عنوان مقاله [English]
This research aims at ecological classification and identifying diagnostic species of box tree (Buxus hyrcana Pojark) sites in Farim forests, Sari, Iran. The data were collected from 60- 400m2 sample releves in systematic- selective design (200 × 400 meter) by considering the indicator stand concept. Floristic data were analyzed using modified TWINSPAN and consequently, six ecological groups were classified. The results of diagnostic species analysis based on Phi fidelity index revealed that of the 77 species recorded, only 47 species had positive fidelity to presented groups (P<0.05) and therefore they were determined as diagnostic species of box tree plant communities in the study area. Detrended Correspondence Analysis (DCA) was used to represent the distribution pattern of ecological groups based on two data matrix of plant species canopy cover, full data and only diagnostic species. Comparisons of scatter plots of ecological groups along two first axes in DCA, as well as, result of cumulative traced variances in two series data (37.1% in only diagnostic species data and 26.2% in total species data) in this analysis clearly indicated that ordination of six groups based on only diagnostic species data presents more appropriate distribution pattern compared to full species data. It is concluded that application of diagnostic species may lead to improve the results of ecological group ordination in determining the effective environmental variables precisely on considered groups.