GASEOUS REFRACTIVE INDICES 279 



tubes, one of which contains the gas under examination. When 

 the pressure of the gas in one tube is changed, the bands which 

 pass the cross-wire in the telescope are counted, and the refrac- 

 tive index of the gas is found from the equation 



_ NX 760 t 

 91 * " L P 273 



where N is the number of bands, \ the wave length, L the 

 length of the tube, t the absolute temperature, and P the change 

 of pressure in millimetres. 



This method is capable of extreme accuracy. Apart from the 

 impurity of the gas, which is generally much the largest source 

 of error, the limits of its effectiveness are to be found in the 

 difficulty of reading the bands to less than a tenth, in counting a 

 large number of bands without mistake, and in the drift of the 

 zero point during the experiment owing to unequal changes of 

 temperature in different parts of the apparatus. And of these 

 the last is the most important. In order to ascertain the density of 

 the substance under examination, various methods are employed. 

 With permanent gases which do not attack the materials of the 

 apparatus the temperature and pressure are directly observed. 

 If the gas attacks mercury, it is led into the refractometer tube 

 at atmospheric pressure, and the bands read as it gradually 

 displaces the air. 



With vapours the temperature and pressure are sometimes 

 read, and the density calculated from previously ascertained 

 data. But this method is not accurate, and it is much better to 

 have a density bulb in connection with the refractometer, and 

 to take the density directly. With solids and liquids which 

 require high temperatures to vaporise, such as mercury, 

 the most fruitful method has been to introduce a weighed 

 quantity into the refractometer tube, which is evacuated and 

 sealed off. The temperature is then raised till the whole of the 

 substance is vaporised and the density is calculated from the 

 weight of the substance present and the volume of the tube. 

 Hitherto this plan -has not been much used for compounds, and 

 a rich harvest of important results would probably reward 

 any one who would take up this branch and work it out 

 methodically. 



There are three great problems upon which it seems probable 

 that the study of refractive indices will throw light : the relative 



