SPECTRAL ASYMPTOTICS AND LAMÉ SPECTRUM FOR COUPLED PARTICLES IN PERIODIC POTENTIALS

The paper’s Theorem 2 proves tr F_E = a cos((ω/λ) ln E − ϕ) + o(1) with a ≥ 2 as E → 0+, by replacing p(t;E) with the heteroclinic p0, factoring X(t) = e^{At}C(t), applying an “exponential death” lemma, and using T(E) = (1/λ) ln(1/E) + K + o(1). There is, however, a sign/notation slip in the definition of ω in (13): the text writes ω^2 = 2κ + λ^2 and even equates 2κ + V''(π) with 2κ + λ^2, although V''(π) = −λ^2; the decomposition used in the proof (A + B(t)) implicitly requires ω^2 = 2κ + V''(π) = 2κ − λ^2. With this correction, the proof is coherent and yields the stated asymptotics and amplitude bound (see the statement of Theorem 2 and definitions (12)–(15) and the A+B(t) split in (22)–(25) . The model’s solution reaches the same asymptotic law via a two-bump SU(1,1) scattering factorization and explicitly identifies the correct local frequency as α = √(2κ + V''(π)), thereby flagging the sign issue. Hence, both are correct, with the paper needing minor notational fixes and the model offering a different proof sketch consistent with the corrected ω. The amplitude bound a ≥ 2 is obtained in the paper from det N = 1 and in the model from SU(1,1) structure; the separatrix-time estimate T(E) = (1/λ) ln(1/E) + K + o(1) is the same in both .