Let us then consider the specific characteristics of sleep. The definition of sleep is an acknowledged _crux_ in physiology. And I would point out that the increased experience of hypnotic sleep which recent years have afforded has made this difficulty even more striking than before. A physiological explanation must needs assume that some special bodily condition,--such, for instance, as the clogging of the brain by waste-products,--is at least the usual antecedent of sound sleep. But it is certain, on the other hand, that with a large percentage of persons profound and prolonged sleep can be induced, in _any_ bodily condition, by simple suggestion. Hypnosis, indeed (as Wetterstrand and others have shown) may be prolonged, with actual benefit to the sleeper, far beyond the point which the spontaneous sleep of a healthy subject ever reaches. A good subject can be awakened and thrown into hypnosis again almost at pleasure, and independently of any state either of nutrition or of fatigue. Such sleep belongs to those phenomena which we may call nervous if we will, but which we can observe or influence from the psychological side alone. We can hardly hope, from the ordinary data, to arrive at a definition of sleep more satisfactory than others have reached. We must defer that attempt until we have collected something more than the ordinary evidence as to what occurs or does not occur during the abeyance of waking life. One point, however, is plain at once. We cannot treat sleep,--as it has generally been treated,--in its purely _negative_ aspect. We cannot be content merely to dwell, with the common text-books, on the mere _absence_ of waking faculties;--on the diminution of external perception, the absence of controlling intelligence. We must treat sleep _positively_, so far as we can, as a definite phase of our personality, co-ordinate with the waking phase. Each phase, as I believe, has been differentiated alike from a primitive indifference;--from a condition of lowly organisms which merited the name neither of sleep nor of waking. Nay, if there were to be a contest as to which state should be deemed primary and which secondary, sleep might put forward its claim to be regarded as the more primitive phase. It is sleep rather than vigilance which prenatal and infantile life suggest; and even for us adults, however much we may associate ourselves in thought with the waking state alone, that state has at least thus much of secondary and adventitious that it is maintained for short periods only, which we cannot artificially lengthen, being plainly unable to sustain itself without frequent recourse to that fuller influx of vitality which slumber brings.