Phenomena of the Rainbow. 291 



failure to do this be not perceptible in a bulb one inch in dia- 

 meter and near to the eye, it is not surprising that the light 

 from the rain-drops should appear colourless. The light at 

 length becomes coloured, partly in consequence of the coloured 

 rays arriving successively at the position of parallel emergence 

 and greatest brilliancy, as stated by M. Biot, and partly in 

 consequence of the drops failing successively to send to the 

 eye certain of the colours, the blue failing first, and the red 

 last. 



From these observations it appears to me, that the pheno- 

 mena of this beautiful apparition are no where detailed as per- 

 fectly and comprehensively as they might be ; and I have 

 endeavoured, in the diagram, Plate 4., to convey a clearer 

 and more popular notion of them than is usually to be found. 

 Here the irregular figure A B C D represents a section of 

 a shower of rain, taken in any plane passing through the 

 eye of the spectator and the sun ; the former being at E, the 

 latter infinitely distant on the line E C. Under these circum- 

 stances, coloured rays, formed by two refractions and one 

 internal reflexion, will reach the spectator from every drop in 

 the cone whose section is E F G ; but the colours will be 

 nearly neutralized by superposition everywhere, except at and 

 near the surface of the cone, where they will give the impres- 

 sion of the primary bow ; the cone of red rays being larger 

 than that of the orange rays, this larger than the cone of 

 yellow, this again larger than that of green, and so on. 



Coloured rays, formed by two internal reflexions and two 

 refractions, will reach the spectator from every part of the 

 shower, except the cone E H I, the colour being, as before, 

 neutralized by superposition in every part, except at and near 

 the extreme edges of the illuminated space, where each colour 

 will successively overlap the last, in the same order as before, 

 producing the secondary bow. 



The space between the two cones, E F G and E H I, returns 

 no light after one or two internal reflexions, and is therefore 

 comparatively dark, though the difference is by no means so 

 great as, for the sake of distinctness, it has been made in the 

 engraving. 



In confirmation of the view here taken of the causes which 



