They revive in us only as disjointed fragments, and only for brief moments when some accidental resemblance in the present happens to stir the latent trace they have left on our minds.

We get sudden flashes out of our dream-world, and the process is too rapid, too incomplete for us to identify the region whence the flashes come.
It is highly probable that our dreams are, to a large extent, answerable for the sense of familiarity that we sometimes experience in visiting a new locality or in seeing a new face.

If, as we have found some of the best authorities saying, we are, when asleep, always dreaming more or less distinctly, and if, as we know, dreaming is a continual process of transformation of our waking impressions in new combinations, it is not surprising that our dreams should sometimes take the form of forecasts of our waking life, and that consequently objects and scenes of this life never before seen should now and again wear a familiar look.
That some instances of this puzzling sense of familiarity can be explained in this way is proved.

Thus, Paul Radestock, in the work _Schlaf und Traum_, already quoted, tells us: "When I have been taking a walk, with my thoughts quite unfettered, the idea has often occurred to me that I had seen, heard, or thought of this or that thing once before, without being able to recall when, where, and in what circumstances.
This happened at the time when, with a view to the publication of the present work, I was in the habit of keeping an exact record of my dreams.

Consequently, I was able to turn to this after these impressions, and on doing so I generally found the conjecture confirmed that I had previously dreamt something like it." Scientific inquiry is often said to destroy all beautiful thoughts about nature and life; but while it destroys it creates.