Stop signaling by foragers not encountering danger at a food source
The stop signal is an acoustic negative feedback signal employed by honey bees that plays a key role in foraging regulation and group decision-making. In this study, foragers trained to visit a feeding station were either pinched on the hind femur by an observer in a simulation of a predator attack or left to forage uninhibited. Pinched bees were significantly less likely to perform the waggle dance, a recruitment signal, upon their return to their hive and also significantly more likely to perform the stop signal than bees that had not been pinched. Of note is the observation that approximately 70% of the stop signals recorded during the course of this experiment were performed by bees that had never visited the feeding station at all. This may have occurred because the bees interpreted the presence of a high-reward food source, the feeding station, as a honey-robbing situation. Alternatively, the stop signaling bees may have been unloader bees responsible for taking food from incoming foragers and storing it elsewhere in the hive. These unloader bees may have been overwhelmed by the influx of food from the feeding station and used the stop signal in an effort to modulate the colony’s foraging efforts.
KeywordsBehavior Decision-making Honey bee Negative feedback Recruitment
Honey bees (Apis mellifera L.) regulate their colony’s foraging through a repertoire of signals. The best known of these is the waggle dance, which is used to communicate the distance and direction to food and other resources (von Frisch 1967) and is part of a suite of other signals involved in recruitment (see Kietzman and Visscher 2015). One such signal is the stop signal, a brief acoustic signal known as a negative feedback signal, because it has been shown to cause bees to stop waggle dancing significantly more often than could be ascribed to chance alone (Kirchner 1993; Nieh 1993; Pastor and Seeley 2005). Using the stop signal, bees that have had a negative experience at a food source can decrease waggle dances advertising that location and consequently inhibit foraging (Lau and Nieh 2010; Nieh 2010). This note reports unexpected results found during an experiment employing a method similar to that used by Nieh (2010) of pinching foragers at a feeding station to simulate a predator attack.
We conducted this experiment in January 2013 at UC Riverside’s Agricultural Operations, a research farm adjacent to the main campus in Riverside, California. We established a two-frame observation hive on 15 January. Beginning on 17 January, we trained the bees to visit a feeding station containing a 2 M solution of sucrose lightly scented with anise. Its final distance from the hive was 100 m. Data collection occurred on 21–22 January. Conditions were sunny and breezy with temperatures from 19–24 °C. Data on the first day was collected from 09.30 h to approximately 16.30 h, and from 10.30 to 14.30 h on the second day.
Bees in the control group were gently caught in a small insect net at the feeding station (by JKL on Day 1 and PKV on Day 2), marked on the thorax with a paint pen, and released. Those in the treatment group were additionally squeezed with forceps on the rear femur and marked with a different color of paint. Treatments were randomized in pairs to avoid temporal bias and to ensure equal numbers (n = 64) in each group. Some bees may have visited the feeder and left before they could be marked; however, much care was taken to ensure that this number was as small as possible by using a feeder that could only be accessed by a small number of bees at a time. Any marked bees that returned to the feeder were treated correspondingly with their color, i.e., bees that had previously been pinched were pinched again and bees that had not been pinched were left untouched.
The number of waggle dances and stop signals performed by three groups of bees
Waggle dances (n)
Mean duration (m:s)
Standard deviation (s)
Stop signals (n)
Not pinched (control)
Unmarked (no feeder visit)
Using Z tests we assessed the number of waggle dances performed by bees in each of the treatment groups, the mean length of time spent dancing by bees in each of the treatment groups, the number of stop signals performed by bees of each group, and the number of stop signals received by dancers of each group. Overall, significantly fewer pinched than unpinched marked bees performed dances (Z test: Z = 11.5047, p < 0.0001). Furthermore, the mean dance length was significantly longer among bees that were not pinched (Z test: Z < −1000, p < 0.0001). A significantly higher proportion of dances were performed by unpinched bees on Day 1 (2-proportion Z test: Z = −3.0977, p = 0.002). No bees in the treatment group performed waggle dances on Day 2, but bees in the control and treatment groups performed stop signals on both days. There was no significant difference between days in the proportion stop signals performed by the two groups (2-proportion Z test: Z = −0.8323, p = 0.4052). We found no significant difference in the number of stop signals received by bees of either treatment group (Z test: Z = 1.1992, p = 0.2305).
Bees in the treatment group did not perform waggle dances on Day 2. This may have resulted from a difference in the researcher performing the treatment or the smaller data set because of the shorter collection time. The bees in the treatment group produced 21% of the total number of stop signals recorded, while bees that were part of the control group produced only 8%. This was a significant difference in proportions (2-proportion Z test: Z = −4.149, p < 0.0001). However, about 70% of the stop signals were performed by unmarked bees. These bees probably did not visit our feeder (very few unmarked bees may have, but not been marked). Compared to the percentage of stop signals produced by marked bees, there were significantly more stop signals performed by bees that were not part of either experimental group (2-proportion Z test: Z = −8.6162, p ≪ 0.0001).
In this analysis we assumed independence of events. In reality, the bees were not uniquely marked, so it was impossible to determine whether a bee was a repeat dancer or if a bee performed multiple stop signals. For the purposes of this study, we chose to assess the dances and stop signals as independent events, because bees that returned to the feeding station were re-treated according to their treatment group, thereby “refreshing” their experience at the feeder.
Our observations confirm the results of Nieh (2010): a statistically significant higher proportion of pinched bees used the stop signal upon their arrival to the colony than bees that had not been pinched. Our finding that 70% of stop signaling bees were unmarked and therefore unlikely to have visited the feeder was unexpected, and has several possible explanations. One is that the unmarked bees may have been foraging at some other location and were not unloaded promptly due to the influx of food from the feeding station. Stop signaling commonly increases when a feeding station is present (Nieh 1993; Thom et al. 2003; Pastor and Seeley 2005), and most of the stop signaling comes from tremble dancers (Thom et al. 2003). Tremble dancing occurs when foragers are not unloaded quickly by other bees (Seeley 1992). Thus, if the unloader bees were overwhelmed by the amount of food brought in by bees visiting the feeding station they would not have been available to unload bees that had foraged at other locations. Alternatively, the unmarked bees may have been unloader bees attempting to inhibit foraging due to an unmanageable number of foragers. To date the stop signal has not been observed in use in this capacity; however, a study on this area would be highly relevant and likely yield fruitful results.
A further explanation pertains to the idea that having unfettered access to a rich food source such as a feeding station is more similar to a honey-robbing situation than to foraging on natural food sources such as flowers, which offer a small nectar reward and are located across a patchy landscape. Johnson and Nieh (2010) showed that the stop signal might be used to rapidly shut down a robbing event by countering the waggle dance. The ability to quickly stop a robbing event would be important if the robbed colony were very strong and an excessively high number of robbers were killed by guard bees (Johnson and Nieh 2010). It is possible that the pinched bees that visited the feeding station may have been emitting alarm pheromone, and that as a result other bees in the colony interpreted the presence of the feeding station as an unfavorable robbing situation and started using the stop signal accordingly. The stop signal is a versatile communication tool and it is clear that all of its uses and meanings have not yet been described.
- von Frisch K (1967) The dance language and orientation of bees. Transl. by L. E. Chadwick. Harvard University Press, CambridgeGoogle Scholar