500 HOWARD. [Von. XIX. 
in fact, regularity in the breaking is very frequently not 
obvious. 
In addition to the superficial striations visible during one 
to three hours after the removal of the rods from the eye, 
a central core is discernible; this is sometimes of slightly 
different color or shade from the rest, is regular in outline, 
and from one-fourth to one-third the diameter of the rod. 
It is very frequently seen at middle focus. If this appearance 
were the only evidence for the existence of such an internal 
structure, I should, on account of its inconstancy, look for its 
explanation in simple optical effects as many others have 
done. But additional evidence has been found in fresh mate- 
rial in the form of an occasional projection of a central 
portion beyond the broken proximal end of the outer seg- 
ment. In one such case (Pl. 5, Fig. 42, med.) this axis took 
a differential stain in toluidin blue (see pp. 604, 605). . 
To aid in an interpretation of the appearance of the outer 
segments as seen lengthwise in the fresh condition, it is of 
advantage to have them in optical cross-section. This can 
be done by placing a fresh retina, proximal surface down, 
on a slide and employing a support for the cover-giass. 
The outer segments thus viewed (Pl. 2, Fig. 17) are. gen- 
erally circular, but frequently with considerable variation from 
a true circle. The outline appears crenate and in some 
instances nothing more is discernible. In others a distinct 
border can be made out, and this border in favorable light 
is distinctly resolved into small circles of highly refractive 
substance. It is evident, then, that these circles represent the 
highly refractive portions between the dark lines and are fibrils 
(Pl. 2, Fig. 13). The portion between the fibrils, of smaller 
diameter and less refractive, appears as dark lines when in 
focus (Pl. 2, Fig. 13) and as light lines on lower focus (PI. 2, 
Bigs). 
On first inspection one would think the dark lines of Figure 
13 (Pl. 2) were identical with the staining fibril of Figure 1 
(Pl. 1) because of relative sizes, but this seems not to be the 
