These cases of invasion by the spirits of living persons pass on into cases of invasion by the dying, the impression being generally that of the presence of the visitant in the percipient's surroundings.[54] Sometimes the phantasm is seen as nearly as can be ascertained at the time of death. But there is no perceptible break in the series at this point. Some appear shortly after death, but before the death is known to the percipient. [See Appendix IV. F]. Finally, there are cases when the appearance takes place some time after death, but presents features unknown to the percipient.[55] We have now briefly reviewed certain phenomena of sleep from a standpoint somewhat differing from that which is commonly taken. We have not (as is usual) fixed our attention primarily on the _negative_ characteristics of sleep, or the extent to which it lacks the capacities of waking hours. On the contrary, we have regarded sleep as an independent phase of personality, existing with as good a right as the waking phase, and dowered with imperfectly expressed faculties of its own. In investigating those faculties we have been in no wise deterred by the fact of the apparent uselessness of some of them for our waking ends. _Useless_ is a pre-scientific, even an anti-scientific term, which has perhaps proved a greater stumbling-block to research in psychology than in any other science. In science the _use_ of phenomena is to prove laws, and the more bizarre and trivial the phenomena, the greater the chance of their directing us to some law which has been overlooked till now. In reviewing the phenomena of sleep, then, we found in the first place that it possesses a specific recuperative energy which the commonly accepted data of physiology and psychology cannot explain. We saw that in sleep there may be an increased co-ordination or centralisation of muscular control, and also an increased vividness of entencephalic perception, indicating a more intimate appreciation of intra-peripheral changes than is manifest in waking life. In accordance with this view, we found that the dreaming self may undergo sensory and emotional experiences apparently more intense than those of vigilance, and may produce thereby lasting effects upon the waking body and mind. Similarly again, we saw that that specific impress on body and mind which we term memory may in sleeping or hypnotic states be both wider in range and fuller in content than the evocable memory of the waking day. Nay, not memory only, but power of inference, of argument, may be thus intensified, as is shown by the solution in sleep of problems which have baffled waking effort.