Abstract: Excentrodendron hsienmu is a species endemic to tropical limestone areas. It is one of the constructive species in north tropical limestone seasonal rainforest. The mixed forest dominated by E.hsienmu is very important to the karst ecosystem. So far, little is know about the main environmental factors affecting the plant diversity in this community. To explore the effects of variations in soils and topography on the species diversity in this community, woody plants with diameter at breast height (DBH) ≥1.0 cm, shrubs, herbs, soils, and topographic factors were investigated in 15 plots in the karst mountains of Pinguo County, located in Guangxi, south China. α diversity indexes were calculated, and 7 soil properties were measured. Principal component analysis (PCA) and canonical correlation analysis (CCA) were employed in this study. It was observed that the community composition was simple, and the α diversity indexes were low. All the species richness, Shannon-Wiener and Pielou index in tree layer were the greatest, followed by shrub and herb layer. Results of PCA showed the accumulative variance contribution of the first three principal components was 78.8%. The eigenvalue of first principal component was 5.29, and it encompassed 44.0% of the total variation, which mainly reflected the variation in soil nutrients. Results of CCA showed α diversity of tree layer was sensitive to the variation of soil nutrients. Shannon-Wiener diversity index of tree layer was negatively correlated with total phosphorus (TP) and positively with total nitrogen. Shannon-Wiener diversity index of shrub layer was sensitive to the topography gradient. It was positively related to both slope position and elevation. Meanwhile, there was close coupling relationship between soils and topography. TP was positively correlated with position, and pH negatively with per. of covered rock.

Key words: soil, topography, principal component analysis, canonical correlation analysis, Excentrodendron hsienmu