The pharmacokinetic theory of intra-arterial drug administration has been clearly articulated. It provides a useful guide to the development and interpretation of preclinical studies and clinical trials. Despite the clarity and usefulness of the theory, there are misunderstandings of its implications and technical problems associated with its implementation. Few studies have been properly designed to validate the theory experimentally. Independence of steady-state pharmacokinetic advantage on intraregion blood flow differences predicted for a nonextracted drug is counterintuitive. Extrapolation from laboratory animals to humans raises some important questions of allometry, particularly for the brain, which does not follow the scaling rules applicable to other organs. Finally, drug streaming from the site of infusion has been observed in vitro and in vivo. The extent of the resulting clinical problem has not been adequately characterized, but it may be severe under some circumstances.