66 



THE ROYAL SOCIETY OF CANADA 



increase of experience, this difficulty was surmounted o-nd it became 

 possible to make large pressure changes and keep the fringes of 

 countable clearness and size all the time. 



To measure the changes of density as the oxygen is released the 

 outlet valve was connected to a Hempel gas burette B by means of 

 a glass T piece provided with a stop cock (Fig. 2). The burette 

 contained water and readings were always taken when the water 

 surfaces were level. By letting equal volumes of gas escape from 

 the steel tube to the burette, the density of the gas was decreased 

 by equal amounts. Correction could be applied for slight variations 

 from this, also for temperature, humidity and so on. 



The volume of the steel tube and connections up to the two 

 needle valves was determined by a volumenometer method. It was 

 about 75.0 cc. so that it could be arranged with each run of the 

 burette that the pressure drop was practically one atmosphere. 



Series I. 



Table I. gives the num.ber of fringes crossing the centre of the 

 field of view of the spectrometer for one atmosphere pressure drop 

 at the pressures named and the wave-length setting of the spectro- 

 meter. 



TABLE L 



The numbers in each wave-length column are practically con- 

 stant, thus varifying Gladstone and Dale's law. Corrections were 

 not made for variation of atmospheric pressures, temperature and 

 humidity; this might be the reason for the slight variations observed. 

 The mean values were used to work out the refractive index of oxygen 

 at normal pressure. 



Table II. gives the values and those obtained by other experi- 

 menters. 



