Little is known about how information diffuses to clinicians and influences their purchase and use of new technology. This is especially true about the role of the scientific literature. As a case study, we examined the literature for magnetic resonance imaging (MRI) during the years preceding and the first five years following its clinical introduction. Using a computerized retrieval system, we identified approximately 1,700 citations in which MRI was the major topic. The clinical literature on MRI was heavily concentrated in radiology journals. Less than 28% of articles compared MRI with alternative diagnostic technologies. During the first five years of clinical availability, the diffusion patterns of scientific articles and operational units mirrored the example set by computerized tomography (CT), in that a substantial number of units were purchased in both research and nonresearch settings before studies were available comparing them to alternative diagnostic technologies. These patterns of diffusion, combined with other studies of the MRI literature's content and methodology, suggest that less comprehensive and objective sources of information were important in early purchasing decisions. This study also suggests that the present readership and publication patterns of professional journals may not facilitate effective, rapid information dissemination about innovations to a broad spectrum of clinicians.