We examined mortality and feeding inhibition response of Lymantria dispar L. (Lepidoptera: Lymantriidae) larvae to ingested doses of Bacillus thuringiensis subsp. kurstaki as a function of dose, instar and temperature. We developed generalized (logistic) linear mixed models and a mixture survival model, commonly used in medical statistics, to analyze the complex data set. We conducted bioassays of Foray 48B with larvae from the NJSS laboratory stock, using droplet imbibing or force-feeding to ensure dose ingestion. The dose causing mortality in 50% of the test population (LD(50)) under standard test conditions (22 degrees C) ranged from 0.019 International Units (IU)/larva for first instar larvae (L(1)) to 1.6 IU/larva for L(4). Temperature affected larval mortality in two ways. Mortality occurred sooner and progressed more rapidly with increasing temperature (13-25 degrees C) at each dose level and instar, while the maximum level of mortality attained by each instar decreased with increasing rearing temperature. The mechanisms underlying this effect are being investigated. Larvae that survived exposure to B. thuringiensis resumed feeding after a period that was dependent on instar, dose, and temperature. The equations describing observed mortality and feeding recovery responses were used to construct a simulation model, which was able to predict both processes, and which forms the basis for a process-oriented model that can be used as a decision support tool in aerial sprays.