156 Electrons conveying the Conduction Currents in Metals. 
list contains no value of p sensibly less than one must have 
its significance. I have not included sodium in the list, 
though Drude has investigated to some extent its optical 
properties, and found the ‘remarkable fact that light having 
a wave-length equal to that of sodium light is propagated i in 
it with a velocity 220 times greater than in empty space. 
If these results are confirmed we should have to conclude that 
for sodium p=*22, or that there is only one conducting 
electron in it for every four or five atoms. But Drude 
admits the possibility of a large error possibly diminishing 
the velocity of propagation twelve times, and this would 
aise the value of p to°76. There is also some uncertainty 
as to the electric conductivity of metallic sodium, so that for 
the present we may leave this metal out of account. 
If we once more ask ourselves to what extent these 
numbers may a trusted we may put down the assumptions 
made as follows 
(1) The conta of e/m in all metals and its value as 
determined by experiments on cathode rays ; (2) the absence 
of opacity due to selective absorption. 
As regards the first point, assuming the constancy of e/m, 
an error in the actual value would increase or diminish p in 
the same ratio for all metals. It is not altogether satisfactory 
that a quantity the constancy of which is asserted with con- 
siderable confidence is not known with greater accuracy, 
different methods disagreeing in their result by more than 
100 per cent. We know in fact more about the charge e 
carried by the electron than about the more easily measured 
value of e/m. Those experiments in which the method first 
devised by myself is used, and e/m calculated from the 
magnetic deflexion and potential-difference, give, in the hands 
of later observers working under more favourable conditions, 
the higher values which I have adopted above. The combi- 
nation of electrostatic and magnetic deflexions devised by J.J. 
Thomson gives a number about half as large. A great deal 
is to be said for Wiechert’s method, in which the velocities 
are observed directly, and which gives as a mean about 
13x10’, a result intermediate between Simon’s value of 
1:86 and the lower values in which electrostatic deflexions 
are used. If Wiechert’s number be adopted the above values 
of p ought all to be multiplied by 1-4. 
Experiments with electric discharges and radioactivity have 
made us familiar with ions carrying positive electricity and 
possessing a mass much greater “than that of the negative 
electron. But if the mass of the electron were increased say 
five times, its effect upon v« would be diminished 25 times, 
: 

