When the sun shines on bare rock its surface is heated and expanded, and the expanded particles crowd one another with great force. Since rock is a poor conductor of heat its surface is heated and expanded notably more than parts beneath the surface. It follows that strains are set up between the expanded outer portion and the cooler and less expanded parts within. In the cooling of the same rock mass it is the outermost portion which cools first and fastest, and, contracting as it cools, strains are again set up between the outer part, which is cooled more, and the inner part, which is cooled less. The result may be illustrated by the effect of cold water on hot glass, or of hot water on cold glass. In either case the fracture is the result of the sudden and considerable differential expansion or contraction. Since the heating and cooling of rock are much slower than the heating and cooling of glass under the conditions mentioned, the rupturing effects are less conspicuous, but none the less real. The actual effects of temperature changes are illustrated by familiar phenomena. The surface portions of bowlders exposed to the sun are frequently seen to be shelling off (Fig. 26). The loosened concentric shells may be a fraction of an inch, or sometimes even several inches in thickness. This process of _exfoliation_ affects not only bowlders, but bare rock surfaces wherever exposed to the sun (Figs. 27, 28). It is often conspicuous on the faces of cliffs. [Illustration: +Fig.+ 26.—Exfoliation. A bowlder of weathering, the rock being granite. Wichita Mountains, Oklahoma.] [Illustration: +Fig.+ 27.—A weathered summit of granite in the Wichita Mountains. Oklahoma. (Willis, U. S. Geol. Surv.)]