This is Exercise II.4.4 in Shafarevich's book Basic Algebraic Geometry, second edition.

Consider the rational map $ \phi:\mathbb{P}^2\to \mathbb{P}^4 $ given by $$ \phi(x_0:x_1:x_2)=(x_0x_1:x_0x_2:x_1^2:x_1x_2:x_2^2). $$ Prove that $ \phi $ is a birational map to a surface $ \overline{\phi(\mathbb{P}^2)} $, and that the inverse map $ \overline{\phi(\mathbb{P}^2)}\to \mathbb{P}^2 $ coincides with the blowup.

Here is what I have done. I called the variables in $ \mathbb{P}^2 $ $ z_0, \cdots, z_4 $ and obtained the equations \begin{align*} z_0&=x_0x_1\\ z_1&=x_0x_2\\ z_2&=x_1^2\\ z_3&=x_1x_2\\ z_4&=x_2^2. \end{align*} Then I have tried to solve this for $ x_0, x_1 $ and $ x_2 $ to obtain a map $ \psi:\overline{\phi(\mathbb{P}^2)}\to \mathbb{P}^2 $. However, it doesn't seem to be possible to do this, I always get expressions for $ x_i^2 $, not $ x_i $. Of course, I can't take roots. Then I tried to make some assumptions. Since $ \phi $ is not defined for $ p=(1:0:0) $, I assume that $ x_1=1 $ and I obtain $$ \psi(z_0:z_1:z_2:z_3:z_4)=(z_0:1:z_3). $$ By assuming that $ x_2=1 $ instead, I get a similar (but different) map. In both cases, $ \psi $ is not rational and they "use" only two of the five variables. So I doubt that what I have done is the right way to go. So, now I am stuck. Can someone please help me?


$\newcommand{\P}{\Bbb P}$Outside $[1:0:0]$, $\phi$ is an isomorphism. For example, an partial inverse is given by (with your notation) $(z_0:z_1:z_2:z_3:z_4) \mapsto (z_0:z_1:z_3)$. This works only if $x_1 \neq 0$ (or equivalently $z_0,z_1,z_3$ are not all zero). Other cases are similar.

So we just need to understand what is the closure of the image. Since we understand the image of $v_2(\P^2 \backslash \{p\})$ we just need to see what is $\lim\limits_{x \to p} \phi(x)$. For this, I take the path $x$ defined by $$x_0(t) = 1, x_1(t) = at, x_2(t) = bt$$ for some $a,b \in k$. It is easy to get that $\lim\limits_{t \to 0}\phi(x(t)) = [a:b:0:0:0]$. This shows that $\phi$ is remplacing a point by a projective line : this is a good start for the blow-up !

For verify it we just need to understand $\phi$ around $[1:0:0]$. Recall that the local form of the blow-up is the map $\Bbb A^2 \to \Bbb A^2 \times \P^1$ $(x,y) \to ((x,y), [x:y])$. Back to your map, take local coordinates around $[1:0:0]$, namely $s=\frac{x_1}{x_0}$ and $t = \frac{x_2}{x_0}$. The map is just $(s,t) \mapsto [s:t:s^2:st:t^2]$. Now the three last coordinates corresponds to $(x,y)$ in our local model (i.e exercise : $\Bbb A^2 \backslash \{(0,0)\} \to \P^2, (x,y) \mapsto (x^2;xy:y^2)$ is an embedding) and of course the first part is remembering the direction of $(s,t)$. So this is exactly the blow-up !

There is a very nice interpretation in term of the Veronese embedding and linear systems of conics.


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.