Abstract
In vegetation survey and analysis methods using the power law, which was developed by the Ecological Laboratory, Faculty of Science, Ibaraki University, Japan, statistical tests cannot be used to compare occurrence frequencies (p) and spatial heterogeneity values (δ), which indicate differences from the random distribution of each species in a community. We introduced a two-way layout (randomized block design) with replications (blocks) to allow statistical tests on the means of these parameters. We also introduced a new parameter (ε) to represent the difference in the mean spatial heterogeneity of each species from that of the surveyed plant community. The design was tested using a survey of a grazed, semi-natural grassland in central Japan. The plant species were divided into four groups based on life form ; erect-type with a large crown, erect-type with a small crown, creeping type, and bunch type. These groups differed in each of the two types of spatial heterogeneity. Differences in the δ- and ε-values among species within each group were also detected and were explained based on a more precise life form analysis using analysis of variance.
