3.3 Effects of P. kansuensis on the stoichiometry of its community
The unitary linear recursive analysis results for the stoichiometry between soil nutrients (SN, SP, and SK) and the leaves (LN, LP, and LK) showed that there was no linear or nonlinear relationship between the stoichiometry of the paired factors (SN and LN, Fig. 3A; SP and LP, Fig. 3B; SN:SP and LN:LP, Fig. 3D; and SN:SK and LN:LK, Fig. 3E) except for SK. Indeed, SK had a significant impact on the amount of LK (P< 0.05, Fig. 3C and Fig. 3F). There was a negative correlation between SN and LN and SP and LP. However, SK and LK had a significant and positively correlated relationship. Figure 4 shows that SW controlled the leaf stoichiometry of P. kansuensis rather than soil nutrients. There were no statistically significant relationships between LN, LK, or LN:LK and SW, respectively (P ≥ 0.05, Figs. 4A, C, E). It mainly affected LP (Figs. 4B, D, and F). However, SW greatly affected LP (R2 = 0.88, P = 0.0047, Fig. 4B), LN:LK (R2 = 0.88, P = 0.0053, Fig. 4D) and LN:LP (R2 = 0.77, P = 0.0266, Fig. 4F). The SW decreased LN:LP and this was probably due to the fact that the nitrogen increase in leaves was lower than that for phosphorus in leaves as SW rose and this led to a decrease in the LN:LP.
The second axis explains 28.39% of the variation, which meant that the studied factors explained the stoichiometry well. Soil K had a large impact on LN:LP, whereas SN had a considerable influence on LN:LP and LN:LK (Fig. 5A). Figure 5B shows the variance decomposition of the stoichiometry for plant leaves. Leaf stoichiometry is mainly affected by the combined action of soil and plants. The soil accounted for 46.50%, which meant that the main factor affecting soil was SW (Fig. 5B). Plants accounted for 22.50% and was the other major factor affecting community disturbance and competition by P. kansuensis (Fig. 5B).
Discussion