A further research was made to the 13 forest communities on Hong Kong island by using fuzzy cluster

reciprocal averaging (RA) and R-Principal component analysis (R-PCA). The results show that fuzzy cluster is suitable for the classfication of the lower subtropical forests

especially

the higher classficational units. The RA ordination indicates the relations between 13 forest communities and 60 species distribution in these forest communities and it seems better and more significant than the other methods in this field. And R-PCA illustrated the relationship between the dominant and subordinate species.