Abstract: A method to study population dynamics based on raw data and nonlinear model by stepwise regression and stepwise prediction is developed in this paper．The data on Rattus nitidus were collected in our study site in Qionglai county，Sichuan province by markrecapture，trap and dissection technique from 1989 to 1991．The relationship between populatign density and age structure，sex ratios，reproductive eharaotera，temprature and precipitation are analysed and quantified to predict population density of Rattus nitidus and determine to what extent the present day ecological factora are influencing the future populaton density．Stepwise regression results reveal that the relationship between population density of Rattus nitidus and age structure is conspicuous，but the relationship between population density and sex ratios is not Reproductive characteristics showed significant effects on population density with reproductive time delay．

Key words: Nonlinear model, Population dynamics, Stepwise regression, Rattus nitidus

摘要： 本文提出了一种从原始数据中获取关于种群动态机制的基本信息的、以非线性模型为基础的逐步回归逐步预测的种群动态分析法。并以作者1989-1991年间于四川邛崃定点进行标志重捕以及夹捕解剖所得的关于农田大足鼠(Rattus nitlidus)种群生态学特征的数据为例，对大足鼠的种群密度与年龄组成、各年龄组性比、繁殖特征以及气象等因子的关系进行了初步的分析，定量地确定了当前的各生态因子对未来的大足鼠种群密度的作用，并对数量预测问题进行了初步的讨论。结果发现，大足鼠的种群密度与其年龄结构显著相关，而与性比无关。繁殖特征对密度也有显著的作用，并有明显的生殖时滞。

关键词: 种群动态, 非线性模型, 逐步回归, 大足鼠