514 Mr. R. H. M. Bosanquet on the 



Rubenson's observations are much complicated by the varia- 

 tion of the time of day, so that there is a difficulty in dis- 

 entangling the effects; and Brewster's impression seems to have 

 been in favour of a smaller correction. Some of his observa- 

 tions, however, favour a larger one than he employs. The point 

 will repay investigation. Experiments made in TyndalPs man- 

 ner can hardly deal with this point. In fact there can be no 

 doubt that when a beam is observed obliquely, the line of vision 

 travelling through a greater thickness of the beam than in the 

 normal position, the stage of the phenomenon is slightly more 

 advanced than in the normal position. I have endeavoured to 

 imagine an arrangement of the experiments which should elimi- 

 nate this complication*. The case of the sky alone can at 

 present supply information on this point. There we observe in 

 the zenith under the same conditions as in the position normal 

 to the experimental beam ; and in the horizon, at right angles 

 to the sun^ we have the observation through an increased thick- 

 ness, but still in a direction normal to the incident light. This 

 advancement of the stage when the beam is regarded obliquely, 

 probably accounts in some measure for the fact that the direc- 

 tion of emission for maximum negative polarization always forms 

 so small an angle with the incident beam. The smaller this 

 an gle i s, the greater will it be necessary to take n in the term 

 \/s'm 9 in the above empirical formulae. 



The theory of these phenomena has been treated by Lord 

 Kayleigh (see Phil. Mag. vol. xli. pp. 107, 274, 447, 519). 

 The few remarks I propose to make on the general theory are 

 offered with diffidence. At present I shall allude to only two 

 points. 



The theory which has seemed most satisfactory rests on the 

 assumption that the effect of matter on the aether is simply to 

 load it, i, e. to increase the inertia. Lord Eayleigh has em- 

 ployed an illustration (p. 521), in which he shows, from the ana- 

 logy of the motion of an ellipsoid in a fluid, that there is nothing 

 absurd in the idea of an inertia varying with the direction of 

 motion (p. 523). But the physical hypothesis on which the 

 theory then proposed is built up is not very precisely defined. 

 Now it seems to me that there is a great difficulty in employing 

 the hypothesis of the loading of the aether unless the mode in 

 which the loading is originated is defined. I take it that the 



* This could be done by employing a beam whose section is of the form 

 of a long and narrow rectangle, i. e. a beam forming a thin sheet of light. 

 By raiding the inclination of the line of vision to the sheet in a plane at 

 right angles to the lines of light, the effect here alluded to might be ob- 

 served. But it would be difficult to arrange such a beam unless, perhaps, 

 with direct sunlight. 



