Site icon Conservation news

Camera-wielding robot records effects of pesticide on bees’ behavior

  • Bee populations are on the decline, and studies have linked this to the use of pesticides containing neonicotinoid compounds, which can impact insect behavior.
  • Researchers built a robotic platform that allowed them to observe the impacts of neonicotinoid compounds on bumblebee behavior inside bee colonies over a 12-day period.
  • The robotic observation platform held computer-programmed movable cameras that could monitor up to 12 colonies at a time, which included foraging and nesting chambers with simulated “daytime” and “nighttime” conditions.
  • The team found that bumblebees exposed to environmentally realistic amounts of neonicotinoid compounds reduced their nursing and caretaking activities at night and were less able to regulate the colony’s temperature, among other behavioral changes that may impact their population.

“It’s easy to quantify if insects are dead,” said Harvard post-doctoral fellow James Crall, lead author of the recent study “Neonicotinoid exposure disrupts bumblebee nest behavior, social networks, and thermoregulation.” What’s more challenging, according to Crall, is studying behavioral changes in living insects, specifically bees in this case.

It’s now well-known that bees, vital to crop production and the survival of flowering plants, are in trouble. Neonicotinoid compounds, the most popular type of insecticides, have been shown in various studies to negatively impact wild bee populations as well, most notably by reducing colony sizes. However, the specific ways that these compounds shrink a colony’s size is still not well-understood. That’s where the robot comes in.

Crall and his colleagues wanted to see how neonicotinoid exposure affected social behaviors within bumblebee (Bombus impatiens) colonies. Bumblebees are social insects that work as a team to keep their colony functioning, so changes in their behaviors may affect a colony’s survival.

A bumblebee (Bombus impatiens) worker outfitted with a unique tag (BEEtag), foraging outdoors. Image by James Crall.

To monitor each neonicotinoid-exposed colony as a unit, the researchers built a robot that would allow them to observe an entire colony without disturbing it. While they had in the past tracked individual worker bees tagged with simplified QR codes after exposing them to the pesticide compounds, developing the robotic camera system allowed them to record and track the bees and their interactions across multiple colonies.

The team began at the fruit fly neuroscience lab at the Harvard Center for Brain Science, reworking the chassis, essentially the backbone, of a robot that was used for fruit fly monitoring. The robot they built was a bridge-like structure carrying a pair of video cameras, one to record behavior in the nest and the other to record behavior in a separate foraging chamber. The cameras were able to move freely across the planar structure suspended above the colonies, allowing the researchers to monitor a given colony underneath.

The researchers programmed a computer to tell the cameras how long to look at which colonies; they were able to monitor up to 12 at a time. The robot moved these cameras over the different colonies and automated the filming of the tagged bees over multiple days.

Chronic exposure to a neonicotinoid pesticide alters nest behavior and social interactions in bumblebee colonies. Schematic diagrams of (A) the robotic platform for tracking bee behavior inside multiple colonies and (B) inside a single colony chamber. (C) A sample tracking of worker bees inside the nest, each with a unique ID number shown in green. The range dotted line shows the nest structure. Image is Figure 1 of Crall et al (2018).

Each of 18 study colonies had a foraging chamber, where the bees found pollen and artificial nectar made of sugar water. In nine colonies, the nectar contained the amount of pesticide the team would have expected to find in flower nectar that was also exposed. The other nine colonies had access to pesticide-free nectar. Each foraging chamber maintained “daytime” and “nighttime” periods like out in the real world.

In each colony, the researchers also crafted a nesting chamber, which was kept dark through the use of a plastic that let through infrared light for filming but blocked out visible light. They could then remotely observe the behavioral impacts of exposure to the pesticide using the robot, keeping the nest environment as close to natural conditions as possible.

The robotic data collection platform for studying the effects of chronic pesticide exposure on bees’ behavior in their nests. The artificial nests are in the white boxes. Each has a __ chamber with a clear top and a nesting chamber with a darkened top that let infrared light through for filming but blocked out visible light. Image by James Crall.

The researchers found that exposing the colony to these neonicotinoids in amounts that could realistically be expected in the natural environment affected bees’ ability to thermoregulate, meaning they were less able to control the temperature inside their nest.

“We think of insects in general as being cold-blooded and their body temperatures fluctuate a lot,” Crall said. “But they do a remarkable job of maintaining a stable temperature inside of the nest.”

Monitoring of the internal temperature of outdoor bumblebee colonies with digital sensors showed that chronic neonicotinoid pesticide exposure disrupted the bees’ thermoregulation. Image by James Crall.

The team found that the bees exposed to the neonicotinoid were less able to control the temperatures of the colony, and specifically the developing young and larvae within the nest, when outside temperatures dropped. Additionally, these workers failed to construct a wax canopy over the developing brood (eggs, larvae, and pupae) that they typically build during times of prolonged cold temperatures to keep them warm.

“That behavior was completely wiped out in pesticide-exposed colonies,” Crall said. On the other hand, the next-door control colonies that were not exposed to the chemical built the wax caps as expected.

The pesticide affected other behaviors. Bees exposed to it stayed along the edges of their nest more frequently than bees that consumed the pure nectar. The exposed bees also spent less time nursing, taking care of the nest, and interacting with nest-mates, and were less active overall than bees that had access to uncontaminated nectar.

A bumblebee colony with tagged workers constructing an insulating wax canopy to keep the developing larvae and pupae warm. Image by James Crall.

Some of these impacts were significant during the day while others were not, but the researchers saw the greatest change in all of these behaviors at night. A lot happens at night in bee nests, including nursing and caring for young. Yet nighttime is often under-studied when it comes to looking at the impacts of these chemicals on bumblebee populations. Some of the contaminated worker bees that remained immobile at night seemed to recover during the day.

“Every time we’re making a decision about the safety of these compounds for insects we care about and insects in general, we tend to look at them during the day,” Crall said.

According to Crall, the robot allowed the researchers to experiment with the bumblebees in an environment that better approximated realistic field exposures than other experiments had been able to in the past.

There were still limitations, however. “The lab is still the lab, and our little foraging chamber is nothing like foraging out over kilometers,” Crall said. For example, the robot was not built to withstand the wind or rain or even too much sunshine and was kept indoors.

The robotic data collection platform, with just a few nests, with data collection monitor showing the inside of a nest on the right. The system was kept indoors, so researchers would like to monitor the effects of pesticides on bees in a more natural conditions. Image by James Crall.

Crall said he would like to continue to look at the varied impacts of these chemicals on bee populations, examining not just their direct behavioral effects but also their effects combined with other stressors. For example, when the air temperature got really cold, the team saw strong differences between control and treated colonies. They would like to look further at the interactions between these stressors to better understand what is happening in the overall system. They also plan to assess impacts from other pesticides.

“Now that we think we have a good tool for studying these kinds of behaviors at low levels,” Crall said, “we’d like to start studying different compounds at different concentrations.”

Bumblebees at work covering the colony with an insulating wax canopy to maintain its temperature in cold conditions. Image by James Crall.

In addition, while bumblebees are social, only about 5 percent of bee species are. Crall said solitary bees were probably affected more than social bees by the impacts of the neonicotinoids, but they aren’t yet sure why.

“One of the goals of this work is of course not just to understand in the abstract what the effects of these compounds are,” Crall said, “but at these concentrations in the field and in the outdoors, do these effects matter?”


Crall, J. D., Switzer, C. M., Oppenheimer, R. L., Versypt, A. N. F., Dey, B., Brown, A., … & de Bivort, B. L. (2018). Neonicotinoid exposure disrupts bumblebee nest behavior, social networks, and thermoregulation. Science, 362(6415), 683-686.


FEEDBACK: Use this form to send a message to the editor of this post. If you want to post a public comment, you can do that at the bottom of the page.