THREE FIFTH-GRADE CLASSES OF ECONOMICALLY DEPRIVED NEGRO CHILDREN, EQUATED ON INTELLIGENCE AND READING ACHIEVEMENT, PARTICIPATED IN A STUDY OF THE EFFECTS OF VARYING AMOUNTS AND TYPES OF RACIAL CUEING ON AFFECTIVE SETS TOWARD THE PROTAGONIST AND ANTAGONIST IN A CLOZE TYPE READING SELECTION. ALL THREE CLASSES READ THE SELECTION, BUT CLASS A WAS PRESENTED PICTURES OF THE INTERACTION OF THE TWO MAIN CHARACTERS WITH EACH OTHER AND WITH PEERS. CLASS B HAD SINGLE PICTURES OF THE TWO MAIN CHARACTERS ONLY WHO WERE IDENTIFIED AS A NEGRO PROTAGONIST AND A WHITE ANTAGONIST, AND CLASS C DID NOT RECEIVE ANY RACIAL CUEING. EACH CHILD RATED THE PROTAGONIST AND ANTAGONIST ON A SEMANTIC DIFFERENTIAL SCALE ADAPTED FROM OSGOOD'S SEMANTIC DIFFERENTIAL. IT WAS FOUND THAT THE CLEAR IDENTIFICATION OF THE RACIAL BACKGROUND OF THE PROTAGONIST AND ANTAGONIST AND THE USE OF SINGLE PICTURE CUES RESULTED IN A POSITIVE RATING OF THE PROTAGONIST AND A NEGATIVE RATING OF THE ANTAGONIST. CLASS B ALSO USED MORE EXACT LANGUAGE TO ACHIEVE CLOSURE. THE USE OF PICTURE CUES ONLY AND THE LACK OF RACIAL CUEING, WHETHER PICTORIAL OR TEXTUAL, RESULTED IN PERCEPTION OF THE ANTAGONIST AS AN IDEAL TYPE HAVING POSITIVE EVALUATION AND STRENGTH AND OF THE PROTAGONIST AS HAVING POISTIVE AND STRONG TYPE CHARACTER. CLASSES A AND C USED LANGUAGE LIMITED TO SYNONYMS AND THE GROUPS' OWN LANGUAGE TO OBTAIN CLOSURE. THIS PAPER WAS PRESENTED AT THE AMERICAN EDUCATIONAL RESEARCH ASSOCIATION CONFERENCE (CHICAGO, FEBRUARY 6-10, 1968). (NS)