506 HOWARD. [Vou. XIX. 
have been demonstrated, the deduction would be natural, that 
the outer segments of rods are bodies possessing, not a trans- 
verse, but a longitudinal structure comparable to that of the 
axis cylinder of a single nerve fiber, in which the substance is 
differentiated for longitudinal conduction. 
The question might arise as to whether polarization here 
were due to certain qualities inherent in individual fibers or 
whether it were due simply to the gross grouping together of 
fibers. It seems probable that the former is true and that 
a certain arrangement of substance, perhaps molecular and 
characteristic of organic fibers, accounts for the effect on light. 
A case of double refraction in a glass rod may help to 
illustrate the point. Such a rod ordinarily will not show 
double refraction; 7. e., the glass is isotropic, but if the rod be 
subjected to a stress (pressure or tension) the glass will cause 
polarization of light which passes through it. In the one case 
the rod (or fiber) is not doubly refractive, in the other the 
particles of glass are so disposed by the stress that they effect 
the light differently from what they did before. Tension in 
the direction of the axis will have the effect of determining 
the axis of maximum elasticity as transverse; while pressure 
in the longitudinal direction will change that (optical) axis 
to longitudinal. The mere cylindrical form of the glass rod 
will not produce polarization nor will polarization occur when 
several isotropic glass rods or fibers are placed together; the 
phenomenon depends upon a condition of the particles of glass 
in individual rods or fibers. 
The view (based on comparison of reactions with fibrous 
bodies) that the outer segment of the visual rod is longi- 
tudinally fibrous may be opposed by another hypothesis, which 
is in accord with many observed facts. This view is, that 
the outer segments of the rods are made up of transverse 
plates of medullary substance and that between these plates 
are transverse neurofibrils which connect with longitudinal 
fibrils on the periphery of the rod. This hypothesis is sug- 
gested by a marked similarity in appearance, under polarized 
