Significance of chemical recognition cues is context dependent in ants

Abstract

Recognition of group members is of fundamental importance in social animals, allowing individuals to protect resources against intruders and parasites, as well as ensuring social cohesion within the group. In ants and other social insects, social recognition relies on multicomponent chemical signatures, composed primarily of long-chain cuticular hydrocarbons. These signatures are colony specific and allow discrimination between nestmates and non-nestmates. Nevertheless, the mechanisms underlying detection, perception and information processing of chemical signatures are poorly understood. It has been suggested that associative learning might play a role in nestmate recognition. We investigated whether Camponotus aethiops ants can associate a complete cuticular hydrocarbon profile, consisting of about 40 compounds, with a food reward and whether the new association, developed in an appetitive context, affects aggression against non-nestmates carrying the hydrocarbon profile associated with food. Individual ant workers were able to associate the non-nestmate chemical profile with food. However, conditioned ants were still aggressive when encountering a non-nestmate carrying the odour profile used as training odour in our experiments. This suggests that ants, like some, but not all other insects, show interactions between different modalities (i.e. olfactory and visual), and can treat complex chemical cues differently, according to the context in which they are perceived. This plasticity ensures that learning in an appetitive context does not interfere with the crucial task of colony defence.

Full text

Significance of chemical recognition cues is context dependent in ants
Nick Bos
a
,
*
, Fernando J. Guerrieri
a
,
b
, Patrizia d’Ettorre
a
,
c
a
Centre for Social Evolution, Department of Biology, University of Copenhagen
b
Max Planck Institute for Chemical Ecology, Department of Evolutionary Neuroethology, Germany
c
Laboratoire d’Ethologie Expérimentale et Comparée (LEEC), Université Paris 13, France
article info
Article history:
Received 10 March 2010
Initial acceptance 7 May 2010
Final acceptance 2 August 2010
Available online 9 September 2010
Keywords:
ant
associative learning
Camponotus aethiops
conditioning
context
cuticular hydrocarbon
perception
recognition
Recognition of group members is of fundamental importance in social animals, allowing individuals to
protect resources against intruders and parasites, as well as ensuring social cohesion within the group. In
ants and other social insects, social recognition relies on multicomponent chemical signatures, composed
primarily of long-chain cuticular hydrocarbons. These signatures are colony specific and allow discrimi-
nation between nestmates and non-nestmates. Nevertheless, the mechanisms underlying detection,
perception and information processing of chemical signatures are poorly understood. It has been suggested
that associative learning might play a role in nestmate recognition. We investigated whether Camponotus
aethiops ants can associate a complete cuticular hydrocarbon profile, consisting of about 40 compounds,
with a food reward and whether the new association, developed in anappetitive context, affects aggression
against non-nestmates carrying the hydrocarbon profile associated with food. Individual ant workers were
able to associate the non-nestmate chemical profile with food. However, conditioned ants were still
aggressive when encountering a non-nestmate carrying the odour profile used as training odour in our
experiments. This suggests that ants, like some, but not all other insects, show interactions between
different modalities (i.e. olfactoryand visual), and can treat complex chemical cues differently, according to
the context in which they are perceived. This plasticity ensures that learning in an appetitive context does
not interfere with the crucial task of colony defence.
Ó2010 The Association for the Study of Animal Behaviour. Published by Elsevier Ltd. All rights reserved.
Associative learning is a widespread phenomenon in the animal
kingdom (Giurfa 2007), allowing individuals to extract important
information from their environment by establishing predictive
relationships between different stimuli (i.e. honeybees, Apis melli-
fera:Bitterman et al. 1983;cuttlefish, Sepia officinalis:Cole & Adamo
2005;housemice,Mus domesticus:Watkins et al. 19 98). Although
insects have been generally neglected in studies addressing higher-
order cognitive processes, during the last three decades the
honeybee has become a model organism for the study of learning
and memory (Giurfa 2007; de Brito-Sánchez et al. 2008), since they
live in organized complex societies and show an amazing capacity
for learning.
All ants are eusocial, having cooperative brood care, reproductive
division of labour and overlapping generations. Ants often live in
very complex societies, and can learn to solve a variety of problems,
such as navigating in complex environments (Cataglyphis:Wehner
2009), visiting feeding places at specific times during the day
(Paraponera:Harrison & Breed 1987) or learning to avoid plants that
are detrimental for their fungus garden (Atta:Saverschek et al. 2010).
Some ants can learn and remember individual recognition cues
(d’Ettorre & Heinze 2005; Dreier et al. 2007) and selectively police or
punish colony members (e.g. Monnin et al. 2002; Van Zweden et al.
2007). However, ants have been largely ignored in studies on
learning, memory and cognition other than studies on spatial
memories (reviewed in Collett et al. 2006). A pioneering work by
Dupuy et al. (2006) showed that individual Camponotus ants can
learn to associate volatile substances with either a positive stimulus
(sucrose solution) or an aversive stimulus (quinine). In their study,
foraging ants had to make a choice between an odour associated
with sucrose solution (appetitive conditioned stimulus) and another
odour associated with quinine (aversive conditioned stimulus) in
a Y-maze. Individual Camponotus aethiops ants can be trained to
associate single synthetic long-chain hydrocarbons with a sucrose
reward (S. Dreier et al., unpublished data). These hydrocarbons have
a low volatility, are present on the ant cuticle and thus not usually
found in a foraging context. Instead, cuticular hydrocarbons (CHCs),
especially some classes of hydrocarbons, such as methyl-branched
alkanes, are important for nestmate recognition (e.g. Guerrieri et al.
2009; reviewed in d’Ettorre & Lenoir 2010). Ant colonies are good
targets for predators and parasites, since they contain many worker
*Correspondence: N. Bos, Centre for Social Evolution, Department of Biology,
University of Copenhagen, Universitetsparken 15, bygning 12, DK-2100 Copenha-
gen, Denmark.
E-mail address: nbos@bio.ku.dk (N. Bos).
Contents lists available at ScienceDirect
Animal Behaviour
journal homepage: www.elsevier.com/locate/anbehav
0003-3472/$38.00 Ó2010 The Association for the Study of Animal Behaviour. Published by Elsevier Ltd. All rights reserved.
doi:10.1016/j.anbehav.2010.08.001
Animal Behaviour 80 (2010) 839e844

ants, eggs, larvae, pupae and stored food. To maintain these valuable
resources, ants need to defend theircolonies from potential invaders.
Also, because ant workers usually do not reproduce, fitness benefits
are gained only if cooperative acts are directed towards nestmates,
which are usually related, and not towards alien individuals. This
requires an effective recognition system, allowing reliable discrimi-
nation of nestmates from non-nestmates.
Nestmate recognition is therefore essential for the stability and
success of insect societies. Because ant nests are usually dark,
a nonvisual recognition system is needed. Recognition cues are
chemical in nature and, as mentioned above, CHCs appear to be
important substances for chemical recognition in social insects
(i.e. bees: Breed 1998; wasps: Dani 2006; termites: Dronnet et al.
2006; ants: Hefetz 2007; d’Ettorre & Lenoir 2010). One of the first
studies showing the role of CHCs in ant recognition was performed
by Bonavita-Cougourdan et al. (1987), who removed the chemical
profile of freshly killed workers of Camponotus vagus and replaced
it with a chemical cuticular extract of a different colony. Workers
were significantly more aggressive against these treated nestmates
than against nontreated ones. In a subsequent study, a single
synthetic hydrocarbon was added to the cuticle of individual
C. vagus ants, which were then attacked by their nestmates
(Meskali et al. 1995); however, a recent study showed that this
might not be the case for all classes of hydrocarbons, although it is
the presence and not the absence of specific hydrocarbons on the
cuticle that promotes aggression (Guerrieri et al. 2009).
According to the currently accepted hypothesis, which has
recently been questioned (Ozaki et al. 2005; Guerrieri et al. 2009),
nestmate recognition occurs following a ‘labeletemplate matching
model’: each worker carries a set of recognition cues (the label) and
when an individual detects this label, it compares the label and its
inner template (a neural representation of the colony odour, stored
in the long-term memory). If the label is dissimilar to the template,
the worker will reject the encountered individual (Vander Meer &
Morel 1998; Lenoir et al. 1999). Previous studies specifically
showed that ants behave aggressively against the CHC extracts
from non-nestmates (Lahav et al. 1999; Lucas et al. 2005; Ozaki
et al. 2005). Guerrieri & d’Ettorre (2008) recently introduced
a controlled protocol for recording a clear binary response, named
the mandible opening response (MOR): harnessed ants (genus
Camponotus and Formica) will immediately open their mandibles,
as a sign of aggression, when presented with a glass rod coated with
the cuticular extracts of non-nestmates. Conversely, ants keep their
mandibles closed when presented with the cuticular extract of
nestmates. These studies demonstrate that the chemical stimulus
itself, that is, the CHC extract of non-nestmates, is sufficient to elicit
aggression in several ant species.
Knowing that ants behave aggressively against CHC extracts of
non-nestmates, and also that ants can be conditioned to associate
asynthetic chemical compound, including long-chain hydrocarbons,
with food (Dupuy et al. 2006; S. Dreier et al., unpublished data), we
investigated whether individual focal ants were able to associate the
chemical extract of the CHC profile of a non-nestmate with food, thus
transforming this non-nestmate cuticular hydrocarbon profile from
an aversive stimulus (see Guerrieri & d’Ettorre 2008)intoanappe-
titive stimulus. The question then arises, how would an ant that
previously associated a non-nestmate CHC profile with food react to
a real non-nestmate individual bearing the same CHC profile? We
predicted two mutually exclusive scenarios: (1) the non-nestmate
CHC profile changes its meaning, now indicating the presence of
food and not of a potential enemy, and thus the focal ant will not
behave aggressively towards the non-nestmate individual; (2) even
if in a foraging context the CHC profile indicated the presence of
a food source, the focal ant will still behave aggressively against
a non-nestmate bearing the food-associated CHC profile. Our
experimental design allowed us to distinguish between these two
predictions, although it remains to be investigated whether the ants
learn the entire cuticular hydrocarbon mixture or a subset of it.
METHODS
Study Organism
Twelve queenright colonies of C. aethiops (Latr.) were collected in
April 2008 in the Italian Apennines. Six colonies were collected in
Castel del Rio (44

21
0
62.83
0
N, 11

52
0
30.92
0
E) and six in Moraduccio
(44

10
0
32.75
0
N, 11

29
0
3.08
0
E). Each colony was housed in a plastic
box (27 17 cm and 9.5 cm high) with a plaster floor, serving as
a nest. This was connected to another plastic box of the same size,
serving as a foraging arena. The ants were fed twice a week with
diluted honey and mealworms, Tenebrio molitor; water was provided
ad libitum. The nests were kept in a climate room, at 25 2

C, and
a 12:12 h light:dark regime. Ants were deprived of honey at least 1
week before the experiment to increase motivation for foraging on
sucrose food sources.
Preparation of CHC Extracts
To condition individual ants to the chemical profile of non-
nestmates, we used cuticular extracts in pentane as a solvent
(SigmaeAldrich); this was the ‘training odour’. For each colony
source of training odour, we prepared five extracts in the
following way: 11 foragers (six major and five media workers)
were frozen and after 30 min their abdomen was cut off to prevent
any possible extraction of pheromones produced by glands. The
remaining parts of the bodies were inserted into a glass vial and
covered with 1 ml of pentane. After 10 min, the extract was
transferred to a new vial and the solvent allowed to evaporate;
chemical extracts were then redissolved in 50 0
m
lofpentane
before use in the following experiments.
Conditioning Set-up and Procedure
The ant nests used in conditioning trials were provided with
a vertical wooden stick in their foraging arena for the ants to climb
on. For each ant to be tested, seven petri dishes (100 mm diame-
ter 15 mm high) were prepared bycoating their walls with Fluon
and covering their floor with clean filter paper. Two microscope
cover slips (18 18 mm) were placed on each filter paper. One
cover slip was treated with 20
m
l of non-nestmate extract
(the training odour, as described above) deposited on its borders
and the solvent was allowed to evaporate completely. In the centre
of the cover slip, a droplet of sucrose solution (1
m
l, 30% w/w) was
deposited. On the other cover slip, we put 1
m
l of water, so that the
two slips looked exactly alike. The slips were positioned on oppo-
site sides of the petri dish (see Fig. 1a).
At the start of the first conditioning trial, a medium foraging
worker climbing on the verticalstick was gently allowed to walk onto
a small piece of paper, whichwasthen transferred into the petri dish
where theant was allowed toget off the paper at approximately equal
distance from both cover slips (see Fig. 1a). We then recorded the
time the ant required to find the sucrose reward. After the ant
finished drinking the sugar reward, it was picked up withsoft forceps,
and marked with a dot of enamel paint on its abdomen. The ant was
then allowed to walk on the piece of paper again, and transferred
back to the vertical stick inits colony of origin. Once back in the nest,
the ants usually transferred the food to their nestmates by trophal-
laxis and returned to the stick. After a minimum of 1 min, the same
ant was picked up again from thestick and transferred to a new petri
dish for a subsequent conditioning trial. A total of six conditioning
N. Bos et al. / Animal Behaviour 80 (2010) 839e844840

trials were performed per ant. The location of the slide with the
sucrose reward was randomized in each trial, but the cover slip was
never in the same quadrant for more than two consecutive trials to
prevent the ant from associating a given spatial direction with the
reward.
Following conditioning, a choice test was conducted to see
whether the ant had developed a positive association between
non-nestmate CHC extract and the sucrose reward. This choice test
consisted of presenting the focal ant, in a new petri dish, with one
cover slip coated with the CHC extract used as training odour, and
another one treated with solvent only (pentane, P). In this test
neither sucrose nor water was provided. The cover slips represent
the contextual stimulus in these tests. The petri dish was divided
into four quadrants by drawing a light cross in the middle of the
filter paper with a pencil, and the two slips were placed randomly
in two opposite quadrants (see Fig. 1b). The ant was transferred
from the stick to the petri dish, analogously to the conditioning
procedure, and was allowed to search for the expected reward for
2 min (although only the chemical stimulus, and not the reward,
was present during this time). The time spent by the focal ant in
each quadrant was recorded using the software EthoLog (Ottoni
2000). Afterwards, 1
m
l of sucrose solution was provided on the
CHC-coated cover slip to avoid any possible extinction effect. The
ant was then transferred back to the vertical stick in the foraging
arena of the colony.
As a control, naïve ants were subjected directly to the choice
test, without conditioning. Ten naïve ants were tested from each
colony used for conditioning (40 control ants in total).
Aggression Tests
Once the choice test was completed, the conditioned ant was left
in its colony for 10 min and then underwent an aggression test. We
measured aggression by placing the focal ant from the stick into
a neutral arena (50 mm diameter 60 mm height), where she was
allowed to habituate to the new environment for 1 min. After this,
a dead ant (freshly killed by freezing), from its own colony (nestmate
control), the colony carrying the CHC profile used as the training
odour or a novel colony (non-nestmate control), was placed in the
centre of the arena. In these tests, the dead ant represents the
contextual stimulus. The small size of the arena allowed for rapid
contact between the focal ant and the freshly killed ant (average
delay for first contact SE ¼7. 2 1.4 s). The behaviour of the focal
ant towards the dead ant was recorded continuously for 3 min using
the software EthoLog (Ottoni 2000). We evaluated any possible
aggression elicited by the target (dead ant) by recording six cate-
gorical behaviours: (1) no contact between the focal ant and the
target; (2) antennation; (3) grooming; (4) mandible opening;
(5) biting; and (6) gaster flexing (attempting to spread formic acid).
We established a baseline of aggression before starting the
conditioning experiments. Ten unconditioned focal ants and 10
target ants were used for each set of dyadic encounters (with an ant
from the focal ant’s own colony, the colony carrying the CHC profile
used as the training odour or a novel colony), giving a total of 30
aggression tests. This was replicated four times, using three new
colonies for each replicate, for a total of 12 colonies and 120
aggression tests.
In summary, 15 ants were conditioned. The aggression tests
involving these conditioned focal ants were repeated five times with
target nestmate ants, five times with target ants originating from the
colony whose cuticular extracts were used as the training odour and
five times with target ants from a novel non-nestmate colony. This
was replicated four times, involving a total of 12 colonies and 60
conditioned ants. As a control for the conditioning experiment, 40
additional naïve ants were used in choice tests, and to establish the
baseline of aggression,120 additional unconditioned ants were used.
Data Analysis
There were no significant differences in searching time between
colonies (KruskaleWallis ANOVA:
c
4
2
<9.49, P>0.05), so the data
could be pooled. Searching time in the course of the six condi-
tioning trials (1e6) were analysed using a Friedman ANOVA
followed by multiple comparisons (Siegel & Castellan 1988).
For the choice tests, a preference index (PI) was calculated using
the following formula, where t
TO
and t
P
is the time spent in quad-
rant TO (training odour) and P (pentane), respectively.
PI ¼t
TO
t
P
t
TO
þt
P
PI data were normally distributed (KolmogoroveSmirnov test:
P>0.2 in all cases). There was no colony effect on PI (ANOVA:
conditioned ants: F
3,56
¼1.47, P¼0.23; naïve ants: F
3,36
¼1.08,
P¼0.37); hence the data were pooled. If ants preferred one
quadrant over the other, the PI would differ significantly from zero.
This was analysed using a ttest for single means. Preference for
quadrants was also compared between conditioned and uncondi-
tioned (naïve) ants using a ttest for independent samples.
The different behaviours observed during aggression tests were
scored so that the most aggressive behaviour was assigned the
highest score: 0 ¼antennation and grooming; 1 ¼mandible
opening; 2 ¼biting; 3 ¼gaster flexing. The maximum level of
aggression was recorded, giving a value between 0 and 3 for every
ant tested.
Since there was no colony effect on overall aggression
(KruskaleWallis ANOVA: P>0.05), data could be pooled. The
difference in baseline aggression of unconditioned workers
towards nestmates, non-nestmate ants bearing the CHC profile
used for conditioning and novel non-nestmate ants was analysed
with a Wilcoxon signed-ranks tests.
The data were not normally distributed (KolmogoroveSmirnov
test: P<0.05) and thus a generalized linear model (GLM; Poisson
Sucrose
Non-nestmate
CHC profile
Non-nestmate
CHC profile
Water
Focal ant
Focal ant
Solvent
(a)
(b)
Figure 1. (a) Set-up for conditioning. (b) Set-up for choice test. For (a) the time the
focal ant spent finding the reward was recorded. For (b) the time the focal ant spent in
each of the four quadrants was recorded for 2 min.
N. Bos et al. / Animal Behaviour 80 (2010) 839e844 841

distribution, log link function) was used for analysing aggression
levels among naïve ants (the baseline of aggression) and condi-
tioned ants.
All analyses were performed using Statistica 7.1 (StatSoft, Tulsa,
OK, U.S.A.).
RESULTS
Conditioning
The searching time decreased significantly in the course of the
six conditioning trials (Friedman ANOVA:
c
5
2
¼97.17, N¼60,
P<0.05; Fig. 2). In particular, searching time decreased markedly
from the second trial on, showing that ants directed their search
more promptly towards the training odour.
In the control group (naïve ants), the PI did not differ signifi-
cantly from zero (one-sample ttest: t
39
¼0.27, P¼0.79), meaning
that ants had no spontaneous preference for a particular quadrant
during the choice test. However, for the conditioned ants, the PI
differed significantly from zero (one-sample ttest: t
59
¼19.59,
P<0.001). During the choice test, conditioned ants spent signifi-
cantly more time in the quadrant containing the training odour
than control (naïve) ants (ttest: t
98
¼10.60, P<0.001; Fig. 3).
Aggression
When the baseline of aggression was established, aggression of
unconditioned ant workers was significantly higher towards non-
nestmate ants bearing the CHC profile used for conditioning (b) or
novel non-nestmate ants (c) than towards nestmates (a; Wilcoxon
signed-ranks test: aaeab: T¼0.00, P<0.01; aaeac: T¼8.00,
P<0.01; abeac: T¼7. 50 , P>0.05; Bonferroni correction, adjusted
significance level
a
¼0.025). The results of the aggression tests are
shown in Fig. 4. Aggression levels of conditioned ants against either
ants from their own colony, non-nestmate ants bearing the CHC
profile used for conditioning or novel non-nestmate ants did not
differ significantly from the baseline of aggression of unconditioned
ants (GLM: Wald
c
2
2
¼2.783, P¼0.249), showing that conditioning
did not interfere with recognition abilities and expression of
aggressive behaviour.
DISCUSSION
We investigated whether ants are able to associate a cuticular
hydrocarbon mixture with food, and whether this association,
developed in an appetitive context, might affect nestmate discrimi-
nation, which is usually expressed in an aggressive context. We
conditioned freely walkingindividual C. aethiops workers to asso ciate
a cuticularextract of a non-nestmate with a sucrose reward,showing
that stimuli that originally elicit aggression in ants can be associated
with food. This suggests plasticity in the significance of cues/signals
that are typically important in the modulation of social interactions.
The fact that the conditioned ant kept coming back to the stick, and
that during the trials more ants were present on the stick, shows
clearlythat our set-up wasappropriate to simulate a foraging context.
Ants formed the association between the chemical blend and reward
very rapidly; right after the first training trial, the searching time
needed by the ant to find the reward decreased significantly.
A subsequent choice test showed that the association between the
280
260
240
220
200
180
160
140
120
100
80
60
40
20
0123456
Searching time (s)
Trial
Figure 2. Average searching time for each conditioning trial (N¼60) showing median,
quartiles and range.
0.4
0.35
0.3
0.25
0.2
0.15
0.1
0.05
0
Conditioned
Unconditioned
Preference index
Figure 3. Preference index of conditioned and unconditioned (naïve) ants (meanSE).
3.5
3
2.5
2
1.5
1
0.5
0Nestmate Familiar Novel
Aggression index
Conditioned
Unconditioned
Figure 4. Aggression index of conditioned and unconditioned ants versus nestmates,
familiar non-nestmates and novel non-nestmates, respectively (mean SE). Note that
the ‘familiar non-nestmates’are novel non-nestmates for the unconditioned ants, since
they have not been familiarized by conditioning.
N. Bos et al. / Animal Behaviour 80 (2010) 839e844842

food and the non-nestmate CHC profile was robustly established,
although we do not know whether the ants learned the entire CHC
profile or a subset of it (see below).
Individual olfactory learning in ants was shown for the first time
only recently (Dupuy et al. 2006), by conditioning individual ants to
pure volatile substances using a Y-maze. Although the substances
tested were present in either flowers or honeybee pheromones
(see Balderrama et al. 2002), they were not potential cues involved
in nestmate recognition. In addition, (S. Dreier et al., unpublished
data) showed that individual Camponotus ants can be conditioned
to associate single synthetic long-chain hydrocarbons with a sugar
reward. Unlike these two studies, in which only pure synthetic
chemical substances were used as conditioned stimuli, we used
a multicomponent blend, namely the cuticular extract of C. aethiops
workers, consisting of about 40 different hydrocarbons
(Van Zweden et al. 2009), as training odour. The extracted cuticular
hydrocarbons have the same properties as the natural mix on the
ant’s cuticle, as shown by Bonavita-Cougourdan et al. (1987), Morel
et al. (1988) and Nowbahari et al. (1990). These authors washed
ants in solvent to remove the cuticular hydrocarbons, and then
applied on these washed ants the cuticular extract of a nestmate or
non-nestmate. This restored the expected levelof aggression, while
washed ants did not elicit an aggressive response. In addition,
Ozaki et al. (2005) and Guerrieri & d’Ettorre (2008) showed that
a non-nestmate cuticular extract, when applied to a glass bead (or
rod), elicits aggression, proving that the CHC extract alone is
sufficient to promote aggression.
It remains to be determined whether the ants indeed learned the
entire CHC profile or only some specific compounds within the
profile. If the ants learned the entire CHC mixture of the profile, or
a large part of it, learning could have happened in various ways.
According to the elemental theory of learning (i.e. Rescorla & Wagner
1972), a mixture is processed as the sum of its components
(AB ¼AþB), while the configural theory (i.e. Pearce 1987) suggests
that the properties of a mixture are different from the properties of
the components (AB ¼XsAþB; Giurfa 2003). In honeybees,
a new model has been suggested: the key odorant hypothesis, which
encompasses features of both elemental and configural learning
(Reinhard et al. 2010). Even though honeybees were able to learn all
individual odorants used in the experiments; when conditioned to
a complex mixture, only certain key compounds were learned,
suggesting that some odorants are more representative of the
mixture than others.
In our experiment, we cannot distinguish between these
models. However, recent evidence indicates that methyl-branched
alkanes are more important in nestmate recognition than linear
alkanes (i.e. Guerrieri et al. 2009), suggesting that only key
compounds of the cuticular hydrocarbon blend present on the ant
might be used in recognition. Future studies should focus on which
classes of hydrocarbons are learned when a mixture is used as the
training odour.
In our experiments, the association of the CHC profile with food
did not affect the social meaning of the non-nestmate cuticular
profile. Indeed, after conditioning, the focal ants were as aggressive
towards an individual bearing the familiar odour (the cuticular
profile associated with food) as towards an individual bearing
a novel non-nestmate cuticular hydrocarbon profile.
Context-dependent learning in the visual modality has been
shown in a variety of organisms (e.g. ants: Chameron et al. 1998;
honeybees: Collett et al. 1993; humans: Smith & Vela 2001;for
a review on insects, see Collett et al. 2003). However, context-
dependent olfactory learning in insects has only been shown in the
cricket Gryllus bimaculatus (Matsumoto & Mizunami 2004) and the
cockroach Periplaneta americana (Sato et al. 2006). For instance,
individual crickets were conditioned to select one odour and to avoid
another odour in one context (darkness), and to do the opposite in
anothercontext (light). In thisexperiment, neitherthe light condition
nor the odour could predict a reward (water) or non-reward
(saline solution); only a combination of the two could. In contrast,
Drosophila melanogaster is not capable of solving a similar task (Yara li
et al. 2008). Groups of flies were conditioned in a ‘biconditional
discrimination’design. One odour was paired with an electric shock
in darkness, but not in light;another odour waspaired with a shock in
light, but not darkness. Flies failed to show any preference for the
right odour in the specific light conditions. Therefore, interactions
between different modalities (i.e. visual and olfactory), which are
necessary for context-dependent learning, are not a general feature
of all insects. Two explanations of context-dependent learning have
been proposed (Pearce & Bouton 2001). In the occasion-setting
theory, the contextual stimulus (in our case, the cover slip or the
target ant) ‘sets the occasion’for responding to another conditioned
stimulus (in our case, the training odour: the non-nestmate CHC
profile), without forming an association with the unconditioned
stimulus (the food reward) itself. In the configural theory, an indi-
vidual perceives different CSecontext combinations as different
stimuli. Although learning in our experimental design was clearly
context dependent, further studies will be needed to distinguish
between these two theories.
Living in a complex and changing environment requires being
aware that the same cues/signals might have different significance
according to context and the role of individual experience is
extremely important. This learning capability and flexibility is an
adaptive trait for the ants, as it maximizes their fitness and survival:
on the one hand, ants can eventually find food by following the
smell of non-nestmates, but still avoid or attack enemies that they
encounter. Our study shows that ants have complex cognitive
abilities and are capable of distinguishing the meaning of identical
cues according to the context in which they are perceived.
Acknowledgments
This work was supported by the Marie Curie Excellence grant
CODICES (MEXT-CT-2004-014202) and a Freia grant from the
Faculty of Science, University of Copenhagen, both assigned to P.d’E.
We are grateful to all members of the Copenhagen Centre for Social
Evolution for a stimulating work environment. We thank three
anonymous referees for useful comments and suggestions.
References
Balderrama, J., Núñez, J., Guerrieri, F. J. & Giurfa, M. 2002. Different functions of
two alarm substances in the honeybee. Journal of Comparative Physiology A:
Neuroethology, Sensory, Neural, and Behavioral Physiology,188 , 485e491.
Bitterman, M. E., Menzel, R., Fietz, A. & Schafer, S. 1983. Classical conditioning of
proboscis extension in honeybees (Apis mellifera). Journal of Comparative
Psychology,97,107e119.
Bonavita-Cougourdan, A., Clément, J. L. & Lange, C. 1987. Nestmate recognition:
the role of cuticular hydrocarbons in the ant Camponotus vagus Scop. Journal of
Entomological Science,22,1e10.
Breed, M. D. 1998. Pheromones of the honey bee. BioScience,48, 463e470.
de Brito-Sánchez, M. G., Deisig, N., Sandoz, J. C. & Giurfa, M. 2008. Neurobiology
of olfactory communication in a social insect: the honeybee. In: Sociobiology of
Communication: an Interdisciplinary Perspective (Ed. by P. d’Ettorre &
D. P. Hughes), pp. 119e138. Oxford: Oxford University Press.
Chameron, S., Schatz, B., Pastergue Ruiz, I., Beugnon, G. & Collett, T. S. 1998. The
learning of a sequence of visual patterns by the ant Cataglyphis cursor.
Proceedings of the Royal Society B,265, 2309e2313.
Cole, P. & Adamo, S. A. 2005. Cuttlefish (Sepia officinalis: Cephalopoda) hunting
behavior and associative learning. Animal Cognition,8,27e30.
Collett, T. S., Fry, R. & Wehner, R. 1993. Sequence learning by honeybees. Journal of
Comparative Physiology A: Neuroethology, Sensory, Neural, and Behavioral Physi-
ology,172, 693e706.
Collett, T. S., Fauria, K. & Dale, K. 2003. Contextual cues and insect navigation. In:
The Neurobiology of Spatial Behaviour (Ed. by K. J. Jeffery), pp. 67e82. Oxford:
Oxford University Press.
N. Bos et al. / Animal Behaviour 80 (2010) 839e844 843

Collett, T. S., Graham, P., Harris, R. A. & Hempel De Ibarra, N. 2006. Navigational
memories in ants and bees: memory retrieval when selecting and following
routes. Advances in the Study of Behavior,36,123e17 2.
Dani, F. R. 2006. Cuticular lipids as semiochemicals in paper wasps and other social
insects. Annales Zoologici Fennici,43,500e514.
Dreier, S., Van Zweden, J. S. & D’Ettorre, P. 2007. Long-term memory of individual
identity. Biology Letters,3, 459e462.
Dronnet, S., Lohou, C., Christides, J. P. & Bagneres, A. G. 2006. Cuticular
hydrocarbon composition reflects genetic relationship among colonies of the
introduced termite Reticulitermis santonensis Feytaud. Journal of Chemical
Ecology,32, 1027e1042.
Dupuy, F., Sandoz, J. C., Giurfa, M. & Josens, R. 2006. Individual olfactory learning
in Camponotus ants. Animal Behaviour,72, 1081e1091.
d’Ettorre, P. & Heinze, J. 2005. Individual recognition in ant queens. Current
Biology,15,2170e2174.
d’Ettorre, P. & Lenoir, A. 2010. Nestmate recognition. In: Ant Ecology (Ed. by L. Lach,
C. Parr & K. Abbot), pp. 194e209. Oxford: Oxford University Press.
Giurfa, M. 2003. Cognitive neuroethology: dissecting non-elemental learning in
a honeybee brain. Current Opinion in Neurobiology,13,726e735.
Giurfa, M. 2007. Behavioral and neural analysis of associative learning in the
honeybee: a taste from the magic well. Journal of Comparative Physiology
A: Neuroethology, Sensory, Neural, and Behavioral Physiology,193,801e824.
Guerrieri, F. J. & d’Ettorre, P. 2008. The mandible opening response: quantifying
aggression elicited by chemical cues in ants. Journal of Experimental Biology,211,
110 9e1113.
Guerrieri, F. J., Nehring, V., Jørgensen, C., Nielsen, J., Galizia, G. & d’Ettorre,
P. 2009. Ants recognize foes and not friends. Proceedings of the Royal Society B,
276, 2461e2468.
Harrison, J. M. & Breed, M. D. 1987. Temporal learning in the giant tropical ant,
Paraponera clavata.Physiological Entomology,12,317e320.
Hefetz, A. 2007. The evolution of hydrocarbon pheromone parsimony in ants
(Hymenoptera: Formicidae): interplay of colony odor uniformity and odor
idiosyncrasy. A review. Myrmecological News,10,59e68.
Lahav, S., Soroker, V., Hefetz, A. & Vander Meer, R. K. 1999. Direct behavioral
evidence for hydrocarbons as ant recognition discriminators. Natur-
wissenschaften,86, 246e249.
Lenoir, A., Fresneau, D., Errard, C. & Hefetz, A. 1999. Individuality and colonial
identity in ants. In: Information Processing in Social Insects (Ed. by C. Detrain,
J. L. Deneubourg & J. Pasteels), pp. 219e237. Basel: Birkhauser Verlag.
Lucas, C., Pho, D. B., Jallon, J. M. & Fresneau, D. 20 05. Role of cuticular
hydrocarbons in the chemical recognition between ant species in the
Pachycondyla villosa species complex. Journal of Insect Physiology,51,
114 8e115 7.
Matsumoto, Y. & Mizunami, M. 2004. Context-dependent olfactory learning in an
insect. Learning & Memory,11, 288e293.
Meskali, M., Provost, E., Bonavita-Cougourdan, A. & Clément, J. L. 1995. Behav-
ioural effects of an experimental change in the chemical signature of the ant
Camponotus vagus (Scop.). Insectes Sociaux,42,347e358.
Monnin, T., Ratnieks, F. L., Jones, G. R. & Beard, R. 2002. Pretender punishment
induced by chemical signalling in a queenless ant. Nature,419,61e65.
Morel, L., Vander Meer, R. K. & Lavine, B. K. 1988. Ontogeny of nestmate recog-
nition cues in the red carpenter ant (Camponotus floridanus). Behavioral and
chemical evidence for the role of age and social experience. Behavioral Ecology
and Sociobiology,22,175e183.
Nowbahari, E., Lenoir, A., Clément, J. L., Lange, C., Bagneres, A. G. & Joulie, C.
1990. Individual, geographical and experimental variation of cuticular hydro-
carbons of the ant Cataglyphis cursor (Hymenoptera: Formicidae): their use in
nest and subspecies recognition. Biochemical Systematics and Ecology,18,63e73.
Ottoni, E. B. 2000. Etholog 2.2: a tool for the transcription and timing of behavior
observation sessions. Behavior Research Methods, Instruments & Computers,32,
446e449.
Ozaki, M., Wada-Katsumata, A., Fujikawa, K., Iwasaki, M., Yokohari, F., Satoji, Y.,
Nisimura, T. & Yamaoka, R. 2005. Ant nestmate and non-nestmate discrimi-
nation by a chemosensory sensillum. Science,309,311e314.
Pearce, J. M. 1987. A model for stimulus generalization in Pavlovian conditioning.
Psychological Review,94,61e73.
Pearce, J. M. & Bouton, M. E. 2001. Theories of associative learning in animals.
Annual Review of Psychology,52,111e139.
Reinhard, J., Sinclair, M., Srinivasan, M. V. & Claudianos, C. 2010. Honeybees learn
odour mixtures via a selection of key odorants. PloS ONE,5, e9110.
Rescorla, R. A. & Wagner, A. R. 1972. A theory of Pavlovian conditioning: variations
in the effectiveness of reinforcement and nonreinforcement. In: Classical
Conditioning II (Ed. by A. H. Black & W. F. Prokasy), pp. 64e99. New York:
Appleton-Century-Crofts.
Sato, C., Matsumoto, Y., Sakura, M. & Mizunami, M. 2006. Contextual olfactory
learning in cockroaches. Neuroreport,17, 553e557.
Saverschek, N., Herz, H., Wagner, M. & Roces, F. 2010. Avoiding plants unsuitable
for the symbiotic fungus: learning and long-term memory in leaf-cutting ants.
Animal Behaviour,79,689e698.
Siegel, S. & Castellan, N. J. 1988. Nonparametric Statistics for the Behavioral Sciences.
2nd edn. Singapore: McGraw-Hill.
Smith, S. M. & Vela, E. 2001. Environmental context-dependent memory: a review
and meta-analysis. Psychonomic Bulletin & Review,8, 203e220.
Van Zweden, J. S., Fürst, M. A., Heinze, J. & d’Ettorre, P. 2007. Specialization in
policing behaviour among workers in the ant Pachycondyla inversa.Proceedings
of the Royal Society B,274, 1421e1428.
Van Zweden, J. S., Dreier, S. & d’Ettorre, P. 2009. Disentangling environmental and
heritable nestmate recognition cues in a carpenter ant. Journal of Insect
Physiology,55,159e164.
Vander Meer, R. K. & Morel, L. 1998. Nestmate recognition in ants. In: Pheromone
Communication in Social Insects (Ed. by R. K. Vander Meer, M. Breed, M. Winston
& K. E. Espelie), pp. 79e103. Boulder, Colorado: Westview Press.
Watkins, R. W., Gurney, J. E. & Cowan, D. P. 1998. Taste-aversion conditioning of
house mice (Mus domesticus) using the non-lethal repellent, cinnamamide.
Applied Animal Behaviour Science,57,171e177.
Wehner, R. 2009. The architecture of the desert ant’s navigational toolkit
(Hymenoptera: Formicidae). Myrmecological News,12,85e96.
Yarali, A., Mayerle, M., Nawroth, C. & Gerber, B. 2008. No evidence for visual
context-dependency of olfactory learning in Drosophila.Naturwissenschaften,
95,767e774.
N. Bos et al. / Animal Behaviour 80 (2010) 839e844844