29 
divide it into 3£>0 parts, or degrees ; set one foot of your compasses in the 
centre, extend the other so as to take in 48 degrees, that is, 24 on each side 
the centre. On the same centre extend it again, so as to take in the proper 
portion of orange, do the same with red, and so on with the rest of the 
colours, filling each circle with the colour belonging to it, and blending 
each into the contiguous ones. Thus you will perceive all the colours 
arranged as it were on the surface of a ball *, in a sort of perspective suc¬ 
cession, each coming forwarder retiring from the eye according to its na¬ 
tural brilliancy : and if any one intrudes upon the eye, before its superior, 
you may be sure there is a fault in its situation. But if these colours were 
arranged in this order, in a picture, I believe there will not be found any 
greater error than what the necessary introduction of light and shade would 
sufficiently reduce to harmony, provided a nice attention is paid to the 
choice of the tints we make use of to imitate those in the Prismatic Spectrum : 
for it must be considered, that the pigments we have to express these colours, 
are far inferior in brilliancy—we have not any, I believe, equal to them. As 
to the red, vermillion will not express it, neither will red-lead, carmine, nor 
lake separately but those that appear to me the nearest to 
all, will be mentioned hereafter. 
It will be said, cplours arranged in circles in a picture would be very 
ridiculous, which it certainly would; but that is not meant: it is only in¬ 
tended to shew the proper situation for objects of such and such colours 
in respect to the others, not that there should be entire circles of them. If 
you admit white into your picture, in its pure state, it must come before 
yellow, and a space be left for it: as to the proportion, or quantity of it, 
that will be decided by the portion of pure light (or •white to express it) 
supposed to be on a ball, and by the same rule in your picture; so that the 
size of your picture must be your guide. The portion of each of the other 
colours must be regulated by the proportion they bear to one another in the 
Prismatic Spectrum, which is sometimes a difficulty, when the objects you 
* If they were thus arranged on the surface of a real ball, their perspective situations would be 
more readily perceived, as the prominence of the ball would really give each its proper distance. 
