THE ABSORPTION AND EMISSION OF AIR. 13 



useful with quartz, but did not work well with Hiior-spar. In this case, the layer 

 of luting loosens itself, usually in a short time, from the fluor-spar and ci-nnibles 

 away. Of course, this thi-ows doubts upon the security of the closure. It is 

 possible that a luuch thicker coat than that I used would answer bettei-. I have 

 not, however, tried it, but have employed as luting for the fluor-spar a mixture of 

 white wax and clock-maker's oil, which is spread in a very thin layer forming a 

 band half a millimetei- broad, in a ring round the end of the tube. Now pi'essing 

 tlowu the fluor-spar disk a neat and pei'fect closure is obtained ; and it is loosened 

 even easier than the watei'-glass. It has, however, the inconvenience of soiling the 

 gas in the tube after long illumination, by producing fatty vapors. 



The source of light for the absorption experiments. — The best thing for this 

 purpose would be a Geissler tube which should give a continuou.s spectrum of 

 e(pial intensity of energy in all parts. Unfortunately, no single tube meets this 

 lequirement completely. The hydrogen tube, however, does so in part, giving 

 a continuous spectrum down to 170 l■*^^, where begins a I'ich line spectrum of toler- 

 ably uniform distribution of energy. Such a tube closed with fluor-spar has given 

 me good service in my absorption experiments, particularly by its continuous spec- 

 trum. A merit of it not to be undervalued when long exposures are required, as 

 they sometimes are, is that it has an unconunon photographic l)rightness. For 

 comparison spectra I have employed the spark-spectra of hydi'ogen, aluminum, and 

 cadmium, which were photographed beside the absoi-ption-spectra. 



The source of liglit for the emission-spectra. — These sjjcctra were given by 

 Geissler tubes of different shapes and ari'angements, closed with (juartz. Some of 

 these tubes, after being filled, remain in connection with the air-i)ump, so that they 

 can readily be emptied and refilled. Others are melted off like ordinary Geissler 

 tubes, before they are attached to the spectrograph. For some investigations, how- 

 ever, none of these tubes will answer. In such cases, Paalzow-Vogel tubes of the 

 Paschen-Euuge pattern were prefei'red. In these, the place for the dischai'ge com- 

 municates at both ends through licpiid valves, of which one consists of sulphuric 

 acid and the other of a solution of bichromate mixed with sulphuric acid, with 

 which the air-pum[) is so connected that the pumping out of the tubes can go on 

 unhindered, while the vapors from the mercury and from the grease of the pump- 

 cocks are kept away from it. At the same time, the sulphuric-acid valve of such a 

 tube, of which the receptacle is arranged as an apparatus for evolving hydrogen and 

 oxygen electrolytically, serves to produce the oxygen and hydrogen to fill the tube. 



The filling of the tubes has hitherto been limited to 



Nitrogen 



Oxygen 



Water 



Carbon monoxide 



Carbon dioxide 



Hydrogen. 



These erases have been in part obtained from the atmosphere, and in part have 



