362 ON THE DISTRIBUTION OF THE POLAR!SING FORCE 
When T = T’ the line of no-polarisation is a complete para- 
bola, passing through the angles of the parallelogram, and ha- 
ving its vertex at the very edge of the bent plate. The curves 
within it will also be complete parabolas, but all those without 
it will be only parabolic segments. 
4. On the Distribution of the Polarising Force in Tubes and 
Cylinders of Glass. 
We have already seen, that circular plates, or cylinders of 
glass, have one negative axis, like Quartz ; but when the cylin- 
der has the form of a tube, like AB, Plate VII. Fig. 3. the pola- 
rising force is distributed in a very remarkable manner. A 
black circular fringe mp no forms the line of no-polarisation, 
and the coloured fringes are placed on each side of this dark 
ring, and concentric with it. The structure on the outside of 
mp no is negative like Calcareous spar ; and the structure on the 
‘inside positive, like Quartz, &c.; and the effect is exactly the 
same as if a plate of glass had been bent into a circular form, 
and kept in that position by force. 
The breadth of the positive annulus ao is always less than 
that of the negative annulus Ao, the former decreasing, and 
the latter increasing, as the bore of the tube diminishes ; and 
when the bore becomes extremely small, as in Fig. 4. the posi- 
tive structure is also extremely small, and sometimes can 
scarcely be seen without the aid of a microscope. 
In comparing the values of the tints with their distances 
from the line m pn o, I have found that they vary as the distan- 
ces, in the same manner as in bent glass. Hence it is obvious, 
that the glass is in a state of compression within the black 
ring mpno, and in a state of dilatation without that ring, and 
that the particles are held in a state of violent constraint, en- 
tirely different from that position of equilibrium in which they 
are placed in plates of glass with one or two axes. 
With 
