The following theorem appears in "Princeton Lectures in Analysis, Complex Analysis," by Stein & Shakarchi:

Suppose $\{F_n\}$ is a sequence of holomorphic functions on the open set $\Omega$.
If there exists constants $c_n \gt 0$ s.t. $\sum c_n < \infty$ and $|F_n(z) - 1| \le c_n$, for all $z \in \Omega$, then
The product $\prod_{n=1}^\infty F_n(z)$ converges uniformly in $\Omega$ to a holomorphic function $F(z)$

The idea is that we can write $\prod_{n=1}^N F_n(z)$ as $\prod_{n=1}^Ne^{\text{log}F_n(z)} = e^{\sum_{n=1}^N\text{log}F_n(z)}$
Using the fact that $|\text{log} f| \le 2|f-1|$ for $f$ close enough to $1$, together with $|F_n(z) - 1| \le c_n$ and $\sum c_n < \infty$, it follows that $e^{\sum_{n=1}^N\text{log}F_n(z)}$ converges to some function $F(z)$,
but how do we know that it converges uniformly?


Hint: bound $\left|\sum_{n=M}^N \log F_n(z)\right|$

  • $\begingroup$ I may have some clue at what you're getting at. So I want to write $e^{\sum_{n=1}^{M-1}\text{log}F_n}e^{\sum_{n=M}^{N}\text{log}F_n}$ and then I can show that the second exponential term is close to $1$. But that only allows me to get within a multiplicative constant of my limit function, not an additive constant, right? $\endgroup$ – Mark Sep 20 '12 at 1:52
  • $\begingroup$ If the function is bounded, a multiplicative factor is as good as an additive one. $\endgroup$ – Robert Israel Sep 20 '12 at 7:46
  • $\begingroup$ That's a great point! $\endgroup$ – Mark Oct 5 '12 at 2:24

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.