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



They appear to have been taken straight from the note- book, 

 if we may judge from several uncorrected mistakes; this, of course, 

 materially enhances their value as observations. I have no doubt 

 that they constitute by far the largest and most reliable body of 

 evidence accessible on the most interesting parts of the subject ; 

 also it is clear that they actually were the evidence on which 

 his map and discussion were founded. 



The considerable work of Rubenson (Memoire sur la Polari- 

 sation de la lumiere atmospherique, Upsal, 1864) is devoted almost 

 entirely to the variations of the maximum intensity of the pola- 

 rization of the sky, the time of day and year being connected 

 with it by empirical formula?. Although his instrument was 

 almost of an astronomical character as regards the attempt at 

 accuracy, and his series of observations are very extensive, these 

 are not for the most part applicable to the determination of the 

 points to which I have so far principally attended; in a more 

 advanced stage of the subject his work may probably be of value. 



The summary of general results, as stated by Brewster, refers 

 only to the plane through sun, zenith, and observer; and it is 

 clear from the observations that points in this circle, and the 

 point in the horizon at right angles to it, were alone generally 

 observed. The normal arrangement of the polarization is taken 

 to be when the sun is in or near the horizon (generally setting) . 



Polarization is said to be positive when light is polarized in a 

 plane through sun, point, and observer ; negative when polarized 

 at right angles to that plane. 



Thus, in Tyndall's experiments, the polarization of the first 

 stage, where the blue light first appears, and the polarization is 

 parallel to the beam, is positive ; in the subsequent stage the 

 polarization of rays forming an acute angle with the beam is 

 negative. 



Now Brewster and all the observers are agreed in this, that 

 the polarization of the region in and about a great circle at right 

 angles to the sun is always positive, and that the position of 

 points of maximum positive polarization is always at right angles 

 to the sun. 



They are also agreed that there are in the plane through sun, 

 zenith, and observer at least three neutral points : — the point of 

 Arago, between the zenith and antisolar point ; the point of 

 Babinet, between the zenith and the sun ; and a point called by 

 Brewster's name, below the sun. 



Also that between Arago' s point and the antisolar point the 

 polarization is generally negative ; that between Babinet's point 

 and the sun the polarization is negative ; and between the sun 

 and Brewster's point the polarization is negative ; though the 

 observation of the negative polarization in the last case, and 



