The introduction of bamboo to montane rain forests of the Luquillo Mountains, Puerto Rico in the 1930s and 1940s has led to present-day bamboo monocultures in numerous riparian areas. When a non-native species invades a riparian ecosystem, in-stream detritivores can be affected. Bamboo dynamics expected to influence stream communities in the Luquillo Experimental Forest (LEF) were examined. Based on current distributions, bamboo has spread downstream at a rate of 8 m y−1. Mean growth rate of bamboo culms was 15.3 cm d−1. Leaf fall from bamboo stands exceeded that of native mixed-species forest by c. 30% over a 10-mo study. Bamboo leaves (k = −0.021), and leaves from another abundant riparian exotic, Syzygium jambos (Myrtaceae) (k = −0.018), decayed at relatively slow rates when submerged in streams in fine-mesh bags which excluded macro-invertebrate leaf processors. In a second study, with leaf processors present, bamboo decay rates remained unchanged (k = −0.021), while decay rates of S. jambos increased (k = −0.037). Elemental losses from bamboo leaves in streams were rapid, further suggesting a change in riparian zone / stream dynamics following bamboo invasion. As non-indigenous bamboos spread along Puerto Rico streams, they are likely to alter aquatic communities dependent on leaf input.

The species richness of pimpline ichneumonids was examined using catches from a network of Malaise traps operated throughout Costa Rica. More species were caught at some sites than at reasonably well-collected sites in other countries. When corrected for area, the species richness of ‘Pimplinae’ and Pimplini in Costa Rica as a whole is greater than that for virtually any other region recorded to date. The local richness of both groups appears to be a function of their regional richness. Many of the species found in Costa Rica apparently have large geographic ranges.

The species present, and the frequencies and intensities of termite attack on 600 paper baits exposed at the soil surface were compared over three seasonal exposures and between two savanna sites of contrasting vegetation structure in northern Australia. Eleven species were recorded, with Microcerotermes nervosus and Schedorhinotermes actuosus comprising 43% and 27% of collections respectively. The most commonly sampled species nest underground or build epigeal mounds and are known to feed on sound and decaying wood. Changes in species dominance occurred between seasons and the two forest types. In both vegetation types, the number of species active and the frequency of attack increased with the duration of bait exposure and decreased in the order: transitional > wet > dry. Bait consumption was greater in the site with higher canopy cover, and did not differ significantly between seasons. No direct relationships were noted between rainfall recorded at the sites and species richness, frequency and intensity of attack on baits. We recommend exposure of paper baits for at least 2 mo during the transitional period as the optimal protocol for sampling at the time of greatest activity and diversity of those species within the guild of wood-feeding species regularly attracted to paper baits.

The analysis of beta diversity (inter-habitat diversity) of very species-rich and incompletely sampled tropical arthropod communities requires the choice of appropriate statistical tools. The performance of the three commonly employed ordination methods, correspondence analysis (CA), detrended correspondence analysis (DCA), and non-metric multidimensional scaling (NMDS), was compared on a large empirical data set of geometrid moths sampled along an altitudinal gradient in an Andean montane rain forest. Despite the high species richness and incompleteness of the ensembles, all methods depicted the same, readily interpretable patterns. Both CA and NMDS showed an arch-like structure, which hints at an underlying coenocline, whereas this arch was computationally eliminated in DCA. For this particular data set, CA and NMDS both provided convincing results while the detrending algorithm of DCA did not improve the interpretability of the data. Of the large number of similarity indices available to be used in combination with NMDS, the binary Sørensen and the abundance-based Normalized Expected Species Shared (NESS) index were tested. Performance of the indices was measured by comparing stress, a measure of poorness-of-fit in NMDS. NMDS ordinations with lowest values of stress were achieved by the NESS index with the parameter m set to its maximum (mmax). In contrast, ordinations based on NESS values with the parameter m set to 1 (identical with Morisita's index), had consistently higher stress values and performed worse than ordinations using Sørensen's index. Hence, if high values of m can be achieved in similar data sets, the NESS index with mmax is recommended for ordination purposes and Morisita's index should be avoided.

Although levels of parasitism can vary greatly among individual bats of the same species, little is known about the characteristics of hosts that affect such variation. Bats were captured via mist nets from June 1995 to July 1997 from 28 localities throughout Paraguay. Over 17 500 ectoparasites were collected from 2909 bats; however, analyses of ectoparasite abundance were restricted to more abundant taxa of host and ectoparasite. We quantified the abundances of 29 taxa of ectoparasite on 19 species of bat host, as well as total abundance of ectoparasites regardless of taxonomic affiliation for 22 species of bat from Paraguay. The effects of host sex and host body size on these estimates of ectoparasite abundance were evaluated separately for each species of host. Ectoparasites did not respond consistently to host body size: ectoparasite abundance increased with host body size in 12 instances and decreased with host body size in 11 instances. Regardless of the existence or direction of effects of host body size on ectoparasite abundance, female hosts generally harboured more ectoparasites than did male hosts. Differences in host quality associated with the sex of bats, especially those related to behaviour, may be a more important determinant of ectoparasite abundance than are differences in size. Opportunities for host transfer are critical for species persistence of ectoparasites; consequently, ectoparasite populations on host individuals that form social groups or colonies should be larger, less prone to stochastic extinction, and have greater opportunity for speciation.

With 86 species, Iwokrama Forest in central Guyana has the highest reported bat biodiversity for a protected area in the world. Using standardized capture data for 73 of these 86 species, we document community structure of bats in terms of species diversity, relative abundance, gross biomass, feeding guilds, vertical stratification and a trophic-size niche matrix. Based on faunal surveys in 1997, with similar amounts of effort in the forest canopy and at ground level, the greater fruit-eating bat (Artibeus lituratus) was by far the most ecologically dominant species in terms of frequency of capture and biomass. In total, frugivores comprised 70% of the species diversity and 78% of the biomass. The most common species of bat were fully partitioned in a resource niche matrix of size and trophic guild when vertical stratification was included as a variable. We conclude that resource partitioning and species packing differentially affect relative size in tropical bats, and are better summarized and analysed in three dimensions.

The composition of anuran assemblages was studied in 16 streams at the Serra do Cipó, south-eastern Brazil, in which 26 anuran species were found. Volume and isolation level of stream sections of 150 m were estimated; tadpoles, adult anurans, and potential tadpole predators were searched for over 16 consecutive months. Stream size, isolation, richness and diversity of tadpole predators, and some additional physical and biotic variables were measured and related to anuran species richness. Smaller streams tended to shelter more anuran species than larger ones, but neither stream size nor stream isolation explained variations in anuran species richness. Anuran species richness tended to increase with tadpole predator diversity. Tadpole species richness was influenced by predator species richness, and was highest when there were seven types of tadpole predators present. The distribution of few anuran species could be clearly related to particular physical and biotic variables. The behavioural flexibility of anurans and the similarity among streams may have contributed to this pattern.

Evidence from southern African semi-arid savannas suggests that high elephant densities in nature reserves cause the over-utilization of woodlands. Northern Botswana, with its prolific elephant population, is expected to follow such a pattern unless the densities of elephants which could be sustained by indigenous woodlands are determined and maintained at carrying capacity. Above ground biomass production was estimated for mopane (Colophospermum mopane) woodlands, a principal food source for elephants, which grows over much of northern Botswana. Densities of trees and shrubs, dimensions of plants and elephant densities were recorded within stratified plots situated according to a regional rainfall gradient. Mean estimates of above ground biomass (foliage and twigs) were 9.41 and 7.83 t ha−1 for shrubs and trees respectively. A logistic model described the maximum levels of biomass removal from plants before over-utilization of mopane occurred. Variables incorporated in the model included above ground biomass of mopane shrubs and trees, growth rates of plants and expected off-take by elephants. The model predicted a complete biomass regain within 10 y if no elephant browsing occurs. Intensive elephant browsing in woodlands containing 15 elephants km−2, however, can suppress biomass production if growth rates of plants fall below 70% of the maximum annual rate. Nonetheless, there was no substantial evidence to suggest that elephants will reduce the biomass of mopane woodlands in northern Botswana below a sustainable level if their numbers are allowed to increase considerably beyond the current estimate.

The sampling of epiphytes is fraught with methodological difficulties. We present a protocol to sample and analyse vascular epiphyte richness and abundance in forests of different structure (SVERA). Epiphyte abundance is estimated as biomass by recording the number of plant components in a range of size cohorts. Epiphyte species biomass is estimated on 35 sample-trees, evenly distributed over six trunk diameter-size cohorts (10 trees with dbh > 30 cm). Tree height, dbh and number of forks (diameter > 5 cm) yield a dimensionless estimate of the size of the tree. Epiphyte dry weight and species richness between forests is compared with ANCOVA that controls for tree size. SChao1 is used as an estimate of the total number of species at the sites. The relative dependence of the distribution of the epiphyte communities on environmental and spatial variables may be assessed using multivariate analysis and Mantel test. In a case study, we compared epiphyte vegetation of six Mexican oak forests and one Colombian oak forest at similar elevation. We found a strongly significant positive correlation between tree size and epiphyte richness or biomass at all sites. In forests with a higher diversity of host trees, more trees must be sampled. Epiphyte biomass at the Colombian site was lower than in any of the Mexican sites; without correction for tree size no significant differences in terms of epiphyte biomass could be detected. The occurrence of spatial dependence, at both the landscape level and at the tree level, shows that the inclusion of spatial descriptors in SVERA is justified.

The effect of canopy species dominance on understorey light availability in secondary forests was examined for Pentaclethra macroloba and Goethalsia meiantha, two common tree species with contrasting canopy architecture at La Selva Biological Station in the Atlantic lowlands of Costa Rica. In each of six 12- to 16-year-old successional stands dominated by one of these species, relative abundance and basal area of the focal species were determined within a 20 m × 50 m plot. Light availability at 1 m height was measured within each plot using a Licor-2000 Plant Canopy Analyzer at 27 locations and by analysing hemispherical photographs taken at 10 locations. Across stands, mean LAI values were significantly negatively correlated with canopy openness. Understorey microsites beneath Pentaclethra had significantly higher light availability than microsites beneath Goethalsia. Across stands, however, light availability was not correlated with either total basal area or stem density of trees ≥10 cm DBH. These results indicate that basal area and stem density of trees in the upper levels of the forest canopy are poor predictors of light penetration to the understorey. One hypothesis suggested here is that dense subcanopy and understorey vegetation in tropical secondary forests can strongly influence understorey light levels. Alternatively, light transmittance characteristics of the upper forest canopy may be influenced by species-specific differences in canopy architecture or foliage density that vary independently from tree diameter and density.

Bat communities in the understorey of primary lowland dipterocarp forest in peninsular Malaysia and Sabah were sampled with mist nets and harp traps to estimate their trophic structure. Overall, 4 of 26 species (15%) at one site, and 6 of 33 species (18%) at the other site were frugivores, while the remainder were insectivores. In terms of individuals, 7–12% of all captures were frugivores. However, the true proportion of frugivores at both sites was probably closer to 1%, because the sampling procedures were strongly biased towards frugivores. In contrast, previously published studies indicate that frugivores comprise 34–48% of species, and up to 80% of individuals in bat communities in the understorey of Neotropical forest. These results support the hypothesis, developed from studies of bird communities, that the under-storey of Malaysian dipterocarp forest supports relatively few obligate frugivores, at least in non-masting years. However, further data are required on bats in the canopy, and on the foraging ranges of bats to compare the absolute densities of bats in Malaysian and Neotropical forests.

The potential effect of forest type on the structuring of animal communities in western Amazonia remains poorly understood. In this study, we tested the hypothesis that amphibian species richness, composition and abundance differ across forest types in the lowland rain forest of south-eastern Peru. By using 320 individual transects, we compared the amphibian assemblages across four major forest types (floodplain, terra firme, bamboo and palm swamp) at each of four sites separated by 3.5–105 km. We identified 1967 individuals of 65 species in 11 families and found that a large proportion of the amphibian diversity in this region is attributed to habitat-related beta diversity. Overall, we found that forest type is more important than site in predicting both species composition and abundance. We also found that, when analyses are conducted separately for each forest type and include species abundance data, similarity between assemblages decreases with increasing geographic distance. In contrast to studies that considered species presence/absence but ignored species abundances, our results highlight the importance of including abundance data in the assessment of animal diversity patterns in western Amazonia. We conclude that evaluating community structure across forest types can improve our understanding of diversity patterns in this region.

Sympatric lizards in a transitional forest of Roraima, Brazil, dividing Amazon rain forest from savanna, contained a lizard assemblage of 16 species representing six families. Lizards varied in size, activity (diurnal versus nocturnal), microhabitats used, exposure to various conditions of light availability, prey types, and prey sizes. Overlaps in microhabitat occurrence varied from 0 (no overlap) to 1.0 (total overlap) whereas overlaps in prey types varied from 0.005 to 0.607. Microhabitat overlaps were higher overall than dietary overlaps. Pseudocommunity analyses on microhabitat data indicate that the community is not randomly assembled and that two distinct guilds exist, a leaf-litter guild and an arboreal guild, each with four species. Similar analyses on diet data revealed no apparent guild structure ast the first rank (nearest neighbour). Lizard diets did not differ from a random assortment based on prey type. At lower levels, the assemblage was structured with respect to food. Variation in prey use among lizard species was tied more closely to the effect of lizard body size on prey size (lizards ate different-sized prey items). Although exploitative competition among species may maintain structure within this assemblage it does not necessarily cause the observed differences. Several species are nearly identical ecologically to sister taxa in other environments and within different lizard assemblages suggesting that composition of the local assemblage limits the species that can enter the assemblage. Finally, lack of structure at the lowest (most similar) neighbour ranks may reflect the impact of a transitional habitat on stability of species interactions at the local level.

Studies of predation on propagules of the mangroves Avicennia marina, Bruguiera exaristata, Ceriops tagal and Rhizophora stylosa were made in a forest in northern Australia to test the generality of the dominance-predation model. This model states that an inverse relationship exists between the dominance of a species in the canopy of mangrove forests and the rate of predation on the propagules of that species. Significant differences in predation were found among the four species, and among patches of forest dominated by the different species. Predators attacked more than 50% of the propagules of all species except R. stylosa, so are likely to significantly affect forest structure. The intensity of predation did not, however, vary as the dominance-predation model predicted. Instead, predation on the propagules of a species appeared to depend on the availability of propagules of other, more highly preferred, species.

The insular distribution and abundance of the spiny rat (Proechimys semispinosus) were studied by live-trapping rats on 50 small forested islands in Gatun Lake, Republic of Panama. Rats were found on 29 of these islands. Rat abundance was not correlated with island size or isolation, in contrast to temperate rodents, and may be due to the highly patchy nature of tropical forests and the differing floristic compositions of the study islands. Rat occurrence on islands was related positively to island size and negatively to island isolation. All islands greater than 2.1 ha contained spiny rats, whereas most islands less than 0.9 ha did not contain rats. The relationship between occurrence, size, and isolation are illustrated by a three-dimensional response surface. Results suggest that colonization and extinction regularly occur on these islands and that (1) larger islands with year-round fruit production, regardless of isolation, will have persistent populations of rats, (2) small isolated islands will never or only very rarely have rats owing to a lack of a year-round food source and the difficulty of immigration, and (3) small near islands frequently will have rats when fruits arc present. No other species of mouse-or rat-like rodents were captured on the islands. We suggest that P. semispinosus is a better colonist than other sympatric forest species, is more of a habitat generalist, and has a more flexible demographic structure allowing persistence in heterogeneous and changing environments.

Habitat alteration can have far-reaching consequences for natural communities and can alter existing species interactions in profound ways. Working in a tropical dry forest ecosystem in the Guanacaste Province of Costa Rica, we measured seed removals and seed dispersal distances for Acacia collinsii and papaya seeds to determine if habitat disturbance affects ant community composition and the associated interactions between ants and seeds. Two hundred and forty experimental seed depots were observed in four sites that differed in land-use history and disturbance severity: secondary forest, forest edge, Acacia collinsii edge, and open-pasture. Both seed removals and average dispersal distances achieved by ants differed among habitats. Habitat disturbance did not negatively affect seed removals by ants in our study; seed removals were highest in the most disturbed habitat, the open pasture (38.3% of seeds removed within 2 h). Ant community composition and the relative abundance of key seed dispersers also differed among habitats. In all sites combined, ten species of ant were observed dispersing seeds, with two species, Pheidole fallax and Ectatomma ruidum, in combination being responsible for 92% of all observed seed removals. The abundance of P. fallax and E. ruidum differed among habitats, with E. ruidum being the most abundant ant species collected in the open-pasture habitat and P. fallax being the most abundant in the edge habitats. Pheidole fallax ants dispersed seeds significantly further (mean ± SD = 2.11 ± 1.35 m, maximum = 5.2 m) than did E. ruidum (mean ± SD = 0.7 ± 0.81 m, maximum = 3.85 m), a difference we attribute to recruitment behaviour and not body size. Habitat disturbance thus alters the ant community and the relative abundance of key seed-dispersing ant species, with cascading effects on seed removals and seed dispersal distances.

Plant responses to herbivory vary depending on herbivory type, yet the comparative effects of defoliation (e.g. by insects) and pruning (e.g. by large mammals) on a single tree species are poorly documented. We investigated this in the Northern Province of South Africa by comparing the regrowth of Colophospermum mopane trees previously defoliated by caterpillars or pruned by elephants, the two main browsers of C. mopane foliage. Shoots were up to 160% and 125% longer after natural (elephant) and simulated pruning and leaves ~25% longer in regrowth after natural pruning (n = 13–15 trees per treatment). Shoot density and chemical defences in leaves (tannin:protein ratio and total polyphenolic concentration) were, however, no different from control trees. Simulated defoliation resulted in statistically insignificant changes to regrowth in terms of leaf and shoot size (both slightly decreased) and shoot density (slightly increased). Natural (caterpillar) defoliation, however, resulted in regrowth with significantly decreased shoot and leaf size (about 50% and 20% of control lengths, respectively), as well as decreased leaf chemical defence. Shoot and leaf length were longer on trees flushing for the first time after pruning and late-season defoliation had a greater negative impact than mid-season defoliation. Despite the differences in regrowth characteristics after pruning and defoliation, mopane plants showed no apparent trade-off in investment between tolerance and resistance after either herbivory type, as neither regrowth nor chemical defence occurred at the expense of the other.

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.

Extrafloral nectar is an indirect, generally ant-mediated, defence mechanism that is particularly common in tropical plants. This study focuses on interactions among different groups of arthropods visiting extrafloral nectaries of the South-East Asian myrmecophilic plant, Macaranga tanarius. The diurnal activity patterns of arthropods on nectaries were recorded on two occasions, each plant being used once in an untreated state and once after chemical induction of extrafloral nectar flow. Ants, widely regarded as the most important consumers of extrafloral nectar, made up only 60% of all nectary-visiting arthropods. Striking negative relations became obvious between ants and ‘non-ants’, among which two (morpho)species of fly (Grammicomya sp. and Mimegralla sp., both Micropezidae) were most important. On most leaves, either ants or ‘non-ants’, but not both groups, were present at one time. Behavioural observations revealed that many flies actively excluded other arthropods, including ants, from the nectaries or leaves. However, the flies had no detectable defensive effect against herbivores. The presence of these ‘thieves’ of extrafloral nectar therefore can cause ecological costs in terms of reduced presence of ants that consume extrafloral nectar and defend plants.

Soil microbial biomass (Cmic.) and is relation with soil carbon and nutrient contents in a tropical rain forest (Bawangling Nature Reserve, Hainan Island, P.R. China) were studied. Cmic. was determined by the substrate induced respiration (Sir) method; bacterial and fungal fractions were measured using a selective inhibition technique. For the A horizon, Cmic ranged between 350 and 700μgg-1 soil in April (beginning of the rainy season) and 250–600μg g-1 in October (end of the rainy season). Cmic. contents of the AB and B horizons were lower, ranging from 80 to 350 μg Cmic g-1 soil. On average, 75% of the biomass was bacterial, in the AB and B horizon this was slightly higher than in the A horizon. Cmic was closely correlated with Corg and N content. N and P contents as well as the respiratory response upon glucose addition indicated that towards the end of the rainy season compared with the start the soils were depleted in nutrients. These observations stress the importance of microbial biomass as a factor for nutrient retention in tropical forest ecosystems.