Frugivorous dacine fruit flies were studied in a lowland tropical rain forest in Papua New Guinea to determine their host specificity, abundance, and the number of species attacking various plant species. Plant species hosted 0–3 fruit fly species at median (1–3 quartile) densities of 1 (0–17) fruit flies per 100 fruits. Fruit flies were mostly specialized to a single plant family (83% species) and within each family to a single genus (88% species), while most of the species (66%) were able to feed on >1 congeneric plant species. Only 30 from the 53 studied plant species were colonized by fruit flies. The plant–fruit fly food web, including these 30 plant species and the total of 29 fruit fly species feeding on them, was divided into 14 compartments, each including 1–8 plant species hosting mutually disjunct assemblages of fruit flies. This structure minimizes indirect interactions among plant species via shared herbivores. The local species pool was estimated at 152±32 (±SE) fruit fly species. Forty per cent of all taxonomically described species known from Papua New Guinea were reared or trapped in our study area. Such a high proportion indicates low beta-diversity of fruit flies. Steiner traps were highly efficient in sampling the lure-responsive fruit fly species as they re-collected 84% of all species trapped in the same area 5 y before. Fruit fly monitoring by these traps is a cheap, simple and efficient method for the study of spatial and temporal changes in rain-forest communities.

Relative size at onset of maturity (RSOM) is defined as size at first reproduction divided by asymptotic maximal size. RSOM is remarkably constant among species within many higher clades of animals, but varies widely among tree species from the Pasoh Forest Reserve, Malaysia according to the work of S. C. Thomas. RSOM was examined for 16 mid-storey and canopy tree species from a second tropical forest at Barro Colorado Island (BCI), Panama. Interspecific variation in RSOM was equally large for BCI and Pasoh and was unrelated to gap dependence or life form for BCI species. The shape of the relationship between size and the proportion of individuals that were reproductive differed between forests, with an abrupt increase over a narrow range of sizes at Pasoh and a more gradual increase over a wider range of sizes at BCI. Both overtopping trees and heavy liana infestation reduced the probability that BCI trees were reproductive. This presumably reflects reduced availability of carbon for reproduction. We speculate that greater liana loads and a greater abundance of large, shade-casting trees at BCI may increase variation among individuals and contribute to the relatively wide range of sizes characterized by a mixture of sterile and fertile individuals observed for most BCI species.

Variation among forests in environmental and biotic conditions may strongly influence seed fate with important consequences for the abundance and distribution of plant species. Here we examine the post-dispersal seed removal rates of six pioneer species (Cecropia peltata, Miconia argentea, Luehea seemannii, Trema micrantha, Apeiba aspera and Jacaranda copaia) from the soil surface at five sites in Panama varying in elevation (0–1100 m) and seasonality (0–4 mo dry season). We compared removal rates of washed seeds placed in vertebrate exclosures, invertebrate exclosures, and unprotected controls in January and June. Overall, removal rates of unprotected seeds were similar among sites. Almost all seed removal could be attributed to litter ants in two subfamilies (Myrmicinae and Ponerinae). Little or no removal was recorded from invertebrate exclosures while vertebrate exclosures had no effect on removal either in lowland and montane forests. Seed removal rates were high for four animal-dispersed species (mean 45% removed over 2 d), whereas two wind-dispersed species were largely untouched (mean 2% removed). These results indicate that seed dispersal characteristics, rather than site characteristics, may be the strongest determinant of the post-dispersal seed fate of pioneers.

To investigate whether seedlings of ten dipterocarp species differed significantly in terms of growth and mortality or whether species were not significantly different and could be considered ecologically similar, seedlings were grown, two per pot, in two experiments: (1) where the two seedlings were of equal height (30 cm); and (2) where one seedling was 10 cm shorter than the other. Seedlings were grown in a shade house with 15% above-canopy light in a 50:50 forest soil–sand mixture and were watered frequently; pots were placed so that seedling density was 130 seedlings m−2 of ground. In the first experiment there were 45 pairwise combinations of species when seedlings were 30 cm tall (AB, AC, AD,…. BC, BD…IJ; where A, B, C…J signify different species); each combination was replicated 10 times so there were 450 pots with 900 seedlings. In the second experiment there were 100 pairwise combinations of species and size e.g. Aa (30 cm A with 20 cm a), Ab (30 cm A with 20 cm b), each combination was replicated 10 times hence there were 1000 pots with 2000 seedlings. After 22 mo 79% of the initial 2900 seedlings survived; on average they had grown 42 cm (i.e. to 72 cm tall from their initial 30 cm). The most frequent outcome of competition-trials between different sized individuals (784 of 1000 trials) was that the initially taller seedling of each pair ‘won’ (it was the taller or surviving seedling). When 900 of these trials (setting aside, Aa, Bb, Cc etc.) were analysed as 45 comparisons between species with different sized individuals (Ab and aB are one interspecific comparison for these purposes), initial height determined the outcome in 23 cases (even in some competitions between light hardwood species and heavy hardwood species); in 6 cases a species (mostly light hardwoods) behaved as a ‘dominant’ – they usually won even if they were smaller initially. We found few significant differences between species in: initial seedling heights; leaf nitrogen concentrations; and specific leaf areas when they were grown in similar conditions, and these attributes were not correlated with growth rates. The similarity of seedlings of different species meant that often a height difference of just 10 cm was enough to determine the outcome of a pairwise competition-trial in high seedling densities and light equivalent to that in forest gaps.

The induction of carbon-based secondary metabolites in leaves following damage has been proposed to be a result of a shift in the carbon:nutrient balance, when growth is limited by nutrients in relation to carbon. Here we test this hypothesis using seedlings of a tropical tree, Shorea leprosula (Dipterocarpaceae). In the short term, we analysed the phenolic content of leaves 7 d after damage on seedlings grown under differing light and nutrient treatments. In the long term, we examined the effect of nutrients, over 12 mo, on leaf phenolic concentration and seedling growth. In both the long and short term, levels of phenolics increased in damaged leaves under low nutrient treatments. No changes in leaf phenolics were detected under high nutrient regimes, or in the short term under low light. In addition, it was found that defoliation of seedlings in high-nutrient environments led to greater rates of leaf production than in undamaged seedlings, suggesting compensation.

Although fire is occurring at greater frequencies and spatial scales in the moist tropics, few studies have examined the ecological impacts of fire in tropical montane cloud forest (TMCF). This study, conducted in the Chimalapas region of Oaxaca, Mexico, documents changes in live tree biomass, live fine-root biomass, and fallen and standing dead wood 4 y following deep ground fires occurring in TMCF during the 1997–98 El Niño Southern Oscillation event. Forests growing on two different substrates (metamorphic and sedimentary) and having three different statures (mean canopy heights: 20–30 m, 15–20 m and 4–6 m) were assessed within six paired plots established on adjacent burned and unburned forest sites. Total live tree biomass was 82% and 88% lower for burned TMCF growing on metamorphic and sedimentary substrates, respectively, compared with unburned TMCF. Nearly 100% of the living biomass was killed in elfin TMCF located on exposed sedimentary limestone at the highest elevations. Live fine-root biomass in the upper organic soil horizon of burned TMCF sites was 49% lower on metamorphic substrates and 77% lower on sedimentary substrates compared with unburned sites. The amount of total dead wood was 3- to 14-fold greater in burned forests compared with unburned forests. These results suggest that first-time fires in relatively undisturbed TMCF can cause dramatic changes in live above- and below-ground biomass at levels greatly exceeding values reported for most lowland tropical rain forests. These patterns may be attributed to the slower decomposition rates and thick organic soils typical of TMCF, combined with the relatively fast drainage associated with steep topography and, in some locations, sedimentary limestone-derived substrates.

Photosynthesis was limited by low-intensity photosynthetically active radiation (PAR) and leaf wetness in a lower montane cloud forest (LMCF) of Cauca, Colombia. Mean PAR intensity remained below the saturation level for leaf-scale net photosynthesis (Pn) throughout the solar day during the wet season and for most of the solar day during the dry season. PAR represented a smaller fraction of total solar radiation (K↓) in LMCF than in lowland rain forest (LRF). In LMCF trees and shrubs, mean PAR-saturated Pn ranged from 4.3–10.6 μmol C m−2 s−1 at 1450 m, and from 3.5–10.2 μmol C m−2 s−1 at 2150 m. Pn was reduced by abaxial wetness in leaves of some trees and shrubs, and eliminated in others. This study indicates that persistent cloudiness and interception of cloud water by leaves limit LMCF productivity.

Studies of mixed-species bird flocks have found that ‘nuclear’ species, those important to flock coherence, are either intraspecifically gregarious or are ‘sentinel’ species highly sensitive to predators. Both types of species are present in flocks of a Sri Lankan rain forest: orange-billed babblers (Turdoides rufescens Blyth) are highly gregarious, whereas greater racket-tailed drongos (Dicrurus paradiseus Linnaeus) are less so, but more sensitive and reliable alarm-callers. We hypothesized that flock participants would be attracted to the playback of both species more than to the clearly non-nuclear yellow-fronted barbet (Megalaima flavifrons Cuvier). Further, we hypothesized that insectivores would prefer babbler vocalizations, as babblers could facilitate their foraging in several ways. We found that the response of insectivores was three times greater during babbler or drongo playback, and eight times greater during playback of these two species together, than during barbet playback or silence. Insectivores did not show, however, any difference in their response to babbler as compared to drongo playback; omnivores and frugivores responded relatively equally to all treatments. Our results show that birds with high propensity to flock, such as insectivores, use the vocalizations of nuclear species to locate flocks and that a sentinel species may be as attractive as a highly gregarious species.

Guiera senegalensis tolerates repeated cutting and pruning to which it is increasingly subjected at the onset of each cropping period, and one to three times during the annual cropping period (June–September). It responds by profusely branching at the shrub base. The topographic and geomorphic influences, and the effect of clearing on the relationships between shrub density, mean individual above-ground biomass and leaf water status were analysed for seven fallow sites. They were sampled 75 km east of Niamey in Niger. Surprisingly, shrub density and mean individual above-ground biomass were highest in the sites that had been recently in fallow and intensively cultivated prior to crop abandonment. Stand above-ground biomass was also positively related to mean predawn and daily minimum leaf water potentials. Consequently, in the studied G. senegalensis stands, above-ground biomass appeared to be controlled by water availability – rainfall, runoff, infiltration – rather than by cropping intensity. Anisohydric stomatal regulation, resulting in large safety margins from critical transpiration, was inferred from the diurnal amplitude of leaf water potential during the dry season. The plant physiological resistance to water stress combined with its population adjustment in density and growth contributes to the sustainable dominance of G. senegalensis.

To study vertical gradients of arthropod species richness in tropical forests, adult chrysomelids were surveyed with similar sampling effort by beating in four plots of 0.8 ha, representative of the canopy and understorey of one wet and one dry forest in Panama. Samples included in total 4615 individuals representing 253 species, and were of similar species richness at the two study sites. At both sites, chrysomelids were significantly more species-rich in the canopy than in the understorey. The proportion of species shared between the two study sites was 24%, whereas 16% and 28% of species were shared between the canopy and understorey of the wet and dry sites, respectively. Mature trees supported more and different chrysomelid species than conspecific saplings. A higher proportion of liana feeders vs. tree feeders occurred at the dry site than at the wet site. Multivariate analyses confirmed the faunal differences between the wet and dry sites and that stratification was more marked at the wet site than at the dry site. The latter observation may relate to differences in forest physiognomy (a tall and closed canopy at the wet site) and to the high interconnectivity via lianas between the understorey and canopy at the dry site.

Anthropogenic wildfires are becoming increasingly frequent in wet tropical forests. This trend follows that of other anthropogenic disturbances, which are now acute and widespread. Fires pose a potentially serious threat to tropical forests. However, little is known about the impact of unintended forest fires on below-ground resources in these ecosystems. This study investigated the influence of fires on the distribution and variability of soil resources on two sets of 50×50-m burned and unburned plots in a Nicaraguan rain forest. Samples were collected at 5-m intervals throughout each plot as well as subsamples at 50-cm intervals. Geostatistical techniques as well as univariate statistics were used to quantify the spatial autocorrelation and variability of selected nutrients (N, P and K), carbon and standing leaf litter. Most variability in this forest was spatially dependent at a scale of 30 m or less. However the average range of autocorrelations varied greatly between properties and sites. Burning altered soil heterogeneity by decreasing the range over which soil properties were autocorrelated. Overall the average patch size (range) for nitrogen was reduced by 7%, phosphorus by 52%, potassium by 60% and carbon by 43%. While phosphorus and leaf litter increased in the burned plots compared to unburned plots, potassium was not different. Nitrogen and carbon did not display a consistent pattern between burning regimes and this may be explained by variation in fire intensity. Leaf litter measurements did not correlate with measured soil nutrients within plots. Observed changes in the burned forest were likely a result of both the intensity of burning and change in vegetative cover between the time of the fires and soil sampling.

Upper Amazonian tree communities are famous for their very high alpha-diversity. This paper describes an anomalous forest just 6 km south of the equator in lowland Ecuador that is structurally mature, surrounded by hyperdiverse forest, but strikingly poor in species. To investigate the anomaly, a 1-ha tree inventory and soil analysis were carried out and compared with 15 similar surveys of upland forest in the same region. The anomalous forest contains only 102 tree species ha−1, compared with a regional mean of 239±28 species ha−1. It is structurally indistinguishable from richer forests, and closest in composition to upland forest, but lacks the uplands' typically rich understorey tree community. Three hypotheses for its origin are examined: recovery from anthropogenic disturbance, unique soil conditions and recovery from a large-scale natural catastrophe. The third hypothesis is the best supported. Mineralogical, geological and remote-sensing evidence, and 14C-dating suggest that the forest grows on a vast debris plain left by a catastrophic flooding event roughly 500 y ago. The forest's low diversity today is most likely due to the failure of a full complement of the region's tree species – especially understorey taxa – to recolonize the outwash plain in the time since the disaster.

Abundance of floor-dwelling frogs was studied in the humid subtropical lowland forest of Iriomote Island, the Ryukyu Archipelago, southern Japan. Field survey was conducted monthly for 1 y, and the seasonal pattern of abundance was examined for each species. We observed 251 frogs belonging to four species of three families in a total sampling of 600 m2. Of these species, Microhyla ornata (66.5% of total frogs) and Fejervarya limnocharis (30.3%) were dominant in number. Total abundance of the frogs was strongly seasonal and peaked in the hottest month, August, while the number of frog species was one to three throughout the year. The seasonal pattern of abundance showed considerable variation among species, most likely due to their differences in reproductive season, and in physiological characteristics under the influence of varying temperature and rainfall. These patterns are more similar to those of temperate species than those of tropical species. The estimated density of all frogs ranged from 0.08 to 1.60 individuals m−1, and their average annual density and biomass were 4180 individuals ha−1 and 11460 g ha−1, respectively. These values, among the highest so far reported for floor-dwelling anurans, may be a feature of amphibian fauna in the humid subtropical forests on a small island.

The distribution and formation of foraging trails have largely been neglected as factors explaining harvesting patterns of leaf-cutting ants. We applied fractal analysis, circular, and conventional statistics to published and newly recorded trail maps of seven Atta colonies focusing on three aspects: permanence, spatio-temporal plasticity and colony life stage. In the long term, trail patterns of young and mature Atta colonies revealed that foraging activities were focused on distinct, static sectors that made up only parts of their potentially available foraging range. Within these foraging sectors, trails were typically ephemeral and highly variable in space and time. These ephemeral trails were concentrated around permanent trunk trails in mature and around nest entrances in young colonies. Besides these similarities, the comparison of trail systems between the two life stages indicated that young colonies exploited fewer leaf sources, used smaller and less-complex systems of foraging trails, preferred different life forms as host plants, and switched hosts more often compared with mature colonies. Based on these analyses, we propose a general hypothesis which describes the foraging pattern in Atta as a result of initial foraging experiences, spatio-temporal distribution of suitable host plants, energetic constraints, and other factors such as seasonality and interspecific predation.

Ecological theory predicts that interspecific interactions can affect population and community dynamics. Two experiments were conducted on Mantidactylus bicalcaratus and M. punctatus, two sympatric frog species from Madagascar that live and breed in rain-forest plants (Pandanus spp.), to test for interspecific competition. The first experiment examined larval growth rates and survivorship with and without conspecifics. While survivorship did not differ among treatments, mean growth rates for M. bicalcaratus were significantly reduced in the presence of M. punctatus larvae. The second experiment manipulated the presence and density of adults in Pandanus plants. Emigration from and immigration to experimental plants tended to be higher and lower, respectively, for M. bicalcaratus in the presence of M. punctatus, but these differences were not significant. These results demonstrate asymmetric competition (at least as larvae) and indicate that M. punctatus is the superior competitor. Field data showed that M. bicalcaratus was found significantly more frequently in the absence of M. punctatus. Also, M. bicalcaratus populations were significantly more likely to go locally extinct in the presence of M. punctatus. These data suggest that asymmetric competitive interactions are important influences on the dynamics of these populations.

Seasonality in tropical forests can be more subtle than that of temperate forests but still affects the resources available to wildlife. Much work has been done describing changes in fruit availability and dietary composition but the nutritional quality of any particular food item is assumed to be relatively constant. We investigated seasonal changes in the quality of the ripe fruit of Celtis durandii, a common tree that produces fruit year-round and is important in the diets of many species. The lipid content of the ripe fruit was found to be highly variable (0.3–30.8% dry matter) among months and this variation was positively correlated with the summed daily rainfalls of the previous and concurrent months. The amount of this fruit in the diets of three frugivorous primate species (Cercopithecus mitis, Cercopithecus ascanius and Lophocebus albigena) was positively related to measured or estimated lipid levels in the fruit. Such predictable changes in the quality of a constantly available fruit have not been previously reported and suggest that the resources provided by tropical forests may be more seasonal than shown by common measures of fruit availability.

We studied diet choice by a generalist herbivore, the red howler monkey (Alouatta seniculus) under conditions of high and normal population density. Densities equivalent to 800–1000 km−2 (roughly 20–40 times normal) occurred in populations trapped on small, predator-free islands in Lago Guri, Venezuela. For three successive years, we studied one such population, a group of six animals living on a 0.6-ha island and compared its feeding ecology to that of two groups living at normal densities on a 190-ha island. The 0.6-ha island supported a total of 351 trees >10 cm dbh of 46 species, whereas >100 species probably occurred within the 16- and 23-ha home ranges of the two large-island howler troops. Small-island howlers were thus predicted to consume fewer resources, in particular less fruit, and to be less selective in diet choice than large-island howlers. As predicted, small-island howlers consumed fewer resources and obtained a smaller fraction of their intake from rare tree species (those contributing <1% of basal area). Small-island howlers consumed less fruit (2% of feeding time vs. 22%) and more foliage (73% vs. 55%) than the large-island groups. Diet breadth of small-island howlers was markedly less than that of their large-island counterparts. Tree species not present on the small island contributed ≥60% of leaf consumption by large-island howlers. Foliage sources preferred by large-island howlers were different in each of 3 years, whereas foliage of the same species of tree consistently ranked first on the small island. Long-term persistence (17 y) of self-perpetuating howler groups on Lago Guri islets at >20 times normal density strongly suggests that food availability does not limit mainland populations.

Habitat selection can be detected at different scales. For small mammals, habitat studies normally evaluate micro- or meso-habitat selection, but rarely are multiple scales analysed simultaneously. These studies use quantitative data of structural characteristics around the trap stations to evaluate selection. We test an alternative method to evaluate habitat selection in three didelphid marsupials (Didelphis aurita, Philander frenatus and Metachirus nudicaudatus), using a spool-and-line device, and measure habitat selection at two spatial scales. Habitat was characterized by seven habitat variables measured at ten points of direction change along the path traversed by each individual, and at 25 trap stations distributed in a 1-ha area. Micro-habitat selection was estimated at each point by comparing habitat measures between the direction selected against other directions. Meso-habitat selection was estimated by comparing habitat measurements on the whole path of an individual against habitat measures on the 25 trap stations. Patterns of selection for each species could only be detected at the meso-habitat scale, although a few individuals in each species were selective at the micro-habitat scale. Studies of habitat selection need to address the scale of study quantitatively, if possible comparing two or more scales to determine the scale of selection.

Due to the unbalanced distribution of their fauna and flora, which leads to the creation of a niche opportunities, it is generally accepted that island communities offer weak biotic resistance to biological invasion. In order to empirically test this statement, we compared resource use by ants in the understorey of an undisturbed New Caledonian rain forest recently invaded by the little fire ant, Wasmannia auropunctata. We tested the exploitation of: (1) food sources by placing baits on all trees with trunks greater than 5 cm in diameter; and (2) nesting sites on two tree species likely to shelter ant colonies. In non-invaded areas, the native ants occupied only 44.6% of the baits after 2 h of exposure, while in invaded areas all the baits were occupied by numerous W. auropunctata workers. Similarly, in non-invaded areas only 48.9% of Meryta coriacea (Araliaceae) trees and 64.5% of Basselinia pancheri (Arecaceae) sheltered ants, while in invaded areas W. auropunctata nested in 92.6–98.3% of these trees. Also, workers attended native Margarodidae (Hemiptera) for which they promoted the development of populations significantly larger than those attended by native ants. Thus native ants appear unable to efficiently exploit and defend several of the available food sources and nesting sites, providing a niche opportunity for an invader like W. auropunctata.

Spatial variation in canopy structure induces highly variable patterns of light availability for juveniles and understorey trees (Nicotra et al. 1999, Oberbauer et al. 1988, Tang et al. 1999). In the heterogeneous light environment, plasticity in resource use is essential for survival of juveniles and hence maintenance of populations (Valladares et al. 2000). One of the most striking characteristics of plants in their response to a given resource availability is their capacity to change morphological traits (Küppers 1994). Previous studies have attempted to understand the significance of morphological adaptations in their contribution to competitive ability, focusing on crown form (Kohyama 1987, Sterck et al. 2003), branching pattern (King 1998), shoot architecture (Canham 1988) and leaf shape (King & Maindonald 1999, Takenaka 1994, Yamada et al. 2000). Several studies have investigated the changes in morphology of saplings or seedlings growing under different light availability across canopy gaps (Ackerly & Bazzaz 1995, King 1994, Valladares & Pearcy 1998). However, the phenotypic plasticity in morphology of saplings adapting to the fluctuation of light availability has been investigated less (Blundell & Peart 2001, Poorter & Werger 1999).