FORMATION OF SMALL CLEAR SPACES IN DUSTY AIR. 261 



These experiments naturally suggested the possibility of constructing 

 an air filter on thermic principles. They showed that the visible particles 

 of dust could be thrown out of the air, as the particles tended to move from 

 the hot parts, and to attach themselves to cold surfaces. But the question 

 which naturally suggested itself was, Are the very small invisible particles also 

 arrested ? If the thermic filter turned out to be a success, it appeared to 

 me it would also be the best way to get an answer to this question. In order 

 to filter air on thermic principles, all that appeared necessary was to pass the 

 air through a space or channel, the two sides of which were kept at different 

 temperatures. In this way I hoped the dust would be driven from the hot 

 side and attach itself to the cold one. Practically to carry out this idea, the 

 simplest method that suggested itself was to pass the air through the space 

 between two concentric tubes, the one tube being kept hot, and the other cold. 

 In the preliminary instruments which have been made, the distance between 

 the tubes forming the space through which the air passes, is in one instrument 

 less than 1 mm., but in other instruments this space is nearly as much as 3 mm. 

 The length of the passage in the different instruments is about 35 cm. One of 

 these instruments has the outer tube jacketed by means of a larger pipe for the 

 purpose of heating it with steam. The other instruments were heated simply 

 by means of a gas flame. The filter is shown in section, fig. 14. A is a tube 

 about 13 mm. diameter. B is another tube slightly larger, and allowing a space 

 C, between the two for the passage of the air to be filtered, which enters and 

 leaves by the tubes D, D. The outer tube E forms a steam jacket round B. 

 F, F are pipes for steam entering, and for condensed water leaving the jacket. 

 The pipe A is kept cold by means of a stream of water. In working the instru- 

 ment it is not, however, necessary to keep to this arrangement ; steam may be 

 admitted to the centre tube A, and cold water to the outside jacket ; both 

 arrangements do equally well. For the purpose of cleaning and examining the 

 surfaces of the air channel, the centre tube was not permanently fixed in its 

 place, but was so arranged that it could be easily taken out, and the joints 

 were made tight by means of the short pieces of india-rubber tube H, H. The air, 

 after passing through the space C, was conveyed by means of a tube to a glass 

 flask, in which there was a little water. The flask in turn was connected by 

 means of another tube to an air-pump, in order to test the condition of the 

 air after passing through the instrument. If cloudy condensation is produced 

 when the pressure is reduced in the flask, we know that the air is not filtered; 

 and, on the other hand, if the air remains perfectly clear on exhausting, we 

 know that no dust, not even the invisible particles, have passed into it. 



The apparatus was fitted up for trial, all the connections being made and 

 tested. Using the instrument heated with flame, the first effect of the heat, 

 as expected, was a great increase in the fogging. The temperature was raised 



