THE ABSORFnON AND EMISSION OF All;. 29 



account of the bauds of carbon monoxide which tiius far have made up the whole. 

 Nor is there ground for confident hope that better success will be met with in the 

 sequel. For with low pressure, where success was most confidently to be antic- 

 ipated, these bands appear irrepressibly in their full force; and with higher pres- 

 sure the absoi-ptiou, as I shall presently show, is so marked that the attempt to get 

 an extended air-spectrum in this way promises little. The increased absorption 

 which follows higher pressure might, indeed, be avoided by reducing the thickness 

 of the stratum ; but in that case the window of the tube would be so near the 

 capillary orifice that in a very short time the deposit 'upon it would produce com- 

 plete opacity. 



Information regarding the transparency of air is easier to obtain. All the old 

 investigations upon this subject were limited to strata of considerable thickness, but 

 I have extended my observations to very thin layers. This was suggested by the 

 fact, long ago repeatedly ascertained by me, that the rays beyond 185 ^n were re- 

 markably weakened even by strata less than 1 cm. thick. Since then, notable data 

 concerning the behavior of thin layers of air have come to light, and since I have 

 occupied myself more with this mattei' than with others, it is proper that I should 

 here confine myself to this. 



The above desci'ibed absorption-apparatus. Fig. 10, was used for the production 

 of the thin strata of air. From the negatives which I obtained with it, I take two 

 and give photogi'apliic reproductions of them magnified fourfold. 



In one plate (No. 5941), A, Plate IV, the thickness of the stratum of air was 

 15 mm., 14 mm., ... 3 mm., 2 mm., 1 mm., 0.5 mm., 0.25 mm., 0.10 mm.; while 

 the time of exposure was in every case 1 minute, and the width of the slit 0.020 mm. 

 The Geissler tube was filled with dry air at a pressure of \ mm., and before photo- 

 gi'aphing, electric discharges were passed through it for some time. Without this 

 precaution the source of light would be wanting in the requisite constancy. The 

 fir.st eight spectra, with thicknesses of 15 mm. to 8 mm., end, without any material 

 gain in length, at the wave-length 178 iajj. The exposures following extend beyond 

 that wave-length ; yet the stivatum of thickness 4 mm. first allows the band at 170 /<// 

 to appear. From that point the gain in length of spectrum with diminution of 

 thickness of the stratum is more rapid ; and with a stratum of 0.5 mm. the spectrum 

 runs to the end of the plate, corresponding to wave-length 163 yuy". The last two 

 strata of the series of thicknesses 0.25 mm. and 0.10 mm. give spectra of consider- 

 ably increasing intensity, and the inference that the photographed spectra would 

 be longer is confirmed by C, Plate IV. These two specti'a show more than any 

 of the foregoing how energetically the air absorbs these rays. For better orienta- 

 tion, I add to A, Plate IV, a small series of spectra (No. 6066), B, of \vhich the 

 uppermost band-spectrum is similai- to the uppermost of A, while beneath the 

 banded spectra appear the most I'efrangible lines of the spectrum of aluminum, 

 at 1989.90, 1935.29, 1862.20, and 1854.09 Angstrom units. 



Plainly, a longer exposure would extend the photographic spectra to shorter 

 wave-lengths; and this is confirmed by another plate (No. 5935), C, where the 

 exposure was for two minutes, the width of slit 0.020 mm. as before, and the 



