In the psychological literature on eyewitness testimory, the classic Allport and Postman (1945, 1947) study of rumor has consistently been described inaccurately. In the inaccurate accounts both the procedures and implications of the study have been substantially distorted. The erroneous descriptions have found their way into the legal literature, apparently as the results of legal scholars' reliance on the inaccurate secondary reports in the psychological literature. Furthermore, psychologists testifying as experts in court have offered inaccurate accounts of the study in support of contentions about effects of prejudice on eyewitness perception. This note contrasts the actual methods and results of the Allport and Postman study with the descriptions in the eyewitness testimony literature, with the aims of averting future errors concerning the study, and of emphasizing the importance of consulting original sources and reading them carefully.