Vertical structure in plant communities is well recognized, but questions remain about the response of different stratal layers in a plant community to resource use. This study provides a quantitative approach to compare the relationships of nine edaphic factors and species distributions in shrub and herb layers, respectively, in an arid valley of the upper Minjiang River, Sichuan Province, China. Classification (TWINSPAN) determined three groups of shrubs and four groups of herbs. Ordination (both DCA and DCCA) found a correlation among species and edaphic variables. Borcard's method [Borcard, D., Legendre, P., Drapeau, P., 1992. Partialling out the spatial component of ecological variation. Ecology 73, 1045-1055.] was employed to partition the variation of species abundance data into independent components: pure spatial, pure environmental, spatial component of environmental influence, and undetermined. Based on those four components, a quantitative comparison of ecological variation between shrubs and herbs further eluidated their different response to the environment. Although soil moisture (SM) content, pH, soil organic matter (OC) and available potassium (AK) were the most important factors to both, the explained ecological variation of herbs was lower than that of shrubs. Understanding differences between stratal layers in resource use can shed light on plant-environment relationships. r

A pot experiment was performed to study the light-response curve of photosynthesis (PN-PAR curve) of Mangifera indica and the applicability of light-response models under different soil water conditions. The experimental data were fitted and analyzed using the rectangular hyperbola model, the nonrectangular hyperbola model, the exponential model, the modified rectangular hyperbola model, and the kinetic model. The results showed that the optimal range of relative soil water content (RSWC) for the normal photosynthesis of M. indica was 45.1-77.3%. The modified rectangular hyperbola model could well fit the PN-PAR curves and photosynthetic parameters under wide range of soil water conditions (RSWC 23.3-77.3%). The rectangular hyperbola model, the nonrectangular hyperbola model, the exponential model, and the kinetic model could only be used to fit the PN-PAR curves of M. indica under mild and moderate drought stress (RSWC 45.1-77.3%).