We evaluated the reliability of 20 cognitive tasks as part of a clinical mental status examination. Twenty-five adult psychiatric inpatients were selected at random and interviewed before three board-certified psychiatrists who, without any knowledge of each other's ratings, completed a form determining the presence or absence of abnormal responses. We found 19 of the 20 items to yield R values of greater than .50, with 14 of these having R values of .80 or better. All correlations were significant at the <.01 level. We suggest that these cognitive tasks with demonstrated reliability be included in the standard mental status examination. As a group, the tasks are easily and rapidly administered and should provide more accurate clinical screening of patients suspected of cortical dysfunction. These tasks will enable clinicians to make more precise and cost-effective referrals for elaborate, timeconsuming, and expensive neuropsychological testing.