1 70 Mr. Knox on some pheriomena of colours , 
transmitted or reflected sets, provided the two, between which 
they are formed, are unequal in dimensions. 
Exp. 8. The four sets of rings represented in fig. 3, ( pi. VIII. ) 
were produced by an arrangement consisting of two convex 
lenses laid on each other, and a piece of plane looking-glass 
plate laid on the uppermost lens. The large imperfect set, 
whose centre is at D, was evidently produced by the intersec- 
tions of the two primary sets, A and B,(the former was observed 
to have a dark and the other a light centre), but the set whose 
centre is at C cannot be accounted for on the same principle ; 
but may be owing to another cause, which will be better un- 
derstood by a future experiment. 
Notwithstanding these intersectionaries seem to owe their 
origin to a few rings only, adjoining that which divides the 
two classes ; yet they are always filled up to the centre with 
prismatic colours ; unless when very large, as that whose 
centre is at D, (fig. 3. pi. VIII.) or when they appear in part 
only, as segments of circles ; in which case the central spaces 
* 
appear blank. 
Exp. 9. It is perfectly certain that these intersectionary 
rings are always formed between the two contiguous surfaces 
of the two lower pieces of glass when three pieces are em- 
ployed ; this is easily proved by the test of the shadow. But 
if four or more pieces are laid on each other, the number of 
spectra may be increased indefinitely, and these again may be 
multiplied by internal reflections ; so that to pursue these 
phenomena through all the varieties of which they are capable 
would be an endless task. 
, Let it be observed here, that the several drawings of pris- 
