Temperature on Spontaneous Ionization. 619 



raised to about twice its normal value. Similar effects were 

 noticed when the air was ionized by Rontgen rays or by 

 /3-rays from radium placed outside the vessel. 



13. The magnitude of Q increases, both absolutely and 

 relatively to the normal leak q, with an increase in the volume 

 of the vessel and with an increase of the ratio of the heated 

 surface to the whole surface of the vessel. In the very large 

 Mat vessel (a cylinder of 1 metre diameter and ID cms. 

 height) used in the former experiments, Q' may be so large 

 that the direction of the apparent current during the initial 

 cooling is reversed and there is a leak against the electric 

 field (which amounted to some 50 volts per cm.). The 

 reversal of the leak was so surprising that every effort was 

 made to discover its cause in some peculiarity of the experi- 

 mental arrangements. However, it has been observed in two 

 different vessels with no material (except the air inside them) 

 common to both, in which the leak was measured by different 

 methods. 



14. Q is increased by the presence of dust in the air ; 

 but I have never been able to reduce its value below a certain 

 limit by blowing in air through a plug of cotton-wool two feet 

 long. Moreover, if dusty air is used, the effects obtained on 

 heating up the vessel a second time will be less intense than 

 those observed at the first heating; but if dust-free air is 

 used, alternate heating and cooling for days will not stop the 

 change. Xo difference could be observed between air 

 saturated with moisture and air dried by a long tube of 

 calcium chloride, except that the former spoilt the insulation. 



15. A few experiments were made in which the air was 

 kept at constant volume instead of at constant pressure. The 

 changes observed were of the same nature as before. 



16. The above summary by no means includes all the 

 observations that have been taken, but includes all that 

 I have been able to reduce to any sort of order. On several 

 occasions a wholly new set of heat-changes has appeared ; 

 but when an attempt was made to investigate them later, 

 they could not be reproduced. 



Explanation of the Changes. 



17. OmV one explanation can be suggested which seems at 

 all competent to account for the effects which have just been 

 described. It may be supposed that the air of the vessel 

 contains charged particles much heavier than the ordinary 

 ions, possessing, in fact, amass so large that the gravitational 

 forces acting on them are larger than the electrical forces 

 due to the field in the vessel. When the vessel is cold 



