Students were pretested, then 27 first graders were trained using a highly verbal procedure emphasizing class addition activities. On posttests, trained subjects significantly improved their class inclusion and number conservation scores, but no effects were found for substance conservation or transitivity. (DT)