ABSTRACT
Smooth-tracking in the praying mantis was regarded by Rossel (1980, p. 325) as being driven by a velocity input signal in the same way that the optokinetic reaction is governed by the retinal image velocity of the surroundings. However, the following results indicate that the angular error between target position and midline of sight is also important in the control of smooth-tracking since there were increases in both saccadic and smooth torsion responses towards a stationary eccentric target in the open loop arrangement.
This paper is based on experiments with Sphodromantis lineola. Experiments on Mantis religiosa and Tenodera aridifolia sinensis gave comparable results. The torque was measured by means of strain gauges, one pair each for the torque of the head against the prothorax, of the prothorax against the abdomen and of the abdomen against the ground. The stiffness was chosen to allow the head to move by 16% of the expected amount. More rigid coupling suppressed the animal’s responsiveness to visual targets. Forces were transmitted through stiff wires flexibly inserted on both sides of the joint (Fig. 1). Visual stimuli were a black and white vertical grating of 0·05 cycles degree−1 spatial wavelength or a live cricket nymph against a homogeneous background. Both patterns were fixed to the inner wall of a cylinder which could be moved sinusoidally around the mantis in the frequency range of 0·08 to 1·9 cycles s−1. The movement covered an angle of 15°–30°; target movement was symmetrical about the midline of the mantis.
Mantis with three pairs of strain gauges I, II, III in the arrangement for open loop experiments. P, plastic disc. Capture legs omitted.
When the prey target is moved sinusoidally and then stopped close to a turning point, torsion is not only sustained at the level reached at this time but may sometimes increase still further in the direction of the target. This increase usually starts after a delay. One or more torsion saccades are followed by a sustained smooth torsion level which may lie below the over-shooting saccadic torsion as in Fig. 2 or may gradually increase as in Fig. 3. Head, prothorax and legs produce a coordinated response with synchronous changes of torsion (Fig. 3). This smoothly continued or increasing torsion must be driven by the target’s eccentric position and not by its motion. In contrast, the mantid’s response to a grating which stops moving is quite different. In this case the torsion slowly declines and never increases.
Torsion course of head (H), prothorax (P) and legs (L) before and after stop of sinusoidal target movement (T). S, two torsion saccades towards the target followed by slowly relaxing increased smooth torsion. Leg torsion is recorded inversely to the target movement direction. Torsion scales given in Joules.
Torsion course of head (H), prothorax (P) and legs (L) before and after stop of sinusoidal target movement (T). S, two torsion saccades towards the target followed by slowly relaxing increased smooth torsion. Leg torsion is recorded inversely to the target movement direction. Torsion scales given in Joules.
ACKNOWLEDGEMENTS
Lässig & Kirmse (1972, p. 243), using the closed loop condition, provided direct evidence for position-dependent action in smooth-tracking: after an overshooting correction saccade the line of sight is returned to the target by smooth-tracking, which in this case is opposite to the target motion.
