Random logistic maps. II: The critical case (Q1426849)

[For part I (by \textit{K. B. Athreya} and \textit{J. Dai}) see ibid. 13, No. 2, 595--608 (2000; Zbl 0969.60069).] The authors study iterations of i.i.d. random logistic maps of the interval which generate a Markov chain on the interval. If the expectation of logarithm of the parameter is zero (called the critical case), it was shown before that the Markov chain converges in probability to zero. The authors show that in spite of this the Markov chain does not converge to zero with probability one. Moreover, with probability one the Markov chain exceeds certain fixed positive valued infinitely many times.