The effects of inducible defenses and constitutive defenses on population dynamics were investigated in a freshwater plankton system with rotifers as predators and different algal strains as prey. We made predictions for these systems using a chemostat predator–prey model and focused on population stability and predator persistence as a function of flow-through rate. The model exhibits three major types of behavior at a high nutrient concentration: (1) at high dilution rates, only algae exist; (2) at intermediate dilution rates, algae and rotifers show stable coexistence; (3) at low dilution rates, large population fluctuations occur, with low minimum densities entailing a risk of stochastic rotifer extinctions. The size and location of the corresponding areas in parameter space critically depend on the type of algal defense strategy. In an 83-day high-nutrient chemostat experiment we changed the dilution rate every 3 weeks, from 0.7 to 0.5 to 0.3 to 0.1 per day. Within this range of dilution rates, rotifers and algae coexisted, and population fluctuations of algae clearly increased as dilution rates decreased. The CV of herbivore densities was highest at the end of the experiment, when the dilution rate was low. On day 80, herbivorous rotifers had become undetectable in all three chemostats with permanently defended algae (where rotifer densities had already been low) and in two out of three chemostats where rotifers had been feeding on algae with inducible defenses (that represented more edible food). We interpret our results in relation to the paradox of enrichment.