Two patients' sessions were recorded for several months. Two analysts were asked to judge the material and two to predict future behaviour, according to a prescribed record-form. The results were then assessed statistically. A high degree of agreement was found among all the participants concerning the nature of the case. The ability to agree on the nature of short-range changes was not unequivocally demonstrated. None the less, in spite of the differences in conceptualization and focus of interest between the analysts, the overall agreement was impressive. The experimental procedure is discussed in detail, with its usefulness for the study and validation of specific hypotheses in psycho-analysis.