(II.) Now turn to our experiments in thought-transference. Here again the agent is a living man; but he is no longer operating by ordinary means,--by spoken words or visible gestures. He is operating on the percipient's subconscious self by means of a telepathic impulse, which he desires, indeed, to project from himself, and which the percipient may desire to receive, but of whose _modus operandi_ the ordinary waking selves of agent and percipient alike are entirely unaware. Here again we may divide the messages sent into the same three main classes. First come the hallucinatory figures--always or almost always of himself--which the agent causes the percipient to see. Secondly come impulses to act, telepathically impressed; as when the hypnotiser desires his subject to come to him at an hour not previously notified. And thirdly, we have a parallel to the post-hypnotic writing of definite words or figures in our own experiments on the direct telepathic transmission of words, figures, cards, etc., from the agent, using no normal means of communication, to the percipient, either in the hypnotised or in the waking state. (III.) We come next to the spontaneous phantasms occurring during life. Here we find the same three broad classes of messages, with this difference, that the actual apparitions, which in our telepathic experimentation are thus far unfortunately rare, become now the most important class. I need not recall the instances given in Chapters IV. and VI., etc., where an agent undergoing some sudden crisis seems in some way to generate an apparition of himself seen by a distant percipient. Important also in this connection are those apparitions of the _double_, where some one agent is seen repeatedly in phantasmal form by different percipients at times when that agent is undergoing no special crisis. Again, among our telepathic impressions generated (spontaneously, not experimentally) by living agents, we have cases, which I need not here recapitulate, of pervading sensations of distress; or impulses to return home, which are parallel to the hypnotised subject's impulse to approach his distant hypnotiser, at a moment when that hypnotiser is willing him to do so. And thirdly, among these telepathic communications from the living to the living, we have definite sentences automatically written, communicating facts which the distant person knows, but is not consciously endeavouring to transmit.