382 Scientific Proceedings. Roi/al Diibliu Hociefii, 



them outside with " Plastacene." The radiuiii was phiced so tliat its lays 

 penetrated this paper tube between the ends of the glass tubes. Thus the 

 radium produced a strong ionisation due principally to j3 radiation. The 

 fig. 1 represents this arrangement, the paper tube being at A, represented by 

 dotted lines. 



In the case of the narrowest tubes (1"1 cms. diameter), the ionisation 

 would have been very small at B, and the quantity of air passing into 

 the gasometer would also have been rather too small to measure accurately. 

 Consequently two tubes were used, both of the same diameter, each being 

 in two lengths, and joined together with paper tubes as before. This pair 

 of tubes was placed so that the radium affected each equally, and they 

 entered BC through two holes in a cork, one at either side of the electrode i?. 

 The double tube delivered twice the quantity of air, and also gave twice 

 the ionisation current. It was assumed, of course, that similar conditions 

 obtained in eaoli tube. 



It was found that the largest velocities required for tubes .S'OS cms. 

 in diameter were just obtainable by the gasometer, the whole drum V 

 becoming filled in about a minute. In order to work witli larger tubes, the 

 gasometer was out of the question, and auotlier arrangement was adopted. 



The method then employed was to draw the air through the system ABC 

 by means of a fan. A small sei-ies-wound electric motor was used to 

 drive the fan; and a large range of Sf)eeds was obtainable by using a 

 variable resistance in the motor circuit. To measure the velocity of the 

 air an air meter was fitted at the end (C) of the chamber BC. The air meter 

 was tlie ordinary fan-type used for measuring air-currents and was first 

 calibrated over a large part of its range b}' means of the gasometer. 

 As it was feared tliat tlie fan might have possibly produced some 

 irregularities in the flow, a few experiments were made with a fine gauze 

 screen placed between the ionisation chamber and the air-meter. This 

 noticeably retarded the flow ; but the discontinuity in the experimental 

 curve occurred for the same mean velocity of the air. From this it was 

 concluded that the effects noticed were due to turbulence of tlie ordinary 

 kind setting in, and were not due in any way to the action of tlie fan. 



When using these arrangements it was occasionally found difficult to 

 obtain a smooth curve. This was thought to be due to drauglits in the 

 laboratory (a very large room, 30 feet by 90 feet, and 40 feet high), where 

 the experiments were being carried out. It was generally found, indeed, 

 possible to obtain very " good " numbers on a calm day — and one when the 

 room was unoccupied. The numbers were then usually quite steady, and 

 produced smooth curves like the majority of those shown. Subsequently, for 

 work with carbon dioxide gas, the gas was transferred from one gasometer, 



