732 Professor Sir James Deivar [June ll, 



was heated to near a low red heat for a short time. After these 

 alterations no increase in pressure was observed, from which it may 

 be inferred that the occlusion of the helium takes place mainly in 

 that part of the apparatus where the radium chloride is situated. 

 ' "^ On two occasions the charcoal was cooled in liquid hydrogen, 

 viz., after 165 hours, and again after 6.50 hours. The proportionate 

 reduction of pressure was the same in both cases, tending to show 

 that the composition or nature of the gas remaining uncondensed by 

 the charcoal in liquid air remained the same throughout, although 

 steadily increasing in quantity. 



In reference to this last point a separate experiment was made 

 in which pure helium produced by heating ()-5 grm. uranite and 

 passing the gas produced over 1 grm. charcoal cooled in liquid air, 

 was subjected, at a small tension measured on a McCleod gauge, to 

 the action of \ grm, of clean exhausted charcoal at the temperatures 

 of liquid air and liquid hydrogen respectively. The ratio of the 

 two pressures so obtained was in close agreement with that observed 

 in the radium experiment. 



A further test of the purity of the gas producing the permanent 

 pressure observed in the radium experiment with the charcoal cooled 

 in liquid air was made by simply cooling the bulb containing the 

 radium in liquid hydrogen, allowing the charcoal meanwhile to warm 

 up to 0° C. If any hydrogen had been present in the gas it is certain 

 that there would have been an increase of pressure recorded, since 

 although hydrogen is partially absorbed by charcoal in liquid air, yet 

 it would not be materially reduced in pressure by cooling in liquid 

 hydrogen. On allowing the charcoal therefore to warm up, any 

 hydrogen thus expelled would remain and cause an increased pres- 

 sure. Inasmuch as an increase was not recorded, it can be safely 

 assumed that no hydrogen was present, and thus the gas pressure 

 measured consisted entirely of helium. 



A confirmation of this was obtained spectroscopically as follows : 

 Two tin foil electrodes were placed round the narrow capillary 

 measuring-tube of the gauge, near the closed end. These were about 

 3 cm. long and about 1^ cm. apart, and were wired on with thin 

 copper wire. The gas was compressed into this capillary space, as in 

 taking an ordinary measure, to any pressure of the order of 2 or 3 mm., 

 while an induction discharge passed in the gas. The spectroscopic 

 examination of this discharge revealed only the six principal helium 

 lines, mercury, and a trace of the carbonic oxide spectrum. I have 

 shown that the carbonic oxide spectrum always occurs in electrode-less 

 tubes.* 



The curve showing the rate of production of helium is clearly 

 linear within experimental errors, as shown in Fig. I. The volume 

 of the gauge given was unfortunately erroneously estimated, and the 



* Proc. Roy. Soc, Ixiv., 237. 



