504 Mr. R. H. M. Bosanquet on_ the 



(in the plane of the beam) ; the portion polarized negatively 

 (passing through a Nicol which stops the positively polarized 

 light) affords the phenomenon called by Tyndall residual blue. 



(j3) If a point on the beam be observed from positions such 

 that the light examined diverges at an acute angle with the in- 

 cident beam, a position can be found in which the emitted 

 light is neutral, or devoid of polarization. 



(7) If a point on the beam be observed from positions such 

 that the light examined diverges at a more acute angle than in 

 the last case, the light is polarized negatively, or at right angles 

 to the beam. 



In Stage II. there is a progressive change, which will be as 

 follows, if we suppose the density, or, as we may say with great 

 probability, the mean size, of the diffracting particles continually 

 to increase. 



(a) If the beam be regarded normally, the light, which 

 throughout Stage I. was perfectly polarized, ceases to be so, the 

 proportion of polarized light thus normally emitted diminishes, 

 and eventually the light normally emitted is neutral; this I take 

 to mark the conclusion of Stage II. 



(f$) The angle at which neutral light is emitted, which 

 throughout Stage I. was evanescent, gradually increases, till at 

 the conclusion of Stage II. it becomes a right angle, viz. when 

 the light normally emitted is neutral. 



(y) The negative polarization of light emitted at angles less 

 than the neutral angle, increases continually in intensity. 

 Towards the close of Stage II. this negative polarization is 

 stronger than the normal positive polarization. According to 

 the experiments the direction of emission for maximum negative 

 polarization seems always to have formed a very acute angle with 

 the beam. 



Stage III. A third stage appears to have been reached, in 

 which the polarization was everywhere negative. Here, 



(a) The polarization of light normally emitted, after vanishing 

 at the conclusion of Stage II., reappears with changed sign ; 

 i. e. it is now negative. 



(/3) No neutral light is emitted in any direction. 



(y) The direction of emission of the light with maximum 

 polarization forms a very acute angle with the beam. 



As these deductions are stated in a different form from that 

 employed by Tyndall, I must point out how they are derived 

 from his facts. 



As to Stage I. there is no difficulty. See ' Radiant Heat/ 

 p. 353. (The sentence at the end of (I.) is rather a stumbling- 

 block. "The direction of maximum polarization is at right 

 angles to the beam." This means "The direction of emission 

 for maximum polarization.") 



