TROWBRIDGE. — HIGH ELECTROMOTIVE FORCE. 21.*> 
spark discharge, the pressure 
the light was a glow or luminescence and not a white 
in the tube being from one to two centimeters. 
A Rowland grating was employed, and an eyepiece was fixed on the C line <>f 
hydrogen. The second spark gap gave a fine bright line of the apparent length 
of the slit, the capillary a continuous spectrum, and where the line bright line crossed 
this continuous spectrum, it was reversed. 
Kirchhoffs law of radiation thus applies to the radiation in Geissler tubes, and 
Pringsheim's contention is justified. If the solar corona is an electrical phenomenon 
of the nature of luminescence it can exhibit either bright lines or dark lines according 
as it is hotter or colder than the background. 
In this study of the upper limit of temperature which one can reach by electrio 
discharges through rarefied gases, we perceive that spectrum analysis is one of the 
most difficult analyses which modern science has revealed. There are a few broad 
facts, such as Doppler's principle and the reversal of spectral lines according to Kirch- 
« 
hoffs law; on the other hand there is ionization, dissociation, adsorption, and absorp- 
tion, all modified by the glass or quartz vessels which must be employed. 
M. Cantor calls attention to the fact that Ilittorf failed also to observe reversals of 
spectral lines in the case of electric discharges in Geisder tubes. Ilittorf speaks of 
a first series of hydrogen lines which are seen with feeble discharges. This feeble 
spectrum with its bands seems to be a peculiarly luminescent effect in which any 
translatory or colliding effect of the molecules is a minimum. The new theories in 
regard to the composite nature of the atom seem to demand an extension of our 
views in regard to the nature of the light emitted by atoms and their aggn gates 
under the stimulus of an electric discharge. The phosphorescent and fluorescent 
light of a gas under this stimulus may arise from the mechanism of the atom, and 
therefore may not give sensible heat. The combination of atoms into molecules, and 
their dissociation and formation of new combinations, may give the spectra we usually 
observe under the effect of fairly strong electric discharges, and provide the sensible 
heat which can be measured by the bolometer or the thermal junction. 
Spectrum analysis of the future thus becomes more and more difficult of applica- 
tion, and one of its most important fields is in the study of phosphorescent and fluores- 
cent light emitted by gases. We seem to be on the point of regarding the light and 
heat of the sun more from the electrical standpoint. And the study of discharges 
of electricity in rarefied gases assumes a great importance. 
My attention was first directed to the phenomenon of reversals in Geissler tubes 
filled with rarefied gases by the reversal of two lines apparently closely coinciding with 
