Sucking or lapping: facultative feeding mechanisms in honeybees (Apis mellifera)

Abstract

Nectarivorous insects generally adopt suction or lapping to extract nectar from flowers and it is believed that each species exhibits one specific feeding pattern. In recent literature, large groups of nectarivores are classified as either ‘suction feeders', imbibing nectar through their proboscis, or ‘lappers', using viscous dipping. Honeybees (Apis mellifera) are the well-known lappers by virtue of their hairy tongues. Surprisingly, we found that honeybees also employ active suction when feeding on nectar with low viscosity, defying their classification as lappers. Further experiments showed that suction yielded higher uptake rates when ingesting low-concentration nectar, while lapping resulted in faster uptake when ingesting nectar with higher sugar content. We found that the optimal concentration of suction mode in honeybees coincided with the one calculated for other typical suction feeders. Moreover, we found behavioural flexibility in the drinking mode: a bee is able to switch between lapping and suction when offered different nectar concentrations. Such volitional switching in bees can enhance their feeding capabilities, allowing them to efficiently exploit the variety of concentrations presented in floral nectars, enhancing their adaptability to a wide range of energy sources.

1. Introduction

Recent discoveries on natural drinking strategies highlight the need to deepen our understanding of this vital process across diverse taxa [1–4]. Nectarivorous insects regularly consume floral nectar as their principal energy source and they generally employ two distinct feeding techniques, which can be classified as suction or lapping [5]. A single predominant drinking mode is believed to be used by particular species as a specific strategy to feed on nectar, owing to the specific structure of their feeding organs [6]. Suction feeders like butterflies, moths, or orchid bees, suck nectar through their long and thin proboscides [7]. However, those elongated mouthparts do not deal well with high-viscosity nectars which present greater energy rewards [8]. Contrastingly, honeybees consume high-viscosity nectar by lapping it with their hairy tongue (glossa) [9–11]; after retracting the tongue full of nectar they suck it using a pump in their heads [12]. Honeybees prefer to collect nectar with 30–50% sugar content [13], and the nectar concentrations of flowers pollinated by bees (around 35%) are generally higher than the ones of those pollinated by suction feeders (approx. 19–25%) [14]. However, since honeybees also visit flowers with low nectar concentrations [15,16], and they possess comparatively short mouthparts capable of forming a tube encircling the tongue, we aimed to unveil their drinking mechanism at a wide range of nectar concentrations.

Attaching a high-speed camera to a microscope, we showed that honeybees also adopt active suction to imbibe liquid directly, without using lapping. We conducted further experiments to examine the occurrence of suction and lapping patterns across different nectar concentrations. We also fed bees nectars with two different concentrations, with a large viscosity difference, to examine if the feeding mechanism was facultative at the individual level. To figure out the key stimuli of choosing a feeding method (lapping versus sucking), we increased the nectar viscosity while maintaining the sugar concentration constant. We estimated nectar intake rates for different feeding methods and derived the corresponding energy intake rates.

2. Material and methods

(a). High-speed microscopic imaging

We used honeybees (Apis mellifera ligustica) from two outdoor beehives, which were purchased from Jingzhou (Huashanfengye, Hubei, China) and kept at Sun Yat-Sen University, Guangzhou, China (23° N, 113° E). All the bees were starved for 2 h before the experiments. No individuals were used more than once in all experiments. Our experimental set-up consisted of a positioner, a light source (XD-300, Yanan, China), a glass feeder, and a microscope (CX33, Olympus, Japan) coaxially connected to a high-speed camera (VEO E, Phantom, USA) [17]. Twelve bees were held in a truncated conical tube (basal diameter: 10.8 mm; total length: 36.7 mm) that allowed their heads to move freely outside of the tube. The glass feeder was filled with 10 and 30% (wt/wt) sucrose solution mixed with blue dye (Neon brite blue, Chefmaster, USA) for observing sucking and lapping behaviors respectively. When we positioned a bee close to the glass feeder, it started to ingest the sucrose solution and the feeding process was recorded by the high-speed camera at 500 frames per second (fps).

(b). Ingestion modes

To examine the occurrence rates of each drinking mode under different nectar concentrations, we kept honeybees inside a syringe (diameter = 14 mm, length = 60 mm) with the front tube truncated and offered them five different sucrose solutions (10, 20, 30, 40, 50% wt/wt) at around 25°C via a transparent tube (diameter = 2 mm). Bees were free to ingest the artificial nectar and the feeding process was recorded by a high-resolution camera (Canon EOS 6D) at 50 fps. We examined the feeding patterns of 200 bees, which were equally divided into five groups and treated with 10, 20, 30, 40 and 50% (wt/wt) sucrose solutions, respectively. If a bee ingested nectar by reciprocating its tongue, we classified the feeding method as lapping. If a bee imbibed nectar directly with its tongue staying still, we classified the feeding method as suction. The feeding method of each bee was classified based on the predominant feeding mode (above 75% of the nectar consumed) in each observation, since some bees would perform two methods in the same feeding bout.

(c). Behavioral flexibility in feeding mechanism

To further examine the determinants of the voluntary switching between feeding modes at the individual level, we established two groups of 20 bees, and offered to the first group 10 and then 50% nectar, and to the second group first 50 and then 10% nectar. Considering that nectar viscosity increases exponentially with sucrose concentration, we further examined the effect of viscosity on feeding patterns by increasing the viscosity of a 10% wt/wt sucrose solution to the viscosity equivalent of a 50% wt/wt sucrose solution at 25°C using Tylose (HS 30 000 YP, Wuxi Handesen Channl, China), and the added Tylose amount (the viscosities of unadjusted and adjusted nectar are 1.2 and 11.7 mPa s, respectively) was determined by a rotational viscometer (NDJ-5S, MinCe, China). Tylose is a polysaccharide that can greatly change the viscosity of the solution without affecting its density when added in very low quantities [18]. It has no nutritional value for bees and solutions with Tylose are well-accepted by bees [19]. We offered the viscosity-increased nectars to 20 bees and used the high-resolution camera to record the feeding patterns at 50 fps.

(d). Caloric ingestion rate

We measured the nectar intake rate Q (μl s⁻¹) of suction and lapping modes for five concentrations at 10% increments ranging from 10 to 50%. The nectar intake rate was determined by the ingested nectar volume divided by the corresponding ingesting time, both measured from the videos (in Adobe Premiere 2019). The energy intake rate E′ was calculated by E′ = ρscQ/100, where $\rho$ is the density of the nectar, s the sugar concentration and c the energy content per unit mass of sugar [12]. A total of 47 individuals were used in the trial.

3. Results

We found two distinctive drinking patterns in honeybees. Figure 1a shows the suction pattern when a bee ingested sucrose solution of 10%. Its tongue first protruded from the proboscis base at 0 s with the glossal hairs fully deployed, and the liquid level went down while the tongue stayed extended and did not reciprocate (until 8.11 s); this is evidence of active suction. The blue tone at the base of the proboscis evidences liquid moving through the interior of the mouthparts. Figure 1b shows the lapping pattern when a bee ingested sucrose solution of 30%. The tongue was protracted from 0 to 118 ms, and the glossal hairs were erected near the maximum extension. The cycle finished with the total retraction of the tongue at 254 ms. The liquid level went down slowly with these lapping movements (electronic supplementary material, movie S1).

Figure 1.

(a) Four micrographs of a honeybee sucking the artificial nectar. (b) Six micrographs of a honeybee lapping the artificial nectar.

We found that the rates of occurrence of the two distinctive ingestion patterns were significantly different for different sucrose concentrations (${\chi }^2=48.3,p<0.001$, figure 2a; electronic supplementary material, table S1 and movie S2). The occurrence rate of the suction mode reached up to 70.0% when the bees were feeding on nectar with 10% concentration, and the highest occurrence rate of lapping mode was 87.5% when the bees fed on 50% concentration.

Figure 2.

(a) Occurrence rate of the two feeding modes in honeybees, when feeding on sucrose solutions with various concentrations. (b) Rates of switching between feeding modes when offered extreme nectar concentrations. Each encircled number represents a different individual. (c) Nectar intake rates of suction Q_suction and lapping Q_lapping under different concentrations of nectar; dashed line denotes the equivalence point of feeding rate and the corresponding sugar concentration. (d) Energy intake rates of suction E′_suction and lapping E′_lapping for different nectar concentrations: blue dashed line depicts the optimal concentration for suction mode; green dashed line indicates that the optimal concentration for lapping mode is around 50% or above.

The two groups of bees consecutively fed with nectar at the extreme concentrations, first 10 and then 50% or first 50 and then 10%, both showed a significant tendency to switch between the two ingesting modes (figure 2b; electronic supplementary material, table S2), from suction to lapping $({\chi }^2=19.8,p<0.001)$ or from lapping to suction $({\chi }^2=22.6,p<0.001)$, respectively. When feeding on nectar with 10% concentration of sucrose and increased viscosity equivalent to 50% concentration (by adding Tylose), 16 out of 20 tested individuals predominantly exhibited lapping mode, the other four individuals only employed suction. The average nectar intake rate of bees using lapping was 0.71 ± 0.11 µl s⁻¹ (mean ± s.d., n = 16), while the rate of bees using suction was 0.43 ± 0.02 µl s⁻¹ (n = 4).

We found a significant negative association between the nectar intake rate and the sucrose concentration in both suction ($F_{4,16}=43.98,p<0.001$, figure 2c; electronic supplementary material, table S3) and lapping ($F_{4,21}=\hspace{0.17em}12.82,p<0.001$). The theoretical energy intake rate of the suction mode is higher than that of lapping when the concentration is below 25–30%, and the reverse is true when the concentration is 30% or above (figure 2d; electronic supplementary material, table S4). The concentration bringing the highest energy reward for the suction mode was around 30% but the optimum for the lapping mode is predicted to be at 50% or higher.

4. Discussion

Our study provides the first explicit evidence of suction drinking in honeybees. They tended to employ suction more than lapping when the nectar concentration was lower than 30%. Conversely, lapping was predominantly used in sucrose solutions of 30% concentration or higher. When we offered nectar to bees at first 10 and then 50% concentration, they presented a strong tendency to switch their initial suction mode to lapping. Inversely, if the offered nectar was first 50 and then 10%, most bees switched their initial lapping pattern to suction. Moreover, most bees exhibited lapping instead of suction when the offered nectar had 10% sugar concentration but viscosity equivalent to nectar with 50% sugar concentration (with added Tylose), and the average ingesting rate of the lapping bees was in accordance with the rate of bees lapping 50% nectar. We thus suggest that the key stimulus to switch feeding modes is viscosity, rather than sugar content.

The nectar intake rate of lapping did not change much between 10 and 30% concentration, and it decreased with a shallow slope between 40 and 50% (figure 2c), which could be a consequence of the tongue being filled up to its maximum volume at a low concentration and the reduction in volume per lick at higher concentrations [20]. In contrast to lapping, the nectar intake rate for suction strongly decreased with increasing concentration from 10 to 50% (figure 2c); however, the nectar intake rate for suction was higher than for lapping when the concentration was below 30%, the point at which we also registered the switch to lapping as the predominant feeding mechanism employed by bees (figure 2a).

Among various nectarivorous insects classified as ‘suction feeders', the decrease in nectar intake rates across concentrations generally present even steeper slopes than the one presented here, owing to the effects of exponentially increasing viscosity and the morphology of longer mouthparts with channels of smaller radius [6,21]. Moreover, the concentrations for optimal energy rewards that have been calculated for suction feeders are generally much lower than those calculated for lapping feeders [5,21]. Our study illustrates how two drinking mechanisms are used facultatively by the same species and that they follow the same trends that have been proposed for other nectarivores (optimality achieved at lower concentrations for suction than for lapping), because they follow the same physical rules. Furthermore, we verified the prediction of lapping being less dependent on fluid viscosity than suction [5]. The concentration at which honeybees preferred to use suction coincides with the average nectar concentration of flowers pollinated by most suction feeders, such as hawkmoths (19%) and butterflies (25%) [21]. Moreover, the concentration (around 30%) bringing the greatest energy rewards in suction pattern of honeybees falls within the optimal range of suction feeders [21], such as ants [22], butterflies [23], hawkmoths [18] and orchid bees [24]. The present study adds a new perspective to the idea that the sugar concentration in the nectar offered by the plants that use animals as pollinators may have coevolved with the specific drinking technique employed by those animals [21].

The capillary-based lapping mechanism that allows honeybees to achieve high energy intake rates when feeding on highly concentrated nectar also yields slower uptake than suction when feeding on less viscous liquids. Therefore, exhibiting facultative use of drinking mechanisms offers bees the possibility to exploit a larger spectrum of resources more efficiently. In fact, bee colonies can focus on rich sources when they are abundant, but would spread their workers to a wider range of sources when the rewards are scarce [16]. A wide spectrum of flowers containing comparatively low concentration nectars are also visited by honeybees [15,16,25] and the resulting nectar mixture in their crops after foraging might represent an average viscosity and energy values among nectars of all of the plant species visited [26,27]. Bees would also alter their foraging choices to sources with diluted nectar when accessing the same quantity of sweeter nectar requires extra efforts that do not compensate for the additional energy investment [27]. Moreover, some of the workers would directly forage for water to bring back to the beehive and use in thermoregulation on hot days, and for the dilution of mature honey on cool days when the nectar sources are scarce [28,29]. Thus the efficient collection of less viscous liquids is of great importance for their communal living. Our findings of a behavioral switching between predominant feeding modes in honeybees match their resulting energy intake rates across concentrations. This facultative use of drinking methods behaviorally adjusted to fluid viscosity has potentially enhanced the adaptability of honeybees to a wider range of resources.

Data accessibility

Data are available as electronic supplementary material.

Authors' contributions

J. Wei conceptualization, data curation, formal analysis, investigation, methodology, writing original draft. Z.H. data curation, formal analysis, investigation, writing original draft. S.N.G. methodology, writing review and editing. A.R.-G. funding acquisition, methodology, writing review and editing. Z.W. writing review and editing, supervision. J. Wu funding acquisition, methodology, project administration, resources, supervision, writing original draft. All authors approve the final version of the manuscript and agree to be held accountable for the content therein.

Competing interests

We declare we have no competing interests.

Funding

This work was supported by the research grant of Sun Yat-Sen University for Bairen Plan (grant no. 76200-18841223), and the National Natural Science Foundation of China (grant no. 51905556). A.R-G. is supported by the Walt Halperin Endowed Professorship and the Washington Research Foundation as Distinguished Investigator.