386 Mr Bevan, The Influence on Light reflected from and 
values ( loc . cit. p. 43) both a and 2 h for several metals are greater 
than 10. For a rough approximation we can, in the expressions 
for v x and v 2) neglect l 2 compared with a/ 2 and hf 2 , and write 

'y (a + Va 2 + 
b% 
f 2 
vi= y(-«+ Va 2 + 6 2 ), 
or v x = fn, 
v 2 = fnk. 
That is, v x , v 2 have approximately the same values as for light 
incident perpendicularly on the surface. 
We thus obtain 
_ 2 nj> {l-n' 2 (l -k 2 ) + 2n' 2 k 2 } 
ripf* (1 + A; 2 )[{1 — n' 2 (1 — A; 2 )} 2 + 4 n 4 k 2 ] 
- xPln 1 -n' 2 (l- k 2 ) + 2 n 2 k 2 
'pf i (rV n (1 + k 2 ) [{1 - n' 2 ( 1 — A; 2 )] 2 + 4 n'*k 2 \ 
If now Sr is the angle of incidence, l — f sin S-, n =f cos f= ~ , 
2?j- 
and p — , where t is the period of the vibration. 
T 
We obtain therefore 
where 
A = 
HI — P sin ^ cos S-, 
47 rcr 
n 2 — k 2 ) -f- 2n' 2 k 2 
n (1 + & 2 ) [{1 - n 2 (1 - A; 2 )} 2 + 4?i /4 & 2 ] ’ 
The fraction of the time of vibration or the fraction of the 
wave-length, if we regard the change of phase as a length, is owing 
to the current 
A\t 
47 ra 
P sin ^ cos S-. 
The units adopted so far are Gaussian ; to express this result 
in electromagnetic units we observe that a is in these units 
the same as in electrostatic units, and the unit of resistance in 
the electrostatic system is v 2 the electromagnetic unit; if then g 
is the number expressing the specific resistance in electromagnetic 
units 
v 2 
is the number expressing the same quantity in Gaussian 
