LYMAN. 
TI1E SPECTRUM OF HYDROGEN. 
129 
each end by a stop-cock, thus the gas does not flow through the system directly but 
stands over the material for some minutes before entering the spectre- :ope. The pei 
feet dry 
of the gas is necessary for the success of 
All connections l>e 
tween hydrogen apparatus, tubes, and spectroscope are of glass. The exhaustion is 
effected by a " Geryk" oil pump driven by an electric motor, the pressure is read by a 
McLeod gauge properly protected by drying tubes. Here again all connections arc of 
glass. All air admitted to the spectroscope is passed through a separate set of drying 
tubes. These precautions have been found necessary to prevent the appearance of ab> 
sorption bands. The joint between the brass receiver and the system of glass tubing 
is made by a glass sleeve sealed with De Kotenski cement. Though this form of joint 
leaves something to be desired, nothing better has as yet been devised. 
The use of a discharge tube separated from the receiver by a fluorite window neces- 
sitates a separate pumping system, for the tube must be exhausted apart from the 
receiver and filled with the gas to be studied. For this purpose a mercury pump by 
Kiss of Buda-Pesth has been used. The hydrogen is made electrolytically from a solu- 
tion of barium hydrate and is dried over phosphorous pentoxide. 
The form of the discharge tube depends upon the manner of 
making the experiment. If the tube is to communicate directh 
with the receiver so that the whole apparatus is fdled from the 
receiver with hydrogen, the usual form of capillary tube with 
ring electrodes 
were 
follow 
is employed. The dimensions in a typical case 
: Length of capillary 6.4 cm., internal diameter 
2.5 mm., diameter of electrodes 1.6 
distance of mouth of tube 
to electrode 4.5 cm. This last dimension is of special importance, 
since if it be made too small the discharge from the tube spreads 
into the receiver and produces fog, and if it be made too large 
intensity of illumination is sacrificed. 
If the tube is to be separated from the receiver by a window and 
is to be separately exhausted, a special form is used. (Fig. 1.) Here 
the end of the internal capillary is brought as near the fluorite window as may be 
without undue heating. A device of this type not only brings the ource of light near 
to the slit of the spectroscope but reduces the absorption in the tube itself to a 
Fig. 1- 
min- 
imum. 
The 
last advantage is a most important one in dealing with gases such as 
air which absorb strongly. 
The electrodes in both forms of tube were usually of alu- 
minum, but iron and copper have also been tried. 
9 
