How many species of host-specific insects feed on a species of tropical tree?

The assumptions on the host specificity of beetles that led Terry Erwin to suggest that there may be over 30 million arthropod species were tested for 10 species of trees and their insect associates at a rainforest site in Papua New Guinea. The data included 391 species and 4696 individuals of herbivorous beetles collected during a one year period using hand collecting, beating, branch clipping, intercept flight traps and pyrethrum knockdown. Insect host specificity was assessed by feeding trials in captivity. The data suggest that between 23 and 37 monophagous leaf-feeding species are most likely to be present in this system, whereas Erwin's method yields an estimate of 138 monophagous species. The major factors responsible for the discrepancy between our observations and Erwin's assumptions appears to be (a) the importance of transient species; (b) the insect fauna that is shared among tree species; (c) some generalist wood-eating species may inflate the apparent species richness of leaf-feeding beetles; and (d) the proportion of specialist species varies significantly among tree species. We conclude that studies reporting the proportion of specialist insect herbivores associated with particular tropical tree species will yield only a portion of the information needed to estimate global arthropod species richness, but may be useful for elucidating certain aspects of food-web ecology in tropical rain forests.     

Comparing four simple,inexpensive methods for sampling forest arthropod communities

Terrestrial arthropods are diverse, and quantifying their availability to consumers is important for understanding both consumer and insect distribution, abundance, and communities. However, characterizing arthropod communities in complex forest ecosystems is challenging. We compared arthropod communities in a wet‐limestone forest in Jamaica during the dry season sampled by four methods: branch clips, sweep netting, and sticky traps applied to tree trunks and hanging free of vegetation. We found no effect of relative height in the canopy for the two methods that could be used at different heights, i.e., hanging sticky traps and branch clips. In addition, the arthropod community sampled changed over time (season) for sweep nets and branch clips. We also found that branch clips and sweep nets sampled more arthropod taxa than the two sticky‐trap methods. In addition, branch clips and sweep nets sampled more ants and spiders than the two sticky‐trap methods, whereas collar sticky traps on tree trunks sampled more bark lice (Psocoptera), and hanging sticky traps more flies (Diptera) than the other methods. Percentages of flying insects and strong‐flying insects sampled did not differ between sweep netting and branch clipping, but a higher percentage of both groups were captured with collar and hanging sticky traps. Because we found that the different methods sampled different subsets of the arthropod community, both taxonomically and in terms of aerial versus non‐aerial taxa, investigators should choose the arthropod sampling methods that most closely align with their focal species and study questions. For example, investigators might use collar traps for studies of bark gleaners, hanging sticky traps for aerial foragers, and branch clips or sweep nets for foliage gleaners. Alternatively, if a focal species is known to prefer certain prey items, investigators may instead select a method that effectively samples those prey taxa. Finally, for some studies, using multiple sampling methods may be the best option.     

Comparative Study of Ant Communities of Rehabilitated Mineral Sand Mines and Heathland, Western Australia

We conducted a survey of ants in April 1997 in seven rehabilitated (2- to 20-year-old) and three native heathland reference sites of the Renison Goldfields Corporation (RGC) Mineral Sands operations at Eneabba, Western Australia. We employed a variety of collecting methods, including pitfall trapping, litter and soil sampling, sweeping and beating of vegetation, and collecting by hand in the day and after dark, replicating those used by previous researchers in a similar study conducted in the same area in 1980. We found a total of 96 ant species representing 30 genera in 1997, compared with 46 species from 18 genera in 1980. Ant species richness increased with age of rehabilitation and exceeded that of heath controls by the time it reached 11 years. Ant species richness related to rehabilitation age was represented by a logarithmic curve, and the slope for current rehabilitation was steeper than that for the earlier rehabilitation studied in 1980. This probably reflects improved rehabilitation practices. Ordination of the sites in terms of ant species composition indicated differences between ant species in the rehabilitated sites studied in 1980 and 1997, and also between all rehabilitated sites and heath controls. Classification of the sites in terms of the distribution of ants across functional groups also indicated differences between rehabilitation and heath control sites. According to species composition and functional group profiles from rehabilitated and control sites, we concluded that although the current rehabilitation allows for a rapid return of ant species, even after 20 years the fauna still had not attained the composition of the original heath fauna.     

The Amazonas‐trap: a new method for sampling plant‐inhabiting arthropod communities in tropical forest understory

Methods to quantify plant‐insect interactions in tropical forests may miss many important arthropods and can be time consuming and uneven in capture efficiency. We describe the Amazonas‐trap, a new method that rapidly envelops the target plant for sampling arthropods. We evaluated the efficiency of the Amazonas‐trap by comparing it with two commonly used sampling methods to collect arthropods from plants: the beating tray and manual collection. Samples were collected in 10 permanent plots, in the Ducke forest reserve, Manaus (Amazonas, Brazil). In each plot we sampled 18 plant individuals of Protium sp. (Burseraceae): six by a beating tray, six by manual collection, and six using the Amazonas‐trap. All insects were identified to the family level and those belonging to the order Hymenoptera were identified to the species and morphospecies level. The new method sampled more insect families and more Hymenoptera species than tree beating and manual collection. Of the 75 total families collected, 20 were sampled exclusively by the Amazonas‐trap, seven were only collected with a beating tray, and seven were sampled exclusively with manual collecting. A similar pattern was found for abundance: Amazonas‐trap sampled more individuals, followed by the beating tray and manual collection. Small and winged arthropods were more abundant in Amazonas‐trap, explaining the highest richness of Hymenoptera and insect families sampled with this method. The new method sampled more spiders, wood‐fungi feeders, sap suckers, omnivorous, parasitoids, and insect predators than the other methods, but was equally effective in sampling leaf‐feeders and ants. Amazonas‐trap was more time consuming in the field, but for all diversity parameters evaluated, the new method showed better performance for collecting invertebrates on plants.     

Stratification and diel activity of arthropods in a lowland rainforest in Gabon

The abundance, activity and species richness of arthropods, particularly of insect herbivores, were investigated in the upper canopy and understorey of a lowland rainforest at La Makandé, Gabon. In total 14 161 arthropods were collected with beating, flight interception and sticky traps, from six canopy sites, during the day and at night, from mid-January to mid-March 1999. The effects of stratum were most important, representing between 40 and 70% of the explained variance in arthropod distribution. Site effects represented between 20 and 40% of the variance and emphasized the néed for replication of sampling among canopy sites. Time effects (diel activity) explained a much lower percentage of variance (6–9%). The density and abundance of many arthropod taxa and species were significantly higher in the upper canopy than in the understorey. Arthropod activity was also higher during the day than at night. In particular, insect herbivores were 2.5 times more abundant and twice as speciose in the upper canopy than in the understorey, a probable response to the greater and more diverse food resources in the former stratum. Faunal overlap between the upper canopy and understorey was low. The most dissimilar herbivore communities foraged in the understorey at night and the upper canopy during the day. Further, a taxonomic study of a species-rich genus of herbivore collected there (Agrilus , Coleoptera Buprestidae) confirmed that the fauna of the upper canopy was different, diverse and very poorly known in comparison to that of the understorey. Herbivore turnover between day and night was rather high in the upper canopy and no strong influx of insect herbivores from lower foliage to the upper canopy was detected at night. This suggests that insect herbivores of the upper canopy may be resident and well adapted to environmental conditions there.     

Efficient monitoring of maize orange leafhopper, <Emphasis Type="Italic">Cicadulina bipunctata</Emphasis> (Hemiptera: Cicadellidae), and small brown planthopper, <Emphasis Type="Italic">Laodelphax striatellus</Emphasis> (Hemiptera: Delphacidae), in forage maize fields using yellow sticky traps

The monitoring of insect pests in fields of forage maize is difficult because plants are tall and grow at a high density. We investigated the effectiveness of colored sticky traps and appropriate conditions for monitoring insect pests in forage maize fields. Large numbers of the maize orange leafhopper, Cicadulina bipunctata Melichar (Hemiptera: Cicadellidae), and the small brown planthopper, Laodelphax striatellus Fallen (Hemiptera: Delphacidae), were collected during the experimental period with yellow and blue sticky traps placed in summer crop forage maize fields. A greater number of insects were trapped in yellow traps relative to blue traps. Traps located at a lower height (40 cm above the ground) attracted larger numbers of C. bipunctata, whereas L. striatellus did not demonstrate a height-dependent preference. These results indicated that yellow-colored sticky traps located at low height are effective for collecting C. bipunctata and L. striatellus simultaneously. Seasonal occurrence data obtained by the yellow sticky traps showed clearer seasonal occurrences than that obtained by two previously developed methods, suction and light traps, indicating that sticky traps are effective for monitoring the seasonal occurrence of these two insects in forage maize fields.     

Plant Species Redundancy and the Restoration of Faunal Habitat: Lessons from Plant‐Dwelling Bugs

Restoring disturbed lands is essential for conserving biodiversity. In floristically diverse regions, restoring all plant species following anthropogenic disturbance is financially costly and it is unknown if this can be achieved. However, re‐creating faunal habitat may not require reinstating all plant species if there is a high degree of redundancy. Here, we assess whether there is redundancy among a subset of native plant species chosen to restore fauna habitat following a severe disturbance. Additionally, we determine if reestablished plants support similar faunal assemblages as the same plant species in less disturbed forest. We sampled plant‐dwelling Hemiptera from 1,800 plants across 16 species. We found 190 species of Hemiptera, with most plant species in the forest having distinct hemipteran assemblages. Returning these plant species to areas undergoing restoration reinstated 145 hemipteran species, including the dominant species. Recalcitrant plant species (difficult to propagate and reestablish in restored areas) had different hemipteran assemblages from all other species. There was only one plant species that did not have a distinct assemblage and thus was considered redundant. We conclude that there is little redundancy in this study. For plant‐dwelling Hemiptera (with good powers of dispersal) to recolonize restored areas, restoration efforts will need to reinstate at least 13 of the 16 species of host plant of appropriate age and structure. Consequently, to meet the goal of restoring fauna habitat when there is no knowledge of which plant species are redundant, restoration projects should aim to reinstate all plant species present in less disturbed reference areas.     

Comparing the precision,accuracy, and efficiency of branch clipping and sweep netting for sampling arthropods in two Jamaican forest types

ABSTRACT Devising methods for sampling arthropods presents many challenges, including understanding possible differences in results obtained by different individuals (precision), investigating differences between estimates and the actual variable of interest (accuracy), and assessing the effort and cost of a given method (efficiency). We assessed the precision, accuracy, and efficiency of sweep netting and branch clipping, two common methods of sampling arthropods, in mangrove and second‐growth scrub forests in Jamaica, West Indies, in 2009. Three individuals used both methods sequentially to sample arthropods in the territories of American Redstarts (Setophaga ruticilla). We found that both branch clipping and sweep netting lacked precision because different individuals produced different estimates of either arthropod abundance (number of individuals per sample) or biomass. In both forests, more arthropods were sampled with sweep netting, in terms of biomass and abundance, and several orders of arthropods were collected that were missed by branch clipping. We also detected the absence of a predictable habitat‐based difference in arthropod biomass with sweep netting, but not with branch clipping. Sweep netting took longer overall (field and processing time combined) and was therefore less efficient. Despite problems with precision and efficiency, our results suggest that sweep netting may be a more accurate method than branch clipping for sampling foliage arthropods in some forest habitats. Our study also reveals the importance of recognizing and controlling for individual bias and of choosing arthropod sampling methods most appropriate to each study species and habitat type.     

The Hemiptera (Insecta) communities of tropical rain forest in Sulawesi

The structure and composition of Hemiptera communities of tropical rain forest in Dumoga-Bone National Park, Sulawesi, Indonesia were investigated over a 1-year period. The aim was to investigate the extent to which insect samples, obtained using six different techniques, were representative of the whole community, and to explore temporal and spatial variation in Hemiptera community composition. Sampling techniques employed were Rothamsted light trapping, canopy fogging, malaise trapping at both ground and canopy level, flight interception trapping and yellow water pan trapping. Overlap between faunas collected using different techniques was surprisingly low emphasizing the limitations of using any single method to sample and thereby describe the Hemiptera community. Differences were observed at the species, family and suborder levels. Similarly, there were major differences in faunal composition between sites and diversity appeared to peak at elevations between 600–1000 m. Seasonal changes were also significant but generally lower than between methods or sites. Results are discussed in relation to factors such as food availability and host specialization and the efficacy of each sampling method is reviewed in relation to the biology of the different Hemiptera groups and the composition of the samples obtained.     

Trap color and placement affects estimates of insect family-level abundance and diversity in a Nebraska salt marsh

Sampling programs to establish baseline ecosystem information (e.g., species abundance and diversity) often fail to consider the potential influence of sampling techniques on results. Research on sampling economically important insects has demonstrated the possible influences of trap color and trap placement on results, but few data have been collected from natural environments. Consequently, we examined the effects of color (yellow and blue) and placement (exposed and shaded by plants) of sticky traps on insect captures and diversity estimates from a Nebraska inland salt marsh community. We identified 1913 specimens from 67 insect families collected during five trapping dates in July 1996. More Cicindelidae were collected on exposed traps, and more Staphylinidae, Dolichopodidae, Cicadellidae, and Thripidae were collected on shaded traps. More Dolichopodidae were collected on yellow traps, while more Syrphidae and Thripidae were collected on blue. Shannon and alpha diversity measures were significantly higher for shaded traps than exposed traps, but were not affected by trap color. Our results highlight the importance of characterizing sampling techniques when establishing diversity estimates. These data provide the first complete accounting of community-level insect response to colored sticky traps and provide new information for color preference of non-economic insect species.     

Diversity of canopy and understorey spiders in north-temperate hardwood forests

1 We characterized and compared diversity patterns of canopy and understorey spiders (Arachnida: Araneae) on sugar maple ( Acer saccharum Marsh.) and American beech ( Fagus grandifolia Ehrh.) in hardwood forests of southern Québec, Canada.
2 We sampled canopies of 45 sugar maple and 45 American beech trees and associated understorey saplings in mature protected forests near Montréal. Samples were obtained by beating the crown foliage at various heights and by beating saplings around each tree.
3 Eighty-two species were identified from 13 669 individuals. Forty-eight species and 3860 individuals and 72 species and 9809 individuals were collected from the canopy and the understorey, respectively.
4 Multivariate analyses (NMDS ordination and NPMANOVA) showed the composition of canopy and understorey assemblages differed significantly, and canopy assemblages differed between tree species. Rank-abundance distribution models fitted to the canopy and understorey data indicated that different mechanisms structure the assemblages in both habitats. Three abundant spider species were significantly more common in the canopy; ten species were collected significantly more often in the understorey.
5 The forest canopy was shown to be an important reservoir for spider diversity in north-temperate forests.     

Hand searching versus pitfall trapping: how to assess biodiversity of ground beetles (Coleoptera: Carabidae) in high altitude equatorial Andes?

The use of ground beetles (Coleoptera: Carabidae) as bioindicators of environmental change depends on the reliability and the effectiveness of the sampling methods. Those that have been tested in the temperate zone and in tropical forests still await experimentation in tropical high-altitude environments. For the first time, pitfall trapping and hand searching have been compared in Ecuadorian páramo above 4000 m a.s.l., in terms of practical effectiveness. The study was performed on six volcanoes and was based on the comparison of 28 sampling sessions (pitfall trapping and hand searching) performed along two different elevational belts [lower superpáramo (LSP) and upper superpáramo (USP)]. Analyses of sampling sessions showed that detected species richness is slightly higher with hand searching than with pitfall trapping, regardless of the elevation. Additionally, hand searching is more time-effective than pitfall trapping. The performance of the sampling method slightly varies when species assemblage composition is analysed in relation to elevational belts. In the LSP, hand searching and pitfall trapping should be simultaneously used to obtain exhaustive inventories of carabid biodiversity, since different species are likely to be collected by each method. In the USP, hand searching and pitfall trapping efficiency is very similar, but hand searching allows to collect a slightly larger number of species. Lastly, the sample-based rarefaction curves showed that four temporal replicates are mandatory to obtain a robust dataset and an exhaustive inventory of the true species richness and species assemblages composition. Our findings suggest a combined use of hand searching and pitfall trapping in the LSP, while both methods can be used alone for surveying carabids in the USP. Furthermore, hand searching is recommended if the aim is to obtain an inventory of species diversity, whereas pitfall trapping seems more convenient for fine grain ecological and comparative studies.     

Sampling Rhodnius neglectus in Mauritia flexuosa palm trees: a field study in the Brazilian savanna

Two sampling methods (manual capture and live-baited adhesive traps) were compared for collecting the bug Rhodnius neglectus Lent (Hemiptera: Reduviidae: Triatominae) from palm trees, Mauritia flexuosa L. (Arecaceae), in the savanna of Brasília DF. R. neglectus was found in 19/50 (38%) of palm trees sampled. The detection rate was much higher by visual inspection and manual capture (18/50=36%) than by our trapping method (5/50=10%), although one tree was found to be positive by trapping but not by manual capture. Bugs collected manually were mostly (146/154=95%) found among the dead organic material in palm crowns. In combination, these sampling techniques are useful for quick detection of triatomine bug infestation in palm trees, especially in areas of high ecological value where the palms should not be cut and dissected, but arboreal Rhodnius are suspected to transmit enzootic Trypanosoma cruzi that might represent a risk of causing human cases of Chagas disease.     

Phylogenomics of Hemiptera (Insecta: Paraneoptera) based on mitochondrial genomes

Hemiptera is the largest order in Paraneoptera and the fifth largest in Insecta. Disputes about hemipteran phylogeny have concerned the monophyly of Auchenorrhyncha and relationships between the suborders Fulgoromorpha, Cicadomorpha, Coleorrhyncha and Heteroptera. In a phylogenomic study of Hemiptera, we add two new mitochondrial genomes of Peloridiidae (Coleorrhyncha) to those reported in GenBank, to complete the taxon sampling of all suborders. We used two types of data – amino acid sequences and nucleotides of various combinations between protein coding genes, tRNAs and rRNAs – to infer the phylogeny of Hemiptera. In total 27 taxa of Paraneoptera were sampled, 24 of them being hemipterans. Bayesian inference, maximum likelihood and maximum parsimony analyses were employed. The relationship of Cicadomorpha + Heteroptera is always stable in the results with different combinations between data types and phylogenetic methods, but our results challenge the monophyly of ‘Homoptera’ and Auchenorrhyncha. In evaluating the relative contribution of each gene, the phylograms generated by single genes CO1, ND1, ND2, ND4 and ND5, respectively, closely matched the tree yielded by the combined datasets. In light of the taxon‐sampling sensitivity of trees based on mitochondrial genomes, the results need to be tested with further data from nuclear genes.     

北京市珍珠泉乡不同昆虫采集方法采集效果对比分析

为了探究科研调查中不同昆虫采集方法的采集效率差异,在北京珍珠泉乡不同的生境类型中,采用样线法、马氏网法、灯诱法、陷阱法(糖醋液)4种采集方法进行昆虫采集,按生境类型对不同采集方法采集的昆虫进行统计分析.在实验中共采集昆虫3996头,隶属12目87科,其中在昆虫种类数量上,鳞翅目最多,其次是半翅目和鞘翅目;在数量上,半翅...     

A comparison of three methods to survey saproxylic beetles in hollow oaks

One of the most endangered assemblages of species in Europe is insectsassociated with old trees. For that reason there is a need of developing methodsto survey this fauna. This study aims at comparing three methods – windowtrapping, pitfall trapping and wood mould sampling – to assess speciesrichness and composition of the saproxylic beetle fauna in living, hollow oaks.We have used these methods at the same site, and to a large extent in the sametrees. Useful information was obtained from all methods, but they partiallytarget different assemblages of species. Window trapping collected the highestnumber of species. Pitfall trapping collected beetles associated with treehollows which rarely are collected by window traps and therefore it isprofitable to combine these two methods. As wood mould sampling is the cheapestmethod to use, indicator species should preferably be chosen among specieswhich are efficiently collected with this method.     

Spatial and temporal variation in butterfly biodiversity in a West African forest: lessons for establishing efficient rapid monitoring programmes

Because tropical forests face serious threats and are usually situated in developing countries, cheap and easy Rapid Biodiversity Assessment (RBA) seems an ideal strategy for documenting changes in habitat quality and setting conservation priorities. Fruit‐feeding butterflies are often used as focal taxa to document biodiversity trends in tropical forests. However, the choice of inappropriate sampling strategies may nullify its value. More extensive data sets that sample multiple months and seasons and incorporate the vertical dimension are needed as a baseline for designing reliable sampling regimes. Therefore, we performed a butterfly trapping study in Ghana, using both understorey and canopy traps, and sampling seven periods covering dry and wet seasons. Both individual numbers and species richness were on average three to four times higher in the understorey than in the canopy traps with strikingly different species composition (only 11% overlap in species). The number of species and individual abundance recorded in a month varied extensively. This study underlined the importance of taking into account temporal variation and vertical stratification when designing RBAs of fruit‐feeding butterflies in West African forests. We recommend the use of both understorey and canopy traps and inclusion of both ‘wet’ and ‘dry’ seasons into RBA sampling regimes.     

Unnecessary roughness? Testing the hypothesis that predators destined for molecular gut‐content analysis must be hand‐collected to avoid cross‐contamination

Molecular gut‐content analysis enables detection of arthropod predation with minimal disruption of ecosystem processes. Mass‐collection methods, such as sweep‐netting, vacuum sampling and foliage beating, could lead to regurgitation or rupturing of predators along with uneaten prey, thereby contaminating specimens and compromising resultant gut‐content data. Proponents of this ‘cross‐contamination hypothesis’ advocate hand‐collection as the best way to avoid cross‐contamination. However, hand‐collection is inefficient when large samples are needed, as with most ecological research. We tested the cross‐contamination hypothesis by setting out onto potato plants immature Coleomegilla maculata and Podisus maculiventris that had been fed larvae of either Leptinotarsa decemlineata or Leptinotarsa juncta, or unfed individuals of these predator species along with L. decemlineata larvae. The animals were then immediately re‐collected, either by knocking them vigorously off the plants onto a beat cloth and capturing them en masse with an aspirator (‘rough’ treatment) or by hand‐searching and collection with a brush (‘best practice’). Collected predators were transferred in the field to individual vials of chilled ethanol and subsequently assayed by PCR for fragments of cytochrome oxidase I of L. decemlineata and L. juncta. Ten to 39 per cent of re‐collected fed predators tested positive by PCR for DNA of both Leptinotarsa species, and 14–38% of re‐collected unfed predators contained L. decemlineata DNA. Overall levels of cross‐contamination in the rough (31%) and best‐practice (11%) samples were statistically different and supported the cross‐contamination hypothesis. A pilot study on eliminating external DNA contamination with bleach prior to DNA extraction and amplification gave promising results.     

Species richness of phytophagous beetles in the tropical tree Brosimum utile (Moraceae): the effects of sampling strategy and the problem of tourists

Abstract.  1. Phytophagous beetles on six mature living trees and two dead trees of Brosimum utile (Moraceae) were surveyed during 1 year in a tropical wet forest in Panama. The dead trees were surveyed both as suspended in the canopy and after falling down to the understorey.
2. Canopy access was provided by a construction crane and sampling was performed by beating and hand-collecting. The same amount of time was spent on each tree in order to standardise sampling effort. A list of all species associated with the tree is presented.
3. A total of 3009 individuals representing 364 species were collected. Tourists were excluded from the analyses by recording host associations directly and by probability assessments of host associations based on abundance categories. A total of 2603 individuals and 244 species were associated with the tree. The proportion of tourists in the trees increased with sample size.
4. A single mature living tree had on average 58.5 ± 6.5 species. The local species richness of B. utile was estimated as 2.5 times higher than in a single mature tree; however, a substantial increase in species richness was attained when dead wood habitats were included. Saproxylic species made up 82% of the total material.
5. The investigated habitat types of B. utile constituted distinct, complementary species assemblages. Similarity between saproxylic species of dead suspended wood and dead understorey wood of the same tree was 0.2 (Morisita–Horn index), confirming a prominent vertical stratification among this guild.     

Ant-hemipteran trophobioses in a Bornean rainforest – diversity, specificity and monopolisation

Trophobiotic interactions between ants, hemipterans and plants play an important role for all three partners. This study compared a broad spectrum of trophobiotic associations in a tropical rainforest in Sabah, Borneo. We studied partner specificity, ant recruitment, temporal continuity and monopolisation in 218 trophobioses, comprising 58 ant species, 62 hemipteran and over 31 plant species. The most common associations involved Dinochloa trichogona (Poaceae) with coreids and delphacids in the forest understorey, and the invasive weed Chromolaena odorata (Asteraceae) with Aphis gossypii and A. spiraeola in the open vegetation; both associations were attended by a broad spectrum of ant species. In general, associations between hemipterans and plants were highly and significantly specialised, while ants were more opportunistic in their choices of partners, although partitioning was also significant between ant versus hemipteran species and consequently between ant versus plant species. The number of ant workers increased significantly, but at a declining rate, with the number of hemipterans at a trophobiosis. Most trophobioses (96%) were only tended by a single ant species at a time and thus effectively monopolised. Occasionally these guards were replaced by another ant species after a few weeks (11%) or during the night (34%). In order to test whether other sugar-seeking ants as potential competitors occurred in the vicinity of trophobioses, sugar baits were placed next to the trophobioses, on a different branch of the same plant, and on a neighbouring plant. While the hemipteran-tending ant colony mostly monopolised the nearest sugar bait, the number of ant species on more distant baits was significantly higher. Our results show that ant associations with honeydew-producing hemipterans may be relatively opportunistic at the community level, but highly predictable on a smaller spatio-temporal scale in respect to recruitment to, and long term securing of this important resource. Received 23 August 2005; revised 22 November 2005; accepted 19 December 2005.