72 



THE ROYAL SOCIETY OF CANADA 



scarcely perceptible. The intervals between successive spectrometer 

 settings were sometimes as small as 20 A.U. 



Fig. 4 shows the curve obtained by plotting the change in refrac- 

 tive index for 1 atmosphere change in pressure against the wave- 

 length of the light used. This curve shows dips at the wave-lengths 

 where the absorption bands occur but there are dips at other places 



I 



i 



\NAVELENGTHS IN HUNDREDS OF ANGSTROM UNITS . 



and it looks as if the observed variations in the change in refractive 

 index for 1 atmosphere change in pressure close up to the absorption 

 band are of the same order as the experimental errors; therefore no 

 conclusions can be drawn from these results. 



It follows from this that in order to detect any variation in the 

 change in refractive index corresponding to that shown in Fig. 3, it 

 would be necessary to find the absolute values of the refractive index 

 for ranges of, say, 5 A.U. right across the estimated position of the 

 absorption band. This would entail counting about 6000 fringes for 

 each reading and clearly a photographic method might with advan- 

 tage be used. 



