Learning and perceptual similarity among cuticular hydrocarbons in ants

Abstract

Nestmate recognition in ants is based on perceived differences in a multi-component blend of hydrocarbons that are present on the insect cuticle. Although supplementation experiments have shown that some classes of hydrocarbons, such as methyl branched alkanes and alkenes, have a salient role in nestmate recognition, there was basically no information available on how ants detect and perceive these molecules. We used a new conditioning procedure to investigate whether individual carpenter ants could associate a given hydrocarbon (linear or methyl-branched alkane) to sugar reward. We then studied perceptual similarity between a hydrocarbon previously associated with sugar and a novel hydrocarbon. Ants learnt all hydrocarbon-reward associations rapidly and with the same efficiency, regardless of the structure of the molecules. Ants could discriminate among a large number of pairs of hydrocarbons, but also generalised. Generalisation depended both on the structure of the molecule and the animal's experience. For linear alkanes, generalisation was observed when the novel molecule was smaller than the conditioned one. Generalisation between pairs of methyl-alkanes was high, while generalisation between hydrocarbons that differed in the presence or absence of a methyl group was low, suggesting that chain length and functional group might be coded independently by the ant olfactory system. Understanding variations in perception of recognition cues in ants is necessary for the general understanding of the mechanisms involved in social recognition processes based on chemical cues.

Full text

Learning and perceptual similarity among cuticular hydrocarbons in ants
Nick Bos
a,1,
⇑
, Stephanie Dreier
a,1
, Charlotte G. Jørgensen
a,b
, John Nielsen
c
, Fernando J. Guerrieri
a,d
,
Patrizia d’Ettorre
a,e
a
Centre for Social Evolution, Department of Biology, University of Copenhagen, Universitetsparken 15, 2100 Copenhagen, Denmark
b
Department of Medicinal Chemistry, University of Copenhagen, Jagtvej 162, 2100 Copenhagen, Denmark
c
Department of Life Sciences, Bioorganic Chemistry, University of Copenhagen, Thorvaldsensvej 40, DK-1871 Frederiksberg, Denmark
d
Max Planck Institute for Chemical Ecology, Department of Evolutionary Neuroethology, Jena, Germany
e
Laboratory of Experimental and Comparative Ethology (LEEC), 13, University of Paris, France
article info
Article history:
Received 25 July 2011
Received in revised form 24 October 2011
Accepted 25 October 2011
Available online xxxx
Keywords:
Ants
Camponotus
Conditioning
Generalisation
Learning
Nestmate recognition
abstract
Nestmate recognition in ants is based on perceived differences in a multi-component blend of hydrocar-
bons that are present on the insect cuticle. Although supplementation experiments have shown that
some classes of hydrocarbons, such as methyl branched alkanes and alkenes, have a salient role in nest-
mate recognition, there was basically no information available on how ants detect and perceive these
molecules. We used a new conditioning procedure to investigate whether individual carpenter ants could
associate a given hydrocarbon (linear or methyl-branched alkane) to sugar reward. We then studied per-
ceptual similarity between a hydrocarbon previously associated with sugar and a novel hydrocarbon.
Ants learnt all hydrocarbon-reward associations rapidly and with the same efficiency, regardless of the
structure of the molecules. Ants could discriminate among a large number of pairs of hydrocarbons,
but also generalised. Generalisation depended both on the structure of the molecule and the animal’s
experience. For linear alkanes, generalisation was observed when the novel molecule was smaller than
the conditioned one. Generalisation between pairs of methyl-alkanes was high, while generalisation
between hydrocarbons that differed in the presence or absence of a methyl group was low, suggesting
that chain length and functional group might be coded independently by the ant olfactory system. Under-
standing variations in perception of recognition cues in ants is necessary for the general understanding of
the mechanisms involved in social recognition processes based on chemical cues.
Ó2011 Elsevier Ltd. All rights reserved.
1. Introduction
The ability to discriminate among different stimuli is funda-
mental in many aspects of an animal’s life, from food location
and predator avoidance to social behaviour and communication.
Social insects are notable models for studying stimulus detection,
perception, learning and memory because they combine behav-
ioural plasticity and experimental accessibility. It has been long
known that honeybees can learn relationships between different
sensory cues (e.g. colour or odour) with a reward of sugar solution
(Frisch, 1919). Since then, the honeybee has become a model
organism for a plethora of cognitive studies analysing the func-
tional properties of sensory systems (review in Menzel and Giurfa,
2006). In natural conditions, honeybees encounter floral odour
blends consisting of tens to hundreds of components (Knudsen
et al., 1993) and individual components of odorants can be catego-
rised according to the selectivity of receptor neurons and behav-
ioural discrimination abilities (Akers and Getz, 1992, 1993).
Recently, Reinhard et al. (2010) showed that when honeybees are
conditioned to a complex blend of odours, they only learn specific
key odorants. No correlation was found with molecular structure,
but concentration appeared to have a positive effect on whether
an odorant would become a key odorant. However, other studies
suggested that the perceptual qualities of odours do correlate with
specific molecular features. For example, changes in carbon chain
length or structure, or changes in functional groups, affect general-
isation from conditioned to test odorants (Laska et al., 1999;
Guerrieri et al., 2005a). There is also evidence for a neuro-chemical
modulation of perception, for instance, bees injected with octopa-
mine agonists significantly increased their ability to discriminate
nestmates from non-nestmates based on chemical cues (Robinson
et al., 1999).
Nestmate recognition is a fundamental feature of insect socie-
ties and thus the identification of social recognition cues and the
study of recognition mechanisms are essential for an integrated
understanding of their advanced social organisation. Social recog-
nition cues are encoded in the complex pattern of hydrocarbons
present on the cuticle of social insects; these cuticular hydrocar-
0022-1910/$ - see front matter Ó2011 Elsevier Ltd. All rights reserved.
doi:10.1016/j.jinsphys.2011.10.010
⇑
Corresponding author. Tel.: +45 353 21257; fax: +45 353 21250.
E-mail address: Nbos@bio.ku.dk (N. Bos).
1
These authors contributed equally to this study.
Journal of Insect Physiology xxx (2011) xxx–xxx
Contents lists available at SciVerse ScienceDirect
Journal of Insect Physiology
journal homepage: www.elsevier.com/locate/jinsphys
Please cite this article in press as: Bos, N., et al. Learning and perceptual similarity among cuticular hydrocarbons in ants. Journal of Insect Physiology
(2011), doi:10.1016/j.jinsphys.2011.10.010

bons generally range from about 20 to over 40 carbons in chain-
length, with three principal structural classes: linear alkanes,
methyl-branched alkanes and alkenes (review in e.g. Lenoir et al.,
1999; Howard and Blomquist, 2005; Hefetz, 2007; Martin and Dri-
jfhout, 2009; d’Ettorre and Lenoir, 2010). It has been suggested that
certain hydrocarbon classes, such as methyl-alkanes and alkenes,
have evolved a signal function while other classes, such as alkanes,
serve little role in nestmate recognition (cf. in wasps: Dani et al.,
2001; in honeybees: Dani et al., 2005; in ants: Akino et al., 2004;
Lucas et al., 2005; Martin et al., 2008a; Guerrieri et al., 2009; but
see Greene and Gordon, 2007). Indeed, honeybees tested with
the classical ‘proboscis extension response’ conditioning paradigm
appear to learn certain alkenes better than alkanes when con-
fronted with a discriminatory task (Châline et al., 2005).
Ants, like honeybees, can learn to solve different tasks (e.g. Weh-
ner, 2003), and particularly Camponotus ants can be trained in con-
trolled laboratory conditions to investigate learning and odour
discrimination. This has been done with free walking ants (Dupuy
et al., 2006); and with harnessed ants (Guerrieri and d’Ettorre,
2010). Dupuy et al. (2006) studied free walking workers of two dif-
ferent Camponotus species that were presented with two volatile
substances. One substance was positively reinforced with sugar
solution, while the other was negatively reinforced with quinine
solution. Individual ants were shown to learn single odours effi-
ciently. The substances tested were volatile odour compounds either
present in flowers or in honeybee pheromones (see Balderrama
et al., 2002).
Chemical communication is especially important in nestmate
recognition and ourunderstanding of what part of the multi-compo-
nent chemical cues of the ant’s cuticle might contain information
about species, nestmate, task or fertility recognition is increasing
(e.g. Greene and Gordon, 2003, 2007; d’Ettorre et al., 2004; Hefetz,
2007; d’Ettorre and Lenoir, 2010; Holman et al., 2010). In contrast,
we know very little about how ants detect and perceive such infor-
mation, and so far, no study has systematically investigated percep-
tual similarity among cuticular hydrocarbons in ants. Meskali et al.
(1995) have supplemented the cuticle of Camponotus vagus workers
with (Z)-9-tricosene, an alkene naturally absent in this species. This
change in the composition of cuticular hydrocarbons was followed
by a significant increase in antennation level and threats expressed
by non-treated individuals towards treated individuals. In C. vagus,it
has also been established that the proportions of certain hydrocar-
bons vary among worker subcastes suggesting that cuticular hydro-
carbons serve as cues allowing discrimination between subcastes
(Bonavita-Cougourdan et al., 1993).
Fine-tuned discrimination of cuticular hydrocarbons is expected
to increase the accuracy of nestmate and within-colony recognition.
Conversely, ants need to categorise cuticular profiles by taking inter-
individual variation into account. This individual variation requires
generalisation between cuticular profiles that are similar but not
identical to decrease errors in discrimination between nestmates
and non-nestmates (van Zweden and d’Ettorre, 2010). Indeed,
Argentine ants show similar levels of aggression toward nestmate
cuticular profiles supplemented with hydrocarbons with the same
branch position but differing in chain length, suggesting that they
are perceived as similar (van Wilgenburg et al., 2010).
Here we studied in individual Camponotus aethiops ants: (i) asso-
ciative learning and perception of single hydrocarbons usually found
on the cuticle of ants; (ii) the relationship between structure of
hydrocarbons and perceptual similarity (generalisation). Using an
original experimental procedure, we conditioned individual foragers
of C. aethiops to associate a given single hydrocarbon with sugar solu-
tion (reward), and subsequentlytested the conditioned ants for gen-
eralisation between the conditioned hydrocarbon and a novel
hydrocarbon.We used an array of eight hydrocarbons combining dif-
ferent carbon chain lengths and presence/absence of a methyl group.
2. Materials and methods
2.1. Study organism
Nine queenright colonies of the carpenter ant C. aethiops (Latr.)
were collected in spring 2006 in the Apennines, Italy (Moraduccio,
44°10
0
32.75
00
N, 11°29
0
3.08
00
E) and brought to our laboratory in
Copenhagen. Each colony was housed in a plastic box
(27 17 9.5 cm) with a plaster floor, which was connected to
an equally sized plastic box serving as a foraging arena. The colo-
nies were kept at 22 ± 2°C under a 12:12 LD photoperiod. Each for-
aging arena contained a vertical wooden stick (platform) on which
ants could be collected and put back (cf. Dupuy et al., 2006;Fig. 1A
and B). Ants were fed with honey water and mealworms (Tenebrio
molitor, Linn.). However, in order to increase ant motivation for su-
gar rewards, each colony was deprived of honey 10 days before the
start of the experiments. All experiments were performed with
individually marked workers. Each test ant was marked with a
dot of enamel paint on its thorax after the first training trial. Test
ants were typically foragers, i.e. medium-size workers present in
the foraging arena.
2.2. Experimental setup
We trained individual ants to associate a sugar solution reward
(30% w/w) with a hydrocarbon (S+), and later tested them for dis-
crimination between the trained stimulus (S+) and a novel hydro-
carbon (N) in the absence of a reward. Each hydrocarbon was
diluted in pentane (Sigma–Aldrich, HPLC grade, >99.9% purity) to
a final concentration of 10
l
g/ml.
Training was conducted in Petri dishes (100 mm diame-
ter 15 mm high), and a clean Petri dish was used for each trial.
The bottom of each Petri dish was covered with filter paper and
the side walls were coated with Fluon
Ò
to prevent the ants from
escaping. One of the quadrants contained a glass cover slide
(18 18 mm) treated with a hydrocarbon (S+), by applying 20
l
l
of hydrocarbon solution on the edges (making sure that the solu-
tion did not touch the filter paper). This solution contained
0.2
l
g of hydrocarbon, which is within the natural range found
on the cuticle of ants of similar size (Holman et al., 2010). The trea-
ted cover slide was baited with a 1
l
l droplet of sugar solution
placed on its centre. In the opposite quadrant, a similar cover slide
held a 1
l
l droplet of water, but no hydrocarbon (Fig. 1C). The loca-
tion of the cover slides was pseudo-randomised. The slides were
never positioned in the same location for more than two consecu-
tive trials to prevent the ants from using any directional prefer-
ences for locating the sugar solution.
2.3. Training procedure
At the beginning of each experiment, an ant was gently taken
from the platform placed within the foraging arena and transferred
to the Petri dish using a piece of filter paper the ant could climb on.
The ant was randomly released in one of the two empty quadrants
and allowed to search for the reward (Fig. 1C). When the sugar
solution was found, the ant was allowed to ingest it. The time
needed to find the sugar reward was recorded. Subsequently, the
ant was gently recaptured with the help of the filter paper and re-
leased on the platform in its colony, so that it could perform troph-
allaxis with nestmates. Once its intake was unloaded, a motivated
forager typically came back to the platform within 3–4 min, and
was at that point picked up for the next trial. The ant was always
left for a minimum of 1 min in its colony. Each ant performed six
consecutive training trials, in order to build an association between
a given hydrocarbon (S+) and the reward.
2N. Bos et al. / Journal of Insect Physiology xxx (2011) xxx–xxx
Please cite this article in press as: Bos, N., et al. Learning and perceptual similarity among cuticular hydrocarbons in ants. Journal of Insect Physiology
(2011), doi:10.1016/j.jinsphys.2011.10.010

2.4. Test phase
After the training phase, two choice tests were conducted. The
surface of the Petri dish was divided by a light pencil line into four
quadrants of equal size. The test Petri dish gave the ant a choice be-
tween two cover slides, one treated with the trained stimulus (S+)
and the other one with a novel hydrocarbon (N). This time how-
ever, no reward was present on either slide (Fig. 1D). The cover
slides where placed in opposite quadrants and given that the
hydrocarbons have low volatility the ants had to approach the cov-
er slide to perceive the stimulus. The ant was allowed to walk for
2 min, during which its location was recorded. The time the ant
spent in each of the four quadrants of the Petri dish was recorded
using the software EthoLog v. 2.2.5 (Ottoni, 2000). Afterwards, a
droplet of sucrose solution was provided on the coverslip contain-
ing S+. The ant was transferred back in its colony, and after a min-
imum of 1 min, a second choice test was conducted. During the
second choice test, the location of S+ and N was switched. Any
association between a hydrocarbon and reward would be evinced
by the ant remaining more time in the quadrants containing the
hydrocarbons than in the empty quadrants. Perceptual dissimilar-
ity (low generalisation) between hydrocarbons presented in the
test would be evinced by the ants spending more time in the S+
quadrant than in the N quadrant.
A control test was conducted for each hydrocarbon, where a
naïve ant (without training) had a choice between a slide with a
hydrocarbon, and a slide with pentane. No preference for any
quadrant was expected in this control test, indicating no spontane-
ous preference for a given hydrocarbon.
A total of eight hydrocarbons, including five linear alkanes and
three mono-methyl alkanes were used (Supplementary Table 1S),
some of which are present on the cuticle of C. aethiops (van Zweden
et al., 2009). We chose pairs of hydrocarbons differing in structural
similarity as S+ and N stimuli. For example, when conditioned with
a linear hydrocarbon, the ant was presented in the test with an-
other linear hydrocarbon of different chain-length, or with a
branched hydrocarbon having a methyl group in a certain position
(carbon 3 or carbon 11). All hydrocarbons used as S+ were also
used as N for another, fully independent, set of ants. A total of 24
independent pairs of hydrocarbons were tested. For each pair of
hydrocarbons, 10 replicates were performed representing 10 ants
conditioned and tested individually, for a total of 240 ants from
nine colonies. For the control tests (naïve ants), 10 ants were used
for each hydrocarbon, for a total of 80 ants tested.
2.5. Data analysis and statistics
During the training trials we measured the time required by
each ant to find the reward. The performance over the six trials
was analysed using a Friedman’s ANOVA, followed by post hoc
comparison to assess differences between trial 1 and trial 6 (Wilco-
xon test for matched pairs). During the choice test we measured
Fig. 1. Experimental setup. (A) Platform in the foraging arena. (B) A C. aethiops worker on the top of the platform. (C) Training situation: the ant must find the sugar solution
reward associated with a hydrocarbon. In an opposite side a droplet of water is offered (no hydrocarbon present). (D) Test situation: the ant was offered two different
hydrocarbons in absence of reward, one previously associated with the reward and a novel hydrocarbon.
N. Bos et al. / Journal of Insect Physiology xxx (2011) xxx–xxx 3
Please cite this article in press as: Bos, N., et al. Learning and perceptual similarity among cuticular hydrocarbons in ants. Journal of Insect Physiology
(2011), doi:10.1016/j.jinsphys.2011.10.010

the time that each ant spent in the four quadrants. The time was
subsequently normalised, and differences between the percentage
of time spent in quadrant S+ and the percentage of time spent in
quadrant N were analysed using Wilcoxon test for matched pairs,
and effect sizes were calculated using GPower 3.1.3.
A preference index (PI) was calculated and defined as (t
S+
t
N
)/
(t
S+
+t
N
). This PI varies between 1 and +1; with 0 meaning that
there was no preference between S+ and N (complete generalisa-
tion), positive values meaning preference for S+ and negative val-
ues preference for N.
Comparison between PI from hydrocarbon A to B and PI from B
to A (between symmetric pairs) was done using Mann–Whitney U
tests. In order to establish whether there was a defined pattern for
asymmetric generalisation, the analysis between difference in car-
bon chain length (S+ N) and preference index was performed
separately for pairs containing linear alkanes and pairs containing
at least one methyl-alkane using Spearman rank correlations.
2.6. Synthetic hydrocarbons
The linear alkanes were purchased from Sigma–Aldrich (Stein-
heim, Germany), and the mono-methyl-akanes were all synthes-
ised as described in the Supplementary material.
Time to Fi n d the Reward (s)
0
20
40
60
80
100
0
20
40
60
80
100
n-C
21
n-C
20
n-C
19
n-C
28
n-C
22
0
20
40
60
80
100
3-MeC
27
123456
3-MeC
31
Training Trials
0
20
40
60
80
100
123456
11-MeC
27
n=30
n=20
n=40
n=30
n=40
n=30
n=30 n=20
Z=4.28 Z=3.53
Z=4.19
Z=4.2 3
Z=2.68 Z=4.06
Z=4.09
Z=2.3 0
**
**
**
**
**
**
**
*
Fig. 2. Search time for each hydrocarbon tested. The time (s) until the ants found the sugar solution (reward) was lower in trial 6 than in trial 1 for all eight hydrocarbons
tested. Significant differences between trials 1 and 6 are indicated by stars. Medians and 25th and 75th percentiles are shown. Wilcoxon’s tests for matched pairs (Zvalue are
shown).
⁄
P< 0.05,
⁄⁄
P< 0.01; nis the sample size (number of ants tested).
4N. Bos et al. / Journal of Insect Physiology xxx (2011) xxx–xxx
Please cite this article in press as: Bos, N., et al. Learning and perceptual similarity among cuticular hydrocarbons in ants. Journal of Insect Physiology
(2011), doi:10.1016/j.jinsphys.2011.10.010

3. Results
3.1. Training: Building the association between a hydrocarbon and
sucrose reward
The time needed by the ant to find the sugar reward diminished
significantly over the six trials for all hydrocarbons tested (Fried-
man’s ANOVA,
v
2
= 11.74–39.30, df = 5, n= 20–30, P< 0.05 in all
cases). Indeed, trials 1 and 6 were always significantly different
(Fig. 2; eight hydrocarbons tested, Wilcoxon tests for matched
pairs, Z= 2.30–4.28, n= 20–30, P< 0.05 in all cases).
3.2. Choice test: Specific response to the hydrocarbon associated to
sucrose
In 23 out of 24 cases, the ants spent significantly more time in
the quadrants containing hydrocarbons than in the empty quad-
rants (Wilcoxon test for matched pairs, Z=1.89–2.80, n=10,
P< 0.05). The exception (n-C
19
/n-C
28
) showed an almost significant
trend (Wilcoxon test for matched pairs, Z= 1.89, n= 10, P< 0.06).
So, we performed the subsequent analyses on the time spent by
the ants in the two quadrants containing hydrocarbons.
Ants spent significantly more time in quadrant S+ than in quad-
rant N in 12 out of 24 cases (Wilcoxon test for matched pairs,
Z= 1.99–2.80, n= 10 for each pair of hydrocarbons, P< 0.05 in 12
cases, see Fig. 3). High preference for S+ means that the hydrocar-
bons are perceived as dissimilar and thus there is low generalisa-
tion between hydrocarbons.
The preference of the ants in the second choice test did not dif-
fer from the first one (Wilcoxon test for matched pairs, Z= 0.05–
1.78, n= 10, P> 0.05) in all cases except for one (n-C
28
/n-C
22
), in
which the ants had an even stronger preference for n-C
28
in the
second test than in the first.
By contrast, in the control tests, for each pair of hydrocarbons,
naïve ants spent an equal amount of time in both quadrants
(Wilcoxon test for matched pairs, Z= 0.36–1.69, n= 10 for each
0% 20% 40% 60% 80% 100%
*
n-C
19
/ n-C
21
n-C
20
/ 11-MeC
27
11-MeC
27
/n-C
20
11meC
27
/3meC
27
3meC
27
/11meC
27
3meC
31
/11meC
27
11meC
27
/3meC
31
3meC
31
/3meC
27
3meC
27
/3meC
31
S+
N
**
**
**
*
**
**
**
*
**
**
*
n-C
21
/ n-C
19
n-C
19
/ n-C
22
n-C
22
/ n-C
19
n-C
19
/ n-C
28
n-C
28
/ n-C
19
n-C
20
/ n-C
22
n-C
22
/ n-C
20
n-C
20
/ n-C
28
n-C
28
/ n-C
20
n-C
21
/ n-C
22
n-C
22
/ n-C
21
n-C
22
/ n-C
28
n-C
28
/ n-C
22
n-C
28
/ 3-MeC
27
3-MeC
27
/n-C
28
dz: 0.49
dz: 2.81
dz: 2.88
dz: 4.64
dz: 0.82
dz: 1.94
dz: 0.74
dz: 3.94
dz: 0.99
dz: 3.30
dz: 1.31
dz: 4.35
dz: 0.27
dz: 1.35
dz: 1.83
dz: 0.66
dz: 4.95
dz: 3.84
dz: 1.72
dz: 0.08
dz: 1.40
dz: 0.74
dz: 0.24
dz: 0.93
Fig. 3. Choice test. Relative amount of time spent in each quadrant during the first choice test for each pair of hydrocarbons. Dark gray bars denote S+ quadrant; light grey
bars denote N quadrant. The first hydrocarbon represents S+, while the second represents N in each pair (e.g. in the first line n-C
19
is S+ and n-C
21
is N). Stars denote
significance level
⁄
P< 0.05,
⁄⁄
P< 0.01. Effect sizes are given as Cohen’s dz.
N. Bos et al. / Journal of Insect Physiology xxx (2011) xxx–xxx 5
Please cite this article in press as: Bos, N., et al. Learning and perceptual similarity among cuticular hydrocarbons in ants. Journal of Insect Physiology
(2011), doi:10.1016/j.jinsphys.2011.10.010

pair of hydrocarbons, P> 0.05), suggesting that there is no sponta-
neous preference (or avoidance) for any specific hydrocarbon.
3.3. Asymmetry in generalisation
In some cases generalisation among stimuli depended on the
animal’s experience, i.e. ants conditioned to a stimulus A general-
ised to a stimulus B, whilst ant conditioned to stimulus B did not
generalise to stimulus A (cf. Guerrieri et al., 2005a,b). When hydro-
carbon pairs included only linear alkanes, the preference index (PI)
from hydrocarbon A to B was significantly different from the PI
from B to A in five out of seven cases (Fig. 4A, Mann–Whitney U
test, Z= 1.97–2.34, n= 10, P< 0.05). When tested hydrocarbon
pairs contained at least one methyl-alkane, none of the PI were sig-
nificantly different (Fig. 5A; Mann–Whitney U test, Z= 0.60–1.59,
n= 10, P> 0.05).
Among linear alkanes, there was a significant relationship be-
tween difference in carbon chain length (S
+
minus N) and PI
Preference Index
n-C
19
n-C
28
n-C
28
n-C
19
n-C
20
n-C
28
n-C
28
n-C
20
n-C
22
n-C
28
n-C
28
n-C
22
n-C
19
n-C
22
n-C
22
n-C
19
n-C
20
n-C
22
n-C
22
n-C
20
n-C
19
n-C
21
n-C
21
n-C
19
n-C
21
n-C
22
n-C
22
n-C
21
-0.6 -0.4 -0.2 0.0 0.2 0.4 0.6 0.8 1.0
-1.0 -0.8
*
*
** *
Z=2.34; P=0.02
Z=2.34; P=0.02
Z=1.66; P=0.09
Z=-2.19; P=0.03
Z=2.19; P=0.03
Z=-1.97; P=0.05
Z=1.74; P=0.08
A
-10-8-6-4-20246810
Difference in Carbon Chain Length (S
+
-0.1
0.0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
Preference Index
-10-8-6-4-20246810
Difference in Carbon Chain Length (S
+
minus N)
-0.1
0.0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
Preference Index
B
NS+
Fig. 4. Perceptual similarity of linear alkanes. (A) The preference index for quadrant S+ over N for each alkane pair tested. Values around 0 mean that there was no preference
between S
+
and N; positive values mean that there was a preference for S+. Exact P-values and Z-test values for independent samples are shown. (B) Among linear alkanes
there was a significant relationship between difference in carbon chain length (S+ minus N) and preference index (P< 0.05, Spearman’s correlation). Black bars and black dots
show the cases where S+ is the longer hydrocarbon of the pair tested; white bars and white dots show the cases where S+ is the shorter hydrocarbon.
6N. Bos et al. / Journal of Insect Physiology xxx (2011) xxx–xxx
Please cite this article in press as: Bos, N., et al. Learning and perceptual similarity among cuticular hydrocarbons in ants. Journal of Insect Physiology
(2011), doi:10.1016/j.jinsphys.2011.10.010

(Fig. 4B; Spearman’s correlation, r
s
=0.58, P< 0.05), whilst among
pairs containing at least one methyl-alkane no such relationship
was found (Fig. 5B; Spearman’s correlation, r
s
= 0.04, P> 0.05), sug-
gesting that they are perceived as similar.
4. Discussion
We developed an experimental protocol to study whether ants
can associate single hydrocarbons to another stimulus. The first
important result of our study is that individual C. aethiops workers
can associate long-chain hydrocarbons to a sugar reward, support-
ing the results of Bos et al. (2010), where ants learned to associate
a complete chemical profile to a sucrose reward. The fact that all
eight hydrocarbons (five linear alkanes and three methyl-alkanes)
were efficiently learnt by the ants suggests that both classes of sub-
stances can be detected and may have a potential role in chemical
communication and recognition. In ants, the evolution of similar
learning performances for hydrocarbons of different classes can be
understood in terms of the relative importance of chemical cues
for the identification of several categories of individuals. For exam-
ple, in C. vagus, cuticular hydrocarbons are thought to be important
to discriminate both nestmates from non-nestmates (Bonavita-Cou-
gourdan et al., 1987) and functional subcastes among nestmates
(Bonavita-Cougourdan et al., 1993). Similarly, in Camponotus fellah,
cuticular hydrocarbons are involved in nestmate recognition (e.g.
Boulay and Lenoir, 2001), while some hydrocarbons have a possible
role as fertility signals and are abundant on the queen’s cuticle and
on queen-laid eggs in C. floridanus (Endler et al., 2006).
The second important result of our study is that ants can dis-
criminate most of the hydrocarbons learnt, but their discrimina-
tion performance is dependent on the structure of the
hydrocarbon molecule and the previous experience of the individ-
ual. In general, for linear alkanes ants usually searched for sugar
close to the conditioned hydrocarbon (S+) when discrimination
was from lower to higher chain length molecules (e.g. when n-
C
19
was S+ and n-C
28
was N). In the reciprocal situation, ants had
no preference for either of two hydrocarbons presented in the
choice test, showing that they were generalising from higher to
lower chain length molecules. This suggests both that the signifi-
cance of the stimulus is dependent on the animal’s experience ac-
quired during conditioning and that odour detection and
perception might follow an inclusion criterion: if molecules pre-
3-MeC
31
11-MeC
27
11-MeC
27
3-MeC
31
Preference Index
11-MeC
27
n-C
20
n-C
20
11-MeC
27
n-C
28
3-MeC
27
3-MeC
27
n-C
28
11-MeC
27
3-MeC
27
3-MeC
27
11-MeC
27
3-MeC
31
3-MeC
27
3-MeC
27
3-MeC
31
-0.6 -0.4 -0.2 0 .0 0.2 0.4 0.6 0 .8 1.0-1.0 -0.8
Z=-0.91; P=0.36
Z=0.68; P=0.50
Z=-0.60; P=0.54
Z=-0.83; P=0.41
Z=1.58; P=0.11
A
B
-0.1
0.0
0.1
0.2
0.3
0.4
0.5
0.6
Difference in Carbon Chain Length (S
+
-
Preference Index
-8 -6 -4 -2 0 2 4 6 8-8 -6 -4 -2 0 2 4 6 8
-0.1
0.0
0.1
0.2
0.3
0.4
0.5
0.6
Difference in Carbon Chain Length (S
+
minus N)
Preference Index
NS+
Fig. 5. Perceptual similarity of methyl alkanes. (A) The preference index for quadrant S+ over N for each pair of hydrocarbons where at least one hydrocarbon is methylated.
Values around 0 mean that there was no preference between S+ and N; positive values mean that there was a preference for S+. Exact P-values and Z-test values for
independent samples are shown. (B) Among pairs containing at least one methyl-alkane, no relationship between difference in carbon chain length (S+ N) and preference
index was found (P> 0.05, Spearman’s correlation). Black bars and black dots show the cases where S+ is the longer hydrocarbon of the pair tested; white bars and white dots
show the cases where S+ is the shorter hydrocarbon.
N. Bos et al. / Journal of Insect Physiology xxx (2011) xxx–xxx 7
Please cite this article in press as: Bos, N., et al. Learning and perceptual similarity among cuticular hydrocarbons in ants. Journal of Insect Physiology
(2011), doi:10.1016/j.jinsphys.2011.10.010

sented in the tests are similar in structure to those presented as
conditioned stimuli and carry a shorter chain, these molecules
are perceived as similar to the conditioned stimulus (cf. Guerrieri
et al., 2009). Ants would then use an inclusion rule that produced
the observed asymmetry in generalisation. However, when the dif-
ference in chain length is too large (for example n-C
28
/n-C
20
) the
ants did not generalise. Asymmetry in generalisation has previ-
ously been found in honeybee (e.g. Guerrieri et al., 2005a,b; Sandoz
et al., 2001); for instance, generalisation responses of bees condi-
tioned to the alarm compound were much higher than those of
bees conditioned to the floral odour. Such asymmetry was further
linked to different biological significance of these compounds for
bees (Sandoz et al., 2001). Asymmetry in generalisation has been
hypothesised to be due to experience-dependent differences in sal-
iency of odours or innate odour preferences. However, it could also
be due to glomerular activation patterns and/or lateral inhibition
by local interneurons in the antennal lobes, the first integration
centres of olfactory stimuli in the insect brain (cf. Guerrieri et al.,
2005a).
Finally, in the present study, we found that ants that learnt to
associate food to a methyl-branched hydrocarbon (S+) also
searched for food around the novel methyl-branched hydrocarbon
(N), meaning there was high generalisation among this class of
molecules. Also, ants perceived hydrocarbons that differed among
each other by the presence or absence of one methyl group (i.e. lin-
ear versus methyl-alkanes) as dissimilar, especially when there
was also a difference in chain-length (e.g. when n-C
20
was S+ and
11-meC
27
was N, and vice versa), and thus expressed low general-
isation. This confirms the importance of carbon chain length as a
key feature for generalisation among hydrocarbons, and suggests
that chain length and functional group might be coded indepen-
dently by the ant olfactory system. Chain length of hydrocarbons
differs greatly on the cuticle of ants (e.g. van Zweden et al.,
2009), thus distinguishing between them might be important.
We start elucidating which molecular features are significant
for perceptual similarity in ants (see also van Wilgenburg et al.,
2010). The substances we assayed in the present study have long
carbon chains, low volatility and are typically present on the ants’
cuticle. However, cuticular hydrocarbon profiles are blends of
many molecules in different ratios. When the ratios of certain
hydrocarbons in the profile are highly correlated with each other,
they might be interpreted as a single variable (Martin et al.,
2008b), thus simplifying an apparent complex profile. Our results
suggest the possibility that ants might indeed perceive multiple
compounds as a single variable, through generalisation. General-
isation might also blur the line between different chemical profiles,
as differences present in the profile, apparent to us by use of GC–
MS, might not be perceived as different by the ant. To fully under-
stand how generalisation affects nestmate recognition, and possi-
bly further improve multivariate analysis of chemical data in this
species, a comprehensive study using hydrocarbons of all different
classes, present on the cuticle of the species is needed. Recent work
suggests that olfactory generalisation allows honeybees to adjust
their sensitivity to differences in concentrations of chemical stim-
uli according to the animal’s experience (Wright et al., 2008). Hon-
eybees generalised between two concentrations of the same
stimulus when both were paired with sucrose reward. Conversely,
generalisation was low when one concentration was paired with
sucrose reward and the other one was paired with salt solution
(Wright et al., 2008). Flowers, the main food sources for honeybees
are recognised by their perfumes. These perfumes are complex
odorant blends composed by similar (but not always equal) pro-
portions of volatile substances. Therefore, generalising among rel-
ative concentrations allows bees to precisely locate exploitable
food sources. Further studies should address whether ants are
capable of generalising among complex cuticular hydrocarbon
blends, since this might be of great importance in the context of
nestmate recognition.
Acknowledgements
We are grateful to Bernhard Seifert and Alain Lenoir for con-
firming the identity of ant specimens. Many thanks to Martin Giur-
fa for discussion on pheromone perception and Gösta Nachman for
help with statistical analysis. We thank all the members of the Cen-
tre for Social Evolution, University of Copenhagen, for a stimulating
working environment. This work was supported by the EU-Marie
Curie Excellence Grant, EXT-CT-2004-014202 to PdE. FJG was
supported as a post-doctoral fellow by the Danish National Re-
search Foundation.
Appendix A. Supplementary data
Supplementary data associated with this article can be found, in
the online version, at doi:10.1016/j.jinsphys.2011.10.010.
References
Akers, R.P., Getz, W.M., 1992. A test of identified response classes among olfactory
receptor neurons in the honey-bee worker. Chemical Senses 17, 191–209.
Akers, R.P., Getz, W.M., 1993. Response of olfactory receptor neurons in honeybees
to odorants and their binary mixtures. Journal of Comparative Physiology A 173,
169–185.
Akino, T., Yamamurea, N., Wakamuru, S., Yamaoka, R., 2004. Direct behavioral
evidence for hydrocarbons as nestmate recognition cues in Formica japonica
(Hymenoptera: Formicidae). Applied Entomology and Zoology 39, 381–387.
Balderrama, N., Núñez, J., Guerrieri, F.J., Giurfa, M., 2002. Different functions of two
alarm substances in the honeybee. Journal of Comparative Physiology A 168,
485–491.
Bonavita-Cougourdan, A., Clément, J.-L., Lange, C., 1987. Nestmate recognition: the
role of cuticular hydrocarbons in the ant Camponotus vagus Scop. Journal of
Entomology Science 22, 1–10.
Bonavita-Cougourdan, A., Clément, J.-L., Lange, C., 1993. Functional subcaste
discrimination (foragers and brood-tenders) in the ant Camponotus vagus
Scop: polymorphism of the cuticular hydrocarbon pattern. Journal of Chemical
Ecology 19, 1461–1477.
Bos, N., Guerrieri, F.J., d’Ettorre, P., 2010. Significance of chemical recognition cues is
context dependent in ants. Animal Behaviour 80, 839–844.
Boulay, R., Lenoir, A., 2001. Social isolation of mature workers affects nestmate
recognition in the ant Camponotus fellah. Behavioural Processes 55, 67–73.
Châline, N., Sandoz, J.-C., Martin, S.J., Ratnieks, F.L.W., Jones, G.R., 2005. Learning and
discrimination of individual cuticular hydrocarbons by honey bees (Apis
mellifera). Chemical Senses 30, 327–333.
Dani, F.R., Jones, G.R., Destri, S., Spencer, S.H., Turillazzi, S., 2001. Deciphering the
recognition signature within the cuticular chemical profile of paper wasps.
Animal Behaviour 62, 165–171.
Dani, F.R., Jones, G.R., Corsi, S., Beard, R., Pradella, D., Turillazzi, S., 2005. Nestmate
recognition cues in the honey bee: differential importance of cuticular alkanes
and alkenes. Chemical Senses 30, 477–489.
d’Ettorre, P., Heinze, J., Schulz, C., Francke, W., Ayasse, M., 2004. Does she smell like
a queen? Chemoreception of a cuticular hydrocarbon signal in the ant
Pachycondyla inversa. Journal of Experimental Biology 207, 1085–1091.
d’Ettorre, P., Lenoir, A., 2010. Nestmate recognition. In: Lach, L., Parr, C., Abbott, K.
(Eds.), Ant Ecology. Oxford University Press, pp. 194–209.
Dupuy, F., Sandoz, J.-C., Giurfa, M., Josens, R., 2006. Individual olfactory learning in
Camponotus ants. Animal Behaviour 72, 1081–1091.
Endler, A., Liebig, J., Hölldobler, B., 2006. Queen fertility, egg marking and colony
size in the ant Camponotus floridanus. Behavioral Ecology and Sociobiology 59,
490–499.
Frisch, K.Von., 1919. Über den Geruchsinn der Biene und seine blütenbiologische
Bedeutung. Zoologischer Jahrbücher (Physiologie) 37, 1–238.
Greene, M.J., Gordon, D.M., 2003. Cuticular hydrocarbons inform task decisions.
Nature 423, 32.
Greene, M.J., Gordon, D.M., 2007. Structural complexity of chemical recognition
cues affects the perception of group membership in the ants Linephithema
humile and Aphaenogaster cockerelli. Journal of Experimental Biology 210, 897–
905.
Guerrieri, F.J., Schubert, M., Sandoz, J.-C., Giurfa, M., 2005a. Perceptual and neural
olfactory similarity in honeybees. PLoS Biology 3, e60.
Guerrieri, F.J., Lachnit, H., Gerber, B., Giurfa, M., 2005b. Olfactory blocking and
odorant similarity in the honeybee. Learning and memory 12, 86–95.
Guerrieri, F.J., Nehring, V., Jørgensen, C.G., Nielsen, J., Galizia, C.G., d’Ettorre, P., 2009.
Ants recognize foes and not friends. Proceeding of Royal Society B 276, 2461–
2468.
8N. Bos et al. / Journal of Insect Physiology xxx (2011) xxx–xxx
Please cite this article in press as: Bos, N., et al. Learning and perceptual similarity among cuticular hydrocarbons in ants. Journal of Insect Physiology
(2011), doi:10.1016/j.jinsphys.2011.10.010

Guerrieri, F.J., d’Ettorre, P., 2010. Associative learning in ants: Conditioning of the
maxilla-labium extension response in Camponotus aethiops. Journal of Insect
Physiology 56, 88–92.
Hefetz, A., 2007. The evolution of hydrocarbon pheromone parsimony in ants
(Hymenoptera: Formicidae) – Interplay of colony odor uniformity and odor
idiosyncrasy – A review. Myrmecolical News 10, 59–68.
Holman, L., Dreier, S., d’Ettorre, P., 2010. Selfish strategies and honest signalling:
reproductive conflicts in ant queen associations. Proc. Roy. Soc. B 277, 2007–
2015.
Howard, R.W., Blomquist, G.J., 2005. Ecological, behavioral and biochemical aspects
of insect hydrocarbons. Annual Review of Entomology 50, 371–393.
Knudsen, J.T., Tollsten, L., Bergstrom, L.G., 1993. Floral scents – A checklist of
volatile compounds isolated by headspace techniques. Phytochemistry 33,
253–280.
Laska, M., Galizia, C.G., Giurfa, M., Menzel, R., 1999. Olfactory discrimination ability
and odor structure–activity relationships in honeybees. Chemical Senses 22,
457–465.
Lenoir, A., Fresneau, D., Errard, C., Hefetz, A., 1999. Individuality and colonial
identity in ants: the emergence of the social representation concept. In: Detrain,
C., Deneubourg, J.L., Pasteels, J. (Eds.), Information Processing in Social Insects.
Birkhauser, Basel, Switzerland, pp. 219–237.
Lucas, C., Pho, D.B., Jallon, J.M., Fresneau, D., 2005. Role of cuticular hydrocarbons in
the chemical recognition between ant species in the Pachycondyla villosa species
complex. Journal of Insect Physiology 51, 1148–1157.
Martin, S.J., Vitikainen, E., Helanterä, H., Drijfhout, F.P., 2008a. Chemical basis of
nestmate discrimination in the ant Formica exsecta. Proc. Roy. Soc. B 275, 1271–
1278.
Martin, S.J., Helanterä, H., Drijfhout, F.P., 2008b. Colony-specific hydrocarbons
identify nest mates in two species of Formica ant. Journal of Chemical Ecology
34, 1072–1080.
Martin, S.J., Drijfhout, F.P., 2009. A review of ant cuticular hydrocarbons. Journal of
Chemical Ecology 35, 1151–1161.
Menzel, R., Giurfa, M., 2006. Dimensions of cognition in an insect, the honeybee.
Behavioral and Cognitive Neuroscience Reviews 5, 24–40.
Meskali, M., Provost, E., Bonavita-Cougourdan, A., Clément, J.L., 1995. Behavioral
effects of an experimental change in the chemical signature of the ant
Camponotus vagus (Scop.). Insectes Sociaux 42, 347–358.
Ottoni, E.B., 2000. EthoLog 2.2: a tool for the transcription and timing of behavior
observation sessions. Behavior Research Methods, Instruments & Computers 32,
446–449.
Reinhard, J., Sinclair, M., Srinivasan, M.V., Claudianos, C., 2010. Honeybees learn
odour mixtures via a selection of key odorants. PLoS ONE 5, e9110.
Sandoz, J.C., Pham-Delègue, M.H., Renou, M., Wadhams, L.J., 2001. Asymmetrical
generalisation between pheromonal and floral odours in appetitive olfactory
conditioning of the honey bee (Apis mellifera L.). Journal of Comparative
Physiology A 187, 559–568.
Robinson, G.E., Heuser, L.M., Leconte, Y., Lenquette, F., Hollingworth, R.M., 1999.
Neurochemicals aid bee nestmate recognition. Nature 399, 534–535.
Van Wilgenburg, E., Sulc, R., Shea, K.J., Tsutsui, N.D., 2010. Deciphering the chemical
basis of nestmate recognition. Journal of Chemical Ecology 36, 751–758.
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, 159–164.
Van Zweden, J.S., d’Ettorre, P., 2010. Nestmate recognition in social insects and the
role of hydrocarbons. In: Blomquist, G.J., Bagnères, A.G. (Eds.), Insect
Hydrocarbons. Cambridge University Press, pp. 222–243.
Wehner, R., 2003. Desert ant navigation: how miniature brains solve complex tasks.
Journal of Comparative Physiology A 189, 579–588.
Wright, G.A., Kottcamp, S.M., Thomson, M.G.A., 2008. Generalization mediates
sensitivity to complex odor features in the honeybee. PLoS ONE 3 (2), e1704.
N. Bos et al. / Journal of Insect Physiology xxx (2011) xxx–xxx 9
Please cite this article in press as: Bos, N., et al. Learning and perceptual similarity among cuticular hydrocarbons in ants. Journal of Insect Physiology
(2011), doi:10.1016/j.jinsphys.2011.10.010