Several years ago Professor Dewar gave me a specimen of silica in a state of exceedingly minute division, which had been produced in Dr Playfair's laboratory in the preparation of fluosilicic acid. I noticed at the time how much its great mobility is increased by heating—so that it behaves almost like a liquid. And I fancied that I observed close to the surface a thin stratum of what might by the same analogy be called a vapour; consisting of particles thrown up and falling back again, like the little drops thrown up at the surface of soda-water. I was inclined to ascribe these phenomena to heat directly—supposing that the particles were fine enough to behave, though in a very imperfect way, as the kinetic theory assumes the particles of a gas to behave.