Bee Olfactory Conditioning Apparatus

Honey bees are not only fascinating for providing honey and pollinating flowers, they are also a well-established model organism in many fields of research. When I was working at the University of Osnabrück, Germany, there were several bee hives in our institute. As a behavioral biologist with a strong interest in learning and memory I decided that it was timely to get some student projects started to investigate the learning behavior of our bees.

There are many different fascinating tasks that honey bees can learn. For example honey bees have advanced navigation skills, can communicate position of resources by means of waggle dance, and they can learn to differentiate colors, shapes, and fragrances. In 2011 it was shown that there is evidence that whether bees make an "optimistic" or "pessimistic" decision depends upon an emotion analogues inner state. In that experiment the bees were trained to associate a reward (sugared water) with a certain odor and a punishment (unenjoyable quinine solution) with a different smell. By presenting mixtures of both odors it could be tested if the bees were optimistically expecting a reward or pessimistically expecting an unpleasant taste. Honey bees were more likely to expect a punishment instead of a reward when they were vigorously shaken before the experiment. Those bees that did not experience this hassle were more likely to expect a reward (Bateson et al. 2011). We successfully conducted cognitive bias tests in honey bees, too and could show that a treatment against the varroa mites does not render the bees to be more pessimistic (Schlüns et al. 2017).

The test paradigm used here relies on the so called proboscis extension reflex (PER). This reflex describes folding out the "tongue" of an insect after antennal contact with something palatable. Touching a bees antenna with sucrose solution is an unconditioned stimulus which elicits a unconditioned reaction, namely the PER. Using techniques of classical Pavlovian conditioning the unconditioned reaction (PER) can be associated with a neutral stimulus (odor) which as a consequence become a conditioned stimulus eliciting the now conditioned reaction without the presence of the unconditioned stimulus.

In order to do this experimentally we build an apparatus that allows to process the conditioning in a standardized way. The apparatus consists of a mount for a restrained bee, a fan that absorbs air behind the bee, a mini pump that provides an air stream containing the odor which is pumped through a tube, a servomotor rotating the outlet of the tube in front of the honey bee, and a sound and light signal indicating the time point for presenting the unconditioned stimulus.

This instructable shows how to build such an apparatus.

When the first version of the apparatus was built, we used parts that were at hand in our workshop. This parts list is thus a mixture of recommended items and the items we used. It should be no problem to use different parts as long as the basic electronic rules regarding voltage and amperage are followed.

- ARDUINO UnoRev. 3, ATmega328, USB (19,50 €)

Alternatively a cheap clone e.g., Dswy_Robot (<3 €). If you have no experience with ARDUINO programming I would recommend starting with the original a) because of supporting the developers and b) finding optimal support - in case anything is not working, these are the best documented models. See also links below for getting started with ARDUINO programming.

- Air pump 12V KPM27H (2,95 €) Pumping volume is 2 liter per minute

- Top-Line Standard Servo RS 2 MG/BB (7,99 €)

- Step down converter LM2596 (1,00 €)

- Power supply 12v, 1000ma (10,00 €)

- Relay (3,00 €)

- Push button

- Piezo buzzer

- LED

- Resistors (330 and 10 k Ohm)

- a few cable, solder, a blank perfboard (DOT PCB)

There are some electronic parts that use 12 volts and others that work on 5v. Therefore two isolated circuits are realized to fulfill the voltage needs for the different devices. The ARDUINO clone used here works on 5 volts and we thus need a voltage regulator to drop down from our 12v power supply. If you use an ARDUINO UNO, it can be powered by 12v as it already includes a voltage regulator.

From the original 12v source we also power the fan and the mini pump which is connected to the power line via a relay. All other parts are powered by 5v.

The scheme shown above shows how everything should be connected. Before soldering the parts you might want to make sure that the wires are long enough to put the parts to their final place of the apparatus.

In order to start the the conditioning procedure, first a button is pressed. Then the servomotor is rotating in place and the mini pump is started. A tone as well as an LED is signaling the need for the experimenter to provide the stimulus. Finally the mini pump is stopped and the servo motor revolves back to the starting position. This all is realized by a small piece of software that has to be uploaded to the ARDUINO. The source code is richly documented and attached to this instructable.

If you are new to programming microcomputers you might want to follow these links to get started:

https://www.instructables.com/id/Arduino-Basics/

https://www.instructables.com/id/Arduino/

https://www.instructables.com/class/Arduino-Class/

In the original version we hid the chaotic looking electronic parts in a box (see picture above). In this new version which was re-build for the purpose of this instructable I just glued everything into place on a wooden platform.

The purpose of the apparatus is to provide a constant air stream behind the bee in order to erase any cues from prior testing. This is what the fan (found in a pile of old computer parts) does. A defined amount of scented air stream should be presented. Therefore a mini pump is started that will pump air through a tube. In order to odorize the air stream, we used a modified infusion set. I cut a small window into the reservoir of the infusion set where a small tissue with a few micro liters of odor could be inserted. Then a pipette tip was cut open a little bit and attached to the tube. This outlet of the air stream was mounted to the servo motor using polymorph formable plastic. If you are working with different odors, you should use a different infusion set for each odor in order to prevent mixing leftovers from previous odors with the new odor. The mount for the bee is made of an aluminum tube which was cut as shown below. This tube is held in place in front of the pipette tip by a construction made of perforated steel sheet.

For presenting an olfactory stimulus the bee have to be be restricted in her movement. This can be achieved by fastening the bee in a bee holder. In order to have a reasonable throughput in experimentation about 20 bee holders should be available. The bee holders we used were made of an aluminum tube with 1 cm diameter. The tube was cut into pieces of 4 cm. Each of the small tubes was cut 1 cm from top and at an angle of 60° from the side. The edges were filed and smoothed using steel wool. For mounting the bee a small duct tape strip of 1 mm is used. In order to prevent the bee from sticking to the tape a thin plastic strip is attached in the contact zone. We used the back side of a plastic folder for the thin plastic. The desired piece of duct tape with the isolator can be build by gluing a strip of duct tape to the backside of the plastic folder, cutting thin strips and removing the isolating plastic at both ends of the strip (see figures).

The strip is attached to the side of the tube and in addition a band aid can be used to prevent the bee from stinging the experimenter.

Actually the process of conditioning a bee is including a number of steps that would not fit here. An accompanying publication is in preparation for the detailed procedure. Please see our recent paper for a short description of the procedure while waiting on the detailed version.

Schlüns, H; Welling, H; Federici, JR; Lewejohann, L (2017): The glass is not yet half empty: Agitation but not Varroa treatment causes cognitive bias in honey bees. Animal Cognition. 20 (2): 233–241.

Also see the first published version of a cognitive bias test:

Bateson, M; Desire, S; Gartside, SE; Wright, GA (2011): Agitated honeybees exhibit pessimistic cognitive biases. Current Biology 21(12):1070–1073