Chin J Plant Ecol ›› 2007, Vol. 31 ›› Issue (1): 50-55.DOI: 10.17521/cjpe.2007.0007

• Research Articles • Previous Articles     Next Articles


SHEN You-Xin1,2(), LIU Wen-Yao1, CUI Jian-Wu1   

  1. 1Kunming Section of Xishuangbanna Tropical Botanical Garden, Chinese Academy of Sciences, Kunming 650223, China, and
    2Graduate University of Chinese Academy of Sciences, Beijing 100049, China
  • Received:2005-11-17 Accepted:2006-04-08 Online:2007-11-17 Published:2007-01-30


Aims Seeds stored in soil are important for plant regeneration and ecological restoration. Seeds are dispersed from parent plants stochastically, and this results in a non-uniform distribution of seeds across the litter and soil. Soil samples must be collected to overcome this variation and to obtain a reasonable richness and density of seeds. However, extracting seeds from soil is time-consuming and studying the germination of seeds from soil samples requires much space. Therefore, it is important to determine a reasonable sample size for soil seed bank studies. However, the relationship between soil sampling area and seeds has rarely been explored in China.

Methods We collected 100 soil samples of 10 cm×10 cm in a Karst forest in central Yunnan. Species and seed density were determined with the seedling emergence method after washing soil samples through 4 and 0.21 mm mesh sieves to eliminate coarse and fine materials, respectively. Soil samples were treated as quadrats and arrayed in a regular but non-contiguous grid to establish a species-area relationship.

Important findings A total of 2 536 seedlings from 69 species (7 trees, 6 shrubs and 56 herbs) emerged from the samples. There were 3-67 seeds from 1-14 species in a soil sample, with an average of 23.4 seeds and 7.1 species. The percentage of samples in which a species germinated ranged from 1% to 90%. The number of total, tree, shrub and herb species had significant quadratic and logarithmic relationships with area. The bases of total species and herb species were large and the slopes of their species-area curves were steep. In contrast, the bases for tree species and shrub species were small and showed little increase. The increase of total species leveled out at 0.15-0.2 m2, i.e., 15-20 samples (10 cm×10 cm). This study indicates that 15-20 samples with total surface area of 0.15-0.2 m2 can serve as a reference for soil seed bank studies in areas similar to this study. However, the seedling emergence method, even with the total sample area in this study, needs to be improved or replaced to determine the sampling requirement for tree and shrub species.

Key words: Karst, forest, soil seed bank, species-area relation, sample area