Results

Interspecific variation in leaf NSC

In this study, there were significant species variations in leaf soluble sugar, starch, and NSC concentrations (p < 0.001 for all, Table 2). In the CK treatment, the leaf soluble sugar concentration of S. octophylla was generally higher than that of the other three species, with concentrations of 124.54, 174.28, and 179.58 mg g-1, in 2012, 2015, and 2017, respectively. In addition, the starch concentration of S. octophylla in the CK treatment was approximately twice as high as the other three species (Fig. 1, e-h). The NSC concentrations of S. octophylla in the CK plot were generally higher than those of the other three species (Fig. 1 i-l), with concentrations of 179.81, 219.00, and 199.44 mg g-1, in 2012, 2015, and 2017, respectively.

Leaf NSC concentration in response to N and P addition

Neither N addition nor N×P interactions had significant effects on the total leaf NSC concentrations (p =0.746 and p =0.320, respectively), while P addition significantly affected leaf NSC concentrations (p <0.001). P addition significantly influenced the NSC concentrations of U. macrophylla , S. octophylla, and P. rubra in 2012, those of S. octophyllain 2015, and those of P. rubra in 2017. Overall, +P reduced the NSC concentrations by 18.14% (except for U. macrophylla ), 15.13%, and 24.68% (except for U. macrophylla and S. octophylla ), in 2012, 2015, and 2017, respectively (Fig. 2).
In this tropical forest, neither N addition nor N×P significantly affected leaf soluble sugar concentrations (p =0.719 andp =0.866, respectively, Table 2). However, the leaf soluble sugar concentrations were significantly changed by P addition (p =0.016). Overall, +P decreased the leaf soluble sugar concentrations by an average of 12.78% across the three sampled years (except for S. bullockii in 2012 and U. macrophylla in 2017, Fig. 2). Neither N nor P addition had significant effects on leaf soluble sugar concentrations in any of the species in 2012 (Table S1). However, there was a significant reduction in the soluble sugar concentrations of S. octophylla in 2015 and those of P. rubra in 2017 under P addition (Table S1).
Nitrogen addition did not change the leaf starch concentrations of the sample species (Table 2). However, both P addition and N×P interactions significantly affected starch concentrations (p =0.001 andp =0.030, respectively). Overall, +P reduced leaf starch concentrations by 38.10% (except for S. octophylla in 2017, Fig. 2), +NP also decreased the starch concentrations by 28.28% (except forU. macrophylla and P. rubra in 2015, Fig. 2). Additionally, P addition also had significant effects on the leaf starch concentrations of P. rubra in 2012, and those of S. bullockii in 2015 (Table S1).
Although the ratios of soluble sugar to starch showed significant interspecific variation (p <0.001), this ratio was generally not affected by N, P, nor N×P interaction at the intra-species level in this study (Table S1).

Relationships between leaf NSC and other traits

Leaf soluble sugar and total NSC concentrations, and the ratios of soluble sugar to starch were significantly negatively related to the LMA (Table 3). Leaf starch concentrations was significantly positively correlated with photosynthesis (p <0.05). In addition, both the soluble sugar and NSC concentrations were significantly related to the leaf N concentrations (p <0.05). Although there was no clear relationship between NSC (soluble sugar and starch) and leaf P concentrations, the structural P concentration had significantly positive correlations, while residual P concentration had significantly negative correlations, with soluble sugar and NSC concentrations, and the ratios of soluble sugar to starch (p <0.05, Table 3).