Agglomerative clustering works by merging the two most similar clusters at each iteration.
Step through the algorithm one iteration at a time, or select the auto box to automatically animate it. The similarity metric used for this algorithm is the euclidean distance between the centroids of clusters.