Foveal fixation and tracking in the praying mantis

Summary

• 1. Visually guided head and body movements of restrained and freely moving mantids (Tenodera australasiae) have been studied by means of closed circuit television. Interest was concentrated on the different roles of the fovea and the periphery of the eye in controlling visuo-motor behaviour.

• 2. The peripheral eye is mainly responsible for the detection of novel objects (preferably potential prey) and the generation of ballistic (open-loop) saccadic head movements (Fig. 2) which bring the target image to the fovea (Figs. 3, 4, 5, 6).

• 3. Measurements on monocular animals show that the fovea ofeach eye is encircled by a saccade sensitive periphery (Figs. 5, 6). In other words each eye is capable of measuring any retinal position of the target image in a coordinate system whose origin is at the fovea. Based on this finding, a hypothesis, which is outlined in Sect. IV.7, suggests that the binocular coordination during fixation, tracking and distance estimation is based on the comparison of angular coordinates extracted by each eye from the position vector of the target.

• 4. Moving targets which have been fixated are held in the fovea either by smooth or saccadic tracking eye movements. The degree to which either tracking strategy is employed depends mainly on the features of the background, but to some extent also on the velocity of the target.

• 5. Targets which move against a homogeneous background are tracked by smooth eye movements (Fig. 7). Low target angular velocities are closely matched by the eye velocity. At high target speeds the head lags increasingly behind the target and saccades are periodically required to reduce the position error relative to the fovea.

• 6. Smooth pursuit eye movements, evoked either by a single target (Fig. 7) or a disrupted background (Fig. 8), are affected primarily by the velocity of the retinal image. While the effects of target and background are similar in this respect, they differ in others. Small objects in the foreground, subtending an angle of only a few degrees on the retina, evoke strong pursuit responses only when they resemble typical prey and project onto the fovea (Fig. 9). On the other hand, the image stabilisation of the background is a stereotyped response that can be evoked whenever a large part of the background moves across the visual field (Fig. 10a). Moreover, responses caused by a moving target in the fovea, and movements of the background in the periphery, are not combined additively (Fig. 10b). The foveal tracking response is weighted more strongly, but because the target is usually small, compared with the background, competing background motion can suppress smooth foveal tracking almost completely.

• 7. This limitation imposed upon the smooth pursuit system by the presence of a disrupted background (either a stripe pattern in the experimental set-up or grass and other plants in a natural setting) is avoided by the adoption of a strategy of saccadic tracking (Figs. 11, 12, 15, 16). This also applies for the tracking which immediately precedes the catching of prey (Fig. 15). Therefore, the stabilisation of the target image in place on the fovea is not a prerequisite for a successful strike.

• 8. Up to target angular velocities of about 100 °/s, saccadic tracking is predictive, i.e. the saccades have adequate amplitudes to bring the fovea right on target at the instant the saccade is completed (Fig. 13). This implies that the saccadic system processes not only position information of the target but velocity information as well. It is suggested that this velocity information is provided by the smooth pursuit system. Saccadic tracking would then reflect interactions of two circuits, the velocity coding circuit which in the presence of a homogeneous background also generates smooth pursuit head movements, and the position coding circuit which in the absence of target movement is able to generate saccades on its own.