Questions: How do defences expressed only in the presence of predators (inducible defences) affect the relationship between the number of prey eaten and prey density (the predator’s functional response)? What is the relationship between prey defence level and vulnerability? Do inducible defences show features that are likely to stabilize predator–prey dynamics on theoretical grounds? Organisms: We conducted experiments in a laboratory system. The prey was the protist Euplotes octocarinatus, exposing different levels of a morphological inducible defence. The predator was the turbellarian Stenostomum virginianum. Methods: We analysed the data using non-linear mixed effects models that combine non-linear curve fitting with random effects. Results: This predator’s functional response was a sigmoid Holling-Type III. The induced defence lowered the maximum number of prey that predators ate, suggesting that the defence increased the predator’s handling time. The level of defence expressed by Euplotes depended on the level of predator cue. There was a negative exponential decline in the number of prey eaten with increasing level of defence. Low levels of defence were thus effective and further increases in defence improved prey survival only slightly. The possibility for prey to become nearly invulnerable, the effectiveness of low levels of defence, and variation in vulnerability all suggest that this inducible defence can stabilize community dynamics.