154 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY. 



ependyma, but has shown it as lying wholly in the median plane. In 

 P. marinas this ependyma is bilateral, as already described, and nowhere 

 united. Gage ('93, p. 266) first noticed this characteristic ependyma 

 in cyclostomes, and described it as follows: "In lamprey (Fig. 110) 

 similar cells (endyma) are found, but the masses are further separated 

 by the post-commissure." She presents no further description, and offers 

 no suggestion as to function further than to cite the opinion of Rabl- 

 Iliickhard ('87), that similar cells in Amphibia represent the torus, 

 and the suggestion of Edinger ('92), that similar cells in selachians are 

 secretory. 



h. Optic Reflex Cells. Immediately dorsal to the posterior commis- 

 sure there are two groups of large cells symmetrically placed on either 

 side of, and at some little distance from, the median plane. These are 

 probably the cells somewhat vaguely referred to by Mayer ('97). 

 Ji)lii)ston (:02) failed to notice these cells. In another place Johnston 

 (:03, p. 1072) states that the cells corresponding to those of the 

 ' Dachkern ' are not met with in Petromyzon, although Reissner's fibre 

 is present. They lie in immediate proximity to the ends of the horns 

 of the recessus and partially surround them, though the greater number 

 are ventral (Plate 1, Figs. 6, 7). The number of these cells is not great, 

 probably from 8 to 12 in each group. They are conspicuous in size, 

 having a diameter two or three times that of the suri'ounding nerve 

 cells. The nucleus is large, and the cytoplasm takes the iron-hematoxy- 

 lin stain but lightly. These cells are undoubtedly the homologues of 

 the tectal reflex cells which I have described in Amia and other gnatho- 

 stomes. Tliey are multipolar, and some of their neurites pass laterad 

 with the fibres of the posterior commissure into the lateral part of the 

 mesencephalon to the tectum, where it is probable that they are in 

 connection with the terminals of the retinal fibres, as in other forms. 



The axons of these tectal reflex cells are sharply defined non-medul- 

 lated nerve-fibres, which run toward the ependymal thickening sur- 

 rounding the horns of the mesocoelic recess (Fig. 6, trt.tct.ffrr.Reis.). 

 This tliey penetrate, passing between the elongated cylindrical ependymal 

 cells, and into the lumen of the ventricle, into which they may be seen 

 projecting radially (Figs, fi, 7). Within the lumen they unite, form- 

 ing trunks, which continue caudad in the ependymal groove into the 

 recessus and mesocoele. 



There is another source of the axons which furm Reissner's fibre in 

 Petromyzon. Sharply defined unmedullated axis-cylinders come from 

 the lateral anterior portion of the tectum opticum and run toward the 



