29 



method. Hardestt attributes the oversight to the incomplete preci- 

 pitation of the silver salt. Kolster^) has, however, seen these fine 

 extensions as well as the larger ones in Golgi preparations of the 

 spinal cord of the salmon. As the radial axes are followed out toward 

 the region of nerve cells and fibers, they branch and become attenuated. 

 Large branches running side by side and others, smaller and irregu- 

 larly disposed, join with the fine filaments in a wide meshed network 

 in which the radial columns are finally lost. This open network can 

 be followed over into the region of nerve cells where it is continuous 

 with the neuroglia mass 

 surrounding them. The 

 continuity of the neu- 

 roglia mass is easily ob- 

 served even here, since 

 the nervous elements 

 are separated by evident 

 intervals. 



Whereas the ar- 

 rangement of the neu- 

 roglia is like that shown 

 by Gierke 2) in Tafel 

 XXI, Fig. 10 of his 

 work, the structure of 

 the columns is different. 

 In this instance the 

 columns are composed 

 of a fibrillated proto- 

 plasm ; not of closely 

 applied fibers. Large 

 leaf-like masses of proto- 

 plasm lying between and 

 connected with the ra- 

 dial axes contain a few 

 irregular fibrils. There 

 is no cell membrane nor differentiated exoplasm in any of the proto- 

 plasmic structures. 



The nuclei of the neuroglia columns are located for the most 



Fig. 2. Neuroglia Syncytium of Batrachus tau. 

 Lateral wall of the third ventricle. X ^'^'^' 



1) KoLSTER, RuD., Studien über das zentrale Nervensystem. L lieber 

 das Rückenmark einiger Teleostier. Berlin, A. Hirschwald, 1898. 



2) Gierke, Hans, Die Stützsubstanz des Zentralnervensystems. Arch. 

 f. mikrosk. Anatomie, Bd. 25, 1885. 



