中文摘要：

群落构建机制是土壤动物群落生态学的重点内容之一.相关研究证明,扩散限制、环境筛选和生物间相互作用可能是土壤动物群落构建的重要调控机制,但少有基于小尺度的系统研究.本文在黑龙江省帽儿山森林生态系统研究站的温带落叶阔叶林内,于2012年8月设置3个5 m×5 m的样地,探讨扩散限制、环境筛选和生物间相互作用在小尺度土壤跳虫群落构建中的作用.基于环境主分量因子（PCs）和Moran特征向量图（MEMs）,用变差分解的方法来区分环境筛选和扩散限制的调控作用,并基于偏Mantel检验进一步分析环境变量和空间变量的调控作用.用零模型模拟的方法来说明土壤跳虫群落物种的共存格局、识别群落水平的显著物种对,同时计算环境生态位重叠度指标和体型大小比例来说明生物间相互作用和环境筛选在跳虫群落构建中的作用.变差分解结果表明,扩散限制对土壤跳虫群落具有较大的显著方差解释量,而环境筛选和空间环境结构的解释量相对较小且不显著;偏Mantel检验进一步表明,扩散限制对样地Ⅰ跳虫群落具有重要的贡献;RDA分析表明,土壤有机质、土壤含水量、凋落物干重等环境变量对土壤跳虫群落具有显著的影响,表明环境变量对跳虫群落的重要作用.群落水平零模型分析表明,土壤跳虫群落表现为非随机的共存格局;群落中发现较少的显著物种对,且显著负相互作用的物种对群落构建可能没有显著的作用;体型大小比例和环境生态位重叠度指标表明,土壤跳虫不是竞争性的群落结构,说明环境筛选具有重要调控作用.本研究表明温带落叶阔叶林小尺度土壤跳虫群落主要受到扩散限制和环境筛选的共同调控,生物间相互作用的驱动作用并不显著.

英文摘要：

The underlying processes affecting community structuring in the soil animal community is an important research topic. Related publications advise that community structuring is influenced by dispersal limitation, environmental filtering and biotic interactions. However, the relative contributions of these mechanisms in the soil animal community on the fine scale are less known. In this study, the relative contributions of dispersal limitation, environmental filtering and biotic interactions influencing the fine-scale regulation of a soil collembolan community from a temperate deciduous forest in the Maoershan Ecosystem Research Station in northeastern China are investigated. In August 2012, three plots were established and 100 samples were captured from each plot in a 5 m × 5 m area. To identify the relative roles of dispersal limitation and environmental filtering, principal component analysis, Moran＇s eigenvector maps, variation partitioning and partial Mantel tests were used. To quantify the influence of biotic interactions, null mode analyses were performed for non-random patterns of species co-occurrence, significant species pairs, environmental niche overlap and body size ratios in the soil collembolan community. This determined whether the observed pattern was the result of interspecific competition or environmental filtering. The results of variation partitioning indicated that the largest and most significant variance was ascribed to dispersal limitation in the three plots. The contributions of environmental filtering and spatially structured environmental variation to the collembolan community were relatively low and non-significant in the three plots. Additionally, the results of the partial Mantel tests inferred an obvious contribution of dispersal limitation to the soil collembolan community in one of the plots. The environmental variables （such as soil organic matter contents, soil water contents and litter dry weight） explained a significant fraction of the variance that was detected in the s