Hyatt.] 120 [March 5, 



expected to occur ; and that it is perfectly consistent with the 

 general structure and mode of formation of the neural infolding 

 and of the chorda dorsalis, that none should exist. The extra 

 growth of the ectoblast on the ventral side turns the blastopore 

 dorsally. The first aspect of the medullary groove is an invag- 

 ination leading to the blastopore ; and the folds, according to 

 Kowalevsky, arise on either side and from behind the blastopore 

 forming a part of its border. They then grow forwards extending 

 into a groove which is subsequently closed by the junction of the 

 posterior parts over the blastopore, and the progress of this junc- 

 tion is carried forwards until only a small anterior orifice is left 

 (Arch. Mikr. Anat., vol. 13). 



This looks to us like the formation of a primitive stomodeum 

 by an outgrowth of the circumoral parts; and the elonga- 

 tion of the cup-like stomodeum into a groove could easily be 

 accounted for as due to the need of a respiratory and vortex pro- 

 ducing trap-like vestibule, which might be at first open and then 

 closed, but not necessarily having any anterior connection with 

 the archenteron. Thus the neural canal does not necessarily 

 represent an actinostome at all, but a primitive imagination, 

 becoming a fore and aft tube by the elongation of the body, 

 and the necessary transference of the principal opening to the 

 forward end of the same body. 



Hubrecht's and Lieberktihn's opinion that the hypophysis was 

 the vestibule of a hollow notochord may be used to show that 

 this is the representative of another fore and aft tube, which has 

 arisen from the archenteron. Whether this can be considered as 

 indicating a vermian ancestor for the Vertebrata (Hubrecht), 

 we cannot discuss in detail ; but we are not aware that any inver- 

 tebrate embryos exhibit organs which can be considered as 

 homologous with the medullary plate and groove. Until such 

 characteristics are exhibited, the positive data afforded by the suc- 

 cession of characteristics during the development of Ascidia and 

 Amphioxus are decisive against any supposition that the noto- 

 chord could have occurred in the evolution of the phylum before 

 the medullary plate and fold. The ancestors of the Vertebrata 

 must have acquired their characteristics in some succession, and 

 that succession is plainly recorded in the embryos of the forms 

 mentioned. The notochord, whether primitively a tube or a solid 



