Mechanisms of Body-Temperature Regulation in Honeybees, Apis Mellifera: I. Regulation of Head Temperature

Studies on the physiology of body-temperature regulation in insects have been concerned almost exclusively with the regulation of thoracic temperature (Heinrich, 1974). The thorax contains not only the heat-producing muscles but also the ganglia that regulate their contraction. The neural output from the thoracic ganglia upon which flight control depends is highly temperature-dependent (Kammer, 1968; Hanegan, 1973). Additionally, a high thoracic muscle temperature is necessary to achieve the power output for flight (Leston, Pringle & Tamasige, 1962; Esch, 1976).

The insect head contains negligible muscle mass that could be used for endothermic heat production. However, the anterior end of the aorta should release heated blood near the brain. (The blood would have been heated during its passage through the active flight muscles in the thorax.) Is the temperature of the head (and by implication the temperature of the neural control centres it contains) regulated in those insects which, like honeybees (Esch, 1960; Bastian & Esch, 1970; Heinrich, 1979 a), regulate thoracic temperature (T_Th) over some ranges of ambient temperature (T_A)?

Body temperatures of free-flying bees were examined in animals transferred directly from the entrance of an observation hive (having access to the outside) to a Modulab room whose temperature was controlled to ± 0.5 °C. Relative humidity (measured by wet-bulb temperature) varied from 14% to 40%. The bees flew toward the fluorescent lights at the ceiling, and they were nudged with a pencil to stimulate them to resume flight if they landed. In addition, continuous flight was also achieved (where indicated) by tarsectomizing bees. Tarsectomized bees were unable to land on the walls or ceiling. After the bees had been in continuous flight for 3-5 min they were grasped between thumb and forefinger of a gloved hand. Body temperatures were measured by inserting a 40-gauge thermocouple (threaded into the tip of a hypodermic needle with o.d. = 0.05 mm). Temperatures were read to the nearest 0-5 °C from an Omega Engineering Thermocouple Thermometer within 3 s of capture.

Heat would flow from the bee into the glove and the thermocouple when body temperature is elevated, and from the glove into the bee when T_B is below ambient. The amount of this passive heat flow, which is a direct function of the temperature difference and the duration of contact, constitutes a source of error that was estimated by taking body-temperature measurements of bees with arbitrary but known internal temperatures. Body temperatures were continuously recorded in live endothermic bees, and these bees were then grasped and probed with the thermocouple as under experimental conditions. Average measurements of T_H were 0.06 °C lower for every °C of head temperature excess (T_H−T_A), while T_Th measurements were indistinguishable from actual T_Th. Therefore T_H was actually slightly higher ( < 1 °C) than when measured at the low T_A, and slightly lower ( < 1 °C) at the high T_A, where T_H was below T_A. The thermoregulatory capacity here demonstrated is thus a conservative estimate of the bees’ capabilities.

Heat transfer to the head was investigated in bees fastened on to a styrofoam pad with pairs of insect pins crossing between head and thorax and between thorax and abdomen. An incandescent lamp was used to focus a narrow beam of heat on to the thorax (or the head) while the head (or abdomen) was shielded with tin foil. Several body temperatures were simultaneously recorded with a Honeywell multichannel potentiometric recorder, using 46-gauge copper-constantan thermocouples insulated with cotton, except for the tip. In some experiments heart and/or aortic activity were also concurrently recorded with body temperatures using a Beckman R411 Dynograph recorder. In some experiments, body temperatures were continuously monitored with the Omega Thermocouple Thermometer and recorded, along with heart or aortic activity, on the same chart of the Dynograph recorder, in order to provide a continuous record at high magnification, and to examine interrelationships of events.

Mechanical activities of both the heart and the aorta were measured from paired electrodes on each side of the respective organs. Electrical resistance changes, resulting from mechanical disturbance or fluid flow between the paired 36-gauge copper electrodes, were measured using Biocom Model 2991 Impedance converter and recording with the Dynograph. The electrodes and thermocouples were inserted about 1 mm into holes pierced with an insect pin, and they were glued in place using a mixture of melted beeswax and resin applied from the head of an insect pin.

Operative procedures were done while the animals were under light CO₂ narcosis. The bees were generally not narcotized longer than 1 min. After coming out of CO₂ narcosis the bees were offered a 30% sugar solution. They generally fed. Measurements on each live bee were completed in 20 min or less. In order to distinguish active from passive heat transfer, bees that had been heated on the head (or thorax) while body temperature was monitored were then killed by injecting 1 μl ethyl acetate into the thorax. They were again heated (after 10-15 niin) with the thermocouples and heat-lamp remaining in the same positions. The differences in body temperature changes between live and dead bees were used to determine the extent of active cooling and to differentiate passive from active processes of heat transfer.

During free flight at T_A from 15 to 25 °C head temperature averaged 7 °C above T_A (Fig.1). However, at 36 °C T_H averaged only 2 °C above T_A, and at T_A> 40 °C T_H was usually below T_A. At 46 °C the highest T_A at which the bees were tested in flight, T_H was up to 5 °C below T_A, averaging about 3 °C below T_A (Fig. 1). These data suggest that T_H is not a passive function of ambient temperature. Head temperature is physiologically regulated by being held < 45 °C at high T_A.

Is T_H actively or passively elevated at the lower T_A? In bees with thermocouples implanted in both head, thorax, and abdomen, T_H was always close to but below T_Th. As in bees that had been in free flight, T_H generally rose and fell in parallel with changes of T_Th, while T_Ab remained close to T_A (Figs. 2 and 3). Nevertheless, T_H showed some independence from T_Th in that it sometimes decreased while T_Th increased (Figs. 2 and 3).

In bees that maintained endothermically elevated and stable body temperatures (at T_A = 23-24 °C) while pinned down on to a Styrofoam pad, the average ratio in temperature excess of head (T_H− T_A) υ. temperature excess thorax (T_Th − T_A) was

The droplet could not be observed directly in bees during free flight. However, it was clearly visible in tethered bees flying in place at high T_A. No droplet was observed at any time during flight at room temperature or lower T_A. Bees in free flight at high T_A presumably also engaged in droplet extrusion, because their mandibles were usually open and their tongues were in motion (‘tongue lashing’) after landing. A droplet was visible when these bees, which had just been flying, were examined under the microscope.

Bees that had been ‘soiled’ with the regurgitated sugar solution from the honeycrop were cleaned by hivemates when they were caged together; their pile again became fluffy.

The regurgitated droplets showed large fluctuations in temperature (Fig. 4). Droplet temperature was always several °C below T_H, and the fluctuations in droplet temperature were not correlated with tongue movements alone. It was difficult to remove samples of the fluid for analysis because the bees rapidly retracted the fluid. It is probable that the droplet is kept in motion in and out of the body. It could thus take up heat in the body, and lose it by evaporation when it contacts the atmosphere.

The appearance of a droplet in tethered bees that were heated on the head was always followed by a decline, and stabilization, of T_H (Fig. 5). When the tethered bees were killed and heated again with the same heat input and with the thermocouples and heat-lamp in the same positions, T_H typically increased to 50 °C, or as much as 10 °C above that observed in heated live bees which stabilized their T_Th (Figs. 5 and 6).

The heart in the abdomen, and its extension the aorta in the thorax and head, pumps blood anteriorly. The aorta empties near the brain, and since it traverses the sometimes hot muscles in the thorax, the pumping activity of either the heart or the aorta could have considerable effect on head temperature.

For the most part there was relatively little apparent mechanical activity of the aorta in the head in the 27 bees examined when they were not heat-stressed. However, on occasion distinct low-amplitude pulses were evident. These pulses were generally, but not always, at the same frequency as the heart pulsations in the abdomen as well as the abdominal pumping movements. Fluid visible from pinholes punched into the top of the head revealed in-out movements in synchrony with this cycle.

A dramatic change occurred, generally within a few seconds, when the head was heated to temperatures > 47 °C. Heating of the head resulted in high-amplitude aortic pulses in the head, although there was little or no response in the activity of the heart in the abdomen (Fig. 7; Table 2). Heart pulsations in many bees decreased in amplitude at the same time that aortic pulses were increasing in frequency and amplitude.

Thoracic heating resulted sometimes in an increased pumping of the head aorta (9 instances), a decrease (4 instances), or no apparent change (8 instances). Abdominal heating, on the other hand, resulted in increased frequency of abdominal heart pulsations, but it did not change the aortic pulsations in the head (Table 2). These data suggest that the blood circulation from abdomen to thorax is, in part, independent from that between thorax and head.

Sometimes after a hole was punched into the head prior to the insertion of either electrodes or thermocouples it was possible to observe in-out movement of fluid from the wound. These movements were in synchrony with abdominal respiratory movements, which were in turn in synchrony with aortic pulses (observed either with the recording equipment or visually as throbbing in the flexible membrane of the neck). These observations suggested that heating of the heart and/or the aortic, or abdominal pumping movements, functioned to move blood and possibly heat between thorax and head.

During head heating the aortic pulses were often independent of the heart pulsations, and the heart pulsations were, in turn, sometimes independent of the abdominal (respiratory) pumping movements. Which of the three pumping rhythms are primarily responsible for the heat transfer between head and thorax?

Highly amplified signals from thermocouples implanted in the head indicated that the heart activity was often matched, beat by beat, with temperature changes in the head, although independent of aortic pulsations and abdominal pumping movements (Figs. 8 and 9). When the head was heated these temperature changes (of the order of 0.001 °C) registered by the 46-gauge thermocouple indicated an influx of a pulse of medium (presumably blood) at lower temperature.

It is possible that the aortic pulses and the abdominal pumping modify or amplify the heat transfer response of the heart, but the present methods were not sensitive enough to detect this directly. However, indirect observations indicate it. For example, vigorous heart pulsations do not necessarily lead to temperature pulsations in the head. When the activity of the aorta is selectively obliterated (by overheating the head to near lethal temperatures) without affecting the abdominal heart, this results in the cessation of temperature pulses in the head (Fig. 9).

The present data indicate that honeybees regulate head temperature, but they do so apparently using only one upper set-point. They prevent overheating of the head at high T_A by evaporative cooling using regurgitated honeycrop contents (Fig. 10). Heat is redistributed by conduction, by in-out movements of the fluid droplet, and by aortal pulsations which presumably stir the blood. The temperature sensor for this response appears to be located in the head itself, since heating of the thorax to near lethal temperatures generally elicits neither the regurgitation nor the associated response of the aorta, unless T_H also becomes elevated above at least 45 °C.

At low T_A head temperature is not independently regulated. Instead, T_H at low T_A is primarily a function of Head temperature becomes elevated primarily due to passive conduction of heat from the thorax, and head temperature is thus ‘regulated’ only to the extent that T_Th is regulated. Active regulation of T_Th in honeybees has been well studied. At low T_A the bees warm up, shivering as needed (Bastian & Esch, 1970), to maintain T_Th near or above 30 °C (Esch, 1960).

The low-temperature sensor(s) may reside in the thorax itself, rather than in the head, the apparent site of the high-temperature sensor in honeybees. Heating of the thoracic muscles necessarily also heats the ganglia in the thorax. At least in the moth, Hyalophora cecropia, the temperature of the thoracic ganglia determines whether or not the animals engage in warm-up or flight behaviour (Hanegan, 1973). Temperature of the thorax also affects thermoregulatory behaviour of other insects. In the sphinx moths, Manduca sexta (Heinrich, 1971), and in the bumblebee, Bombus vosnesenskii (Heinrich, 1976), overheating of the thorax stimulates the heart in the abdomen to pump cool blood through the thorax, thus causing T_Th to decline.

In the honeybees thoracic heating by itself did not greatly stimulate heart activity, but pulses of low temperature recorded in the head were in synchrony with heart pulsations and not necessarily in synchrony with pulsations of the aorta in the head. These data demonstrate that the activity of the aorta can be independent of the activity of the heart. This suggests further that the blood circulation between abdomen and thorax may be, in part, independent of the circulation between thorax and head. Possibly the aortic pulsations modify blood flow to the bead that is driven by the abdominal pump. However, the present data do not allow one to distinguish among several alternate mechanisms of fluid-flow mechanisms.

It is of interest that the regurgitation of fluid has several functions. First, it is used to empty the honeycrop to feed hivemates (in honeybees) and larvae, and to fill empty storage cells. Secondly (in bumblebees), it is used by young queens (at both low and high T_A) prior to hibernation apparently to concentrate nectar to ensure a high calorific content of the honeycrop (Heinrich, unpublished). Lindauer (1954) reported the behaviour for honeybees in an overheated hive and studied it relative to hive temperature regulation. Esch (1976) first reported the behaviour in bees flying at high T_A and speculated that it might be used for the control of body temperature of individual bees. The present study has confirmed and extended these observations and speculations.

Support by NSF grant DEB 77-08430.