Some investigators have used evidence that a drug may induce conditioned taste aversion to dismiss other behavioral effects of the drug as due to general toxicity rather than specific pharmacological action. To test the validity of this position, behaviorally relevant doses of alcohol, chlordiazepoxide, and morphine were studied in an aversive taste conditioning paradigm. Drug injections were paired with exposure to saccharin solution on repeated trials. Each drug produced saccharin aversion at one or more doses. Since comparable doses are known to facilitate rather than impair behavior in some situations, the taste aversion test is insufficient to establish the general behavioral toxicity of particular drug doses.