364 ON THE DISTRIBUTION OF THE POLARISING FORCE 
ternal tints do not rise so high, from the diminution in the 
thickness of the glass. 
The same phenomena are exhibited by a glass-tube AD, 
Fig. 6. similarly placed. In this case, however, the maximum 
tint does not appear along the line mm, the axis of the tube; 
but it is seen both in the lines 6d and dd, equidistant from 
mm. This effect is obviously occasioned by the greater thick- 
ness of the glass in these directions ; for a and cd are each 
greater than Ao-+pm, and the difference is sufficiently great 
to overbalance the diminution of the tints at a greater distance 
from the principal axis. 
In examining very carefully the structure of glass tubes, 
when exposed to polarised light, it will be found that they are 
as it were divided into different elementary concentric tubes ; 
and that in some cases there is an actual separation between 
them. Hence, there arises a remarkable singularity in the 
progression of the tints. Instead of shading into one another 
by imperceptible gradations, each elementary tube has an uni- 
form tint of its own, as is represented in Fig. 7. where the 
tube AB is divided into four tubes 1,1; 2,2; 3,3; 4,4; the 
tube 4,4 is in every part of it a white of the first order ; the 
tube 3,3 is every where equally dark, being the black cir- 
cular fringe mpno of Fig. 3.; the tubes 2,2 and 4,4 are 
a white of the first order; and the tube 1,1 is a yellow of the 
first order. Hence it follows, that in tubes which possess this 
peculiarity of structure, each elementary tube is uniformly 
dense throughout, and that the variation of density takes place 
by leaps. 
If a portion of a glass tube perfectly annealed, is exposed to 
pressure, it exhibits the tints shown in Fig. 8. when the line 
mn, joining the points of pressure, is parallel or perpendicular 
to the plane of primitive polarisation ; and the tints shown in 
Fig. 9. when the same line is inclined 45°to that plane. 
5, On 
