I'm having trouble completing a homework question which will produce an alternative proof for Liouville's Theorem. The question reads

Let $f$ be an entire function. Evaluate, for $|a|<R,|b|<R,\int_{|z|=R}\frac{f(z)dz}{(z-a)(z-b)}$. When $f$ is bounded, let $R\to \infty$ and deduce another proof for Liouville's Theorem.

Liouville's Theorem: If a function f is entire and bounded in the complex plane, then $f(z)$ is constant throughout the plane.

So I begin by evaluating the integral. Using the Cauchy-Residue theorem (and skipping some methodic computation) I have that $\int_{|z|=R}\frac{f(z)dz}{(z-a)(z-b)}=2\pi i[\frac{f(a)-f(b)}{a-b}].$Now, it is given that $f$ is entire (analytic everywhere) and bounded. Thus all I must show is that it is constant. Since $f$ is bounded there exists some non-negative number $M$ such that $|f(z)|<M$.

I proceed by using the M-L Lemma in attempts to try and generate some ideas.

$|\frac{f(z)dz}{(z-a)(z-b)}|\leq\frac{M}{(R-a)(R-b)}\times 2\pi R$, by the conditions given in the question and the use of the reverse triangle inequality.

Hence, $|\int_{|z|=R}\frac{f(z)dz}{(z-a)(z-b)}|\leq\frac{M}{(R-a)(R-b)}\times 2\pi R$. Now as $R\to\infty$, the RHS of the inequality clearly approaches 0. Thus $\int_{|z|=R}\frac{f(z)dz}{(z-a)(z-b)}=0?$ I think I saw this in one of my text books. Does this lead on to the conclusion that $f(z)$ is constant? But then I realise I have not made use of the equation I established above.

Thank you to all in advanced for you help.


Fix $a$. Take $b$ arbitrarily and you showed that $$ \big| f(a) - f(b) \big| \leq \text{constant}\cdot \frac{M}{R} $$ taking $R \to \infty$ you get that $f(a) = f(b)$. Therefore for any $b \in \mathbb{C}$ you have $f(b) = f(a)$, implying that $f$ is constant.

  • $\begingroup$ I'm not quite on track. So I may understand that $|f(z)-f(b)|\leq 0$ and $R \to \infty$, but the whole idea behind fixing $a$, $b$ and the equality established is a little confusing. Can you please elaborate? Thanks. @Tom $\endgroup$ – Gustavo Louis G. Montańo Oct 31 '13 at 13:09
  • 1
    $\begingroup$ Sure! Let $a \in \mathbb{C}$ and set $P = f(a)$. Now take any arbitrary $z \in \mathbb{C}$ and as you have written you have shown that for $R$ large enough $|f(z) - f(a)|\leq \text{constant}\cdot R^{-1}$. But since this is true for any $R$ large enough, then you can take it as large as you want, which forces the inequality $|f(z) - f(a)| \leq 0$ to be true for any $z$. Since $0 \leq |f(z)-f(a)| \leq 0$ then $f(z)=f(a)$ for every $z \in \mathbb{C}$. That is $f(z) = P$ for every $z \in \mathbb{C}$, meaning that $f$ is constant. $\endgroup$ – Tom Oct 31 '13 at 13:17
  • 1
    $\begingroup$ Brilliant, @Tom. Thanks a lot mate! That made perfect sense. $\endgroup$ – Gustavo Louis G. Montańo Oct 31 '13 at 13:19
  • $\begingroup$ @eXtremiity Glad to help! $\endgroup$ – Tom Oct 31 '13 at 13:27

You just computed the result in two ways so you have that for all a,b complex numbers $$f(a)=f(b)$$ This happens only when $f$ is constant


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.