STUDIES HAVE SHOWN THAT IT IS POSSIBLE TO ESTABLISH DISCRIMINATORY RESPONSES TO PAIRS OF STIMULI WITHOUT ERROR RESPONSES TO THE UNREINFORCED (S-) STIMULUS. THE PURPOSE OF THIS STUDY WAS TO DETERMINE THE EFFECTS OF 2 METHODS OF INTRODUCING S- (FADING AND CONSTANT) AND 2 RESPONSE CONTINGENCIES (DELAY AND NO DELAY) ON THE OCCURRENCE OF S- RESPONSES. THE SUBJECTS WERE 27 KINDERGARTEN CHILDREN WHO WERE DIVIDED INTO 3 GROUPS WHICH RECEIVED 1 OF 3 TRAINING PROCEDURES--(1) FADING-NO DELAY, (2) CONSTANT-NO DELAY, AND (3) CONSTANT-DELAY. THE STIMULI WERE A HORIZONTAL AND A VERTICAL LINE PROJECTED ONTO A PLEXIGLASS SCREEN. THE RESPONSE CONSISTED OF PUSHING ON THE SCREEN, AND THE S+ STIMULUS WAS REINFORCED WITH CANDY. IN THE INITIAL SESSION THE S+ STIMULUS WAS PROJECTED 5 TIMES TO ESTABLISH THE RESPONSE, AND THEN THE S- STIMULUS WAS INTRODUCED IN A RANDOM FASHION. IN THE FADING PROCEDURE, ITS INTENSITY AND DURATION WERE GRADUALLY INCREASED UNTIL IT MATCHED THAT OF THE S+ STIMULUS. IN THE CONSTANT PROCEDURE IT WAS KEPT AT THE SAME INTENSITY AS S+. IF THERE WAS NO RESPONSE TO S-, IT TERMINATED IN 5 SECONDS. IN THE DELAY CONTINGENCY, 5 SECONDS HAD TO ELAPSE WITHOUT A RESPONSE BEFORE IT WOULD TERMINATE. IN THE NO-DELAY CONTINGENCY A RESPONSE IMMEDIATELY TERMINATED IT, AND IT WAS THEN REPROJECTED. THE SUBJECTS WERE TRAINED UNTIL THEY WENT THROUGH ONE SESSION OF 20 PRESENTATIONS OF EACH STIMULUS WITH NO S- RESPONSES. SUBJECTS TRAINED WITH THE FADING PROCEDURE MADE SIGNIFICANTLY FEWER RESPONSES IN THE COURSE OF TRAINING THAN SUBJECTS TRAINED WITH THE CONSTANT PROCEDURE. SUBJECTS TRAINED WITH THE CONSTANT-DELAY PROCEDURE MADE SIGNIFICANTLY FEWER RESPONSES THAN THE SUBJECTS TRAINED WITH THE CONSTANT-NO DELAY PROCEDURE. IN THE LATTER, IT WAS FELT THAT THE OFFSET OF THE S- STIMULUS WITH A RESPONSE WAS ACTING AS A REINFORCEMENT. (DR)