Trowbridge and Sabine — Steam in Spectrum Analysis. 115 



impinge directly upon the spark. No effect could be perceived 

 when there were no condensers in the secondary circuit, and 

 with the introduction of small condensers the effect was not 

 marked ; but when the number of Leyden jar condensers, was 

 increased to four the effect of the jet of steam upon the elec- 

 tric spark was surprising. . Its light immediately became com- 

 parable with that of the electric arc, enabling us to see the 

 metallic spectra with the naked eye upon the ground glass of 

 the photographic camera without the use of an eye piece. The 

 chamber in which the spark and steam jet were placed became 

 rosy red from the hydrogen arising from the dissociation of 

 the steam. The hydrogen and oxygen lines in the air spectra 

 became very much strengthened, a continuous spectrum showed 

 itself in the neighborhood of the C line and also in the yellow, 

 and a photograph of the air line and metallic line of the ter- 

 minals employed could be taken in a third of the time which 

 was necessary when the steam jet was not employed. 



The apparatus consists merely of a tin box which is placed 

 opposite the slit of the spectroscope. Steam enters at one side 

 and is blown across the terminals of the Ruhmkorff coil which 

 are placed in the box opposite the slit, an outlet on the side op- 

 posite from the place of entrance of the steam allows the waste 

 steam to escape into the outer air. 



The change of color of the spark is undoubtedly due to hy- 

 drogen. The light filling the box above referred to is decidedly 

 red, and the hydrogen line C flashes out with great brilliancy 

 in the midst of a continuous band of red in the spectrum. 

 The metallic line from the terminals are greatly strengthened. 

 The light from iron terminals is especially brilliant. Without 

 the steam the spark between iron terminals seemed to consist 

 of a single line of discharge. When the steam was turned on 

 a great bundle of sparks appeared in the midst of a flaring 

 light and the noise of the spark was greatly increased. This 

 effect can undoubtedly be traced to increased conduction of the 

 air space between the terminals of the Ruhmkorff coil. 



The appearance of the spectra led us to examine the ques- 

 tion of the spectrum of the Aurora Borealis and its connection 

 with that of aqueous vapor. We believe that the theory that 

 the shifting nature of the northern lights may be due to elec- 

 trical discharges following strata of air more or less laden with 

 aqueous vapor has been advocated. The appearance of the 

 speetra of the electric spark in steam certainly leads one at first 

 to favor this hypothesis. We have spoken of the marked brill- 

 iancy of the hydrogen line and of a continuous red band near 

 this line. The continuous spectrum in the yellow is no less 

 prominent. The observations which have been made on the 

 northern lights do not enable one to make exact comparisons. 



