Sukanta Das, Tarakanta Nayak
10 min
This survey paper by Sukanta Das and Tarakanta Nayak explores Baker wandering domains (BWDs), a fascinating class of Fatou components in the dynamics of transcendental meromorphic functions—holomorphic maps from the complex plane to the Riemann sphere with an essential singularity at infinity (and possibly poles). These functions partition the Riemann sphere into the Fatou set (where iterates behave nicely, forming a normal family) and the Julia set (the chaotic complement, containing poles and the backward orbit of infinity). Fatou components are the connected pieces of the Fatou set, and BWDs are a special wandering type: their iterates never overlap, stay bounded and multiply connected (surrounding holes), encircle the origin for large iterates, and drift to infinity (distance to 0 grows unboundedly).
Why care? BWDs, first constructed by Baker in 1963 for entire functions, reveal exotic dynamics impossible in rational maps (per Sullivan's theorem: no wandering domains there). The paper reviews constructions, examples, their impact on singular values and overall dynamics, classes of functions without BWDs, and open problems. It's a roadmap for studying 'tractable' transcendental dynamics where chaos is tamed in bounded regions.
A wandering domain is a Fatou component W where all iterates W^n = f^n(W) are pairwise disjoint. Baker's innovation: make them multiply connected (c(U) > 1, i.e., complement has multiple components), bounded, surrounding 0 with positive winding number for large n, and dist(W^n, 0) → ∞. Baker built this via nested annuli A_n where f(A_n) ⊂ A_{n+1} and f^n(z) → ∞ for z in A_1—initially unsure if components were distinct, but proved so in 1976. Crucially, these are bounded (Baker proved no unbounded multiply connected Fatou components in entire functions).
Contrast with Baker domains: periodic Fatou components where iterates tend to the essential singularity (unbounded, absent in rationals). BWDs 'wander' to infinity without periodicity.
Baker's 1963 entire function used conformal annuli mapping outward, escaping to ∞. Later examples expanded to meromorphic functions. The survey revisits these, highlighting how BWDs enable systematic study: their presence forces all Fatou components bounded (no Baker domains), allowing rational-map tools like Riemann-Hurwitz. Julia components simplify—essential singularity and backward orbit become 'singleton buried components,' curbing dynamical chaos from ∞.
Singular values (whose preimages aren't dense) are pivotal; BWDs restrict them—no finite asymptotic values, so singular set is just critical values and limit points (unbounded singular set overall). This shapes the escaping set {z : f^n(z) → ∞ without hitting ∞}. Boundedness tames infinity's pull, making dynamics more predictable. Paper discusses functions without BWDs and proposes problems, e.g., on omitted values.
Transcendental dynamics generalize polynomials/rationals but explode with ∞'s essential singularity. BWDs offer a 'model case'—bounded wandering regions escaping slowly—bridging to complex dynamics. Intuition: imagine donuts (annuli) orbiting 0 ever farther, mapped outward by f, staying 'stable' (Fatou) amid Julia chaos. Perfect for building intuition before harder cases.
Let $f:\mathbb C\to \widehat{\mathbb C}=\mathbb C \cup\{\infty\}$ be a transcendental meromorphic function (possibly without any pole) with a single essential singularity, and that is chosen to be at $\infty$. The set of points $z\in\mathbb{\widehat{C}}$ such that the family of iterates $\{f^n\}_{n\geq 0}$ is defined and forms a normal family in a neighborhood of $z$ is known as the Fatou set of $f$. For a Fatou component $W$, let $W_j$ denote the Fatou component containing $f^j(W)$. A Fatou component $W$ is called wandering if $W_m\bigcap W_n=\emptyset$ for all $m \neq n$. A wandering domain $W$ of $f$ is called a Baker wandering domain, if each $W_n$ is bounded, multiply connected, and $W_n$ surrounds $0$ for all large $n$ and, dist$(W_n,0)\to\infty$ as $n\to\infty$. This paper surveys the current state of knowledge on Baker wandering domains. We revisit the first example of the Baker wandering domain followed by other examples. The influence of Baker wandering domain on the singular values and dynamics of the function is presented. We also discuss some classes of functions that do not possess any Baker wandering domain. Several problems are proposed throughout the article at relevant places.
Alex: Wait—super-exponential growth of the radii... does that bury the wild infinity point completely?
Sam: Yes. The distances climb so rapidly that infinity gets isolated as a single buried point, with no paths from finite areas leading to calm spots asymptotically. All calm regions stay trapped in these expanding but finite annuli, and the chaotic Julia parts become singletons too. No finite values act as attractors from infinity, so singular points are just critical values—where the function flattens, derivative zero—and their clusters. Later examples tweak this for infinitely connected domains, full of holes like endless nested donuts, and even control the function's overall growth speed to any prescribed rate.
Alex: The paper mentions extending this to meromorphic functions too—with poles. How do those still get Baker wandering domains?
Sam: Meromorphic functions can have poles, points where they blow up to infinity, unlike entire functions that stay finite everywhere. For ones with only finitely many poles, the paper notes a clear rule: they have a Baker wandering domain if and only all their chaotic Julia components, except the one holding infinity, stay bounded. This mirrors the entire case but accounts for poles, as shown by Zheng in 2002.
Alex: Okay—so adding a few poles doesn't ruin the bounded setup. But what about functions with infinitely many poles?
Sam: Rippon and Stallard in 2005 built the first meromorphic example with a single pole, multiplying an entire function by a factor that stays close to one for points far out in the annuli. For points far out in the annuli, this preserves the ring-to-ring stretch into Fatou components that wander and escape. They even constructed one with infinitely many poles in 2005, using an infinite sum that adds poles but approaches one along the annuli, keeping the images inside the next rings—proving wandering non-periodic components.
Alex: And with these domains, what happens to the singular values—the special points driving the dynamics?
Sam: Singular values are points where the inverse function fails to be one-to-one nearby, like bottlenecks in undoing the map. They split into critical values, from points where the function flattens—or multiple poles—and asymptotic values, where paths to infinity approach the value without hitting it. In Baker wandering domains, no finite asymptotic values exist; singular sets are just critical values and their limits, simplifying analysis since infinity stays buried.
Alex: So the domains block those infinite paths to finite attractors. That really cleans up the picture.
Sam: The paper builds on known links between special points—singular values—and stable regions. Limit behaviors in wandering domains land in the post-singular set, which tracks where those singular values go under repeated mapping, or at infinity. Zheng extended this to meromorphic functions with few poles.
Alex: Post-singular set... so it's like following the family tree of trouble spots forward in time?
Sam: Yes. To connect this to Baker domains, they use the exponent of convergence for zeros—measuring how quickly zeros cluster near zero. If this exponent is less than the function's growth order, a theorem by Cao and Wang says some forward image of the domain must hit a singular value. This guarantees singular values enter the domains, shaping their connections.
Alex: Okay—and omitted values? Functions skipping some point entirely?
Sam: An omitted value is one the function never hits. Iversen showed every such is asymptotic—paths to infinity approach it without reaching. For Baker domains, neighborhoods around infinity have pre-images that are infinitely connected with bounded boundaries, defining a Baker omitted value at infinity for entire functions. Chakra's theorem links this: Baker domains imply such an omitted value, though the reverse fails, like for e^z plus linear terms.
Alex: Okay, so these buried singletons tidy up the Julia set. But what about points that escape to infinity—does a Baker wandering domain force them to behave in a certain way?
Sam: Some points under repeated mapping head straight to infinity fast enough to outpace any growing bound. Researchers track these in the escaping set, a collection that connects across the plane. The paper notes Rippon and Stallard showed in 2005 that with a Baker wandering domain, this set and a related one called A(f)—points escaping even faster—both become connected and unbounded, pulling in all the domains themselves.
Alex: Wait—so the domains get swept into this escaping web, linking Julia bits between them?
Sam: Yes. Their 2005 theorem says the closure of each domain sits inside A(f), and the escaping set joins them with paths hitting the Julia set. Zheng in 2000 extended this: in entire functions with these domains, every other wandering domain also escapes to infinity along some subsequence, making their union unbounded.
Alex: Huh. Like forcing all calm spots to flee outward eventually.
Sam: Precisely. This setup lets rational-map tools apply, like Riemann-Hurwitz for connectivity changes. The survey defines eventual connectivity: the fixed hole count in domains after many steps—either 2, infinite, or for non-Baker wanderers, simply connected at 1 in finite-pole cases.
Alex: Eventual connectivity... so the holes either settle to a simple ring or explode forever?
Sam: Yes. Kisaka-Shishikura in 2011 showed if finite holes start, they drop to 2 without critical points there; infinite stays infinite across the orbit. Rippon-Stallard in 2008 confirmed for finite-pole meromorphic: Baker ones end at 2 or infinite, others at 1. The paper focuses mainly on entire functions and meromorphic ones with finitely many poles, where clear rules hold—like all calm regions staying bounded if a Baker wandering domain exists. Cases with infinitely many poles are less explored, though examples exist; finding general sufficient conditions remains an open question.
Alex: Huh. So not every wild function at infinity fits this tidy bounded picture.
Sam: Right. For instance, adding a polynomial to e^z can skip Baker domains despite growth conditions that might suggest otherwise. The survey unifies what's known but highlights these gaps.
Alex: Okay, so it provides a solid foundation for studying unbounded chaos by taming it into bounded pieces analyzable like simpler maps.
Sam: Precisely. This bridges transcendental functions to rational map tools, like counting connections via Riemann-Hurwitz, enabling systematic progress on singularity dynamics. Many questions persist, but it's a meaningful step forward.
Alex: That's a clear picture of where the field stands. Thanks, Sam—this has been a thoughtful dive into Baker wandering domains.
Sam: My pleasure, Alex. Thanks for listening to this ResearchPod episode.