# What “trapped loops near infinity” means

[project] This makes the earlier infinity claim precise. It concerns the intrinsic xi zero set, not a new zeta function, a proved horned embedding, or an RH criterion. Multiplicities do not affect the underlying complement topology.

## The rotational reduction works without RH

[known/project] Let d=4 or 8. Write q=x+rI, r>0, I in S^(d-2). Let P be the distinct upper-half-plane zeros beta+i gamma of xi and Z their spherical lifts. Xi has no real zeros, P is locally finite, its real coordinates lie in (0,1), and its imaginary coordinates are unbounded. No assumption beta=1/2 is made. Removing the real axis gives the homeomorphism

    (R^d \ Z) \ R = ((R × (0,infinity)) \ P) × S^(d-2).

[project — proof] S2 and S6 are simply connected. Restoring the real axis does not change pi1: in an open d-manifold, a loop and a disk homotopy can be perturbed off a smooth line, whose codimension is d-1 >=3. The same argument works outside a sphere centred at the origin. Thus the fundamental group is that of a punctured upper half-plane, or its exterior-half-disk version. Van Kampen gives F_m for m finite punctures, and the free group on countably many generators for a locally finite infinite set. Loops and homotopies are compact; only finite words occur. This is NOT the Hawaiian-earring group. Background: Hatcher, Algebraic Topology, §1.2 (van Kampen); xi zero facts are classical, independent of RH.

## The inverse system, not just “an infinite group”

[project — exact end calculation] Choose increasing radii R_n tending to infinity, none passing through a zero sphere, and a proper base ray on the positive real side, away from the zeros. Put

    U_n = {q : |q| > R_n} \ Z,
    P_n = {p in P : |p| > R_n}.

These U_n are connected. They give a cofinal neighborhood system for the ambient-infinity end of the complement. To see this in the usual compact-exhaustion definition of ends, excise tiny disjoint tubes around the finitely many spheres inside each ball. The ball minus those open tubes is compact in the complement; its exterior component is U_n. Shrink the tube thicknesses as n grows. Other ends approach individual deleted finite spheres; we are singling out the end escaping every ambient ball, not claiming the complement has only one end.

[project] With compatible meridian paths along the base ray,

    pi1(U_n) = F(P_n),
    pi1(U_(n+1)) --> pi1(U_n)

is the inclusion sending each surviving meridian to itself (up to the standard base-path choices). It omits the finitely many earlier generators. Radii can omit a batch if several spheres have the same norm. A schematic basis version is

    F<g1,g2,g3,...> <-- F<g2,g3,...> <-- F<g3,...> <-- ...

All groups are abstractly countably generated free groups; the bonding maps, not their abstract isomorphism types, retain the tail information.

[project — important distinction] This pro-group is NOT pro-trivial: for every n and every farther m, some still farther sphere has a meridian whose image in pi1(U_n) is nontrivial. Yet the ordinary inverse limit in this compatible model IS trivial: an element in every tail subgroup would be a finite reduced word using no finite-index generator, hence the empty word. A trivial inverse limit does not imply a trivial end system. This is the precise sense in which new trapped loops keep appearing, without one fixed finite word surviving forever.

## The local-connectivity test at the compactified point

[project — proof] In S^d=R^d union {infinity}, let K=Z union {infinity}. K fails the complement-side 1-LCC condition at infinity: there is no smaller neighborhood V inside a neighborhood U of infinity such that every loop in V\K contracts in U\K. A meridian of a sufficiently distant sphere fits inside V; its winding about the orbit puncture is 1 and cannot contract even in the whole complement. Its nonzero integer homology class already detects it.

[project] K is not locally connected at infinity. Every individual sphere is a clopen component of K. Consequently the connected component of infinity is just that point, but every neighborhood of infinity contains infinitely many spherical components. There is no connected open neighborhood basis there. K is therefore not a topological sphere or a manifold at infinity.

## Comparison with Alexander's horned sphere: a genuine difference

[known] Alexander's 1924 construction is a single topological S2 embedded in S3; its recursively interlocking horns give a non-simply-connected exterior. The original three-page paper was inspected: J. W. Alexander, “An Example of a Simply Connected Surface Bounding a Region which is not Simply Connected,” PNAS 10 (1924), 8–10, https://doi.org/10.1073/pnas.10.1.8. The problematic limiting points lie on a still-locally-connected topological sphere.

[known/project] Alexander duality gives H1(S3\S2)=0 for this embedded topological sphere, even though its exterior pi1 is nontrivial. In our configuration H1 is the direct sum of one Z per zero sphere: the meridians are already visible after abelianizing. This is a global homology comparison, not a claim to compute the horned sphere's entire local pro-group. Source: Hatcher, Algebraic Topology, Corollary 3.45 (Alexander duality), https://pi.math.cornell.edu/~hatcher/AT/ATch3.pdf.

[project] Both examples warn against assuming loops disappear in smaller neighborhoods. But their mechanisms differ: infinitely many separate, ordinary codimension-two holes versus a single codimension-one sphere with interlocking horns. Even a smooth codimension-two sheet has a local meridian, so failure of the complement's 1-LCC test alone is not a certificate of wild embedding. Calling xi's compactified zero set an Alexander horned sphere would be false.

## Numerical sanity checks and illustration

[project — finite checks only] `infinity.py` checks inversion of four actual on-line zero spheres and their orbit-plane meridian windings. Inversion maps

    beta + gamma S^(d-2)
      to beta/(beta^2+gamma^2) + gamma/(beta^2+gamma^2) S^(d-2).

The spheres shrink toward the compactified point while keeping a meridian. `figures/infinity-tail-system.svg` illustrates the orbit holes, inversion and the tail inclusions. Numerical samples illustrate the theorem; they do not prove RH or the infinite statement. Further work should examine actual norm-level tubes rather than mistaking the end calculation for arithmetic progress.
