Recent developments in the study of nonlinear systems have shown that simple theoretical models are capable of extremely complicated dynamics that cannot be predicted over the long term but that are nonetheless deterministic. The dynamics of such systems are "chaotic," rather than random. The emergence of the behavior of a social unit from the actions of the individuals that comprise it is one of the central problems in the study of behavior. Dr. Cole is studying how the patterns of movement activity of ant colonies are derived from the actions of individuals. Individual worker ants appear to have chaotic activity patterns, spontaneously becoming active and reverting to inactivity. Colonies go through periodic episodes of activity about every 30 minutes. The transition from the chaotic dynamics of individual workers to the periodic dynamics of colonies provides an opportunity to study the mechanisms by which a social phenotype is produced. Furthermore, the fact that the activity of an individual is chaotic, and therefore deterministic, suggests that short-term predictions can be made about the activity of workers. Comparing the actual activity of ants and computer models of chaotic systems, Dr. Cole will study the transition from chaos to periodicity in two situations: first, in artificial aggregates of workers constructed from mature colonies, and second, during the pattern of normal colony growth as the colony passes from a single individual (the queen) to a mature colony with many workers.
