Introduction

Previous chapters have detailed the responses of Peter and Tuesday Lakes to fish manipulations. Some of the changes were anticipated, while others were surprises. Our research was guided by models of the trophic cascade in lakes. To what extent did these models forecast the experimental results? The purpose of this chapter is to assess how our predictions fared, and how our view of the trophic cascade has been modified by the experimental outcome. First, we must explain why we developed simulation models of the trophic cascade, how the models were structured, and the predictions that derived from the models.

The trophic cascade is, in essence, a simple idea. In the complexity of real lakes, however, it involves the collective outcome of life history, predator–prey, and physical–chemical processes that cannot be adequately represented by simple verbal, graphical or mathematical models. Computer simulations are one way of integrating these complex interactions. They elaborate the conceptual framework and develop specific, testable predictions. We began simulation studies of the trophic cascade in 1981, three years before initiating the ecosystem experiments. Many of the predictions of those models can now be evaluated.

This chapter has four parts. First, we review three simulation models that produced the hypotheses we tested in the field. Second, we address predictions specific to the outcome of our ecosystem experiments. Third, we turn to more general expectations that should apply to trophic cascades in many lakes.