politics (compare the warming-pan lie of 1688), in the telling of
ghost stories a different plan has its merits. Beginning with the
common-place and familiar, and therefore credible, with the thin end
of the wedge, in fact, a wise narrator will advance to the rather
unusual, the extremely rare, the undeniably startling, and so arrive
at statements which, without this discreet and gradual initiation, a
hasty reader might, justly or unjustly, dismiss as "great swingeing
falsehoods".

The nature of things and of men has fortunately made this method at
once easy, obvious, and scientific. Even in the rather fantastic
realm of ghosts, the stories fall into regular groups, advancing in
difficulty, like exercises in music or in a foreign language. We
therefore start from the easiest Exercises in Belief, or even from
those which present no difficulty at all. The defect of the method is
that easy stories are dull reading. But the student can "skip". We
begin with common every-night dreams.

Sleeping is as natural as waking; dreams are nearly as frequent as
every-day sensations, thoughts, and emotions. But dreams, being
familiar, are credible; it is admitted that people do dream; we reach
the less credible as we advance to the less familiar. For, if we
think for a moment, the alleged events of ghostdom--apparitions of all
sorts--are precisely identical with the every-night phenomena of
dreaming, except for the avowed element of sleep in dreams.

In dreams, time and space are annihilated, and two severed lovers may
be made happy. In dreams, amidst a grotesque confusion of things
remembered and things forgot, we _see_ the events of the past (I have
been at Culloden fight and at the siege of Troy); we are present in