White subjects were asked to evaluate toddler behavior after viewing a 14-minute film of a white child and a black child playing spontaneously. Each toddler was presented as both a male and a female to 208 well-educated adult subjects. In Treatment I, the children were called John and Mike; in Treatment II, Jane and Mary; in Treatment III, John and Mary; and in Treatment IV, Jane and Mike. Subjects rated filmed toddler behaviors on a specially designed instrument consisting of 12 adjective pairs placed at opposite ends of a seven-choice Likert-type scale. Thirteen pairs centered on gender, 13 pairs on ethnicity, 10 pairs on both, and 10 pairs on neither. The toddlers were generally seen as significantly different on all variables except "bright-dull" and "handsome-beautiful." On six variables, the gender of the subject made a significant difference in responses. For two variables, the treatment situation made a difference in how the toddlers were perceived. The dependent variable "handsome-beautiful" showed an interaction depending upon whether the toddlers were believed to be male or female. This interaction acted as a rough validity check to demonstrate that the procedure and instrument were sensitive enough to pick up statistically significant differences in viewer perceptions. It could not be ascertained whether the subjects perceived or were reacting to ethnic stereotypes. The need for black subjects was also noted. (Author/RH)