Content uploaded by Elise Nowbahari
Author content
All content in this area was uploaded by Elise Nowbahari on Mar 19, 2016
Content may be subject to copyright.
In social insects, nestmate recognition plays an essen-
tial role in social organization. Given the importance of
nestmate discrimination in social behavior, it is not surpris-
ing that many investigators have attempted to determine the
nature of the signals that insects use for such recognition
(called labels or cues), their sites of production, and the dis-
crimination ability of individuals throughout ontogenesis
(see Breed, 1998, for a review). These studies have dem-
onstrated that nestmate recognition is based on chemical
cues—mainly cuticular hydrocarbons—which are species
and colony specific in both composition and concentration
(see, e.g., Breed & Bennett, 1987; Howard, 1993; Lahav,
Soroker, Hefetz, & Vander Meer, 1999; Lenoir, Fresneau,
Errard, & Hefetz, 1999; Nowbahari et al., 1990; Soroker,
Vienne, Hefetz, & Nowbahari, 1994). Individuals within a
colony often differ from heterospecific individuals as well
as conspecific non-nestmates in both genetic and environ-
mental cues. Recognition cues are learned by workers and
serve as a template to determine the colonial membership
of each individual encountered (Breed & Bennett, 1987;
Breed & Julian, 1992; Greenberg, 1988; Pfenning, Gamboa,
Reeve, Shellman-Reeve, & Ferguson, 1983; Stuart, 1987).
In contrast to the important advances in the determina-
tion of the nature of the signals involved in nestmate recog-
nition, less experimental progress has been achieved in the
attempt to understand the mechanisms, or decision rules,
underlying the behavior of individuals during encounters.
The mechanisms underlying amicable or aggressive behav-
iors performed by one individual facing another could be
based on the comparison of the perceived signal with the
individual’s own signal or template, as has been revealed
mostly by studies focusing on territorial behavior. Notable
among invertebrate studies are those that investigate a phe-
nomenon known as “dear–enemy” which has been studied
largely in vertebrates (see, e.g., Temeles, 1994). In eusocial
Hymenoptera and particularly in ants, the “dear–enemy”
phenomenon refers to a pattern of aggression where neigh-
bors are discriminated from strangers and are the object of
less aggression (Heinze, Foitzik, Hippert, & Hölldobler,
1996; Knaden & Wehner, 2003; Pfenning et al., 1983). The
mechanisms by which neighbors are discriminated from
strangers and the function of neighbor–stranger discrimi-
nation has not been clearly determined, but it has been sug-
gested that individual recognition based on odor learning is
involved (Obin & Vander Meer, 1988).
Another largely unsolved problem regarding nestmate
recognition is the impact of memory on the integration of
perceived signals. Adoption experiments point out the exis-
tence of a sensitive learning period during which the sensory
imaging of perceived signals is formed. In ants, as in other
social insects, during this sensitive period the sensory imag-
ing of a template of colonial odor is learned, and it appears
to be an imprinting-like phenomenon starting in early life at
the larval stage (Carlin, Halpern, Hölldobler, & Schwartz,
1987; Isingrini, Lenoir, & Jaisson, 1985) or during the first
days of adult life (Fénéron & Jaisson, 1995; Isingrini et al.,
1985; Morel, 1982), with long-term memory effects last-
ing 1 year (Errard, 1994). In Camponotus vagus ants, Morel
showed that emerging within the colony is important for
colonial odor recognition. Le Moli and Mori (1985) also
found that early experience in Formica cunicularia strongly
influenced intra- and interspecific relationships. In Cata-
glyphis cursor, the sensitive learning period corresponds
to the 4 days following emergence (Nowbahari & Lenoir,
87 Copyright 2007 Psychonomic Society, Inc.
Learning of colonial odor in the ant
Cataglyphis niger (Hymenoptera; Formicidae)
E
LISE
N
OWBAHARI
UMR CNRS 7153, Université Paris Nord, Villetaneuse, France
Ants learn the odors of members of their colony early in postnatal life, but their ability to learn to recognize
noncolony conspecifics and heterospecifics has never been explored. We used a habituation–discrimination
paradigm to assess individual recognition in adult Formicine ants, Cataglyphis niger. Pairs of workers from
different colonies were placed together for repeated trials, and their ability to discriminate the ant that they en-
countered from another familiar or unfamiliar ant was observed. Some ants were isolated between encounters,
and others were returned to their home colonies. Our results suggest for the first time in ants that C. niger adults
learn about individual ants that they have encountered and recognize them in subsequent encounters. Ants are
less aggressive toward non-nestmates after they are familiar with one another, but they are aggressive again
when they encounter an unfamiliar individual. Learning about non-nestmates does not interfere with an ant’s
memory of members from its own colony.
Behavior Research Methods
2007, 35 (2), 87-94
E. Nowbahari, elise.nowbahari@leec.univ-paris13.fr
88 N
OWBAHARI
1989). Early experience may thus induce a preference for
a particular social environment and explain observed natu-
ral associations between different species of ants (Errard,
1994). But learning and recognition memory of any encoun-
tered ants has to be plastic rather than fixed, and thus not
limited to the early postnatal period. Experimental evidence,
obtained, for example, in the study of hydrocarbons of indi-
viduals isolated from their colonies, has demonstrated that
their chemical labels deviate from that of the mother colony
(Boulay & Lenoir, 2001; Lenoir, Cuisset, & Hefetz, 2001).
Therefore, an efficient learning system has to be flexible and
must continue throughout the entire lifetime of the ant.
This article reports non-nestmate odor learning dur-
ing encounters between mature adult ants. It describes
individual odor discrimination and the capacity of Cata-
glyphis niger adult ants to modulate their reactions toward
non-nestmates. Its particular objective was to investigate
the following questions: Do the adult workers of C. niger
change their reactions toward a foreign individual that
they encounter repeatedly? In other words, do they recog-
nize familiar non-nestmates? After they learn to recognize
the non-nestmate during successive encounters, do they
change their behavior toward members of their colony?
EXPERIMENT 1
Conspecific Encounters
Two experiments were conducted. Experiment 1 used
habituation/discrimination trials in which repeated en-
counters between 2 individuals resulted in a decrease of
aggressiveness (habituation) followed by an increase in
responsiveness (discrimination) when a novel individual
was encountered. In this experiment, encounters between
2 ants (dyadic encounters) from two distant colonies of
C. niger were performed, followed by a trial with a novel
ant to confirm the discrimination between 2 ants.
Method
Ants and rearing conditions
. Nine polygynous (including more
then one reproductive female) colonies of C. niger were collected
from two areas to the north (Netanya) and south (Rishon LeZion)
of Tel Aviv. These sites extend over a distance of some 15 km and
are more than 50 km (31.07 miles) apart. In a previous study of
intercolonial recognition in C. niger, we observed a high level of
aggression between individuals from different colonies from these
sites (Nowbahari, Fénéron, & Malherbe, 1999).
In the laboratory, colonies were maintained under homogeneous
conditions (a temperature of 25º–28ºC) to reduce as much as pos-
sible the influence of temperature and nutrition on behavior. They
were reared in similar artificial nests consisting of a round plastic
box (15 cm in diameter) colored in black and connected to a forag-
ing arena (28 27.5 8.5 cm). They were fed on mealworm larvae
and a honey/apple mixture twice a week.
Behavioral observations and data analysis
. We used a classi-
cal habituation/discrimination paradigm (see Manning & Dawkins,
1998; Todrank & Heth, 2003). The behavioral trials consisted of dy-
adic encounters between ants from two C. niger colonies (Figure 1).
One ant was from a colony collected in Netanya (Colony A), and the
other was from a colony collected in Rishon le Zion (Colony B).
The habituation portion consisted of 10 repeated encounters. To
stage an encounter, with cleaned forceps we transferred 2 work-
ers from the origin colony into a neutral arena: a round plastic box
9 cm in diameter, the bottom of which was covered with filter paper
changed for each encounter and the walls of which were coated with
fluon to prevent climbing. Where possible, the tested workers were
sampled from among the forager workers in the foraging arena of
the colony or within the nest. The ants were marked on the thorax
Figure 1. Method used in Experiment 1: Conspecific tests. A and B, nestmates
for A and B encountered ants. Palmahim and Tel Aviv were the two different
sites where ants were obtained.
Dishabituation Trials
AA, BB or AB, BA
- 10-min interval
- Isolated from mother colony
- 10 dyadic encounters (AB)
- 10 dyadic encounters (AB)
- 10-min interval
- Return to mother colony
between trials.
Colony A
Palmahim
Colony B
Tel Aviv
Condition 1
Habituation Trials
Condition 2
Habituation Trials
L
EARNING
IN
A
NTS
89
with a distinct spot of odorless indelible paint (Uni Paint Marker PX
20, Mitsubishi Pencil Co.) according to their colonial origin. Before
each encounter, both workers were placed for 2 min in separate glass
cylinders inside the arena to keep them calm. The test began with
careful removal of the cylinders. All behaviors performed by each
worker were recorded, by direct visual scoring, according to the scan
sampling method, every 5 sec for 3 min. The habituation dyadic
encounter trials were performed with the same individuals 10 times,
with an intertrial interval of 10 min. At the end of each series of 10
encounters, we performed a discrimination trial, in which the subject
ants were confronted with novel ants. During this trial, each ant from
Colony A or B was confronted with a nestmate or with an unfamiliar
non-nestmate issued from the same colony as that of the encounter
ants. This permitted us to verify whether or not the ant (1) distin-
guished between non-nestmate odors and its colonial odor, which
was probably learned during the early sensitive period, and (2) dif-
ferentiated a familiar non-nestmate from an unfamiliar one (i.e.,
differentiated individually 2 sisters that had very similar odors).
In these trials, we investigated the modification of the ant’s behav-
ior over successive encounters with previously unfamiliar ants. The
tests were conducted in two different conditions.
Condition 1: Successive trials with isolated workers. During the
10-min interval between trials, the opponent workers remained isolated
from their colony and were kept in two separate cylinder boxes. A total
of 27 series of tests were performed, so that 54 workers were tested.
Condition 2: Successive trials with nonisolated workers. During
the interval between trials, the opponent workers were replaced into
their respective colonies. A total of 33 series of tests were performed,
so that 66 workers were tested. These tests were performed to deter-
mine whether access to the worker’s colonial odor and eventually
the comparison of that odor with the encountered non-nestmate odor
would modify the worker’s behaviors during the tests. We hypothe-
sized that the ants that returned to their mother colony between trials
would be less stressed or might refresh their colony odor memory,
as was observed previously in Camponotus fellah (Boulay, Katzav-
Gozansky, Hefetz, & Lenoir, 2004).
For each trial, all behaviors displayed by each worker were recorded
every 5 sec for 3 min (Table 1). We obtained a total of 36 behavioral
records for each ant. Only agonistic behaviors and self-grooming
that changed during the successive trials were analyzed here. These
included aggressive behaviors (i.e., biting, flexing the gaster forward,
and spraying formic acid), threatening behaviors (i.e., wide opening
of mandibles), and nonaggressive behaviors (i.e., self-grooming).
The frequency with which these behaviors were performed during
each trial (3 min) was calculated for each individual.
Comparisons between the successive trials in both conditions were
made using Friedman tests. Comparisons between the two conditions
were made using a permutation test for two independent samples.
To study changes in behaviors across the 10 trials, we used Helmert
contrasts to compare the mean of the 1st trial with the means of the
subsequent trials. If there were significant differences, we compared
the mean of Trial 2 with the mean of the remaining trials, and so forth.
Contrasts were no longer significant when asymptote was reached.
We also used permutation tests for two related samples to com-
pare the discrimination trial (Trial 11) with Trial 10. The statistics
were considered significant at p .05. All statistics were performed
with StatXact 7 software.
Results
As a whole, whether isolated or not, C. niger ants con-
fronted with a heterocolonial (conspecific) non-nestmate
modified their behaviors during the 10 successive dyadic
encounters. Aggressive behaviors, usually considered an
accurate measure of nestmate recognition, changed the
most (Figure 2). In general, the tested ants significantly
changed their level of aggression over the trials (Friedman
test, p .005). We observed a significant change of ag-
gressive behaviors during the 10 trials, for both nonisolated
ants ( p .001) and isolated ants ( p .03). But, the mean
level of aggression over the 10 trials was higher in isolated
ants (Condition 1) than in nonisolated ants (Condition 2)
(permutation test for two independent samples, p .03).
For nonisolated ants (Condition 1), aggressive behav-
iors were more frequent during the first encounter in
comparison with the mean of the subsequent encounters
(Helmert contrasts, permutation test for related samples,
p .001). They were also more frequent during the second
encounter in comparison with the mean of the subsequent
encounters ( p .01). But the following Helmert contrasts
were not significant. So the asymptote level was reached
in the third trial. For isolated ants (Condition 2), Helmert
contrasts were not significant for aggressive behaviors.
Similarly to aggressive behaviors, threatening behav-
iors changed over all the successive trials ( p .001, Fig-
ure 3A). We observed a significant change of threatening
behaviors during the 10 trials for nonisolated ants ( p
.001), but no significant change for isolated ants ( p
.34). There was no significant difference between isolated
and nonisolated ants for the mean level of threatening be-
haviors over the 10 trials ( p .37).
For nonisolated ants, threatening behaviors were more
frequent during the first encounter in comparison with the
mean of the subsequent encounters (Helmert contrasts,
permutation test for two related samples, p .001). They
Tabl e 1
Description of All Behaviors Observed During Trials
Behavior Description
Nonaggressive
Exploration movement The ant moves in test area.
No interaction with the other ant.
Inactivity The ant is motionless.
No interaction with the other ant.
Antennal contacts Inspection with antennation contact of different body parts of the encountered ant.
Grooming Self-grooming of different parts of body, such as the antenna, legs and mandibles.
No interaction with the other ant.
Trophallactic exchanges Oral food exchange.
Aggressive Biting (seizing the opposite ant with the mandibles), threatening
(opening of mandibles), spraying of formic acid.
Other Submission (motionless in nymphal position), running away.
90 N
OWBAHARI
were also more frequent during the second encounter in
comparison with the mean of the subsequent encoun-
ters ( p .03) and during the third in comparison with
the mean of the remaining encounters ( p .05). But
the subsequent Helmert contrasts were not significant.
The asymptote level was reached in the fourth trial. For
isolated ants, Helmert contrasts were not significant for
threatening behaviors. Threatening behavior usually pre-
ceded overt aggression (i.e., biting and venom spraying),
but sometimes it occurred alone or was followed by either
intensive antennation or self-grooming.
Self-grooming changed significantly during the succes-
sive trials ( p .0005). We observed a significant change
of this behavior during the 10 trials for both nonisolated
ants ( p .0001) and isolated ants ( p .0001). The mean
level of self-grooming was significantly higher for the iso-
lated than for the nonisolated ants ( p .02) (Figure 3B).
For nonisolated ants, it was more frequent during the first
encounter in comparison with the mean of the subsequent
encounters ( p .003). They were also more frequent dur-
ing the second encounter in comparison with the mean of
the following encounters ( p .03). But the subsequent
Helmert contrasts were not significant. The asymptote
level was reached in the third trial.
For isolated ants, we observed asymptote level in the
fourth trial. The mean level of self-grooming was higher
during the first encounter in comparison with the mean of
the subsequent encounters ( p .02), during the second
encounter in comparison with the mean of the subsequent
encounters ( p .001), and during the third in comparison
with the mean of remaining encounters ( p .009). The
subsequent Helmert contrasts were not significant.
Following the habituation trials, we performed a dis-
crimination trial, in which the subject ants were con-
fronted with novel ants from the same colony as that of
the previously encountered ant. During this trial, each ant
from Colony A or B was confronted with a nestmate or an
unfamiliar non-nestmate.
The results of the discrimination trials (Trial 11) showed
that the frequencies of displayed aggressive behaviors dif-
fered when ants encountered their nestmates and unfamil-
iar non-nestmates: There was virtually no aggressiveness
toward nestmates (Figure 4).
The nonisolated and isolated ants confronted with a
heterocolonial unfamiliar non-nestmate increased their
aggressive behaviors during the discrimination trial (two
related-samples permutation tests: nonisolated ants, n
35, p .001; isolated ants, n 22, p .05). In addition,
the mean level of aggression in the discrimination trial
was higher for isolated ants than for nonisolated ants ( p
.05). These ants, when they encountered a homocolonial
ant, actually decreased their aggressive behaviors signifi-
cantly (nonisolated ants, n 31, p .03; isolated ants,
n 32, p .04). There were no differences between iso-
lated ants and nonisolated ants in these trials ( p .97)
Discussion
Our results from Experiment 1 show that C. niger adult
ants are able to discriminate between conspecific non-
nestmates. The hostility toward ants of alien colonies dur-
ing the first encounter and the lack of aggression when
they encounter their colony members outside the colony is
well-known evidence of colony recognition. The decrease
in aggressive and threatening behaviors, exhibited by both
socially isolated (Condition 1) and socially nonisolated
(Condition 2) ants, demonstrates that ants habituate to
individual foreign odors over successive encounters. A
recent study in ants has reported a peripheral recognition
mechanism in detecting colony-specific chemical signals
by chemosensory sensillum (Ozaki et al., 2005).
However, isolated and nonisolated workers did not be-
have similarly with regard to non-nestmate recognition. We
Figure 2. Mean (SE) aggressive reactions/3 min of conspecific
ants’ encounters during 10 successive trials in two conditions:
isolated and nonisolated condition. Aggressive reactions were re-
corded by sampling every 5 sec.
0
1
2
3
4
5
6
12345678910
Trials
Aggressive Reactions
Mean/3 min (±SE )
Isolated ants
Nonisolated ants
Figure 3. Mean (SE) threatening (A) and self-grooming (B)
reactions/3 min of conspecific ants’ encounters during 10 succes-
sive trials in two conditions: isolated and nonisolated condition.
Threatening and self-grooming reactions were recorded by sam-
pling every 5 sec.
0
1
2
3
4
5
6
12345678910
Trials
Threatening Reaction
Mean/3 min (±SE )
0
1
2
3
4
5
7
6
12345678910
Trials
Self-Grooming
Mean/3 min (±SE )
Isolated ants
Nonisolated ants
A
B
L
EARNING
IN
A
NTS
91
observed a more rapid decline in the level of aggression
(i.e., habituation) in ants that were returned to their mother
colony and less of a decline in the level of aggression when
the ants were maintained in isolation from their colony.
We noted that only some of the ants were aggres-
sive. But all of the aggressive ants remained aggressive
throughout the trials even though the frequency of aggres-
sion decreased. Similarly, workers that did not display any
aggression during the first encounter showed consistent
behavior and remained nonaggressive throughout the tri-
als. This result suggests that in C. niger there is a division
of labor in defensive behavior, which is usually performed
by foragers. Nowbahari et al. (1999) demonstrated that in
C. niger, only large workers that were also the foragers
evolved alternative fighting strategies. It seems that, in
social insects, foragers who generally have the defensive
role possess higher nestmate recognition ability (Breed,
Smith, & Torres, 1992).
The results from the discrimination trials indicate
clearly that the C. niger ants after successive encounters
with a non-nestmate recognize their nestmate and can dis-
criminate the odor of an unfamiliar non-nestmate, which
theoretically has a very similar odor to that of the previ-
ously encountered ant.
EXPERIMENT 2
Heterospecific Encounters
To determine whether ants discriminate between dif-
ferent heterospecific individuals, Experiment 2 was con-
ducted with successive dyadic encounters between two
different species: C. niger and Cataglyphis cursor. This
experiment permitted us to determine how ants react to a
variety of non-nestmates in consecutive encounters.
Method
Ants and rearing conditions
. In Experiment 2 we used the same
polygynous colonies as in Experiment 1, which were collected in Is-
rael, as well as three monogynous colonies (each including one repro-
ductive female) of C. cursor, collected near Perpignan (in southern
France). Preliminary observations indicated that these colonies did not
show intercolonial aggression (see Nowbahari et al., 1990).
The colonies were maintained in artificial nests and conditions
similar to those in Experiment 1.
Behavioral observations and data analysis
. This experiment
was conducted in the same way as was Experiment 1, to determine
how an ant would react in heterospecific dyadic encounters with 1 ant
from a C. niger colony collected in Israel and a 2nd from a C. cursor
colony collected in France. Preliminary observations indicated that
ants confronting heterospecific individuals exhibit overt aggression,
perhaps because of the substantial differences in the chemical signa-
ture between these ants (for the chemical analysis of C. cursor, see
Nowbahari et al., 1990, and for C. niger, see Soroker et al., 1994).
Here nearly all tested ants exhibited aggressive behaviors. A total
of 17 series of tests were performed, but because of high aggression
levels and the mortality of 1 of the tested ants during the trials, only 7
series could be analyzed, corresponding to 7 tested workers.
We also assessed the reactions of an ant toward different
heterospecific ants; that is, the partners were changed in successive
trials. Here we also observed high aggressiveness, and we analyzed
the reactions of 10 C. niger ants encountering successively 7 differ-
ent C. cursor ants. In Experiment 2, as in Experiment 1, all behav-
iors displayed during the tests were recorded (Table 1), but only the
results of aggressive behaviors are presented and analyzed here.
Comparisons between the successive trials in both conditions
were made using Friedman tests. Comparisons between two con-
ditions were made using a permutation test for two independent
samples. To study changes in behaviors across the 10 trials, we used
Helmert contrasts.
Results
When C. niger ants encountered a heterospecific ant
(C. cursor), they exhibited overt aggression. These aggres-
sive reactions were followed by a significant reduction over
10 successive trials with the same individual ( p .001),
followed by an increase in aggression during the discrimina-
tion trials ( p .03; n 7; Figure 5). We observed a stron-
ger aggressive reaction in the first trial in heterospecific
encounters (average of aggressive behavior in the first
encounter 18.8 2.2) in comparison with the dyadic
conspecific encounters presented in Figure 2 (average of
aggressive behavior in first encounter 4.07 1.03).
As in homospecific encounters (Experiment 1), in
heterospecific encounters aggressive behaviors were more
frequent during the first encounter in comparison with
the mean of the subsequent encounters ( p .01). They
were also more frequent during the second encounter in
comparison with the mean of the following encounters
( p .03). But the following Helmert contrasts were not
significant.
When C. niger encountered different heterospecific in-
dividual ants (C. cursor) in successive trials, the aggressive
level did not change significantly, and there was no reduc-
tion in their aggressive behaviors ( p .5) (see Figure 5).
Discussion
These results confirm the results of Experiment 1. But in
Experiment 2, with heterospecific encounters, we observed
higher aggressiveness in the first trial and a rapid reduc-
tion of this reaction during the 10 successive encounters.
The asymptote level was reached in the second encounter
in the third trial for nonisolated ants, and it was not es-
tablished for isolated ants (Experiment 1). This difference
could have been due to the greater odor similarity between
the ants of the conspecific as opposed to the heterospecific
Figure 4. Mean (SE) aggressive reactions/3 min of the ha-
bituated ants when they encountered a known conspecific non-
nestmate in the 10th trial (unfilled bars) and then an unknown
non-nestmate ant in the 11th
trial or dishabituation trial (hatched
bars) or a nestmate in test trial (filled bars), in two conditions:
isolated and nonisolated conditions. Aggressive reactions were
recorded by sampling every 5 sec.
*
p .05.
***
p .001.
0
1
2
3
4
5
6
7
8
9
Isolated Ants Nonisolated Ants
Aggressive Actions
Mean/3 min (±SE )
*
*
***
***
***
92 N
OWBAHARI
colonies. The process of comparing the learned template of
the colonial odor with the individual foreign odors could
be the reason for the stronger modification of the level of
aggression observed in heterospecific encounter trials.
The results of dyadic encounters of an ant facing two
different heterospecific partners indicate that the ants did
not react identically when faced with the different indi-
viduals or the same individuals in successive trials. We
did not observe a reduction of aggressive behaviors. This
result confirms that the reduction in aggressiveness ob-
served across encounters with the same individual was a
habituation process and not the result of tiredness.
GENERAL DISCUSSION
The results of our experiments showed that C. niger
adult ants can learn individual new odors of conspecific or
heterospecific ants. Individual learning in ants is known
only in the queen recognition of Pachycondyla villosa, a
primitive Ponerine ant that forms very small societies. It
is now well known that in these ants the queen odor is a
fertility signal very different from the odor of workers and
is therefore probably easy to discriminate (D’Ettorre &
Heinze, 2005). Our data are the first obtained for large so-
cieties in which individual discrimination has been consid-
ered improbable. They confirm the large olfactory learn-
ing possibilities of ants, as demonstrated in Camponotus
workers (Dupuy, Sandoz, Giufra, & Josens, 2006).
Although more investigations will be necessary for
researchers who wish to understand the mechanisms of
recognition systems in ants, it seems that ants modulate
their defensive actions according to environmental con-
text (Carlin & Johnston, 1984; Temeles, 1994). The study
of nestmate recognition and spatial variation in aggressive
behavior in Camponotus chilensis has shown that C. chil-
ensis individuals are able to discriminate nestmates from
intruders, with no aggression toward nest companions but
with aggressiveness toward heterocolonial conspecifics
that decreases significantly with distance (Velásquez,
Gómez, González, & Vásquez, 2006).
In ants, the modification of reactions through learning is
observed in field and experimental situations of territorial
defense as an increase in aggressive behaviors, a phenom-
enon known as enemy specification (Carlin & Johnston,
1984; Obin & Vander Meer, 1988; Wilson, 1975). Enemy
specification presents evidence for strong discrimination
between different types of intruders in the intensity of the
stimulus required for activation of a response. The intensity
of the stimulus can be increased by repeated introductions
of alien ants. The decrease in the level of aggression through
habituation, as in C. niger, and its increase through sensi-
tization, as in other ant species such as Pheidole dentata
(Carlin & Johnston, 1984; Velásquez et al., 2006; Wilson,
1975), are probably some of the simple types of learning in
invertebrates as described by Abramson (1997). In our ex-
periments, we considered the decrease in the frequencies of
aggression as non-nestmate odor learning and as recogni-
tion of a previously encountered individual. The differences
in aggressiveness during the habituation and discrimination
trials in two conditions, as well as the observed differences
when C. niger encountered a conspecific in comparison
with a heterospecific ant, confirm these results of context
dependence of recognition processes. This result could cor-
respond to the “dear–enemy” phenomenon, which may be
mediated by simple recognition learning as has been sug-
gested by Langen, Tripet, and Nonacs (2000) in Pheidole
ants and reported by Heinze et al. (1996) within and be-
tween species of Leptothorax. In C. niger, workers meeting
in a neutral arena probably reduce their motivation to be
aggressive toward non-nestmates over time. In addition, our
analysis suggests the efficacy of future field studies in natu-
ral territories in order to assess whether or not the intensity
of a worker’s responses is an adaptive consequence of the
economics of territorial defense. In the field, aggressive
interactions with less threatening neighbors can represent
wasted energy, which would be better utilized in the defense
of the territory against strangers. The variation in aggres-
sion between spatially distant colonies also suggests that
additional genetic or environmental factors are involved in
differential discriminative responses (Heinze et al., 1996).
Figure 5. Mean (SE) aggressive reactions/3 min of ants in encounters with
the same heterospecific ant during 10 successive trials and then with a native
heterospecific in dishabituation trial (Trial 11), or when encountering different
individuals during 7 trials. Aggressive reactions were recorded by sampling
every 5 sec.
***
p .001.
0
5
10
15
20
25
30
1234567891011
Trials
Aggressive Reactions
Mean/3 min (± SE )
***
Different individuals
Same individuals
L
EARNING
IN
A
NTS
93
C. niger adult ants seem to learn new odors. Although
historically habituation was dismissed as a functionally
insignificant form of behavior, most contemporary scien-
tists view habituation as a form of adaptive modification
of behavior, as in learning (Wyers, Peeke, & Herz, 1973).
This process seems to be more accurate when C. niger
workers have the opportunity to compare the opponent’s
individual odor with their own reference (i.e., their colonial
odor, Experiment 1). Indeed, we observed a more rapid
decrease in the level of aggression (i.e., habituation) in
ants that were returned to their colony. Ants isolated from
their colony appear to use their long-term memory ca-
pacity to distinguish their opponents, most likely by com-
paring the opponents’ odors with the memorized colonial
odor learned during the sensitive period. This behavioral
difference between isolated and nonisolated ants seems
to indicate that short-term memory is involved in non-
nestmate odor learning in C. niger adult ants as has previ-
ously been observed in bees (Erber, Masuhr, & Menzel,
1980). C. niger ants act as if they do not exclusively use
their own odor as a reference for nestmate recognition.
Isolated ants seem to learn the odor of non-nestmates as
well as that of the nonisolated ants. But the ants that have
the possibility of returning to their nest seemed to have
better learning of non-nestmate odors. They rapidly de-
creased their agonistic level during the habituation trials
and reacted significantly more aggressively with a novel
non-nestmate in the discrimination test. As Lenoir et al.
(2001) reported, the isolated condition could induce some
changes in cuticular hydrocarbon profiles and cause a dif-
ferent reaction toward non-nestmates. These authors ob-
served overt aggression toward isolated ants when these
were reintroduced into their mother colonies.
Therefore, the ability to learn the non-nestmate odors
is not necessarily limited to early and/or preimaginal life
(i.e., the larval stage) by means of an imprinting process
(Carlin et al., 1987; Fénéron & Jaisson, 1995; Isingrini
et al., 1985; Morel, 1982; Nowbahari & Lenoir, 1989).
Learning also occurs at the adult stage, and it may con-
tinue during the entire lifetime of a bee (Barrows, Bell, &
Michener, 1975) or of an ant, as shown here in C. niger.
At this stage, the learning process consists of a continu-
ous habituation to slight variations in the odor emanating
from nestmates. As Vander Meer, Saliwanchik, and Lavine
(1989) have demonstrated in Solenopsis invicta, and as
several authors have since demonstrated in various other
species (see review by Lenoir et al., 1999), the cuticular
patterns of workers change continuously over time. Odor
is a dynamic phenomenon. Adult workers are then able to
adapt their behavior to the changing cuticular patterns by a
process of odor learning. They can learn to recognize indi-
vidual and/or colonial odors. In an artificially mixed col-
ony (i.e., a colony comprising different ant species), ants
acquire some of the hydrocarbon components character-
istic of their allospecific nestmates, thus achieving a uni-
fied profile corresponding to a mixture of odors from both
species (Errard, 1994; Errard, Hefetz, & Jaisson, 2006). In
such experimental conditions, ants from different colonies,
which are usually very aggressive, can cohabit if they are
grouped during their early life and remain nonaggressive
throughout their adult life. Nevertheless, recognition learn-
ing and subsequent reductions in aggressive behavior are
not always the result of similarities in hydrocarbon patterns.
In naturally occurring parabiotic societies, ants are able to
recognize nestmate and non-nestmate individuals of the
associated species even though their cuticular profiles are
different (Orivel, Errard, & Dejean, 1997). In such a case,
recognition can take place only through memorization or
through familiarization with parabiotic individuals during
the sensitive period. In C. niger, recognition also seems
to be malleable in the adult stage by means of a sensory
template encoding the labels (memory or familiarity). The
hydrocarbon pattern modification that usually results from
passive exchanges between individuals (Soroker et al.,
1994) does not seem to have occurred in our experiments.
Indeed, no trophallaxis or allogrooming was observed dur-
ing the encounters.
Furthermore, the discrimination trial showed that ants
subjected to repeated tests, in conspecific or heterospecific
encounters, did not “forget” the colonial odor or template
learned in early adult life. They distinguished nestmates
from non-nestmates. This type of recognition necessarily
implies long-term memory like that demonstrated in other
ant species (see Errard, 1994). C. niger ants could also dif-
ferentiate familiar from unfamiliar non-nestmates belong-
ing to the same colony, even though they both presented
some similarities in their odor profiles. Therefore, the re-
tention of non-nestmate odors by C. niger adults might
result from a narrowly specific individual odor learning,
which would permit distinction between the slight differ-
ences in individual odors.
Our results show that ants can learn to recognize their
colony odor early in life, but that when mature adults en-
counter non-nestmates repeatedly, they can learn to rec-
ognize non-nestmates individually as well. This process
of learning as a consequence of confrontations seems to
be context dependent and may constitute an individual
strategy that increases fitness in ants.
AUTHOR NOTE
I am especially grateful to R. Fénéron, A. Lenoir, J. Todrank, and A. Hef-
etz for their helpful suggestions and comments. I thank J.-L. Durand for
helping in statistics analysis and J. Theau for his participation in performing
this study, A. Pezon for English revision of a first version of this article,
my friend C. Guignard for collecting ants in the field, and M.-C. Malherbe
for rearing ants. This study was supported by the CNRS (France) and the
MOST (Israel). Correspondence should be addressed to E. Nowbahari,
Laboratoire d’Ethologie Expérimentale et Comparée, UMR CNRS 7153,
Université Paris Nord, 99 Avenue J.-B. Clément, 93430 Villetaneuse,
France (e-mail: elise.nowbahari@leec.univ-paris13.fr).
REFERENCES
Abramson, C. I. (1997). Where have I heard it all before? Some ne-
glected issues of invertebrate learning. In G. Greenberg & E. Tobach
(Eds.), Comparative psychology of invertebrates: The field and labo-
ratory study of insect behavior (pp. 55-78). New York: Garland.
Barrows, E. M., Bell, W. J., & Michener, C. D. (1975). Individual
odor differences and their social function in insects. Proceedings of
the National Academy of Sciences, 7, 2824-2828.
Boulay, R., Katzav-Gozansky, T., Hefetz, A., & Lenoir, A. (2004).
Odour convergence and tolerance between nestmates through troph-
allaxis and grooming in the ant Camponotus fellah (Della Torre). In-
sectes Sociaux, 51, 55-61.
94 N
OWBAHARI
Boulay, R., & Lenoir, A. (2001). Social isolation of mature workers af-
fects nestmate recognition in the ant Camponotus fellah. Behavioural
Processes, 55, 67-73.
Breed, M. D. (1998). Chemical cues in kin recognition: Criteria for
identification, experimental approaches, and the honeybees as an ex-
ample. In R. K. Vander Meer, M. D. Breed, K. E. Espelie, & M. L.
Wiston (Eds.), Pheromone communication in social insects: Ants,
wasps, bees, and termites (pp. 57-78). Boulder, CO: Westview.
Breed, M. D., & Bennett, B. (1987). Kin recognition in highly eusocial
insects. In D. J. C. Fletcher & C. D. Michener (Eds.), Kin recognition
in animals (pp. 243-285). New York: Wiley.
Breed, M. D., & Julian, G. E. (1992). Do simple rules apply in honey-
bee nestmate discrimination? Nature, 357, 685-686.
Breed, M. D., Smith, T. A., & Torres, A. (1992). Role of guard hon-
eybees (Hymenoptera: Apidae) in nestmate discrimination and re-
placement of removed guards. Annals of the Entomological Society of
America, 85, 633-637.
Carlin, N. F., Halpern, R., Hölldobler, B., & Schwartz, P. (1987).
Early learning and the recognition of conspecific cocoons by carpen-
ter ants (Camponotus spp.). Ethology, 75, 306-316.
Carlin, N. F., & Johnston, A. B. (1984). Learned enemy specification
in the defense recruitment system of an ant. Naturwissenschaften,
71, 156-157.
D’Ettorre, P., & Heinze, J. (2005). Individual recognition in ant
queens. Current Biology, 15, 2170-2174.
Dupuy, F., Sandoz, J. C., Giufra, M., & Josens, R. (2006). Individ-
ual olfactory learning in Campontotus ants. Animal Behaviour, 72,
1081-1091.
Erber, J., Masuhr, T. H., & Menzel, R. (1980). Localization of short-
term memory in the brain of the bee, Apis mellifera. Physiological
Entomology, 5, 343-358.
Errard, C. (1994). Long-term memory involved in nestmate recogni-
tion in ants. Animal Behaviour, 48, 263-271.
Errard, C., Hefetz, A., & Jaisson, P. (2006). Social discrimination
tuning in ants: Template formation and chemical similarity. Behav-
ioral Ecology & Sociobiology, 59, 353-363.
Fénéron, R., & Jaisson, P. (1995). Ontogeny of nestmate brood rec-
ognition in a primitive ant, Ectatomma tuberculatum Olivier (Poneri-
nae). Animal Behaviour, 50, 9-14.
Greenberg, L. (1988). Kin recognition in the sweet bee, Lasiogloosum
zephyrum. Behavior Genetics, 18, 425-437.
Heinze, J., Foitzik, S., Hippert, A., & Hölldobler, B. (1996). Ap-
parent dear–enemy phenomenon and environment-based recognition
cues in the ant Leptothorax nylanderi. Ethology, 102, 510-522.
Howard, R. W. (1993). Cuticular hydrocarbons and chemical commu-
nication. In D. W. Stanley-Samuelson & D. R. Nelson (Eds.), Insect
lipids: Chemistry, biochemistry, and biology (pp. 179-229). Lincoln:
University of Nebraska Press.
Isingrini, M., Lenoir, A., & Jaisson, P. (1985). Preimaginal learning
as a basis of colony-brood recognition in the ant Cataglyphis cursor.
Proceedings of the National Academy of Sciences, 82, 8545-8547.
Knaden, M., & Wehner, R. (2003). Nest defense and conspecific
enemy recognition in the desert ant Cataglyphis fortis. Journal of In-
sect Behavior, 16, 717-730.
Lahav, S., Soroker, V., Hefetz, A., & Vander Meer, R. K. (1999).
Direct behavioral evidence for hydrocarbons as ant recognition dis-
criminators. Naturwissenschaften, 86, 246-249.
Langen, T. A., Tripet, F., & Nonacs, P. (2000). The red and the black:
Habituation and the dear–enemy phenomenon in two desert Pheidole
ants. Behavioral Ecology & Sociobiology, 48, 285-292.
Le Moli, F., & Mori, A. (1985). The influence of the early experience of
worker ants on enslavement. Animal Behaviour, 33, 1384-1386.
Lenoir, A., Cuisset, D., & Hefetz, A. (2001). Effects of social iso-
lation on hydrocarbon pattern and nestmate recognition in the ant
Aphaenogaster senilis (Hymenoptera, Formicidae). Insectes Sociaux,
48, 101-109.
Lenoir, A., Fresneau, D., Errard, C., & Hefetz, A. (1999). Individu-
ality and colonial identity in ants: The emergence of the social repre-
sentation concept. In C. Detrain, J. L. Deneubourg, & J. M. Pasteels
(Eds.), Information processing in social insects (pp. 219-223). Basel:
Birkhäuser.
Manning, A., & Dawkins, M. S. (1998). An introduction to animal
behaviour (5th ed.). Cambridge: Cambridge University Press.
Morel, L. (1982). Mise en place des processus de régulation du com-
portement agressif et de la reconnaissance entre ouvrières d’une
société de Camponotus vagus Scop. (Hymenoptera: Formicidae).
In A. De Harom & X. Espadaler (Eds.), La communication chez
les sociétés d’insectes (pp. 127-136). Barcelona: Press Universitat
Autònoma de Barcelona.
Nowbahari, E., Fénéron, R., & Malherbe, M.-C. (1999). Effect of
body size on aggression in the ant Cataglyphis niger (Hymenoptera;
Formicidae). Aggressive Behavior, 25, 369-379.
Nowbahari, E., & Lenoir, A. (1989). Age-related changes in aggres-
sion in ant Cataglyphis cursor (Hymenoptera, Formicidae): Influence
on intercolonial relationships. Behavioural Processes, 18, 173-181.
Nowbahari, E., Lenoir, A., Clément, J.-L., Lange, C., Bagnères,
A.-G., & Joulie, C. (1990). Individual, geographical and experimen-
tal variation of cuticular hydrocarbons of the ant Cataglyphis cursor
(Hymenoptera; Formicidae): Their use in nest and subspecies recogni-
tion. Biochemical Systematics & Ecology, 18, 63-73.
Obin, M. S., & Vander Meer, R. K. (1988). Sources of nestmate recog-
nition cues in the imported fire ant Solenopsis invicta Buran (Hyme-
noptera: Formicidae). Animal Behaviour, 36, 1361-1370.
Orivel, J., Errard, C., & Dejean, A. (1997). Ant gardens: Interspecific
recognition in parabiotic ant species. Behavioral Ecology & Sociobi-
ology, 40, 87-93.
Ozaki, M., Wada-Katsumata, A., Fujikawa, K., Iwasaki, M.,
Yokohari, F., Satoji, Y., et al. (2005). Ant nestmate and non-nestmate
discrimination by a chemosensory sensillium. Science, 309, 311-314.
Pfenning, D. W., Gamboa, G. J., Reeve, H. K., Shellman-Reeve, J.,
& Ferguson, I. D. (1983). The mechanism of nestmate discrimina-
tion in social wasps (Polistes, Hymenoptera: Vespidae). Behavioral
Ecology & Sociobiology, 13, 299-305.
Soroker, V., Vienne, C., Hefetz, A., & Nowbahari, E. (1994). The
postpharyngeal gland as a “Gestalt” organ for nestmate recognition in
the ant Cataglyphis niger. Naturwissenschaften, 81, 510-513.
Stuart, R. J. (1987). Transient nestmate recognition cues contrib-
ute to a multicolonial population structure in the ant Leptothorax
curvispinosus. Behavioral Ecology & Sociobiology, 21, 229-235.
Temeles, E. J. (1994). The role of neighbors in territorial systems: When
are they “dear enemies”? Animal Behaviour, 47, 339-350.
Todrank, J., & Heth, G. (2003). Odor–genes covariance and genetic
relatedness assessments: Rethinking odor-based “recognition” mech-
anisms in rodents. In P. Slater, J. Rosenblatt, C. Snowdon, & T. Roper
(Eds.), Advances in the study of behavior (Vol. 32, pp. 77-130). Am-
sterdam: Elsevier.
Vander Meer, R. K., Saliwanchik, D., & Lavine, B. (1989). Tem-
poral changes in colony cuticular hydrocarbon patterns of Solenopsis
invicta: Implications for nestmate recognition. Journal of Chemical
Ecology, 15, 2115-2125.
Velásquez, N., Gómez, M., González, J., & Vásquez, R. A. (2006).
Nest-mate recognition and the effect of distance from the nest on the
aggressive behaviour of Camponotus chilensis (Hymenoptera: Formi-
cidae). Behaviour, 143, 811-824.
Wilson, E. O. (1975). Enemy specification in the alarm-recruitment
system of an ant. Science, 190, 798-800.
Wyers, E. J., Peeke, H. V. S., & Herz, M. J. (1973). Behavioral ha-
bituation in invertebrates. In H. V. S. Peeke & M. J. Herz (Eds.), Ha-
bituation: Vol. I. Behavioral studies (pp. 1-57). New York: Academic
Press.
(Manuscript received February 10, 2006;
revision accepted for publication January 10, 2007.)
