Journal of Tropical Ecology

Published by Cambridge University Press (CUP)
Online ISSN: 1469-7831
Print ISSN: 0266-4674
Because of its position on the northern margin of the tropics (22° 17′N) and the southern coast of a huge continent, Hong Kong has a climate in which both temperature and rainfall are highly seasonal. Although summer temperatures are equatorial, the January mean is only 15.8 °C,and the absolute minimum recorded at sea level is 0 °C (Dudgeon & Corlett 1994). As a result, all aspects of the ecology of Hong Kong show seasonal changes. The most dramatic changes occur in the bird fauna, with the majority of species migratory (Carey et al. 2001). The winter fruiting peak in secondary shrublands and the forest understorey coincides with the arrival of partially frugivorous migrant robins and thrushes (Corlett 1993). However, while resident insectivore-frugivores consume almost entirely fruit during this period (Corlett 1998), all the winter visitors continue to eat insects and some (e.g. Phylloscopus warblers) are entirely insectivorous. The study of insect seasonality reported here formed part of a 30-mo study of the seasonality of a forest bird community in Hong Kong (Kwok & Corlett 1999, 2000). Plant names follow Corlett et al. (2000).
Waza-Logone area and the position of the transect where the study took place. 
Differentiating soil characteristics of sampled sites in Waza-Logone, Cameroon.
Maximum flood depth and flood duration in the transect Zina-Zwang-Tchikam. For position, see also Figure 1. Flood situation: f: flooded (before and after 1994), r: reflooded from 1994 onwards, d: dry (before and after 1994).
Correlations between maximum flood depth, above-ground and below-ground biomass and intermediate soil parameters in Waza-Logone, Cameroon.
Relation between maximum flood depth and total above-ground biomass in the seasonally flooded grasslands of the Sudd, Sudan (MefitBabtie 1983), Kafue Flats, Zambia (Ellenbroek 1987) and Inner Niger Delta, Mali (Hiernaux & Diarra 1984a). Biomass measurements are based on 10 (Sudd), an unknown (Kafue Flats) and 24 (Inner Niger Delta) samples of 1-m 2 collected at the end of the flooding season.
Flood depth has been frequently used to explain the distribution of plant species in seasonally flooded grasslands, but its relation with vegetation production has remained ambiguous. The relationship between flooding and above-ground biomass at the end of the flooding season and during the dry season was studied to assess the impact of reflooding on the Logone floodplain, Cameroon. Above-ground biomass of a combination of all species and of the individual perennial grasses Oryza longistaminata and Echinochloa pyramidalis showed a positive linear relationship with maximum flood depth up to 1 m. The gradient of these relationships became steeper and their fit better during the 2 y following the installation of the flooding, showing the response lag to floodplain rehabilitation. Flood duration only explained the above-ground biomass of the combination of all species and not of the individual species. Above-ground biomass data from other floodplains in the three main African geographic regions showed a similar relationship with maximum flood depth less than 1 m. Dry-season regrowth, important because of its high nutrient quality during forage scarcity, was not directly related to maximum flood depth, possibly because of its dependency on the period of burning and soil moisture. Presented data indicate that a rise of water level of 1 cm corresponds to an increase in above-ground biomass of c. 150 kg DM ha[minus sign]1.
Details of the five vertical-transect sites. All were within Brisbane Forest Park (see text).
Understanding the pattern in which adult drosophilids of different species are distributed across and within different vegetation types is necessary for accurate interpretation of their local ecology and diversity. Such studies have been conducted mainly in temperate regions, and there is no basis for extrapolating their conclusions to tropical areas. This study describes the vertical distribution (0-20 m) of drosophilids attracted to banana baits in five different vegetation types in subtropical eastern Australia including open woodland, and rain-forest types. The distribution of most of the 15 common species could be characterized three-dimensionally by vegetation type and height above forest floor. Only one species, Scaptodrosophila lativittata, was common in all vegetation types and it was a canopy species in rain forests and a ground-level species in open woodland. Vertical distribution of some species clearly matched that of their larval hosts, but it did not in others. For example, the fungivore Leucophenga scutellata was mostly trapped well above the forest floor, yet it breeds at ground level, suggesting behavioural mode can influence vertical distributions. We conclude that the vertical dimension, although still poorly understood in relation to drosophilid habitats, needs to be taken into account when conducting and interpreting studies aimed at understanding drosophilid populations and communities in the subtropics.
Schematic representation of four types of competition treatments between tree and liana Byttneria grandifolia seedlings. RSC, root and shoot competition, RC, root competition, SC, shoot competition, NC, no competition. 
Relative growth rate (mean ± SE, n = 8–10) in seedlings of three species in response to a competing liana Byttneria grandifolia under different light and competition treatments. Significant differences (at P < 0.05) among the competition treatments in the same light level are indicated by different letters. RGR B , relative growth rate in biomass, RGR H , relative growth rate in height, RGR D , relative growth rate in 
The light-saturated photosynthetic rate in leaves ( P n ), 
Abstract: In tropical forests, trees compete not only with other trees, but also with lianas, which may limit tree growth and regeneration. Liana effects may depend on the availability of above- and below-ground resources and differ between tree species. We conducted a shade house experiment to test the effect of light (4% and 35% full sun, using neutral-density screen) on the competitive interactions between seedlings of one liana (Byttneria grandifolia) and three tree species (two shade-tolerant trees, Litsea dilleniifolia and Pometia tomentosa, and one light-demanding tree, Bauhinia variegata) and to evaluate the contribution of both above- and below-ground competition. Trees were grown in four competition treatments with the liana: no competition, root competition, shoot competition and root and shoot competition. Light strongly affected leaf photosynthetic capacity (light-saturated photosynthetic rate, Pn), growth and most morphological traits of the tree species. Liana-induced competition resulted in reduced Pn, total leaf areas and relative growth rates (RGR) of the three tree species. The relative importance of above- and below-ground competition differed between the two light levels. In low light, RGR of the three tree species was reduced more strongly by shoot competition (23.1¿28.7% reduction) than by root competition (5.3¿26.4%). In high light, in contrast, root competition rather than shoot competition greatly reduced RGR. Liana competition affected most morphological traits (except for specific leaf area and leaf area ratio of Litsea and Pometia), and differentially altered patterns of biomass allocation in the tree seedlings. These findings suggest that competition from liana seedlings can greatly suppress growth in tree seedlings of both light-demanding and shade-tolerant species and those effects differ with competition type (below- and above-ground) and with irradiance
Species diversity of butterflies in primary and selectively logged forest during wet and dry seasons in Sabah, Borneo. See text for rainfall during the two seasons. 
Indices of temporal clumping and wet-season preference for butterflies in primary forest in Sabah, Borneo. 
We used traps baited with fruit to examine how the temporal variation of butterflies within primary forest in Sabah, Borneo differed between species. In addition, we compared patterns of temporal variation in primary and selectively logged forest, and we tested the hypothesis that selective logging has different recorded impacts on species diversity of adults during the wet monsoon period and the drier remaining half of the year. Species of Satyrinae and Morphinae had significantly less-restricted flight periods than did species of Nymphalinae and Charaxinae, which were sampled mainly during the drier season, especially in primary forest. Species diversity of adults was significantly higher during the drier season in primary forest, but did not differ between seasons in logged forest. As a consequence, logging had opposite recorded impacts on diversity during wetter and drier seasons: primary forest had significantly higher diversity than logged forest during the drier season but significantly lower diversity than logged forest during the wetter monsoon season. The results of this study have important implications for the assessment of biodiversity in tropical rain forests, particularly in relation to habitat disturbance: short-term assessments that do not take account of seasonal variation in abundance are likely to produce misleading results, even in regions where the seasonal variation in rainfall is not that great.
Representative actograms displaying activity over 48-h period for a subadult male ocelot (a) and an adult male agouti (b) from mid-May to the end of August 2003. Data were collected from animals on Barro Colorado Island, Panama. A 48-h period is shown to better display both diurnal and nocturnal activity. The grey areas represent periods of missing data which are caused by several factors such as: animals venturing outside of the range of the system, the system being shut down for maintenance or the system’s capacity being devoted to other species or system testing. Note that the ocelot is generally active at night with occasional activity bouts during day-time, whereas the agouti shows the opposite activity pattern. 
An animal's fitness is in part based on its ability to manage the inherent risks (foraging costs, predation, exposure to disease) with the benefits (resource gain, access to mates, social interactions) of activity (Abrams 1991, Altizer et al. 2003, Lima & Bednekoff 1999, Rubenstein & Hohmann 1989, Wikelski et al. 2001). Thus, understanding an animal's pattern of activity is key to understanding behavioural and ecological processes. However, while numerous laboratory methodologies are available to continuously quantify activity over long periods of time, logistical difficulties have greatly hindered activity studies of animals in the field (DeCoursey 1990).
Germination of Acacia tortilis seeds strongly depends on micro-site conditions. In Lake Manyara National Park, Tanzania, Acacia tortilis occurs abundantly in recently abandoned arable fields and in elephant-mediated gaps in acacia woodland, but does not regenerate in grass swards or beneath canopies. We examined the germination of Acacia tortilis using field and laboratory experiments. Seeds placed on top of the soil rarely germinated, while seeds covered with elephant dung or buried under the soil surface had a germination success between 23¿43%. On bare soil 39% of both the dung-covered and buried seeds germinated, in perennial grass swards 24¿43%, and under tree canopies 10¿24% respectively. In laboratory experiments, seed water absorption correlated positively with temperature up to 41 °C, while subsequent germination was optimal at lower (21¿23 °C) temperatures. Seeds that had absorbed water lost their viability when kept above 35.5 °C. The absence of light did not significantly influence germination success. Acacia tortilis does not actively disperse its seeds, but regeneration outside tree canopies was substantial. The regeneration potential thus strongly depends on the physiognomy of the vegetation
A comparison is made of the light acclimation potential of seedlings of three canopy species of the tropical rain forest of Los Tuxtlas, Mexico: Cordia megalantha, Lonchocarpus guatemalensis, and Omphalea oleifera. These species showed similar growth rates in a range of microhabitats. Gap dynamics were simulated by transferring plants between three environments: beneath a closed canopy, a small gap, and a large gap. Plants of all three species were able to adjust their morphology and growth rates in response to changes in light availability. Growth rates increased when plants were moved to a (larger) gap, and decreased when plants were moved to a more shaded environment. Shade-grown plants were able to acclimate faster to increasing light availability than sun-grown plants to decreasing light availability. Also, plants moved from shady to sunny conditions showed higher relative growth rates than sun control plants, whereas sun-grown plants when moved to the shade showed lower relative growth rates than shade control plants. Species differed in their response to gap dynamics. Omphalea could not acclimate morphologically to shading, but reacted faster than the other species in response to the occurrence of a large gap. Acclimation potential seemed to be related to plasticity in physiological rather than in morphological traits. Suppressed seedlings of all three species performed well in the shade, and were able to acclimate rapidly to gap-conditions.
We examined the seasonal and inter-annual variation of leaf-level photosynthetic characteristics of three C4 perennial species, Cyperus articulatus, Panicum repens and Imperata cylindrica, and their response to environmental variables, to determine comparative physiological responses of plants representing particular microhabitats within a seasonal tropical floodplain in the Okavango River Delta, Botswana. Five measurement campaigns were carried out over a period of 2 y which covered two early rainy seasons, two late rainy seasons and one dry season. For all three species, light-saturated net photosynthetic rates (Asat) and stomatal conductance (gsat) decreased with decreasing soil water content with a seasonal range for Asat of approximately 5¿45 ¿mol m¿2 s¿1, and for gsat of 0.03¿0.35 mol m¿2 s¿1. The species representing the wettest microhabitat (Cyperus) had the highest gsat at low leaf-to-air vapour pressure deficits (Dl), the highest ratio of intercellular to ambient CO2 concentration (Ci/Ca), as well as the highest degree of variation in Ci/Ca from season to season. We interpret this as being indicative of its adaptation to a moist growth environment allowing for non-conservative water use strategies as soil moisture is usually abundant. For all three species there was significant variation in photosynthetic fluxes from one year to another that was related to variation in leaf nitrogen and phosphorus. This study shows that when assessing the role of savanna stands in large-scale carbon balance models, the remarkable inter-annual variation in leaf photosynthesis reported in this study should be taken into account
Map of the study area in the Okavango Delta, Botswana (adapted from Ellery & McCarthy 1994) and approximate location of the study site. The 'impala study site' is provided in more detail in Figure 2.  
Study site, April (a) and August (b) Landsat image (black and white version of Landsat band combination 4, 5, and 3) of the impala study site in the Okavango Delta, Botswana, including transect locations.
Width of the observation zones (m), determined by Distance 3.5, for each transect (Bushcamp, Croc, Lion and Mopane) and for each physiognomic vegetation type (grass on floodplain, woodland, and grass on dryland), based on impala observations.
Selection of vegetation subclasses by impala for the April and August 2000 periods, with Bonferroni simultaneous probability intervals.
Classification accuracy assessment. Overall, producer and user accuracy of the number of pixels correctly classified as landcover types (i.e. physiognomic vegetation types and open water) through a multi-temporal classification of the study area in the Okavango Delta, Botswana.
We studied the spatial and temporal habitat use of impala in Botswana's Okavango Delta at landscape level with the aid of satellite imagery, with minimal fieldwork. We related remotely sensed vegetation to impala habitat preferences, by first distinguishing three vegetation types through a multi-temporal classification, and dividing these into subclasses on the basis of their Normalized Difference Vegetation Index (NDVI). This indicator for abundance and greenness of biomass was assessed for wet and dry season separately. Similarly, habitat use was assessed for both seasons by allocating vegetation classes to bimonthly impala observations. Impala distribution patterns coincided with NDVI-based subclasses of the landscape, nested within broad vegetation types, to which impala did not show a marked seasonal response. We suggest that this methodology, using limited field data, offers a functional habitat classification for sedentary herbivores, which appears particularly valuable for application in extensive areas with high spatial variability, but with restricted access.
Changes in density, structure and species composition of the woody component of the two predominant savanna types in Tarangire National Park, northern Tanzania, during a period of 25 y were investigated. The park is known for its large, increasing elephant numbers and high frequency of fires. In 1996 a study on woody species density, composition and age structure, which was first performed in 1971, was repeated, using the same transects and method. Access to the original data of 1971 allowed for a full comparison of the changes and an investigation whether these could be related to changes in elephant numbers and fire frequency. The total tree density declined during the 25 y, but the decline was not evenly distributed over the different height classes. Although the density of trees taller than 5 m declined significantly, the greatest decline occurred in the density of trees shorter than 1 m. The density of trees in the intermediate height class of 1–5 m did not decline. Although damage to trees by elephants increased during the 25-y period, c. 25␜howed no browse damage and, except for some severely damaged trees, elephant damage was not found to reduce tree vigour. Elephants affected the size distribution of the savanna woody component much more than the density, while the data suggest no significant effect of fire on changes in tree density. The large decline in density of small trees was attributed to a severe drought in 1993. Based on large numbers of elephants during the past decades and on relatively low elephant impact on the total tree density, the present study suggests that the current elephant number of 2300 can be sustained in the park without causing detrimental effects, provided that their current range is maintained.
Design of the exclosure experiment in the Hluhluwe-iMfolozi Park, South Africa. Herbivore species of different size classes are stepwise permanently excluded from 40 m × 40-m blocks of savanna vegetation by using fences with different height and mesh width. Exclosure treatments include (from left to right) unfenced control, rhino fence, zebra fence, impala fence and hare fence. Herbivore species that are able to feed within the different exclosure treatments are listed below each. Animal pictures are copyright of O. Bonnet and A.M. Shrader. 
Frequency of occurrence of dominant grass species for the different exclosure treatments in (a) Hluhluwe and (b) iMfolozi measured in March 2003 in the Hluhluwe-iMfolozi Park, South Africa. The grass species composition was significantly different between the exclosure treatments in both Hluhluwe (n = 200, χ 2 16 = 432, 
Mean vegetation height for the different exclosure treatments in (a) Hluhluwe and (b) iMfolozi measured in March 2003 in the Hluhluwe- iMfolozi Park, South Africa. Error bars represent 1 SE. Different upper-case letters show significant differences in vegetation height between exclosure treatments (n = 200, one-way ANOVA, F 2, 3 = 53.4, P = 0.005). 
Total number of murid rodents captured in the different exclosure treatments from July 2002 to December 2004 in the Hluhluwe- iMfolozi Park, South Africa. The different exclosure treatments are presented as: solid diamonds = hare fence; open squares = rhino fence; solid triangles = unfenced control. The dashed lines indicate a hypothesized trend in the number of murid rodents due to a missing trapping session. Murid rodent numbers were significantly higher in the absence of all larger herbivores throughout the course of the study (n = 9, repeated-measures ANOVA, F 16, 48 = 2.1, P = 0.03). 
Total number of murid rodent species captured in the different exclosure treatments from July 2002 to December 2004 in the Hluhluwe- iMfolozi Park, South Africa. The exclusion of large herbivores resulted in a higher number of rodent species in the hare fence. In addition, the composition of the rodent species assemblage differed significantly between the exclosure treatments (n = 9, three-way ANOVA, F 10, 36 = 3.1, P = 0.006). 
Our study presents experimentally based results on how large herbivore species affect savanna vegetation and thus murid rodents in the Hluhluwe-iMfolozi Park in KwaZulu-Natal, South Africa. We permanently excluded groups of large herbivore guilds of various body sizes (ranging from white rhino to hares) from sixteen 40 x 40-m plots of vegetation by using different fence types. We determined grass species composition and vegetation height and collected capture-mark-recapture data on murid rodents. Nutrient concentrations of the dominant grass species and rodent diet compositions were analysed. We found that herbivore species of different body sizes had different effects on murid rodents. The exclusion of medium-sized herbivores, such as warthog, impala and nyala increased the abundance of high-quality grass species, especially Panicum maximum. However, the dominant rodent species Lemniscomys rosalia preferred the most abundant grass species, rather than high-quality grasses. The absence of large bulk feeders, such as zebra, buffalo and white rhino led to an increase in vegetation height. In response. tall vegetation promoted both rodent abundance and species diversity and altered rodent species composition. Ultimately, our results indicate that the greatest effect on murid rodents came from the reduction of vegetation cover by large bulk feeders, which likely increased rodent predation risk.
Per capita net growth rate of the small-herbivore population (solid line) and the large-herbivore population (dashed line) as a function of vegetation density. The growth rate of the large herbivore is zero at V 1 *, the growth rate of the small herbivore is zero at V 2 * and V T . 
Current hypotheses to explain dynamic transitions between savanna grasslands and woodlands in Africa focus on grazing by elephant or the influence of fire. Using a simple mathematical model, this paper argues that interactions between small herbivores such as impala or buffalo and large herbivores such as elephant or giraffe may provide a plausible alternative hypothesis. The interplay of competition and facilitation between these types of herbivores could explain transitions between grassland and woodland and vice versa. A review of the literat- ure is presented in support of this hypothesis.
Patches of wooded vegetation in nutrient-poor grassland characteristically contain high soil moisture and nutrient availability compared with surrounding grassland. These `islands of fertility¿ appear stable in size, suggesting that tree recruitment at the patch boundary is limited. We hypothesize that tree establishment in adjacent grassland is limited by (1) competition for resources, (2) fire or (3) herbivory. In a South African grassland, we measured moisture availability and conducted a bioassay experiment to analyse whether soil nutrient limits tree recruitment at the patch boundary. We thus measured nutrient concentrations of maize plants grown in patch, boundary and grassland soil. To investigate whether browsing or fire affected tree seedlings at the patch boundary, we burned plots including patches, and used fences to exclude browsers. Neither soil moisture nor nutrient availability at the boundary differed from inside the patches, suggesting that tree recruitment at the boundary is not resource limited. Both fire and browsing combined caused a significantly lower tree seedling growth at the patch boundary, suggesting that these two factors can impede tree recruitment. The balance between positive feedback mechanisms facilitating tree recruitment, and the negative impact of fire and browsing can explain the apparent stability of these islands of fertility.
Nutrient fluxes in rainfall and throughfall of different tropical montane rain-forest ecosystems. For details on forest types, elevation, annual rainfall amounts and soil types, see footnotes.
Volume weighted mean nutrient concentrations in the Oi litter layer and in fine litterfall collected in a montane rain forest at Mt. Kilimanjaro and annual nutrient fluxes via fine litterfall (± SE, n = 3).
In contrast to their well-studied counterparts in the Neotropics and in Asia, East African montane rain forests are surrounded by semi-arid savanna plains. These plains have a high erosion potential for salt crusts accumulated at the soil surface. Hence it may be hypothesized that East African montane forest ecosystems experience strongly enhanced nutrient inputs via dry deposition, which alters their overall biogeochemistry. The aim of our study was to test this hypothesis by investigating K, Mg, Ca, Na and N-forms in rainfall, throughfall, fine litter, litter percolate and soil solution of a montane rain forest at Mt. Kilimanjaro. Four forest plots situated at elevations between 2250 and 23 50 in asl on the south-western slopes of Mt. Kilimanjaro were studied for 2 y. In contradiction to our hypothesis, inputs of K, Mg, Ca and Na via rainfall (7.5, 0.9, 2.3 and 6.2 kg ha(-1) y(-1)) and throughfall (3 5, 2.0, 3.5 and 11 kg ha(-1) y(-1)) were low on Mt. Kilimanjaro. Fluxes of NH4-N and NO3-N were within the range observed at other montane rain forests. with NO3-N being the only nutrient partly absorbed in the forest canopies (2.9 kg ha(-1) y(-1) in rainfall, 0.9 kg ha(-1) y(-1) in throughfall). The highest overall nutrient concentrations in water samples occurred in litter percolate (1.4 mg l(-1) K, 0.3 mg l(-1) Mg, 0.8 mg l(-1) Ca, 0.3 mg l(-J) NH4-N, 0.9 mg l(-1) NO3-N), with values still being low compared to other sites. Nutrient concentrations in seepage water strongly declined with increasing soil depth. Thus, both inputs and losses of base cations from the forest by water pathways are assumed to be low. N or P limitation of growth is not expected since high fluxes of N and P in fine litter (119 and 5.9 kg ha(-1) y(-1) for N and P respectively) indicate low within-stand efficiency.
Home range (100% Minimum Convex Polygon) of the agouti ( Dasyprocta punctata ) on Barro Colorado Island, Panama, Central America. 
This study investigates the movements and home range of the agouti (Dasyprocta punctata) on Barro Colorado Island, Panama. We captured and tracked 12 agoutis from January to December 2003. Home-range size (95% kernel) ranged from 1.56–2.45 ha (n = 6) for males and 1.34–1.97 ha (n = 5) for females. Agouti ranges overlapped and we estimated a density of approximately 100 agoutis km−2. We compared agouti movement with the locations of refuges and food trees, and the results suggest that the agoutis are central-place foragers. Agoutis moved an average of 850 m d−1 covering approximately 35% of their range daily. These movement data help us understand the potential impacts of agoutis as seed dispersers, predicting that D. punctata will encounter and hoard fallen fruit within 10–200 m (i.e. radius of home range) of its source, and move seeds towards refuges such as ground holes and dense vegetation around recent tree falls.
The agouti (Dasyprocta punctata) meticulously defleshes Astrocaryum standleyanum palm seeds before scatter hoarding. On Barro Colorado Island, Panama, we experimentally tested three hypotheses on how this behaviour could reduce seed predation to the mutual benefit of the tree and the rodent. The first and established hypothesis – that defleshing reduces seed predation by bruchid beetles by intercepting larvae – was rejected. Experiments in which manually defleshed seeds or entire fruits were incubated at different times showed that defleshing reduced bruchid infestation before fruit fall but not after fruit fall. The second hypothesis – that defleshing reduces cache pilferage by making seeds less conspicuous – was supported. An experiment in which intact fruits and manually defleshed seeds were placed in mimicked agouti caches and followed showed that seeds with flesh were pilfered at higher rates than defleshed seeds. The third hypothesis – that defleshing reduces post-dispersal infestation of cached seeds – was rejected. An experiment in which intact fruits and manually defleshed seeds were placed in mammal exclosures and later collected to assess infestation showed that burial reduced seed infestation but defleshing did not. Thus, seed defleshing reduced palm seed predation, but in a different way than previously believed. We also found that (1) bruchid beetles can be pre-dispersal rather than post-dispersal seed predators, (2) seed infestation by scolytid beetles may control bruchid larvae, and (3) scolytids rather than bruchids are the main invertebrate seed predators of this palm
Ordination by multidimensional scaling of 70 Amazonian terra firme (TF) plots on the basis of their similarity in numbers of individuals in 16 major plant families. Similarity was calculated with correlation coefficients. Symbols: v, central Amazonia, „, eastern Amazonia; , Guiana Shield; , western Amazonia and Rondônia.
Location of Amazonian terra firme (TF) plots used in the analysis of tree alpha-diversity. Dot size indicates the value of Fisher's alpha (in case of overlap only the highest may be visible). Grey-scales indicate interpolated values of Fisher's alpha (with interdistance weighting up to 500 km).  
The relationship between annual rainfall (mm y −1 ) and Fisher's alpha for (a) all plots in western, central, eastern Amazonia and the Guiana Shield, (b) plots in western Amazonia and Rondônia, (c) central Amazonia, (d) eastern Amazonia, (e) Guiana Shield. Symbols: , terra firme; „, floodplain; v, podzol; , swamp; b, dry forest.
A large number of newly published and unpublished hectare plots in Amazonia and the Guiana Shield area allow an analysis of family composition and testing of hypotheses concerning alpha-diversity in the south American rain forest. Using data from 94 plots the family-level floristic patterns in wet tropical South America are described. To test diversity patterns, 268 plots are used in this large area. Contrary to a widely held belief, western Amazonian plots are not necessarily the most diverse. Several central Amazonian plots have equal or even higher tree diversity. Annual rainfall is not a good estimator for tree diversity in the Amazonia area and Guiana shield. Plots in the Guiana Shield area (and eastern Amazonia) usually have lower diversity than those in central or western Amazonia. It is argued that this is not because of low rainfall or low nutrient status of the soil but because of the small area of the relatively isolated rain forest area in eastern Amazonia and the Guiana Shield. The low diversity on nutrient-poor white sand soils in the
In the Central Amazonian floodplains, several hundred tree species grow in areas that are periodically flooded by nutrient-rich while-water rivers (varzea) and by nutrient poor black-water rivers (igapo). Seed masses of 31 species from varzea and 27 species from igapo were compared taking into consideration their taxonomic relatedness. Overall average seed mass was higher (mean = 7.08 g) in nutrient-poor igapo than in nutrient-rich varzea (mean = 1.16 g). In igapo, the species growing at high elevations on the flooding gradient had significantly higher seed masses than the species growing at low elevations. In varzea, no difference was found between species growing at high and low elevations. Four large-seeded species from igapo occurring at high elevations on the flooding gradient were responsible for most of the difference in average seed mass between forest types. These data suggest that at low positions in the flooding gradient in igapo, selection pressure on seed size is probably the same as in varzea. At sites with short periods of flooding in igapo forests, on high levels in the flooding gradient, the need for rapid height growth may have selected for species with larger seeds which enable seedlings to be less dependent on soil nutrients.
In central Amazonian white-water floodplains (varzea), different forest types become established in relation to the flood-level gradient. The formations are characterized by typical patterns of species composition, and their architecture results in different light reflectance patterns, which can be detected by Landsat TM image data. Ground checking comprised a detailed forest inventory of 4 ha, with Digital Elevation Models (DEM) being generated for all sites. The results indicate that, at the average flood level of 3 m, species diversity and architecture of the forests changes, thus justifying the classification into tile categories of low varzea (varzea baixa) and high varzea (varzea alta). In a first step to scale up, the study sites were observed by aerial photography. Tree heights, crown sizes, the projected crown area coverage and the gap frequencies provide information, which confirms a remotely sensed classification into three different forest types. The structure of low varzea depends on the successional stage, and species diversity increases with increasing age of the formations. In high varzea, only one successional stage was found and species diversity is higher than in all low-varzea formations. The more complex architecture of the high-varzea forest results in a more diffuse behaviour pattern in pixel distribution, when scanned by TM image data.
Study site, Ilha da Marchanteria. 
Investigated tree species classified into different ecotypes listed by scientific names, family, number of individuals and Importance-Value Indices (IVI) after Klinge et al. (1996) for a stand in the same study area. Species with dendrometer bands (d) were studied for a period of 26 mo between June 1998 and August 2000, those without for 17 mo from April 1999 until August 2000. 
To study the impact of the annual long-term flooding (flood- pulse) on seasonal tree development in Amazonian floodplains, the phenology and growth in stem diameter of various tree species with different leaf-change patterns were observed over a period of 2 y. The trees of the functional ecotypes, evergreen, brevi-deciduous, deciduous and stem-succulent showed a periodic behaviour mainly triggered by the flood-pulse. Trees have high increment during the terrestrial phase. Flooding causes a shedding of some or all leaves leading to a cambial dormancy of about 2 mo and the formation of an annual ring. Studies carried out in tropical dry forests verify a strong relationship between the phenological development and the water status of the trees, strongly affected by seasonal drought. The comparison of the phenology and the diameter growth of the corresponding ecotypes in floodplain forest and a semi- deciduous forest in Venezuela shows a displacement of at least 2 mo in the periodicity, except for stem-succulent tree species. For stem-succulent trees it remains unclear which factors influence phenology and stem diameter growth.
The nature of the relationship between amblyopinine staphylinid beetles and the mammals upon which they are found has been an enigma since the group was first described in 1875 (Solsky 1875). The Amblyopinini include some forty species placed in five genera, and have been found almost exclusively in the für of small mammals in the Neotropics and Australia (Seevers 1955). This habit differs significantly from that of other members of the large and diverse family Staphylinidae, most of which are free-living predators. As a result of this, and reports of epidermal damage to the host, amblyopinines have been accepted to be obligate blood-feeding ectoparasités (Seevers 1955, Marshall 1981, Kim & Adler 1985). However, many reported behavioural facets appear to be inconsistent with this conclusion. For example, it has been reported that hosts ignore both the presence and movements of these beetles through their für and across highly sensitive areas such as the eyes and vibrissae (P. Hershkovitz, quoted in Seevers 1955). This suggests that the interaction between amblyopinines and their hosts cannot be simply interpreted as adversarial.
Striga asiatica (Scrophulariaceae) is an obligate root hemiparasite of mainly C-4 grasses (including cereals). It is the most widespread of the 42 Striga species occurring in many semi-tropical, semi-arid regions of mainly the Old World. Examination of herbaria specimens revealed that S. asiatica has a wider geographical distribution, is present at higher altitudes and occurs in a more diverse range of habitats than previously reported. The host range is also larger than previously reported and is likely to include a large number of C-3 plants. Morphology of examined specimens revealed variation in size and corolla colour suggesting the existence of ecotypes. Climate may exert a significant influence on the distribution of S. asiatica given the diversity of potential host plants and their distribution beyond the current recorded range of S. asiatica.
The fruits of Cola lizae, an endemic tree with a limited geographical distribution, have been a major food source for lowland gorillas in the Lope Reserve during part of each year over a six-year period. Faecal analysis indicated that 11,000- 18,000 Cola seeds km-2 were deposited by gorillas during the 4- month season in 1989. Gorillas are the only important dispersers of this species: other primates consume the succulent mesocarp, but do not swallow the large seed; elephants do not eat Cola fruits. Observations of Cola seeds in gorilla faeces showed a very high germination rate and, despite initial high mortality, 18% of seedlings still survived six months after deposition. Survival of seedlings was significantly better in faeces left at nest-sites than in other areas of the forest: 40% of seedlings were viable at nest- sites six months after deposition. This suggests that the open areas of forest, preferred by gorillas as nest-sites, are advantageous to the propagation of this species.
Cumulative maps of all Atta study colonies. Lines represent foraging trails and dots the corresponding leaf sources. The first row shows young colonies with (E) indicating nest entrances: (a) Y-1, (b) Y-2, (c) Y-3. The second and third row show mature colonies with (N) defining nest areas: (d) M-1, (e) M-2 (Wirth et al . 2003), (f) M-4 (Vasconcelos 1990), and (g) M-3 (Farji-Brener & Sierra 1993). The upper scale refers to the young and the lower to mature colonies. 
Assessment of temporal changes in the geometrical distribution of exploited leaf sources by calculating mean angles for each observation day. Data are presented as box-and-whisker plots, which indicate the overall distribution of the data by showing the first and third quartile (lower and upper margin of the box), the median (line within the box) and the 5% as well as the 95% percentile. Colonies of the two age classes young (left side) and mature (right side) are shown. Mean angles were calculated on a daily (A-3), 2-3 d (A-1 and Y-1, 2 and 3) or monthly scale (A-2).
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 trailmaps 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
The ant species were sampled in one campo (grassy shrubland), one mata (semideciduous rain forest) and 11 rehabilitated bauxite mine plots at Poços de Caldas, in the humid sub-tropical climatic region of Brazil. Rehabilitation was either by planting Australian Eucalyptus spp, the Brazilian Mimosa scabrella tree, or by planting mixed mata trees. Sixty-eight ant species were recorded, of which 26 were exclusively found in the native vegetation and 16 were confined to the rehabilitated plots. Ant species richness built up most rapidly in areas rehabilitated with mixed mata species and least rapidly in areas with Eucalyptus. Younger rehabilitated plots appeared to be developing a campo-type ant community, although evidence indicates that more mata ant species will colonise once tree canopy closure takes place. Rate of ant return in Australia is positively correlated with the quantity and distribution of rainfall – the rates in the current study concur with those from humid sub-tropical climatic zones within Australia, suggesting that similar constraints to the succession may be operating.
Ant species were sampled in three rainforest and 10 rehabilitated bauxite mine plots at Trombetas, in the tropical monsoonal region of Brazil. Rehabilitation ranged from 0 to II years in age and was mainly performed by planting mixed native forest tree species. One plot supported single-species blocks of Australian Eucalyptus and Acacia species. Two hundred and six ant species were recorded, of which 82 were exclusively found in the native vegetation, 54 were confined to the rehabilitation and 70 were found in both situations. In contrast with other studies, ant species richness in the Eucalyptus/Acacia plantation was as great as in the areas rehabilitated with native vegetation. The overall rate of return of ant species was considerably greater than in mines situated within subtropical regions of Brazil, Africa and Australia. However, if the greater richness of ants in the native vegetation at Trombetas was accounted for, the proportional return of the original ant fauna was not particularly rapid. The return of ant species slowed as the rehabilitated areas aged. In comparison with forest, the rehabilitation was characterized by proportionately more generalist species and fewer specialists, especially from the soil and litter layers. The full range of habitat requirements for the ant community has not been restored by the eleventh year of rehabilitation and further management may be required to enhance the degree of colonization. It is suggested that the findings for ants may apply to other components of the biota as well.
Ants recruited to a termite bait placed with a needle on an internode of M. bancana. White arrow: stipule (food bodies are produced on the inner surface). Right panel: Longitudinal cut, showing the domatium inside the twig, scale insects used by the ants as additional food source, and the food bodies produced under the stipule (white arrow).
Relative contribution of leaves at different positions to a plant's net carbon gain (in per cent of total CO 2 fixation per 24 h). Gas exchange was measured for six M. bancana plants in the vicinity of the Ulu Gombak Field Studies Centre in Selangor, Malaysia, in 1996. We used a CO 2 /H 2 O porometer system (CQP 130, Walz, Effeltrich, Germany) in an open-flow mode to quantify net CO 2 fixation (µmol CO 2 m −2 s −1 ) repeatedly (once per leaf and per hour during day time) for the three youngest mature and every other of the older leaves (see Zotz, 2001, for a detailed description of methods). Assimilation rates of the non-measured of the older leaves were estimated as the average of the two adjacent ones, and a whole plant's net assimilation per 24 h was calculated from these data and the areas of the respective leaves (data from Heil 1998).
The optimal defence hypothesis predicts that plant parts characterized by a high value and/or a high risk of being attacked should exhibit the highest level of defence. We tested this hypothesis with Macaranga bancana ant-plants, which are protected efficiently by resident, mutualistic ants from herbivores, parasites and encroaching vegetation. Because cost-effective defence of the host by ants increases ant fitness, selection should act on ant behaviour to produce patterns of distribution of defence as predicted for direct chemical defence traits. Termites and pieces of tape were equally distributed over the uppermost ten leaves and over the leaf-bearing part of the stems (with termites mimicking a transient herbivore, while tape mimics a long-term stress caused by a climber or plant parasite). This arrangement allowed a separation of putative coevolutionary adaptations in the ants' behaviour from other potential sources of spatial patterns in ant defence, such as differences in herbivore pressure, in the vulnerability of different herbivores, or in direct plant defences. Ant activity dropped rapidly at termite baits, but remained high at tapes for at least 5 h, thereby demonstrating adaptive differential responses to the differences between the two stressors. Most importantly, ants preferentially defended young leaves and shoot parts. The temporal and spatial patterns of ant attention to exogenous stressors thus were clearly adaptive, varying with likely costs and benefits of defence as formulated in the optimal defence theory.
Data (mean ? SD) collected from 21 tents constructed by Artibeus watsoni and 152 uncut leaves of Anthurium ravenii (measure- ments in cm)
Artibeus watsoni in Costa Rica alter the shape of 11 species of broad-leafed plants to form diurnal roost sites. The plant most commonly used for tent construction was Anthurium ravenii. Bats create a tent on A. ravenii by severing the basal 2-5 lateral nerves at 5-10 mm from the midrib; the entire margin of the leaf then collapses downward to form a pyramid-shaped tent. The number of altered leaves per plant ranges from 1-4 (mean 2.1). A. watsoni appears to select leaves of medium size and low within the plant. The size, shape, and location of leaves selected is consistent with the hypothesis that tent construction provides bats with roost sites that will be available to them for a long period of time and provides protection from both predators and the elements. -from Authors
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(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.
Mean total P content of O. alata ( v ) and T. moreliana ( ࡯ ) in relation to P application treatments at harvest three. The replicates of 0, 3, 15 mg kg − 1 rock P applied are combined within a ‘6 mg kg − 1 ’ rock P 
Root fraction and specific root length of O. alata and T. moreliana at harvests one and three.
The relationship between mycorrhizal colonisation and phosphorus acquired by seedlings of the arbuscular mycorrhizal tree Oubanguia alata Bak f. (Scytopetalaceae) and the ectomycorrhizal tree Tetraberlinia moreliana Aubr. (Caesalpiniodeae) was evaluated at low and high inorganic phosphorus availability. AM colonisation was positively correlated with phosphorus uptake by O. alata at low, but not at high phosphorus availability. Seedlings growth was positively related to arbuscular mycorrhizal colonisation at both low and high phosphorus availability, suggesting that growth promotion by arbuscular mycorrhizas is not simply related to an increase of phosphorus uptake. In contrast, phosphorus uptake by T. moreliana was correlated with EM colonisation at both low and high phosphorus availability, but there was no relationship between growth and ectomycorrhizal colonisation. Promotion of phosphorus uptake by arbuscular mycorrhizas and ectomycorrhizas at low phosphorus availability is consistent with the co-occurrence of the two types of mycorrhiza in tropical rain forests where available soil phosphorus is low. However, ectomycorrhizal colonisation may also be of advantage where inputs of phosphorus rich litter raise the phosphorus status of the soil, as seen in the groves of ectomycorrhizal trees in Korup National Park, and may be one of the factors reinforcing local dominance by these trees.
Mycorrhizae, symbioses between plant roots and fungi, are found in virtually all terrestrial plant ecosystems. Substantial evidence indicates that mycorrhizal associations are beneficial to both partners. Journal Article
Measures of the amount and quality of PFD for the six light treatments (see Ashton 1995 for details). 
This study investigated the effects of light and soil fertility, on arbuscular mycorrhizal fungi (AMF) colonization, and the growth responses (height and dry mass) of Syzygium seedlings. Seedlings of four Syzygium spp. were grown for 2 y in six different light treatments at the research station of the Sinharaja Forest, Sri Lanka. The light treatments exposed seedlings to: (1) 3%; (2) 16%; (3) 50%; (4) 100% of full sun (control); (5) short periods (2 h d−1) of direct sunlight; and (6) long periods (6 h d−1) of direct sunlight. In the 16% of full sun treatment five sets of fertilizer applications supplied: (1) magnesium; (2) potassium; (3) phosphorus; (4) all three nutrients combined; and (5) no fertilizer (control). The Syzygium species had the greatest mycorrhizal colonization in brighter treatments that provided direct light. Comparison across species revealed S. firmum to have moderate mycorrhizal colonization but high total dry mass. Syzygium operculatum had high percentages of mycorrhizal colonization while S. rubicundum had low percentages of mycorrhizal colonization especially in deep shade. Syzygium makul showed moderate levels of mycorrhizal colonization and dry mass, but low height growth. Among fertilizer applications, phosphorus enhanced seedling growth and mycorrhizal colonization for all species. However, species showed decreased growth with high amounts of potassium and combined fertilizer applications. Results suggest that AMF colonization will be highest, and Syzygium spp. growth greatest, beneath canopy openings large enough to receive direct sun in phosphorus-rich soils.
A total of 200 variable-area plots covering 20 3 ha containing 4000 trees with a basal area of 921 4 m2 were assessed over a combined elevational range of 410-2180 m in the moist forests of three Tanzanian Eastern Arc mountains: West Usambara, Nguru and Udzungwa. Plot data were ordinated on the basis of species presence/absence, frequency and basal area Axis 1 of ordinations based on species presence/absence are correlated with elevation Axis 1 of the frequency-weighted ordination was correlated with elevation in the Nguru and Udzungwa mountains, but plots from the West Usambara showed a rainfall-related discontinuity Axis 1 of the West Usambara basal area-weighted ordination showed evidence of long-term dynamics of Ocotea usambarensis and in the Udzungwa mountains was determined by presence of Parinari excelsa Plot diversity was not correlated with elevation or latitude, but was lower in disturbed, low rainfall or more seasonal forest. Stem density was positively correlated with elevation and was greater on ridge tops than valley sides and valley bottoms
There are two trade-offs at the levels of leaves and crowns, i.e. assimilation capacity per leaf mass is greater for shorter-lived leaves, and unbranched species grow faster in height by allocating carbon more to trunk than to leaves and branches compared with highly branched species. The hypotheses were tested that the degree of branching (LTB) correlates with leaf traits and that height growth rate is negatively correlated with the degree of branching and leaf life span (LLS) by examining saplings of five canopy and subcanopy species, two shrub species and one invasive subshrub species (Clidemia hirta) in a tropical rain forest, West Java, Indonesia. Of the eight species, the most and least branched species were Castanopsis acuminatissima and Macaranga semiglobosa, respectively. Leaf traits examined were leaf size, LLS, leaf mass per area (LMA), leaf nitrogen concentration per mass (Nmass) and per area. LLS tended to be positively correlated with LMA, and negatively correlated with Nmass. Leaf size was negatively correlated with LTB, but the other leaf traits were not correlated with LTB. The height growth of the eight species was low, irrespective of LTB and LLS, for understorey individuals. The height growth of gap individuals was negatively correlated with LLS for the eight species, and also negatively with LTB for the seven species other than one subshrub species. Thus, the degree of branching was correlated with leaf size only among the five leaf traits, and both leaf life span and the degree of branching affected the height growth of gap individuals, except for the subshrub species.
Cordia elaeagnoides is locally important within a large area of southern México as a highly gregarious canopy tree of dry forest and as a timber species. Its recruitment is problematic so studies of its reproduction and population dynamics were undertaken. The distylous flowers are apparently self-compatible, and pollination and seed dispersal are by wind. Pre-dispersal mortality is largely from embryo abortion, and post-dispersal predation is mostly by generalist rodents. Some seeds may survive in enforced dormancy for two or more years. Recruitment was apparently absent for more than a decade, but had previously occurred within established stands, where size classes were interspersed at random. Population size structure suggests that recruitment over the last century occurred in several pulses. Growth rates determined from ring counts and remeasurement of marked trees support a size-age relationship. The youngest reproductive trees are estimated to be about 18 years old, and about 5% of the present population is over 95 years old.
Prevalence (per cent infected) of nematode orders and cestodes in the different host species. The numbers of faecal samples are given in parentheses.
Chao2 species richness estimate and Simpson's 0 diversity indices based on nematode morphotypes found in hosts captured in unlogged (UP) and logged (LF) forest. Calculations are based on a randomized order of the minimum standardized number of samples per forest type (given in parentheses). Niviventer cremoriventer is excluded from this analysis because of small sample size
Parasites are ubiquitous in wild animals, with host-specific life histories considered as major determinants of prevalence and parasite assemblage patterns. It is predicted that habitat differences in logged rain forests influence population performances of small mammals and consequently may change the infection patterns of local animal populations with regard to endo- and ectoparasites. We investigated patterns of helminth species assemblages (Nematoda, Platyhelminthes) in two rat species (Leopoldamys sabanus, Niviventer cremoriventer) and two tree shrew species (Tupaia tana, T. longipes) in three logged and three unlogged rain forests in Borneo by examining 337 faecal samples with non-invasive faecal egg count (FEC). Nematode eggs prevailed in 95% of all samples with up to five (mean 1.9 ± 1.1) morphotypes. Whereas members of Strongylida were most prevalent in L. sabanus, T. tana and T. longipes, Spirurida dominated in N. cremoriventer that revealed at the same time the lowest average nematode prevalence and FEC. Cestode eggs were only found in L. sabanus and T. tana. Composition and abundance patterns of the parasitic helminth assemblages were influenced by logging. As hypothesized, species richness of nematode morphotypes and mean number of infections per host of T. longipes were larger in logged than in unlogged forest. In contrast, L. sabanus was more heavily infected with cestodes in unlogged than in logged forest and also revealed larger egg counts for strongylids and spirurids in unlogged forest. Our results suggest that forest degradation and altered environmental conditions influence helminth diversity and infection patterns of small mammals with contrasting trends among host species. The inconsistent logging-induced changes in helminth assemblages from different hosts indicate that specific sets of habitat-host-parasite interactions are uniquely influenced by the effects of logging. Consequently, predictions on changes of parasite diversity and prevalence with regard to habitat disturbance need to be based on the individual life histories of the hosts (and the parasites).
Monthly distribution of the relative abundance of fig fruit fall events at Reserva Ducke, Amazonia, observed throughout the study period of July 1991–August 1996. 
Number of fruit falls for different Ficus species studied at the Reserva Ducke, Amazonia, with their abbreviations as used in Figure 2.
Numbers of individuals of carabid species collected at 61 fig fruit falls at Reserva Ducke, Amazonia. The genus is not known for four of the species. Unknown spp. were not identified but probably represent some of the predatory species listed.
Differences in abundance of eight Notiobia species at the times of different fruit fall events at Reserva Ducke, Amazonia, for (a) one F. subapiculata tree (no. 8), and (b) three F. guianensis trees. (a)
The carabid beetle assemblage found feeding on fig fruit falls at night was studied in a terra firme rain forest near Manaus (Amazonia) from July 1991 to August 1996. A total of 8926 carabid beetles were collected on 64 fruit falls from 10 fig species. The most abundant genus was Notiobia with eight species, N. pseudolimbipennis being the most abundant. The Notiobia species comprised 92% of all specimens collected and all feed on small fig seeds. Their species abundance patterns varied considerably between individual fruit falls and during the course of a single fruit fall. However, the species abundance patterns for all Notiobia at all observed fruit falls for each of the two commonest fig species (Ficus subapiculata, F. guianensis), as well as for fruit falls of the remaining fig species, were very similar. Through feeding and breeding experiments and observations of reproductive success by dissection of females, only two of the eight Notiobia species were found to be specialized fig seed feeders, being able to reproduce only on fig fruit falls. The remaining six species of this genus use fig fruit falls as alternate hosts or 'stepping stones' between fruit falls of their host trees, which are widely separated both in time and space
Habit comparisons in neotropical moist and wet forest lowland florulas for sites > 1 km 2 .
Relationship between sample size (number of independent 0.1-ha floristic samples in Madre de Dios) and the ability to detect a significant relationship between tree alpha-diversity and soil conditions. 0.1-ha samples were randomly subsampled 100 times to determine the range of P-values for the best-fit simple or multiple linear regression equations between tree alpha diversity and soil PCA factors 1 to 4. Solid line represents the moving-average of the median values; vertical lines indicate the upper and lower 95% confidence intervals of the mean for the given sample size; dotted line represents the point at which P (accept H 0 ) = 0.05.
Protocols compared in terms of inventory efficiencies, under different assumptions about the relative importance of shrubs and trees in the target flora. See text for details. All values are expressed × 10 2 ; comparisons are with Kruskall-Wallis non-parametric tests.
Site soil PCA scores.
Habitat indicator tree species (following Dufrene & Legendre 1997) revealed by 0.1-ha and 1-ha inventory protocols standardized for field effort and target flora. The matrix shows the number of self-supporting species 10 cm dbh that are significant habitat indicators.
The tropical flora remains chronically understudied and the lack of floristic understanding hampers ecological research and its application for large-scale conservation planning. Given scarce resources and the scale of the challenge there is a need to maximize the efficiency of both sampling strategies and sampling units, yet there is little information on the relative efficiency of different approaches to floristic assessment in tropical forests. This paper is the first attempt to address this gap. We repeatedly sampled forests in two regions of Amazonia using the two most widely used plot-based protocols of floristic sampling, and compared their performance in terms of the quantity of floristic knowledge and ecological insight gained scaled to the field effort required. Specifically, the methods are assessed first in terms of the number of person-days required to complete each sample (‘effort’), secondly by the total gain in the quantity of floristic information that each unit of effort provides (‘crude inventory efficiency’), and thirdly in terms of the floristic information gained as a proportion of the target species pool (‘proportional inventory efficiency’). Finally, we compare the methods in terms of their efficiency in identifying different ecological patterns within the data (‘ecological efficiency’) while controlling for effort. There are large and consistent differences in the performance of the two methods. The disparity is maintained even after accounting for regional and site-level variation in forest species richness, tree density and the number of field assistants. We interpret our results in the context of selecting the appropriate method for particular research purposes.
We tested the hypothesis that tree species in a subtropical rain forest in south-east Queensland are ecologically equivalent and therefore have identical environmental requirements for their regeneration. We assessed the evidence that juveniles of species differed in their distributions in treefall gap microsites and along gradients of light availability, soil pH, soil PO4-P availability and soil NO3-N availability. Pairwise comparisons were made on a subset of the common species selected on the basis that they showed a relatively high level of positive association, and would therefore, a priori, be expected to have similar regeneration requirements. Detailed comparisons between the species failed to demonstrate evidence for species differentiation with respect to their tolerance of the disturbance associated with gap microsites or to the gradient of NO3-N availability. However, species differed markedly in their distributions along the soil pH gradient and along the gradients of light availability and soil PO4-P availability. The overall level of ecological differentiation between the species is high: seven out of the 10 possible species pairings showed evidence for ecological differentiation. Such niche differentiation amongst the juveniles of tree species may play an important role in maintaining the species richness of rain-forest communities.
All stems greater than or equal to 1 cm dbh were measured, tagged, mapped and identified on a 1-ha plot of rain forest at Gambubal State Forest, south-east Queensland, Australia. The spatial patterns and size class distributions of 11 common tree species on the plot were assessed to search for mechanisms determining their distribution and abundance. The forest was species-poor in comparison to many lowland tropical forests and the common species are therefore present at relatively high densities. Despite this, only limited evidence was found for the operation of density-dependent processes at Gambubal. Daphnandra micrantha saplings were clumped towards randomly spaced adults, indicating a shift of distribution over time caused by differential mortality of saplings in these adult associated clumps. Ordination of the species composition in 25-m x 25-m subplots revealed vegetation gradients at that scale, which corresponded to slope across the plot. Adult basal area was dominated by a few large individuals of Sloanea woollsii but the comparative size class distributions and replacement probabilities of the 11 common species suggest that the forest will undergo a transition to a more mixed composition if current conditions persist. The current cohort of large S. woollsii individuals probably established after a large-scale disturbance event and the forest has not attained an equilibrium species composition.
The factors that influence food choice have implications for animal survival, reproduction and population growth. We conducted a 1-y study of food choice by four mountain gorilla groups that consumed herbs and fruit at two locations differing spatially and temporally in food availability in Bwindi Impenetrable National Park, Uganda. We collected data on 45 important foods consumed by the gorillas, the availability of those foods in each gorilla group's home range and their corresponding nutrient and phenolic concentrations. Employing a linear multiple regression, we tested three hypotheses regarding the influence of food availability and the nutritional and phenolic concentrations of food on food choice. Regardless of changes in herb availability, the choice of herbs was positively influenced by their abundance and sugar concentrations and negatively influenced by their fibre, condensed tannin and protein concentrations. Furthermore, regardless of changes in fruit availability, the choice of fruit was positively influenced by its abundance and negatively influenced by its condensed tannin concentrations. During periods of low fruit availability, the gorillas did not increase the consumption of herbs high in fibre and sugar. The choice of herbs low in fibre had less of an influence on food choice at the location with lower fruit availability than the other location. Our results underscore the importance of incorporating both availability and nutrient concentrations into studies of food choice; by doing so we found Bwindi gorillas were able to choose abundant, relatively high-quality foods year round.
Nutrient inputs into tank bromeliads were studied in relation to growth and productivity, and the abundance, diversity and biomass of their animal inhabitants, in three forest types along an elevational gradient. Concentrations of phosphorus, potassium and calcium in canopy-derived debris, and nitrogen and phosphorus in phytotelm water, declined with increasing elevation. Dwarf forest bromeliads contained the smallest amounts of debris/plant and lowest concentrations of nutrients in plant tissue. Their leaf turnover rate and productivity were highest and, because of high plant density, they comprised 12.8and contained 3.3 t ha -1 of water. Annual nutrient budgets indicated that these microcosms were nutrient-abundant and accumulated < 5dwarf forest, where accumulation was c. 25biomass/plant peaked in the intermediate elevation forest, and were positively correlated with the debris content/bromeliad across all forest types. Animal species richness showed a significant mid-elevational peak, whereas abundance was independent of species richness and debris quantities, and declined with elevation as forest net primary productivity declined. The unimodal pattern of species richness was not correlated with nutrient concentrations, and relationships among faunal abundance, species richness, nutrient inputs and environment are too complex to warrant simple generalizations about nutrient resources and diversity, even in apparently simple microhabitats.
We studied the effects of savanna fires on the structure of local ungulate communities in a West African woodland savanna. The distribution of 11 ungulate species over 9¿15 burned sites (the number of which increased as burning activity continued during the dry season) and 7¿13 unburned sites was compared with a variety of null models or randomized `virtual communities¿. Five different parameters of community structure were examined: body mass distribution, co-occurrence patterns, species richness, species density and guild dominance. Overall, ungulate species were not randomly distributed over burned and unburned sites. The regular spacing of body masses in the set of species recorded on burned and unburned sites indicated competition, since species similar in body mass are more likely to compete than species of different size. However, co-occurrence patterns on burned sites were random, indicating absence of competition at fine spatial scales due to differential habitat use within the burned landscape. Although the attractiveness of the regrowth on burned sites resulted in higher numbers of ungulates compared with unburned sites, species richness was not different. Grazers were the dominant guild on burned sites, but there were no differences in species richness or species density between grazers and browsers on unburned sites.
Climate diagram of Riberalta with annual course of rainfall and temperature, based on a 25-y average. The perhumid period (>100 mm mo − 1 , black), relative humid period (<100 mm mo − 1 , hatched) 
Annual course of rainfall (filled bars) and soil moisture content for gap (open circles) and understorey (filled circles) sites in ‘El Tigre’. One-sided error bars represent 1 SE. Data are for 1997. 
Root length (a) and rooting depth (b) of understorey (open bars) and gap (hatched bars) and plants of Cedrela odorata , Schizolobium amazonicum , and Brosimum lactescens . Error bars represent 1 SE. For plant numbers see the legend to Figure 4. 
In tropical moist forests, length of the dry period may have a profound influence on leaf dynamics, plant growth and survival. To evaluate the role of light and water availability on seedling performance, a 1-y experiment was carried out in a tropical moist forest in the Bolivian Amazon in which seedlings of three tree species (Brosimumlactescens, Cedrelaodorata and Schizolobiumamazonicum) were planted in gaps and the understorey. Variation in length of the dry period was simulated by subjecting part of the seedlings to a water treatment at the end of the dry period. Gaps and understorey had a similar soil moisture content, which varied between 39% in the wet season and 16% in the dry season. Height and leaf growth rates were higher in gap compared to understorey plants, and in the wet compared to the dry season. A high growth during the wet season provided gap plants with a decisive size advantage over understorey plants during the dry season. Their larger root system allowed gap plants to explore a larger surface area and deeper soil layers for water. Consequently, gap plants of Cedrela experienced a shorter deciduous period (22 d) compared to understorey plants (61 d). Watering at the end of the dry season cued the flushing of new leaves by Cedrela, although it did not lead to a higher plant growth.
Seasonal variation in precipitation, soil matric potentials and midday leaf water potentials. The graph shows the precipitation in the moist (a) and dry forest (b). The bars represent the sum of the precipitation of March 2007 and the months covered by this study (April–October 2007) (data from the nearby towns of Ascención de Guarayos (a) and Concepción (b)). The central graphs give the seasonal and topographical variation in soil matric potentials (logarithmic scale) in the moist (c) and dry forest (d) respectively, with different topographical positions represented by the different lines. Valleys are represented by the continuous black lines, slopes by the small dotted lines and crests by the large dotted lines. The lower graphs give the seasonal variation in the mean midday leaf water potential (logarithmic scale) of three moist forest (e) and three dry forest species (f). Amp.r = Ampelocera ruizii ; Swe.f = Sweetia fruticosa ; Tre.m = Trema micrantha ; Aco.c = Acosmium cardenasii ; and Sol.r = Solanum riparium . Midday leaf water potentials could not be measured for Swe.f in August as the species sheds its leaves in the dry season. Whiskers give the standard error for every month. 
Differences in soil matric potentials within forests along a topographical gradient. The graph describes soil matric potentials (logarithmic scale) in the moist (a) and dry forest (b) at different topographical positions; valleys, slopes and crests. The median (black horizontal bar), interquartile range (upper and lower limits of the boxes; 75 and 25 percentile), and the total variation in soil matric potentials are given. Within forests, topographical positions with different letters vary significantly ( α = 0.05) (Tukey test); ∗∗∗ P < 0.001. 
Changes in soil matric potential with soil depth in the moist (a) and dry forest (b) spanning the dry season (April, early dry season; August, mid-dry season; and October, first month of the wet season). 
Differences in plant water availability among species within a tropical moist and dry forest. The graph describes predawn leaf water potentials (logarithmic scale) in the moist (a) and dry forest (b) of five different species at the onset of the dry season (April; hatched boxes) and halfway through the dry season (July; white boxes). Amp.r = Ampelocera ruizii ; Swe.f = Sweetia fruticosa ; Tre.m = Trema micrantha ; Aco.c = Acosmium cardenasii ; and Sol.r = Solanum riparium . The median (black horizontal bar), interquartile range (upper and lower limits of the boxes; 75 and 25 percentile), and the total variation in predawn leaf water potentials are given. Within forests and within months, species with different letters vary significantly (ANOVA with post hoc Tukey test); ∗∗ P < 0.01, ∗∗∗ P < 0.001. 
Among-species differences in midday leaf water potentials (at a standardized predawn leaf water potential of − 0.98 MPa). The graph describes midday leaf water potentials (logarithmic scale) of three moist- (a) and dry-forest tree species (b). Among species, different letters indicate significant differences (P < 0.001) as resulted from an F-test ( α = 0.05), based on independent linear pairwise comparisons of standardized species means (with a Bonferroni adjustment for multiple comparisons). 
We determined seasonal variation in soil matric potentials (ψsoil) along a topographical gradient and with soil depth in a Bolivian tropical dry (1160 mm y−1 rain) and moist forest (1580 mm y−1). In each forest we analysed the effect of drought on predawn leaf water potentials (ψpd) and drought response (midday leaf water potential at a standardized ψpd of −0.98 MPa; ψmd) of saplings of three tree species, varying in shade-tolerance and leaf phenology. ψsoil changed during the dry season and most extreme in the dry forest. Crests were drier than slopes and valleys. Dry-forest top soil was drier than deep soil in the dry season, the inverse was found in the wet season. In the moist forest the drought-deciduous species, Sweetia fruticosa, occupied dry sites. In the dry forest the short-lived pioneer, Solanum riparium, occupied wet sites and the shade-tolerant species, Acosmium cardenasii drier sites. Moist-forest species had similar drought response. The dry-forest pioneer showed a larger drought response than the other two species. Heterogeneity in soil water availability and interspecific differences in moisture requirements and drought response suggest great potential for niche differentiation. Species may coexist at different topographical locations, by extracting water from different soil layers and/or by doing so at different moments in time
Top-cited authors
Ariel E. Lugo
  • International Institute of Tropical Forestry
Valerie Kapos
  • UNEP World Conservation Monitoring Centre
Sandra Brown
  • Winrock International
Carlos A. Peres
  • University of East Anglia
M D Swaine
  • University of Aberdeen