Abstract
The nonlinear dynamics view of population interactions emphasizes three critical points: 1) instability does not imply extinction, 2) very complicated behavior is possible from very simple systems, and 3) if density-dependence occurs as lagging nonlinear feedback, then it is the primary cause of instability and chaos and does not stabilize populations in contrast to the traditional view. Discrete time ("Nicholsonian") host-parasitoid models are used to illustrate that a patch-efficient parasitoid is destabilizing when searching a patchy host distribution. When some very crude genetics are added in terms of parameter phenotypes (patchy and random hosts and patchy and random parasitoid search strategies), then there is a much wider variety of dynamic behavior. In general, selection does not seem to work for or against chaotic parameter phenotypes. Instead, it appears to increase the likelihood of chaos (i.e., at lower parameter values) and decrease the likelihood of extinction of chaotic systems if they occur.