|26 LYMAN. — THE SPECTRUM OF HYDROGEN. 
spectrum, in fact it is to its brilliancy that the success of the research is due. In 
spite of the feebleness of the other spectra, however, it has been found possible to 
obtain many of the stronger lines between X1550 and X1250 in the second spectrum. 
Their measurement therefore forms a valuable check on the numbers obtained by 
the two-slit method. 
The elimination of the lines due to impurities from the spectrum of hydrogen 
necessitates the study of the spectrum of air. As has been set forth in the earlier 
paper it is found most convenient to fill the spectroscope itself with pure hydrogen ; 
in f act if the lines of the shortest wave-length are to be obtained the light-path must 
be entirely in this medium. No window between discharge tube and spectroscope is 
permissible. When, however, the spectrum of a gas other than hydrogen is to be 
studied a window of fluorite must separate the discharge tube from the spectroscope. 
The extent of the spectrum is limited, therefore, by the transparency of colorless 
lluorite and the absorption of this substance has formed a necessary part of this 
research. As a matter of fact even fluorite of the best quality was found to absorb 
all light below wave-length X 1200 ; the study of the spectra of gases other than 
hydrogen therefore terminates with this value. 
In view of the fact that Schumann made use of two fluorite lenses and a fluorite 
prism it seemed extremely probable that his spectrum does not extend below wave- 
length X 1200. To test the matter the plates published in the Smithsonian Contribu- 
tions to Knowledge No. 1413, have been compared with the normal spectrum obtained 
during this research, and it has been found possible to identify a great majority of 
the hydrogen lines in this prism spectrum with lines measured by the author. Two 
important results follow. First, a scale of wave-lengths has been attached to the 
Schumann spectrum, as shown in the half-tone reproductions at the end of this 
memoir. 
Second, as the line of lowest wave-length visible in Schumann's plat 
the value X 1267, the present limit X1030 establishes a considerable 
spectrum 
of 
Since the effect of change in the electrical conditions under which a spectrum is 
produced is extremely important, the question of the existence of a secondary spectrum 
trumf 0g6n m the regl ° n ° f Sh0rt wave - len S ths ha * ^en examined. No such spec- 
X ^OOO^ndT 1200 eX1St ; that ^ t0 Bay ' th6re ^ bUt ° ne hydr °2 en s P ectrum between 
para?ltl l0Wing **** ^^ * detailed accounfc of the work of which the foregoing 
paragraphs may serve as an outline 
