326 On the Interference of Light near a Caustic. 



pressures produces a slight but sensible effect upon the distance 

 to which the light extends beyond the caustic. 



The side of the tube being oiled, with a little care in adding 

 s:nall quantities of water after the stream had ceased to run, I 

 obtained on the 10th of April a good drop at the extremity of 

 the tube, which remained perfect to the end of the experiments. 

 The drop obtained in this manner is larger than one would wish, 

 but the telescope being carefully adjusted to distant objects, or 

 for parallel rays, and the observations being alw 7 ays taken at the 

 centre of the field of view, there is no correction of the observa- 

 tions required on that account. The iris being steady, was much 

 better seen than in the running stream, and the maxima or 

 bright bars were counted to upwards of forty when a piece of 

 red glass was held before the eye. 



The drop being T \jth of an inch in diameter, and the other 

 circumstances as in the last experiments, the following measures 

 were obtained for the limit beyond which no light could be dis- 

 cerned, from the perpendicular ray : — 



42 22 

 42 23i 

 42 24i 

 42 71 



A good set of measures of the places of the first six maxima 

 were also obtained, whose distances, counted from the caustic, 

 approximated in progression to those of Newton's reflected rings, 

 counting from a black centre as ordinarily considered, namely 

 they approximated to the progression of the values of 1, \/3, 

 Vo, V7, &c. 



As the extreme limit of the visible red was taken, it is clear 

 that since the asymptote of the caustic or place ordinarily 

 assigned to the primary rainbow, for the fixed line B is 42° 22' 

 from the perpendicular ray, therefore no light extends beyond 

 this ; and that the rays which have traversed equal spaces pro- 

 duce dai'kness in interfering, contrary to the fundamental require- 

 ment of the undulatory theory. 



The support which that theory has so long had, from experi- 

 ments made with running water compared with calculations for 

 still water, must now cease. 



London, April 13, 1855. 



