jected to various pressures and temperatures. In the case of 

 such substances as benzene and carbon tetrachloride a tem- 

 perature of from 60° to 90° is necessary to ensure a conveni- 

 ent gas-density. There is, however, this drawback to the use 

 of high temperatures, that the insulators begin to lose their 

 efficiency, and the joints cease to be quite airtight. I find it 

 necessary to use glass as an insulator instead of sulphur, for 

 the latter cracks under the unequal expansions due to altera- 

 tion of temperature. In the case of vapours, a certain quan- 

 tity of air usually finds its way into the apparatus, for, as 

 just mentioned, the joints leak somewhat at the higher tem- 

 peratures. The amount so entering is sometimes determined 

 by opening a communication between the vessel and an evacu- 

 ated bulb, and weighing the quantity of mixture drawn off. 

 The bulb and connections are placed within the oven, and 

 communication is made by opening a pinchcock, worked by a 

 key, projecting outside the oven. In this way condensation 

 in cold tubes is avoided. This method is not always em- 

 ployed, for as soon as the stopping power of the gas is suffi- 

 ciently well known the proportion of the mixture is much 

 more easily, and, I think, at least as accurately determined 

 by observation of the range of the a particles therein. 



The insulation leak is determined by measuring the de- 

 flection of the electrometer first for ten seconds, and then 

 for twenty. With no leak the latter should be double the 

 former. This is never quite the case, and the correction fac- 

 tor can be obtained from a comparison of the two values. 

 The factor to be applied to a ten-second leak varies from 

 about 1-03 at 40° C. to about I'lO at 70° C. ; at 90° C. it is 

 much higher. 



The total ionisation is measured in terms of the product 

 RI. The ordinate R is the range of the a particle, due to 

 that product of radium whose speed is next to that of RaC. 

 In air at 760 mm. pressure and 20° C, R = 4-83 cm. very 

 nearly. The abscissa I is the ionisation produced in the 

 chamber described when the radium layer is at a distance of 

 4*83 cm. from the middle of the chamber ; or, more correctly, 

 it is proportional to the ionisation that would be produced in 

 a very shallow chamber at that distance. The effect is 

 wholly due to the a particles from RaC, the chamber being 

 out of range of all the others. 



These two quantities R and I differ materially from each 

 other in two respects. To take the less important considera- 

 tion first, the former quantity lends itself readily to exact 

 measurement, the latter does not. The range of the a 

 particle in a gas can be measured to an accuracy of 1 or 2 % 



