Current Biology Vol 16 No 15
R570
social relationships or dominance
styles to variation in food
resource characteristics, such
as the distribution, abundance
and quality of food resources. If
resources are of high quality and
spatially clumped, for example,
they are monopolizable by an
individual, which usually leads
to a dominance effect on food
intake, because higher ranking
individuals have preferred access.
In this case, it will pay to be of
high rank and maintain despotic
dominance relationships. More
egalitarian relationships are
expected, for example, if low
quality resources are more evenly
scattered in the environment and
hence fighting for access is not
worth the effort.
Given the lack of long-term
field studies on macaque
species -- with only very few
examples -- a thorough test
across species is still pending.
An alternative model, the
phylogenetic hypothesis, on
the contrary proposes that the
variation in social relationships
is a consequence of phylogeny,
with more closely related species
showing more similar social
patterns. So far, three broad
species groups have been
identified, the silenus-sylvanus
group, which presumably
is the oldest lineage, the
sinica-arctoides group and
the fascicularis group, which
presumably includes all the nasty,
despotic species. Cross-species
comparison of captive macaques
indeed shows evidence for
conservative traits, such as rank
acquisition or dominance relations
between the sexes, which suggest
at least some phylogenetic inertia.
The possible influence of ecology,
however, has not been tested in
these studies on captive groups.
So the debate continues.
Is there a macaque genome
project? After the chimpanzee,
rhesus macaques are the
second non-human primate
whose genome sequence now is
available. A multi-centered team
just recently deposited a draft
version into databases accessible
to the public. The rhesus genome
shares about 92­95% of its
sequence with that of humans
and more than 98% with the
chimpanzee genome.
Where can I find out more?
Fa, J.E., and Lindberg, G.E. (1996). Evolution
and Ecology of Macaque Societies.
(Cambridge, Cambridge University Press).
Barr, C.S., Newman, T.K., Shannon, C.,
Parker, C., Dvoskin, R.L., Becker, M.L.,
Schwandt, M., Chmapoux, M., Lesch,
K.P., Goldman, D., et al. (2004). Rearing
condition and rh5-HTTLPR interact to
influence limbic-hypothalamic-pituary-
adrenal axis response to stress in infant
rhesus macaques. Biol. Psychiatry 55,
733­738.
Thierry, B., Singh, M., and Kaumanns, W.
(2004). Macaque Societies ­ A Model
for the Study of Social Organization.
(Cambridge, Cambridge University Press).
Huffman, M.A. (1996). Acquisition of
innovative cultural behaviors in
nonhuman primates: a case study of
stone handling, a socially transmitted
behaviour in Japanese macaques. In
Social Learning in Animals, C.M. Heyes
and B.G. Galef, Jr., eds. (New York:
Academic Press), pp. 267­289.
Max Planck Institute for Evolutionary
Anthropology, Deutscher Platz 6, 04103
Leipzig, Germany. E-mail: ostner@eva.
mpg.de
Male stumptail macaques from Thailand. Despite a clear dominance hierarchy stump-
tail macaques are tolerant and reconcile frequently after conflicts. (Photos by Oliver
Schülke, MPI-EVA.)
Communication in
ants
Duncan E. Jackson1 and Francis
L.W. Ratnieks2
Social insects have several
advantages over solitary insects.
The presence of many individuals
can increase system reliability,
and work can also be organized
more efficiently through division
of labour and task partitioning.
Another advantage is the possibility
of sharing information, especially
communicating where food
can be found. But research is
increasingly showing that foraging
communication does more than
merely direct nestmates to food.
It also allows the colony to regulate
total foraging activity, to retain a
memory of previously rewarding
locations, and to select among
locations of different profitability.
In this primer, we first provide a
brief historical perspective, then
focus on recent research that has
uncovered remarkable richness
and sophistication in ant foraging
communication, and finally identify
some key questions for further
research.
The study of foraging
communication in social insects
has a long history. In the 1880s the
eminent Victorian John Lubbock
(Baron Avebury) showed that
ants used odour trails in foraging.
His contemporary Wassmann
even believed that ants had a
sophisticated language encoded
by antennal tapping, somewhat
like Morse code. Far-fetched as
Wassmann's idea may seem, the
subsequent discovery by Karl von
Frisch that honeybee foragers use
waggle dances to communicate
both direction and distance of food
sources showed sophistication in
communication that seemed barely
credible for an animal, let alone an
insect. Von Frisch went on to win
the 1973 Nobel Prize for physiology
or medicine for this discovery.
Research into chemical
communication developed
rapidly in the 1960s following
the identification of the first two
Primer
Magazine
R571
pheromones: queen substance
in the honeybee and the male
attractant of female silk moths.
Investigation of numerous ant
species demonstrated that a
wide range of chemicals are
used to mark pheromone trails
and are produced by several
different glands. Research
published by E.O. Wilson in 1962
demonstrated that ant pheromone
trails provide positive and negative
feedback to organise foraging at
the colony level. A colony forms
a trail when successful foragers
deposit pheromone on their return
to the nest, with the trail gaining
in strength as more and more
workers add pheromone to it, so
providing positive feedback. The
trail decays when the food runs
out because foragers refrain from
reinforcing it on their return and the
existing pheromone evaporates, so
providing negative feedback.
In the 1980s the new field of
self-organisation adopted ant
pheromone trails as a paradigm
to illustrate emergent processes,
where the activities of many
`agents' responding only to local
information leads to a global
adaptive process. Mathematical
and computational models showed
how worker ants, which were
credited with minimal individual
intelligence, could work together to
solve problems such as selecting
the shorter of two paths between
food and nest, or selecting
the better food source when
presented with two of differing
quality. These models showed
that adaptive global solutions can
arise in a system with a single trail
pheromone providing positive
feedback. From a biological
perspective, however, this may
have oversimplified things. Most
trail-using ants employ multiple trail
pheromones secreted from one or
more glands. But why use many
pheromones if one is apparently
sufficient? Recent research into
the roles of multiple pheromones
has uncovered remarkable
sophistication of communication
in ant foraging trail networks. In
Pharaoh's ants, for example, a suite
of trail pheromones complement
each other by providing a
long-term memory of previously
used trails, short-term attraction
to currently rewarding trails, and
a `no entry' signal to unrewarding
branches. The geometry of the trail
system also provides information,
and there is worker specialization in
trail laying and detection.
Foraging communication
Natural selection will favour
communication if it helps
nestmates to forage more
efficiently. In social insects, workers
collect the food for the colony.
So if worker A helps worker B to
collect more food, this is as good
to worker A as if she collected
it herself, because the food is
brought back to the same nest to
feed the same larvae. Nevertheless,
many species of social insects do
not share foraging information. In
some cases this may be because
foragers have no useful information
to share. For example, desert ants
(Cataglyphis spp.) collect dead
insects, but there would be little
point in directing nest-mates to
the site of a discovery if no food
remains. Communication is most
useful when food resources are
found that are larger than can
be exploited by a single forager,
or that need defending. Large or
renewable feeding sites would
be well worth communicating to
nest-mates, such as the location
of a group of aphids secreting
honeydew or a patch of flowers.
In a general sense, social
insect colonies live in a dynamic,
competitive environment in which
food sources of variable quality are
constantly changing in location.
Most ant species are dependent
upon ephemeral food finds. In
such an environment, there is an
advantage to sharing information
if it can help the colony direct its
workers quickly to the best food
sources. Persistent or recurring
food sources may also be available,
such as the aphids or scale insects
`farmed' by many ant species. The
best strategy is often to remember
rewarding foraging sites but also to
be flexible enough to exploit newly
discovered food and to select the
better sources from those available.
To this end, information directing
nestmates to food also enables
them to select the highest quality
food find when multiple resources
are available.
Different ant species employ a
range of communication methods
for directing nestmates to foraging
sites. The simplest is `tandem
running', where a successful forager
leads a recruit. Recruitment is faster
when the successful forager leads
a group of recruits. The recruit
or recruits follow the leader by
physical contact or pheromone
from the leader. The most
spectacular use of trail pheromones
is in mass foraging. Here the
recruitment and guiding aspects of
foraging communication are usually
decoupled. The pheromone trail
provides only the route to food,
whilst recruitment of additional
foragers is caused by other
behaviours, such as dances or
direct physical contact in the nest.
In honeybees, the waggle dance
recruits additional foragers but also
directs them to the food. However,
Figure 1. In the Malaysian ponerine army ant, Leptogenys distinguenda, emigration
takes place along a pheromone trail to a new temporary nest site.
Workers communicate the initiation of emigration with audible `clicking' sounds made
by rattling their mandibles against each other, the sound resonating on dry leaves.
(Image courtesy of Alex Wild.)
Current Biology Vol 16 No 15
R572
honeybees have another dance,
the vibratory signal, which helps
recruit more foragers but does not
guide them to food. Decoupling
means that mass foraging ants
broadcast guidance information
widely, potentially to all foragers, in
the form of a trail network marked
with varying amounts and types
of pheromone. In contrast, the
broadcast range of the honeybee
waggle dance is limited to workers
in contact with the dancer.
Multi-pheromone trails
Ant pheromone trails contain many
chemicals that differ greatly in their
persistence. Trail pheromones
are also secreted from a diverse
range of glandular sources, such
as the Dufour's gland, poison
gland, anal glands, glands on the
feet, and glands on the thorax
or abdomen. The use of multiple
trail pheromones by a single ant
species means that foraging
communication can be more
complex than is possible with a
single pheromone.
Many foraging insects, for
example a worker honeybee,
can individually remember where
they have foraged and can return
to rewarding sites. However, for
trail-following ants this memory
need not be an individual memory
encoded in the brain. Instead, it
can be a group memory encoded
externally in the pheromone trail
system. The use of several trail
pheromones that differ in their
persistence provides memory over
differing time scales. In particular,
a non-volatile pheromone can
provide a longer-term memory,
while a volatile pheromone can
allow rapid choice among potential
feeding locations by quickly
`forgetting' depleted locations.
The traditional view of ant
pheromone trails as short-lived
signals designed for rapid effect
is often illustrated by the swarm
raids of army ants. Raiding army
ants certainly use short-lived
trails to coordinate their lightning
raids. But recent research has
detected a more complex array of
pheromone signals. For example,
in the Malaysian ponerine army ant,
Leptogenys distinguenda (Figure 1),
distinct roles have been assigned
to trail pheromones from two
glands (poison and pygidial).
Temporal and spatial variation in
the use of three trail pheromones
communicates context-specific
information in directing and
organizing raids. The poison
gland of L. distinguenda contains
two pheromone components.
One elicits a strong short-term
attraction to prey items. The other
guides workers from foraging sites
to the colony, but only weakly. The
prey-attraction component directs
more ants to prey encountered
during raiding to ensure that
the prey is swiftly overwhelmed.
The number of foragers attracted is
a non-linear function of pheromone
concentration, such that a trail
laid by just a few ants leads to a
rapid increase in workers attacking
the prey. In this way a small
number of workers encountering
prey can rapidly attract enough
nestmates to capture the prey. This
prey-attraction pheromone is highly
volatile and lasts only 5 minutes,
ensuring that ants are not attracted
long after the prey item has been
captured.
In contrast, the pygidial gland
of L. distinguenda produces a
longer-lasting trail pheromone
(approximately 25 minutes). When
attacking prey, workers often
become detached from the trail
network and this pheromone
guides them back to the trail, or
the colony. The pygidial gland is
responsible for maintaining the
spatial organisation of raiding
ants, helping them explore
the environment for prey in a
systematic manner. Raiding parties
advance in a single direction on the
trail, only departing when locating
prey or when signalled to do so by
the poison gland pheromone. Thus,
the longer-lived trail pheromone
forms a well-connected network
from which all raiding excursions
are made. The trail network ensures
rapid and reliable communication
between foragers and enables the
rapid transport of prey items back
to the colony.
Our second example is the
Pharaoh's ant, Monomorium
pharaonis (Figure 2). Pharaoh's
ants are common in human
habitations and are an introduced
pest worldwide. They are
generalist foragers, taking a wide
range of foods. Recent research
has shown distinct short-lived
and long-lived attractive trail
pheromone effects, and also a
short-lived repellent pheromone
effect. The short-lived attractive
trail pheromone (approximately 20
minutes) is used to guide foragers
to currently rewarding feeding
sites. Again, like L. distinguenda,
Pharaoh's ants also make use
of a longer-lived trail network to
organise foraging and maintain
foraging cohesion. In contrast
to L. distinguenda, however, the
long-lived trails of Pharaoh's ants
can persist for several days. The
long-lived pheromone means that
the trail network can be explored
from day to day. Sections of the
network leading to food can be
reinforced with the short-lived
trail pheromone. The negative
pheromone is placed locally in the
network, immediately after trail
bifurcations on the non-rewarding
branch. These three effects
seem to have complementary
Figure 2. Pharaoh's ants, Monomorium pharaonis, form branching networks of phero-
mone trails.
Here the network has been formed on a smoked glass surface to aid visualisation.
(Image courtesy of Duncan Jackson.)
Magazine
R573
roles. The long-lived attractive
pheromone is a memory. The
short-lived attractive pheromone
marks out routes to current food
sources. The short-lived repellent
pheromone is a `no entry' signal to
unrewarding branches in rewarding
trail sections. There are probably
additional complementary effects.
For example, the presence of both
attractive and repellent short-lived
trail pheromones may help ants
chose the more rewarding branch
at a trail bifurcation, and may
also allow more rapid changes
in directing foragers to particular
locations.
Caste-specific communication
Division of labour, in which different
workers do different tasks, is
universal in insect societies. For
example, some workers forage and
others nurse the brood. Within the
foragers there is also specialization.
In the honeybee, most foragers
collect nectar but some specialize
in collecting pollen, water or tree
resin. Most are guided by waggle
dances to known food sources
but some scout out new sources.
Recent research shows that
individual specializations also exist
in relation to ant pheromone trails.
In Pharaoh's ants, only workers
that walk with their antennae
in contact with the substrate
can detect the long-lived trail
pheromone. Although it is not
possible to individually mark
Pharaoh's ant workers because
they are so small (body length
approximately 2 mm), ants that are
individually confined for several
hours show consistent behaviour
with approximately 17% being able
to detect a previously established
trail that has been unused for 24
hours. These `pathfinder' ants
are probably a behaviourally
specialized sub-caste of foragers
that help re-establish existing trails.
That is, they convert a long-term
memory into a more easily detected
signal. In addition, approximately
40% of the Pharaoh's ant foragers
on an active trail make repeated
U-turns. They walk with their sting
extended indicating that they are
maintaining the trail by laying
additional pheromone. Thus, in
Pharaoh's ant trail networks there
is specialization for both laying and
detecting trail pheromones.
Pharaoh's ant workers are all of
the same size. But some cases of
individual differences in relation to
trail pheromones involve different
size castes. In the physically
dimorphic Pheidole embolopyx the
minor workers specialize in laying
trail pheromone (from their poison
gland), but both major and minor
workers follow trails in foraging.
Major workers do not lay trails but
do most of the food transporting.
Both castes actively cooperate
in defending food finds. The two
castes also have different defensive
roles. Minors bite the legs of
competitors whilst majors attack
the heads. During foraging minor
workers also guard food finds
whilst majors transport food back
to the nest.
Specialization in pheromone
communication among different
worker castes also extends beyond
foraging trails. For example, Atta
leafcutter ants (Figure 3) use alarm
pheromones to signal predators or
other dangers. The different size
castes in Atta possess different
blends of the same overall alarm
pheromone components, but
worker castes respond differently
to the blends produced by other
worker castes.
Multimodal communication
Chemical communication is of
great importance in ant foraging
organization. But foraging ants
also use other modalities to
communicate, and signals of
different modalities may combine
in promoting the organization
of a colony's foraging system,
and in other areas of colony life
such as defence. Close behind
chemical communication in overall
Figure 3. Leafcutter ants, Atta cephalotes, form dense foraging columns when trans-
porting leaves back to the nest along pheromone trails. (Image courtesy of Alex Wild.)
importance is the use of tactile
communication, either through
substrate-borne vibration or
direct contact. Direct contacts
may take the form of ritualised
movements in communication,
such as displays, dances,
waggling and jerking. Physical
displays by returning foragers
of many ant species often serve
a similar excitatory/recruitment
role to that observed in honeybee
waggle dances. The commonest
form of physical contact is mutual
antennation. This is seen very
frequently when ants pass in
opposite directions on a trail but
has yet to be assigned a purpose.
It probably does not comprise
a `language', as suggested by
Wassmann, but it is hard to
believe that no information is
transmitted.
In contrast to the widely
broadcast information of
pheromone trails, the use of
sounds, physical contacts and
displays are primarily mechanisms
whereby information can be
communicated to near neighbours.
In some situations, however, the
message is passed from ant to ant
and so travels further. Camponotus
senex live in large arboreal nests
built from larval silk. If a small area
of the nest is disturbed physically,
or by carbon dioxide, then the
ants affected produce an alarm
response by drumming their
abdomens on the nest substrate.
This stimulates other ants to follow
suit, resulting in the communication
of alarm throughout the entire nest,
which can be up to 1m in length.
The volume of a medium-sized
colony drumming is greater than
human speech.
Current Biology Vol 16 No 15
R574
Multiple pheromones, displays,
contacts and sounds are often
used in combination. This is
probably not to provide backup
mechanisms (redundancy) but to
communicate a wider repertoire
of messages. For example,
Aphaenogaster albisetosus
(Figure 4) modulate the recruitment
pheromone by rubbing their
abdominal tergites together to
make sound. When individual
A. albisetosus workers locate
large prey items, such as dead
insects, they release a poison
gland pheromone and audibly
stridulate to attract workers in the
locale. The stridulation encourages
other workers to release further
pheromone and this feedback
leads to rapid trail recruitment
to the prey site. A. albisetosus
retrieves prey items significantly
faster when stridulation is present.
Future research
Research shows that ants, and
also honeybees and other species
of social insects, use several
pheromones or other signals in
organizing their foraging system.
Two important and connected
questions, therefore, are to
determine why multiple signals
are needed and how they work
together. Some progress has been
made in the honeybee, where
four dances and a pheromone are
known to be involved. The role of
most of these signals is known. The
waggle dance directs foragers to
food and recruits them to foraging,
the vibratory dance prepares
foragers for work by causing them
to move into the dance floor area
where returning foragers make
waggle dances, and the tremble
dance recruits additional workers
to the task of unloading nectar
foragers. But what is not known is
precisely how they work together
and why five signals (and maybe
more) are needed. Why not four?
One possibility is that some
of these signals are fine tuning.
Multiple signals may be needed
because of inherent limitations
in the signals used. For example,
the short-lived attractive and
repellent trail pheromones used
by Pharaoh's ants can direct
foragers to the rewarding branch
at a trail bifurcation but a single
one of these pheromones can only
direct about 75% to the rewarding
branch. Perhaps the presence of
two pheromones can increase this
to 90%.
We also need to understand
communication mechanisms in
relation to the foraging method.
The solid substrate upon which
ants walk from nest to food
is suitable for depositing trail
pheromones to guide nestmates.
A trail pheromone is obviously less
useful for flying social insects, such
as honeybees. However, some
stingless bees do mark routes to
food with pheromone, which they
deposit on vegetation as discrete
beacons rather than a continuous
trail. The use of a trail pheromone
means channels of communication
may be continuously open for
ants, because they are capable of
a continual, reactive exchange of
information with nestmates whilst
foraging. This is in marked contrast
with the honeybee, where the
signals used to organize foraging
are communicated in the nest.
Foraging is a dynamic process and
an important role of communication
is to recruit or direct nestmates
rapidly to a food source. But the
need to do more than this, for
example to retain a longer term
memory, may require additional
pheromones or signals.
The multiple signals used in
social insect communication
provide shared information and
enable the colony (or system)
to be more responsive or better
regulated, so that it functions
better. Similar complexity is found
at other biological levels, such as
cell signalling pathways, where
positive and negative feedback
provide the capacity for control at
multiple levels and enable greater
flexibility in system responsiveness.
A major interdisciplinary challenge
in modern biology is to understand
how complex adaptive systems
function, and how they function
robustly yet flexibly. One goal in
this research is to determine if
there are any general principles
underlying adaptive biological
systems. The focus of this research,
and the funding, is usually directed
at the organismal level or below,
particularly cells in a multicellular
organism, or molecules within cells.
Insect societies provide another
level of organization for comparison.
Further reading
Hölldobler, B., and Wilson, E.O. (1990). The
Ants. (Cambridge, Mass.: The Belknap
Press of Harvard University).
Jackson, D.E., Martin S.J., Holcombe M., and
Ratnieks F. (2006). Longevity and detection
of persistent foraging trails in Pharaoh's
ants, Monomorium pharaonis. Anim.
Behav. 71, 351­359.
Sumpter, D. J.T. (2006). The principles of
collective animal behaviour. Phil. Trans. R.
Soc. Lond. B 361, 5­22.
Witte, V., and Maschwitz, U. (2002).
Coordination of raiding and emigration
in the ponerine army ant Leptogenys
distinguenda: a signal analysis. J. Insect
Behav. 15, 195­217.
1Department of Computer Science,
University of Sheffield, Western Bank,
Sheffield S1 4DP, UK. 2Department of
Animal and Plant Sciences, University
of Sheffield, Western Bank, Sheffield S1
4DP, UK. E-mail: duncan@dcs.sheffield.
ac.uk
Figure 4. Foragers of Aphaenogaster albisetosus use stridulation and a poison gland
pheromone to attract additional foragers when they locate prey they cannot capture
alone. Transport of the prey back to the nest is via highly volatile trails. (Image courtesy
of Alex Wild.)