FORMATION OF SMALL CLEAR SPACES IN DUSTY AIR. 263 



while the hot tube is quite clean. If we send the smoke of a cigar through the 

 filter, nothing but perfectly transparent gases come out at the other end. The 

 effect of coating the cold surface with glycerine has been tried, as it seemed pos- 

 sible that the dust deposited on the clean surface might be carried on by the air 

 current. The dust, however, seems to be firmly held on a cold clean surface, and 

 no decided improvement was got by the addition of the glycerine. No accurate 

 experiments have been made to determine the best size of the filtering channel. 

 The filters with very narrow passages and those with much wider ones all work 

 well, but no quantitative experiments have been made as to their relative values. 

 It is not easy to determine what influence difference of temperature has on 

 the action of a cotton- wool filter. Heating the cotton- wool has little effect in 

 reducing its filtering powers. "We might expect this, as the cotton and the air 

 passing through it rapidly acquire the same temperature ; and it is extremely 

 difficult to say how much of the action of this filter depends on the slight 

 differences of temperature produced by the air in passing through the cotton. 



Diffusion Effects. 



I shall now describe two experiments on diffusion, which were made in the 

 hope they would throw some light on this repelling action of hot bodies. For this 

 purpose a tube similar to those used in the previous experiments was taken, 

 and an opening made in the side of it, at the front end. Into this opening was 

 fitted a thin plug of plaster of paris. The surface of the plug was made flat, 

 and when put in the dust-box was placed vertically, as in the experiments on 

 the heat effect, to get rid of the distribution due to gravitation. This diffusion 

 diaphragm was blackened, to enable the effect to be better observed, as a white 

 surface reflects so much light, it makes it difficult to see what is taking place. 



After the diffusion apparatus was fitted in its place, the dust-box was filled 

 with sulphate dust, and left till everything had acquired the same temperature. 

 Carbonic acid gas was then introduced into the tube. At once a downward 

 current was produced in front of the diaphragm, the dust particles kept 

 close up to its surface, and if there was any tendency to the formation of a 

 clear space the carbonic acid at once closed it. The apparatus for supplying 

 the carbonic acid gas was then removed, and a small pipe connected with the 

 gas pipes was then led into the diffusion tube, so as to get the effect due to the 

 diffusion of gases lighter than air. The effect in this case was the opposite of 

 that given by the carbonic acid. An upward current at once started, and a 

 thin clear space formed in front of the diffusion diaphragm. These experiments 

 prove that the dust particles move in the direction in which the greatest rate of 

 diffusion takes place. This at first sight looks very self-evident; but we must 

 remember that in front of the diffusion diaphragm, when hydrogen is coming 



VOL. XXXII. PART II. 2 X 





