Answers are grounded in this paper and cite their sources.
The gaps of the index-two subgroup U₀ of rational points on the unit circle are governed by the missing coset: every gap is controlled by an odd–odd vector, yielding an explicit five-mark law, hard edge 2/π², determinant law 4/(k(k+1)(k+2)), and a non-Markov colour process whose runs are arithmetic progressions. Let U be the group of rational points of the unit circle, μ₄ = {±1, ±i}, and 𝒜 = {Arg z : z ∈ U \ μ₄}. The Pythagorean angle lattice is PAL = {cot(α/D) : α ∈ 𝒜, D ≥ 1}; its angles are the arguments Arg(W²) of primitive Gaussian integers W, and its classical layer is carried by the sources, the primitive W = a + bi with a, b > 0 and odd norm, through the angles θ_W = 2 arg W of the primitive Pythagorean triangles. The points ei θ_W are the points in the upper half-plane of the index-two subgroup U₀ of U consisting of the points with even second numerator, and the complementary coset i U₀ is the second cusp of the theta group. Korolev and Ustinov determined the gap density of the whole group U ordered by denominator. We give a complete statistical theory of the gaps of U₀, that is, of the angles of the primitive Pythagorean triangles with hypotenuse at most R², and we show that it is governed throughout by the missing coset. Every gap is controlled by a unique primitive vector with both coordinates odd. A five-mark law gives the joint limit law of the length, determinant, endpoint ratio, colour word and fluctuation of a gap as an explicit pushforward of Lebesgue measure; it rests on an equidistribution theorem for modular inverses on parity classes, proved with the Weil–Estermann bound. The normalized gaps have an explicit law with hard edge 2/π² and tail 16/(π⁴ T²), and the determinant has the law 4/(k(k+1)(k+2)). Along the sequence of gaps, two consecutive determinants exceeding one form one of the pairs (2,2), (2,3), (3,2), the colours form a stationary process that is not Markov of any finite order, in which L ≥ 3 consecutive sources share a colour with probability 1/(8L−10), and monochromatic runs are arithmetic progressions with length spectrum R/(4|z|) and phases governed by square roots of −1. The second moments of the determinant and of the gap length are 4 log R + C_det and 32 log R + C_G with both constants in closed form, the higher moments are values of Epstein zeta functions, and at finite height the largest gaps and the largest determinant are determined exactly.
No takes yet. Share an insight, caveat, or question.
Chetansing K. Rajput (2026) studied this question.
Synapse has enriched 8 closely related papers on similar clinical questions. Consider them for comparative context: