Flight energetics in sphinx moths: heat production and heat loss in Hyles lineata during free flight

Sphinx moths maintain elevated thoracic temperatures (T_th) during flight that are independent of ambient temperature (Heath & Adams, 1965; Heinrich, 1970). These temperatures result from high rates of heat production associated with flight (Zebe, 1954; Heinrich, 1971a; Heinrich & Casey, 1973) and lower rates of heat loss than most insects due to a dense layer of scales on the thorax (Church, 1960; Heinrich, 1971 a). In the sphinx moth Manduca sexta, heat loss in flight appears to be regulated by varying rates of blood flow to the poorly insulated abdomen (Heinrich, 1970; 1971 b).

In the white-lined sphinx moth, Hyles (formerly Celerio) lineata, regulation of heat production has been inferred (Heath & Adams, 1965). Under confinement during periods of intermittent activity, H. lineata can maintain a relatively constant thoracic temperature by alternate periods of cooling and pre-flight warm-up (Heath & Adams, 1967). Under natural conditions, however, H. lineata regularly engages in activity which requires sustained flight, making such a pattern of temperature regulation unfeasible; during foraging it flies continuously, and during migration this species flies for hours at a time (Grant, 1937).

H. lineata is particularly interesting in terms of temperature regulation because, unlike the situation in most sphinx moths, the body weight of different individuals in the field may vary fourfold. The present study examines the effects of body weight and air temperature on the flight energetics of H. lineata and the importance of the abdomen as a site of heat exchange.

Moths were collected in the Mojave desert near Pearblossom, Los Angeles county, California, by attracting them to a u.v. lamp powered by a portable generator. Fifth-instar larvae were also collected in the same area of the desert and reared on various foodplants in the laboratory.

Thoracic and abdominal temperatures were measured with a small thermistor (GC₃₂SM2, Fenwal Instrument Co.) and read to the nearest 0·2 °C on a telethermometer. Thoracic temperature (T_th) was measured by thrusting the probe into the centre of the thorax. Abdominal temperature was measured in the approximate centre of the abdomen by inserting the probe through the intersegmental membrane behind the third abdominal segment. The data were discarded if the temperatures were not obtained within 58 of capture. Laboratory measurements of body temperatures were made within 2 days of capture. The moths were caused to fly in an 8·7 1 jar in a temperature-controlled room for at least three minutes before measurement of body temperatures. Moths were killed immediately after temperatures were taken by injecting a small amount of ethyl acetate into the thorax. In the laboratory they were weighed immediately. In the field, they were placed (in numbered paper triangles) in a box saturated with water vapour to prevent weight loss by evaporation, and weighed later on the night of capture.

For measurements of oxygen consumptions during flight, moths were placed in an 8·7 1 jar immediately following pre-flight warm-up. The moths hovered readily, and if they began to settle the jar was shaken to cause them to continue flight. Flight duration was timed with a stopwatch.

Before and after flight air samples were taken from the jar with a 60 ml syringe through a three-way stopcock cemented in the jar’s lid. Chamber air was mixed by pumping the syringe several times before the sample was collected. Air currents set up by the moths’ beating wings also helped to mix the air. Gas samples were injected through a tube of drierite and ascarite into a Beckman E-2 oxygen analyser. Oxygen consumption (s.t.p.) was calculated by the method of Depocas & Hart (1957). Data are presented only for continuous flights lasting longer than two minutes.

Cooling rates of dead moths were measured in a wind tunnel equipped with a fan attached to a variable-speed motor. Air flow was laminar. Moths, with 40-gauge copper−constantan thermocouples inserted into the centre of the thorax and abdomen, were held in the airstream by a toothpick glued to the thorax and attached to the floor of the wind tunnel. The moths were heated to approximately 45 ° C in a constant-temperature cabinet and quickly transferred to the wind tunnel at 22 ° C. Air flow in the tunnel was monitored with a hot-wire anemometer. Body temperatures and air temperature in the tunnel were continuously monitored with a multi-channel recorder.

Body weights were measured to the nearest milligram. The moths were mounted with wings in flight position. When dry, the moths were photographed together with a millimetre scale, or the wings were traced. Photographs were enlarged to life size and wing areas were measured to the nearest 10 mm² with a planimeter. Calipers accurate to o-1 mm were used to measure thoracic and abdominal dimensions and scale lengths. To calculate surface area, the thorax was assumed to be a cylinder and the abdomen to be a combination of a cylinder and a right cone (Fig. 1).

During free flight, thoracic temperatures were relatively constant over a 16 ° C range of air temperatures (Fig. 2). Thoracic temperatures of flying moths in the laboratory and in the field were about 40 ° C except at the highest T_a (32 ° C) T_th where increased slightly to 42·5 ° C. The difference between T_th and T_a varied from about 33 ° C at an air temperature of 16 ° C to 10·5 ° C at an air temperature of 32 ° C. At air temperatures of 35 ° C or more the moths were reluctant to fly, and even when repeatedly stimulated did not fly for more than a minute.

At a given T_a, the thoracic temperatures of flying moths in the field were not correlated with body weight (r = 0·06, P > 0·1 ; Fig. 3), but were weakly correlated with wing loading, the ratio of body weight to effective wing area (r = 0·53, P < 0·05 ; Fig. 4). These data are similar to those reported by Dorsett (1962) for sphinx moths from three species. However, the low value of the correlation coefficient indicates that much of the variability of T_th is not explained by wing loading.

Unlike T_tb, abdominal temperatures of H. lineata immediately following flight varied directly with air temperature. The difference between abdominal temperature and air temperature was small (maximum T_ab – T_a = 6·0 ° C) and the mean difference increased with increasing air temperature from 2 ° C at T_a. of 18 ° C to 4·2 ° C at T_a. = 32 ° C (Fig. 2).

At a given T_a, the abdominal temperature in flying moths was directly related to the thoracic temperature (r = 0·62, P < 0·05; Fig. 5). Heat production by the abdomen is negligible during warm-up and flight, and elevated abdominal tempera-tures result primarily because heat is transferred from the thorax to the abdomen. The correlation of T_ab with T_th suggests that T_th is stabilized, at least in part, by heat transfer to the abdomen (see Heinrich, 1970).

Rates of oxygen consumption should offer an accurate index to the total energy expenditure of flying insects because they do not appear to meta-bolize anaerobically during flight (see discussion by Weis-Fogh, 1961 ; Sacktor, 1970). In flying H. lineata, weight-specific oxygen consumption is inversely proportional to body weight (Fig. 6) and both linear and power regressions are significant (P < 0 ·01). Metabolism varied from about 73 ml O₂(g.h) ⁻¹ in a 0·5 g moth to about 38 ml O₂ (g.h) ⁻¹ in a 1·5 g moth. The linear least-squares relation between log-transformed values of metabolism and body weight (g) is :

In dead moths, thoracic and abdominal cooling was linear when plotted on semi-logarithmic co-ordinates (Fig. 7). The correlation coefficients (r) of linear regression equations for each individual cooling curve were 0-99. Such cooling is passive in accordance with Newton’s law of cooling which states :

In still air, the abdomen cooled only slightly faster than the thorax, but as wind speed increased the difference between the cooling rates of the thorax and the abdomen also increased (Fig. 8). At a wind speed of 1 m/s or more, thoracic cooling was twice as rapid as in still air. Above 1 m/s thoracic cooling rates were similar. The cooling rates of the abdomen increased sharply at wind speeds of 0·5–1·0 m/s and only slightly between 1 ·0 and 2·5 m/s. At 2·5 m/s, abdominal cooling rates were about three times as rapid as in still air. Cooling rates of the thorax and abdomen are inversely proportional to body weight (Fig. 9).

To determine the relative rates of thoracic and abdominal heat loss in moths of different sizes, it is necessary to know how the weights of these tagmata vary with body weight. Since heat transfer is calculated on the basis of surface area, the relations of calculated thoracic and abdominal surface area to body weight are included (Table 2).

Scale lengths differ on different parts of the body. From Table 3 several facts are apparent. The thorax is more heavily insulated than the abdomen, and on different parts of the thorax or abdomen the length of scales is not the same. In general, the scales on any particular part of the body are longer on larger moths.

On logarithmic co-ordinates, weight-specific conductance of the thorax (C_th) and of the abdomen (C_ab) are linearly and inversely related to their respective weights (Fig. 10). Abdominal conductance is greater than thoracic conductance in still air and at all wind speeds tested (Table 4). The slope of C_th is greater than the slope of C_ab as a function of weight, both in still and moving air (Table 5).

Under steady-state conditions the heat lost by a flying moth must equal the heat produced if T_th is to remain constant. Energy exchange takes place from a variety of avenues in flying moths. Only about 20 % of the meta-bolic energy is converted to power. A portion of this 20 % (aerodynamic power) is transferred to the air and represents the rate of energy expenditure necessary to keep the animal aloft. The remaining portion (inertial power) is necessary to accelerate and decelerate the wings and is eventually lost as heat through the wings (see Weis-Fogh, 1972).

About 80 % of the energy expended is immediately degraded to heat which must be lost from the respiratory passages or from the surface of the body. An attempt will be made to quantify heat loss through various avenues to evaluate their importance for thermoregulation during flight. Pertinent data used in the calculations are given in Table 6.

Conductance values for the thorax and abdomen obtained from the cooling curves should be applicable to moths flying in the laboratory because (1) substrate conduction is negligible in the wind tunnel and does not exist in hovering moths, (2) radiative temperature of the environment (T_r) is equal to the air temperature (in flying moths, T_r = temperature of the glass jar; in dead moths, T_r = temperature of plexiglass wind tunnel), (3) cooling rates at wind speeds approximating those in flight were obtained (see Bakken & Gates, 1974). However, convective heat exchange in flying moths is very complex (see, for example, Chance, 1975) and is only crudely approximated by a dead moth in a wind tunnel.

Heat may be lost in respiratory passages by evaporation and convection. Although a detailed analysis of respiration during flight in sphinx moths is lacking, a knowledge of oxygen consumption and the temperature of the thorax under various conditions may be used to estimate the rates of heat loss by evaporation from the respiratory passages during flight (Weis-Fogh, 1967; Heinrich, 1974 b). Relative humidity of 0% will be assumed so that calculated values will represent maximum rates of water and heat loss. The high rates of oxygen consumption associated with hovering flight (Fig. 6) necessitate substantial ventilation of the primary tracheae. The efficiency of oxygen extraction during flight in insects, birds and bats is about 4–7% (Weis-Fogh, 1967; Tucker, 1968; Thomas & Suthers, 1972). If H. Eneaia extracts oxygen from the incurrent air with an efficiency of 5 %, ventilation rates will vary from about 1460 ml (g.h) ⁻¹ in a 0·5 g moth to 760 ml (g.h) ^{− 1} in a 1·5 g moth.

The amount of water evaporated depends in part on the thoracic temperature during flight. At higher T_th, the saturation deficit is greater and more water is evaporated. Assuming that the incurrent air is warmed to T_th and saturated while in the respiratory system, the water loss may be calculated using the gas laws. Respiratory water loss at T_a’s of 16–32 ° C vary from o·6–0·7 mg/min in a 0·5 g moth to 0·9–1·1 mg/min in a 1·5 g moth. The heat dissipated by respiratory evaporation (2·4 J/ mg H₂O) amounts to 17–21 % of the heat production of moths flying at various air temperatures (Table 7a). Since oxygen uptake does not appear to vary as a function of T_a, it is unlikely that heat loss is controlled by regulation of respiratory water loss from the thorax. If O₂ consumption were lower at high T_a, less water would be lost by evaporation. Abdominal ventilation of sphingids is slight and probably of little importance for regulation of heat loss during flight (Heinrich, 1974 b). Cuticular water loss is also too low to contribute significantly to the heat budget (Church, 1960; Casey, unpublished observations).

In addition to heat loss by evaporation, heat is lost by respiration because the air is warmed while it is in the thorax. The specific heat of air is low, however, and respiratory heat loss by convection represents only a few per cent of the total heat production in flight and therefore may be neglected.

Even when a moth hovers in still air, there is a considerable air flow around its body because a mass of air equal to body weight must be accelerated downward through the wing to balance the weight. The rate of air flow (induced velocity) at the level of the beating wings, using body dimensions of H. lineata (Casey, 1976) and Pennycuick’s (1968) equation (3), is about 0 ·7–0 ·8 m/s. To calculate heat loss, conductance at 0·5 m/s (Table 5) is used. Thoracic and abdominal con-ductance will be underestimated by about 5 and 10% respectively at this slightly lower wind speed.

Thoracic heat loss (Table 7 b) calculated from equation (3), is inversely proportional to T_a; moths lose about twice as much heat from the thorax at T_a of 16 than at T_a of 32 ° C. At any given T_a, smaller moths lose a greater fraction of their heat production from the surface of the thorax than large moths. However, over the range of T_a employed in this study, more heat is produced than can be lost passively from the thorax.

The flight muscles are responsible for about 99% of the metabolism during flight in insects (Weis-Fogh, 1964) and it is unlikely that the abdomen produces sufficient heat to raise T_ab above T_a. Abdominal temperature is usually above T_a during flight, however (Fig. 2; Bartholomew & Heinrich, 1973), because heat is transferred from the thorax to the abdomen via blood circulation (Heinrich, 1970; 1971 b). Although the moths cannot change their insulation, variation of heat loss can occur because added heat input to the abdomen causes local changes in the Δ T, particularly in the ventral abdomen. Temperatures measured in the centre of the abdomen should represent an average abdominal temperature because the temperature in the ventral portion is higher than that in the dorsal portion (see Heinrich, 1971b). If the system is in steady state, with T_ab stabilized and the rate of heat input constant, abdominal heat loss may be estimated using equation (3).

Calculated abdominal heat loss increases from about 10% of the total heat pro-duction at T_a = 16 to about 21 % at T_a = 32°C for moths of various sizes (Table 7 c). The increase in heat loss by the abdomen at high T’a is not sufficient to offset the decrease in thoracic heat loss, and unless heat loss is regulated at some other point T_tb would continue to rise.

It should be emphasized that these data for abdominal heat loss rest on several assumptions. For example, the abdomen is not losing heat equally at all points due to variable T_ab, (Heinrich, 1971b) and differences in insulation (Table 3). In addition, heat loss may be greater than calculated values because the heat is delivered near the surface of the abdomen. It is likely that the major resistance to heat flow is provided by the scales and the muscles and cuticle which compose the body wall, but heat is not flowing from the centre of the abdomen outward to its surface, as occurs in the thorax (see Church, 1960).

Thermoregulation in hovering sphinx moths must involve regulation of heat loss because metabolism apparently is not varied as a function of T_a. In those species where regulation of heat loss is not involved, such as locusts (Weis-Fogh, 1956) and Monarch butterflies (Kammer, 1970), a constant rate of heat production results in a constant temperature difference (T_th–T_a) over a wide range of T_a.

The rate of heat loss by the abdomen at any given T_a is limited by its conductance. If all heat not lost by the thorax and by respiration (Table 8 a) is dissipated from the abdomen, by rearrangement of equation (3), it is possible to calculate the effective temperature difference of the whole abdomen. These temperatures (Table 8,b) are much higher than T_ab of moths in flight at various T_a (Fig. 2). At high T_a the moths need an average abdominal temperature similar to T_tb in order to dissipate the heat. During flight at T_a = 32 ° C, T_ab is about 7 ° C lower than T_tb. Although heat loss is facilitated at high T’a by blood flow to the abdomen, particularly when T_th is high (Figs. 2, 3), these calculations suggest that C_ab and T_ab are too low in H. lineata for all excess heat to be dissipated from the abdomen. However, there is no reason to assume that the abdomen is the only site at which regulation of heat loss can occur. To remain in thermal balance during flight at high T_a, some active cooling should occur in the head, wings or legs. The wings would make excellent thermal windows because they have a high surface-to-volume ratio and are poorly insulated. In addition, pulsating organs, which facilitate blood flow in the wings, are particularly well developed in sphinx moths (Brocher, 1919). Small size and poor insulation reduce the risk of overheating in most insects, but in sphinx moths high rates of heat pro-duction and a well-insulated thorax make overheating during flight possible even at moderate air temperatures. Heat loss by variable blood flow to the wings represents a potentially important avenue of heat exchange which should be examined in endothermic insects during flight.

Data from the present study and that of Heinrich (1971a) suggest that heat dissi-pation rather than heat conservation is important for sphinx moths during flight at ecologically relevant ambient temperatures. These data do not support a previously proposed model for heat exchange of endothermic insects during flight (McCrea & Heath, 1971; Heath et al. 1971). That model predicts that at T_as below 30 ° C, heat loss from the thorax is so great that heat production must be increased by varying flight performance. However, calculated metabolic rates based on flight mechanics (McCrea & Heath, 1971) are several times lower than measured values (for discussion see Heinrich, 1974a; Casey, 1976). Sphinx moths fly at T_as as low as 7 ° C (Bartholomew & Heinrich, 1973). If the model were correct, the moths would be unable to fly at such low T_as due to excessive heat loss. Moreover, there are no data to show that insects vary their flight performance for thermoregulation. In bumblebees (Heinrich, 1975) and in several species of sphinx moths (Casey, 1976) the metabolic rate during hovering is related to wing loading, regardless of the air temperature.

Sphinx moths are capable of flight only when they maintain T_th within a relatively narrow range. The ability to fly at high T_a depends on active heat loss at sites other than the surface of the thorax. Since most sphinx moths are nocturnal, they probably do not see temperatures as high as 32 ° C in nature. The heat production of moths during forward flight should be much lower than during hovering, and since moths probably do not hover continuously in nature, heat stress during foraging flights is probably not so extreme as to incapacitate them.

The lowest T_a at which flight is possible may well be determined by the ability to elevate thoracic temperature by pre-flight warm-up. Heat production during warm-up is proportional to thoracic temperature (Heinrich & Bartholomew, 1971 ; Bartholomew & Casey, 1973). If T_a is too low, moths produce insufficient heat to elevate T_th even when warm-up is continuous for up to thirty minutes (Heinrich, 1971 a)

The energy budget in this paper assumes that oxygen consumption during flight accounts for all the energy mobilized. There is in vitro evidence to suggest that flight metabolism is aerobic (Sacktor, 1970), and indirect evidence to suggest that oxygen availability is not limiting during flight in locusts (Weis-Fogh, 1967). However, in the locust Schistocerca gregaria (Krogh & Weis-Fogh, 1951) and in the sphinx moth Metopsilus porcellus (Zebe, 1954) there is a significantly elevated post-flight meta-bolism, lasting over an hour. The significance of this phenomenon has not been resolved, and further in vivo data are desirable.

This paper forms part of a Ph.D. dissertation submitted to the University of California, Los Angeles. Support for this study was provided by National Science Foundation grant GB32947 to G. A. Bartholomew. I thank Dr Bartholomew for his guidance and support throughout my graduate career and particularly for his editorial comments and criticisms. Thanks are due to Dr F. Engelmann, Dr E. B. Edney, Dr B. Heinrich and Mr P. Withers for helpful comments and stimulating discussions, and to Dr K. A. Nagy and Dr F. N. White for reading the manuscript. My wife Kathy helped in all phases of this study.