A Banana-Like Smell Can Signal Alarm to Honeybees

A banana-like smell can be a signal that recruits honeybees to defend their colony. Its main component is isoamyl acetate. A 2021 study found that increasing the concentration of this alarm compound increased the proportion of bees that stung, but at very high concentrations that proportion fell instead.

Alongside behavioral experiments, the researchers built a computer model evaluated by the survival of the group. The study explored the environments in which responses might evolve not just to escalate an alarm, but also to limit unnecessary losses.

An alarm’s concentration, not just its presence, can carry information

The alarm pheromone associated with a honeybee’s sting is a mixture of more than 40 compounds. Because it is released whenever another bee stings, its concentration may provide a clue not only to the presence of danger but also to how many bees have already joined the defense.

The study connected two questions: how concentration changes an individual bee’s stinging response, and how that response contributes to collective defense and is shaped by evolutionary pressures.

Testing real stingers and a synthetic version of the main component

The experiments used western honeybees kept at the University of Konstanz in Germany that were responding to a threat to their hive. Each bee was placed alone in an arena and tested for whether it stung a rotating target covered in black leather. Researchers observed its posture and also checked for a stinger left in the leather at the end.

In the first experiment, zero to seven collected stingers were placed on the target. A total of 756 bees from three colonies were tested. The proportion that stung rose from about 20% with the target alone to about 60% with seven stingers. Although colonies differed in their overall tendency to sting, the tendency for the response to increase with more stingers was shared.

Because adding stingers made it difficult to test high concentrations, the researchers next used the main component, isoamyl acetate. Liquid diluted in mineral oil was placed on filter paper, and its odor was carried into the arena by airflow. Eight liquid concentrations from 0% to 100% were tested. These numbers are not the concentrations in the air the bees breathed.

The reported sample counts for this experiment are inconsistent: the Methods give a total of 384 bees, while the figure caption states 72 bees per concentration. Rather than settling on a count for each condition, this article focuses on the reported direction of the responses and comparisons.

The relationship between concentration and stinging changed partway

Between 0% and 25% isoamyl acetate, the proportion of bees that stung increased with concentration. It then fell at 50% and 100%, and the lower proportion at 100% than at 25% was statistically significant.

The response to the alarm was therefore not a simple upward progression. It showed two phases: increasing recruitment at low to intermediate concentrations, followed by a weaker response at very high concentrations.

Each bee was tested only once. The researchers did not observe the same individual following a rise in concentration and then ceasing to sting. Instead, they compared the proportion of individuals that stung at each concentration.

A model of collective survival explores what the alarm might mean

In the basic model, 100 virtual bees choose whether to sting according to the alarm concentration. When a bee stings, the alarm increases and that individual is lost. Meanwhile, a predator kills bees and stops attacking after receiving a specified number of stings.

The outcome was evaluated by how many bees remained alive at the end. Computations were repeated so that responses associated with better collective performance were passed on. Learning here was not an individual bee learning from experience, but a mechanism representing selection across generations.

This approach allowed the researchers to vary the frequency of false alarms without an enemy present, the predator’s resistance, how early it was detected, and the delay before its departure was recognized, separating their effects on the response.

False alarms and predator resistance shaped the response

In environments with frequent false alarms, the probability of stinging before alarm compounds were present became lower. Responding every time to a falling branch or a passing animal would unnecessarily deplete the colony. At the same time, the model developed a response that rapidly intensified defense when other bees really did sting and the alarm increased.

The concentration range over which responses stayed high was strongly influenced by the predator that required the most stings to drive away. When weak predators were encountered more often, responses at high concentrations weakened, and survival against resistant predators also fell.

When visually confirming the predator’s departure took longer, the model also developed a tendency to curb stinging in response to high alarm concentrations. This can be explained as a way to reduce losses from continuing an attack after enough stings have already been delivered but before departure can be seen.

The authors also retained an alternative explanation. Such high concentrations might rarely be encountered in nature, leaving no opportunity for an appropriate response to evolve, so the model could revert to its initial response. The decrease at high concentrations was not assigned to one definitive mechanism.

Connections to defended territory and regional differences

With earlier predator detection, bees could drive the predator away before being killed even with a lower probability of stinging. Defending a larger area, however, also meant waiting longer for the predator to leave that area. In the model, a larger defended territory was advantageous against resistant predators, while a smaller territory was advantageous against predators that were easier to deter.

The researchers also compared the model with previously published data on the responses of European and African honeybees. Settings involving more frequent encounters with more resistant predators helped explain the strong defensive responses known in African bees. This was a comparison of existing evidence with a model, not a new regional comparison experiment. Gathering actual predation rates and other environmental data to test the predictions remains a next step.

Summary

Honeybees become more likely to sting as alarm compounds increase, but in this experiment the proportion that stung fell at very high concentrations. In addition to recruiting nestmates, alarm concentration may convey how far the collective defense has already progressed.

In a model evaluated by collective survival, frequent false alarms shaped initial caution, the most resistant predator shaped the range over which responses stayed high, and delayed recognition of departure shaped responses that limited excessive stinging. The study provides a way to think about odor-mediated collective behavior through both individual responses and evolution.

The research received funding from organizations including the German Research Foundation, the Austrian Science Fund, the University of Konstanz, research support from the state of Baden-Württemberg, and the Volkswagen Foundation. The authors declared no competing interests.

Reference

López-Incera A, Nouvian M, Ried K, Müller T, Briegel HJ. Honeybee communication during collective defence is shaped by predation. BMC Biology. 2021;19:106. Original paper