Abstract:
Let $\left\{ Z_{i},i=0,1,2,...\right\} $ be a critical branching process in
random environment and the variance $\sigma ^{2}$ of a step of the
associated random walk is finite and positive. Consider a sequence of
branching processes $\mathbf{Z}^{(n,x)}=\left\{
Z_{i}^{(n,x)},i=0,1,...\right\} $, where $Z_{i}^{(n,x)}=\left\{
Z_{i}|Z_{0}=m_{n}(x)\right\} $ and $\log m_{n}(x)\sim \sigma x\sqrt{n}$ as $
n\rightarrow \infty $ for some $x>0$.
Three limit theorems are proved: on the extinction time of the process $
\mathbf{Z}^{(n,x)}$, on the properties of a normalized process constructed
by $\mathbf{Z}^{(n,x)}$, and on the normalized logarithm of the process.