An iterative approach from the standpoint of the additive hypothesis to the dendrogram problem posed by molecular data sets