal{O}(1)$. Why?

I understand that in the case of $\mathbb{P}^1_k$, we have an isomorphism of Weil and Cartier divisors, as well as an isomorphism of Cartier divisors and invertible sheaves. I understand that these isomorphisms descend onto isomorphism of $Cl$, $CaCl$ and $Pic$. I understand that $Cl(\mathbb{P}^1_k)$ is isomorphic to $\mathbb{Z}$ and is generated by the class of a hyperplane. I know how to compute the corresponding Cartier divisor, but am struggling to show that any sheaf $\mathcal{O}(D)$ generated by a representative of $\text{CaCl}(\mathbb{P}^1_k)$ having $\text{deg}(D) = 1$ is isomorphic to $\mathcal{O}(1)$.

Any help?

Thanks in advance!

Edit: I managed to prove this fact.

  • 3
    $\begingroup$ There is some serious confusion here. $\mathscr O(1)$ generates the Picard group. A sheaf (or line bundle) is not a group! $\endgroup$ Jun 8 at 17:49
  • 2
    $\begingroup$ I can't believe the legendary Theodore Shifrin replied to my stack post! While I do not protest the fact that I am generically confused, I meant to ask why is it true that all sheaves in the isomorphism class of $\mathcal{O}(D)$ for $D\in\text{Cl}(\mathbb{P}^1_k)$ with $\text{deg}(D) = 1$ are isomorphic to $\mathcal{O}(1)$. $\endgroup$
    – kindasorta
    Jun 8 at 18:35
  • 2
    $\begingroup$ If you've managed to prove it, then you should write up your solution and accept the answer so others can benefit from what you've learned! $\endgroup$
    – Stahl
    Jun 8 at 19:21
  • $\begingroup$ My argument is not Hartshorne style. The exponential sheaf sequence gives $H^1(X,\mathscr O^*) \cong H^2(X,\Bbb Z)$ for $X=\Bbb P^1$. $\endgroup$ Jun 8 at 19:22
  • 1
    $\begingroup$ You may also wish to correct the title, too (it looks like you missed this in your last edit clarifying the question). $\endgroup$
    – KReiser
    Jun 8 at 19:33

1 Answer 1


Hartshorne already proves in II.6.13.c that if $D_1\sim D_2$ are two Cartier divisors, then $\mathcal{L}(D_1)$ is isomorphic to $\mathcal{L}(D_2)$. Therefore, in order to prove that the generator of the class of $\text{CaCl}(\mathbb{P}^1_k)$ is isomorphic to $\mathcal{O}(1)$, it is enough to show that any invertible sheaf in the linear equivalence class of the generator of $\text{CaCl}(\mathbb{P}^1_k)$ is isomorphic to $\mathcal{O}(1)$.

Let's consider the representative corresponding to the Weil divisor $D = 1\cdot (1,0)$. The corresponding Cartier divisor is given by $\{1, U_x\}, \{x/y,U_y\}$, which corresponds to the sheaf $\mathcal{L}(D)$, given by the trivialization $\mathcal{L}(D)|_{U_x} = \mathcal{O}_{\mathbb{P}^1_k}(U_x)$ and $\mathcal{L}(D)|_{U_y} = \mathcal{O}_{\mathbb{P}^1_k}(U_y)\cdot \dfrac{y}{x}$.

To see that the sheaf $\mathcal{L}(D)$ is isomorphic to the sheaf $\mathcal{O}(1)$, consider the sheaf morphism, which over the open $U$ sends the section $s\in \mathcal{L}(D)(U)$ to $x\cdot s\in \mathcal{O}(1)$. Injectivity follows from the fact that $\mathbb{P}^1_k$ is integral. Surjectivity can be checked locally on $U_x$ and $U_y$. For example, for $U_y$ we have $\dfrac{y}{x}\mapsto y$ and $1\mapsto x$, which generate $\mathcal{O}(1)|_{U_y}$, and similarly for $U_x$.

Our sheaf morphism is injective and locally surjective and therefore an isomorphism, which completes the proof.

  • 1
    $\begingroup$ Just a picky remark. By $(1,0)$ you mean the divisor $P=[1,0]$. $\endgroup$ Jun 9 at 0:14

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.