Ordination is the procedure of reducing the dimensionality (that is the number of variables) of multivariate data by deriving the small number of new variables which contain much of the information in the original data. The decreased data set is many times more useful for investigating possible structure in the observations.