433 
«* These experiments indicate three things, as exerting a 
potent influence on this peculiar stratification of the light, the 
chemical character of the medium, its density, and the inten- 
sity of the induced current which is discharged through it. 
These three may be reduced to one, namely, the quantity of 
electricity which is transmitted in a given instance; and which 
depends on the conducting power of the circuit and the elec- 
tro-motive force. The superiority of hydrogen and coal gas 
over common air arises from their being better conductors; and 
as any gas becomes a better conductor by rarefaction, this ex- 
plains why, through the whole series, the phenomena are most 
distinct when the vacuum is best.* Indeed, for each gas there 
seems to be a limit of density above which stratification does 
not occur. Thus, for hydrogen, the bands were not seen (10) 
till its pressure = 0':70, and in another instance (11) where 
the electric tension was higher, till 1-70. In air I have never 
seen them when its density was more than 01.30. This limit 
will be highest for the best conductors; but it is seen by com- 
—_—-- =_ 
paring (14) with (7) that inferior conducting power may be 
_ more than compensated by higher exhaustion. 
«The proofs of the influence of electric tension are still 
more numerous. In (3), (6), (13), and (14), where Smee’s 
battery was used, the strata, though distinct at first, faded 
away in a few seconds, reappearing when the circuit was 
broken for a short period. The power of this battery, like 
that of other single fluid ones, declines rapidly when the cir- 
‘cuit has small resistance, and recovers when that is broken; 
and in these two states the induction spark in air was 0:+20 and 
0:05: this decline caused the disappearance. ‘The trifling 
| effect of frictional electricity (1), as compared-to that of the 
_ * As the rarefaction proceeds, the intensity of the spark at the contact 
breaker decreases; and so also we should suppose the vibration there; yet 
