»9 
proportion constant we become capable of a certain accuracy : we are 
furthermore led to reduce them to any proportion at pleasure, as we 
can repeatedly produce them: (see Tab. 3. and 4.) 
Seeinjif them produced in any place at pleasure, it will be found 
that they may be brought or placed in contact with each other, and 
be mixed : (see 7«Z'. 4. d. e. &,c.) Thus we are enabled to show, with 
as much nicety and calculation as we think proper, w hat may be the 
primitive rays ; and considering them as reduced to three, we can 
produce by means of a few strokes properly placed any possible tint, 
naturally and truly, without equivocation. In the present instance, 
for simplicity' sake, I determine the proportion by the breadth of the 
black. Thus a given tint may be represented by a given proportion 
of black upon white, as in Jig. d. &c. We have therefore a very 
simple means of operation for accomplishing what has long been 
thought beyond our reach. 
It appears that the primaries chiefly depend on the edges of the 
white and black of a certain proportion in contact, and that orange 
and green depend on the narrowness of the white between two black 
edges : thus a broad light upon black efluses yellow, red and blue, 
the red and blue always bordering more or less on the black back 
ground, and the yellow appearing upon some part of the space of 
light or white, whatever light it is, even the shining or gloss light 
on any black or dark coloured surface ; and should the light be nar- 
row, the blue on one side and the red on the other mix with the 
yellow, independent of the motion of the prism * ; and that the 
light being still narrower, the yellow is lost in green formed by 
its mixing with the blue; and when the red reaches the blue 
* I do not put a black page with white streaks as an example of this ; those who have 
a mind to examine it with a prism may easily procure such. The light spaces in Tab. 3. 
f. e. d. answer nearly the same purpose, showing the space of tlie lights producing the 
tintSj which are put on the side for better explanation : the lights being left in these ex- 
amples may be made use of, and will show the colours nearly as figured in the proper 
places for comparison. As we intend to use a white ground as most convenient, the pro- 
portion ofblack producuig the tints will chiefly be considered. 
D 2 
