The use of an alternative methodology to study transfer of learning in clinical instruction during medical school was investigated. The environment in which clinical instruction takes place was examined, after which hypotheses were proposed and tested in a quasi-experimental design. The first phase of the study, an ethnographic analysis of the teaching and learning environment, revealed that different disciplines (medicine, pediatrics, and surgery) provided quite different environments for the students. Different types of learning (factual knowledge, problem-solving, technical skill, interpersonal skill, and attitude toward health care) were emphasized in each discipline; students were involved in varying activities for varying amounts of time; and the roles of the individuals involved in the teaching process (residents, interns, attending staff) differed among the rotations. Both of the following hypotheses were partially confirmed: (1) the different disciplines in the clerkship program would facilitate different types of learning, and (2) there would be no significant transfer of learning among the disciplines. Factual knowledge and problem-solving were facilitated by different disciplines. These two types of learning were also the ones that appeared to transfer from one discipline to the other. The remaining types (interpersonal skill, technical skill, and attitude toward health care) showed no evidence of being transferred. These results indicate that the clinical teaching and learning process is complex, and it is suggested that educators consider the variety of tasks that students undertake, the situations encountered, and the types of abilities required. A bibliography is appended. (SW)