Carbon and nutrient stocks in below-ground biomass have rarely been investigated in tropical montane forests. In the present study, the amounts of carbon, nitrogen, phosphorus, sulphur, potassium, calcium and magnesium in root biomass were determined by soil coring and nutrient analysis in forests at three altitudes (1900, 2400 and 3000 m) in the Ecuadorian Andes. Root biomass increased markedly from 2.8 kg m−2 at 1900 m and 4.0 kg m−2 at 2400 to 6.8 kg m−2 at 3000 m. The contribution of coarse roots (> 2 mm in diameter) to total root biomass increased from about 70% at 1900 m to about 80% at higher altitudes. In fine roots (≤ 2 mm in diameter), concentrations of nutrients except calcium markedly decreased with altitude. Therefore, the nutrient stocks in fine roots were similar at 1900 m and 3000 m for nitrogen and sulphur, and were even lower at higher altitudes for phosphorus, potassium and magnesium. In coarse roots of Graffenrieda emarginata concentrations of nutrients were substantially lower than in fine roots, and were little affected by altitude. The data suggest that the importance of coarse roots for long-term carbon and nutrient accumulation in total plant biomass increases with increasing altitude.

The mountain gorilla (Gorilla beringei beringei) lives in two geographically separated populations, Bwindi Impenetrable National Park, Uganda and in three national parks spanning the Virunga mountain region in Rwanda, Democratic Republic of Congo, and Uganda. The altitude, climate and plant composition of these habitats differ. Our goal was to compare the diets of gorillas living in each of these habitats. The nutrients in staple foods and in the diets of individuals in a group of gorillas in Bwindi (N = 12 individuals) and a group in the Virungas (N = 7 individuals) were compared to determine if differences in dietary composition affected concentrations of nutrients in their diets. At both sites gorilla diets consisted primarily of herbaceous leaves, but the diet of Bwindi gorillas contained more tree leaves, fruit, pith and dry wood, and fewer stems. Despite differences in habitat and dietary composition, the nutrient concentrations in both gorilla diets were remarkably similar. On a dry matter basis, the diets and staple foods of Bwindi and Virunga gorillas contained similar concentrations of crude protein (CP), fibre (NDF) and non-structural carbohydrates (TNC). Bwindi gorillas ate diets containing 18% CP, 43% NDF and 19% TNC on a dry-matter basis, while the diets of the Virunga gorillas contained 17% CP, 41% NDF and 18% TNC. Our results demonstrate that gorillas consume diets that differ by plant species and part, but contain similar concentrations of nutrients. This suggests that classifying animals by broad dietary strategy (e.g. frugivory and folivory) does not provide a reliable indicator of the nutritional quality of their diet, and that our previous assumptions about these categories should be re-evaluated.

Differentiation of limiting nutrients within small spatial scales has been observed in coastal mangrove forests, but research on other tropical peatlands suggests it is a more widespread phenomenon. In the Changuinola mire of coastal Panama, oligotrophy was hypothesized to increase along a gradient of peat development (peat doming). Nutrient and carbon concentration of leaf tissue, soil, and soil porewater were characterised over a successive sequence of plant communities along the gradient. Soil phosphorus (P) and nitrogen (N) concentrations decreased from 1200 μg P g−1 and 27 mg N g−1 to 377 μg P g−1 and 22 mg N g−1 within 2.7 km into the mire interior. These changes coincided with an increase in soil and average leaf N:P molar ratios from 52–128 and 24–41, respectively. Soil P was strongly related to leaf P and soil N:P to foliar N:P. There was a wide range in δ15N values for canopy (4.0 to −9.4‰), Campnosperma panamense (4.0 to −7.8‰) and understorey (4.8 to −3.1‰) species. Foliar δ15N values of canopy species were strongly related to soil N:P, soil P and leaf P. The depleted foliar δ15N values appeared to be an effect of both the N atmospheric source and P limitation. Here, P limitation is likely associated with ombrotrophic conditions that developed as hydrologic inputs became dominated by precipitation.

Resolving the carbon (C) balance in the Amazonian forest depends on an improved quantification of production and losses of particulate C from forested landscapes via stream export. The main goal of this work was to quantify litterfall, the lateral movement of litter, and the export of coarse organic particulate matter (>2 mm) in four small watersheds (1–2 ha) under native forest in southern Amazonia near Juruena, Mato Grosso, Brazil (10°25′S, 58°46′W). Mean litterfall production was 11.8 Mg ha−1 y−1 (5.7 Mg C ha−1 y−1). Litterfall showed strong seasonality, with the highest deposition in the driest months of the year. About two times more C per month was deposited on the forest floor during the 6-mo dry season (0.65 Mg C ha−1 mo−1) compared with the rainy season (0.3 Mg C ha−1 mo−1). The measured C concentration of the litterfall samples was significantly greater in the dry season than in the rainy season (49% vs. 46%). The lateral movement of litter increased from the plateau (upper landscape position) towards the riparian zone. However, the trend in C concentration of laterally transported litter samples was the opposite, being highest on the plateau (44%) and lowest in the riparian zone (42%). Stream-water exports of particulate C were positively correlated with streamflow, increasing in the rainiest months. The export of particulate C in streamflow was found to be very small (less than 1%) in relation to the amount of litterfall produced.

Barn owl pellet content was studied on seven occasions over a 2-y period during which terrestrial small-mammal populations were assessed via a capture-mark-recapture (CMR) programme in a Sahelian agro-ecosystem of the Inner Delta of Niger River in Mali. Rodents (especially Mastomys huberti representing 78.5% of the total number of prey) were the major prey of the barn owl on all but one occasion, when bats were dominant. This exception coincided with the period of lowest abundance of M. huberti at the study site. Distribution of M. huberti prey into four age classes was assessed through analysis of tooth wear in remains from the seasonal pellet samples. Comparisons with age structure of the CMR population indicate that the barn owl tended to prey on smaller-than-average (thus younger) individuals, especially when these are rare in the population (non-reproductive period between June and October). The spectrum of prey consumed is compared with data previously reported in Sahelian Africa, showing for the first time in this region a major shift in prey choice by the barn owl when its preferred prey becomes rare. At the rodent population level, the apparent choice of younger M. huberti prey at some periods is interpreted in the light of our knowledge on population dynamics of the species in this habitat.

Army ants (Formicidae: Ecitoninae) are top predators in neotropical forests. Army ant raids support a community of diverse organisms, including birds that attend the raids to collect prey. While elevation and forest fragmentation influence army ant and insectivorous bird communities, their effects on the interaction between army ants and bird species is unknown. We studied the size and species composition of bird flocks attending army ant swarms in forest fragments and continuous forest across an elevational gradient (1100–1680 m asl) in a neotropical montane region (Monteverde, Costa Rica). We observed a total of 41 bird species attending army ant swarms. Neither the number of birds, nor the total body mass of birds, nor the number of bird species in attending bird flocks was related to elevation. However, we found a higher bird species richness, larger flock size and greater total body mass of birds attending army ant swarms in continuous forest. Continuous and fragmented forest shared many attending bird species in common, but there was elevational segregation of attending bird species. Some montane endemic birds, and neotropical migrants, attend swarms regularly and use army ant raids as a food source.

Understanding the causes of savanna–forest dynamics is vital as small but widespread changes in the extent of tropical forests can have major impacts on global climate, biodiversity and human well-being. Comparison of aerial photographs for 50 rain-forest patches in Kakadu National Park had previously revealed a landscape-wide monotonic expansion of rain-forest boundaries between 1964 and 2004. Here floristic, structural, environmental and disturbance attributes of the changes were investigated by sampling 588 plots across 30 rain-forest patches. Areas that had changed from savanna to rain forest were associated with a significantly higher abundance of rain-forest trees and less grasses, relative to stable savanna areas. Ordination analyses showed that overall floristic composition was not significantly different between newly established rain forest and longer established rain forest. Generalized linear models also indicated that contemporary levels of disturbance (fire and feral animal impact) and environmental variables (slope and soil texture) were poor predictors of historical vegetation change. We concluded that (1) the rain-forest boundaries are highly dynamic at the decadal scale; (2) rain-forest expansion is consistent with having been driven by global environmental change phenomena such as increases in rainfall and atmospheric CO2; and (3) expansion will continue if current climatic trends and management conditions persist.

Gallery and floodplain forests in monsoonal northern Australia are mostly sclerophyllous and dominated by five closely related species of Melaleuca (Myrtaceae) amongst which niche differentiation is unclear. We present a floristic and environmental analysis of ‘the flooded forest’ using data from 340 plots distributed across 450 000 km2 of the Top End of the Northern Territory. Melaleuca argentea was confined to streams and occurred on sandier substrates, whereas M. cajuputi mostly occurred in the near-coastal lowlands on clay soils. The greater basal area of M. cajuputi suggests an association with productive sites. Melaleuca dealbata, M. viridiflora and M. leucadendra occurred on a wide range of soils. More deeply floodprone sites were occupied by M. argentea and M. leucadendra along streams and by M. leucadendra and M. cajuputi on floodplains and in swamps. A general deficiency but occasional abundance of Melaleuca seedlings suggests that regeneration is episodic. Seedlings were more frequent in recently burnt areas and especially where fires had been severe. We propose that Melaleuca forests occur where disturbance by fire and/or floodwater is too great for rain forest to persist, rendering them the wetland analogue to the eucalypts that dominate well-drained portions of the north Australian environment.

Although indigenous farmers and other traditional peoples have disturbed Amazonian forests for centuries, few studies investigate the effects of these disturbances on biodiversity. This short-term study investigates how bird assemblages are affected by agricultural practices adopted by the residents of a national park in the Brazilian Amazon. Twelve sites in secondary forest (four sites in three age categories) and 12 sites in primary forest were selected for bird sampling. Audio-visual censuses of birds were conducted in small plots (1 ha) of young secondary growth (4–5 y), middle-aged (7–15 y), and old (20–35 y) secondary forests. Each site in secondary forest had a corresponding control site in adjacent primary forest. Young secondary growth had fewer species than old secondary and primary forests. Bird species richness was similar in middle-aged, old secondary and primary forests. Bird communities of young secondary habitat were dominated by species normally found in open habitats, such as agricultural fields and igapó flooded forests. The bird species composition of young secondary growth was more similar to that of middle-aged forest, which in turn was more similar to that of old secondary forests. Bird species composition differed between young secondary growth and old secondary forests and primary forests and between middle-aged secondary and primary forests. Nectarivore/insectivores and frugivore/insectivores of monospecific flocks are significantly associated with secondary forests, especially in initial stages of forest regeneration. In contrast, specialized insectivorous birds that join mixed-flocks show a significant association with primary forest.

Precipitation and temperature are known to have important effects on forest productivity, but these effects may be strongly mediated through their influence on soil and leaf nutrients. We measured indicators of forest productivity and soil and leaf nutrients across independent gradients of precipitation and elevation/temperature in lower montane Hawaiian forests dominated by a single overstorey species, Acacia koa, situated on 1500–3000-y-old soils that were mixtures of volcanic ash and basalt. Stand basal area was highest at the wettest site, 2000 mm mean annual precipitation (MAP), and leaf N and P were lowest at the driest site, 1000 mm MAP. Soil N availability and leaf N concentration declined across an 850-m elevation gradient, but this was not correlated with stand basal area or soil organic matter content. Across all stands, basal area was negatively correlated with the exchangeable soil P fraction. As well, the soil C:N ratio was negatively correlated with both soil P availability and the size of the primary mineral P fraction. Soil P availability and weathering appear to be important determinants of soil organic matter quantity and quality. Overall, precipitation is the major driving force for forest productivity, but P weathering and availability play important roles in limiting productivity in wetter sites and in controlling soil organic matter dynamics in these N-fixing forests.

The long-term dynamics of epiphyte communities are little studied although such baseline data are urgently needed, in particular in the context of global change. Census data of a vascular epiphyte community from 0.4 ha of undisturbed lowland forest in Panama were used to infer future changes in community composition by deducing population growth from the current size class structure of populations. The study includes 11 387 individuals out of 45 species, ranging in abundance from 16 to 1568 individuals. There was a significant negative correlation between the size of a population and the steepness of the size distribution, indicating that more common species are likely to increase in abundance in the future, while rarer species apparently depend on immigration from other populations to allow local persistence.

Ten leopard cats (Prionailurus bengalensis borneoensis) were captured and radio tracked in an agricultural landscape in Sabah, Malaysia. Seventy-two leopard cat scats were analysed for diet while information on prey distribution and abundance was obtained from a concurrent study on small mammals. Mammals, namely murids, were the major prey with Whitehead's rat (Maxomys whiteheadi) being the principal prey species. Leopard cats significantly preferred the relatively open oil palm habitat over both selectively logged dipterocarp forest and secondary forest fragments. Although relative murid abundance was highest in selectively logged dipterocarp forest, oil palm harboured a higher relative abundance of Maxomys whiteheadi. Visibility and ease of movement for leopard cats was also better in oil palm, thereby possibly increasing their hunting success. We suggest that the significantly higher use of oil palm by leopard cats is related to their preference for areas with high prey ‘catchability’ rather than high prey density. Although secondary-forest fragments were least selected, they were important to leopard cats for resting and possibly breeding, highlighting the importance of forest fragments for the conservation of Bornean leopard cats in agricultural landscapes.

This study investigates whether increases in elephant populations may influence the structure of African savannas, and consequently may affect other herbivores through changes in habitats. Two contrasting periods in terms of elephant population densities were compared in Hwange National Park, Zimbabwe: the early 1980s and the late 1990s. Elephant population density and other ungulate population densities were estimated for a c. 400-km2 area from road counts. Vegetation structure at the landscape scale was assessed using aerial photographs for the same area. All browsers and grazers declined between the early 1980s and the late 1990s, whereas elephants experienced a 16-fold increase. At the landscape scale, vegetation structure changed little with no evidence of an opening of the habitats. These results do not support any kind of medium-term facilitation between elephants and other herbivores. They rather suggest a negative effect of elephants on other herbivore species when elephants are present at high densities. This study rules out a scenario where the decrease of the different herbivore populations was caused by large changes in vegetation structure due to elephant activity.

Behavioural sex-role reversal occurs when males and females exchange their standard roles in territorial defence or parental care. One circumstance under which sex-role reversal may occur is when males are a limiting resource, so that females have to compete for access to mates. Here we report on male rarity and male and female behaviour of species within the damselfly genus Nesobasis, endemic to Fiji. Earlier reports suggested that, in some members of this genus, males were seldom observed and that females of these species were consequentially territorial, a phenomenon described as ‘sex-role reversal’. Quantitative estimation of the ratio of adult males to females at 15 localities in 13 Nesobasis species (1489 individuals) indicated that males were extremely rare in some species, yet common in others. This interspecific variability in male rarity cannot be explained by elevation or habitat. Formal observations of three species with abundant males revealed that males of these species were highly territorial: they physically challenged intruders while remaining within a confined area. By contrast, in three species where males were consistently rare or absent, females were not territorial: instead, they moved widely and were primarily engaged in oviposition. While we do not know the underlying reason for the unusual rarity of males at oviposition sites in some species, it is clear that this rarity has not provided sufficient selection pressure to generate genuine sex-role reversal.

Consumption of terrestrial leaf litter by stream invertebrates is an important process, but little attention has been paid to the influence of leaf diversity on the process. Tropical forests are known to have much greater diversity of plant species than their temperate counterparts, but studies of litter processing in tropical streams have not explicitly addressed this issue. This paper documents the streambed leaf litter composition and diversity of an Australian tropical rain-forest stream and the leaf preferences of consumers in the stream. We hypothesized that there would be high diversity of litter in the stream and that the shredders would have broad preferences, given that litterfall of any one species would occur over a restricted period. Leaf litter was characterized by high species diversity (81 species from one stream reach sampled on two occasions). Leaf consumers (‘shredders’) were associated with a relatively broad suite of leaf species (38 species) and did not indicate clear leaf preferences. However, in a laboratory feeding experiment, using the three most common shredder species and some of the most abundant leaf species in the stream, all shredder species exhibited clear preference for a single leaf species (Endiandra bessaphila). Preference for this and other species was affected by the conditioning age of leaves (i.e. the length of time leaves were exposed to leaching and microbial colonization), with conditioned leaves usually being preferred, and previously non-selected leaves becoming more palatable with conditioning. Thus, different successional stages were more important than the identity of leaf species in determining the distribution of shredders among the leaves.

We studied the response of vegetation and vertebrate assemblages to fire and grazing, and their interacting effects, in Eucalyptus woodland in north-eastern Australia. In this vegetation type, many pastures remain free of cattle grazing due to the occurrence of a native shrub poisonous to livestock. Vegetation (floristic data and 22 habitat variables) and vertebrate fauna (birds, mammals, reptiles) were sampled in 29 standardized 50 × 50-m quadrats in the 2001 wet season, representing four treatments: sites burnt recently (within 2 y) and grazed by cattle (4–8 ha per livestock unit); sites unburnt (last burnt >2 y ago) and grazed; sites burnt recently and ungrazed; and unburnt and ungrazed sites. Fire and grazing had a significant influence on vegetation: both grazing and fire reduced ground cover (fire in grazed sites 51–23%, fire in ungrazed sites 68–39%) and increased the cover of forbs (8% in burnt and grazed sites, 3% if ungrazed) and tussock grasses (20% in grazed and unburnt sites and 5% when ungrazed). Grazing caused a shift in floristic composition from the perennial hummock grass Trioda pungens to tussock grasses (e.g. Aristida spp., Enneapogon spp.), forbs (e.g. Phyllanthus spp.) and shrubs (e.g. Acacia spp.). Of the vertebrate groups, birds responded more to fire effects (9 species), reptiles to grazing effects (6 species) and mammals to the interaction (2 species). Species reacted to increases in bare ground (e.g. crested pigeon Ocyphaps lophotes, hooded robin Melanodryas cucullatus, Ctenophorus nuchalis) and to the dominant ground cover (e.g. Ctenotus pantherinus) or change in vegetation architecture (e.g. singing honeyeater Lichenostomus virescens, variegated fairy-wren Malurus lamberti). The clearest example of an interacting effect was the cycle of complementary dominance between the rodents Pseudomys delicatulus and P. desertor, the latter's post-fire recovery becoming more muted in sites where cattle grazed (modelled time for population recovery twice as long as in ungrazed sites).

Most seedlings and saplings remain in the forest understorey for decades before becoming adults or dying, and thus may be exposed to various sources of physical disturbance (Clark & Clark 1991, 2001). Because tree species vary in their ability to recover after physical damage (Gillman et al. 2003, Guariguata 1998), this damage can act as an ecological filter and influence the juvenile community structure and species composition (Peters et al. 2004). Studies have demonstrated the occurrence and magnitude of stem breakage in juveniles caused by falling branches and other canopy debris (Clark & Clark 1989, 1991; Gillman & Ogden 2001, Scariot 2000). Surprisingly, little is known about the magnitude and ecological consequences of physical damage to juvenile plants by mammals, particularly large ungulates, including herbivory, trampling and uprooting (Gillman & Ogden 2003, Roldán & Simonetti 2001).

This study provides the first quantitative description of vertical stratification in calling heights of ensiferan species constituting most of the dry-season nocturnal acoustic community of an evergreen forest in Kudremukh National Park in south-west India. Calling heights of an average of 26 individuals of each of the 20 ensiferan species were measured and subjected to an analysis of variance to test for differences in mean calling height between species and a cluster analysis to check for the presence of discontinuous calling height layers. There were significant differences in mean calling heights between species. Calling heights of different gryllid and tettigoniid species ranged from the ground to the canopy. More gryllid than tettigoniid species occupied the ground and herb layer. Our study revealed vertical stratification of calling heights, with discrete layers corresponding to the canopy, understorey and the ground layer. These clusters emerged from the raw data of calling heights of individuals without a priori distinction of layers. We found no significant correlation between the calling heights of species and call features, including mean dominant frequency of narrow band calls, mean syllable rate and mean duty cycle.

Ants frequently interact with non-myrmecochorous diaspores on the ground in neotropical savannas. Our objective was to quantify the removal rate of these diaspores by ants and vertebrates in order to test the predator avoidance hypothesis, and to test how diaspore traits influence removal by ants and dispersal distance. We also investigated whether seed cleaning (removal of fruit matter simulating ant activity) can influence seed germination. We performed removal experiments with nine diaspore species in a reserve of cerrado savanna in south-east Brazil. Considerable differences in removal rates were found among the nine species. We found a positive linear relationship between lipid content and removal rates for five diaspore species. Vertebrate predation pressure was low for most species, limiting the benefits that ants can provide to plants to escape predators. Ants displace diaspores up to 25 m, which may increase the chance of a seed hitting a safe site. Smaller diaspores attain longer distances of dispersal than large ones. Seed cleaning increased the germination rate for five out of six species tested in greenhouse experiments. Ant activity can have relevant and possibly lasting effects on seed fate of plants adapted for vertebrate dispersal in the cerrado savanna.

Drying floodplains that were not inundated for more than 10 y were investigated to relate plant species and vegetation structure to time-since-fire for the southern Okavango Delta wetland in Botswana. The aims of the research were to study post-fire succession on savanna habitats in the Okavango Delta and to investigate the mechanisms behind vegetation changes with time-since-fire. We hypothesize that vegetation structure rather than species occurrence is affected by the fire event and that the driving forces of post-fire succession change from fire-driven to competition-driven with increasing time-since-fire. A total number of 153 study sites were investigated, representing areas of 2–12 y without a fire. Both woody and herbaceous plant species were recorded and cover/abundance values were measured. The vegetation changes with time-since-fire could be assigned to three different successional stages: (1) the post-fire conditions shortly after fire, supporting high species numbers due to open space and little competition; (2) the dominance of competitive perennial grass species (e.g. Urochloa mosambicensis) after 6–8 y; and (3) the enhanced establishment of woody species after about 10 y, creating new habitats in their vicinity. High species numbers were found after short and long fire-free periods, and different species occurred shortly after the fire event and at late stages of the post-fire succession. A variety of different inter-fire intervals is therefore expected to be responsible for sustaining plant species richness and habitat diversity in the study area. Post-fire succession could be described as not only being driven by the life history of the species, but also by seedbank dynamics and plant mobility in early successional stages and by competition and niche differentiation in late-successional stages.