Bales' Interaction Process Analysis, which is an observational method for the study of social behavior of persons in small groups, was applied to attorney-witness interactions in actual trials in Townsville, Queensland over 3 months. Profiles of responses from courtroom interactions were compared with those from other environments. Trial participants were found to experience few tension-reduction behaviors when compared to other environments. It was also found that when counsel were higher on suggestion-giving, witnesses were higher on both agreeing and negative responses; whereas, when more information was requested, more information was supplied. Further, information-request and information-supply was lower during cross-examination. During cross-examination, attorney power in the interaction increased while amicability decreased. The failure of trial interactions to reduce tension levels may indicate a serious reduction in the effectiveness of the communication process. The tension in the interaction between witness and attorney can reduce recall and the accuracy of recall. The study lends support to those who claim that the adversary process is a poor method for producing evidence, since the polarizing efforts to admit, exclude, restrict, and interpret information-giving are clouded by the biases inherent in the win-lose postures of the competing attorneys. Tabular data and 21 references are provided.