§ 10. I believe conditions of form like those in Fig. 82 (turn the book with its outer edge down) may not unfrequently be thus, owing to stratification, when they occur in the nearer sky. This line of cloud is far off at the horizon, drifting towards the left (the points of course forward), and is, I suppose, a series of nearly circular eddies seen in perspective. Which question of perspective we must examine a little before going a step farther. In order to simplify it, let us assume that the under surfaces of clouds are flat, and lie in a horizontal extended field. This is in great measure the fact, and notable perspective phenomena depend on the approximation of clouds to such a condition. [Illustration: 64. Cloud Perspective. (Rectilinear.)] [Illustration: 65. Cloud Perspective. (Curvilinear.)] § 11. Referring the reader to my Elements of Perspective for statements of law which would be in this place tiresome, I can only ask him to take my word for it that the three figures in Plate 64 represent limiting lines of sky perspective, as they would appear over a large space of the sky. Supposing that the breadth included was one-fourth of the horizon, the shaded portions in the central figure represent square fields of cloud,[2] and those in the uppermost figure narrow triangles, with their shortest side next us, but sloping a little away from us. In each figure, the shaded portions show the perspective limits of cloud-masses, which, in reality, are arranged in perfectly straight lines, are all similar, and are equidistant from each other. Their exact relative positions are marked by the lines connecting them, and may be determined by the reader if he knows perspective. If he does not, he may be surprised at first to be told that the stubborn and blunt little triangle, _b_, Fig. 1, Plate 64, represents a cloud precisely similar, and similarly situated, to that represented by the thin triangle, _a_; and, in like manner, the stout diamond, _a_, Fig. 2, represents precisely the same form and size of cloud as the thin strip at _b_. He may perhaps think it still more curious that the retiring perspective which causes stoutness in the triangle, causes leanness in the diamond.[3]