Starting with Charcot, diagnostic criteria for multiple sclerosis (MS) have evolved to reflect advances in our understanding of the disease and the development of new diagnostic techniques, and from purely clinical considerations to increasing dependency upon imaging of the central nervous system. The MS diagnostic process was revolutionized by the 1981 introduction of magnetic resonance imaging (MRI), but the increasing reliance upon this technique has led to a surge in erroneous diagnoses, mostly because of the failure to distinguish between MS and disseminated encephalomyelitis (DEM), as well as mounting disregard for the data obtained from the traditional history and physical examination. The most recent scheme of McDonald et al. incorporated quantitative MRI criteria of dubious origin and reliability, but failed to provide qualitative, illustrative ones that would help differentiate between MS and DEM. The choice will have to be made by the neurological community between basing the diagnosis of MS on the MRI alone, or to use it as one aspect of a comprehensive clinical diagnostic algorithm. There will never be a substitute for the experienced and astute clinician's 'feel' for the patient.