244 Mil JOHN AITKEN ON THE 



of the tubs meet underneath it, and form a dark plane in the centre of the 

 descending current, as represented in fig. 2. 



It might be thought that gravitation would not act quickly enough to keep up 

 a supply of dustless air sufficient for this purpose. This however does not seem 

 to be the case, and gravitation appears to be the only cause of the distribution of 

 the dust, causing this dark plane in the descending current. One reason for 

 supposing this is, that if we only cool the tube very slightly, the dark plane is 

 very thick and well marked; but the more we cool the tube the thinner does 

 the dark plane become, intead of thicker, which would be the result if it was 

 due to difference of temperature. The effect of the increased cold is to increase 

 the velocity of the descending current, and draw out and thin down the dark 

 plane. Further, if we closely examine the air round the tube with the strong 

 magnifying glass, we shall see the particles of dust descending and settling on 

 the horizontal part on the top of the tube, while the particles which fall a little 

 to each side of the centre line are carried on by the current, and continue to clasp 

 the tube closely till the current begins to turn under the tube, where the particles 

 being free to fall, drop away from the tube, and leave a clear space (see fig. 2). 

 This clear space only begins to be perceptible when the current begins to turn 

 underneath the tube, and gradually becomes thicker as it travels underneath 

 towards the centre where the two currents join, and form the descending 

 dark plane. 



The rate at which dust settles out of air by gravitation is much quicker 

 than we might imagine. Dust is kept in suspension by ascending currents, and 

 when these are removed it settles remarkably quickly. There was an oppor- 

 tunity for seeing this in these experiments. If the experimental tube was 

 cooled, then the cold gave rise to currents descending on the side of the box 

 where the tube was, and rising on the other side ; but the rising current only 

 came up to the height of the tube, and all the air above the tube was still and 

 currentless, because its temperature increased towards the top of the box, and 

 then was produced a condition of stable equilibrium. Under these circum- 

 stances, I have frequently seen the whole of the upper part of the box above 

 the cold tube become quite clear, and with a sharp line of demarcation between 

 the clear still air above, and the dusty currents underneath. It is, of course, 

 the vertical component of the currents that keeps the dust in suspension, the 

 horizontal component having no such action. This may be seen when we 

 cause a current of dusty air to flow along the under side of a horizontal flat 

 surface. At the point where the current starts, the dust is in contact with the 

 under surface of the body, but falls further and further from it as it flows 

 along. 



In order to study the effect of temperature alone, it was necessary to 

 arrange the experiment so as to get rid of this gravitation effect. For this 

 purpose I prepared another piece of apparatus. The ideal shape of body for 



