OF NITRO-CELLULOSE AND THE LAW OF ITS OPTICAL BEHAVIOUR. 
149 
excellent optical properties and freedom from initial stress. The thickest plate, 
however, is exceptional, as its optical properties are poor. 
Further experimental work on these materials is almost entirely devoted to the 
examination of the optical law of retardation under load, and for convenience all the 
data which follows is expressed as a stress or a strain, the units being pounds and 
inches, in which e is the strain under direct stress f, and the equivalent stress f, is 
obtained from the comparator beam. A typical example of these values is given in 
Table I., for material ^ inch thick, as these measurements are referred to later for 
comparison with values of stress and strain obtained from spectrum observations. 
Table I. 
I-inch. 
f-inch. 
e. 
/• 
A- 
e. 
/• 
u 
0 
0 
73 
0-0123 
3345 
3440 
0-0003 
160 
220 
0-0135 
3505 
3560 
0-0007 
319 
322 
0-0142 
3665 
3900 
0-0013 
478 
440 
0-0153 
3825 
4140 
0-0020 
637 
660 
0-0172 
3980 
4500 
0-0027 
797 
880 
•0-0187 
4140 
4770 
0-0033 
956 
1025 
0-0208 
4300 
5100 
0-0037 
1115 
1173 
0-0212 
4460 
0-0043 
1274 
1320 
0-0237 
4620 
0-0050 
1433 
1495 
0-0287 
4780 
0-0055 
1594 
1642 
0-0320 
4940 
0-0058 
1752 
1760 
0-0370 
5100 
0-0063 
1912 
1910 
0-0553 
5260 
0-0070 
2070 
2050 
0-120 
5420 , 
0-0077 
2230 
2200 
0-157 
5580 
0-0082 
2390 
2350 
0-183 
5740 
0-0087 
2550 
2540 
0-247 
5900 
0-0095 
2710 
2310 
0-280 
6055 
0-0102 
2865 
2860 
0-340 
6215 
0-0108 
3025 
3080 
0-357 
6375 
0-0117 
3185 
3230 
In the earlier experiment on the material xg- inch thick, a fracture was obtained 
near the change of section and before the full extension developed, but still very 
nearly at the full load. It is included here (fig. 7) as, although the later parts o± the 
stress-strain curves are not entirely satisfactory, this does not aifect the problem in 
hand, since the .stress-strain curve is not required very much beyond a pronounced 
yield in the material. 
As a purely mechanical problem, however, there is a considerable amount of interest 
attaching to the accurate measurement of stress and strain over the whole of the 
y 2 
