A brain homogenate prepared from a terminally ill hamster infected with scrapie strain 263K was serially diluted and administered orally to groups of hamsters. The undiluted brain homogenate led to clinical scrapie in all animals inoculated. The attack rate decreased with dilutions of the homogenate, and subclinical infections were identified among the healthy survivors at 520 days post-infection by Western blotting. The number of animals succumbing to disease and the combined number of Western blot-positive survivors plus diseased hamsters were used to calculate the LD(50) and ID(50) of the inoculum. The model system represents an approximation to the transmission of TSEs such as new variant Creutzfeldt-Jakob disease (vCJD) via dietary exposure to the infectious agent and suggests that, due to the rather small difference between the calculated LD(50) and ID(50), the number of clinical cases will not be vastly exceeded by the number of subclinical carriers of the disease.