In this paper we study the aggregate behaviour of subjects who participated in two well-known ex-periments (Holt, 1985; Fouraker-Seigel, 1963) on oligopoly markets, and we analyse whether it is possible to explain such behaviour on the basis of Evolutionary Game Theory. For this purpose, we compare the simulated results of a generalised version of the "Replicator Dynamics" model with the evolution of strategies selected by participants during the experiments. The evolutionary model predicts, in the presence of noise, a dynamic equilibrium which does not correspond to the Cournot outcome but is closer to Collusion. Our work shows that the "Replicator Dynamics" describes the individual behaviour better than the models of rationality originally tested in the experiments. Furthermore, our results suggest that collusive outcomes might be considered as a possible consequence of the strategic interaction in oligopoly markets if we introduce the concept of bounded rationality adopted in the evolutionary models.