Rigid Quadrangulations: Structure & Applications

• Rigid quadrangulations are quadrangulations with strictly controlled combinatorial and geometric properties defined by flatness, topology, boundary data, or hinge geometry.
• They establish bijections with colorful labeled maps and decorated trees, enabling precise enumerative series and moduli space representations for discrete flat disks.
• Their rigidity ensures unique reconstruction from boundary metrics and controls topological invariants, underpinning stability in both surface graph theory and polyhedral mesh structures.

Rigid quadrangulations are quadrangulations whose admissible local and global combinatorics are sharply constrained by flatness, topology, boundary data, or hinge geometry. In recent work, the phrase has acquired several precise meanings. For rooted planar maps, it denotes flat quadrangulations of the disk with directed rays constrained to run from concave to straight boundary vertices (Budd, 29 Sep 2025). For surface quadrangulations tied to the Lovász complex, it denotes non-bipartite quadrangulations in which every $4$-cycle is facial, under hypotheses making this equivalent to the associated Lovász complex being a closed surface (Arana et al., 4 Oct 2025). In other strands of the literature, rigidity describes uniqueness from boundary distances, extremality of order, genus, and chromatic number, or the generic absence of nontrivial flexes in quadrilateral mesh neighborhoods (Haslegrave, 2021, Lawrencenko, 2012, Izmestiev, 2014).

1. Terminology and principal notions

The term is used in several adjacent but non-identical senses. In the disk setting, rigidity is a combinatorial-flatness condition on a rooted planar map. In topological graph theory, it is a facial $4$-cycle condition that forces a surface structure on the Lovász complex. In inverse and extremal problems, it refers either to reconstruction from boundary distances or to quadrangulations whose parameters are tightly controlled by a spine. In polyhedral geometry, it refers to the local or global absence of nontrivial isometric deformations of quadrilateral meshes.

Setting	Defining constraint	Representative consequence
Disk quadrangulations	Inner vertices of degree $4$; ray constraints	Bijections to colorful quadrangulations and decorated trees
Lovász-complex surfaces	Non-bipartite; every $4$-cycle facial	$\Lo(G)$ is a closed surface
Disc quadrangulations from boundary data	Internal degrees at least $4$	Boundary distances determine the map
Spinal quadrangulations	$G[:]\hookrightarrow \Sigma_{\beta(G)}$	Genus, order, chromatic number are controlled
Quad-mesh mechanisms	Kokotsakis patch conditions	Generic rigidity; flexible families are exceptional

This multiplicity of usage is structural rather than accidental. Each meaning isolates a regime in which quadrangular faces do not admit arbitrary rearrangement: the combinatorics of rays, the topology of $4$-cycles, the boundary metric, the spine, or the hinge geometry suppresses alternative configurations.

2. Flat disk models and the combinatorial notion of rigidity

A quadrangulation of the disk of perimeter $2k$ is a rooted planar map in which all faces have degree $4$, except the root face, which has degree $4$0, and the boundary is oriented counterclockwise. It is called flat if every inner vertex has degree $4$1, the boundary is simple, and the root vertex is convex. Boundary vertices are classified by degree: convex if the degree is $4$2, straight if the degree is $4$3, and concave if the degree is at least $4$4. A ray is a path of inner edges starting and ending at boundary vertices, visiting only inner vertices in between, and going “straight” at each inner vertex; each inner edge belongs to a unique ray. A rigid quadrangulation of the disk is then a flat quadrangulation with two extra constraints: all concave boundary vertices are of degree $4$5, and each ray connects a concave boundary corner to a straight boundary vertex (Budd, 29 Sep 2025).

This definition concentrates curvature on the boundary. The interior is locally modeled on the square lattice, while the boundary encodes the deviation from Euclidean flatness. For $4$6, the class $4$7 consists of rigid quadrangulations having $4$8 convex boundary corners, including the root corner, and $4$9 concave boundary corners. The union $4$0 is the main enumerative object in the recent discrete-flat-disk literature.

A closely related formulation uses rooted quadrangulations of the disk in which the root face is simple, the boundary is decomposed into sides, and rays are classified as closed or open. In that language, a rigid quadrangulation is a flat quadrangulation of the disk such that the root edge starts at a convex corner and every ray is either a closed ray from a concave corner to a straight boundary vertex or an open ray oriented from an open side. A rigid quadrangulation is complete if it has no open sides other than possibly the base; Budd’s original definition corresponds to the complete case with no open sides at all. This framework introduces complete base-$4$1 rigid quadrangulations and the generating functions

$4$2

for objects with $4$3 boundary corners (Zonneveld, 29 Sep 2025).

These constructions are not only combinatorial. A rectilinear metric on the unit disk is flat in the interior and has piecewise geodesic boundary with corner angles $4$4 or $4$5. Extending rays from concave corners subdivides such a disk into rectangles, and for generic disks the resulting combinatorial type is a rigid quadrangulation. This suggests that rigid quadrangulations are the discrete combinatorial types of top-dimensional cells in a moduli space of rectilinear flat disks.

3. Bijections, generating functions, and random geometry

The central structural result for disk-type rigid quadrangulations is a bijection

$4$6

between rigid quadrangulations of the disk and colorful $4$7-labeled quadrangulations of the sphere. On the labeled side, every edge joins labels differing by exactly $4$8, the root edge is oriented from label $4$9 to label $4$0, and every face has cyclic label pattern $4$1. The bijection is built by labeling non-root convex corners by turning numbers, doubling the disk to a sphere, and converting rows and columns of the rigid quadrangulation into even and odd edges of a colorful quadrangulation. Its dictionary is explicit: non-root convex corners correspond to vertices, rows to even edges, columns to odd edges, right-tangential convex corners to local minima, left-tangential convex corners to local maxima, right-tangential sides to decreasing level lines, and left-tangential sides to increasing level lines (Budd, 29 Sep 2025).

The same paper gives the basic counting series

$4$2

and characterizes it through the unique formal power series $4$3 satisfying

$4$4

The resulting identity is

$4$5

with asymptotic behavior

$4$6

The moduli-space interpretation is equally explicit: if $4$7 denotes rectilinear disks with $4$8 corners and half-perimeter $4$9, then

$\Lo(G)$0

Thus rigid quadrangulations index the top-dimensional cells of $\Lo(G)$1.

A second bijective theory replaces colorful sphere maps by decorated trees. For every $\Lo(G)$2 and $\Lo(G)$3, there is a bijection between complete base-$\Lo(G)$4 rigid quadrangulations with $\Lo(G)$5 boundary corners and H-trees of degree $\Lo(G)$6 and base-length $\Lo(G)$7. The decomposition is canonical: every complete base-$\Lo(G)$8 rigid quadrangulation with $\Lo(G)$9 contains exactly one minimal submap of an allowed type, and recursive removal of these submaps yields a binary tree whose vertices are decorated by signatures $4$0. H-trees are then transferred to Q-trees, where the generating functions become explicit in the same inverse series $4$1 (Zonneveld, 29 Sep 2025).

A related but distinct decorated model is furnished by rigid loop-$4$2 quadrangulations with a boundary, where each internal face is either empty or crossed by a loop through opposite edges. In the non-generic critical regime, the volume $4$3 admits explicit scaling limits: for $4$4,

$4$5

where $4$6 is the limit of the additive Malthusian martingale of a multiplicative cascade; for $4$7,

$4$8

where $4$9 is the limit of the derivative martingale and $G[:]\hookrightarrow \Sigma_{\beta(G)}$0 is exponential with parameter $G[:]\hookrightarrow \Sigma_{\beta(G)}$1 (Aïdékon et al., 2024). This places rigid quadrangulations in direct contact with multiplicative cascades and Liouville quantum gravity.

4. Surface rigidity via the Lovász complex

For a finite graph $G[:]\hookrightarrow \Sigma_{\beta(G)}$2, the neighborhood complex is

$G[:]\hookrightarrow \Sigma_{\beta(G)}$3

and the Lovász complex $G[:]\hookrightarrow \Sigma_{\beta(G)}$4 is the induced subcomplex of $G[:]\hookrightarrow \Sigma_{\beta(G)}$5 on the vertex set

$G[:]\hookrightarrow \Sigma_{\beta(G)}$6

It carries a canonical free $G[:]\hookrightarrow \Sigma_{\beta(G)}$7-action given by $G[:]\hookrightarrow \Sigma_{\beta(G)}$8. The central classification theorem states that, under mild assumptions, $G[:]\hookrightarrow \Sigma_{\beta(G)}$9 is homeomorphic to a closed surface if and only if $4$0 is a non-bipartite quadrangulation of $4$1 in which every $4$2-cycle is facial (Arana et al., 4 Oct 2025).

The graph-theoretic hypotheses are explicit: $4$3 is finite, connected, not isomorphic to $4$4, no neighborhood dominates another, and $4$5 is $4$6-free. Under the forward implication, the vertices of $4$7 are exactly of three kinds: a pair of opposite vertices in some face of $4$8, a neighborhood $4$9, or a singleton $2k$0. Maximal simplices are triangles, and the quotient by the canonical $2k$1-action recovers the original embedded graph. Under the reverse implication, the same local structure forces $2k$2 to be a quadrangulation of $2k$3, and all $2k$4-cycles must bound faces.

The facial $2k$5-cycle condition is the rigidity mechanism. Observation 3.5 states that if a non-bipartite quadrangulation has a non-facial $2k$6-cycle, then $2k$7 is not a surface: an edge in $2k$8 lies in three different triangles, so a link is not a circle. This makes “every $2k$9-cycle is facial” exactly the condition preventing local singularities. The paper therefore treats these quadrangulations as rigidity-like objects: the only possible $4$0-closed sets are singletons, neighborhoods, and face diagonals, and the induced symmetric quadrangulation on $4$1 rigidly encodes $4$2.

The same framework gives a topological strengthening of chromatic statements. If $4$3 is an odd quadrangulation of a non-orientable surface, then

$4$4

hence

$4$5

For such surface Lovász complexes, the topological obstruction is stronger than the chromatic inequality alone.

5. Boundary-distance and extremal rigidity on surfaces

A different notion of rigidity arises for disc quadrangulations determined by their boundary metric. A near-quadrangulation with a simple closed boundary is a plane graph whose outer face is bounded by a simple cycle and whose other faces are $4$6-cycles. If all internal vertices have degree at least $4$7, then the distances between boundary vertices determine the quadrangulation uniquely up to a boundary-fixing isomorphism. The proof is inductive: one cuts along detectable chords, or finds a “nice configuration” consisting of a strip of quadrangles along the boundary, removes it, computes the new boundary distances, and reconstructs the map recursively. The degree bound is sharp: if degree-$4$8 internal vertices are allowed, one can glue a cube into a face without changing the boundary distances (Haslegrave, 2021).

In an extremal direction, spinal quadrangulations make rigidity a matter of constrained parameters. For a connected graph $4$9, the first Betti number is

$4$00

and the $4$01-fold interlacement $4$02 is formed from two disjoint copies of $4$03 by retaining all original edges and adding cross-edges between each vertex in one copy and the neighbors of its twin in the other. The White–Craft theorem asserts that for any nontrivial connected graph $4$04, there exists a quadrangulation

$4$05

Moreover,

$4$06

These spinal quadrangulations therefore lock genus, order, and chromatic number to the spine (Lawrencenko, 2012).

The extremal content becomes explicit for $4$07: if $4$08 is connected, then there exists a quadrangulation

$4$09

and it is minimal in $4$10 whenever $4$11 and

$4$12

The associated lower bound for any quadrangulation of $4$13 is

$4$14

so these constructions are rigid in the extremal sense of realizing the smallest possible order.

A homological variant of the spinal method expresses the same control through regular neighborhoods in $4$15. For a connected graph $4$16,

$4$17

and $4$18 quadrangulates $4$19. If $4$20 has $4$21 vertices and $4$22 edges, then the resulting quadrangulation has

$4$23

while

$4$24

This suggests a rigid template: once the spine is fixed, the genus and basic combinatorial counts are fixed as well (Lawrencenko, 2013).

6. Local metric rigidity in quadrilateral meshes

In polyhedral geometry, rigidity is studied through Kokotsakis polyhedra with quadrangular base, that is, neighborhoods of a quadrilateral in a quad surface. A simply connected piece of a quadrangular mesh is flexible if and only if each of its Kokotsakis subpolyhedra is. The dihedral angles are related by Euler–Chasles correspondences, leading to a diagram of elliptic curves covering complex projective planes. Generically, the solution set of the corresponding algebraic system is discrete, so the polyhedron is rigid. The complete classification of flexible cases consists of orthodiagonal, isogonal, equimodular, conjugate-modular, linear compounds, linearly conjugate, chimeras, and trivial types (Izmestiev, 2014).

A recent extension allows skew quadrilateral faces and formulates the same local mechanism as a $4$25 quadrilateral mesh whose faces are rigid bodies joined by hinges at common edges. At each corner one obtains a spherical quadrilateral and a Bricard equation

$4$26

together with hinge relations

$4$27

Eliminating the $4$28 yields

$4$29

and flexibility is equivalent to the four-equation matching system having an infinite zero set. In the reducible-quadrilateral regime, the flexible cases are exactly the isogonal family, constant matchings, singular non-constant matchings with two isograms, and deltoidal matchings with four deltoids, reducible or irreducible. The paper emphasizes that generic $4$30 meshes are rigid and that flexibility requires isogram/antiisogram or deltoid/antideltoid spherical vertex quads together with global Möbius closure conditions (Liu, 6 Mar 2026).

A still more local metric statement comes from convex quadrangular pyramids. A convex quadrangular pyramid is strongly rigid: there is no continuous family of pairwise non-congruent quadrangular pyramids with the same edge lengths. A self-intersecting quadrangular pyramid can be strongly flexible. Since a quadrangular pyramid is a local cone over a quadrilateral face, this implies that local metric flexibility in quadrangulated polyhedra cannot come from a single convex quadrilateral star; any such flexibility must arise from a larger global mechanism or from self-intersection (Kochetkov, 2020).

Taken together, these literatures show that rigid quadrangulations form a family of sharply constrained quadrilateral structures rather than a single universally fixed class. In the planar-map setting, rigidity means flat interior geometry with boundary-controlled rays; in topological graph theory, it means facial $4$31-cycles rigid enough to force a surface Lovász complex; in inverse and extremal problems, it means determination from boundary distances or from a spine; and in polyhedral geometry, it means the generic absence of nontrivial motions in quadrilateral mesh neighborhoods. This suggests a unifying theme: quadrangulations become rigid precisely when local quadrilateral data, together with one global compatibility principle, eliminate hidden degrees of freedom.