This study provides a comprehensive picture of experimental Kreps-Scheinkman markets with capacity choice in the first stage and subsequent price competition at the second. We conduct seven different treatments of such markets, varying the number of firms, the demand rationing scheme, the subject matching, and subjects' knowledge about the market mechanism. We find that only the number of firms entails a persistent effect on capacity choices. Price choices are affected by both the number of firms and the rationing scheme. Subjects in the high-knowledge condition behave in the same manner from the first periods as subjects with low knowledge do in later periods after having gained experience. In all treatments conduct is generally more competitive than the Cournot outcome, irrespective of whether the Cournot outcome is the Nash equilibrium or not. Nevertheless, the Cournot model entails some predictive power. Exact Cournot choices are more likely to occur for both capacities and prices under efficient demand rationing, where the Cournot outcome is the equilibrium, than under proportional rationing.