The first mistake, committed by the distinguished members of the Sidgwick group, was to suppose that psychical phenomena can be obtained at will. Whenever a paid medium gives regular seances, there are a hundred chances to one of downright fraud. If there be a positive feature in these supernormal facts, that feature in my opinion is their apparent irregularity. I have been able to experiment with intelligent, well-educated mediums anxious for a thorough investigation of their powers: I have made very many experiments with them, and I have observed that often whole weeks passed away without a good seance; at other times, the force was so abundant that phenomena were forthcoming without seance. I have related some curious facts in this respect, _e.g._ the table moving spontaneously in the course of a conversation bearing upon psychical phenomena (p. 106). What are the conditions which impede or favour the production of this unknown mode of energy? I cannot specify them; but I think I have noticed concordances, which confirm in a measure the conclusions of Ochorowicz (_Annales des Sciences Psychiques_, vi. 115):— 1. Action of temperature. Dry cold weather is the most favourable. Damp or close weather is most unfavourable. 2. Health of the medium and sitters. If the medium does not feel well, things happen as though he exteriorised no force whatever. It is the same thing with the sitters, but in a lesser degree; in the latter case it suffices to eliminate the experimenter who feels ill. 3. Mental condition of the medium and sitters.[40] Ill-humour, anxiety, sadness—especially a sadness without any specific cause, a kind of mental discomfort—are prejudicial. Joy, gaiety are often favourable. [40] There are _apparent_ exceptions to this rule. 4. Nervous exhaustion. This condition is too often overlooked. I have not unfrequently had occasion to conduct several series of experiments at one and the same time. I generally noticed that the results were not good. I have not been able to understand the cause of this want of success; it is probably other than that of simple nervous exhaustion, although this may have an action in prolonged series of seances. Neither do seances held too frequently with the same medium give good results; in this case, nervous exhaustion is certainly in play. The English experimenters do not appear to have taken these diverse elements into consideration; I am persuaded the results of their investigations would have been different had they shunned ‘paid mediums,’ and sought for fresh or undeveloped mediums, persons uninfluenced by private considerations, intelligent and capable of bringing a correct analysis of their subjective impressions into the research. These mediums are rare, but they are to be found.