We investigated the flowering phenology, pollinator visitation, and fruit set of 25 animal-pollinated woody species in a warm temperate secondary forest in Japan. Various species flowered sequentially from February to October. The principal pollinators were bumblebees, honey-bees, flies and/or beetles and birds; bumblebees and flies/beetles pollinated most trees. The duration of flowering was shorter for species that bloomed in the middle of the season than it was for species that bloomed earlier or later in the season. The timing of flowering was more synchronous within species that had a shorter flowering duration; this was also detected when phylogenetically independent contrasts were calculated. This could be important for the effective pollination of species with a short flowering duration because such species bloom sequentially over a short period of less than 1 month around May. Fruit set was related not to pollinator type, sex expression, flowering sequence (in order of the date of peak flowering) or flowering duration, but to the relative abundance of the species in the forest. This correlation was detected for fly-and beetle-pollinated species but not for bumblebeepollinated species. Thus, relatively rare plant species with opportunistic pollinators might experience limited fruit set because of insufficient pollinator services. Bagging experiments conducted on eight hermaphrodite species revealed that the fruit set of bagged flowers was nearly zero, lower than that of control flowers. These results indicate the importance of pollinators for successful reproduction and thus for the coexistence of plants in this secondary forest.

The suitability of a color acetate film for estimating photosynthetic photon flux density (PPFD) in a forest understory was examined. The fading ratio of the film (F), the total PPFD (PPFD total ) to which the film was exposed, and the average daily maximum temperature during exposure (T) were obtained from measurements at multiple sampling points throughout an entire year within a natural secondary forest (n ϭ 42). The ranges of the recorded values were as follows: F 35%-99%, PPFD total 1.4-28.3 mol m Ϫ2 , and T 6°-32°C. PPFD total was regressed by F and T with a high r 2 (ϭ0.94; P Ͻ 0.0001): PPFD total ϭ (100 Ϫ F)/(1.085 ϩ 0.051 T). The absolute error (|estimated PPFD total Ϫ measured PPFD total |) averaged 1.3 mol m Ϫ2 with a maximum of 5.7 mol m Ϫ2 , indicating a good fit. These results indicated broad applicability of the film, both spatially and temporally, for estimating forest understory PPFD.

This study quantitatively compared the sapling (height 62-289 cm) architecture and growth of Castanopsis cuspidata and Quercus glauca, both of which are major components in the temperate zone of western Japan, under shaded light conditions in secondary forest. When the sapling architectures were compared at the same support mass (trunk ϩ branch mass), C. cuspidata had a larger crown area but a smaller height gain than did Q. glauca owing to the allocation of more biomass to lateral branches in C. cuspidata. The above-ground relative growth rate (RGR) of C. cuspidata (0.442 g g Ϫ1 year Ϫ1 ) was nearly twice that of Q. glauca (0.256 g g Ϫ1 year Ϫ1 ), primarily as a result of a greater total leaf area per above-ground biomass (LAR) in C. cuspidata (56 cm 2 g Ϫ1 ) as compared to Q. glauca (33 cm 2 g Ϫ1 ).Because it has a disadvantage in height gain, related to its allocation pattern of biomass, C. cuspidata realized the same height growth (RGR H ) as Q. glauca, despite the large biomass production. The great potential for photosynthesis in C. cuspidata, which results from its vigorous lateral spreading, is presumed to give it a long-term advantage over Q. glauca in the shaded forest understory. Q. glauca invests preferentially in trunk biomass, possibly giving it an advantage in sunny sites as opposed to a shaded forest understory.