THE RAINBOW DUE TO A CIRCULAR SOURCE OF LIGHT. 27 



SO on. If ill tliis int(>i-v;il between the two stationary values the 

 maxima of {'2) corresponil to those of (1), then in the next in- 

 terval the maxima of (2) correspond to the minima of (1). 



((/) For larger values of r (radius of the drop) the intensity 

 of (1) and (2) increases by r^^. But at the same time for (2), 

 the diiference between the maximum and minimum values is 

 diminished by another factor F. 



(e) The above is more manifestly shown in the case of 

 the laboratory experiment with a cylindrical glass rod and a 

 straight slit as the source of light. The stationary points of (2) 

 at which the maximum value coincides with the minimum are 

 easily found by 



•/A) = mx interval of the maximum and minimum of (1), 

 where 7)i represents an integer. 



(/) According to Airy's tlieor}^, the lasv of the distribu- 

 tion of the colours of the rainbow is independent of the magni- 

 tude of the drop. But in the case of the finite source, the 

 colour distributions are changed by the magnitude of the drop, 

 especially in the supernumerary bows. 



iff) The supernumerary bows almost lose their colour as 

 the consequence of the finiteuess of the source. This effect is 

 more remarkable when the drop becomes larger. 



11. Note on the experimental side. 



To shoW' the above-mentioned results, we repeated rough 

 experiments witli gUiss rods and a straight slit as the source 

 of light. 



Using homogeneous light, we see that when the slit is very 

 narrow the phenomena nearly coincide with Airy's theory, and 



