[Effects of different vegetation restoration of degraded red soil on earthworm population dynamics]

This study was conducted at the long-term experimental plots in Ecological Experimental Station of Red Soil in Yujiang County (28 degrees 15'30''N, 116 degrees 55'30''E ), Jiangxi Province, subtropical China. Earthworm population was investigated seasonally from May 1999 to February 2000, under different vegetations including four artificial woodlands [deciduous broadleaf woodland (Quercus chenii, Qc), evergreen broadleaf woodland (Schima superba, Ss), coniferous woodland (Pinus massonina, Pm) and mixed woodland (Schima superba-Pinus massonina, Sm)], two grasslands [gently-disturbed grassland (G1), undisturbed grassland (G2)] and control wasteland (CK). The results indicated that the population structure was very simple. Only Drawinda gisti characterized by pioneer was found. The seasonal averages of density and biomass were in the order of G2 > G1 > Qc > Ss > Pm > Sm > CK, and those of G2, G1 and Qc were significantly higher than those of the latters (P < 0.05). Seasonal fluctuations were obvious with dry-hot summer depressing the earthworm population sharply, leading to the aestivation of earthworm. Based on the variation coefficients of density and biomass, Qc had the highest ecosystem stability, followed by Sm and Ss, and G1, G2, and Pm had the lowest stability. The overall differentiation of earthworm population could be drawn through canonical discriminant analysis. There were significant correlations between earthworms and some soil properties (P < 0.01). Overall, the differentiation of earthworm population was driven by the quantity and quality of soil organic matter returned by the vegetations. Additionally, based on earthworm population, the importance of selecting appropriate vegetation types during the restoration of degraded red soil was emphasized.