1. The methodology of the initial patient screening of psychopharmacological agents is a relatively neglected but important area for discussion and improvement.2. The main functions of this activity are to uncover safely and efficiently a wide variety of leads about desirable and undesirable pharmacological and psychiatric actions. This requires balance and compromise between the practice of the clinical art of pharmacotherapy and the use of procedures of good scientific control.3. There are advantages in conducting such investigations on a research and teaching-oriented, active and eclectic treatment, inpatient service, using small numbers of acutely ill psychotic patients under intensive observation for periods up to 2 to 3 months. Subtle drug effects can be seen, and such a milieu calls forth a wide range of adjustmental behavior for study.4. Multiple observers of different professional backgrounds and training, using a variety of rating scales and checklists to evaluate independently the same phenomena, enhance objectivity, diminish bias and help to circumvent individual blind spots.5. There are special problems of selection of subjects since one is interested in observing the effects of drugs upon a large spectrum of psychopathological factors while insuring safety.6. In the experimental design the testing of a new drug against a carefully selected active referent agent and inert placebo, and the use of intrapatient and interpatient comparisons are feasible and to be recommended. Although starting with flexible adjustment of dose, a rigid dosage schedule and true crossover design can be employed to advantage safely in the later stages of many studies.7. Along with the collection of data by open experiment, provision can be made for employing double-blind technique.8. There is a special need in preliminary drug testing for systematic handling of individual case data during each investigation in order to uncover and exploit leads, insure safety and to allow for the report of certain kinds of summary data to drug firms and to other investigators.9. The detailed post-experiment report of individual case data allows others to seek significant correlations other than those of interest to the investigator himself.10. This particular kind of research can be effectively incorporated into a larger research program.