Mr. H. A. McTaggarttfrc the 

 rest by using a fine tube fitted with a screw as shown in 



300 



rest 

 fig. 3. 



Fier. 3, 



J 



"Y 



It is made by drawing down an ordinary piece of glass 

 tubing to a fine capillary and fastening a plate with a screw 

 through its centre to the large end. When this apparatus 

 is filled with water, if the screw be turned back the water 

 retreats along the capillary followed by air, any quantity of 

 which may then be expelled into the cell by turning the 

 screw forward again. The cell w r as set in motion by a small 

 electric motor connected to it by a belt of fine thread 

 running over pulleys* This apparatus gave, even when set 

 up in a more or less careless way, a very interesting exhibi- 

 tion of the bubble moving along the axis, now fast, now slow, 

 as the potential between the ends was altered. 



The velocity of a bubble of gas in such a cell is not, how- 

 ever, its real velocity through the liquid. In tubes of such 

 small bore the endosmose sets up a current along the wall 

 which returns along the axis, increasing or diminishing the 

 true velocity according to circumstances. Another error, 

 due to currents in the liquid, is introduced if the speed of 

 rotation is not uniform, an error which it is almost impossible 

 to avoid with an ordinary type of motor. In making actual 

 measurements a cell of the form shown in fig. 4 was used. 



Fi-. 4. 



The glass stoppers AA with platinum electrodes on their 

 ends cc are ground into the tube JB and furnished with metal 

 bearings to run on the pivot points. A wire running through 

 the stoppers connects the electrodes with the bearings. 

 Bubbles of gas are introduced through a small hole at E. 

 To ensure uniform rotation a Rayleigh motor (Scientific 

 Papers, Lord Rayleigh, vol. ii. art, 56, p. 355) was used 

 which drove a beltiof thread over the pulley D. This worked 

 fairly well if the thread was kept moistened with oil to 



