Some of the difficulties of trying to establish an animal model of schizophrenia are first considered. Then, after a review of the evidence on the experimental psychopathology of schizophrenia, particularly that concerned with attention and arousal, it is concluded that the core feature which needs to be modeled in animals is some aspect of "input dysfunction." It is argued that, of the pharmacological strategies, LSD-25 comes nearest to meeting that requirement, for two reasons. First, the phenomenology of an LSD "model psychosis" closely parallels that of the natural disease. Secondly, the experimental effects of the drug, both in animals and man, are very similar to or can be closely aligned theoretically with those of schizophrenia. An example is quoted from work in the author's laboratory where LSD was found to produce psychophysiological effects virtually identical to those observed occurring naturally in acute psychotic patients and in normal subjects high in "psychotic" personality traits. It is suggested that the rejection of LSD as a drug model was premature, especially as the currently popular preference for amphetamine has not been vindicated, either by the latter's ability to mimic an important central feature of the psychotic state or by work on dopamine as a specific common mediator of amphetamine psychosis and of schizophrenia.