134 LYMAN. — THE SPECTRUM OF HYDROGEN 
ELECTRIC APPARATUS. 
METHOD OF TESTING FLUORITE. 
In order to provide a window of the greatest possible transparency for the dis- 
charge tube, in those cases where the spectra of gases other than hydrogen were to be 
examined, it was necessary to test various specimens of fluorite which the author had 
at his disposal. The method was as follows : The piece under trial was attached to 
the plate-holder at the end of an arm in such a way that it projected to the right of 
the ways in which the holder moves. The length and shape of the arm was so ad- 
justed that when the plate-holder was at its highest position the fluorite was just 
above the right hand slit, but when the holder had been allowed to fall the fluorite 
slab fell with it and came between the slit and the mouth of the discharge tube. The 
and filled with hydrogen in the usual way, rather a wide slit 
was used. A photograph was then taken with the plate-holder at its highest position, 
thus the light path lay entirely iu hydrogen. Next, by means of the magnetic device, 
the plate-holder was allowed to fall until the specimen of fluorite came in front of the 
s it, the light from the tube now passed through the fluorite before reaching the slit. 
W comparing the two spectra, obtained one below the other on the same plate, the 
point in the spectrum at which the specimen cut off the light could be easi 
determined. & 
exhausted 
The electric apparatus used to excite the discharge tube has in almost all cases 
consisted of a transformer run from the 60 cycle 110 volt alternating circuit and pro- 
vided with a suitable rheostat in the primary. When such a transformer is used with 
a discharge tube containing gas at pressures from 1 to .1 mm. the addition of capacity 
across the terminals of the tube produces — with most gases — very little effect on 
the nature of the discharge because the low resistance of the tube after the current 
has once begun to pass does not permit the condensers to charge. If a spark gap be j 
introduced in series with the tube this difficulty is of course obviated. In all the 
earlier work no gap was used so the spectra obtained were due to a discharge practi- 
cally without capacity. The capacity when introduced consists of glass plates coated 
with tin-foil and has a value of perhaps .005 microfarads. In some of the work a coil 
with a mechanical break taking 12 volts and 5 amperes in the primary has been sub- 
stituted for the transformer. In the case of the metallic spectra used for comparison 
the spark has of course been brightened by the use of capacity. 
