165 



apex of a conical hole made in a block of lead. By means 

 of a powerful magnet, the poles of which are shown, most 

 of the iS rays could be prevented from passing out of the 

 conical hole. The magnet and lead block containing the Ra 

 were surrounded by an iron case, to prevent the magnetic 

 field from producing any effect in the ionization chamber. 

 Resting upon the iron case and suitably insulated from it 

 is placed a lead plate one inch thick, with a circular por- 

 tion removed from the centre. The ionization chamber stand- 

 ing upon this lead plate is made of lead 125 mm. thick, and 

 contains an insulated electrode in the form of a circular 

 wire ring, suitably protected by sulphur, and connecting to 

 the electrometer or to earth by suitable keys. 



In the path of the y radiation which proceeds from the 

 conical hole, plates of material which are used as radiators 

 can be placed horizontally as at e. When increasing the 

 thickness of radiator the top plate is always kept in the 

 position e, and additional thicknesses are placed immediately 

 below and in contact with it. When large thicknesses are 

 required, plugs of the material may be inserted in the coni- 

 cal hole. 



The ionization chamber was filled with ethyl -chloride ; 

 this increased the effect considerably and worked very satis- 

 factorily. 



It was necessary to balance the leak due to natural ioni- 

 zation in the chamber, and to y radiation which was not 

 completely absorbed by the lead block and plates. For this 

 purpose a balance chamber of about the same size as the 

 ionization chamber was placed against one of the sides of 

 the iron case insulated from it. 



This was connected up in the usual manner, and was 

 found to give quite sufficient balance with the ionization pro- 

 duced by the y radiation which came through the sides of 

 the lead block containing the Ra. The electrometer was a 

 sensitive instrument of the Dolezalek pattern, giving a scale 

 reading of about 4,000 divisions per volt ; each division could 

 easily be subdivided to tenths. The charge was allowed to 

 pass into the electrometer for 30 sec, and the mean of the 

 first and second swings was taken as the deflection. 



In fig. ii., curve A shows the effect of increasing the 

 thickness of a Pb radiator. In an experiment, the zero leak 

 was first obtained as a mean of several determinations, and 

 this was subtracted from the measurement obtained with the 

 radiator in position. It will be seen from the curve that 

 the effect increases until a thickness of between 5 and 6 

 mm. of Pb is reached, after which it decreases gradually. 



