September 29, 1905.] 



SCIENCE. 



387 



by MacCullagh (1837), Briot, Sarrau 

 (1868) ; but there is naturally no difficulty 

 in accounting for rotary polarization by 

 the electromagnetic theory of light, as was 

 shown by Drude (1892). 



Among investigational apparatus of 

 great importance the Soleil (1846, 1847) 

 saccharimeter may be mentioned. 



THEORIES. 



In conclusion, a brief summary may be 

 given of the chief mechanisms proposed to 

 account for the undulations of light. 

 Fresnel suggested the first adequate optical 

 theory in 1821, which, though singularly 

 correct in its bearing on reflection and 

 refraction in the widest sense, was merely 

 tentative in construction. Cauchy (1829) 

 proposed a specifically elastic theory for 

 the motion of relatively long waves of light 

 in continuous media, based on a reasonable 

 hypothesis of molecular force, and deduced 

 therefrom Fresnel's reflection and refrac- 

 tion equations. Green (1838), ignoring 

 molecular forces and proceeding in accord- 

 ance with his own method in elastics, pub- 

 lished a different theory, which did not, 

 however, lead to Fresnel's equations. 

 Kelvin (1888) found the conditions im- 

 plied in Cauchy 's theory compatible with 

 stability if the ether were considered as 

 bound by a rigid medium. The ether 

 implied throughout is to have the same 

 elasticity everywhere, but to vary in den- 

 sity from medium to medium, and vibra- 

 tion to be normal to the plane of polariza- 

 tion. 



Neumann (1835), whose work has been 

 reconstructed by Kirchhoff (1876), and 

 MacCullagh (1837), with the counter- 

 hypothesis of an ether of fixed density but 

 varying in elasticity from medium to me- 

 dium, also deduced Fresnel's equations, 

 obtaining at the same time better surface 

 conditions in the case of seolotropic media. 



The vibrations are in the plane of polariza- 

 tion. 



All the elastic theories essentially predict 

 a longitudinal light wave. It was not until 

 Kelvin in 1889, 1890 proposed his remark- 

 able gyrostatic theory of light, in which 

 force and displacement become torque and 

 twist, that these objections to the elastic 

 theory were wholly removed! MacCullagh, 

 without recognizing their bearing, seems 

 actually to have anticipated Kelvin's equa- 

 tion. 



With the purpose of accounting for dis- 

 persion, Cauchy in 1835 gave greater 

 breadth to his theory by postulating a 

 sphere of action of ether particles com- 

 mensurate with wave-length, and in this 

 direction he was followed by F. Neumann 

 (1841), Briot (1864), Rayleigh (1871) and 

 others, treating an ether variously loaded 

 with material particles. Among theories, 

 beginning with the phenomena observed, 

 that of Boussinesq (1867, et seq.) has re- 

 ceived the most extensive development. 



The difficult surface conditions met with 

 when light passes from one medium to an- 

 other, including such subjects as ellipticity, 

 total reflection, etc., have been critically 

 discussed, among others, by Neumann 

 (1835) and Rayleigh (1888) ; but the dis- 

 crimination between the Fresnel and the 

 Neumann vector was not accomplished 

 without misgiving before the advent of the 

 work of Hertz. 



It appears, therefore, that the elastic 

 theories of light, if Kelvin's gyrostatic 

 adynamic ether be admitted, have not been 

 wholly routed. Nevertheless, the great 

 electromagnetic theory of light propounded 

 by Maxwell (1864, 'Treatise,' 1873) has 

 been singularly apt not only in explaining 

 all the phenomena reached by the older 

 theories and in predicting entirely novel 

 results, but in harmoniously uniting as 

 parts of a unique doctrine, both the electric 



