Gaussian moat

About 64 years old · traced to

Let Z[i]={a+bi:a,b∈Z}\mathbb{Z}[i]=\{a+bi : a,b\in\mathbb{Z}\} be the ring of Gaussian integers, with norm N(a+bi)=a2+b2N(a+bi)=a^{2}+b^{2} and Euclidean absolute value ∣a+bi∣=a2+b2|a+bi|=\sqrt{a^{2}+b^{2}}. Call π∈Z[i]\pi\in\mathbb{Z}[i] a Gaussian prime if π\pi is a nonzero non-unit whose only divisors in Z[i]\mathbb{Z}[i] are units and unit multiples of π\pi; equivalently, π\pi is a Gaussian prime exactly when N(π)N(\pi) is a rational prime, or π=up\pi=u p with u∈{1,−1,i,−i}u\in\{1,-1,i,-i\} and pp a rational prime with p≡3(mod4)p\equiv 3 \pmod 4.

Then there is no infinite sequence π1,π2,π3,…\pi_{1},\pi_{2},\pi_{3},\dots of pairwise distinct Gaussian primes such that

sup⁡n≥1 ∣πn+1−πn∣<∞.\sup_{n\ge 1}\,\bigl|\pi_{n+1}-\pi_{n}\bigr|<\infty .

Equivalently, for each real k>0k>0 let GkG_{k} be the graph whose vertex set is the set of Gaussian primes and in which two distinct Gaussian primes π,π′\pi,\pi' are joined by an edge when ∣π−π′∣≤k|\pi-\pi'|\le k; then every connected component of GkG_{k} is finite. In particular, for every k>0k>0 the component of GkG_{k} containing the Gaussian prime 1+i1+i is finite, so that only finitely many Gaussian primes can be reached from the origin by steps of length at most kk through Gaussian primes.

References

Primary source

Wikipedia

Additional references

  1. Wikipedia, Gaussian moat, the article this problem comes from.

Progress summary

Refreshed
Claimed progress

The problem remains open: an unverified 2019 paper claims a proof only for Gaussian primes off the coordinate axes.

Basil Gordon posed the question in 1962: can one walk arbitrarily far through Gaussian primes using steps bounded by one fixed length? The general negative answer has not been established.

Known results

  • Jordan and Rabung constructed a 10\sqrt{10}-moat in 1970.
  • Gethner et al. constructed 44-, 18\sqrt{18}-, and 26\sqrt{26}-moats in 1998.
  • Tsuchimura constructed a 66-moat in 2004. These results exclude walks with step length at most 55, but do not settle whether some finite bound permits an infinite walk.

2019 restricted claim

A 2019 arXiv note, “A Note on The Gaussian Moat Problem,” claims that no infinite bounded-step sequence exists for Gaussian primes a+bia+bi with a,b≠0a,b\ne0. This is only a restricted version of the stated problem, and the claim is unverified; the paper also acknowledges an unresolved issue about whether its constructed moats separate the origin from infinity.

Current status (as of October 2026): The general Gaussian moat problem remains open; classical results rule out bounded walks through step length 55, while the 2019 restricted proof claim is unverified and does not cover all Gaussian primes.

Sources

Solutions 0

No solutions have been posted yet.