TROWBRIDGE. — HIGH ELECTROMOTIVE FORCE. 207 
the tube in each case had been exhausted to the same degree. It was b] <Hly seen 
that a great excess of oxygen did not result in dimming or decreasing the brilliancy 
of the red line of hydrogen. On the contrary, up to a certain point the addition of an 
excess of oxygen increased the brilliancy of the hydrogen line. The mass law ol 
physical chemistry, therefore, does not seem to apply in the spectrum 
water vapor and oxygen. If one plots the intensity of the hydrogen red 
abscissa and the nressure of the exhausted tube as an ordinate, one can re 
b 
changes in intensity of the line by 
The intensity 
proceeds, until, the current being kept constant, when a certain pressure is reached the 
brilliancy of the red line in the tuhe connected to the pump b< pns to diminish. The 
curve representing the run of intensity has a point of inflexion and returns to the 
axis of Y; thus between the origin and this point of inflexion there are two vain- oi 
pressure giving the same intensity. The hydrogen line diminishes in intensity as 
the amount of water vapor diminishes, the strength of exciting current remaining 
constant. 
This experiment illustrates the fact that oxygen can be pr< cut with an exce s of 
water vapor and yet does not manifest itself spectroscopically. One should not, there 
fore, conclude from the absence of oxygen lines in the solar protuberance and in the 
hydrogen types of stars that oxygen is not present. It is maintained by some that 
hydrogen in such a mixture ai 
)gen in such a mixture and water vapor and ox\ 
oxygen, on account of its resistance, does not pa 
However this may be 
from the absence of oxygen 
the spectrum of the solar protuberance that oxygen is not present 
REVERSALS OF METALLIC LINI -. 
discharge of powerful condenser sparks between metallic terminals 1 cm 
Tl~ — s 
glass or quartz tubes of 1.2 mm. diameter, fdled with rarefied gases, afford 
pe 
In general, the strongest lines of the metals employed 
nals are reversed. These reversals (dark lines on the positives) coincide in position 
with that of the same lines in air, while the general broadening of the line is toward 
the red. This fact is clearly seen in the spectrum of the line of silicon wave len th 
2881 (Fieure 9, Plate XXVII). Thus we have the effect noticed in the spectrum of 
n line dark on one side and bright on the other. A slight 
able stars, 
ily evident, might lead 
one 
de that there had been a shift 
on the whole of the line in question toward the red. A changing thickness, or 
