An experiment was conducted to determine whether eyewitnesses questioned under hypnosis are more likely to answer leading questions incorrectly than eyewitnesses questioned in a normal waking state. Ss viewed a videotape-recording depicting a car-bicycle accident and were questioned about the details of the accident. Half of Ss were questioned under hypnosis and half of Ss were questioned in a normal waking state and in each of these conditions, half of Ss were questioned 15 minutes after seeing the videotape-recording and half were questioned after a 24-hour delay. The results revealed that Ss in the hypnosis condition made significantly more errors on leading questions (questions that suggested an incorrect answer) than Ss in the normal waking condition. There were no significant differences between the groups of Ss on objective, non-leading questions. The results are discussed in terms of their implications for the use of hypnosis in police investigations.