In this paper we verify usefulness of a variance marker of collusion in an application to Lysine prices during well known cartel/competition episode. As an econometric method of detection of structural change in price volatility during collusive and competitive phase wavelet analysis was applied. This method, relatively new in the context of cartel detection has some attractive features for such an applications. It is not data demanding and has very good localization power in a time domain. Our work was motivated by theoretical supergame model of collusion based on fixed cartel quota exogenously provided by cartel members’ agreement we developed in first part of our work.