30 



THE ABSORPTION AND EMISSION OK AIR. 



tliii'knessea of .strata of air, 0.05 mm., 0.1 mm., 0.2 mm., 0.4 mm., O.G ram., 0.8 mm., 

 1.0 mm. Unfortunately, the wave-lengths bejond lOU n-^i being unknown, I am 

 not in condition to state exactly tlie limits of these spectra. But it will suffice to 

 compai-e their lengths to show how enormously thin strata of air absorb lays more 

 refrangible than those of wave-length 160/<//. This fact is mo.st evident in the 

 comparison of the effects of strata of thicknesses of 0.1 mm. and 0.2 mm.; for this 

 difference corresponds to a considerable dift"erence in the extensicm of the spectra. 

 Even the reduction of the thickness from 0.1 mm. to 0.05 mm. results in a very 

 perceptible increase of the energy. It is true that this is not very well shown by 

 6' because the most lefraiigible lines, for which the increase of energy is most 

 marked, have not sufficient intensity to bear the fourfold magnification, so that they 

 disappear entirely from C. I Vjelieve, howevei", that I shall not exaggerate if I say 

 that the increase of the thickness of the stratum from 0.05 mm. to 0.1 mm. involves 

 a loss of 50 <o of the energy. 



The spectrum of the Geissler tube tilled with air, owing to the discontinuity 

 of its spectrum, never shows the absoiptiou-bands near 185 /</< due to the oxygen, 

 which appear so clearly when the tube is filled with hydi-ogen. 



Considering the value which an acquaintance with continuous spectra has for ab- 

 sorption-specti'oscopy, and in view of the difficulty of searching out suitable sources 

 of lifht, let me refer to a discharge-tui)e with which I made many emission-e.vperi- 

 ments in 188G. It was an end-on tul)t' witli an uncommonly narrow capillary — 

 about \ or ^ mm. wide. If such a tube is exhausted until no discharge will pass 

 through it, or even until it fluoresces strongly, then as soon as a Leyden jar and a 

 spark-gap are inserted in the secondary circuit of a powerful Ruhmkorff coil, it 

 gives an uncommonly blight light of a white color. If the spectrum of this longi- 

 tudinally emergent light is photogi-aphed with cori-ect arrangements, one obtains, 

 instead of a line-s[)ectrum, a continuous spectrum in which, at most, a few traces 

 of washed-out lines are I'emarked. Sometimes the continuous spectrum enters only 

 as a stripe, several milliraetei's broad, running the length of the line-spectrum. In 

 both cases this continuous spectrum is well adapted to absorption-experiments in 

 which occur lines and especially groups of lines like tho.se of oxygen. The })hoto- 

 gi'aphie action of this capillary light is very powerful. T^nha{)pily these tubes are 

 of shorter life than those of the usual caliber. The spark widens the narrow oi)eu- 

 ing by degrees, and with the widening the photograjihic energy diminishes, and 

 worst of all, it then ceases to be so easy to convert the line-spectrum into a continu- 

 ous spectrum. 



I.Kii'zir,, .T:imiary 19, 1001. 



