We resolve the three open problems stated in Paper 10 (Bird 2026, DOI: 10. 5281/zenodo. 19228226), §6. 2. (1) Class B Borel singularity. The Class B iterated-logarithm series Σ Pₙ (ln ln K) / (ln K) ⁿ is not a Gevrey-1 series of any fixed order in v = 1/ln K: its Borel-transform radius of convergence satisfies ρ (K) ~ 1/ln ln K → 0 as K → ∞, so no Borel–Laplace re-summation exists in the standard sense. The obstruction is the non-Gevrey growth aₙ ~ (ln ln K) ⁿ, consistent with Écalle-type doubly-iterated resurgence (Costin 2008). (2) Full boundary-layer series. All higher-order coefficients dₖ of the Class A boundary-layer series Σₖ dₖ W^ (k+3/2) / (k+3/2) are computed analytically via a direct binomial convolution and confirmed numerically to relative error 7. 2×10⁻¹⁵ at K = 10⁹ for k ≤ 10, all three benchmark α-values. (3) Stokes constant amplitude. Via the convergent amplitude ratio |S₁A| = lim|₊|→∞ |ANP (K) | / (CA |K|^−γA) along the Stokes ray arg K = π/ (1+α), we establish |S₁A| to five significant figures for α ∈ 1/4, 1/2, 3/4. The sign of S₁A is identified as an open problem requiring Picard–Lefschetz orientation analysis.
Michael Bird (Thu,) studied this question.