The role of fungus in the diet of the leaf-cutting ant Atta cephalotes (L.)

Abstract. 1. The workers and queen of the leaf-cutting ant Atta cephalotes fed on the juice of swollen hyphae (staphylae) produced by their cultivated fungus, but neither obtained sufficient energy from this source for their respiratory needs. The number of staphylae eaten by workers increased with worker size but was not enough to satisfy their energy requirements.
2. Larvae fed on whole staphylae and staphylae previously chewed by workers, and obtained sufficient energy from this source for respiration and growth. No evidence of feeding on fungus hyphae or of trophallaxis between worker and larvae was found. Larvae preferred staphylae to hyphae when fed them artificially and they gained more weight on the former.
3. Worker ants imbibed plant sap during the preparation of plant material for the fungus garden and the uptake of carbohydrate during this process was sufficient to supply their energy needs for approximately 24 h.
4. Staphylae were richer in lipid and carbohydrate, and poorer in protein than ant fungal hyphae.
5. The number of staphylae produced by the fungus gardens of two small nests was comparable with the observed consumption rate but would provide only about 4% of the nest's respiratory requirements.
6. In the light of these findings, a revised view of the role of the fungus in the diet of the ant is discussed.     

Ecdysteroid production during caste differentiation in larvae of the ant, Plagiolepis pygmaea

ABSTRACT. Variations in ecdysteroids were measured by radio-immunoassay in worker-biased and queen-biased larvae of the ant Plagiolepis pygmaea Latr. (Hymenoptera, Formicidae) during late larval development, i.e. from the end of winter diapause up to the prepupal period. At the end of diapause, larvae are bipotential and, depending on culture conditions, can become either queens or workers. Ecdysteroid profiles revealed that there are striking differences between the two castes: worker larvae showed high titres during their development, queen larvae had low titres over the same period.     

A duration constant for worker-to-larva trophallaxis in fire ants

Summary The allocation of liquid food by workers to larvae, a central process in ant biology, could be regulated by the frequency of trophallaxis, its duration, or both. In 4th-instar fire ant larvae, the duration of trophallaxis, bolus size, and the rate at which boluses were swallowed were all constant, indicating that the volume of food ingested during each worker-larva trophallaxis was both small and constant. Neither larval size over a 20-fold volume range nor larval starvation had a significant effect on duration of trophallaxis (mean = 11s, SD = 2s), bolus swallowing rate (mean = 2/s, SD = 0.5/s), or bolus volume (mean = 0.0675 nl, SD = 0.0002 nl, based on the assumption that the stomodaeum's epithelial layer is not expandable). Larval body orientation and larval location within the brood pile also had no effect on duration. Durations of trophallaxis by workers of different sizes were similar. Durations of trophallaxis for 1st-, 2nd-, and 3rd-instar worker larvae were also constant but greater than that for 4th-instar worker larvae. Fourth-instar minim larvae (from founding colonies) and 4th-instar worker larvae (from mature colonies) were fed for the same duration by workers but for different durations by founding queens. Founding queens fed minim larvae longer than they fed worker larvae. The durations of feedings to 4th-instar sexual larvae were more variable than those to worker larvae. Altogether, these findings indicated that 4th-instar worker larvae ingested a small, nearly constant volume of food (mean = 1.50 nl, SD = 0.005 nl) during each trophallactic event. Consequently, the long-term allocation of liquid food by workers to these larvae is regulated by the frequency of trophallaxis. Several other ant species showed a similar brevity and constancy in the duration of worker-larva trophallaxis. This brevity of worker-larva trophallaxis is in contrast to the duration of worker-worker trophallaxis.Although the duration of worker-larva trophallaxis appears to be determined by the worker, the data are not totally consistent with this interpretation.     

Cuticular body hairs mediate clumping of small Camponotus floridanus larvae

The body surface of insects usually carries cuticular hairs. Commonly, important functions of these structures are to prevent drowning and to defend against predators. Here, we report on our studies on hairs at the surface of larvae of the ant species Camponotus floridanus and Camponotus sericeiventris. First, we present data supporting the hypothesis that anti-drowning properties of the surface might rely on cuticular hairs. Second, we show that especially in young larvae body hairs serve as attachment and interlocking devices mediating clumping of larvae facilitating transport by workers. Based on our observations, we speculate that clumping also enhances larval perceptibility. Taken together, larval cuticular hairs seem to have at least two important functions augmenting chances of larval survival. Obviously, despite their immobility, young Camponotus larvae support childcare in the ant colony providing an arsenal of cuticular hairs on their body surface.     

Fungal hyphae as a source of nutrients for the leaf-cutting ant Atta sexdens

Abstract. The mutualistic fungus of leaf-cutting ants produces both ordinary hyphae and specialized ant rewards: the staphylae. Workers of Atta sexdens (L.) lived longer on diets which included nutritive staphylae than on those which provided only hyphae. However, hyphae were a better diet than sucrose solution or water alone. Small workers lived longer than large workers when receiving fungus garden, whether or not staphylae were available, probably because they are specialized to care for the garden and can exploit it. Large workers lived longer than small ones when only water was available. This may be a consequence of scale, larger workers containing proportionately more nutritional reserves in their bodies.
When starved workers were exposed to fungus for 3h, they gained weight. This weight gain represented the amount of food ingested during the test period. Workers gained 7.3 times more weight on natural fungus garden bearing staphylae than on garden with only hyphae. They gained only 1.5 times as much weight when the fungus was artificially cultured on agar. The results suggested that workers found hyphae attractive, but difficult to obtain in natural fungus gardens.
Material from hyphae and staphylae stained with a fluorescent dye was detected in worker crop contents using fluorescence microscopy. The crop contents of workers fed on staphylae fluoresced 1.14 times more than those of workers fed on hyphae.
Hyphae may provide a small source of food for workers, and the fungus as a whole may provide up to 9.0% of the respiratory energy requirements of workers, the remainder presumably being provided by plant sap.     

Behavioral and developmental homeostasis in the fire ant, Solenopsis invicta

The von Bertalanffy rule (1960) predicts that low incubation temperature during larval development will result in larger adult body size. If larval development in social insects followed this rule, then low incubation temperature would induce the development of larger workers and possibly even sexuals. To test this prediction, the effect of incubation temperature on larval development, larval meal size, larval tending and worker recruitment to food in the fire ant, Solenopsis invicta was investigated. Temperatures tested where within the range at which brood remains viable.Contrary to the predictions of the von Bertalanffy rule, worker size was unaffected by incubation temperature, and sexuals were reared at the high rather than the low incubation temperature. Moreover, larval meal size, the rate of larval tending by workers and the total number of workers recruited to food were unaffected by temperature.Mechanisms regulating developmental and behavioral homeostasis were as follows: the duration of larval development and the rate of larval growth changed proportionately with temperature such that the mean and variation of pupal size was unaffected by incubation temperature. Larvae solicited at the same rate, swallowed at the same rate and swallowed for the same duration such that meal size was unaffected by incubation temperature. On the brood pile, fewer workers tended brood at higher incubation temperatures, but worker tempo increased; as a result, brood tending was not adversely affected by incubation temperature. The rate of worker recruitment to food sites outside the nest increased with temperature, but the duration of the recruitment effort decreased such that, over time, the same total number of workers was employed to retrieve food.Incubation humidity was also investigated. When brood chambers were less than 100% humid, workers recruited to food and tended larvae (retrieved, sorted and groomed them), but did not feed larvae. Eventually, larvae died of starvation and were cannibalized.     

Larval isolation and brood care in Acromyrmex leaf-cutting ants

The larvae of leaf-cutting ants are maintained within the fungus gardens of their colonies and are fed pieces of fungus by the adult workers. However, little else is known about the nature of the worker-larva interaction in these ecologically important ants. To examine whether workers can gauge the needs of individual larvae, we isolated larvae without adult workers for different lengths of time. We then placed workers with the larvae and recorded the type and frequency of the subsequent behaviours of the workers. Workers scraped the mouthparts of larvae, ingested their faecal fluid, fed them with fungal hyphae, transported them around the fungus garden and, most frequently, licked their bodies. The workers were also observed to ‘plant’ fungal hyphae on the bodies of larvae. Workers interacted more frequently with larvae that had been isolated without workers than with those that had not, but there was no effect of the length of isolation. The results suggest that the interactions are complex, involving a number of behaviours that probably serve different functions, and that workers are to some extent able to assess the individual needs of larvae. Received 8 November 2004; revised 31 March 2005; accepted 22 April 2005.     

Influence of larvae and workers on egg production of queens of the pharaoh's ant,Monomorium pharaonis (L.)

Summary Social regulation of egg production and weight in queens was studied in relation to presence and absence of larvae and workers in the pharaoh's ant,Monomorium pharaonis (L.).Results were obtained by counting eggs and weighing queens under various conditions.The results confirm the existence of a positive feed-back loop between mated queens and their larvae as evident from a correlation (Y = 4.575 * X + 6.452) between the number of large worker larvae (X) and the egg yield (Y). This correlation seems to relate to the queens preferential feeding on larval secretions. Queens without larvae maintained a low level of egg production of about 6 eggs/day. Queens deprived of larvae as well as workers stopped producing eggs within 24 hours.Repletes, special workers, with greatly distended gasters functioned as a buffer retarding decline of egg production by feeding the queens during short periods without larvae.     

Interaction between the nematophagous fungus Duddingtonia flagrans and infective larvae of Haemonchus contortus (Nematoda: Trichostrongyloidea)

The interaction between Duddingtonia flagrans and infective larvae of Haemonchus contortus was studied in vitro under optical and scanning electron microscopy. Trap formation by the fungus started 9 hours after inoculation and first larvae were found 11 hours after larval inoculation on colonies grown on the surface of dialysis membranes. Scanning electron micrographs were taken 12, 24, 36 and 48 h after larval predation. Details of predation structures and fungus-larvae interaction are described. A mucilaginous substance occurred at the points of adherence of traps to nematode cuticle. Bacteria were also found at some points of interaction between fungus and larval cuticle. Cuticle penetration by fungus hyphae occurred only 48 h after predation.     

The role of leaf-cutting ant workers (Hymenoptera: Formicidae) in fungus garden maintenance

Abstract.

• 1 We studied the role of leaf-cutting ant workers (Atta sexdens (L.) in fungus garden maintenance, by temporarily excluding workers from the garden. This increased its subsequent attractiveness, as expressed by an increase in the numbers of workers licking it.
• 2 The length of free mycelia on areas of the garden from which workers were excluded increased but was reduced again when workers were returned. Workers therefore removed hyphae from the garden surface.
• 3 The maximum‘isolation effect’was obtained by preventing ant access for 2–3 days, after which the effect declined. Removing staphylae from portions of garden kept ant-free for 4 and 6 days restored the effect, as the ants were not distracted by harvesting staphylae. Portions of garden kept ant-free for longer than this were no more attractive than non-isolated control garden.
• 4 Workers were highly efficient in detecting and removing contaminants from their fungus garden. Samples of garden could be isolated from workers for up to 12 days before major growth of contaminants occurred, and this contrasted with the maximum of 6 days for the isolation effect on licking. The isolation effect was therefore not a response to contaminant growths on the garden.
• 5 Workers on the garden surface may remove hyphae for nutritional reasons, or to‘prune’their fungus and stimulate its growth. In either case, the result is a regulation of fungal growth.

     

Lethal infection thresholds of Paenibacillus larvae for honeybee drone and worker larvae (Apis mellifera)

We compared the mortality of honeybee (Apis mellifera) drone and worker larvae from a single queen under controlled in vitro conditions following infection with Paenibacillus larvae, a bacterium causing the brood disease American Foulbrood (AFB). We also determined absolute P. larvae cell numbers and lethal titres in deceased individuals of both sexes up to 8 days post infection using quantitative real‐time PCR (qPCR). Our results show that in drones the onset of infection induced mortality is delayed by 1 day, the cumulative mortality is reduced by 10% and P. larvae cell numbers are higher than in worker larvae. Since differences in bacterial cell titres between sexes can be explained by differences in body size, larval size appears to be a key parameter for a lethal threshold in AFB tolerance. Both means and variances for lethal thresholds are similar for drone and worker larvae suggesting that drone resistance phenotypes resemble those of related workers.     

Nutritional function of replete workers in the pharaoh's ant, Monomorium pharaonis (L.)

Summary: Queens of the pharaoh's ant Monomorium pharaonis (L.), like several other ant species, feed on larval secretions as their main nourishment and their fecundity is positively correlated with the number of large larvae present in the nest. The surplus of secretions produced by larvae is stored in a temporary caste of replete workers, which comprises young workers who remain in the nest and store liquid nourishment. Repletes are characterised by a conspicuously large gaster, caused by large amounts of liquid food stored in the crop, from which it may be regurgitated and distributed among colony members. In this study, repletes of pharaoh's ants were demonstrated to be functioning as buffers, smoothing fluctuations in availability of high quality food to the reproductive queens when larvae are scarce or missing, thus temporarily keeping up the egg production of queens.¶In undisturbed two-queen colonies with 20 large worker larvae and 30 workers (15 young and 15 old workers), approximately 10 repletes developed (one replete per two larvae). Development of older workers into repletes, when some or all repletes had been removed from the colonies, demonstrated that their temporal polyethism exhibits great plasticity in this trait.¶This study confirmed that, in pharaoh's ants, the regulation of fecundity depends not only on the food flow to the queen from larvae or from repletes but also on an unknown larval stimulus.¶The term crop repletes is suggested for replete workers which use their crop to store nourishment, as opposed to fat-body repletes, which store nourishment in their fat body.¶The presence of brood tending crop repletes in nests in several European ant species of Leptothorax, Myrmica, and Lasius, show that repletism is a common trait in ants, and that it may play an important role in regulation of nutrition in ant colonies, as demonstrated in Monomorium pharaonis.     

Ultrastructural observations of Beauveria bassiana infection in Xylotrechus rusticus larvae

In the present study we investigated the infection process of Beauveria bassiana on Xylotrechus rusticus larvae using scanning electron microscopy (SEM) and transmission electron microscopy (TEM). The SEM results showed that B. bassiana spores germinated on the surface of the larval body and invaded the larva as an appressorium. The hyphae then germinated from the spores and spread throughout the larval body. After the death of the larva, conidiophores formed at one end of the hypha on the surface of the larval body and prepared for a new round of infection. The TEM results showed severe damage to the larval cuticle after hyphae infection. The structure of the cuticle became thinner and eventually flocculent; muscle tissues were dissociated and eventually stuck to the hyphae, and the corpus adiposum was loose and deformed, and eventually degraded.     

Clash of kingdoms or why Drosophila larvae positively respond to fungal competitors

BACKGROUND: Competition with filamentous fungi has been demonstrated to be an important cause of mortality for the vast group of insects that depend on ephemeral resources (e.g. fruit, dung, carrion). Recent data suggest that the well-known aggregation of Drosophila larvae across decaying fruit yields a competitive advantage over mould, by which the larvae achieve a higher survival probability in larger groups compared with smaller ones. Feeding and locomotor behaviour of larger larval groups is assumed to cause disruption of fungal hyphae, leading to suppression of fungal growth, which in turn improves the chances of larval survival to the adult stage. Given the relationship between larval density, mould suppression and larval survival, the present study has tested whether fungal-infected food patches elicit communal foraging behaviour on mould-infected sites by which larvae might hamper mould growth more efficiently. RESULTS: Based on laboratory experiments in which Drosophila larvae were offered the choice between fungal-infected and uninfected food patches, larvae significantly aggregated on patches containing young fungal colonies. Grouping behaviour was also visible when larvae were offered only fungal-infected or only uninfected patches; however, larval aggregation was less strong under these conditions than in a heterogeneous environment (infected and uninfected patches). CONCLUSION: Because filamentous fungi can be deadly competitors for insect larvae on ephemeral resources, social attraction of Drosophila larvae to fungal-infected sites leading to suppression of mould growth may reflect an adaptive behavioural response that increases insect larval fitness and can thus be discussed as an anti-competitor behaviour. These observations support the hypothesis that adverse environmental conditions operate in favour of social behaviour. In a search for the underlying mechanisms of communal behaviour in Drosophila, this study highlights the necessity of investigating the role of inter-kingdom competition as a potential driving force in the evolution of spatial behaviour in insects.     

The organisation of larval feeding in bumblebees (Hymenoptera,Apidae) and its significance to caste differentiation

Summary. In pollen-storing bumblebees, the rate at which workers nourish larvae has been proposed to be the main factor influencing caste differentiation since workers feed prospective queens more frequently and longer than worker larvae during the last instars. In order to determine how the frequency of feedings is established small groups of Bombus terrestris larvae were either subjected to starvation or nourished regularly by workers. Experimentally starved larvae were fed significantly earlier and more often than control larvae. Behavioural observations provide evidence for the existence of a stimulus of larval origin that releases feeding behaviour in workers. Intentional inspections in the sense of a recognisable and functional behaviour intended to organise the feeding process were never observed. We argue that worker inspections are not required for the adequate maintenance of larvae and that a simple stimulus-response chain appears to be sufficient to regulate feeding behaviour at the individual and the colony level in bumblebees. Furthermore, hand-rearing experiments with female larvae indicate that queen rearing is not dependent on a high frequency of feedings in itself, nor is worker development induced by larval deprivation. This makes it unlikely that workers impose a caste-specific frequency of feedings on larvae in order to actively control or manipulate caste development. Since caste fate seems to be predetermined early in larval life, we propose that early caste-specific differences in development are reflected in the physiology of the larvae and the associated nutritional needs to which workers respond accordingly. Consequently, caste-specific differences in feeding frequencies are a result, but not the cause of differences in development.     

A comparison of honeybee (Apis mellifera) queen,worker and drone larvae by RNA‐Seq

Honeybees (Apis mellifera) have haplodiploid sex determination: males develop from unfertilized eggs and females develop from fertilized ones. The differences in larval food also determine the development of females. Here we compared the total somatic gene expression profiles of 2-day and 4-day-old drone, queen and worker larvae by RNASeq. The results from a co-expression network analysis on all expressed genes showed that 2-day-old drone and worker larvae were closer in gene expression profiles than 2-day-old queen larvae. This indicated that for young larvae (2-day-old) environmental factors such as larval diet have a greater effect on gene expression profiles than ploidy or sex determination. Drones had the most distinct gene expression profiles at the 4-day larval stage, suggesting that haploidy, or sex dramatically affects the gene expression of honeybee larvae. Drone larvae showed fewer differences in gene expression profiles at the 2-day and 4-day time points than the worker and queen larval comparisons (598 against 1190 and 1181), suggesting a different pattern of gene expression regulation during the larval development of haploid males compared to diploid females. This study indicates that early in development the queen caste has the most distinct gene expression profile, perhaps reflecting the very rapid growth and morphological specialization of this caste compared to workers and drones. Later in development the haploid male drones have the most distinct gene expression profile, perhaps reflecting the influence of ploidy or sex determination on gene expression.     

Social control of egg-laying rate in queens of the fire ant, Solenopsis invicta*

ABSTRACT Social control of egg-laying rate in queens of the fire ant (Solenopsis invicta Buren) was studied by experimental manipulation of the number of larvae, pupae and workers in colonies, and the age and size of larvae and workers. Workers and pupae do not stimulate oviposition by queens. The number of fourth instar larvae, on the other hand, bears a positive log-log relationship to the queen's egg-laying rate. Such larvae are needed both to stimulate and maintain oviposition. Their withdrawal results, within 48 h, in a decline in queen oviposition almost to zero. Their addition to broodless nests results in peak laying in about 4 days. Larvae in the first three stadia and early in the fourth stadium have a much lower effect upon queen fecundity. Sexual larvae are only c. 5% as stimulating on a weight basis, but equivalent on an individual basis. Several associated measures are positively correlated to egg-laying rate: weight of the queen, the number of her vitellogenic follicles per ovariole, total vitellogenic follicles, the time she spends feeding and (usually) the number of workers in the retinue that cares for her. The egg volume is negatively correlated with laying rate, so that queens lay more eggs for the same expenditure of material as laying rate increases. As body size of workers increases, they become less effective in transmitting the larval stimulation to the queen, but worker age has no effect on this ability. For a given number of larvae, queens in small, naturally growing colonies lay fewer, larger eggs than do queens in experimental colonies, but their fecundity increases more rapidly in relation to number of larvae. When larvae are fed vital-dyed food in one experimental colony, and then transferred to an undyed colony, the dye is rapidly transferred to worker crops, and into the queen's eggs, indicating bulk movement of material from larvae to workers to the queen and eggs. Large larvae are more effective at this than small larvae. Fourth instar larvae may be a digestive and metabolic caste that processes protein for egg production by the queen. If that is the case, the queen and fourth instar larvae are linked in a positive feedback loop. Either the logarithmic relation of fecundity to larval numbers or physical limits of the queen may set the maximum egg-laying rate, and thus determine maximum colony size. The data do not allow a clear choice between these alternatives.     

Absence of cuticular alkenes allows lycaenid larvae to avoid predation by Formica japonica ants

Chemical mimicry and camouflage based on cuticular hydrocarbons (CHCs) are adaptive strategies that are frequently observed in myrmecophilous insects. The larvae of several lycaenid butterfly species that exhibit obligate associations with specific ant species have been reported to use chemical mimicry. However, little is known about the strategies used by the larvae of species that have facultative associations with multiple ant species. We attempted to reveal the effects of larval CHC profiles on interactions with Formica japonica workers, using three lycaenid species, two facultative ant‐associated (Lycaeides argyrognomon and Zizeeria maha) and one non‐ant‐associated (Lycaena phlaeas), which commonly possess n‐alkanes as the major CHCs. In field bioassays, the lycaenid larvae were attacked by ant workers less often than larvae of Papilio polytes (Papilionidae), the CHCs of which were rich in 7‐alkenes. Treating the lycaenid larvae with 7‐heptacosene and 9‐heptacosene significantly activated ant aggression (biting), whereas treating them with n‐heptacosane, n‐octacosane and 13‐methylheptacosane had little effect. Furthermore, larvae of Pieris rapae (Pieridae), possessing n‐alkanes as the dominant CHCs, suffered an intermediate level of ant biting between the lycaenid and Pa. polytes larvae. However, treatments of the P. rapae larvae with 7‐heptacosene and 9‐heptacosene significantly affected the frequency of ant biting. These findings suggest that the absence of alkenes in larval CHC profiles is an effective means of circumventing predation by ants and allows lycaenid larvae to inhabit the foraging territory of predaceous ants, at least to some extent.     

Brood adoption in the leaf-cutting ant <Emphasis Type="Italic">Acromyrmex echinatior</Emphasis>: adaptation or recognition noise?

The ability to discriminate between nestmates and non-nestmates is an important prerequisite for the evolution of eusociality. Indeed, social insect workers are typically able to discriminate between nestmate and non-nestmate workers. Adult non-nestmate workers are readily detected and rejected from the colony. Whether social insects can discriminate between nestmate and non-nestmate brood, however, is less clear. Here, we show that workers of the leaf-cutting ant Acromyrmex echinatior discriminate between nestmate and non-nestmate brood, and among brood of different stages. Initially, non-nestmate brood is attacked, but it is adopted after a delay. Adoption could occur due to inefficiency of the recognition system, or it could be adaptive because it is an inexpensive way to increase the workforce. Our results suggest that brood adoption may occur accidentally. We also report how workers replace fungal hyphae on the brood’s surface before transporting the brood into their fungus garden.     

Larval cannibalism and worker-induced separation of larvae in <Emphasis Type="Italic">Hypoponera</Emphasis> ants: a case of conflict over caste determination?

Reproductive altruism and cooperative brood care are key characteristics of eusocial insects and reasons for their ecological success. Yet, Hymenopteran societies are also the stage for a multitude of intracolonial conflicts. Recently, a conflict between adult and larval colony members over caste fate was described and evidence for overt conflict was uncovered in several bee species. In theory, diploid larvae of many Hymenopteran species should experience strong fitness benefits, if they would be able to change their developmental pathway towards the queen caste. However, larval self-determination potential is low in most advanced eusocial Hymenopterans, because workers often control larval food intake and queenworker caste dimorphisms are generally high. In the ant genus Hypoponera, larvae actively feed on food provided by workers and here we show extremely low queenworker size differences in these ants: the lowest in H. opacior, where fertile wingless (intermorphic) queens weigh on average only 13% more than workers. Thus, slightly better nutrition during development might change the fate of a Hypoponera larva from a completely sterile worker to a fertile queen. One possibility to obtain extra food for Hypoponera larvae with their well-developed mandibles would be to cannibalise adjacent larvae. Indeed, we observed frequently larval cannibalism in ant nests. Yet, adult workers apparently try to prohibit larval cannibalism by carefully separating larvae in the nest. Larvae, which were experimentally brought into close contact, were rapidly set apart. Workers further sorted larvae according to size and responded swiftly to decreasing food levels, by increasing inter-larval distance. Still, an experimental manipulation of the larval cannibalism rate in H. schauinslandi failed to provide conclusive evidence for the link between larval cannibalism and caste development. Hence, further experiments are needed to determine whether the widespread larval cannibalism in Hypoponera and the untypical brood distribution can be explained by an overt caste conflict. Received 18 December 2006; revised 2 August and 20 September 2007; accepted 21 September 2007.