25 



middle gauze. Readings were taken — (a) With the copper 

 ring active, and air which had previously passed over the 

 uranium surface drawn through the chamber, (h ) With the 

 ring active, but the uranium removed, and still using the 

 draught. ( c ) With air passed over the uranium surface after 

 the activity of the ring had disappeared. The results showed 

 that the amount of ionization produced by the copper ring 

 in air was the same as in the air which had just pre- 

 viously passed under the influence of uranium. 



Again, had the want of saturation, shown in the curves 

 of fig. 6, been due to the presence of clusters, or partially com- 

 bined pairs of ions, an alteration of the shape of these curves 

 might have been, expected if the number of ions was reduced 

 without altering the number of clusters. No alteration in 

 the shape of the curves could, however, be detected when the 

 uranium was moved some distance further from the gauze, 

 or even when the ionization was still further reduced by plac- 

 ing along the axis of the tube leading to the gauze a 

 wire maintained at a few volts difference of potential from 

 its surroundings. 



Recently a paper by M. de Broglie <^"^ has appeared, in 

 which tests similar to those just described have been pear- 

 formed upon the gases coming from flames. His results are 

 in agreement with the foregoing in so far that no clusters 

 are found when water vapour is not present. 



Having found no support to the cluster theory, we must 

 look to diffusion for an explanation of the curves 

 in fig. 6. The distortion of the field in the imme- 

 diate neighbourhood of the wires of the middle gauze and its 

 penetration through the wider gauzes have accounted for the 

 results shown in fig. 5. The came cause, in gradually pre- 

 venting diffusion of the ions to the lower side of the bottom 

 gauze, will explain the difference in shape of the ionization 

 curves obtained with broad- and with narrow-mesh gauzes, 

 also the approach, in man}- cases, to the same final value of 

 current, as the field strength is increased. 



One difficulty which arises in the foregoing explanation 

 is that while the field exerts its influence through the air 

 spaces of the bottom gauze, if the gauze be positively electri- 

 fied, negative ions will, of course, be more readily drawn to 

 the lower surfaces and sides of the wires of the gauze than 

 if the gauze were uncharged; and positive ions should, appa- 

 rently, be similarly repelled with a force which depends upon 

 the strength of the field. Now, it is known dO) that when a 



(9a) Le Radium, tome iv., No. 7, July, 1907. 



(10) J, .J. Thomson, Elements of Electricity and Magnetism. 



