115 
exhihlted hy Moilter-qf-Pcarl. 
These nebulous masses are usually about twice the distance 
of the coloured images, described by Dr Brewster, from the 
centre ; and, except the plate be very thin, are much larger 
and more conspicuous^ and particularly distinguished from them 
by the equable gradation and softening of their colours, which 
are those of the prismatic spectrum, the red being outermost. 
Their angle of deviation, or distance from the central image, 
increases on inclining the plate in the plane passing through 
them, while their brightness rapidly diminishes, the former be- 
ing a minimum, and the latter a maximum, at a perpendicular 
incidence. This angle, as well as the shape and colour of the 
nebulae, is the same, or nearly so, in all the specimens I have 
examined ; nor does any marked variation in these particulars 
arise by a variation in the thickness of the plate, or by cutting 
it at any moderate angle with the natural surface; only, in the 
latter case^ the maximum of their intensity^ and minimum of 
distance^ takes ylace at such an angle of incidence^ that the or- 
dinary ray traverses the substance in a direction perpendicular 
to the natural surface of the shelly which is also the direction 
of its greatest transparency. The minimum angle of deviation 
in the nebulous images appears, by the best measures I have 
been able to procure to be 
For the extreme red, about - - 10° 29' 
For the mean rays, - - - 6 59 
For the extreme violet, - - - 4 10 
The extraordinary images seen by reflection, and their com- 
plementary pair by transmission, are completely obliterated by 
surrounding the mother-of-pearl with oil, or varnishing its sur- 
faces. This, however, is so far from impairing the nebulous 
masses, that it heightens them a little, by perfecting the polish ; 
and should any doubt arise as to the identity of either pair, 
(which will sometimes happen), it may thus be immediately re- 
moved. From all these circumstances, compared with what Dr 
Brewster has demonstrated respecting the former pair of images, 
we might reasonably conclude that the latter are entirely un- 
* By passing a sun-beam, of an inch in diameter, through a thin plate, 
and measuring the images projected on a screen. 
n 2 
