Abstract
Small-scale spatial heterogeneity of individual plant species in sown grasslands was investigated using a new experimental and statistical method proposed by SHIYOMI et al. (2000). The frequency of occurrence of each species was used to evaluate the spatial heterogeneity of species distribution. A discrete probability distribution, beta-binomial series (BB) was introduced as a model to describe the heterogeneity. One of the parameters of the BB, p, was used as a heterog.eneity index. The objectives of this study were: (1) to determine whether or not the use of the BB is suitable for describing the spatial distribution of individual species grown in sown grasslands, and (2) to determine how different the spatial heterogeneity is among plant species. It was empirically demonstrated that the BB is a good model to describe the frequency of occurrence of each species population composing sown grasslands. The spatial heterogeneity of the clonal species was significantly higher than that of the non-clonal species. The spatial heterogeneity of the creeping species was significantly higher than that of the erect species. The use of the experimental and statistical method for detecting patchiness of grassland vegetation is suitable for grassland diagnosis.
