from a CaH.^, and are more liable to initial recombination, Tliis 

 is in agreement with experiment : it is far harder to saturate 

 benzene than acetylene. 



The secondary ionisation would appear to take place 

 within rather than without tlie molecule, because the amount 

 of it does not depend upon the distance of the molecules from 

 one another. The total ionisation is independent of the pres- 

 sure. It is certainly not due to the electric field, for if it 

 were there would be no saturation value of the current. 



I subjoin the details of one or two of the many experi- 

 ments which Mr. Madsen and I have made. We hope to give 

 a fuller description at some future time. 



Determination of Stopping Power and of RI in Pentane. 



Electrodes, 3 mm. apart (nearly). Volts applied = 300. 



Temperature of apparatus = 35° C. 



Apparatus charged with vapour from standard pentane. 



These being plotted, it is found that R = 2-95, 1 = 1044, 

 the copper leak having been deducted. 



Thus, R = 2-95 in this mixture of pentane and air, at a 

 pressure of 41*15 cm., and a temperature (observed) of 308° 

 (absolute). But at a pressure of 760 cm. and 293° absolute, 

 R in air is 4*83. Hence the mixture stops — 



4-83 7600 308 ^ ^ ^ ^. 



= 3-14 times as much as air. 



2-95 4115 293 



A special set of readings at 3'2 cm. is now taken, three 

 for ten seconds and three for twenty seconds. The means are 

 1196 and 2325 respectively. Comparing these, it is found 

 that the ten-second reading should be multiplied by 1-03 in 

 order to allow for leakage by the insulators. 



Again, a set is taken with 600 volts between the plates, 

 and it is found that the mean reading, when the copper leak 

 is deducted, is 1134. At the same time the reading for 300 

 volts, copper leak being deducted, is 1088. Thus saturation 

 is nearly complete. 



