Abstract:Ecological stoichiometry has become the focus of research in ecological sciences in recent years and many studies have examined the carbon:nitrogen:phosphorus (C:N:P) stoichiometry of plants, soil, or litter in forest communities. However, few studies have explored the stoichiometric characteristics of herbs at different ages. Elephant grass, Pennisetum purpureum cv. Guimu-1, is a tall perennial C4 grass that can withstand repeated cutting and regenerates rapidly, and has been widely cultivated in northwest Guangxi, China because of the "grain for green" policy. For enhanced understanding of the nutrient cycling characteristics of artificial forage at different time scales in fragile karst ecosystems, we investigated the C:N:P stoichiometric characteristics of plant-soil-microbe interactions in Guimu-1 pastures of different growth ages. Corn and three different restoration stage pastures (1-, 5-, and 7-y-old Guimu-1 pastures, representing initial restoration, vigorous restoration, and degenerating stages, respectively, after returning farmland to grassland) were chosen in northwest Guangxi, China to determine the C, N, and P contents and the C:N:P stoichiometry. The results showed that:(1) the C, N, P contents and C:N:P ratios in the aerial portion of plants varied with different artificial Guimu-1 pasture ages. The N and P contents of the plant aerial portion in different years all followed the same order:5-y-old < 1-y-old < 7-y-old pastures, whereas the C content followed the reverse order. Plant C:N, C:P, and N:P ratios, which varied over a large range from 26.50 to 33.91, 631.70 to 2254.33, and 23.89 to 67.21, respectively, were all in the order of 7-y-old < 1-y-old < 5-y-old pastures. (2) Soil C, N, and P in the topsoil layer (0-10 cm) all showed the same order:corn field < 5-y-old < 1-y-old < 7-y-old pastures. However, there were no significant differences among the soil C:N, C:P, and N:P ratios for the four herb types (P > 0.05), with average values of 9.20, 27.88, 3.38, respectively. (3) Soil microbial C biomass (MBC), microbial N biomass (MBN), and microbial P biomass (MBP) showed significant differences among the different plant types (P < 0.05). The corn field had the highest MBC/soil organic C, MBN/total N, and MBP/total P ratios among the four plant types, while the 5-y-old pasture had the highest ratios among the three different aged pastures. (4) Correlation analysis showed that:(a) there were significant positive linear correlations between MBC and soil C, and MBP and soil P (P < 0.05); (b) there were highly significant negative linear correlations between plant and soil C, plant C and soil N, plant C:N and soil C, and plant C:N and soil N (P < 0.01); and (c) there were highly significant positive linear correlations between plant N and soil C, and plant and soil N (P < 0.01). Thus, the determination of C, N, and P stoichiometric characteristics of plant-soil-microbe interactions in the present study showed that in karst ecosystems, croplands were more limited by N than by P, whereas artificial pastures in rehabilitated land were mainly limited by P, especially during the vigorous restoration stage. Plant C:N:P stoichiometric characteristics in pastures at different ages showed a higher temporal variability than those of soil.