Resource Depletion and Habitat Segregation by Competitors Under Predation Hazard

Individuals of different sizes in size-structured populations often differ greatly in the use of resources and/or space. Spectacular examples include shifts, within a lifetime, from carnivory to herbivory (e.g., some turtles; Clark and Gibbons 1969), or from herbivory to carnivory (e.g., some copepods; Neill and Peacock 1980), or from aquatic to terrestrial habitats (e.g., many amphibians). Many other species show large shifts in prey size or habitat use as they grow, but yet other species show little difference across sizes (Fraser 1976; Polis 1~84; Werner and Gilliam 1984; Persson this Vol.). Thus, patterns of resource and habitat use within a species vary from complete segregation between two given size classes, to partial overlap, to complete overlap. The presence or absence of such diet or habitat segregation between different size classes can greatly affect population structure and dynamics. For example, the degree of overlap between different size classes influences population stability, and the intensity of competition between classes influences, in interesting ways, whether increases in the density of a given class increases or decreases the density of other classes (Tschumy 1982; Mittelbach and Chesson 1987; Ebenman this Vol.). Some of the shifts in diet or habitat use might be rather inflexible in the on­ togenyof a given individual. However, a large literature demonstrates that many mobile animals show rapid behavioral changes in diet or use of space as resources change across time or space (Stephens and Krebs 1986). Further, several recent ex­ perimental studies show that alteration of predation hazard alters habitat selec­ tion or feeding behavior by mobile animals (e.g., reviewsby Helfman 1986; Milin­ ski 1986; Dill 1987; and Sih 1987 for aquatic systems). Flexible behavior by in­ dividuals greatly enriches the range of population structures and dynamics that we might expect to observe in nature, because behavioral shifts in habitat use can rapidly alter the degree of size class (or species) overlap, and may also suddenly alter size-specific mortality and growth rates. A hypothetical example is illustrated in Fig. 1. Mortality rate is usually taken to be an increasing function of predator density, but this need not be true if in­ dividuals can shift habitats in response to the hazard of predation (e.g., Iwasa 1982; Werner and Gilliam 1984). In Fig. 1, the mortality rate of a given size class is assumed to increase as a function of predator density over some interval, but