that occur; and if that percentage is decidedly exceeded, there is reason to surmise that some other cause than chance--in other words, some objective origin for the phantasm--is present." But on the other hand, a phantasm representing a person whose death is recent is specially likely to arouse interest and, in cases where the death is previously known to the percipient, his emotional state may be considered a sufficient cause of the hallucination. "If, then," Gurney continues, "we are to draw any probable conclusion as to the objective nature of _post-mortem_ appearances and communications (or of some of them) from the fact of their special frequency soon after death, we must confine ourselves to cases where the fact of death has been unknown to the percipient at the time of his experience. Now, in these days of letters and telegrams, people for the most part hear of the deaths of friends and relatives within a very few days, sometimes within a very few hours, after the death occurs; so that appearances of the sort required would, as a rule, have to follow very closely indeed on the death. Have we evidence of any considerable number of such cases? "Readers of _Phantasms of the Living_ will know that we have. In a number of cases which were treated in that book as examples of telepathic transference from a dying person, the person was actually dead at the time that the percipient's experience occurred; and the inclusion of such cases under the title of _Phantasms of the Living_ naturally occasioned a certain amount of adverse criticism. Their inclusion, it will be remembered, required an assumption which cannot by any means be regarded as certain. We had to suppose that the telepathic transfer took place just before, or exactly at, the moment of death; but that the impression remained latent in the percipient's mind, and only after an interval emerged into his consciousness, whether as waking vision or as dream or in some other form. Now, as a provisional hypothesis, I think that this assumption was justified. For in the first place, the moment of death is, in time, the central point of a cluster of abnormal experiences occurring to percipients at a distance, of which some _precede_, while others follow, the death; it is natural, therefore, to surmise that the same explanation will cover the whole group, and that the motive force in each of its divisions lies in a state of the 'agent' prior to bodily death. In the second place, some of the facts of experimental thought-transference countenance the view that 'transferred impressions' may be latent for a time before the recipient becomes aware of them; and recent discoveries with respect to the whole subject of automatism and 'secondary intelligence' make it seem far