14 TfrK ARSORPTION AND KMI8SION OF All:, 



been chemically or electrolytically pnuliiceil. After being carefully imritied. they 

 were dried with i)hosphorus-j)entoxide and in some cases with snl|ihuric acid. 



The pre-SHKre in the absor|ition-tubes was e(|iial to the atmospheiic pressure at 

 tlie time. In the emission-tubes, it varit-d from (>.i>o to one centimeter of mercury- 

 pressure. 



The electric discharge was effected either by the Kuhmkorff coil alone, oi- with 

 a Leyden jar, or Leyden jai- and spark gap interposed into the circuit, or a Leydeu 

 jar and a Ilemsalech self-induction coil. The piimaiy current never exceeded 2.5 

 amperes and 10.5 volts. 



Dlffioilties. — Before I touch upon the results thus far obtained, I must allude 

 to a phenomenon which I encountered as soou as I took my first emission-spec- 

 trograms. The fact substantially was this. I photographed the spectrum of a 

 Gei.ssler tube filled with air, and obtained a spectrum of considerable )thotogra[)hic 

 effect compounded of numerous thickly successive bands shading off on the red 

 side. Numerous repetitions of the ex[)eriment confirmed the result. I concluded 

 from this that I was dealing with the S[)ectrura of aii- or, at any rate, of one or more 

 of its ingredients. Rut when I came to examine the latter separately, as I had 

 dune the aii', I found that the same supposed air-bands dominated every one of them. 

 With carbon monoxide and caibon dioxide, however, they were much stronger; 

 while with water and hydrogen they were weaker. After compaiing these photo- 

 graphs with one another, no loom for doubt seemed to remain that these bands be- 

 longed to a carbon compound which presumably had been derived from the fat of 

 the air-pump cocks and fioin the occluded and aksorbed gases of the inner wall 

 of the tube. However, increased can- in cleaning the tubes and in the proihiction 

 of the gases to fill them, from which 1 piomised myself better success, made no 

 dift'erence; the bands remained as strong as before. Now, in order to keep the 

 air-pump vapors of tlie cock-fat and mercury completely away from the place of 

 discharge, I exchanged the Geissler tubes, which had up to that time been em- 

 ployed, for l'aalzow-^^)gel tubes. By way of greatei' precaution, these tubes, 

 Ijefore being used, were cleaned for twelve hours with a solution of bichromate of 

 jiotassium in concentrated sulphuric acid, to destroy any impurities of organic 

 origin which might be present, and were then repeatedly rinsed out with distilled 

 water, and were forthwith closed with a quaitz w indow and water-glass, and dried 

 by tlie air-pump, though the two sulphuric-acid valves of the tube itself had in the 

 main effected this. In this way, the penetration into the tubes of particles of dust, 

 which might have occasioned the formation of carbon compounds, was as far as 

 jiossible prevented. By theieupon washing out the place of discharge with oxygen 

 which was evolved from the sulidiuric acid of the electrolytic apparatus belonging 

 to the tulte, I promised myself, judging by the visible bands of carbon compounds, 

 the sure attainment of my object. Still, these means too proved without effect. 

 Thereupon the suggestion presented itself that occluded gases from the aluminum 

 electrodes miifht have caused the bands, and the electrodes were laid aside and re- 

 placed by glass cylinders covered with tin-foil. But in .spite of all, the bands after 

 this alteration remained intact. The tul)es were now furnished with new aluminum 



