Nuclei in Air hy Intense Cooling. 565 



presence of nuclei after the cooling process. H was cou- 

 nected to FG through the long capillary tube M, the object 

 of which was to prevent diffusion of moisture from H into F 

 during the operation o£ filling the latter with dry air. 

 Before the P 2 5 was placed in the bulb E the whole appa- 

 ratus to the left of G was thoroughly dried by allowing a 

 current of dry air (obtained from the liquid air stored in A) 

 to stream through for a considerable time, the tubes and 

 bulbs meanwhile being strongly heated with a flame. Finally 

 P 2 5 was introduced through K which was sealed off:, and 

 the taps D and G closed. The efficiency of this method of 

 drying the apparatus was shown by the fact that the P 2 5 

 preserved indefinitely its dry and powdery condition. Hence 

 no appreciable amount of moisture could have remained in 

 the apparatus, even before the P 2 5 was introduced. 



After the air in E, F had been in contact with the P 2 5 

 for three weeks, the bulb F was surrounded by liquid air for 

 about one minute and then allowed to regain atmospheric 

 temperature. Its contents were then quickly driven into the 

 cloud-chamber in the following way. Some liquid air had 

 previously been placed in A and allowed to boil away through 

 C until the space between A and D was wholly free from 

 dusty air. The cloud-chamber and expansion apparatus had 

 also been previously exhausted to a pressure of about half an 

 atmosphere. Thus when F had regained atmospheric tem- 

 perature, the taps D and G were opened and the air in the 

 bulbs driven by a stream of: dust-free air into the cloud- 

 chamber. G was then closed and the pressure in H adjusted 

 so as to admit of an expansion of about 1*15, corresponding 

 to a pressure fall of some 10 cm. of mercury. The result 

 of tins expansion was a coloured cloud, showing that the 

 severe cooling process through which the well-dried air had 

 passed had resulted in the formation of a very large number 

 of nuclei. 



Control experiments were carried out in exactly the same 

 way, omitting only the cooling of the bulb F. Under these 

 conditions no drops were seen in the cloud-chamber on 

 expanding. 



One other point required testing beforehand. Simple 

 calculation showed that the pressure of the air in EF would 

 fall to about half an atmosphere when F was surrounded by 

 the liquid air. Now this fall of pressure might conceivably 

 liberate dusty air or other particles from the P 2 5 . It was 

 shown, however, in the following way that this complication 

 did not occur. The dried air in EF was expanded from 

 atmospheric pressure to half an atmosphere by connecting 



Phil. Mag. S. 6. Vol. 22. No. 130. Oct. 1911. 2 P 



