769 
Now negative values of v and of x have no meaning and must be 
ignored; accordingly, from (6) and (7) we conclude that 
v 1 — v t = <3 (8) 
— ^2 — ^ (9) 
Thus, as was to be expected, the differences v x — v 2 and x x — x 2 
are independent of ® and <<J3. 
Having regard to (5) in Art. 1., we may write (8) and (9) in 
the form 
8 nn ^ e 2 N 
*1 — V 2 =- 2j - 
c m 
h(n o 2 — n 2 ) 
(10) 
' (V — » n*f + 4&»n* 
Snn %?e 2 N 
x 1 — x 2 = - 
c m 
h. 2kn 
(11) 
( n Q 2 — n 2 ) 2 -|- 4:k 2 n 2 
If, as in the paper alluded to, we consider the behaviour of a 
«mono-electronic» substance, supposed to contain only one class of 
movable electrons, our result can be expressed very simply; we 
have then 
8 Tin 
/II - 
e 2 N 
h (n 0 2 — n 2 ) 
* 2 e ' 
m 
(V 
— n 2 ) 2 -)- 4 k 2 n 2 
8 nn 
e 2 N 
h . 2 kn 
c ' 
m 
(V 
— w 2 ) 2 + 4& 2 » 2 
These lead to 
v x — v 2 _ n 0 2 — n 2 
x x — x 2 2kn 
( 12 ) 
(13) 
(14) 
a remarkable result in which the constant e 2 N/m : as well as the 
specific coefficient h has disappeared. Perhaps the most interesting 
point of view from which formula (14) may be regarded is to con¬ 
sider it as affording a means of calculating, at least in the case of 
certain substances, the magnitude of that important constant, the 
dissipation coefficient k which has hitherto eluded numerical eval¬ 
uation. 
§ 5. In deducing the foregoing results from equation (10) of 
Article 3 we have ignored the presence of one term, that involving 
the quantity In order to avoid much complication, other effects* 
of the same order of magnitude as the‘term in have been left 
out of account at a former stage; there is not much object there- 
