Pictures of objects were recalled significantly better than their names on the first two of four free recall trials. Recall for the two modes did not differ in intertrial organization but striking differences occurred as a function of input serial order. Picture superiority occurred for terminal input items on Triali, and both terminal and early items on Trial 2. The findings are discussed in terms of verbal and nonverbal (concrete) memory codes.Free verbal recall is generally higher for objects or pictures than for their labels (e.g., Ducharme & Fraisse, 1965; Kirkpatrick, 1894; Lieberman & Culpepper, 1965). Although rarely discussed, the finding is of uncommon theoretical interest because verbal coding processes alone seem insufficient to account for it. If only verbal processes were involved, one would expect either no difference between words and objects, or superior recall for the former because they more readily evoke the appropriate (implicit) verbal response: words can be read faster than objects can be named (Fraisse, 1964). Since objects are better recalled, nonverbal processes must somehow contribute to the retrieval process. How migh t this occur?Four theoretical possibilities suggest themselves. The effective variable might be some physical stimulus characteristic such as vividness, or it might be some coding (meaning) process aroused by the stimulus. Within either of these are two further alternatives: the variable affects the recall of the items as independent units, or provides a basis for organizing them into higher-order units (Tulving, 1968). Prior research does not permit a choice among the alternatives. Bousfield, Esterson, & Whitmarsh (1957) found recall to be highest for nouns presented along with colored pictures, next for nouns with uncolored pictures, and lowest for nouns alone. They interpreted their findings in terms of the compounding of stimulus elements, in a conditioning framework, which implies that the effective variables were physical stimulus characteristics. Their results and interpretation may not be relevant to the present problem, however, because words were not compared to pictures alone.The present investigation compared colored and uncolored sets of both pictures and their noun labels. To the extent that either physical vividness or compounding of stimulus elements is effective, recall should be better for colored versions of words as well as pictures. To the extent that color is ineffective, concrete meaning may be the important factor. Organizational processes should be reflected in intertrial (subjective) organization.Materials. The items consisted of the pictures and the names of 25 familiar objects, selected according to the criterion that the pictures are reliably coded by the particular words used (see Paivio & Yarmey, 1966). The items were: hand, cigar, ladder, bread, horse, soldier, microscope, telephone, kettle, scissors, pencil, fish, lobster, apple, star, leopard, bottle, tree, clock, flag, radio, umbrella, book, stove, and knife. For the black and white co...