In this state Mr. Clark and I took the matter up last summer, and critically examined all sorts of hypotheses that suggested themselves, Mr. Clark following up the phenomena experimentally with great ingenuity and perseverance. One hypothesis after another suggested itself, seemed hopeful for a time, but ultimately had to be discarded. Some died quickly, others lingered long. In the examination of one electrical hypothesis which suggested itself we came across various curious phenomena which we hope still to follow up.[2] It was some months before what we now believe to be the true explanation began to dawn upon us. Meanwhile we had acquired various new facts, and first and foremost we found that the dark plane rising from a warm body was only the upstreaming portion of a dust-free _coat_ perpetually being renewed on the surface of the body. Let me describe the appearance and mode of seeing it by help of a diagram. (For full description see _Philosophical Magazine_ for March, 1884.) [Footnote 2: For instance, the electric properties of crystals can be readily examined in illuminated dusty air; the dust grows on them in little bushes and marks out their poles and neutral regions, without any need for an electrometer. Magnesia smoke answers capitally.] Surrounding all bodies warmer than the air is a thin region free from dust, which shows itself as a dark space when examined by looking along a cylinder illuminated transversely, and with a dark background. At high temperatures the coat is thick; at very low temperatures it is absent, and dust then rapidly collects on the rod. On a warm surface only the heavy particles are able to settle--there is evidently some action tending to drive small bodies away. An excess of temperature of a degree or two is sufficient to establish this dust-free coat, and it is easy to see the dust-free plane rising from it. The appearances may also be examined by looking along a cylinder _toward_ the source of light, when the dust-free spaces will appear brighter than the rest. A rod of electric light carbon warmed and fixed horizontally across a bell-jar full of dense smoke is very suitable for this experiment, and by means of a lens the dust-free regions may be thus projected on to a screen. Diminished pressure makes the coat thicker. Increased pressure makes it thinner. In hydrogen it is thicker, and in carbonic acid thinner, than in air. We have also succeeded in observing it in liquids--for instance, in water holding fine rouge in suspension, the solid body being a metal steam tube. Quantitative determinations are now in progress. [Illustration: Fig. 1 and Fig. 2] Fig. 1 shows the appearance when looking along a copper or carbon rod laterally illuminated; the paths of the dust particles are roughly indicated. Fig. 2 shows the coat on a semi-cylinder of sheet copper with the concave side turned toward the light.