Spontaneous vs. Reflexive Turns

Following are bottom and side views of a tilapia (22 cm long) spontaneously tuning left by 45°.

Fish employ different swimming styles in different situations: constant-velocity cruising, for example, or sudden starts for prey capture and escape. The kinematics and hydrodynamics of steady swimming have been studied extensively. However, so-called unsteady movements, during which fish change their swimming speed and direction, are more difficult to measure and analyze. Unsteady maneuvers include escape responses (which can be provoked in the lab by a visual or auditory stimulus) and voluntary turns. The aim of our study is to explore the mechanisms that fish use to turn. We expect that turning performance depends on body shape, turning mechanism, and motivation (spontaneous versus reflexive).

We have selected a cooperative individual, and trained him to swim through hoops by positive conditioning. By arranging the hoops along various points of an arc, we can encouraged the fish to turn through predetermined angles and radii. However, after filming it has become clear that useful data can also be obtained as the fish spontaneously turns through the hoops and around them (the trained fish has a tendency to swim in the area where the hoops are positioned). The fish now routinely performs voluntary turns that have been filmed from two perspectives with high-speed video. Of particular interest are small-radius turns that superficially resemble the reflexive escape response, which we have also filmed. It is hoped that quantitative comparison of these recordings will reveal fundamental differences in the turning mechanisms.

The kinematics of its center of mass and midline curvature are are tracked by computer:

Finally, the time-course of the midline curvature can be analyzed. In the plots below, the curvature is plotted as a function of time upward (0-400 ms). Each horizontal slice represents the distribution of midline curvatures at the corresponding frame of the video. Red and blue indicate opposite extremes of local curvature.

In the spontaneous turn (left), a well-defined wave with 2-3 nodes moves down the tail. This pattern is similar to steady swimming in a straight line. In the reflexive turn (right), the rear 70% of the fish curves strongly nd simultaneously after the stimulus (at ~100 ms). At 150 ms, the fish has assumed the characteristic “C” shape of panic start.

Reflexive turns are marked by a conspicuously large peak in center-of-mass acceleration. Because this is brief, and occurs normal to the ultimate displacement, it has little effect on the position or speed of the fish. The fish merely bends about its (relatively stationary) center of mass. This “preparatory” motion is not followed by further escape behavior in the present example (where the threat was transient).

Larval fish escape responses

Schematic diagram of the experimental setup:  the fish larva experiences a sudden onset of flow, induced by the piston to simulate the suction field of a predator.

Photograph of the laboratory setup, on loan from the McHendry lab.

Example data:  a single excape trajectory (left) and a compilation of multiple runs (right).  The distribution of escape directions is clearly not isotropic.