Since the emanation continuously breaks up, and is transformed into a solid type of matter, which is deposited on the surface of bodies, the volume of the emanation, when separated from radium, should contract at the same rate as it loses its activity, _i.e._ it should decrease to half value in about four days. The amount of emanation to be obtained from a given quantity of radium is a maximum when the rate of production of new emanation balances its rate of change. This condition is practically attained when the emanation has been allowed to collect for an interval of one month. The probable volume of the emanation to be obtained from 1 gram of radium was early calculated on certain assumptions, and from data then available the writer[265] deduced that the volume of the emanation from 1 gram of radium lay between ·06 and ·6 cubic millimetre at atmospheric pressure and temperature, and was probably nearer the latter value. The volume to be expected on the latest data has been discussed in the preceding section and shown to be about ·82 cubic mm. The volume of the emanation is thus very small, but not too small to be detected if several centigrams of radium are available. This has been proved to be the case by Ramsay and Soddy[266] who, by very careful experiment, finally succeeded in isolating a small quantity of the emanation and in determining its volume. The experimental method employed by them will now be briefly described. [Illustration: Fig. 61.] The emanation from 60 milligrams of radium bromide in solution was allowed to collect for 8 days and then drawn off through the inverted siphon _E_ (Fig. 61) into the explosion burette _F_. This gas consisted for the most part of hydrogen and oxygen, produced by the action of the radiations on the water of the solution. After explosion, the excess of hydrogen mixed with emanation was left some time in contact with caustic soda, placed in the upper part of the burette, in order to remove all trace of carbon dioxide. In the meantime the upper part of the apparatus had been completely evacuated. The connection _C_ to the pump was closed, and the hydrogen and emanation were allowed to enter the apparatus, passing over a phosphorous pentoxide tube _D_. The emanation was condensed in the lower part of the capillary tube _A_, by surrounding it with the tube _B_ filled with liquid air. The process of condensation was rendered manifest by the brilliant luminosity of the lower part of the tube. The mercury from the burette was then allowed to run to _G_, and the apparatus again completely evacuated. The connection of the pump was again closed, the liquid air was removed and the volatilized emanation forced into the fine capillary tube _A_. Observations were then made, from day to day, of the volume of the emanation. The results are given in the table below. Time Volume Time Volume Start 0·124 cub. 7 days 0·0050 cub. mm. mm.