(IV.) Passing on to phantasms which cluster about the moment of death, we find our three main classes of cases still meeting us. Our readers are familiar with the _visual_ cases, where there is an actual apparition of the dying man, seen by one or more persons; and also with the _emotional and motor_ cases, where the impression, although powerful, is not definitely sensory in character. And various cases also have been published where the message has consisted of definite words, not always externalised as an auditory hallucination, but sometimes automatically _uttered_ or automatically _written_ by the percipient himself, as in the case communicated by Dr. Liébeault (see Appendix VIII. C), where a girl writes the message announcing her friend's death at the time when that friend is, in fact, dying in a distant city. (V.) And now I maintain that in these post-mortem cases also we find the same general classes persisting, and in somewhat the same proportion. Most conspicuous are the actual _apparitions_, with which, indeed, the following pages will mainly deal. It is very rare to find an apparition which seems to impart any verbal message; but a case of this kind has been given in Appendix IV. F. As a rule, however, the apparition is of the apparently automatic, purposeless character, already so fully described. We have also the _emotional and motor_ class of post-mortem cases;[131] and these may, perhaps, be more numerous in proportion than our collection would indicate; for it is obvious that impressions which are so much less definite than a visual hallucination (although they may be even more impressive to the percipient himself) can rarely be used as evidence of communication with the departed. But now I wish to point out that, besides these two classes of post-mortem manifestations, we have our _third_ class also still persisting; we have definite verbal messages which at least purport, and sometimes, I think, with strong probability, to come from the departed. I have, indeed, for the reader's convenience, postponed these motor cases to a subsequent chapter, so that the evidence here and now presented for survival will be very incomplete. Yet, at any rate, we are gradually getting before us a fairly definite task. We have in this chapter to record and analyse such sensory experiences of living men as seem referable to the action of some human individuality persisting after death. We have also obtained some preliminary notion as to the kind of phenomena for which we can hope, especially as to what their probable limitations must be, considering how great a gulf between psychical states any communication must overpass.