VERTEBRATES FROM A ORUSTAOEAN-LIKB ANCESTOR. 401 



(3, p. 244), " Sehr auffallend ist ihre grosse Ahnlichkeit mit 

 den Epithelzellen des centralkanals die sich von ihnen uur 

 dureh die oberflaehliche Lage untersclieideu/' Although he 

 looks upon it as highly probable that they are derived directly 

 from the epithelium surrouudiug the central canal, he con- 

 siders that the proof must be left to embryological investiga- 

 tion. 



Shipley (10) does not deal with this point. Scott (11, 

 p. 274), however, describes how the shape of the spinal cord 

 is at first like that of other Vertebrates with a large central 

 canal, and how a portion of the canal closes up when the cord 

 by its elongation forms its well-known flattened ribbon-like 

 shape. He does not, however, attempt to trace the fate of 

 the cells of this closed-off portion. In connection with this 

 paper of Scott I may mention that he still holds to the belief 

 that the epithelium of the central canal is formed by the 

 epidermic layer of epiblast passing down into the nervous 

 layer, notwithstanding the criticism of Shipley. Naturally, as 

 far as I am concerned, Scott's explanation of the appearances 

 presented by his sections is entirely in accord with my theory, 

 that this epithelium lining represents the original intestinal 

 wall. Although these observations of Scott require extension, 

 they make it possible, if not probable, that the diminution of the 

 central canal gives rise to the formation of groups of outlying 

 elements, derived from its epithelial walls in a similar manner 

 to that noticed by Corning (12), who describes the substantia 

 gelatinosa Rolando as formed by a lateral and nipped-ofi" 

 fold of the original embryonic tube. A still further piece of 

 evidence in favour of considering these cells to be derived from 

 the epithelium of the central canal, i.e. from the walls of the 

 old intestine, is afforded by osmic specimens, which show the 

 presence of fat-globules in the epithelial cells of the central 

 canal, and the extension of those fat-globules in two lateral 

 lines, corresponding exactly to the position of these non- 

 nervous cells. 



Here, then, if anywhere, we find a tissue corresponding to 

 the myelospongium of His (13), and it is possible that in the 



