CLASSIFICATION OF THE ANIMAL KINGDOiT. 221 



out by the mouth. Nevertheless there are no such canals. The 

 ventro-lateral folds in question begin on each side of the front 

 part of the mouth, and are continued along-side it, as Goodsir 

 rightly states, becoming deeper as they pass back. At the sides 

 of the abdominal pore, they terminate without uniting, one on each 

 side of the'prseanal fin. In the living state, as well as in spirit 

 specimens, these ventro-lateral laminae are strongly curved in- 

 wards ; and they meet, or nearly meet, in the middle line, more or 

 less covering the proper ventral aspect of the body, between the 

 mouth and the respiratory pore. And it is simply the semicanals 

 enclosed by these infolded ventro-lateral laminse which Eathke 

 took for abdominal canals, open only in front and behind. The 

 superficial layer of the integument, with its epiderm, is continued 

 from the outer margin of each ventro-lateral lamina, over its edge, 

 on to the inner surface of the lamina, and, in the normal state, 

 is closely adherent to the greater part of that surface, becoming 

 detached, to be reflected on to the proper ventral face of the 

 body, only at the reentering angle between the ventro-lateral 

 lamina and the ventral face. But, in spirit specimens, this super- 

 ficial layer, which coats the inner face of the ventro-lateral lamina, 

 sometimes becomes detached, along with more or less of its conti- 

 nuation on to the ventral surface of the body, and leaves a wide 

 space, ^which is the abdominal canal described by Stieda, and 

 erroneously supposed by him to be Kathke's canal. The floor of 

 the respiratory chamber is formed by a layer of transversely 

 disposed fibres, chiefly composed of muscular tissue and coated 

 on the dorsal face by a layer of cells, forming part of the epithe- 

 lium of the chamber. In the middle line these fibres are more 

 or less interrupted by the raphe described by Stieda ; the dorsal 

 aspect of the floor is longitudinally grooved in correspondence 

 with the raphe; and, not unfrequently, the epithelial cells dip 

 down into this groove for a greater or less distance. 



On the ventral face of the thick floor of the respiratory cham- 

 ber the superficial layer of the integument is naturally separated 

 by a narrow interspace from the transverse fibres of the floor, ex- 

 cept in the middle line, where it is attached along a depression or 

 groove corresponding with the raphe, like that of the dorsal aspect 

 of the floor. This layer of integument is thrown into regular and 

 close-set longitudinal plaits, which have been described as muscular 

 fibres by Eathke, Miiller, Groodsir, and Quatrefages. Stieda dis- 

 covered the true nature of these longitudinal fibres ; but his 



