The application proposes a continuation of statistical research into problems arising in laboratory experiments designed to assess the potential hazards of food additives, drugs, cosmetics, and other chemicals. The long-term objectives are to develop methodologies for designing and interpreting these experiments. The specific problems to be addressed in the current proposal include: (1) Biases and efficiencies of existing methods for assessing carcinogenicity; (2) Theoretical and empirical studies of cause-of-death data; (3) Use of pathology information in analysis of tumor onset; (4) Adaptive sacrifice methods; and (5) Methods for rodent teratology experiments. Each of these topics will involve theoretical considerations, as well as extensive exploratory analyses of the ED01 experiment and other data sets.