554 



SCIENCE 



[N. S. Vol. LIV. No. 1405. 



appeared on the film, the second order being 

 superimposed on the first order comparison 

 spectrum. By the use of the foregoing method, 

 an estensive spectrum was obtained with 

 oxygen in the receiver and is attributed to 

 that gas. A spectrum was also found for 

 air. (See table.) 



TABLE 



Explanation of Symiols of Taile: \ indicates 

 doublet; } t triplet; | 1 doublet and probably 

 triplet; d diffuse; 2 violet edge of band; s con- 

 tinuous maximum. 



In almost every instance the wave-lengths 

 given in the table are tlie averages of two 

 or more plates, hence judging from the con- 

 sistency of the several measurements they 

 are believed to be accurate to about 0.5 

 Angstrom. The first column contains the 

 lines that occur in both air and oxygen. 

 Column 2 gives the lines that were not regis- 

 tered on the films of air spectra, but were on 

 those of the spectrum of oxygen. They are 

 faint. Some of the more intense lines that 

 occur in air only are indicated in column 3; 

 the fainter lines, of which there are about 

 thirty, were omitted from the list. Column 



4 is a record of the wave lengths produced in 

 oxygen with direct current discharge. The 

 spectra listed in columns 1, 2 and 3 were ob- 

 tained with disruptive discharge. In column 

 4, lines in the Schumann region are included; 

 similar spectra were also present in the cases 

 of disruptive discharge but were omitted 

 from the tabulations. 



The other columns are lists of the rela- 

 tive intensities of the wave-lengths in the 

 columns immediately preceding. Wliere two 

 values of intensity are given in the same 

 column, the first refers to the spectrum with 

 oxygen and the second with air. 



It is worthy of note that the line A1215.9 

 is very strong in the spectrum of oxygen and 

 air even when a direct current was used. 

 This wave length is very near to A1215.6, 

 the fundamental line in the hydrogen spec- 

 trum, and probably is that line. This was 

 found to be present in most of the spectra 

 obtained by Millikan in his investigations on 

 the spectra of metals. The transparency of 

 oxygen and air (1340-916 for air and 1336- 

 990 for oxygen) in this region is proved from 

 the fact that these spectrograms were ob- 

 tained. It is evident that the absorbing 

 layer of gas in these experiments amounted to 

 more than 0.5 mm. at atmospheric pressure, 

 and judging from the intensities of the spec- 

 tra, these gases are transparent in layers of 

 even much greater thickness. The films of 

 the spectrum of air were badly fogged, and 

 in some cases the entire spectrum appeared 

 reversed. However, since other films of this 

 spectrum were obtained without this reversal, 

 it is believed to be of a chemical nature and 

 due to the corrosive gases formed by the 

 radiation or the discharge. This point will 

 be investigated more thoroughly in the near 

 future. The work of getting the spectra of 

 electrolytic oxygen and pure nitrogen is now 

 on the way, and the thorough search for 

 series lines and for ionization and resonance 

 potential relations is postponed until this 

 new data is available. 



j. j. hopfield 

 dlpabtuent of physics, 

 University of California 



