Some models, based on the latitudinal variation in sun angle distribution, predict that trees at high latitudes have narrowly conical crowns and constitute simple‐layered forests, whereas trees at low latitudes have shallowly dome‐shaped and form more structurally complex multilayered forests. There is a hypothesis that structurally complex habitats can harbor potentially more species than simple ones. In this study, we examined latitudinal correlations between tree shape, forest structure and diversity in drosophilid communities, comparing boreal and cool‐temperate forests. We selected secondary birch forests with a common canopy tree species, white birch (Betula platyphylla Sukatchev), as study sites. The crown shape of white birch tended to be spherical in the cool‐temperate forest, but narrowly conical in the boreal forest. The foliage structure differed between the two forests. The cool‐temperate forest was characterized by a clearly two‐layered structure, whereas foliage in the boreal forest was less clearly stratified, being distributed somewhat continuously from the ground to the canopy at lower densities. The structural complexity expressed by foliage height diversity was greater in the cool‐temperate forest than in the boreal forest. Various measures of drosophilid diversity were higher in the cool‐temperate forest than in the boreal forest, probably resulting from the impoverishment of the canopy subcommunity in the boreal forest. Thus, a physical environmental factor (i.e. the angle of solar inclination) could be a potentially important factor in structuring latitudinal patterns of sylvan animal communities through changes in plant structure at the individual and community levels.

Abstract. 1. The spatial distribution pattern of forest-dwelling drosophilid flies was compared among species and among four forests with markedly different vertical foliage structures, including secondary and primary forests, with special reference to vertical stratification.2. All 20 drosophilid species analysed showed vertically stratified distribution patterns, which were detected statistically in at least one forest site during the 2-year study period. As a whole, the vertical distribution patterns were stable, with or without stratification, over the 2 years.3. The ratio of species with stratified distribution to the total number of species changed in a similar trend with the vertical foliage complexity of forests, both showing a large gap between secondary and primary forests, even for species common to all four forests.4. Many species, most of which were tree-sap feeders, showed highly predictable patterns of vertical stratification across forests and years. This was associated strongly with regular preference for microhabitat, mostly for the canopy, suggesting that the unvarying nature of canopy environments in any forest is very important in producing, maintaining, and promoting the vertical stratification in organism abundance, and contributes to the ubiquity of the pattern.5. Some species with a wider range of food resources than other species, most of which showed no clear preference for the canopy, were characterised by large between-forest differences in vertical distribution patterns and tended to show nonstratified patterns in secondary forests, contributing to the lower ratio of species with stratified distribution in the two secondary forests than in the primary forests.6. The role of forest structure in organising the environment is discussed; it is suggested that vertical complexity of foliage structure affects the prevalence of vertical stratification in animal communities indirectly through the vertical heterogeneity of environmental conditions.

Satoyama, the traditional agricultural landscape in Japan, has drawn much attention from the viewpoint of biodiversity conservation. It is composed of diverse vegetation types, including secondary forests, paddy fields and cultivated fields in a narrow area (ca. 1 km 2 ). To clarify the characteristics of butterfly assemblage and the relative contribution of each vegetation type to butterfly diversity and endemism in satoyama, we conducted a line-transect survey along a 1.1-km route with five sections (two forest interior, one interior-edge, one edgeopenland and one openland). Life history features (voltinism, host plant range and host plant type) and endemism of butterflies were discussed in relation to their abundance and vegetation association. Fifty-one species and 856 individuals were recorded. Species richness was highest in the two sections with edges, 32 and 36 species, respectively. Analysis of vegetation association of each butterfly species showed that 16 species were forest interior species, 24 forest edge species and 11 openland species, indicating that vegetation diversity enhanced butterfly diversity. However, the point is that forest interior and edge species contained many specialists sensitive to human impact (univoltine and/or oligophagous species) and many species within temperate East Asia including all endemic species. In contrast, most openland species were generalists (multivoltine and/or polyphagous species) feeding on herbs/grasses with the widest geographic range. Since secondary forests kept Ó Springer Science+Business Media B.V. 2006 more butterfly diversity and specialists than openlands, and had all endemic species, they must be maintained without over fragmentation in order to avoid loss of specialists and endemics in satoyama.

No abstract

Kanazawa Castle Park (KCP) is an isolated urban green area in Kanazawa, Japan. A mature forest had developed there since 1949, but it was partially deforested from 1996 to 1999. Moreover, most of the grasslands in KCP were removed from 2000 to 2002. In this study, we report the damage to the butterfly assemblage by these anthropogenic disturbances and subsequent recovery after the disturbances, using previous literature on the butterfly assemblage in KCP from 1990 to 1994 and our monitoring data from 1999 to 2005. In addition, to understand urbanization effects, we also compared the butterfly assemblage of KCP with that of Tawara, a satoyama area 6 km away from KCP. Before the urbanization around KCP, KCP might have had a butterfly assemblage similar to Tawara. The previous articles indicate that 33 species lived in KCP before the partial deforestation. The number of generalist species (multivoltine and polyphagous species) was similar in KCP and Tawara (22 vs. 19), but that of specialist species (uni‐bivoltine and oligophagous species) was lower in KCP than in Tawara (5 vs. 15). Six species, composed of a bamboo grass feeder, mantle feeders, herb/grass feeders, and a multi‐feeder, disappeared from KCP after the deforestation and have not yet recolonized. The grassland removal also affected herb/grass feeders negatively, but they have been recovering quickly following the partial grassland recovery. Some species that disappeared after the deforestation were sporadically found from 2000 to 2005 probably due to immigration from the surrounding rural areas (the nearest one is 2 km away). For recolonization of the lost species, recovery of the interior and edge of the forest, which are the habitats for these species, is necessary.