FORMATION OF SMALL CLEAR SPACES IN DUSTY AIR. 267 



the repulsion of the discs by a hot surface placed inside the radiometer bulb, as 

 in the apparatus described by Mr Crookes in Nature, vol. xv. p. 301. In this 

 radiometer the vanes were made of very clear mica, and they did not rotate 

 when light fell on them. Inside the bulb, and just clear of the discs, was fixed 

 in a vertical plane a blackened plate of mica. When the light was allowed to 

 fall on this fixed and black plate, the vanes instantly rotated as if a wind were 

 issuing from this surface. The energy which causes the repulsion of the dust 

 and of the discs of this radiometer is transferred in both cases from the 

 hot surface to the repelled surface by the kinetic energy of the gas molecules, 

 and not by radiation. 



Another consideration which indicates that the force causing the movement 

 of the dust is not transferred by radiation is the well-known fact that radiant 

 heat is not much intercepted by dust. When we concentrate a strong beam of 

 light and heat in dusty air by means of a lens, perhaps one of the things which 

 strikes us most is the very slight heating effect which is developed at the focus. 

 The dust is not destroyed, and no rapid upward current is formed. But if we 

 place a piece of paper at the same focus it is at once charred, and a rapid 

 current of air rises from its heated surface. 



The rate at which vapour molecules diffuse under the conditions existing 

 in the experiments is very great, and seems to be quite sufficient to account for 

 the results. Take, for instance, the water molecules when they pass into 

 vapour. Vapour molecules are selected because we can follow their move- 

 ments. In a small fraction of a second they diffuse to a distance of nearly 

 1 cm. This can be seen in the experiment described with the flat test-surface 

 when moistened. With a slight rise of temperature, fog particles are seen 

 forming in the current, rising in front of the wet surface. Even at the lower 

 edge of the plate these particles are seen at some distance from the plate, and 

 separated from it by a dark space, showing that even at that point the 

 vapour molecules have already diffused outwards to a distance and far beyond 

 the dark space, while probably other molecules have gone further than the fog 

 boundary, but are under conditions which keep them in the state of vapour. 



Or take the reverse of this diffusion process, seen in the evaporation of fog 

 particles. Let us blow some steam into the dust-box, so as to form a regular 

 fog, but without adding any dust. Into this fog introduce a piece of very 

 dry wood ; if it is charred so much the better, as its blackness enables us to see 

 more easily what is taking place. It will be observed that there is formed all 

 round the wood a clear space, in which not a particle of fog can be seen. If 

 we watch the air currents we shall see the particles approaching, but vanishing 

 at some distance from the wood, and over the wood the particles will be seen 

 falling into the clear space and disappearing. This clear space is caused by 

 the wood absorbing the vapour in the air near it, thus surrounding itself with 



