54 Journal of Comparative Neurology. 



species with large optic nerves the development of the genicu- 

 latus is apparently somewhat more prominent than in others. 

 In Cycleptus, for example, it is better developed than in the buf- 

 falo-fish. In Hyodon, it is enormous and there is an imperfect 

 concentric arrangement of the cells. In higher vertebrates, as 

 all reptiles, the geniculate nidulus is entirely cut off from the 

 surface but retains the curious peripheral pseudo-epithelium and 

 is enveloped by a cortical mass of large switch cells which are 

 likewise prominent in birds. The geniculatus itself contains the 

 same plexiform reticulum of blood-vessels so characteristic of the 

 nidulus ruber. The latter is also well-developed in the turtles 

 and has a peripheral zone of pseudo-epithelium of bifid nerve- 

 cells. The fibre connection with both ruber and geniculatus 

 are. the same in reptilia as in fishes. 



The microscopic structure of the nidulus genicuhxtus can be 

 more clearly made out in horizontal sections of the eel, 

 whose plane is oblique laterally, but appro.ximately horizon- 

 tal longitudinally, so that the connection with the fibre tracts 

 caudad can be readily seer;. The optic fibres cephalad have no 

 connection with the nidulus, but are separated by a con- 

 nective partition. The concentration of the brain base in 

 this fish has caused the hypoarium to close about the geniculatum 

 latero-caudad, but a sharp line of demarkation remains within. 

 The epithelium-like arrangement of the marginal cells is very 

 obvious and the cells composing it are much larger than the sim- 

 ilar ones of the hypoarium. They give rise to two processes, 

 one of which passes cephalo-dorsad, the other caudo ventrad. 

 The latter undoubtedly connects with the dorsal peduncular 

 bundle. The interior consists of cells of similar character, 

 though they are frequently tripolar, the third branch dividing 

 into numerous branchlets associated with the lateral twigs of other 

 cells. The suggestion that the ganglion is a switch station, is 

 the first to arise, and if the fibres of the ventral commissure 

 (caudal brachium tract) actually end in the nidulus, the physio 

 logical significance is not difficult to make out. In transverse 

 sections of the same region fibres from the dorso-cephalic region 

 can be followed to their union with one of the processes, while 

 in the meso-ventral portion connection with the ventral commis- 

 sure is scarcely less obvious. 



