In order to provide an estimate of encoding within the active mode and a comparison between three modes of representation (enactive, imagery, and verbal), 36 second grade children from the high socioeconomic community of Beverly Hills, California, were presented with object pairs under one of three conditions (haptic-object versus visual-object versus aural-label). The pairs were presented for old/new recognition response, and it was found that the children responded better to visual-object than to haptic-object and aural-label presentation. In a second experiment, 54 kindergarden children were presented single objects in the haptic-object mode of study, in order to assess retrieval from the enactive mode. Old/new recognition testing was conducted for objects presented in one of three modes. The pattern of test mode performance suggested that each of the incongruent test conditions produced a decrement in performance relative to the congruent test mode; however, reliable differences between the test modes were not detected. The pattern of performance suggests that visual imagery is a more hospitable mode for the storage of concrete information than either the enactive or verbal, while the similarity of performance in the second experiment suggests that subjects were reasonably efficient at accessing information stored within the enactive mode. (JF)