Abstract: [Objective] This paper explores understory seedlings of different provenances of Machilus pauhoi and their response to geographical environment changes with the intention of providing theoretical and practical basis for artificial reforestation of natural forest and selection of suitable cultivation environment selection.[Method]Leaf samples were collected from understory seedlings of 9 provenances of M. pauhoi and variations of leaf C, N, P stoichiometry of different provenances were analyzed by one-way ANOVA and the correlations were calculated using Pearson method. Representative environmental factors were selected by combining both entropy-weight method and redundancy analysis method. Based on stepwise regression, fitting models of leaf stoichiometry and environmental factors were established. Using akaike information criterion, the best fitting models were chosen to analyze the influence of environmental factors on leaf stoichiometry of M. pauhoi understory seedlings. [Result] Firstly, leaf P concentrations, C:P and N:P of M. pauhoi understory seedlings of different provenances presented extremely significant differences (P < 0.01), while C and N concentrations as well as C:N presented little. For the variation coefficients of leaf stoichiometry, P concentrations, C:P, N:P, N concentrations, C:N and C concentrations were in descending order. Secondly, N concentrations were extremely significantly positively correlated to C and P concentrations (P < 0.01), but extremely significantly negatively correlated to C:N and C:P (P < 0.01). P concentrations demonstrated significant positive correlations with C concentrations (P < 0.05), but extremely significant negative with C:N, C:P and N:P (P < 0.01). And C:P displayed extremely significant positive correlations with C:N and N:P (P < 0.01). Thirdly, leaf C, N, P stoichiometry of M. pauhoi understory seedlings were mutually affected by longitude and mean annual temperature. Leaf C and N concentrations were significantly and extremely significantly positively correlated with longitude (P < 0.05 and P < 0.01, respectively), while C:N showed a significantly negative correlation (P < 0.05). Leaf P concentrations were extremely significantly positively correlated with longitude and mean annual temperature (P < 0.01), while C:P and N:P extremely significantly and significantly negatively correlated (P < 0.01 and P < 0.05, respectively). Longitude rather than mean annual temperature had greater effects on the three. [Conclusion] Differences were detected in leaf C, N and P stoichiometry of M. pauhoi understory seedlings of different provenances, which were mainly resulted from longitude and mean annual temperature. As a consequence, it is crucial to consider the adaptability of different provenances and environmental heterogeneity during the process of artificial reforestation of natural M. pauhoi forest and selection of suitable cultivation environment.

Key words: C, N, and P stoichiometry, geographical environment, understory seedlings, provenance, Machilus pauhoi