Using the materials and general procedure described by Rock and Gutman (1981), 160 subjects were required to attend selectively to a series of 10 red or green outlines of two overlapping contours. Immediately afterwards, six groups of 20 subjects were administered various kinds of recognition tests involving attended, unattended, or new figures, and two groups of 20 were asked to provide affect ratings of the test stimuli. Consistent with the findings of Rock and Gutman in 1981, attended outlines were selected as old more often than unattended or new ones. Although some data in a single-stimulus condition suggested that the unattended forms were "discriminated negatively," this effect was not replicated with forced-choice testing. Over-all, there was little evidence that, on unattended forms, subjects were significantly more accurate, differentially confident or displayed more liking than on new ones. These results are interpreted as supporting the view that attention is required for the cognitive representation of form.