Present statistical analysis of survivorship data obtained from bioassays using nonpersistent toxicants does not explicitly take into account that exposure decreases with time due to degradation of the toxicant. Such a situation typically occurs in static aquatic toxicity tests and in soil tests. We develop a model for the analysis of survival data obtained in such experiments and show how estimates for the initial LC50, the degradation rate constant, and the elimination rate constant may be obtained from observations on time‐dependent survival rates. The model assumes linear one‐compartment toxicokinetics with exponentially decreasing input. The probability of dying is assumed to be directly proportional to the internal concentration. An explicit expression for time‐ and concentration‐dependent survival is obtained. The model predicts that, with increasing exposure time, survival will approach a nonzero baseline value for certain initial concentrations. Application of the model is illustrated by curve‐fitting to experimental observations, made over 6 weeks, for the effect of diazinon on the terrestrial isopod Porcellio scaber. The fit to data is fairly good, but the parameter estimates tend to have a high coefficient of variation. A considerable increase of precision may be obtained if the degradation rate constant is given a fixed value (for example, following from chemical residue analysis of the medium). The approach is applicable to all situations where, due to loss of the toxicant during the test, mortality shows no further increase after a certain exposure period.