Fig. 1.10 (a) Replicate two in this sample of macrobenthos from a mangrove forest forms an outlier
in the Non-metric MultiDimensional Scaling (nMDS) plot because the data for 3 of the 37 taxa in the set
of data were omitted from this sample (and were, therefore, erroneously given a value of 0). When this
was corrected, the replicates grouped into a single cluster. (b) If the points from a number of replicates
in a single sample cluster into two or more distinct groups, it often suggests that the replicates are
coming from two distinct strata and the sampling design should be stratified.