A number field has a shape. Take the ring of integers, view it as a lattice in Euclidean space, and mod out by rotations, reflections, and scaling. What remains is a point in a shape space — a geometric invariant that captures the lattice's proportions while discarding its size and orientation.
Note (2026-11-13; read against the paper's full text): the paper's theorem is an equidistribution statement in a regularized sense. Ordered by discriminant, the shapes of totally real multiquadratic fields of degree 2^n in which 2 is unramified are equidistributed, with respect to the natural measure, in the subset of the shape space that such fields occupy (parametrized by shape parameters lambda_2 to lambda_l), and the count is stated for compact continuity sets of that measure. The words curve, closed and periodic do not occur in the paper; the essay's picture of the shapes following a closed periodic orbit, a geometric curve, is the essay's.
For totally real multiquadratic fields of degree 2^n, the natural question is how these shapes distribute as you vary the field. Are they scattered uniformly across the shape space? Concentrated near special points? Governed by some hidden geometric structure?
Jakhar and Ray (arXiv:2603.11443) resolve a conjecture of Haidar: the shapes are governed by the restriction of a natural measure to a specific torus orbit in the shape space. Not uniform, not random, not concentrated at isolated points. The distribution lives on a particular geometric curve — a torus — and the shapes of number fields are constrained to orbit along it.
The structural claim: the arithmetic of number fields doesn't explore shape space freely. The algebraic structure (multiquadratic extensions, the condition that 2 is unramified) confines the lattice geometry to a single orbit in the space of all possible shapes. Different fields can have different shapes, but all those shapes lie on the same geometric track. The algebra writes itself into the geometry as a constraint, and the constraint is a torus — a closed, periodic orbit. The number fields don't choose their shapes arbitrarily; they inherit an orbit from their algebraic origin, and no field of this type can produce a shape off that orbit.