Species assembly and niche differentiation were studied, and future species composition was predicted by simple Markov models, in an old-growth deciduous forest at the Ogawa Forest Reserve in central Japan. The dominant species in our 6ha study site are Quercus serrata, Fagusjaponica, and F. crenata.An ordination by population parameters revealed four different combination 3 of life forms and regeneration niches. Cluster analysis based on interspecific spatial correlation revealed three groups of species. The species in cluster A, such as F. japonica, occurred at the bottom of the valley, while those in cluster B, such as Q. serrata, occurred along ridges. Species in cluster C, such as F. crenata, did not show any particular habitat preference. Clusters B and C were further divided into three smaller clusters (a-c). Both clusters Ba and Bb included shade intolerant species. Species in cluster Ba had large clump sizes (> 1500 m2), reflecting regeneration following large-scale disturbances. Species in cluster Bb had smaller clump sizes (< 400 m 2) reflecting regeneration following local disturbances. Clusters Caand Cb mainly included shade tolerant species and shade intolerant species, respectively. Markov models predicted that shade intolerant species, particularly those in cluster Ba, would be eliminated. Thus, species coexist by differentiation of both habitats and regeneration niche in this forest community. Some species such as Quercus serrata, however, regenerate following large-scale disturbances or human activity.

To investigate synchronized annual fluctuation of seed production and its advantage for regeneration at the community level, for nine years (1987–1995) we monitored the flowering, seed production, and seedling emergence of the 16 principal tree species in a temperate deciduous forest, Ogawa Forest Reserve, in central Japan. We found that the species with higher synchronized flowering within a population had larger fluctuation of annual seed production at the population level. The coefficient of concordance of flowering and the coefficient of variation of annual seed production were continuously distributed among species, making it difficult to distinguish masting from nonmasting species. The annual seed production patterns of the 16 species were classified, by cluster analysis, into groups that synchronize their fluctuation of annual seed production. This analysis showed a highly synchronized annual seed production, not only among congeneric species, but also among species of different families. Although our results have some insufficiency of statistical significance, they did show that predator satiation, both in a population and a guild, effectively operated for many species to enhance seed survival at the pre‐dispersal stage. They also showed that pollination efficiency was likely to be operating at the population level for half of the wind‐pollinating species. However, generalist predator satiation at the postdispersal seed stage may not operate in a simple, detectable manner in this species‐rich forest community. It is highly probable that there are combined effects of several factors: limited weather triggers for flowering, common flowering physiology among taxonomically related species, and the ecological advantages at the population and guild levels, may cause multiple species to have synchronized fluctuation patterns of seed production.

Abstract. Structural dynamics of a natural tropical seasonal – mixed deciduous – forest were studied over a 4‐yr period at Mae Klong Watershed Research Station, Kanchanaburi Province, western Thailand, with particular reference to the role of forest fires and undergrowth bamboos. All trees > 5 cm DBH in a permanent plot of 200 m × 200 m were censused every two years from 1992 to 1996. The forest was characterized by a low stem density and basal area and relatively high species diversity. Both the bamboo undergrowth and frequent forest fires could be dominant factors that prevent continuous regeneration. Recruitment, mortality, gain (growth of survival tree plus ingrowth) and loss in basal area (by tree death) during the four years were 6.70%/yr, 2.91%/yr, 1.22%/yr and 1.34%/yr, respectively. Mortality was size dependent; middle size trees (30–50 cm) had the lowest mortality, while the smallest (5–10 cm) had the highest mortality. Tree recruitment was observed particularly in the first two years, mostly in the area where die‐back of undergrowth bamboo occurred. The bias of the spatial distribution of recruitment to the area of bamboo die‐back was significant and stronger than that to the forest canopy gaps. Successful regeneration of trees which survive competition with other herbs and trees after dieback of bamboo could occur when repeated forest fires did not occur in subsequent years. It is suggested that both the fire disturbance regime and bamboo life‐cycle greatly influence the structure and dynamics of this seasonal tropical forest.

No abstract

JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms of scholarship. For more information about JSTOR, please contact support@jstor.org.. Ecological Society of America is collaborating with JSTOR to digitize, preserve and extend access to Ecology.Abstract. Aerial photographs were used to analyze long-term canopy dynamics over 15 yr in a large area (25.25 ha) for a mixed deciduous forest in the Ogawa Forest Reserve, central Japan. Digital elevation models of canopy surface were made from aerial photographs taken in four summers at 5-yr intervals from 1976 to 1991. A digital elevation model of ground surface was also made from the aerial photograph taken in winter 1991. The canopy height at each location was calculated from the elevational difference between the canopy and the ground, and the "vegetation profile technique" was applied to quantify canopy dynamics. The size distributions of gaps at each observation time and of gaps that were created between observation times followed a power function model. While the rates of gap formation and closure throughout the whole period were nearly balanced, they showed large temporal variations among the observation periods. Spatial distribution of newly created gaps was significantly biased toward the edges of old gaps. Probability of repeated gap creation was also significantly high. Analysis of changes in the canopy height profile by a transition probability matrix showed that gap area distribution of this forest may be nearly in a steady state.