Given a sequencef1,f2,… of rational functions, there are two obvious ways of forming a sequence of composites, namely the sequences (fn∘ ∘f2∘f1) and (f1∘f2∘fn). If thefjrandom perturbbations of some fixedfboth nonnormal sets associated with these sequences are perturbations of the Julia set off. The first of the above orders has been considered by Fornaess and Sibonv and it has been shown that the (random) Julia sets are of area zero with probability one. We consider the Julia set corresponding to the second order, it turns out that they depend onfalone and are of positive area almost surely.