Ethanol Self-Administration in Free-Flying Honeybees (Apis mellifera L.) in an Operant Conditioning Protocol.

Abstract

This study examines the effect of ethanol (EtOH) on continuous reinforcement schedules in the free-flying honeybee (Apis mellifera L.). As fermented nectars may be encountered naturally in the environment, we designed an experiment combining the tools of laboratory research with minimal disturbance to the natural life of honeybees.
Twenty-five honeybees were trained to fly from their colonies to a fully automated operant chamber with head poking as the operant response. Load size, intervisit interval, and interresponse times (IRTs) served as the dependent variables and were monitored over the course of a daily training session consisting of many visits. Experimental bees were tested using an ABA design in which sucrose only was administered during condition A and a 5% EtOH sucrose solution was administered during condition B. Control bees received sucrose solution only.
Most bees continued to forage after EtOH introduction. EtOH significantly reduced the load size and the intervisit interval with no significant effect on IRTs. However, a look on individual data shows large individual differences suggesting the existence of different kinds of behavioral phenotypes linked to EtOH consumption and effects.
Our results contribute to the study of EtOH consumption as a normal phenomenon in an ecological context and open the door to schedule-controlled drug self-administration studies in honeybees.

Full text

Page 1

Ethanol Self-Administration in Free-Flying Honeybees
(Apis mellifera L.) in an Operant Conditioning Protocol
Michel B. C. Sokolowski, Charles I. Abramson, and David Philip Arthur Craig
Background: This study examines the effect of ethanol (EtOH) on continuous reinforcement sched-
ules in the free-flying honeybee (Apis mellifera L.). As fermented nectars may be encountered naturally
in the environment, we designed an experiment combining the tools of laboratory research with mini-
mal disturbance to the natural life of honeybees.
Methods: Twenty-five honeybees were trained to fly from their colonies to a fully automated oper-
ant chamber with head poking as the operant response. Load size, intervisit interval, and interresponse
times (IRTs) served as the dependent variables and were monitored over the course of a daily training
session consisting of many visits. Experimental bees were tested using an ABA design in which sucrose
only was administered during condition A and a 5% EtOH sucrose solution was administered during
condition B. Control bees received sucrose solution only.
Results: Most bees continued to forage after EtOH introduction. EtOH significantly reduced the
load size and the intervisit interval with no significant effect on IRTs. However, a look on individual
data shows large individual differences suggesting the existence of different kinds of behavioral pheno-
types linked to EtOH consumption and effects.
Conclusions: Our results contribute to the study of EtOH consumption as a normal phenomenon in
an ecological context and open the door to schedule-controlled drug self-administration studies in hon-
eybees.
Key Words: Ethanol,Foraging,Continuous
Administration.
ReinforcementSchedule,Honeybees,Self-
F
be consumed regularly (Ehlers and Olesen, 1997; Jakubska
et al., 2005). The fact that EtOH consumption may be based
on some heritable variation (Guarnieri and Heberlein, 2002),
and possibly be linked to fitness (Etges and Klassen, 1989;
Geer et al., 1991), provides an opportunity for natural selec-
tion to shape traits related to EtOH consumption (Dudley,
2002; Sullivan and Hagen, 2002; Wiens et al., 2008). Conse-
quently, genes involved in EtOH metabolism or tolerance
may be the result of a long evolution beginning from differ-
ent taxa (Dudley, 2004).
Fruit-eating animals like birds and mammals consume
EtOH from naturally fermenting fruits (Dudley, 2002). How-
ever, EtOH can also be found in fermenting floral nectar.
OR SOME ANIMAL species, ethanol (EtOH) is
readily available in the natural environment and so may
Fermented floral nectar has only recently been taken into
consideration as a source of EtOH (Ehlers and Olesen, 1997;
Wiens et al., 2008).
EtOH in plants can be produced by nectar that is
contaminated by microorganisms (Kevan et al., 1988).
However, there are few studies that have investigated how
plants manufacture EtOH. Yeast and fungi can produce up
to 0.6% EtOH (Ehlers and Olesen, 1997), and concentrations
of EtOH produced by yeast have been measured in the
bertam palm of up to 3.8% (Wiens et al., 2008).
Some birds and small mammals consume floral nectar
(Mazeh et al., 2008; Sanchez et al., 2008), but pollinator
insects such as honeybees are probably the species most
exposed to fermenting nectar. If linked to fitness, we would
expect that animals actively search and consume alcoholic
nectar through a process of reinforcement and operant
conditioning.
Natural flowers generally produce only low doses of
EtOH. Consequently, if there are benefits of consuming low
doses of EtOH, there may be negative effects of consuming
large doses of EtOH. Previous research on the honeybee
model indicates that EtOH reduces the frequency of locomo-
tion and impairs Pavlovian conditioning in harnessed bees as
a function of concentration (Abramson et al., 2000; Bozic
et al., 2006; Maze et al., 2006). In addition to disruptions in
learned behavior and activity, drinking EtOH increases
aggression in free-flying honeybees as measured by stings on
a leather patch (Abramson et al., 2004a,b). Honeybees
From the Jules Verne (MBCS), INSERM 24 (ERI24), Groupe de
Recherche sur l’Alcool et les Pharmacode´pendances, Universite´ de Picar-
die, Amiens Cedex 1, France; and Laboratory of Comparative Psychology
and Behavioral Biology (CIA, DPAC), Departments of Psychology and
Zoology, Oklahoma State University, Stillwater, Oklahoma.
Received for publication October 6, 2011; accepted January 9, 2012.
Reprint requests: Michel Sokolowski, Jules Verne, Department de
Psychologie, Universite´ de Picardie, Chemin du Thil, 80025 Amiens
Cedex 1, France; Tel.: 33-(0)-322-82-89-11; Fax: 33-(0)-322-82-74-08;
E-mail: michel.sokolowski@u-picardie.fr
Copyright © 2012 by the Research Society on Alcoholism.
DOI: 10.1111/j.1530-0277.2012.01770.x
1568Alcohol Clin Exp Res, Vol 36, No 9, 2012: pp 1568–1577
ALCOHOLISM: CLINICAL AND EXPERIMENTAL RESEARCH
Vol. 36, No. 9
September 2012

Page 2

influenced by EtOH fail to recognize caloric nectar rewards
under conditions in which unimpaired foragers readily do so
(Abramson et al., 2004a,b, 2005). Honeybee social behavior,
such as the waggle dance used to communicate among
foragers, is disrupted as is hygienic behavior and behavior of
the queen (Bozic et al., 2006; Cakmak et al., 2009).
Recent experiments show that the metabolism of EtOH in
honeybees shares temporal and dose-dependent responses in
common with vertebrate models (Bozic et al., 2007). Blood
EtOH concentrations rise immediately following consump-
tion, peak after 30 minutes, and remain elevated for more
than 4 hours postconsumption. Moreover, levels of HSP70
(heat shock proteins) in bees feed various EtOH solutions
showed an inverted U-shape curve of HSP70 concentration,
indicating that ingesting 2.5% EtOH and 5% EtOH stimu-
lated the stress response whereas ingestion of 10% EtOH
inhibited the stress response (Hranitz et al., 2010).
The purpose of this study is to extend our understanding
of the effect of EtOH on the social insect model of EtOH-
induced behavior by examining the effect of EtOH on sche-
dule-controlled behavior of honeybees. The rationale behind
these experiments is 4-fold. First, we wanted to determine
whether it is feasible to study controlled behavior in free-
flying honeybees that are under the influence of EtOH.
Previous studies from this laboratory have shown that it is
possible to obtain lawful data in free-flying honeybees when
bees are trained to fly to 1 or more targets containing a large
drop of EtOH, but an operant response reinforced with
microliter amounts had not previously been used. It is
possible that exposure to EtOH would inhibit or otherwise
interfere with tube entering, which was the operant used in
these experiments.
Second, we hoped to provide the first comparative data
using an invertebrate on the effect of EtOH on a traditional
schedule of reinforcement in an EtOH self-administration
protocol. Much information has been obtained on the effect
of EtOH on schedules with vertebrate models but not inver-
tebrate models (Henningfield and Meisch, 1976; Meisch and
Thompson, 1973).Continuous and intermittent positive rein-
forcement schedules have also been shown to be effective in
flies and honeybees (Boisvert and Sherry, 2006; Grossman,
1973; Sigurdson, 1981; Sokolowski et al., 2010). However,
despite the well-developed learning capacity of insects, the
usefulness of reinforcement schedules has surprisingly not
been recognized until now. This study would lay the founda-
tion necessary for more analytical experiments of the type
routinely used for vertebrate studies of schedule-controlled
behavior.
Third, we wanted to provide further data on the effect of
EtOH in an ecological context. Our previous data have
shown that EtOH interferes with honeybee decision-making
(Abramson et al., 2005). Such ecological studies are rare in
the EtOH literature with animals in which the majority of
studies use captive animals under laboratory conditions
(Meisch, 2001; Wiens et al., 2008). Finally, most learning
or pharmacological studies on animals rely on traditional
experimental group designs where data are averaged across
subjects (Yang et al., 2008). However, it is well known that
an average curve may hide the main features of a behav-
ioral phenomenon when individual differences are the norm
(Gallistel et al., 2004; Sidman, 1960; Stepanov and Abram-
son, 2008). Our final goal was to analyze individual data to
detect individual differences in EtOH self-administration
behavior.
MATERIALS AND METHODS
Subjects
The subjects were 25 European honeybees (Apis mellifera L.)
randomly selected from 1 of 6 colonies. The bees were captured at
an artificial feeder filled with an 8% sucrose solution. Bees captured
this way are known to be foragers between the age of 21 and
31 days. Whenoutsideofthe operantconditioning chamber, honey-
bees were free to fly and interact with nest mates.
Apparatus
The Skinner box was located in a garden about 25 m from the
colonies. Midway between the colonies and the apparatus was an
artificial feeder continuously providing a 10% sucrose solution.
Bees were first trained to come from their colonies to the artificial
feeder and eventually from their colonies directly to the Skinner
box. The Skinner box was composed of a computer-controlled
feeder(an artificialflower)
(42 9 34 9 117 cm) and connected to a Plexiglas conditioning
chamber (44 9 35 9 51 cm). Honeybees could enter the condition-
ing chamber, but only 1 bee could be in the chamber at a time. A
clear plastic door was used to keep the bee inside of the chamber
until it was finished. The space within the chamber was large, and a
bee could readily fly around the response tube. A sketch of the
apparatus is shown in Fig. 1.
The apparatus was connected to a USB control unit and from the
control unit to a portable computer (Windows XP; Microsoft
Corp., Redmond, WA). A Philips (Amsterdam, NL) SPC 900NC
locatedona woodentable
Fig. 1. Sketch of the apparatus. The apparatus includes a wood table,
an automatic feeder with a landing platform, and a conditioning chamber
with 2 doors (from Sokolowski and Abramson, 2010).
THE EFFECT OF ETHANOL IN THE HONEYBEE
1569

Page 3

webcam was located inside the chamber and fixed above the feeder.
All contingencies and data recording were performed automatically.
A joystick was connected to the computer, and by activating the
appropriate button, data would be directed to the file of a particular
bee.
The operant response was defined as entering a 6-mm-diameter
hole (the response hole) and was automatically detected with a set
of infrared emitter/detector pairs. The reinforcer was a 5-ll drop of
solution stored in a glasssyringe. When the appropriate contingency
was met, the syringe was activated and the solution appeared at the
bottom of the response hole. Details of the apparatus can be found
in Sokolowski and Abramson (2010). Test data revealed that the
apparatus provided accurate and reliable flow rates and response
detection.
Procedure
Recruitment, Shaping Phase, and Marking. After numerous hon-
eybees visited the artificial feeder, we placed small drops of 50%
solution in a Petri dish and positioned the dish close to the bees.
When a bee landed on the dish and began to feed upon the solution,
we moved the dish to the conditioning chamber. Subsequently, a
honeybee would return to the chamber on its own and recruit other
foragers. During the shaping phase, small droplets of 50% sucrose
solution (w/w) were deposited closer and closer to the response hole
with some droplets deposited inside the response hole. After a few
minutes, a bee would enter the hole, and we marked it with a
colored numbered tag (Catalog Item: Queen Number Set; Betterbee,
Inc., Greenwich, NY). The experiment started when the marked bee
returned several times within few minutes.
Experimental Design. We restricted access to the conditioning
chamber to only 1 bee per day. When the experimental bee arrived
at the operant chamber, we opened the door and immediately
closed it when the bee entered. When the crop of the bee was
filled, it was permitted to leave, and the experimenter waited for
its return. Data were collected on an individual bee throughout
the day and over the course of several days until the experiment
was completed.
During the experiment, we used a continuous reinforcement sche-
dule, and all hole-entering responses were reinforced. Following a
response, the bee was required to back out of the hole for a mini-
mum of 100 ms. Any response during this time-out was not rein-
forced and reset the time-out clock.
The experimental group consisted of 14 that performed the over-
all experiment. Eight bees stopped the experiment before the end of
the EtOH phase, and their results are not included in the analysis.
We used a traditional ABA design (Long and Hollin, 1995; Sidman,
1960). During the baseline (phase A1), the reinforcer was a 50%
sucrose solution (w/w). After 6 or 7 crop fillings (i.e., visits), we
shifted to phase B (treatment phase). The program was stopped,
and a new syringe was filled with a 50% sucrose solution mixed with
5% EtOH (v/v). When the solution was in place, the software was
reactivated. This manipulation could easily be performed during the
intervisit period when the bee was unloading its crop inside
the hive. To prevent evaporation, the mix was prepared just before
the start of phase B. This phase was continued for an additional 6 to
8 crop fillings. Following phase B, we returned to the no EtOH
baseline condition (phase A2). Phase A2 was continued for the
remainder of the afternoon. Their results are not included in the
analysis. The control group consisted of 11 bees that have been
exposed to the same conditions as the experimental group with the
exception that no EtOH was available during phase B. These bees
were employed to study the effect of long-term exposure to the
contingencies over the course of a day. The control bees received a
continuous reinforcement schedule with a 50% sucrose solution as
the reinforcer.
Data Collection and Analysis
For each visit, the software automatically recorded the data. We
measured 3 dependent variables: the load size, the intervisit interval,
and the interresponse times (IRTs). The load size is defined as the
amount of solution that the bee brings to the hive. As the software
precisely controls the amount of reinforcer, crop load is easily and
accurately measured by summing all of the reinforcer amounts col-
lected by the bee during a visit. An IRT is defined as the time
between the end of the last response and the beginning of the next
response. Long IRTs correspond to intervisit intervals when the
honeybee went back to the hive and unload her crop. It has been
observed that when going out from the conditioning chamber, the
honeybees go directly to the hive. As session length was limited by
amount of available sunlight, the amount of data collected with
each bee is limited to daylight hours.
RESULTS
Load Size
Figure 2 shows the average load size during successive
visits of the flower and across conditions for experimental
and control groups. Without EtOH, the average load sizes
are between 50 and 60 ll and fall inside the range of
previously reported values (Sokolowski and Abramson,
2010). Figure 2 shows that load size remains constant in the
control group during each of the 3 phases. The performance
of experimental bees is similar to control bees during the
non-EtOH phases (A1 and A2). However, EtOH consump-
tion decreased load size during the EtOH phase. Load size
immediately increased after the removal of EtOH with base-
line levels re-established after only 2 visits. Two 1-way analy-
ses of variance were performed to assess the effects of phases
separately in control and experimental groups. Analysis con-
firmed that there is no effect of phases in the control group, F
(2, 20) = 0.2924, p = 0.7496, but effect of phase proved to be
significantin theEtOH
p = 0.04464. t-Tests were subsequently performed to com-
pare experimental and control groups during each phase.
Load sizes are very similar in both groups during the baseline
conditions (phase A1: t = 0.7165, p = 0.482 and phase A2:
t = 1.1104, p = 0.28) but differ during phase B (t = 2.6641,
group,F(2,22) = 3.5931,
0
10
20
30
40
50
60
0510 152025
) s r e t i l
o r c i
m
(
e z i s
d
a
o
L
Successive visits
Control group
Ethanol group
ethanol
Fig. 2. Average load size (ll) as a function of successive visits for the
control and the ethanol (EtOH) group. The EtOH was introduced at the
seventh visit and removed at the 14th visit. The phases are separated by
the vertical dash lines.
1570
SOKOLOWSKI ET AL.

Page 4

p = 0.007259). When returning to baseline condition follow-
ing EtOH consumption, the load size returned to the first
baseline level after 2 visits. The difference between phase A1
and phase A2 is not significant (paired t-test, t = 0.6437,
p = 0.533).
As several behavioral phenotypes may be involved in
EtOH consumption, we carefully examined individual data.
Representative examples are shown in Fig. 3. In bee 20 rep-
resenting the control bees, load size did not vary significantly
over the course of the day. However, individual data for the
experimental bees show that the shift from phase A1 to phase
B or from phase B to phase A2 had an opposite effect in
some bees. When starting to consume EtOH, several bees
increased their load size during the first visit (see, e.g., bee 25)
when others decreased their load size (bee 24). When return-
ing to phase A2, some bees increased their load size (e.g., bee
27) when others decreased load size (bee 25). The amplitude
of the change combined with the constancy of the load size
level observed in the control group seems to show that hon-
eybees may react very differently when discovering EtOH in
the flower for the first time. This fact is supported by the
observation that the standard deviation is higher just after
the shift to the EtOH condition (Fig. 3, lower right panel).
Intervisit Interval
Figure 4 shows the intervisit interval as a function of
visits, phases, and EtOH condition. The intervisit interval
includes the time the honeybee spends inside the hive and the
time she takes to fly from the hive to the flower. Without
EtOH, the average intervisit intervals fall between 4 and
5 minutes. It is important to note that the intervisit interval
includes the time necessary to return to the hive, unload her
crop, and return to the conditioning chamber. Intervisit
intervals remain approximately constant in the control group
during all phases of the experiment. Similar levels are mea-
sured in the experimental group during the non-EtOH
phases (A1 and A2). However, EtOH consumption greatly
increased intervisit intervals. Intervisit interval immediately
returned to baseline levels after EtOH removal.
0
10
20
30
40
50
60
70
80
05 10
ethanol
152025 30
) s r e t i l
o r c i
m
(
e z i s
d
a
o
L
Bee 20
0
10
20
30
40
50
60
70
05 101520 25
control group
0
20
40
60
80
100
05101520 2530
) s r e t i l
o r c i
m
(
e z i s
d
a
o
L
Bee 25
0
20
40
60
80
100
0510 15202530
ethanol
Bee 21
0
20
40
60
80
100
05 10
Successive visits
15202530
) s r e t i l
o r c i
m
(
e z i s
d
a
o
L
ethanol
Bee 27
0
10
20
30
40
50
60
70
0510 152025
Successive visits
ethanol group
ethanol
Fig. 3. The lefttoppanel shows individualload size data for 1 typical control bee (bee20). Thetopright panel shows the averagedataforcontrolgroup
with standard deviation. The middle left and right and bottom left panels show selected individual load size data. The bottom right panel shows the
average data for experimental group with standard deviation. The phases are separated by the vertical dash lines.
THE EFFECT OF ETHANOL IN THE HONEYBEE
1571

Page 5

Two 1-way analyses of variance were performed to assess
the effects of phases separately in control and experimental
groups. There is no significant effect of phases in the control
group, F(2, 20) = 1.9434, p = 0.1693, but the effect is signifi-
cant in the EtOH group, F(2, 22) = 4.2328, p = 0.02785.
t-Tests were performed to compare experimental and control
group during each phase. The differences between experi-
mental and control groups conditions are not significant for
phases A1 and A2 (phase A1: t = 0.4746, p = 0.64 and phase
A2: t = ?0.5227, p = 0.6067) and also not significant for
phase B (t = 1.6269, p = 0.06486). However, if we compare
experimental and control groups during the 3 last visits of
phase B, the intervisit intervals in the experimental group are
significantlylonger(t = ?2.4614,
returning to the baseline condition after EtOH consumption,
the load size returned to the first baseline level after 2 visits.
The difference between phase A1 and phase A2 is not signifi-
cant (paired t-test, t = ?0.106, p = 0.9175).
Selected individual data about intervisit interval patterns
are shown in Fig. 5. As illustrated with control bee 18, we
visually detected no trends in her pattern of intervisit inter-
val. As shown in the top right panel of Fig. 5, at the group
level, the variability of intervisit interval remained low. How-
ever, individual data for the experimental group show that
the shift from phase A1 to phase B or from phase B to phase
A2 had the opposite effects in some bees. When starting to
consume EtOH, several bees increased their intervisit interval
(see, e.g., bee 5), others decreased it (bee 27), and some bees
did not change it (see bee 29). When returning to phase A2,
most bees returned to their initial intervisit interval level.
p = 0.01477).When
Interresponse Times
Figure 6 shows average IRT as a function of visits, phases,
and EtOH. The average IRT for the control group has low
variation and oscillates around 1-second. The introduction
of EtOH has no immediate effect, but curiously, IRTs started
to increase only after EtOH removal.
Two 1-way analyses of variance were performed to assess
the effects of phases separately in control and experimental
groups. There is no significant effect of phases in the control
group, F(2, 18) = 0.4484, p = 0.6456, and in the experimen-
tal group, F(2, 22) = 2.5877, p = 0.09788.
t-Tests were performed to compare experimental and
control group during each phase. The differences between
experimental and control groups are not significant for
phases A1 and A2 (phase A1: t = ?0.3536, p = 0.7274 and
phase A2: t = ?1.5969, p = 0.126) and not significant for
phase B (t = ?1.0928, p = 0.2875).
Inspection of individual data in the control group (see
Fig. 7) clearly shows 2 behavioral phenotypes. Some honey-
bees have very short IRTs (bee 17), while others have longer
and more variable IRTs (bee 21). One bee exhibited a
decreasing IRT curve with successive visits that seems to
show a learning process during the course of the experiment
(bee 16 not shown here).
As with other dependent variables, individual IRT data in
the experimental group show various patterns of responding
(see Fig. 7). For some bees, EtOH seems to increase IRTs
(e.g., bee 25) but not for others (see bee 22). The delayed
effect of alcohol observed in phase B is observable only in
some bees but is particularly important in bee 5.
CONCLUSIONS
Self-administration of EtOH on a continuous reinforce-
ment schedule was investigated in honeybees. The honeybees
were completely free to visit our Skinner box. All bees spent
most of their time gathering the sucrose solution at the auto-
matized flower. As previously observed with continuous
access experiments, most bees continued to collect solution
when contaminated with 5% EtOH (Abramson et al., 2000,
2004a,b), and they did so with a high rate of responding.
Contrary to most EtOH self-administration studies carried
out with rodents or monkeys living in isolation and artificial
conditions, our procedure is unique in the sense that it only
minimally disturbs the animals from their natural life. Except
in the well-controlled conditioning chamber where the tested
animals were alone, all bees were free to interact and commu-
nicate with other bees. This point indicates that honeybees
may gather and process small doses of EtOH, if encountered
in their environment, and suggests that our results are not
limited to the artificial environment of a laboratory.
Our study shows that after few visits, EtOH reduces the
honeybees’ load size and increases the intervisit interval.
Contrary to this immediate effect, EtOH does not seem to
affect IRTs. Such results suggest that different biological
mechanisms are involved in each kind of effect. The particu-
larity of the protocol is the fact that honeybees can continue
to interact with other bees during the process of the experi-
ment. Consequently, it is possible that the increase in
intervisit intervals comes from a difficulty of EtOH foragers
to find a receiver bee to unload her crop. Another possibility
could be an increase in rest time before a new foraging trip.
0
100
200
300
400
500
600
05 1015 2025
) c
e s ( l a
v r e t n i t i s i v r e t n I
Successive visits
Control group
Ethanol group
ethanol
Fig. 4. Average intervisit interval (seconds) as a function of successive
visits for the control and the ethanol (EtOH) group. The EtOH was intro-
duced at the seventh visit and removed at the 14th visit. The phases are
separated by the vertical dash lines. However, only 6 visits are shown here
because with 7 visits, we can compute only 6 intervals.
1572
SOKOLOWSKI ET AL.

Page 6

It is also possible for EtOH to increase flying times. It is unli-
kely that this hypothesis could account for the data because
in no case, we observed a disoriented bee looking for her
way. After leaving the conditioning box, most bees flew with
no delay and a rapid speed in the direction of the hives.
Only a direct observation of bees inside the hive between 2
foraging trips could help us to answer the question.
The fact that a continuous schedule of EtOH reinforce-
ment can maintain responding in honeybees extends the use-
fulness of such a model for the study of alcoholism. This
study opens the door to more detailed studies about sche-
dule-controlled behavior. Previously, most studies have been
performed with rodents (Elmer et al., 1986; Meisch and
Thompson, 1973; Risinger et al., 1998) or monkeys
(Henningfield and Meisch, 1976). However, at the behavioral
level, the honeybee model has some unique characteristics.
While rats and monkeys consume EtOH for their own bene-
fit, honeybees store most of the collected solution for future
usage in the hive. Consequently, we observed no satiation in
our bees. With rats, it has been shown that most of EtOH
consumption happens during the first part of the session
(Meisch and Thompson, 1973). No such phenomenon
0
100
200
300
400
500
0 102030
Intervisit interval(sec)
Bee 18
0
100
200
300
400
500
05 1015 2025
control group
0
200
400
600
800
05 1015 2025
Intervisit intervals (sec)
ethanol
Bee 29
0
200
400
600
800
05 1015 20 25
ethanol
Bee 5
0
200
400
600
800
05 10 152025
Intervisit interval (sec)
Successive visits
ethanol
Bee 27
0
200
400
600
800
05 1015 2025
Successive visits
ethanol
ethanol group
Fig. 5. The left top panel shows individual intervisit interval data for 1 typical control bee (bee 18). The right panel shows the average data for control
group with standard deviation. The middle left and right and bottom left panels show selected individual intervisit interval data. The bottom right panel
shows the average data for experimental group with standard deviation. The phasesare separated bythe vertical dash lines.
0
2
4
6
8
10
0510 15 2025
Average IRT (sec)
Successive visits
Control group
Ethanol group
ethanol
Fig. 6. Average interresponse time (IRT; seconds) as a function of suc-
cessive visits for the control and the ethanol (EtOH) group. The EtOH was
introduced at the seventh visit and removed at the 14th visit. The phases
are separated by the vertical dash lines.
THE EFFECT OF ETHANOL IN THE HONEYBEE
1573

Page 7

happened in honeybees: They worked continuously during
the entire intoxication phase.
As shown with individual data, some effects of EtOH on
behavior are clearly detectable at the individual level with the
use of an ABA experimental design. This result opens the
door to a more systematic use of such design in insect behav-
ior and learning studies to assess the effect of various toxics
or chemicals at the individual level (Cohn and MacPhail,
1996; Laties, 1978). Moreover, reinforcement schedules
appear to be a promising tool to model work and searching
time during honeybee foraging trips (Hodos and Trumbule,
1967; Kamil, 1983; Zeiler, 1984).
One important question our study raises is the impor-
tance of behavioral interindividual differences about the
effect of EtOH on nectar-foraging behaviors. For some
bees, EtOH had no effect, while for others we observed
opposite results. Other individual variations have already
been reported, for example about sucrose responsiveness
0
1
2
3
4
5
0 10 2030
Average IRT (sec)
Bee 17
0
1
2
3
4
5
05 10152025 30
Bee 21
0
1
2
3
4
0510 1520 25
Average IRT (sec)
control group
0
1
2
3
010 2030
ethanol
Bee 25
0
1
2
3
0 10
Successive visits
20 30
Average IRT (sec)
ethanol
Bee 22
0
2
4
6
8
10
12
0510
Successive visits
15202530
ethanol
Bee 5
0
2
4
6
8
10
05 10 152025
Average IRT (sec)
Successive visits
ethanol
Ethanol group
Fig. 7. The 3 top panel graphics shows individual interresponse time (IRT) data for 2 typical control bees (bee 17 with short IRTs and bee 21 with long
IRTs). The third panel shows the average data for control group with standard deviation. The 4 bottom left panels show selected individual IRT data. The
bottom right panel shows the average data for experimental group with standard deviation. The phases are separated by the vertical dash lines.
1574
SOKOLOWSKI ET AL.

Page 8

(Scheiner et al., 2004), pumping rate (Sokolowski and
Abramson, 2010), or aversive responsiveness (Roussel
et al., 2009). Several hypotheses can be advanced to explain
the observed behavioral variations. First, the possibility
exists that individual honeybees differ in rate of EtOH
metabolism; some bees may be able to tolerate larger
amounts (Bozic et al., 2007). Such large individual differ-
ences would not be surprising with respect to the genetic
structure and variation inside honeybee colonies (Tarpy
and Nielsen, 2002). Second, we do not know whether our
honeybees have already been exposed to fermenting nectar
and/or preexposed to EtOH. As many of the bees were able
to gather EtOH in the Skinner box, it is possible that some,
but not all, of the bees had already been exposed to EtOH
in empty alcoholic beverage bottles. Consequently, some
honeybees may have become more tolerant (Siegel et al.,
2000). Even if not exposed to EtOH before the experiment,
bees may also differ in learned tolerance if we suppose that
tolerance starts to develop from the first visit and grows
with each successive visit (Krank and O’Neill, 2002; Siegel,
2005). In any case, a close description of individual differ-
ences and a search of parameters and mechanisms that
could affect them will be the goal of our future research.
We did not use an isocaloric EtOH solution, and so the
changes observed in dependent variables may have been
caused by an increased caloric content rather than because of
a qualitative change of the reinforcer. We discount this inter-
pretation for several reasons. First, because of the high
sucrose content of the solution that we used, the caloric
increase is only small (<2%). Second, honeybees have a
preference for high sucrose concentration and high caloric
solutions. If the increase in caloric content explains observed
behavioral changes, we cannot explain why several bees
stopped to return to the apparatus after several visits. Third,
small drops of sucrose or caloric solution consumption are
known to be associated with increased activity in insects
(Dethier, 1976). Sucrose and caloric concentration affects
also honeybee choice between flowers (Banschbach, 1994;
Cnaani et al., 2006) and dances inside the hive (Waddington,
1982). However, when switching from 20 to 50% sucrose
solutions, honeybees decrease the time they took to collect
small drop of solution from one place to another adjacent
one (Loo and Bitterman, 1994). All these results lead to the
prediction that IRTs and intervisit intervals should decrease
with increasing caloric content. In our experiment, we
observed exactly the reverse.
At the psychological level, several mechanisms may be
involved and influence the results. It is possible that some
honeybees have been subjected to EtOH aversion learning—
illness occurring progressively and being associated with
EtOH taste (Davis and Riley, 2010). Another possibility is
that the decreased response rate with increasing IRTs is a
result of the punitive effect of EtOH, the sucrose solution
continuing to serve as a reinforcer. However, it is also
possible for changes in the dependent variables to be the
result of an unconditioned effect of EtOH consumption
(Maze et al., 2006). Future research is needed to assess all
these hypotheses.
Another question is whether EtOH can serve as a rein-
forcer for honeybees. Our study was not designed to answer
that question. At this point in our investigation, we believe
that the reinforcer is the sucrose solution. What we can say
is that we know that EtOH is not aversive to honeybees
(Abramson et al., 2004a,b). Several comments can be made.
First, from an evolutionary point of view, EtOH does not
need to be a reinforcer for the evolution of EtOH metabo-
lism. Only EtOH exposure and self-administration is
required. However, the problem is different if the selective
consumption of EtOH offers particular fitness advantages
leading to active EtOH search in the environment. Second,
we usedcontinuousreinforcement,
amounts, and high EtOH content (5%). Such large
amounts and high concentrations of EtOH are probably
rare in the environment as fermented nectar has only a low
EtOH content (<1%). It is possible that foraging in a less
rich environment (that could be simulated with intermittent
or progressive schedules) in association with low honey
storage in the colony could yield a more active search for
EtOH. Moreover, testing lower EtOH concentrations
(<1%) could perhaps give interesting results if the reinforce-
ment process follows dose–effect rules. Finally, to test the
reinforcing properties of EtOH, choice protocols would be
an interesting tool. In this case, the honeybee could, for
example, choose between only sucrose reinforcement and
sucrose plus EtOH reinforcement. With such a protocol
that could be implemented with our device, we could then
easily perform parametric studies to detect true reinforce-
ment effects of EtOH in honeybees. If EtOH reinforcement
could be demonstrated, that would considerably expand the
usefulness of invertebrates and especially insects and honey-
bees in the study of drug and EtOH consumption (Wolf
and Heberlein, 2003).
At a procedural level, we presented honeybees with a mix-
ture of 50% sucrose/5% EtOH solution following a period
of reinforcement with only the 50% sucrose solution. This
procedure is in some way equivalent to the sucrose substitu-
tion procedure. In this case, to encourage bees to consume
EtOH, it is first presented with a sweetener that is subse-
quently removed later in the experiment (Koob, 2000;
Roberts et al., 1999). The rationale for using such a proce-
dure in rodents or monkeys is that EtOH is not consumed
easily by unprepared animals. We did not remove the sucrose
in our experiment. Our rationale for not doing so is that we
wanted to place our experiment in an ecological context of
honeybees foraging on fermenting nectar. Honeybees could
move freely from the hive and continue to interact with all
other members of the colony. We believe that this point is
especially important if we consider the consumption of
EtOH as a natural phenomenon with specific behavioral and
neurological adaptations. However, the efficiency with which
honeybees forage alcoholic beverages in a natural context
still needs to be investigated.
largereinforcer
THE EFFECT OF ETHANOL IN THE HONEYBEE
1575

Page 9

ACKNOWLEDGMENTS
This research was funded by the Socie ´ te ´ Picarde pour
l’Etude du Comportement Animal (http://speca.free.fr), the
‘Unite ´ Dynamiques des Syste ` mes Anthropise ´ s (JE 2532)’, the
‘INSERM-ERI 24, Groupe de Recherche sur l’Alcool et les
Pharmacode ´ pendances’, and the Laboratory of Comparative
Psychology and Behavioral Biology. Parts of the data have
been collected during the completion by the first author of a
Fulbright scholarship program at the Oklahoma State
University. We would like to express our appreciation to
Bradley R. Gibson, BA for his help in data collection.
Special thanks to Mickael Naassila for his support and
encouragement to conduct that research.
REFERENCES
Abramson CI, Kandolf A, Sheridan A, Donohue D, Boz ˇ ič J, Meyers JE,
Benbassat D (2004a) Development of an ethanol model using social
insects: III. Preferences for ethanol solutions. Psychol Rep 94:227–239.
Abramson CI, Place AJ, Aquino IS, Fernandez A (2004b) Development of
an ethanol model using social insects: IV. Influence of ethanol on the
aggression of africanized honeybee (Apis mellifera L.). Psychol Rep
94:1107–1115.
Abramson CI, Sanderson C, Painter J, Barnett S, Wells H (2005) Develop-
ment of an ethanol model using social insects: V. Honeybee foraging
decisions under the influence of alcohol. Alcohol 36:187–193.
AbramsonCI, Stone SM, Ortez RA, Luccardi A, VannKL, Hanig KD, Rice
J (2000) The development of an ethanol model using social insects I:
behavior studies of the honey bee (Apis mellifera L.). Alcohol Clin Exp
Res 24:1153–1166.
Banschbach VS (1994) Colour association influences honey bee choice
between sucrose concentrations. J Comp Physiol A 175:107–114.
Boisvert M, Sherry DF (2006) Interval timing by an invertebrate, the bumble
bee Bombus impatiens. Curr Biol 16:1–5.
Bozic J, Abramson CI, Bedencic M (2006) Reduced ability of ethanol drink-
ers for social communication in honeybees (Apis mellifera carnica Poll.).
Alcohol 38:178–183.
Bozic J, DiCesare J, Wells H, Abramson CI (2007) Ethanol levels in
honeybee hemolymph resulting from alcohol ingestion. Alcohol 41:281–
284.
Cakmak I, Abramson CI, Seven-Cakmak S, Nentchev P, Well H (2009)
Observations of ethanol exposure on the queen honey bee Apis mellifera
anatoliaca(preliminary note). Bull Insectology 62:99–101.
Cnaani J, Thomson JD, Papaj DJ (2006) Flower choice and learning in for-
aging bumblebees: effects of variation in nectar volume and concentration.
Ethology 112:278–285.
Cohn J, MacPhail RC (1996) Ethological and experimental approaches to
behavior analysis: implications for ecotoxicology. EnvironHealth Perspect
104:299–305.
Davis CM, Riley AL (2010) Conditioned taste aversion learning:
implications for animal models of drug abuse. Ann NY Acad Sci 1187:
247–275.
Dethier VG (1976) The Hungry Fly: A Physiological Study of the Behavior
Associated with Feeding. Harvard University Press, Cambridge, MA.
DudleyR (2002) Fermentingfruit andthehistorical ecologyofethanolinges-
tion: is alcoholism in modern humans an evolutionary hangover? Addic-
tion 97:381–388.
Dudley R (2004) Ethanol, fruit ripening, and the historical origins of human
alcoholism in primate frugivory. Integr Comp Biol 44:315–323.
Ehlers B, Olesen JM (1997) The fruit-wasp route to toxic nectar in Epipactis
orchids? Flora 192:223–229.
Elmer GI, Meisch RA, George FR (1986) Oral ethanol reinforced behavior
in inbredmice. Pharmacol Biochem Behav 24:1417–1421.
Etges WJ, Klassen CS (1989) Influences of atmospheric ethanol on adult
Drosophila mojavensis: altered metabolic rates and increase in fitness
among populations. Physiol Zool 62:170–193.
Gallistel CR, Fairhurst S, Balsam P (2004) The learning curve: implications
of a quantitative analysis. Proc Natl Acad Sci USA 1001:13124–13131.
Geer BW, McKechnie SW, Heinstra PWH, Pyka MJ (1991) Heritable
variation in ethanol tolerance and its association with biochemical traits in
Drosophila melanogaster. Evolution 45:1107–1119.
Grossman KE (1973) Continuous, fixed-ratio, and fixed-interval reinforce-
ment in honey bees. J Exp Anal Behav 20:105–109.
Guarnieri DJ, Heberlein U (2002) Drosophila melanogaster, a genetic model
system for alcohol research. Int Rev Neurobiol 54:199–228.
Henningfield JE, Meisch RA (1976) Ethanol as a positive reinforcer via the
oral route for rhesus monkeys: maintenance of fixed-ratio responding.
Pharmacol Biochem Behav 4:473–475.
Hodos W, Trumbule GH (1967) Strategies of schedule preference in chim-
panzees. J Exp Anal Behav 10:503–514.
Hranitz JM, Abramson CI, Carter RP (2010) Ethanol increases HSP70
concentration in honey bee (Apis mellifera L.) brain tissue. Alcohol 44:275
–282.
Jakubska A, Przado D, Steininger M, Aniol-Kwiatkowska J, Kadej M
(2005) Why do pollinators become “sluggish”? Nectar chemical constitu-
ents from Epipactis helleborine (L.) crantz (orchidaceae). Appl Ecol Env
Res 3:29–38.
Kamil AC (1983) Optimal foraging theory and the psychology of learning.
Am Zool 23:291–302.
Kevan PG, Eisikowitch D, Fowle S, Thomas K (1988) Yeast-contaminated
nectar and itseffects on bee foraging. J Apicult Res 27:26–29.
Koob GF (2000) Animal models of craving for alcohol. Addiction 95:S73–
S81.
Krank MD, O’Neill S (2002) Environmental context conditioning with etha-
nol reducesthe aversiveeffects ofethanol in the acquisition ofself-adminis-
tration in rats. Psychopharmacology 159:258–265.
Laties VG (1978) How operant conditioning can contribute to behavioral
toxicology. Environ Health Perspect 26:29–35.
Long CG, Hollin CR (1995) Single case design: a critique of methodology
and analysis of recent trends. Clin Psychol Psychother 2:177–191.
Loo SK, Bitterman ME (1994) Learning in the honeybees (Apis mellifera) as
a function of sucrose concentration. J Comp Psychol 106:29–36.
Maze IS, Wright GA, Mustard JA (2006) Acute ethanol ingestion produces
dose-dependent effects on motor behavior in the honey bee (Apis mellifer-
a). J Insect Physiol 52:1243–1253.
Mazeh S, Korine C, Pinshow B, Dudley R (2008) The influence of ethanol
on feeding in the frugivorous yellow-vented bulbul (Pycnonotus xanthopy-
gos). Behav Process 77:369–375.
Meisch RA (2001) Oral drug self-administration: an overview of laboratory
animal studies. Alcohol 24:117–128.
MeischRA, ThompsonT (1973)Ethanol as a reinforcer:effects of fixed-ratio
size and food deprivation. Psychopharmacologia 28:171–183.
Risinger FO, Brown MM, Doan AM, Oakes RA (1998) Mouse strain differ-
ences in oral ethanol reinforcement under continuous access conditions.
Alcohol Clin Exp Res 22:677–684.
Roberts A, Heyser CJ, Koob GF (1999) Operant self-administration of
sweetened versus unsweetened ethanol: effects on blood alcohol levels.
Alcohol Clin Exp Res 23:1151–1157.
Roussel E, Carcaud J, Sandoz JC, Giurfa M (2009) Reappraising social
insect behavior through aversive responsiveness and learning. PLoS ONE
4:1–9.
Sanchez F, Kotler BP, Korine C, Pinshow B (2008) Sugars are complemen-
tary resources to ethanol in foods consumed by Egyptian fruit bats. J Exp
Biol 211:1475–1481.
Scheiner R, Page RE, Erber J (2004) Sucrose responsiveness and behavioral
plasticity in honey bees (Apis mellifera). Apidologie 35:133–142.
Sidman M (1960) Tactics of Scientific Research: Evaluating Experimental
Data in Psychology. Basic Books, Inc., New York.
Siegel S (2005) Drug tolerance, drug addiction, and drug anticipation. Curr
Dir Psychol Sci 14:296–300.
1576
SOKOLOWSKI ET AL.

Page 10

Siegel S, Baptista MAS, Kim JA, McDonald RV, Weise-Kelly L (2000)
Pavlovian psychopharmacology: the associative basis of tolerance. Exp
Clin Psychopharm 8:276–293.
Sigurdson JE (1981) Measurement of consummatory behavior in honey bees.
Behav Res Methods 13:308–310.
Sokolowski MBC, AbramsonCI (2010) From foragingtooperantcondition-
ing: a new computer-controlled Skinner box to study free-flying nectar
gathering behavior in bees. J Neurosci Methods 188:235–242.
Sokolowski MBC, Disma G, Abramson CI (2010) A paradigm for operant
conditioning in blow flies (Phormia terrae novae Robineau-Desvoidy,
1830). J Exp Anal Behav 93:81–89.
Stepanov II, Abramson CI (2008) The application of an exponential mathe-
matical model for 3-arm radial maze learning. J Math Psychol 52:309–319.
Sullivan RJ, Hagen EH (2002) Psychotropic substance-seeking: evolutionary
pathology or adaptation? Addiction 97:389–400.
Tarpy DR, Nielsen DI (2002) Sampling error, effective paternity, and
estimating the genetic structure of honey bee colonies (Hymenoptera:
Apidae). Ann Entomol Soc Am 95:513–528.
Waddington KD (1982) Honey bee foraging profitability and round dance
correlates. J Comp Physiol 148:297–301.
Wiens F, Zitzmann A, Lachance MA, Yegles M, Pragst F, Wurst FM, von
Holst D, Guan SL, Spanagel R (2008) Chronic intake of fermented floral
nectar by wild treeshrews. Proc Natl Acad Sci USA 105:10426–10431.
Wolf FW, Heberlein U (2003) Invertebrate models of drug abuse. J Neuro-
biol 54:161–178.
Yang EC, Chuang YC, Chen YL, Chang LH (2008) Abnormal foraging
behavior induced by sublethal dosage of imidacloprid in the honey bee
(Hymenoptera: Apidae). J Econ Entomol 101:1743–1748.
Zeiler MD (1984) The sleeping giant: reinforcement schedules. J Exp Anal
Behav 42:485–496.
THE EFFECT OF ETHANOL IN THE HONEYBEE
1577