MUSEUM OF COMPARATIVE ZOOLOGY. 29 



diately below the basement membrane. Here two groups of four fibres, 

 and one of three, can be distinguished. Below this level the groups of 

 three aud four fibres are no longer to be recognized. The fibres diminish 

 rapidly in diameter as they leave the retina. Figure 23 represents a 

 transverse section of the fibres at one fourth the distance from the base- 

 ment membrane to the distal face of the ganglion. Figure 24 represents 

 a similar section midway between retina and ganglion. Figure 25 repre- 

 sents the fibres as they enter the ganglion. The same kind of fibres 

 have been identified in the substance of the ganglion. These fibres, I 

 believe, are the true fibres of the optic nerve, and, as I have shown, 

 they connect the seven proximal retinulaB of each ommatidium with the 

 optic ganglion. 



The termination of the nerve-fibres in the proximal retinulse is so 

 directly opposed to the method of termination which Patten has de- 

 scribed, that, before making a final statement based on a study of the 

 lobster only, it seemed prudent to seek confirmation in other species. 

 This I have done, and I can now state that the termination of the nerve- 

 fibres in the retinulse has been demonstrated in Eupagurus, Cambarus, 

 and Gammarus. From this I conclude that it is highly probably that 

 in the compound eyes of all Crustacea the nerve-fibres terminate in the 

 retinulse. 



This method of nerve-termination, namely, the direct continuity of 

 the nerve-fibre and the perceptive cell, has also been demonstrated by 

 Watase ('89, pp. 34-36) in the eyes of Limulus. 



There are some interesting phases in the transition from the nerve- 

 fibre to the retinula. On examining the transverse sections of nerve- 

 fibres at a level slightly below the basement membrane, one observes 

 that it is in the form of a transparent cylinder the periphery of which 

 has scattered over it a few pigment granules. As the fibre passes 

 through the basement membrane the pigment increases in quantity, and 

 when the retinula is reached the nerve-fibre is represented by a trans- 

 parent axis in its centre (Fig. 19, ax. n.). The nervous axis is transpar- 

 ent, because it contains no pigment granules, while the peripheral portion 

 of the retinula is densely pigmented. In this way the nerve-fibre proper 

 is continuous as a transparent axial shaft from below the basement 

 membrane to a level in the retina, which corresponds to the middle of 

 the rhabdome. From this level to its distal end the retinula is com- 

 pletely filled with pigment, no trace of a transparent central axis being 

 visible. (Compare Figs. 19 and 14 with Figs. 11, 10, and 9.) The trans- 

 parent nervous axis of each retinula terminates, as I have said before, 



