In two studies, the role of biomedical knowledge in the diagnosis of clinical cases by physicians and medical students was explored. Experiment 1 demonstrated a decrease in the use of biomedical knowledge with increasing expertise. This result appeared to be at variance with some findings reported in some but not all of the literature. In Experiment 2, three possible explanations for this phenomenon were investigated: rudimentation of biomedical knowledge, inertia, or compilation. Using a combined think-aloud and post-hoc-explanation methodology, it was shown that experts have more in-depth biomedical knowledge than novices and subjects of intermediate levels of expertise, but use this knowledge in a tacit way, the result of a process of knowledge compilation. The findings generally support a three-stage model of expertise development in medicine: (1) during the preclinical stage biomedical and clinical knowledge develop separately with an emphasis on biomedical knowledge; (2) during the clerkship stage, emphasis is shifted toward the acquisition of clinical knowledge; and (3) after both knowledge bases have matured to a sufficient extent, biomedical knowledge is compiled and integrated into clinical knowledge. Contains 30 references. (DB)