THE STRUCTURE OF THE NUCLEUS. 13 



relatively so enormous that the electric currents could not be measured by the 

 same electrometer and the occurrence of an internal metallic contact or bi'eak in 

 the insulation was immediately suggested. I therefore overhauled the condenser 

 carefully, inserted an internal bushing, replaced the internal i-od l)y a new one, 

 etc., all without effect. The condenser showed good insulation after putting it 

 together, but became a conductor immediately after the passage of the first phos- 

 phoius emanation. Permanent leakage due to dislocation of the solid parts was 

 thus out of the question. 



This conduction vanished over night. It was ]'e[)i'oduced as soon as fresh 

 phosphorus air passed through the condenser. It then I'emained permanent, 

 though gradually diminishing for liours, and was nearly gone again next day. 

 Hence two causes are suggested : eithei' a film of residual moisture aspirated off 

 from the phosphorus gi'id (which, however, was as usual carefully dried by squeez- 

 ing in a press between folds of blotting paper, and then exposed to the air, 

 so that only traces of moisture can be in question) was precipitated in the con- 

 denser to the detriment of the hard rubber insulation ; or else some form of 

 emanation given off from the phosphorns made the condenser radio-active. In- 

 cidentally, I may advert to the extreme caution needed before such radio-activity 

 can be assumed, the behavior in both cases being essentially alike. 



Warming the condenser seemed to be useless. Moderate amounts of dry air 

 (say 7 liters flowing out in about 10 minutes) passing over the phosphorus were 

 nearly ineffective. It was no I'emedy to remove the phosphorus and pass dry aii' 

 alone in the forward direction. Separating the condenser from the water bath 

 did not change its conduction. Thus I found, for instance, for the condenser 

 alone and free fiom air current, ch/dt = 33 ; an hour later, ds/dt = 25 ; next day 

 ds/dt = 7 ; good insulation, d,s/dt = 2. 



At = 30° (water bath temperature), the tendency of the condenser to con- 

 duct permanently was at first accentuated but soon completely wiped out. The 

 electric current reached a normal value. This appeared so much like a moisture 

 error that I further tested it by passing the air cui-rent backwai'ds, through the 

 condenser first and then over the phosphorus into the atmosphere, in this way 

 drying both parts. Insulation of the condenser was thus at once restored. Again, 

 on passing a considerable volume of diy air (say 15 litei's, slowly) over the 

 phosphorus, this too lost its powei- to make the condenser permanently con- 

 ducting. Hence in the experiments of the following table the phosphorus was 

 first dried in this way in a current of dry air. The work then progressed 

 smoothly, showing the relation of the emanating activity of phosphorus to 

 temperature in a new light. 



I may add again that in none of my earlier experiments were like discrep- 

 ancies encountered. Possibly a corroded copper grid may be hygroscopic, some- 

 thing like platinum black; or the room may have been exceptionally diy. 



16. Specially dried j)hosphor us grids. — In tables 6 and 7, 6 is the temperatui-e 

 of the water bath, d Vjdt the volume of dry air in liters passed per minute over 

 the phosphorus: ds/dt is the corresponding initial radial electric current in the 



