CLASSIFICATION OF THE ANIMAL KINGDOM. 225 



growth before the olfactory sacs are distinguishable. The regions 

 whence the olfactory nerves will be developed are thus widely 

 separated, and thrown to the ventral and lateral aspect of the 

 head, before the Schneiderian membrane is differentiated. It 

 must also be recollected that, when the naso-frontal process 

 of the embryo appears, the olfactory sacs become connected with 

 one another by a transverse groove, which is persistent in the 

 Bays, and has the same relations as the middle of the olfactory 

 sac of the Marsipobranchii would have if it were supposed to 

 be transversely elongated. 



' Eecent investigations lead me to think that the lower jaw is 

 by no means wanting in the Marsipobranchii, though it presents 

 a very curious modification. In the Ammocoete the hyoidean 

 cleft, which has been overlooked, is present ; and the manner in 

 which the branchial filaments are developed leads me to believe 

 that those which are first formed represent the external gills of 

 the Elasmobranchii, Granoidei, Dipnoi, and Amphibia. 



I have formerly expressed the opinion that the naso-palatine 

 canal of the Marsipobranchii represents the " primitive mouth " 

 of the Vertebrata. The resemblance of the mouth of AmpMoxus 

 to that of an Ascidian renders this comparison questionable ; but, 

 on the other hand, it is a remarkable circumstance that the median 

 nasal involution of AmpMoxus corresponds very nearly, in its rela- 

 tion to the segmented mesoblast, with the oral aperture of an 

 Arthropod or an Annelid ; and it may be that the canal represents 

 the ordinary invertebrate oral passage. 



The dorso-median brain-lobe of the Marsipob ranch appears to 

 me to be represented in the higher Vertebrata by the peduncle 

 of the pineal gland, which in the embryo is a hollow process of the 

 roof of the anterior cerebral vesicle. It is particularly conspi- 

 cuous in young Elasmobranchs. 



In a few Metazoa, as in some small Eotifera and in the Gor- 

 diacese, the alimentary canal never becomes developed, although 

 these animals clearly belong to groups in which the alimentary 

 apparatus is normally formed, and may be safely regarded as 

 modified Gastrese. "Whether the like is true of the Cestoidea, 

 which are so closely allied with the Trematoda, and of the Acan- 

 thocephala, is not certain. Probable as it may be that these are 

 Gastrese with aborted digestive cavities, it may be well to bear in 

 mind the possibility of their never having passed through the 



