522 Professor Lodge [May 28, 



In trying this experiment one often finds a copy of the appearance 

 imprinted upon the glass wall of the box in smoke, and this dust- 

 graph also serves to illustrate the phenomenon. 



[Projection on screen of dust-graph formed on a vertical glass 

 plate, where the end of a hot rod is held against it in still dusty 

 air.] 



It may be convenient here to state, at least roughly, the line of 

 explanation to which we have ultimately been impelled. We regard 

 it as quite certain that the facts serve as another illustration of the 

 kinetic theory of gases, of the same kind as Mr. Crookes's radiometer. 

 The dust particles are kept out of contact with the warm body by 

 means of a differential molecular bombardment on their surfaces. 

 True, the thickness of the coat is far larger than the mean free path 

 of a molecule at the corresponding pressure ; and accordingly no body 

 as big as a radiometer vane could at that distance be acted upon ; but 

 Prof. Osborne Eeynolds has shown that very minute bodies are sus- 

 ceptible to a bombardment at much greater distances ; and it is, I 

 think, in accord with his theory that dust particles should be repelled 

 from hot bodies over a measurable distance even at ordinary atmo- 

 spheric pressure, although radiometer vanes would only be acted upon 

 at a pressure of a thousandth, or something approaching a millionth, 

 of an atmosphere. I do not say that this explanation is yet complete, 

 and I am not aware that any one has adopted it. It is, however, 

 pretty clear that there will be an outward pressure from the surface 

 of a hot body,* so long as the air near it is getting warmer : and if 

 convection currents are allowed, the air near a hot body is continually 

 in the state of getting warmer. 



The main facts which we assured ourselves of may be thus briefly 

 summarised. 



1. The dark plane becomes visible over a cylindrical thermometer 

 bulb indicating a temperature only half a degree warmer than the air. 



2. The dark coat becomes just visible round a body one degree 

 hotter than the air ; at two degrees it is distinct ; and at five degrees it 

 is fairly thick. 



3. The coat enlarges with diminished pressure, and narrows when 

 the pressure is increased. 



4. In hydrogen the coat is thicker, and in carbonic acid thinner, 

 than in air, 



5. If volatile smoke like volatilised camphor is used, the coats 

 become thicker, though less sharply defined. 



6. Eound a rod of camphor in ordinary smoke the dust-free 

 coat is extra thick, owing to the extra bombardment of evaporation. 



7. liound cold bodies the coat is practically absent, esjiecially on 

 the top of the body ; and if the cold is too great no descending dark 

 plane is formed : only a bright one. 



* See for instance some experiments of Mr. Crookcs ou rcpulaion by hot 

 surfaces, described in ' Nature,' vol. xv. page 301. 



