Change in management regimes invariably result in an alteration in the vegetation landscape. The present study seeks to elucidate socially and economically induced changes in landscape pattern in terms of spatial distribution and structure of vegetation at Minamikata, Chiyoda-cho, Hiroshima Pref., a typical farm village in southwestern Japan. In 1966,the scrub community and the forests of all height classes of deciduous oak and Japanese red pine were distributed mosaically, because the trees had been periodically felled for charcoal and timber. Additionally, a sparse woodland of pine was maintained for mowing to produce compost. In 1987,the total number of units identified by dominant species and community height (vegetation types) had decreased, and larger vegetational units had appeared. Due to the cessation of disturbance by cutting and mowing following the replacement of fuel and fertilizer from forest sources by fossil and chemical products. Practically all the vegetation types of 1966 had developed into tall tree forests of oak and pine. Throughout this process, the boundaries of several vegetational units became fused into the same vegetation type. In 1987,tall pine forest was divisible into Sasa veitchii var. hirsuta type and the Eurya japonica type by dominance in forest floor vegetation. These two types are respectively distributed only in forests near to, or far from hamlets, clearly showing that the recent human impact is limited to rural forests near settlements. The progressive succession of secondary vegetation has been occurring in all of the remaining area.
Aquatic macrophytes were surveyed at 23 stations from the head stream through to the rivermouth of the Kako River, Hyogo Prefecture, southwestern Japan. A total of 36 species were recorded, including the submerged form of 4 wetland species. Based on the floristic composition and riverbed morphology the watercourse was divided into four zones as follows : I. An uppermost zone, where only aquatic moss occurred. II. An upper zone, characterized by the occurrence of Ranunculus nipponicus var. submersus and the submerged forms of a few wetland species. III. A middle zone, where Potamogeton crispus, Hydrilla verticillata, Egeria densa and P. malaianus (in the lower part only) were dominant species, abundance varying from site to site. IV. A lower zone, lacking macrophyte species except Phargmites australis due to tidal influence. Most of the reaches of the Kako River (St. 3 to 21) belonged to the middle zone.
Relations between variations of species composition and ecological factors were investigated at lucidophyllous forests on the Wakasa Bay coast and its adjacent islands in Japan. The community data were classified following the Zurich-Montpelier School, and species and plots were ordinated by principal component analysis (PCA). One community, and two associations, which were divisible into seven lower units, were recognized as follews : typical group (A-1), Carex matsumurae group (A-2), Quercus myrsinaefolia group (A-3) and Ilex pedunculosa group (A-4) of Polysticho-Perseetum thunbergii ; typical group (B-1), Rumohra simplicior var. major group (B-2) and Rhododendron reticulatum group (B-3) of Ardisio-Castanopsietum sieboldii ; Persea thunbergii-Castanopsis cuspidata var. sieboldii community (C). The phytosociological units (A-1), (A-2) and (B-1) were at one extremity of axis 1 extracted by PCA, whereas (A-3) and (B-3) were at the opposite. For axis 2,(A-4) and (B-3) were at one extremity, whereas (A-3) and (B-2) at the opposite. The phytosociological unit (C) was scattered around the origin. The first axis reflected the gradients from insular to inland habitat, while the second, that from disturbed to the non-distrubed habitat. The behaviour of the phytosociological units in the ordination diagram successfully explained relations between varlations of species composi tion and ecological factors.
A study was made of branch and leaf biomass in an 83-year-old man-made Cryptomeria japonica stand in the Tokyo University Forest in Chiba. In the study area, four plots with different topographical conditions were established, i.e., R-1 on steep ridge, R-2 on flat ridge, M on mid-slope, and B on bottom. Branch and leaf biomass were estimated by dimension analysis. Leaf biomass was estimated at 28 in R-1,26 in R-2,17 in M, and 13 t/ha in B. Plots with high tree density had tended to give large leaf biomass. Branch biomass was estimated at 40 in R-1,14 in R-2,25 in M, and 28 t/ha in B. Except in R-2,branch biomass exceeded leaf biomass by a factor of two or more, especially in B. In old stands, the very long life span of branches seems one reason for the apparently large branch biomass when compared with younger stands.
Correlations among nine indices of species diversity, HURLBERT's S(100), ITOW's b, FISHER's α, three varieties of SIMPSON's d, MCINTOSH's index, SHANNON's H', and PIELOU's J', were studied using measurements taken in 57 forests, from the humid tropical to the humid warm-temperate regions of East Asia, Pacific islands and Amazonian upstream. The indices were categorized into two groups according to their correlations. In first group, consisting of S(100), b, α and 1/d, values are sensitive to changes in species richness. (1-d), (1-d)/(1-d)max, MCINTOSH's index, H' and J' belong to the second group. The indices studied showed intra-but not inter-group correlation. The number of species occurring once in the community is significantly correlated with FISHER's α.