NTCLKI OF POSITIVELY AND NEGATIVELY CHARGED IONS. 



303 



As already pointed out, the least expansion necessary to produce such showers 

 in air initially saturated is identical with that required to cause condensation on ions 

 really on negative ions, as the experiments described in this paper show. This is a 

 remarkable coincidence if we are really concerned with nuclei of entirely different 

 kinds. Further, it was during experiments made in the absence of ionising agents 

 that I first noticed indications of an increase in the number of the drops about the 

 point v t /v l = 1*3 1, the expansion now proved to be that required to cause water to 

 condense on positive ions. 



These considerations seemed to furnish strong ground for believing that the very 

 few nuclei always present actually are ions. In a recent paper,* it is true, I 

 described some unsuccessful attempts which I had made to remove the nuclei by 

 applying a strong electric field ; I did not, however, consider these experiments to 

 be conclusive evidence against the ionic nature of the nuclei. I have, therefore, 

 recently subjected them to a much more severe test by means of a differential appa- 

 ratus (fig. 3). 



Fig. 3. 



The mechanism for causing the sudden expansion was the same as in the other 

 experiments, and is not shown in the figure. A front view of the apparatus is given 

 on the left, a side view on the right. It consists of a short glass cylinder, 4 centims. 

 in diameter and 2 centims. long, the ends being ground smooth and closed by plates, 

 that forming the front face being of glass, the other of quartz. (The quartz was for 

 experiments with ultra-violet light described below.) A thin brass plate, 2 centims. 

 wide (reaching, therefore, from back to front of the apparatus), divided the vessel 

 into two equal chambers. On each side of this, at a distance of 8 millims., was a 

 parallel brass plate of the same width, 2 centims., but not reaching to the lower wall 

 of the cylinder. The brass plates were covered with wet filter paper. 



A difference of potential of 320 volts could be maintained, by means of a series of 



* ' Phil. Trans., 1 A, vol. 192, pp. 403-453, 1899. 



