SMITH : EYES OF HJLMONATE GASTEKOrODS. 249 



core appears to lie in a loose sac. But this condition is artificial. I 

 obtained all of my methylen-blue preparations of sensory cells from 

 Limax incidentally, while using methylen blue as a neurofibrillar stain. 

 The less special methods in common use might give better results, but 

 those which I have used are sufficient to serve in identifying the rod as 

 the light-recipient structure, and also to give a needed emphasis to the 

 distinction between the two parts of the rod. 



Although ordinary methods of fixing and staining bring out the struc- 

 ture of the rod, only brief mention is here made of its finer structure, 

 since it is to be described in detail in connection with the neurofibrils of 

 the remainder of the cell. It is sufficient to point out here that the 

 essential feature of the mantle is the fine, radially ai*ranged fibrils, which 

 give it a striated appearance (Plate 1, Fig. 9 ; Plate 2, Fig. 14 ; Plate 

 3, Fig. 21). In Figure 14 the mantle on either side of the core appears 

 to rest upon the distal ends of the pigment cells. But the mantle is in 

 no sense a part of the pigment cells, nor does it result from their secre- 

 tive activity ; it is structurally connected to the core of the rod, into 

 which the fibrils of the mantle can be traced. 



The core is cylindrical or club-shaped and has its greatest diameter at 

 the base, where it has the form of a funnel, suddenly narrowing into a 

 cylinder, a condition which may be understood more easily by an inspec- 

 tion of Figure 14. The core seems to contain little visible material 

 besides the many parallel fibrils which pass into the body of the cell. 



In macerated preparations the rods are found in various states of pres- 

 ervation. The core is always present. Sometimes the mantles are in- 

 tact, although here, again, they are much separated from the cores, 

 giving the impression that the macerating fluid causes them to swell and 

 push off from the cores. The fluid used was not concentrated enough to 

 plasmolize the rods, neither were the cores shrunken. Often the mantle 

 has been more or less completely torn from the core, a condition which 

 is easily explained in view of the swelling of the mantle. In most cases, 

 if any part of the mantle is present, it is still faintly striated (Plate 1, 

 Fig. 9). The swollen condition of the mantle is important, because it 

 may help us to understand better the nature of this structure. In these 

 maceration preparations the core shows a conspicuously granular con- 

 dition. The granules appear to be arranged in longitudinal rows, which, 

 no doubt, are due to the fibrillar structure of the core in the living 

 condition. 



Even in depigmented sections, if the decolorizer has not acted too long, 

 the mantle still shows the characteristic radial striations (Fig. 14). The 



