32 



clearness. The field within it was notably darker than the- 

 surrounding portion, which was lighter even than the sky 

 beyond. The limb of the halo became wider as the suu 

 approached the horizon. Presently the underneath side 

 blending, vanished, whilst the remaining portion could well be 

 mistaken for an iridescent belt of cloud. This at last became 

 quenched by the brilliant crimson clouds. No glow existed. 

 Moonlight during evening displayed a similar halo. 



It will be seen from the two descriptions just given that 

 frequent mistakes might be made in taking a portion of some 

 halo for iridescent clouds; and as the character of polariza- 

 tion is very obscure, it is almost impossible to distinguish 

 the difference by the aid of the polariscope, and experience in 

 the various forms the halo assumes becomes necessary. 



In many instances short pieces of rainbow might pass for 

 iridescent clouds, but they can be instantly detected if viewed 

 through a nicol-prism, as they can be completely quenched witb 

 the nicol-prism in radial position. 



* The following note is important, as bearing directly on 

 this part of the subject : — A tinted sunrise ; sky overcast, with 

 clouds and patch}^ colours to north-east, but to the west there 

 was a fine exhibition of tinted clouds. The fine high clouds- 

 were tinted pink, and their light was polarised in planes passing 

 through the suu, the clouds, and myself. Near the east no 

 change was produced when the nicol-prism was rotated. The 

 colour of the clouds was pink, and very bright towards the west 

 portions of sky, opposite to the sun, which had not yet risen 

 above the horizon to my sight ; and a phenomenon existed that 

 might easily have been mistaken for an iridescent cloud. A 

 band appeared, with red to the outside and green to the 

 inside of the circle of which this band formed a. part. 

 The inside area of this circle was tinted red, and it was- 

 nearly as bright as the red of the portion of ring described. 

 This ring was evidently part of a rainbow, for the nicol-prism 

 completely extinguished it when the long diagonal was at right 

 angles to its circumference. But what was more important 

 was the fact that the red within this ring, and also the red 

 without — though there was not much outside — suffered extinc- 

 tion with the nicol in same position. It occurred to me that 

 if the nicol were directed to the diff'erent points of sky corres- 

 ponding to the position of the unseen rainbow, the light of the 

 red haze would be extinguished with the diagonal of the 

 crystal at right angles to the tangent. Trying this, I found 

 it to be the case. 



This is an important note, because of the character of polari- 



* This note was inserted after the paper vras delivered. 



