Fig. 2

Clustering of ECMf species based on the similarities in their potential enzymatic activity profiles. Clustering of ECMf species based on their enzymatic activities was done by comparing similarities in their functional traits (eight enzymatic activities). Data coming from both sites (Humont, Champenoux), all sampling dates (3 for Humont, 15 for Champenoux), all treatments (four for Humont: Beech limed, Beech Untreated, Spruce limed, Spruce Untreated, two for Champenoux: Horizon A₁ and Horizon A₂) and all replicates (Humont: n = 4) were pooled altogether. Then a minimal spanning tree (“Minimal spanning tree”) was computed from a clustering of the species–species Euclidean distance matrix. The tree was then split into ten clusters according to the k-means method following algorithm of Hartigan and Wong (“K-means clustering”). The center-reduced average enzymatic activities (centroids) of each cluster are presented in the table below the tree. The whiter the cell is, the higher the average potential activity in the cluster (lowest activities are in black, highest ones in white). The size of each cluster and the number of species present in each cluster are reported on the 2 bottom rows. The size of a cluster is the number of enzymatic profiles, an enzymatic profile being measurement of eight enzyme activities for 1 species at a given date (Humont: 3; Champenoux: 15) in a given replicate (Humont: 4) of a given treatement (Humont: 4; Champenoux: 2)