1: \begin{abstract}

2:     This paper concerns $\mu$-limit sets of cellular automata: sets of

3:     configurations made of words whose probability to appear does not

4:     vanish with time, starting from an initial $\mu$-random

5:     configuration. More precisely, we investigate the computational

6:     complexity of these sets and of related decision problems. Main

7:     results: first, $\mu$-limit sets can have a $\Sigma_3^0$-hard

8:     language, second, they can contain only $\alpha$-complex

9:     configurations, third, any non-trivial property concerning them is

10:     at least $\Pi_3^0$-hard. We prove complexity upper bounds, study

11:     restrictions of these questions to particular classes of CA, and

12:     different types of (non-)convergence of the measure of a word

13:     during the evolution.

14:   \end{abstract}

15: