122. W. Huggins—Spectrum of the Flame of Hydrogen. 
arranged in very close pairs. There is a pair of fine, but very 
3100 3200 
it rer errs 
IM 
ae ee 
Spectrum of Water. 
I then introduced oxygen into the flame, leaving a small ex- 
cess of hydrogen. A spectrum in all respects similar came out 
upon the plate. I repeated the experiment, taking both spectra 
on the same plate. Through one-half of the slit the spectrum 
of the oxyhydrogen flame was taken. This flame was about 
seven inches long, and the spectrum taken of a part of the 
flame twoinches from the jet. The oxygen was then turned 
off, and the quantity of hydrogen allowed to remain unaltered. 
A second spectrum with an exposure of the same duration was 
then taken through the second half of the slit. On the plate the 
two spectra are in every respect similar, and have so exactly 
the same intensity, that they appear as one broad spectrum. 
these experiments a platinum jet which had been 
carefully cleaned was use 
n these experiments the two gases met within the blowpipe 
and issued in a mixed state. 
The jet was removed, and a flame of hydrogen was sur- 
rounded with oxygen. The spectrum (No. 2) shows some addi- 
tional lines, In this case the jet was brass, and ix this or some 
other way impurities may have been introduced ; and I should, 
at present, incline to the view that the additional lines about 
43429 and 43473, and the groups more refrangible than 4 3062, 
do not belong to the water spectrum, but to impurities. 
oal-gas was substituted for hydrogen in the oxyhydrogen 
blowpipe, and oxygen admitted in as large a proportion as pos- 
sible. The inner blue flame rising about two inches above the 
jet showed in the visible part of the spectrum the usual “ five- 
fingered spectrum.” The light from this part of the flame 
was projected upon the slit. The ce (No. 3), contains 
the water groupalready described, and in addition a very strong 
line close to G, and two lines, A 3872 and A 3890; this latter 
