8
$\begingroup$

On page 24 of Krantz's Complex Analysis, there is the following proof:

Proposition 2: If $F$ is entire and $F$ has a removable singularity at $\infty$, then $F$ is constant.

Proof: By examining $F(1/z)$, we see that $F$ must have a finite limit at $\infty$. Thus $F$ is bounded. By Liouville's theorem, $F$ is constant.

It's mysterious to me what is meant by the first sentence. What does he mean by "examining $F(1/z)$"? I tried expanding $F$ has a Laurent series around $0$, and then using the fact that $F(1/z)$ has the origin as a removable singularity, but I didn't get the conclusion.

Is the idea to do this?

Let $$ F(z)=\sum_{n=-\infty}^{-1}a_nz^n+\sum_{n=0}^\infty a_nz^n $$ be the Laurent expansion around the origin. Then $$ F(1/z)=\sum_{n=-\infty}^{-1}a_nz^{-n}+\sum_{n=0}^\infty a_nz^{-n} $$ but since $F(1/z)$ has a removable singularity at the origin, we really have $$ F(1/z)=\sum_{n=-\infty}^{-1}a_nz^{-n}+a_0. $$ So $a_n=0$ for $n>0$, and thus $F(z)=\sum_{n=-\infty}^{-1}a_nz^n+a_0$, so $\lim_{z\to\infty}F(z)=a_0<\infty$?

$\endgroup$
  • 1
    $\begingroup$ Note you are essentially considering the transition map between coordinate patches on the Riemann sphere. $\endgroup$ – Potato Jun 20 '12 at 10:02
6
$\begingroup$

Your idea is definitely going in the right direction. You've found that $F(z)$ has a finite limit as $|z| \to \infty$, so you can find constants $M$ and $r$ such that $|F(z)| \leq M$ for all $|z| > r$. Now, since $F$ is entire, what can you say about $|F(z)|$ for $|z| \leq r$? Can you conclude that $|F(z)|$ is bounded for all $z$?

| cite | improve this answer | |
$\endgroup$
  • $\begingroup$ Does that just follow from the fact that $|z|\leq r$ is compact, and so $|F(z)|$ achieves a maximum on the disk? Since it's also bounded outside the disk, it must be bounded on all of $\mathbb{C}$. $\endgroup$ – Hana Bailey May 6 '12 at 5:19
  • $\begingroup$ Yup, that's it! $\endgroup$ – Antonio Vargas May 6 '12 at 5:19
  • $\begingroup$ Thanks for confirming Antonio. $\endgroup$ – Hana Bailey May 6 '12 at 5:20
2
$\begingroup$

Here's an approach via Laurent expansion:

Say we have the Taylor expansion around zero: $F(w)=\sum_{n=0}^{\infty} a_nw^n$, then if $w=1/z$ for $z\neq 0$, and by uniqueness of the Laurent expansion we have that this is exactly the Laurent series. Now, a removable singularity at infinity of $F(z)$ means a removable singularity for $F(1/z)$ at $z=0$, but this can only happen if $a_n=0$ for $n>0$.

| cite | improve this answer | |
$\endgroup$

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.