A new mathematical preprint sets out an exact combinatorial way to count plane hypermaps whose boundaries alternate between two colors. The work focuses on the first two alternating cases and turns the counting problem into identities between formal generating functions, rather than an analysis of a sampled dataset.
The paper’s central move is to combine slice decomposition with accessibility and a marked vertex. In practical terms, the authors break a pointed hypermap into pieces whose boundary information can be tracked, then match those pieces with a second family of objects called slices. The resulting correspondence preserves the weights assigned to the maps.
That matters because the formulas are exact within the formal system being studied. There are no participants, measurements or statistical error bars here: the analysis covers the full class of k-alternating plane hypermaps under the stated definitions.
Turning maps into manageable pieces
The objects are weighted according to their vertices and according to the degrees and colors of their inner faces. The boundary is described through alternating intervals, and the marked vertex helps determine which parts of the map are accessible from the boundary. This setup gives the authors a way to organize a complicated family without listing each map one by one.
The key bijection links accessibly pointed k-alternating hypermaps with fixed boundary-interval lengths to slices whose base word records those same lengths. Because the mapping is weight-preserving, the generating function on one side can be translated directly into the generating function on the other.
The paper also uses bridges to handle connectivity. For plane hypermaps, it states that strong connectivity is equivalent to having no bridge. That equivalence lets the decomposition separate maps that contain such a bottleneck from the strongly connected class needed for later formulas.
From the slice correspondence, the accessibly pointed generating function can be written as a sum of one-boundary terms, with denominators built from differences between boundary variables. The result provides a general formula for the accessibly pointed k-alternating family before the paper specializes to particular values of k.
A closed form for the first case
For the 1-alternating case, the authors derive a differential identity linking the generating function A1 to the derivative, with respect to the vertex variable t, of the logarithm of H1. The same quantity is also related to the derivative of B1 divided by 1 minus B1. These are formal identities, so the symbols encode weighted classes of maps rather than measured quantities.
Integrating those relations gives a compact expression for H1: H1(x; y) = 1/(1 − B1(x; y)). In the paper’s notation, H1 and B1 are generating functions for the relevant one-alternating classes, and the relation connects the broader class with its strongly connected component.
When the degrees of both white and black faces are bounded, the work goes further. It identifies A1 with the t-derivative of the logarithm of a ratio built from the spectral curve, written as E(x,y)/((x − X(y))(y − Y(x))). The bounded-degree condition is part of the result; the formula is not presented as an unrestricted statement.
The monochromatic-boundary case supplies related building blocks. The paper records W◦(x) = Y(x) − V◦′(x) and W•(y) = X(y) − V•′(y), linking those boundary generating functions to the functions used in the spectral-curve description.
The second case is assembled from the first
For the 2-alternating family, the generating function H2 is expressed through an antisymmetric cross-product of four H1 terms, divided by the product (x1 − x2)(y1 − y2). The structure means that the two-alternating count is built from one-alternating data while retaining the distinction between the two boundary-variable pairs.
The strongly connected counterpart, B2, has a parallel form: it uses the difference between reciprocal products of H1 terms over the same boundary-variable denominator. Together, the two formulas show how the decomposition and accessibility machinery carries information from the simpler boundary pattern into the next case.
The authors present this as a purely combinatorial route to formulas that had previously been reached through algebraic methods. The contribution is therefore methodological as well as enumerative: it supplies a bijective explanation for the identities instead of relying only on algebraic manipulation.
A first step, not a general solution
The preprint does not solve the full alternating-boundary problem. Its detailed results cover k = 1 and k = 2, while the general case for k ≥ 3 is left as an open direction. Higher-genus extensions are also proposed rather than established.
Asymptotic analysis and distance-distribution questions remain open as well. Those would address how the enumerated families behave at large size and how distances are distributed within the maps, but the supplied results do not report such analyses.
The characterization of strong connectivity used in the paper also has a defined scope: it relies on planarity and on there being only one face whose contour is not a directed cycle. That condition is part of why the bridge-based decomposition cannot automatically be treated as a result for every broader map setting.
The work is identified as arXiv version 1, dated 20 August 2026, and no journal publication is reported in the supplied metadata. It reports support from the ERC-SyG ReNewQuantum project under Horizon 2020 and from the Agence Nationale de la Recherche grant ANR-23-CE48-0018 CartesEtPlus.
Paper data and sources
Original title: Enumeration of plane hypermaps with a mixed boundary I
Authors: Jérémie Bouttier, Bertrand Eynard, Thomas Lejeune
Journal/Repository: arXiv
Status: Preprint, not yet peer-reviewed
First online: 2026-08-20
DOI: Not available
Original paper · Full text