How is eyewitness testimony affected by the presence of an audience (e.g., in a court of law)? Relevant social psychological research (Zajonc, 1965) leads to testing the hypothesis that people will report more central and less peripheral details in the presence of an audience, compared to a no-audience condition. Students watched four short video sequences and freely recalled them one week later, either in an individual questioning condition or in a courtroom-like situation with one questioning and four observing persons. Central (peripheral) details were operationalized as details with a high (low) overall frequency of reporting. A content analysis of the recall protocols supported the hypothesized differences between the questioning conditions. This also held true if only correct details were considered. As a speculative explanation, I suggest that retrieval processes are affected as a consequence of being distracted by the audience. Possible implications for the utility of eyewitness testimony in court, its perceived credibility, and for eyewitness interrogation are discussed.