C. S. Hastings — Double Refraction in Iceland Spar. 61 



standpoint is that it led Stokes to the first careful investigation 

 of the accuracy of Huyghens's construction.* 



In these investigations Professor Stokes found that the error 

 in the construction could hardly exceed a unit in the fourth 

 place of decimals, which was quite sufficient to disprove Ran- 

 kine's hypothesis. This study, the details of which have not 

 been published, remains unexcelled to the present time ; for the 

 investigations since made by Abria, Glazebrook and Kohl- 

 rausch, whether by the prism method or by total reflection, do 

 not present a closer accordance between theory and observa- 

 tion. The results of earlier observers, cited in most treatises 

 on double refraction, are all of quite inferior accuracy. 



Of all these investigations, Glazebrook's, given in the Trans. 

 Roy. Soc, vol. clxxi, 1880, is the most extensive. His method 

 was to measure the deviations produced by four different 

 prisms, so cut from the same piece of Iceland spar that the di- 

 rections of the propagation of the light varied from an angle 

 of — 3° to + 94° to the crystalline axis, the relation of this 

 axis for each prism to its faces being determined by reference 

 to planes of cleavage. The observations were made with con- 

 siderable accuracy, indicating a probable error in the deduced 

 indices of refraction considerably less than fifty units in the 

 sixth place of decimals. The reductions, however, show a sys- 

 tematic deviation from Huyghens's construction, varying from 

 100 to 200 in the sixth decimal in the three hydrogen lines ob- 

 served — the wave-surface for the more refrangible ray deviat- 

 ing most widely. This result would be of great theoretical 

 interest if the values derived from observation were not 

 vitiated by an important oversight in the details of the experi- 

 ment, which the author himself points out. In view of this 

 source of error the conclusion from the investigation is, that 

 Huyghens's construction is true within the limit of error of 

 these observations. 



Briefly, then, the state of the case is this. The law of double 

 refraction in Iceland spar as given by Huyghens is known to 

 be true to within about one part in ten thousand, but no rea- 

 son, dependent on the theories of elasticity, can be assigned 

 why it should be as accurate as this, or how much more ac- 

 curate we may expect to find it. The labor of testing the law 

 to the last degree of refinement possible with modern instru- 

 mental means seems well worth while ; for, except its sim- 

 plicity, there is no reason in the world why it should not break 

 down just at the limit assigned by Stokes's observations. I am 

 quite willing to admit, also, that the systematic deviations of 

 Glazebrook's observations, so near the limit of magnitude set 



* Proceedings of the Royal Society, June, 1872; quoted by Sir Wm. Thomson 

 in his Baltimore Lectures, p. 273. 



