66 BULLETIN OF THE LABORATORIES 



2d. Our preparations show with perfect clearness in Purkinje's 

 cells a basal process and one or more strong peripheral processes, 

 which latter seem to us to be continuous with a system of transverse 

 fibres lying in the molecular layer. The molecular layer contains 

 sparsely scattered Deiter's ("glia") cells \vhich stain deeply with 

 haematoxylin and cells of the connective tissue system, which include 

 those mentioned by Koelliker. 



3d. The fine branching processes of peripheral and basal pro- 

 cesses of Purkinje's cells are probably not nervous. 



4th. The "basket cells" of Koelliker are almost certainly of 

 connective tissue character. 



5th Haematoxylin staining brings out (juite evidently the large 

 nerve cells of the granular layer, but we suspect that the ramifications 

 described by Koelliker are, in this case also, the result of differentia- 

 tion of the connective-tissue frame work of the cell. 



6th. ?]ranching of the nerve fibres of the white central layer is 

 rare. 



We are driven to reject the results of the Golgi method from the 

 resemblance of the structures produced to those constantly encoun- 

 tered and demonstrably the result of the intricate meshwork of con- 

 nective tissue in the form of hollow sheaths enveloping the cells. 

 Each nerve cell is loosely enveloped in such a sheath, which can be 

 separated l^y projjer treatment, and this sheath is produced in numer- 

 ous directions into tubular or conical projections which may branch 

 and ramify extensively. The same sheath may expand in a nodular 

 manner to envelop another cell or embrace a fibre and may ultimately 

 terminate in a number of brush-like fibrils attached to some membrane 

 or fibrous strand. We believe this explanation is competent to account 

 for many of the complicated structures described by Golgi. 



In the great majority of Piirkinje's cells, for example, we can 

 trace a single strong process (these are numerous in the opossum), per- 

 ipherad nearly to the surface, with little diminution of size, and perfect 

 distinctness and with no lateral l)ranches whatever. Hundreds of such 

 cases are before us where a perfectly distinct uniform branch can be 

 followed to near the surfiice and another of smaller diameter medianly 

 through the granular layer. The nervous protoplasm in the cell and 

 in both fibres is colored similarly. 



Furthermore, no rational interpretation of the structure described 

 by Koelliker seems possible, while the arrangemexit we observe seems 



