We compared the understory vegetations of a hinoki （Chamaecyparis obtusa） plantation at before thinning, immediately after thinning and one year after thinning to clarify short-term effects of different thinning methods. We established four experimental treatments;ordinary （point） thinning （scattered section of thinned trees）, line thinning （one line cut with three lines retained）, point and line thinning （point thinning was applied to the three retained lines） and control （no thinning）.　The understory vegetation in the thinned stands suffered from damages by disturbances associated with the thinning practice;the damaged rate in the shrub layer in terms of decrease of stem number was equivalent to or more than the thinning intensity （proportion of the number of thinned trees）.　In particular, two sites where the line thinning applied tended to decrease their number of stems in the shrub layer and total coverage in the herb layer severely compared to the thinning intensity.　No clear difference was observed in the decrease in species number both in the shrub and the herb layers.　One year after thinning, the herb layer recovered their coverage and species number to the same level before thinning or more, while the shrub layer didn’t recover their stem number.　In the herb layer, species composition changed from those before thinning.　Whereas about 56% lucidophyllus forest species disappeared after thinning, only 24% recovered and 19% newly colonized one year after thinning, indicating that lucidophyllus forest species were replaced by open habitat species.　These results suggested that intensive thinning （i.e., high thinning rate） which can maintain better light condition rather than frequent and less intensive thinning which gives repeated disturbances in understory might be suitable for conserving lucidophyllus forest species because they would require a longer period for recolonization.　On the other hand, to conserve the open habitat species in the herb layer, frequent thinning as well as intensive thinning may be effective by keeping preferable light environment at the ground surface.

We investigated the occurrence of plant species in the understories of three agroforestry (AF) patches dominatedby pine (Pinus merkusii), teak (Tectona grandis) and eucalypt (Eucalyptus camaldulensis) (hereafter, P, T, and E)in East Java, Indonesia, in order to observe relative plant occurrences in relation to the AF types and their adjacency.We recorded the presence/absence of all vascular plants (height < 1 m) occurring in sample quadrats (1m × 1 m) placed along transects across the border of P-T and P-E patches with different distance from the patchborders. At each quadrat, coverage of litter and vegetation, the sky factor (SF) and soil water content (SWC)were also measured. Altogether 52 species including 29 native species were recorded in the three AFtypes. P, T and E consisted of 32, 20 and 35 species, including 16, 11, and 18 native species and 8, 6 and 13species that were specific to each type, respectively. The difference in the specific species were partly attributedby the relatively high SF in E and thick litter cover of the forest floor in T. These results indicated that thevariability among different AF types contributed to more than half of the total and native species richness in thestudied AF patches, suggesting an advantage of the having landscape consist of mosaics of different AF typesfor the conservation of plant species diversity. On the other hand, the species that were common to differentAF types were found in the patch interior (6 and 11 species for P-T and P-E combinations) rather than in thepatch edge (4 species for each of the P-T and the P-E edge). These results suggested that the low contrast ofenvironmental differences between AFs can provide only limited edge effects in terms of promoting the coexistenceof plants with different characteristics within a patch, compared to typical forest edges such as thosebetween closed forests and open grasslands.

Kuchinoerabujima Island is a volcanic island in Kagoshima Prefecture, Japan.　Although secondary evergreen broad-leaved forests dominated by Machilus thunbergii are distributed in various locations on the island, information on their ecology is very limited.　To examine the characteristics of the species composition and richness of such forests, i.e. M. thunbergii secondary forests （MSF）, we conducted surveys of MSF and Castanopsis sieboldii secondary forests（ CSF） on the island.　We also compared data obtained via surveys of the MSF with published data from M. thunbergii natural forests （MNF） and MSF distributed in neighboring regions of the island.　The species composition of MSF on the island differed from that of CSF on the island as well as that of MNF and MSF in neighboring regions.　The habitat condition of MSF on the island was substantially different from that of MNF in neighboring regions.　We considered that, as a result of succession, the MSF on the island would change in the future to a climax forest dominated by C. sieboldii rather than a climax forest dominated by M. thunbergii.　Species richness （the number of species per 100-m² plot） was slightly higher in MSF on the island than in CSF on the island or MSF in neighboring regions, but much lower in MSF on the island than in MNF in neighboring regions.