DISPERSION IN ARTIFICIAL DOUBLE REFRACTION. 



2'J1 



TABLE IV. 



In this case there seems to be a progressive increase of 7-AX/AW,. with the order, for 

 both tension and pressure. This excludes a correction for sinking, since the latter 

 must act opposite ways for tension and pressure. It is here probably due to a failure 

 of HOOKE'S law, which the observations have shown otherwise, and which will be 

 discussed in a later section. 



Under the circumstances no real advantage could be derived here by attempting 

 to reduce the various sets by means of a single formula. The sets have, therefore, 

 been independently reduced, using the formula 



X = A, + W(AX/AW r ). 



The actual observations of all glasses are given in Table V. for purposes of reference. 

 Rich column corresponds to a single set of observations. As a rule the order of the 

 band observed will be clear from the place of the observation in the series. Wherever 

 this is not so, or where observations of different orders correspond to the same load, 

 Roman numerals have been added to indicate the order of the band. 



2 P 2 



