Certainly, the attacks made on Eusapia Paladino by a badly informed press and public are not encouraging to the more highly gifted mediums. I owe it to Eusapia to say that, in my experiments with her, she has always submitted to the exigencies of the most severe test conditions. If she has sometimes given me suspicious phenomena, she did so only under especial psychological conditions.[1] [1] See Appendix B. Though I have not employed any registering apparatus, I have used instruments of weight and measure—particularly a letter-balance—an article as convenient as it is easily employed. Each experimenter can and ought to vary the conditions of experimentation according to his wishes, within the limits which frequent experimentation will very quickly give him. The results obtained must be definite. To be satisfied with approximate results in such a matter would be absolute loss of time. In concluding my remarks about the paraphernalia of the seance-room, I will give one more recommendation which may seem extraordinary, but which, I have reason to believe, is useful; this is that there should be no metal about the table: it is better to fasten it together with pegs rather than with nails. This is not an absolute condition, however, for I have obtained good results with nailed tables; yet my impression is that the absence of all metal is an element of success. Mediums are sometimes extremely sensitive to metals. Certain sensitives complain of their rings, which seem to make them feel uncomfortable, giving them, at times, a sensation of exaggerated heat. This brings to mind certain facts met with from time to time in our neurotic cliniques. II. COMPOSITION OF THE CIRCLE The most important thing in the organisation of a series of experiments is the choice of persons with whom we intend to operate. First of all, it must be remembered that without a _medium_ no phenomena will be forthcoming. The presence of some one, gifted with the power of producing psychical phenomena, is perhaps the only necessary and indispensable condition of their realisation. Therefore, experimentation ought only to be seriously thought of when in possession of that _rara avis_. What, then, is a medium? By what distinguishing features can he be recognised? It is very difficult to answer these questions.