JOURNAL OF EXPERIMENTAL ZOOLOGY 284:119–129 (1999) Analysis of Free Forward Flight of Schistocerca gregaria Employing Telemetric Transmission of Muscle Potentials W. KUTSCH,* M. VAN DER WALL, AND H. FISCHER Fakultät für Biologie, Universität, 78457 Konstanz, Germany ABSTRACT The free flight of mature female Schistocerca gregaria loaded with a transmitter device has been video-monitored for short periods of time in the laboratory. Several flight parameters, such as body angles to horizontal and to flight path, ascent angle, and flight speed, have been analysed. These flight parameters and their interrelationships coincide with those revealed during steady free flight in the field (Baker et al. 1981. J Comp Physiol 141:233–237). Therefore, these short free flights in the laboratory represent free autonomous flight performance, but offer the possibility of controlling environmental conditions. During free flight, the firing patterns of four selected powerful depressor and elevator muscles of the fore- and hindwings were obtained by a transmitter device fixed to the prothorax of the locust. Based on the recruitment of individual motor units, a correlation of the muscle-specific activity with the aerodynamic flight parameters can be accomplished. Wingbeat frequency is positively correlated to ascent angle and to speed. The recruitment of the motor units of the forewing muscles shows a correlation with some of the flight parameters, whereas the hindwing muscles are activated in a rather immutable pattern. This argues for the importance of the forewings for maneuvering, whereas the hindwings mainly serve to provide the basic power output in flight. J. Exp. Zool. 284:119–129, 1999. © 1999 Wiley-Liss, Inc. Locust flight has been extensively analysed for its aerodynamic and physiological basis. Thus, two main approaches have been chosen: (1) analysis of freely flying locusts (e.g., Baker and Cooter, ’79; Baker et al., ’81; Kutsch and Stevenson, ’81) and (2) analysis of tethered flight (e.g., Weis-Fogh, ’56; Gewecke, ’75; Wortmann and Zarnack, ’93). However, even the most sophisticated tethered arrangement (e.g., Baader et al., ’92; Spork and Preiss, ’93) represents a rather static system missing the dynamics characterizing the natural closed loop situation, in which the sensory inflow is unimpeded allowing an unbiased sensorimotor integration. Undoubtedly, flight behaviour is influenced by tethering (Zarnack and Wortmann, ’89). For example, one of the consequences of such an open loop paradigm (loss of adequate sensory feedback) is that the wing-beat frequency is lower in tethered compared to free flight (Baker et al., ’81; Kutsch and Stevenson, ’81). Even a recommendation to study flight by suspending locusts “freely” by wires (Wootten and Sawyer, ’54; Möhl, ’88) does not overcome the restriction impeded by the wires’ length. In a further advance, a telemetric transmitter mounted to the locust has been developed (Kutsch et al., ’93) which allows the recording of physiological signals from freely flying animals. This © 1999 WILEY-LISS, INC. technique was used here to understand the recruitment of selected muscles (depressor and elevator muscles of both wing pairs) during forward flight (horizontal, ascent, descent). Therefore, the classical work of Wilson and Weis-Fogh (’62), then employing tethered locusts, is extended to freely flying locusts. For the first time we can correlate aerodynamic parameters with motor patterns for a better understanding of the neuronal control underlying locust flight. This advance allows us to study dynamic processes underlying natural behaviour. MATERIALS AND METHODS Animals and flight conditions For our experiments, female desert locusts, Schistocerca gregaria Forskål (adult age 20–25 days) were chosen, selected from a crowded culture (temperature, 42°C; rel. humidity, c. 75%; light-dark cycle, 12 hr–12 hr). Immediately before the flight test, the animals were separated in small perspex cages and warmed to T = 35°C. Grant sponsor: DFG; Grant number: Ku 240/17-1. *Correspondence to: W. Kutsch, Fakultät für Biologie, Universität, 78457 Konstanz, Germany. Received 6 May 1998; Accepted 5 October 1998. 120 W. KUTSCH ET AL. The dimensions of the flight room were: length 10 m, width 5.7 m, height 2.2 m. The room was illuminated by several rows of light tubes along the ceiling with a mean light intensity of c. 1400 lux. The room temperature was 28–30°C and relative humidity was 65–70%. Flight activity of the locust was induced by accelerating them by hand with a rather constant launching angle (c. 35°) and an initial speed (v) of c. 4 m/s. (For Schistocerca some other values for take-off velocity have been published: v = 2.5 m/s [Katz and Gosline, ’93] or v = 3.2 m/s [Bennett-Clark, ’75]; however, in our lab, take-off velocities, following a jump, ranged from v = 4.0–4.4 m/s [Fischer, ’94].) A stationary video-camera was placed at 5 m beyond the start point, with a distance to the opposite wall of 4 m. To avoid possible wall effects (Stevenson et al., ’95) only forward flights within a corridor of 8 × 2 × 2 m were analysed. The focus width of the camera covered an area of 2.8 × 2.0 m. The distance from the start point to that of entering the camera focus was sufficient to enable free, autonomous flight performance. The locusts could fly 2 m further ahead after leaving the camera focus. Telemetric device The mass of the transmitter device, including the battery, was m = 0.5 g, which is equivalent to c. 20% of the locust’s mass. This additional load does not impede the flight parameters (Fischer et al., ’97). The transmitter was fixed to the pronotum by a beeswax-resin mixture. Each wing is driven by about 12 muscles (Snodgrass, ’29; Albrecht, ’53). From these, we recorded the activity of the powerful direct depressor muscles (M99, M129) and indirect elevator muscles (M83/ M84 and M113) of the fore- and hindwing, respectively. The electrode (Manganin, Isabellenhütte, Dillenburg; insulated to the tip, diameter 50 µm) was inserted into the chosen muscle, and the reference electrode was inserted into the abdominal haemocoel. Both electrodes were connected to the modulator part of the transmitter (Kutsch et al., ’93). The telemetric electromyograms (EMGs) were detected by a HB9CV aerial and demodulated by a wide-band receiver system (for details, see Kutsch et al., ’93). Data acquisition and analysis The EMGs were displayed on an ocilloscope and stored on tape. In parallel, the oscilloscope trace was filmed by a second video camera. Both the recordings of the flight (see above) and the oscilloscope camera were synchronously stored online on a video mastertape. The frame frequency of both cameras was 50 Hz (shutter opening time 1 ms), giving a frame interval of 20 ms. The video mastertapes were analysed frame by frame (Panasonic AG 7380 recorder, Sony Trinitron monitor). The silhouettes of the animals were successively drawn onto clear overhead films. Using conventional analysis (e.g., Baker and Cooter, ’79; Kutsch and Stevenson, ’81) several flight parameters (see Fig. 1) could be determined. The locust’s centre of gravity is midpoint between front and abdominal tip (Weis-Fogh, ’56), which was not shifted significantly by the transmitter (Fischer, ’94). All flight parameters are considered to be “instantaneous” with respect to the given time interval of 20 ms. The video mastertapes allowed for the correlation of the instantaneous flight parameters with EMGs. First of all, the cycle length (or its inverse, the instantaneous wing-beat frequency, Hz) can be determined. Given a mean cycle length (tc) of 43 ms (see Results), there are 2 or 3 individual frames related to each wing-beat cycle; therefore, for a correlation of the optical data and cycle length, the instantaneous flight parameters had to be averaged and associated to the concomitant muscle cycle. The size of the viewing field allowed an analysis of up to 15 consecutive wing cycles. By a precise evaluation of the EMGs, the recruitment of motor units can be assessed, taking into account the parameters of signal form, relative amplitude, and possible changes in relative size within a burst (see Wilson and Weis-Fogh, ’62 and Fig. 2). Flight parameters The following flight parameters were analysed (compare with Baker et al., ’81): ascent angle (α, [°]); body angle to horizontal (ϕ, [°]); body angle to flight path (δ, [°]); flight speed along the flight path (v, [m/s]) (see Fig. 1). Altogether, 335 flights of 75 animals were investigated and the parameters were correlated to each other. Statistics Data analysis and conventional statistical procedures (e.g., Sachs, ’78) were employed. They were aided by the computer programs Origin 3.0 and SAS for Windows. All data were tested for normal distribution (P < 0.01) and homogeneity. Analysis of correlation and linear regression was based on the least square method. Binominal distribution and their confidence levels concerning relative occurrence have been determined after EMGS IN FREELY FLYING LOCUSTS Fig. 1. Determination of flight parameters by analysing consecutive video-frames. t, cumulative time interval (∆t = 20 ms) of successive frames; S, translatory path of the centre of gravity, indicated by a black dot; α, ascent angle: angle between horizontal plane and translatory path of S; ϕ, body angle to horizontal: angle between horizontal plane and lon- 121 gitudinal axis of the locust; δ, body angle to flight path, calculated by the difference between ϕ and α; v, flight speed: instantaneous translatory speed of the centre of gravity within two frames; SL, silhouette length of the animal, the mean natural length of the female population is 54 mm. Clopper and Pearson (’34). Comparisons between data were carried out by a one-tailed test. RESULTS Analysis of flight parameters Fig. 2. Criteria for the discrimination of motor unit recruitment, derived from telemetric EMGs of M99. AI, AII: Sequences of a continuous flight showing the activity of the first unit (1), double discharges of this unit (1,1′), and the additional recruitment of the second unit (2). For this example, the second unit is seen to either add up to the after discharge of the first unit (1′/2) or to overlap completely with the first single spike (1/2, see last potential). B: Sequence from a different animal, exhibiting constant recruitment of both motor units (1,2). The ascent angle (α) varied from –20° to almost 40°, whereas for the body angle to horizontal (ϕ) we found values spanning from –20 to 60° (Fig. 3A). The ascent angle (α) was positively correlated with body angle to horizontal (ϕ) (Fig. 3A). On average, an increase of the ascent angle of 10 degrees was paralleled by a 9 degree increase of ϕ. During horizontal flight (α ± 0 degrees) the body angle (ϕ) was in the range of 16 degrees, during descent negative body angles to horizontal were observed. To understand whether the position of the animal’s longitudinal axis changed in relation to the general flight direction (translatory movement of its centre of gravity), the ascent angle (α) and body angle to flight path (ϕ) had to be compared (Fig. 3B). The negative regression line indicated that the steeper the animal’s ascent, the more congruent was its longitudinal axis to the flown flight path. A significant relationship was found when com- 122 W. KUTSCH ET AL. suppose that the general flight speed represents a “scalar” flight parameter independent of the ascent angle (α). The flight speed (v) of the animals was in the range of 1.8 to 6.6 m/s (mean ± SD: v = 3.7 ± 1.0 m/s). Correlation of flight parameters and motor pattern Altogether 1,610 wing-beat periods (tc) have been analysed. The mean cycle length (±SD) was 43.4 ± 3.2 ms. This corresponded to a mean wingbeat frequency of 23 Hz (range: 18.2 to 31.3 Hz). Analysis of correlation and linear regression was performed for paired values: tc vs. α and tc vs. v (Fig. 4). The observed cycle length (tc) was significantly reduced with increasing ascent (α). Depending on the regression, the wing-beat frequency (Fig. 4A) increased by c. 0.8 Hz with each degree of ascent (α). At horizontal flight, wingbeat frequency was in the range of 22 Hz. When flight speed was related to cycle length (tc), then a decrease of tc was correlated with a reduction in speed (Fig. 4B). A recalculation of cycle length to wing-beat frequency indicated that flight speed increased by 0.1 m/s when wing-beat frequency increased by 1 Hz. Recruitment of muscle units during flight Fig. 3. Relationship between (A) ascent angle (α) and body angle to horizontal (ϕ), (B) ascent angle (α) and body angle to flight path (δ), and (C) ascent angle (α) and flight speed (v); n = number of samples. All correlation coefficients (r) are significant (P < 0.05). paring instantaneous flight speed (v) to the ascent angle (α) (Fig. 3C). A linear regression indicated a reduction of the speed by 0.04 m/s when the ascent angle (α) changed by 10 degrees. Due to this minimal change in speed related with a wide change in α we assume that the statistically shown relationship is of low importance in free flight; we rather An interpretation of the relevant EMGs corroborates previous data regarding the number of motor units involved in flight (e.g., Wilson and Weis-Fogh, ’62; Tyrer and Altman, ’74; Zarnack and Möhl, ’77). For both depressor muscles (M99, M129) two (fast) motor units are reported. For the elevator muscle of the hindwing (M113) only one unit is apparent. Both forewing elevator muscles (M83/M84) consist of only one unit, and since a difference in the recruitment of both muscles was not seen, the activity of these two elevator muscles is lumped together. To allow a correlation of muscle recruitment with the kinematic flight parameters (see above), data for the ascent angle (α) were grouped into the following classes: –15° to –6° → descent, –5° to 4° → horizontal flight, 5° to 14° → slight ascent, above 15° → steep ascent (see Fig. 6). Since data analysis of EMGs vs. body angle to horizontal (ϕ) revealed similar trends to that shown for the ascent angle (α) (see also the positive correlation of α vs. ϕ, Fig. 3A), we did not duplicate our analysis for ϕ. Flight speed (v) and cycle duration (tc) were also grouped into four and five classes, respectively, by partitioning the observed range of data into equal subdivisions (see Figs. EMGS IN FREELY FLYING LOCUSTS Fig. 4. Relationship between (A) cycle length (tc) and ascent angle (α) and (B) cycle length (tc) and speed (v), accumulated for all muscles. Both flight parameters decrease with increasing cycle length; n = samples. Both correlation coefficients (r) are significant (P < 0.05). 7, 8). The activation of the different muscle units as well as their mode of recruitment is given as their relative occurrence within the relevant classes. Complete omission of potentials for individual wing-beat cycles within a longer series for any complete total occurred only rarely, for such a case an effect upon the instantaneous flight performance was not obvious. Recruitment of M99 (forewing depressor) During flight both motor units of this muscle (also known as subalar muscle) could be recruited (see Figs. 2, 5A), which, based on their temporal recruitment, were termed first and second unit. Firing of these two units depended on the flight path. During descent flight (instantaneous values –15°< α <–5°), both motor units were activated (Fig. 5A) with a slight preponderance of the first unit (Fig. 6A). Thereby the following modes of activation became obvious with almost equal prob- 123 Fig. 5. Examples of telemetric EMGs of two forewing muscles, right side (r): depressor (M99r) (A, B) and elevator (M83/M84r) (C, D). For both muscles a change of muscle activity is correlated with different ascent angles. A: Recruitment of both motor units, varying from close succession to complete overlap; flight data: α = –6°, v = 3.6 m/s, f = 24 Hz. B: Same animal as A, exhibiting only single spikes of one unit with steep ascent; flight data: α = 16°, v = 3.2 m/s, f = 24 Hz. C: Single firing of one unit of this muscle complex; flight data: α = 8°, v = 3.5 m/s, f = 22 Hz. D: Same animal as C, exhibiting an afterdischarge per each cycle with steep ascent; flight data: α = 19°, v = 3.0 m/s, f = 24 Hz. ability (Fig. 6B): only the first unit was recruited with a single spike; the first unit discharged twice; both units fired once whereby both could summate completely. Occasionally the first unit showed a double discharge together with a single activation of the second unit. With a transition from horizontal to steep flight path (α > 15°), the relative activation of the second unit diminished (Fig. 6A). This was paralleled by a reduction in double discharges of the first unit. Therefore, a steep ascent was usually characterized by a single discharge of the first unit of M99 (Figs. 5B, 6B). There was no significant change in the activation of the forewing subalar muscle in correlation with flight speed (Fig. 7). For all classes, at least when above 2.5 m/s, both motor units were recruited, with the typical preponderance of the first unit (Fig. 7A). Thereby, the mode of activation, whether employing both motor units or after-discharges of the first unit, did not show a significant change with increasing flight speed (Fig. 7B). For the lowest instantaneous flight speed class (v < 2.5 m/s), we registered only a small number of values (n = 10); therefore, the apparent reduction 124 W. KUTSCH ET AL. Fig. 6. Analysis of motor unit recruitment of the forewing depressor muscle (M99) in relation to ascent angle (α). A: Indication of the overall activity of both motor units. B: Individual recruitment of both units. The number (n) of values per class, the relative occurrence per class and the 95% confidence interval (bars) (P < 0.05) are shown; * indication of groups which are not significantly different (P > 0.05). (This presentation of data is valid also for Figs. 7, 8, 10, 11.) in repetitive discharges for low speeds was of no statistical significance. Similarly, no trend was observed in the recruitment of the subalar’s motor units with respect to cycle length (tc) (Fig. 8). There was a more or less constant relation in the activity of both units with a superiority of the first unit (Fig. 8A). The grade of activation of both units remained almost constant independent of the cycle length (Fig. 8B). Therefore, wing-beat frequency and grade of subalar recruitment were not related to each other. Recruitment of M129 (hindwing depressor) Due to the size of this muscle, also known as the hindwing subalar muscle, it became difficult to separate reliably the two known motor units in a common EMG. We therefore, restricted our analysis to the recruitment of the rostral unit, be- Fig. 7. Analysis of motor unit recruitment of M99 in relation to instantaneous flight speed (v). There is a preponderance of the first motor unit, and there is no change of motor unit recruitment related to a change of speed. For the presentation of data, see Fig. 6. cause this unit is reported to be more active than the caudal one (Stevenson, ’88). The rostral unit fired mostly once per cycle (Fig. 9A, B), while double firing was rare (mean: 15%) (Fig. 10). Apparently, each animal had its “individual” activation pattern exhibiting either single or double firings, likewise covering the broad spectrum of flight parameters. This is in agreement with the independence of the mode of activation of these motor units versus flight kinematics (Fig. 10). Recruitment of M83/M84 (forewing elevators) Each muscle of this elevator group, M83/M84, comprised only one motor unit (see Fig. 5C, D) for which a tight coupling was reported (Baker, ’79b). Our recordings allowed the interpretation of only one unit, prohibiting an association to either of the two muscles. When the ascent angle (α) changed from descending to a steeply ascending flight, the overall relation of single vs. double discharges changed (Fig. 11A). While during de- EMGS IN FREELY FLYING LOCUSTS Fig. 8. Analysis of motor unit recruitment of M99 in relation to cycle length (tc). The first unit is activated with a significantly higher relative occurrence compared to the second unit (P < 0.05), except for the small group representing a short cycle length (30–34 ms). For the presentation of data, see Fig. 6. scent, a single spike usually occurred (Fig. 5C), and when the locust steeply ascended, there was a strong preponderance of double discharges (Fig. 5D). When the activity of this muscle group was correlated with speed (v), the relative occurrence of single and double discharges did not alter significantly (Fig. 11B). However, muscle recruitment changed when considering cycle length. At short wing-beat cycles (tc) there was a preponderance of two spikes, whereas longer cycles were almost exclusively correlated with single excitations (Fig. 11C). Recruitment of M113 (hindwing elevator) Muscle M113 consists of only one unit. This unit discharged usually only once per wing beat. Consequently, there was no correlation between the recruitment of this motor unit and the instantaneous flight parameters (e.g., see changes in ascent angle, Fig. 9C, D). 125 Fig. 9. Examples of telemetric EMGs of two hindwing muscles, left side (l): depressor (M129l) (A, B) and elevator (M113l) (C, D). For both muscles a constant (single) firing of a motor unit is exhibited independent of ascent angle. Flight parameters: A, α = –2°, v = 4.5 m/s, f = 21 Hz; B, α = 12°, v = 4.2 m/s, f = 20.5 Hz (same animal as A); C, α = –8°, v = 4.4 m/s, f = 24.5 Hz; D, α = 20°, v = 4.5 m/s, f = 25 Hz (same animal as C). DISCUSSION Flight parameters Only a few reports are available in which flight parameters of freely flying individual locusts have been analysed. For S. gregaria, Waloff (’72) determined mean airspeeds ranging from 3.2 to 8.1 m/s, whereas for freely flying L. migratoria in the field, Baker et al. (’81) gave a distribution of the flight speed spanning 3.3 to 6.1 m/s. In the present study of freely flying S. gregaria under laboratory conditions, the range of flight speeds coincides with that observed in the field. As observed for L. migratoria (Baker et al., ’81) there is a positive correlation of ascent angle (α) with body angle to horizontal (ϕ). Furthermore, the body angle to flight path (δ) is negatively correlated with ascent angle (α), strengthening a slight tendency as mentioned by Baker et al. (’81). Whether this body adjustment is passive or actively controlled by changes in wing-stroke parameters needs to be investigated. Altogether, our restricted conditions representing short flights in the laboratory (analysis of flights over a distance 126 W. KUTSCH ET AL. Fig. 10. Analysis of the recruitment of the rostral motor unit of the hindwing depressor muscle (M129) in relation to A, ascent angle (α); B, speed (v); and C, cycle length (tc). For this motor unit a correlation of its recruitment with any of the flight parameters is not obvious. For the presentation of data, see Fig. 6. of 2–2.5 m, within about 1 sec after release, low light intensity) do not result in a completely changed flight performance when compared to a flight in the field. Cycle length (wing-beat frequency) and its correlation to other flight parameters Wing-beat frequency in locusts has often been determined under different conditions. For teth- Fig. 11. Analysis of motor unit recruitment of the forewing elevator muscles (M83/M84) in relation to A, ascent angle (α); B, speed (v); and C, cycle length (tc). There is an intensification of double discharges correlated with an increment of the ascent angle (α), no correlation of muscle recruitment related with speed (v), and a decrease of the relative occurrence of double discharges with lengthening of the wingbeat cycle length (tc). For the presentation of data, see Fig. 6. ered female S. gregaria a value of 16.8 Hz is given (Weis-Fogh, ’56). Furthermore, it is known that wing-beat frequency decreases after flight start, at least in tethered flight (Weis-Fogh, ’56; Gewecke and Kutsch, ’79) and the wing-beat frequency is higher in free flight compared to EMGS IN FREELY FLYING LOCUSTS tethered flight (Baker et al., ’81; Kutsch and Stevenson, ’87). Therefore, the presently measured mean wing-beat frequency of 23 Hz for female S. gregaria seems to represent the frequency of free flight soon after take-off. The major aspects of wing movements are the production of two dynamic forces, lift and thrust. While Baker et al. (’81) did not find a correlation of wing-beat frequency with any of the flight variables (except flight speed, see below), the present investigation points to a positive correlation of wing-beat frequency with ascent angle (α). This observation corresponds with the notion that in tethered flight lift and wing-beat frequency are positively correlated (Weis-Fogh, ’56; Kutsch and Gewecke, ’79). However, under tethered flight conditions, animals produced a lift carrying only about 70% of their own body weight (Gewecke,’75; Kutsch and Gewecke, ’79; see Zarnack and Wortmann, ’89). Extrapolating the previously found relationship of wing-beat frequency to lift, the body weight is almost fully compensated by the higher wing-beat frequency in free flight—almost 23 Hz independent of additional loading; however, when additional loading surpasses a value of about 20% of the animal’s weight, active flight (horizontal or ascent) is impeded (unpublished results). Probably in combination with other wing-stroke parameters, increase in wing-beat frequency will result in a steeper ascent. A positive correlation of wing-beat frequency with speed agrees with previous results for tethered (Gewecke, ’75) and natural flight (Baker et al., ’81). The agreement of our present findings with those of previous investigators, therefore, allows us to consider that the flight performance of Schistocerca carrying a transmitter is comparable to that of a freely flying animal. Recruitment of muscles Wing mechanics controls power production in flight (e.g., Jensen, ’56; Brodsky, ’94). Previous studies of tethered locusts have shown that by varying certain wing-stroke parameters, flight performance can be adjusted (e.g., Weis-Fogh, ’56; Gewecke, ’75; Zarnack, ’83, ’88; Waldmann and Zarnack, ’88). Modulation of wing-stroke parameters is also the main principle for flight steering, when the theory of unsteady aerodynamics is applied to insect flight (Send ’92; Dickinson and Götz, ’93; Zarnack and Send, ’94; see also Ellington et al., ’96). The present advance allows us to study muscle activities accompanying free adjustment of flight parameters previously impeded by the tethered 127 approach. Hence, it becomes possible to extend the study of insect flight from a rather “steady” to a “dynamic” approach. Aerodynamic forces can be adjusted by wing-beat frequency and/or by modulation of the angle of attack, stroke amplitude, and stroke plane angle. During flight both motor units of the forewing depressor (M99) are usually recruited but the first unit is more active than the second one (Figs. 6–8). However, there is no change in the relative activation of the two motor units, whether single or double discharges, with respect to flight speed (Fig. 7). Speed-related activity changes are also not observed for the identified (rostral) unit of the hindwing depressor (M129) (Fig. 10B), the forewing elevator muscle complex (M83/M84; Fig. 11B), or the hindwing elevator (M113). Modification of flight speed, therefore, appears not to depend on a modification of the recruitment of the motor units serving these important flight muscles. Consequently, the higher speed, generated at start (e.g., Kutsch and Gewecke, ’79) cannot be explained by a higher activity of the (depressor) muscles (e.g., Wilson and Weis-Fogh, ’62; Gettrup, ’66). In free flight, cycle length and recruitment of the different motor units seem not to be coupled (except M83/M84). This is in agreement with a previous statement for tethered flight (Wilson and Weis-Fogh, ’62; for the forewing subalar muscle), whereas Waldron (’67) found both a positive and a negative correlation of the two parameters, depending on the stimulus conditions. Only in the forewing elevators (M83/M84) is a decrease in activity correlated with a lengthening of the wingbeat cycle (Fig. 11C). Therefore, in free forward flight no specific neuronal mechanism is involved to control the grade of muscle recruitment during different wing-beat frequencies. In tethered flight, body angle to horizontal (ϕ) is correlated with the angle of attack of the beating wings (e.g., Weis-Fogh, ’56; Wortmann and Zarnack, ’93). Measurements of the ascent angle (α), however, require a free flight approach. A higher angle of attack of the wings, equivalent to an enlarged pronation with respect to the animal’s flight path, must result in a growing lift production (Jensen, ’56). Assuming conditions are otherwise constant, this will cause a progressive ascent angle. Among other kinematic parameters not yet studied, such as stroke amplitude and stroke plane, muscle activity controls the angle of attack. The pleuroalar muscle (M85/M114) has been argued to influence the angular setting of the wing (Pfau and Nachtigall, ’81; Wolf, ’90) during 128 W. KUTSCH ET AL. tethered flight; whether this muscle functions similarly in free flight requires further study. In tethered flight higher lift production seems to depend mainly on the activity of the forewing basalar muscles (M97/M98) in relation to the ipsisegmental subalar muscle (M99); for the hindwings (representing 67.5% of the total wing area; Weis-Fogh, ’52)—which cover almost 70% of the produced lift (Jensen, ’56)—such differences in the recruitment of both segmental depressors are not apparent (see Thüring, ’86; Zarnack, ’88). In the present study of free flight, there is a clear correlation of the recruitment of the two motor units of the forewing subalar muscle (M99) with respect to the flight angles, whether it be ascent angle (α) (Fig. 6) or body angle to horizontal (ϕ) (not shown here). Gettrup (’66) associated a reduction of M99 activity with an effect on relative lift. Given the original idea that lift is dependent on the wings’ pronation/supination during downstroke (Wilson, ’62; Wilson and Weis-Fogh, ’62), the present result indicates that a steeper ascent (more lift) may be associated with decreased supination (increased pronation). Because both angles (α and ϕ) are positively related to each other (Fig. 3A), it becomes difficult to decide whether wing kinematic during ascent requires an active regulation of body position or whether this angle is passive. The presently applied video monitoring did not give detailed information with regard to an analysis of wing position, stroke angle, or stroke plane. Such measurements are urgently needed to fully understand the aerodynamics of freely flying locusts and its underlying neuronal control. There is no unequivocal trend in recruitment of the hindwing subalar muscle (M129) units in relation to flight angles. This conforms to the observation that pronation/ supination is more pronounced in forewings than in hindwings (Baker, ’79a; see also Möhl and Zarnack, ’77). In regard to the elevators, Wilson and WeisFogh (’62) mentioned a variable activity depending on lift, while Zarnack (’88) reports a switching off of M83 resulting in a reduction of upstroke. Our study of free flight shows that the recruitment of the forewing elevator group (M83/M84) is strengthened with steeper ascent (Fig. 11A). How the total power output can be effected by this muscle group is unknown, since wing elevation in tethered flight is only 20 % of the amount required for carrying the body weight (Jensen, ’56). No consistent change of the equivalent hindwing muscle (M113) was observed with respect to the instantaneous flight parameters. In free forward flight the chosen flight muscles are routinely recruited independent of the wingbeat frequency or flight speed. However, the flight path (horizontal, ascent, descent) seems to be mainly regulated by the activity of the forewing muscles. But for a complete description of the motor pattern, which must also consider rolling and yawing maneuvers, other muscles have to be tested. And not only the relative strength of muscle recruitment but also the relative timing of the muscles (e.g., Zarnack and Möhl, ’77; Möhl and Zarnack, ’77; Thüring, ’86; Waldmann and Zarnack, ’88) need consideration. We hope that this can be tackled by employing a newly developed 2-channel recording unit (Fischer et al., ’96). ACKNOWLEDGMENTS We thank Dipl.-Ing. H. Kautz for technical assistance. The helpful suggestions of two anonymous referees are appreciated. LITERATURE CITED Albrecht FO. 1953. The anatomy of the migratory locust. London: Athlone University Press. Baader A, Schäfer M, Rowell CHF. 1992. The perception of the visual flow field by flying locusts: a behavioural neuronal analysis. J Exp Biol 165:137–160. Baker PS. 1979a. The wing movements of flying locusts during steering behaviour. J Comp Physiol 131:49–58. Baker PS. 1979b. The role of forewing muscles in the control of direction in flying locusts. J Comp Physiol 131:59–66. Baker PS, Cooter RJ. 1979. The natural flight of the migratory locust, Locusta migratoria L: I. wing movements. J Comp Physiol 131:79–87. Baker PS, Gewecke M, Cooter RJ. 1981. The natural flight of the migratory locust, Locusta migratoria L: III. wing-beat frequency, flight speed and attitude. J Comp Physiol 141:233–237. Bennett-Clark HC. 1975. The energetics of the jump of the locust Schistocerca gregaria. J Exp Biol 63:53–83. Brodsky AK. 1994. The evolution of insect flight. Oxford: Oxford University Press. Clopper CJ, Pearson ES. 1934. The use of confidence or fiducial limits illustrated in the case of the binomial. Biometrika 26:404–413. Dickinson MH, Götz KG. 1993. Unsteady aerodynamic performance of model wings at low Reynolds numbers. J Exp Biol 174:45–64. Ellington CP, van den Berg C, Willmott AP, Thomas ALR. 1996. Leading-edge vortices in insect flight. Nature 384:626–630. Fischer H. 1994. Analyse des freien Fluges der Wanderheu schrecke Schistocerca gregaria mit optischen und radio telemetrischen Methoden. Diploma thesis. Konstanz: University Konstanz. Fischer H, Ebert E, Salva J, Kutsch W. 1997. Adjustments of the free flight system of desert locust (Schistocerca gregaria) to changes in body mass. In: Elsner N, Wässle H, editors. EMGS IN FREELY FLYING LOCUSTS Göttingen neurobiology report. Proc 25th Göttingen Neurobiol Conf., vol II. Stuttgart: Thieme. p 270. Fischer H, Kautz H, Kutsch W. 1996. A radiotelemetric 2channel unit for transmission of muscle potentials during free flight of the desert locust, Schistocerca gregaria. J Neurosci Methods 64:39–45. Gettrup E. 1966. Sensory regulation of wing twisting in locusts. J Exp Biol 44:1–16. Gewecke M. 1975. The influence of the air-current sense organs on the flight behaviour of Locusta migratoria. J Comp Physiol 103:79–95. Gewecke M, Kutsch W. 1979. Development of flight behaviour in maturing adults of Locusta migratoria: I. flight performance and wing-stroke parameters. J Insect Physiol 25:249–253. Jensen M. 1956. Biology and physics of locust flight: III. the aerodynamics of locust flight. Philos Trans R Soc Lond Biol 239:511–552. Katz SL, Gosline JM. 1993. Ontogenetic scaling of jump performance in the African desert locust (Schistocerca gregaria). J Exp Biol 177:81–111. Kutsch W, Gewecke M. 1979. Development of flight behaviour in maturing adults of Locusta migratoria: II. aerodynamic parameters. J Insect Physiol 25:299–304. Kutsch W, Schwarz G, Fischer H, Kautz H. 1993. Wireless transmission of muscle potentials during free flight of a locust. J Exp Biol 185:367–373. Kutsch W, Stevenson P. 1981. Time-correlated flights of juvenile and mature locusts: a comparison between free and tethered animals. J Insect Physiol 27:455–459. Möhl B. 1988. Short-term learning during flight control in Locusta migratoria. J Comp Physiol 163:803–812. Möhl B, Zarnack W. 1977. Activity of direct down-stroke flight muscles of Locusta migratoria (L.) during steering behaviour in flight: II. dynamics of the time shift and changes in burst length. J Comp Physiol 118:235–247. Pfau HK, Nachtigall W. 1981. Der Vorderflügel groβer Heuschrecken als Luftkrafterzeuger. II. Zusammenspiel von Muskeln und Gelenkmechanik bei der Einstellung der Flügelgeometrie. J Comp Physiol 148:135–140. Sachs L. 1978. Angewandte Statistik. Berlin: Springer. Send W. 1992. The mean power of forces and moments in unsteady aerodynamics. Z Angew Math Mech 72:113–132. Snodgrass RE. 1929. The thoracic mechanism of a grasshopper, and its antecedents. Smithson Misc Coll 82:1–122. Spork P, Preiss R. 1993. Control of flight by means of lateral visual stimuli in gregarious desert locusts, Schistocerca gregaria. Physiol Entom 18:195–203. Stevenson P. 1988. Central nervous aspects of flight pattern generation and its ontogenesis in the migratory locust. PhD thesis. Konstanz: University Konstanz. 129 Stevenson RD, Corbo K, Baca LB, Le QD. 1995. Cage size and flight speed of the tobacco hawkmoth Manduca sexta. J Exp Biol 198:1665–1672. Thüring DA. 1986. Variability of motor output during flight steering in locusts. J Comp Physiol [A] 158:653–664. Tyrer NM, Altman JS. 1974. Motor and sensory flight neurones in a locust demonstrated using cobalt chloride. J Comp Neurol 157:117–138. Waldmann B, Zarnack W. 1988. Forewing movements and motor activity during roll manoeuvers in flying desert locusts. Biol Cybern 59:325–335. Waldron I. 1967. Neural mechanisms by which controlling inputs influence motor output in the flying locust. J Exp Biol 47:213–228. Waloff Z. 1972. Observations on the airspeeds of freely flying locusts. Anim Behav 20:367–372. Weis-Fogh T. 1952. Fat combustion and metabolic rate of flying locusts (Schistocerca gregaria Forskål). Philos Trans R Soc Lond Biol 237:1–36. Weis-Fogh T. 1956. Biology and physics of locust flight: II. flight performance of the desert locust (Schistocerca gregaria). Philos Trans R Soc Lond Biol 239:459–510. Wilson DM. 1962. Bifunctional muscles in the thorax of grasshoppers. J Exp Biol 39:669–677. Wilson DM, Weis-Fogh T. 1962. Patterned activity of coordinated motor units, studied in flying locusts. J Exp Biol 39:643–667. Wolf H. 1990. On the function of a locust flight steering muscle and its inhibitory innervation. J Exp Biol 150:55–80. Wootten NW, Sawyer KF. 1954. The pick-up of spray droplets by flying locusts. Bull Entom Res 45:177–197. Wortmann M, Zarnack W. 1993. Wing movements and lift regulation in the flight of desert locusts. J Exp Biol 182:57–69. Zarnack W. 1983. Untersuchungen zum Flug von Wanderheu schrecken. Die Bewegungen, räumlichen Lagebeziehungen sowie Formen und Profile von Vorder- und Hinterflügeln. In: Nachtigall W, editor. Biona report, vol 1. Stuttgart: Fischer. p 79–102. Zarnack W. 1988. The effect of forewing depressor activity on wing movement during locust flight. Biol Cybern 59:55–70. Zarnack W, Möhl B. 1977. Activity of the direct downstroke flight muscles of Locusta migratoria (L.) during steering behaviour in flight: I. patterns of time shift. J Comp Physiol 118:215–233. Zarnack W, Send W. 1994. Application of unsteady aerodynamics to the locust flight. In: Elsner N, Breer H, editors. Sensory transduction. Proc. 22nd öttinger Neurobiology Conference, vol II. Stuttgart: Thieme. p 274. Zarnack W, Wortmann M. 1989. On the so-called constantlift reaction of migratory locusts. J Exp Biol 147:111–124.