Content uploaded by Brian A. Whyte
All content in this area was uploaded by Brian A. Whyte on Dec 21, 2020
Content may be subject to copyright.
The Ant Who Cried Wolf? Short-Term Repeated
Exposure to Alarm Pheromone Reduces Behavioral
Response in Argentine Ants
Jessica J. Maccaro 1, *,†, Brian A. Whyte 2, †and Neil D. Tsutsui 2
1Department of Entomology, University of California, Riverside, CA 92521, USA
Department of Environmental Science, Policy, and Management, University of California, 130 Mulford Hall,
#3114, Berkeley, CA 94720, USA; email@example.com (B.A.W.); firstname.lastname@example.org (N.D.T.)
*Correspondence: email@example.com; Tel.: +1-(949)-939-1417
†These authors contributed equally to this study.
Received: 25 October 2020; Accepted: 3 December 2020; Published: 8 December 2020
A signiﬁcant challenge of chemical communication between ants is to maintain
accurate communication of information in a variety of contexts. Argentine ants use volatile (airborne)
compounds for a variety of functions, but one very important function is to elicit alarm via alarm
pheromones. Given the importance of accurately responding to this signal, we expected Argentine
ants to consistently show an alarm response to repeated exposure of alarm pheromones from their
nestmates. However, we instead observed a reduction in their alarm behaviors over time. We speculate
that a consistent response to repeated alarm signaling might require reinforcement from an actual
alarming stimulus (e.g., the presence of predators or rival colonies). Argentine ants are considered
a pest and several integrated pest management regimes use pheromones (i.e., mating disruption,
aggregation pheromones, etc.) to reduce pest populations. Our results could be important to consider
in the development of such control strategies because if ants habituate to their alarm pheromone over
continuous exposure (without actually alarming stimuli) it might prove to be an ineﬀective strategy
to repel them.
In this study we test whether Argentine ants (Linepithema humile) progressively reduce their
response to a salient stimulus (alarm pheromone) with increased exposure over time. First, we used
a two-chamber olfactometer to demonstrate three focal behaviors of Argentine ants that indicate
an alarmed state in response to conspeciﬁc alarm pheromone and pure synthetic iridomyrmecin
(a dominant component of L. humile alarm pheromone). We then measured how these behaviors
changed after repeated exposure to conspeciﬁc alarm pheromone from live ants. In addition,
we investigate whether there is a diﬀerence in the ants’ behavioral response after “short” (3 min)
versus “long” (6 min) intervals between treatments. Our results show that Argentine ants do exhibit
reduced responses to their own alarm pheromone, temporarily ceasing their response to it after four or
ﬁve exposures, and this pattern holds whether exposure is repeated after “short” or “long” intervals.
We suggest alarm pheromones may be perceived as false alarms unless threatening stimuli warrant a
continued state of alarm. These results should be kept in mind while developing pheromone-based
integrated pest management strategies.
Keywords: short-term; alarm pheromone; Linepithema humile; behavioral assay; Argentine ants
The societies of eusocial insects are typiﬁed by diverse and complex interactions among members
of the colony. These interactions are facilitated by sophisticated systems of chemical communication
Insects 2020,11, 871; doi:10.3390/insects11120871 www.mdpi.com/journal/insects
Insects 2020,11, 871 2 of 12
that can be useful for myriad purposes, including recruitment [
], colony recognition [
], mate attraction
and recognition [
], reproductive and caste recognition [
] task identiﬁcation [
], repellency [
brood care , or alarm [10,11].
A signiﬁcant challenge of a heavily chemosensory lifestyle is the maintenance of olfactory acuity
across a wide variety of contexts and circumstances. Desensitization to olfactory cues could play a
role in olfactory acuity and be explained by a variety of mechanisms including sensory adaptation,
habituation, or fatigue. Sensory adaptation is generally thought of as a peripheral nervous system
process, whereas habituation is attributed to central nervous system processes [
]. Olfactory fatigue is
considered a subcategory of severe sensory adaptation, where it becomes challenging and sometimes
impossible to smell an odor for a given time [
]. These mechanisms can be diﬃcult to disentangle
from behavioral data alone, and this paper does not aim to do so. Rather, this paper will reveal a
behavioral assay for testing insect response to olfactory cues over time, as well as demonstrate that
Argentine ants lessen their response to alarm pheromone with increased intermittent exposure.
The Argentine ant (Linepithema humile) is a globally widespread and ecologically damaging
invasive species [
]. As is true for ants generally, the coordination of their activities is primarily
regulated by chemical signals and the study of their chemical communication has helped elucidate
the mechanisms for their success (e.g., [
]). Since chemicals mediate insect behavior and
coordination, there is interest in developing integrated pest control strategies that exploit these chemical
signals to lower or repel pest populations (i.e., mating disruption [
] and mass trapping [
Our results might prove important to consider in the development of pheromone-based control
strategies for Argentine ants.
Here, we describe the behavioral responses of Argentine ant workers to airborne alarm pheromones,
test whether repeated exposure produces lessened olfactory acuity, and quantify behavioral responses
under two diﬀerent stimulation intervals. Since the alarm pheromone is an important signal of
danger, one might expect that there would be strong selective pressure to ensure that this chemical
is always perceived. Two volatile compounds (iridomyrmecin and dolichodial) that are found in
the Argentine ant pygidial gland have been implicated as pheromones for alarm and defense [
We quantify three distinct behaviors that characterize the Argentine ant state of alarm by using synthetic
] and live ant alarm pheromone. After verifying that these behavioral responses are
qualitatively similar, we use live ants’ pheromones (as they are more biologically relevant) in short
and long exposure interval tests to monitor ant response over time. Based on results from other
systems, we hypothesize that these behavioral responses will dampen with repeated exposure to alarm
pheromones, whether it be over short or long interstimulus intervals [21–24].
2. Materials and Methods
2.1. Ant Colony Collection
For all experiments, Linepithema humile were collected at Albany Bulb peninsula (
Albany, CA, USA
37.887329784–122.321498714) two days before each treatment. All ants were collected within a ten-meter
radius and kept in a plastic bucket with substrate from the ﬁeld. All collected ants were queenright.
They were given 10% sucrose and distilled water on cotton balls once daily. All collections occurred
between January and March of 2018.
To ensure that visual or mechanical cues did not inﬂuence the ants’ alarm behavior, we constructed
a two-chamber olfactometer (Figure 1). The receiver chamber (15
9 mm; Sistema
New Zealand) was connected to a vacuum valve with clear Poly Vinyl-Chloride (PVC) tubing (0.65 mm
0.9 mm O.D.
5.5 mm L) and the sender chamber (15
9 mm; Sistema
Plastic) was connected
to the receiver chamber via a 3-way forged brass ball valve with a lever handle (Nigo 180SS Series,
Nigo Industrial Co., Ltd., Chang Hua City, Taiwan), with the 3rd opening of the valve open to the
Insects 2020,11, 871 3 of 12
ambient air. This valve was held in place by a clamp so that when turned there would be no mechanical
movement that would alarm the ants. For every experiment, an air ﬂow meter was used to ensure the
air ﬂow rate in the sender chamber was always 2000 mL/min. The sender chamber was 380 cm
, so the
equivalent of one chamber volume could be pulled to the receiver chamber in about 11.4 s. Since our
measurement intervals were 30 s, this would guarantee that all air from the sender chamber would be
present in the receiver chamber.
Olfactometer schematic. (
) Air ﬂow during interstimulus points, when ambient air was being
introduced into the receiver chamber. (
) Air ﬂow during stimulus, when synthetic iridomyrmecin or
alarm pheromone from living ants was being introduced into the receiver chamber.
2.3. Deﬁning Alarm Behaviors
In preliminary experiments, we observed the behavior of ants in response to several types of
disturbance including mechanical (tapping vials that contained workers against a hard surface),
exposure to other alarmed ants, or simply breathing on them. In all cases, we noted three speciﬁc
behaviors that were associated with the alarm response, which correspond to previously determined
ant alarm behaviors [
]. We quantiﬁed these behaviors in the subsequent experiments (below)
to assess the alarm response of worker ants to stimuli. The ﬁrst focal behavior was grooming (GROM),
quantiﬁed as the number of grooming events within the designated measuring period. A grooming
event was deﬁned as any time an ant cleaned its antennae or legs, by placing them and/or drawing them
through their mandibles. The second focal behavior was antennation frequency (AF), characterized
by an ant touching any part of another ant with its antennae. The third focal behavior was antennal
raising (AR), in which an ant would raise its antenna up into the air “in an exploratory way”, sniﬀtheir
headspace, then lower them.
2.4. Testing Eﬀect of Alarm Pheromone
2.4.1. Single Exposure, Live Ant Pheromone
We ﬁrst tested how ants responded to a single stimulation of alarm pheromone from alarmed
nestmates. Twenty-ﬁve receiver ants were removed from the bucket with an aspirator and placed into
a Petri dish with Fluon on the sides and a moist cotton ball to prevent dehydration. This was replicated
for 20 separate Petri dishes (10 for the experimental treatment, and 10 for the control treatment).
To begin the experiment, one Petri dish with 25 ants was placed in the receiver chamber, the lid was
sealed, and airﬂow initiated (Figure 1). The video recording began after 5 min of acclimation in the
receiver chamber. During this time, 100 stimulus ants were aspirated from the bucket and kept in
the vial directly attached to the aspirator. Fifteen seconds before the stimulus interval, 100 stimulus
ants were agitated by striking the vial against the counter ten times, then immediately placed into a
Fluon-coated Petri dish inside the sender chamber. After closing the lid, the valve was switched to
send airﬂow from the sender chamber to the receiver chamber for a 30 s stimulus interval (Figure 1B).
The three alarm behaviors of the receiver chamber ants were recorded before the stimulus, during the
stimulus, and 3 min after the stimulus (Figure 2A). Afterward, all ants from the receiver chamber and
Insects 2020,11, 871 4 of 12
the sender chamber were discarded so that no ants would be re-used in further trials. These trials
were interspersed with ten controls treatments performed in a random order, determined by a random
number generator. The controls were performed in the same manner, but with an empty Fluon-coated
Petri dish in the sender chamber instead of 100 agitated ants.
Designs for recording behavioral responses. Bolded boxes represent the 30 s time intervals
when the three alarm behaviors were recorded, and the large “X” at the end of each line indicate
when the experiment was ended. On the y-axis, “stim” represents when air passes through the sender
chamber (Figure 1B), and “air” represents when ambient air is ﬂowing directly into the receiver chamber
(Figure 1A). (
) Single exposure to alarm pheromone from agitated ants or synthetic iridomyrmecin.
) Multiple exposures to live ant alarm pheromone separated by “short” interstimulus intervals of
3 min. (
) Multiple exposures to live ant alarm pheromone separated by “long” interstimulus intervals
of 6 min.
2.4.2. Single Exposure, Synthetic Pheromone
We used the same method as above for this iridomyrmecin treatment, but the sender chamber
contained synthetic iridomyrmecin instead of 100 irritated ants. Iridomyrmecin is one of the two
components that comprise the argentine ant pygidial gland products, which have been shown to
function as their alarm pheromone [
]. Synthetic iridomyrmecin in a solid state was received from
the USDA ARS Invasive Insect Biocontrol and Behavior Laboratory (IIBBL). This solid isolate was
dissolved in 95% ethanol at a concentration of 0.2 mg/mL, and 4 mL of this solution was pipetted into a
glass vial (equal to 100 ants; following Choe et al. 2012 [
], although Welzel et al. 2018 [
conﬂicting estimates). The ethanol was used as the solvent because it evaporates more quickly than the
iridomyrmecin under nitrogen air ﬂow, so that all that would be left in the vial was the iridomyrmecin.
Once the ethanol evaporated, a thin waxy ﬁlm of pure iridomyrmecin remained in the vial. The vial’s
mouth rested on the valve in the sender chamber. The vial, being much larger, did not seal the valve;
it was only placed this way to allow the air to be pulled out of the vial before pulling from other
air in the sender chamber. For the control, an empty vial was ﬁlled with 4 mL of 95% ethanol and
evaporated before placing it in the sender chamber in the same orientation. Data were collected in the
same method as above (Figure 2A).
2.5. Repeated Exposures to Alarm Pheromone
To monitor the ants’ behavioral responses to alarm pheromone with increased exposure intervals
we performed experiments in which 25 Argentine ant workers were stimulated four times for 30 s
each (Figure 2). Two diﬀerent interstimulus intervals (“short” and “long”) were chosen to test if the
duration of time between stimuli changes the ants’ responses. This was replicated 10 times for long
Insects 2020,11, 871 5 of 12
intervals as well as 10 times for short intervals. We hypothesized that longer interstimulus intervals
would allow for greater recovery from odorant stimulation, and thus produce more responsive ants
across an equivalent number of stimulation events.
2.5.1. Short Intervals
We performed a “short interval” experiment in which Argentine ant workers were provided only
a short rest period between exposures to alarm pheromone. For this experiment, the alarm stimulus
was introduced at 3 min intervals, and behavioral data were recorded before, during, and after the
stimulus (Figure 2B). During each interstimulus interval (Figure 2B, odd timepoints), 100 new ants
were collected, alarmed, and then placed into the sender chamber at each stimulus point (Figure 2B,
even timepoints). The ants in the receiver chamber remained the same throughout the experiment
to monitor change their behavioral responses, across repeated stimulation events. The control was
performed following the same design, but with an empty Fluon-coated Petri dish in the sender chamber.
2.5.2. Long Intervals
We next performed a “long interval” experiment in which Argentine ants were exposed to the same
number of alarm pheromone stimulations, but at 6 min intervals, twice as long as in the “short interval”
experiment. Although these intervals were still fairly short, we wished to test whether the longer
rest period provided enough time to recover following successive pheromone exposures, relative to
behaviors observed in the short-interval experiments. As before, behavioral data were recorded before,
during, and after the stimulus points (Figure 2C). Because the interstimulus intervals in this experiment
were longer than in the short-interval habituation experiment, interstimulus behavioral data were
recorded three minutes after each stimulus point (Figure 2C). The 100 new ants for each stimulus
interval were collected during the last three minutes of the six-minute interval to ensure that the
sender ants between both treatments were in the vial for the same amount of time. The control for this
test was performed following the same procedure, but with an empty Fluon-coated Petri dish in the
2.6. Data Collection and Analysis
The receiver chamber was ﬁlmed throughout the entire duration of each experiment using a
Canon EOS 6D camera with an Electro-Focus (EF) 100 mm macro lens. Each video was later viewed in
a VLC Media Player. Each of the three focal behaviors were quantiﬁed within 30 s intervals before,
during, and after each stimulus point. Within each interval of measurement, the number of behavioral
events were recorded, not the number of individuals performing the behavior. A summary of the data
used in Figures 3and 4can be found in the Supplemental Materials (Tables S1 and S2). The remaining
ﬁgures (Figures 1and 2) were created in Microsoft PowerPoint 2016.
To test if stimulus events elicited signiﬁcantly diﬀerent behaviors in the single exposure
experiments, we used a Welsh two-sample t-test (results reported in Figure 3). For the
repeated exposure experiments, we used a repeated measures analysis of variance (ANOVA)
testing the interaction of condition groups and timepoints and a between-trial eﬀect
Model =Behavior ~ Condition ×Timepoint +Error (Trial)
) (Table 1). The between-trial eﬀect tests
was used if the condition groups showed signiﬁcantly diﬀerent behaviors between trials, while the
within-trial eﬀect tests was used if the condition groups showed signiﬁcantly diﬀerent behavioral
changes over timepoints (Table 1).
Insects 2020,11, 871 6 of 12
Response to single exposure. Behavioral responses to alarm pheromone emitted from agitated
) or synthesized iridomyrmecin (
) in the sender chamber. Timepoint 2 (TP 2) is the stimulus
point, when the experimental subjects received air from the sender chamber. Results from a Welsh
two-sample t-test between the negative control (CON; ambient air) and stimulus (EXP; sender chamber
air) at timepoint 2 are shown above each graph (** p-value
*** p-value ≤0.0001
). Shaded ribbons
around each line show standard error around the mean. For both the live ant pheromone and synthetic
pheromone experiments, n=10 replicate groups, each with 25 ants.
Patterns of reducing alarm response. Behavioral responses to repeated exposure of alarm
pheromone over short (
) and long (
) intervals. Solid lines are linear models of the change in
each behavior over the experimental duration. Dashed lines represent mean behavior surrounded by a
shaded standard error ribbon. Only stimulus timepoints are included in these graphs. CON =ambient
air, EXP =sender chamber air. For both the short- and long-interval experiments, n=10 replicate
groups, each with 25 ants.
Insects 2020,11, 871 7 of 12
Repeated Measures ANOVA on repeated exposure data. Condition refers to the treatment
group (experiment versus control), while timepoint refers to the timepoints shown in Figure 4.
Df =degrees of freedom
, Sum.sq. =sum of squares, Mean.sq. =mean of squares, Pr|>F| = p-value.
ANOVA Formula =(Behavior ~ Condition ×Timepoint).
Behavior (Short Intervals) Df Sum.sq Mean.sq. F. Value Pr|>F|
Grooming Error: Trial Condition 1 505.00 505.00 13.81 0.00158
Residuals 18 658.40 36.60
Error: Within Timepoint 3 29.60 9.88 0.89 0.45
Condition: Timepoint 3 156.20 52.08 4.17 0.00541
Residuals 54 596.90 11.05
Antennation freq. Error: Trial Condition 1 1814.00 1814.50 14.82 0.00117
Residuals 18 2203.00 122.40
Error: Within Timepoint 3 1207.00 402.40 7.63 0.000244
Condition: Timepoint 3 1074.00 358.00 6.78 0.000578
Residuals 54 2850.00 52.80
Antennal raising Error: Trial Condition 1 112.80 112.81 10.55 0.00446
Residuals 18 192.40 10.69
Error: Within Timepoint 3 26.94 8.98 3.03 0.03732
Condition: Timepoint 3 52.54 17.51 5.90 0.00147
Residuals 54 160.27 2.97
Behavior (Long Intervals) Df Sum.sq. Mean.sq. F. Value Pr|>F|
Grooming Error: Trial Condition 1 304.20 304.20 10.82 0.00408
Residuals 18 506.30 28.13
Error: Within Timepoint 3 49.10 16.35 1.15 0.337
Condition: Timepoint 3 410.10 136.70 9.63
Residuals 54 766.40 14.19
Antennation freq. Error: Trial Condition 1 1720.50 1720.50 35.41 1.2 ×10−5
Residuals 18 874.60 48.60
Error: Within Timepoint 3 1192.00 397.20 12.81
Condition: Timepoint 3 1302.00 434.10 14
Residuals 54 1675.00 31.00
Antennal raising Error: Trial Condition 1 270.10 270.11 9.44 0.00657
Residuals 18 515.30 28.63
Error: Within Timepoint 3 82.20 27.41 3.82 0.0149
Condition: Timepoint 3 88.50 29.51 4.11 0.0107
Residuals 54 388.00 7.19
3.1. Single Exposure
When exposed to a single 30 s introduction of alarm pheromone from 100 sender ants, there was a
signiﬁcant increase in the number of receiver ants antennating each other (Figure 3C, black triangles)
and raising their antennae (Figure 3D, black triangles) compared to their control groups at the same
timepoint. At the same time, grooming appeared to be interrupted, as signiﬁcantly fewer receiver ants
performed this behavior after the stimulation (Figure 3B, black triangles). All behaviors at the ﬁnal
timepoint (3 min later) were not substantially diﬀerent from the behaviors at the ﬁrst timepoint (3 min
before exposure), indicating a relaxation to a pre-disturbance behavioral state in this amount of time.
When receiver ants were exposed to the synthetic alarm pheromone component, iridomyrmecin,
a single 30 s stimulus produced qualitatively similar behavioral responses to those seen after exposure
to living agitated ants, but the magnitude of the behavioral changes were larger (Figure 3D–F).
All behaviors were signiﬁcantly diﬀerent from those displayed in the ambient air control groups.
The frequency of ants grooming decreased during stimulus (Figure 3F, black triangle), whereas the
frequency of ants antennating each other (AF), and raising their antennae (AR) increased during stimulus
(black triangle lines in Figure 3D–F, respectively). Similar to the live ant stimulus, the magnitude of all
behaviors fell back to pre-stimulus levels 3 min after the stimulus.
While all ants used in these experiments were colony mates collected from the same ten-meter
radius ﬁeld site, there appear to be innate behavioral diﬀerences between these collected groups,
even when sampled under the same circumstances. For instance, grooming behavior for control groups
Insects 2020,11, 871 8 of 12
in Figure 3A,D were diﬀerent, as well as antennation frequency in the single exposure (Figure 3B,
timepoint 2) and repeated exposure groups (Figure 4B,E, timepoint 2). However, within experiment,
the behavioral diﬀerences between control and experiment were consistent, supporting the
3.2. Short Intervals
In this experiment, the ants experienced four successive exposures to alarm pheromone,
each interspersed by a 3 min rest period (ambient air ﬂow, Figure 1). At the ﬁrst stimulus, the behavioral
response of ants to alarm pheromones from 100 sender ants was substantially diﬀerent from their
corresponding ambient air control groups (Figure 4A–C). However, the magnitude of the behavioral
response diﬀerence between control and experiment groups decreased across successive exposures
to alarm pheromone, approaching the levels seen in the ambient air control groups (Figure 4A–C).
Speciﬁcally, AF and AR decreased across successive stimulation events (Figure 4B,C), whereas increased
(Figure 4A). When compared using a Repeated Measures ANOVA, all behaviors showed signiﬁcant
interaction between condition groups over timepoints (Table 1), meaning the condition groups were
signiﬁcantly diﬀerent in how their behaviors changed over time.
Grooming (GROM) started oﬀless frequent in the experimental groups, but increased in frequency
after repeated exposure, projected to occur at the same frequency as the control group after ﬁve
stimulus points (Figure 4A, Table 2). The experiment only incorporated four stimulus points, but the
linear models of experiment and control groups predicted the grooming behavior between the two
groups would become similar if a ﬁfth stimulus point was introduced. As for antennation (AF, AR)
behaviors, these started oﬀat high frequencies, decreasing to levels similar to their control groups in
ﬁve or fewer stimulus points (Figure 4B,C, Table 2).
Points at which alarmed ant behavior returns to resting ant behavior. “Time pt. int.” refers to the
xaxis value where Control and Experiment linear models (shown on Figure 4) intercept. “
Stim pt. int
refers to the nearest stim point that the linear model intercepts round up to.
Behavior Interval Condition Slope Time pt. int. Stim pt. int.
Grooming Short Control −0.31 10.05 5
Short Experiment 0.68
Long Control −1.20 7.15 4
Long Experiment 0.73
Antennation freq. Short Control −0.06 8.58 5
Short Experiment −2.71
Long Control 0.16 7.63 4
Long Experiment −3.35
Antennal raising Short Control 0.24 9.92 5
Short Experiment −0.36
Long Control 0.67 9.68 5
Long Experiment −0.12
3.3. Long Intervals
The response of ants to stimulation in the experiment with longer intervals between stimuli (6 min)
was similar overall to the results seen in the short-interval experiment. Speciﬁcally, the condition
groups were signiﬁcantly diﬀerent in how their behaviors changed over time (Table 1), always starting
at substantially diﬀerent levels and then, after successive stimulations, approaching an estimated
intercept around 4–5 stimulus points (Figure 4, Table 2). The frequency of grooming (GROM) behavior
increased with repeated stimulus, whereas antennation (AF, AR) behaviors decreased (Figure 4).
However, the sharp increase in antennal raising at the last stimulus point (Figure 4F) means we cannot
be conﬁdent that the antennal raising behavior would continue to decrease over time as the linear
Insects 2020,11, 871 9 of 12
Diﬀerent from the short-interval results though, experiment and control linear models in the longer
interval test usually intercepted after four stimulus points, with only antennal raising intercepting
closer to ﬁve stimulus points (Table 2), likely due to the sharp increase in the last stimulus point
(Figure 4F). Overall, the pattern of reducing behavioral responses to alarm stimulus over time is the
same for both the short- and long-interval experiments.
Outside of an experimental setting, a reduced response to alarm pheromone is likely inﬂuenced by
many factors. There could be diﬀerent mechanisms enabling a colony to remain aware of iridomyrmecin
even after sustained exposure. Following the research of “response thresholds” in ants [
], there could
be asymmetry in the sensitivity of ants collectively perceiving this chemical. Some ants may stop
responding to alarm while others remain responsive. The ants could also only need a single induction
of alarm response to alarm pheromone in order to appropriately react to alarm stimuli, and additional
stimulation may not be useful, or could even be costly. Studies of the aphid Myzus periscope [
shown lower ﬁtness after habituation to their alarm pheromone in the context of predators, but if
predation was removed, habituation led to ﬁtness increases. Lastly, it has been noted that D. melanogaster
can dishabituate with mechanical stimulation [
] and their larva can dishabituate by being exposed to
noxious fumes [
]. If habituation underlies the behavioral changes we observed in this experiment,
Argentine ants may resist or reverse habituation to alarm when other stimuli of a threat to their nest are
also present (i.e., exposure to light, rival colonies, predators, etc.). Further experimentation is needed
to determine if our results demonstrate habituation, sensory deprivation, or fatigue.
When alarm pheromone is presented without an actual threat, as it was in the present study,
the exposed workers may not continue responding to the signal (i.e., “the ant who cried wolf”),
nor amplify it by producing additional alarm pheromone in response. Following this line of reasoning,
it is not so surprising that ants respond less to such an important chemical signal, because if there
is no context of danger to follow the signal, the pheromone is essentially a false alarm. Context or
concentration could be necessary for alarm pheromone to mean “alarm” and not something else.
However, we did not test how diﬀerent concentrations of the alarm pheromone inﬂuenced alarm
response, so we cannot deduce the interactions of quantity and context in determining the Argentine
ant response to alarm pheromone. We can only suggest that context could prove to be necessary for
this alarm pheromone to consistently induce an alarm and future research should aim to address this.
The results we report here show that Argentine ants decrease their response towards pheromones
that are vital to their survival after repeated short-term exposure. In this study we solely demonstrate
that the ants lessen their behavioral response to alarm pheromone over repeated exposure, but we do
not directly address the mechanism which could be due to sensory adaptation, habituation, or fatigue.
Regardless of alarm pheromone exposure, we can expect ants to show changing behavioral responses
after being transported to a new environment (hence, the non-zero slopes we see in the control
groups of Figure 4). However, these control groups show opposing, not similar, behavioral changes
to the experimental groups, suggesting a diﬀerent mechanism to explain these behavior trends over
time. Future research would greatly complement this study by using electroantennography (EAG) to
disentangle if ants reduce their alarm response due to habituation versus sensory adaptation to alarm
pheromone. Additionally, future studies could consider how pairing an actual threat (e.g., intra- or
interspeciﬁc competitors, predators, or parasitoids) with alarm pheromone aﬀects behavioral responses.
The evolutionary success of social insects generally, and the invasive success of the Argentine ant
in particular, stem in part from their chemical communication abilities. Consequently, our results could
be important for the development of pheromone-based control methods in integrated pest management
(IPM). Many pheromone-based baits are used for mating disruption or attract and kill techniques,
using sex pheromone and aggregation pheromone, respectively [
]. However, research into ant
pheromone baits, using trail pheromone to attract ants, has already been developed [
]. The trail
pheromone has been shown eﬀective when paired with an insecticide to attract and kill ants [
Insects 2020,11, 871 10 of 12
Trail pheromones are considered useful for short range communication and so alarm pheromone, as a
method for long range communication, may overcome some of the current challenges to disrupting
]. This has motivated researchers to develop alarm pheromone baits that can act at a distance
to attract ants [
]. This may sound counter-intuitive that ants would be attracted to alarm signals,
but alarm pheromones are often used by ants to initiate coordinated group defense behaviors [
Therefore, this appears to be a very promising avenue to manage ants in an agricultural setting [
However, if alarm pheromone (or any pheromone) is to be considered in the future to control ant
populations, it will be important to take into account that continuous exposure might weaken the
ant’s response and thus the eﬃcacy of the bait. This weakened response has been shown in our study,
as well as in several other insect species including other hymenopterans [21–24].
In this study, we hypothesized that Argentine ant alarm behaviors would remain consistent even
after repeated exposure to their alarm pheromone. However, we found that Argentine ants reduce
their response to alarm pheromone after repeated exposure in a matter of minutes. All three recorded
behaviors (GROM, AF, AR) displayed by worker ants approached the control values after repeated
exposure in both short-interval and long-interval tests. These results are qualitatively very similar to
observations from Drosophila melanogaster. In response to ethanol vapor, for example, D. melanogaster
displays a stereotypical olfactory “startle” response, but it is signiﬁcantly reduced after four repeated
pulses with interstimulus periods ranging from 3–18 min [
]. Similarly, the D. melanogaster olfactory
jump response that occurs after 4 s exposure to benzaldehyde is reduced after 2–15 pulses separated by
0.25–20 min intervals .
Additionally, our results conﬁrm that a behaviorally similar alarm response can be induced
by either introducing air from a chamber of alarmed ants or using synthetic iridomyrmecin alone.
These ﬁndings support the recent proposal that Argentine ant pygidial products (iridomyrmecin
and dolichodial) function as the Argentine ant alarm pheromone [
]. In fact, the alarm response to
synthetic iridomyrmecin appears even stronger than the live ant release of alarm pheromone (Figure 3).
This could be due to the fact that the amount used in this experiment was equal to 100 ant-equivalents
of iridomyrmecin [
], not necessarily the amount they release when alarmed. Therefore, we did
not attempt to make quantitative comparisons between response to synthetic iridomyrmecin and
live ant alarm pheromone, but rather show qualitatively that the behavioral response is similar.
Personal observations using solid phase microextraction (SPME) as well as other reports [
that iridomyrmecin is certainly present in the headspace of alarmed ants and synthetic iridomyrmecin
solutions. However, while we can indirectly estimate how much iridomyrmecin was available in the
sender chamber during stimulus points, we did not record the quantity in the receiver chamber head
space during each experiment.
The following are available online at http://www.mdpi.com/2075-4450/11/12/871/s1,
Table S1: Summary of stimulus point data used in Figure 3, Table S2: Summary of data used in Figure 4.
Conceptualization, N.D.T.; Data curation, J.J.M.; Formal analysis, B.A.W.;
Funding acquisition, N.D.T.; Methodology, J.J.M.; Supervision, N.D.T.; Visualization, B.A.W.; Writing—original
draft, J.J.M.; Writing—review and editing, B.A.W. and N.D.T. All authors have read and agreed to the published
version of the manuscript.
This work was supported by the US National Science Foundation (IOS-1557934/1557961), USDA National
Institute of Food and Agriculture (2016-67013-24749), and USDA Hatch Project (CA-B-INS-0087-H).
Conﬂicts of Interest:
The funders had no role in the design of the study; in the collection, analyses, or interpretation
of data; in the writing of the manuscript, or in the decision to publish the results.
Insects 2020,11, 871 11 of 12
Czaczkes, T.J.; Grüter, C.; Ratnieks, F.L.W. Trail Pheromones: An Integrative View of Their Role in Social
Insect Colony Organization. Annu. Rev. Entomol. 2015,60, 581–599. [CrossRef]
Brandt, M.; van Wilgenburg, E.; Sulc, R.; Shea, K.J.; Tsutsui, N.D. The scent of supercolonies: The discovery,
synthesis and behavioural veriﬁcation of ant colony recognition cues. BMC Biol. 2009,7, 71. [CrossRef]
Greenberg, L.; Johnson, C.A.; Trager, J.C.; Mcelfresh, J.S.; Rodstein, J.; Millar, J.G. Sex Attractant Pheromones
of Virgin Queens of Sympatric Slave-Making Ant Species in the Genus Polyergus, and their Possible Roles in
Reproductive Isolation. J. Chem. Ecol. 2018,44, 547–555. [CrossRef]
Singer, T.L. Roles of hydrocarbons in the recognition systems of insects. Am. Zool.
Keller, L.; Nonacs, P. The role of queen pheromones in social insects: Queen control or queen signal?
Anim. Behav. 1993,45, 787–794. [CrossRef]
Leonhardt, S.D.; Menzel, F.; Nehring, V.; Schmitt, T. Ecology and Evolution of Communication in Social
Insects. Cell 2016,164, 1277–1287. [CrossRef] [PubMed]
Greene, M.J.; Gordon, D.M. Cuticular hydrocarbons inform task decisions. Nature
,423, 32. [CrossRef]
Stout, J.C.; Goulson, D.; Allen, J.A. Repellent scent-marking of ﬂowers by a guild of foraging bumblebees
(Bombus spp.). Behav. Ecol. Sociobiol. 1998,43, 317–326. [CrossRef]
Den Boer, S.P.A.; Duchateau, M.J.H.M. A larval hunger signal in the bumblebee Bombus terrestris.Insectes Soc.
2006,53, 369–373. [CrossRef]
Welzel, K.F.; Lee, S.H.; Dossey, A.T.; Chauhan, K.R.; Choe, D. Veriﬁcation of Argentine ant defensive
compounds and their behavioral eﬀects on heterospeciﬁc competitors and conspeciﬁc nestmates. Sci. Rep.
2018, 1–15. [CrossRef] [PubMed]
Hölldobler, B.; Wilson, E.O. The Ants; Belknap Press of Harvard University Press: Cambridge, MA, USA,
1990; ISBN 9780674040755.
Dalton, P. Psychophysical and Behavioral Characteristics of Olfactory Adaptation. Chem. Senses
487–492. [CrossRef] [PubMed]
Moncrieﬀ, R.W. Olfactory adaptation and odour likeness. J. Physiol.
,133, 301–316. [CrossRef] [PubMed]
Holway, D.A.; Lach, L.; Suarez, A.V.; Tsutsui, N.D.; Case, T.J. The causes and consequences of ant invasions.
Annu. Rev. Ecol. Syst. 2002,33, 181–233. [CrossRef]
Torres, C.W.; Brandt, M.; Tsutsui, N.D. The role of cuticular hydrocarbons as chemical cues for nestmate
recognition in the invasive Argentine ant (Linepithema humile). Insectes Soc. 2007,54, 363–373. [CrossRef]
Choe, D.H.; Villafuerte, D.B.; Tsutsui, N.D. Trail Pheromone of the Argentine Ant, Linepithema humile (Mayr)
(Hymenoptera: Formicidae). PLoS ONE 2012,7, e45016. [CrossRef]
17. Minks, A.K. Mating Disruption of the Codling Moth. Insect Pheromone Res. 1997, 372–376. [CrossRef]
Baker, T.C. Chemical control of behaviour. In Comprehensive Insect Physiology, Biochemistry, and Pharmacology;
Kerkut, G.A., Gilbert, L.S., Eds.; Pergamon Press: Oxford, UK, 1985; pp. 621–672.
Maruthadurai, R.; Ramesh, R. Mass trapping of red palm weevil and rhinoceros beetle in coconut with
aggregation pheromone. Indian J. Entomol. 2020,82, 439–443. [CrossRef]
Kirk, W. The aggregation pheromones of thrips (Thysanoptera) and their potential for pest management.
Int. J. Trop. Insect Sci. 2017,37, 41–49. [CrossRef]
Twick, I.; Lee, J.A.; Ramaswami, M. Olfactory Habituation in Drosophila—Odor Encoding and its Plasticity
in the Antennal Lobe. Prog. Brain Res. 2014,208, 3–38. [CrossRef]
Larkin, A.; Karak, S.; Priya, R.; Das, A.; Ayyub, C.; Ito, K.; Rodrigues, V.; Ramaswami, M. Central synaptic
mechanisms underlie short-term olfactory habituation in Drosophila larvae. Learn. Mem.
De Vos, M.; Cheng, W.Y.; Summers, H.E.; Raguso, R.A.; Jander, G. Alarm pheromone habituation in Myzus
persicae has ﬁtness consequences and causes extensive gene expression changes. Proc. Natl. Acad. Sci. USA
2010,107, 14673–14678. [CrossRef] [PubMed]
Jernigan, C.M.; Birgiolas, J.; McHugh, C.; Roubik, D.; Wcislo, W.; Smith, B. Colony-level non-associative
plasticity of alarm responses in the stingless honey bee Tetragonisca angustula.Behav. Ecol. Sociobiol.
Insects 2020,11, 871 12 of 12
Regnier, F.E.; Wilson, E.O. The alarm-defence system of the ant Acanthomyops claviger.J. Insect Physiol.
14, 955–970. [CrossRef]
Löfqvist, J. Formic acid and saturated hydrocarbons as alarm pheromones for the ant Formica rufa.
J. Insect Physiol. 1976,22, 1331–1346. [CrossRef]
Theraulaz, G.; Bonabeau, E.; Deneubourg, J.L. Response threshold reinforcement and division of labour in
insect societies. Proc. R. Soc. B Biol. Sci. 1998,265, 327–332. [CrossRef]
Asztalos, Z.; Arora, N.; Tully, T. Olfactory jump reﬂex habituation in Drosophila and eﬀects of classical
conditioning mutations. J. Neurogenet. 2007,21, 1–18. [CrossRef]
Sunamura, E.; Suzuki, S.; Nishisue, K.; Sakamoto, H.; Otsuka, M.; Utsumi, Y.; Mochizuki, F.; Fukumoto, T.;
Ishikawa, Y.; Terayama, M.; et al. Combined use of a synthetic trail pheromone and insecticidal bait provides
eﬀective control of an invasive ant. Pest. Manag. Sci. 2011,67, 1230–1236. [CrossRef]
Tanaka, Y.; Nishishue, K.; Sunamura, E.; Suzuki, S.; Sakamoto, H.; Fukumoto, T.; Terayama, M.; Tatsuki, S.
Trail-following disruption in the invasive Argentine ant with a synthetic trail pheromone component
(Z)-9-hexadecenal. Sociobiology 2009,54, 139–152.
Tatsuki, S.; Terayama, M.; Tanaka, Y.; Fukumoto, T. Behavior-Disrupting Agent and Behavior Disrupting
Method of Argentine Ant. U.S. Patent 8,278,360, 2 October 2012.
Suckling, D.M.; Stringer, L.D.; Corn, J.E. Argentine Ant Trail Pheromone Disruption is Mediated by Trail
Concentration. J. Chem. Ecol. 2011,37, 1143. [CrossRef]
Nishishue, K.; Sunamura, E.; Tanaka, Y.; Sakamoto, H.; Suzuki, S.; Fukumoto, T.; Terayama, M.; Tatsuki, S.
Long-term ﬁeld trial to control the invasive argentine ant (Hymenoptera: Formicidae) with synthetic trail
pheromone. J. Econ. Entomol. 2010,103, 1784–1789. [CrossRef]
Hamzah, H.; Rixson, D.; Paul-Taylor, J.; Doan, H.; Dadswell, C.; Roﬀe, G.; Sridhar, A.; Hobday, C.; Wedd, C.;
Duren, T.; et al. Inclusion and release of ant alarm pheromones from metal–organic frameworks.
2020,49, 10334–10338. [CrossRef] [PubMed]
Hughes, W.; Goulson, D. The use of alarm pheromones to enhance bait harvest by grass-cutting ants.
Bull. Entomol. Res. 2002,92, 213–218. [CrossRef] [PubMed]
Hughes, W.; Howse, P.; Vilela, E.; Knapp, J.; Goulson, D. Field Evaluation of Potential of Alarm Pheromone
Compounds to Enhance Baits for Control of Grass-Cutting Ants (Hymenoptera: Formicidae).
J. Econ. Entomol.
2002,95, 537–543. [CrossRef] [PubMed]
Hughes, W.O.H.; Howse, P.E.; Vilela, E.F.; Goulson, D. The response of grass-cutting ants to natural and
synthetic versions of their alarm pheromone. Physiol. Entomol. 2001,26, 165–172. [CrossRef]
MDPI stays neutral with regard to jurisdictional claims in published maps and institutional
2020 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access
article distributed under the terms and conditions of the Creative Commons Attribution
(CC BY) license (http://creativecommons.org/licenses/by/4.0/).