The Fermat cubic surface is a classical geometric object defined by a simple symmetric equation in four variables. Like all smooth cubic surfaces, it contains exactly 27 straight lines, and these lines interact with each other in a rich combinatorial and geometric way. A "sixer" is a specially chosen set of six mutually non-intersecting lines on such a surface that together encode a way to "unfold" the cubic surface back into a flat plane by reversing a standard geometric construction. The Fermat cubic has exactly 72 such sixers, and this paper studies their structure in detail.
The authors first sort these 72 sixers into groups based on the symmetries of the Fermat cubic surface itself. They find that the symmetry group of the surface organizes the sixers into two families: one containing 18 sixers and another containing 54. To understand what geometrically distinguishes these two families, they look at the corresponding flat plane pictures that each sixer produces. In each case, the six lines on the surface correspond to six points in the plane, and those six points have their own symmetry group. The two families turn out to correspond to six-point configurations in the plane whose symmetry groups have sizes 36 and 12, respectively. Larger symmetry in the plane corresponds to the smaller family of 18 sixers, and smaller symmetry corresponds to the larger family of 54.
The authors then push further by analyzing these configurations using tools from algebra and number theory, working over a specific number field built from a cube root of unity. They define an invariant, called a determinant square-class character, that measures a subtle algebraic property of the symmetry groups involved. This invariant cleanly separates the two families: it produces a one-dimensional output for the family of 18 and a two-dimensional output for the family of 54. When the authors reduce the problem modulo the prime 13, the symmetry groups specialize to well-known finite groups, and the same distinction persists. This robustness across different choices of normalization gives the arithmetic separation a particularly solid footing.