Consider first the single, a - integral. It can be deformed to the integral over the segment [0, k^{-1}] so that we would have,
\frac{1}{2\pi i} \int_{\Gamma_0}a(\tau;n)
\frac{d\tau}{\tau-z} = -\frac{1}{\pi } \int_{0}^{k^{-1}}\tau^n(k^{-1} - \tau)^{-1/2}\Phi(\tau)d\tau,
where
\Phi(\tau) = \sqrt{\tau k^{-1}}\sqrt{k-\tau}\frac{\alpha^{2}(\tau)}{\tau - z}
is holomorphic at \tau = k^{-1}. Let
\Phi(\tau) = \sum_{l=0}^{\infty}d_l(\tau-k^{-1})^l, \quad d_0 = \Phi(k^{-1}) =
\frac{1}{k}\sqrt{k-k^{-1}}\frac{\alpha^{2}(k^{-1})}{k^{-1} - z},
be the Taylor series of \Phi(\tau) at \tau = k^{-1}. Then, according to the standard Watson Lemma type arguments, we arrive at the asymptotic formula,
\frac{1}{2\pi i} \int_{\Gamma_0}a(\tau;n)
\frac{d\tau}{\tau-z} \sim -\frac{1}{\pi} \sum_{l=0}^{\infty}d_l\int_{0}^{k^{-1}}\tau^n(k^{-1}-\tau)^{l-1/2}d\tau,
and this asymptotic is uniform in any compact subset of \{z: |z| > \rho^{-1}\} and, in particular, for z \in \Gamma_1. For the integrals in the right hand side of (Proof), we have,
\begin{split}
\int_{0}^{k^{-1}}\tau^n(k^{-1}-\tau)^{l-1/2}d\tau & = k^{-n-l-1/2}\int_{0}^{1}t^n(1-t)^{l-1/2}dt = k^{-n-l-1/2}B(n+1, l+1/2) \\ & = k^{-n-l-1/2}\frac{\Gamma(n+1)\Gamma(l+1/2)}{\Gamma(n+l+3/2)} = k^{-n-l-1/2}\Gamma(l+1/2)n^{-l-1/2}\left(1+O\left(\frac{1}{n}\right)\right),
\end{split}
as n \to \infty, and hence
\frac{1}{2\pi i} \int_{\Gamma_0}a(\tau;n)
\frac{d\tau}{\tau-z} =
-\frac{1}{\sqrt{\pi}}\alpha^{2}(k^{-1})\frac{\sqrt{k-k^{-1}}}{k}
\frac{1}{k^{-1} -z}k^{-n -1/2}n^{-1/2}\left(1+O\left(\frac{1}{n}\right)\right),
as n \to \infty, uniformly in any compact subset of \{z: |z| > \rho^{-1}\}. This is the estimate (2.10).
Consider now the double integral (2.10). In view of (Proof) we have at once that( We are also taking into account that \int_{\Gamma_1}b(\mu)f(\mu) d\mu = O(k^{-n}) for any bounded f(\mu) analytic in the annulus, \rho -\epsilon
<|\mu| < k+\epsilon.)
\begin{split}
\displaystyle \frac{1}{4\pi^2}\int_{\Gamma_1}b(\mu;n)\left[\int_{\Gamma_0}a(\tau;n)
\frac{d\tau}{\tau-\mu}\right]\frac{d\mu}{\mu-z} & = -\frac{i}{2\pi^{3/2}}\alpha^{2}(k^{-1})\frac{\sqrt{k-k^{-1}}}{k}k^{-n -1/2}n^{-1/2}
\\ & \times \left[\int_{\Gamma_1}b(\mu;n)\frac{d\mu}{(k^{-1} - \mu)(\mu - z)} + O\left(\frac{k^{-n}}{n}\right)\right], \quad n \to \infty.
\end{split}
Applying to the b -integral in the right hand side of the last equation the same arguments as we used for derivation of the estimate (Proof), we obtain that
\int_{\Gamma_1}b(\mu;n)\frac{d\mu}{(k^{-1} - \mu)(\mu - z)} = -i\sqrt{\pi}\alpha^{-2}(k)\frac{1}{(k-z)\sqrt{k-k^{-1}}}
\frac{1}{k^{-1} -k}k^{-n +3/2}n^{-3/2}\left(1+O\left(\frac{1}{n}\right)\right),
as n \to \infty, uniformly in any compact subset of \{z: |z| < \rho\}. This in turns yields the estimate,
\displaystyle \frac{1}{4\pi^2}\int_{\Gamma_1}b(\mu;n)\left[\int_{\Gamma_0}a(\tau;n)
\frac{d\tau}{\tau-\mu}\right]\frac{d\mu}{\mu-z} = -\frac{1}{2\pi}
\frac{\alpha^2(k^{-1})\alpha^{-2}(k)}{k^{-1} - k}\frac{1}{k-z}
k^{-2n}n^{-2}\left( 1+ O\left(\frac{1}{n}\right)\right), \quad n \to \infty,
which, taking into account that
\alpha^{-1} (z)\alpha(z^{-1}) =1,
implies (2.10)