We tested the hypothesis that seed size influences which frugivores eat fruits, and the size and nature of disperser assemblages in Pakke Tiger Reserve, India. Four tree species had large seeds (> 18 mm width) that could be handled by birds with large gape widths, while two tree species had smaller seeds (< 7 mm width) falling within the gape size range of many frugivores. We tested whether (1) disperser assemblage and activity reflected seed size, and (2) large-gaped hornbills were more effective at fruit removal of tree species with large seeds, than of those with smaller seeds and many dispersers. Day-long watches were conducted in 2005 at trees of Dysoxylum binectariferum (three in 2005 and nine in 2006), Chisocheton cumingianus (nine), Aglaia spectabilis (seven), Polyalthia simiarum (nine), Litsea monopetala (four) and Cinnamomum bejolghota (two in 2005 and six in 2006) to determine which frugivores visited trees and ate fruit. Disperser visitation and species diversity per tree to species with medium-sized seeds averaged 85 visits and 10 species d−1, contrasted with five visits by one frugivore species to large-seeded tree species. Seed removal rates per tree averaged 486 seeds d−1 from medium-seeded tree species, but 10 seeds d−1 from large-seeded species. Hornbills (Bucerotidae) and Ducula badia (Columbidae) removed large seeds from capsules of Aglaia spectabilis, Chisocheton cumingianus and Dysoxylum binectariferum (Meliaceae). Primates, civets and bats also consumed drupes of large-seeded Polyalthia simiarum (Annonaceae). Anthracoceros albirostris hornbills were important dispersers of Litsea monopetala (Lauraceae), a medium-seeded tree with a large disperser assemblage. Conversely, hornbills were quantitatively inconsequential for Cinnamomum bejolghota (Lauraceae), another medium-seeded tree species with several dispersers. Results suggest that the size and activity of disperser assemblages accurately reflects seed size. While hornbills were quantitatively important dispersers of large-seeded tree species, their effectiveness for trees with small- to medium-sized seeds depended on the tree species.

Soil fertility and plant productivity are known to vary across the Amazon Basin partially as a function of geomorphology and age of soils. Using data on herpetofaunal abundance collected from 5 × 5 m and 6 × 6 m plots in mature tropical forests, we tested whether variation in community biomass of litter frogs and lizards across ten Neotropical sites could be explained by cation exchange capacity, primary productivity or stem turnover rate. About half of the variation in frog biomass (48%) could be attributed to stem turnover rate, while over two-thirds of the variation in lizard biomass (69%) was explained by primary productivity. Biomass variation in frogs resulted from variation in abundance and size, and abundance was related to cation exchange capacity (45% of variation explained), but size was not. Lizard biomass across sites varied mostly with individual lizard size, but not with abundance, and size was highly dependent on primary productivity (85% of variation explained). Soil fertility and plant productivity apparently affect secondary consumers like frogs and lizards through food webs, as biomass is transferred from plants to herbivorous arthropods to secondary consumers.

Although predation is generally seen as one of the key factors determining the abundance and composition of insect herbivore communities in tropical rain forests, quantitative estimates of predation pressure in rain-forest habitats remain rare. We compared incidence of attacks of different natural enemies on semi-concealed and exposed caterpillars (Lepidoptera) in lowland and montane tropical rain forests, using plasticine models of caterpillars. We recorded attacks on caterpillars in four habitats: primary forest, secondary forest and forest fragment in lowlands (200 m asl), and montane primary forest (1700 m asl). We used 300 exposed and 300 semi-concealed caterpillars daily, and conducted the experiment for 6 d in every habitat. Daily incidence of attacks was higher on exposed caterpillars (4.95%) than on semi-concealed (leaf-rolling) caterpillars (2.99%). Attack pressure of natural enemies differed also among habitats. In the lowlands, continuous primary and secondary forests had similar daily incidence of attacks (2.39% and 2.36%) which was however lower than that found in a primary forest fragment (4.62%). This difference was caused by higher incidence of attacks by birds, ants and wasps in the forest fragment. The most important predators were birds in montane rain forests (61.9% of identified attacks), but insect predators, mostly ants, in the lowlands (58.3% of identified attacks). These results suggest that rapid decrease in the abundance of ants with altitude may be compensated by increased importance of birds as predators in montane forests. Further, it suggests that small rain-forest fragments may suffer from disproportionately high pressure from natural enemies, with potentially serious consequences for survival of their herbivorous communities.

Following the 2003 invasion of the armoured scale Aulacaspis yasumatsui to Guam, changes to population traits of the dominant Cycas micronesica were determined. Belt transects with a width of 4 m and an average length of 120 m were established in October 2004 to document plant mortality until January 2011. Stem height, basal diameter and leaf number were also measured for each plant and used to determine density, demography and allometric relationships. Allometric traits and a left-skewed demographic structure of the pre-invasion C. micronesica habitat documented a thriving population with high recruitment potential. Aulacaspis yasumatsui dispersed into the study site 4 mo after the initial census. All seedlings were killed within 9 mo and all juvenile plants were killed within 40 ± 10 mo. Mortality reached 92% by 6 y after chronic scale infestations. Allometry and demography of the 2011 survivors described a collapsing C. micronesica population of stressed and reproductively challenged trees with no recruitment. This classic example of the enemy release hypothesis has resulted in a homogeneous decline in plant density from 2007–2011. The trend predicts extirpation of C. micronesica from west Guam habitats by 2019.

In heterotrophic streams the retention and export of coarse particulate organic matter and associated elements are fundamental biogeochemical processes that influence water quality, food webs and the structural complexity of forested headwater streams. Nevertheless, few studies have documented the quantity and quality of exported organic matter over multiple years and under a range of conditions that includes both droughts and hurricanes. This study quantifies the export of coarse particulate organic matter (CPOM, > 12.7 mm), over 18 y in two headwater streams in north-east Puerto Rico. Daily exports ranged from 0 to over 170 g ha−1 d−1 and averaged 7.39 g ha−1 d−1, with similar amounts coming from leaves (3.5 g ha−1 d−1) and wood (3.2 g ha−1 d−1). Export of coarse particulate organic carbon was 3.0 g ha−1 d−1 which constitutes only 1.32% of carbon exports. Most litter falling into the streams was processed in place as only 2.3% of the leaf litter falling directly into these perennial channels was exported as CPOM. On average, 6 wk y−1 had no exports while events transporting more than 10 g ha−1 d−1 occurred every 2.8 mo. Instead of a single annual pulse as observed in deciduous systems, there were annual peaks in CPOM exports during May and September and less export during the drier period from December to February. Ratios of C:N in the exported material were highest in the driest month and lowest during rainy months, while leaf fluxes for nitrogen, phosphorus and calcium were highest in rainy months and lowest during February. Although median daily exports and exports during low- and base-flow periods were similar before and after Hugo, after 16 y exports during moderate- and high-flow periods were still less than those in the 2 y prior to the hurricane. Our observations indicate a system with high rates of internal processing that quickly returns to median daily conditions following hurricanes but requires several decades for storm-flow exports to return to pre-disturbance conditions and indicates that the long-term pattern of CPOM export is associated with the level of maturity of watershed vegetation.

Bulbul species (family Pycnonotidae) are important seed dispersers in Asian forests, but almost nothing is known of their movement patterns inside intact forest, which are likely to impact forest dynamics. We examined the movement patterns of the forest-dwelling puff-throated bulbul (Alophoixus pallidus) in relation to fruit productivity and distribution of fruiting trees/lianas in an evergreen forest in north-eastern Thailand. Movement patterns of 10 groups were precisely mapped by following colour-ringed individuals in each group 4 h mo−1 for 1 y. We evaluated fruit productivity and dispersion of fruiting trees/lianas based on monthly phenologies. There were clear seasonal fluctuations in fruit availability, which appeared to affect movement patterns, particularly distance moved between fruiting trees, time spent feeding and food selection. When fruit availability was low, bulbuls spent more time on average at a given food plant and moved longer distances between fruiting plants than compared with periods of higher fruit availability (low availability: 58 s, 83.2 m; high availability: 10 s, 43.4 m). This study points to the importance of seasonal availability of fruit resources on frugivore movement patterns. Seasonal dynamics of movement may be useful for understanding interactions between fruiting trees and their dispersers, and forest tree recruitment patterns.

The Neotropics house two guilds of large arboreal vertebrate seed predators: parrots and the pitheciin primates. Both have diets dominated by immature fruits. The possibility of members of the Pitheciinae (genera Cacajao, Chiropotes and Pithecia) acting as occasional seed dispersers has been mooted, but not experimentally shown. We combined primate behavioural data and seed germination data from three separate field studies in the Brazilian states of Amazonas and Pará to analyse patterns of post-consumption seed survivorship for seeds discarded by three pitheciin species (Cacajao melanocephalus ouakary, Chiropotes chiropotes and Chiropotes albinasus). We then calculated the frequency of dispersal events for four species eaten by C. m. ouakary. All three primate species dropped intact seeds while feeding, and 30.7% of 674 dropped seeds germinated ex situ. Undamaged seeds from unripe and ripe samples germinated (29.3% and 42.7%, respectively), and all three primate species carried some fruits up to 20 m from the parent tree before consuming them. Potential seed-dispersal events varied from 1 (Macrolobium acaciifolium) per fruiting cycle to more than 6500 (Duroia velutina), suggesting that there are differences in dispersal potential. In summary, although they are highly specialized seed predators, these primates may also act as important dispersers for some plant species, and effective dispersal is not restricted to ripe fruits, as immature fruits removed from a tree may continue to mature and the seeds later germinate, a much-neglected aspect of dispersal ecology. The possibility that similar events occur in parrots should be experimentally investigated.

The fruits of Melastomataceae are consumed by many Neotropical frugivorous birds. Several studies have reported segregated fruiting seasons of melastomes, but this pattern is not widespread. The segregated fruiting phenologies of congeneric sympatric species may be an evolutionary response to reduce competition for seed dispersers. Alternatively, aggregated fruiting phenologies may be favoured if local fruit abundance attracts more frugivores, thus enhancing seed dispersal. We monitored melastome fruiting in transects over a 2-y period at a cloud-forest site in the Colombian Andes. Fruiting periods of nine melastome species were aggregated and fruiting peaks coincided with rainy seasons. In a separate 6-mo study, observations at focal plants revealed that 47 of 61 bird species fed on 10 species of melastome, representing 37.4% of feeding events observed. Melastomes were consumed by birds in a higher proportion than expected from their availability and peak melastome fruit abundance coincided with the breeding season of the frugivore community, when melastomes constituted 54% of feeding records. Melastomes interact with a large number of bird species throughout their annual cycles, and seem to constitute pivotal elements that sustain the frugivore community in montane forests.

The correlation between the biomass of forest primates and a chemical index of the average nutritional quality of leaves in tropical forests has been repeatedly documented since 1990. We tested the role played by protein : fibre on lemur biomass in a gallery forest in southern Madagascar. Plant species abundance was determined based on transect censuses. We calculated an average ratio of protein-to-fibre in leaves and an abundance-weighted ratio, i.e. the mean weighted by the basal area of tree species, to be compared with the figures available for other forest ecosystems in Madagascar and a number of anthropoid habitats. Lemur densities were evaluated through compilation of previous studies made from prior to 1975 and up until 2011 based on strip censuses and/or identification of all groups supplemented with new censuses. A high mean ratio of protein to fibre (> 0.4) supports high folivore biomass at 390 kg km−2 (reaching 630 kg km−2 in the closed-canopy forest area) compared with primate communities in other Malagasy forests (protein : fibre: < 0.5; folivore biomass: < 440 kg km−2), as predicted. However, the data corroborate the finding that the total biomass of lemur communities as well as the biomass of folivorous lemur species are low compared with those of African and Asian primate communities for a given protein : fibre ratio. Tree diversity and leaf production do not consistently explain this pattern. In contrast, the extinction of large folivorous lemurs during the past two millennia presumably allowed too little time for smaller-sized species to evolve equally effective morphological and physiological specializations for processing a large range of fibrous foods.

The decomposition of leaf litter is governed, in part, by litter invertebrates. In tropical rain forests, ants are dominant predators in the leaf litter and may alter litter decomposition through the action of a top-down control of food web structure. The role of ants in litter decomposition was investigated in a Costa Rican lowland rain forest with two experiments. In a mesocosm experiment, we manipulated ant presence in 50 ambient leaf-litter mesocosms. In a litterbag gradient experiment, Cecropia obtusifolia litter was used to measure decomposition rate constants across gradients in nutrients, ant density and richness, with 27 separate litterbag treatments for total arthropod exclusion or partial arthropod exclusion. After 2 mo, mass loss in mesocosms containing ants was 30.9%, significantly greater than the 23.5% mass loss in mesocosms without ants. In the litter bags with all arthropods excluded, decomposition was best accounted by the carbon: phosphorus content of soil (r2 = 0.41). In litter bags permitting smaller arthropods but excluding ants, decomposition was best explained by the local biomass of ants in the vicinity of the litter bags (r2 = 0.50). Once the microarthropod prey of ants are permitted to enter litterbags, the biomass of ants near the litterbags overtakes soil chemistry as the regulator of decomposition. In concert, these results support a working hypothesis that litter-dwelling ants are responsible for accelerating litter decomposition in lowland tropical rain forests.

Insect-pollinated Laguncularia racemosa has a variable breeding system; some populations are androdioecious, with male and hermaphroditic plants, while others lack male plants. We observed the foraging behaviours of insects in three androdioecious and three hermaphrodite-only populations of L. racemosa in Florida. In each population, insect visitation rates were estimated from 30–108 timed intervals. We recorded the number of flowers visited by 144–224 insects during foraging bouts made to 15–40 male and hermaphroditic plants. Male plants in androdioecious populations had significantly more visitors than hermaphroditic plants, increasing the number of vectors carrying pollen from male plants. Further, many insects visited few flowers during foraging bouts, which should increase outcrossing frequency. According to mathematical models, male plants benefit from these combined factors. Plants in hermaphrodite-only populations had significantly more visitors than hermaphroditic plants in androdioecious populations. Proportionately more insects visited many flowers during foraging bouts in hermaphrodite-only versus androdioecious populations. The increased likelihood of geitonogamous self-pollination could help explain the lack of male plants in hermaphrodite-only populations. Differences in pollinator assemblages and the relative abundances of several species were responsible for differences in foraging behaviours: Apis mellifera, Bombus sp., Melissodes sp., Xylocopa sp., Euodynerus sp. and a calliphorid species.

A climate model, based on effects of water availability and temperature, was recently proposed to explain global variation in bat species richness along altitudinal gradients. Yet such studies are sparse in the tropics and near-absent in Africa. Here we present results from an altitudinal study of bat diversity from Mount Mulanje, Malawi. Using ground nets, canopy nets and harp traps, we sampled eight sites across three habitat zones from 630 m to 2010 m asl. We assessed the influence of climatic, geographic and biotic variables on measures of estimated species richness, Fisher's α, and an unbiased index of compositional turnover. We recorded 723 individuals and 30 species along the gradient, revealing a ‘low plateau’ pattern in estimated species richness, peaking at 1220 m, which is congruent with the global climate model. Measures of local habitat structure significantly explained a large degree of variation in species richness and compositional turnover between sites. Fisher's α was further significantly correlated to mean annual relative humidity, suggesting a background climatic influence.

In this study we investigated the importance of two species of fruit bat (Rousettus leschenaulti and Cynopterus sphinx) as seed dispersers for a species of fruit tree (Syzygium oblatum) found in the Xishuangbanna Tropical Botanical Garden in South-West China. We found that although R. leschenaulti and C. sphinx were the two primary seed dispersers of S. oblatum over half of the fruit produced by the tree (65%) fell to the ground. Out of the fruit collected, R. leschenaulti and C. sphinx were able to disperse seeds up to 73 m from the parent tree with the highest density of feeding roosts occurring at 21.3 m (SE = 5.2 m). We found no signs that either species of bat used the parent tree as a feeding roost, instead choosing specific trees that were at lower densities compared with other trees in the forest that were not used. When comparing the viability of seeds in three different habitats (under parent tree, in forest gap, under feeding roost) survival analysis revealed that seedling survival was significantly higher in the forest gap (91.7% ± 4.41%) than under the parent tree (78.3% ± 1.67%), but was not significantly different to seedling survival underneath feeding roosts (86.7 ± 1.67%). Further work also showed that the seeds did not have to be removed from the fruit or ingested by the bat in order to germinate. We conclude that although S. oblatum is not dependent on R. leschenaulti and C. sphinx for successful germination of its seeds, these two species of bat are important seed dispersers and can move seeds to areas where there is a greater chance of germination success and survival.

We conducted a study in 30 plots distributed uniformly in an area of 25 km2 at Ducke Reserve, Manaus, to test the hypothesis that understorey herb richness and abundance are mediated by litter, manipulating experimentally the amount of litter in the field. Over 10 mo, we followed the emergence of herbaceous seedlings and sporophytes in control, litter-addition and litter-exclusion treatments, covering an area of 1.2 m2 per plot in each treatment. We also assessed the relationship between topography and litter depth and frequency of bare-soil patches; and the influence of density of reproductive individuals on the emergence of herbs. Litter depth decreased, and the frequency of bare-soil patches increased with terrain slope in the wet season, but were not related with the soil clay content. Neither was related to the topography in the dry season. Emergence of pteridophytes was four times higher in the litter-exclusion treatment (3.7 ± 1.1 individuals m−2) than in the litter-addition treatment (0.9 ± 0.28 indiv. m−2) and increased with soil clay content. Seedlings from monocot herbs emerged twice more frequently in the litter exclusion (0.71 ± 0.25 indiv. m−2) than in the litter-addition treatment (0.33 ± 0.11 indiv. m−2), and also more in sites with high density of fruiting plants. The results are consistent with the hypothesis that regeneration of herbs with very small propagules is strongly affected by the physical barrier imposed by litter. Given that litter is shallower on slopes during the wet season, this creates a pattern of higher density and richness of pteridophytes in these areas. Monocot herbs, although also limited by litter, were more highly limited by availability of propagules, and their distribution patterns are at least in part explained by dispersal limitation. We conclude that litter is an important causal factor behind the herb distribution patterns along topographical gradients.

Primates commonly consume leaves that are high in protein but low in digestion-inhibiting fibre. Due to the fact that mature leaves do not meet these criteria, they are typically avoided and many leaf-eating primates select for leaves high in protein and low in fibre leading to the theory that food selection is based on protein maximization. However, feeding records for a population of black howler monkey (Alouatta pigra) in Monkey River, Belize, collected over a 5-y period, together with synchronous phenological data, indicate that this population does not meet the expectation and actually prefer mature leaves. This study aims to describe the nutritional composition of the food supply and investigate the possibility that, rather than to maximize protein ingestion, mature leaves are eaten to balance nutrient intake. Macronutrient analyses (moisture, lipids, protein, NDF, ADF and simple sugars) were conducted on a sample of 96 plant samples from 18 food species of this population of black howler. Results reported here show that mature leaves eaten by howlers in this forest contain sufficient protein to meet minimum metabolic requirements (range: 11.6–24%; mean: 16.4% ± 3.8%) and have significantly higher concentrations of simple sugars than young leaves (means of 7.2% ± 2.7% vs. 4.4% ± 2.3% respectively). Thus, it appears that mature leaf ingestion is likely serving to balance energy and protein intake. This result may be due to the disruptive effects of a hurricane in 2001 that resulted in a loss of 80% of the howler population, changed forest composition and may have affected plant chemistry. Despite this, the data reported here suggest that the accepted view that mature leaves are simply fallback foods for primates, eaten only in times of preferred food scarcity, may have to be revised.

Reproductive success of many plant species declines in fragmented habitat, but this effect is little studied in trees of tropical rain forest understorey. Paypayrola blanchetiana (Violaceae) is a continuous-flowering treelet endemic to the Atlantic rain forest of north-east Brazil. Plants are distributed in localized patches. Flower, fruit and seed production of a total of 86 trees was quantified in six forest fragments, each belonging to one of two categories of size (> 300 ha vs. < 50 ha) and fragmentation history (isolated for c. 25–30 y vs. at least c. 50 y). Relative fruit set (fruits/flower) and seed set (seeds/ovule) were calculated for a spatial (total fruit set of tree individuals) and a temporal analysis (combined month-by-month fruit set in patches in response to different flowering intensities). Fruit set (1%) and seed set (0.6%) were very low, but variable among trees. Plants in large fragments had significantly higher fruit set and seed set than plants in small fragments. Trees in older fragments, however, displayed similar fruit and seed set to those in recently created ones. We found no interaction effect on seed set of fragment size and isolation time. Seed set was a negative function of patch flower density. Possible drivers of the observed patterns are discussed.

The spatial heterogeneity of tropical forest epiphytes has rarely been quantified in terms of biomass. In particular, the effect of topographic variation on epiphyte biomass is poorly known, although forests on ridges and ravines can differ drastically in stature and exposure. In an Ecuadorian lower montane forest we quantified epiphytic biomass along two gradients: (1) the twig–branch–trunk trajectory, and (2) the ridge–ravine gradient. Twenty-one trees were sampled in each of three forest types (ridge, slope, ravine positions). Their epiphytic biomass was extrapolated to stand level based on basal area–epiphyte load relationships, with tree basal areas taken from six plots of 400 m2 each per forest type. Our results document the successional addition and partial replacement of lichens by bryophytes, angiosperms and finally dead organic matter along the twig–branch–trunk trajectory. Despite having the highest tree basal area, total epiphytic biomass (mean ± SD) of ravine forest was significantly lower (2.6 ± 0.7 Mg ha−1) than in mid-slope forest (6.3 ± 1.1 Mg ha−1) and ridge forest (4.4 ± 1.6 Mg ha−1), whereas maximum bryophyte water storage capacity was significantly higher. We attribute this pattern to differences in forest dynamics, stand structure and microclimate. Although our study could not differentiate between direct effects of slope position (nutrient availability, mesoclimate) and indirect effects (stand structure and dynamics), it provides evidence that fine-scale topography needs to be taken into account when extrapolating epiphytic biomass and related matter fluxes from stand-level data to the regional scale.

To assess the effects of soil water status on the spatial variation in soil carbon dioxide (CO2), methane (CH4) and nitrous oxide (N2O) fluxes, we examined these gas fluxes and environmental factors in a tropical rain forest in Peninsular Malaysia. Measurements of soil CO2, CH4 and N2O fluxes were taken ten, nine, and seven times, respectively over 30 mo at 15 or 39 sampling point within 2-ha plot. Mean (± SE) value of spatially averaged CO2 flux was 4.70 ± 0.19 μmol CO2 m−2 s−1 and observed spatial variation in CO2 flux was negatively related to the volumetric soil water content (VSWC) during the dry period. Over the wet period, extremely high CO2 emissions were positively correlated with VSWC at some locations, suggesting that no spatial structure of CO2 flux was because of such hot-spot CO2 emissions. Flux of CH4 was usually negative with little variation, with a mean value of –0.49 ± 0.15 mg CH4 m−2 d−1, resulting in the soil at our study site functioning as a CH4 sink. Spatial variation in CH4 flux was positively related to the VSWC throughout the entire study period (dry and wet). Some CH4 hot spots were observed during dry periods, probably due to the presence of termites. Mean value of spatially averaged N2O flux was 98.9 ± 40.7 μg N m−2 h−1 and N2O flux increased markedly during the wet period. Spatially, N2O flux was positively related to both the VSWC and the soil N concentration and was higher in wet and anaerobic soils. These findings suggest that denitrification is a major contributor to high soil N2O fluxes. Additionally, analysis by adjusting confounding effects of time, location and interaction between time and location in mixed models, VSWC has a negative effect on CO2 flux and positive effects on CH4 and N2O fluxes. We found that soil water status was related temporally to rainfall and controlled greenhouse gas (GHG) fluxes from the soil at the study site via several biogeochemical processes, including gas diffusion and soil redox conditions. Our results also suggest that considering the biological effects such as decomposer activities may help to explain the complex temporal and spatial patterns in CO2 and CH4 fluxes.

Species commonly exist in sympatry, yet ecological studies are often based on a single species approach while ignoring the impact of sympatric competitors. Over 13 mo we used 24 remote video-camera traps to monitor habitat use of sympatric chimpanzee, gorilla and elephant in four different habitat types in Loango National Park, Gabon. Habitat use by each species was predicted to vary according to seasonal changes in food availability and precipitation. Increased interspecific competition between the three species was expected at times of reduced resource availability, leading to exclusion of the inferior competitor. Supporting the predictions, species abundance per habitat showed seasonal variation: all three species responded positively to increased fruit availability in all habitats, but the response was only significant for gorilla in mature forest and elephant in coastal forest. Responses to rainfall varied, with the chimpanzee responding negatively to rainfall in swamp forest, the gorilla responding positively to rainfall in coastal and secondary forest, and the elephant responding positively to rainfall in mature forest. Elephant presence resulted in competitive exclusion of the apes under certain conditions: the chimpanzee was excluded by the elephant where fruit availability was low, whereas the gorilla was excluded by the elephant in areas of low herb density despite high fruit availability. Our results emphasize the value of applying a multi-species, longer-term approach to studying variation in habitat use among sympatric species and highlight the impact competitors can exert on one another's distribution.

Secondary forests promote an increased abundance of leaf-cutting ants (LCA) but the consequences on forest dynamics have been poorly addressed. Here we examine seedling assemblage attributes, seed germination and seedling survival across foraging zones of Atta cephalotes colonies inhabiting 15 second-growth patches (25–47 y old) of Atlantic forest. Seeds and seedlings were monitored within foraging zones and control sites over 1 y, including spots around ant nests. Overall, 1862 seedlings from 108 plant species were recorded. Seedling density decreased by 53% in foraging-zone plots (3.31 ± 0.23 seedlings m−2) when compared with control plots (7.02 ± 0.44 seedlings m−2) and a similar decrease was observed for species richness. Ant-induced alterations in the seedling assemblage were further indicated by segregation between foraging-zone and control plots (NMDS), habitat effects on species similarity (ANOSIM), and indicator tree species associated with control plots. While seed germination and seedling survivorship were uncorrelated to either nest distance or age of second-growth stands (with the exception of Tapirira guianensis), defoliation by LCAs was a significant cause of seedling mortality. Our results suggest that LCAs interfere with successional trajectories of Atlantic forest as foraging zones filter seedling establishment, supporting less-dense, impoverished and convergent seedling assemblages.