TRANSACTIONS OF SECTION G. 585 
3. Note on the Application of Polarised Light to Determine the Con- 
dition of a Body under Stress. By Professors S. P. THompson, 
r.R.S., and E. G. Coxrr, D.Sc. 
The application of polarised light to observe optically the strained condition 
of glass and other transparent materials when subjected to stresses has been 
known since the days of Brewster and Biot. When an ordinarily isotropic body, 
such as glass, is subjected to pressure or tension, it becomes in fact doubly- 
refractive, the axis of double-refraction being along the line of maximum stress ; 
and the presence of such double-refraction is made evident by examining in 
the polariscope the light transmitted through the object. 
Hitherto the usual disposition has been the following: An ordinary polariser, 
such as a large Nicol prism, or a black-glass reflector set at the polarising angle, 
is employed to produce plane-polarised light. An ordinary Nicol prism is used 
as analyser ; and habitually their principal planes are crossed so as to produce 
the dark field. Then the piece of glass to be examined is placed between them, 
and means are provided, by pinching screws or compressors, to subject the glass 
to stresses in a plane normal to the axis of the beam of light through the Nicols. 
No effect is produced if the axis of double-refraction, that is, the axis of the 
strain, lies parallel to the plane of polarisation of either the polariser or the 
analyser. If the axis of double-refraction is at all oblique to such plane then 
more or less light is transmitted. As a result of this disposition the apparent 
amount of light seen through the polariscope varies according to the square of 
the sine of half the angle which the axis of stress makes with the plane of polarisa- 
tion of the polariser ; and the observed figure in the stressed specimen changes 
if it is rotated in its own plane in the instrument. This is further complicated, 
if the light be not monochromatic, by the chromatic differences in the double- 
refraction. Again, as is well known, and as a result of this maximum effect at 
45°, if any object be taken in which the stresses are radial in direction—as is 
the case when a short cylinder or circular disc of glass has been subjected to 
sudden cooling, so that the peripheral portion contracts with great force upon 
the central portion—the object exhibits in the polariscope a black cross with 
its pairs of arms respectively parallel to the planes of polarisation of polariser 
and analyser, the light being a maximum in directions at 45° between the arms 
of the black cross. These arms remain fixed, even though the object be rotated 
in the apparatus. If other objects are used in which these naturally occur, equal 
strains will apparently produce unequal effects, because the different parts 
cannot be so placed that everywhere the axis of strain shall be all at the same 
angle with the principal axis of the apparatus, and there is an optical inequality, 
due to the very arrangement of the polariscope that does not correspond to 
any inequality in the specimen, or in the forces to which it is subjected. 
Furthermore, in order to produce any very evident effects in glass, either 
the specimen must be very thick (as is the case of the pieces of unannealed glass 
generally employed in polariscopic demonstrations) or else the forces to which 
the glass is subject by the compressing screws must be so great as to come 
perilously near to breaking the object. Indeed, it is quite common, in the 
ordinary apparatus for showing these effects, for the little bars or squares of 
glass to be broken or crushed in showing the experiment. 
Two things, therefore, need to be done in improving the means for studying 
and exhibiting optically the stresses in transparent solids: (1) to use a material 
more compressible than glass; (2) to devise optical means for getting rid of 
the black cross, and of making the optical effect independent of the particular 
angle at which the specimen, or any part of it, might be placed. 
The first of these ends was obtained by the adoption of a particular kind 
of xylonite instead of glass. Xylonite is a preparation of nitro-cellulose in com- 
mercial use. It is not quite as transparent as glass, but it is transparent enough, 
even when of a thickness of ten millimetres, to be entirely useful. It is slightly 
tinted and slightly turbid. Sheets three or four millimetres thick are practically 
transparent. From such sheets or selected portions of sheets the objects to be 
experimented upon are cut out. 
The second improvement, which is entirely successful in eliminating the 
