Another group of conditions depends on variation not in the size but in the nature of the retinal impression produced by the same line or curve looked at from different points of view. If we look straight at a rectangular cross, with the lines of vision parallel or equally converging, the impression on the retina also has the form of a rectangular cross. The same is the case if we look straight up at it or straight down at it, or directly to the right, or directly to the left; but if we turn our eyes obliquely upwards and rightwards, upwards and leftwards, downwards and rightwards, downwards and leftwards, the legs of the retinal cross no longer make a right angle with each other, but are distorted so as to slant in varying degrees and in varying ways according to the point of view. The corresponding modifications of retinal sensations are not normally attended to, because they answer to no real differences in the shape of the cross itself. Thus a slanting retinal cross, when the slant is merely produced by the point of view, gives rise to the perception of a rectangular cross lying in a certain direction. Conversely, a rectangular retinal cross, produced by an object in the same direction and position, gives rise to the perception of a slanting cross, because a slanting cross under such conditions would actually produce a square retinal impression. Afterimages admirably illustrate this point. If we have obtained an afterimage of a rectangular cross by looking straight at it, and if we then look straight at the wall of the room in front of us, the afterimage is outlined on the wall as a rectangular cross. If we turn our eyes to the upper lefthand part of the wall, we see a slanting cross, as represented in Fig. 3; if we then turn our eyes to the upper righthand corner, we see a slanting cross, as in Fig. 4.

Fig. 3.                                       Fig. 4.

In all cases, the retinal impression and the visual sensation remain unchanged. The above is only a specimen of what is continually taking place. The nature of the retinal impressions produced by straight or curved lines and their combinations, is constantly varying in a regular and systematic manner with the position of the eyes relatively to the object looked at. The variation depends in a very large measure on the shape of surfaces and on their relative position in the third dimension. Thus the afterimage of the square cross f it is seen on a perpendicular wall directly in front of the eye, produces a square cross on the retina. But if the plane on which it is projected is tilted away from or towards the spectator, the retinal cross is distorted. Even an "inclined wall, in a picture, will, if an afterimage be thrown upon it, distort the shape thereof, and make us see a form of which our afterimage would be the natural projection on the retina, were that form laid upon the wall. Thus a signboard is painted in perspective on a screen, and the eye, after steadily looking at a rectangular cross, is turned to the painted signboard. The afterimage appears as an obliquelegged cross upon the signboard."* In looking at any solid figure from a given point of view, the lines and contours presented by its bounding surfaces produce varying retinal images according to the shape of the surfaces and their position relatively to the eye. These retinal differences correspond to no real differences in the shape of the lines and contours themselves. Our tendency is therefore as far as possible to ignore them, except in so far as they mark position in the third dimension. In so far as the variation in optical sensation as such is disregarded, it fulfils the function of determining our perception of depth and solid figure. The artist avails himself of these perspective distortions in producing stereoscopic effect.

* Op. cit., p. 254.

The play of light and shade also contributes in a very large measure to determine our perception of depth. The mode in which light is intercepted varies with the shape of the solid object on which it falls. The distribution of shade among the parts of the object itself is also determined by its shape. Thus the play of light and shade is exactly opposite in the case of a hollow mask and a projecting face. This "modelling," as it is called by the artist, takes the most subtle gradations, according to the various minute hollows and elevations in the surface of an object, as for instance in the folds of drapery. Besides this what is called the "castshadow," viz. the shadow thrown by an object as a whole, plays a very important part. "Objects in a landscape stand out much better in morning and evening light when strong and distinct castshadows are thrown, than in noonday light."*

Other factors have a peculiar importance in the case of very remote objects. These are covered by the term aerial perspective, and include indistinctness of outline and modification of colouring. If two mountains are seen in the distance, and one appears bluish, and the other green, the green is perceived as nearer. The green of the vegetation is only visible at a certain distance; at a greater distance it gives place to a blue tint derived from the intervening air. These associative conditions do not themselves enter into the constitution of the perception of depth, but are able to reproduce it when it has once been formed by other means. It is* frequently said that they are signs which the mind interprets. Such phrases are only permissible if we are careful to explain the nature of the signs and of their interpretations. Usually when we speak of interpreting a sign, it is implied that the sign is itself distinctly and separately noticed, and that the interpretation is an additional distinct act of thought. But this is not the case with the perceptual signs which we have been discussing. They themselves are in the main ignored, and only their meaning comes before consciousness. They have no independent existence for consciousness apart from their meaning. The meaning being inseparably one with the sensations that are its signs, has the immediacy, the obtrusiveness, the fixity, and the detailed definiteness of impressional experience. In other words, it is a percept and not an idea. The connexion between sign and meaning is one of complication and not of free revival.

*Sully, Human Mind, vol. i., p. 252.