Here's my updated attempt:

Write$$f(z) = \sum_{n=-1}^{\infty} a_n(z-z_1)^n + ...+\sum_{n=-1}^{\infty} m_n(z-z_m)^n+\sum_{n=+1}^{-\infty} \psi_n(z)^n$$

with the last series being an expansion about the origin, and with $\large z=\frac{1}{w}$.

Now I subtract off the principal part of each Laurent series, getting

$$g:=\sum_{n=-1} a_n(z-z_1)^n + ...+\sum_{n=-1} m_n(z-z_m)^n - \frac{a_{-1}}{z-z_1} - ... - \frac{m_{-1}}{z-z_m}- {\psi_{1}}{z}$$

Now $g$ becomes entire and bounded, hence constant by Liouville's Theorem.

Moving the principal parts that I subtracted over to the L.H.S. gives me

$$ M + \frac{a_{-1}}{z-z_1} + ... + \frac{m_{-1}}{z-z_m}- {\psi_{1}}{z}= \sum_{n=-1}^{\infty} a_n(z-z_1)^n + ...+\sum_{n=-1}^{\infty} m_n(z-z_m)^n+\sum_{n=+1}^{-\infty} \psi_n(z)^n$$


So, I have shown that this meromorphic function is a rational function, as required.

What do you think? Thanks,

The problem statement is:

Let $f(z)$ be analytic in the whole complex plane apart from simple poles at $z_1,…,z_m$ and let $ f(\large \frac{1}{z})$ have a simple pole at $z=0$.

Part 1: Show that $f$ is a rational function $\large \frac{Q(z)}{P(z)}$ of polynomials $Q$ and $P$.

Part 2: What can you say about the degrees of $Q$ and $P$?

My work:

We know that $g(z)=(z-z_1)...(z-z_m)f(z)$ is an entire function.

This seems like a job for Liouville's Theorem, but I'm not sure how to utilize it right now, since I don't have a boundedness condition.

How could I use the assumption that $f(\frac{1}{z})$ has a simple pole at $z=0$?

I know that this means $f(z)$ has a simple pole at $\infty$, but this doesn't seem to help. I feel that this is the most important assumption in the question...

Any suggestions are welcome.



Subtract off the principal part of each pole, including the one at $\infty$, and you have a bounded entire function, which by Liouville is constant. What you subtracted was a rational function...

  • $\begingroup$ Hi Professor Israel - how does a pole have a principal part? I know that Laurent series has a principal part = all the negative-power terms in z. Are you referring to the same thing? Thanks, $\endgroup$ – User001 Dec 1 '15 at 8:24
  • $\begingroup$ Do you mean to subtract off the principal part of $m$ different Laurent series, each series being an expansion about one of the poles? Thanks @RobertIsrael, $\endgroup$ – User001 Dec 1 '15 at 8:55
  • 1
    $\begingroup$ Yes, the principal part at $z=a$ of a function (analytic in a deleted neighbourhood of $a$) is the principal part of the Laurent series about $z=a$. After subtracting that, you have a function whose singularity at $z=a$ is removable; all the other singularities are the same as for the original function. $\endgroup$ – Robert Israel Dec 1 '15 at 16:39
  • 1
    $\begingroup$ If $f(1/w) = c/w + \ldots$ as $w \to 0$, writing $z = 1/w$ gives you $f(z) = c z + \ldots$. The principal part at $\infty $ is then $c z$. $\endgroup$ – Robert Israel Dec 2 '15 at 3:42
  • 1
    $\begingroup$ Your notation is not good, since the $c_n$ for different $m$ are different. $\endgroup$ – Robert Israel Dec 2 '15 at 3:43

As $f(\frac1z)$ has no zeroes near $z=0$, the set of zeroes of $f$ is bounded, hence finite, say (with multiplicity) they are $w_1,\ldots,w_n$. Then $$h(z)=\frac{(z-z_1)\cdots (z-z_m)}{(z-w_1)\cdots (z-w_n)} f(z)$$ is entire, has no zeroes and $h(\frac1z)$ has a pole of order $m+1-n$ (if $m+1>n$) or a zero of order $n-1-m$ (if $m+1<n$) or a removable discontinuity (iff $n=m+1$)at $z=0$. As $h$ has no zeroes, we can also consider $\frac1{h(\frac1z)}$ which is entire except possibly a pole at $z=0$ (if $m+1<n$). At any rate, at least one of $h(\frac1z)$, $\frac1{h(\frac1z)}$ is entire, hence bounded near $z=0$. Correspondingly, one of $h(z)$, $\frac1{h(z)}$ is entire and bounded (near $\infty$ and hence throughout), hence constant. We conclude $$ f(z)=c\cdot \frac{(z-w_1)\cdots (z-w_n)} {(z-z_1)\cdots (z-z_m)}$$ with $c\ne 0$ and $n=m+1$.

  • $\begingroup$ Hi @HagenvonEitzen, would it be easier / valid if instead I could show that my $g$ is a polynomial? My $g$ I think should have a pole at infinity, since on the R.H.S., $f$ has a pole at infinity. I am a little unfamiliar with your technique, so I wonder whether I could still proceed in my current direction ... thanks $\endgroup$ – User001 Dec 1 '15 at 7:26

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.