I am looking for a function $f:\mathbb C^2 \rightarrow \mathbb C^2$ that satisfies the two equations

$$\partial_{z_2}f_1(z_1,z_2) + \partial_{z_1} f_2(z_1,z_2)=0 \text{ and }$$ $$\partial_{\bar z_1}f_1(z_1,z_2) - \partial_{\bar z_2} f_2(z_1,z_2)=0$$

and in addition, is doubly-periodic in both its complex variables $z_1,z_2$. Does such a function exist and if not, why? I would not even know how to start building such a function.

In particular, I would like to have

$$f_1(z_1+1,z_2)=f_1(z_1,z_2+1)=f_1(z_1,z_2)$$ and

$$f_1(z_1+i,z_2) = e^{2\pi i k_1}f_1(z_1,z_2)$$ and

$$f_1(z_1,z_2+i) = e^{2\pi i k_2}f_1(z_1,z_2)$$

for some fixed $k_1,k_2 \in \mathbb R.$ Please let me know if you do have any questions. I had some typos in there, but hopefully everything is coherent now.

  • 2
    $\begingroup$ You, probably, want more because as written, you can just take any function $f=(f_1,f_2)$ in which $f_1$ depends only on $z_1$ and $f_2$ only on $z_2$ with any periodicity in each variable you like. $\endgroup$
    – fedja
    Jan 28, 2021 at 21:59
  • $\begingroup$ @fedja you are right, there were some additional constraints missing to give this question a meaning. $\endgroup$
    – Sascha
    Jan 28, 2021 at 22:31
  • 1
    $\begingroup$ By (an analog of) Cauchy-Riemann, $f_1+if_2$ only depends on $z_1+iz_2$ and $f_1-if_2$ only on $z_1-iz_2$, is this intended? $\endgroup$ Jan 31, 2021 at 15:39
  • $\begingroup$ @მამუკაჯიბლაძე what does it mean for $f_1+if_2$ to only depend on $z_1+iz_2$? What kind of independent coordinate system are you considering here? $\endgroup$ Feb 1, 2021 at 19:53
  • $\begingroup$ @PritamBemis If I am not confused (and I believe something similar is done in the answer by Robert Bryant below), if you change variables to $z_1+iz_2=z$, $z_1-iz_2=w$, and denote $f_1+if_2=g$, $f_1-if_2=h$, then $\partial g/\partial w=\partial h/\partial z=0$, no? $\endgroup$ Feb 1, 2021 at 19:55

1 Answer 1


The answer is that the only solutions have the form $$ f = (f_1,f_2) = \bigl(c, h(\,\overline{z}_1, z_2)\bigr) $$ where $h:\mathbb{C}^2\to\mathbb{C}$ is holomorphic and $c$ is a constant, which must equal zero unless $k_1$ and $k_2$ are integers.

The argument is as follows: The first equation implies that there exists a function $g:\mathbb{C}^2\to\mathbb{C}$ such that $$ f_1 = \frac{\partial g}{\partial z_1} \quad\text{and}\quad f_2 = -\frac{\partial g}{\partial z_2}. $$ Substituting this into the second equation implies that $g$ must satisfy $$ \frac{\partial^2 g}{\partial z_1\partial\overline{z}_1} + \frac{\partial^2 g}{\partial z_2\partial\overline{z}_2} = 0. $$ In other words $g$ is a harmonic function on $\mathbb{C}^2$. Since $g$ is harmonic, so is its derivative with respect to $z_1$, i.e., $f_1$.

The periodicity conditions imposed on $f_1$ imply that $f_1$ is bounded, and a bounded harmonic function on $\mathbb{C}^2$ is constant. Thus, $f_1 = c$ for some constant $c\in\mathbb{C}$. Obviously, $c$ must be zero unless $k_1$ and $k_2$ are integers.

Since $f_1$ is constant, the given equations on $f_2$ reduce to $$ \frac{\partial f_2}{\partial z_1} = \frac{\partial f_2}{\partial\overline{z}_2} = 0. $$ Hence $f_2 = h(\overline{z}_1,z_2)$ for some homorphic function $h:\mathbb{C}^2\to\mathbb{C}$.

Remark: It wasn't clear from the OP's question whether the OP wanted $f$ to be 'doubly-periodic' or just $f_1$, nor was it clear exactly what the OP meant by 'doubly-periodic' because, normally, the 'doubly-periodic' condition wouldn't have the exponential factors in its definition.


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.