An international project, DIWPA‐IBOY, took place for simultaneously observing biodiversity throughout the Western‐Pacific and Asian regions in 2001–2003, as one of the core projects for International Biodiversity Observation Year, a crosscutting network activity of DIVERSITAS (an international programme of biodiversity science). DIWPA‐IBOY provides extensive data on species diversity obtained by the standardized method. Under this project, 51,742 individuals of Lepidoptera and 11,633 of Coleoptera were collected by light traps from the Tomakomai Experimental Forest of Hokkaido University, one of the core DIWPA‐IBOY sites, in the cool‐temperate region of northern Japan. Based on these data, this study examined the relative abundance distribution (RAD) to evaluate the amount of rare species in the Lepidoptera and Coleoptera communities. The beta diversities between sampling seasons, forest strata, and trap sites were also assessed to evaluate the spatio‐temporal variability of species composition in these communities. In the analysis of the RAD, the best‐fit model was selected from the log‐Normal, Zipf–Mandelbrot, and Zipf models differing in the tail length of the RAD, i.e., the proportion of rare species. To explore the beta diversity between samples, the abundance‐based Jaccard index with an unseen species estimator was calculated, and then a hierarchical clustering analysis was conducted. As a result of RAD analysis, the Coleoptera community was regarded as containing a larger proportion of rare species than the Lepidoptera community. The seasonal compartmentalization of the community, deduced from the beta‐diversity analysis, was finer in Lepidoptera (seven assemblages recognized) than in Coleoptera (three assemblages). The spatial (vertical and horizontal) compartmentalization was negligible in both communities. The coincidence of the larger proportion of rare species and the lower beta diversity between seasons in the Coleoptera community was explained by the longer life spans of beetles compared to moths, based on the assumption that the length of life span acts as a temporal agent for mass effect on the analogy of the migration rate as a spatial agent for mass effect.

1. Forest entomofauna retain high diversity, and examining beta diversity, or species turnover, among assemblages in a forest community is vital to elucidate the source of this diversity.2. Under the DIVERSITAS in Western Pacific and Asia -International Biodiversity Observation Year (DIWPA -IBOY) project for simultaneously documenting biodiversity throughout the Western Pacific and Asian Region, 892 lepidopteran species (51 742 specimens) and 355 coleopteran species (11 633 specimens) were collected in 2001 by light traps in a cool -temperate forest in northern Japan.3. This study evaluated the beta diversity of lepidopteran and coleopteran communities by ecological categories (i.e. trap location, forest strata, sampling days, and months), and assessed the habitat preferences of lepidopteran and coleopteran species.4. anova -like additive apportioning models were used to quantify the beta diversity among the categories. The models simultaneously provide assessments of whether species distributions are biased in favour of particular habitats. 5. Significantly high beta diversity was observed among months for both Lepidoptera and Coleoptera. The category of months corresponded fairly well to the number of specialist species detected in the category, although a remarkably large number of significant specialists in Coleoptera were observed on strata.6. The high beta diversity and number of specialist species among strata in both communities indicate that stratification between canopy and ground, and seasonal variation, played major roles in species composition and the rich entomofauna in the forest. Highly mobile adults were influenced by the vertical spatial scale, as previously suggested for larvae.

Abstract. Rural biodiversity in East Asia is at risk due to the loss of habitat diversity, and good indicators are needed to evaluate diverse habitats in rural landscapes. We examined whether the higher taxa (classes and orders) of soil invertebrates discriminated among several types of secondary forests such as broad-leaved deciduous forests, conifer forests and bamboo forests, primary forests, grasslands and/or wetlands, better than species assemblages of a well-established indicator, ground beetles (Coleoptera, Carabidae and/or Staphylinidae), in three East Asian regions (Japan, South Korea and the Russian Far East). We collected soil invertebrates with pitfall traps and used community composition and an ordination technique to test their performance as indicators. In Japan, the higher taxa of soil invertebrates discriminated finely among a wide range of habitats, and soil moisture seemed to be an important factor underlying habitat arrangement by these taxa along an ordination axis. While species assemblages of ground beetles detected large faunal differences among grasslands, wetlands and a composite group of three forest-type habitats (oak, conifer and bamboo forests), it failed to discriminate among any of the three forest-type habitats. When the analysis included only these types of forests, ground beetles were found to be able to discriminate finely among them, indicating limited performance in relation to the range of habitats 164 S.-I. TANABE ET AL.covered. In the other two countries, the higher taxa of soil invertebrates showed a performance similar to that of species assemblages of ground beetles, possibly because of the narrow range of habitats analyzed. We conclude that the higher taxa of soil invertebrates are an effective tool for assessing the diversity of rural habitats across the East Asian region, where taxonomic knowledge at the species level is still insufficient. Our results may be applied broadly to other regions where agricultural intensification and land abandonment have caused quantitative and qualitative changes in rural landscapes.