(1) Floristic data were collected from forty quadrats (1 x 1 m) at each of fifty-six mire sites in Great Britain. These data were used to derive ordination scatter diagrams (using reciprocal averaging) and classification hierarchies (using indicator-species analysis) at both site and community levels of differentiation. (2) Ordination of the sites showed a major line of variation from ombrotrophic mires with acid waters and peats to geotrophic sites with circumneutral waters and peats. In floristic terms, these two divisions were reflected in the types of plant comprising the majority of species recorded: in ombrotrophic mires bryophytes were more numerous, and in geotrophic sites vascular plants predominated. A second line of variation, shown particularly among ombrotrophic sites, was from wet sites with extensive pool systems to those in which pools were absent and surface peats were drier. (3) Different floristic groupings corresponded with different mire types as defined on the basis of hydrological, morphological and chemical criteria: this was shown particularly in the correspondence of floristic classification units and hydromorphological m,ire types. (4) Classification of individual quadrats to produce floristically similar units (noda) again revealed a major line of variation, from those noda found in the most acid peats to those of more basic, mineral-enriched locations. The change from one extreme to the other was, for the most part, continuous, with a number of noda containing common species, the proportions of which varied from one nodum to another. (5) By combining site and nodum data, a clearer picture emerged of the changes taking place between the extremes. In the most acid and the most basic sites there were high proportions of quadrats falling into a few noda, but in the intermediate sites a large number of noda were represented, each by few quadrats. (7) A series was produced which linked floristic classifications on the basis of speciescontent of sites and the proportions of different communities (noda) within sites, and related these to hydromorphological mire types and overall chemical conditions, as indicated by pH. By the use of the classification keys produced by indicator-species analysis new sites may be fitted into this graded series without recourse to re-analysis of the full data-set for all sites. The assigning of a new site to indicate its relationships with other British mires may be by use of species-content of the site as a whole, or on the basis of the proportions of different noda present within that site.