Abstract
A method of deriving vegetation mapping units from quantitative data is described based on results from an ecological survey of Tongariro National Park. A particular aim was to develop a repeatable procedure. The method of classifying the samples uses a polythetic agglomerative technique m which the sortmg strategy has as a priority the combining of similar entities that are closest together in the field. This allows class boundaries to be made more nearly coincidental with map boundaries. A naming system for vegetation mapping units is further refined from an earlier published system. The names convey both structural and compositional information about the vegetation in such a way that diagnostic field criteria for most mapping units are summansed by the unit names. Although emphasising cover estimates, both the classificatory method and naming system are independent of the sampling technique used to estimate cover. The method is suitable for a wide range of terrestrial habitats.