It is difficult to distinguish variations from modifications; intuitive insight is often the guide to deviations from the rules. Variations in classical technique must be directed toward establishing and maintaining the optimal distance ('rapprocher') between patient and analyst. Both activity and passivity are determined by this need, especially for the psychotic; the level of anxiety and the role of projection are guides to what is needed. Deliberate technical variations should be introduced only when indispensable for fostering (and later reducing) the transference neurosis; such variations should be subject to the rules laid down by Eissler concerning parameters. If a particular variation in technique produces the desired 'distance', it may be viewed as correct analytically. Bouvet illustrates in precise fashion, by a case history, his successful use of seven parameters during the analysis of a seriously disturbed (probably psychotic) patient.