Eight experiments using as Ss either retarded children, normal children, or normal adults studied the relations of retardation and normal development to the perceptual process of identification. Two experiments were reported on the identification of stimuli varying in either one, two, or three dimensions. Retardates did not perform as well as normals of equal chronological or mental age. Multidimensionally varying stimuli were easier to identify than those varying undimensionally. In three experiments efficiency in identification and discrimination tasks was evaluated. The effects of dimensional combination and stimulus distinctiveness were similar both in normal adults and retarded children. Success was reported in increasing the efficiency of retardate performance by first presenting a less demanding memory task (discrimination) and moving to a more demanding task (match-to-sample identification). Three experiments evaluated short term memory coding. Memory effects were found in all studies in which delay between stimulus presentation and response was varied. Explicit coding was not found to be effective at very short presentation times when the code was in octal form. The coding involved in remembering word-like material as opposed to non-word-like material was effective at even ten milliseconds presentation time. The rapid coding appeared to be under the subject's conscious control. (Author/DB)