The ethical norms established in various codes and regulations are inadequate to resolve some of the ethical problems presented by clinical trials. They are stated too vaguely to provide unequivocal answers to many specific questions. In order to remedy this situation, many commentators have proposed the development of more specific and complex regulations. We propose that a more fruitful approach would be to examine the ethical principles underlying the norms and to apply these principles to the specific problems. We apply this approach to two questions: (1) Is it ethical to select subjects for a randomized clinical trial (RCT) exclusively from Veterans Administration (VA) hospitals? (2) In the conduct of a RCT is it necessary to disclose the fact that therapy will be determined by chance? We conclude that problems of justice arise not only because of the vulnerability of patients in VA hospitals but also because of the loss of the physician-patient relationship in an RCT. However, the use of patients in a VA hospital is not always unjust; in most cases such use can be made more just through various modifications in design. We also conclude that the fact of randomization should be disclosed in any situation in which it might materially affect the prospective subject's decision, and that the values and preferences of the subjects should be taken into account in determining what information might be material. This work is only a preliminary step toward analyzing ethical issues in clinical trials. While some would challenge our conclusions, we hope that our methods will facilitate clarity about the locus of disagreement in current controversies and about the value questions that must be answered in order to set an ethical context for the conduct of clinical trials.