Kissing number

The kissing number counts how many congruent spheres can simultaneously touch one central sphere without overlapping. After scaling, the spheres may all be taken as unit spheres, and their centers must lie at mutual distance at least 2 while each lies at distance 2 from the center of the middle sphere. Equivalently, the problem asks for the largest finite arrangement of points on an (n−1)-dimensional sphere whose pairwise angular separation is at least 60 degrees.¹ In one dimension the answer is 2, while in two dimensions it is 6: six equal circles fit around a seventh. In three dimensions the answer is 12, although the apparent possibility of a thirteenth sphere created a long-standing controversy.

The three-dimensional kissing number is 12, and proving that twelve is maximal is substantially harder than constructing twelve contacts. Newton and Gregory discussed the question in the seventeenth century, with Newton conjecturing that 13 might be possible and Gregory arguing for 12. The issue was finally settled in favor of Newton's numerical guess being false: no arrangement of 13 non-overlapping unit spheres can touch a central one. A modern proof combines spherical geometry, local density constraints, and careful case analysis; the result is often called the three-dimensional sphere-kissing theorem.² The problem illustrates a recurring feature of discrete geometry: a simple packing picture can conceal a difficult global optimization over continuously varying configurations.

Exact kissing numbers are known in dimensions 1, 2, 3, 4, 8, and 24, while many other dimensions have only upper and lower bounds. In dimension 4, the answer is 24, proved by work culminating in a rigorous solution using linear-programming and geometric methods. The exceptional values in dimensions 8 and 24 arise from highly structured configurations: the roots of the E8 lattice give 240 contacts, and the minimal vectors of the Leech lattice give 196,560 contacts.³⁴ These arrangements are closely related to error-correcting codes, modular forms, and optimal sphere packing. The kissing problem is therefore connected not only with geometry but also with algebra and number theory.

The kissing number is related to, but is not identical with, the sphere-packing density: it measures contacts around one sphere, whereas density measures the proportion of space occupied throughout a packing. A packing can have a large local kissing number without being globally densest. The contact points themselves form a spherical code, and their geometry can reveal symmetries of lattices or other extremal configurations. In dimensions 8 and 24, the same exceptional structures that determine the kissing arrangements also support proofs of optimal sphere-packing density.³⁴ More generally, bounds for τ(n) are obtained through linear programming, semidefinite programming, algebraic constructions, and computational certification. Determining exact values in all dimensions remains an open problem.

Unit sphere

A sphere of radius 1; in n-dimensional space, the analogous object is an n-dimensional ball boundary.

Spherical code

A finite set of points on a sphere subject to a prescribed minimum angular separation.

E8 lattice

An exceptional eight-dimensional lattice whose minimal vectors form a 240-point kissing configuration.

Leech lattice

An exceptional 24-dimensional lattice whose minimal vectors produce 196,560 contacts around a central sphere.

Sphere packing

The arrangement of non-overlapping congruent spheres, often studied through its maximum density or local contact structure.