"But is not that just what they do?" No. They float at different heights, and with definite forms, in the body of the air itself. If they rose like foam, the sky on a cloudy day would look like a very large flat glass of champagne seen from below, with a stream of bubbles (or clouds) going up as fast as they could to a flat foam-ceiling. "But may they not be just so nicely mixed out of something and nothing, as to float where they are wanted?" Yes: that is just what they not only may, but must be: only this way of mixing something and nothing is the very thing I want to explain or have explained, and cannot do it, nor get it done. § 5. Except thus far. It is conceivable that minute hollow spherical globules might be formed of water, in which the enclosed vacuity just balanced the weight of the enclosing water, and that the arched sphere formed by the watery film was strong enough to prevent the pressure of the atmosphere from breaking it in. Such a globule would float like a balloon at the height in the atmosphere where the equipoise between the vacuum it enclosed, and its own excess of weight above that of the air, was exact. It would, probably, approach its companion globules by reciprocal attraction, and form aggregations which might be visible. This is, I believe, the view usually taken by meteorologists. I state it as a possibility, to be taken into account in examining the question--a possibility confirmed by the scriptural words which I have taken for the title of this chapter. § 6. Nevertheless, I state it as a possibility only, not seeing how any known operation of physical law could explain the formation of such molecules. This, however, is not the only difficulty. Whatever shape the water is thrown into, it seems at first improbable that it should lose its property of wetness. Minute division of rain, as in "Scotch mist," makes it capable of floating farther,[1] or floating up and down a little, just as dust will float, though pebbles will not; or gold-leaf, though a sovereign will not; but minutely divided rain wets as much as any other kind, whereas a cloud, partially always, sometimes entirely, loses its power of moistening. Some low clouds look, when you are in them, as if they were made of specks of dust, like short hairs; and these clouds are entirely dry. And also many clouds will wet some substances, but not others. So that we must grant farther, if we are to be happy in our theory, that the spherical molecules are held together by an attraction which prevents their adhering to any foreign body, or perhaps ceases only under some peculiar electric conditions. § 7. The question remains, even supposing their production accounted for,--What intermediate states of water may exist between these spherical hollow molecules and pure vapor?