Five actors were videotaped and audiotaped during 12 performances, in order to identify variables and to observe actors' behaviors and attitudes in positive, negative, and "unclaqued" audience response conditions. Audience response was effected by actors trained in nonverbal cues and placed in the audience as "claquers," to subtly influence observable response. The actors, the director, and expert evaluators rated all performances. Videotapes were analyzed for physical behaviors, and audiotapes were analyzed for dysfluencies and rate, as suggested by speech communciation feedback research. The data are biased by the actors' awareness of experimental observation; consequently, performances were highly consistent and had no significant differences in nonverbal behavior. Actors' and experts' evaluations correlated significantly: ratings generally were favorable in the positive set, slightly less favorable in the "unclaqued" set, and distinctly less favorable in the negative set. The variable "in character" was differentiated from all other variables. These findings support the conclusion that, in terms of the source/receiver relationship, rhetoric and theatre are not comparable communication events. (Author/RL)