624 Profs. Rutherford and Barnes on the 



previously employed to measure the heating effect of the 

 radium emanation. 



It was necessary to use larger flasks than were previously 

 employed *, in order to admit the lead cylinders surrounding 

 the radium bromide tube. 



Our method was to surround the two glass tubes, into 

 which the small radium bromide capsule was lowered, by 

 exactly similar lead cylinders, 3 cm. high and o cm. in dia- 

 meter. A hole was bored to the centre of the lead cylinder, 

 into which the glass tube was fixed. The leads, supported by 

 the glass tube, were accurately centred in the flasks. A second 

 pair of cylinders, 5 cm. high, were cast and fixed in exactly 

 similar glass tubes. The surface of the leads was wrapped 

 around with a thin sheet of polished aluminium, to present 

 the same radiation surface in all cases. 



A pair of aluminium cylinders were cast in the same way 

 around similar glass tubes, and the surfaces of the cylinders 

 wrapped with polished aluminium. These cylinders were 

 exactly equal in size to the first pair of leads. 



It was a matter of a very few minutes to change the leads 

 for the aluminiums and vice versa, and the greatest care was 

 taken to have the cylinders always at the same level in the 

 flasks. Great reliance can be placed on comparative measure- 

 ments where a change of conditions can be made quickly. 

 We think that the weak point of the differential ice-calori- 

 meter measurements lay in the fact that apparently only one 

 side was used for the lead, as the conditions could not easily 

 be reversed. 



In our experiments the radium capsule was placed in each 

 flask alternately, and a change from the lead to aluminium 

 cylinders gave a direct comparison of the relative absorption 

 in the two metals. It can readily be shown that at least 50 

 per cent, more of the 7 rays must have been absorbed in 

 the leads than in the aluminiums. The heating effect of 

 the radium could also be obtained without a metal envelope, 

 as in our original experiments. In order to be independent 

 of such difficulties as are introduced by the difference in 

 thermal conductivity of the two metals, and the relative 

 temperature of the two surfaces, which seems to exert the 

 greatest influence in raising the temperature of the air in the 

 flasks, we used our original heating-coil to calibrate the 

 readings in every case. This consisted of 65 ohms of fine 

 silk-covered manganin wire, wound in a coil of approximately 

 the same area as the radium, and the heating current was 

 adjusted so as to give approximately the same heat emission 

 as the radium bromide used. 



* See previous description, loc. cit. 



