Ordination is a multivariate technique that arranges vegetation samples in relation to each other based on compositional similarity and relative species-abundances. Ordination procedures summarize multidimensional data in a reduced coordinate system, extracting those axes that explain the most variation in the data. DCR-DNH ecologists use non-metric multidimensional scaling (NMDS) , an ordination technique based on indirect gradient analysis that maximizes, to the extent possible, the rank-order (i.e ., non-parametric) correlation between inter-sample dissimilarity and inter-sample distance in ordination space. The results of ordination analyses are depicted by diagrams, in which each point represents a plot and the distance between points roughly indicates the degree of compositional similarity. Statistically significant correlations between measured environmental variables and sample coordinates on each axis may be plotted as vectors and overlain on the diagram. The direction of a vector indicates the direction of maximum correlation through ordination space, while vector line lengths are determined by the strength of the correlation. This diagram shows a two-dimensional ordination of the same dataset used to illustrate cluster analysis. Symbols indicate the four groups identified in the dendrogram and significant environmental gradients are plotted.