Another well-observed fact, opposed to the telepathic theory, is the selection made amongst incidents by the communicator. If we were dealing with telepathy, the secondary personalities of the medium would sometimes be mistaken, make blunders, record facts which the so-called communicator could never have known, but which the sitter alone knows well. Now this never happens. The reported facts are always common to at least two consciousnesses, that of the communicator and that of the sitter, or that of the communicator and that of a distant person. The inaccuracies prove nothing against this argument; if they are wilful falsehoods they simply prove that the communicator is a liar, and not that he is a secondary personality of Mrs Piper. If the reported facts are unverifiable, this does not prove that they are inexact. If the telepathic theory expresses the truth, we must grant an almost infinite power to telepathy. This supposition is indispensable to account for the facts. Then how shall we understand the errors and confusions of the communicators? How can an infinite power seem at times so limited, so finite, when the conditions remain unchanged? On the other hand, the lapses of memory and confusions are quite explicable on the spiritualistic theory; we cannot reasonably think that a change so great as death should not induce some disturbance of mind, at least temporarily, or should not greatly weaken certain groups of memories which, in the new surroundings, have no longer any practical use. A change of communicators has always been frequent, but was especially so in Professor Hyslop's sittings. Mr Robert Hyslop constantly says to his son, "James, I am getting weak; wait for me, I am coming back." And then another communicator appears on the spot. The telepathic hypothesis cannot explain this fact; it would seem quite natural that the communicator should be always the same. To explain it, another hypothesis--that of suggestion on the part of the sitter--must be added to the telepathic hypothesis. But the spiritualistic hypothesis, on the other hand, explains this perfectly well, even though we may be compelled to reckon with the complications which the admission of the existence of another world may introduce.