168 Mr. Knox on some phenomena of colours, 
seemed to originate; that the dividing ring between the classes, 
passed through a point, whose distance from the centre of each 
primary set was in proportion to its largest diameter. 
This will be better understood by referring to fig. 2. pi. VII. 
where A and B are the primary sets of rings : C will represent 
oneof the newly discovered sets, which were denominated inter - 
sectionaries* from their apparent origin. The fourth ring from 
the centre will be the division between the two classes. Those 
rings within the division having the red on their insides, and 
those without having the red on their outsides, as represented 
by the figure; where the same rule has been observed in 
shadowing these rings, as was observed with respect to the 
parallel fringes in fig. 1, namely, that the shaded side repre- 
sents the violet, and the unshaded side the red of each ring, 
and the dark lines the divisions between them. 
Admitting that parallel fringes are necessarily rectilinear, 
in consequence of being drawn through intersections of cir- 
cles that are perfectly equal in dimensions, it follows, that 
where two sets of circles differ in dimensions, the correspond- 
ing intersections cannot lie in straight lines, but must neces- 
sarily be circular, as will appear evident on inspection of the 
figure ; where the dividing ring D E could not pass through 
the several intersections in the points 1, 2, 3, 4, 5, 6, 7, 8, g, 
10, 11, 12, unless it were circular : the same thing will hold 
true of all the other rings of the intersectionary set. 
It also appears by the foregoing experiments, that rectili- 
near fringes and intersectionary rings are coloured exactly 
alike, and are alike divided into two classes with respect to 
the order of the colours. There is also an exact similarity in 
* The best apology for using a new word is, that it expresses a new idea, 
