In recent controlled trials, conducted by the authors (1956), we were interested to note the proportion of patients improving with the administration of placebos. A placebo is defined in the Shorter Oxford Dictionary (1944) as “A medicine given more to please than to benefit the patient”. It is presumed that, implicit in this definition, is the qualification that the doctor is aware of the relatively inert qualities of the substance, whilst the patient is not. There is, of course, a type of patient who derives considerable benefit mainly by virtue of the fact that he has been pleased. In spite of the discovery of increasing numbers of drugs with specific properties, the prescription of placebos is still common practice. Dunlop et al. (1952) analysed over 17,000 prescriptions of which about one-third were tonics, stomachics or sedatives; although barbiturates cannot strictly be called placebos, the spirit in which they are prescribed frequently puts them near to this category.