The initiation of flight in unrestrained locusts,Schistocerca gregaria

Abstract

1. A technique is described for taking multiple exposure photographs of the first 100 msec of the jump of unrestrained and unstimulated locusts,Schistocerca gregaria (phasis gregaria) (Fig. 1). 2. Photographs were taken of locusts jumping under twelve different experimental conditions (Table 1). Eight features of the jump were measured (Table 2). 3. The direction of take-off is variable under all conditions. In intact locusts the mean take-off angle is approximately 45 ° but the mean increases significantly in all the groups of operated insects (Table 3). 4. The body angle at the moment of take-off is always variable but neither the mean nor the standard deviation of the performance changed significantly in different experimental groups. There is no correlation between the take-off angle and the body angle at the instant of take-off. 5. 70 msec after take-off a correlation develops between the body angle and the direction of motion in all the experimental groups except that in which the central nervous system is severed between the brain and the thoracic ganglia (Table 4). The correlation coefficient is always higher in the experimental groups which were exposed to a headwind. 6. Experiments using localised jets of air show that the increased correlation between the body angle and the direction of motion is not due to the passive aerodynamic effect of the air flowing over the body. The evidence suggests that the insect obtains information about the direction of the headwind and decisions about the form of the later stages of the jump and the timing of the onset of wing movements are made before the jump begins. 7. The rotation of the body in the plane of pitch during the first 100 msec of the jump is caused mainly by movements of the legs which alter the balance of the body. 8. The delay between the metathoracic tarsi leaving the ground and the start of visible wing movements is never constant and may be negative. The mean delay in certain experimental groups was found to be significantly different from the controls. Cutting the nerve cord anterior to the thoracic ganglia and destruction of the cephalic wind sensitive organs shortens the delay, but damage to the nervous system posterior to the thoracic ganglia has no effect (Table 5). 9. In both the intact and injured groups, the stimulus of an airstream caused a significant decrease in the mean latency of the flight response. There was no correlation between the degree of alignment of the body to its direction of motion and the latency of the flight response. 10. Cutting the nerve cord anterior to the thoracic ganglia and destruction of the cephalic wind sensitive organs also caused a significant increase in the period of the first wingbeat cycle but none of the other experimental conditions altered it (Table 6). 11. The timing of the beginning of visible wing movements did not correlate with any feature of the jump which was measured. 12. It is concluded that variation in the form of the jump is an inherent property of the system and may be due to mechanical instability as well as to variability in the neural program. The brain is involved in adapting the form of the jump so that less energy is wasted as drag before wing movements take over as the method of propulsion. It is suggested that the neural centres which control jumping and flight are mutually inhibitory and that the strength of the inhibition depends upon input from the rest of the nervous system.