Abstract
Self-organizing motion is an important yet inadequately understood phenomena in the field of collective behavior. For birds flocks, insect swarms, and fish schools, group behavior can provide a mechanism for defense against predators, better foraging and mating capabilities and increased hydro/aerodynamic efficiency in long-distance migration events. Although collective motion has received much scientific attention, more work is required to model and understand the mechanisms responsible for school initiation and formation, and information transfer within these groups. Here we investigate schooling of black tetra (Gymnocorymbus ternetzi) fish triggered by startle stimuli in the form of approaching objects. High-speed video and tagging techniques were used to track the school and individual members. We then measured several variables including reaction times, group formation shapes, fish velocity, group density, and leadership within the group. These data reveal three things: (1) information propagates through the group as a wave, indicating that each fish is not reacting individually to the stimulus, (2) the time taken for information to transfer across the group is independent of group density, and (3) information propagates across large groups faster than would be expected if the fish were simply responding to the motion of their nearest neighbor. A model was then built wherein simulated fish have a simple 'stimuli/escape' vector based on a hypothetical field of vision. The model was used to simulate a group of individual fish with initial conditions, size, and stimuli similar to the biological experiments. The model revealed similar behavior to the biological experiments and provide insights into the observed patterns, response times, and wave speeds.