How plants shape the ant community in the Amazonian rainforest canopy: the key role of extrafloral nectaries and homopteran honeydew

Ant-plant interactions in the canopy of a lowland Amazonian rainforest of the upper Orinoco, Venezuela, were studied using a modified commercial crane on rails (Surumoni project). Our observations show a strong correlation between plant sap exudates and both abundance of ants and co-occurrence of ant species in tree canopies. Two types of plant sap sources were compared: extrafloral nectaries (EFNs) and honeydew secretions by homopterans. EFNs were a frequent food source for ants on epiphytes (Philodendron spp., Araceae) and lianas (Dioclea, Fabaceae), but rare on canopy trees in the study area, whereas the majority of trees were host to aggregations of homopterans tended by honeydew-seeking ants (on 62% of the trees examined). These aggregations rarely occurred on epiphytes. Baited ant traps were installed on plants with EFNs and in the crowns of trees from three common genera, including trees with and without ant-tended homopterans: Goupia glabra (Celastraceae), Vochysia spp. (Vochysiaceae), and Xylopia spp. (Annonaceae). The number of ant workers per trap was significantly higher on plants offering one of the two plant sap sources than on trees without such resources. Extrafloral nectaries were used by a much broader spectrum of ant species and genera than honeydew, and co-occurrence of ant species (in traps) was significantly higher on plants bearing EFNs than on trees. Homopteran honeydew (Coccidae and Membracidae), on the other hand, was mostly monopolised by a single ant colony per tree. Homopteran-tending ants were generally among the most dominant ants in the canopy. The most prominent genera were Azteca, Dolichoderus (both Dolichoderinae), Cephalotes, Pheidole, Crematogaster (all Myrmicinae), and Ectatomma (Ponerinae). Potential preferences were recorded between ant and homopteran species, and also between ant-homopteran associations and tree genera. We hypothesize that the high availability of homopteran honeydew provides a key resource for ant mosaics, where dominant ant colonies and species maintain mutually exclusive territories on trees. In turn, we propose that for nourishment of numerous ants of lower competitive capacity, Philodendron and other sources of EFNs might be particularly important.     

The contribution of cacao agroforests to the conservation of lower canopy ant and beetle diversity in Indonesia

The ongoing destruction of tropical rainforests has increased the interest in the potential value of tropical agroforests for the conservation of biodiversity. Traditional, shaded agroforests may support high levels of biodiversity, for some groups even approaching that of undisturbed tropical forests. However, it is unclear to what extent forest fauna is represented in this diversity and how management affects forest fauna in agroforests. We studied lower canopy ant and beetle fauna in cacao agroforests and forests in Central Sulawesi, Indonesia, a region dominated by cacao agroforestry. We compared ant and beetle species richness and composition in forests and cacao agroforests and studied the impact of two aspects of management intensification (the decrease in shade tree diversity and in shade canopy cover) on ant and beetle diversity. The agroforests had three types of shade that represented a decrease in tree diversity (high, intermediate and low diversity). Species richness of ants and beetles in the canopies of the cacao trees was similar to that found in lower canopy forest trees. However, the composition of ant and beetle communities differed greatly between the agroforest and forest sites. Forest beetles suffered profoundly from the conversion to agroforests: only 12.5% of the beetle species recorded in the forest sites were also found in the agroforests and those species made up only 5% of all beetles collected from cacao. In contrast, forest ants were well represented in agroforests, with 75% of all species encountered in the forest sites also occurring on cacao. The reduction of shade tree diversity had no negative effect on ants and beetles on cacao trees. Beetle abundances and non-forest ant species richness even increased with decreasing shade tree diversity. Thinning of the shade canopy was related to a decrease in richness of forest ant species on cacao trees but not of beetles. The contrasting responses of ants and beetles to shade tree management emphasize that conservation plans that focus on one taxonomic group may not work for others. Overall ant and beetle diversity can remain high in shaded agroforests but the conservation of forest ants and beetles in particular depends primarily on the protection of natural forests, which for forest ants can be complemented by the conservation of adjacent shaded cacao agroforests.     

The invasive Yellow Crazy Ant and the decline of forest ant diversity in Indonesian cacao agroforests

Throughout the tropics, agroforests are often the only remaining habitat with a considerable tree cover. Agroforestry systems can support high numbers of species and are therefore frequently heralded as the future for tropical biodiversity conservation. However, anthropogenic habitat modification can facilitate species invasions that may suppress native fauna. We compared the ant fauna of lower canopy trees in natural rainforest sites with that of cacao trees in agroforests in Central Sulawesi, Indonesia in order to assess the effects of agroforestry on occurrence of the Yellow Crazy Ant Anoplolepis gracilipes, a common invasive species in the area, and its effects on overall ant richness. The agroforests differed in the type of shade-tree composition, tree density, canopy cover, and distance to the village. On average, 43% of the species in agroforests also occurred in the lower canopy of nearby primary forest and the number of forest ant species that occurred on cacao trees was not related to agroforestry characteristics. However, A. gracilipes was the most common non-forest ant species, and forest ant richness decreased significantly with the presence of this species. Our results indicate that agroforestry may have promoted the occurrence of A. gracilipes, possibly because tree management in agroforests negatively affects ant species that depend on trees for nesting and foraging, whereas A. gracilipes is a generalist when it comes to nesting sites and food preference. Thus, agroforestry management that includes the thinning of tree stands can facilitate ant invasions, thereby threatening the potential of cultivated land for the conservation of tropical ant diversity.     

Trees as islands: canopy ant species richness increases with the size of liana‐free trees in a Neotropical forest

The physical characteristics of habitats shape local community structure; a classic example is the positive relationship between the size of insular habitats and species richness. Despite the high density and proximity of tree crowns in forests, trees are insular habitats for some taxa. Specifically, crown isolation (i.e. crown shyness) prevents the movement of small cursorial animals among trees. Here, we tested the hypothesis that the species richness of ants (S_a) in individual, isolated trees embedded within tropical forest canopies increases with tree size. We predicted that this pattern disappears when trees are connected by lianas (woody vines) or when strong interactions among ant species determine tree occupancy. We surveyed the resident ants of 213 tree crowns in lowland tropical forest of Panama. On average, 9.2 (range = 2–20) ant species occupied a single tree crown. Average (± SE) S_a was ca 25% higher in trees with lianas (10.2 ± 0.26) than trees lacking lianas (8.0 ± 0.51). S_a increased with tree size in liana‐free trees (S_a = 10.99A^0.256), but not in trees with lianas. Ant species composition also differed between trees with and without lianas. Specifically, ant species with solitary foragers occurred more frequently in trees with lianas. The mosaic‐like pattern of species co‐occurrence observed in other arboreal ant communities was not found in this forest. Collectively, the results of this study indicate that lianas play an important role in shaping the local community structure of arboreal ants by overcoming the insular nature of tree crowns.     

Ant visitation of extrafloral nectaries of Passiflora: the effects of nectary attributes and ant behavior on patterns in facultative ant-plant mutualisms

Extrafloral nectary (EFN) plants are widespread and can be quite species-rich in some communities. Thus, ants that utilize extrafloral nectar may have the opportunity to discriminate among a wide variety of nectar sources, resulting in variation in the ant attention EFN plants receive. In this study, we compare ant visitation rates of three Passiflora species that coexist in an early successional neotropical forest. These three vine species (Passiflora auriculata, P. biflora, and P. oerstedii) differ in their extrafloral nectary structure and placement, and thus may attract different numbers or species of ants. Through censuses of ants tending extrafloral nectaries, we found that P. auriculata received significantly higher numbers of ant visitors than P. oerstedii, but did not differ significantly from P. biflora in its attractiveness to ants. We also found that termite worker baits (simulating herbivores) placed on P. auriculata and P. biflora were discovered by ants significantly more quickly than baits placed on P. oerstedii. In both ant visitation censuses and in termite bait trials, we found no significant associations between Passiflora species and the species of ant visitors. We also performed experimental manipulations of several characteristics of P. auriculata, which resulted in changes in levels of ant visitation. When petiolar nectaries of P. auriculata were experimentally blocked, visitation by the common ant Ectatomma ruidum declined, even though nectaries on the leaf surfaces were still functional. Connections with other vegetation also had an effect on ant visitation. Though experimental creation of connections between growing P. auriculata shoots and other vegetation did not enhance ant visitation, eliminating connections resulted in a significant decline in the number of ant visitors. The results of this study suggest factors that may contribute to variation in ant visitation of extrafloral nectary plants. In addition, this study demonstrates that extrafloral nectary plants co-occurring in a habitat and available to the same ants may differ in patterns of visitation by ants and perhaps in the quality of protection from herbivores that they receive.     

Patterns of the <Emphasis Type="Italic">Crematogaster-Macaranga</Emphasis> association: The ant partner makes the difference

Summary. One of the most species-rich ant-plant mutualisms worldwide is the palaeotropical Crematogaster-Macaranga system. Although the biogeography and ecology of both partners have been extensively studied, little is known about the temporal structuring and the dynamics of the association. In this study we compared life-history traits of the specific Crematogaster (Decacrema) partner-ants and followed the development of ant colonies on eight different Macaranga host plant species, from colony founding on saplings to adult trees in a snapshot fashion. We found differences in the onset of alate production, queen number and mode of colony founding in the ant species and examined the consequences of these differences for the mutualism with the host plant. The lifespan of some host plants and their specific ant partners seemed to be well matched whereas on others we found an ontogenetic succession of specific partner ants. The partner ants of saplings or young plants often differed from specific partner ants found on larger trees of the same species. Not all specific Crematogaster species can re-colonize the crown region of adult trees, thus facilitating a change of ant species. Therefore lifespan of the ant colony as well as colony founding behaviour of the different partner ant species are important for these ontogenetic changes. The lifespan of a colony of two species can be prolonged via secondary polygyny. For the first time, also primary polygyny (pleometrosis) is reported from this myrmecophytic system.     

The arboreal ants of a Neotropical rain forest show high species density and comprise one third of the ant fauna

In tropical rain forests, the ant community can be divided into ground and arboreal faunas. Here, we report a thorough sampling of the arboreal ant fauna of La Selva Biological Station, a Neotropical rain forest site. Forty-five canopy fogging samples were centered around large trees. Individual samples harbored an average of 35 ant species, with up to 55 species in a single sample. The fogging samples yielded 163 observed species total, out of a statistically estimated 199 species. We found no relationship between within-sample ant richness and focal tree species, nor were the ant faunas of nearby trees more similar to each other than the faunas of widely spaced trees. Species density was high, and beta diversity was low: A single column of vegetation typically harbors at least a fifth of the entire arboreal ant fauna. Considering the entire fauna, based on 23,326 species occurrence records using a wide variety of collecting methods, 182 of 539 observed species (196 of 605, estimated statistically) were entirely arboreal. The arboreal ant fauna is thus about a third of the total La Selva ant fauna, a robust result because inventory completeness was similar for ground and arboreal ants. The taxonomic history of discovery of the species that make up the La Selva fauna reveals no disproportionately large pool of undiscovered ant species in the canopy. The "last biotic frontier" for tropical ants has been the rotten wood, leaf litter, and soil of the forest floor.     

Ant protection of a Heteropteran trophobiont against a parasitoid wasp

Trophobiotic associations between Caternaultiella rugosa (Heteroptera; Plataspidae) and two ant species, Camponotus brutus (Formicidae) and Myrmicaria opaciventris (Myrmicinae), were recorded at the base of the trunks of Bridelia micrantha (Euphorbiaceae). The bug colonies were generally sheltered in pavilions built by the ants, but during periods of proliferation, part of the colonies developed outside the pavilions. We examined the protective activity of the two associated ant species against an encyrtid parasitoid wasp that parasitized egg masses of the bug. Egg masses situated outside the pavilions were significantly less parasitized on trees occupied by either ant species than on experimental trees where ants were excluded by a sticky barrier. Egg masses were significantly less parasitized on trees occupied by C. brutus than on those occupied by M. opaciventris, both inside and outside the pavilions. Moreover, they were less parasitized inside than outside the pavilions when protected by C. brutus. Inside the pavilions, parental care of the egg masses and first-instar nymphs was abandoned to the ants. In contrast, outside the pavilions, the females of C. rugosa guarded egg masses and clusters of first-instar nymphs. Furthermore, adults and last-instar nymphs grouped around and above first-instar nymphs, forming a protective shield. These groups formed and disintegrated daily, according to the attending ant's rhythm of activity.     

Sources of variation in the incidence of ant–aphid mutualism in boreal forests

• 1 The mutualism between wood ants of the Formica rufa group and aphids living in the canopy of trees is a widespread phenomenon in boreal forests, and it can affect tree growth. However, not all trees in the forest are involved in this interaction.
• 2 To assess the incidence of host trees involved in this ant–aphid mutualism and its spatial distribution in boreal forests, we inventoried sample plots with a radius of 10–15 m around wood ant mounds in 12 forest stands of two age classes (5–12‐year‐old sapling stands and 30–45‐year‐old pole stands) and two dominant tree species (Scots pine and silver birch) in Eastern Finland from 2007 to 2009.
• 3 The proportion of trees visited by ants out of all trees on the individual study plots were in the range 4–62%, and 1.5–39% of the trees on the plots were consistently visited by ants during all 3 years. The percentage of host trees increased with the ant mound base area on the plots. Trees visited by ants were larger and closer to the mound than trees not visited by ants. Within the group of visited trees, more ants were found on bigger trees and on trees close to the ant mounds.
• 4 Extrapolated from plot to stand level, we estimated that 0.5–6.6% of the trees were host trees in at least one of the three study years, and that only 0.01–2.3% of all the trees were consistently visited by ants during all 3 years. It is concluded that ant–aphid mutualism is a minor occurrence at the stand level.

     

Structure of ant assemblages on street trees in urban Busan,Korea

Street trees planted in urban areas are one of the smallest urban green habitats; however, their role as a valuable refuge for local biota is poorly understood. The aim of this study was to investigate urban street trees as habitat for ants. We found ants on 195 (92%) of 211 street trees studied. Seventeen ant species were collected from five tree species. The two most common ant species were Tetramorium tsushimae (on 112 trees) and Lasius japonicus (on 93 trees). The number of ant species found per tree ranged from 0 to 5, with a mean of 1.55 ± 0.92. The species composition differed significantly among tree species. Nine ant species (Camponotus vitiosus, Lasius alienus, L. japonicus, Nylanderia flavipes, Crematogaster matsumurai, Crematogaster vagula, Monomorium intrudens, T. tsushimae, and Brachyponera chinensis) nested on the trees, mostly in the root zone. Nests were found on 150 trees but were relatively rare on trees with diameter at breast height of less than 40 cm.     

Disentangling the Diversity of Arboreal Ant Communities in Tropical Forest Trees

Tropical canopies are known for their high abundance and diversity of ants. However, the factors which enable coexistence of so many species in trees, and in particular, the role of foragers in determining local diversity, are not well understood. We censused nesting and foraging arboreal ant communities in two 0.32 ha plots of primary and secondary lowland rainforest in New Guinea and explored their species diversity and composition. Null models were used to test if the records of species foraging (but not nesting) in a tree were dependent on the spatial distribution of nests in surrounding trees. In total, 102 ant species from 389 trees occurred in the primary plot compared with only 50 species from 295 trees in the secondary forest plot. However, there was only a small difference in mean ant richness per tree between primary and secondary forest (3.8 and 3.3 sp. respectively) and considerably lower richness per tree was found only when nests were considered (1.5 sp. in both forests). About half of foraging individuals collected in a tree belonged to species which were not nesting in that tree. Null models showed that the ants foraging but not nesting in a tree are more likely to nest in nearby trees than would be expected at random. The effects of both forest stage and tree size traits were similar regardless of whether only foragers, only nests, or both datasets combined were considered. However, relative abundance distributions of species differed between foraging and nesting communities. The primary forest plot was dominated by native ant species, whereas invasive species were common in secondary forest. This study demonstrates the high contribution of foragers to arboreal ant diversity, indicating an important role of connectivity between trees, and also highlights the importance of primary vegetation for the conservation of native ant communities.     

The Importance of Primary Tropical Rain Forest For Species Diversity: An Investigation Using Arboreal Ants as an example

The arboreal ant communities of a primary lowland rain forest and three differently disturbed forests that lie close together forming an anthropogenic disturbance gradient were collected with insecticide fogging. Combined samples from all trees (87 foggings) comprised 153,504 ants sorted to 331 morphospecies. The primary forest ant fauna was characterized by high species richness and 53 foggings were necessary to collect communities representatively. Another 63 species of ants were found in the disturbed forests indicating a large regional species pool that might exceed 420 species of arboreal ants. Anthropogenic disturbance caused a change in the taxonomic composition, diversity and structure of ant communities. Community size was a predictor of species richness in the severely disturbed forest types but not in the old secondary or primary forest. Ant abundance had declined significantly in the disturbed forests and only 10% of the primary forest’s species were collected in the most disturbed forest type. In each of the secondary forests a change in the frequency distribution of species was observed and a small number of species had gained numerical dominance. Analysis of species associations indicates that the strength of species interactions changed with the degree of forest disturbance. These changes were still clearly recognizable after 40 years of forest regeneration despite optimal conditions for colonization from the adjacent primary forest, demonstrating that the time scale needed for forest recovery after anthropogenic disturbance is very long.     

Arboreal ants as key predators in tropical lowland rainforest trees

Ants numerically dominate the canopy fauna of tropical lowland rain forests. They are considered to be key predators but their effects in this regard have only rarely been studied on non-myrmecophytes. A conspicuously low abundance of less mobile, mainly holometabolous arthropods like Lepidoptera larvae corresponds with ant dominance, while hemimetabolous highly mobile nymphs occur regularly and in large numbers in the trees. This is in contrast to the temperate regions where ants are mostly lacking on trees and holometabolous larvae are frequent. In this study we experimentally measured ant predation in the trees by offering caterpillars as baits. Fifty-four ant species were tested, of which 46 killed caterpillars and carried them away to their nests while only eight species ignored the offered larvae. Insecticidal knockdown fogging of ten trees after finishing the prey experiments showed that on average 85% of ant individuals per tree were predacious. With the analysis of another 69 foggings and meticulous observations in many other trees this suggests that arboreal ants are responsible for the low abundance of less mobile arthropods in tropical lowland rain forest canopies. Ant predation was significantly lower in a disturbed forest indicating that human disturbance induces a change in the functional interactions in these ecosystems.     

Why are there more arboreal ant species in primary than in secondary tropical forests?

1.?Species diversity of arboreal arthropods tends to increase during rainforest succession so that primary forest communities comprise more species than those from secondary vegetation, but it is not well understood why. Primary forests differ from secondary forests in a wide array of factors whose relative impacts on arthropod diversity have not yet been quantified. 2.?We assessed the effects of succession-related determinants on a keystone ecological group, arboreal ants, by conducting a complete census of 1332 ant nests from all trees with diameter at breast height?≥?5?cm occurring within two (unreplicated) 0·32-ha plots, one in primary and one in secondary lowland forest in New Guinea. Specifically, we used a novel rarefaction-based approach to match number, size distribution and taxonomic structure of trees in primary forest communities to those in secondary forest and compared the resulting numbers of ant species. 3.?In total, we recorded 80 nesting ant species from 389 trees in primary forest but only 42 species from 295 trees in secondary forest. The two habitats did not differ in the mean number of ant species per tree or in the relationship between ant diversity and tree size. However, the between-tree similarity of ant communities was higher in secondary forest than in primary forest, as was the between-tree nest site similarity, suggesting that secondary trees were more uniform in providing nesting microhabitats. 4.?Using our rarefaction method, the difference in ant species richness between two forest types was partitioned according to the effects of higher tree density (22·6%), larger tree size (15·5%) and higher taxonomic diversity of trees (14·3%) in primary than in secondary forest. The remaining difference (47·6%) was because of higher beta diversity of ant communities between primary forest trees. In contrast, difference in nest density was explained solely by difference in tree density. 5.?Our study shows that reduction in plant taxonomic diversity in secondary forests is not the main driver of the reduction in canopy ant species richness. We suggest that the majority of arboreal species losses in secondary tropical forests are attributable to simpler vegetation structure, combined with lower turnover of nesting microhabitats between trees.     

Communities of insect herbivores foraging on saplings versus mature trees of Pourouma bicolor (Cecropiaceae) in Panama

The arthropod fauna of 25 saplings and of three conspecific mature trees of Pourouma bicolor (Cecropiaceae) was surveyed for 12 months in a tropical wet forest in Panama, with particular reference to insect herbivores. A construction crane erected at the study site provided access to tree foliage in the upper canopy. A similar area of foliage (ca. 370 m²) was surveyed from both saplings and trees, but samples obtained from the latter included 3 times as much young foliage as from the former. Arthropods, including herbivores and leaf-chewing insects with a proven ability to feed on the foliage of P. bicolor were 1.6, 2.5 and 2.9 times as abundant on the foliage of trees as on that of saplings. The species richness of herbivores and proven chewers were 1.5 (n=145 species) and 3.5 (n=21) times higher on trees than on saplings, respectively. Many herbivore species preferred or were restricted to one or other of the host stages. Host stage and young foliage area in the samples explained 52% of the explained variance in the spatial distribution of herbivore species. Pseudo-replication in the two sampling universes, the saplings and trees studied, most likely decreased the magnitude of differences apparent between host stages in this forest. The higher availability of food resources, such as young foliage, in the canopy than in the understorey, perhaps combined with other factors such as resource quality and enemy-free space, may generate complex gradients of abundance and species richness of insect herbivores in wet closed tropical forests.     

Pruning of host plant neighbours as defence against enemy ant invasions: <Emphasis Type="Italic">Crematogaster </Emphasis>ant partners of <Emphasis Type="Italic">Macaranga</Emphasis> protected by "wax barriers" prune less than their congeners

Many ant partners of tropical ant-plants prune the leaves and shoot tips of other plants growing around their hosts. According to the hypothesis proposed by Davidson et al. (Ecology 69:801-808), this specialized behaviour not only protects the host plants against overgrowth, but it also conveys a direct benefit to the ant colony as it removes contact points to the neighbouring vegetation where invasions of enemy ants could occur. Here we test this hypothesis by comparing pruning intensity in five closely related Crematogaster (subgenus Decacrema) plant-ant species (and one species of Technomyrmex) that differ in their exposure to competition by other ants. Pruning intensity was quantified by measuring the area loss of paper tape pieces wrapped around the stems of Macaranga host plants. All Crematogaster (Decacrema) ants tested but not Technomyrmex sp. pruned, but the intensity of the behaviour varied strongly between and within species. Pruning was significantly weaker in the three tested Crematogaster species inhabiting Macaranga host plants with a slippery, waxy stem surface, which functions as a mechanical barrier protecting the specific ant partners against generalist competitors. Pruning was generally stronger on more densely ant-populated trees. Even though the number of ants per twig length was lower in associations of ants with glaucous Macaranga hosts, only part of the variation of pruning activity could be explained by "ant density". When corrected for ant density, "wax-running" Crematogaster (Decacrema) ants still pruned more weakly than their congeners inhabiting non-glaucous Macaranga hosts. Pruning is obviously most important when an ant-plant is potentially accessible to intruders, but less necessary when the ant colony is isolated by a protective wax barrier. Our results support the hypothesis that "selfish" defence against invasions is the major selective pressure that has led to the development and maintenance of pruning behaviour in weakly competitive plant-ants.     

Observations on the Ecology of Weaver Ants (Oecophylla smaragdina Fabricius) in a Thai Mangrove Ecosystem and Their Effect on Herbivory of Rhizophora mucronata Lam.

Ants of the genus Oecophylla are predators of other insects and are able to protect a variety of terrestrial plants against pest insects; however, observations on the ecology of these ants in mangrove forests are lacking. General observations on the ecology of Oecophylla smaragdina were carried out in a Thai mangrove forest to determine if these ants can protect their host plants in less favorable mangrove habitats. Leaf herbivory and the density of O. smaragdina ants were measured on Rhizophora mucronata trees at two sites. The results showed a negative correlation between ant density and herbivory. At both sites, the mean percent damaged leaf area was more than four times higher on trees without ants compared to “ant‐trees.” A significant negative correlation was found between tree mean percent leaf damage and the density of ants on the tree. Furthermore, on trees with ants, there was less herbivory on leaves close to ant nests compared to other leaves on the tree. Most damage was caused by chrysomelid beetles (62%) and sesarmid crabs (25%) and both types of herbivory were significantly reduced on ant‐trees.     

Arboreal thorn-dwelling ants coexisting on the savannah ant-plant, Vachellia erioloba, use domatia morphology to select nest sites

Nest site selection in arboreal, domatia-dwelling ants, particularly those coexisting on a single host plant, is little understood. To examine this phenomenon we studied the African savannah tree Vachellia erioloba, which hosts ants in swollen-thorn domatia. We found four ant species from different genera (Cataulacus intrudens, Tapinoma subtile, Tetraponera ambigua and an unidentified Crematogaster species). In contrast to other African ant plants, many V. erioloba trees (41 % in our survey) were simultaneously co-occupied by more than one ant species. Our study provides quantitative field data describing: (1) aspects of tree and domatia morphology relevant to supporting a community of mutualist ants, (2) how ant species occupancy varies with domatia morphology and (3) how ant colony size varies with domatia size and species. We found that Crematogaster sp. occupy the largest thorns, followed by C. intrudens, with T. subtile in the smallest thorns. Thorn age, as well as nest entrance hole size correlated closely with ant species occupant. These differing occupancy patterns may help to explain the unusual coexistence of three ant species on individual myrmecophytic trees. In all three common ant species, colony size, as measured by total number of ants, increased with domatia size. Additionally, domatia volume and species identity interact to predict ant numbers, suggesting differing responses between species to increased availability of nesting space. The proportion of total ants in nests that were immatures varied with thorn volume and species, highlighting the importance of domatia morphology in influencing colony structure.     

Experimental suppression of ants foraging on rainforest vegetation in New Guinea: testing methods for a whole‐forest manipulation of insect communities

1. Ants are extremely abundant in lowland tropical forests where they are important predators, plant mutualists, and herbivores. Their complex role in tropical plant–insect food webs can be best assessed by experimental manipulation of their abundance. Historically, ant exclusion experiments have had a small‐scale focus, such as single trees. Here, we test a new ‘whole‐forest' method of ant exclusion, using treated canopy bait stations, in a diverse tropical rainforest in New Guinea. 2. We conducted a 10‐month manipulative experiment in primary and secondary rainforests. In each forest type, a 625 m² treatment plot was isolated from the surrounding forest and 135 bait stations treated with fipronil, S‐methoprene, and hydramethylnon were placed in trees to suppress ants. Ant activity was monitored in the forest canopy and understorey with an additional 65 stations in treatment and control plots. 3. We achieved a dramatic decline in ant abundance in treatment plots compared with controls in both forest types, with an average decrease in ant numbers per station of 82.4% in primary and 91.2% in secondary forest. In particular, native dominant species Oecophylla smaragdina, Anonychomyrma cf. scrutator in primary forest, and invasive Anoplolepis gracilipes in secondary forest were greatly affected. In contrast, Tapinoma melanocephalum flourished in treatment plots, perhaps benefiting from reduced competition from other ant species. 4. Our study demonstrates that it is possible to selectively eradicate most of the foraging ants in a structurally complex tropical forest. We propose whole‐forest manipulation as a novel tool for studying the role of ants in shaping plant–insect food webs.