224 



Profs. G. D. Liveing and J. Dewar. [June 6, 



ultra-violet rays does not extend quite so far down as the limit of the 

 solar spectrum, though it approaches that limit. There is a diffuse 

 edge of gradually diminishing absorption succeeding the complete 

 absorption, and this fact, together with the rapid diminution of the 

 extent of the complete absorption with decrease of pressure, lead us 

 to class this absorption of the extreme rays with the diffuse bands, 

 which, according to Janssen, increase in intensity as the square of the 

 density of the gas. If that be so, it is unlikely that the limit of the 

 solar spectrum is due to the absorption of ordinary oxygen. For 

 though we may suppose interplanetary space to be pervaded by 

 materials similar to our atmosphere, yet they must be in such a state 

 of tenuity that, although they may to some extent reinforce A and B, 

 they will not add sensibly to the strength of the diffuse bands. 

 Moreover, these bands, though identical in position, so far as the blue 

 and less refrangible part of the spectrum is concerned, with bands 

 observed by Brewster and others in the solar spectrum, are seen much 

 more strongly through our tubes of compressed oxygen than they 

 appear in the solar spectrum with a low sun. The ultra-violet bands, 

 of which the one near "N appears in our photographs nearly as strong 

 as the band just above F, and that in the indigo, have not, so far as 

 we are aware, been noticed in the solar spectrum. Probably they 

 would appear if photographs were taken with small dispersion when 

 the sun was low. 



As the pressure in the tube diminished, the bands rapidly faded ; 

 that in the indigo, with an oscillation-frequency or wave-number 

 about 2240, was the first to disappear, then those near L and and 

 that near E. At the same time the limit of the transmitted ultra- 

 violet light advanced from an oscillation-frequency of about 3575 at 

 97 atmospheres, to 3710 at 50 atmospheres and 3848 at 23 atmo- 

 spheres. At 20 atmospheres the three bands above C, D, and F, 

 respectively, were still visible, though faint. B remained visible until 

 the pressure was reduced to 2 atmospheres, and A could still be seen, 

 but with difficulty, when the pressure of the 18 metres of oxygen was 

 reduced to 1 atmosphere. 



When atmospheric air was substituted for oxygen we found that 

 7 atmospheres was the limit of pressure at which we could certainly 

 distinguish A, and 18 atmospheres the limit at which we could see B. 

 It is a difficult matter to say exactly when an absorption becomes 

 invisible, but the observations on air were made under the same cir- 

 cumstances as those on oxygen, and the two sets of observations were 

 fairly comparable. With air at 75 atmospheres the three bands above 

 C, D, and F, respectively, could all be seen, but that near C only 

 with difficulty. The mass of oxygen and its partial pressure in the 

 tube was in this case less by about one quarter than that which was 

 required to bring out the bands when oxygen alone was used. The 



