Cause of Iridescence in Clouds, 429 



gives the angular distance and relative brightness of the 

 first four maxima for wave-length X. 



Table II. 

 Order of Spectrum. Angular Distance. Brightness. 



1 1*43 X/a 1 



2 2-46 X/a '37 



3 3-47 X/a -185 



4 .... . 4-48 X/a *109 



9 9-50 X/a -024 



For the sake of comparison I give the corresponding figures 

 for spheres of diameter D. 



1 1*64 X/D 1 



2 2-67 X/D -24 



3 3-69 X/D -095 



4 4-72 X/D -045 



Though the first ring is rather further out, the succeeding 

 intervals are much the same, but the intensity falls off much 

 more rapidly. 



To return to the filaments, (1) if the first red were at 4° 

 from the sun, the second and third would be at 7° and 9|°. 

 Observation points to something of this kind. (2) In obser- 

 vation we find the spectrum colours show themselves tolerably 

 distinct, in favourable cases, as far as the green of the second 

 order ; but outside this there is decided evidence of over- 

 lapping. And this is what we might expect from the theory, 

 when we remember, on the one hand, that only the brighter 

 parts of the spectrum need be considered, and, on the other, 

 that there are three important causes of blurring — the finite 

 diameter of the source of light, the spreading out of the 

 diffracted light on either side of the maximum, and the want 

 of uniformity among the particles, both as regards size, and, 

 though of less importance, as regards orientation. 



It will be remarked that only a small proportion of the 

 filaments are sufficiently near the "reflecting plane " to send 

 diffracted light. And it might be concluded that the light 

 sent would be very feeble. But this conclusion is by no 

 means borne out by some rough calculations I have made, 

 and which lead to the result that a cloud of ice filaments is 

 but little inferior in diffracting power to a cloud of water 

 drops. (This surprising result is chiefly due to the advantage 

 a filament gains from its length. The following is a general 

 explanation : — If a long filament and a sphere give equal 

 diffraction to the first spectrum, the area of the filament must 



