“…Just differentiate (3.3) with respect to z, which is permissible by uniform convergence on compact subsets of C\{E j } 2g j=0 . We remark that if we require our perturbation to satisfy the usual short-range assumption as in [1], [9,10,11] (i.e., the first moment is integrable, see (5.1)), then we even have e ∓ixk(z) (ψ ± (z, x) −ψ q,± (z, x)) → 0.…”