286 Journal of Comparative Neurology. 



After these statements we understand how broad a line of 

 ' facts ' //if /di' attacks with his publications. The first attack goes 

 against the interpretation of the structure of the cell. On sections 

 of I micron, Held could not convince himself of the fibrillary 

 character of axones or cell-plasma. He corroborated the view of 

 Biitschli. He sees the ' fibrils ' connected by cross-fibrils. A thin 

 section through a honey-comb would produce just the same ap- 

 pearance. In the axone Held sees this axo-spongium (simulat- 

 ing the fibrils) and the neurosomes, small granules embedded 

 in the axo-spongium and especially numerous in the end-plates 

 of the axones. Held has observed in the trapezoid nucleus of 

 the medulla that the a.Kones entering it form a sort of end-plate 

 as they approach the cells. In the cells he distinguished the 

 cytospongium, the neurosomes and the Nissl-granules, and he 

 finds it possible to distinguish the axo-spongium from the cyto- 

 spongium by a difference of stain. In an animal one or two 

 days old, or in the new-born, the axone-ends form a basket-like 

 surface of contact with the cell-body separated from it merely by 

 a homogeneous line probably of ectoplasm of the nerve-cell (or 

 axone-plasin ?). In the animal nine days old, this limiting line 

 cannot be seen. In this and in the adult the basket branches 

 of tl.e axone are quite plainly distinguished from the cytoplasm 

 by the number of neurosomes and the stain ; but in many 

 places it is impossible to deny a concrescence between some of 

 them and the cytoplasm. In a third part of this study (Arch, 

 f. Anat. und Entw. 1897, Supplement Band, p. 273-312, plates 

 XII-XIV), Held establishes the same facts with a more delicate 

 method, insists on the Axencylindet-endfldche being in contact 

 largely with the dendrites, consisting of a real net, not of inter- 

 lacing but anastomosing axone-terminations, even so that ax- 

 ones from several neurones should enter into the mesh-work. 

 He corroborates Beta Mailer's findings in principle at least, the 

 claim of anastomoses of dendrites in the spinal cord of adult 

 teleosts, and the findings of true meshes by Ballowitz in the 

 electric organ of torpedo, but refuses Apathy's net work of 

 nerve-fibrils in worms. Why ? He also examines Golgi speci- 

 mens of the adult with the same result and attributes the gene- 



