1. Based on the slow‐growth high‐mortality (SGHM) hypothesis, which predicts that prolonged larval development increases mortality from their natural enemies, studies have often assumed that low quality of plants that slows larval development would function as a defence against insect herbivores. However, empirical support for theSGHMhypothesis has been limited, especially in natural and ecologically relevant contexts.2. In a leafminerAmauromyza flavifronsMeigen (Agromyzidae,Diptera), theSGHMhypothesis was tested along with four other hypotheses (e.g. prey size, mine appearance, density‐dependent parasitism, and plant quality hypotheses) to control for spurious associations between development time and parasitism that are primarily driven by other larval traits. Two host plant species,Saponaria officinalisandSilene latifolia, were grown under varying nitrogen levels, and leafminers developing on these plants were exposed to, or protected from, a natural assembly of parasitoids across the entire course of larval development.3. On both host plant species, leafminers that survived to an adult stage in the presence of parasitoids had a shorter development time than those in the absence of parasitoids, indicating that parasitoids disproportionately kill leafminers with longer larval development. The results provided concrete evidence for theSGHMhypothesis within the natural ecological context for these interacting species. Moreover, reduced plant quality was associated with higher larval mortality onSa. officinalisonly in the presence of parasitoids, suggesting that low quality could function as indirect plant resistance viaSGHMunder some tri‐trophic inter