ANIMAL BEHAVIOUR, 2000, 59, 1177–1185 doi:10.1006/anbe.1999.1418, available online at http://www.idealibrary.com on
Trophallaxis in the honeybee Apis mellifera (L.): the interaction between flow of solution and sucrose concentration of the exploited food sources ALEJANDRO J. WAINSELBOIM & WALTER M. FARINA
Departamento de Ciencias Biolo´gicas, Facultad de Ciencias Exactas y Naturales, Universidad de Buenos Aires (Received 16 June 1999; initial acceptance 2 September 1999; final acceptance 16 December 1999; MS. number: A8521)
Forager bees arriving at the hive after visiting a nectar source, unload the collected liquid food to recipient hivemates through mouth-to-mouth contact (trophallaxis). We analysed whether the main characteristics that define nectar in energetic terms, that is, rate of production (flow of solution), sucrose concentration and rate of sucrose production (sucrose flow) influence trophallactic behaviour. Individual bees trained to feed at a regulated-flow feeder offering sucrose solution were captured once the foraging visit was complete and placed in an acrylic arena with a recipient bee that had not been fed. The rate at which liquid was transferred during the subsequent trophallactic contact (transfer rate) was analysed as a function of the different solution flows and sucrose concentrations offered at the feeder. A relationship was found between transfer rate during trophallaxis and the flow of solution previously presented at the feeder. This relationship was independent of sucrose concentration when above a certain threshold value (ca. 22% weight on weight). We also analysed whether the rate of sucrose deliverance of the food source (sucrose flow) influenced the rate at which the solution was transferred. No clear relationship was found between the rate of sucrose deliverance during trophallactic events (sucrose transfer rate) and the sucrose flow presented at the feeder. The possibility that trophallaxis could be a communication channel through which quantitative information on food source profitability is transmitted among hivemates is discussed.
1983). In the case of short-tongued insects, viscosity does not influence the dynamics of nectar collection in such a high degree, and therefore higher sugar concentrations are favoured. Additionally ambient conditions such as temperature and humidity may strongly affect nectar characteristics even within the same day (Nu´n ˜ ez 1977a; Vogel 1983; Kearns & Inouye 1993). These factors not only influence the concentration of the saccharide components of nectar, but also its flow. For instance, water stress greatly reduces the rate of nectar production (Fahn 1949), while increasing temperatures favour it (Huber 1956). Due to these ambient factors, the most relevant variable to represent the profitability of an exploited nectar source is the energy yielded over time, represented by the sugar flow (i.e. the mass of sugar delivered per unit time, independently of the water content of the nectar). One of the most common insect pollinators are honeybees, Apis mellifera, which regularly forage on nectar in order to fulfil the nutritional requirements of the colony. Honeybee foragers must adapt their behaviour to these continuous changes in nectar availability and quality to achieve an energetically efficient gathering. Nectar
Nectars collected by insects are composed by different mono- and oligosaccharides as major constituents (Kearns & Inouye 1993). The concentration of the saccharide nectar components (Wolf et al. 1984) and the rate of nectar delivery depend on factors such as plant age, species and environmental conditions (Vogel 1983; Kearns & Inouye 1993). Nectar quality and quantity along with flower morphology influence the type of pollinator visiting each plant species. For example, whereas birds, bats and hawk moths visit mostly large flowers that offer relatively large volumes of low sugar concentration nectar, honeybees and bumblebees prefer smaller nectaries with solutions of higher sugar concentrations (Vogel 1983). These species-specific preferences are also a consequence of the different anatomical characteristics involved in nectar ingestion. Thus, nectarfeeding insects with long proboscises will prefer diluted nectars of low viscosity that reduce the effort required in their ingestion (e.g. hawk moths, butterflies; Heyneman Correspondence: W. M. Farina, Departamento de Ciencias Biolo´gicas, Facultad de Ciencias Exactas y Naturales, Universidad de Buenos Aires, Pabello´n. II, Ciudad Universitaria, 1428, Buenos Aires, Argentina (email:
[email protected]). 0003–3472/00/061177+09 $35.00/0
2000 The Association for the Study of Animal Behaviour
1177
2000 The Association for the Study of Animal Behaviour
1178
ANIMAL BEHAVIOUR, 59, 6
foraging behaviour by honeybees has been extensively studied in the past decades, leading to the discovery that different aspects of the foraging cycle, both at the food source and at the hive, are modified by the characteristics of the nectar collected by the bee. Nu´ n ˜ ez (1966), working with regulated flow-rate feeders, found that foraging behaviour at the food source changed markedly depending on the offered flow of solution. In this way he was able to characterize four different ranges for a 50% weight on weight (w/w) sugar solution. (1) Flows above 60–65 l/ min exceed the maximal intake rate of bees, being therefore unlimited (Nu´ n ˜ ez 1966). Under these conditions, bees attained maximal crop loads in approximately 1 min. (2) For flows between 10 and 60–65 l/min, foragers attained maximal crop loads with different crop filling times depending on the flow of solution of the food source (Nu´ n ˜ ez 1971). (3) Between approximately 1 and 8–10 l/min, there was a positive relationship between crop load and delivery rate (Nu´ n ˜ ez 1966). (4) This same dependency was found for values below 1 l/min, but in this case there was an additional negative dependency between flow of solution and the duration of the subsequent foraging pause (Grosclaude & Nu´ n ˜ ez 1998). Changing the flow of a solution of a given sucrose concentration affects the rate at which sucrose itself is delivered by the food source (i.e. its sucrose flow). Thus, data can be analysed as a function of the offered flow of solution and as a function of the flow of sucrose. Nu´ n ˜ ez (1966), for instance, described that crop load values of foraging bees were influenced not only by the offered flow of solution, but by sugar concentration values as well. Thus, bees exploiting a particular flow increased their final crop load when a higher sucrose concentration was offered. Nevertheless, when conditions at the food source were described in terms of sucrose flow, final crop load values fitted a single curve, independently of solution flow or sucrose concentration (Varju´ & Nu´ n ˜ ez 1991). Thus, energy deliverance, which in the case of nectar is mostly due to the amount of saccharides rendered per unit time by the source, affects the attractiveness of the food site. Similarly, Greggers et al. (1993), working with an array of four simultaneous feeders, found that the relative choice frequency to a particular feeder depended on the sucrose concentration and the flow of the delivered solution. Thus, feeders had significantly higher probabilities of being visited when delivering higher sucrose concentrations or high flow. Yet probabilities were similar when feeders presented the same sucrose flow, irrespective of sucrose concentration or flow of solution. These results suggest the possibility that honeybees can detect sucrose flow and determine the number of sucrose molecules passing through the digestive system over a certain period of time. When returning from a successful bout, foragers transfer the collected food stored in their crops to recipient hivemates through trophallactic contacts (i.e. mouth-tomouth food exchanges). This food unloading is sometimes associated with recruitment behaviour such as dance manoeuvres, and recruits nestmates to known food sources (von Frisch 1923, 1968). Given that the profitability of the food source, defined mainly by sucrose
concentration and flow of solution, exerts a strong influence on foraging behaviour and dance manoeuvres (von Frisch 1965; Nu´ n ˜ ez 1966, 1970; Waddington & Kirchner 1992; Farina 1996), is trophallactic behaviour also modified by these parameters? To address this question, Farina & Nu´ n ˜ ez (1993), working with pairs of bees within an experimental arena, simulated a flow of 1 to 8–10 l/min by feeding donor bees different amounts of a 50% w/w sucrose solution through a graduated capillary tube. The latency to initiate an unloading contact (i.e. transfer delay) decreased when donor bees ingested higher amounts of solution and the number of unloading events increased with increasing crop load (Farina & Nu´ n ˜ ez 1993). Additionally, there was a strong correlation between the amount of solution ingested and the rate at which the solution was transferred from donor to recipient during the subsequent trophallactic encounter, henceforth: transfer rate (Farina & Nu´ n ˜ ez 1991). When bees were fed a fixed volume of solution (ca. 55 l), transfer rate increased with increasing sucrose concentration, reaching a maximum for 30% w/w solutions (Farina & Nu´ n ˜ ez 1991). Yet the sucrose transfer rate, that is, the speed at which sucrose was delivered during a trophallactic contact, showed a linear increase up to the 50% w/w concentration, the maximum value employed in the experiment. A similar result was found when the viscosity of the solution was kept constant (Tezze & Farina 1999). Moreover, the amount of sucrose delivered during trophallactic contacts was related to the mass of sucrose carried in the donor bee’s crop, regardless of whether the donor bee attained its crop load from varying amounts of a fixed sucrose concentration or a fixed quantity of varying concentration. Because the donor bees were fed through a graduated capillary, they attained a maximal intake rate while collecting the solution, and therefore the modulation of transfer rate with the sucrose concentration of the solution suggests that intake rate actually increased with sucrose flow, because with increasing sucrose concentration more sucrose molecules were ingested per unit time. In the present study, we examined whether changes in the solution intake rate (i.e. flow of solution) of the donor bees influence trophallactic behaviour, and if so, how sucrose concentration and flow of solution interact when manipulated independently. Additionally, by manipulating these two factors, we sought to determine the possible role of different sucrose flows on trophallactic behaviour. METHODS This study was carried out from February to May 1996 and 1997 (at the end of the nectar season), in the experimental field of the Faculty of Exact and Natural Sciences (3432 S, 5826 W), University of Buenos Aires.
Experimental Procedure We trained two groups of honeybees from a hive located 50 m away from the laboratory to feed at artificial feeders placed at two sites ca. 30 m apart. The first feeder
WAINSELBOIM & FARINA: TROPHALLAXIS IN HONEYBEES
(site 1) consisted of an acrylic syringe driven by a pump by means of a synchronous motor (see Nu´ n ˜ ez 1971). The motor’s revolutions could be adjusted at will, changing thereby the flow of the scented (80 l of vanilla essence per litre) sucrose solution delivered by the feeder. A plastic tube (0.75 mm internal diameter) connected the syringe to an artificial flower consisting of a perforated wooden cube with a coloured disc on top. The artificial flower was placed inside a wooden box that had one side replaced by a cardboard door. The second feeder (site 2) consisted of a petri dish filled with scented (80 l of vanilla essence per litre) sucrose solution ad libitum, on top of which a perforated plastic cover was placed. Arriving bees introduced their proboscis in the perforations to imbibe the offered solution. When the trained bee (henceforth donor bee) entered the box at site 1, we closed the door and turned on the pump, and recorded the time the bee spent foraging at the feeder (henceforth: feeding time). To allow observation of the bee inside the box and to assure that the animal was indeed ingesting the offered solution throughout the recorded time, we made a window on the door and covered it with a transparent plastic sheet. When the foraging visit was complete, the animal left the box through a perforation on the door in which a transparent acrylic vial (1 cm diameter2.5 cm long) had been placed. In this way the bee completed the foraging visit without perturbations but was trapped when returning to the hive. Once the donor bee had been trapped, we captured a bee from site 2 (henceforth: recipient bee) in a vial upon its arrival at the petri dish before it began to drink, having therefore an empty crop. We then weighed both bees to the nearest 0.1 mg (Wi) and attached the vials holding each bee to opposite walls of the experimental arena. The arena consisted of a transparent plastic cube (5.552 cm) divided in half by a sliding door (see Farina & Nu´ n ˜ ez 1991). Once both bees entered the arena, we allowed them to interact freely, and separated them after the first trophallactic encounter or after 10 min if no trophallactic contact occurred. Each bee was reweighed (Wf) and then killed. Trophallactic contacts lasting less than 1 s were not considered in subsequent analysis (see Korst & Velthuis 1982). We recorded the following parameters: (1) trophallactic responsiveness (number of trophallactic trials100/total trials); (2) exchanged volume (l), that is, the mean difference between final (Wf) and initial weights (Wi) of both bees divided by the density of sucrose solution; and (3) trophallactic time (s), time spent engaged in food exchange. As in previous works (Farina & Nu´ n ˜ ez 1991, 1993), we estimated the transfer rate (l/s) as the slope, b, of the linear regression between exchanged volume (l) and trophallactic time (s) for each experimental condition presented at the food source. The relationship between the exchanged volumes and their corresponding trophallactic times adjusted very accurately to simple linear equations, therefore, the estimation of the transfer rate through the slope b of this equation was more representative of the population data obtained in each experiment. We calculated sucrose transfer rate (mg of
Table 1. Flows of solution and sucrose concentrations employed for experiment 1 Flow of solution (µl/min)
Concentration (% w/w) Sucrose flow (mg/min)*
1.2
2.4
4.8
1.2
2.4
50 0.8
50 1.5
50 3.0
64 1.0
36 1.0
*Sucrose flow was obtained as flow of solution × density × concentration/100. Table 2. Flows of solution and sucrose concentrations employed for experiment 2 Flow of solution (µl/min)
Concentration (% w/w) Sucrose flow (mg/min)*
2.5
4.5
3.0
4.5
8.5
38 1.1
22 1.1
54 2.0
38 2.0
22 2.0
*Sucrose flow was obtained as flow of solution × density × concentration/100.
sucrose/s) as transfer rate (l/s) density of solution concentration/100.
Experimental Series Experiment 1 In this experimental series we addressed separately the possible influence of both solution flow and sucrose concentration on trophallactic behaviour. First, we offered a sucrose solution of constant concentration (50% w/w) at site 1 at three different flows: 1.2, 2.4 or 4.8 l/ min. We then changed the sucrose concentration to 36% w/w and presented it at 2.4 l/min or 64% w/w offered at 1.2 l/min. We chose these two concentrations so that in both cases the same sucrose flow would be obtained (1 mg of sucrose/min; Table 1), because we also wished to see whether sucrose flow as an independent parameter influenced trophallactic behaviour. For this study, carried out from March to May 1996, we used 90 pairs of bees.
Experiment 2 In this experiment we also separated both solution flow and sucrose concentration as in experiment 1, but with the additional goal of obtaining two different sucrose flows: 1.1 mg/min (similar to the 1.0-mg/min value of the previous season) and 2.0 mg/min. These two values were obtained by combining several solution flows and sucrose concentrations (Table 2). For this study (March–May 1997) we used 75 pairs of bees.
Statistical Analysis Statistical analysis was done through analysis of frequency, one-way analysis of variance (ANOVA), linear
1179
ANIMAL BEHAVIOUR, 59, 6
regression model I and slope comparison tests (Sokal & Rohlf 1981).
When sucrose solution was presented at 1.2, 2.4 or 4.8 l/min, feeding time significantly decreased with increasing flow of solution (one-way ANOVA: F4,38 =13.34, P
