342 



POLARISATION. 



i.e., that it may be compared with a simple crystal. If, for 

 instance, we are dealing with small horizontally situated cells 

 or fibres consisting of concentric layers of molecules whose 

 ellipses of elasticity are inclined towards the longitudinal 

 direction of the cells, the portions of the walls situated in the 

 median zone act singly as simple crystals, though we should 

 be compelled to split up the fibres in order to be able to observe 

 these effects separately ; and such a division of the fibres is in 

 many cases impracticable. The same holds good also for some- 

 what thickened membranes which form walls between adjoining 

 cells, and which are therefore to be looked upon as composed 

 of two lamellae. For since the thickenings of these membranes 

 (e.g., spiral threads, obliquely-situated pores, &c.) are, as a rule, 

 un symmetrically placed, which indicates a corresponding situation 

 of the series of molecules, it is also probable that the axes 

 of elasticity of the two lamellae do not lie in the same plane. 



The observer has not unfrequently to deal with two inseparable 

 superposed media, whose axes of elasticity intersect at various angles. 

 He is thus presented with a new problem, which is apparently 

 more complicated than the foregoing one ; and the first question to 

 be decided is whether it is really capable of solution by the same 

 method. We will endeavour to answer this question experimentally. 

 Let A B and CD (Fig. 194) be the ellipses of elasticity of the 

 two superposed bodies, and e the angle between their major axes 

 (assumed to be variable). The focal adjustment should always 

 be so chosen that the bi-sectional line m n of that angle is 

 always at 45 to the crossed Nicols. On this assumption we get 



experimentally the following results for 

 the different values of e : If e is very 

 small the effects of the two bodies increase 

 almost as if they were observed in the 

 position of acceleration, where e ; the 

 resulting colour undergoes no considerable 

 alteration up to an angular magnitude 

 of several degrees. On further increase 

 of e, however, a marked degradation in 

 the interference colour soon follows, 

 FIG. 194 which expresses itself up to a mean 



value of 45 by a distinct change in 

 the tint. From this point, however, the colour remains ap- 



