This habitual interpretation of the flat in art as answering to objects in relief, or having depth, can only be understood when it is remembered that our daily experience gives us myriads of instances in which the effect of such flat representations answers to solid receding forms.  That is to say, in the case of all distant objects, in the perception of which the dissimilarity of the retinal pictures and the feeling of convergence take no part, we have to interpret solidity, and relations of nearer and further, by such signs as linear perspective and cast shadow.  On the other hand, it is only in the artificial life of indoors, on our picture-covered walls, that we experience such effects without discovering corresponding realities.  Hence a deeply organized habit of taking these impressions as answering to the solid and not to the flat.  If our experience had been quite different; if, for example, we had been brought up in an empty room, amid painted walls, and had been excluded from the sight of the world of receding objects outside, we might easily have formed an exactly opposite habit of taking the actual mountains, trees, etc., of the distant scene to be pictures laid on a flat surface.

[Illustration:  FIG. 4.]

It follows from this that, with respect to the distant parts of a scene, pictorial art possesses the means of perfect imitation; and here we see that a complete illusory effect is obtainable.  I need but to refer to the well-known devices of linear and aerial perspective, by which this result is secured.[40] The value of these means of producing illusion at the command of the painter, may be illustrated by the following fact, which I borrow from Helmholtz.  If you place two pieces of cardboard which correspond to portions of one form at the sides and in front of a third piece, in the way represented above, so as just to allow the eye to follow the contour of this last, and then look at this arrangement from a point at some little distance with one eye, you easily suppose that it stands in front of the side pieces.  The explanation of the illusion is that this particular arrangement powerfully suggests that the outline of the whole figure, of which the two side pieces are parts, is broken by an intervening object.  Owing to the force of these and other suggestions, it is easy for the spectator, when attending to the background of a landscape painting, to give himself up for a moment to the pleasant delusion that he is looking at an actual receding scene.

In connection with pictorial delusion, I may refer to the well-known fact, that the eye in a portrait seems to follow the spectator, or that a gun, with its muzzle pointing straight outwards, appears to turn as the spectator moves.[41] These tricks of art have puzzled many people, yet their effect is easily understood, and has been very clearly explained by Sir D. Brewster, in the work already referred to (letter v.).  They depend on the fact that a painting, being a flat