Seventy-two children, 36 each from the first and sixth grades, were given a70-item study list followed by a recognition test list of equal length. Items were line drawings of common objects accompanied by orally presented verbal labels. To provide an indication of the relative dominance of features encoded at the two ages, labels, pictures, and referents were manipulated in various combinations to form visual, semantic, and acoustic distractors, which were included in the test list. Significant numbers of false recognitions (as measured against false-recognition rates for control words) were elicited by visual distractors for both age groups, by acoustic distractors for first-grade subjects and by semantic distractors for sixth graders and first-grade girls. Implications of these findings for hypotheses concerning the role of visual representation and the developmental onset of semantic encoding are discussed.