1
$\begingroup$

I have a question a step in the example demonstating the blowing up in Liu's "Algebraic Geometry and Arithmetic Curves" in the excerpt below (or look up at page 320):

enter image description here

We blow up the surface $X= Spec k[S,T,W]/(ST-W^2)= Spec k[s,t,w)$ in the point $x_0$ corresponding to max ideal $m=(s,t,w)$.

The point of my interest is the fiber $E:=\pi^{-1}(x_)$

I don't understand why the observation that

$$E \cap A_1= V(s)=Spec k[ws^{-1}]$$ $$E \cap A_2= V(t)=Spec k[wt^{-1}]$$ $$E \cap A_3= V(w)=Spec k[tw^{-1},sw^{-1}]/((sw^{-1}) \cdot (tw^{-1})-1)$$

leads to conclusion that $E= \mathbb{P}^1_k$?

Indeed, by definition $\mathbb{P}^1_k$ glues together as $Spec k[x/y] \cup Spec k[y/x]$ on the intersection via $k[x/y,y/x] \to k[x/y,y/x]$

by $ x/y \mapsto y/x$.

Why the calculation in the excerpt imply that $E= \mathbb{P}^1_k$ hold?

$\endgroup$
2
  • 1
    $\begingroup$ Because he's realized $E$ as the cofibered product of $\mathbb{A}^1$ with itself along $\mathbb{G}_m$ where the gluing map is inversion. This cofibered product is just $\mathbb{P}^1$. $\endgroup$ May 12 '19 at 22:13
  • $\begingroup$ @AlexYoucis: I'm not sure if I understood you completely so please correct me if I wrong. Since it's not allowed to draw diagrams in comments I outsourced it as an answer: $\endgroup$
    – KarlPeter
    May 12 '19 at 23:08
1
$\begingroup$

your cofibered product along $\mathbb{G}_m$ seems to be

$$ \require{AMScd} \begin{CD} Spec( k[tw^{-1},sw^{-1}]/((sw^{-1}) \cdot (tw^{-1})-1)) @>{g} >> Spec( k[ws^{-1}]) \\ @VVfV @VVV \\ Spec( k[wt^{-1}]) @>{}>> Spec (k[ws^{-1}]) \coprod Spec( k[wt^{-1}]) \end{CD} $$

where we identified $E \cap A_1$ and $E \cap A_2$ with $\mathbb{A}^1$ and $E \cap A_3$ with $\mathbb{G}_m$.

Futhermore $g$ is induced by ring map $k[ws^{-1}] \to k[tw^{-1},sw^{-1}]/((sw^{-1}) \cdot (tw^{-1})-1), ws^{-1} \mapsto tw^{-1}$ and

$f$ by $k[wt^{-1}] \to k[tw^{-1},sw^{-1}]/((sw^{-1}) \cdot (tw^{-1})-1), wt^{-1} \mapsto sw^{-1}$

Is this the construction which you mean?

$\endgroup$
1
  • $\begingroup$ Yes. That's correct. $\endgroup$ May 12 '19 at 23:11

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.