THE EXPERIMENTS WERE UNDERTAKEN TO EXPLAIN THE PHENOMENON OF MISTAKEN IDENTIFICATION AND TO PERFECT TECHNIQUES FOR THE EFFECTIVE INTERROGATION OF WITNESSES. IN ONE STUDY, 100 BLACK AND WHITE FULL-FACE PHOTOGRAPHS OF MEN AGED 17-60 YEARS WERE COLLECTED, AND SAMPLES WERE THEN SHOWN TO 40 STUDENTS WHO WROTE DESCRIPTIONS OF THE FACES. A MAJOR RESULT OF THE EXPERIMENT WAS THE FINDING THAT SOME FEATURES OF THE FACE ARE MENTIONED MANY MORE TIMES THAN OTHERS. HAIR WAS THE FEATURE MOST FREQUENTLY EMPLOYED (IN PARTICULAR COLOR, LENGTH, TEXTURE AND PARTING) FOLLOWED BY EYES AND EYEBROWS, NOSE AND GENERAL FACIAL STRUCTURE. BY CONTRAST, THE CHEEKS, MOUTH, AND FOREHEAD WERE RARELY USED AS A MEANS OF DISCRIMINATING ONE FACE FROM ANOTHER. RESULTS INDICATE THAT, IN GENERAL, UPPER FEATURES OF THE FACE WERE MORE IMPORTANT IN DESCRIBING FACES THAN LOWER FEATURES. AN INCIDENTAL FINDING SUGGESTED THAT THE BEST DESCRIPTIONS WERE THOSE PROVIDED AFTER 1 HOUR, IMPLYING THAT THE BEST TIME TO INTERROGATE A WITNESS MAY NOT BE IMMEDIATELY AFTER THE CRIME. ANOTHER EXPERIMENT INVOLVED SORTING FACES BY SIMILARITY USING A MULTIDIMENSIONAL SCALING TECHNIQUE. THREE FACTORS WERE FOUND TO BE IMPORTANT IN THE SORTING PROCESS: AGE, FACE SHAPE, AND HAIR LENGTH. THE FINAL EXPERIMENT USED A HIERARCHICAL CLUSTERING ANALYSIS AND PROVED THAT FACES WHICH WERE SIMILAR IN TERMS OF SHAPE, HAIR LENGTH, AND AGE, BUT DIFFERENT IN MANY OTHER RESPECTS, WERE READILY CONFUSED BY SUBJECTS. THE WORK THUS DEMONSTRATED THAT ABILITY TO REMEMBER FACES DEPENDS ON, AND CAN BE CONFUSED BY, A MUCH SMALLER RANGE OF CUES AND THAT IN TAKING AND STORING INFORMATION ABOUT FACES, THE UPPER FEATURES AND THE SHAPE OF THE FACE APPEAR TO BE GIVEN PRIORITY. PHOTOGRAPHS, TABLES, AND FOOTNOTES ARE PROVIDED. (MJW)