The conversion to psychoanalysis from psychotherapy with the same analyst is a recent phenomenon. In 1976 only 12% of analytic cases had first been in psychotherapy with their analyst; 40% had had psychotherapy with a different analyst. Horwitz discusses the desirability of psychotherapy as a trial for analysis, rather than a trial analysis in those cases where there is serious doubt about a patient's analyzability. He points out that although a trial analysis was recommended by Freud, psychoanalytic psychotherapy as we now know it was not then available. He suggests that a failed "trial analysis" may be unnecessarily traumatic for these questionable cases. Some recent literature regarding the distinction between psychotherapy and psychoanalysis, examples of converting from therapy to analysis, and therapy as preparatory for analysis is reviewed. Horwitz distinguishes three broad categories of patients started in therapy and later considered for analysis: those with questionable psychological makeup, those in whom reality factors (schedule, finances) play a