1 Introduction

Honey bees (Apis mellifera) are eusocial insects. A honey bee colony is characterized by division of labor and the distinction of reproductive (queens and drones) and sterile (workers) individuals. Being aware of the life cycle and society structure of honey bees is fundamental to understand the complex communication system they have developed. It is known that the workers are fundamental for the conduction of the colony (Sagakami 1958; Seeley 1982, 1985; Pankiw et al. 1998; Schmickl and Crailsheim 2004); the queen is designed to lay eggs in order to maintain a continuous generational population renewal (Jay 1970; Wanner et al. 2007; Slater et al. 2020; Mao et al. 2024) and the drones play an important role in the spreading of the family genome (Hrassnigg and Crailsheim 2005; Neubauer et al. 2023; Zarić et al. 2024). Honey bees live in a nest constituted of multiple combs situated in a cavity. In the darkness of the nest, communication does not rely on visual signals, but on chemical (Jay 1970; Pankiw et al. 1998; Slessor et al. 2005; Wanner et al. 2007), tactile (Esch 1967; Lindauer 1955; Tautz and Rohrseitz 1998), electric (Greggers et al. 2013) and mechanical signals (airborne sounds, Michelsen et al. 1987, Kirchner et al. 1991; substrate borne vibrations, Kirchner et al. 2022).

Vibrational communication is defined as the transfer of signals and/or cues, which convey information, through a substrate, from a sender to a receiver (Hill 2008), and is extensively used by workers and queens. The vibrational signals are detected through the subgenual organ, a chordotonal organ located in the tibia of their legs (Autrum and Schneider 1948). It has been studied as far as now exclusively in workers.

1.1 Workers

Workers make use of vibrational signals in various context and as modulator of other bees’ activity (stop signal, dorso-ventral-abdominal vibration (DVAV), and low-frequency vibrations). The amplitudes of these signals can be low as 0.2–0.5 m/s2 (DVAV) or high as ~ 6 m/s2 (stop signal), with frequencies spanning from around 20 Hz (DVAV and low-frequency vibrations) to 320 Hz (stop signal) to high frequencies as 500–800 Hz (DVAV) (Michelsen et al. 1986a; Ramsey et al. 2018; Hrncir et al. 2019; Kirchner et al. 2022). Kilpinen and Storm (1997) found in workers a higher sensitivity for the vertical component of the vibration with an electrophysiological threshold of 0.06–0.15 mm/s pp at 300 Hz (~ 0.1 m/s2), while behavioral experiments conducted in the hive showed a vibrational threshold of the freezing response at about 4 m/s2 at 200 Hz (Rohrseitz and Kilpinen 1997) and about 5 m/s2 at 300 Hz (Michelsen et al. 1986a). The freezing response consists of an immediate stop of all the bees detecting a substrate borne vibration of a sufficient amplitude (Michelsen et al. 1986a).

1.2 Queens

If for workers, vibrational communication is necessary for an everyday efficient work, for honey bee queens, vibrational signals are essential for their survival. Indeed, during the swarming period, a deadly fight among newly hatched virgin queens take place for the supremacy in the colony. The first virgin queen who emerges from the royal cell emits what is known as tooting, to which the queens which are still enclosed in the royal cells respond with another signal called quacking. The two signals are clearly audible to humans and can be distinguished for frequency (tooting ~ 300–500 Hz, quacking ~ 200–300 Hz; Michelsen et al. 1986b) and temporal structure (tooting is characterized by a first longer syllable followed by shorter ones, while quacking is made of a series of very short syllables; Michelsen et al. 1986b); however, bees detect them as substrate borne vibrations (Wenner 1962; Simpson 1964). Despite the role of these two signals is still under debate, it is of interest to notice how relevant it is for any of the involved queens to be able to detect any of the other rivals´ signals to prevail and not to be killed and, in case, to evaluate the possible competitors (Visscher 1993; Kirchner 1993). It is possible that also workers would take advantage of this exchange of signals to be aware of the presence of a queen (or spare queens) and/or manipulate the queen’s actions toward the royal cells or toward other hatched queens during tournaments (Gilley 2001). It is known that not only workers but also enclosed queens freeze (freezing response) when an emerged queen toots. Virgin queens may exploit this freezing to keep workers away and protect themselves and delay the emergence of rival queens (Michelsen et al. 1986b; Grooters 1987). This competition among queens lasts until one queen kills all the rivals and remains as the new indisputable queen of the colony. At this point, the queen mates and begins to lay eggs. A laying queens is still able to toot (Woods 1950; Allen 1956); however, the need of tooting and the risk for her life is lower, since she does not have competitors anymore and her daughters protect, feed and attend her everywhere (van der Blom 1992).

1.3 Drones

Up to date, there are no data describing vibrational signals generated by drones. However, it has been found that workers vibrate drones, especially when sexually immature, before their mating flight (Boucher and Schneider 2009). The consequent effect consists in a higher activation of drones and more time spent in trophallaxis and grooming activities with workers. Slone et al. (2012) also observed that drones with lower thorax weight and smaller thorax-body weight ratio (which may affect flight ability and mating success) tended to receive much more vibrations, hypothesizing higher care for undeveloped drones in view of the competitive mating flight. It was suggested that these vibrational signals are promoting the sexual development of drones. Indeed, trophallaxis provides nutrients, which are essential for the development and for sexual maturation of the drone, but also for maintaining mature drones heathy. Specifically, proteins are necessary for flight muscles’ and sexual organs’ development (Slone et al. 2012). Grooming is a form of caring and one of the first activities promoted by workers toward drones after they hatch (Ohtani 1974) and the authors suggest that it may contribute to the physical conditions of the drone.

1.4 Aim of the study

Based on this knowledge, we decided to conduct a behavioral experiment that could discern differences in the vibrational sensitivity among the castes, for sex-dependence and relatively to their role in honey bee society. We conducted the experiment in the frequency range used by emerged queens (350–500 Hz), since we hypothesized virgin queens to be much more sensitive than laying queens. We also predicted workers to be sensitive to this frequency range. Indeed, workers are sensitive to a wide range of frequencies (which definitely includes the frequencies chosen for this study; Michelsen et al. 1986a) and use different frequencies (low and high) for communication (Kirchner et al. 2022). On the contrary, this work constitutes the first study about vibrational sensitivity in drones. We hypothesized workers to be also highly sensitive, taking into account their considerable daily vibrational communication.

2 Materials and methods

For the behavioral experiment the bees were divided in four groups: workers, virgin queens, laying queens and drones. Queens were taken from a total of 28 nucleus colonies; workers and drones were taken from 4 different colonies. All groups were tested to assess the vibrational sensitivity threshold and workers were in addition tested to estimate the threshold of the freezing response. We observed the presence of any behavioral reaction (or freezing) by bees placed inside an arena to a pseudorandomized sequence of real stimuli (audio for the experimenter and vibration for the bee both present) and sham stimuli (audio for the experimenter present and vibration for the bee absent) at different amplitudes and frequencies, until a threshold was found. It was defined as reaction any change in behavior in respect to the baseline (without external stimuli). The threshold was defined as the lowest tested amplitude [m/s2] at which significantly more behavioral reactions were shown to real stimuli than to sham stimuli. The chosen frequencies were 350, 400, 450, and 500 Hz, which are the most used by honey bee queens for vibrational communication. The stimuli were created with a waveform generator (RIGOL DG1022) and the intensity of the signal was adjusted through an amplifier (FeinTech AVS00100) placed between the generator and a vibrational exciter, onto which the arena was placed. Vibrational amplitudes of signals were measured using an LDV-calibrated accelerometer (KD37 MMF). The signals were amplified using an IEPE100 (MMF) preamplifier and M29 IEPE sensor supply (MMF) and visualized on an oscilloscope (Tektronix TDS1001B). The observer wore headphones (connected to the wave generator) not to be influenced by the possible sound of the exciter vibrating and focusing only on the visual detection of behavioral reactions (the volume could also be adjusted through another FeinTech AVS00100 amplifier). A switch (Speaka professional), also connected to the wave generator, was used to alternate real stimuli and sham stimuli, while always maintaining the sound in the headphones. The duration of the interval between the stimuli was of 1 min.

For every series of trials of the same amplitude, a Chi-square-four-field test was performed to assess if the probability of the bee to react when the stimulus was present was higher than when absent. If the test was significant (χ2 > 3.84, p < 0.05), the amplitude was decreased; if not significant, the amplitude was increased. This procedure was repeated until a threshold was found. Each queen (both virgin and laying) and each drone were tested along with 14 trials for every amplitude. The workers were tested along with 30 trials. The difference in number of trials is due to the resistance of the various groups to vibrations and experimental conditions: queens and drones tended to get stressed much faster than workers. The subjects of drones, virgin queens, and laying queens were tested alone in the petri-dish, while workers were always paired. This choice was dictated by the sociality of honey bees: the presence of a companion made workers more reassured inside the tiny space of a petri-dish and allowed to perform the experiment. The experiment was run in a completed dark and isolated room, in order to exclude any distraction for the bees, with the only presence of a red light to permit to the experimenter to see (the red wavelength is out from the bee visual spectrum). The animals were collected from colonies situated in our university botanical garden and gently chilled with ice to position them inside the petri-dish and then transferred into the dark room. Afterward, a time span of 30–45 min was used to let the animals familiarize and acclimatize with the new environment, before the experiment was started (Figure 1). The same group of queens was tested for each frequency: 15 virgin queens and 13 laying queens, respectively. Conversely, different groups of 20 workers and 30 drones were tested for every frequency. In addition, different groups of 23 workers were tested for their freezing response for every frequency. The virgin queen age spanned from 2 to 5 days old. The age of laying queens instead spanned between 2 months old and 2 years old. The age of workers and drones is unknown, since they were taken randomly from the colonies.

Figure 1.
figure 1

Two workers in the petri-dish mounted on a vibrational exciter. Observations were made under dim red light.

Full size image

The mean thresholds for every group and frequency were calculated and were successively compared with one another using U-test (for independent groups) to highlight any significant difference in sensitivity. The U-test was run using Matlab.

3 Results

Honey bees reacted behaviorally with an initial freezing response at the highest amplitudes and then with a stand-by antenna position or with a lowering of the antenna toward the ground, when the amplitudes were getting lower. The lowering of the antennas was particularly evident in drones. The vibrational sensitivity was higher for all the groups at 500 Hz (Figure 2). The lowest thresholds found in virgin queens and in workers at 500 Hz were 0.25 m/s2 pp. At all frequencies, workers and virgin queens were significantly more sensitive than laying queens and drones (U-test, p < 0.01). At 350 Hz and 400 Hz, laying queens were in addition significantly more sensitive than drones (U-test, p < 0.01), and at 350 Hz, virgin queens were in addition slightly but significantly more sensitive than workers (U-test, p < 0.01). In workers and drones, there was a significant difference in sensitivity between the lowest and the highest frequency (350 and 500 Hz; U-test, p < 0.01). Virgin queens present a constant threshold along all frequencies (U-test, p > 0.01). Details of the analysis and significant differences (p-values) among the four groups can be visualized in Table I in the Appendix.

Figure 2.
figure 2

Mean vibrational sensitivity of drones, laying queens, workers, and virgin queens along the four tested frequencies (350 Hz, 400 Hz, 450 Hz, and 500 Hz). Virgin queens and workers are the most sensitive. All groups show maximal sensitivity at 500 Hz. Standard error bars indicate variability within samples. Sample size: 13 laying queens, 15 virgin queens, 20*4 workers, and 30*4 drones. Difference between subgroups is pairwise significantly different on the 1% level, when labels are not overlapping (U-test; see also Table I in Appendix).

Full size image

The threshold of the freezing response of workers, tested in a petri-dish, was found constant along the four frequencies at 0.5 m/s2 pp (U-test; p > 0.01) and slightly but significantly higher than their threshold of general behavioral responses (U-test, p < 0.01 at all frequencies: p = 3.8*10¯5 at 350 Hz, 2.9*10¯⁹ at 400 Hz, 1.4*10¯⁸ at 450 Hz, and 4.7*10¯⁹ at 500 Hz).

4 Discussion

We can conclude that caste and sex have effects on the vibrational sensitivity in honey bees and that they may reflect the role that each group cover within the colony life and colony cycle.

4.1 Queens

A clear difference in sensitivity exist between virgin and laying queens, with virgin queens being significantly more sensitive than laying queens. The vibrational sensitivity of laying queens become more and more finest at the highest frequencies, while the sensitivity of virgin queens is almost constant along the four tested frequencies. This could potentially reflect the increasing of tooting’s frequency with age (Michelsen et al. 1986b). Indeed, at a young age and in the context of swarming, a sharp sensitivity is crucial for survival and queens need to be sensitive to all frequencies. With aging and with the queen becoming the only queen of the colony, the sensitivity would tune in turn more to the high frequencies than the low frequencies. It could be interesting to monitor the temporal change of the threshold of vibrational sensitivity in young queens and assess if it starts with fertilization or depends on age.

4.2 Workers

The high sensitivity for the workers is equivalent to efficient everyday work. Their sensitivity is similar to the one of virgin queens. Indeed, vibrational communication is at the base of the relations among these two groups. The vibrational sensitivity of workers turned out to be higher than what reported in previous behavioral studies run in the hive (Kirchner et al. 2022), which were measuring the freezing threshold. This is understandable, since the hive, differently from an isolated arena, is characterized by a lot of background noise, that may consequently increase the threshold. Indeed, the freezing is a response of the bee to substrate borne vibrations at a certain threshold (Michelsen et al. 1986a), which does not exclude that bees may be more sensitive to substrate borne vibrations than this. In the current study, the freezing response threshold is clearly higher than the general vibrational sensitivity threshold and this finest vibrational sensitivity is represented by reactions as the stand by position and the lowering of the antenna.

The freezing threshold can be expected to be higher than the general vibrational sensitivity also for the other groups. Moreover, specifically for laying queens and drones, the freezing threshold can be presumed also higher than the worker freezing threshold, while for virgin queen quite similar. At 350 Hz, the general vibrational sensitivity is slightly lower than the other frequencies. Considering that DVAV signals, which are above 500 Hz, can have amplitude as low as 0.2–0.5 m/s2 and other signals below 500 Hz (stop signal, low-frequency vibrations) have generally much higher amplitude, it could be deducted that workers possess higher vibrational sensitivity at the higher frequencies. This intuition is coherent with the previous results (Kirchner et al. 2022), which found a flat freezing threshold curve [m/s] for frequencies above 300–400 Hz, suggesting that honey bees’ receptors are tuned to detect the velocity component of the signal. It is important to underline that the vibrational threshold, both behavioral and electrophysiological, also depends on the position and inclination of the bee’s legs on the substrate (Sandeman et al. 1996; Kilpinen and Storm 1997; Strauß et al. 2019), since bees are more sensitive to the vertical component of the vibration in respect to the horizontal component (Rohrseitz and Kilpinen 1997).

4.3 Drones

This is the first study exploring vibrational sensitivity in drones. Drones resulted to be the least sensitive among the tested groups for the chosen frequency range. There is an abrupt escalate in sensitivity between 400 and 450 Hz, where the sensitivity of drones increases and becomes similar to the one of laying queens. This suggests that also drones, as workers and laying queens, may generally be more sensitive to the higher frequencies. In support of this, for example, drones can be recipient of DVAV, which amplitude is very low and the signal is characterized of high frequency (Ramsey et al. 2018). Overall, however, these results reflect the role of drones in the colony, where they are mainly present for reproductive purposes. Drones spend their youth inside the nest, without essential role, until sexual maturity, when they leave to reach aggregation arenas and mate. Therefore, a fine vibrational sensitivity does not appear to be relevant. Further studies could examine the vibrational sensitivity of drones with different genetic characteristics: diploid versus haploid drones generated by queens and haploid drones generated by workers. And again, sexually immature drones, vibrated by workers, with sexually mature drones.

4.4 General conclusions

The current study shows that vibrational sensitivity diverges among groups of different caste and sex in eusocial insects as honey bees. It displays a general trend of higher sensitivity for high frequencies in respect to low frequencies, with the exception of virgin queens, which require a sharp sensitivity for all frequencies for survival. Virgin queens, together with workers, result to be the most sensitive to substrate borne vibrations. Confirming our hypothesis, evident is the difference in sensitivity between virgin and laying queens.

We do not know yet, in as much environmental factors affect the vibrational sensitivity in addition to sex and caste. A recent study by Farina (2023) using microphone recordings of the background sound inside beehives indicates that colony activity and external factors have effects on the background noise level. It would be interesting to see whether the sensitivity to substrate vibrations is adjusted accordingly.