This study is a partial replication of Rosenthal's (1968) study of teacher expectation in which students falsely identified to teachers as being capable of doing better work showed greater intellectual growth than control students. The present study sought to determine whether or not this result could be duplicated and whether or not teacher-student classroom interaction would be affected by the information. False information was leaked to first and second grade teachers about certain of their students who were supposed to have special late-blooming academic abilities. Four experimental students and four control students were randomly selected from each of 18 classes. A total of 144 students were pre- and posttested on the Test of General Ability and the reading and arithmetic subtests of the Metropolitan Achievement Tests. Classroom interaction was pre- and post-measured by use of the Interaction Analysis Scale. After posttesting, teachers also completed questionnaires concerning student expectancy. Results revealed that the experimental group did not make significantly higher gains than the control group on any of the tests. In addition, classroom interaction measures showed no special treatment accorded the "special" students. (MH)