FORMATION OF SMALL CLEAR SPACES IN DUSTY AIR. 245 



this purpose would be one having some length and breadth, but infinitely thin 

 and flat, so that when placed vertically, the air in passing over it would never 

 have to move in a horizontal direction. The nearest approach I could make to 

 this was made with a piece of copper foil folded on itself, soldered all round 

 its edges, and fixed to the end of a brass tube. It was heated and cooled by 

 passing into it hot or cold water. This instrument presented at the front edge 

 extremely little thickness, and was found to answer well, but was rather delicate 

 and easily put out of shape. As it is only necessary to examine one side of the 

 test plane or surface, a different form of apparatus was afterwards adopted. It 

 was made of a piece of brass tube the same as used in the previous experi- 

 ments, and a flat plate of copper was soldered to one side of it at the front 

 end. This plate was filed perfectly flat and smooth, and sharpened at the top 

 and bottom edges, all the bevel being on the tube side of the plate. The side 

 of the plate presented towards the source of illumination was thus a perfectly 

 flat surface, and when placed vertically, the air passing over the front surface 

 could not have its dust separated from it by gravitation, as all the horizontal 

 movement went to the back of the plate. 



Placing either of these test surfaces in the dust-box with the plate vertical, 

 cold was applied. At once a downward current was produced, but no dark 

 space was formed on the vertical test surface ; and if the copper foil apparatus, 

 which is flat on both sides, is used, no dark plane whatever is formed, as 

 shown in fig. 3. More intense cold was tried, and a temperature of — 10° C. in 

 air of a temperature of 15° C. was found to produce no effect save an increased 

 rate of current, and an increased brightness in the particles near the plate, 

 due to water vapour being deposited on them by the lowering of the tem- 

 perature, an effect observed by Lord Rayleigh on the dust bounding his cold 

 dark plane. Different dusts were tried, and the experiment varied in many ways, 

 but when the gravitation effect was removed, not the slightest tendency to the 

 formation of a dark plane by cold could be detected. The tendency seemed 

 to be the other way. The dust particles in all cases tended to keep close to 

 the cold body. This indicates that Lord Rayleigh's dark plane formed in the 

 descending current from a cold body is not an effect of the cold, but is due to 

 the separating action of gravitation. 



What I am about to state may at first seem a contradiction of this conclusion. 

 When varying the conditions of the experiment, and altering the amount of water 

 vapour present, I was much surprised to find that under certain conditions the 

 dark plane had a decided tendency to make its appearance in the descending 

 current even from a thin vertical surface. On repeating the experiment and 

 varying it, it was found that the conditions best suited for getting this dark plane 

 were when there was nothing but ordinary atmospheric dust in the box, and 

 the air was saturated with water vapour. Under these conditions, there was 

 generally all through the box a haziness, but in the space in front of the cold test 



