RKFIIACTIVK IXDHT.s OF THK KI.KMKNTS. 327 



calculate the ratio of the refractive index of the substance to that of air at the 

 same temperature and pressure, and thence to deduce the index under standard 

 conditions. 



It is obvious that this method is open to many objections, the gravity of which 

 was fully recognised by the experimenter. Indeed, in the case of arsenic, no great 

 accuracy was even claimed. It seemed probable, therefore, that, by using a more 

 accurate method, better results might be obtained. The plan adopted by nearly all 

 students of the refractivity of gases since LE Rorx has been that of interference, 

 and the instrument employed is JAMIN'S refractometer. M. MASCART, however, 

 made use of TALBOT'B bands in his researches on the refractivities of vapours, and 

 preliminary experiments were accordingly made to determine the comparative 

 advantages of the two plans. In the end JAMIN'S method was selected. 



The principle of the instrument, and its adaptation for measuring the indices of 

 gases, are too well known to need description. In the case of vapours the procedure 

 which seemed to promise best was to introduce into one of the two parallel tubes 

 some of the element whose index was sought, and to evacuate and seal off both tubes. 

 When this system is placed in position in the refractometer, and the tubes are 

 heated by means of an electric furnace, the path of that ray which passes through 

 the empty tube is unaltered within the tube. But the ray which traverses the tube 

 containing the substance is gradually retarded by the increasing density of the 

 vapour produced by heating. If the number of bands which pass over the field and 

 the density of vapour in the tubes are known, the refractive index can l>e calculated. 



In adapting JAMIN'S apparatus to the measurement of the indices of substances 

 which are not gaseous, except at high temperatures, several difficulties presented 

 themselves. The first, and most serious, was to find some way of effecting so perfect 

 a junction between the tube and the glass plate with which its end was closed that 

 it would hold a vacuum for several hours at a temperature of, at least, 300 C. 

 M. MASCART, in his researches on the indices of vapours, did not use any temperature 

 higher than 50 C., and the ends of his tubes were made air-tight by pressing the 

 plates of glass against a ring of caoutchouc. For higher temperatures this plan is 

 useless, and, consequently, new means had to be sought. Various devices were tried, 

 and, after about two months' work, it was found that the desired result could be 

 obtained by stopping the ends of the tube with discs of plate glass, ground 

 conical, and covering the junction with sealing-wax, or, better still, with shellac. 

 Tubes fitted with these ends have acted well ; but it is necessary that the tube 

 itself should I* of moderately hard glass, otherwise the unequal expansion of the 

 tube and plate glass end causes the former to crack with unfailing regularity. It 

 was calculated that, at any rate for Hg, S and P, a temperature of not more 

 than 300 C. would produce such a vapour density as would suffice ; and, to obtain this, 

 a primitive form of electric furnace was constructed which could raise the temperature 

 of the enclosed space to 270 C. in about an hour, and could attain 300 C. with 



