Tracheal Gases, Respiratory Gas Exchange, Body Temperature and Flight in Some Tropical Cicadas

Fidicina mannifera, a large (∼3g) cicada that occurs in the lowland tropics of central America, is able to take off and fly with a body temperature as low as 22 °C. This contrasts sharply with the situation in many beetles and most moths of similar mass, which require thoracic temperatures greater than 30 °C for flight (Heinrich, 1981).

The present study examines the oxygen consumption, body temperature, tracheal gas concentrations and flight characteristics of F. mannifera, as well as the energy metabolism of other cicadas at rest. Previously published research on the energetics and body temperatures of cicadas has been directed at the mechanism of sound production (Josephson & Young, 1979) and aspects of behavioural thermoregulation (Heath, Wilkin & Heath, 1972).

Our studies were carried out during July, 1982 at the Barro Colorado Island Station of the Smithsonian Tropical Research Institute in the Republic of Panama.

Cicadas were captured at night at lights, or during the day on the trunks and branches of trees, particularly Zanthoxylum panamense on which they commonly fed. The cicadas were weighed to the nearest mg and housed individually in ventilated plastic boxes which contained moistened paper and were kept in an air-conditioned room. Physiological measurements were made within 24 h of capture. Species identifications were confirmed by Dr Thomas E. Moore of the Museum of Zoology, University of Michigan, Ann Arbor where the specimens are deposited.

Experiments were carried out at 23·0-24·5 °C either in an air-conditioned room or a refrigerated incubator in which ambient temperature (T_a) could be controlled to within half a degree and which was equipped with a window, and a reach-through port.

All temperatures were measured with copper-constantan thermocouples connected to Bailey Bat thermometers. Thoracic temperature was measured with a 40 gauge thermocouple inserted dorsolaterally into the flight muscles to a depth of about 2 mm. The thermocouple was secured to the cuticle with beeswax. A shortened insect pin was inserted transversely through the heavy cuticle on the dorsal part of the caudal end of the thorax and secured with beeswax. A pipe cleaner was looped around the pin and served as a handle for positioning and lifting the cicada and for securing the thermocouple leads (Fig. 1).

Rates of oxygen consumption (⁠$V·O2$⁠) were measured with an Applied Electrochemistry S3-A Oxygen Analyzer. Measurements during fixed flapping and non-flapping warm-up were made in a flow-through system. Measurements of inactive cicadas were made in a closed system. In both systems the air was dried and the CO₂ removed before it was introduced into the oxygen sensor. Rates of air flow were measured with a flow meter calibrated against a Brooks mass flow meter.

The cicada was placed in a cylindrical Lucite respirometer chamber equipped with input and output manifolds to ensure adequate air mixing. The thermocouple and the pipe cleaner that were attached to the cicada were passed through a small hole in the respirometer lid. Air was drawn successively through respirometer chamber, drying train (Drierite), CO₂ absorbent (NaOH pellets), sensor, flowmeter and pump. Flow was controlled at 73 cm³ min⁻¹ by a needle valve. Tygon tubing was used throughout. The system was calibrated before each run. An AIM-65 microcomputer (Rockwell) controlled a multichannel switching and analogue-to-digital conversion system, recorded the outputs of the sensors and converted these voltages to temperature and to instantaneous rates of oxygen consumption.

Tracheal gases were sampled through a polyethylene tube 1 cm long (o.d. = 1 mm; i.d. = 0·5 mm) which was inserted into the dorsal air sac through a hole drilled in the cuticle. The tube was sealed to the cuticle with beeswax to ensure an airtight connection. The cicada was attached, as previously described, to a pipe cleaner, one end of which was embedded in a post of plasticine clay (Fig. 1). The cicada was orientated horizontally and given a foam rubber ball to hold so that it could make normal walking movements. A mercury-filled micropipette (Scholander & Evans, 1947) was inserted into the polyethylene tube. The plunger of the pipette was depressed, filling the polyethylene tube with a column of mercury that was continuous with that in the pipette. To obtain a gas sample the plunger was slowly withdrawn, pulling all of the mercury back into the pipette and also drawing 3-5 μd of air from the air sac into the tip of the pipette. A drop of water, slightly acidified with HO, was placed on the junction of the pipette and the tube. When the pipette was withdrawn from the tube a small quantity of the acidified water entered the tip, preventing exchange between room air and the gas sample. The sample was then stored behind mercury in a glass tube, one end of which was sealed. The pipette was reinserted into the polyethylene tube and the sampling process was repeated as desired. The concentrations of CO₂ and O₂ in the samples were determined as described in Scholander & Evans (1947) except that we used the reagents described by Scholander (1947). A series of 25 determinations of room air gave mean values of 20·88 % ± S.D. 0·09 for O₂ and 0·03 % ± S.D. 0·09 for CO₂.

Each cicada was weighed to the nearest mg. The volume of the abdominal air sac was measured by making a small slit in the ventral wall of the abdomen, carefully filling the air sac with water and then reweighing the animal. The cicada was then put in a 60 cm³ plastic syringe that was equipped with a valve and filled with water to which a detergent had been added. The air in the syringe was ejected, the valve was closed, and a vacuum pulled by hand to draw air out of the tracheae and the air sacs. This air was ejected from the syringe. The process was repeated until no more air could be obtained. The water-filled cicada was then blotted dry and weighed. The increase in the mass was used as a measure of the volume of the respiratory system The flight muscles were dissected out of the thorax, blotted, and then weighed to the nearest mg.

The left wings were removed and arranged on translucent paper in the spread position typical of flight, with front and hind wings interlocked. The outline of the wings was traced on opaque paper and cut out. The area of the cut-out was measured with a model L1-3000 Licor portable area meter and multiplied by two.

Cicadas with thermocouples implanted in the thorax were suspended above a microphone by a pipe cleaner secured to the thorax as previously described. By jiggling the cicada and directing a stream of air at it from a blower it was induced to flap its wings until exhausted while the sound of the wing beats was recorded with a tape recorder. The thoracic temperature was recorded verbally on the tape at intervals. Wing-beat frequency was determined from the recordings with a Kay Sonagraph.

The relationship between body mass and daytime resting $V·O2$ was investigated using three species of cicadas (Carineta viridicata Distant, Zammara smaragdula Walker andFidicina mannifera) ranging in mass from 0·2 to 3·0 g. Body temperatures did not differ from ambient (23–24·5 °C). The equation for the linear regression of log $V·O2$ (Y) on log mass (X) is Y = 0·633X^0·89 (r² = 0·89). The slope does not differ significantly from those of beetles and moths measured under similar conditions but the intercept is higher (see Discussion). Z. smaragdula and F. mannifera feed and call during daylight. Consequently, the values of $V·O2$ represent resting metabolism during a time of day when they were normally active.

F. mannifera is the largest of the cicadas on Barro Colorado Island. The males and females do not differ in mass, wing area or wing-loading (Table 1). Wing-loading increases with body mass in F. mannifera (Y = 0·104+0·056X; r² = 0·77). A similar relationship exists interspecifically (Fig. 2).

The flight muscles of F. mannifera constitute about 35 % of the total mass. The volume of the tracheal system is large, comprising about 41 % of the total body volume in females and 48 % in males. The abdominal air sac accounts for about half the volume of the tracheal system. Tracheal volume is significantly greater in females than in males (Table 2).

All of the F. mannifera we observed in the field, and at 22–24·5 °C in the laboratory, were capable of immediate flight when startled. However, in the laboratory, spontaneous take-offs (those in which the insect was not startled into immediate flight) were always preceded by non-flapping warm-up, which was accompanied by barely visible wing movements of low frequency (l–2s⁻¹). It was usually possible to induce warm-up in the restrained cicadas by brief, gentle prodding. Non-flapping warm-up was sometimes, but not always, followed by an attempt to fly. Mean T_th immediately before spontaneous take-off was 29·3 °C (S.D., 2·5; range, 25·4–33·0; N= 12). Some individuals that did not attempt to fly after warm-up maintained elevated T_th for 20 min or longer (Fig. 3).

Like the elevation in T_th, the increase in $V·O2$ during non-flapping warm-up we variable. Peak values of $V·O2$ (mean of the two highest readings during each warm-up) were about 16 times the resting rate at the same ambient temperature (Table 3).

F. mannifera apparently never makes prolonged flights. The flights we observed were usually no longer than a few tens of metres from one tree to another. We were never able to force them to remain airborne in enclosed areas for more than 100s. They readily attempted to fly while restrained (see Fig. 1) but would not flap continuously for more than 90s.

We measured instantaneous $V·O2$ and thoracic temperature during fixed flapping to exhaustion 11 times on five female and four male F. mannifera (Table 3). Both $V·O2$ and T_th rose rapidly as soon as wing-flapping began. T_th continued to increase after $V·O2$ had peaked and started to decrease. Neither reached equilibrium during the period of flapping. As soon as flapping stopped, $V·O2$ fell precipitously and T_th declined slowly and exponentially (Fig. 4).

The relationship between $V·O2$ and Tth differs during fixed flapping and post-flight cooling (Fig. 5). The factorial metabolic scope of a 2·84gF. mannifera, calculated as the mean peak $V·O2$ during flapping (Table 3) divided by the mean resting $V·O2$ (calculated from mean mass and the relation between $V·O2$ and mass), is 69.

By analogy with vertebrates, the rapidity of exhaustion in F. mannifera, suggested that flight might be supported anaerobically. We tested this possibility by exposing cicadas to pure nitrogen in a flow-through chamber and then attempting to induce flapping. The animals showed no overt response when the chamber was flushed with nitrogen and remained completely immobile when we attempted to induce flapping. When returned to a normal atmosphere, the cicadas could flap vigorously within less than 1 min.

We cooled F. mannifera in the temperature-controlled chamber while monitoring T_th with an implanted thermocouple. When the desired T_th was reached, the thermocouple leads were cut and the cicada was tossed into the air to a height of about 2 m. Even when T_th was as low as 16 °C, the cicadas attempted to fly. Cicadas which could not maintain altitude for a distance of 4 to 5 m were judged incapable of effective flight. The cut-off temperature for effective flight was quite sharp. In 15 trials on four cicadas, no successful flights occurred when Tth was below 21 °C. When T_th was 22°C or higher the cicadas always flew effectively.

When F. mannifera were motionless or walking, $V·O2$ in the thoracic air sacs remained near 17 % and $V·O2$ remained near 3 %. During non-flapping warm-up $V·O2$ fell to as low as 1 % and $V·O2$ rose to as high as 21 %. When wing-flapping began, both $V·O2$ and $V·O2$ quickly returned toward the resting levels, and reached about 16 % and 5%, respectively (Table 4). Immediately after wing-flapping ended, $V·O2$ decreased somewhat and $V·O2$ increased; both then slowly returned toward resting levels over a period of about 20 min. Changes in $V·O2$ and $V·O2$ were inverse and roughly symmetrical. However, $V·O2$ changed more rapidly than $V·O2$ during the first part of warm-up and immediately after flapping (Fig. 6).

Telescoping movement of the abdomen at 15-36 cycles min⁻¹ was characteristic of non-flapping warm-up. This pumping also occurred intermittently during and immediately after flapping, but was not necessary for flight. Two unrestrained specimens flew strongly with the abdomen immobilized with wax to prevent pumping and with the abdominal spiracles sealed. Moreover, sealing either the first or third pair of thoracic spiracles in addition to sealing the abdominal spiracles did not prevent flight. When only the second pair of thoracic spiracles was left open, the cicadas could only flap weakly.

Wing-beat frequency (f) was positively correlated with T_th between 25 and 34 °C (f = 19·47+0· 58Tth r² = 0·36). However, under the conditions of measurement, high T_th occurred only towards the end of flapping as the animals approached exhaustion. Thus, the relationship between frequency and temperature may have been partially offset by fatigue. Mean wing-beat frequency between 25 and 34 °C was 36·6 Hz (S.D., 8; N= 55 on seven specimens).

The slope of the regression of log resting $V·O2$ on log mass in cicadas is similar to the slopes reported for other orders of insects. However, at a temperature of 22-24 °C the one gram intercept for cicadas (0·633 ml h⁻¹) is substantially higher than that for either beetles (0·23 ml h⁻¹) or heterothermic moths (0·402 ml h⁻¹) (Bartholomew & Casey, 1977, 1978). By analogy with vertebrates, the relatively elevated $V·O2$ of the cicadas may be related to the fact that their $V·O2$ was measured during an active phase of their daily cycle, while that of the moths and many of the beetles was measured during their inactive phase.

As in beetles (Bartholomew & Heinrich, 1978) and moths (Bartholomew & Casey, 1978; Casey & Joos, 1983), wing-loading in cicadas increases with increasing body mass (Fig. 2). The wing-loading for F. mannifera is higher than for sphingids but lower than for beetles of similar size. Unlike beetles and sphingids of similar size, F. mannifera is not conspicuously endothermic, need not warm up prior to flight, and cannot remain airborne for long periods.

Pre-flight warm-up has not previously been reported in cicadas. F. mannifera can warm itself by endothermy but need not do so in order to fly, as long as its body temperature exceeds 22°C. Its ability to fly at Tth as low as 22°C, together with its diurnal habits and lowland tropical distribution means that it can usually take off without pre-flight warm-up. Nevertheless, in the laboratory F. mannifera always warmed up prior to spontaneous take-off. The dependence of wing-beat frequency on body temperature in F. mannifera is slight, but it is within the range reported for insects of other orders (May, 1981). A warm-up from 22 to 30 °C would increase wing beat frequency by about 15 %, which might be selectively advantageous if it resulted in increased flight speed or manoeuvrability. However, the functional significance of endogenous warm-up in F. mannifera remains problematic.

The peak $V·O2$ during warm-up in a 2·63 g F. mannifera (Table 3) is only 11·6% that of a sphinx moth of similar mass (Bartholomewet al. 1981). This difference may be related to relatively limited autoventilatory gas exchange during warm-up in cicadas (see below). Warm-up in cicadas is not accompanied by substantial wing movements, whereas sphinx moths vibrate the wings and thorax vigorously and at high frequency during warm-up and presumably are strongly autoventilated.

Cicadas exhaust quickly during flight, probably due to depletion of muscle stores of substrate for oxidative metabolism in the flight muscles. Exhaustion is evidently not due to any insufficiency in the rate of oxygen delivery. During flapping, F_O2 in the thoracic air sacs is high and relatively uniform (Table 4, Fig. 6). Post-flapping oxygen debt (Figs 4, 5) is negligible in comparison with post-flight oxygen debt in sphingid and saturniid moths (Bartholomew et al. 1981).

The rapidity of exhaustion during flight indicates that fuel reserves in the flight Mescles are small and that the rate of delivery of substrate to the muscles is low, relative to the rate of utilization during flight. In contrast, cicadas do not become exhausted during warm-up, although the total volume of oxygen consumed during a bout of warm-up exceeds that during a bout of flapping (Table 3). Warm-up is metabolically less intense, but more prolonged, than flapping. After warm-up the animals were capable of flights of normal duration. Thus, substrate was not depleted during warm-up and must have been delivered at least as fast as it was being used.

We did not measure the cost of free flight in cicadas. In the sphinx moth, Manduca, $V·O2$ during fixed flapping is less than half that during free flight (Heinrich, 1971). However, in F. mannifera, it appears unlikely that such a large difference exists between $V·O2$ during fixed flapping and free flight. The observed $V·O2$ of flapping cicadas in this study is about 70 % of that predicted for a free-flying sphingid of similar size. Moreover, the cicadas exhausted equally rapidly during fixed flapping and free flight, suggesting that $V·O2$ under the two conditions may be similar. However, measurements of $V·O2$ during free-flapping flight are needed.

We have calculated G_O2 for F. mannifera using the resting $V·O2$⁠, mean peak $V·O2$ during warm-up and during fixed flapping (Table 3), and the thoracic F_O2 (Table 4). G02 is about 0·007 ml min⁻¹ Torr⁻¹ in resting individuals and increases to about 0·032 ml min⁻¹ Torr⁻¹ during warm-up. During fixed flapping G_O2 increases by about 0·391 ml min⁻¹ Torr⁻¹, or about 50 times the resting value.

The moderate increase of G_O2 during warm-up relative to rest is presumably due to opening of the spiracles and possibly to ventilation by abdominal pumping. High F_CO2 and low F_O2 in the tracheal system during warm-up presumably induce maximal opening of the spiracles so that resistance to diffusion is minimized (cf. Burkett & Schneiderman, 1974). The larger increase in G_O2 during wing-flapping is attributable to autoventilation. If all of the difference in G_O2 between warm-up and flapping is due to autoventilation, then autoventilation accounts for about 92% of gas exchange during wing-flapping. Abdominal pumping often continues during and after flapping, but it is not critical for flight.

Ventilation and locomotor movements can be coupled in birds and mammals as well as in insects (Bramble & Carrier, 1983; Butler & Woakes, 1980; Weis-Fogh, 1967 In vertebrates, the mechanism and functional significance of this coupling have not been clearly demonstrated. However, in locusts and dragonflies, direct measurements have shown that the movements of the thorax during flapping flight ventilate the thorax at a rate sufficient to sustain flight without the support of any other ventilatory mechanism (Weis-Fogh, 1967). This also appears to be the case in F. mannifera.

The concentrations of O₂ and CO₂ in the thoracic air sacs of F. mannifera at rest and during fixed flapping are similar to values recorded by Weis-Fogh (1967) under similar conditions in the locust Schistocerca. For example, during tethered flapping in Schistocerca, F_O2 was 13·4% and F_CO2 was 5·7%. We are not aware of any previously published data on tracheal gas tensions in an insect during an episode of endothermy. The low F_O2 and high F_CO2 in F-mannifera during warm-up indicate that, in the absence of autoventilation, gas exchange imposes a ceiling on metabolic rate. Therefore, gas exchange may limit the rate of non-flapping warm-up. The rapid restoration of F_O2 and F_CO2 to near resting levels as soon as wing-flapping commenced, indicates that in F. mannifera autoventilation closely matches oxygen supply with demand during flight.