March 14, 18S9] 



NA TURK 



475 



an indigo-violet tint than ordinary so-called sky-blue. There is 

 nothing more beautiful in Nature than the blueness of the 

 heavens. 



The limitation of the solar spectrum has been the subject of 

 elaborate investigation by M. Cornu.^ He proved by direct ex- 

 periment that the ultra violet rays are absorbed vfc'ilh energy by 

 the atmo-phere, and showed that there is a variation in the 

 amount of absorption corresponding with different altitudes, so 

 that the absorbent matter is at each elevation proportional to 

 the barometric pressure, and consequently in constant relation to 

 the mass of the atmosphere. This fact alone is sufficient to ex- 

 clude water-vapour fiom consideration as being the medium of 

 Misorption. Moreover, water- vapour, while it absorb < the red 

 and infra-red rays, transmits the ultra-violet very completely. 



Photographs taken in 1879 on the Riflfelbcrg, at an altitude 

 therefore of 8414 feet, reached to wave-length 2932 ; but at Viege, 

 an altitude of 2165 feet, to only 2954. 



In short, it was shown that within the limits of altitude at 

 which Cornu's observations were made there was a difference of 

 10 tenth-metrets of wave-length for every 984 feet of dry atmo- | 

 sphere, the short enin;^ of the spectrum being due to the gaseous 1 

 constituents. Notwithstanding this, it was stated by Prof. Live- 

 ing, in a lecture delivered at the Royal Institution on March 9, 

 1883, that we must suppose the absorbent substance, whatever 

 it may be, is not in our atmosphere, because, when observations 

 are made upon the summit of an elevation like the RifTelberg, 

 the lengthening of the .spectrum reaches only a very trifle be- 

 yond U. "The same reason will lead us to reject the notion 

 that the absorption can be due to matter in interplanetary space, 

 for it is not easy to suppose that the gases which pervade that space 

 in extreme tenuity can differ much from those in our atmo- 

 sphere, because the earth in i s annual course mu.-t pick up | 

 vwhatever they are, and they must then diffuse into our atmo- i 

 sphere. The absorption is, therefore, probably neither in our , 

 atmosphere nor in interplanetary space, and we must look for | 

 it in the solar atmosphere." | 



It is then suggested that the blotting out of the sun's light \ 

 beyond U is caused by something in the solar atmosphere higher I 

 up than the metallic vapours which give rise to Fraunhofer's lines, ; 

 but at a lower temperature. It has been shown by Young that j 

 the Fraunhofer lines are bright lines, but appear black against 

 the brilliant light of the photo-phere. It never appeared to me i 

 that Prof. Liveing's objection to Cornu's explanation of the I 

 limitation of the solar spectrum by our atmosphere was valid, \ 

 because it was proved beyond question that the atmosphere 

 absorbs the ultra-violet rays, and also that on the same day and 

 hour, at different elevations, the extent of rays absorbed was pro- ! 

 p^rtional to the barometric pressure — that is, to the quantity of air 

 through which the rays ipas^ed. A considerable acquaintance 

 with absorption-spectra in the ultra-violet region has proved to 

 me that when an absorption band has been blotted out by in- ! 

 creasing the proportion of substance, or by increase of the thick- 

 ness of the absorbent layer, a stage is soon reached at which any 

 further increase only causes a trifling difference in absorbent 

 action, and in fact that many substances attain a maximum of 

 absorption beyond which there is no change unless we increase 

 the density of the substance, and so probably a'ter its molecular 

 structure. Under the same conditions of pressure, increased , 

 thickness of the absorbent layer only very slightly increases the ' 

 absorbent action, and that in a degree which is by no means 

 proportional to the layer of material. 



There is a difficulty in accepting Prof. L'veing's views, because ^ 

 we know nothing, as he remarks, in the solar atmosphere capable , 

 of causing such absorption, and at the same time of transmit- j 

 ting the Fraunhofer lines of the less refrangible portions of the 

 spectrum in the condition in which we observe them. 



The matter was very fully considered by me two years pre- 

 viously—that is to say, in the year 1881. The absorption spectra 

 of various ga-es were exammed by photographing the ultr.i- violet 

 rays which were transmitted by carefully measured quantities of 

 gas at the atmospheric pressure, and one of these gases was 

 ozone. It proved to be a substance with most extraordinary 

 absorptive powers, so that even when very much diluted it 

 exhibited an absorption-band of great intensity which was 

 carefully investigated. 15y examining the spectrum transmitted 

 by increasing quantities of ozone the band disappeared, and 

 there was a complete and total absorption of rays extending to 

 about wave length 3160. Any further increase did not cause a 

 « " Sur I'Absorption Atmosp'i<£riine des Radiations ulira-violef.es/'/o//;-*/. 

 de Physique, t. .\. i38i. 



corresponding shortening of the spectrum. The band was- 

 observed between wave-lengths about 2850 and 2320 ; but with 

 increased proportions of ozone the rays transniitt*;d were re- 

 stricted to about 2920, and became more restricted in presence- 

 of greater quantities of gas. ^ Thus : 



Actu.il volume of ozone 



of o lumn. 

 Cubic cent. metres. 

 I 01 2 

 2-162 

 3-175 



„ » • J Length of C( lumn 



E.xtreme ray transmuted ^^ ^^ traversed 



by ozjn.zed o.'cygen. .^^ ^^^ ^^^^ 



\. Centimetres. 



3035 ••■ 92 



3150 ... 1965 



3160 ... 2885 



It was found that a quantity of ozone proportional to the 

 average quantity present in a vertical column of the atmosphere 

 caused an absorption similar to that observed in the solar spec- 

 trum — that is to say, terminating about 2950. Largely incre ised 

 quantities did not largely, but only in a trifling degree, increase 

 the absorption. Furthermore, it was shown that the atmo- 

 sphere contained ozone as a normal constituent, and that it was- 

 present in greater proportion in the upper regions than near the 

 earth's surface. It was proved that all the other minute con- 

 stituents of the atmosphere were either non absorbent or exerted 

 absorption in a manner different from that of ozone, and that the- 

 qnantity of ozone commonly present in the atmosphere i> quite 

 sufficient to account for the limitation of the solar spectrum, 

 without taking into account the possible absorption caused by 

 the great thickness of oxygen and of nitrogen. The possibility 

 of oxygen being the absorptive substance seemed very great, 

 considering the small difference in constitution between the 

 molecules of ozone and oxygen. It must be understood that 

 the conclusions were arrived at by reasoning from strictly quanti- 

 lative experiments, and seemed almost incontrovertible, and it 

 may be stated that none of the facts alleged have ever beei> 

 questioned. It was impossible to deal with the matter further 

 without costly appliances for the compression of oxygen and- 

 nitrogen into tubes capable of holding a quantity comparable 

 with the amount of oxygen and of nitrogen in a vertical column 

 of the atmosphere, and for this reason the investigation fell into 

 abeyance. 



It was al-^o considered that the problem might be attacked in. 

 another manner. 



Messrs. Liveing and Dewar have rtcently made a very inter- 

 esting and important communication to the Chemical Nen's tvol. 

 Iviii. p. 163), on the absorption-spectrum of oxygen. In a tube 

 1-6 metre in length, filled with the gas at a pressure of i6o^- 

 atmospheres, all rays were absorbed beyond wave-length 2665, 

 but they began to diminish at 2705. With a tube 6 metres or 

 20 feet long, and with a pressure of 90 atmospheres, it seems 

 that an absorption-band is to be traced at wave-length 3640 to- 

 3600, and there is a complete absorption beyond 3360. 



The gas seen in quantity corresponding to that in a vertical 

 column of the atmosphere appears to have a faint blue tint. 

 There can be no doubt whatever that the oxygen of the air exerts 

 a powerful absorption on the rays of the sun, but it does not 

 appear from these experiments that this absorption is exactly 

 the cause of the limitation of the spectrum, as described 

 by Cornu, since when observed in tubes it is carried into 

 a region of longer wave-length than is observed at the 

 level of the sea ; thus at Dublin the limit in summer is 

 usually about 3130."- It is no doubt the high density of 

 the gas which causes the absorption to be stronger than that of 

 the atinosphere. It should be noted that the oxygen in the 

 20-foot tube was the quantity in a vertical column of the atmo- 

 sphere. It is probable that there are several substances in 

 interplanetary space, or in the solar atmosphere, which, besides 

 oxygen and ozone in the air, cause an absorption of the sun's 

 rays and a limitation of the length of the spectrum, but as- 

 Messrs. Liveing and Dewar point out, our atmosphere places a 

 limit to the observations we can make on the rays of other 

 heavenly bodies. 



Touching the colour of the sky, Prof. Tyndall has told us 

 that four centuries ago it was believed that the floating particles- 

 in the atmosphere render it a turbid medium through which we 



• " On the Absorption-Specirum of Ozone, and o" the Ab<orption of Solar 

 Rays by Atmospheric Ozone," Journ. Chem. Soc. i88i, xxxix., pp. 57, iii, 



^ In the rhil. Mag., September i3S8, p. 288, Messrs. Liveing and Dewar 

 refer only to my first paper on the .absorption-spectrum of ozone, Journ. 

 Chera Soc, xxxix, p. 57, but not to the more complete paper on this 

 spectrum at pp. 111-119, loc. cit., xvh'ch indicates the possible limits of the- 

 oz ne in the atmo.Lhere. 



