* Departamento de Biología de Organismos, Universidad Simón Bolivar, Apartado 80659, Caracas 1080, Venezuela.




Ambas especies tienen habitos de forrajeo diurnos. Conornyrma sp es monogínica y construye nidos en la arena hasta unos 50 cm de profundidad consistentes de numerosas cámaras pequeñas. A. foreli es poligínica y anida en ramas huecas en árboles. Experimentos de laboratorio revelaron lo siguiente:a.- Ambas especies utilizan reclutamiento químico masivo con una feromona de camino producida por una glándula abdominal que informa sobre la presencia y localidad del alimento. La concentración de la feromona regula la intensidad del reclutamiento. La duración de la actividad de la feromona es de mas de una hora para Aforeli y menos de 30 min. para Conomyrma sp. b.- Los territorios son reconocidos en forma colonia-específica, gracias principalmente a señales químicas. A. foreli marca su territorio con secreciones del tórax, mientras que Conomyrma sp. lo hace con secreciones del abdomen. La duración de la actividad de la marca es de cerca de una hora para A. foreli y de unos 30 min. para Corlomyrma sp. c.- Compañeras de nido son reconocidas gracias a un olor colonia específico, absorbido sobre la cutícula. El olor es producido en la región cefálica, probablemente por la glándula mandibular. d.- Ambas especies producen feromonas de alarma cefálicas y abdominales. Los resultados se comparan con los sistemas de comunicación de otras hormigas y se discuten algunos aspectos ecoetológicos.


Both species have diurnal foraging habits. Conomyrma sp is monogynic and build nests with many small chambers up to 50 cm deep in the sand. A. foreli is poligynic and uses holes in trees as nest. Experiments in the laboratory revealed the following:

a.- Both species use chemical mass recruitment with a trail pheromone from an abdominal gland which gives information about the presence and localization of food. Its concentration regulates the amount of recruitment. The fade out time of a trail is over one hour for A. foreli and less than 30 min for Conomyrma sp.

b.- The ants have a territorial behaviour in which territories are recognized colony-specifically, mainly by chemical cues. A. foreli marks its territories with a pheromone from a thoracic gland, whereas Conomyrma sp. does it with a pheromone from a gland in the "aster. The fade-out time of the pheromone of A. foreli is about 1 h whereas for Canomyrma sp. it is about 30 min.

c.- Nestmates are recognized by colony specific odour, which can be adsorbed on the cuticle of the insects. The odour is produced in the head, probably by the mandibular gland.

d.- Both ants produce cephalic and abdominal alarm pheromones.

Results are compared to the communication systems of other Formicidae and some ecoetilological aspects are discussed.


Azteca spp. are one of the most common ants in the neotropics. They are reported to form close association with plants, protecting them from herbivores whereas the plant provide them with shelters and/or food (WHEELER 1942). Besides casual observations, little is known about their behaviour. Even less is known about the behaviour of Conomyrma spp., in spite of the fact that they seem to be common in the neotropics.

Few works on recruitment and alarm-defense systems, mainly on Tapinama spp. and Iridomyrmex spp. (CAVILL & PHYLLIS 1965, TRAVE & PAVAN 1956, SZLEP & JACOBI 1967, COURET & PASSERA 1979, ROBERTSON et al., 1980) tell us about the behavior of the ants of the subfamily Dolichoderinae, in spite of the fact that they are one the most important members of neotropical ecosystems, competing in importance, regarding biomass, with the Myrmicinae.

In this work we make a preliminary survey of the more conspicuous communication systems and characterize in more detail the recruitment system and the agonistic communication systems of two Dolichoderinae ants. The results are compared to the behavior known for ants from other subfamilies.


Nests of Conomyrma sp. (types of which were deposited at Museo de Entomología, Instituto de Zoología Agrícola, U.C.V., Maracay, Venezuela) were excavated from the beach of Higuerote, whereas nests A. foreli were collected from dry cocoa pods in Caucagua, Estado Miranda, Colonies were established in the laboratory at 25.6 ° C and 40-70% RH, using natural illumination. The size of the colonies varied between 500 and 10,000 individuals. Twenty-two colonies of each species were kept in plastic containers (50 x 50 cm). Ants were offered a plaster of Paris nest, cast a 12 cm diameter Petri dish. This nest had excavated chambers which were covered with a glass plate and a dark piece of paper. Ants occupied these nests with their brood and queen. They were offered water, honey and dead crickets. Details of experimental procedures will be given together with the results.


General Observations:

Conomyrma sp. is monomorphic and builds its nests in the sand, next to the roots of trees and bushes. The main activity observed in the field was foraging on extrafloral nectaries of mangrove trees. The nests had 3 to 5 entrances, separated 50 to 60 cm from each other, interconnected through a complex system of galleries (2-4 mm internal diameter) which lead to several oval chambers (up to 20) 10 to 50 cm. deep. Sexual brood were always found in separate chambers. Colonies in the field produced elates during February and March. A. foreli, which is also monomorphic, uses natural cavities of trees to establish puts of their colony. Trails run through the tree and between trees, from one cavity to another, were the brood is kept. Dry cocoa pods are nearly always occupied by this ant. The dry seeds of the fruits are perforated with 2 to 3 small holes and used by the ants to keep pupae. We found up to 7 physogastric queens in one pod. Sexuals and sexual brood occupied separate pods. In the lab. queens were found always in the same chamber. Field colonies produced elates all year around, with an increased elate production between April and June.

The daily foraging rhythm, measured as the number of ants on a 90 cm2 area on the foraging table of the laboratory colonies, was counted once each 60 min. Colonies were kept with food at libitum beginning 24 h prior to the counting. Results showed that A. foreli had its foraging activity mainly during the day with a peak of activity two hours before sunset. At night, activity levels dropped to about 15% of that of peak hours. Conomyrma sp. had a more uniform foraging activity pattern. The activity peaked at sunset and never dropped below 30% of the peak value during the 24 h. cycle.

Recruitment to food:

In both species, workers finding the food returned to the nest dipping the gaster on the ground. A few seconds afterward, nestmates emerged from the nest and walked to the food following the route the first scout made for homing. No leader ant was seen guiding the ant s to the food. Removal of the first ants in the column did not interrupt recruitment. Thus, chemical mass recruitment was suspected to be used for recruitment to food by both species (JAFFE et al., 1985).

Bioassay I: Food was offered to the colony on a specific place on the foraging table. The number of ants going to the food and the number returning to the nest was counted during I min intervals for 60 min. Time zero was the moment when food was made available to the colony. The data obtained served to graphically (Fig. 1) determine the rapid increase phase (see also JAFFE & HOWSE, 1979), and to calculate the parameters on Table 1.

Results: Data on Table I suggest that both species use chemical mass recruitment (WILSON 1962, JAFFE 1980). Both regulate the amount of nestmates recruited in accordance to food quality and/or the length of the previous starvation period. That is, more workers are recruited to the food (greater Max. Nr. of ants recruited) or they are recruited faster (steeper slope of the recruitment curve) when the starvation period was longer or if the food quality was better.

Conomyrma sp. differs from A. foreli in that the recruitment curve of the former showed two distinct phases (Fig. I ). For Conomyrma sp. the slope of the first phase, but not that of the second phase (Fig. 1, Table I), differed statistically when the food-nest distance was varied. Close observation showed that during the first phase, the recruited ants came mainly from the foraging area, whereas during the second phase, the main body of workers came from the nest. In the case of A. Morels ants from the nest participated in recruitment from the beginning.

Exocrine glands producing the trail pheromone:

In order to find the source of the trail pheromone, the following bioassay was developed:

Bioassay II: Whatmann 42, 4.25 cm filter papers, were placed at the nest entrance for 48 h. A. thinned Pasteur pipes was used to make a 3.5 cm diameter circle on the paper with one ant equivalent (the extract of the glands of one ant in 0.5 ml methanol) of the different extracts. The number of ants leaving the nest, and the number of them following the artificial trail was counted during I min intervals, as well as the number of consecutive 1/4 segments of the circumference followed by those ants following the artificial trail.

Results: Data on table II showed that in both species, the trail pheromone comes from the gaster. It is probably produced by the Pavan's gland as in other dolichoderine (PAVAN & RONCHETTI 1955, WILSON & PAVAN 1959, WILSON 1965, 1AFFE 1984). Gaster extracts of each species have an attractive effect on nestmates and orientate them at the same time. No interspecific trail following between both species was observed.

Additional tests showed that higher concentrations of gaster extract recruited a greater number of workers. No saturation effect or repulsive effect of concentrated gaster extracts could be detected for the concentration range tested (0.5 to 5 ant equivalents/paper). Thus, the trail pheromone gives information about the presence of food, its spatial position and its quality, i.e. pheromone concentration regulates amount of recruitment.

The duration of the orientation effect of the trail pheromone, if applied as extract, is about one hour in A. foreli and less than 30 min Conomyrma sp. (Table III).

Territorial behaviour:

When a worker al any of the two species was placed on the foraging area of a different conspecific colony, an increase in the speed of movements of this intruder could be observed. As the foreign ant was often but not always attacked, Bioassay III was developed in order to test the responses of ants to new and to alien conspecific territories and to study the reaction of workers to different types of territories.

Bioassay III: Two Whatmann 42, 4.25 cm diameter filter papers (A and B in Table IV), were placed for 48 h on the foraging table, next to the nest entrance of the colonies. For the lest, both filter papers ("simulated territories") were removed, and 10 min. later (unless stated otherwise in Table IV) were replaced by two other papers. These papers could be new (N), the same as the ones just removed (C), or they could come from a different nonspecific colony (D), which had explored them for at least 48 h. Five min after placing the papers in the arena the number of ants on the two papers was counted during 1 min.

When testing extracts for territorial activity, the papers were impregnated methanol extracts (E) of one ant equivalent of the respective body part, or with pure solvent (S), and left at room temperature for 5 min. (or more, see Table II) in order to allow the solvent to evaporate. The papers were then tested as describe above.

Results: The results (Table IV) show that workers of both species distinguish between a territory previously explored by themselves, and a new territory or one previously explored by another nonspecific colony. Conomyrma sp. was also able to distinguish between a new paper and one previously explored by another nonspecific colony.

Experiments using extracts (Table V) suggest that the recognition signal is produced by the ants and thus, that they mark their territory with a pheromone. The possible source of the territorial pheromone seems to be different for the two species. A. foreli seems to produce the pheromone from a thoracic gland, whereas Conomyrma sp. seems to produce it from an abdominal gland. The effect of a territory marked with extracts of these body parts, was equivalent to that of territories previously explored by them. as measured by Bioassay III. No dissection of glands was possible due to the small size of the workers.

Experiments with Conomyrma sp. showed that filter papers, placed in the foraging arena near a permanent food source, were treated differently by the ants than papers placed in the arena, in places far from the nest entrance and from the food (p < 0.01, Mann-Whitney's U- test). The latter were treated like new filter papers (p = 0.443, Wilcoxon's matched pairs test ). No differences in the response to simulated territories placed on different parts of the foraging arena could be observed with A. foreli.

When two Conomyrma sp. workers from different colonies were placed in a Petri dish outside their colonies on a filter paper from Bioassay III, they showed aggression. The ant initiating the fight was in 11 out of 12 cases the ant on its own territory, i.e. on the paper previously explored by its colony (p < 0.008, Sign test). When testing papers with extracts in the same way, the number of tests in which fights were initiated by an ant on its own territory: by an ant on a foreign territory; and in which there was no tight, were as follow: Head extract, 3:5:4; thorax, 2:3:7; gaster 6:1:5. The only statistically meaningful ratio is the one for the gaster extract (p = 0.06, Sign test). Similarly, for A. foreli for 15 replicates, the number of fights initiated by the ant on its own territory was 7, and 2 for the foreign ant whereas on a new paper, the same workers initiated only 2 fights each. This difference is highly significant (p = 0.001, Fisher's exact probability test).

Nestmate recognition:

Bioassay IV: Two workers drawn at random, one from the nest where the test was performed (C, for control), and the other from a different nonspecific colony (E, for experimental), were picked up with deodorized plastic forceps and placed anywhere on the foraging area of the control colony. The behaviour of the resident workers toward the intruders was observed and recorded on a tape recorder for later quantification.

Results: When workers were placed on the foraging area of a foreign colony, they were frequently attacked. Not all the intruders were attacked to the same extent, but E ants were always grasped by workers and carried to the borders of the foraging arena (Table VI, live ants).

No colony specific brood recognition (Nymphal stage) was detected for either of the species, as the time of removal, time of the duration of antennation by workers, numbers of workers attracted and number of workers alarmed was the same for nestmates and intruders (p > 0.1 in all cases, Wilcoxon's matched pairs test).

Foreign Conomyrma sp. queens and males were attacked more vigorously than workers. They were also hindered from reentering their own nest, although not attacked by their nestmates once they left the nest. A. foreli a polygynic species, attacked foreign queens only if the colony had an active queen. Queenless colonies accepted in 4 out of 5 cases a foreign queen. That is, the queen was seen alive for at least 6 weeks after introduction to the colony.

Recognition Signals:

Bioassay V: Dead workers or parts of them were placed near the nest entrance on the foraging area of the colonies. Always two workers or parts were presented simultaneously to the colonies. One of them originated from a nestmate (C) whereas the other came from a different nonspecific calmly (E). The time resident workers took before removing each of the presented objects was measured for A. foreli. For Conomyrma sp. the occurrence or not of removal, during a 10 min. observation period was noted.

Dead workers were freeze dried by applying pressures of 0.001 mm Hg at -20° C to them using a vacuum pump. Freeze dried workers were impregnated with odors from different body parts in the following way: Freeze dried ants were placed together with two freshly crushed body parts of workers from a specific colony in a 2 ml glass vial. Care was taken in order to avoid direct contact between the freeze dried ant and the crushed body part. The vial was sealed and left at room temperature for 30 min. After this period, the freeze dried ant was taken out and used for tests.

Results: The thin a dead ant remained on the foraging area before h was picked up, was different for foreign nonspecific ants compared to nestmates, when both wale presented simultaneously (Table VI). Only in the case of freeze dried ants, was the probability of removal ova resident suit the same as for a foreign ant. These experiments indicated that freeze drying removed the nestmate recognition signal. This signal was present only in the head of the workers. Other body parts were not recognized colony-specifically. When a freeze dried A. foreli ant was impregnated with vapors from head, thorax or gaster, and assayed, the recognition signal reappeared on the freeze-dried ant, only in the case of vapors from the head, reproducing the colony-specific behaviour.

These experiments indicate the presence a colony-specific odour, which has to be a complex of volatile substances in order to evaporate and impregnate a freeze dried ant colony-specifically. The source of the odour is in both cases a cephalic gland.

Alarm pheromones:

Bioassay VI: One body part (head, thorax or gaster) was crushed over the center of a t) cm diameter Whatmann 42 filter paper. The paper had concentric circles, one cm apart, drawn with a pen, in order to facilitate observation. The paper was placed in the foraging arena of a colony and the behaviour of the workers was observed during 10 min. and recorded on a tape recorder for posterior quantification Eight replicate observations for each experimental situation were perfumed.

Results: For Conomyrma sp., alarm behaviour, characterized by last movements of the workers and lifting of the head and antennae, was induced by crushed heads and abdomens. Crushed thoraces produced no visible reaction by the workers. The crushed heads produced alarm for 3 min. (range 2-3 min.) and to distances up to 6 cm. from the source. After this time an increased exploration of the paper was observed for up to 9 min. Crushed abdomens induced alarm for only 0.5 min (range 0.4-0.55 min) at distances of up to 2 cm, but attracted ants, i.e. orientated to the center of the paper, for up to 15 min.

A. foreli also showed alarm behaviour toward crushed heads and abdomens, but not toward thoraces. Alarm behaviour in this species was characterized by fast movements of the workers and lifting of the gaster. Heads induced alarm at distances from the source of up to 1 cm. for 10 min. (range 8-10 min.), Gasters alarmed workers at distances up to 2 cm. for 15 min. (range 10-16 min.). Heads induced orientation toward the center of the paper but gasters did not.


The communication system used for recruitment to food of the two Dolichoderinae species studied is very similar. Both have colony- specific recognition systems, probably based on inter-colony differences of the different components of the cephalic alarm pheromones. as is the case in the advanced Myrmicinae (JAFFE 1983) and Formicinae (JAFFE & SANCHEZ 1984).

Territorial Behaviour

Both have territorial marking behaviour fulfilling the criteria for a territorial pheromone (JAFFE & PUCHE 1984), which are:

1.- Workers secrete a pheromone on the territory.

2.- Workers recognize the territorial mark colony- specifically.

3.- Workers on territories with the pheromonal mark from its colony have some kind of advantage, compared to foreign nonspecific workers on the same territory.

4.- A marked area should be respected by conspecifics. The end result should be a reduction in aggressive interactions between colonies, because colonies are spaced out more than would be expected from a random occupation of suitable habitats.

Criteria 4 is a sufficient but not necessary condition for territory. I.e., criteria 3 could be fulfilled without fulfillment of criteria 4.

The presence of territorial marking behaviour seems to he a characteristic shared only by advanced Formicidae (HOLLDOBLER & WILSON 1977, JAFFE & PUCHE 1984, JAFFE & SANCHEZ, 1984, JAFFE, 1987).

The two species of ants use different glands to produce their territorial pheromone. This supports previous suggestions (JAFFE &c PUCHE 1984, JAFFE & SANCHEZ 1984, JAFFE 1987) that territorial marking behaviour is a relatively recent feature in ant phylogeny. Different ant species have different territorial pheromone sources, even if they belong to the seine subfamily, contrary to what occurs with trail pheromones, in wich the same gland is used by a broad spectrum of related species (JAFFE 1984). However, we can not exclude other glands as sources of territorial pheromones, as our bioassays were performed in the lab. and as different markers could be used by the species to mark different types of territories.

Nestmate Recognition:

The nestmate recognition system seems to be independent of the number of queens in the colony. The polygynic species A. foreli has the same recognition mechanism as monogynic species such as Conomyrma sp. or Atta spp. (JAFFE 19X3). This contradicts the suggestion of CROZIER & DIN (1979) and HOLLDOBLER & MICHENER (1980), who proposed that polygynic species should use a kind of gestalt odour in order to recognize their nestmates. This is certainly not the case for Aforeli where only the cephalic alarm pheromone is used as a recognition signal. Chemical composition of pheromones are very probably genetically fixed, and learning of the odour of the colony (i.e. the odour of the nestmates) is required in order to recognize nestmates

Alarm Communication:

A. foreli seems to have a more complex alarm communication system comparad to Conomyrma sp. as both, cephalic and abdominal secretions, complement each other. The cephalic pheromone orientates the, ants but has a short-lasting effect, whereas the abdominal alarm pheromone is long lasting but does not orientate the workers toward the source of emission. In Conomyrma sp. the role of the abdominal alarm pheromone is not clear. Its short-lasting effect makes it doubtful to be a true alarm pheromone. The attractive effect produced by the crushed gasters may be due to the trail pheromone. In any case, a more detailed study of the alarm behaviour of these species is needed.

Recruitment to food.

Also the recruitment system of these Dolichoderinae seems to be fairly advanced. That is, the chemical mass recruitment system (WILSON 1962, JAFFE 1980) found is similar to that which had been described for the Dolichoderinae Iridomyrmex humilis (See Robertson et al., 1980), which has been classified as of the more advanced among the Formicidae (JAFFE 1984). As in the majority of ants, the Dolichoderinae studied seem to use the democratic decision making system during chemical mass recruitment (JAFFE et al, 1985).

Ethoecological Considerations:

The two species differ in the fade out time of the different pheromones. A. foreli has longer lasting trail, alarm and territorial pheromones, compared to Conomyrma sp. This is likely to be related to the different habitats they exploit. A. fin elf lives exclusively on trees. whereas Conomyrma sp. is a terrestrial ant, although foraging also on trees. A. foreli exploits mainly spatially and temporally predictable resources (Aphid colonies and extrafloral nectaries), whereas Conomyrma sp. seems to be a more generalized forager, collecting dead arthropods. and extrafloral nectar of mangrove trees and food residues left behind by beach tourists. A. foreli seems to have a more developed territorial behaviour comparad to Conomyrma sp. Evidence for this is the fact that Conomyrma sp. does not advertise all their foraging arena in the laboratory with territorial pheromone, as A. foreli does. It is known that A. foreli is a dominant ant in cocoa plantation ecosystems, excluding nearly all other ant species from the trees they inhabit (JAFFE et al, 1986). Thus, territorial behaviour seconds to be related to the type of resources exploited by the species, i.e. each species seems to optimize energy expenditure in marking their territory.

Tabla 1


Tablas 2 y 3


Tabla 3


Tablas 5 y 6


Tabla 6


Thanks are given to Alberto Mona for his technical assistance, to Dr. C. Bosque for critically reading an early draft of the manuscript, and to CONICIT, Venezuela for financial support through the grant Sl-1336. Azteca foreli was identified by John Lattke and Conomyrma sp. by the late Dr. W.F. Buren, who thought it was probably an as yet undescribed Conomyrma species, close to Corurmyrrna biconnis.

Fig 1
Fig 1
Figure 1. Recruitment curves obtained with Bioassay I.


BOSSERT, W.H. & WILSON, E.O. 1963. The analysis of olfactory communication among animals. J. theor. Biol., 5, 443-469.

CAVILL, G.W.K. & PHYLLIS, L.R. 1965. Ant venoms, attractants and repellents. Science, 149, 1337-1345.

COIJRET, H. & PASSERA, L. 1979. Food-trail laying hi 3 species of Dolichoderinae ants. Ann. Sci. Nat. Biol. Anim., 1, 39- 48.

CROZIER, R.H. & DIX, M. W. 1979. Analysis of two genetic models for the innate components of colony odor in social hymenoptera. Behav. Ecol. Sociobiol., 4, 217-224.

HOLLDOHLER, B. & MICHENER, C.1). 19X(). Mechanisms of identification arid discrimination in social hymenoptera. In Evolution of Social Behaviour, Markt ed. Weinheim, Verlag Chemie Gmbh, pp. 35-58.

HOLLDOHLER, B. & WILSON E.O. 1977. Colony specific territorial pheromone in the African weaver ant. Pro. Nat. Acad. Sci. USA., 74, 2072-2075.

JAILING, K. 1980. A theoretical analysis of the communication system or chemical mass recruitment in ants. J. Theor. Biol., 84, 587-609.

JAFFE, K. 1983. Chemical communication among workers of leaf-culling ants Social Insects in the Tropics, Jaisson Ed., Univ. Paris Vol 2, 165- 180.

JAFFE, K. 1984. Negentropy and the evolution of chemical recruitment in auks J. Theor. Biol., 106, 587-604.

JAFFE, K. 1987. The evolution of territorial behaviour and nestmate recognition systems in ants. Experientia supp. Vol 54, 295-311.

JAFFE, K. & HOWSE, P.E. 1979. The mass recruitment system of the leaf cutting ant Atta cephalotes. Anim. Behav., 27, 930- 939.

JAFFE, K & PUCHE, H. 1984. Territorial marking with metapleural gland secretion in the ant geminate. J. Insect Physiol., 30, 265-270.

JAFFE, K. & SANCHEZ, C.1984. Nestmate recognition and territorial behaviour in the ant Campanotus ruppes. Insectes Sociaux, 31, 302-315.

JAFFE, K., SANCHEZ, P. & TABLANTE, A. 1986. Ecología de Formicidae en plantaciones de cacao en Barlovento, Venezuela. Rev. Theobroma 16.

JAFFE, K., VILLEGAS, C;., COLMENARES, D., PUCHE, H., ZABALA, N., ALVAREZ, M., NAVARRO, J.G. & PINO, E. 1985. Two different decision making systems in recruitment to food in ant societies. Behaviour 92,9-21.

PAVAN, M. & RONCHETTI, G. 1955. Studi sulfa morfología externa e anatomia intema dell'operaria di Iridomyrmex hurnilis e richerche chimichi e biologiche sulla iridomirmecina. Atte Soe. Ital. Scien. nat. Milano, 94, 379 477.

ROBERTSON, P., DIIDZINXT(Y, M. & MORTON. C. 1980. Exocrine gland involvement in trailing behaviour in the Argentine ant. Anim. Behav., 28, 1255-1273.

SZLEP, R. & JACOBI, T. 1967. The mechanism of recruitment to mass foraging of Monomorium venustum, M. subopacum, Tapinoma israelis T. simonthy. Insectes Sociaux., 14, 25-40.

TRAVE, R. & PAVAN, M. 1956. Veleni degli Insectti principe estratti della hormici Tapinoma nigerrium. Chimica e l'Industria, 3X, 1015-10.

WHEELER,W.M.1942. Studies of neotropical ant-plants and their ants. Bul. Museum Comp. Zoology. Harvad, 90, 1-262.

WILSON, E.O. 1962. Chemical commucation among workers of the fire ant Solenopsis saevissima. Anim. Behav., 10, 134-147.

WILSON, E. O. 1965. Chemical commucation in social insects. Science, 149, 1064-1071.

WILSON, E.O. & PAVAN, M. 1959, Glandular sources and specificity of some chemical releasers of social behaviour in Dolichoderinae ants. Psyche, 66, 70-76.