Singular Rainbow. 545 



A B was a perfect and very splendid primary bow. c D re- 

 presents a secondary bow, having its colours of course in an 

 inverse order to those of the primary. But from e, in the 

 direction e f, came the segment of a third bow, commencing, 

 as nearly as could be guessed, between 1 5° and 20° from the 

 vertex of the secondary, and descending in the direction shown 

 in the figure. This segment was equally vivid with the second- 

 ary, but had its colours in the order of the primary. No con- 

 tinuation of it could be discerned above or on the left of the 

 secondary, and its centre evidently lay considerably to the left 

 of that of the other bows. The termination at the lower ex- 

 tremity (f) was abrupt and well defined. It remained visible 

 for more than ten minutes ; the extremity (f) appearing lat- 

 terly to have moved somewhat nearer to the primary bow, 

 from which, however, it was never less than a bow's breadth 

 distant. The side c of the upper bow first became invisible, 

 but so long as the other side could be seen the segment could 

 also. 



A somewhat similar phenomenon is recorded in the PhiL 

 Trans, p. 1793., as having been seen on the coast of Hamp- 

 shire ; and the explanation there offered is, that the additional 

 bow was caused by the reflection of the sun from the sea when 

 perfectly still. But, if this were the cause, would not such 

 appearances be more usual ? Besides, if a ray of light pro- 

 is 4. "^^^^^^ ceeding from the earth falls on 

 the lower side of a drop, must 

 it not be refracted into the air 

 (as in^^. 314.), and not towards 

 the earth, s being considered 



^^^ the reflection of the sun ? And 



what again becomes of the received opinion, that the eye of 

 the observer must be in the apex of the cone, of which the 

 bow forms the base ? 



I am, Sir, &c. 

 Coventry y Oct. 12. 1829. E. G. 



