Most medical communications are difficult to read. To determine why, contributions to three issues of the New England Journal of Medicine were studied, and the prose analyzed.Materials and MethodsArticles were taken from the Journal issues dated April 4, 1974; February 6, 1975; and October 16, 1975. These issues lay buried under a pile of papers on my desk. Articles were read at random.ResultsI identified 10 recurring faults in the Journal articles I read.1. Poor Flow of IdeasThis is the most common and subtle fault plaguing medical communication. Within a single sentence, or from one . . .