Ten experiments examined the effect on the performance of 265 nursery schoolers through first graders using collection versus class terms to describe sets of familiar objects. No facilitative effect of collection terms occurred on number tasks; in a class-inclusion task, performance was better with collection than class terms. (SLD)