330 MK. CLIVK CUTHBKUTSOX ON* THE 



The tube which remained in connection with the pump was then exhausted, and 

 the height of the mercury in the manometer noted. The observer, having taken his 

 seat at the telescope and noted the position of the bands, turned on the electric 

 current. The tubes slowly warmed, and in about an hour had reached a temperature 

 of 270 C. As the substance grew hot it began to volatilize, and the bands which 

 slowly crossed the field were counted. When that temperature was reached at which 

 all the substance present had passed into vapour, the bands ceased to move with 

 remarkable abruptness. When this had occurred the temperature was allowed to 

 rise about 10 C. further, to make s\ire that no more of the substance was still lurking 

 unevaporated in any cool corner of the tube. The manometer was again noted, and 

 was seldom found to have varied as much as a millimetre, in spite of several rubber 

 connections. The supply of heat was then cut off, and the bands again counted, as 

 they retreated, owing to the condensation of the vapour. Lastly the manometer was 

 again read. Thus, in each experiment two readings were obtained.* After making 

 the necessary corrections these two readings coincided very closely, and thus furnished 

 a valuable proof of the absence of leaks. 



When the tubes had cooled to such a point that no further movement of the bands 

 could be detected, the tube containing the substance was taken out of the furnace 

 and examined. In a good experiment the whole of the element volatilized was 

 found condensed in the portions of the tube which had successively become coolest as 

 the temperature fell, and unless the deposit looked perfectly clean and unoxidised 

 the experiment was discarded. This tube was then opened at the extremity of the 

 appendix, and its volume measured by weighing it empty and full of distilled water. 

 Its length was also measured. 



From these data the calculation of the refractive index is simple. What we wish 

 to compare is the retardation of light caused by the presence of equal numbers of the 

 atoms of different elements. In the case of the permanent gases the indices are 

 calculated for standard temperature and pressure, and it is assumed that the numbers 

 of molecules are then equal. If the state of aggregation of the molecule is also 

 known, the materials for comparison are complete. In the case of vapours we 

 measure the retardation of light produced by a known density of vapour, and 

 calculate what the retardation would be if the number of atoms per unit volume were 

 the same as they are in the case of the permanent gases at normal temperature and 

 pressure. This is the case when the density of the vapour bears to that of a gas 

 the ratio of their respective atomic weights. Hence, we have only to multiply the 

 refractivity observed in an experiment by the ratio of the standard to the observed 

 density in order to obtain the index of refraction. In the present case the density 

 of the vapour employed is calculated from the weight of the element evaporated and 

 the volume of the tube, and the standard density chosen is that of H 2 ('0000899) 



* At least in the later experiments. In the earlier the upward run was lost owing to moisture con- 

 densing on the diaphragms. This was afterwards obviated by a preliminary heating of the apparatus. 



