Similar yet different: differential response of a praying mantis to ant‐mimicking spiders

Batesian mimics typically dupe visual predators by resembling noxious or deadly model species. Ants are unpalatable and dangerous to many arthropod taxa, and are popular invertebrate models in mimicry studies. Ant mimicry by spiders, especially jumping spiders, has been studied and researchers have examined whether visual predators can distinguish between the ant model, spider mimic and spider non‐mimics. Tropical habitats harbour a diverse community of ants, their mimics and predators. In one such tripartite mimicry system, we investigated the response of an invertebrate visual predator, the ant‐mimicking praying mantis (Euantissa pulchra), to two related ant‐mimicking spider prey of the genus Myrmarachne, each closely mimicking its model ant species. We found that weaver ants (Oecophylla smaragdina) were much more aggressive than carpenter ants (Camponotus sericeus) towards the mantis. Additionally, mantids exhibited the same aversive response towards ants and their mimics. More importantly, mantids approached carpenter ant‐mimicking spiders significantly more than often that they approached weaver ant‐mimicking spiders. Thus, in this study, we show that an invertebrate predator, the praying mantis, can indeed discriminate between two closely related mimetic prey. The exact mechanism of the discrimination remains to be tested, but it is likely to depend on the level of mimetic accuracy by the spiders and on the aggressiveness of the ant model organism.     

Territories of the African Weaver Ant (Oecophylla longinoda [Latreille]); A Field Study

The African weaver ant (Oecophylla longinoda) is known to be highly aggressive toward conspecific aliens. In the study area Shimba Hills Reserve (Kenya) individual territories sometimes covered an area of up to approximately 1600 m², comprising 17 major trees. The territorial defense is organized by an elaborate defense recruitment system, which can also be aimed against several other ant species, which are potential competitors of Oecophylla for essential resources. This selective “enemy identification” seems to be the major behavioral mechanism by which the mosaic distribution of ecologically dominant ants is regulated.     

Territoriality in the Malaysian giant ant Camponotus gigas (Hymenoptera/Formicidae)

In a 5-ha area of primary lowland rain forest in Borneo, we observed four polydomous colonies of the night-active giant ant Camponotus gigas. The nonoverlapping, three-dimensional territories in the canopy had a ground size up to 0.8 ha. C. gigas showed a distinct territorial behavior: (1) specific “barrack” nests, especially containing many majors, were situated at the borders and were established during long-term territorial conflicts; (2) trunk trails were regularly patrolled by majors that attacked alien conspecifics and some other ant species violently; and (3) sentinels, often involved in long-enduring conflicts with neighboring ant colonies, defended the borders at bridgeheads. Interspecific conflicts with sympatric Camponotus species always led to violent, “bloody” fights of all castes. Intraspecific conflicts, however, were solved by ritual fights (“front leg boxing”) of majors. C. gigas performed a de-escalation strategy to end short periods of true intraspecific “ant war” that we provoked experimentally. Artificially induced ritualized combats continued over weeks also in the absence of baits, indicating that borders may become established by long-term conflicts of attrition. We discuss the differences between ritual fights in desert and rain-forest ants and apply Lanchester's theory of battles to our findings. Received: October 17, 2000 / Accepted: May 14, 2001     

中国蚁科研究的展望

<正> 蚂蚁是地球上数量最多、分布甚广、营群体生活的昆虫。据 B.Holldobler和 E.O.Wilson(1990)的统计,目前已定名的有8804种,如加上未定名的种类,估计在2万种以上。E.O.Wilson教授(1971)曾估计地球上每时每刻至少有10~(15)个蚂蚁在行走。因此,蚂蚁对人类的关系非常密切。     

Effect of supplementary feeding of Oecophylla longinoda on their abundance and predatory activities against cashew insect pests

Many studies have shown the efficiency of using weaver ants (Oecophylla species) as natural biocontrol agents against agricultural pests. Supplementary feeding could promote fast growth of this ant's population and discourage them from moving away. However, such artificial feeding might slow down ants search rates and in this way make them less efficient bio-agents. The experiments were conducted for two consecutive seasons at Naliendele Research Station. Cashew trees planted at a spacing of 12 m × 12 m in 2002 were used to investigate whether supplementary feeding could enhance foraging behaviour of Oecophylla longinoda. Fed O. longinoda colonies fed weaver ants (FWA) were supplemented with a 30% sugar solution and approximately 22 g of finely ground fish meat at two-week intervals while the unfed colonies unfed weaver ants (UWA) had access to only naturally occurring food sources. Weaver ant densities and pest damage were monitored fortnightly on newly damaged shoots, panicles and fruits and nut yields assessed after each harvest season. The results revealed that there was a significant difference (P < 0.05) with higher weaver ant densities in the FWA compared to UWA colonies and significantly lower (P < 0.05) pest damage levels were recorded on weaver ant treatments compared to plots without weaver ants. No significant differences (P > 0.05) in yields and mean damage levels were recorded between the two weaver ant treatments. Highest nut yield (4.22 ± 0.30 kg/tree and 5.37 ± 0.27 kg/tree) was recorded in the fed colonies, followed by non-fed colonies (4.20 ± 0.30 kg/tree and 4.88 ± 0.24 kg/tree) and the least (2.66 ± 0.19 kg/tree and 2.99 ± 0.19 kg/tree) was recorded from the untreated controls in 2012/2013 and 2013/2014, respectively. The studies indicated that supplementary feeding could boost weaver ants to higher population levels without reducing their effectiveness as biocontrol agents.     

Bottom-up control and co-occurrence in complex communities: honeydew and nectar determine a rainforest ant mosaic

Complex distribution patterns of species-rich insect communities in tropical rainforests have been intensively studied, and yet we know very little about processes that generate these patterns. We provide evidence for the key role of homopteran honeydew and plant nectar in structuring ant communities in an Australian tropical rainforest canopy and understorey. We also test the ant visitation of these resources against predictions derived from the 'ant-mosaic' hypothesis. Two ant species were highly dominant in terms of territorial behaviour and abundance: Oecophylla smaragdina and Anonychomyrma gilberti . Both dominant ant species monopolised large aggregations of honeydew-producing homopterans. Attended homopteran species were highly segregated between these two ant species. For the use of extrafloral and floral nectar (involving 43 ant species on 48 plant species), partitioning of ant species among plant species and between canopy and understorey was also significant, but less pronounced. In contrast to trophobioses, simultaneous co-occurrence of different nectar foraging ant species on the same plant individuals was frequent (23% of all surveys). While both dominant ant species were mutually exclusive on honeydew and nectar sources, co-occurrence with non-dominant ant species on nectaries was common. The proportion of visits with co-occurrences was low for dominant ants and high for many sub-ordinate species. These findings support the ant mosaic theory. The differential role of honeydew (as a specialised resource for dominant ants) and nectar (as an opportunistic resource for all ants including the co-occurring non-dominant species) provides a plausible structuring mechanism for the Australian canopy ant community studied.     

Comparing different methods for trapping mated queens of weaver ants (Oecophylla longinoda; Hymenoptera: Formicidae)

The predatory efficiency of African weaver ants Oecophylla longinoda and their utilisation in protein production is a function of ant abundance. Reliable control of insect pests in tropical crops is achieved when ant populations are constantly high. Transplanted populations of weaver ant colonies containing egg-laying queens are more stable than those without. Achieving such stability through collection of colonies established in the wild is usually difficult because of uncertainty in locating the nest containing the egg-laying queen. In this study, we investigated four methods that may be used to collect mated queens that subsequently can be used to stock ant nurseries. The catch efficiencies of (1) leaf traps, (2) paper traps (both types providing a refuge for founding queens), (3) random search for queens and (4) light trapping were compared. Light trapping was the most efficient way to collect queens followed by leaf traps, random search and, last, paper traps. Light trapping and random search, though, required the presence of a person throughout the ant's mating season (several months), whereas this was not required when using leaf and paper traps.     

Molecular phylogenies reveal host-specific divergence of Ophiocordyceps unilateralis sensu lato following its host ants

Ophiocordyceps unilateralis (Hypocreales, Ascomycetes) is an entomopathogenic fungus specific to formicine ants (Formicinae, Hymenoptera). Previous works have shown that the carpenter ant Camponotus leonardi acts as the principal host with occasional infections of ants from the genus Polyrhachis (sister genus of Camponotus). Observations were made on the permanent plots of Mo Singto, Khao Yai National Park of Thailand according to which O. unilateralis was found to occur predominantly on three host species: C. leonardi, C. saundersi and P. furcata. Molecular phylogenies of the elongation factor 1-α and β-Tubulin genes indicate a separation of O. unilateralis samples into three clades, reflecting specificity to each of the three different ant species. Samples collected from P. furcata and from C. leonardi were found to form sister groups with samples from C. saundersi forming an outgroup to the latter. Additional samples collected from unidentified ant species of Camponotus and Polyrhachis were positioned as outgroups to those samples on identified species. These results demonstrate that O. unilateralis is clearly not a single phylogenetic species and comprises at least three species that are specific to different host ant species. These cryptic species may arise through recent events of speciation driven by their specificity to host ant species.     

Highly similar microbial communities are shared among related and trophically similar ant species

Ants dominate many terrestrial ecosystems, yet we know little about their nutritional physiology and ecology. While traditionally viewed as predators and scavengers, recent isotopic studies revealed that many dominant ant species are functional herbivores. As with other insects with nitrogen-poor diets, it is hypothesized that these ants rely on symbiotic bacteria for nutritional supplementation. In this study, we used cloning and 16S sequencing to further characterize the bacterial flora of several herbivorous ants, while also examining the beta diversity of bacterial communities within and between ant species from different trophic levels. Through estimating phylogenetic overlap between these communities, we tested the hypothesis that ecologically or phylogenetically similar groups of ants harbor similar microbial flora. Our findings reveal: (i) clear differences in bacterial communities harbored by predatory and herbivorous ants; (ii) notable similarities among communities from distantly related herbivorous ants and (iii) similar communities shared by different predatory army ant species. Focusing on one herbivorous ant tribe, the Cephalotini, we detected five major bacterial taxa that likely represent the core microbiota. Metabolic functions of bacterial relatives suggest that these microbes may play roles in fixing, recycling, or upgrading nitrogen. Overall, our findings reveal that similar microbial communities are harbored by ants from similar trophic niches and, to a greater extent, by related ants from the same colonies, species, genera, and tribes. These trends hint at coevolved histories between ants and microbes, suggesting new possibilities for roles of bacteria in the evolution of both herbivores and carnivores from the ant family Formicidae.     

Ant mosaics occur in SE Asian oil palm plantation but not rain forest and are influenced by the presence of nest‐sites and non‐native species

Interaction networks within biotic communities can be dramatically altered by anthropogenic habitat modification. Ants, an important ecological group, often interact competitively to form mosaic‐like patterns in disturbed plantation habitats, in which dominant species form mutually exclusive territories. However, the existence of these ant mosaics in pristine forests is contentious. Here we assess the relative strengths of ant competitive interactions in oil palm plantation and primary rain forest in Sabah, Malaysia, using null models of species co‐occurrence. We use two metrics: the C‐score, which measures mean degree of overall co‐occurrence, and a novel metric, the C_var‐score, which measures the variance in degree of co‐occurrence. We also investigate the role of nest sites by collecting ants from canopy and leaf litter microhabitats, and from epiphytic ferns, an important nest site for canopy ants. Furthermore, we assess whether non‐native species, which were widespread in oil palm plantation (61 occurrences vs five in rain forest) are important in driving the formation of ant mosaics. We found no evidence for ant mosaics in any primary forest microhabitat. In oil palm plantation, segregation between species was pronounced in epiphytes, weak in the rest of the canopy and absent in leaf litter communities. Intriguingly, exclusion of non‐native ant species from analyses increased the degree of negative species co‐occurrence in all three microhabitats, with species segregation in the oil palm canopy becoming statistically significant. Our results suggest that invasion of plantation habitats by non‐native species does not drive increased species segregation in ant communities. Rather, high degrees of species segregation might relate to changes in the importance of canopy nest sites, with colonies competing more strongly for these in plantations. In primary forests, weaker nest‐site limitation and the highly complex, more vertically stratified, non‐uniform canopy could lead to random co‐occurrence between ant species at the scales studied here.     

Inter-specific aggression generates ant mosaics in canopies of primary tropical rainforest

The ant mosaic is a concept of the non-random spatial distribution of individual ant species in trees built upon the assumption of interspecific behavioural associations. However, colony identity and environmental variance may also play a role in species distribution. Here we assess the presence of ant mosaics in a primary forest ecosystem and whether they are structured by species' aggressive behaviours or by habitat filtering. We sampled arboreal ants from vertically stratified baits exposed in 225 canopy trees in a 9-ha plot of primary lowland forest in Papua New Guinea, the largest forest area surveyed to detect ant mosaics. We performed behavioural tests on conspecific ants from adjacent trees to determine the territories of individual colonies. We explored the environmental effects on the ant communities using information on the plot vegetation structure and topography. Furthermore, we created a novel statistical method to test for the community non-random spatial structure across the plot via spatial randomisation of individual colony territories. Finally, we linked spatial segregation among the four most common species to experimentally assessed rates of interspecies aggression. The ant communities comprised 57 species of highly variable abundance and vertical stratification. Ant community composition was spatially dependent, but it was not affected by tree species composition or canopy connectivity. Only local elevation had a significant but rather small effect. Individual colony territories ranged from one tree to 0.7 ha. Species were significantly over-dispersed, with their territory overlap significantly reduced. The level of aggression between pairs of the four most common species was positively correlated with their spatial segregation. Our study demonstrates the presence of ant mosaics in tropical pristine forest, which are maintained by interspecific aggression rather than habitat filtering, with vegetation structure having a rather small and indirect effect, probably linked to microclimate variability.     

Competition with insectivorous ants as a contributor to low songbird diversity at low elevations in the eastern Himalaya

Competitive interactions between distantly related clades could cause complementary diversity patterns of these clades over large spatial scales. One such example might be ants and birds in the eastern Himalaya; ants are very common at low elevations but almost absent at mid‐elevations where the abundance of other arthropods and insectivorous bird diversity peaks. Here, we ask if ants at low elevations could compete with birds for arthropod prey. Specifically, we studied the impact of the Asian weaver ant (Oecophylla smaragdina), a common aggressive ant at low elevations. Diet analysis using molecular methods demonstrate extensive diet overlap between weaver ants and songbirds at both low and mid‐elevations. Trees without weaver ants have greater non‐ant arthropod abundance and leaf damage. Experimental removal of weaver ants results in an increase in the abundance of non‐ant arthropods. Notably, numbers of Coleoptera and Lepidoptera were most affected by removal experiments and were prominent components of both bird and weaver ant diets. Our results suggest that songbirds and weaver ants might potentially compete with each other for arthropod prey at low elevations, thereby contributing to lower insectivorous bird diversity at low elevations in eastern Himalaya. Competition with ants may shape vertebrate diversity patterns across broad biodiversity gradients.     

Effects of an African weaver ant, Oecophylla longinoda, in controlling mango fruit flies (Diptera: Tephritidae) in Benin

Six mango, Mangifera indica L., plantations around Parakou, northern Benin, were sampled at 2-wk intervals for fruit fly damage from early April to late May in 2005. Mean damage ranged from 1 to 24% with a weaver ant, Oecophylla longinoda (Latreille), being either abundant or absent. The fruit fly complex is made up of Ceratitis spp. and Bactrocera invadens Drew et al., a new invasive species in West Africa. In 2006, Ceratitis spp. peaked twice in the late dry season in early April and early May, whereas B. invadens populations quickly increased at the onset of the rains, from mid-May onward. Exclusion experiments conducted in 2006 with 'Eldon', 'Kent', and 'Gouverneur' confirmed that at high ant abundance levels, Oecophylla significantly reduced fruit fly infestation. Although fruit fly control methods are still at an experimental stage in this part of the world, farmers who tolerated weaver ants in their orchard were rewarded by significantly better fruit quality. Conservation biological control with predatory ants such as Oecophylla in high-value tree crops has great potential for African and Asian farmers. Implications for international research for development at the Consultative Group on International Agricultural Research level are discussed.     

The effect of weaver ants Oecophylla smaragdina on the shoot borer Hypsipyla robusta on African mahoganies in Australia

• 1 African mahogany Khaya senegalensis is a high‐value timber tree species widely grown in central Africa, south‐east Asia and northern Australia. Pilot plantings show that the tree grows well in the wet‐dry tropical areas of northern Australia, and the shoot borer Hypsipyla robusta (Lepidoptera: Pyralidae) is a potential pest of the tree. The weaver ant Oecophylla smaragdina is an efficient biocontrol agent in some horticulture crops. To investigate whether the ants control shoot borers, field experiments were conducted at two sites near Darwin, Australia from April 2006 to January 2009.
• 2 In the weaver ant treatments, the overall percentage of trees damaged by shoot borers was 0–2.7% at Berrimah Farm and 0–4.2% at Howard Springs, and the damaged trees were attacked once only. In the treatments without weaver ants, however, 9.9–52.1% trees were damaged at Berrimah Farm, and 6.3–64.6% at Howard Springs, and the damaged trees were generally attacked more than once.
• 3 At both sites, significantly fewer trees on each monitoring occasion were damaged in weaver ant treatments than in treatments without weaver ants.
• 4 The mean percentage of overall flushing shoots damaged by the pest at both sites was significantly lower in weaver ant treatments compared with treatments without weaver ants.
• 5 Fewer shoots were damaged per damaged tree in weaver ant treatments compared with treatments without weaver ants.
• 6 The data obtained suggest that weaver ants were effective biological control agents of the shoot borer, and that the ants can be used to manage the pest on African mahogany trees.

     

Urea in Weaver Ant Feces: Quantification and Investigation of the Uptake and Translocation of Urea in <Emphasis Type="Italic">Coffea arabica</Emphasis>

Weaver ants are tropical insects that nest in tree canopies, and for centuries these ants have been used for pest control in tropical orchards. Trees hosting weaver ants might benefit not only from the pest protective properties of these insects but also an additional supply of nutrients from ant feces deposited on the leaves. In a recent study, we demonstrated that Coffea arabica plants hosting Oecophylla smaragdina weaver ants under laboratory conditions experienced enhanced nitrogen availability compared with plants grown without ants. Therefore, the aim of the present study was to further investigate the interactions of weaver ants with the host plants with respect to plant nutrition. Here, we report the identification and quantification of urea, a highly effective foliar nutrient present in the fecal depositions of O. smaragdina. Feces samples obtained from six O. smaragdina colonies were analyzed, and urea concentrations ranging from 1.98 to 31.05 μg/mg ant feces were detected. Subsequently, we investigated the uptake and translocation of ¹⁵N₂-urea in amounts corresponding to the estimated urea contribution via feces depositions on single host plant leaves under laboratory conditions. The results clearly demonstrated that fecal urea was not only assimilated but also translocated within the plant. This evidence strongly supports the hypothesis that the fecal urea of weaver ants is a source of nitrogen for the host trees. Thus, weaver ant feces likely contribute to an improved nutritional status of ant-hosting trees in tropical orchards, thereby adding value to the use of weaver ants for the biocontrol of insect pests.     

Nest and colony-specific spectra in the weaver ant Oecophylla smaragdina

Animals in social groups need to differentiate between group members and others. In very large groups, such as those formed by many ant species, it is not possible to rely on individually specific cues to identify colonymates. Instead, recognition must be based on the colony-specific cues. Individual ant colonies tend to have a specific chemical gestalt that is maintained by the continual exchange of chemicals between workers. In very large polydomous colonies, the exchange of chemicals may be limited between nests within the colony, resulting in inter-nest variation in colony odour that might hinder identification of colonymates or conspecific intruders. We used near-infrared spectroscopy to explore variation in the chemical profile between and within colonies of the weaver ant Oecophylla smaragdina. We found that differences between colonies were reflected in the position, amplitude and width of spectral peaks, while differences between nests within colonies were reflected mainly in amplitude. Furthermore, in the context of colonymate recognition, the behaviour of the ants themselves was positively correlated with colony-specific spectral characteristics, rather than with nest-specific characteristics. Thus, colony spectra have features that are not obscured by intra-colonial variation and may potentially encode the chemical characteristics used by workers to identify colonymates.     

Spatial Variation in Army Ant Swarm Raiding and its Potential Effect on Biodiversity

Invertebrate communities of the tropical rain forest floor are highly diverse, characterized by patchy species distribution patterns and high variation in species density. Spatial variation in the foraging activity of swarm raiding army ants, prime invertebrate predators in tropical rain forests, is discussed as a mechanism contributing to these patterns, but highly resolved long‐term data on army ant raiding on the local and landscape scale are hitherto lacking. In this study, 196 positions in 11 study sites in a tropical rain forest in western Kenya were continuously monitored over ~4 mo for the occurrence of swarm raids of army ants. Using population simulation analyses, the consequences of army ant raiding for prey communities were assessed. We found an unexpectedly high variation in raid rates at the study site and landscape scale. The weekly chance of communities to become raided by army ants was on average 0.11, but ranged from 0 to 0.50 among the 196 positions. Simulating population developments of two Lotka–Volterra species—showing slight trade‐offs between competitive strength and resistance to army ant raids—in the real raiding landscapes showed that the observed spatial variation in raid rates may produce high prey diversity at larger spatial scales (due to high β‐diversity) and strong variation in species density. Our results indicate that high spatial variation in army ant swarm raiding is a mechanism capable of generating patchy species distribution patterns and maintaining the high biodiversity of invertebrate communities of the tropical rain forest floor.     

Diversity of entomopathogenic fungi near leaf-cutting ant nests in a neotropical forest, with particular reference to Metarhizium anisopliae var. anisopliae

We investigated the prevalence of entomopathogenic fungi associated with leaf-cutting ant colonies in a small area of tropical forest in Panama. There was a high abundance of Metarhizium anisopliae var. anisopliae near the colonies. Beauveria bassiana was also detected in the soil, Aspergillus flavus in dump material, and six Camponotus atriceps ants were found infected with Cordyceps sp. Based on a partial sequence of the IGS region, almost all of the M. anisopliae var. anisopliae isolates fell within one of the three main clades of M. anisopliae var. anisopliae, but with there still being considerable diversity within this clade. The vast majority of leaf-cutting ants collected were not infected by any entomopathogenic fungi. While leaf-cutting ants at this site must, therefore, regularly come into contact with a diversity of entomopathogenic fungi, they do not appear to be normally infected by them.     

Entedoninae wasps (Hymenoptera, Chalcidoidea, Eulophidae) associated with ants (Hymenoptera, Formicidae) in tropical America, with new species and notes on their biology

Three new species of Eulophidae associated, or presumed to be associated with ants are described: two species of Horismenus Walker and one species of Microdonophagus Schauff. Information on the biology is also included. The two Horismenus species are from Chiapas, Mexico. Horismenus myrmecophagussp. n. is known only from females and is a gregarious endoparasitoid in larvae of the weaver ant Camponotus sp. ca. textor. The parasitoids pupate inside the host larva, and an average of 6.7 individuals develops per host. This is the second time a species of genus Horismenus is found parasitizing the brood of a formicine ant of genus Camponotus. Horismenus microdonophagussp. n. is described from both males and females, and is a gregarious endoparasitoid attacking the larvae of Microdon sp. (Diptera: Syrphidae), a predator on ant brood found in nests of Camponotus sp. ca. textor. The new species of Microdonophagus, Microdonophagus tertius, is from Costa Rica, and known only from the female. Nothing is known about its biology but since another species in same genus, Microdonophagus woodleyi Schauff, is associated with ants through its host, Microdon larva (with same biology as Horismenus microdonophagus), it is possible that also Microdonophagus tertius has this association. A new distributional record for Microdonophagus woodleyi is also reported, extending its distribution from Panama and Colombia to Brazil.     

Detection of Mitochondrial COII DNA Sequences in Ant Guts as a Method for Assessing Termite Predation by Ants

Termites and ants contribute more to animal biomass in tropical rain forests than any other single group and perform vital ecosystem functions. Although ants prey on termites, at the community level the linkage between these groups is poorly understood. Thus, assessing the distribution and specificity of ant termitophagy is of considerable interest. We describe an approach for quantifying ant-termite food webs by sequencing termite DNA (cytochrome c oxidase subunit II, COII) from ant guts and apply this to a soil-dwelling ant community from tropical rain forest in Gabon. We extracted DNA from 215 ants from 15 species. Of these, 17.2 % of individuals had termite DNA in their guts, with BLAST analysis confirming the identity of 34.1 % of these termites to family level or better. Although ant species varied in detection of termite DNA, ranging from 63 % (5/7; Camponotus sp. 1) to 0 % (0/7; Ponera sp. 1), there was no evidence (with small sample sizes) for heterogeneity in termite consumption across ant taxa, and no evidence for species-specific ant-termite predation. In all three ant species with identifiable termite DNA in multiple individuals, multiple termite species were represented. Furthermore, the two termite species that were detected on multiple occasions in ant guts were in both cases found in multiple ant species, suggesting that ant-termite food webs are not strongly compartmentalised. However, two ant species were found to consume only Anoplotermes-group termites, indicating possible predatory specialisation at a higher taxonomic level. Using a laboratory feeding test, we were able to detect termite COII sequences in ant guts up to 2 h after feeding, indicating that our method only detects recent feeding events. Our data provide tentative support for the hypothesis that unspecialised termite predation by ants is widespread and highlight the use of molecular approaches for future studies of ant-termite food webs.