A novel method of a priori sequentially ordered samples has been developed for vegetation transect evaluation. Some new coefficients of associated similarity (homotoneity) are suggested for its application. The HSCA procedures operate on the principle of agglomeration and can be used to identify typical sample groups on a gradient. The method has been applied to data of four short quadrat transects of pond littoral vegetation and adjacent meadows of the Kratochvile pond in South Bohemia. Vegetation zonation has been studied by use of this method in dependence on the complex moisture gradient. Results are confronted with those following from the classical cluster analysis and common classification as carried out by the Zurich-Montpellier school.