The use of optokinetic nystagmus as a criterion of vivid, hypnotic imagery was re-examined. A slowly moving stripe was used to induce nystagmus. Eye movements were recorded whileSobserved the stimulus and then imagined it, in the waking state and as an hypnotic hallucination. A nystagmus-like pattern was observed in all three cases, but eye movements during imagery contained significantly more saccades than eye movements during actual observation. There was no significant difference between waking imagery and hypnotic hallucinations.