A GIS (geographic information systems) database was constructed from aerial photographs, a vegetation map, and topographic map data of the Ngeremeduu Bay Drainage Area (NBDA), Palau, to examine relationships between upland land cover dynamics, environmental variables, and past land use. In 1992, 82.9 percent of the NBDA was forest, 16.6 percent was grassland, and 0.5 percent consisted of village areas. Between 1947 and 1992, there was a 11.2 percent reduction of grassland area primarily due to a 10.9 percent increase in forest cover. These land cover changes led to larger, more continuous stretches of forest and numerous, highly fragmented grassland patches. Significant relationships (P < 0.001) were found between the spatial distribution of forest and grassland cover and slope, elevation, soil pH, and percent soil organic matter. These patterns, however, may have resulted from past farm site selection rather than from ecological relationships. Our results indicate that areas of forest expansion were significantly (P < 0.001) associated with the location of abandoned agricultural communities. In addition, over 92 percent of areas of forest expansion occurred within 100 m of established forest. These results suggest that the proximity of established forest facilitate forest recovery following human disturbance.

Resistance and resilience of Najas marina to disturbance by flash floods were studied in a permanent fluvial pool of a Brazilian semiarid intermittent stream. A total of 21 macrophyte samples was collected in the high-rainfall season during two annual cycles (1996–1997). Decreases in macrophyte biomass were positively correlated with flood magnitude (Pearson, P = 0.047), varying from 25 to 53 percent when discharges were lower than or equal to 0.5 m³/sec and between 70 and 100 percent when discharges were higher than 1.0 m³/sec. Macrophyte resilience was greater after floods of low magnitude. After floods of 0.5 m³/sec, three weeks were necessary to re-establish 88 percent of biomass lost, and after a flood of 1.4 m³/sec, six months were needed to initiate N. marina regrowth. This population of N. marina in Avelós stream has higher stability in response to small disturbances, but as expected, its resistance and resilience decreases with larger disturbances. In general, the high resistance and resilience of N. marina in response to small disturbances have been observed in intermittent tropical streams. The absence of large floods during the study period and the low variability of water temperature in this tropical region may have influenced these results.

We addressed the interacting effects of a natural large-scale fire and a subsequent major hurricane on relative positions of epiphytes in a subtropical forest. In Everglades National Park, subtropical hammocks (hardwood tree “islands”; burned and unburned) during the Ingraham Fire (1989) were surveyed for trees and epiphytic bromeliads (Tillandsia spp.) one year before, as well as one and five years after, Hurricane Andrew (1992). We measured trees (species, diameter, and status [alive/dead]) and epiphytes (species, height, host tree characteristics, substrate life status, and density). The fire decreased the height of epiphytes during the hurricane because branches and bark of trees killed by the fire were unstable epiphyte substrates in the high winds. Proportions of epiphytes on Quercus virginiana were equally increased after the hurricane in both unburned and burned hammocks; the large size and bark characteristics resulted in greater proportional survival of epiphytes on this species. During the five years following the hurricane, changes in the distributions of epiphytes generally were toward pre-hurricane distributions, but recovery was faster in unburned than burned hammocks. We conclude that disturbances that kill trees are likely to amplify the vertical reduction of epiphytes during a subsequent hurricane and that effects of a single disturbance on plant populations can be influenced by the disturbance history of the system, including different types of disturbances.

We examined the interaction between a palm and two bruchid beetles along with several mammal species to explore how poachers and habitat fragmentation may indirectly alter the spatial pattern of seed dispersal, seed predation, and seedling recruitment in central Panama. The large, stony endocarps of Attalea butyraceae decay slowly and bear distinctive scars when opened by rodents or beetles. We determined the final distance between endocarps and reproductive trees (which we call an ecologically effective dispersal distance), the predation status of each endocarp, and the distance between seedlings and reproductive trees. The 68 focal trees were divided among 14 sites and four levels of anthropogenic disturbance. Levels of disturbance included full protection from poachers, light and heavy pressure from poachers, and small island habitat fragments. Ecologically effective seed dispersal distances were greatest for protected sites, intermediate for lightly poached sites, and shortest for heavily poached sites and habitat fragments. Seed predation by rodents increased with distance to the nearest reproductive Attalea and was greatest for fully protected sites, intermediate for lightly poached sites, and least for heavily poached sites and habitat fragments. Seed predation by beetles reversed the patterns described for seed predation by rodents. Total seed predation by beetles and rodents combined was independent of distance, greatest for fully protected sites, and lower for poached sites and habitat fragments. Seedling densities were always greatest close to reproductive trees; however, the increase in seedling densities close to reproductive trees was minimal for fully protected sites, clearly evident for poached sites, and pronounced for habitat fragments. Increased seedling recruitment near conspecific trees may in time reduce tree diversity where humans disrupt mammal communities.

We studied the phenology and reproductive biology of the stained glass palm Geonoma epetiolata in a premontane rain forest of Costa Rica. This understory species is endemic to Costa Rica and Panama. We followed 102 plants for two years and found that G. epetiolata is monoecious and protandrous. Flowering duration at the inflorescence level is three to six months for the male phase and two to three weeks for the female phase; this is one the most extended flowering periods reported for palms. Flowering in this palm is continuous throughout the year, both at the individual and population level. This long and continuous flowering pattern may represent an adaptation to increase the chances of reproduction in a environment of high and unpredictable rainfall. The species is self-compatible, but self-pollination is possible only when there is overlap of male and female inflorescences (38% of the cases). Only 28 percent of the stigmas of 1020 flowers received pollen grains and 20 percent developed pollen tubes in the styles. Of all the flowers produced by an inflorescence, 14 percent initiated fruit development but only 3 percent develop into mature fruits. Seventy-six percent of the initiated fruits were aborted.

Most tropical plants produce fleshy fruits that are dispersed primarily by vertebrate frugivores. Behavioral disparities among vertebrate seed dispersers could influence patterns of seed distribution and thus forest structure. This study investigated the relative importance of arboreal seed dispersers and seed predators on the initial stage of forest organization—seed deposition. We asked the following questions: (1) To what degree do arboreal seed dispersers influence the species richness and abundance of the seed rain? and (2) Based on the plant species and strata of the forest for which they provide dispersal services, do arboreal seed dispersers represent similar or distinct functional groups? To answer these questions, seed rain was sampled for 12 months in the Dja Reserve, Cameroon. Seed traps representing five percent of the crown area were erected below the canopies of 90 trees belonging to nine focal tree species: 3 dispersed by monkeys, 3 dispersed by large frugivorous birds, and 3 wind-dispersed species. Seeds disseminated by arboreal seed dispersers accounted for ca 12 percent of the seeds and 68 percent of the seed species identified in seed traps. Monkeys dispersed more than twice the number of seed species than large frugivorous birds, but birds dispersed more individual seeds. We identified two distinct functional dispersal groups, one composed of large frugivorous birds and one composed of monkeys, drop dispersers, and seed predators. These groups dispersed plants found in different canopy strata and exhibited low overlap in the seed species they disseminated. We conclude it is unlikely that seed dispersal services provided by monkeys could be compensated for by frugivorous birds in the event of their extirpation from Afrotropical forests.

Variation in post-dispersal seed predation between mature forest and fallow swidden agriculture sites (young forest) was examined for 26 tree and liana species in Peruvian lowland tropical forest. Exclosures were also used to determine the relative importance of vertebrate and invertebrate seed predators. Although species differed significantly in levels of predation, predation of seeds exposed to vertebrates was high (>75%) for all species, while seed predation by invertebrates was lower and more variable. Seed and seedling predation by insects was significantly higher in mature forest, while seed predation by rodents was higher in fallow areas. Other sources of mortality (primarily desiccation) were higher for seeds in fallow areas. The percentage of seeds moved was significantly different among species but not between forests.

In Amazonian rain forest trees of Vochysia vismiaefolia (Vochysiaceae), ants were found to induce twig structures that resembled classical ant domatia. This phenomenon is novel for ant-plants, which commonly develop domatia without the activity of ants. Eight species of ants were recorded inside the domatia of six individual trees, but domatia were most numerous and characteristic when induced and inhabited by an undescribed species of Pseudomyrmex on two trees. To investigate the mechanism of domatium growth, we drilled holes into young twigs: the expansion of the twig diameter surrounding the holes was significantly accelerated, comparable to domatia formation. Domatia induction is discussed as a putative step in the evolution of ant-plants.

This paper describes the ant assemblages sampled from rain forest canopies ranging from southern Victoria through to Cape York Peninsula, Australia, and also in Brunei. Specifically, it examines the influence of decreasing latitude and variations in elevation on the character, richness, and abundance of the arboreal rain forest ant fauna, and also the relative contribution of ants to the total arthropod community. The sites that were examined included: cool temperate Nothofagus cunninghamii forest from a range of locations in Victoria; cool temperate N. moorei forest at both Werrikimbe and Styx River, New South Wales; notophyll vine forest in Lamington National Park, southeast Queensland; high elevation notophyll vine forest in Eungella National Park, central Queensland; complex notophyll vine forest at Robson Creek, Atherton Tablelands, north Queensland; complex mesophyll vine forest at Cape Tribulation, north Queensland; and mixed dipterocarp forest in Brunei. Although these sites represent a gradient increasingly tropical in character, botanically speaking, Eungella is less tropical than Lamington because of its high elevation. All samples were obtained by fogging the canopy with a rapid-knockdown pyrethrin pesticide. In all cases, circular funnels were suspended beneath the foliage of individual trees or small plots of mixed canopy. Arthropods were collected four hours after fogging. Following ordinal sorting, ants were identified and counted to morphospecies level. The resulting catch were then standardized across sites as numbers caught per 0.5 m² sampling funnel. Generic and species richness were higher at the lowland tropical Cape Tribulation sites than at the sites to the south and was comparable with values in the Brunei site. Species richness was negatively correlated with latitude and elevation. Within the Australian rain forest, the lowland/highland break appears to be the strongest predictor of ant relative abundance, with a weaker latitudinal relationship superimposed.

I observed 22 species of birds visiting flowering Erythrina fusca trees at Matamatá, Amazonas, Colombia. The large orange flowers of E. fusca are adapted for pollination by birds and are protected from illegitimate visits by a petal that covers the nectaries and anthers until displaced by a foraging bird. Experiments with flowers bagged to exclude potential pollinators demonstrated that the flowers do not open without assistance. At Matamatá, parrots are the most frequent visitors to flowering E. fusca, and two species, Dusky-headed Parakeet (Aratinga weddellii) and Cobalt-winged Parakeet (Brotogeris cyanoptera), appear to be the main pollinators. This is only the fourth report of Neotropical parrots acting as pollinating agents. At least five other parrot species fed on the nectar or flowers of E. fusca but destroyed the flowers in the process. Orange-backed Troupials (Icterus jamacaii) were the only other species observed opening E. fusca flowers nondestructively and are likely to be pollinators. Hummingbirds were common visitors to E. fusca flowers and some species were found to carry E. fusca pollen; however, hummingbirds were unable to open the flowers themselves and relied on other visitors to open the flowers for them. The number of hummingbird visits to a flowering E. fusca tree was positively correlated with the number of visits by parrots and icterids, but not with the number of mature flowers, indicating that legitimate visitors facilitate access by hummingbirds.

Although bats are nocturnal, many species emerge from roosts to forage during twilight, despite a presumed high risk of predation at this time. Here, we describe twilight foraging by a maternity colony of Schneider's leafnosed bat (Hipposideros speoris) in the dry zone of Sri Lanka and determine the dietary benefits of such behavior. Bats usually began foraging during dusk, sometimes before sunset, and also foraged during twilight in the morning. Mean use of available twilight by four radio-tagged bats was 75 percent. Twilight foraging made up, on average, 47 percent of the total foraging time of these bats (range = 25–96%), although twilight consisted of only 12 percent of the available time between sunset and sunrise the next morning. Eight species of potential predators (7 birds and 1 mammal) were observed within a 1 km radius of the colony, of which 5 species are predicted to regularly capture bats. Bats took a wide diversity of prey (11 insect orders, including at least 27 families, and spiders) that ranged in wing length from 2.0 to 54.0 mm. Major orders in the diet were Coleoptera, Lepidoptera, and Diptera. Prey of secondary importance included Hemiptera, Hymenoptera, Isoptera, and Neuroptera. Bats captured large numbers of insects that were only available or had marked peaks in abundance during twilight. These groups included small, swarming insects (especially flies) that have peaks in flight activity at dusk and dawn, large diurnal species (especially dragonflies) that have crepuscular activity, and winged termites that emerge in swarms at dusk. Access to these insects was a clear benefit of twilight foraging.

This study reports extraordinarily high density estimates for the wild pig (Sus scrofa) from an aseasonal tropical forest site within the species' native range. At Pasoh Forest Reserve, a 2500 ha area of lowland dipterocarp rain forest in Peninsular Malaysia, line transects were used to estimate pig density from May to October in 1996 and 1998. In 1996, 44 sightings of S. scrofa consisting of 166 individuals were recorded along 81 km of transects. In 1998, 39 sightings documented 129 individuals along 79.9 km of transects. Estimated population density was 47.0 pigs/km² in 1996 and 27.0 pigs/km² in 1998. Sus scrofa biomass in this forest was estimated at 1837 kg/km² and 1346 kg/km² in 1996 and 1998, respectively. Differences between years were attributed to changes in the density of young pigs, coincident with a mast-seeding year of dipterocarp trees in 1996. Pig densities at Pasoh Forest Reserve were much higher than at other forest locations within the species' native range in Europe and Asia. Because Pasoh Forest Reserve is a forest fragment, two factors likely account for the extremely high pig densities: (1) local extinction of natural predators (mainly tigers and leopards) and (2) an abundant year-round food supply of African oil palm fruits from extensive plantations bordering the reserve.

Between January and December 1996, the food habits of a relict population of jaguars were studied in 220 km² Linhares Forest Preserve, which comprises much of the remaining old-growth Atlantic Forest of Espírito Santo, Brazil. Fecal analysis indicated opportunistic feeding on 24 prey species (N = 101 scats). Mammals represented 87 percent of the total items, followed by reptiles (9.8%) and birds (2.8%). Considering prey weight, 23.4 percent of the items weighed 1–3 kg, 40.5 percent were 3–10 kg, and 27.7 percent weighed more than 10 kg. Analysis of relative prey frequency and biomass indicated that the diet was concentrated in two prey types: long-nosed armadillo and white-lipped peccary. Literature data suggest that forest jaguars rely on the same mammal prey over their entire geographic range.