Spherical Skeletons: Theory & Applications

• Spherical skeletons are abstract combinatorial structures that classify inscribed polyhedra through precise graph-theoretic representations and color-coded edge cycles.
• They serve as essential invariants in spherical varieties, encapsulating Luna–Vust data and enabling reconstruction of Cox rings and divisor class groups.
• Utilizing combinatorial, linear programming, and metric criteria, spherical skeletons provide actionable tests for geometric realizability and topological decompositions.

A spherical skeleton is an abstract combinatorial structure originating in both convex polyhedral geometry and the theory of spherical varieties. The term encapsulates, depending on context, either (1) a purely graph-theoretic encoding of which planar graphs arise as the 1-skeleta of convex polyhedra whose vertices lie on a sphere (inscribed polyhedra), or (2) a combinatorial invariant which encodes essential data governing the structure and classification of spherical varieties under a reductive group action, capturing their Luna–Vust invariants, toricness, and other properties. In both settings, the spherical skeleton plays a decisive role in classifying realizable geometric objects, specifying their algebraic invariants, and providing explicit combinatorial or numerical criteria for geometric and topological properties.

1. Spherical Skeletons in Polyhedral Geometry

The notion of a "spherical skeleton" for convex polyhedra concerns the classification of those combinatorial graphs (specifically 3-connected planar graphs) which serve as the 1-skeleton (vertex and edge graph) of a polyhedron inscribed to a sphere in projective space, i.e., all vertices lying on a fixed sphere $S \subset \mathbb{R}\mathbb{P}^3$ (Chen et al., 2017).

This characterization fundamentally distinguishes between strongly inscribed polyhedra (all vertices on the same side of $S$) and weakly inscribed polyhedra (vertices distributed on both sides). The combinatorial characterization is as follows:

• The vertex set $V$ decomposes as $V^+ \sqcup V^-$, covered by exactly two vertex-disjoint cycles of lengths $p = |V^+|$ and $q = |V^-|$.
• Edges are colored red (within $V^+$ or $V^-$) or blue (between $V^+$ and $V^-$). There must exist a closed cycle in the skeleton graph, traversing all edges, in which the red/blue edge-color pattern alternates in a prescribed way.

This alternation reflects the geometric transition between hyperbolic and de Sitter regions in the model $\mathbb{H}^3 \cup dS^3$ where the sphere forms the ideal boundary. The spherical skeleton thus encodes the ability to patch hyperbolic and de Sitter "faces" while guaranteeing geometric realizability as a weakly inscribed polyhedron.

2. Combinatorial and Linear Programming Characterizations

The classification admits an equivalent formulation via real weight functions on edges, interpreted as exterior dihedral angles. For a graph to arise as a spherical skeleton of a weakly inscribed polyhedron, there must exist $w: E \to \mathbb{R}$ such that:

• $w_e > 0$ for every red edge, $w_e < 0$ for every blue edge.
• Vertex angle sums: $\sum_{e \ni v} w_e = 0$ for all $v$ except in the isolated $p=1$ case, where $\sum_{e \ni v_0} w_e = -2\pi$.

Feasibility of this system is necessary and sufficient for combinatorial realizability, with no additional constraints required, in contrast to the strong (hyperbolic) case which requires extra cut-inequalities. The correspondence between combinatorics and geometry is made precise through an infinitesimal rigidity argument, showing that the assignment from geometric parameters (dihedral angles) to combinatorial data (the skeleton) is bijective and local homeomorphism (Chen et al., 2017).

3. Spherical Skeletons and Spherical Varieties

In representation theory and algebraic geometry, the spherical skeleton appears as a coarse combinatorial invariant that encapsulates the essential Luna–Vust classification data of a spherical $G$-variety $X$ (Gagliardi, 2016, Gagliardi et al., 10 Jan 2026). Specifically, the skeleton comprises:

• The set of spherical roots $\Sigma$ (primitive generators of the tail cone in the weight lattice $M$).
• The subset $S^p$ of simple roots whose minimal parabolics fix every color.
• The set of colors $D^a$ (type $a$ $B$-invariant prime divisors).
• The set of $G$-invariant prime divisors $\Gamma$.
• Associated maps: $\varsigma$ (from colors $D^a$ to subsets of $S$) and $\rho'$ (restriction of valuation maps to the sublattice generated by $\Sigma$).

The spherical skeleton serves as a minimal combinatorial "fingerprint," sufficient to reconstruct the Cox ring $R(X)$ and, via associated maps and cones, to determine further geometric/algebraic properties.

4. Numerical and Smoothness Criteria

Recent advances show that the spherical skeleton enables purely combinatorial and numerical criteria for toricness and local smoothness of spherical varieties (Gagliardi et al., 10 Jan 2026). Introducing the functional $\widetilde{\wp}(\mathcal{R}_X)$, defined from anticanonical coefficients, valuation cones, and spherical roots, yields:

• $\widetilde{\wp}(\mathcal{R}_X) = 0$ if and only if the variety is toric.
• $\widetilde{\wp}(\mathcal{R}_{X,I}) < 1$ if and only if the variety is smooth along a $G$-orbit corresponding to skeleton elements $I$.

These criteria dispense with the need for detailed Luna–Vust colored fan data or external reference tables, streamlining verification and classification directly via the skeleton.

5. Spherical Skeletons in Skeletal Factorizations and Topological Combinatorics

The theory of spherical skeletons extends to topological combinatorics through decomposition theorems for skeleta of Platonic polytopes into canonical spheres (Hammack et al., 2021). For example, the $k$-skeleton of a cross-polytope $O_n$ can be partitioned into the boundaries of $(k+1)$-cross-polytopes, and similar factorizations are established for simplices and hypercubes via the existence of combinatorial block designs (Keevash’s theorem). Each spherical skeleton here encodes a decomposition into disjoint subcomplexes, each topologically a piecewise-linear sphere, tightly linking geometric realization and combinatorial design theory.

6. Geometric Stretch Factors and Metric Properties

In polyhedral geometry, the spherical skeleton directly governs metric properties, such as the stretch factor of the skeleton graph of convex polyhedra with vertices on a sphere (Bose et al., 2015). For the unit-sphere case, the skeleton is a $(0.999\cdot\pi)$-spanner: for any pair of vertices $p, q$, the shortest path in the skeleton does not exceed $0.999\cdot\pi$ times the Euclidean distance. When vertices are only approximately spherical, stretch factors can be arbitrarily large. These results are derived using convex cycle dilation bounds and surface-unfolding lemmas, further connecting combinatorial structure to geometric performance.

7. Examples, Counterexamples, and Unifying Framework

The spherical skeleton framework explicates both realizable and non-realizable cases with explicit combinatorial witnesses (Chen et al., 2017). For instance, specific cycle covers and alternation patterns guarantee realizability as a weakly inscribed polyhedron, whereas failures of the alternation or LP feasibility criteria provide obstructions. The approach generalizes classical problems, such as Steiner’s question on inscribed polytopes, and unifies the combinatorics underlying inscribability in hyperbolic, de Sitter, and mixed spaces. In spherical varieties, skeleton data suffices for complete Cox ring reconstruction, divisor class group analysis, and iteration stabilization (Gagliardi, 2016).

Spherical skeletons thus form a versatile and unifying concept spanning polyhedral geometry, algebraic group actions, metric graph theory, combinatorial design, and topological manifold decompositions, precisely encoding the structure and properties of spherical configurations in multiple mathematical domains.