Production of Helium by Radium. 593 



Calibration of Measuring Portion of Apparatus. 



As already explained the vessel A was in its general design 

 similar to an ordinary MacLeod gauge. For the purposes 

 of the experiments, it was calibrated in the following manner. 

 The entire apparatus was completely freed from air, the 

 vessel C being heated to a low red heat to remove the gases 

 absorbed in the coconut charcoal, and the pressure in the 

 apparatus was reduced to a very low one by continued 

 pum; ing with the large pump P. The level of the mercury 

 in A was raised to the point a, the stop-cocks/^ d, a, and c 

 were closed, and a small quantity of oxygen mixed with pure 

 helium (obtained from a crystalline uraninite) was introduced 

 into A through the tube L and its connexions. The stop- 

 cock b was now closed and the liquid air was placed around 

 the tube C. After about twenty minutes the mercury was 

 raised in A and the volume and pressure of gas in the 

 capillary tube E was measured. With the mercury still 

 raised to the top of A, the stop-cocks a, b, and / w T ere 

 opened, the liquid air was removed from around C, which 

 was allowed to warm up and was finally heated to a low red 

 heat, and the tubes C, U, M, &c. were pumped out thoroughly. 

 The stop-cocks a and b were now closed, liquid air was placed 

 around C and U, and the mercury in A was lowered to the 

 point a. This permitted the helium which had been pre- 

 viously trapped in E to expand into the tubes U and C. The 

 mercury was again raised in A and the volume of helium in 

 E was measured. The ratio of the volume of helium in E 

 in the second measurement to the volume of helium in E in 

 the first, was the, ratio of the amount of helium trapped in E, 

 on raising the mercury in A, to the total volume of helium 

 contained in A and its connexions. With the mercury still 

 raised in A, the other tubes were again pumped out and the 

 expansion of the helium in E into the connected tubes U, C, 

 was again repeated. Several operations were carried out in 

 this manner giving results in excellent agreement with one 

 another, and indicating that 71*5 per cent, of the total volume 

 of helium in the apparatus was trapped in the tube E when 

 the mercury in A was raised under standard conditions. 



By pumping out the tube A and its connexions to the 

 highest possible vacuum obtainable by the use of the pump P, 

 closing the stop-cocks a, b } c, and immersing the tube C in 

 liquid air for twenty minutes or more, it was possible to 

 obtain so low a vacuum in the tube A that, on raising the 

 mercury in A, no measurable increase of pressure in E was 

 indicated when the mercury was raised in E to within a 



