A clustering method operates on some measure of resemblance (similarity/dissimilarity/distance) among objects. It uses those resemblances to produce a result, be it a dendrogram or some other result. Cophenetic correlation is a measure of how well the clustering result matches the original resemblances. So, as an example, similarities among samples are clustered using a method like UPGMA to produce a dendrogram. The distances among samples are calculated through the dendrogram (actually to a common node, but the idea holds) to give cophenetic distances. The between-sample original resemblances are correlated with the cophenetic distances to give cophenetic correlation. If the value is high (near 1) the clustering result is an excellent representation of the original distances, if it is