The study aimed to estimate the observers' impression after watching two kinds of animation: the sequence of two figures chasing each other, similar to that described by Heider & Simmel (1944), and the sequence in which two figures walk through a pedestrian crossing. The order in which two different sized figures appeared was also manipulated. To assess the observer's impression, systematically the author used comprehensive adjectives which appeared in the descriptive sentences more than two times in the preceding experiment (Mori, Tuchida, & Jingu, 2000). After watching each animation, the subjects (130 students) evaluated the impression using 19 available adjectives. A factor analysis of the ratings found four factors such as "attractiveness'', "cleverness", "aversion" , and "potency". These factors captured the stories of animation and the characters of the figure recognized as a person. Results of the analysis of variance showed a significant interaction between the order of appearance of two figures and the kinds of animation. The order of appearance of two different sized figures affected the observers' impressions differently in each animation.