I am working through section IV.2 about blow-ups of the book Geometry of schemes by Eisenbud and Harris. I am having some trouble understanding the details of the following example, which they discuss in this section. I have very little expercience with schemes and I am sure my problem has to do with this.

Pick some algebraically closed field $K$ and define the affine plane $\mathbb{A}_K^2=\mathrm{Spec}(K[x,y])$. Set $U'=\mathrm{Spec}(K[x',y'])$ and $U''=\mathrm{Spec}(K[x'',y''])$. Consider the morphisms of schemes $$ \phi':U'\rightarrow \mathbb{A}_K^2 \quad\mbox{and}\quad \phi'':U''\rightarrow \mathbb{A}_K^2 $$ induced by the ring morphisms $$ \begin{matrix} (\phi')^{\#}:K[x,y]\rightarrow K[x',y']\\ x \mapsto x'\\ y \mapsto x'y' \end{matrix} \quad\mbox{and}\quad \begin{matrix} (\phi'')^{\#}:K[x,y]\rightarrow K[x'',y'']\\ x \mapsto x''y''\\ y \mapsto y''. \end{matrix} $$

Observe that $(\phi')^{\#}\vert_{K[x,y,x^{-1}]}:K[x,y,x^{-1}]\rightarrow K[x',y',{x'}^{-1}]$ is an isomorphism with inverse $$ \begin{matrix} K[x',y',{x'}^{-1}]\rightarrow K[x,y,x^{-1}]\\ x'\mapsto x\\ y'\mapsto yx^{-1}. \end{matrix} $$ Similarly we see that $(\phi'')^{\#}\vert_{K[x,y,x^{-1}]}:K[x,y,x^{-1}]\rightarrow K[x'',y'',{y''}^{-1}]$ is an isomorphism with inverse $$ \begin{matrix} K[x'',y'',{y''}^{-1}]\rightarrow K[x,y,x^{-1}]\\ x''\mapsto xy^{-1}\\ y''\mapsto y. \end{matrix} $$ Therefore, we have the dual isomorphisms $$ \phi'\vert_{U'_x}:U'_x=\mathrm{Spec}(K[x',y',x'^{-1}])\rightarrow U_x=\mathrm{Spec}(K[x,y,x^{-1}]) $$ and $$ \phi''\vert_{U''_y}:U''_y=\mathrm{Spec}(K[x'',y'',y''^{-1}])\rightarrow U_y=\mathrm{Spec}(K[x,y,y^{-1}]). $$ We write $U_{xy}=U_x\cap U_y=\mathrm{Spec}(K[x,y,x^{-1},y^{-1}])$ and observe that $U_{xy},U'_{xy}$ and $U''_{xy}$ can be identified via these isomorphisms.

Let's compute the morphisms of schemes $\psi=(\phi')^{-1}\circ\phi'':U_{xy}''\rightarrow U_{xy}'.$ This map is induced by the ring morphism $$ \begin{matrix} K[x',y',x'^{-1},y'^{-1}] & \rightarrow & K[x,y,x^{-1},y^{-1}] & \rightarrow & K[x'',y'',x''^{-1},y''^{-1}]\\ x' & \mapsto & x & \mapsto & x''y''\\ y' & \mapsto & yx^{-1} & \mapsto & {x''}^{-1}. \end{matrix} $$ Define $Z= \frac{U'\cup U''}{U_{xy}'\simeq U_{xy}''}$ and let $\phi:Z\rightarrow \mathbb{A}_K^2$ be the `gluing' of $\phi'$ and $\phi^{''}$.

The idea is that this scheme $Z$, together with the map $\phi$, is the blow-up of $\mathrm{A}_K^2$ in the origin. In particular $\phi^{-1}(x,y)$ should be isomorphic to $\mathbb{P}_K^1$ and $\phi$ should be an isomorphism everywhere else.

I am having trouble verifying this statement. This is what I tried in order to compute $\phi^{-1}(x,y)$:

First I wanted to consider $U'$ and $U''$ individually, expecting that $\phi'^{-1}(x,y)$ and $\phi''^{-1}(x,y)$ would both be isomorphic to an affine line. Then by gluing I expected this to become the projective line when considering $\phi^{-1}(x,y)$.

In order to compute $\phi'^{-1}(x,y)$, recall that $\phi'$ sends $p \in \mathrm{Spec}(K[x',y'])$ to the preimage of $pK[x',y']_{p}$, under the composite map $$ \begin{matrix} \alpha: K[x,y] & \rightarrow & K[x',y'] & \rightarrow & K[x',y']_{p}\\ x & \mapsto & x' & \mapsto & \frac{x'}{1}\\ y & \mapsto & x'y' & \mapsto & \frac{x'y'}{1}. \end{matrix} $$ Hence we are interested in finding $p \in \mathrm{Spec}(K[x',y'])$ such that $\phi'(p)=(x,y)$, i.e. $\alpha^{-1}(p)$ should equal $(x,y)$. So we want $p=\alpha(x,y)$. Compute $\alpha(x,y)=(x',x'y')=(x')$. Therefore, $\phi'^{-1}(x,y)$ consists just of one point, the ideal $(x')$. This is clearly not as expected.

Does anyone know what mistakes I am making? Thank you in advance.


I am not sure why you are complicating things with $\alpha$ (but it could be me it is pretty late here as I am writing this and maybe I forgot I need to localize).

By definition, $\phi'(p) = (\phi'^\#)^{-1}(p)$ (where $^{-1}$ means the inverse image of a set and not an inverse map).

So you want to find the prime ideals $p$ such that $(\phi'^\#)^{-1}(p) = (x,y)$, which means $\{g \in K[x,y] \mid \phi'^\#(g) \in p\} = (x,y)$

$\phi'^\#$ is injective but not surjective, so this means that $p \cap Im \phi'^\# = \phi'^\#((x,y))$, and thus
$p \cap K[x',x'y'] = (x',x'y')$.

In fact you can show that (because $(x',x'y')$ is a maximal ideal of $K[x',x'y']$)
this is equivalent to $x' \in p$ and $1 \notin p$.

And there are many nontrivial prime ideals of $K[x',y']$ containing $x'$, such as all the $(x',y'-k)$ for $k \in K$


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.