342 Lord Kelvin and Messrs. Maclean and Gait. [Feb. 21, 



minutes. The rate at which the air was blown in in these experi- 

 ments was such as to displace the entire volume of air in the vat in 

 half an hour. 



12. Curve (8) shows the rate of electrification of air, in one of 

 the fifteen experiments, when thus bubbled through the water in the 

 (J-tube and then admitted into the vat. 



13. Two U -tubes, in series, with water in each, did not seem to 

 give a perceptibly cumulative effect. 



14. The effect of one or more wire gauze strainers between the 

 U-tube and the vat, or between the (J^be and the bellows, was 

 next tested. The gauzes were placed between short lengths of lead 

 tube, which were held together by a rubber tube slipped over them. 

 The arrangement is shown by longitudinal and cross sections in fig. 5. 



India Rubber 



Twelve wire gauzes, with or without cotton- wool between them, 

 placed between the bellows and the U^ube, did not prevent the sub- 

 sequent electrification by bubbling of the air thus filtered. But when 

 placed between the U -tube and the vat they almost entirely diselec- 

 trified the air, even without the cotton- wool, and still more decidedly 

 when cotton-wool was loosely packed between the wire gauzes. A 

 single wire-gauze strainer produced but little of dis- electrifying effect. 



15. The interpretation of these experiments is complicated, and 

 the time required for each is lengthened, on account of the large 

 mass of air in the vat to start with, whether uncharged or retaining 

 electricity from previous experiments, and also on account of the 

 effect of the water-dropper itself. Hence, in our later experiments, 

 we fell back on the arrangement shown in fig. 2, by which we test 

 the electrification of the liquid, and not directly that of the gas blown 

 through it. 



16. In our first experiments with this apparatus the amount of 

 the electrification did not seem much affected when a paper cover 

 was put on the jar, or when we tilted the jar as shown in fig. 3. 

 We now made a large number of tests with different covers and 

 screens (chiefly of sheet copper or sheet zinc, or brass wire gauze) at 

 different heights above the liquids, and we concluded that, if the 

 screens are not within a centimetre and a half of the liquid surface, 

 they do not directly affect the magnitude of the electrification obtained. 

 In nearly all of the subsequent experiments a horizontal circular screen 



