L 139 ] 



XYI. On the Use of Steam in Spectrum Analysis. 

 By John Trowbkidge and W. C. Sabine*. 



AMONG the difficulties with which the investigator in the 

 subject of spectrum analysis must contend is that of 

 obtaining a source of light which is free from constituents 

 other than those which are under examination, and is at the 

 same time sufficiently powerful to enable one to photograph 

 the spectra of the latter. The voltaic arc gives a sufficiently 

 strong light to enable one to photograph throughout the visible 

 spectrum ; but the carbons between which the arc is formed 

 are full of impurities, and it is difficult to interpret the spectra 

 obtained by this means ; moreover it is not easy to employ 

 the arc-light for researches in the ultra-violet portion of the 

 spectrum. On the other hand, the spark from a Ruhmkorff 

 coil taken between terminals of metals the spectra of which 

 one wishes to examine, gives us, in general, spectra which are 

 comparatively free from impurities; but its light is very 

 feeble compared with that of the electric arc ; and even when 

 the spark is obtained by means of a powerful coil excited by 

 an alternating dynamo machine, an hour is necessary to obtain, 

 with a concave grating of 21 feet radius of curvature, a pho- 

 tograph on the most sensitive plate we can obtain of the 

 ultra-violet spectrum of copper at wave-length 2100. 



It becomes an important question, therefore, to ascertain 

 whether the time of exposure of the sensitive plate can be 

 shortened by any process; for the outlay in obtaining one 

 photograph in the ultra-violet by the means hitherto at our 

 command is very large, involving the running of an engine 

 of at least two-horse power for an hour. In our experiments 

 with a jet of steam, we find that the time of exposure of the 

 sensitive plate can be shortened to at least one third. 



We were led to employ steam for the purpose of obtaining 

 the spectra of oxygen and hydrogen with a more powerful 

 electrical excitation than is possible in Geissler tubes. During 

 the winter of 1886 one of us_, in connexion with Mr. 0. 0. 

 Hutchins, of Bowdoin College, while engaged upon the inves- 

 tigation of oxygen-spectra, tried to obtain a powerful electric 

 spark in an atmosphere of steam ; but the experiments were 

 unsatisfactory. The difficulties were chiefly in the way of 

 defective insulation of the terminals of the Ruhmkorff coil. 

 Our experiments showed that no containing-vessel could be 

 employed ; for the sides of the vessel conducted the electrical 



* Communicated by Prof, John Trowbridge. 



