a great deal in
common with these so-called abstract ideas. The composite portraits
consist, as was explained, of numerous superimposed pictures,
forming a cumulative result in which the features that are common to
all the likenesses are clearly seen; those that are common to a few
are relatively faint and are more or less overlooked, while those
that are peculiar to single individuals leave no sensible trace at
all.

This analogy, which I pointed out in a Memoir on Generic Images,
[11] has been extended and confirmed by subsequent experience of the
process. One objection to my view was that our so-called
generalisations are commonly no more than representative cases, our
recollections being apt to be unduly influenced by particular events,
and not by the totality of what we have seen; that the reason why
some one recollection has prevailed is that the case was sharply
defined, or had something unusual about it, or that our frame of
mind was at the time of observation susceptible to that particular
kind of impression. I have had exactly the same difficulties with
the composites. If one of the individual portraits has sharp outlines,
or if it is unlike the rest, or if the illumination is temporarily
strong, it will assert itself unduly in the result. The cases seem
to me exactly analogous. I get over my photographic difficulty very
easily by throwing the sharp portrait a little out of focus, by
eliminating such portraits as have exceptional features, and by
toning down the illumination to a standard intensity.

[Footnote 11: "Generic Images," _Proc. Royal Institute_, Friday,
April 25, 1879, partly reprinted in the Appendix.]

PSYCHOMETRIC EXPERIMENTS.

When we attempt to trace the first steps in each operation of our
minds, we are usually baulked by the difficulty of keeping watch,
without embarrassing the freedom of its action. The difficulty is
much more than the common and well-known one of attending to two
things at once. It is especially due to the fact that the elementary
operations of the mind are exceedingly faint and evanescent, and
that it requires the utmost painstaking to watch them properly. It
would seem impossible to give the required attention to the
processes of thought, and yet to think as freely as if the mind had
been in no way preoccupied. The peculiarity of the experiments I am
about to describe is that I have succeeded in evading this difficulty.
My method consists in allowing the mind