the colour of the under sheet the grey scrap took the
complementary hue. If, however, the piece of grey paper is put outside
the thin sheet, it looks grey; and what is most remarkable is that when
a second piece is put outside, the scrap inside no longer wears the
complementary hue.

There is here evidently something more than a change of organic
conditions; there is an action of experience and suggestion. The reason
of our seeing the scrap rose-red in one case and neutral grey in
another, is that in the first instance we vividly represent to ourselves
that we are looking at it through a greenish veil (which is, of course,
a part of the illusion); for rose-red seen through a greenish medium
would, as a matter of fact, be light grey, as this scrap is. Even if we
allow that there always exists after an impression of colour a temporary
organic disposition to see the complementary hue, this does not suffice
as an explanation of these cases; we have to conclude further that
imagination, led by the usual run of our experience, is here a
co-operant factor, and helps to determine whether the complementary tint
shall be seen or not.

_Misinterpretation of Form._

More complex and circumscribed associations take part in those errors
which we occasionally commit respecting the particular form of objects.
This has already been touched on in dealing with artistic illusion. The
disposition of the eye to attribute solidity to a flat drawing is the
more powerful in proportion to the familiarity of the form. Thus, an
outline drawing of a building is apt to stand out with special force.

Another curious illustration of this is the phenomenon known as the
conversion of the concave mould or matrix of a medal into the
corresponding convex relief. If, says Helmholtz, the mould of a medal be
illuminated by a light falling obliquely so as to produce strong
shadows, and if we regard this with one eye, we easily fall into the
illusion that it is the original raised design, illuminated from _the
opposite side_. As a matter of fact, the visual impression produced by a
concave form with the light falling on one side, very closely resembles
that produced by a corresponding convex form with the light falling on
the other side. At the same time, it is found that the opposite mode of
conversion, that is to say, the transformation of the raised into the
depressed form, though occurring occasionally, is much less frequent.
Now, it may be asked, why