414. I pass on to the still more novel and curious questions involved in the apparent existence of a dream-memory which, while accompanying the memory of ordinary life, seems also to have a wider purview, and to indicate that the record of external events which is kept within us is far fuller than we know.

Let us consider what stages such a memory may show.

I. It may include events once known to the waking self, but now definitely forgotten.

II. It may include facts which have fallen within the sensory field, but which have never been supraliminally "apperceived" or cognised in any way. And thus also it may indicate that from this wider range of remembered facts dream-inferences have been drawn; - which inferences may be retrospective, prospective, or, if I may use a word of Pope's with a new meaning, circumspective, - that is to say, relating not to the past or to the future, but to the present condition of matters beyond the range of ordinary perception. It is plain that inferences of this kind (if they exist) will be liable to be mistaken for direct retrocognition, direct premonition, direct clairvoyance; while yet they need not actually prove anything more than a perception on the part of the subliminal self more far-reaching, a memory more stable, - than is the perception or the memory of the supraliminal self which we know.

These hypermnesic dreams, then, may afford a means of drawing our lines of evidence more exactly; of relegating some marvellous narratives to a realm of lesser marvel, and at the same time of realising more clearly what it is in the most advanced cases which ordinary theories are really powerless to explain.

415. As to the first of the above-mentioned categories no one will raise any doubt. It is a familiar fact - or a fact only sufficiently unfamiliar to be noted with slight surprise - that we occasionally recover in sleep a memory which has wholly dropped out of waking consciousness. As an example, we may take the dream of M. Delboeufs, discussed in his interesting book, Le Sommeil et les Rives. In that dream the name of the "Asplenium Ruta Muralis" figured as a familiar phrase. On waking, he puzzled himself in vain to think where he could have learnt that botanical appellation. Long afterwards he discovered the name "Asplenium Ruta Muraria" in his own handwriting, - in a little collection of flowers and ferns to which he had added their designations, under the dictation of a botanical friend.

In this and similar cases the original piece of knowledge had at the time made a definite impress on the mind, - had come well within the span of apprehension of the supraliminal consciousness. Its reappearance after however long an interval is a fact to which there are already plenty of parallels. But the conclusion to which the cases about to be cited seem to me to point is one of a much stranger character. I think that there is evidence to show that many facts or pictures which have never even for a moment come within the apprehension of the supraliminal consciousness are nevertheless retained by the subliminal memory, and are occasionally presented in dreams with what seems a definite purpose.1

The same point, as we shall hereafter see, is illustrated by the phenomena of crystal-vision. Miss Goodrich-Freer,2 for example, saw in the crystal the announcement of the death of a friend; - a piece of news which certainly had never been apprehended by her ordinary conscious self. On referring to the Times, it was found that an announcement of the death of some one of the same unusual name was contained in a sheet with which she had screened her face from the fire; - so that the words may have fallen within her range of vision, although they had not reached what we broadly call her waking mind.

This instance was of value from the strong probability that the news could never have been supraliminally known at all; - since it was too important to have been merely glanced at and forgotten.