352 P. C. on the Colours that enter into the [Nov. 



by varying the distance of the screen it may be observed in 

 different stages of its developement. If we look at the 

 spectrum thus formed, through an orange-yellow glass, 

 which we have already stated is nearly opaque to violet 

 light, we shall find that the violet part of the spectrum will 

 disappear, and that the blue which is found in it previous 

 to its full developement, will invariably be converted to a 

 green. If the spectrum be completely developed, the violet 

 only will disappear, leaving the red and green very little 

 altered in their appearance by the interposition of the 

 coloured glass. 



Now, if blue be a simple colour, upon what principle is 

 this conversion made 1 The yellow glass may refuse trans- 

 mission to one of the constituents of a compound colour, 

 but it cannot add to a simple colour ; it may weaken it, but 

 it cannot change its character. 



If we look through the yellow glass at the spectrum in 

 an early stage of its developement, before the white centre 

 has entirely disappeared, this white will be converted to a 

 yellow : the violet will disappear, and the blue be changed 

 to green, as usual. 



The yellow centre, then, is evidently white light from 

 which the violet has been withdrawn, or the complement of 

 violet light : it follows, that if green be a compound of blue 

 and yellow, it must enter into the composition of one of its 

 own constituents ; for green is transmitted by the coloured 

 glass, and, of course, must form part of the yellow centre. 



We may vary the experiment, by placing the coloured 

 glass which refuses transmission to the violet light, either 

 before or behind the prism ; and we shall then have a spec- 

 trum formed of only two of the primitive colours ; which, 

 in its early stages, will exhibit red, the least refrangible of 

 these colours, on one side; and green, the most refrangible, 

 on the other ; separated by a yellow, formed by their super- 

 position . 



This experiment alone would decide the question at issue; 

 the spectrum which is formed by it never exhibits more 

 than three colours, red, yellow, and green ; one of which, 

 the central yellow, is evidently, from its disappearance in 

 proportion as they are developed, a compound of the other 

 two. 



It is probable that the communication of this orange- 



