I studied the coordination of collective behavior in
seven species of the genus Leptogenys, based on the communication
between the individuals of a colony. Influences of environmental factors
were also considered, so that the coordination of behavior could be
fully understood in the light of a species ecology.
The communication systems:
In all Leptogenys-species several pheromones,
located in the pygidial and the poison gland, were involved in trail
communication. In six species with army ant behavior (L. distinguenda,
L. mutabilis, L. borneensis, L. myops, L.
sp. 2 and L. sp. 3), pygidial gland components functioned in
the initial forming of foraging formations in the absence of prey. These
pheromones were also responsible for swarm cohesion during its further
progress. "Slow" recruitments, which played a role in the exploration
of new terrain, are coordinated with pygidial gland contents through
summation of pheromone concentrations in the trails, supported by a
medium-termed persistence. Furthermore, the ants used these substances
as orientation cues to locate their nest. Furthermore, emigrations were
guided exclusively by pygidial gland secretions (see below).
In species with army ant behavior, poison gland contents
released intense prey recruitment and induced aggressiveness. Quick
attacks were coordinated with signals of this gland,
enabling the ants to overwhelm even large and mobile prey objects. The
recruitment trails by these pheromones could sum up and become strong
enough to influence the spatial development of entire raid formations.
Furthermore, raid swarms reached a considerable flexibility to switch
to new environmental conditions due to the extraordinary short persistence
of this excitement pheromone. A second compound of the poison gland
played an additional role in the cohesion of the swarm. This pheromone
probably provides an extra orientation cue during and after prey-attack,
when the pygidial gland might not find use.
In L. diminuta, a species not showing army ant
behavior, the same pheromone glands were used in a different way. Pygidial
gland contents were used in orientation and prey recruitment. The poison
gland, however, had its primarily function in the orientation of scout
ants. A second component released excitement in raid-groups and is assumed
to coordinate the synchronized attack of prey.
In addition to chemical trails, which are by far the
most important orientation cues in army ants, optical cues played a
minor role in the orientation of L. distinguenda, although this
species is strictly nocturnal. Within the extensive network of trails,
individual ants integrate optical information to distinguish mayor directions.
The role of individual behavior:
In L. distinguenda, the most comprehensively
studied species, individuals did not react equally to identical stimuli.
Individual motivation was crucial for the behavioral response observed.
The motivation itself depended not only on the age of workers, which
led to temporal polyethism, but also on the current status of the colony
like nutrition, humidity, temperature etc. as well as the actual task
performed by an individual. Workers on their way to the swarm front,
for instance, were highly sensitive to poison gland extracts if the
colony was starved. This pheromone leads ants which are motivated to
hunt to the actual raiding front. On the contrary, the same pheromone
is ignored by ants retrieving booty to the nest. These ants reacted
concentration-dependent to pygidial gland extracts, which guides them
safely to the nestsite. Finally, workers participating in emigrations
did not ignore rather than avoid poison gland extracts. This behavior
enables the strict separation of raiding and emigration activities,
both of which are often performed synchronously in this ant species.
If, however, a colony was overfed, the reactions on pheromones were
generally low and no raid formations were formed at all.
Different ways of orientation on individual level were
also found in L. diminuta, a member of the genus displaying scout-induced
group raiding. Only high concentrations of both poison and pygidial
gland extracts led entire raid-groups safely to a prey object. However,
when the trails were low concentrated (or aged), they were followed
only by the scouts, which took over the leadership of the group under
these conditions. On their way back to the nest, all ants followed all
trail concentrations equally.
Pheromone properties and ecology:
Physical properties of pheromone components were also
investigated and discussed in an ecological context. Comparing different
species revealed that physical differences of components correlated
with differences in ecology. Without exception, the persistences of
pheromone trails were short in species with army ant behavior (poison
max. 5 min and pygidial max. 30 min). The high dynamics of raiding formations
typical for these ants are clearly based on the use of volatile communication
substances. On the contrary, trails of L. diminuta persist considerably
longer (poison max. 24 h and pygidial max. 1,5 h). Singly foraging scouts
are able to orientate on these stable trails during long lasting foraging
Some of the ecological differences are not correlated
to different pheromone properties. The structure of raiding formations
e.g. is highly dependent on the type of prey and its distribution in
space. Species which are partly or completely specialized in locally
concentrated food resources like social insect colonies (e.g. L.
mutabilis), forage preferably in columns in contrast to swarm hunters,
which accept a wide range of arthropod prey or even small vertebrates
(e.g. L. distinguenda). In this respect, the spatial development
of army ant raids is self-organizing due to interactions between individual
ants on the one hand and their environment on the one hand. An army
ant colony can be viewed as a complex adaptive system with regard to
its adaptation to outer influences, its self-organizing structure and
the ability to change system properties through evolution. I visualized
the raiding and emigration behavior of army ants as a complex system
in form of a schematic illustration. Furthermore, I modeled a possible
evolutionary scenario of the main communication systems and the corresponding
predatory behavior in the genus Leptogenys. This approach contributes
to hypotheses existing in literature.
Characteristics of army ant
By use of a theoretical model, I studied the role of
different pheromone components for the recruitment efficiency of L.
distinguenda. In this model, a communication-system of at least
two pheromone signals meets the requirement for a trail system like
it is found in army ants best. Such a system includes not only a highly
efficient recruitment but also a remarkable flexibility to switch behavior
according to changing conditions. Other models of swarming systems in
ants based on one single pheromone cannot satisfactorily explain this
In a comprehensive survey including numerous ant species,
I established basic principles, which are important for the coordination
of army ant behavior. Some behavioral patterns, which are displayed
by army ants, are also found in ants of other ecotypes and only the
combination of all patterns is unique to army ants. The classical definition
of the army ant syndrome is not completely unambiguous, resulting in
a confusion in the literature with ant species, which do not show true
army ant behavior. For this reason I suggest a modified and unequivocal
definition: The expressions "group-predation" or "group-raiding" should
not be used anymore for army ant like hunting strategies, rather than
exclusively for scout-induced foraging. If raiding behavior of army
ants is consequently referred to as "mass-predation" or "mass-raiding",
any misunderstanding is excluded. Furthermore, these definitions refer
better to the communication systems realized in each group and consequently
transfer the expressions "group-recruitment" and "mass-recruitment"
to the corresponding foraging behavior.
I also took the interesting biology of myrmecophiles
into consideration, which are extraordinarily abundant in army ants,
by examining their integration into the host colonies. This was done
with special regard to the chemical communication. In L. distinguenda,
part of the guest fauna is able to detect pheromone trails of their
hosts and follow emigrations actively (springtails, a bristletail, a
spider, staphylinid beetles, phorid flies). Other species are not able
to detect any pheromone trails but they nevertheless participate in
emigrations, riding on the brood which is carried by workers to the
new nest (Acari, a woodlouse, nitidulid beetles). Regularly, part of
these myrmecophiles stay behind when they miss an emigration of their
Bonabeau, E., 1998. Social insect colonies as complex
adaptive systems. Ecosystems 1: 437-443.
Dooley, K., 1996. A nominal definition of complex adaptive
systems. The Chaos Network 8: 2-3.
Gell-Mann, M., 1994. The Quark and the Jaguar.
Freeman & Co, New York, NY. 392 pp.
Gotwald, W. H. J., 1982. Army Ants. In: Social
insects (H. R. Hermann, Ed.), Academic Press, New York. pp. 157-254.
Schneirla, T. C., 1971. Army Ants: A Study in Social
Organization. W. H. Freeman, San Francisco. 349 pp.
Wilson, E. O., 1958. The beginnings of nomadic and group-predatory
behavior in the ponerine ants. Evolution 12: 24-31.
Wilson, E. O., 1971. The Insect Societies. The
Belknap Press of Harvard University Press, Cambridge, Mass. 548 pp.