College students and adults were asked to make random patterns by either arranging patches of colored paper or filling in small graph-paper squares. Variations both in number of constituent elements and serial ordering were administered to separate groups. The most frequently administered instructions required Ss to randomize the sequence order as well as the constructed pattern. Significant exact balancing of pattern elements around both vertical and horizontal axes was found together with a dispersal of elements in excess of mathematical expectation. These results were summarized as the “tendency toward equal density of distributed elements.” Ss performing sequence randomization fitted a distribution of expected run-lengths very well. It is argued that whether either the sequential or pattern results are considered to have met the criteria for randomness depends on the standpoint from which randomness is referenced. Three possible reference points are considered for present results: (a) the total frequency distribution, (b) the individual S, and (c) the individual event S performs. Visual randomness for the artist does not imitate the visual randomness consensus of the nonartist.
