Let $X\subset \mathbb C^2$ be the Riemann surface given by (the multivalued function) $y=(1-x^3)^{1/3}$, and let $\phi:X \to \mathbb C $ be the induced map. Let $X'\subset P(\mathbb C^2)$ be the complex curve $x^3+y^3=z^3$ (in homogeneous coordinates). I want to define a complex manifold structure on $X'$ and a holomorphic $\phi':X'\to \mathbb C$ that extends $\phi$.

So far I know the following,

(i) $(x,y) \mapsto [x:y:1]$ gives a homeomorphism $\psi$ of $\mathbb C^2$ onto its image.

(ii) $\psi(X)\bigcup \{[1:-\zeta_3:0], [1:-\zeta_3^2:0], [1:-1:0]\}=X'$.

So I want to first define a manifold structure for $X$ and push it to $X'$ by $\psi$ and use another chart to cover $\{[1:-\zeta_3:0], [1:-\zeta_3^2:0], [1:-1:0]\}$. I think $x$ is a coordinate chart around $x_0$ if we fix a branch (fix a $y_0$). So we need three charts to cover $X$ because we have three branches. Is this correct?

Also, how do I construct a compatible chart to cover the three points at infinity?


If I got well, $X$ is the zero locus $x^3+y^3=1$ in $\mathbb{C}^2$ and your $\phi $ is the projection on one of the coordinate (I will assume is the $y$ just to be clear).

Now you take $X'$ that is ,as you can easily prove, the projective closure of $X$ in $\mathbb{P}^2(\mathbb{C})$.

As you correctly said , you need three charts because intuitively there are three branches. A standard way to proceed in the projective spaces is to use the standard charts $x_i \neq 0$ that are (in the projective plane) biholomorphic to $\mathbb{C}^2$.

If you restrict $X'$ to this charts , you will get three varieties: the first one (in the chart $z \neq0$) is your $X$ (and so you get your immersion) and the other two are very closely related to $X$ (so you can easily find complex structures on them in the same way you did with $X$).

  • $\begingroup$ One of my concern is that when we consider a point$x_0,0$, could x still be a local coordinate even though $x_0$ is a branch point? $\endgroup$ – user136592 Mar 25 '18 at 1:16
  • $\begingroup$ I would use the general fact that the partial derivatives of $x^3+y^3-1$ do not both vanish on $X$ and use the Implicit Function theorem $\endgroup$ – Tommaso Scognamiglio Mar 25 '18 at 7:03
  • $\begingroup$ I can differentiate to get $3x^2 dx +3y^2dy=0$ which I can rearrange to get $\frac{dx}{y}=-\frac{y}{x^2}$. From this I can determine the local coordinate depending on whether $x=0$ or $y=0$, but how would you construct the other two varieties? $\endgroup$ – Jade Mar 26 '18 at 4:19
  • $\begingroup$ Start writing your curve in the other two standard charts and look if you can make something similar $\endgroup$ – Tommaso Scognamiglio Mar 26 '18 at 6:36
  • $\begingroup$ If you take another chart where $x \neq 0$, then let the affine coordinate in that chart be $(y',z')=(y/x,z/x)$ and homogenize the equation wrt to that, you get $1+y'^3=z'^3$. You can differentiate to get $3y'^2=3z'^2$ and when $y_0' \neq 0$, take $z'$ is a local coordinate with differential $w=dz'/y'$. When $y_0' =0$, take $y'$ to be the local coordinate and let $w=y'dy'/dz'^2$. Is this what you mean? Is this how you are supposed to approach this problem? $\endgroup$ – Jade Mar 27 '18 at 4:03

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.