A key assumption when inferring floristic composition and related properties of a mapped unit of vegetation is that the unit consistently represents a suite of species, and that the mapping units have some type of equivalence.  An important class of vegetation for such maps is the alliance, a physiognomically uniform group of plant associations sharing one or more diagnostic species, which are usually found in the canopy.  This study explores methods for measuring the relationship between the dominant and subdominant species in defined vegetation alliances of the Inland Northwest region in the USA.  A set of 8,625 field plot records were assembled and classified to 49 alliances.  A comparative data set for the classified records was derived by removing the dominant species.  A nonmetric multidimensional scaling ordination and a multi-response permutation procedure were used to examine the distinction and separation among and between classified field plots of each alliance and all other field plots, as well as between the plot data set containing all species and the one containing only the subdominant species.  The role of the dominant species and the subdominant species in determining the floristic identity of each alliance was measured by applying a multi-response permutation procedure to (a) an alliance’s entire plot data and (b) the derived plot data where the dominant species had been removed.  The ratio of the amount of floristic distinction found in these two data sets represents the extent to which the dominants or the subdominants determine the alliance’s floristic identity, and it provides a measure of how general or specific an alliance is in floristic composition.