on the Hypothesis of Undulations. 335 



which points to such a difference. It is possible that, in the 

 prismatic colours, the condition m—m! may be less exactly ful- 

 filled than in the colours of substances. A probable reason may 

 be given for such a circumstance. Sunlight is generated within 

 the sun's atmosphere at a great depth below the exterior bound- 

 ary, and the ray- undulations have consequently to traverse a 

 medium which may absorb them in different degrees according 

 to the value of X. The absorption might stifle particular 

 rays, and change the value of m for others ; and if one colour, 

 as yellow, were transmitted more readily than the rest, the 

 value of m for this colour would be larger than for the others, 

 supposing that originally m was the same for all. Thus the 

 excess of brightness in the yellow part of the spectrum may be 

 owing, not only to a greater number of yellow rays, but also 

 to greater intrinsic brightness of each yellow ray. Now the 

 coloured light of substances, though derived from sunlight, is yet 

 new light, because it is produced at their surfaces, or within a 

 small depth below their surfaces, by disturbances of the solar 

 undulations similar to those already described with reference to 

 absorption, which is, in fact, a phsenomenon of the like kind, 

 differing only in degree. This terrestrial light, not having passed 

 through a large extent of atmosphere, may satisfy the condition 

 m=m', and in this respect differ from solar light, for which the 

 value of //< in a mixture of blue and yellow rays may be consider- 

 able. Such a difference would account for the more complete 

 production of green by the blue and yellow of terrestrial sub- 

 stances, than by the blue and yellow of solar light. 



Our theory shows that there would be no reason to conclude, 

 from even a perfect imitation to the sense of green by a mixture 

 of yellow and blue, that the green of the spectrum is a compound 

 colour. 



(6) " Red with green gives a yellow, which, less saturated, is 

 paler than the simple yellow." "Green gives with indigo a 

 bright blue, which, however, is much duller and whiter than that 

 of the spectrum ; with violet also it gives a bright blue." (Helm- 

 holtz, Phil. Mag. pp. 525 and 526.) These facts, which relate 

 to the colours of the spectrum, are in accordance with the theory. 

 The dullness and whiteness of the resulting colour is owing, not 

 only to the difference of wave-lengths of the components, but 

 also to difference in the intrinsic brightness of the individual rays. 

 For it may be remarked, that in the combination of colours, the 

 wave-lengths of which differ considerably, the result is more de- 

 cidedly coloured in the above instances than in that of the mix- 

 ture of yellow and blue, although in this combination the differ- 

 ence of wave-lengths is not so great as in the others. The 

 excess of whiteness of the blue and yellow compound may there- 



