=Vision of Solidity.=--This description of binocular vision follows what is called the theory of identical points. On the whole it formulates the facts correctly. The only odd thing is that we should be so little troubled by the innumerable double images which objects nearer and farther than the point looked at must be constantly producing. The answer to this is that _we have trained ourselves to habits of inattention_ in regard to double images. So far as things interest us we turn our foveæ upon them, and they are necessarily seen single; so that if an object impresses disparate points, that may be taken as proof that it is so unimportant for us that we needn't notice whether it appears in one place or in two. By long practice one may acquire great expertness in detecting double images, though, as some one says, it is an art which is not to be learned completely either in one year or in two. [Illustration: FIG. 11.] Where the disparity of the images is but slight it is almost impossible to see them as if double. They give rather the perception of a solid object being there. To fix our ideas, take Fig. 11. Suppose we look at the dots in the middle of the lines _a_ and _b_ just as we looked at the spots in Fig. 8. We shall get the same result--i.e., they will coalesce in the median line. But the entire lines will not coalesce, for, owing to their inclination, their tops fall on the temporal, and their bottoms on the nasal, retinal halves. What we see will be two lines crossed in the middle, thus (Fig. 12): [Illustration: FIG. 12.] The moment we attend to the tops of these lines, however, our foveæ tend to abandon the dots and to move upwards, and in doing so, to converge somewhat, following the lines, which then appear coalescing at the top as in Fig. 13. [Illustration: FIG. 13.] [Illustration: FIG. 14.] If we think of the bottom, the eyes descend and diverge, and what we see is Fig. 14. Running our eyes up and down the lines makes them converge and diverge just as they would were they running up and down some single line whose top was nearer to us than its bottom. Now, if the inclination of the lines be moderate, we may not see them double at all, but single throughout their length, when we look at the dots. Under these conditions their top does look nearer than their bottom--in other words, we see them stereoscopically; and we see them so even when our eyes are rigorously motionless. In other words, the slight disparity in the bottom-ends which _would_ draw the foveæ divergently apart makes us see those ends farther, the slight disparity in the top ends which _would_ draw them convergently together makes us see these ends nearer, than the point at which we look. The disparities, in short, affect our perception as the actual movements would.[12]