The second portion of the heat sent us by the moon is that which she emits on her own account as a warm body--warmed, of course, mainly, if not entirely, by the action of the sun. The amount of _this_ heat will depend upon the temperature of the moon's surface and its radiating power; and the temperature will depend upon a number of things (chiefly heat-absorbing power of the surface, and the nature and density of the lunar atmosphere, as well as the supply of heat received from the sun), being determined by a balance between give and take. So long as more heat is received in a second than is thrown off in the same time, the temperature will rise, and _vice versa_. It is to be noted, further, that this second component of the moon's thermal radiance must be mainly what is called "obscure" or dark heat, like that from a stove or teakettle, and characterized by the same want of penetrative power. No one knows why at present; but it is a fact that the heat-radiations from bodies at a low temperature--radiations of which the vibrations are relatively slow, and the wave-length great--have no such power of penetrating transparent media as the higher-pitched vibrations which come from incandescent bodies. A great part, therefore, of this contingent of the lunar heat is probably stopped in the upper air, and never reaches the surface of the earth at all. Now, the thermopile cannot, of course, discriminate directly between the two portions of the lunar heat; but to some extent it does enable us to do so indirectly, since they vary in quite a different way with the moon's age. The simple _reflected_ heat must follow the same law as moonlight, and come to its maximum at full moon. The _radiated_ heat, on the other hand, will reach its maximum when the average temperature of that part of the moon's surface turned toward the earth is highest; and this must be some time after full moon, for the same sort of reasons that make the hottest part of a summer's day come two or three hours after noon. The conclusion early reached by Lord Rosse was that nearly all the lunar heat belonged to the second category--dark heat _radiated_ from the moon's warmed surface, the _reflected_ portion being comparatively small--and he estimated that the temperature of the hottest parts of the moon's surface must run as high as 500 deg. F.; well up toward the boiling-point of mercury. Since the lunar day is a whole month long, and there are never any clouds in the lunar sky, it is easy to imagine that along toward two or three o'clock in the lunar afternoon (if I may use the expression), the weather gets pretty hot; for when the sun stands in the lunar sky as it does at Boston at two P.M., it has been shining continuously for more than two hundred hours. On the other hand, the coldest parts of the moon's surface, when the sun has only just risen after a night of three hundred and forty hours, must have a temperature more than a hundred degrees below zero.