Clustering of data has always been an open problem in the field of data mining. That is so because even though different techniques have been developed over the years, each of them bringing some new improvement to best classification, they all fall short of some weaknesses. In this paper yet another technique is proposed, that is a genetic chromodynamics-based clustering method. Genetic chromodynamics is a metaheuristics for maintaining population diversity and for detecting multiple optima. Its strategy is to form and maintain stable subpopulations that co-evolve and lead, at convergence, each to an optimum. It uses a variable sized solution population, a stepping stone search mechanism in connection with a local interaction principle, and a special operator for merging very similar individuals. Taking into account all these facts, it seems that the underlying principles of genetic chromodynamics should be perfect for the idea of clustering, each optimum representing a cluster, no number of clusters given in advance and thus obtaining both the optimal number  of groups and the optimal classification.