The present study evaluated the effectiveness of a simple successive (go/ no-go) discrimination procedure in producing functional class formation as a systematic replication of a previous study with simple simultaneous discrimination. Five preschool children learned 3 simple discriminations and were given repeated reversal training with all 6 stimuli simultaneously. Four of these children learned a fourth discrimination and a new series of reversals with 8 stimuli. Rapid reversals (learning set) occurred with 6 stimuli in 3 children and 8 stimuli in 4 children. Two children presented performance compatible with functional class formation. The results were similar to those reported in the previous study. Similar to the simultaneous procedure, the successive procedure effectively generated functional class formation with some, but not all, of the participants. The occurrence of errors inherent to the reversal procedure may be a source of variability. Future research should investigate other strategies to promote stimulus class formation on the basis of shared discriminative functions.