214 TROWBRIDGE. — HIGH ELECTROMOTIVE FORCE 
I 
the II and K lines of the solar spectrum. It did not seem possible that these lines 
could be due to calcium, for I obtained them in quartz tubes. 
lave, however, made a more extended examination of the possible origin of these 
lines from the calcium contained in glass, using quartz tubes, and although I occa- 
ionaily obtained these lines with quartz tubes, I am now persuaded that these lines 
are due to calcium; for using the utmost care in luting pure silver terminals into 
quartz tubes, the luting being at a great distance from the point of discharge I 
found difficulty in obtaining the lines, and in the majority both lines did 
There was, however, in the case of quartz tub 
appear 
th one of 
cium lines and nearly coincident with the II line of the solar spectrum. It seems 
dent that in the immediate region of the H and K lines, perhaps superimposed 
At one time I thought that oxvo-pti mioht unntmh 
o " A Jo tu ,J,1 o 
this region, but I am less sure of this from subseq _ 
It is evident, I think, that the characteristic spectrum show n7n° Fig ur^ 9, Plate 
XXVII 
for it appears in qua 
very powerful discharges in rarefied oxygen, hydrogen, and 
The region, however, in the immediate nmVlihn 
as in tdass tubes 
hood, or included in the broad 
of the H and K lines, is a composite one, and must be investigated „, . 
of greater dispersion than that I have employed. The one I have used 
by 
space of approximately 4 mm. b 
& 
Fig 
10 shows the change in the spectrum of rarefied 
of disc hare: 
quartz tub 
diminish in number, and the final 
spec run, . characterized principally by two very broad band,. 
from alUr" 1 ^^^ '* U ' e GeisS ' er tubes were Ponced by discharges 
T d s IT T and Varied b W " nber fr ° m ° ne di «*»'ge to six discharges. 
s: sa =ti: ^z^-t:: (F t ? PIate xxv h 
tnest trom the solar spectrum. It 
evident from the studv nf +1.« r u — * *ruiu me soiar spectrum, i 
dy weak dis, , *»***»■ that Certain ]i » es W*» with compara 
jr weak discharges winch disappear with stronger discharges 
*i<r Ure 9 show* ti, Q «„„i.. f ... ° uiscoarges. 
o of a silicon line, 2881, toward the red end of the 
gure y snows the fl 
pectrum. This fkring toward th*, ^ , ' *"* ' red en 
i~ •** „ „C££E: s= :r"* ~ - - ■— 
1. Hydrogen lines do not 
RESULTS. 
cWge. which gU* or quartz ^,rr b " ^ the -ost powerful electric dis- 
withstand. 
