416 Prof. A. P. Chattock on the Velocity and 



and take this momentum as spread over the area 2*3 sq. centim., 

 we have the means of obtaining an upper limit to v, the 

 average velocity of the air. 



40 

 For + discharge ^^ = v 2 x density of air, 



or ?•+ =120 centim per sec. 



Similarly v— = 95 „ „ 



The upper limit of v is thus only 2 per cent, of the actual 

 velocity of the + ions, and 1*4 per cent, that of the — ions ; 

 and the percentage is no doubt much less in reality, as 

 one cannot suppose that the ionic drag is all turned into 

 momentum. 



In the "point and ring" apparatus the field, and therefore 

 the velocity of the ions, is unknown ; still we may infer that 

 v is negligible here also from the straightness of the pressure- 

 curves ; for the lines of flow of gas must be very different to 

 those of the electricity on the average, and the motion of the 

 gas would thus alter that of the ions in different proportions 

 at different points in their path, if it were appreciable. 



The conclusion therefore is that V is practically equal to the 

 velocity of the ions relatively to the air they travel through. 



Spread of Current between Point and Plate. 



The close equality of the actual velocities of the positive 

 and negative ions makes it a simple matter to explain the 

 fact, noted above, that the negative ions spread more than 

 the positive in passing from the point to the plate. For, the 

 velocities of the ions being the same, the amount of free 

 electricity between the point and the plate is the same for 

 given values of z and c. But the repulsion of this electricity 

 tor itself, which is the cause of the spreading, is thus also the 

 same for both ions if they spread equally ; hence, as the 

 negative ions move faster than the positive under a given 

 force, they will spread farther. 



By altering the strength of the field we affect the spread 

 in two ways. If the field is increased the time taken by an 

 ion to pass from point to plate is decreased in proportion. 

 Also the quantity of electricity between point and plate, i. e. 

 the spreading force, is decreased in the same proportion. On 

 these lines it is easy to show that if the spread is not too 

 great, and if c and z are both constant, V cc 1/F 2 for con- 

 stant spread. This suggests a simple method of comparing 

 the values of V for different physical conditions of the gas, 



