Investigated were the effects of varying instructional explicitness (minimal, general, and explicit) and types of reinforcement (none, extrinsic, and intrinsic or relational) on the learning of an oddity discrimination task by 48 nursery school children. Ss were randomly assigned to six groups where general or minimal instructional explicitness was reinforced in one of three ways. Twelve other children were assigned to a contrast group that received particularly explicit instructions and intrinsic reinforcement. No significant differences in learning were found among the six groups of the major design, but the contrast group was found to perform significantly better when compared with groups receiving minimal and general instruction. No significant difference emerged from an analysis of the contrast condition with the three reinforcement conditions. Different stimuli were used in a test of transfer of training in which 32 of 34 Ss who had earlier reached criterion made successful transfers. A significant correlation of chronological age and original learning was found across the total subject sample, but not within any group or condition except the minimal instruction-no reinforcement condition. Data from two other samples involved in a replication study and an initial pilot study were also reported. (Author/GW)