I was reading some slides from a lecture. In a proof, there arose the need to show a certain function $f : \mathbb{C} \to \mathbb{C}$ was constant. The argument proceeded by checking that
- $f$ was entire
- $f(z+i) = f(z)$ for all $z$
- $f$ was bounded on $\mathbb{R}$
and then concluding that $f$ must be constant. I followed the proofs of the three claims no problem, but my complex analysis is weak enough that I'm unsure how one is supposed to conclude that $f$ is constant. Clearly Liouville's theorem does not apply directly. My guess is that some kind of boundary principle is being applied to the "rectangle" $R = \{ z \in \mathbb{C} : 0 \leq \Im(z) \leq 1 \}$. For instance, if the maximum of $|f|$ on $R$ must occur on the boundary, then the result follows. My complex analysis is patchy enough, though, that I'm unaware if such a result.
Added:
- Relevant: Lindelof's theorem
- The above article contains an enlightening example. If we take $f(z) = \exp(\exp(2 \pi z))$, then (1) and (2) are satisfied. (3) is "half satisfied" in the sense that $\lim_{t \to -\infty} f(z) = 1$ here.