The facts, moreover, condemn Janet’s theory. I have too high an opinion of the distinguished man whose ideas I criticise, but whose works I admire sincerely, not to be convinced that he has only observed undeveloped subjects. What demonstrates this in my eyes is his timid affirmation, that ‘nearly always (I do not say always in order not to prejudice an important question) these mediums are neurotics, when they are not downright hysterics.’ It is difficult to discuss an opinion expressed with so much reserve, and I can only commend him for his circumspection, for my personal observations contradict his. I have seen many mediums: the best were not _neurotics_ in the medical sense of the word. The finest experiments I have made have been with persons appearing to present none of the stigmæ of hysteria. Up to the present Janet seems to have operated with invalids only, and I am not surprised, therefore, that he should assimilate the automatic phenomena of sensitives with those of his hysterical patients. It would be surprising were it otherwise. I am not going to defend spiritistic mediums; they appear to me to present very poor interest—at least in ordinary seances—but my duty is to protest against the generality of the judgment which Janet brings to bear upon automatic phenomena. Those facts, which are worthy of careful observation, differ essentially from those which ordinary hysterics present. They indicate no _misère psychologique_—quite the contrary, and I will state the reasons why. The discussion, in order to be clear, must be divided: 1. The phenomena observable with good mediums are not those we observe in hysterical patients. I said I had obtained raps and movements without contact under conditions of control, which appeared to me to be convincing. I added that I had obtained by raps, or by the rappings of a table without contact, words and phrases which were extremely coherent. This is not quite the kind of phenomena to which hospital patients have accustomed us. What does Janet say on this point?