An Analysis Of Pre-Flight Warm-Up In The Sphinx Moth, Manduca Sexta

It has been known for many years that some insects of the orders Coleoptera, Hymenoptera, and Lepidoptera initiate flight only at body temperatures above 30 °C (Dotterweich, 1928; Krogh & Zeuthen, 1941; Sotavalta, 1954; Dorsett, 1962), and that some species maintain thoracic temperatures above 35 °C while flying. The heat which causes the rise in body temperature during pre-flight warm-up is produced by contractions of the flight muscles.

This study examines the inter-relations of ambient temperature with cardiac performance, wing movements, insulation, convection, and rates of change in thoracic and abdominal temperatures during pre-flight warm-up and post-flight cooling in sphinx moths weighing from 1 to 3 g.

The moths were reared using a slightly modified version of the method of Hoffman, Lawson & Yamamoto (1966). The eggs and larvae were collected on Jimson weed (Datura stramonium) in the Mohave Desert of California. In the days between emergence and experimentation the moths were kept at an environmental temperature of 21–23 °C. Measurements were made in an insulated room in which temperature could be controlled to within 0·5 °C over a range from 0 to 60 °C.

All temperatures were measured to within 0·1 °C using copper-constantan thermocouples connected to a multichannel recording potentiometer. Thoracic temperatures were measured with 40-gauge thermocouples inserted slightly off centre through small holes pierced in the tergum. Care was taken to avoid disturbance of the scales. The thermocouples remained securely in place without further attention. Abdominal temperatures were measured ventrally with 40-gauge thermocouples slipped through the intersegmental membranes.

Wing movements and heart action were recorded from pairs of silver electrodes implanted dorsally on either side of the heart in both thorax and abdomen. The electrodes were wired to two Biocom impedance converters which in turn were connected to different channels of a Grass Polygraph. The changes in impedance which were recorded resulted from alterations in the distance between electrodes and from changes in electrical conduction across the dorsal vessel. In favourable preparations both wing movements and heart action were clearly shown. All measurements of heating and cooling were made in still air.

Thermocouples and impedance electrodes were implanted 16–20 h prior to measurement in animals anaesthetized with CO₂. Between implantation and measurement the moths were placed on a moist sponge in a transparent plastic container with room lights on, and kept at a temperature at which they were to be studied. To initiate warm-up, illumination was reduced. At the higher temperatures no further stimulation was needed, but at temperatures of 15 °C and below gentle prodding was sometimes required to arouse the animals.

At rest the thoracic (T_Th) and abdominal temperatures (T_Ab) of Manduca sexta are indistinguishable from each other and from ambient temperature (T_A), as is the case with most insects (Krogh & Zeuthen, 1951). However, during pre-flight warm-up T_Th rose rapidly while T_Ab showed little change.

The increase in T_Th during warm-up was approximately linear until flight temperature was approached (Fig. 1). The mean rates of increase in T_Th during warm-up increased linearly with T_A. At T_A of 15 °C the rate of increase was approximately 2 °C/min, while at 30 °C it was about 7 · 6 °C/min (Fig. 2). The difference between maximum and minimum rates of increase in T_Th was greatest (4 · 5 °C/min) at a T_A of 30 °C, and least (1 · 2 °C/min) at a T_A of 15 °C. At the latter temperature the moths sometimes seemed reluctant to initiate warm-up even after repeated tactile stimulation.

Mean thoracic temperatures immediately after initiation of flight varied from 36 · 8 to 38 · 8 °C with the lowest figures occurring at ambient temperatures between 12 and 20 °C (Fig. 3).

During warm-up sphinx moths produce heat by contractions of the thoracic muscles which vibrate the wings rapidly through a narrow angle. These wing vibrations are sometimes called shivering by analogy with endothermic vertebrates. The rate wing vibration increased linearly (r = 0 · 97) with thoracic temperature, from about 8/sec at 15 °C to approximately 25/sec at 35 °C (Fig. 4).

When a resting M. sexta is held by its legs, it immediately begins to move its wings synchronously through a wide angle as if trying to fly. These ‘flying movements’ cause a rapid rise in body temperature. The rate of increase in T_Th during this forced warmup closely approximates that observed during normal warm-up (Figs. 1 and 5).

Although the rate of wing movement during forced warm-up increased with increasing T_Th, at any given T_Th it was less than the rate of wing vibration at the same temperature during normal warm-up. In addition, the increase in wing beat tended to show a curvilinear relation to T_Th, falling off at temperatures above 30 °C (Fig. 6).

Although the dense covering of scales of sphinx moths affords insulation (Church, 1960; Heinrich, 1971a), in still air M. sexta is able to warm up to flight temperature even with all thoracic scales removed (Fig. 7).

The cooling curves of moths, immediately after being killed with ether, are straight lines when plotted on semi-logarithmic coordinates (Figs. 8 and 9). The cooling curves of live animals are sometimes linear on such a plot, but are often curved as shown in Fig. 8. In all cases live animals cooled more rapidly than dead ones, and their rates of cooling at a given difference between T_Th and T_A (Δ T) tended to be greater at high T_A than at low. Using a Δ T of 15 °C as a starting point, the decrease in mean T_Th of nine intact live animals in still air during 4 min was 11 · 5 °C in a T_A of 25 °C and 8 · 5 °C in a T_A of 15 °C (Fig. 9).

These differences indicate that the rate of heat loss from the thorax in M. sexta during post-flight cooling is not completely passive but is affected by physiological factors, one of which appears to be an accelerated transfer of heat to the abdomen. During warm-up the T_Ab remained at or near T_A, but after T_Th reached the level characteristic of flight, T_Ab rose several degrees above T_A and continued to rise as T_Th decreased during cooling (Fig. 10).

Under the conditions of measurement the performance of the heart of resting M. sexta was extremely variable. Both rate and amplitude of beating varied widely at any given body temperature. In a general way heart rate tended to increase direction with resting body temperature, but quantitative characterization of the relation between heart rate and body temperature is difficult because of long periods without beats, periods of reversed direction of contraction, and variable amplitudes of contraction (Fig. 11). Despite the variability of heart performance seen in most animals, in some individuals at rest the contractions travelled the length of the heart regularly and smoothly for many consecutive minutes. As previously reported (Heinrich, 1971 b) records of impedance changes indicate that contractions of the heart travel along the length of the abdomen and continue through the thorax.

During warm-up the records of thoracic heart rate were often either absent or obscured by the movements associated with wing vibrating. However, in some cases both the rates of wing vibration and of thoracic heart beat were clearly recorded Figs. 12 and 13). The signals from the abdominal electrodes coincided with thoracic heart beats in resting animals, but characteristically became arrhythmic and got out of phase with the thoracic beats during warm-up. In those cases where regularity of heart beat was maintained the frequency differed markedly from that recorded simultaneously from the thorax (Fig. 12). Temperatures in the abdomen remained near ambient during warm-up. We assume that we were observing fibrillations rather than beats when the signals recorded from the abdomen were at a higher frequency than those from the much warmer thorax (see upper record, Fig. 12). At the completion of warm-up wing vibrating ceased abruptly and the underlying thoracic heart beat was then more apparent (Fig. 13). Sometimes immediately, and always within several minutes after the termination of wing vibration, synchrony between thoracic and abdominal heart beats was re-established (Fig. 13).

The new information which we have obtained concerning the effects of ambient temperature on pre-flight warm-up and post-flight cooling, the relative effectiveness of wing vibrating and wing beating in heat production, together with data on heart activity and regional differences of body temperature can be used to extend the understanding of the physiology of endothermy in large insects.

In the sphingids and satumiids that have been studied effective flight is possible only if the temperature of the thorax is at least 32 °C. The reasons for this have not been analysed in detail in moths. However, in isolated nerve-muscle preparations of Schistocerca it has been shown (Neville & Weis-Fogh, 1963) that at low temperatures the duration of the contractions of the upstroke and downstroke muscles is prolonged, causing them to work primarily against each other, rather than on the wings. The maximum mechanical energy for flight becomes available only when thoracic temperatures are high enough to ensure that the twitches of the antagonistic muscles no longer overlap significantly and so act mostly on the wings.

In M. sexta the rate of wing vibration during warm-up increases directly with thoracic temperature (Fig. 4), and the frequency of wing movements reaches that characteristic of normal flight, 25 – 27/sec (Heinrich, 1971a), at a thoracic temperature of about 37 °C which approximates the temperature at take-off (Fig. 3). Kammer (1968) has found that during warm-up the flight muscles of sphingids and other lepi-dopterans contract synchronously, which causes the wings to vibrate rather than beat. However, when resting M. sexta are held by the legs they invariably execute wing beats of large amplitude even at low thoracic temperatures. This wing flapping causes a rapid rise in thoracic temperature. The rate of wing flapping increases directly with thoracic temperature but averages slightly lower than the rate of wing vibration. During wing vibration the upstroke and downstroke muscles contract synchronously and work mostly against each other; during wing flapping they contract alternately and work mostly to move the wings. The rate of warm-up is about the same in both situations (Figs. 1 and 5). From these data we infer (1) that in the intact animal the timing of contraction of the antagonistic flight muscles can be either alternate or synchronous over a range of temperatures extending from at least 13 °C to 35 °C, and (2) that the amount of heat produced is very nearly the same whether the muscles are working against each other or are moving the wings.

Using the data for mean rates of warm-up of live moths and the mean rates of cooling of both dead and live moths it is possible to compute approximate upper and lower values for the rates of heat production during warm-up, if it is assumed that all the heat is produced in the thorax and the weight and specific heat of the thorax are known.

From Fig. 14 the following points are apparent:

1. In all cases the rate of heat production increases with increasing thoracic temperatures.

2. The calculated rates of heat production are higher when the cooling rates of live animals are used than when the cooling rates of dead ones are used, with this difference increasing directly with Δ T.

3. The apparent rate of heat production at any given body temperature is greater when T_A is 25 °C than when it is 15 °C.

On the basis of these relationships we can further analyse pre-flight warm-up and add to our understanding of the physiological processes involved.

It appears that heat is actively moved from the thorax to the abdomen, at least during the initial few minutes of cooling, by the circulation of blood. The fact that the rise in T_Ab is most conspicuous in the period during which T_Th is declining implies that something more than passive conduction is involved. From the data in Figs. 11 to 13 we have previously concluded that during warm-up the action of the abdominal heart is erratic and out of phase with the pulsations of the thoracic heart. From these data we infer that the movement of blood between abdomen and thorax during warm-up is restricted. This should sequester heat in the thorax, and is consistent with the observation that the abdomen remains at or near ambient temperature during warm-up. It has previously been shown (Heinrich, 1970) that during flight at high thoracic temperatures the movement of heat from thorax to abdomen by blood circulation plays an important role in thermoregulation in M. sexta. In Fig. 14 the difference between the upper and lower heat-production curves at both 15 and 25 °C is accounted for in part by the absence of circulation in the dead animals. This gives a rough indication of the amount of heat retained in the thorax by the minimizing of abdominal circulation during warm-up.

Adams and Heath (1964) suggested that the evaporation from a droplet of fluid extruded from the proboscis of a sphinx moth Pholus achemon could account for the accelerated cooling seen in this moth. It seems possible that the accelerated cooling could equally well be accounted for by the circulation of blood between thorax and abdomen. We did not at any time observe the extrusion of fluid droplets from the proboscis of M. sexta.

The difference between the calculated rates of heat production during warm-up at the same thoracic temperature when T_A = 15 °C and T_A = 25 °C is a measure of the heat lost by respiratory evaporation, and by forced convection due to the air movements associated with wing vibration. (Heat loss due to conduction, radiation and passive evaporation should be the same in freshly killed and in live moths and can therefore be ignored in our calculations of heat production.) Because the wings are motionless during cooling, the cooling curves used in the calculation of heat production include only the effects of passive convection. The heat loss due to forced convection should be a major factor in causing the difference between the heating rates at a given thoracic temperature but differing ambient temperatures. (We have shown in Fig. 4 that the frequency of wing-vibration - and by inference, heat production - are dependent on thoracic temperature and independent of ambient temperature.) For any given thoracic temperature the ΔT during warm-up from 15 °C is 10 °C greater, than the ΔT during warm-up at 25 °C. Both upper and lower limits of the calculated rates of heat production for the 15 °C and 25 °C sets of curves shown in Fig. 14 differ by about 0 · 8 cal/min. From this we conclude that during warm-up at any given thoracic temperature the heat loss due to forced convection and respiratory evaporation is about 0 · 08 cal/min/°C Δ T. Our data do not allow us to distinguish between heat loss due to evaporation and heat loss due to forced convection.

The sets of curves for heat production during warm-up at 15 °C and warm-up at 25 °C can be corrected for heat loss due to evaporation and forced convection by adding to them 0 · 08 cal/min/°C Δ T. The resulting corrected curves (Fig. 14) give our computation of the relation of rate of heat production to thoracic temperature during warm-up. From the set of corrected curves it appears that rate of heat production during warm-up increased 4 · 6 to 5 · 4 times between thoracic temperatures of 15 and 35 °C.

Extrapolation of the corrected rates of heat production during warm-up to a thoracic temperature of 41 °C, which is close to that normally maintained by M. sexta in free and unsupported flight, yields values between 4 · 35 and 5 · 10 cal/min. Assuming a weight of 1 · 5 g and that the utilization of 1 ml of O₂ yields 4 · 8 calories, these figures indicate an oxygen consumption of 36 · 3 – 42 · 6 ml O₂ (g h) ^{− 1}. These values are slightly below the 40 – 50 ml O₂ (g h) ^{− 1} measured during free flight at ambient temperatures between 15 and 25 °C (Heinrich, 1971a), which is to be expected because oxygen consumption is strongly correlated with the mechanical work required to remain airborne (Weis-Fogh, 1964; Heinrich, 1971a).

When M. sexta was stimulated to initial warm-up after having been allowed to come to equilibrium with ambient temperature, the rate of increase in thoracic temperature increased linearly with ambient temperature (Fig. 2). However, it has been reported in the sphingid Celerio lineata (Heath & Adams, 1967) and in the saturniid Hyalophora cecropia (Hanegan & Heath, 1970) that rates of warm-up are independent of ambient temperature. We do not know whether this discrepancy is caused by species differences or by differences in experimental methods.

The pre-flight warm-up of large moths shows physiological similarities to the performance of birds and mammals during arousal from the dormancy associated with daily torpor, hibernation, and aestivation. It has long been known that small heterothermic mammals warm up more rapidly than large ones, and a similar situation exists in heterothermic birds (Bartholomew et al. 1957; Lasiewski & Lasiewski, 1967). The generalization that rate of warm-up is inversely related to body weight can be extended to include the small sample of insects for which data are available (Fig. 15). For example, in an ambient temperature of 20 –25 °C a 100 g mammal warms up about 0 ·3 °C per minute while a 100 mg insect warms up almost 10 °C per minute. No regular relationship between body weight and rate of warm-up is apparent with the small sample of insects. The larger moths (numbers 7 to 10 in Fig. 15) warm up much more rapidly than do humming-birds and bats which are only slightly heavier.

This difference may be explained by the fact that the moths warm only the thorax whereas their rates of warm-up are plotted in terms of total body weight. Although regional differences in body temperature during warm-up are well known in large mammals, this situation is much less conspicuous in the smallest heterothermic mammals and birds.

This study was supported in part by U.S. National Science Foundation Grant GB-18744.