§ XX. And observe, also, knowledge is not only very often unnecessary, but it is often _untrustworthy_. It is inaccurate, and betrays us where the eye would have been true to us. Let us take the single instance of the knowledge of aerial perspective, of which the moderns are so proud, and see how it betrays us in various ways. First by the conceit of it, which often prevents our enjoying work in which higher and better things were thought of than effects of mist. The other day I showed a line impression of Albert Durer's "St. Hubert" to a modern engraver, who had never seen it nor any other of Albert Durer's works. He looked at it for a minute contemptuously, then turned away: "Ah, I see that man did not know much about aerial perspective!" All the glorious work and thought of the mighty master, all the redundant landscape, the living vegetation, the magnificent truth of line, were dead letters to him, because he happened to have been taught one particular piece of knowledge which Durer despised. § XXI. But not only in the conceit of it, but in the inaccuracy of it, this science betrays us. Aerial perspective, as given by the modern artist, is, in nine cases out of ten, a gross and ridiculous exaggeration, as is demonstrable in a moment. The effect of air in altering the hue and depth of color is of course great in the exact proportion of the volume of air between the observer and the object. It is not violent within the first few yards, and then diminished gradually, but it is equal for each foot of interposing air. Now in a clear day, and clear climate, such as that generally presupposed in a work of fine color, objects are completely visible at a distance of ten miles; visible in light and shade, with gradations between the two. Take, then, the faintest possible hue of shadow, or of any color, and the most violent and positive possible, and set them side by side. The interval between them is greater than the real difference (for objects may often be seen clearly much farther than ten miles, I have seen Mont Blanc at 120) caused by the ten miles of intervening air between any given hue of the nearest, and most distant, objects; but let us assume it, in courtesy to the masters of aerial perspective, to be the real difference. Then roughly estimating a mile at less than it really is, also in courtesy to them, or at 5000 feet, we have this difference between tints produced by 50,000 feet of air. Then, ten feet of air will produce the 5000th part of this difference. Let the reader take the two extreme tints, and carefully gradate the one into the other. Let him divide this gradated shadow or color into 5000 successive parts; and the difference in depth between one of these parts and the next is the exact amount of aerial perspective between one object, and another, ten feet behind it, on a clear day.