ice, from time to time, of the sketch as a whole. I can trace no
likeness between what I draw and the images that present themselves
to me in dreams, and I find that a very trifling accident, such as a
chance dot on the paper, may have great influence on the general
character of any one of these automatic sketches.

Visions, like dreams, are often mere patchworks built up of bits of
recollections. The following is one of these:--

"When passing a shop in Tottenham Court Road, I went in to order a
Dutch cheese, and the proprietor (a bullet-headed man whom I had
never seen before) rolled a cheese on the marble slab of his counter,
asking me if that one would do. I answered 'Yes,' left the shop, and
thought no more of the incident. The following evening, on closing
my eyes, I saw a head detached from the body rolling about slightly
on a white surface. I recognised the face, but could not remember
where I had seen it, and it was only after thinking about it for
some time that I identified it as that of the cheesemonger who had
sold me the cheese on the previous day. I may mention that I have
often seen the man since, and that I found the vision I saw was
exactly like him, although if I had been asked to describe the man
before I saw the vision I should have been unable to do so."

Recollections need not be combined like mosaic work; they may be
blended, on the principle of composite portraiture. I suspect that
the phantasmagoria may be in some part due to blended memories; the
number of possible combinations would be practically endless, and
each combination would give a new face. There would thus be no limit
to the dies in the coinage of the brain.

I have found that the peculiarities of visualisation, such as the
tendency to see Number-Forms, and the still rarer tendency to
associate colour with sound, is strongly hereditary, and I should
infer, what facts seem to confirm, that the tendency to be a seer of
visions is equally so. Under these circumstances we should expect
that it would be unequally developed in different races, and that a
large natural gift of the visionary faculty might become
characteristic not only of certain families, as among the
second-sight seers of Scotland, but of certain races, as that of the
Gipsies.

It happens that the mere acts of fasting, of want of sleep, and of
solitary musing, are severally conducive to visions. I have myself
been told of cases in which persons accidentally long dep