CH. XXL] MODIFICATION OF OLD VIEWS 199 



ionised gas falls off' if we call upon it to carry more 

 than a certain current, called the saturation current. 

 See investigations by Town send and others briefly 

 referred to in Chapter VII. But I am not aware of 

 any experimental indication of such a limit in solids 

 or liquids, at present. 



In solids the pace of travel is unknown, though it 

 has been ingeniously surmised, and is thought to be 

 very great ; considerations of centrifugal force would 

 make the speed of each electron during an atomic 

 encounter equal to e/J(Kmr) or about 10 8 centimetres 

 per second ; views based on Maxwell's theorem about 

 equal distribution of energy among the particles of 

 mixed gases suggest 10 7 for the average speed of 

 electrons at ordinary temperatures in a solid where 

 they are free, that is, a hundred kilometres or sixty 

 miles per second; though since each particle is subject 

 to constant changes of direction, this is by no means 

 the pace of straightforward progression. But in 

 liquids they are attached to atoms, and the pace 

 of progression is known both theoretically and 

 experimentally with considerable accuracy, and is 

 comparable to an inch an hour for customary 

 gradients of potential. 



The total current is neu ; and to give a unit c.g.s. 

 current at so low a speed we can reckon how many 

 ions there must be. For e= L0~ 20 electromagnetic 

 units; so if we take u=W~ z centimetre per second, 

 the number of ions engaged in conveying the 

 c.g.s. unit of 10 amperes is n=10 23 . But, after all, 

 that is nothing very great ; it is only about the 

 number of atoms in a cubic centimetre of liquid. 

 By applying a greater gradient of potential the ions 

 can be made to move faster. By gradually narrowing 

 down the section of a liquid conductor under a given 



