The Experimental Approach to Rodent Population Dynamics

Abstract

Four paradigms can be followed to help us understand the dynamics of rodent populations. Natural observations are the traditional way to do ecology and it is important to have a good description of the population changes that are to be explained. Laboratory experiments were used in the 1940s and 1950s but provided little insight into events in the field. Field experiments began in the 1950s and have provided the greatest insights into the factors causing changes in numbers. Six factors have been manipulated in field experiments: food, predators, cover, density, physiological condition, and genetic composition. Field experiments can be used to test single-factor or multi-factor hypotheses of population regulation. They have the additional advantage of defining operationally verbal expressions like food-limitation that provoke fuzzy thinking. There is no need to separate natural observations from field experiments because they are the control for any manipulation done in the field. Mathematical modelling would seem to be the best of all possible worlds, but in rodent population dynamics it has been the worst because we do not know what the relevant variables to model are. Mathematical models are not a means of discovery and cannot be used to reject logically impossible ideas in the real world. Stenseth's attempt to reject the Chitty hypothesis as logically impossible is itself logically impossible. Whether the Chitty hypothesis is correct or not can be determined only by field experiments. Nor are mathematical models useful for defining concepts operationally in ecology, and I doubt if models ever produce new principles in ecology. I argue that the only useful goal for mathematical modellers in rodent ecology is to analyze how particular variables (like female territoriality) may affect population dynamics. This goal can be achieved only by a closer liaison between modellers and field ecologists at every step in the analysis.