[10] I have just said, Sec. 12, that if, _quitting hold_ of this original impression, the artist tries to compose something prettier than he saw, it is all over with him; but, retaining the first impression, he will, nevertheless, if he has invention, instinctively modify many lines and parts of it--possibly all parts of it--for the better; sometimes making them individually more pictorial, sometimes preventing them from interfering with each other's beauty. For almost all natural landscapes are redundant treasures of more or less confused beauty, out of which the human instinct of invention can by just choice arrange, not a better treasure, but one more fitted to human sight and emotion, infinitely narrower, infinitely less lovely in detail, but having this great virtue, that there shall be absolutely nothing which does not contribute to the effect of the whole; whereas in the natural landscape there is a redundancy which impresses only as redundance, and often an occurrence of marring features; not of ugliness only, but of ugliness _in the wrong place_. Ugliness has its proper virtue and use; but ugliness occurring at the wrong time (as if the negro servant, instead of standing behind the king, in Tintoret's picture, were to thrust his head in front of the noble features of his master) is justly to be disliked and withdrawn. "Why, this," exclaims the idealist, "is what _I_ have always been saying, and _you_ have always been denying." No; I never denied this. But I denied that painters in general, when they spoke of improving Nature, knew what Nature was. Observe: before they dare as much as to _dream_ of arranging her, they must be able to paint her as she is; nor will the most skilful arrangement ever atone for the slightest wilful failure in truth of representation; and I am continually declaiming against arrangement, not because arrangement is wrong, but because our present painters have for the most part nothing to arrange. They cannot so much as paint a weed or a post accurately; and yet they pretend to improve the forests and mountains. [11] For instance, even in my topographical etching, Plate 20, I have given only a few lines of the thousands which existed in the scene. Those lines are what I considered the leading ones. Another person might have thought other lines the leading ones, and his representation might be equally true as far as it went; but which of our representations went furthest would depend on our relative degrees of knowledge and feeling about hills. [12] Or the best photograph. The question of the exact relation of value between photography and good topographical drawing, I hope to examine in another place. CHAPTER III. OF TURNERIAN LIGHT.