Dependence of Flight On Temperature Regulation in the Moth, Manduca Sexta

Some insects markedly increase their internal temperature during flight. Locusts (Church, 1960) maintain an internal thoracic temperature 6°C in excess of the ambient temperature. The sphinx moth Celerio lineata raises its temperature to a minimum level before take-off, and during flight it maintains an internal temperature within a range of 4°C independent of the external temperature (Adams & Heath, 1964). The regulation of thoracic temperature presumably optimizes the responsiveness of the flight system as well as allowing the moth independence of ambient temperature for its active periods.

Large moths use two distinct patterns of flight muscle activity in thermoregulation, flight and a special warm-up pattern. Celerio and other insects (Adams & Heath, 1964; Dorsett, 1962; Dotterweich, 1928; Kammer, 1967; Krogh & Zeuthen, 1941; Moran & Ewer, 1966; Soltavalta, 1954) vibrate their wings through a small angle until the muscular activity has raised the temperature of the thorax, and then, take off. Most investigators who have studied warm-up behaviour distinguish it from flight activity by the small amplitude of the stroke and the rising thoracic temperature.

Our study of Manduca sexta investigates the effects of ambient temperature upon the flight system during both flight and warm-up and the internal temperature regulation exhibited behaviourally and during muscular activity by this moth. Finally, we propose a general model for heat balance in endothermic insects.

Local sphinx moths, M. sexta, were raised at the University of Illinois and kept in diapause at 15°C till needed. They were then placed in a warm room at 25°C in a cycle of 14·75 h of light and 9·25 h of darkness per day. The moths emerged within 21–22 days after transfer. Moths were flown on the evening of their emergence to correspond to natural conditions; they were not fed prior to flight.

Each moth was implanted with a thermocouple of copper and constantan wire of 0·2 mm diameter and approximately 2 mm long. The leads of the thermocouple were approximately 1 m long and coated with enamel. The thermocouples were calibrated using a standard mercury thermometer prior to use. The thermocouple leads led to a Honeywell penwriting recorder (Electronik 16) incorporating a Zener diode reference junction.

The thermocouple was inserted through and glued to a small cork which was then glued to the surface of the thorax; no correction for heat loss through the thermocouple leads and cork support was made. The scales of the moth were removed from the dorsal thoracic cuticle. The thermocouple was inserted approximately 2 mm into the dorsal longitudinal muscle, slightly off-centre and just anterior to the scutellum. No anaesthesia was required.

All experiments were conducted in a darkened room at 22–23°C except as stated.

Weight, and wingloading were calculated in order to determine whether behaviour or temperature regulation were dependent upon these properties of the insect. The outline of the wings was traced on paper of a known mass per unit area. The paper was weighed and the area of the wings was computed.

A variable strobe light (Strobotac 1521-A) was used to measure wingbeat frequency.

Moths were attached to the temperature recorder overnight. No stimulation, such as lights or sound, was provided to arouse the moth. Under these conditions the moths awoke and flew spontaneously within the range of the leads. Seven moths were used; they were flown from one to seven nights in succession. Data obtained on later days were not used if the moth had reduced activity. Moths were flown in ambient temperatures between 15 and 26°C.

Moths show a pattern of activity consisting of four phases (Fig. 1). There is a ‘warm-up period’ during which the internal thoracic temperature rises steadily from slightly above ambient temperature to approximately 33°C. Next, a ‘flight period’ of activity, which may include brief landings, is characterized by a thoracic temperature which varies within a range of approximately 4°C. A moth may have periods of 5 or more minutes of almost constant temperature. The third, or ‘cooling period’ following activity shows a steady decline in temperature to slightly above ambient. These three active phases are followed by a period of inactivity during which the thoracic temperature is slightly above ambient temperature because of resting metabolism. In some cases, a moth warms and then cools without an intervening period of activity.

The records obtained from the moths were analysed to determine the rate of heating, the average maximum body temperature and the average minimum body temperature during continuous activity, and the temperatures during longer flights, when body temperature was almost constant.

During warm-up the thoracic temperature rises nearly linearly. In general, the rate of warm-up decreases slightly at high thoracic temperatures. The rate of warm-up was measured graphically and averaged for each warm-up period. The average for all flights was 2·69°C/min (N = 72; s.D. = 0·77; range: 1·14–5·5). The rate of warm-up varied for different trials of a particular animal. The experiments took place at ambient temperatures ranging from 15 to 26°C. Calculation of the correlation coefficient showed the warm-up rate and ambient temperature were independent (P < 0·05) over this temperature range.

The maximum and minimum temperatures during active periods were determined from the records as shown by the arrows in Fig. 1. The values for these levels vary. For example, during one evening one animal showed maximum temperatures of 35·8, 39·8, 36·8, 37·2, 36·4, 36·6, and 36·6°C and minimum temperatures of 34·1, 30·8, 35·0, 34·6, and 35·1°C. The average of the average maximum for each animal was 36·76°C (N = 7; S.D. = 0·58; range: 35·2–37·56). The average of the average minimum for each animal was 34·03°C (N = 7; S.D. = 0·72; range: 33·1–35·1). Calculation of the correlation coefficients indicated that the average maximum temperature of an animal is independent of its average minimum (P < 0·05).

Fig. 2 is a histogram of the number of records for each maximum and each minimum temperature. Each dot represents the average minimum or maximum for one animal. The wide range of values indicates that thermoregulation during flight activity cannot be described by simple internal set-points. Thermoregulation is probably influenced by moment-to-moment changes in the environment, such as available landing or take-off sites or light intensity.

On continuous flights, such as that shown in Fig. 1, the body temperature is almost constant for 5 min or more. The average for all such flights obtained in our records is 36·5°C (N = 38; S.D. = 1·58; range: 33·0–39·1). This temperature is closer to the average maximum during activity, 36·93°C than to the average minimum, 33·45°C.

Fig. 3 shows the relationship between thoracic temperature during flight and the ambient temperature. They are not independent (P < 0·05). The relationship between ⁠, the mean of the distribution of thoracic temperatures for a given T_α, ambient temperature, is:

Experiments were conducted to determine whether cooling after flight is active or passive. In each of these the moth was twice stimulated to warm up, and its thoracic temperature and the ambient temperature were recorded during the subsequent cooling period. The moth was then chloroformed and the dead moth was twice heated to a temperature approximately equal to that following flight and allowed to cool.

Cooling was examined as a function of the difference between the thoracic tem-perature and the ambient temperature, both for live and dead moths. The rate of cooling was determined graphically from the record of the thoracic temperature as a function of time.

The rates of cooling were also determined for dead and live moths treated as before but placed in a wind tunnel at speeds simulating flight of 3 m s ^{− 1} and 6 m s ^{− 1}. The moths were mounted with their wings extended in flight position. Fig. 4 compares the cooling rates of a moth at 3 m s^{− 1} and in still air. In this example, as in other cases studied, the live and dead moths cool at identical rates, indicating that the live moths cool passively rather than employ some active method. The cooling of live moths follows Newton’s law of cooling:

If M. sexta becomes exposed to direct insolation during the day, it will, after a time, move to a new location. Failure to move could result in overheating or death. In Celerio lineata, Adams & Heath (1964) showed that movement to shade occurred whenever the moth was warmed above 38°C. This shade-seeking or maximum voluntary tolerance was used to locate temperature responsiveness in the central nervous system.

A moth with a thermocouple implanted in the thorax was placed on a vertical cheesecloth substrate. A 125 W infra-red lamp was shone on the animal to heat it at a rate approximately equal to the normal rate of warm-up. After a time the moth walked up the cloth and out of the heated area. This behaviour is analogous to shade-seeking which occurs in the field. The temperature when it began to walk was recorded as the temperature of maximum voluntary tolerance. These experiments were conducted in a lighted room, because shade-seeking would normally occur during the day. This procedure was also followed with moths whose nerve cord was severed in one of three locations: (1) between the head and the thorax to eliminate nervous signals from the head; (2) between the thorax and the abdomen to eliminate nervous signals from the abdomen; or (3) between the prothoracic ganglion and the pterothoracic ganglion.

Transections of the ventral nerve cord were carried out on moths immobilized by chilling and anaesthetized by carbon dioxide. The scales were cleaned from the surgical field, and a small flap was cut in the cuticle. The nerve cord was severed, and the adjoining sections were separated. The cuticle was sealed with paraffin. The animal was dissected after experimentation to verify that the cord was severed and separated.

Both operated and unoperated moths initiated walking movements after a period of heating. The average thoracic temperatures for operated and control moths are shown in Table 2.

Nervous connexions between the thorax and either the head or the abdomen are unnecessary to produce this response. The internal temperature receptors responsible for the shade-seeking response therefore are located in the thorax, but we cannot distinguish from these data whether many diffusely distributed temperature receptors are involved, or whether temperature sensitivity of the central nervous system itself triggers the behaviour. Each animal had a range of approximately 4°C between the highest and lowest measurements, which might indicate that this shade-seeking is not simply cued but possibly is the result of the input from many temperature receptors warming at different rates. When the thoracic ganglia are separated survival is poor. The failure to avoid high temperature may indicate interruption of synergistic nervous connexions.

M. sexta restricts its body temperature during activity to a small range of temperatures, presumably to optimize flight. We therefore studied the relationship of temperature regulation to several features of the flight system and flight behaviour of M. sexta.

Thoracic temperature and wingbeat frequency were monitored during six to ten warm-up periods for each of seventeen moths over a range of ambient temperature of 22·5−27·5°C. In half the experiments the moths were suspended by the thermo-.couple leads. Removal of the tarsal support caused each moth to begin warm-up vibrations of small amplitude until it either abruptly ceased activity or began flying. In the remaining experiments the animals stood supporting their own weight. Pinching the abdomen or antennae or manoeuvring an object in front of the animal was sufficient stimulus to initiate vibration of the wings. No wind tunnel or other aerodynamic device was necessary.

The average weight of freshly emerged, but hardened M. sexta was 1·63 g (S.D. = 0·25). Males (N = 8; range: 1·21–1·89)averaged 1·51 g, slightly less than females (N = 6; range: 1·46–2·13) which averaged 1·80 g. Females are expected to weigh more because of their eggs. However, they have less fat available for fuel. For all animals, wingloading averaged 0·073 g/cm² (S-D- = 0·0047). Females (N = 6; range: 0·66–0·82) averaged 0·0731 g/cm², slightly more than the average of the males, 0·0726 g/cm²(N = 8; range: 0·066–0·080).

Moths stimulated as described vibrate their wings through a small angle of 5–20°. They usually hold their wings vertically. The frequency of vibration increases with increasing thoracic temperature during this period. Fig. 5 includes all wingbeat frequencies recorded during ten trials for one animal. The wingbeat frequency levels off just before take-off. At this time the amplitude of the wing stroke increases.

The Q₁₀ of wingbeat frequency with internal thoracic temperature was calculated for each moth using the average of at least three values at the low temperatures at the beginning of the warm-up period, and the average of at least three values at high temperatures shortly before take-off, but before the wingbeat frequency levelled off. The average Q₁₀ of wingbeat frequency with thoracic temperature was 1·41 (N = 17; S.D. = 0·10; range: 1·27–1·69). The rate of rise of wingbeat frequency with thoracic temperature is linear over the temperature range studied. The Q₁₀ of wingbeat frequency with thoracic temperature averaged 1·44 for females and 1·39 for males. Suspended animals averaged 1·43 and standing animals averaged 1·38. No significance was attached to these differences.

The Q₁₀ of the rise in wingbeat frequency with thoracic temperature is independent of ambient temperature (P < 0·05). Thus, it is independent of the initial temperature of the moth or the gradient between the thorax and the air. The Q₁₀ of the wingbeat frequency with thoracic temperature is independent of both weight and wing-loading (P <0·05).

Since there is a wide range of wingbeat frequencies at take-off for any one animal, one would not expect that wingbeat frequency at take-off would be dependent upon a specific physical parameter of the insect’s body. Calculation of the correlation coefficients indicates that wingbeat frequency at take-off is independent of both weight and wingloading (P < 0·05).

The wingbeat frequency at take-off was 26·03 Hz for females and 26·75 Hz for males. No significance has been attached to the small difference between the average value for males and for females. The average wingbeat frequency at take-off was 25·64 Hz for animals warming while suspended from a tether but 28·48 Hz for standing animals. Perhaps the wire suspension provided additional lift for the suspended animals and thus allowed them to fly at a lower wingbeat frequency.

The thoracic temperature at take-off was defined as that recorded just before full flight for the suspended animals, and that recorded just before take-off for standing animals. The amplitude of the stroke increased and frequency levelled before take-off in the latter case. Each animal exhibited a range of take-off temperatures in successive trials. For example, one moth had temperatures of 34·8, 35·7, 35·9, 35·5, 36·2, 37·0, 37·2, 36·0, and 37·8°C in a series of take-offs. Thus the temperature of the flight muscles cannot be a unique cue for take-off. Since each moth showed a range of temperatures marking the transition from warm-up to flight rather than a specific set-point, take-off probably also depends upon other environmental factors such as site, light intensity, etc.

The average of the average temperatures at take-off for all animals was 31·97°C (N = 15; S.D. = 1·90; range: 26·6–35·4). The thoracic temperature at take-off is less than the average minimum thoracic temperature during activity. The average temperature of the thorax at take-off is independent of ambient temperature, weight, and wingloading (P < 0·05).

The temperature at take-off averaged 31·7°C for males and 31·6°C for females. Suspended animals averaged 31·3°C and standing animals averaged 32·98°C.

The temperature at which the animal would maintain horizontal flight was defined as the minimum temperature for flight. Three moths were heated externally with an infra-red lamp at a rate approximating that of warm-up. At intervals of 1·5°C during the heating of the animal the leads of the thermocouple were removed from the recorder and the animal was released. At low thoracic temperatures a moth either falls or glides to the floor. At high temperatures a moth is able to climb in flight. An average temperature of 28·8°C is necessary for horizontal flight. This is very close to the average temperature at take-off of 31·9°C. The average minimum temperature maintained during activity is at least 2°C higher than the temperature necessary for horizontal flight. Therefore, the moth keeps its temperature high enough during its active period to manoeuvre effectively.

The sphinx moth, M. sexta, maintains its body temperature during activity within the temperature range of 33–37°C. Fig. 6 summarizes the temperature responses and temperature-dependent behaviour of M. sexta. The thermoregulatory responses of M. sexta tend to occur over a range of temperatures rather than at precise set-points. For example, the temperature at which moths cease warming (maximum temperature during spontaneous activity) has a standard deviation of 0·58°C. Thus, the mean value should be thought of as the temperature of a moth with a 50% probability of cessation of warming, rather than as a set-point like a bimetallic thermostat (Hardy, 1961). Nevertheless, such responses can lead to effective thermoregulation (Heath, 1970).

The thermoregulatory responses of M. sexta imply that this insect can sense its internal temperature. However, no internal temperature receptors have been detected in insects, although the isolated nerve cord of the cockroach is temperature sensitive (Kerkut & Taylor, 1956). One might infer from the range of temperatures for each response that in M. sexta a general temperature sensitivity of the nerve cord was responsible for thermoregulation rather than specific temperature receptors. On the other hand, thermal sensitivity of the nervous system may be localized. Shade-seeking occurs in animals without input from the head, so that a cephalic integrating centre is unnecessary. Insects may lack higher integrating centres and have differing sensitivity along the nerve cord. Celerio lineata exhibits shade-seeking at 34°C if only the abdomen is heated, but moves to shade at 38°C if the entire body is heated (Heath & Adams, personal communication).

Hanegan &Heath (1970b) have shown that transition from warm-up to flight depends upon the temperature of the thoracic ganglia of Hyalophora cecropia. Using tiny thermodes, they heated the ganglia and induced the transition to flight. When the heating stopped, the moth returned to the warm-up motor pattern. They proposed that the transition is not controlled by a temperature receptor; rather, the transition occurs because nonspecific activity in the nervous system is heightened by warming releasing inhibition on the flight-pattern generator. Hanegan and Heath’s model emphasizes that set-points based upon temperature receptors, as they are modelled in vertebrates (Hammel, 1968), are not necessary in insects. Their model also supports our conclusion that thermoregulation in moths does not behave as if controlled by precise set-points.

Warm-up behaviour. The warm-up motor pattern consists of synchronous firing of antagonist flight muscles (Kammer, 1968). As a result of this activity the body temperature rises. Hanegan & Heath (1970b) have proposed that the warm-up mechanism in the moth, H. cecropia, involves a different pattern generator in the central nervous system from that controlling flight. Their study showed that the frequencies of the two patterns are not reconcilable; hence, both patterns cannot be produced by the same generator. Kammer (1968) rejected the hypothesis of multiple pattern generators as unnecessary to explain warm-up patterns in sphinx moths because the stroke frequency in warm-up and flight were very similar at the same temperature.

The Q₁₀ of wingbeat frequency against thoracic temperature is about 1·4 for Manduca sexta. However, the change in frequency is linear with change in thoracic temperature for this moth, as Dorsett (1962) reported for four other species of sphinx moths. The warm-up mechanism causes a linear rise in temperature at a fixed rate. Both stroke frequency and rate of rise of thoracic temperature are independent of ambient temperature over a wide range of external temperature. Wingbeat frequency in warm-up is dependent only on thoracic temperature. Heat production during warm-up is high and increases as the thoracic temperature increases (Heath & Adams, 1967). This adjusts for the increasing heat loss experienced by the moth as its gradient to the outside increases.

In Table 3 we have calculated the heat required to raise the thoracic temperature 2·69°C in a minute by a 0·5 gm moth with a specific heat of 0·8 cal g^{− 1}°C (2·69 × 0·5 × 0·8 = 1·08 cal). We then divided through by the wingbeat frequency at 24°C of 20 Hz to obtain the heat generated per minute by a single stroke (0·054 cal). We multiplied this value by the wingbeat frequency of a moth at 30°C, 26·5 Hz, to obtain the heat production generated by increasing the wingbeat frequency (1·43 cal). This value is compared to the heat production required to maintain a 6°C gradient and continue the warming rate of 2·69°C min⁻¹. At any beginning ambient temperature the increase in wingbeat frequency accompanying a rise in thoracic temperature will account for about 75% of the increase in heat production needed to maintain a linear rise of thoracic temperature. The remaining 25% could come from the increased metabolism of warmer tissues. Thus, the heat production per muscle twitch need not change appreciably during warm-up.

When warm-up from several ambient temperatures is considered, the rate of heat production can no longer be accounted for by a uniform heat production by each twitch of the muscle. A moth warming from an ambient temperature of 20°C has a higher heat production at 30°C body temperature than one warming from an ambient of 25°C at 30°C body temperature. Both insects have the same stroke frequency. Thus, the heat produced depends on the gradient between the thorax and ambient and not on stroke frequency or body temperature. This difference in heat production is not due to any simple intrinsic property of the muscle, but must arise from changes in the control of muscle activity. Some of the additional heat could be generated by multiple firing of each muscle during a single stroke, or by recruitment of muscles as the gradient increases. Kammer (1968) reports examples of both activities. However, systematic study of such activities with respect to the heat production of the muscle are lacking.

After a period of warm-up M. sexta either increases its stroke amplitude and locates a suitable site for take-off, or it takes off directly. Both the stroke frequency at take-off and the thoracic temperature at take-off are lower in moths that are suspended than in moths that must generate lift to leave the ground. Moths do not gain the initial velocity required for flight by jumping as do many insects. Rather, they generate sufficient lift to rise by hovering from a surface. Since hovering requires more power than direct flight, the moths take advantage of suspension to begin flying directly.

Both the temperature at take-off and the wingbeat frequency at take-off are independent of weight and wingloading. Apparently, voluntary take-off includes a considerable safety factor in estimating the power available to generate lift for flight, because M. sexta will not voluntarily begin flight until its temperature is several degrees above the minimum temperature for horizontal flight.

Like other responses to temperature, the temperature and wingbeat frequency at take-off vary among moths and among different trials for one moth. It would seem that after the moth has achieved the minimum temperature required for flight, other conditions, such as site or location, determine when take-off occurs.

Contrary to our findings on Manduca, Dorsett (1962) found with Deleiphilia nerii that each individual had a constant temperature at take-off which was independent of ambient temperature but dependent upon wingloading. From his data one might conclude that in Deleiphilia the take-off temperature triggers the switch from warm-up to flight, and that this temperature was set by physical characteristics of the animal.

Some function of a moth directly concerned with flight must operate too slowly at low temperatures to permit flight. Wilson (1968) investigated relative refractoriness in locust motor neurons, and he found it to be about 10 ms at 25°C. If moth motor neurones have similar properties, they can fire rapidly enough to drive the wing muscles without warm-up. Thus, the thermal sensitivity of the motor neurones does not set the minimum temperature for flight.

The amplitude of the wing stroke could be a limiting factor. However, moths at low temperatures occasionally flap their wings through a large angle when handled. Thus, the small angle of the wings during warm-up is not due to inability to produce large strokes.

Moths may warm-up to achieve sufficient lift for flight. The lift required to support an insect is set, in the case of locusts, both by frequency of wingbeat and changes in the angle of attack of the wing (Gettrup & Wilson, 1964). Frequency is limited by the duration of the muscular contraction at a given temperature. It is metabolically expensive for an animal to attempt flight if the contractions of the antagonistic muscles, depressors and elevators, overlap. For maximum efficiency, the contraction time of each muscle should be approximately half the maximum period (inverse of the minimum frequency necessary for flight). Although no data are available for M. sexta,

Neville & Weis-Fogh (1963) have studied the effect of temperature upon locust flight muscle. The contraction time for the dorsolongitudinal muscle of the locust decreases approximately linearly from 52 ms at 23°C to 26 ms at 40°C.

The average period at take-off for M. sexta was 37·3 ms. If the dorsolongitudinal muscles of M. sexta are similar to those of S. gregaria, the contractions of the antagonistic muscles would overlap at any of the temperatures considered by Neville and Weis-Fogh. The muscles of M. sexta are probably faster than those of S. gregaria. Nevertheless, these data suggest that at high thoracic temperature alternating muscles contract without significant overlap. Neville and Weis-Fogh have shown that at low thoracic temperatures the overlap between the contractions of antagonistic flight muscles would cause the locust to fly inefficiently and generate extra heat causing its thoracic temperature to rise to a level where such interference would not occur. It is an incipient form of thermoregulation. Similar interference may occur in moths at the temperature of take-off or below. Thus, not only is there a minimum temperature for horizontal flight, but also a temperature below which horizontal flight cannot occur efficiently. Thus, take-off might be at the optimal temperature, maximum efficiency at a minimum frequency, of the flight system.

There is sufficient information on the channels of heat gain and loss to construct in some detail the energy balance of an endothermic insect in flight at a range of ambient temperatures. The only source of heat for a nocturnal insect is that generated in his flight muscles and basal metabolism. The channels of heat loss are loss by radiation, conduction, and convection from the surface of the thorax, convective heat loss due to the relative wind in flight, evaporation from the thoracic trachea, and heat transfer by the blood from the warm thorax to cool parts of the animal. The primary restraints are the temperatures regulated by the moth in activity.

The heat loss from the thorax in still air and at various flight speeds can be determined from the rate of cooling of the moth (Heath & Adams, 1967; Heath & Adams, 1969). The rate of loss predicts the necessary heat production to maintain a constant or high internal temperature. Since the moths maintain higher body temperatures in flight at high ambient temperatures, the linear relationship between heat production and ambient temperature must be corrected for heat storage in the body (Fig. 7).

The rate of heat loss due to evaporation can be estimated from Weis-Fogh’s (1967) study of the locust. He showed that the locust ventilates the thoracic tracheal system largely through changes in thoracic volume accompanying wing movements. In the locust this respiration is adequate to supply the metabolic needs of the flight muscles. He also indicates that sphinx moths show large thoracic movements like those of the locust. Using Weis-Fogh’s estimates of the volume of air moved through the tracheal system of the thorax and adjusting it for the high wingbeat frequency of the moth, we have calculated the heat loss due to evaporation for a moth maintaining a body temperature of 39°C. The amount of water evaporated is proportional to the saturation deficit between the warm thoracic air spaces and the exterior air (Fig. 8). We calculated these values by assuming a minute-volume of 30 ml, a body temperature of 39°C, a specific heat for insect blood of 1·0, and a latent heat of evaporation of 580 cal gm^{− 1} of water.

The metabolism required for level flight in M. sexta was calculated by C. J. Pennycuick for P. J. Wilkin of our laboratory from measurements of the mechanical system of the moth. These calculations are based on Pennycuick’s (1968) model of flight in the pigeon. Pennycuick’s values were converted to calories per minute and are presented in Fig. 9. These values correspond closely to the energy requirements of flight in the locust, S. gregaria (Weis-Fogh, 1952). M. sexta is of about the same size as the locust.

A model summarizing these channels of heat loss and gain is shown in Fig. 10. Whenever the heat production is less than the rate of loss, the body temperature will decrease. For example, a moth flying at 6 m s^{− 1} is in thermal equilibrium at about 30°C in a saturated atmosphere. If it decides to fly slower or if the ambient temperature falls, it must either increase its flight speed or increase its heat production by changing the efficiency of flight. It may land intermittently and enter the warm-up pattern, or it may adjust its flight muscles to cause overlap of the contraction of antagonist muscles. It may engage in climbing flight and powered descent. In any case, level flight at 6 m s^{− 1} at ambient temperatures below 30°C is not compatible with maintenance of body temperature.

Whenever the heat production due to flight at any speed is greater than the rate of heat loss, the moth will incur a thermal excess which will cause a rise in body temperature. A moth can continue flying while generating a heat excess by increasing the flow of blood between the thorax and abdomen (Heinrich, 1970 a). Heat is dissipated in the relatively cool abdomen and cool fluid returns to the thorax. The amount of heat transported in this manner can be calculated from the excess heat expected in level flight at high ambient temperatures. The calories lost equal the volume of fluid transported times the gradient between the thorax and abdomen. Only 0·1 ml min^{− 1} of blood will dissipate 1·5 calories at a gradient of 15°C.

In Fig. 10 a moth flying at 6 m sec^{− 1} at 35°C has a heat excess of 0·75 cal min^{− 1}. This would cause the thoracic temperature to rise 1·5°C in a minute. A cardiac output of 0·4 ml at a 4°C gradient will dissipate the entire heat excess. The required cardiac output necessary to dissipate excess heat from flying at 6 m s⁻¹ over the range of 30 to 35°C is shown in Fig. 11.

The heart cannot be stopped completely if a moth is to fly continuously. A moth must receive substrate for metabolism of the flight muscles. Thus, at body temperatures below 30°C at 6 m s^{− 1} flight speed, the circulatory system becomes a channel of undesirable heat loss for which the moth must compensate by increased heat production.

It is possible to estimate the cardiac output required to meet the metabolic needs of M. sexta at several flight speeds. Heinrich (1970b) has shown that the heart rate depends upon thoracic temperature and remains nearly the same even when the cardiac output changes by an order of magnitude. Thus, the cardiac output is adjusted by changes in the stroke-volume of the heart. Assuming a heart rate of 100 min^{− 1}, we have calculated the stroke volumes required to produce the cardiac output needed to dissipate excess heat generated by exercise in the moth. Fig. 11 shows the stroke volume predicted plotted against ambient temperature. At a flight speed of 6 m s^{− 1} the stroke volume decreases with decreasing ambient temperature until, at 30°C, the moth is in heat balance. At that temperature the circulatory loss of heat is no longer needed. However, the line predicts a stroke volume of about 0·07 μl. This may be taken as the minimum volume required to transport substrate for metabolism of the muscles. A similar estimate for 8 m s^{− 1} is about 0·12 μl. The ratio of metabolism at the two flight speeds is 1·6 and of stroke volume is 1·7. A line connecting these points and extrapolated to 35°C, the thermoneutral temperature for a moth flying at 3 m s^{− 1}, predicts a minimum stroke volume of 0·05 μl. A line then connecting from 35° C and 0·05 μl to the intersection of the slopes of stroke volumes for 6 and 8 m s^{− 1} flight speeds accurately predicts the stroke volume required at 36°C and 3 m s^{− 1}. The minimum cardiac output at 6 m s^{− 1} predicts an unavoidable heat loss due to circulation at temperatures below 30°C (Fig. 10).

The following would seem to hold for the role of circulation in exercise and heat loss. Thoracic temperature determines the heart rate independent of the cardiac output (Heinrich, 1970b). The rate of convective transfer within the animal is adjusted by changes in stroke volume of the heart. Heinrich (1970b) observed large changes in stroke volume during heat stress in M. sexta, although he did not quantify them. The cardiac output is adjusted to changes in level of exercise by changes in stroke volume rather than in heart rate. Because of the large changes in cardiac output required of insects to adjust for heat loss, increasing the heart rate, even by 100%, could only account for a small percentage of the blood flow. These conclusions are consistent with Heinrich’s (1970a, b) observations in M. sexta, and Hanegan’s observations (unpubl.) on Hyalophora cecropia.

If the cardiac bolus is spherical in shape, a change in radius of × 4 to × 5 will account for the range of cardiac outputs predicted. Hanegan’s unpublished observations on the heart of H. cecropia show them to be of that size when that moth is under severe heat stress.

The temperature responses of M. sexta can be compared to those for other insects. Regulation of temperature in M. sexta seems to differ from that in S. gregaria in two ways. First, M. sexta is influenced less by ambient temperature during periods of continuous flight. Using the regression line shown in Fig. 3, one can calculate that if the ambient temperature varied between 25 and 35°C, thoracic temperature of the locust would vary between 31°C and 41°C while that of M. sexta would vary from 37·2 to 39·9°C. This shows that M. sexta exercises greater control of its thoracic temperature than does S. gregaria.

Regulation of temperature during flight and activity is also found in other Lepidoptera (Adams & Heath, 1964; Soltavalta, 1954). M. sexta exhibits the same pattern of temperature regulation during spontaneous flight as Celerio lineata (Heath & Adams, 1965). Table 4 lists the average maximum and average minimum temperatures during activity, the take-off temperatures, and the shade-seeking temperature, in both M. sexta and C. lineata (Adams & Heath, 1964; Heath & Adams, 1965). The range is approximately the same for the two moths. In C. lineata, however, the shade-seeking temperature and the take-off temperature are closer to the average maximum temperature than to the average minimum temperature as they are in M. sexta. The average rate of warm-up, 2·68°C min^{− 1}, of M. sexta is less than that of Deleiphilia nerii, 4·2°C min^{− 1} (Dorsett, 1962), or that of C. lineata, 4·06°C min^{− 1} (Adams & Heath, 1964)_:

Heinrich (1970 a) has also studied temperature in M. sexta during flight. His study, performed on moths from the Mojave desert, show a similar dependence of thoracic temperature upon ambient temperature during flight. However, Heinrich’s moths maintain much higher temperatures (36−44°C) than our moths from Illinois (33–39°C). Although our techniques differ from his, the discrepancy more probably represents adaptation to the respective climates.

Hanegan & Heath (1970,a) with H. cecropia, and Heath & Adams ( 1967) with C. lineata, found marked decrease of O₂ consumption, hence also in heat production, in response to high ambient temperatures. Both of these moths show more precision in thermoregulation than M. sexta, in that both C. lineata and H. cecropia show less variability of thoracic temperature for any given behavioural parameter than M. sexta. On the other hand, there is no reason to relate variability of temperature regulation to differing mechanisms of heat loss and production as Heinrich (1970 a) has suggested. Rather, it seems likely that all of the moths studied use both as means of regulating temperature. The variability probably derives from differences between species in the general sensitivity to temperature of non-specific reflexive inputs (Hanegan & Heath, 1970 b).

This work was supported in part by NSF GB 6303 and NSF GB 13797 and a NIH traineeship in Biophysics to Mary Jacobs McCrea. Dr James L. Hanegan contributed much time to the development of the model presented here and to calculations of cardiac output. Dr P. J. Wilkin contributed advice on the metabolic cost of flight. We are especially indebted to Dr C. J. Pennycuick for his calculations on the cost of flight in M. sexta.