TWENTY-FOUR MALES AND 24 FEMALES FROM A UNIVERSITY POPULATION INSPECTED 2 FACES SEPARATELY FOR 20 SECONDS AND WERE THEN ASKED TO DESCRIBE THEM FROM MEMORY. HALF DESCRIBED FACE A IMMEDIATELY AFTER SEEING IT, AND THEN DESCRIBED FACE B AFTER ONE OF THREE DELAY PERIODS, AMOUNTING TO 1 HOUR, AND 1 DAY, OR 1 WEEK. THE OTHER HALF DESCRIBED FACE B IMMEDIATELY AND FACE A AFTER A DELAY. DESCRIPTIONS GIVEN FOR EACH FACE BY EVERY SUBJECT WERE USED IN A SORTING TEST IN ORDER TO SEE HOW WELL THEY COULD BE USED TO DIFFERENTIATE THE TWO FACES EMPLOYED. AN ANALYSIS OF VARIANCE (WITH SEX OF SUBJECT AND DELAY INTERVAL AS THE TWO BETWEEN FACTORS) REVEALED A SIGNIFICANT MAIN EFFECT AMONG THE THREE DELAY CONDITIONS. THERE WAS NO DIFFERENCE BETWEEN THE SEXES, HOWEVER, NOR WAS THERE AN INTERACTION BETWEEN DELAY CONDITION AND SEX. FURTHER ANALYSIS OF THE DELAY EFFECT BY THE NEWMAN-KEULS METHOD SHOWED THAT THE MEAN NUMBER OF DESCRIPTIONS GIVEN FOR ALL THREE DELAY CONDITIONS DIFFERED SIGNIFICANTLY FROM ONE ANOTHER. THE RESULTS OF THE SORTING TASK REVEALED THAT THE MEAN ERROR RATE FOR DESCRIPTIONS MADE 1 WEEK LATER WAS SIGNIFICANTLY GREATER THAN EITHER THE MEAN FOR 1 DAY OR THE MEAN FOR THE 1 HOUR DELAY. THE 1-DAY AND THE 1-HOUR DELAY CONDITIONS DID NOT PRODUCE SIGNIFICANTLY GREATER MEAN ERROR RATES. THERE WAS NO SIGNIFICANT MAIN EFFECT FOR SEX. IN CONCLUSION, THE EXPERIMENT HAS REVEALED THAT DESCRIPTIONS OF FACES ARE LESS DETAILED AND CONVEY LESS USEFUL INFORMATION AFTER AN INTERVAL OF MORE THAN 24 HOURS THAN WHEN MADE WITHIN A SHORT TIME AFTER SEEING A FACE. AT ALL INTERVALS, UPPER FACE FEATURES ARE MENTIONED MORE FREQUENTLY THAN LOWER FACE FEATURES, AND THERE IS NO SEX DIFFERENCE FOR DESCRIBERS REQUIRING THE ACCURACY OF FACE DESCRIPTIONS. FOOTNOTES AND TABLES ARE PROVIDED. (DEG)