730 Dr. E. W. Boyle on the Behaviour of 



upper part of the tube. This slow increase of emanation in 

 the upper part of the tube, with the consequent growth of 

 Ra A and Ra C, causes the ionization to increase, but much 

 less rapidly than before, and thus gives the final, long- 

 continued, portion of the curve. 



Within the range of temperature covered by the initial, 

 flat, portion of the curve, the emanation was certainly vola- 

 tilizing, though not in sufficient quantity to affect the galva- 

 nometer. But a more sensitive instrument should be affected 

 within this range, and this is shown to be the case by curve 

 B. This curve corresponds to a quantity of 0*30 m.r.e. of 

 emanation and an electroscope with a pressure of a few 

 mm. Ho- in the ionization chamber for measuring the ioniza- 

 tion. Instead of the flat portion extending from the lowest 

 condensing temperature to — 163° C, as in the case of 

 curve A, it extends here only to —171° C, after which the 

 carve rises, and then bends towards the temperature axis, as 

 explained. The measuring device in this case was so much 

 more sensitive than the galvanometer, that it could detect the 

 changes in the vapour phase of the emanation at temperatures 

 as low as —171° C, whereas the galvanometer could only do 

 this as low as —163° C. 



A still more sensitive device, viz. an electroscope with 

 atmospheric pressure in the ionization chamber, could detect 

 the changes in the vapour phase at a lower temperature still. 

 This is the case of curve C, where 0*01 m.r.e. was employed. 

 From the above it follows that if it were feasible to condense 

 a very large quantity of emanation, and employ at the same 

 time only one measuring instrument possessing the required 

 ranges of sensibility, we could obtain a single ionization 

 curve of the form already shown. But this curve would 

 rise immediately from the lowest temperature of condensatiou, 

 and would cover a wide range of temperature before bending 

 towards the axis of temperature. In other words, the con- 

 densed emanation would begin to volatilize at the lowest 

 temperature of condensation, and would continue to volatilize 

 through a wide range of temperature until the emanation was 

 entirely free from the condensing surface. 



In the experiments we cannot be sure that the amount of 

 emanation volatilized at any temperature was the exact 

 amount required to saturate the space of the containing tube 

 at that temperature, and from the curves given we cannot 

 calculate the vapour pressures. The experiments were only 

 qualitative, and the complications introduced by the rate of 

 rise of temperature, the growth of RaA and Ra C with their 

 different ranges of a-particles. effectively prevent our utilizing 



