The second research report in a continuing investigation of group communication via computer networks describes a series of field tests designed to identify the basic parameters of human interaction in computer-based teleconferencing. The method used in the tests is described, and the research approach used in analyzing the tests is outlined. Ten conferences involving small groups ranging from 3 to 30 participants are described in detail and analyzed according to a taxonomy of group communication developed in collaboration with other research groups. The possible impact of computer-based communication on work patterns and its potential as a medium for networks of disseminated persons are explored and evaluated. A statement of statistical results and a bibliography are appended. (Author/SK)